Exploring the Use of Tyrosine Kinase Inhibitors in Glioblastoma

EXPLORING THE USE OF TYROSINE KINASE INHIBITORS IN GLIOBLASTOMA

AND RET-REARRANGED LUNG CANCER TO CURTAIL CELL MIGRATION

AND IMPROVE THERAPY BY UNDERSTANDING DRUG RESISTANCE

SARAH KAY NELSON

B.S., University of Arizona, 2009

A thesis submitted to the

Faculty of the Graduate School of the

University of Colorado in partial fulfillment

of the requirements for the degree of

Doctor of Philosophy

Cancer Biology Program

2017

This thesis for the Doctor of Philosophy degree by

Sarah Kay Nelson

Has been approved for the

Cancer Biology Program

Arthur Gutierrez-Hartmann, Chair

Amy Keating

Lynn Heasley

Rebecca Schweppe

Lee Niswander

Robert Doebele, Advisor

Date: May 19, 2017

Nelson, Sarah Kay (Ph.D., Cancer Biology)

Exploring the Use of Tyrosine Kinase Inhibitors in Glioblastoma and RET-Rearranged

Lung Cancer to Curtail Cell Migration and Invasion and Improve Therapy by

Understanding Resistance

Thesis directed by Associate Professor Robert C. Doebele

ABSTRACT

The exponential increase in targeted tyrosine kinase inhibitor (TKI) therapies approved for use in cancer or in late-stage clinical trials has revolutionized the treatment of solid tumors over the last decade. However, many questions regarding the fundamental biology of oncogenic and wild-type receptor tyrosine kinases (RTKs) and how cancer cells adapt to become resistant to TKI therapies have been raised. This work addresses some of these questions specifically as they relate to glioblastoma (GBM) and RET fusion positive lung adenocarcinoma (LAC).

First, I demonstrate the feasibility of targeting the Mer receptor tyrosine kinase

(MERTK) in order to decrease glioblastoma cell migration and invasion. Using genetic knockdown and the MERTK TKIs foretinib and UNC2025, I demonstrate that MERTK inhibition decreases the migration and invasion of glioblastoma cells in two- and three- dimensional in vitro assay systems. Further, I demonstrate that siRNA knockdown of

MERTK inhibits expression of focal adhesion kinase (FAK), a master regulator of cell migration and invasion.

Second, I demonstrate that mechanisms of acquired resistance to the RET- inhibitor ponatinib RET positive LAC result in reactivation of RAS/MAPK signaling either through the acquisition of oncogenic NRAS p.Q61K signaling or through increased

iii

dependence upon wild-type EGFR and AXL signaling. And third, I investigate the mechanisms through which combination mTOR and RET inhibition is improving patient responses to targeted therapy in RET positive LAC patients and demonstrate that while

RET inhibition partially suppresses signaling through mTOR, maximal inhibition of this signaling pathway is not achieved without the addition of mTOR inhibition with everolimus.

Overall, these pre-clinical studies indicate that clinically available targeted therapies can be used to enhance the inhibition of the glioblastoma cell migration and invasion, as well as to target intrinsic and acquired resistance to targeted RET inhibition in RET fusion positive glioblastoma.

The form and content of this abstract are approved. I recommend its publication.

Approved: Robert C. Doebele

“Nobody in life gets exactly what they thought they were going to get. But, if you work

really hard and you are kind, amazing things will happen.”

Conan O’Brien

I dedicate this to my friend,

Lindsay Marlene Weiss (1987-2009), who fought Hodgkin’s lymphoma with bravery, style, and sass.

ACKNOWLEDGEMENTS

The work presented here is a testament to the love and encouragement of a great number of individuals without whom this journey would not have been possible. Over the last six years, I have grown tremendously both personally and professionally, and to those listed below I am eternally grateful.

My success in graduate school would not have been possible without my mentors,

Amy K. Keating and Robert C. Doebele. I am acutely aware of how fortunate I have been to have such exemplary individuals invested in my training as a clinician-scientist. Under their guidance, I have learned to think critically, embrace collaboration, and persevere at times of adversity, and I am utterly grateful for the time and enthusiasm they have invested in my training.

Next, I would like to thank Arthur Gutierrez-Hartmann—for the scientific guidance he has provided as the chair of my thesis committee and for the unwavering commitment he has made to my training as a physician-scientist. I thank the remaining members of my thesis committee, whose insights and criticisms have immeasurably improved the quality of my work: Lynn Heasley, Rebecca Schweppe, and Lee

Niswander. And I am incredibly grateful for the support of the leadership and students of the Cancer Biology Graduate Program and the Medical Scientist Training Program; in particular, Cancer Biology program director Mary Reiland and former MSTP director

Angeles Ribera.

I would like to thank the colleagues and collaborators whose generous contributions of time, expertise, reagents, and equipment was essential to the completion of this work; especially Aik-Choon Tan, Minjae Yoo, Marileila Varella-Garcia, and John

vii

Heymach. A great deal of thanks is owed to Anh Le for the innumerable ways in which his vast expertise and knowledge has informed my training as a scientist including, but not limited to, teaching me everything I know about cloning and allowing me to continuously solicit his opinions on experimental design and data analysis. The same debt of gratitude is extended to all of my colleagues in the Keating and Doebele Labs for fostering a warm, supportive, and most importantly, fun work environment: Angela

Pierce, Kristina Knubel, Alexandra Sufit, Benjamin Pernu, Anh Le, Adriana Estrada-

Bernal, Aria Vaishnavi, Laura Schubert, and Andie Doak. I would like to thank those who supported my initial graduate work in the Department of Biochemistry at the

University of Colorado, Boulder: Natalie J. Ahn, Kasey Couts, Mary Katherine Tarrant-

Connacher, and Jennifer Liddle. I would be remiss to not thank the many inspiring teachers and mentors I have had over the years who have nurtured and cultivated my love of science and, in particular, would like to thank my undergraduate research advisor

Heddwen Brooks.

I would like to thank the incredibly supportive network of friends I have made during my time thus far in Denver. Especially, my fellow MSTPs Mira Estin, Genevieve

Park, Christopher Knoeckel, Sally Peach, Hannah Scarborough, and Alexandra

Antonioli; medical school classmates Julie Knoeckel, Chloe Hughes, Nathan Lamborn,

Daniel Balk, and Betty Plate; and graduate school classmates Trisha Sippel and Claire

Gustafson. I also extend my gratitude to those whose friendships have sustained me from afar: Sara Button and Kendal Nystedt.

Last, but definitely not least, I would like to thank the members of my extended family. To my aunts, uncles, and cousins in Colorado and Wyoming, thank you for

viii

providing me with a home away from home. To my parents, Bob and Kim Nelson, thank you for teaching me to love science and to never stop asking questions. To my brother and sister in law, Michael Nelson and Kristin Bratton Nelson, you are two of the smartest and snarkiest people I know. And finally, to my fiancé, Daniel Taylor, your love and support over the last three years has immeasurably benefitted the quality of my work and life, and I truly do not know if it would have been possible without you.

TABLE OF CONTENTS

CHAPTER

I. INTRODUCTION AND BACKGROUND ...... 1

Non-Small Cell Lung Cancer 1

Current standard of care 2

Oncogenic drivers of lung adenocarcinoma 3

Receptor Tyrosine Kinases 5

Receptor tyrosine kinases and cancer 7

Gain-of-function mutations 7

Chromosomal rearrangements 8

Amplification and overexpression 8

Autocrine and paracrine stimulation 9

Signaling pathways regulated by receptor tyrosine kinases 9

The mitogen-activated protein kinase (MAPK) signaling pathway 10

The PI3K/AKT signaling pathway 12

Tyrosine kinase inhibitors 15

Resistance to Targeted Therapies 16

Secondary mutations 17

Bypass-signaling 17

Phenotype switching 18

Strategies to overcome TKI resistance 19

The RET Receptor Tyrosine Kinase 20

Ligand-dependent RET signaling 20 x

RET mutations and rearrangements in cancer 22

Epidermal Growth Factor Receptor 25

Ligand-dependent activation and signaling 26

Ectodomain shedding and GPCR transactivation 27

EGFR heterodimerization 28

EGFR and lung cancer 29

The TAM Family of Receptor Tyrosine Kinases 30

AXL and cancer 31

MERTK and cancer 31

Scope of Thesis 32

II. MATERIALS AND METHODS ...... 33

Chapter III Materials and Methods 33

Cell lines and reagents 33

Production of shRNA clones 33

xCELLigence migration assay 34

Neurosphere 3D collagen invasion assay 34

Chapter IV Materials and Methods 35

Cell lines and reagents 35

Derivation of ponatinib-resistant cell lines 35

Cellular proliferation 36

Fluorescence in-situ hybridization 36

Immunoblotting 37

Phospho-arrays 37

NRAS activation assay 37

RNA isolation and Sequencing 37

EMT signature analysis 38

DNA Isolation and Sequencing 38

Retroviral constructs and transduction 39

RNAi-mediated silencing 39

Chapter V Materials and Methods 39

Cell lines and reagents 39

Cellular proliferation 40

Clonogenic growth assays 40

Immunoblotting 40

III. TARGETING MERTK TO INHIBIT CELL MIGRATION IN GLIOBLASTOMA ...... 42

Introduction 42

Glioblastoma: incidence, standard of care, and the need for improved treatment strategies. 42

MERTK is overexpressed in glioblastoma and is a novel therapeutic target for the inhibition of migration. 43

Results 44

Stable knockdown of MERTK decreases glioblastoma cell invasion. 44

Foretinib inhibits glioblastoma cell migration in vitro. 45

UNC2025 inhibits glioblastoma cell migration in vitro. 49

MERTK regulates FAK expression. 50

Discussion 51

xii

IV. ACQUIRED RESISTANCE TO RET INHIBITION IN RET-FUSION POSITIVE NSCLC IS DRIVEN BY REACTIVATION OF THE RAS/MAPK PATHWAY...... 54

Introduction 54

Results 55

Ponatinib inhibits RET signaling and cell growth in the RET- rearranged NSCLC cell line LC-2/ad. 55

Generation of ponatinib-resistant derivatives of the LC-2/ad cell line. 58

MET phosphorylation is increased in the PR1 cell line but does not mediate resistance to ponatinib. 61

Oncogenic NRAS p.Q61K confers resistance to RET-inhibition in the PR1 cell line. 63

NRAS p.Q61K persists as the dominant oncogene in the PR1 cell line in the absence of RET inhibition. 68

PR2 cells have undergone EMT-like changes in mRNA expression, however this does not contribute to the resistance phenotype. 70

Activation of wild-type EGFR and AXL signaling mediates acquired resistance to ponatinib in PR2 cells. 73

In the absence of chronic RET-inhibition, PR2 cells are co-dependent upon EGFR and RET. 81

EGFR and AXL are increased and capable of driving resistance in LC-2/ad cells following RET inhibition 81

Discussion 84

V. PRE-CLINICAL ASSESSMENT OF COMBINATION MTOR AND RET INHIBITORS IN RET FUSION POSITIVE NSCLC ...... 91

Introduction 91

RET inhibitors underperform TKI therapies employed in other fusion-positive LACs. 91

Everolimus. 93

xiii

Combination RET and mTOR inhibition significantly decreases tumor burden in RET fusion positive NSCLC. 93

Results 94

mTOR inhibition decreases proliferation of RET-dependent cancer cells, but does not increase sensitivity to vandetanib. 94

Everolimus enhances the inhibition of colony formation in RET-dependent cells by vandetanib. 96

Addition of everolimus to vandetanib results in maximal inhibition of signaling through RAS/MAPK and the PI3K/AKT/mTOR axis . 98

Discussion 99

VI. DISCUSSION ...... 103

Summary of Findings and Key Conclusions 103

MERTK inhibition is a promising pharmacologic strategy for the inhibition glioblastoma cell migration and invasion 104

Resistance to RET-inhibition in RET-rearranged LAC is mediated by reactivation of RAS/MAPK signaling 104

RET-rearranged cancer cells are sensitive to the mTOR inhibitor everolimus and the addition everolimus to the RET TKI vandetanib results in maximal suppression of RAS/MAPK and PI3K/AKT/mTOR signaling. 107

Future Directions 108

Is MERTK inhibition a viable therapeutic strategy in vivo? 108

How does FAK contribute to MERTK-dependent GBM migration and invasion? 109

By what mechanism are EGFR and AXL activated in the PR2 cell line? 110

Is there an early persister state that emerges following targeted inhibition of RET in RET positive LAC cells? 112

How is mTOR signaling regulated in RET fusion positive LAC? 114

xiv

Significance and Implications of This Work 115

REFERENCES 118

LIST OF FIGURES

FIGURE

1.1 Mutational landscape of lung adenocarcinoma. 4

1.2 General mechanism of receptor tyrosine kinase activation. 6

1.3 The RAS/MAPK signaling pathway. 10

1.4 The PI3K/AKT signaling pathway. 13

1.5 RET signaling. 21

1.6 RET fusions in lung adenocarcinoma. 24

1.7 EGFR signaling. 26

1.8 EGFR transactivation and ectodomain shedding. 28

3.1 Stable knockdown of MERTK decreases glioblastoma cell invasion. 45

3.2 Foretinib decreases transwell migration of glioblastoma cells. 46

3.3 Foretinib inhibits in vitro cell migration of adherent glioblastoma cells in a scratch assay. 47

3.4 Foretinib inhibits glioblastoma cell invasion. 48

3.5 UNC2025 inhibits glioblastoma cell migration in vitro. 49

3.6 MERTK regulates FAK expression. 50

4.1 Ponatinib inhibits RET and decreases cell proliferation of LC-2/ad cells. 56

4.2 PR1 and PR2 cells are ponatinib-resistant derivatives of the LC-2/ad cell line. 57

4.3 PR1 and PR2 cells are resistant to RET TKIs alectinib and cabozantinib. 58

4.4 PR1 and PR2 cells maintain expression of the CCDC6-RET fusion protein. 60

4.5 Downstream signaling in the PR1 and PR2 cell lines is resistant to RET inhibition. 61

4.6 MET phosphorylation is increased in the PR1 cell line, but does not mediate ponatinib resistance. 62

xvi

4.7 Oncogenic NRAS p.Q61K is expressed and drives cell proliferation in the PR1 cell line. 64

4.8 Oncogenic NRAS p.Q61K confers resistance to RET inhibition in the CCDC6-RET expressing TPC1 papillary thyroid cancer cell line. 67

4.9 NRAS pQ61K persists as the dominant oncogene in the PR1 cell line in the absence of chronic RET inhibition. 69

4.10 PR2 cells express changes in mRNA expression similar to EMT, but are not dependent on TGFB signaling. 71

4.11 EGFR regulates downstream signaling and cell proliferation in the PR2 cell line. 76

4.12 PR2 sensitivity to afatinib is not enhanced by RET inhibition and vice versa. 77

4.13 AXL inhibition enhances sensitivity to EGFR inhibition in the PR2 cell line. 79

4.14 AXL inhibition sensitizes HCC78-TAER cells to EGFR inhibition 80

4.15 Removal of PR2 cells from ponatinib partially restores RET-dependence. 82

4.16 LC-2/ad cells are primed to utilize AXL and EGFR signaling when RET is inhibited. 83

5.1 mTOR inhibition decreases cell proliferation in RET-dependent cell lines, but does not sensitize cells to the RET inhibitor vandetanib. 95

5.2 Addition of everolimus to vandetanib enhances inhibition of colony formation in TPC1 cells. 96

5.3 Addition of everolimus to vandetanib results in maximal inhibition of MAPK and mTOR signaling. 97

5.4 Combination vandetanib and everolimus maintains suppression of S6RP compared to vandetanib or everolimus alone. 98

6.1 PR2 cells are more sensitive to JQ1 than LC-2/ad and express increased Fra-1. 113

xvii

ABBREVIATIONS

4E-BP1 eukaryotic translation initiation factor 4E-binding protein 1 ABL Abelson murine leukemia viral oncogene homologue 1 ADAM a disintegrin and metalloproteinase ALK anaplastic lymphoma receptor tyrosine kinase AKT1 protein kinase B ARAF serine/threonine-protein kinase A-raf ARG amphiregulin ARTN artemin ATP adensosine triphosphate AXL tyrosine-protein kinase receptor UFO BAD Bcl-2 associated agonist of cell death BCR breakpoint cluster region protein BIM Bcl-2-like protein 11 BMK big mitogen activated kinase BRAF serine/threonine-protein kinase B-raf BTC betacellulin C carboxy CCDC6 coil coiled domain containing 6 CRAF serine/threonine-protein kinase C-raf DEPTOR DEP domain-containing mTOR-interacting protein EGF epidermal growth factor EGFR epidermal growth factor receptor EGN epigen EML4 echinoderm microtubule-associated protein-like 4 EPR epiregulin ERK extracellular signal-related kinase ETS protein C-ets 1 ETV6 ETS translocation variant 6 FGFR3 fibroblast growth factor receptor 3 FKBP12 FK-binding protein 12 FLT3 Fms-like tyrosine kinase 3 FOS proto-oncogene c-FOS GAB GRB2-associated-binding protein GAP GTPase activating protein GAS6 growth arrest-specific protein 6 GBM glioblastoma GDNF glial cell line-derived neurotrophic factor GDP guanine diphosphate GFLs glial cell line-derived neurotrophic factor family of ligands GFRα glial cell line-derived neurotrophic factor family receptor alpha GOF gain-of-function GPCR G-protein coupled receptor

xviii

GRB2 growth factor receptor-bound protein 2 GRB7 growth factor receptor-bound protein 7 GSK3 glycogen synthase kinase 3 GTP guanine triphosphate HB-EGF heparin-binding EGFR-like growth factor HER2 receptor tyrosine-protein kinase erbB-2 HER3 receptor tyrosine-protein kinase erbB-3 HGF hepatocyte growth factor HRAS GTPase HRas IGF insulin-like growth factor 1 IGF1R insulin-like growth factor 1receptor IRS1 insulin receptor substrate 1 JAK janus kinase JNK c-Jun N-terminal kinase JUN transcription factor AP-1 KDR vascular endothelial growth factor receptor KIF5B kinesin family member 5B KIT mast/stem cell growth factor receptor Kit KRAS GTPase KRas LAC lung adenocarcinoma MAPK mitogen activated protein kinase MEN2 multiple endocrine neoplasia type 2 MEN2A multiple endocrine neoplasia type 2A MEN2B multiple endocrine neoplasia type 2B MET hepatocyte growth factor receptor mLST8 target of rapamycin complex subunit LST8 MST1R macrophage stimulating protein receptor mTOR mammalian target of rapamycin mTORC mammalian target of rapamycin complex MCL1 induced myeloid leukemia cell differentiation protein Mcl-1 MDM2 E3 ubiquitin protein ligase Mdm2 MEK dual specificity mitogen-activated protein kinase kinase 1 MERTK tyrosine-protein kinase Mer MET hepatocyte growth factor receptor MTC medullary thyroid cancer MYC myc proto-oncogene N amino NCOA4 nuclear receptor coactivator 4 NRAS GTPase NRas NRG1 neuregulin-1 NRTN neuturin NSCLC non-small cell lung cancer NTRK1 neurotrophic tyrosine kinase receptor 3 NTRK3 NT-3 growth factor receptor PDGFRA platelet-derived growth factor receptor alpha PDK1 3-phosphoinositide-dependent protein kinase 1

xix

PDL1 programmed cell death 1 ligand 1 PH pleckstrin homology PI3K phosphoinositide 3-kinase PIK3CA phosphatidylinositol (4,5)-bisphosphate 3-kinase catalytic subunit alpha isoform PIP2 phosphatidylinositol (4,5)-bisphosphate PIP3 phosphatidylinositol (3,4,5)-trisphosphate PKC protein kinase C PLC 1-phosphatidylinositol 4,5-bisphosphate phosphodiesterase gamma-1 PRAS40 proline-rich AKT1 substrate 1 PROTOR proline-rich protein PSPN persephin PTB phosphotyrosine binding PTEN phosphatase and tensin homolog RAF Raf kinase RAS Ras GTPase RBD Ras binding domain RET ret proto-oncogene RAPTOR regulatory associated protein of mTOR RICTOR rapamycin-insensitive companion of mTOR RIT1 GTP-binding protein Rit1 RTK receptor tyrosine kinase ROS1 ROS proto-oncogene 1, receptor tyrosine kinase S6K1 ribosomal protein S6 kinase beta-1 S6RP ribosomal protein S6 SCLC small cell lung cancer SH2 Src-homology 2 SH3 Src-homology 3 SRC proto-oncogene tyrosine-protein kinase Src STAT3 signal transducer and activator of transcription 3 SOS son of sevenless TGFα transforming growth factor α TKD tyrosine kinase domain TKI tyrosine kinase inhibitor TRIM33 tripartite motif containing protein 33 TSC2 tuberin TUB tubby protein homolog TYRO3 tyrosine-protein kinase receptor TYRO3 VEGF vascular endothelial growth factor WHO World Health Organization Y tyrosine

CHAPTER I

INTRODUCTION AND BACKGROUND

Across the globe, fourteen million people are diagnosed with cancer and over eight million deaths are attributed to the disease each year.1 In the United States, cancer is the second leading cause of death following heart disease and the number one killer of adults ages 40 to 79.2 The number of cancer deaths each year has been slowly decreasing since peaking in the early nineties likely due to advances in cancer prevention, early detection, and the development of better treatment options. However, despite this progress almost six hundred thousand people in the United States alone will have lost their lives to cancer in 2016.2

Non-Small Cell Lung Cancer

Cancer is an all-encompassing term used to describe over one-hundred different diseases that are characterized by uncontrollable cellular proliferation and distinguished from one another by the organ and type of cell from which they arise. Lung cancer is the leading cause of cancer-related deaths in the United States in both men and women.2

Each year, over two-hundred thousand new cases of lung cancer are diagnosed in the

United States leading to approximately one-hundred and fifty thousand deaths, which is more than the number of deaths attributable to breast, colon, prostate, and pancreatic cancer combined.2

There are two major categories of lung cancer: small cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC). NSCLC accounts for over 85% of lung cancer

diagnoses and is categorized into histopathologic subtypes, the most common of which are squamous cell carcinoma (SCC) and lung adenocarcinoma (LAC).3 SCC is strongly associated with a history of tobacco use or chronic lung inflammation and usually presents in the upper airway whereas LAC, which most commonly arises in the distal airway, is the predominant lung cancer diagnosis made in never-smokers.4-6 Tobacco use is the biggest risk factor associated with lung cancer however, as rates of tobacco use have fallen in recent decades so too have the number of lung cancer diagnoses made in patients with a smoking history.7 Over the same period of time the rate of LAC has slowly risen.8 The natural consequence of these paradoxical trends is a steady increase in the proportion of lung cancer patients who are never-smokers with LAC. In fact, it is currently estimated that over 15% of lung cancers diagnosed in men and 53% of those in women cannot be directly attributed to smoking.9

Current standard of care.

The current standard of care recommended for patients with NSCLC is dependent upon histopathology and disease staging at the time of diagnosis. Surgical resection is the most effective therapy for patients diagnosed with stage I-IIIA lung cancer, and is the only recommended treatment for patients diagnosed with stage IA NSCLC.10 For patients diagnosed with stage II-IIIA, surgical resection is followed by adjuvant platinum-based chemotherapy and radiation if complete surgical resection cannot be obtained.10

However, even with treatment, five-year survival rates range from 67% (IA) to only 23%

(IIIA).11 A diagnosis of stage IV NSCLC is considered terminal and the recommended treatment is platinum-based combination chemotherapy; however, the goal of therapy is to prolong life and not to cure.12 There are two major exception to the above guidelines.

In patients with sensitizing mutations in the epidermal growth factor receptor (EGFR) or anaplastic lymphoma receptor tyrosine kinase (ALK) or ROS proto-oncogene 1, receptor tyrosine kinase (ROS1) gene rearrangements, survival can be significantly prolonged with the addition of a tyrosine kinase inhibitor (TKI) such as afatinib, erlotinib, or gefitinib

(which inhibit EGFR) or crizotinib (which inhibits ALK and ROS1).12 Additionally, pembrolizumab, a programmed cell death 1 ligand 1 (PDL1) inhibitor, can also significantly increase progression-free survival in advanced NSCLC patients with PDL1- positive tumors.13

Oncogenic drivers of lung adenocarcinoma.

At its most basic, cancer is the result of changes in the genome. When these changes result in a gain-of-function the affected gene is termed an oncogene and when they produce a loss of function, a tumor suppressor. Oncogenic drivers are genetic changes that are sufficient to induce oncogenic transformation.

Twelve years ago, mutations in and amplifications of KRAS, which are present in one-third of LAC patients, were the only known oncogenic drivers of LAC. Rapid advancements in next-generation sequencing technologies and improved access to sequencing in the clinic over the interceding decade has resulted in a dramatic expansion of our understanding of the oncogenic drivers of lung cancer. Since 2004, upwards of twenty additional driver oncogenes have been identified in LAC14 (Figure I.1). These include mutations in EGFR,15-17 BRAF,18,19 PIK3CA,20 HRAS,21 and NRAS,21 HER2 and

MET gene amplifications,22,23 and chromosomal rearrangements involving ALK,24-26

ROS1,26,27 RET,24,26,28 NTRK1,29 BRAF,30 NRG1,30,31 AXL,32 and PDGFRA.32

Interestingly, driver mutations in LAC appear to be mutually exclusive, and generally do

Figure I.1: Mutational landscape of lung adenocarcinoma. Summary of known oncogenic drivers in lung adenocarcinoma in 2004 (left) and 2016 (right). Adapted from Hirsch et al. “New and emerging targeted treatments in advanced non-small cell lung cancer”. Lancet (2016). not co-occur within one tumor.21,32,33 While next-generation sequencing of tumors has dramatically expanded our understanding of the molecular underpinnings of LAC, we are still unable to identify oncogenic drivers in nearly one-quarter of LAC patients.21,32,34

Important differences in the genomic landscape of lung adenocarcinoma exist between smokers and never-smokers. The overall mutational burden in smokers with

LAC is much higher than in never-smokers.34 Additionally, KRAS mutations are strongly associated with a smoking history.35-37 Polycyclic aromatic hydrocarbons contained in tobacco smoke are associated with transversion mutations (G>T and G>C), which produce the KRAS p.G12C and KRAS p.G12V mutant proteins.37 When KRAS mutations are found in patients without a smoking history, they are most commonly KRAS p.G12D mutations which result from a transition mutation (G>A), which is not associated with exposure to cigarette smoke.36 Aside from KRAS, all other driver oncogenes, including

EGFR, that have been identified in LAC are more commonly found in patients with a

non-smoking history.21 Consequently, over 40% of smokers with LAC lack an identifiable oncogenic driver, which is in stark contrast to the only 12% of non-smokers for which this is true.38

Receptor Tyrosine Kinases

Receptor tyrosine kinases (RTKs) are a highly conserved family of proteins present among all eukaryotic organisms, having emerged over six-hundred million years ago prior to the evolutionary emergence of multicellularity.39 RTKs are transmembrane receptor proteins that play a critical role in intercellular communication and function by phosphorylating tyrosine residues on intracellular signal transduction molecules in response to extracellular cues.40 In humans, RTK signaling is required for the execution of normal developmental processes and the maintenance of tissue homeostasis.41

In total, there are 58 human receptor tyrosine kinases which are grouped into 20 different families based on the ligands they bind and the structure of their extracellular domains.41 All RTKs share the same basic structure consisting of four major components: a ligand-binding extracellular domain, a single transmembrane helix, a conserved cytoplasmic protein tyrosine kinase domain (TKD), and additional cytoplasmic amino acid residues which function as a regulatory domain.42 In general, RTK signaling is ligand-dependent and initiated upon binding of the cognate ligand to the extracellular domain of the monomeric RTK.43,44 In the absence of ligand binding, the catalytic activity of the TKD is suppressed by intramolecular interactions unique to each RTK45

(Figure I.2a). Ligand-induced receptor activation and subsequent dimerization reverses

Figure I.2: General mechanism of receptor tyrosine kinase activation. (A) In the absence of ligand, RTKs are maintained in a cis-autoinhibited state by the juxtamembrane region (orange). (B) Ligand-binding results in homodimerization and trans-autophosphorylation of tyrosine residues (open circles). (C) Phosphorylation of tyrosine residues in the kinase domain (yellow) and juxtamembrane domain (orange) relieves cis-autoinhibition and results in full activation of the kinase domain allowing for trans-autophosphorylation of tyrosine residues (black circles) that can act as scaffolding sites for adaptor proteins. Adapted from Hubbard, J.S. et al. “Juxtamembrane autoinhibition in receptor tyrosine kinases” Nat. Rev. Mol. Cell. Biol. (2004). the cis-autoinhibition of the TKD, permitting an initial phase of trans- autophosphorylation of tyrosine residues in the activation loop46 (Figure I.2b). This stabilizes the activation loop in an open conformation thus allowing ATP to access the nucleotide binding site such that trans-autophosphorylation of tyrosine residues outside of the kinase domain can occur47 (Figure I.2c). The resulting phosphotyrosines recruit

Src-homology 2 (SH2) and phosphotyrosine binding (PTB) domain-containing proteins, which initiate signaling through downstream signaling cascades. These recruited proteins fall into three categories: (1) catalytic enzymes that are activated by RTK binding, (2) adaptor proteins that contain extra protein-binding domains with which they recruit and interact with additional signaling proteins, and 3) docking proteins localized to the membrane that interact directly with the RTK via a PTB domain or with an adaptor protein previously recruited to the RTK.48

Receptor tyrosine kinases and cancer.

Receptor tyrosine kinases are one of the most overrepresented class of genes mutated in human cancers and mutations in protein kinase domains occur at a rate approximately four times higher than that expected by chance alone.49 This is not surprising considering that aberrant RTK signaling can lead to pathologic activation of signaling pathways that positively regulate cell proliferation and survival. Aberrant activation of RTK signaling in cancer can occur via four distinct mechanisms45: 1) gain- of-function mutations, 2) chromosomal rearrangements that result in the expression of oncogenic fusion kinases, 3) gene amplification and overexpression, and 4) autocrine activation of RTK signaling.

Gain-of-function mutations. RTK gain-of-function (GOF) mutations result in increased catalytic activity of the tyrosine kinase domain in the absence of the cognate ligand. GOF mutations that occur outside of the TKD occur in a wide variety of locations within RTKs and confer kinase domain activation via many different mechanisms. Two such examples are the fibroblast growth factor receptor 3 (FGFR3) and the mast/stem cell growth factor receptor Kit (KIT). Amino acid substitutions in the extracellular and transmembrane domains FGFR3 encoding cysteine residues introduce abnormal disulfide bond formation between two FGFR3 monomers resulting in constitutive dimerization and activation.50,51 In contrast, small deletions in the juxtamembrane domain of KIT relieve cis-autoinhibition of the TKD resulting in increased catalytic activity.52 In and near the

TKD, however, there are residues conserved across the RTK family that are frequently the site of activating mutations. These “hotspots” are primarily located on the activation loop and nucleotide binding region of the kinase domain.53 For example, the EGFR

L858R mutation substitutes a bulky arginine residue into the activation loop, preventing the TKD from adopting the inactive conformation resulting in constitutive EGFR kinase activity.54

Chromosomal rearrangements. The generation of novel kinase fusion proteins occur as the result of chromosomal translocations or rearrangements that couple the kinase domain of an RTK with an unrelated 5’ fusion partner. A wide variety of 5’ fusion partner genes have been documented however, in general, they are defined by two properties: (1) an active promoter region that can drive expression of the fusion kinase, and (2) an oligomerization domain, such as a coiled coil or leucine zipper domain that can produce constitutive ligand-independent kinase activity.55,56 For example, a chromosomal rearrangement recently identified in LAC produces a fusion that combines an N-terminal portion of the echinoderm microtubule-associated protein-like 4 (EML4) gene with the C-terminal intracellular domain, including the TKD, of the gene encoding

ALK.25 Expression of the fusion kinase protein can be attributed to active transcription of the EML4 promoter, and subsequent activation of the ALK TKD in the cytoplasm can be ascribed to a basic region at the very N-terminus of EML4 that facilitates dimerization of the fusion.25 In addition to ALK, chromosomal rearrangements involving the RTKs

NTRK1, RET, and ROS1 have been identified in LAC and collectively, oncogenic fusion kinases are present in over 10% of lung adenocarcinoma patients.56

Amplification and overexpression. Increased RTK expression, whether mediated by genetic amplification or enhanced transcription, can result in ligand-independent receptor activation. Under physiologic conditions, RTK localization in the membrane is tightly regulated in order to prevent aberrant receptor signaling. However, in the context

of oncogenic amplification or transcriptional upregulation, monomers of the overexpressed RTK are locally concentrated in the plasma membrane resulting in ligand- independent homodimerization and abnormally promiscuous heterodimerization.57 The

RTK MET is a prime example as it is amplified or transcriptionally overexpressed in a wide variety of cancers, including LAC.58 When overexpressed in the absence of the cognate ligand, hepatocyte growth factor (HGF), MET dimerization and signaling can be activated by cell adhesion.59 Further, MET amplification can activate receptor tyrosine- protein kinase erB-3 (HER3), an RTK with which it does not traditionally heterodimerize, simply as a consequence of increased physical interaction between the two receptors secondary to the increased concentration of MET in the membrane.60

Autocrine and paracrine stimulation. Increased secretion of RTK ligands can lead to increased ligand-mediated activation of wild-type RTKs. This is particularly important in vivo, where stromal cells in the microenvironment can contribute to the growth factor milieu to which cancer cells are exposed. For instance, cancer-associated fibroblasts can secrete insulin-like growth factor 1(IGF1) which activates IGF1R signaling and promotes the maintenance of plasticity and a stem cell-like phenotype.61 Up-regulation of proteins that regulate secretion of RTK ligands, such as the ADAMs family of metalloproteinases, can also contribute to increased autocrine and paracrine signaling as well as contribute to resistance.62

Signaling pathways regulated by receptor tyrosine kinases.

