Pharmacologic Inhibition of Ubiquitin Activation As a Novel Therapeutic Strategy in Acute Myeloid Leukemia

Pharmacologic Inhibition of Ubiquitin Activation as a Novel

Therapeutic Strategy in Acute Myeloid Leukemia

Samir Hamouda Barghout

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

Pharmacologic Inhibition of Ubiquitin Activation as a Novel Therapeutic Strategy in Acute Myeloid Leukemia

Samir Hamouda Barghout

Doctor of Philosophy

Department of Medical Biophysics University of Toronto

2020 Abstract

Acute myeloid leukemia (AML) is an aggressive hematologic malignancy for which new therapeutic approaches are required. One such potential therapeutic strategy is to target ubiquitin- like modifier activating enzyme 1 (UBA1), the initiating enzyme in the ubiquitylation cascade in which proteins are tagged with ubiquitin moieties to regulate their degradation or function. Here, we evaluated TAK-243, a first-in-class UBA1 inhibitor, in preclinical models of AML and identified potential determinants of sensitivity to this drug. In AML cell lines and primary AML samples, TAK-243 induced cell death and inhibited clonogenic growth. In contrast, normal hematopoietic progenitor cells were more resistant. TAK-243 preferentially bound to UBA1 over the related enzymes UBA2, UBA3 and UBA6 in intact AML cells. Inhibition of UBA1 with

TAK-243 decreased levels of ubiquitylated proteins, increased markers of proteotoxic stress and

DNA damage stress. In vivo, TAK-243 reduced leukemic burden and targeted leukemic stem cells without evidence of toxicity. We selected populations of AML cells resistant to TAK-243 and identified missense mutations in the adenylation domain of UBA1. To identify the determinants of TAK-243 sensitivity, we conducted a genome-wide CRISPR/Cas9 knockout screen in AML cells in the presence of TAK-243. We identified BEN domain-containing protein

3 (BEND3), a transcriptional repressor and a regulator of chromatin organization, as the top gene ii whose knockout conferred resistance to TAK-243 in vitro and in vivo. BEND3 knockout dampened TAK-243 effects on ubiquitylation, proteotoxic stress and DNA damage response. It also upregulated the ABC efflux transporter breast cancer resistance protein (BCRP) resulting in increased efflux of TAK-243 out of the cells. In addition, TAK-243 sensitivity correlated with

BCRP expression in cancer cell lines of different origin and chemical inhibition of BCRP sensitized intrinsically resistant high-BCRP cells to TAK-243. Thus, our data demonstrate that

TAK-243 targets AML cells and stem cells and support a clinical trial of TAK-243 in this patient population. We also provide insights into potential mechanisms of acquired resistance to TAK-

243. Moreover, our data demonstrate that BEND3 regulates the expression of BCRP for which

TAK-243 is a substrate. Therefore, BCRP expression could serve as a predictor of TAK-243 sensitivity in different malignancies.

iii

Dedication

My parents, Yasmin, Salma, Adam and Maha

Acknowledgments

I would like to express my gratitude to my supervisor, Dr. Aaron Schimmer, for his academic guidance and mentorship during my PhD research. He provided me with a great opportunity to work in a well-structured laboratory and a friendly environment which significantly boosted my research. I am also thankful to my supervisory committee members, Dr. Linda Penn and Dr. Marianne Koritzinsky, for their continued support, critical discussions and insightful suggestions that considerably helped move my project forward. I would like to thank my laboratory members, G. Wei Xu, Neil MacLean, Xiaoming Wang, Rose Hurren, Marcela Gronda and Yulia Jitkova, for their technical and scientific support. Without their efforts, the completion of this project would not have been possible. I would like to thank all my laboratory colleagues for their scientific input in weekly meetings and for daily discussions on personal, social and career issues. I will always remember them as great fellows without whom the PhD journey would have become longer and harsher. I owe special thanks to Simon Kavanagh, Zack Blatman, Karen Arevalo, Geethu Thomas and Sarah Zarabi for their appreciable contributions to this work. I would like to thank our collaborators from University Health Network (UHN)—Parasvi Patel, Dr. Razq Hakem, Dr. Mark Minden, Ondrej Halgas and Dr. Emil Pai, from Takeda Pharmaceuticals—Dr. Marc Hyer, Dr. Allison Berger, Dr. Tary Traore, Dr. Michael Sintchak, Dr. Michael Milhollen and Dr. James Brownell, from Ontario Institute of Cancer Research (OICR)—Dr. Ahmed Aman and Dr. Rima Al-Awar, and from Zewail City of Science and Technology—Moustafa Abohawya, for their invaluable support and significant contributions to this research. I would like to thank Jing Xu at the Applied Molecular Profiling Laboratory (AMPL) facility for immunohistochemistry staining and technical staff at the Advanced Optical Microscopy Facility (AOMF) at Princess Margaret Cancer Centre for imaging. I would like to thank Jill Flewelling and Francesca Pulice at Princess Margaret Cancer Centre for administrative assistance. I also thank the Department of Medical Biophysics, University of Toronto (UofT) and its administrative staff for their continued support throughout the PhD program particularly for their financial support to attend the on-site admission interviews. I also thank Dr. Linda Penn and Dr. Rama Khokha for rotations in their laboratories.

I would like to acknowledge financial support by the Ontario Trillium Scholarship, Department of Medical Biophysics fellowship, Graduate Studies Endowment Fund (GSEF) scholarship, and Queen Elizabeth II Graduate Scholarship from the Faculty of Medicine, University of Toronto. In addition, I would like to thank the Canadian Cancer Society (CCS), American Society of Hematology (ASH), Canadian Hematology Society (CHS), Office of Research Trainees (ORT) at UHN, School of Graduate Studies (SGS) and Division of Hematology at UofT for travel awards to present my research at international conferences. This work was supported by research grants from Takeda Pharmaceuticals Inc., the Princess Margaret Cancer Centre Foundation, Canadian Institutes of Health Research (CIHR), OICR, the Canada Research Chairs program, the Ontario Ministry of Research and Innovation, the Ministry of Long-Term Health and Planning in the Province of Ontario, and the Leukemia & Lymphoma Society of Canada.

Contributions

The majority of work presented in this thesis was conducted by Samir H. Barghout. Technical assistance was contributed as follows: Rose Hurren provided technical assistance with RT-qPCR. Rose Hurren and Xiaoming Wang performed and analyzed animal studies. Neil MacLean provided technical assistance with the production of lentivirus and CRISPR/Cas9 screens. Marcela Gronda provided technical assistance with colony-forming assays. G. Wei Xu provided technical assistance with cloning and sequencing experiments. Parasvi Patel, Dr. Razq Hakem contributed to immunofluorescence experiments outlined in Figure 2.7 and 2.8. Ondrej Halgas, Dr. Emil Pai and Sarah Zarabi contributed to structural modeling experiments outlined in Figure 2.13F. Simon Kavanagh contributed to cytotoxicity experiments outlined in Figure 2.1E. Geethu Thomas contributed to clonogenic experiments outlined in Figure 3.2F. Ahmed Aman and Rima Al-Awar contributed to LC-MS experiments outlined in Figure 3.5F. Moustafa Abohawya and Troy Ketela contributed to bioinformatic analysis of CRISPR/Cas9 screen data. Zachary Blatman and Karen Arevalo performed immunoblotting experiments as part of their summer studentship programs. They contributed to Figure 3.5. Marc Hyer, Allison Berger, Tary Traore, Michael Sintchak, and Michael Milhollen from Takeda Pharmaceuticals provided TAK-243 and technical advice with animal studies. Chapter 1 is to be submitted as a review article. The full citation will be: Barghout SH, Schimmer AD. E1 enzymes as therapeutic targets in cancer. To be submitted. Chapter 2 has been published as a research article in Leukemia journal. The full citation of this article is as follows: Barghout SH, Patel PS, Wang X, Xu GW, Kavanagh S, Halgas O, Zarabi SF, Gronda M, Hurren R, Jeyaraju DV, MacLean N, Brennan S, Hyer ML, Berger A, Traore T, Milhollen M, Smith AC, Minden MD, Pai EF, Hakem R, Schimmer AD. Preclinical evaluation of the selective small-molecule UBA1 inhibitor, TAK-243, in acute myeloid leukemia. Leukemia. 2019;33(1):37-51.

Chapter 3 has been prepared as a manuscript for submission to an appropriate journal. The full author list will be as follows: Barghout SH, Aman A, Blatman Z, Arevalo K, Thomas G, MacLean N, Wang X, Hurren R, Ketela T, Abohawya M, Al-Awar R, and Schimmer AD. BEND3 controls sensitivity to the selective UBA1 inhibitor, TAK-243, via regulating BCRP expression. To be submitted.

vii

Table of Contents

Dedication ...... iv

Acknowledgments ...... v

Contributions ...... vii

Table of Contents ...... viii

List of Tables ...... xiii

List of Figures ...... xv

List of Appendices ...... xvii

List of Abbreviations ...... xviii

Chapter 1 Introduction ...... 1

1.1 UBL conjugation system ...... 1

1.2 E1 enzymes ...... 3

1.2.1 Members of the E1 activating enzyme class ...... 3

1.2.2 Biochemical and structural mechanisms of E1-catalyzed UBL activation ...... 6

1.2.3 Druggability of E1 enzymes ...... 7

1.2.4 Rationale for targeting E1 enzymes in cancer ...... 10

1.3 NEDD8-activating enzyme (NAE) ...... 16

1.3.1 Role of NAE in cancer ...... 17

1.3.2 NAE inhibitors ...... 17

1.4 Ubiquitin-like modifier-activating enzyme 1 (UBA1; UBE1) ...... 39

1.4.1 Role of UBA1 in cancer ...... 42

1.4.2 UBA1 inhibitors ...... 42

1.5 SUMO-activating enzyme (SAE) ...... 53

1.5.1 Role of SAE in cancer ...... 53

1.5.2 SAE inhibitors ...... 54

1.6 Adenosine sulfamates ...... 67 viii

1.6.1 Discovery and structure-activity relationship ...... 67

1.6.2 Substrate-assisted inhibition ...... 70

1.6.3 Selectivity and potency of adenosine sulfamates ...... 72

1.6.4 Resistance to adenosine sulfamates ...... 72

1.6.5 Pharmacodynamic activity of adenosine sulfamates ...... 75

1.6.6 Cell cycle effects of adenosine sulfamates ...... 76

1.6.7 Clinical adenosine sulfamates: pevonedistat and TAK-243 ...... 76

1.7 Acute myeloid leukemia ...... 79

1.7.1 Overview ...... 79

1.7.2 Pathobiology of AML ...... 79

1.7.3 Mutational landscape of AML ...... 80

1.7.4 Therapy of AML ...... 84

1.8 Rationale for targeting UBA1 in AML ...... 87

Chapter 2 Preclinical evaluation of TAK-243 in AML ...... 91

2.1 Introduction ...... 91

2.2 Methods ...... 92

2.2.1 Compounds and reagents ...... 92

2.2.2 Cell lines ...... 92

2.2.3 Primary AML and normal hematopoietic cells ...... 93

2.2.4 Cytotoxicity assays ...... 93

2.2.5 Cellular thermal shift assay (CETSA) ...... 93

2.2.6 Immunoblotting ...... 94

2.2.7 Overexpression of GRP78 ...... 94

2.2.8 Immunofluorescence ...... 94

2.2.9 Immunohistochemistry (IHC) ...... 95

2.2.10 Animal studies ...... 98 ix

2.2.11 Generation of TAK-243-resistant AML cell lines ...... 99

2.2.12 Detection of UBA1 mutations in TAK-243-resistant AML cell line ...... 99

2.2.13 Statistical and data analysis ...... 99

2.3 Results ...... 102

2.3.1 TAK-243 induces cell death and decreases clonogenic growth in AML cell lines and primary cells ...... 102

2.3.2 TAK-243 preferentially binds to UBA1 in AML cell lines and primary cells ... 102

2.3.3 TAK-243 reduces ubiquitylation of cellular proteins in AML cells and primary AML samples ...... 110

2.3.4 TAK-243 induces endoplasmic reticulum (ER) stress in AML cells ...... 110

2.3.5 TAK-243 inhibits the DNA damage response ...... 115

2.3.6 TAK-243 reduces the leukemic burden in a mouse xenograft model of AML .. 118

2.3.7 TAK-243 reduces engraftment of human primary AML cells in the bone marrow of mice ...... 118

2.3.8 TAK-243 preferentially targets UBA1 in vivo ...... 123

2.3.9 Missense mutations in the adenylation domain of UBA1 confer resistance to TAK-243 ...... 123

2.4 Discussion ...... 137

2.5 Summary and conclusion ...... 138

Chapter 3 Determinants of TAK-243 sensitivity in AML ...... 142

3.1 Introduction ...... 142

3.2 Methods ...... 142

3.2.1 Chemicals and reagents ...... 143

3.2.2 Cell culture ...... 143

3.2.3 Positive-selection genome-wide CRISPR/Cas9 knockout screen ...... 143

3.2.4 CRISPR/Cas9 knockout experiments ...... 144

3.2.5 Cytotoxicity assays ...... 144

3.2.6 Cellular thermal shift assay (CETSA) ...... 145 x

3.2.7 Quantitative reverse transcription polymerase chain reaction (RT-qPCR) ...... 145

3.2.8 Immunoblotting ...... 147

3.2.9 Determination of intracellular ATP levels ...... 147

3.2.10 Measurement of intracellular TAK-243 concentrations ...... 147

3.2.11 Animal studies ...... 148

3.2.12 Statistical and data analysis ...... 148

3.3 Results ...... 150

3.3.1 A positive-selection genome-wide CRISPR/Cas9 knockout screen identifies BEND3 as a top hit ...... 150

3.3.2 BEND3 knockout confers resistance to TAK-243 in AML cells ...... 150

3.3.3 BEND3 knockout confers resistance to TAK-243 in vivo ...... 160

3.3.4 BEND3 knockout dampens TAK-243 effects on ubiquitylation, proteotoxic stress and DNA damage response in AML cells ...... 160

3.3.5 BEND3 knockout reduces intracellular transport of TAK-243 into AML cells . 160

3.3.6 Upregulation of BCRP mediates TAK-243 resistance upon BEND3 knockout in AML cells ...... 165

3.3.7 BEND3 knockout confers partial cross-resistance to related adenosine sulfamates and selected MDR substrates ...... 169

3.3.8 TAK-243 is a substrate for BCRP in cell lines of different origin ...... 169

3.4 Discussion ...... 173

3.5 Summary and conclusions ...... 175

Chapter 4 Summary and future directions ...... 177

4.1 Summary ...... 177

4.2 Future directions ...... 178

4.2.1 Mechanistic insights into BEND3-mediated regulation of BCRP expression .... 178

4.2.2 The role of KMT5B/C in TAK-243 resistance ...... 179

4.2.3 Correlative biomarkers of TAK-243 response ...... 179

4.2.4 SAE as a potential therapeutic target in AML ...... 180 xi

References ...... 182

Appendix Copyright and permissions ...... 212

xii

List of Tables

Page

Table 1.1 E1 enzymes, their cognate E2 enzymes, UBLs and UBL proteases 8

Table 1.2 NAE inhibitors, their chemical structures and pharmacological properties 19

Table 1.3 Preclinical and clinical studies of pevonedistat 25

Table 1.4 Clinical trials of pevonedistat 32

Table 1.5 Properties of ubiquitin-activating enzymes: UBA1 and UBA6 41

Table 1.6 UBA1 inhibitors, their chemical structures and pharmacological 43 properties

Table 1.7 Clinical trials of TAS4464, TAK-243 and TAK-981 52

Table 1.8 SAE inhibitors, their chemical structures and pharmacological properties 55

Table 1.9 Multitarget E1 inhibitors, their structures and pharmacological properties 66

Table 1.10 Activity of pevonedistat and TAK-243 against E1 enzymes as assessed 78 by cell-free transthiolation assays

Table 1.11 Currently used and investigational agents in AML therapy 85

Table 1.12 New drug approvals in AML 86

Table 2.1 Antibodies used in immunoblotting, IHC and immunofluorescence, their 96 sources and dilution

Table 2.2 Primers used for amplifying UBA1 exons 101

Table 2.3 Conditions used for amplifying UBA1 exons 101

Table 2.4 Characteristics of patient samples used in the study 106

Table 3.1 BEND3-targeting gRNA sequences 146

xiii

Table 3.2 Primers used for RT-qPCR 146

Table 3.3 Antibodies used in Western blotting, their sources and dilution 149

Table 3.4 Top hits in the IC90 arm of the CRISPR/Cas9 knockout screen 153

Table 3.5 Top hits in the IC99 arm of the CRISPR/Cas9 knockout screen 154

Table 3.6 Top depleted hits in the IC90 arm of the CRISPR/Cas9 knockout screen 155

Table 3.7 Top depleted hits in the IC99 arm of the CRISPR/Cas9 knockout screen 156

xiv List of Figures

Page Figure 1.1 Ubiquitin and UBL conjugation system 4

Figure 1.2 E1 enzymes and their structural domains 5

Figure 1.3 Cascade of E1-catalyzed UBL activation and substrate-assisted 9 inhibition by adenosine sulfamates

Figure 1.4 Inhibitors of E1enzymes and their sites of action 11

Figure 1.5 Cancer dependency of E1 genes 13

Figure 1.6 Essentiality of E1 genes as determined from pan-cancer genome-wide 14 loss-of-function screens

Figure 1.7 Essentiality of E1 genes in cancer cell lines of different origin 15

Figure 1.8 Structural features of adenosine sulfamates 68

Figure 1.9 Discovery of adenosine sulfamates 69

Figure 1.10 Pathobiology of AML 81

Figure 1.11 Mutational landscape of AML 82

Figure 1.12 Functional categories of genes frequently mutated in AML 83

Figure 2.1 Anti-leukemic activity of TAK-243 in AML cell lines and primary cells 103

Figure 2.2 Anti-leukemic activity of TAK-243 in AML cell lines 105

Figure 2.3 TAK-243 preferentially binds UBA1 and reduces levels of global 111 protein ubiquitylation in AML cell lines and primary cells

Figure 2.4 TAK-243 preferentially targets UBA1 in AML cell lines and primary 113 cells

Figure 2.5 Effect of TAK-243 on global protein ubiquitylation and Bcl-2 levels 114

Figure 2.6 TAK-243 induces proteotoxic stress in AML cells 116

Figure 2.7 TAK-243 inhibits DNA double-strand break repair in AML cells 119

Figure 2.8 TAK-243 inhibits histone ubiquitylation in primary patient samples and 121 DNA double-strand break repair in AML cells

Figure 2.9 TAK-243 reduces the leukemic burden in mouse models of AML, and 124 preferentially binds and inhibits UBA1 in AML cells in vivo

Figure 2.10 TAK-243 is effective and tolerable in an AML mouse xenograft model 127

Figure 2.11 TAK-243 binds and inhibits UBA1 in AML cells in vivo 128

Figure 2.12 Y583C and A580S UBA1 mutations confer resistance to TAK-243 in 131 AML cells

Figure 2.13 Y583C mutation in UBA1 confers resistance to TAK-243 133

Figure 2.14 A580S mutation in UBA1 confers resistance to TAK-243 135

Figure 2.15 Working model of TAK-243 cytotoxicity in AML 140

Figure 3.1 A genome-wide CRISPR/Cas9 knockout screen identifies BEND3 as a 151 top hit

Figure 3.2 BEND3 knockout confers resistance to TAK-243 in AML cells 158

Figure 3.3 BEND3 knockout cells are resistant to TAK-243 in an AML mouse 161 xenograft model

Figure 3.4 BEND3 knockout cells are resistant to TAK-243 in vitro and in vivo 162

Figure 3.5 BEND3 knockout dampens TAK-243 effects by reducing the 163 intracellular transport of the drug into AML cells

Figure 3.6 Upregulation of BCRP mediates TAK-243 resistance upon BEND3 166 knockout in AML cells

Figure 3.7 Correlation of ABC transporter expression and TAK-243 sensitivity 168

Figure 3.8 BEND3 knockout confers partial cross-resistance to related adenosine 170 sulfamates and selected MDR substrates

Figure 3.9 BEND3 knockout does not affect response to proteasome inhibitors or 171 ER stressors

Figure 3.10 TAK-243 is a substrate for BCRP in cell lines of different origin 172

xvi

List of Appendices

Page

Appendix: Copyright and permissions 212

xvii

List of Abbreviations

Abbreviation Full name 53BP1 TP53-binding protein 1 AAD Active adenylation domain ABC ATP-binding cassette ABCG2 ATP-binding cassette super-family G member 2 AD Adenylation domain ADS Adenosine sulfamate ALP Alkaline phosphatase AML Acute myeloid leukemia AML-MRC AML with myelodysplasia-related change AMP Adenosine monophosphate ANOVA Analysis of variance APL Acute promyelocytic leukemia APN Adenosyl-phospho-NEDD8 APPBP1 Amyloid beta precursor protein-binding protein 1 APU Adenosyl-phospho-ubiquitinol AST Aspartate transaminase ATF4/6 Activating transcription factor 4/6 ATG12 Autophagy-related protein 12 ATG7 Autophagy-related protein 7 ATG8 Autophagy-related protein 8 ATO Arsenic trioxide ATP Adenosine triphosphate ATRA All-trans retinoic acid BACH2 BTB Domain And CNC Homolog 2 BCL B-cell lymphoma Bcl-2 B-cell lymphoma 2 BCRP Breast cancer resistance protein BEND3 BEN domain-containing protein 3 BFU-E Burst-forming unit-erythroid

xviii

BID Twice daily BSA Bovine serum albumin Cas9 CRISPR associated protein 9 CC Cervical carcinoma CCD Catalytic cysteine domain CCLE Cancer Cell Line Encyclopedia CCRCC Clear cell renal cell carcinoma CDK Cyclin-dependent kinase CDT1 Chromatin licensing and DNA replication factor 1 CETSA Cellular thermal shift assay CFU-GM Colony-forming unit-granulocyte, monocyte cGAS Cyclic GMP-AMP synthase CHIP Clonal hematopoiesis of indeterminate potential CHOP CCAAT-enhancer-binding protein (C/EBP) homologous protein CI Confidence interval CK Creatinine kinase CLL Chronic lymphocytic leukemia CML Chronic myeloid leukemia CMML Chronic Myelomonocytic Leukemia CNVs Copy number variations CR Complete response/remission CRC Colorectal cancer CRISPR Clustered regularly interspaced short palindromic repeats CRL Cullin-RING ubiquitin ligase CSC Cancer stem cell CSN5 COP9 signalosome complex subunit 5 DAB 3,3’-Diaminobenzidine tetrahydrochloride DAXX Death domain-associated protein 6 DC Dendritic cells DEN1 Deneddylase 1 DLBCL Diffuse large B-cell lymphoma DMSO Dimethyl sulfoxide

xix

DOT1L Disruptor of telomeric silencing 1-like DSB Double-strand break DTT Dithiothreitol DUB Deubiquitinating enzyme

EC50 Half-maximal effective concentration EGFR Epidermal growth factor receptor ELN European LeukemiaNet EMT Epithelial-to-mesenchymal transition EOC Epithelial ovarian cancer ER Endoplasmic reticulum ES Ewing sarcoma ESCC Esophageal squamous cell carcinoma EZH2 Enhancer of zeste homolog 2 FAB French-American-British FCCH First catalytic cysteine half-domain FDA Food and Drug Administration FDR False discovery rate FLT3 FMS-like receptor tyrosine kinase-3 GBM Glioblastoma G-CSF Granulocyte-colony stimulating factor GFP Green fluorescent protein GO Gene ontology gRNA Guide RNA GRP78 78 kDa glucose-regulated protein GSH Glutathione GST Glutathione S-transferases Gy Gray (unit of ionizing radiation dose) HCC Hepatocellular carcinoma HDAC Histone deacetylase HiDAC High-dose cytarabine HIF Hypoxia-inducible factor HNSCC Head and neck squamous cell carcinoma

HPBCD 2-hydroxypropyl-β-cyclodextrin HR Homologous recombination HSC Hematopoietic stem cell HSCT Hematopoietic stem cell transplantation HTS High-throughput screen IAD Inactive adenylation domain

IC50 Half-maximal inhibitory concentration ICC Intrahepatic cholangiocarcinoma IDH1/2 Isocitrate dehydrogenase 1/2 IFN Interferon IFNAR IFN α/β receptor IHC Immunohistochemistry IP Intraperitoneal IR Ionization radiation IRE1α Inositol-requiring transmembrane kinase/endoribonuclease 1α IRF9 IFN-regulatory factor 9 ISG15 Interferon-stimulated gene 15 ISGF3 IFN-stimulated gene factor 3 ISRE IFN-stimulated response element ITD Internal tandem duplication IV Intravenous IκBα NF-kappa-B inhibitor alpha JNK c-Jun N-terminal kinase 1 K Lysine kDa Kilodalton Ki Inhibitory constant KMT5B/C Lysine N-methyltransferase 5B KO Knockout LBCL Large B-cell Lymphoma LC-MS Liquid chromatography–mass spectrometry LRH-1 Liver receptor homolog-1 LSC Leukemic stem cell

xxi

LSD1 Lysine-specific demethylase 1A MAGeCK Model-based Analysis of Genome-wide CRISPR/Cas9 Knockout MCL Mantle cell lymphoma MDR Multi-drug resistance MDS Myelodysplastic syndrome MHC Major histocompatibility complex MM Multiple myeloma MOCS3 Molybdenum cofactor synthesis protein 3 MPN Myeloproliferative neoplasms MRP Multidrug resistance-associated protein MTAs Molecularly targeted agents mTOR Mammalian target of rapamycin NAE NEDD8-activating enzyme NB Nuclear body NCI US National Cancer Institute NEDD8 Neural precursor cell-expressed developmentally downregulated protein 8 NEMO NF-κB essential modulator NES Non-canonical E1-specific domain NF2 Neurofibromatosis type 2 NF-κB Nuclear factor-kappa B NHEJ Non-homologous end-joining NHL Non-Hodgkin Lymphoma NK Natural killer NoRC Nucleolar-remodeling complex NPC Nasopharyngeal carcinoma NPM1 Nucleophosmin NSCLC Non-small cell lung cancer NuRD Nucleosome remodeling and deacetylase ORR Overall response rate OS Osteosarcoma Pan-ub Pan-ubiquitin PARP Poly [ADP-ribose] polymerase 1

xxii

PBS Phosphate-buffered saline PCR Polymerase chain reaction PD Pharmacodynamics PDAC Pancreatic ductal adenocarcinoma PD-L1 Programmed death-ligand 1 PDX Patient derived xenograft PE Plating efficiency PERK PKR-like endoplasmic reticulum kinase P-gp P-glycoprotein PI Proteasome inhibitor PICH PLK1-interacting checkpoint helicase PK Pharmacokinetics PML Promyelocytic Leukemia PPi Inorganic pyrophosphate PPI Protein-protein interaction PR Partial response PRC2 Polycomb repressive complex 2 PRL Prolactin PTM Post-translational modifications PVDF Polyvinylidene difluoride q.o.d Every other day QD Once a day QTc Corrected QT QW Once weekly RANGAP1 Ran GTPase Activating Protein 1 RARA Retinoic Acid Receptor-Alpha rDNA Ribosomal DNA rh Recombinant human RHD Rhodanese homology domain rh-SCF Recombinant human stem cell factor RING Really Interesting New Gene RIPA Radio-immunoprecipitation assay

xxiii

RNAi RNA interference RNAseq RNA sequencing RNF4 Ring Finger Protein 4 ROS Reactive oxygen species RP2D Recommended phase 2 dose RRM2 Ribonucleoside-diphosphate reductase RT-qPCR Quantitative reverse transcription polymerase chain reaction SAE SUMO-activating enzyme SAR Structure-activity relationship sc Subcutaneous SCC Squamous cell carcinoma SCCH Second catalytic cysteine half-domain SCID Severe combined immunodeficiency SCLC Small-cell lung cancer SD Standard deviation SDS-PAGE Sodium dodecyl sulfate-polyacrylamide gel electrophoresis SEM Standard error of the mean SENP Sentrin-specific protease sgRNA Single guide RNA shRNA Short hairpin RNA STING Stimulator of interferon genes protein SUMO Small ubiquitin-like modifier SUV4-20H1/2 Suppressor of variegation 4-20 homolog ½ T-ALL T-cell acute lymphoblastic leukemia t-AML Therapy-related AML TGFβR Tumor-growth factor beta receptor TIW 3 times weekly TKD Tyrosine kinase domain Tregs Regulatory T cells UBA1 Ubiquitin-like modifier-activating enzyme 1 UBA4 Ubiquitin-like modifier-activating enzyme 4 UBA6 Ubiquitin-like modifier-activating enzyme 6

xxiv

UBA7 Ubiquitin-like modifier-activating enzyme 7 UBC12 Ubiquitin carrier protein 12 UBC9 Ubiquitin carrier protein 9 UBL Ubiquitin-like protein UC Urothelial carcinoma UFC1 Ubiquitin-fold modifier-conjugating enzyme 1 UFD Ubiquitin-fold domain UFM1 Ubiquitin-fold modifier 1 UFSP1/2 UFM1-specific protease 1/2 ULP UBL-specific protease UM Uveal melanoma UPR Unfolded protein response UPS Ubiquitin-proteasome system URM1 Ubiquitin-related modifier 1 USP18 Ubiquitin specific peptidase 18 VHL Von Hippel–Lindau XBP1 X-box binding protein 1 γH2AXSer139 Ser139 phosphorylated H2AX

xxv Chapter 1

Introduction

Chapter 1 Introduction

1.1 UBL conjugation system

Ubiquitin is an 8.5 kDa, 76-amino acid protein that serves as a post-translational modifier of cellular proteins1. Similar to other post-translational modifications (PTMs), the process of reversible enzymatic attachment of ubiquitin to proteins is called ubiquitylation. This PTM is involved in modulating turnover, function, interaction, or localization of cellular proteins and therefore regulates a wide range of biological processes2. These include, among other functions, protein homeostasis, cell cycle regulation, DNA repair, transcriptional regulation and endocytosis. Given its indispensable biological roles, ubiquitin is evolutionarily conserved among eukaryotes and, as the name suggests, is ubiquitously expressed in most tissues3. In eukaryotes, a group of proteins, collectively known as ubiquitin-like proteins (UBLs), share sequence homology and have a similar three-dimensional structure to ubiquitin4, 5. There are more than a dozen UBLs which are classified into 8 families including: Neural precursor cell- expressed developmentally downregulated protein 8 (NEDD8), Small ubiquitin-like modifier (SUMO), FAT10 (ubiquitin D), Interferon-stimulated gene 15 (ISG15), Autophagy-related protein 8 (ATG8), ATG12, Ubiquitin-fold modifier 1 (UFM1) and Ubiquitin-related modifier 1 (URM1) protein families. These UBLs, known as type I UBLs, are similarly involved in PTMs that modulate a wide variety of cellular processes. In contrast, type II UBLs are not conjugated to protein substrates, but rather exist as a part of proteins with multiple domains and these include HUB1, ESC2 and FUBI. Here, our focus will be on ubiquitin and type I UBLs to which we will simply refer to as UBLs6. PTMs with UBLs are brought about by a sequential enzymatic cascade that involves three classes of enzymes: E1 activating enzymes, E2 conjugating enzymes and E3 ligases7. The function of E1 enzymes is to activate UBLs in an adenosine triphosphate (ATP)-dependent multi-step enzymatic reaction whereby they become attached to the catalytic cysteine site of the E1 enzyme via a high-energy thioester bond (symbolized by tilde [∼]), forming a UBL-loaded E1 enzyme (E1∼UBL). The E1-bound UBL is then transferred from the catalytic cysteine of the E1 enzyme to the catalytic cysteine of a cognate E2 enzyme via a transthiolation reaction, forming a UBL-loaded E2 enzyme (E2∼UBL)6. In concert with their cognate E3 ligases, the

1 2

UBL is conjugated to protein substrates by the E2 enzyme, forming a UBL-modified substrate usually via an isopeptide linkage with a lysine residue8. E3 ligases may function as either adaptors that bring E2s and protein substrates in close proximity for UBL conjugation (represented by U-box E3s and RING finger E3s), or UBL acceptors that form thioester intermediates with UBLs for subsequent transfer to protein substrates (represented by HECT domain E3s)9. In the human proteome, there exist 8 known E1s, > 40 E2s, and > 600 E3s, with E3 ligases controlling substrate specificity10. As UBL conjugation is a reversible process, the UBL signal can be removed by another set of enzymes, known collectively as deubiquitinating enzymes (DUBs) and UBL-specific proteases (ULPs), which catalyze proteolytic cleavage of these PTMs from their substrates. There are >100 DUBs and ULPs in the human proteome10, 11. A subset of ubiquitylated proteins, specifically those tagged with K48-linked polyubiquitin chains, are identified by the proteasome, deubiquitylated and degraded into peptides to maintain protein homeostasis in the cell. The cellular machinery orchestrating ubiquitin-dependent protein degradation is collectively known as the ubiquitin-proteasome system (UPS)7. Apart from degradative ubiquitylation, conjugation with other forms of ubiquitin and UBLs serves non- degradative functions in many aspects of cellular signaling1 (Fig. 1.1). Historically, drug discovery within the ubiquitin and UBL systems started with two parallel programs targeting the proteasome and E3 ligases. Despite the non-selective nature of proteasome inhibition, this program has led to the discovery of clinically useful drugs such as bortezomib, while the E3 ligase program is still stumbling12. Therefore, it is tempting to explore other signaling nodes in these systems especially E1 enzymes that lie at the apex of UBL conjugation cascade and affect proteasomal degradation as well as other non-degradative signaling processes in the cell8. In this introduction, we highlight different E1 enzymes, their biological roles and pathobiological alterations particularly in the context of cancer. In addition, we discuss in more detail the biochemical mechanisms of UBL activation and therapeutic strategies known so far to target different E1 enzymes with a focus on mechanism-based E1 inhibitors that have been advanced to clinical trials.

1.2 E1 enzymes

1.2.1 Members of the E1 activating enzyme class

In the human proteome, eight E1 enzymes are known to activate UBLs8, 10. These include Ubiquitin-like modifier-activating enzyme 1 (UBA1), NEDD8-activating enzyme (NAE), SUMO-activating enzyme (SAE), UBA6, UBA7, UBA4 (MOCS3), UBA5, and Autophagy- related protein 7 (ATG7). Based on structural and biochemical properties, E1s have been subdivided into canonical (UBA1, NAE, SAE, UBA6, UBA7) and non-canonical (UBA4, UBA5, and ATG7) E1s (reviewed in Ref. 8). Structurally, E1 enzymes adopt a monomeric (UBA1, UBA6 and UBA7), heterodimeric (NAE and SAE), or homodimeric (UBA4, UBA5 and ATG7) architecture (Fig. 1.2). To carry out their multi-step catalytic function, eukaryotic E1 enzymes possess multiple domains including: the adenylation domain (AD), the catalytic cysteine domain (CCD), and the ubiquitin- fold domain (UFD), with several variations existing among different E1 enzymes13. In canonical E1s, the AD is pseudo-symmetric with one active adenylation domain (AAD) involved in identifying and adenylating the C-terminus of cognate UBLs, and another inactive adenylation domain (IAD) involved in structural stability8, 14. Similarly, the CCD is divided into the first (FCCH) and second (SCCH) half-domains and is involved in thioester bond formation with UBLs14. The UFD is involved in the interaction with cognate E2s for UBL transfer via transthiolation. Of these, the AD is the most conserved domain and is homologous to the ancestral prokaryotic MoeB and ThiF domains, suggesting adenylation of UBLs is the most conserved functionality of all E1s13. Other domains might have evolved to accommodate the more complex catalytic activity carried out by eukaryotic E1s as opposed to their prokaryotic homologs6. Non-canonical E1s are homodimeric with symmetric ADs as well as other non- canonical E1-specific domains. Their catalytic cysteines are situated close to the adenylation pocket without having CCD6. UBA4 possesses a distinctive Rhodanese homology domain (RHD) at its C-terminus which is involved in sulfur transfer to URM1, the cognate UBL of UBA415. Each of these E1s activate specific UBL(s) and transfer them to one or more of their cognate E2s, establishing distinct UBL→E1→E2→E3→substrate cascades to influence a broad range of cellular functions. Of these, the UBA1-initiated cascade is the most branched with tens of E2s, hundreds of E3 and thousands of substrates10 (Table 1.1).

Figure 1.1 Ubiquitin and UBL conjugation system. Ubiquitin and UBL conjugation system comprises 3 enzyme classes that act sequentially to catalyze UBL conjugation: UBL activating enzymes (E1), UBL conjugating enzymes (E2), and UBL E3 ligases. Protein substrates are conjugated with different forms of ubiquitin and UBLs. Proteins conjugated with Lys48(K48)- linked polyubiquitin chains are recognized by the proteasome, the major proteolytic machinery in the cell that degrades such proteins into smaller peptides. Other forms of ubiquitin and UBL conjugation are involved in different pathways of cellular signaling. Ubiquitin and UBL conjugation is reversed by other classes of enzymes including deubiquitinating enzymes (DUBs) that deconjugate ubiquitin signals and UBL-specific proteases that deconjugate other UBL signals.

Figure 1.2 E1 enzymes and their structural domains. There are 8 E1 enzymes known so far including canonical (UBA1, UBA6, UBA7, SAE and NAE) and non-canonical (UBA4, UBA5 and ATG7) E1 enzymes. Of these, UBA1, UBA6 and UBA7 are monomeric, SAE and NAE are heterodimeric, and UBA4, UBA5 and ATG7 are homodimeric. AAD: active adenylation domain; IAD: inactive adenylation domain; CCD: catalytic cysteine domain; UFD; ubiquitin- fold domain; NES: non-canonical E1-specific domain; RHD: rhodanese homology domain.

1.2.2 Biochemical and structural mechanisms of E1-catalyzed UBL activation

UBL activation is a multi-step ATP-dependent enzymatic process whereby the UBL moiety is attached to the catalytic cysteine of E1 enzyme via a thioester bond for subsequent transfer to cognate E2s8. With their multiple functional domains, particularly the AD, E1s are catalytically competent to perform such activation as opposed to the simpler E2 enzymes (reviewed in Ref. 16). The catalytic steps of UBL activation have been well characterized with canonical E1s particularly the archetypal UBA1 enzyme. There exist, however, some variations in UBL activation mechanisms by non-canonical E1s (reviewed in Ref. 6, 8 ). In general, the catalytic cascade of UBL activation and transfer by canonical E1s includes four major steps: 1) first adenylation; 2) thioester formation; 3) second adenylation; and 4) UBL transfer to E2 by transthiolation. The cascade starts with C-terminal adenylation of a free UBL molecule in the presence of ATP and Mg2+ followed by binding of the adenylated UBL to the AAD of E1, forming an E1-bound

UBL∼adenylate (UBL∼AMP) intermediate and releasing inorganic pyrophosphate (PPi). This is followed by a nucleophilic inter-domain reaction whereby the catalytic cysteine of the CCD attacks the adenylated UBL forming a UBL-bound E1 via a covalent thioester bond, coupled with the release of free AMP. Subsequently, E1 catalyzes another round of adenylation of a second UBL molecule releasing PPi and forming a double-loaded E1 with two UBL molecules at two sites: one bound to the CCD via a covalent thioester bond and another bound to the AAD via a non-covalent bond17. This form is likely more competent from energetic and conformational perspectives for subsequent transfer of UBL6, 8. The double-loaded E1 then interacts with the cognate E2 enzyme through the UFD to transfer the thiol-bound UBL molecule to the corresponding thiol in the CCD of E2 via a transthiolation reaction, leaving an E1-bound UBL~AMP intermediate which is used in other rounds of UBL activation8 (Fig. 1.3). The catalytic activity of E1s is enabled by several conformational changes that facilitate different steps of UBL activation. Of a particular importance is the thioester bond formation enabled by rotation of the catalytic cysteine to come in close proximity to the UBL~AMP and attack the UBL moiety18. This rotation is associated with other conformational changes that lead to remodeling of the AAD of E1 and switching from an open to a closed conformation19. Such conformational changes are transient with the E1 assuming the open conformation again after forming the E1∼UBL thioester. These structural insights were partly derived by using E1

7 inhibitors that serve as UBL~AMP mimetics such as SUMO-AMSN and SUMO-AVSN18 (see below). While all steps of UBL activation are reversible, the progression of these reactions is maintained in one direction by several biochemical and structural factors including the abundance of ATP, active site remodeling to promote thioester bond formation and PPi release, and double loading with another UBL molecule to regain the open conformation of the AAD and drive subsequent UBL transfer to E2s6.

1.2.3 Druggability of E1 enzymes

As highlighted above, E1 enzymes catalyze the multi-step UBL activation exploiting their multi- domain structure, with at least two active sites that are amenable to therapeutic targeting (Fig. 1.4). The first active site is the nucleotide-binding pocket in the AAD to which the UBL~AMP binds. Historically, small-molecule ATP mimetics targeting mutant tyrosine kinases have been among the first classes of molecularly targeted agents (MTAs) to be developed for cancer therapy20, 21. Therefore, E1 enzymes with such ATP-binding pockets can serve as typical druggable targets for modulating different UBL conjugation pathways. Moreover, the unique catalytic mechanism involving the interaction of UBL~AMP rather than ATP to the nucleotide- binding pocket sets a clear distinction between E1s and other enzymes and offers an opportunity to develop inhibitors selectively targeting this class of enzymes with little impact on the kinome or other ATP-dependent enzymes22. The challenge lies, however, in defining selectivity among the eight E1 enzymes that utilize conserved mechanisms of UBL activation. In this respect, the involvement of UBLs at the nucleotide-binding pocket can be perceived as another source of selectivity even among related E1s given the structural variations among different UBLs. Therefore, structural information of these UBLs can be exploited to develop inhibitors specific to individual E1s. As an example, semisynthetic UBL~AMP analogs have been used as selective chemical probes to inhibit and interrogate structural biology of E1s18, 23. With more structural and biochemical information revealed on different E1s, small-molecule E1 inhibitors with more favorable drug-like properties have been developed. Standing out among these are the adenosine sulfamate E1 inhibitors that form a UBL~AMP-like intermediate in situ after permeation into the cells overcoming the drug delivery issue experienced with semisynthetic analogs24 (see below).

Table 1.1 E1 enzymes, their cognate E2 enzymes, UBLs and UBL proteases E1 Structure UBL E2s UBL Amino acid MW (kDa) proteases number UBA1 Monomer Ubiquitin Multiple DUBs 1058 117.85 (UBE1)

UBA6 Monomer Ubiquitin USE1 DUBs 1052 117.97 (UBE1L2) FAT10 (UBE2Z) FAT10 (unknown)

UBA7 Monomer ISG15 UBCH8 USP18 1012 111.69 (UBE1L) UBCH6

SAE Heterodimer SUMO 1/2/3 UBC9 SENPs SAE1: 346 SAE1: 38.45 (SAE1-UBA2) UBA2: 640 UBA2: 71.22

NAE Heterodimer NEDD8 UBC12 CSN5 NAE1: 534 NAE1: 60.25 (NAE1–UBA3) UBE2Fs DEN1 UBA3: 463 UBA3: 51.85

UBA4 Homodimer URM1 - - 460 49.67 (MOCS3)

UBA5 Homodimer UFM1 UFC1 UFSP1 404 44.86 UFSP2 ATG7 Homodimer ATG12 ATG10 - 703 77.96 ATG8 family ATG3

Figure 1.3 Cascade of E1-catalyzed UBL activation and substrate-assisted inhibition by adenosine sulfamates. E1-catalyzed UBL activation is a multi-step catalytic process that involves UBL adenylation in an ATP-dependent manner forming a UBL~AMP-E1 intermediate and releasing inorganic pyrophosphate (PPi). The UBL at the adenylation domain is then attacked by the sulfhydryl (-SH) group to form a UBL~E1 thioester intermediate associated with the release of adenosine monophosphate (AMP). The adenylation step is then repeated with another ATP molecule to form E1 doubly-loaded with 2 UBL molecules at 2 distinct sites. The UBL at the catalytic cysteine is then attacked by the –SH group of the cognate E2 enzyme to transfer the UBL in a transthiolation reaction. Adenosine sulfamates (ADS) inhibit E1 enzymes by attacking the E1-S~UBL intermediate (highlighted in a green dashed box) and forming covalent adduct with the UBL that binds to the nucleotide-binding site of E1 preventing its utilization in subsequent reactions.

9 10

The second active site is the catalytic cysteine in the CCD to which the UBL binds via a thioester bond. This catalytic residue with its redox-sensitive thiol group offers another great opportunity for developing thiol-reactive electrophilic inhibitors that covalently modify the active site25. Such cysteine-directed covalent agents are expected to exert irreversible and potent inhibition which is desirable in several contexts including cancer therapy. While many of these covalent inhibitors may have off-target effects due to their promiscuous reactivity, a number of FDA-approved agents including the kinase inhibitors rociletinib and osimertinib target thiol-containing residues with a high level of selectivity25. For E1 enzymes, several drugs that target the CCD have been reported, most prominently the nitro-pyrazone-based UBA1 inhibitors PYR-41 and PYZD-4409 (Ref. 26, 27). Interestingly, COH000 is a recently discovered SAE inhibitor that targets a cysteine residue in the AAD without affecting the catalytic cysteine28, 29. The third, and yet more challenging, site to target in E1 enzymes is the UFD through which they interact with their cognate E2s for subsequent UBL transfer. The difficulty of targeting E1-E2 interaction, like other protein-protein interactions (PPIs), lies in the large surface area implicated in such an interaction and the lack of pockets to which small-molecule inhibitors can bind30. Added to these difficulties, many inhibitors targeting PPIs are peptide-based with less favorable drug-like properties31. An example of these is UBC12N26, a 26-amino acid peptide that inhibits the interaction between NAE and its E2 enzyme, UBC12. UBC12N26 corresponds to the N- terminus of UBC12 and has been used to gain structural insights into NAE-UBC12 interaction32.

1.2.4 Rationale for targeting E1 enzymes in cancer

Cancer is a genetic disease whose initiation and progression is promoted by tumor-specific oncogenic alterations in cellular proteins (oncogene addiction)33. However, the cellular stresses created as a result of malignant transformation such as proteotoxic, replicative, oxidative and metabolic stresses, need to be supported by other broadly-acting cellular machineries that are essential for both normal and cancer cells34. Cancer cells are much more dependent on these machineries, and thus are more vulnerable to their inhibition compared to normal cells (non- oncogene addiction). While MTAs directed mostly against tumor-specific oncogenes are relatively safer compared to cytotoxic agents, emergence of resistance is a common pitfall that compromises the efficacy of such agents35. Therefore, it has become increasingly important to explore targeting non-oncogene addictions of cancer with agents that are potentially more effective than MTAs and less toxic compared to cytotoxic therapy (reviewed in Ref. 35 and 34 ).

Figure 1.4 Inhibitors of E1enzymes and their sites of action. A diagram of the E1 enzyme showing different E1 inhibitors and the structural domains or active sites they are reported to target. AAD: active adenylation domain; CCD: catalytic cysteine domain; UFD: ubiquitin-fold domain.

Proteasome inhibitors, currently in clinical use for the treatment of multiple myeloma and mantle cell lymphoma, constitute a classical example of such therapeutics. They target the proteasome which is an essential cellular machinery with broad cellular functions in both normal and cancer cells including a key role in protecting against proteotoxic stress particularly in myelomas that are engaged in immunoglobulin production36. Despite this essentiality and the expected toxicity of proteasome inhibition, bortezomib—the prototype of this class— was found to have a clinically acceptable safety profile with adverse effects that can be managed by dose optimization37. Similarly, E1s are cellular enzymes that are not frequently affected by oncogenic mutations and are essential for a broad spectrum of cellular functions in both normal and tumor cells8. Of all E1s, UBA1 is the most essential enzyme whose loss-of-function is anticipated to be most deleterious to the survival and growth of cancer cells10, 38. As assessed by the analysis of publicly available datasets of cancer cell line dependencies, only UBA1 and UBA2 (encoding the active subunit of SAE) genes are considered essential upon knockdown in large pan-cancer RNAi screens. However, the analysis of CRISPR/Cas9 knockout screens extends essentiality to UBA3 (encoding the active subunit of NAE), UBA4 and UBA5 as well as SAE1 and NAE1 encoding the non-active subunits of SAE and NAE, respectively (Fig. 1.5-1.7). E1 enzymes lie at the apex of UBL conjugation cascade and their targeting is expected to disrupt biological processes regulated by UBL conjugation8. For example, UBA1 activates ubiquitylation required for both degradative and non-degradative cellular functions39. Compared to proteasome inhibitors, UBA1 inhibitors are anticipated to be less selective, yet more efficacious, as they will disrupt proteasomal degradation as well as ubiquitin-regulated signaling pathways such as DNA repair and nuclear factor-kappa B (NF-κB) signaling1. Such higher efficacy may be needed in malignancies where proteasome inhibitors suffer from intrinsic or acquired resistance37. Targeting SAE is similarly expected to induce broad, yet less profound, effects as UBA1 inhibitors given the indispensable role of SUMOylation in many signaling pathways40, 41.

