Role of Tafazzin in Hematopoiesis and Leukemogenesis

Home , Gilteritinib, PIGS (gene), Tafazzin

Ayesh Seneviratne

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

Role of Tafazzin in Hematopoiesis and Leukemogenesis

Ayesh Seneviratne

Doctor of Philosophy

Institute of Medical Science University of Toronto

2020 Abstract

Tafazzin (TAZ) is a mitochondrial transacylase that remodels the mitochondrial cardiolipin into its mature form. Through a CRISPR screen, we identified TAZ as necessary for the growth and viability of acute myeloid leukemia (AML) cells. Genetic inhibition of TAZ reduced stemness and increased differentiation of AML cells both in vitro and in vivo. In contrast, knockdown of

TAZ did not impair normal hematopoiesis under basal conditions. Mechanistically, inhibition of

TAZ decreased levels of cardiolipin but also altered global levels of intracellular phospholipids, including phosphatidylserine, which controlled AML stemness and differentiation by modulating toll-like receptor (TLR) signaling (Seneviratne et al., 2019).

Acknowledgments

Firstly, I would like to thank Dr. Aaron Schimmer for his guidance and support during my PhD studies. I really enjoyed our early morning meetings where he provided much needed perspective to navigate the road blocks of my project, whilst continuing to push me. It was a privilege to be mentored by such an excellent clinician scientist. I hope to continue to build on the skills I learned in Dr. Schimmer’s lab as I progress on my path to become a clinician scientist.

Working in the Schimmer lab was a wonderful learning environment. I would like to especially thank Dr. Mingjing Xu who helped with critical experiments for the Cell Stem Cell paper that came out of this thesis. Her work ethic and love for science was very inspirational. I also want to thank Dr. Wei Xu, Dr. Danny Jeyaraju, Rose Hurren, Dr. Marcela Gronda, Mr. Neil MacLean, and Mrs. Xiaoming Wang for assisting me with experiments. I really appreciate your tireless effort to ensure that I have the resources to do my work, as well as your advice and help during the troubleshooting process.

I also want to acknowledge our collaborators Dr. Juan J. Aristizabal, Dr. Ken Stark, Dr. Val Fajardo, and Dr. Paul Leblanc for your commitment to this project. Your expertise in lipid biology really pushed this work forward.

Thanks to my advisory committee, Dr. Steven Chan, and Dr. Catherine O’Brien for reading all of my progress reports, and providing me with insightful comments, which helped to carry my project forward.

I am very grateful for Dr. Auro Viswabandya, and Dr. Frank Michelis for letting me attend morning rounds, as well as bone marrow transplant clinics. I am also very appreciative of the late Dr. Hans Messner for giving me the opportunity to get involved in clinical research. The exposure to this extremely complicated medical specialty has inspired me to seriously pursue a carrier as a bone marrow transplant physician.

I very grateful for the Canadian Institute of Health and Research and the MD/PhD program for their funding, without which this work would not be possible.

I would like to especially thank my family, my father Dr. Charitha Seneviratne, my mother Mrs. Dilrukshi Seneviratne, and my sister Miss. Nayantara Seneviratne for their continuous guidance, emotional support, and encouragement during these four very testing years of graduate school.

iii

Your countless trips to Toronto with food to fill my fridge, and words of encouragement to fuel my spirit really kept me going. All my achievements reflect your commitment to my education.

I also want to thank my new family Mr. Ravi Senanayake, Dr. Dhammika Senanayake, Mr. Jaliya Jayawardena, Mrs. Ramila Jayawardena, Master. Kanishka Jayawardena, Master. Kevaan Jayawardena, and Mr. Thimila Senanayake for their words of encouragement and support.

Finally, I want to thank my wife Mrs. Uthumi Senanayake for her your support. I really appreciate the time you took to listen to my countless theories regarding my project, even though you were not familiar with the content, some of the pressing questions you asked pushed me to look at my project differently. I also want to thank you for taking the time-off from work to attend all of my talks. I appreciated your attendance not only for the sentiment, but also because it motivated me practice more, so I ended up giving a better presentation. Your tireless effort to make sure that I take time to: reflect on my achievements, travel, spend time with family and friends, and eat well, made the arduous task of completing a Ph.D. more pleasurable. I really do mean it when I say that this thesis is as much yours as it is mine.

Contribution Statement

The work in this thesis was published in Cell Stem Cell (Seneviratne et al., 2019). The majority of work presented in this thesis was done by Ayesh Seneviratne. Technical assistance was contributed as follows:

Dr. Minjing Xu helped with investigating the role of phosphatidylserine decarboxylase in AML. She assisted with all experiments pertaining to this aspect of the project.

Veronique Voisin performed all the bioinformatic analysis, under the supervision of Dr. Gary D. Bader.

Rose Hurren provided technical assistance for qRT-PCR, under the supervision of Dr. Aaron Schimmer

Mass spectrometric experiments were performed by. Dr. Juan J. Aristizabal, under the supervision of Dr. Ken Stark.

Densitometric analysis of phospholipids were performed by. Dr. Val Fajardo, under the supervision of Dr. Paul Leblanc.

Neil Maclean (Princess Margaret Cancer Centre, University of Toronto, Ontario, Canada) provided technical assistance with production of lentivirus for genetic knockdown experiments, under the supervision of Dr. Aaron Schimmer

Rose Hurren and Xiaoming Wang (Princess Margaret Cancer Centre, University of Toronto, Ontario, Canada) performed and analyzed data for in vivo experiments, under the supervision of Dr. Aaron Schimmer

Table of Contents

Acknowledgments ...... iii

Contribution Statement ...... v

Table of Contents ...... vi

List of Tables ...... xiii

List of Figures ...... xiv

List of Abbreviations ...... xvii

Chapter 1 ...... 1

Literature Review ...... 1

1.1 Hematopoiesis ...... 1

1.1.1 Discovery of the Hematopoietic Hierarchy ...... 1

1.1.2 Current Models of Hematopoiesis ...... 2

1.1.3 Emergency Hematopoiesis ...... 6

1.2 Leukemia ...... 6

1.3 Acute Myeloid Leukemia ...... 8

1.3.1 AML Pathogenesis ...... 8

1.3.2 Leukemia Stem Cells ...... 10

1.3.3 Classification of AML ...... 12

1.3.4 Treatment Modalities of AML ...... 17

1.3.4.1 Cytotoxic Therapies ...... 17

1.3.4.2 Hematopoietic Cell Transplant ...... 18

1.3.4.3 All-trans retinoic Acid and Arsenic Trioxide (ATRA-ATO) ...... 27

1.4 Energy Metabolism ...... 28

1.4.1 Catabolism ...... 28

1.4.2 Anabolism ...... 31

1.4.2.1 Nucleoside Synthesis ...... 32

1.4.2.2 Lipid Synthesis ...... 33

1.4.2.3 Amino Acid Synthesis ...... 34

1.5 Cancer Metabolism ...... 34

1.6 Oncogenic Pathways Drive Cancer Metabolism ...... 35

1.7 Mitochondria Biology ...... 37

1.7.1 Biogenesis and Quality Control of Mitochondrial Proteins ...... 38

1.7.1.1 Mitochondrial DNA Replication ...... 38

1.7.1.2 Mitochondrial DNA Transcription ...... 39

1.7.1.3 Mitochondrial DNA Translation into Proteins ...... 39

1.7.1.4 Quality Control by Mitochondria Proteases ...... 40

1.7.1.5 Mitochondrial Biogenesis and Protein Quality Control in AML ...... 41

1.7.2 Amino Acids and Fatty Acid Oxidation by the Mitochondria ...... 42

1.7.2.1 Mitochondrial Amino Acids Oxidation in AML ...... 43

1.7.2.2 Mitochondrial Fatty acid Oxidation in AML ...... 44

1.7.3 Mitochondrial Redox and Antioxidant Systems ...... 44

1.7.3.1 Mitochondria Redox/Antioxidant Systems and Signaling in AML ...... 45

1.7.4 Retrograde Signaling ...... 46

1.7.4.1 Mitochondrial Dynamics and Mitophagy ...... 46

1.7.4.2 Mitochondrial Dynamics and Mitophagy Signaling in AML ...... 48

1.7.4.3 Mitochondrial Metabolites ...... 49

1.7.4.4 Metabolite Signaling in AML ...... 49

1.7.4.5 Dihyroorotate Dehydrogenase and Pyrimidine Synthesis ...... 49

1.7.4.6 DHODH and AML Cell State ...... 50

1.8 CRISPR Screens ...... 52

1.9 Tafazzin and Cardiolipin ...... 54

vii

1.9.1 Cardiolipin Functions...... 54

1.9.2 Cardiolipin Synthesis and Remodeling ...... 55

1.10 Objective and Aims ...... 57

Chapter 2 ...... 58

Methods ...... 58

2.1 Experimental Model and Subject Details ...... 58

2.1.1 Human Cell Lines ...... 58

2.1.2 Animals ...... 59

2.1.3 TAZ-KD Induction in iDOX-TAZ-KD Mice ...... 61

2.1.4 Primary AML and Normal Hematopoietic Cells ...... 61

2.1.5 Primary AML Cell Cultures for Transduction ...... 63

2.2 Method Details ...... 64

2.2.1 Plasmids ...... 64

2.2.2 Lentiviral Packing ...... 66

2.2.3 CAS9-OCI-AML2 Cell Line Generation ...... 66

2.2.4 CRISPR Screen ...... 67

2.2.5 MAGeCK Analysis ...... 67

2.2.6 Viral Infection ...... 67

2.2.6.1 CRISPR-sgRNA Knockout ...... 67

2.2.6.2 shRNA knockdown of AML Cell Lines ...... 68

2.2.6.3 shRNA knockdown of Primary AML Cell Samples ...... 68

2.2.6.4 PISD Overexpression ...... 68

2.2.7 Mitochondrial Protein Lysates ...... 69

2.2.8 Whole Cell Protein Lysates ...... 69

2.2.9 Immunoblotting...... 69

2.2.10 Basal Apoptosis ...... 70

viii

2.2.11 Cell Cycle Analysis ...... 70

2.2.12 Colony Formation Assays ...... 70

2.2.12.1 Leukemia Cell Lines ...... 70

2.2.12.2 Primary AML Patient Samples ...... 71

2.2.13 RNA-Sequencing ...... 72

2.2.14 Differential Gene Expression Analysis ...... 73

2.2.15 Pathway Analysis ...... 73

2.2.16 TCGA Unsupervised Clustering ...... 73

2.2.17 LSC+/LSC- Signature Analysis ...... 74

2.2.18 DMAP- Signature Analysis ...... 74

2.2.19 PERT Deconvolution Analysis ...... 75

2.2.20 Non-Specific Esterase Staining ...... 75

2.2.21 NSE Staining Analysis Using ImageJ ...... 75

2.2.22 Cell Surface Phenotype of OCI-AML2 ...... 75

2.2.23 RNA Isolation and qRT-PCR ...... 75

2.2.24 Lipid Extraction Protocols ...... 76

2.2.24.1 Ultra-Hight Performance Liquid Chromatography/Mass Spectrometry (UHPLC/MS) ...... 76

2.2.24.2 Densitometric Analysis ...... 76

2.2.25 UHPLC/MS Characterization of MLCL and CL ...... 77

2.2.26 Densitometric Characterization of Phospholipids ...... 80

2.2.27 Densitometric Phospholipids Analysis ...... 80

2.2.28 Quantification of Extracellular PS ...... 80

2.2.29 Intracellular PS Quantification by Confocal Microscopy ...... 80

2.2.30 Intracellular PS Quantification by ImageJ ...... 81

2.2.31 Sensitivity to Extrinsic Apoptosis ...... 81

2.2.32 Seahorse ...... 81 ix

2.2.33 Cellular ROS ...... 82

2.2.34 Mitochondrial Mass ...... 82

2.2.35 Live Imaging ...... 82

2.2.36 Aspect Ratio ...... 82

2.2.37 Electron Microscopy ...... 83

2.2.38 Lipid Overlay Assay ...... 83

2.2.39 Lipid Supplementation and Growth Analysis ...... 83

2.2.40 MMV Treatment and Growth Analysis ...... 84

2.2.41 8227 Flow Cytometry ...... 84

2.2.42 PS Quantification by Flow Cytometry ...... 84

2.2.43 MMV and CL075 Synergism Studies ...... 84

2.2.44 Animal Studies ...... 84

2.2.44.1 Hematopoiesis in WT and Taz-KD mice ...... 84

2.2.44.2 TEX and 8227 Engraftment ...... 85

2.2.44.3 Stability and Pharmacokinetics of MMV007285 ...... 85

2.2.44.4 Subcutaneous AML Xenografts ...... 86

2.2.44.5 Primary AML Engraftment Models ...... 86

2.2.45 Calculation of Engraftment Potential ...... 88

2.3 Quantification and Statistical Analysis ...... 88

2.4 Data Availability ...... 88

Chapter 3 ...... 89

Results ...... 89

3.1 CRISPR Screen to Identify Mitochondrial Proteins Essential for Leukemia Cell Growth ...... 89

3.1.1 CRISPR screen identifies TAZ as essential for the growth and viability of AML cells ...... 89

3.1.2 TAZ knockdown reduced the growth of leukemia cells ...... 92

3.2 Role of TAZ in LSC Function ...... 94

3.2.1 TAZ knockdown reduced the expression of genes associated with LSCs ...... 94

Cellular phenotype of AML cells after TAZ knockdown ...... 100

3.2.2 TAZ knockdown reduced the function of AML stem cells ...... 101

3.3 Role of TAZ in Hematopoiesis ...... 103

3.3.1 Levels of blood cells after TAZ knockdown ...... 103

3.3.2 Role of TAZ in hematopoietic stem and progenitor cells ...... 104

3.3.3 Role of TAZ in hematopoietic stem and progenitor cells during stress ...... 106

3.4 Mechanism by which TAZ knockdown Reduces AML Growth ...... 110

3.4.1 The effect of TAZ knockdown on MLCL levels and CL composition ...... 110

3.4.2 Mitochondria structure and function after TAZ knockdown ...... 112

3.4.3 TAZ knockdown alters cellular phospholipids ...... 114

3.4.4 Functional effects of altered phospholipids ...... 115

3.4.5 Effects of TAZ knockdown on phosphatidylserine decarboxylase (PISD) ...... 119

3.4.6 Anti-leukemic activity of a PISD inhibitor ...... 122

3.4.7 Increasing intracellular phosphatidylserine increases toll-like receptor activity .124

3.5 Pre-clinical anti-leukemic activity of strategies that increase PS ...... 127

3.5.1 Effect of PISD inhibition on AML stem cell function in vitro ...... 127

3.5.2 Effect of TAZ knockdown on AML stem cell function in vivo ...... 127

3.5.3 Stability, anti-leukemic activity, and toxicity of MMV007285 in vivo ...... 131

3.5.4 Effect of MMV007285 on stem cell function of AML patient samples in vivo ..131

Chapter 4 ...... 134

Discussion ...... 134

4.1 Summary ...... 134

4.2 TAZ a Novel Mitochondrial a Therapeutic Target for AML ...... 136

4.3 Altered Cardiolipin does not Impair Mitochondria Structure or Function ...... 137

4.4 Neutropenia of Barth Syndrome ...... 139

4.5 Cardiolipin and Phospholipid Synthesis ...... 141

4.6 Cardiolipin and PISD ...... 142

4.7 Phospholipids and Cell Signaling Pathways ...... 142

4.8 PISD Inhibitor MMV007285 ...... 144

Chapter 5 ...... 146

Future Directions ...... 146

5.1 Mechanistic Studies ...... 146

5.1.1 Determine how Increase in PS Stimulates TLR activation ...... 146

5.1.2 Determine if metabolic changes after TAZ-KD stimulates differentiation ...... 147

5.1.3 Cardiolipin Lipid Network ...... 147

5.2 Pre-clinical studies ...... 148

5.2.1 Characterization of Taz knockdown in Barth Syndrome ...... 148

5.2.2 Basis for the AML specific effect of TAZ knockdown and PS increase ...... 148

5.2.3 Improve Efficacy of PISD inhibitor ...... 149

5.2.4 Phospholipid Modulation and Viral Mimicry ...... 149

References or Bibliography (if any) ...... 150

xii

List of Tables

Table 1. French-American-British classification of AML.

Table 2. World Health Organization classification of AML.

Table 3. European Leukemia Network classification of AML.

Table 4. GVHD prophylaxis strategies.

Table 5 . Adjuvant treatments for GVHD.

Table 6. Cell lines.

Table 7. PCR primers.

Table 8. Clinical characteristics of primary AML patient Samples.

Table 9. sgRNA or shRNA sequences.

Table 10. Detection of lipids by MS/MS.

Table 11. CRISPR Hits of mitochondrial proteins.

Table 12. Mouse bone marrow mononuclear cell count.

Table 13. 5-FU treated mouse bone marrow mononuclear cell count.

xiii

List of Figures

Figure 1 Mouse hematopoietic hierarchy.

Figure 2 Human hematopoietic hierarchy.

Figure 3 Leukemogenesis.

Figure 4 AML pathogenesis and hierarchy.

Figure 5 Conditioning regiments for allogenic-HCT.

HCT trends by type and recipient age, 2002-2006 and 2007-2011, Figure 6 CIBMTR data.

Allogenic-HCT trends by conditioning regimen intensity and age, 2001- Figure 7 2011, CIBMTR.

Figure 8 Glucose catabolism in the cell

Figure 9 Catabolism of amino acids and fatty acids.

Figure 10 Macromolecular biosynthesis pathways.

Figure 11 The mitochondrial dependencies of AML.

Pooled clustered regularly interspaced short palindromic repeats Figure 12 (CRISPR) screens.

Figure 13 Cardiolipin synthesis and remodeling pathways.

Figure 14. CRISPR screens identify TAZ as an essential gene for the growth and viability of AML cells.

Figure 15. Knockdown of TAZ reduces the growth of leukemia cells lines.

Figure 16. Growth analysis of leukemia cells after TAZ knockdown.

xiv

Figure 17. TAZ knockdown reduces LSC signature in OCI-AML2 cells.

Figure 18. Related to Figure 17. TAZ knockdown reduces stemness signature in OCI-AML2 cells.

Figure 19. Related to Figure 17. Knockdown of TAZ increases the expression of genes associated with granulocytic/neutrophilic populations.

Figure 20. Knockdown of TAZ increases the expression of mature cell markers.

Figure 21. Knockdown of Tafazzin reduces AML stem cell function.

Figure 22. Taz Knockdown mice are viable with a normal blood count.

Figure 23. Taz Knockdown mice have normal frequencies of hematopoietic stem and progenitor cells.

Figure 24. Blood counts of WT and Taz-KD mice after the induction of hematopoietic stress.

Figure 25. Frequencies and function of hematopoietic stem and progenitor cells in Taz-KD mice.

Figure 26. Knockdown of Tafazzin in OCI-AML2 cells reduces TAZ activity.

Figure 27. Knockdown of Tafazzin in OCI-AML2 cells does not affect mitochondrial structure.

Figure 28. Knockdown of Tafazzin in OCI-AML2 cells does not affect mitochondrial function.

Figure 29. TAZ knockdown increases phosphatidylserine levels in leukemia cells.

Figure 30. Effect of PE and LPE supplementation on TAZ knockdown OCI-AML2 cells.

Figure 31. The increase PS levels after TAZ knockdown is functionally important in the reduction of AML stemness.

Figure 32. TAZ knockdown reduces phosphatidylserine decarboxylase (PISD) levels.

Figure 33. PISD decrease after TAZ knockdown is functionally important in the decrease in AML growth and stemness.

Figure 34. The PISD Inhibitor MMV007285 reduces AML growth and stemness.

Figure 35. TAZ-KD and PISD Inhibitor MMV007285 activates TLR signaling.

Figure 36. TLR8 agonist CL075 reduces stemness in OCI-AMl2 cells and synergizes with MMV007285.

Figure 37. The PISD Inhibitor MMV007285 demonstrates specific anti-leukemic activity.

Figure 38 TAZ knockdown reduces stem cell function of primary AML patient samples.

Figure 39. Related to Figure 38 TAZ knockdown reduces stem cell function of primary AML patient samples.

Figure 40. MMV007285 reduced leukemia burden in xenograft models of human leukemia.

Figure 41. MMV007285 reduced the function of primary AML patient samples.

Figure 42. Summary of Major Conclusions.

Figure 43. Illustration of phospholipid synthesis.

Figure 44. Illustration of the TLR signaling pathway.

xvi

List of Abbreviations

5-FU 5-Fluorouracil

A-site Acceptor site

Acetyl-CoA Acetyl coenzyme A

ADI-PEG20 Pegylated arginine deiminase

ADP Adenosine diphosphate

Akt Akt serine/threonine kinase 1

ALCAT1 Acyl-CoA lysocardiolipin acyltransferase-1

ALL Acute lymphoblastic leukemia

ALP Alkaline phosphatase a-KG a-Ketogluterate

AML Acute Myeloid Leukemia

APL Acute promyelocytic leukemia

ASS1 Arginninosuccinate synthase-1

AST aspartate transaminase

ASXL1 Additional sex comb-like 1

ATP Adenosine triphosphate

ATRA-ATO All-trans retinoic Acid and Arsenic Trioxide

AUP Animal use protocol

BCL-2 B-cell lymphoma

xvii

Bili Bilirubin

BIT Bovine serum albumin-insulin-transferrin

BSA Bovine serum albumin

CAS9 CRISPR associated nuclease 9

CD Cluster of differentiation cDNA Complementary DNA

CDP-DAG Cytidine diphosphate diacylglycerol

CH Clonal Hematopoiesis

CHIP Clonal hematopoiesis of indeterminant potential

CK Creatine kinase

CLL Chronic lymphoblastic leukemia

CLP Chronic lymphocytic leukemia

CML Chronic myeloid leukemia

CMP Common myeloid progenitors

CMPK2 Cytidine monophosphate kinase 2

CMV Cytomegalovirus

CO2 Carbon dioxide

CoA Coenzyme A

COX Cytochrome c oxidase

CPT Carnitine palmitoyltransferase

xviii

Cr Creatinine

CRISPR Clustered regularly interspaced palindromic repeats

CTP Cytidine triphosphate

CYTB Cytochrome bc1 dATP Deoxyadenosine triphosphate dCTP Deoxycytidine triphosphate dGTP Deoxyguanosine triphosphate dGUOK Deoxyguanosine kinase

DHF Dihydrofolate

DHODH Dihydroorotate dehydrogenase

DMEM Dulbecco's Modified Eagle Medium

DNA Deoxyribonucleic acid

DNMT3A DNA methyltransferase 3

DRP1 Dynamin-related protein-1 dTMP Deoxythymidine monophosphate

DTT 1,4-dithiothreitol dTTP Deoxythymidine triphosphate dUMP Deoxyuridine monophosphate

ECAR Extracellular acidification rate

EDTA Ethylenediaminetetraacetic Acid

xix

ELN European Leukemia Net

EOBA Excess over bliss additivism

FBS Fetabl Bovine Serum

FCCP carbonyl cyanide p-trifluoromethoxyphenylhydrazone

FCgR FCg receptors

FIS1 Fission, mitochondrial 1

FITC Fluorescein isothiocyanate

FLT3 Fms related tyrosine kinase 3

Flt3-L Fms-related tyrosine kinase 3 ligand fMET-tRNA Formylated methionyl-transfer-RNA

G-CSF Granulocyte colony-stimulating factor

G3P Glycerol 3-phosphate

GFP Green fluorescent protein

GLS Glutaminase

GM-CSF Granulocyte-macrophage colony-stimulating factor

GPx Glutathione peroxidases

GrMP Granulocyte/macrophage progenitors

GRP56 G protein-coupled receptor 56

GSEA Gene set enrichment analysis

GSH Reduced glutathione

GTP Guanosine triphosphate

GVHD Graft-versus-host disease

GVT Graft-versus-tumor effect

H2O2 Hydrogen peroxide

HBSS Hank's balances salt solution

HCT Hematopoietic cell transplant

HDR Homologous directed repair

HEPES 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid

HPC Hematopoietic progenitor cell

HPTLC High-performance thin layer chromatography

HRP Horseradish peroxidase

HSC Hematopoietic stem cell

IDH Isocitrate dehydrogenase iDOX-Taz-KD B6.Cg-Gt (ROSA) 26Sortm37(H1/tetO- RNAi:Taz)Arte/ZkhuJ doxycycline-Inducible-Tafazzin- Knockdown

IFNb Interferon β

IL Interleukin

IL1RAP Interleukin 1 Receptor Accessory Protein

IMDM Iscove’s modified Dulbecco’s medium

IMP Inosine monophosphate

xxi

IP Intraperitoneally

JAK2 Janus kinase 2

LC3 Microtubule-associated proteins (1A/1B) light chain 3B

Lin Lineage

LMPP Lymphoid-primed multipotent progenitors

LPE lysophosphatidylethanolamine

LSC Leukemic Stem Cell

LT-HSC Long-term hematopoietic stem cells

LYZ Lysozyme

MAGeCK Model-based Analysis of Genome-wide CRISPR/Cas9 Knockout

MAVs Mitochondrial antiviral signaling proteins

MCL-1 Myeloid leukemia cell differentiation protein

MDS Myelodysplastic syndrome

MEP Megakaryocyte/erythroid progenitors

MFI Mean fluorescent intensity

MFN1 Mitofusin 1

MIRO1 Mitochondrial Rho GTPase 1

MLCL Monolysocardiolipin

MMP Mitochondrial membrane permeabilization

MnSODs Manganese-dependent superoxide dismutases

xxii

MPP Multipotent progenitors mtDNA Mitochondrial DNA mtEEFG1 Mitochondrial elongation factor G1 mtEFG2-GTP Mitochondrial elongation factor G2-GTP mtEFTU Mitochondrial elongation factor thermo unstable mtIF Mitochondrial initiation factors mtLSU Mitochondrial large subunit mTOR Mammalian target of rapamycin complex 1 mTORC1 mTOR complex 1 mTORC2 mTOR complex 2 mtRNA Mitochondrial RNA mtRRF Mitochondrial ribosome releasing factor

MTSSB Mitochondrial single-stranded binding protein mtSSU Mitochondrial small subunit

NADH Reduced nicotinamide adenine dinucleotide

NADP Nicotinamide adenine dinucleotide phosphate

NADPH Reduced nicotinamide adenine dinucleotide phosphate

ND NADH dehydrogenase

NGS Next generation sequencing

NHEJ Non-homologous end joining

xxiii

NOD/SCID NOD.CB17-Prkdcscid/J

NOD/SCID-GF NOD.Cg-Prkdcscid Il2rgtm1Wjl Tg (CMV- IL3,CSF2,KITLG)1Eav/MloySzJ

NOX2 NADPH dependent oxidase 2

NOXA BH3 interacting-domain death agonist

NPM1 Nucleophosmin 1

OCR Oxygen consumption rate

OPA1 Optic atrophy 1 mitochondrial dynamin like GTPase

P-site Peptidyl site

PAM Protospacer-adjacent

PARL Presenilin-associated rhomboid-like

PBS Phosphate-buffered saline

PBSCs Peripheral blood stem cells

PC phosphatidylcholine

PCR polymerase chain reaction

PE Phosphatidylethanolamine

PG Phosphatidylglycerol

PGS1 Phosphatidylglycerol phosphate synthase

PI phosphatidylinositol

PI3K Phodphoinositide 3-kinase

PINK1 PTEN-induced kinase 1

xxiv

PISD phosphatidylserine decarboxylase

POLG Polymerase g

POLRMT RNA polymerase mitochondrial

PPARg Peroxisome proliferator associated receptor g

PRPP Phosphoribosyl-a-phosphate

PRPP Phosphoribosyl phosphate

Prx Peroxiredoxins

PS Phosphatidylserine qRT-PCR Quantitative reverse transcriptase-real time polymerase chain reaction

RARa Retinoic acid receptor-a

RNA Ribonucleic acid

ROS Reactive Oxygen Species

RUNX1T1 or AML- Runt-related transcription factor 1, translocated to, 1 ETO

RXR Retinoid X receptor

Sca-1 Stem cells antigen-1

SCF Stem Cell Factor

SCID Prkdcscid

SD Standard deviation

SDHA Succinate dehydrogenase complex subunit A

xxv

SDHB Succinate dehydrogenase complex subunit B

SEM Standard error of the mean

SF3B1 Splicing factors including splicing factor 3b subunit 1 sgRNA Single guide RNA

SRSF2 Serine and arginine rich splicing factor 2

ST-HSC Short-term hematopoietic stem cells

STR short-tandem repeat

TAG Triacylglycerol

TAZ Tafazzin

TBI Total body irradiation

TBS Tris-buffered saline

TBST Tris-buffered saline with Tween 20

TEFM Transcription elongation factor

TEM Transmission electron microscopy

TET2 Tet methylcytosine dioxygenase

TFAM Transcription factor A

TFB2M Mitochondrial transcription factor B

TIM3 T-cell immunoglobulin and mucin-domain containing-3

TK2 Thymidine kinase 2

TLRs Toll-like receptors

xxvi

TMPK2 Thymidine monophosphate kinase 2

TPO Thrombopoietin

U2AF1 U2 small nuclear RNA auxiliary factor 1

UHPLC/MS Ultra-Hight Performance Liquid Chromatography/Mass Spectrometry

UMP Uridine monophosphate

UTP Uridine triphosphate

VDAC1 Voltage dependent anion channel 1

VOD Veno-occlusive disease

WT Wildtype

xxvii

Chapter 1

Literature Review 1.1 Hematopoiesis

1.1.1 Discovery of the Hematopoietic Hierarchy

The ancient Greeks living in Homeric times (1100-800 BC) described blood as being the center of life itself (Meletis and Konstantopoulos, 2010). We now understand the true meaning of this statement, since we now know that blood plays a critical role in the transport of nutrients, the excretion of waste, wound healing and the fighting of infections. Healthy, red blood cells, white blood cells, and platelets are necessary for blood to carry out these respective functions. ~4.9 × 1011 senescent blood cells are eliminated from the body per day. Hematopoiesis, the process by which senescent blood cells are replenished, is therefore essential for the maintenance of a functional blood system (Sou, 2012).

The earliest understanding of hematopoiesis began with Artur Pappenheim’s, Wera Dantschakoff’s, and Alexander Maximow’s postulation that all blood cells are derived from a common, hematopoietic stem cell (HSC) (Konstantinov, 2000; Maehle, 2011). Immune tolerance studies conducted in the 1940s and 1950s supported this ‘unitarian’ view of hematopoiesis (Weissman and Shizuru, 2008). The first of these studies were conducted by Ray Owen in 1945. In this seminal study, he observed that adult dizygotic cattle twins were more likely to have a common blood type compared to siblings, indicating the transfer of erythrocyte progenitors through the placenta in utero (Owen, 1945; Weissman and Shizuru, 2008). Then in the 1950's Medwar and Billingham found that dizygotic cattle twins accepted skin grafts to the same degree as monozygotic twins, and that immune tolerance in fetal as well as neonatal mice can be induced by a hematopoietic cell transplant (HCT), suggesting that the immune system is generated from a common precursor (R E Billingham, 1952; Weissman and Shizuru, 2008).

Scientist in the atomic era began to understand the functional importance of human hematopoietic cells (Doulatov et al., 2012; Lorenz et al., 1952). Hematopoietic failure was found to be the reason underlying the lethal consequences of radiation toxicity after the nuclear bombs in Hiroshima and Nagasaki (Doulatov et al., 2012; Lorenz et al., 1952; Weissman and Shizuru, 2008). Mice experiments investigating the cause of bone marrow failure after radiation showed

1 2 radiation syndrome could be prevented by shielding the spleen with lead, and by injecting spleen as well as marrow cells into irradiated mice from unirradiated donors (Doulatov et al., 2012; E.L.; and J.H., 1951; Jacobson et al., 1950; Weissman and Shizuru, 2008). The aforementioned studies established the regenerative, and functional potential of the hematopoietic system which led to the establishment of the bone marrow transplantation field (Thomas, 1999). However, the identity and properties of the cells responsible for hematopoietic regeneration at this point were still not known.

In 1961, the studies published by Till and McCulloch were the first to experimentally identify and characterize the properties of hematopoietic cells (Till and McCulloch, 1961). Data Till and McCulloch generated using murine repopulation assays suggested the existence of a multipotent hematopoietic stem cell (HSC) that has the ability to self-renew and differentiate. From this time on, murine repopulation assays remain the gold standard to functionally characterize human/murine hematopoietic stem and progenitor cells (HSC/HPC) (Doulatov et al., 2012). Using murine models, the multipotency of HSC’s are demonstrated by the ability of a single HSC to repopulate the myeloid and lymphoid lineages of irradiated hosts. Whereas, the self- renewal potential of HSC’s are demonstrated by the ability of a single HSC to give rise to other HSC’s that can differentiate into all blood lineages in a secondary host (Weissman and Shizuru, 2008). In addition, clonal in vitro assays are used to functionally characterize the stem/progenitor functions of cells in the hematopoietic system, multipotent hematopoietic cells form a greater number, as well as more diverse types of colonies compared to differentiated unipotent hematopoietic cells (Challen et al., 2009; Doulatov et al., 2012).

1.1.2 Current Models of Hematopoiesis

Murine repopulation assays, as well as clonal in vitro assays, resulted in the establishment of the hematopoietic hierarchy, with HSC sitting at the apex and mature blood cells at the base (Figure 1). Furthermore, using these assays the expression of cell surface molecules were found to be highly regulated during the gradual fate restriction of HSC’s/HPC’s, therefore cell surface markers can be used to distinguish between different subpopulations of hematopoietic cells (Civin and Loken, 1987). In the murine system cell surface markers such as CD5, CD45RA, CD11b, Anti-Gr-1, 7/4, and TER-119 are expressed on thymocytes, B-cells, macrophages, monocytes, neutrophils, as well as erythrocytes respectively, and not on hematopoietic stem/progenitor cells. Hence, these lineage (Lin) specific markers are used to delineate mature

3 unipotent effector cells from multipotent hematopoietic stem/progenitor cells as well as oligopotent progenitor cells. Furthermore, the expression of c-KIT and stem cells antigen-1 (Sca- 1) differ between multipotent progenitors and oligopotent progenitors. Self-renewing multipotent long-term hematopoietic stem cells (LT-HSC), short-term hematopoietic stem cells (ST-HSC), as well as non-self-renewing multipotent progenitors (MPP), are c-KIT+ Sca-1+ whereas oligopotent common myeloid progenitors (CMP), Megakaryocyte/erythroid progenitors (MEP), and granulocyte/monocyte progenitors (GrMP) are c-KIT+Sca-1- (Kiel et al., 2005; Oguro et al., 2013; Wilson et al., 2008).

The human hematopoietic hierarchy is very similar to the mouse hematopoietic hierarchy with a few key differences (Figure 2). Firstly, there is a lack of congruence between mouse hematopoietic and human hematopoietic cell surface markers (Doulatov et al., 2012). For instance, human HSC are CD34+ CD38- Thy1+ CD45RA-, whereas mouse HSCs are c-KIT+ Sca- 1+ CD150+ CD48- (Doulatov et al., 2012; Notta et al., 2016; Oguro et al., 2013). Also, in mice CMPs give rise to MEPs and GrMPs. MEPs ultimately differentiate into megakaryocytes and erythrocytes, while GrMPs become granulocytes (Notta et al., 2016). Conversely, in humans, megakaryocytes arise from multipotent progenitor cells instead of oligopotent common myeloid progenitor cells. Despite these differences, the murine model is a good system to begin to understand the molecular and biochemical pathways that underline HSC/HPC functions.

Figure 1 Mouse hematopoietic hierarchy. Phenotypic markers expressed by mouse hematopoietic stem/progenitor cells (HSC/HPC) and mature blood cells adapted from Doulatov et al, 2012. Cell surface markers can be used to separate HSCs from more mature cells, since the expression of cell surface molecules as HSCs mature is a highly regulated process. Abbreviations: CLP, common lymphoid progenitor; CMP, common myeloid progenitors; DC, dendritic cell; ETP, earliest thymic progenitors; Gran, granulocyte; GrMP, granulocyte- macrophage progenitors; LT-HSC, long-term hematopoietic stem cells; MEP, megakaryocyte/erythroid progenitors; Meg, megakaryocyte; Mon, monocyte; MPP, multipotent progenitors; Plt, platelet; RBC, red blood cell; ST-HSC, short- term hematopoietic stem cells.

Figure 2. Human hematopoietic hierarchy. Phenotypic markers expressed by human hematopoietic stem/progenitor cells (HSC/HPC) and mature blood cells Doulatov et al, 2012. Like in the mouse the expression of cell surface molecules during differentiation is highly regulated. However, the cell surface markers expressed on human hematopoietic cells is different to those expressed on mouse hematopoietic cells. Abbreviations: B/NK, B/natural killer cell progenitor, CMP, common myeloid progenitors; DC, dendritic cell; ETP, earliest thymic progenitors;

Gran, granulocyte; GrMP, granulocyte-macrophage progenitors; HSC, hematopoietic stem cells; Meg, megakaryocyte; MLP, immature lymphoid progenitor; Mon, monocyte; MPP, multipotent progenitors; NK, natural killer; RBC, red blood cell.

1.1.3 Emergency Hematopoiesis

HSC/HPC's are critical for the augmentation of blood cells during conditions of natural stress, such as bleeding or infection as well as iatrogenic stress such, as chemotherapy or irradiation (Boettcher and Manz, 2017). Despite these varying etiologies, the term emergency hematopoiesis has been coined to distinguish stress induced hematopoiesis from steady-state hematopoiesis, since cellular and molecular mechanisms are shared during hematopoietic recovery in response to the natural and iatrogenic insults mentioned above (Boettcher and Manz, 2017). It is well established that HSCs are usually quiescent during steady-steady state hematopoiesis, however in response to chemotherapy-induced cytopenia HSCs rapidly self-renew and differentiate until normal blood counts are achieved (Busch et al., 2015; Wilson et al., 2008). Similarly, during an infection pathogen sensing toll-like receptors (TLRs) on HSC/HPC detect the presence of pathogenic components, then proliferate and differentiate to supply the leukocytes, neutrophils and monocytes required to combat the invading microbes (Boettcher and Manz, 2016; Manz and Boettcher, 2014; Takizawa et al., 2012; Yáñez et al., 2013; Zhao and Baltimore, 2015). Despite these immediate benefits of acute HSC/HPC activation, chronic HSC/HPC activation has been shown to impair HSC self-renewal and stemness, which in turn affects blood production (de Bruin et al., 2013; Esplin et al., 2011; Essers et al., 2009; Pietras et al., 2016; Sato et al., 2009). However, the molecular mechanisms that govern the suppression of HSC function are not known (Boettcher and Manz, 2017).

1.2 Leukemia

Cancer arises when a single cell, among the trillions of cells in our body, starts to grow uncontrollably (Mukherjee, 2010). Leukemia refers to cells in the hematopoietic system growing uncontrollably (Ruhl, 2018; Swerdlow, 2016). The types of leukemia are grouped according to the rate at which the disease progresses and the blood lineage that is affected (Ruhl, 2018; Swerdlow, 2016). In acute leukemias, leukemic blasts (malignant blood cells) are progenitor cell-like, and not functional. In addition, blasts proliferate at a fast rate, resulting in the rapid progression of the disease (Ruhl, 2018; Swerdlow, 2016). Whereas in chronic leukemias, the blast cells are more mature, and are able to partially function. Also, blasts increase at a slower rate compared to acute leukemias, resulting in the gradual progression of the disease (Ruhl, 2018; Swerdlow, 2016). Accordingly, there are four main types of leukemia-acute lymphoblastic leukemia (ALL), acute myeloid leukemia (AML), chronic lymphocytic leukemia (CLL), and chronic myeloid leukemia (CML) (Ruhl, 2018; Swerdlow, 2016). The biological distinction

7 between ALL, AML, CLL, and CML is further illustrated by the difference in the cell of origin for each type of leukemia (Figure 3). On average, AML has the worst prognosis compared to other leukemias, with a 5 years survival of approximately 24%. Therefore, our lab is interested in understanding AML biology in order to develop novel therapeutic strategies for AML.

Figure 3. Leukemogenesis. The cell of origin in each type of leukemia. The myeloid leukemias acute myeloid leukemia (AML), and chronic myeloid leukemia (CML) originate from multipotent progenitor cells (MPP). Whereas, the lymphoid leukemias acute lymphoblastic leukemia (ALL), and chronic lymphocytic leukemia (CLL) originate from immature lymphoid progenitors (MLP), and marginal zone (MZ) B cells respectively. Abbreviations: ALL, acute lymphoblastic leukemia; AML, acute myeloid leukemia; CLL, chronic lymphocytic leukemia; CML, chronic myeloid leukemia; HSC, hematopoietic stem cell; MLP, immature lymphoid progenitor; MPP, multipotent progenitors; MLP, immature lymphoid progenitors; MZ B cells, marginal zone B cell.

1.3 Acute Myeloid Leukemia

AML is a heterogenous disease that is characterized by the proliferation of clonal myeloid progenitor cells, that have a reduced ability differentiate into mature blood cells (Dohner et al., 2015; Lowenberg et al., 1999; Thomas and Majeti, 2017). In General, AML is diagnosed when the number of undifferentiated leukemic blast in the bone marrow or peripheral blood exceeds 20% (De Kouchkovsky and Abdul-Hay, 2016). The accumulation of blasts, in addition to the lack of red blood cells, platelets, and granulocytes contributes to the morbidity and mortality of the disease (Thomas and Majeti, 2017). AML is a disease of age-in patients above 65 the incidence of AML increases by four-fold (De Kouchkovsky and Abdul-Hay, 2016). Unfortunately, with current diagnosis algorithms and treatment strategies the prognosis for older AML patients are poor, with 70% of patients over 65 years dying within 1 year of diagnosis (Shah et al., 2013). A better understanding of AML pathogenesis can lead to early detection, and an improvement in AML outcomes.

1.3.1 AML Pathogenesis

AML is a multi-factorial disease that can arise as a consequence of an underlying hematological malignancy such as myelodysplastic syndrome (MDS); or as a result of prior exposure to agents that cause DNA strand breaks-topoisomerase II inhibitors, alkylating agents, as well as radiation (De Kouchkovsky and Abdul-Hay, 2016; Sill et al., 2011). However, in 65-70% of cases AML appears as a de novo malignancy in healthy individuals (Hulegårdh et al., 2015; Renaud et al., 2016). Regardless of the etiology, the pathogenesis of AML involves the abnormal proliferation, and differentiation of a clonal population of HSC/HPCs, resulting in the accumulation of immature myeloid leukemic blast cells in the bone marrow and peripheral blood (De Kouchkovsky and Abdul-Hay, 2016).

The development of AML from a normal blood cell is a complicated process. Recent studies have provided insight into the genetic events responsible for leukemic transformation (Bowman et al., 2018; Gibson and Steensma, 2018; Sellar et al., 2018). For some individuals, clonal hematopoiesis (CH), a mutational events causing the growth of an HSC clone, is a precursory permissive state of AML (Bowman et al., 2018). In this model of leukemogenesis, mutations in an HSC promotes its increased self-renewal, proliferation, and/or reduced cell death, which facilitates the rapid growth expansion of a particular HSC clone compared to other clones. After which, secondary or tertiary mutations occur in MPPs, derived from mutated HSC clones,

9 resulting in the initiation of AML (Figure 4) (Bowman et al., 2018). When CH occurs due to mutations in leukemia driver genes, and is detected at a variable allele frequency (VAF) of at least 0.02 it is referred to as clonal hematopoiesis of indeterminant potential (CHIP) (Gibson and Steensma, 2018). Interestingly, recent studies have found that CHIP is more common in individuals that subsequently develop AML (Abelson et al., 2018; Desai et al., 2018). Risk factors for AML such as, age, genotoxic drugs, and smoking all create a landscape enabling the selection of HSC clones that have a growth advantage thereby facilitating CHIP (Buscarlet et al., 2017; Busque et al., 2012; Coombs et al., 2017; Genovese et al., 2014; Gibson et al., 2017; Gillis et al., 2017; Jaiswal et al., 2014; Jaiswal et al., 2017; Takahashi et al., 2017; Wong et al., 2015; Xie et al., 2014; Zink et al., 2017)

The most common somatic CHIP mutations are present in genes associated with epigenetic modifiers such as DNA methyltransferase 3 a (DNMT3A), tet methylcytosine dioxygenase (TET2) and additional sex comb-like 1 (ASXL) as well as splicing factors including splicing factor 3b subunit 1 (SF3B1), serine and arginine rich splicing factor 2 (SRSF2), and U2 small nuclear RNA auxiliary factor 1 (U2AF1) (Gibson and Steensma, 2018; Inoue et al., 2016). DNMT3A and TET2 loss of function mutations are frequently seen in hematological malignancies (Papaemmanuil et al., 2016). DNMT3A methylates DNA, while TET2 removes methylation marks by converting 5-methylcytosine to 5-hydroxymethylcytosine (Ito et al., 2011; Smith and Meissner, 2013). Loss of DNMT3A function is associated with hypomethylation, expansion of HSC, and the development of lymphoid and myeloid leukemias (Challen et al., 2014). Whereas, loss of TET2 leads to hypermethylation and differentiation blockade in HSCs (Moran-Crusio et al., 2011; Quivoron et al., 2011). Both DNM3TA and TET2 mutations cooperate with oncogenic mutations to facilitate malignant transformation. Large AML cohort studies have identified that DNM3TA is commonly co-mutated with receptor tyrosine kinase fms related tyrosine kinase 3 (FLT3); RAS signaling pathway members NRAS, KRAS, as well as the multifunctional nuclear protein nucleophosmin 1 (NPM1) mutations (Cancer Genome Atlas Research, 2013; Papaemmanuil et al., 2016). In mouse studies, DMNT3A mutations combined with NRAS mutations, FLT3 mutations with or without mutated NPM1 lead to the development of AML (Guryanova et al., 2016; Mayle et al., 2015; Yang et al., 2016) On the other hand, TET2 loss of function mutations have been shown to cooperate with mutations in FLT3, KIT, runt- related transcription factor 1, translocated to, 1 (RUNX1T1, commonly known as, AML- ETO), and Janus kinase 2 (JAK2) during leukemogenesis (De Vita et al., 2014; Ortmann et al., 2015;

Rasmussen et al., 2015; Shih et al., 2015). Evidently, mutational events in HSCs and MPPs give rise to AML. A better understanding of these mutational events can lead to strategies that prevent AML. In addition, to preventative strategies cancer biologist are trying to understand the mechanisms that initiate and propagate AML to develop new treatments for AML.

1.3.2 Leukemia Stem Cells

Like hematopoiesis AML is also hierarchical with the leukemic stem cell (LSC) at the apex (Figure 4) (Tan et al., 2006; Thomas and Majeti, 2017). Similar to HSCs, LSCs self-renew and differentiate, two properties necessary for driving the initiation and propagation of AML. The earliest evidence alluding to the hierarchical organization of AML came from the in vitro clonogenic studies performed in the 1960s and 1970s, by the Metcalf and Warner groups (Donald et al., 1969; Moore et al., 1973). The self-renewal and differentiation capacity of LSCs in vivo was only established in the 1990s by Dick and colleagues (Bonnet and Dick, 1997; Dick, 2005; Lapidot et al., 1994). These seminal studies by the Dick group established the utilization of primary and secondary engraftment assays in immunodeficient mice to demonstrate the differentiation and self-renewal capacity of LSCs, respectively (Kreso and Dick, 2014).

Similar to normal blood, cell surface markers can be used to delineate LSCs from AML blasts. The majority of LSCs reside in the CD34+38- cell fraction (Bonnet and Dick, 1997). Interestingly, LSCs are also present in the CD34+CD38+, as well as in the 34- fractions (Kreso and Dick, 2014). Both the CD34+CD38-, and CD34+CD38+ LSC populations have been shown to coexist in AML patient samples (Goardon et al., 2011). The CD34+CD38- fraction resemble normal lymphoid-primed multipotent progenitors (LMPP), while the CD34+CD38+ fraction resembles normal GrMP (Goardon et al., 2011). Interestingly, the LMP-like LSC population was able to give rise to the GrMP-like LSC population, however the converse was not possible, suggesting that LMP-like LSC represent a more primitive LSC population. In contrast, the relationship between the 34+ and 34- LSC subpopulations are not hierarchal, but instead represent the same population of cells with CD34 plasticity (Quek et al., 2016).