RTKs regulate a variety of downstream signaling pathways that control a diverse array of cellular functions including cell proliferation, survival, differentiation, motility, and metabolism. The precise downstream signaling program elicited by an individual

RTK is defined by the cytoplasmic tyrosine phosphorylation pattern generated by ligand- mediated receptor activation, as well as the unique set of proteins that are subsequently recruited to these sites.

The mitogen-activated protein kinase (MAPK) signaling pathway. Mitogen- activated protein kinase (MAPK) signaling pathways are activated by extracellular stimuli and comprised of a linear cascade of at least three protein kinases—the last of which phosphorylates regulatory proteins, which elicit changes in cellular functions.63

There are four unique MAPK pathways (ERK, p38, JNK, and BMK) which contribute to

Figure I.3: The RAS/MAPK signaling pathway. Following ligand binding, the GRB2- SOS adaptor complex is recruited to phosphorylated tyrosine residues on the RTK where it is able to catalyze the guanine nucleotide exchange reaction required to catalyze RAS activation. RAS-GTP can then initiate a kinase phosphorylation cascade resulting in the phosphorylation and activation of RAF, which phosphorylates and activates MEK, which can then phosphorylate and activate ERK. ERK can then phosphorylate and activate substrates in the cytoplasm, such as BIM and MCL1, which inhibit apoptosis or translocate to the nucleus where it can phosphorylate and activate transcription factors like ETS, JUN, MYC, and FOS which positively regulate proliferation, survival and cell differentiation. 10

oncogenic RTK signaling, the most characterized of which is the RAS-RAF-MEK-ERK or RAS-MAPK signaling cascade, which will be discussed here.

The first step in canonical RAS-MAPK signaling is activation of the GTPase

RAS. The adaptor protein growth factor receptor-bound protein 2 (GRB2) is recruited to the plasma membrane following receptor activation. There, the GRB2 SH2 domain binds a phosphotyrosine residue either on the RTK64 or on adaptor protein previously recruited and phosphorylated by the activated receptor.65-67 Two GRB2 Src-homology 3 (SH3) domains constitutively interact with a proline-rich region of the guanine nucleotide exchange factor Son of Sevenless homologue (SOS) which is therefore also recruited to the plasma membrane following receptor activation. There, SOS encounters its substrate,

RAS-GDP, tethered to the membrane and catalyzes a nucleotide exchange reaction that produces activated RAS-GTP68-70 (Figure I.3).

Once RAS is activated, it initiates a series of phosphorylation events culminating in the activation of transcription factors, which can modulate cellular functions. First,

RAS binds to the Ras-binding domain (RBD) of RAF (Rapidly Accelerated

Fibrosarcoma) serine/threonine kinase recruiting it to the membrane where it is phosphorylated and activated.71 RAF, a MAPK kinase kinase, then phosphorylates and activates the MAPK kinase, MEK1, which then phosphorylates and activates extracellular related kinase 1 and 2 (ERK1/2)48 (Figure I.3). Activated ERK1/2 can then phosphorylate and activate dozens of signaling molecules and transcription factors in the cytoplasm and nucleus in order to regulate a diverse array of cellular functions including: proliferation, survival, differentiation, apoptosis and stress response.72 Considering the

nature of the cellular functions regulated by RAS/MAPK signaling, it is not surprising that components of this pathway are frequently mutated in cancer.

Activation of RAS signaling whether by mutation, amplification, or overexpression, has long been understood to have transforming properties in cells mediated primarily by increased activation and signaling through the RAF-MEK-ERK pathway. Oncogenic mutations have been identified in all three isoforms of RAS (HRAS,

KRAS, and NRAS) and primarily occur at codons G12, G13, and Q61.73-76 G12 and G13 substitutions decrease the affinity of RAS-GAPs (GTPase activating proteins) which catalyze the hydrolysis of GTP to GDP and substitutions at Q61 interfere with the stabilization of a water molecule that is required for the hydrolysis reaction to occur.77-79

The end result, regardless of mutation, is the stabilization of RAS in its active GTP- bound form and constitutive activation of RAS-dependent signaling.

The PI3K/AKT signaling pathway. Phosphatidylinositol 3-kinases (PI3Ks) are a heterogeneous family of lipid kinases that, like the RAS-MAPK pathway, are activated following RTK activation and tyrosine autophosphorylation. While there are three classes of PI3Ks, the transforming properties of class IA PI3Ks, which will be discussed here, are the most fully characterized.

Class IA PI3Ks are heterodimers composed of a regulatory p85 subunit (which contains an SH2 domain) and a p110 catalytic subunit. Following receptor activation, p85 is recruited to the plasma membrane where it binds a phosphorylated tyrosine on the activated RTK or on an adaptor protein, such as insulin receptor substrate 1 (IRS1)80-82

(Figure I.4). The results in the subsequent activation of the p100 subunit which catalyzes the phosphorylation of phosphatidylinositol (4,5) bisphosphate (PIP2) into

Figure I.4: The PI3K/AKT signaling pathway. Following ligand-binding p85 PI3K can bind directly to phosphorylated tyrosine residues on the RTK or to an adaptor protein like IRS1. This results in activation of the catalytic p110 PI3K subunit which can phosphorylate PIP2 to PIP3, a reaction antagonized by the phosphatase PTEN. PIP3 activates AKT signaling either via PDK1 which phosphorylates thr308 on AKT or via the MTORC2 complex which phosphorylates the ser473 residue. Activated AKT can then promote protein synthesis and metabolism via activation of the MTORC1 complex or promote cell survival by activating MDM2 (which negatively regulates p53) or STAT3 signaling and inhibiting signaling through pro-apoptotic proteins such as GSK3 and BAD.

83 phosphatidylinositol (3,4,5) trisphosphate (PIP3). Alternatively, p100 can also be activated directly by RAS without involvement of p85.84-86 The tumor suppressor PTEN

(phosphatase and TENsin homolog) catalyzes the hydrolysis of PIP3 back to PIP2 and is a key negative regulator of PI3K activity.87-89 Loss of PTEN function, either through loss- of function (LOF) mutations or deletions, is oncogenic and has been observed in a variety of tumor types.90

Once PIP2 is phosphorylated pleckstrin-homology (PH) domain-containing proteins, including the protein serine-threonine kinase AKT and phosphoinositide- dependent kinase 1 (PDK1), are recruited to and bind PIP3. The resulting juxtaposition of

AKT and PDK1 at the membrane facilitates AKT thr308 phosphorylation by PDK1.91

The other major phosphorylation site on AKT, ser473, is phosphorylated by the mammalian target of rapamycin complex 2 (mTORC2); in fact, phosphorylation of AKT at ser473 is actually required to prime AKT for phosphorylation at thr308 by PDK1.92 mTORC2 is a protein complex composed of mammalian target of rapamycin (mTOR),

DEP domain-containing mTOR-interacting protein (DEPTOR), mammalian lethal with

SEC13 protein 8 (mLST8), Pro-rich protein 5 (PROTOR), and rapamycin insensitive

93 companion of mTOR (RICTOR). Once phosphorylated, AKT dissociates from PIP3, disperses throughout the cytoplasm and nucleus, and phosphorylates a multitude of downstream signaling proteins that regulate a variety of cellular functions (cell proliferation, survival, and cell cycle entry)94 (Figure I.4).

One particularly important consequence of AKT signaling is the activation of mammalian target of rapamycin complex 1(mTORC1) signaling. AKT-mediated inhibitory phosphorylation of TSC2 (tuberous sclerosis protein 2) relieves inhibition of the mTORC1 activator RHEB (Ras homologue enriched in brain) resulting in activation of mTORC1.95-97 mTORC1, like mTORC2, is a multi-protein complex comprised of mTOR, RAPTOR (regulatory-associated protein of mTOR), and PRAS40 (40 kDa Pro- rich Akt substrate), as well as DEPTOR and mLST8 (which are also found in mTORC2).98 Whereas the primary substrate of mTORC2 is AKT, mTORC1 phosphorylates a variety of signaling molecules including S6K1 (S6 kinase 1) and eIF4E- binding protein (4E-BP1) which together promote mRNA translation and protein synthesis.99

Tyrosine kinase inhibitors.

Tyrosine kinase inhibitors (TKIs) are rationally designed pharmacologic agents which inhibit the catalytic activity of specific protein tyrosine kinases. When given to a patient with a known mutated or activated RTK, TKIs can rapidly block cancer cell proliferation, induce apoptosis, and produce a dramatic clinical response. The first FDA approved TKI, imatinib, was developed to inhibit the BCR-ABL fusion protein in chronic myelogenous leukemia (CML) has dramatically improved the rate of overall survival in

CML patients.100 In the interceding fifteen years since imantinib’s approval, over 25 additional TKIs have been approved for use in a variety of different malignancies.101 The growing list of oncogenic drivers associated with LAC has revolutionized treatment options for patients in whom a targetable mutation can be detected. Currently, targeted therapies for virtually all known oncogenic drivers of LAC (with the exception of RAS mutations) are clinically available or in phase II or III clinical trials.102

The dramatic responses seen in patients treated with TKIs are due to a phenomenon termed “oncogene addiction:” the idea that the genetic insult responsible for the initial transformation of a cell is also essential to preserve the malignant phenotype of the resultant tumor.103 Therefore, under this paradigm targeting the original mutation or aberration to which cancer cells are “addicted” would be a viable therapeutic strategy.

The biological mechanisms that explain oncogene addiction in cancer cells have yet to be fully elucidated, however the current consensus is that inhibiting an oncogene to which a cell is addicted induces “oncogenic shock.”104 Many oncogenes, in particular RTKs, activate signaling through pathways that are pro-apoptotic in addition to conventional pro-proliferation and pro-survival signaling, however, when RTK signaling is acutely

inhibited pro-apoptotic signals persist longer than those promoting survival.104 For example, in vitro phosphorylation of RAS/MAPK and PI3K/AKT effectors is decreased within thirty minutes of RTK inhibition, however because many pro-apoptotic effectors are activated via feedback mechanisms not immediately interrupted these pro-apoptotic signals can persist for a period of time long enough after oncogene inhibition to commit the cell to an irreversible apoptotic response.105

Resistance to Targeted Therapies

Despite providing meaningful increases in progression-free and overall survival,

TKIs are not curative in patients with NSCLC and resistance is inevitable.106 Resistance to TKI therapy can be intrinsic, and present when the therapy is initiated, or acquired, only emerging after an initial period of clinical tumor response. Failure to achieve a therapeutic kinase-inhibiting concentration of the TKI at the site of the tumor, either due to suboptimal pharmacokinetics or dose-limiting toxicity, is a common mechanism of intrinsic resistance to TKI therapies. Less common is the pre-existence of a mutation in the target RTK rendering the TKI ineffective. For example, a small percentage of patients with somatic EGFR mutations also have a co-existing germline EGFR p.T790M gatekeeper mutation, which is not sensitive to first-generation EGFR TKIs. Mechanisms of acquired resistance generally fall in to one of three broad categories, each of which will be discussed below: (1) secondary mutations in the RTK that nullify TKI activity, (2) the activation of alternate or by-pass signaling mechanisms, or (3) a change in cell phenotype that renders individual cancer cells less dependent on the original driver oncogene.

Secondary mutations.

The vast majority of TKIs in use are ATP-mimetic compounds that bind to the nucleotide-binding pocket in the TKD of the targeted RTK, competitively inhibit ATP binding and, thus, constrain catalytic activity.107 Substitution mutations at the location of a conserved “gatekeeper residue” in the ATP binding pocket can confer resistance either by increasing the binding affinity for ATP or by sterically inhibiting TKI binding

(examples include EGFR T790M108 and RET V804L109), thus restoring constitutive kinase signaling.110 For some, but not all RTKs, mutations at additional residues throughout the TKD can also prevent TKI binding; for instance, crizotinib-resistance mutations in ALK have been identified at more than five distinct sites within the tyrosine kinase domain.111

Bypass-signaling.

Bypass signaling mechanisms restore the intracellular signaling activated by the now-inhibited oncogenic RTK in the absence of target mutations, rendering the TKI incapable of suppressing cellular programs responsible for proliferation and survival.110

Activating mutations, increased expression, or autocrine stimulation of proto-oncogenes as well as deletions or loss of function mutations in tumor suppressors are all possible mechanisms through which bypass signaling can be acquired. These changes can occur in parallel to the TKI target or in series. In the context of EGFR-mutant NSCLC, examples of parallel activation include MET amplification or FGFR1 activation secondary to increased secretion of the ligand FGF2.60,112 The emergence of oncogenic mutant RAS or

RAF proteins, which are downstream effectors of, and thus in series with, oncogenic

RTKs re-activate MAPK signaling and restore cell proliferation.113,114 Loss of function

mutations in tumor suppressor genes can also be responsible for re-activation of downstream signaling as exemplified by the reactivation of PI3K/AKT signaling due to

PTEN deletions—which has also been demonstrated to confer resistance to EGFR

TKIs.115

Phenotype switching.

Within a population of cells at any given point in time, there exists some degree of stochastic heterogeneity and cell-to-cell variations in protein expression. This heterogeneity is not genetically determined, but the result of natural variations in transcription and translation elicited by normal cellular functions.116 Thus, even in a population of genetically identical, clonally-derived cancer cells, a range of cell phenotypes may be observed. The basal level of heterogeneity that exists in a tumor prior to exposure to targeted TKI therapy is likely responsible for equipping a portion of the cells in a tumor to persist despite inhibition of the dominant oncogene.117 These “persister cells” likely play an important role in the mechanisms of acquired resistance described above—buying time before a mutation conferring a secondary mutation or a upregulation of bypass signaling emerges.118 TKI-resistant cancer cell populations that lack these readily identifiable resistance mechanisms are likely to have undergone a phenotype switch.

Phenotype switching is the least well understood of the three acquired resistance mechanisms and derives from the observation that cancer cells emerge resistant to targeted therapies often with a different cellular phenotype.117 This change is not genetically driven but instead, a result of reversible epigenetic and transcriptional changes.119 In epithelial-derived cancers, resistant populations often emerge with a more

mesenchymal phenotype, having undergone an epithelial to mesenchymal transition

(EMT).120,121 In EGFR mutant NSCLC, a small percentage of LAC patients who become resistant to EGFR TKIs relapse with cancer population that has transformed into small- cell lung cancer (SCLC).122 These cancer cells retain the original dominant oncogene, but are no longer “oncogene addicted.”

Strategies to overcome TKI resistance.

Identifying mechanisms of resistance to TKI therapies is critical to improving the efficacy of these drugs and increasing overall survival in patients with oncogene-driven cancers; particularly because many acquired resistance mechanisms are easily targetable with the current arsenal of targeted therapies. The identification of gatekeeper mutations has led to the rational design of second- and third-generation TKIs that can effectively bind RTKs that can overcome resistance due to acquisition of secondary mutations; such is the case with the EGFR TKI osimertinib that works in patients with the T790M gatekeeper mutation.123 Likewise, understanding the bypass signaling pathways likely to be upregulated or mutated in response to a particular targeted inhibitor can inform pharmacologic strategies in patients who develop resistance and lead to up-front combination therapies designed to prevent this type of resistance from emerging. An initial example would be the addition of the MEK inhibitor cobimetinib to the BRAF inhibitor vemurafenib in order to prevent resistance due to MAPK pathway reactivation.124 Given our limited understanding of how phenotype-switching leads to

TKI failure, strategies to overcome this mechanism of resistance remain limited. There is, however, preliminary evidence that EMT-driven TKI resistance can be reversible.

Several groups have shown that inhibiting AXL and/or TGFβ, known drivers of EMT, can restore cells to their original epithelial state as well as reinstate TKI sensitivity.121,125

The RET Receptor Tyrosine Kinase

RET was first identified in 1985 due to its transformative properties following

DNA rearrangement.126 The Ret proto-oncogene (RET) is a RTK normally expressed during development that is required for spermatogenesis, nephrogenesis, and peripheral nervous system differentiation.127 Activating mutations and chromosomal rearrangements of RET are prevalent in thyroid cancer as well as other cancers, including LAC and inactivating mutations are responsible for Hirschsprung’s disease, a developmental syndrome characterized by the absence of normal ganglionic innervation of the colon.128,129

Ligand-dependent RET signaling.

RET signaling is activated by the glial cell line-derived neurotrophic factor

(GDNF) family of ligands (GFLs) which include GDNF, neurturin (NRTN), artemin

(ARTN), and persephin (PSPN).130 Activation of RET signaling differs from that of other

RTKs in that instead of binding to its ligand directly it depends upon a cell surface coreceptor, GDNF family receptor alpha (GFRα), to mediate this interaction. There are four

GFRα receptors (GFRα1, GFRα2, GFRα3, and GFRα4), each of which binds preferentially to a homodimer of one of the GFLs to form a heterodimeric GFRα-GFL

Figure I.5: RET signaling. (A) Ligand-dependent RET signaling is activated when a GFL family ligand, such as GDNF, binds its corresponding GFRα co-receptor, such as GFRα1, promoting dimerization of the GFL-GFRα complex. This heterodimeric complex, in turn, can recruit two in-active RET monomers resulting in homodimerization of RET and trans-autophosphorylation of cytoplasmic tyrosine residues. (B) Summary of cytoplasmic RET phosphotyrosine residues and the downstream signaling pathways they activate. Adapted from Mulligan, L.M “RET Revisited: expanding the oncogenic portfolio” Nat. Rev. Cancer (2014). complex.131-134 This complex can then bind to RET, promote homodimerization, and trans-autophosphorylation of intracellular tyrosine residues135,136 (Figure I.5a). There are two isoforms of RET, RET9 and RET51, which are the result of alternative splicing of 3’ exons. RET9 and RET51 share the first 1063 residues and differ in their C-termini with 9 and 51 unique amino acids, respectively.137,138

Unlike the archetypal model of RTK signaling, in which phosphorylation of activation loop tyrosine residues relieves the cis-autoinhibitory mechanism that prevents access to the nucleotide binding pocket, phosphorylation of Y905 in the activation loop of RET does not result in a conformational change of the TKD and only mildly increases catalytic activity.139 Instead, the activation loop of the RET TKD is in a persistently

active conformation leaving the nucleotide-binding pocket open to bind ATP regardless of receptor phosphorylation status.139 Further, the C-terminal tail of RET also lacks any cis-inhibitory properties and emerging evidence suggests that RET relies on an autoinhibitory mechanism mediated by a glycine-rich loop flanking the kinase domain that regulates optimal RET kinase substrate presentation in trans.139,140

There are eighteen tyrosine residues located in the intracellular domain of RET, fourteen of which have been demonstrated to be phosphorylated.139,141,142 The most important of these being phosphotyrosine 1062 (pY1062) which is required for activation of RAS/MAPK.143-146 RET pY1062 binds SHC1 which, once phosphorylated by the RET

TKD, can activate RAS/MAPK signaling with the recruitment of GRB2 and SOS or

PI3K/AKT signaling with the recruitment of GRB2 and GAB which can bind directly to the p85 regulatory subunit of PI3K143 (Figure I.5b). In vivo, mice with point mutations at

Y1062 exhibit bilateral renal agenesis, thus underscoring the importance of this residue to

RET signaling.147 In addition to pY1062, several other pY residues play a role in the activation of RET signaling. These include pY905 which can recruit GRB7,148 pY928 which can signal to the JAK/STAT pathway via STAT3,149 pY981 which can bind

150 151 SRC, pY1015 which can activate PKC signaling through PLC, and pY1096 (only present in the RET51 isoform) which can also recruit GRB2/GAB in order to activate

PI3K/AKT143 (Figure I.5b).

RET mutations and rearrangements in cancer.

Mutations and chromosomal rearrangements involving RET are prevalent in thyroid cancer. There is a clear distinction between germline and somatic RET mutations in thyroid cancer with regards to clinical presentation: germline RET mutations are

generally point mutations associated with medullary thyroid cancer (MTC) whereas somatic mutations are generally chromosomal rearrangements of the RET gene associated with papillary thyroid cancer.152 Germline RET mutations are responsible for the hereditary cancer syndrome multiple endocrine neoplasia type 2 (MEN2), a family of three autosomal dominant diseases associated with endocrine tumors. There are two subtypes of MEN2, MEN2A, MEN2B, each characterized by early onset medullary thyroid carcinoma (MTC) and mutations in the RET gene.153 There is a distinct distribution of RET mutations associated with each MEN2 subtype. Mutations in the extracellular domain of RET at codons encoding cysteine residues are present in over

95% of patients MEN2A;154 these mutations produce an unpaired cysteine residue on each RET monomer which can form intermolecular di-sulfide bonds which result in the constitutive, ligand-independent formation of activated RET homodimers.155 MEN2B is associated exclusively with mutations near the RET TKD including M918T, which dramatically increases ATP binding affinity of the nucleotide binding pocket.156

Somatically occurring RET mutations are most often chromosomal rearrangements and are associated with papillary thyroid cancer. RET rearrangements are associated with exposure to ionizing radiation and occur at an increased incidence in regions of the world where there is documentation of environmental radiation exposure like Hiroshima, Japan and Chernobyl, Ukraine,157,158 as well as in patients previously treated with radiation.159

While RET fusions are classically associated with papillary thyroid cancer, they were recently identified as oncogenic drivers of approximately 1-2% of lung adenocarcinomas (LAC).26,28 Similar to other fusion driven lung cancers, patients with

Figure I.6: RET fusions in lung adenocarcinoma. The most common RET fusion partners are KIF5B, CCDC6, NCOA4 and TRIM33. The RET TKD is in dark blue and the dimerization domain of the fusion partner (RET) is in orange. Adapted from Mulligan, L.M “RET Revisited: expanding the oncogenic portfolio” Nat. Rev. Cancer (2014).

RET-rearranged LAC tend to be younger, lack a smoking history, and negative for EGFR or KRAS mutations.160 These chromosomal rearrangements pair the C-terminal RET TKD with an unrelated 5’ partner gene which drives expression, dimerization, and constitutive activation of RET signaling. More than two dozen 5’ RET fusion partners have been identified however kinesin family member 5B (KIF5B) coil-coiled domain containing 6

(CCDC6), nuclear receptor coactivator 4 (NCOA4), and tripartite motif containing protein

33 (TRIM33) are the most frequently seen in LAD161 (Figure I.6).

RET TKIs vandetanib and cabozantinib are already FDA approved for use in

MTC patients with disease refractory to radioiodine therapy having demonstrated objective response rates (ORRs) in patients of 45% and 28%, respectively162,163

(summarized in Table 1.1). However, the use of RET TKIs in LAC patients is still under investigation, with the results of phase II studies of vandetanib and cabozantinib in this patient population only recently released.164,165 Vandetanib is a multi-kinase inhibitor that inhibits RET, EGFR, and vascular endothelial growth factor receptor (VEGFR) signaling166,167 and cabozantinib is a multi-kinase inhibitor that inhibits RET, in addition to ROS1, MET, VEGFR2, AXL, TIE2, and KIT.168 Preliminarily, vandetanib appears to be more effective in RET-fusion positive LAC patients having demonstrated an ORR of

47% compared to the 28% ORR achieved with cabozantinib.164,165

RET-fusion positive LAC patients do not appear to benefit as much from currently available TKI therapies as patients with other fusion-driven lung tumors; patients with advanced ALK- or ROS-fusion positive LAD treated with crizotinib were

65% and 72%, respectively.169,170 There are many possible explanations for this discrepancy. First, vandetanib and cabozantinib are both multi-kinase inhibitors repurposed as RET-TKIs and, therefore, achieving therapeutic RET-inhibiting doses in patients is likely limited by side effects mediated by off-target effects. Currently, only one RET-selective TKI is in Phase I clinical trials, RXDX-105.171 However, trials of multi-kinase inhibitors that more potently inhibit RET than cabozantinib and vandetanib may represent better alternatives are currently underway. Second, the basic mechanisms via which RET regulates tumor development, especially in LAC, have been understudied and may potentially present a more complex biological system than other oncogene- addicted cancers. One of the major goals of the work presented here is to add to the current understanding of RET in LAC in order to more accurately anticipate challenges that may prevent the successful use of targeted therapies in this patient population.

Epidermal Growth Factor Receptor

The epidermal growth factor receptor (EGFR; also known as ERBB1 or HER1) is a member of the ErbB family of RTKS, which also includes ERBB2, ERBB3, and

ERBB4. Like most RTKs, physiologic activation of EGFR signaling is important for the regulation of developmental processes, particularly development of the nervous, cardiovascular, and gastrointestinal systems.172

Ligand-dependent activation and signaling.

EGFR can be activated by one of seven known ligands: epidermal growth factor

(EGF), transforming growth factor α (TGFα), betacellulin (BTC), heparin-binding

EGFR-like growth factor (HB-EGF), amphiregulin (ARG), epiregulin (EPR), and epigen

(EGN).173,174 Unlike other RTKs, which rely upon bivalent ligands that effectively tether two RTK monomers together, once monomeric EGFR is bound to its ligand, the dimerization that occurs is mediated purely by receptor-receptor interactions175,176 (Figure

I.7a). Subsequent activation of the kinase domain and autophosphorylation of tyrosine residues provide scaffolding sites for adaptor proteins which mediate downstream signaling through multiple pathways including RAS/MAPK, PI3K/AKT, PKC, and

JAK/STAT signaling64,80,177-181 (Figure I.7b). Several tyrosine residues are

Figure I.7: EGFR signaling. (A) EGFR signaling is activated when a ligand (EGF, TGFa, BTC, HB-EGF, ARG, or EGN) binds promoting dimerization of EGFR monomers and phosphorylation of cytoplasmic tyrosine residues. (B) Summary of the cytoplasmic phosphotyrosine residues on EGFR and the downstream signaling pathways they activate. Adapted from Sebastian S. et al. “The complexity of targeting EGFR signaling in cancer: from expression to turnover” Biochim. Biophys. Acta. (2006). 26

phosphorylated by SRC as opposed to autophosphorylated by EGFR, including Y845 in the activation loop which regulates catalytic activity of the EGFR TKD as well as signaling through signal transducer and activator of transcription 5B (STAT5B).179,182

Tyrosine autophosphorylation at specific residues also serves to diminish EGFR signaling: pY974 mediates receptor internalization via adipocyte protein 2 (AP2), pY1045 promotes degradation via Cbl proto-oncogene (CBL)-mediated ubiquitylation of

EGFR, and pY1173 can bind SH2-domain containing phosphatase 1 (SHP1) which dephosphorylates EGFR.183,184

Ectodomain shedding and GPCR transactivation.

The EGF family of ligands are transmembrane proteins cleaved into their soluble form at the cell membrane in a process termed ectodomain shedding. Ectodomain shedding is a critical regulator of ligand-availability and is mediated by the ADAMs (a disintegrin and metalloproteinase) family of proteins.174 EGF ligands can participate in juxtacrine, autocrine, and paracrine mechanisms of signaling depending on their processing at the membrane. In the absence ectodomain shedding, ligands remain membrane-anchored leaving them only available to bind to receptors on neighboring cells.185 In the presence of ADAMs proteins (primarily ADAM9, ADAM10, ADAM12 and ADAM17) the membrane-anchored ligand is cleaved releasing a soluble ectodomain which can then bind to receptors expressed by the cell from which it was produced

(autocrine signaling) or cells that are further away (paracrine signaling).186 ADAMs proteins are a key mediator of EGFR transactivation by G-protein coupled receptors

(GPCRs).187,188 While the process is not fully understood, GPCR agonists activate

ADAMs protein activity resulting in ligand shedding and the activation of EGFR

Figure I.8: EGFR transactivation and ectodomain shedding. GPCR signaling activates ADAMs family of metalloproteinases which cleave the membrane-anchored pro-form of EGFR ligands. In this example, pro-EGF is cleaved releasing soluble EGF into the cytoplasm where it can bind to and activate EGFR signaling. Adapted from Lappano R. et al. “G protein coupled receptors: novel targets for drug discovery in cancer”. Nat. Rev. Drug Discov. (2011).

signaling189 (Figure I.8). Dysregulation of ADAMs protein expression has been

implicated in many cancer types and may play a role in the upregulation of EGF family

ligands in the tumor microenvironment that leads to autocrine and paracrine activation of

EGFR signaling.190

EGFR heterodimerization.

The ErbB family of RTKs is unique in that in addition to forming homodimers

following ligand binding, heterodimerization with other family members can also occur.

In fact, unique structural properties of ERBB2 and ERBB3 suggest that

heterodimerization with other ErbB family members is a primary mechanism through

which these two receptors signal.191 There is no known ligand for ERBB2 and its

extracellular domain is persistently in an active conformation (which, in the case of

EGFR, ERBB3, and ERBB4 requires ligand-binding) leaving ERBB2 constitutively able

to heterodimerize.192 ERBB3, which lacks a catalytically active kinase domain, relies

upon heterodimerization for trans-phosphorylation and successful activation.193,194 With

each ErbB family member displaying a unique adaptor binding profile, heterodimerization allows a means to diversify the signaling pathways activated following receptor activation.191,195 Further, ligand-dependent activation of downstream signaling can be demonstrably stronger when originating from an ErbB heterodimer as compared to signaling induced by a homodimer.196,197

EGFR heterodimerization is not limited to ErbB family members and many interactions with other RTKs including have been documented.191 For the purposes of this work heterodimerization with RET and AXL are particularly relevant. EGFR can bind to and transactivate RET-fusion proteins, resulting in phosphorylation of RET

Y905;198 and recent work from our lab has shown that this may mediate early adaptive resistance to RET inhibition in LAC.199 Heterodimerization with AXL can diversify ligand-dependent EGFR signaling and mitigate sensitivity to EGFR inhibition in EGFR- driven cancer cell lines.200,201 These heterodimers appear to be particularly important in the context of cancer, as they can mitigate the effect of single-agent EGFR TKI therapies.

EGFR and lung cancer.

Aberrant EGFR signaling contributes to a wide variety of cancer types, particularly those associated with epithelial and mesenchymal lineages.202 EGFR is overexpressed in over 60% of NSCLC tumors and its expression has negative prognostic value.203 Initial clinical trials of EGFR TKIs in unselected NSCLC demonstrated that only 10% of patients experienced a partial response.204,205 The search for predictive biomarkers in this patient population led to the identification of activating mutations which confer constitutive ligand-independent EGFR signaling and were present in the majority of patients who responded to EGFR TKIs.15-17 However, even in patients with

EGFR mutations the efficacy of EGFR TKIs is only temporary, with the majority of patients experiencing relapse within 6-12 months.206 The biological mechanisms responsible for the disparate responses to EGFR TKIs seen in NSCLC patients with mutant and wild-type EGFR have yet to be fully elucidated, however, activating EGFR mutations are more likely to incur an oncogene addiction phenotype and further, these mutations often increase the binding affinity for ATP and ATP mimetics including EGFR

TKIs.207

The TAM Family of Receptor Tyrosine Kinases

The TAM (TYRO3, AXL, and MERTK) family of receptor tyrosine kinases are one of the most recently evolved families of RTKs and the least oncogenic.208 The TAM family play an important physiologic role in the maintenance of macrophage and platelet function. TAM RTKs are activated primarily by growth arrest specific 6 (GAS6) and the vitamin K-dependent Protein S (PROS1), which induce receptor dimerization, kinase domain activation, and trans-autophosphorylation of intracellular tyrosine residues.209

Unlike other RTKs, members of the TAM family are rarely amplified or mutated in human malignancies, although AXL rearrangements have been reported.32,210 AXL rearrangements have While AXL and MERTK are oncogenic, they are rarely oncogenic drivers in solid tumors and they predominantly regulate survival via PI3K/AKT and

STAT3 signaling.211,212 This is likely why the study of the TAM RTKs in the context of cancer has focused primarily on the role they play in innate and acquired resistance to conventional cytotoxic and targeted therapies.

AXL and cancer.