R N A i C R IS P R

U B A 1 U B A 2 U B A 1 U B A 2

9 6.8% 4 8.4% 100% 9 9.8%

U B A 3 U B A 4 U B A 3 U B A 4

1 2.3% 0 % 9 8.7% 9 9.1%

U B A 5 U B A 6 U B A 5 U B A 6

0.8% 2.4% 7 4.6% 1 0.9%

U B A 7 A T G 7 U B A 7 A T G 7

0 % 0.4% 0 % 0 %

E s s e n tia l N o n -e s s e n tia l

Figure 1.5 Cancer dependency of E1 genes. Essentiality of E1 enzymes as assessed by pan- cancer genome-wide loss-of-function screens conducted in 710 (RNAi) and 582 (CRISPR) cancer cell lines. Donut charts display the percentage of cell lines dependent on the indicated E1 gene. A cell line is regarded as dependent when it has a probability of dependency greater than 0.5. These data were obtained by the analysis of publicly available datasets on depmap portal at https://depmap.org/portal.

d e p m a p P ro je c t s c o re

1 2

r g

o 0

c -

R -1

E l

T -2

g M

-2 o

E L D -4

-3

1 2 3 4 5 6 7 7 1 2 3 4 5 6 A A A A A A A G A A A A A A B B B B B B B T B B B B B B U U U U U U U A U U U U U U

1 4

e r

o 2 c

e 0

s c

s 0

-1 n

E f

-2

C -2 s

s -4

o L

-3 -6

2 3 4 5 6 7 7 1 2 3 4 5 6 A A A A A A G A A A A A A B B B B B B T B B B B B B U U U U U U A U U U U U U

Figure 1.6 Essentiality of E1 genes as determined from pan-cancer genome-wide loss-of- function screens. Essentiality metrics of E1 genes were obtained from depmap (RNAi and CRISPR/Cas9 screens) and Project score (CRISPR/Cas9 screens) databases. DEMETER2 and CERES are dependency scores used on depmap portal for RNAi and CRISPR screens, respectively. Both scores are based on data from cell depletion assays with lower values indicating higher likelihood of gene essentiality in a given cell line. A score of 0 corresponds to non-essentiality and -1 corresponds to the median of all pan-cancer essential genes. The gene fitness metrics used on project score are fold-change and loss-of-fitness score. Fold-change is the copy number-corrected depletion fold-change calculated from average representation of targeting sgRNAs 14 days post-transfection compared to basal representation in the plasmid library. Loss- of-fitness score is a quantitative measure of cell depletion and significantly depleted genes (at FDR < 5%) are assigned a score < 0. Data points represent cell lines with horizontal blue dashes corresponding to the median score of all cell lines. Dashed lines are drawn at the dependency cutoffs set by these databases. Data were obtained from depmap (accessed at https://depmap.org/portal) and Project Score (accessed at https://score.depmap.sanger.ac.uk) databases.

U B A 1 U B A 1

2 4

r o

e 2

c g

0 s

s s

h 0

i l

-2 f

o -2

o s L -4

-4 o L

-6

e t s h k e g y s S n h e t s h k e g y s S n h n s S m u p c n n r a i c n s S m u p c n n r a i c a N u e i a e N k a a N u e i a e N k a o e i g m t u v r P S o e i g m t u v r P S B r C tr a N s L B r C tr a N s L h y e O c m h y e O c m B e L & t n o B e L & t n o p n a t p n a t m o & d i S m o & d i S o s P o s P d e a e d e a e E e g E e g n m H r n m H r E e a E e a H L H L U B A 2 U B A 2

2 4

r o

e 2

c g

0 s

s s

h 0

i l

-2 f

o -2

o s L -4

-4 o L

-6

2 4

r o

e 2

c g

0 s

s h

s 0

l i

-2 f o

-2

o s L -4

-4 o L

-6

2 4

r o

e 2

c g

0 s

s s

h 0

i l

-2 f

o -2

o s L -4

-4 o L

-6

2 4

r o

e 2

c g

0 s

s h

s 0

l i

-2 f o

-2

o s L -4

-4 o L

-6

e t s k e y s e t y s S m h g r S in h s S m s h k e g r s S in h n a u p c n n a N k c n u p c n n a N c o N iu g e ti u a e a o a N iu e i a e k a e r m v r P S e r g m t u v r P S B r C t a y N s L c m B r C t a y N s L c m B e h e O B e h e O L & t n to L & t n o m p n a p n a t o & d i S m o & d i S o s P o s P d e a e d e a e n E e g E e g m H r n m H r E e a E e a H L H L U B A 6 U B A 6

2 4

r o

e 2

c g

0 s

s s

h 0

i l

-2 f

o -2

o s L -4

-4 o L

-6

Figure 1.7 Essentiality of E1 genes in cancer cell lines of different origin. CRISPR/Cas9 screen data from the Project Score database were analyzed for E1 dependency in cell lines of 14 different tissues. Data points represent cell lines with horizontal blue dashes corresponding to the median score of all cell lines.

Since neddylation is required to regulate a subset of ubiquitin ligases (see below), targeting NAE is expected to be much more selective, which is consistent with the development of pevonedistat as the prototype of mechanism-based E1 inhibitors42. In the following sections, we will discuss in more detail the different E1s, their role in cancer, and the inhibitors developed so far to target these enzymes. In our discussion, we will emphasize mechanism-based E1 inhibitors particularly those advanced to clinical trials and the common principles shared among these agents. We will start with NAE against which the prototypical clinical E1 inhibitor was developed.

1.3 NEDD8-activating enzyme (NAE)

Neddylation is the process of conjugating NEDD8 to cellular proteins. NEDD8 is 59% identical to ubiquitin—the highest level of similarity observed among all UBLs, albeit with structural differences that are sufficient to mediate distinct functions43, 44. As with ubiquitylation, neddylation cascade is initiated by NAE that catalyzes NEDD8 activation. NAE is a heterodimeric enzyme composed of a regulatory subunit, NAE1 (APPBP1), and a catalytic subunit, NAE2 (UBA3), with similar domains to UBA1 (Fig. 1.2)8. While the AAD is located on NAE2, the CCD half-domains are located on both subunits6, 8. Interestingly, the ubiquitin- specific UBA1 enzyme can activate NEDD8 under stress conditions such as heat shock and oxidative stress45. Two neddylation E2 enzymes have been reported in metazoans, UBC12 and UBE2F, which function in concert with E3 ligases including ubiquitin ligases43. Neddylation targets many substrates, of which cullins are the best characterized and the most established class of neddylated proteins43. Cullins serve as scaffold proteins upon which the largest class of RING ubiquitin E3 ligases known as cullin-RING ubiquitin ligases (CRLs) are assembled44. Cullin neddylation plays a pivotal role in ubiquitylation of a subset of cellular proteins by inducing conformational changes that increase the activity of CRLs and enhance ubiquitin conjugation46. Other non-cullin substrates include signaling molecules such as p53, p73, E2F1, epidermal growth factor receptor (EGFR), tumor-growth factor beta receptor (TGFβR), NF-κB essential modulator (NEMO), Von Hippel–Lindau (VHL) tumor suppressor and Hypoxia-inducible factors (HIFs) which play diverse roles in normal and cancer cell biology43, 44. Therefore, neddylation regulates numerous pathways including proteasomal degradation, cell cycle

17 progression, receptor tyrosine kinase signalling, apoptosis, DNA damage response, inflammatory/immune responses, oxidative stress, hypoxia, and nucleolar stress signalling43, 47-51.

1.3.1 Role of NAE in cancer

As highlighted above, neddylation plays a salient role in numerous signaling pathways; therefore, dysregulation of NAE and/or downstream components of neddylation cascade is anticipated to contribute to the development and progression of several malignancies44, 52. In most cases, expression data from published studies or publicly available datasets indicate a negative correlation between expression of either NAE1 or NAE2 or both and patient outcomes. For instance, NAE1/2 are upregulated in hepatocellular carcinoma and are associated with an aggressive phenotype and poor overall and relapse-free survival in patients53-55. Similar data has been reported in lung cancer56, glioblastoma57, squamous cell carcinoma56, intrahepatic cholangiocarcinoma58, multiple myeloma59, esophageal squamous cell carcinoma60, uveal melanoma61, pancreatic cancer62, 63, clear cell renal cell carcinoma64, and chronic myeloid leukemia65. In contrast, thyroid cancer patients show a favorable prognosis with high NAE1 expression54. It appears, however, that no significant mutations are known to implicate NAE1/2 in cancer pathogenesis66.

1.3.2 NAE inhibitors

NAE was one of the first human E1 enzymes to have its crystal structure elucidated and this provided valuable structural insights and a common framework for mechanisms of activation and transfer of NEDD8 as well as other UBLs32, 67-69. These structural information contributed, in part, to guiding drug discovery endeavors for E1 inhibitors and it was not surprising that the first reported mechanism-based E1 inhibitor was targeted against NAE42, 70. The hyperactive neddylation pathway or high reliance on one or more of its components in cancer is a therapeutic avenue that has garnered an immense attention in recent years particularly after the development of selective NAE inhibitors71. Compounds targeting NAE comprise the adenosine sulfamate pevonedistat, the related adenosine aminosulfonamide TAS4464—both are selective clinical candidates, and several experimental inhibitors including NEDD8 adenylates and NEDD8- mimicking peptides (APN, pND20 and pND22), natural and semisynthetic NAE inhibitors (biapigenin, flavokawain A and piperacillin), as well as other inhibitors with diverse chemical structures (M22, LZ3, ZM223, and rhodium (III)-based complexes). So far, pevonedistat is the

18 most extensively studied E1 inhibitor with a large number of preclinical and clinical studies in numerous malignancies. Apart from pevonedistat and TAS4464, most known NAE inhibitors have been discovered by in silico approaches and possess a less favorable potency and selectivity profile with activity in the micromolar range. Nonetheless, they can potentially serve as useful chemotypes to boost drug discovery endeavors and develop new NAE inhibitors (Table 1.2). Pevonedistat Pevonedistat (MLN4924; TAK-924) is the prototypical adenosine sulfamate E1 inhibitor42. A high-throughput screen (HTS) identified N6-benzyl adenosine as an NAE inhibitor, and further optimization by iterative medicinal chemistry led to the development of pevonedistat, a first-in- class mechanism-based NAE inhibitor24, 42. Pevonedistat is structurally related to AMP and binds to the nucleotide binding site on NAE70. While compounds of this structural class may have inhibitory effects on adenylate-forming enzymes such as tRNA synthetases72-74, or potentially other ATP-dependent enzymes such as kinases, the action of pevonedistat has been shown to be highly selective for NAE even compared to other E1 enzymes42. In this context, it selectively inhibits neddylation with little or no effect on conjugation of related UBLs with half-maximal inhibitory concentration (IC50) values in the nanomolar range (Table 1.2). Given the role of neddylation in regulating the activity of CRLs, pevonedistat induces a potent inhibition of these enzymes with a subsequent reduction in the turnover of a subset of cellular proteins42, 47. Prominent among these proteins is the DNA replication licensing factor CDT1 which is essential for pre-replication complex assembly and plays a key role in DNA replication and mitosis75. Stabilization of CDT1 in the S phase of the cell cycle has been shown to be primarily responsible for the re-replication phenotype and mitotic defects observed after pevonedistat treatment42, 75, 76. DNA re-replication is characterized by ≥4N DNA content as a result of unscheduled repeated rounds of DNA synthesis and leads to replication stress, DNA damage and apoptosis77. Consistent with these cell cycle-specific effects, actively proliferating cells are more susceptible to pevonedistat cytotoxicity42. TAK-243, the related UBA1 inhibitor, does not induce re-replication possibly due to the stabilization of both CDT1 and its endogenous inhibitor geminin78 (see below).

Table 1.2 NAE inhibitors, their chemical structures and pharmacological properties Inhibitor Structure Type Malignancies Class; IC50 Mechanism Pevonedistat Clinical: 33 Many solid Adenosine EC50 (cell-free (MLN4924; clinical trials up and sulphamate; thioester TAK-924) to phase 3 hematologic substrate- formation) = malignancies assisted 0.0047 μM inhibition EC50 (cell-based neddylation) < 0.1 μM IC50 (cytotoxicity in cell lines) = 0.05 – 1.03 μM In vivo dose = 30-90 mg/kg QD or BID TAS4464 Clinical: phase Many solid Adenosine EC50 (cell-free (HY-128586) 1/2 clinical trial and amino- thioester (multiple hematologic sulfonamide formation) = myeloma and malignancies 0.955 nM non-Hodgkin IC50 (cytotoxicity lymphoma) in cell lines) ≤ 10 nM In vivo dose = 6.3

– 100 mg/kg QW

19 20

6,6’’- Experimental Cell-free and Semisynthetic EC50 (cell-free Biapigenin cell-based flavonoid UBC12-NEDD8 activity in thioester Caco-2 cells79 formation) =20 μM EC50 (cell-based UBC12-NEDD8 thioester formation) =5 μM Flavokawain Experimental Prostate, Chalcone EC50 (cell-free A bladder and flavonoid UBC12-NEDD8 urothelial thioester carcinoma80-82 formation) =5 μM

M22 Experimental Several cancer Piperidin-4- IC50 (viability cell lines; amine assays) = 8.98- mouse gastric 26.6 μM carcinoma In vivo dose = 60 model83 mg/kg IP QD Piperacillin Experimental Cell-free and Semisynthetic EC50 (cell-free cell-based β-lactam UBC12-NEDD8 activity in antibiotic thioester Caco-2 cells84 formation) =1 μM

Rhodium (III)- Experimental Cell-free and Cyclo- EC50 (cell-free based complex cell-based metallated UBC12-NEDD8 activity in rhodium(III) thioester Caco-2 cells85 complex formation) =1.5 μM IC50 (viability in Caco-2 cells) = 0.3 μM

LZ3 Experimental Cell-free and Sulfonamide EC50 (cell-free cell-based derivative UBC12-NEDD8 activity in thioester Caco-2, MCF7 formation) =1.06

and Bel-7402 μM 86 cells IC50 (viability) = 12.3-29.5 μM ZM223 Experimental Cell-free and Benzothiazole IC50 (viability) = cell-based derivative 0.1-1.22 μM activity in colon cancer and osteosarcoma cells87 Dipeptide- Experimental Cell-free and Alkaloid EC50 (cell-free conjugated cell-based derivative UBC12-NEDD8 Deoxy- activity in thioester vasicinone Caco-2 cells88 formation) =0.8 μM EC50 (cell-based UBC12-NEDD8 thioester formation) =0.8

μM

Adenosyl- Experimental Cell-free NEDD8 IC50 (NAE- phospho- activity70 adenylate UBC12 NEDD8 transthiolation) = (APN) 0.05 μM pND20 Experimental Cell-free Heptameric EC50 (NAE- (VI LTFGG) activity89 peptides NEDD8 thioester formation) = 133 μM pND22 Experimental Cell-free Heptameric EC50 (NAE- (VRLMFGG) activity89 peptides NEDD8 thioester formation) = 215 μM

pN1 Experimental Cell-free Heptameric NA (VWLSYGG) activity90 peptides

pN7 Experimental Cell-free Heptameric NA (VMLFYGG) activity90 peptides

23 pN26 Experimental Cell-free Heptameric NA (VLALRGG) activity90 peptides

NA: Not available

With the narrower spectrum of proteins influenced by disrupting neddylation, it is conceivable that NAE inhibitors will exhibit a wider therapeutic window compared to proteasome inhibitors91. Pevonedistat has been shown to be well-tolerated in vivo with a broad anti-tumor activity in solid and hematologic malignancies (Table 1.3). Apart from induction of re-replication and DNA damage64, 92-95, a plethora of diverse mechanisms have been reported to mediate the anti-tumor effects of pevonedistat. These include inhibition of the NF-κB pathway92, 96-99, stabilization and upregulation of pro-apoptotic proteins such as NOXA62, 100-102, generation of ROS92, stabilization of cell cycle regulators such as p21, p27 and WEE159, 65, 103-106, inhibition of the mTOR pathway99, 107, induction of ER stress108, 109, and activation of death receptor signaling60, 110. In addition to apoptosis, pevonedistat induces senescence in several malignancies111-113. Moreover, it exerts anti-angiogenic and anti-metastatic effects in preclinical models of pancreatic cancer114 and uveal melanoma61. While inhibiting migration in human urothelial108, clear cell renal64 and gastric cancer cells115, pevonedistat has been reported to display pro-migratory effects in glioblastoma and prostate cancer cells through induction of caveolin-1 phosphorylation116, suggesting control of migration by neddylation is context- dependent. With the multiple effects exerted by pevonedistat, it is not surprising that additive or synergistic anti-tumor effects are observed with mechanistically diverse antineoplastic agents including conventional cytotoxic drugs56, 117-123, radiotherapy105, 124-126, differentiation therapy127, and targeted therapies such as Bcl-2 inhibitors101, epigenetic modulators128, 129, monoclonal antibodies120, immunomodulatory drugs130, 131, PARP inhibitors94, BRAF inhibitors113, and other investigational therapies132-135. In certain malignancies, the cytotoxicity of pevonedistat is ameliorated by induction of cytoprotective autophagy136—an effect that can be overcome by combination with autophagy inhibitors leading to synergistic anti-tumor effects134, 137. Since its discovery in 2009, pevonedistat has been and is currently evaluated in different phases of over 30 clinical trials alone and in combination with other drugs (Table 1.4). Approximately half of these are in acute myeloid leukemia (AML) and/or MDS.

Table 1.3 Preclinical and clinical studies of pevonedistat Cancer Findings Reference Lymphoma Preclinical efficacy in DLBCL via NF-κB pathway inhibition and DNA re- Milhollen et al., 201096 replication Anti-tumor effects in MCL cells via stabilization of NOXA Dengler et al., 2014100 Anti-tumor effects in lymphoma cell lines associated with apoptosis and Wang et al., 2015111 senescence Preclinical efficacy in MCL models and additive/synergistic antitumor effects with Czuczman et al., conventional therapies 2016120 Phase 1 clinical trial showing tolerability and pharmacodynamic effects in Shah et al., 2015138 relapsed/refractory lymphoma Sensitization of DLBCL cells to the death receptor agonists TRAIL and FasL by Paiva et al., 2017110 enhancing death receptor signaling Acute myeloid Preclinical efficacy via NF-κB pathway inhibition, DNA damage, and ROS Swords et al., 201092 leukemia generation Sensitization of AML cell lines to retinoic acid differentiation therapy Tan et al., 2011127 Preclinical efficacy in FLT3-ITD AML models via NF-κB inhibition and Khalife et al., 201597 downregulation of miR-155 Anti-tumor effects in AML cells via Noxa upregulation as well as synergistic Knorr et al., 2015101 effects with Bcl-2 inhibitors Synergistic anti-leukemic effects with cytarabine in AML cells and mouse models Nawrocki et al., 2015139 via disruption of nucleotide metabolism Phase 1 clinical trial showing tolerability and modest clinical activity in patients Swords et al., 2015140 with AML and MDS Synergistic effects with the IAP antagonist T-3256336 in cell lines and in vivo Sumi et al., 2016132 Synergistic combination with azacytidine in preclinical models of AML via Visconte et al., 2016128 antagonizing RRM2 Synergistic combination with belinostat in preclinical models of AML/MDS by Zhou et al., 2016129 reciprocal inhibition of specific DNA repair pathways Synergistic combination with the LSD1 inhibitor T-3775440 in preclinical models Ishikawa et al., 2017133 of AML via trans-differentiation and DNA re-replication Expanded safety analysis of pevonedistat in AML/MDS patients with serious Swords et al., 2017141 toxicity (e.g., hepatotoxicity and sepsis syndromes with multi-organ failure) only

25 26

observed at doses beyond 100 mg/m2 Phase 1b clinical trial of pevonedistat in combination with azacytidine showing Swords et al., 2018142 tolerability and potential benefit Colorectal cancer Anti-tumor effects via DNA re-replication in CRC cell lines Lin et al., 201076 Preclinical efficacy in models of poorly differentiated, clinically aggressive CRC Picco et al., 2017143 with a transcriptional signature associated with pevonedistat sensitivity Sensitization of CRC cells to oxaliplatin via induction of DNA damage and CHK2 Zheng et al., 201793 phosphorylation Solid cancers Chemical/biochemical characterization and preclinical evaluation in models of Soucy et al., 200942 CRC and lung cancer Anti-tumor effects via induction of senescence in CRC, lung and GBM cell lines Jia et al., 2011112 Anti-tumor effects via DNA re-replication in CRC and breast cancer cell lines Milhollen et al., 201175 Synergistic effects with DNA inter-strand cross-linking agents in CRC and breast Kee et al., 2012117 cancer cell lines Inhibition of autophagy sensitizes solid cancer cell lines to pevonedistat Zhao et al., 2012137 Synergistic effects with mitomycin C, cisplatin, cytarabine, UV radiation, SN-38, Garcia et al., 2014118 and gemcitabine Selective anti-tumor effects in solid CRC and melanoma cell lines as well as Malhab et al., 2016144 zebrafish embryos in a p53-based cyclotherapy to selectively target p53-mutant cells Phase 1 clinical trial showing tolerability and pharmacodynamic effects of Sarantopoulos et al., pevonedistat 2016145 Phase 1b clinical trial showing tolerability in combination with either docetaxel or Lockhart et al., 2018146 carboplatin + paclitaxel with sustained clinical responses in the latter combination Head & neck cancer Anti-tumor effects in combination with TRAIL via promoting c-FLIP degradation Zhao et al., 2011147 in HNSCC cell lines Preclinical efficacy in HNSCC models and radio-sensitization of HNSCC cells Vanderdys et al., 2018125 Liver cancer Preclinical efficacy associated with induction of protective autophagy and Luo et al., 2012136 apoptosis in HCC models Preclinical efficacy in Phb-KO mouse models of HCC by destabilizing LKB1 and Barbier-Torres et al., Akt 201555 Synergistic effects with autophagy inhibitors in cell lines and mouse models of Chen et al., 2015134

HCC Myeloma Preclinical efficacy and additive effects with conventional drugs in MM models McMillin et al., 2012148 Anti-tumor effects in MM cells via upregulation of REDD1 and subsequent Gu et al., 2014107 inhibition of AKT and mTOR signaling Anti-tumor effects in CKS1B-overexpressing MM cell lines via p21 stabilization Huang et al., 201559 Phase 1 clinical trial showing tolerability and pharmacodynamic effects in Shah et al., 2015138 relapsed/refractory MM Synergistic cytotoxicity with pomalidomide using sequential treatment schedule in Liu et al., 2019130 MM cells Pediatric cancers Preclinical efficacy in a subset of pediatric tumors tested in Pediatric Preclinical Smith et al., 2012149 Testing Program (PPTP) Pancreatic cancer Radio-sensitization of preclinical models of pancreatic cancer Wei et al., 2012126 Anti-angiogenic and anti-metastatic effects in pancreatic cancer models Yao et al., 2014114 Sensitization of preclinical models of PDAC to gemcitabine via accumulation of Li et al., 201762 NOXA and ERBIN Potentiating the anti-tumor effects of pevonedistat by the Chk1 inhibitor SCH Li et al., 201863 900776 in preclinical models of pancreatic cancer Breast cancer Radio-sensitization of breast cancer cells via induction of p21 accumulation Yang et al., 2012103 Melanoma A genome-wide siRNA screen identified the p21-dependent intra-S-phase Blank et al., 2013104 checkpoint as a key mediator of pevonedistat cytotoxicity in melanoma cells independent of CDT1 stabilization Preclinical efficacy in melanoma models via DNA re-replication and senescence Benamar et al., 2016113 and synergistic effects with the BRAF kinase inhibitor vemurafenib Phase 1 clinical trial showing tolerability and durable stable disease in a subset of Bhatia et al., 2016150 patients Preclinical efficacy and anti-metastatic effects in UM models via anti-angiogenic Jin et al., 201761 and anti-CSC effects Preclinical efficacy in models of melanoma and gene expression differences Wong et al., 2017151 between sensitive and resistant cell lines Ovarian cancer Synergistic effects with platinum in ovarian cancer cells Jazaeri et al., 2013119 Synergistic effects with cisplatin in ovarian cancer cells and mouse xenograft Nawrocki et al., 2013121 models Anti-tumor effects alone and in combination with conventional therapies in EOC Pan et al., 2013152

cells Synergistic effects with the IAP antagonist T-3256336 in an ovarian cancer cell Sumi et al., 2016132 line Antagonistic effects with paclitaxel in ovarian cancer cells Hong et al., 2018153 Ewing sarcoma Preclinical efficacy in ES models via WEE1 accumulation and deregulation of S- Mackintosh et al., phase proteins 2013106 Cholangiocarcinoma Preclinical efficacy in ICC models and additive effects with cisplatin Gao et al., 201458 Chronic Anti-tumor effects in primary CLL B cells co-cultured with stromal cells via NF- Godbersen et al., 201498 lymphocytic κB pathway inhibition leukemia Anti-tumor effects in primary CLL B cells co-cultured with stromal cells via Paiva et al., 2015154 CDT1 accumulation and sensitization of cells to alkylating agents Sensitization of primary CLL cells co-cultured with stromal cells to death receptor Paiva et al., 2017110 agonists by enhancing death receptor signaling Lung cancer Anti-tumor effects in KrasG12D-driven lung cancer models via NF-κB and mTOR Li et al., 201499 pathways inhibition Preclinical efficacy and synergistic anti-tumor effects with platinum in lung Li et al., 201456 adenocarcinoma models Synergistic effects with the PARP inhibitor olaparib in NSCLC cell lines via Guo et al., 201794 inhibiting BRCA1 recruitment to DNA damage sites Anti-tumor effects in paclitaxel-resistant human lung adenocarcinoma cells with Xu et al., 2018122 no synergism with paclitaxel Urothelial Synergistic effects with cisplatin in urothelial cancer cell lines and mouse Ho et al., 2015123 carcinoma xenograft models Preclinical efficacy in UC models associated with ER stress induction, cell cycle Kuo et al., 2015108 arrest and apoptosis Glioblastoma Preclinical efficacy in models of GBM associated with induction of apoptosis and Hua et al., 201557 senescence Pevonedistat promotes migration of GBM cells via induction of caveolin-1 Park et al., 2018116 phosphorylation Upregulation of PD-L1 expression and synergism with anti-PD-L1 therapy in Zhou et al., 2019131 GBM models Cervical carcinoma Preclinical efficacy in models of CC as well as potentiation of cisplatin Lin et al., 2015155 cytotoxicity

Gastric cancer Anti-tumor effects in gastric cancer cells associated with protective p27 Zhang et al., 201595 upregulation Anti-tumor effects in gastric cancer cells associated with DNA damage induction, Lan et al., 2016115 senescence and autophagy Esophageal cancer Preclinical efficacy in ESCC models via ATF4–CHOP–DR5 axis-mediated Chen et al., 201660 extrinsic apoptosis Acute lymphoblastic Preclinical efficacy in T-ALL models via induction of nucleolar stress signaling Han et al., 2016156 leukemia Preclinical efficacy in ALL models via ER stress and synergistic effects with Leclerc et al., 2016109 dexamethasone, doxorubicin and cytarabine Inhibition of the MEK/ERK pathway sensitizes ALL to pevonedistat in vitro and Zheng et al., 2018157 in vivo Prostate cancer Anti-tumor effects in prostate cancer cells with apoptosis observed only at high Rulina et al., 2016158 concentrations Radio-sensitization of hormone-resistant prostate cancer cells via accumulation of Wang et al., 2016105 WEE1, p21 and p27 Pevonedistat promotes migration of prostate cancer cells via induction of caveolin- Park et al., 2018116 1 phosphorylation Osteosarcoma Preclinical efficacy in OS models associated with induction of senescence and Zhang et al., 2016159 apoptosis Genetic and chemical inhibitors of Mcl-1 sensitize OS cells to pevonedistat Zhang et al., 2017160 Mesothelioma Preclinical efficacy in NF2-mutant mesothelioma models and synergistic effects Cooper et al., 2017135 with mTOR inhibitors Renal cell Anti-tumor effects in CCRCC cell lines via induction of DNA damage and Tong et al., 201764 carcinoma suppression of EMT Anti-tumor effects in RCC cell lines via upregulation of NOXA Wang et al., 2017102 Preclinical efficacy in RCC models associated with induction of cell-cycle arrest, Xu et al., 2018161 senescence, and apoptosis Nasopharyngeal Preclinical efficacy in NPC models associated with stabilization of c-Jun and Xie et al., 2017124 carcinoma synergistic effects with radio- and chemotherapy Chronic myeloid Preclinical efficacy in CML models associated with p27 accumulation and Liu et al., 201865 leukemia overcoming mutation- and LSC-driven imatinib resistance AML: Acute myeloid leukemia; CC: Cervical carcinoma; CCRCC: Clear cell renal cell carcinoma; CLL: Chronic lymphocytic leukemia; CML: Chronic myeloid leukemia; CRC: colorectal cancer; CSC: Cancer stem cell; DLBCL: Diffuse large B-cell

30 lymphoma; EMT: Epithelial-to-mesenchymal transition; EOC: Epithelial ovarian cancer; ER: Endoplasmic reticulum; ES: Ewing sarcoma; ESCC: esophageal squamous cell carcinoma; GBM: glioblastoma; HCC: Hepatocellular carcinoma; HNSCC: Head and neck squamous cell carcinoma; ICC: Intrahepatic cholangiocarcinoma; KO: Knockout; LSD1: Lysine-specific demethylase 1A; MCL: Mantle cell lymphoma; MDS: Myelodysplastic syndromes; MM: Multiple myeloma; NF2: Neurofibromatosis type 2; NPC: Nasopharyngeal carcinoma; NSCLC: Non-small cell lung cancer; OS: Osteosarcoma; PARP: Poly (ADP-ribose) polymerase; PDAC: Pancreatic ductal adenocarcinoma; PD-L1: Programmed death-ligand 1; ROS: reactive oxygen species; RRM2: Ribonucleoside- diphosphate reductase; T-ALL: T-cell acute lymphoblastic leukemia; UC: Urothelial carcinoma; UM: Uveal melanoma

Phase 1 clinical trials in AML/MDS has shown pevonedistat is generally well-tolerated with pyrexia as the most common adverse event and some dose-limiting toxicities such as hepatotoxicity and sepsis syndromes with multi-organ failure that are mostly observed at doses beyond 50 mg/m2 (Ref. 140, 141). In AML, the rates of complete (CR) and partial responses (PR) ranged from 10-17%, with an overall response rate (ORR) of 13%140, 141. When evaluated in combination with azacitidine, the ORR increased up to 83% with a CR/PR rate of 80% in patients with TP53 mutations142. These encouraging results with azacitidine prompted several phase 2 clinical trials, and a phase 3 clinical trial, PANTHER, is currently ongoing to evaluate pevonedistat-azacitidine combination versus azacitidine a lone as a frontline therapy for specific forms of MDS and AML (ClinicalTrials.gov Identifier: NCT03268954). In other hematologic (relapsed/refractory multiple myeloma and lymphoma) and advanced solid malignancies, pevonedistat has also shown a similar tolerability with potential therapeutic benefit in these populations138, 145, 146, 150.

TAS4464 TAS4464 is an adenosine aminosulfonamide discovered by Taiho Pharmaceuticals. Generally, it has structural features similar to adenosine sulfamates, but differs in having an aminosulfonamide group in lieu of the sulfamate moiety162. Still, it needs to form a covalent adduct with NEDD8 through this functional group, and this adduct in turn serves as the potent inhibitory species of NAE162. TAS4464 retains the ribose sugar moiety found in AMP as is the case with several dual and pan E1 inhibitors (see below). Despite this similarity, it exhibits a highly selective activity toward NAE. Accordingly, it induces similar biological effects on neddylation and downstream signaling, yet with up to 64-fold higher potency and more durable effects compared to pevonedistat. TAS4464 displays a broad-spectrum activity on cancer cell lines of different origin, particularly those of hematologic malignancies. It also displays a superior activity on primary patient samples as well as PDXs, including treatment-resistant samples. TAS4464 has been advanced to a phase 1/2 clinical trial in multiple myeloma and non- Hodgkin lymphoma, but the trial was terminated (ClinicalTrials.gov Identifier: NCT02978235).

Table 1.4 Clinical trials of pevonedistat Identifier Tile Phase Disease Single/Combo Combined drug (s) Status Ref. NCT01814826 Study of MLN4924 Plus 1 AML Combo Azacitidine C 142 Azacitidine in Treatment- Naïve Patients with AML Who Are 60 Years or Older NCT03772925 Pevonedistat and Belinostat 1 AML/MDS Combo Belinostat N - in Treating Patients with Relapsed or Refractory AML or MDS NCT01011530 Dose Escalation Study of 1 Melanoma Single - C MLN4924 in Adults with 150 Melanoma NCT01862328 Dose Escalation, Multi-arm 1 Solid Tumors Combo • Docetaxel C 146 Study of MLN4924 Plus • Paclitaxel + Docetaxel, Gemcitabine, or Carboplatin Combination of Carboplatin • Gemcitabine and Paclitaxel in Patients with Solid Tumors NCT03330106 A Study to Evaluate the 1 Advanced Safety (QTc Docetaxel R - Effects of Pevonedistat on Solid Tumors interval) Paclitaxel + Carboplatin the QTc Interval in Participants with Advanced Solid Tumors NCT03770260 Ixazomib and Pevonedistat 1 Plasma Combo Ixazomib N - in Treating Patients with Cell Myeloma Multiple Myeloma That Has Come Back or Does Not Respond to Treatment NCT00677170 Study of MLN4924 in 1 Advanced Single - C 140, Adult Patients with Solid Tumors 145 Nonhematologic

32 33

Malignancies NCT02122770 Effects of Fluconazole and 1 Advanced Combo • Fluconazole C - Itraconazole CYP3A- Solid Tumors DDI study • Itraconazole Mediated Inhibition on the • Docetaxel Pharmacokinetics, Safety, • Paclitaxel + and Tolerability of Carboplatin MLN4924 in Participants with Advanced Solid Tumors NCT03814005 A Study of Pevonedistat in 1 MDS Combo Azacitidine N - Combination with CMML Azacitidine in Participants AML with Higher-risk MDS, CMML, or Relapsed/Refractory AML With Severe Renal impairment or Mild Hepatic Impairment NCT03323034 Pevonedistat, Irinotecan 1 Solid tumors; Combo Irinotecan R - Hydrochloride, and Lymphoma Temozolomide Temozolomide in Treating Patients with Recurrent or Refractory Solid Tumors or Lymphoma NCT03486314 A Study to Evaluate the 1 Advanced Combo; • Rifampin R - Effects of Rifampin on PK Solid Tumors DDI Study • Docetaxel of Pevonedistat in • Paclitaxel + Participants Carboplatin With Advanced Solid Tumors NCT03459859 Pevonedistat and Low Dose 1 AML; MDS Combo Cytarabine R - Cytarabine in Adult Patients With AML and MDS

NCT00722488 Study of MLN4924, a 1 Lymphoma; Single - C 138 Novel Inhibitor of Nedd8 MM Activating Enzyme, in Adult Patients With Lymphoma or MM NCT00911066 MLN4924 for the 1 AML; MDS; Single - C 140 Treatment of ALL AML, MDS, and ALL NCT03745352 Pevonedistat with 2 AML Combo Azacitidine N - Azacitidine Versus Azacitidine Alone in Treating Patients with Relapsed or Refractory AML NCT03057366 A Study of [14 C]- 1 Advanced Combo • Docetaxel C 146 Pevonedistat in Participants Solid Tumors (excretion • Paclitaxel + with Advanced Solid study) Carboplatin Tumors NCT03386214 Pevonedistat in 1 Myelofibrosis Combo • Ruxolitinib R - Combination With Ruxolitinib for Treatment of Patients with Myelofibrosis NCT03813147 Pevonedistat, Azacitidine, 1 AML; MDS Combo • Azacitidine R - Fludarabine Phosphate, and • Cytarabine Cytarabine in Treating • Fludarabine Patients with Relapsed or • Methotrexate Refractory AML or • Hydrocortisone Relapsed High-Risk MDS NCT01415765 MLN4924 Compared With 1/2 Diffuse LBCL Combo • Etoposide W - MLN4924 Plus (+EPOCH-R) • Prednisone Chemotherapy for LBCL • Vincristine • Cyclophosphamide

• Doxorubicin • Rituximab • Filgrastim NCT03479268 Pevonedistat and Ibrutinib 1 CLL; NHL Combo Ibrutinib R - in Treating Participants With Relapsed or Refractory CLL or NHL NCT03009240 Pevonedistat and Decitabine 1 AML Combo Decitabine R - in Treating Patients with High-Risk AML

NCT03330821 Pevonedistat, Cytarabine, 1/2 AML Combo Pevonedistat R - and Cytarabine Idarubicin in Treating Idarubicin Patients With AML NCT03709576 Pevonedistat and 2 AML Combo Azacitidine R - Azacitidine as Maintenance Therapy After Allogeneic Stem Cell Transplantation for Non- Remission AML NCT02782468 A Study of Pevonedistat in 1 AML/MDS Combo Azacitidine ANR - Adult East Asian Participants NCT03268954 Pevonedistat Plus 3 MDS; CMML; Combo Azacitidine R - (PANTHER) Azacitidine AML Versus Single-Agent Azacitidine as First-Line Treatment for Participants with Higher-Risk MDS, CMML, or Low-Blast AML

NCT02610777 An Efficacy and Safety 2 MDS; CMML; Combo Azacitidine ANR - Study of Pevonedistat Plus AML Azacitidine Versus Single- Agent Azacitidine in Participants with HR-MDS, CMML and Low-Blast AML NCT03319537 Pevonedistat with VXLD 1 ALL Combo Vincristine R - Chemotherapy for Dexamethasone Adolescent/Young Adults PEG-asparaginase with Relapsed/Refractory Doxorubicin ALL NCT03228186 Trial of Pevonedistat Plus 2 NSCLC Combo Docetaxel R - Docetaxel in Patients With Previously Treated Advanced NSCLC NCT03238248 Pevonedistat and 2 MDS; Combo Azacitidine R - Azacitidine MDS/MPN in MDS or MDS/MPN Patients Who Fail Primary Therapy With DNA Methyl Transferase Inhibitors NCT03013998 Study of Biomarker-Based 1/2 AML Combo Azacitidine R - Treatment of AML NCT03965689 Testing the Combination of 2 NSCLC Combo Carboplatin N - MLN4924 (Pevonedistat), Paclitaxel Carboplatin, and Paclitaxel for NSCLC NCT03862157 Azacitidine, Venetoclax, 1/2 AML Combo Azacitidine R - and Pevonedistat in Venetoclax Treating Patients With Newly diagnosed Acute

Myeloid Leukemia AML: Acute myeloid leukemia; ANR: Active, not recruiting; C: Completed; CLL: Chronic Lymphocytic Leukemia; CMML: Chronic Myelomonocytic Leukemia; LBCL: Large B-cell Lymphoma; N: Not yet recruiting; NHL: Non- Hodgkin Lymphoma; NSCLC: Non- small cell lung cancer; QTc: Corrected QT; R: Recruiting; W: Withdrawn

NEDD8 adenylates and NEDD8-mimicking peptides Adenosyl-phospho-NEDD8 (APN) has been synthesized as a chemical probe to assist in elucidating the mechanism of NAE inhibition by pevonedistat70. Similar to the ubiquitin- adenylate analogue APU (see below), APN acts as a non-hydrolyzable mimetic of NEDD8- AMP. It also resembles pevonedistat-NEDD8 adduct in tightly binding to NAE and inhibiting its activity in an ATP-competitive manner; however, NAE exhibits a slower recovery from pevonedistat-NEDD8 adduct which is consistent with a tighter binding compared to APN70. As the C-terminal sequence of NEDD8 plays a fundamental role in its recognition by NAE, phage display technology has been exploited to screen mutant versions of this sequence for variants with different kinetics and enzyme reactivities89. In this respect, pND20 and pND22 have been identified as NEDD8-mimicking heptameric peptides that can still be recognized and activated by NAE, yet with higher affinity to the enzyme compared to the wild-type NEDD8. Despite successful activation by NAE, these mutant variants cannot be subsequently utilized in the neddylation cascade, thereby serving as mechanism-based inhibitors that block NEDD8 activation and conjugation. NEDD8 is the closest UBL to ubiquitin in terms of sequence homology and UBA1 can activate NEDD8 under certain conditions45. Therefore, NEDD8- mimicking peptides have also been derived from mutant C-terminal sequences of ubiquitin resulting in similar inhibitory properties against NAE90.

Natural and semisynthetic NAE inhibitors 6,6’’-Biapigenin is a semisynthetic flavonoid derivative whose NAE inhibitory activity was identified through virtual screening of a 20,000-compound library using the crystal structure of the quaternary NAE1–UBA3–NEDD8–ATP complex79. Based on molecular modeling studies, 6,6’’-biapigenin is predicted to bind to NAE at a binding site distinct from pevonedistat resulting in reversible inhibition. Another flavonoid with anti-NAE activity is Flavokawain A80. Specifically, it belongs to the chalcone subclass of flavonoids and is isolated from the Kava extract81. Flavokawain A inhibits NAE after binding to the enzyme at the nucleotide-binding pocket, resulting in deneddylation of cullins and inducing proteasome-dependent degradation of Skp2. As a result, it shows a selective anti-tumor activity against Rb-deficient cells that are dependent on Skp2 for survival, as well as anti-tumor and anti-metastatic effects in vivo80. Piperacillin—an FDA-approved semisynthetic β-lactam antibiotic—is also reported to competitively inhibit NAE at the nucleotide-binding pocket suppressing neddylation of

38 39 downstream targets and inducing accumulation of p27kip1 in cells84. Another virtual screen of 90,000-compound library identified a dipeptide-conjugated alkaloid derivative with anti-NAE properties in cell-free and cell-based enzymatic assays88.

Other NAE inhibitors M22 is a piperidin-4-amine-based NAE inhibitor identified through a structure-based virtual screening of 50,000 compounds83. Molecular modeling studies predict M22 binds to the nucleotide-binding site in a similar conformation to that of ATP. As assessed by biochemical analyses, M22 reversibly inhibits NAE in an ATP-competitive manner, without affecting other E1 enzymes, resulting in stabilization of CRL substrates. These effects are associated with antiproliferative activity in cancer cells of different origin both in vitro and in vivo, as well as synergistic effects with bortezomib. Using zebrafish as an acute toxicity model, M22 shows minimal toxicity even with high micromolar concentrations83. To identify covalent NAE inhibitors, Zhang et al. leveraged structural information on NAE and pharmacologic data available on pevonedistat’s mode of action86. Specifically, they applied a ligand- and structure-based pharmacophore modeling approach combined with covalent docking and screened a focused library of free sulfamoyl-containing compounds for such irreversible inhibitors. These endeavors led to the identification of LZ3 as the top hit. LZ3 is a sulfonamide derivative that covalently binds to NAE inhibiting its activity in cell-free and cell-based assays. It is unclear, however, whether it forms a covalent adduct with NEDD8 to exhibit such irreversible inhibition. On the other hand, target-based virtual screening approaches have been performed using structural data of the non-covalent interaction between pevonedistat and NAE to identify reversible NAE inhibitors. This led to the identification of a series of reversible non-sulfamate- based inhibitors including ZM223, a benzothiazole with an activity in the low micromolar range87. A rhodium (III)-based complex is reported to possess a similar inhibitory activity against NAE in cell-free assays and Caco-2 cells, with a low micromolar potency partly attributed to its octahedral coordination geometry85.

1.4 Ubiquitin-like modifier-activating enzyme 1 (UBA1; UBE1)

UBA1 is the major ubiquitin-activating enzyme that catalyzes the first step in ubiquitin conjugation, and is involved in the ubiquitylation of more than 99% of cellular proteins78. UBA1

40 is an evolutionarily conserved protein and plays an indispensable role in regulating numerous cellular functions38. Structurally, it is a canonical monomeric E1 enzyme with several domains including the AAD, IAD, UFD, and CCD8, 163. Among the eight E1 enzymes identified so far, only UBA1 and UBA6 can activate ubiquitin, but the functions of UBA6-mediated ubiquitylation are not well-characterized compared to those mediated by UBA1164-166. In addition, several differences exist between UBA1 and UBA6 (Table 1.5). Ubiquitin itself can be modified on one or more of its 7 lysine residues, its N-terminus and/or 11 phosphorylation sites with ubiquitin or other PTMs creating highly diverse ubiquitin signals with multiple topologies (reviewed in Ref. 167). These distinct forms of ubiquitylation can be simply divided into mono- and poly-ubiquitylation, and this latter comprises K48- and K63-linked polyubiquitin chains as the most common forms of ubiquitylation167. The diverse ubiquitin-based modifications constitute a ‘ubiquitin code’ that is written by the concerted action of ubiquitin E1- E2-E3 enzymes, erased by a set of DUBs, and read by cellular proteins that specifically recognize this code to translate it to biological response167. Ubiquitylation is important for protein homeostasis through proteasomal degradation and autophagy, and is implicated as a PTM in regulating and fine-tuning a myriad of non-degradative cellular processes such as cell cycle progression, DNA replication and repair, protein translation, cell death, endocytosis, chromatin architecture and epigenetic regulation, cellular differentiation, inflammation and immune signaling2, 39, 167-170. Given its role in almost every aspect of cell biology, ubiquitylation is tightly regulated by multiple mechanisms including other PTMs and DUBs to specifically conjugate target proteins in a spatio-temporal manner. The ubiquitin system is ‘pyramidal’ in design as two E1 enzymes transfer ubiquitin to approximately over 30 E2s which in turn interact with hundreds of E3 ligases to conjugate the respective protein substrate7. The specificity of ubiquitin conjugation is mainly controlled downstream at the level of ubiquitin E3 ligases. Therefore, pharmacologic manipulation of the upstream enzyme UBA1 is anticipated to indiscriminately influence numerous biological processes.

Table 1.5 Properties of ubiquitin-activating enzymes: UBA1 and UBA6 (Ref. 10, 78, 166, 171, 172) Enzyme UBA1 UBA6 Identity* 100 % 40 % Molecular weight 118 kDa 118 kDa UBL Ubiquitin Ubiquitin; FAT10# Contribution to cellular > 99% < 1% ubiquitylation E2 enzymes** All ubiquitin E2 except USE1 Ubiquitin: Few E2s and USE1 is UBA6-specific FAT10: Unknown Preferential activity toward ubiquitin under basal conditions; cytokines (e.g IFNγ) increase activity toward FAT10 Knockout phenotype Lethality in C. elegans and Embryonic lethality in miceφ yeastφ Expression£ Constitutively expressed in Expressed in low amounts relatively large amounts in all (compared to UBA1) in all tissues tissues, with specifically higher expression in the testis

Ubiquitin charging in Fully charged with ubiquitin 50% charged proliferating cells *Percent amino acid identity relative to UBA1 **Nine E2s were reported to be charged equally by UBE1 and UBA6 #FAT10 was reported to serve as a PTM to direct proteins to proteasomal degradation φ Embryonic lethality of knockout phenotypes may indicate that UBA1 and UBA6 are not merely redundant versions of each other and that each of them is required for distinct essential biological functions. £ UBA1 is ranked among the top 2% of abundant proteins in HeLa cells (>3x106 copies per cell), and the relative UBA1:UBA6 abundance ratio is approximately 10:1

1.4.1 Role of UBA1 in cancer

Although many mutations and copy number variations (CNVs) were reported to affect UBA1, there is no reported evidence that these genetic aberrations are functionally important in the pathogenesis or progression of cancer66, 173. Physiologically, UBA1 is among the most abundant proteins in the cell; however, variations in expression levels may occur in several cancers10. Analysis of patient survival data demonstrated that UBA1 expression may correlate with the prognosis of cancer. While high expression of UBA1 is associated with poor prognosis in liver cancer, it is associated with a favorable prognosis in prostate cancer54. On the other hand, AML cell lines and primary samples exhibited equal levels of UBA1 protein compared to normal hematopoietic cells, despite having higher abundance of ubiquitylated proteins27. This increased ubiquitylation is likely attributable to the more active utilization of UBA1 in AML versus normal cells constituting a potential therapeutic vulnerability in AML and other malignancies26, 78. Ubiquitin conjugation is a multi-enzymatic process and cancer-associated dysregulation of ubiquitylation may occur at the level of downstream E2 or E3 enzymes with no changes in the ubiquitin E1 enzymes. The role of these other components of the ubiquitin system in cancer has been extensively reviewed9, 174-176.

