1 Targeting the Mitochondrial Dna Polymerase Gamma in Acute Myeloid Leukemia

TARGETING THE MITOCHONDRIAL DNA POLYMERASE GAMMA IN ACUTE MYELOID LEUKEMIA

By Sanduni U. Liyanage

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

1 TARGETING THE MITOCHONDRIAL DNA POLYMERASE GAMMA IN ACUTE MYELOID LEUKEMIA

Sanduni U. Liyanage

Doctor of Philosophy Department of Medical Biophysics University of Toronto 2017 ABSTRACT

Part I

Mitochondrial DNA (mtDNA) biosynthesis requires replication factors and adequate nucleotide

pools from the mitochondria and cytoplasm. We performed gene expression profiling analysis of

542 human AML samples and identified 55% with upregulated mtDNA biosynthesis pathway

expression compared to normal hematopoietic cells. Genes that support mitochondrial nucleotide

pools, including mitochondrial nucleotide transporters and a subset of cytoplasmic nucleoside

kinases, were also increased in AML compared to normal hematopoietic samples. Knockdown of

cytoplasmic nucleoside kinases reduced mtDNA levels in AML cells, demonstrating their

contribution in maintaining mtDNA. To assess cytoplasmic nucleoside kinase pathway activity,

we employed a nucleoside analog 2’3’-dideoxycytidine (ddC), which is phosphorylated to the

activated anti-metabolite, 2’3’-dideoxycytidine triphosphate (ddCTP) by cytoplasmic nucleoside

kinases. ddC is a selective inhibitor of the mitochondrial DNA polymerase, POLG. ddC was

preferentially activated in AML cells compared to normal hematopoietic progenitor cells. ddC

treatment inhibited mtDNA replication, oxidative phosphorylation, and induced cytotoxicity in a

panel of AML cell lines. Furthermore, ddC preferentially inhibited mtDNA replication in a subset

of primary human leukemia cells and selectively targeted leukemia cells while sparing normal

progenitors cells. In animal models of human AML, treatment with ddC decreased mtDNA,

electron transport chain proteins, and induced tumor regression without toxicity. ddC also targeted

leukemic stem cells in secondary AML xenotransplantation assays. Thus, AML cells have

ii increased cytidine nucleoside kinase activity that regulates mtDNA biogenesis and can be leveraged to selectively target oxidative phosphorylation in AML.

Part II

Human mitochondrial DNA is replicated by the mitochondrial DNA polymerase gamma. Using proximity dependent biotin labelling (BioID), we characterized the POLG interactome and identified new interaction partners involved in mtDNA maintenance, transcription, translation and protein quality control. We also identified interaction with the nuclear AAA+ ATPase Ruvbl2, suggesting mitochondrial localization for this protein. Ruvbl2 was detected in mitochondria- enriched fractions in leukemic cells. Additionally, transgenic overexpression of Ruvbl2 from an alternative translation initiation site resulted in mitochondrial co-localization. Overall, POLG interactome mapping identifies novel proteins which support mitochondrial biogenesis and a potential novel mitochondrial isoform of Ruvbl2.

iii Acknowledgements

Throughout the course of my studies, I have been extremely fortunate to be surrounded by supportive and caring colleagues, family and friends. Despite the twists and turns during this journey, knowing that I have a supportive environment has been a great strength and I am forever grateful for the relationships and memories created these past few years.

To my supervisor, Dr. Aaron Schimmer, thank you for providing me with the opportunity to grow and develop as a scientist through your constant support and mentorship. Your kind, enthusiastic and positive nature has made it a pleasure to have worked with you. I hope to continue these meetings and collaborate on projects in the future.

The completion of this project would not have been possible without the excellent work of many colleagues, in particular, Rose. I truly admire the dedication, hard work and care you placed into every project you have participated it. I am grateful for your friendship, laughs, conversations and consistent support provided during my time at the Schimmer lab. To Marcela, thank you for all your advice, encouragement, support and friendship during this time. You both have served as great role models to me for your passion, commitment and caring attitude towards the people you have worked with. To the other staff in the Schimmer lab, Neil and XiaoMing, thank you for your technical assistance with this project. It’s a blessing to have worked with a great team, who can be consistently relied upon on for support and help.

To the post-doctoral fellows in the Schimmer Lab, Danny J., Danny B. and Ian, thank you for serving as my mentors. In particular, thank you to Danny J. for your time and companionship.

Our discussions both intellectual and personal, have been valuable to my growth on a personal and career level. To the other members of the Schimmer Lab, Lianne, Yan, Thirushi, Dilshad, Ayesh,

Geethu, Dana, Samir, Iulia and others; thank you for your friendship, encouragement and support.

iv I will cherish the many laughs, coffee breaks and conversations in between experiments, and otherwise. Each of you have made my time at the Schimmer lab a memorable experience.

To my committee members, Rebecca Laposa and David Hedley, thank you for providing a positive and supportive framework to progress my project in a productive manner. Thank you in particular to Rebecca for your enthusiasm, understanding and mentorship during my graduate studies. The opportunity to collaborate with members of other programs on research projects has been a valuable training experience. Additionally, I extend my thanks to our collaborators at the

Donnelly Centre and the Metabolomics Facility at McGill University, your technical expertise has been essential to the progress of this study.

Lastly, but most importantly, I am grateful to my parents and my brother. The unconditional love and support provided by my parents have been the foundation for any success

I have achieved. Thank you for all your sacrifices and efforts to ensure that I received the best education possible. Moving forward, I hope to use my scientific training to serve all beings to the best of my ability.

v TABLE OF CONTENTS

ABSTRACT………………………………………………………………………………………………..ii

ACKNOWLEDGEMENTS…………………………………………..………………...... ……………….iv

TABLE OF CONTENTS……….…………………………………...……………………………………vi

LIST OF FIGURES…………....…………………………………..….…………………………………..x

LIST OF ABBREVIATIONS…...……………………………………………………………….….…xii

1: INTRODUCTION ...... 1

1.1. ACUTE MYELOID LEUKEMIA ...... 1

1.1.1 CLASSIFICATION ...... 4

1.1.2. STEM CELL CHARACTERISTICS OF AML ...... 5

1.1.2. MECHANISMS ...... 6

1.1.3. CURRENT TREATMENT STRATEGIES FOR AML ...... 9

1.1.4. EMERGING THERAPIES FOR TREATMENT OF AML ...... 9

1.2 MITOCHONDRIA ...... 11

1.2.1. STRUCTURE AND FUNCTION ...... 11

1.2.2. MITOCHONDRIAL DNA ORGANIZATION ...... 14

1.2.3. MITOCHONDRIAL BIOGENESIS REGULATION ...... 15

1.2.4. OXIDATIVE PHOSPHORYLATION ...... 16

1.2.5. QUALITY CONTROL OF MITOCHONDRIA ...... 19

1.2.6. MITOCHONDRIAL REGULATION OF CELL DEATH ...... 20

1.3. MITOCHONDRIAL DNA ...... 23

1.3.1. STRUCTURE AND FUNCTION ...... 23

1.3.2. MITOCHONDRIAL DNA REPLICATION ...... 26

1.3.3. MITOCHONDRIAL TRANSCRIPTION AND TRANSLATION ...... 27 vi 1.3.4. REGULATION OF MITOCHONDRIAL DNA CONTENT ...... 29

1.3.5. MITOCHONDRIAL NUCLEOTIDE POOL MAINTENANCE ...... 30

1.3.5. Cytoplasmic nucleotide metabolism ...... 34

1.3.5.1. De novo biosynthesis ...... 34

1.3.5.2. Cytoplasmic nucleotide salvage pathway ...... 34

1.3.6. MITOCHONDRIAL NUCLEOTIDE TRANSPORT ...... 36

1.3.7. MTDNA MUTATIONS ...... 37

1.3.8. MTDNA DEPLETION SYNDROMES ...... 38

1.3.9. MITOCHONDRIAL DNA MUTATIONS IN HEMATOLOGICAL DISORDERS ...... 38

1.4. MITOCHONDRIAL DNA POLYMERASE GAMMA ...... 39

1.4.1. STRUCTURE AND FUNCTION ...... 39

1.4.2. FUNCTIONAL EFFECTS OF TARGETING POLG ...... 40

1.4.2.1. Mouse models of POLG ...... 40

1.4.2.2. POLG related diseases...... 41

1.4.2.3. 2’3’-dideoxycytidine (ddC) ...... 42

1.4.2.4. Inhibition of POLG by anti-viral nucleoside analogs ...... 45

1.5. TARGETING MITOCHONDRIA AS A THERAPEUTIC STRATEGY ...... 46

1.5.1. Targeting mitochondria in cancer ...... 46

1.5.2. Targeting mitochondria in AML ...... 48

1.6 REFERENCES ...... 51

2. MATERIALS AND METHODS ...... 59

2.1 BIOINFORMATIC ANALYSIS ...... 59

2.2 PRIMARY AML AND NORMAL HEMATOPOIETIC CELLS ...... 60

2.3 CELL LINES ...... 61

2.4 IMMUNOBLOTTING ...... 61

2.5 SHRNA KNOCKDOWN ...... 62

vii 2.6 QUANTITATIVE REAL-TIME POLYMERASE CHAIN REACTION ...... 63

2.7 MITOCHONDRIAL DNA QUANTIFICATION ...... 63

2.8 COX I AND COX II MRNA QUANTIFICATION ...... 63

2.9 INTRACELLULAR QUANTIFICATION OF DDC AND DDCTP ...... 64

2.10 RNA EXPRESSION PROFILING ...... 66

2.11 METABOLIC ANALYSIS...... 66

2.12 CELL PROLIFERATION AND VIABILITY ASSAYS ...... 66

2.13 FLOW CYTOMETRY ...... 67

2.13.1 MITOCHONDRIAL MASS ...... 67

2.14 ELECTRON MICROSCOPY ...... 68

2.15 XENOGRAFT MODELS OF HUMAN AML ...... 68

2.16 BIOID CLONING ...... 69

2.17 BIOID INTERACTOME MAPPING IN CELL LINES ...... 70

2.18 BIOID MASS SPECTROMETRY AND DATA ANALYSIS ...... 71

2.19 PREPARATION OF PURIFIED- MITOCHONDRIAL FRACTIONS FOR RUVBL2 IMMUNOBLOTTING ...... 72

2.20 IDENTIFICATION OF AN ALTERNATIVE TRANSLATION INITIATION ISOFORM OF RUVBL2 BY

IMMUNOFLUORESCENCE ...... 72

2.20 STATISTICAL ANALYSIS ...... 73

2.3 REFERENCES ...... 74

3. LEVERAGING INCREASED CYTOPLASMIC NUCLEOSIDE KINASE ACTIVITY TO TARGET

MITOCHONDRIAL DNA AND OXIDATIVE PHOSPHORYLATION IN AML ...... 76

Contributions ...... 76

3.0 INTRODUCTION ...... 77

3.1 RESULTS...... 78

3.1.1 A subset of primary AML cells display upregulated mtDNA biosynthesis ...... 78

3.1.2. Leukemia cell lines display upregulated mtDNA biosynthesis expression ...... 81

3.1.3. AML cells display upregulated mitochondrial nucleotide transporter expression ...... 84

viii 3.1.4. A subset of cytoplasmic nucleoside kinases are upregulated in AML ...... 86

3.2 CYTOPLASMIC NUCLEOSIDE KINASES REGULATE MTDNA BIOSYNTHESIS ...... 91

3.2.1 Knockdown of cytoplasmic nucleoside kinases deplete mtDNA in AML ...... 91

3.3 ASSESSMENT OF CYTOPLASMIC NUCLEOSIDE KINASE ACTIVITY ...... 94

3.3.1 Cytoplasmic nucleoside kinase activity is elevated in AML ...... 94

3.3.2. ddC is activated by cytoplasmic nucleoside kinases ...... 96

3.4 INVESTIGATING THE EFFECT OF DDC ON MTDNA REPLICATION AND CELLULAR BIOENERGETICS ...... 98

3.4.1 ddC depletes mtDNA in AML cells...... 98

3.4.2 ddC depletes mtDNA encoded transcripts, proteins and oxidative phosphorylation ...... 100

3.4.3 ddC alters mitochondrial structure independent of changes in mitochondrial mass ...... 104

3.5. ASSESSING THE EFFICACY DDC IN VITRO AND IN VIVO ...... 107

3.5.1 ddC inhibits proliferation and viability of AML cell lines ...... 107

3.5.2. HEK 293 cells are resistant to ddC in comparison to AML cells ...... 109

3.5.3. mtDNA-replication is functionally important for ddC mechanism of action ...... 111

3.5.4. ddC preferentially targets mtDNA replication in primary AML cells and induces anti-leukemic effects in

vitro ...... 113

3.5.5 ddC displays efficacy in mouse models of human AML ...... 116

3.5.6. ddC is well-tolerated in mouse models of human AML...... 119

3.5.7. ddC targets bulk AML and LSC population in vivo ...... 123

4. DISCUSSION AND FUTURE DIRECTIONS ...... 125

4.1 Discussion ...... 125

4.2 References ...... 130

5. CHARACTERIZING THE INTERACTOME OF POLG ...... 135

Contributions ...... 135

5.1 INTRODUCTION ...... 136

5.2 RESULTS...... 137

5.2.1 POLG interacts with proteins which support mitochondrial metabolism and biogenesis ...... 137

ix 5.2.2. Ruvbl1 and Ruvbl2 are detected in mitochondria-enriched fractions ...... 141

5.2.3. Ruvbl2 has a potential mitochondrial isoform generated by alternative translation initiation...... 142

5.3 DISCUSSION...... 145

5.4 REFERENCES ...... 147

6. FUTURE DIRECTIONS...... 148

6.2 References ...... 155

LIST OF FIGURES

Figure 1: Normal and leukemic hematopoiesis ...... 3

Figure 1.2: Mitochondria function as bioenergetic, biosynthetic and signaling organelles ...... 13

Figure 1.3: Mitochondrial DNA encoded genes in oxidative phosphorylation ...... 18

Figure 1.4: Mitochondrial regulation of apoptosis ...... 22

Figure 1.5: Human mitochondrial genome map...... 25

Figure 1.6: Mitochondrial nucleotide biosynthesis ...... 33

Figure 1.7: ddC (2’3’-dideoxycytidine) metabolism ...... 44

Figure 3.1: Subsets of AML display upregulated mtDNA biosynthesis expression ...... 80

Figure 3.2: mtDNA biosynthesis genes are upregulated in leukemia cell lines ...... 83

Figure 3.3: Mitochondrial nucleotide transporters are upregulated in primary human AML samples...... 85

Figure 3.4: Cytoplasmic nucleoside kinases are upregulated in a subset of AML ...... 88

Table 1: Leukemia patient characteristics ...... 89

Figure 3.5: Cytoplasmic nucleoside kinases regulate mtDNA biosynthesis...... 93

Figure 3.6: Cytoplasmic nucleoside kinase activity is elevated in AML ...... 95

Fig 3.7: DCK knockdown depletes levels of activated ddC...... 97

Figure 3.8: ddC inhibits mtDNA replication in AML cell lines...... 99

Figure 3.9: ddC selectively depletes mtDNA-encoded transcripts ...... 101

Table 2: Nanostring analysis of ETC subunits in OCI-AML2 cells following ddC treatment ...... 102

Figure 3.10: ddC targets oxidative metabolism proteins and activity ...... 103

Figure 3.11: Effect of targeting mtDNA replication on mitochondrial content and ultrastructure...... 106

x Figure 3.12: ddC induces anti-leukemic effects in a panel of AML cell lines...... 108

Figure 3.13: ddC is preferentially active in AML cell lines compared to HEK 293 cells...... 110

Figure 3.14: MtDNA-deficient Rho (0) cells are resistant to ddC treatment...... 112

Figure 3.15: ddC preferentially targets mtDNA replication and induces anti-AML effects in primary AML samples in vitro .... 114

Table 3: Effect of ddC treatment on primary cells ...... 115

Figure 3.16: ddC displays efficacy in OCI-AML2 xenograft model of human AML ...... 118

Figure 3.17: Safety profile of ddC in mouse models of human AML ...... 122

Figure 3.18: ddC targets LSC populations in mouse models of human AML ...... 124

Table 5.1: Identification of POLG interaction partners by BioID ...... 138

Figure 5.1: POLG BioID interactome mapping reveals novel interaction candidates ...... 140

Figure 5.2: Ruvbl2 is detected in mitochondria-enriched fractions of leukemic cells ...... 142

Figure 5.3: Alternative translation initiation generates a mitochondrial isoform of RUVBL2 in human cultured cells ...... 144

List of copyright permissions ...... 156

xi List of Abbreviations

AML: Acute myeloid leukemia

AraC: Cytarabine

ATI: Alternative translation initiation

ATP: Adenosine triphosphate

ATP6: ATP synthase 6

BioID: Proximity-based biotinylation

BirA: Biotin ligase

CD34: Cluster of differentiation 34

Complex IV: Cytochrome c oxidase

CMPK1: Cytidine monophosphate kinase 1

CO: Cytochrome c oxidase subunit

COX: Cytochrome c oxidase

CSC: Cancer Stem Cell

DCK: Deoxycytidine kinase dCTP: Deoxycytidine triphosphate ddC: 2’3’-dideoxycytidine ddCTP: 2’3’-dideoxycytidine triphosphate

DGUOK: Mitochondrial deoxyguanosine kinase dNTP: Deoxynucleotide triphosphate

ETC: Electron transport chain

FITC: Fluoroscein isothiocyanate

xii FBS: Fetal bovine serum

G-CSF: Granulocyte colony-stimulating factor

GO: Gene Ontology

HEK: Human embryonic kidney

IL-3: Interleukin 3

I.P.: Intra-peritoneal

LC-MS: Liquid chromatography mass-spectrometry

LSC: Leukemia stem cell mtDNA: Mitochondrial DNA

ND: NADH dehydrogenase

NME: Nucleoside diphosphate kinase

NOD-SCID: Non-obese diabetic-severe combined immunodeficient

OCR: Oxygen consumption rate

PBSC: Peripheral blood stem cell

PGC-1α: Proliferator-activated receptor gamma co-activator 1α

PI: Propidium Iodide

Pi: Inorganic phosphate

POLA: Nuclear DNA polymerase alpha

POLG: Mitochondrial DNA polymerase gamma

POLRMT: Mitochondrial RNA polymerase

RUVBL: RuvB like AAA ATPase

SAINT: Significance Analysis of INTeractome

SCID: Severe combined immunodeficient

xiii SCF: Stem-cell factor

SDS: Sodium dodecyl sulphate shRNA: Short hair-pin RNA

SLC: Solute carrier transporters

TFAM: Mitochondrial transcription factor A

TK: Thymidine kinase

xiv 1. Introduction

1.1. Acute myeloid leukemia

Acute myeloid leukemia (AML) is a biologically heterogeneous hematopoietic malignancy arising in the bone marrow and spreading to peripheral blood (Rowe, 2010). This results in a decreased production of three main lineages of myeloid cells (granulocytes-mononocytes, red blood cells and platelets), accompanied by proliferation of non-functional progenitor cells termed blasts with an impaired capacity for differentiation (Corey et al., 2007) (Figure 1).

Clinically, AML has a rapid onset of symptoms representative of bone marrow failure such as fatigue, recurrent infections as a result of neutropenia, shortness of breath following exertion due to anemia, a tendency to bleed or bruise as a cause of thrombocytopenia and nonspecific symptoms such as tiredness and loss of appetite.

AML is the most common form of acute leukemia, presenting at all ages yet increasing in incidence with age(Löwenberg and Rowe, 2016). It occurs most commonly beyond 60 years of age with a median age of diagnosis at 72. The age-adjusted incidence rate of AML among adult leukemia >18 years of age lies between 30 to 40 per million, childhood leukemia (0-14 years) is approximately 7.7 per million (Khwaja et al., 2016). For adults under the age of 60, survival has significantly improved in the past 30 years, but similar advances have not been made in older patients. For patients younger than 60, 70-80% achieve complete remission. Yet most patients undergo relapse and 5-year overall survival rate stands at 40-50%. For patients over 65 years of age, overall survival (OS) has not improved in the past 30 years, with a 5-year survival rate less than 10% and median survival less than 1 year. (Khwaja et al., 2016; Roboz, 2012; Rowe and

Tallman, 2010).

1 For a majority of AML patients, risk factors associated with increased susceptibility to AML remain to be identified. However, the risk of developing AML is moderately higher by exposure to agents which damage DNA, such as cigarette smoke and therapy-related agents such as ionizing radiation and cytotoxic chemotherapy. Alkyating agents such as chlorambucil, cyclophosphamide and melphalan, and topoisomerase II inhibitors such as etoposide, mitoxantrone and anthracyclines are associated with development of therapy-related AML (Khwaja et al., 2016).

2 Figure 1: Normal and leukemic hematopoiesis A) In normal hematopoiesis, long-term hematopoietic stem cells have self-renewal potential and give rise to several hematopoietic progenitor cells. These progenitors have decreased self- renewal potential and can differentiate to form several mature myeloid or lymphoid lineage cells. B) In acute myeloid leukemia, leukemic stem cells are present at the top of the hierarchy with extensive self-renewal capacity, giving rise to leukemic progenitors and the bulk tumor cells termed blasts. These LSC’s have a block in differential potential and results in a reduction of normal myeloid lineage cells (Khwaja et al., 2016).

3 1.1.1 Classification

AML is a clonal, yet molecularly and cytogenetically heterogeneous disease. Patients diagnosed with AML undergo risk classification to guide therapeutic decision making. Until the past decade, prognostic classification of AML was based primarily on cytogenetic and morphological analysis of blast cells at time of diagnosis and clinical features (age, response to initial therapy, de novo versus secondary AML) (Reese and Schiller, 2013). However, given that a large majority of AML patients were cytogenetically normal (55%), treatment of this subgroup remained challenging.

Advances in next-generation sequencing approaches have generated a significant understanding of the molecular landscape of AML with normal karyotype. Following a landmark genome wide mutational analysis of cytogenetically normal AML patients study by Patel et. al. (Agaronyan et al., 2015), mutational analysis has been incorporated into prognostic risk classification. Prognostic risk classification is currently based on both mutational and cytogenetic profiles and can be stratified according to three risk groups: adverse-risk, intermediate-risk I, II and favorable-risk leukemia. (Komanduri and Levine, 2016) .

1.1.2. Stem cell characteristics of AML

The hematopoietic lineage is hierarchically organized, whereby hematopoietic stem cells (HSC’s) lie at the top of the pyramid and have the capacity for self-renewal and multipotency. Self-renewal is the ability to give rise to identical daughter cells. Multipotency is the capacity to differentiate into defined progenitor cells which give rise to multiple types of blood cells. AML is organized similarly with rare leukemic stem cells (LSC’s) at the apex of the hierarchy. This subpopulation has self-renewal capacity to give rise to the bulk tumor cells and displays resistance to cytotoxic chemotherapy (Dick and Lapidot, 2005). Leukemic stem cells share similar characteristics with

HSC’s, including quiescent cell cycle state, self-renewal capacity and resistance to apoptosis

(Basak and Banerjee, 2015).

The existence of LSC’s has been demonstrated over two decades ago in xenograft mouse models.

The primitive Lin-/CD34+/CD38- subpopulation of AML cells were able to give rise to the disease phenotype in recipient mice, while the more mature Lin-/CD34+/CD38+ or CD34+ cells were unable to generate tumor xenografts. The LSC subpopulation gave rise to the disease phenotype of the patient, and was also able to propagate secondary tumors in mice, indicating self-renewal capacity(Bonnet and Dick, 1997). However, a recent study has shown that the LSC fraction is not limited to the CD34+ subpopulation. Primary AML cells were sorted based on CD34 and CD38 surface marker expression and assayed for the ability to generate tumor xenografts. Most samples contained LSC’s in the CD34+/CD38- fraction. However, LSC’s were observed in atleast one other fraction in a majority of samples, indicating heterogeneity in AML LSC surface marker expression (Eppert et al., 2011).

Leukemic stem cell fractions display extensive heterogeneity in primary AML samples, functionally defined as cells capable of engraftment in NOD/SCID/IL2Rγc-deficient mice. LSC’s were enriched in Lin-, CD38-, fraction. However, LSC’s were also present in fractions associated with normal committed myeloid progenitors, such as Lin+/C38+ or Lin+/CD45RA+ cells.

Additionally, the LSC fractions were assessed for long term self-renewal capacity by engraftment in secondary recipient mice. Long term self-renewal capacity was present in all Lin, CD38 and

CD34 dim/high fractions. (Sarry et al., 2011)

Although the cell-of-origin of the LSC is currently debated, AML can arise out of either the HSC or hematopoietic progenitor cells which incur mutations and gradually incur additional genetic and epigenetic changes and progress to acute transformation (Khwaja et al., 2016). Clinically, analysis of phenotypically normal stem cells from AML patients showed the presence of DNMT3A mutations, an epigenetic modifier implicated in leukemogeneis which allows for a competitive survival advantage over normal HSC’s. This supports the hypothesis of the existence of pre- leukemic HSC’s which carry early driver mutations that allow for positive selection and expansion of this pool (Shlush et al., 2014). Subsequent additional pathogenic changes in these pre-leukemic

HSC’s may give rise to LSC’s and lead to AML.

1.1.2. Mechanisms

AML is characterized by numerous genetic, epigenetic and microenvironment changes resulting in deregulation of proliferation, cell survival and differentiation capacity. Although a significant proportion of AML patients have gross chromosomal abnormalities, over 50% of AML patients have a normal karyotype. Global analysis of mutational patterns associated with AML have discovered fewer coding sequence mutations compared to other solid epithelial cancers. Similar to

6 other cancers, multiple co-operating mutations are required for tumorigenesis. Mutations associated with disease pathogenesis can be functionally divided into eight categories with the following frequencies in patients: 1) Signaling pathways (59%): FLT3, KIT, KRAS and serine/threonine kinases , 2) DNA methylation (44%): DNMT3A, TET2, IDH1 and IDH2, 3)

Chromatin modifiers (30%): MLL, ASXL1 and EZH2, 4) Nucleophosmin (27%): NPM1, 5)

Myeloid transcription factors (22%): RUNX1, CEBPA1, 6) Transcription factors(18%): PML-

RAR1, MYHI1-CFP and RUNX1-RUNX1T1, 7) Tumor suppressors (16%): TP53, WT1 and

PHF6, 8) Spliceosome complex (14%): SRF2 and U2Af1, 9) Cohesin complex(13%): STAG2,

RAD21, SMC1 and SMC3 (Khwaja et al., 2016).

Genetic mutations in epigenetic regulators DNMT3A, TET2, WT1, IDH1 and IDH2 are functionally important for disease progression. These mutations are primarily implicated in regulation of DNA methylation and histone modifications and can affect mitochondrial function.

Briefly, DNMT3A is implicated in self-renewal of stem cells or progenitor cells and regulation of myeloid differentiation. TET2 enzymes regulate the first step of DNA demethylation via α- ketoglutarate and Fe(II)-dependent conversion of the DNA base 5-methylcytosine (5mC) to 5- hydroxymethyl cytosine (5-hmC). Although aberrant 5mC and 5-hmC levels contribute to leukemic transformation, the mechanisms are yet to be fully elucidated. Similarly, WT1 mutations are associated with lower levels of 5-hmC, suggesting functional similarities between TET2 and

WT1 mutations. IDH1 and IDH2 mutations are mutually exclusive with TET2 mutations, and mutations in IDH1 and IDH2 produce the oncometabolite 2-hydroxyglutarate (2-HG). 2-HG contributes to hematopoietic transformation partly by competition with α- ketoglutarate to inhibit

TET2 activity, leading to a hypermethylated state (Figueroa et al., 2010). IDH mutations are

7 associated with decreased mitochondrial function, as (R)-2-HG inhibited mitochondrial cytochrome c oxidase activity(Chan et al., 2015).

Mutations in signaling genes such as receptor tyrosine kinases (FLT3) promote leukemogenesis through inducing a proliferative advantage via the RAS– RAF, JAK–STAT, and PI3K–AKT signaling pathways. Mutations in myeloid transcription factors (RUNX1-RUNX1T1) regulate transcription and induces impaired hematopoietic differentiation. (center left box). Mutant NPM1, a multifunctional phoshoprotein which shuttles between the nucleus and cytoplasm, leads to aberrant cytoplasmic localization of NPM1 and NPM1-interacting proteins. Mutations of spliceosome associated lead to deregulation in RNA processing. Cohesin-complex gene mutations are associated with deregulated chromosome segregation and transcription. (Dohner et al., 2015)

In addition to intrinsic cellular deregulation, the extracellular niche environment plays an important role in AML disease pathogenesis. The bone marrow niche is a significant microenvironment that regulates many stem cell functions such as self-renewal, mobilization, engraftment, and differentiation. Mutations of non-hematopoietic cells of the marrow microenvironment are sufficient to cause hematopoietic transformation. For example, deletion of the RNase III endonuclease, Dicer1, in bone marrow osteoprogenitor cells produced myelodysplasia and an increased predisposition to AML in vivo (Yu and Scadden, 2016). Overall,

AML is a highly heterogenous disease characterized by various genetic and molecular changes.

Currently, novel therapeutic approaches are moving towards a personalized treatment approach to target the unique vulnerabilities of different AML subpopulations.

1.1.3. Current treatment strategies for AML

The standard of care in AML has remained largely unchanged for the past three decades and can be divided into induction, consolidation and maintenance therapy. Standard induction therapy consists of a combination treatment of the nucleoside analog cytarabine(AraC) and an anthracycline such as daunorubicin to induce remission. The standard 7+3 regimen consists of 7- day continuous infusion of cytarabine at 100 or 200mg/m2 daily on days 1 to 7 and daunorubicin at 60 mg/m2 on days 1-3 (Dombret and Gardin, 2016). The goal of induction therapy is to deplete the bone marrow of both malignant and benign cells (< 5% blast count), allowing hematopoiesis to repopulate the marrow with normal cells and leading to remission. Once remission is observed, additional treatments are required to target undetectable leukemic cells and lead to long-term cure.