AXL is overexpressed in a panoply of human cancers, including NSCLC, and is generally associated with poor prognosis and chemoresistance;209 for this reason, AXL appears to be an attractive pharmacologic target. AXL is a transcriptional target of the

FOS family member transcription factor Fos-related antigen 1 (Fra-1), which is regulated by ERK and contributes to the malignant phenotype observed oncogene-driven cancers with pathologic activation of RAS/MAPK signaling.213 AXL is a key regulator of the epithelial to mesenchymal transition (EMT), a process by which epithelial cells switch to a more mesenchymal phenotype, which is associated with increased migration, metastasis, and therapeutic resistance. Stimulation of AXL signaling increases expression of TWIST, SNAIL, and SLUG, canonical transcriptional regulators of EMT.121,214,215 As mentioned previously, AXL can also heterodimerize with EGFR and diversify signaling elicited by EGFR ligands and has been shown to mitigate sensitivity to EGFR TKIs.200

AXL-EGFR crosstalk has also been implicated in acquired resistance to PI3K inhibitors.216

MERTK and cancer.

Like AXL, MERTK is upregulated and associated with a poor prognosis in a variety of cancer types, including NSCLC.209 In particular, MERTK appears to confer chemoresistance and an increased migratory phenotype to cancer cells in which it is expressed.217,218 Physiologically, MERTK regulates many functions that require cytoskeletal reorganization including, the engulfment of apoptotic material by macrophages;219 therefore it is not surprising that MERTK has been found to regulate cell

migration and metastasis, functions which also require cytoskeletal reorganization, in several solid tumor types including NSCLC, glioblastoma, and melanoma.

Scope of Thesis

The exponential increase in targeted TKI therapies approved for use in cancer patients or in late-stage clinical trials has revolutionized the treatment of solid tumors over the last decade. However, in the wake of this expansive paradigm shift many questions regarding the fundamental biology of oncogenic and wild-type RTKs and how cancer cells adapt to become resistant to TKI therapies have been raised. The work presented here addresses some of these questions. First, in Chapter III, I will pivot to the work I performed under Dr. Amy Keating exploring the role of wild-type TAM RTKs in the regulation of glioblastoma cell migration. Next, in Chapter IV and Chapter V, I will address the use of targeted inhibitors in RET-fusion positive LAC. Specifically, in

Chapter IV, I will demonstrate that the TKI ponatinib is a potent inhibitor of RET- fusions in a patient-derived RET-fusion positive LAC model as well as describe the development of two, distinct ponatinib-resistant LAC cell lines dependent on mechanisms of acquired resistance that re-activate RAS/MAPK signaling. In Chapter V,

I will explore the molecular effects of combination mTOR and RET inhibitor treatment in patient-derived RET-fusion driven cancer cell lines. Taken together, these experiments highlight the complexities of RTK signaling in oncogene-driven cancers and provide insight into therapeutic strategies that may improve overall survival in these patient populations.

CHAPTER II

MATERIALS AND METHODS

Chapter III Materials and Methods

Cell lines and reagents.

The U251 and A172 cell lines were obtained from American Type Culture

Collection (Manassas, VA, U.S.A) while the SF188 cell line was obtained from the

UCSF Brain Tumor Bank; all were maintained in DMEM + 10% fetal bovine serum per culture guidelines, unless otherwise indicated. Cell lines were authenticated by DNA fingerprinting bi-annually through the CU Cancer Center DNA Sequencing and Analysis

Core Facility.220 All antibodies were obtained from Cell Signaling Technology (Danvers,

MA, U.S.A) unless otherwise noted. Foretinib was kindly provided by GlaxoSmithKline

(Brentford, Middlesex, UK) via MTA. Stock solutions were prepared in dimethyl sulfoxide (DMSO) and stored in aliquots at -20°C. All reagents were obtained from Life

Technologies, unless otherwise noted.

Production of shRNA clones.

Lentiviral vectors (pLKO.1) containing shRNA sequences targeting MerTK

(shMerTK, oligo ID: TRCN0000000862) or non-silencing control green fluorescent protein (GFP; shGFP) were obtained from Open Biosystems. Replication incompetent viral particles were generated using the 293FT cell line and the third-generation packaging system (two packaging plasmids and one envelope plasmid) developed by the laboratory of Dr. Didier Trono. Puromycin-resistant colonies were typically observed on days 9 to 13. Stable, clonal lines were developed from heterogeneous MerTK shRNA 33

knockdown cell populations by single-cell flow cytometry sorting for low MerTK expression. xCELLigence migration assay.

The migration assays were conducted as previously described.221 Briefly, adherent cells at 70% confluency were collected using 0.02% EDTA and resuspended at a

4 concentration of 3 x 10 cells/50 L. Cells were plated in triplicate with media in the

CIM-plate 16 (ACEA Biosciences, San Diego, CA) and allowed to adhere to the plates free of therapeutics for 1 h. Cells were treated with either DMSO or serial dilution of foretinib in DMEM + 10% FBS (final concentrations of 900 nM, 300 nM, and 100 nM).

Migration was quantified every 10 min for 20 h. Three independent experiments were performed.

Neurosphere 3D collagen invasion assay.

1 x 104 cells were cultured in Neurobasal Medium A supplemented with B27, 0.5 mM L-glutamine, 20 ng/mL hEGF (Sigma-Aldrich, St. Louis, MO), 20 ng/mL bFGF

(Sigma-Aldrich) in ultra-low adherent 24-well plates (Corning Life Sciences, Corning,

222 NY) for 24 hours to allow for formation of neurospheres. Neurospheres of 75-150 m were resuspended in type I collagen (2.17 mg.mL)/DMEM + 10% FBS and appropriate treatments and plated in a 96-well plate pre-coated with type I collagen (Advanced

BioMatrix, San Diego, CA). Neurospheres were imaged immediately (Time 0) with an

Olympus CKX41 fitted with a Qicam Fast 1394 camera and then again at 15 and 24 h.

Absolute maximal radius of each neurosphere was measured using ImageJ software and

this distance was used to determine the relative change in radius. Three independent experiments were performed.

Chapter IV Materials and Methods

Cell lines and reagents.

LC-2/ad cells were obtained from Sigma (cat no. 94072247), TPC1 cells obtained from R.E. Schweppe 223; H2228 cells obtained from J.D. Minna. HCC78-TAER were previously described 224. Cells were maintained in RPMI-1640 (Invitrogen) with 10%

FBS at 37°C in a humidified 5% CO2 incubator. Fingerprint analysis of cell lines was performed by the Molecular Biology Service Center at the Barbara Davis Center for

Diabetes at the University of Colorado Anschutz Medical Campus in Aurora, CO to ensure authenticity. Alectinib was provided by Chugai Pharmaceuticals. Ponatinib, cabozantinib, crizotinib, trametinib, gefitinib, afatinib, and foretinib were obtained from

Selleck Chemicals. SB525334 was obtained from L.E. Heasley. Antibodies used were as follows: pEGFR Y1068 (D7A5), pEGFR Y1173 (53A5), total RET (D3D8R), pMET, total MET, pERK1/2 XP T202/Y204 (D13.14.4E), total ERK1/2 (L34F12), pAKT S473

XP (D9E), total AKT (40D4), and pSHC1 Y239/Y240 (2434) from Cell Signaling; pTYR

(4G10 Platinum), GAPDH (6C5) and GAPDH (ABS16) from Millipore; α-tubulin (TU-

02) and NRAS (F155) from Santa Cruz Biotechnology.

Derivation of ponatinib-resistant cell lines.

The PR1 cell line was derived by initially culturing LC-2/ad cells in a continuous dose of 200 nM ponatinib. Once cells were able to proliferate normally at 200 nM, the dose was increased to 400 nM ponatinib. Once cells were able to proliferate normally in 35

400 nM ponatinib. The PR2 cell line was derived by continuous culture in ponatinib gradually increasing from an initial dose of 5 nM until cells could proliferate normally in

400 nM ponatinib. Once established, the PR1 and PR2 cell lines were maintained in 200 nM ponatinib. To generate PR1-OP and PR2-OP cells, PR1 and PR2 cells were removed from ponatinib for 6 weeks.

Cellular proliferation.

Cells were plated in 96-well tissue culture plates (2,500 cells per well for LC-

2/ad, PR1, PR2, PR1-OP, PR2-OP, H2228, and HCC78-TAER cells; 1,000 cells per well for TPC1) and removed from ponatinib, if indicated, 24 hours prior to drug treatment or siRNA transfection for the time periods indicated. Cell numbers were assessed using

CyQUANT Direct Cell Proliferation Assay (Thermo Scientific) according to the manufacturer’s instructions or using 3-(4,5-dimethylthiazol-2-yl)-5-(3- carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium) assays as previously described.225

Fluorescence in-situ hybridization.

FISH assays and analyses were conducted as described previously with minor modifications 226. The RET break-apart probe set includes a 3’RET (Spectrum Red [R]) probe recognizing a genomic region 3’ end of exon 8, and a 5’RET (Spectrum Green[G]) probe recognizing a genomic region 5’ end of exon 12. Samples were positive for RET- rearrangement if more than 15% of cells showed split signals, 3’RET and 5’RET apart by

≥2x the signal diameter.

Immunoblotting.

Immunoblotting was performed as previously described 225. Briefly, cells were lysed in modified radioimmunoprecipitation assay (RIPA) buffer supplemented with Halt

Protease and Phosphatase Inhibitor Cocktail (Thermo Scientific) and diluted in 4X

Protein Sample Loading Buffer (LI-COR). Total protein was separated by SDS-PAGE, transferred to nitrocellulose, and stained with the indicated primary antibodies and IRDye anti-mouse or anti-rabbit IgG (LI-COR). Membranes were scanned with an Odyssey

Imager and Odyssey Image Studio Software (LI-COR). When appropriate, membranes were stripped with 3X NewBlot Nitro Stripping buffer (LI-COR) and re-probed with additional primary antibodies.

Phospho-arrays.

PathScan RTK Signaling Antibody Arrays were obtained from Cell Signaling

Signaling and performed per the manufacturer’s instructions.

NRAS activation assay.

Ras Pull-down Activation Assay Biochem Kit was obtained from Cytoskeleton,

Inc. and performed as per the manufacturer’s instructions, with the exception that the

Raf-RBD/GTP-Ras complexes pulled down were probed with an NRAS-specific antibody.

RNA isolation and sequencing.

RNA isolation from cells was performed using the RNeasy Kit (Qiagen) per the manufacturer’s instructions. High-throughput mRNA sequencing (RNA-seq) of each sample (three replicates per cell line) was performed as previously described 224.

Sequencing was performed on the Illumina HiSeq2000. On average, 50M (45M – 59M) single-end 126 bp sequencing reads were obtained per sample. Reads were mapped against the human genome using Tophat (version 2.0.13)227, NCBI reference annotation

(build 37.2) was used as a guide, allowing 3 mismatches for the initial alignment and 2 mismatches per segment with 25 bp segments. On average, 43M (37M-49M) of the reads aligned to the human genome. Transcripts were assembled using Cufflinks (version v2.2.1)228 using the RefSeq annotation as the guide, but allowing for novel isoform discovery in each sample. The data were fragment bias corrected, multi-read corrected, and normalized by the total number of reads. Differentially expressed genes were identified by Cuffdiff (version v2.2.1) after merging the transcript assemblies. Samtools

(version 1.2)229,230 was used to convert sequence alignment map (SAM) files to binary alignment map (BAM) files, and variant calling was performed for each sample (a variant is defined as minimum 5 reads detected in 20 reads depth). Integrative Genomics Viewer

(IGV) was used to visualize the aligned reads and variants.231

EMT signature analysis.

Normalized gene expression of genes associated with EMT in LC-2/ad and PR2 cells was calculated and filtered for those significantly changed in PR2 cells as determined by a Student’s T-Test (P<0.05) and ranked according to fold change.

DNA isolation and sequencing.

Mutations identified by mRNA sequencing were confirmed using standard Sanger sequencing of genomic as previously described 224.

Retroviral constructs and transduction.

TPC1 cells were stably transduced with retroviral particles containing pBabe N-

Ras 61K, a gift from Channing Der (Addgene plasmid # 12543), or an empty vector control and subject to puromycin selection.

RNAi-mediated silencing.

Cells were transfected with 20 nM siRNA targeting RET (cat no 4392420,

Ambion), NRAS (L-003919-00, GE Dharmacon) or a non-targeting control (SIC001,

Sigma) using DharmaFECT 1 transfection reagent (T-2001, GE Dharmacon).

Chapter V Materials and Methods

Cell lines and reagents.

LC-2/ad cells were obtained from Sigma (cat no. 94072247), TPC1 cells obtained from R.E. Schweppe.223 Cells were maintained in RPMI-1640 (Invitrogen) with 10%

FBS at 37°C in a humidified 5% CO2 incubator. Fingerprint analysis of cell lines was performed by the Molecular Biology Service Center at the Barbara Davis Center for

Diabetes at the University of Colorado Anschutz Medical Campus in Aurora, CO to ensure authenticity. Vandetanib and everolimus were obtained from J.V. Heymach.

AZD8055 was obtained from L.E. Heasley. Antibodies used were as follows: total RET

(D3D8R), pERK1/2 XP T202/Y204 (D13.14.4E), total ERK1/2 (L34F12), pAKT S473

XP (D9E), total AKT (40D4), pS6K1, total S6K1, pRPS6, total RPS6 from Cell

Signaling; pTYR (4G10 Platinum), GAPDH (6C5) and GAPDH (ABS16) from

Millipore.

Cellular proliferation.

Cells were plated in 96-well tissue culture plates (2,500 cells per well for LC-

2/ad; 1,000 cells per well for TPC1) 24 hours prior to drug treatment. Cell numbers were assessed using CyQUANT Direct Cell Proliferation Assay (Thermo Scientific) according to the manufacturer’s instructions or using 3-(4,5-dimethylthiazol-2-yl)-5-(3- carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium) assays as previously described.225

Clonogenic growth assays.

TPC1 cells were seeded in 6-well plates at 100 cells/well. After 24 hours, media was changed and drug was added. Drug-containing media was subsequently replaced every 72 hours. After 14 days, cells were fixed with 7% (vol/vol) glacial acetic acid and stained with 6% (vol/vol) glutaraldehyde and 0.5% (wt/vol) crystal violet solution for 15 minutes at room temperature. Plates were rinsed with water, allowed to dry, and then photographed. Images were quantified with the MetaMorph software program (Molecular

Devices, Downington, PA) to quantify the total area of the colonies in each well.

Immunoblotting.

Immunoblotting was performed as previously described 225. Briefly, cells were lysed in modified radioimmunoprecipitation assay (RIPA) buffer supplemented with Halt

Protease and Phosphatase Inhibitor Cocktail (Thermo Scientific) and diluted in 4X

Imager and Odyssey Image Studio Software (LI-COR). When appropriate, membranes

were stripped with 3X NewBlot Nitro Stripping buffer (LI-COR) and re-probed with additional primary antibodies.

CHAPTER III

TARGETING MERTK TO INHIBIT CELL MIGRATION IN GLIOBLASTOMA1

Introduction

Glioblastoma: incidence, standard of care, and the need for improved treatment strategies.

Glioblastoma (GBM) is the most common primary tumor of the central nervous system, accounting for over 15% of brain tumors.232 Even with aggressive treatment the median survival is less than 15 months, with tumor recurrence occurring in over 90% of patients and less than 5% of patients surviving longer than five years.233 According to the

World Health Organization’s (WHO) classification system for tumors of the central nervous system, GBM is a grade IV glioma which can result from disease progression of a lower-grade astrocytoma, termed secondary GBM, or develop de novo with no history of malignancy prior to diagnosis, termed primary GBM.234 Primary GBM accounts for over 95% of cases and likely reflects a distinct malignant process compared to secondary

GBM.233 The current standard of care for GBM is total surgical resection, temozolomide chemotherapy, and radiation.233 Total surgical resection is usually impossible due to the highly infiltrative nature of glioblastoma, and the profound migratory phenotype of GBM cells is a hallmark of the disease. GBM tumors most often reoccur within 2-3 cm of the

1Figure 5.2 and Figure 5.4 are included in the following publication and are reprinted here with permission: Knubel, K. H. et al. “MerTK inhibition is a novel therapeutic approach for glioblastoma multiforme.” Oncotarget (2014). 42

primary resection cavity, however it is not uncommon for recurrent disease to be located in the contralateral hemisphere.235

New pharmacologic agents and therapies are clearly needed to GBM treatment and to extend patient survival. Unlike NSCLC, GBM has yet to benefit from the surge in development of targeted molecular therapies with many agents demonstrating potential in pre-clinical studies only to fail in patients. A particularly pointed example of this being the anti-angiogenic VEGF inhibitor bevacizumab. Bevacizumab did not increase overall survival in GBM patients, and it is now widely accepted that the reduced blood supply to the tumor following VEGF inhibition increases tumor cell invasion into the brain parenchyma.236 This development has highlighted the important role that cell migration and invasion plays in the pathogenesis of glioblastoma, and the need for therapies that specifically inhibit tumor invasion.

MERTK is overexpressed in glioblastoma and is a novel therapeutic target for the inhibition of migration.

MERTK is a member of the TAM (TYRO3, AXL, and MERTK) family of RTKs that is expressed normally in macrophages and platelets, where it contributes to phagocytosis of apoptotic cells and platelet aggregation.209 Upon ligand binding

(MERTK ligands include GAS6, PROS1, tubby protein homolog (TUB), tubby-related protein 1 (TULP1), and galectin-3) MERTK homodimerizes and autophosphorylates cytoplasmic tyrosine residues which bind and recruit effector proteins that elicit activation of anti-apoptotic and pro-survival pathways, including PI3K/AKT and

RAS/MAPK.209 MERTK regulates cancer cell survival and chemoresistance in a variety of human malignancies, including GBM.221,237,238 In addition to the anti-apoptotic and

pro-survival phenotype elicited by MERTK signaling, the Keating lab has demonstrated that MERTK signaling is required to maintain normal cell migration in GBM cells grown in vitro.221 Further, MERTK inhibition alters FAK and Rho GTPase signaling and induces morphological changes in the actin cytoskeleton of glioblastoma cells.221 In agreement with this finding, MERTK has since been shown to similarly regulate cell migration in melanoma.217

The highly infiltrative nature of glioblastoma is currently the primary obstacle to successful patient outcomes from this devastating disease. MERTK inhibition has been identified as a potential avenue through which GBM cell migration and invasion can be targeted. The goal of the work presented here is to assess the translational relevance of

MERTK inhibition in glioblastoma with small molecule TKIs.

Results

Stable knockdown of MERTK decreases glioblastoma cell invasion.

Previous studies from the Keating lab demonstrated that stable knockdown of

MERTK in glioblastoma cells decreases transwell migration, in a manner that can be reversed by re-expression of MERTK.221 In order to assess whether three-dimensional cell invasion was also impaired in these cells, I optimized a three-dimensional glioblastoma neurosphere invasion assay. Parental U251 cells, shGFP, and shMERTK cells were induced to form neurospheres, plated in collagen in 96-well plates, and incubated at 37C for 24 hours. Migration of cells from the core of individual neurospheres were tracked and imaged at 15 and 24 hours after plating in collagen

(Figure III.1). The absolute maximal radius of each neurosphere (core and furthest

reaching cell) was measured and normalized to the radius at the time of plating.

Foretinib inhibits glioblastoma cell migration in vitro.

Foretinib is a multi-kinase inhibitor designed to inhibit the RTKs hepatocyte

growth factor receptor (MET) and vascular endothelial growth factor receptor 2 (KDR)

RTK families, which has broad inhibitory activity against many RTKs at low-nanomolar

concentrations.239 Foretinib is currently being assessed clinically in a variety of cancer

types, with preliminary efficacy demonstrated in hepatocellular carcinoma and papillary

renal cell carcinoma.240,241 Foretinib potently inhibits MERTK at low-nanomolar A. B.

0 h 24 h

350 WT 300 shGFP 250 shMERTK 200 150 100

50 shGFP

Change in radius (uM) radius in Change 0 0 3 6 9 12 15 18 21 24 Time (hours)

shMERTK

Figure III.1: Stable knockdown of MERTK decreases glioblastoma cell invasion. Parental U251 cells (WT) and U251 cells constitutively expressing shRNA targeting GFP (shGFP) or MERTK (shMERTK) were cultured in neurosphere media for 24 hours, embedded in a type I collagen matrix and imaged immediately (t = 0) and at 15 and 24 hours. Absolute maximal radius of each neurosphere was measured using ImageJ software. The change in radius was calculated by subtracting the absolute maximal radius of t=0 from t=15 and t=24. Data are the average of three independent experiments (A) or representative images of t=0 and t=24 from three independent experiments with green circles delineating the absolute maximal radius of invasion. Error bars are SEM.

A. B. C. U251 SF188 A172 100 100 100

50 50 50

% Cell Migration % Cell Migration % Cell Migration % Cell

DMSO 100 300900 DMSO 100 300900 DMSO 100 300900 foretinib (nM) foretinib (nM) foretinib (nM) Figure III.2: Foretinib decreases transwell migration of glioblastoma cells. (A) U251, (B) SF188, and (C) A172 cells were plated in triplicate in xCELLigence CIM-Plate transwell cultures with 8 m pores and treated with DMSO or foretinib at the indicated doses. After 10 hours migration was measured as electrical impedance at the point of the transwell and normalized to the DMSO control. Data are the average of three independent experiments. Error bars are standard deviations. Statistics are repeated measurement ANOVA with a Dunnet’s multiple comparison test (*P<0.05; ***P<0.001).

concentrations and at lower concentrations than it inhibits other TAM family members,

TYRO3 and AXL, or MET.242

Previous studies from the Keating lab demonstrated that stable knockdown of

MERTK in glioblastoma cells decreases transwell migration, in a manner that can be

reversed by re-expression of MERTK.221 Therefore, I hypothesized that pharmacologic

inhibition of MERTK with foretinib would have the same effect. Using the Xcelligence

real-time cell analysis system, cellular transwell migration of U251, A172, and SF188

glioblastoma cells was measured after 14 hours in the presence of increasing doses of

foretinib. Foretinib decreased cell migration in all three cell lines with maximal inhibition

of U251 cell migration achieved at 100 nM while both the A172 and SF188 cells

exhibiting a dose-dependent decrease in cell migration (Figure III.2A-C). Foretinib also

inhibited U251 and A172 cell migration in a dose-dependent manner in an in vitro wound

healing assay. (Figure III.3).

The translational relevance of the cell migration assays described above is limited given that cell migration in vivo takes place in a three-dimensional environment not

A. U251 A172 0h 24h 0h 24h

DMSO

100 nM

300 nM

900 nM

B. 40 U251 80 A172

30 60

20 40

10 20

% Wound Closure Wound % 0 Closure Wound % 0 DMSO 100 300 900 DMSO 100 300 900 foretinib (nM) foretinib (nM) Figure III.3: Foretinib inhibits in vitro cell migration of adherent glioblastoma cells in a scratch assay. U251 and A172 confluent cell monolayers were scratched and imaged 24 hours after plating (0 hr). DMSO or foretinib was added immediately to the cells at the concentrations indicated. After 24 hours, each well was then re-imaged. The area of each image not occupied by cells (the “wound”) was quantified using TScratch software and the percent wound closure at 24 hours was calculated for each condition. Data shown are representative images of t-=0 and t=24 from two independent experiments (A) or the average of two independent duplicate experiments (B). Error bars are SEM.

Figure III.4: Foretinib inhibits glioblastoma cell invasion. (A) U251, SF188, and A172 neurospheres were plated in a collagen matrix and imaged immediately (0 hr), treated with DMSO or foretinib at the doses indicated and re-imaged at 15 hr (not shown) and 24 hr. Green circle indicates that absolute maximal radius of invasion. (B) Absolute maximal radius of each neurosphere was measured and the change in radius from t=0 to t=15 and t=24 was calculated. Data shown are representative of three independent experiments (A) or the average of three independent quintuplicate experiments (B). Error bars are SEM. Statistics are a two-sided t-test comparing the average absolute maximal radius of invasion of foretinib treatment to DMSO control (*P<0.05).

recapitulated by two-dimensional cell culture conditions. In order to address this shortcoming, I utilized the three-dimensional collagen invasion assay described above.

Briefly, U251, SF188 and A172 cells were induced to form neurospheres, plated in a collagen matrix and after 1 hour treated with increasing doses of foretinib. Migration of cells from the core of individual neurospheres were tracked and imaged at 15 and 24 hours after addition of the drug (Figure III.4). The absolute maximal radius of each neurosphere was measured and normalized to the radius at the time drug treatment (0 hr).

Foretinib decreased glioblastoma cell migration through a collagen in all three cell lines both visually (Figure III.4A) and quantitatively (Figure III.4B), with 100 nM foretinib decreasing cell migration in A172 and SF188 cells.

Figure III.5: UNC2025 inhibits glioblastoma cell migration in vitro. 25 Confluent U251 cell monolayers were 20 scratched and imaged 24 hours after plating (0 hr). DMSO or UNC2025 was added 15 immediately to the cells at the 10 concentrations indicated. After 24 hours, each well was then re-imaged. The area of 5 each image not occupied by cells (the % Wound Healing Wound % “wound”) was quantified using TScratch 0 software and the percent wound closure at DMSO 50 100 400 24 hours was calculated for each condition. UNC2025 (nM) Data shown represents the mean of two technical replicates in a pilot experiment.

UNC2025 inhibits glioblastoma cell migration in vitro.

UNC2025 is a TKI with low-nanomolar activity against MERTK and Fms-like tyrosine kinase 3 (FLT3).243 Unlike other MERTK inhibitors, UNC2025 is approximately

15-fold more specific for MERTK than AXL, making it a useful tool compound to tease out the role of MERTK inhibition.244 In order to verify that MERTK inhibition was

responsible for the inhibition of cell migration seen with foretinib, we performed an in

vitro scratch assay on U251 cells treated with increased concentrations of UNC2025 for

12 hours. Like foretinib, UNC2025 inhibited U251 cell migration in a dose-dependent

manner (Figure III.5). Given the distinct kinase inhibition profiles of UNC2025 and

foretinib, this strengthens the conclusion that the inhibitory effects of foretinib are due to

inhibition of MERTK.

MERTK regulates FAK expression.

Focal adhesion kinase (FAK) is a non-receptor tyrosine kinase that integrates pro-

migratory signals from integrins and RTKs in order to control and regulate cell

migration.245 Previously, the Keating lab has shown that stable MERTK knockdown in

U251 glioblastoma cells increased total FAK protein expression and phosphorylation.221

However, because these experiments were performed in cells with MERTK stably

knocked down, we were unable to conclude whether this dysregulation was a direct effect

of MERTK knockdown, or an adaptive response to the loss of MERTK signaling.

Therefore, in order to determine if MERTK directly regulates FAK, we utilized siRNA to

transiently knockdown MERTK in U251 cells. Interestingly, contrary to our previous

findings, siRNA knockdown of MERTK resulted in a robust decrease in total FAK

siMERTK Figure III.6: MERTK regulates FAK L NT 1 2 expression. U251 cells were treated with transfection reagent (L), non-targeting siRNA FAK (NT), or one of two (1 and 2) siRNA constructs targeting MERTK. 72 hours after transfection, MERTK lysates were collected and analyzed via immunoblot with the antibodies indicated. Data AXL shown represents a single pilot experiment. Tubulin

expression (Figure III.6). This suggests that MERTK regulates total FAK expression and is not necessarily incongruent with the previous finding of increased FAK expression in shMERTK cells—as this may be reflect a compensatory response required for cell survival following the loss of MERTK signaling.

Discussion

Taken together, data presented in this chapter provide pre-clinical evidence to support the hypothesis that MERTK inhibition is a viable strategy for inhibiting cell migration and invasion in GBM. First, I demonstrate that stable genetic knockdown of

MERTK decreases GBM cell invasion in a three-dimensional collagen invasion model

(Figure III.1), further enhancing the clinical relevance of the MERTK-dependent migration phenotype originally described in two-dimensional cell culture conditions by the Keating Lab.221 Next, using a clinically available multi-kinase inhibitor, foretinib, I demonstrate that pharmacologic inhibition of MERTK also inhibits cell migration and invasion in two- and three-dimensional pre-clinical in vitro models (Figure III.2, Figure

III.3, and Figure III.4). The decrease in cell migration I observed (Figure III.5) in U251

GBM cells treated with the MERTK-specific TKI, UNC2025, reinforces the conclusion that the decrease in cell migration observed with foretinib was due to MERTK inhibition.

Finally, my finding that FAK protein expression is decreased following siRNA knockdown of MERTK (Figure III.6) provides a hint of a possible mechanistic explanation as to how MERTK may regulate GBM cell migration.

The most compelling evidence supporting MERTK’s role in the regulation of

GBM migration and invasion that I present in this chapter relies on the use of a multi-

kinase inhibitor, foretinib. Foretinib is a potent inhibitor of MERTK, with nanomolar specificity for MERTK.242 However, I would be remiss to overlook the possible unintended consequences of using a multi-kinase inhibitor as a tool-compound. At the same nanomolar concentrations that MERTK is inhibited, foretinib also inhibits many additional RTKs including AXL (as described in Chapter IV), MET, and VEGFR. While the MERTK-specific TKI, UNC2025, also decreases GBM cell migration, the experiments described in this chapter do not rule out the possibility that the decrease in cell migration observed with foretinib may be due to inhibition of an alternate RTK. In order to prove that foretinib decreases GBM cell migration and invasion via MERTK inhibition, I would need to employ genetic strategies to individually silence additional foretinib targets and assess the effect on cell migration. However, if foretinib were to be acting via a mechanism other than MERTK inhibition, the clinical relevance of my findings would not be lost. Foretinib remains a clinically-available TKI which crosses the blood-brain barrier;246 that I show here to be effective at decreasing cell migration and invasion in patient-derived GBM cell lines. Given the abysmal therapeutic options currently available to patients with GBM, further investigation into the use of foretinib in this patient population is warranted.

The highly invasive nature of GBM remains the primary obstacle to improving patient outcomes and overall survival. Despite the large number of targeted cancer therapeutics that have been approved over the last decade, successful total surgical resection remains the primary predictor of survival in GBM patients. Understanding the molecular mechanisms that drive GBM cell migration and invasion will facilitate the development of improved pharmacologic strategies and treatment options for GBM

patients. Together, my findings support the hypothesis that MERTK inhibition is a viable strategy for inhibiting cell migration and invasion in multiple pre-clinical models of

GBM. Further, preliminary data suggesting that MERTK is required for expression of

FAK, a master regulator of cell migration, hints at a possible molecular mechanism by which MERTK regulates the invasive phenotype of GBM cells.

CHAPTER IV

ACQUIRED RESISTANCE TO RET INHIBITION IN RET-FUSION POSITIVE

NSCLC IS DRIVEN BY REACTIVATION OF THE RAS/MAPK PATHWAY2

Introduction

Lung cancer is the leading cause of cancer-related deaths in the United States in both men and women.2 Patients diagnosed with advanced stage non-small cell lung cancer (NSCLC)—which accounts for over 85% of lung cancer diagnoses—have only a

6% chance of survival at 5 years.247 Over the last decade many new oncogenes, primarily activated receptor tyrosine kinases (RTKs), capable of driving NSCLC have been identified leading to the development and successful clinical use of a number of small- molecule therapies.248 An early, well-established example of this paradigm is the use of the tyrosine kinase inhibitor (TKI) gefitinib in patients with activating mutations in the epidermal growth factor receptor (EGFR).16,249

Unfortunately, the use of targeted TKI therapies in NSCLC is not curative and relapse secondary to acquired TKI resistance is universal.106 Mechanisms of acquired resistance generally fall into one of three broad categories: 1) the acquisition of a mutation that prevents the TKI from binding its intended target (e.g. resistance to first- generation EGFR TKIs due to the EGFR T790M gatekeeper mutant)108; 2) the activation of alternate or bypass signaling pathways (e.g. MET amplification in patients resistant to

EGFR TKIs);60,250 and 3) a change in cell phenotype that decreases dependency on the

2Portions of this chapter has been submitted for publication to Molecular Cancer Therapeutics and are reprinted here with permission. 54

original driver oncogene (e.g. conversion of EGFR-driven NSCLC to small-cell lung cancer (SCLC) or EGFR TKI-induced transition to a mesenchymal phenotype).251,252

Oncogenic rearrangements of the gene encoding the RTK RET (rearranged during transfection) have been identified in NSCLC, papillary thyroid cancer (PTC), and colorectal cancer.253 Approximately 1-2% of NSCLCs are driven by RET fusions, which now account for as many as 20% of lung cancers of never-smokers in whom no other known NSCLC-driving mutations have not been identified.24,26,28 These chromosomal rearrangements link the intracellular 3’-RET kinase domain to the 5’-dimerization domain of an unrelated gene (most commonly CCDC6 (coiled-coil domain containing 6),

KIF5B (kinesin family member 5b), and NCOA4 (nuclear receptor co-activator 4),254 resulting in constitutive expression of the RET fusion protein, homodimerization, and ligand-independent activation of pro-survival and pro-proliferation signaling.