1.4.2 UBA1 inhibitors

The crystal structure of human UBA1 has not been elucidated until recently177 and most of the structural information on UBA1 and ubiquitin activation were derived from published structures of yeast, humanized yeast, mouse Uba1 and related bacterial enzymes14, 78, 163, 178-181. Nonetheless, drug discovery endeavors started earlier using several approaches including structure-based rational design relying on the close similarity between human UBA1 and related orthologs78. These efforts led to the identification of several inhibitors that act by diverse mechanisms (Table 1.6). These inhibitors can be structurally classified into ubiquitin adenylates, natural compounds, disulfides, nitro-based compounds and adenosine sulfamates. Of these, the adenosine sulfamate TAK-243 (MLN7243) is the only drug that has been advanced to phase 1 clinical trials.

Table 1.6 UBA1 inhibitors, their chemical structures and pharmacological properties Inhibitor Structure Type Malignancies Class/ IC50/Dose Mechanism TAK-243 Clinical - Phase 1 Solid & Adenosine IC50 (cell-free (MLN7243; - NCT02045095 hematologic sulphamate; UBCH10 AOB87172) (advanced solid malignancies78; substrate- thioester tumors) AML182; assisted assay) = 0.001 - NCT03816319 Cutaneous inhibition µM (Relapsed/refrac- SCC183; IC50 (cell tory AML/MDS) BCL184; lines) = 0.006 Myeloma185 - 1.31 µM PYR-41 Experimental Transformed Nitro-pyrazone 10 µM (cell- cells26; breast free ubiquitin cancer cells186; loading); several cell lines 10-25 µM (MM, MCL and (cellular CML)187 thioester formation) PYZD-4409 Experimental Leukemia (cell Nitro-pyrazone IC50 (cell-free lines and in vivo), E1 activity) = MM and solid 20 µM tumors (cell IC50=3-20 µM lines)27 (cell lines) Dose= 10 mg/kg i.p q.o.d. for 8 d JS-K Experimental Transformed Nitro- IC50 (48h): cells188; piperazine 0.3-1.2 µM Prostate cancer (MM cell cells189; MM190 lines) IC50 (72h): 2- 2.5 µM (MM primary cells)

43 44

IC50: 2 µM (cellular thioester formation) Dose in mice: 4 µM/kg TIW IV for 9 W Largazole Experimental Breast and colon Macrocyclic IC50 (cell-free cancer cells, natural thioester neuroblastoma compound; formation) = and osteosarcoma inhibits 29 µM 191 cell lines ; Lung adenylation IC50 (cancer epithelial cell step (and cell lines): 7.7 line192 several reports – 55 nM of HDACi IC50 (non- activity) transformed cell lines): 122-480 nM Adenosyl-phospho- Experimental Cell-free Ubiquitin Ki (ATP) = 50 ubiquitinol (APU) activity23 adenylate; nM Nonhydrolyza- Ki (Ub) = 35 ble analogue nM Ki (Ub conjugation) = 750 nM Panepophenanthrin Experimental Cell-free Natural IC50: 40 µM activity193 compound (cell-free thioester formation)194

NSC624206 Experimental Lung cell line and Disulfide IC50 (cell-free HCC cell line195 thioester cell-free activity formation) = 9-13 µM Himeic acid A Experimental Cell-free Natural IC65: 50 µM activity196 compound (cell-free thioester formation)

Hyrtioreticulins A Experimental Cell-free Natural IC50: 2.4 and & B activity194 compounds 35 µM, respectively (cell-free thioester formation) Ub-AMSN Experimental Cell-free Semisynthetic NA activity18, 197 ubiquitin adenylate

Ub-AVSN Experimental Cell-free Semisynthetic NA activity18, 197 ubiquitin adenylate

q.o.d = every other day; IP=intraperitoneal; TIW; 3 times weekly; SCC: Squamous cell carcinoma

Ubiquitin adenylate analogs Adenosyl-phospho-ubiquitinol (APU) was synthesized in 1990 and is reportedly the first experimental inhibitor of UBA123. It is a stable analogue of the ubiquitin-adenylate intermediate, formed during the adenylation step of ubiquitin activation, with a methylene group replacing the carbonyl of the C-terminal glycine residue70. Biochemically, it inhibits ubiquitin activation in a competitive manner with ATP and a non-competitive manner with ubiquitin, and tightly binds to the AAD of the free rather than ubiquitin-bound form. APU also inhibits protein degradation in in vitro assays; however, its structural features are not suggestive of appreciable cell permeability70. As such, the biochemical properties of this inhibitor, particularly affinity and ATP competitiveness, are not sufficiently favorable to produce a physiologically relevant UBA1 inhibition70. However, APU or related analogs could potentially serve as chemical probes to interrogate ubiquitin biology. In this respect, Ub-AMSN and Ub-AVSN have been developed as semisynthetic mechanism-based UBA1 inhibitors by chemical ligation of truncated ubiquitin to a synthetic cysteylglycylglycyl tripeptide with a C-terminal 5′-sulfonyladenosine and an N- terminal cysteine197. While structurally related, Ub-AMSN and Ub-AVSN possess distinct features that allow them to inhibit different steps of ubiquitin activation. Specifically, Ub-AMSN inhibits the adenylation step due to its non-hydrolyzable sulfamide that mimics the phosphate group in the ubiquitin-AMP intermediate. On the other hand, Ub-AVSN inhibits the ubiquitin thioester formation step due to its electrophilic vinyl sulfonamide that traps the nucleophilic cysteine of UBA1. In cell-free assays, Ub-AMSN and Ub-AVSN selectively inhibits Uba1- ubiquitin thioester formation in a concentration-dependent manner. In addition, Ub-AVSN forms a covalent adduct with Uba1 that is resistant to hydrolysis by dithiothreitol (DTT) as opposed to the native Uba1-ubiquitin adduct. Importantly, these inhibitors have been exploited to trap ubiquitin activation intermediates to gain insights into the biochemical and structural basis of ubiquitin activation18.

Natural compounds This class includes panepophenanthrin, himeic acid A, hyrtioreticulins A/B, and largazole. The UBA1 inhibitory activities of these compounds were not extensively characterized or verified in other studies. In addition, all of them except largazole were not tested in cellular contexts. Panepophenanthrin is a bioactive compound derived from the fermented broth of the mushroom strain, Panus rudis Fr. IFO 8994. It is reported to possess UBA1 inhibitory activity in a cell-free

46 47 assay193. Himeic acid A is another natural compound isolated from the marine fungus Aspergillus sp196. Similar to panepophenanthrin, it inhibits ubiquitin loading of UBA1 in vitro at a high micromolar concentration, with no further characterization of its UBA1 inhibitory activity in independent studies. Hyrtioreticulins A and B are indole alkaloids isolated from the marine sponge Hyrtios reticulatus194. When tested in vitro, they showed a more potent UBA1 inhibitory activity compared to panepophenanthrin and himeic acid A; however, they displayed no activity against HeLa cells. While himeic acid A inhibits ubiquitin binding to the AAD of UBA1, it appears that hyrtioreticulins A and B act at the ubiquitin-UBA1 thioester formation step194. Largazole is a macrocyclic compound isolated from the Floridian marine Cyanobacterium Symploca sp. It demonstrates a potent and preferential antiproliferative activity against breast, osteosarcoma, colon and neuroblastoma cell lines with relatively less activity against non- transformed cells191. While the mechanism of this antiproliferative action has been attributed to histone deacetylase (HDAC) inhibition in several studies198-202, one study has reported that largazole and select analogs also possess a UBA1 inhibitory activity192. Specifically, largazole was identified as a top hit through a cell-based screen for compounds that stabilize GFP-labeled p27 in a lung epithelial cell line with a potent activity in the nanomolar range. Further investigation using cell-free assays demonstrated that largazole and ketone/ester analogs inhibited the adenylation step of UBA1 activation resulting in decreased ubiquitin charging of CDC34 and reduced ubiquitylation of p27 with no effect on SUMOylation. Nonetheless, the higher micromolar IC50 required for in vitro UBA1 inhibition as opposed to the nanomolar cellular activity suggests that largazole-induced cytotoxicity is mediated through HDAC inhibition and subsequent cell cycle arrest rather than UBA1 inhibition192, 203.

Nitro-pyrazones and -piperazines Nitro-based UBA1 inhibitors include pyrazones (PYR-41 and PYZD-4409) and piperazines (JS- K). They share a nitro group that is susceptible to displacement by nucleophilic attack and covalent interaction with the active site cysteine of the CCD resulting in irreversible inhibition. PYR-41—the prototype of this class—was reported in 2007 as the first cell permeable inhibitor of UBA126. It was discovered through an in vitro HTS for compounds that inhibit ubiquitin thioester bond formation with E1 and E2 enzymes. In vitro, PYR-41 demonstrates an activity against the E1 enzyme with some activity against HECT E3s. Biochemically, it inhibits ubiquitin loading to free UBA1 with no effect on ubiquitin transfer to E2s from ubiquitin-loaded enzymes.

It also inhibits degradative and non-degradative forms of ubiquitylation and cytokine-induced activation of NF-κB. p53 is an important transcription factor and tumor suppressor protein whose stability is prone to regulation by degradative ubiquitylation204. PYR-41 induces stabilization of p53 with a subsequent increase in transcriptional activity and preferential cytotoxicity to transformed cells with wild-type p53. Cellular effects of PYR-41 include reductions in both ubiquitin loading of UBA1 and abundance of ubiquitylated proteins. Interestingly, increases in high-molecular weight (>250 kDa) ubiquitylated proteins as well total SUMOylation is observed after PYR-41 treatment with no effect on neddylation26, 187. This paradoxical accumulation of ubiquitylated proteins is ascribed to the protein cross-linking activity of PYR-41 resulting in covalent interaction with several DUBs (e.g. USP5), kinases (BCR-ABL and JAK2) and possibly other proteins, forming high-molecular weight adducts. Therefore, PYR-41 is a promiscuous inhibitor with multiple targets, and its antineoplastic effects are likely due to the combined inhibition of UBA1 as well as other off-target effects particularly after prolonged treatment187. PYZD-4409 is a structural analog of PYR-41 discovered through a focused screen of a pyrazolidine pharmacophore-based chemical library27. In addition to cancer cell lines, it has been tested in primary cancer cells and a mouse model of murine leukemia. It is cytotoxic to cell lines of hematologic and solid malignancies with MM cell lines being particularly sensitive. It also displays preferential cytotoxicity to AML versus normal hematopoietic primary cells. Mechanistically, it stabilizes the short-lived proteins p53 and cyclin D3, and induces ER stress which is functionally important for cell death. Nonetheless, the structural similarities to PYR-41 suggest that PYZD-4409 may also induce cytotoxicity through off-target effects besides UBA1 inhibition. JS-K is a nitric oxide (NO•)-releasing prodrug that is metabolically activated by glutathione S- transferases (GSTs) in the presence of glutathione (GSH). Therefore, it is preferentially cytotoxic to GST-overexpressing malignancies such as MM190. In this context, it is cytotoxic to MM cell lines and primary cells with a lesser effect on normal hematopoietic cells. It has also been tested in a mouse xenograft model of MM without significant toxicity. Mechanistically, JS-K induces apoptosis, DNA double-strand breaks (DSBs), activation of DNA damage response and JNK pathways. The apoptotic effects of JS-K in preclinical models of MM were attributed to NO• generation without implicating UBA1 (Ref. 190). A later study, however, showed that JS-K induced S-nitrosylation of UBA1 in immortalized human cells, resulting in UBA1 inhibition along with other effects similar to those reported with nitro-pyrazones in transformed cells188.

Effects on ubiquitylation and p53 levels have been further corroborated in prostate cancer cells189. Therefore, the nitro group of JS-K serves as a source of NO• that causes multi-target inhibition and is likely a leaving group at which covalent adduct formation occurs with UBA1 in a similar manner to nitro-pyrazones.

Disulfides As with largazole, the disulfide compound NSC624206 was discovered through a cell-based screen of an NCI library for compounds that prevent degradation of p27 (Ref. 195). NSC624206 was selected as a top hit based on the results of a subsequent cell-free p27 ubiquitylation assay. Through in vitro E1/E2 thioester formation assays, NSC624206 showed a selective inhibition of UBA1 over several E2 enzymes. As the inhibitory activity is lost with an analog lacking the disulfide group, this group is thought to be essential for activity. In a similar manner to nitro- pyrazones, NSC624206 inhibits UBA1~ubiquitin thioester formation, with little or no effect on the adenylation step. It is speculated that that the disulfide group of NSC624206 may undergo a nucleophilic attack by the -SH group of the active cysteine in a thiol-disulfide exchange reaction, forming a covalent adduct coupled with UBA1 inhibition. NSC624206 is a cell-permeable inhibitor as it stabilizes p27 in lung epithelial and liver cancer cells.

TAK-243 TAK-243 (MLN7243) is a first-in-class mechanism-based UBA1 inhibitor78. It was identified through a screen of more than 700 chemical candidates using an in vitro transthiolation assay that assesses ubiquitin loading onto the ubiquitin-specific E2 enzyme UBCH10. It is a second-in- class adenosine sulfamate which comes after the NAE inhibitor pevonedistat, the prototype of this structural class42. TAK-243 has been reported to be partially selective for UBA1 as it displays activity against UBA6 and NAE in the low nanomolar range but with several-fold 24, 78 higher IC50 in transthiolation assays . Nonetheless, it displays minimal or no activity against kinases and carbonic anhydrases. It forms a covalent adduct with ubiquitin (through a covalent linkage between the sulfamate nitrogen of TAK-243 and glycine 76 of ubiquitin), which tightly binds to the nucleotide-binding site and inhibits UBA1 by a mechanism known as substrate- assisted inhibition78, 205 (see below). In cells of different origins, TAK-243 reduces the abundance of mono- and poly-ubiquitylated proteins and induces the accumulation of ubiquitin-regulated short-lived proteins including p53, c-Jun, c-Myc, MCL1, and XIAP78, 182, 184, 185. These changes are associated with a multitude of

50 ubiquitylation-dependent biological effects that lead to cell death including cell cycle arrest, proteotoxic stress and DNA damage stress. In this respect, TAK-243 induces cell cycle arrest at the G2/M phase or both G1 and G2/M phases at high concentrations. Notably, the cell cycle phenotype observed after TAK-243 treatment resembles that observed after proteasome inhibitors and differs from NAE inhibitors as TAK-243 exhibits no induction of re-replication phenotype78, or a slight induction of this phenotype at high concentrations184. Given the role of UPS in the degradation of misfolded and unfolded proteins in the ER, TAK-243 induces accumulation of such proteins leading to sustained ER stress associated with morphologic ER changes, a terminal unfolded protein response (UPR) and apoptosis78, 184, 185, 206. Moreover, ubiquitylation plays a pivotal role in several DNA repair pathways such as translesion synthesis, Fanconi anemia pathway, DSB repair and nucleotide excision repair207; thus TAK-243 impairs DNA repair and induces DNA damage stress under irradiated and unirradiated conditions78, 182. Of note, genomic and proteomic analysis of TAK-243 effects demonstrates a pleiotropic impact on diverse signaling pathways including proliferative signaling, inflammatory responses, apoptosis, hypoxia, oxidative stress, and oxidative phosphorylation185. As expected with these pleiotropic effects, TAK-243 demonstrates a broad anti-tumor activity in several malignancies including solid tumors78, AML78, 182, myeloma78, 185, BCL184, cutaneous SCC183, and small-cell lung cancer (SCLC)208. Most tested cell lines of these malignancies are sensitive with low nanomolar IC50 values and robust responses in vivo and only few show intrinsic resistance78, 185, 208. Nonetheless, data with normal human fibroblasts and in vivo studies show a lesser effect on normal compared to cancer cells, suggesting TAK-243 possesses a therapeutic window in such malignancies78, 182. While seemingly narrow, this therapeutic window may lead to beneficial therapeutic outcomes in selected malignancies with dose optimization and careful patient selection as is the case with proteasome inhibitors78, 209. Based on preclinical data including efficacy and toxicology studies, TAK-243 has been advanced to phase 1 clinical trials in advanced solid tumors and AML (Table 1.7). TAK-243 also exhibits additive and/or synergistic effects with radiotherapy and DNA damaging agents in preclinical models of breast cancer and lung cancer78, 207, 208, as well as HDAC inhibitors in MM185. Due to the different mechanism of action and biological effects that are distinct from proteasome inhibitors, TAK-243 can overcome specific forms of resistance to these drugs183, 185. Interestingly, BCL cells with Myc overexpression are more sensitive to TAK-243 suggesting higher dependency on ubiquitylation in such cells, and resembling in that the effects

51 observed with SUMOylation inhibitors184 (see below). While many aspects of TAK-234 action in several malignancies have been studied, the determinants of sensitivity and resistance remain largely unknown210. With the drug being tested in clinical trials, it is important to identify patient cohorts who are likely to respond and to maximize the efficacy and minimize unnecessary toxicity from the drug. Such endeavors require a deeper understanding of the drug action at the cellular and molecular levels to identify transporters, metabolic factors, signaling molecules and/or (epi)genetic mechanisms that may influence response. In this context, unbiased large-scale genetic and chemical screens can serve as a rapid and adaptable tool to achieve this goal and provide, with appropriate validation, useful insights into TAK-243 response211. Of note, it has been attempted to correlate TAK-243 sensitivity with UBA1 expression levels in cell lines; however, the data are conflicting, and the number of cell lines tested is not large enough for conclusive findings78, 183. Equally important is the development and optimization of biomarkers to monitor drug response in these clinical trials. In this thesis, the focus will be on the preclinical evaluation and determinants sensitivity of TAK-243 in AML (see next chapters).

Table 1.7 Clinical trials of TAS4464, TAK-243 and TAK-981 Identifier Tile Phase Disease Single/Combo Status NCT02045095 A Phase 1, Dose Escalation 1 Advanced Single T Study of MLN7243 in Adult Solid Tumors Patients with Advanced Solid Tumors NCT03816319 TAK-243 in Treating Patients 1 AML; MDS; Single N with Relapsed or Refractory CMML AML or Refractory MDS or CMML NCT03648372 A Study to Evaluate the 1 Metastatic Single R Safety, Tolerability and PK of Solid Tumors; TAK-981 in Adult Lymphomas Participants with Metastatic Solid Tumors or Lymphomas. NCT02978235 A Dose Finding Study 1 Multiple Single T Followed by a Safety and Myeloma; Efficacy Study of TAS4464 Lymphoma for Patients with MM or Lymphoma N: Not yet recruiting; R: Recruiting; T: Terminated

1.5 SUMO-activating enzyme (SAE)

SUMOylation is a PTM that regulates several cellular processes including nuclear transport, gene expression, chromosome segregation, quality control and proteasomal degradation of proteins, DNA damage response and cell cycle progression41, 212. SAE is the E1 enzyme that catalyzes SUMO activation, which is followed by SUMO conjugation to a variety of cellular proteins8. Structurally, SAE is a canonical heterodimeric enzyme composed of two subunits: SAE1 and SAE2 (UBA2), with several domains similar to UBA1 and NAE (Fig. 1.2). Of the four SUMO isoforms known so far, SUMO1-3 are activated by SAE and it is not known if SUMO4 is conjugated to cellular proteins41. Compared to the ubiquitylation cascade that involves tens of E2 enzymes, hundreds of E3 ligases and over 100 DUBs, SUMOylation cascade is much less complex as it involves a single E2 enzyme, UBC9, as well as few E3 ligases and SUMO-specific proteases (SENPs)41.

1.5.1 Role of SAE in cancer

Given its critical role in many cellular processes, SUMOylation is tightly coordinated to maintain normal cellular functions40. In cancer, however, SUMOylation may be dysregulated and mostly activation of the SUMOylation pathway is associated with an adverse outcome. The expression, activity or stability of SAE1/2 are among the key regulators of SUMOylation, and either SAE1, SAE2 or both may be altered in several malignancies. For example, gastric cancer and SCLC show a high expression of SAE2, which promotes tumor progression and correlates with a poor prognosis213, 214. SAE1 is upregulated in glioblastoma tumors compared to normal brain samples215, and is a part of a 75-gene signature associated with lymph node metastasis in lung adenocarcinomas216. SAE expression is also reported to be important for Myc-driven tumor progression in solid and hematologic malignancies. For example, SAE1/2 expression was found to be high in Myc-overexpressing lymphoma, which was functionally important for Myc-driven tumorigenesis as assessed by genetic and pharmacologic approaches217. Similarly, low expression of SAE1/2 in Myc-high breast cancers is associated with favorable survival outcomes in patients218. In this context, targeting SAE is synthetically lethal with hyperactive Myc, demonstrating SUMOylation is crucial to tolerate aberrant Myc signaling and maintain tumor progression possibly through alleviating mitotic stress. Similarly, higher dependence on

53 54

SUMOylation enzymes including SAE is observed in breast cancer cells with Notch1 activation219. High expression of SAE1 is associated with unfavorable prognosis in liver, renal and thyroid cancers, with similar prognostic outcome with SAE2 in liver and renal cancers54. In preclinical models of CRC, SAE has been found to play a critical role in CSC self-renewal and maintenance220. The role of other components of SUMO pathway in cancer has been reviewed elsewhere40, 212. These data provide a compelling evidence that SUMOylation may be hyperactive in cancer compared to normal cells in order to support cellular stresses associated with the malignant phenotype. This form of non-oncogene addiction can be therapeutically targeted at different levels of the SUMOylation machinery including SAE34.

1.5.2 SAE inhibitors

The structural basis of SUMO activation by human SAE and interaction with UBC9 has been elucidated since 2005, and most SAE inhibitors reported later were natural compounds whose SAE-inhibiting activity was discovered without regard to the crystal structure of SAE18, 221, 222. These endeavors started with Gam1, an adenoviral protein, that interacts with SAE and enhances SAE-UBC9 proteasomal degradation223. This was followed by the discovery of a series of small- molecule SAE inhibitors with diverse structures until the first mechanism-based selective SAE inhibitor, ML-792, was developed more than a decade later224. SAE inhibitors reported so far can be classified into SUMO adenylates/peptide analogs, natural compounds, urea-based derivatives, oxabicycloheptadienes and adenosine sulfamates (Table 1.8). A first-in-class SAE activator, N106, was reported in 2015 as a potential modulator of heart failure225.

SUMO adenylates and peptide analogs SUMO-AMSN and SUMO-AVSN are semisynthetic mechanism-based SAE inhibitors that are structurally and biochemically analogous to Ub-AMSN and Ub-AVSN. They were similarly exploited as chemical probes to interrogate the structural basis and biochemistry of SUMO activation in cell-free assays18, 197. Based on the SUMO C-terminal sequences that can be activated by SAE as profiled by phage display technology, Zhao et al. developed pS50 and pS90, other two mechanism-based SAE inhibitors226.

Table 1.8 SAE inhibitors, their chemical structures and pharmacological properties Inhibitor Structure Type Malignancies Class/ IC50 Mechanism TAK-981 Clinical: Phase BCL and Adenosine EC50(SUMOylati 1 - colorectal sulphamate; on assays) = NCT03648372 carcinoma227- substrate- 0.005 – 0.026 μM (metastatic 230 assisted In vivo dose = 5- solid tumors/ inhibition 10 mg/kg IV lymphoma BIW; 15 mg QW

ML-792 Experimental Colorectal, Adenosine IC50 (cytotoxicity breast and sulphamate; in cell lines) = melanoma substrate- 0.03 – 0.45 μM cancer cell assisted EC50 lines224 inhibition (SUMOylation assays) = 0.003 - 0.095 μM

COH000 Experimental Cell-free Oxabicyclohe- IC50 (cell-free activity; ptadiene; RanGAP1 lymphoma, allosteric SUMOylation) = colorectal & inhibition 0.2 µM ovarian cancer In vivo dose = 10 cell lines in mg/kg QD vitro and in vivo28, 29

55 56

Davidiin Experimental Cell-free and Natural IC50 (cell-free cell-based compound; RanGAP1-C2 activity Ellagitannin SUMOylation) = (gastric, 0.15 µM prostate & IC50 (viability) = lung cancer 8.3 – 16.4 µM cell lines)231; In vivo dose = 10 hepatocellular mg/kg TIW cancer232

Ginkgolic acid Experimental Cell-free and Natural 3 µM (in vitro cell-based compound; substrate activity233; alkyl phenol SUMOylation) derivative Kerriamycin B Experimental Cell-free and Natural IC50 (cell-free (Urdamycin cell-based compound; RanGAP1-C2 A) activity234 polycyclic SUMOylation) = antibiotic 11.7 µM

Tannic acid Experimental Cell-free and Natural IC50 (cell-free cell-based compound; rhLRH-1 activity (no Gallotannin SUMOylation) = cytotoxicity 12.8 µM with liver cancer cell line)235

Anacardic acid Experimental Cell-free and Natural 2.2 µM cell-based compound; (in vitro substrate activity233; alkyl phenol SUMOylation) derivative 27-80 µM (AML

AML236; cell lines) Median IC50= 42 µM (primary AML samples) In vivo dose = 2 mg/kg/d Compound 21 Experimental Cell-free Biaryl urea IC50 (cell-free activity237 derivative RanGAP1-C2 SUMOylation) = 14.4 µM

SUMO- Experimental Cell-free Semisynthetic NA AMSN activity18, 197 SUMO adenylate

SUMO-AVSN Experimental Cell-free Semisynthetic NA activity18, 197 SUMO adenylate pS50 Experimental Cell-free Synthetic 125 µM (Cell-free (YSFVSGG) activity226 heptameric thioester bond SUMO- formation) mimicking peptides pS90 Experimental Cell-free Synthetic 131 µM (YQYVSGG) activity226 heptameric (Cell-free SUMO- thioester bond mimicking formation) peptides

OD: once daily; QW: once weekly; BIW: twice weekly; TIW: 3 times weekly

pS50 and pS90 are heptameric SUMO-mimicking peptides that can be activated by SAE, conjugated to UBC9 and transferred to protein substrates at a higher rate compared to the wild- type SUMO, but with little or no functionality. Therefore, these peptides competitively inhibit SUMOylation at every step of the cascade starting with the SAE-catalyzed activation step226.

Natural compounds This class of SAE inhibitors includes ginkgolic acid, anacardic acid, davidiin, tannic acid, and kerriamycin B. They have diverse chemical structures and molecular mechanisms of SAE inhibition. A cell-based screen of herbal extracts and nutraceuticals for compounds that inhibit SUMOylation identified ginkgolic acid, anacardic acid and davidiin as SAE inhibitors231, 233. Ginkgolic acid is an alkylphenol derivative and one of the active ingredients of gingko (Gingko biloba) leaves extract, and anacardic acid is a structurally related analog found in cashew nut (Anacardium occidentale) shells and immature gingko seeds231, 233. Both compounds reduce cell- free protein-specific and cell-based global SUMOylation in a concentration- and time-dependent manner, with no effect on ubiquitin conjugates. Structure-activity relationship (SAR) and molecular docking studies suggest the carboxylic acid and long aliphatic side chains are essential for activity and that they potentially target the AAD of the enzyme238. While these compounds are reported to possess multiple molecular activities, it appears that effects on SAE are achievable at lower concentrations especially in cells233. A subset of these other targets may themselves be subject to regulation by SUMOylation. For instance, NF-κB pathway, regulated partly by SUMOylation of NEMO and IκBα40, is modulated by anacardic acid resulting in antitumor effects in cancer cell lines of solid and hematologic malignancies239. Inhibition of NEMO SUMOylation by ginkgolic acid has also been reported in cell-free assays238. Therefore, it is possible that inhibition of SAE and subsequent effects on global SUMOylation mediate the pleiotropic effects of these two compounds on other targets. Ginkgolic and anacardic acids have also been reported to display antitumor effects in cancers addicted to oncogenes that promote a higher reliance on SUMOylation. In this respect, anacardic acid selectively induces growth arrest, apoptosis and aneuploidy in BCL cell lines with Myc activation217. Similarly, ginkgolic acid reduces proliferation and induces apoptosis in engineered and patient-derived breast cancer cell lines with Notch1 activation, without cytotoxicity on isogenic cell lines with normal Notch1 activity219. Consistent with these findings, ginkgolic acid

58 59 only reduced migration in wound-healing assays of MCF7 and MDA-MB 231 breast cancer cell lines without inducing cytotoxicity238. Inhibition of SUMOylation by anacardic acid has been reported to overcome resistance of AML to conventional therapies. In this context, treatment with daunorubicin, cytarabine or etoposide induced ROS-dependent deSUMOylation that was associated with transcriptional alterations and apoptosis in chemosensitive AML cells236. Intrinsically chemoresistant AML cells, however, failed to undergo deSUMOylation after chemotherapy. Treatment with anacardic acid in vitro and in vivo induced deSUMOylation and circumvented AML chemoresistance. Most AML subtypes, except for acute promyelocytic leukemia (APL), are not responsive to differentiation therapy with all-trans retinoic acid (ATRA). Baik et al. reported this was partly due to SUMOylation that repressed ATRA-induced expression of genes involved in myeloid differentiation, proliferation and apoptosis240. Treatment with anacardic acid among other SUMOylation inhibitors sensitized non-APL AML cells to ATRA, possibly reviving its utility in subtypes of AML other than APL240. ATRA is mostly used in APL in combination with arsenic trioxide (ATO) that acts partly by inducing hyperSUMOylation and subsequent degradation of PML/RARA (Box 1.1). Therefore, SUMOylation inhibitors may be beneficial when combined with ATRA therapy in non-APL AML, while in theory may cause adverse outcomes when combined with ATO therapy in APL.

Box 1.1 Role of SUMOylation in differentiation therapy of APL Acute promyelocytic leukemia (APL) is a subtype of AML characterized by a chromosomal translocation [t(15;17)(q22;q12)] involving the Promyelocytic Leukaemia (PML) gene and the Retinoic Acid Receptor-Alpha (RARA) gene, producing the fusion oncoprotein PML/RARA241. Biologically, PML/RARA contributes to APL pathogenesis by deregulation of RARα-mediated transcriptional control and disruption of PML nuclear bodies (PML NBs)—spherical, stress- sensitive nuclear structures involved in regulating many cellular functions and are mainly organized by PML protein241, 242. These effects lead to resistance to apoptosis, differentiation block, and enhancement of self-renewal of myeloid progenitors243. This oncoprotein is the main oncogenic driver in several types of APL, providing a unique opportunity for targeted therapy with the aim of inducing differentiation. Currently, differentiation therapy comprised of ATO plus ATRA is a frontline therapeutic option that has transformed the outlook of APL in many patients with cure rates up to 90%244-246. Mechanistically, ATO and ATRA bind to PML/RARA

60 to induce its degradation, resulting in the reformation of PML NBs, p53 activation, loss of self- renewal and differentiation241, 247-249. ATO-induced degradation of PML/RARA is dependent on both SUMOylation and ubiquitylation. Specifically, ATO targets the PML moiety of the fusion oncoprotein inducing oxidative and conformational changes that enhance recruitment and tight binding of UBC9 leading to hyperSUMOylation of PML. This form of PML actively recruits the SUMO-dependent ubiquitin E3 ligase RNF4 followed by K48-linked polyubiquitylation and proteasomal degradation250, 251. Of note, genetic and chemical inhibition of SAE in colon cancer cells has been reported to reduce SUMOylation and cause disruption of PML NB organization and dissociation of the PML-interacting protein death-associated protein 6 (DAXX), leading to antitumor effects224, 252. Therefore, it appears that induction of either hyper- or hypo- SUMOylation can be leveraged in different contexts to obtain beneficial therapeutic outcomes.

Davidiin is an ellagitannin found in the botanical extract of Davidia involucrata and Polygonum capitatum232. Compared to other reported naturally occurring SUMOylation inhibitors, it possesses a more potent inhibition of SUMOylation in vitro with an IC50 in the nanomolar range, and an antiproliferative activity in several cancer lines in the micromolar range. The antitumor activity of davidiin in vivo is reported to be associated with downregulation on EZH2, a histone- lysine N-methyltransferase that itself is a target for SUMOylation232, 253. Suzawa et al. performed a cellular, gene-expression-based, phenotypic screen of a 1,600-drug library for compounds that inhibit human Liver Receptor Homolog-1 (hLRH-1) SUMOylation using two genes whose expression undergoes a robust induction upon hLRH-1 deSUMOylation235. They identified tannic acid, a polyphenolic gallotannin related to davidiin and is extracted from several botanical sources. Compared to ginkgolic acid, tannic acid exhibits much less cytotoxicity in liver cancer cells; however, its effects on SAE appear to be specific and not due to promiscuous binding or modulation of ROS levels known with similar polyphenols. An earlier cell-based screen of samples from microbial cultured broth identified kerriamycin B as a potential inhibitor of SUMOylation234. Kerriamycin B is a polycyclic antibiotic produced by certain strains of actinomycetes and is reported to reduce substrate-specific and global SUMOylation by targeting SAE and inhibiting the formation of E1~SUMO thioesters.

Urea-based derivatives The discovery of this class of SAE inhibitors was informed by the crystal structure of SAE as opposed to natural compounds reported earlier237. A structure-based virtual screening of a

78,000-compound library for molecules that docked to the nucleotide-binding site of SAE identified a series of hits with IC50 values comparable to ginkgolic acid including Compound 21. However, structural features of these compounds were suggestive of unfavorable solubility properties. Computational interrogation of chemical databases and drug libraries identified structurally related compounds with a better solubility profile and comparable potency254, 255. Notwithstanding the potential of this class, further chemical optimization is still needed to generate more potent compounds with enhanced drug-like properties.

Oxabicycloheptadienes Most SAE inhibitors reported so far target either the nucleotide-binding site or the catalytic cysteine of the enzyme. Through an HTS of ∼ 300,000 small-molecule library for compounds that inhibit RanGAP1 SUMOylation, Li et al. discovered COH000, a first-in-class inhibitor that targets a novel, highly conserved allosteric site in SAE29. COH000 is an oxabicycloheptadiene- based compound that selectively and potently inhibits SUMOylation in cell-free assays with no effects on ubiquitylation. COH000 inhibits SAE~SUMO thioester formation in a non- competitive manner with either ATP or SUMO, and this is associated with a very slow off-rate suggesting irreversible inhibition via covalent binding to SAE. Mass spectrometry data show covalent adduct formation after binding to a hidden cysteine residue (Cys30), without modifying the other SAE’s non-disulfide-bonded cysteines including the catalytic cysteine (Cys173). Chemically, a Michael addition reaction is involved in the covalent bond formation between the Cys30 and an electrophilic center within the 7-oxabicyclohept-2-ene group of COH000. However, the selectivity observed with COH000 implies non-covalent interactions precede covalent adduct formation. Cys30 is located in the AAD at a site proximal to the active adenylation site; however, as per published SAE crystal structures, it is among the inaccessible cysteines, suggesting COH000 binding is associated with a set of conformational changes that expose this residue29. This has been confirmed by a detailed structural analysis of SAE in complex with COH000, where the drug was shown to target a novel cryptic allosteric site located specifically between the AAD and IAD28. While the associated conformational changes involving AAD, IAD and CCD are extension of those observed during the normal catalytic mechanism of SAE, they are differently orchestrated in a way that leads ultimately to locking the enzyme in an inactive conformation. Interestingly, COH000 induces ubiquitylation and degradation of SAE2, suggesting this inactive conformation is susceptible to proteasomal

62 degradation. The anti-SAE activity of COH000 is associated with reduced proliferation, apoptosis, increased miR-34b expression and decreased c-Myc expression in lymphoma and CRC cell lines and primary cells as well as reduced tumor growth in an esterase-deficient mouse xenograft model of colorectal cancer29. Since COH000 contains two ester bonds, it is susceptible to hydrolysis by plasma esterases; however, it is unclear whether cleavage of these bonds will compromise the efficacy of the drug in vivo. Apart from the unique mechanism of action and potent biological activities, COH000 has enabled a chemically-assisted discovery of a novel cryptic pocket that was not previously reported and provided new insights into the structural biology of SAE and possibly other E1 enzymes29.

Adenosine sulfamate SAE inhibitors This class comprises ML-792 and TAK-981 that share similar profiles but differ in terms of their in vivo properties256. In pursuit of developing other E1 inhibitors similar to pevonedistat, ML-792 was discovered through a drug development program that utilized pyrazole-carbonylpyrimidine- based scaffold224. Similar to other adenosine sulfamates, it acts by substrate-assisted inhibition and forms a covalent adduct with SUMO (see below). In cell-free and cell-based assays, ML-792 demonstrates a potent and selective SAE inhibitory activity, with little or no activity against related E1s or other ATP-dependent enzymes. ML-792 is also cytotoxic to CRC, breast and melanoma cancer cell lines, with Myc-overexpressing cell lines being particularly sensitive. Despite the importance of SUMOylation pathway in regulating gene expression, ML-792 only induces modest transcriptional and proteomic changes, with no consistent impact on splicing in different cell lines. As opposed to TAK-243, ML-792 has little or no effect on DNA repair suggesting DNA damage plays no role in the observed cytotoxicity224. Instead, ML-792 induces mitotic disruption (multinucleation, cell enlargement, and endoreduplication), senescence, and chromosome segregation defects (anaphase/telophase decrease and formation of DNA bridges). Cytotoxicity and cell cycle changes are evident only in actively cycling cells and can be rescued by overexpression of ML-792-resistant UBA2 S95N M97T mutant, suggesting mitotic disruption is functionally important for ML-792 cytotoxicity224. Despite the excellent profile of ML-792 ex vivo, it shows a poor stability in vivo making it suitable only as a chemical probe for interrogating SUMOylation biology. Specifically, high doses (150-200 mg/kg) are required to maintain in vivo responses as ML-792-SUMO adducts are rapidly released from SAE and degraded after in vivo administration256. Conversely, TAK-981—

63 developed thereafter to overcome this problem, forms SUMO adducts that are stable for 48 h and causes sustained inhibition of SUMOylation (16-24 h) following administration of small doses (10 mg/kg)256. TAK-981 selectively inhibits SAE-SUMO and UBC9-SUMO thioester formation as well as global SUMOylation with IC50’s in the low nanomolar range, with no significant effects on neddylation or ubiquitylation228, 256. Interestingly, prolonged effects of TAK-981 in vivo allowed the observation of several immune responses that are associated with the antitumor effects of the drug, particularly against tumor models with intrinsic resistance to TAK-981 in vitro256. The molecular basis of these responses partly derives from the role of SUMOylation in regulating type I interferon (IFN) signaling257, 258 (Box 1.2). This is consistent with the reduced antitumor effects of TAK-981 with prior administration of anti-IFN α/β receptor 1 (IFNAR1) antibody228. These effects mediated by type I IFN signaling are pleiotropic and implicate several effectors of innate and adaptive arms of immunity227-229. Direct micro-injection of TAK-981 into tumors is also associated with both local and abscopal immune responses that are consistent with those observed after systemic dosing230. In addition, TAK-981 exerts synergistic antitumor effects when administered in combination with other immunotherapies such as rituximab, an anti-CD20 monoclonal antibody, and immune checkpoint inhibitors227, 256. It appears, however, that these immune effects are not tumor- agnostic as treatment with TAK-981 in mice bearing specific tumors confers immunity only against those tumors but not other types256. Based on these preclinical data, TAK-981 is currently evaluated in a phase 1 clinical trial for patients with metastatic solid tumors and lymphomas (Table 1.7).

Box 2 Immunoregulatory roles of SUMOylation Interferons (IFNs) are small secreted proteins that play a crucial role in innate and adaptive immunity. They include three classes: type I (IFN-α, IFN-β, IFN-ε, IFN-κ, IFN-ω), type II (IFN- γ), and type III INFs (IFN-λ, IL-28 and IL-29) that initiate three distinct signaling cascades259. SUMOylation has been reported to regulate type I IFN signaling. Type I IFN signaling is initiated by binding of IFNs to a heterodimeric cell surface receptor complex composed of IFN α receptor 1 (IFNAR1) and IFNAR2 leading to activation of JAK-STAT signaling and formation of heterotrimeric complex termed IFN-stimulated gene factor 3 (ISGF3) and is composed of STAT1, STAT2 and IFN-regulatory factor 9 (IRF9). ISGF3 undergoes nuclear translocation where it activates expression of interferon stimulated genes (ISGs) by modulating IFN-

64 stimulated response elements (ISREs). Recently, SUMOylation has been reported to negatively regulate the innate inflammatory response by silencing IFN-β expression and attenuating priming of inflammatory cytokines production by type I IFNs. Accordingly, inhibiting SUMOylation by UBC9 knockdown enhanced the innate immune responses257. Loss of SUMOylation after SAE or UBC9 knockout was also found to induce a type I IFN response, however, through a non- canonical mechanism independent of known IFN-regulating pathways258. Among the inducers of type I IFN signaling is the cGAS-STING pathway that acts as a sensor of aberrant cytosolic DNA and activates the expression of type I IFNs259, 260. SUMOylation has been reported to suppress the activity of cGAS, and this suppression was reversed by the SUMO-specific protease SENP7 that catalyzes deSUMOylation of cGAS triggering type I IFN-dependent innate immune responses261. The importance of SUMOylation is not confined to regulating innate immunity, as it has been reported to play a key role in adaptive immune responses as well. In this context, SENP3 was found to maintain the stability and function of regulatory T cells (Tregs) through deSUMOylation of BACH2, a Tregs-specific transcription factor, facilitating its nuclear retention and function262. Therefore, it appears that SUMOylation restrains a subset of innate and adaptive immune responses, and thus targeting the SUMOylation machinery including SAE can serve, in specific contexts, as a potential therapeutic strategy to obtain useful antitumor and antimicrobial immune effects.

Dual and multi-E1 inhibitors E1 inhibitors discussed so far have variable selectivity profiles toward E1 enzymes or other targets of the proteome. Of these, most adenosine sulfamates particularly those in clinical trials, TAS4464 and COH000 exhibit a high level of selectivity toward a specific E1 enzyme, with little or no activity against other E1 members. On the other hand, other E1 inhibitors particularly those with relatively high EC50 values are expected to exhibit multiple off-target effects besides their E1 activities. In between, there exists a third subclass whose members possess a potent activity toward two or more E1 enzymes. These include the non-nucleoside derivative LP0040 and the adenosine sulfamates Compound 1, ABP1 and ABPA3 (Table 1.9). In addition, several investigational adenosine sulfamates with such activity—but are not as well characterized— have been identified during the chemical optimization stage of pevonedistat and TAK-243 development such as Compound 15 and 46 (Ref. 24).

Multi-E1 adenosine sulfamates Members of this class retain several structural features of AMP including the ribose sugar and the adenine base. Compound 1 is a pan-E1 inhibitor that was exploited to elucidate important mechanistic aspects of substrate-assisted inhibition70. While having variable potencies against E1 enzymes in ATP-PPi exchange assays, with the highest potency observed with NAE, Compound 1 is equipotent against UBA1, NAE, and SAE in E1-E2 transthiolation assays70, 263. Of note, Compound 1 displays no appreciable activity against ATG7 in ATP-PPi exchange assays. Replacing the indane moiety of Compound 1 with a propargyl moiety generates another non- selective E1 inhibitor, ABP1, with a free alkyne group that has been leveraged for covalent conjugation to fluorescent or biotin tags using click chemistry264. This advantage has enabled the use of ABP1 as a probe to quantify the activity of E1 enzymes as the adducts are formed by the enzymes themselves and thus the rate of adduct formation is dependent on E1 activity (see below). In addition, it has been used to provide further insights into the biochemistry of UBL conjugation and evaluate the potency and selectivity of investigational E1 inhibitors264. Experience with ABP1 has been exploited to develop a related analog, ABPA3, that acts as a dual UBA1/NAE inhibitor. ABPA3 is also structurally related to Compound 1 with a phenylacetylene replacing the indane moiety265. ABPA3 displays anti-UBA1/NAE activity in lung cancer cells which induces UPR and prevents cytoprotective aggresome formation leading to cell death. Of note, ABPA3 paradoxically causes increases in SUMOylation and ufmylation which are likely compensatory responses activated in response to UPR induction265.

LP0040 LP0040 is a non-nucleoside dual UBA1/NAE inhibitor developed by rational drug design and molecular modeling based on the structure of the NAE inhibitor M22 (Ref. 266). Replacing the benzyl group of M22 with biphenyl-substituted 2H-chromen-2-one in LP0040 led to an improved anti-proliferative activity associated with cell cycle arrest and apoptosis in cancer cell lines of different origin with a low micromolar IC50. LP0040 has also displayed synergistic cytotoxicity with bortezomib266.

Table 1.9 Multitarget E1 inhibitors, their chemical structures and pharmacological properties Inhibitor Structure Type Malignancies Class/Mechanism IC50 Compound 1 Experimental Chemical probe Adenosine IC50= 0.01 – 6.4 pan-E1 with cell-based sulphamate; μM (ATP-PPi inhibitor activity70, 263 substrate-assisted exchange assays) inhibition IC50 (UBA1/NAE/ SAE-E2 cell-free transthiolation) = 0.005 μM ABP1 Experimental Chemical probe Adenosine NA pan-E1 with cell-based sulphamate; inhibitor activity in lung substrate-assisted cancer cells 264 inhibition

ABPA3 Experimental Cell-based activity Adenosine IC50 (viability in UBA1/NAE in lung sulphamate; A549 cells) = 2.5 inhibitor cancer cells265 substrate-assisted μM inhibition

LP0040 Experimental Cell-based activity 2H-chromen-2-one IC50 (viability in UBA1/NAE in cancer cells266 derivative cell lines) = 0.76- inhibitor 3.29 μM

1.6 Adenosine sulfamates

This class comprises pevonedistat, TAK-243, ML-792, TAK-981, Compound 1, ABPA3, and ABP1 which act as mechanism-based inhibitors of E1 enzymes. These compounds were first developed by Millennium Pharmaceuticals, which is currently an oncology subsidiary of Takeda. They are structurally related to AMP that tightly binds to E1 enzymes during the catalytic cascade involved in UBL activation42. The structures of adenosine sulfamates share three major chemical features: 1) a ribose or cyclopentane moiety, 2) a sulfamate group analogous to the phosphate group in AMP, and 3) a nucleobase extension analogous to the adenine base in AMP (Fig. 1.8).

1.6.1 Discovery and structure-activity relationship

The discovery and SAR of adenosine sulfamates has been reviewed in detail and here we briefly highlight the main features of these programs24. An HTS identified N6-substituted adenosine derivatives as initial chemotypes with weak NAE inhibitory activity. Further chemical optimization was required to enhance potency, cellular activity, selectivity and stability24 (Fig. 1.9). Although phosphate derivatives demonstrated sub-micromolar potency, they did not permeate into cells and thus lacked cellular activity24. Replacement with the neutral sulfamate group retained potency and yielded a series of cell-active E1 inhibitors. The sulfamate functional group is the closest bioisostere to phosphate and does not undergo promiscuous thioester formation in cells as a result of its low intrinsic reactivity265. Compared to sulfamate, sulfonamide and sulfamide groups are less favorable in terms of cellular potency; therefore, the sulfamate group remained as a common chemical feature essential for the cellular activity, potency and selectivity of adenosine sulfamates. In this series, certain compounds were prone to an intramolecular cyclization reaction resulting in the loss of activity. In addition, several compounds lacked selectivity for individual E1 enzymes. A combination of chemical modifications in the ribose sugar, purine base and stereochemical orientation of the sulfamate group solved these stability and selectivity issues and retained or even improved potency and cellular activity, leading to pevonedistat as the prototype of this class24, 42.

67 68

Figure 1.8 Structural features of adenosine sulfamates. The chemical structures of TAK-243, pevonedistat, ML-792 and TAK-981 are displayed compared to AMP. Adenosine sulfamates share common chemical features mainly: 1) sulfamate group, 2) cyclopentane/ribose moiety, and 3) nucleobase extension.

Figure 1.9 Discovery of adenosine sulfamates. A diagram showing the main features of the drug discovery program of pevonedistat and TAK-243. Compounds, their chemical structures and pharmacological properties are shown. Chemical moieties changed between intermediate compounds are shaded in similar colors. UAE: ubiquitin-activating enzyme; NAE: NEDD8- activating enzyme.