Figure 4. AML pathogenesis and hierarchy. Mutations implicated in the generation of a leukemia stem cell (LSC) that both initiates and propagates the disease. Mutations in epigenetic modifiers and splicing factors cooperate with oncogenic mutations during leukemogenesis. that of Adapted from Tan et al, 2006. Abbreviations: AML-ETO, Runt- related transcription factor 1, translocated to, 1; ASXL additional sex comb-like 1; CMP, common myeloid progenitors; DC, dendritic cell; DNM3TA, DNA methyltransferase 3 a; FLT3, Fms related tyrosine kinase 3; Gran, granulocyte, HSC, hematopoietic stem cell; JAK2, Janus kinase 2; Mon, monocyte; Meg, megakaryocyte; MPP, multipotent progenitors, MLP, immature lymphoid progenitor, NK, natural killer; NPM1, nucleophosmin 1; RBC, red blood cell; SF3B1, splicing factor 3b subunit 1; SRSF2, serine and arginine rich splicing factor 2; TET2, tet methylcytosine dioxygenase; U2AF1, U2 small nuclear RNA auxiliary factor 1.

AML patient derived LSCs have been compared to normal hematopoietic HSC in order to identify markers that are differentially upregulated on LSCs. Although several makers have been identified, none of which is unique to LSCs due to close resemblance of LSCs to HSCs. In spite of these challenges, markers such as: CD123, CD44, CD47, T-cell immunoglobulin and mucin- domain containing-3 (TIM3), CD96, CD99, c-type lectin-like molecule-1, CD32, CD25, Interleukin 1 Receptor Accessory Protein (IL1RAP), G protein-coupled receptor 56 (GRP56), and CD93 have been shown to be upregulated LSCs compared to normal HSPCs (Ågerstam et al., 2015; Askmyr et al., 2013; Bonardi et al., 2013; Chung et al., 2017; Hosen et al., 2007;

Iwasaki et al., 2015; Jaiswal et al., 2009; Jan et al., 2011; Jin et al., 2006; Jin et al., 2009; Kikushige et al., 2010; Majeti et al., 2009; Moshaver et al., 2008; Pabst et al., 2016; Saito et al., 2010; van Rhenen et al., 2007). Further investigation into LSC specific markers is vital for the improvement of LSC purification as well as the assessment of LSC specific activity of anti- leukemic agents.

Since LSCs have the capacity to initiate and propagate leukemia, patient outcome should be linked to the amount and the activity of LSCs. Initial evidence demonstrated that the overall survival of patients was negatively correlated with the frequency of LSCs present in their disease (Pearce et al., 2006; van Rhenen et al., 2005). Patients whose samples had a high frequency of cells with a stem cell phenotype (CD34+,CD38-), and whose samples engrafted in NOD/SCID mice had a poor overall survival (Pearce et al., 2006; van Rhenen et al., 2005). Furthermore, studies that established LSC gene-signatures using AML patient LSC samples functionally validated in immunodeficient mice, showed that stemness gene signature scores were independent predictors of overall survival (Ng et al., 2016). However, LSC signatures are not used to clinically characterize AML. Instead, AML is currently characterized using chromosomal abnormalities, genetic mutations, and cell morphology.

1.3.3 Classification of AML

The French-American-British classification system was the first attempt to distinguish between different subtypes of AML (Table 1) (De Kouchkovsky and Abdul-Hay, 2016; Lowenberg et al., 1999). It divides AML into 9 distinct subtypes based on the morphologic appearance of leukemic blasts, and their reactivity with the histochemical stains (myeloperoxidase, Sudan black, and the nonspecific esterases α-naphthylacetate and naphthylbutyrate) (Lowenberg et al., 1999). However, the French-American-British classification system is limited in assessing the molecular characteristics of the disease (Dohner et al., 2015). Consequently, in 2001, the World Health Organization, introduced a new classification system to incorporate improvements in the diagnosis and management of AML (De Kouchkovsky and Abdul-Hay, 2016). The WHO revised its classification system in 2008, and 2016 (Arber et al., 2016; Vardiman et al., 2009). The most recent version (Table 2) uses genetic information, morphology, immunophenotyped, and clinical presentation to divide AML into six major disease entities: (1) AML with recurrent genetic abnormalities; (2) AML with myelodysplasia-related features; (3) therapy-related AML; (4) AML not otherwise specified; (5) myeloid sarcoma; (6) myeloid proliferation related to

Down syndrome. AML with recurrent genetic abnormalities can be further divided into 11 subtypes, according to distinct chromosomal translocations as well as genetic mutations (Arber et al., 2016). Genetic abnormalities also inform the diagnosis AML with myelodysplasia-related changes. Finally, the European Leukemia Network’s recommendations integrates both cytogenic and molecular features to divide AML patients into favorable, intermediate, and adverse risk prognostic groups that vary according to rates of complete remission, disease-free survival, and overall survival (Dohner et al., 2017; Rotin, 2016). Current treatment decisions are informed by both the World Health Organization and the European Leukemia Network’s characterization systems.

FAB Classification Definition Description M0 Undifferentiated acute Leukemia cells that are myeloblastic leukemia minimally differentiated M1 Acute myeloblastic leukemia Myeloblasts are the with minimal maturation dominant leukemia cell M2 Acute myeloblastic leukemia There are many with maturation myeloblasts present, but there are some cells differentiating towards fully formed blood cells M3 Acute promyelocytic When Leukemic cells have leukemia a translocation between chromosome 15 and 17 M4 Acute myelomonocytic Leukemic cells that usually leukemia have a translocation, an inversion, or deletion of chromosome 16 M4eo Acute myelomonocytic Leukemic cells that usually leukemia with eosinophilia have a translocation or an inversion of chromosome 16, and have features of developing esonophils M5 Acute monocytic leukemia Leukemic cells have features of maturing monocytes M6 Acute erythroid leukemia Leukemic cells have features of developing red blood cells M7 Acute megakaryoblastic Leukemic cells have leukemia features of developing platelets Table 1. French-American-British classification of AML

AML Disease Entity Subtype AML with recurrent genetic abnormalities AML with t(8;21) (q22;q22); RUNX1- RUNX1T1 AML with inv(16) (p13.1q22) or t(16;16)(p13.1;q22); CBFB-MYH11 Acute promyelocytic leukemia (APL) with PML-RARA AML with t(9;11) (p21.3;q23.3); MLLT3- KMT2A AML with t(6;9)(p23;q34.1); DEK- NUP214 AML with inv(3) (q21.3q26.2) or t(3;3)(q21.3;q26.2); GATA2,MECOM AML (megakaryoblastic) with t(1;22) (p13.3;q13.3); RBM15-MKL1 AML with BCR-ABL1 AML with mutated NPM1 AML with biallelic mutations of CEBPA AML with mutated RUNX1 AML with myelodysplasia-related features N/A Therapy-related AML AML not otherwise specified M0 M1 M2 M4 M5 M6 M7 Acute basophilic leukemia Acute panmyelosis with myelofibrosis Myeloid Sarcoma N/A Myeloid proliferations related to Down Transient abnormal myelopoiesis Syndrome AML of Down syndrome

Table 2 World Health Organization classification of AML. Reproduced with permission of the American Society of Hematology, from Dohner et al, 2017, license ID: 4635961338760.

Risk Group Genetic Abnormality Favourable • t(8;21) (q22;q22.1); RUNX1-RUNX1T1 • inv(16) (p13.1;q22) or t(16;16) (p13.1;q22); CBFB- MYH11 • Mutated NPM1 without FLT3-ITD or with low allelic ratio (<0.5) of FLT3-ITD • Biallelic mutated CEBPA Intermediate • Mutated NPM1 and high allelic ratio (>0.5) of FLT3- ITD • Wild-type NPM1 without FLT3-ITD or with low allelic ratio (<0.5) of FLT3-ITD (without adverse-risk genetic lesions) • t(9;11)(p21.3;q23.3); MLLT3-KMT2A • Cytogenetic abnormalities not classified as favorable or adverse Adverse Risk • t(6;9)(p23;q34.1); DEK-NUP214 • t(v;11q23.3); KMT2A rearranged • t(9;22)(q34.1;q11.2); BCR-ABL1 • inv(3)(q21.3;q26.2) or t(3;3)(q21.3;q26.2); GATA2, MECOM (EVI1) • Monosomy 5 or del(5q); monosomy 7; monosomy 17/abn(17p) • Complex karyotype, monosomal karyotype • Wild-type NPM1 and high allelic ratio (>0.5) of FLT3- ITD • Mutant RUNX1, ASXL1, or TP53

Table 3 European Leukemia Network classification of AML. Reproduced with permission of the American Society of Hematology, from Dohner et al, 2017, license ID: 4635971363793.

1.3.4 Treatment Modalities of AML

Broadly, AML treatment is divided into remission and post-remission therapy (Kanate et al., 2014). Cytotoxic therapies are the main remission therapy for AML patients, Whereas, cytotoxic therapies or allogenic-hematopoietic cell transplantation are used as post-remission therapies (Dombret and Gardin, 2016; Kanate et al., 2014). APL is the exception, it is successfully managed with the differentiation therapy retinoic acid in combination with arsenic trioxide.

1.3.4.1 Cytotoxic Therapies

Currently, the continuous intravenously administration of the anti-metabolite cytarabine for seven days in combination with daunorubicin given on days 1, 3, and 5 is the front-line induction therapy for the majority of favorable and intermediate risk AML patients (Dombret and Gardin, 2016). Cytarabine is a synthetic pyrimidine nucleoside that is converted intracellularly to its active form cytarabine triphosphate, once converted it competes with deoxycytidine triphosphate for the active site of DNA polymerase to inhibit DNA replication. As a result, cells in the s-phase of the cell cycle that are undergoing a high degree of DNA replication are preferentially targeted.

Daunorubicin is an anthracycline isolated from the soil bacteria specie Streptomyces and works by targeting topoisomerase II. Daunorubicin binding to topoisomerase II stimulates the formation of the topoisomerase II-cleavable complex, a transient configuration of topoisomerase II where it is covalently attached to DNA and causes cytotoxic DNA double-strand breaks (Thakur, 2011). In addition, daunorubicin also induces cell death by intercalating into DNA, generating free radicals, DNA cross-linking, and the disrupting of DNA helicase activity. Unlike cytarabine the cytotoxic activity of daunorubicin is not restricted to cells in a specific-phase of the cell cycle.

AML patients who are in the adverse risk group are brought into remission with the FLAG-IDA treatment regimen. As part of the FLAG-IDA protocol, patients are treated with the cytotoxic agents: idarubicin an analogue of daunorubicin for 3 days; the adenine analogue fludarabine for 5 days; and cytarabine for 5 days. As part of the intensive FLAG-IDA protocol, AML patients are given granulocyte-colony stimulating factor, a stimulator of granulocyte production, for 6 days to prevent infection and neutropenic fevers.

Sixty-80% of AML patients achieve remission after induction therapy, which is clinically defined as having-<5% blast in the bone marrow, no blast with Auer rods, normal maturation of all cellular components in the bone marrow, no extramedullary disease, a neutrophil count of

>1,00/ul, a platelet count of >100,000/ul, as well as transfusion independence for at least four weeks (CIBMTR, 2010/2011; Döhner et al., 2015). However, without additional cytotoxic therapy the majority of patients will relapse within an average of four to eight months (Cassileth et al., 1988). Thus, post-remission therapy is used to prevent relapse and prolong survival.

The goal of post-remission therapy is to kill leukemia cells that survive induction chemotherapy but are undetectable, by conventional investigations that measure disease burden. The specific post-remission treatment used depends on the cytogenetic profile, functional performance status, and the age of the patient. Intermediate-dose cytarabine is most widely used in favourable risk AML patients less than 60 years of age with good performance status, it cures 60-70% of these patients (Döhner et al., 2015).While, hematopoietic cell transplantation (HCT) is recommended for intermediate and high risk AML patients, since they are unlikely to have an extended complete remission with conventional chemotherapeutic approaches (Döhner et al., 2015). Moreover, intermediate and adverse risk AML patients have a significant survival benefit when treated with HCT compared to other consolidation therapies (Koreth et al., 2009; Ossenkoppele et al., 2016).

1.3.4.2 Hematopoietic Cell Transplant

HCT was first performed in 1968, by Donald E. Thomas’ group in Seattle, it was used as a treatment of last resort for hematological malignancies (Negrin, 2014; Savani et al., 2011; Thomas et al., 1975). As a result of the improved safety, and an increase in the number of HCT donors, it is now a standard treatment for intermediate and poor risk AML patients (Gluckman et al., 1997; Majhail, 2017; Majhail et al., 2012; Negrin, 2014). The HCT procedure involves removing the patients diseased hematopoietic cells with chemotherapeutic drugs, then transplanting healthy hematopoietic cells into the patient’s bone marrow. There are two types of HCT based on the source of hematopoietic cells: 1) in allogenic-HCT hematopoietic cells come from a matched donor; 2) in autologous-HCT hematopoietic cells are from the patient who has the disease. Out of the two types, allogenic-HCT is used for the treatment of AML. The hematopoietic cells used in allogenic-HCT can come from the bone marrow, peripheral blood or cord blood of either a twin, related, or unrelated donor.

Allogenic-HCT treats AML by rescuing AML patients from hematological toxicity induced by high-dose chemotherapy, and by attacking the patient’s tumor, a phenomenon known as graft- versus-tumor effect (GVT). There are 5 district phases in the allogenic-HCT procedure,

19 conditioning, stem cell infusion, neutropenia, engraftment, and post-engraftment, each phase has a unique set of complications.

1.3.4.2.1 Conditioning Phase

During the conditioning phase patients are treated with chemotherapy, radiation or a combination of the two to eliminate malignancy, create space for the new cells, and to prevent rejection of the donor’s hematopoietic cells. Conditioning regiments can be either total body irradiation (TBI)- based or chemotherapy-based. The combination of conditioning modalities used dictates the intensity of the regiment, and the GVT effect required to cure the disease. Higher the intensity of the conditioning regiment, higher the toxicity, and less GVT is required to cure the disease (Fig 5) (Deeg and Sandmaier, 2010; Gyurkocza and Sandmaier, 2014). Currently, the most common high-intensity, TBI-based, conditioning regiment used is TBI in combination with cyclophosphamide. Agents such as busulfan, cytarabine (AraC), etopside, melphalan, have been used instead of cyclophosphamide, in TBI-based conditioning regiments (Moore, 2017).There are both acute and long-term complication associated with TBI-based regimens. The acute complications include: nausea, vomiting, transient acute parotiditis, xerostomia, mucositis, and diarrhea. Chronic complications such as infertility, cataract formation, hyperthyroidism, thyroiditis, and secondary malignancies are also observed (Gyurkocza and Sandmaier, 2014). Chemotherapy-based regiments are used in some centers to avoid the toxicities associated with high-intensity, TBI-based regimens. Alkylating agents such as busulfan, cyclophosphamide, and melphalan are used in specific combinations-busulfan + cyclophosphamide or busulfan + melphalan-due to their favourable toxicity profile and cytotoxic effect on malignant cells (Gyurkocza and Sandmaier, 2014). It should be noted that treatments with alkylating agents can cause hepatic veno-occlusive disease (VOD), a potentially life-threatening complication (Dalle and Giralt, 2016; Ho et al., 2008).

Figure 5. Conditioning regiments for allogenic-hematopoietic cell transplant (allogenic -

HCT). Conditioning regimens for allogenic-HCT is categorized according to intensity, the requirement of graft-versus-tumor (GVT) effect for disease control, and toxicity. The higher the toxicity of the conditioning regiment the lower the toxicity required to cure the disease. Adapted from Gyurkocza and Sandmaier, 2014. Abbreviations: ATG, anti-thymocyte globulin; BU busulfan, TBI*, high-dose total body irradiation, TBI†, low-dose total body irradiation.

Studies in the 1980’s and 1990’s demonstrated that the presence of immune cells in hematopoietic grafts were associated with lower relapse rates, indicating that the GVT effect can be successfully utilized to target the tumor (Fefer et al., 1987; Gorin et al., 1996; Horowitz et al., 1990). As a result, reduced-intensity and non-myeloablative regiments were developed to promote GVT effects by preserving the immune cells of the patient as well as the graft. The reduced toxicity associated with reduced intensity and non-myeloablative regimens made allogenic-HCT accessible to older and medically infirm patients, who previously were not considered candidates for high-intensity conditioning. Both reduced-intensity/non-myeloablative conditioning regimens and high intensity conditioning regimens have very similar rates of relapse free survival in older and medically infirm patients (Appelbaum, 2009). As a result of the establishment of these regiments the number of patients >50 years treated with allogenic-HCT increased more than any other age group (Figure 6). Reduced-intensity/non-myeloablative

21 conditioning regiments were created by replacing the toxic chemotherapeutic agents in high-dose regiments with less toxic immunosuppressive agents (Figure 7). For instance, the toxicity associated with fludarabine + busulfan is less than that of cyclophosphamide + busulfan. Commonly used reduced intensity regiments include: fludarabine + melphalan, fludarabine + busulfan + anti-thymocyte globulin (ATG, antibodies against human T cells); whereas, low-dose TBI, low-dose TBI + fludarabine are commonly used non-myeloablative regiments (Gyurkocza and Sandmaier, 2014).

Figure 6. HCT trends by type and recipient age. The proportion of patients >50 treated with allogenic -HCT has increased. Data is segmented into 2002-2006 and 2007-2011, Center for International Blood and Marrow Transplant Research (CIBMTR). Reproduced with approval of CIBMTR from CIBMTR summary slides 2013. (Pasquini and Wang, 2013).

Figure 7. Allogenic-HCT trends by conditioning regimen intensity and age. More reduced intensity transplants are performed in patients who are ≥50 compared to those <50. CIBMTR data from 2001-2011. Reproduced with approval of CIBMTR from CIBMTR summary slides 2013. (Pasquini and Wang, 2013).

1.3.4.2.2 Stem Cell Infusion

Stem cell infusion proceeds the conditioning phase. The cell product used is either infused directly or processed before infusion. In the haplotype-matched transplant setting T cells are usually depleted before transplant (Moore, 2017). Some centers are investigating the ex vivo expansion of a specific population of cells before transplant in order to improve engraftment. In general, patients are premedicated with acetaminophen and diphenhydramine before the infusion of cells to prevent an anaphylactic reaction. Complications of the procedure include anaphylaxis, volume overload, transient graft-versus-host disease (GVHD), a condition whereby the donor’s immune cells attack the transplant recipients’ tissues. Furthermore, the dimethyl sulfoxide, the preservative used in the cryopreservation of stem cells has the potential to cause renal failure (Moore, 2017).

1.3.4.2.3 Neutropenia Phase

Before the donor’s cells engraft, the recipient is neutropenic and has no immune system for about 2-4 weeks, making the transplant recipients susceptible to infections (Moore, 2017). Early in the neutropenic phase, HCT recipients are susceptible to reactivation of the herpes simplex virus, as well as infections caused by endogenous gut and skin flora. Hospital-acquired nosocomial infections are the most concerning infection, because they are resistant to standard antibiotic regimens. To prevent infection broad-spectrum antibiotic therapy, as well as anti-fungals such as amphotericin are administered to transplant recipients. In addition, ensuring that the transplant recipient’s nutritional needs are met is another concern during the neutropenic phase. Oral intake is restricted during this period because of the severe mucositis that most transplant recipients develop. As a result, total parenteral nutrition is necessary to ensure that HCT recipients are sufficiently nourished.

1.3.4.2.4 Engraftment Phase

After 2-4 weeks of neutropenia the donor’s cells engraft in the recipient’s bone marrow (Moore, 2017). During this period patients are at a very high risk for GVHD, viral infections such as cytomegalovirus (CMV), as well as VOD.

GVHD which was first described in 1962 remains a significant complication of allogenic-HCT. Clinicians arbitrarily refer to GVHD that takes place prior to 100 days as acute GVHD, and GVHD occurring after 100 days as chronic GVHD (Toubai et al., 2008). Accordingly, acute

GVHD manifests in the transplant phase of the allogenic-HCT procedure. During acute GVHD, the donors T cells are activated by disparate host antigens, that are released by the host’s tissues, after allogenic-HCT conditioning regimens (Ferrara and Reddy, 2006; Sun et al., 2007). Activated T cells then launch an immune attack against the epithelial cells of the skin, gastrointestinal tract, and liver (Ferrara and Reddy, 2006; Sun et al., 2007). As a result, the clinical features of GVHD are epithelial cell damage in the skin, gastrointestinal tract, and liver (Sale, 1996). Prophylaxis is the most effective approach to manage acute GVHD. Although there is no unanimity in the strategies used, the aim during GVHD prophylaxis is to directly or indirectly reduce T cell activity (Table 4) (Ram and Storb, 2013). It is important to note that clinicians don’t completely abrogate acute GVHD because patients who developed acute GVHD have an improved relapse-free survival through GVT (Weiden et al., 1979; Weiden et al., 1981). Strategies to prevent GVHD while preserving GVT is currently an active area of research (Chang et al., 2018). Drug Mechanism of Action Steroids (Methylprednisolone or • Activates glucocorticoid response elements prednisone) • Down-regulates proinflammatory cytokines • Has direct lyphotoxic effects (including T cell toxicity) Methotrexate • Inhibits dihydrofolate reductase to interfere with purine and thymidylate synthesis • Targets proliferating cells including antigen activated T cells Cyclosporine • Inhibits the production of IL-2 downstream of the T cell receptor • As a result, T cell proliferation is reduced Tacrolimus • Inhibits the production of IL-2 downstream of the T cell receptor • As a result, T cell proliferation is reduced Sirolimus • Inhibits cell cycle progression of cells including T cells • Promotes the apoptosis of T cells by inhibiting cell cycle entry in response to IL-2 Mycophenolate mofetil • Inhibits purine metabolism • Targets T and B cell proliferation specifically Antithymocyte Globulin • Depletes T cell lymphocytes Aletuzumab • Antibody against CD52, an antigen on T and B lymphocytes • Binds to T and B lymphocytes and stimulates antibody dependent cellular mediated lysis

Table 4 GVHD prophylaxis strategies. Adapted from Ram and Strob, 2013.

VOD has an incidence of 8-14%, during which the toxic metabolites generated by alkylating agents used in the conditioning phase causes injury to sinusoidal endothelial cells (Coppell et al., 2010). Consequently, local cytokines are released resulting in the activation of the coagulation pathway (Ho et al., 2008). Fibrosis of sinusoids, perivascular hepatocyte necrosis, and venular blockage ensues. Severe hepatic VOD is characterized by ascites and hypertension is life threatening, with a mortality rate of 80% (Coppell et al., 2010). VOD management primarily involves supportive care using, diuretics, dialysis and oxygen, with the goal of minimizing extracellular fluid overload without deteriorating renal function (Coppell et al., 2010). The antithrombotic agent defibrotide is used in Europe for VOD prophylaxis/treatment. Surprisingly, there are currently no VOD prophylaxis/treatments approved by the FDA, highlighting an unmet clinical need (Richardson et al., 2010).

CMV is observed in 30-50% allogenic-HCT and contributes to morbidity and mortality during the engraftment phase (Sissons and Carmichael, 2002). The majority of CMV infections in allogenic-HCT is because of the reactivation of a latent CMV (Pillay et al., 1992). CMV infects epithelial cells, hematopoietic cells, and connective tissue causing the systemic spread of the infection (Minton, 2010). Clinically, CMV can manifest as pneumonia, a gastrointestinal infection, central nervous system infection, or retinitis (Bhat et al., 2015). To prevent CMV during transplant, CMV immunoglobulin G levels are measured in both the donor and the recipient, so that CMV negative donors can be selected for CMV negative recipients, while CMV positive donors can be selected for CMV positive recipients (Bowden et al., 1995). As an additional precaution, CMV DNA is monitored once a week from days 10-100 post-transplant. The anti-virals ganciclovir, acyclovir or cidofovir are administered if CMV DNA rises five-fold above baseline levels (Ljungman et al., 2011; Reusser et al., 2002). In the event that CMV infection occurs gangcyclovir or foscarnet is used to treat gastrointestinal CMV(Boeckh and Ljungman, 2009). While, central nervous system CMV, and retinitis CMV are treated with gangcyclovir and foscarnet (Bhat et al., 2015). Pneumonia CMV on the other hand is treated by intravenous immune globulin in combination with gangcyclovir or foscarnet (Bhat et al., 2015).

1.3.4.2.5 Post-Engraftment Phase

In the post-engraftment phase the patient’s immune system gradually becomes reconstituted. Most patients require reimmunization during this phase, which usually begins with tetanus. If a protective titer is obtained from the tetanus vaccine haemophilus influenzae, pneumococcal, and

25 hepatitis B vaccine series is then administered. Also, within the first year the inactivated polio vaccine is administered. Then two years after transplant, allogenic-HCT recipients are given the measles, mumps, and rubella (MMR) vaccine.

A prominent complication during the post-engraftment phase of transplant is Chronic GVHD. It occurs either as an extension of acute GVHD, following a GVHD free interval, or in the absence of prior GVHD (Toubai et al., 2008). Clinically, chronic GVHD is characterized by fibrosis in either the skin, mouth, eyes, vagina, esophagous, liver, lung, and/or kidneys (Shulman et al., 2006). Chronic GVHD is typically managed with prednisone along with an adjunct treatment in Table 5, depending on the organ affected (Couriel et al., 2006; Lee et al., 2003).

Allogenic-HCT is an extremely high-risk procedure due to complication of the procedure as well as the risk of relapse. Complications such as GVHD and infections contributes to a non-relapse mortality risk ranging from 5-34%. In addition, the chemotherapeutic agents used to condition AML patients before allogenic-HCT cannot completely inhibit the disease propagative function of AML stem cells, as evident by a 15-50% risk of AML relapse (Cornelissen et al., 2012). Moreover, 90% of allogenic-HCT recipients, who have survived the procedure, have more than one chronic condition, that negatively impacts their morbidity and mortality (Hilgendorf et al., 2015; Inamoto and Lee, 2017). Thus, allogenic-HCT although curative, has a poor risk to benefit ratio, which is in contrast to the all-trans retinoic acid (ATRA) and arsenic trioxide (ATO) differentiation treatment strategy for a sub-type of AML, APL.

Organ Symptom Ancillary Treatments Used

Mouth Caries Topical fluorides

Ulceration and Clobetasol leukoplakia

Eyes Keratoconjunctivitis Artificial tears, topical steroids sicca syndrome

Vagina Vulvovaginal Topical steroids GVHD

Vulvovaginal Topical estrogen symptoms and low estrogen status

Gastrointestinal Dysphagia Lubrication, esophageal dilation Tract Diarrhea Pancreatic enzyme replacement

Abnormal liver Ursodeoxycholic acid function test

Pulmonary Pulmonary GVHD Inhaled corticosteroids, inhaled bronchodilators

Musculoskeletal Fasciitis, Physical therapy Contractures, and Steroid Myopathy

Osteoporosis Calcium, vitamin D, antiresorptive therapy

Table 5: Adjuvant treatments for GVHD.

1.3.4.3 All-trans retinoic Acid and Arsenic Trioxide (ATRA-ATO)

APL, characterized by a specific t(15,17) chromosomal translocation, accounts for 10-15% of AMLs (de The, 2018). The translocation of t(15,17) codes for the promyelocytic leukemia (PML)-retinoic acid receptor-a (RARa) fusion protein that drives APL progression (Rowley et al., 1977). In normal physiological conditions RARa binds with its coreceptor retinoid X receptor (RXR) and then binds retinoic acid response elements in DNA to modulate gene expression (Das et al., 2014). In the presence of retinoids, the RXR-RARa complex activates histones acetylated transferases, and thereby promotes the acetylation of histones as well as the expression of differentiation genes (Gudas and Wagner, 2011; Wei, 2003). In the absence of retinoids the RXR-RARa complex binds histone deacetylase and inhibits differentiation by facilitating chromatin condensation in regions of DNA that contain differentiation genes (Das et al., 2014; Gudas and Wagner, 2011; Wei, 2003). PML on the other hand, activates p53 which stimulates the apoptosis of cells, in response to apoptotic stimuli such as DNA damage (Wendel et al., 2006; Zhao et al., 2010). In APL, the PML coiled-coil domain causes the RARa dimer to have less DNA-binding specificity and increases its binding with RXR (Martens et al., 2010; Sunami et al., 2017). Furthermore, PML-RARa fusion protein is insensitive to physiological concentrations of retinoids (Yuan et al., 2016). As a result, the PML-RARa-RXR complex recruits histone deacetylase and represses differentiation genes (Das et al., 2014; de The, 2018). In addition, the PML domain of the PML-RARa fusion protein blunts PML driven p53 apoptosis by binding wild type PML (Insinga et al., 2004). Reductions in p53 driven apoptosis contributes to the increased self-renewal of APL stem cells.

ATRA-ATO cures over 95% of APL patients by stimulating differentiation and reducing the self-renewal of APL (Burnett et al., 2015; Cicconi et al., 2016; Lo-Coco et al., 2013; Lo-Coco et al., 2016). Mechanistically, ATRA binds to the RARa portion of the PML-RARa fusion protein (de The, 2018; Huang et al., 1988). This leads the dissociation of PML-RARa with histone deacetylases allowing for chromatin opening in regions containing differentiation genes (Das et al., 2014; de The, 2018; Huang et al., 1988). ATO then degrades the PML-RARa complex which promotes p53 driven inhibition of APL stem cell self-renewal (de The, 2018).The success of the ATRA-ATO therapy highlights the efficacy of differentiation therapies in the clinic. Improving our understanding of biological mechanisms that increases the self-renewal and blocks the differentiation of cancer cells can lead to novel cancer therapies. Interestingly, there is a large

28 body of work showing that metabolic processes and mitochondrial pathways play a critical role in regulating cancer self-renewal and differentiation.

1.4 Energy Metabolism

Energy is essential for the maintenance of cellular homeostasis. Adenosine triphosphate (ATP), the main energy currency of cells, is produced by catabolic processes. Glucose, proteins, and fats are the main sources of ATP. The series of catabolic reactions the cell uses to produce ATP depends on the ATP source.

1.4.1 Catabolism

Glucose catabolism or oxidation is composed of three main steps – glycolysis, citric acid cycle, and oxidative phosphorylation. Glycolysis begins with glucose entering the cells through a glucose transporter, a series of enzymatic reactions then subsequently (Figure 8) breaks down glucose into two pyruvate molecules, while generating ATP and reduced nicotinamide adenine dinucleotide. (NADH) (Alberts, 2014). In the presence of oxygen, the mitochondrial pyruvate carrier, transports pyruvate from the cytoplasm into the mitochondria, where it is converted to acetyl coenzyme A (acetyl-CoA). Then in the mitochondrial matrix, the acetyl group in acetyl-CoA is transferred to an oxaloacetate molecule to form citrate, which enters the citric acid cycle. The eight-step citric acid cycle generates ATP as well as the electron carriers, NADH and reduced flavin adenine dinucleotide (FADH2). Respiratory complexes I-IV in the inner mitochondrial membrane uses the electrons from NADH, and FADH2 to create an electrochemical gradient, by transferring protons from the matrix to the intermembrane space. The electrochemical gradient created powers complex

V, which catalyzes the synthesis of ATP by reducing oxygen into H2O (Alberts, 2014; Weinberg and Chandel, 2015)

Oxidative Phosphorylation

III I II V IV

Citric Acid Cycle

Figure 8. Glucose catabolism in the cell. Biochemical reactions in three main steps of glucose catabolism- (1) glycolysis, (2) citric acid cycle, and (3) oxidative phosphorylation (Green, 2014). Abbreviations: ATP, adenosine triphosphate; CoA,

coenzyme A; CO2, carbon dioxide; FAD, flavin adenine dinucleotide; FADH2

reduced flavin adenine dinucleotide; NAD nicotinamide adenine dinucleotide; NADH reduced nicotinamide adenine dinucleotide.

Fatty acids and proteins on the other hand are not broken down by glycolysis, but instead directly enter the citric acid cycle to be converted into ATP (Figure 9). Similar to that of glucose, the conversion of fat, and protein into ATP requires the generation of an electrochemical gradient in the mitochondria. The ATP produced by catabolism is then consumed by cells to carryout unfavorable anabolic reactions, such as, the synthesis of nucleic acids, proteins, and lipids.

Figure 9. Catabolism of amino acids and fatty acids. Unlike glucose amino acids and fatty acids are not broken down by glycolysis, instead the oxidation amino acids and fatty acids begins at the level of the citric acid cycle. Reproduced with approval of the W.W. Norton & Company, Inc. from Alberts, 2014.

Abbreviations: ATP, adenosine triphosphate; CoA, coenzyme A; CO2, carbon dioxide, H2O water; NADH, reduced nicotinamide adenine dinucleotide.

1.4.2 Anabolism

Glycolysis and the citric acid cycle are the two primary pathways that make the building blocks for macromolecules such as, nucleic acids, lipids, and proteins (Lunt and Vander Heiden, 2011). Our current understanding of carbon metabolism through glycolysis as well as the citric acid cycle, and its connections to macromolecular biosynthesis is depicted below (Figure 12).

Figure 10 Macromolecular biosynthesis pathways. Intermediates of both glycolysis and the citric acid cycle are used to synthesize nucleotides, lipids, and amino acids. Reproduced with permission of the Annual Reviews, Inc., from Lunt and Vander Heiden, 2011, license ID: 4627040581668.

Abbreviations: ACP, acyl carrier protein; ADP, adenosine diphosphate; Asp, asparate; ATP, adenosine triphosphate; CoA, coenzyme A; CTP, cytidine triphosphate; dATP, deoxyadenosine triphosphate; dCTP, deoxycytidine triphosphate; dGTP, deoxygunosine triphosphate; DHF,

dihydrofolate; dTMP, deoxythymidine monophosphate; dTTP, deoxythymidine triphosphate; dUDP,

deoxyuridine diphosphate; dUMP, deoxyuridine monophosphate; dUTP, deoxyuridine triphosphate; FAD, flavin adenine dinucleotide; FADH2, reduced falvin adenine dinucleotide; GMP, guanosine monophosphate; GTP, guanosine triphosphate; IDH1, isocitrate dehydrogenase 1; IDH2 isocitrate dehydrogenase 2; IMP inosine monophosphate; NAD, nicotinamide adenine dinucleotide; NADH,

reduced nicotinamide adenine dinucleotide; NADP, nicotinamide adenine dinucleotide phosphate; NADPH, reduced nicotinamide adenine dinucleotide phosphate; P, phosphate; PDH, pyruvate dehydrogenase; PRPP, phosphoribosyl phosphate; SDH, succinate dehydrogenase; THF,

tetrahydrofolate; UDP, uridine diphosphate; UMP, uridine monophosphate; UTP, uridine

In triphosphate;general, the intermediates a-KG, a-ketoglutarate. of glycolysis and the citric acid cycle are directed into biosynthetic pathways where they are converted into macromolecules (Lunt and Vander Heiden, 2011; Weinberg and Chandel, 2015).

1.4.2.1 Nucleoside Synthesis

Ribonucleic acid (RNA) and DNA account for a significant portion of cell mass. The nucleoside triphosphates- ATP, guanosine triphosphate (GTP), cytidine triphosphate (CTP), and uridine triphosphate (UTP) are required for RNA synthesis, while deoxyadenosine triphosphate (dATP), deoxyguanosine triphosphate (dGTP), deoxycytidine triphosphate (dCTP), and deoxythymidine triphosphate (dTTP) are required for DNA synthesis. Structurally, both nucleoside triphosphates and deoxynucleoside triphosphates are composed of a purine or pyrimidine nitrogenous base, a five-carbon sugar and three phosphate group. Nucleosides with purine nitrogen base, referred to as purine nucleosides are derived from inosine monophosphate; whereas, nucleosides with a pyrimidine nitrogen base, referred to as pyrimidines nucleosides, are derived from uridine monophosphate. Interestingly, glycolysis is the most prominent carbon source of both purines and pyrimidines.

The de novo synthesis of purine nucleosides – ATP, GTP, dATP, and dGTP requires the assimilation of 10 carbon atoms from the extracellular environment (Lunt and Vander Heiden, 2011). Half of the 10 carbons come from phosphoribosyl-a-phosphate (PRPP), which is derived from glycolysis intermidiates-glucose-6-phosphate, fructose-6-phosphate and glyceraldehyde-3- phosphate. Two of the remaining carbons are from glycine, which can come from the extracellular environment or be synthesized from glyceraldehyde-3-phosphate or tetrahydrofolate (Lunt and Vander Heiden, 2011). Tetrahydrofolate acquire their carbons from either serine or glycine both of which can be synthesized from 3-phosphoglycerate (Lunt and Vander Heiden, 2011). Alternatively, purine nucleotides can be synthesized by the salvage pathway. The salvage pathway similar to the de novo pathway uses the glycolysis derivative PRPP as a carbon source, however it requires less energy (Lunt and Vander Heiden, 2011).

Five of the 9 carbons in pyrimidines-CTP, UTP, dCTP, dTTP-come from PRPP. The amino acid aspartate is the source of 3 of the remaining carbons (Lunt and Vander Heiden, 2011). Aspartate is acquired either from the environment or the citric acid cycle intermediate oxaloacetate. dTTP has an extra carbon compared to other pyrimidines (Lunt and Vander Heiden, 2011). The folate species 5,10-methylene-tetrahydrofolate, which acquires its carbon from serine, adds a carbon to dUMP, making it dTMP. Afterwards, dTMP is phosphorylated to form dTTP.

Experimental evidence suggest that glycolysis is essential for DNA synthesis. For instance, in mitogen-stimulated mouse lymphocytes, the increase in lactate levels parallels the increase in 3H-thymidine incorporation in DNA (Wang et al., 1976). Furthermore, in stimulated human lymphocytes, glycolytic enzyme activities, lactate production, and DNA synthesis reach maximum levels in the S phase of the cell cycle (Marjanovic et al., 1988).

1.4.2.2 Lipid Synthesis

Intermediates of glycolysis and the citric acid cycle serves as a carbon source for lipid precursors. The glycolytic intermediate, dihydroxyacetone phosphate gives rise to glycerol-3- phosphate, the precursor of phospholipids and triacylglycerols (Lunt and Vander Heiden, 2011). 3-phosphoglycerate is the precursor of the sphingolipid molecules - ceramides, sphingomyelins, and cerebrosides (Lunt and Vander Heiden, 2011). Finally, acetyl-CoA is the carbon source of fatty acyl chain components of a number of lipid classes, as well as mevalonate, a precursor for cholesterol and related molecules (Lunt and Vander Heiden, 2011). Note that, mitochondrial acetyl-CoA is not used for lipid synthesis. Instead, citrate derived acetyl-CoA serves as the

34 precursor for fatty acyl chains and mevalonate (Weinberg and Chandel, 2015). To participate in lipid biosynthesis citrate must be first transported from the mitochondria into the cytosol where ATP citrate lyase converts citrate into acetyl-CoA.

1.4.2.3 Amino Acid Synthesis

A number of non-essential amino acids are direct precursors of glycolytic and citric acid cycle intermediates. 3-phosphoglycerate is the carbon source cysteine, glycine, and serine (Lunt and Vander Heiden, 2011). Whereas, pyruvate gives rise to alanine, and oxaloacetate from the citric acid cycle is the precursor of aspartate (Lunt and Vander Heiden, 2011). In addition to forming proteins, both serine and glycine are also involved the synthesis of 5,10-methylenetetrahydrofolate which as mentioned above supplies carbon to dTTP. Moreover, serine contributes to phospholipid metabolism by being the head group of phosphatidylserines and the precursor for ethanolamine and choline, which are the head groups of phosphatidylethanolamine and phosphatidylcholine respectively. Interestingly, numerous studies have shown that the balance between anabolism and catabolism of rapidly growing cells such as cancer differs from that of normal cells.

1.5 Cancer Metabolism

Warburg’s seminal discovery in 1956 showed that cancer cells convert the majority of their glucose carbon to lactate even in oxygen-rich condition, referred to as aerobic glycolysis, suggesting that the metabolism of cancer cells is district to that of normal cells (Vyas et al., 2016; Ward and Thompson, 2012). He hypothesized that altered metabolism of cancer was because of dysfunctional mitochondria. However, it was later found that damaged mitochondria are not the root of aerobic glycolysis since most tumor cells are not defective in their ability to perform oxidative phosphorylation (Ward and Thompson, 2012).

Instead, compared to nonproliferating differentiated cells rapidly proliferating cells like cancer utilizes the mitochondria differently (Ward and Thompson, 2012). Differentiated cells depend on oxidative phosphorylation to generate ATP to maintain their integrity. These cells adapt a catabolic metabolism program using glycolysis, the TCA cycle, and the electron transport chain to produce ATP. In contrast, rapidly proliferating tumor cells adapt an anabolic metabolic program, using glycolysis, and the TCA cycle, to produce biomolecules required for cell growth (Lunt and Vander Heiden, 2011; Ward and Thompson, 2012). Classical, oncogenic pathways

35 such as phodphoinositide 3-kinase (PI3K) / Akt serine/threonine kinase 1 (Akt) / mammalian target of rapamycin (mTOR), 5' AMP-activated protein kinase (AMPK), and MYC alter the metabolic program of the cell to support, macromolecular synthesis, and bioenergetic demand.

1.6 Oncogenic Pathways Drive Cancer Metabolism

The PI3K/Akt signaling cascade is a common cell proliferative pathway chronically activated in most cancers either due to an activating mutation in PI3K, oncogenic mutations in kinases that activate PI3K, or loss of function mutations in phosphatase, like phosphatase and tensin homolog, that shuts off the PI3K pathway (Fruman and Rommel, 2014). The activation of PI3K/Akt leads to enhanced glucose uptake through increased expression glucose transporters, and also the upregulation of glycolysis by the activation of hexokinase, and phophofucktokinase- 1 (Buzzai et al., 2005; Elstrom et al., 2004). The PI3K/Akt pathways also activates ATP- synthase lyase to promote the flux of carbons from glucose into the lipid synthesis pathway (Bauer et al., 2005; Berwick et al., 2002; Hatzivassiliou et al., 2005). ATP-synthase lyase converts citrate, a metabolic product of the Krebs cycle, to oxaloacetate, and acetyl-CoA, the precursor of fatty acyl chains. Genetic or chemical inhibition of ATP-synthase lyase result in diminished Akt-driven tumorigenesis, which demonstrates that the modulation of mitochondrial citrate mediated lipid synthesis is central to PI3K/Akt oncogenic activity (Bauer et al., 2005; Hatzivassiliou et al., 2005). mTOR is a major downstream effector of PI3K/Akt signaling and is best known for enhancing protein synthesis (Ward and Thompson, 2012). The signaling functioning of mTOR is carried out by the two mTOR complexes, mTOR complex 1 (mTORC1) and mTOR complex 2 (mTORC2) (Jhanwar-Uniyal et al., 2019). Interestingly, mTORC1 regulates multiple mitochondrial metabolic pathways. Glutaminolysis, the process cells use to convert glutamine into citric acid cycle intermediates is regulated by mTORC1 (Ward and Thompson, 2012; Yang et al., 2017). Specifically, mTORC1 increases the activity of an enzyme in the glutaminolysis pathway, glutamate dehydrogenase that converts glutamate to a-ketoglutarate, by inhibiting the inhibitor of glutamate dehydrogenase, sirtuin 4 (Csibi et al., 2013). mTORC1 also enhances purine synthesis by increasing tetrahydrofolate production (Ben-Sahra et al., 2016; Pedley and Benkovic, 2017). Mechanistically, mTORC1 activates the transcription factor that stimulates transcription factor 4, which then upregulates methylenetetrahydrofolate

36 dehydrogenase/cyclohydrolase, a mitochondria enzyme that produces formate, which is then used by the cell to synthesize tetrahydrofolate.

AMPK, a regulator of energy homeostasis, also plays a role in tumor initiation and propagation. It is composed of an a subunit, two regulatory subunits b and g, and four nucleotide binding clefts (Jeon, 2016). Two of the nucleotide binding clefts bind AMP, ADP, or ATP in a competitive manner (Jeon, 2016). In normal cells, AMPK is activated during low energy conditions when ATP levels are low. Once activated, AMPK promotes ATP production by stimulating catabolic pathways, and also initiates mitophagy to increase the efficiency of energy production. The role of AMPK in tumors depends on the presence of other oncogenic drivers as well as the stage of the tumor (Faubert et al., 2015). When tumor growth is driven by oncogenic drivers such as MYC, AMPK levels are low (Faubert et al., 2013). As a result, cell energy- sensing is uncoupled with proliferation, leading to cell proliferation independent of ATP levels. Whereas, in established tumors AMPK promotes mitophagy through the phosphorylation of unc- 51 like autophagy activating kinase 1, and Fission, mitochondrial 1 (FIS1) (Kim et al., 2011). Furthermore, loss of AMPK leads to the loss of TET2 function resulting in the accumulation of 5-hydoxymethylcytosine, a signature that predisposes normal cells to cancer (Wu et al., 2018).

Finally, the transcriptional factor MYC classically known for being a cell cycle re-programmer has now been shown to control genes encoding proteins involved in cell metabolism. Similar to the PI3/Akt pathway MYC promotes an anabolic metabolic profile by up regulating proteins that increase glucose uptake, glycolysis, and glutaminolytic programs (Haikala et al., 2017). Specifically, MYC upregulates glucose transporter 1, hexokinase 2, and lactate dehydrogenase A to increase the amount of pyruvate being shunted into the citric acid cycle (Kim et al., 2004; Osthus et al., 2000; Spady et al., 1976; Stine et al., 2015). However, enhanced glycolysis alone is not enough to compensate for the efflux of citric acid cycle intermediates to biosynthetic reactions in rapidly growing cancer cells (Estefania Ochoa-Ruiz, 2012). Therefore, MYC also upregulates the glutamine consumption and glutamate production to maintain citric acid cycle intermediates. Mechanistically, glutamine transporters such as ASC amino-acid transporter 2 and system N transporter 2 (SN2), are upregulated by MYC (Wise et al., 2008). Once in the cell, glutamine can be used for nucleotide synthesis in the cytoplasm or can be converted to glutamate (Lane and Fan, 2015; Wise and Thompson, 2010). MYC increases glutamate production by upregulating glutaminase (GLS), the enzyme that converts glutamine to glutamate, which can be used to replenish a ketoglutarate (Wise and Thompson, 2010).

A number of these oncogenic pathways are altered in AML in a complex pattern (Gao and Estey, 2015). Each AML patients has an average of 13 mutations, 8 of which are random passenger mutations, while 5 are recurrent driver mutations (Cancer Genome Atlas Research, 2013; Gao and Estey, 2015). In a study performed by the Eastern Cooperative Oncology Group that analyzed somatic mutations across 18 genes in 398 AML patients younger that 60 years old, an astounding 97.3% of patients had at least one identifiable mutations independent of cytogenic abnormalities (Patel et al., 2012). Consequently, drugs such as midosaturin, gilteritinib, and quizartinib have been developed to target FLT3 activating mutations, a common mutation in AML (Kavanagh et al., 2017; Short et al., 2019). However, the presence of multiple coexisting mutations forms a complex network of interacting molecular pathways, with adapting feedback and crosstalk loops that are difficult to be targeted with agents, inhibiting a single overexpressed oncogenic pathway (Gao and Estey, 2015). An alternative approach is to target metabolic pathways that the oncogenes control. The mitochondria are a hub for many of these metabolic processes (Basak and Banerjee, 2015). We and others have demonstrated that AML initiation and growth are uniquely dependent on mitochondrial functions.

1.7 Mitochondria Biology

Eukaryotic cells first acquired mitochondria by engulfing a-proteobacterium about 2 billion years ago (Friedman and Nunnari, 2014; Lane and Martin, 2010). Scientists such as Richard Altman, Carl Benda, and Otto Warburg showed that as the a-proteobacterium evolved into mitochondria that are integrated with eukaryotic cells, and that it continued to maintain a double membrane structure as well as the core ATP production processes of its ancestors (Friedman and Nunnari, 2014). Thus, the mitochondria was coined as the ‘powerhouse of the cell,’ due to its central role in energy production.

Since, its origination the genetic composition of the mitochondria changed (Friedman and Nunnari, 2014). The a-proteobacterium lost most of its genomic material or transferred it to the nuclear genome, and retained only a 16 kilobase circular genome as it evolved into the mitochondria of eukaryotic cells. In addition, the mitochondria have acquired additional functions such as, biogenesis and turnover, biosynthetic metabolism, fatty acid oxidation, cell death, and signaling that support the cellular activity (Vyas et al., 2016). Several recent studies have gained insight into how the aforementioned mitochondrial functions support the initiation and propagation of AML.