RET TKIs are clinically available and multiple agents are currently in clinical trials for RET+ NSCLC. In this study, we demonstrate that ponatinib is active in a pre- clinical model of RET-driven NSCLC and report two distinct mechanisms of ponatinib resistance, both of which restore signaling through the RAS/MAPK pathway: oncogenic

NRAS and upregulation of wild-type EGFR signaling.

Results

Ponatinib inhibits RET signaling and cell growth in the RET-rearranged NSCLC cell line LC-2/ad.

Ponatinib is a multi-kinase TKI originally designed as a second-generation ABL inhibitor, which is FDA-approved as a second-line therapy for patients with Philadelphia-

Figure IV.1: Ponatinib inhibits RET and decreases cell proliferation of LC-2/ad cells.(A) Cell viability of LC-2/ad (blue), TPC1 (pink), and H2228 (green) cells treated with increasing doses of ponatinib for 72 hours. (B) Western blot analysis of LC-2/ad cells treated with increasing doses of ponatinib for 2 hours. Lysates were analyzed via Western blotting with the antibodies indicated. Data are the average of three independent triplicate experiments (A) or immunoblots representative of three independent experiments (B). Error bars are SEM. positive leukemia.255 Ponatinib inhibits the kinase domain of wild-type RET and the gatekeeper mutant, RET V804M/L, with low-nanomolar affinity.256 Clinical trials of ponatinib in RET-fusion positive NSCLC patients are ongoing (NCT01935336,

NCT01813734). The LC-2/ad cell line (which expresses CCDC6-RET and is the only published patient-derived NSCLC cell line driven by a RET fusion)257,258 is exquisitely sensitive to ponatinib in vitro with an IC50 of approximately 25 nM (Figure IV.1A). This is comparable to the effect of ponatinib on cell proliferation in the TPC1 cell line (a papillary thyroid cancer cell line that also expresses and is driven by a CCDC6-RET fusion and has been previously shown to be sensitive to RET inhibition with ponatinib).259,260 Notably, LC-2/ad and TPC1 cells are considerably more sensitive to

Figure IV.2: PR1 and PR2 cells are ponatinib-resistant derivatives of the LC-2/ad cell line. (A) Dose escalation schema for the derivation of the PR1 and PR2 cell lines. (B) Cell viability of LC-2/ad (blue), PR1 (pink), and PR2 (green) cells treated with increasing doses of ponatinib for 72 hours. Error bars: mean ± SEM for 3 triplicate experiments.

ponatinib than H2228 cells (IC50>1M, Figure IV.1A), which harbor an oncogenic

EML4-ALK fusion.27 Ponatinib does not inhibit ALK kinase activity in vitro at

261 concentrations less than 1 M, thus the high IC50 in the H2228 cell line likely reflects the accumulation of known kinase targets inhibited by ponatinib at concentrations greater than 1 M—supporting the conclusion that the low-nanomolar sensitivity of LC-2/ad cells is due to RET inhibition. Further, when LC-2/ad cells were treated with increasing doses of ponatinib for 2 hours, tyrosine phosphorylation of RET was decreased in a dose dependent manner (Figure IV.1B). This was accompanied by decreased phosphorylation of SHC1, which mediates activation of PI3K/AKT and RAS/MAPK signaling (Figure

IV.1B).143

Generation of ponatinib-resistant derivatives of the LC-2/ad cell line.

In an attempt to generate two discrete in vitro models of resistance to RET inhibition, LC-2/ad cells were cultured in increasing concentrations of ponatinib using two dose escalation strategies (Figure IV.2A). The PR1 (Ponatinib Resistant-1) cell line was generated by initially culturing LC-2/ad cells in a fixed dose of 200 nM ponatinib that was increased to 400 nM after resistant clones successfully proliferated at a normal rate. The PR2 (Ponatinib Resistant-2) cell line was generated using a traditional dose escalation strategy, with LC-2/ad cells initially treated with a sub-nanomolar concentration of ponatinib that was incrementally increased to 400 nM over 10 months.

The LC-2/ad derivatives were deemed resistant when proliferation at 400 nM ponatinib matched that of parental cells. Unless otherwise noted, PR1 and PR2 cells were subsequently maintained in 200 nM ponatinib.

PR1 and PR2 cell proliferation was significantly less sensitive to ponatinib when compared to parental LC-2/ad cells, with IC50’s of approximately 660 nM and 400 nM,

Figure IV.3: PR1 and PR2 cells are resistant to RET TKIs alectinib and cabozantinib. (A) Cell proliferation of LC-2/ad (blue), PR1 (pink), and PR2 (green) cells treated with increasing doses of alectinib for 72 hours. Error bars: mean ± SEM for 3 triplicate experiments. (B) Cell viability of LC-2/ad (blue), PR1 (pink), and PR2 (green) cells treated with increasing doses of cabozantinib for 72 hours. Error bars: mean ± SEM for 3 triplicate experiments. 58

respectively (Figure IV.2b). This dramatic shift in IC50, as opposed to a complete loss of sensitivity to ponatinib, is likely due to the cumulative inhibition of multiple kinase targets by ponatinib at high concentrations.261 Further, the resistance phenotype extends to two additional RET TKIs, cabozantinib and alectinib (Figure IV.3), both of which are currently being assessed in clinical trials accruing RET-fusion positive NSCLC patients.262-264

Break-apart FISH analysis revealed that the PR1 and PR2 cell lines still harbor the RET fusion gene with no notable change in copy number (Figure IV.4A). Expression of CCDC6-RET mRNA was confirmed using qRT-PCR (Figure IV.4B). Sequencing revealed a wild-type RET kinase domain in the parental LC-2/ad, PR1, and PR2 cell lines, establishing that a RET kinase domain mutation did not account for the resistance phenotype of the PR1 or PR2 cells (data not shown). This lack of kinase domain mutation was not entirely surprising as ponatinib is active again RET gatekeeper mutations

V804L/M.256

In contrast to parental LC-2/ad cells, which are dependent upon RET kinase activity, RET is expressed but not phosphorylated in the PR1 and PR2 cell lines (Figure

IV.4C). PR1 and PR2 cells express similar levels and activation of ERK1/2 and AKT to the parental LC-2/ad cells (Figure IV.4C). However, unlike the parental LC-2/ad cells ponatinib failed to perturb signaling through either ERK1/2 or AKT (Figure IV.1B and

Figure IV.5A). The same was true with alectinib which inhibited phosphorylation of RET and ERK1/2 in a dose dependent manner in the LC-2/ad cells, but not in either the PR1 or

PR2 cell line (Figure IV.5B). Thus providing further evidence that both the PR1 and PR2

cell lines have acquired a bypass signaling mechanism that drives PI3K/AKT and

RAS/MAPK signaling independent of RET.

Figure IV.4: PR1 and PR2 cells maintain expression of the CCDC6-RET fusion protein. (A) LC-2/ad, PR1, PR2, and H2228 cells were analyzed via break-apart fluorescence in-situ hybriziation assay utilizing probes for 5’-RET (pink) and 3’-RET (green). Pink and green signal separated by >1 signal diameter indicates a positive fusion. (B) CCDC6-RET mRNA expression quantified via real-time qRT-PCR analysis using primers that span the fusion break-point, (C) Western blot analysis of untreated LC-2/ad, PR1 and PR2 cells. Lysates were analyzed via western blotting with the antibodies indicated.

Figure IV.5: Downstream signaling in the PR1 and PR2 cell lines is resistant to RET inhibition. (A) Western blot analysis of PR1 and PR2 cells removed from ponatinib for 24 hours prior to 2 hours treatment with increasing doses of ponatinib. Lysates were analyzed via western blotting with the antibodies indicated. (B) Western blot analysis of LC-2/ad, PR1, and PR2 cells treated with increasing concentrations of alectinib for 2 hours. PR1 and PR2 cells were removed from ponatinib for 24 hours prior to treatment. Lysates were analyzed via western blotting with the antibodies indicated.

Figure IV.6: MET phosphorylation is increased in the PR1 cell line, but does not mediate ponatinib resistance. (A) Phospho-RTK antibody array performed on lysates obtained from untreated LC-2/ad and PR1 cells with green boxes indicating the location of phospho-MET. (B) MTS cell proliferation assay of PR1 cells treated with increasing concentrations of ponatinib plus DMSO control or 200 nM crizotinib for 72 hours. PR1 cells were removed from ponatinib 24 hours prior to the onset of the assay. (C) Western Blot analysis of LC-2/ad, PR1 and PR2 cells treated with DMSO control or 200 nM crizotinib for 2 hours. Lysates were analyzed via Western blotting with the antibodies indicated.

MET phosphorylation is increased in the PR1 cell line but does not mediate resistance to ponatinib.

In order to determine if an alternate receptor tyrosine kinase was driving the resistance phenotype in the PR1 cell line, a phospho-antibody array was used to screen for any differentially phosphorylated RTKs. Phosphorylated MET was the only RTK with significantly increased phosphorylation in the PR1 cells compared to the LC-2/ad cells (Figure IV.6A), which was confirmed via western blot (Figure IV.6C). Interestingly,

despite increased phospho-MET, the PR1 cells were not sensitive to the MET inhibitor crizotinib, nor did a fixed MET-inhibiting dose of crizotinib restore sensitivity to ponatinib (Figure IV.6B). Further, inhibition of MET did not decrease phosphorylation of

ERK1/2 (Figure IV.6C), suggesting that MET does not regulate signaling RAS/MAPK signaling in the PR1 cells. Increased expression and phosphorylation of MET has often been used as a proxy for MET amplification-mediated acquired resistance to TKIs. This data suggests that the increased phosphorylation and expression of MET alone should not be considered a biomarker for functional MET signaling.

Oncogenic NRAS p.Q61K confers resistance to RET-inhibition in the PR1 cell line.

Next-generation RNA sequencing was performed on all three cell lines in order to identify differentially expressed genes and to detect mutations in known oncogenes or tumor suppressors. EGFR, BRAF, and KRAS which are commonly mutated in NSCLC were found to be wild-type in all three cell lines, however a previously described activating mutation in NRAS was identified in the PR1 cell line.

DNA sequencing confirmed that a single base pair substitution in NRAS encoding the NRAS p.Q61K mutant was present in the PR1 cell line, but not in the parental LC-

2/ad or PR2 derivative (Figure IV.7A). siRNA knockdown of RET and/or NRAS protein expression in the LC-2/ad cells reinforced the RET-dependence of the LC-2/ad cells, with decreased cell proliferation seen with knockdown of RET, but not NRAS (Figure IV.7B).

In contrast, PR1 cell proliferation is more sensitive to silencing of NRAS than RET

(Figure IV.7D), consistent with an NRAS-dependent phenotype. Knockdown of RET and

NRAS in combination did not further decrease cell proliferation above single knockdown of the dominant oncogene in either parental LC-2/ad or PR1 cells (Figure IV.7B,D).

Figure IV.7: Oncogenic NRAS p.Q61K is expressed and drives cell proliferation in the PR1 cell line.

Figure IV.7: Oncogenic NRAS p.Q61K is expressed and drives cell proliferation in the PR1 cell line. (A) DNA sequencing of NRAS revealed the PR1 cell line is heterozygous for an A to C single base pair substitution responsible for the NRAS p.Q61K mutation. This was not found in either LC-2/ad cells or PR2 cells. (B) LC-2/ad cells were transfected with siRNA targeting RET, NRAS, RET and NRAS, or a non- targeted control. Cell viability was measured after 7 days using the CyQuant Direct Cell Proliferation Assay and normalized to the non-targeting control. Error bars: mean ± SEM for 3 triplicate experiments. (C) Western Blot analysis of LC-2/ad cells transfected under the same conditions collected lysed 48 hours after transfection. Lysates were analyzed via Western blotting with the antibodies indicated. (D) PR1 cells were transfected with siRNA targeting RET, NRAS, RET and NRAS, or a non-targeted control. Cell viability was measured after 7 days using the CyQuant Direct Cell Proliferation Assay and normalized to the non-targeting control. Error bars: mean ± SEM for 3 triplicate experiments. (E) Western Blot analysis of PR1 cells transfected under the same conditions collected lysed 48 hours after transfection. Lysates were analyzed via Western blotting with the antibodies indicated. (F) Cell viability of LC-2/ad (blue) and PR1 (pink) cells (which had been removed from ponatinib for 24 hours) treated with increasing doses of trametinib for 72 hours at which point cell viability was measured using the CyQuant Direct Cell Proliferation Assay. Error bars: mean ± SEM for 3 triplicate experiments. (G) Pull-down assay for activated NRAS in untreated LC-2/ad, PR1 and PR1-OP cells. Lysates from LC-2/ad, PR1 and PR1-OP cells were collected and the GTP-bound form of RAS was pulled down using glutathione beads conjugated to the RAF-RBD (Ras binding domain). Western Blot analysis using NRAS and pan-RAS antibodies was performed on the pulled-down sample, and total RAS expression in the whole cell lysate was used as a loading control.

Consistent with the acquisition of a RAS-driven phenotype, PR1 cells were more sensitive to the MEK-inhibitor trametinib (IC50=32 nM) compared to parental LC-2/ad cells (IC50=195 nM) (Figure IV.7F). Not surprisingly, the proportion of activated GTP- bound NRAS in the PR1 cells was greater in the PR1 cells than in the LC-2/ad parental cells as assessed by a RAS-GTP pull-down assay (Figure IV.7G).

We attempted to express mutant NRAS p.Q61K via both retroviral transduction and transient transfection in the parental LC-2/ad cells, but LC-2/ad cells did not tolerate expression of this oncogene. However, we were able to confirm that stable expression of

NRAS p.Q61K can induce resistance to ponatinib in the CCDC6-RET expressing TPC1 papillary thyroid cancer cell line. Resistance to ponatinib was only observed in

NRASQ61K-expressing TPC1 cells maintained in 10 nM ponatinib. TPC1-NRASQ61K cell proliferation was less sensitive to ponatinib compared to empty-vector transduced cells (TPC1-EV) cultured in the same conditions (Figure IV.8A). This was supported by the observation that EV and NRASQ61K cells not cultured in ponatinib expressed approximately the proportion of GTP-NRAS, and only NRASQ61K cells cultured in ponatinib demonstrated a robust increase in GTP-bound active NRAS (Figure IV.8B).

Phosphorylation of ERK1/2 remained sensitive to RET-inhibition in TPC1-EV cells but not TPC1-NRASQ61K cells, ruling out the possibility that the ponatinib treatment required to induce the switch to NRAS-dependence was responsible for the resistance phenotype (Figure IV.8C).

Figure IV.8: Oncogenic NRAS p.Q61K confers resistance to RET inhibition in the CCDC6-RET expressing TPC1 papillary thyroid cancer cell line. (A) TPC1 cells were stably transduced with retrovirus containing NRAS p.Q61K expressing vector (NRASQ61K) or an empty vector control (EV). Cells were selected with puromycin and maintained in 10 nM ponatinib. Cells were removed from ponatinib for 24 hours and treated with increasing doses of ponatinib for 72 hours at which point cell viability was measured using the CyQuant Direct Cell Proliferation Assay. Error bars: mean ± SEM for 3 triplicate experiments. (B) Pull-down assay for activated NRAS-GTP in TPC1-EV, -NRASQ61K, and -NRASQ61K-IP cells. Lysates from TPC1-EV, -NRASQ61K, and -NRASQ61K-IP cells were collected and the GTP-bound form of RAS was pulled down using glutathione beads conjugated to the RAF-RBD (Ras binding domain). Western Blot analysis using NRAS and pan-RAS antibodies was performed on the pulled-down sample, and total RAS expression in the whole cell lysate was used as a loading control. (C) Western blot analysis of TPC1 NRASQ61K or EV cells treated with 10 nM ponatinib for 2 hours. Cells were removed from ponatinib 24 hours prior to the onset of the experiment. Western blot analysis of lysates was performed using the antibodies indicated.

NRAS p.Q61K persists as the dominant oncogene in the PR1 cell line in the absence of RET inhibition.

To test the permanence of the oncogene switch observed in the PR1 cell line, we removed the cells from ponatinib for six weeks. PR1 cells removed from RET inhibition

(PR1-Out of Ponatinib or PR1-OP) remain ponatinib-resistant, with no change in IC50

(Figure IV.9A). Interestingly, PR1-OP cells regained phospho-RET and phospho-SHC1, which were inhibited by ponatinib in a dose dependent manner (Figure IV.9C). Ponatinib did not, however, inhibit phospho-AKT or phospho-ERK (Figure IV.9C)—suggesting that while RET is phosphorylated and capable of signaling through SHC1, its ability to regulate downstream signaling is negated in the presence of oncogenic NRAS. siRNA knockdown of NRAS decreased cell proliferation in PR1-OP cells, whereas knockdown of RET had no effect (Figure IV.9B). This confirms that PR1 cells have undergone a stable switch to NRAS-dependence that is independent of RET.

Figure IV.9: NRAS pQ61K persists as the dominant oncogene in the PR1 cell line in the absence of chronic RET inhibition. (A) Cell viability of PR1 cells (closed circle) and PR1 cells maintained without ponatinib for > 4 weeks (PR1-OP) (open circle) were treated with increasing doses of ponatinib for 72 hours at which point cell viability was measured using the CyQuant Direct Cell Proliferation Assay. Error bars: mean ± SEM for 3 triplicate experiments. (B) PR1-OP cells were transfected with siRNA targeting RET, NRAS, RET and NRAS, or a non-targeted control. Cell viability was measured after 7 days using the CyQuant Direct Cell Proliferation Assay and normalized to the non- targeting control. Error bars: mean ± SEM for 3 triplicate experiments. (C) Western Blot analysis of PR1-OP cells treated with increasing doses of ponatinib for 2 hours and analyzed with the antibodies indicated.

PR2 cells have undergone EMT-like changes in mRNA expression, however this does not contribute to the resistance phenotype.

With no identifiable oncogenic mutations detectable in the PR2 cell line, we instead looked for any identifiable changes in gene expression indicative of a switch in cellular programming that could be driving the resistance phenotype in the PR2 cells.

Using an EMT geneset265 we filtered for EMT genes significantly up or down regulated in the PR2 cell line and then ranked according to the fold change compared to the parental LC-2/ad cells. Interestingly, genes that are increased in EMT were generally increased in PR2 cells, and genes down regulated with EMT were decreased in the PR2 cells compared to the parental LC-2/ad cells (Figure IV.10A). Further, when we looked at genes associated with TGFβ-mediated EMT signaling, many of these genes were also significantly changed in the PR2 cell line including TGFB1 and TGFB2. Interestingly,

TGFBR1 and TGFBR2 were also up regulated in the PR2 cell line, although not significantly. This led us to hypothesize that the resistance phenotype of the PR2 cell line may be due to TGFβ-mediated signaling responsible for driving EMT. In order to test this hypothesis, we assessed the sensitivity of the PR2 cell line to the TGFβR inhibitor

SB525334. However, PR2 cells were not sensitive to SB525334 and this was not enhanced by the addition of a fixed, RET-inhibiting dose of ponatinib (Figure IV.10C).

Further, SB52534 did not restore sensitivity of the PR2 cell line to ponatinib (Figure

IV.10D).

Figure IV.10: PR2 cells express changes in mRNA expression similar to EMT, but are not dependent on TGFB signaling

Figure IV.10. PR2 cells express changes in mRNA expression similar to EMT, but are not dependent on TGFB signaling. (A) Normalized gene expression of EMT genes in LC-2/ad and PR2 cells significantly changed in the PR2 cells (p<0.05) and ranked according to fold change (red=increased compared to average expression of all replicates; green = decreased compared to average expression of all replicates). (B) mRNA expression of genes associated with TGFB-mediated EMT in the LC-2/ad and PR2 cells. (C) Cell viability of PR2 cells treated with increasing doses of SB525334 plus DMSO (light green) or 10 nM ponatinib (dark green) for 72 hours at which point cell viability was measured using the CyQuant Direct Cell Proliferation Assay. Data represents the average of three technical replicates. (D) Cell viability of PR2 cells treated with increasing doses of ponatinib plus DMSO (light green) or 1 uM SB525334 (dark green) for 72 hours at which point cell viability was measured using the CyQuant Direct Cell Proliferation Assay. Data represents the average of three technical replicates.

Activation of wild-type EGFR and AXL signaling mediates acquired resistance to ponatinib in PR2 cells.

In the absence of any known, detectable oncogenic mutations, we hypothesized that resistance in the PR2 cell line may be mediated by increased expression or activation of a wild-type bypass signaling mechanism. Previously, our lab has shown that wild-type

EGFR signaling can mediate acquired resistance to ROS1 inhibition in the HCC78

NSCLC cell line224 and we hypothesized that increased expression and activation of a wild-type RTK in the PR2 cell line may be contributing to the resistance phenotype in a similar manner. Candidate RTKs were identified via the RNA sequencing data. Both

AXL and EGFR, RTKs previously implicated in acquired resistance in NSCLC,121,224 were upregulated in the PR2 cells (Table IV.1). Further, mRNA expression of the AXL ligand GAS6 and the EGFR ligands HB-EGF and NRG1 were also upregulated (Table

IV.2)—suggesting that these two RTKs may be actively signaling in the PR2 cell line.

Given that PR2 cells are resistant to cabozantinib (Figure IV.3B), which also inhibits

AXL, we hypothesized that EGFR was the primary driver of resistance in the PR2 cell line. To test this, LC-2/ad and PR2 cells were treated with 10 nM ponatinib, 500 nM afatinib (a second-generation EGFR TKI), or a combination of both ponatinib and afatinib for four hours. Consistent with previous experiments, ponatinib decreased phosphorylation of RET, ERK1/2, and AKT in the parental LC-2/ad cells but not the PR2 cells (Figure IV.11A). The opposite was true of afatinib, which robustly decreased phospho-ERK1/2 and phospho-AKT in PR2 cells but had no effect of phospho-ERK1/2 or phospho-AKT in LC-2/ad cells (Figure IV.11A). We did detect low levels of total

EGFR protein in the LC-2/ad cells, although phospho-EGFR was virtually absent, in

contrast to the robust phospho-EGFR signal detected in PR2 cells (Figure IV.11A). A similar effect was observed in PR2 cells treated with gefitinib, an EGFR TKI with a distinct mechanism of inhibition (Figure IV.11B), minimizing the possibility that this observation was due to off-target effects of either TKI.

Table IV.1

LC-2/ad PR1 PR2 Gene Fold Fold avg FPKM avg FPKM p-value avg FPKM p-value Change change AXL 1.61 38.78 24.09 <0.05 81.52 50.63 <0.005 KIT 0.36 1.12 3.10 <0.005 2.67 7.42 <0.005 MERTK 1.42 3.11 2.19 n.s. 6.02 4.24 <0.05 CSF1R 0.05 0.02 -2.99 <0.05 0.21 4.20 n.s. FGFR1 0.7 1.86 2.65 <0.05 2.31 3.30 <0.005 EGFR 8.27 15.70 1.90 <0.005 23.39 2.83 <0.005 DDR2 0.11 0.16 1.42 n.s. 0.21 1.91 n.s. EPHB2 13.08 15.53 1.19 n.s. 24.53 1.88 n.s. IGF2R 22.54 24.00 1.06 n.s. 35.28 1.57 n.s. FGFR2 3.47 3.40 -1.02 n.s. 4.87 1.40 n.s. ERBB2 80.9 56.63 -1.43 <0.05 108.79 1.34 n.s. MST1R 12.95 19.15 1.48 n.s. 15.32 1.18 n.s. FGFR4 2.05 1.07 -1.92 <0.05 2.41 1.18 n.s. IGF1R 8.13 12.71 1.56 <0.05 9.26 1.14 n.s. TYRO3 3.3 3.13 -1.06 n.s. 3.73 1.13 n.s. EPHB4 35.2 36.33 1.03 n.s. 38.13 1.08 n.s. EPHA1 11.39 9.84 -1.16 n.s. 12.08 1.06 n.s. MET 87.24 110.14 1.26 n.s. 89.69 1.03 n.s. KDR 2.05 3.79 1.85 n.s. 2.07 1.01 n.s. EPHA2 57.63 59.18 1.03 n.s. 52.59 -1.10 n.s. EPHA4 0.87 1.58 1.81 <0.05 0.74 -1.18 n.s. ERBB3 26.5 33.56 1.27 n.s. 21.23 -1.25 n.s. EPHB3 17.77 14.21 -1.25 n.s. 10.02 -1.77 <0.05 INSR 3.73 0.71 -5.24 <0.005 1.37 -2.72 <0.05 FGFR3 11.04 11.20 1.01 n.s. 3.89 -2.84 <0.05 PDGFRA 1.48 0.49 -3.05 <0.05 0.45 -3.29 <0.05 ROS1 0.34 0.15 -2.27 n.s. 0.01 -34.00 n.s. EPHB1 5.01 0.34 -14.60 <0.005 0.05 -100.20 <0.005

Table IV.2

LC-2/ad PR1 PR2

Gene Fold avg FPKM avg FPKM p-value avg FPKM Fold change p-value Change NRG1 0.22 1.50 6.82 n.s. 4.36 19.82 <0.05 GAS6 2.69 15.33 5.70 <0.05 45.74 17.00 <0.05 PDGFB 0.68 1.71 2.52 <0.05 9.10 13.38 <0.05 BDNF 0.87 5.56 6.39 <0.05 7.97 9.16 <0.005 PROS1 4.44 10.74 2.42 n.s. 24.24 5.46 <0.05 PDGFD 0.99 4.76 4.81 <0.05 4.42 4.46 <0.005 HBEGF 5.44 9.34 1.72 n.s. 17.26 3.17 <0.005 PDGFA 1.64 1.50 -1.09 n.s. 4.88 2.98 <0.005 NGF 2.30 2.27 -1.02 n.s. 6.81 2.96 <0.05 FGF11 0.95 0.62 -1.52 n.s. 2.54 2.67 <0.05 VEGFC 5.34 16.60 3.11 n.s. 13.83 2.59 <0.05 EFNA2 1.40 1.84 1.32 n.s. 2.67 1.91 n.s. PDGFC 7.86 16.42 2.09 <0.005 14.22 1.81 <0.005 EFNB2 11.22 13.47 1.20 n.s. 18.44 1.64 <0.05 BTC 4.43 2.93 -1.51 n.s. 6.38 1.44 <0.05 VEGFB 29.50 46.19 1.57 <0.05 39.89 1.35 n.s. EFNA3 15.93 16.89 1.06 n.s. 19.24 1.21 n.s. EFNB1 5.24 3.57 -1.47 <0.05 5.99 1.14 n.s. NRG4 0.37 0.42 1.12 n.s. 0.42 1.14 n.s. EREG 20.87 14.15 -1.47 n.s. 22.76 1.09 n.s. PIGF 23.97 21.41 -1.12 n.s. 26.06 1.09 n.s. AREG 1.81 2.29 1.26 n.s. 1.88 1.04 n.s. VEGFA 78.36 53.82 -1.46 n.s. 75.30 -1.04 n.s. EGF 1.87 0.42 -4.43 <0.005 1.79 -1.04 n.s. FGF13 6.45 12.71 1.97 <0.05 5.41 -1.19 n.s. EFNA5 9.56 11.14 1.17 n.s. 7.86 -1.22 n.s. CSF1 10.03 18.14 1.81 n.s. 7.97 -1.26 n.s. FGF8 62.63 58.29 -1.07 n.s. 48.36 -1.30 n.s. KITLG 18.60 15.25 -1.22 n.s. 13.30 -1.40 n.s. FLT3LG 10.94 9.59 -1.14 n.s. 7.46 -1.47 <0.05 MST1 2.11 1.03 -2.05 <0.05 1.42 -1.49 n.s. EFNA1 43.17 41.65 -1.04 n.s. 28.92 -1.49 n.s.

Figure IV.11. EGFR regulates downstream signaling and cell proliferation in the PR2 cell line. (A) Western Blot analysis of LC-2/ad and PR2 cells treated with a DMSO control, 10 nM ponatinib, 500 nM afatinib, or 10 nM ponatinib and 500 nM ponatinib for 4 hours. Lysates were analyzed via Western blotting with the antibodies indicated. (B) Western Blot analysis of LC-2/ad and PR2 cells treated with a DMSO control, 10 nM ponatinib, 1 uM gefitinib, or 10 nM ponatinib and 1 uM gefitinib for 4 hours. Lysates were analyzed via Western blotting with the antibodies indicated. (C) Cell viability of LC-2/ad (blue) and PR2 (green) cells treated with increasing doses of afatinib for 72 hours. Error bars: mean ± SEM for 3 triplicate experiments.

Given that control of oncogenic signaling appears to have switched from RET to

EGFR in the PR2 cell line, we posited that PR2 cell proliferation would be more sensitive to EGFR inhibition than that of the parental LC-2/ad cells. As expected, PR2 cell proliferation was more sensitive to afatinib than proliferation of the parental LC-2/ad cells (Figure IV.11C). However, addition of 10 nM ponatinib did not further sensitize the

PR2 cells to afatinib, nor did afatinib restore PR2 sensitivity to ponatinib (Figure IV.12A,

B), suggesting that EGFR-driven cell proliferation is independent of RET.

We hypothesized that AXL signaling may contribute to the EGFR-mediated resistance phenotype we observed in the PR2 cells given that AXL has been shown to diversify signaling and mitigate resistance to EGFR inhibition.200 Inhibition of AXL with the TKIs cabozantinib or foretinib enhanced the degree to which afatinib inhibited cell

Figure IV.12. PR2 sensitivity to afatinib is not enhanced by RET inhibition and vice versa. (A) Cell viability of PR2 cells treated with increasing doses of afatinib plus DMSO (closed circle) or 1 uM geftinib (open circle) for 72 hours. Error bars: mean ± SEM for 3 triplicate experiments. (B) Cell viability of PR2 cells treated with increasing doses of ponatinib plus DMSO (closed circle) or 500 nM afatinib (open circle) for 72 hours. Error bars: mean ± SEM for 3 triplicate experiments.

proliferation in the PR2 cells (Figure IV.13A). Consistent with earlier data, cabozantinib had no effect on cell proliferation as a single agent (Figure IV.3B and Figure IV.13A).

Interestingly, EGF stimulation increased phosphorylation of ERK1/2 and AKT in the

PR2 cells, which was blocked by afatinib, while GAS6 stimulation only increased phosphorylation of AKT in a manner that was blocked by the AXL inhibitor foretinib

(Figure IV.13B, C)—suggesting that AXL signals primarily via the pro-survival

PI3K/AKT pathway. GAS6 failed to increase phospho-AKT in the parental LC-2/ad cells, perhaps due to the low expression of AXL in these cells, and foretinib (which is also a RET inhibitor) inhibited phosphorylation of both AKT and ERK (Figure IV.13C), but EGF treatment did increase phospho-ERK1/2 (Figure IV.13B)—indicating that the

RET-driven LC-2/ad cells are primed to utilize EGFR to drive MAPK signaling. These findings were replicated in an additional fusion-kinase inhibition resistance model,

SLC34A2-ROS1 lung adenocarcinoma cells with acquired resistance to the ROS1 inhibitor TAE-684 (TAER).224 Cabozantinib and foretinib dramatically sensitized TAER cells to afatinib and GAS6 stimulation also preferentially activated AKT signaling

(Figure IV.14), suggesting that AXL contributes to wild-type EGFR driven cell proliferation, regardless of a cell’s original molecular driver.

Figure IV.13. AXL inhibition enhances sensitivity to EGFR inhibition in the PR2 cell line. (A) Cell viability of PR2 cells treated with increasing doses of afatinib plus DMSO (green), 500 nM cabozantinib (pink), or 500 nM foretinib (blue) for 72 hours. Error bars: mean ± SEM for 3 triplicate experiments (n=9). (B) Western blot analysis of LC-2/ad and PR2 cells serum starved for 2 hours with DMSO (-) or 500 nM afatinib (+) b and then treated with vehicle (-) or 100 ng/ml EGF (+) for 15 minutes. Lysates were analyzed via Western blotting with the antibodies indicated. (C) Western blot analysis of LC-2/ad and PR2 cells serum starved for 2 hours with DMSO (-) or 500 nM foretinib (+) and then treated with vehicle (-) or 100 ng/ml GAS6 (+) for 30 minutes. Lysates were analyzed via western blotting with the antibodies indicated.