The success of NAE inhibitor program inspired the initiation of another program to develop UBA1 inhibitors by exploring the adenosine sulfamate series and building on the previous SAR and mechanistic studies done with pevonedistat70. While chemical modifications of the ribose sugar and aromatic side chain led to improved selectivity to UBA1, the pharmacodynamic profile in vivo was not optimal. The replacement of purine base with a pyrazolo-pyrimidine scaffold and the placement of the N9-nitrogen in an exo position improved the in vivo profile, and significantly enhanced selectivity over SAE; however, the selectivity over NAE was only partial24. Further modifications enhanced selectivity to UBA1, however, at the expense of potency which was significantly compromised. As NAE regulates the activity of only a subset of ubiquitin E3 ligases downstream of UBA1, it is anticipated that a dual UBA1/NAE inhibitor will demonstrate the same biologic activity and phenotype as a selective UBA1 inhibitor24. Therefore, chemical optimization was pursued with priority given to potency over selectivity, leading ultimately to the selection of TAK-243 as a clinical candidate.

1.6.2 Substrate-assisted inhibition

The development of pevonedistat involved an extensive characterization of its mode of action using various approaches including crystallographic studies to elucidate the structural basis of NAE inhibition by this drug70. These studies showcased a continuous electron density between the C terminus of NEDD8 and the sulfamate group of pevonedistat when bound to NAE, suggesting a pevonedistat-NEDD8 adduct is formed and possibly it is the species that inhibits NAE70. Intrigued by this interesting finding and knowing UBL activation mechanism is highly conserved among E1 enzymes, further studies have been conducted and revealed this mode of inhibition, termed substrate-assisted inhibition, is a general mechanism shared by other adenosine sulfamates including those targeting UBA1 and SAE78, 224, 263. Substrate-assisted inhibition is unique in that the UBL substrate of E1 enzymes is involved in the enzyme inhibition through chemical reaction with adenosine sulfamates to form a covalent adduct that itself serves as the active inhibitory species70. This mode of inhibition may be pharmacologically beneficial as it confers selectivity among E1 enzymes because each enzyme has its cognate UBL. In addition, it reduces the likelihood of interaction with other ATP- dependent enzymes, including kinases, which have different catalytic mechanisms and thus are less susceptible to such UBL-based species70.

To form these adducts, adenosine sulfamates act as AMP-mimetics and exploit the multi-step catalytic mechanism of UBL activation by E1 enzymes (see before). They specifically target the E1~UBL thioester intermediate—a binary complex in which E1 is loaded with a single UBL molecule at the catalytic cysteine, and bind to the nucleotide-biding site in the AAD. This is followed by a nucleophilic attack of the thioester bond by the sulfamate amino group of adenosine sulfamates to form a covalent adduct with UBL (Fig. 1.3). Consistent with a mechanism-based inhibitory mode, the adduct formation is catalyzed by the E1 enzyme itself and requires the presence of magnesium, ATP and an active catalytic cysteine, mimicking the conditions of AMP-UBL formation and implicating most sites of interaction with the E1 enzyme70, 263. While E1~UBL intermediate formation is the rate-limiting step in UBL activation under physiological conditions, biochemical studies suggest adduct formation is the rate-limiting step in E1 inhibition by adenosine sulphamates263. As opposed to the binary complex, the ternary complex loaded with two ubiquitin molecules is not susceptible to inhibition by adenosine sulfamates as the nucleotide-binding site, to which these drugs need to bind first, is occupied by AMP-UBL. After formation, adenosine sulfamate adducts are stable and tightly bind, with a picomolar affinity, to E1 enzymes in an ATP-competitive manner. By competing for E1 binding, high ATP concentrations are thus anticipated to slow the binding rate of adenosine sulfamates 263 with a resultant increase in their IC50 values . Of note, the inhibitory activity of stable AMP-UBL mimetics against E1 enzymes has been known several years before the development of adenosine sulfamates as with APU and later with UBL-AMSN, which are preformed semisynthetic UBL phospho- and sulfamoyl-adenosine derivatives18, 23. Adenosine sulfamates, however, differ in that they form adducts in situ after binding to their cognate E1 enzymes, allowing for a preferential activity against cancer cells that display higher activity of these enzymes to support their cellular stresses34, 35. It is also noteworthy that, in cell-free assays, preformed adenosine sulfamate adducts can inhibit E1 enzymes with much higher potency compared to unbound adenosine sulfamates263. This may be ascribed to their different inhibition kinetics as the adducts rapidly bind to and inhibit unloaded E1 enzymes halting all subsequent steps, whereas unbound adenosine sulfamates bind at a slower rate to the E1~UBL thioester intermediate with no effect on the first step of UBL activation263.

1.6.3 Selectivity and potency of adenosine sulfamates

Biochemical studies suggest that the potency of adenosine sulfamates is a function of the rate of adduct formation and the affinity of this adduct to E1 enzymes, and these biochemical parameters in turn depend on their structure263. While adenosine sulfamates share general chemical features as highlighted above, they still have several variations that dictate, among other properties, potency and selectivity to E1 enzymes. These include chemically diverse nucleobase extensions, different stereochemical orientation of the sulfamate group, as well as varying number and stereochemical orientation of the hydroxyl groups on the cyclopentane/ribose moiety205. For instance, TAK-243 possesses a characteristic thio(trifluoromethyl) group in its nucleobase extension which is absent in other adenosine sulfamates (Figure 1.8). This group serves as a hook that engages TAK-243 in several tight contacts with several amino acid residues within a hydrophobic pocket in UBA1, contributing to the drug selectivity compared to related analogs. On the other hand, NAE enzyme has an extended hydrophobic pocket, and the presence of a hydrophobic indane cycle in the structure of pevonedistat partly contributes to the selectivity profile of this drug through hydrophobic interactions with this extended pocket205. To exhibit a wider selectivity profile, dual- or pan- specific E1 inhibitors such as ABPA3 and Compound 1 retain the adenine base and the ribose sugar with its two hydroxyl groups that are present in the natural substrate AMP, thereby increasing their capacity to interact with multiple E1 enzymes with comparable affinity205. Although the interactions with AMP are generally conserved among different E1 enzymes, there still exist some variations that are reflected in their interactions with pan-specific E1 inhibitors with different nucleobase extensions, leading to narrower/wider selectivity profile. For example, ABP1 possesses a small propargyl group linked to the exocyclic amino group of the adenine as opposed to a larger phenylacetylene in the related adenosine sulfamate ABPA3. With its small size, this propargyl can fit into the SAE pocket leading to an inhibitory activity against SAE which is lacking with ABPA3205.

1.6.4 Resistance to adenosine sulfamates

Emergence of acquired anticancer drug resistance is a common problem particularly with MTAs including drugs targeting the ATP-dependent tyrosine kinases35, 267. Mechanisms of such resistance include the activation of adaptive/secondary signaling pathways that maintain survival

73 by compensating for the inhibitory effects on the targeted pathway, overexpression of multi-drug resistance proteins that reduce the intracellular concentrations of the drug by extracellular efflux, and the emergence of on-target mutations that interfere with binding or preclude inhibition of the intracellular target268. While the first mechanism is less likely to develop with adenosine sulfamates especially those targeting UBA1— given their pleiotropic effects, the other two mechanisms are still potential contributors to resistance due to the on-target selectivity and structural similarity of adenosine sulfamates to ATP 35. Preclinical resistance studies on adenosine sulfamates involved the development of laboratory- evolved resistance models under the selective pressure of exposure to the drug in tissue culture and in vivo. Such models led to the identification of several on-target mutations in the AADs of UBA1 and NAE that preclude binding of TAK-243 and pevonedistat, respectively182, 269-271. Milhollen et al. and Toth et al. independently identified A171T missense mutation as a mechanism of resistance to pevonedistat269, 270. Milhollen et al. also identified other mutations such as A171D, C324Y, G201V, E204K, and N209K. These mutations map to the AAD either within the nucleotide-binding pocket (A171T and A171D) or within/close to the NEDD8-binding cleft (C324Y, G201V, E204K, and N209K), reducing the binding of pevonedistat or NEDD8 to NAE, respectively269. Consistent with the heterozygosity of these mutations, pevonedistat- NEDD8 adduct can still be formed by the wild-type copy of the enzyme; however, the mutant form displays a dominant effect resulting in a resistant phenotype. Of these mutation sites, A171 is particularly important as it constitutes a hotspot where two- thirds of the detected mutations are located. A171 serves as a gatekeeper residue that plays a key role in controlling ATP and pevonedistat access to the nucleotide-binding site270. Specifically, the alanine residue with its small side chain contributes to shaping an edge of the nucleotide- binding pocket without engaging in direct contact with ATP270. Structural and computational studies suggest the replacement of alanine with threonine (A171T) or aspartic acid (A171D) that possess larger side chains leads to conformational changes and steric clashes with the aminoindane group of pevonedistat particularly with the A171D-mutant NAE which is completely insensitive to pevonedistat due to the bulkier aspartic acid residue269, 270, 272. These mutations alter two fundamental biochemical properties required for potent NAE inhibition. Specifically, they slow the rate of pevonedistat-NEDD8 adduct formation and reduce the adduct affinity to NAE resulting in the loss of tight binding and faster recovery of enzyme activity with 2- to > 2,000-fold reduction in pevonedistat potency. Compound 1, which exhibits

74 tighter binding properties to several E1 enzymes including NAE, could circumvent pevonedistat- resistant NAE mutations, suggesting second-generation NAE inhibitors with higher affinities may offer a potential opportunity to overcome acquired pevonedistat resistance. Using a similar approach, Xu et al. identified two other mutations, I310N and Y352H, that decrease the affinity of NAE for NEDD8 and confer resistance of leukemia cell lines to pevonedistat271. The A171 residue of NAE is highly conserved among E1 enzymes and thus mutation of the corresponding alanine residues in other E1s is expected to alter the response to adenosine sulfamates selective for these enzymes. In this respect, A580T mutation generated in recombinant UBA1 conferred resistance to TAK-243 in cell-free assays205. Consistent with these findings, Barghout et al. identified A580S mutation in a laboratory-evolved TAK-243-resistant leukemia cell line182. They also identified another mutation in a close residue, Y583C, which is specific for TAK-243. This mutation has not been reported with pevonedistat because TAK-243 possesses a thio(trifluoromethyl) hook through which it extends in the nucleotide-binding pocket further than pevonedistat and undergoes a favorable interaction with Y583 (Ref. 182, 205(. While not highly conserved among E1 enzymes, the 583 position is mostly occupied with amino acid residues that possess bulky side chains. Therefore, replacement with cysteine is anticipated to perturb the hydrophobic core and eliminate hydrogen bonds with neighboring residues leading to changes in the binding pocket that preclude efficient binding of TAK-243 to UBA1 (see Section 2.3.9 and 2.4). Drug resistance is known to be multifactorial and the resistance phenotype observed after clonal selection with adenosine sulfamates can result from a combination of several mechanisms. In this context, moderate upregulation of ATP-binding cassette (ABC) transporter proteins is observed in pevonedistat-resistant cells, suggesting they may contribute to drug resistance particularly at higher expression levels269. While A580T-mutant UBA1 exhibits a 6-fold resistance to TAK-243 in cell-free assays, TAK-243-resistant cell lines with A580S and Y583C mutations exhibit 28- to 33-fold resistance, suggesting other resistance mechanisms may coexist in these cells182, 205. Apart from these mechanisms, genetic knockout of the transcriptional regulator BEND3 is also reported to confer resistance to TAK-243 in a leukemia cell line; however, it remains unknown how it alters the response to this drug211.

1.6.5 Pharmacodynamic activity of adenosine sulfamates

Adenosine sulfamates targeting UBA1, NAE and SAE are currently investigated in different phases of clinical trials in various malignancies (Table 4 and 7). To maximize the benefit from such trials, activity biomarkers need to be assessed either as an evidence of mechanism or to inform clinical decision-making in the trial273. These biomarkers are usually pathway-specific, and in case of adenosine sulfamates, adduct formation, E1~UBL thioesters and/or reduction in the levels of UBL conjugates could serve as potential biomarkers. For preclinical research, qualitative or semi-quantitative approaches can be used to assess the pharmacodynamic effects of these drugs as a proof of principle. However, for clinical trials, it is important to develop reliable pharmacodynamic assays that can assess and monitor the activity of these investigational agents, preferably in a quantitative manner. Pharmacodynamic evaluation of adenosine sulfamate activity has been performed at different levels pre- and post-E1 inhibition. For example, the cellular thermal shift assay (CETSA)—a biophysical assay that depends on thermostabilization of proteins upon binding chemical ligands, has been used to assess the binding of TAK-243 to UBA1 in cells and in tumors after drug injection into mice182, 274. As the formation of adenosine sulfamate adducts is a prerequisite for potent inhibition of E1 enzymes, antibodies raised against adducts with ubiquitin (MIL90 antibody), NEDD8 and SUMO have been used to monitor drug activity in vitro and in vivo78, 224, 228, 275. In addition, levels of E1~UBL thioesters, global protein conjugates and conjugates with specific proteins such as ubiquitylated H2A, neddylated Cul1, and SUMOylated RanGAP1 have been used78, 182, 224, 275. In such preclinical studies, semi-quantitative assays such as immunoblotting and immunohistochemistry have been utilized as a proof of principle for E1 engagement and pathway inhibition. To achieve a more quantitative evaluation of these, other platforms have been developed such as the AlphaScreen format, a non-radioactive amplified luminescent proximity homogenous assay, that provides faster and more robust assessment as well as higher throughput275. Moreover, a mass spectrometry-based method applied on SDS-PAGE fractionated lysates of biological samples has been used to provide an absolute quantification of NAE and NEDD8-pevonedisatat adducts in these samples276. In addition to gaining quantitative insights into the dynamics of UBL conjugation, the molar ratio of adduct to E1 enzyme can be calculated to provide an indirect pharmacodynamic measure of pathway inhibition.

Since the adduct is formed by the E1 enzyme itself70, the rate of adduct accumulation can provide valuable information on the activity of E1 enzymes, which in turn can serve as a predictive biomarker of response to these drugs. In this context, the availability of specific antibodies against these adducts can be exploited to develop flow cytometry-based other quantitative assays that can—with appropriate analytical and clinical validation—be employed to quantitatively measure the rate of adduct formation after administration of a single dose of the drug. This may help with decision-making as to whether patients should continue therapy with these agents particularly in hematologic malignancies where blood samples can be easily obtained. It remains challenging, however, to determine the cut-offs of enzyme activity that will be used to make such decisions.

1.6.6 Cell cycle effects of adenosine sulfamates

PTMs particularly with ubiquitin, SUMO and NEDD8 play an important role in regulating cell cycle progression212, 277. Therefore, disruption of these PTMs is anticipated to result in deregulation of the cell cycle. However, the ultimate phenotype varies among different adenosine sulfamates. While inhibition of neddylation with pevonedistat causes cell cycle arrest and accumulation of cells in the S-phase, selective inhibition of ubiquitylation with a specific UBA1 inhibitor causes cell cycle arrest in the G2/M phase. Interestingly, dual inhibition of ubiquitylation and neddylation with TAK-243 also results in G2 arrest, suggesting the effect on ubiquitylation is more dominant24. While TAK-243 resembles pevonedistat in inducing accumulation of CDT1, it also stabilizes its endogenous inhibitor—geminin—offsetting CDT1 effects78. As a result, pevonedistat induces CDT1-mediated re-replication which is not observed with TAK-243. ML-792, on the other hand, causes mitotic disruption with a decrease in the number of cells in the anaphase/telophase subphases of mitosis, without affecting the number of those in the prometaphase and metaphase. These mitotic defects are associated with induction of endoreduplication and formation of DNA bridges224.

1.6.7 Clinical adenosine sulfamates: pevonedistat and TAK-243

Pevonedistat is the first-in-class prototypical adenosine sulfamate and serves as a mechanism- based NAE inhibitor42. Pevonedistat has been preclinically studied in several malignancies and evaluated in over 30 clinical trials alone and in combination with other therapeutic agents (Table

1.3 and 1.4). The discovery of pevonedistat inspired the initiation of drug discovery programs aimed at exploring adenosine sulfamates that selectively inhibit other E1 enzymes including UBA1 and SAE (Fig. 1.9). These endeavors has led to the discovery of TAK-243 and TAK-981 that target UBA1 and SAE, respectively which are currently evaluated in phase 1 clinical trials in hematologic and solid malignancies (Table 1.7)228, 278. While pevonedistat is highly specific for NAE, TAK-243 exhibits partial selectivity for UBA1 due to its ability to bind both NAE and UBA6 (Ref. 278). In cell-based transthiolation assays in colon cancer cells, TAK-243 displayed equipotent inhibition of UBA1 and UBA6, and 10-fold lower potency against NAE78. Consistent with these data, TAK-243 displayed no significant effects on neddylated proteins at concentrations known to significantly reduce the abundance of ubiquitylated proteins in AML cell lines182. Table 1.10 lists potencies of pevonedistat and TAK-243 against different E1s as 78 assessed by IC50 values obtained from cell-free transthiolation assays .

Table 1.10 Activity of pevonedistat and TAK-243 against E1 enzymes as assessed by cell- free transthiolation assays (compiled from Ref. 42, 278) Pevonedistat IC50 (nM) NAE 4.76 ± 1.5 UBA1 1,500 ± 710 SAE 8,200 ± 6,200 UBA6 1,800 ATG7 >10,000

TAK-243 IC50 (nM) UBA1 1 ± 0.2 UBA6 7 ± 3 NAE 28 ± 11 SAE 850 ± 180 UBA7 5,300 ± 2,100 ATG7 >10,000

1.7 Acute myeloid leukemia

1.7.1 Overview

AML is the most common acute leukemia in adults279. It is a clonal myeloid neoplasm characterized by the blockade of differentiation of hematopoietic stem cells (HSCs) leading to the accumulation of abnormally or poorly differentiated hematopoietic cells called blasts279. Leukemic blasts interfere with the normal function and production of healthy hematopoietic cells in the bone marrow280. As a result, a state of cytopenias (anemia, neutropenia and thrombocytopenia) develops and is responsible for the clinical manifestations of the disease. In addition to its cytogenetic heterogeneity that has long been used in prognostic and therapeutic stratification of AML patients281, molecular heterogeneity of AML is becoming increasingly appreciated due to deeper understanding of the mutational profile of this disease282-285. Therapy of AML consists of induction and consolidation chemotherapy together with HSC transplantation in selected patients279. The major chemotherapeutic agents used in AML include cytarabine and daunorubicin282. These agents have been used over the past four decades and the improvement in survival rates particularly in young adults is mostly attributable to supportive care and HSC transplantation279. Despite these improvements, more than 50% of young adults and 90% of older patients still die from AML280. The major hurdle to improving survival rates is disease recurrence. While the majority of AML patients achieve initial CR, most patients will experience a relapse285. Among the mechanisms underpinning relapse is the selection of a subpopulation of progenitor cells that are resistant to conventional chemotherapy286. Therefore, it is important to identify novel therapeutic strategies by leveraging our increased understanding of the molecular vulnerabilities of AML. In this respect, a plethora of novel therapeutics are currently investigated in patients with this disease283. These novel approaches might circumvent resistance to conventional therapies, reduce recurrence rates, and improve treatment outcomes in AML.

1.7.2 Pathobiology of AML

In normal hematopoiesis, cells maintain a hierarchical organization with all blood cell types arising from primitive non-committed cells called HSCs279. They lie at the apex of the hematopoietic hierarchy and are characterized by self-renewal and the capacity to differentiate into progenitor cells which are more committed to one or more lineages279. These in turn can differentiate into precursor cells that give rise to mature and functional blood cells of either

80 myeloid (erythrocytes, eosinophils, neutrophils, basophils, dendritic cells and macrophages) or lymphoid (lymphocytes and nature killer cells) lineage287. In AML, it is thought that malignant transformation occurs at the level of HSCs and/or progenitor cells by accumulating driver mutations that concertedly lead to malignant primitive cells called leukemic stem cells (LSCs) and/or leukemic progenitors288. These cells possess higher competitive advantage compared to normal counterparts and thus expand at the expense of normal cells resulting in clonal hematopoiesis and giving rise to AML blast cells that are not fully differentiated, non-functional, rapidly proliferating and form the bulk of AML neoplasms (Fig. 1.10). LSCs and leukemic progenitors are less susceptible to cytotoxic agents and thus are the major root of relapse289. As LSCs differentiate into leukemic progenitors, these latter cells may acquire further mutations with distinct proliferative potential, diversifying the resultant leukemic clones and leading to molecularly and cytogenetically heterogeneous AML neoplasms. Of note, several mutations particularly in the epigenetic machinery (DNMT3A, TET2, IDH2, ASXL1) have been observed in preleukemic HSCs before the development of AML, suggesting they serve as early genetic events that contribute to malignant transformation of the normal cells of origin279, 282, 285. A subset of AML, known as therapy-related AML, develops as a complication of genotoxic chemo- and radiotherapy281, 290. Clonal hematopoiesis has also been reported in a subset of patients without the development of AML—termed clonal hematopoiesis of indeterminate potential (CHIP), and those patients exhibit a higher risk of atherosclerotic coronary heart disease291.

1.7.3 Mutational landscape of AML

With the advent of next-generation sequencing, it has become increasingly possible to gain deeper insights into the mutational profiles of different cancers including AML. The number of mutations in the genome of AML patients is fewer compared to most solid malignancies285. AML is associated with a number of recurrent mutations and in most cases two or more mutations coexist to drive leukemogenesis in a cooperative manner (Fig. 1.11). These mutations influence genes implicated in many biological pathways, and thus have been categorized into functional groups as follows: signaling pathways, DNA methylation, chromatic modifiers, nucleophosmin (NPM1), transcription factors, tumor suppressors, spliceosome machinery and cohesion complex (Fig. 1.12)279, 282.

Figure 1.10 Pathobiology of AML. a) In normal hematopoiesis, blood components follow a hierarchical organization where hematopoietic stem cells (HSCs) characterized by self-renewal and capacity to differentiate into progenitor cells that differentiate into precursor cells which, in turn, produce myeloid and lymphoid blood components. b) In AML, malignant transformation at the level of HSCs results in leukemic stem cells (LSCs) that show higher competitive advantage and produce leukemic progenitors and poorly differentiated AML blasts that rapidly proliferate and do not differentiate into functional blood cells. Aberrations are restricted to myeloid lineage without affecting lymphoid lineages. NK: Natural killer; DC: Dendritic cells. Reproduced with permission from Ref. 279.

Figure 1.11 Mutational landscape of AML. Recurrent mutations reported in AML as depicted by a Circos diagram (left) showing the relative frequency and co-existence of mutations. The frequency of the mutation is represented by the length of the arc and the percentage of patients with co-existing mutations is represented by the width of the ribbon. Color codes are used to represent different genes. The overall frequency of these mutations is shown in the table (right). ITD: internal tandem duplication; TKD: tyrosine kinase domain. Reproduced with permission from Ref. 292, Copyright Massachusetts Medical Society.

Figure 1.12 Functional categories of genes frequently mutated in AML. Recurrent mutations in AML affect genes involved various signaling pathways classified into 8 functional categories (outlined in boxes). These include transmembrane receptors (FLT3), DNA modifying machineries (tumor suppressors, DNA methylation, myeloid transcription factors, chromatin modifiers, spliceosome complex, cohesion complex and nucleolar molecules (nucleophosmin; NPM1). Reproduced with permission from Ref. 282, Copyright Massachusetts Medical Society.

In addition, this mutational landscape has provided a framework for classifying AML patients into 11 subtypes of AML with variable prognostic significance284. As per the European LeukemiaNet (ELN) guidelines, a subset of these mutations have been integrated with cytogenetic abnormalities to stratify AML patients into risk groups with favorable, intermediate and adverse prognostic outcomes293.

1.7.4 Therapy of AML

The therapeutic modalities involved in AML treatment comprise chemotherapy/targeted therapy, HSC transplantation (HSCT) and supportive care (transfusions and antimicrobial therapy) which are delivered to patients in a phasic manner279, 282, 285. The phases of AML therapy include: induction of remission, consolidation (post-remission) and maintenance therapies. The aim of induction is to achieve CR preferably without residual disease using chemotherapy of appropriate intensity. This is followed by consolidation chemotherapy with the aim of eliminating any residual disease and achieving cure. Maintenance therapy is a low-intensity non- myelosuppressive treatment indicated in selected patients for longer periods with the aim of sustaining CR and preventing relapse294. The therapeutic options in these phases rely on the functional status and comorbidities of patients which are mostly dependent on their age. Thus, the regimens currently employed vary among young adults and older patients (> 60 years old)279, 282, 285. While numerous agents have been used in AML therapy (Table 1.10), the standard regimen (termed “7+3” regimen) typically involves induction of remission with 7 days of cytarabine in combination with 3 days of an anthracycline (daunorubicin, idarubicin, or mitoxantrone), followed by consolidation with high-dose cytarabine (HiDAC). A third agent may be added to the standard induction regimen in certain settings including fludarabine, etoposide, topotecan, thioguanine, vorinostat, clofarabine, midostaurin, and gemtuzumab ozogamicin295. The therapeutic arsenal of AML has not changed significantly over decades until recently as 8 new agents have been approved in 2 years (Table 1.11).

Table 1.11 Currently used and investigational agents in AML therapy279, 282, 285 Currently used agents Investigational agents Class Drugs Class Drugs Antimetabolites Cytarabine Hypomethylating Guadecitabine agents Fludarabine DOT1L inhibitor Pinometostat Clofarabine Bromodomain Birabresib Mercaptopurine inhibitors Molibresib Thioguanine LSD1 inhibitors GSK2879552 Methotrexate HDAC inhibitors Vorinostat Hydroxyurea Panobinostat Anthracyclines Daunorubicin Pracinostat Idarubicin Kinase inhibitors Sunitinib Mitoxantrone Sorafenib Hypomethylating Azacitidine Quizartinib agents Decitabine Crenolanib Topoisomerase Etoposide Dasatinib inhibitors Toptecan Trametinib Alkylating agents Cyclophosphamide Cobimetinib G-CSF Filgrastim Selumetenib Newly approved Table 1.12 Ruxolitinib agents MDM2 inhibitor Idasanutlin PLK inhibitor Volasertib Rigosertib Aurora kinase Barasertib inhibitors Alisertib CDK inhibitors Alvocidib Palbociclib PI3K inhibitors Buparlisib Hedgehog pathway Vismodegib inhibitors XPO1 inhibitors Selinexor Immune checkpoint Ipilimumab blockade Durvalumab Nivolumab Antibody-drug Vadastuximab conjugates talirine Bispecific antibodies AMG 330

G-CSF: Granulocyte colony-stimulating factor

Table 1.12 New drug approvals in AML296, 297 Drug Mechanism Indication Approval date Company Ref. Midostaurin Multitarget tyrosine kinase Newly diagnosed FLT3- Apr 2017 Novartis 298 (Rydapt®) inhibitor with activity mutated AML (plus standard against wild-type and mutant chemotherapy) FLT3

Enasidenib Selective inhibitor of mutant Relapsed/refractory IDH2 Aug 2017 Celgene 299, 300 (Idhifa®) IDH2 enzyme variants (R140Q, mutated AML R172S, and R172K)

CPX-351 Liposomal fixed-dose Newly diagnosed t-AML or Aug 2017 Jazz 301, 302 (Vyxeos®) combination of Daunorubicin and AML-MRC Ara-C (1:5)

Gemtuzumab Anti-CD33 monoclonal antibody Newly diagnosed and Sep 2017 Pfizer 303, 304 ozogamicin conjugated to the cytotoxic agent relapsed/refractory CD33- (Mylotarg®) calicheamicin positive AML Ivosidenib Selective inhibitor of mutant Newly diagnosed AML July 2018 Agios 305 (Tibsovo®) IDH1 patients ≥ 75 year old; relapsed/ refractory AML Venetoclax Bcl-2 inhibitor Newly diagnosed AML (plus Nov 2018 AbbVie 306, 307 (Venclexta®) azacitidine or decitabine or low-dose Ara-C)

Gilteritinib Multitarget tyrosine kinase Relapsed/ refractory FLT3- Nov 2018 Astellas 308, 309 (Xospata®) inhibitor with activity against mutated AML Pharma mutant FLT3 Glasdegib Hedgehog pathway inhibitor Newly diagnosed AML Nov 2018 Pfizer 310, 311 (Daurismo®) patients ≥ 75 year old (plus low-dose cytarabine) t-AML: therapy-related AML; AML-MRC: AML with myelodysplasia-related change

1.8 Rationale for targeting UBA1 in AML

The first clinically approved drug that targets the UPS is the proteasome inhibitor (PI) bortezomib which was approved in 2003 (Ref.36). The proteasome is important for nuclear and cytosolic protein degradation as well as the proteolysis of misfolded proteins in the ER312. Proteasomal inhibition leads to antitumor activity by multiple mechanisms including NF-κB suppression, NOXA activation, anti-angiogenic effects, activation of p38-MAPK pathway, generation of ROS and induction of ER stress313-315. The differential sensitivity of certain malignancies to bortezomib relies on their dependency on these signaling pathways for survival316. In addition, it correlates with reliance on protein turnover and proteasome workload rather than expression317. Therefore, hematologic malignancies arising from cells engaged in immunoglobulin production (e.g. MM) and antigen presentation via MHC class I (e.g. acute leukemias) are especially vulnerable to PIs318. In this context, bortezomib demonstrated preclinical antitumor activity in many cell lines and animal models of several malignancies; however, it was clinically approved only for the treatment of MM and MCL319-321. These clinical successes in hematologic malignancies reveal the susceptibility of these tumors to proteasomal inhibition37. Given the dependency of MM and other hematologic malignancies on the ER- associated protein quality control and turnover, bortezomib was reported to exert its cytotoxic effects in these malignancies partly via induction of ER stress and terminal UPR206, 322-324. Proteasomal inhibition is theoretically a targeted therapy; however, the proteasome is essential cellular machinery with broad cellular functions in normal cells. Despite this essentiality and the expected toxicity of proteasome inhibition, bortezomib was found to have a clinically acceptable safety profile with adverse effects such as peripheral neuropathy and fatigue that can be managed by dose optimization37, 325, 326. AML has a number of vulnerabilities that make them susceptible to PIs including reliance on NF-κB signaling for survival and intensive engagement in antigen presentation12. NF-κB signaling was reported to be aberrantly activated in a large fraction of AML patients327-329. Proteasomal inhibition leads to NF-κB inhibition by preventing degradation of its inhibitor, IκB, together with impaired antigen presentation leading to accumulation of ubiquitylated proteins and proteotoxic stress. In addition, immunoproteasome expression is markedly increased in AML and higher immuno-to-constitutive proteasome ratios correlated with sensitivity to PIs330. PIs have shown preclinical anti-leukemic effects alone and in combination with other

87 88 therapeutics by multiple mechanisms including modulation of ER stress and UPR331-334. Moreover, they have been tested in a number of clinical trials of AML335-342. In this respect, single-agent bortezomib therapy was reported to have evidence of biological activity, but with limited or no clinical benefit337, 338. Although bortezomib showed encouraging results in a subset of AML patients in combinatorial regimens with standard therapies, it failed to induce prolonged remissions and its contribution to rates of response seen with the combination therapy is uncertain. Of note, the therapeutic index of bortezomib was reported to be narrow (1.3-1.5 mg/m2) which reflects the important biological roles of proteasomes in normal cells209. Another major hurdle to the efficacy of bortezomib is the development of resistance. To overcome bortezomib resistance, a number of strategies were proposed including targeting enzymes upstream of the proteasome (E1, E2 and E3 enzymes). As highlighted before, drug discovery within the UPS commenced with PIs and E3 ligase inhibitors, with the PI program leading to clinical successes with bortezomib12. Therefore, it is tempting to explore other signaling nodes especially the UBA1 enzyme that lies at the apex of ubiquitylation cascade and affects proteasomal degradation as well as other non-degradative signaling processes in the cell. Although inhibiting UBA1 was not initially considered due to expected toxicity, a number of inhibitors were reported to be tolerated in animal models including PYZD-4409 discovered by our laboratory27. Given the essentiality of UBA1 in normal cells, UBA1 inhibition was tolerated in mice either due to either incomplete UBA1 inhibition or the capacity of UBA6, a less- characterized ubiquitin-activating enzyme, to compensate for UBA1 inhibition. Since deletion of UBA1 was reported to cause embryonic lethality in C. elegans and yeast38, it appears that UBA1- dependent ubiquitylation is required for biological processes that are distinct from those mediated by UBA6. Therefore, tolerability of UBA1 inhibition in mice is more likely due to incomplete inhibition rather than compensatory effects of UBA6. Nonetheless, it remains important to confirm this hypothesis when investigating response to drugs targeting UBA1. Therefore, several factors provide a rationale for targeting UBA1 in AML. These include: 1) dependency of leukemic cells on the UPS due to active engagement in antigen presentation and constitutive activity of NF-κB signaling, 2) data suggesting efficacy of UPS-targeting therapeutics such as PIs and NAE inhibitors (affecting a subset of ubiquitin E3 ligases) in preclinical models of AML, and 3) the anticipated pleiotropic effects of UBA1 inhibition that may overcome some forms of clinical resistance to existing therapeutics. Another compelling factor is the data reported by our laboratory that support efficacy and tolerability of UBA1

inhibition in preclinical models of AML27. Specifically, our laboratory showed that primary AML cells and leukemia cell lines (K562, OCI-AML2 and Jurkat) exhibit higher protein ubiquitylation levels compared to normal hematopoietic cells27. Furthermore, genetic knockdown and chemical inhibition of UBA1 by PYZD-4409 induced cell death in AML cells, preferentially inhibited clonogenic growth of primary AML cells, and delayed tumor growth in a mouse xenograft model. Recently, we established collaboration with Takeda Pharmaceuticals and have obtained their first-in-class UBA1 inhibitor, TAK-243. In this thesis, we evaluated the preclinical efficacy of TAK-243 in cell-based and mouse models of AML. In addition, we identified the determinants of sensitivity/resistance to this drug.

Chapter 2

Preclinical evaluation of TAK-243 in AML

Chapter 2 Preclinical evaluation of TAK-243 in AML

2.1 Introduction

Despite improvements in survival rates through the use of targeted agents such as midostaurin and enasidenib298, 343, relapse rates remain high and more than 50% of young adults and 90% of older patients still die from AML280. In addition, most patients do not have genetic mutations that are directly druggable279, 282. Therefore, it is important to identify novel therapeutic strategies for patients with AML. These new therapies must be active in patients with high-risk molecular and cytogenetic mutations and capable of targeting the leukemic stem cells that are frequently responsible for disease relapse344. One such therapeutic approach is to target the ubiquitin-proteasome system (UPS) and particularly the ubiquitin-like modifier activating enzyme 1 (UBA1)36. The UPS is the major cellular machinery responsible for the regulated degradation of cellular proteins and for maintaining cellular homeostasis1, 345. The process of protein degradation via the UPS consists of two major steps: ubiquitylation and proteasomal degradation. Ubiquitylation is a post- translational modification that involves a reversible enzymatic attachment of an 8-kDa, 76- amino-acid, evolutionarily conserved polypeptide called ubiquitin to substrate proteins. Ubiquitylation is mediated by the sequential action of members of three enzyme classes: ubiquitin-activating enzymes (E1), ubiquitin-conjugating enzymes (E2) and ubiquitin ligases (E3). UBA1 is the major ubiquitin-activating E1 enzyme and initiates the ubiquitylation cascade10, 346. UBA1 adenylates a free ubiquitin molecule to form a ubiquitin-adenylate complex bound to its active adenylation domain8. This ubiquitin is then transferred to the catalytic cysteine domain followed by adenylation of another ubiquitin molecule to produce UBA1 loaded with two ubiquitin molecules at two different sites. The loaded UBA1 then interacts with the cognate E2 enzyme to transfer ubiquitin from the catalytic cysteine domain to the active cysteine domain of E2. In concert with E3 enzymes, E2s then transfer ubiquitin to the corresponding protein substrates8. The fate of ubiquitylated proteins depends on the pattern of ubiquitylation. Proteins tagged with a lysine-48 (K48)-linked polyubiquitin chain are recognized and degraded by the proteasome. Through proteasomal degradation, the cell maintains protein homeostasis and degrades excess and misfolded proteins that may adversely affect its survival347. Ubiquitylation also affects

91 92 cellular processes beyond proteasomal protein degradation. For example, ubiquitylation patterns including K63-linked polyubiquitylation and monoubiquitylation constitute non-degradative post-translational modifications that regulate cell signaling348, DNA repair39, cell cycle control9 and endocytosis349. Thus, ubiquitylation affects a broad range of non-degradative cellular processes. Previously, we measured levels of ubiquitylated proteins in lysates from leukemia cell lines and primary AML patient samples as well as normal hematopoietic cells27. We showed that levels of UBA1 protein did not differ between AML and normal cells. However, UBA1 was more actively utilized in AML cells and may be a potential therapeutic target for patients with this disease. TAK-243 is a first-in-class, small-molecule UBA1 inhibitor that bears structural features similar to adenosine sulphamate (Fig. 2.1A) which acts as a mechanism-based inhibitor of E1 enzymes1, 24, 42, 70, 263, 278. Here, we evaluated the preclinical efficacy and biological activity of TAK-243 in AML. In addition, we identified two acquired genetic mutations in the adenylation domain of UBA1 which confer resistance to TAK-243.

2.2 Methods

2.2.1 Compounds and reagents

TAK-243 (MLN7243) was provided by Takeda Pharmaceuticals. Pevonedistat was provided by Dalton Medicinal Chemistry (Toronto, Canada), Bortezomib and Nelfinavir mesylate were purchased from Selleckchem (Catalog# S1013 and S4282, respectively), Daunorubicin from Sequoia (Catalog# SRP01005d), Mitoxantrone from Sigma-Aldrich (Catalog# M6545), GSK2606414 from Tocris (Catalog# 5107), and 4µ8C from Calbiochem (Catalog# 412512).

2.2.2 Cell lines

OCI-AML2 and K562 cells were cultured in Iscove’s modified Dulbecco’s medium (IMDM), and NB4 and U937 cells in Roswell Park Memorial Institute (RPMI) medium. Both media were supplemented with 10% fetal bovine serum (FBS) and appropriate antibiotics. Mouse 32D cells were cultured in RPMI medium supplemented with 10% FBS and murine IL-3 from conditioned medium of the mouse X63 cell line (provided as gift from Dr. N. Iscove, University Health Network). TEX cells were cultured in IMDM medium supplemented with 15% FBS, 2 mM L- glutamine, 20 ng/ml recombinant human (rh) stem cell factor (SCF), 2 ng/ml rh-IL-3 and

93 appropriate antibiotics350. Cells were not tested for Mycoplasma contamination. U937 and K562 cells were authenticated by STR profiling.

2.2.3 Primary AML and normal hematopoietic cells

Primary human AML cells were obtained from peripheral blood or bone marrow samples isolated from consenting AML patients, with blast count of at least 80% among low-density cells, using Ficoll density gradient centrifugation. Primary normal hematopoietic cells were obtained from healthy consenting volunteers who donated peripheral blood stem cells for allogeneic stem cell transplantation following mobilization by granulocyte-colony stimulating factor (G-CSF). All primary cells were cultured in MyeloCult™ H5100 media (Stem Cell Technologies) supplemented with rh-IL7 (20 ng/ml), rh-IL6 (20 ng/ml), rh-FLT3-L (10 ng/ml), rh-GM-CSF (20 ng/ml), rh-SCF (100 ng/ml), rh-IL3 (10 ng/ml) and rh-G-CSF (20 ng/ml; Miltenyi). The collection and use of human samples was done based on informed consent from all subjects and approval of the University Health Network Institutional Review Board.

2.2.4 Cytotoxicity assays

CellTiter 96® AQueous MTS Reagent Powder was purchased from Promega (Catalog# G1111), Annexin V-FITC apoptosis kit from Biovision (Catalog# K101-400), and CellTiter-Fluor kit from Promega (Catalog# G6080). The MTS, Annexin V/propidium iodide (PI), and CellTiter- Fluor assays were conducted as per the manufacturer’s guidelines. Flow cytometric analysis of Annexin V/PI-treated samples was performed using BD FACSCANTO flow cytometer (BD Biosciences). Colony formation was assessed by the clonogenic assay as described previously351. Briefly, primary AML cells (4x105 cells) were treated with DMSO or TAK-243 for 48 hours, followed by washing and plating by volume in duplicate in 35 mm dishes (Nunclon, Rochester, USA) to a final volume of 1 mL/dish in MethoCult GF H4434 medium (Stem Cell Technologies, ° Vancouver, Canada). Cells were then incubated at 37 C, 5% CO2 with 95% humidity to form colonies. Colonies of at least 10 cells were then counted as previously described351.

2.2.5 Cellular thermal shift assay (CETSA)

CETSA was conducted as previously described352. Briefly, OCI-AML2 cells were treated with either DMSO or increasing concentrations of TAK-243 for 30 min. Cells were then washed and

94 re-suspended in phosphate-buffered saline (PBS) containing protease inhibitors (Thermo Fisher Scientific). Cells were then heated to 54°C for 3 min using a thermal cycler to assess the stability of UBA1 and UBA3 and 52°C to assess the stability of UBA6. These temperatures were experimentally derived to produce the optimal thermal shift of the proteins. To assess thermal stability on tumors harvested from TAK-243-treated mice, samples were cut into small pieces and re-suspended in PBS with protease inhibitors followed by heating for 3 min. Cell lysates were then prepared by 6 freeze-thaw cycles with vigorous vortexing in between, followed by centrifugation twice at 20,000 g for 20 min.

2.2.6 Immunoblotting

Whole cell lysates were prepared as previously described353. Briefly, cells were washed with phosphate-buffered saline (PBS; pH=7.4) and lysed with RIPA buffer followed by sonication and centrifugation at 13,000 RPM for 20 min at 4°C. Supernatants were then collected followed by protein quantification using Bradford assay (Bio Rad, Hercules, CA), and denaturation by heating at 95°C for 5 min. For CETSA assays, lysates were heated at 70°C for 10 min. Equal amounts of proteins were fractionated by 10% gels using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), followed by transfer to PVDF membranes. Tested proteins were then probed using appropriate primary and secondary antibodies (Table 2.1).

2.2.7 Overexpression of GRP78 pcDNA3.1(+)-GRP78 plasmid was obtained as a gift from Richard C. Austin (Addgene plasmid #32701)354. pcDNA control and pcDNA-GRP78 plasmids were transfected into K562 cells using Effectene transfection kit as per the manufacturer’s guidelines (Qiagen, catalog#301425). The transfected cells were then selected with G418 (800 µg/mL) for 3 weeks.

2.2.8 Immunofluorescence OCI-AML2 cells were used for immunofluorescence studies. Cells were subjected to irradiation with 3 Gy or left unirradiated. For indicated conditions, cytospins were prepared using Cytospin™ 4 Cytocentrifuge (Thermo Scientific). Cells were fixed using 4% paraformaldehyde for 15 minutes, blocked in dilution buffer (5% BSA, 0.1% Coldwater fish skin gelatin, 5% FBS, 0.1% Triton X-100), and incubated in anti-H2AXSer139 and anti-53BP1 overnight. Anti-mouse IgG conjugated with Alexa Fluor 488 (Invitrogen) and anti-rabbit conjugated with Alexa Fluor

594 (Invitrogen) were used as secondary antibodies for γH2AX and 53BP1 staining, respectively (Table 2.1). Subsequently, slides were counterstained with DAPI (Invitrogen) and mounted using Mowiol (Sigma-Aldrich). Cells were visualized and quantified for their subnuclear foci using a Leica DM 4000 B microscope with a 100X oil immersion objective. Cells displaying equal to or greater than 10 foci were scored. Image acquisition and overlay were performed using Leica Application Suite V 4.0 software (Leica Microsystems, Switzerland).

2.2.9 Immunohistochemistry (IHC)

Tumors, normal tissues or cells (pellets of 20x106 cells) were fixed in 10% buffered formalin overnight followed by 70% ethanol. Paraffin cell block sections at 4µm thickness were dried in 60°C oven for 1h before staining. IHC was performed as per the manufacturer’s guidelines using BenchMark XT-an automated slide stainer (Ventana Medical Systems) with standard antigen retrieval (CC1, Tris/Borate/EDTA pH 8.0, #950-124). The FK2, Ub-H2A and Ub-H2B antibodies were incubated for 60 minutes (Table 2.1). Ultraview Universal DAB Detection Kit (Ventana Medical Systems #760-500) was utilized. The kit contains a cocktail of enzyme-labeled secondary antibodies that locate the bound primary antibody. The complex was then visualized with hydrogen peroxide substrate and 3,3’-diaminobenzidine tetrahydrochloride (DAB) chromogen, which produces a dark brown precipitate readily detected by light microscope. The slides were counterstained with Gill Hematoxylin and Bluing in PBS, dehydrated in graded alcohol, cleared in xylene and cover-slipped in PermountTM. For primary AML samples, cytospins of 75,000-100,000 cells were prepared using Cytospin™ 4 Cytocentrifuge (Thermo Scientific). Cytospin slides were then fixed in 2% formaldehyde for 10 min. IHC was done as described above but without antigen retrieval.

Table 2.1 Antibodies used in immunoblotting, IHC and immunofluorescence, their sources and dilution. Primary antibody Source Clone Catalog# Dilution 53BP1 Bethyl NA A300-272 IF: Primary: 1:1000 Secondary: 1:2000 ABCG2 Cell Signaling NA 4477 Primary: 1:1000 Secondary: 1:1000 ATF4 Santa Cruz C-20 sc-200 Primary: 1:100 Secondary: 1:1000 ATF6α Santa Cruz F-7 sc-166659 Primary: 1:100 Secondary: 1:1000 CHOP Sigma NA G6916 Primary: 1:500 Secondary: 1:1000 Cleaved Cell Signaling NA 9661 IHC: Caspase 3 Primary: 1:500 Secondary: 1:200 GAPDH Cell Signaling 14C10 2118 Primary: 1:4000 Secondary: 1:2000 GRP78 Sigma GL-19 G8918 Primary: 1:5000 Secondary: 1:3000 IRE1α Cell Signaling 14C10 #3294 Primary: 1:1000 Secondary: 1:1000 Mono- and poly- Enzo Life FK2 BML- IHC: ubiquitinylated Sciences PW8810-0100 Primary: 1 :100 conjugates Secondary: Ultraview detection kit NEDD8 Cell Signaling NA 2745 Primary: 1: 200 Secondary: 1:1000 PARP Cell Signaling NA 9542 Primary: 1:1000 Secondary: 1:1000 p-IRE1α Ser724 Abcam NA ab48187 Primary: 1:1000 Secondary: 1:1000 p-JNKThr183/Tyr185 Cell Signaling NA 9251 Primary: 1:1000 Secondary: 1:1000 p-PERKThr981 Santa Cruz NA sc-32577 Primary: 1:100 Secondary: 1:1000 UBA1 Santa Cruz 2G2 sc-53555 Primary: 1:300 Secondary: 1:1000 UBA2 Santa Cruz 28 sc-136359 Primary: 1:500 Secondary: 1:1000 UBA3 Santa Cruz E-5 sc-377272 Primary: 1:400 Secondary: 1:1000 UBA6 Cell Signaling NA 13386 Primary: 1:1000 Secondary: 1:1000 Ubiquitin Cell Signaling NA 3933 Primary: 1: 500 Secondary: 1:1000

Ubiquityl-Histone Cell Signaling D27C4 8240 Western: H2A (Lys119) Primary : 1:1000 Secondary: 1:1000 IHC: Primary: 1:400 Secondary: Ultraview detection kit Ubiquityl-Histone Cell Signaling D11 5546 Western: H2B (Lys120) Primary: 1:500 Secondary: 1:1000 IHC: Primary: 1:400 Secondary: Ultraview detection kit XBP1 Santa Cruz M-186 sc-7160 Primary: 1:200 Secondary: 1:1000 β-actin Santa Cruz AC-15 sc-69879 Primary: 1:10,000 Secondary: 1:5,000 γH2AXSer139 EMD JBW301 05-636 IF: Millipore Primary: 1:600 Secondary: 1:1000

2.2.10 Animal studies

OCI-AML2 cells (1x106) were injected subcutaneously (sc) into the flanks of male SCID mice (Ontario Cancer Institute, Toronto, Canada). When the tumors became palpable, mice were randomly divided into 2 groups (n=10 per group) and treated with vehicle (10 % 2- hydroxypropyl-β-cyclodextrin [HPBCD] in water) or TAK-243 (20 mg/kg) sc twice weekly (BIW) for 3 weeks without blinding. Tumor volume was assessed by caliper measurements every 2-3 days using the following equation355: tumor volume (mm3) = tumor length (mm) × width2 (mm)× 0.5236. At the end of the experiment, serum was collected from mice by cardiac puncture for assessment of serum chemistries. Mice were then euthanized, tumors excised and weighed, and organs were harvested and processed for histopathology. Tumors were fixed with 10% buffered formalin, embedded, sectioned, and stained with hematoxylin and eosin (H&E). The stained samples were then scanned using Aperio Scanscope XT and analyzed using Aperio ImageScope. To assess TAK-243 effect on primary human AML samples in a bone marrow transplant model, primary AML samples were thawed, washed and re-suspended in PBS. 2.5x106 viable trypan blue-negative cells were injected into the right femurs of 10-week-old female NOD-SCID mice which have been sub-lethally irradiated with 208 rad from an X-RAD 320 system (PXi Precision X-ray) and pre-conditioned with 200 mg anti-mouse CD122. Two weeks post-injection, mice were randomly divided into 2 groups and treated with either 10% HPBCD or TAK-243 (20 mg/kg) sc BIW for 4 weeks. At the end of the experiment, mice were sacrificed and the bone marrow flushed from the femurs. The percentage of AML cells present in the non-injected left femora was measured by flow cytometry using anti-human CD45, anti-human CD33 and anti- human CD19 antibodies in a BD Biosciences Canto flow cytometer. To assess secondary engraftment, primary human AML cells isolated from the bone marrow of vehicle- and TAK- 243-treated mice were pooled, and equal numbers of viable cells were injected into the right femur of untreated mice. Five weeks post-transplantation, mice were sacrificed, and secondary transplantation was assessed as described above. All animal studies were carried out according to the regulations of the Canadian Council on Animal Care and with the approval of the local ethics review board.