1.7.1 Biogenesis and Quality Control of Mitochondrial Proteins

Mitochondrial DNA (mtDNA) encodes thirteen essential subunits of respiratory enzyme complexes, which are: (1) NADH dehydrogenase (ND)1, ND2, ND3, ND4, ND4L, ND5, ND6 of complex I; (2) cytochrome bc1(CYTB) of complex III; (3) cytochrome c oxidase (COX) I, COXII, COXIII of complex IV; (4) ATP6, ATP8 of complex V. The mitochondria have its own independent mechanisms of replication, transcription and translation as well as protein trafficking and quality control, each of which support the oxidative phosphorylation function of AML cells and stem cells (Liyanage, 2017; Stewart and Chinnery, 2015).

1.7.1.1 Mitochondrial DNA Replication

Mitochondrial DNA (mtDNA) is replicated by polymerase g (POLG) holoenzyme, a heterodimer consisting of a 140 kDa catalytic subunit POLGA, and two 55 kDa accessory subunit POLGB (Lee and St John, 2015; Stumpf and Copeland, 2011). The catalytic subunit and accessory subunits of POLG is encoded by the nuclear genes POLG and POLG2 respectively (Lee and St John, 2015).

The replication of mtDNA begins with Transcription factor A (TFAM) unwinding mtDNA and RNA polymerase mitochondrial (POLRMT) synthesizing an RNA primer (Clayton, 2000). POLGA subunit of POLG then binds the replication primer to synthesize mtDNA. The mtDNA helicase twinkle unwinds double stranded mtDNA, while mitochondrial single-stranded binding protein (MTSSB) stabilizes the single stranded genome to support replication (Falkenberg et al., 2007; Korhonen et al., 2003). mtDNA Replication terminates with a pair of catenated circles, each circle with an old parental strand and a new daughter strand (Stumpf and Copeland, 2011). Then mtDNA topisomerase decatenates the circles, releasing the two daughter molecules (Stumpf and Copeland, 2011). In addition to DNA replication, POLG is responsible for DNA recombination and repair, all of which are required for the maintenance of the thousands of mtDNA copies in a cell. mtDNA replication is supported by nucleotide pools generated by mitochondrial and cytosolic pathways. In the mitochondrial nuclear salvage pathway, nucleosides are first transported into the mitochondria from the cytosol, then series of phosphorylation reactions involving the kinases, thymidine kinase 2 (TK2), deoxyguanosine kinase (dGUOK), cytidine monophosphate kinase 2 (CMPK2), and thymidine monophosphate kinase 2 (TMPK2) convert nucleoside

39 precursors to nucleotides (Gandhi and Samuels, 2011). Alternatively, the mitochondria can also import nucleotides made in the cytosol into the mitochondria, using the human equilibrative nucleoside transporters 1-4, and the solute carrier family 25 proteins 33 and 36 (Choi and Berdis, 2012; Di Noia et al., 2014; Liyanage, 2017).

1.7.1.2 Mitochondrial DNA Transcription

POLRMT a DNA-dependent RNA polymerase drives the transcription of mtDNA (Schwinghammer et al., 2013). The complex that initiates transcription by binding to specific regions in the light and heavy strands of mtDNA is composed of, POLRMT, mitochondrial TFAM, and mitochondrial transcription factor B (TFB2M) (D'Souza and Minczuk, 2018). Recent, evidence suggest that TFAM first binds to mtDNA and then recruits POLRMT to the promoter using its N-terminal extension. (Posse and Gustafsson, 2017; Ramachandran et al., 2017) TFB2M then modifies the structure of POLRMT to induce the opening of the promoter.

The elongation stage of transcription requires the association of POLRMT with transcription elongation factor (TEFM) (Minczuk et al., 2011). Structural studies have shown that c-terminal domain of TEFM interacts with POLRMT and forms a sliding clamp around mtDNA downstream of POLRMT binding sites (Hillen et al., 2017). The pseudonuclease core of TEFM prevents the formation of G-quadruplexes that ceases premature transcriptional termination (Agaronyan et al., 2015; Hillen et al., 2017). Finally, transcription is terminated by the mitochondria termination factor 1, and then all mitochondrial RNA (mtRNA) transcripts except ND6 undergo 3’ polyadenylation (Martin et al., 2005; Tomecki et al., 2004). The mature polycistronic transcripts produced from both the heavy and light strands are then translated by mitochondria ribosomes in the mitochondrial matrix.

1.7.1.3 Mitochondrial DNA Translation into Proteins

Mitochondrial ribosomes that translate mtRNA contain ribosomal RNAs, and mitochondrial ribosomal proteins which are composed of a 28S small subunit (mtSSU) and a 39S large subunit (mtLSU) (Smits et al., 2007). mtRNA translation is divided into 3 stages: initiation, elongation, and termination. During initiation, mtRNA binds to the mtSSU and formylated methionyl- transfer-RNA (fMET-tRNA) binds to the peptidyl site (P-site) of mtSSU (Christian and Spremulli, 2009). The initiation of mitochondrial translation is coordinated by the mitochondrial initiation factors (mtIF) – mtIF2 and mtIF3 (Christian and Spremulli, 2009; Koc and Spremulli,

2002). Studies have shown that mtlF3 makes mtSSU accessible to mtRNA by preventing the premature association of the mtLSU and mtSSU complexes. mtIF3 also correctly positions the start codon of mtRNA into the P-site of the mitoribosome (Christian and Spremulli, 2009). mtlF2 on the other hand directs the association of fMet-tRNA with mtRNA. The association of mtSSU with mtLSU together with the release of mtIF3 and mtIF2 signals the end of initiation. During elongation, mitochondrial elongation factor thermo unstable (mtEFTU) complexes with GTP and aminoacyl tRNA to direct aminoacyl tRNA to the acceptor site (A-site) of the mitochondrial ribosome (Woriax et al., 1997). Once, aminoacyl tRNA pairs with the mtRNA, at the condon- anticodon site, hydrolysis of GTP in the mtEFTU complex stimulated by the mitochondrial ribosome, results in peptide bond formation between the amino acid in the A-site and P-site and simultaneous deacylation of tRNA at the P-site. The mtEFTU-GDP complex is then released from the ribosomal complex, only to be recycled back to mtEFTU-GTP by mitochondrial elongation factor Ts (mtEFTs). Mitochondrial elongation factor G1 (mtEEFG1) causes the release of the deacylated tRNA from the P-site to the E-site, and the translocation of peptidyl- tRNAs from A-site to the P-site, as well as moving the A-site to the next mRNA codon. The termination of translation is triggered by mitochondrial release factors recognizing a stop codon at the A-site. Mitochondrial release factors facilitate the release of the synthesized polypeptide by hydrolyzing the bond between the tRNA and the polypeptide, resulting in the dissociation of the polypeptide from the tRNA molecule. After the release of the newly formed polypeptide, mitochondrial recycling factors-mitochondrial ribosome releasing factor (mtRRF) and mitochondrial elongation factor G2-GTP (mtEFG2-GTP)-catalyze the disassociation of mRNA, deacylated tRNA, and ribosomal subunits, making these components available for a new round of protein synthesis (Bhargava et al., 2004).

1.7.1.4 Quality Control by Mitochondria Proteases

The mitochondria have their own proteases that perform highly regulated proteolytic reactions to maintain mitochondrial homeostasis (Quiros et al., 2015). The 25 mitochondrial proteases that have been currently identified can be grouped into three different catalytic classes – cysteine proteases, metalloproteases, serine proteases. Mitochondrial proteases support mitochondrial homeostasis by performing a diverse array functions in the cytosol of the cell, the mitochondrial intermembrane space, inner membrane, and matrix. For instance, the Ser protease HtrA2 is located in the mitochondrial intermembrane space and acts as a chaperone regulating the folding of mitochondrial proteins (Vande Walle et al., 2008). In response to cellular insults such as

41 nuclear damage, HtrA2 is released from the mitochondria into the cytosol where it activates caspases that trigger apoptosis. The mitochondrial processing peptidases are another type of mitochondrial proteases that cleave mitochondrial-targeting sequences that signal the entry of proteins into the mitochondrion (Voos, 2013). The cleavage of mitochondrial-targeting sequences is necessary for the generation of functionally active proteins. Finally, Lon and caseinolytic mitochondrial matrix peptidase proteolytic subunit (ClpP) are two mitochondrial serine proteases that play an important role in degrading damaged proteins in the mitochondrial matrix (Baker et al., 2011). Impaired or dysregulated mitochondrial proteases are associated with aging and many pathobiological conditions such as, neurodegenerative disorders, metabolic syndromes, and cancer. Understanding the biological functions of mitochondrial proteases in ageing and disease is an important area investigation.

1.7.1.5 Mitochondrial Biogenesis and Protein Quality Control in AML

AML initiation and propagation has a unique reliance on oxidative phosphorylation (Baccelli et al., 2019; Baran et al.; Bralha et al., 2015; Cole et al., 2015; Lagadinou et al., 2013; Liyanage, 2017; Molina et al., 2018; Skrtić et al., 2011). Both small molecules targeting electron transport chain complexes – complex I inhibitors, IACS-010759 and mubritinib – and compounds targeting biological processes that support oxidative phosphorylation have anti-leukemic activity.

Our lab has shown that oxidative phosphorylation in AML is supported by mitochondrial biogenesis, which includes biological processes such as mtDNA replication, transcription and translation (Figure 11A). A subset of AML cells demonstrated increased mitochondrial biogenesis compared to normal hematopoietic cells, which is evident by the increased mitochondrial mass and DNA content in AML (Škrtić, 2013; Skrtić et al., 2011). Moreover, cytoplasmic nucleoside kinases, mitochondrial nucleotide transporters, and the mtDNA biogenesis genes required for increased mitochondrial mass and mtDNA content are also upregulated in AML (Liyanage et al., 2017). The depletion of mtDNA, and the inhibition of mitochondrial transcription and translation, using both genetic and chemical approaches, selectively targeted AML cells and stem cells by impairing oxidative phosphorylation (Bralha et al., 2015; Liyanage et al., 2017; Skrtić et al., 2011; Yehudai et al., 2019). The genetic knockdown of cytosolic nucleotide synthesis enzymes- deoxycytidine kinase, thymidine kinase, UMP-CMP kinase-reduced mtDNA in AML cells. Moreover, chemical depletion of mtDNA using nucleotide analogs ddC and alovudine reduced the expression of electron transport chain

42 subunits encoded by the mitochondria, impaired oxidative phosphorylation, and reduced the viability of AML cells and stem cells both in vitro and in vivo (Liyanage et al., 2017; Yehudai et al., 2019). Similarly, targeting POLRMT using 2-c-methyladenosine, and mitochondrial protein translation using tigecycline caused AML cell death, without affecting normal tissues, by inhibiting the expression of electron transport chain subunits encoded by mtDNA (Bralha et al., 2015; Skrtić et al., 2011).

Oxidative phosphorylation in AML cells and stem cells are also dependent on the stable activity of the mitochondrial protease ClpP. 45% of primary AML samples, with varying cytogenic and molecular profiles, overexpress ClpP (Cole et al., 2015). The complex I subunit NDUA12, as well as the complex II subunits, succinate dehydrogenase complex subunit A (SDHA) and succinate dehydrogenase complex subunit B (SDHAB), are putative substrates of ClpP in AML cells (Cole et al., 2015; Ishizawa et al., 2019). Stable activity of ClpP is vital for maintaining the integrity of oxidative phosphorylation. The inhibition of ClpP lead to the accumulation of degraded and misfolded SDHA and SDHA subunits, and a reduction in the activity of complex II (Cole et al., 2015). While ClpP hyperactivation resulted in an increased degradation of the complex I subunit NDUA12, and the inhibition in complex I activity (Cole et al., 2015; Ishizawa et al., 2019). Both ClpP inhibition and activation reduced the overall activity of the electron transport chain. Furthermore, AML cells and stem cells demonstrated an increased sensitivity to small molecules that activate and inhibit ClpP, when compared to normal cells.

1.7.2 Amino Acids and Fatty Acid Oxidation by the Mitochondria

As mentioned previously, both amino acids and fatty acids enter the citric acid cycle to support oxidative phosphorylation (Figure 11B). Amino acid oxidation begins with the amino acids transport across the cell membrane and then the mitochondria membrane by cell and mitochondrial membrane amino acids transporters respectively. Once in the mitochondria, the amino acids are subsequently deaminated by the urea cycle before entering the citric acid cycle. Fatty acids on the other hand enter the cell through the fatty acid translocase CD36. Once, in the cell CoA groups are added to the fatty acids by fatty-acyl-CoA synthase (FAS). Fatty-acyl-CoA must be then converted to acyl carnitine by carnitine palmitoyltransferase (CPT) 1, before being transferred across the mitochondrial membrane. Once in the mitochondria, carnitine acyltransferase is converted back to fatty acyl-CoA by CPT2. The b-oxidation pathway then breaks down fatty acyl-CoA into acetyl-CoA molecules, which then enters the citric acid cycle. AML

43 has a unique dependence on both amino acids and fatty acids as an energy sources to fuel oxidative phosphorylation. Moreover, amino acids and fatty acids have unique ability to support AML cell survival by activating cell signaling pathways.

1.7.2.1 Mitochondrial Amino Acids Oxidation in AML

Glutamine arginine are critical for the growth and survival of AML cells. Firstly, glutamine has demonstrated the ability to funnel into the citric acid cycle and activate cell signaling pathways in AML. Glutaminase C, an isoform of GLS that facilitates glutamine entry into the TCA cycle, is highly expressed in a subset of AML patient samples (Jacque et al., 2015). Removal of glutamine reduces oxidative phosphorylation and ATP levels in AML, which can be rescued by supplementing cells with a-ketoglutarate. Furthermore, the chemical inhibitor of glutaminase C CB-839 induces AML cell death through its inhibition of oxidative phosphorylation. Secondly, glutamine also activates the amino acid/Rag/Rac/mTORC1 pathway in AML cells (Jewell et al., 2015). The removal of glutamine or the genetic inhibition of the glutamine high-affinity transporter SLCA1A5 inhibits mROTC1 signaling and induces apoptosis in AML. Thirdly, glutaminolysis contributes to the resistance of AML cells to tyrosine kinase inhibitors (Gallipoli et al., 2018). The dependency of FLT-ITD+ AML on glutaminolysis can targeted in vivo by combining CB-839 with tyrosine kinase inhibitors (Gregory et al., 2018).

The dependency of AML cell on arginine stems from its deficiency in argininosuccinate synthase-1 (ASS1), an enzyme that converts citrulline to arginine. ASS1 deficiency in AML confers a proliferative advantage to cancer cells. Consequently, arginine depletion by the enzyme pegylated arginine deiminase (ADI-PEG20), that degrades arginine to citrulline, resulted in leukemia cell cytotoxicity (Miraki-Moud et al., 2015). Clinical trials for ADI-PEG20 is currently ongoing and results thus far has shown that ADI-PEG20 was able to induce complete remission in only some patients with ASS1 deficiency, suggesting that ASS1 deficiency is not sufficient to predict response to ADI-PEG20 monotherapy (Tsai et al., 2017).

AML stem cell functions are also are dependent on amino acids. Even compared to bulk AML, AML stem cells have a high level of amino acids and are enriched for genes involved in amino acid metabolism (Jones et al., 2018). Amino acid depletion reduced oxidative phosphorylation and AML stem cell initiation as well as propagative functions. Specifically, AML stem cells require cysteine to drive the synthesis of glutathione, which is important for the modification of SDHA, and the activity of complex II of the electron transport chain (Jones et al., 2019; Pollyea

44 et al., 2018). Inhibition of cysteine metabolism alone is sufficient to substantially impair oxidative phosphorylation in AML stem cells, but not hematopoietic stem and progenitor cells.

1.7.2.2 Mitochondrial Fatty acid Oxidation in AML

Interestingly, AML disease characteristics are influenced by the adipocyte content in the bone marrow. Adipocytes in the bone marrow influences the sensitivity of AML to chemotherapy, as obese leukemia mice had a higher rate of relapse compared to normal weight mice, when treated with chemotherapy (Behan et al., 2009). An increased adipocyte content in bone marrow was associated with enhanced AML engraftment, suggesting that adipocytes also improve AML stem cell functions in vivo (Battula et al., 2013). AML cells can enhance the lipolysis of adipocytes in the bone marrow into fatty acids, and then take up fatty acid released by adipocytes through fatty acid transporters like CD36 (Ye et al., 2016). Once in the cell fatty acids influence AML disease characteristics. For instance, AML stem cells that have the capacity of oxidize fatty acids are resistant to therapeutic strategies that inhibit amino acid uptake (Jones et al., 2018). Fatty acids can also stimulate peroxisome proliferator associated receptor g (PPARg) (Tabe et al., 2017). The activation of PPARg promotes AML cell survival by upregulating the anti-apoptotic protein BCL-2, and also reduces oxidative stress by upregulating uncoupling protein 3 (Samudio et al., 2010). Additionally, when AML cells use fatty acids as an energy source, an increased production of antioxidants are observed (Samudio et al., 2010; Tabe et al., 2017). The disease characteristics that fatty acids promote in AML can be exploited for therapeutic gain. Inhibiting fatty acid uptake by sulfosuccinimidyl oleate reduced disease burden in an AML mouse model (Ye et al., 2016), and the inhibition of fatty acid oxidation by etomoxir sensitized cells to the BCL-2 inhibitor ABT-737 (Samudio et al., 2010). Furthermore, chemical inhibition of fatty acid oxidation by avocatin B decreased levels of the antioxidant reduced glutathione (GSH), which resulted in reactive oxygen species (ROS)-dependent leukemia cell death (Lee et al., 2015). Similar to avocatin B, direct inhibitors of the AML antioxidant system target AML by increasing ROS.

1.7.3 Mitochondrial Redox and Antioxidant Systems

- ROS are partially reduced forms of molecular oxygen such as superoxide anions (O2 ) that are continually being produced in cells. Specifically, ROS originate from either incomplete reduction reactions that occur in the mitochondrial electron transport chain, or through reductions of molecular oxygen by the enzymes nitric oxide synthase, NADPH-dependent

45 oxidases, and a-ketoglutarate dehydrogenase (Handy and Loscalzo, 2012). Of note, the role of NADPH-dependent oxidases is to produce ROS, whereas the electron transport chain complexes and a-ketoglutarate dehydrogenase produce ROS as a by-product. ROS produced by each of these processes participates in cell signaling. However, if ROS levels exceed a certain threshold it damages proteins, lipids, and DNA which can lead to cell death (Di Meo et al., 2016). Consequently, the mitochondria and other components of the cell have developed antioxidant mechanisms to control levels of ROS so that cellular homeostasis can be maintained (Handy and Loscalzo, 2012).

The mitochondrial antioxidant system is composed of the enzymes manganese-dependent superoxide dismutases (MnSODs), glutathione peroxidases (GPx), and peroxiredoxins (Prx) - (Handy and Loscalzo, 2012). The O2 species generated by the mitochondria is first converted to hydrogen peroxide (H2O2) by MnSOD in the mitochondrial matrix. H2O2 has a longer half-life than superoxides and modulates cellular processes by reversibly oxidizing protein thiols.

However, when in excess H2O2 overoxidizes protein thiols, reacts with free metals, and causes lipid peroxidation, all of which results in cell stress. Enzymes in the GPx and Prx family neutralize these ROS mediated cellular processes by converting H2O2 to water (H2O) (Jiang et al., 2011).

1.7.3.1 Mitochondria Redox/Antioxidant Systems and Signaling in AML

AML stem cells depend on antioxidant systems to keep ROS levels below a detrimental threshold (Lagadinou et al., 2013; Pei et al., 2013). The antioxidant, glutathione is an important component of the mitochondrial antioxidant system. Glutathione synthesis begins with the ligation of glutamate to cysteine to make g-glutmyl-cysteine by the enzyme glutamate-cysteine ligase. After which, glutathione synthase ligates glycine to g-glutmyl-cysteine to produce the reduced form of glutathione-GSH. GPx uses reduced GSH to maintain low levels of ROS. AML stem cells have higher levels of glutathione synthesis enzymes, glutamate-cysteine ligase, and glutathione synthase, as well as a higher rate glutathione synthesis and consumption compared to normal hematopoietic stem cells. Interestingly, depleting cellular glutathione chemically (parathenolide) or genetically results in reductions in AML cell viability and engraftment suggesting that glutathione is essential for AML cell and stem cell survival (Pei et al., 2013).

Of note, ROS play an important role in the initiation and propagation functions of leukemia stem cells. In a murine model of leukemia, genetic inhibition of NADPH dependent oxidase 2 (NOX2), reduced ROS, attenuated disease development, and depleted functionally defined leukemia stem cells (Adane et al., 2019). Mechanistic studies revealed that reductions in ROS levels reduced self-renewal transcriptional programs and activated differentiation transcriptional programs. Furthermore, genetic depletion of NOX2 in primary AML patient samples reduced AML propagation in immunodeficient mice, highlight the importance of ROS signaling in AML stem and progenitor cells. The mitochondria have developed the capacity to signal to the nucleus by utilizing molecules like ROS, which is referred to as retrograde signaling.

1.7.4 Retrograde Signaling

Cell homeostasis depends on the structure and functions of its mitochondria (Vyas et al., 2016; Ward and Thompson, 2012). To maintain cellular homeostasis, the mitochondria uses a process called retrograde signaling in order to communicate with the cell. The nucleotides, metabolites, free radicals, peptides, ions and lipids generated from the mitochondria change depending on mitochondrial activity. These molecules in turn activate signaling pathways or alter the epigenome of the cell to modulate cell state. Retrograde signaling closely links mitochondrial function to cellular activity. In regard to AML, mitophagy, mitochondrial metabolites, and mitochondrial pyrimidine synthesis influences cell state.

1.7.4.1 Mitochondrial Dynamics and Mitophagy

The shape of the mitochondria in a cell are the result of the balance between fusion and fission mechanisms. Rates of both mitochondria fusion and fission respond to changes in metabolism. Mitochondria becomes more fused when they rely on oxidative phosphorylation, during starvation induced autophagy, or in the event of mTOR inhibition-induced autophagy (Rossignol et al., 2004; Youle and van der Bliek, 2012). In contrast, fission is inhibited during starvation in order to maximize oxidative phosphorylation (Rambold et al., 2011; Youle and van der Bliek, 2012). These processes work with mitophagy to maintain an optimal mitochondrial network that can meet the metabolic demands of the cell.

During fusion the inner and outer mitochondria membranes of two distinct mitochondria fuse with each other. The Outer mitochondrial membrane proteins, mitofusin 1 (MFN1) and mitofusin 2 (MFN2) together with the inner mitochondrial membrane protein, optic atrophy 1

47 mitochondrial dynamin like GTPase (OPA1) regulate mitochondrial fusion (Bordi et al., 2017; Kasahara and Scorrano, 2014). Initially, MFN1 and MFN2 bring the two outer membranes together to foster their fusion. Following outer mitochondrial membrane fusion, OPA1 drives inner mitochondrial membranes fusion (Bordi et al., 2017; Kasahara and Scorrano, 2014). OPA1 activity is regulated by the mitochondrial proteases OMA1, Yme1L, and presenilin-associated rhomboid-like (PARL). These mitochondrial proteases control mitochondrial quality by preventing the fusion of defective mitochondria. For instance, when mitochondrial membrane potential or ATP levels decrease, OMA1 is stabilized which then cleaves OPA1, generating the short OPA1 isoform. As a result, membrane fusion is inhibited since the long OPA1 isoform is required for this process.

Mitochondria fission is the process by which the two lipid bilayers of the mitochondria divide. Interestingly, when fusion is inhibited, fission is initiated. Dynamin-related protein-1 (DRP1) mediates the mitochondrial fission process, by constricting and cutting the inner mitochondrial membrane (Cereghetti et al., 2008; van der Bliek et al., 2013). DRP1 is mainly localized in the cytosol, and only initiates fission when it is translocated into the outer mitochondrial membrane so it can interact with the mitochondrial fission factors: FIS1, Mid49/mief1, Mid41/mief2 (Loson et al., 2013). In parallel, the endoplasmic reticulum (ER) is recruited to the mitochondria, and IMM constriction occurs at ER at mitochondrial contact sites. When the GTP in DRP1 is hydrolyzed, mitochondrial membrane constriction is further enhanced. Finally, dynamin 2 is recruited to the region of constriction where is terminates membrane scission, leading to two daughter mitochondria (Tilokani et al., 2018).

Autophagy is a mechanism established to compensate for nutrient depletion and to protect cells from deleterious protein aggregation (Youle and van der Bliek, 2012). Damaged mitochondria are also selectively removed by mitophagy - autophagy specific to mitochondria, to maintain a healthy mitochondrial network. The most characterized mitophagic pathway is the PTEN- induced kinase 1 (PINK1)-Parkin (PRKN)-dependent pathway. In basal healthy conditions, mitochondrial processing protease (MPP) and PARL constitutively cleaves and degrades PINK1. When mitochondria are damaged to the point that the mitochondrial membrane is depolarized, MPP and PARL are inhibited, resulting in the stabilization of PINK1 in the outer mitochondrial membrane. PINK1 autophosphorylation ensues, leading to the recruitment of the E3 ligase PRKN. Once in the outer mitochondrial membrane, PRKN ubiquitinates a variety of mitochondria membrane proteins - MFN1 and MFN2, TOM20, mitochondrial Rho GTPase

1(MIRO1) - to promote their degradation (Chan et al., 2011). PARKIN also ubiquitinates voltage dependent anion channel 1(VDAC1), which acts as a signal for the recruitment of microtubule- associated proteins (1A/1B) light chain 3B (LC3) expressing autophagosomes to the mitochondria; and ruptures the outer mitochondrial membrane facilitating the interaction of prohibitin-2 with LC3, committing the mitochondria to mitophagy (Geisler et al., 2010; Wei et al., 2017).

Mitophagy, is intimately linked to fusion and fission processes (Twig et al., 2008). During most mitochondrial fission events, one daughter mitochondrial is hyperpolarized while the sister mitochondria is hypopolarized, under these conditions if a single daughter mitochondrion functions sub-optimally, it becomes depolarized and then eliminated by mitophagy (Youle and van der Bliek, 2012). Dominant negative DRP1 prevents mitophagy, suggesting that fission is required for mitophagy. Fusion on the other hand can compensate for damaged/missing components of the mitochondria and thereby prevent mitophagy. Another link between mitophagy and mitochondrial dynamics is illustrated by the observation that faulty mitochondria in PINK1 and PRNKN deficient flies can be corrected by either inhibiting of mitochondrial fusion or by promoting mitochondrial fission (Burman et al., 2012; Clark et al., 2006; Pimenta de Castro et al., 2012). Importantly, AML stem cells are enriched for mitochondrial dynamics and mitophagy pathways compared to more differentiated bulk AML cells (Pei et al., 2018).

1.7.4.2 Mitochondrial Dynamics and Mitophagy Signaling in AML

AML stem cells have more compact mitochondria as well as an increased expression of mitochondrial dynamics regulators FIS1, PINK1, and MiD49. Out of all these genes the mitochondrial fission gene, FIS1 showed the most consistent upregulation in AML stem cells, suggesting that mitochondria AML stem cells undergo more mitochondrial fission compared to bulk leukemia cells. The genetic inhibition of FIS1 in AML cells increased mitochondrial content and impaired mitophagy in response to mitochondrial stress. Mouse engraftment studies revealed that the stemness of leukemia was affected more than the stemness of normal hematopoietic cells when mitochondrial FIS1 was inhibited. Mechanistically, levels of FIS1 are regulated by the energy homeostasis sensor, AMPK, since the activation of AMPK mimics the effects seen in the loss of FIS1. The impairment in mitophagy leads to the inhibition of glycogen synthase kinase (GSK) which induces stem cells to differentiate, suggesting that AML stem cells sense changes in mitophagy by GSK signaling (Pei et al., 2018; Schimmer, 2018). These

49 findings highlight the essential role of mitophagy in the maintenance of AML stem cell functions. Similar mitochondrial structure mitochondrial metabolic pathways have the capacity to regulate AML cell state.

1.7.4.3 Mitochondrial Metabolites

The mitochondria house a number of different metabolites, such as metabolites of the citric cycle which include succinate, fumarate and a-ketogluterate. Depending on the energy balance of the cell the citric acid cycle metabolites functions to either oxidize acetyl-COA to CO2 or aid in biosynthetic processes (Owen et al., 2002). Citric acid cycle metabolites are tightly controlled in the cell. Specifically, the intermediates of the cycle are replenished by anaplerosis and removed by cataplerosis. Reductions or increases in metabolites disrupts cell homeostasis. Failure to replenish citric acid cycle metabolites by anaplerosis results in the inability of the cell to carry out both its oxidative and biosynthetic functions. The buildup of citric acid cycle metabolites, on the other hand, have been associated with the aberrant activation of cell proliferation pathways (Yang et al., 2013).

1.7.4.4 Metabolite Signaling in AML

In normal cells, isocitrate dehydrogenase (IDH) 1 or 2 is an anaplerosis enzyme that converts isocitrate to a-ketoglutarate in the citric acid cycle (Yang et al., 2013). In some AML patients a mutation IDH1/2, increases the levels of the oncometabolite D–2-hydroxyglutarate (2HG). The accumulation of 2HG, alters the genome, and epigenome, to shift the AML cells from a quiescent differentiated state to an actively proliferating malignant state (Yang et al., 2013). Inhibitors of IDH1/2, promote the differentiation of AML cells that carry the mutation in these enzymes (Wang et al., 2013). Clinically, relapse and recovery AML patients with IDH1/2 mutations, show a 40% response, and 20% complete remission when treated with IDH1/2 inhibitors (Stein et al., 2017).

1.7.4.5 Dihyroorotate Dehydrogenase and Pyrimidine Synthesis

Dihydroorotate dehydrogenase (DHODH) is an enzyme located in the inner mitochondrial membrane that catalyzes the conversion of dihydroorotate to orotate, which is necessary for the de novo synthesis of uridine monophosphate (UMP) (Sykes, 2018). Both DNA and RNA synthesis require UMP. Beyond the synthesis of DNA and RNA UMP also feeds into the production of phosphatidylcholine, phosphatidylserine, phosphatidylinositol, glycogen,

50 hyaluronic acid, proteoglycans, glucuronidation, and the post-translational modification of proteins by O-liked N-acetylglucosamine, making UMP a key biosynthetic molecule.

1.7.4.6 DHODH and AML Cell State

DHODH inhibitor brequinar was a hit in a high-throughput screen phenotypic screen that aimed to identify small molecules that promote differentiation in AML (Sykes et al., 2016). Out of the 330,000 compounds tested 11 out of the 12 hits were DHODH inhibitors which illustrates the importance of this pathway in AML stemness. DHODH inhibition induced cell death and differentiation of AML cells, as measured by gene expression, cell morphology, cell surface marker expression, and in vivo engraftment studies. Furthermore, mice treated with DHODH did not show organ toxicity. Based on these pre-clinical studies phase I clinical trials for the DHODH inhibitors, BAY2402234 and ASLAN003, in patients with relapsed refractory AML are currently underway (Sykes, 2018). However, the reason for the therapeutic window observed is still not known (Sykes, 2018). The DHODH gene is not mutated or translocated, and the DHODH proteins is not overexpressed in AML cells. It is plausible that AML cells have a lower tolerance for pyrimidine depletion compared to non-malignant counter parts. It is still not clear if this decrease in pyrimidine tolerance is due to a lower pool of pyrimidines at baseline or a higher rate of flux through the de novo pyriminde synthesis pathway. More studies investigating the therapeutic window of DHODH inhibition in AML will give us insight into how mitochondrial biosynthetic processes regulate AML stemness. We also need to understand more about AML mitochondrial biology to determine the basis the therapeutic window of mitochondrial targets. Therefore, to understand more about AML mitochondrial biology, we conducted a genome-wide CRISPR screen and searched for new mitochondrial pathways that inhibited the growth and viability of AML cells.

Figure 11. The mitochondrial dependencies of AML. Mitochondrial pathways in AML that control - mitochondrial biogenesis and protein quality control (A), fatty acid/amino acid oxidation (B), redox/antioxidant balance (C), and retrograde signaling (D) - can be exploited as indicated for therapeutic gain. Abbreviations: BCL-2, b-cell lymphoma 2; UCP3, uncoupling protein 3; FtACS fatty-acyl-CoA synthase change; CPT, carnitine palmitoyltransferase; CAT, carnitine acyltransferase; UCP3, uncoupling protein 3; MnSOD, manganese-dependent superoxide dismutases; GPx, glutathione peroxidases; Prx, peroxiredoxins; DHODH, dihydroorotate dehydrogenase; IDH, isocitrate dehydrogenase.

1.8 CRISPR Screens

We performed a genome-wide CRISPR screen to identify mitochondrial pathways that affect AML growth. The aim of any genetic screen is to recognize a causal relationship between genotype and phenotype. In the past, genetic screens were conducted with the aid of RNA interference or complementary DNA (cDNA) libraries (Boutros and Ahringer, 2008). Both loss of function RNA interference and gain of function cDNA screens have a number of limitations. In the former, RNA interference only partially suppresses gene expression and has pervasive off- target effects. In the latter, it is impossible for the cDNA library used to cover the full complexity of the cell’s transcriptome, and cDNAs are often expressed at non-physiological aberrantly high levels, which reduces the certainty of biological relationships identified. The new clustered regularly interspaced palindromic repeats – CRISPR associated nuclease 9 (CRISPR-CAS9) technology has yielded a versatile platform for genetic screen studies (Hsu et al., 2014).

The CRISPR-CAS9 system is a programmed nuclease derived from the microbial immune system. In CRISPR, a single guide RNA (sgRNA) directs CAS9 to a particular gene. When the gene target contains a short sequence termed the protospacer-adjacent (PAM) motif on the complementary DNA strand, sgRNA binds to the target strand by complementarity and then guides CAS9 to generate a site specific double stranded break on the target sequence. Subsequently, the double strand break is repaired by either non-homologous end joining (NHEJ) or homologous directed repair (HDR). NHEJ is more active than HDR because HDR requires a homologous template and is mainly restricted to the S and G2 phase of the cell cycle. Since HDR uses a template, NHEJ is more error prone than HDR, usually inducing frameshift indel mutation that may abrogate target gene function (Heyer et al., 2010).

CRISPR can used in two screen formats: arrayed screen and pooled screen (Xue et al., 2016). Arrayed screens enable a wide range of cellular phenotypes to be screened, since known genetic perturbation are separated and arrayed into individual wells. In the pooled screen format, the sgRNA library is massively synthesized, cloned, and delivered into cells to introduce genetic perturbations in a number of genes. After cells a specific phenotype has been separated, their genetic perturbations are read out to identify causal links between genotype and phenotype.

Currently, the pooled format is the most common CRISPR screen (Chen et al., 2015; Shalem et al., 2014). Procedurally, the sgRNA library is massively synthesized and then delivered to the initial cell population. Afterwards, cells that have taken up the plasmid must be selected. Then,

53 there are two ways of screening for an intended phenotype: enrichment screen or depletion screen. An enrichment screen aims to identify mutations that confer resistance to a condition such as, a drug treatment, pathogen infection or hypoxia (Chen et al., 2015; Shalem et al., 2014). Generally, there are relatively few protective gene perturbations and cells with protective gene perturbations continue to proliferate making them easy to identify. In contrast, depletion selection aims to identify genetic perturbations that cause cells to be depleted from the cell population over time (Shalem et al., 2014). Depletion screens require a more sensitive readout because the magnitude of sgRNA enrichment is more modest compared to positive selection, and the number of depleted genetic perturbations in negative selection is usually bigger than in positive selection.

In both positive and negative selection, genomic DNA is extracted from the cell, and the sgRNA are subjected to PCR amplification, and next generation sequencing (NGS) (Xue et al., 2016). Subsequently, these regions are mapped to a pre-existing sgRNA library to readout sgRNA representation in the selected cell population. By comparing the sgRNA profile at the end of the screen to profile at the beginning of the screen, a causal link between the genetic perturbation and phenotype can be identified.

Figure 12. Pooled clustered regularly interspaced short palindromic repeats (CRISPR) screens. The 6 main steps of a pooled CRISPR screen. Reproduced with permission of the BMJ publishing group ltd., from Hsu et al., 2014, license ID: 4627040581668. Abbreviations: PCR,

polymerase chain reaction; sgRNA, single guide RNA; NGS, next generation sequencing.

In our CRISPR depletion screen we tested a total of 17,232 nuclear-encoded proteins using 91,320 single guide RNA (sgRNA), out of which tafazzin (TAZ) was a top hit.

1.9 Tafazzin and Cardiolipin

Out of the 1,049 genes that encode mitochondrial proteins tested, TAZ was in the top 1% of depleted genes. TAZ is a transacylase, non-integrally associated with intermembrane space facing leaflets of both the inner and outer mitochondrial membranes; and is responsible for converting nascent cardiolipin into its mature form (Lu et al., 2016). Cardiolipin is the only mitochondrion-specific a phospholipid in the cell that is predominantly localized to the inner mitochondrial membrane, where it is up to 20% of the phospholipid content (Gonzalvez et al., 2008; Paradies et al., 2014). Structurally, cardiolipin is classified as a non-bilayer lipid, since its glycerol head group has a smaller diameter than its four acyl chains. The unique structure of non- bilayer forming phospholipids induces a negative membrane curvature and supports membrane fusion events, membrane bending, and imparts order to surrounding lipids (Lu and Claypool, 2015; Paradies et al., 2014).

1.9.1 Cardiolipin Functions

Cardiolipin interacts with electron transport chain complexes and is required for the proper localization and efficient function of respiratory chain enzymes (Paradies et al., 2014). In metabolically active tissues such as the liver and heart, the reduction of cardiolipin levels or the alteration of its structure have been shown to reduce the rate of oxidative phosphorylation (Claypool and Koehler, 2012; Lu and Claypool, 2015; Mejia et al., 2015; Mejia et al., 2014; Sparagna and Lesnefsky, 2009). Cardiolipin also provides a platform for caspase 8 activation, and regulates cytochrome c release, both of which lead to apoptosis (Gonzalvez et al., 2008). Furthermore, cardiolipin has been implicated to play a role in: 1) mitochondrial fusion/fission (Ban et al., 2010; Bustillo-Zabalbeitia et al., 2014; Montessuit et al., 2010); 2) the assembly of mitochondrial protein transporters (Gebert et al., 2009; Jiang et al., 2000; van der Laan et al., 2007); 3) establishing mitochondrial cristae (Weber et al., 2013); 4) initiating mitophagy (Chu et al., 2013). As such, while relatively low in abundance in the membrane, cardiolipin is fundamental in a number of cellular processes

1.9.2 Cardiolipin Synthesis and Remodeling

Cardiolipin is synthesized by a cascade of enzymes located in the cytoplasm, and the inner mitochondrial membrane (Lu and Claypool, 2015). It begins with phosphatidic acid, a common substrate in triacylglycerol, and glycerolipid metabolism. Glycerophosphate acyltransferases synthesize phosphatidic acid from glycerol 3-phosphate (G3P) and acyl CoA. Once synthesized, phosphatidic acid moves from the cytoplasm into the matrix of the mitochondria. Here, it is converted to cytidine diphosphate diacylglycerol (CDP-DAG) by CDP-DAG synthase. Afterwards, phosphatidylglycerol phosphate synthase (PGS1) catalyzes the committed step of this pathway, which is the conversion of CDP-DAG and G3P to produce phosphatidylglycerol phosphate. Subsequently, phosphatase localized to mitochondrion rapidly dephosphorylates phosphatidylglycerol phosphate to phosphatidylglycerol (PG). Finally, cardiolipin synthase condenses PG with CDP-DAG to generate unremodeled nascent cardiolipin.

Once synthesized, nascent cardiolipin is then remodeled into a functional cardiolipin molecule (Figure 1B). To initiate this process, phospholipases remove an acyl chain from cardiolipin generating monolysocardiolipin (MLCL). Afterwards, MLCL is reacylated by one of: (1) TAZ; (2) monolysocardiolipin acyltransferase (MLCAT1); or (3) acyl-CoA lysocardiolipin acyltransferase-1 (ALCAT1) to produce fatty acids that vary in their fatty acyl side chains. TAZ uses phosphatidylcholine (PC) or phosphatidylethanolamine (PE), as the acyl chain donor whereas ALCAT1 and MLCAT1 use acyl-CoA. Absence of TAZ causes alterations in cardiolipin molecular species in every model tested to date (Acehan et al., 2009; Acehan et al., 2011; Baile et al., 2014; Bissler et al., 2002; Dudek et al., 2013; Gonzalvez et al., 2013; Gu et al., 2004; Houtkooper et al., 2009; Schlame et al., 2003; Valianpour et al., 2005; Valianpour et al., 2002; van Werkhoven et al., 2006; Vaz et al., 2003; Vreken et al., 2000; Wang et al., 2014; Xu et al., 2009). In contrast, the loss of MLCAT1, or ALCAT1 does not consistently change the steady state acyl composition of cardiolipin (Li et al., 2010; Richter-Dennerlein et al., 2014; Schlame et al., 1993). Thus, among the three remodeling enzymes TAZ is responsible for the majority of cardiolipin remodeling under physiological conditions. Whereas, MLCAT1, and ALCAT1 acylate MLCL only in the absence of TAZ. Currently, the role that TAZ plays in hematopoiesis and leukemogenesis is not known.

Cardiolipin Synthesis

Cytoplasm Mitochondria Glycerol-3-Phosphate CDP-DAG acyl transferase synthase G3P PGS1 PTPMT1 PG CLS G3P Phosphatidic + PGP + Nascent Acid CDP-DAG CDP-DAG Cardiolipin Acyl-CoA

Cardiolipin Remodeling

(PC or PE) Phospholipase’s TAZ Nascent MLCL Mature Cardiolipin MLCAT1/ALCAT1 Cardiolipin (Acyl-COA)

Figure 13. Cardiolipin synthesis and remodeling pathways. Cardiolipin is synthesized and remodeled by a series of enzymes located in the cytoplasm and the mitochondria. Abbreviations: ALCAT1, acyl-CoA lysocardiolipin acyltransferase-1; CDP-DAG, cytidine diphosphate diacylglycerol; CLS, cardiolipin synthase; G3P, glycerol 3-phosphate; MLCAT1, monolysocardiolipin acyltransferase; MLCL, monolysocardiolipin; PG, phosphatidylglycerol; PGP, phosphatidylglycerol phosphate; PGS1, phosphatidylglycerol phosphate synthase; PTPMT1, phosphatase localized to mitochondrion 1; TAZ, tafazzin.

1.10 Objective and Aims

The central objective of this project is to understand the role of TAZ in hematopoiesis and leukemogenesis. In addition, we wanted to determine if TAZ or proteins that are part of the TAZ signaling axis can be targeted therapeutically.

The specific aims of this study were:

1. Understand the role of TAZ in hematopoietic stem cell function. 2. Characterize the role of TAZ in AML initiation and propagation and elucidate the molecular mechanism by which TAZ regulates AML propagation and self-renewal. 3. Determine if TAZ or proteins that are part of the TAZ signaling axis can be targeted therapeutically.

Chapter 2

Methods 2.1 Experimental Model and Subject Details

2.1.1 Human Cell Lines

OCI-AML2, and K562 cells were maintained in IMDM (Iscove’s modified Dulbecco’s medium), supplemented with 10% FCS (fetal calf serum), 100 units/mL penicillin and 100 µg/mL of streptomycin. TEX cells obtained from John Dick’s lab (Warner et al., 2005) were cultured in IMDM, with 20% FCS, 2 mM L-glutamine, 100 units/mL penicillin and 100 µg/mL of streptomycin, 20 ng/mL stem cell factor (SCF), and 2 ng/mL interleukin (IL)-3. U-937 cell were maintained in RPMI 1640 medium, with 10% FCS, 100 units/mL penicillin and 100 µg/mL of streptomycin. K562 and U-937 cells were authenticated via STR (short-tandem repeat) profiling. 8227 cells were also provided by John Dick’s lab (Princess Margaret Cancer Centre, University Health Network, Toronto, Canada) (Lechman et al., 2016) and were cultured in X-VIVOÔ-10, with 20% bovine serum albumin-insulin-transferrin (BIT), Fms-related tyrosine kinase 3 ligand (Flt3-L, 50 ng/ml), IL-6 (10 ng/ml), SCF (50 ng/ml), thrombopoietin (TPO, 25 ng/ml), IL-3 (10 ng/ml), granulocyte colony-stimulating factor (G-CSF, 10 ng/ml). Lentiviral packing cells (293T) were cultured in Dulbecco's Modified Eagle Medium (DMEM) with 10% FCS for seeding, and DMEM with 10% FCS, 100 units/mL penicillin and 100 µg/mL of streptomycin, as well as 1% bovine serum albumin (BSA) for harvesting of virus. All cell lines were maintained at 37°C, supplemented with 5% CO2. The sex and age of the patients from whom the cell lines were generated are indicated in table 6.

58 59

Cell line Species Tissue Sex Age

OCI-AML2 Homo sapiens, human Peripheral Blood Male 65

TEX Homo sapiens, human Cord Blood n/a n/a

U-937 Homo sapiens, human Pleural Effusion Male 37

K562 Homo sapiens, human Bone Marrow Female 53

8227 Homo sapiens, humans n/a n/a n/a

293T Homo sapiens, humans Fetal Kidney n/a n/a

Table 6. Cell Lines.

2.1.2 Animals

Six-twelve week Male or female immunodeficient NOD.Cg-Prkdcscid Il2rgtm1Wjl Tg (CMV- IL3,CSF2,KITLG)1Eav/MloySzJ (NOD/SCID-GF) mice used to transplant TEX and 8227 cells were obtained from Dr. Connie J. Eaves and bred in our facility (Nicolini et al., 2004). Eight- twelve week female immunodeficient NOD.CB17-Prkdcscid/J (NOD/SCID) and 6-8 week immunodeficient male Prkdcscid (SCID) mice, used for the transplantation of primary AML and OCI-AML2 cells respectively were obtained from the Ontario Cancer Institute. Mice were randomly assigned to each experimental group.

Immunocompetent B6.Cg-Gt (ROSA) 26Sortm37(H1/tetO-RNAi:Taz)Arte/ZkhuJ doxycycline-Inducible- Tafazzin-Knockdown (iDOX-Taz-KD) transgenic mice were acquired from Dr. Zaza Khuchua, at the University of Cincinnati. In collaboration with TaconicArtemis GmbH Dr. Khuchua generated the iDOX-Taz-KD mice, in a C57BL/6J genetic background as previously described (Acehan et al., 2011). iDOX-Taz-KD transgenic mice transferred to our facility, were backcrossed with female immunocompetent C57BL/6J mice for 7-10 generations. To maintain iDOX-Taz-KD transgenic mice on a C57BL/6J genetic background, iDOX-Taz-KD male mice were bred with either wildtype littermates or C57BL/6J female mice. iDOX-Taz-KD mice were weaned onto a standard chow diet and genotyped for the iDOX-Taz-shRNA transgene by polymerase chain reaction (PCR) of the tail genomic DNA (Table 7). Note that, iDOX-Taz-KD

60 transgenic mice were born viable at the expected Mendelian ratio and displayed normal fertility. Litter mates were randomly assigned to experimental groups. Twenty-one to twenty-nine week C57BL/6J mice were used to assess the effect MMV007285 has on hematopoietic cells subjected to stress.

During all experiments, the weights of the mice were approximately 18-30g with no animals loosing greater than 10% body weight. All animals were housed in microisolator cages with temperature-controlled conditions under a 12-hour light/dark cycle with free access to drinking water, and food. Only one experimental procedure was performed in each mouse, and all mice used were drug naïve prior to the experiment. Furthermore, all animal studies were performed in accordance with the Ontario Cancer Institute Animal Use Protocol (AUP): # 1251.33 (NOD/SCID-GF, NOD/SCID, and SCID) and AUP # 2244.12 (iDOX-Taz-KD transgenic mice).