Figure IV.14: AXL inhibition sensitizes HCC78-TAER cells to EGFR inhibition (A) Cell viability of TAER cells treated with increasing doses of afatinib plus DMSO (green), 500 nM cabozantinib (pink), or 500 nM foretinib (blue) for 72 hours at which point cell viability was measured using the CyQuant Direct Cell Proliferation Assay. Error bars: mean ± SEM for 3 triplicate experiments (n=9). (B) Western Blot analysis of TAER cells serum starved for 2 hours with DMSO or 500 nM foretinib and treated with 100 ng/ml GAS6 for the time points indicated. Western blot analysis of lysates was performed using the antibodies indicated. (B) Western blot analysis of TAER cells serum starved for 2 hours with DMSO or 500 nM afatinib and treated with 100 ng/ml EGF for the time points indicated. Western blot analysis of lysates was performed using the antibodies indicated.

In the absence of chronic RET-inhibition, PR2 cells are co-dependent upon EGFR and RET.

Having demonstrated the permanence of the oncogene switch in the PR1 cell line, we sought to determine whether the same was true of EGFR signaling in the PR2 cells.

PR2 cells were removed from ponatinib-containing media for six weeks (PR2-OP), resulting in minimal restoration of ponatinib sensitivity (Figure IV.15A) and a concomitant loss of sensitivity to afatinib (Figure IV.15B). We hypothesized that in acquiring resistance to RET-inhibition the PR2 cells had simply gained a degree of plasticity with regards to the source of oncogenic signaling; in the presence of a single agent RET or EGFR inhibitor the PR2-OP cells can rapidly adapt to rely upon signaling through the remaining uninhibited pathway. Therefore, it is not surprising that maximal inhibition of PR2-OP cell proliferation is observed when both RET and EGFR signaling is inhibited (Figure IV.15C). This observation is further supported by what we observed at the protein level: maximal inhibition of phospho-ERK and phospho-AKT is only observed in PR2-OP cells treated with combination ponatinib plus afatinib and RET and

EGFR are both fully inhibited (Figure IV.15D).

EGFR and AXL are increased and capable of driving resistance in LC-2/ad cells following RET inhibition

We hypothesized that the EGFR-driven resistance phenotype observed in the PR2 cells may reflect an outgrowth of a small population of LC-2/ad cells capable of engaging early adaptive resistance mechanisms to increase expression of EGFR and/or AXL in order promote cell proliferation and survival even in the context of RET inhibition.

Figure IV.15: Removal of PR2 cells from ponatinib partially restores RET- dependence.(A) Cell viability of PR2 cells (closed circle) and PR2 cells maintained without ponatinib for > 6 weeks (PR2-OP) (open circle) were treated with increasing doses of ponatinib for 72 hours at which point cell viability was measured using the CyQuant Direct Cell Proliferation Assay. Error bars: mean ± SEM for 3 triplicate experiments (n=9). (B) Cell viability of PR2 (closed circle) and PR2-OP (empty circle) cells treated with increasing doses of afatinib for 72 hours at which point cell viability was measured using the CyQuant Direct Cell Proliferation Assay. Error bars: mean ± SEM for 3 triplicate experiments (C) Cell viability of PR2-OP cells treated with increasing doses of afatinib plus DMSO (closed circle) or 10 nM ponatinib (open circle) for 72 hours at which point cell viability was measured using the CyQuant Direct Cell Proliferation Assay. Error bars: mean ± SEM for 3 triplicate experiments. (D) Western Blot analysis of PR2-OP cells treated with DMSO, 10 nM ponatinib, 500 nM afatinib, or 10 nM ponatinib and 500 nM afatinib for 4 hours. Western blot analysis of lysates was performed with the antibodies indicated.

Interestingly, after only 24 hours of ponatinib treatment, AXL expression is increased in parental LC-2/ad cells (Figure IV.16A). Further, treatment with EGF induced resistance to ponatinib in the LC-2/ad cells (Figure IV.16B), demonstrating that EGFR signaling may represent an early adaptive resistance mechanism capable of promoting cell proliferation. Stimulation of AXL with GAS6 was not able to induce resistance to ponatinib, nor did it enhance the resistance phenotype observed with EGF (Figure

IV.16B). Therefore, AXL signaling is likely not part of the immediate early adaptive response, but can contribute to acquired drug resistance after longer term cellular reprogramming.

Figure IV.16. LC-2/ad cells are primed to utilize AXL and EGFR signaling when RET is inhibited. (A) LC-2/ad cells were treated with 10 nM ponatinib for the 0, 4, 24, and 48 hours. Lysates were analyzed via Western blotting with the antibodies indicated. (B) Cell viability of LC-2/ad cells treated with increasing doses of ponatinib plus vehicle control (blue), 100 ng/mL EGF (gold), 100 ng/mL GAS6 (pink), or 100 ng/mL EGF + 100 ng/mL GAS6 (green) for 72 hours. Error bars: mean ± SEM for 3 triplicate experiments (n=9).

Discussion

Relapse is universal in NSCLC patients treated with targeted therapies; understanding the molecular mechanisms that contribute to resistance, and subsequent

TKI failure, will facilitate the development of improved pharmacologic strategies and treatment options for these patients. In particular, RET-fusion positive NSCLC patients do not appear to benefit as much from TKI therapies as patients with other fusion-driven

NSCLCs. While the objective response rates (ORR) of patients with advanced ALK- or

ROS1-fusion positive NSCLC treated with crizotinib were 65% and 72%, respectively.170

RET-fusion positive patients with advanced disease treated with cabozantinib experienced an ORR of only 28%.164 This, underscores the need to better understand the basic biology of RET-fusions in NSCLC and the mechanisms of intrinsic and acquired resistance that limit patient responses. Here, we demonstrate that the multi-kinase inhibitor ponatinib is active in the LC-2/ad cell line, a pre-clinical patient-derived model of RET-driven NSCLC, and report on the development of two ponatinib-resistant derivatives with distinct mechanisms of resistance. Our decision to specifically study resistance to the RET-TKI ponatinib was guided by the low-nanomolar affinity with which ponatinib inhibits the RET kinase domain, its approval for use in patients with

CML, as well as the ongoing clinical studies assessing its efficacy in RET-positive

NSCLC. Further, a recent study in KIF5B-RET-dependent BaF3 cells demonstrated that the vandetanib-resistant RET G810A mutation exhibited increased sensitivity to ponatinib, concluding that ponatinib is the current “drug of choice” for targeted inhibition of RET in the clinic.266

The scarcity of RET-driven lung cancer cell model systems represents a significant limitation. Currently, the LC-2/ad cell line is the only publically-available patient-derived RET-fusion driven NSCLC cell line. The TPC1 PTC cell line, also utilized here, represents the only other patient-derived cancer cell line harboring a RET- fusion. Ba/F3 and NIH-3T3 cells engineered to express RET-fusions provide adequate proof-of-concept model systems to examine the oncogenicity and pharmacologic inhibition of the fusion protein, however they do not adequately recapitulate the biology of a human lung cancer cell, which is critical for studies on drug resistance. The derivation of additional patient-derived RET-fusion models is critical to further our understanding of resistance to RET inhibition and remains an ongoing priority for our laboratory.

Despite this limitation, the findings presented here may illuminate our understanding of resistance to RET TKIs in the clinic. We demonstrated that ponatinib is a RET inhibitor that inhibits cell proliferation in the LC-2/ad cell line with similar potency as previously reported in RET-dependent thyroid cancer models.259 We described the development of two ponatinib-resistant LC-2/ad derivatives, PR1 and PR2, which both maintain expression of the WT CCDC6-RET protein. Interestingly, RET phosphorylation was undetectable in either PR1 or PR2 cell lines, remaining absent even when cells were removed from ponatinib for 24 hours. Thus reinforcing the lack of requirement for RET signaling and suggesting that chronic RET inhibition can suppress phosphorylation independent of ponatinib’s ability to bind to the RET kinase domain; perhaps through increased expression of phosphatases or increased turnover of phosphorylated RET. We are not the first to report this phenomenon: dovitinib-resistant

LC-2/ad cells with Src activation were also reported to have lost phosphorylation of the

RET-fusion protein.267 This finding brings into question our fundamental understanding of fusion-kinase signaling. It is assumed that the dimerization-domain containing 5’- fusion partner drives constitutive expression, homodimerization, and activation of the fusion protein. However, our data suggests that expression alone is not sufficient for dimerization-mediated activation and that, perhaps, cellular reprogramming that suppresses fusion kinase activity (for example, by the up-regulation of phosphatases or increased turnover of the fusion protein) may regulate the oncogenicity of an expressed fusion kinase.

Using next-generation RNA sequencing we identified an oncogenic NRAS pQ61K mutation which constitutively activates MAPK and PI3K/AKT signaling, freeing the cell from RET-dependent proliferation and survival.268 While we were not able to detect this mutation in parental LC-2/ad or PR2 cells, we did not exhaust single-cell sequencing approaches to rule out the existence of a pre-existing population of parental

LC-2/ad cells harboring this mutation. We confirmed that PR1 cell proliferation is dependent upon NRAS p.Q61K expression by knocking down NRAS, and demonstrated that PR1 cells are more sensitive to the MEK inhibitor trametinib than parental LC-2/ad cells. Further, NRAS p.Q61K remains the dominant oncogene driving PR1 cell proliferation and survival even in the absence of chronic RET inhibition. While we were able to induce ponatinib resistance with exogenous expression of NRAS p.Q61K in TPC1 cells, we were unable to achieve expression of this oncogene in LC-2/ad cells. Oncogenic

RAS can promote apoptosis and senescence through a variety of downstream effector mechanisms including the activation of the MAPK pathway and JNK signaling,269,270

possibly explaining our failure to induce ponatinib-resistance via expression of NRAS p.Q61K in the LC-2/ad cell line.

The emergence of oncogenic NRAS as the mechanism of resistance in the PR1 cell line buttresses a recent study in which NRAS was shown to induce resistance to

ROS1 inhibition.113 It is interesting that mutations in NRAS, and not KRAS, are increasingly being identified as mechanisms of resistance to TKI therapy in NSCLC, despite the fact that KRAS is a far more common primary oncogenic driver of NSCLC.271

Assessment of NRAS mutational status in post-progression, drug-resistant patient tumor samples may provide insight into the clinical significance of this finding. The requirement for RET-inhibition in order for NRAS-expressing TPC1 cells to exhibit resistance to ponatinib and for NRAS to be activated suggests that oncogenic signaling in these cells is tightly regulated and that RET-driven cells must be forced to undergo an

“oncogene switch” in order to transition to a RAS-driven state. This is not without precedent: pharmacologic inhibition of ROS1 was required for a novel activating KIT mutation to acquire control of oncogenic signaling in NSCLC cell lines HCC78 and

CUTO-2 and induce resistance to crizotinib,272 and loss of EGFR T790M mutation has been shown to coincide with acquisition of MET amplification as a mechanism of resistance to second-generation EGFR inhibitors.273 Considering the current lack of pharmacological RAS inhibitors, the emergence of a RAS-dependent resistance mechanism in vitro provides a rationale for the evaluation of upfront combination MEK and RET TKI therapy in RET-positive NSCLC patients to prevent RAS-dependent resistance. In fact, in both ALK- and EGFR-driven NSCLC, upfront combination

ALK/MEK or EGFR/MEK inhibition has been shown to prevent the emergence of

MAPK pathway-addicted resistance mechanisms.274,275

Although next-generation sequencing failed to identify any oncogenic mutations in the PR2 cell line, expression of EGFR and AXL as well as their ligands, HBEGF,

NRG1, and GAS6, were increased. We demonstrated that PR2 cells are more sensitive to

EGFR inhibition compared to the LC-2/ad cell; and while AXL inhibition did not decrease cell proliferation, it did enhance sensitivity to EGFR inhibition. While we did not rule out the possibility that other ERBB isoforms activated by NRG1 were contributing to resistance in the PR2 cells, we did determine that EGF-dependent EGFR signaling regulates MAPK pathway signaling in PR2 cells, suggesting that EGFR signaling is sufficient to promote proliferation in the PR2 cells. Conversely, AXL primarily signaled through the pro-survival PI3K/AKT pathway. Interestingly, stimulation of EGFR, but not AXL, signaling in LC-2/ad cells increased activation of downstream signaling, suggesting that RET-driven cells are primed to utilize EGFR to regulate oncogenic signaling. Here, we have shown that upregulation of both wild-type

EGFR and wild-type AXL signaling contribute cooperatively to the resistance phenotype observed in the PR2 cell line. The differential sensitivities to single agent EGFR or AXL inhibition may be due to the specific oncogenic signaling pathways they regulate. PR2 cell proliferation is sensitive to EGFR inhibition due to the loss of pro-proliferative signaling elicited by decreased MAPK pathway activation. Conversely, AXL appears to signal primarily through PI3K/AKT, which is a key regulator of cell survival and anti- apoptotic functions.

Individually, both EGFR and AXL are well-documented mediators of acquired resistance in NSCLC. Activation of EGFR signaling is a common bypass signaling mechanism responsible for acquired resistance to targeted therapies in fusion-driven

NSCLC226,276-279 and AXL mediates resistance via canonical upregulation of downstream signaling and by inducing EMT.120,121,280-283 Here, we demonstrate that AXL and EGFR independently regulate downstream signaling through PI3K/AKT and MAPK, respectively, and together contribute to the acquired resistance phenotype seen in the PR2 cells.

The activation of wild-type signaling in the PR2 cells provides a glimpse into how

RET-dependent cells may be primed for resistance to targeted therapies. Parental LC-2/ad cells express low levels of both EGFR and AXL, though they do not rely on either of these RTKs for regulation of oncogenic signaling. LC-2/ad cells with higher EGFR and/or AXL expression may be over-represented amongst early persisters of RET- inhibition and further, cells that are able to transition to EGFR- and/or AXL-driven oncogenic signaling are more likely to survive and proliferate amidst RET inhibition. The partial restoration of RET-dependence in PR2-OP cells further supports the notion that

RET-driven lung cancer cells can flexibly rely on different oncogenic drivers. From a clinical standpoint, this suggests that a brief hiatus from RET TKI therapy in patients with acquired resistance due to the activation of WT RTKs may result in the restoration of drug-sensitivity.

Few studies have been performed with the aim of understanding RET biology in the context of NSCLC and it is possible that RET-driven NSCLC cells may be primed for particular modes of resistance to targeted therapies. Here, we demonstrate that LC-2/ad

cells are capable of circumventing RET inhibition via reactivation of the RAS/MAPK pathway via the acquisition of oncogenic NRAS p.Q61K or upregulation of signaling through wild-type receptors expressed in ponatinib-naïve LC-2/ad cells. These studies also underlie the need for improved proteomic tools to detect signaling dependence on wild type RTKs, rather than relying solely on the detection of acquired drug resistance through gene mutations. Understanding acquired mechanisms of resistance to RET- inhibitor therapy in vitro may help design improved combination therapy strategies that could improve patient outcomes by preventing or delaying drug resistance.

CHAPTER V

PRE-CLINICAL ASSESSMENT OF COMBINATION MTOR AND RET

INHIBITORS IN RET FUSION POSITIVE NSCLC

Introduction

Oncogenic rearrangements of the gene encoding the RTK RET have been identified in NSCLC, papillary thyroid cancer (PTC), and colorectal cancer.253

Approximately 1-2% of NSCLC patients harbor a RET fusions, which now account for as many as 20% of lung cancers of never-smokers in whom no other known NSCLC-driving mutations have not been identified.24,26,28 These chromosomal rearrangements link the intracellular 3’-RET kinase domain to the 5’-dimerization domain of an unrelated gene

(most commonly CCDC6, KIF5B, or NCOA4),254 resulting in constitutive expression of the RET fusion protein, homodimerization, and ligand-independent activation of pro- survival and pro-proliferation signaling.

RET inhibitors underperform TKI therapies employed in other fusion-positive

LACs.

RET fusion positive LAC patients do not appear to benefit as much from currently available TKI therapies as patients with other fusion-driven lung tumors, with responses rates ranging from 18% - 47%; patients with advanced ALK or ROS fusion positive LAD treated with crizotinib were 65% and 72%, respectively.169,170 There are many possible explanations for this discrepancy. First, most clinically available RET

TKIs are multi-kinase inhibitors repurposed for use in RET- positive patients and

therefore, accompanying therapeutic RET-inhibiting doses in patients is likely limited by side effects mediated by off-target effects.

Vandetanib is a multi-kinase inhibitor originally developed as a VEGFR inhibitor which also nanomolar affinity for the kinase domains of EGFR and RET. Vandetanib is currently approved for use in patients with advanced medullary thyroid cancer (which is primarily driven by oncogenic RET signaling) and has been shown to be effective at inhibiting RET-dependent cell proliferation in vitro in LC-2/ad cells.258 However, studies of RET fusion positive LAC patients treated with vandetanib have thus far yielded mixed results. Multiple case reports have been published describing clinical responses to vandetanib in RET fusion positive LAC patients however, a retrospective analysis of

RET copy number gain among unselected NSCLC patients treated with vandetanib did not reveal a differential benefit for this patient population.284-286 Further, recently released data from phase II studies of vandetanib are similarly conflicting with the LURET trial reporting 9 of 19 (47%) RET fusion positive patients achieving an objective response, while the RET GLORY trial reported an objective response rate of only 18% (2 of 11 patients) among RET positive patients treated with vandetanib.165,287 The large discrepancy in these two studies can be attributed to the low number of patients and large confidence intervals.

Developing more effective pharmacologic strategies for treating RET fusion- positive LAC patients is a priority, with research efforts focused on developing more specific RET TKIs and investigating combination therapies that enhance the potency of

RET inhibitors. Currently, only one RET-selective TKI is in Phase I clinical trials,

RXDX-105.171,288 However, the vast array of targeted therapies approved over the last

decade provides a plethora of pharmacologic candidates to investigate as candidates for combination therapies with RET inhibitors. The focus of this chapter will be on investigating one such pharmacologic combination featuring the mTOR inhibitor everolimus.

Everolimus.

The serine/threonine kinase mTOR is a downstream effector of the PI3K/AKT signaling pathway that regulates a diverse array of cellular processes including macromolecule biosynthesis, autophagy, cell cycle progression, growth, metabolism, cytoskeletal organization, and cell survival. Everolimus is a rapamycin analog that indirectly inhibits mTORC1 signaling by binding and forming a complex with FK- binding protein 12 (FKBP12) which can then bind mTOR near its kinase domain and inhibit catalytic activity. Interestingly, for reasons that remain unclear the everolimus-

FKBP12 complex only inhibits mTOR activity when it is bound to RAPTOR as part of the mTORC1. In the absence of mTOR kinase activity, signaling through S6K1 and its substrates RPS6 and 4E-BP1 is abrogated preventing ribosomal biogenesis and formation of the pre-initiation translation complex (eIF4F) required for cap-dependent mRNA translation.

Combination RET and mTOR inhibition significantly decreases tumor burden in

RET fusion positive NSCLC.

Everolimus is orally available and currently approved for use in renal cell carcinoma and as an immunosuppressive agent in cardiac and renal transplant patients98 and has also been shown to be moderately effective as a single agent in late-stage LAC patients previously treated with chemotherapy or an EGFR TKI.289 Given the prominent

role the PI3K/AKT/mTOR signaling axis plays in regulating cancer cell survival a phase

1 trial of combination vandetanib and everolimus therapy was initiated by colleagues at

M.D. Anderson (NCT01582191). While enrollment in this study has not been limited to

RET-fusion positive patients, preliminary data does suggest that this patient population is disproportionately benefiting from the combination therapy with 5 out of 6 RET positive patients achieving a partial response, one of whom had previously progressed while being treated with the RET-inhibitor cabozantinib and one of whom experienced tumor regression of brain metastases.290,291 Pre-clinical studies assessing this pharmacologic combination are limited, and it is not fully understood why RET-patients appear to respond more favorably to TKI therapy when combined with the mTOR inhibitor everolimus. I therefore set out, in collaboration with our colleagues at M.D. Anderson, to perform an in vitro characterization of combination vandetanib and everolimus therapy in patient-derived RET-driven cancer cell lines. The findings of which are presented in this chapter.

Results mTOR inhibition decreases proliferation of RET-dependent cancer cells, but does not increase sensitivity to vandetanib.

Considering the promising clinical results thus far seen in RET positive NSCLC patients treated with combination vandetanib and everolimus, I hypothesized that everolimus may increase the sensitivity of RET-dependent cancer cells to targeted RET inhibition. In order to assess whether everolimus increases the potency of vandetanib, I treated LC-2/ad and TPC1 cells, which both harbor the CCDC6-RET fusion, with

increasing concentrations of vandetanib plus DMSO or a fixed dose of 1 nM or 10 nM everolimus for 72 hours. Interestingly, low nanomolar concentrations of everolimus are in and of themselves capable of inhibiting cell proliferation in both LC-2/ad and TPC1 cells, but did not decrease the IC50 of vandetanib in either cell line (Figure V.1A). In order to confirm that the mTOR inhibitory properties of everolimus were responsible for

Figure V.1: mTOR inhibition decreases cell proliferation in RET-dependent cell lines, but does not sensitize cells to the RET inhibitor vandetanib. (A) Cell viability of LC-2/ad and TPC1 cells treated with increasing doses of vandetanib plus DMSO (dark blue), 1 nM everolimus (light blue), or 10 nM everolimus (orange) for 72 hours at which point cell viability was measured using the CyQuant Direct Cell Proliferation Assay. (B) Cell viability of LC-2/ad and TPC1 cells treated with increasing doses of vandetanib plus DMSO (dark blue), 10 nM AZD8055 (light blue), or 100 nM AZD8055(orange) for 72 hours at which point cell viability was measured using the CyQuant Direct Cell Proliferation Assay. Data are the mean of one triplicate experiment (A, B).

the decrease in cell proliferation observed as a single agent, I repeated this experiment with the mTOR inhibitor AZD8055, which also decreased cell proliferation as a single agent but did not enhance sensitivity to vandetanib (Figure V.1B).Unlike everolimus and other rapamycin analogues, AZD8055 inhibits mTOR activity by directly binding the

ATP-binding pocket of the mTOR tyrosine kinase domain. Therefore, because everolimus and AZD8055 inhibit mTOR via distinct mechanisms, it is unlikely that the decrease in cell proliferation observed is due to an off-target effect of either agent.

Everolimus enhances the inhibition of colony formation in RET-dependent cells by vandetanib.

Clonogenicity, or the potential for a single, isolated cancer cell to proliferate indefinitely, is an oncogenic property distinct from the proliferation of a population of cells. In order to test whether everolimus enhanced the degree to which RET-inhibition A. B. everolimus (nM) n.s. DMSO 0.1 1 100

(% DMSO) (% * 50 DMSO

Colony area area Colony 0 everolimus (nM): 0 0.1 1 0 0.1 1

500 nM DMSO 500 nM vandetanib vandetanib Figure V.2: Addition of everolimus to vandetanib enhances inhibition of colony formation in TPC1 cells. TPC1 cells were plated at 100 cells/well in a 6 well plate. After 24 hours, media was removed and replaced with media containing DMSO, vandetanib, everolimus, or vandetanib plus everolimus at the concentrations indicated above. Drug-containing media was replaced every 72 hours for 14 days, at which point cells were fixed, stained with a crystal violet solution, imaged, and the total area of each well occupied by colonies was quantified. Data are the average of three independent experiments (A) or representative images from one experiment (B). Error bars are SEM and statistics are a Student’s t-test (*P<0.05). 96

with vandetanib inhibited the oncogenicity of RET-dependent cancer, I performed a

clonogenic assay in which TPC1 cells were seeded at very low density and treated with

500 nM vandetanib, 0.1 or 1 nM everolimus, or a combination of vandetanib and

everolimus for 14 days. Not surprisingly, vandetanib alone inhibited the clonogenicity of

TPC1 cells, however unlike what was observed with cell proliferation, single agent

everolimus did not decrease clonogenicity (Figure V.2). In combination, however,

everolimus did significantly enhance the degree to which vandetanib inhibited colony

formation of TPC1 cells (Figure V.2). Due to the morphology of the colonies generated

by LC-2/ad cells, clonogenicity was unable to be evaluated.

A. B.

LC-2/ad TPC1 LC-2/ad vandetanib – + – + – + – + ponatinib – + – + everolimus – – + + – – + + everolimus – – + + P-ERK1/2 P-ERK1/2 ERK1/2 ERK1/2 P-AKT P-AKT AKT AKT P-P70S6K P-P70S6K P70S6K P70S6K P-S6RP P-S6RP S6RP S6RP GAPDH GAPDH Figure V.3: Addition of everolimus to vandetanib results in maximal inhibition of MAPK and mTOR signaling. (A) Western blot analysis of LC-2/ad and TPC1 cells were treated with DMSO, 500 nM vandetanib, 1 nM everolimus, or 500 nM vandetanib and 1 nM everolimus for 2 hours. (B) Western blot analysis of LC-2/ad cells treated with DMSO, 10 nM ponatinib, 1 nM everolimus, or 10 nM ponatinib and 1 nM everolimus for 2 hours. Lysates were analyzed with the antibodies indicated.

Addition of everolimus to vandetanib results in maximal inhibition of signaling through RAS/MAPK and the PI3K/AKT/mTOR axis .

In order to understand the molecular mechanisms explaining the clinical efficacy of combination vandetanib and everolimus treatment in patients, I treated LC-2/ad and

TPC1 cells with vandetanib, everolimus, vandetanib plus everolimus or a DMSO control for two hours and then assessed phosphorylation of downstream signaling proteins by

Western Blot (Figure V.3A). Not surprisingly, vandetanib decreased phosphorylation of

RET and ERK1/2; this was not seen with everolimus treatment alone. Vandetanib also decreased phospho-AKT, which interestingly was increased by everolimus. This increase in phospho-AKT is not without precedent, as mTOR inhibition has been demonstrated to result in paradoxical upregulation of phospho-AKT due to the loss of mTOR-mediated inhibition of IRS1, which signals through AKT.292 Everolimus significantly inhibited phosphorylation of mTORC1 effectors p70S6K and S6RP, whereas single agent vandetanib only partially decreased phosphorylation of these proteins. Consequently, combination vandetanib and everolimus was the only condition where maximal inhibition of phospho-ERK1/2, phospho-p70S6K, and phospho-S6RP was observed. When the

500 nMvandetanib + 500 nMvandetanib 1 nMeverolimus 1 nMeverolimus

0 15” 2’ 8’ 24’ 48’ 0 15” 2’ 8’ 24’ 48’ 0 15” 2’ 8’ 24’ 48’ P-S6RP S6RP GAPDH Figure V.4: Combination vandetanib and everolimus maintains suppression of S6RP compared to vandetanib or everolimus alone. Western blot analysis of LC-2/ad cells were treated with 500 nM vandetanib, 1 nM everolimus, or 500 nM vandetanib and 1 nM everolimus for the time course indicated. Protein lysates were analyzed with the antibodies indicated. 98

same experiment was performed using ponatinib instead of vandetanib, the same patterns in protein expression were observed (Figure V.3B).

In order to assess the long-term effects of combination vandetanib plus everolimus on downstream signaling, I performed a time-course experiment on LC-2/ad cells treated with vandetanib, everolimus, or vandetanib plus everolimus (Figure V.4).

Interestingly, everolimus did decrease phospho-S6RP initially, however after 48 hours it appears to be somewhat restored. However, in cell treated with vandetanib and everolimus phospho-S6RP remained decreased at the end of the experiment, suggesting that the combination of both drugs is more effective at fully suppressing mTORC1 activity (Figure V.4).

Discussion

Unlike other fusion-driven LACs which respond well to targeted inhibition of the dominant fusion oncogene, RET fusion positive LAC patients do not appear to be benefitting from TKI therapy to the same degree. As discussed more thoroughly in

Chapter 3, this may in part due to differences in the biology of RET-driven lung cancer cells such that they are not as susceptible to RET-inhibition as other oncogene-driven cancers. One approach to overcoming lack of treatment response in RET positive patients is the development of rational combination therapies which pair other available targeted agents with RET TKIs.

Our collaborators at M.D. Anderson are already employing this clinical strategy in lung cancer patients and have preliminarily shown that RET positive LAC patients disproportionately benefit from the combination therapy with the mTOR inhibitor

everolimus and the RET TKI vandetanib. The work presented here represents an initial attempt to characterize this drug combination in patient-derived RET-fusion positive pre- clinical models in order to tease out the mechanism via which this treatment strategy is effecting tumor responses in patients. While I hypothesized that proliferation of RET- dependent cancer cells would be sensitized to the RET-inhibitor vandetanib in the presence of everolimus, this was not what I found. Instead, mTOR inhibition alone, either with everolimus or the mTOR kinase inhibitor AZD8055, was capable of inhibiting cell proliferation but this did not increase the potency of vandetanib. Interestingly, the converse was true with regards to clonogenicity: TPC1 cells colony formation was not inhibited by everolimus as a single agent, but the addition of everolimus to RET- inhibiting concentrations of vandetanib did decrease colony formation compared to vandetanib alone. With regards to protein activation, vandetanib as a single agent inhibited phosphorylation of ERK1/2 and AKT, as well as partially decreasing phosphorylation of mTORC1 effects p70S6K and S6RP—suggesting that in addition to canonical RAS/MAPK and PI3K/AKT signaling, oncogenic RET also contributes to the activation of the mTOR pathway. The complete loss of phospho-p70S6K and phospho-

S6RP with everolimus alone likely explains the decrease in cell proliferation observed in with single-agent everolimus treatment. Further, while preliminary, my finding that combination everolimus and vandetanib sustains the inhibition of mTORC1 signaling more effectively than either agent alone is a plausible, yet underdeveloped, mechanistic explanation for the success of this combination treatment strategy in RET positive lung cancer patients.

100

Together, the findings I present in this chapter reaffirm the critical importance of the PI3K/AKT/mTOR signaling axis in the context of a RET-dependent cancer cell growth and survival. While these findings are preliminary, they are consistent with previous observations in RET-driven thyroid cancer models that RET-dependent oncogenicity is primarily dependent upon activation of the PI3K/AKT/mTOR signaling axis and that mTOR inhibition is viable therapeutic strategy in medullary thyroid cancer patients.144,293 While I did not observe any obvious or dramatic synergism with the combination of vandetanib and everolimus in vitro, I only assessed cell proliferation and clonogenicity. It is possible that such a phenotype may emerge in pre-clinical models assessing other hallmarks of cancer. Further, it is possible that other targeted inhibitors of the PI3K/AKT pathway may also produce the same clinical benefit in RET positive patients when used in concert with RET TKIs. Future experiments that assess vandetanib combined with targeted inhibition of different PI3K/AKT pathway components will address the generalizability of the clinical benefit seen with combination everolimus and vandetanib.

One clear limitation of the findings presented here is the lack of model systems in which the combination therapy was investigated. As discussed in Chapter 3, the lack of publically available patient-derived RET-driven LAC cell lines clearly restricts the findings that can be drawn from these investigations. More critically, the only available

RET-driven cancer cell lines express CCDC6-RET not the more common, and potentially more aggressive,165 KIF5B-RET fusion. While several proprietary patient-xenograft models have been developed commercially, there has been no attempt to generate cell lines from these models that may benefit the wider academic community.

101

Rationally designed combination targeted therapy may present one approach to overcoming the lack of response to single agent RET inhibition currently observed in

RET positive LAC patients. Vandetanib and everolimus present one such possible combination treatment strategy demonstrating therapeutic potential in RET positive LAC patients treated in phase I studies. Here I demonstrate that maximal downstream inhibition of the PI3K/AKT/mTOR axis may contribute to the potency of vandetanib and everolimus when used in concert. Together, these results support the continued investigation of this combination treatment strategy in the clinical setting and provide preliminary insights into the mechanism via which they induce tumor regression in RET positive LAC patients.

102

CHAPTER VI

DISCUSSION

Summary of Findings and Key Conclusions

The rapid increase in targeted TKI therapies approved for use in cancer patients or in late-stage clinical trials has revolutionized the treatment of solid tumors over the last decade. However, in the wake of this paradigm shift many questions regarding the fundamental biology of oncogenic and wild-type RTKs has been raised. Particularly with regards to how resistance to TKI therapies occurs. The work presented above addresses some of these questions. First, in Chapter III I summarized the findings of work performed in the laboratory of Dr. Amy Keating exploring the role of wild-type TAM

RTKs in the regulation of glioblastoma cell migration. And secondly in Chapter IV, and

Chapter V, I investigated the use of targeted inhibitors in RET-fusion positive LAC.

Specifically, in Chapter IV, I demonstrated that the TKI ponatinib is a potent inhibitor of

RET-fusions in a patient-derived RET-fusion positive LAC model and described the development of two, distinct ponatinib-resistant LAC cell lines with acquired resistance mechanisms that re-activate RAS/MAPK signaling. In Chapter V, I explored the molecular effects of combination mTOR and RET inhibitor treatment in patient-derived

RET-fusion driven cancer cell lines. Taken together, my findings highlight the complexities of RTK signaling in oncogene-driven cancers and provide insight into therapeutic strategies that will improve overall survival in these patient populations.

103

MERTK inhibition is a promising pharmacologic strategy for the inhibition glioblastoma cell migration and invasion.