2.2.11 Generation of TAK-243-resistant AML cell lines

TAK-243-resistant cells were generated by two approaches: continuous IC90 and stepwise cyclic treatment. In the continuous IC90 approach, OCI-AML2 cells were cultured continuously in media containing TAK-243 starting at a concentration corresponding to its IC90 (30 nM). Over 3 months, the concentration was doubled to 60, 120 and then to 250 nM. Control cells treated with DMSO were maintained in parallel for the same duration. The resistant cell line was designated

OCI-AML2-RIC90. In the stepwise approach, the resistant OCI-AML2 cells were generated in a stepwise cyclic manner by maintaining the cells in media containing TAK-243 at a concentration of 0.5 nM for 3 days followed by culture in drug-free media for 3 days. The concentration was then doubled until the cells could normally grow in a concentration of 250 nM. The resistant cell line was designated OCI-AML2-RSW. Both resistant cell lines were periodically cultured in TAK-243 to maintain selection of the resistant population. UBA1 is on the X chromosome, and by FISH analysis, over 90% of the cells had two X-chromosomes.

2.2.12 Detection of UBA1 mutations in TAK-243-resistant AML cell line

Genomic DNA was isolated from OCI-AML2 cells using Gentra Puregene Cell Kit (Qiagen, catalog#158745) as per the manufacturer’s guidelines. Briefly, 107 cells were washed, lysed and protein was precipitated. Genomic DNA was then precipitated with isopropanol followed by washing with 70% ethanol. DNA was then dissolved in hydration solution and quantified. Exons 12-16 and 23-24, spanning the UBA1 adenylation domain, were then amplified by PCR using appropriate primers and conditions (Table 2.2 and 2.3). PCR amplicons were then purified using a PCR purification kit (Qiagen, catalog# 28104), and sequenced by Sanger sequencing.

2.2.13 Statistical and data analysis

All statistical analyses were performed with GraphPad Prism software version 6.03 (GraphPad Software Inc.) Statistical significance of differences between means was calculated by unpaired t-test (2 groups), one-way ANOVA and appropriate multiple comparisons test (˃ 2 groups), and two-way repeated measures ANOVA (for the tumor growth rate in mice). Median inhibitory concentration (IC50) and IC90 values were calculated using the nonlinear regression function in GraphPad Prism. All experiments were performed in triplicate with at least 3 biological replicates. The structure of TAK-243 was drawn with ChemDraw Professional 16.0 Suite (PerkinElmer Informatics). Flow cytometry data were analyzed with BD FACSDiva Software

100

6.0 (BD Biosciences) and FlowJo version 7.7.1 (FlowJo, LLC). ImageJ Software version 1.48v (National Institutes of Health, USA) was used for calculation of percent staining of IHC images. DNA sequences were viewed with Chromas software 2.6.4. Image of TAK-243 interaction with mutant E1 was generated using PYMOL software (DeLano Scientific).

101

Table 2.2 Primers used for amplifying UBA1 exons UBA1 Exon F/R Primer pair (5’→3’) Conditions 12/13 F TCTGTCACTTGCTCTCTGTCTG AT = 56°C R GGAAACGAGGGATGCACTGG CN = 35 X 14 F GCAACCTGAAGATGTGGTGTTAGCC AT = 54.1°C R AAGGAGGAAAGTGAGACGACAAAGCC CN = 40 X 15 F GTGTGATGGAGAGAGACCCTG AT = 56.7°C R AAACATCAGGAGAGGACAGAAAG CN = 40 X 16 F CGGGCAAAGTTGTGTGTGTT AT = 54.1°C R CAACATCTGAAGGCAAGGCA CN = 35 X 23 F TTAGCCCCTCTGTAGACCCT AT = 54.1°C R GGTGACAGGGTATGACGTGA CN = 40 X 24 F GGTTGTGATCTGACTAAACACGT AT = 54.1°C R GGTCATGGAGGTGGACAGG CN = 40 X AT: annealing temperature; CN: number of cycles

Table 2.3 Conditions used for amplifying UBA1 exons Stage Temperature Duration Notes

A. Initial denaturation 95°C 5 min B. Cycles 1. Cyclic denaturation 95°C 60 Sec Number of cycles 2. Annealing Shown in the table 60 Sec shown in the table 3. Extension 72°C 42 Sec above C. Final extension 72°C 10 min D. Hold 4°C Indefinitely

2.3 Results

2.3.1 TAK-243 induces cell death and decreases clonogenic growth in AML cell lines and primary cells

TAK-243 is a selective UBA1 inhibitor278. To determine the effects of TAK-243 on the growth and viability of AML cells, we treated OCI-AML2, TEX, U937, NB4 and 32D cell lines with increasing concentrations of TAK-243. Cell growth and viability were measured with the MTS assay 16 and 48 h after incubation. TAK-243 decreased the growth and viability of AML cell lines in a concentration- and time-dependent manner with a median inhibitory concentration

(IC50) values ranging from 20.4 to 83.5 nM after 16 h and 15.1 to 41.5 nM after 48 h (Fig. 2.1B and 2.2A). We confirmed cell death in OCI-AML2 and TEX cells by Annexin V/PI staining and flow cytometry (Fig. 2.1C and 2.2B). We also assessed the effects of TAK-243 in primary AML cells. Primary samples (n=19) were treated with increasing concentrations of TAK-243 for 48 h, and viability measured by Annexin

V/PI staining. IC50 values ranged from 30-480 nM with a median IC50 of 74 nM (Figure 2.1D). An additional cohort of 7 primary samples was treated with TAK-243 for 72 hours and growth and viability measured by CellTiter Fluor assay. IC50 values ranged from 13.8- 126.2 nM with a median IC50 of 31.4 nM (Fig. 2.1E). Sensitivity was not related to morphologic subtype or cytogenetic/molecular mutations as TAK- 243 induced cell death in primary AML samples with high-risk cytogenetics, FLT3 mutations, and samples from patients refractory to induction chemotherapy (Table 2.4). Next, we assessed the effects of TAK-243 on clonogenic growth of AML and normal hematopoietic progenitor cells. Primary AML cells (n=6) and normal hematopoietic cells from consenting individuals donating G-CSF-mobilized peripheral blood stem cells for allotransplants (n=6) were treated with vehicle or TAK-243 for 48 h and then plated into colony-forming assays. TAK-243 nearly abolished the clonogenic growth of all tested primary AML samples, but normal hematopoietic samples were more resistant (Fig. 2.1F).

2.3.2 TAK-243 preferentially binds to UBA1 in AML cell lines and primary cells

We assessed the binding of TAK-243 to UBA1 in intact cells using CETSA. It is based on the principle of thermal stabilization of proteins upon binding chemical ligands352.

102 103

104

Figure 2.1 Anti-leukemic activity of TAK-243 in AML cell lines and primary cells. A) Chemical structure of TAK-243. B) OCI-AML2, TEX, U937, NB4 and 32D cell lines were treated with increasing concentrations of TAK-243 for 48 h. Cell growth and viability was measured by MTS assay. Insert: the IC50 and 95% confidence interval (CI) values (nM) are shown. C) OCI-AML2 and TEX AML cells were treated with increasing concentrations of TAK- 243 for 48 h. Cell viability was measured by Annexin V/PI staining and flow cytometry. All data points in B and C represent the mean ± SEM of at least 3 independent experiments. D and E) Primary AML samples were treated with increasing concentrations of TAK-243 for 48 h (D) and 72 h (E). Cell viability was measured by Annexin V/PI staining (n=19) (D) and cell growth and viability were measured by CellTiter-Fluor (n=7) (E). The IC50 values (nM) were calculated from 3 technical replicates for each sample and represented on the Y-axis with the corresponding sample numbers on the X-axis. F) Primary AML (n=6) and normal hematopoietic samples (n=6) were treated for 48 h with TAK-243 (250 nM) and then plated into colony-forming assays. Growth of leukemic colonies was measured in the AML cells and burst-forming unit-erythroid (BFU-E) and colony-forming unit-granulocyte, monocyte (CFU-GM) was assessed in the normal hematopoietic cells. The Y-axis shows the number of colonies as a percentage of the vehicle- treated controls from individual samples with 2-4 technical replicates per sample. Each data point represents one primary sample and the horizontal lines represent the means of 6 samples. **p < 0.01; ****p < 0.0001 using one-way ANOVA and Tukey's multiple comparisons test.

105

Figure 2.2 Anti-leukemic activity of TAK-243 in AML cell lines. A) OCI-AML2, TEX, U937, NB4 and 32D cells were treated with increasing concentrations of TAK-243 for 16 h. Cell growth and viability was measured by MTS assay. Insert: the IC50 and 95% confidence interval values are shown. B) OCI-AML2 and TEX AML cells were treated with increasing concentrations of TAK-243 for 16 h. Cell viability was measured by Annexin V/PI staining and flow cytometry. All data points in A and B represent the mean ± SEM of at least 3 independent experiments.

Table 2.4 Characteristics of patient samples used in the study

Age at Status of IC50 IC50 Patient # Disease Gender Cytogenetics Molecular Assay collection sample (48h) (72 h)

AML, FLT3-ITD neg, Annexin 8243 34 F inconclusive Diagnostic 104 NA M4 FLT-3-TKD neg V/PI

46,XX,del(11)(q14)[3]/46,ide AML, FLT3-ITD neg, Annexin 8314 38 F m,+1,dic(1;9)(p13;q34)[17].is Diagnostic 218 NA M0 FLT-3-TKD neg V/PI h del(11)(q23q23)(MLL-)

AML, NPM1 pos, FLT3- Annexin 160757 61 F 46,XX[20] Diagnostic 89 NA M1 ITD neg V/PI

AML, FLT3-ITD neg, Annexin 8812 38 M 46,XY[20] Diagnostic 74 NA M2 FLT-3-TKD neg V/PI

nuc Annexin 160818 AML 79 F 46,XX [29] ish(D13S319,D13S Diagnostic 112 NA V/PI 25,LAMP1)x2[200]

43~44,XX,add(2)(q33)[2],del( 3)(p21)[2],- 5[8],del(5)(q13q33)[2],add(6)( q21)[2],add(7)(q32)[6], AML- Annexin 160853 76 F del(12)(p11.2p13)[2],del(12)( None performed Diagnostic >480 NA MRC V/PI q13q24.1)[6],- 17[3],dic(17;21)(p13;p11.2)[5 ],-18[10], +1~3mar[8],dmin[2] [cp10]

AML, FLT3-ITD neg, Annexin 9784 37 M 46,XY[20] Diagnostic 406 NA M1 FLT-3-TKD neg V/PI

AML nuc with 47,XY,+X,t(9;11)(p22;q23[10 ish(5’KMT2A,3’K Annexin 161820 30 M Diagnostic 30 NA t(9;11)(p ] MT2A)x2(5’KMT2 V/PI 22;q23) A sep

106 107

3’KMT2Ax1)[190/ 200]

AML- Annexin 130624 79 M inconclusive None performed Diagnostic 73 NA MDC V/PI

AML with Annexin 130354 31 M inv (16) CBFB-MYH11 pos Diagnostic 57 NA CBFB- V/PI MYH11

Annexin 162027 APL 62 F NA PML-RARA pos Diagnostic 82 NA V/PI

AML CBFB-MYH11 with Annexin 162111 28 M 45,X,-Y[9]/46,XY[11] neg, NPM1 pos, Diagnostic 36 NA mutated V/PI FLT3-ITD neg NPM1

AML- NPM1 neg, FLT3- Annexin 162131 64 F 46,XX[20] Diagnostic 48 NA MDC ITD pos V/PI

AML PML/RARA neg, with Annexin 162252 25 F 46,XX[20] NPM1 pos, FLT3- Diagnostic 53 NA mutated V/PI ITD neg NPM1

AML with NPM1 neg, FLT3- Annexin 162335 68 M 46,XY[20] Diagnostic 45 NA monocyti ITD neg V/PI c diff.

46,XY,t(1;12)(q21;p12~13),?a Annexin 130714 AML 83 M None performed Diagnostic 175 NA dd(4)(p16)[3]/47,idem,+9[7] V/PI

AML, Annexin 130607 74 F 46,XX [20] None performed Diagnostic 83 NA M5 V/PI

130537 AML, 52 M 46,XY[20] NPM1 pos, FLT3- Diagnostic Annexin 53 NA

108

M2 ITD neg, FLT3- V/PI TKD neg

NPM1 neg, FLT3- AML- Annexin 130752 68 M 46,XY[20] ITD neg, FLT3- Diagnostic 57 NA MDC V/PI TKD neg

NPM1 neg, FLT3- AML, ITD neg, FLT3- Mouse 130578 62 M 46,XY[20] Diagnostic marrow NA NA M4 TKD neg, engraftment BCR/ABL1 neg

45,XY,der(6;7)t(6;7)(p21;q22) BCR-ABL1 CellTiter- 150256 M2 23 M Diagnostic NA 31.36 del(6)(q13q21)[17]/46,XY[3] negative Fluor™

46,XY,t(3;3)(q21;q26.2),ins(7 CellTiter- 160376 M4 45 M ;12)(p15;q1?1q13),der(12)del( None performed Diagnostic NA 126.2 Fluor™ 12)(p11.2p12)ins(7;12)[13]

CellTiter- 140320 M0 72 M 46,XY[16] None performed Diagnostic NA 31.87 Fluor™

41,XY,-4,- 5,?del(9)(p11p13),inv(13)(q14 CellTiter- 130616 M2 62 M None performed Diagnostic NA 20.1 q32),-16,-17,- Fluor™ 18[10]/46,XY[2]

42~43,XX,add(1)(p34),der(3) add(3)(p21)inv(3)(q21q27),de r(5;14)(p10;q10),-8,-9,- 13,add(14)(p11.2),- CellTiter- 151620 M1 59 F None performed Diagnostic NA 73.48 17,der(18)?t(18;21)(p11.2;p11 Fluor™ .2),-19,- 21,add(21)(q22),add(22)(p11. 2),+3~4mar[cp10]

FLT3-ITD positive; CellTiter- 160121 M5 63 M 46,XY[20] FLT3-TKD Diagnostic NA 13.75 Fluor™ negative; NPM1

109

negative

45,X,-Y,- 9,add(17)(p11.1),+mar1[4]/45, X,-Y,-9,- CellTiter- 160326 M5 67 M None performed Diagnostic NA 14.73 17,+20,+mar1[2]/45,X,- Fluor™ Y,+1,der(1;17)(q10;q10)[2]/4 6,XY[4]

NPM1 pos, FLT3- Mouse 90239 AML 74 M inconclusive ITD pos, FLT3- Diagnostic marrow NA NA TKD neg engraftment

OCI-AML2 and primary AML cells were treated with TAK-243 at increasing concentrations and the thermal shifts of UBA1, UBA2, UBA3, and UBA6 were measured. In OCI-AML2 cells and primary AML samples, TAK-243 bound UBA1 as evidenced by increased thermostabilization of UBA1 protein with increasing concentrations of TAK-243. In contrast, thermostabilization of UBA3, UBA6 or UBA2 was only observed at much higher concentrations (Fig. 2.3A-B and 2.4). As a control, pevonedistat, a selective inhibitor of the related NEDD8-activating enzyme (NAE), preferentially bound the NAE subunit UBA3 over UBA1 as expected (Fig. 2.3C). Thus, TAK-243 preferentially binds to UBA1 in AML cells and primary samples at concentrations corresponding to its IC50 (See Section 1.6.7).

2.3.3 TAK-243 reduces ubiquitylation of cellular proteins in AML cells and primary AML samples

UBA1 is the initiating enzyme in the ubiquitylation cascade, so we tested whether UBA1 inhibition by TAK-243 affects ubiquitylation of cellular proteins. OCI-AML2 and 32D cell lines and primary AML samples (n=3) were treated with increasing concentrations of TAK-243. In a time- and concentration-dependent manner, TAK-243 reduced mono- and polyubiquitylation of global cellular proteins in cell lines and primary samples (Fig. 2.3D-F and 2.5A). In contrast, TAK-243 inhibited neddylation only at much higher concentrations exceeding those required to induce cell death and at a later time point (Fig. 2.3G). Taken together, these data highlight the specificity of TAK-243 in reducing global ubiquitylation of proteins.

2.3.4 TAK-243 induces endoplasmic reticulum (ER) stress in AML cells

Failure to degrade misfolded and damaged proteins may cause ER stress and cell death356, 357. Given the role of UBA1 as the major ubiquitin-activating enzyme, we tested whether inhibition of UBA1 by TAK-243 induces ER stress and an unfolded protein response (UPR). Treatment of AML cells with TAK-243 produced changes consistent with increased ER stress. TAK-243 increased PERKThr981 phosphorylation, and induced expression of CHOP and ATF4 indicating activation of the PERK axis of UPR. It also induced IRE1αSer724 phosphorylation, JNKThr183/Tyr185 phosphorylation and XBP1 splicing to the active form, XBP1s, indicating activation of the IRE1α axis of UPR.

110 111

112

Figure 2.3 TAK-243 preferentially binds to UBA1 and reduces levels of global protein ubiquitylation in AML cell lines and primary cells. A) OCI-AML2 cells were treated with increasing concentrations of TAK-243 for 30 min followed by heating the intact cells at 54°C. After heating, whole cell lysates were prepared and levels of UBA1, UBA3, UBA6, and GAPDH were measured by immunoblotting. B) AML patient sample #161820 was treated with increasing concentrations of TAK-243 for 30 min followed by heating the intact cells at 54°C. After heating, whole cell lysates were prepared and levels of UBA1, UBA3, and GAPDH were measured by immunoblotting. C) AML patient sample #161820 was treated with increasing concentrations of pevonedistat for 30 min followed by heating the intact cells at 54°C. After heating, whole cell lysates were prepared and levels of UBA1, UBA3, and GAPDH were measured by immunoblotting. D) OCI-AML2 cells were treated with increasing concentrations of TAK-243 for increasing times. After treatment, whole cell lysates were prepared and levels of global mono- and poly- ubiquitylated proteins (anti-ubiquitin antibody) and GAPDH were measured by immunoblotting. E) OCI-AML2 cells were treated with DMSO or TAK-243 (100 and 1000nM) for 3 h. After treatment, levels of global mono- and poly-ubiquitylated proteins (FK2 antibody) were measured by immunohistochemistry (IHC). Scale bar= 200 µm. F) Primary AML cells were treated with TAK-243 (100 and 1000nM) for 3 h. After treatment levels of global mono- and poly-ubiquitylated proteins (FK2 antibody) were measured by IHC. Sample numbers are shown in the figure. Scale bar= 200 µm. G) OCI-AML2 cells were treated with increasing concentrations of TAK-243 for increasing times. After treatment, whole cell lysates were prepared and levels of neddylated proteins were measured by immunoblotting.

113

Figure 2.4 TAK-243 preferentially targets UBA1 in AML cell lines and primary cells. A) OCI-AML2 cells were treated with DMSO (UT) or 40 µM TAK-243 (T) for 1h followed by heating the intact cells at temperatures between 50-58°C. After heating, whole cell lysates were prepared by 2-4 freeze-thaw cycles in liquid nitrogen and levels of UBA1, UBA3, UBA6, and GAPDH were assessed by immunoblotting. B) Primary AML sample #130624 was treated with DMSO (UT) or 40 µM TAK-243 (T) for 1h followed by heating the intact cells at temperatures between 50-58°C. After heating, whole cell lysates were prepared by 2-4 freeze-thaw cycles in liquid nitrogen and levels of UBA1, UBA3 and GAPDH were assessed by immunoblotting. C) OCI-AML2 cells were untreated, treated with DMSO or 100 nM TAK-243 for 30 min followed by direct processing (unheated) or heating cells at temperatures 54°C (for UBA1) or 56°C (for UBA2). Whole cell lysates were prepared by 2-4 freeze-thaw cycles in liquid nitrogen and levels of UBA1, UBA2 and GAPDH were assessed by immunoblotting.

114

Figure 2.5 Effect of TAK-243 on global protein ubiquitylation and Bcl-2 levels. A) Mouse 32D cells were treated with either DMSO or TAK-243 (100 and 1000 nM) for 4h. After treatment, whole cell lysates were prepared and levels of global ubiquitylated proteins and GAPDH were measured by immunoblotting. B) OCI-AML2 cells were treated with increasing concentrations of TAK-243 for increasing times. After treatment, whole cell lysates were prepared and levels of Bcl-2 and β-tubulin were measured by immunoblotting.

115

Little or no changes were observed in the ATF6 axis or in GRP78 expression levels. Similarly, no changes were observed in UBA1 or UBA6 expression after treatment with TAK-243. Increased ER stress was related to induction of apoptosis as measured by PARP cleavage (Fig. 2.6A). Of note, no changes were observed in levels of Bcl-2 protein up to 8h post-treatment (Fig. 2.5B). To investigate the functional importance of ER stress in TAK-243-induced cell death, we assessed the cytotoxicity of TAK-243 after treatment with modulators of protein loading and UPR. First, co-treatment of OCI-AML2 cells with cycloheximide (2.5 µg/ml), a protein synthesis inhibitor that reduces ER loading, abrogated cell death induced by TAK-243 as assessed by annexin V/PI staining after 16 h of incubation (Fig. 2.6B). Second, we used genetic and chemical approaches to assess the roles of ER stress signaling molecules in mediating cell death by TAK- 243. Overexpression of GRP78, a repressor of UPR358, reduced TAK-243 cytotoxicity and apoptosis (Fig. 2.6C-D). Likewise, co-treatment of OCI-AML2 cells with the PERK inhibitor GSK2606414 (Ref. 359) increased TAK-243 cytotoxicity, while co-treatment with the IRE1α inhibitor 4µ8C (Ref. 360) rescued TAK-243-mediated cell death (Fig. 2.6E). In contrast, nelfinavir, an inhibitor of ATF6 activation361, 362, did not alter cell death consistent with the minimal activation of ATF6 after TAK-243 treatment (Fig. 2.6A). Taken together, these results show that TAK-243 induces ER stress and activates the PERK and IRE1α arms of UPR, which may, in part, mediate TAK-243 effects.

2.3.5 TAK-243 inhibits the DNA damage response

Apart from proteolytic functions through UPS, ubiquitylation also regulates the DNA damage response39, 363, 364. For example, mono-ubiquitylation of histones H2A and H2B is associated with DNA repair and transcriptional regulation365, 366. Therefore, we examined the effects of TAK-243 on DNA damage repair. Treatment of OCI-AML2 cells and primary AML samples (n=3) with TAK-243 decreased H2A and H2B ubiquitylation (Fig. 2.7A-B and 2.8A-B). To determine the effect of TAK-243 on the repair of DNA double strand breaks (DSBs), we examined OCI-AML2 cells pretreated with TAK-243 or DMSO, for irradiation (IR)-induced foci formation of Ser139 phosphorylated H2AX (γH2AX) and 53BP1, two markers of DSBs364.

116

117

Figure 2.6 TAK-243 induces proteotoxic stress in AML cells. A) OCI-AML2 cells were treated with increasing concentrations of TAK-243 for increasing times. After treatment, whole cell lysates were prepared and levels of UBA1, UBA6, PARP, cleaved PARP (c. PARP) and ER stress related proteins were measured by immunoblotting. GAPDH and β-actin were used as loading controls. B) OCI-AML2 cells were treated with DMSO, Cycloheximide (CHX; 2.5 µg/ml), TAK-243 (50 nM or 100 nM) or a combination of CHX+TAK-243 for 16 h followed by annexin V/PI staining. Bars represent mean ± SEM of 3 independent experiments. C) Left: Whole cell lysates of K562 cells overexpressing GRP78 and their control cells, K562-pcDNA, were prepared and levels of GRP78 were measured by immunoblotting. Right: K562-GRP78 and K562-pcDNA were treated with either DMSO or TAK-243 (100 nM) for 24 h followed by immunoblotting to measure PARP and c. PARP. GAPDH was used as a loading control. D) Wild-type and transfected K562 cells were treated with increasing concentrations of TAK-243 for 16 h followed by the MTS assay to assess growth and viability. Insert: the IC50 values are shown. E) OCI-AML2 cells were treated with TAK-243 alone or in combination with the UPR modulators GSK2606414 (10 µM), 4µ8C (10 µM) or Nelfinavir (10 µM) for 16 h followed by the MTS assay. The relative IC50 compared to TAK-243 alone (considered as 1) are shown. Each bar represents the mean±SEM of the IC50 values calculated from 3 independent experiments.

118

First, we examined γH2AX and 53BP1 foci formation 1h post-IR of OCI-AML2 cells pre-treated with DMSO or TAK-243 for 4h. While all samples exhibited similar γH2AX foci formation, recruitment of downstream effector 53BP1 was impaired in cells pretreated with TAK-243, in a concentration-dependent manner, with 1000 nM abolishing all 53BP1 foci formation (n=3) (Fig. 2.7C-D). Interestingly, OCI-AML2 cells treated with TAK-243 without irradiation showed increased γH2AX foci suggesting the ability of TAK-243 to induce DNA damage stress under unirradiated conditions (Fig. 2.8C-E). However, we cannot exclude the possibility that increased γH2AX foci may be secondary to cell death. Furthermore, OCI-AML2 cells pretreated for 2h with 100 nM of TAK-243 prior to IR, displayed reduced ability to resolve DSBs (γH2AX foci) 24h post-IR. In addition, 24h post-IR, these TAK-243 pretreated cells also exhibited reduced foci formation of 53BP1 compared to DMSO pretreated controls (Fig. 2.7E-F).

2.3.6 TAK-243 reduces the leukemic burden in a mouse xenograft model of AML

To evaluate the preclinical efficacy and toxicity of TAK-243 in vivo, we used a mouse xenograft model of AML in which OCI-AML2 cells were injected subcutaneously (sc) into the flanks of SCID mice. After tumors became palpable, mice were treated with either vehicle or TAK-243 at a dose of 20 mg/kg sc twice weekly (BIW). Systemic administration of TAK-243 significantly delayed tumor growth rate and reduced final tumor weights in treated mice as compared to vehicle control (Fig. 2.9A-B, and 2.10A). The dose and schedule were well-tolerated with no significant changes in body weight (Fig. 2.10B). Serum chemistries measured from the blood obtained from mice at the end of treatment showed no increase in levels of total bilirubin, alkaline phosphatase (ALP), creatinine and aspartate transaminase (AST), or creatinine kinase (CK) between vehicle- and TAK-243-treated mice (Fig. 2.10C). In addition, no significant changes were observed in gross or histological examination of organs collected at the end of treatment (Fig. 2.10D).

2.3.7 TAK-243 reduces engraftment of human primary AML cells in the bone marrow of mice

To evaluate the preclinical efficacy of TAK-243 in a more clinically relevant model of AML, primary AML cells were injected into the right femurs of sublethally irradiated NOD-SCID mice. Two weeks after injection of the primary cells, mice were treated with vehicle or TAK-243 (20 mg/kg sc BIW) for 3 weeks.

119

Figure 2.7 TAK-243 inhibits DNA double-strand break repair in AML cells. A, B) OCI- AML2 cells were treated with increasing concentrations of TAK-243 for increasing times. After treatment, whole cell lysates were prepared and levels of mono-ubiquitylated histones H2A and H2B were assessed by immunoblotting (A) and immunohistochemistry (B). Scale bar= 200 µm. C, D) OCI-AML2 cells were treated with increasing concentrations of TAK-243 for 4 h. Cells were then subjected to irradiation (3 Gy) and cytospins were prepared 1 h later. Cells were

120 examined for subnuclear γH2AX and 53BP1 foci by immunofluorescence. Foci were then quantified (C). Each bar represents mean ± SEM. Representative images of γH2AX and 53BP1 foci are shown (D). ns: non-significant; *p ≤ 0.05; ****p < 0.0001 using two-way ANOVA and Sidak's multiple comparisons test. Scale bar = 20 µm. E, F) OCI-AML2 cells were treated with DMSO or TAK-243 (100 nM) for 2 h. Cells were then irradiated (3 Gy) and cytospins were prepared 1 and 24 h post-IR. Cells were examined for subnuclear γH2AX and 53BP1 foci by immunofluorescence. Foci were then quantified (E) and bars represent mean ± SEM. Representative images of γH2AX and 53BP1 foci are shown (F). ****p < 0.0001 using two-way ANOVA and Sidak's multiple comparisons test.

121

Figure 2.8 TAK-243 inhibits histone ubiquitylation in primary patient samples and DNA double-strand break repair in AML cells. A and B) Primary AML samples were treated with TAK-243 (100 and 1000nM) for 3 h. After treatment levels of mono-ubiquitylated H2A (A) or H2B (B) were measured by immunohistochemistry. Scale bar= 200 µm. C and D) OCI-AML2 cells were treated with increasing concentrations of TAK-243 for 4 h. As a control for irradiated

122 cells, these cells were not subjected to ionizing radiation and cytospins were prepared 1 h later. Cells were examined for subnuclear γH2AX and 53BP1 foci by immunofluorescence. Foci were then quantified (C). Each bar represents mean±SEM of 3 independent experiments. Representative images of γH2AX and 53BP1 foci are shown (D). ***p < 0.001 using two-way ANOVA and Sidak's multiple comparisons test. Scale bar = 20 µm. E) OCI-AML2 cells were treated with increasing concentrations of TAK-243 for 2h. After treatment, whole cell lysates were prepared and levels of γH2AX and β-tubulin were measured by immunoblotting.

123

TAK-243 treatment reduced levels of leukemia as assessed by flow cytometric quantification of human CD45+CD19-CD33+ cells collected from the non-injected left femurs (Fig. 2.9C). Secondary transplant experiments demonstrated that TAK-243 also effectively targeted leukemic stem cells in vivo (Fig. 2.9D).

2.3.8 TAK-243 preferentially targets UBA1 in vivo

We next sought to assess TAK-243 binding to UBA1 and the effects on the ubiquitylation pathway in vivo. Mice with established subcutaneous OCI-AML2 xenograft tumors were treated with vehicle or TAK-243 at a dose of 20 mg/kg sc for 2 doses. Based on our CETSA data showing that TAK-243 remained bound to UBA1 up to 8 h after washout of the drug from OCI- AML2 cells treated in culture, we sacrificed the mice 8 h after the second dose (Fig. 2.11A). Tumors were collected after sacrifice to assess target engagement of TAK-243. As assessed by CETSA, TAK-243 preferentially bound UBA1 over UBA6 in OCI-AML2 tumors (Fig. 2.9E). We also examined the effects of TAK-243 on ubiquitylation in tumors and normal tissues after systemic administration to mice. At doses that produce an anti-leukemic effect in vivo, TAK-243 preferentially reduced levels of global ubiquitylated proteins and mono-ubiquitylated H2A and H2B in the leukemic cells compared to normal tissues (Fig. 2.9F-H and 2.11B, C and E-I). Finally, TAK-243 preferentially induced caspase 3 cleavage in the leukemic cells compared to normal tissues (Fig. 2.9G and 2.11D-I).

2.3.9 Missense mutations in the adenylation domain of UBA1 confer resistance to TAK- 243

To gain insight into potential mechanisms of resistance to TAK-243, we developed TAK-243- resistant cells by two approaches. First, we cultured OCI-AML2 cells continuously in increasing concentrations of TAK-243 starting from the IC90 of the drug. We selected a population of TAK-

243-resistant cells (OCI-AML2-RIC90) that were 33-fold more resistant to the drug compared to the parental cells (IC50 757 vs 23.4 nM; Fig. 2.12A). Second, we cultured cells in a cyclic manner by treating cells with stepwise increasing concentrations of TAK-243 starting from 0.5 nM. We selected a population of TAK-243-resistant cells (OCI-AML2-RSW) that were 28-fold more resistant compared to the parental cells (930.2 vs 33.8 nM; Fig. 2.12B). Compared to their parental cell lines, both OCI-AML2-RIC90 and OCI-AML2-RSW cells were equally sensitive to pevonedistat, bortezomib, daunorubicin, and mitoxantrone (Fig. 2.13A-D and 2.14A-D).

124

125

126

Figure 2.9 TAK-243 reduces the leukemic burden in mouse models of AML, and preferentially binds and inhibits UBA1 in AML cells in vivo. A) OCI-AML2 cells (1x106) were injected subcutaneously into the flanks of SCID mice. When the tumors became palpable, mice were randomly divided into 2 groups (n=10 per group) and treated with vehicle (10 % 2- hydroxypropyl-β-cyclodextrin [HPBCD] in water) or TAK-243 (20 mg/kg) subcutaneously (sc) twice weekly (BIW) for 3 weeks. Tumor volume was assessed by caliper measurements every 2- 3 days.****p < 0.0001 using repeated-measure two-way ANOVA and Sidak's multiple comparisons test. B) Tumors were harvested at the end of the study and weighed. The bars represent means ± SD of 10 mice. ****p < 0.0001 using unpaired t-test. C) Two AML patient samples (2.5 x106 viable trypan blue-negative cells) were injected into the right femurs of NOD- SCID mice that had been sub-lethally irradiated 24 h previously with 208 rad. Two weeks after injection of the primary cells, mice were treated with TAK-243 (20 mg/kg sc BIW) or vehicle control (10 % HPBCD in water). After 4 weeks of treatment, engraftment of human AML cells into the marrow of the non-injected left femurs was assessed by measuring the percentage of human CD45+CD19-CD33+ cells by flow cytometry. D) Primary human AML cells were isolated from the bone marrow of vehicle- and TAK-243-treated mice in (C). Equal numbers of viable cells were injected into the right femurs of untreated secondary mice. Mice were sacrificed 5 weeks post-injection and engraftment of human cells was assessed by measuring the percentage of human CD45+CD19-CD33+ cells by flow cytometry. Each data point in (C) and (D) represents one mouse with the horizontal bars representing the means. Unpaired t-test was used to determine significance. p values and primary sample numbers are shown on the graphs. E) Mice with established OCI-AML2 tumors were treated with two doses of vehicle or TAK-243 (20 mg/kg) on days 1 and 4. Eight hours after the second dose, mice were sacrificed, and tumors were collected and heated at 54°C followed by lysate preparation. Levels of UBA1, UBA6 and GAPDH were measured by immunoblotting. F) OCI-AML2 tumors were harvested from mice treated with vehicle or TAK-243 as in (E). Whole cell lysates were prepared and levels of global ubiquitylated proteins, Ub-H2A, Ub-H2B and β-actin were measured by immunoblotting. G, H) OCI-AML2 tumors, lungs, kidneys, liver, heart, and muscle were harvested from mice treated with vehicle or TAK-243 as in (E). Levels of global ubiquitylated proteins, Ub-H2A, Ub-H2B, and cleaved caspase 3 in OCI-AML2 tumors were measured by IHC (G). Levels of global ubiquitylated proteins in tumors and normal tissues were quantified by taking 3 random images from different areas of tumors or organs collected from each mouse (H). The Y-axis represents the area of positively stained cells as a percentage of the total area of the image. % staining was quantified by ImageJ software. Four mice were assessed per treatment. Each data point in the graph represents a single mouse and horizontal bars represent the means. Unpaired t-test was used for determining significance and corresponding p values are shown on the graph.

127

Figure 2.10 TAK-243 is effective and tolerable in an AML mouse xenograft model. A) Images of tumors harvested from mice. In the TAK-243-treated group, only two mice had detectable tumors. B) Mice were weighed every 2-3 days. The data points represent means ± SD of 10 mice. BIW: twice weekly. C) Serum was collected from 4 mice per group at the end of the study and serum chemistry was measured. All data points represent mean±SD of 4 mice. ALP, Alkaline phosphatase; AST, Aspartate transaminase. *p < 0.05 using unpaired t-test. D) Histology images of organs collected from mice. Scale bar = 200 µm.

128

129

Figure 2.11 TAK-243 binds and inhibits UBA1 in AML cells in vivo. A) OCI-AML2 cells were treated with TAK-243 (200 nM) for 1 h followed by washing the cells thoroughly with PBS and adding fresh drug-free media. The cells were incubated for 0, 1, 2, 4 and 8 h followed by heating the intact cells at 54°C. After heating, whole cell lysates were prepared and levels of UBA1 and GAPDH were measured by immunoblotting. B-D) Levels of ubiquitylated H2A, ubiquitylated H2B and cleaved caspase 3 in tumors and normal tissues were quantified by taking 3 random images from different areas of tumors or organs collected from each mouse. The Y- axis represents the area of positively stained cells as a percentage of the total area of the image. % staining was quantified by ImageJ software. 4 mice were assessed per treatment. Each data point in the graph represents a single mouse and horizontal bars represent the means. Unpaired t- test was used for determination of significance. E-I) Representative IHC images for lungs, kidneys, liver, heart and muscle collected from vehicle- and TAK-243-treated mice. Scale bar = 200 µm.

130

Using CETSA, binding of TAK-243 to UBA1 was reduced in both OCI-AML2-RIC90 and OCI-

AML2-RSW, while binding to UBA3 and UBA6 remained unchanged (Fig. 2.12C-D). Levels of UBA1 did not differ between the parental and resistant cells. Likewise, expression of XBP1s, GRP78 and breast cancer resistance protein (BCRP) was not different in the resistant cells (Fig. 2.13E and 2.14E). To identify the mechanism of resistance to TAK-243, we sequenced the exons spanning the adenylation domain of UBA1 (exons 12-16 and 23-24) in the resistant cell lines and their parental sensitive cells. We identified different mutations in the two resistant cell lines. In OCI-AML2-

RIC90 cells, we identified a missense mutation in exon 16 at codon 583 that results in substitution of tyrosine with cysteine (Y583C) (Fig. 2.12E). In OCI-AML2-RSW cells, we identified a missense mutation in exon 15 at codon 580 that results in substitution of alanine with serine (A580S) (Fig. 2.12F). We confirmed the presence of the mutations by repeating the sequencing in independent cell populations. Mutations in the highly conserved residue A580 have been reported to confer resistance to TAK-243 using recombinant UBA1 and cell-free enzymatic assays367. Therefore, we focused on understanding the potential mechanism by which the Y583C mutation confers resistance to TAK-243. As the structure of human UBA1 has not been reported yet, we used recently published structure of highly homologous yeast Uba1 (ScUba1) in complex with the drug to model the interactions of TAK-243 with the Y583C human UBA1 (HsUBA1) mutant; Fig. 2.13F)367. Unlike much of the protein, the 583 position is not well conserved in the E1 family of enzymes. However, it is usually occupied by larger side chains of tyrosine (ScUba1, HsUBA1, 6, 7), tryptophan (HsUBA3) or histidine (HsUBA2). Introducing cysteine into this position likely perturbs the hydrophobic core and creates a cavity between C583 and its neighboring residues. Moreover, it also eliminates putative hydrogen bonds that Y583 of human UBA1 can form with G553 and E557 (G521 and E525 in ScUba1; Fig. 2.13F). Thus, the Y583C mutation likely destabilizes this side of the binding pocket and may prevent efficient binding of the thio(trifluoromethyl) hook of TAK-243 to HsUBA1 surface.

131

Figure 2.12 Y583C and A580S UBA1 mutations confer resistance to TAK-243 in AML cells. A, B) Sensitive and resistant OCI-AML2IC90 and OCI-AML2SW cell lines were treated with increasing concentrations of TAK-243 for 48 h. After treatment, cell growth and viability were

132

measured by the MTS assay. Insert: IC50 values of the parental and resistant cells. C, D) Sensitive (S) and resistant (R) OCI-AML2IC90 and OCI-AML2SW cell lines were treated with increasing concentrations of TAK-243 for 30 min. For CETSA, cells were heated to 54°C followed by lysate preparation. Lysates of unheated controls were prepared in parallel. Levels of UBA1, UBA3, UBA6 and GAPDH were then determined by immunoblotting. E, F) Sanger sequencing chromatograms showing Y583C and A580S missense mutations of UBA1 in OCI- AML2IC90 and OCI-AML2SW cell lines. Corresponding UBA1 exons are shown in the figure. Y583C: substitution of tyrosine with cysteine at UBA1 codon 583; A580S: substitution of alanine with serine at UBA1 codon 580.

133

134

Figure 2.13 Y583C mutation in UBA1 confers resistance to TAK-243. A-D) Sensitive and resistant OCI-AML2IC90 cells were treated with increasing concentrations of pevonedistat, bortezomib, daunorubicin and mitoxantrone for 48 h. After treatment, cell growth and viability was measured by the MTS assay. Insert: IC50 values of sensitive and resistant cells. All data points represent the mean ± SEM of at least 3 independent experiments. E) Whole cell lysates of sensitive and resistant OCI-AML2IC90 cells were prepared and levels of breast cancer resistance protein (BCRP), XBP1, UBA1, GRP78 and GAPDH were assessed. F) Y583C (Y551C in ScUba1) mutation likely destabilizes the binding pocket by eliminating hydrogen bonds (dashed lines) and perturbing the hydrophobic core (modeled using ScUba1, PDB:5L6J, chain A) – generated using PYMOL (DeLano Scientific).

135

136

Figure 2.14 A580S mutation in UBA1 confers resistance to TAK-243. A-D) Sensitive and resistant OCI-AML2SW cells were treated with increasing concentrations of bortezomib, pevonedistat, daunorubicin and mitoxantrone for 48 h. After treatment, cell growth and viability was measured by the MTS assay. Insert: IC50 values of sensitive and resistant cells. All data points represent the mean ± SEM of at least 3 independent experiments. E) Whole cell lysates of sensitive and resistant OCI-AML2SW cells were prepared and levels of breast cancer resistance protein (BCRP), XBP1, UBA1, GRP78 and GAPDH were assessed.

137

2.4 Discussion

In this study, we evaluated the UBA1 inhibitor TAK-243 in AML. By the CETSA assay, TAK- 243 preferentially bound UBA1 in AML cell lines and primary AML samples over the related E1 enzymes UBA2, UBA3, and UBA6. Moreover, TAK-243 preferentially inhibited UBA1- mediated ubiquitylation in AML over the activity of the other E1 enzymes. Thus, these data support biochemical studies demonstrating TAK-243 as a preferential UBA1 inhibitor278. TAK- 243 inhibits UBA1 by a mechanism referred to as substrate-assisted inhibition, a unique mechanism that imparts potency and selectivity to members of this class263, 278. As such, it is similar to pevonedistat that binds UBA3 and acts as a selective mechanism-based inhibitor of neddylation42, 70. Our studies in vitro and in vivo demonstrated that TAK-243 preferentially targeted AML cells and stem/progenitors over normal cells and normal hematopoietic progenitors. Moreover, TAK- 243 preferentially inhibited ubiquitylation in AML cells over normal cells in vivo. These data are consistent with our previous studies demonstrating that UBA1 enzymes have reduced reserve capacity in AML compared to normal hematopoietic cells27. In addition, these data help explain the therapeutic window we observed with TAK-243 in our mouse models of leukemia. Hyer et al. evaluated the pharmacokinetics of TAK-243 in mice and reported that a maximal plasma concentration of 2.69 µM was achieved after a dose of 18.8 mg/kg278. Future clinical studies in AML will be necessary to determine anti-leukemic concentrations of TAK-243 that can be obtained in humans and to define the efficacy and toxicity of this drug in patients with AML. TAK-243 induced ER stress in AML cells. van Galen et al. reported that hematopoietic stem cells are more sensitive to ER stress stimuli compared to more differentiated cells368. Therefore, the activity of TAK-243 against leukemic stem and progenitor cells may derive in part from its ability to induce ER stress and an increased sensitivity of leukemic stem cells to ER stressors. Ubiquitylation is critical for DNA damage response364. Indeed, several proteins must be ubiquitylated for efficient DSB repair. For instance, ubiquitylation of H2A-type histones and Histone H1 at DSB flanking sites triggers orchestrated recruitment to these damage sites of a number of DSB signaling and repair proteins including 53BP1 and BRCA1; thus, allowing the repair through non-homologous end-joining (NHEJ) and homologous recombination (HR), respectively. Our data indicate that TAK-243 restrains levels of Ub-H2A and Ub-H2B and impairs DSB repair. Notably, the recruitment of 53BP1 and its retention at DSB flanking sites

138 was significantly impaired. Given the importance of 53BP1 for NHEJ, the impaired IR-induced 53BP1 foci formation after TAK-243 treatment suggests that TAK-243 impairs NHEJ efficiency and this likely contributes to impaired DSB repair. Several classes of proteins are known to be regulated by ubiquitylation such as NF-κB, cell cycle proteins (e.g. cyclins and CDK inhibitors), and short-lived proapoptotic proteins (e.g. p53)1, 26. Ubiquitylation is also involved in regulating DNA damage response (e.g. double-strand break repair, translesion synthesis and Fanconi anemia pathway)39. In the context of AML, a number of these proteins play important biological roles and their perturbation by UBA1 inhibition may account for the observed cytotoxicity. For example, NF-κB was reported to be constitutively active in AML and in leukemic stem cells compared to normal progenitors327, 369. This higher activity is maintained by proteasomal degradation of the NF-κB inhibitor IκBα, which may explain the higher requirement of UBA1 and the observed therapeutic window in AML. Although UBA1 inhibition is expected to stabilize p53, proteasomal inhibition was reported to be efficacious in AML cells independent of p53 status370. To understand potential mechanisms of resistance to TAK-243, we selected a population of resistant cells and identified Y583C and A580S missense mutations in the adenylation domain of UBA1. Based on our analysis of available E1 enzyme structures, resistance of the Y583C mutant is likely caused by elimination of hydrogen bonds and destabilization of the protein hydrophobic core. Only further structural studies will be able to fully explain the mechanism of resistance of

Y583C and A580S mutants to TAK-243. Of note, Y583 in human UBA1 corresponds to Y551 in yeast Uba1, which was reported to be uniquely involved in binding TAK-243 but not pevonedistat or ABPA3, a dual inhibitor of UBA1 and NAE265, 367. On the other hand, mutation in A580 has been reported to confer TAK-243 resistance in recombinant UBA1367. In addition, A580 corresponds to A171 in NAE, which was also involved in pevonedistat resistance269, 270. Thus, we report for the first time Y583C mutation as a TAK-243-specific resistance mechanism in a cell-based model, which may have clinical relevance in AML and potentially other malignancies.

2.5 Summary and conclusion

In summary, our study demonstrates TAK-243 is a selective UBA1 inhibitor with a potent anti- leukemic activity in cell-based and animal models of AML. It decreases the abundance of

139 ubiquitylated proteins, induces proteotoxic stress and inhibits DNA repair in AML (Fig. 2.14). In addition, TAK-243 exhibits preferential activity towards leukemic versus normal progenitors in vitro and inhibits primary and secondary engraftment of patient-derived leukemic cells in bone marrow of mice confirming its activity towards leukemic stem and progenitor cells. It also displays preferential inhibition of UBA1 in tumor xenografts over normal tissues, explaining in part its tolerability in mice. Our study also reports missense mutations in UBA1 that confer acquired resistance to TAK-243 and may have clinical relevance. In conclusion, our findings reveal preclinical efficacy and tolerability of TAK-243 and support advancement of this drug to a phase 1 clinical trial in AML.