Consolidation and maintenance therapies include additional chemotherapy treatment or allogeneic hematopoietic stem cell transplantation of the bone marrow, yet this technique carries a high morbidity risk due to chronic graft-versus-host diseases (Bonawitz et al., 2006). Consequently, allogeneic HSCT is applied based on risk-benefit ratio and considered standard care in patients with intermediate II-risk and adverse-risk AML after first complete remission, but is not advised for favorable-risk AML.

However, treatment often fails for a majority of patients due to relapse from complete remission instead of primary resistance to therapy or treatment-related mortality. This highlights the critical need for development of novel therapeutic strategies for AML.

1.1.4. Emerging therapies for treatment of AML

Novel agents in the treatment of AML target a variety of processes such as oncogenic signaling through tyrosine kinases, epigenetic deregulation, nuclear export of proteins, and antigens that are preferentially on leukemic cells by antibody-based therapy. A promising new targeted

9 approach is promoting differentiation through reprogramming metabolic and epigenetic changes of mitochondrial TCA cycle enzymes isocitrate dehydrogenase enzymes (IDH). Mutations in

IDH1 and 2 result in the overproduction of onco-metabolite 2-hydroxyglutarate (2-HG) instead of

α-ketoglutarate. 2-HG competes with native alpha-ketoglutarate to inhibit the activity of α- ketoglutarate-dependent DNA demethylation enzyme TET2. TET promotes DNA demethylation of methylcytosine to 5-hydroxymethylcytosine (Figueroa et al., 2010). This results in increased expression of stem/progenitor markers and impaired differentiation, leading to leukemogenesis.

Targeted IDH inhibitors of IDH enzymes have shown encouraging activity in phase 1 trials. For example, an IDH2 inhibitor, AG-221, displayed 40% overall response rate in relapsed or refractory AML by inducing differentiation in leukemic blasts of IDH mutated AML (Stein and

Tallman, 2016).

Novel drugs and biologics have been recently approved in the USA for treatment in relapsed/refractory AML. These include Volasertib, (a selective and potent Plk inhibitor which inhibits the cell cycle), Gemtuzumab / Mylotarg- (monoclonal antibody against CD33, conjugated with cytotoxic antibiotic calicheamicin), CPX-351 (a liposomal formulation cytarabine:daunorubicin), Venetoclax (BH3-mimetic Bcl2 inhibitor), Pracinostat (histone deacetylase inhibitor) and Midostaurin (FLT3 inhibitor) (Shafer and Grant, 2016). Additional emerging therapies include tyrosine kinase inhibitors such as FLT3 and KIT inhibitors; cell cycle and signaling inhibitors, such as PI3K, CDK and aurora kinase inhibitors; antibody drug conjugates against cell surface proteins such as CD33 (Dohner et al., 2015).

Survival rates in relapsed/refractory AML remain poor, despite variations in used salvage therapies. In a large international Phase III clinical trial of relapsed/refractory (R/R) AML, the efficacy of elacytarabine (a novel ester of cytarabine) was compared to one of 7 investigator’s

10 choice, including high dose cytarabine, multiagent chemotherapy (mitoxantrone,etoposide, and cytarabine), fludarabine, granulocyte colony-stimulating and idarubicin (FLAG-Ida), hypomethylating agents, hydroxyurea, or supportive care. No difference was observed between overall survival, response rate or relapse free survival between elacytarabine and the control arms.

Additionally, no differences between the OS of investigators choice was observed. (Roboz et al.,

2014) This study highlights the urgent need for novel treatment strategies for R/R AML.

1.2 Mitochondria

1.2.1. Structure and Function

Mitochondria are believed to have arisen two billion years ago when an α-proteobacterial cell began to live inside an archaeal host cell, an endosymbiotic relationship that was mutually beneficial (Margulis, 1970). Although the structure and composition has drastically changed, the core ancestral characteristic of the double-membraned wall and ATP production of the bacterium remained. Through the evolutionary process, a majority of the genetic material of the original α- proteobacterium was lost and the organelle acquired numerous new functions over time

(Bullerwell and Gray, 2004).

The diverse and essential roles of mitochondria include coordinating energy production through oxidative phosphorylation and the production of biosynthetic intermediates for cell proliferation including nucleotides, lipids and amino acids through the tricarboxylic acid (TCA) cycle (Figure

1.2). Mitochondria also regulate metal metabolism, synthesizing heme and iron-sulfur clusters which are important for hemoglobin to transport oxygen, oxidative phosphorylation and DNA repair (Stehling and Lill, 2013). Mitochondria also maintain cellular and mitochondrial redox state

11 to prevent accumulation of toxic levels of reactive oxidative species (ROS) generated through oxidative metabolism. One-carbon metabolism generates NADPH, which maintains the activity of anti-oxidant enzymes to regulate mitochondrial ROS (Gruning et al., 2011). Additional roles encompass regulating stress responses such as apoptosis and autophagy through remodeling mitochondrial structure (Twig and Shirihai, 2011)

Mitochondrial function is tightly linked to its form and architecture, which is dynamic and varies in response to meet cellular demands. Mitochondria consist of a double-membraned lipid bilayer wall, the outer membrane (OM) and the inner membrane (IM) separated by the intermitochondrial space. The inner membrane has two separate regions: the inner boundary membrane (IBM) and the cristae membrane (CM). The IBM is adjacent to the outer membrane whereas the CM encompasses the invaginations or tight folds of the IBM which face the mitochondrial matrix. The invaginations of the CM are proposed to increase efficiency of oxidative phosphorylation, and the surface area of cristae positively correlates with ATP-production by oxidative phosphorylation in numerous tissues (Zick et al., 2009). Additionally, components of respiration complexes such as

Complex V play a critical role in maintenance of inner membrane topology, and downregulation of respiration subunits altered mitochondrial cristae morphology. (Arselin et al., 2004).

Figure 1.2: Mitochondria function as bioenergetic, biosynthetic and signaling organelles Cancer cells catabolize pyruvate and glutamine through the TCA cycle, resulting in the generation of reducing equivalents such as NADH that donate electrons to the ETC. The ETC generates a proton gradient used for production of ATP, i.e., oxidative phosphorylation (blue). TCA cycle intermediates are also directed into biosynthetic pathways (green) that allow for the production of macromolecules (lipids, amino acids and nucleotides). Finally, mitochondrial production of ROS and metabolites act as signaling molecules to alter protein function. Mitochondrial one-carbon metabolism produces NADPH to prevent accumulation of ROS in the mitochondrial matrix. NADPH maintains antioxidant activity of glutathione peroxidase (GPX) and thioredoxin reductases (TrxRs). TA, aminotransferase; VDAC, voltage-dependent anion channel. TRX, thioredoxin; GSH, glutathione; SOD2, superoxide dismutase 2, mitochondrial; 5,10-CH2-THF: 5,10- methylene-tetrahydrofolate. (Weinberg and Chandel, 2015)

1.2.2. Mitochondrial DNA organization

The circular mitochondrial genome (5uM) is highly compacted and organized in large multiprotein-DNA complexes termed nucleoids. Nucleoids are structures with diameters between

100-200nm and are tethered between adjacent foldings of the inner mitochondrial membrane

(Bogenhagen, 2012; Gerhold et al., 2015; Huang da et al., 2009b). Although the composition of the mitochondrial nucleoid is currently poorly defined and a matter of debate, several studies have identified functional patterns of nucleoid-associated proteins. These include mtDNA replication and maintenance proteins, mitochondrial transcription and translation processes, chaperones and quality control proteases, mitochondrial ribosomal proteins, and lipid metabolism enzymes

(Bogenhagen, 2012).

Of the factors involved in mtDNA replication and maintenance, the most commonly identified proteins include TFAM, MTSSB and ETu. This is probable due to the high abundance of TFAM, which is involved in binding and unwrapping duplex mtDNA. Less commonly identified proteins include the core nucleoid proteins POLG, POLG2 and Twinkle (Hensen et al., 2014). The presence of abundant proteins can hinder the ability to detect other low-abundance proteins involved in mtDNA maintenance and repair. In addition to mtDNA replication, nucleoid preparations have detected proteins with mRNA and tRNA binding ability and transcription, such as LRPPRC

(Leucine-Rich Pentatricopeptide Repeat Containing) and EF-Tu. This suggests that mtDNA replication and transcription processes are linked and occur in close proximity to each other.

Similarly, numerous mitochondrial ribosomal proteins have been discovered in nucleoid preparations, indicating that mitochondrial translation may proceed adjacent to mitochondrial nucleoids. A recent study (Huang da et al., 2009a) has confirmed mtDNA nucleoids as sites which

14 initiate and support the assembly of mitochondrial ribosomes. Overall, these functional classes indicate mtDNA nucleoid is a critical centre which promotes expression of mtDNA-encoded genes and support mitochondrial biogenesis.

1.2.3. Mitochondrial biogenesis regulation

As mitochondria form a key metabolic hub, mitochondrial biogenesis is tightly regulated to facilitate a physiological response in response to various cellular signals. The components that detect energy status include transcription factors, hormones, cofactors, nuclear receptors, and kinases to detect levels of mitochondrial function and activity. One important factor is the AMP- activated protein kinase (AMPK), whose activity is upregulated by an increase in AMP/ATP ratios and increased ADP, which signifies an increase in energy expenditure and decreased intake of calories. AMPK phosphorylates multiple targets to upregulate oxidative phosphorylation, catabolic pathways such as gluconeogenesis and autophagy, while downregulating anabolic pathways to decrease energy expenditure, such as cell growth and proliferation pathways. The second metabolic sensor, Sirtuin 1 (Sirt1), is an NAD+ -dependent deacetylase, which senses changes in status of NAD+/NADH. NAD+ levels are increased upon starvation, leading to upregulation of Sirtuin, which co-operates with AMPK to activate the PGC-1 (peroxisome proliferator-activated receptor (PPAR)-γ) family of transcriptional co-activators (Hardie et al.,

2011; Jager et al., 2007). Interactions between PGC-1 and nuclear respiratory transcription factors

(NRF1, NRF2) upregulate the expression of nuclear-encoded genes involved in the regulation of mitochondrial mass and ATP production (Jeninga et al., 2010; Scarpulla et al., 2012).

1.2.4. Oxidative phosphorylation

Mitochondria are primarily known for ATP production through oxidative phosphorylation

(Oxphos). This involves the metabolism of multiple carbon-containing fuels such as glucose through glycolysis, amino acids such as glutamine and fatty acids to generate electrons. These fuels channel into the mitochondrial matrix to feed into the tricarboxylic acid cycle (TCA) enzymes. This anaplerotic cycle consists of the a series of enzymes (citrate synthase, aconitase, isocitrate dehydrogenase, a-ketoglutarate dehydrogenase,succinyl-CoA synthetase, succinate dehydrogenase, fumarase and malate dehydrogenase) which catalyze the oxidation of acetyl-CoA to produce carbon dioxide, guanosine triphosphate and two reducing agents, nicotinamide adenine dinucleotide and Flavin adenine dinucleotide (NADH and FADH2). These electron carriers transport electrons to electron transport chain (ETC) complexes. The ETC consists of four protein machines (I–IV), which undergo sequential redox reactions and conformational changes to pump protons across the matrix into the IMS. NADH is oxidized by transfer of electrons to complex I

(NADH dehydrogenase). FADH2 is oxidized by donating electrons to complex II (succinate dehydrogenase). Electrons from complexes I and II are transferred to coenzyme Q, then on to complex III (bc1 cytochrome c complex), complex IV (cytochrome c oxidase/COX) to the final electron acceptor oxygen to produce H2O. As electrons are transferred between complexes, the free energy released is used to pump protons across the mitochondrial inner membrane, creating an electrochemical gradient. The electrochemical gradient generated by the pumping of protons into the intermembrane space drives the phosphorylation of ADP to ATP by Complex V (F1-F0

ATP synthase) (Cermakian et al., 1997). (Figure 1.3)

In addition to ATP production, the electrochemical gradient produced by ETC activity is an essential mitochondrial feature. Membrane potential is required for other functions such as

16 mitochondrial protein import of nuclear-encoded proteins, and mitochondrial membrane depolarization acts as a trigger for mitochondrial degradation through mitophagy in response to mitochondrial dysfunction (Wallace, 2012) (Twig and Shirihai, 2011).

Figure 1.3: Mitochondrial DNA encoded genes in oxidative phosphorylation (Mishra and Chan, 2014)

1.2.5. Quality control of mitochondria

In response to environmental and metabolic stress, mitochondria undergo fission and fusion events to maintain functionality. Fusion is the process of merging of the outer and inner membrane of two distinct mitochondria to form one, while fission is the process of division of one mitochondria to form two. Fusion allows healthy mitochondria to compensate for damaged mitochondria by sharing functional mitochondrial components. However, if a particular threshold of damage is reached, mitochondria are eliminated by mitophagy. Fission separates the most dysfunctional and damaged mitochondrial components by segregation, prevents fusion with healthy mitochondria and induces mitophagy (Devineni and Gallo, 1995).

Mitochondrial fission and fusion are mediated by highly conserved large guanosine triphosphatases (GTPases) of the dynamin family. Fusion of mitochondrial outer membranes is regulated by membrane-anchored mitofusins (Mfn1 and Mfn2), while fusion between mitochondrial inner membranes is mediated by Optic Atrophy 1 (Opa1). Mitochondrial fission is catalyzed by dynamin-related protein 1 (Drp1) in combination with mitochondrial fission factor

(Mff). Drp1 wraps around constriction sites of the outer membrane and result in mitochondrial division (Clinton 2016)(Mishra and Chan, 2014) .

Factors involved in mitochondrial dynamics work in combination with the mitochondrial quality control factors to promote mitophagy. To induce removal of dysfunctional mitochondria, damaged mitochondria are sensed by a loss of mitochondrial membrane potential. This results in switch in localization of PINK1 (PTEN-induced putative kinase 1), an inner membrane space serine/threonine kinase which is imported into the mitochondria and constitutively degraded. In

19 response to mitochondrial depolarization, PINK1 import is inhibited and relocalized to the mitochondrial outer membrane, where it recruits the cytoplasmic E3 ubiquitin ligase Parkin. This results in ubiquitination of outer membrane mitochondrial proteins. This process flags the dysfunctional mitochondria for degradation through fusion with autophagosomes via mitophagy

(Devineni and Gallo, 1995).

1.2.6. Mitochondrial regulation of cell death

Mitochondria play a critical role in cell fate decisions through regulation of apoptosis, the process of programmed cell death. Apoptosis can be triggered through several signaling pathways including endogenous (such as DNA damage) or exogenous (such as growth factor withdrawal, exposure to cell death ligands, steroid hormones) triggers. These triggers can lead to the activation of caspases, a family of cysteine proteases, which are normally expressed at inactive enzymatic states (De Silva et al., 2015). The activation of initiator caspases (1, 2, 4, 5, 8, 9, 10, 11 and 12 in mammals) results in the cleavage of executioner caspases (3, 6, 7 and 14 in mammals).

Executioners caspases cleave a cascade of cellular proteins which can lead to apoptosis.

During intrinsic apoptosis, cell death signals cause the mitochondria to release pro-apoptotic proteins from the intermembrane space through mitochondrial outer membrane permeabilization

(MOMP). Key regulators of MOMP is the balance of the anti-apoptotic Bcl-2 (B-cell lymphoma

2) family proteins and the pro-apoptotic BH3 proteins. Inhibition of Bcl2 proteins results in the assembly of pro-apoptotic proteins, Bax, Bak and Bid, at the mitochondrial outer membrane space(Lee et al., 2009). At the membrane surface, these factors undergo conformational changes and oligomerization to form a mitochondrial pore resulting in MOMP. This membrane

20 permeabilization releases several proteins from the intermembrane space, including cytochrome c to form the apoptosome complex in the cytosol in combination with apoptosis activating factor-1

(Apaf-1) and pro-caspase-9. This leads to activation of caspase-9 and triggers the caspase cascade through by activation of pro-caspase-3 resulting in apoptosis (Figure 1.4).

The intrinsic pathways also runs through caspase-independent mechanisms. This process involves release of mitochondrial Apoptosis Inducing Factor (AIF) AIF induces nuclear chromatin condensation and production of high weight DNA fragments, leading to cell death.

Figure 1.4: Mitochondrial regulation of apoptosis Apoptosis is regulated by an interplay between the anti-apoptotic BCL-2 subfamily of proteins and the pro-apoptotic BH3-only subfamily. During apoptosis, BH3-only proteins facilitate a BAX (BCL-2-associated X protein)- and BAK (BCL-2 antagonist/killer)-dependent release of cytochrome c from mitochondria; cytochrome c binds to APAF1 and gives rise to the apoptosome. In parallel, X-linked IAP antagonists such as second mitochondria-derived activator of caspase (SMAC), HTRA2 and apoptosis-related protein in the TGF-β signalling pathway (ARTS) translocate from mitochondria and release caspases from negative regulation by IAPs. In particular, caspase 9 is liberated from X-linked IAP (XIAP) and activated by the apoptosome, which stimulates the executioner caspases 3 and 7. (Fuchs and Steller, 2015)

1.3. Mitochondrial DNA

1.3.1. Structure and function

The mitochondrial genome is a maternally inherited 16.6kb double stranded circular genome. It encodes 13 genes which form essential subunits of respiratory enzyme Complexes I/NADH dehydrogenase (ND) (7 peptides- ND1, ND2, ND3, ND4, ND4L, ND5 and ND6), Complex

III/cytochrome bc1 complex (1 peptide- CYTB), Complex IV/Cytochrome c oxidase (COX) (3 peptides- COXI, COXII, COXIII), and Complex V/F1 F0 ATP synthase (2 peptides- ATP6, ATP8) and non-protein coding genes, including 22 transfer RNA’s (tRNA’s) and 2 ribosomal RNA’s

(rRNA’s) required for intra-mitochondrial protein synthesis. (Figure 1.5). Unlike the nuclear genome, the mitochondrial genome lacks histones and introns and possesses a non-coding region termed the D-loop. (Gaziev et al., 2014), (Lee and St John, 2015). The D-loop also consists of two hypervariable regions, which have been employed to detect patterns of ancestry due to its strict maternal inheritance pattern. This region acts to regulate binding of nuclear-encoded transcription and replication factors to initiate replication and transcription in both strands of genes involved in mitochondrial metabolism. (St John, 2016).

In addition to the essential role of mitochondrial DNA (mtDNA) in regulation of mitochondrial metabolism, recent reports have implicated mtDNA with novel non-metabolic roles. MtDNA displays similarities with bacterial DNA, which contains inflammatogenic unmethylated CpG

(cytosine base followed immediately by a guanine) regions. Damaged mitochondria and its encompassing mtDNA are degraded by autophagy. However, if mtDNA fails to be degraded by autophagy, this leads to the induction of Toll-like receptor (TLR) 9-mediated inflammation in cardiomyocytes, leading to myocarditis and cardiomyopathy (Oka et al., 2012). Similarly,

23 deregulation of mtDNA packaging and mtDNA stress can also lead to anti-viral responses.

Deregulated mtDNA packaging promoted the escape of mtDNA into the cytoplasm, where it induced signaling and upregulated interferon-stimulated gene expression, leading to type I interferon responses and generating broad anti-viral resistance. (West et al., 2015) Overall, mitochondrial DNA plays critical roles in regulation of cellular metabolism and immunity.

Figure 1.5: Human mitochondrial genome map. Replication is initiated by transcription within the non-coding mtDNA displacement (D) loop and proceeds along the origin of replication of the heavy strand (OH) and the origin of the light-strand (OL). Human mtDNA has three proposed transcription promoters: heavy strand promoter 1 (HSP1), HSP2 and light stranded promoter (LSP). Single letters indicate the positions of the corresponding tRNA genes. MT: mitochondria; ND: NADH dehydrogenase genes; CYB: cytochrome b gene; CO, cytochrome c oxidase genes; ATP6/8, ATP synthase genes 6 and 8; 12S/16S, RNR: ribosomal RNA genes (Stewart and Chinnery, 2015)

1.3.2. Mitochondrial DNA replication

Mitochondrial DNA replication occurs independently of cell cycle status and is packaged within multi-protein structures termed nucleoids(Wang, 2010). Several factors have been implicated in mtDNA replication. The Mitochondrial transcription factor A (TFAM) binds enhancer regions located upstream of the heavy strand promoter and the light strand promoter to enhance mitochondrial DNA bending and unwinding. This reveals the transcription initiation site, where mitochondrial DNA replication is initiated by synthesis of an RNA primer by mt-RNA polymerase

POLRMT. The RNase mitochondrial RNA processing endonuclease and endonuclease G are involved in processing the precursor RNA primers for heavy-strand replication. Once the RNA primer is formed, it binds to the DNA strand forming a DNA-RNA-hybrid which is extended by

POLG and its accessory subunit POLG2.

Replication proceeds in tandem with mitochondrial DNA unwinding by the DNA helicase

Twinkle. The mitochondrial single-stranded binding protein (MTSSB) acts to stabilize single- stranded DNA at the replication fork and promotes the functional activities of the replication factors POLG and Twinkle(Lee and St John, 2015). The mitochondrial topoisomerase I

(TOPOImt) relaxes negative supercoils to remove positive supercoils created by DNA unwinding(Graziewicz et al., 2006).

The mechanism for mtDNA replication is currently debated and two main models prevail: the strand-displacement model and the strand-coupled model. In the strand-displacement model, replication begins at the origin of the leading strand (OH), elongating the heavy chain as a single strand. Replications proceeds two thirds of the way to expose the origin of the light chain (OL)

26 and initiates replication in the light chain (Lee and St John, 2015). The strand-coupled model is similar to nuclear DNA replication, where DNA synthesis is bidirectional and continuous, consisting of a leading and lagging strand with the formation of okazaki fragments. (Bailey and

Anderson, 2010). A complementary model to the strand-displacement model was proposed given a high proportion of ribonucleotides in the lagging strand. Termed RITOLS (RNA incorporated throughout the lagging strand) , long patches of RNA act to protect the displaced strand during one directional replication. The RNA transcript is then processed to leave behind short RNA templates that can be used as primers to complete replication of the lagging strand. This theory may explain the delay in synthesis between the heavy and light chains (Bailey and Anderson, 2010).

1.3.3. Mitochondrial transcription and translation

Mitochondrial DNA replication occurs in close proximity to mitochondrial transcription and translation events (De Silva et al., 2015). Transcription initiation requires the heavy-strand promoter (HSP) and light-strand promoter (LSP) present on opposite strands of mtDNA. The promoters are divided into two regions, one responsible for transcription initiation and the second serving as the upstream binding site for the transcription factor. (Lee 2015). This results in two polycistronic transcripts of almost genome length that are processed before addition of polyadenylation tail and translation. Excision of tRNAs from these polycistronic fragments results in the release of mature mRNAs and rRNAs.(Agaronyan et al., 2015).

The core mitochondrial transcription is carried out by the mitochondrial RNA polymerase

(POLRMT) which displays similarity to RNA-polymerase of bacteriophage (Cermakian et al.,

1997), the transcription factor TFAM, and TFBM1 and TFBM2 which display similarity to rRNA methyltransferases. TFAM-mediated transcription is activated by its ability to induce a “U-turn” in the TFAM responsive element of mtDNA once bound. TFB2M activates transcription by assisting in maintaining an open complex formation during transcription initiation (Sologub et al.,

2009). To carry out transcription, POLRMT ﬁrst binds the mtDNA promoter and initiates transcription de novo by catalysis of phosphodiester bond formation between two ribonucleotides(Arnold et al., 2012). Termination of transcription is regulated by several termination factors including MTERF1,2 and 3.

After transcription, mitochondrial translation progresses on mitochondrial ribosomes which reside in the matrix and associate with the inner mitochondrial membrane. Mitoribosomes are composed of ribosomal RNAs and mitochondrial ribosomal proteins (MRP) and comprise two subunits, the small (28S, SSU) and the large subunit(39S/LSU). Translation can be divided into three phases similar to the cytosolic process: initiation, elongation and termination. Mitochondrial transcripts are unique in that they lack the Shine Delgarno sequence of prokaryotes nor the 7-methylguanlyate cap structure found in eukaryotic transcripts to initiate ribosome binding. Initiation begins with binding of mRNA to the small subunit, followed by fMet-tRNA binding to the peptidyl site (P site). With the assistance of translation initiation factors MtIF1,2,3, the mRNA is correctly place the peptidyl site of the mitoribosome with the start codon(Smits et al., 2010). Elongation factor mtEFTu forms a complex with GTP and an aminoacylated tRNA, carries it to the mitoribosomal acceptor site (A site). Once codon-anticodon binding occurs with the mRNA template strand, GTP hydrolysis of mtEFTu is stimulated by the mitoribosome to form mtEfTu/GDP. This catalyzes the aminoacyl-tRNA to move into the peptidyl transferase center of the LSU. (Chiron et al., 2005) At

28 this site, peptide bond formation is catalyzed, adding on one amino acid to the polypeptide chain to elongate the chain (Smits et al., 2010). mtEF-Tu/GDP is recycled by the guanine nucleotide exchange factor EF-Ts. Lastly, termination of translation begins when a stop codon enters the A site. A mitochondrial release factor promotes protein release from the P site. After release of the newly synthesized protein, mitochondrial release factors (mtRRF) catalyze the dissociation of mitoribosomal subunits, tRNA and mRNA, allowing the machinery to be free to perform of another round of protein synthesis.

1.3.4. Regulation of mitochondrial DNA content

Each mitochondrion contains between 2-10 copies of mtDNA and each cell can contain between

102 to 105 copies. The large variation in mitochondrial copy number depends on multiple factors which are poorly understood(Lee and St John, 2015). Currently, information on regulation of mtDNA content in physiologically relevant tissue models is lacking (Spelbrink, 2010). Some reports suggest mtDNA is tailored to match the cellular energy demands, which is dependent on tissue type and developmental stage. In tissue types highly dependent on energy, such as the heart, liver,muscle and neurons, mtDNA is essential for function. In contrast, other cell types which do not require large respiratory activity and rely on glycolysis, such as spleen and sperm cells, have lower mtDNA content. MtDNA content regulation is also essential for embryogenesis and development. If mtDNA copy number does not exceed a critical threshold, the resulting effect is a failure of oocytes to complete fertilization or early-stage arrest in development (Lee and St John,

2015; St John, 2016).

MtDNA content is also regulated by mitochondrial replication and biogenesis factors. These include the core mtDNA replication factors POLG2, Twinkle and TFAM, whereby overexpression of these factors upregulated mtDNA content in cell culture models(Ikeda et al., 2015; Spelbrink,

2010). TFAM expression is also controlled by factors which regulate mitochondrial biogenesis such as peroxisome proliferators-activated receptor gamma coactivator 1 alpha (PGC1a) and nuclear respiratory factors (NRF).

Access to mitochondrial replication factors are also affected by mitochondrial fission and fusion events. Additional mtDNA replication regulation include environmental factors such as exposure to oxidative damage and mtDNA damage, resulting in a compensatory upregulation of mtDNA content (Gaziev et al., 2014). Lastly, replication of the mitochondrial genome is critically dependent on the availability of sufficient nucleotide pools (Carling et al., 2011).

1.3.5. Mitochondrial nucleotide pool maintenance

Mitochondrial DNA biosynthesis requires sufficient nucleotide pools to carry out DNA replication. Nucleotide pools within the cell are generated by mitochondrial and cytoplasmic pathways. Nucleotides can be imported from the cytosol into the mitochondria, and secondly nucleoside precursors are converted to nucleotides within the mitochondria through phosphorylation by the mitochondrial nucleotide salvage pathway (Carling et al., 2011) (Figure

1.6). This pathway contains two of the four cellular deoxynucleoside kinases- Mitochondrial thymidine kinase 2(TK2) and deoxyguanosine kinase (dGUOK). These enzymes catalyze the conversion of deoxynucleosides to deoxynucleoside monophosphates in an ATP-dependent fashion to provide all 4 deoxynucleoside monophoshate (dNMP) precursors. TK2 is specific to

30 thymidine and deoxycytidine and dGUOK is specific for deoxyguanosine and deoxyadenosine

(Mathews 2007, Mathews 2015). Following monophosphorylation, dNMP’s undergo di- and triphosphorylation, however the enzyme kinase cascade has been poorly investigated. It is suggested that putative candidates for monophosphate kinase activity include a mitochondrial adenylate kinase (AK) isoform, cytidine monophosphate kinase 2 (CMPK2), and thymidine monophosphate kinase 2 (TMPK2) which act on dAMP, dCMP, and dTMP, respectively (Gandhi and Samuels,

2011b). Collectively, the mitochondrial nucleotide salvage kinases support mitochondrial nucleotide pools for mtDNA replication.

In addition to mitochondrial nucleotide salvage, nucleotide import from the cytosol is necessary to support mitochondrial DNA biogenesis. Multiple lines of evidence support cytoplasmic nucleotide supplementation towards mitochondrial nucleotide pools in normal cells. To synthesize nucleotide pools at physiologically relevant rates, nucleotide salvage enzymes must catalyze products rapidly to support mitochondrial DNA biosynthesis. Analysis of mitochondrial nucleoside kinase enzyme kinetic constants does not support rapid enzyme-substrate catalysis of nucleotides required for this process. Secondly, mitochondrial nucleoside kinases have broad range specificity and can also bind ribonucleotide precursors. As ribonucleosides are present in the cell at larger concentrations than deoxyribonucleotides, ribonucleosides compete with deoxyribonucleosides as substrates for NSP enzymes, leading to slower rates of nucleotide biosynthesis. (Gandhi and Samuels, 2011a). Thus, supplementation from additional cytoplasmic nucleotide pools are required for maintenance of mitochondrial nucleotide pools.