Taken together, the data presented in Chapter III provides pre-clinical evidence supporting the hypothesis that MERTK inhibition is a viable strategy for inhibiting cell migration and invasion in GBM. First, I demonstrated that stable genetic knockdown of

MERTK decreases GBM cell invasion in a three-dimensional collagen invasion model, further enhancing the clinical relevance of the MERTK-dependent migration phenotype originally described in two-dimensional cell culture conditions by the Keating Lab.221

Next, using a clinically available multi-kinase inhibitor, foretinib, I demonstrated that pharmacologic inhibition of MERTK also inhibits cell migration and invasion in two- and three-dimensional pre-clinical in vitro models. Using the MERTK-specific TKI,

UNC2025, I provided additional evidence that MERTK regulates glioblastoma cell migration thus buttressing the conclusion that the decrease in cell migration and invasion observed with foretinib is likely due to MERTK inhibition. Finally, my finding that FAK protein expression is decreased following siRNA knockdown of MERTK provides a hint of a possible mechanistic explanation as to how MERTK may regulate GBM cell migration, which will be discussed in greater detail below.

Resistance to RET-inhibition in RET-rearranged LAC is mediated by reactivation of RAS/MAPK signaling.

In Chapter IV, I demonstrated that ponatinib is a potent RET TKI capable of inhibiting cell proliferation and oncogenic signaling in the patient-derived RET fusion positive LC-2/ad LAC cell line. Similar to previous findings in pre-clinical models of

RET-dependent thyroid cancer, ponatinib inhibited RET phosphorylation at low

104

nanomolar concentrations resulting in the disruption of signaling through the

RAS/MAPK and PI3K/AKT signaling pathways and inhibition of cell proliferation. I then described the development of two, distinct ponatinib-resistant LC-2/ad derivatives,

PR1 and PR2. DNA sequencing confirmed that both PR1 and PR2 cells retained expression of wild-type CCDC6-RET, thus ruling out the possibility that a mutation in the RET kinase domain was responsible for either resistance phenotype. Interestingly,

PR1 and PR2 cells both retained expression, but not phosphorylation of the RET fusion protein and phosphorylated RET remained absent even when PR1 and PR2 cells were removed from ponatinib-containing media for 24 hours. Considering that ponatinib is a reversible, non-covalent inhibitor RET this period time is sufficient for drug wash out.

This not only reinforced the lack of requirement for RET signaling in either the PR1 or

PR2 cell lines but also raised questions as to the basic biologic properties of RET fusion proteins. It is widely assumed that the dimerization domain of the 5’-fusion partner drives the constitutive expression, dimerization, and activation of kinase fusions, therefore it is unclear as to the mechanism by which the RET fusion protein can be expressed, but not activated in these cells in the absence of pharmacologic inhibition.

Next-generation RNA sequencing of PR1 cells identified an oncogenic NRAS p.Q61K mutation not present in either the parental LC-2/ad cells or the PR2 derivative, which was confirmed using DNA sequencing. I confirmed that PR1 cell proliferation is dependent upon NRAS p.Q61K expression by knocking down NRAS, and demonstrated that PR1 cells are more sensitive to the MEK inhibitor trametinib than parental LC-2/ad cells—suggesting PR1 cell proliferation exhibits increased dependence upon

RAS/MAPK signaling than the parental LC-2/ad cells. Further, NRAS p.Q61K remains

105

the dominant oncogene driving PR1 cell proliferation and survival even in the absence of chronic RET inhibition. Next-generation sequencing of the PR2 cells did not reveal the acquisition of any known oncogenic mutations, but did reveal upregulation of the RTKs

EGFR and AXL, as well as their ligands HBEGF, NRG1, and GAS6, which led to the hypothesis that resistance in the PR2 cell line was being driven by the upregulation of wild-type EGFR signaling. I determined that PR2 cell proliferation is more sensitive to

EGFR inhibition than the proliferation of the parental LC-2/ad cells and that EGFR inhibition in the PR2 cells disrupts signaling through the RAS/MAPK and PI3K/AKT signaling pathways, whereas single-agent EGFR inhibition has no effect on the activation of these pathways in parental LC-2/ad cells. Interestingly, while AXL inhibition alone did not inhibit PR2 cell proliferation, it did enhance sensitivity to EGFR inhibition— suggesting that AXL signaling contributes cooperatively to the EGFR-driven resistance phenotype in the PR2 cells. Further, when PR2 cells were stimulated with the AXL ligand GAS6 I observed activation of the PI3K/AKT signaling pathway which was not did not occur in LC-2/ad cells also treated with GAS6. Finally, I demonstrated that the

EGFR ligand, EGF, was capable of inducing ponatinib resistance in LC-2/ad cells and that RET inhibition with ponatinib induced expression of AXL in parental LC-2/ad cells, thus priming them for the acquisition of EGFR/AXL driven resistance.

Interestingly, in the absence of chronic RET inhibition, RET phosphorylation was restored in both the PR1 and PR2 cells. However, only the PR2 cells were able to partially restore RET-dependent cell proliferation, whereas the PR1 cells retained a strictly NRAS-dependent phenotype—suggesting that in the context of a resistance phenotype driven by wild-type RTK signaling, RET-dependent LAC cells gain a degree

106

of flexibility as to the source of oncogenic signaling. It is tempting to speculate that this may be especially relevant in vivo, where the tumor stroma produces a milieu of growth factors capable of stimulating wild-type RTKs expressed by a cancer cell. If this proves to be a property common among RET-dependent LAC, it may provide insight into the inherent resistance currently being observed in RET positive LAC patients treated with single agent TKI therapies.

Preliminary results from a phase 1 trial of combination vandetanib and everolimus in patients with oncogene driven LAC suggest that RET fusion positive patients are disproportionately benefiting from this combination treatment strategy. In Chapter V, I presented my findings from experiments characterizing combination drug treatment with the RET inhibitor vandetanib and the mTOR inhibitor everolimus in patient-derived RET fusion positive cancer cell lines.

While I did not observe increased sensitivity to the RET inhibitor vandetanib when combined with a fixed dose of everolimus, I did determine that RET-dependent cell proliferation is inherently sensitive to single-agent mTOR inhibition. Further, the addition of everolimus to vandetanib enhanced the degree to which clonogenicity of RET- dependent cells was inhibited by vandetanib alone. Immunoblot analysis of expression and phosphorylation of downstream signaling components suggested that RET inhibition alone does partially decrease mTOR signaling as evidenced by a moderate inhibition of phospho-p70S6K and phospho-S6RP. However, the addition of the mTOR inhibitor

107

everolimus to vandetanib results in the complete suppression of mTOR signaling in addition to maximal inhibition of the RAS/MAPK pathway. Further, S6RP phosphorylation was suppressed to a greater degree and for a longer period of time with the combination of vandetanib and everolimus than was observed with either agent alone.

Future Directions

Is MERTK inhibition a viable therapeutic strategy in vivo?

In vivo, the glioblastoma cell migration and invasion responsible for tumor infiltration takes place in a three-dimensional environment vastly more complex than the conditions simulated in the two- and three-dimensional in vitro models utilized in the work presented here. Cell grown in two-dimensional cultures display altered morphology, cell-cell interactions, and gene expression when compared to their counterparts grown in three-dimensional substrates. While invasion in a three-dimensional collagen matrix is a more translationally relevant assay for GBM infiltration than the conditions recapitulated by a plastic tissue culture dish, it remains a poor representation of the extra-cellular matrix components found in the brain parenchyma.

A more appealing approach would be to study glioblastoma cell biology in the context of the normal brain architecture and extracellular matrix constituents using an organotypic brain slice culture model. Brain slices 100-300 μm in thickness can be cultured in vitro for four to six weeks before exhibiting signs of necrosis294 and when co- cultured with glioblastoma cells or neurospheres allow for invasion to be studied in the context of normal brain architecture, vasculature, and functional neuronal and supporting cell.294-297 Studying the effects of MERTK inhibition on glioblastoma cell migration and

108

invasion using this model is an appropriate next step to understanding the translational the findings presented here. Further, obtaining brain slice cultures from an intracranial glioblastoma mouse model treated with MERTK inhibitors such as foretinib and

UNC2025 may shed light on how long term MERTK inhibition in vivo changes the invasion phenotype in GBM.

How does FAK contribute to MERTK-dependent GBM migration and invasion?

FAK is overexpressed in many cancers, including glioblastoma, and has been shown to correlate with invasive disease.298,299 Further, pharmacologic inhibition of FAK autophosphorylation has also been shown to decrease viability and sensitize glioblastoma cells to temozolomide in vitro.300 Interestingly, in Chapter III, I demonstrated that total

FAK expression is decreased following siRNA knockdown of MERTK, in a manner which correlated with the decreased migration and invasion phenotype that I also observed. However, this is far from a conclusive demonstration that the decrease in FAK expression observed with siRNA knockdown of MERTK is responsible for the decrease in cell migration and invasion seem with pharmacologic MERTK inhibition. Therefore, further investigation into how FAK-signaling contributes to MERTK regulation of cell migration and invasion is warranted.

First, FAK’s contribution to the overall regulation of glioblastoma cell migration and invasion must be evaluated in the absence of MERTK inhibition. siRNA knockdown and pharmacologic inhibition of FAK would be appropriate experimental strategies to evaluate the manner in which FAK regulates two- and three-dimension GBM cell migration and invasion. FAK has been shown to be a critical regulator of cell migration and invasion in a wide variety of cellular contexts, therefore I would hypothesize that loss

109

of FAK expression and/or signaling would decrease GBM cell migration and invasion.

These experiments do not, however, contribute to our understanding of FAK as a downstream component of MERTK signaling. If FAK is a critical mediator of MERTK- dependent GBM cell migration and invasion, then overexpression of FAK in the context of MERTK inhibition or knockdown should rescue the cell migration phenotype.

By what mechanism are EGFR and AXL activated in the PR2 cell line?

My findings regarding the PR2 cell line in Chapter IV demonstrate that activation of wild-type EGFR and AXL signaling can mediate acquired resistance to ponatinib.

However, I have demonstrated mechanistically how EGFR and AXL signaling is activated in these cells. Considering that transcript expression of EGFR ligands NRG1 and HBEGF as well as the AXL ligand GAS6 are all increased in PR2 cells, my hypothesis is that EGFR and AXL signaling is being activated in an autocrine fashion as a consequence of increased ligand expression and/or secretion.

To assess the role of autocrine activation of EGFR and/or AXL in mediating the acquired resistance phenotype seen in the PR2 cells, I would first assess whether conditioned media from PR2 cells can confer resistance to ponatinib in the LC-2/ad cells.

My finding that treatment of LC-2/ad cells with EGF induces resistance to ponatinib indicates that LC-2/ad cells are programmed to utilize EGFR signaling to regulate cell proliferation, therefore if PR2 cells are actively secreting EGFR ligands, I would expect conditioned media from the PR2 cells to induce resistance in parental LC-2/ad cells in a cell proliferation. Further, if the induced resistance is truly due to increased autocrine activation of EGFR, I would also expect PR2 conditioned media to transiently increase phosphorylation of EGFR in serum starved parental LC-2/ad cells when compared to LC-

110

2/ad conditioned media. The next step would be to identify exactly which EGFR ligands are being secreted by PR2 cells. Based on the RNA sequencing data I would anticipate that of all the EGFR ligands, NRG1 and HBEGF are the most likely to be mediating this phenotype considering that their transcript expression is increased. Experimentally, this could be confirmed by geneticically knocking down expression of NRG1 and/or HBEGF in the PR2 cell line, and assessing for a shift in ponatinib sensitivity in a cell prolfieration assay. It would be critical to perform this experiment in PR2-OP cells, as EGFR inhibition only restored sensitivity to ponatinib in PR2 cells maintained in the absence of chronic RET inhibition. However, assessing EGFR ligand concentrations in PR2 media, via ligand-specific ELISA assays will be an important step in confirming that increased expression of EGFR ligands is resulting in autocrine activation of EGFR signaling.

In addition to increased ligand expression driving autocrine activation of EGFR signaling, is is possible that increased ectodomain shedding of ligands at the cell membrane is also contributing to this phenotype in the PR2 cells. Interestingly, increased

RNA expression of EGFR, AXL, HBEGF, NRG1, and GAS6 is not limited to PR2 cells as these proteins were also upregulated in PR1 cells, albeit to a lesser extent (Table IV.1 and Table IV.2). What is unique to the PR2 cell line, compared to parental LC-2/ad and

PR1 cells, is the upregulation of the ADAMs metalloproteinases ADAM9, ADAM10,

Table VI.1: Normalized gene expression of ADAMs proteins implicated in ectodomain shedding.

LC-2/ad PR1 PR1 PR2 PR2 Gene p-value p-value avg FPKM avg FPKM Fold Change avg FPKM Fold Change

ADAM9 25.38 39.89 1.57 n.s. 87.90 3.46 <0.005 ADAM12 0.45 0.73 1.61 n.s. 2.30 5.13 <0.05 ADAM10 24.55 32.69 1.33 n.s. 37.85 1.54 <0.05 ADAM17 9.48 10.30 1.09 n.s. 14.20 1.50 <0.05

111

ADAM12, and ADAM17 (Table VI.1)—all of which are known to contribute to EGFR transactivation and ectodomain shedding and which may represent an additional pharmacologic approach to reverse EGFR-mediated ponatinib resistance in RET positive

LAC.186

Is there an early persister state that emerges following targeted inhibition of RET in

RET positive LAC cells?

I have demonstrated that the resistance phenotype in the PR1 cell line is dependent upon oncogenic NRAS p.Q61K signaling. However, as mentioned in the previous section, increased transcript expression of EGFR and AXL is not exclusive to the PR2 cell line (Table S1), as these RTKs were also upregulated in PR1 cells. One might therefore speculate that both PR1 and PR2 cells evolved from an early persister state, with the PR1 cells subsequently acquiring the NRAS p.Q61K mutation and the PR2 resistance phenotype reflecting the outgrowth of dependence upon RTKs that support the maintenance of an early persister population following TKI exposure. Parental LC-2/ad cells express low levels of both EGFR and AXL, though they do not rely on either of these RTKs for regulation of oncogenic signaling. RET inhibition may simply select for

LC-2/ad cells that happen to express higher levels of EGFR and AXL at the time of pharmacologic inhibition and are therefore better able to survive following loss of downstream signaling generated by RET as the dominant oncogene. Conversely, RET signaling may transcriptionally or epigenetically down-regulate EGFR and/or AXL expression, such that when RET is inhibited expression of these two RTKs is increased.

In either case, it is tempting to speculate that increased expression of EGFR and AXL may allow cells to survive the loss of RET signaling long enough to acquire an

112

alternative oncogenic driver. In the case of the PR1 cell line, this would be the emergence

of oncogenic NRAS p.Q61K, whereas in the PR2 cell line this would be the transition to

EGFR- and/or AXL-dependent cell proliferation and survival.

Interestingly, I found preliminary evidence suggesting that expression of the

oncogenic transcription factor, Fra-1, was increased in PR2 cells compared to parental

LC-2/ad cells (Figure VI.1). Fra-1, has been shown to known to positively regulate both

AXL and EGFR; 301,302 further, sensitivity to the BET bromodomain inhibitor JQ1 in

LAC cells, has been shown to be mediated by suppression of Fra-1expression.

Interestingly, in addition to increased protein expression of Fra-1 in the PR2 cells; I also

observed a decrease in Fra-1 expression with JQ1 treatment in PR2 cells as well as

increased sensitivity to JQ1 compared to parental LC-2/ad cells as measured in a cell

proliferation assay (Figure VI.1). It would be logical therefore to further assess whether

JQ1 treatment or Fra-1 knockdown resulted in decreased expression of EGFR and AXL.

125 LC-2/ad PR2-IP Figure VI.1: PR2 cells are more 100 sensitive to JQ1 than LC-2/ad and 75 express increased Fra-1. (A) Cell viability of LC-2/ad (blue) and PR2

uorescence (%) uorescence ﬂ 50 cells (green) treated with increasing 25 doses of JQ1 for 72 hours at which point cell viability was measured

CyQuant CyQuant 0 -4 -3 -2 -1 0 1 using the CyQuant Direct Cell 10 10 10 10 10 10 Proliferation Assay. (B) Western blot JQ1 (uM) B. analysis of LC-2/ad and PR2 cells treated with 1 uM JQ1 for the time LC-2/ad PR2 points indicated. Lysates were analyzed with the antibodies JQ1: 0’ 2’ 6’ 24’ JQ1: 0’ 2’ 6’ 24’ indicated. Fra-1 Fra-1 GAPDH GAPDH

113

However, given that the PR1 cells have acquired a dependence upon oncogenic NRAS p.Q61K signaling, I would not expect PR1 cells to also exhibit increased sensitivity to

JQ1. It is tempting to speculate that Fra-1 knockdown may also decrease EGFR and AXL expression. Further, if increased Fra-1-dependent transcription is responsible for driving an EGFR and/or AXL-dependent persister state, one could logically postulate that the upfront addition of JQ1 to target RET inhibition may suppress the persister cell state, thereby preventing the emergence of a resistant cell population.

How is mTOR signaling regulated in RET fusion positive LAC?

Rationally designed combination treatment strategies may present the best therapeutic options for patients with RET positive LAC. In Chapter V, I describe a pre- clinical assessment of one such combination therapy, vandetanib plus the mTOR inhibitor everolimus. My finding that RET inhibition alone only partially suppresses mTOR signaling as measured via phosphorylation of mTOR signaling components p70S6K and

S6RP suggest that this pathway may be subject to activation via RET-independent mechanisms. Considering the degree to which RET-dependent cancer cells were observed to be sensitive to single-agent mTOR inhibition, elucidation of how mTOR signaling is upregulating and contributing to the oncogenic phenotype in these cells will likely inform improved treatment strategies in patients with RET-dependent cancers.

There is some evidence in the literature supporting the role of mTOR signaling in mediating sensitivity to targeted therapies in lung cancer. One study reported that among patients with EGFR mutant LAC treated with gefitinib, patients with co-mutations in the

PI3K/AKT/mTOR signaling pathway were significantly less likely to respond to EGFR

TKI therapy.303 Further, the degree to which gefitinib treatment decreases

114

phosphorylation of mTOR and p70S6K has been shown to correlated with intrinsic gefitinib sensitivity.304 In line with the findings presented here, targeted inhibition of

ALK with alectinib in ALK positive lung cancer cells has also been shown to insufficiently decrease phosphorylation of p70S6K with full suppression of mTOR signaling only observed when the mTOR inhibitor temsirolimus was used in combination with an ALK TKI.305 However, this study provides fails to ascertain why mTOR signaling is not fully suppressed by ALK inhibition other than to suggest it is merely a consequence of mTOR being an “indirect downstream target”. If this is the case, then combination therapy utilizing targeted PIK3CA inhibition may also produce similar decreases in RET-dependent cell proliferation as targeted mTOR inhibition. It is possible that not all RET driven LAC patients will benefit from combination mTOR and RET TKI therapies. Therefore, it is important to identify biomarkers associated with a positive clinical response to this treatment strategy. A starting point may be baseline phosphorylation of S6RP, which has already been shown to be correlated with poor survival in patients with lung cancer.306

Significance and Implications of This Work

Together, the work described here represents an investigation in to the use of tyrosine kinase inhibitors in glioblastoma and RET-rearranged lung cancer to curtail cell migration and invasion and improve oncogene-targeted therapy by understanding drug resistance. Glioblastoma and LAC are cancers with dismal prognosis which have seen some improvements in patient survival with the accelerated identification of novel oncogenic drivers and the parallel development of targeted therapeutic approaches.

115

Glioblastoma is a disease characterized by a profound ability of cancer cells to migrate and invade into the surrounding brain parenchyma, making full surgical excision of the tumor nearly impossible. The development of novel adjuvant treatment approaches to decrease cell migration and invasion prior to surgical intervention is therefore a logical approach to improving survival in this patient population. The work presented in Chapter

III presents a strong pre-clinical argument for further investigation and development of

MERTK inhibition as a promising therapeutic approach to inhibit the invasive phenotype of glioblastoma. Further, the TKI foretinib, which inhibits MERTK and crosses the blood-brain barrier, is currently in clinical trials for other cancer types. Considering the absolutely dismal prognosis that accompanies a GBM diagnosis, clinical investigation of foretinib in this patient population is warranted.

Unlike glioblastoma, patients diagnosed with LAC have benefited significantly from the surge in identification of oncogenic driver mutations and the accompanying development of novel targeted pharmacologic agents. However, there are still significant obstacles preventing durable and curative responses to targeted therapies—primarily newly emerging paradigms of intrinsic and acquired resistance. RET-rearranged LAC appears to be a particularly pointed example of an oncogenic driver that does not appear to confer a strong susceptibility to targeted therapies. Here, I present several compelling lines of evidence which support the further investigation of combination therapy strategies that may enhance the clinical effects of RET TKI inhibition and consequently delay the time to resistance and relapse or improve the degree and duration of initial tumor regression. In Chapter IV I describe two mechanisms of resistance to the RET TKI ponatinib, both of which rely upon the acquisition of alternative mechanisms of

116

RAS/MAPK pathway activation. Luckily, there are several clinically available TKIs currently being used in patients that could easily be combined with RET TKIs in order to prevent the outgrowth of RAS/MAPK-dependent resistance mechanisms such as the

MEK inhibitor trametinib and the EGFR inhibitors afatinib, erlotinib and gefitinib.

Intrinsic resistance to RET TKI therapies is also an increasingly recognized barrier to the successful use of targeted therapies in this patient population. And in Chapter V, I present pre-clinical evidence that strategies that increase suppression of mTOR signaling in conjunction with pharmacologic RET inhibition may increase upfront efficacy of targeted treatment approaches in RET positive patients.

Finally, while the findings presented here highlight the complexity of successfully exploiting molecular drivers of cancer, they also suggest that the pharmacologic tools necessary to overcoming these roadblocks are already clinically available. Well-designed studies assessing the benefits of rational, combination therapies are likely to accelerate our understanding of how best to treat oncogene-driven cancers and to dramatically improve the overall survival in this patient population.

117

REFERENCES

1. Stewart, B. W. & Wild, C. P. World Cancer Report 2014. Report No. 978-92-832- 0443-5, (International Agency for Research on Cancer, Lyon, France, 2014).

2. Siegel, R. L., Miller, K. D. & Jemal, A. Cancer statistics, 2016. CA Cancer J Clin 66, 7-30 (2016).

3. Chen, Z., Fillmore, C. M., Hammerman, P. S., Kim, C. F. & Wong, K. K. Non- small-cell lung cancers: a heterogeneous set of diseases. Nat Rev Cancer 14, 535- 546 (2014).

4. Langer, C. J., Besse, B., Gualberto, A., Brambilla, E. & Soria, J. C. The evolving role of histology in the management of advanced non-small-cell lung cancer. J Clin Oncol 28, 5311-5320 (2010).

5. Davidson, M. R., Gazdar, A. F. & Clarke, B. E. The pivotal role of pathology in the management of lung cancer. J Thorac Dis 5 Suppl 5, S463-478 (2013).

6. Toh, C. K., Gao, F., Lim, W. T., Leong, S. S., Fong, K. W., Yap, S. P. et al. Never-smokers with lung cancer: epidemiologic evidence of a distinct disease entity. J Clin Oncol 24, 2245-2251 (2006).

7. Siegel, R. L., Jacobs, E. J., Newton, C. C., Feskanich, D., Freedman, N. D., Prentice, R. L. et al. Deaths Due to Cigarette Smoking for 12 Smoking-Related Cancers in the United States. JAMA Intern Med 175, 1574-1576 (2015).

8. Lee, P. N. & Forey, B. A. Indirectly estimated absolute lung cancer mortality rates by smoking status and histological type based on a systematic review. BMC Cancer 13, 189 (2013).

9. Parkin, D. M., Bray, F., Ferlay, J. & Pisani, P. Global cancer statistics, 2002. CA Cancer J Clin 55, 74-108 (2005).

10. Pisters, K. M., Evans, W. K., Azzoli, C. G., Kris, M. G., Smith, C. A., Desch, C. E. et al. Cancer Care Ontario and American Society of Clinical Oncology adjuvant chemotherapy and adjuvant radiation therapy for stages I-IIIA resectable non small-cell lung cancer guideline. J Clin Oncol 25, 5506-5518 (2007).

11. Mountain, C. F. Revisions in the International System for Staging Lung Cancer. Chest 111, 1710-1717 (1997).

12. Masters, G. A., Temin, S., Azzoli, C. G., Giaccone, G., Baker, S., Jr., Brahmer, J. R. et al. Systemic Therapy for Stage IV Non-Small-Cell Lung Cancer: American Society of Clinical Oncology Clinical Practice Guideline Update. J Clin Oncol 33, 3488-3515 (2015).

118

13. Reck, M., Rodriguez-Abreu, D., Robinson, A. G., Hui, R., Csoszi, T., Fulop, A. et al. Pembrolizumab versus Chemotherapy for PD-L1-Positive Non-Small-Cell Lung Cancer. N Engl J Med 375, 1823-1833 (2016).

14. Hirsch, F. R., Suda, K., Wiens, J. & Bunn, P. A., Jr. New and emerging targeted treatments in advanced non-small-cell lung cancer. Lancet 388, 1012-1024 (2016).

15. Pao, W., Miller, V., Zakowski, M., Doherty, J., Politi, K., Sarkaria, I. et al. EGF receptor gene mutations are common in lung cancers from "never smokers" and are associated with sensitivity of tumors to gefitinib and erlotinib. Proc Natl Acad Sci U S A 101, 13306-13311 (2004).

16. Lynch, T. J., Bell, D. W., Sordella, R., Gurubhagavatula, S., Okimoto, R. A., Brannigan, B. W. et al. Activating mutations in the epidermal growth factor receptor underlying responsiveness of non-small-cell lung cancer to gefitinib. N Engl J Med 350, 2129-2139 (2004).

17. Paez, J. G., Janne, P. A., Lee, J. C., Tracy, S., Greulich, H., Gabriel, S. et al. EGFR mutations in lung cancer: correlation with clinical response to gefitinib therapy. Science 304, 1497-1500 (2004).

18. Marchetti, A., Felicioni, L., Malatesta, S., Grazia Sciarrotta, M., Guetti, L., Chella, A. et al. Clinical features and outcome of patients with non-small-cell lung cancer harboring BRAF mutations. J Clin Oncol 29, 3574-3579 (2011).

19. Paik, P. K., Arcila, M. E., Fara, M., Sima, C. S., Miller, V. A., Kris, M. G. et al. Clinical characteristics of patients with lung adenocarcinomas harboring BRAF mutations. J Clin Oncol 29, 2046-2051 (2011).

20. Kawano, O., Sasaki, H., Endo, K., Suzuki, E., Haneda, H., Yukiue, H. et al. PIK3CA mutation status in Japanese lung cancer patients. Lung Cancer 54, 209- 215 (2006).

21. Cancer Genome Atlas Research, N. Comprehensive molecular profiling of lung adenocarcinoma. Nature 511, 543-550 (2014).

22. Stephens, P., Hunter, C., Bignell, G., Edkins, S., Davies, H., Teague, J. et al. Lung cancer: intragenic ERBB2 kinase mutations in tumours. Nature 431, 525- 526 (2004).

23. Zhao, X., Weir, B. A., LaFramboise, T., Lin, M., Beroukhim, R., Garraway, L. et al. Homozygous deletions and chromosome amplifications in human lung carcinomas revealed by single nucleotide polymorphism array analysis. Cancer Res 65, 5561-5570 (2005).

119

24. Lipson, D., Capelletti, M., Yelensky, R., Otto, G., Parker, A., Jarosz, M. et al. Identification of new ALK and RET gene fusions from colorectal and lung cancer biopsies. Nat Med 18, 382-384 (2012).

25. Soda, M., Choi, Y. L., Enomoto, M., Takada, S., Yamashita, Y., Ishikawa, S. et al. Identification of the transforming EML4-ALK fusion gene in non-small-cell lung cancer. Nature 448, 561-566 (2007).

26. Takeuchi, K., Soda, M., Togashi, Y., Suzuki, R., Sakata, S., Hatano, S. et al. RET, ROS1 and ALK fusions in lung cancer. Nat Med 18, 378-381 (2012).

27. Rikova, K., Guo, A., Zeng, Q., Possemato, A., Yu, J., Haack, H. et al. Global survey of phosphotyrosine signaling identifies oncogenic kinases in lung cancer. Cell 131, 1190-1203 (2007).

28. Kohno, T., Ichikawa, H., Totoki, Y., Yasuda, K., Hiramoto, M., Nammo, T. et al. KIF5B-RET fusions in lung adenocarcinoma. Nat Med 18, 375-377 (2012).

29. Vaishnavi, A., Le, A. T. & Doebele, R. C. TRKing down an old oncogene in a new era of targeted therapy. Cancer Discov 5, 25-34 (2015).

30. Nakaoku, T., Tsuta, K., Ichikawa, H., Shiraishi, K., Sakamoto, H., Enari, M. et al. Druggable oncogene fusions in invasive mucinous lung adenocarcinoma. Clin Cancer Res 20, 3087-3093 (2014).

31. Fernandez-Cuesta, L., Plenker, D., Osada, H., Sun, R., Menon, R., Leenders, F. et al. CD74-NRG1 fusions in lung adenocarcinoma. Cancer Discov 4, 415-422 (2014).

32. Seo, J. S., Ju, Y. S., Lee, W. C., Shin, J. Y., Lee, J. K., Bleazard, T. et al. The transcriptional landscape and mutational profile of lung adenocarcinoma. Genome Res 22, 2109-2119 (2012).

33. An, S. J., Chen, Z. H., Su, J., Zhang, X. C., Zhong, W. Z., Yang, J. J. et al. Identification of enriched driver gene alterations in subgroups of non-small cell lung cancer patients based on histology and smoking status. PLoS One 7, e40109 (2012).

34. Govindan, R., Ding, L., Griffith, M., Subramanian, J., Dees, N. D., Kanchi, K. L. et al. Genomic landscape of non-small cell lung cancer in smokers and never- smokers. Cell 150, 1121-1134 (2012).

35. Ahrendt, S. A., Decker, P. A., Alawi, E. A., Zhu, Y.-r., Sanchez-Cespedes, M., Yang, S. C. et al. Cigarette smoking is strongly associated with mutation of the K- ras gene in patients with primary adenocarcinoma of the lung. Cancer 92, 1525- 1530 (2001).

120

36. Riely, G. J., Kris, M. G., Rosenbaum, D., Marks, J., Li, A., Chitale, D. A. et al. Frequency and distinctive spectrum of KRAS mutations in never smokers with lung adenocarcinoma. Clin Cancer Res 14, 5731-5734 (2008).

37. Dogan, S., Shen, R., Ang, D. C., Johnson, M. L., D'Angelo, S. P., Paik, P. K. et al. Molecular epidemiology of EGFR and KRAS mutations in 3,026 lung adenocarcinomas: higher susceptibility of women to smoking-related KRAS- mutant cancers. Clin Cancer Res 18, 6169-6177 (2012).

38. Saito, M., Shiraishi, K., Kunitoh, H., Takenoshita, S., Yokota, J. & Kohno, T. Gene aberrations for precision medicine against lung adenocarcinoma. Cancer Sci 107, 713-720 (2016).

39. King, N. & Carroll, S. B. A receptor tyrosine kinase from choanoflagellates: molecular insights into early animal evolution. Proc Natl Acad Sci U S A 98, 15032-15037 (2001).

40. Miller, W. T. Tyrosine kinase signaling and the emergence of multicellularity. Biochim Biophys Acta 1823, 1053-1057 (2012).

41. Robinson, D. R., Wu, Y. M. & Lin, S. F. The protein tyrosine kinase family of the human genome. Oncogene 19, 5548-5557 (2000).

42. Hubbard, S. R. & Till, J. H. Protein tyrosine kinase structure and function. Annu Rev Biochem 69, 373-398 (2000).

43. Ullrich, A. & Schlessinger, J. Signal transduction by receptors with tyrosine kinase activity. Cell 61, 203-212 (1990).

44. Lemmon, M. A. & Schlessinger, J. Regulation of signal transduction and signal diversity by receptor oligomerization. Trends Biochem Sci 19, 459-463 (1994).

45. Lemmon, M. A. & Schlessinger, J. Cell signaling by receptor tyrosine kinases. Cell 141, 1117-1134 (2010).

46. Hubbard, S. R. Juxtamembrane autoinhibition in receptor tyrosine kinases. Nat Rev Mol Cell Biol 5, 464-471 (2004).

47. Johnson, L. N., Noble, M. E. & Owen, D. J. Active and inactive protein kinases: structural basis for regulation. Cell 85, 149-158 (1996).

48. McKay, M. M. & Morrison, D. K. Integrating signals from RTKs to ERK/MAPK. Oncogene 26, 3113-3121 (2007).

49. Futreal, P. A., Coin, L., Marshall, M., Down, T., Hubbard, T., Wooster, R. et al. A census of human cancer genes. Nat Rev Cancer 4, 177-183 (2004).