140

Figure 2.15 Working model of TAK-243 cytotoxicity in AML. TAK-243 binds to and inhibits UBA1 resulting in reductions in ubiquitylation and disruption of several signaling pathways. These biological effects lead to the induction of proteotoxic stress and DNA damage and consequently cell death. Structure of TAK-243 is depicted. NAE: NEDD8-activating enzyme; UPS: ubiquitin-proteasome system. Adapted with permission from Ref. 210.

Chapter 3

Determinants of TAK-243 sensitivity in AML

141

Chapter 3 Determinants of TAK-243 sensitivity in AML

3.1 Introduction

TAK-243 is a first-in-class inhibitor of the ubiquitin-like modifier activating enzyme 1 (UBA1) that catalyzes the first step of ubiquitin conjugation cascade8. Through this cascade, protein substrates are tagged with mono- or poly-ubiquitin to induce their proteasomal degradation or to modulate their functions1, 7. This process is executed through multi-step enzymatic reactions whereby ubiquitin is initially activated by the ubiquitin-activating enzyme (E1) in an ATP- dependent manner. This step is followed by transfer of the activated ubiquitin from the catalytic cysteine site of E1 to the corresponding catalytic cysteine in one of the cognate ubiquitin- conjugating E2 enzymes (E2s). Ubiquitin is then transferred to protein substrates by E2s and this step is facilitated by ubiquitin ligases (E3s). While UBA1 is the major ubiquitin E1 in the cell, there are over 30 ubiquitin E2s and hundreds of ubiquitin E3 that mediate ubiquitylation of substrates in a highly coordinated and specific manner10. We previously reported that AML cell lines and primary patient samples are more dependent on the activity of UBA1 compared to normal hematopoietic cells, and thus are more vulnerable to UBA1 inhibition27. UBA1 was also reported by others to serve as a therapeutic target in cancer26. Accordingly, we evaluated the selective UBA1 inhibitor, TAK-243, in preclinical models of AML and found that it displayed a potent anti-leukemic activity in vitro and in vivo182, 210. Similar findings have also been reported with TAK-243 in solid tumors and other hematologic malignancies78, 183-185. Nonetheless, the determinants of sensitivity to TAK-243 are still largely unknown. To gain further insights into mechanisms of sensitivity and resistance to TAK-243, we conducted a genome-wide CRISPR/Cas9 knockout screen in AML cells and identified the transcriptional repressor, BEN domain-containing protein 3 (BEND3) as the top gene whose knockout confers resistance to TAK-243.

3.2 Methods

142 143

3.2.1 Chemicals and reagents

TAK-243 (formerly known as MLN7243) was provided by Takeda Pharmaceuticals Inc. and purchased from Active Biochem (catalog# A-1384), and pevonedistat was provided by Dalton Medicinal Chemistry (Toronto, Canada). Bortezomib (catalog# S1013) was purchased from Selleckchem, mitoxantrone (catalog# M6545) and (2-Hydroxypropyl)-β-cyclodextrin (HPBCD; catalog# H107) from Sigma-Aldrich, tunicamycin (catalog# 3516), thapsigargin (catalog# 1138), zosuquidar (catalog# 5456), and Ko143 (catalog# 3241) from Tocris.

3.2.2 Cell culture

OCI-AML2, K562, MV4-11 and RPMI 8226 cells were cultured in Iscove’s modified Dulbecco’s medium (IMDM), NB4, U937, MDAY-D2 and Jurkat cells in Roswell Park Memorial Institute (RPMI) medium, and A549, Hep G2 and MCF7 cells in Dulbecco's Modified Eagle Medium (DMEM). All culture media were supplemented with 10% fetal bovine serum

° (FBS) and appropriate antibiotics. All cells were incubated at 37 C, 5% CO2 with 95% humidity.

3.2.3 Positive-selection genome-wide CRISPR/Cas9 knockout screen

The genome-wide CRISPR/Cas9 knockout screen was performed in OCI-AML2 cells stably expressing Cas9 (OCI-AML2-Cas9). OCI-AML2 cells were transduced with a Cas9 in a lentiviral vector, Lenti-Cas9-2A-Blast, (Addgene plasmid#73310)371. The cells were then selected with blasticidin (10 μg/mL) for 6 days. Single-cell clones were obtained by plating in a 96-well plate at a density of 0.4 cell/well, and a clonal population with high Cas9 expression was selected. OCI-AML2-Cas9 cells were transduced with a pooled library (90k library) of 91,320 guide RNAs (gRNAs) in lentiviral vectors targeting 17,232 genes at a ratio of 6 gRNAs per gene371. These cells were transduced at a multiplicity of infection (MOI) of approximately 0.3- 0.4 to obtain a coverage of at least 200-fold per gRNA. One day post-transduction, cells were treated with puromycin (2μg/mL) for 48h to select transduced cells. Cells were then treated with

DMSO or TAK-243 at its IC90 (25 nM) or IC99 (30 nM) for 32 days. Genomic DNA was then extracted from cell populations, gRNA sequences were amplified by PCR and sequenced on an Illumina Hiseq2500. Data were analyzed using Model-based Analysis of Genome-wide CRISPR/Cas9 Knockout (MAGeCK) method372.

144

3.2.4 CRISPR/Cas9 knockout experiments

OCI-AML2-Cas9 cells (5x106) were resuspended in 5 mL of fresh media containing protamine sulfate (5 µg/mL). Viral supernatants (2 mL) of distinct BEND3-targeting gRNAs encoded in pLCKO lentiviral vectors (crBEND3 #1 and #3) were added to cells to achieve an MOI of 0.3 (Addgene #73311; Ref. 371). After 24h of incubation, cells were centrifuged and resuspended in fresh media containing puromycin (1.5 mg/mL). After 3d of selection, cell lysates were collected, and knockout was then confirmed by immunoblotting. BEND3 was also knocked out using a single-plasmid system encoding additional gRNAs. To do so, OCI-AML2 cells were transduced with lentiCRISPR v2 vectors encoding Cas9 and 3 distinct BEND3-targeting gRNAs (crV2-BEND3 #3, 5 and 8) as described above (Addgene# 52961; Ref. 373). Sequences of BEND3-targeting gRNAs are listed in Table 3.1.

3.2.5 Cytotoxicity assays

CellTiter 96® AQueous MTS Reagent Powder was purchased from Promega (catalog# G1111), and Annexin V-FITC apoptosis kit from Biovision (catalog# K101-400). The MTS and Annexin V/propidium iodide (PI) assays were performed as per the manufacturer’s instructions. For the MTS assay, the cells were counted and seeded in 96-well plates at the following densities: OCI- AML2 (25,000/well), K562 (10,000/well), MV4-11 (25,000/well), RPMI 8226 (25,000/well), NB4 (25,000/well), U937 (10,000/well), MDAY-D2 (10,000/well) and Jurkat (10,000/well), and then treated with increasing concentrations of the drug(s) under investigation. After 72h of incubation, the MTS solution was directly added to the media at a ratio of 1:5 and absorbance was measured at 490 nm using SpectraMax Microplate Reader (Molecular Devices, USA). The growth and viability was then calculated as a percentage of the untreated cells and concentration- response curves were constructed and IC50 calculated using the non-linear regression function in GraphPad Prism (Version 6.03, GraphPad Software Inc). For the Annexin V/PI assay, OCI- AML2 cells were seeded in a 24-well plate at a plating density of 1x105 cells/mL and treated with increasing concentrations of TAK-243. After 96h of incubation, media were collected, and cells washed with phosphate-buffered saline (PBS), centrifuged at 2,000 rpm for 10 min and then re-suspended in the binding buffer containing Annexin V-FITC and PI. Unstained and single- stained cells were used as compensation controls. Flow cytometric analysis of samples was performed using BD FACSCANTO flow cytometer (BD Biosciences).

145

For the colony-forming assay, control and BEND3 knockout OCI-AML2 cells were seeded in MethoCult™ H4100 medium (Stem Cell Technologies, Vancouver, Canada) in 35 mm gridded dishes at a plating density of 400 cells/dish (DMSO-treated) or 1000 cells/dish (TAK-243- treated) and were incubated for 7d to allow colonies to form. After incubation, colonies of at least 50 cells were counted and plating efficiency (PE) was calculated from DMSO-treated controls using this equation: (#colonies formed/#cells seeded). The % viability of TAK-243- treated cells was then calculated using this equation: [#colonies formed/(#cells seeded x PE) x100]374. For proliferation assays, DMSO- and TAK-243-treated OCI-AML2 cells were seeded at a density of 104 cells/mL and viable trypan blue-negative cells were counted every 2-3 days using a hemocytometer.

3.2.6 Cellular thermal shift assay (CETSA)

We conducted CETSA as previously described375. In brief, cells were treated with increasing concentrations of TAK-243 for 0.5 and 1h. Cells were then washed with PBS and re-suspended in PBS containing protease inhibitor cocktail (Thermo Fisher Scientific). Cells were heated at 54°C for 3 min in a thermal cycler (SimpliAmp, Applied Biosystems). This temperature corresponds to the maximal thermal shift of UBA1 as previously described182. Cell lysates were prepared by 4 freeze–thaw cycles in liquid nitrogen and a thermal cycler set at 25°C, respectively with vigorous vortexing in between. Lysates were then centrifuged at 20,000 g for 20 min and supernatants were collected and frozen at -70°C until immunoblotting.

3.2.7 Quantitative reverse transcription polymerase chain reaction (RT- qPCR)

Total RNA was isolated using the RNeasy Plus Mini Kit (QIAGEN), and reverse transcribed into cDNAs using SuperScript IV Reverse Transcriptase (ThermoreFisher, MA, USA). Equal cDNA amounts were then added to a PCR master mix (Power SYBR Green PCR Master mix; Applied Biosystems, CA, USA). RT-qPCR reactions were conducted using an ABI Prism 7900 sequence detection system (Applied Biosystems, CA, USA). The relative gene expression was calculated by the 2–∆∆Ct method using 18s rRNA as a control. Sequences of the primers used in the study are listed in Table 3.2.

146

Table 3.1 BEND3-targeting gRNA sequences ID F/R Sequence (5’→3’) crBEND3 #1 F ACCGTCAGGAGCCGCAGTGACGAG R AAACCTCGTCACTGCGGCTCCTGA crBEND3 #3 F ACCGGGATGAGCTTGATGCGGGAG R AAACCTCCCGCATCAAGCTCATCC crV2 BEND3 #3 F CACCGGGATGAGCTTGATGCGGGAG R AAACCTCCCGCATCAAGCTCATCCC crV2 BEND3 #5 F CACCGTGAACAGTACAGCTGCTACG R AAACCGTAGCAGCTGTACTGTTCAC crV2 BEND3 #8 F CACCGAGTAGACCTCCACATAGTTG R AAACCAACTATGTGGAGGTCTACTC

Table 3.2 Primers used for RT-qPCR ID F/R Sequence (5’→3’) ABCG2 F GTGGCCTTGGCTTGTATGAT R GATGGCAAGGGAACAGAAAA ABCB1 F GCTCCTGACTATGCCAAAGC R TCTTCACCTCCAGGCTCAGT

147

3.2.8 Immunoblotting

To prepare whole cell lysates, cells were washed with PBS (pH=7.4) and lysed with radio- immunoprecipitation assay (RIPA) buffer followed by sonication and centrifugation at 13,000 rpm for 20 min at 4°C. Supernatants were collected and total protein quantified using the Bradford assay (Bio Rad, Hercules, CA). Samples were then denatured by boiling at 95°C for 5 min. For CETSA lysates, samples were not sonicated and were heated at 70°C for 10 min. Proteins were loaded in equal amounts and then fractionated by 10% gels (except otherwise specified) using sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE). Proteins were transferred to polyvinylidene difluoride (PVDF) membranes and then probed using appropriate primary and secondary antibodies (Table 3.3).

3.2.9 Determination of intracellular ATP levels

Intracellular ATP levels were measured using a highly sensitive ATP Bioluminescence Assay Kit HS II (Sigma-Aldrich; catalog# 11-699-709-001) as per the manufacturer’s guidelines. In brief, control and BEND3KO OCI-AML2 cells were washed with PBS and re-suspended in the manufacturer’s dilution buffer, and then seeded in triplicate in white 96-well microtiter plates at a plating density of 25,000 cells and a volume of 25 μL per well. Cells were then lysed by adding an equal volume of cell lysis buffer and incubating for 5 min at RT. A 50 μL of the luciferase reagent was then dispensed by automated injection and luminescence was measured after a 1s delay and integration for 1s using Hidex Sense Microplate Reader (Hidex Inc.). Relative ATP levels in BEND3KO OCI-AML2 cells were calculated by normalizing the luminescence intensities obtained from the assay to control OCI-AML2 cells.

3.2.10 Measurement of intracellular TAK-243 concentrations

To assess TAK-243 concentrations in the cells, BEND3KO and control OCI-AML2 cells were seeded in triplicate in a 12-well plate at a density of 10x106/well and then treated with increasing concentrations of the drug. After 1h of incubation, cells were collected, centrifuged at 3,000 rpm for 5 min, and media removed by aspiration. The cells were then washed twice with drug-free PBS and kept on ice during processing. Cell pellets were then extracted with 50 µL of ice-cold acetonitrile containing internal standard. Cell extracts were centrifuged at 14,000 rpm for 10 min,

148 followed by careful collection of 40 µL of the supernatant in HPLC vials and were stored at - 20°C until liquid chromatography–mass spectrometry (LC-MS) analysis. To measure TAK-243 by LC-MS, we used an Acquity UPLC BEH C18 (2.1 X 50 mm, 1.7 µm) column using ACQUITY UPLC II system. The mobile phase was 0.1% formic acid in water (solvent A) and 0.1% formic acid in acetonitrile (solvent B). A gradient starting at 95% solvent A going to 5% in 4.5 min., holding for 0.5 min., going back to 95% in 0.5 min. and equilibrating the column for 1 min. was employed. A Waters Synapt G2S QTof mass spectrometer equipped with an electrospray ionization source was used for mass spectrometric analysis.

3.2.11 Animal studies

Control and BEND3KO OCI-AML2 cells (1x106 tryban-negative viable cells) were injected subcutaneously (sc) into the right and left flanks of male SCID mice (Ontario Cancer Institute, Toronto, Canada), respectively. After the tumors became palpable, mice were randomly divided into 4 groups (n=5 per group) and treated with vehicle (10 % HPBCD in water) or TAK-243 at doses of 10, 15, and 20 mg/kg sc twice weekly (BIW) for 3 weeks. Mice were weighed and tumor volumes were measured by caliper measurements every 2-3 days using the following equation355: tumor volume (mm3) = tumor length (mm) × width2 (mm)× 0.5236. At the end of the experiment, mice were euthanized, and tumors excised for weighing. All animal studies were carried out according to the regulations of the Canadian Council on Animal Care and with the approval of the local ethics review board.

3.2.12 Statistical and data analysis

GraphPad Prism software version 6.03 (GraphPad Software Inc.) was used to perform all statistical analyses. To calculate significance of differences between means, unpaired t-test (2 groups), one-way ANOVA and appropriate multiple comparisons test (˃ 2 groups), and two-way repeated measures ANOVA (for the tumor growth rate in mice) were used. Half-maximal inhibitory concentration (IC50) and IC90 values were calculated using the nonlinear regression function in GraphPad Prism. All experiments were performed in triplicate with at least 3 biological replicates unless otherwise specified. To analyze flow cytometry data, BD FACSDiva Software 6.0 (BD Biosciences) and FlowJo version 7.7.1 (FlowJo, LLC) were used.

149

Table 3.3 Antibodies used in Western blotting, their sources and dilution Primary antibody Source Clone Catalog# Dilution ATF4 Santa Cruz C-20 sc-200 Primary: 1:100 Secondary: 1:1000

BEND3 Proteintech - 23101-1-AP Primary: 1:500 Secondary: 1:1000

Cas9 Cell Signaling 7A9-3A3 #14697 Primary: 1:1000 Secondary: 1:1000

CHOP Sigma - G6916 Primary: 1:500 Secondary: 1:1000

GAPDH Cell Signaling 14C10 2118 Primary: 1:4000 Secondary: 1:2000

PARP Cell Signaling - 9542 Primary: 1:1000 Secondary: 1:1000 p-JNKThr183/Tyr185 Cell Signaling - 9251 Primary: 1:1000 Secondary: 1:1000

UBA1 Santa Cruz 2G2 sc-53555 Primary: 1:300 Secondary: 1:1000

UBA2 Santa Cruz 28 sc-136359 Primary: 1:500 Secondary: 1:1000

UBA3 Santa Cruz E-5 sc-377272 Primary: 1:400 Secondary: 1:1000

UBA6 Cell Signaling - 13386 Primary: 1:1000 Secondary: 1:1000

Ubiquitin Cell Signaling - 3933 Primary: 1: 500 Secondary: 1:1000

Ubiquityl-Histone Cell Signaling D27C4 8240 Primary: 1:1000 H2A (Lys119) Secondary: 1:1000

β-actin Santa Cruz AC-15 sc-69879 Primary: 1:10,000 Secondary: 1:5,000

γH2AXSer139 EMD Millipore JBW301 05-636 Primary: 1:600 Secondary: 1:1000

150

3.3 Results

3.3.1 A positive-selection genome-wide CRISPR/Cas9 knockout screen identifies BEND3 as a top hit To identify genes that influence the cytotoxicity of TAK-243, we performed a genome-wide CRISPR/Cas9 knockout screen in OCI-AML2 cells treated with TAK-243 at concentrations of the drug corresponding to the IC90 and IC99. We focused our analysis on genes whose knockout conferred resistance to TAK-243. Using the MAGeCK algorithm to rank enriched genes376 and a false discovery rate (FDR) of < 0.2, 33 and 11 genes were identified as enriched in the populations of cells treated with TAK-243 at its IC90 and IC99 arms, respectively (Fig. 3.1A and Table 3.4-3.7). At both concentrations, BEND3 ranked as the top hit (FDR = 0.001238). In addition, all 6 gRNAs targeting BEND3 were enriched up to 1,222- and 9,136-fold after selection with TAK-243 at IC90 and IC99 concentrations, respectively (Fig. 3.1B). Pathway enrichment analysis demonstrated that significantly enriched gRNAs corresponded to genes involved in diverse processes including transcriptional regulation, chromatin organization, ATP metabolism, mTOR and NF-κB signaling pathways (Fig. 3.1C). To explore the role of other potential hits in TAK-243 sensitivity, we knocked out a number genes including KMT5B, FLCN, CDK2, DAZAP1, PNKD, SETD7, and FBXL19; however, none of these displayed a significant impact on TAK-243 response upon knockout with the exception of KMT5B. Therefore, we focused our investigation on the top hit BEND3. The role of KMT5B in TAK-243 sensitivity is planned in future work (See Section 4.2.2). 3.3.2 BEND3 knockout confers resistance to TAK-243 in AML cells To validate the screen results, we knocked out BEND3 using independent gRNAs. OCI-AML2 cells were stably transduced with gRNAs targeting BEND3 or control sequences, and knockout of BEND3 was confirmed by immunoblotting (Fig. 3.2A and C). As assessed by the MTS assay, BEND3 knockout cells were treated with increasing concentrations of TAK-243 and growth and viability measured by the MTS assay. BEND3 knockout conferred resistance to TAK-243 with up to an 8.9 increase in the IC50 of the drug (Fig. 3.2B and D). Moreover, BEND3 knockout conferred resistance to TAK-243 as measured by Annexin V/PI staining and proliferation assays (Fig. 3.2E). Finally, knockout of BEND3 also reduced the ability of TAK-243 to target the leukemia initiating cells in clonogenic assays (Fig. 3.2F and 3.4A). Of note, BEND3 knockout had little or no impact on cell proliferation rate in the absence of TAK-243 (Fig. 3.2G).

151

A IC arm 90 BEND3 BEND3 BEND3

10 KMT5B KMT5B

n a

h Number of guide RNAs c

- 0 d

l 20,000 40,000 60,000 80,000 100,000

2 g

o UBE2A

L DEPDC5 USP15 -10 NPRL2

IC99 arm

15 BEND3 BEND3 BEND3 KMT5B

10 KMT5B

g n

a 5 h

c Number of guide RNAs -

d 0

l o

f 20,000 40,000 60,000

2 -5

g o

L DEPDC5 -10 WDR11 USP15

152

Figure 3.1 A genome-wide CRISPR/Cas9 knockout screen identifies BEND3 as a top hit. A) Enrichment/depletion of guide RNAs (gRNAs) after treatment with concentrations corresponding to the IC90 (top) and IC99 (bottom) of TAK-243 as assessed by fold-change analysis compared to untreated control cells. B) Enrichment of gRNAs targeting BEND3 in the IC90 and IC99 arms of the screen. C) Gene Ontology (GO) analysis for genes whose gRNAs were enriched in the IC90 arm of the screen. A dashed vertical line represents the cut-off significance level at p < 0.001. Genes of relevance to GO processes are shown in the blue horizontal bars.

153

Table 3.4 Top enriched hits in the IC90 arm of the CRISPR/Cas9 knockout screen Rank Gene gRNAs Score p-value FDR Good LFC gRNA 1 BEND3 6 6.55E-21 2.87E-07 0.001238 6 10.853 2 FLCN 6 7.30E-13 2.87E-07 0.001238 6 6.3585 3 SUV420H1 6 3.83E-11 2.87E-07 0.001238 5 8.3798 4 DAZAP1 6 7.16E-11 2.87E-07 0.001238 6 6.4072 5 FBXL20 6 2.19E-08 8.62E-07 0.002475 6 4.5263 6 EPC2 6 2.87E-08 8.62E-07 0.002475 4 5.7689 7 CDK2 6 6.33E-08 1.44E-06 0.003536 6 5.9522 8 ZNF800 5 2.48E-07 2.58E-06 0.005569 4 4.5446 9 BAHD1 6 7.91E-07 5.46E-06 0.010451 6 3.9898 10 chr10Rand 584 1.11E-06 7.18E-06 0.012376 260 0.20079 11 RC3H1 6 1.55E-06 1.12E-05 0.014851 5 4.9588 12 RRAGA 6 1.64E-06 1.12E-05 0.014851 6 4.3055 13 ZNF787 6 1.67E-06 1.12E-05 0.014851 6 3.0131 14 HDAC1 6 5.85E-06 3.30E-05 0.040594 6 3.6472 15 LINS 6 6.18E-06 3.53E-05 0.040594 6 2.9762 16 FOXP3 6 1.03E-05 5.00E-05 0.053837 4 3.4924 17 SFMBT2 6 1.18E-05 5.95E-05 0.058581 5 3.5042 18 ACSL4 6 1.25E-05 6.12E-05 0.058581 6 3.7833 19 QKI 6 2.70E-05 0.000123 0.109586 5 4.296 20 ZFR 6 2.81E-05 0.00013 0.109586 5 3.9118 21 TFPI2 6 2.95E-05 0.000136 0.109586 6 2.4544 22 CCDC43 6 3.00E-05 0.00014 0.109586 6 2.814 23 DROSHA 6 3.95E-05 0.000193 0.144641 3 0.74809 24 UBE2F 3 5.17E-05 0.000254 0.18255 3 4.0303 25 SRM 6 5.65E-05 0.000277 0.187069 4 4.2371 26 MAT2B 6 5.86E-05 0.000285 0.187069 6 3.1659 27 SMAD7 6 6.22E-05 0.000296 0.187069 6 2.5751 28 SLC30A10 6 6.88E-05 0.000333 0.187069 4 3.4307 29 CALHM3 6 7.09E-05 0.000339 0.187069 4 3.6409 30 TIAL1 6 7.15E-05 0.000341 0.187069 5 3.0422 31 SAFB2 6 7.33E-05 0.000346 0.187069 5 3.9446 32 AZIN1 3 7.57E-05 0.000357 0.187069 3 6.5417 33 HEXB 6 7.59E-05 0.000358 0.187069 3 1.8724 34 ZC3HAV1 6 8.32E-05 0.000386 0.195836 6 2.9386 Good guide RNAs (gRNAs): gRNAs whose ranking is below the alpha cutoff; Score: Robust Ranking Aggregation (RRA) algorithm score of enriched/depleted gRNAs; FDR: False discovery rate; LFC: Log2 (fold change) which is Log2 of the average read count of treated cells divided by the average read count of untreated cells.

154

Table 3.5 Top enriched hits in the IC99 arm of the CRISPR/Cas9 knockout screen Rank Gene gRNAs Score p-value FDR Good gRNA LFC 1 BEND3 6 1.03E-19 2.87E-07 0.001238 6 11.258 2 SUV420H1 6 4.47E-12 2.87E-07 0.001238 4 7.6959 3 DAZAP1 6 1.23E-10 2.87E-07 0.001238 5 3.7274 4 CDK2 6 2.13E-09 2.87E-07 0.001238 5 2.4161 5 TIAL1 6 3.58E-07 1.44E-06 0.00495 4 1.3476 6 ACSL4 6 1.18E-06 3.73E-06 0.010726 4 0.81102 7 OPRK1 6 9.33E-06 2.50E-05 0.061528 3 0.65749 8 RRAGA 6 2.27E-05 6.17E-05 0.133045 3 0.95433 9 FLCN 6 2.79E-05 7.50E-05 0.143564 3 0.41866 10 EIF4G1 6 3.98E-05 0.000101 0.174752 3 0.36052 11 SAFB2 6 5.01E-05 0.000124 0.193969 3 1.0937

155

Table 3.6 Top depleted hits in the IC90 arm of the CRISPR/Cas9 knockout screen Rank Gene gRNAs Score p-value FDR Good LFC gRNA 1 DEPDC5 6 5.84E-15 2.87E-07 0.00099 6 -4.779 2 NPRL2 6 8.06E-10 2.87E-07 0.00099 5 -3.4036 3 KCTD5 4 3.64E-09 2.87E-07 0.00099 4 -4.2317 4 ATP8B4 6 1.92E-08 2.87E-07 0.00099 6 -4.8429 5 NANS 6 4.49E-08 2.87E-07 0.00099 6 -2.2525 6 UBE2A 6 9.73E-08 8.62E-07 0.002122 6 -3.9618 7 USP15 6 1.01E-07 8.62E-07 0.002122 4 -1.531 8 C12orf66 6 1.41E-07 1.44E-06 0.00275 6 -4.2062 9 TBC1D7 4 2.35E-07 1.44E-06 0.00275 4 -2.6609 10 SNRK 6 1.31E-06 7.18E-06 0.012376 4 -1.9391 11 KIAA1967 6 1.86E-06 1.06E-05 0.016652 6 -4.5925 12 CHP1 5 3.67E-06 2.10E-05 0.030116 5 -3.4998 13 ZFP91 6 5.28E-06 2.67E-05 0.035415 6 -3.2056 14 ZBTB39 5 1.05E-05 5.03E-05 0.061881 5 -2.697 15 FAM13B 6 1.19E-05 5.95E-05 0.067141 6 -3.45 16 GOLGA1 6 1.28E-05 6.23E-05 0.067141 5 -3.2728 17 FIBP 6 1.51E-05 7.38E-05 0.07484 4 -2.3146 18 NPRL3 6 1.76E-05 8.53E-05 0.07995 6 -2.9513 19 ZNF292 6 1.88E-05 9.22E-05 0.07995 5 -1.8471 20 RCBTB1 6 1.89E-05 9.28E-05 0.07995 5 -3.0836 21 TXK 6 2.02E-05 0.0001 0.082273 6 -2.9687 22 HNRNPD 5 2.28E-05 0.000108 0.084833 5 -5.2242 23 RAB7A 6 2.75E-05 0.000131 0.098364 4 -3.793 24 SZT2 6 3.15E-05 0.000152 0.107578 5 -2.3441 25 MORN1 6 3.29E-05 0.000157 0.107578 2 0.22232 26 ZNF496 6 3.47E-05 0.000162 0.107578 6 -4.1532 27 SLCO6A1 6 3.74E-05 0.000178 0.113495 3 -0.53046 28 PGP 6 5.02E-05 0.000233 0.143388 5 -6.193 29 TAS1R3 6 6.61E-05 0.0003 0.178047 4 -2.8178 30 REST 6 7.18E-05 0.000324 0.185974 6 -3.7764 31 FBXO27 6 8.05E-05 0.000368 0.198793 6 -3.952 32 RPS6KL1 6 8.09E-05 0.000369 0.198793 6 -3.3892 33 C11orf30 6 8.41E-05 0.000381 0.19907 6 -4.6306 34 TCF12 6 8.78E-05 0.000397 0.201077 6 -1.8531

156

Table 3.7 Top depleted hits in the IC99 arm of the CRISPR/Cas9 knockout screen Rank Gene gRNAs Score p-value FDR Good LFC gRNA 1 chr10Rand 584 3.87E-17 2.87E-07 0.000825 389 -5.2826 2 FOXA1 6 3.10E-16 2.87E-07 0.000825 6 -4.3456 3 NPRL2 6 2.95E-11 2.87E-07 0.000825 6 -8.6211 4 CAPRIN1 6 7.82E-10 2.87E-07 0.000825 6 -5.8658 5 NANS 6 1.27E-09 2.87E-07 0.000825 6 -8.1914 6 DEPDC5 6 3.91E-09 2.87E-07 0.000825 6 -9.1186 7 SZT2 6 8.80E-09 8.62E-07 0.001856 6 -7.1938 8 KCTD5 4 5.47E-08 8.62E-07 0.001856 4 -8.0817 9 SNRK 6 9.00E-08 1.44E-06 0.002475 6 -7.9576 10 ZBTB33 4 9.02E-08 1.44E-06 0.002475 4 -7.8773 11 RRAGC 5 1.75E-07 2.01E-06 0.00315 4 -5.4472 12 MARK3 6 6.29E-07 3.16E-06 0.004538 6 -5.6666 13 RAB41 6 1.53E-06 8.33E-06 0.01021 5 -5.7413 14 LacZ 96 1.64E-06 8.33E-06 0.01021 60 -4.8502 15 NPRL3 6 1.78E-06 8.90E-06 0.01021 5 -7.8757 16 C18orf8 6 1.85E-06 9.48E-06 0.01021 5 -6.2566 17 USP15 6 2.53E-06 1.18E-05 0.011939 6 -7.1794 18 PLOD2 4 6.94E-06 3.19E-05 0.030528 4 -4.7831 19 LEPR 6 8.44E-06 3.99E-05 0.036217 6 -7.2505 20 TBC1D7 4 1.11E-05 5.43E-05 0.046782 4 -7.6484 21 SLX4IP 6 1.48E-05 7.09E-05 0.056931 6 -6.4967 22 TPRG1L 6 1.51E-05 7.27E-05 0.056931 5 -6.4166 23 C2CD5 6 1.63E-05 7.96E-05 0.059621 6 -7.3736 24 CADPS2 6 1.78E-05 8.47E-05 0.06085 6 -6.0668 25 NRAS 2 1.99E-05 9.62E-05 0.066337 2 -5.6664 26 ORMDL2 4 2.19E-05 0.000105 0.069878 3 -6.906 27 PEF1 6 3.05E-05 0.000152 0.096535 3 -5.6264 28 FIBP 6 3.29E-05 0.000163 0.096535 6 -7.9116 29 SLC1A4 6 3.33E-05 0.000166 0.096535 4 -3.4202 30 RBM41 6 3.40E-05 0.000168 0.096535 5 -5.4648 31 CXorf38 5 3.89E-05 0.000198 0.110029 5 -5.8156 32 EPC1 6 4.29E-05 0.000214 0.115254 5 -3.9017 33 WDR31 6 4.51E-05 0.000224 0.115276 5 -6.1993 34 ZNF202 6 4.59E-05 0.000228 0.115276 6 -6.5758 35 ITPR1 6 4.74E-05 0.000234 0.115276 6 -4.4119 36 ZNF292 6 5.60E-05 0.000271 0.129675 6 -7.7994 37 ABCC1 6 6.12E-05 0.000296 0.137945 6 -6.8988 38 UBE2A 6 6.90E-05 0.00033 0.149687 6 -7.3422 39 SLC39A9 6 7.37E-05 0.00035 0.154735 5 -7.281 40 PI4K2A 6 8.83E-05 0.000415 0.178807 6 -5.7963

157

41 GSKIP 4 9.10E-05 0.000432 0.178807 4 -7.0041 42 FAM49B 4 9.20E-05 0.000436 0.178807 4 -6.2474 43 ELF1 6 9.66E-05 0.000451 0.180806 5 -5.0925 44 MORN1 6 9.86E-05 0.000462 0.180806 3 -3.8167 45 ESRRA 6 0.000109 0.000512 0.19615 5 -6.384 46 MAN1C1 6 0.000113 0.000535 0.200495 4 -5.784

158

Figure 3.2 BEND3 knockout confers resistance to TAK-243 in AML cells A) OCI-AML2 cells overexpressing Cas9 were stably transduced with either non-targeting gRNAs or gRNAs targeting BEND3. After transduction, whole cell lysates were prepared and levels of BEND3 and β-actin serving as a loading control were measured by immunoblotting. B) Control and BEND3KO OCI-AML2 cells were treated with increasing concentrations of TAK-243 for 72 h.

159

Cell growth and viability was measured by MTS assay. Insert: the IC50 values (nM) are shown. Data points represent means ± SEM of at least 3 independent experiments. C) Wild-type OCI- AML2 cells were stably transduced with a single-plasmid system encoding Cas9 and either non- targeting gRNAs or gRNAs targeting BEND3 (#3, 5, and 8). After transduction, whole cell lysates were prepared and levels of BEND3, Cas9 and GAPDH serving as a loading control were measured by immunoblotting. D) Control and BEND3KO OCI-AML2 cells were treated with increasing concentrations of TAK-243 for 72 h. Cell growth and viability was measured by MTS assay. Insert: the IC50 values (nM) are shown. Data points represent means ± SEM of at least 3 independent experiments. E) Control and BEND3KO OCI-AML2 cells were treated with 2 concentrations of TAK-243 for 96 h. Cell viability was measured by Annexin V/PI staining and flow cytometry. Data points represent means ± SEM of at least 3 independent experiments. F) Control and BEND3KO OCI-AML2 cells were treated with TAK-243 (30 nM) and then plated into colony-forming assays. After 7d of incubation, colonies of at least 50 cells were counted. The Y-axis shows the number of colonies as a percentage of the DMSO-treated controls taking into account plating efficiency as detailed in methods. G) Control and BEND3KO OCI-AML2 cells were seeded with or without TAK-243 (30 nM) and counted every 2-3 days. Data points represent means ± SEM of 2-3 counts.

160

3.3.3 BEND3 knockout confers resistance to TAK-243 in vivo

Next, we determined whether BEND3 regulates sensitivity of AML cells to TAK-243 in vivo. Control or BEND3 knockout OCI-AML2 cells were injected into severe combined immunodeficiency (SCID) mice. After the tumors became palpable, mice were treated with increasing doses of TAK-243 sc twice weekly (BIW). As previously described, TAK-243 produced dramatic reductions in tumor growth in wild-type OCI-AML2 cells182. In contrast, BEND3 knockout rendered the tumors resistant to TAK243 and thus grew at rate similar to control. (Fig. 3.3 and 3.4B-C). Of note, BEND3 knockout cells exhibited a tumor growth rate in vivo comparable to that of control cells in vehicle-treated mice, which is consistent with proliferation data observed in vitro.

3.3.4 BEND3 knockout dampens TAK-243 effects on ubiquitylation, proteotoxic stress and DNA damage response in AML cells

TAK-243 inhibits UBA1 leading to reductions in poly- and mono-ubiquitylation, with the resultant induction of proteotoxic and DNA damage stress and subsequent cell death182. To determine how BEND3 influences sensitivity to TAK-243 resistance, we treated control and BEND3 knockout OCI-AML2 cells with TAK-243 and measured changes in levels of UBA1, abundance of ubiquitylated proteins and markers of proteotoxic and DNA double-strand break repair. Knockout of BEND3 did not change protein levels of UBA1 or other related E1 enzymes (Fig. 3.5A). However, knockout of BEND3 attenuated TAK-243-induced reductions in both pan- ubiquitylation and H2A mono-ubiquitylation (Fig. 3.5B). In keeping with this finding, TAK- 243-treated BEND3 knockout cells exhibited a little or no induction of markers of proteotoxic stress (ATF4, CHOP and p-JNK), DNA damage (γH2AX) and apoptosis (PARP cleavage) (Fig. 3.5C).

3.3.5 BEND3 knockout reduces intracellular transport of TAK-243 into AML cells

TAK-243 is an adenosine monophosphate (AMP)-mimetic that inhibits UBA1 by forming a covalent adduct with ubiquitin, which tightly binds to the nucleotide-binding site of the enzyme inhibiting its activity in an ATP-competitive manner78.

161

Figure 3.3 BEND3 knockout cells are resistant to TAK-243 in an AML mouse xenograft model A and B) Control (A) and BEND3KO (B) OCI-AML2 cells (1x106) were injected subcutaneously into the right and left flanks of SCID mice, respectively. When the tumors became palpable, mice were randomly divided into 4 groups (n=5 per group) and treated with vehicle (10 % 2-hydroxypropyl-β-cyclodextrin [HPBCD] in water) or TAK-243 (10, 15 and 20 mg/kg) subcutaneously (sc) twice weekly (BIW) for 3 weeks. Tumor volume was assessed by caliper measurements every 2-3 days. C and D) After 3 weeks, mice were euthanized and tumors of control (C) and BEND3KO (D) OCI-AML2 cells harvested and weighed. Data points in A-D represent means ± SD of a representative experiment.

162

Figure 3.4 BEND3 knockout cells are resistant to TAK-243 in vitro and in vivo. A) Representative images of the colony-forming assay. B) Mice treated with vehicle and TAK-243 (10-20 mg/kg) were weighed every 2-3 days. C) Tumors were harvested at the end of the in vivo study and images of these tumors were taken.

163

Figure 3.5 BEND3 knockout dampens TAK-243 effects by reducing the intracellular transport of the drug into AML cells. A-C) Control and BEND3KO OCI-AML2 cells were treated with DMSO or TAK-243 (30 nM) for 24h. After treatment, whole cell lysates were prepared and levels of UBA1, UBA3, UBA6, UBA2, polyubiquitylated proteins, ATF4, PARP, cleaved PARP (c. PARP), CHOP, phospho-JNK (p-JNK), and Ser139 phosphorylated H2AX

164

(H2AX) were measured by immunoblotting. GAPDH and β-actin were used as loading controls. D) Control and BEND3KO OCI-AML2 cells were treated with DMSO or increasing concentrations of TAK-243 at 15-120 nM for 1h and at 300-1200 nM for 0.5 h followed by heating the intact cells at 54°C. After heating, whole cell lysates were prepared and levels of UBA1 and GAPDH were measured by immunoblotting. E) Control and BEND3KO OCI-AML2 cells were washed, seeded in equal numbers and then lysed. Luminescence was then measured after adding an ATP-dependent luciferase reagent. Relative luminescence obtained from BEND3KO OCI-AML2 cells was calculated by normalizing to control cells. Data points represent means ± SEM of at least 3 independent experiments. F) Control and BEND3KO OCI-AML2 cells were treated with increasing concentrations of TAK-243 (300-1200 nM) for 1h, washed and pellets were then extracted with acetonitrile. TAK-243 concentrations were then measured by LC-MS. Data points represent means ± SD of one representative experiment.

165

We used the cellular thermal shift assay (CETSA) assay to evaluate the binding of TAK-243 to UBA1 in control versus BEND3 knockout OCI-AML2 cells. Control and BEND3 knockout cells were treated with increasing concentrations of TAK-243 followed by measuring the thermal shift of UBA1 by immunoblotting. BEND3 knockout reduced TAK-243 binding to UBA1 (Fig. 3.5D). In contrast, BEND3 knockout did not change intracellular levels of ATP, indicating that resistance to TAK-243 could not be explained by increased levels of ATP that competes for UBA1 binding (Fig. 3.5E). To assess the accumulation of TAK-243 into OCI-AML2 cells, we measured intracellular TAK-243 concentrations following treatment with increasing concentrations of the drug for 1h. As assessed by LC-MS, BEND3KO cells displayed significantly reduced intracellular concentrations of TAK-243 compared to control (Fig. 3.5F).

3.3.6 Upregulation of BCRP mediates TAK-243 resistance upon BEND3 knockout in AML cells

Emergence of multi-drug resistance (MDR) is a common problem with antineoplastic agents including cytotoxic drugs and molecularly targeted therapeutics377. A major class of proteins mediating MDR is the ATP-binding cassette (ABC) transporters that act as efflux pumps to extrude drugs and xenobiotics out of the cells in an ATP-dependent manner378. Since BEND3 knockout reduced the accumulation of TAK-243 into OCI-AML2 cells, we hypothesized that upregulation of one or more of ABC transporters may be responsible for the resistance phenotype. Of the 49 known human ABC transporters, 12 have been reported to be commonly implicated in MDR. To determine the most likely transporter for which TAK-243 might serve as a substrate, we correlated publicly available mRNA expression data of these 12 transporters across 30 cancer cell lines for which TAK-243 sensitivity has been reported. Breast cancer resistance protein (BCRP) displayed the strongest correlation between expression and TAK-243 sensitivity with cells having the highest expression of BCRP being most resistant to TAK-243 (r = 0.83; p < 0.0001). MRP2 also displayed a weaker but statistically significant correlation (r = 0.51; p < 0.0038). All the other transporters in our analysis showed a weak and statistically non-significant correlation (Fig. 3.6A-B and 3.7). These data suggest BCRP (encoded by ABCG2) is likely a transporter that mediates TAK-243 efflux and its expression may be upregulated after BEND3 knockout. To test this hypothesis, we measured the changes in mRNA expression of ABCG2 as well as ABCB1 (encoding P-gp) in BEND3 knockout versus control OCI-AML2 cells.

166

Figure 3.6 Upregulation of BCRP mediates TAK-243 resistance upon BEND3 knockout in AML cells. A) RNAseq expression data of 12 ABC transporters were obtained from the Cancer Cell Line Encyclopedia (CCLE) and correlated with TAK-243 sensitivity (as measured by IC50) of 30 cell lines. The X-axis represents the ABC transporters and the Y-axis represents the value of the linear Pearson correlation coefficient (r) ± upper and lower confidence intervals (CI) for

167 each transporter. Significance of correlation is shown on the graph. ns: non-significant; **p ≤ 0.01; ****p ≤ 0.0001. B) Correlation curve of the mRNA expression of BCRP (ABCG2) and TAK-243 sensitivity (as measured by IC50). Data points represent the 30 cell lines used in the analysis. A logarithmic scale was used for the X-axis to display all the data points over a wide range. Insert: the Pearson correlation coefficient (r), CI and significance of correlation (as assessed by p value). C) Relative mRNA expression of BCRP and P-gp in control and BEND3KO OCI-AML2 cells as assessed by qRT-PCR. Data points represent mean+SD of a representative experiment. D, E) Control and BEND3KO OCI-AML2 cells were treated with increasing concentrations of TAK-243 alone and in combination with 0.5 µM Ko143 (D) or 0.5 µM zosuquidar (E) for 72 h. Cell growth and viability was measured by MTS assay. Insert: the IC50 values (nM) are shown. Data points represent means ± SEM of at least 3 independent experiments. F) MDAY-D2 mouse lymphosarcoma cells were treated with increasing concentrations of doxorubicin alone and in combination with 0.5 µM zosuquidar for 72 h. Data points represent means ± SEM of 2 independent experiments.

168

Figure 3.7 Correlation of ABC transporter expression and TAK-243 sensitivity. A-K) Correlation curves of the mRNA expression of ABC transporters commonly involved in drug multidrug resistance and TAK-243 sensitivity (as measured by IC50). Data points represent the 30 cell lines used in the analysis. A logarithmic scale was used for the X-axis to display all the data points over a wide range. Insert: the Pearson correlation coefficient (r), CI and significance of correlation (as assessed by p value). Transporter names are shown on the graphs.

169

As assessed by RT-qPCR, BEND3 knockout increased ABCG2 mRNA expression by 15-fold, while having no statistically significant effect on ABCB1 expression (Fig. 3.6C). To further validate the functional importance of BCRP in explaining resistance to TAK-243 after BEND3 knockout, we treated BEND3 knockout and control OCI-AML2 cells with increasing concentrations of TAK-243 alone and in combination with either the selective BCRP inhibitor Ko143379, 380, or the selective P-gp inhibitor zosuquidar381. BCRP but not P-gp inhibition re- sensitized BEND3 knockout cells to TAK-243 (Fig. 3.6D-E). As a control, zosuquidar sensitized MDAY-D2 cells that express high levels of P-gp to doxorubicin, a known substrate of P-gp (Fig. 3.6F).

3.3.7 BEND3 knockout confers partial cross-resistance to related adenosine sulfamates and selected MDR substrates

To determine whether BEND3 knockout confers resistance to other cytotoxic agents, we treated BEND3 knockout and control OCI-AML2 cells with increasing concentrations of 7 related and unrelated drugs. The drugs evaluated include the NAE inhibitor pevonedistat, the SAE inhibitor TAK-981, the proteasome inhibitor bortezomib, the ER stressors thapsigargin and tunicamycin, as well as the chemotherapeutic agents mitoxantrone and cytarabine. BEND3 knockout conferred partial cross-resistance to pevonedistat, TAK-981, cytarabine and mitoxantrone with a 2.6-, 3.3-,

2.5- and 1.85-fold increase in their IC50 values (Fig. 3.8). In contrast, knockout of BEND3 displayed no cross-resistance to bortezomib, thapsigargin or tunicamycin (Fig. 3.9).

3.3.8 TAK-243 is a substrate for BCRP in cell lines of different origin

To determine if BCRP mediates resistance to TAK-243 in other cell lines, we treated A549 lung cancer cells, MCF7 breast cancer cells, MDAY-D2 mouse lymphosarcoma cells, and RPMI- 8226 myeloma cells with TAK-243 alone and in combination with Ko143 or zosuquidar. Inhibition of BCRP with Ko143 sensitized all cell lines to TAK-243 with a potentiation up to 113-fold, while P-gp inhibition with zosuquidar had no impact on the response to TAK-243 (Fig. 3.10).

170

Figure 3.8 BEND3 knockout confers partial cross-resistance to related adenosine sulfamates and selected MDR substrates A-E) Control and BEND3KO OCI-AML2 cells were treated with increasing concentrations of pevonedistat (A), TAK-981 (B), mitoxantrone (C) and Ara-C (D) for 72 h. Cell growth and viability was measured by MTS assay. Insert: the IC50 values (nM) are shown. Data points represent means ± SEM of at least 3 independent experiments.

171

Figure 3.9 BEND3 knockout does not affect response to proteasome inhibitors or ER stressors. A-C) Control and BEND3KO OCI-AML2 cells were treated with increasing concentrations of bortezomib (A), thapsigargin (B), and tunicamycin (C) for 72 h. Cell growth and viability was measured by MTS assay. Insert: the IC50 values (nM) are shown. Data points represent means ± SEM of at least 3 independent experiments.

172

Figure 3.10 TAK-243 is a substrate for BCRP in cell lines of different origin. A-D) A549 (A), MCF7 (B), MDAY-D2 (C) and RPMI 8226 (D) cells were treated with increasing concentrations of TAK-243 for 72 h. Cell growth and viability was measured by MTS assay. Insert: the IC50 values (nM) are shown. Data points represent means ± SEM of at least 3 independent experiments. F) Potentiation of TAK-243 effect on A549, MCF7, MDAY-D2, and RPMI 8226 upon combination with Ko143 (BCRPi) or zosuquidar (P-gpi). The Y-axis represents fold-potentiation calculated from this equation: IC50 (TAK-243)/ IC50 (combination).