Gene Strand Sequence

TAZ (Human AML) Forward 5’-TTGCTGCCTTCTGGATTCTT-3’

Reverse 5’-CCCTGCCTAAGCTTCTTCCT-3’

PISD (Human AML) Forward 5’-CAACCTCAGCGAGTTCTTCC-3’

Reverse 5’-CGACTCCAGGGAGTAGGTGA-3’

18srRNA (housekeeping, Forward 5’-AGGAATTGACGGAAGGGCAC-3’ human)

Reverse 5’-GGACATCTAAGGGCATCACA-3’

LYZ Forward 5’-GCCAAATGGGAGAGTGGTTA-3’

Reverse 5’-ATCACGGACAACCCTCTTTG-3’

iDOX-Taz-shRNA Forward 5’- CCATGGAATTCGAACGCTGACGTC-3’ Transgene

Reverse 5’- TATGGGCTATGAACTAATGACCC-3’

ND1 Forward 5’-AACATACCCATGGCCAACCT-3’

Reverse 5’-AGCGAAGGGTTGTAGTAGCCC-3’

HGB Forward 5’-GAAGAGCCAAGGACAGGTAC-3’

Reverse 5’-CAACTTCATCCACGTTCACC-3’

TLR4 Forward 5’-TGAGCAGTCGTGCTGGTATC-3’

Reverse 5’-CAGGGCTTTTCTGAGTCGTC-3’

TLR8 Forward 5’-CAGAGCATCAACCAAAGCAA-3’

Reverse 5’-CTGTAACACTGGCTCCAGCA-3’

IL6 Forward 5’-GGAGACTTGCCTGGTGAAAA-3’

Reverse 5’-GTCAGGGGTGGTTATTGCAT-3’

IFNβ Forward 5’-CAACTTGCTTGGATTCCTACAAAG-3’

Reverse 5’-TATTCAAGCCTCCCATTCAATTG-3’

Table 7. PCR Primers. Abbreviations: TAZ, tafazzin; PISD, phosphatidylserine decarboxylase; rRNA, ribosomal RNA, LYZ, lysozyme; ND1, NADH: ubiquinone oxidoreductase core subunit 1; HGB, human globulin gene; TLR4, toll-like receptor 4; TLR8, toll-like receptor 8; IL6, interleukin 6; IFNb, interferon b.

2.1.3 TAZ-KD Induction in iDOX-TAZ-KD Mice

TAZ-knockdown was induced by feeding 7.6-14.3 week old iDOX-Taz-KD transgenic mice with DOX-containing chow (625 mg/kg chow), formulated by Purina Mills, for 12.9-19.4 weeks. Non-transgenic littermates were fed DOX-containing chow and used as wildtype (WT) controls.

2.1.4 Primary AML and Normal Hematopoietic Cells

Primary human AML samples were obtained from peripheral blood or the bone marrow of consenting male or female AML patients, with a malignant cell frequency of 80% among

62 mononuclear cells. Differential density centrifugation was used to isolate AML cells. Peripheral blood stem cells (PBSCs) were obtained from healthy consenting male or female volunteers, donating PBSCs for allogenic stem cell transplantation. PBSCs were isolated by G-CSF stimulation, and leukopheresis. Both primary AML cells, and PBSCs were frozen in alpha MEM + 5% FBS or 90% FBS +15U/mL of heparin + 10% DMSO. The University Health Network institutional review boards approved the collection and use of human tissue for this study (Research Ethics Board protocol #15-9324). As per regulation, all specimens were de-identified. Each experiment was performed using a single aliquot from a donor. Information about the patients who were the source of the cells are indicated in Table 8.

Patient Disease Age at Sex Cytogenetics Molecula Status of Sample ID Collection r

110839 AML, 86 M 58-59, CYY,- Not Done Diagnostic undifferentiated 2,-3,-4,-5,-6,- 7,-9,-12,- 13,add(15)(q2 4)x2,-16,-16,- 17,-17,-19,- 20,+22,+r,+2- 7mar[cp14]

0676 AML, 79 M 46~48, XY,+6, Not Done Diagnostic undifferentiated del(13) (q12q22),+del( 13) (q12q22)[cp9]/ 46,XY[11]

120021 AML, with 88 F 46,XX,1~5dmi Not Done Diagnostic maturation n[13]/46,XX[7

160556 AML with 57 46,XY,t(3;5) Not Done Diagnostic myelodysplasia- (q21;q35)[10] related features

162111 AML with 18 M 45,X,- NPM1+, Diagnostic mutated NPM1 Y[9]/46,XY[1 FLT3- 1] ITD-

100565 AML 65 M 46,XY,t(6;11) Not Done Diagnostic (q27;q23)[19]/ 46,XY[1].nuc ish(MLLx2)(5' MLL sep 3'MLLx1)[148 /200]

120541 AML with 51 F 46,XX[20] NPM1+, Diagnostic mutated NPM1 FLt3- ITD+

120287 AML with 77 M 46,XY[20] NPM1+, Diagnostic mutated NPM1 FLt3- ITD+

120860 AML with 31 F 46,XX,t(9;11)( Not Done Diagnostic t(9;11) p22;q23)[10] (p22;q23); MLLT-MLL

Table 8. Clinical Characteristics of Primary AML Patient Samples.

2.1.5 Primary AML Cell Cultures for Transduction

Cells were thawed in a 37°C water bath, washed once in media composed of IMDM 10% FBS, 100 units/mL penicillin, and 100 µg/mL of streptomycin. Primary AML cells were resuspended in X-VIVOÔ 10 supplemented with 20% BIT, 50 ng/ml Flt3-L, 10 ng/mL IL-6, 50 ng/mL SCF, 25 ng/mL TPO, 10 ng/mL IL-3, 10 ng/mL G-CSF at a concentration of 5.0 x 105 cells/mL before being transduced.

2.2 Method Details

2.2.1 Plasmids

The 90K gRNA library developed in Hart et al., 2015 was obtained from the Moffat lab. Whereas, the TAZ, phosphatidylserine decarboxylase (PISD), and control guide RNA (sgRNA) oligonucleotides were synthesized and cloned into the pLCKO lentiviral vector using the BspAI/BfuAI sites as described previously (Hart et al., 2015). The coding sequence of sgRNA’s targeting TAZ (gene id: 6901), PISD (gene id: 23761), and the LacZ genes (Control) are listed in table 9.

For the transduction of AML cell lines, the TAZ shRNA constructs in the hairpin-pLKO.1 vector were purchased from Sigma-AldrichÒ as glycerol bacterial stocks. The coding sequence of shRNAs targeting TAZ (accession no. NM_000116), and the control shRNA targeting the green fluorescent protein (GFP) sequence (GFP, accession no. clonetechGfp_587s1c1) are listed in Table S6.

To transduce 8227 cell and primary AML patient samples, TAZ shRNA sequences were first modified in order to be cloned into the hairpin-pRS19 vector using the restriction enzyme BsbI. The coding sequence of shRNAs targeting tafazzin (accession no. NM_000116), and the control shRNA non-targeting sequence are listed in Table 9.

The PISD gene (Accession no. NM_014338) from the Myc-DDK-tagged PISD ORF (OriGene #RC200269) was cloned into the pLenti-EF1a-C-Myc-DDK-IRES-Bsd plasmid (OriGene, #PS100085), using restriction enzymes AsisI and MluI.

All sgRNA, shRNA, and overexpression plasmids were validated by Sanger sequencing before its use in downstream experiments.

Gene shRNA Referenc Sequence name e

Control N/A 5’-CCCGAATCTCTATCGTGCGG-3’ sgRNA

TAZ Gene ID: 5’-TACGAGCTCATCGAGAAGCG-3’ sgRNA1 6901

TAZ Gene ID: 5’-GCTCATCGAGAAGCGAGGCC-3’ sgRNA2 6901

PISD sgRNA Gene ID: 5’-AGCTGCCACACTGGCTGCGC-3’ 23761

Control clonetech 5’-CCGG TGC CCG ACA ACC ACT ACC TGA CTCGAG TCA shRNA Gfp_587s GGT AGT GGT TGT CGG GCA TTTTT-3’ (pLKO.1) 1c1

TAZ NM_000 5’-CCGG TCC TAA CAG TCC GCC CTA CTT CTCGAG AAG shRNA1 116 TAG GGC GGA CTG TTA GGA TTTTTG-3’ (pLKO.1)

TAZ NM_000 5’-CCGG TGC TTC CTC AGT TAC ACA AAG CTCGAG CTT shRNA2 116 TGT GTA ACT GAG GAA GCA TTTTTG-3’ (pLKO.1)

Control (Chan et 5’- ACCG GCA CTA CCA GAG CTA ACT CAG ATA GTA CT shRNA al., 2015) TCAAGAG AGTA CTA TCT GAG TTA GCT CTG GTA GTGC (pRS19) TTTT-3’

TAZ NM_000 5’-ACCG TGC TTC CTC AGT TAC ACA AAG TCAAGAG shRNA2 116 CTT TGT GTA ACT GAG GAA GCA TTTT -3’ (pRS19)

Table 9. sgRNA or shRNA Sequences.

2.2.2 Lentiviral Packing

For library virus production lentivirus was made in a 175 cm2 flask format, by transfecting 11 x 106 packaging cells (293T) with a three-plasmid system (lentiviral pLCKO vector containing the library, packing plasmid with: gag, pol, and rev genes, and envelope plasmid) (Hart et al., 2015; Moffat et al., 2006). sgRNA and shRNA plasmids used in AML cell lines were isolated using the E.N.Z.A.® Plasmid Midi Kit system (Omega bio-tek, GA, USA) from glycerol bacteria stocks, and then quantified by the NanoDrop™ (ThermoScientific, MA, USA) spectrophotometer. Lentivirus was made in a 25 cm2 flask format, by 293T cells with a three-plasmid system (hairpin-pLKO.1 vector/ guide- pLCKO vector, packing plasmid with: gag, pol, and rev genes, and envelope plasmid) (Cole et al., 2015; Simpson et al., 2012).

Whereas, cloned shRNA plasmids used in primary AML patient samples were isolated using the E.N.Z.A® Plasmid Midi Kit system (Omega bio-tek, GA, USA), and then quantified by the NanoDrop™ (ThermoScientific, MA, USA) spectrophotometer. Lentivirus was made in a 175 cm2 flask format, by transfecting 293T cells with a three-plasmid system (hairpin-pRS19 vector, packing plasmid with: gag, pol, and rev genes, and envelope plasmid) (Cole et al., 2015; Simpson et al., 2012). The virus was concentrated using the Lenti-X™ Concentrator as per manufacturer’s instructions.

2.2.3 CAS9-OCI-AML2 Cell Line Generation

CAS9-OCI-AML2 cells were generated by adapting a method previously described (Hart et al., 2015). We first infected OCI-AML2 cells with virus containing a CAS9-2A-Basticidin expressing cassette (Addgene plasmid#73310) and then selected with blasticidin (10 µg/mL) for six days. After selection period CAS9-OCI-AML2 clones were sorted by manual seeding at a concentration of 0.4 cells/well in 96-well plates. Independent clones were isolated, and the CAS9 mRNA levels for each clone was quantified by quantitative reverse transcriptase-real time polymerase chain reaction (qRT-PCR). The clone with the highest level of CAS9 mRNA (designated clone 7), was further analyzed by immunoblotting for CAS9 protein levels. Subsequently, the Cas9-expressing clonal population was subject to another round of single cell cloning procedures described above. The clone (designated clone 7.2) from the second round of

67 selection was then characterized by immunoblotting for Cas9 protein expression and selected for screening as well as TAZ, and PISD knockout studies as described below.

2.2.4 CRISPR Screen

To identify gRNA that reduced the growth and viability of AML cells, we performed a pooled lentiviral gRNA screen cells as described previously (Hart et al., 2015). CAS9-OCI-AML2 cells were transduced with a pooled lentiviral library consisting of 91,320 sgRNAs in barcoded lentiviral vectors targeting 17,232 nuclear-encoded genes. The day after transduction, cells were treated with puromycin (2 μg/mL) to select transduced cells. Resistant clones were passaged at regular intervals (3-4 days). Cells were harvested at 3 and 17 days post-transduction. Genomic DNA was extracted from cell pellets obtained on day 3 as well as day 17 and sequenced on an Illumina NextSeq500 (Illumina, CA, USA). sgRNA depletion was characterized by model-based Analysis of Genome-wide CRISPR/Cas9 Knockout (MAGeCk) analysis as described below.

2.2.5 MAGeCK Analysis sgRNA sequences were first extracted from FASTQ files by trimming and then aligned to a reference library via Bowtie. Mapped sgRNA counts for samples were analyzed using MAGeCK 0.5.5, using default parameters (Li et al., 2014). Essential genes were identified at a FDR < 5%, by comparing of D17 sgRNA to D0 sgRNA.

2.2.6 Viral Infection

2.2.6.1 CRISPR-sgRNA Knockout

1.0 x 106 CAS9-OCI-AML2 cells were centrifuged and resuspended in 3 mL of medium containing 5 µg/mL of protamine sulfate. 1 mL of virus was added to cells, followed by a 24- hour incubation (37°C, 5% CO2). On the following day, cells were centrifuged and washed, then resuspended in fresh medium with puromycin (1.5 µg/mL). After at least 3 days of selection, equal numbers of cells were plated for growth assay, and counted by trypan blue exclusion staining for a period of 12-15 days post-transduction. To confirm knockout, 5-20 x 106 cells were collected at day 8 post-transduction for immunoblot analysis, as described in the mitochondria protein lysate, and immunoblotting sections below.

2.2.6.2 shRNA knockdown of AML Cell Lines

Lentiviral infections with shRNA constructs were performed as previously described in Cole et al., 2015. 5 x 106 cells were centrifuged and resuspended in 5 mL of medium containing 5 µg/mL of protamine sulfate. 2 mL of virus was added to cells, followed by a 24-hour incubation

(37°C, 5% CO2). On the following day, cells were centrifuged washed, and resuspended in fresh medium with puromycin (1.5 µg/mL for OCI-AML2, 2 µg/mL for TEX, K562, and U-937 cells). After 3 days of selection, equal numbers of cells were plated for growth assay, and counted by trypan blue exclusion staining for a period of 12 days post-transduction. To confirm target knockdown, 5-20 x 106 cells were collected at day 4 post-transduction for immunoblot analysis, as described in mitochondria protein lysate and immunoblotting sections below.

2.2.6.3 shRNA knockdown of Primary AML Cell Samples

Lentiviral infections with shRNA constructs were performed as previously described (Chan et al., 2015). Wells in a 24-well non-tissue-culture-treated plate were coated with 500 µl of RetroNectin® (20 µg/mL in phosphate-buffered saline, PBS) per well for 2 hours at room temperature. The wells were blocked with 1 mL of 2% BSA (W/V) for 30 minutes at room temperature. The BSA solution was then aspirated, and concentrated lentiviral particles in Hank’s balanced salt solution (HBSS) with 25 mM 4-(2-hydroxyethyl)-1- piperazineethanesulfonic acid (HEPES) were added to each well at a volume of 0.5 mL. The plate was then centrifuged at 3,000 rpm for 5 hours at room temperature to promote the attachment of lentiviral particles to RetroNectin®. After aspirating 0.4 mL of the viral supernatant, 5 x 105 cells were added to each well in 1 mL of X-VIVOÔ 10 (20% BIT, 50 ng/ml Flt3-L, 10 ng/mL IL-6, 50 ng/mL SCF, 25 ng/mL TPO, 10 ng/mL IL-3, 10 ng/mL G-CSF). The plate was then centrifuged again at 1,300 rpm for 10 minutes at room temperature to promote the interaction between the cells and lentiviral particles, and then transferred to a 37°C incubator to initiate lentiviral infection. Twenty-four hours afterwards, the cells were resuspended in fresh media at a concentration of 1 x 106 cells/mL and seeded in a 24-well plate at 1 mL per well. TAZ knockdown was confirmed by quantitative reverse qRT-PCR, 5-7 days after transduction, as described in the qRT-PCR section.

2.2.6.4 PISD Overexpression

OCI-AML2 cells were transduced with lentiviral particles containing PISD or control vector sequences, and then selected with 10 μg/ml blasticidin for 8 days. After selection, PISD or empty

69 vector overexpressing cells were transduced with TAZ or control shRNA constructs as described in shRNA knockdown of AML cell lines section. Eleven days after transduction with the PISD vector sequence PISD overexpression was confirmed by qRT-PCR, and cell phenotype was characterized by non-specific esterase staining, as described in the qRT-PCR and non-specific esterase sections below.

2.2.7 Mitochondrial Protein Lysates

Mitochondrial lysates were made using the cytochrome c releasing apoptosis kit (Abcam, Cambridge, UK) to quantify TAZ, as per manufacturer’s instructions, with minor modifications. We harvested confluent cells and froze these cells on dry ice. Cell pellets (10-20 x 106 cells) were then thawed on ice and lysed in distilled water containing protease inhibitors, by hypotonic shock for 15 minutes at 4ºC. The Cytosolic Extraction Buffer Mix containing 1,4-dithiothreitol (DTT) and protease inhibitors were added to the cell lysate, and samples were then centrifuged at 800 g at 4ºC for 20 minutes. The supernatant containing mitochondria was collected and centrifuged at 10,000 g for 30 minutes at 4ºC. Next, the supernatant was discarded, and the mitochondrial pellet was resuspended in ice-cold Mitochondrial Extraction Buffer, with DTT and protease inhibitors. Finally, mitochondria were lysed by adding 2X Laemmli buffer. The protein concentration in mitochondrial lysates was quantified using the DC protein determination kit (Bio-Rad, CA, USA) before immunoblotting.

2.2.8 Whole Cell Protein Lysates

For all other immunoblots whole cell lysates were used. To make whole cell lysates, cells (5 x 106) were washed with PBS followed by cell lysis in RIPA buffer. Protein concentration was measured by Bradford method (Protein assay dye, Bio-Rad, CA, USA).

2.2.9 Immunoblotting

Equal amounts of protein were loaded and fractionated on 10-12% sodium dodecyl sulfate- polyacrylamide gels. Proteins were separated on sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes or polyvinylidene difluoride prior to antibody treatment. Blots were blocked with 5% milk tris-buffered saline with Tween-20 (TBST) then incubated overnight with an appropriate primary antibody. Membranes were then blocked with 5% milk TBST, and further probed using an appropriate secondary antibody conjugated to horseradish peroxidase, and then developed using enhanced chemiluminescence detection

(ThermorFisher Scientific, MA, USA). The antibodies used for immunoblotting can be found in the key resource table.

2.2.10 Basal Apoptosis

OCI-AML2 cells were transduced with shRNA in lentiviral vectors targeting TAZ or control sequences. Seven days after transduction, cells were harvested, and cell death was measured by flow cytometry with annexin V, fluorescein isothiocyanate (FITC) and propidium iodide, (PI) R- Phycoerythrin (PE, Biovision Research Products, CA, USA) staining according to the manufacturer’s instructions. Flow cytometry data were acquired using a LSRFORTESSA X20 (BD Biosciences, FL, USA) flow cytometer and frequency of annexin V, PI negative cells were quantified with the FlowJo software (TreeStar, OR, USA)

2.2.11 Cell Cycle Analysis

OCI-AML2 cells were transduced with shRNA in lentiviral vectors targeting TAZ or control sequences. Seven days after transduction, cells were harvested, washed with cold PBS, and fixed in absolute ethanol and PBS (80% ethanol, 20% PBS) for a minimum of one hour. Afterwards, cells were washed in cold PBS, and treated with 5 µg/mL of DNase-free RNase A (Invitrogen, CA, USA) at 37°C for 30 minutes, then incubated in PBS containing 5 µg/mL of propidium iodide (PI) for 5-10 minutes at room temperature. DNA content was measured using flow cytometry (FACs CANTO, BD, FL, USA) and analyzed with FlowJo software (TreeStar, OR, USA).

2.2.12 Colony Formation Assays

2.2.12.1 Leukemia Cell Lines

Five (TAZ knockdown OCI-AML2, and TEX cells) or nine (PISD knockout CAS9-OCI-AML2 cells) days after transduction cells were plated at equal concentrations (CAS9-OCI-AML2 and OCI-AML2 = 750 cells; TEX cells = 2,000) in duplicate 35mm dishes. (Nunclon, Rochester, USA) to a final volume of 1 mL/dish in MethoCultä H4100 media (StemCell Technologies, BC, Canada) supplemented with 30% FCS (CAS9-OCI-AML2 and OCI-AML2) or MethoCultä H4100 media (StemCell Technologies, BC, Canada) supplemented with 30% FCS, 20 ng/mL SCF, and 2 ng/mL IL-3 (TEX cells). After incubating the dishes for 5 (CAS9-OCI-AML2), 7 (OCI-AML2), or 10 (TEX) days at 37°C, 5% CO2 with 95% humidity, the number of colonies

71 containing 10 or more cells were counted on an inverted microscope. The mean of the duplicate plates for each condition are presented. During serial re-plating colonies were removed from MethoCultä by vortexing and washing in PBS, cells were counted and then re-plated in MethoCultä every 5 (PISD knockout CAS9-OCI-AML2 cells) or 7 (TAZ knockdown TEX cells) days. The mean of the duplicate plates for each condition are presented.

To assess clonogenic growth after PS treatment, OCI-AML2 cell were pre-treated with 25 µM of PS or the vehicle control for 9 days. Afterwards, treated OCI-AML2 cells were platted at equal volumes in duplicate 35mm dishes. (Nunclon, NY, USA) to a final volume of 1 mL/dish in MethoCult H4100 media (StemCell Technologies, BC, Canada) supplemented with 30% FCS without PS. After 5 days of incubation at 37°C, 5% CO2 with 95% humidity, the number of colonies containing 10 or more cells were counted on an inverted microscope. The mean of the duplicate plates for each condition are presented.

2.2.12.2 Primary AML Patient Samples

Fresh AML mononuclear cells and normal peripheral blood stem cells (PBSCs) (4 × 105 AML cells, 2 x 105 PBSCs) were incubated with MMV007285, PS or vehicle control for 48 hours in MyelocultäH5100, supplemented with 100 ng/mL SCF, 10 ng/mL Flt3-L, 20 ng/mL IL-7, 10 ng/mL IL-3, 20 ng/mL IL-6, 20 ng/mL G-CSF, 20 ng/mL granulocyte-macrophage colony- stimulating factor (GM-CSF). Treated PBSCs or AML patient samples were platted in MethoCultä H4434 medium (StemCell Technologies, BC, Canada). After incubating the dishes for 7 days (AML) or 2 weeks (normal hematopoietic cells) at 37 °C with 5% CO2 and 95% humidity, AML colonies containing 10 or more cells and PBSC colonies containing more than 50 cells were counted. The mean of duplicate plates for each condition are presented.

To assess the clonogenic growth of mouse hematopoietic progenitor cells after TAZ knockdown and hematopoietic stress, 40000 bone marrow cells cells/ml/dish from WT or Taz-KD mice treated with 0 or 200 mg/Kg of 5-FU were platted on MethoCult GF M3534, and incubated at

37°C, 5% CO2 with 95% humidity, 15 days after 5-FU treatment. Six days after, colonies containing 50 or more cells were enumerated. During serial re-plating, colonies were removed from MethoCult GF M3534 by vortexing and washing in PBS. Cells were counted and 40000 cells/ml/dish replated on MethoCult GF M3534. After 6 days in culture, colonies containing 50

72 cells or more cells were enumerated. The mean of the duplicate plates for each condition are presented.

The effect MMV007285 has on hematopoietic cells subjected to stress was characterized by incubating 160000 bone marrow cells isolated from C57BL/6J mice 15 days after treatment with 0 or 200 mg/Kg of 5-FU in StemSpan™ SFEM media (50 ng/mL mouseSCF, 50 ng/mL mouseIL-6, 1:320 mouseIL-3 conditioned media) containing 0 or 12.5 µM MMV007285 for 72 hours. After pre-treatment the cells were washed and plated on MethoCult GF M3534 with 0 or 12.5 µM of MMV007285, then incubated at 37°C, 5% CO2 with 95% humidity for 6 days. Finally, colonies containing 50 or more cell were enumerated. The mean of the duplicate plates for each condition are presented.

2.2.13 RNA-Sequencing

Total RNA was isolated from OCI-AML2 cells using the RNeasy Plus Mini Kit (Qiagen, Hilden, Germany) 7 days after transduction with shRNA targeting TAZ or control sequences. The experimental design consisted of 2 biological replicates of control cells and 2 biological replicates of TAZ knockdown samples. Quality of total RNA samples were checked on an Agilent Bioanalyzer 2100 RNA Nano chip (CA, USA) following Agilent Technologies’ recommendation. Concentration was measured by Qubit RNA HS Assay on a Qubit fluorometer (ThermoFisher Scientific, MA, USA). RNA library preparation was performed following the NEBNext Ultra Directional Library (MA, USA) Preparation Protocol. Briefly, 800 ng of total RNA was used as the input material and enriched for poly-A mRNA, fragmented into the 200- 300-bases range for 4 minutes at 94°C and converted to double stranded cDNA, end-repaired and adenylated at the 3’ to create an overhang A to allow for ligation of Illumina adapters with an overhang T; library fragments were amplified under the following conditions: initial denaturation at 98°C for 10 seconds, followed by 13 cycles of 98°C for 10 seconds, 60°C for 30 seconds and 72°C for 30 seconds, and finally an extension step for 5 minutes at 72°C; at the amplification step, each sample were amplified with a different barcoded adapters to allow for multiplex sequencing. One µl of the final RNA libraries was loaded on a Bioanalyzer 2100 DNA High Sensitivity chip (Agilent Technologies, CA, USA) to check for size; RNA libraries were quantified by qPCR using the KAPA Library Quantification Illumina/ABI Prism Kit protocol (MA, USA). Libraries were pooled in equimolar quantities and paired-end sequenced on 1 lane of a High Throughput Run Mode flowcell with the V4 sequencing chemistry on an Illumina

HiSeq 2500 platform (CA, USA) following Illumina’s recommended protocol to generate paired- end reads of 126-bases in length. RNA-sequencing data was used to perform differential gene expression analysis as described below.

2.2.14 Differential Gene Expression Analysis

Prior to analysis, read adapters and low quality ends were removed using Trim Galore v. 0.4.0. Reads were aligned against hg19 using Tophat v. 2.0.11. Read counts per gene were obtained through htseq-count v.0.6.1p2 in the mode “intersection_nonempty”. After removing genes whose cpm (counts per million reads) were less than 0.5 in at least one sample, edgeR R package v.3.16.5 was used to normalize the data using the TMM (trimmed mean of M values) method and to estimate differential expression by applying the generalized linear model likelihood ratio test between the control OCI-AML2 samples and the TAZ-KD OCI-AML2 samples. A score that ranks genes in TAZ knockdown samples from the most up-regulated to the most down-regulated compared to control shRNA samples was calculated using the formula `-log10(pvalue) * sign(logFC)`. Pathway, LSC+/LSC- signature, DMAP signature, and PERT deconvolution analyses was performed using the differential expression data generated.

2.2.15 Pathway Analysis

The rank list genes generated for TAZ knockdown and control shRNA samples was used for Gene set enrichment analysis (GSEA) with a setting of 1000 permutations, and default parameters. The pathway database from the Bader lab (http://baderlab.org/GeneSets, version of September 2017) which contains the Gene Ontology, Biological Process, as well as Msigdb-c2 and –c3, IOB, NetPAth, HumanCyc, Reactome and Panther was used. Results were visualized in Cytoscape 3.6.1 using EnrichmentMap 3.1, clustered and annotated using AutoAnnotate 1.2. Cluster labels were manually edited for clarity. Pathway analysis was validated by measuring mRNA levels of TLR4, TLR8, IL6, and IFNb in TAZ knockdown or MMV treated samples, by qRT-PCR as described in the qRT-PCR section.

2.2.16 TCGA Unsupervised Clustering

The TCGA AML RNAseq data were retrieved from the GDC portal (https://portal.gdc.cancer.gov/). Read counts were normalized using the TMM method in the edgeR R package. The TCGA samples were hierarchically clustered using the Ward linkage method and Euclidean distance on the basis of the scaled log2(CPM) of the 5,334 genes with

74 standard deviation > 1. AU (approximately unbiased) cluster probability, computed by multiscale bootstrap resampling, was calculated using the R package pvclust.

2.2.17 LSC+/LSC- Signature Analysis

Illumina beadchip transcriptomics data containing LSC+ and LSC- sorted AML fractions were obtained from the Gene Expression Omnibus data portal (GSE76008) (Ng et al., 2016) and differential expression between LSC+ and LSC- fractions was calculated using a moderated t-test available in the limma R package 3.28.21, incorporating array batch effects into the linear model. Firstly, the 500 most highly up-regulated genes in TAZ knockdown and control shRNA samples were compared to normalized gene expression of LSC+/LSC- fractions, then subsequently visualized using the heatmap.2 function in R by setting row normalization as true. In addition, the top up-regulated in the LSC+ fractions to top down-regulated in comparison to LSC- fractions was scored using the formula `-log10(pvalue) * sign(logFC)`. The 400 highest scoring in LSC+ and the 400 highest scoring LSC- genes were used as signature gene set for LSC+ and LSC- samples respectively. Then gene set enrichment analysis (GSEA) (Broad Institute, MA, USA) with 1000 permutations and default parameters was used to measure the enrichment of genes upregulated in TAZ knockdown and control samples in the LSC+ or LSC- gene signatures.

2.2.18 DMAP- Signature Analysis

Changes in gene expression after TAZ knockdown were mapped to the Gene Expression Omnibus dataset GSE24759 (DMAP) (Novershtern et al., 2011), containing Affymetrix GeneChip HT-HG_U133A Early Access Array gene expression data of 20 distinct hematopoietic cell states. GSE24759 data were background corrected using Robust Multi-Array Average (RMA), quantile normalized using the expresso function of the affy Bioconductor package (affy_1.38.1, R 3.0.1), array batch corrected using the ComBat function of the sva package (sva_3.6.0), and then standardized using Z score for each gene across samples. Bar graphs were created by selecting genes that were up- (top 500) TAZ knockdown cells as well as control shRNA cells, and summing the number of standardized data points that were above (>0) or below (<0) the mean for each DMAP cell population, corrected by the number of samples per population and results obtained from 1,000 random selections of DMAP genes.

2.2.19 PERT Deconvolution Analysis

The PERT deconvolution was run on the TMM-normalized CPM data from TAZ knockdown and control samples (Qiao et al., 2012). The batch-corrected linear RMA-normalized data from the GSE24759 (DMAP) data were used as the reference profile. The vector theta from the PERT output was used to estimate the percentage of reference populations within each TAZ knockdown and control samples. The theta percentage of each biological replicate of the same populations was summed. A high theta value for one population means a high estimated presence of the profile in the TAZ knock down or control sample.

2.2.20 Non-Specific Esterase Staining

Cytospin slides were prepared by centrifuging OCI-AML2 cells (1 x 105) 7-11 days after transduction (TAZ shRNA and PISD over expression plasmids respectively) onto glass slides, which were then subsequently stained for nonspecific esterase (NSE) activity, using the staining kit from Sigma-Aldrich, as per manufacturer’s instructions. The monocyte non-specific esterase inhibitor sodium fluoride was used to confirm the specificity of the reaction. Finally, glass slides were dried at room temperature, mounted in Clear Mount and scanned using the Aperio ScanScope AT2 (Leica, Wetzlar, Germany), and then analyzed by ImageJ as described below.

2.2.21 NSE Staining Analysis Using ImageJ

Aperio ImageScope (Leica, Wetzlar, Germany) was used to select 5 random sections from the scanned AML cell images. ImageJ was used to quantify the intensity of NSE staining. The average of the 5 sections for each condition, from three independent experiments is presented.

2.2.22 Cell Surface Phenotype of OCI-AML2

OCI-AML2 cells were co-immunostained with the viability dye 7AAD and anti-human CD11b or 7AAD and anti-human CD14. Flow cytometry data were acquired using a LSRFORTESSA X20 (BD Biosciences, FL, USA) flow cytometer and analyzed with the FlowJo software (TreeStar, OR, USA).

2.2.23 RNA Isolation and qRT-PCR

Total RNA was isolated from leukemia cells using the RNeasy Plus Mini Kit (Qiagen), and cDNA was prepared using SuperScript IV Reverse Transcriptase (ThermoreFisher, MA, USA). Equal amounts of cDNA for each sample were added to a prepared master mix (Power SYBRä

Green PCR Master mix; Applied Biosystemsä, CA, USA). qRT-PCR reactions were performed on an ABI Prism 7900 sequence detection system (Applied Biosystems, CA, USA). The relative abundance of a transcript was represented by the threshold cycle of amplification (CT), which is inversely correlated to the amount of target RNA/first-strand cDNA being amplified. To normalize for equal amounts of cDNA we assayed the transcript levels of 18s ribosomal RNA gene. The comparative CT method was calculated as per manufacturer’s instructions. Primers that were used are listed in Table 7.

2.2.24 Lipid Extraction Protocols

2.2.24.1 Ultra-Hight Performance Liquid Chromatography/Mass Spectrometry (UHPLC/MS)

Frozen leukemia cells (day 13 post-transduction), or mouse bone marrow samples (12.9-19.4 weeks after doxycycline treatment) were resuspended in PBS, and then added to 2:1 chloroform: methanol (v/v) on ice. Samples were vortexed, and 400 µL of 0.2 M sodium-phosphate buffer was added to induce layer separation. After inversion, samples were centrifuged for 5 minutes at 1734 g. The organic layer, which contains the lipids, was collected. An additional 2 mL of chloroform was added to the aqueous layer. Samples were re-vortexed, re-centrifuged, and the organic layer was collected and combined with the organic layer of the first extraction. The methanol/water layer was discarded. Samples were stored in chloroform at 4°C until further analysis.

2.2.24.2 Densitometric Analysis

Lipids from TAZ knockdown (day 8 post-transduction), PISD knockout (15 days post- transduction), PS supplemented (day 7 post supplementation), PE/LPE supplemented (day 4 post supplementation) or MMV treated (day 7 post treatment) leukemia cell lines were extracted for densitometric analysis as described previously (FOLCH et al., 1957). Briefly, Samples were resuspended in PBS, and mixed with 1:2 chloroform:methanol (v/v). H20 (HPLC grade) was added to increase separation. Afterwards, samples were mixed then centrifuged for 6 minutes at 1700g. The organic layer, which contains all the lipids was collected for the characterization of phospholipid distribution, using densitometric analysis.

2.2.25 UHPLC/MS Characterization of MLCL and CL

UHPLC/MS analysis was performed as previously described in (Bradley et al., 2017), with modification to detect MLCL and CL. Samples were dried under N2 gas and resuspended in 100 μL 65:35:5 acetonitrile: isopropanol: water (v/v/v) with 0.1% formic acid. A Dionex UltiMate 3000 UHPLC System was used (Dionex Corporation, Bannockburn, IL, USA), and was coupled to a Thermo Q-Extractive Quadrupole-Orbitrap mass spectrometer (ThermoFisher Scientific, Waltham, MA, USA). A reversed-phase, binary multistep, ultrahigh-performance liquid chromatography (UHPLC) protocol was used with a C18 Ascentis Express column with dimensions of 15cm x 2.1mm x 2.0μm (Sigma-Aldrich®, MO, USA). The mobile phase consisted of A: 60:40 acetonnitrile: water (v/v), 10mM ammonium formate, and 0.1% formic acid, and B: 90:10 isopopanol: acetonitrile (v/v), 10mM ammonium formate and 0.1% formic acid. The gradient protocol used was as follows: from 0 – 1.5 minutes it was 32% B, from 1.5 – 4 minutes 45% B, from 4 – 8 minutes 50% B, from 8 – 18 minutes 55% B, from 18 – 20 minutes 60% B, from 20 – 35 minutes 70% B, from 35 – 40 minutes 95% B, from 40 – 45 minutes 95% B, from 45 – 47 minutes B was decreased to 32%, and allowed to equilibrate until the 48 minute mark. The flow was set to 260μL/minute, column temperature at 45°C, and tray temperature at 4°C. The injection volume was 5μL. The mass spectrometer was operated in negative electrospray ionization mode, 35,000 resolution, with a scan range of m/z 200 to 2,000, spray voltage of -3.0 kV, sheath gas flow rate of 35 units, and capillary temperature of 300°C. MS/MS experiments were done under data-dependent conditions with top 5 ions and 17,500 resolution, and the normalized collision energy was 17.5. Thermo X calibur QualBrowser software (version 2.1; ThermoFisher Scientific, MA, USA) was used for extracting ion chromatograms, integrating peak areas, and exporting MS-MS spectra. Chromeleon Xpress (version 7.2; ThermoFisher Scientific, MA, USA) was used to monitor and control the Dionex UHPLC settings. MS/MS spectra were analyzed using the NIST MS search program (version 2.0; National Institute of Standards and Technology, MD, USA) and LipidBlast library. Extracted ion profiles were integrated for 20 cardiolipin species and 8 mono-lyso cardiolipin species using their deprotonated molecular ion m/z ratios ([M-H]-) within a 0.2Da window (Table 10).

Lipids Polarity Molecular Mode m/z ion NCE Approximate RT Ion (min)

52:2 - [M-H]- Top-5 DDA 1165.7666 17.5 11.55 - 11.88 MLCL

52:3 - [M-H]- Top-5 DDA 1163.7509 17.5 10.85 - 10.96 MLCL

52:4 - [M-H]- Top-5 DDA 1161.7353 17.5 8.79 - 9.53 MLCL

54:2 - [M-H]- Top-5 DDA 1193.7979 17.5 12.46 - 12.76 MLCL

54:3 - [M-H]- Top-5 DDA 1191.7822 17.5 11.45 - 11.90 MLCL

54:4 - [M-H]- Top-5 DDA 1189.7666 17.5 10.41 - 10.90 MLCL

54:5 - [M-H]- Top-5 DDA 1187.7509 17.5 9.26 - 9.71 MLCL

54:6 - [M-H]- Top-5 DDA 1185.7353 17.5 10.32 - 10.66 MLCL

64:3 CL - [M-H]- Top-5 DDA 1345.9188 17.5 14.51 - 14.74

66:2 CL - [M-H]- Top-5 DDA 1375.9659 17.5 16.29 - 16.45

66:3 CL - [M-H]- Top-5 DDA 1373.9502 17.5 15.41 - 15.70

66:4 CL - [M-H]- Top-5 DDA 1371.9345 17.5 14.57 - 14.85

68:3 CL - [M-H]- Top-5 DDA 1401.9816 17.5 16.23 - 16.55

68:4 CL - [M-H]- Top-5 DDA 1399.9659 17.5 15.42 - 15.76

68:5 CL - [M-H]- Top-5 DDA 1397.9493 17.5 14.71 - 14.98

70:4 CL - [M-H]- Top-5 DDA 1427.9973 17.5 16.21 - 16.55

70:5 CL - [M-H]- Top-5 DDA 1425.9817 17.5 15.54 - 15.85

70:6 CL - [M-H]- Top-5 DDA 1423.9660 17.5 14.85 - 15.20

70:7 CL - [M-H]- Top-5 DDA 1421.9503 17.5 15.43 - 15.73

72:5 CL - [M-H]- Top-5 DDA 1454.0131 17.5 16.40 - 16.59

72:6 CL - [M-H]- Top-5 DDA 1451.9974 17.5 15.69 - 15.97

72:7 CL - [M-H]- Top-5 DDA 1449.9817 17.5 16.24 - 16.54

72:8 CL - [M-H]- Top-5 DDA 1447.9661 17.5 15.55 - 15.81

72:9 CL - [M-H]- Top-5 DDA 1445.9504 17.5 14.87 - 15.07

74:6 CL - [M-H]- Top-5 DDA 1480.0288 17.5 16.52 - 16.59

74:7 CL - [M-H]- Top-5 DDA 1478.0131 17.5 15.91 - 16.08

74:8 CL - [M-H]- Top-5 DDA 1475.9975 17.5 16.36 - 16.61

74:9 CL - [M-H]- Top-5 DDA 1473.9818 17.5 15.72 - 15.94

Table 10. Detection of lipids by MS/MS. Abbreviations: MLCL = mono-lyso cardiolipin, CL – cardiolipin, DDA = data-dependent acquisition, m/z = mass-to-charge ratio, NCE = normalized collision energy, RT = retention time.

2.2.26 Densitometric Characterization of Phospholipids

Lipids extracted as described above, were spotted onto high-performance thin layer chromatography plates (HPTLC; 5633-5, EMD Chemicals, Darmstadt, Germany) and individual phospholipids were separated using a chloroform: methanol: acetic acid: water (100:75:7:4) solvent system (Stefanyk et al., 2010). A reference standard containing a mixture of sphingomyelin, phosphatidylcholine (PC), phosphatidylserine (PS), phosphatidylinositol (PI), phosphatidylethanolamine (PE), and cardiolipin was spotted alongside each plate to correctly identify the phospholipid species. After allowing the solvent to run up each plate for 45 minutes, the plates were then charred at 180°C with a 10% (w/v) copper (II) sulfate in 8% phosphoric acid solution for 15 minutes (Churchward et al., 2008). Images of the HPTLC plates were captured using a CCD camera on a Fluorchem 5500 imaging station (Alpha Innotech, CA, USA) under reflective white light, and then quantified as described in densitometric phospholipid analysis section below.

2.2.27 Densitometric Phospholipids Analysis

Densitometry analyses were then performed using Image StudioTM (LI-COR, NB, USA) and the percent distribution of each phospholipid specie was calculated by dividing the densitometric value for the individual phospholipid specie by the sum of the densitometric values for all phospholipid species combined.

2.2.28 Quantification of Extracellular PS

OCI-AML2 cells were transduced with shRNA in lentiviral vectors targeting TAZ or control sequences. Seven days after transduction, cells were harvested, and cell death was measured by flow cytometry with annexin V, FITC and PI, PE (Biovision Research Products, CA, USA) staining according to the manufacturer’s instructions. Flow cytometry data were acquired using a LSRFORTESSA X20 (BD Biosciences, FL, USA) flow cytometer and frequency of annexin V positive cells were quantified by with the FlowJo software (TreeStar, OR, USA).

2.2.29 Intracellular PS Quantification by Confocal Microscopy

Cytospin slides were prepared by centrifuging 10,000 TAZ knockdown 8227 (day 11 post- transduction) or primary AML (day 7 post-transduction) cells onto glass slides, which were then fixed in 4% paraformaldehyde for 5 minutes, permeabilized (0.2% saponin) for 20 minutes, and then stained with annexin V Alexa-647 (1:15) for 1 hour, and DAPI (1:500) for 2 minutes, and

81 then mounted with antifade solution. Finally, images were acquired using the Olympus FV1000 (Tokoyo, Japan) confocal microscope, and the amount of PS was quantified as described below.

2.2.30 Intracellular PS Quantification by ImageJ

Olympus FV1000 was used to select 10-13 random sections from acquired AML cell images. ImageJ was used to quantify the number of DAPI+ cells, and intensity of PS staining. Then, the average PS intensity per cell of each image acquired was quantified by ImageJ using the formula:

�� ÷ #�� + ��

The average of the 10-13 sections for each condition is presented.

2.2.31 Sensitivity to Extrinsic Apoptosis

OCI-AML2 cells transduced with shRNA in lentivrial vectors targeting TAZ or control sequences were then co-treated with 2.5µg/mL of cycloheximide and increasing concentrations of TRAIL for 16 hours, 13 days after transduction. Cells were stained with annexin V and PI staining (BD Biosciences, FL, USA) according to the manufacturer’s instructions. Flow cytometry data were acquired using a LSRFORTESSA X20 (BD Biosciences, FL, USA) flow cytometer and frequency of annexin V, PI negative cells were quantified with the FlowJo software (TreeStar, OR, USA)

2.2.32 Seahorse

Oxygen consumption rate (OCR), and extracellular acidification rate (ECAR) were measured in AML cells after TAZ or PISD knockdown using the Seahorse XF-96 analyser (Seahorse Bioscience, MA, USA). OCR was measured as a marker of electron transport chain activity. Six days (TAZ knockdown) or 8 days (PISD knockout) after transduction cells were resuspended in unbuffered medium and seeded at 1 x 105 cells/well in Cell-Tak coated (0.15 µg/well) XF96 plates. Cells were equilibrated in the unbuffered a-MEM medium (supplemented with 2% FCS) or XF Assay medium (supplemented with 4.5 g/l (25mM) glucose + 1mM pyruvate) for 45 minutes at 37°C in a CO2-free incubator before being transferred to the XF96 analyser. Spare reserve capacity of the mitochondrial respiratory chain was measured by treating cells with oligomycin, and carbonyl cyanide p-trifluoromethoxyphenylhydrazone (FCCP) in succession.

2.2.33 Cellular ROS

Four days after transduction with TAZ or control shRNA sequences cells were stained with Carbocy-H2DCFDA in PBS buffer at 37°C for 30 minutes and resuspended in PBS with the viability dye PI to assess the reactive oxygen species produced by viable cells. Data was acquired by LSRFortessaX20 (BD Biosciences, FL, USA), and mean fluorescent intensity (MFI) of Carbocy-H2DCFDA in PI negative cells were analyzed using FloJo (TreeStar, OR, USA).

2.2.34 Mitochondrial Mass

Eight days after transduction with TAZ or control shRNA sequences mitochondrial mass was characterized by quantifying the mtDNA to nuclear DNA ratio. mtDNA and nuclear DNA was extracted using the DNeasy Blood and Tissue kit (Qiagen, MD, USA) from AML cell lines transduced with lentiviral vectors with shRNA targeting TAZ, or control sequences. The relative amount of mtDNA to nuclear DNA was measured using qRT-PCR, using primer pairs for mitochondrial ND1, and nuclear HGB outlined in Table 7.

2.2.35 Live Imaging

The mitochondrial morphology of cells transduced with shRNA in lentiviral vectors targeting TAZ or control sequences, was assessed in live cells by labeling the mitochondria with 1 µM MitoTracker and Red CMXRos (ThermoFisher Scientific, MA, USA) for 30 minutes at 37°C, 7 days after transduction. Cells were washed twice with complete media and transferred to a poly- L-lysine (Sigma-Aldrich®, MO, USA) coated imaging plate. Images were acquired on spinning disc confocal microscope VisiScope CSU-W1 Olympus IX83 (Tokyo, Japan). Mitochondria shape was characterized by quantifying mitochondrial aspect ratio as described below.

2.2.36 Aspect Ratio

For mitochondrial shape assessment an aspect ratio of each mitochondrion was measured, confocal z-stacks were collected and projected as a z-project in ImageJ (NIH, Bethesda, MD). The population median, quartiles, minimum, and maximum of fields (40-60 cells) are expressed as a boxplot.

2.2.37 Electron Microscopy

The mitochondrial morphology of cells transduced with shRNA in lentiviral vectors targeting TAZ or control sequences, was assessed by transmission electron microscopy (TEM). To perform TEM, cells were harvested at day 8 post-transduction, and fixed with a modified Graham-Karnovsky’s fixative-4% paraformaldehyde plus 1% glutaraldehyde in 0.1 M phosphate buffer pH 7.2 (PB)-for 30 minutes at room temperature, and 4°C overnight. Cells were then washed with phosphate buffer (PB, NaH2P04. H2O + Na2H2PO4) 3 times for 15 minutes. Next, cells were post-fixed with 1% osmium tetroxide buffered with PB for 1 hour and washed again using PB twice for 20 minutes, dehydrated with an ethanol series then infiltrated with propylene oxide. The samples were then resin embedded with Epon, which was polymerized at 40°C for 48 hours. Solid epoxy blocks were sectioned on a Reichert Ultracut E microtome to 90 nm thickness, collected on 300 mesh copper grids and counterstained with uranyl acetate and lead citrate. A Hitachi H7000 (Hitachi, Tokyo, Japan) transmission electron microscope with a beam current of 25 µm, was used to view the sections. Images of the sections were acquired by the Advanced Microscopy Techniques (AMT)-CCD camera (AMT, MA, USA). A composite of representative images acquired are shown in Figure 27

2.2.38 Lipid Overlay Assay

Binding of PISD to lipid strips (Echelon Biosciences) was performed as per manufacturer's instructions. Briefly, strips were blocked in 3% fatty acid-free BSA in Tris-buffered saline (TBS, 25 mM Tris, 150 mM NaCl pH7.2) followed by incubation with 1 µg/mL of PISD recombinant protein in 1% fatty acid-free BSA in TBS for 16 hours at 4°C. Strips were washed extensively in TBST (0.05% Tween-20 in TBS) buffer to remove unbound proteins, and incubated with PISD antibody for 16 hours at 4°C. PISD protein bound to lipid strips were detected by immunoblotting.

2.2.39 Lipid Supplementation and Growth Analysis

All lipids were purchased from Avanti Polar lipids, lysophosphatidylethanolamine (LPE), PE, PS. A stock solution of LPE (18mM) and PE (25mM) were made by dissolving lipid powder in 65 chloroform: 35 methanol: 8 water (v:v:v), whereas a 25mM stock solution of PS was made by dissolving PS powder in 70% methanol: 20% TWIN-80: 10% propylene glycol (PEG) (v:v:v). After lipid supplementation, equal numbers of cells were plated for growth assay, and counted by trypan blue exclusion staining for a period of 8-14 days.