121

50. d'Avis, P. Y., Robertson, S. C., Meyer, A. N., Bardwell, W. M., Webster, M. K. & Donoghue, D. J. Constitutive activation of fibroblast growth factor receptor 3 by mutations responsible for the lethal skeletal dysplasia thanatophoric dysplasia type I. Cell Growth Differ 9, 71-78 (1998).

51. Adar, R., Monsonego-Ornan, E., David, P. & Yayon, A. Differential activation of cysteine-substitution mutants of fibroblast growth factor receptor 3 is determined by cysteine localization. J Bone Miner Res 17, 860-868 (2002).

52. Roskoski, R., Jr. Structure and regulation of Kit protein-tyrosine kinase--the stem cell factor receptor. Biochem Biophys Res Commun 338, 1307-1315 (2005).

53. Lahiry, P., Torkamani, A., Schork, N. J. & Hegele, R. A. Kinase mutations in human disease: interpreting genotype-phenotype relationships. Nat Rev Genet 11, 60-74 (2010).

54. Yun, C. H., Boggon, T. J., Li, Y., Woo, M. S., Greulich, H., Meyerson, M. et al. Structures of lung cancer-derived EGFR mutants and inhibitor complexes: mechanism of activation and insights into differential inhibitor sensitivity. Cancer Cell 11, 217-227 (2007).

55. Shugay, M., Ortiz de Mendibil, I., Vizmanos, J. L. & Novo, F. J. Genomic hallmarks of genes involved in chromosomal translocations in hematological cancer. PLoS Comput Biol 8, e1002797 (2012).

56. Shaw, A. T., Hsu, P. P., Awad, M. M. & Engelman, J. A. Tyrosine kinase gene rearrangements in epithelial malignancies. Nat Rev Cancer 13, 772-787 (2013).

57. Casaletto, J. B. & McClatchey, A. I. Spatial regulation of receptor tyrosine kinases in development and cancer. Nat Rev Cancer 12, 387-400 (2012).

58. Peters, S. & Adjei, A. A. MET: a promising anticancer therapeutic target. Nat Rev Clin Oncol 9, 314-326 (2012).

59. Wang, R., Ferrell, L. D., Faouzi, S., Maher, J. J. & Bishop, J. M. Activation of the Met receptor by cell attachment induces and sustains hepatocellular carcinomas in transgenic mice. J Cell Biol 153, 1023-1034 (2001).

60. Engelman, J. A., Zejnullahu, K., Mitsudomi, T., Song, Y., Hyland, C., Park, J. O. et al. MET amplification leads to gefitinib resistance in lung cancer by activating ERBB3 signaling. Science 316, 1039-1043 (2007).

61. Chen, W. J., Ho, C. C., Chang, Y. L., Chen, H. Y., Lin, C. A., Ling, T. Y. et al. Cancer-associated fibroblasts regulate the plasticity of lung cancer stemness via paracrine signalling. Nat Commun 5, 3472 (2014).

62. Higashiyama, S. & Nanba, D. ADAM-mediated ectodomain shedding of HB-EGF in receptor cross-talk. Biochim Biophys Acta 1751, 110-117 (2005).

122

63. Shaul, Y. D. & Seger, R. The MEK/ERK cascade: from signaling specificity to diverse functions. Biochim Biophys Acta 1773, 1213-1226 (2007).

64. Lowenstein, E. J., Daly, R. J., Batzer, A. G., Li, W., Margolis, B., Lammers, R. et al. The SH2 and SH3 domain-containing protein GRB2 links receptor tyrosine kinases to ras signaling. Cell 70, 431-442 (1992).

65. Rozakis-Adcock, M., McGlade, J., Mbamalu, G., Pelicci, G., Daly, R., Li, W. et al. Association of the Shc and Grb2/Sem5 SH2-containing proteins is implicated in activation of the Ras pathway by tyrosine kinases. Nature 360, 689-692 (1992).

66. Skolnik, E. Y., Batzer, A., Li, N., Lee, C. H., Lowenstein, E., Mohammadi, M. et al. The function of GRB2 in linking the insulin receptor to Ras signaling pathways. Science 260, 1953-1955 (1993).

67. Skolnik, E. Y., Lee, C. H., Batzer, A., Vicentini, L. M., Zhou, M., Daly, R. et al. The SH2/SH3 domain-containing protein GRB2 interacts with tyrosine- phosphorylated IRS1 and Shc: implications for insulin control of ras signalling. EMBO J 12, 1929-1936 (1993).

68. Egan, S. E., Giddings, B. W., Brooks, M. W., Buday, L., Sizeland, A. M. & Weinberg, R. A. Association of Sos Ras exchange protein with Grb2 is implicated in tyrosine kinase signal transduction and transformation. Nature 363, 45-51 (1993).

69. Li, N., Batzer, A., Daly, R., Yajnik, V., Skolnik, E., Chardin, P. et al. Guanine- nucleotide-releasing factor hSos1 binds to Grb2 and links receptor tyrosine kinases to Ras signalling. Nature 363, 85-88 (1993).

70. McCormick, F. Signal transduction. How receptors turn Ras on. Nature 363, 15- 16 (1993).

71. Wellbrock, C., Karasarides, M. & Marais, R. The RAF proteins take centre stage. Nat Rev Mol Cell Biol 5, 875-885 (2004).

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

73. Reddy, E. P., Reynolds, R. K., Santos, E. & Barbacid, M. A point mutation is responsible for the acquisition of transforming properties by the T24 human bladder carcinoma oncogene. Nature 300, 149-152 (1982).

74. Tabin, C. J., Bradley, S. M., Bargmann, C. I., Weinberg, R. A., Papageorge, A. G., Scolnick, E. M. et al. Mechanism of activation of a human oncogene. Nature 300, 143-149 (1982).

123

75. Taparowsky, E., Suard, Y., Fasano, O., Shimizu, K., Goldfarb, M. & Wigler, M. Activation of the T24 bladder carcinoma transforming gene is linked to a single amino acid change. Nature 300, 762-765 (1982).

76. Capon, D. J., Chen, E. Y., Levinson, A. D., Seeburg, P. H. & Goeddel, D. V. Complete nucleotide sequences of the T24 human bladder carcinoma oncogene and its normal homologue. Nature 302, 33-37 (1983).

77. Scheffzek, K., Ahmadian, M. R., Kabsch, W., Wiesmuller, L., Lautwein, A., Schmitz, F. et al. The Ras-RasGAP complex: structural basis for GTPase activation and its loss in oncogenic Ras mutants. Science 277, 333-338 (1997).

78. Buhrman, G., Holzapfel, G., Fetics, S. & Mattos, C. Allosteric modulation of Ras positions Q61 for a direct role in catalysis. Proc Natl Acad Sci U S A 107, 4931- 4936 (2010).

79. Scheidig, A. J., Burmester, C. & Goody, R. S. The pre-hydrolysis state of p21(ras) in complex with GTP: new insights into the role of water molecules in the GTP hydrolysis reaction of ras-like proteins. Structure 7, 1311-1324 (1999).

80. Hu, P., Margolis, B., Skolnik, E. Y., Lammers, R., Ullrich, A. & Schlessinger, J. Interaction of phosphatidylinositol 3-kinase-associated p85 with epidermal growth factor and platelet-derived growth factor receptors. Mol Cell Biol 12, 981- 990 (1992).

81. McGlade, C. J., Ellis, C., Reedijk, M., Anderson, D., Mbamalu, G., Reith, A. D. et al. SH2 domains of the p85 alpha subunit of phosphatidylinositol 3-kinase regulate binding to growth factor receptors. Mol Cell Biol 12, 991-997 (1992).

82. Myers, M. G., Jr., Backer, J. M., Sun, X. J., Shoelson, S., Hu, P., Schlessinger, J. et al. IRS-1 activates phosphatidylinositol 3'-kinase by associating with src homology 2 domains of p85. Proc Natl Acad Sci U S A 89, 10350-10354 (1992).

83. Hawkins, P. T., Jackson, T. R. & Stephens, L. R. Platelet-derived growth factor stimulates synthesis of PtdIns(3,4,5)P3 by activating a PtdIns(4,5)P2 3-OH kinase. Nature 358, 157-159 (1992).

84. Chan, T. O., Rodeck, U., Chan, A. M., Kimmelman, A. C., Rittenhouse, S. E., Panayotou, G. et al. Small GTPases and tyrosine kinases coregulate a molecular switch in the phosphoinositide 3-kinase regulatory subunit. Cancer Cell 1, 181- 191 (2002).

85. Rodriguez-Viciana, P., Warne, P. H., Dhand, R., Vanhaesebroeck, B., Gout, I., Fry, M. J. et al. Phosphatidylinositol-3-OH kinase as a direct target of Ras. Nature 370, 527-532 (1994).

124

86. Rodriguez-Viciana, P., Warne, P. H., Vanhaesebroeck, B., Waterfield, M. D. & Downward, J. Activation of phosphoinositide 3-kinase by interaction with Ras and by point mutation. EMBO J 15, 2442-2451 (1996).

87. Myers, M. P., Pass, I., Batty, I. H., Van der Kaay, J., Stolarov, J. P., Hemmings, B. A. et al. The lipid phosphatase activity of PTEN is critical for its tumor supressor function. Proc Natl Acad Sci U S A 95, 13513-13518 (1998).

88. Stambolic, V., Suzuki, A., de la Pompa, J. L., Brothers, G. M., Mirtsos, C., Sasaki, T. et al. Negative regulation of PKB/Akt-dependent cell survival by the tumor suppressor PTEN. Cell 95, 29-39 (1998).

89. Maehama, T. & Dixon, J. E. The tumor suppressor, PTEN/MMAC1, dephosphorylates the lipid second messenger, phosphatidylinositol 3,4,5- trisphosphate. J Biol Chem 273, 13375-13378 (1998).

90. Li, J., Yen, C., Liaw, D., Podsypanina, K., Bose, S., Wang, S. I. et al. PTEN, a putative protein tyrosine phosphatase gene mutated in human brain, breast, and prostate cancer. Science 275, 1943-1947 (1997).

91. Alessi, D. R., James, S. R., Downes, C. P., Holmes, A. B., Gaffney, P. R., Reese, C. B. et al. Characterization of a 3-phosphoinositide-dependent protein kinase which phosphorylates and activates protein kinase Balpha. Curr Biol 7, 261-269 (1997).

92. Sarbassov, D. D., Guertin, D. A., Ali, S. M. & Sabatini, D. M. Phosphorylation and regulation of Akt/PKB by the rictor-mTOR complex. Science 307, 1098-1101 (2005).

93. Wullschleger, S., Loewith, R., Oppliger, W. & Hall, M. N. Molecular organization of target of rapamycin complex 2. J Biol Chem 280, 30697-30704 (2005).

94. Cantley, L. C. The phosphoinositide 3-kinase pathway. Science 296, 1655-1657 (2002).

95. Inoki, K., Li, Y., Zhu, T., Wu, J. & Guan, K. L. TSC2 is phosphorylated and inhibited by Akt and suppresses mTOR signalling. Nat Cell Biol 4, 648-657 (2002).

96. Inoki, K., Li, Y., Xu, T. & Guan, K. L. Rheb GTPase is a direct target of TSC2 GAP activity and regulates mTOR signaling. Genes Dev 17, 1829-1834 (2003).

97. Inoki, K., Zhu, T. & Guan, K. L. TSC2 mediates cellular energy response to control cell growth and survival. Cell 115, 577-590 (2003).

98. Zoncu, R., Efeyan, A. & Sabatini, D. M. mTOR: from growth signal integration to cancer, diabetes and ageing. Nat Rev Mol Cell Biol 12, 21-35 (2011).

125

99. Ma, X. M. & Blenis, J. Molecular mechanisms of mTOR-mediated translational control. Nat Rev Mol Cell Biol 10, 307-318 (2009).

100. Druker, B. J., Guilhot, F., O'Brien, S. G., Gathmann, I., Kantarjian, H., Gattermann, N. et al. Five-year follow-up of patients receiving imatinib for chronic myeloid leukemia. N Engl J Med 355, 2408-2417 (2006).

101. Wu, P., Nielsen, T. E. & Clausen, M. H. Small-molecule kinase inhibitors: an analysis of FDA-approved drugs. Drug Discov Today 21, 5-10 (2016).

102. Swanton, C. & Govindan, R. Clinical Implications of Genomic Discoveries in Lung Cancer. N Engl J Med 374, 1864-1873 (2016).

103. Weinstein, I. B. Cancer. Addiction to oncogenes--the Achilles heal of cancer. Science 297, 63-64 (2002).

104. Sharma, S. V., Fischbach, M. A., Haber, D. A. & Settleman, J. "Oncogenic shock": explaining oncogene addiction through differential signal attenuation. Clin Cancer Res 12, 4392s-4395s (2006).

105. Sharma, S. V. & Settleman, J. Oncogene addiction: setting the stage for molecularly targeted cancer therapy. Genes Dev 21, 3214-3231 (2007).

106. Camidge, D. R., Pao, W. & Sequist, L. V. Acquired resistance to TKIs in solid tumours: learning from lung cancer. Nat Rev Clin Oncol 11, 473-481 (2014).

107. Zhang, J., Yang, P. L. & Gray, N. S. Targeting cancer with small molecule kinase inhibitors. Nat Rev Cancer 9, 28-39 (2009).

108. Kobayashi, S., Boggon, T. J., Dayaram, T., Janne, P. A., Kocher, O., Meyerson, M. et al. EGFR mutation and resistance of non-small-cell lung cancer to gefitinib. N Engl J Med 352, 786-792 (2005).

109. Carlomagno, F., Guida, T., Anaganti, S., Vecchio, G., Fusco, A., Ryan, A. J. et al. Disease associated mutations at valine 804 in the RET receptor tyrosine kinase confer resistance to selective kinase inhibitors. Oncogene 23, 6056-6063 (2004).

110. Lovly, C. M. & Shaw, A. T. Molecular pathways: resistance to kinase inhibitors and implications for therapeutic strategies. Clin Cancer Res 20, 2249-2256 (2014).

111. Choi, Y. L., Soda, M., Yamashita, Y., Ueno, T., Takashima, J., Nakajima, T. et al. EML4-ALK mutations in lung cancer that confer resistance to ALK inhibitors. N Engl J Med 363, 1734-1739 (2010).

112. Terai, H., Soejima, K., Yasuda, H., Nakayama, S., Hamamoto, J., Arai, D. et al. Activation of the FGF2-FGFR1 autocrine pathway: a novel mechanism of acquired resistance to gefitinib in NSCLC. Mol Cancer Res 11, 759-767 (2013).

126

113. Cargnelutti, M., Corso, S., Pergolizzi, M., Mevellec, L., Aisner, D. L., Dziadziuszko, R. et al. Activation of RAS family members confers resistance to ROS1 targeting drugs. Oncotarget 6, 5182-5194 (2015).

114. Ohashi, K., Sequist, L. V., Arcila, M. E., Moran, T., Chmielecki, J., Lin, Y. L. et al. Lung cancers with acquired resistance to EGFR inhibitors occasionally harbor BRAF gene mutations but lack mutations in KRAS, NRAS, or MEK1. Proc Natl Acad Sci U S A 109, E2127-2133 (2012).

115. Sos, M. L., Koker, M., Weir, B. A., Heynck, S., Rabinovsky, R., Zander, T. et al. PTEN loss contributes to erlotinib resistance in EGFR-mutant lung cancer by activation of Akt and EGFR. Cancer Res 69, 3256-3261 (2009).

116. Losick, R. & Desplan, C. Stochasticity and cell fate. Science 320, 65-68 (2008).

117. Emmons, M. F., Faiao-Flores, F. & Smalley, K. S. The role of phenotypic plasticity in the escape of cancer cells from targeted therapy. Biochem Pharmacol 122, 1-9 (2016).

118. Frank, S. A. & Rosner, M. R. Nonheritable cellular variability accelerates the evolutionary processes of cancer. PLoS Biol 10, e1001296 (2012).

119. Sharma, S. V., Lee, D. Y., Li, B., Quinlan, M. P., Takahashi, F., Maheswaran, S. et al. A chromatin-mediated reversible drug-tolerant state in cancer cell subpopulations. Cell 141, 69-80 (2010).

120. Zhang, Z., Lee, J. C., Lin, L., Olivas, V., Au, V., LaFramboise, T. et al. Activation of the AXL kinase causes resistance to EGFR-targeted therapy in lung cancer. Nat Genet 44, 852-860 (2012).

121. Byers, L. A., Diao, L., Wang, J., Saintigny, P., Girard, L., Peyton, M. et al. An epithelial-mesenchymal transition gene signature predicts resistance to EGFR and PI3K inhibitors and identifies Axl as a therapeutic target for overcoming EGFR inhibitor resistance. Clin Cancer Res 19, 279-290 (2013).

122. Oser, M. G., Niederst, M. J., Sequist, L. V. & Engelman, J. A. Transformation from non-small-cell lung cancer to small-cell lung cancer: molecular drivers and cells of origin. Lancet Oncol 16, e165-172 (2015).

123. Skoulidis, F. & Papadimitrakopoulou, V. A. Targeting the Gatekeeper: Osimertinib in EGFR T790M Mutation-Positive Non-small Cell Lung Cancer. Clin Cancer Res (2016).

124. Larkin, J., Ascierto, P. A., Dreno, B., Atkinson, V., Liszkay, G., Maio, M. et al. Combined vemurafenib and cobimetinib in BRAF-mutated melanoma. N Engl J Med 371, 1867-1876 (2014).

127

125. Kleczko, E. K., Kim, J., Keysar, S. B., Heasley, L. R., Eagles, J. R., Simon, M. et al. An Inducible TGF-beta2-TGFbetaR Pathway Modulates the Sensitivity of HNSCC Cells to Tyrosine Kinase Inhibitors Targeting Dominant Receptor Tyrosine Kinases. PLoS One 10, e0123600 (2015).

126. Takahashi, M., Ritz, J. & Cooper, G. M. Activation of a novel human transforming gene, ret, by DNA rearrangement. Cell 42, 581-588 (1985).

127. Wang, X. Structural studies of GDNF family ligands with their receptors-Insights into ligand recognition and activation of receptor tyrosine kinase RET. Biochim Biophys Acta 1834, 2205-2212 (2013).

128. Romeo, G., Ronchetto, P., Luo, Y., Barone, V., Seri, M., Ceccherini, I. et al. Point mutations affecting the tyrosine kinase domain of the RET proto-oncogene in Hirschsprung's disease. Nature 367, 377-378 (1994).

129. Edery, P., Lyonnet, S., Mulligan, L. M., Pelet, A., Dow, E., Abel, L. et al. Mutations of the RET proto-oncogene in Hirschsprung's disease. Nature 367, 378-380 (1994).

130. Kotzbauer, P. T., Lampe, P. A., Heuckeroth, R. O., Golden, J. P., Creedon, D. J., Johnson, E. M., Jr. et al. Neurturin, a relative of glial-cell-line-derived neurotrophic factor. Nature 384, 467-470 (1996).

131. Jing, S., Wen, D., Yu, Y., Holst, P. L., Luo, Y., Fang, M. et al. GDNF-induced activation of the ret protein tyrosine kinase is mediated by GDNFR-alpha, a novel receptor for GDNF. Cell 85, 1113-1124 (1996).

132. Buj-Bello, A., Adu, J., Pinon, L. G., Horton, A., Thompson, J., Rosenthal, A. et al. Neurturin responsiveness requires a GPI-linked receptor and the Ret receptor tyrosine kinase. Nature 387, 721-724 (1997).

133. Baloh, R. H., Tansey, M. G., Lampe, P. A., Fahrner, T. J., Enomoto, H., Simburger, K. S. et al. Artemin, a novel member of the GDNF ligand family, supports peripheral and central neurons and signals through the GFRalpha3-RET receptor complex. Neuron 21, 1291-1302 (1998).

134. Masure, S., Cik, M., Hoefnagel, E., Nosrat, C. A., Van der Linden, I., Scott, R. et al. Mammalian GFRalpha -4, a divergent member of the GFRalpha family of coreceptors for glial cell line-derived neurotrophic factor family ligands, is a receptor for the neurotrophic factor persephin. J Biol Chem 275, 39427-39434 (2000).

135. Tansey, M. G., Baloh, R. H., Milbrandt, J. & Johnson, E. M., Jr. GFRalpha- mediated localization of RET to lipid rafts is required for effective downstream signaling, differentiation, and neuronal survival. Neuron 25, 611-623 (2000).

128

136. Amoresano, A., Incoronato, M., Monti, G., Pucci, P., de Franciscis, V. & Cerchia, L. Direct interactions among Ret, GDNF and GFRalpha1 molecules reveal new insights into the assembly of a functional three-protein complex. Cell Signal 17, 717-727 (2005).

137. Myers, S. M., Eng, C., Ponder, B. A. & Mulligan, L. M. Characterization of RET proto-oncogene 3' splicing variants and polyadenylation sites: a novel C-terminus for RET. Oncogene 11, 2039-2045 (1995).

138. Tahira, T., Ishizaka, Y., Itoh, F., Sugimura, T. & Nagao, M. Characterization of ret proto-oncogene mRNAs encoding two isoforms of the protein product in a human neuroblastoma cell line. Oncogene 5, 97-102 (1990).

139. Knowles, P. P., Murray-Rust, J., Kjaer, S., Scott, R. P., Hanrahan, S., Santoro, M. et al. Structure and chemical inhibition of the RET tyrosine kinase domain. J Biol Chem 281, 33577-33587 (2006).

140. Plaza-Menacho, I., Barnouin, K., Goodman, K., Martinez-Torres, R. J., Borg, A., Murray-Rust, J. et al. Oncogenic RET kinase domain mutations perturb the autophosphorylation trajectory by enhancing substrate presentation in trans. Mol Cell 53, 738-751 (2014).

141. Liu, X., Vega, Q. C., Decker, R. A., Pandey, A., Worby, C. A. & Dixon, J. E. Oncogenic RET receptors display different autophosphorylation sites and substrate binding specificities. J Biol Chem 271, 5309-5312 (1996).

142. Kawamoto, Y., Takeda, K., Okuno, Y., Yamakawa, Y., Ito, Y., Taguchi, R. et al. Identification of RET autophosphorylation sites by mass spectrometry. J Biol Chem 279, 14213-14224 (2004).

143. Besset, V., Scott, R. P. & Ibanez, C. F. Signaling complexes and protein-protein interactions involved in the activation of the Ras and phosphatidylinositol 3- kinase pathways by the c-Ret receptor tyrosine kinase. J Biol Chem 275, 39159- 39166 (2000).

144. Segouffin-Cariou, C. & Billaud, M. Transforming ability of MEN2A-RET requires activation of the phosphatidylinositol 3-kinase/AKT signaling pathway. J Biol Chem 275, 3568-3576 (2000).

145. Coulpier, M., Anders, J. & Ibanez, C. F. Coordinated activation of autophosphorylation sites in the RET receptor tyrosine kinase: importance of tyrosine 1062 for GDNF mediated neuronal differentiation and survival. J Biol Chem 277, 1991-1999 (2002).

146. Hayashi, H., Ichihara, M., Iwashita, T., Murakami, H., Shimono, Y., Kawai, K. et al. Characterization of intracellular signals via tyrosine 1062 in RET activated by glial cell line-derived neurotrophic factor. Oncogene 19, 4469-4475 (2000).

129

147. Jain, S., Encinas, M., Johnson, E. M., Jr. & Milbrandt, J. Critical and distinct roles for key RET tyrosine docking sites in renal development. Genes Dev 20, 321-333 (2006).

148. Pandey, A., Liu, X., Dixon, J. E., DiFiore, P. P. & Dixit, V. M. Direct association between the Ret receptor tyrosine kinase and the Src homology 2-containing adapter protein Grb7. Journal of Biological Chemistry 271, 10607-10610 (1996).

149. Schuringa, J. J., Wojtachnio, K., Hagens, W., Vellenga, E., Buys, C. H., Hofstra, R. et al. MEN2A-RET-induced cellular transformation by activation of STAT3. Oncogene 20, 5350-5358 (2001).

150. Encinas, M., Crowder, R. J., Milbrandt, J. & Johnson, E. M., Jr. Tyrosine 981, a novel ret autophosphorylation site, binds c-Src to mediate neuronal survival. J Biol Chem 279, 18262-18269 (2004).

151. Borrello, M. G., Alberti, L., Arighi, E., Bongarzone, I., Battistini, C., Bardelli, A. et al. The full oncogenic activity of Ret/ptc2 depends on tyrosine 539, a docking site for phospholipase Cgamma. Mol Cell Biol 16, 2151-2163 (1996).

152. Mulligan, L. M. RET revisited: expanding the oncogenic portfolio. Nat Rev Cancer 14, 173-186 (2014).

153. Romei, C., Ciampi, R. & Elisei, R. A comprehensive overview of the role of the RET proto-oncogene in thyroid carcinoma. Nat Rev Endocrinol 12, 192-202 (2016).

154. Eng, C., Clayton, D., Schuffenecker, I., Lenoir, G., Cote, G., Gagel, R. F. et al. The relationship between specific RET proto-oncogene mutations and disease phenotype in multiple endocrine neoplasia type 2. International RET mutation consortium analysis. JAMA 276, 1575-1579 (1996).

155. Asai, N., Iwashita, T., Matsuyama, M. & Takahashi, M. Mechanism of activation of the ret proto-oncogene by multiple endocrine neoplasia 2A mutations. Mol Cell Biol 15, 1613-1619 (1995).

156. Gujral, T. S., Singh, V. K., Jia, Z. & Mulligan, L. M. Molecular mechanisms of RET receptor-mediated oncogenesis in multiple endocrine neoplasia 2B. Cancer Res 66, 10741-10749 (2006).

157. Ricarte-Filho, J. C., Li, S., Garcia-Rendueles, M. E., Montero-Conde, C., Voza, F., Knauf, J. A. et al. Identification of kinase fusion oncogenes in post-Chernobyl radiation-induced thyroid cancers. J Clin Invest 123, 4935-4944 (2013).

158. Hamatani, K., Eguchi, H., Ito, R., Mukai, M., Takahashi, K., Taga, M. et al. RET/PTC rearrangements preferentially occurred in papillary thyroid cancer among atomic bomb survivors exposed to high radiation dose. Cancer Res 68, 7176-7182 (2008).

130

159. Bounacer, A., Wicker, R., Caillou, B., Cailleux, A. F., Sarasin, A., Schlumberger, M. et al. High prevalence of activating ret proto-oncogene rearrangements, in thyroid tumors from patients who had received external radiation. Oncogene 15, 1263-1273 (1997).

160. Tsuta, K., Kohno, T., Yoshida, A., Shimada, Y., Asamura, H., Furuta, K. et al. RET-rearranged non-small-cell lung carcinoma: a clinicopathological and molecular analysis. Br J Cancer 110, 1571-1578 (2014).

161. Lin, C., Wang, S., Xie, W., Chang, J. & Gan, Y. The RET fusion gene and its correlation with demographic and clinicopathological features of non-small cell lung cancer: a meta-analysis. Cancer Biol Ther 16, 1019-1028 (2015).

162. Elisei, R., Schlumberger, M. J., Muller, S. P., Schoffski, P., Brose, M. S., Shah, M. H. et al. Cabozantinib in progressive medullary thyroid cancer. J Clin Oncol 31, 3639-3646 (2013).

163. Wells, S. A., Jr., Robinson, B. G., Gagel, R. F., Dralle, H., Fagin, J. A., Santoro, M. et al. Vandetanib in patients with locally advanced or metastatic medullary thyroid cancer: a randomized, double-blind phase III trial. J Clin Oncol 30, 134- 141 (2012).

164. Drilon, A., Rekhtman, N., Arcila, M., Wang, L., Ni, A., Albano, M. et al. Cabozantinib in patients with advanced RET-rearranged non-small-cell lung cancer: an open-label, single-centre, phase 2, single-arm trial. Lancet Oncol (2016).

165. Yoh, K., Seto, T., Satouchi, M., Nishio, M., Yamamoto, N., Murakami, H. et al. Vandetanib in patients with previously treated RET-rearranged advanced non- small-cell lung cancer (LURET): an open-label, multicentre phase 2 trial. Lancet Respir Med (2016).

166. Carlomagno, F., Vitagliano, D., Guida, T., Ciardiello, F., Tortora, G., Vecchio, G. et al. ZD6474, an orally available inhibitor of KDR tyrosine kinase activity, efficiently blocks oncogenic RET kinases. Cancer Res 62, 7284-7290 (2002).

167. Wedge, S. R., Ogilvie, D. J., Dukes, M., Kendrew, J., Chester, R., Jackson, J. A. et al. ZD6474 inhibits vascular endothelial growth factor signaling, angiogenesis, and tumor growth following oral administration. Cancer Res 62, 4645-4655 (2002).

168. Yakes, F. M., Chen, J., Tan, J., Yamaguchi, K., Shi, Y., Yu, P. et al. Cabozantinib (XL184), a novel MET and VEGFR2 inhibitor, simultaneously suppresses metastasis, angiogenesis, and tumor growth. Mol Cancer Ther 10, 2298-2308 (2011).

131

169. Shaw, A. T., Kim, D. W., Nakagawa, K., Seto, T., Crino, L., Ahn, M. J. et al. Crizotinib versus chemotherapy in advanced ALK-positive lung cancer. N Engl J Med 368, 2385-2394 (2013).

170. Shaw, A. T., Ou, S. H., Bang, Y. J., Camidge, D. R., Solomon, B. J., Salgia, R. et al. Crizotinib in ROS1-rearranged non-small-cell lung cancer. N Engl J Med 371, 1963-1971 (2014).

171. Drilon, A., Fakih, M., Fu, S., Patel, M. R., Olszanski, A. J., Lockhart, A. C. et al. A Phase 1/1b Study of RXDX-105, an Oral RET and BRAF Inhibitor, in Patients with Advanced Solid Tumors. EORTC-NCI-AACR 2016 2016. Munich, Germany.

172. Casalini, P., Iorio, M. V., Galmozzi, E. & Menard, S. Role of HER receptors family in development and differentiation. J Cell Physiol 200, 343-350 (2004).

173. Harris, R. C., Chung, E. & Coffey, R. J. EGF receptor ligands. Exp Cell Res 284, 2-13 (2003).

174. Schneider, M. R. & Wolf, E. The epidermal growth factor receptor ligands at a glance. J Cell Physiol 218, 460-466 (2009).

175. Garrett, T. P., McKern, N. M., Lou, M., Elleman, T. C., Adams, T. E., Lovrecz, G. O. et al. Crystal structure of a truncated epidermal growth factor receptor extracellular domain bound to transforming growth factor alpha. Cell 110, 763- 773 (2002).

176. Ogiso, H., Ishitani, R., Nureki, O., Fukai, S., Yamanaka, M., Kim, J. H. et al. Crystal structure of the complex of human epidermal growth factor and receptor extracellular domains. Cell 110, 775-787 (2002).

177. Pelicci, G., Lanfrancone, L., Grignani, F., McGlade, J., Cavallo, F., Forni, G. et al. A novel transforming protein (SHC) with an SH2 domain is implicated in mitogenic signal transduction. Cell 70, 93-104 (1992).

178. Bjorge, J. D., Chan, T. O., Antczak, M., Kung, H. J. & Fujita, D. J. Activated type I phosphatidylinositol kinase is associated with the epidermal growth factor (EGF) receptor following EGF stimulation. Proc Natl Acad Sci U S A 87, 3816- 3820 (1990).

179. Kloth, M. T., Laughlin, K. K., Biscardi, J. S., Boerner, J. L., Parsons, S. J. & Silva, C. M. STAT5b, a Mediator of Synergism between c-Src and the Epidermal Growth Factor Receptor. J Biol Chem 278, 1671-1679 (2003).

180. Fedi, P., Pierce, J. H., di Fiore, P. P. & Kraus, M. H. Efficient coupling with phosphatidylinositol 3-kinase, but not phospholipase C gamma or GTPase- activating protein, distinguishes ErbB-3 signaling from that of other ErbB/EGFR family members. Mol Cell Biol 14, 492-500 (1994).

132

181. Sebastian, S., Settleman, J., Reshkin, S. J., Azzariti, A., Bellizzi, A. & Paradiso, A. The complexity of targeting EGFR signalling in cancer: from expression to turnover. Biochim Biophys Acta 1766, 120-139 (2006).

182. Lombardo, C. R., Consler, T. G. & Kassel, D. B. In vitro phosphorylation of the epidermal growth factor receptor autophosphorylation domain by c-src: identification of phosphorylation sites and c-src SH2 domain binding sites. Biochemistry 34, 16456-16466 (1995).

183. Grovdal, L. M., Stang, E., Sorkin, A. & Madshus, I. H. Direct interaction of Cbl with pTyr 1045 of the EGF receptor (EGFR) is required to sort the EGFR to lysosomes for degradation. Exp Cell Res 300, 388-395 (2004).

184. Keilhack, H., Tenev, T., Nyakatura, E., Godovac-Zimmermann, J., Nielsen, L., Seedorf, K. et al. Phosphotyrosine 1173 mediates binding of the protein-tyrosine phosphatase SHP-1 to the epidermal growth factor receptor and attenuation of receptor signaling. J Biol Chem 273, 24839-24846 (1998).

185. Anklesaria, P., Teixido, J., Laiho, M., Pierce, J. H., Greenberger, J. S. & Massague, J. Cell-cell adhesion mediated by binding of membrane-anchored transforming growth factor alpha to epidermal growth factor receptors promotes cell proliferation. Proc Natl Acad Sci U S A 87, 3289-3293 (1990).

186. Sahin, U., Weskamp, G., Kelly, K., Zhou, H. M., Higashiyama, S., Peschon, J. et al. Distinct roles for ADAM10 and ADAM17 in ectodomain shedding of six EGFR ligands. J Cell Biol 164, 769-779 (2004).

187. Prenzel, N., Zwick, E., Daub, H., Leserer, M., Abraham, R., Wallasch, C. et al. EGF receptor transactivation by G-protein-coupled receptors requires metalloproteinase cleavage of proHB-EGF. Nature 402, 884-888 (1999).

188. Lappano, R. & Maggiolini, M. G protein-coupled receptors: novel targets for drug discovery in cancer. Nat Rev Drug Discov 10, 47-60 (2011).

189. Daub, H., Weiss, F. U., Wallasch, C. & Ullrich, A. Role of transactivation of the EGF receptor in signalling by G-protein-coupled receptors. Nature 379, 557-560 (1996).

190. Murphy, G. The ADAMs: signalling scissors in the tumour microenvironment. Nat Rev Cancer 8, 929-941 (2008).

191. Kennedy, S. P., Hastings, J. F., Han, J. Z. & Croucher, D. R. The Under- Appreciated Promiscuity of the Epidermal Growth Factor Receptor Family. Front Cell Dev Biol 4, 88 (2016).

133

192. Klapper, L. N., Glathe, S., Vaisman, N., Hynes, N. E., Andrews, G. C., Sela, M. et al. The ErbB-2/HER2 oncoprotein of human carcinomas may function solely as a shared coreceptor for multiple stroma-derived growth factors. Proc Natl Acad Sci U S A 96, 4995-5000 (1999).

193. Guy, P. M., Platko, J. V., Cantley, L. C., Cerione, R. A. & Carraway, K. L., 3rd. Insect cell-expressed p180erbB3 possesses an impaired tyrosine kinase activity. Proc Natl Acad Sci U S A 91, 8132-8136 (1994).

194. Jura, N., Shan, Y., Cao, X., Shaw, D. E. & Kuriyan, J. Structural analysis of the catalytically inactive kinase domain of the human EGF receptor 3. Proc Natl Acad Sci U S A 106, 21608-21613 (2009).

195. Yarden, Y. & Sliwkowski, M. X. Untangling the ErbB signalling network. Nat Rev Mol Cell Biol 2, 127-137 (2001).

196. Pinkas-Kramarski, R., Soussan, L., Waterman, H., Levkowitz, G., Alroy, I., Klapper, L. et al. Diversification of Neu differentiation factor and epidermal growth factor signaling by combinatorial receptor interactions. EMBO J 15, 2452- 2467 (1996).

197. Riese, D. J., 2nd, van Raaij, T. M., Plowman, G. D., Andrews, G. C. & Stern, D. F. The cellular response to neuregulins is governed by complex interactions of the erbB receptor family. Mol Cell Biol 15, 5770-5776 (1995).

198. Croyle, M., Akeno, N., Knauf, J. A., Fabbro, D., Chen, X., Baumgartner, J. E. et al. RET/PTC-induced cell growth is mediated in part by epidermal growth factor receptor (EGFR) activation: evidence for molecular and functional interactions between RET and EGFR. Cancer Res 68, 4183-4191 (2008).

199. Vaishnavi, A., Marek, L. A., Schubert, L. R., U., Le, A. T., Keysar, S. B., Glogowska, M. J. et al. EGFR rescues drug-inhibited fusion kinases via fusion kinase reactivation and adaptor switching. Cancer Res (submitted ).

200. Meyer, A. S., Miller, M. A., Gertler, F. B. & Lauffenburger, D. A. The receptor AXL diversifies EGFR signaling and limits the response to EGFR-targeted inhibitors in triple-negative breast cancer cells. Sci Signal 6, ra66 (2013).

201. Gusenbauer, S., Vlaicu, P. & Ullrich, A. HGF induces novel EGFR functions involved in resistance formation to tyrosine kinase inhibitors. Oncogene 32, 3846- 3856 (2013).

202. Krause, D. S. & Van Etten, R. A. Tyrosine kinases as targets for cancer therapy. N Engl J Med 353, 172-187 (2005).

134

203. Hirsch, F. R., Varella-Garcia, M., Bunn, P. A., Jr., Di Maria, M. V., Veve, R., Bremmes, R. M. et al. Epidermal growth factor receptor in non-small-cell lung carcinomas: correlation between gene copy number and protein expression and impact on prognosis. J Clin Oncol 21, 3798-3807 (2003).

204. Thatcher, N., Chang, A., Parikh, P., Rodrigues Pereira, J., Ciuleanu, T., von Pawel, J. et al. Gefitinib plus best supportive care in previously treated patients with refractory advanced non-small-cell lung cancer: results from a randomised, placebo-controlled, multicentre study (Iressa Survival Evaluation in Lung Cancer). Lancet 366, 1527-1537 (2005).

205. Shepherd, F. A., Rodrigues Pereira, J., Ciuleanu, T., Tan, E. H., Hirsh, V., Thongprasert, S. et al. Erlotinib in previously treated non-small-cell lung cancer. N Engl J Med 353, 123-132 (2005).

206. Sharma, S. V., Bell, D. W., Settleman, J. & Haber, D. A. Epidermal growth factor receptor mutations in lung cancer. Nat Rev Cancer 7, 169-181 (2007).

207. Linardou, H., Dahabreh, I. J., Bafaloukos, D., Kosmidis, P. & Murray, S. Somatic EGFR mutations and efficacy of tyrosine kinase inhibitors in NSCLC. Nat Rev Clin Oncol 6, 352-366 (2009).

208. Lapraz, F., Rottinger, E., Duboc, V., Range, R., Duloquin, L., Walton, K. et al. RTK and TGF-beta signaling pathways genes in the sea urchin genome. Dev Biol 300, 132-152 (2006).

209. Graham, D. K., DeRyckere, D., Davies, K. D. & Earp, H. S. The TAM family: phosphatidylserine sensing receptor tyrosine kinases gone awry in cancer. Nat Rev Cancer 14, 769-785 (2014).

210. Sharma, A. V., Steven, L. W., Hariprasad, V., David, J. B. & Sunil. Targeting Axl and Mer Kinases in Cancer. (2011).

211. Melaragno, M. G., Cavet, M. E., Yan, C., Tai, L. K., Jin, Z. G., Haendeler, J. et al. Gas6 inhibits apoptosis in vascular smooth muscle: role of Axl kinase and Akt. Journal of molecular and cellular cardiology 37, 881-887 (2004).

212. Guttridge, K. L., Luft, J. C., Dawson, T. L., Kozlowska, E., Mahajan, N. P., Varnum, B. et al. Mer receptor tyrosine kinase signaling: prevention of apoptosis and alteration of cytoskeletal architecture without stimulation or proliferation. J Biol Chem 277, 24057-24066 (2002).

213. Sayan, A. E., Stanford, R., Vickery, R., Grigorenko, E., Diesch, J., Kulbicki, K. et al. Fra-1 controls motility of bladder cancer cells via transcriptional upregulation of the receptor tyrosine kinase AXL. Oncogene 31, 1493-1503 (2011).

135

214. Asiedu, M. K., Beauchamp-Perez, F. D., Ingle, J. N., Behrens, M. D., Radisky, D. C. & Knutson, K. L. AXL induces epithelial-to-mesenchymal transition and regulates the function of breast cancer stem cells. Oncogene 33, 1316-1324 (2013).

215. Lee, H. J., Jeng, Y. M., Chen, Y. L., Chung, L. & Yuan, R. H. Gas6/Axl pathway promotes tumor invasion through the transcriptional activation of Slug in hepatocellular carcinoma. Carcinogenesis 35, 769-775 (2013).

216. Elkabets, M., Pazarentzos, E., Juric, D., Sheng, Q., Pelossof, R. A., Brook, S. et al. AXL mediates resistance to PI3Kalpha inhibition by activating the EGFR/PKC/mTOR axis in head and neck and esophageal squamous cell carcinomas. Cancer Cell 27, 533-546 (2015).

217. Tworkoski, K. A., Platt, J. T., Bacchiocchi, A., Bosenberg, M., Boggon, T. J. & Stern, D. F. MERTK controls melanoma cell migration and survival and differentially regulates cell behavior relative to AXL. Pigment Cell Melanoma Res 26, 527-541 (2013).

218. Rogers, A. E., Le, J. P., Sather, S., Pernu, B. M., Graham, D. K., Pierce, A. M. et al. Mer receptor tyrosine kinase inhibition impedes glioblastoma multiforme migration and alters cellular morphology. Oncogene 31, 4171-4181 (2011).

219. Lemke, G. Biology of the TAM receptors. Cold Spring Harb Perspect Biol 5, a009076 (2013).

220. Masters, J. R., Thomson, J. A., Daly-Burns, B., Reid, Y. A., Dirks, W. G., Packer, P. et al. Short tandem repeat profiling provides an international reference standard for human cell lines. Proc Natl Acad Sci U S A 98, 8012-8017 (2001).

221. Rogers, A. E., Le, J. P., Sather, S., Pernu, B. M., Graham, D. K., Pierce, A. M. et al. Mer receptor tyrosine kinase inhibition impedes glioblastoma multiforme migration and alters cellular morphology. Oncogene 31, 4171-4181 (2012).

222. Reynolds, B. A. & Weiss, S. Generation of neurons and astrocytes from isolated cells of the adult mammalian central nervous system. Science 255, 1707-1710 (1992).

223. Schweppe, R. E., Klopper, J. P., Korch, C., Pugazhenthi, U., Benezra, M., Knauf, J. A. et al. Deoxyribonucleic acid profiling analysis of 40 human thyroid cancer cell lines reveals cross-contamination resulting in cell line redundancy and misidentification. J Clin Endocrinol Metab 93, 4331-4341 (2008).

224. Davies, K. D., Mahale, S., Astling, D. P., Aisner, D. L., Le, A. T., Hinz, T. K. et al. Resistance to ROS1 inhibition mediated by EGFR pathway activation in non- small cell lung cancer. PLoS One 8, e82236 (2013).

136

225. Davies, K. D., Le, A. T., Theodoro, M. F., Skokan, M. C., Aisner, D. L., Berge, E. M. et al. Identifying and targeting ROS1 gene fusions in non-small cell lung cancer. Clin Cancer Res 18, 4570-4579 (2012).

226. Doebele, R. C., Pilling, A. B., Aisner, D. L., Kutateladze, T. G., Le, A. T., Weickhardt, A. J. et al. Mechanisms of resistance to crizotinib in patients with ALK gene rearranged non-small cell lung cancer. Clin Cancer Res 18, 1472-1482 (2012).

227. Trapnell, C., Pachter, L. & Salzberg, S. L. TopHat: discovering splice junctions with RNA-Seq. Bioinformatics 25, 1105-1111 (2009).

228. Trapnell, C., Williams, B. A., Pertea, G., Mortazavi, A., Kwan, G., van Baren, M. J. et al. Transcript assembly and quantification by RNA-Seq reveals unannotated transcripts and isoform switching during cell differentiation. Nat Biotechnol 28, 511-515 (2010).

229. Li, H., Handsaker, B., Wysoker, A., Fennell, T., Ruan, J., Homer, N. et al. The Sequence Alignment/Map format and SAMtools. Bioinformatics 25, 2078-2079 (2009).

230. Li, H. A statistical framework for SNP calling, mutation discovery, association mapping and population genetical parameter estimation from sequencing data. Bioinformatics 27, 2987-2993 (2011).

231. Robinson, J. T., Thorvaldsdottir, H., Winckler, W., Guttman, M., Lander, E. S., Getz, G. et al. Integrative genomics viewer. Nat Biotechnol 29, 24-26 (2011).

232. Dolecek, T. A., Propp, J. M., Stroup, N. E. & Kruchko, C. CBTRUS statistical report: primary brain and central nervous system tumors diagnosed in the United States in 2005-2009. Neuro Oncol 14 Suppl 5, v1-49 (2012).

233. Wen, P. Y. & Kesari, S. Malignant gliomas in adults. N Engl J Med 359, 492-507 (2008).

234. Sturm, D., Bender, S., Jones, D. T., Lichter, P., Grill, J., Becher, O. et al. Paediatric and adult glioblastoma: multiform (epi)genomic culprits emerge. Nat Rev Cancer 14, 92-107 (2014).

235. Giese, A., Bjerkvig, R., Berens, M. E. & Westphal, M. Cost of migration: invasion of malignant gliomas and implications for treatment. J Clin Oncol 21, 1624-1636 (2003).

236. Keunen, O., Johansson, M., Oudin, A., Sanzey, M., Rahim, S. A., Fack, F. et al. Anti-VEGF treatment reduces blood supply and increases tumor cell invasion in glioblastoma. Proc Natl Acad Sci U S A 108, 3749-3754 (2011).

137

237. Linger, R. M., Keating, A. K., Earp, H. S. & Graham, D. K. TAM receptor tyrosine kinases: biologic functions, signaling, and potential therapeutic targeting in human cancer. Adv Cancer Res 100, 35-83 (2008).

238. Keating, A. K., Kim, G. K., Jones, A. E., Donson, A. M., Ware, K., Mulcahy, J. M. et al. Inhibition of Mer and Axl receptor tyrosine kinases in astrocytoma cells leads to increased apoptosis and improved chemosensitivity. Mol Cancer Ther 9, 1298-1307 (2010).

239. Qian, F., Engst, S., Yamaguchi, K., Yu, P., Won, K. A., Mock, L. et al. Inhibition of tumor cell growth, invasion, and metastasis by EXEL-2880 (XL880, GSK1363089), a novel inhibitor of HGF and VEGF receptor tyrosine kinases. Cancer Res 69, 8009-8016 (2009).

240. Yau, T. C., Lencioni, R., Sukeepaisarnjaroen, W., Chao, Y., Yen, C. J., Lausoontornsiri, W. et al. A Phase I/II Multicenter Study of Single-Agent Foretinib as First-Line Therapy in Patients with Advanced Hepatocellular Carcinoma. Clin Cancer Res (2016).

241. Choueiri, T. K., Vaishampayan, U., Rosenberg, J. E., Logan, T. F., Harzstark, A. L., Bukowski, R. M. et al. Phase II and biomarker study of the dual MET/VEGFR2 inhibitor foretinib in patients with papillary renal cell carcinoma. J Clin Oncol 31, 181-186 (2013).

242. Knubel, K. H., Pernu, B. M., Sufit, A., Nelson, S., Pierce, A. M. & Keating, A. K. MerTK inhibition is a novel therapeutic approach for glioblastoma multiforme. Oncotarget 5, 1338-1351 (2014).

243. Zhang, W., DeRyckere, D., Hunter, D., Liu, J., Stashko, M. A., Minson, K. A. et al. UNC2025, a potent and orally bioavailable MER/FLT3 dual inhibitor. J Med Chem 57, 7031-7041 (2014).

244. Cummings, C. T., Zhang, W., Davies, K. D., Kirkpatrick, G. D., Zhang, D., DeRyckere, D. et al. Small Molecule Inhibition of MERTK Is Efficacious in Non- Small Cell Lung Cancer Models Independent of Driver Oncogene Status. Mol Cancer Ther 14, 2014-2022 (2015).

245. Mitra, S. K., Hanson, D. A. & Schlaepfer, D. D. Focal adhesion kinase: in command and control of cell motility. Nat Rev Mol Cell Biol 6, 56-68 (2005).

246. Faria, C. C., Golbourn, B. J., Dubuc, A. M., Remke, M., Diaz, R. J., Agnihotri, S. et al. Foretinib is effective therapy for metastatic sonic hedgehog medulloblastoma. Cancer Res 75, 134-146 (2015).

247. Wang, T., Nelson, R. A., Bogardus, A. & Grannis, F. W., Jr. Five-year lung cancer survival: which advanced stage nonsmall cell lung cancer patients attain long-term survival? Cancer 116, 1518-1525 (2010).

138

248. Berge, E. M. & Doebele, R. C. Targeted therapies in non-small cell lung cancer: emerging oncogene targets following the success of epidermal growth factor receptor. Semin Oncol 41, 110-125 (2014).

249. Mok, T. S., Wu, Y. L., Thongprasert, S., Yang, C. H., Chu, D. T., Saijo, N. et al. Gefitinib or carboplatin-paclitaxel in pulmonary adenocarcinoma. N Engl J Med 361, 947-957 (2009).

250. Bean, J., Brennan, C., Shih, J. Y., Riely, G., Viale, A., Wang, L. et al. MET amplification occurs with or without T790M mutations in EGFR mutant lung tumors with acquired resistance to gefitinib or erlotinib. Proc Natl Acad Sci U S A 104, 20932-20937 (2007).

251. Sequist, L. V., Waltman, B. A., Dias-Santagata, D., Digumarthy, S., Turke, A. B., Fidias, P. et al. Genotypic and histological evolution of lung cancers acquiring resistance to EGFR inhibitors. Sci Transl Med 3, 75ra26 (2011).

252. Ware, K. E., Hinz, T. K., Kleczko, E., Singleton, K. R., Marek, L. A., Helfrich, B. A. et al. A mechanism of resistance to gefitinib mediated by cellular reprogramming and the acquisition of an FGF2-FGFR1 autocrine growth loop. Oncogenesis 2, e39 (2013).

253. Stransky, N., Cerami, E., Schalm, S., Kim, J. L. & Lengauer, C. The landscape of kinase fusions in cancer. Nat Commun 5, 4846 (2014).

254. Wang, R., Hu, H., Pan, Y., Li, Y., Ye, T., Li, C. et al. RET fusions define a unique molecular and clinicopathologic subtype of non-small-cell lung cancer. J Clin Oncol 30, 4352-4359 (2012).

255. Cortes, J. E., Kim, D. W., Pinilla-Ibarz, J., le Coutre, P., Paquette, R., Chuah, C. et al. A phase 2 trial of ponatinib in Philadelphia chromosome-positive leukemias. N Engl J Med 369, 1783-1796 (2013).

256. Mologni, L., Redaelli, S., Morandi, A., Plaza-Menacho, I. & Gambacorti- Passerini, C. Ponatinib is a potent inhibitor of wild-type and drug-resistant gatekeeper mutant RET kinase. Mol Cell Endocrinol 377, 1-6 (2013).

257. Suzuki, M., Makinoshima, H., Matsumoto, S., Suzuki, A., Mimaki, S., Matsushima, K. et al. Identification of a lung adenocarcinoma cell line with CCDC6-RET fusion gene and the effect of RET inhibitors in vitro and in vivo. Cancer Sci 104, 896-903 (2013).

258. Matsubara, D., Kanai, Y., Ishikawa, S., Ohara, S., Yoshimoto, T., Sakatani, T. et al. Identification of CCDC6-RET fusion in the human lung adenocarcinoma cell line, LC-2/ad. J Thorac Oncol 7, 1872-1876 (2012).

139

259. De Falco, V., Buonocore, P., Muthu, M., Torregrossa, L., Basolo, F., Billaud, M. et al. Ponatinib (AP24534) is a novel potent inhibitor of oncogenic RET mutants associated with thyroid cancer. J Clin Endocrinol Metab 98, E811-819 (2013).

260. Ishizaka, Y., Ushijima, T., Sugimura, T. & Nagao, M. cDNA cloning and characterization of ret activated in a human papillary thyroid carcinoma cell line. Biochem Biophys Res Commun 168, 402-408 (1990).

261. O'Hare, T., Shakespeare, W. C., Zhu, X., Eide, C. A., Rivera, V. M., Wang, F. et al. AP24534, a pan-BCR-ABL inhibitor for chronic myeloid leukemia, potently inhibits the T315I mutant and overcomes mutation-based resistance. Cancer Cell 16, 401-412 (2009).

262. Drilon, A., Wang, L., Hasanovic, A., Suehara, Y., Lipson, D., Stephens, P. et al. Response to Cabozantinib in patients with RET fusion-positive lung adenocarcinomas. Cancer Discov 3, 630-635 (2013).

263. Kodama, T., Tsukaguchi, T., Satoh, Y., Yoshida, M., Watanabe, Y., Kondoh, O. et al. Alectinib shows potent antitumor activity against RET-rearranged non-small cell lung cancer. Mol Cancer Ther 13, 2910-2918 (2014).

264. Lin, J. J., Kennedy, E., Sequist, L. V., Brastianos, P. K., Goodwin, K. E., Stevens, S. et al. Clinical Activity of Alectinib in Advanced RET-Rearranged Non-Small Cell Lung Cancer. J Thorac Oncol 11, 2027-2032 (2016).

265. Groger, C. J., Grubinger, M., Waldhor, T., Vierlinger, K. & Mikulits, W. Meta- analysis of gene expression signatures defining the epithelial to mesenchymal transition during cancer progression. PLoS One 7, e51136 (2012).

266. Huang, Q., Schneeberger, V. E., Luetteke, N., Jin, C., Afzal, R., Budzevich, M. M. et al. Preclinical Modeling of KIF5B-RET Fusion Lung Adenocarcinoma. Mol Cancer Ther 15, 2521-2529 (2016).

267. Kang, C. W., Jang, K. W., Sohn, J., Kim, S. M., Pyo, K. H., Kim, H. et al. Antitumor Activity and Acquired Resistance Mechanism of Dovitinib (TKI258) in RET-Rearranged Lung Adenocarcinoma. Mol Cancer Ther 14, 2238-2248 (2015).

268. Taparowsky, E., Shimizu, K., Goldfarb, M. & Wigler, M. Structure and activation of the human N-ras gene. Cell 34, 581-586 (1983).

269. Kennedy, N. J., Sluss, H. K., Jones, S. N., Bar-Sagi, D., Flavell, R. A. & Davis, R. J. Suppression of Ras-stimulated transformation by the JNK signal transduction pathway. Genes Dev 17, 629-637 (2003).

270. Kauffmann-Zeh, A., Rodriguez-Viciana, P., Ulrich, E., Gilbert, C., Coffer, P., Downward, J. et al. Suppression of c-Myc-induced apoptosis by Ras signalling through PI(3)K and PKB. Nature 385, 544-548 (1997).

140

271. Prior, I. A., Lewis, P. D. & Mattos, C. A comprehensive survey of Ras mutations in cancer. Cancer Res 72, 2457-2467 (2012).

272. Dziadziuszko, R., Le, A. T., Wrona, A., Jassem, J., Camidge, D. R., Varella- Garcia, M. et al. An Activating KIT Mutation Induces Crizotinib Resistance in ROS1-Positive Lung Cancer. J Thorac Oncol 11, 1273-1281 (2016).

273. Mizuuchi, H., Suda, K., Murakami, I., Sakai, K., Sato, K., Kobayashi, Y. et al. Oncogene swap as a novel mechanism of acquired resistance to epidermal growth factor receptor-tyrosine kinase inhibitor in lung cancer. Cancer Sci 107, 461-468 (2016).

274. Tricker, E. M., Xu, C., Uddin, S., Capelletti, M., Ercan, D., Ogino, A. et al. Combined EGFR/MEK Inhibition Prevents the Emergence of Resistance in EGFR-Mutant Lung Cancer. Cancer Discov 5, 960-971 (2015).

275. Hrustanovic, G., Olivas, V., Pazarentzos, E., Tulpule, A., Asthana, S., Blakely, C. M. et al. RAS-MAPK dependence underlies a rational polytherapy strategy in EML4-ALK-positive lung cancer. Nat Med 21, 1038-1047 (2015).

276. Sasaki, T., Koivunen, J., Ogino, A., Yanagita, M., Nikiforow, S., Zheng, W. et al. A novel ALK secondary mutation and EGFR signaling cause resistance to ALK kinase inhibitors. Cancer Res 71, 6051-6060 (2011).

277. Song, A., Kim, T. M., Kim, D. W., Kim, S., Keam, B., Lee, S. H. et al. Molecular Changes Associated with Acquired Resistance to Crizotinib in ROS1-Rearranged Non-Small Cell Lung Cancer. Clin Cancer Res 21, 2379-2387 (2015).

278. Miyawaki, M., Yasuda, H., Tani, T., Hamamoto, J., Arai, D., Ishioka, K. et al. Overcoming EGFR bypass signal-induced acquired resistance to ALK tyrosine kinase inhibitors in ALK-translocated lung cancer. Mol Cancer Res, molcanres.0211.2016 (2016).

279. Tani, T., Yasuda, H., Hamamoto, J., Kuroda, A., Arai, D., Ishioka, K. et al. Activation of EGFR Bypass Signaling by TGFalpha Overexpression Induces Acquired Resistance to Alectinib in ALK-Translocated Lung Cancer Cells. Mol Cancer Ther 15, 162-171 (2016).

280. Rho, J. K., Choi, Y. J., Kim, S. Y., Kim, T. W., Choi, E. K., Yoon, S. J. et al. MET and AXL inhibitor NPS-1034 exerts efficacy against lung cancer cells resistant to EGFR kinase inhibitors because of MET or AXL activation. Cancer Res 74, 253-262 (2014).

281. Wang, Y., Xia, H., Zhuang, Z., Miao, L., Chen, X. & Cai, H. Axl-altered microRNAs regulate tumorigenicity and gefitinib resistance in lung cancer. Cell Death Dis 5, e1227 (2014).

141

282. Brand, T. M., Iida, M., Stein, A. P., Corrigan, K. L., Braverman, C. M., Luthar, N. et al. AXL mediates resistance to cetuximab therapy. Cancer Res 74, 5152-5164 (2014).

283. Bae, S. Y., Hong, J. Y., Lee, H. J., Park, H. J. & Lee, S. K. Targeting the degradation of AXL receptor tyrosine kinase to overcome resistance in gefitinib- resistant non-small cell lung cancer. Oncotarget 6, 10146-10160 (2015).

284. Falchook, G. S., Ordonez, N. G., Bastida, C. C., Stephens, P. J., Miller, V. A., Gaido, L. et al. Effect of the RET Inhibitor Vandetanib in a Patient With RET Fusion-Positive Metastatic Non-Small-Cell Lung Cancer. J Clin Oncol 34, e141- 144 (2016).

285. Gautschi, O., Zander, T., Keller, F. A., Strobel, K., Hirschmann, A., Aebi, S. et al. A patient with lung adenocarcinoma and RET fusion treated with vandetanib. J Thorac Oncol 8, e43-44 (2013).

286. Platt, A., Morten, J., Ji, Q., Elvin, P., Womack, C., Su, X. et al. A retrospective analysis of RET translocation, gene copy number gain and expression in NSCLC patients treated with vandetanib in four randomized Phase III studies. BMC Cancer 15, 171 (2015).

287. Gautschi, O., Wolf, J., Milia, J., Filleron, T., Carbone, D. P., Camidge, D. R. et al. Targeting RET in patients with RET-rearranged lung cancers: Results from a global registry. 2016 ASCO Annual Meeting

288. Li, G. G., Somwar, R., Joseph, J., Smith, R. S., Hayashi, T., Martin, L. et al. Antitumor Activity of RXDX-105 in Multiple Cancer Types with RET Rearrangements or Mutations. Clin Cancer Res (2016).

289. Soria, J. C., Shepherd, F. A., Douillard, J. Y., Wolf, J., Giaccone, G., Crino, L. et al. Efficacy of everolimus (RAD001) in patients with advanced NSCLC previously treated with chemotherapy alone or with chemotherapy and EGFR inhibitors. Ann Oncol 20, 1674-1681 (2009).

290. Cascone, T., Subbiah, V., Hess, K. R., Nelson, S. K., Nilsson, M. B., Subbiah, I. M. et al. Significant systemic and CNS activity of RET inhibitor vandetanib combined with mTOR inhibitor everolimus in patients with advanced NSCLC with RET fusion. 2016 ASCO Annual Meeting 2016. Chicago, USA.

291. Subbiah, V., Berry, J., Roxas, M., Guha-Thakurta, N., Subbiah, I. M., Ali, S. M. et al. Systemic and CNS activity of the RET inhibitor vandetanib combined with the mTOR inhibitor everolimus in KIF5B-RET re-arranged non-small cell lung cancer with brain metastases. Lung Cancer 89, 76-79 (2015).

292. O'Reilly, K. E., Rojo, F., She, Q. B., Solit, D., Mills, G. B., Smith, D. et al. mTOR inhibition induces upstream receptor tyrosine kinase signaling and activates Akt. Cancer Res 66, 1500-1508 (2006).

142

293. Manfredi, G. I., Dicitore, A., Gaudenzi, G., Caraglia, M., Persani, L. & Vitale, G. PI3K/Akt/mTOR signaling in medullary thyroid cancer: a promising molecular target for cancer therapy. Endocrine 48, 363-370 (2015).

294. Jung, S., Kim, H. W., Lee, J. H., Kang, S. S., Rhu, H. H., Jeong, Y. I. et al. Brain tumor invasion model system using organotypic brain-slice culture as an alternative to in vivo model. J Cancer Res Clin Oncol 128, 469-476 (2002).

295. Ohnishi, T., Matsumura, H., Izumoto, S., Hiraga, S. & Hayakawa, T. A novel model of glioma cell invasion using organotypic brain slice culture. Cancer Res 58, 2935-2940 (1998).

296. Matsumura, H., Ohnishi, T., Kanemura, Y., Maruno, M. & Yoshimine, T. Quantitative analysis of glioma cell invasion by confocal laser scanning microscopy in a novel brain slice model. Biochem Biophys Res Commun 269, 513-520 (2000).

297. Aaberg-Jessen, C., Norregaard, A., Christensen, K., Pedersen, C. B., Andersen, C. & Kristensen, B. W. Invasion of primary glioma- and cell line-derived spheroids implanted into corticostriatal slice cultures. Int J Clin Exp Pathol 6, 546-560 (2013).

298. Owens, L. V., Xu, L., Craven, R. J., Dent, G. A., Weiner, T. M., Kornberg, L. et al. Overexpression of the focal adhesion kinase (p125FAK) in invasive human tumors. Cancer Res 55, 2752-2755 (1995).

299. Hecker, T. P., Grammer, J. R., Gillespie, G. Y., Stewart, J., Jr. & Gladson, C. L. Focal adhesion kinase enhances signaling through the Shc/extracellular signal- regulated kinase pathway in anaplastic astrocytoma tumor biopsy samples. Cancer Res 62, 2699-2707 (2002).

300. Golubovskaya, V. M., Huang, G., Ho, B., Yemma, M., Morrison, C. D., Lee, J. et al. Pharmacologic blockade of FAK autophosphorylation decreases human glioblastoma tumor growth and synergizes with temozolomide. Mol Cancer Ther 12, 162-172 (2013).

301. Scaltriti, M., Elkabets, M. & Baselga, J. Molecular Pathways: AXL, a membrane receptor mediator of resistance to therapy. Vol. 22 (2016).

302. Adiseshaiah, P., Vaz, M., Machireddy, N., Kalvakolanu, D. V. & Reddy, S. P. A Fra-1-dependent, matrix metalloproteinase driven EGFR activation promotes human lung epithelial cell motility and invasion. J Cell Physiol 216, 405-412 (2008).

303. Lim, S. M., Kim, E. Y., Kim, H. R., Ali, S. M., Greenbowe, J. R., Shim, H. S. et al. Genomic profiling of lung adenocarcinoma patients reveals therapeutic targets and confers clinical benefit when standard molecular testing is negative. Oncotarget 7, 24172-24178 (2016).

143

304. Karachaliou, N., Codony-Servat, J., Teixido, C., Pilotto, S., Drozdowskyj, A., Codony-Servat, C. et al. BIM and mTOR expression levels predict outcome to erlotinib in EGFR-mutant non-small-cell lung cancer. Sci Rep 5, 17499 (2015).

305. Redaelli, S., Ceccon, M., Antolini, L., Rigolio, R., Pirola, A., Peronaci, M. et al. Synergistic activity of ALK and mTOR inhibitors for the treatment of NPM-ALK positive lymphoma. Oncotarget (2016).

306. Chen, B., Tan, Z., Gao, J., Wu, W., Liu, L., Jin, W. et al. Hyperphosphorylation of ribosomal protein S6 predicts unfavorable clinical survival in non-small cell lung cancer. J Exp Clin Cancer Res 34, 126 (2015).

144