173

3.4 Discussion

TAK-243 is a selective mechanism-based UBA1 inhibitor with a broad preclinical efficacy in solid and hematologic malignancies and has entered phase 1 clinical trials78, 182-185. In this study, we evaluated the determinants of sensitivity to TAK-243 in AML with potential implications in other malignancies using a genome-wide CRISPR/Cas9 knockout screen. From this screen, we identified BEND3 as the top hit whose knockout conferred resistance to TAK-243. BEND3 is a transcriptional repressor that interacts with chromatin-modifying complexes and induces repressive histone and DNA methylation changes resulting in transcriptional repression382, 383. While BEND3 knockout conferred resistance to TAK-243 in vitro and in vivo, it did not alter basal cell proliferation, consistent with publicly available data from pan-cancer RNAi and CRISPR/Cas9 dropout screens showing BEND3 is not an essential gene with no significant cell depletion upon knockdown or knockout384. Our study demonstrated that knockout of BEND3 attenuated TAK-243 effects on poly- and mono-ubiquitylation of protein substrates and alleviated ER stress. Previous studies have shown that induction of ER stress by TAK-243 is functionally important for TAK-243-induced cell death78, 182, 184, 185. Through subsequent studies, we demonstrated that knockout of BEND3 upregulates the MDR protein BCRP resulting in increased efflux of the drug, reduced binding to UBA1 and consequently reduced UBA1 inhibition. Upregulation of MDR proteins leads to excessive efflux of structurally and mechanistically diverse drugs and is an important mechanism of drug resistance385. BCRP has been reported to mediate the resistance of many unrelated anticancer drugs including etoposide386, imatinib387, methotrexate388 and mitoxantrone389, 390, among others377, 378, 385. In keeping with this, our results showed the TAK-243-resistant BEND3 knockout cells were cross-resistant to known BCRP substrates including mitoxantrone and cytarabine. In AML, high expression of BCRP has been correlated to chemotherapy resistance, poor prognosis and unfavorable therapeutic outcomes391-395. To our knowledge, no prior studies have implicated drug efflux pumps as mechanisms of resistance to TAK-243 or the related adenosine sulfamates such as the UBA3 inhibitor, pevonedistat. Pevonedistat has been extensively studied in preclinical settings and in over 30 clinical trials; however, upregulation of MDR proteins has not been reported as a mechanism of resistance to this drug. Instead, on-target missense mutations in UBA3 (the gene encoding the

174 active NAE subunit) have been reported to mediate acquired resistance to pevonedistat269-271. Analogous on-target missense mutations in UBA1 have also been associated with TAK-243 resistance182, 205. Here, we report for the first time that TAK-243 serves as a substrate for BCRP whose upregulation upon BEND3 knockout confers resistance to the drug and potentially related adenosine sulfamates. TAK-243 has been preclinically evaluated in several malignancies; however, the determinants of sensitivity remain largely unknown78, 182-185. Hyer et al. investigated whether sensitivity of TAK- 243 was related to UBA1 expression levels or cell line proliferation rates as assessed by doubling time, but found no significant correlation78. In our study, we demonstrated that TAK-243 sensitivity strongly correlated with BCRP expression levels in cancer cell lines of different origin. We also found that selectively targeting BCRP with chemical inhibitors sensitized cell lines intrinsically resistant to TAK-243 as a result of their high BCRP expression. Modulation of MDR proteins with inhibitors such as zosuquidar and tariquidar has been investigated in clinical trials as a strategy to sensitize certain malignancies to chemotherapy396, 397. In such settings, it should be noted that while BCRP inhibitors may sensitize cancer cells to TAK-243, they may also lead to a narrower therapeutic window by exposing cells, normally protected from xenobiotics by high BCRP expression, to higher concentrations of the drug398, 399. Therefore, this strategy may be used with caution in cases where toxicity can be managed or minimized. Expression of BCRP and other MDR proteins is regulated by multiple transcriptional and post- transcriptional mechanisms as well as alterations in cellular signaling. In this respect, promoter methylation status of ABCG2 under basal conditions or in response to chemotherapy was reported to control BCRP expression levels in multiple myeloma cell lines and patient samples400. MicroRNAs have also been implicated in regulating BCRP and other MDR proteins387, 401-403. In addition, hormonal alterations have been reported to alter cell signaling and subsequently BCRP expression in breast cancer404, 405. In our study, we demonstrated that BEND3 is important for regulating BCRP expression. Given its role as a transcriptional repressor, we speculate BEND3 regulates BCRP expression by inducing histone and DNA methylation changes at the promoter region of ABCG2. As per our CRISPR/Cas9 screen data, the histone methyl transferase KMT5B (SUV4-20H1) ranked as a second hit after BEND3. A related enzyme, KMT5C (SUV4-20H2), has been reported to interact with BEND3 in co- immunoprecipitation assays382. Loss of BEND3 has also been reported to increase histone H3K4 trimethylation and DNA methylation of the ribosomal DNA (rDNA) promoter, silencing rDNA

175 expression383. Therefore, it is possible that BEND3 may interact with KMT5B to alter the methylation of ABCG2 promoter, resulting in expression changes.

3.5 Summary and conclusions

In summary, our study demonstrates TAK-243 is a substrate for the ABC efflux transporter BCRP. Moreover, TAK-243 sensitivity correlates with BCRP expression levels in cancer cell lines of different origin, suggesting BCRP expression can serve as a predictor of TAK-243 response and a potential therapeutic target for synergistic combinations with TAK-243. Our study also reports, for the first time, BEND3 as a transcriptional regulator of BCRP expression and lack of BEND3 expression confers resistance to TAK-243 and potentially other BCRP substrates.

Chapter 4

Summary and future directions

176

Chapter 4 Summary and future directions 4.1 Summary

In this thesis, we evaluated different aspects of the selective UBA1 inhibitor, TAK-243, in AML with potential implications in other malignancies. First, we evaluated the efficacy and tolerability of TAK-243 in preclinical models of AML. In this respect, TAK-243 displayed a potent anti-leukemic activity in vitro and in vivo, which was associated with reductions in different forms of ubiquitylation and induction of proteotoxic and DNA damage stress. In vivo, TAK-243 exhibited preferential targeting of UBA1 in leukemic versus normal cells accounting, at least in part, for the observed tolerability of the drug in animal models. Based on these data, TAK-243 has been advanced to a phase 1 clinical trial in relapsed/refractory AML, refractory MDS, or chronic myelomonocytic leukemia (CMML) patients at Princess Margaret Cancer Center (ClinicalTrials.gov Identifier: NCT03816319). The primary objective of this trial is to determine the recommended phase 2 dose (RP2D) of TAK-243 after twice-weekly intravenous administration. Secondary objectives include assessment of safety profile, preliminary anti-leukemic activity, correlative response studies, pharmacokinetic and pharmacodynamic evaluation of the drug. Second, we identified potential mechanisms of acquired resistance to TAK-243 in AML and possibly other malignancies. In this respect, we developed a TAK-243-resistant AML cell line and identified two missense mutations mapping to the adenylation domain of UBA1. As per our biophysical and structural modeling studies, these mutations are predicted to interfere with TAK- 243 binding to UBA1. Whether these mutations play any role in acquired resistance to TAK-243 in clinical settings is yet to be determined. Third, we performed a genome-scale CRISPR/Cas9 knockout screen in AML cells to gain further insights into the biological determinants of TAK-243 sensitivity. Through this screen, we found that the lack of BEND3 expression is associated with upregulation of the ABC efflux transporter BCRP, for which TAK-243 is a substrate, and consequently resistance to the drug. Therefore, the basal expression of BCRP itself may serve as a predictor of sensitivity to TAK- 243 in AML and other malignancies. However, it remains unclear how BEND3 regulates the expression of BCRP, a direction which will be investigated in future work to gain a better understanding of transcriptional/epigenetic regulation of MDR protein expression.

177 178

4.2 Future directions

The findings reported in this thesis opened the way to 4 future directions, 3 of which are related to TAK-243 and the 4th is related to evaluating adenosine sulfamates recently developed against the related SAE enzyme in AML. These directions include: 1. Mechanistic insights into BEND3-mediated regulation of BCRP expression 2. The role of KMT5B/C in TAK-243 resistance 3. Correlative biomarkers of TAK-243 response 4. SAE as a potential therapeutic target in AML In this section, we briefly highlight the preliminary data we have for each of these directions and the experimental plan to tackle the remaining open questions.

4.2.1 Mechanistic insights into BEND3-mediated regulation of BCRP expression

BEND3 is a transcriptional repressor that contains four BEN domains and regulates gene expression and chromatin organization by interacting with chromatin-remodeling complexes382, 406. In this respect, it was reported to associate with the nucleolar-remodeling complex (NoRC) inducing repression of rDNA transcription383. Another complex with which BEND3 was reported to associate is the nucleosome remodeling and deacetylase (NuRD) complex to trigger switch from constitutive (DNA hypermethylation and increased H3K9Me3) to facultative heterochromatin (DNA hypomethylation and increased H3K27Me3)407. It has been proposed to interact with the polycomb repressive complex 2 (PRC2) complex to induce such a switch382, 407. BEND3 may also play a role in genome stability and mitosis as it was reported to associate with the ATP-dependent DNA translocase PICH408, 409. Based on these reports, BEND3 may transcriptionally repress several proteins including BCRP either directly or indirectly. To investigate this possibility, we plan to assess histone and DNA methylation status in control versus BEND3 knockout cells. Methylation changes in the promoter of ABCG2 were reported to mediate expression levels of BCRP and thus response to several chemotherapeutic agents in MM400. Therefore, we will also assess promoter methylation of ABCG2 after BEND3 knockout. BEND3 may indirectly control the response to TAK-243 by altering the expression of other signaling molecules that modulate BCRP expression. In this respect, we plan to conduct RNAseq in order to identify transcriptional changes in response to

179

BEND3 knockout. These experiments may provide an insight into the mechanism by which BEND3 alters BCRP expression.

4.2.2 The role of KMT5B/C in TAK-243 resistance

Our CRISPR/Cas9 knockout screen has identified KMT5B as a second hit after BEND3. KMT5B is a histone methyltransferase involved in the methylation of H4K20 (Ref. 410). In addition to KMT5B, H4K20 is also methylated by KMT5C and SETD8 (Ref. 411, 412). H4K20 methylation has been reported to be functionally important in transcriptional regulation, DNA repair, and heterochromatin organization413-415. As highlighted above in Section 3.4, KMT5C has also been reported to interact with BEND3 in co-immunoprecipitation assays382. As per our preliminary data, KMT5B knockout conferred resistance to TAK-243, but to a lesser extent compared to BEND3 knockout. We hypothesized that combinatorial inactivation of KMT5B and BEND3 will result in synergistic/additive resistance to TAK-243. To do so, we treated BEND3 knockout OCI-AML2 cells with A-196, a chemical inhibitor of KMT5B/C416. Interestingly, A-196 paradoxically sensitized both control and BEND3 knockout cells to TAK- 243. This potentiation was associated with reversal of the effects induced by BEND3 knockout on TAK-243 transport into the cells and subsequent binding to UBA1. A-196 also potentiated TAK-243 effects on cell lines intrinsically resistant to the drug, with no cross-potentiation of bortezomib, pevonedistat or ML-792. These paradoxical findings may be explained by the dual targeting of KMT5B and C and potential antagonistic roles of these histone methyltransferases. In addition, the epigenetic changes induced by these enzymes may result in changes in the expression of an MDR transporter that is distinct from BCRP and is also important for TAK-243 transport. We plan to investigate these hypotheses in our future work mainly by individual and combinatorial loss-of-function studies of KMT5B and C. We also plan to explore the potential of A-196 to modulate drug transport independent of its epigenetic effects.

4.2.3 Correlative biomarkers of TAK-243 response

TAK-243 has been advanced to a phase 1 clinical trial in AML, and one of the secondary objectives is to identify correlative response biomarkers that can be used to assess and monitor the response to TAK-243. As highlighted in Section 1.6.5, several approaches can be used for pharmacodynamic evaluation of TAK-243 response. We explored the development of quantitative

180 flow-cytometry based assays for potential use in this clinical trial. For this purpose, we started testing three antibodies raised against TAK-243-ubiquitin adduct (MIL90), mono- and poly- ubiquitylated proteins (pan-ub) and γH2AX. We obtained MIL90 antibody from Takeda and optimized the antibody first in IHC-based assays. We have already tested the anti-pan-ub antibody in IHC-based assays as highlighted in Section 2.3.3. We also optimized both antibodies using OCI-AML2 cells treated with TAK-243. We then tested 4 samples of peripheral blood collected from healthy donors treated with increasing concentrations of TAK-243 for 4h. We observed a concentration-dependent reduction in pan-ub levels up to approximately 50% in whole blood and different cell populations. In contrast, we observed a 2.5-fold increase in the MIL90 signal that plateaued at 50 nM. The variability among normal samples as assessed by the % coefficient of variation did not exceed 20% at most data points, suggesting the inter-sample variability for these two pharmacodynamic biomarkers are within acceptable ranges. We also observed an increase in γH2AX after TAK-243 treatment of OCI-AML2 cells as was the case with daunorubicin and Ara-C used as positive controls. However, changes in γH2AX are not direct biological effects that are specific to TAK-243; therefore, we decided to pursue our investigation with MIL90 and pan-ub. Further work is planned to fully assess different measures of analytical validity and clinical utility of these 2 assays in normal and patient samples. These measures include sample collection and processing, assay range, drug kinetics, precision (intra- and inter-donor reproducibility), robustness, and statistical modeling to determine cutoff levels of true drug effects.

4.2.4 SAE as a potential therapeutic target in AML

As highlighted before in Section 1.5.1, SUMOylation is dysregulated in several malignancies including AML. While hyperSUMOylation is required for the response of APL to differentiation therapy, induction of deSUMOylation has been therapeutically exploited in non-APL AML (Box 1)417. Inhibition of SUMOylation at the level of SAE or UBC9 using genetic or chemical approaches has been reported to sensitize non-APL AML to differentiation therapy and chemoresistant AML to conventional chemotherapy236, 240, 418. This therapeutic strategy has also been evaluated in vitro and in vivo in other malignancies224, 252. Takeda has developed a clinical SAE inhibitor, TAK-981, which is currently evaluated in phase 1 clinical trials (Table 1.8 and Section 1.5.2). Based on these reports, we sought to preclinically evaluate SAE as a therapeutic target in AML. As per our preliminary data, shRNA-mediated

181 knockdown of UBA2 in 3 cell lines reduced abundance of SUMOylated proteins and proliferation of AML cells. In addition, pharmacologic inhibition of SAE using ML-792 and TAK-981 induced cytotoxicity of AML cells, and TAK-981 was 2-3-fold more potent than ML-792. Further work is planned to fully evaluate TAK-981 in preclinical models of AML.

References

1. Bedford, L., Lowe, J., Dick, L.R., Mayer, R.J. & Brownell, J.E. Ubiquitin-like protein conjugation and the ubiquitin-proteasome system as drug targets. Nat Rev Drug Discov 10, 29-46 (2011).

2. Rape, M. Ubiquitylation at the crossroads of development and disease. Nat Rev Mol Cell Biol 19, 59-70 (2018).

3. Goldstein, G. et al. Isolation of a polypeptide that has lymphocyte-differentiating properties and is probably represented universally in living cells. Proc Natl Acad Sci U S A 72, 11-5 (1975).

4. Hochstrasser, M. Origin and function of ubiquitin-like proteins. Nature 458, 422-9 (2009).

5. van der Veen, A.G. & Ploegh, H.L. Ubiquitin-like proteins. Annu Rev Biochem 81, 323- 57 (2012).

6. Cappadocia, L. & Lima, C.D. Ubiquitin-like Protein Conjugation: Structures, Chemistry, and Mechanism. Chem Rev 118, 889-918 (2018).

7. Nalepa, G., Rolfe, M. & Harper, J.W. Drug discovery in the ubiquitin-proteasome system. Nat Rev Drug Discov 5, 596-613 (2006).

8. Schulman, B.A. & Harper, J.W. Ubiquitin-like protein activation by E1 enzymes: the apex for downstream signalling pathways. Nat Rev Mol Cell Biol 10, 319-31 (2009).

9. Nakayama, K.I. & Nakayama, K. Ubiquitin ligases: cell-cycle control and cancer. Nat Rev Cancer 6, 369-81 (2006).

10. Clague, M.J., Heride, C. & Urbe, S. The demographics of the ubiquitin system. Trends Cell Biol 25, 417-26 (2015).

11. Ronau, J.A., Beckmann, J.F. & Hochstrasser, M. Substrate specificity of the ubiquitin and Ubl proteases. Cell Res 26, 441-56 (2016).

12. Mattern, M.R., Eddins, M.J., Agarwal, S., Sterner, D.E. & Kodrasov, M.P. in Resistance to Proteasome Inhibitors in Cancer (ed. Dou, Q.P.) (Springer International Publishing Switzerland, 2014).

13. Huang, D.T., Walden, H., Duda, D. & Schulman, B.A. Ubiquitin-like protein activation. Oncogene 23, 1958-71 (2004).

14. Lv, Z. et al. Domain alternation and active site remodeling are conserved structural features of ubiquitin E1. J Biol Chem 292, 12089-12099 (2017).

182 183

15. Termathe, M. & Leidel, S.A. The Uba4 domain interplay is mediated via a thioester that is critical for tRNA thiolation through Urm1 thiocarboxylation. Nucleic Acids Res 46, 5171-5181 (2018).

16. Stewart, M.D., Ritterhoff, T., Klevit, R.E. & Brzovic, P.S. E2 enzymes: more than just middle men. Cell Res 26, 423-40 (2016).

17. Haas, A.L. & Rose, I.A. The mechanism of ubiquitin activating enzyme. A kinetic and equilibrium analysis. J Biol Chem 257, 10329-37 (1982).

18. Olsen, S.K., Capili, A.D., Lu, X., Tan, D.S. & Lima, C.D. Active site remodelling accompanies thioester bond formation in the SUMO E1. Nature 463, 906-12 (2010).

19. Streich, F.C., Jr. & Lima, C.D. Structural and functional insights to ubiquitin-like protein conjugation. Annu Rev Biophys 43, 357-79 (2014).

20. Capdeville, R., Buchdunger, E., Zimmermann, J. & Matter, A. Glivec (STI571, imatinib), a rationally developed, targeted anticancer drug. Nat Rev Drug Discov 1, 493-502 (2002).

21. Cohen, P. Protein kinases--the major drug targets of the twenty-first century? Nat Rev Drug Discov 1, 309-15 (2002).

22. Hubbard, S.R. & Till, J.H. Protein tyrosine kinase structure and function. Annu Rev Biochem 69, 373-98 (2000).

23. Wilkinson, K.D. et al. A specific inhibitor of the ubiquitin activating enzyme: synthesis and characterization of adenosyl-phospho-ubiquitinol, a nonhydrolyzable ubiquitin adenylate analogue. Biochemistry 29, 7373-80 (1990).

24. Ciavarri, J. & Langston, S. in Comprehensive Medicinal Chemistry III (eds. Rotella, D. & Ward, S.E.) 95-112 (Elsevier, Oxford, 2017).

25. Visscher, M., Arkin, M.R. & Dansen, T.B. Covalent targeting of acquired cysteines in cancer. Curr Opin Chem Biol 30, 61-67 (2016).

26. Yang, Y. et al. Inhibitors of ubiquitin-activating enzyme (E1), a new class of potential cancer therapeutics. Cancer Res 67, 9472-81 (2007).

27. Xu, G.W. et al. The ubiquitin-activating enzyme E1 as a therapeutic target for the treatment of leukemia and multiple myeloma. Blood 115, 2251-9 (2010).

28. Lv, Z. et al. Molecular mechanism of a covalent allosteric inhibitor of SUMO E1 activating enzyme. Nat Commun 9, 5145 (2018).

29. Li, Y.J. et al. Allosteric Inhibition of Ubiquitin-like Modifications by a Class of Inhibitor of SUMO-Activating Enzyme. Cell Chem Biol (2018).

30. Jin, L., Wang, W. & Fang, G. Targeting protein-protein interaction by small molecules. Annu Rev Pharmacol Toxicol 54, 435-56 (2014).

184

31. Vlieghe, P., Lisowski, V., Martinez, J. & Khrestchatisky, M. Synthetic therapeutic peptides: science and market. Drug Discov Today 15, 40-56 (2010).

32. Huang, D.T. et al. A unique E1-E2 interaction required for optimal conjugation of the ubiquitin-like protein NEDD8. Nat Struct Mol Biol 11, 927-35 (2004).

33. Weinstein, I.B. & Joe, A.K. Mechanisms of disease: Oncogene addiction--a rationale for molecular targeting in cancer therapy. Nat Clin Pract Oncol 3, 448-57 (2006).

34. Luo, J., Solimini, N.L. & Elledge, S.J. Principles of cancer therapy: oncogene and non- oncogene addiction. Cell 136, 823-37 (2009).

35. Dobbelstein, M. & Moll, U. Targeting tumour-supportive cellular machineries in anticancer drug development. Nat Rev Drug Discov 13, 179-96 (2014).

36. Hoeller, D. & Dikic, I. Targeting the ubiquitin system in cancer therapy. Nature 458, 438-44 (2009).

37. Manasanch, E.E. & Orlowski, R.Z. Proteasome inhibitors in cancer therapy. Nat Rev Clin Oncol 14, 417-433 (2017).

38. Groen, E.J.N. & Gillingwater, T.H. UBA1: At the Crossroads of Ubiquitin Homeostasis and Neurodegeneration. Trends Mol Med 21, 622-632 (2015).

39. Ulrich, H.D. & Walden, H. Ubiquitin signalling in DNA replication and repair. Nat Rev Mol Cell Biol 11, 479-89 (2010).

40. Seeler, J.S. & Dejean, A. SUMO and the robustness of cancer. Nat Rev Cancer 17, 184- 197 (2017).

41. Gareau, J.R. & Lima, C.D. The SUMO pathway: emerging mechanisms that shape specificity, conjugation and recognition. Nat Rev Mol Cell Biol 11, 861-71 (2010).

42. Soucy, T.A. et al. An inhibitor of NEDD8-activating enzyme as a new approach to treat cancer. Nature 458, 732-6 (2009).

43. Enchev, R.I., Schulman, B.A. & Peter, M. Protein neddylation: beyond cullin-RING ligases. Nat Rev Mol Cell Biol 16, 30-44 (2015).

44. Watson, I.R., Irwin, M.S. & Ohh, M. NEDD8 pathways in cancer, Sine Quibus Non. Cancer Cell 19, 168-76 (2011).

45. Leidecker, O., Matic, I., Mahata, B., Pion, E. & Xirodimas, D.P. The ubiquitin E1 enzyme Ube1 mediates NEDD8 activation under diverse stress conditions. Cell Cycle 11, 1142-50 (2012).

46. Lydeard, J.R., Schulman, B.A. & Harper, J.W. Building and remodelling Cullin-RING E3 ubiquitin ligases. EMBO Rep 14, 1050-61 (2013).

185

47. Soucy, T.A., Smith, P.G. & Rolfe, M. Targeting NEDD8-activated cullin-RING ligases for the treatment of cancer. Clin Cancer Res 15, 3912-6 (2009).

48. Brown, J.S. et al. Neddylation promotes ubiquitylation and release of Ku from DNA- damage sites. Cell Rep 11, 704-14 (2015).

49. Mathewson, N. et al. Neddylation plays an important role in the regulation of murine and human dendritic cell function. Blood 122, 2062-73 (2013).

50. Mathewson, N.D. et al. SAG/Rbx2-Dependent Neddylation Regulates T-Cell Responses. Am J Pathol 186, 2679-91 (2016).

51. Balachandran, R.S. et al. The ubiquitin ligase CRL2ZYG11 targets cyclin B1 for degradation in a conserved pathway that facilitates mitotic slippage. J Cell Biol 215, 151- 166 (2016).

52. Soucy, T.A., Dick, L.R., Smith, P.G., Milhollen, M.A. & Brownell, J.E. The NEDD8 Conjugation Pathway and Its Relevance in Cancer Biology and Therapy. Genes Cancer 1, 708-16 (2010).

53. Yu, J. et al. Overactivated neddylation pathway in human hepatocellular carcinoma. Cancer Med (2018).

54. Uhlen, M. et al. A pathology atlas of the human cancer transcriptome. Science 357 (2017).

55. Barbier-Torres, L. et al. Stabilization of LKB1 and Akt by neddylation regulates energy metabolism in liver cancer. Oncotarget 6, 2509-23 (2015).

56. Li, L. et al. Overactivated neddylation pathway as a therapeutic target in lung cancer. J Natl Cancer Inst 106, dju083 (2014).

57. Hua, W. et al. Suppression of glioblastoma by targeting the overactivated protein neddylation pathway. Neuro Oncol 17, 1333-43 (2015).

58. Gao, Q. et al. Neddylation pathway is up-regulated in human intrahepatic cholangiocarcinoma and serves as a potential therapeutic target. Oncotarget 5, 7820-32 (2014).

59. Huang, J. et al. NEDD8 Inhibition Overcomes CKS1B-Induced Drug Resistance by Upregulation of p21 in Multiple Myeloma. Clin Cancer Res 21, 5532-42 (2015).

60. Chen, P. et al. Neddylation Inhibition Activates the Extrinsic Apoptosis Pathway through ATF4-CHOP-DR5 Axis in Human Esophageal Cancer Cells. Clin Cancer Res 22, 4145- 57 (2016).

61. Jin, Y. et al. Neddylation Blockade Diminishes Hepatic Metastasis by Dampening Cancer Stem-Like Cells and Angiogenesis in Uveal Melanoma. Clin Cancer Res (2017).

186

62. Li, H. et al. Inhibition of Neddylation Modification Sensitizes Pancreatic Cancer Cells to Gemcitabine. Neoplasia 19, 509-518 (2017).

63. Li, J.A. et al. Chk1 inhibitor SCH 900776 enhances the antitumor activity of MLN4924 on pancreatic cancer. Cell Cycle 17, 191-199 (2018).

64. Tong, S. et al. MLN4924 (Pevonedistat), a protein neddylation inhibitor, suppresses proliferation and migration of human clear cell renal cell carcinoma. Sci Rep 7, 5599 (2017).

65. Liu, C. et al. Antitumor Effects of Blocking Protein Neddylation in T315I-BCR-ABL Leukemia Cells and Leukemia Stem Cells. Cancer Res 78, 1522-1536 (2018).

66. Tate, J.G. et al. COSMIC: the Catalogue Of Somatic Mutations In Cancer. Nucleic Acids Res 47, D941-D947 (2019).

67. Walden, H., Podgorski, M.S. & Schulman, B.A. Insights into the ubiquitin transfer cascade from the structure of the activating enzyme for NEDD8. Nature 422, 330-4 (2003).

68. Walden, H. et al. The structure of the APPBP1-UBA3-NEDD8-ATP complex reveals the basis for selective ubiquitin-like protein activation by an E1. Mol Cell 12, 1427-37 (2003).

69. Huang, D.T. et al. Basis for a ubiquitin-like protein thioester switch toggling E1-E2 affinity. Nature 445, 394-8 (2007).

70. Brownell, J.E. et al. Substrate-assisted inhibition of ubiquitin-like protein-activating enzymes: the NEDD8 E1 inhibitor MLN4924 forms a NEDD8-AMP mimetic in situ. Mol Cell 37, 102-11 (2010).

71. Ying, J., Zhang, M., Qiu, X. & Lu, Y. Targeting the neddylation pathway in cells as a potential therapeutic approach for diseases. Cancer Chemother Pharmacol 81, 797-808 (2018).

72. Brown, P. et al. Molecular recognition of tyrosinyl adenylate analogues by prokaryotic tyrosyl tRNA synthetases. Bioorg Med Chem 7, 2473-85 (1999).

73. Forrest, A.K. et al. Aminoalkyl adenylate and aminoacyl sulfamate intermediate analogues differing greatly in affinity for their cognate Staphylococcus aureus aminoacyl tRNA synthetases. Bioorg Med Chem Lett 10, 1871-4 (2000).

74. Lux, M.C., Standke, L.C. & Tan, D.S. Targeting adenylate-forming enzymes with designed sulfonyladenosine inhibitors. J Antibiot (Tokyo) 72, 325-349 (2019).

75. Milhollen, M.A. et al. Inhibition of NEDD8-activating enzyme induces rereplication and apoptosis in human tumor cells consistent with deregulating CDT1 turnover. Cancer Res 71, 3042-51 (2011).

187

76. Lin, J.J., Milhollen, M.A., Smith, P.G., Narayanan, U. & Dutta, A. NEDD8-targeting drug MLN4924 elicits DNA rereplication by stabilizing Cdt1 in S phase, triggering checkpoint activation, apoptosis, and senescence in cancer cells. Cancer Res 70, 10310- 20 (2010).

77. Petropoulos, M., Champeris Tsaniras, S., Taraviras, S. & Lygerou, Z. Replication Licensing Aberrations, Replication Stress, and Genomic Instability. Trends Biochem Sci (2019).

78. Hyer, M.L. et al. A small-molecule inhibitor of the ubiquitin activating enzyme for cancer treatment. Nat Med 24, 186-193 (2018).

79. Leung, C.H. et al. A natural product-like inhibitor of NEDD8-activating enzyme. Chem Commun (Camb) 47, 2511-3 (2011).

80. Li, X. et al. Flavokawain A induces deNEDDylation and Skp2 degradation leading to inhibition of tumorigenesis and cancer progression in the TRAMP transgenic mouse model. Oncotarget 6, 41809-24 (2015).

81. Zi, X. & Simoneau, A.R. Flavokawain A, a novel chalcone from kava extract, induces apoptosis in bladder cancer cells by involvement of Bax protein-dependent and mitochondria-dependent apoptotic pathway and suppresses tumor growth in mice. Cancer Res 65, 3479-86 (2005).

82. Liu, Z. et al. Kava chalcone, flavokawain A, inhibits urothelial tumorigenesis in the UPII- SV40T transgenic mouse model. Cancer Prev Res (Phila) 6, 1365-75 (2013).

83. Lu, P. et al. Discovery of a novel NEDD8 Activating Enzyme Inhibitor with Piperidin-4- amine Scaffold by Structure-Based Virtual Screening. ACS Chem Biol 11, 1901-7 (2016).

84. Zhong, H.J. et al. Structure-based repurposing of FDA-approved drugs as inhibitors of NEDD8-activating enzyme. Biochimie 102, 211-5 (2014).

85. Zhong, H.J. et al. A metal-based inhibitor of NEDD8-activating enzyme. PLoS One 7, e49574 (2012).

86. Zhang, S. et al. Effective virtual screening strategy toward covalent ligands: identification of novel NEDD8-activating enzyme inhibitors. J Chem Inf Model 54, 1785-97 (2014).

87. Ma, H. et al. Discovery of benzothiazole derivatives as novel non-sulfamide NEDD8 activating enzyme inhibitors by target-based virtual screening. Eur J Med Chem 133, 174-183 (2017).

88. Zhong, H.J. et al. Discovery of a natural product inhibitor targeting protein neddylation by structure-based virtual screening. Biochimie 94, 2457-60 (2012).

89. Zhao, B. et al. Phage display to identify Nedd8-mimicking peptides as inhibitors of the Nedd8 transfer cascade. Chembiochem 14, 1323-30 (2013).

188

90. Zhao, B. et al. Profiling the cross reactivity of ubiquitin with the Nedd8 activating enzyme by phage display. PLoS One 8, e70312 (2013).

91. Buac, D. et al. From bortezomib to other inhibitors of the proteasome and beyond. Curr Pharm Des 19, 4025-38 (2013).

92. Swords, R.T. et al. Inhibition of NEDD8-activating enzyme: a novel approach for the treatment of acute myeloid leukemia. Blood 115, 3796-800 (2010).

93. Zheng, W. et al. Neural precursor cell expressed, developmentally downregulated 8activating enzyme inhibitor MLN4924 sensitizes colorectal cancer cells to oxaliplatin by inducing DNA damage, G2 cell cycle arrest and apoptosis. Mol Med Rep 15, 2795-2801 (2017).

94. Guo, Z.P. et al. MLN4924 suppresses the BRCA1 complex and synergizes with PARP inhibition in NSCLC cells. Biochem Biophys Res Commun 483, 223-229 (2017).

95. Zhang, Q. et al. The novel protective role of P27 in MLN4924-treated gastric cancer cells. Cell Death Dis 6, e1867 (2015).

96. Milhollen, M.A. et al. MLN4924, a NEDD8-activating enzyme inhibitor, is active in diffuse large B-cell lymphoma models: rationale for treatment of NF-{kappa}B- dependent lymphoma. Blood 116, 1515-23 (2010).

97. Khalife, J. et al. Pharmacological targeting of miR-155 via the NEDD8-activating enzyme inhibitor MLN4924 (Pevonedistat) in FLT3-ITD acute myeloid leukemia. Leukemia 29, 1981-92 (2015).

98. Godbersen, J.C. et al. The Nedd8-activating enzyme inhibitor MLN4924 thwarts microenvironment-driven NF-kappaB activation and induces apoptosis in chronic lymphocytic leukemia B cells. Clin Cancer Res 20, 1576-89 (2014).

99. Li, H. et al. Inactivation of SAG/RBX2 E3 ubiquitin ligase suppresses KrasG12D-driven lung tumorigenesis. J Clin Invest 124, 835-46 (2014).

100. Dengler, M.A. et al. Discrepant NOXA (PMAIP1) transcript and NOXA protein levels: a potential Achilles' heel in mantle cell lymphoma. Cell Death Dis 5, e1013 (2014).

101. Knorr, K.L. et al. MLN4924 induces Noxa upregulation in acute myelogenous leukemia and synergizes with Bcl-2 inhibitors. Cell Death Differ 22, 2133-42 (2015).

102. Wang, J. et al. Targeting neddylation pathway with MLN4924 (Pevonedistat) induces NOXA-dependent apoptosis in renal cell carcinoma. Biochem Biophys Res Commun 490, 1183-1188 (2017).

103. Yang, D., Tan, M., Wang, G. & Sun, Y. The p21-dependent radiosensitization of human breast cancer cells by MLN4924, an investigational inhibitor of NEDD8 activating enzyme. PLoS One 7, e34079 (2012).

189

104. Blank, J.L. et al. Novel DNA damage checkpoints mediating cell death induced by the NEDD8-activating enzyme inhibitor MLN4924. Cancer Res 73, 225-34 (2013).

105. Wang, X. et al. Radiosensitization by the investigational NEDD8-activating enzyme inhibitor MLN4924 (pevonedistat) in hormone-resistant prostate cancer cells. Oncotarget 7, 38380-38391 (2016).

106. Mackintosh, C. et al. WEE1 accumulation and deregulation of S-phase proteins mediate MLN4924 potent inhibitory effect on Ewing sarcoma cells. Oncogene 32, 1441-51 (2013).

107. Gu, Y. et al. MLN4924, an NAE inhibitor, suppresses AKT and mTOR signaling via upregulation of REDD1 in human myeloma cells. Blood 123, 3269-76 (2014).

108. Kuo, K.L. et al. MLN4924, a novel protein neddylation inhibitor, suppresses proliferation and migration of human urothelial carcinoma: In vitro and in vivo studies. Cancer Lett 363, 127-36 (2015).

109. Leclerc, G.M. et al. The NEDD8-activating enzyme inhibitor pevonedistat activates the eIF2alpha and mTOR pathways inducing UPR-mediated cell death in acute lymphoblastic leukemia. Leuk Res 50, 1-10 (2016).

110. Paiva, C. et al. Pevonedistat, a Nedd8-activating enzyme inhibitor, sensitizes neoplastic B-cells to death receptor-mediated apoptosis. Oncotarget 8, 21128-21139 (2017).

111. Wang, Y. et al. Targeting protein neddylation with an NEDD8-activating enzyme inhibitor MLN4924 induced apoptosis or senescence in human lymphoma cells. Cancer Biol Ther 16, 420-9 (2015).

112. Jia, L., Li, H. & Sun, Y. Induction of p21-dependent senescence by an NAE inhibitor, MLN4924, as a mechanism of growth suppression. Neoplasia 13, 561-9 (2011).

113. Benamar, M. et al. Inactivation of the CRL4-CDT2-SET8/p21 ubiquitylation and degradation axis underlies the therapeutic efficacy of pevonedistat in melanoma. EBioMedicine 10, 85-100 (2016).

114. Yao, W.T. et al. Suppression of tumor angiogenesis by targeting the protein neddylation pathway. Cell Death Dis 5, e1059 (2014).

115. Lan, H., Tang, Z., Jin, H. & Sun, Y. Neddylation inhibitor MLN4924 suppresses growth and migration of human gastric cancer cells. Sci Rep 6, 24218 (2016).

116. Park, S.Y., Park, J.W., Lee, G.W., Li, L. & Chun, Y.S. Inhibition of neddylation facilitates cell migration through enhanced phosphorylation of caveolin-1 in PC3 and U373MG cells. BMC Cancer 18, 30 (2018).

117. Kee, Y. et al. Inhibition of the Nedd8 system sensitizes cells to DNA interstrand cross- linking agents. Mol Cancer Res 10, 369-77 (2012).

190

118. Garcia, K. et al. Nedd8-activating enzyme inhibitor MLN4924 provides synergy with mitomycin C through interactions with ATR, BRCA1/BRCA2, and chromatin dynamics pathways. Mol Cancer Ther 13, 1625-35 (2014).

119. Jazaeri, A.A. et al. Overcoming platinum resistance in preclinical models of ovarian cancer using the neddylation inhibitor MLN4924. Mol Cancer Ther 12, 1958-67 (2013).

120. Czuczman, N.M. et al. Pevonedistat, a NEDD8-activating enzyme inhibitor, is active in mantle cell lymphoma and enhances rituximab activity in vivo. Blood 127, 1128-37 (2016).

121. Nawrocki, S.T. et al. Disrupting protein NEDDylation with MLN4924 is a novel strategy to target cisplatin resistance in ovarian cancer. Clin Cancer Res 19, 3577-90 (2013).

122. Xu, Q. et al. MLN4924 neddylation inhibitor promotes cell death in paclitaxel-resistant human lung adenocarcinoma cells. Oncol Lett 15, 515-521 (2018).

123. Ho, I.L. et al. MLN4924 Synergistically Enhances Cisplatin-induced Cytotoxicity via JNK and Bcl-xL Pathways in Human Urothelial Carcinoma. Sci Rep 5, 16948 (2015).

124. Xie, P. et al. Promoting tumorigenesis in nasopharyngeal carcinoma, NEDD8 serves as a potential theranostic target. Cell Death Dis 8, e2834 (2017).

125. Vanderdys, V. et al. The Neddylation Inhibitor Pevonedistat (MLN4924) Suppresses and Radiosensitizes Head and Neck Squamous Carcinoma Cells and Tumors. Mol Cancer Ther 17, 368-380 (2018).

126. Wei, D. et al. Radiosensitization of human pancreatic cancer cells by MLN4924, an investigational NEDD8-activating enzyme inhibitor. Cancer Res 72, 282-93 (2012).

127. Tan, M., Li, Y., Yang, R., Xi, N. & Sun, Y. Inactivation of SAG E3 ubiquitin ligase blocks embryonic stem cell differentiation and sensitizes leukemia cells to retinoid acid. PLoS One 6, e27726 (2011).

128. Visconte, V. et al. Comprehensive quantitative proteomic profiling of the pharmacodynamic changes induced by MLN4924 in acute myeloid leukemia cells establishes rationale for its combination with azacitidine. Leukemia 30, 1190-4 (2016).

129. Zhou, L. et al. The NAE inhibitor pevonedistat interacts with the HDAC inhibitor belinostat to target AML cells by disrupting the DDR. Blood 127, 2219-30 (2016).

130. Liu, J. et al. A genome-scale CRISPR-Cas9 screening in myeloma cells identifies regulators of immunomodulatory drug sensitivity. Leukemia 33, 171-180 (2019).

131. Zhou, S. et al. Neddylation inhibition upregulates PD-L1 expression and enhances the efficacy of immune checkpoint blockade in glioblastoma. Int J Cancer (2019).

191

132. Sumi, H. et al. An inhibitor of apoptosis protein antagonist T-3256336 potentiates the antitumor efficacy of the Nedd8-activating enzyme inhibitor pevonedistat (TAK- 924/MLN4924). Biochem Biophys Res Commun 480, 380-386 (2016).

133. Ishikawa, Y. et al. Synergistic anti-AML effects of the LSD1 inhibitor T-3775440 and the NEDD8-activating enzyme inhibitor pevonedistat via transdifferentiation and DNA rereplication. Oncogenesis 6, e377 (2017).

134. Chen, P. et al. Synergistic inhibition of autophagy and neddylation pathways as a novel therapeutic approach for targeting liver cancer. Oncotarget 6, 9002-17 (2015).

135. Cooper, J. et al. Combined Inhibition of NEDD8-Activating Enzyme and mTOR Suppresses NF2 Loss-Driven Tumorigenesis. Mol Cancer Ther 16, 1693-1704 (2017).

136. Luo, Z. et al. The Nedd8-activating enzyme inhibitor MLN4924 induces autophagy and apoptosis to suppress liver cancer cell growth. Cancer Res 72, 3360-71 (2012).

137. Zhao, Y., Xiong, X., Jia, L. & Sun, Y. Targeting Cullin-RING ligases by MLN4924 induces autophagy via modulating the HIF1-REDD1-TSC1-mTORC1-DEPTOR axis. Cell Death Dis 3, e386 (2012).

138. Shah, J.J. et al. Phase I Study of the Novel Investigational NEDD8-Activating Enzyme Inhibitor Pevonedistat (MLN4924) in Patients with Relapsed/Refractory Multiple Myeloma or Lymphoma. Clin Cancer Res 22, 34-43 (2016).

139. Nawrocki, S.T. et al. The NEDD8-activating enzyme inhibitor MLN4924 disrupts nucleotide metabolism and augments the efficacy of cytarabine. Clin Cancer Res 21, 439- 47 (2015).

140. Swords, R.T. et al. Pevonedistat (MLN4924), a First-in-Class NEDD8-activating enzyme inhibitor, in patients with acute myeloid leukaemia and myelodysplastic syndromes: a phase 1 study. Br J Haematol 169, 534-43 (2015).

141. Swords, R.T. et al. Expanded safety analysis of pevonedistat, a first-in-class NEDD8- activating enzyme inhibitor, in patients with acute myeloid leukemia and myelodysplastic syndromes. Blood Cancer J 7, e520 (2017).

142. Swords, R.T. et al. Pevonedistat, a first-in-class NEDD8-activating enzyme inhibitor, combined with azacitidine in patients with AML. Blood 131, 1415-1424 (2018).

143. Picco, G. et al. Efficacy of NEDD8 Pathway Inhibition in Preclinical Models of Poorly Differentiated, Clinically Aggressive Colorectal Cancer. J Natl Cancer Inst 109 (2017).

144. Malhab, L.J., Descamps, S., Delaval, B. & Xirodimas, D.P. The use of the NEDD8 inhibitor MLN4924 (Pevonedistat) in a cyclotherapy approach to protect wild-type p53 cells from MLN4924 induced toxicity. Sci Rep 6, 37775 (2016).

192

145. Sarantopoulos, J. et al. Phase I Study of the Investigational NEDD8-Activating Enzyme Inhibitor Pevonedistat (TAK-924/MLN4924) in Patients with Advanced Solid Tumors. Clin Cancer Res 22, 847-57 (2016).

146. Lockhart, A.C. et al. Phase Ib study of pevonedistat, a NEDD8-activating enzyme inhibitor, in combination with docetaxel, carboplatin and paclitaxel, or gemcitabine, in patients with advanced solid tumors. Invest New Drugs (2018).

147. Zhao, L., Yue, P., Lonial, S., Khuri, F.R. & Sun, S.Y. The NEDD8-activating enzyme inhibitor, MLN4924, cooperates with TRAIL to augment apoptosis through facilitating c- FLIP degradation in head and neck cancer cells. Mol Cancer Ther 10, 2415-25 (2011).

148. McMillin, D.W. et al. Molecular and cellular effects of NEDD8-activating enzyme inhibition in myeloma. Mol Cancer Ther 11, 942-51 (2012).

149. Smith, M.A. et al. Initial testing of the investigational NEDD8-activating enzyme inhibitor MLN4924 by the pediatric preclinical testing program. Pediatr Blood Cancer 59, 246-53 (2012).

150. Bhatia, S. et al. A phase I study of the investigational NEDD8-activating enzyme inhibitor pevonedistat (TAK-924/MLN4924) in patients with metastatic melanoma. Invest New Drugs 34, 439-49 (2016).

151. Wong, K.M. et al. Targeting the protein ubiquitination machinery in melanoma by the NEDD8-activating enzyme inhibitor pevonedistat (MLN4924). Invest New Drugs 35, 11- 25 (2017).

152. Pan, W.W. et al. Ubiquitin E3 ligase CRL4(CDT2/DCAF2) as a potential chemotherapeutic target for ovarian surface epithelial cancer. J Biol Chem 288, 29680-91 (2013).

153. Hong, X. et al. Disruption of protein neddylation with MLN4924 attenuates paclitaxel- induced apoptosis and microtubule polymerization in ovarian cancer cells. Biochem Biophys Res Commun (2018).

154. Paiva, C., Godbersen, J.C., Berger, A., Brown, J.R. & Danilov, A.V. Targeting neddylation induces DNA damage and checkpoint activation and sensitizes chronic lymphocytic leukemia B cells to alkylating agents. Cell Death Dis 6, e1807 (2015).

155. Lin, W.C. et al. MLN4924, a Novel NEDD8-activating enzyme inhibitor, exhibits antitumor activity and enhances cisplatin-induced cytotoxicity in human cervical carcinoma: in vitro and in vivo study. Am J Cancer Res 5, 3350-62 (2015).

156. Han, K. et al. The NEDD8-activating enzyme inhibitor MLN4924 induces G2 arrest and apoptosis in T-cell acute lymphoblastic leukemia. Oncotarget 7, 23812-24 (2016).

157. Zheng, S., Leclerc, G.M., Li, B., Swords, R.T. & Barredo, J.C. Inhibition of the NEDD8 conjugation pathway induces calcium-dependent compensatory activation of the pro-

193

survival MEK/ERK pathway in acute lymphoblastic leukemia. Oncotarget 9, 5529-5544 (2018).

158. Rulina, A.V. et al. Distinct outcomes of CRL-Nedd8 pathway inhibition reveal cancer cell plasticity. Cell Death Dis 7, e2505 (2016).

159. Zhang, Y., Shi, C.C., Zhang, H.P., Li, G.Q. & Li, S.S. MLN4924 suppresses neddylation and induces cell cycle arrest, senescence, and apoptosis in human osteosarcoma. Oncotarget 7, 45263-45274 (2016).

160. Zhang, Y. et al. Inhibition of Mcl-1 enhances Pevonedistat-triggered apoptosis in osteosarcoma cells. Exp Cell Res 358, 234-241 (2017).

161. Xu, B. et al. A first-in-class inhibitor, MLN4924 (pevonedistat), induces cell-cycle arrest, senescence, and apoptosis in human renal cell carcinoma by suppressing UBE2M- dependent neddylation modification. Cancer Chemother Pharmacol 81, 1083-1093 (2018).

162. Yoshimura, C. et al. TAS4464, A Highly Potent and Selective Inhibitor of NEDD8- Activating Enzyme, Suppresses Neddylation and Shows Antitumor Activity in Diverse Cancer Models. Mol Cancer Ther 18, 1205-1216 (2019).

163. Lee, I. & Schindelin, H. Structural insights into E1-catalyzed ubiquitin activation and transfer to conjugating enzymes. Cell 134, 268-78 (2008).

164. Jin, J., Li, X., Gygi, S.P. & Harper, J.W. Dual E1 activation systems for ubiquitin differentially regulate E2 enzyme charging. Nature 447, 1135-8 (2007).

165. Chiu, Y.H., Sun, Q. & Chen, Z.J. E1-L2 activates both ubiquitin and FAT10. Mol Cell 27, 1014-23 (2007).

166. Groettrup, M., Pelzer, C., Schmidtke, G. & Hofmann, K. Activating the ubiquitin family: UBA6 challenges the field. Trends Biochem Sci 33, 230-7 (2008).

167. Swatek, K.N. & Komander, D. Ubiquitin modifications. Cell Res 26, 399-422 (2016).

168. Bhoj, V.G. & Chen, Z.J. Ubiquitylation in innate and adaptive immunity. Nature 458, 430-7 (2009).

169. Varshavsky, A. The Ubiquitin System, Autophagy, and Regulated Protein Degradation. Annu Rev Biochem 86, 123-128 (2017).

170. Dikic, I. Proteasomal and Autophagic Degradation Systems. Annu Rev Biochem 86, 193- 224 (2017).

171. McGrath, J.P., Jentsch, S. & Varshavsky, A. UBA 1: an essential yeast gene encoding ubiquitin-activating enzyme. EMBO J 10, 227-36 (1991).

194

172. Kulkarni, M. & Smith, H.E. E1 ubiquitin-activating enzyme UBA-1 plays multiple roles throughout C. elegans development. PLoS Genet 4, e1000131 (2008).

173. UniProt, C. UniProt: a worldwide hub of protein knowledge. Nucleic Acids Res 47, D506-D515 (2019).

174. Senft, D., Qi, J. & Ronai, Z.A. Ubiquitin ligases in oncogenic transformation and cancer therapy. Nat Rev Cancer 18, 69-88 (2018).