2.2.40 MMV Treatment and Growth Analysis

Cell lines or primary samples were seeded at 1 x 105 cells/mL in 20 mL of medium in T75 flasks. Cells were then treated with increasing concentrations of MMV007285 or DMSO control. Fresh MMV or vehicle control was added every 3 days during the incubation period. After MMV treatment, equal numbers of cells were plated for growth assay, and counted by trypan blue exclusion staining for a period of 14 days

2.2.41 8227 Flow Cytometry

8227 cells were co-immunostained with the viability dye Aqua Dead (ThermoFisher Science, MA, USA), anti-human antibodies recognizing CD34, CD38, Flow cytometry data were acquired using a LSRFORTESSA X20 (BD Biosciences, FL, USA) flow cytometer and frequency of viable 34+,38- cells were analyzed with the FlowJo software (TreeStar, OR, USA).

2.2.42 PS Quantification by Flow Cytometry

Five days after MMV treatment primary AML cells were fixed in 4% paraformaldehyde for 5 minutes, permeabilized (0.2% saponin) for 20 minutes, and then stained with Annexin V FITC (1:25) for 1 hour. Flow cytometry data were acquired using a LSRFORTESSA X20 (BD Biosciences, FL, USA) flow cytometer and annexin V MFI was analyzed with the FlowJo software (TreeStar, OR, USA).

2.2.43 MMV and CL075 Synergism Studies

OCI-AML2 leukemia cells were seeded in 96 well plates, and then treated with increasing concentrations of CL075 with and without increasing concentrations of MMV007285. Cells were incubated at 37°C with 5% CO2 for 72 hours. Cell growth and viability was determined using the MTS assay (Promega), as per manufacturer’s instructions. For each combination excess over Bliss additivism (EOBA) values were calculated as previously described (Borisy et al., 2003).

2.2.44 Animal Studies

2.2.44.1 Hematopoiesis in WT and Taz-KD mice

In order to understand the role of TAZ in normal hematopoiesis under basal conditions, adult (7.6-14.3) week iDOX-Taz-KD mice were treated with doxycycline for 12.9-19.4 weeks, and then sacrificed.

To understand the role of TAZ in hematopoietic stress, doxycycline treated iDOX-Taz-KD mice were injected intraperitoneally (IP) with 200mg/kg of 5-FU (Sigma-Aldrich®, MO, USA), dissolved in phosphate-buffered saline (Gibco, MA, USA). Peripheral blood was collected at days 0, 3, 7, 10, and 15 days after 5-fluorouracil (5-FU) injection from the tail vein into ethylenediaminetetraacetic acid (EDTA)-containing tubes. White blood cell, neutrophil, lymphocyte, red blood cell, and platelet counts were enumerated using the HEMAVET 950FS (Drew Scientific Inc, FL, USA). All mice were sacrificed 15 or 23 days after 5-FU treatment.

At the time of sacrifice, blood was isolated by cardiac puncture for complete blood counts, and hematopoietic cells were flushed from the femur and tibiae of WT and iDOX-TAZ-KD mice, using DMEM + 2% FCS. Red cell were lysed, and the remaining cells were passed through a 100 µM strainer. Bone marrow samples were depleted of mature myeloid and lymphoid lineages (Lin-) using magnetic bead separation with biotin-conjugated antibodies CD5, CD45R (B220), CD11b, Anti-Gr-1 (Ly-6G/C), 7-4 and Ter-119 (Miltenyi Biotec, mouse lineage depletion kit).

Lin- cells were co-immunostained with anti-mouse antibodies recognizing Sca1, CKIT, CD48, CD150, CD34, and Fc receptor (FCyR), and the viability dye 7AAD. Flow cytometry data were acquired using a FACSCanto II (BD Biosciences) flow cytometer and analyzed with the FlowJo software (TreeStar, OR, USA).

2.2.44.2 TEX and 8227 Engraftment

Equal numbers of TEX or 8227 cells (1-2 x 106) transduced with shRNA sequences in lentiviral vectors targeting TAZ or control sequences or treated with vehicle control or PS, were injected into the right femur of sublethally irradiated NOD/SCID-GF mice with human IL-3, GM-CSF, and Steel factor (Nicolini et al., 2004). Five weeks after injection, mice were sacrificed, and the percentage of human CD45+ was enumerated in TEX and PS supplemented 8227 cells by flow cytometry. Whereas in 8227 cells transduced with TAZ shRNA, engraftment of cells into the bone marrow was assessed by measuring GFP+ CD45+ cells by flow cytometry, and then calculating relative human cell engraftment as described in the calculation of engraftment potential section.

2.2.44.3 Stability and Pharmacokinetics of MMV007285

The concentration of intact MMV007285 in blood plasma was measured using Waters Xevo Quadruple Time-of-flight hybrid MS system coupled with ACQUITY UPLC (Waters Limited,

Mississauga, Canada). Chromatographic separations were carried out on an ACQUITY UPLC BEH C18 (2.1 X 50 mm, 1.7 µm) column (Waters Limited, ON, Canada). The mobile phase was 0.1% formic acid in water (solvent A) and 0.1% formic acid in acetonitrile (solvent B). Compound concentrations were measured relative to a freshly prepared calibration curve.

2.2.44.4 Subcutaneous AML Xenografts

To understand the role of TAZ in AML differentiation and PISD in AML growth in vivo, OCI- AML2 or CAS9-OCI-AML2 cells were first transduced with shRNA targeting TAZ/control sequences, or sgRNA targeting PISD/control sequences respectively. 1 × 106 transduced cells were then in injected subcutaneously into the flanks of male SCID mice (Ontario Cancer Institute). Differentiation after TAZ knock down was evaluated by measuring differentiation markers in sub-cutaneous tumors 8 days after injection, whereas role of PISD in tumor growth was characterized by measuring tumor volume every 2-3 days.

The in vivo anti-tumor activity and toxicity of MMV007285 was also assessed by injecting OCI- AML2 human leukemia cells (1 × 106) subcutaneously into the flanks of male SCID mice (Ontario Cancer Institute). When the tumors were palpable, mice were treated with MMV007285 or vehicle control (5% DMSO, 47.5% PEG400, 10% Tween-80, 37.5% H20) by oral gavage (300 mg/kg twice daily for 5 of 7 days) for 10 days (n = 10/group). Tumor volume was measured every 2-3 days. Nineteen after injection mice body weight was measured, afterwards mice were sacrificed, peripheral blood was collected then alkaline phosphatase (ALP), aspartate transaminase (AST), bilirubin (Bili), creatine kinase (CK), creatinine (Cr) levels were measured by Idexx Laboratories (Ontario, Canada).

2.2.44.5 Primary AML Engraftment Models

The role of TAZ in leukemia stem cells in vivo was characterized by transducing primary AML patient samples with TAZ shRNA or control shRNA containing lenti-viral vectors. NOD/SCID mice were conditioned with 208 rad of irradiation from a 137 rad source, and a 200 μg of anti- mouse CD122 48 hours before transplantation. Forty-eight hours after transduction 3 ´ 105-1 ´ 106 cells were injected into the right femurs of conditioned NOD/SCID mice, 6-8 weeks after, mice were sacrificed, cells were flushed from the femora, and then stained with CD33. The percentage of human GFP+, CD33+ cells in the bone marrow was determined by flow cytometry. The engraftment of transduced human AML cells into the bone marrow was assessed

87 by enumerating relative human cell engraftment as described in the calculation of engraftment potential section.

To asses effect of MMV on leukemia initiating cells in vivo, frozen aliquots of a primary AML patient sample were thawed, and pre-treated with MMV007285 or DMSO for 48-hours, in IMDM supplemented with 20% FCS, 100 units/mL penicillin and 100 µg/mL of streptomycin before transplantation. The day before transplantation, NOD/SCID mice conditioned as described above. Then, 2 ´ 106 viable treated cells were injected into the right femurs of 10 week-old female NOD/SCID mice. Six weeks after, mice were sacrificed, cells were flushed from the femora. Engraftment of human AML cells into the bone marrow of was assessed by enumerating the percentage of human hematopoietic myeloid cells with the cell surface signature of CD45+CD33+CD19− by flow cytometry using the BD Biosciences FACSCanto. Data were analyzed with FlowJo version 7.7.1 (TreeStar, OR, USA).

Whereas, to characterize the of MMV007285 on AML patient stem cell propagation properties in vivo, frozen aliquots of a Primary AML patient sample were thawed, counted and re-suspended in PBS, and 2 ´ 106 viable trypan blue negative cells were injected into the right femurs of NOD/SCID mice, conditioned as described above. Eleven days after transplantation mice were treated with MMV007285 (150mg/kg) by oral gavage or vehicle control (n = 9 mice vehicle control group, n = 10 mice MMV007285 group) once daily for 5 of 7 days for 5 weeks. Mice body weight was measured at the end of the experiment (41 days after AML patient sample transplantation). Afterwards mice were sacrificed, and AML cell engraftment bone marrow was measured as described above. Also, at sacrifice, peripheral blood was collected and ALP, AST, Bili, CK, Cr, levels were measured by Idexx Laboratories (Ontario, Canada).

The effects of MMV007285 on leukemia stem cell disease initiating properties were further characterized by injecting equal numbers of human leukemia cells isolated from the bone marrow of MMV007285 or DMSO treated mice, into secondary NOD/SCID mice that were conditioned as described above, but not treated with MMV007285. Six week later, human leukemia cell engraftment in the bone marrow was measured by flow cytometric analysis as described above.

2.2.45 Calculation of Engraftment Potential

Relative engraftment was measured as described in Pei et al., 2018. Using flow cytometry, we first measured % of GFP+ cells within the 7AAD- gate to determine the % of shRNA-expressing cells injected (%GFP-injected). After engraftment, we also measured % of viable GFP+, human CD33+ (primary AML samples) or % of GFP+, human CD45+ (8227) cells to determine the % of shRNA-expressing cells engrafted (%GFP-engrafted). For each experimental group, we first calculated engraftment value TAZ shRNA or control shRNA using the following formula: TAZ shRNA or control shRNA = %GFP-engrafted ÷ %GFP-injected for each mouse. This calculation yielded an array of engraftment values for both control shRNA and TAZ shRNA groups. To calculate the final relative engraftment potential score, each engraftment value was normalized to the mean engraftment of the control shRNA group. These scores were plotted to compare the relative engraftment potential between control and TAZ shRNA.

2.3 Quantification and Statistical Analysis

In all figures, n represents technical replicates and N represents independent biological replicates. Blinded analysis was implemented into the experimental design when possible. Data are mean ± standard error of the mean (SEM) or mean ± standard deviation (SD). Prism Graph Pad 6.0 was used to perform statistical analysis and data plotting unless otherwise specified in figure legends. A Two-Way ANOVA followed by a post-hoc Dunnett’s test was used to compare two or more variables between three groups, and a Two-Way ANOVA followed by a post-hoc Bonferroni test was used to compare two or more variables between two groups. A One-Way ANOVA followed by a post-hoc Dunnett’s test was used to compare one variable between three groups unless otherwise stated. Finally, an unpaired Student’s t-test was used to compare the mean between two groups. The level of significance is indicated as follows: *p<0.05, **p<0.01, ***p<0.001. A detailed description of the experimental setup and statistical analysis is in the figure legends.

2.4 Data Availability

AML gene expression after TAZ knock down has been deposited in the Gene Expression Omnibus database under the accession code GSE107045.

Chapter 3

Results 3.1 CRISPR Screen to Identify Mitochondrial Proteins Essential for Leukemia Cell Growth

3.1.1 CRISPR screen identifies TAZ as essential for the growth and viability of AML cells

CAS9-expressing human OCI-AML2 leukemia cells were transduced with a library of 91,320 sgRNAs in barcoded lentiviral vectors targeting 17,232 nuclear-encoded genes. Cells were harvested, genomic DNA was isolated, and the relative abundance of sgRNAs was determined by sequencing barcodes 17 days after transduction. sgRNAs able to reduce the viability or growth of OCI-AML2 cells were inferred to be those not represented in the final cell population. In analyzing the data, we focused on the sgRNAs targeting the 1049 nuclear-encoded mitochondrial proteins. Top hits included sgRNAs targeting dihydroorotate dehydrogenase (DHODH), B-cell lymphoma 2 (BCL2), and components of the mitochondrial ribosome, all of which have been previously shown to be necessary for the growth and viability of AML cells and stem cells (Table 11) (Skrtić et al., 2011; Sykes et al., 2016). We also identified the cardiolipin remodelling enzyme TAZ among the top 1% of mitochondrial hits (Figure 14A, Table 11). Of note, TAZ was also in the top 0.5% of hits in the whole genome. We also analyzed data from previously published CRISPR screens (Tzelepis et al., 2016) and identified TAZ as essential for the growth and viability of the leukemic cell lines OCI-AML2, OCI-AML3, MOLM-13, MV4- 11 and HL-60 (Figure 14B). Using individual sgRNA, we confirmed that knockout of TAZ reduced the growth of CAS9-OCI-AML2 cells, thus validating the findings from our screen (Figure 14C).

Figure 14. CRISPR screens identify TAZ as an essential gene for the growth and viability of AML cells.

(A) Results of a dropout screen in Cas9-OCI-AML2 cells. Positive-hits were identified at a False Discovery Rate (FDR) <5%.

(B) The rank of TAZ in screens of OCI-AML2, OCI-AML3, MOLM-13, MV4-11, and HL-60 cells from the published CRISPR dropout screens by Tzeplepis et al. (2016).

(C) Proliferation of CAS-9-OCI-AML2 cells after CRISPR mediated knockout of TAZ. The relative area under the curve (AUC) of viable cell counts over 12 days are shown (control sgRNA = 100%). Data are Mean ± SEM (N = 3). ****p≤0.0001 by one-way ANOVA and Dunnett's post hoc test 91

id neg|score neg|p-value neg|fdr Mitochondri a neg|rank

SLC25A1 4.39E-09 2.87E-07 0.000381 1

COX10** 1.57E-08 8.62E-07 0.000464 2

SARS2 2.61E-08 8.62E-07 0.000464 3

PPWD1 6.74E-08 1.44E-06 0.000495 4

DHODH** 9.55E-08 1.44E-06 0.000495 5

CCT7 1.19E-07 2.01E-06 0.000587 6

SHMT2 1.41E-07 2.58E-06 0.000601 7

KARS 2.22E-07 3.16E-06 0.000601 8

BCL2** 2.26E-07 3.16E-06 0.000601 9

MRPS24 2.34E-07 3.16E-06 0.000601 10

TAZ 2.51E-07 3.16E-06 0.000601 11

UQCRC2 3.66E-07 3.73E-06 0.000601 12

MRPS15 4.61E-07 3.73E-06 0.000601 13

POLRMT** 5.07E-07 4.31E-06 0.00064 14

IBA57 5.79E-07 4.88E-06 0.000679 15

ISCU 1.10E-06 9.48E-06 0.001011 16

AARS2 1.15E-06 9.48E-06 0.001011 17

FECH 1.18E-06 1.01E-05 0.001011 18

LIAS 1.39E-06 1.06E-05 0.001011 19

MTCH2 1.39E-06 1.06E-05 0.001011 20

Table 11. CRISPR Hits of Mitochondrial Proteins.

3.1.2 TAZ knockdown reduced the growth of leukemia cells

To understand how TAZ affects the growth of leukemia, OCI-AML2, TEX, U-937, and K562 leukemia cells were transduced with two independent shRNA targeting TAZ or control sequences in lentiviral vectors. Similar to knockout of TAZ by CRISPR, TAZ knockdown inhibited the growth of the tested leukemia cells (Figure 15A-D). TAZ knockdown induced cell cycle arrest but not apoptosis as measured by annexin V staining (16A-D).

Figure 15. Knockdown of TAZ reduces the growth of leukemia cells lines. (A-D) Proliferation and TAZ protein expression of OCI-AML2 (A), TEX cells (B), (C) K562 and (D) U937 cells after shRNA mediated TAZ knockdown. The relative AUC of viable cell counts over 12 days are shown (control shRNA = 100%). (A-B) Data are Mean ± SD (n = 2) of a representative experiment from 3 independent experiments, (C-D) Data are relative mean ± SEM (N = 4, K562; N = 3, U-937). **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001 by one-way ANOVA and Dunnett's post-hoc test 93

Figure 16. Growth analysis of leukemia cells after TAZ knockdown. (A) Cell viability of OCI-AML2 cells after TAZ knockdown. Data are mean ± SEM of three independent experiments. (B-D) Cell cycle status (B) (G0/G1)-S ratio (C), or (G0/G1)-G2 ratio (D), after TAZ knockdown. Data are mean ± SEM of three independent experiments. **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001 by two-way ANOVA (cell cycle status) or one-way ANOVA (all other graphs) and Dunett's post hoc test.

3.2 Role of TAZ in LSC Function

3.2.1 TAZ knockdown reduced the expression of genes associated with LSCs

LSCs initiate and drive the propagation of AML, so we assessed the effects of TAZ knockdown on the LSCs of AML. First, we analyzed global changes in gene expression by RNA-sequencing after TAZ knockdown in OCI-AML2 cells. The gene expression profile of TAZ knockdown AML cells was compared with genes that characterize primary AML stem cells (LSC+) and bulk cells (LSC-) (Ng et al., 2016). TAZ knockdown increased expression of genes associated with LSC- cells and decreased expression of genes associated with LSC+ cells (Figure 17A-B). As shown in Shlush et al., 2017 unsupervised clustering can be used to divide the TCGA AML cohort into two distinct clusters-one TCGA cluster is similar to stem/progenitor-like cells, and the other TCGA cluster is similar to more mature myeloid cells (Figure 18A). Per the LSC gene set enrichment analysis, TAZ knockdown increased the expression of genes associated with the TCGA myeloid like cluster and decreased the expression of genes in the TCGA stem/progenitor- like cluster (Figure 18B). Besides, the gene expression after TAZ knockdown was compared to 20 different sub-fractions of normal hematopoietic cells from 38 human samples (Novershtern et al., 2011). Genes associated with more differentiated myeloid populations such as granulocyte- neutrophils were increased, after TAZ knockdown (Figure 19A-C).

Figure 17. TAZ knockdown reduces LSC signature in OCI-AML2 cells.

(A) Heatmap of standardized Z score expression of the 500 most highly upregulated genes in OCI-AML2 cells after TAZ shRNA knockdown. Rows represent genes and columns represent LSC+ (pink bars) or LSC- (black bars) fractions.

(B) Gene set enrichment analysis (GSEA) of OCI-AML2 cells from (A). The normalized enrichment scores (NES), and false discovery rates (FDRs) are indicated in each GSEA plot.

Figure 18. Related to Figure 17. TAZ knockdown reduces stemness signature in OCI- AML2 cells.

(A) Unsupervised hierarchical clustering of gene expression data from the TCGA AML cohort (n = 179). Values above each branching represent percent approximate unbiased p values calculated by 1000 boostrap resampling using the R package pvclust. The most differentially expressed genes (top 500) genes between the two TCGA clusters were compared with Gene Expression Omnibus dataset GSE24759 (DMAP) populations. Each bar is a measure of the expression of these genes in each DMAP population compared to the mean of all DMAP populations (mean is equal to 0). Numbers beside bars indicate the percentage of time for which the observed value was better represented in a specific DMAP population than random values (equal number of randomly selected genes based on 1,000 trials).

(B) Gene set enrichment analysis (GSEA) measure of the TCGA stem-like or myeloid-like genes in the TAZ shRNA or control shRNA gene signatures. The normalized enrichment scores (NES), and false discovery rates (FDRs) are indicated in each plot.

Figure 19. Related to Figure 17. Knockdown of TAZ increases the expression of genes associated with granulocytic/neutrophilic populations.

(A) The most differentially expressed genes (top 500) between TAZ shRNA2 and control samples were compared with Gene Expression Omnibus dataset GSE24759 (DMAP) populations. Each bar is a measure of the expression of these genes in each DMAP population compared to the mean of all populations (mean is equal to 0). Numbers beside bars indicate the percentage of time for which the observed value was better represented in a specific DMAP than random values (equal number of randomly selected genes based on 1,000 trials).

(B-C) Proportions of control shRNA or TAZ-KD samples that possess gene expression profiles of undifferentiated HSC or mature myeloid cells, as determined by the perturbation (PERT) deconvolution analysis using gene expression data from normal hematopoietic subsets.

100

Cellular phenotype of AML cells after TAZ knockdown

Given the changes in gene expression, we characterized the cellular phenotype of OCI-AML2 cells after TAZ knockdown. Non-specific esterase (Figure 20A), a marker of monocytic differentiation in vitro, was increased, although no change in macrophage markers CD11b and CD14 were seen (Figure 20B-C). Increased differentiation was also observed when TAZ knockdown cells were grown in vivo, with increased expression of the myeloid granule protein, lysozyme (LYZ) (Sykes et al., 2016) (Figure 20D).

101

Figure 20. Knockdown of TAZ increases the expression of mature cell markers.

(A) Non-specific esterase (NSE) staining of OCI-AML2 cells after TAZ knockdown. Data are relative mean ± SEM (N = 4, control shRNA = 1.00 ABU). ****p ≤ 0.0001 by one-way ANOVA and Dunnett's post hoc test.

(B-C) CD11b (B) and CD14 (C) expression in OCI-AML2 cells after TAZ knockdown. Data are relative mean fluorescent intensity (MFI) ± SD (n = 3, control shRNA = 1.0).

(D) Lysozyme (LYZ) levels in OCI-AML2 cells transplanted into SCID mice after TAZ knockdown. Data represent relative mean ± SD (n = 3 mice/group, control shRNA = 1.0). *p ≤ 0.5 by Student’s t-test.

3.2.2 TAZ knockdown reduced the function of AML stem cells

We then evaluated the effects of TAZ knockdown on stem and progenitor cell functions both in vitro and in vivo. In methylcellulose assays, TAZ knockdown reduced the clonogenic growth of OCI-AML2 and TEX cells. The reduction in clonogenic growth persisted upon serial re-plating (Figure 21A-C). In xenotransplantation assays, TAZ knockdown in TEX cells reduced bone marrow engraftment in NOD/SCID-GF mice, suggesting that TAZ knockdown reduces AML propagation by reducing AML stemness both in vitro and in vivo (Fig 21D).

102

Figure 21. Knockdown of Tafazzin reduces AML stem cell function.

(A-B) Clonogenic growth of OCI-AML2 (A) or TEX (B) after TAZ knockdown. Data are relative mean ± SEM (N = 3, control shRNA = 100%). ****p ≤ 0.0001 by one-way ANOVA and Dunnett's post hoc test.

(C) Clonogenic growth of TEX cells after TAZ knockdown upon serial replating. Data are relative mean ± SEM of three independent experiments (control shRNA = 100%). ****p ≤ 0.0001 by Student’s t-test.

(D) CD45+ TEX cell engraftment after TAZ knockdown. Bar represents mean (n = 10 mice control shRNA group, n = 9 mice TAZ shRNA2 group). **p ≤ 0.01 by Student’s t-test.

103

3.3 Role of TAZ in Hematopoiesis

3.3.1 Levels of blood cells after TAZ knockdown

To assess the effects of TAZ inhibition on normal hematopoiesis in vivo, we analyzed doxycycline-inducible tafazzin-knockdown (iDOX-Taz-KD) mice. (Acehan et al., 2011; Soustek et al., 2011). 7.6-14.3 week old adult iDOX-Taz-KD were fed DOX to induce Taz knockdown. After 12-19.4 weeks of DOX induction, we measured levels of Taz and phospholipid levels. We observed reduced Taz protein levels and an expected increase of the monolysocardiolipin to cardiolipin (MLCL/cardiolipin) ratio in bone marrow mononuclear cells (Figure 22A). We then analyzed the hematopoietic system in these mice. Taz knockdown mice had normal levels of hemoglobin, white blood cells, neutrophils, and platelets (Figure 22C-F).

104

Figure 22. Taz Knockdown mice are viable with a normal blood count.

(A-B) TAZ expression (A) and the MLCL/Cardiolipin (B) ratio of 7.6-14.3 week old wild-type (WT) and iDox-Taz-shRNA transgenic mice treated with

doxycycline for 12.9-19.4 weeks to induce TAZ-knockdown. Data are mean ±

SEM (N = 3 mouse groups, n = 7 WT mice, and n = 6 Taz-KD mice). ****p≤0.0001 by Student’s t-test.

(C-F) Hemoglobin (C), white blood cells (D), neutrophils (E), and platelets (F) of WT and Taz-KD mice treated with doxycycline as in A. Data are mean ± SEM (N

= 6 mouse groups, n = 13 WT mice, and n = 14 Taz-KD mice)

3.3.2 Role of TAZ in hematopoietic stem and progenitor cells

We also enumerated the frequency of hematopoietic stem and progenitor cells by flow cytometry. Taz-knockdown mice had normal levels of LT-HSC, ST-HSC, MPP, MEP, CMP, and GrMP (Figure 23A-C, Table 12).

Figure 23. Taz Knockdown mice have normal frequencies of hematopoietic stem and progenitor cells.

(A-C) The frequency of (A) Lin- cKit+ Sca1+ (LSK), Lin- cKIT+ (LK), (B) multipotent progenitor cells (MPP, CD48+, CD150-), short-term hematopoietic stem cells (ST-HSC, CD48+, CD150+), long-term hematopoietic stem cells (LT-HSC, CD48-, CD150+), (C) common myeloid progenitor cells (CMP, CD34+, FcgR-), granulocyte-macrophage progenitor cells (GrMP, CD34+, FcgR+), megakaryocyte-erythrocyte progenitors (MEP, CD34-, FCgR-) of WT and Taz-KD mice treated with doxycycline as in figure 9. Data are mean ± SEM (N = 4 mouse groups, n = 10 WT mice, and n = 8 Taz-KD mice)

105

MOUSE TYPE MOUSE ID CELL NUMBER (1 ´ 105)

WT 100 58.2

114 75.9

115 51.9

121 39.1

124 41.8

130 50.8

131 44.5

132 36.7

133 31.6

134 31.9

Average 46.2

Standard Deviation 12.9

TAZ-KD 104 50.0

105 49.2

108 62.0

113 57.9

116 66.0

122 52.9

123 53.2

136 28.3

Average 52.4

Standard Deviation 11.4

Table 12. Mouse bone marrow mononuclear cell count

106

3.3.3 Role of TAZ in hematopoietic stem and progenitor cells during stress

To determine the role of TAZ in hematopoietic stress, WT or Taz-knockdown mice were treated with 200mg/kg of 5-FU, and changes in white blood cells, neutrophils, lymphocytes, erythrocytes and platelets were monitored (Figure 24A-F). WT and Taz-KD mice demonstrated reversible cytopenias after 5-FU treatment. Compared to controls, Taz-knockdown mice showed decreased numbers of ST-HSC and increased numbers of GrMP after 5-FU treatment (Figure 25A-D, Table 13). However, we observed little or no change in clonogenic growth in primary and serial re-plating experiments (Figure 25E). Thus, the lack of effect on stem cell function contrasts with the effects on AML cells.

107

Figure 24. Blood counts of WT and Taz-KD mice after the induction of hematopoietic stress.

(A) Schematic outlining the induction of hematopoietic stress by 5-FU.

WT and Taz-KD mice were treated with 200mg/kg of 5-FU i.p. Arrows indicate the days that peripheral blood of WT and Taz-KD mice were collected.

(B-F) White blood cells (B), neutrophils (C), lymphocytes (D), erythrocytes (E), and platelets (F) levels in the mice in A. Data are relative mean ± SEM of 2 independent mouse groups (n = 7 WT mice, and n = 6 Taz-KD mice, D0 = 100%).

108

Figure 25. Frequencies and function of hematopoietic stem and progenitor cells in Taz-KD mice.

(A) Schematic outlining the characterization of hematopoiesis is 5-fluorouracil (FU) treated Taz-KD mice.

(B-D) The frequency of (B) LSK, LK, MPP, (C) ST-HSC, LT-HSC, (D) CMP, GrMP, MEP in WT and Taz-KD 15 days after being injected with 5-FU. Both WT and Taz-KD mice were treated with doxycycline as in figure 11. Data are mean ± SEM (N = 2 mouse groups, n = 5 WT mice, and n = 6 Taz-KD mice). *p ≤ 0.05, **p ≤ 0.01 by two-way ANOVA and Bonferroni post hoc test.

(E) Clonogenic growth of hematopoietic cells in 5-FU treated Taz-KD mice upon serial replating. Data represent relative mean ± SEM of 2 independent mouse groups (n = 7 WT mice, and n = 5 Taz-KD mice, WT mice CFU-GM=100%).

109

Mouse Type Mouse ID Cell Number (1 ´ 105)

WT 154 7.2

156 10.0

195 22.0

197 35.0

198 33.0

154 7.2

Average 21.4

Standard Deviation 12.8

TAZ-KD 152 23.0

153 12.5

155 7.1

159 8.5

192 19.0

193 45.0

Average 19.2

Standard Deviation 6.8

Table 13. 5-FU treated Mouse bone marrow mononuclear cell count.

110

3.4 Mechanism by which TAZ knockdown Reduces AML Growth

3.4.1 The effect of TAZ knockdown on MLCL levels and CL composition

To understand the effects of TAZ knockdown on cardiolipin levels and composition in leukemia, we knocked down TAZ in OCI-AML2 cells and measured levels of MLCL and cardiolipin. As expected, TAZ knockdown increased MLCL/cardiolipin ratio (Figure 26A) in AML cells. TAZ- based cardiolipin remodelling also determines the final acyl composition of cardiolipin (Lu and Claypool, 2015; Lu et al., 2016; Paradies et al., 2014). Interestingly, TAZ knockdown AML cells had more cardiolipin acyl species with >5 double bonds (Figure 26B), as well as longer MLCL and cardiolipin acyl chain lengths (Figure 26C-D). Also, the predominant cardiolipin species in AML switched from 68:4 to 70:5 (Figure 26E-J). Cardiolipin binds to and activates caspase-8, which cleaves pro-apoptotic BID into t-bid, and facilitates extrinsic apoptosis (Gonzalvez et al., 2013; Gonzalvez et al., 2008). Mutation or deletion of TAZ renders cells resistant to extrinsic apoptosis. Similarly, TAZ knockdown rendered OCI-AML2 cells resistant to extrinsic apoptosis (Figure 26K). Thus, taken together, TAZ knockdown produced functionally significant reductions in levels of the phospholipid cardiolipin.

111

Figure 26. Knockdown of Tafazzin in OCI-AML2 cells reduces TAZ activity.

(A-H) The relative MLCL/Cardiolipin ratio (A), cardiolipin double bonds (B) MLCL given chain lengths (C), cardiolipin chain lengths (D), and the distribution of double bonds per acyl chain length (E-J). Data are mean ± SEM of three independent experiments. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001 by Student’s t-test (MLCL/Cardiolipin ratio) one-way ANOVA (64:3 graph), two-way ANOVA (all other graphs), and a post hoc Dunett's test.

(K) Cell viability of OCI-AML2 cell after TAZ knockdown when treated with cyclohexamide, and TRAIL for 16 hours. Data are relative mean ± SD of a representative experiment from 3 independent experiments (0ng/mL TRAIL = 100%).

112

3.4.2 Mitochondria structure and function after TAZ knockdown

Cardiolipin plays a role in forming the shape of the mitochondria (Dudek, 2017). Surprisingly, mitochondrial mass (Figure 27A), mitochondrial structure (Fig 27B), or mitochondrial length (Fig 27C) was not affected by TAZ knockdown.

Figure 27. Knockdown of Tafazzin in OCI-AML2 cells does not affect mitochondrial structure.

(A) Mitochondrial mass of OCI-AML2 cell after TAZ knockdown by the measurement of ND1 levels. Data represent mean ± SD (n = 3, Control shRNA = 1.0).

(B) Mitochondrial morphology of cells from (D). Composite of representative images are shown. Scale bar = 2µm or 500nM

113

Furthermore, cardiolipin is required for the stability, proper localization, and efficient function of respiratory chain enzymes (Paradies et al., 2014). However, we observed no decrease in basal oxygen consumption (Figure 28A), electron transport chain reserve capacity (Figure 28B), cellular ROS production (Figure 28C). It is noteworthy that, we did observe a decrease in glycolysis which is consistent with differentiation (Figure 28D) (Gu et al., 2016).

Figure 28. Knockdown of Tafazzin in OCI-AML2 cells does not affect mitochondrial function.

(A-B) Basal OCR (A) and reserve OCR (B) in OCI-AML2 cells after TAZ knockdown. Data are relative mean ± SD of a representative experiment from 3 independent experiments (Control shRNA = 100%). ****p ≤ 0.0001 by one-way ANOVA, and a Dunnett's post hoc test.

(C) Cellular ROS of OCI-AML2 cells after TAZ knockdown. Data are mean ± SD of a representative experiment from 3 independent experiments (Control shRNA = 100%).

(D) ECAR of cells from (A). Data are relative mean ± SD of a representative experiment from 3 independent experiments (Control shRNA=100%). *** p ≤ 0.001, ****p ≤ 0.0001 by one-way ANOVA, and a Dunnett's post hoc test 114

Thus, the change in cardiolipin levels after TAZ knockdown was not sufficient to impact mitochondrial structure or oxidative phosphorylation.

3.4.3 TAZ knockdown alters cellular phospholipids

Phospholipids are organized in a highly interconnected network, and perturbations in one species affect levels of other phospholipids (Koberlin et al., 2015). Therefore, we measured changes in cellular phospholipids by high-performance thin-layer chromatography (HPLC) after TAZ knockdown in OCI-AML2 and observed decreased levels of cardiolipin, as well as PE, increased levels of PS, but no change in PC, PI and SM (Figure 29A). Of note, the increase in PS was intracellular, and no increase in PS was detected on the cell surface (Figure 29B). We also evaluated changes in PS after TAZ knockdown in primary AML cells and the primary AML culture system 8227 (Lechman et al., 2016). TAZ knockdown increased intracellular PS in primary AML and 8227 cells (Fig 29C-E).

115

Figure 29. TAZ knockdown increases phosphatidylserine levels in leukemia cells.

(A) Composition of sphingomyelin (SM), phosphatidylcholine (PC), phosphatidylserine (PS), phosphatidylinositol (PI), phosphatidylethanolamine (PE), and cardiolipin (CL) in

OCI-AML2 cells after TAZ knockdown. Data are mean ± SEM (N = 3). *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001 by two-way ANOVA, and Dunnett's post-hoc test.

(B) Extracellular PS levels in OCI-AML2 cells following TAZ knockdown. Data are

mean ± SEM of 3 independent experiments.

SEM of two independent experiments. ****p ≤ 0.0001 Student’s t-test.

(D-E) Intracellular phosphatidylserine levels of 8227 (D), and primary AML cells (E).

Data are mean integrated PS staining intensity/DAPI+ cells ± SD (n = 8-12 images). *p ≤ 0.5, ***p ≤ 0.001 by Student’s t-test.

3.4.4 Functional effects of altered phospholipids

To investigate if the changes in PE and PS were functionally essential to explain the effects of TAZ knockdown on AML, we first tested whether the supplementation of PE or the PE substrate LPE could protect cells from TAZ knockdown. Despite increasing PE levels, PE or LPE supplementation did not protect AML cells from TAZ knockdown (Figure 30A-C).

Next, we tested whether supplementing cells with PS could increase differentiation and decrease stemness in AML. Supplementing AML cells with PS increased intracellular levels of PS and the PS/PE ratio (Figure 31A), and decreased AML cell growth (Figure 31B). Furthermore, pre- treatment of leukemia cells with PS reduced AML clonogenic growth (Figure 31C-D) and engraftment into mice (Figure 31E-G).

116

Figure 30. Effect of PE and LPE supplementation on TAZ knockdown OCI-AML2 cells.

(A) Phosphatidylethanolamine (PE) levels in OCI-AML2 cells after PE or lysophosphatidylethanolamine (LPE) supplementation. Data are mean ± SD (n = 3, Control = 1.0). *p < 0.05 by one-way ANOVA, and Dunnett's post hoc test.

(B-C) Proliferation TAZ knockdown OCI-AML cells after PE (E) or LPE (F) supplementation. The relative area under the curve (AUC) of viable cell counts 12 days after transduction is shown (Control shRNA = 100%). Data represents mean ± SD of representative experiment from 3 independent experiments. *p < 0.05, **p ≤ 0.01, ****p ≤ 0.001 by one-way ANOVA and Tukey’s post hoc test.

117

118

Figure 31. The increase PS levels after TAZ knockdown is functionally important in the reduction of AML stemness.

(A) PS/PE ratio in OCI-AML2 cells supplemented with 25µM PS or vehicle control. (Vehicle Control = 1.0). Data are mean ± SD (n = 4). ***p < 0.001 by Student’s t-test.

(B) Cell proliferation of OCI-AML2 cells supplemented with 25µM PS or vehicle control. The relative AUC of viable cell counts over 14 days are shown. Data are relative mean ± SD of a representative experiment from 3 independent experiments (Vehicle Control = 100%). **p < 0.01 by Student’s t-test.

(C-D) Clonogenic growth of OCI-AML2 cells (C), and primary AML cells (D), pre-treated with 25µM PS or vehicle control before being seeded in methylcellulose medium without PS. Data are relative mean ± SD (representative experiment from N = 3, OCI-AML2 cells; n = 2 primary AML cells, vehicle Control=100%). *p ≤ 0.05, **p ≤ 0.01 by Student’s t-test.

(E-F) Engraftment of TEX cells (I) or 8227 cells (J) treated with PS (25µM) in NOD/SCID-GF mice. Bar represents mean (n = 9-10 mice vehicle control group, n = 9-10 mice). **p ≤ 0.01 by Student’s t-test.

(K) Engraftment of primary AML cells treated with PS (75µM) or vehicle controls in NOD/SCID mice. Bar represents mean. (n = 10 mice vehicle control group, n = 10 mice PS group). ***p ≤ 0.001 by Student’s t-test.

119

3.4.5 Effects of TAZ knockdown on phosphatidylserine decarboxylase (PISD)

Phosphatidylserine decarboxylase (PISD) converts PS to PE in the inner mitochondrial membrane (Figure 32A) and we discovered that recombinant PISD binds preferentially to cardiolipin and the cardiolipin moiety PG (Figure 32B). Knockdown of TAZ decreased levels of PISD protein but not mRNA (Figure 32C) and did not change the levels of the PS synthesis enzymes PSS1 and PSS2 (Figure 32D).

To further characterize the effects of increasing PS on AML stemness and differentiation we evaluated the effects of PISD inhibition in AML. First, we knocked out PISD in CAS9-OCI- AML2 cells using sgRNA sequences in lentiviral vectors. Knockout of PISD was confirmed by immunoblotting (Figure 33A). The knockout of PISD increased the PS/PE ratio (Figure 33B), reduced the growth (Figure 33C), clonogenic growth (Figure 33D-E), and tumor formation capacity of CAS9-OCI-AML2 cells (Figure 33F), without decreasing basal oxygen consumption or electron-transport chain reserve capacity in OCI-AML2 cells (Figure 33G-H). Furthermore, the overexpression of PISD rescued the differentiation phenotype of TAZ knockdown AML cells (Figure 33I-J).

120

Figure 32. TAZ knockdown reduces phosphatidylserine decarboxylase (PISD) levels.

(A) The role of phosphatidylserine decarboxylase (PISD), where PISD decarboxylates phosphatidylserine (PS) to produce phosphatidylethanolamine (PE).

(B) Immunoblots measuring recombinant PISD protein bound to lipids. Data from 3 independent experiments are shown.

(C) Protein and mRNA levels of the PE synthesizing enzyme PISD in OCI-AML2 cells after TAZ knockdown. An immunoblot from three independent experiments is shown. qRT-PCR data are mean ± SD (n = 3, control shRNA = 1.0).

(D) Protein levels of the PS synthesizing enzymes PS synthase 1 (PSS1) and PS synthase 2 (PSS2) in OCI-AML2 cells after TAZ knockdown. An immunoblot from four independent experiments is shown.

121

Figure 33. PISD decrease after TAZ knockdown is functionally important in the decrease in AML growth and stemness.

( A-D) PISD protein expression (A) and PS/PE ratio (B) of Cas9-OCI-AML2 cells after CRISPR mediated PISD knockout. Data are mean ± SD (n = 4). ***p < 0.001 by Student’s t- test. (C) Proliferation of Cas9-OCI-AML2 cells after CRISPR mediated PISD knockout. The relative AUC of viable cell counts over 15 days are shown. Data are mean ± SD, of a representative experiment from three independent experiments (Control sgRNA = 100%).

**p < 0.001 by Student’s t-test. (D) Clonogenic growth of Cas9-OCI-AML2 cells after CRISPR mediated PISD knockout. Data are relative mean ± SD of a representative experiment from three independent experiments (Control sgRNA = 100%). **p ≤ 0.01 by Student’s t-test.

(E) Clonogenic growth of Cas9-OCI-AML2 after PISD knockout upon serial replating. Data are relative mean ± SEM of 2 independent experiments. (Control sgRNA = 100%). **p ≤

0.01 ***p ≤ 0.001, ****p ≤ 0.001 by Student’s t-test.

122

(F) Tumor growth of Cas9-OCI-AML2 cells after PISD knockout. Data represent mean ±

SD (n = 8 mice/group). ****p ≤ 0.0001 by two-way ANOVA and Bonferroni’s post hoc test.

(G -H) Basal OCR (G) and reserve OCR (H) in of Cas9-OCI-AML2 after PISD knockout. Data are relative mean ± SD of a representative experiment from 2 independent experiments

(Control shRNA = 1.0).

(I-J) PISD mRNA expression (I) and NSE staining (J) of TAZ knockdown OCI-AML2 cells after PISD over expression. In (I) Data are mean ± SD (n = 3). ***p ≤ 0.001, ****p ≤ 0.0001 by one-way ANOVA and Tukey’s post hoc test. In (J) data represent relative mean ± SD of a representative experiment from 2 independent experiments (control shRNA = 100%). ****p ≤ 0.0001 by one-way ANOVA and Tukey’s post hoc test.

3.4.6 Anti-leukemic activity of a PISD inhibitor

As a chemical approach to inhibit PISD and increase intracellular levels of PS, we tested the reported inhibitor of Plasmodium falciparum PISD, MMV007285 (Figure 34A) (Choi et al., 2016; Lucantoni et al., 2013). Consistent with its effects as a PISD inhibitor, MMV007285 increased the PS/PE ratio in OCI-AML2 cells (Figure 34B). Concentrations of MMV007285 that increased PS also decreased growth (Figure 34C) and increased differentiation as evidenced by increase in non-specific esterase staining and CD11b expression (Figure 34D-E).

123

Figure 34. The PISD Inhibitor MMV007285 reduces AML growth and stemness.

A) Chemical structure of MMV007285.

(B) PS/PE ratio in OCI-AML2 cells treated with MMV007285 (10µM) or vehicle controls. Data are relative mean ± SEM of three independent experiments (0µM = 1.00). **p ≤ 0.01 by Student’s t-test.

(C) Cell proliferation of OCI-AML2 cells treated with increasing concentrations of MMV007285 or vehicle control. Relative AUC of viable cell counts over 14 days are shown. Data are the relative mean ± SD (0µM = 100%) of one of three independent experiments. ***p ≤ 0.001, ****p ≤ 0.0001 by one-way ANOVA and Dunnett’s post hoc test.

(D) NSE staining of OCI-AML2 cells incubated with MMV007285 or the vehicle control. Data represent relative mean ± SEM of three independent experiments (0µM = 1ABU). **p ≤ 0.01, ****p ≤ 0.0001 by Student's t-test.

(E) CD11b expression of OCI-AML2 cells incubated with MMV007285 or vehicle control. Data represent relative mean fluorescent intensity (MFI) ± SEM of two independent experiments (0µM=1.0). ***p ≤ 0.001 by Student’s t-test. 124

3.4.7 Increasing intracellular phosphatidylserine increases toll-like receptor activity

To investigate how TAZ knockdown and increases in PS reduce AML stemness and increase differentiation we characterized the genes that changed after the knockdown of TAZ in OCI- AML2 cells. Among the genes upregulated after TAZ knockdown, we observed a strong enrichment in TLR and other immune activation pathways (FDR ≤ 0.01, Figure 35A). TAZ knockdown increased TLR4 and TLR8 expression as well as expression of their downstream cytokine mediators IL6 and INFb (Figure 35B). Similar changes in TLR4 and TLR8 signaling were observed after inhibiting PISD at time-points before AML differentiation (Figure 35C).

Treatment of AML cells with the TLR8 agonist, CL075 increased differentiation and decreased growth of AML cells similar to TAZ knockdown (Figure 26A-C). Moreover, the combination of the TLR8 agonist CL075 with the PISD inhibitor MMV007285 increased expression of TLR8 more than either drug alone, and synergistically decreased growth of AML cells (Figure 36D-E). Collectively, these data suggest that the observed TLR activation is functionally important in promoting AML differentiation.

125

Figure 35. TAZ-KD and PISD Inhibitor MMV007285 activates TLR signaling.

(A) Pathway enrichment analysis using GSEA, and visualization using Cytoscape EnrichmentMap in OCI-AML2 cells after TAZ knockdown. Circles (nodes) represent pathways, clusters represent biological processes, and lines connect pathways with common genes. Red nodes represent pathways that are upregulated, and blue nodes represent pathways that are down-regulated in TAZ shRNA samples compared to control (FDR ≤ 0.01). Red arrows indicate up-regulated immune pathways.

(B) qRT-PCR of toll-like receptor 4 (TLR4), toll-like receptor 8 (TLR8), interleukin-6 (IL6), interferon-b (IFNβ) in OCI-AML2 cells after TAZ knockdown. Data are relative mean ± SD (n=2- 3, control shRNA = 1.0). *p ≤ 0.05, **p ≤ 0.01, ****p ≤ 0.0001 by one-way ANOVA and Dunnett’s post-hoc test.

(C) qRT-PCR of TLR4, TLR8, IL6, and IFNb in OCI-AML2 cells treated with MMV007285 or vehicle control. Data are mean ± SD (n = 2-3, 0µM = 1.0). *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, by Student’s t-test 126

Figure 36. TLR8 agonist CL075 reduces stemness in OCI-AMl2 cells and synergizes with MMV007285.

(A) Proliferation of OCI-AML2 cells after treatment with increasing concentrations of

CL075 or vehicle controls. The relative AUC of viable cell counts over 14 days are shown

(control shRNA = 100%). Data are the relative mea n ± SEM (0µM = 100%) of two independent experiments. ***p≤0.001, ****p≤0.0001 by one-way ANOVA and Dunnett’s post hoc test.

(B-C) CD11b expression in OCI-AML2 cells seven (B) or fourteen (C) days after treatment with CL075. Data are relative mean fluorescent intensity (MFI) ± SEM of 2-3 independent experiments (0µM = 1.0). **p≤0.01, ***p≤0.001, **** p≤0.0001 by one-way ANOVO and Dunett’s post hoc test.

(D) qRT-PCR of toll-like receptor 8 (TLR8) of OCI-AML2 cells were treated with 6 or 24 hours after treatment with MMV007285 alone, CL075 alone or MMV007285 + CL075. Data represent mean ± SD (n = 3, Control, 0µM MMV + 0µg/mL CL075 = 1.0). **** p≤0.0001 by two-way ANOVA and Tukey’s post hoc test.

(E) Excess over bliss additivism score of the combination of MMV007285 and CL075 on cell viability. A representative experiment of two independent experiments is shown. 127

3.5 Pre-clinical anti-leukemic activity of strategies that increase PS

3.5.1 Effect of PISD inhibition on AML stem cell function in vitro

The effects of inhibiting PISD and increasing PS on stemness in primary AML cells was evaluated by testing MMV007285 in the 8227 primary AML culture system. 8227 cells are arranged in a hierarchy of bulk and stem cells with the stem cells enriched in the CD34+38- fraction. Treatment with MMV007285 decreased the abundance of the CD34+38- stem cells (Figure 37A).

We also examined the effects of MMV007285 on the clonogenic growth of primary AML (N = 3) and normal hematopoietic cell (N = 3) samples. MMV007285 increased PS and inhibited the clonogenic growth of primary AML cells (Figure 37B-C) but did not affect the clonogenic growth of normal hematopoietic cells (Figure 37D). MMV007285 also did not alter the clonogenic growth of normal murine hematopoietic cells under conditions of hematopoietic stress (Figure 37E).

3.5.2 Effect of TAZ knockdown on AML stem cell function in vivo

We knocked down TAZ in 8227 cells as well as primary AML samples (N = 4) and injected control and knockdown cells into the femurs of immunodeficient mice. TAZ knockdown reduced the engraftment potential (Pei et al., 2018) of these primary cells into the mouse marrow (Figure 38A-E, 39A-E).

128

Figure 37. The PISD Inhibitor MMV007285 demonstrates specific anti-leukemic activity.

(A) Frequency of viable CD34+,38- in MMV007285 and vehicle treated samples. Data represent relative mean ± SEM of three independent experiments (0µM = 100%). **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001 by one-way ANOVA and Dunnett’s post-hoc test.

(B) Intracellular PS levels of primary AML cells treated with MMV007285 or the vehicle control. Data represent relative MFI ± SD (n = 3, Control = 100%). ***p ≤ 0.001 by Student’s t-test.

(C) Clonogenic growth of primary AML cells treated with MMV007285 or vehicle control for 48 hours. Data represent mean ± SEM of three independent experiments (0µM = 100%). ****p ≤ 0.0001 by a Student’s t-test. 129

(D) Burst forming unit-erythroid (BFU-E) and colony forming unit-granulocyte/monocyte

(CFU-GM) of normal hematopoietic cells treated with MMV007285 or vehicle control. Data represent mean ± SEM of three independent experiments (0µM = 100%).

(E) Clonogenic growth of bone marrow cells from 21-29 week C567BL/6J mice injected with 200mg/kg of 5-fluorouracil (FU) and treated with MMV007285 or vehicle control. Data represent mean ± SEM of two independent experiments (0µM = 100%).

Figure 38 TAZ knockdown reduces stem cell function of primary AML patient samples.

(A) Engraftment of 8227 cells in NOD/SCID-GF mice after TAZ knockdown. Bar represents mean engraftment potential (n = 6 mice/group, control shRNA = 100%), **p ≤ 0.01 by Student’s t-test.

(B-E) Engraftment of primary AML cells from patients from patients 100565 (B), 120541 (C), 120287 (D), and 120860 (E) in NOD/SCID mice after TAZ knockdown. Bar represents mean engraftment potential (n = 6-8 mice/group, control shRNA = 100%), **p ≤ 0.01, **** p ≤ 0.0001 by Student’s t-test.

130

Figure 39. Related to Figure 38 TAZ knockdown reduces stem cell function of primary AML patient samples.

(A-E) Raw data of the engraftment experiment performed in 8227 (A), and AML patient sample cells (B-E) evaluating the impact of TAZ loss on engraftment potential. The y-axis (%GFP+ cells in viable cells) indicates % of shRNA-positive cells in injected or engrafted human cells. In the engrafted group each dot represents an individual mouse, lines represent mean. This figure is related to the normalized relative engraftment potential data presented in Figure 38A-E.

131

3.5.3 Stability, anti-leukemic activity, and toxicity of MMV007285 in vivo

Our pharmacokinetic studies suggested that MMV007285 has a short half-life in vivo (Figure 40A). In spite of this, we still evaluated the anti-leukemic efficacy and toxicity of systemic treatment with MMV007285 in mouse models of leukemia. SCID mice xenografted with OCI- AML2 cells were treated with MMV007285 or vehicle control. Treatment with MMV007285 delayed the growth of OCI-AML2 cells in mice (Figure 40B) without a more than 10% reduction in body weight (Figure 40C) or alterations in serum chemistry measuring liver, muscle or renal toxicity (Figure 40D-H).

3.5.4 Effect of MMV007285 on stem cell function of AML patient samples in vivo

Primary AML cells were first treated with the PISD inhibitor MMV007285 or vehicle control for 48-hours and injected an equal number of viable cells into sub-lethally irradiated NOD-SCID mice. Primary AML cells pre-treated with MMV007285 had significantly lower levels of leukemic engraftment compared to primary AML cells pre-treated with the vehicle control (Figure 41A), suggesting that pre-treatment with MMV007285 was sufficient to reduce the leukemia-initiating cell frequency in AML.

We also treated mice engrafted with primary AML cells with MMV007285. AML cells were injected intrafemorally into NOD/SCID mice. Eleven days after injection mice were treated with MMV007285 or vehicle control for 4 weeks. After treatment, levels of AML engraftment in the marrow of the mice was determined by flow cytometry. Treatment of mice with MMV007285 reduced levels of leukemic engraftment compared to controls with no evidence of toxicity (Figure 41B-H). We then harvested leukemia cells from the marrow of primary mice and injected these cells into secondary mice. The leukemia grafts from MMV00285 treated mice demonstrated an 80% reduction in engraftment compared to vehicle-treated controls, indicating that MMV007285 inhibits both leukemia propagation and targets leukemia stem cells (Figure 41I).

132

Figure 40. MMV007285 reduced leukemia burden in xenograft models of human leukemia.

(A) Plasma concentration of intact MM007285 in SCID mice treated with a single dose of

MMV007285 (100mg/kg) dissolved in 5% DMSO, 47.5% PEG400, and 47.5% H20 containing 10% Tween80, via oral gavage. Data represent mean ± SD (n = 3 mice).

(B) Tumor growth of OCI-AML2 cells xenografts into SCID mice treated with 300 mg/kg of MMV007285 or vehicle control twice daily for 5 days/week orally (n = 10 mice/group). Data represent mean ± SD, **p≤0.01. ****p≤0.0001 by two-way ANOVA and Bonferroni’s post hoc test.

(C-H) Mice body weights (C), biochemical markers of liver (alkaline phosphatase D, aspartate transaminase E, bilirubin F), muscle (creatine kinase, G), and renal (creatinine, H) toxicity of SCID mice xenografted with OCI-AML2 cells MMV007285 or vehicle control. Data represent mean ± SD (n = 10 mice per group for body weights, and n = 4 mice/group for biochemical markers).

133

Figure 41. MMV007285 reduced the function of primary AML patient samples.

(A) Primary AML cell engraftment in NOD/SCID after pre-treatment in vitro of cells for 48 hours with 12.5uM of MMV00728. Data represent mean ± SD (n = 10 mice/group). ***p≤0.01 by Student’s t-test.

(B) Engraftment of primary AML cells in NOD/SCID mice treated once daily with MMV007285 (150 mg/kg orally) for 5 days/week Bar represents mean (n = 9 control group, n = 10 MMV007285 group). ***p≤0.001, as determined by Student’s t-test

(C-H) Mice body weight (C), biochemical markers of liver (alkaline phosphatase D, aspartate transaminase E, bilirubin F), muscle (creatine kinase, G), and renal (creatinine, H) toxicity of NOD/SCID mice injected with primary AML patient samples treated with MMV007285 or vehicle control. Data represent mean ± SD (n = 9-10 mice/group for body weights, n = 4 mice/group for biochemical markers). **p<0.01, as determined by Student’s t-test

(I) Secondary engraftment of primary AML cells from (I). Bar represents mean (n = 7 mice/group). ***p≤0.001, as determined by Student’s t-test.

Chapter 4 Discussion 4.1 Summary

Our main objective was to identify novel targets in the mitochondrial proteome whose knockout reduced AML growth and viability. Through a CRISPR screen we tested genes necessary for the growth and viability of AML cells, and identified TAZ as a top hit. TAZ is a transacylase responsible for the generation of mature cardiolipin (Lu and Claypool, 2015; Lu et al., 2016). Using individual sgRNA, we confirmed that the knockout of TAZ reduced the growth of CAS9- OCI-AML2 cells by >70%. We also knocked down TAZ with 2 independent shRNA and demonstrated reductions in growth and viability in a panel of AML cells: OCI-AML2, TEX, K562, and U937. TAZ knockdown also reduced the engraftment of TEX leukemia cells, and primary leukemia patient samples in vivo, indicating that TAZ knockdown targets leukemia initiating cells. In contrast, knockdown of TAZ in mouse models did not impair normal hematopoiesis nor reduced the abundance of hematopoietic stem cells, although a subtle decrease in ST-HSCs in conditions of stress was observed, which might explain transient episodes of neutropenia seen in Barth’s syndrome, a congenital condition associated with X-linked TAZ mutations (Seneviratne et al., 2017).

As previously reported in other systems, TAZ knockdown in both AML and normal hematopoietic cells increased the substrate (MLCL) to product (cardiolipin) ratio of TAZ. However, knockdown of TAZ did not alter mitochondrial structure, or reduce basal oxygen consumption or respiratory reserve capacity.

Interestingly, the inhibition of TAZ reduced the levels of PISD, the enzyme responsible for converting PS to PE, reduced levels of PE, and increased levels of PS. Both the reduction of PISD and the increase in PS were sufficient to reduce AML growth and stemness, implying that TAZ suppresses AML stemness by reducing PISD and increasing PS (Figure 42).

We also discovered that TAZ knockdown or PS increase up-regulated genes in the TLR pathway, and that the activation of the TLR pathway was functionally important to explain the effects of TAZ knockdown on AML stemness and differentiation (Seneviratne et al., 2019; Seneviratne, 2019).

134 135

Finally, we evaluated the pre-clinical activity of MMV007285, a previously published PISD inhibitor. MMV007285 reduced AML stemness similar to that of TAZ knockdown. In mouse models of leukemia, MMV007285 reduced AML disease burden without toxicity to normal tissues, and targeted AML stem cells as evidence by reductions in AML engraftment in secondary transplant models.

Thus, in this study we identified a new mechanism by which mitochondrial pathways regulate cellular phospholipids and control AML cell fate and stemness. Moreover, PISD inhibition may be a novel therapeutic strategy to reduce AML stemness.

Figure 42. Summary of Major Conclusions. We performed a CRISPR screen and identified tafazzin (TAZ) as an important for the growth of leukemia cells. The inhibition of TAZ specifically reduced the stemness of leukemia cells by increasing phosphatidylserine levels and activating toll-like receptor signaling (TLR).

136

4.2 TAZ a Novel Mitochondrial a Therapeutic Target for AML

DNA-targeted chemotherapy such as cytarabine, daunorubicin, idarubicin and HCT form the back backbone of current AML therapy. The long-term survival rates of AML patients under 60 years, treated with a combination of these regimens are 25-50%, and only 5-15% in older patients (Gao and Estey, 2015). It is increasingly clear that these classical therapies cause genetic damage to leukemia cells, which contributes to the development of resistance (Gao and Estey, 2015). Moreover, 90% of the AML patients that survive HCT have more than one chronic condition, that negatively impacts their morbidity and mortality (Hilgendorf et al., 2015; Inamoto and Lee, 2017). In an attempt to develop a new therapeutic strategy for older AML patients, who have a particularly poor prognosis, DiNardo et al., conducted a phase I clinical trial that combined venetoclax with hypomethylating agents azacitidine and decitabine in newly diagnosed AML patients, older than 65 years (DiNardo et al., 2019; DiNardo et al., 2018). Four- hundred, 800, and 1,200 mg of venetoclax was assessed, and the patients were divided into subgroups based on the hypomethylating agent received. Data from 145 patients was analyzed in the study, 49% of which had poor-risk cytogenetics. The combination of 400 mg venetoclax and azacitidine group showed the most favourable results with a response rate of 76%. The duration of remission was 17.5 months, and the median overall survival had not been reached. Most notably, subgroup analysis revealed that patients with poor risk cytogenetics demonstrated a good response rate. The findings from this study has led to an accelerated FDA approval of venetoclax and a hypomethylating agent for elderly de novo AML patients (McBride et al., 2019).

Mechanistic studies that have attempted to understand the role of BCL-2 in AML suggest that venetoclax and azacitidine reduced AML disease burden by affecting cell metabolism. Langadinou et al., 2013 was the first to demonstrate BCL-2’s unique role in cell metabolism. Oxidative phosphorylation of LSCs was shown to be dependent on BCL-2. Consequently, the inhibition of BCL-2 in LSCs reduced ATP levels, as well as AML initiation and propagation in vitro, as well as in AML mouse models. Subsequent work by Pollyea et al., 2018 and Jones et al., 2019 found that the venetoclax and azacitidine combination reduced AML disease burden by inhibiting complex II activity, further supporting BCL-2’s role in oxidative phosphorylation.

These studies also highlight that the disease initiation and propagation functions of AML stem cells are dependent on oxidative phosphorylation, which is in contrast to normal HSCs. In this

137 background, we wanted to identify additional biological pathways in the mitochondria that were necessary for the growth and viability of AML cells and stem cells. To that end, we performed a genome-wide clustered regularly interspaced palindromic repeats (CRISPR) screen to identify genes whose knockout reduced AML growth. We focused our analysis on the mitochondrial proteins encoded by nuclear DNA, translated in the cytoplasm and imported into the mitochondria. From this screen, we identified the mitochondrial transacylase Tafazzin (TAZ) as a top hit (Seneviratne et al., 2019; Seneviratne, 2019).

4.3 Altered Cardiolipin does not Impair Mitochondria Structure or Function

Cardiolipin is a mitochondrial phospholipid, in the inner mitochondrial membrane that contributes to 20% of the mitochondrial lipidome. Structurally, cardiolipin is composed of a glycerol head and two phosphatidylglyceride backbones. It has a total of four fatty acid chains that differ in length and saturation (Dudek, 2017). The metabolism and energy requirements of the tissue determine the specific fatty acid composition of cardiolipin (Paradies et al., 2014). For instance, in heavily respiring tissue like the mammalian heart and skeletal muscle linoleic acid is the predominant fatty acid in cardiolipin, suggesting that cardiolipin influences both mitochondrial structure and function.

Mitochondrial fusion and fission work together to maintain the reticular network of the mitochondria. Fusion is primarily mediated by MFN1, MFN2, and OPA1. Whereas, DRP1 is the predominant protein that mediates fission. Studies have implicated that cardiolipin is required for fusion and fission processes (Dudek, 2017; Sesaki and Jensen, 1999). The induction GTPase activity of OPA1, which is necessary for mitochondrial fusion depends on cardiolipin (DeVay et al., 2009; Dudek, 2017; Meglei and McQuibban, 2009). The IMS facing domain of OPA1 contains a cardiolipin binding site. It was proposed that after the fusion of the outer mitochondrial membrane, the cardiolipin binding domain of OPA1 allows OPA1 to interact with cardiolipin of the inner membrane of the second mitochondria to induce inner mitochondrial membrane fusion (Ban et al., 2017; Dudek, 2017). Mitochondrial fission on the other hand, requires the activation of the GTPase of DRP1 which mediates the constriction of the outer mitochondrial membrane. Cardiolipin has a strong affinity to DRP1 and enhances its GTPase activity (Bustillo-Zabalbeitia et al., 2014; Stepanyants et al., 2015).

138

Given the role of cardiolipin in mitochondrial fusion and fission, cardiolipin deficiency has been shown to alter the mitochondrial structure of drosophila and human lymphoblast (Acehan et al., 2007; Xu et al., 2006a). We characterized the mitochondrial cristae of TAZ knockdown leukemia cells by electron microscopy. We also measured the aspect ratio of the mitochondria using fluorescent microscopy to characterize mitochondrial fusion and fission after tafazzin knockdown. The mitochondrial aspect ratio is a measure of mitochondrial length and form. A high aspect ratio indicates increased mitochondrial length and branching, a characteristic of more mitochondrial fusion. Whereas, a decrease in this value indicates mitochondrial fragmentation, a characteristic of more mitochondrial fission. We quantified the aspect ratio of 3000 mitochondria in both the control and tafazzin knockdown samples. These studies revealed that tafazzin knockdown did not alter mitochondrial morphology although it decreased cardiolipin levels.

Cardiolipin has also been shown to play a role in oxidative phosphorylation (Dudek et al., 2013; Dudek et al., 2016; Huang et al., 2015; Jussupow et al., 2019). Structural analysis of electron transport chain complexes found cardiolipin binding sites on complex I, III and IV as well as 200 cardiolipin molecules associated with mitochondrial supercomplexes. The formation of mitochondrial supercomplexes increases the efficiency of electron transfer between respiratory chain components by substrate channeling mechanisms, thereby maximizing oxidative phosphorylation (Dudek, 2017). Moreover, mitochondrial super complexes minimize the distance that mobile electron carriers have to travel reducing the risk of reactive oxygen species generation. The analysis of cardiolipin deficient mitochondria reveled the remodeling of supercomplexes from a higher to lower molecular weight complexes (Dudek et al., 2013; Dudek et al., 2016; Huang et al., 2015). The transition to lower molecular weight complexes was associated with lower oxidative phosphorylation and an increase in the generation of reactive oxygen species. However, our data demonstrate that the cardiolipin decrease after TAZ knockdown did not increase reactive oxygen species or reduce oxidative phosphorylation in leukemia cells.

Similar to other biological systems TAZ is responsible for altering the acyl composition of nascent cardiolipin to produce mature cardiolipin. (Lu and Claypool, 2015). Interestingly, both nascent and mature cardiolipin can support oxidative phosphorylation (Baile et al., 2014). The remodelling intermediate MLCL is a byproduct of this reaction, and an optimal MLCL cardiolipin ratio, which is normally very low, is essential for the stability of respiratory chain supercomplexes, (Baile et al., 2014; Cheneval et al., 1983; Claypool and Koehler, 2012;

139

Claypool et al., 2006; Claypool et al., 2008; Eble et al., 1990; Lu and Claypool, 2015; Lu et al., 2016; Paradies et al., 2014; Shinzawa-Itoh et al., 2007). We showed that knockdown of TAZ in AML cells increased MLCL levels, and decreased cardiolipin levels, consistent with reductions in TAZ function. However, the increase in MLCL and reductions in cardiolipin after inhibiting TAZ were not sufficient to impair mitochondrial structure or function, in keeping with the ability of nascent cardiolipin to support respiratory chain function. Alternatively, monolysocardiolipin acyltransferase 1 (MLCLAT1) and acyl-CoA:lysocardiolipin acyltransferase-1 (ALCAT1) also contribute to cardiolipin remodelling, so the redundancy in the cardiolipin acyl chain remodelling pathway potentially maintained sufficient levels of cardiolipin to preserve both mitochondrial structure and function (Claypool and Koehler, 2012; Taylor and Hatch, 2003, 2009; Taylor et al., 2012; Xu et al., 2006b).

4.4 Neutropenia of Barth Syndrome

Barth syndrome is an inherited mitochondrial disease due to inactivating mutation in the X- linked TAZ gene. Barth et al., first described this disease in 1980 as affecting cardiac muscle, skeletal muscle and neutrophil leukocytes. Cardiomyopathy, skeletal myopathy and neutropenia are still the cardinal clinical features of Barth syndrome. It is estimated that the incidence of Barth syndrome is ~1/300,000-400,000 live birth, with fewer that 500 individuals worldwide. Of note, there is no racial or ethnic predilection. The initial clinical presentation of the majority of patients with Barth syndrome is dilated cardiomyopathy with a component of left ventricular noncompaction. In addition, hypertrophic cardiomyopathy has also been described. Skeletal muscle fatigue and myopathy are also important features of Barth syndrome. Decreased exercise endurance, weakness of proximal leg muscles, and overall lower activity are demonstrated by affected individuals. Both skeletal muscle weakness and fatigue are independent of cardiac dysfunction. Neutropenia yet another clinical feature of Barth syndrome that can present as severe chronic neutropenia, cyclic neutropenia, or intermittent/non-cyclical neutropenia. The low neutrophil counts increase the risk of affected individuals developing bacterial infections that range from mouth ulcers and gingival inflammation, to sepsis and multi-organ system failure. Bacterial infections that develop as a result of neutropenia are the primary cause of death in Barth syndrome patients (Acehan et al., 2011; Lu and Claypool, 2015; Lu et al., 2016). Currently, there is no specific therapy for Barth syndrome, but instead treatments are directed towards individual symptoms including medical treatment for cardiac failure and granulocyte-

140 colony stimulating factor therapy for neutropenia. Cardiac transplant is performed in the case of intractable heart failure (Anzmann, 2019; Clarke et al., 2013).

Mitochondrial abnormalities caused by lower cardiolipin levels have been implicated in the heart and skeletal muscle defects observed in Barth Syndrome (Acehan et al., 2011). Studies have shown that the neutropenia in Barth syndrome is due to the dissipation of the mitochondrial membrane potential, aberrant release of cytochrome c, activation of caspase-8, and accelerated apoptosis in myeloid cells (Makaryan et al., 2012). Furthermore, the bone marrow aspirates of Barth Syndrome patients show that myelocytes (immediate precursors of neutrophils) are arrested in maturation. However, the role of TAZ in HSC cell self-renewal and differentiation, as well as in differentiation of HPCs and myeloid precursors during development is unknown.

To gain insight into the role of Taz in hematopoiesis, we utilized the doxycycline inducible Taz- KD mouse model (Acehan et al., 2011). In this model Taz shRNA is under the control of a tetracycline operator. Under basal conditions, the tetracycline repressor binds the tetracycline operator inhibiting the expression of Taz shRNA. When mice are fed the tetracycline derivative, doxycycline it binds to the tetracycline repressor, changing its confirmation, resulting in the removal of the teracycline repressor from the tetracycline operator and the subsequent expression of Taz shRNA. Taz shRNA expression in this manner, was sufficient to recapitulate many of the cardiac phenotypes in Barth syndrome such as left ventricular dilation, reduced ventricular mass, and reduced ejection fraction.

We found that knockdown of Taz did not affect normal hematopoiesis under basal conditions in mice. Interestingly, when subjected to hematopoietic stress, Taz-KD mice had a 40% decrease in ST-HSC and 30% increase in GMP cells, although no change in clonogenic growth was observed. Thus, our results provide a potential mechanism for the transient episodes of neutropenia seen in Barth syndrome. As reported for other conditions that lead to chronic or cyclic neutropenia, repeated episodes of hematopoietic stress can reduce the ST-HSC reserve and produce neutropenia (Bartels et al., 2016; Colijn and Mackey, 2005; Haurie et al., 1998; Mitroulis et al., 2017).

However, the doxycycline used to induce Taz knockdown in this model has been reported to reduce neutrophil counts. Although, our experimental design controlled for this limitation, the subtle hematopoietic phenotype observed may have been due to the effect of doxycycline on

141

neutrophils. Future studies should examine hematopoiesis in a more clinically relevant model of Barth syndrome (Add ASG abstract here?).

4.5 Cardiolipin and Phospholipid Synthesis

Biological lipids are divided into 8 categories, fatty acyls, glycerolipids, glycerophospholipids, sphingolipids, sterol lipids, prenol lipids, saccharolipids, and polyketides . Out of which, glycerophospholipids, sphingolipids, and sterols are the most abundant membrane lipids (Koberlin et al., 2015). Each class of lipid has distinct biosynthetic processes. Recently, Koberlin et al., genetically perturbed sphingomyelin metabolism and then used mass spectrometry based lipidomics to construct a lipid coregulatory network. Similar to metabolic pathways such as the citric acid cycle and urea cycle, the lipid metabolic pathway also demonstrates a circular motif. Notably, there was a tight coregulation of sphingolipids with glycerolipids. However, the molecular mechanism by which these classes of lipids are co-regulated remains unknown.

PC, PE, triacylglycerol (TAG), PI, PG, PS, and cardiolipin are examples of glycerolipids. All glycerolipids are derived from phosphatidic acid (Figure 43). PI, PG, and cardiolipin synthesis requires PA to be first converted to CDP-DAG, while PC, PE, TAG, and PS synthesis requires PA to be first converted to DAG (Kelly and Jacobs, 2018). Interestingly, the Koberlin study also showed how the synthesis of these lipids are regulated. PS and PE are co-regulated, while PC synthesis is co-regulated with PG (Koberlin et al., 2015). In this study, we demonstrated that the levels of PS are influenced by cardiolipin metabolism and gained insight into the enzymes that regulate the cardiolipin-PS axis.

Figure 43. Illustration of phospholipid synthesis. Biochemical steps involved in the synthesis of PC, PE, PS, TAG, PI, PG, and cardiolipin. Adapted from Kelly and Jacobs 2018. GPAT, glycerol 3-phosphate acyltransferase; AGPAT, acylglycerol-3-acyltransferase; PAP, phosphatidic acid phosphatase; PSS1, phosphatidylserine synthase 1; PSS2, phosphatidylserine synthase 1; PISD, phosphatidylserine decarboxylase; CDS, CDP-diacylglycerol synthase; PA, phosphatidic acid; DAG, diacylglycerol; PC, phosphatidylcholine; PE, phosphatidylethanolamine; TAG, triacylglycerol; PI, phosphatidylinositol; PG, phosphatidylglycerol.

142

4.6 Cardiolipin and PISD

We showed that knockdown of TAZ increased levels of PS by reducing levels of functional PISD protein (~38kDA) but not mRNA. Of note, PS synthesis enzymes PSS1 and PSS2 were not affected after TAZ knockdown. PISD is a decarboxylase that converts PS to PE. Our data are thus consistent with prior studies where PISD inhibition increases PS and decreases PE. Future studies will help define how inhibiting TAZ and decreasing cardiolipin affects levels of PISD protein. Potentially, mitochondrial PISD stability and function requires cardiolipin (Friedman et al., 2017; Gebert et al., 2009; Jiang et al., 2000; van der Laan et al., 2007). It is also plausible that the reduction of cardiolipin reduces the transport of PISD into the mitochondria by impairing the function of inner and outer mitochondrial membrane translocases.

4.7 Phospholipids and Cell Signaling Pathways

Altering cellular lipid composition modulates the activity of some cell signalling pathways (Bi et al., 2019; Huang et al., 2011; Kay and Grinstein, 2013; Koberlin et al., 2015). Firstly, the interaction of Akt with the cell membrane plays a critical role in the activation of Akt signaling. PS in the cell membrane promotes: Akt binding to PIP3-another cell membrane phospholipid, participates in PIP3-induced Akt interdomain conformational changes for T308 phosphorylation, and causes an open confirmation that allows for S473 phosphorylation by mTORC2 (Huang et al., 2011). Secondly, sphingomyelin composition in the cell membrane regulates TLR trafficking, signaling, and cytokine release. Finally, saturated phosphatidylcholine required for the transduction of growth factor signaling pathways such as EGFR (Bi et al., 2019).

We have shown that increasing PS levels activates TLR signalling. TLRs are expressed on stem cells and myeloid cells (Hayashi et al., 2010; Ignatz-Hoover et al., 2015; Koberlin et al., 2015; Megias et al., 2012; Nagai et al., 2006). They are classified into two subfamilies based on its cellular localization, cell surface TLRs and intracellular TLRs. The TLRs expressed on the cell surface include: TLR1, TLR2, TLR4, TLR5, TLR6 and TLR10. Intracellular TLRs on the other hand include: TLR3, TLR7, TLR8, TLR9, TLR11, TLR12, TLR13 (Kawasaki and Kawai, 2014). TLRs are stimulated by a number of viral components. The TLRs on the cell surface recognizes lipids, lipoproteins, and proteins in the microbial membrane. Intracellular TLRs on the other hand are stimulated by nucleic acid from bacteria and viruses, and self-nucleic acids during conditions such as autoimmunity. TLR signaling can be divided into MYD88 and TRIF dependent pathways (Kawasaki and Kawai, 2014). MYD88 is used by the majority of TLRs to

143 activate MAPKs, NF-kB, and inflammatory cytokine genes. Alternatively, TLRs such as TLR4 and TLR3 can recruit TRIF to activate MAPKs, IRFs, and NF-kb to induce type I interferon signaling. Notably, TRAF3 plays a critical role in the regulating MYD88-dependent and TRIF- dependent signaling, the degradation of TRAF3 activates MYD88-dependent signaling while supressing TRIF dependent signaling (Figure 3).

Figure 44. Illustration of the TLR signaling pathway. TLR signal either through MYD88 or TRIF. Illustration reproduced courtesy of Cell Signaling Technology, Inc. (www.cellsignal.com).

144

PS predominately stimulated the TLR4 and TLR8 pathway. Interestingly, TLR signalling is a known regulator of hematopoietic and AML differentiation (Hayashi et al., 2010; Ignatz-Hoover et al., 2015; Koberlin et al., 2015; Megias et al., 2012; Nagai et al., 2006). The stimulation of TLR4 drives myeloid differentiation during emergency hematopoiesis in response to infection (Hayashi et al., 2010; Megias et al., 2012; Nagai et al., 2006). Also, the TLR8 agonist R848 promotes AML differentiation and reduces tumour burden both in vivo and in vitro (Ignatz- Hoover et al., 2015).

We did not show how PS increase regulates TLR signaling in this study. Nuclear PS has been shown to impact chromatin accessibility (Cocco et al., 1986; Cocco et al., 1988; Jones and Divecha, 2004; Kuvichkin, 2002). Also, PE on the plasma membrane has been shown to act as a methyl sink, by consuming S-adenosylmethionine (SAM) and thereby reducing the methylation of DNA (Ye et al., 2017). PE has to be methylated to be converted into PC. The reduction PE methylation has been shown to results in the accumulation of SAM. Higher SAM levels lead to the increase in histone trimethylation. Since TAZ inhibition reduces levels of PE, TAZ may increase TLR expression by increasing the trimethylation of histones. Future studies should examine if PS directly activates TLR receptors or if PS increases TLR gene expression by epigenetic mechanisms.

RIG-I and MDA5 are cytoplasmic receptors that also recognize pathogenic components (Jacobs and Coyne, 2013; Seth et al., 2005; Wu and Hur, 2015). Once, activated both RIG-I and MDA5 bind to and activate the mitochondrial antiviral signaling proteins (MAVS), on the outer mitochondrial membrane. The RIG-I/MDA5, MAVs complex then recruit TRAFs which stimulate interferon production. Protein-protein interactions, alternations in mitochondrial dynamics, and post-translational modifications modulate MAVS signaling. Mitochondrial phospholipids such as cardiolipin has been shown to have immune activation functions (Balasubramanian et al., 2015). Moreover, MAVS aggregation has been linked to disruption of mitochondrial pathways and increased anti-cancer immune response signalling and may explain the activation of the TLR pathway. Future investigations should determine if TAZ knockdown and PS increase stimulates TLR pathways by activating MAVS.

4.8 PISD Inhibitor MMV007285

Increasing intracellular levels of PS by supplementing cells with this phospholipid decreased AML growth and stemness, and thus phenocopied the effects of TAZ inhibition. These results

145 are consistent with a prior study where PS supplementation promoted the osteogenic differentiation of mesenchymal stromal cells (Xu et al., 2013). We also increased intracellular PS and reduced AML stemness by genetically or chemically inhibiting PISD. As a chemical approach, we used the PISD inhibitor, MMV007285 (Choi et al., 2016; Lucantoni et al., 2013). MMV007285 was initially identified through screen containing 200 drug-like and 200 probe-like compounds, for inhibitors of malarial PISD. Our results show that MMV007285 also targets the mammalian form of PISD. In breast cancer inhibiting PISD induced the differentiation in these cells (Keckesova et al., 2017). Thus, similar to breast cancer inhibiting PISD in AML reduces stemness and promoted differentiation and could be a novel therapeutic strategy for AML.

Chapter 5 Future Directions 5.1 Mechanistic Studies

5.1.1 Determine how Increase in PS Stimulates TLR activation

We showed that TAZ stimulates TLR activation by increasing PS. However, additional experiments are required to functionally establish that the TLR pathway is activated, and to understand the mechanism by which PS increases TLR pathway. In order to further demonstrate that the TLR pathway is functionally important following PS increase after TAZ studies should attempt to block the most highly upregulated TLR signaling pathways, TLR4 and TLR8, by treating cells with the reported TLR4 or TLR8 antagonists LPS-RS and CU-CPT-8m respectively. Furthermore, a TLR signaling deficient cell line should be constructed by knocking down the TLR signaling adaptor proteins, MYD88 and TRIF and its sensitivity to the inhibition of TAZ and the increase in PS should be tested.

To determine if this occurs through changes in membrane composition or epigenetics, we will fractionate the cell membrane from the nucleus and measure levels of membrane, and nuclear PS after TAZ knockdown, MMV007285 and PS treatment. Increase in PS in the cell membrane and not the nucleus suggest that PS stimulates the activation of the TLR pathway by increasing PS in the cell membrane. Alternatively, the increase in PS in the cell nucleus suggests that PS increases the chromatin accessibility of the TLR pathway. These investigations should be paired with ATAC-seq to measure chromatin accessibility.

We could not rescue the TAZ knockdown phenotype by supplementing cells with PE or its substrate LPE. However, it is still possible decrease in PE was functionally important for the effects of TAZ knockdown. The inability of PE or LPE to rescue the effects of TAZ knockdown may have been due to PE and LPE not being able to increase PE levels beyond the biological threshold required to have a functional effect. In order to investigate the functional importance of PE’s activity as a methyl sink after TAZ knockdown, future studies should measure levels of SAM, and trimethylated histone marks after TAZ knockdown. The ability to rescue these effects by PE and LPE liposomes should also be investigated.

146 147

The role of MAVS in TAZ knockdown and PS upregulation should also be investigated by measuring MAVs aggregation in TAZ knockdown, PS supplemented, and MMV treated samples. The functional role of MAVS aggregation should be determined by stimulating AML cells with the RIG-I specific stem-loop RNA molecule and measuring MAVS aggregation, and AML stem cell properties both in vivo and in vitro (Jacobs and Coyne, 2013; Linehan et al., 2018). It also should be attempted to rescue the phenotype of TAZ knockdown, and PS increase by genetically inhibiting MAVS (Chiappinelli et al., 2015).

5.1.2 Determine if metabolic changes after TAZ-KD stimulates differentiation

The rate of glycolytic metabolism changes during the differentiation of human embryonic stem cells as well as the reprogramming of somatic cells to pluripotency (Gu et al., 2016). Both embryonic stem cells and induced pluripotent stem cells have higher rates of glycolysis that diminish during differentiation. More differentiated cells also have higher oxygen consumption rates suggesting that differentiation is associated with a shift away from glycolytic metabolism towards oxidative metabolism. Moreover, abating glycolytic metabolism by inhibiting the pyruvate and lactate transporter monocarboxylate transporter 1 induced the differentiation of embryonic stem cells. We observed that TAZ knockdown reduced glycolysis, and one of the TAZ shRNA clones increased oxidative phosphorylation of leukemia cells. However, other approaches that shift cells to oxidative phosphorylation, such as culturing cells in galactose media, did not increase AML differentiation (unpublished data). Also, the inhibition of glycolysis has not been shown to stimulate AML differentiation. Therefore, we think that shifts in metabolism observed after TAZ knockdown are manifestation and not causes of differentiation. We did not adequately address the relationship between TAZ and metabolism in this study, so future studies to comprehensively address this relationship are warranted.

5.1.3 Cardiolipin Lipid Network

In this study we showed that cardiolipin influences levels of both PE and PS, however the relationship of cardiolipin with other lipids is not known. We also do not know if the cardiolipin lipid network in normal cells is that same as that in malignant cells. In order to study the role of cardiolipin in membrane lipid composition in AML, enzymes in the cardiolipin synthesis and remodeling pathways should be first genetically perturbed and then mass spectrometry based lipidomics should be used to characterize lipid changes. It will also be interesting characterize

148 the cellular phenotypes after the perturbation of both cardiolipin synthesis and remodeling enzymes. Cardiolipin synthesis or remodeling enzymes that drastically alter the leukemia lipidome, as well as reduce AML growth and viability should be genetically perturbed in normal hematopoietic cells to understand if the cardiolipin lipid network in malignant cells differs from that in normal cells.

5.2 Pre-clinical studies

5.2.1 Characterization of Taz knockdown in Barth Syndrome

Edwards et al., have developed a new mouse model of Barth Syndrome by using CRISPR, to introduce the D75H point mutation into the murine Taz gene (Edwards, 2017). This mouse model recapitulated cardiac, skeletal muscle, and hematopoietic defects of the disease. Future studies should investigate the lipidomic profile, frequencies and functions of hematopoietic stem cells in this mouse model of Barth syndrome. If the increase in PS is also seen in this mouse model, it will be interesting to determine if the inhibition of one of the PS synthesis enzymes- PSS1 or PSS2-rescues the neutropenia observed in this mouse model. Furthermore, the role of TAZ knockdown in human HSC and HPC function should also be evaluated by transducing CD34+ human HSC/HPC with lentiviral vectors targeting TAZ. Subsequently, PS, CL levels and stem cell function should be characterized.

5.2.2 Basis for the AML specific effect of TAZ knockdown and PS increase

According to the dependency map portal, AML cells are more dependent on TAZ compared to normal hematopoietic cells (Broad, January, 2020). Moreover, we have shown that TAZ knockdown and PISD inhibition specifically target the initiation of propagation properties of AML stem cells. However, the basis of this therapeutic window is not clear. TAZ and PISD are not typical AML targets. Both these proteins are expressed on both normal hematopoietic cells and AML cells, and it is not clearly overexpressed in malignant cells. Interestingly, PS is 1.8- fold lower in a subset of AML patients when compared to normal hematopoietic cells. To determine if this is the basis of therapeutic window, PS should be quantified in a larger cohort of patient samples, and the stem cell function of these samples after TAZ knockdown, and MMV treatment should be characterized. Finally, the correlation between PS levels and sensitivity to TAZ knockdown and MMV treatment should be determined.

149

5.2.3 Improve Efficacy of PISD inhibitor

Although a useful tool compound to study PISD in vitro, MMV007285 has poor stability in aqueous media and a short half-life when administered systemically to mice. It is quite straightforward to make structural analogs of MMV007285 due to its structure (Choi et al., 2016). Therefore, developing more potent and stable PISD inhibitors based on alternate scaffolds is a priority to fully assess the efficacy and toxicity of PISD inhibition in vivo and evaluate PISD as a potential therapeutic target for both AML and breast cancer.

5.2.4 Phospholipid Modulation and Viral Mimicry

The increase in PS after TAZ and PISD inhibition increases TLR receptor activation. The activation of innate immune genes such as TLRs in cancer is referred to as viral mimicry (Alcami, 2003; Roulois et al., 2015). The state of viral mimicry has been shown to increase the sensitivity of other cancers to immune check point inhibitors. It will be interesting to determine if PS increase after TAZ knockdown enhances the sensitivity of AML to immune checkpoint inhibitors as well other immune therapies like chimeric antigen receptor T cells. In addition, the GVT effect is essentially an immune reaction of the donor cells against the tumor. Currently, there are no therapies that help potentiate GVT while not enhancing the GVHD effect. Thus, it will also be interesting to determine if an increase in innate immune pathways after TAZ knockdown or PISD inhibition increases the efficacy of GVT without increasing GVHD.

150

References or Bibliography (if any)

Tutorials and lectures on lipids (15).

Abelson, S., Collord, G., Ng, S.W.K., Weissbrod, O., Mendelson Cohen, N., Niemeyer, E., Barda, N., Zuzarte, P.C., Heisler, L., Sundaravadanam, Y., et al. (2018). Prediction of acute myeloid leukaemia risk in healthy individuals. Nature.

Acehan, D., Khuchua, Z., Houtkooper, R.H., Malhotra, A., Kaufman, J., Vaz, F.M., Ren, M., Rockman, H.A., Stokes, D.L., and Schlame, M. (2009). Distinct effects of tafazzin deletion in differentiated and undifferentiated mitochondria. Mitochondrion 9, 86-95.

Acehan, D., Vaz, F., Houtkooper, R.H., James, J., Moore, V., Tokunaga, C., Kulik, W., Wansapura, J., Toth, M.J., Strauss, A., et al. (2011). Cardiac and skeletal muscle defects in a mouse model of human Barth syndrome. J Biol Chem 286, 899-908.

Acehan, D., Xu, Y., Stokes, D.L., and Schlame, M. (2007). Comparison of lymphoblast mitochondria from normal subjects and patients with Barth syndrome using electron microscopic tomography. Lab Invest 87, 40-48.

Adane, B., Ye, H., Khan, N., Pei, S., Minhajuddin, M., Stevens, B.M., Jones, C.L., D'Alessandro, A., Reisz, J.A., Zaberezhnyy, V., et al. (2019). The Hematopoietic Oxidase NOX2 Regulates Self-Renewal of Leukemic Stem Cells. Cell Rep 27, 238-254 e236.

Agaronyan, K., Morozov, Y.I., Anikin, M., and Temiakov, D. (2015). Mitochondrial biology. Replication-transcription switch in human mitochondria. Science 347, 548-551.

Ågerstam, H., Karlsson, C., Hansen, N., Sandén, C., Askmyr, M., von Palffy, S., Högberg, C., Rissler, M., Wunderlich, M., Juliusson, G., et al. (2015). Antibodies targeting human IL1RAP (IL1R3) show therapeutic effects in xenograft models of acute myeloid leukemia. Proc Natl Acad Sci U S A 112, 10786-10791.

Alberts, B.B., D; Hopkin, K; Jhonson, A; Lewis, J; Raff, M; Roberts, K; Walter, P (2014). Essential Cell Biology (New York: W. W. Norton & Company, Inc).

Alcami, A. (2003). Viral mimicry of cytokines, chemokines and their receptors. Nat Rev Immunol 3, 36-50.

Anzmann, A.C., S.M.; Vernon, H (2019). Handbook of Mitochondrial Dysfunction (CRC Press, Taylor & Francis Group).

Appelbaum, F.R. (2009). Optimising the conditioning regimen for acute myeloid leukaemia. Best Pract Res Clin Haematol 22, 543-550.

Arber, D.A., Orazi, A., Hasserjian, R., Thiele, J., Borowitz, M.J., Le Beau, M.M., Bloomfield, C.D., Cazzola, M., and Vardiman, J.W. (2016). The 2016 revision to the World Health Organization classification of myeloid neoplasms and acute leukemia. Blood 127, 2391-2405.

151

Askmyr, M., Ågerstam, H., Hansen, N., Gordon, S., Arvanitakis, A., Rissler, M., Juliusson, G., Richter, J., Järås, M., and Fioretos, T. (2013). Selective killing of candidate AML stem cells by antibody targeting of IL1RAP. Blood 121, 3709-3713.

Baccelli, I., Gareau, Y., Lehnertz, B., Gingras, S., Spinella, J.F., Corneau, S., Mayotte, N., Girard, S., Frechette, M., Blouin-Chagnon, V., et al. (2019). Mubritinib Targets the Electron Transport Chain Complex I and Reveals the Landscape of OXPHOS Dependency in Acute Myeloid Leukemia. Cancer cell 36, 84-99 e88.

Baile, M.G., Sathappa, M., Lu, Y.W., Pryce, E., Whited, K., McCaffery, J.M., Han, X., Alder, N.N., and Claypool, S.M. (2014). Unremodeled and remodeled cardiolipin are functionally indistinguishable in yeast. J Biol Chem 289, 1768-1778.

Baker, M.J., Tatsuta, T., and Langer, T. (2011). Quality control of mitochondrial proteostasis. Cold Spring Harb Perspect Biol 3.

Balasubramanian, K., Maeda, A., Lee, J.S., Mohammadyani, D., Dar, H.H., Jiang, J.F., St Croix, C.M., Watkins, S., Tyurin, V.A., Tyurina, Y.Y., et al. (2015). Dichotomous roles for externalized cardiolipin in extracellular signaling: Promotion of phagocytosis and attenuation of innate immunity. Sci Signal 8, ra95.

Ban, T., Heymann, J.A., Song, Z., Hinshaw, J.E., and Chan, D.C. (2010). OPA1 disease alleles causing dominant optic atrophy have defects in cardiolipin-stimulated GTP hydrolysis and membrane tubulation. Hum Mol Genet 19, 2113-2122.

Ban, T., Ishihara, T., Kohno, H., Saita, S., Ichimura, A., Maenaka, K., Oka, T., Mihara, K., and Ishihara, N. (2017). Molecular basis of selective mitochondrial fusion by heterotypic action between OPA1 and cardiolipin. Nat Cell Biol 19, 856-863.

Baran, N., Han, L., Herbrich, S., Sweeney, S.R., Lodi, A., Gay, J., Feng, N., Molina, J.R., Kaminski, M., Guzman, M.L., et al. Novel Complex I inhibitor IACS-010759 Targets Leukemia Initiating Cells (LICs) in AML Patients. Clinical Lymphoma, Myeloma and Leukemia 17, S297.

Bartels, M., Murphy, K., Rieter, E., and Bruin, M. (2016). Understanding chronic neutropenia: life is short. Br J Haematol 172, 157-169.

Basak, N.P., and Banerjee, S. (2015). Mitochondrial dependency in progression of acute myeloid leukemia. Mitochondrion 21, 41-48.

Battula, V.L., Chen, Y., Cabreira Mda, G., Ruvolo, V., Wang, Z., Ma, W., Konoplev, S., Shpall, E., Lyons, K., Strunk, D., et al. (2013). Connective tissue growth factor regulates adipocyte differentiation of mesenchymal stromal cells and facilitates leukemia bone marrow engraftment. Blood 122, 357-366.

Bauer, D.E., Hatzivassiliou, G., Zhao, F., Andreadis, C., and Thompson, C.B. (2005). ATP citrate lyase is an important component of cell growth and transformation. Oncogene 24, 6314- 6322.

Behan, J.W., Yun, J.P., Proektor, M.P., Ehsanipour, E.A., Arutyunyan, A., Moses, A.S., Avramis, V.I., Louie, S.G., Butturini, A., Heisterkamp, N., et al. (2009). Adipocytes impair leukemia treatment in mice. Cancer Res 69, 7867-7874.

152

Ben-Sahra, I., Hoxhaj, G., Ricoult, S.J.H., Asara, J.M., and Manning, B.D. (2016). mTORC1 induces purine synthesis through control of the mitochondrial tetrahydrofolate cycle. Science 351, 728-733.

Berwick, D.C., Hers, I., Heesom, K.J., Moule, S.K., and Tavare, J.M. (2002). The identification of ATP-citrate lyase as a protein kinase B (Akt) substrate in primary adipocytes. J Biol Chem 277, 33895-33900.

Bhargava, K., Templeton, P., and Spremulli, L.L. (2004). Expression and characterization of isoform 1 of human mitochondrial elongation factor G. Protein Expr Purif 37, 368-376.

Bhat, V., Joshi, A., Sarode, R., and Chavan, P. (2015). Cytomegalovirus infection in the bone marrow transplant patient. World J Transplant 5, 287-291.

Bi, J., Ichu, T.A., Zanca, C., Yang, H., Zhang, W., Gu, Y., Chowdhry, S., Reed, A., Ikegami, S., Turner, K.M., et al. (2019). Oncogene Amplification in Growth Factor Signaling Pathways Renders Cancers Dependent on Membrane Lipid Remodeling. Cell Metab.

Bissler, J.J., Tsoras, M., Goring, H.H., Hug, P., Chuck, G., Tombragel, E., McGraw, C., Schlotman, J., Ralston, M.A., and Hug, G. (2002). Infantile dilated X-linked cardiomyopathy, G4.5 mutations, altered lipids, and ultrastructural malformations of mitochondria in heart, liver, and skeletal muscle. Lab Invest 82, 335-344.

Boeckh, M., and Ljungman, P. (2009). How we treat cytomegalovirus in hematopoietic cell transplant recipients. Blood 113, 5711-5719.

Boettcher, S., and Manz, M.G. (2016). Sensing and translation of pathogen signals into demand- adapted myelopoiesis. Curr Opin Hematol 23, 5-10.

Boettcher, S., and Manz, M.G. (2017). Regulation of Inflammation- and Infection-Driven Hematopoiesis. Trends Immunol 38, 345-357.

Bonardi, F., Fusetti, F., Deelen, P., van Gosliga, D., Vellenga, E., and Schuringa, J.J. (2013). A proteomics and transcriptomics approach to identify leukemic stem cell (LSC) markers. Mol Cell Proteomics 12, 626-637.

Bonnet, D., and Dick, J.E. (1997). Human acute myeloid leukemia is organized as a hierarchy that originates from a primitive hematopoietic cell. Nat Med 3, 730-737.

Bordi, M., Nazio, F., and Campello, S. (2017). The Close Interconnection between Mitochondrial Dynamics and Mitophagy in Cancer. Front Oncol 7, 81.

Borisy, A.A., Elliott, P.J., Hurst, N.W., Lee, M.S., Lehar, J., Price, E.R., Serbedzija, G., Zimmermann, G.R., Foley, M.A., Stockwell, B.R., et al. (2003). Systematic discovery of multicomponent therapeutics. Proc Natl Acad Sci U S A 100, 7977-7982.

Boutros, M., and Ahringer, J. (2008). The art and design of genetic screens: RNA interference. Nat Rev Genet 9, 554-566.

Bowden, R.A., Slichter, S.J., Sayers, M., Weisdorf, D., Cays, M., Schoch, G., Banaji, M., Haake, R., Welk, K., Fisher, L., et al. (1995). A comparison of filtered leukocyte-reduced and

153 cytomegalovirus (CMV) seronegative blood products for the prevention of transfusion- associated CMV infection after marrow transplant. Blood 86, 3598-3603.

Bowman, R.L., Busque, L., and Levine, R.L. (2018). Clonal Hematopoiesis and Evolution to Hematopoietic Malignancies. Cell Stem Cell 22, 157-170.

Bradley, R.M., Mardian, E.B., Bloemberg, D., Aristizabal Henao, J.J., Mitchell, A.S., Marvyn, P.M., Moes, K.A., Stark, K.D., Quadrilatero, J., and Duncan, R.E. (2017). Mice Deficient in lysophosphatidic acid acyltransferase delta (Lpaatdelta)/acylglycerophosphate acyltransferase 4 (Agpat4) Have Impaired Learning and Memory. Mol Cell Biol 37.

Bralha, F.N., Liyanage, S.U., Hurren, R., Wang, X., Son, M.H., Fung, T.A., Chingcuanco, F.B., Tung, A.Y., Andreazza, A.C., Psarianos, P., et al. (2015). Targeting mitochondrial RNA polymerase in acute myeloid leukemia. Oncotarget 6, 37216-37228.

Broad Institute (January, 2020). Dependency Map (DepMap).

Burman, J.L., Yu, S., Poole, A.C., Decal, R.B., and Pallanck, L. (2012). Analysis of neural subtypes reveals selective mitochondrial dysfunction in dopaminergic neurons from parkin mutants. Proc Natl Acad Sci U S A 109, 10438-10443.

Burnett, A.K., Russell, N.H., Hills, R.K., Bowen, D., Kell, J., Knapper, S., Morgan, Y.G., Lok, J., Grech, A., Jones, G., et al. (2015). Arsenic trioxide and all-trans retinoic acid treatment for acute promyelocytic leukaemia in all risk groups (AML17): results of a randomised, controlled, phase 3 trial. Lancet Oncol 16, 1295-1305.

Buscarlet, M., Provost, S., Zada, Y.F., Barhdadi, A., Bourgoin, V., Lépine, G., Mollica, L., Szuber, N., Dubé, M.P., and Busque, L. (2017). DNMT3A and TET2 dominate clonal hematopoiesis and demonstrate benign phenotypes and different genetic predispositions. Blood 130, 753-762.

Busch, K., Klapproth, K., Barile, M., Flossdorf, M., Holland-Letz, T., Schlenner, S.M., Reth, M., Hofer, T., and Rodewald, H.R. (2015). Fundamental properties of unperturbed haematopoiesis from stem cells in vivo. Nature 518, 542-546.

Busque, L., Patel, J.P., Figueroa, M.E., Vasanthakumar, A., Provost, S., Hamilou, Z., Mollica, L., Li, J., Viale, A., Heguy, A., et al. (2012). Recurrent somatic TET2 mutations in normal elderly individuals with clonal hematopoiesis. Nat Genet 44, 1179-1181.

Bustillo-Zabalbeitia, I., Montessuit, S., Raemy, E., Basanez, G., Terrones, O., and Martinou, J.C. (2014). Specific interaction with cardiolipin triggers functional activation of Dynamin-Related Protein 1. PLoS One 9, e102738.

Buzzai, M., Bauer, D.E., Jones, R.G., Deberardinis, R.J., Hatzivassiliou, G., Elstrom, R.L., and Thompson, C.B. (2005). The glucose dependence of Akt-transformed cells can be reversed by pharmacologic activation of fatty acid beta-oxidation. Oncogene 24, 4165-4173.

Cancer Genome Atlas Research, N. (2013). Genomic and epigenomic landscapes of adult de novo acute myeloid leukemia. N Engl J Med 368, 2059-2074.

154

Cassileth, P.A., Harrington, D.P., Hines, J.D., Oken, M.M., Mazza, J.J., McGlave, P., Bennett, J.M., and O'Connell, M.J. (1988). Maintenance chemotherapy prolongs remission duration in adult acute nonlymphocytic leukemia. J Clin Oncol 6, 583-587.

Cereghetti, G.M., Stangherlin, A., Martins de Brito, O., Chang, C.R., Blackstone, C., Bernardi, P., and Scorrano, L. (2008). Dephosphorylation by calcineurin regulates translocation of Drp1 to mitochondria. Proc Natl Acad Sci U S A 105, 15803-15808.

Challen, G.A., Boles, N., Lin, K.K., and Goodell, M.A. (2009). Mouse hematopoietic stem cell identification and analysis. Cytometry A 75, 14-24.

Challen, G.A., Sun, D., Mayle, A., Jeong, M., Luo, M., Rodriguez, B., Mallaney, C., Celik, H., Yang, L., Xia, Z., et al. (2014). Dnmt3a and Dnmt3b have overlapping and distinct functions in hematopoietic stem cells. Cell Stem Cell 15, 350-364.

Chan, N.C., Salazar, A.M., Pham, A.H., Sweredoski, M.J., Kolawa, N.J., Graham, R.L., Hess, S., and Chan, D.C. (2011). Broad activation of the ubiquitin-proteasome system by Parkin is critical for mitophagy. Hum Mol Genet 20, 1726-1737.

Chan, S.M., Thomas, D., Corces-Zimmerman, M.R., Xavy, S., Rastogi, S., Hong, W.J., Zhao, F., Medeiros, B.C., Tyvoll, D.A., and Majeti, R. (2015). Isocitrate dehydrogenase 1 and 2 mutations induce BCL-2 dependence in acute myeloid leukemia. Nat Med 21, 178-184.

Chang, Y.J., Zhao, X.Y., and Huang, X.J. (2018). Strategies for Enhancing and Preserving Anti- leukemia Effects Without Aggravating Graft-Versus-Host Disease. Front Immunol 9, 3041.

Chen, S., Sanjana, N.E., Zheng, K., Shalem, O., Lee, K., Shi, X., Scott, D.A., Song, J., Pan, J.Q., Weissleder, R., et al. (2015). Genome-wide CRISPR screen in a mouse model of tumor growth and metastasis. Cell 160, 1246-1260.

Cheneval, D., Muller, M., and Carafoli, E. (1983). The mitochondrial phosphate carrier reconstituted in liposomes is inhibited by doxorubicin. FEBS Lett 159, 123-126.

Chiappinelli, K.B., Strissel, P.L., Desrichard, A., Li, H., Henke, C., Akman, B., Hein, A., Rote, N.S., Cope, L.M., Snyder, A., et al. (2015). Inhibiting DNA Methylation Causes an Interferon Response in Cancer via dsRNA Including Endogenous Retroviruses. Cell 162, 974-986.

Choi, J.S., and Berdis, A.J. (2012). Nucleoside transporters: biological insights and therapeutic applications. Future Med Chem 4, 1461-1478.

Choi, J.Y., Kumar, V., Pachikara, N., Garg, A., Lawres, L., Toh, J.Y., Voelker, D.R., and Ben Mamoun, C. (2016). Characterization of Plasmodium phosphatidylserine decarboxylase expressed in yeast and application for inhibitor screening. Mol Microbiol 99, 999-1014.

Christian, B.E., and Spremulli, L.L. (2009). Evidence for an active role of IF3mt in the initiation of translation in mammalian mitochondria. Biochemistry 48, 3269-3278.

Chu, C.T., Ji, J., Dagda, R.K., Jiang, J.F., Tyurina, Y.Y., Kapralov, A.A., Tyurin, V.A., Yanamala, N., Shrivastava, I.H., Mohammadyani, D., et al. (2013). Cardiolipin externalization to the outer mitochondrial membrane acts as an elimination signal for mitophagy in neuronal cells. Nat Cell Biol 15, 1197-1205.

155

Chung, S.S., Eng, W.S., Hu, W., Khalaj, M., Garrett-Bakelman, F.E., Tavakkoli, M., Levine, R.L., Carroll, M., Klimek, V.M., Melnick, A.M., et al. (2017). CD99 is a therapeutic target on disease stem cells in myeloid malignancies. Sci Transl Med 9.

Churchward, M.A., Brandman, D.M., Rogasevskaia, T., and Coorssen, J.R. (2008). Copper (II) sulfate charring for high sensitivity on-plate fluorescent detection of lipids and sterols: quantitative analyses of the composition of functional secretory vesicles. Journal of chemical biology 1, 79-87.

CIBMTR (2010/2011). Comprehensive Disease Specific Manuals-Acute Myeloid Leukemia (Center for International Blood & Marrow Transplant Research ).

Cicconi, L., Divona, M., Ciardi, C., Ottone, T., Ferrantini, A., Lavorgna, S., Alfonso, V., Paoloni, F., Piciocchi, A., Avvisati, G., et al. (2016). PML-RARalpha kinetics and impact of FLT3-ITD mutations in newly diagnosed acute promyelocytic leukaemia treated with ATRA and ATO or ATRA and chemotherapy. Leukemia 30, 1987-1992.

Civin, C.I., and Loken, M.R. (1987). Cell surface antigens on human marrow cells: dissection of hematopoietic development using monoclonal antibodies and multiparameter flow cytometry. Int J Cell Cloning 5, 267-288.

Clark, I.E., Dodson, M.W., Jiang, C., Cao, J.H., Huh, J.R., Seol, J.H., Yoo, S.J., Hay, B.A., and Guo, M. (2006). Drosophila pink1 is required for mitochondrial function and interacts genetically with parkin. Nature 441, 1162-1166.

Clarke, S.L., Bowron, A., Gonzalez, I.L., Groves, S.J., Newbury-Ecob, R., Clayton, N., Martin, R.P., Tsai-Goodman, B., Garratt, V., Ashworth, M., et al. (2013). Barth syndrome. Orphanet J Rare Dis 8, 23.

Claypool, S.M., and Koehler, C.M. (2012). The complexity of cardiolipin in health and disease. Trends Biochem Sci 37, 32-41.

Claypool, S.M., McCaffery, J.M., and Koehler, C.M. (2006). Mitochondrial mislocalization and altered assembly of a cluster of Barth syndrome mutant tafazzins. J Cell Biol 174, 379-390.

Claypool, S.M., Oktay, Y., Boontheung, P., Loo, J.A., and Koehler, C.M. (2008). Cardiolipin defines the interactome of the major ADP/ATP carrier protein of the mitochondrial inner membrane. J Cell Biol 182, 937-950.

Clayton, D.A. (2000). Transcription and replication of mitochondrial DNA. Hum Reprod 15 Suppl 2, 11-17.

Cole, A., Wang, Z., Coyaud, E., Voisin, V., Gronda, M., Jitkova, Y., Mattson, R., Hurren, R., Babovic, S., Maclean, N., et al. (2015). Inhibition of the Mitochondrial Protease ClpP as a Therapeutic Strategy for Human Acute Myeloid Leukemia. Cancer cell 27, 864-876.

Colijn, C., and Mackey, M.C. (2005). A mathematical model of hematopoiesis: II. Cyclical neutropenia. J Theor Biol 237, 133-146.

Coombs, C.C., Zehir, A., Devlin, S.M., Kishtagari, A., Syed, A., Jonsson, P., Hyman, D.M., Solit, D.B., Robson, M.E., Baselga, J., et al. (2017). Therapy-Related Clonal Hematopoiesis in

156

Patients with Non-hematologic Cancers Is Common and Associated with Adverse Clinical Outcomes. Cell Stem Cell 21, 374-382.e374.

Coppell, J.A., Richardson, P.G., Soiffer, R., Martin, P.L., Kernan, N.A., Chen, A., Guinan, E., Vogelsang, G., Krishnan, A., Giralt, S., et al. (2010). Hepatic veno-occlusive disease following stem cell transplantation: incidence, clinical course, and outcome. Biol Blood Marrow Transplant 16, 157-168.

Cornelissen, J.J., Gratwohl, A., Schlenk, R.F., Sierra, J., Bornhauser, M., Juliusson, G., Racil, Z., Rowe, J.M., Russell, N., Mohty, M., et al. (2012). The European LeukemiaNet AML Working Party consensus statement on allogeneic HSCT for patients with AML in remission: an integrated-risk adapted approach. Nat Rev Clin Oncol 9, 579-590.

Couriel, D., Carpenter, P.A., Cutler, C., Bolanos-Meade, J., Treister, N.S., Gea-Banacloche, J., Shaughnessy, P., Hymes, S., Kim, S., Wayne, A.S., et al. (2006). Ancillary therapy and supportive care of chronic graft-versus-host disease: national institutes of health consensus development project on criteria for clinical trials in chronic Graft-versus-host disease: V. Ancillary Therapy and Supportive Care Working Group Report. Biol Blood Marrow Transplant 12, 375-396.

Csibi, A., Fendt, S.M., Li, C., Poulogiannis, G., Choo, A.Y., Chapski, D.J., Jeong, S.M., Dempsey, J.M., Parkhitko, A., Morrison, T., et al. (2013). The mTORC1 pathway stimulates glutamine metabolism and cell proliferation by repressing SIRT4. Cell 153, 840-854.

D'Souza, A.R., and Minczuk, M. (2018). Mitochondrial transcription and translation: overview. Essays Biochem 62, 309-320.

Dalle, J.H., and Giralt, S.A. (2016). Hepatic Veno-Occlusive Disease after Hematopoietic Stem Cell Transplantation: Risk Factors and Stratification, Prophylaxis, and Treatment. Biol Blood Marrow Transplant 22, 400-409.

Das, B.C., Thapa, P., Karki, R., Das, S., Mahapatra, S., Liu, T.C., Torregroza, I., Wallace, D.P., Kambhampati, S., Van Veldhuizen, P., et al. (2014). Retinoic acid signaling pathways in development and diseases. Bioorg Med Chem 22, 673-683. de Bruin, A.M., Demirel, Ö., Hooibrink, B., Brandts, C.H., and Nolte, M.A. (2013). Interferon-γ impairs proliferation of hematopoietic stem cells in mice. Blood 121, 3578-3585.

De Kouchkovsky, I., and Abdul-Hay, M. (2016). 'Acute myeloid leukemia: a comprehensive review and 2016 update'. Blood Cancer J 6, e441. de The, H. (2018). Differentiation therapy revisited. Nat Rev Cancer 18, 117-127.

De Vita, S., Schneider, R.K., Garcia, M., Wood, J., Gavillet, M., Ebert, B.L., Gerbaulet, A., Roers, A., Levine, R.L., Mullally, A., et al. (2014). Loss of function of TET2 cooperates with constitutively active KIT in murine and human models of mastocytosis. PLoS One 9, e96209.

Deeg, H.J., and Sandmaier, B.M. (2010). Who is fit for allogeneic transplantation? Blood 116, 4762-4770.

157

Desai, P., Mencia-Trinchant, N., Savenkov, O., Simon, M.S., Cheang, G., Lee, S., Samuel, M., Ritchie, E.K., Guzman, M.L., Ballman, K.V., et al. (2018). Somatic mutations precede acute myeloid leukemia years before diagnosis. Nat Med 24, 1015-1023.

DeVay, R.M., Dominguez-Ramirez, L., Lackner, L.L., Hoppins, S., Stahlberg, H., and Nunnari, J. (2009). Coassembly of Mgm1 isoforms requires cardiolipin and mediates mitochondrial inner membrane fusion. J Cell Biol 186, 793-803.

Di Meo, S., Reed, T.T., Venditti, P., and Victor, V.M. (2016). Harmful and Beneficial Role of ROS. Oxid Med Cell Longev 2016, 7909186.

Di Noia, M.A., Todisco, S., Cirigliano, A., Rinaldi, T., Agrimi, G., Iacobazzi, V., and Palmieri, F. (2014). The human SLC25A33 and SLC25A36 genes of solute carrier family 25 encode two mitochondrial pyrimidine nucleotide transporters. J Biol Chem 289, 33137-33148.

Dick, J.E. (2005). Acute myeloid leukemia stem cells. Ann N Y Acad Sci 1044, 1-5.

DiNardo, C.D., Pratz, K., Pullarkat, V., Jonas, B.A., Arellano, M., Becker, P.S., Frankfurt, O., Konopleva, M., Wei, A.H., Kantarjian, H.M., et al. (2019). Venetoclax combined with decitabine or azacitidine in treatment-naive, elderly patients with acute myeloid leukemia. Blood 133, 7-17.

DiNardo, C.D., Pratz, K.W., Letai, A., Jonas, B.A., Wei, A.H., Thirman, M., Arellano, M., Frattini, M.G., Kantarjian, H., Popovic, R., et al. (2018). 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.

Dohner, H., Estey, E., Grimwade, D., Amadori, S., Appelbaum, F.R., Buchner, T., Dombret, H., Ebert, B.L., Fenaux, P., Larson, R.A., et al. (2017). Diagnosis and management of AML in adults: 2017 ELN recommendations from an international expert panel. Blood 129, 424-447.

Dohner, H., Weisdorf, D.J., and Bloomfield, C.D. (2015). Acute Myeloid Leukemia. N Engl J Med 373, 1136-1152.

Döhner, H., Weisdorf, D.J., and Bloomfield, C.D. (2015). Acute Myeloid Leukemia. N Engl J Med 373, 1136-1152.

Dombret, H., and Gardin, C. (2016). An update of current treatments for adult acute myeloid leukemia. Blood 127, 53-61.

Donald, M., S., M.M.A., and L., W.N. (1969). Colony Formation In Vitro by Myelomonocytic Leukemic Cells. Journal of the National Cancer Institute 43, 983–1001.

Doulatov, S., Notta, F., Laurenti, E., and Dick, J.E. (2012). Hematopoiesis: a human perspective. Cell Stem Cell 10, 120-136.

Dudek, J. (2017). Role of Cardiolipin in Mitochondrial Signaling Pathways. Front Cell Dev Biol 5, 90.

Dudek, J., Cheng, I.F., Balleininger, M., Vaz, F.M., Streckfuss-Bomeke, K., Hubscher, D., Vukotic, M., Wanders, R.J., Rehling, P., and Guan, K. (2013). Cardiolipin deficiency affects

158 respiratory chain function and organization in an induced pluripotent stem cell model of Barth syndrome. Stem Cell Res 11, 806-819.

Dudek, J., Cheng, I.F., Chowdhury, A., Wozny, K., Balleininger, M., Reinhold, R., Grunau, S., Callegari, S., Toischer, K., Wanders, R.J., et al. (2016). Cardiac-specific succinate dehydrogenase deficiency in Barth syndrome. EMBO Mol Med 8, 139-154.

E.L.;, J.L.O.E.S., and J.H., M.E.K.E. (1951). Recovery from Radiation Injury. Science 113, 510- 511.

Eble, K.S., Coleman, W.B., Hantgan, R.R., and Cunningham, C.C. (1990). Tightly associated cardiolipin in the bovine heart mitochondrial ATP synthase as analyzed by 31P nuclear magnetic resonance spectroscopy. J Biol Chem 265, 19434-19440.

Edwards, D.P., E.S.; Huang, X; Sun, Z; Capitano, M; Miller, C; He, Y; Broxmeyer, H.E.; Nalepa, G (2017). Patient-Tailored Mouse Genome Editing Recapitulates Hematopoietic and Systemic Manifestations of Barth Syndrome. Paper presented at: American Societ of Hematology Annual Meeting (Atlanta, GA: Blood).

Elstrom, R.L., Bauer, D.E., Buzzai, M., Karnauskas, R., Harris, M.H., Plas, D.R., Zhuang, H., Cinalli, R.M., Alavi, A., Rudin, C.M., et al. (2004). Akt stimulates aerobic glycolysis in cancer cells. Cancer Res 64, 3892-3899.

Esplin, B.L., Shimazu, T., Welner, R.S., Garrett, K.P., Nie, L., Zhang, Q., Humphrey, M.B., Yang, Q., Borghesi, L.A., and Kincade, P.W. (2011). Chronic exposure to a TLR ligand injures hematopoietic stem cells. J Immunol 186, 5367-5375.

Essers, M.A., Offner, S., Blanco-Bose, W.E., Waibler, Z., Kalinke, U., Duchosal, M.A., and Trumpp, A. (2009). IFNalpha activates dormant haematopoietic stem cells in vivo. Nature 458, 904-908.

Estefania Ochoa-Ruiz, R.D.-R. (2012). Anaplerosis in cancer: Another step beyond the warburg effect. American Journal of Molecular Biology 2.

Falkenberg, M., Larsson, N.G., and Gustafsson, C.M. (2007). DNA replication and transcription in mammalian mitochondria. Annu Rev Biochem 76, 679-699.

Faubert, B., Boily, G., Izreig, S., Griss, T., Samborska, B., Dong, Z., Dupuy, F., Chambers, C., Fuerth, B.J., Viollet, B., et al. (2013). AMPK is a negative regulator of the Warburg effect and suppresses tumor growth in vivo. Cell Metab 17, 113-124.

Faubert, B., Vincent, E.E., Poffenberger, M.C., and Jones, R.G. (2015). The AMP-activated protein kinase (AMPK) and cancer: many faces of a metabolic regulator. Cancer Lett 356, 165- 170.

Fefer, A., Sullivan, K.M., Weiden, P., Buckner, C.D., Schoch, G., Storb, R., and Thomas, E.D. (1987). Graft versus leukemia effect in man: the relapse rate of acute leukemia is lower after allogeneic than after syngeneic marrow transplantation. Prog Clin Biol Res 244, 401-408.

Ferrara, J.L., and Reddy, P. (2006). Pathophysiology of graft-versus-host disease. Semin Hematol 43, 3-10.

159

FOLCH, J., LEES, M., and SLOANE STANLEY, G.H. (1957). A simple method for the isolation and purification of total lipides from animal tissues. J Biol Chem 226, 497-509.

Friedman, J.R., Kannan, M., Toulmay, A., Jan, C.H., Weissman, J.S., Prinz, W.A., and Nunnari, J. (2017). Lipid Homeostasis Is Maintained by Dual Targeting of the Mitochondrial PE Biosynthesis Enzyme to the ER. Dev Cell.

Friedman, J.R., and Nunnari, J. (2014). Mitochondrial form and function. Nature 505, 335-343.

Fruman, D.A., and Rommel, C. (2014). PI3K and cancer: lessons, challenges and opportunities. Nat Rev Drug Discov 13, 140-156.

Gallipoli, P., Giotopoulos, G., Tzelepis, K., Costa, A.S.H., Vohra, S., Medina-Perez, P., Basheer, F., Marando, L., Di Lisio, L., Dias, J.M.L., et al. (2018). Glutaminolysis is a metabolic dependency in FLT3(ITD) acute myeloid leukemia unmasked by FLT3 tyrosine kinase inhibition. Blood 131, 1639-1653.

Gandhi, V.V., and Samuels, D.C. (2011). A review comparing deoxyribonucleoside triphosphate (dNTP) concentrations in the mitochondrial and cytoplasmic compartments of normal and transformed cells. Nucleosides Nucleotides Nucleic Acids 30, 317-339.

Gao, W., and Estey, E. (2015). Moving toward targeted therapies in acute myeloid leukemia. Clin Adv Hematol Oncol 13, 748-754.

Gebert, N., Joshi, A.S., Kutik, S., Becker, T., McKenzie, M., Guan, X.L., Mooga, V.P., Stroud, D.A., Kulkarni, G., Wenk, M.R., et al. (2009). Mitochondrial cardiolipin involved in outer- membrane protein biogenesis: implications for Barth syndrome. Curr Biol 19, 2133-2139.

Geisler, S., Holmstrom, K.M., Skujat, D., Fiesel, F.C., Rothfuss, O.C., Kahle, P.J., and Springer, W. (2010). PINK1/Parkin-mediated mitophagy is dependent on VDAC1 and p62/SQSTM1. Nat Cell Biol 12, 119-131.

Genovese, G., Kähler, A.K., Handsaker, R.E., Lindberg, J., Rose, S.A., Bakhoum, S.F., Chambert, K., Mick, E., Neale, B.M., Fromer, M., et al. (2014). Clonal hematopoiesis and blood- cancer risk inferred from blood DNA sequence. N Engl J Med 371, 2477-2487.

Gibson, C.J., Lindsley, R.C., Tchekmedyian, V., Mar, B.G., Shi, J., Jaiswal, S., Bosworth, A., Francisco, L., He, J., Bansal, A., et al. (2017). Clonal Hematopoiesis Associated With Adverse Outcomes After Autologous Stem-Cell Transplantation for Lymphoma. J Clin Oncol 35, 1598- 1605.

Gibson, C.J., and Steensma, D.P. (2018). New Insights from Studies of Clonal Hematopoiesis. Clin Cancer Res.

Gillis, N.K., Ball, M., Zhang, Q., Ma, Z., Zhao, Y., Yoder, S.J., Balasis, M.E., Mesa, T.E., Sallman, D.A., Lancet, J.E., et al. (2017). Clonal haemopoiesis and therapy-related myeloid malignancies in elderly patients: a proof-of-concept, case-control study. Lancet Oncol 18, 112- 121.

Gluckman, E., Rocha, V., Boyer-Chammard, A., Locatelli, F., Arcese, W., Pasquini, R., Ortega, J., Souillet, G., Ferreira, E., Laporte, J.P., et al. (1997). Outcome of cord-blood transplantation

160 from related and unrelated donors. Eurocord Transplant Group and the European Blood and Marrow Transplantation Group. N Engl J Med 337, 373-381.

Goardon, N., Marchi, E., Atzberger, A., Quek, L., Schuh, A., Soneji, S., Woll, P., Mead, A., Alford, K.A., Rout, R., et al. (2011). Coexistence of LMPP-like and GMP-like leukemia stem cells in acute myeloid leukemia. Cancer cell 19, 138-152.

Gonzalvez, F., D'Aurelio, M., Boutant, M., Moustapha, A., Puech, J.P., Landes, T., Arnauné- Pelloquin, L., Vial, G., Taleux, N., Slomianny, C., et al. (2013). Barth syndrome: cellular compensation of mitochondrial dysfunction and apoptosis inhibition due to changes in cardiolipin remodeling linked to tafazzin (TAZ) gene mutation. Biochim Biophys Acta 1832, 1194-1206.

Gonzalvez, F., Schug, Z.T., Houtkooper, R.H., MacKenzie, E.D., Brooks, D.G., Wanders, R.J., Petit, P.X., Vaz, F.M., and Gottlieb, E. (2008). Cardiolipin provides an essential activating platform for caspase-8 on mitochondria. J Cell Biol 183, 681-696.

Gorin, N.C., Labopin, M., Fouillard, L., Meloni, G., Frassoni, F., Iriondo, A., Brunet Mauri, S., Goldstone, A.H., Harousseau, J.L., Reiffers, J., et al. (1996). Retrospective evaluation of autologous bone marrow transplantation vs allogeneic bone marrow transplantation from an HLA identical related donor in acute myelocytic leukemia. A study of the European Cooperative Group for Blood and Marrow Transplantation (EBMT). Bone Marrow Transplant 18, 111-117.

Green, H. (2014). ATP and Respiration. In Biochemistry.

Gregory, M.A., Nemkov, T., Reisz, J.A., Zaberezhnyy, V., Hansen, K.C., D'Alessandro, A., and DeGregori, J. (2018). Glutaminase inhibition improves FLT3 inhibitor therapy for acute myeloid leukemia. Exp Hematol 58, 52-58.

Gu, W., Gaeta, X., Sahakyan, A., Chan, A.B., Hong, C.S., Kim, R., Braas, D., Plath, K., Lowry, W.E., and Christofk, H.R. (2016). Glycolytic Metabolism Plays a Functional Role in Regulating Human Pluripotent Stem Cell State. Cell Stem Cell 19, 476-490.

Gu, Z., Valianpour, F., Chen, S., Vaz, F.M., Hakkaart, G.A., Wanders, R.J., and Greenberg, M.L. (2004). Aberrant cardiolipin metabolism in the yeast taz1 mutant: a model for Barth syndrome. Mol Microbiol 51, 149-158.

Gudas, L.J., and Wagner, J.A. (2011). Retinoids regulate stem cell differentiation. J Cell Physiol 226, 322-330.

Guryanova, O.A., Shank, K., Spitzer, B., Luciani, L., Koche, R.P., Garrett-Bakelman, F.E., Ganzel, C., Durham, B.H., Mohanty, A., Hoermann, G., et al. (2016). DNMT3A mutations promote anthracycline resistance in acute myeloid leukemia via impaired nucleosome remodeling. Nat Med 22, 1488-1495.

Gyurkocza, B., and Sandmaier, B.M. (2014). Conditioning regimens for hematopoietic cell transplantation: one size does not fit all. Blood 124, 344-353.

Haikala, H.M., Anttila, J.M., and Klefstrom, J. (2017). MYC and AMPK-Save Energy or Die! Front Cell Dev Biol 5, 38.

161

Handy, D.E., and Loscalzo, J. (2012). Redox regulation of mitochondrial function. Antioxidants & redox signaling 16, 1323-1367.

Hart, T., Chandrashekhar, M., Aregger, M., Steinhart, Z., Brown, K.R., MacLeod, G., Mis, M., Zimmermann, M., Fradet-Turcotte, A., Sun, S., et al. (2015). High-Resolution CRISPR Screens Reveal Fitness Genes and Genotype-Specific Cancer Liabilities. Cell 163, 1515-1526.

Hatzivassiliou, G., Zhao, F., Bauer, D.E., Andreadis, C., Shaw, A.N., Dhanak, D., Hingorani, S.R., Tuveson, D.A., and Thompson, C.B. (2005). ATP citrate lyase inhibition can suppress tumor cell growth. Cancer cell 8, 311-321.

Haurie, C., Dale, D.C., and Mackey, M.C. (1998). Cyclical neutropenia and other periodic hematological disorders: a review of mechanisms and mathematical models. Blood 92, 2629- 2640.

Hayashi, E.A., Granato, A., Paiva, L.S., Bertho, A.L., Bellio, M., and Nobrega, A. (2010). TLR4 promotes B cell maturation: independence and cooperation with B lymphocyte-activating factor. J Immunol 184, 4662-4672.

Heyer, W.D., Ehmsen, K.T., and Liu, J. (2010). Regulation of homologous recombination in eukaryotes. Annu Rev Genet 44, 113-139.

Hilgendorf, I., Greinix, H., Halter, J.P., Lawitschka, A., Bertz, H., and Wolff, D. (2015). Long- term follow-up after allogeneic stem cell transplantation. Dtsch Arztebl Int 112, 51-58.

Hillen, H.S., Parshin, A.V., Agaronyan, K., Morozov, Y.I., Graber, J.J., Chernev, A., Schwinghammer, K., Urlaub, H., Anikin, M., Cramer, P., et al. (2017). Mechanism of Transcription Anti-termination in Human Mitochondria. Cell 171, 1082-1093 e1013.

Ho, V.T., Revta, C., and Richardson, P.G. (2008). Hepatic veno-occlusive disease after hematopoietic stem cell transplantation: update on defibrotide and other current investigational therapies. Bone Marrow Transplant 41, 229-237.

Horowitz, M.M., Gale, R.P., Sondel, P.M., Goldman, J.M., Kersey, J., Kolb, H.J., Rimm, A.A., Ringden, O., Rozman, C., Speck, B., et al. (1990). Graft-versus-leukemia reactions after bone marrow transplantation. Blood 75, 555-562.

Hosen, N., Park, C.Y., Tatsumi, N., Oji, Y., Sugiyama, H., Gramatzki, M., Krensky, A.M., and Weissman, I.L. (2007). CD96 is a leukemic stem cell-specific marker in human acute myeloid leukemia. Proc Natl Acad Sci U S A 104, 11008-11013.

Houtkooper, R.H., Rodenburg, R.J., Thiels, C., van Lenthe, H., Stet, F., Poll-The, B.T., Stone, J.E., Steward, C.G., Wanders, R.J., Smeitink, J., et al. (2009). Cardiolipin and monolysocardiolipin analysis in fibroblasts, lymphocytes, and tissues using high-performance liquid chromatography-mass spectrometry as a diagnostic test for Barth syndrome. Anal Biochem 387, 230-237.

Hsu, P.D., Lander, E.S., and Zhang, F. (2014). Development and applications of CRISPR-Cas9 for genome engineering. Cell 157, 1262-1278.

162

Huang, B.X., Akbar, M., Kevala, K., and Kim, H.Y. (2011). Phosphatidylserine is a critical modulator for Akt activation. J Cell Biol 192, 979-992.

Huang, M.E., Ye, Y.C., Chen, S.R., Chai, J.R., Lu, J.X., Zhoa, L., Gu, L.J., and Wang, Z.Y. (1988). Use of all-trans retinoic acid in the treatment of acute promyelocytic leukemia. Blood 72, 567-572.

Huang, Y., Powers, C., Madala, S.K., Greis, K.D., Haffey, W.D., Towbin, J.A., Purevjav, E., Javadov, S., Strauss, A.W., and Khuchua, Z. (2015). Cardiac metabolic pathways affected in the mouse model of barth syndrome. PLoS One 10, e0128561.

Hulegårdh, E., Nilsson, C., Lazarevic, V., Garelius, H., Antunovic, P., Rangert Derolf, Å., Möllgård, L., Uggla, B., Wennström, L., Wahlin, A., et al. (2015). Characterization and prognostic features of secondary acute myeloid leukemia in a population-based setting: a report from the Swedish Acute Leukemia Registry. Am J Hematol 90, 208-214.

Ignatz-Hoover, J.J., Wang, H., Moreton, S.A., Chakrabarti, A., Agarwal, M.K., Sun, K., Gupta, K., and Wald, D.N. (2015). The role of TLR8 signaling in acute myeloid leukemia differentiation. Leukemia 29, 918-926.

Inamoto, Y., and Lee, S.J. (2017). Late effects of blood and marrow transplantation. Haematologica 102, 614-625.

Inoue, D., Bradley, R.K., and Abdel-Wahab, O. (2016). Spliceosomal gene mutations in myelodysplasia: molecular links to clonal abnormalities of hematopoiesis. Genes Dev 30, 989- 1001.

Insinga, A., Monestiroli, S., Ronzoni, S., Carbone, R., Pearson, M., Pruneri, G., Viale, G., Appella, E., Pelicci, P., and Minucci, S. (2004). Impairment of p53 acetylation, stability and function by an oncogenic transcription factor. EMBO J 23, 1144-1154.

Ishizawa, J., Zarabi, S.F., Davis, R.E., Halgas, O., Nii, T., Jitkova, Y., Zhao, R., St-Germain, J., Heese, L.E., Egan, G., et al. (2019). Mitochondrial ClpP-Mediated Proteolysis Induces Selective Cancer Cell Lethality. Cancer cell 35, 721-737 e729.

Ito, S., Shen, L., Dai, Q., Wu, S.C., Collins, L.B., Swenberg, J.A., He, C., and Zhang, Y. (2011). Tet proteins can convert 5-methylcytosine to 5-formylcytosine and 5-carboxylcytosine. Science 333, 1300-1303.

Iwasaki, M., Liedtke, M., Gentles, A.J., and Cleary, M.L. (2015). CD93 Marks a Non-Quiescent Human Leukemia Stem Cell Population and Is Required for Development of MLL-Rearranged Acute Myeloid Leukemia. Cell Stem Cell 17, 412-421.

Jacobs, J.L., and Coyne, C.B. (2013). Mechanisms of MAVS regulation at the mitochondrial membrane. J Mol Biol 425, 5009-5019.

Jacobson, L.O., Simmons, E.L., Marks, E.K., Robson, M.J., Bethard, W.F., and Gaston, E.O. (1950). The role of the spleen in radiation injury and recovery. J Lab Clin Med 35, 746-770.

163

Jacque, N., Ronchetti, A.M., Larrue, C., Meunier, G., Birsen, R., Willems, L., Saland, E., Decroocq, J., Maciel, T.T., Lambert, M., et al. (2015). Targeting glutaminolysis has antileukemic activity in acute myeloid leukemia and synergizes with BCL-2 inhibition. Blood 126, 1346-1356.

Jaiswal, S., Fontanillas, P., Flannick, J., Manning, A., Grauman, P.V., Mar, B.G., Lindsley, R.C., Mermel, C.H., Burtt, N., Chavez, A., et al. (2014). Age-related clonal hematopoiesis associated with adverse outcomes. N Engl J Med 371, 2488-2498.

Jaiswal, S., Jamieson, C.H., Pang, W.W., Park, C.Y., Chao, M.P., Majeti, R., Traver, D., van Rooijen, N., and Weissman, I.L. (2009). CD47 is upregulated on circulating hematopoietic stem cells and leukemia cells to avoid phagocytosis. Cell 138, 271-285.

Jaiswal, S., Natarajan, P., Silver, A.J., Gibson, C.J., Bick, A.G., Shvartz, E., McConkey, M., Gupta, N., Gabriel, S., Ardissino, D., et al. (2017). Clonal Hematopoiesis and Risk of Atherosclerotic Cardiovascular Disease. N Engl J Med 377, 111-121.

Jan, M., Chao, M.P., Cha, A.C., Alizadeh, A.A., Gentles, A.J., Weissman, I.L., and Majeti, R. (2011). Prospective separation of normal and leukemic stem cells based on differential expression of TIM3, a human acute myeloid leukemia stem cell marker. Proc Natl Acad Sci U S A 108, 5009-5014.

Jeon, S.M. (2016). Regulation and function of AMPK in physiology and diseases. Exp Mol Med 48, e245.

Jewell, J.L., Kim, Y.C., Russell, R.C., Yu, F.X., Park, H.W., Plouffe, S.W., Tagliabracci, V.S., and Guan, K.L. (2015). Metabolism. Differential regulation of mTORC1 by leucine and glutamine. Science 347, 194-198.

Jhanwar-Uniyal, M., Wainwright, J.V., Mohan, A.L., Tobias, M.E., Murali, R., Gandhi, C.D., and Schmidt, M.H. (2019). Diverse signaling mechanisms of mTOR complexes: mTORC1 and mTORC2 in forming a formidable relationship. Adv Biol Regul 72, 51-62.

Jiang, F., Ryan, M.T., Schlame, M., Zhao, M., Gu, Z., Klingenberg, M., Pfanner, N., and Greenberg, M.L. (2000). Absence of cardiolipin in the crd1 null mutant results in decreased mitochondrial membrane potential and reduced mitochondrial function. J Biol Chem 275, 22387- 22394.

Jiang, F., Zhang, Y., and Dusting, G.J. (2011). NADPH oxidase-mediated redox signaling: roles in cellular stress response, stress tolerance, and tissue repair. Pharmacol Rev 63, 218-242.

Jin, L., Hope, K.J., Zhai, Q., Smadja-Joffe, F., and Dick, J.E. (2006). Targeting of CD44 eradicates human acute myeloid leukemic stem cells. Nat Med 12, 1167-1174.

Jin, L., Lee, E.M., Ramshaw, H.S., Busfield, S.J., Peoppl, A.G., Wilkinson, L., Guthridge, M.A., Thomas, D., Barry, E.F., Boyd, A., et al. (2009). Monoclonal antibody-mediated targeting of CD123, IL-3 receptor alpha chain, eliminates human acute myeloid leukemic stem cells. Cell Stem Cell 5, 31-42.

Jones, C.L., Stevens, B.M., D'Alessandro, A., Culp-Hill, R., Reisz, J.A., Pei, S., Gustafson, A., Khan, N., DeGregori, J., Pollyea, D.A., et al. (2019). Cysteine depletion targets leukemia stem cells through inhibition of electron transport complex II. Blood.

164

Jones, C.L., Stevens, B.M., D'Alessandro, A., Reisz, J.A., Culp-Hill, R., Nemkov, T., Pei, S., Khan, N., Adane, B., Ye, H., et al. (2018). Inhibition of Amino Acid Metabolism Selectively Targets Human Leukemia Stem Cells. Cancer cell 34, 724-740.e724.

Jussupow, A., Di Luca, A., and Kaila, V.R.I. (2019). How cardiolipin modulates the dynamics of respiratory complex I. Sci Adv 5, eaav1850.

Kanate, A.S., Pasquini, M.C., Hari, P.N., and Hamadani, M. (2014). Allogeneic hematopoietic cell transplant for acute myeloid leukemia: Current state in 2013 and future directions. World J Stem Cells 6, 69-81.

Kasahara, A., and Scorrano, L. (2014). Mitochondria: from cell death executioners to regulators of cell differentiation. Trends Cell Biol 24, 761-770.

Kavanagh, S., Murphy, T., Law, A., Yehudai, D., Ho, J.M., Chan, S., and Schimmer, A.D. (2017). Emerging therapies for acute myeloid leukemia: translating biology into the clinic. JCI Insight 2.

Kawasaki, T., and Kawai, T. (2014). Toll-like receptor signaling pathways. Front Immunol 5, 461.

Kay, J.G., and Grinstein, S. (2013). Phosphatidylserine-mediated cellular signaling. Adv Exp Med Biol 991, 177-193.

Keckesova, Z., Donaher, J.L., De Cock, J., Freinkman, E., Lingrell, S., Bachovchin, D.A., Bierie, B., Tischler, V., Noske, A., Okondo, M.C., et al. (2017). LACTB is a tumour suppressor that modulates lipid metabolism and cell state. Nature 543, 681-686.

Kelly, K., and Jacobs, R. (2018). Phospholipid Biosynthesis (AOCS Lipid Library).

Kiel, M.J., Yilmaz, O.H., Iwashita, T., Yilmaz, O.H., Terhorst, C., and Morrison, S.J. (2005). SLAM family receptors distinguish hematopoietic stem and progenitor cells and reveal endothelial niches for stem cells. Cell 121, 1109-1121.

Kikushige, Y., Shima, T., Takayanagi, S., Urata, S., Miyamoto, T., Iwasaki, H., Takenaka, K., Teshima, T., Tanaka, T., Inagaki, Y., et al. (2010). TIM-3 is a promising target to selectively kill acute myeloid leukemia stem cells. Cell Stem Cell 7, 708-717.

Kim, J., Kundu, M., Viollet, B., and Guan, K.L. (2011). AMPK and mTOR regulate autophagy through direct phosphorylation of Ulk1. Nat Cell Biol 13, 132-141.

Kim, J.W., Zeller, K.I., Wang, Y., Jegga, A.G., Aronow, B.J., O'Donnell, K.A., and Dang, C.V. (2004). Evaluation of myc E-box phylogenetic footprints in glycolytic genes by chromatin immunoprecipitation assays. Mol Cell Biol 24, 5923-5936.

Koberlin, M.S., Snijder, B., Heinz, L.X., Baumann, C.L., Fauster, A., Vladimer, G.I., Gavin, A.C., and Superti-Furga, G. (2015). A Conserved Circular Network of Coregulated Lipids Modulates Innate Immune Responses. Cell 162, 170-183.

165

Koc, E.C., and Spremulli, L.L. (2002). Identification of mammalian mitochondrial translational initiation factor 3 and examination of its role in initiation complex formation with natural mRNAs. J Biol Chem 277, 35541-35549.

Konstantinov, I.E. (2000). In search of Alexander A. Maximow: the man behind the unitarian theory of hematopoiesis. Perspect Biol Med 43, 269-276.

Koreth, J., Schlenk, R., Kopecky, K.J., Honda, S., Sierra, J., Djulbegovic, B.J., Wadleigh, M., DeAngelo, D.J., Stone, R.M., Sakamaki, H., et al. (2009). Allogeneic stem cell transplantation for acute myeloid leukemia in first complete remission: systematic review and meta-analysis of prospective clinical trials. JAMA 301, 2349-2361.

Korhonen, J.A., Gaspari, M., and Falkenberg, M. (2003). TWINKLE Has 5' -> 3' DNA helicase activity and is specifically stimulated by mitochondrial single-stranded DNA-binding protein. J Biol Chem 278, 48627-48632.

Kreso, A., and Dick, J.E. (2014). Evolution of the cancer stem cell model. Cell Stem Cell 14, 275-291.

Lagadinou, E.D., Sach, A., Callahan, K., Rossi, R.M., Neering, S.J., Minhajuddin, M., Ashton, J.M., Pei, S., Grose, V., O'Dwyer, K.M., et al. (2013). BCL-2 inhibition targets oxidative phosphorylation and selectively eradicates quiescent human leukemia stem cells. Cell Stem Cell 12, 329-341.

Lane, A.N., and Fan, T.W. (2015). Regulation of mammalian nucleotide metabolism and biosynthesis. Nucleic Acids Res 43, 2466-2485.

Lane, N., and Martin, W. (2010). The energetics of genome complexity. Nature 467, 929-934.

Lapidot, T., Sirard, C., Vormoor, J., Murdoch, B., Hoang, T., Caceres-Cortes, J., Minden, M., Paterson, B., Caligiuri, M.A., and Dick, J.E. (1994). A cell initiating human acute myeloid leukaemia after transplantation into SCID mice. Nature 367, 645-648.

Lechman, E.R., Gentner, B., Ng, S.W., Schoof, E.M., van Galen, P., Kennedy, J.A., Nucera, S., Ciceri, F., Kaufmann, K.B., Takayama, N., et al. (2016). miR-126 Regulates Distinct Self- Renewal Outcomes in Normal and Malignant Hematopoietic Stem Cells. Cancer cell 29, 214- 228.

Lee, E.A., Angka, L., Rota, S.G., Hanlon, T., Mitchell, A., Hurren, R., Wang, X.M., Gronda, M., Boyaci, E., Bojko, B., et al. (2015). Targeting Mitochondria with Avocatin B Induces Selective Leukemia Cell Death. Cancer Res 75, 2478-2488.

Lee, S.J., Vogelsang, G., and Flowers, M.E. (2003). Chronic graft-versus-host disease. Biol Blood Marrow Transplant 9, 215-233.

Lee, W.T., and St John, J. (2015). The control of mitochondrial DNA replication during development and tumorigenesis. Ann N Y Acad Sci 1350, 95-106.

Li, J., Romestaing, C., Han, X., Li, Y., Hao, X., Wu, Y., Sun, C., Liu, X., Jefferson, L.S., Xiong, J., et al. (2010). Cardiolipin remodeling by ALCAT1 links oxidative stress and mitochondrial dysfunction to obesity. Cell Metab 12, 154-165.

166

Li, W., Xu, H., Xiao, T., Cong, L., Love, M.I., Zhang, F., Irizarry, R.A., Liu, J.S., Brown, M., and Liu, X.S. (2014). MAGeCK enables robust identification of essential genes from genome- scale CRISPR/Cas9 knockout screens. Genome Biol 15, 554.

Linehan, M.M., Dickey, T.H., Molinari, E.S., Fitzgerald, M.E., Potapova, O., Iwasaki, A., and Pyle, A.M. (2018). A minimal RNA ligand for potent RIG-I activation in living mice. Sci Adv 4, e1701854.

Liyanage, S.U. (2017). TARGETING THE MITOCHONDRIAL DNA POLYMERASE GAMMA IN ACUTE MYELOID LEUKEMIA. In Medical Biophysics (Tspace: University of Toronto), pp. 162.

Liyanage, S.U., Hurren, R., Voisin, V., Bridon, G., Wang, X., Xu, C., MacLean, N., Siriwardena, T.P., Gronda, M., Yehudai, D., et al. (2017). Leveraging increased cytoplasmic nucleoside kinase activity to target mtDNA and oxidative phosphorylation in AML. Blood 129, 2657-2666.

Ljungman, P., Hakki, M., and Boeckh, M. (2011). Cytomegalovirus in hematopoietic stem cell transplant recipients. Hematol Oncol Clin North Am 25, 151-169.

Lo-Coco, F., Avvisati, G., Vignetti, M., Thiede, C., Orlando, S.M., Iacobelli, S., Ferrara, F., Fazi, P., Cicconi, L., Di Bona, E., et al. (2013). Retinoic acid and arsenic trioxide for acute promyelocytic leukemia. N Engl J Med 369, 111-121.

Lo-Coco, F., Di Donato, L., Gimema, Schlenk, R.F., German-Austrian Acute Myeloid Leukemia Study, G., and Study Alliance, L. (2016). Targeted Therapy Alone for Acute Promyelocytic Leukemia. N Engl J Med 374, 1197-1198.

Lorenz, E., Congdon, C., and Uphoff, D. (1952). Modification of acute irradiation injury in mice and guinea-pigs by bone marrow injections. Radiology 58, 863-877.

Loson, O.C., Song, Z., Chen, H., and Chan, D.C. (2013). Fis1, Mff, MiD49, and MiD51 mediate Drp1 recruitment in mitochondrial fission. Mol Biol Cell 24, 659-667.

Lowenberg, B., Downing, J.R., and Burnett, A. (1999). Acute myeloid leukemia. N Engl J Med 341, 1051-1062.

Lu, Y.W., and Claypool, S.M. (2015). Disorders of phospholipid metabolism: an emerging class of mitochondrial disease due to defects in nuclear genes. Front Genet 6, 3.

Lu, Y.W., Galbraith, L., Herndon, J.D., Lu, Y.L., Pras-Raves, M., Vervaart, M., Van Kampen, A., Luyf, A., Koehler, C.M., McCaffery, J.M., et al. (2016). Defining functional classes of Barth syndrome mutation in humans. Hum Mol Genet 25, 1754-1770.

Lucantoni, L., Duffy, S., Adjalley, S.H., Fidock, D.A., and Avery, V.M. (2013). Identification of MMV malaria box inhibitors of plasmodium falciparum early-stage gametocytes using a luciferase-based high-throughput assay. Antimicrob Agents Chemother 57, 6050-6062.

Lunt, S.Y., and Vander Heiden, M.G. (2011). Aerobic glycolysis: meeting the metabolic requirements of cell proliferation. Annu Rev Cell Dev Biol 27, 441-464.

167

Maehle, A.H. (2011). Ambiguous cells: the emergence of the stem cell concept in the nineteenth and twentieth centuries. Notes Rec R Soc Lond 65, 359-378.

Majeti, R., Chao, M.P., Alizadeh, A.A., Pang, W.W., Jaiswal, S., Gibbs, K.D., van Rooijen, N., and Weissman, I.L. (2009). CD47 is an adverse prognostic factor and therapeutic antibody target on human acute myeloid leukemia stem cells. Cell 138, 286-299.

Majhail, N.S. (2017). Long-term complications after hematopoietic cell transplantation. Hematol Oncol Stem Cell Ther.

Majhail, N.S., Rizzo, J.D., Lee, S.J., Aljurf, M., Atsuta, Y., Bonfim, C., Burns, L.J., Chaudhri, N., Davies, S., Okamoto, S., et al. (2012). Recommended screening and preventive practices for long-term survivors after hematopoietic cell transplantation. Rev Bras Hematol Hemoter 34, 109-133.

Makaryan, V., Kulik, W., Vaz, F.M., Allen, C., Dror, Y., Dale, D.C., and Aprikyan, A.A. (2012). The cellular and molecular mechanisms for neutropenia in Barth syndrome. Eur J Haematol 88, 195-209.

Manz, M.G., and Boettcher, S. (2014). Emergency granulopoiesis. Nat Rev Immunol 14, 302- 314.

Marjanovic, S., Wielburski, A., and Nelson, B.D. (1988). Effect of phorbol myristate acetate and concanavalin A on the glycolytic enzymes of human peripheral lymphocytes. Biochim Biophys Acta 970, 1-6.

Martens, J.H., Brinkman, A.B., Simmer, F., Francoijs, K.J., Nebbioso, A., Ferrara, F., Altucci, L., and Stunnenberg, H.G. (2010). PML-RARalpha/RXR Alters the Epigenetic Landscape in Acute Promyelocytic Leukemia. Cancer cell 17, 173-185.

Martin, M., Cho, J., Cesare, A.J., Griffith, J.D., and Attardi, G. (2005). Termination factor- mediated DNA loop between termination and initiation sites drives mitochondrial rRNA synthesis. Cell 123, 1227-1240.

Mayle, A., Yang, L., Rodriguez, B., Zhou, T., Chang, E., Curry, C.V., Challen, G.A., Li, W., Wheeler, D., Rebel, V.I., et al. (2015). Dnmt3a loss predisposes murine hematopoietic stem cells to malignant transformation. Blood 125, 629-638.

McBride, A., Houtmann, S., Wilde, L., Vigil, C., Eischen, C.M., Kasner, M., and Palmisiano, N. (2019). The Role of Inhibition of Apoptosis in Acute Leukemias and Myelodysplastic Syndrome. Front Oncol 9, 192.

Megias, J., Yanez, A., Moriano, S., O'Connor, J.E., Gozalbo, D., and Gil, M.L. (2012). Direct Toll-like receptor-mediated stimulation of hematopoietic stem and progenitor cells occurs in vivo and promotes differentiation toward macrophages. Stem Cells 30, 1486-1495.

Meglei, G., and McQuibban, G.A. (2009). The dynamin-related protein Mgm1p assembles into oligomers and hydrolyzes GTP to function in mitochondrial membrane fusion. Biochemistry 48, 1774-1784.

168

Mejia, E.M., Ibdah, J.A., Sparagna, G.C., and Hatch, G.M. (2015). Differential reduction in cardiac and liver monolysocardiolipin acyltransferase-1 and reduction in cardiac and liver tetralinoleoyl-cardiolipin in the α-subunit of trifunctional protein heterozygous knockout mice. Biochem J 471, 123-129.

Mejia, E.M., Nguyen, H., and Hatch, G.M. (2014). Mammalian cardiolipin biosynthesis. Chem Phys Lipids 179, 11-16.

Meletis, J., and Konstantopoulos, K. (2010). The beliefs, myths, and reality surrounding the word hema (blood) from homer to the present. Anemia 2010, 857657.

Minczuk, M., He, J., Duch, A.M., Ettema, T.J., Chlebowski, A., Dzionek, K., Nijtmans, L.G., Huynen, M.A., and Holt, I.J. (2011). TEFM (c17orf42) is necessary for transcription of human mtDNA. Nucleic Acids Res 39, 4284-4299.

Minton, K. (2010). Viral immunity: How CMV bypasses immune memory. Nat Rev Immunol 10, 288.

Miraki-Moud, F., Ghazaly, E., Ariza-McNaughton, L., Hodby, K.A., Clear, A., Anjos-Afonso, F., Liapis, K., Grantham, M., Sohrabi, F., Cavenagh, J., et al. (2015). Arginine deprivation using pegylated arginine deiminase has activity against primary acute myeloid leukemia cells in vivo. Blood 125, 4060-4068.

Mitroulis, I., Chen, L.S., Singh, R.P., Kourtzelis, I., Economopoulou, M., Kajikawa, T., Troullinaki, M., Ziogas, A., Ruppova, K., Hosur, K., et al. (2017). Secreted protein Del-1 regulates myelopoiesis in the hematopoietic stem cell niche. J Clin Invest 127, 3624-3639.

Moffat, J., Grueneberg, D.A., Yang, X., Kim, S.Y., Kloepfer, A.M., Hinkle, G., Piqani, B., Eisenhaure, T.M., Luo, B., Grenier, J.K., et al. (2006). A lentiviral RNAi library for human and mouse genes applied to an arrayed viral high-content screen. Cell 124, 1283-1298.

Molina, J.R., Sun, Y., Protopopova, M., Gera, S., Bandi, M., Bristow, C., McAfoos, T., Morlacchi, P., Ackroyd, J., Agip, A.A., et al. (2018). An inhibitor of oxidative phosphorylation exploits cancer vulnerability. Nat Med.

Montessuit, S., Somasekharan, S.P., Terrones, O., Lucken-Ardjomande, S., Herzig, S., Schwarzenbacher, R., Manstein, D.J., Bossy-Wetzel, E., Basanez, G., Meda, P., et al. (2010). Membrane remodeling induced by the dynamin-related protein Drp1 stimulates Bax oligomerization. Cell 142, 889-901.

Moore, M.A.S., Williams, N., and Metcalf, D. (1973). In Vitro Colony Formation by Normal and Leukemic Human Hematopoietic Cells: Characterization of the Colony-Forming Cells. Journal of the National Cancer Institute 50, 603–623.

Moore, T.I.A. (2017). Bone Marrow Transplantation (Medscape).

Moran-Crusio, K., Reavie, L., Shih, A., Abdel-Wahab, O., Ndiaye-Lobry, D., Lobry, C., Figueroa, M.E., Vasanthakumar, A., Patel, J., Zhao, X., et al. (2011). Tet2 loss leads to increased hematopoietic stem cell self-renewal and myeloid transformation. Cancer cell 20, 11-24.

169

Moshaver, B., van Rhenen, A., Kelder, A., van der Pol, M., Terwijn, M., Bachas, C., Westra, A.H., Ossenkoppele, G.J., Zweegman, S., and Schuurhuis, G.J. (2008). Identification of a small subpopulation of candidate leukemia-initiating cells in the side population of patients with acute myeloid leukemia. Stem Cells 26, 3059-3067.

Mukherjee, S. (2010). Emperor of All Maladies: A Biography of Cancer. (New York: Scribner ).

Nagai, Y., Garrett, K.P., Ohta, S., Bahrun, U., Kouro, T., Akira, S., Takatsu, K., and Kincade, P.W. (2006). Toll-like receptors on hematopoietic progenitor cells stimulate innate immune system replenishment. Immunity 24, 801-812.

Negrin, R.S. (2014). Introduction to the review series on "Advances in hematopoietic cell transplantation". Blood 124, 307.

Ng, S.W., Mitchell, A., Kennedy, J.A., Chen, W.C., McLeod, J., Ibrahimova, N., Arruda, A., Popescu, A., Gupta, V., Schimmer, A.D., et al. (2016). A 17-gene stemness score for rapid determination of risk in acute leukaemia. Nature 540, 433-437.

Nicolini, F.E., Cashman, J.D., Hogge, D.E., Humphries, R.K., and Eaves, C.J. (2004). NOD/SCID mice engineered to express human IL-3, GM-CSF and Steel factor constitutively mobilize engrafted human progenitors and compromise human stem cell regeneration. Leukemia 18, 341-347.

Notta, F., Zandi, S., Takayama, N., Dobson, S., Gan, O.I., Wilson, G., Kaufmann, K.B., McLeod, J., Laurenti, E., Dunant, C.F., et al. (2016). Distinct routes of lineage development reshape the human blood hierarchy across ontogeny. Science 351, aab2116.

Novershtern, N., Subramanian, A., Lawton, L.N., Mak, R.H., Haining, W.N., McConkey, M.E., Habib, N., Yosef, N., Chang, C.Y., Shay, T., et al. (2011). Densely interconnected transcriptional circuits control cell states in human hematopoiesis. Cell 144, 296-309.

Oguro, H., Ding, L., and Morrison, S.J. (2013). SLAM family markers resolve functionally distinct subpopulations of hematopoietic stem cells and multipotent progenitors. Cell Stem Cell 13, 102-116.

Ortmann, C.A., Kent, D.G., Nangalia, J., Silber, Y., Wedge, D.C., Grinfeld, J., Baxter, E.J., Massie, C.E., Papaemmanuil, E., Menon, S., et al. (2015). Effect of mutation order on myeloproliferative neoplasms. N Engl J Med 372, 601-612.

Ossenkoppele, G.J., Janssen, J.J., and van de Loosdrecht, A.A. (2016). Risk factors for relapse after allogeneic transplantation in acute myeloid leukemia. Haematologica 101, 20-25.

Osthus, R.C., Shim, H., Kim, S., Li, Q., Reddy, R., Mukherjee, M., Xu, Y., Wonsey, D., Lee, L.A., and Dang, C.V. (2000). Deregulation of glucose transporter 1 and glycolytic gene expression by c-Myc. J Biol Chem 275, 21797-21800.

Owen, O.E., Kalhan, S.C., and Hanson, R.W. (2002). The key role of anaplerosis and cataplerosis for citric acid cycle function. J Biol Chem 277, 30409-30412.

Owen, R.D. (1945). Immunogenetic Consequences of Vascular Anastomoses between Bovine Twins. Science 102, 400-401.

170

Pabst, C., Bergeron, A., Lavallée, V.P., Yeh, J., Gendron, P., Norddahl, G.L., Krosl, J., Boivin, I., Deneault, E., Simard, J., et al. (2016). GPR56 identifies primary human acute myeloid leukemia cells with high repopulating potential in vivo. Blood 127, 2018-2027.

Papaemmanuil, E., Gerstung, M., Bullinger, L., Gaidzik, V.I., Paschka, P., Roberts, N.D., Potter, N.E., Heuser, M., Thol, F., Bolli, N., et al. (2016). Genomic Classification and Prognosis in Acute Myeloid Leukemia. N Engl J Med 374, 2209-2221.

Paradies, G., Paradies, V., Ruggiero, F.M., and Petrosillo, G. (2014). Cardiolipin and mitochondrial function in health and disease. Antioxidants & redox signaling 20, 1925-1953.

Pasquini, M.C., and Wang, Z. (2013). Current use and outcome of hematopoietic stem cell transplantation: CIBMTR Summary Slides (Center for Internaitonal Blood and Marrow Transplant Research).

Patel, J.P., Gonen, M., Figueroa, M.E., Fernandez, H., Sun, Z., Racevskis, J., Van Vlierberghe, P., Dolgalev, I., Thomas, S., Aminova, O., et al. (2012). Prognostic relevance of integrated genetic profiling in acute myeloid leukemia. N Engl J Med 366, 1079-1089.

Pearce, D.J., Taussig, D., Zibara, K., Smith, L.L., Ridler, C.M., Preudhomme, C., Young, B.D., Rohatiner, A.Z., Lister, T.A., and Bonnet, D. (2006). AML engraftment in the NOD/SCID assay reflects the outcome of AML: implications for our understanding of the heterogeneity of AML. Blood 107, 1166-1173.

Pedley, A.M., and Benkovic, S.J. (2017). A New View into the Regulation of Purine Metabolism: The Purinosome. Trends Biochem Sci 42, 141-154.

Pei, S., Minhajuddin, M., Adane, B., Khan, N., Stevens, B.M., Mack, S.C., Lai, S., Rich, J.N., Inguva, A., Shannon, K.M., et al. (2018). AMPK/FIS1-Mediated Mitophagy Is Required for Self-Renewal of Human AML Stem Cells. Cell Stem Cell.

Pei, S., Minhajuddin, M., Callahan, K.P., Balys, M., Ashton, J.M., Neering, S.J., Lagadinou, E.D., Corbett, C., Ye, H., Liesveld, J.L., et al. (2013). Targeting aberrant glutathione metabolism to eradicate human acute myelogenous leukemia cells. J Biol Chem 288, 33542-33558.

Pietras, E.M., Mirantes-Barbeito, C., Fong, S., Loeffler, D., Kovtonyuk, L.V., Zhang, S., Lakshminarasimhan, R., Chin, C.P., Techner, J.M., Will, B., et al. (2016). Chronic interleukin-1 exposure drives haematopoietic stem cells towards precocious myeloid differentiation at the expense of self-renewal. Nat Cell Biol 18, 607-618.

Pillay, D., Webster, A., Prentice, H.G., and Griffiths, P.D. (1992). Risk factors for viral reactivation following bone marrow transplantation. Ann Hematol 64 Suppl, A148-151.

Pimenta de Castro, I., Costa, A.C., Lam, D., Tufi, R., Fedele, V., Moisoi, N., Dinsdale, D., Deas, E., Loh, S.H., and Martins, L.M. (2012). Genetic analysis of mitochondrial protein misfolding in Drosophila melanogaster. Cell Death Differ 19, 1308-1316.

Pollyea, D.A., Stevens, B.M., Jones, C.L., Winters, A., Pei, S., Minhajuddin, M., D'Alessandro, A., Culp-Hill, R., Riemondy, K.A., Gillen, A.E., et al. (2018). Venetoclax with azacitidine disrupts energy metabolism and targets leukemia stem cells in patients with acute myeloid leukemia. Nat Med 24, 1859-1866.

171

Posse, V., and Gustafsson, C.M. (2017). Human Mitochondrial Transcription Factor B2 Is Required for Promoter Melting during Initiation of Transcription. J Biol Chem 292, 2637-2645.

Qiao, W., Quon, G., Csaszar, E., Yu, M., Morris, Q., and Zandstra, P.W. (2012). PERT: a method for expression deconvolution of human blood samples from varied microenvironmental and developmental conditions. PLoS Comput Biol 8, e1002838.

Quek, L., Otto, G.W., Garnett, C., Lhermitte, L., Karamitros, D., Stoilova, B., Lau, I.J., Doondeea, J., Usukhbayar, B., Kennedy, A., et al. (2016). Genetically distinct leukemic stem cells in human CD34- acute myeloid leukemia are arrested at a hemopoietic precursor-like stage. J Exp Med 213, 1513-1535.

Quiros, P.M., Langer, T., and Lopez-Otin, C. (2015). New roles for mitochondrial proteases in health, ageing and disease. Nat Rev Mol Cell Biol 16, 345-359.

Quivoron, C., Couronné, L., Della Valle, V., Lopez, C.K., Plo, I., Wagner-Ballon, O., Do Cruzeiro, M., Delhommeau, F., Arnulf, B., Stern, M.H., et al. (2011). TET2 inactivation results in pleiotropic hematopoietic abnormalities in mouse and is a recurrent event during human lymphomagenesis. Cancer cell 20, 25-38.

R E Billingham, G.H.L., P B Medawar & H LL Williams (1952). Tolerance to homografts, twin diagnosis, and the freemartin condition in cattle. Heredity 6, 201-212.

Ram, R., and Storb, R. (2013). Pharmacologic prophylaxis regimens for acute graft-versus-host disease: past, present and future. Leuk Lymphoma 54, 1591-1601.

Ramachandran, A., Basu, U., Sultana, S., Nandakumar, D., and Patel, S.S. (2017). Human mitochondrial transcription factors TFAM and TFB2M work synergistically in promoter melting during transcription initiation. Nucleic Acids Res 45, 861-874.

Rambold, A.S., Kostelecky, B., Elia, N., and Lippincott-Schwartz, J. (2011). Tubular network formation protects mitochondria from autophagosomal degradation during nutrient starvation. Proc Natl Acad Sci U S A 108, 10190-10195.

Rasmussen, K.D., Jia, G., Johansen, J.V., Pedersen, M.T., Rapin, N., Bagger, F.O., Porse, B.T., Bernard, O.A., Christensen, J., and Helin, K. (2015). Loss of TET2 in hematopoietic cells leads to DNA hypermethylation of active enhancers and induction of leukemogenesis. Genes Dev 29, 910-922.

Renaud, L., Nibourel, O., Berthon, C., Roumier, C., Rodriguez, C., Frimat, C., Roche-Lestienne, C., Renneville, A., Quesnel, B., and Preudhomme, C. (2016). De Novo and Secondary Acute Myeloid Leukemia, Real World Data on Outcomes from the French Nord-Pas-De-Calais Picardie Acute Myeloid Leukemia Observatory. In 55th American Society of Hematology Annual Meeting (San Diego, Califonia ).

Reusser, P., Einsele, H., Lee, J., Volin, L., Rovira, M., Engelhard, D., Finke, J., Cordonnier, C., Link, H., Ljungman, P., et al. (2002). Randomized multicenter trial of foscarnet versus ganciclovir for preemptive therapy of cytomegalovirus infection after allogeneic stem cell transplantation. Blood 99, 1159-1164.

172

Richardson, P.G., Soiffer, R.J., Antin, J.H., Uno, H., Jin, Z., Kurtzberg, J., Martin, P.L., Steinbach, G., Murray, K.F., Vogelsang, G.B., et al. (2010). Defibrotide for the treatment of severe hepatic veno-occlusive disease and multiorgan failure after stem cell transplantation: a multicenter, randomized, dose-finding trial. Biol Blood Marrow Transplant 16, 1005-1017.

Richter-Dennerlein, R., Korwitz, A., Haag, M., Tatsuta, T., Dargazanli, S., Baker, M., Decker, T., Lamkemeyer, T., Rugarli, E.I., and Langer, T. (2014). DNAJC19, a mitochondrial cochaperone associated with cardiomyopathy, forms a complex with prohibitins to regulate cardiolipin remodeling. Cell Metab 20, 158-171.

Rossignol, R., Gilkerson, R., Aggeler, R., Yamagata, K., Remington, S.J., and Capaldi, R.A. (2004). Energy substrate modulates mitochondrial structure and oxidative capacity in cancer cells. Cancer Res 64, 985-993.

Rotin, L.E. (2016). Preclinical Evaluation of Synergistic Drug Combinations in Acute Myeloid Leukemia (TSpace University of Toronto ), pp. 184.

Roulois, D., Loo Yau, H., Singhania, R., Wang, Y., Danesh, A., Shen, S.Y., Han, H., Liang, G., Jones, P.A., Pugh, T.J., et al. (2015). DNA-Demethylating Agents Target Colorectal Cancer Cells by Inducing Viral Mimicry by Endogenous Transcripts. Cell 162, 961-973.

Rowley, J.D., Golomb, H.M., and Dougherty, C. (1977). 15/17 translocation, a consistent chromosomal change in acute promyelocytic leukaemia. Lancet 1, 549-550.

Ruhl, J.A., M; Dickie, L; Negoita, S. (2018). Hematopoietic and Lymphoid Neoplasm Coding Manual (Bethesda, MD).

Saito, Y., Uchida, N., Tanaka, S., Suzuki, N., Tomizawa-Murasawa, M., Sone, A., Najima, Y., Takagi, S., Aoki, Y., Wake, A., et al. (2010). Induction of cell cycle entry eliminates human leukemia stem cells in a mouse model of AML. Nat Biotechnol 28, 275-280.

Sale, G.E. (1996). Does graft-versus-host disease attack epithelial stem cells? Mol Med Today 2, 114-119.

Samudio, I., Harmancey, R., Fiegl, M., Kantarjian, H., Konopleva, M., Korchin, B., Kaluarachchi, K., Bornmann, W., Duvvuri, S., Taegtmeyer, H., et al. (2010). Pharmacologic inhibition of fatty acid oxidation sensitizes human leukemia cells to apoptosis induction. J Clin Invest 120, 142-156.

Sato, T., Onai, N., Yoshihara, H., Arai, F., Suda, T., and Ohteki, T. (2009). Interferon regulatory factor-2 protects quiescent hematopoietic stem cells from type I interferon-dependent exhaustion. Nat Med 15, 696-700.

Savani, B.N., Griffith, M.L., Jagasia, S., and Lee, S.J. (2011). How I treat late effects in adults after allogeneic stem cell transplantation. Blood 117, 3002-3009.

Schimmer, A.D. (2018). Mitochondrial Shapeshifting Impacts AML Stemness and Differentiation. Cell Stem Cell 23, 3-4.

173

Schlame, M., Brody, S., and Hostetler, K.Y. (1993). Mitochondrial cardiolipin in diverse eukaryotes. Comparison of biosynthetic reactions and molecular acyl species. Eur J Biochem 212, 727-735.

Schlame, M., Kelley, R.I., Feigenbaum, A., Towbin, J.A., Heerdt, P.M., Schieble, T., Wanders, R.J., DiMauro, S., and Blanck, T.J. (2003). Phospholipid abnormalities in children with Barth syndrome. J Am Coll Cardiol 42, 1994-1999.

Schwinghammer, K., Cheung, A.C., Morozov, Y.I., Agaronyan, K., Temiakov, D., and Cramer, P. (2013). Structure of human mitochondrial RNA polymerase elongation complex. Nat Struct Mol Biol 20, 1298-1303.

Sellar, R.S., Jaiswal, S., and Ebert, B.L. (2018). Predicting progression to AML. Nat Med 24, 904-906.

Seneviratne, A.K., Henao, J.J.A., Xu, G.W., Hurren, R., Kim, S., MacLean, N., Wang, X., Xu, M., Fajado, V., Hao, Z., et al. (2017). Tafazzin (TAZ) Regulates the Differentiation of AML Cells by Reducing Levels of the Phospholipid Phosphatidylethanolamine Blood 130.

Seneviratne, A.K., Xu, M., Henao, J.J.A., Fajardo, V.A., Hao, Z., Voisin, V., Xu, G.W., Hurren, R., Kim, S., MacLean, N., et al. (2019). The Mitochondrial Transacylase, Tafazzin, Regulates for AML Stemness by Modulating Intracellular Levels of Phospholipids. Cell Stem Cell 24, 621- 636 e616.

Seneviratne, A.K.X., M.; Schimmer, A.D. (2019). Tafazzin modulates cellular phospholipid composition to regulate AML stemness. Molecular & Cellular Oncology.

Sesaki, H., and Jensen, R.E. (1999). Division versus fusion: Dnm1p and Fzo1p antagonistically regulate mitochondrial shape. J Cell Biol 147, 699-706.

Seth, R.B., Sun, L., Ea, C.K., and Chen, Z.J. (2005). Identification and characterization of MAVS, a mitochondrial antiviral signaling protein that activates NF-kappaB and IRF 3. Cell 122, 669-682.

Shah, A., Andersson, T.M., Rachet, B., Bjorkholm, M., and Lambert, P.C. (2013). Survival and cure of acute myeloid leukaemia in England, 1971-2006: a population-based study. Br J Haematol 162, 509-516.

Shalem, O., Sanjana, N.E., Hartenian, E., Shi, X., Scott, D.A., Mikkelson, T., Heckl, D., Ebert, B.L., Root, D.E., Doench, J.G., et al. (2014). Genome-scale CRISPR-Cas9 knockout screening in human cells. Science 343, 84-87.

Shih, A.H., Jiang, Y., Meydan, C., Shank, K., Pandey, S., Barreyro, L., Antony-Debre, I., Viale, A., Socci, N., Sun, Y., et al. (2015). Mutational cooperativity linked to combinatorial epigenetic gain of function in acute myeloid leukemia. Cancer cell 27, 502-515.

Shinzawa-Itoh, K., Aoyama, H., Muramoto, K., Terada, H., Kurauchi, T., Tadehara, Y., Yamasaki, A., Sugimura, T., Kurono, S., Tsujimoto, K., et al. (2007). Structures and physiological roles of 13 integral lipids of bovine heart cytochrome c oxidase. EMBO J 26, 1713-1725.

174

Short, N.J., Kantarjian, H., Ravandi, F., and Daver, N. (2019). Emerging treatment paradigms with FLT3 inhibitors in acute myeloid leukemia. Ther Adv Hematol 10, 2040620719827310.

Shulman, H.M., Kleiner, D., Lee, S.J., Morton, T., Pavletic, S.Z., Farmer, E., Moresi, J.M., Greenson, J., Janin, A., Martin, P.J., et al. (2006). Histopathologic diagnosis of chronic graft- versus-host disease: National Institutes of Health Consensus Development Project on Criteria for Clinical Trials in Chronic Graft-versus-Host Disease: II. Pathology Working Group Report. Biol Blood Marrow Transplant 12, 31-47.

Sill, H., Olipitz, W., Zebisch, A., Schulz, E., and Wölfler, A. (2011). Therapy-related myeloid neoplasms: pathobiology and clinical characteristics. Br J Pharmacol 162, 792-805.

Simpson, C.D., Hurren, R., Kasimer, D., MacLean, N., Eberhard, Y., Ketela, T., Moffat, J., and Schimmer, A.D. (2012). A genome wide shRNA screen identifies α/β hydrolase domain containing 4 (ABHD4) as a novel regulator of anoikis resistance. Apoptosis 17, 666-678.

Sissons, J.G., and Carmichael, A.J. (2002). Clinical aspects and management of cytomegalovirus infection. J Infect 44, 78-83.

Škrtić, M. (2013). INHIBITION OF MITOCHONDRIAL TRANSLATION AS A THERPEUTIC STRATEGY FOR ACUTE MYELOID LEUKEMIA

. In Institute of Medical Sciences (Tspace: University of Toronto), pp. 197.

Skrtić, M., Sriskanthadevan, S., Jhas, B., Gebbia, M., Wang, X., Wang, Z., Hurren, R., Jitkova, Y., Gronda, M., Maclean, N., et al. (2011). Inhibition of mitochondrial translation as a therapeutic strategy for human acute myeloid leukemia. Cancer cell 20, 674-688.

Smith, Z.D., and Meissner, A. (2013). DNA methylation: roles in mammalian development. Nat Rev Genet 14, 204-220.

Smits, P., Smeitink, J.A., van den Heuvel, L.P., Huynen, M.A., and Ettema, T.J. (2007). Reconstructing the evolution of the mitochondrial ribosomal proteome. Nucleic Acids Res 35, 4686-4703.

Sou, K. (2012). Advanced Drug Carriers Targeting Bone Marrow, Recent Advances in Novel Drug Carrier Systems IntechOpen.

Soustek, M.S., Falk, D.J., Mah, C.S., Toth, M.J., Schlame, M., Lewin, A.S., and Byrne, B.J. (2011). Characterization of a transgenic short hairpin RNA-induced murine model of Tafazzin deficiency. Hum Gene Ther 22, 865-871.

Spady, D.W., Payne, P.R., Picou, D., and Waterlow, J.C. (1976). Energy balance during recovery from malnutrition. Am J Clin Nutr 29, 1073-1088.

Sparagna, G.C., and Lesnefsky, E.J. (2009). Cardiolipin remodeling in the heart. J Cardiovasc Pharmacol 53, 290-301.

Stefanyk, L.E., Coverdale, N., Roy, B.D., Peters, S.J., and LeBlanc, P.J. (2010). Skeletal muscle type comparison of subsarcolemmal mitochondrial membrane phospholipid fatty acid composition in rat. The Journal of membrane biology 234, 207-215.

175

Stein, E.M., DiNardo, C.D., Pollyea, D.A., Fathi, A.T., Roboz, G.J., Altman, J.K., Stone, R.M., DeAngelo, D.J., Levine, R.L., Flinn, I.W., et al. (2017). Enasidenib in mutant IDH2 relapsed or refractory acute myeloid leukemia. Blood 130, 722-731.

Stepanyants, N., Macdonald, P.J., Francy, C.A., Mears, J.A., Qi, X., and Ramachandran, R. (2015). Cardiolipin's propensity for phase transition and its reorganization by dynamin-related protein 1 form a basis for mitochondrial membrane fission. Mol Biol Cell 26, 3104-3116.

Stewart, J.B., and Chinnery, P.F. (2015). The dynamics of mitochondrial DNA heteroplasmy: implications for human health and disease. Nat Rev Genet 16, 530-542.

Stine, Z.E., Walton, Z.E., Altman, B.J., Hsieh, A.L., and Dang, C.V. (2015). MYC, Metabolism, and Cancer. Cancer Discov 5, 1024-1039.

Stumpf, J.D., and Copeland, W.C. (2011). Mitochondrial DNA replication and disease: insights from DNA polymerase gamma mutations. Cell Mol Life Sci 68, 219-233.

Sun, Y., Tawara, I., Toubai, T., and Reddy, P. (2007). Pathophysiology of acute graft-versus- host disease: recent advances. Transl Res 150, 197-214.

Sunami, Y., Araki, M., Kan, S., Ito, A., Hironaka, Y., Imai, M., Morishita, S., Ohsaka, A., and Komatsu, N. (2017). Histone Acetyltransferase p300/CREB-binding Protein-associated Factor (PCAF) Is Required for All-trans-retinoic Acid-induced Granulocytic Differentiation in Leukemia Cells. J Biol Chem 292, 2815-2829.

Swerdlow, S.C., E; Harris, N; Jaffe, E; Pileri, S;Stein, H; Thiele, J; Vardiman, JW. (2016). The rules, guidelines, and the Hematopoietic Database follow the World Health Organization (WHO) Classification of Tumours of Haematopoietic and Lymphoid Tissues., Vol 2, 4th edn (IARC Press ).

Sykes, D.B. (2018). The emergence of dihydroorotate dehydrogenase (DHODH) as a therapeutic target in acute myeloid leukemia. Expert Opin Ther Targets 22, 893-898.

Sykes, D.B., Kfoury, Y.S., Mercier, F.E., Wawer, M.J., Law, J.M., Haynes, M.K., Lewis, T.A., Schajnovitz, A., Jain, E., Lee, D., et al. (2016). Inhibition of Dihydroorotate Dehydrogenase Overcomes Differentiation Blockade in Acute Myeloid Leukemia. Cell 167, 171-186 e115.

Tabe, Y., Yamamoto, S., Saitoh, K., Sekihara, K., Monma, N., Ikeo, K., Mogushi, K., Shikami, M., Ruvolo, V., Ishizawa, J., et al. (2017). Bone Marrow Adipocytes Facilitate Fatty Acid Oxidation Activating AMPK and a Transcriptional Network Supporting Survival of Acute Monocytic Leukemia Cells. Cancer Res 77, 1453-1464.

Takahashi, K., Wang, F., Kantarjian, H., Doss, D., Khanna, K., Thompson, E., Zhao, L., Patel, K., Neelapu, S., Gumbs, C., et al. (2017). Preleukaemic clonal haemopoiesis and risk of therapy- related myeloid neoplasms: a case-control study. Lancet Oncol 18, 100-111.

Takizawa, H., Boettcher, S., and Manz, M.G. (2012). Demand-adapted regulation of early hematopoiesis in infection and inflammation. Blood 119, 2991-3002.

Tan, B.T., Park, C.Y., Ailles, L.E., and Weissman, I.L. (2006). The cancer stem cell hypothesis: a work in progress. Lab Invest 86, 1203-1207.

176

Taylor, W.A., and Hatch, G.M. (2003). Purification and characterization of monolysocardiolipin acyltransferase from pig liver mitochondria. J Biol Chem 278, 12716-12721.

Taylor, W.A., and Hatch, G.M. (2009). Identification of the human mitochondrial linoleoyl- coenzyme A monolysocardiolipin acyltransferase (MLCL AT-1). J Biol Chem 284, 30360- 30371.

Taylor, W.A., Mejia, E.M., Mitchell, R.W., Choy, P.C., Sparagna, G.C., and Hatch, G.M. (2012). Human trifunctional protein alpha links cardiolipin remodeling to beta-oxidation. PLoS One 7, e48628.

Thakur, D.S. (2011). Topoisomerase II Inhibitors in Cancer Treatment. International Journal of Pharmacutical Science and Nanotechnology 3, 1173-1181.

Thomas, D., and Majeti, R. (2017). Biology and relevance of human acute myeloid leukemia stem cells. Blood 129, 1577-1585.

Thomas, E., Storb, R., Clift, R.A., Fefer, A., Johnson, F.L., Neiman, P.E., Lerner, K.G., Glucksberg, H., and Buckner, C.D. (1975). Bone-marrow transplantation (first of two parts). N Engl J Med 292, 832-843.

Thomas, E.D. (1999). Bone marrow transplantation: a review. Semin Hematol 36, 95-103.

Till, J.E., and McCulloch, E.A. (1961). A direct measurement of the radiation sensitivity of normal mouse bone marrow cells. Radiat Res 14, 213-222.

Tilokani, L., Nagashima, S., Paupe, V., and Prudent, J. (2018). Mitochondrial dynamics: overview of molecular mechanisms. Essays Biochem 62, 341-360.

Tomecki, R., Dmochowska, A., Gewartowski, K., Dziembowski, A., and Stepien, P.P. (2004). Identification of a novel human nuclear-encoded mitochondrial poly(A) polymerase. Nucleic Acids Res 32, 6001-6014.

Toubai, T., Sun, Y., and Reddy, P. (2008). GVHD pathophysiology: is acute different from chronic? Best Pract Res Clin Haematol 21, 101-117.

Tsai, H.J., Jiang, S.S., Hung, W.C., Borthakur, G., Lin, S.F., Pemmaraju, N., Jabbour, E., Bomalaski, J.S., Chen, Y.P., Hsiao, H.H., et al. (2017). A Phase II Study of Arginine Deiminase (ADI-PEG20) in Relapsed/Refractory or Poor-Risk Acute Myeloid Leukemia Patients. Sci Rep 7, 11253.

Twig, G., Elorza, A., Molina, A.J., Mohamed, H., Wikstrom, J.D., Walzer, G., Stiles, L., Haigh, S.E., Katz, S., Las, G., et al. (2008). Fission and selective fusion govern mitochondrial segregation and elimination by autophagy. EMBO J 27, 433-446.

Tzelepis, K., Koike-Yusa, H., De Braekeleer, E., Li, Y., Metzakopian, E., Dovey, O.M., Mupo, A., Grinkevich, V., Li, M., Mazan, M., et al. (2016). A CRISPR Dropout Screen Identifies Genetic Vulnerabilities and Therapeutic Targets in Acute Myeloid Leukemia. Cell Rep 17, 1193- 1205.

177

Valianpour, F., Mitsakos, V., Schlemmer, D., Towbin, J.A., Taylor, J.M., Ekert, P.G., Thorburn, D.R., Munnich, A., Wanders, R.J., Barth, P.G., et al. (2005). Monolysocardiolipins accumulate in Barth syndrome but do not lead to enhanced apoptosis. J Lipid Res 46, 1182-1195.

Valianpour, F., Wanders, R.J., Overmars, H., Vreken, P., Van Gennip, A.H., Baas, F., Plecko, B., Santer, R., Becker, K., and Barth, P.G. (2002). Cardiolipin deficiency in X-linked cardioskeletal myopathy and neutropenia (Barth syndrome, MIM 302060): a study in cultured skin fibroblasts. J Pediatr 141, 729-733. van der Bliek, A.M., Shen, Q., and Kawajiri, S. (2013). Mechanisms of mitochondrial fission and fusion. Cold Spring Harb Perspect Biol 5. van der Laan, M., Meinecke, M., Dudek, J., Hutu, D.P., Lind, M., Perschil, I., Guiard, B., Wagner, R., Pfanner, N., and Rehling, P. (2007). Motor-free mitochondrial presequence translocase drives membrane integration of preproteins. Nat Cell Biol 9, 1152-1159. van Rhenen, A., Feller, N., Kelder, A., Westra, A.H., Rombouts, E., Zweegman, S., van der Pol, M.A., Waisfisz, Q., Ossenkoppele, G.J., and Schuurhuis, G.J. (2005). High stem cell frequency in acute myeloid leukemia at diagnosis predicts high minimal residual disease and poor survival. Clin Cancer Res 11, 6520-6527. van Rhenen, A., van Dongen, G.A., Kelder, A., Rombouts, E.J., Feller, N., Moshaver, B., Stigter-van Walsum, M., Zweegman, S., Ossenkoppele, G.J., and Jan Schuurhuis, G. (2007). The novel AML stem cell associated antigen CLL-1 aids in discrimination between normal and leukemic stem cells. Blood 110, 2659-2666. van Werkhoven, M.A., Thorburn, D.R., Gedeon, A.K., and Pitt, J.J. (2006). Monolysocardiolipin in cultured fibroblasts is a sensitive and specific marker for Barth Syndrome. J Lipid Res 47, 2346-2351.

Vande Walle, L., Lamkanfi, M., and Vandenabeele, P. (2008). The mitochondrial serine protease HtrA2/Omi: an overview. Cell Death Differ 15, 453-460.

Vardiman, J.W., Thiele, J., Arber, D.A., Brunning, R.D., Borowitz, M.J., Porwit, A., Harris, N.L., Le Beau, M.M., Hellstrom-Lindberg, E., Tefferi, A., et al. (2009). The 2008 revision of the World Health Organization (WHO) classification of myeloid neoplasms and acute leukemia: rationale and important changes. Blood 114, 937-951.

Vaz, F.M., Houtkooper, R.H., Valianpour, F., Barth, P.G., and Wanders, R.J. (2003). Only one splice variant of the human TAZ gene encodes a functional protein with a role in cardiolipin metabolism. J Biol Chem 278, 43089-43094.

Voos, W. (2013). Chaperone-protease networks in mitochondrial protein homeostasis. Biochim Biophys Acta 1833, 388-399.

Vreken, P., Valianpour, F., Nijtmans, L.G., Grivell, L.A., Plecko, B., Wanders, R.J., and Barth, P.G. (2000). Defective remodeling of cardiolipin and phosphatidylglycerol in Barth syndrome. Biochem Biophys Res Commun 279, 378-382.

Vyas, S., Zaganjor, E., and Haigis, M.C. (2016). Mitochondria and Cancer. Cell 166, 555-566.

178

Wang, F., Travins, J., DeLaBarre, B., Penard-Lacronique, V., Schalm, S., Hansen, E., Straley, K., Kernytsky, A., Liu, W., Gliser, C., et al. (2013). Targeted inhibition of mutant IDH2 in leukemia cells induces cellular differentiation. Science 340, 622-626.

Wang, G., McCain, M.L., Yang, L., He, A., Pasqualini, F.S., Agarwal, A., Yuan, H., Jiang, D., Zhang, D., Zangi, L., et al. (2014). Modeling the mitochondrial cardiomyopathy of Barth syndrome with induced pluripotent stem cell and heart-on-chip technologies. Nat Med 20, 616- 623.

Wang, T., Marquardt, C., and Foker, J. (1976). Aerobic glycolysis during lymphocyte proliferation. Nature 261, 702-705.

Ward, P.S., and Thompson, C.B. (2012). Metabolic reprogramming: a cancer hallmark even warburg did not anticipate. Cancer cell 21, 297-308.

Warner, J.K., Wang, J.C., Takenaka, K., Doulatov, S., McKenzie, J.L., Harrington, L., and Dick, J.E. (2005). Direct evidence for cooperating genetic events in the leukemic transformation of normal human hematopoietic cells. Leukemia 19, 1794-1805.

Weber, T.A., Koob, S., Heide, H., Wittig, I., Head, B., van der Bliek, A., Brandt, U., Mittelbronn, M., and Reichert, A.S. (2013). APOOL is a cardiolipin-binding constituent of the Mitofilin/MINOS protein complex determining cristae morphology in mammalian mitochondria. PLoS One 8, e63683.

Wei, L.N. (2003). Retinoid receptors and their coregulators. Annu Rev Pharmacol Toxicol 43, 47-72.

Wei, Y., Chiang, W.C., Sumpter, R., Jr., Mishra, P., and Levine, B. (2017). Prohibitin 2 Is an Inner Mitochondrial Membrane Mitophagy Receptor. Cell 168, 224-238 e210.

Weiden, P.L., Flournoy, N., Thomas, E.D., Prentice, R., Fefer, A., Buckner, C.D., and Storb, R. (1979). Antileukemic effect of graft-versus-host disease in human recipients of allogeneic- marrow grafts. N Engl J Med 300, 1068-1073.

Weiden, P.L., Sullivan, K.M., Flournoy, N., Storb, R., Thomas, E.D., and Seattle Marrow Transplant, T. (1981). Antileukemic effect of chronic graft-versus-host disease: contribution to improved survival after allogeneic marrow transplantation. N Engl J Med 304, 1529-1533.

Weinberg, S.E., and Chandel, N.S. (2015). Targeting mitochondria metabolism for cancer therapy. Nat Chem Biol 11, 9-15.

Weissman, I.L., and Shizuru, J.A. (2008). The origins of the identification and isolation of hematopoietic stem cells, and their capability to induce donor-specific transplantation tolerance and treat autoimmune diseases. Blood 112, 3543-3553.

Wendel, H.G., de Stanchina, E., Cepero, E., Ray, S., Emig, M., Fridman, J.S., Veach, D.R., Bornmann, W.G., Clarkson, B., McCombie, W.R., et al. (2006). Loss of p53 impedes the antileukemic response to BCR-ABL inhibition. Proc Natl Acad Sci U S A 103, 7444-7449.

179

Wilson, A., Laurenti, E., Oser, G., van der Wath, R.C., Blanco-Bose, W., Jaworski, M., Offner, S., Dunant, C.F., Eshkind, L., Bockamp, E., et al. (2008). Hematopoietic stem cells reversibly switch from dormancy to self-renewal during homeostasis and repair. Cell 135, 1118-1129.

Wise, D.R., DeBerardinis, R.J., Mancuso, A., Sayed, N., Zhang, X.Y., Pfeiffer, H.K., Nissim, I., Daikhin, E., Yudkoff, M., McMahon, S.B., et al. (2008). Myc regulates a transcriptional program that stimulates mitochondrial glutaminolysis and leads to glutamine addiction. Proc Natl Acad Sci U S A 105, 18782-18787.

Wise, D.R., and Thompson, C.B. (2010). Glutamine addiction: a new therapeutic target in cancer. Trends Biochem Sci 35, 427-433.

Wong, T.N., Ramsingh, G., Young, A.L., Miller, C.A., Touma, W., Welch, J.S., Lamprecht, T.L., Shen, D., Hundal, J., Fulton, R.S., et al. (2015). Role of TP53 mutations in the origin and evolution of therapy-related acute myeloid leukaemia. Nature 518, 552-555.

Woriax, V.L., Bullard, J.M., Ma, L., Yokogawa, T., and Spremulli, L.L. (1997). Mechanistic studies of the translational elongation cycle in mammalian mitochondria. Biochim Biophys Acta 1352, 91-101.

Wu, B., and Hur, S. (2015). How RIG-I like receptors activate MAVS. Curr Opin Virol 12, 91- 98.

Wu, D., Hu, D., Chen, H., Shi, G., Fetahu, I.S., Wu, F., Rabidou, K., Fang, R., Tan, L., Xu, S., et al. (2018). Glucose-regulated phosphorylation of TET2 by AMPK reveals a pathway linking diabetes to cancer. Nature 559, 637-641.

Xie, M., Lu, C., Wang, J., McLellan, M.D., Johnson, K.J., Wendl, M.C., McMichael, J.F., Schmidt, H.K., Yellapantula, V., Miller, C.A., et al. (2014). Age-related mutations associated with clonal hematopoietic expansion and malignancies. Nat Med 20, 1472-1478.

Xu, C., Zheng, Z., Fang, L., Zhao, N., Lin, Z., Liang, T., Zhai, Z., and Zhou, J. (2013). Phosphatidylserine enhances osteogenic differentiation in human mesenchymal stem cells via ERK signal pathways. Mater Sci Eng C Mater Biol Appl 33, 1783-1788.

Xu, Y., Condell, M., Plesken, H., Edelman-Novemsky, I., Ma, J., Ren, M., and Schlame, M. (2006a). A Drosophila model of Barth syndrome. Proc Natl Acad Sci U S A 103, 11584-11588.

Xu, Y., Malhotra, A., Ren, M., and Schlame, M. (2006b). The enzymatic function of tafazzin. J Biol Chem 281, 39217-39224.

Xu, Y., Zhang, S., Malhotra, A., Edelman-Novemsky, I., Ma, J., Kruppa, A., Cernicica, C., Blais, S., Neubert, T.A., Ren, M., et al. (2009). Characterization of tafazzin splice variants from humans and fruit flies. J Biol Chem 284, 29230-29239.

Xue, H.Y., Ji, L.J., Gao, A.M., Liu, P., He, J.D., and Lu, X.J. (2016). CRISPR-Cas9 for medical genetic screens: applications and future perspectives. J Med Genet 53, 91-97.

Yáñez, A., Goodridge, H.S., Gozalbo, D., and Gil, M.L. (2013). TLRs control hematopoiesis during infection. Eur J Immunol 43, 2526-2533.

180

Yang, L., Rodriguez, B., Mayle, A., Park, H.J., Lin, X., Luo, M., Jeong, M., Curry, C.V., Kim, S.B., Ruau, D., et al. (2016). DNMT3A Loss Drives Enhancer Hypomethylation in FLT3-ITD- Associated Leukemias. Cancer cell 29, 922-934.

Yang, L., Venneti, S., and Nagrath, D. (2017). Glutaminolysis: A Hallmark of Cancer Metabolism. Annu Rev Biomed Eng 19, 163-194.

Yang, M., Soga, T., and Pollard, P.J. (2013). Oncometabolites: linking altered metabolism with cancer. J Clin Invest 123, 3652-3658.

Ye, C., Sutter, B.M., Wang, Y., Kuang, Z., and Tu, B.P. (2017). A Metabolic Function for Phospholipid and Histone Methylation. Mol Cell 66, 180-193 e188.

Ye, H., Adane, B., Khan, N., Sullivan, T., Minhajuddin, M., Gasparetto, M., Stevens, B., Pei, S., Balys, M., Ashton, J.M., et al. (2016). Leukemic Stem Cells Evade Chemotherapy by Metabolic Adaptation to an Adipose Tissue Niche. Cell Stem Cell 19, 23-37.

Yehudai, D., Liyanage, S.U., Hurren, R., Rizoska, B., Albertella, M., Gronda, M., Jeyaraju, D.V., Wang, X., Barghout, S.H., MacLean, N., et al. (2019). The thymidine dideoxynucleoside analog, alovudine, inhibits the mitochondrial DNA polymerase gamma, impairs oxidative phosphorylation and promotes monocytic differentiation in acute myeloid leukemia. Haematologica 104, 963-972.

Youle, R.J., and van der Bliek, A.M. (2012). Mitochondrial fission, fusion, and stress. Science 337, 1062-1065.

Yuan, R., Zhu, X., Radich, J.P., and Ao, P. (2016). From molecular interaction to acute promyelocytic leukemia: Calculating leukemogenesis and remission from endogenous molecular-cellular network. Sci Rep 6, 24307.

Zhao, J.L., and Baltimore, D. (2015). Regulation of stress-induced hematopoiesis. Curr Opin Hematol 22, 286-292.

Zhao, Z., Zuber, J., Diaz-Flores, E., Lintault, L., Kogan, S.C., Shannon, K., and Lowe, S.W. (2010). p53 loss promotes acute myeloid leukemia by enabling aberrant self-renewal. Genes Dev 24, 1389-1402.

Zink, F., Stacey, S.N., Norddahl, G.L., Frigge, M.L., Magnusson, O.T., Jonsdottir, I., Thorgeirsson, T.E., Sigurdsson, A., Gudjonsson, S.A., Gudmundsson, J., et al. (2017). Clonal hematopoiesis, with and without candidate driver mutations, is common in the elderly. Blood 130, 742-752.

181

Figure 6. and Figure 7.

182 183

Figure 8. Image was open source, however could not get in touch with author to confirm.

184

Figure 9.

185

Figure 10.

186

Figure 12.

187

Table 2

188

Table 3

189

Figure in Results

190

Excerpts in discussion