Additionally, mass spectrometric analysis of cytoplasmic and mitochondrial dNTP pools in normal and cycling human lung and skin fibroblasts indicate a strong correlation between mitochondrial dNTP pool sizes and cytoplasmic dNTP pool sizes (dATP, dTTP and dGTP) and cell cycle (Ferraro et al., 2005; Pontarin et al., 2006). These results demonstrate nucleotide flux between the cytoplasm and mitochondria. (Gandhi and Samuels, 2011b).

In transformed cells, the contribution of cytoplasmic nucleotide pools to maintenance of mitochondrial nucleotide pools is unclear. Although the association between mitochondrial and cytoplasmic pools are significant in normal cycling cells, a clear association is not observed in transformed cells (Gandhi and Samuels, 2011b). Analysis of human osteosarcoma (HOS), mouse

ﬁbroblasts (3T3) (Rampazzo et al., 2004) and Hela Cells (Bestwick et al., 1982) demonstrated no significant correlation in cytoplasmic and mitochondrial pool sizes. In contrast, a disproportionally higher dNTP pool size was observed in the cytoplasm compared to the mitochondria. These results suggest that communication between cytoplasmic and mitochondrial pools are deregulated in cancer.

Further evidence for cytoplasmic nucleotide supplementation is observed in cycling cells. In TK1 proficient HOS cells, knockdown of TK2 resulted in no change in total phosphorylation of radiolabeled thymidine triphosphate in cytoplasmic and mitochondrial dTTP pools, reflecting a compensatory nucleotide salvage by TK1. In TK1 knockdown cells HOS cells, knockdown of TK2 inhibited the salvage of thymidine to thymidine triphosphate, however the size of the mitochondrial dTTP remained unaltered. This is further evidence that in cycling cells, a significant source of mitochondrial nucleotides is cytoplasmic de novo synthesis of dTTP (Rampazzo et al.,

2007). Thus, cytoplasmic nucleotides contribute to mitochondrial nucleotide metabolism.

Figure 1.6: Mitochondrial nucleotide biosynthesis Nucleotides are synthesised in the parallel in both the cytosol or mitochondria to support mtDNA replication. In the cytosol, small molecule precursors such as amino acids and carbohydrates are reduced to deoxyribonucleotide diphosphates by R1/p53R2 in the de novo biosynthesis pathway, which then feed into the nucleotide salvage pathway. Deoxynucleosides (dTMP, dCMP, dGMP, dAMP) are monophosphorylated in the cytosol by TK1 and dCK, in the mitochondria by dGK and TK2. The monophosphates (dNMP) are converted to dNTPs by two additional phosphorylation steps, nucleoside monophosphate kinases (NMPK) and nucleoside diphosphate kinases (NDPK). Deoxynucleosides and deoxynucleotides are transported across the inner mitochondrial membrane by specialized nucleotide transporters (Wang, 2010)

1.3.5. Cytoplasmic nucleotide metabolism

1.3.5.1. De novo biosynthesis

Maintenance of the genome requires sufficient pools of deoxyribonucleotide triphosphate (dNTP) pools to support both nuclear and mitochondrial DNA biosynthesis. It is highly regulated by two processes: the de novo biosynthetic pathway (which occurs in the cytoplasm) and the nucleotide salvage pathway (NSP) (which occurs concurrently in the cytoplasm and mitochondria). The de novo biosynthetic pathway is responsible for the majority of nucleotide biosynthesis during S- phase (Rampazzo et al., 2010). In the cytoplasm, de novo nucleotide biosynthesis utilizes amino acids and carbohydrates to synthesize ribonucleoside diphosphates (NDP’s) in an energy consuming process. These NDP’s are then reduced by ribonucleotide reductase (RNR) to deoxyribonucleosides (dNDP’s). RNR is a heterotetramer of large subunit (protein R1) and two a small subunit (R2 or p53R2). Lastly, dNDP’s synthesized from the de novo biosynthesis pathway feed into the nucleoside salvage pathway. Here, dNDP’s are converted to dNTP’s through the final phosphorylation step through NSP kinases.

1.3.5.2. Cytoplasmic nucleotide salvage pathway

Similar to the mitochondria, cytoplasmic nucleoside precursors are converted to nucleotides through a cascade of kinases in the cytoplasmic termed the cytoplasmic nucleotide salvage pathway. This process includes nucleoside kinases deoxycytidine kinase (DCK) and thymidine kinase 1 (TK1). DCK converts deoxycytidine, deoxyguanosine and deoxyadenosine to their monophosphate forms, while TK1 converts deoxythymidine to deoxythymidine-monophosphate.

Of the nucleoside kinases, TK1 is highly cell cycle regulated, while the activity of the other kinases appear to be constitutively expressed and partially influenced by cell proliferation state and

34 metabolic stress. An upregulation of dCK activity was observed following treatment of cells with nucleoside derivatives and genotoxic agents, such as aphidicolin, etoposide, taxol, and gamma irradiation. The increased dCK activity following toxic treatment suggests a compensatory mechanism induced by metabolic stress to upregulate production of nucleotides (Staub and

Eriksson, 2007) .

dNMP’s undergo sequential di- and triphosphorylation by nucleoside monophosphate kinases

(NMPK) and nucleoside diphosphate kinases (NDPK) to form dNTP’s. There are 4 classes categories of monophosphate kinases: thymidylate kinase (DTMPK), UMP-CMP kinases

,adenylate kinases (AK1, AK2, AK4, AK5,AK7, AK8) and guanylate kinases or

(GMPK/GUK1). Lastly, dNDP’s are converted to dNTP’s by nucleoside diphosphate kinases

(NME1, NME2, NME3, NME4, NME5, NME6, NME7, NME1-2) (Fan et al., 2013; Viale et al.,

2014). In addition to nucleotide biosynthesis, additional levels of nucleotide pool regulation include nucleotide degradation by 5-nucleotidaseswhich facilitates dephosphorylation of nucleotides to form nucleosides (Mathews, 2015).

In AML cells, the contribution of the de novo biosynthesis pathway and nucleotide salvage pathway to maintain cellular nucleotide pools is poorly characterized. There is evidence to suggest increased rates of nucleotide biosynthesis in AML, analysis of purine nucleotides (such as ATP,

ADP, GTP, GDP) in three leukemic cell lines (HL-60, THP1, HEL) were two to ten-fold higher in AML cell lines relative to healthy human lymphocytes cells (Baranowska-Bosiacka et al.,

2005). Other reports suggest that AML cells rely on de novo nucleotide biosynthesis for generation of up to 70% of nucleic acids, while bone marrow cells in comparison account for approximately

30% (Sugiura et al., 1986). Overall, increased nucleotide biosynthesis is observed in actively replicating tumor cells.

1.3.6. Mitochondrial nucleotide transport

Several studies have demonstrated evidence for transport of nucleotides from the cytoplasm to support mitochondrial nucleotide metabolism. In vitro, proteoliposomes which contain the mitochondrial compartment of human acute lymphocytic leukemic cells displayed deoxycytidine triphosphate transport activity in proteoliposomes containing the mitochondrial protein fraction of human AML cells (Bridges et al., 1999). The nucleotide precursors deoxythimidine monophosphate displayed ability for transportation into and out of mitochondria isolated from mouse liver cells in vitro (Ferraro et al., 2006), suggesting that cytoplasmic and mitochondrial pools are co-dependent.

As the mitochondrial nucleotide pool is separated by a hydrophobic double membrane, mitochondria require specialized transporters for import of deoxynucleosides and deoxynucleotides. The human equilibrative nucleoside transporter (hENT) family contains 4 isoforms hENT1-4 specialized for facilitated diffusion of nucleosides and nucleotides down a concentration gradient and are distributed at the plasma membrane and mitochondrial membrane

(Choi and Berdis, 2012). hENT1(Lai et al., 2004) and hENT3 (Govindarajan et al., 2009) have been detected in the mitochondrial membrane and facilitate nucleotide transport. hENT4 is also localized in intracellular organelles ((Endo et al., 2007), however is yet to be functionally characterized.

Additional mitochondrial pyrimidine transporters discovered recently include the solute carrier family 25 proteins SLC25A33 and SLC25A36. SLC25A33 expression is upregulated by the insulin-like growth factor signaling factor and its expression is upregulated in transformed fibroblasts, cancer cell lines, and primary prostate cancer compared to normal tissues (Di Noia et al., 2014). Additionally, genetic knockdown of SLC25A33/PNC1 decreased mtDNA content, oxidative phosphorylation, mitochondrial biogenesis and invasive cancer phenotypes in cell culture models (Favre et al., 2010). These data demonstrate that mitochondrial nucleotide metabolism is supported by transport of cytoplasmic macromolecules and is tightly linked to maintenance of mitochondrial function.

1.3.7. mtDNA mutations

Mitochondrial DNA is present in multiple copies within the mitochondria at varying proportions of wild-type and mutant forms known as heteroplasmy. Heteroplasmy is observed in almost all healthy individuals at low basal rates. Increased rates of mtDNA mutations (germline mutations) are considered pathogenic if observed at rates beyond a threshold required to induce respiratory chain dsyfunction, typically beyond 80% (Stewart and Chinnery, 2015) Germline mitochondrial

DNA mutations can be induced by POLG-associated diseases (see section 1.4.2.2) Somatic mutations in the mitochondrial genome have also been reported in cancer. Although some reports suggest mutations are associated with cancer progression and pathogenesis, a study by Ju et. al. performed mtDNA mutation profiling of 1675 human tumors and revealed a strong bias towards mutations in the leading strand of replication. This indicates that mutations arise primarily out of errors during the replication process, not as a consequence of exposure to reactive oxygen species, contrary to conventional wisdom. The study also revealed mtDNA mutations that are detrimental

37 to respiratory activity are selected against in cancer cells, and cells with functional mitochondria maintain a survival advantage (Ju et al., 2014). In this study, fewer mtDNA substitutions were observed in AML tumors compared to other cancer types, suggesting a higher dependence on functional mitochondrial activity in AML.

1.3.8. mtDNA depletion syndromes

Reductions in mtDNA copy number with no qualitative defects such as deletions or point mutations is a hereditary human mitochondrial disorder known as mitochondrial depletion syndrome (MDS). This disorder is generally observed in specific tissues such as the muscle, liver and central nervous system. In addition to POLG, mutations in 8 other nuclear-encoded genes have been associated with MDS; these include mitochondrial DNA salvage enzymes thymidine kinase

2 (TK2) and deoxyguanosine kinase (dGUOK), p53 inducible ribonucleotide reductase small subunit (p53R2- involved in de novo DNA nucleotide biosynthesis) , succinyl-CoA ligase beta subunit (SUCLA2-), succinyl-CoA ligase alfa subunit (SUCLG1), Twinkle gene (mitochondrial

DNA helicase), and MPV17 (unknown function). These genes are primarily responsible for mtDNA biogenesis or maintenance of deoxynucleotide pools (Wang, 2010).

1.3.9. Mitochondrial DNA mutations in hematological disorders

Pearson syndrome a multisystem mitochondrial disorder caused by rearrangements of mitochondrial DNA, leading to defects in the mitochondrial electron transport chain. Pearson syndrome is associated with refractory, hypoplastic macrocytic anemia, lactic acidosis, anemia is associated with thrombocytopenia and neutropenia, and pancreatic dysfunction (Farruggia et al.,

2016).

1.4. Mitochondrial DNA polymerase gamma

1.4.1. Structure and function

The mitochondrial DNA polymerase was first identified in Saccharomyces cerevisiae (Foury,

1989) and subsequently been determined in mammalian species. Mitochondrial DNA is replicated in mammals uniquely by the mitochondrial DNA polymerase gamma (POLG) in humans. The holoenzyme is a heterotrimer which consists of the primary subunit POLG and a homodimeric form of its accessory subunit POLG2 (Chan and Copeland, 2009). POLG belongs to the family A

DNA polymerases, which include Klenow Poymerase I from E. coli and the T7 bacteriophage

DNA polymerase, which have roles in both replication and repair. (Spelbrink, 2000).

The nuclear-encoded 140kDa primary subunit (encoded by POLG at 15q25) contains a C-terminal catalytic polymerase domain and N-terminal exonuclease domain separated by a linker/spacer region (Lee et al., 2009). The polymerase domain is responsible for extension of the DNA strand by catalyzing the addition of deoxynucleotide triphosphates (dNTP’s) to the growing chain. POLG also performs repair roles to increase fidelity of DNA replication, which is the ability to discriminate against the incorporation of improper base pairs at the enzymatic active site. If errors occur, the 3′–5′ exonuclease activity cleaves mismatched nucleotides from the end of the polynucleotide chain. Additionally, POLG has also 5'-deoxyribose-5-phosphate (dRP) lyase activity to increase fidelity of DNA replication, by releasing 5’- dRP sugar moieties from sites of genetic lesions (Graziewicz et al., 2006).

The accessory 55kDa subunit POLG2 is involved in regulation of DNA binding to increase processivity from 50 to hundreds and thousands of nucleotides compared to catalytic subunit alone

(Yakubovskaya et al., 2006). Alignments of the primary protein sequence reveal that the catalytic subunit sequence is highly conserved across species, with all genes comprising conserved sequence motifs for catalytic and 3′–5′ exonuclease functions.(Chan and Copeland, 2009).

As the mitochondrial genome encodes genes essential for electron transport chain activity, POLG indirectly regulates oxidative phosphorylation. The expression of POLG is regulated through epigenetic mechanisms, decreased intragenic DNA methylation at CpG islands of exon 2 correlates with increased POLG mRNA expression (Kelly et al., 2012). Additionally, the myc transcription factor stimulates expression by binding at POLG promoter regions (Kim et al., 2008)

1.4.2. Functional effects of targeting POLG

1.4.2.1. Mouse models of POLG

The physiological effects of targeting POLG have been investigated in several mouse models. In the homozygous knockout of POLG mouse model, early developmental arrest of mouse embryos with significant reductions in mtDNA content were observed, indicating that POLG is essential for embroyogenesis. In contrast, POLG+/- heterozygous mice undergo normal development with slight reductions in mtDNA content, indicating that cells can tolerate depletion in mtDNA unless a critical pathogenic threshold is reached (Cermakian et al., 1997; Hance et al., 2005).

A second model termed the mtDNA-mutator mice are homozygous for a mutation in the POLG exonuclease domain, leading to inactivation of the proofreading and exonuclease function of

POLG. These mice display 10-fold higher increase in mtDNA mutation frequencies compared to normal tissue, an age-related decline in respiratory complexes and premature aging phenotypes.

Thus, increased mtDNA mutations leads to respiratory deficiency and associated with accelerated

40 aging in mouse models(Foury, 1989). The mtDNA mutator phenotype has also been characterized in the hematopoietic lineage of these mice. Although accumulation of mtDNA mutations did not functionally affect the HSC pool, defects in hematopoiesis was observed. Intermediately aged mutator mice displayed increased pre-megakaryocytes and erythroid committed progenitors, indicating an inhibition of differentiation at early stages of erythroid development and a defect in erythropoiesis. Developmental defects induced by the mtDNA mutator phenotype was associated with decreased mitochondrial membrane potential and increased rates of apoptosis when erythroid committed progenitors cells were cultured in vitro. Thus, maintenance of mitochondrial function through the activity of POLG is essential for development of multipotent stem cell differentiation and mitochondrial metabolism (Cloutier and Coulombe, 2010; Norddahl et al., 2011).

1.4.2.2. POLG related diseases

Germline mutations in POLG are common in a wide range of human mitochondrial diseases.

Currently, over 150 mutations in POLG have been associated with disease (http:// tools.niehs.nih.gov/polg/). POLG-associated diseases are characterized by mtDNA depletion, deletions and point mutations, resulting in dysfunctional respiratory activity (Tang et al., 2011).

Patients with POLG mutations also display abnormal changes in mitochondrial morphology, such as enlarged mitochondria with tubular or concentric cristae (Müller‐Höcker et al., 2011; Wallace and Fan, 2009) . Thus, POLG plays a critical role in maintenance of mtDNA, respiration, mitochondrial structure and function.

The effects of POLG-related disorders in humans include clinical manifestations such as lactic acidosis, seizures, ataxia, neuropathy, myopathy and liver failure. The most common POLG- related diseases associated with these symptoms are forms of familial autosomal dominant progressive external ophthalmoplegia, Alpers syndrome, spinocerebellar ataxia with epilepsy

(SCAE), sensory ataxia neuropathy with or without ophthalmoplegia (SANDO) and chronic progressive external ophthalmoplegia (CPEO)(Hance et al., 2005; Tang et al., 2011). Similar clinical features are also observed in NRTI-associated toxicities, which include lactic acidosis, hepatoxicity, neuropathy and cardiomyopathies. These mitochondrial toxicities may be attributed to higher rates of mtDNA turnover and higher mitochondrial mass in the liver, brain and muscle, indicating a higher dependence on mitochondrial function (Bienstock and Copeland, 2004; Wang,

2010).

1.4.2.3. 2’3’-dideoxycytidine (ddC) ddC/zalcitabine is a pyrimidine nucleoside analog of deoxycytidine which lacks the 3’-hydroxyl group. ddC acts as an anti-metabolite and mimics native nucleosides where it is transported into the cytoplasm by several nucleoside transporters, including human equilibrative transporters hENT2 and human concentrative transporter hCNT1 (Yao et al., 2001) . In the cytoplasm, ddC undergoes a series of sequential phosphorylation steps by nucleoside kinases to its active triphosphate(TP) form ddCTP (Figure 1.7). The rate limiting step in activation of ddC is monophosphorylation to ddC-monophosphate by cytoplasmic deoxycytidine kinase (DCK), an enzyme of the nucleotide salvage pathway(Chen and Cheng, 1992). It is then phosphorylated by

CMPK1, followed by nucleoside diphosphate kinases (NME proteins). Following activation, it is imported to the mitochondria through transporters that have yet to be functionally characterized and exerts its inhibitory effect on POLG catalytic activity. ddC can be inactivated by deamination to dideoxyuridine by cytidine deaminase, however is not significantly metabolized in humans as high rates of ddC were recovered unmetabolized in urine (Devineni and Gallo, 1995). ddC can also be inactivated by dephosphorylation through the activity of phosphorylases.

POLG displays increased sensitivity to ddC compared to nuclear DNA polymerases. The inhibitory constant of ddCTP for POLG over nuclear polymerases α,β and Ɛ is atleast 100 fold greater as determined by in vitro primer extension assays (Martin et al., 1994). The molecular basis for heightened sensitivity is due to differences in primary amino acid sequence of the polymerase active site involved in dideoxynucleotide (ddNTP) discrimination. Specifically, DNA Polymerase

I, which is able to discriminate against ddNTP incorporation, has a phenylalanine residue on the highly conserved O-helix, opposite the 3′-OH of the bound dNTP. In contrast, Family A polymerases that incorporate ddNTPs, such as T7 DNA polymerase, have a tyrosine at the equivalent position on the O-helix. The phenol group of the tyrosine residue 951 of POLG appears to substitute for the missing hydroxyl group in ddNTP’s, allowing for its incorporation. Additional mechanisms for ddC sensitivity and toxicity include rates of activation, inactivation, and rates of cellular and mitochondrial transport (Bailey and Anderson, 2010).

Figure 1.7: ddC (2’3’-dideoxycytidine) metabolism ddC is transported into the cytoplasm through nucleotide transporter proteins. Following import, ddC is activated by three phosphorylation steps to its triphosphate form ddC-trisphosphate (ddC-PPP) through the activity of cytoplasmic nucleotide salvage pathway kinases. ddC can also be inactivated by nucleotidases and deaminases. Following activation, ddC-PPP is imported into the mitochondria where it inhibits replication activity of POLG.

1.4.2.4. Inhibition of POLG by anti-viral nucleoside analogs

POLG is uniquely sensitive to FDA-approved anti-viral drugs termed nucleoside reverse transcriptase inhibitors (NRTI’s), originally designed to target HIV reverse transcriptase. HIV patients treated with NRTI’s commonly experienced mitochondrial toxicities, leading to the discovery of cross-reactivity to POLG (Bienstock and Copeland, 2004). The mechanisms of

POLG inhibition of NRTI’s include 1) competition with native nucleotides causing chain termination due to lack of the 3’-OH (most prevalent), 2) direct inhibition of POLG, 3) inefficient excision of NRTI’s and persistence in mtDNA, 4) impairing fidelity of DNA synthesis by inhibiting POLG exonuclease activity. NRTI-induced POLG inhibition is primarily evaluated using mtDNA depletion as a surrogate marker and has been characterized in vitro according to decreasing severity as: 1) zalcitabine (2'-3'-dideoxycytidine/ ddC), 2) didanosine (DDI), 3) stavudine (2'-3'-didehydro-2'-3'-dideoxythymidine, d4T), 4) lamivudine (2',3'-dideoxy-3'- thiacytidine, 3TC) , 5) zidovudine (AZT/ZDV), 6) abacavir (Setzer et al., 2005).

Defects in mtDNA replication and mitochondrial function have been clinically observed in a subset of HIV patients treated with anti-viral nucleoside analogs/NRTI’s (Darbyshire, 1996)(Kohler et al., 2006). Anti-viral nucleoside analogs designed to target HIV reverse transcriptase enzyme also cross-reacted with human POLG. HIV patients treated with anti-viral regimens displayed mtDNA depletion, loss of mtDNA-encoded proteins, defective respiratory chain enzymes and ETC activity in tissues preferentially targeted by NRTI’s (Kallianpur and Hulgan, 2009; Lewis et al., 2003). A majority of NRTI-associated clinical manifestations are similar to mitochondrial diseases. Dose- limited toxicities associated with ddC primarily consisted of peripheral neuropathy in a subset of patients (30%), a primary issue associated with discontinuation of ddC as a HIV

45 treatment(Anderson et al., 2004). Pharmacogenomic analysis of patients treated with NRTI’s revealed peripheral neuropathy was associated with the incidence of mitochondrial genomic mutations. Specifically, mutations in haplogroup T, characterized by point mutation 7028C>T,

10398G>A, and 13368G>A and two non-synonymous mitochondrial DNA polymorphisms,

4216C and 4917G were independently associated with peripheral neuropathy. Thus, mutations in the mitochondrial genome can increase susceptibility to NRTI`s. Although the molecular mechanisms underlying functional differences between haplogroups are poorly elucidated, the polymorphisms in the mitochondrial genome can deregulate efﬁciency of oxidative phosphorylation, triggering the clinical onset of peripheral neuropathy following exposure to some anti-viral agents(Tozzi, 2010).

1.5. Targeting mitochondria as a therapeutic strategy

1.5.1. Targeting mitochondria in cancer

A hallmark characteristic of cancer cells is reprogramming metabolism to support indefinite growth and proliferation. One of the most common tumor metabolic phenotypes observed is rapid glucose uptake and production of lactate in the presence of oxygen, known as the Warbug effect or aerobic glycolysis. This is characterized by the shift of ATP production from oxidative phosphorylation to glycolysis even under aerobic conditions (Ruzzenente et al., 2012). This observation led to the long-standing assumption that tumors contain dysfunctional mitochondria, and require glycolysis as the primary source of ATP for proliferation. However, emerging studies have challenged this notion. The current consensus states that tumors use both glycolysis and oxidative phosphorylation for production of ATP and macromolecule biosynthesis, where most of tumor ATP is produced in the mitochondria (Weinberg and Chandel, 2015). Specifically, ATP

46 production routes was assessed using isotope-tracer studies in immortalized murine cell culture models. Oxidation of glutamine, glucose, and acetyl-CoA generated 60, 30, and 10% of the total

NADH/FADH2 content respectively, the reducing equivalents used for ATP production through oxidative phosphorylation(Fan et al., 2013). Hence, glutamine-driven TCA cycle intermediates coupled to oxidative phosphorylation is a primary driver of bioenergetic processes in subtypes of cancer.

An essential role for oxidative phosphorylation in carcinogenesis has been demonstrated in vivo

(Tan et al., 2015). In this study, two murine cancer cell lines (B16 melanoma and 4T1 breast carcinoma) were depleted of their mtDNA content to generate rho zero cells in vitro, characterized by loss of functional respiration enzymes and oxidative phosphorylation. When tumor-forming capacity was assessed in vivo, the rho zero lines formed tumors with a longer latency of several weeks compared to mtDNA-containing (rho +) parental control tumors. Once the rho zero tumors emerged, they divided as rapidly as rho+ controls. Interestingly, when the rho-zero primary tumors were analyzed, the increase in tumor-initiating potential correlated with a re-establishment of mtDNA, functional respiratory enzyme activity and oxidative phosphorylation activity. This was attributed to acquiring of mtDNA from the host cell to the initially mtDNA-depleted rho-zero tumors. Thus, the mitochondrial metabolic capacity attained by re-acquiring mtDNA was essential for tumorigenesis in vivo.

Moreover, a critical role for mitochondrial function was demonstrated in a subpopulation of pancreatic cancers cells with cancer stem cell characteristics. Characteristics of this subpopulation included survival of oncogene-ablation of a key driver mutation KRAS and drivers of tumour relapse. Analysis of the transcriptome and metabolome of this cancer stem-like subpopulation revealed an upregulation of genes governing ETC and oxidative phosphorylation, and an increased

47 reliance on mitochondrial respiration compared to glycolysis. Additionally, this subpopulation was highly sensitive to inhibitors of oxidative phosphorylation. Xenograft regressed tumors following

KRAS activation were treated with oligomycin, a complex V inhibitor, resulting in decreased tumor recurrence and increased survival of tumor-bearing mice in vivo (Viale et al., 2014). Thus, targeting mitochondrial bioenergetics through inhibition of mitochondrial DNA replication represents an attractive strategy to target a subpopulation of cancer cells and cancer stem cells.

1.5.2. Targeting mitochondria in AML

A subset of AML cells have unique mitochondrial characteristics such as increased mitochondrial mass and mitochondrial DNA content compared to normal hematopoietic progenitors (Lagadinou et al., 2013; Skrtic et al., 2011; Sriskanthadevan et al., 2015b). A higher reliance on oxidative metabolism and lower spare reserve capacity (Sriskanthadevan et al., 2015a) are likely key factors for increased sensitivity to agents which target mitochondrial function. mtDNA damaging agents such as bleomycin increased mtDNA mutations, impaired oxidative metabolism and caused anti- leukemic effects (Yeung et al., 2015). Inhibitors of mitochondrial transcription(Bralha et al., 2015) and translation (Skrtic et al., 2011) through chemical and genetic approaches preferentially target bulk AML and leukemic stem cells. Targeting the mitochondrial RNA polymerase POLRMT through either shRNA knockdown and chemical inhibition with the anti-viral nucleoside analog

2-C-methyladenosine decreased transcription of mtDNA encoded subunits of electron transport complexes, inhibited oxidative phosphorylation and preferentially induced cell death in leukemic cell lines and xenograft models (Bralha et al., 2015).

Similarly, genetic knockdown of mitochondrial translation elongation factor Ef-Tu decreased leukemic cell viability and proliferation. Inhibition of mitochondrial translation with the anti-

48 microbial agent Tigecycline depleted mtDNA-encoded proteins required for oxidative metabolism, decreased mitochondrial respiration complex activity and inhibited oxidative phosphorylation. Functional inhibition of mitochondrial translation selectively targeted bulk AML and LSCs in cell culture models, targeted primary AML cells preferentially in vitro and displayed efficacy in xenograft and patient-derived primary and secondary engraftment models of human

AML (Skrtic et al., 2011).

Additional approaches to deregulating mitochondrial function as a therapeutic strategy include targeting the mitochondrial protease ClpP, whose substrates include subunits of ETC Complex II.

Inactivation of ClpP resulted in deregulation of degradation of Complex II subunits, a substrate of

ClpP protease activity. This lead to inhibition of Complex II activity and oxidative metabolism, leading to anti-leukemic effects in bulk and LSC populations in several models of human AML

(Cole et al., 2015).

The dependence of LSC’s on mitochondrial function has also been interrogated. In a study by

Lagadinou et. al., (2013), the bioenergetic properties of LSC’s in primary AML cells were functionally defined using a xenograft assay as low-cellular redox status (ROS-low). The ROS- low subset was characterized as a chemotherapy-resistant, quiescent subpopulation with low mitochondrial energy metabolism and decreased basal oxygen consumption rate compared to

ROS-high cells. However, the LSC’s were preferentially dependent on oxidative phosphorylation with an impaired ability to adapt by glycolysis when mitochondrial metabolism was inhibited.

Inhibition of oxidative phosphorylation with the Bcl2 inhibitor ABT -263 selectively targeted the

LSC fraction, suggesting that targeting oxidative phosphorylation can be amenable as a therapeutic approach for AML. Overall, these studies suggest that bulk and LSC subpopulations have a preferential dependence on mitochondrial function that may be targeted as therapeutic approach

49 strategy for AML. Currently, mitochondria-targeted therapeutic strategies have remained primarily at the preclinical stage (excluding IDH inhibitors). Translating these studies to human clinical trials should focus on identifying predictive and diagnostic biomarkers of mitochondrial dependence in AML cells to promote personalized treatment approaches.

1.6 References

1. Agaronyan, K., Morozov, Y.I., Anikin, M., and Temiakov, D. (2015). Replication- transcription switch in human mitochondria. Science 347, 548-551. 2. Anderson, P.L., Kakuda, T.N., and Lichtenstein, K.A. (2004). The cellular pharmacology of nucleoside- and nucleotide-analogue reverse-transcriptase inhibitors and its relationship to clinical toxicities. Clin Infect Dis 38, 743-753. 3. Arnold, J.J., Smidansky, E.D., Moustafa, I.M., and Cameron, C.E. (2012). Human mitochondrial RNA polymerase: structure-function, mechanism and inhibition. Biochim Biophys Acta 1819, 948-960. 4. Arselin, G., Vaillier, J., Salin, B., Schaeffer, J., Giraud, M.F., Dautant, A., Brethes, D., and Velours, J. (2004). The modulation in subunits e and g amounts of yeast ATP synthase modifies mitochondrial cristae morphology. J Biol Chem 279, 40392-40399. 5. Bailey, C.M., and Anderson, K.S. (2010). A mechanistic view of human mitochondrial DNA polymerase gamma: providing insight into drug toxicity and mitochondrial disease. Biochimica et biophysica acta 1804, 1213-1222. 6. Baranowska-Bosiacka, I., Machalinski, B., and Tarasiuk, J. (2005). The purine nucleotide content in human leukemia cell lines. Cellular & molecular biology letters 10, 217-226. 7. Basak, N.P., and Banerjee, S. (2015). Mitochondrial dependency in progression of acute myeloid leukemia. Mitochondrion 21, 41-48. 8. Bestwick, R.K., Moffett, G.L., and Mathews, C.K. (1982). Selective expansion of mitochondrial nucleoside triphosphate pools in antimetabolite-treated HeLa cells. J Biol Chem 257, 9300-9304. 9. Bienstock, R.J., and Copeland, W.C. (2004). Molecular insights into NRTI inhibition and mitochondrial toxicity revealed from a structural model of the human mitochondrial DNA polymerase. Mitochondrion 4, 203-213. 10. Bogenhagen, D.F. (2012). Mitochondrial DNA nucleoid structure. Biochimica et Biophysica Acta (BBA)-Gene Regulatory Mechanisms 1819, 914-920. 11. Bonawitz, N.D., Clayton, D.A., and Shadel, G.S. (2006). Initiation and beyond: multiple functions of the human mitochondrial transcription machinery. Mol Cell 24, 813-825. 12. Bonnet, D., and Dick, J.E. (1997). Human acute myeloid leukemia is organized as a hierarchy that originates from a primitive hematopoietic cell. Nature medicine 3, 730-737. 13. 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. 14. Bridges, E.G., Jiang Z Fau - Cheng, Y.C., and Cheng, Y.C. (1999). Characterization of a dCTP transport activity reconstituted from human mitochondria. J Biol Chem. 15. Bullerwell, C.E., and Gray, M.W. (2004). Evolution of the mitochondrial genome: protist connections to animals, fungi and plants. Current opinion in microbiology 7, 528-534. 16. Carling, P.J., Cree, L.M., and Chinnery, P.F. (2011). The implications of mitochondrial DNA copy number regulation during embryogenesis. Mitochondrion 11, 686-692. 17. Cermakian, N., Ikeda, T.M., Miramontes, P., Lang, B.F., Gray, M.W., and Cedergren, R. (1997). On the evolution of the single-subunit RNA polymerases. Journal of molecular evolution 45, 671-681.

18. Chan, S.S., and Copeland, W.C. (2009). DNA polymerase gamma and mitochondrial disease: understanding the consequence of POLG mutations. Biochim Biophys Acta 1787, 312-319. 19. 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. 20. Chen, C.H., and Cheng, Y.C. (1992). The role of cytoplasmic deoxycytidine kinase in the mitochondrial effects of the anti-human immunodeficiency virus compound, 2',3'- dideoxycytidine. J Biol Chem 267, 2856-2859. 21. Chiron, S., Suleau, A., and Bonnefoy, N. (2005). Mitochondrial translation: elongation factor tu is essential in fission yeast and depends on an exchange factor conserved in humans but not in budding yeast. Genetics 169, 1891-1901. 22. Choi, J.S., and Berdis, A.J. (2012). Nucleoside transporters: biological insights and therapeutic applications. Future Med Chem 4, 1461-1478. 23. Cloutier, P., Al-Khoury, R., Lavallee-Adam, M., Faubert, D., Jiang, H., Poitras, C., Bouchard, A., Forget, D., Blanchette, M., and Coulombe, B. (2009). High-resolution mapping of the protein interaction network for the human transcription machinery and affinity purification of RNA polymerase II-associated complexes. Methods 48, 381-386. 24. Cloutier, P., and Coulombe, B. (2010). New insights into the biogenesis of nuclear RNA polymerases? Biochem Cell Biol 88, 211-221. 25. 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. 26. Corey, S.J., Minden, M.D., Barber, D.L., Kantarjian, H., Wang, J.C., and Schimmer, A.D. (2007). Myelodysplastic syndromes: the complexity of stem-cell diseases. Nat Rev Cancer 7, 118-129. 27. Darbyshire, J.H. (1996). Delta: a randomised double-blind controlled trial comparing combinations of zidovudine plus didanosine or zalcitabine with zidovudine alone in HIV- infected individuals. The Lancet 348, 283-291. 28. De Silva, D., Tu, Y.T., Amunts, A., Fontanesi, F., and Barrientos, A. (2015). Mitochondrial ribosome assembly in health and disease. Cell Cycle 14, 2226-2250. 29. Devineni, D., and Gallo, J.M. (1995). Zalcitabine. Clinical pharmacokinetics and efficacy. Clinical pharmacokinetics 28, 351-360. 30. 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. 31. Dick, J.E., and Lapidot, T. (2005). Biology of normal and acute myeloid leukemia stem cells. Int J Hematol 82, 389-396. 32. Dohner, H., Weisdorf, D.J., and Bloomfield, C.D. (2015). Acute Myeloid Leukemia. N Engl J Med 373, 1136-1152. 33. Dombret, H., and Gardin, C. (2016). An update of current treatments for adult acute myeloid leukemia. Blood 127, 53-61. 34. Endo, Y., Obata, T., Murata, D., Ito, M., Sakamoto, K., Fukushima, M., Yamasaki, Y., Yamada, Y., Natsume, N., and Sasaki, T. (2007). Cellular localization and functional

characterization of the equilibrative nucleoside transporters of antitumor nucleosides. Cancer Sci 98, 1633-1637. 35. Eppert, K., Takenaka, K., Lechman, E.R., Waldron, L., Nilsson, B., van Galen, P., Metzeler, K.H., Poeppl, A., Ling, V., Beyene, J., et al. (2011). Stem cell gene expression programs influence clinical outcome in human leukemia. Nat Med 17, 1086-1093. 36. Fan, J., Kamphorst, J.J., Mathew, R., Chung, M.K., White, E., Shlomi, T., and Rabinowitz, J.D. (2013). Glutamine-driven oxidative phosphorylation is a major ATP source in transformed mammalian cells in both normoxia and hypoxia. Mol Syst Biol 9, 712. 37. Favre, C., Zhdanov, A., Leahy, M., Papkovsky, D., and O'Connor, R. (2010). Mitochondrial pyrimidine nucleotide carrier (PNC1) regulates mitochondrial biogenesis and the invasive phenotype of cancer cells. Oncogene 29, 3964-3976. 38. Farruggia, P., Di Cataldo, A., Pinto, R.M., Palmisani, E., Macaluso, A., Valvo, L.L., Cantarini, M.E., Tornesello, A., Corti, P., Fioredda, F., et al. (2016). Pearson Syndrome: A Retrospective Cohort Study from the Marrow Failure Study Group of A.I.E.O.P. (Associazione Italiana Emato-Oncologia Pediatrica). JIMD reports 26, 37-43. 39. Ferraro, P., Nicolosi, L., Bernardi, P., Reichard, P., and Bianchi, V. (2006). Mitochondrial deoxynucleotide pool sizes in mouse liver and evidence for a transport mechanism for thymidine monophosphate. Proc Natl Acad Sci U S A 103, 18586-18591. 40. Ferraro, P., Pontarin, G., Crocco, L., Fabris, S., Reichard, P., and Bianchi, V. (2005). Mitochondrial deoxynucleotide pools in quiescent fibroblasts: a possible model for mitochondrial neurogastrointestinal encephalomyopathy (MNGIE). J Biol Chem 280, 24472-24480. 41. Figueroa, M.E., Abdel-Wahab, O., Lu, C., Ward, P.S., Patel, J., Shih, A., Li, Y., Bhagwat, N., Vasanthakumar, A., Fernandez, H.F., et al. (2010). Leukemic IDH1 and IDH2 mutations result in a hypermethylation phenotype, disrupt TET2 function, and impair hematopoietic differentiation. Cancer Cell 18, 553-567. 42. Foury, F. (1989). Cloning and sequencing of the nuclear gene MIP1 encoding the catalytic subunit of the yeast mitochondrial DNA polymerase. J Biol Chem 264, 20552-20560. 43. Gandhi, V.V., and Samuels, D.C. (2011a). Enzyme kinetics of the mitochondrial deoxyribonucleoside salvage pathway are not sufficient to support rapid mtDNA replication. PLoS Comput Biol 7, e1002078. 44. Gandhi, V.V., and Samuels, D.C. (2011b). 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. 45. Gaziev, A.I., Abdullaev, S., and Podlutsky, A. (2014). Mitochondrial function and mitochondrial DNA maintenance with advancing age. Biogerontology 15, 417-438. 46. Gerhold, J.M., Cansiz-Arda, S., Lohmus, M., Engberg, O., Reyes, A., van Rennes, H., Sanz, A., Holt, I.J., Cooper, H.M., and Spelbrink, J.N. (2015). Human Mitochondrial DNA- Protein Complexes Attach to a Cholesterol-Rich Membrane Structure. Sci Rep 5, 15292. 47. Govindarajan, R., Leung, G.P., Zhou, M., Tse, C.M., Wang, J., and Unadkat, J.D. (2009). Facilitated mitochondrial import of antiviral and anticancer nucleoside drugs by human equilibrative nucleoside transporter-3. American journal of physiology Gastrointestinal and liver physiology 296, G910-922. 48. Graziewicz, M.A., Longley, M.J., and Copeland, W.C. (2006). DNA polymerase gamma in mitochondrial DNA replication and repair. Chemical reviews 106, 383-405.

49. Gruning, N.M., Rinnerthaler, M., Bluemlein, K., Mulleder, M., Wamelink, M.M., Lehrach, H., Jakobs, C., Breitenbach, M., and Ralser, M. (2011). Pyruvate kinase triggers a metabolic feedback loop that controls redox metabolism in respiring cells. Cell Metab 14, 415-427. 50. Hance, N., Ekstrand, M.I., and Trifunovic, A. (2005). Mitochondrial DNA polymerase gamma is essential for mammalian embryogenesis. Hum Mol Genet 14, 1775-1783. 51. Hardie, D.G., Carling, D., and Gamblin, S.J. (2011). AMP-activated protein kinase: also regulated by ADP? Trends in biochemical sciences 36, 470-477. 52. Hensen, F., Cansiz, S., Gerhold, J.M., and Spelbrink, J.N. (2014). To be or not to be a nucleoid protein: a comparison of mass-spectrometry based approaches in the identification of potential mtDNA-nucleoid associated proteins. Biochimie 100, 219-226. 53. Huang da, W., Sherman, B.T., and Lempicki, R.A. (2009a). Bioinformatics enrichment tools: paths toward the comprehensive functional analysis of large gene lists. Nucleic Acids Res 37, 1-13. 54. Huang da, W., Sherman, B.T., and Lempicki, R.A. (2009b). Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nature protocols 4, 44- 57. 55. Ikeda, M., Ide, T., Fujino, T., Arai, S., Saku, K., Kakino, T., Tyynismaa, H., Yamasaki, T., Yamada, K., Kang, D., et al. (2015). Overexpression of TFAM or twinkle increases mtDNA copy number and facilitates cardioprotection associated with limited mitochondrial oxidative stress. PLoS One 10, e0119687. 56. Jager, S., Handschin, C., St-Pierre, J., and Spiegelman, B.M. (2007). AMP-activated protein kinase (AMPK) action in skeletal muscle via direct phosphorylation of PGC- 1alpha. Proc Natl Acad Sci U S A 104, 12017-12022. 57. Jeninga, E.H., Schoonjans, K., and Auwerx, J. (2010). Reversible acetylation of PGC-1: connecting energy sensors and effectors to guarantee metabolic flexibility. Oncogene 29, 4617-4624. 58. Ju, Y.S., Alexandrov, L.B., Gerstung, M., Martincorena, I., Nik-Zainal, S., Ramakrishna, M., Davies, H.R., Papaemmanuil, E., Gundem, G., Shlien, A., et al. (2014). Origins and functional consequences of somatic mitochondrial DNA mutations in human cancer. Elife 3. 59. Kallianpur, A.R., and Hulgan, T. (2009). Pharmacogenetics of nucleoside reverse- transcriptase inhibitor-associated peripheral neuropathy. Pharmacogenomics 10, 623-637. 60. Kelly, R.D., Mahmud, A., McKenzie, M., Trounce, I.A., and St John, J.C. (2012). Mitochondrial DNA copy number is regulated in a tissue specific manner by DNA methylation of the nuclear-encoded DNA polymerase gamma A. Nucleic Acids Res 40, 10124-10138. 61. Khwaja, A., Bjorkholm, M., Gale, R.E., Levine, R.L., Jordan, C.T., Ehninger, G., Bloomfield, C.D., Estey, E., Burnett, A., Cornelissen, J.J., et al. (2016). Acute myeloid leukaemia. Nature Reviews Disease Primers 2, 16010. 62. Kim, J., Lee, J.H., and Iyer, V.R. (2008). Global identification of Myc target genes reveals its direct role in mitochondrial biogenesis and its E-box usage in vivo. PLoS One 3, e1798. 63. Komanduri, K.V., and Levine, R.L. (2016). Diagnosis and Therapy of Acute Myeloid Leukemia in the Era of Molecular Risk Stratification. Annu Rev Med 67, 59-72. 64. 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. 65. Lai, Y., Tse, C.M., and Unadkat, J.D. (2004). Mitochondrial expression of the human equilibrative nucleoside transporter 1 (hENT1) results in enhanced mitochondrial toxicity of antiviral drugs. J Biol Chem 279, 4490-4497. 66. 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. 67. Lee, Y.S., Kennedy, W.D., and Yin, Y.W. (2009). Structural insight into processive human mitochondrial DNA synthesis and disease-related polymerase mutations. Cell 139, 312- 324. 68. Lewis, W., Day, B.J., and Copeland, W.C. (2003). Mitochondrial toxicity of NRTI antiviral drugs: an integrated cellular perspective. Nat Rev Drug Discov 2, 812-822. 69. Löwenberg, B., and Rowe, J.M. (2016). Introduction to the review series on advances in acute myeloid leukemia (AML). Blood 127, 1. 70. Margulis, L. (1970). Origin of Eukaryotic cells : evidence and research implications for a theory of the origin and evolution of microbial, plant, and animal cells on the precambrian earth (New Haven [etc.]: [s.n.]). 71. Martin, J.L., Brown, C.E., Matthews-Davis, N., and Reardon, J.E. (1994). Effects of antiviral nucleoside analogs on human DNA polymerases and mitochondrial DNA synthesis. Antimicrobial agents and chemotherapy 38, 2743-2749. 72. Mathews, C.K. (2015). Deoxyribonucleotide metabolism, mutagenesis and cancer. Nat Rev Cancer 15, 528-539. 73. Mishra, P., and Chan, D.C. (2014). Mitochondrial dynamics and inheritance during cell division, development and disease. Nat Rev Mol Cell Biol 15, 634-646. 74. Müller‐Höcker, J., Horvath, R., and Schäfer, S. (2011). Mitochondrial DNA depletion and fatal infantile hepatic failure due to mutations in the mitochondrial polymerase γ (POLG) gene: a combined morphological/ …. Journal of cellular …. 75. Norddahl, G.L., Pronk, C.J., Wahlestedt, M., Sten, G., Nygren, J.M., Ugale, A., Sigvardsson, M., and Bryder, D. (2011). Accumulating mitochondrial DNA mutations drive premature hematopoietic aging phenotypes distinct from physiological stem cell aging. Cell Stem Cell 8, 499-510. 76. Oka, T., Hikoso, S., Yamaguchi, O., Taneike, M., Takeda, T., Tamai, T., Oyabu, J., Murakawa, T., Nakayama, H., Nishida, K., et al. (2012). Mitochondrial DNA that escapes from autophagy causes inflammation and heart failure. Nature 485, 251-255. 77. Pontarin, G., Ferraro, P., Valentino, M.L., Hirano, M., Reichard, P., and Bianchi, V. (2006). Mitochondrial DNA depletion and thymidine phosphate pool dynamics in a cellular model of mitochondrial neurogastrointestinal encephalomyopathy. J Biol Chem 281, 22720-22728. 78. Rampazzo, C., Fabris, S., Franzolin, E., Crovatto, K., Frangini, M., and Bianchi, V. (2007). Mitochondrial thymidine kinase and the enzymatic network regulating thymidine triphosphate pools in cultured human cells. J Biol Chem 282, 34758-34769. 79. Rampazzo, C., Ferraro, P., Pontarin, G., Fabris, S., Reichard, P., and Bianchi, V. (2004). Mitochondrial deoxyribonucleotides, pool sizes, synthesis, and regulation. J Biol Chem 279, 17019-17026.

80. Rampazzo, C., Miazzi, C., Franzolin, E., Pontarin, G., Ferraro, P., Frangini, M., Reichard, P., and Bianchi, V. (2010). Regulation by degradation, a cellular defense against deoxyribonucleotide pool imbalances. Mutat Res 703, 2-10. 81. Reese, N.D., and Schiller, G.J. (2013). High-dose cytarabine (HD araC) in the treatment of leukemias: a review. Curr Hematol Malig Rep 8, 141-148. 82. Roboz, G.J. (2012). Current treatment of acute myeloid leukemia. Curr Opin Oncol 24, 711-719. 83. Roboz, G.J., Rosenblat, T., Arellano, M., Gobbi, M., Altman, J.K., Montesinos, P., O'Connell, C., Solomon, S.R., Pigneux, A., Vey, N., et al. (2014). International randomized phase III study of elacytarabine versus investigator choice in patients with relapsed/refractory acute myeloid leukemia. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 32, 1919-1926 84. Rowe, J.M. (2010). The evolving paradigm of prognostic factors in AML: Introduction to the Acute Leukemia Forum 2010. Best Pract Res Clin Haematol 23, 453-456. 85. Ruzzenente, B., Metodiev, M.D., Wredenberg, A., Bratic, A., Park, C.B., Camara, Y., Milenkovic, D., Zickermann, V., Wibom, R., Hultenby, K., et al. (2012). LRPPRC is necessary for polyadenylation and coordination of translation of mitochondrial mRNAs. EMBO J 31, 443-456. 86. Sarry, J.E., Murphy, K., Perry, R., Sanchez, P.V., Secreto, A., Keefer, C., Swider, C.R., Strzelecki, A.C., Cavelier, C., Recher, C., et al. (2011). Human acute myelogenous leukemia stem cells are rare and heterogeneous when assayed in NOD/SCID/IL2Rgammac-deficient mice. The Journal of clinical investigation 121, 384- 395. 87. Scarpulla, R.C., Vega, R.B., and Kelly, D.P. (2012). Transcriptional integration of mitochondrial biogenesis. Trends in endocrinology and metabolism: TEM 23, 459-466. 88. Setzer, B., Schlesier, M., Thomas, A.K., and Walker, U.A. (2005). Mitochondrial toxicity of nucleoside analogues in primary human lymphocytes. Antiviral therapy 10, 327-334. 89. Shlush, L.I., Zandi, S., Mitchell, A., Chen, W.C., Brandwein, J.M., Gupta, V., Kennedy, J.A., Schimmer, A.D., Schuh, A.C., Yee, K.W., et al. (2014). Identification of pre- leukaemic haematopoietic stem cells in acute leukaemia. Nature 506, 328-333. 90. Skrtic, 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. 91. Smits, P., Smeitink, J., and van den Heuvel, L. (2010). Mitochondrial translation and beyond: processes implicated in combined oxidative phosphorylation deficiencies. J Biomed Biotechnol 2010, 737385. 92. Sologub, M., Litonin, D., Anikin, M., Mustaev, A., and Temiakov, D. (2009). TFB2 is a transient component of the catalytic site of the human mitochondrial RNA polymerase. Cell 139, 934-944. 93. Spelbrink, J.N. (2010). Functional organization of mammalian mitochondrial DNA in nucleoids: history, recent developments, and future challenges. IUBMB Life 62, 19-32. 94. Sriskanthadevan, S., Jeyaraju, D.V., Chung, T.E., Prabha, S., Xu, W., Skrtic, M., Jhas, B., Hurren, R., Gronda, M., Wang, X., et al. (2015a). AML cells have low spare reserve capacity in their respiratory chain that renders them susceptible to oxidative metabolic stress. Blood 125, 2120-2130.

95. Sriskanthadevan, S., Jeyaraju, D.V., Chung, T.E., Prabha, S., Xu, W., Skrtic, M., Jhas, B., Hurren, R., Gronda, M., Wang, X., et al. (2015b). AML cells have low spare reserve capacity in their respiratory chain that renders them susceptible to oxidative metabolic stress. Blood 125, 2120-2130. 96. St John, J.C. (2016). Mitochondrial DNA copy number and replication in reprogramming and differentiation. Semin Cell Dev Biol. 97. Staub, M., and Eriksson, S. (2007). The Role of Deoxycytidine Kinase in DNA Synthesis and Nucleoside Analog Activation. In Deoxynucleoside Analogs In Cancer Therapy, G.J. Peters, ed. (Totowa, NJ: Humana Press), pp. 29-52. 98. Stehling, O., and Lill, R. (2013). The role of mitochondria in cellular iron-sulfur protein biogenesis: mechanisms, connected processes, and diseases. Cold Spring Harbor perspectives in medicine 3, 1-17. 99. Stein, E.M., and Tallman, M.S. (2016). Emerging therapeutic drugs for AML. Blood 127, 71-78. 100. 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. 101. Sugiura, Y., Fujioka, S., and Yoshida, S. (1986). Biosynthesis of pyrimidine nucleotides in human leukemic cells. Japanese journal of cancer research : Gann 77, 664-673. 102. Tang, S., Wang, J., Lee, N.C., Milone, M., Halberg, M.C., Schmitt, E.S., Craigen, W.J., Zhang, W., and Wong, L.J. (2011). Mitochondrial DNA polymerase gamma mutations: an ever expanding molecular and clinical spectrum. J Med Genet 48, 669-681. 103. Tozzi, V. (2010). Pharmacogenetics of antiretrovirals. Antiviral Res 85, 190-200. 104. Twig, G., and Shirihai, O.S. (2011). The interplay between mitochondrial dynamics and mitophagy. Antioxid Redox Signal 14, 1939-1951. 105. Viale, A., Pettazzoni, P., Lyssiotis, C.A., Ying, H., Sanchez, N., Marchesini, M., Carugo, A., Green, T., Seth, S., Giuliani, V., et al. (2014). Oncogene ablation-resistant pancreatic cancer cells depend on mitochondrial function. Nature 514, 628-632. 106. Wallace, D.C. (2012). Mitochondria and cancer. Nat Rev Cancer 12, 685-698. 107. Wallace, D.C., and Fan, W. (2009). The pathophysiology of mitochondrial disease as modeled in the mouse. Genes Dev 23, 1714-1736. 108. Wang, L. (2010). Deoxynucleoside salvage enzymes and tissue specific mitochondrial DNA depletion. Nucleosides Nucleotides Nucleic Acids 29, 370-381. 109. Weinberg, S.E., and Chandel, N.S. (2015). Targeting mitochondria metabolism for cancer therapy. Nat Chem Biol 11, 9-15. 110. West, A.P., Khoury-Hanold, W., Staron, M., Tal, M.C., Pineda, C.M., Lang, S.M., Bestwick, M., Duguay, B.A., Raimundo, N., MacDuff, D.A., et al. (2015). Mitochondrial DNA stress primes the antiviral innate immune response. Nature 520, 553-557. 111. Yakubovskaya, E., Chen, Z., Carrodeguas, J.A., Kisker, C., and Bogenhagen, D.F. (2006). Functional human mitochondrial DNA polymerase gamma forms a heterotrimer. J Biol Chem 281, 374-382. 112. Yao, S.Y.M., Ng, A.M.L., Sundaram, M., Cass, C.E., Baldwin, S.A., and Young, J.D. (2001). Transport of antiviral 3′-deoxy-nucleoside drugs by recombinant human and rat equilibrative, nitrobenzylthioinosine (NBMPR)-insensitive (ENT2) nucleoside transporter proteins produced in Xenopus oocytes. Molecular Membrane Biology 18, 161-167.

113. Yeung, M., Hurren, R., Nemr, C., Wang, X., Hershenfeld, S., Gronda, M., Liyanage, S., Wu, Y., Augustine, J., Lee, E.A., et al. (2015). Mitochondrial DNA damage by bleomycin induces AML cell death. Apoptosis 20, 811-820. 114. Yu, V.W.C., and Scadden, D.T. (2016). Hematopoietic Stem Cell and Its Bone Marrow Niche. 115. Zick, M., Rabl, R., and Reichert, A.S. (2009). Cristae formation-linking ultrastructure and function of mitochondria. Biochim Biophys Acta 1793, 5-19.

2. MATERIALS AND METHODS

2.1 Bioinformatic analysis

Affymetrix gene expression data of AML (542 samples) and healthy bone marrow samples

(73 samples) from the Haferlach dataset was downloaded from the leukemia gene atlas portal

(http://www.leukemia-gene-atlas.de) on March 2016 (Haferlach et al., 2010). The platform used is Affymetrix Human Genome U133 Plus 2.0 Array and the data values correspond to Robust

Multichip Average (RMA) expression measure. Official gene symbols from the Hugo Gene

Nomenclature Committee were retrieved from the Affymetrix probe identifiers using the R package ‘biomaRt’ (biomaRt_2.26.1 and R version 3.2.3) and searched against Ensembl Genes version 84. Data was reduced at the gene level by selecting the probe with the highest median absolute deviation (MAD) across samples per gene. In one case (POLG), two of three probes

(203366_at, 217635_at) displayed a similar POLG expression pattern, while the highest MAD score probe (217636_at) did not and was excluded from the analysis. Instead, 203366_at was used for analysis. In order to study the gene expression pattern within the AML samples, and between

AML and normal patients, data were centered, scaled (z-score) and clustered using the heatmap.2 function available from the gplots R package (gplots_2.17.0). Using the default parameters, each row (gene) in the result has mean 0 and sample standard deviation 1. The z score is a normalized value, indicating how many standard deviations away the gene expression is compared to the mean expression of all samples for the same gene. z = (x – μ) / σ, where x is gene expression, μ is mean gene expression across samples, and σ is standard deviation of the population. See supplemental information for additional methods.

Hierarchical clustering of AML mtDNA biosynthesis gene expression was done with the

Euclidean distance metric and the complete linkage method on the scaled data on the AML samples. The clusters are defined using the dynamic branch cutting method, which is implemented using the R package dynamicTreeCut with parameter k set to 4 to retrieve the four main clusters.

The z-score scaled data generated using heatmap.2 were retrieved for each normal samples and the four AML cluster groups and their distribution visualized using boxplots for the mtDNA biosynthesis genes. Distribution of these groups were compared by looking at both the median values and two-sided t-tests with the null hypothesis being that there is no difference in the z-score distribution between the groups.

Furthermore, for some genes of interest (CMPK1, NME1-NME1, DCK, TK1, SLC25A33,

SLC25A36, SLC29A1, SLC29A3), boxplot distributions of z-score values were generated for all

AML and normal samples and distribution differences were assessed by comparing both median values and two-sided t-tests. Heatmaps of z-score data were generated by ranking these genes using the same sample order (column) as for the mtDNA biosynthesis genes. In this order, the correlation with the median value of the z-scores of the mtDNA biosynthesis genes were assessed using Pearson correlation with the R function cor.test() which also calculates a paired t-test of r values across samples.

2.2 Primary AML and normal hematopoietic cells

Primary mononuclear cells were isolated from peripheral blood of consenting patients diagnosed with AML, who had at least 80% malignant cells among low-density cells isolated by

Ficoll density gradient centrifugation. Normal peripheral blood stem cells were obtained from healthy consenting volunteers donating peripheral blood for allogeneic stem cell transplantation

60 after granulocyte-colony stimulating factor (G-CSF) mobilization. Normal CD34+ cells were isolated from PBSC’s by immunomagnetic positive isolation, EasySep™ Human CD34 Positive

Selection Kit 18056 (Stem cell technologies, Vancouver, Canada). Primary cells were cultured at

37°C in Iscove modiﬁed Dulbecco medium and supplemented with 20% fetal bovine serum

(FBS), 2 mM L-glutamine, 2 ng/mL Interleukin-3 (IL-3) cytokines, and 20 ng/mL stem cell factor (SCF). The collection and use of human tissue was approved by the University Health

Network institutional review.

2.3 Cell lines

All cell lines were maintained at 37°C and 5% CO2. OCI-AML2, K562 cells and HL-60 were cultured in Iscove’s modified Dulbecco’s medium (IMDM) supplemented with 10% fetal bovine serum (FBS) and appropriate antibiotics. TEX cells (Warner et al., 2005) were cultured in

IMDM supplemented with 20% FBS, 2 mM L-glutamine (Life Technologies, Carlsbad, CA), 2 ng/mL IL-3 (R&D Systems, Minneapolis, MN), 20 ng/mL SCF (Miltenyi Biotec, San Diego, CA), and appropriate antibiotics. HEK 293 Flp-In TRex cell lines were cultured in Dulbecco’s H21 modified eagle’s medium (DMEM) supplemented with 10% FBS and appropriate antibiotics.

2.4 Immunoblotting

DCK, CMPK1, NME2, POLG, TFAM, TK1 proteins were detected from whole cell lysates prepared with RIPA buffer. COX I, II, IV proteins were detected from whole cell lysates prepared in 1.5% (w/v) n-dodecyl-β-maltoside (n-DBM) (Sigma Aldrich, St. Louis, MO) supplemented with protease inhibitor cocktail (Roche, Mississauga, CA ). For detection of COX mitochondrial proteins, lysates were resolved by SDS-PAGE and transferred with (3-

[Cyclohexylamino]-1-propanesulfonic acid) CAPS buffer (pH 10.5) to a PVDF membrane.

Membranes were blocked with 5% milk in PBST for 1 hour at room temperature, following by incubation with primary antibody overnight at 4°C. HRP-conjugated secondary antibodies (GE

Healthcare, Buckinghamshire, UK) were incubated for 1 hour at room temperature. Proteins were detected by HRP chemiluminescence and bands were quantified by Image J software. Antibodies were purchased as indicated: DCK (Ab186128), NME2 (Ab60602), VDAC (Ab15895) - Abcam

(Cambridge, Massachusetts); TFAM (7495)-Cell Signaling (Whitby, Ontario); Actin (AC15, SC

69879), β-tubulin (H235, SC9104), mt-COX I (SC 58347), mt-COX II (SC 65239) - Santa Cruz

Biotechnology Inc. (Santa Cruz, California); CMPK1 (41523) - Signalway Antibodies (Maryland,

USA); POLG (TA310770) - Origene (Rockville, Maryland); and COX IV (A21347) - Molecular

Probes (Burlington, Canada).

2.5 shRNA knockdown

Genetic knockdown of DCK, CMPK1, TK1 and TFAM was achieved using lentiviral transduction of shRNAs inserted into the pLKO.1 vector. shRNA against green fluorescent protein inserted in the pLKO.1 vector was employed as a negative control. Assays were performed following a 3 day puromycin selection for transduced cells. Bacterial stocks containing pLKO.1 vectors with shRNA against target genes were purchased from Sigma Aldrich (Oakville, Ontario,

Canada). To assess mtDNA levels following shRNA knockdown, cells were collected at day 5 post transduction for TK1 and CMPK1 knockdown, and day 7 post transduction for DCK and

TFAM knockdown.

2.6 Quantitative real-time polymerase chain reaction

Equal amounts of genomic DNA or cDNA for each sample were added to a prepared master mix (SYBR Green PCR Master mix Part # 436759; Applied Biosystems, Foster City,

CA). Quantitative reverse-transcriptase polymerase chain reaction (qRT-PCR) reactions were performed on an ABI Prism 7900 sequence detection system (Applied Biosystems, Foster City,

CA) as described previously (Skrtic, 2010). The relative abundance was represented by the cycle threshold of amplification (CT), which is inversely proportional to the level of first strand cDNA being amplified. The comparative ΔCT method was used for calculation of relative abundance.

– ΔΔ The expression level of a gene relative to the baseline level was calculated as 2 CT, where ΔCT

= (average CT (treated sample) – average CT–(control sample)) and ΔΔ CT = (average CT (target gene) – average CT (reference gene)).

2.7 Mitochondrial DNA quantification

Genomic DNA was isolated using DNeasy Blood and Tissue Kit (Qiagen, MD, USA).

Relative human mtDNA levels were assessed by quantitative PCR using the mtDNA-encoded human NADH dehydrogenase 1 (ND1) gene normalized to the nuclear-encoded human beta- globulin gene (HGB). Primer sequences for mt-ND1-F: 5’-AACATACCCATGGCCAACCT-3’ and ND1-R: 5’-AGCGAAGGGTTGTAGTAGCCC-3’, and for HGB-F: 5’-

GAAGAGCCAAGGACAGGTAC-3’ and HGB-R: 5’-CAACTTCATCCACGTTC ACC-3’.

2.8 COX I and COX II mRNA quantification

RNA was isolated from cell pellets using the RNEasy Plus Mini kit (Qiagen, MD, USA).

First strand cDNA was synthesized with random primers using the SuperScript III Kit

(Invitrogen, Burlington, ON) as per manufacturer’s instructions. The cDNAs encoding human cytochrome C oxidase (COX I and COX II), ribosomal RNA subunit 18 (18S) were amplified using the following primer pairs: COXI-F: 5’-CTATACCTATTATTCGGCGCATGA-3’, COXI-

R: 5’-CAGCTCGGCTCGAATAAGGA-3’, COXII-F: 5’-CTGAACCTACGAGTACACCG-3’,

COXII-R: 5’-TTAATTCTAGGACGATGGGC-3’, 18s-F: 5’-

AGGAATTGACGGAAGGGCAC-3’, and 18s-R: 5’-GGACATCTAAGGGCATCACA-3’. The cDNAs encoding mouse COX I, COX II and 18S were amplified using the primers: m-COXI-F:

5’-GCCCCAGATATAGCATTCCC-3’, m-COXI-R: 5’-GTTCATCCTGTTCCTGCTCC-3’, m-

COXII-F: 5’-ACGAAATCAACAACCCCGTA-3’, m-COXII-R: 5’-

GGCAGAACGACTCGGTTATC-3’, m-18S-F: 5’-CTTAGAGGGACAAGTGGCG-3’, and m-

18S-R: 5’- ACGCTGAGCCAGTCAGTGTA-3’.

2.9 Intracellular quantification of ddC and ddCTP

All liquid chromatography/mass spectrometry (LC/MS) grade solvents and salts were purchased from Fisher (Ottawa, Ontario Canada): dichloromethane (DCM), water (H2O), acetonitrile (ACN), methanol (MeOH), and ammonium acetate. The authentic metabolite standard for ddC and ddCTP (2’3’-dideoxycytidine triphosphate) was purchased from Sigma-Aldrich Co.

(St. Loius, MO). 5x106 cultured cells were first washed with phosphate buffered saline, three times with cold ammonium formate buffer (pH 7.4) and pelleted. Cell pellets were extracted with 600

L of 31.6% MeOH/36.3% ACN in H2O (v/v). Cells were lysed and homogenized by bead-beating for 2 minutes at 30 Hz using a 5 mm metal bead (TissueLyser II – Qiagen). Cellular extracts were partitioned into aqueous and organic layers following DCM treatment and centrifugation. Aqueous supernatants were dried by vacuum centrifugation with sample temperature maintained at -4oC

(Labconco, Kansas City MO, USA). Pellets were subsequently resuspended in 50 μl of H2O as the injection buffer.

For targeted metabolite analysis and relative concentration determination of ddC, ddCTP and dCTP, samples were injected onto an Agilent 6430 Triple Quadrupole (QQQ)-LC-MS/MS

(Agilent Technologies, Santa Clara, CA, USA). Chromatography was achieved using a 1290

Infinity ultra-performance LC system (Agilent Technologies, Santa Clara, CA, USA) consisting of vacuum degasser, autosampler and a binary pump. Separation was performed on a Scherzo SM-

C18 column 3 μm, 3.0×150 mm (Imtakt Corp, JAPAN) maintained at 10°C. The chromatographic gradient started at 100% mobile phase A (5 mM ammonium acetate in water) with a 5 min gradient to 100% B (200 mM ammonium acetate in 20% ACN / 80% water) at a flow rate of 0.4 mL/min.

This was followed by a 5 min hold time at 100% mobile phase B and a subsequent re-equilibration time (6 min) before next injection. A sample volume of 5 or 10 L of sample were injected for analysis.

The mass spectrometer was equipped with an electrospray ionization (ESI) source and samples were analyzed in positive mode. Multiple reaction monitoring (MRM) transitions were optimized on standards for each metabolite quantitated. Transitions for quantifier and qualifier ions were respectively 452.0 → 112.1 and 452.0 → 83.1 for ddCTP, 212.1 → 112.1; 212.1 →

95.1; 212.1 → 55.2 and 212.1 → 57.2 for ddC. Gas temperature and flow were set at 350°C and

10 L/min respectively, nebulizer pressure was set at 40 psi and capillary voltage was set at 3500

V. Relative (semi-quantitative) concentrations were determined by integrating area under the curve for the quantifying MRM transition and compared to external calibration curves. For each assay, a 12 point calibration curve for both ddC and ddCTP was constructed. The calibration curve was constructed each time the assay was run.

2.10 RNA expression profiling

Total RNA was isolated from cells using the RNeasy Plus kit (Qiagen, MD, USA). RNA expression was analyzed for 13 mtDNA-encoded genes and 13 nuclear encoded genes which code for subunits of respiration complexes. Relative transcript abundance was quantified using a custom-built Nanostring nCounter chip (Nanostring Technologies, Seattle, WA) and analyzed as per the manufacturer’s instructions. Briefly, the raw data was normalized to positive controls, followed by normalization to 4 housekeeping genes and lastly, normalized to the vehicle control.

Normalized data was used for analysis.

2.11 Metabolic Analysis

Basal oxygen consumption rate (OCR) was measured using the XF96 Seahorse extracellular flux analysis (Seahorse Biosciences, Billerica, MA). XF96 well plates were coated with Cell Tak at 0.16 g/well according to the manufacturer’s instructions (BD Biosciences,

Mississauga, CA). On the day of the assay, cells were resuspended in unbuffered Alpha-MEM media supplemented with 2% FCS and plated at 1.9×105 cells/well. Basal OCR (pMol/min) was taken as the average of the first three readings measured 3 minutes apart.

2.12 Cell proliferation and viability assays

Cell proliferation assays were set up at a density of 50,000 cells/mL for cell lines and at

600,000 cells/mL for primary AML cells. ddC was purchased from Sigma Aldrich (St. Louis, MO), dissolved in DMSO and added fresh to culture media at a final DMSO concentration of 0.1% v/v each time cells were passaged (twice weekly). Viable cells were counted by trypan blue exclusion

66 staining. P143B and P1669 Rho (0) cells were cultured for 7 days in tissue culture flasks with ddC.

On day 7 of treatment, cells were passaged and reseeded into 96-well plates for 3 days with fresh drug and cell viability was assessed by Sulforhodamine B (SRB) colorimetric assay as previously described (Vichai and Kirtikara, 2006). Normal peripheral blood stem cells (PBSC’s) were seeded at a density of 600,000 cells/mL and viability of cells was assayed by Cyquant direct cell proliferation assay kit (Sigma Aldrich, Oakville, Ontario, CA) after ddC treatment. Normal

PBSC’s were seeded at a density of 600,000 cells/mL and viability of cells treated with ddC was assayed by Cyquant direct cell proliferation assay kit (Sigma Aldrich, Oakville, Ontario, CA) as per the manufacturer’s instructions.

2.13 Flow Cytometry

Flow cytometry analysis was performed on FACSCANTO II (BD BioSciences, Mississauga, ON,

Canada) or LSRFortessa X-20 (BD Biosciences) using FACS DIVA software. Data was analyzed post-acquisition with FloJO Software Version 7.7.1 (Treestar, Ashland, OR). Cell death was quantified by Annexin V-fluoroscein isothiocyanate (FITC+) and Propidium Iodide (PI)-staining

(Biovision Research Products, Mountain View, CA) according to the manufacturer’s instructions by flow cytometry. For analysis of CD34+ cell viability, normal PBSC’s were co-stained with

CD34 (8G12) FITC antibody, 348053 (BD Biosciences, Mississauga, CA) and PI

2.13.1 Mitochondrial Mass

Mitochondrial mass was assessed by Mitotracker Deep Red FM staining (Invitrogen,

Burlington, Canada). Cells were stained at 100 nM for 30 minutes, pelleted and resuspended in

Hanks Buffered Solution (HBS) with Annexin V-FITC stain to exclude non-viable cells from the

67 analysis. Mitochondrial mass was quantified as mean fluorescence intensity of the viable (Annexin

V-) cell population.

2.14 Electron microscopy

Transmission electron microscopy of mitochondria was performed of OCI-AML2 cells treated with increasing doses of ddC for 3, 6 or 10 days as previously described (Jhas et al., 2013).

Representative images are shown at 8000× magnification.

2.15 Xenograft models of human AML

For in vivo studies, ddC was purchased from Seqiuoa Research products (Pangbourne,

UK). OCI-AML2 human leukemia cells (1×106) were injected subcutaneously into the flanks of

SCID mice (Ontario Cancer Institute, Toronto, ON). After the appearance of a palpable tumor

(9-11 days), the mice were treated with ddC by i.p. once daily or vehicle control (n=7 per group) at a treatment schedule of 5 out of 7 days for a total of 11 days at 35 and 75 mg/kg or for 9 days at 150 and 300 mg/kg ddC. Tumor volumes were measured 3 times per week based on caliper measurements of tumor length and width (volume= tumor length × width2 × 0.5236). At the end of treatment, mice were sacrificed and tumor volumes and mass was measured from excised tumors. To evaluate ddC efficacy in a primary AML engraftment mouse model, a frozen aliquot of primary AML cells was thawed, counted and resuspended in PBS. 2-3 × 106 viable trypan blue- negative cells were injected into the right femur of 10 week-old female NOD-SCID mice that were previously irradiated, and injected with 200 µg anti-mouse CD122. Mice were treated once daily with ddC at 75 mg/kg or vehicle control (n=7 per group) for 3 weeks at a treatment schedule of 5 of 7 days. Following treatment, mice were sacrificed primary AML engraftment (CD45+ CD33+

CD19– cells) in left femur was quantified by flow cytometry. To assess secondary engraftment, primary human AML cells were isolated from the bone marrow of control and ddC treated mice.

Cells were pooled and equal numbers of viable cells were transplanted into the right femur of secondary untreated mice. After 5 weeks, mice were sacrificed and human CD45+ CD33+ CD19– cells were assessed by flow cytometry. All in vivo studies were carried out according to the regulations of the Canadian Council on Animal Care and with the approval of the Ontario

Cancer Institute Animal Ethics Review board.

SCID mice were treated with ddC (300 mg/kg once daily by i.p. injection) or vehicle control (3 mice per treatment group) for 11 days. Following treatment, mice were sacrificed and organs were collected and embedded with 10% buffered formalin, sectioned and stained with hematoxylin and eosin. The stained samples were scanned using Aperio Scanscope X and analyzed using Aperio ImageScope at 15× magnification.

2.16 BioID Cloning

PCR reactions were performed to insert full-length coding sequences for POLG and Ruvbl2 into pcDNA5 FRT/TO FlagBirA* expression vector to generate C-terminal FlagBirA* tagged

POLG, and C-terminal FlagBirA* tagged Ruvbl2 constructs. Briefly, PCR fragments were amplified with Kapa Hifi Hotstart Taq polymerase (Kapa Biosystems, Boston, USA) according to manufacturer’s instructions. Primers for designed for entry into destination plasmids as follows,

POLG forward: CGGGATCCA TGAGCCGCCTGCTCTGGAG, POLG reverse:

TAAGAATAGCGGCCGCTGGTCCAGGCTGGCTTCGT. Destination vector and PCR products were digested with BamHI and NotI for POLG construct. Digested products were gel purified using gel purification kit (Qiagen, Maryland, USA) and ligated into the destination plasmid using

T4 DNA ligase following manufacturer’s instructions. All restriction and ligation enzymes were purchased from New England Biolabs (Whitby, Ontario).

2.17 BioID interactome mapping in cell lines

Using the Flp-In system (Invitrogen), HEK 293 T-REx Flp-In cells stably expressing

POLG- FlagBirA*, Ruvbl2-FlagBirA*, and the mitochondrial matrix localizing negative control ornithine transcarbamylase (OTC)- FlagBirA*, as well as Flag-BirA*- HEK 293 T-REx Flp-In cells were generated as previously described (Huai et al., 2012; Lozano et al., 2013). Cells were incubated for 24 hours in 1 μg/mL tetracycline (Bioshop) and 50 μM biotin (Sigma) to induce transgene expression and biotin labelling. The cell pellet was resuspended in 10 mL of lysis buffer

(50 mM Tris-HCl pH 7.5, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, 0.1%

SDS, 1:500 protease inhibitor cocktail (Sigma-Aldrich), 1:1000 benzonase nuclease (Novagen) and incubated on an end-over-end rotator at 4°C for 1 hour, briefly sonicated to disrupt any visible aggregates, then centrifuged at 16,000  g for 30 min at 4°C. Supernatant was transferred to a fresh

15 mL conical tube. 30 μL of packed, pre-equilibrated Streptavidin sepharose beads (GE) was added and the mixture and incubated for 3 hours at 4°C with end-over-end rotation. Beads were pelleted by centrifugation at 2000 rpm for 2 min and transferred with 1 mL of lysis buffer to a fresh Eppendorf tube. Beads were washed once with 1 mL lysis buffer and twice with 1 mL of 50 mM ammonium bicarbonate (pH=8.3). Beads were transferred in ammonium bicarbonate to a fresh centrifuge tube, and washed two more times with 1 mL ammonium bicarbonate buffer.

Tryptic digestion was performed by incubating the beads with 1 μg MS-grade TPCK trypsin

(Promega, Madison, WI) dissolved in 200 μL of 50 mM ammonium bicarbonate (pH 8.3) overnight at 37°C. The following morning, 0.5 μg MS-grade TPCK trypsin was added and beads

70 were incubated 2 additional hours at 37°C. Beads were pelleted by centrifugation at 2000 g for 2 min, and the supernatant was transferred to a fresh Eppendorf tube. Beads were washed twice with

150 μL of 50 mM ammonium bicarbonate, and these washes were pooled with the first eluate. The sample was lyophilized, and resuspended in buffer A (0.1% formic acid). 1/5th of the sample was analyzed per MS run.

2.18 BioID mass spectrometry and data analysis

Analytical columns (75 m inner diameter) and pre-columns (150 m inner diameter) were made in-house from fused silica capillary tubing from InnovaQuartz (Phoenix, AZ) and packed with 100 Å C18–coated silica particles (Magic, Michrom Bioresources, Auburn, CA). Peptides were subjected to liquid chromatography (LC)-electrospray ionization-tandem mass spectrometry, using a 120 min reversed-phase (100% water–100% acetonitrile, 0.1% formic acid) buffer gradient running at 250 nL/min on a Proxeon EASY-nLC pump in-line with a hybrid LTQ-Orbitrap Velos mass spectrometer (Thermo Fisher Scientific, Waltham, MA). A parent ion scan was performed in the Orbitrap using a resolving power of 60,000, then up to the twenty most intense peaks were selected for MS/MS (minimum ion count of 1000 for activation), using standard collision induced dissociation fragmentation. Fragment ions were detected in the LTQ. Dynamic exclusion was activated such that MS/MS of the same m/z (within a range of 15 ppm; exclusion list size = 500) detected twice within 15 seconds were excluded from analysis for 30 seconds. For protein identification, Thermo .RAW files were converted to the .mzXML format using Proteowizard

(Kessner et al., 2008), then searched using X!Tandem (Craig and Beavis, 2004) against the human

(Human RefSeq Version 45) database. X!Tandem search parameters were: 15 ppm parent mass error; 0.4 Da fragment mass error; complete modifications, none; cysteine modifications, none;

71 potential modifications, +16@M and W, +32@M and W, +42@N-terminus, +1@N and Q. Each of the two biological replicates of POLG and Ruvbl2 samples was analyzed using two technical replicates. Data were analyzed using the trans-proteomic pipeline (TPP) (Deutsch et al., 2010;

Pedrioli, 2010) via the ProHits software suite (Liu et al., 2010). Proteins identified with a Protein

Prophet cut-off of 0.9 were analyzed with the SAINT express algorithm (v3.3) (Teo et al., 2014).

24 control runs (consisting of 20 Flag-BirA*only and 4 ornithine transcarbamylase (OTC-

FlagBirA*) were collapsed to the 4 highest spectral counts for each prey, and the SAINT score cut-off value was set to 0.70 (Govindarajan et al., 2009). Cell compartment localization of the hits was obtained from Uniprot database, functional classification was determined from Gene-ontology annotatation using the DAVID bioinformatics resource 6.7 and manual literature searches if unannotated (Huang da et al., 2009a, b) Interactome data was presented as functional classes using

Cytoscape software v. 3.0.1 (Smoot et al., 2011).

2.19 Preparation of purified- mitochondrial fractions for Ruvbl2 immunoblotting

Crude mitochondrial fractions were prepared as previously described by differential centrifugation. Mitochondrial fractions were further purified using a single step 30% Percoll gradient or a three step 50%,22% and 15% gradient (Jeyaraju et al., 2009).

2.20 Identification of an Alternative translation initiation isoform of Ruvbl2 by immunofluorescence

Human Ruvbl2 cDNA corresponding to isoform 1 (Accession: NP_006657.1) was tagged with C- terminal hemagglutinin (HA) and cloned into pcDNA5 (Invitrogen). The codon specifying the first

(M1) or second (M46) methionine was converted to alanine (ATG – GCG) by site directed

72 mutagenesis and the constructs (Ruvbl2M1A.HA, and Ruvbl2M46A.HA) transfected into HOS cells using lipofectamineTM (Thermofisher), and imaged by confocal microscopy as previously described(Kazak et al., 2013).

2.20 Statistical analysis

All statistical analyses were performed on Graph Pad Prism 6.03 (La Jolla, CA, USA).

2.3 References

1. Craig, R., and Beavis, R.C. (2004). TANDEM: matching proteins with tandem mass spectra. Bioinformatics 20, 1466-1467. 2. Deutsch, E.W., Mendoza, L., Shteynberg, D., Farrah, T., Lam, H., Tasman, N., Sun, Z., Nilsson, E., Pratt, B., Prazen, B., et al. (2010). A guided tour of the Trans-Proteomic Pipeline. Proteomics 10, 1150-1159. 3. Govindarajan, R., Leung, G.P., Zhou, M., Tse, C.M., Wang, J., and Unadkat, J.D. (2009). Facilitated mitochondrial import of antiviral and anticancer nucleoside drugs by human equilibrative nucleoside transporter-3. American journal of physiology Gastrointestinal and liver physiology 296, G910-922. 4. Haferlach, T., Kohlmann, A., Wieczorek, L., Basso, G., Kronnie, G.T., Bene, M.C., De Vos, J., Hernandez, J.M., Hofmann, W.K., Mills, K.I., et al. (2010). Clinical utility of microarray- based gene expression profiling in the diagnosis and subclassification of leukemia: report from the International Microarray Innovations in Leukemia Study Group. J Clin Oncol 28, 2529-2537. 5. Huai, L., Wang, C., Zhang, C., Li, Q., Chen, Y., Jia, Y., Li, Y., Xing, H., Tian, Z., Rao, Q., et al. (2012). Metformin induces differentiation in acute promyelocytic leukemia by activating the MEK/ERK signaling pathway. Biochem Biophys Res Commun 422, 398-404. 6. Huang da, W., Sherman, B.T., and Lempicki, R.A. (2009a). Bioinformatics enrichment tools: paths toward the comprehensive functional analysis of large gene lists. Nucleic Acids Res 37, 1-13. 7. Huang da, W., Sherman, B.T., and Lempicki, R.A. (2009b). Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nature protocols 4, 44-57. 8. Jeyaraju, D.V., Cisbani, G., De Brito, O.M., Koonin, E.V., and Pellegrini, L. (2009). Hax1 lacks BH modules and is peripherally associated to heavy membranes: implications for Omi/HtrA2 and PARL activity in the regulation of mitochondrial stress and apoptosis. Cell death and differentiation 16, 1622-1629. 9. Jhas, B., Sriskanthadevan, S., Skrtic, M., Sukhai, M.A., Voisin, V., Jitkova, Y., Gronda, M., Hurren, R., Laister, R.C., Bader, G.D., et al. (2013). Metabolic adaptation to chronic inhibition of mitochondrial protein synthesis in acute myeloid leukemia cells. PLoS One 8, e58367. 10. Kazak, L., Reyes, A., Duncan, A.L., Rorbach, J., Wood, S.R., Brea-Calvo, G., Gammage, P.A., Robinson, A.J., Minczuk, M., and Holt, I.J. (2013). Alternative translation initiation augments the human mitochondrial proteome. Nucleic Acids Res 41, 2354-2369. 11. Kessner, D., Chambers, M., Burke, R., Agus, D., and Mallick, P. (2008). ProteoWizard: open source software for rapid proteomics tools development. Bioinformatics 24, 2534-2536. 12. Liu, G., Zhang, J., Larsen, B., Stark, C., Breitkreutz, A., Lin, Z.Y., Breitkreutz, B.J., Ding, Y., Colwill, K., Pasculescu, A., et al. (2010). ProHits: integrated software for mass spectrometry- based interaction proteomics. Nat Biotechnol 28, 1015-1017. 13. Lozano, E., Herraez, E., Briz, O., Robledo, V.S., Hernandez-Iglesias, J., Gonzalez- Hernandez, A., and Marin, J.J. (2013). Role of the plasma membrane transporter of organic cations OCT1 and its genetic variants in modern liver pharmacology. Biomed Res Int 2013, 692071. 14. Pedrioli, P.G. (2010). Trans-proteomic pipeline: a pipeline for proteomic analysis. Methods Mol Biol 604, 213-238. 15. Smoot, M.E., Ono, K., Ruscheinski, J., Wang, P.L., and Ideker, T. (2011). Cytoscape 2.8: new features for data integration and network visualization. Bioinformatics (Oxford, England) 27, 431-432.

16. Teo, G., Liu, G., Zhang, J., Nesvizhskii, A.I., Gingras, A.C., and Choi, H. (2014). SAINTexpress: improvements and additional features in Significance Analysis of INTeractome software. J Proteomics 100, 37-43. 17. Vichai, V., and Kirtikara, K. (2006). Sulforhodamine B colorimetric assay for cytotoxicity screening. Nature protocols 1, 1112-1116. 18. 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.

3. Leveraging increased cytoplasmic nucleoside kinase activity to target mitochondrial DNA and oxidative phosphorylation in AML

Contributions

The majority of work presented in this chapter was performed by the author of the thesis. Technical assistance was contributed as follows: Veronique Voisin and ChangJiang Xu performed all bioinformatics analyses, under the supervision of Dr. Gary D. Bader (Donnelly Centre for Cellular and Biomolecular Research, Toronto, Ontario, Canada) Rose Hurren provided technical assistance for qRT-PCR and Nanostring experiments, under the supervision of Dr. Aaron Schimmer Mass spectrometric experiments were performed and analyzed with technical assistance by Gaëllle Bridon (Goodman Cancer Research Centre, McGill University, Montreal, Quebec, Canada). under the supervision of Dr. Daina Avizonis Thirushi Siriwardena (Department of Medical Biophysics, Faculty of Medicine, University of Toronto, Ontario, Canada) and Shrivani Sriskanthadevan (Princess Margaret Cancer Centre, University of Toronto, Ontario, Canada) provided assistance with genetic knockdown experiments for TK1 and TFAM, under the supervision of Dr. Aaron Schimmer 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

Note: Sections of this work has been accepted for publication with Blood (Liyanage, S. U., Hurren, R., Voisin, V., Bridon, G., Wang, X., Xu, C., MacLean, N., Siriwardena, T. P., Gronda, M., Yehudai, D., Sriskanthadevan, S., Avizonis, D., Shamas-Din, A., Minden, M. D., Bader, G. D., Laposa, R., & Schimmer, A. D. (2017). Leveraging increased cytoplasmic nucleoside kinase activity to target mtDNA and oxidative phosphorylation in AML. Blood, https://doi.org/10.1182/blood-2016-10-741207).

3.0 Introduction

AML cells have unique mitochondrial characteristics such as increased mitochondrial biogenesis, decreased spare reserve capacity, and increased reliance on oxidative phosphorylation compared to normal hematopoietic progenitors,(Lagadinou et al., 2013; Skrtic et al., 2011; Sriskanthadevan et al., 2015b) that make a subset of AML cells susceptible to agents that target mitochondrial function. Mitochondria contain multiple copies of their own 16.6 kB genome which encodes 13 subunits of electron transport chain complexes necessary for oxidative phosphorylation. These genes are replicated, transcribed, and translated within the mitochondria using specialized machinery. Mitochondrial DNA (mtDNA) is synthesized independent of cell cycle status and copied solely by mitochondrial DNA polymerase gamma (POLG), accompanied by replication factors such as mitochondrial transcription factor A (TFAM), mitochondrial RNA polymerase

(POLRMT), and the mitochondrial helicase Twinkle.(Copeland, 2012; Nunnari and Suomalainen,

2012) mtDNA biosynthesis is also dependent on an adequate supply of nucleotide pools, which are derived from mitochondrial and cytoplasmic pathways. Within the mitochondria, the mitochondrial nucleotide salvage pathway converts nucleoside precursors to nucleotides by a cascade of kinases.(Carling et al., 2011) Nucleotides are also imported into the mitochondria from the cytoplasm by specialized nucleotide transporters.(Gandhi and Samuels, 2011b; Kakuda, 2000;

Lam et al., 2005) In the cytoplasm, nucleoside diphosphates synthesized from the de novo biosynthesis pathway funnel into the cytoplasmic nucleotide salvage pathway. Kinases in this pathway catalyze the phosphorylation of nucleosides to nucleotides, similar to the mitochondria.(Lane and Fan, 2015; Mathews, 2015) In this chapter, we assessed the activity of the cytoplasmic nucleotide salvage pathway in AML and its contribution to mtDNA biosynthesis.

3.1 Results

3.1.1 A subset of primary AML cells display upregulated mtDNA biosynthesis

Our previous studies demonstrated that a subset of AML cells display increased mitochondrial biogenesis and mtDNA content compared to normal hematopoietic progenitors.

(Skrtic et al., 2011; Sriskanthadevan et al., 2015b) Here, we examined the expression of the 8 core mtDNA biosynthesis genes in AML: mitochondrial DNA polymerase gamma 1, 2 (POLG,

POLG2), mitochondrial RNA polymerase (POLRMT), mitochondrial transcription factor A

(TFAM), mitochondrial single-stranded binding protein (MTSSB), Twinkle, that collectively initiates and performs mtDNA replication,(Graziewicz et al., 2006; Ikeda et al., 2015) and mitochondrial nucleoside kinases deoxyguanosine kinase (DGUOK), and thymidine kinase 2

(TK2), involved in biosynthesis of mitochondrial nucleotide pools.(Carling et al., 2011)

We compared the mRNA expression of the mtDNA biosynthesis gene signature between 542 human primary AML samples at diagnosis and 73 normal non-leukemia samples (mononuclear cells isolated from healthy bone marrow and peripheral blood) (GSE13159).(Haferlach et al.,

2010) Unsupervised hierarchical clustering analysis of AML samples identified 4 subsets of AML with significantly different mtDNA biosynthesis expression pattern, designated as AML clusters

1-4 (Figure 3.1A-B) (P < 2x10-16, pair-wise two-sided t-tests adjusted for multiple hypothesis testing using Benjamini-Hochberg method).

Next, we compared the mtDNA biosynthesis pattern between all 4 AML clusters and normal samples. AML clusters 3 and 4, which comprise 55% of the total population, have higher mtDNA biosynthesis expression compared to median expression of normal samples; AML cluster 2 is not

78 different than normal (p> 0.05), and AML cluster 1 has lower mtDNA biosynthesis expression compared to normal samples. (Figure 3.1B) (P < 2x10-16 for all pair-wise two-sided t-tests). These results suggest an increased mtDNA biogenesis in a subset of AML samples and cell lines.

Figure 3.1: Subsets of AML display upregulated mtDNA biosynthesis expression A) Expression pattern and hierarchical clustering of microarray data from 542 primary human AML and 73 normal non-leukemic and healthy bone marrow samples(Haferlach et al., 2010) for 8 mtDNA biosynthesis genes. The four main AML mtDNA biosynthesis clusters (purple, green, red, cyan) were designated as 1, 2, 3 and 4. A red color indicates a higher expression compared to the mean of all AML and normal samples and a blue color indicates a lower expression. B) Boxplot of scaled data distribution (z-score) of the 4 AML clusters and normal samples for each of the 8 mtDNA biosynthesis genes from panel A. Values displayed on boxplot indicate median z-score values. Two-sided t-tests were applied to estimate significance of differences between AML and normal clusters.

3.1.2. Leukemia cell lines display upregulated mtDNA biosynthesis expression

Next, we investigated the expression of mtDNA biosynthesis factors in AML cell lines using publicly available datasets. POLG, POLG2, Twinkle and TFAM mRNA expression was upregulated preferentially in leukemia cell lines compared to other cancer cell lines based on the

Cancer cell line encyclopedia (Figure 3.2). These results suggest an increased mtDNA biogenesis in a subset of AML samples and cell lines.

Figure 3.2: mtDNA biosynthesis genes are upregulated in leukemia cell lines (A)POLG, (B)POLG2, (C) TFAM and (D) Twinkle (D) are upregulated in leukemia cell lines (Cancer Cell line encyclopedia).

3.1.3. AML cells display upregulated mitochondrial nucleotide transporter expression

As increased mtDNA biosynthesis is associated with a larger requirement of available nucleotide pools, this need can be compensated by importing cytoplasmic pools through mitochondrial nucleotide transporters. Therefore, we investigated pathways which support mitochondrial nucleotide pools and observed higher overall gene expression levels of three of four known mitochondrial nucleotide transporters(Di Noia et al., 2014; Govindarajan et al., 2009) SLC25A33,

SLC25A36, and SLC29A3 but not SLC29A1 in AML compared to normal cells (Figure 3.3A, P

< 1x10-8, t-test), suggesting increased nucleotide import from the cytoplasm. We also observed a positive correlation (p<0.05, paired t-test of r values) between mtDNA nucleotide transporters

SLC29A3 (r=0.42) and SLC26A36 (r=0.2), but not SLC29A1, SLC25A33 and mtDNA expression in AML (Figure 3.3B).

Z score

B -3 -2 0 2 3 AML mtDNA biosynthesis

1 2 3 4 R p

SLC25A33 0.06 0.16 SLC25A36 0.2* 2.5e-6

SLC29A1 0.08 0.06 SLC29A3 0.42* <2.2e-16

Figure 3.3: Mitochondrial nucleotide transporters are upregulated in primary human AML samples. A. Boxplots of scaled data distribution (z-score) of mitochondrial nucleotide transporters (SLC25A33, SLC25A36, SLC29A3, and SLC29A) in 542 primary human AML and 73 normal samples. Two-sided t-test was applied to estimate significance of differences between these 2 groups. B. Expression pattern of mitochondrial nucleotide transporters in AML mDNA biosynthesis clusters 1-4. Pearson correlation (r) between scaled gene expression and median mtDNA biosynthesis expression (z-scores) was determined and statistical significance was assessed by paired t-test of r- values.

3.1.4. A subset of cytoplasmic nucleoside kinases are upregulated in AML

Next, we characterized the expression of kinases involved in biosynthesis of cytoplasmic nucleotide pools.(Mathews, 2015) Collectively, this kinase cascade catalyzes three sequential phosphorylation steps to convert deoxynucleosides to deoxynucleotides. Nucleoside monophosphate kinase CMPK1 and nucleoside diphosphate kinases NME1-NME2, which perform the second and third step of nucleoside phosphorylation, respectively, were upregulated in AML samples compared to normal hematopoietic cells by gene expression analysis . In contrast, nucleoside kinases deoxycytidine kinase (DCK) and thymidine kinase 1 (TK1), which perform the first phosphorylation step, were not upregulated in AML compared to normal (Figure 3.4A).

Among AML samples, there was a positive correlation between NME1-NME2 (r=0.57, p<0.05) and DCK (r=0.27, p<0.05), but not TK1 and CMPK1 and mtDNA expression in AML (Figure

3.4B). We also assessed protein levels in an independent set of AML patient samples. Higher levels of POLG, CMPK1 and NME2 protein were seen in subset of primary AML samples (Table

1 for patient characteristics) compared to normal hematopoietic progenitors (G-CSF mobilized peripheral blood stem cells (PBSC’s) from healthy donors) by immunoblotting (Figure 3.4C).

B AML mtDNA biosynthesis

1 2 3 4 R p Z score

CMPK1 0.08 0.05 -3 -2 0 2 3

NME1-NME2 0.57* <2.2e-16

DCK 0.27* 7.5e-11

TK1 -0.01 0.89

C AML (n=12) PBSC (n=5) POLG

CMPK1 NME2

GAPDH

Figure 3.4: Cytoplasmic nucleoside kinases are upregulated in a subset of AML A) Boxplot of scaled data distribution of cytoplasmic nucleoside kinases CMPK1 between the 542 primary human AML and the 73 normal samples. Two-sided t-test was applied to estimate significance of differences between these 2 groups. B) Expression pattern of mitochondrial nucleotide transporters in AML mDNA biosynthesis clusters 1-4. Pearson correlation (r) between scaled gene expression and median mtDNA biosynthesis expression (z-scores) was determined and statistical significance was assessed by paired t-test of r-values. C) Total proteins were extracted and immunoblotted for POLG and cytoplasmic nucleoside kinases CMPK1 and NME2 in 12 primary human AML cells and 5 normal G-CSF mobilized peripheral blood stem cells (PBSC’s).

Table 1: Leukemia patient characteristics

Sample # Patient Age of Gender Diagnosis Cytogenetics Molecular ID Diagnosis A1 NPM1 pos, FLT3- AML with NPM1 141065 58 Male 46,XY[20] ITD pos, FLT3- mutated TKD neg A2 NPM1 neg, FLT3- 140176 79 Male AML,M5 46,XY[20] ITD pos, FLT3- TKD neg A3 NPM1 pos, FLT3- Acute monocytic 140301 60 Female 46,XX[20] ITD pos, FLT3- leukemia TKD neg A4 NPM1 pos, FLT3- AML with NPM1 160004 67 Female 46,XX[20] ITD neg, FLT3- mutated TKD neg A5 AML with NPM1 neg, FLT3- 140372 74 Male myelomonocytic 46,XY[20] ITD pos, FLT3- differentiation TKD neg A6 46,XY,7, AML with MDS 151258 78 Male +mar[5]/ not done related changes 46,XY[2] A7 NPM1 pos, FLT3- 161476 56 Female AML unsuccessful ITD pos A8 NPM1 pos, FLT3- AML with NPM1 46,XY[20] 151257 58 Male ITD neg, FLT3- mutated TKD neg A9 NPM1 pos, FLT3- AML with NPM1 160406 64 Female unsuccessful ITD pos, mutated PML/RARA neg A10 AML with MDS 46,XY,t(3;5) 160556 55 Male not done related features (q21;q35)[10] A11 NPM1 pos, FLT3- AML with 1411104 67 Female 46,XX[10] ITD pos, FLT3- mutated NPM1 TKD neg A12 45~46,X,add( X)(q22),-2,- 4,add(7) Secondary AML, (q31),der(19)t( 140175 66 Female post PRV, JAK2 not done 2;19)(q11.2;q1 V617F pos 3.1),add(20)(q 11.2),+1~2mar [cp9] B1 151258 78 Male AML with MDS 46,XY,7,+mar PML/RARA neg related changes [5]/ 46,XY[2]

B2 151123 58 Female CML blast crisis 46,XX,t(9;22)( BCR/ABL1 pos q34;q11.2)[20] B3 151149 69 Male AML, NOS 46,XY[20] NPM1 neg, FLT3- TKD neg, FLT3- ITD pos B4 162235 68 Male AML with 46,XY[20] NPM1 neg, FLT3- monocytic ITD neg differentation B5 162252 26 Female AML with 46,XX[20] NPM1 pos, FLT3- mutated NPM1 ITD neg, PML/RARA neg B6 151257 NPM1 pos, FLT3- AML with NPM1 46,XY[20] 58 Male ITD neg, FLT3- mutated TKD neg C1 151258 78 Male AML with MDS 46,XY,7,+mar PML/RARA neg related changes [5]/ 46,XY[2] C2 151123 58 Female CML blast crisis 46,XX,t(9;22)( BCR/ABL1 pos q34;q11.2)[20] C3 151149 69 Male AML, NOS 46,XY[20] NPM1 neg, FLT3- TKD neg, FLT3- ITD pos C4 151536 76 Male CMML-2 47,XY,+8[20] C5 Unknow n C6 150549 57 Female AML 46,XX[20] NPM1 neg, FLT3- (PB) ITD neg, FLT3- TKD neg C7 Unknow n C8 151257 58 Male AML with NPM1 46,XY[20] NPM1 pos, FLT3- mutated ITD neg, FLT3- TKD neg C9 151311 77 Female AML, monocytic 46,XX,del(5)( without q13q31)[6]/47, maturation idem,+11[14] Xenograft 1 110102 50 Female AML 45,XX,inv(3)( q21q26),-7[20] Xenograft 2 080315 35 Male AML 45,XY,inv(3)( CBFB-MYH11 myelomonocytic q21q26.2),- neg 7[20] Xenograft 3 090240 54 Female AML without 52,XX,+2,+9, maturation +10,+13,+14,+ 15[20]

3.2 Cytoplasmic nucleoside kinases regulate mtDNA biosynthesis

3.2.1 Knockdown of cytoplasmic nucleoside kinases deplete mtDNA in AML

To determine whether cytoplasmic nucleoside kinases are functionally important for mtDNA biosynthesis, we performed lentiviral-mediated shRNA knockdown of DCK, TK1, and CMPK1 cytoplasmic nucleoside kinases in TEX and OCI-AML2 AML cell lines. In TEX leukemia cells, knockdown of DCK reduced levels of mtDNA to 63% compared to control cells, while knockdown of TK1 and CMPK1 did not affect mtDNA levels. In contrast, knockdown of CMPK1 and TK1 in

OCI-AML2 cells reduced mtDNA levels to 60% and 55% respectively compared to controls, but

DCK knockdown did not reduce mtDNA levels (Figure 3.5A-C). Knockdown of the mtDNA replication factor TFAM decreased mtDNA content in both TEX and OCI-AML2 cells (Figure

3.5D). Target knockdown was confirmed by immunoblotting (Figure 3.5E). Thus, cytoplasmic nucleoside kinases contribute to mtDNA biogenesis, but contribution of individual kinases is cell- dependent.

A TEX O C I-A M L 2

1 . 0 2 . 0

**** *** 1 . 5

0 . 5 1 . 0

0 . 5 Relative mtDNA Relative mtDNA 0 . 0 0 . 0 2 2 1 1 n t ro l n t ro l C K C K C o C o h D h D h D C K h D C K s s s s B TEX O C I-A M L 2 1 . 5 1 . 0

* 1 . 0 *

0 . 5 0 . 5 Relative mtDNA Relative mtDNA 0 . 0 0 . 0 1 2 1 2 n t ro l n t ro l K 1 K 1 C o C o h T h T h TK 1 h TK 1 s s s s

C TEX O C I-A M L 2

2 .0 1 . 0

** 1 .5 **

1 .0 0 . 5

0 .5 Relative mtDNA Relative mtDNA 0 .0 0 . 0 l 1 2 1 1 2 n t ro l P K 1 PK C o n tro C o C M P K 1 CM C M P K 1 h C M s h s h s s h

D TEX O C I-A M L 2

1 . 0 1 . 0

0 . 5 0 . 5 ** **** **** *** Relative mtDNA Relative mtDNA 0 . 0 0 . 0 1 2 2 ro l t ro l M 1 FAM FAM C o n t C o n h T h TF A M h T s h TF A s s s

E TEX OCI-AML2

DCK DCK

β-Tubulin β-Tubulin

Control Control shDCK 1 shDCK 2 shDCK 1 shDCK 2

TK1 TK1 β-Actin β-Actin

Control Control shTK1 2 shTK1 1 shTK1 2 shTK1 1

CMPK1 CMPK1

β-Actin β-Actin

Control Control shCMPK1 shCMPK1 1 shCMPK1 2 shCMPK1 1

TFAM TFAM β-Actin β-Actin

Control Control shTFAM 1 shTFAM shTFAM 2 shTFAM shTFAM 2 shTFAM shTFAM 1 shTFAM

Figure 3.5: Cytoplasmic nucleoside kinases regulate mtDNA biosynthesis. A) DCK, (B)TK1, (C) CMPK1 and (D) TFAM were knocked down with lentiviral-mediated shRNA in TEX and OCI-AML2 cells as described in the materials and methods. After target knockdown, mtDNA content was assessed by qRT-PCR using primers for mt-ND1 relative to nuclear-encoded HGB. E) Protein knockdown was confirmed by immunoblotting in whole cell extracts. Representative Immunoblots are displayed. Data are shown as mean ± SD of three independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 using Bonferroni post-test after one-way ANOVA.

3.3 Assessment of cytoplasmic nucleoside kinase activity

3.3.1 Cytoplasmic nucleoside kinase activity is elevated in AML

2’3’-dideoxycytidine (ddC) mimics native nucleosides and undergoes activation to its triphosphorylated state ddC-triphosphate (ddCTP) in the cytoplasm. It is activated by nucleoside kinases DCK, CMPK1, and nucleoside diphosphate kinases (NME), which perform the first, second, and third phosphorylation steps, respectively.(Chen and Cheng, 1992; Liou et al., 2002)

The activated form is imported into the mitochondria where it inhibits the catalytic activity of

POLG through chain termination or competition with native nucleotides.(Anderson et al., 2004;

Ray, 2005) To assess cytoplasmic nucleoside kinase activity, primary AML and normal hematopoietic progenitor cells (G-CSF mobilized peripheral blood stem cells (PBSC’s)) were treated with ddC and total levels of ddC and ddCTP were measured by liquid chromatography mass-spectrometry (LC-MS/MS). Levels of ddC were increased in 1 of 6 AML cells compared to the mean levels in normal hematopoietic cells. ddCTP content was higher in 6 of 6 AML samples compared to mean levels in normal hematopoietic cells (p<0.05, one-way ANOVA), and ddCTP levels were below the limit of quantification in the normal hematopoietic cells (Figure 3.6A). Of note, in all samples, increased ddCTP levels were not solely related to higher levels of ddC as increased ddCTP was seen in samples with lower levels of ddC (Figure 3.6B), these results indicate that cytoplasmic nucleoside kinase pathway activity is upregulated in AML compared to normal PBSC’s.

A * * 0 . 6 8

6 0 . 4 g p r o te in ) g p r o te in )   4 0 . 2 2 BLQ 0 . 0 0 [ddC] (pmol/ 1 2 1 2 L 2 [ddCTP] (pmol/ M L 1 M L 3 M L 1 A AM A A M L 2 A M L 3 A o rm a l o rm a l o rm a l o rm a l N N N N

* * * 8 4

6 3 g p r o te in )  4 2

2 1

BLQ 0 0 [ddC] (pmol/ 4 l 3 l 3 4 [ddCTP] (pmol/ug protein) L 4 L 4 AM A M L 5 AM A M L 5 o rm a l N o rm a N o rm a N o rm a l N

0 . 8 B * 8 5 2 R = 0 .1 g )

0 . 6  4 6 g p r o te in ) g p r o te in )  0 . 4  3 4 2 0 . 2 2 1 **

0 . 0 BLQ ddCTP (pmol/ 0 [ddC] (pmol/ 6 0 0 2 4 6 [ddCTP] (pmol/ l 5

A M L d d C (p m o l/ g ) A M L 6 o rm a l 5 N N o rm a Figure 3.6: Cytoplasmic nucleoside kinase activity is elevated in AML A) Relative quantification of total intracellular 2’3’-dideoxycytidine (ddC) and ddC-triphosphate (ddCTP) by LC-MS/MS in primary human AML (samples B1-B6) and normal hematopoietic progenitor samples following treatment with 2 µM ddC for 6 days. ddC and ddCTP content were normalized to total protein input and represented as mean ± SD (n=2) of technical replicates within each independent experiment. “BLQ” indicates below the limit of quantification. Levels of ddC and ddCTP in each panel represent semi-quantitative values. Each panel represents a separate experiment. Differences between ddC or ddCTP levels in each AML sample compared to mean of normal samples within an experiment was assessed using Bonferroni post-test after one-way ANOVA or student’s t-test. *P < 0.05, ***P < 0.001. B) Correlation between levels of ddC and ddCTP from samples in panel A was determined using Pearson’s correlation method.

A - B) Quantification of total intracellular 2’3’-95dideoxycytidine (ddC) and ddC-triphosphate (ddCTP) by LC/MS/MS in OCI-AML2, 3 primary human AML (samples B1-B3), and 2 normal PBSC samples following treatment with 2 µM ddC for 6 days. ddC and ddCTP content were normalized to total protein input and represented as mean ± SD. BLQ= below limit of quantification. Mean ddC or ddCTP levels between AML and normal samples were analyzed by student’s t-test for significance. 3.3.2. ddC is activated by cytoplasmic nucleoside kinases

To confirm that ddC activation is regulated by cytoplasmic nucleoside kinases, we perfomed lentiviral DCK knockdown in TEX and OCI-AML2 cells. DCK knockdown was confirmed by immunoblotting (Figure 3.7A). DCK knockdown decreased levels of ddCTP, independent of changes in levels of ddC in TEX cells. In contrast, knockdown of DCK in OCI-AML2 cells decreased levels of both ddC and ddCTP (Figure 3.7B-C).

TEX

A B TEX C TEX 1 5 0 1 5 0 NS shDCK 1 shDCK shDCK 2 Control 1 0 0 1 0 0 DCK d d C ****

5 0 d d C T P 5 0 ****

β-Tubulin **** **** % Total intracellular 0 % Total intracellular 0 l o l 1 1 K 2 K 2 o n t ro DCK DC DCK C o n t r C h D C s s h s h s h

OCI-AML2 O C I-A M L 2 O C I-A M L 2 1 5 0

1 5 0

1 0 0 1 0 0 Control shDCK

shDCK 2 ** *** d d C DCK 5 0 d d C T P 5 0 **** **** ****

β-Tubulin % Total intracellular 0 % Total intracellular 0

2 ro l 1 2 K 1 C K o n t ro l DC C o n t DCK C DCK h D s h s s h s h

Fig 3.7: DCK knockdown depletes levels of activated ddC. A) DCK was knocked down in TEX and OCI-AML2 cells using shRNA. Levels of DCK were measured in whole cell lysates by immunoblotting 7 days post transduction. A representative immunoblot is displayed B, C) Quantification of total intracellular ddC and ddCTP by LC-MS/MS in DCK knockdown or control cells following treatment with 1 µM ddC for 4 days. ddC and ddCTP levels were normalized to 106 cells input. Data are shown as mean ± SD of two biological replicates performed at least in triplicate.

3.4 Investigating the effect of ddC on mtDNA replication and cellular bioenergetics

3.4.1 ddC depletes mtDNA in AML cells

Previously, we and others demonstrated that AML cells and stem cells have increased reliance on oxidative phosphorylation due to decreased spare reserve capacity and an inability to upregulate glycolysis (Lagadinou et al., 2013; Sriskanthadevan et al., 2015a). ddCTP inhibits the sole mtDNA polymerase POLG, but not nuclear DNA polymerases. Given the increased activity of nucleoside kinases in AML cells over normal, we examined the effects of ddC treatment on mtDNA content and cellular bioenergetics. AML cell lines were treated with increasing concentrations of ddC that depleted mtDNA content in a panel of leukemia cell lines (n=4) (Figure 3.8). In OCI-AML2 and

TEX cell lines, we observed greater than 90% mtDNA depletion following 3 days of treatment with 500 nM ddC. Similar reductions in mtDNA levels were observed in HL-60 and K562 cells following ddC treatment.

O C I-A M L 2 TEX

1 .0 1 .0 D a y 3 0 .8 D a y 6

0 .6 0 .5

0 .4 *** ****

Relative mtDNA **** ****

Relative mtDNA0 .2 0 .0 0 .0 2 0 1 2 0 1 2 .2 .5 2 5 0 1 2 0 1 2 .2 0 0 .5 0 . 0 0 0 .5 0 . 0 . 0 .0 2 5 [d d C ]  M [d d C ]  M

H L -6 0 K 5 6 2

1 .5 D a y 3 1 .5 D a y 6

1 .0 1 .0

0 .5 0 .5

Relative mtDNA **** **** Relative mtDNA **** **** **** 0 .0 **** 0 .0 0 2 0 2 0 2 0 2 0 .2 0 .2 0 .2 0 .2 [d d C ]  M [d d C ]  M

Figure 3.8: ddC inhibits mtDNA replication in AML cell lines. OCI-AML2 cells, TEX, HL-60 and K562 cells were treated with ddC for 3 and 6 days. Relative mtDNA content was assessed by qRT-PCR. Mean ± SD, n = 3.

3.4.2 ddC depletes mtDNA encoded transcripts, proteins and oxidative phosphorylation

Subsequently, we evaluated the effect of ddC on mtDNA-encoded transcripts and proteins involved in mitochondrial bioenergetics. ddC treatment decreased mRNA expression of all 13 mtDNA-encoded genes which form subunits of electron transport chain (ETC) complexes. In contrast, no significant changes in mRNA expression of nuclear-encoded ETC complexes were observed (Figure 3.9, Table S2).

Similarly, ddC treatment selectively depleted mtDNA-encoded COX I and COX II protein subunits, which form the catalytic core of ETC complex IV/cytochrome c oxidase (Fontanesi et al., 2008) in a dose and time-dependent manner, with minor changes in the nuclear-encoded subunit COX IV (Figure 3.10A).

We next assessed the effect on mitochondrial respiration by measuring the basal oxygen consumption rate (OCR). ddC treatment decreased basal OCR in a dose dependent manner in OCI-

AML2 cells and TEX cells (Figure 3.10B).

100

1.0

0.5

0.0 Relative mRNA expression mRNA Relative TFAM SDHA SDHB SRP14 COX4I2 COX4I1 UQCRC MT-ND3 MT-ND4 MT-ND5 MT-ND6 MT-ND1 MT-ND2 MT-CO2 MT-CO1 MT-CO3 MT-CYB NDUFV2 NDUFS1 UQCRC1 MT-ATP8 MT-ATP6 MT-ND4L NDUFAB1 Mitochondrial Nuclear

Figure 3.9: ddC selectively depletes mtDNA-encoded transcripts OCI-AML2 cells were treated with 2 µM ddC for 10 days and relative mRNA levels of mitochondrial and nuclear-encoded subunits of ETC complexes was assessed using a custom-Nanostring chip by normalizing to DMSO-treated control cells. Additional doses and timepoints displayed in table below

101

Table 2: Nanostring analysis of ETC subunits in OCI-AML2 cells following ddC treatment

102

O C I-A M L 2 5 0 0

4 0 0

3 0 0 *

2 0 0

1 0 0 ** Basal OCR (pMol/min) 0 0 0 .2 2 [d d C ]  M

TEX

1 5 0 *

1 0 0 **

5 0 Basal OCR (pMol/min) 0 0 0 .2 2 [d d C ] M

Figure 3.10: ddC targets oxidative metabolism proteins and activity A) Effect of ddC treatment on protein levels of mt-COX I, mt-COXII, nu-COX IV, and β-tubulin in whole cell extracts of OCI-AML2 and TEX cells; mt: mitochondrial and nu: nuclear. The immunoblot shown is from one representative experiment. B) Basal oxygen consumption rate (OCR) in OCI-AML2 cells after treatment with ddC for 6 days was assessed using Seahorse XF96 metabolic flux assay. Basal OCR in TEX cells was assessed after 3 days of treatment Mean ± SD, n = 3.

103

3.4.3 ddC alters mitochondrial structure independent of changes in mitochondrial mass

Inhibition of oxidative phosphorylation can be associated with reductions in mitochondrial mass, thus we investigated the effect of ddC on mitochondrial mass. ddC treatment did not decrease mitochondrial mass (Figure 3.11A). ddC treatment altered mitochondrial morphology in AML cells, such as decreased cristae density, abnormal concentric cristae formation, and enlarged mitochondria, consistent with previous reports of mitochondrial disorders associated with POLG mutations (Figure 3.11B)(Bourgeois and Tarnopolsky, 2004; Wallace and Fan, 2009).

104

105

Figure 3.11: Effect of targeting mtDNA replication on mitochondrial content and ultrastructure. A) Effect of ddC on mitochondrial mass in OCI-AML2 and TEX cells treated for 3 and 6 days. Cells were stained with Mitotracker Deep Red FM and analyzed by flow cytometry. Data is shown as mean fluorescence intensity ± SD of three independent experiments. B) Mitochondrial morphology was assessed by transmission electron microscopy in OCI-AML2 cells following ddC treatment at increasing concentrations and time periods. Representative images are shown. Scale bar = 2 µm.

106

3.5. Assessing the efficacy ddC in vitro and in vivo

3.5.1 ddC inhibits proliferation and viability of AML cell lines

Given that inhibiting mtDNA replication with ddC targeted oxidative phosphorylation, we assessed the efficacy of ddC in inducing anti-leukemic effects. ddC treatment resulted in decreased cell proliferation and induced cell death as observed by Annexin V+ staining in a panel of AML cell lines (Figure 3.12A-F). Reductions in cell proliferation and viability were associated with significant reductions in mtDNA-encoded proteins and inhibition of oxidative phosphorylation.

For instance, in OCI-AML2 cells, reductions in cell proliferation and viability were observed after

6 days of treatment with ddC at doses greater than 0.5µM, which corresponded to greater than

95% reduction in mtDNA content, significant reductions of mtDNA-encoded ETC subunits COX

I, COXII and inhibition of oxidative phosphorylation.

107

Figure 3.12: ddC induces anti-leukemic effects in a panel of AML cell lines. A-D) Cell viability and proliferation of OCI-AML2, TEX, HL-60 and K562 cells treated with ddC for 6-10 days. Viable cells were assessed by trypan blue exclusion staining. Representative data are shown as the mean ± SD. E-F) OCI-AML2 and TEX cells were treated with ddC for 3-10 days. Annexin V staining was assessed by flow cytometry. Data are shown as the mean ± SD of three independent experiments. *P < 0.05, **P < 0.01, and ***P < 0.001 using Bonferroni post-test after one- way ANOVA.

108

3.5.2. HEK 293 cells are resistant to ddC in comparison to AML cells

In contrast to leukemic cells, HEK 293 cells displayed decreased sensitivity to ddC; a 10-fold greater dose (5 µM) was required to observe significant reductions in cell proliferation and mtDNA compared to AML cells(Figure 3.13A-B). We characterized the mechanism of resistance in HEK

293 by quantifying total intracellular levels of ddC and ddCTP. Despite 2 µM ddC treatment for 6 days, ddC and ddCTP were below the limit of quantification in HEK 293 cells compared to OCI-

AML2 cells (Figure 3.13C-D). Thus, increased resistance to ddC in HEK 293 cells is likely due to decreased ddC import.

109

Figure 3.13: ddC is preferentially active in AML cell lines compared to HEK 293 cells. A) Cell viability and proliferation of HEK 293 cells treated with ddC for 10 days. Viable cells were assessed by trypan blue exclusion staining. Data are shown as the mean ±SD of two independent experiments. B) Effect of ddC on mtDNA content in HEK 293 cells treated for 3, 7, and 10 days. Relative mtDNA was assessed by qRT-PCR. Data are shown from one of two representative experiments as mean ± SD. C and D) Quantification of total intracellular ddC (C) and ddCTP (D) by LC/MS/MS in OCI-AML2 and HEK 293 cells following treatment with 2 µM ddC for 6 days. ddC and ddCTP levels are shown as mean ± SD of absolute peak area per input of 5x106 cells. BLQ: below limit of quantification. **P < 0.01 and ***P < 0.001 using Bonferroni post-test after one-way ANOVA.

110

3.5.3. mtDNA-replication is functionally important for ddC mechanism of action

We next tested whether the effects of ddC on mtDNA replication and mitochondrial bioenergetics was functionally important for the observed cell death in leukemia. We observed that mtDNA- deficient P16669 Rho(0) cells, which display an oxidative-phosphorylation defective phenotype, were resistant to ddC treatment while the parental wild type osteosarcoma P143B cells were sensitive to ddC (Figure 3.14). Taken together, these data show that ddC preferentially inhibits

AML cell viability and proliferation by targeting the mtDNA replication activity of POLG.

111

WT R h o ( 0 ) 1 2 0

8 0

4 0 % Cell Viability 0 0 5 1 0 0 5 1 0 [d d C ]  M

Figure 3.14: MtDNA-deficient Rho (0) cells are resistant to ddC treatment. Osteosarcoma WT P143B and Rho (0) P16669 cells were treated with ddC for 10 days and cell viability was assessed by Sulforhodamine B assay. Mean ± SD, n = 2.

112

3.5.4. ddC preferentially targets mtDNA replication in primary AML cells and induces anti- leukemic effects in vitro

Given the effects in AML cell lines, we investigated the effect of ddC on primary leukemia cells and normal hematopoietic progenitors (Table 1 for patient characteristics). ddC depleted mtDNA content in 7 of 9 primary leukemia samples, whereas, a smaller reduction in mtDNA content was observed in all tested hematopoietic progenitors (n=7), (Figure 3.15A and Table 3) including the

CD34+ fraction (Figure 3.15B). We also assessed the effect of ddC on the viability of primary

AML and normal hematopoietic cells. Six of nine primary leukemia samples were sensitive to ddC treatment in vitro as empirically defined as a cell viability less than 75% following 6 days of treatment with 2 µM ddC compared to controls. In contrast, 7 out of 8 samples of normal hematopoietic cells were resistant to ddC treatment in vitro (Figure 3.15C). Normal CD34+ hematopoietic cells were also resistant to ddC (Figure 3.15D). Preferential anti-leukemic activity was associated with greater reductions in mtDNA content as sensitive leukemia samples displayed on average, a two-fold lower mtDNA content after ddC treatment compared to normal hematopoietic cells (Table 3). Thus, ddC preferentially inhibits mtDNA biosynthesis in AML cells and selectively targets a subset of AML cells in vitro.

113

A B

2 . 2 Normal CD34+ 2 . 0 1 . 5 0 . 6 1 . 0 0 . 4 0 . 5 0 . 2 Relative mtDNA Relative mtDNA 0 . 0 0 . 0 A M L N o rm a l C o n t ro l d d C

C D Normal CD34+ * 2 0 0 1 0 0

1 5 0 7 5

1 0 0 5 0

5 0 2 5 % Cell Viability % Cell Viability 0 0 A M L N o rm a l C o n t ro l d d C

Figure 3.15: ddC preferentially targets mtDNA replication and induces anti-AML effects in primary AML samples in vitro A) Primary leukemia and normal hematopoietic progenitor cells (G-CSF mobilized peripheral blood stem cells (PBSC’s)) were treated with 2 µM ddC for 6 days. mtDNA content was assessed by qRT-PCR. Leukemia samples C1-C9 were used for analysis. B) Normal PBSC’s were treated with 2 µM ddC for 6 days and sorted for CD34+ subpopulation using immunomagnetic selection. mtDNA content was assessed in CD34+ population by qRT-PCR. C) Cell viability was assessed by trypan blue exclusion staining in primary AML cells and Cyquant DNA staining for PBSCs from panel (G). Dotted line indicates the cut-off to stratify samples as ddC-sensitive or ddC-resistant. D) Normal PBSC’s were treated with 2 µM ddC for 6 days and cell viability was assessed by PI staining in CD34+ subpopulation by flow cytometry. For all experiments, *P < 0.05, Student’s t-test.

114

Table 3: Effect of ddC treatment on primary cells Effect of ddC on primary leukemia and normal hematopoietic progenitors on cell viability and mtDNA content following 6 days of treatment at 2 µM ddC

Sample ddC response % cell viability % mtDNA AML C1 75 21 AML C2 47 41

AML C3 70 22 sensitive AML C4 53 21 AML C5 65 3 AML C6 30 20 AML C7 113 210 AML C8 82 210 AML C9 118 10 Normal 1 146 31

Normal 2 87 57 resistant Normal 3 115 30 Normal 4 132 43 Normal 5 84 47 Normal 6 199 37 Normal 7 89 24 Normal 8 sensitive 60 -

115

3.5.5 ddC displays efficacy in mouse models of human AML

We next investigated whether inhibiting mtDNA biosynthesis with ddC can target AML cells in vivo. In the xenograft tumor model, OCI-AML2 cells were injected subcutaneously into the flank of severe combined immune deficient (SCID) mice. After tumors were palpable, mice were treated with ddC once daily by intra-peritoneal (i.p.) injection for 11 days. Treatment with low doses of ddC (35 and 75 mg/kg) induced tumor regression and decreased tumor mass by greater than 75% relative to vehicle controls (Figure 3.16A and B, P < 0.0001, Bonferroni post-test after one-way

ANOVA). Higher doses of ddC (150 and 300 mg/kg) produced tumor regression and reductions in tumor mass by greater than 90% compared to vehicle controls (Figure 3.16D and E, P < 0.0001,

Bonferroni post-test after one-way ANOVA). Similar to the in vitro results, we observed greater than 90% mtDNA depletion in OCI-AML2 tumors excised at the end of treatment at all doses of ddC (Figure 3.16C, F, G). Consequently, we observed reductions in mitochondrial COX I and

COX II mRNA (Figure 3.16H) and in COX II protein levels in OCI-AML2 xenografts (Figure

3.16I).

116

117

Figure 3.16: ddC displays efficacy in OCI-AML2 xenograft model of human AML. A - B) OCI-AML2 cells were injected subcutaneously into the flank of SCID mice. Once tumors were palpable (day 9), mice were treated with ddC (35 or 75 mg/kg once daily by i.p. injection) or vehicle control for 11 days (n = 8 per group). At the end of the treatment, mice were sacrificed and tumor volume (A) and weight (B) were assessed from excised tumors. Data represents mean ± SD. C) Relative mtDNA was assessed from xenograft tumors excised from ddC or vehicle treated mice in panel A by qRT-PCR. Data represents mean ± SD (n = 3 per group). Student’s t-test was performed for control vs ddC. D-E) OCI-AML2 cells were injected subcutaneously into the flank of SCID mice. Once tumors were palpable (day 11), mice were treated with ddC (150 or 300 mg/kg/day by i.p. injection) or vehicle control (n = 7 per group). After 9 days of treatment, mice were sacrificed and tumor volume and mass was assessed from excised tumors. ***P < 0.001 using Bonferroni post-test after one-way ANOVA. F-G) Relative mtDNA was assessed from xenograft tumors excised from ddC or vehicle treated mice in panel D by qRT-PCR. Data represents mean ± SD (n = 3 per group). Student’s t-test was performed for control vs ddC. H) Relative mRNA expression for mt-COX I and mt-COX II was assessed by qRT-PCR in tumors excised from SCID mice treated with 300 mg/kg/day of ddC or vehicle control for 11 days. Data represents mean ± SD (n = 3 per group). I) Tumors were excised from vehicle and ddC treated mice (35 and 75 mg/kg/day) following 11 days of treatment and assessed for protein levels of mt-COX II, nu-COX IV, and VDAC from whole cell extracts was assessed by immunoblotting.

118

3.5.6. ddC is well-tolerated in mouse models of human AML

We evaluated the toxicity of ddC in the SCID-mouse model at doses which produced anti-leukemic effects in vivo. Treatment with ddC did not affect mouse body weight or behavior, activities of liver and enzymes in serum and histology of organs known to have high mitochondrial mass

(Wang, 2010) (Figure 3.17A-C). Importantly, ddC did not affect normal mouse hematopoiesis as no significant decrease in mouse leukocyte counts was observed (Figure 3.17D). Thus, ddC may be clinically effective in a subset of AML as a novel therapeutic agent.

119

120

121

Figure 3.17: Safety profile of ddC in mouse models of human AML A) SCID mice were injected with 35, 75, 150 or 300 mg/kg/day ddC by i.p. injection and body weight was assessed (n=7 per group). B) SCID mice were treated with 300 mg/kg ddC once daily by i.p. injection. Liver enzymes aspartate transaminase, bilirubin, creatinine, and creatine kinase activity from the serum was assessed after 11 days of treatment. Line represents mean (n=3 per group). C) Organ histology of SCID mice following treatment with ddC at 300 mg/kg/day by i.p. injection for 11 days. Representative sections of heart, liver, lung, muscle, and kidney tissue stained with hematoxylin and eosin are shown. Scale bar = 300 µm. D) SCID mice were injected with 35, 75 or 150 mg/kg/day ddC by i.p. injection. Leukocytes were profiled after 11 days of treatment (n=3 per group).

122

3.5.7. ddC targets bulk AML and LSC population in vivo

We also assessed whether ddC can target primary AML cells in vivo (Pearce et al., 2006). Primary cells from 3 patients diagnosed with AML were injected intra-femorally into NOD-SCID mice

(See Table 1 for patient characteristics). Eleven days after injection, mice were treated for 5 out of 7 days for 3 weeks at 75 mg/kg ddC by i.p. injection once daily. ddC treatment decreased levels of primary engraftment in the bone marrow, as assessed by quantification of human CD45+ CD19-

CD33+ cells (n=3, ****P < 0.0001 and *P < 0.05, Student’s t-test) (Figure 3.18A-C). Next, we determined whether ddC targets the leukemic stem cell (LSC) fraction by evaluating secondary engraftment. Primary AML cells harvested from the bone marrow of ddC-treatment mice were engrafted in untreated secondary recipient mice, and human leukemic cell engraftment was assessed after five weeks. ddC significantly reduced AML secondary engraftment (P < 0.0001,

Student’s t-test) (Figure 3.18D). Overall, ddC targets bulk AML and LSCs and inhibits mtDNA replication and ETC function in vivo.

123

Figure 3.18: ddC targets LSC populations in mouse models of human AML A-C) Three primary human AML cell samples were injected intra-femorally into irradiated female NOD/SCID mice. Mice were treated with 75 mg/kg/day of ddC by i.p. injection or vehicle control on day 11 for three weeks (n = 7/group). Following treatment, human leukemia cell engraftment in the left femur was assessed by flow cytometry of human CD45+CD33+ CD19- cells. D) Secondary engraftment was assessed by injecting viable leukemia cells from the bone marrow of ddC-treated and vehicle mice and injected into the right femur of irradiated female NOD/SCID mice, which remained untreated. Five weeks later, human leukemia cell engraftment in the left femur was measured by ﬂow cytometry of human CD45+CD33+ CD19- cells. Line represents mean engraftment of human cells. For all experiments, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 using Bonferroni post-test after one-way ANOVA in panels A-C, and Student’s t-test in panels D, F-I.

124

4. Discussion and Future Directions

4.1 Discussion

In our current study, gene expression analysis demonstrated an increase in mtDNA biogenesis genes in a subset of primary human AML samples compared to normal hematopoietic samples. A limitation of this experiment is the comparison between non-leukemic mononuclear cells, which comprise of more differentiated cells, to AML samples, which resemble stem-cell like fractions.

However, increased mitochondrial biosynthesis in LSC’s have been previously reported.

Upregulation of mitochondrial biogenesis factors, including TFAM, c-myc and NRF were observed in LSC’s compared to HSC’s (Sriskanthadevan et al., 2015a). Additionally, our results are also consistent with previous reports of upregulated mtDNA levels in a subset of primary AML cells.(Boultwood et al., 1996; Skrtic et al., 2011) Given that increased mitochondrial nucleotide metabolism is associated with larger nucleotide pools, we hypothesized that this need is supported by import of cytoplasmic nucleotide pools in AML. Confirming this, we detected significant mRNA upregulation of mitochondrial nucleotide transporters, which transport nucleotides and macromolecules from the cytoplasm into the mitochondria. We did not detect upregulation of the mitochondrial nucleotide transpoter, SLC29A1. This may be due to redundancies in mitochondrial nucleoside transporters, such as similar substrate profiles. Additionally, upregulation of a subset of cytoplasmic nucleoside kinases, involved in the biosynthesis of cytoplasmic nucleotide pools was observed in AML at both the mRNA and protein level. We detected an upregulation of

CMPK1 and NME2, but not TK1 and DCK. The differential expression of nucleoside monophosphate and disphosphate kinases compared to nucleoside kinases may be attributable to substrate flux by other pathways. A larger substrate load of nucleoside diphosphates, generated by de novo biosynthesis, feed into the nucleotide salvage pathway at the nucleoside diphosphate step.

125

Hence NME2 may be upregulated to compensate for an increased amount of nucleoside diphosphate precursors. Similarly, CMPK1 also phosphorylates dUMP, required for RNA synthesis, while thymidine precursors are not required for RNA synthesis. As RNA intermediates are more abundant in the cell compared to DNA precursors, CMPK1 may be overexpressed to support increased cellular transcription and cell proliferation

Demonstrating a functional link between cytoplasmic and mitochondrial nucleotide pools, we also showed that knockdown of nucleoside kinases depleted mitochondrial DNA content. Our study is the first to report the contribution of cytoplasmic nucleoside kinases to mtDNA levels in human cells. These findings are in line with a recent study in Drosophila, which demonstrated that overexpression of cytoplasmic deoxynucleoside kinase increased mtDNA content and oxidative metabolism proteins.(Tufi et al., 2014) Through these experiments, we highlighted the contribution of cytoplasmic nucleotide metabolism, in particular, the nucleotide salvage pathway to support mitochondrial nucleotide metabolism.

During nucleotide biosynthesis, nucleoside precursors undergo sequential phosphorylation to form deoxynucleotide triphosphates (dNTP’s) through catalysis by nucleoside kinases. To assess the activity of nucleoside kinases in AML, we tested the conversion of ddC to its activated triphosphorylated form, ddCTP. We demonstrated increased cytoplasmic nucleoside kinase activity in AML by detecting increased levels of conversion of ddC to ddCTP in AML cells. Since ddC acts as an anti-metabolite that is minimally incorporated into nuclear DNA once triphosphorylated, levels of ddCTP are readily quantifiable by mass spectrometric approaches.7 In contrast, using an endogenous nucleoside to assess levels of nucleoside kinase activity is challenging due to constant turnover of nucleotide pools during genomic replication and degradation to nucleoside precursors

126 such as nucleosides and nucleoside mono and diphosphates (Ferraro et al., 2006). However, the binding affinities of nucleotide salvage enzymes differ between substrates, such as native nucleosides and nucleoside analogs, hence rates of nucleotide biosynthesis cannot be extrapolated from this study (Liou et al., 2002).

Having demonstrated increased cytoplasmic nucleoside kinase activity in AML, we then leveraged this vulnerability to preferentially activate the POLG inhibitor, ddC, in AML cells. mtDNA depletion with ddC inhibits ETC function and targets both bulk AML and LSC populations in mouse models of human leukemia without overt toxicity to normal cells. Our findings reflect a dependence on oxidative phosphorylation to support leukemic stem cell survival and tumor growth. In addition, these results support our previous findings that AML cells have decreased spare reserve capacity and increased sensitivity to strategies that target oxidative phosphorylation.(Sriskanthadevan et al., 2015b) Moreover, ATP depletion may further impair mtDNA replication through inadequate ATP-dependent topoisomerase II activity and resultant formation of mtDNA catenanes.(Gattermann and Aul, 1997) Contrary to nuclear DNA, mtDNA replication occurs independent of cell cycle status,(Wang, 2010) hence is active in quiescent cells and can be therapeutically targeted by the nucleoside analog ddC.

Selectively targeting mtDNA polymerase activity is a characteristic feature of anti-viral nucleoside analogs. The most potent inhibitor of this class reported is ddC,(Setzer et al., 2005) with increased sensitivity of approximately 100-1000 fold greater for POLG over nuclear DNA polymerases, as determined by in vitro enzymatic primer extension assays.(Martin et al., 1994) In contrast, chemotherapeutic nucleoside analogues used in AML such as cytarabine (AraC), clofarabine, and

127 cladribine target nuclear DNA polymerases and are less active against POLG. For example, AraC has 50-fold less affinity for POLG compared to nuclear POLA, whereas, clofarabine and cladribine are 54 and 9 fold-less potent, respectively.(Parker et al., 1991; Ross et al., 1990)

6-34% of patients treated with ddC as an anti-HIV treatment regimen experienced grade 3 or higher mitochondrial toxicities similar to those observed in patients with POLG associated mitochondrial disorders, such as reversible peripheral neuropathy and lactic acidosis, although lactic acidosis was very rare, occurring in 1 in 10,000 patients.(1993; Anderson et al., 2004; Dalakas et al., 2001;

Fischl et al., 1995; Kallianpur and Hulgan, 2009; Lewis et al., 2003) In contrast, treatment with current non-nucleoside anti-viral agents such as indinavir and nevirapine, which do not inhibit

POLG, do not display POLG related mitochondrial toxicities.(Apostolova et al., 2011; van Leth et al., 2004) The mitochondrial toxicities seen with ddC suggest that anti-leukemic effects could be observed at clinically achievable concentrations.

POLG inhibitors, such as ddC, offer advantages compared to other inhibitors of oxidative phosphorylation in AML cells, such as BH3-mimetics. As ddC requires an activation step, it displays a higher safety profile as this step is preferentially upregulated in AML. In contrast, BH3 mimetics do not require an activation step, thus a smaller therapeutic window between AML and normal cells may exist. More importantly, our recent data shows ddC induces markers of myeloid differentiation. It is possible that the combination of targeting oxidative metabolism and inducing differentiation contribute to its potent anti-leukemic effect.

128

In summary, our study demonstrates that a subset of AML cells have elevated cytidine nucleoside kinase pathway activity and supports mitochondrial nucleotide metabolism. We leveraged this unique biological vulnerability to preferentially target mitochondrial bioenergetics in AML using ddC. ddC was preferentially converted from its pro to active form where it depleted mtDNA, inhibited oxidative phosphorylation and selectively targeted AML cells and stem cells in vitro and in vivo.

129

4.2 References

1. (1993). Zalcitabine IARC Monograph 76, 129-151. 2. Anderson, P.L., Kakuda, T.N., and Lichtenstein, K.A. (2004). The cellular pharmacology of nucleoside- and nucleotide-analogue reverse-transcriptase inhibitors and its relationship to clinical toxicities. Clin Infect Dis 38, 743-753. 3. Apostolova, N., Blas-Garcia, A., and Esplugues, J.V. (2011). Mitochondrial interference by anti-HIV drugs: mechanisms beyond Pol-gamma inhibition. Trends Pharmacol Sci 32, 715- 725. 4. Boultwood, J., Fidler C Fau - Mills, K.I., Mills Ki Fau - Frodsham, P.M., Frodsham Pm Fau - Kusec, R., Kusec R Fau - Gaiger, A., Gaiger A Fau - Gale, R.E., Gale Re Fau - Linch, D.C., Linch Dc Fau - Littlewood, T.J., Littlewood Tj Fau - Moss, P.A., Moss Pa Fau - Wainscoat, J.S., et al. (1996). Amplification of mitochondrial DNA in acute myeloid leukaemia. Br J Haematol 95, 426-431. 5. Bourgeois, J.M., and Tarnopolsky, M.A. (2004). Pathology of skeletal muscle in mitochondrial disorders. Mitochondrion 4, 441-452. 6. Carling, P.J., Cree, L.M., and Chinnery, P.F. (2011). The implications of mitochondrial DNA copy number regulation during embryogenesis. Mitochondrion 11, 686-692. 7. Chen, C.H., and Cheng, Y.C. (1992). The role of cytoplasmic deoxycytidine kinase in the mitochondrial effects of the anti-human immunodeficiency virus compound, 2',3'- dideoxycytidine. J Biol Chem 267, 2856-2859. 8. Copeland, W.C. (2012). Defects in mitochondrial DNA replication and human disease. Critical reviews in biochemistry and molecular biology 47, 64-74. 9. Dalakas, M.C., Semino-Mora, C., and Leon-Monzon, M. (2001). Mitochondrial alterations with mitochondrial DNA depletion in the nerves of AIDS patients with peripheral neuropathy induced by 2'3'-dideoxycytidine (ddC). Laboratory investigation; a journal of technical methods and pathology 81, 1537-1544. 10. 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. 11. Estey, E., Plunkett, W., Dixon, D., Keating, M., McCredie, K., and Freireich, E.J. (1987). Variables predicting response to high dose cytosine arabinoside therapy in patients with refractory acute leukemia. Leukemia 1, 580-583. 12. Ferraro, P., Nicolosi, L., Bernardi, P., Reichard, P., and Bianchi, V. (2006). Mitochondrial deoxynucleotide pool sizes in mouse liver and evidence for a transport mechanism for thymidine monophosphate. Proc Natl Acad Sci U S A 103, 18586-18591. 13. Fischl, M.A., Stanley, K., Collier, A.C., Arduino, J.M., Stein, D.S., Feinberg, J.E., Allan, J.D., Goldsmith, J.C., and Powderly, W.G. (1995). Combination and monotherapy with zidovudine and zalcitabine in patients with advanced HIV disease. The NIAID AIDS Clinical Trials Group. Annals of internal medicine 122, 24-32. 14. Fontanesi, F., Soto, I.C., and Barrientos, A. (2008). Cytochrome c oxidase biogenesis: new levels of regulation. IUBMB Life 60, 557-568. 15. 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.

130

16. Gattermann, N., and Aul, C. (1997). Mitochondrial DNA amplification in AML: a hypothesis. Br J Haematol 97, 242. 17. Govindarajan, R., Leung, G.P., Zhou, M., Tse, C.M., Wang, J., and Unadkat, J.D. (2009). Facilitated mitochondrial import of antiviral and anticancer nucleoside drugs by human equilibrative nucleoside transporter-3. American journal of physiology Gastrointestinal and liver physiology 296, G910-922. 18. Graziewicz, M.A., Longley, M.J., and Copeland, W.C. (2006). DNA polymerase gamma in mitochondrial DNA replication and repair. Chemical reviews 106, 383-405. 19. Haferlach, T., Kohlmann, A., Wieczorek, L., Basso, G., Kronnie, G.T., Bene, M.C., De Vos, J., Hernandez, J.M., Hofmann, W.K., Mills, K.I., et al. (2010). Clinical utility of microarray- based gene expression profiling in the diagnosis and subclassification of leukemia: report from the International Microarray Innovations in Leukemia Study Group. J Clin Oncol 28, 2529- 2537. 20. Herold, N.A.-O.h.o.o., Rudd, S.G., Ljungblad, L., Sanjiv, K., Myrberg, I.H., Paulin, C.B., Heshmati, Y., Hagenkort, A., Kutzner, J., Page, B.D., et al. (2017). Targeting SAMHD1 with the Vpx protein to improve cytarabine therapy for hematological malignancies. Nat Med. 21. Ikeda, M., Ide, T., Fujino, T., Arai, S., Saku, K., Kakino, T., Tyynismaa, H., Yamasaki, T., Yamada, K., Kang, D., et al. (2015). Overexpression of TFAM or twinkle increases mtDNA copy number and facilitates cardioprotection associated with limited mitochondrial oxidative stress. PLoS One 10, e0119687. 22. Kakuda, T.N. (2000). Pharmacology of nucleoside and nucleotide reverse transcriptase inhibitor-induced mitochondrial toxicity. Clin Ther 22, 685-708. 23. Kallianpur, A.R., and Hulgan, T. (2009). Pharmacogenetics of nucleoside reverse-transcriptase inhibitor-associated peripheral neuropathy. Pharmacogenomics 10, 623-637. 24. 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. 25. Lam, W., Chen, C., Ruan, S., Leung, C.H., and Cheng, Y.C. (2005). Expression of deoxynucleotide carrier is not associated with the mitochondrial DNA depletion caused by anti-HIV dideoxynucleoside analogs and mitochondrial dNTP uptake. Mol Pharmacol 67, 408- 416. 26. Lane, A.N., and Fan, T.W. (2015). Regulation of mammalian nucleotide metabolism and biosynthesis. Nucleic Acids Res 43, 2466-2485. 27. Lewis, W., Day, B.J., and Copeland, W.C. (2003). Mitochondrial toxicity of NRTI antiviral drugs: an integrated cellular perspective. Nat Rev Drug Discov 2, 812-822. 28. Liou, J.Y., Dutschman, G.E., Lam, W., Jiang, Z., and Cheng, Y.C. (2002). Characterization of human UMP/CMP kinase and its phosphorylation of D- and L-form deoxycytidine analogue monophosphates. Cancer research 62, 1624-1631. 29. Martin, J.L., Brown, C.E., Matthews-Davis, N., and Reardon, J.E. (1994). Effects of antiviral nucleoside analogs on human DNA polymerases and mitochondrial DNA synthesis. Antimicrobial agents and chemotherapy 38, 2743-2749. 30. Mathews, C.K. (2015). Deoxyribonucleotide metabolism, mutagenesis and cancer. Nat Rev Cancer 15, 528-539. 31. Nunnari, J., and Suomalainen, A. (2012). Mitochondria: in sickness and in health. Cell 148, 1145-1159.

131

32. Panuwet, P., Hunter, R.E., D’Souza, P.E., Chen, X., Radford, S.A., Cohen, J.R., Marder, M.E., Kartavenka, K., Ryan, P.B., and Barr, D.B. (2016). Biological Matrix Effects in Quantitative Tandem Mass Spectrometry-Based Analytical Methods: Advancing Biomonitoring. Critical reviews in analytical chemistry / CRC 46, 93-105. 33. Parker, W.B., Shaddix Sc Fau - Chang, C.H., Chang Ch Fau - White, E.L., White El Fau - Rose, L.M., Rose Lm Fau - Brockman, R.W., Brockman Rw Fau - Shortnacy, A.T., Shortnacy At Fau - Montgomery, J.A., Montgomery Ja Fau - Secrist, J.A., 3rd, Secrist Ja 3rd Fau - Bennett, L.L., Jr., and Bennett, L.L., Jr. (1991). Effects of 2-chloro-9-(2-deoxy-2-fluoro-beta- D-arabinofuranosyl)adenine on K562 cellular metabolism and the inhibition of human ribonucleotide reductase and DNA polymerases by its 5'-triphosphate. Cancer Res 9 2386- 2394. 34. Pearce, D.J., Taussig, D., Zibara, K., Smith, L.-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. 35. Ray, A.S. (2005). Intracellular interactions between nucleos(t)ide inhibitors of HIV reverse transcriptase. AIDS reviews 7, 113-125. 36. Ross, D.D., Chen, S.R., and Cuddy, D.P. (1990). Effects of 1-beta-D-arabinofuranosylcytosine on DNA replication intermediates monitored by pH-step alkaline elution. Cancer research 50, 2658-2666. 37. Setzer, B., Schlesier, M., Thomas, A.K., and Walker, U.A. (2005). Mitochondrial toxicity of nucleoside analogues in primary human lymphocytes. Antiviral therapy 10, 327-334. 38. Skrtic, 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. 39. Sriskanthadevan, S., Jeyaraju, D.V., Chung, T.E., Prabha, S., Xu, W., Skrtic, M., Jhas, B., Hurren, R., Gronda, M., Wang, X., et al. (2015a). AML cells have low spare reserve capacity in their respiratory chain that renders them susceptible to oxidative metabolic stress. Blood 125, 2120-2130. 40. Sriskanthadevan, S., Jeyaraju, D.V., Chung, T.E., Prabha, S., Xu, W., Skrtic, M., Jhas, B., Hurren, R., Gronda, M., Wang, X., et al. (2015b). AML cells have low spare reserve capacity in their respiratory chain that renders them susceptible to oxidative metabolic stress. Blood 125, 2120-2130. 41. Tufi, R., Gandhi, S., de Castro, I.P., Lehmann, S., Angelova, P.R., Dinsdale, D., Deas, E., Plun- Favreau, H., Nicotera, P., Abramov, A.Y., et al. (2014). Enhancing nucleotide metabolism protects against mitochondrial dysfunction and neurodegeneration in a PINK1 model of Parkinson's disease. Nat Cell Biol 16, 157-166. 42. van Leth, F., Phanuphak, P., Ruxrungtham, K., Baraldi, E., Miller, S., Gazzard, B., Cahn, P., Lalloo, U.G., van der Westhuizen, I.P., Malan, D.R., et al. (2004). Comparison of first-line antiretroviral therapy with regimens including nevirapine, efavirenz, or both drugs, plus stavudine and lamivudine: a randomised open-label trial, the 2NN Study. Lancet (London, England) 363, 1253-1263. 43. Wallace, D.C., and Fan, W. (2009). The pathophysiology of mitochondrial disease as modeled in the mouse. Genes Dev 23, 1714-1736. 44. Wang, L. (2010). Deoxynucleoside salvage enzymes and tissue specific mitochondrial DNA depletion. Nucleosides Nucleotides Nucleic Acids 29, 370-381.

132

45. Zittoun, R., Marie, J.P., Delanian, S., Suberville, A.M., and Thevenin, D. (1987). Prognostic value of in vitro uptake and retention of cytosine arabinoside in acute myelogenous leukemia. Seminars in oncology 14, 269-275.

133

5 Characterizing the interactome

of POLG by BioID

134

5. Characterizing the interactome of POLG

Contributions

The majority of work presented in this chapter was performed by the author of the thesis.

Technical assistance was provided as follows: qRT-PCR experiments were performed with technical assistance by Rose Hurren

BioID experiments were performed and analyzed with technical assistance by Etienne Coyaud and

Estelle Laurent (Princess Margaret Cancer Centre, University of Toronto, Ontario, Canada), under the supervision of Dr. Brian Raught

Overexpression and immunofluorescence studies of Ruvbl2 were performed by Stuart R. Wood and Lawrence Kazak (Medical Research Council, Mitochondrial Biology Unit, Cambridge, UK), under the supervision of Dr. Ian J. Holt

Note: Sections of this work has been published with Mitochondrion (Liyanage SU, Coyaud E,

Laurent EM, Hurren R, Maclean N, Wood SR, Kazak L, Shamas-Din A, Holt I, Raught B,

Schimmer A. Characterizing the mitochondrial DNA polymerase gamma interactome by BioID identifies Ruvbl2 localizes to the mitochondria. Mitochondrion. 2017 ;32:31-35)

135

5.1 Introduction

Human mitochondrial DNA is a 16.6kb genome which encodes genes essential for oxidative metabolism(Chan and Copeland, 2009). It is replicated by the nuclear-encoded mitochondrial

DNA polymerase gamma (POLG). Human POLG consists of a C-terminal catalytic polymerase domain and an N-terminal exonuclease domain separated by a linker region. The holoenzyme consists of the primary subunit POLG and a homodimeric form of its accessory subunit

POLG2(Graziewicz et al., 2006). POLG is located within poorly defined multi-protein-DNA complexes termed nucleoids, a crucial hub conducive to packaging and supporting mtDNA maintenance, transcription and translation(Bogenhagen, 2012).

We reasoned that a novel proximity-based biotinylation (BioID) established for identification of protein-protein interactions can be used to identify low abundant, transient interactors of POLG.

To identify protein-protein interactions with BioID, the protein of interest is fused with a mutant

E. coli biotin conjugating enzyme (BirA R118G, or BirA*) that promiscuously biotinylates nearby proteins, Following cell lysis, biotinylated proteins are affinity purified and identified using mass spectrometry(Roux et al., 2012). Currently, the interactome profile of POLG is not characterized.

Here, we report novel protein-protein interactions of POLG using the BioID method.

136

5.2 Results

5.2.1 POLG interacts with proteins which support mitochondrial metabolism and biogenesis

We investigated protein-protein interactions of POLG using BioID. Expression of the fusion

POLG-FlagBirA* following tetracycline induction in HEK 293 T-REx Flp-In cells was confirmed by immunoblotting (Figure 5.1A). POLG BioID interactome mapping yielded 39 high confidence interaction candidates over negative control ornithine transcarbamylase (OTC), a mitochondrial matrix localized protein. Mitochondria localized interactors comprised 66% (26 out of 39) of the identified interacting proteins (Table 2.1, Figure 5.1B). Of the mitochondria localized hits, 42.3%

(11/26) represented reported nucleoid-associated proteins involved in mtDNA maintenance and the mitochondrial ribosome(Hensen et al., 2014). Novel interaction candidates comprised 53.8%

(14/26) and functional analysis revealed enrichment for mitochondrial DNA replication, transcription, translation, protein processing and quality control, and metabolism processes.

137

Table 5.1: Identification of POLG interaction partners by BioID Reported POLG interaction partners and mt-nucleoid associated proteins are displayed in yellow.

POLG- BioID interactome map Total Spectral counts SAINT (n=4) Score Gene Gene Full name Localization POLG- Controls (Max ID FlagBirA* Score 1.0) 81570 ClpB homolog, mitochondria 57 0 1.00 mitochondrial AAA ATPase CLPB chaperonin 28973 mitochondrial ribosomal mitochondria 50 0 1.00 MRPS18B protein S18B 11232 polymerase (DNA) gamma mitochondria 45 2 1.00 POLG2 2, accessory subunit 2926 G-rich RNA sequence mitochondria 34 0 1.00 GRSF1 binding factor 1 9529 BCL2 associated athanogene mitochondria 31 4 1.00 BAG5 5 28977 mitochondrial ribosomal mitochondria 31 0 1.00 MRPL42 protein L42 8604 solute carrier family 25 mitochondria 26 3 1.00 SLC25A12 member 12 5165 pyruvate dehydrogenase mitochondria 21 3 1.00 PDK3 kinase 3 92667 mitochondrial genome mitochondria 64 21 0.99 MGME1 maintenance exonuclease 1 51081 mitochondrial ribosomal mitochondria 75 10 0.99 MRPS7 protein S7 3420 isocitrate dehydrogenase 3 mitochondria 25 1 0.99 IDH3B (NAD+) beta 9512 peptidase (mitochondrial mitochondria 18 8 0.97 PMPCB processing) beta 23078 von Willebrand factor A mitochondria 106 0 0.96 KIAA0564 domain containing 8 64965 mitochondrial ribosomal mitochondria 92 0 0.95 MRPS9 protein S9 65993 mitochondrial ribosomal mitochondria 57 4 0.78 MRPS34 protein S34 81892 SRA stem-loop interacting mitochondria 22 3 0.75 SLIRP RNA binding protein 65080 mitochondrial ribosomal mitochondria 20 0 0.75 MRPL44 protein L44

138

55178 RNA methyltransferase like mitochondria 15 22 0.75 RNMTL1 1 9238 transforming growth factor mitochondria 15 0 0.75 TBRG4 beta regulator 4 20020 IBA57 homolog, iron-sulfur mitochondria 13 331 0.74 5 IBA57 cluster assembly 55149 mitochondrial poly(A) mitochondria 11 29 0.74 MTPAP polymerase 51253 mitochondrial ribosomal mitochondria 15 19 0.74 MRPL37 protein L37 55210 ATPase family, AAA mitochondria 92 11 0.73 ATAD3A domain containing 3A 5442 polymerase (RNA) mitochondria 30 0 0.73 POLRMT mitochondrial 211 5'-aminolevulinate synthase mitochondria 13 4 0.72 ALAS1 1 51021 mitochondrial ribosomal mitochondria 8 4 0.71 MRPS16 protein S16 8607 RUVBL1 RuvB like AAA ATPase 1 cytoplasm, 2180 360 1.00 nucleus 10856 RUVBL2 RuvB like AAA ATPase 2 cytoplasm, 1903 331 1.00 nucleus 9532 BAG2 BCL2 associated athanogene cytoplasm 101 22 1.00 2 10910 SUGT1 SGT1 homolog, MIS12 cytoplasm, 119 29 1.00 kinetochore complex nucleus assembly cochaperone 23386 NUDCD3 NudC domain containing 3 Golgi 38 0 1.00 5211 PFKL phosphofructokinase, liver cytoplasm 77 19 0.98 type 25911 DPCD deleted in primary ciliary nucleus 74 11 0.96 dyskinesia homolog (mouse) 22824 HSPA4L heat shock protein family A cytoplasm 13 0 0.94 (Hsp70) member 4 like 10294 DNAJA2 DnaJ heat shock protein cytoplasm 67 20 0.91 family (Hsp40) member A2 83481 EPPK1 epiplakin 1 cytoplasm 46 4 0.90 10644 IGF2BP2 insulin like growth factor 2 nucleus 21 4 0.81 mRNA binding protein 2 11047 ADRM1 adhesion regulating plasma 17 8 0.75 molecule 1 membrane 1488 CTBP2 C-terminal binding protein 2 nucleus 9 0 0.73

139

A -Tet + Tet POLG-FLAGBirA* FLAG

POLG-FLAGBirA* 180 kDa POLG 140 kDa POLG

Aconitase

Figure 5.1: POLG BioID interactome mapping reveals novel interaction candidates (A) Overexpression of POLG-FlagBirA* was induced by 24 hour tetracycline treatment in HEK 293 T-Rex Flp-In cells. Total proteins were extracted and probed for POLG and FLAG. Aconitase was used as a loading control. (B) POLG BioID hits were sorted based on functional categories and cellular localization as mitochondrial (left) or non-mitochondrial (right). Previously reported nucleoid associated proteins are displayed in circles and novel interaction candidates are displayed in rectangles

140

5.2.2. Ruvbl1 and Ruvbl2 are detected in mitochondria-enriched fractions

In addition to mitochondrial proteins, we identified 13 proteins not previously reported to localize to the mitochondria (Table 2.1, Figure 2.1B). The top hits based on total peptide count and SAINT score were Ruvbl2 and Ruvbl1(Grigoletto et al., 2011; Kanemaki et al., 1999). Ruvbl2 and Ruvbl1 are AAA+ ATPases which act as essential chaperones and oncoproteins in regulation of diverse cellular processes. These proteins can function independently or as a Ruvbl1/2 oligomers to assemble multi-protein complexes which regulate processes such as energy metabolism, chromatin remodelling, transcription and DNA damage repair and sensing (Fuchs and Steller, 2015). Given their ability to form diverse protein-protein interactions in multiple cellular sub-compartments and the high spectral counts determined by BioID, we investigated their potential mitochondrial localization.

Analysis of Cancer Cell line encyclopedia database indicated a preferential upregulation of Ruvbl2 in leukemia cell lines (Figure 5.2A). Hence, we assessed whether Ruvbl2/1 are present in the mitochondria of leukemic OCI-AML2 cells. Ruvbl2 and to a lesser extent, Ruvbl1 were detected in mitochondria-enriched fractions generated by Percoll gradient fractionation in OCI-AML2, further supporting mitochondrial localization of these proteins (Figure 5.2B).

141

Mitochondrial enrichment Whole cell B extracts

Ruvbl2

Ruvbl1

VDAC - mitochondria

POLG- mitochondria

LAMP1- lysosomes

Calnexin-ER

β-Tubulin- cytosol

Figure 5.2: Ruvbl2 is detected in mitochondria-enriched fractions of leukemic cells (A) Analysis of Ruvbl2 mRNA expression in Cancer cell line encyclopedia (B) Whole cell extracts (WCE) and mitochondria-enriched fractions were isolated from OCI- AML2 cells and probed for Ruvbl2. Lane 1) 10 µg WCE, 2) 5 µg WCE, 3) 10 µg crude mitochondria, 4)10 µg of crude mitochondria further purified by 30% Percoll gradient fractionation, 5) 10 µg of crude mitochondria further purified by 3 step Percoll gradient (50, 22 and 15%). VDAC, POLG, LAMP1, Calnexin and β-tubulin were probed to indicate level of purification.

142

5.2.3. Ruvbl2 has a potential mitochondrial isoform generated by alternative translation initiation.

To further interrogate Ruvbl2 localization to the mitochondria, we assessed whether Ruvbl2 contains a mitochondria targeting signal (MTS) using prediction software. Of four algorithms tested (Mitoprot, TargetP, PSORTII, iPSORT,), only the least stringent suggested that the first portion of the protein (starting from the first methionine, M1) might function as a mitochondrial targeting signal (MTS). However, a number of mitochondrial isoforms are produced by means of downstream alternative translation initiation (ATI)(Kazak et al., 2013), and in the case of human and murine Ruvbl2, the second downstream methionine, M46, was recognized as a potential MTS by all four programs (Figure 5.3A). Therefore, variants of human Ruvbl2 with alanine in place of

M1 or M46 were cloned and expressed in human 143B osteoscarcoma (HOS) cells. Ruvbl2.M46A was distributed throughout the cytoplasm, In contrast, Ruvbl2.M1A co-localized with the mitochondria (Figure 5.3B). Although the cDNA overexpression results demonstrate that a truncated Ruvbl2 protein starting from M46 is produced in living cells, an alternatively spliced form of Ruvbl2, 45 residues shorter than isoform 1, has been reported as isoform 3 (Accession:

NP_001308120.1). Hence there are two potential mechanisms, alternative splicing and ATI, of generating the Ruvbl2 isoform that is capable of being targeted to mitochondria.

143

Figure 5.3: Alternative translation initiation generates a mitochondrial isoform of RUVBL2 in human cultured cells (A)In silico mitochondrial targeting prediction scores of human and mouse RUVBL2 for Mitoprot, TargetP, PSORTII, and iPSORT. “M” followed by a number, indicate the methionine residue location; red-box indicates strong prediction for mitochondrial targeting. (B) Full-length RUVBL2 cDNAs with C-terminal HA-tag were overexpressed in HOS cells, with a methionine(M) to alanine (A) mutation at the putative translation start codons: RUVBL2M46A.HA and RUVBL2M1A.HA. Cells were labeled with nuclei (blue, DAPI stain), mitochondria (red, Mitotracker stain) and anti-HA antibody (green).

144

5.3 Discussion

In this chapter, we report the first screen of the POLG interactome by BioID, yielding known and novel candidates which promote mitochondrial biogenesis. Known interactors consisted of its bona fide interactor POLG2 and proteins previously reported to localized to mtDNA nucleoids by mass spectrometric approaches.(Hensen et al., 2014) These include factors associated with mtDNA, mtRNA and ribosome metabolism such as ATAD3A, POLRMT and MTPAP.(He et al., 2012b)

We also identified novel interaction partners with POLG including genes associated with RNA granules (SLIRP, GRSF1, RNMTL1), a compartment where newly synthesized mRNA undergo processing, storage, sorting or translation. Additional interactome candidates included several mitochondrial ribosome proteins, suggesting a close proximity between the mitochondrial ribosome and mtDNA replication events. This is in line with the functional role of mtDNA nucleoids as sites of mitochondrial ribosome assembly.(Bogenhagen et al., 2014; He et al., 2012a)

Additionally, we detected several novel mitochondrial protein processing chaperones such as

ClpB, which may play roles in processing and maintenance of the DNA-protein nucleoid complex.

Hence, the identification of known POLG interaction partners and proteins located in the vicinity of mtDNA validate the sensitivity of the BioID technique in revealing known interaction partners.

The POLG BioID interactome map also revealed proteins not previously known to localize to the mitochondria, including Ruvbl2 and Ruvbl1. Ruvbl1 and Ruvbl2 are AAA+ ATPases, essential chaperones and oncoproteins involved in regulation of diverse cellular processes.(Kanemaki et al.,

1999; Kanemaki et al., 1997) These proteins can function independently or as a Ruvbl1/2 oligomers to assemble multi-protein complexes which regulate processes such as energy metabolism, chromatin remodelling, transcription and DNA damage repair and sensing. They also

145 functionally interact with numerous transcription factors important for carcinogenesis, including c-myc and β-catenin.(Rosenbaum et al., 2013) Given their ability to form diverse protein-protein interactions in multiple cellular sub-compartments, we reasoned these chaperones are top candidates for further investigation in the mitochondrial proteome.

We confirmed the presence of Ruvbl2 in enriched mitochondrial fractions by immunoblotting. We recognize however, that the interaction of Ruvbl2 with POLG may reflect a transient interaction during the translation of POLG in the cytoplasm prior to mitochondrial import. Alternatively, it is possible that Ruvbl2/1 may have interacted with POLG in the cytoplasm under conditions of protein aggregation when overexpressed.(Zaarur et al., 2015) However, we also detected Ruvbl2 mitochondrial localization by immunofluorescence when Ruvbl2 was synthesized at the alternative translation start site in HOS cells. Given that the putative alternative mitochondrial isoform of Ruvbl2 is 45 amino acids shorter than native Ruvbl2, it is possible that isoform 3 was undetectable by immunoblotting due to differences in post-translational modifications which affect protein size.

Thus, further mechanistic studies are required to elucidate its putative function and targeting within the mitochondria. Knockdown of Ruvbl2 in leukemic cells resulted in cell death, thus mechanistic studies of Ruvbl2 mitochondrial function through this approach remained limited. In summary, characterizing the POLG interactome identified novel candidates which support mitochondrial metabolism, including a novel mitochondrial localization role for Ruvbl2.

146

5.4 References

1. Chan SS, Copeland WC. DNA polymerase gamma and mitochondrial disease: understanding the consequence of POLG mutations. Biochim Biophys Acta. 2009;1787(5):312- 319. 2. Graziewicz MA, Longley MJ, Copeland WC. DNA polymerase gamma in mitochondrial DNA replication and repair. Chemical reviews. 2006;106(2):383-405. 3. Bogenhagen DF. Mitochondrial DNA nucleoid structure. Biochimica et Biophysica Acta (BBA)-Gene Regulatory Mechanisms. 2012;1819(9):914-920. 4. Roux KJ, Kim DI, Raida M, Burke B. A promiscuous biotin ligase fusion protein identifies proximal and interacting proteins in mammalian cells. J Cell Biol. 2012;196(6):801-810. 5. Coyaud E, Mis M, Laurent EM, et al. BioID-based Identification of Skp Cullin F-box (SCF)beta-TrCP1/2 E3 Ligase Substrates. Mol Cell Proteomics. 2015;14(7):1781-1795. 6. Jeyaraju DV, Cisbani G, De Brito OM, Koonin EV, Pellegrini L. Hax1 lacks BH modules and is peripherally associated to heavy membranes: implications for Omi/HtrA2 and PARL activity in the regulation of mitochondrial stress and apoptosis. Cell Death Differ. 2009;16(12):1622-1629. 7. Kazak L, Reyes A, Duncan AL, et al. Alternative translation initiation augments the human mitochondrial proteome. Nucleic Acids Res. 2013;41(4):2354-2369. 8. Hensen F, Cansiz S, Gerhold JM, Spelbrink JN. To be or not to be a nucleoid protein: a comparison of mass-spectrometry based approaches in the identification of potential mtDNA- nucleoid associated proteins. Biochimie. 2014;100:219-226. 9. Kanemaki M, Kurokawa Y, Matsu-ura T, et al. TIP49b, a new RuvB-like DNA helicase, is included in a complex together with another RuvB-like DNA helicase, TIP49a. The Journal of biological chemistry. 1999;274(32):22437-22444. 10. Grigoletto A, Lestienne P, Rosenbaum J. The multifaceted proteins Reptin and Pontin as major players in cancer. Biochim Biophys Acta. 2011;1815(2):147-157. 11. Fuchs Y, Steller H. Live to die another way: modes of programmed cell death and the signals emanating from dying cells. Nat Rev Mol Cell Biol. 2015;16(6):329-344. 12. He J, Cooper HM, Reyes A, et al. Mitochondrial nucleoid interacting proteins support mitochondrial protein synthesis. Nucleic Acids Res. 2012;40(13):6109-6121. 13. Bogenhagen DF, Martin DW, Koller A. Initial steps in RNA processing and ribosome assembly occur at mitochondrial DNA nucleoids. Cell Metab. 2014;19(4):618-629. 14. He J, Cooper HM, Reyes A, et al. Human C4orf14 interacts with the mitochondrial nucleoid and is involved in the biogenesis of the small mitochondrial ribosomal subunit. Nucleic Acids Res. 2012;40(13):6097-6108. 15. Kanemaki M, Makino Y, Yoshida T, et al. Molecular Cloning of a Rat 49-kDa TBP- Interacting Protein (TIP49) That Is Highly Homologous to the Bacterial RuvB. Biochemical and Biophysical Research Communications. 1997;235(1):64-68. 16. Zaarur N, Xu X, Lestienne P, et al. RuvbL1 and RuvbL2 enhance aggresome formation and disaggregate amyloid fibrils. EMBO J. 2015;34(18):2363-2382.

147

6. FUTURE DIRECTIONS

In chapter 3, we leveraged the increased activity of cytoplasmic nucleoside kinase pathway to convert ddC to its active form, inhibit mtDNA replication and target leukemic cells in AML in vitro and in vivo. Although the preclinical efficacy of ddC as an anti-leukemic agent is promising, activity of ddC in vitro is in the low micromolar range. Thus identifying new compounds with greater potency will potentially improve their efficacy as a potential treatment for a subset of AML patients.

To develop novel POLG inhibitors, we have partnered with Medivir, a Swedish pharmaceutical company with expertise in anti-virals and nucleoside analog compounds. Through this collaboration, we aim to identify and validate novel POLG inhibitors with a greater potency compared to ddC. Currently, lead compounds are being assessed for their effects on inhibition of mtDNA replication, mitochondrial metabolism and cell viability in multiple cell culture and animal models of AML. If preclinical studies display a favorable efficacy and safety profile, the project will progress towards assessing the efficacy of novel POLG inhibitors in AML through clinical trials.

Given the potential of prodrugs such as ddC as a therapeutic strategy, identifying predictive biomarkers of response to ddC will allow for stratification of the patient population and a personalized treatment approach. We hypothesize that subsets of AML with upregulated cytoplasmic nucleoside kinase activity and a greater dependence on oxidative phosphorylation are most susceptible to ddC treatment. Currently, we have attempted to identify biomarkers of sensitivity to ddC using mRNA expression of individual cytoplasmic nucleoside kinase genes

148 which activate ddC such as DCK, CMPK1, NME. However, expression of these genes individually failed to predict sensitivity to ddC in vitro in AML patient samples. This maybe attributed to limited sample size and the complexity of the drug metabolism process. It is likely that expression of a panel of genes act as a better predictor of ddC response. A high throughput approach such as microarray or RNA-sequencing using a large sample size of AML patient samples may elucidate potential biomarkers of ddC sensitivity. Additionally, gene knockdown or knockout screens may provide novel target genes which regulate ddC activity, implicated in drug transport, conversion and metabolism. Lastly, it is possible that substrate competition by dCTP contributes to ddC efficacy. Inhibitors of de novo biosynthesis pathways may diminish levels of native dCTP, thus enhancing anti-leukemic activity. Alternatively, diminishing dCTP levels may act to reduce cell viability due to metabolic feedback checkpoints, and slow rates of mitochondrial biogenesis.

To further identify genes implicated in ddC metabolism, genes implicated in metabolism of similar prodrugs, such as AraC can be explored. A recent study has shown that levels of a dNTP triphosphohydrolase gene, SAMHD1, inversely correlates with response to AraC in vitro and in patients(Herold et al., 2017). SAMHD1 may be investigated as a biomarker candidate for sensitivity to ddC. Additionally, studies have shown that levels of the activated form of AraC,

AraCTP in AML blasts correlate with cytotoxicity in vitro and clinical response in patients (Estey et al., 1987; Zittoun et al., 1987). Developing a functional assay to assess nucleoside kinase pathway activity may predict levels of pro-drug activation, which correlates with drug sensitivity.

Upon validation of these studies, predictive biomarkers can be used to stratify patients as responders or non-responders to tailor treatment options.

149

Currently, the rate-limiting step in ddC metabolism in AML cells is unknown. Levels of initial transport of the ddC nucleoside across the plasma membrane to the cytoplasm was assessed in our study. Levels of total intracellular ddC were similar between primary AML cells and normal hematopoietic progenitors, but significantly different between AML and HEK cell lines. Thus, rates of intracellular ddC transport may be cell line dependent. Additional work is needed to determine whether plasma membrane nucleoside transporter expression/activity correlates with sensitivity to ddC.

Secondly, the rate limiting kinases are not well understood for ddC. For a majority of nucleoside analogs, DCK performs the primary rate-limiting initial phosphorylation step. As CMPK1 and

NME2 is upregulated in AML, it is possible that DCK monophosphorylation is a rate limiting activation step, combined with initial plama membrane intracellular transport.

The final transport of ddCTP from the cytoplasm to the mitochondria is carried out by mitochondrial nucleoside transporters. Currently, the specific mitochondrial transporter for ddCTP is unknown. It is intriguing that 3 of 4 known mitochondrial transporters are upregulated in AML cells. Additionally, these transporters correlate with mtDNA biosynthesis expression Further characterization of the role of mitochondrial nucleoside transporters in mediating ddC efficacy is required. Studies include overexpression or knockdown of mitochondrial nucleoside transporters in the presence of ddC.

Additionally, research on the role of ddC catabolism, e.g. by dephosphorylases and deaminases, is warranted as they can confer resistance to nucleoside analogs through drug inactivation. In the plasma, ddC displays high bioavailability, and 75% remains unexchanged as ddC, indicating very low rates of deamination to dideoxyuridine (ddU). Additional studies are needed to assess rates of

150 ddC intracellular deamination. LC-MS assessment of ddC, ddCTP and its catabolic products may provide insight into levels of ddC inactivation.

In our study, we assessed levels of ddC and ddCTP in primary AML and normal samples by

LCMS. A limitation of our mass spectrometry assay is that intracellular levels of ddC and ddCTP can be compared between samples within an experiment, but was not designed for comparisons between experiments. As we did not spike into each sample isotopically labeled (13C or 2H) internal standards for ddC and ddCTP, levels of ddC and ddCTP should be considered semi- quantitative but not absolute. Additionally, matrix effects can contribute to differences in metabolite quantification between AML and normal samples. The matrix consists of all components of a sample except for the analyte investigated. A matrix effect may occur when a cellular component co-elutes with the analyte of interest, altering the ionization efficiency compared to the analyte eluting in the absence of the matrix. (Panuwet et al., 2016). Future work should focus on further validating the conversion of ddC to ddCTP between AML and normal samples using internal standard controls.

Our study also demonstrated a role for cytoplasmic nucleoside kinases in mitochondrial nucleotide metabolism. We observed regulation of mtDNA levels, likely through production of nucleotides within in the cytoplasm, followed by import into the mitochondria. Currently, mitochondrial nucleotide metabolism remains a poorly investigated area in AML. To further explore the cross- talk between mitochondrial and cytoplasmic nucleotide metabolism, knockdown of key enzymes in the nucleotide salvage cascade in AML cell lines, followed by quantitation of nucleotide pools

151 in cytoplasmic and mitochondrial fractions can be assessed. Additionally, measuring nucleotide flux between mitochondrial and cytoplasmic components through nucleotide radiolabelling approaches may provide additional insights into mechanisms of nucleotide metabolism in AML.

Our studies suggest that cross talk between cytoplasmic nucleoside kinases and plasma membrane nucleotide transporters occur. We observed DCK and CMPK1 knockdown (data not shown) decreased levels of intracellular ddC in some leukemic cell lines, suggesting that drug activation and drug import undergo a feedback process in a subset of AML cells. Assessing whether AML cells with higher expression of plasma membrane nucleotide transporters correlate with drug activation and sensitivity may provide additional candidates for predictive biomarkers.

Given that cytoplasmic nucleotide metabolism is upregulated in a subset of AML, targeting nucleotide salvage metabolism can be explored as a novel therapeutic strategy for AML. Currently, the functional importance of nucleotide salvage kinases on disease progression is poorly explored.

We hypothesize that if AML cells are more dependent on nucleotide salvage to produce nucleotides compared to de novo biosynthesis, inhibiting this pathway may preferentially deplete cytoplasmic and mitochondrial nucleotide pools in AML cells causing anti-leukemic effects. To identify genes important for AML cell viability, genetic knockdown or knockout screens of nucleotide salvage pathway kinases may provide novel targets genes to further characterize in preclinical studies.

Our study has the potential to be extended towards understanding response to existing AML therapies. Novel prodrugs approved for AML, such as azacytidine and sapacitabine, despite

152 targeting diverse functional pathways such as nuclear DNA replication or epigenetic modifiers, require activation by the same nucleoside kinase pathway. We speculate that upregulation of the nucleoside kinase pathway activity, including the upregulation of CMPK1, may contribute to the activation and efficacy of these drugs in patients. It will be interesting to retrospectively test whether patients which responded favorably to nucleoside analog therapy correlate with increased levels of activated prodrug, and nucleoside kinase pathway activity. If correlations between levels of activated prodrug and nucleoside kinase pathway activity are observed, this pathway can be further explored as a biomarker towards patient selection for existing AML therapies.

Lastly, our preliminary data links mitochondrial metabolism with AML differentiation. We observed that ddC increases expression of differentiation markers in AML cell lines(data not shown). These data suggests cross-talk between mitochondrial and nuclear fractions to regulate cellular differentiation. We hypothesize that inhibition of oxidative phosphorylation alters levels of TCA cycle metabolites, leading to a deregulation of epigenetic enzymes whose activity is dependent on TCA cycle substrates. To further explore this mechanism of cross-talk between mitochondrial and nuclear components, metabolomic analysis of TCA cycle intermediates within mitochondrial and nuclear compartments can be performed. If correlations between metabolite levels in mitochondrial and nuclear components occur, the enzymes regulated by specific TCA cycle metabolites can be further characterized. Additionally, Chip-Seq and RNA-Seq experiments may identify differentially expressed genes which regulate stem cell and differentiation states in response to ddC treatment in AML cells.

153

The link between mitochondrial DNA and differentiation can be further explored by evaluating factors which shuttle between the mitochondria and the nucleus. In chapter 5, we identified novel interactome partners of POLG, including Ruvbl2, a nuclear chaperone identified to localize to the mitochondria. It is possible that Ruvbl2 plays a role in mediating cross-talk between mitochondrial and nuclear fractions.

Currently, it is unknown whether Ruvbl2 and POLG are direct or transient interactors. Within the mitochondria, we hypothesize that Ruvbl2 acts as a scaffolding protein to assemble protein complexes involved in mitochondrial nucleoids and metabolism, similar to its role in other cellular compartments. Thus, characterizing the interactome partners of Ruvbl2 isoform 3 may reveal a novel functional role for Ruvbl2 in the mitochondria. Additional approaches would include knockdown studies of Ruvbl2 with or without overexpression of Ruvbl2 isoform 3 to assess the functional effects on mitochondria. We hypothesize that overexpression of mitochondrial Ruvbl2 relative to wild type Ruvbl2 may preferentially affect mitochondrial function. Alternatively, assessing the localization patterns of POLG interactome candidates in response to POLG knockdown, or ddC treatment, may elucidate candidates which can act as messengers between cellular fractions.

Overall, our study suggests that targeting nucleotide metabolism has far reaching implications on multiple cellular activities, such as mitochondrial function and cellular differentiation. Exploring these unique and diverse pathways will provide novel insights into mitochondrial function and its impact on AML progression and treatment.

154

6.2 References

1. Estey, E., Plunkett, W., Dixon, D., Keating, M., McCredie, K., and Freireich, E.J. (1987). Variables predicting response to high dose cytosine arabinoside therapy in patients with refractory acute leukemia. Leukemia 1, 580-583. 2. Herold, N.A.-O.h.o.o., Rudd, S.G., Ljungblad, L., Sanjiv, K., Myrberg, I.H., Paulin, C.B., Heshmati, Y., Hagenkort, A., Kutzner, J., Page, B.D., et al. (2017). Targeting SAMHD1 with the Vpx protein to improve cytarabine therapy for hematological malignancies. Nat Med. 3. Panuwet, P., Hunter, R.E., D’Souza, P.E., Chen, X., Radford, S.A., Cohen, J.R., Marder, M.E., Kartavenka, K., Ryan, P.B., and Barr, D.B. (2016). Biological Matrix Effects in Quantitative Tandem Mass Spectrometry-Based Analytical Methods: Advancing Biomonitoring. Critical reviews in analytical chemistry / CRC 46, 93-105. 4. Zittoun, R., Marie, J.P., Delanian, S., Suberville, A.M., and Thevenin, D. (1987). Prognostic value of in vitro uptake and retention of cytosine arabinoside in acute myelogenous leukemia. Seminars in oncology 14, 269-275.

155

List of copyright permissions

Figure 1

156

Figure 2

157

Figure 3

158

Figure 4

159

Figure 5

160

161

162