175. Fulda, S., Rajalingam, K. & Dikic, I. Ubiquitylation in immune disorders and cancer: from molecular mechanisms to therapeutic implications. EMBO Mol Med 4, 545-56 (2012).

176. Alpi, A.F., Chaugule, V. & Walden, H. Mechanism and disease association of E2- conjugating enzymes: lessons from UBE2T and UBE2L3. Biochem J 473, 3401-3419 (2016).

177. Lv, Z., Williams, K.M., Yuan, L., Atkison, J.H. & Olsen, S.K. Crystal structure of a human ubiquitin E1-ubiquitin complex reveals conserved functional elements essential for activity. J Biol Chem (2018).

178. Szczepanowski, R.H., Filipek, R. & Bochtler, M. Crystal structure of a fragment of mouse ubiquitin-activating enzyme. J Biol Chem 280, 22006-11 (2005).

179. Olsen, S.K. & Lima, C.D. Structure of a ubiquitin E1-E2 complex: insights to E1-E2 thioester transfer. Mol Cell 49, 884-96 (2013).

180. Lv, Z. et al. S. pombe Uba1-Ubc15 Structure Reveals a Novel Regulatory Mechanism of Ubiquitin E2 Activity. Mol Cell 65, 699-714 e6 (2017).

181. Lake, M.W., Wuebbens, M.M., Rajagopalan, K.V. & Schindelin, H. Mechanism of ubiquitin activation revealed by the structure of a bacterial MoeB-MoaD complex. Nature 414, 325-9 (2001).

182. Barghout, S.H. et al. Preclinical evaluation of the selective small-molecule UBA1 inhibitor, TAK-243, in acute myeloid leukemia. Leukemia 33, 37-51 (2019).

183. McHugh, A. et al. Preclinical comparison of proteasome and ubiquitin E1 enzyme inhibitors in cutaneous squamous cell carcinoma: the identification of mechanisms of differential sensitivity. Oncotarget 9, 20265-20281 (2018).

184. Best, S. et al. Targeting ubiquitin-activating enzyme induces ER stress-mediated apoptosis in B-cell lymphoma cells. Blood Adv 3, 51-62 (2019).

185. Zhuang, J. et al. Ubiquitin-activating enzyme inhibition induces an unfolded protein response and overcomes drug resistance in myeloma. Blood (2019).

195

186. Pesiri, V., Totta, P., Marino, M. & Acconcia, F. Ubiquitin-activating enzyme is necessary for 17beta-estradiol-induced breast cancer cell proliferation and migration. IUBMB Life 66, 578-85 (2014).

187. Kapuria, V. et al. Protein cross-linking as a novel mechanism of action of a ubiquitin- activating enzyme inhibitor with anti-tumor activity. Biochem Pharmacol 82, 341-9 (2011).

188. Kitagaki, J. et al. Nitric oxide prodrug JS-K inhibits ubiquitin E1 and kills tumor cells retaining wild-type p53. Oncogene 28, 619-24 (2009).

189. Tan, G. et al. JS-K, a nitric oxide pro-drug, regulates growth and apoptosis through the ubiquitin-proteasome pathway in prostate cancer cells. BMC Cancer 17, 376 (2017).

190. Kiziltepe, T. et al. JS-K, a GST-activated nitric oxide generator, induces DNA double- strand breaks, activates DNA damage response pathways, and induces apoptosis in vitro and in vivo in human multiple myeloma cells. Blood 110, 709-18 (2007).

191. Taori, K., Paul, V.J. & Luesch, H. Structure and activity of largazole, a potent antiproliferative agent from the Floridian marine cyanobacterium Symploca sp. J Am Chem Soc 130, 1806-7 (2008).

192. Ungermannova, D. et al. Largazole and its derivatives selectively inhibit ubiquitin activating enzyme (e1). PLoS One 7, e29208 (2012).

193. Sekizawa, R. et al. Panepophenanthrin, from a mushroom strain, a novel inhibitor of the ubiquitin-activating enzyme. J Nat Prod 65, 1491-3 (2002).

194. Yamanokuchi, R. et al. Hyrtioreticulins A-E, indole alkaloids inhibiting the ubiquitin- activating enzyme, from the marine sponge Hyrtios reticulatus. Bioorg Med Chem 20, 4437-42 (2012).

195. Ungermannova, D. et al. Identification and mechanistic studies of a novel ubiquitin E1 inhibitor. J Biomol Screen 17, 421-34 (2012).

196. Tsukamoto, S. et al. Himeic acid A: a new ubiquitin-activating enzyme inhibitor isolated from a marine-derived fungus, Aspergillus sp. Bioorg Med Chem Lett 15, 191-4 (2005).

197. Lu, X. et al. Designed semisynthetic protein inhibitors of Ub/Ubl E1 activating enzymes. J Am Chem Soc 132, 1748-9 (2010).

198. Ghosh, S.K., Perrine, S.P., Williams, R.M. & Faller, D.V. Histone deacetylase inhibitors are potent inducers of gene expression in latent EBV and sensitize lymphoma cells to nucleoside antiviral agents. Blood 119, 1008-17 (2012).

199. Liu, Y. et al. Anticolon cancer activity of largazole, a marine-derived tunable histone deacetylase inhibitor. J Pharmacol Exp Ther 335, 351-61 (2010).

196

200. Hong, J. & Luesch, H. Largazole: from discovery to broad-spectrum therapy. Nat Prod Rep 29, 449-56 (2012).

201. Law, M.E. et al. Glucocorticoids and histone deacetylase inhibitors cooperate to block the invasiveness of basal-like breast cancer cells through novel mechanisms. Oncogene 32, 1316-29 (2013).

202. Poli, G., Di Fabio, R., Ferrante, L., Summa, V. & Botta, M. Largazole Analogues as Histone Deacetylase Inhibitors and Anticancer Agents: An Overview of Structure- Activity Relationships. ChemMedChem 12, 1917-1926 (2017).

203. Wu, L.C. et al. Largazole Arrests Cell Cycle at G1 Phase and Triggers Proteasomal Degradation of E2F1 in Lung Cancer Cells. ACS Med Chem Lett 4, 921-6 (2013).

204. Vucic, D., Dixit, V.M. & Wertz, I.E. Ubiquitylation in apoptosis: a post-translational modification at the edge of life and death. Nat Rev Mol Cell Biol 12, 439-52 (2011).

205. Misra, M. et al. Dissecting the Specificity of Adenosyl Sulfamate Inhibitors Targeting the Ubiquitin-Activating Enzyme. Structure 25, 1120-1129 e3 (2017).

206. Wang, M. & Kaufman, R.J. The impact of the endoplasmic reticulum protein-folding environment on cancer development. Nat Rev Cancer 14, 581-97 (2014).

207. Milhollen, M.A. et al. Abstract A164: The small molecule UAE inhibitor TAK-243 (MLN7243) prevents DNA damage repair and reduces cell viability/tumor growth when combined with radiation, carboplatin and docetaxel. Molecular Cancer Therapeutics 14, A164-A164 (2015).

208. Majeed, S. et al. Abstract 2699: Targeting an ubiquitin-activating enzyme in small-cell lung cancer (SCLC). Cancer Research 79, 2699-2699 (2019).

209. Ruschak, A.M., Slassi, M., Kay, L.E. & Schimmer, A.D. Novel proteasome inhibitors to overcome bortezomib resistance. J Natl Cancer Inst 103, 1007-17 (2011).

210. Barghout, S.H. & Schimmer, A.D. The ubiquitin-activating enzyme, UBA1, as a novel therapeutic target for AML. Oncotarget 9, 34198-34199 (2018).

211. Barghout, S.H. et al. A Genome-Wide CRISPR/Cas9 Knockout Screen Identifies BEND3 As a Determinant of Sensitivity to UBA1 Inhibition in Acute Myeloid Leukemia. Blood 132, 1350-1350 (2018).

212. Eifler, K. & Vertegaal, A.C.O. SUMOylation-Mediated Regulation of Cell Cycle Progression and Cancer. Trends Biochem Sci 40, 779-793 (2015).

213. Shao, D.F. et al. High-level SAE2 promotes malignant phenotype and predicts outcome in gastric cancer. Am J Cancer Res 5, 140-54 (2015).

197

214. Liu, X. et al. Knockdown of SUMO-activating enzyme subunit 2 (SAE2) suppresses cancer malignancy and enhances chemotherapy sensitivity in small cell lung cancer. J Hematol Oncol 8, 67 (2015).

215. Bellail, A.C., Olson, J.J. & Hao, C. SUMO1 modification stabilizes CDK6 protein and drives the cell cycle and glioblastoma progression. Nat Commun 5, 4234 (2014).

216. Inamura, K. et al. A metastatic signature in entire lung adenocarcinomas irrespective of morphological heterogeneity. Hum Pathol 38, 702-9 (2007).

217. Hoellein, A. et al. Myc-induced SUMOylation is a therapeutic vulnerability for B-cell lymphoma. Blood 124, 2081-90 (2014).

218. Kessler, J.D. et al. A SUMOylation-dependent transcriptional subprogram is required for Myc-driven tumorigenesis. Science 335, 348-53 (2012).

219. Licciardello, M.P. et al. NOTCH1 activation in breast cancer confers sensitivity to inhibition of SUMOylation. Oncogene 34, 3780-90 (2015).

220. Du, L. et al. Role of SUMO activating enzyme in cancer stem cell maintenance and self- renewal. Nat Commun 7, 12326 (2016).

221. Lois, L.M. & Lima, C.D. Structures of the SUMO E1 provide mechanistic insights into SUMO activation and E2 recruitment to E1. EMBO J 24, 439-51 (2005).

222. Wang, J. et al. The intrinsic affinity between E2 and the Cys domain of E1 in ubiquitin- like modifications. Mol Cell 27, 228-37 (2007).

223. Boggio, R., Colombo, R., Hay, R.T., Draetta, G.F. & Chiocca, S. A mechanism for inhibiting the SUMO pathway. Mol Cell 16, 549-61 (2004).

224. He, X. et al. Probing the roles of SUMOylation in cancer cell biology by using a selective SAE inhibitor. Nat Chem Biol 13, 1164-1171 (2017).

225. Kho, C. et al. Small-molecule activation of SERCA2a SUMOylation for the treatment of heart failure. Nat Commun 6, 7229 (2015).

226. Zhao, B. et al. SUMO-mimicking peptides inhibiting protein SUMOylation. Chembiochem 15, 2662-6 (2014).

227. Nakamura, A. et al. Abstract 1523: Inhibition of SUMOylation by TAK-981 induces antitumor innate immune responses by modulating macrophage and NK cell function through Type I IFN pathway activation. Cancer Research 79, 1523-1523 (2019).

228. Berger, A.J. et al. Abstract 3079: Pharmacodynamic evaluation of the novel SUMOylation inhibitor TAK-981 in a mouse tumor model. Cancer Research 79, 3079- 3079 (2019).

198

229. Khattar, M. et al. Abstract 3252: TAK-981: A first in class SUMO inhibitor in Phase 1 trials that promotes dendritic cell activation, antigen-presentation, and T cell priming. Cancer Research 79, 3252-3252 (2019).

230. Hatton, B.A. et al. Abstract 4136: Direct intratumoral microdosing via the CIVO® platform reveals anti-tumor immune responses induced by the SUMO inhibitor TAK-981. Cancer Research 79, 4136-4136 (2019).

231. Takemoto, M. et al. Inhibition of protein SUMOylation by davidiin, an ellagitannin from Davidia involucrata. J Antibiot (Tokyo) 67, 335-8 (2014).

232. Wang, Y. et al. A potential antitumor ellagitannin, davidiin, inhibited hepatocellular tumor growth by targeting EZH2. Tumour Biol 35, 205-12 (2014).

233. Fukuda, I. et al. Ginkgolic acid inhibits protein SUMOylation by blocking formation of the E1-SUMO intermediate. Chem Biol 16, 133-40 (2009).

234. Fukuda, I. et al. Kerriamycin B inhibits protein SUMOylation. J Antibiot (Tokyo) 62, 221-4 (2009).

235. Suzawa, M. et al. A gene-expression screen identifies a non-toxic sumoylation inhibitor that mimics SUMO-less human LRH-1 in liver. Elife 4 (2015).

236. Bossis, G. et al. The ROS/SUMO axis contributes to the response of acute myeloid leukemia cells to chemotherapeutic drugs. Cell Rep 7, 1815-23 (2014).

237. Kumar, A., Ito, A., Hirohama, M., Yoshida, M. & Zhang, K.Y. Identification of sumoylation activating enzyme 1 inhibitors by structure-based virtual screening. J Chem Inf Model 53, 809-20 (2013).

238. Hamdoun, S. & Efferth, T. Ginkgolic acids inhibit migration in breast cancer cells by inhibition of NEMO sumoylation and NF-kappaB activity. Oncotarget 8, 35103-35115 (2017).

239. Sung, B. et al. Anacardic acid (6-nonadecyl salicylic acid), an inhibitor of histone acetyltransferase, suppresses expression of nuclear factor-kappaB-regulated gene products involved in cell survival, proliferation, invasion, and inflammation through inhibition of the inhibitory subunit of nuclear factor-kappaBalpha kinase, leading to potentiation of apoptosis. Blood 111, 4880-91 (2008).

240. Baik, H. et al. Targeting the SUMO Pathway Primes All-trans Retinoic Acid-Induced Differentiation of Nonpromyelocytic Acute Myeloid Leukemias. Cancer Res 78, 2601- 2613 (2018).

241. de The, H., Pandolfi, P.P. & Chen, Z. Acute Promyelocytic Leukemia: A Paradigm for Oncoprotein-Targeted Cure. Cancer Cell 32, 552-560 (2017).

242. Bernardi, R. & Pandolfi, P.P. Structure, dynamics and functions of promyelocytic leukaemia nuclear bodies. Nat Rev Mol Cell Biol 8, 1006-16 (2007).

199

243. de The, H. & Chen, Z. Acute promyelocytic leukaemia: novel insights into the mechanisms of cure. Nat Rev Cancer 10, 775-83 (2010).

244. Sanz, M.A. et al. Management of acute promyelocytic leukemia: updated recommendations from an expert panel of the European LeukemiaNet. Blood 133, 1630- 1643 (2019).

245. Lallemand-Breitenbach, V., Zhu, J., Chen, Z. & de The, H. Curing APL through PML/RARA degradation by As2O3. Trends Mol Med 18, 36-42 (2012).

246. Lo-Coco, F. et al. Retinoic acid and arsenic trioxide for acute promyelocytic leukemia. N Engl J Med 369, 111-21 (2013).

247. Zhang, X.W. et al. Arsenic trioxide controls the fate of the PML-RARalpha oncoprotein by directly binding PML. Science 328, 240-3 (2010).

248. Ablain, J. et al. Activation of a promyelocytic leukemia-tumor protein 53 axis underlies acute promyelocytic leukemia cure. Nat Med 20, 167-74 (2014).

249. Jeanne, M. et al. PML/RARA oxidation and arsenic binding initiate the antileukemia response of As2O3. Cancer Cell 18, 88-98 (2010).

250. Lallemand-Breitenbach, V. et al. Arsenic degrades PML or PML-RARalpha through a SUMO-triggered RNF4/ubiquitin-mediated pathway. Nat Cell Biol 10, 547-55 (2008).

251. Tatham, M.H. et al. RNF4 is a poly-SUMO-specific E3 ubiquitin ligase required for arsenic-induced PML degradation. Nat Cell Biol 10, 538-46 (2008).

252. He, X. et al. Characterization of the loss of SUMO pathway function on cancer cells and tumor proliferation. PLoS One 10, e0123882 (2015).

253. Riising, E.M., Boggio, R., Chiocca, S., Helin, K. & Pasini, D. The polycomb repressive complex 2 is a potential target of SUMO modifications. PLoS One 3, e2704 (2008).

254. Kumar, A., Ito, A., Hirohama, M., Yoshida, M. & Zhang, K.Y. Identification of quinazolinyloxy biaryl urea as a new class of SUMO activating enzyme 1 inhibitors. Bioorg Med Chem Lett 23, 5145-9 (2013).

255. Kumar, A., Ito, A., Hirohama, M., Yoshida, M. & Zhang, K.Y. Identification of new SUMO activating enzyme 1 inhibitors using virtual screening and scaffold hopping. Bioorg Med Chem Lett 26, 1218-23 (2016).

256. Huszar, D. in AACR Annual Meeting (Atlanta, Georgia).

257. Decque, A. et al. Sumoylation coordinates the repression of inflammatory and anti-viral gene-expression programs during innate sensing. Nat Immunol 17, 140-9 (2016).

258. Crowl, J.T. & Stetson, D.B. SUMO2 and SUMO3 redundantly prevent a noncanonical type I interferon response. Proc Natl Acad Sci U S A 115, 6798-6803 (2018).

200

259. Schneider, W.M., Chevillotte, M.D. & Rice, C.M. Interferon-stimulated genes: a complex web of host defenses. Annu Rev Immunol 32, 513-45 (2014).

260. Chen, Q., Sun, L. & Chen, Z.J. Regulation and function of the cGAS-STING pathway of cytosolic DNA sensing. Nat Immunol 17, 1142-9 (2016).

261. Cui, Y. et al. SENP7 Potentiates cGAS Activation by Relieving SUMO-Mediated Inhibition of Cytosolic DNA Sensing. PLoS Pathog 13, e1006156 (2017).

262. Yu, X. et al. SENP3 maintains the stability and function of regulatory T cells via BACH2 deSUMOylation. Nat Commun 9, 3157 (2018).

263. Chen, J.J. et al. Mechanistic studies of substrate-assisted inhibition of ubiquitin-activating enzyme by adenosine sulfamate analogues. J Biol Chem 286, 40867-77 (2011).

264. An, H. & Statsyuk, A.V. Development of activity-based probes for ubiquitin and ubiquitin-like protein signaling pathways. J Am Chem Soc 135, 16948-62 (2013).

265. An, H. & Statsyuk, A.V. An inhibitor of ubiquitin conjugation and aggresome formation. Chem Sci 6, 5235-5245 (2015).

266. Lu, P. et al. A novel NAE/UAE dual inhibitor LP0040 blocks neddylation and ubiquitination leading to growth inhibition and apoptosis of cancer cells. Eur J Med Chem 154, 294-304 (2018).

267. Holohan, C., Van Schaeybroeck, S., Longley, D.B. & Johnston, P.G. Cancer drug resistance: an evolving paradigm. Nat Rev Cancer 13, 714-26 (2013).

268. Ellis, L.M. & Hicklin, D.J. Resistance to Targeted Therapies: Refining Anticancer Therapy in the Era of Molecular Oncology. Clin Cancer Res 15, 7471-7478 (2009).

269. Milhollen, M.A. et al. Treatment-emergent mutations in NAEbeta confer resistance to the NEDD8-activating enzyme inhibitor MLN4924. Cancer Cell 21, 388-401 (2012).

270. Toth, J.I., Yang, L., Dahl, R. & Petroski, M.D. A gatekeeper residue for NEDD8- activating enzyme inhibition by MLN4924. Cell Rep 1, 309-16 (2012).

271. Xu, G.W. et al. Mutations in UBA3 confer resistance to the NEDD8-activating enzyme inhibitor MLN4924 in human leukemic cells. PLoS One 9, e93530 (2014).

272. Verma, S., Singh, A. & Mishra, A. Molecular dynamics investigation on the poor sensitivity of A171T mutant NEDD8-activating enzyme (NAE) for MLN4924. J Biomol Struct Dyn 32, 1064-73 (2014).

273. Schilsky, R.L., Doroshow, J.H., Leblanc, M. & Conley, B.A. Development and use of integral assays in clinical trials. Clin Cancer Res 18, 1540-6 (2012).

274. Martinez Molina, D. et al. Monitoring drug target engagement in cells and tissues using the cellular thermal shift assay. Science 341, 84-7 (2013).

201

275. Yan, Z.H. et al. Quantifiable analysis of cellular pathway inhibition of a Nedd8-activating enzyme inhibitor, MLN4924, using AlphaScreen. Anal Biochem 439, 109-15 (2013).

276. Yang, X. et al. Absolute quantification of E1, ubiquitin-like proteins and Nedd8- MLN4924 adduct by mass spectrometry. Cell Biochem Biophys 67, 139-47 (2013).

277. Teixeira, L.K. & Reed, S.I. Ubiquitin ligases and cell cycle control. Annu Rev Biochem 82, 387-414 (2013).

278. Hyer, M.L. et al. A small-molecule inhibitor of the ubiquitin activating enzyme for cancer treatment. Nat Med (2018).

279. Khwaja, A. et al. Acute myeloid leukaemia. Nat Rev Dis Primers 2, 16010 (2016).

280. Ferrara, F. & Schiffer, C.A. Acute myeloid leukaemia in adults. Lancet 381, 484-95 (2013).

281. Arber, D.A. et al. The 2016 revision to the World Health Organization classification of myeloid neoplasms and acute leukemia. Blood 127, 2391-405 (2016).

282. Dohner, H., Weisdorf, D.J. & Bloomfield, C.D. Acute Myeloid Leukemia. N Engl J Med 373, 1136-52 (2015).

283. Coombs, C.C., Tallman, M.S. & Levine, R.L. Molecular therapy for acute myeloid leukaemia. Nat Rev Clin Oncol 13, 305-18 (2016).

284. Papaemmanuil, E. et al. Genomic Classification and Prognosis in Acute Myeloid Leukemia. N Engl J Med 374, 2209-2221 (2016).

285. Short, N.J., Rytting, M.E. & Cortes, J.E. Acute myeloid leukaemia. Lancet 392, 593-606 (2018).

286. Welch, J.S. et al. The origin and evolution of mutations in acute myeloid leukemia. Cell 150, 264-78 (2012).

287. Jaiswal, S. & Ebert, B.L. Clonal hematopoiesis in human aging and disease. Science 366 (2019).

288. Vetrie, D., Helgason, G.V. & Copland, M. The leukaemia stem cell: similarities, differences and clinical prospects in CML and AML. Nat Rev Cancer (2020).

289. Shlush, L.I. et al. Tracing the origins of relapse in acute myeloid leukaemia to stem cells. Nature 547, 104-108 (2017).

290. Ganser, A. & Heuser, M. Therapy-related myeloid neoplasms. Curr Opin Hematol 24, 152-158 (2017).

291. Jaiswal, S. et al. Clonal Hematopoiesis and Risk of Atherosclerotic Cardiovascular Disease. N Engl J Med 377, 111-121 (2017).

202

292. Patel, J.P. et al. Prognostic relevance of integrated genetic profiling in acute myeloid leukemia. N Engl J Med 366, 1079-89 (2012).

293. Dohner, H. et al. Diagnosis and management of AML in adults: 2017 ELN recommendations from an international expert panel. Blood 129, 424-447 (2017).

294. Molica, M., Breccia, M., Foa, R., Jabbour, E. & Kadia, T.M. Maintenance therapy in AML: The past, the present and the future. Am J Hematol 94, 1254-1265 (2019).

295. Larson, R.A. in UpToDate (ed. Lowenberg, B.) (UpToDate Inc, Waltham, MA, 2019).

296. Lai, C., Doucette, K. & Norsworthy, K. Recent drug approvals for acute myeloid leukemia. J Hematol Oncol 12, 100 (2019).

297. Kucukyurt, S. & Eskazan, A.E. New drugs approved for acute myeloid leukemia (AML) in 2018. Br J Clin Pharmacol (2019).

298. Stone, R.M. et al. Midostaurin plus Chemotherapy for Acute Myeloid Leukemia with a FLT3 Mutation. N Engl J Med 377, 454-464 (2017).

299. Stein, E.M. et al. Enasidenib in mutant IDH2 relapsed or refractory acute myeloid leukemia. Blood 130, 722-731 (2017).

300. Stein, E.M. et al. Molecular remission and response patterns in patients with mutant- IDH2 acute myeloid leukemia treated with enasidenib. Blood 133, 676-687 (2019).

301. Lancet, J.E. et al. CPX-351 (cytarabine and daunorubicin) Liposome for Injection Versus Conventional Cytarabine Plus Daunorubicin in Older Patients With Newly Diagnosed Secondary Acute Myeloid Leukemia. J Clin Oncol 36, 2684-2692 (2018).

302. Lancet, J.E. et al. Phase 2 trial of CPX-351, a fixed 5:1 molar ratio of cytarabine/daunorubicin, vs cytarabine/daunorubicin in older adults with untreated AML. Blood 123, 3239-46 (2014).

303. Taksin, A.L. et al. High efficacy and safety profile of fractionated doses of Mylotarg as induction therapy in patients with relapsed acute myeloblastic leukemia: a prospective study of the alfa group. Leukemia 21, 66-71 (2007).

304. Amadori, S. et al. Gemtuzumab Ozogamicin Versus Best Supportive Care in Older Patients With Newly Diagnosed Acute Myeloid Leukemia Unsuitable for Intensive Chemotherapy: Results of the Randomized Phase III EORTC-GIMEMA AML-19 Trial. J Clin Oncol 34, 972-9 (2016).

305. DiNardo, C.D. et al. Durable Remissions with Ivosidenib in IDH1-Mutated Relapsed or Refractory AML. N Engl J Med 378, 2386-2398 (2018).

306. DiNardo, C.D. et al. Venetoclax combined with decitabine or azacitidine in treatment- naive, elderly patients with acute myeloid leukemia. Blood 133, 7-17 (2019).

203

307. DiNardo, C.D. et al. Safety and preliminary efficacy of venetoclax with decitabine or azacitidine in elderly patients with previously untreated acute myeloid leukaemia: a non- randomised, open-label, phase 1b study. Lancet Oncol 19, 216-228 (2018).

308. Perl, A.E. et al. Selective inhibition of FLT3 by gilteritinib in relapsed or refractory acute myeloid leukaemia: a multicentre, first-in-human, open-label, phase 1-2 study. Lancet Oncol 18, 1061-1075 (2017).

309. Perl, A.E. et al. Gilteritinib or Chemotherapy for Relapsed or Refractory FLT3-Mutated AML. N Engl J Med 381, 1728-1740 (2019).

310. Savona, M.R. et al. Phase Ib Study of Glasdegib, a Hedgehog Pathway Inhibitor, in Combination with Standard Chemotherapy in Patients with AML or High-Risk MDS. Clin Cancer Res 24, 2294-2303 (2018).

311. Cortes, J.E. et al. Randomized comparison of low dose cytarabine with or without glasdegib in patients with newly diagnosed acute myeloid leukemia or high-risk myelodysplastic syndrome. Leukemia 33, 379-389 (2019).

312. Voorhees, P.M. & Orlowski, R.Z. The proteasome and proteasome inhibitors in cancer therapy. Annu Rev Pharmacol Toxicol 46, 189-213 (2006).

313. Qin, J.Z. et al. Proteasome inhibitors trigger NOXA-mediated apoptosis in melanoma and myeloma cells. Cancer Res 65, 6282-93 (2005).

314. Lee, A.H., Iwakoshi, N.N., Anderson, K.C. & Glimcher, L.H. Proteasome inhibitors disrupt the unfolded protein response in myeloma cells. Proc Natl Acad Sci U S A 100, 9946-51 (2003).

315. Mitsiades, N. et al. Molecular sequelae of proteasome inhibition in human multiple myeloma cells. Proc Natl Acad Sci U S A 99, 14374-9 (2002).

316. Caravita, T., de Fabritiis, P., Palumbo, A., Amadori, S. & Boccadoro, M. Bortezomib: efficacy comparisons in solid tumors and hematologic malignancies. Nat Clin Pract Oncol 3, 374-87 (2006).

317. Bianchi, G. et al. The proteasome load versus capacity balance determines apoptotic sensitivity of multiple myeloma cells to proteasome inhibition. Blood 113, 3040-9 (2009).

318. Niewerth, D., Dingjan, I., Cloos, J., Jansen, G. & Kaspers, G. Proteasome inhibitors in acute leukemia. Expert Rev Anticancer Ther 13, 327-37 (2013).

319. Adams, J. et al. Proteasome inhibitors: a novel class of potent and effective antitumor agents. Cancer Res 59, 2615-22 (1999).

320. Hideshima, T. et al. The proteasome inhibitor PS-341 inhibits growth, induces apoptosis, and overcomes drug resistance in human multiple myeloma cells. Cancer Res 61, 3071-6 (2001).

204

321. LeBlanc, R. et al. Proteasome inhibitor PS-341 inhibits human myeloma cell growth in vivo and prolongs survival in a murine model. Cancer Res 62, 4996-5000 (2002).

322. Clarke, H.J., Chambers, J.E., Liniker, E. & Marciniak, S.J. Endoplasmic reticulum stress in malignancy. Cancer Cell 25, 563-73 (2014).

323. Obeng, E.A. et al. Proteasome inhibitors induce a terminal unfolded protein response in multiple myeloma cells. Blood 107, 4907-16 (2006).

324. Dong, H. et al. Dysregulation of unfolded protein response partially underlies proapoptotic activity of bortezomib in multiple myeloma cells. Leuk Lymphoma 50, 974- 84 (2009).

325. Richardson, P.G. et al. A phase 2 study of bortezomib in relapsed, refractory myeloma. N Engl J Med 348, 2609-17 (2003).

326. Orlowski, R.Z. & Kuhn, D.J. Proteasome inhibitors in cancer therapy: lessons from the first decade. Clin Cancer Res 14, 1649-57 (2008).

327. Bosman, M.C., Schuringa, J.J. & Vellenga, E. Constitutive NF-kappaB activation in AML: Causes and treatment strategies. Crit Rev Oncol Hematol 98, 35-44 (2016).

328. Reikvam, H., Olsnes, A.M., Gjertsen, B.T., Ersvar, E. & Bruserud, O. Nuclear factor- kappaB signaling: a contributor in leukemogenesis and a target for pharmacological intervention in human acute myelogenous leukemia. Crit Rev Oncog 15, 1-41 (2009).

329. Grosjean-Raillard, J. et al. ATM mediates constitutive NF-kappaB activation in high-risk myelodysplastic syndrome and acute myeloid leukemia. Oncogene 28, 1099-109 (2009).

330. Niewerth, D. et al. Higher ratio immune versus constitutive proteasome level as novel indicator of sensitivity of pediatric acute leukemia cells to proteasome inhibitors. Haematologica 98, 1896-904 (2013).

331. Sun, H. et al. Inhibition of IRE1alpha-driven pro-survival pathways is a promising therapeutic application in acute myeloid leukemia. Oncotarget 7, 18736-49 (2016).

332. Dai, Y., Rahmani, M. & Grant, S. Proteasome inhibitors potentiate leukemic cell apoptosis induced by the cyclin-dependent kinase inhibitor flavopiridol through a SAPK/JNK- and NF-kappaB-dependent process. Oncogene 22, 7108-22 (2003).

333. Nencioni, A., Grunebach, F., Patrone, F., Ballestrero, A. & Brossart, P. Proteasome inhibitors: antitumor effects and beyond. Leukemia 21, 30-6 (2007).

334. Fang, J. et al. Cytotoxic effects of bortezomib in myelodysplastic syndrome/acute myeloid leukemia depend on autophagy-mediated lysosomal degradation of TRAF6 and repression of PSMA1. Blood 120, 858-67 (2012).

205

335. Attar, E.C. et al. Phase I and pharmacokinetic study of bortezomib in combination with idarubicin and cytarabine in patients with acute myelogenous leukemia. Clin Cancer Res 14, 1446-54 (2008).

336. Lancet, J.E. et al. A phase I clinical-pharmacodynamic study of the farnesyltransferase inhibitor tipifarnib in combination with the proteasome inhibitor bortezomib in advanced acute leukemias. Clin Cancer Res 17, 1140-6 (2011).

337. Cortes, J. et al. Phase I study of bortezomib in refractory or relapsed acute leukemias. Clin Cancer Res 10, 3371-6 (2004).

338. Sarlo, C. et al. Phase II Study of Bortezomib as a Single Agent in Patients with Previously Untreated or Relapsed/Refractory Acute Myeloid Leukemia Ineligible for Intensive Therapy. Leuk Res Treatment 2013, 705714 (2013).

339. Blum, W. et al. Clinical and pharmacodynamic activity of bortezomib and decitabine in acute myeloid leukemia. Blood 119, 6025-31 (2012).

340. Attar, E.C. et al. Bortezomib added to daunorubicin and cytarabine during induction therapy and to intermediate-dose cytarabine for consolidation in patients with previously untreated acute myeloid leukemia age 60 to 75 years: CALGB (Alliance) study 10502. J Clin Oncol 31, 923-9 (2013).

341. Warlick, E.D., Cao, Q. & Miller, J. Bortezomib and vorinostat in refractory acute myelogenous leukemia and high-risk myelodysplastic syndromes: produces stable disease but at the cost of high toxicity. Leukemia 27, 1789-91 (2013).

342. Horton, T.M. et al. A Phase 2 study of bortezomib combined with either idarubicin/cytarabine or cytarabine/etoposide in children with relapsed, refractory or secondary acute myeloid leukemia: a report from the Children's Oncology Group. Pediatr Blood Cancer 61, 1754-60 (2014).

343. Positive First Trial of Enasidenib for AML. Cancer Discov 7, OF1 (2017).

344. Shlush, L.I. et al. Tracing the origins of relapse in acute myeloid leukaemia to stem cells. Nature (2017).

345. Hoeller, D., Hecker, C.M. & Dikic, I. Ubiquitin and ubiquitin-like proteins in cancer pathogenesis. Nat Rev Cancer 6, 776-88 (2006).

346. Groen, E.J. & Gillingwater, T.H. UBA1: At the Crossroads of Ubiquitin Homeostasis and Neurodegeneration. Trends Mol Med 21, 622-32 (2015).

347. Brodsky, J.L. Cleaning up: ER-associated degradation to the rescue. Cell 151, 1163-7 (2012).

348. Chen, Z.J. & Sun, L.J. Nonproteolytic functions of ubiquitin in cell signaling. Mol Cell 33, 275-86 (2009).

206

349. Clague, M.J., Liu, H. & Urbe, S. Governance of endocytic trafficking and signaling by reversible ubiquitylation. Dev Cell 23, 457-67 (2012).

350. Warner, J.K. et al. Direct evidence for cooperating genetic events in the leukemic transformation of normal human hematopoietic cells. Leukemia 19, 1794-805 (2005).

351. Spagnuolo, P.A. et al. The antihelmintic flubendazole inhibits microtubule function through a mechanism distinct from Vinca alkaloids and displays preclinical activity in leukemia and myeloma. Blood 115, 4824-33 (2010).

352. Jafari, R. et al. The cellular thermal shift assay for evaluating drug target interactions in cells. Nat Protoc 9, 2100-22 (2014).

353. Cole, A. et al. Inhibition of the Mitochondrial Protease ClpP as a Therapeutic Strategy for Human Acute Myeloid Leukemia. Cancer Cell 27, 864-76 (2015).

354. Werstuck, G.H. et al. Homocysteine-induced endoplasmic reticulum stress causes dysregulation of the cholesterol and triglyceride biosynthetic pathways. J Clin Invest 107, 1263-73 (2001).

355. Liyanage, S.U. et al. Leveraging increased cytoplasmic nucleoside kinase activity to target mtDNA and oxidative phosphorylation in AML. Blood 129, 2657-2666 (2017).

356. Christianson, J.C. & Ye, Y. Cleaning up in the endoplasmic reticulum: ubiquitin in charge. Nat Struct Mol Biol 21, 325-35 (2014).

357. Smith, M.H., Ploegh, H.L. & Weissman, J.S. Road to ruin: targeting proteins for degradation in the endoplasmic reticulum. Science 334, 1086-90 (2011).

358. Wang, C.Y. et al. Inhibition of HSP90 by AUY922 Preferentially Kills Mutant KRAS Colon Cancer Cells by Activating Bim through ER Stress. Mol Cancer Ther 15, 448-59 (2016).

359. Moreno, J.A. et al. Oral treatment targeting the unfolded protein response prevents neurodegeneration and clinical disease in prion-infected mice. Sci Transl Med 5, 206ra138 (2013).

360. Cross, B.C. et al. The molecular basis for selective inhibition of unconventional mRNA splicing by an IRE1-binding small molecule. Proc Natl Acad Sci U S A 109, E869-78 (2012).

361. Guan, M., Su, L., Yuan, Y.C., Li, H. & Chow, W.A. Nelfinavir and nelfinavir analogs block site-2 protease cleavage to inhibit castration-resistant prostate cancer. Sci Rep 5, 9698 (2015).

362. Guan, M., Fousek, K. & Chow, W.A. Nelfinavir inhibits regulated intramembrane proteolysis of sterol regulatory element binding protein-1 and activating transcription factor 6 in castration-resistant prostate cancer. FEBS J 279, 2399-411 (2012).

207

363. Al-Hakim, A. et al. The ubiquitous role of ubiquitin in the DNA damage response. DNA Repair (Amst) 9, 1229-40 (2010).

364. Schwertman, P., Bekker-Jensen, S. & Mailand, N. Regulation of DNA double-strand break repair by ubiquitin and ubiquitin-like modifiers. Nat Rev Mol Cell Biol 17, 379-94 (2016).

365. Weake, V.M. & Workman, J.L. Histone ubiquitination: triggering gene activity. Mol Cell 29, 653-63 (2008).

366. Uckelmann, M. & Sixma, T.K. Histone ubiquitination in the DNA damage response. DNA Repair (Amst) 56, 92-101 (2017).

367. Misra, M. et al. Dissecting the Specificity of Adenosyl Sulfamate Inhibitors Targeting the Ubiquitin-Activating Enzyme. Structure (2017).

368. van Galen, P. et al. The unfolded protein response governs integrity of the haematopoietic stem-cell pool during stress. Nature 510, 268-72 (2014).

369. Guzman, M.L. et al. Nuclear factor-kappaB is constitutively activated in primitive human acute myelogenous leukemia cells. Blood 98, 2301-7 (2001).

370. Csizmar, C.M., Kim, D.H. & Sachs, Z. The role of the proteasome in AML. Blood Cancer J 6, e503 (2016).

371. Hart, T. et al. High-Resolution CRISPR Screens Reveal Fitness Genes and Genotype- Specific Cancer Liabilities. Cell 163, 1515-26 (2015).

372. Li, W. et al. MAGeCK enables robust identification of essential genes from genome-scale CRISPR/Cas9 knockout screens. Genome Biol 15, 554 (2014).

373. Sanjana, N.E., Shalem, O. & Zhang, F. Improved vectors and genome-wide libraries for CRISPR screening. Nat Methods 11, 783-784 (2014).

374. Franken, N.A., Rodermond, H.M., Stap, J., Haveman, J. & van Bree, C. Clonogenic assay of cells in vitro. Nature protocols 1, 2315-9 (2006).

375. Jafari, R. et al. The cellular thermal shift assay for evaluating drug target interactions in cells. Nature protocols 9, 2100-22 (2014).

376. Li, W. et al. MAGeCK enables robust identification of essential genes from genome-scale CRISPR/Cas9 knockout screens. Genome biology 15, 554 (2014).

377. Robey, R.W. et al. Revisiting the role of ABC transporters in multidrug-resistant cancer. Nat Rev Cancer 18, 452-464 (2018).

378. Szakacs, G., Paterson, J.K., Ludwig, J.A., Booth-Genthe, C. & Gottesman, M.M. Targeting multidrug resistance in cancer. Nat Rev Drug Discov 5, 219-34 (2006).

208

379. Allen, J.D. et al. Potent and specific inhibition of the breast cancer resistance protein multidrug transporter in vitro and in mouse intestine by a novel analogue of fumitremorgin C. Mol Cancer Ther 1, 417-25 (2002).

380. Weidner, L.D. et al. The Inhibitor Ko143 Is Not Specific for ABCG2. J Pharmacol Exp Ther 354, 384-93 (2015).

381. Dantzig, A.H. et al. Selectivity of the multidrug resistance modulator, LY335979, for P- glycoprotein and effect on cytochrome P-450 activities. J Pharmacol Exp Ther 290, 854- 62 (1999).

382. Khan, A. & Prasanth, S.G. BEND3 mediates transcriptional repression and heterochromatin organization. Transcription 6, 102-5 (2015).

383. Khan, A. et al. BEND3 represses rDNA transcription by stabilizing a NoRC component via USP21 deubiquitinase. Proc Natl Acad Sci U S A 112, 8338-43 (2015).

384. Tsherniak, A. et al. Defining a Cancer Dependency Map. Cell 170, 564-576 e16 (2017).

385. Gottesman, M.M., Fojo, T. & Bates, S.E. Multidrug resistance in cancer: role of ATP- dependent transporters. Nat Rev Cancer 2, 48-58 (2002).

386. Allen, J.D., Van Dort, S.C., Buitelaar, M., van Tellingen, O. & Schinkel, A.H. Mouse breast cancer resistance protein (Bcrp1/Abcg2) mediates etoposide resistance and transport, but etoposide oral availability is limited primarily by P-glycoprotein. Cancer Res 63, 1339-44 (2003).

387. Kaehler, M. et al. MicroRNA-212/ABCG2-axis contributes to development of imatinib- resistance in leukemic cells. Oncotarget 8, 92018-92031 (2017).

388. Volk, E.L. et al. Overexpression of wild-type breast cancer resistance protein mediates methotrexate resistance. Cancer Res 62, 5035-40 (2002).

389. Miyake, K. et al. Molecular cloning of cDNAs which are highly overexpressed in mitoxantrone-resistant cells: demonstration of homology to ABC transport genes. Cancer Res 59, 8-13 (1999).

390. Doyle, L.A. et al. A multidrug resistance transporter from human MCF-7 breast cancer cells. Proc Natl Acad Sci U S A 95, 15665-70 (1998).

391. Benderra, Z. et al. Breast cancer resistance protein and P-glycoprotein in 149 adult acute myeloid leukemias. Clin Cancer Res 10, 7896-902 (2004).

392. Damiani, D. et al. The prognostic value of P-glycoprotein (ABCB) and breast cancer resistance protein (ABCG2) in adults with de novo acute myeloid leukemia with normal karyotype. Haematologica 91, 825-8 (2006).

209

393. Steinbach, D. et al. BCRP gene expression is associated with a poor response to remission induction therapy in childhood acute myeloid leukemia. Leukemia 16, 1443-7 (2002).

394. van den Heuvel-Eibrink, M.M. et al. Increased expression of the breast cancer resistance protein (BCRP) in relapsed or refractory acute myeloid leukemia (AML). Leukemia 16, 833-9 (2002).

395. van den Heuvel-Eibrink, M.M. et al. CD34-related coexpression of MDR1 and BCRP indicates a clinically resistant phenotype in patients with acute myeloid leukemia (AML) of older age. Annals of hematology 86, 329-37 (2007).

396. Gerrard, G. et al. Clinical effects and P-glycoprotein inhibition in patients with acute myeloid leukemia treated with zosuquidar trihydrochloride, daunorubicin and cytarabine. Haematologica 89, 782-90 (2004).

397. Kelly, R.J. et al. A pharmacodynamic study of docetaxel in combination with the P- glycoprotein antagonist tariquidar (XR9576) in patients with lung, ovarian, and cervical cancer. Clin Cancer Res 17, 569-80 (2011).

398. Polgar, O., Robey, R.W. & Bates, S.E. ABCG2: structure, function and role in drug response. Expert opinion on drug metabolism & toxicology 4, 1-15 (2008).

399. Maliepaard, M. et al. Subcellular localization and distribution of the breast cancer resistance protein transporter in normal human tissues. Cancer Res 61, 3458-64 (2001).

400. Turner, J.G. et al. ABCG2 expression, function, and promoter methylation in human multiple myeloma. Blood 108, 3881-9 (2006).

401. Bruckmueller, H. et al. Clinically Relevant Multidrug Transporters Are Regulated by microRNAs along the Human Intestine. Molecular pharmaceutics 14, 2245-2253 (2017).

402. Haenisch, S. et al. Down-regulation of ATP-binding cassette C2 protein expression in HepG2 cells after rifampicin treatment is mediated by microRNA-379. Mol Pharmacol 80, 314-20 (2011).

403. Turrini, E. et al. MicroRNA profiling in K-562 cells under imatinib treatment: influence of miR-212 and miR-328 on ABCG2 expression. Pharmacogenetics and genomics 22, 198-205 (2012).

404. Wu, A.M. et al. Induction of multidrug resistance transporter ABCG2 by prolactin in human breast cancer cells. Mol Pharmacol 83, 377-88 (2013).

405. Arumugam, A. et al. Silencing growth hormone receptor inhibits estrogen receptor negative breast cancer through ATP-binding cassette sub-family G member 2. Experimental & molecular medicine 51, 2 (2019).

210

406. Sathyan, K.M., Shen, Z., Tripathi, V., Prasanth, K.V. & Prasanth, S.G. A BEN-domain- containing protein associates with heterochromatin and represses transcription. J Cell Sci 124, 3149-63 (2011).

407. Saksouk, N. et al. Redundant mechanisms to form silent chromatin at pericentromeric regions rely on BEND3 and DNA methylation. Mol Cell 56, 580-94 (2014).

408. Huang, Y. et al. Loss of PICH promotes chromosome instability and cell death in triple- negative breast cancer. Cell Death Dis 10, 428 (2019).

409. Pitchai, G.P. et al. A novel TPR-BEN domain interaction mediates PICH-BEND3 association. Nucleic Acids Res 45, 11413-11424 (2017).

410. Wu, H. et al. Crystal structures of the human histone H4K20 methyltransferases SUV420H1 and SUV420H2. FEBS Lett 587, 3859-68 (2013).

411. Nishioka, K. et al. PR-Set7 is a nucleosome-specific methyltransferase that modifies lysine 20 of histone H4 and is associated with silent chromatin. Mol Cell 9, 1201-13 (2002).

412. Kapoor-Vazirani, P., Kagey, J.D. & Vertino, P.M. SUV420H2-mediated H4K20 trimethylation enforces RNA polymerase II promoter-proximal pausing by blocking hMOF-dependent H4K16 acetylation. Mol Cell Biol 31, 1594-609 (2011).

413. Brustel, J. et al. Histone H4K20 tri-methylation at late-firing origins ensures timely heterochromatin replication. EMBO J 36, 2726-2741 (2017).

414. Jorgensen, S., Schotta, G. & Sorensen, C.S. Histone H4 lysine 20 methylation: key player in epigenetic regulation of genomic integrity. Nucleic Acids Res 41, 2797-806 (2013).

415. Li, Z., Nie, F., Wang, S. & Li, L. Histone H4 Lys 20 monomethylation by histone methylase SET8 mediates Wnt target gene activation. Proc Natl Acad Sci U S A 108, 3116-23 (2011).

416. Bromberg, K.D. et al. The SUV4-20 inhibitor A-196 verifies a role for epigenetics in genomic integrity. Nat Chem Biol 13, 317-324 (2017).

417. Boulanger, M., Paolillo, R., Piechaczyk, M. & Bossis, G. The SUMO Pathway in Hematomalignancies and Their Response to Therapies. Int J Mol Sci 20 (2019).

418. Zhou, P. et al. 2-D08 as a SUMOylation inhibitor induced ROS accumulation mediates apoptosis of acute myeloid leukemia cells possibly through the deSUMOylation of NOX2. Biochem Biophys Res Commun 513, 1063-1069 (2019).

211

Appendix

Appendix Copyright and permissions

Figure 1.10

212 213

Figure 1.11 and 1.12

Permissions policy of the New England Journal of Medicine is stated below (accessed at https://www.nejm.org/about-nejm/permissions on November 9, 2019):

214

Figure 2.15

Fig. 2.15 is adapted from the author’s published editorial in Oncotarget journal. The full citation of this article is as follows:

Barghout SH, Schimmer AD. The ubiquitin-activating enzyme, UBA1, as a novel therapeutic target for AML. Oncotarget. 2018 Sep 28;9(76):34198-34199. doi: 10.18632/oncotarget.26153.

Below is the policy of reuse of material from Oncotarget journal (accessed online at http://www.oncotarget.com/index.php?journal=oncotarget&page=about on November 9, 2019):

215

Chapter 2

Chapter 2 has been published as a research article in Leukemia journal. The full citation of this article is as follows:

Barghout SH, Patel PS, Wang X, Xu GW, Kavanagh S, Halgas O, Zarabi SF, Gronda M, Hurren R, Jeyaraju DV, MacLean N, Brennan S, Hyer ML, Berger A, Traore T, Milhollen M, Smith AC, Minden MD, Pai EF, Hakem R, Schimmer AD. Preclinical evaluation of the selective small-molecule UBA1 inhibitor, TAK-243, in acute myeloid leukemia. Leukemia. 2019 Jan;33(1):37-51. doi: 10.1038/s41375-018-0167-0. Epub 2018 Jun 8.

Below is the policy of reuse of articles in theses/dissertations: