Investigation of Activated Tyrosine Kinases in Myeloproliferative Neoplasms

Investigation of activated tyrosine kinases in myeloproliferative neoplasms

Michael Ross Marit

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

Investigation of activated tyrosine kinases in myeloproliferative neoplasms

Michael Ross Marit Doctor of Philosophy Graduate Department of Medical Biophysics University of Toronto 2012

Myeloproliferative neoplasms (MPNs) are a group of disorders characterized by an excess production of a speciﬁc, fully functional blood cell type. Many cases involve deregulation of a protein tyrosine kinase. JAK2 is one such kinase, involved in a subset of MPNs. JAK2-selective inhibitors are currently being evaluated in clinical trials. In order to identify inhibitor-resistant JAK2 mutations before they appear in the clinic, we utilized TEL-JAK2 to conduct an in vitro random mutagenesis screen for JAK2 alleles

resistant to JAK Inhibitor-I. Isolated mutations were evaluated for their ability to sustain

cellular growth, stimulate downstream signalling pathways, and phosphorylate a novel

JAK2 substrate in the presence of inhibitor. When testing the panel of mutations in

the context of the Jak2 V617F allele, we observed that a subset of mutations conferred resistance to inhibitor. These results demonstrate that small-molecule inhibitors select for

JAK2 inhibitor-resistant alleles. Chronic myeloid leukemia is an MPN characterized by the presence of the BCR-ABL fusion gene. We determined that a speciﬁc cohort bearing deletions near the ABL gene, which is associated with poor prognosis, do not suﬀer from genomic instability. We also examined the role of a putative tumour suppressor gene

EXOSC2 as an explanation for the reduced survival time, and suggest it may have a role in disease progression.

ii Contents

Abstract ii

List of Tables v

List of Figures vi

List of Abbreviations viii

1 Introduction 1

1.1 Hematopoiesis ...... 1

1.1.1 Epo signalling and erythropoiesis ...... 3

1.1.2 The tyrosine kinase domain and JAK2 ...... 3

1.1.3 JAK-STAT signalling pathway ...... 4

1.1.3.1 Stat activation ...... 5

1.1.3.2 Mitogen-activated protein kinase activation ...... 5

1.1.3.3 Phosphatidylinositol 3 kinase activation ...... 6

1.1.4 Tpo signalling and thrombopoiesis ...... 6

1.2 Myeloproliferative neoplasms ...... 8

1.3 Ph+ myeloproliferative neoplasm – CML ...... 9

1.3.1 CML history and the BCR-ABL translocation ...... 10

1.3.2 Functions of wild-type BCR and ABL ...... 12

1.3.3 Intracellular signalling downstream of BCR-ABL ...... 13

iii 1.3.4 The mouse bone marrow transplant model ...... 15

1.3.5 Deletions at chromosome 9q34 in CML ...... 18

1.3.5.1 The exosome and function of EXOSC2 ...... 19

1.3.5.2 Function of PRDM12 ...... 24

1.3.6 Historical therapies for CML ...... 24

1.3.7 Development and eﬃcacy of imatinib mesylate ...... 25

1.3.8 Development of imatinib mesylate-insensitive relapse ...... 26

1.3.9 Next-generation ABL kinase inhibitors ...... 27

1.4 Ph– myeloproliferative neoplasms – PV, ET, and MF ...... 29

1.4.1 MPN disease phenotypes ...... 30

1.4.2 Discovery of JAK2 V617F ...... 31

1.4.3 Genetic diversity among Ph– MPN disease ...... 34

1.4.3.1 JAK2 exon 12 mutations ...... 34

1.4.3.2 MPL (Tpo-R) mutations ...... 35

1.4.3.3 LNK mutations ...... 35

1.4.3.4 TET2 mutations ...... 36

1.4.3.5 ASXL1 mutations ...... 37

1.4.3.6 EZH2 mutations ...... 37

1.4.3.7 JAK2 haplotype as a disease predictor ...... 38

1.4.3.8 Clonal diversity within disease and individual patients . 38

1.4.4 Inhibitors of JAK2 in the treatment of MPNs ...... 39

1.4.4.1 Ruxolitinib (INCB018424) ...... 40

1.4.4.2 SAR302503 (TG101348) ...... 41

1.4.4.3 Lestaurtinib (CEP-701) ...... 41

1.4.4.4 Momelotinib (CYT387) ...... 42

1.4.4.5 Pacritinib (SB1518) ...... 43

1.4.5 Mouse models for development and treatment of Ph– MPNs . . . 43

iv 1.4.5.1 BMT models ...... 44 1.4.5.2 Transgenic models ...... 45 1.5 Rationale and hypothesis ...... 48 1.6 Thesis statement and study aims ...... 49

2 Random mutagenesis reveals residues of JAK2 critical in evading inhibition by a tyrosine kinase inhibitor 51 2.1 Abstract ...... 52 2.2 Introduction ...... 53 2.3 Materials and methods ...... 55 2.4 Results ...... 60 2.5 Discussion ...... 77

3 Examining Chromosome 9q34 Deletion as a Marker for Genomic Insta- bility in Chronic Myeloid Leukemia 85 3.1 Abstract ...... 86 3.2 Introduction ...... 87 3.3 Materials and methods ...... 89 3.4 Results ...... 96 3.5 Discussion ...... 117

4 Discussion 122

5 Concluding Remarks 136

References 138

v List of Tables

1.1 Exosome cross-kingdom protein sequence conservation ...... 22

2.1 Structural diagrams of selected JAK2 inhibitors ...... 62 2.2 JAK2 kinase domain mutations identiﬁed in an inhibitor-resistance screen 63

3.1 EXOSC2 and PRDM12 exon sequencing primers ...... 97 3.2 Quantitative PCR raw data from CML patient samples ...... 100 3.3 Quantitative PCR relative expression from CML patient samples . . . . . 101 3.4 9q34 and 22q11 status in CML samples as called by aCGH ...... 103 3.5 A genomic map of the 9q34 and 22q11 loci in CML patient samples . . . 106 3.6 CML patient copy number gains from related publications ...... 109 3.7 CML patient copy number losses from related publications ...... 110 3.8 EXOSC2 and PRDM12 exon sequencing ...... 116

vi List of Figures

1.1 Hematopoietic diﬀerentiation ...... 2

1.2 Epo signalling and JAK2 domain structure ...... 7

1.3 BCR-ABL structure and signalling ...... 16

1.4 The mammalian exosome ...... 23

2.1 Location of the putative JAK2 inhibitor-resistant mutations ...... 64

2.2 JAK2 mutations display resistance to JAK inhibitor-I ...... 65

2.3 TEL-JAK2 mutants are not resistant to TG101348 or CEP-701 . . . . . 66

2.4 TEL-JAK2 inhibitor-resistant mutants display enhanced phosphorylation of Stat5, Akt and Erk1/2 - I ...... 68

2.5 TEL-JAK2 inhibitor-resistant mutants display enhanced phosphorylation of Stat5, Akt and Erk1/2 - II ...... 69

2.6 TEL-JAK2 and Jak2 V617F phosphorylate JAK2 substrate activation loop sequences ...... 70

2.7 TEL-JAK2 mutants G935R and R975G display a strong degree of inhibitor resistance ...... 72

2.8 Jak2 V617F G935R is resistant to JAK Inhibitor-I ...... 74

2.9 Jak2 V617F G935R is not resistant to TG101348 or CEP-701 ...... 75

2.10 Jak2 V617F G935R displays enhanced Stat5 and Erk1/2 phosphorylation 76

2.11 Jak2 V617F G935R displays a strong degree of inhibitor resistance . . . . 78

vii 2.12 JAK2 inhibitor-resistant residues mapped to the crystal structure bound to JAK Inhibitor-I ...... 82 2.13 Inhibitor-resistant mutations identiﬁed in mJak1 and hJAK2 mutagenesis screens ...... 84

3.1 EXOSC2 and PRDM12 quantitative PCR standard curves ...... 99 3.2 Nexus software visualization of the 9q34 locus in CML patient samples . 104 3.3 Nexus software visualization of the 22q11 locus in CML patient samples . 105 3.4 Nexus-generated virtual karyotype from CML patient samples ...... 108 3.5 Genomic loss at 12q13.32 in the RB1 tumour suppressor gene ...... 112 3.6 EXOSC2 antibody validation ...... 114 3.7 Exosc2 and L32 expression in hematopoietic progenitors ...... 115

viii List of Abbreviations

5-FU 5-ﬂuorouracil BMT Bone marrow transplant

5mC 5-methylcytosine c-abl ABL1 tyrosine kinase, human orthologue ABL Abelson murine leukemia viral oncogene homo- c-mpl Thrombopoietin recep- logue tor, human orthologue CC Coiled coil aCGH Array comparative genomic hybridization cDNA Complementary DNA

AML Acute myeloid leukemia CFU-E Colony-forming unit, erythroid ARE AU-rich element CLP Common lymphoid pro- ARED AU-rich element genitor database CML Chronic myeloid leu- ATP Adenosine tri-phosphate kemia CMML Chronic monomyelocytic aUPD Acquired uniparental dis- leukemia omy CMP Common myeloid pro- B-ALL B-cell acute lymphoblas- genitor tic leukemia der(9) Derivative chromosome 9 BAC Bacterial artiﬁcial chromosome DH DBL homology DNA Deoxyribonucleic acid BCR Breakpoint cluster region dNTP Dinucleotide triphos- BCR-ABL Fusion gene created with phate BCR and ABL1 ENU N-ethyl-N-nitrosourea BFU-E Burst-forming unit, erythroid EPO Erythropoietin Epo Erythropoietin BID Bis in die (Latin: twice a day) EPO-R Erythropoietin receptor

ix ET Essential thrombo- HSC Hematopoietic stem cell cythemia IFN-α Interferon alpha EXOSC Exosome sub-component IFN-γ Interferon gamma FCS Fetal calf serum IgG Immunoglobulin G FDA Food and drug adminis- IL-3 Interleukin 3 tration IL-3-R Interleukin 3 receptor FERM 4.1 / ezrin / radixin / moesin IL-6 Interleukin 6 FF1 Flip-ﬂop1 IM Imatinib mesylate FISH Fluorescent in situ hy- ITD Internal tandem duplica- bridization tion G-CSF Granulocyte colony stim- JAK Janus kinase ulating factor JH1 JAK kinase domain GAS Gamma activated sequence JH2 JAK dual-speciﬁcity kinase domain GDP Guanine diphosphate JI1 JAK Inhibitor-I GM-CSF Granulocyte macrophage colony stimulating factor LOH Loss of heterozygosity

GMP Granulocyte macrophage LP Lymphoid progenitor progenitor LSK Lineage–, sca-1+, c-kit+ GST Glutathione-S-transferase MAP Mitogen activated protein GST-J2s Fusion of glutathione-S- transferase and the 11 MAPK Mitogen activated pro- amino acid JAK2 sub- tein kinase strate Mast Mast cell GTP Guanine triphosphate MDR Minimal deleted region HCT Hematocrit MEP Megakaryocyte erythro- hmC 5-hydroxymethylcytosine cyte progenitor hopTum-1 HopscotchTumorous- MF Myeloﬁbrosis lethal mJak Murine Jak orthologue HR Homologous repair MMEJ Microhomology-mediated HRP Horseradish peroxidase end joining

x MPL Thrombopoietin receptor PRC2 Polycomb repressive complex 2 MPN Myeloproliferative neoplasm / myeloprolifera- PRDM Positive regulatory do- tive disease main member

NHEJ Non-homologous end PTK Protein tyrosine kinase joining PV Polycythemia vera NK Natural killer cell qPCR Quantitative real-time PCR p190 The 190 kDa isoform of BCR-ABL qRT-PCR Quantitative real-time PCR p210 The 210 kDa isoform of BCR-ABL RB1 Retinoblastoma 1 p230 The 230 kDa isoform of RBC Red blood cell BCR-ABL RNA Ribonucleic acid PBS Phosphate buﬀered RT-PCR Reverse-transcriptase saline PCR PCR Polymerase chain reac- SD Standard deviation tion SDS sodium dodecyl sulfate PDGF-R Platelet derived growth SDS-PAGE Sodium dodecyl sulfate factor receptor polyacrylamide gel elec- PDPK1 3-phosphoinositide de- trophoresis pendent protein kinase-1 SH2 Src homology 2 PH Pleckstrin homology SH3 Src homology 3

Ph+ Philadelphia chromosome siRNA Short interfering RNA positive ST Serine / threonine kinase PI Phosphatidylinositol STAT Signal transducer and PI3K Phosphatidylinositol 3 activator of transcription kinase STI571 Imatinib mesylate pI:pC Polyinosine:polycystine TBS-T Tris-buﬀered saline with tween 20 PMF Primary myeloﬁbrosis TEL-JAK2 Fusion gene created with PNPase Polynucleotide phospho- TEL and JAK2 rylase TNF-α Tumour necrosis factor PNT Pointed alpha

xi v-abl viral ABL1 orthologue TPO Thrombopoietin WBC White blood cells

TPO-R Thrombopoietin receptor Y Tyrosine

xii Chapter 1

Introduction

1.1 Hematopoiesis

Hematopoiesis is the process by which multipotent bone marrow-based stem cells differentiate and mature into fully formed blood cells. These blood cells populate the circulatory system and fight infection within the body’s tissues (Figure 1.1). The multipotent stem cell is maintained at a very low frequency in the bone marrow, and can undergo symmetric or asymmetric division to maintain its own population in the bone marrow or repopulate tissues, respectively. The immature hematopoietic cells receive external stimuli, which are transmitted across the plasma membrane through membrane-spanning proteins. Such extracellular signals can initiate intracellular cascades, which are transmitted primarily through the addition of phosphate moieties to substrates by kinase domain-containing proteins. The addition of phosphate moieties can create co-localization of further members of a pathway, and subsequent propagation and amplification of the extracellular signal. The extracellular signal ultimately activates transcription factors which migrate to the nucleus and stimulate transcription of genes necessary for survival and a specific differentiation path.

1 Chapter 1. Introduction 2

HSC

CMP CLP

LP RBC Mast GMP

Meg NK

B Thrombocyte Basophil Neutrophil EosinophilMono T

Macrophage Plasma

Figure 1.1: Hematopoietic Differentiation. The hematopoietic lineages begin at a multi-potent hematopoietic stem cell (HSC). Differentiation in the lymphocytic pathway proceeds through the common lymphoid progenitor (CLP) and lymphoid progenitor (LP), ultimately ending as a B cell (B), T cell (T), or natural killer cell (NK). Myeloid differentiation occurs through the common myeloid progenitor (CMP), differentiating into megakaryocte (Meg), red blood cell (RBC), mast cell (Mast), or granulcyte- monocyte progenitor (GMP). The GMP can further differentiate into a basophil, neutrophil, eosinophil, or monocyte followed by macrophage. Chapter 1. Introduction 3

1.1.1 Epo signalling and erythropoiesis

One such transmembrane signal transduction protein is the erythropoietin receptor (Epo- R), which is necessary for erythroid survival and diﬀerentiation. The temporal expression of Epo-R in erythroid-committed progenitor cells begins at the burst-forming unit erythroid (BFU-E) stage, and erythropoietin (Epo) is required for growth at the subsequent colony-forming unit erythroid (CFU-E) stage through to the erythroblast stage [1]. The Epo-R is thought to exist as a pre-formed dimer on the surface of erythroid precursors [2]. Epo, the Epo-R ligand, is secreted by interstitial kidney cells in response to reduction in blood oxygen concentration, which stimulates an increased production of red blood cells [3,4]. This response is typically observed in prolonged time at increased altitude [5]. Epo is transported by the circulatory system from the kidneys to the bone marrow where it binds its cognate receptor, Epo-R, and transmits an intercellular signal through a receptor conformational change [6] (Figure 1.2a). The Jak2 FERM domain constitutively binds the Epo-R Box 1 motif [7, 8]. Unligated Epo-R-associated Box 1 motifs exist separated by 73 A,˚ which is a suﬃcient distance to maintain the Jak2 kinase domains in an inactive state [6]. The Epo-induced Epo-R conformational change reduces separation of the Box 1 motifs to 39 A,˚ which facilitates cross-phosphorylation and activation of the Jak2 proteins [6]. Activation of the two Jak2 kinase domains results in phosphorylation of seven Epo-R cytoplasmic tyrosine motifs [9].

1.1.2 The tyrosine kinase domain and JAK2

The tyrosine kinase domain is an enzymatic domain that transfers a phosphoryl group to a substrate protein. Tyrosine kinase domains are common in eukaryotic genomes, and tyrosine kinase-like domains are found in species of bacteria and archaea. Metazoan tyrosine kinase domain-containing proteins are called Protein Tyrosine Kinases (PTKs). This group of proteins has many roles in maintaining cellular function, including the delivery of extracellular mitogenic signals, and controlling cell division, adhesion, and Chapter 1. Introduction 4 migration. As such, they are commonly perturbed in human cancers.

JAK2 is a non-receptor tyrosine kinase and a member of the Janus kinase protein family [10, 11] (Figure 1.2b). The related kinases, JAK1, JAK3, and TYK2, all bear 96% identity to JAK2. Like all four members of this family, JAK2 is composed of four domains. The amino-terminal FERM (4.1/ezrin/radixin/moesin) domain mediates binding to cytokine receptors [12]. The second domain is a putative SH2 domain based on sequence similarity, has no known function [13], but is essential for the transforming ability of mutant JAK2 [14]. The third domain (JH2) has classically been referred to as a ‘pseudo-kinase’ domain because residues critical for kinase function have been mutated over its evolutionary history. Recent work has demonstrated that the JH2 domain is a dual-speciﬁcity kinase and phosphorylates JAK2 Ser523 and Tyr570, acting to negatively regulate the JAK2 kinase domain (JH1), although the details of domain interaction is unknown [15]. The carboxy-terminal domain (JH1) is a tyrosine kinase domain that targets Yxx[L/I/V] amino acid motifs for transfer of a phosphoryl group to the leading tyrosine [16]. JAK2 phosphorylation of tyrosine motifs on the constitutively bound receptor creates a docking site for protein recruitment via SH2 domains, most notably members of the signal transducer and activator of transcription (STAT) protein family [17–19], as discussed below.

1.1.3 JAK-STAT signalling pathway

JAK2 is essential for mammalian development. Jak2 germline knockout mice are homozygous lethal at embryonic day 12.5 [20]. Histological analysis of fetal mice demonstrated anemia in the liver and surrounding tissues, suggesting a defect in hematopoiesis [20]. JAK2 is downstream of multiple receptor families including the type I cytokine receptor family (e.g. interleukin-3 receptor (IL-3-R) [21], granulocyte-macrophage colony stimulating factor receptor (GM-CSF-R) [22], EPO-R [9, 23], and thrombopoietin receptor (TPO-R, c-mpl) [24, 25]), the type II cytokine receptor family (e.g. interferon gamma Chapter 1. Introduction 5

(IFN-γ) [26]), and the gp130 receptor subunit [27, 28]. Jak2 –/– hematopoietic progenitors were not able to form in vitro colonies in response to Epo, interleukin-3 (IL-3), or thrombopoietin (Tpo), which suggested Jak2 has a non-redundant role in the downstream signalling of these cytokines [20].

1.1.3.1 Stat activation

Speciﬁc phosphorylation of Epo-R Y343 and Y401 results in the recruitment and phosphorylation of Stat5a/b [18, 19, 29, 30] (Figure 1.2b), and Y431 results in Stat3 recruitment and phosphorylation [31]. Stat phosphorylation is thought to induce its release from the Epo-R, follwed by hetero or homodimerization and migration to the nucleus. Stat dimers act as transcription factors and bind to IFN-γ activated sequence (GAS) elements [32, 33] to stimulate transcription of genes important in survival and erythroid lineage commitment. Stat1 is downstream of Jak2 activation, but does not require any Epo-R cytoplasmic tyrosine motifs to be targeted for phosphorylation [34]. Stat1–/– mice displayed reduced bone marrow-derived CFU-E and more splenic BFU-E and CFU-E. Splenic colonies displayed Epo hypersensitivity [34].

1.1.3.2 Mitogen-activated protein kinase activation

Epo-R ligation also results in activation of the Ras / mitogen-activated protein (MAP) kinase pathway through recruitment of the Grb2 scaﬀold protein [35] (Figure 1.2b). Grb2 can be indirectly recruited to the Epo-R through interactions with Shc [36], Shp- 2 [37], or Ship [38]. Localization of Grb2 to the cell membrane results in Sos-dependent activation of the Ras GTPase [39], which begins the MAP kinase cascade ultimately leading to the activation of transcription factors important for cell growth and erythroid diﬀerentiation [40]. Chapter 1. Introduction 6

1.1.3.3 Phosphatidylinositol 3 kinase activation

The third critical pathway downstream of the Epo-R is the phosphatidylinositol 3 kinase (PI3K) survival pathway. The PI3K heterdimer can be recruited to the Epo-R through direct binding of Y479 [41] (Figure 1.2b) or through one of several accessory factors such as Gab2 [42]. Recruitment of PI3K to the cell membrane results in phosphorylation of a speciﬁc class of lipids called phosphatidylinositols (PIs) [43]. The phosphorylated version of the PIs are called either PI(3)P, PI(3,4)P2, or PI(3,4,5)P3, depending on the location of the phosphate moiety attachment to the inositol ring. This speciﬁc phosphorylation results in the recruitment and binding pleckstrin homology (PH) domain-containing proteins including the 3-phosphoinositide dependent protein kinase-1 (PDPK1). Pdpk1 activates Akt, a serine / threonine kinase, that stimulates pathways critical in cellular survival, glucose metabolism, and growth [44].

In summary, the binding of Epo to the Epo-R causes a conformational-dependent activation of Jak2, which results initiation of the Stat, MAP kinase, and PI3K pathways. These pathways stimulate the transcription of genes important in erythroid diﬀerentiation and survival.

1.1.4 Tpo signalling and thrombopoiesis

The administration of Epo to murine models can stimulate the growth of megakaryocytes and the production of platelets [45–47], suggesting overlap between Epo- and Tpo-stimulated pathways. It was later shown that thrombopoietin can rescue Epo-R–/– mouse embryos [48], demonstrating the overlap is bi-directional. The thrombopoietin receptor (Tpo-R, c-mpl, mpl) [49] and thrombopoietin (Tpo) [50] were cloned in 1992 and 1994, respectively. The amino-terminal extracellular Tpo-R domain bears 23% identity to Epo-R, but has a very similar tertiary structure including numerous invariant residues critical in ligand binding [51]. Sequence alignment between Epo-R and Tpo-R orthologues Chapter 1. Introduction 7

a) Epo

P P Ras Jak2 Jak2 PDPK1 Sos Shc P P Y343 Y343 Ship Stat5 Grb2 Y401 Y401 P Y429 Y429 Shp2 Y431 Y431 Y443 Y443 PI3K Y460 Y460 Y464 Y464 P MAP kinase Y479 Y479 pathway

Stat5 P

P Erk1/2 Stat5 Akt P

Transcription of genes important in survival and erythroid di!erentiation

b) JAK2 FERM SH2 JH2 JH1

V617F Figure 1.2: Epo Signalling and JAK2 Domain Structure. (a) Epo signalling begins with ligation of the Epo cytokine to the extracellular domain of the homodimeric Epo receptor (Epo-R). This ligation causes a conformational change in the cytoplasmic tail, bringing the constitutively associated JAK2 proteins within close proximity, resulting in cross-phosphorylation and activation. JAK2 activation also results in phosphorylation of up to eight Yxx[L/I/V] motifs on the Epo-R cytoplasmic tail, which creates docking sites for effectors of the MAP kinase, Stat, and PI3K / Akt pathways. Pathway activation results in transcription of genes important in survival and erythroid lineage commitment. (b) JAK2 domain structure consists of an amino-terminal FERM domain, a putative SH2 domain, the dual-specificity kinase JH2 domain (to which the V617F mutation maps), and the JH1 kinase domain. Chapter 1. Introduction 8 identified numerous Tpo-R-conserved residues, which are thought to determine the specificity of Tpo-R to Tpo [51]. It is currently unknown if the Tpo-R exists as pre-formed dimers, as does the Epo-R, on the surface of cells. However, a S368C point mutation in the Tpo-R intracellular domain results in a disulfide bridge, constitutive receptor dimerization, activation of downstream signalling pathways, and induction of tumours in mice injected with BaF3 cells carrying the activated the S368C allele [52]. The primary Tpo- R transcript (the P form) contains Box 1 and Box 2 domains thought to function in Jak2 binding, as demonstrated through Tpo-R truncation mutants [53]. Thrombopoi- etin stimulation of M07e cells [54] results in the phosphorylation of JAK2 [55], suggesting an intracellular signalling pathway very similar to that in Epo-stimulated cells. As in Epo signalling, Tpo stimulation causes the Jak2-dependent phosphorylation of Stat3 and Stat5 [56] and activation of the MAP kinase pathway through recruitment of Grb2 and Sos [57]. Tpo also stimulates the activation of the PI3K / Akt survival pathway indirectly through interaction with Gab2 and Shp2 [58]. Tpo can induce transcription of the pro-survival factor Bcl-XL through Stat5- and PI3K-dependent pathways [59], and the cell cycle control protein p21 through Stat5 alone, which acts as a megakaryocyte differentiation factor [60].

Overall, discovery of Stat, MAP kinase, and PI3K pathway stimulation downstream of the Tpo-R gave a framework to understand the considerable overlap in phenotypic response to Tpo and Epo. Such overlap becomes particularly relevant in Jak2 tyrosine kinase disease models, discussed at length below (section 1.4.5.1).

1.2 Myeloproliferative neoplasms

Tyrosine kinase proteins are essential to signal transduction pathways in hematopoietic progenitors. Many of these pathways promote growth and survival, and therefore the tyrosine kinases that transmit such signals are exquisitely controlled both in activity and Chapter 1. Introduction 9 temporal expression. Unsurprisingly, tyrosine kinase deregulation is a frequent event in the development of hematopoietic disorders. A chromosomal translocation can result in a deregulated tyrosine kinase domain (e.g. BCR-ABL [61], TEL-JAK2 [62, 63]); an internal tandem duplication can relieve the auto-inhibitory role of a particular domain on its tyrosine kinase (e.g. FLT3-ITD [64, 65]); an amino acid substitution can im- pair the binding of negative regulatory elements (FLT3 D835Y [66], JAK2 V617F [67]); an interstitial deletion can fuse two genes (e.g. PDGF-Rα-FIP1L1 fusion [68]); and episomal-based amplification can cause overexpression of a deregulated tyrosine kinase (NUP214-ABL1 fusion [69]). Activation of tyrosine kinase proteins play a prominent role in the development of myeloproliferative neoplasms (MPNs), a group of hematopoietic disorders characterized by an unregulated, clonal expansion of specific hematopoietic lineages. While a myeloproliferative neoplasm is defined as the excess production of any fully differentiated myeloid cell, the major four types are as follows. The sole Philadelphia chromosome-positive MPN is chronic myeloid leukemia (CML). CML is characterized by an excess of granulocytes, including neutrophils, basophils, and eosinophils. The three major Philadelphia chromosome-negative MPNs are polycythemia vera (PV), characterized by an excess of erythrocytes; essential thrombocythemia (ET) by an excess of mature megakaryocytes and platelets; and myelofibrosis (MF) by a deregulation of inflammatory cytokines resulting in excess production of collagen, impaired hematopoietic development in the bone marrow, and pancytopenia.

1.3 Ph+ myeloproliferative neoplasm – CML

CML is a tri-phasic disease of the blood and bone marrow characterized by an unregulated expansion of the myeloid lineage. The median age of onset is 65 years, and 85% of patients present in chronic phase with fatigue, weight loss, and splenomegaly. Peripheral blood analysis reveals an overwhelming expansion of fully differentiated myeloid cells Chapter 1. Introduction 10 with a predominance of neutrophils. Diagnosis of CML is confirmed with detection of the Philadelphia chromosome and/or BCR-ABL fusion gene. The BCR-ABL fusion gene is necessary but not sufficient for a diagnosis of CML, as it can also be observed in pediatric and adult B-cell acute lymphoblastic leukemia (B-ALL) [70], neutrophilic CML [71], and rarely in acute myeloid leukemia (AML) [72]. CML progresses through three phases. Chronic phase is characterized by an elevated white blood cell count, the presence of the Philadelphia chromosome and/or BCR-ABL fusion gene, and generally less than 5% blasts in the blood. Transition to accelerated phase is thought to occur when additional genetic lesions are acquired that interfere the ability of the progenitor clone to produce terminally differentiated cells. Such mutations result in a slow accumulation of blast cells in the bone marrow and peripheral blood. Blast crisis, the final phase of CML, is diagnosed based on a marked increase of blast cells in the peripheral blood. Further evidence of clonal evolution is apparent in karyotypic analysis, and extramedullary blast proliferation and non-hematopoietic tissue infiltration are common. Patients are typically refractory to all therapies and care becomes palliative in nature. The most common causes of death in end-phase CML is sepsis and hemorrhage.

1.3.1 CML history and the BCR-ABL translocation

A detailed physical description of leukemia ﬁrst appeared in the medical literature in three separate 1845 case reports from Virchow (Germany) [73], Bennett (Scotland) [74], and Craigie (Scotland) [75] (reviewed in [76]). Each publication had a variation on a theme of ‘white blood’ [76]. A description of myeloid leukemia followed in 1878 by Neumann (Germany) [77]. Cell staining methods developed and published by Paul Erlich in 1891 allowed physicians to diﬀerentiate types of blood cells and consequently stratify leukemia cell types [78]. The description of the ‘Philadelphia chromosome’ (Ph), so named after the city in which it was discovered, as a consistent marker in human chronic granulocytic leukemia was published by Peter Nowell and David Hungerford in 1960 [79]. Chapter 1. Introduction 11

Evidence for chromosomes 9 and 22 participating in the generation of the Philadelphia chromosome was described by Janet Rowley in 1973 [80, 81]. A potential role for the ABL1 tyrosine kinase in CML was suggested by Annelies De Klein and John Stephenson in 1982, describing the translocation of c-abl from chromosome 9 to 22 [82]. Evidence for a role of ABL1 in a Ph+ cell line was published in by James Konopka and Owen Witte in 1984, where they detail an ‘unmasking’ of the c-abl tyrosine kinase activity in K562 cells [61]. The importance of the coiled-coil domain was discovered in 1991, where deletion experiments localized the colony-forming potential of BCR-ABL to the first exon of BCR [83]. Significantly, CML was the first disease demonstrated to have a genetic origin. Subsequent discovery of the specific genetic lesion and protein product relevant to disease advanced our understanding of hematologic malignancies.

Since the description of the Philadelphia chromosome, evidence for translocations in human cancers been overwhelming. Persistent chromosomal translocations result from the fusion of two separate genes, one of which is typically a proto-oncogene, that causes protein deregulation and gives the cell a proliferative or survival advantage. Chronic myeloid leukemia is caused by the t(9;22)(q34;q11) chromosomal translocation, which results in the fusion of 5’ end of BCR (breakpoint cluster region) and the 3’ end of the potent tyrosine kinase proto-oncogene ABL1 [84]. There are several BCR-ABL alleles associated with hematopoietic malignances. Each allele fuses the same ABL sequence and variable lengths of BCR. CML is associated with the p210 version of the BCR-ABL protein (Figure 1.3), which is generated when the translocation-inducing double strand break occurs between BCR exons 12 and 16 (in the ‘major breakpoint cluster region’). Such a breakpoint causes the inclusion of the coiled-coil (CC), serine / threonine (ST) kinase, DBL-homology (DH), and pleckstrin homology (PH) domains of BCR in the protein product. Two other versions of the BCR-ABL fusion gene can be generated and are associated with their own distinct diseases and pathology. p190 is typically observed in Ph+ B-ALL [70]. Such a protein product results from a double-strand break between Chapter 1. Introduction 12

exons 1 and 2 of BCR (in the ‘minor breakpoint cluster region’), and includes only the CC and ST kinase domains. p230 results from a translocation at the 3’ end of BCR (in the ‘micro breakpoint cluster region’) in which nearly all BCR domains are fused to the 3’ terminus of ABL. The p230-generating allele is observed in neutrophilic chronic myeloid leukemia [85]. Interestingly, it appears that disease severity is inversely related with the length of BCR included in the fusion gene, suggesting that the additional domains serve some regulatory or inhibitory role in the larger fusion proteins [86,87].

1.3.2 Functions of wild-type BCR and ABL

BCR resides in a 138 kb region at chromosome 22q11.23, and produces a 160 kDa protein with six domains (reviewed in [88]). Human BCR displays 93 and 95% sequence similarity and identity to the mouse orthologue. This is considerably above the mouse-human average of 70.1% identity between the approximately 12,000 direct mouse-human orthologues [89]. BCR is ubiquitously expressed at a low level, with higher expression in the hematpoietic lineages, brain, and thymus (GeneAtlas U133A dataset, probe 202315 s at). The N-terminal CC domain promotes oligomerization of BCR proteins [90, 91]; the ST kinase phosphorylates two 14-3-3τ proteins [92–95]; RhoGEF [96] and RAC-GAP (GTPase-activating protein) [97] domains function in rac/rho signalling to modulate cell adhesion and migration, indirectly supported by data demonstrating an siRNA knockdown of BCR reduced epithelial cell migration in a wound healing assay [98] (primary data retrieved from http://www.cellmigration.org/). Early mouse knockout studies demonstrate Bcr is dispensable for survival, but Bcr –/– animals display an increase in reactive oxygen species during neutrophil priming [99]. Considering the high level of conservation between mouse and human BCR othrologues, perhaps there is a yet unknown BCR function in cellular stress or adaptation that maintained selective pressure on the ancestral sequence.

Identiﬁcation of v-abl was made by Herbert Abelson and Louise Rabstein while study- Chapter 1. Introduction 13

ing the development of Moloney murine leukemia virus-induced tumours in mice (reviewed in [100]). One of 153 mice developed an atypical lympho-sarcoma, which was transmittable through ﬁltered extracts. Cell surface markers of transformed cells con- ﬁrmed a B cell lineage expansion. Genome characterization suggested that the new virus was generated through recombination of the Moloney murine leukemia virus with the c-abl proto-oncogene. The new virus was named the Abelson Murine Leukemia Virus. The truncated viral abl, missing the 114-most amino-terminal codons, was named v-abl. The Abelson Murine Leukemia Virus could readily transform hematopoietic cells to IL-3 independence. c-abl (ABL1 ) is located at 9q34.1 and spans approximately 280 kb. Abl expression is ubiquitous through murine development and in all tissues [101]. The ABL protein is a 120 kDa non-receptor tyrosine kinase, which contains SH2 and SH3 domains that act as negative regulators of the tyrosine kinase domain [102, 103]. In the inactive conformation, the SH2 and SH3 domains clamp over the kinase domain and secure it in the inactive state [103]. Ligation of the amino-terminal myristoyl-binding pocket confers stability to the inactive state [103]. The myristoyl binding site is critical for a new class of BCR-ABL inhibitors, discussed below (section 1.3.9). ABL is activated through phosphorylation of Y245, a site between the linker and kinase domains that causes dissociation of the SH2 and SH3 domains from the kinase domain (reviewed in [104]). The activation loop and Y412 are then exposed and a second phosphorylation event is required to fully activate the ABL kinase. ABL also contains nuclear export and localization signals, as well as DNA-binding and actin-binding domains. Abl1 –/– mice display stunted growth and perinatal lethality [101]. The mice that survive the initial two weeks post-partum have defects in thymic and splenic architecture, and a B and T cell lymphopenia [101].

1.3.3 Intracellular signalling downstream of BCR-ABL

The BCR-ABL fusion protein activates many signalling pathways important in hematopoietic cell proliferation and development (Figure 1.3). The BCR-encoded CC domain in- Chapter 1. Introduction 14 duces oligomerization and constitutive ABL tyrosine kinase activity through kinase domain proximity [83]. Co-localization of ABL tyrosine kinase domains causes linker region phosphorylation and relieves the auto-inhibitory role of the SH2 and SH3 domains [102]. The kinase exists in a constant active state. The novel association of BCR-recruited signalling complexes to the constitutively active ABL tyrosine kinase results in non- evolutionary access to substrate proteins and stimulation of highly regulated pro-growth and anti-apoptotic intracellular pathways.

Constitutive Abl kinase activity results in phosphorylation of Y177, which is central to disease development as demonstrated in mouse models discussed below (section 1.3.4). pY177 generates a docking site for the SH2 domain-containing scaffold protein Grb2 [105]. Grb2 contains a pY177-binding SH2 domain flanked by two SH3 domains [106]. The Grb2 SH3 domain recruits Sos1, a Ras guanine nucleotide exchange factor, resulting in constitutive GTP-bound Ras and downstream MAP kinase signalling that promotes cellular proliferation [107, 108]. Ras activity is required for the transformation ability of BCR-ABL, demonstrated through studies in the transforming potential of Ras mutants in fibroblasts and primary bone marrow [109]. Ras is a highly regulated proto-oncogene that is deregulated in a significant proportion of human cancers [110].

Recruitment of Grb2 to Y177 also results in Gab2 binding and phosphorylation [111]. Gab2 is one of the principle activators of the phosphatidylinositol-3 kinase (PI3K) survival pathway [111]. PI3K phosphorylation results in activation of Akt, a serine/threonine kinase, which is required for the transformation of hematopoietic cells by BCR-ABL [112]. Coupling of BCR-ABL to the PI3K pathways confers a pro-survival, anti-apoptotic signal [111].

Cell adhesion and migration is modiﬁed by the recruitment of Vav to BCR-ABL. Vav is a guanine nucleotide exchange factor for the Rho family of GTPases [113]. Upon Vav recruitment, Rac1 is activated through GDP-GTP exchange [114]. This results in altered adhesion and migration of BCR-ABL+ cells [115]. Chapter 1. Introduction 15

The preceding were selected examples of BCR-ABL-activated signalling pathways. Publications describing complete BCR-ABL signalling in hematopoietic cells are available [76,116–118].

1.3.4 The mouse bone marrow transplant model

The importance of signalling pathways and specific proteins has been extensively dis- sected with knockout mice and the murine bone marrow transplant (BMT) model. The BMT is a powerful tool to examine the contribution of specific genes to hematologic disease latency. Briefly, bone marrow cells are collected from a donor mouse of a specific genotype (wild type as a control, and heterozygous or null for a specific gene of interest). These bone marrow cells are then infected with retrovirus expressing a specific oncogene (BCR-ABL, for example) and transplanted into a lethally irradiated mouse. Typical latency for a BCR-ABL-infected wild-type model is 18 to 23 days, with mice succumbing to a myeloid blast crisis-like disease with significant lung infiltration and hemorrhaging [119, 120]. If the gene of interest contributes to disease, survival of the heterozygous or null bone marrow recipient can be considerably extended.

Early bone marrow transplant experiments were used to determine the relative consequence of the p190, p210, and p230 isoforms of BCR-ABL on hematopoietic disease development [121]. When wild-type donor cells were enriched for primitive progenitors with 5-ﬂuorouracil (5-FU), the three BCR-ABL isoforms induced an equally potent CML- like disease (18 to 23 day latency) [121]. Southern blot analysis of proviral integration sites suggested each disease was polyclonal, however, secondary transplantations demonstrated not all clones were capable of inducing disease. Interestingly, when bone marrow recipients were not treated with 5-FU, disease stratiﬁcation according to oncogene became apparent. p190 induced disease in 21 to 28 days much like when 5-FU primed marrow was used, but the predominant disease cell type in that cohort was lymphoid, Chapter 1. Introduction 16 a) p210 breakpoint

CC S/T DH PH BCR Actin binding

ABL SH3 SH2 Y-Kinase b) P P BCR-ABL RAC-GAP

RAC-GAP

P P c) GRB2 Actin binding, altered localization to the cell membrane SOS GAB2 P

PI3K VAV SHP2 P

P Akt RAS RAC1

Altered adhesion Proliferation Survival and migration

Figure 1.3: BCR-ABL Structure and Signalling. (a) Wild-type BCR contains many domains including a coiled-coil oligomerization domain (CC), a serine/threonine kinase (S/T), a DBL-homology domain (DH), a pleckstrin homology domain (PH). Wild- type ABL protein domains include SH3 and SH2 domains, a tyrosine kinase domain (Y-Kinase), nuclear localization and export signals (not indicated), and actin-binding domains. (b) BCR-ABL signalling activates multiple downstream signalling pathways, selected are depicted here. The CC domain induces protein oligimerization, and phosphorylation of Y177 (left-most, docking site) and Y1294 (right-most, activation loop), recruitment, and activation of signalling pathways important in proliferation, survival, and altered adhesion and migration. It is the co-induction of these three signals that mediates the oncogenic signal of BCR-ABL. Figure adapted from Ren [117]. Chapter 1. Introduction 17 with smaller contributions from myeloid and macrophage, and occasionally of mixed lineage. p210 and p230 caused disease with latency up to 70 or 77 days, respectively, with a similar varied cell type [121]. The myeloid-lymphoid mixed disease observed in some recipients suggested a very early multipotent hematopoietic progenitor was the recipient of the transduction [121].

Similar experiments were completed to determine how Y177 helps BCR-ABL induce disease. In vitro experiments demonstrated that Y177 was essential for SH2-mediated recruitment of Grb2 and consequent Ras activation, as discussed above (section 1.3.3). However, the in vivo importance of Y177 was still unknown. To address this question, wild-type mouse bone marrow was transduced with either wild-type BCR-ABL or bearing the Y177F mutation, and transplanted into wild-type lethally irradiated recipient mice [122]. As previously demonstrated, wild-type BCR-ABL induced a polyclonal CML- like disease with a latency of approximately 18 to 23 days. In contrast, BCR-ABL Y177F prolonged latency to 77 to 112 days with almost exclusively mono- or bi-clonal lymphoid disease. These experiments demonstrated that Y177 is critical to inducing a myeloproliferative-like syndrome in mice, but appears to be dispensable for lymphoid leukemias [122].

A ﬁnal example of the utility of bone marrow transplants in BCR-ABL signalling examines the role of the BCR-ABL SH3 domain. The SH3 domain is an evolutionarily conserved protein domain that binds proline-rich motifs (-P-X-X-P-). The SH3 domain of ABL regulates kinase activity by binding the linker region between the SH2 and kinase domains, folding the protein into a inactive state [102]. Phosphorylation of linker-region and activation-loop tyrosine residues disrupt the SH3-linker interaction and results in the active form of ABL. In BCR-ABL, deletion of the CC domain attenuates disease [123] because the SH3-linker interaction can more easily maintain ABL in an inactive state [124]. Bone marrow transplant experiments demonstrate that the SH3 domain is dispensable for induction of CML-like disease because the CC-based oligomerization Chapter 1. Introduction 18

results in a constitutively open conformation [124]. Simultaneous deletion of the CC domain and inactivation of the SH3 domain (through deletion or mutation) restores the oncogenicity of the BCR-ABL fusion protein [124,125].

1.3.5 Deletions at chromosome 9q34 in CML

The majority of Philadelphia translocations are reciprocal, meaning no genomic material is gained or lost. However, deletions of large segments of the derivative chromosome 9 were identified using triple-probe fluorescent in situ hybridization (FISH) system targeting the genes immediately centromeric of the 9q breakpoint site (ASS, BCR, and ABL1 ) [126]. The absence of the 300 kb ASS probe binding to der(9) was observed in 16/56 patients screened [126]. These patients had deletions of heterogeneous size, up to several megabases. Their initial Kaplan-Meier analysis indicated a mean survival time of 36 months in patients with deletions, as compared to 90 months in patients without, giving deletion status a significant prognostic value when considering patients treated with interferon-α (IFN-α) [126]. The shorter survival time reflects a shorter chronic phase. It was determined that poor outcome of a patient possessing a deletion cannot be attributed to ABL-BCR expression [127,128] or presence of a ABL-BCR protein [126,129]. However, a recent study suggested that a stable ABL-BCR protein product has been detected in CML and B-ALL primary patient samples [130]. Regarding 9q34 deletions, one hypothesis suggested that the cause of the decreased survival time is due to hemizygous deletion of one or more tumour suppressor genes [131]. In order to determine if a known tumour suppressor gene mapped to the deleted region, further studies were conducted using finer mapping techniques at the 9q34 locus [132]. As the resolution of traditional FISH techniques was inadequate for detecting sub-microscopic deletions, quantitative PCR probes were anchored over 442 kb of genomic sequence flanking ABL- BCR to determine deletion status among a cohort of patients selected based on adverse outcome, and for whom FISH analysis indicated no deletion [132]. A minimal deleted Chapter 1. Introduction 19 region (MDR) of approximately 120 kb was found among 25 patients previously thought not to possess a deletion at 9q34. The initial analysis indicated that at least two genes mapped to this 120 kb region – PRDM12 and EXOSC2. Deletion of one PRDM12 allele was the most obvious candidate for the accelerated disease progression because it is a member of the PR domain-containing tumour suppressor family and is thought to function as a transcriptional repressor [132,133]. EXOSC2 is a member of the 10-protein exosome, responsible for trimming and degrading of a vast array of RNA species [134]. Detailed descriptions of EXOSC2 and PRDM12 follow below. In patients treated with conventional allogeneic bone marrow transplants and IFN-α therapy, deletion status is a significant indicator of poor prognosis [126,128,135–137]. However, there are conflicting data regarding the prognostic significance of a 9q34 deletion in patients treated with the tyrosine kinase inhibitor imatinib mesylate (IM) [138–142]. In many cases a trend towards a decrease survival time is apparent yet does not reach the threshold of statistical significance.

1.3.5.1 The exosome and function of EXOSC2

Temporal and spatial regulation of RNA copy number is a critical element to controlled cell growth [143, 144]. Eukaryotic RNA decay is completed through two pathways; 5’ - 3’ is controlled by the cytoplasmic exoribonucleases Rat1 and Xrn1 (human homologues XRN1 and XRN2) [145,146], and 3’ - 5’ degradation pathway is controlled by the exosome complex (distinct from exosome microvesicles) [134]. Both 5’ - 3’ and 3’ - 5’ pathways can target transcripts from RNA polymerase I, II, and III, and lie at the end of the nonsense-mediated decay, RNA interference, nuclear RNA surveillance, and other RNA decay pathways (reviewed in [146]).

Exosome terminology can be difficult to follow, as orthologues are named differently. For ease of understanding, the human genes and proteins will be referred to as EXOSC and EXOSC, respectively. The mouse orthologues Exosc and EXOSC. For a complete Chapter 1. Introduction 20 description of names and functions, see Table 1.1. The mammalian exosome core is a ring composed of six protein subunits (EXOSC4, 5, 6, 7, 8, 9; Figure 1.4a), each containing a non-functional RNase PH domain organized to face the interior of the ring [147]. Evidence suggests these modified RNase PH domains function in RNA substrate recognition [148]. Three capping proteins (EXOSC1, 2, 3; Figure 1.4b) control the entry of RNA substrates via KH [149,150] and S1 [151,152] RNA-binding domains. The exosome particle functions in the cytoplasm, nucleus, and nucleolus, with slight variation of interacting and accessory proteins [153–155]. Early studies in yeast demonstrate that all components of the exosome are essential for survival [154], and there is a single documented human disease involving an exosome component [156]. The eukaryotic nine-member core exosome was initially reported to have phosphorolytic exoribonuclease activity [147], but it later demonstrated this result was due to a co-migratory bacterial polynucleotide phos- phorylase (PNPase), an artifact from exosome purification. Additional anion and cation exchange chromatograph purification steps removed the contaminant, and the human nine-member exosome retained no catalytic activity (erratum issued in Cell, Volume 131, Issue 1, 188-189, 5 October 2007). Further studies suggest the RNase PH-like domains of the six core human exosome components have significant evolutionary divergence in key residues of the bacterial RNase PH domain [157, 158]. Yeast experiments revealed that the hydrolytic 3’ - 5’ exoribonuclease activity observed in purified exosomes is due to Rrp44p (DIS3) [157, 159], an accessory protein. These results mean that the eukaryotic core exosome 6-member ring and 3-member cap are catalytically inactive, and serve in substrate specificity and stabilization for the exoribonucleases PM/Scl-100 (Rrp6p), DIS3 (Rrp44p) and DIS3L1.

Early studies on the exosome were conducted in yeast where the orthologue of EX- OSC2 (Rrp4p) was initially identiﬁed [160]. A temperature-sensitive mutant was obtained in which immature 5.8S ribosomal RNA accumulated in the non-permissive temperature [160]. Further study indicated a defect in global 3’ RNA processing, and the Chapter 1. Introduction 21

EXOSC2 yeast orthologue was incorrectly identiﬁed as a 3’ - 5’ exoribonuclease [134], for reasons similar to those discussed above.

The evolutionary conservation of exosome proteins is striking. Homologous genes are present in all tested species of eukaryota and archaea (Table 1.1) [134,161–163], domains of life whose most recent common ancestor lived approximately 1.8 billion years ago [164]. Bacterial species have a similar complexes called the degradasome and PNPase, subunits and domains of which bear some similarity to the eukaryotic exosome [165]. Thus, given the importance of the complex, it is not surprising how sensitive cells are to its disruption.

Many cytokine and oncogene transcripts contain AU-rich elements (AREs) in their 3’ untranslated regions (UTRs) [172], which mediate efficient degradation of their mR- NAs through the exoribonuclease activity of the exosome. The function of the ARE sequence was initially discovered and explored in GM-CSF, demonstrating the AREs mediate its efficient degradation (Figure 1.4c) [172–174]. Since then, the entire human genome has been mined for ARE-containing genes and the results of which are available online in the AU-rich element database [175–177]. ARE-binding proteins have been extensively studied and are divided into both ARE-stabilizing and ARE-destabilizing co- factors (reviewed in [178]). ARE-destabilizing proteins include KSRP and TTP, which bind AREs and interact directly with exosome core components to promote degradation (Figure 1.4) [179]. Stabilization of a synthetic AU-rich element mRNA was observed when HeLa cell extracts were immunodepleted of exosome complexes [180], suggesting a reduction in functional exosome may result in stabilization of specific mRNA species. Support for this observation came from RNAi-based reduction of exosome subunits in Drosophila melanogaster S2 tissue culture cells [181], followed by expression microarray analysis [182]. Depletion of specific exosome subunits results in stabilization of unique mRNA profiles. When grouped according to ontology, D. melanogaster Exosc2 appears to mediate degradation of mRNAs involved in development, cellular signalling, and cell Chapter 1. Introduction 22 . , Sac- , Mtr3p is APE1447 YDR280W , Rrp4p is encoded YGR095C YNL232W , Rrp41 encoded by , and Rrp45p is encoded by APE1448 , Rrp46p is encoded by Core exosome components from yeast ( YCR035C YGR195W , Rrp4 encoded by genome database (http://www.yeastgenome.org/). ‘ts’ denotes APE0445 are listed with identity and similarity compared to human proteins, , Rrp43p is encoded by , Rrp41p is encoded by Saccharomyces YDL111C Aeropyrum pernix) YOL142W . Gene names retrieved from [170]. Table adapted from [171]. ) and archaea ( , Rrp42p is encoded by APE1445 sequence retrieved from the YRC Public Data Repository (http://depts.washington.edu/yeastrc/). Archaeal CSL4RRP4RRP40RRP41 Csl4pRRP46 Rrp4pMTR3 Rrp40p 36RRP42 Rrp41p / 50% 42 37 / / 57%OIP2 Rrp46p 53% 36 Csl4 /PM/Scl-75 57% Rrp4 Mtr3p Rrp4 30 /PM/Scl-100* Rrp42p 53% Rrp41* Rrp45pDIS3* Rrp6p* 26 23 / 23 Rrp41* / 30 42% 25 / 39 37% Rrp43p / / 41%DIS3L1* / 44% 51% 58% 35 32 / 37 / S1 60% Rrp41* / 52% S1 24 47% S1 Rrp42 & / RNase & KH 39% PH-like KH Rrp42 – 34 Rrp44p* RNase Rrp44p / Yes PH-like * 50% [154] Rrp42 33 / 41 33 51% Yes / 31 / [154] 59% / RNase 49% 53% PH-like Yes 35 Yes [160] RNase [154] / PH-like 51% – No Yes – [153,154] RNase – PH-like Yes Yes No [154,166] [154] No RNase PH-like Yes [134,154] No No Yes [156] Yes No [154] No – – RNase D No RNase RNase ts R R [167,168] No Yes Yes [154,169] [154,169] No No , Rrp40p is encoded by YGR158C Humangene HumanEXOSC1 protein Yeast protein Identity / Archaeal similarity Identity / protein Human protein similarity Essential Domains Disease in yeast? role? EXOSC2 EXOSC3 EXOSC4 EXOSC5 EXOSC6 EXOSC7 EXOSC8 EXOSC9 EXOSC10 DIS3 DIS3L1 YHR069C encoded by Aeropyrum pernix protein and gene names are as follows – Csl4 encoded by as called by BLAST.disease. An Yeast sequence asterisk was (*)temperature-sensitive retrieved lethality. denotes from Yeast catalytic protein the and activity. gene names ‘Disease are role’ as follows is – in Csl4p is reference encoded to by documented cases of human Table 1.1: Exosomecharomyces cerevisiae Cross-Kingdom Protein Sequence Conservation. by Rrp42 encoded by Chapter 1. Introduction 23

a) b) 1

1 6 8

7 5

4 9 2 3 2 3

c) GM-CSF mRNA 5’ UU UU A A 3’ KSRP

DIS3 A C G 3’ - 5’ exonuclease activity of DIS3 releases nucleotides U

Figure 1.4: The Mammalian Exosome. (a) A top-down view of the mammalian exosome. The core ring of six proteins (EXOSC4, 5, 6, 7, 8, 9) is visible. (b) A top-down view of the core nine-member exosome with the three cap proteins (EXOSC1, 2, 3) in place. (c) A front-view schematic representation of the RNA-degradation role of the exosome. GM-CSF mRNA is bound by the pro-degradation accessory protein KSRP. KSRP binds directly to the 3’ AU-rich elements in the GM-CSF mRNA. Interaction between KSRP and EXOSC1, 2, or 3 mediates the recognition and transfer of the mRNA into the core exosome. The accessory protein DIS3 possesses 3’ - 5’ exoribonuclease activity, and processively degrades the mRNA, releasing individual nucleotides. DIS3 is depicted, but other 3’ - 5’ exoribonucleases can serve in its place (see Table 1.1 for alternatives). Chapter 1. Introduction 24 cycle [182]. Given the considerable homology both in sequence and function, it is possible that depletion of EXOSC2 in human cells may result in transcript stabilization of the same ontology groups, which are of considerable interest in human leukemias. Experi- ments exploring the consequence of exosome subunit knockdown in human cells have not yet been published.

1.3.5.2 Function of PRDM12

PRDM12 (positive regulatory domain member 12) is one of a 16-member family of PR domain-containing proteins (reviewed in [183, 184]). It possesses an amino-terminal PR domain and three tandem carboxy-terminal C2H2 zinc ﬁnger domains. The PR domain bears homology to the SET family of histone methytransferases, which function in chromatin-mediated regulation of gene expression [185]. PRDM12 is expressed in mouse neuronal development [186], cardiac myocytes, and possibly pancreas and liver cells (GeneAtlas U133A dataset, probe 220894 x at). Although little is known about PRDM12, several PR family members have been studied and have documented roles in hematopoiesis and human cancers. PRDM1 (BLIMP1) is a transcriptional repressor involved in B cell maturation [187], PRDM2 (RIZ) is capable of binding retinoblastoma protein (Rb) and its isoforms are thought to play a role in cancer development [188], PRDM3 (Mds/Evi1) is associated with therapy-related myelodysplasia and acute myeloid leukemia (AML) [189], and PRDM16 (MEL1) is associated with t(1;3) translocations observed in AML [190].

1.3.6 Historical therapies for CML

Treatment for leukemia began in 1865 shortly after its identification as a disease. A German case report documented treatment of leukemia with arsenious oxide, which appeared to lessen some of the physical symptoms including a reduction of splenomegaly, and improvements in anemia and white cell counts, but the effects were not long last- Chapter 1. Introduction 25 ing [191, 192]. Radiotherapy targeting the spleen and bone marrow was introduced in 1903 and induced remission in many patients [193], however, it did not prolong overall survival but did extend ‘efficient life’ by approximately 30% [192, 194]. Clinical trials in the early 1950s were conducted to compare the survival of patients treated with radiotherapy against a new alkylating agent called busulphan [195]. A survival improvement was observed with busulphan, which became the standard treatment for the following 35 years [192]. The DNA synthesis inhibitor hydroxyurea was used to treat busulphan- refractory leukemias starting in the early-to-mid 1960s [196], demonstrating a reduction in white blood cell count, splenomegaly, and thrombosis [197]. Busulphan treatment, and to a lesser extent hydroxyurea, was displaced by IFN-α in 1983 resulting substantial improvement in lifespan in the majority of patients [198]. IFN-α treatment is still the front-line treatment in many countries in which access to small-molecule inhibitors is limited.

1.3.7 Development and eﬃcacy of imatinib mesylate

Kinase inhibitors have faced considerable opposition as therapeutic agents based on claims of poor kinase selectivity, toxicity, and the therapeutic effect of targeting one kinase in a heterogenous disease [199]. In the early 1990s, Nicholas Lydon’s group performed a series of high-throughput screens that identified 2-phenylaminopyrimidine as a class of drug selective against platelet-derived growth factor receptor (PDGF-R) [200]. Lydon’s group collaborated with Brian Druker to test a panel of ATP-mimetic 2-phenylaminopyrimidine compounds against the tyrosine kinase ABL. STI571 (imatinib mesylate, IM) was discovered to have considerable cytotoxicity against BCR-ABL+ CML cells (reviewed in [201]). Early pre-clinical studies confirmed activity against PDGF-R, ABL, and c-kit, and in contrast to claims of broad kinase inhibitor specificity profiles, relatively few off targets (reviewed in [202]). In vitro colony formation and in vivo BCR-ABL+ mouse models demonstrated that IM had low toxicity in non-CML cells and Chapter 1. Introduction 26 efficient reduction of BCR-ABL+ cells [203]. Phase-I clinical trials began in June 1998 with an initial cohort of 54 CML patients who failed IFN-α therapy. After 12 months, 53 of 54 had patients achieved a durable complete hematologic response, and only one patient underwent disease relapse [204]. The phase II clinical trial mirrored the phase I results, and led to FDA-approval of IM in May 2001. IM remains the gold standard in CML treatment into 2012.

1.3.8 Development of imatinib mesylate-insensitive relapse

Imatinib mesylate clinical trial data demonstrated that drug-refractory relapse, although rare, did occur in a minority of patients. Shortly after IM FDA approval in May 2001, a study was published that examined ABL kinase domain mutations that prevent the binding of IM [205]. Before molecular patient relapse data was available, crystal structure- informed mutagenesis of the ABL kinase domain identified T315I as a mutation that abolished IM efficacy [205]. T315 is a residue that controls entry to the hydrophobic pocket; mutation to isoleucine efficiently blocks IM binding [199,205]. This mutation was subsequently discovered in patients [205] and was later demonstrated to prevent binding of all ATP-mimetic inhibitors against which it was tested [206, 207]. High-throughput screens selecting for inhibitor-resistant alleles of BCR-ABL have been tremendously successful in predicting clinical results [208–210]. Azam et al. used a random mutagenesis strategy to saturate BCR-ABL mutations, followed by a selection screen in BaF3 cells to isolate inhibitor-resistant BCR-ABL alleles [208]. Results mapped to the ABL kinase domain crystal structure demonstrated inhibitor-targeted residues were not clustered around the active site. This suggested mechanisms of resistance other than direct interference of inhibitor binding. Modification of SH2 and SH3 contact sites, destabilization of specific tertiary structures, and large-scale conformational changes were all cited as possible mechanisms of resistance [208]. The second screen to identify inhibitor-resistant ABL domain mutations was conducted with BaF3 cells and the mutagen N -ethyl-N - Chapter 1. Introduction 27

nitrosourea (ENU) [209]. Cells were selected in three kinase inhibitors, and mutations were cross-compared between drugs. T315I was the single mutation that conferred substantial resistance in all three inhibitors. Strikingly, the majority of mutations isolated from IM-treated CML relapse were observed in the screen, as well as a small number of novel mutations [209]. Such results demonstrate the power of in vitro screens in predicting clinical results.

1.3.9 Next-generation ABL kinase inhibitors

The ABL kinase crystal structure, combined with kinase domain mutations isolated from patients and mutagenesis screens, signiﬁcantly informed the design and development of next-generation inhibitors [199]. Nilotinib is based on the structure of IM but designed to bind more tightly to the ABL kinase domain [211]. Testing of nilotinib against a large panel of IM-resistant mutations demonstrated a signiﬁcant reduction in cell proliferation and BCR-ABL autophosphorylation, with the exception of T315I [211]. The nilotinib phase II study was conducted on 280 treatment-refractory CML patients with both ABL mutation-based (57%) and ABL mutation-independent (43%) causes of relapse [212]. At six months, 79% of patients achieved either a major or complete cytogenetic response [212].

Crystallographic data of IM-bound ABL depicted an inactive form of the kinase [213]. Mutations were later discovered in patients that resulted in a stabilization of the active conformation, explaining the reduction in IM efficacy [214]. A publication in 2000 described a kinase inhibitor effective against both SRC and ABL kinases [215]. Considering many kinases adopt the same active conformation, Brian Druker suggested that such a compound would likely target the active conformation of ABL, and thus be effective against a number of IM-selected ABL mutations that cause stabilization the the active form [199]. Crystallographic data confirmed the similar active conformations of SRC and ABL kinase domains [216], and biochemical and growth data demonstrated Chapter 1. Introduction 28 dasatinib sensitivity of many mutations, with the exception of T315I [217]. Dasatinib was shown to be effective in IM-refractory or intolerant patients in all stages of CML [218,219].

Despite the eﬃcacy of next-generation drugs on IM-insensitive BCR-ABL, the outstanding issue of T315I prevalence and patient prognosis is still of signiﬁcant concern. There have been several strategies devised to address this problem, three of which will be discussed here.

SGX393 was a rationally designed ATP mimetic, based on the crystal structure of the ABL kinase, to avoid the projection of I315 [220]. Treatment in vitro displays considerable reduction in BCR-ABL I315 autophosphorylation. In vivo eﬃcacy was observed through a reduction in mean tumour volume and silencing of pCrkL, a downstream target of BCR-ABL [220]. Unfortunately, no further studies examining SGX393 have been published.

A second strategy uses a novel class of drug called ‘switch-control’ inhibitors. DCC- 2036 was designed to target ABL R386/E282, residues critical in the ability of the kinase domain to switch between inactive and active conformations [221]. Treatment of BaF3 cells expressing BCR-ABL T315I display a dose-dependent reduction in both Stat5 and CrkL phosphorylation in response to DCC-2036. CML bone marrow transplant models expressing BCR-ABL T315I display a significant, yet not striking, difference in overall survival when treated DCC-2036 as compared to IM or dasatinib [221]. Despite this finding, the switch-control class of drugs and may present a promising new therapeutic opportunity for patients who previously had no options.

The third strategy uses allosteric inhibitors in combination with kinase inhibitors. GNF-2, a drug that binds to the myristoyl-binding pocket in the ABL kinase domain, induces a conformational change of the kinase active site [222]. An in vitro mutagenesis screen demonstrated that a combination of GNF-2 and nilotinib strongly suppresses the emergence of inhibitor-resistant mutations. In order to reduce toxicity and improve drug stability, GNF-5 was developed based on the GNF-2 structure. In vivo experiments Chapter 1. Introduction 29

demonstrate GNF-5 to be extremely eﬀective in prolonging life in a murine BCR-ABL T315I bone marrow transplant model. Treatment with vehicle, nilotinib, or GNF-5 alone resulted in the expected 20 to 30 day latency. However, treatment with GNF- 5 in combination with nilotinib resulted in an 80% survival rate through the 100-day study [222]. Such results are extremely promising for treatment of inhibitor-resistant BCR-ABL mutations.

There are several new drug classes that show promising results against common IM- resistant BCR-ABL mutations. Ultimately, the most eﬀective treatment will likely be combinatorial, targeting diﬀerent functions of BCR-ABL. In order to evade inhibition, BCR-ABL+ clones will need to express multiple simultaneous mutations resisting both (or more?) in a cocktail of inhibitors. Perhaps such a treatment regime will ultimately become front line to reduce development or selection of BCR-ABL inhibitor-resistant mutations.

1.4 Ph– myeloproliferative neoplasms – PV, ET, and

Dr. William Dameshek ﬁrst proposed the idea of a myeloproliferative neoplasm in 1951:

It is possible that these various conditions – ‘myeloproliferative disorders’ – are all somewhat variable manifestations of proliferative activity of the bone marrow cells, perhaps due to a hitherto undiscovered stimulus. This may affect the marrow cells diﬀusely or irregularly with the result that various syndromes, either clear-cut or transitional, result. Among them are the following: chronic granulocytic leukemia, polycythemia vera, idiopathic or ‘agnogenic’ myeloid metaplasia of the spleen (and liver), thrombocythemia, megakaryocytic leukemia and erythroleukemia (Di Guglielmo’s syndrome) [223]. Chapter 1. Introduction 30

Some of the diseases to which Dameshek was referring have become known as the classic Ph– MPNs; polycythemia vera (PV), essential thrombocythemia (ET), and myeloﬁ- brosis (MF).

1.4.1 MPN disease phenotypes

Polycythemia vera is an indolent disease characterized by the excess production of red blood cells. Median age of onset estimates range from 60 to 80 years of age [224, 225] and an incidence of 1.9 per 100,000 person-years [225]. Median patient survival is 19 years [226]. PV symptoms include itching in response to warm temperatures, burning in hands and feet, gouty arthritis, an enlarged spleen, and a high susceptibility to gastic ulcers. Thrombotic complications are more common than in the general population. Analysis of peripheral blood demonstrates a high hematocrit (HCT; packed red cell volume) and low serum EPO levels. Phlebotomy reduces total blood volume and HCT levels. Patients are often prescribed daily aspirin to reduce the chance of thrombic events, and some patients receive a chemotherapeutic such as hydroxurea, which appears to be a safe long-term treatment for MPNs [227]. ET has a median age of onset between 50 and 60 years and an incidence of 2 - 3 per 100,000 person-years [228]. ET is characterized by an elevated platelet concentration in the peripheral blood. The high platelet count results in increased splenomegaly and bleeding from the nose, gums, and gastrointestinal tract. Treatments are similar to PV, which include daily aspirin and hydroxyurea. Long-term patient survival approaches normal life expectancy [229]. MF is the most life-threatening of the Ph– MPNs. Estimates of incidence range from 0.3 to 1.5 per 100,000 person-years. Common symptoms include fatigue, bone pain, an enlarged spleen, and an increased susceptibility to infection [230]. Low-risk MF patients are typically followed without intervention. Intermediate and high-risk MF patients can be treated with splenectomy, radiotherapy, an allogeneic stem cell transplant, or experimental drug therapy [230]. Chapter 1. Introduction 31

1.4.2 Discovery of JAK2 V617F

Support for Dameshek’s hypothesis began to accumulate decades after his 1951 publication [223]. Bone marrow progenitors from PV patients differentiate in vitro to erythroid progenitor cells without supplemented cytokine [231], suggesting an ligand-free stimulation of the EPO pathway. BFU-E cells from MPN patients displayed increased sensitivity to IL-3 (38-fold), EPO (4.3-fold), and GM-CSF (48-fold), as compared to corresponding wild-type patient samples [232]. No differential effect was observed with IL-4, IL-6, or granulocyte colony stimulating factor (G-CSF) stimulation [232]. Megakaryocyte progenitor growth was analyzed in a spectrum of 42 MPN patients representing all four major MPN subtypes [233]. In both bone marrow serum-free agar culture and bone marrow plasma clot cultures, spontaneous megakaryocyte colonies were observed in the majority of samples. The addition of anti-IL-3, anti-IL-6, and anti-GM-CSF neutralizing antibodies did not reduce the number of colonies [233]. These results suggest the involvement of the Tpo-R, and that the megakaryocyte growth signal appears to be cell autonomous. Corroborating evidence for a kinase role in hematopoietic signal deregulation came from studies in the D. melanogaster JAK2 orthologue hopscotch (hop) [234–236]. A mutation called hopscotchTumorous-lethal (hopTum-1) was identified to be hopG341E, which occurs outside of the predicted kinase and kinase-like domains [234]. hopTum-1 expression induced a 4- to 8-fold increase in hemocytes (overall blood cells) and lamellocytes (larval stage-specific blood precursors) in larvae [235], remincient of human blood disorders. Over-expression of the hopE695K mutant protein induced a hyperphosphorylation of D-Stat. Generation of the homologus Jak2 E695K mutation in COS cells resulted in increased Jak2 autophosphorylation and activation of the downstreat Stat5 protein [236]. Both these studies suggested a role for the Janus kinase family in development of blood disorders. As an additional point of support, the molecular cause of CML was known to ABL, a precedent for a tyrosine kinase central to MPN development.

Dameshek’s hypothesis was proven to be correct in 2005 with the simultaneous pub- Chapter 1. Introduction 32

lication of four independent studies suggesting that an acquired mutation in the non- receptor tyrosine kinase JAK2 plays a role in MPNs [237–240]. Tony Green’s group used a candidate gene approach and directly sequenced JAK2 coding exons [237]. Of 140 patients, 104 expressed a G-T transversion at JAK2 1849 (JAK2 1849G>T), which coded for a valine to phenylalanine substitution at codon 617. The breakdown according to JAK2 V617F+ allele presence in MPN is as follows: 71 of 73 (97%) with PV; 25 of 51 (57%) with ET; 8 of 16 (50%) with MF. Gene dosage appeared to be signiﬁcant, as 19 of 73 (26%) of PV and 3 of 16 (19%) MF patients had >80% allele burden. None of the 51 ET patients displayed this phenotype. PV patients typically had a loss of heterozygosity (LOH) resulting in two identical JAK2 mutant alleles [237].

The Vainchenker group previously demonstrated that pharmacological- and siRNA- based attenuation of JAK2 activity suppressed spontaneous erythroid colony formation in vitro [241], suggesting a role for JAK2 in erythroid disorders. JAK2 exons from three PV patients were sequenced and the same 1849G>T mutation was observed [238]. A follow-up cohort of 45 PV patients was analyzed, 40 of whom were identiﬁed to be JAK2 V617F+. γ-2A cells (Jak2 –/–) [242] expressing exogenous Jak2 V617F and Epo-R were hypersensitive to Epo, as demonstrated by a Stat5-induced transcription of luciferase [238]. Cell growth assays suggest a moderate growth advantage in BaF3 Jak2 V617F+ cells in cytokine-free medium. Enhanced phosphorylation of Jak2 and Stat5 was also observed in the absence of stimulation, in contrast to BaF3 Jak2 wild-type cells. HCT was elevated in a Jak2 V617F BMT model (60%), as compared to unmanipulated (42%) and wild-type Jak2 (38%) [238].

Radek Skoda’s laboratory noted that FISH, comparative genomic hybrdization, and microsatellite screening identified LOH at chromosome 9p in PV patients [243–245]. Fur- ther refinement of the 9p LOH region with microsatellite markers allowed the identification of JAK2 as a candidate gene important in PV [239]. They discovered the JAK2 1849G>T mutation in 144 of 244 (50%) of tested MPN patients. LOH at 9p was observed Chapter 1. Introduction 33

predominantly in PV (43 of 128, 34%) and MF (5 of 23, 22%) patients. In contrast, only three of 93 (3%) ET patients displayed 9p LOH.

The Gary Gilliland group identified the JAK2 V617F mutation with high-throughput sequencing of 85 tyrosine kinase activation loops and autoinhibitory domains [240]. The JAK2 1849G>T mutation was a recurrent observation in the initial 93 samples analyzed. In total, 345 samples were sequenced for the JAK2 V617F mutation. Heterozygosity was observed in 126 (36%), and homozygosity in 48 (14%). The breakdown of zygosity (total samples, percent hetero / percent homo) is as follows: PV – 164, 49% / 25%; ET – 115, 29% / 3%; MF – 46, 26% / 9%. In vitro 293T experiments demonstrated JAK2 V617F to be constitutively phosphorylated, and not inhibited by increased expression of the wild- type allele, suggesting no dominant negative activity. Unstimulated BaF3 Epo-R cells that expressed exogenous Jak2 V617F had a low level of growth and were hypersensitive to low concentrations of Epo [240]. This experiment corroborated the hypersensitivity finding of James et al. [238] and early observations of Epo-independent colony growth of primary human PV bone marrow [231]. Sequencing of myeloid leukemia cell lines revealed that HEL, a human erythroleukemia cell line [246], was homozygous for the JAK2 V617F mutation [240]. Constitutive phosphorylation of STAT5 and ERK1/2, two essential downstream effectors of JAK2 signalling, was observed. Incubation of HEL cells, but not K562 (BCR-ABL+) [247], with JAK Inhibitor-I demonstrated a dose-dependent decrease in cellular proliferation, and JAK2 and STAT5 phosphorylation. Treatment with JAK Inhibitor-I also induces selective apoptosis in HEL cells, suggesting they are dependent on activated JAK2 signalling for survival [240].

In summary, these four studies identiﬁed the JAK2 1849G>T and corresponding JAK2 V617F mutation to be prevalent in MPN patients. The majority of PV patients carry the V617F mutation, a subset of whom have undergone mitotic recombination at chromosome 9p resulting in LOH and two JAK2 V617F alleles. The majority of PV patients carried the JAK2 V617F allele, many of which were homozygous; approximately Chapter 1. Introduction 34

half of ET patients carried the mutation, and V617F homozygosity is extremely rare; half of myeloﬁbrosis patients had the mutation and a small minority in a homozygous fashion [237]. BaF3 cells that expressed exogenous JAK2 V617F were able to proliferate independent of cytokine and were hypersensitive to small doses of Epo. JAK2 V617F is constitutively phosphorylated in vitro and responds to increasing doses of a JAK-speciﬁc inhibitor through a reduction in JAK2 and STAT5 phosphorylation. Short-term murine bone marrow transplant experiments recapitulate the PV-like increase in HCT [238].

1.4.3 Genetic diversity among Ph– MPN disease

After the discovery of the JAK2 V617F mutation in the majority of Ph– MPN patients, there may have been an assumption of genetic uniformity, much like the BCR-ABL / CML paradigm. It soon became clear that this group of diseases was far more genetically heterogenous than CML. The following is a review of selected non-V617F mutations that are commonly found in Ph– MPNs. A particularly interesting note is the majority of Ph– mutations fall into one of two categories – activation of the JAK-STAT pathway (JAK2 V617F, JAK2 exon 12, MPL, and LNK ) or aberrant epigenetic modiﬁcation (TET2, ASXL1, and EZH2 ) [248]. A combination of mutations in these genes and environmental factors are likely the determining factors in development one MPN against another.

1.4.3.1 JAK2 exon 12 mutations

A small minority of PV patients do not express the JAK2 V617F mutation. Sequencing of the remainder of the JAK2 gene in this patient group yielded mutations in exon 12 [249,250]. JAK2 F537-K539delinsL, H539QK539L, N542-E543del, and K539L were exon 12 mutations discovered in V617F– PV patients, and all conferred factor-independent growth in a BaF3 Epo-R cell line model [249]. Exon 12 codes for a region linking the SH2 and JH2 domains. Using projected structures, it was hypothesized that exon 12 deletions may relieve the auto-inhibitory role of the JH2 domain, an JH1 activation Chapter 1. Introduction 35

mechanism similar to that suggested for V617F [251]. Approximately 3% of PV patients express JAK2 exon 12 mutations [248].

1.4.3.2 MPL (Tpo-R) mutations

The Epo- and Tpo-stimulated pathways bear some overlap, as discussed in section 1.1.4. This observation led to sequencing of genes important in thrombopoietic signalling in JAK2 V617F-negative ET and MF patients [252]. Sequencing TPO-R yielded a recurrent W515L mutation in a small proportion of MF patients. Expression of the TPO-R W515L mutation in 32D cells [253] demonstrated constitutive phosphorylation of JAK2, STAT3, STAT5, AKT, and ERK1/2, reminiscent of the eﬀect of JAK2 V617F. Exami- nation of TPO-R W515L in a BMT model led to a 17 to 32 day latency, leukocytosis, thrombocytosis, splenomegaly, and bone marrow ﬁbrosis, which recapitulates many of the symptoms of MF and late-ET disease [252]. TPO-R mutations are present in approximately 3% of ET and 10% of MF patients [248].

1.4.3.3 LNK mutations

LNK (SH2B3) is an adaptor protein that contains an amino-terminal dimerization domain, PH and SH2 domains, and a conserved C-terminal tyrosine residue [254]. Lnk was initially identiﬁed to have a role in B cell development [254], but later shown to negatively regulate Jak2 downstream of the Tpo-R in stem cell self-renewal [255, 256]. Lnk –/– mice display aberrant cytokine signalling, an expanded stem cell compartment, megakaryocytic hyperplasia, splenomegaly, leukocytosis, and thrombocytosis [257–259]. This phenotype strongly suggests that Lnk is a tumour suppressor gene in MPN-like disease.

Given the disease developed by Lnk-null mice, LNK was an obvious candidate in JAK2 V617F– human MPNs. A 5 bp deletion (603 607delGCGCT; 613C>G) in LNK was identiﬁed in an MF patient, leading to a premature stop codon and loss of the PH and Chapter 1. Introduction 36

SH2 domains [260]. A second LNK mutation was identiﬁed in an ET patient (622G>C), leading to a E208Q mutation. Both mutations gave a signiﬁcant growth advantage in a BaF3 Tpo-R cell line model [260]. LNK mutations are observed in a small minority of Ph– MPNs, but are acquired as MPNs progress into more serious disorders [248].

1.4.3.4 TET2 mutations

Chromosome 4q24 was initially identified in high-throughput genomic arrays attempting to discover common contributing elements in myeloid malignancies [261, 262]. Fine mapping suggested TET2 was the only gene to map to the region [262]. Upon further investigation, mutations in TET2 have been identified in AML (12% of patients) [263–265], MPNs (8% overall, 17% in JAK2 V617F+, 7% in JAK2 V617F–) [261,264,266,267], MDS (19 - 26%) [267–270], chronic myelomonocytic leukemia (CMML; 42%) [270], and systemic mastocytosis (29%) [271]. Frameshift, nonsense, and missense mutations have been identified and proposed to inactivate TET2 [263, 264]. The TET (ten-eleven translocation) family of proteins consists of three highly conserved members; TET1, TET2, and TET3 [261,262]. TET1 was identified as a fusion partner of the mixed lineage leukemia (MLL) gene, recurring in AML [272, 273]. The function and significance of TET3 is largely unknown. The TET family of proteins was identified as mammalian homologues of the Trypanosome proteins JBP1 and JBP2, which potentially modify DNA though the 5-methyl group of thymine [274]. Evidence suggests TET1 catalyzes the conversion of 5-methylcytosine (5mC) to 5-hydroxymethylcytosine (hmC) in the stem cell compartment, and thus has a potential role in epigenetic regulation [275]. RNA-modulation experiments suggest that downregulation of TET1 in the murine hematopoietic stem cell compartment results in a decrease of hmC6 [275]. TET2 mutations may induce DNA hypermethylation, resulting in gene silencing and impaired myelopoiesis [248]. The role of TET2 in disease progression varies according to disease. A TET2 mutation appears to be a favourable prognostic factor in MDS [270]. TET2 mutation status does not act Chapter 1. Introduction 37 as a prognostic indicator in PV [264], MF [264], systemic mastocytosis [271], MDS [267], or MPNs in general [267]. However, TET2 mutations appear to be a negative prognostic factor in CMML [270].

1.4.3.5 ASXL1 mutations

The D. melanogaster orthologue of ASXL1 (additional sex combs like 1), asx, encodes a protein that functions in chromatin remodelling to both activate and silence drosophila patterning loci [276]. Human ASXL1 is not well studied. Asxl1, the mouse orthologue, repressed expression of adipogenic genes through regulation of peroxisome proliferator- activated receptor gamma [277]. Asxl1 –/– mice display defects in myeloid and lymphoid progenitor diﬀerentiation, but do not display a leukemic phenotype [278], suggesting human ASXL1 mutations occur late in disease progression or collaborate with mutations in other genes. Mutations in ASXL1 are common in MPNs – 3% of ET, 13% of MF, 18% of blast-phase MPN, 11% of AML, 11% of MDS, and 43% of CMML [248]. Further studies will be required to elucidate the role of ASXL1 in hematopoietic malignancies.

1.4.3.6 EZH2 mutations

Chromosome 7q is frequently lost in AML and MDS, and is a poor prognostic factor [279]. Acquired uni-parental disomy (aUPD) of 7q36 is seen in MDS and MPNs, suggesting a role in disease [280]. High-resolution copy number analysis identiﬁed one subject with a 400 kb microdeletion at 7q36.1, the EZH2 locus [280]. Expansion of the initial cohort to 249 individuals yielded a 13% EZH2 mutation frequency in MPNs. EZH2 is a histone methyltransferase and a subunit of the polycomb repressive complex 2 (PRC2). The PRC2 is important in epigenetic silencing of genes important in cell fate decisions [281,282]. Ezh2 is essential for mouse development. Eighty-two percent of Ezh2 –/– embryos fail to implant or cannot complete gastrulation [283]. Ezh2 +/– embryos also develop problems during gestation, but 50% are viable [283]. Sequencing of the hu- Chapter 1. Introduction 38

man EZH2 gene in MPNs identified missense, deletion, and splice site mutations [280]. Ezh2 mutation homozygosity was associated with a worse survival than heterozygous or wild type, although the difference was not significant (p = 0.089) [280]. Expression of patient mutations in cell lines results in a reduction of histone methyltransferase activity [280]. EZH2 mutations have been reported in 6% of MDS [284]. EZH2 mutations are also observed in 3% of PV, 14% of blast-phase MPN, 22% of AML, and 8% of MDS [248].

1.4.3.7 JAK2 haplotype as a disease predictor

MPNs are typically sporadic but there are cases that appear to have a familial component [285–287]. Three publications in 2009 demonstrated that a specific JAK2 haplotype block (46/1, rs10974944, GGCC) was an extremely significant risk factor to development of JAK2 V617F+ disease [288–290]. Inheritence of this block results in a three- to four-fold higher risk of developing an MPN, and it is more likely to acquire the V617F mutation in cis. Two hypotheses to explain this finding were put forward in the studies and summarized [291]. Is it possible that the haplotype block confers a hypermutability to JAK2 [289, 290]. Mutations would be acquired at a higher rate, and those that confer a competitive advantage would be selected. The second was that the JAK2 V617F mutation occurs at equal frequencies across all haplotypes, but the haplotype confers a stronger selective advantage [288]. No current publications to date give preferential support to the first or second hypothesis. A publication examining the 46/1 haplotype block frequency in ET found that 46/1 was over-represented in V617F– disease [292], but does not offer additional evidence to differentiate between the two hypotheses.

1.4.3.8 Clonal diversity within disease and individual patients

In contrast to other malignancies, the order of mutational events in Ph– MPNs can be variable [293]. Mutations in TET2 can be an early [262] or late [294] event in disease. There are also cases of JAK2 V617F+ disease progressing to JAK2 V617F– AML, which Chapter 1. Introduction 39 may be reflective of clonal diversity [293,295]. Seven of 1000 screened MPN patients had bi-clonal disease, which is unlikely to be coincidental and may reflect an underlying pro- neoplastic genotype [296]. There are two models to explain this finding [293]. First, both the AML and MPN arose from a pre-JAK2 mutant clone. Second, the AML and MPN arose from independent stem cells. The particular etiology of disease may be variable among patients, with both explanations represented.

1.4.4 Inhibitors of JAK2 in the treatment of MPNs

Inhibition of JAK2 as a clinical target became attractive after the discovery of JAK2 V617F in 2005. Since, many JAK2-selective compounds have been developed and more than a dozen have shown efficacy in vitro [297]. The model for JAK2 inhibitors is the remarkable success of IM-targeted BCR-ABL. Unfortunately, the considerable genetic diversity in the Ph– MPNs (section 1.4.3) suggests targeting JAK2 to be of less clinical relevance than BCR-ABL in CML [248]. Combined with the essential nature of JAK- STAT signalling in hematopoietic development (section 1.1.3), the probability of JAK inhibitor long-term success is currently under debate. However, the treatment options of MF patients are currently limited, giving JAK2 inhibitors immediate clinical value in the management of symptoms. MF patients may benefit from JAK2 inhibition through directly modulating the pro-growth signals of the JAK-STAT pathway, and from down- regulating specific pro-inflammatory cytokines produced by the affected clone. There are currently five JAK2-selective compounds that have progressed to phase II clinical trials, one of which (ruxolitinib) received FDA approval in November 2011. Below is a review of pre-clinical and clinical efficacy of ruxolitinib (INCB018424), SAR302503 (TG101348), lestaurtinib (CEP-701), CYT387, and SB1518. Chapter 1. Introduction 40

1.4.4.1 Ruxolitinib (INCB018424)

Ruxolitinib (INCB018424, trade name Jakaﬁ) is a compound developed by Incyte Phar-

maceuticals and Novartis, which targets JAK1 (IC50 = 3.3 nM) and JAK2 (IC50 = 2.8 nM) [298]. BaF3 cells that expressed Epo-R and Jak2 V617F displayed a dose-dependent reduction of pJak2, pStat5, and pErk1/2 [298]. In a cell growth assay, BaF3 cells that expressed Epo-R and Jak2 V617F were inhibited at at 3 nM INCB018424, compared to little inhibition of BaF3 cells that expressed BCR-ABL. Propidium iodide and annexin V analysis suggest that the reduction of cell growth is due to INCB018424-induced cytotoxicity. In mice receiving intravenous injection of BaF3 cells expressing Epo-R and JAK2 V617F, treatment with INCB018424 induced a reduction in spleen weight and an increase in overall survival. Splenic architecture was also maintained in the INCB018424-treated cells. Signiﬁcant for MF patients, circulating IL-6 and tumour necrosis factor alpha (TNF-α) levels were signiﬁcantly reduced [298].

The phase I/II study involving 153 JAK2 V617F-positive or negative MF patients began in 2009 and ran 14.7 months [299]. The dose escalation portion identified a 15 mg BID regimen to be optimal, with 50% of patients displaying a rapid reduction in splenomegaly, and alleviation of rapid weight loss, fatigue, and night sweats. Plasma levels of of cytokines and inflammatory factors were reduced, including 1- to 3-log reduc- tions in IL-6 and TNF-α, among others [299]. This finding was JAK2 status independent, suggesting an overall reduction in JAK-STAT signalling and pro-inflammatory cytokines. The initial phase III study was conducted with 309 intermediate-2 or high-risk MF patients [300, 301]. A reduction in spleen size (>35%) was the primary endpoint, which was achieved in 42% of patients. There was no significant difference in overall survival over the course of the trial, however, the drug-treated cohort displayed more anemia and thrombocytopenia. A second phase III trial added 219 patients to the study and achieved similar results. Chapter 1. Introduction 41

1.4.4.2 SAR302503 (TG101348)

TG101348 was intitailly developed by TargeGen, but acquired by Sanoﬁ-Aventis and rebranded as SAR302503. TG101348 inhibits JAK2 (IC50 = 3 nM) and FLT3 kinase

(IC50 = 15 nM) [302]. TG101348 inhibited HEL cells and BaF3 cells expressing JAK2 V617F in a dose-dependent manner with a reduction in pSTAT5 [302]. A Jak2 V617F BMT model displayed increased survival when treated with 120 mg/kg, and had an ac- companying reduction in spleen weight, white blood cells (WBC) per µL, and HCT [302]. Transplantation of human PV progenitors in a xenograft model displayed a reduction in engraftment when treated with TG101348 [303]. Xenograft treatment was accompanied by lower GATA-1 expression, a gene that plays a role in erythroid diﬀerentiation in JAK2 V617F-mediated disease [303].

A phase I/II study of TG101348 included 59 high- or intermediate-risk MF patients in total, with 28 in the dose-escalation phase [304]. Median length of treatment was 12.5 months. More than half the patients report marked improvement of night sweats, fatigue, low appetite, and cough. By 12 treatment cycles, approximately 50% of patients achieved an adequate spleen response, and a reduction in leukocytosis and thrombocytosis. No- tably, a significant decrease in JAK2 allele burden was observed, which was durable at 12 months [304]. Negative side effects include anemia (35%), thrombocytopenia (24%), and neutropenia (10%). Other adverse effects were nausea (70%), diarrhea (64%), and vomiting (35%).

1.4.4.3 Lestaurtinib (CEP-701)

CEP-701 was originally identiﬁed as a FLT3 kinase inhibitor that was cytotoxic to AML blasts harbouring FLT3 internal tandem duplications (ITD) [305]. Further studies demonstrated in vivo eﬃcacy in Balb/c mice injected with BaF3 FLT3-ITD cells [306].

It was later discovered that CEP-701 inhibits JAK2 kinase (IC50 = 1 nM) [307]. An MTT survival assay demonstrated a dose-dependent reduction of HEL cell proliferation, Chapter 1. Introduction 42

with an absence of growth at 1 nM CEP-701 [307]. A reduction in pJAK2, pSTAT5, and pSTAT3 was also observed. The effects of CEP-701 in vitro were examined in a set of primary PV samples, which recapitulated the growth and signalling data from previous cell line experiments, including a reduction in pERK1/2, pAKT, and pSHP2 [307]. A phase-II study was undertaken with 22 JAK2 V617F+ MF patients [308]. Of the 22 patients, four had a spleen size reduction, two became transfusion independent, and one had improvement in WBC count. Unlike either of the two drugs discussed previously, CEP-701 did not induce an improvement in JAK2 V617F allele burden or serum levels of pro-inflammatory cytokines. Substantial side effects were observed, including anemia (14%) and thrombocytopenia (23%).

1.4.4.4 Momelotinib (CYT387)

CYT387 was developed by Cytopia, and targets JAK1 (IC50 = 11 nm), JAK2 (IC50

= 18 nM), and TYK2 (IC50 = 17 nM) [309]. In HEL cells, pSTAT3 and pSTAT5 became undetectable at 5.0 and 1.0 µM, respectively [309]. CYT387 suppressed erythroid colony growth in JAK2 V617F+ primary PV patient cells [309]. CYT387 treatment of BMT mouse models reduced WBC count, HCT, and spleen size. Similar to other JAK2 inhibitors, serum cytokine levels were normalized [310]. In vivo clone re-growth occurred with drug cessation, suggesting that CYT387, like all other JAK2 inhibitors, cannot eliminate the disease-causing clone [310]. Treatment with CYT387 resulted in a reduction of IL-6-induced pSTAT3, a result that was recapitulated when primary samples were examined [311]. The CYT387 phase I/II clinical trial is currently ongoing, but an interim report has been issued [304] and summarized in [248]. At 6.4 months, 92% of patients were still enrolled and receiving treatment. Spleen response was 45% and the majority of patients experienced an easing of itch, night sweats, and bone pain. Thrombocytopenia was observed in 25% of patients. Chapter 1. Introduction 43

1.4.4.5 Pacritinib (SB1518)

SB1518 (trade name Pacritinib, recently acquired by Cell Therapeutics) is inhibitory

to wild-type JAK2 (IC50 = 23 nM), JAK2 V617F (IC50 = 19 nM), and FLT3 (IC50 = 22 nM) [312, 313]. SB1518 induced a reduction of pJAK2, pSTAT3, and pSTAT5 in 293T and HEL cell models [313]. Nude mice were injected with BaF3 JAK2 V617F cells and treated with SB1518, which caused a dose-dependent reduction in splenomegaly and liver weight. Platelet count and hematocrit were not aﬀected. SB1518 reduced tumour volume in a SET-2 (expressing JAK2 V617F [314]) cell xenograft model [313]. Tumour cells were isolated and probed for pJAK2, and pSTAT5, both of which were reduced. SB1518 did not inhibit PV cell viability over cells from healthy volunteers, and did not reduce erythroid colony formation compared to myeloid [313]. Extensive structural modelling has been performed with SB1518 complexed with JAK2 [315], which will be informative should any inhibitor-resistant mutations appear in SB1518-treated patients.

SB1518 has successfully finished phase II clinical trials for myelofibrosis and lymphoma [315] but the results have not yet been published. A JAK2 kinase inhibitor review [248] discussed data presented at the American Society of Hematology annual meeting [316], and will be briefly explained here. The phase I study was completed with 36 MF patients. An in vivo reduction of pJAK2, pSTAT3, and pSTAT5 was observed. Of the 33 MF patients with palpable splenomegaly, 57% had a spleen response as evaluated by MRI. Constitutional symptoms were also reduced by >50%.

1.4.5 Mouse models for development and treatment of Ph– MPNs

After discovery of the JAK2 V617F mutation in 2005, multiple groups examined the consequence of Jak2 V617F bone marrow expression using the BMT model (section 1.3.4), while in parallel generating conditional transgenic models. The initial BMT models used Jak2 V617F-trandsuced bone marrow to reconstitute a lethally irradiated recipient. Peripheral blood, spleen size, tissue infiltration, and hematopoietic tissue fibrosis were Chapter 1. Introduction 44 monitored and compared to the human MPN phenotype. Transgenic models used elegant strategies to induce narrow hematopoietic expression of the Jak2 V617F allele. Below is an detailed analysis of the first BMT models and the later transgenic models of MPN development.

1.4.5.1 BMT models

Wernig et al. used a BMT model to determine the effect of Jak2 V617F hematopoietic expression [317]. They observed a significant increase in spleen weight on both Jak2 V617F+ Balb/c and C57Bl/6 mouse strain backgrounds, which was accompanied by an increase in bone marrow reticulin fibrosis, as compared to wild type [317]. Both strains had significant increases in WBC counts and HCT. Balb/c mice displayed an approximate doubling of platelet count while C57Bl/6 had no change. Jak2 V617F expression resulted in an increased number of myeloid progenitors observed in the bone marrow, effacement of the splenic architecture, and extensive hematopoiesis in the liver. Tri-lineage hyperplasia was common in all three of those tissues. High HCT and low serum Epo levels were observed. Analysis of bone marrow and spleen cells demonstrate constitutive Jak2 and Stat5 phosphorylation. Kaplan-Meier survival plots show 5-FU- treated Balb/c Jak2 V617F latency between 50 and 95 days, and 5-FU-treated C57Bl/6 between 65 and 75 days. These results demonstrated that Wernig et al. generated a murine BMT model that recapitulated many of the features of human PV and MF [317].

Lacout et al. used a similar strategy to examine Jak2 V617F expression in mouse bone marrow [318]. They observed constitutive Jak2 V617F activity in primary spleen cells, and hypersensitivity to Epo or IL-3 stimulation, as determined by Erk1/2 phosphorylation. Tri-lineage hyperplasia in the bone marrow and spleen was common in all primary recipients. Interestingly, after 3 to 4 months post-transplant, a drop in the hemoglobin concentration was accompanied by an increase in reticulin ﬁbres in the bone marrow and spleen. The increased bone marrow ﬁbrosis resulted in thrombocytopenia Chapter 1. Introduction 45

and anemia, characteristic of human MF. Overall, Lacout et al. demonstrated the JAK2 V617F BMT model recapitulated elements of human PV and the transition of post-PV MF [318].

In a third study, Bumm et al. conducted a similar Jak2 V617F BMT experiment [319]. Balb/c mice demonstrated erythrocytosis and leukocytosis caused by expansion of CMP, GMP, and MEP progenitor populations. Extramedullary hematopoiesis and bone marrow fibrosis were also observed, but thrombocytosis was absent [319]. Findings by Zaleskas et al. confirmed the PV-like phenotype observed in the BMT model [320]. Increased erythrocyte count, hematocrit, hemoglobin, and reticulocytosis were observed. Both Balb/c and C57Bl/6 were examined, and the high WBC count was much more apparent in the Balb/c strain. As in Lacout et al. [318], the PV-like phenotype subsided over a few months and evolved into a more MF-like disease with significant collagen deposits in the bone marrow and spleen [320].

Interestingly, none of the early BMT models demonstrated an ET-like phenotype, suggesting a more regulated Jak2 V617F expression is required for megakaryocyte proliferation. The diﬀering results based on strain background means there are additional genetic modiﬁers that contribute to the V617F-induced phenotype, which may also play a role in a lack of ET-like disease. Should drug-resistant JAK2 V617F alleles be discovered in response to treatment, these models would be used to examine their functional consequence.

1.4.5.2 Transgenic models

Transgenic models were required to recapitulate an ET-like phenotype and determine the eﬀect of a physiological JAK2 V617F expression level. Tiedt et al. published an elegant solution to the problem of JAK2 V617F expression regulation [321]. A bacterial artiﬁcial chromosome (BAC) was created that contained human JAK2 V617F driven by the endogenous JAK2 promoter. The JAK2 V617F mutation was contained in an Chapter 1. Introduction 46

inverted region flanked by antiparallel lox71 and lox66 sites, which ensured absence of expression without cre-mediated recombination. BAC microinjection into oocytes resulted in a strain carrying nine tandem repeats of the construct. Such an integration pattern resulted in a unique opportunity – each copy could individually recombine, resulting in very high JAK2 V617F expression (nine copies), or recombination between outer-most lox sites would result in a single copy of JAK2 V617F. This strain, called Flip-Flop1 (FF1 ), was crossed with mice expressing cre recombinase under the vav [322] or Mx1 [323,324] hematopoietic-restricted promoters. The vav-cre / FF1 strain expressed JAK2 V617F at a lower level than wild-type JAK2, which resulted in an ET-like phenotype. Recombination level was modulated with pI:pC dose in the Mx1-cre / FF1 model, which resulted in a higher JAK2 V617F-to-JAK2 wild type ratio and a PV-like phenotype. Both the vav-cre and Mx1-cre FF1 strains were compared to a BMT model of Jak2 V617F hematopoietic expression, which generated the PV-like phenotype observed in the first BMT models [317, 318]. Both FF1 strains demonstrate a high hemoglobin count, platelet concentration, and spleen weight. Tri-lineage hyperplasia in the bone marrow, and fibrosis in the bone marrow and spleen was observed. Perhaps most significantly, Tiedt et al. demonstrate that a ratio between human JAK2 V617F and wild-type JAK2 expression to be highly correlative with the type of MPN that develops. Development of human ET is associated with a ratio of 0.5 - 1 : 1 (V617F : WT), while PV and MF require a much higher ratio of 1 - 10 : 1 [321]. In summary, this transgenic model recapitulates the ET-like phenotype, which was not present in the early BMT models. It also allows for the modulation of JAK2 V617F expression from its endogenous promoter, which closely resembles elements of the human disease.

Akada et al. introduced a wild-type Jak2 cDNA construct into endogenous Jak2 intron 12, which contained a splice acceptor and exons 13 to 24 ﬂanked by loxP sites [325]. Immediately downstream, exon 13 contained the Jak2 V617F mutation. Under normal conditions, transcription would occur of endogenous Jak2 from exons 1 to 12 , followed by Chapter 1. Introduction 47

exons 13 to 24 of the transgenic cDNA, ultimately producing a wild-type Jak2 protein. When crossed with a cre-expressing mouse, the exogenous 13 to 24 cDNA would be spliced out, resulting in transcription of endogenous exons 1 to 12, the transgenic exon 13 containing Jak2 V617F, and endogenous exons 14 to 24, resulting in expression of Jak2 V617F. Mating the Jak2 transgenic with an Mx1-cre mouse, followed by pI:pC treatment to induce recombination, resulted in a heterozygote that recapitulated much of the human PV phenotype. Increased hemoglobin and HCT, tri-lineage hyperplasia, splenomegaly and low serum Epo concentration were observed. Examination of bone marrow Jak2 and Stat5 demonstrate constitutive phosphorylation in the absence of stimulation [325].

Mullally et al. modified the enodgenous Jak2 locus to contain an inverted exon 14, expressing the Jak2 V617F mutation [326]. Cre-mediated recombination resulted in ex- cision of the endogenous exon 14 and flipping of the transgenic exon 14, resulting in expression of the Jak2 V617F allele. Jak2 V617F animals were crossed with E2A-cre transgenics to generate a heterozygous germline Jak2 V617F knock-in model. The median survival was 146 days, and disease was characterized by elevated HCT, splenomegaly, and extramedullary hematopoiesis. Reticulin fibrosis was not observed. Bone marrow fractionation and transplantation demonstrated disease could be transferred to secondary recipients using only the primitive LSK (lineage–, sca-1+, c-kit+) cell population, and not more committed megakaryocyte-erythrocyte progenitor (MEP) or GMP progenitors. This result demonstrated that the Jak2 V617F disease-initiating population is contained within the most primitive progenitors. Bone marrow mixing experiments were conducted in which different ratios of wild-type Jak2 and Jak2 V617F LSK cells were mixed and transplanted into lethally irradiated recipients. Sixteen weeks post-transplant, no Jak2 V617F competitive advantage was observed. Treatment of mice with a JAK2 inhibitor demonstrated a reduction in spleen weight and hematocrit, but not elimination of the disease-initiating cells. This transgenic model does not recapitulate some elements of the human disease, but is significant in understanding the disease-initiating populations and Chapter 1. Introduction 48

their relative growth ability [326].

Development of transgenic models allow us to examine the MPN phenotype in greater depth. An ET-like phenotype was observed for the first time, with evidence to support the gene-dosage model of MPN disease stratification. Cre-inducible models allow for the examination of temporal expression of the Jak2 V617F allele, and its affect in hetero or homozygosity. These models more faithfully recapitulate elements of the human disease, but may not be suitable for more high-throughput experiments such as testing inhibitor- resistant alleles of Jak2.

1.5 Rationale and hypothesis

Progression to better understanding and treatment of myeloproliferative neoplasms requires extensive study from both fundamental and clinical perspectives. This thesis aims to address both areas of research by asking two related but distinct questions. The first concerns the prediction of clinical resistance to small-molecule inhibitors of JAK2, and how that information can be used to inform decisions in drug design. The second explores the underlying cause of poor treatment response in a specific patient cohort, and how those differences are driven on a cellular level.

Chapter 2

Chronic myeloid leukemia patients treated with the rationally designed small-molecule inhibitor imatinib mesylate (IM) can relapse due to BCR-ABL kinase domain mutations that prevent inhibitor binding. The inhibitor exerts considerable selective pressure on the BCR-ABL+ clone to evade inhibition. In vitro screens to isolate inhibitor-resistant alleles have demonstrated a striking overlap with clinical results, which allows for informed design of next-generation ABL kinase inhibitors. We hypothesized that similar to IM-treated BCR-ABL, JAK2 inhibitor-treated myeloproliferative neoplasms will evade inhibition due to selection of an inhibitor-resistant clone. To test this hypothesis, we de- Chapter 1. Introduction 49

veloped methods to screen and test inhibitor-resistant alleles of JAK2.

Chapter 3

At the time of the CML BCR-ABL translocation, 10 - 20% of patients undergo a genomic deletion at the 9q34 breakpoint. Patient segregation based on 9q34 status demonstrates the deletion is associated with a signiﬁcantly worse progression-free and overall survival, as compared to the non-deleted cohort, when treated with traditional chemotherapeutics, IFN-α, or a bone marrow transplant. We hypothesized that this diﬀerence was due to either a pre-existing genomic instability, which was tested by examining copy number aberrations in a 9q34+ and 9q34– patient cohort; or due to hemizygous loss of a tumour suppressor gene, which was tested with molecular and biochemical assays to determine functional relevance in disease progression.

1.6 Thesis statement and study aims

Chapter 2 – JAK2 will evade small molecule inhibition through mutations that modify inhibitor binding.

Chapter aims:

1. Screen for JAK Inhibitor-I-resistant JAK2 alleles using TEL-JAK2 random mutagenesis in a soft agar system.

2. Characterize the impact of TEL-JAK2 mutations on cellular growth, downstream signalling, and kinase function in high doses of inhibitor.

3. Verify inhibitor resistant mutations in the more clinically relevant Jak2 V617F.

4. Map the mutations to the published crystal structure to provide insight as to the mechanism of inhibitor resistance. Chapter 1. Introduction 50

Chapter 3 – 9q34– chronic myeloid leukemia patients fare worse on traditional therapies because of a pre-existing genomic instability or loss of one or more tumour suppressor alleles (EXOSC2 and/or PRDM12 ).

1. Examine genomic copy number aberrations between 9q34+ and 9q34– patients using array comparative genomic hybridization.

2. Determine expression of genes that map to the 9q34 locus in CML cell lines, human patient samples, and mouse hematopoietic progenitors.

3. Develop and characterize an anti-EXOSC2 antibody in order to conduct functional studies on protein expression.

4. Identify whether coding mutations exist in genes that map to 9q34. Chapter 2

Random mutagenesis reveals residues of JAK2 critical in evading inhibition by a tyrosine kinase inhibitor

Michael R. Marit, Manprit Chohan, Natasha Matthew, Kai Huang, Douglas A. Kuntz, David R. Rose, and Dwayne L. Barber. (2012) Random Mutagenesis Reveals Residues of JAK2 Critical in Evading Inhibition by a Tyrosine Kinase Inhibitor. PLoS ONE 7(8): e43437. doi:10.1371/journal.pone.0043437

Author contributions: M.R.M., M.C., N.M., and D.L.B. designed experiments; M.R.M., M.C., and N.M. conducted experiments; K.H. generated the vector constructs; D.A.K. and D.R.R. advised on the structural modeling; M.R.M. and D.L.B. wrote the manuscript

51 Chapter 2. JAK2 mutagenesis and inhibitor resistance 52

2.1 Abstract

The non-receptor tyrosine kinase JAK2 is implicated in a group of myeloproliferative neoplasms including polycythemia vera, essential thrombocythemia, and primary myeloﬁ- brosis. JAK2-selective inhibitors are currently being evaluated in clinical trials. Data from drug-resistant chronic myeloid leukemia patients demonstrate that treatment with a small-molecule inhibitor generates resistance via mutation or ampliﬁcation of BCR-ABL. We hypothesized that treatment with small-molecule inhibitors of JAK2 will similarly generate inhibitor-resistant mutants in JAK2. In order to identify inhibitor-resistant JAK2 mutations before they appear in the clinic, we utilized TEL-JAK2 to conduct an in vitro random mutagenesis screen for JAK2 alleles resistant to JAK Inhibitor-I. Isolated mutations were evaluated for their ability to sustain cellular growth, stimulate downstream signalling pathways, and phosphorylate a novel JAK2 substrate in the presence of inhibitor. Mutations were found exclusively in the kinase domain of JAK2. The panel of mutations conferred resistance to high concentrations of inhibitor accompanied by sustained activation of the Stat5, Erk1/2, and Akt pathways. Using a JAK2 substrate, enhanced catalytic activity of the mutant JAK2 kinase was observed in inhibitor concentrations 200-fold higher than is inhibitory to the wild-type allele. When testing the panel of mutations in the context of the Jak2 V617F allele, we observed that a subset of mutations conferred resistance to inhibitor, validating the use of TEL-JAK2 in the initial screen. These results demonstrate that small-molecule inhibitors select for JAK2 inhibitor-resistant alleles. The design of next-generation JAK2 inhibitors should consider the location of mutations arising in inhibitor-resistant screens. Chapter 2. JAK2 mutagenesis and inhibitor resistance 53

2.2 Introduction

Myeloproliferative neoplasms (MPNs) are blood diseases characterized by an excess production of one or more fully differentiated blood cell types, and can be precursors to more severe disorders including myelodysplastic syndrome and acute leukemia [67,327,328]. Philadelphia chromosome-negative MPNs include polycythemia vera (PV), essential thrombocythemia (ET), and myelofibrosis (MF). The identification of a somatic valine to phenylalanine mutation at codon 617 of JAK2 was made in 90% of PV, 60% of ET, and 50% of PMF patients [237–240]. JAK2 is a cytoplasmic tyrosine kinase that is constitutively associated with members of the cytokine receptor superfamily. Ligation of the upstream receptor results in JAK2 cross-phosphorylation and activation of downstream pathways including the STAT family of transcription factors, the PI3-kinase/AKT survival pathway, and the MAP kinase pathway. Induction of these pathways results in transcription of genes required for cell survival and differentiation. The JAK2 V617F mutation lies in a domain previously thought to be a non-functional kinase domain. Recent work has demonstrated this ‘pseudo-kinase’ domain to be a functional dual-specificity kinase important in the negative regulation of cytokine signalling through phosphorylation of JAK2 Ser523 and Tyr570 [15]. Presence of the Jak2 V617F mutation was demonstrated to reduce phosphorylation on Ser523 and Tyr570, residues important in maintaining a low level of activity in the JAK2 kinase domain. The V617F mutation relieves the negative regulatory role of the dual-specificity kinase domain and makes the kinase weakly oncogenic, able to transform specific cell lines to cytokine independence [329].

Chronic myeloid leukemia (CML) is an Philadelphia chromosome-positive MPN characterized by the presence of the t(9;22)(q34;q11) chromosomal translocation [330] and the consequent expression of the BCR-ABL fusion protein [61]. Treatment of CML was revolutionized in 2001 with the development of the small-molecule inhibitor imatinib mesylate (IM) [203, 204, 331] which binds to the BCR-ABL kinase domain and inhibits its Chapter 2. JAK2 mutagenesis and inhibitor resistance 54 ability to phosphorylate target substrates [203,332]. Patients generally respond very well to IM, demonstrating results ranging from a partial hematologic response to complete cytogenetic remission [204, 333]. However, inhibitor resistance-based patient relapse can occur due to amplification of the BCR-ABL fusion gene or a mutation in the kinase domain that prevents small-molecule inhibitor binding [206,214,334,335]. In order to model BCR-ABL mutant generation, a BCR-ABL / IM in vitro system was developed to identify IM-resistant mutations [208, 209]. The resulting mutation spectrum bore a striking overlap with clinical results [209]. As such, the isolated mutations can be used to design next-generation inhibitors. Patients expressing small-molecule inhibitor-resistant mutations progress to next-generation inhibitors with variable results, largely depending on the specific mutation present [336,337]. Notably, the BCR-ABL T315I mutation is highly resistant to most ATP-competitive inhibitors against which it was tested [206,207], while many other IM-resistant mutations were susceptible to inhibition by second-generation inhibitors such as dasatinib [338]. These data suggest that both inhibitor-specific and ATP competitor-specific mutations can arise in response to drug treatment. Promising new inhibitors targeting different aspects of the BCR-ABL protein function are currently under development [221,222,339].

Discovery of JAK2 V617F and its role in PV, ET, and PMF started the search for a small-molecule inhibitor for JAK2. More than a dozen inhibitors have since been demonstrated to reduce JAK2 V617F kinase activity in vitro [297], some of which show variable efficacy in clinical trials [299, 308, 340]. To date, no inhibitor-resistant JAK2 mutations have been identified in patients. However, as JAK2 inhibitors become more widely used we anticipate a relapse rate that approximates the results observed with IM. We hypothesize that this relapse may be due mutations in the JAK2 kinase domain that prevent inhibitor binding, as is the case with IM-treated BCR-ABL. Using a random mutagenesis approach, we have identified JAK2 kinase domain residues critical in evading small-molecule inhibition. Here we describe the identification and characterization of Chapter 2. JAK2 mutagenesis and inhibitor resistance 55 mutations in the JAK2 kinase (JH1) domain that confer resistance to the presence of small-molecule inhibitors in vitro.

2.3 Materials and methods

Antibodies: The anti-phosphotyrosine antibody 4G10, anti-ERK1/2, and anti-STAT5a/b antibodies were purchased from Upstate Biotechnology (Lake Placid, NY). The anti- phospho-ERK1/2 (pY204) and anti-GST antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). The anti-phospho-STAT5 (pY694) antibody was purchased from Zymed (Carlsbad, CA). The anti-JAK2, anti-phospho-S6 (pS235/236), anti- S6, anti-phospho-Akt (pS473), and anti-Akt antibodies were purchased from Cell Sig- naling (Beverly, MA). The horseradish peroxidase (HRP)-conjugated protein A, donkey anti-rabbit-HRP IgG, and sheep anti-mouse-HRP IgG antibodies were purchased from GE Healthcare UK (Little Chalfont, Buckinghamshire, UK).

Plasmids: Human TEL-JAK2(5-12) cDNA (herein referred to as ‘TEL-JAK2’) and murine full-length Jak2 cDNA were cloned into the retroviral expression vector pMPG2. TEL-JAK2 contains the PNT dimerization domain of TEL, and the dual-speciﬁcity kinase (JH2) and kinase (JH1) domains of JAK2. The indicated point mutations in TEL-JAK2 and Jak2 were induced using the QuikChange site-directed mutagenesis kit (Stratagene; Santa Clara, CA). The JAK2 substrate was modelled after the activation loop of JAK2 (PQDKEYYKVKE) and cloned into pEBG-GST in order to express a GST-JAK2 activation loop fusion protein. KEYY was used to test the phosphorylation ability of Jak2 V617F and its associated mutations. Three other variants were also generated: KEYF was used to test the phosphorylation ability of TEL-JAK2 and its associated mutations, and KEFY and KEFF were used as negative controls in substrate optimization.

Inhibitors: JAK Inhibitor-I was purchased from EMD Chemicals (Gibbstown, NJ). Chapter 2. JAK2 mutagenesis and inhibitor resistance 56

CEP-701 (Lestaurtinib) was purchased from LC Laboratories (Woburn, MA). TG101348 was kindly donated by Ross Levine, Memorial Sloan-Kettering Cancer Center, New York, NY.

Cell lines and cell culture: The mouse hematopoietic cell line BaF3 was cultured in RPMI 1640 medium supplemented with 10% heat-inactivated fetal calf serum (FCS – Thermo Fisher Scientific; Waltham, MA), 50 nM β-mercaptoethanol (Thermo Fisher Scientific), and 10% WEHI-conditioned medium, isolated from WEHI-3B cells [341–343], which is referred to as ‘RPMI complete medium’. BaF3 cells expressing exogenous murine erythropoietin receptor (BaF3 EPO-R) were cultured in RPMI 1640 medium supplemented with 10% heat-inactivated FCS, 50 nM β-mercaptoethanol, and 0.5 units/mL of human recombinant erythropoietin (EPO). Human embryonic kidney 293T cells (herein referred to as ‘293T’) and the 293T-based Phoenix cells (exogenously expressing gag-pol and env genes for ecotropic virus production) [344] were cultured in Dulbecco’s Modi- fied Eagle’s Medium H21 supplemented with 10% heat-inactivated FCS. All cells were

o incubated at 37 C with 5% CO2.

HEK-293T and Phoenix cell transfection: Cells were seeded at 80% (293T) and 50% (Phoenix) conﬂuency on a 10-cm dish (Corning; Lowell, MA) and were transfected the following day with 0.1 µg of pEBG and/or 1.0 µg of pMPG2, unless otherwise indicated, using Lipofectamine-2000 (Invitrogen; Carlsbad, CA), according to the manufacturers instructions. In the case of 293T cells: eighteen hours following transfection, cells were split into 6-cm plates (Sarstedt; Numbrecht, Germany) at 30% conﬂuency and allowed to grow a further 12 hours before experimental manipulation. Phoenix cell treatment follows below.

Phoenix cell virus production and BaF3 cell transduction: Cells were transduced with ecotropic virus produced by the Phoenix cell system as previously described [344]. Brieﬂy, Phoenix cells were seeded at 50% conﬂuency and transfected with the pMPG2 retroviral vector (as above). Twenty-four hours post transfection, the H21 medium was aspirated Chapter 2. JAK2 mutagenesis and inhibitor resistance 57

and replaced with 10% WEHI-conditioned medium. At 48- and 72-hours post transfection, the WEHI retroviral supernatant was collected and passed through a 0.45 µM filter. Immediately after supernatant collection and filtration, 5 x 106 BaF3 cells were resuspended in 6 mL of retroviral supernatant containing 21.4 µM polybrene (Sigma-Aldrich; St. Louis, MO), a solution now referred to as the ‘retroviral cocktail’, and centrifuged at 1800 RPM for 90 minutes at room temperature (Beckman Coulter Spinchron R cen- trifuge). Cells were then resuspended in the same solution and centrifuged again as previously (for a total of 180 minutes). The BaF3 cells were incubated overnight in fresh retroviral cocktail. The following day cells were resuspended in fresh retroviral cocktail, and centrifuged as the previous day (2 x 90 minutes). The retroviral cocktail was then removed and cells resuspended in 25 mL RPMI complete medium. The newly transduced BaF3 cells were given 48 hours for retroviral integration to occur, and then subjected to cytokine-free medium selection or fluorescent-activated cell sorting (FACSCalibur from Becton Dickinson; Franklin Lakes, NJ).

Random mutagenesis and JAK2 mutant screen: pMPG2-TEL-JAK2 was used to transform XL1-Red Competent E. coli (Agilent; Santa Clara, CA), a strain deficient in DNA-repair genes mutS, mutD, and mutT resulting in a mutation rate approximately 5000-fold higher than wild type. A large volume of mutagenized plasmid was isolated from the XL1-Blue strain using a Maxiprep kit (Qiagen; Hilden, Germany). BaF3 cells were transduced with the mutagenized pMPG2-TEL-JAK2 library (as above). Success- fully transduced BaF3 cells were selected in RPMI complete medium lacking WEHI for three days. Cells were then plated at a low concentration in soft agar containing cytokine- free medium plus 1.93 µM JAK Inhibitor-I. Colonies formed were isolated and grown in RPMI 1640 complete medium lacking WEHI and containing 2.5 µM JAK Inhibitor-I. DNA was isolated using a mammalian genomic DNA extraction protocol [345]. Briefly, cell pellets were lysed with 0.5 mL DNA-extraction lysis buffer (100 mM Tris-HCl pH 8.5; 5 mM EDTA; 0.2% (w/v) sodium dodecyl sulfate (SDS); 200 mM NaCl; 100 µg proteinase Chapter 2. JAK2 mutagenesis and inhibitor resistance 58 k/mL). One volume of isopropanol was added to the lysate followed by mild agitation. The solution was centrifuged and the supernatant aspirated. The precipitate (DNA pellet) was re-solubilized in a solution composed of 10 mM Tris-HCl and 0.1 mM EDTA pH 7.5. The solubilized DNA was used to sequence the JAK2 kinase domain to identify mutations: PCR and sequencing primer forward (5’-GTGCCCTAGGGTTTTCTGGT- 3’); PCR primer reverse (5’-GTGTAACGGTTTTCTGCCGT-3’); sequencing primer A (5’-AGCAAGTTTTCTGTGGCCTC-3’). Identified mutations were re-cloned individually into pMPG2-TEL-JAK2, pMPG2-Jak2, pMPG2-Jak2-V617F using QuikChange site-directed mutagenesis kit (see above).

Cell Lysis and protein quantitation: 293T cells were gently washed with magnesium and calcium-free phosphate-buﬀered saline (PBS -MgCl2 -CaCl2), followed by resuspension in 200 µL protein-isolation lysis buﬀer (1 M Tris-HCl pH 8.0; 4 M NaCl; 4% (v/v)

Triton X-100; 0.5 M EDTA; 0.5 M Na4P2O7; 0.5 M NaF; 0.08 M Na3VO4; 0.5 M PMSF; one complete protease inhibitor cocktail tablet (Roche; Mannheim, Germany)). BaF3 cells were washed once with Hanks balanced salt solution buffered with 10 mM HEPES, centrifuged, and resuspended in 200 µL protein-isolation lysis buffer (as above). Cell lysates were incubated on ice for at least two minutes, and then cell debris was pelleted. Protein concentration was determined using the Bradford Colorimetric Assay (Bio-Rad; Hercules, CA) and detected at 595 nm using a Beckman Coulter DU 640 B spectrophotometer. Protein lysates were diluted 1:1 for immunoblot in 2 x sample buffer (140 mM Tris-HCl pH 6.8; 22% (v/v) glycerol; 4.4% (w/v) SDS; 0.04% (w/v) bromophenol blue) containing 100 mM dithiothreitol (DTT from USB; Cleveland, OH) prior to immunoblot analysis.

GST in vitro mixing: 293T cells expressing both pMPG2 and pEBG were lysed in protein-isolation lysis buffer (as above). Fifty microlitres of PBS-washed glutathione sepharose 4B beads (GE Healthcare; Waukesha, WI) and 1 mL protein-isolation lysis buffer were added to the quantified cell lysate and incubated overnight with agitation Chapter 2. JAK2 mutagenesis and inhibitor resistance 59

at 4oC. The beads were then washed three times with PBS and dried with a Hamilton syringe. Fifty microlitres of 1 x sample buﬀer (as above) was added for immunoblot analysis.

SDS-PAGE and immunoblot: Whole-cell lysate and GST pull-down samples were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to a polyvinylidene diuoride (PVDF) membrane (NEN Life Science; Boston, MA). The membranes were blocked at room temperature for one hour with 2.5% (w/v) bovine serum antigen (BSA from Sigma-Aldrich) or 5% (w/v) powdered milk solubilized in tris-buffered saline containing Tween 20 (TBS-T – 50 mM Tris pH 8.0; 150 mM NaCl; 0.1% Tween 20). Membranes were then incubated with an optimal concentration of the indicated primary antibody diluted in TBS-T as per the manufacturer’s directions. Post-primary antibody incubation, membranes were washed three times in TBS-T and incubated for 30 minutes in the suggested dilution of HRP-conjugated secondary antibody in TBS-T. Membranes were washed three times in TBS-T and visualized by enhanced chemi-luminescence (Western Lightning Plus-ECL from Perkin Elmer; Waltham, MA) with auto-radiographic film (Kodak Biomax film from Kodak; Rochester, NY, Amersham Hyperfilm from GE Healthcare, and Biomax film from Cedar Lane; Burlington, ON).

XTT assay: In order to quantify resistance conferred by speciﬁc TEL-JAK2 and Jak2 V617F mutations, an XTT assay was performed. All XTT experiments were performed in 96-well plates (Nunc; Rochester, NY) at an initial seeded cell number of 2 x 103 cells/well. BaF3 and BaF3 EPO-R cells expressing TEL-JAK2 or Jak2 V617F, respectively, bearing the indicated mutation, were diluted into the appropriate medium (as above) containing the indicated inhibitor concentration to a total volume of 100 µL/well. Each cell line and mutation was represented in triplicate, with the values averaged for plotting and

o statistical analysis. Cells were incubated in drug for 48 hours at 37 C and 5% CO2. Post- 48 hour incubation, 25 µL of pre-warmed XTT solution (125 nM phenazine methosulfate; 930 µM XTT diluted in medium) was added to each well, and cells were incubated at Chapter 2. JAK2 mutagenesis and inhibitor resistance 60

o 37 C and 5% CO2 for an additional eight hours. Absorbance at 450 nm was determined using a 96-well plate spectrophotometer (Optimax Turntable Microplate Reader from Molecular Devices; Sunnyvale, CA).

Protein structural analysis: The Jak2 kinase domain structure complexed with JAK Inhibitor-I [346] was obtained from the protein data bank (http://www.rcsb.org/, PDB ID: 2B7A). Structural analysis and image rendering was performed with PyMOL.

Statistical analyses: Data are expressed as mean +/- standard deviation (SD). All graphs were generated with, and the two-way ANOVA with Bonferronis post test was performed using GraphPad Prism version 5.0c for Mac OS X, GraphPad Software, San Diego, CA, www.graphpad.com.

2.4 Results

BaF3 cells transduced with a mutagenized pMPG2-TEL-JAK2 library identify inhibitor- resistant JAK2 kinase domain mutants. XL1-Red competent E. coli were transformed with pMPG2-TEL-JAK2(5-12), producing a mutagenized plasmid library. TEL-JAK2(5- 12), referred to as TEL-JAK2, contains the PNT oligimerization domain of TEL and the kinase and pseudokinase domains of JAK2. BaF3 cells were transduced with the mutagenized TEL-JAK2 library and incubated in soft agar containing 1.93 µM JAK Inhibitor-I (Table 2.1). Colonies presumably expressing mutagenized TEL-JAK2 were observed on the plates. One hundred colonies were selected, expanded, and the kinase domain was sequenced. Nine kinase mutations were identiﬁed (Table 2.2). For ease of interpretation, the wild type human JAK2 amino acid numbering will be used for the remainder of the thesis. Kinase domain mutations were identiﬁed once each in the screen. When mapped on the secondary structure of human JAK2, we do not observe obvious clustering within our panel of mutations (Figure 2.1). Four mutations lie within secondary structure including β2 and β3 sheets, and the hinge region. Five mutations Chapter 2. JAK2 mutagenesis and inhibitor resistance 61

lie within unstructured regions. These results demonstrate that we have optimized a soft-agar assay to identify inhibitor-resistant mutations in the JAK2 kinase domain. We also generated the homologous BCR-ABL T315I mutation in JAK2 (M929I) to determine whether it confers inhibitor resistance as well.

TEL-JAK2 kinase domain mutations are sufficient to support growth and downstream signalling at high concentrations of JAK Inhibitor-I. In order to determine if the identified mutations were responsible for inhibitor resistance and growth in the soft agar system, the mutations were generated in pMPG2-TEL-JAK2 and the mutated plasmid used to transduce BaF3 cells. An XTT assay was conducted with cells expressing selected mutants and treated with increasing concentrations of JAK Inhibitor-I. In BaF3 wild- type TEL-JAK2 cells, death was observed at 0.5 µM (Figure 2.2). In contrast, each of the tested mutants was able to grow at different ability at 0.5 µM. The TEL-JAK2 mutations N909K, G935R, and R975G group together at 0.5 µM JAK Inhibitor-I, and maintain an approximately 40% growth rate at 10 µM, suggesting very strong inhibitor resistance. Interestingly, cells expressing the engineered mutation, TEL-JAK2 M929I (homologous to BCR-ABL T315I), were inhibitor resistant but not to the degree of the strongest three mutants. TEL-JAK2 wild type, G935R, and R975G were also examined by XTT in the presence of TG101348 and CEP-701 (Figure 2.3). A statistically significant difference in growth between wild type and the mutants of TEL-JAK2 was not observed with either inhibitor.

Next we investigated the intracellular signaling downstream of TEL-JAK2. We probed for TEL-JAK2, Stat5, Akt, and Erk1/2 phosphorylation (Figures 2.4 and 2.5). Enhanced TEL-JAK2 phosphorylation was observed when inhibitor-resistant mutations were incubated in JAK Inhibitor-I, compared to wild type JAK2. TEL-JAK2 wild-type subclones displaying variable total expression were isolated and displayed no signiﬁcant Chapter 2. JAK2 mutagenesis and inhibitor resistance 62

Tyrosine kinase inhibitor Formula Structure

NH N

JAK Inhibitor-I C18H16FN3O O

H N O

CEP-701 C26H21N3O4 N O N

CH3 H

HO OH

H N O 2 N N H TG101348 C27H36N6O3S N S N N O O H

Table 2.1: Structural Diagrams of Selected JAK2 Inhibitors. JAK Inhibitor-I, also known as CMP-6, is a laboratory-based inhibitor that inhibits JAK1 (IC50 = 15 nM), JAK2 (IC50 = 1 nM), and Tyk2 (IC50 = 1 nM), values published by EMD Biosciences. CEP-701 (Lestaurtinib), developed by Cephalon, was initially discovered as a FLT3 (IC50 = 2 nM) [306] inhibitor and later determined to inhibit JAK2 (IC50 = 1 nM) [307], and TrkA (IC50 < 4 nM) [347]. TG101348, developed by TargeGen, was identiﬁed as a JAK2 inhibitor (IC50 = 3 nM) and is also active against FLT3 (IC50 = 15 nM) and RET kinase (IC50 = 48 nM) [302]. Chapter 2. JAK2 mutagenesis and inhibitor resistance 63

TEL-JAK2 Jak2 Secondary Mutation mutation mutation structure location type G663R G831R outside kinase non-polar, aliphatic to domain basic, hydrophilic E696K E864K β2 acidic, hydrophilic to basic, hydrophilic V713A V881A β3 non-polar, aliphatic to non-polar, aliphatic N741K N909K unstructured polar, hydrophilic to basic, hydrophilic Y750H Y918H unstructured polar, aromatic to basic, hydrophilic M761I* M929I* β5 non-polar, sulfur-based to non-polar, aliphatic G767R G935R unstructured non-polar, aliphatic to basic, hydrophilic R807G R975G unstructured basic, hydrophilic to non-polar, aliphatic P889S P1057S unstructured non-polar, aliphatic to polar, hydrophilic R959K R1127K α5 basic, hydrophilic to basic, hydrophilic

Table 2.2: JAK2 Kinase Domain Mutations Identiﬁed in an Inhibitor- Resistance Screen. Note: in the remainder of the thesis, mutations are referred to using the JAK2 numbering above prefaced with either Jak2 V617F or TEL-JAK2 to indicate context. To identify these mutants, BaF3 cells were transduced with a TEL- JAK2 mutant library and incubated in soft agar containing 1.96 µM JAK Inhibitor-I, a concentration that caused apoptosis in wild-type TEL-JAK2 cells. Colonies capable of growth were expanded and the kinase domain of TEL-JAK2 was sequenced. The mutations discovered are listed, with the equivalent mutation in the wild-type JAK2 tyrosine kinase domain. Each mutation listed was identiﬁed once in the screen. Methionine 929 isoleucine (*) was generated independently of the screen, modelling the threonine-315- isoleucine gatekeeper mutation found in inhibitor-resistant BCR-ABL. Location of these mutations in the JAK2 kinase domain secondary structure can be see in Figure 2.1, and tertiary structure in Figure 2.12. Secondary structure based on a previous publications [346,348]. Chapter 2. JAK2 mutagenesis and inhibitor resistance 64

E864K V881A

841 3A β1 β2 β3 H1αC

N909K Y918H M929I G935R 900 β4 β5 αD

R975G

959 αEH3 β6 3B β7 β8 β9 β10

P1057S 1018 αF β11 αG 3C αH

R1127K

1077 αI αJ 3D αK

Figure 2.1: Location of the Putative JAK2 Inhibitor-Resistant Mutations. The JAK2 kinase domain secondary structure is labeled as previously published [346,348]. Residues 841 to 1132 of the human JAK2 kinase domain are displayed. Mutations from Table 2.2 are displayed mapped to the JAK2 residue numbers. Identiﬁed mutations do not preferentially map to structured or unstructured regions. Beta sheets (β), alpha helices (α), and 310 alpha helices (3A, 3B, 3C, 3D) are labeled [348]. Chapter 2. JAK2 mutagenesis and inhibitor resistance 65

BaF3 - Vector Control

lity) TEL-JAK2 - Wild Type TEL-JAK2 - E864K TEL-JAK2 - V881A 100 TEL-JAK2 - N909K TEL-JAK2 - M929I TEL-JAK2 - G935R TEL-JAK2 - R975G

50 Percent viability (compared to inhibitor-free viabi (comparedinhibitor-free to viability Percent

0 0.01 0.1 1 10 0.5 μM Log [JAK Inhibitor-I] M

Figure 2.2: JAK2 Mutations Display Resistance to JAK Inhibitor-I. BaF3 hematopoietic cells expressing the construct indicated were treated for 48 hours in cytokine-free medium containing two-fold increasing concentrations of JAK Inhibitor- I. Cell viability was determined using the XTT assay. (n=4) Chapter 2. JAK2 mutagenesis and inhibitor resistance 66

a) 150

100 BaF3 Vector Control TEL-JAK2 - Wild Type TEL-JAK2 - G935R TEL-JAK2 - R975G 50 inhibitor-freeviability)

Percentviability (compared to 0 0.01 0.1 1 10 100 Log [JAK Inhibitor-I] μM b) c) 150 150

100 100

50 50 inhibitor-freeviability) inhibitor-freeviability) Percentviability (compared to

0 Percentviability (compared to 0 0.01 0.1 1 10 0.01 0.1 1 10 100 Log [TG101348] μM Log [CEP-701] μM

Figure 2.3: TEL-JAK2 and Jak2 V617F Mutants are not Resistant to TG101348 or CEP-701. BaF3 hematopoietic cells expressing the TEL-JAK2 construct indicated were treated for 48 hours in cytokine-free medium containing 0.01 – 30 µM of the indicated inhibitor. Cell viability was determined using the XTT assay. (a) BaF3 cells expressing G935R and R975G grow significantly better than wild type in increasing JAK Inhibitor-I concentrations. (b) G935R and R975G do not display a significant difference in growth compared to wild type in increasing TG101348. (c) G935R and R975G do not display a significant difference in growth compared to wild type in increasing CEP-701. (n=3) Chapter 2. JAK2 mutagenesis and inhibitor resistance 67

diﬀerence in overall survival (data not shown), demonstrating total TEL-JAK2 expression does not correlate with survival ability. A substantially stronger Stat5 activation was observed in the mutants as compared to wild type, at all tested inhibitor concentrations. Enhanced Akt phosphorylation was observed in all TEL-JAK2 mutants in the presence of JAK Inhibitor-I, suggesting that Akt activation is coupled to enhanced cell survival in the presence of inhibitor. Erk1/2 phosphorylation was observed at higher concentrations of inhibitor, particularly in TEL-JAK2 E864K (Figure 2.4, lanes 5 – 8), N909K (lanes 9 - 12), G935R (Figure 2.5, lanes 9 - 12), and R975G (lanes 13 - 16). These results suggest we have identiﬁed a panel of JAK2 kinase domain mutants that can sustain growth in high concentrations of inhibitor, perhaps due to activation of Stat5 and Erk1/2 anti-apoptosis or survival pathways.

Speciﬁc TEL-JAK2 kinase domain mutations can support elevated kinase activity at high inhibitor concentrations. To investigate the ability of the TEL-JAK2 mutants to function as a kinase in high concentration of inhibitor, we designed a JAK2 substrate fusion protein combining the glutathione S-transferase protein with an 11 amino acid sequence modelling the JAK2 activation loop sequence (PQDKEYYKVKE [349], referred to as GST-J2s-KEYY. Three additional constructs were generated: PQDKEYFKVKE (GST-J2s-KEYF), PQDKEFYKVKE (GST-J2s-KEFY), and PQDKEFFKVKE (GST- J2s-KEFF). 293T cells were transfected with pMPG2-TEL-JAK2 (Figure 2.6a) or Jak2 V617F (Figure 2.6) and one of the four JAK2 substrate variants in order to assess the ability of the kinase to phosphorylate the tyrosines within these substrate fusion proteins (Figure 2.6). TEL-JAK2 stimulates tyrosine phosphorylation of a doublet in GST-J2s- KEYY, so GST-J2s-KEYF was utilized for intra-cellular kinase assays testing TEL-JAK2 mutants. Neither TEL-JAK2 nor Jak2 V617F phosphorylated the GST-J2s-KEFY or GST-J2s-KEFF fusion proteins.

After substrate optimization, 293T cells expressing pMPG2-TEL-JAK2 and pEBG- Chapter 2. JAK2 mutagenesis and inhibitor resistance 68

TEL-JAK2 Wild Type E864K V881A N909K

[JAK Inhibitor-I] μM 0.00 0.065 0.65 6.50 0.00 0.065 0.65 6.50 0.065 0.65 0.00 6.50 0.065 0.00 0.65 6.50 1 2 3 4 5 6 7 8 910 11 12 13 14 15 16 pTEL-JAK2

TEL-JAK2

pStat5 Stat5

pAkt

Akt

pErk1/2

Erk1/2

Figure 2.4: TEL-JAK2 Inhibitor-Resistant Mutants Display Enhanced Phos- phorylation of Stat5, Akt and Erk1/2. The indicated BaF3 TEL-JAK2 mutant cell lines were cultured in RPMI complete medium containing increasing concentrations of JAK Inhibitor-I for four hours. Lysates were isolated and JAK2, Stat5, Akt, and Erk1/2 phosphorylation and expression was assessed by immunoblot. (n=4) Chapter 2. JAK2 mutagenesis and inhibitor resistance 69

TEL-JAK2 Wild Type M929I G935R R975G [JAK Inhibitor-I] μM 0.00 0.065 0.65 6.50 0.065 0.65 0.00 0.065 0.65 6.50 0.00 0.065 0.65 6.50 0.00 6.50 1 2 3 4 5 6 7 8 910 11 12 13 14 15 16 pTEL-JAK2

TEL-JAK2

pStat5 Stat5

pAkt

Akt

pErk1/2

Erk1/2

Figure 2.5: TEL-JAK2 Inhibitor-Resistant Mutants Display Enhanced Phos- phorylation of Stat5, Akt and Erk1/2. The indicated BaF3 TEL-JAK2 mutant cell lines were cultured in RPMI complete medium containing increasing concentrations of JAK Inhibitor-I for four hours. Lysates were isolated and JAK2, Stat5, Akt, and Erk1/2 phosphorylation and expression was assessed by immunoblot. (n=4) Chapter 2. JAK2 mutagenesis and inhibitor resistance 70

a) Glutathione Sepharose Pull-Down pMPG2 + + + + + pMPG2-TEL-JAK2 + + + + + pEBG-GST + + pEBG-GST-J2s-KEYY + + pEBG-GST-J2s-KEYF + + pEBG-GST-J2s-KEFY + + pEBG-GST-J2s-KEFF + + 1 23 4 5 67 8 910

α-pTyr pJ2s-KEXX-GST

α-GST J2s-KEXX-GST

Lysate

α-pTyr TEL-JAK2

b) Glutathione Sepharose Pull-Down Jak2 Jak2G935R Jak2R975G Vehicle TEL-JAK2 TEL-JAK2 Jak2V617F Jak2V617F/G935R Jak2V617F/R975G pEBG-GST-J2s-KEYY + + + + + + + + pEBG-GST-J2s-KEYF + 1 2 3 4 5 6 7 8 9 α-pTyr pJ2s-KEYX-GST

α-GST J2s-KEYX-GST

Lysate

Jak2 α-JAK2 TEL-JAK2

Figure 2.6: TEL-JAK2 and Jak2 V617F Phosphorylate JAK2 Substrate Acti- vation Loop Sequences. 293T cells were transfected with TEL-JAK2, Jak2 V617F, or empty vector and various GST-JAK2 substrate constructs as indicated. Forty-eight hours post-transfection, cells were lysed, GST fusions were captured on glutathione-sepharose beads, and immunoblotting was performed with anti-phosphotyrosine, anti-GST antibodies, or anti-JAK2 antibodies. (a) KEXX denotes either KEYY, KEYF, KEFY, or KEFF GST-Jak2 substrate fusion constructs. (b) KEYX denotes either KEYY or KEYF. Chapter 2. JAK2 mutagenesis and inhibitor resistance 71

GST-J2s-KEYF were incubated with JAK Inhibitor-I for four hours, lysed, the JAK2 substrate fusion protein was isolated and probed for phosphorylation (Figure 2.7). All tested mutants display phosphorylation of the JAK2 substrate at 0.65 µM, a JAK Inhibitor-I concentration that suppresses wild-type TEL-JAK2 substrate phosphorylation. TEL-JAK2 E864K, V881A, and M929I phosphorylate the substrate slightly at higher JAK Inhibitor-I concentrations. Only TEL-JAK2 G935R (Figure 2.7a, lanes 14- 16) and R975G (lanes 17-19) display substantial kinase activity at 6.5 µM. To test the maximal concentration of inhibitor at which G935R and R975G are able to retain kinase function, we incubated transfected 293T cells in up to 130 µM JAK Inhibitor-I. Wild-type TEL-JAK2 phosphorylation was observed at 0.65 µM JAK Inhibitor-I in a long exposure (Figure 2.7b, lane 3). TEL-JAK2 G935R retains kinase activity exceeding 130 µM JAK Inhibitor-I (Figure 2.7b, lanes 8 - 13), while TEL-JAK2 R975G activity is attenuated but still present (Figure 2.7b, lanes 14 - 19). Interestingly, in 293T cells TEL-JAK2 expression is variable. This result suggests that the isolated TEL-JAK2 mutations disrupt protein stability or turnover. In order to address this issue, we transfected ﬁve-fold more wild-type TEL-JAK2 than G935R or R975G and determined that normalization of TEL-JAK2 expression does not aﬀect its kinase activity at high doses of JAK Inhibitor- I (Figure 2.7b). These results suggest that selected TEL-JAK2 mutations are at least 200-fold more resistant to a JAK Inhibitor-I than wild type.

Speciﬁc TEL-JAK2 mutations confer inhibitor resistance in the context of Jak2 V617F in both growth and downstream signalling. The initial soft agar screen was completed with mutagenized TEL-JAK2. We hypothesized that, due to the identity between the kinase domains of TEL-JAK2 and Jak2 V617F, any inhibitor-resistant mutation discovered in TEL-JAK2 would have similar or identical activity in Jak2 V617F. In order to test this hypothesis, the panel of TEL-JAK2 mutations was generated in the homologous residues of Jak2 V617F (Table 2.2). BaF3 EPO-R cell lines, expressing our panel of mutations in the context of Jak2 V617F, were generated by transducing cells with one of the panel of Chapter 2. JAK2 mutagenesis and inhibitor resistance 72

a) Glutathione Sepharose Pull-Down TEL-JAK2 C WT E864K V881A M929I G935R R975G [JAK Inhibitor-I] μM 0.00 0.00 0.65 6.50 0.00 0.65 0.00 0.65 6.50 0.00 0.65 6.50 0.00 0.65 6.50 0.00 0.65 6.50 6.50 1 2 3 4 5 67 8910 11 12 13 14 15 16 17 18 19

α-pTyr pJ2s-YF-GST

α-GST J2s-YF-GST

Lysate

α-JAK2 TEL-JAK2

b) Glutathione Sepharose Pull Down C TEL-JAK2 WT TEL-JAK2 G935R TEL-JAK2 R975G [JAK Inhibitor-I] μM 0.00 0.00 0.65 6.50 32.5 65.0 130 0.00 0.65 6.50 32.5 65.0 130 0.00 0.65 6.50 32.5 65.0 130 1 2 34 5 6 7 8910 11 12 13 14 15 16 17 18 19

α-pTyr pJ2s-YF-GST

α-GST J2s-YF-GST

Lysate

α-JAK2 TEL-JAK2

Figure 2.7: TEL-JAK2 Mutants G935R and R975G Display a Strong Degree of Inhibitor Resistance. 293T cells were co-transfected with the TEL-JAK2 construct indicated and a GST-JAK2 substrate fusion gene (GST-J2s-KEYF). Post-transfection, cells were incubated with the indicated JAK Inhibitor-I concentration for four hours. Cells were lysed and the GST fusion protein was isolated. Phosphorylation of the JAK2 substrate and the TEL-JAK2 fusion protein were assessed with a phospho-tyrosine spe- ciﬁc antibody. Total GST and total TEL-JAK2 were also examined. (a) JAK2 substrate phosphorylation was assessed between TEL-JAK2 wild type, E864K, V881A, M939I, G935R, and R975G at three concentrations of inhibitor. (n=3) (b) JAK2 substrate phosphorylation was assessed between TEL-JAK2 wild type, G935R, and R975G at seven inhibitor concentrations. Five micrograms of TEL-JAK2 wild type, and 1 µg of both G935R and R975G was transfected. Phosphorylation of the JAK2 substrate is observed at 0.65 µM inhibitor due to a longer immunoblot exposure, as compared to the top panel. (n=3) Chapter 2. JAK2 mutagenesis and inhibitor resistance 73

Jak2 V617F mutants. We chose the BaF3 EPO-R cell line because it has been demonstrated that Jak2 V617F requires a cytokine receptor scaffold to adequately function [350] and presumably display inhibitor resistance. Indeed this was shown to be the case in an XTT growth assay lacking EPO, where Jak2 V617F wild type and mutant cells displayed no difference in growth in JAK Inhibitor-I (data not shown). To test the growth ability of our most inhibitor-resistant mutations, we conducted an XTT assay in 0.1 unit/mL EPO plus increasing concentrations of JAK Inhibitor-I. A statistically significant difference in growth between wild-type Jak2 V617F and Jak2 V617F G935R was observed a JAK Inhibitor-I concentration of 1.25 µM and higher (p < 0.05, Figure 2.8). However, we did not observe a growth difference between Jak2 V617F wild type and R975G. Jak2 V617F wild type, G935R, and R975G were also examined by XTT in the presence of TG101348 and CEP-701 (Figure 2.9). A statistically significant difference in growth was not observed.

Next, we investigated the intracellular signalling downstream of Jak2 V617F wild type, G935R, and R975G. We probed for Stat5, Erk1/2, and S6 kinase activation (Fig- ure 2.10). JAK Inhibitor-I silences Stat5 signalling in the BaF3 EPO-R cell line at all concentrations tested, whereas Stat5 phosphorylation in wild-type Jak2 V617F is suppressed at 8.0 µM (Figure 2.10, lane 9). In contrast, both G935R (lane 14) and R975G (lane 19) show sustained Stat5 phosphorylation up to 8 µM. Erk1/2 phosphorylation in blocked above 1.6 µM JAK Inhibitor-I in BaF3 EPO-R cells. Erk1/2 signalling is also attenuated in wild-type Jak2 V617F and R975G in increasing inhibitor concentrations, but appears to be stronger in G935R. S6 kinase is activated at low concentrations of inhibitor only in G935R. These results suggest the survival diﬀerence observed between Jak2 V617F wild type and Jak2 V617F G935R may be due to enhanced Erk1/2 activation, or perhaps S6 kinase.

Jak2 V617F G935R can support kinase activity at an inhibitor concentration more Chapter 2. JAK2 mutagenesis and inhibitor resistance 74

* Jak2 V617F - Wild Type 100 Jak2 V617F - G935R * Jak2 V617F - R975G ** ** ** ** 50 inhibitor-free viability) inhibitor-free Percent viability (compared to viability Percent 0 0.1 1 10

Log10 [JAK Inhibitor-I] M

Figure 2.8: Jak2 V617F G935R is Resistant to JAK Inhibitor-I. BaF3 hematopoietic cells expressing the construct indicated were treated for 48 hours in RPMI medium plus 0.1 units/mL EPO containing two-fold increasing concentrations of JAK Inhibitor-I. Post-treatment cell viability was determined using the XTT assay. The * indicates p <0.01; ** indicates p <0.001 for Jak2 V617F G935R compared with Jak2 V617F (two-way ANOVA followed by Bonferronis post test). Concentrations up to 30 µM tested and cell lines display the same trend. (n=4) Chapter 2. JAK2 mutagenesis and inhibitor resistance 75

150

100 BaF3-EPO-R Vector Control Jak2 V617F - Wild Type Jak2 V617F - G935R 50 Jak2 V617F - R975G inhibitor-freeviability)

Percentviability (compared to 0 0.01 0.1 1 10 100 Log [JAK Inhibitor-I] μM b) c)

150 150

100 100

50 50 inhibitor-freeviability) inhibitor-freeviability)

0 Percentviability (compared to 0 Percentviability (compared to 0.01 0.1 1 10 0.01 0.1 1 10 100 Log [TG101348] μM Log [CEP-701] μM

Figure 2.9: Jak2 V617F G935R is not Resistant to TG101348 or CEP-701. BaF3 hematopoietic cells expressing the construct indicated were treated for 48 hours in RPMI medium plus 0.1 units/mL EPO containing two-fold increasing concentrations of the inhibitor indicated. Post-treatment cell viability was determined using the XTT assay. (n=3) Chapter 2. JAK2 mutagenesis and inhibitor resistance 76

Jak2 V617F BaF3 EPO-R Wild Type G935R R975G [JAK Inhibitor-I] μM 1.6 0.0 0.3 8.0 0.0 0.3 3.2 1.6 8.0 0.0 0.3 3.2 1.6 8.0 0.0 0.3 3.2 1.6 8.0 1 2 34 5 678 910 11 12 13 14 15 16 17 18 19 pJak2

Jak2

pStat5

Stat5

pErk1/2

Erk1/2

pS6

Figure 2.10: Jak2 V617F G935R Displays Enhanced Stat5 and Erk1/2 Phos- phorylation. BaF3 EPO-R cells expressing Jak2 V617F G935R and R975G were cultured in RPMI complete medium containing 0.1 units/mL EPO and increasing concentrations of JAK Inhibitor-I for four hours. Lysates were isolated and Jak2, Stat5, Erk1/2, and ribosomal S6 kinase phosphorylation and expression were assessed by immunoblot using anti-phospho and anti-total antibodies, respectively. (n=3) Chapter 2. JAK2 mutagenesis and inhibitor resistance 77

than 30-fold higher than wild type. In order to compare the function of the Jak2 mutant kinase in the context of V617F, we used our JAK2 activation loop (KEYY) GST fusion construct to examine Jak2 kinase activity in the presence of JAK Inhibitor-I. 293T cells were co-transfected with a vector expressing Jak2 V617F wild type, G935R, or R975G, and the GST-J2s-KEYY fusion vector. Cells were treated with JAK Inhibitor-I for four hours and lysed. The JAK2 substrate fusion protein was isolated with glutathione sepharose 4B beads and probed for phosphorylation (Figure 2.11). Jak2 V617F G935R displays very strong kinase function up to 26 µM JAK Inhibitor-I, the highest concentration tested, a 30-fold increase over wild type function. Wild-type Jak2 bearing either G935R or R975G does not phosphorylate the substrate (Figure 2.6b).

2.5 Discussion

Inhibitor resistance is currently one of the biggest challenges facing effective treatment of CML. Evidence suggests that BCR-ABL mutations are present at the commencement of treatment, and the inhibitor provides strong selective pressure for mutant clone outgrowth and consequent patient relapse [351, 352]. Considerable effort has been put forth identifying and testing new inhibitors targeting specific BCR-ABL mutations. The in vitro prediction of BCR-ABL mutations against multiple inhibitors was robust and provided the field with significant data to aid in the design of next-generation kinase inhibitors [209].

Identiﬁcation of a single point mutation, Jak2 V617F, thought to play an important role in MPN development and progression, initiated the search for small-molecule inhibitors of the JAK2 tyrosine kinase. We hypothesize that inhibitor-resistant JAK2 alleles will become apparent as large cohorts of MPN patients progress through clinical trials testing JAK2-selective drug therapies. The objective of our study was to identify JAK2 mutations that provide resistance to small molecule inhibitors before patient Chapter 2. JAK2 mutagenesis and inhibitor resistance 78

Glutathione Sepharose Pull Down C Jak2 V617F WT Jak2 V617F G935R Jak2 V617F R975G [JAK Inhibitor-I] μM 0.65 3.30 6.50 13.0 0.00 0.65 3.30 6.50 13.0 26.0 0.00 0.00 0.00 0.65 3.30 6.50 13.0 26.0 26.0 1 2 3 4 5 6 7 8 9 10 1112 13 1415 16 171819 α-pTyr pJ2s-KEYY-GST

α-GST J2s-KEYY-GST

Lysate α-JAK2 Jak2

Figure 2.11: Jak2 V617F G935R Displays a Strong Degree of Inhibitor Re- sistance. 293T cells were co-transfected with the TEL-JAK2 construct indicated and a GST-JAK2 substrate fusion gene (GST-J2s-KEYY). Post-transfection, cells were incubated with the indicated JAK Inhibitor-I concentration for four hours. Cells were lysed and the GST fusion protein was isolated. Phosphorylation of the JAK2 substrate and the Jak2 V617F mutant protein were examined by the phospho-tyrosine speciﬁc antibody 4G10. Total GST and total Jak2 were also assessed. (n=3) Chapter 2. JAK2 mutagenesis and inhibitor resistance 79

resistance and relapse is observed in the clinic.

TEL-JAK2 is a fusion gene created by the t(9;12)(p24;p13) translocation [62, 63]. The identity between the Jak2 and TEL-JAK2 kinase domains has allowed us to directly apply findings in TEL-JAK2 to Jak2 V617F. BaF3 cells expressing each mutation in TEL- JAK2 were evaluated with an XTT assay to indirectly assess viability in the presence of inhibitor. TEL-JAK2 N909K, G935R, and R975G cluster very closely together in their survival profile, followed by M929I, E864K, and V881A (Figure 2.2). This result is closely mirrored in the signalling data in which TEL-JAK2 N909K (Figure 2.4), G935R, and R975G (Figure 2.5) have similar pStat5 and pErk1/2 activation at higher inhibitor concentrations. The weakest mutant, TEL-JAK2 V881A, survives slightly better than wild type at 0.5 µM JAK Inhibitor-I, and the minor difference is evident when comparing wild type and V881A signalling profiles. BaF3 cells bearing TEL-JAK2 mutants display some variation in the activation of Stat5, Akt and Erk1/2 in the absence of inhibitors. Curiously, some TEL-JAK2 mutants with elevated basal phosphorylation of downstream signalling components had lower in vitro kinase activity. For example, TEL-JAK2 V881A had high Erk2 phosphorylation in the absence of JAK Inhibitor-I, but weak kinase activity upon drug addition. We also examined growth ability in the presence of two clinically relevant inhibitors, TG101348 and CEP-701 (TEL-JAK2 in Figure 2.3 and Jak2 V617F in Figure 2.9). The lack of growth difference observed in the XTT data suggests we have isolated compound-specific, not ATP competitor-specific, mutations.

To further understand how the JAK2 kinase domain has been modified by the presence of mutations, we developed a novel intra-cellular assay to directly assess its phosphorylation ability in a system more relevant than a standard in vitro kinase assay. By fusing the glutathione S-transferase gene to the JAK2 activation loop, we were able to isolate and directly probe the JAK2-targeted phosphorylation of a bona fide JAK2 substrate. Our results confirm the XTT and BaF3 TEL-JAK2 signalling data. Wild-type TEL- JAK2 kinase ability is not detectable at 0.65 µM JAK Inhibitor-I. TEL-JAK2 V881A, Chapter 2. JAK2 mutagenesis and inhibitor resistance 80

E864K, and M929I have a low level of phosphorylation, while G935R and R975G have elevated kinase activity up to 130 µM (Figure 2.7b, lanes 13 and 19).

Interestingly, some of our identified mutations in TEL-JAK2 did not translate to resistance in Jak2 V617F. We evaluated our entire panel of mutations in the context of Jak2 V617F with XTT-based survival, downstream signalling, and with our JAK2 substrate kinase assay (data not shown). We observed only Jak2 V617F G935R to display a significant difference in survival, downstream signalling, and substrate phosphorylation in comparison to Jak2 V617F wild type or other mutant alleles. There are at least two possible explanations for this finding. First, the difference may be due to the relative kinase potency of TEL-JAK2 compared to Jak2 V617F. The Jak2 V617F allele is not transforming unless it has a functional FERM domain and is provided with a cytokine scaffold [350], and even then is relatively indolent without other mutations present [294, 327, 353]. In contrast, TEL-JAK2 is a potent oncogene thought to be causative in some cases of acute myeloid leukemia [62,63]. Therefore, even small differences in inhibitor resistance will be evident with TEL-JAK2, while the homologous mutations may have subtle effects in the context of Jak2 V617F. Second, the mechanisms of activation of TEL-JAK2 and Jak2 V617F are different. The PNT domain of TEL causes oligimerization of the TEL-JAK2 protein and constitutive activation. Therefore, the inhibitor resistance observed in some TEL-JAK2 mutations may be due to the oligimerization-specific interaction between the kinase domains.

In order to understand how the panel of identiﬁed mutations contributes to inhibitor resistance, mutations were modelled using the previously published human JAK2 kinase domain crystal structure complexed with JAK Inhibitor-I (Figure 2.12) [346]. The location of the relevant amino acids are displayed on the tertiary structure (Figure 2.12a, b, c). G935 lies within the hinge region between the N-lobe and C-lobe. The G935R mutation introduces a spatial clash resulting from the change from a glycine (Figure 2.12d, e) to an arginine side chain (Figure 2.12f), which prevents inhibitor binding. R975 is located Chapter 2. JAK2 mutagenesis and inhibitor resistance 81

in the catalytic loop region connecting α-helix D with the activation loop. The replace- ment of arginine by glycine (R975G), combined with increased flexibility of the main chain, would influence inter-loop interactions and possibly affect opening of the pocket. E864K changes the side-chain charge and results in a steric clash with the neighbouring lysine. This would result in movement of the β sheet and occlusion of the pocket. N909K introduces a steric clash that may push neighbouring V911 into the binding pocket. The V881A mutation results in loss of the valine in the hydrophobic core, thereby affecting packing and orientation.

A recent publication has identiﬁed activating murine Jak1 (mJak1) mutations selected for by cytokine deprivation [354]. Interestingly, some of these mutations also confer resistance to the JAK Inhibitor-I and INCB018424. In order to compare ﬁndings, the mJak1 and human JAK2 kinase domains were aligned and the relevant mutations highlighted (Figure 2.13). Notably, the JAK2 mutations E864K and V881A from this study cluster with the mJak1 mutations D895H, E897K, T901R, and L910Q in the β2 and β3 loop. JAK2 G935R clusters quite closely with the Jak1 mutations F958V/C/S/L and P960T/S in the kinase domain activation loop. This strong overlap suggests there are common regions in the the JAK family kinase domains that are susceptible to inhibitor- resistant mutations.

Two recent publications utilized a similar approach as this study: the first used mutagenesis of Jak2 V617F and the ruxolitinib (INCB018424) [355], and second used mutagenized Jak2 R683G co-expressed with the Crlf2 receptor in BaF3 cells exposed to the BVB808 JAK2 inhibitor [356]. The results of these mutagenesis screens have also been mapped on the mJak1 / hJAK2 alignment (Figure 2.13). In sum, these studies discovered ten inhibitor-resistant mutations that cluster around the ATP-binding pocket. G935R was identified in all three groups, suggesting that G935 lies at a critical interface for inhibitor binding (Figure 2.12). Weigert et al. demonstrated that G935R has a broad resistance profile by using a wide panel of JAK2-selective inhibitors. These results are in Chapter 2. JAK2 mutagenesis and inhibitor resistance 82

a) b) c) Y918

E864 M929 G935 V881 N909 R1127 R975

P1057 d) e) f)

Figure 2.12: JAK2 Inhibitor-Resistant Residues Mapped to the Crystal Structure Bound to JAK Inhibitor-I. The JAK2 kinase domain in complex with JAK Inhibitor-I has been previously published [346]. (a) The face of the kinase domain displays three of the nine residues with identiﬁed mutations: E864, G935, R975. (b) Rotated 90 degrees counter-clockwise (on the Y plane), three more residues are visible: N909, M929, R1127. (c) Rotated 90 degrees clockwise (on the Y plane) from 9a are the ﬁnal three residues: V881, Y918, P1057. (d) The kinase domain is displayed, the n-lobe in orange and c-lobe in yellow. G935 is displayed in red. (e) The kinase domain binding pocket is displayed, G935 in red. (f) The kinase domain binding pocket is displayed, the R935 mutation is in red. Chapter 2. JAK2 mutagenesis and inhibitor resistance 83

contrast to our results demonstrating inhibitor specificity. Y931C (homologous to Jak1 F958, identified by Hornakova et al. [354]) was isolated the Weinstock group [356] and displayed broad inhibitor resistance. In contrast, the E864K mutation (isolated in this study and by Weigert et al. [356] displayed narrow inhibitor resistance, suggesting that E864 is more inhibitor specific. The importance of the gatekeeper residue, M929, in Jak2 was verified by Deshpande et al. and our study, as the M929I mutation displayed resistance to JAK Inhibitor-I and ruxolitinib [355]. Other mutations were uniquely identified as resistant to JAK Inhibitor-I (V881A, N909K, and R975G) or ruxolitinib (R938L, I960V, and E985K) and may represent inhibitor-specific mutations. It is significant to note that all inhibitor-resistant mutations were identified in the JAK2 JH1 domain and no allosteric mutations were isolated in the pseudokinase or FERM domains. While our approach was a proof-of-concept screen that was not completed to saturation, there is considerable redundancy amongst the three reports, suggesting that fewer JAK2 residues may be critical in mediating inhibitor resistance when compared to the published BCR- ABL studies. Our initial in vitro results outline a framework to identify and test JAK2 alleles capable of small molecule inhibitor resistance. Our choice of inhibitor was based on its commercial availability and the published structure complexed with the JAK2 kinase domain [346]. However, our colony selection scheme and evaluation experiments can be used with any JAK2 inhibitor available. Applied in a high-throughput manner, this experimental procedure may help identify inhibitor-resistant JAK2 mutations before they are observed in the clinic, and therefore allow the development of next-generation inhibitors. Chapter 2. JAK2 mutagenesis and inhibitor resistance 84

G831R ↓ hJAK2 AIIRDLNSLFTPDYELLTENDMLPNMRIGALGFSGAFEDRDPTQFEERHLKFLQQL 855 mJak1 AIMRDINKLEEQNPDIVSE...... KQPTTEVDPTHFEKRFLKRIRDL 881

E864K V881A N909K ↓↓↓ hJAK2 GKGNFGSVEMCRYDPLQDNTGEVVAVKKLQ.HSTEEHLRDFEREIEILKSLQHDNI 910 mJak1 GEGHFGKVELCRYDPEGDNTGEQVAVKSLKPESGGNHIADLKKEIEILRNLYHENI 937 ↑↑↑ D895H E897K L910Q

Y918H M929I* G935R R938L I960V ↓↓↓↓↓ hJAK2 VKYKGVCYSAGRRNLKLIMEYLPYGSLRDYLQKHKERIDHIKLLQYTSQICKGMEY 966 mJak1 VKYKGICMEDGGNGIKLIMEFLPSGSLKEYLPKNKNKINLKQQLKYAIQICKGMDY 993 ↑↑ F958V/C/S/L P960T/S

R975G E985K ↓↓ hJAK2 LGTKRYIHRDLATRNILVENENRVKIGDFGLTKVLPQDKEYYKVKEPGESPIFWYA 1022 mJak1 LGSRQYVHRDLAARNVLVESEHQVKIGDFGLTKAIETDKEYYTVKDDRDSPVFWYA 1049

P1057S ↓ hJAK2 PESLTESKFSVASDVWSFGVVLYELFTYIEKSKSPPAEFMRMIGNDKQGQMIVFHL 1078 mJak1 PECLIQCKFYIASDVWSFGVTLHELLTYCDSDFSPMALFLKMIG.PTHGQMTVTRL 1104

R1127K ↓ hJAK2 IELLKNNGRLPRPDGCPDEIYMIMTECWNNNVNQRPSFRDLALRVDQIRDNMAG 1132 mJak1 VNTLKEGKRLPCPPNCPDEVYQLMRKCWEFQPSNRTTFQNLIEGFEALLK.... 1154

Figure 2.13: Inhibitor-Resistant Mutations Identiﬁed in mJak1 and hJAK2 Mutagenesis Screens. Results from Hornakova et al. [354] are in red, numbering in reference to mJak1. Deshpande et al. [355] in green, numbering in reference to human JAK2. Screened mutations from this study are in blue, which suggests cross-study clustering and overlap of mutations within important secondary and tertiary structures in the consensus JAK kinase domain. M929I* denotes an engineered mutation, mimicking the BCR-ABL T315I gatekeeper mutation. E864K, Y931C, and G935R were also reported by Weigert et al. [356]. Chapter 3

Examining Chromosome 9q34 Deletion as a Marker for Genomic Instability in Chronic Myeloid Leukemia

Michael R. Marit, Stefanie A. Turley, Kai Huang, Jeremy Squire, Suzanne Kamel-Reid, and Dwayne L. Barber

Author contributions: M.R.M. and D.L.B. designed experiments; M.R.M. conducted experiments; S.T. optimized GAPDH and PRDM12 qPCR probes; K.H. generated the vector constructs; J.S. provided the Nexus software suite; S.K.R. provided patient samples; M.R.M. and D.L.B. wrote the manuscript

85 Chapter 3. 9q34 deletions in CML 86

3.1 Abstract

Chronic myeloid leukemia (CML) is a disease of the bone marrow and blood characterized by a deregulated expansion of granulocytic cells. CML is driven by the BCR-ABL fusion gene, a result of the t(9;22)(q34;q11) chromosomal translocation. The majority of CML patients possess a reciprocal translocation generating the Philadelphia chromosome and the derivative 9. However, 20% of patients exhibit a hemizygous deletion at 9q34 (9q34–) that occurs at the time of translocation. Deletions at 9q34 (the ABL-BCR locus) is a poor prognostic factor for patients receiving IFN-α. There are at least two hypotheses regarding this overall survival difference. First, the a 9q34 deletion is a marker of a pre- existing genomic instability. To test this hypothesis, we undertook array comparative genomic hybridization to examine total aberrations in nine diagnostic CML samples. Comparison of 9q34+ and 9q34– samples suggest total genomic aberrations do not differ, but we did identify deletions in the retinoblastoma 1 (RB1 ) gene in two 9q34– samples. These results suggest that the genomic instability hypothesis is likely not correct, but RB1 status may influence disease progression. The second suggests that deletions at 9q34 results in loss of one or more tumour suppressor gene(s), the loss of which contributes to disease progression. Two genes, PRDM12 and EXOSC2, map to this region. To test this hypothesis, we examined expression of both genes in the hematopoietic hierarchy. Exosc2 expression occurs in the macrophage/neutrophil progenitor population, whereas Prdm12 was not expressed. Therefore, EXOSC2 may be a candidate tumour suppressor gene relevant in the initiation and/or progression of CML. We examined a panel of CML cell lines for inactivating mutations in both EXOSC2 and PRDM12. Sequencing results suggested neither gene possess coding mutations. However, we do suggest, based on expression and function, that EXOSC2 may be relevant to 9q34– CML progression. Chapter 3. 9q34 deletions in CML 87

3.2 Introduction

Chronic myeloid leukemia (CML) is a clonal disease of the hematopoietic compartment. The chronic phase is characterized by an overwhelming expansion of the myeloid lineage with a predominance of terminally differentiated granulocytes. Accelerated phase is dis- tinguished by the presence of hematopoietic blasts outside of hematopoietic tissues due to a differentiation block in the progenitor clone. The third and final stage is blast crisis, characterized by an overwhelming proportion of undifferentiated blasts in the peripheral blood. The BCR-ABL fusion gene is a result of the t(9;22)(q34;q11) chromosomal translocation (reviewed in [357]), which generates a smaller chromosome 22 named the Philadelphia chromosome, and a larger chromosome 9 called derivative(9) or der(9). A small minority of CML patients possess a complex translocation involving three or more chromosomes, but the result is ultimately the BCR-ABL fusion [358, 359]. Generation of the BCR-ABL fusion gene is the primary event in human CML initiation and is sufficient to cause a CML-like disease in a murine bone marrow transplant model [119, 360], reviewed in [117, 361]. Historically, CML has been treated with cytarabine [362–364] and/or IFN-α2b [365]. In 2001, imatinib mesylate (IM) was approved for treatment of CML [204,331], reviewed in [366]. IM is a rationally designed ATP mimetic that binds to the tyrosine kinase active site of ABL, resulting in inactivation of the BCR-ABL fusion protein [203, 216, 332]. Patients receiving IM typically respond well and experience a withdrawal of disease symptoms [204,367].

During the 9;22 chromosomal translocation 10 - 20% of patients undergo a hemizygous genomic deletion at the breakpoint of chromosome 9q34 [126,128,135]. The size of the deletion at 9q34 is variable between patients and was mapped in a cohort of 25 using quantitative PCR [132]. A 120 kilobase minimal deleted region (MDR) was common among all 25 patients and contained two genes – EXOSC2 and PRDM12. EXOSC2 is a component of the exosome complex (distinct from exosome microvesicles), an intracellular particle that trims and degrades certain species of RNA [134]. Many exosome targets Chapter 3. 9q34 deletions in CML 88 are highly labile, growth-promoting mRNAs that contain AU-rich elements in their 3’ untranslated regions [173, 177, 180, 368]. AU-rich elements help guide these transcripts for degradation [180]. A single study has been published documenting a role of a core exosome component in human disease [156], demonstrating the importance of a functional exosome in cell growth. Several recessive mutations in EXOSC3 were identified in eight families that result in profound nervous system maldevelopment. PRDM12 (positive regulatory (PR)-domain zinc-finger protein 12) is a member of a putative nuclear histone/protein methyltransferase superfamily, whose members have documented roles in stem cell development (reviewed in [183]), hematologic malignancies [184], as tumour suppressors (PRDM1 (BLIMP-1) [369, 370] and PRDM2 (RIZ) [371] in diffuse large B cell lymphoma), and oncogenes (PRDM3 (EVI1) [372] and PRDM16 [373] in myeloid leukemia). A previous report suggests loss of one PRDM12 allele may play a role in CML disease progression [133].

In patients treated with conventional allogeneic bone marrow transplants and IFN- α therapy, deletion status is a significant predictor of poor prognosis [126, 128, 135– 137]. However, there are conflicting data regarding the prognostic significance of a 9q34 deletion in patients treated with the tyrosine kinase inhibitor IM [138–142,374]. Several hypotheses have been developed to explain the differing response between 9q34+ and 9q34– patients to historical treatments [131]. We attempted to address two hypotheses to explain why patients with a 9q34 deletion experience a shorter progression-free and overall survival. First, it is possible that loss of a single copy of a tumour suppressor in the 9q34 region contributes to disease progression [131]. This hypothesis was addressed through studying the temporal and tissue-specific expression of EXOSC2 and PRDM12. Second, the deletion at 9q34 is a marker for an underlying genomic instability and not a contributor to disease progression. To address this second hypothesis, we undertook array comparative genomic hybridization to examine the genomic copy number in three 9q34+ and five 9q34– diagnostic CML samples. We identified no differences in gains or Chapter 3. 9q34 deletions in CML 89

losses between 9q34+ and 9q34–, but we did observe retinoblastoma 1 loss in two 9q34– patient samples.

3.3 Materials and methods

Chronic myeloid leukemia diagnostic sample preparation: Diagnostic chronic myeloid leukemia samples were obtained by the Toronto General Hospital diagnostic laboratory. Upon receipt by the lab, a fraction of the peripheral white blood cells was used for CML diagnostic tests, including fluorescent in situ hybridization (FISH) to determine 9q34 deletion status. The remainder was subjected to pheno-chloroform nucleic acid extraction and banking. Ten samples (five 9q34+ and five 9q34–) were kindly provided by Dr. Suzanne Kamel-Reid for this study. We used a portion of each sample for quantitative PCR to determine expression of genes that map to 9q34, and another for array comparative genomic hybridization (aCGH).

RNA isolation and cDNA generation from cell lines and CML patient samples: As quantitative real-time PCR (qPCR) positive controls, total RNA was isolated from the rhabdomyosarcoma cell line RH4 (high PRDM12 expression) and from the lung carcinoma cell line H661 (high EXOSC2 expression). RNA was extracted using TRIZOL (Invitrogen; Carlsbad, CA) as per the manufacturer’s directions. Cell lines were pelleted and lysed in 1 mL TRIZOL. The supernatant was isolated and incubated with 200 µL chloroform, and the solution was agitated and re-centrifuged. The aqueous layer was isolated and incubated with 500 µL 100% isopropanol, and the solution was centrifuged and the supernatant aspirated. The pellet was washed in 75% (v/v) RNAse-free ethanol, allowed to dry, and dissolved in autoclaved DEPC-treated water (0.1% (v/v) diethylpyro- carbonate). Quality was conﬁrmed using agarose gel electrophoresis and spectrophome- try. Isolated cell line RNA and patient sample RNA was used to generate complementary DNA (cDNA) with the Thermoscript RT-PCR system (Invitrogen). Three micrograms Chapter 3. 9q34 deletions in CML 90

of total RNA from cell lines and patients was used as template, pairing with the random hexamer or oligo dT primers for ampliﬁcation, as per the manufacturer’s directions.

Plasmids: For antibody generation, EXOSC2 cDNA was cloned into pGEX-2T (GE Healthcare). For qPCR standard curves, PRDM12, EXOSC2, and GAPDH were expressed in pMPG2.

TaqMan quantitative real-time polymerase chain reaction: TaqMan-based quantitative PCR (Applied Biosystems; Foster City, CA) was conducted to determine gene expression level of EXOSC2, PRDM12, GAPDH in cell lines and diagnostic CML samples. The TaqMan system consists of primer pairs flanking a fluorescently labeled probe. A probe contains the complementary nucleotide sequence of the target with covalently at- tached 5’ 6-carboxyfluorescein fluorophore and 3’ minor groove binder quencher. Upon primer binding 5’ of the probe site, the taq DNA polymerase extends the nucleotide sequence towards the probe. The 5’-3’ exonuclease activity of taq cleaves the 5’-most end of the probe and releases the fluorophore from proximity of the quencher, causing the emission of a photon at 518 nm allowing for quantification of the reaction. qPCR results were used to confirm deletion status predicted by the pathology lab’s diagnostic FISH assay and by our aCGH results. Sample cDNA was quantified in technical triplicates using the Sequence Detection Systems ABI Prism 7900HT real time PCR machine (Applied Biosystems). Primers were selected to span intro-exon boundaries to minimize genomic DNA-based false positives. Primers and probes used are as follows:

EXOSC2 forward (5’-GCACACGAGGAGCCTGAAATAT-3’) EXOSC2 reverse (5’-TGGGTCTTCTGCCGTTTCAC-3’) EXOSC2 probe (5’-CCTGACCTAGTTTTCC-3’) PRDM12 forward (5’-ACCACCGGAGCTGGATGAC-3’) PRDM12 reverse (5’-TCCGTACCACACCAGCAGTTC-3’) PRDM12 probe (5’-TTCTACAAGGCCATTGAGATGATCCCACCT-3’) GAPDH forward (5’-TGCACCACCAACTGCTTAGC-3’) Chapter 3. 9q34 deletions in CML 91

GAPDH reverse (5’-CCATCACGCCACAGTTTCC-3’) GAPDH probe (5’-AAGGACTCATGACCACAGTCCATGCCAT-3’)

Four microlites of cDNA from patient samples or cell lines was combined with 600 nM primer mix, 250 nM probe, and a reaction master mix containing AmpliTaq DNA polymerase, uracil-N -glycosylase, dNTPs, and dUTPs. Thermal cycle conditions were as follows:

o 2:00 o 10:00 o 0:15 o 1:00 50 C ; 95 C [95 C ; 60 C ]40

The initial two-minute incubation activates uracil-N -glycosylase, which degrades any uracil-containing nucleic acid contaminates (RNA). The following 10-minute step activates the AmpliTaq DNA polymerase. Forty cycles of 95oC and 60oC stimulate DNA denaturation and annealing / amplification, respectively. Results were analyzed with the SDS software version 2.2 (Applied Biosystems) on the Windows XP operating system. Array comparative genomic hybridization, experimental: The ten diagnostic CML samples were split upon receipt for qPCR and aCGH analysis. For the aCGH portion, sample quality and concentration was confirmed by spectrophotomery using a Nanodrop spectrophotometer (Thermo Scientific, Waltham, MA). Three of the ten samples fell below the concentration threshold for aCGH and were thus subjected to ethanol extraction and re-concentration. Post re-concentration, nine of ten samples were above the minimum cutoff of 50 ng/µL for the Affymetrix SNP 6.0 platform. Sample quality assessment, experimental processing, and data generation was performed by The Centre for Applied Genomics (TCAG; Toronto, ON). Array comparative genomic hybridization, data analysis: Initial data analysis was performed with Affymetrix Genotyping Console version 4.1.2 on the Windows 7 operating system. Further data analysis was completed with Nexus Copy Number version 6.0 on the OS X 10.7 operating system, access kindly provided by Dr. Jeremy Squire, Queen’s University, Kingston, ON. Cell lines and cell culture: Human embryonic kidney 293T (herein referred to as Chapter 3. 9q34 deletions in CML 92

‘293T’) [375] and K562 [247] cells were cultured in Dulbecco’s Modiﬁed Eagle’s Medium H21 supplemented with 10% heat-inactivated fetal calf serum (FCS – Thermo Fisher Scientiﬁc; Waltham, MA). CML-T1 [376], EM2 [377, 378], LAMA-84 [379], MC3 [380], and Meg-01 [381] cells were cultured in RPMI-1640 Medium supplemented with 10%

o heat-inactivated FCS. All cells were incubated at 37 C with 5% CO2.

293T cell transfection: 293T cells were seeded at 80% conﬂuency on a 10-cm dish (Corning; Lowell, MA) and were transfected the following day with 1 µg of the indicated plasmid using Lipofectamine-2000 (Invitrogen), according to the manufacturers instructions. Eighteen hours following transfection, cells were split into 6-cm plates (Sarstedt; Numbrecht, Germany) at 30% conﬂuency and allowed to grow a further 12 hours before experimental manipulation.

Cell Lysis and protein quantitation: Cells from cell lines were gently washed with magnesium and calcium-free phosphate-buﬀered saline (PBS -MgCl2 -CaCl2), followed by resuspension in 200 µL protein-isolation lysis buﬀer (1 M Tris-HCl pH 8.0; 4 M NaCl;

4% (v/v) Triton X-100; 0.5 M EDTA; 0.5 M Na4P2O7; 0.5 M NaF; 0.08 M Na3VO4; 0.5 M PMSF; one complete protease inhibitor cocktail tablet (Roche; Mannheim, Germany)). Cell lysates were incubated on ice for at least two minutes, and then cell debris was pelleted. Protein concentration was determined using the Bradford Colorimetric Assay (Bio-Rad; Hercules, CA) and detected at 595 nm using a Beckman Coulter DU 640 B spectrophotometer. Protein lysates were diluted 1:1 for in 2 x sample buﬀer (140 mM Tris-HCl pH 6.8; 22% (v/v) glycerol; 4.4% (w/v) SDS; 0.04% (w/v) bromophenol blue) containing 100 mM dithiothreitol (DTT from USB; Cleveland, OH) prior to immunoblot analysis.

SDS-PAGE and immunoblot: Whole-cell lysate was resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to a polyvinylidene diuoride (PVDF) membrane (NEN Life Science; Boston, MA). The membranes were blocked at room temperature for one hour with 2.5% (w/v) bovine serum antigen Chapter 3. 9q34 deletions in CML 93

(BSA from Sigma-Aldrich; St. Louis, MO) or 5% (w/v) powdered milk solubilized in tris-buffered saline containing Tween 20 (TBS-T – 50 mM Tris pH 8.0; 150 mM NaCl; 0.1% (v/v) Tween 20). Membranes were then incubated with an optimal concentration of the indicated primary antibody diluted in TBS-T as per the manufacturer’s directions. Post-primary antibody incubation, membranes were washed three times in TBS-T and incubated for 30 minutes in the suggested dilution of HRP-conjugated secondary in TBS-T. Membranes were washed three times in TBS-T and visualized by enhanced chemi-luminescence (Western Lightning Plus-ECL from Perkin Elmer; Waltham, MA) with auto-radiographic film (Kodak Biomax film from Kodak; Rochester, NY, Amersham Hyperfilm from GE Healthcare, and Biomax film from Cedar Lane; Burlington, ON).

Antibody generation: Full-length EXOSC2 cDNA was cloned into the pGEX-2T vector and used to transform Top10 ultracompetent E. coli (Invitrogen) as per the manufacturer’s directions. Successfully transformed E. coli were selected using an agar plate containing lysogeny broth (LB – 1% (w/v) bacto-tryptone; 0.5% (w/v) yeast extract; 1% (w/v) NaCl; pH 7.5) and 575 µM ampicillin. Individual clones were expanded with 250 mL LB with 575 µM ampicillin. During log growth phase, 100 µM IPTG was added to stimulate GST-EXOSC2 expression from the pGEX inducible tac promoter. Cultures were incubated for four hours, centrifuged, and resuspended in 25 mL NETN (20 mM Tris HCl (pH 8.0); 150 mM NaCl; 1 mM EDTA; 1% (w/v) Triton X-100). Cultures were kept at 4oC and sonicated three times at 60 Hz. Cell debris was pelleted at 8000 x g. Supernatant was combined with 500 µL glutathione sepharose 4B beads (GE Healthcare) and incubated overnight at 4oC with agitation. Post-incubation, the beads were washed with NETN followed by 20 mM glutathione to elute the bound GST fusion protein. The purified GST fusion protein was subjected to SDS-PAGE and quality was confirmed using coomassie staining (stain – 50% (v/v) methanol; 10% (v/v) acetic acid; 0.05% (w/v) bromophenol blue, destain – 50% (v/v) methanol; 10% (v/v) acetic acid). After confirmation of presence and quality, the purified GST-EXOSC2 fusion protein was shipped Chapter 3. 9q34 deletions in CML 94 to Biosynthesis(Lewisville, TX) for antibody generation. Two rabbits were inoculated with the full-length peptide. Serum was isolated at day zero, week six, week eight, and an exsanguination bleed at week 10. Immunoblots were performed with week 8 serum unless otherwise indicated.

Antibodies: The anti-FLAG antibody was purchased from Sigma-Aldrich. The anti- α-Tubulin antibody was purchased from Millipore (Billerica, MA). The horseradish peroxidase (HRP)-conjugated protein A, donkey anti-rabbit-HRP IgG, and sheep anti- mouse-HRP IgG antibodies were purchased from GE Healthcare UK (Little Chalfont, Buckinghamshire, UK). All monoclonal antibodies were used according to manufacturer’s directions. The polyclonal anti-EXOSC2-speciﬁc immunoblotting instructions are as follows. Membranes were incubated in 5% (w/v) powdered milk in TBS-T for one hour. Membranes were washed once brieﬂy in TBS-T, and incubated in a 1% (w/v) powdered milk TBS-T solution containing anti-EXOSC2 serum at 1:1000 dilution for one hour at room temperature, or over night at 4oC. Membranes were washed 3 times for 5 minutes each in TBS-T, followed by a one hour incubation in TBS-T containing a 1:5000 dilution of HRP-tagged Protein-A. Membranes were then washed 3 times in TBS-T and visualized as stated previously.

Global cDNA isolation from hematopoietic progenitors and amplification by PCR: Global cDNAs from hematopoietic progenitors were generated as previously outlined [382,383] by members of the Iscove laboratory, and kindly provided for this study. Briefly, CBA/J mouse bone marrow cells were sparsely seeded in methylcellulose medium containing c-kit ligand, IL-1, IL-3, IL-11, erythropoietin, and 5637 bladder carcinoma cell- conditioned medium. A single cell from early colony formation was isolated and immediately subjected to global cDNA amplification, and the differentiation status of that cell was projected from further maturation of sister cells. The final product is a cDNA sample that reflects gene expression at a specific, identified hematopoietic stage of differentiation. Stages include an early multipotent progenitor; early progenitors of the erythrocyte, Chapter 3. 9q34 deletions in CML 95 megakaryocyte, macrophage, neutrophil, and mast cell; further committed progenitors; and terminally differentiated effector cells. cDNA samples from each stage were subjected to PCR to query expression of EXOSC2 (accession number NM 144886), PRDM12 (accession number for incomplete cDNA XM 355325; genomic accession NT 039206), and ribosomal protein L32 (accession number BC046339) as a control. Primer sequences are as follows:

EXOSC2 forward (5’-AGCCTTACCCTAACCCCAGA-3’) EXOSC2 reverse (5’-GAGCAGTCGGGTGCTCTTAC-3’) PRDM12 forward (5’-GGTGCAAGGTCCAAATGTCT-3’) PRDM12 reverse (5’-CACACTGCAGCTGGCTAAAC-3’) L32 forward (5’-GGGAGCAACAAGAAAACCAA-3’) L32 reverse (5’-GGGATTGGTGACTCTGATGG-3’)

Polymerase chain reaction (PCR) was conducted with an Applied Biosystems Ge- neAmp PCR System 9700 thermocycler. Initial PCR conditions were optimized using vector control (data not shown). General PCR reaction mixes (50 µL total volume each) were prepared with Invitrogen reagents as follows: 1 x PCR buﬀer; 1.5 mM MgCl2; 0.1 mM dNTPs; 0.5 µM primer oligos; 1.0 unit recombinant Taq DNA polymerase; 100 ng cDNA template. PCR reactions were run as follows:

o 3:00 o 0:45 o 0:30 o 0:30 o 10:00 95 C ; [95 C ; 55 C ; 72 C ]35 ; 72 C

PCR samples were cooled to 4oC and stored for analysis. Samples were diluted in a DNA loading buﬀer (ﬁnal concentration of 5% (v/v) glycerol; 0.05% (w/v) bromophenol blue) and analyzed using agarose gel electrophoresis (1% (w/v) agarose; 40 mM Tris; 20 mM acetic acid; 1 mM EDTA; 0.5 µg/mL ethidium bromide).

Primer design and PCR for exon sequencing: For both EXOSC2 (accession number NC 000009.11; gene ID 23404) and PRDM12 (accession number NC 000009.11; gene ID 59335) genomic sequence was retrieved from the Gene database in the National Center Chapter 3. 9q34 deletions in CML 96

for Biotechnology Information online repository (http://www.ncbi.nlm.nih.gov/gene/), genome reference consortium human build 37, patch release 5 (GRCh37.p5). Exons were identified by aligning the mRNA sequence (EXOSC2, accession number NM 014285; PRDM12, accession number AY004252) with the total genomic sequence using Sequencher v4.6 (http://genecodes.com/) on the OS X operating system. Three hundred base pairs before and after each exon were used to query primer3 (http://frodo.wi.mit.edu/primer3/) for optimal primer location, using default variable numbers and selections. Sequencing primers were ordered from Sigma-Aldrich Canada (Oakville, Ontario). See Table 3.1 for primer sequence and amplicon size. PCR reactions were conducted as stated previously. Successfully amplified fragments were identified under ultraviolet light according to predicted size (see amplicon size, Table 3.1) and excised from the agarose gel. Purification of PCR fragments from agarose gel was conducted with the Qiagen (Hilden, Germany) QIAquick gel extraction kit. Fragments were sequenced by ACGT (Toronto, Ontario), and results were analyzed by Sequencher v4.6.

3.4 Results

Quantitative real-time PCR is able to identify 9q34 deletions in CML. Prior to commenc- ing array comparative genomic hybridization (aCGH) analysis of our ten patient samples, we sought to determine if expression of EXOSC2 and PRDM12 could predictably indicate 9q34 deletion status. We designed primers anchored in exons of EXOSC2, PRDM12, and GAPDH and conducted a standard curve analysis (Figure 3.1) with known concentrations of plasmid. Lines of best fit for EXOSC2 and PRDM12 were applied, and very similar slopes (0.04 units difference, 0.1 cutoff) indicated these genes could be directly compared. GAPDH was previously compared to PRDM12 by Stefanie Turley, a previous Barber laboratory student, and displayed near-identical slopes (data not shown). These results suggest the primer pairs amplify these three targets at the same rate, and Chapter 3. 9q34 deletions in CML 97

Sequencing target Primer sequence Amplicon size (bp) 5’- TGCTCTGGTTCCTTCTCCAT -3’ PRDM12 exon 1 542 5’- GCCCATGGGGAAACTGAG -3’ 5’- ACATCCGACTTCCTCCACAC -3’ PRDM12 exon 2 501 5’- CTTAGAGTCCCGGCTCAGG -3’ 5’- CCTCTCCAGCCTTCCTTTCT -3’ PRDM12 exon 3 391 5’- TCTGTCCACTGTGGTTCCTG -3’ 5’- AATGGGACTGTGTGGCTTTC -3’ PRDM12 exon 4 337 5’- GAGCCTCTCTGCCTCTCTCA -3’ 5’- CTGTCCCTCCAGTCCGTCT -3’ PRDM12 exon 5 775 5’- GTCCTCATCTGCGACACG -3’

5’- CCATCTCTCTTCCCCTCTCC -3’ PRDM12 exon 1 744 5’- GCTGTAAATGCCGGGAAC -3’ 5’- ACATCCGACTTCCTCCACAC -3’ PRDM12 exon 2 501 5’- CTTAGAGTCCCGGCTCAGG -3’ 5’- CTGTCCCTCCAGTCCGTCT -3’ PRDM12 exon 5 783 5’- CCCTCAGTGTCCTCATCTGC -3’

5’- ATTGGCAGCTGCTCTGAGTC -3’ EXOSC2 exon 1 467 5’- AGGTCAGGAGTGGAGGGTT -3’ 5’- ATGGCACAGTCAAGGTAGGC -3’ EXOSC2 exon 2 518 5’- GAAACCTCTGAAGCCTGCAC -3’ 5’- GGGTTTTTGGGCTAAATGAA -3’ EXOSC2 exon 3 455 5’- CTGAAATCGAACAAGCCACA -3’ 5’- CTCCCAAAGTGCTGGGATTA -3’ EXOSC2 exon 4 381 5’- GAGACCTGGACAGGGTGAAA -3’ 5’- AGATTGGGAGGGGTCTGAAT -3’ EXOSC2 exon 5 458 5’- AGAGCCCGACACTTGGTCTA -3’ 5’- GAGGGGAGGAAAAGATGGAG -3’ EXOSC2 exon 6 314 5’- CACGATCAGTTTGTGGGATG -3’ 5’- CTGAGTCCCAAAGCGTTCTC -3’ EXOSC2 exon 7 601 5’- TCACTGGCTCTTCGGCTATT -3’ 5’- CTCCCAAAGTGCTGGGATTA -3’ EXOSC2 exon 8 441 5’- ACTGACGGATGTTGGGAAAG -3’ 5’- GCCAGCTTTTGGATGAATGT -3’ EXOSC2 exon 9 488 5’- GGCTTGGTGTTAGGACAGGA -3’

Table 3.1: EXOSC2 and PRDM12 Exon Sequencing Primers. Forward (upstream) is listed ﬁrst among each pair, reverse (downstream) second. Two sets of primers for PRDM12 exons 1, 2, and 5 were attempted. See Table 3.8 for sequencing results. Pro- jected amplicon size is listed, which was used as a guideline when excising PCR product fragments. Chapter 3. 9q34 deletions in CML 98

are suitable to comparatively analyze expression of EXOSC2 and PRDM12 relative to GAPDH (the ∆Ct method, explained below).

We applied our real-time quantitative PCR (qPCR) protocol to ten diagnostic CML patient samples kindly provided by Dr. Suzanne Kamel-Reid, director of Molecular Diagnostics, Toronto General Hosptial. Five samples were called 9q34+ and five 9q34– by FISH analysis. Complementary DNA (cDNA) was generated from mRNA transcripts and the expression of EXOSC2, PRDM12, and GAPDH was determined (Table 3.2). CML10 was excluded from further qPCR analysis due to low GAPDH expression (exclusion based on Grubbs’ test for outliers, CML10 GAPDH expression z = 2.84). Excluding CML10, GAPDH was robustly and consistently expressed among our samples (average Ct was 18.64 ± 0.65). EXOSC2 was expressed ∼1550-fold lower than GAPDH (average Ct was 29.25 ± 1.06). PRDM12 was undetected in sample CML01, as well as in one of three technical triplicates in samples 02, 04, 05, 07, 09, and 10 (average Ct was 36.77 ± 1.50), an expression level ∼180-fold lower than EXOSC2 and ∼322,000-fold lower than GAPDH. Comparative threshold cycle ((Ct of gene of interest) – (Ct of GAPDH ) = ∆Ct) was used to determine relative expression of PRDM12 and EXOSC2 (Table 3.3). When grouping Ct values based on 9q34 deletion status, we observe a statistically significant decrease in both PRDM12 and EXOSC2 expression in 9q34– compared to 9q34+ (Figure 3.3; lower ∆Ct represents higher expression). These results suggest that qPCR-based detection of EXOSC2 and PRDM12 correlates with deletion status and that PRDM12 expression in terminally differentiated granulocytes is nearly undetectable.

Array comparative genomic hybridization reveals a stable genome independent of 9q34 or 22q11 deletion status. We hypothesized that the reduced survival time of 9q34– patients may be due to more genomic gains and losses, as compared to patients exhibiting a reciprocal translocation. In order to test this hypothesis, we undertook an array comparative genomic hybridization (aCGH) experiment to examine the gains and losses present Chapter 3. 9q34 deletions in CML 99

EXOSC2 qPCR standard curve PRDM12 qPCR standard curve 30

R2 = 0.999 30 R2 = 0.995 25 y = -4.02x + 20.47 y = -4.06x + 23.16

20 20 Cycle threshold Cycle Cycle threshold Cycle

-2 -1 0 1 -2 -1 0 1

Log10 [vector] pg Log10 [vector] pg

Figure 3.1: EXOSC2 and PRDM12 Quantitative PCR Standard Curves. Log10 plot of quantitative real-time quantitative PCR standard curves using 0.01, 0.1, 1, and 10 picograms of plasmid expressing either (a) EXOSC2 and (b) PRDM12. Each data point is the average of three separate reactions, error bars representing standard deviation. Standard curve slopes are less than 0.1 units diﬀerent, indicating the genes amplify at the same rate. Previous experiments demonstrate PRDM12 and GAPDH primers amplify at the same rate as well (data not shown), validating the use of GAPDH as an internal control. Chapter 3. 9q34 deletions in CML 100

Sample EXOSC2 Ct PRDM12 Ct GAPDH Ct CML01 30.12 ± 0.11 Undet. 18.54 ± 0.05 CML02 29.53 ± 0.16 36.72 ± 0.48 18.06 ± 0.01 CML03 28.92 ± 0.45 36.88 ± 0.13 18.85 ± 0.07 CML04 29.51 ± 0.14 36.99 ± 0.32 18.44 ± 0.10 CML05 28.90 ± 0.05 38.39 ± 2.35 18.07 ± 0.13 CML06 30.31 ± 0.08 35.12 ± 3.48 19.50 ± 0.12 CML07 26.67 ± 0.07 36.23 ± 0.44 18.53 ± 0.07 CML08 29.88 ± 0.14 36.84 ± 0.00 19.90 ± 0.04 CML09 29.41 ± 0.19 37.18 ± 0.02 17.90 ± 0.08 CML10 32.34 ± 0.20 36.64 ± 0.00 25.36 ± 0.21 Average 29.25 ± 1.06 36.77 ± 1.50 18.64 ± 0.65

Table 3.2: Quantitative PCR Raw Data from CML Patient Samples. Average cycle threshold of detection (Ct) for EXOSC2, PRDM12, and GAPDH expression from ten diagnostic CML patient samples. Undet. = undetermined, no reaction. Average Ct is calculated with three separate reactions and expressed plus/minus standard deviation. CML02, 04, 05, 07, 09, and 10 had one of three PRDM12 reactions undetermined. In such cases the average of two separate reactions was used. Average and standard deviation for each gene was calculated with all successful qPCR replicates. Chapter 3. 9q34 deletions in CML 101

Sample 9q34 status EXOSC2 ∆Ct PRDM12 ∆Ct CML01 – 11.58 ± 0.12 N/A CML02 – 11.47 ± 0.16 18.66 ± 0.48 CML03 + 10.07 ± 0.46 18.03 ± 0.15 CML04 – 11.07 ± 0.17 18.55 ± 0.34 CML05 – 10.83 ± 0.14 20.32 ± 2.35 CML06 + 10.81 ± 0.10 15.62 ± 3.48 CML07 + 8.14 ± 0.10 17.70 ± 0.45 CML08 + 9.98 ± 0.15 16.94 ± 0.04 CML09 – 11.51 ± 0.21 19.28 ± 0.08 CML10 – N/A N/A + 9.75 ± 0.33 17.07 ± 1.07 Average – 11.29 ± 1.14 19.02 ± 0.81

EXOSC2 PRDM12

22 * * 12 20

10 18 Ct Ct Δ Δ

16 8

9q34+ 9q34- 9q34+ 9q34- 9q34 status 9q34 status

Table 3.3: Quantitative PCR Relative Expression from CML Patient Samples. Above: qPCR average cycle threshold for EXOSC2, PRDM12, and GAPDH from ten diagnostic CML patient samples. Undet. = undetermined, no reaction. Average Ct is calculated with three separate reactions and expressed plus/minus standard deviation. CML02, 04, 05, 07, 09, and 10 had one of three PRDM12 reactions undetermined. In such cases the average of two separate reactions was used. Average for 9q34+ and 9q34– taken using ∆Ct, plus/minus standard deviation of those listed values (excluding their own standard deviation). Below: Graphical representation of tabular data. Asterisk indicates p <0.02. Note: Lower ∆Ct indicates higher expression. Chapter 3. 9q34 deletions in CML 102

in a cohort of ten CML patients. We initially obtained five FISH-tested 9q34+ and five 9q34– diagnostic CML samples. The status of 22q11 was not examined upon diagnosis. We chose the Affymetrix SNP 6.0 platform due to its extremely high probe density, in- tending to map the translocation breakpoint with very high resolution and detect small copy number aberrations. After DNA concentration, nine of ten samples met the minimum concentration of 50 ng/µL (sample CML08 excluded). The Centre for Applied Genomics (TCAG) performed the experiment and supplied the resulting raw data for analysis. Initial analysis with Affymetrix Genotyping Console suggested a low sample quality (>0.20) for all nine arrays (Table 3.4). Poor quality may be due to the age of the patient samples and/or the banking conditions, as we were limited in our access to samples called by FISH. Samples of low quality exhibit high levels of noise. Initial analysis suggests six of nine samples bear deletions at 9q34. Sample CML01 was labeled as 9q34+ with FISH, but was shown to be 9q34– with our array results (Figure 3.2). The incorrect FISH call may have been due to the relative size of the probe as compared to size of the deletion (Figure 3.2 and Table 3.5). In total, six of nine patient samples bear deletions at 9q34, size ranging from approximately 280 kilobases to two megabases (Figure 3.2 and Table 3.5). Six of nine samples have a deletion at 22q11 (Figure 3.3). Only two samples, CML06 and CML07, exhibit reciprocal translocations. Contiguous genomic loss ranges from 280 kb (CML01) to 3.31mb (CML05) (Table 3.5).

All samples with a 9q34 deletion have a hemizygous loss of the 5’ end of ABL, and complete deletions of EXOSC2, PRDM12, FUBP3, and ASS1. These data conﬁrm the initial report of the minimal deleted region (MDR) as previously reported [132, 135], with no deletions smaller than the reported 120 kb MDR. Deletion breakpoints are heterogeneous upon examination of the breakpoint probe intensities. Initial analysis with the Nexus software suite (default parameters for the SNP 6.0 platform) revealed Chapter 3. 9q34 deletions in CML 103

9q34 22q11 Sample Sample ID status status quality CML01 – + 0.35 CML02 – – 0.29 CML03 + – 0.32 CML04 – – 0.34 CML05 – – 0.31 CML06 + + 0.24 CML07 + + 0.24 CML09 – – 0.34 CML10 – – 0.25

Table 3.4: 9q34 and 22q11 Status in CML Samples as Called by aCGH. Sum- mary of aCGH analysis using Nexus Copy Number 6.0 software suite. Rank segmentation algorithm applied. A (–) represents a single-copy loss, a (+) represents no gain or loss. Significance threshold p < 10-5, minimum number of probes per segment = 5, high gain called at 0.75 and over, gain called at 0.3 and over, loss called at -0.3 and under, big loss called at -0.75 and under. Chapter 3. 9q34 deletions in CML 104 A representation of CML01 CML02 CML03 CML04 CML05 CML06 CML07 CML09 CML10 Figure 3.2:chromosome Nexus 9q34 at Software theproportion Visualization top, of of samples followed bear by thenumber a chromosome deletion variations, base 9q34 at and pair that SNP Locusbottom specific markers 6.0 region. panel. to in probes give Red is CML bars reference. outlined indicate Patient in a Percent deleted the at Samples. region. middle top Location of left of the indicate genes, figure. what exons, copy Specific sample deletions are displayed on the Chapter 3. 9q34 deletions in CML 105 A representation of CML01 CML02 CML03 CML04 CML05 CML06 CML07 CML09 CML10 Figure 3.3:chromosome Nexus 22q11 Software at the Visualizationproportion top, of of followed samples by bear the chromosomenumber a 22q11 base deletion variations, at pair and Locus that markers SNPbottom specific to in 6.0 region. panel. give probes CML Red reference. is bars Patient outlined Percent indicate at in a Samples. deleted top the region. left middle Location indicate of of what the genes, figure. exons, copy Specific sample deletions are displayed on the Chapter 3. 9q34 deletions in CML 106

Chromosome Breakpoint CML patient sample number Address distance (mb) Gene 01 02 03 04 05 06 07 09 10 2.04 ZER1 1.94 CCBL1 1.82 NUP188 1.78 SH3GLB2 1.12 PRRX2 9q34 0.94 USP20 0.60 FREQ 0.28 ASS1 0.16 FUBP3 0.08 PRDM12 0.06 EXOSC2 0.00 ABL1 0.00 BCR 0.35 IGLL1 0.70 GSTT2 1.33 SNRPD3 22q11 1.37 GGT1 1.58 SGSM1 1.97 CRYBB3

Table 3.5: A Genomic Map of the 9q34 and 22q11 Loci in CML Patient Samples. A representation of the 9q34 (from the ABL breakpoint, centromeric) and 22q11 (from the BCR breakpoint, telomeric) genomic regions surrounding the ABL-BCR fusion gene in nine CML diagnostic samples. Grey shaded boxes represent a bi-allelic result with the Aﬀymetrix SNP 6.0 array. White boxes represent a mono-allelic loss. CML06 and 07 had reciprocal translocations and therefore no genomic DNA loss. Chapter 3. 9q34 deletions in CML 107

many genomic aberrations in all nine samples, which contradicts ﬁndings by another group that suggests an average of 0.47 aberrations per chronic phase CML sample [384]. Frequent and highly homogenous gains and losses in repetitive sequence suggested much of the reported aberrations were artifacts of the sample quality and/or RNA isolation method. In order to correctly ﬁlter the data, we chose the size of known deletions at 9q34 as a template for actual genomic loss. The ‘single-copy loss’ threshold was set 0.3, which corresponded to slightly less than a mono-allelic loss, and correctly calls the 9q34 deletions. The size threshold was set to 100 kilobases, which again calls correctly the known 9q34 losses. However, these two parameters combined results in all genomes reported as normal, with the exception of 9q34 and 22q11 (Figure 3.4). Concerning the translocation breakpoint, 1/9 samples was only 9q34– (CML01), 1/9 samples was only 22q11– (CML03), 5/9 samples were 9q34– and 22q11– (CML02, 04, 05, 09, and 10), and 2/9 samples had reciprocal translocations (CML06 and 07).

From these data, we can assert that none of our samples bear gains or losses on the scale of those at 9q34 and 22q11. An alternate approach was required to identify aberrations smaller than those near the 9-22 translocation breakpoint. Manual ﬁltering was not practical considering the nine samples combine for thousands of reported aberrations, particularly when real data can so easily be dismissed as noise or missed entirely. In order to address this issue, we obtained data sets from two related studies [142, 385]. Unfortunately, one study did not make their data set public and the second used a platform incompatible with our analysis programs. Therefore, we manually compared their results to ours (gains reported in Table 3.6; losses reported in Table 3.7).

Two patients, CML09 and CML10, display small deletions of 8 and 18 kilobases, respectively, in the retinoblastoma 1 (RB1 ) gene (Figure 3.5a), a well-documented tumour suppressor [386–389]. These deletions lie in a probe-poor region of RB1, and span a small number of probes (Figure 3.5b). These data suggest the 9q34 and 22q11 breakpoints are Chapter 3. 9q34 deletions in CML 108 samples do not display a – Chromosome 9q34 and 22q11 display samples. + monoallelic loss in 6/9 ofgain samples called each. at 0.3 Otherwise,of there and are 100 over, loss no kb other called or detectable at smaller, aberrations. -0.3 which and correctly High under, gain calls big called all loss at deletions called 0.75 as at and 9q34 -0.75 over, and and 22q11. under. These Data data ﬁltered suggests to 9q34 remove aberrations Figure 3.4: Nexus-Generated Virtual Karyotype from CML Patient Samples. diﬀering number of aberrations than do 9q34 Chapter 3. 9q34 deletions in CML 109

Start End Marit Genes & Study Abnl. Chr no. band band Size (kb) overlap size (kb) Nadarajan et al. Gain 1 p36.33 p36.33 200 No Gain 1 p32.3 p32.2 160 No Gain 3 q27.1 q27.1 350 No Gain 4 q25 q25 350 No Gain 7 q22.1 q22.1 760 No Gain 8 q24.3 q24.3 2,140 No Gain 8 q24.3 q24.3 2,400 No Huh et al. Gain 8 p23.2 p23.2 2,221 No Nadarajan et al. Gain 10 q25.1 q25.1 40 No Gain 11 p15.5 p15.4 3,500 No Gain 12 q23.1 q23.1 480 No Gain 14 q32.33 q32.33 830 No Gain 16 q24.3 q24.3 130 No Gain 16 q24.2 q24.2 80 No Gain 18 q22.3 q22.3 60 No Gain 22 q13.33 q13.33 880 No

Table 3.6: CML Patient Copy Number Gains from Related Publications. Genomic gains, sorted by chromosome number, reported by Nadarajan [385] and Huh et al. [142]. Not listed are aberrations at 9q34 and 22q11, which are covered elsewhere. Losses greater than 15 mb are not listed. Overlap with Marit data requires that the aberration has less than 100% overlap with a copy number variation, is obviously not noise, and is within 100 kb of at least one genetic element (gene and/or miRNA). Chapter 3. 9q34 deletions in CML 110

Start End Marit Genes & Study Abnl. Chr no. band band Size (kb) overlap size (kb) Nadarajan et al. Loss 1 p36.23 p36.22 1,370 No Loss 1 p31.1 p31.1 340 No Loss 2 p16.1 p16.1 40 No Huh et al. Loss 2 q31.1 q31.1 929 No Loss 2 q37.1 q37.3 4,552 No Nadarajan et al. Loss 3 q13.11 q13.11 330 No Loss 3 q11.2 q11.2 330 No Loss 3 q11.2 q11.2 260 No Loss 5 q31.1 q31.1 630 No Loss 6 q16.3 q16.3 2,130 No Loss 7 q21.12 q21.12 100 No Huh et al. Loss 8 q23.1 q24.12 11,224 No Nadarajan et al. Loss 9 p21.3 p21.3 30 No Huh et al. Loss 11 q12.3 q12.3 505 No Loss 11 q13.5 q13.5 637 No Nadarajan et al. Loss 12 p13.32 p13.32 140 CML09, 10 RB1 (8, 18) Loss 13 q14.2 q14.2 7 No Huh et al. Loss 13 q14.3 q14.3 1,176 No Nadarajan et al. Loss 14 q23.2 q23.2 3 No Loss 15 q12 q12 130 No Loss 16 q21 q21 800 No Huh et al. Loss 16 p11.2 p11.2 2,921 No Nadarajan et al. Loss 17 q25.2 q25.3 1,520 No Loss 17 q25.1 q25.1 50 No Loss 17 q21.31 q21.31 60 No Loss 19 q13.2 q13.31 760 No Loss 20 q13.33 q13.33 40 No Loss 21 p11.1 p11.1 40 No Loss 22 q12.1 q12.1 2,980 No

Table 3.7: CML Patient Copy Number Losses from Related Publications. Genomic losses, sorted by chromosome number, reported by Nadarajan [385] and Huh et al. [142]. Not listed are aberrations at 9q34 and 22q11, which are covered elsewhere. Losses greater than 15 mb are not listed. Overlap with Marit data requires that the aberration has less than 100% overlap with a copy number variation, is obviously not noise, and is within 100 kb of at least one genetic element (gene and/or miRNA). Two samples from this study (CML09 and CML10) have copy loss in intron 16 of the retinoblastoma 1 gene. Chapter 3. 9q34 deletions in CML 111 heterogeneous in both location and size, and large-scale genomic aberrations do not diﬀer between patient cohorts. Additionally, RB1 loss may have some signiﬁcance in myeloid disease progression.

Generation and testing of anti-EXOSC2 antibody. In order to examine expression of EXOSC2 we generated two rabbit polyclonal antibodies, herein referred to as 4937 and 4938. Briefly, a GST-EXOSC2 (full length) fusion protein was isolated from E. coli and purified using Glutathoine Sepharose beads. GST-EXOSC2 was eluted from the beads and shipped to Biosynthesis (Lewisville, TX) for antibody generation. We received pre-innoculation, 6-week, 8-week post-innoculation, and exsanguination sera. To test the efficacy of the antibody, we transfected 293T cells with full-length human EXOSC2. Whole-cell lysate was separated on an SDS-PAGE gel, transferred to PVDF membrane and probed with pre-, 6-week, and 8-week serum from both animals (Fig- ure 3.6a). Lysate probed with 4938 demonstrates a strong signal at 33 kDa in the 8-week serum, corresponding to the correct molecular weight for human EXOSC2. Next, we expressed full-length EXOSC2 and EXOSC2 bearing an amino-terminal FLAG tag (DYKDDDDK x 3, ∼1 kDa x 3) [390, 391]. Whole-cell lysates probed with the 4938 antibody demonstrates a band at ∼33 kDa, corresponding to full-length EXOSC2, and a slightly larger band corresponding to FLAG-EXOSC2 (Figure 3.6b). The membrane was stripped and reprobed with an anti-FLAG antibody, demonstrating both adequate expression and detection of FLAG-EXOSC2, and an overlapping signal with the 4938 antibody. Next, we wanted to determine expression of EXOSC2 in a panel of human cell lines. In silico analysis of gene expression suggests that EXOSC2 is highly expressed in all tested human tissues (GeneAtlas U133A GC-RMA, www.biogps.org, probe 216117 at). Probing lysates from 293T (Figure 3.6d, lanes 1–4), human leukemia cell lines (lanes 5–10), and lung cell lines (thought to express EXOSC2 to a high degree, lanes 11–13) revealed comparable expression of EXOSC2 across these cell lines, which confirms the in silico array data. Interestingly, MC3 displays approximately equal EXOSC2 expression, Chapter 3. 9q34 deletions in CML 112

CML09 CML10 b)

CML09

CML10

Figure 3.5: Genomic Loss at 12q13.32 in the RB1 Tumour Suppressor Gene. Our nine CML aCGH samples were manually compared to two related studies [142,385]. (a) Deletions in the retinoblastoma 1 tumour suppressor gene was previously reported and conﬁrmed in our study. (b) CML09 and 10 display 8 and 18 kb deletions, respectively, encompassing a small number of probes. Chapter 3. 9q34 deletions in CML 113

despite being reported as 9q34– [392].

High EXOSC2 expression occurs at an early stage of hematopoietic development. The development and progression of CML is based on the unregulated expansion of hematopoietic progenitors. As the disease progresses from chronic, through accelerated, to blast crisis, the predominant pathologic cell type in the blood represents an increas- ingly earlier stage of development. As such, if EXOSC2 does indeed play a role in disease development, we would expect it to be expressed in early hematopoietic progenitors, as well as more fully differentiated and disease-relevant downstream cells. In order to test this hypothesis, we undertook an experiment looking at Exosc2 (murine orthologue also known as Rrp4 ) expression among murine hematopoietic progenitors. RNA was extracted from murine progenitors as early as the common myeloid progenitor through to fully differentiated effector cells. cDNA was generated and subjected to PCR targeting Exosc2, Prdm12, and L32, a ribosomal housekeeping gene. The expression of L32 was high among all progenitor cells, with the exception of the early myeloid progenitor where expression was low but still detectable (Figure 3.7). Exosc2 was detectable at the erythrocyte/megakaryocyte and downstream in the BFU-E stage. It was also expressed in the macrophage/neutrophil progenitor, through their precursors, and in the differentiated cells. Prdm12 was not detectable in any differentiation stage tested (data not shown). In summary, Exosc2 is expressed at detectable levels in late progenitor cells of the myeloid lineage, suggesting a role in development at these stages.

EXOSC2 and PRDM12 exon sequencing in BCR-ABL+ myeloid leukemia cell lines. If either EXOSC2 or PRDM12 acts as a tumour suppressor gene in CML, we would hypothesize either gene may undergo inactivation through point mutation of one or both copies. In order to test this hypothesis, we isolated genomic DNA from six myeloid leukemia cell lines (CMLT1, EM2, K562, LAMA84, MC3, and Meg01) to sequence EX- OSC2 and PRDM12 exons. We ampliﬁed all nine exons of EXOSC2 and two or three Chapter 3. 9q34 deletions in CML 114 a) 4937 serum 4938 serum Pre week 6 week 8 Pre week 6 week 8 EXOSC2

32 b) c) pcDNA3 EXOSC2 EXOSC2 pFLAG F-EXOSC2 F-EXOSC2 pcDNA3 EXOSC2 EXOSC2 pFLAG F-EXOSC2 F-EXOSC2

FLAG- 4937 4938 EXOSC2 serum serum EXOSC2 32 32

α-FLAG α-FLAG d) 293T Leukemia cell lines Lung cell lines pcDNA3 Mock F-EXOSC2 EXOSC2 CMLT1 EM2 K562 LAMA84 MC3 Meg01 A549 H520 H661 1 2 3 4 5 6 7 8 9 10111213 FLAG- EXOSC2 4938 serum EXOSC2

α-Tub α-Tubulin

Figure 3.6: EXOSC2 Antibody Validation. Antibodies were raised in two rabbits (4937 and 4938) against EXOSC2. Blood plasma was isolated prior to innoculation, at six weeks, and at eight weeks post-innoculation and used to probe for expression of EXOSC2 in 293T cells. Eight-week serum was used unless otherwise noted. (a) Cell lysates from 293T over-expressing EXOSC2 was probed with 4937 and 4938 sera, demonstrated emergence of the anti-EXOSC2 clone. (b) Cell lysates from 293T over- expressing wild-type or FLAG-tagged EXOSC2 and probed with 4937 serum and (c) 4938 serum. (d) Lysates from 293T and a panel of leukemia and lung cell lines demonstrating expression of EXOSC2. Chapter 3. 9q34 deletions in CML 115

L32 expression Exosc2 expression

E/Meg/Mac/N/Mast E/Meg/Mac/N/Mast

E/Meg/Mac/N E/Meg/Mac/N

E/Meg/Mac E/Meg/Mac

E/Meg Mac/N E/Meg Mac/N

BFU-E pMeg pMac pNeut BFU-E pMeg pMac pNeut

CFU-E CFU-E

E Meg Mac N Mast E Meg Mac N Mast

Figure 3.7: Exosc2 and L32 Expression in Murine Hematopoietic Progeni- tors. Mouse bone marrow progenitors were obtained and cultured with hematopoietic cytokine. At the four-cell stage, one cell was used for RNA isolation and global cDNA synthesis and the remaining sister cells were used to project the developmental stage of the single removed cell. Expression of Exosc2, Prdm12, and L32 was analyzed. Left: L32 was ubiquitously expressed among hematopoietic developmental stages. Right: Ex- osc2 was expressed at a detectible level in downstream progenitors and terminally differentiated cells. Not shown: Prdm12 was not successfully ampliﬁed an any progenitor population. Chapter 3. 9q34 deletions in CML 116

EXOSC2 exon PRDM12 exon Cell line 1 2 3 4 5 6 7 8 9 1 2 3 4 5 CMLT1 EM2 K562 LAMA84 MC3 Meg01

Table 3.8: EXOSC2 and PRDM12 Exon Sequencing. Light grey boxes represent successful sequencing result but no mutation. Dark grey represents successful sequencing and a mutation / single nucleotide polymorphism present. All four highlighted mutations / single nucleotide polymorphisms were found in non-coding and non-splice site regions upstream of their respective exons. Empty boxes represent unsuccessful sequencing. Chapter 3. 9q34 deletions in CML 117 of ﬁve PRDM12 exons. Multiple primer sets were used to target PRDM12 exons 1, 2, and 5 (Table 3.1), but none generated a PCR product. We identiﬁed one mutation/polymorphism upstream of EXOSC2 exon 7 in the LAMA84 cell line, and four non- coding polymorphisms/mutations in PRDM12 upstream of exon 4 in LAMA84, MC3, and Meg01 (Table 3.8).

3.5 Discussion

Chronic myeloid leukemia patient progression through chronic phase, accelerated phase, and blast crisis can be highly variable. Additionally, response to treatment can differ widely, with some patients not responding to IFN-α treatment and some display a durable long-term response. Such variability in patient progression and response may be attributed to genetic differences between disease-initiating clones. One such genetic lesion is the atypical translocation breakpoints observed on the derivative 9 chromosome after the t(9;22)(q34;q11) chromosomal translocation [393]. Early reports suggest 10 – 20% of patients exhibit a loss of the 5’ region of ABL, the 3’ region of BCR, or both [126, 393]. Further reports tracked 9q34+ and 9q34– patient survival when treated with IFN-α [126, 136], and determined that 9q34– patients fare significantly worse in long-term survival (56 months vs 37 months, p = 0.0001) [136]. More recent findings demonstrate IFN-α-treated patients with breakpoint-spanning deletions had poorer overall survival as compared with non-deleted patients (4.7 years vs 7.8 years, p = 0.003) [374]. Patients receiving imatinib do not display a significant difference (p = 0.078), however, a trend is apparent [374].

A review published in 2003 discusses four potential hypotheses concerning the survival differences observed in the two patient cohorts [131]. The first involves the lack of ABL-BCR expression that results from a 9q34 deletion. This hypothesis was initially refuted by demonstrating that ABL-BCR expression does not segregate with patient Chapter 3. 9q34 deletions in CML 118 survival [127,128]. However, recent findings suggest both the p40 and p95 forms of ABL- BCR are detectible in chronic phase CML and acute lymphoblastic leukemia primary patient samples [130]. In vitro work demonstrates that expression of either ABL-BCR fusion protein in early hematopoetic progenitors increases their re-plating efficiency and induces a block in differentiation, which is supported by a bone marrow transplant model demonstrating expression of p40 or p95 induces a CML-like disease over 200 or 350 days, respectively [130]. Paradoxically, if the ABL-BCR fusion protein is oncogenic we should expect to see better overall survival in 9q34–. Regardless, such findings will need to be explored in 9q34+ and 9q34– primary samples. The second hypothesis explaining 9q34+ and 9q34– survival difference suggests an increase BCR-ABL expression caused by deletion of regulatory elements upstream of BCR at the time of translocation. Multiple lines of evidence suggest that disease progression and severity are directly correlated with BCR-ABL expression [394,395]. However, quantitative PCR results demonstrate there is no difference in BCR-ABL expression between the two groups [128]. We have attempted to address the remaining two hypothesis in this study.

It is possible that the survival difference observed between 9q34+ and 9q34– patients is due to an underlying genomic instability. The deletion at the ABL-BCR translocation breakpoint would be a marker rather than contributor to progression. We addressed this hypothesis by performing aCGH with three 9q34+ and six 9q34– patient samples, as called by FISH upon diagnosis. We chose the Affymetrix SNP 6.0 platform due to its extremely high probe density, which allowed us to map the deleted region at a very high resolution. The Toronto General Hospital Molecular Diagnostics laboratory stopped examining deletion status upon diagnosis in the early 2000s, meaning access to samples of known 9q34 status would be at least 8-10 years old. The sample quality resulted in extreme noise and frequent artifacts in our raw data, making analysis difficult. To address this problem, we set our thresholds to remove calls smaller than 100 kb and set the call intensity to -0.3 (corresponding to a monoallelic loss). The result was correct Chapter 3. 9q34 deletions in CML 119 calling of all deletions, and otherwise normal genomes (Figure 3.4). Problematic to this approach, micro-deletions were commonly reported in previous studies [142,384,385] and make our analysis incomplete. We were limited in access to comparable data sets, so we analyzed our data relative to organized complete lists from other studies [142, 385]. Unfortunately, this decision relegated our results to confirmatory. Direct comparison of our results with published data suggests common loss in the reintoblastoma 1 (RB1 tumour suppressor gene (Table 3.6 and 3.7)). A role of RB1 in myeloid leukemias has been documented [396–398]. However, our results must be interpreted with caution because the deletions are extremely small and the probes reporting genomic loss in RB1 lie in a probe-poor region of intron 16 (Figure 3.5b). The homozygous deletion discovered in CML09 is based on the loss of a single probe, and hemizygous loss in CML10 is based on five probes. In order to confirm this result, we suggest examining genomic loss in these patients with qPCR probes anchored in the deleted region. Examination of transcript level with qPCR or protein expression with immunoblot would add further support for the result. In summary, our aCGH data demonstrate genomic loss heterogeneity in the 9q34/22q11 region following the t(9;22)(q34;q11) chromosomal translocation. Copy number artifacts make detection of novel micro-aberrations impossible, but we have confirmed RB1 aberrations on chromosome 12. In order to further examine the genomic instability hypothesis, we suggest that optimization of our EXOSC2 qPCR protocol would allow projection of 9q34 deletion status in a more cost-effective and high-throughput manner, as compared to traditional FISH. Such a technique would allow for simple calling of 9q34 status. A follow-up optimization aCGH experiment should be completed with fresh diagnostic samples to confirm these results.

The ﬁnal possibility to explain the diﬀering survival time is the presence of a tumour suppressor gene in the 9q34 deleted region. Deletion of one copy of a tumour suppressor followed by inactivation of the second copy is a frequent event in human cancers. We examined EXOSC2 and PRDM12, the two genes that lie in the previously pub- Chapter 3. 9q34 deletions in CML 120

lished minimal deleted region [132]. EXOSC2 is an integral part of the RNA-degrading exosome and has no published role in human disease [134]. PRDM12 is a member of the nuclear histone/protein methyltransferase superfamily, members of which have documented roles in human disease (reviewed in [184]). Publicly available microarray data suggest PRDM12 is restricted to cardiac myocytes, and possibly pancreas and liver cells (GeneAtlas U133A GC-RMA, www.biogps.org, probe 220894 x at). This result was supported by our qPCR studies that show PRDM12 was nearly undetectable in peripheral blood granulocytes (Table 3.2). We also demonstrated Prdm12 expression was not detectible in murine myeloid progenitors (data not shown). Immunoblot data from a previous student demonstrates lack of PRDM12 protein expression in a panel of CML cell lines. In summary, we suggest that despite initial suspicions, PRDM12 is not the most likely candidate for tumour suppressor loss in 9q34– CML. Therefore, if the tumour suppressor hypothesis is correct, EXOSC2 is the likely candidate gene. We demonstrated low but consistent EXOSC2 expression in peripheral blood granulocytes (Table 3.2), which confirms microarray data demonstrating its ubiquitous expression (GeneAtlas U133A GC-RMA, www.biogps.org, probe 216117 at). Expression of the murine Exosc2 orthologue (also known as Rrp4 ) was observed in the macrophage/neutrophil progenitor and more terminally differentiated cells in that lineage. Previous reports have suggested that the granulocyte-monocyte progenitor (GMP) is an important cytokinse in disease progression [399, 400]. Such temporal expression puts EXOSC2 at a critical stage of CML disease development. We suggest EXOSC2 is the prime target for the tumour suppressor hypothesis. Mechanistically, a reduced expression of EXOSC2 may result in fewer functional exosomes and consequential stabilization of their RNA targets. AU-rich element transcripts include the proto-oncogene c-myc and the cytokine IL-3 [175, 368, 401], both of which have a documented role in myeloid leukemia [402–404]. To confirm this mechanism, we propose a moderate RNAi-based reduction in EXOSC2 expression in cell lines followed by a qPCR query of specific AU-rich element transcripts. Pending positive re- Chapter 3. 9q34 deletions in CML 121

sults, a follow-up in primary 9q34+ and 9q34– CML patient samples could give strength to the tumour-suppressor hypothesis. Alternatively, AU-rich transcripts in expression microarrays could be analyzed to determine whether there is a speciﬁc expression sig- nature in 9q34– as compared to 9q34+. Finally, to support expression results, a serum ELISA could be performed with CML patient samples to examine the secretion of protein products from AU-element mRNA transcripts. Chapter 4

Discussion

122 Chapter 4. Discussion 123

Random Mutagenesis Reveals Residues of JAK2 Crit- ical in Evading Inhibition by a Tyrosine Kinase In- hibitor

Summary of rationale and results

Treatment of disease with a rationally designed small-molecule inhibitor presents significant selection pressure for outgrowth of resistant clones. Emergence of mutation-based patient relapse has been extensively documented with imatinib mesylate (IM)-treated chronic myeloid leukemia. Patient relapse is observed at approximately 2 to 7% per patient per year [405]. Prediction of ABL mutations that confer inhibitor resistance has significantly informed the design of next-generation inhibitors and spurred the development of non-ATP-competitive inhibitors. We hypothesized that the cytoplasmic tyrosine kinase JAK2 will respond similarly when targeted with its own inhibitors, aimed at treating JAK2 V617F+ myeloproliferative neoplasms. To address this hypothesis, we developed an in vitro screen to identify mutant JAK2 alleles that confer significant resistance to JAK Inhibitor-I, an ATP-mimetic inhibitor. Utilizing the TEL-JAK2 fusion oncogene, we identified and tested 10 kinase domain mutations, including the engineered mutation M929I, that mimics the pernicious BCR-ABL T315I mutation. Testing how these mutations support growth, affect downstream signalling, and phosphorylate the JAK2 activation loop sequence in the presence of inhibitor suggested that G935R confers significant resistance to JAK Inhibitor-I. G935R was mutated in the Jak2 V617F allele and the TEL-JAK2 results were recapitulated. Jak2 V617F G935R is at least 30-fold more resistant to JAK Inhibitor-I compared to Jak2 V617F alone. Such results confirm the use of the TEL-JAK2 oncogene in the initial screen, and has demonstrated inhibitor- resistant mutants of JAK2 can be isolated and should be expected as the use of JAK inhibitors in the clinic become more prevalent. Chapter 4. Discussion 124

Outstanding issues and thoughts pertaining to our study

There are a number of outstanding issues and curiosities within the data presented here. The initial screen identified a number of mutations that did not confer resistance to the presence of inhibitor. Indeed, only five of the initial 10 mutations tested displayed inhibitor resistance, which were identified when 66 colonies were expanded and sequenced. Could some or all of the remaining 56 colonies bear mutations elsewhere? The entire TEL-JAK2 fusion gene was mutagenized, creating the possibility of resistance-conferring mutations in other domains. Perhaps colonies that are consistently inhibitor resistant yet bear no kinase domain mutations be sequenced for additional TEL-JAK2 mutations elsewhere in the fusion gene. A second curiosity is the striking increase in protein expression of mutated TEL-JAK2, as compared to wild type, observed in BaF3 cells. This observation is consistent among stable BaF3 cell lines created with the listed constructs. We do observe a similar trend in 293T, although not nearly to the same degree. We observed no difference in TEL-JAK2 stability through lysosomal or proteasomal degradation, as queried through treatment with chloroquine and MG132, respectively (data not shown). We also examined protein half life through a cyclohexamide-based inhibition of the ribo- some (data not shown), which displayed similar degradation patterns between wild-type and mutant TEL-JAK2 proteins. However, the mutations do not affect the stability of Jak2 V617F protein products in the same way. What is causing this substantial difference in TEL-JAK2 allele expression or protein stability? Our working hypothesis is that kinase domain mutations weaken its ability to function, while granting selective advantage in the presence of inhibitor. Thus, in order to support the level of signalling required for BaF3 factor-independent growth, there is strong selective pressure to over-express or stabilize the TEL-JAK2 fusion protein. To test this hypothesis, we suggest transducing BaF3 cells but to support growth with saturating IL-3 stimulation, followed by immediate query of TEL-JAK2 protein expression. If the hypothesis is correct, we should observe consistent expression across TEL-JAK2 alleles. While this topic is interesting Chapter 4. Discussion 125 to consider, it is important to note that considerable differences in TEL-JAK2 protein expression was not predictive of survival in the presence of JAK Inhibitor-I (data not shown), and likely not a cause of differential survival or signalling. A second curious point is that relatively few mutations were inhibitor-resistant in the context of Jak2 V617F. Only one mutation, G935R, was truly inhibitor resistant when examined with survival, signalling, and kinase assays. Like G935R, BaF3 Jak2 V617F R975G displayed sustained Stat5 phosphorylation in our signalling experiments, but R975G cell growth and kinase assays suggest no difference between Jak2 V617F and Jak2 V617F R975G. Such a result is puzzling considering Stat5 is a critical downstream effector of JAK2 signalling and important in myeloproliferative neoplasm development in mice [406]. However, if we compare the activation of Erk1/2 and S6 kinase, we observe subtle increases in cells expressing Jak2 V617F G935R, which may be enough to confer the modest survival advantage observed in our cell survival assays. Overall, the relatively few transferrable mutations may be due to the potency of the kinases being studied or the oligimerization domain-specific interactions, as previously discussed.

The framework for random mutagenesis was outlined by Azam et al. in 2003 [208]. Following in 2006 was a screen identifying BCR-ABL mutations that bore striking overlap with clinical results [209] and gave confidence in the clinical predictability of the study and to its applicability in designing next-generation inhibitors. Given that the JAK2 V617F mutation was identified in 2005, it is likely that many groups were attempting to identify inhibitor-resistant mutations of the JAK2 kinase shortly after its discovery. The four current published examples were in late 2011 and 2012 [354–356,407], six years after the BCR-ABL mutagenesis screen methods were available. Why such an extensive delay between the BCR-ABL and JAK2 studies? The answer likely involves the potency in the oncogenes being studied. BCR-ABL is a robust and promiscuous tyrosine kinase that can readily transform cell lines. BCR-ABL-targeted small molecule inhibitor concentration can vary considerably, while inhibiting wild-type and allowing colonies bearing mutations Chapter 4. Discussion 126 to grow [209, 210]. Conversely, the JAK2 V617F oncogene is weak with limited transformation potential. Our results suggest the G935R mutation confers a modest 30-fold increase in small-molecule resistance, a finding that has been confirmed by other groups. In contrast, mutations in BCR-ABL can confer resistances 800-fold or higher [209, 210]. Our use of soft agar in our Jak2 V617F mutagenesis screen has caused us difficulty in differentiating between growth due to an inhibitor-resistant mutation or to missing the narrow inhibitor concentration window. One is likely to observe zero colonies because inhibitor concentration is too high, or a plate full of colonies all expressing Jak2 V617F with no inhibitor-resistant mutations. In order to address this issue, we utilized the TEL- JAK2 fusion oncogene that has tyrosine kinase activity much closer to BCR-ABL than its counterpart JAK2 V617F. This resulted in our rapid identification of JAK2 kinase domain mutations, eliminating the significant signal-to-noise problem. At least one other group used challenging deep sequencing strategies that allowed for the identification of very rare inhibitor-resistant Jak2 alleles [408]. Fortunately our assumption that, due to kinase domain identity, isolated TEL-JAK2 mutations would be directly transferrable to the JAK2 V617F oncogene was accurate.

This chapter outlined a proof-of-concept framework for identifying and testing JAK2 mutants that confer resistance to small-molecule inhibitor. The next step is to complete a similar screen to achieve amino acid saturation. This was goal was not accomplished in the present study as demonstrated by the lack of multiple clones expressing the same mutation. However, the results of this screen are no less diminished, as three other studies have reported mutations in the same residues with similar consequences [354– 356]. The TEL-JAK2 fusion gene has proved to be a useful tool in identifying inhibitor- resistant mutations and relieves the need for deep sequencing or extreme ﬁne-tuning of the inhibitor concentration. Expansion of these methods into a high-throughput liquid format may produce considerably more data, and sequencing the pseudokinase domain may yet identify inhibitor-resistant mutations. We suggest the continued use of the BaF3 Chapter 4. Discussion 127 cell and TEL-JAK2 fusion gene system, but to conduct the screen with ruxolitinib to yield clinically relevant results. Additionally, isolation of mutations with TEL-JAK2 and exclusive testing in Jak2 V617F may streamline the identiﬁcation and validation process. Testing inhibitor-resistant mutations in mouse models would be the next step in examining Jak2 inhibitor resistance. Use of murine bone marrow transplant models has recapitulated some aspects of the human disease [317–319], and would be an excellent tool in examining Jak2 V617F inhibitor sensitivity. Transduction of mouse bone marrow cells with a inhibitor-resistant Jak2 V617F allele and drug treatment would relatively simple to evaluate, considering the ruxolitinib treatment conditions and mouse health metrics are published [298].

In summary, the cell-based assay presented here is applicable to test inhibitors that block growth-dependent signals from any kinase enzyme. The JAK2 substrate GST fusion provides a simple method of analyzing kinase activity in the presence of inhibitor, and with slight modiﬁcations to the speciﬁc kinase substrate sequence can be applied to any kinase being studied.

Closing thoughts pertaining to JAK2 inhibitors

The probability of success of JAK2 inhibitors is currently an issue of hot debate, and there are at least three main issues. The first is the importance of the JAK2 tyrosine kinase and the JAK-STAT signalling cascade in hematopoietic development. Jak2 –/– mice fail to develop early in embryogenesis, displaying a lack of definitive hematopoiesis in the fetal liver [20]. This result is phenocopied by erythropoietin [1] and erythropoietin receptor [1,48] knockouts. Such results emphasize the fundamental role of the Epo, Epo-R, and Jak2 pathway in mammalian hematopoiesis. Therefore, it is of little surprise to observe the substantial side effects of the current generation of JAK2 inhibitors [409]. Treatment of myelofibrosis with the FDA-approved ruxolitinib can result in anemia and thrombocytopenia [301], two complications that are particularly problematic given that Chapter 4. Discussion 128 myelofibrosis patients are typically anemic at the commencement of treatment [230]. Whether patients can tolerate JAK2 inhibition in the long term remains to be seen. The second issue is whether we are missing too much of the primary MPN disease pathology by targeting JAK2 alone. PV, ET, and MF appear to be of stem cell origin, but unlike BCR-ABL in CML there is considerable genetic heterogeneity present at diagnosis [410]. Mutations in multiple genes have been described, including but not limited to MPL [252], IDH1/2 [411], ASXL1 [284], LNK [260] and TET2 [263]. Mutations in these genes do not exceed 20% of MPN patients [410], resulting in genetic diversity even within a specific disorder. Compounding the problem of heterogeneity, we observe multiple clones emerg- ing in a single individual as demonstrated in the progression of JAK2 V617F+ MPN to JAK2 V617F– AML [293]. Such findings question whether the JAK2 V617F mutation is a primary transforming event, or secondary to other genetic lesions. Perhaps stratification of the patient base according to secondary mutations would allow more success for targeted therapies. The third issue is whether the majority of MPN patients stand to benefit from JAK2 inhibitors at all. First-line care for PV and ET is a combination of hydroxyurea and aspirin [412]. Such treatment regimens contribute to a near-normal life expectancy for ET [229] and a median survival time of 19 years in PV [226]. Given such statistics, it is difficult to support a double-blind clinical trial to examine the efficacy of a JAK2 inhibitor in na¨ıve PV and ET patients, particularly given the side effects. Hydrox- yurea is commonly prescribed for primary and secondary MF however it does not resolve anemia, which is a significant factor in the 5.75-year life expectancy observed in these patients [413]. Therefore, it is of little surprise that the majority of JAK2 inhibitor trials are completed with MF patients, and it is that particular patient population that stands to benefit most from the development of JAK2 inhibitors. Treatment-refractory ET and PV patients are also patient populations that may benefit JAK2-selective inhibitors.

It is not known if the current generation of inhibitors will exert sufficient selective pressure to cause a mutated JAK2 clone to emerge and cause relapse. In order to esti- Chapter 4. Discussion 129 mate when we would observe clinical resistance, a comparison of imatinib mesylate (IM) and ruxolitinib timelines is as follows. IM patient treatment began in June 1998, phase II trials in June 1999 (229 patients), and phase III in June 2000 (1,106 patients) [414]. Mutation-based IM resistance was published in June 2001 [205] shortly after its FDA approval in May 2001. The total time between drug-treatment initiation and observation of mutation-based relapse was approximately 2 - 3 years. Ruxolitinib treatment was initiated in 2007. Results from phase I/II trials, conducted with JAK2 V617F+ or V617F– primary MF, post-ET MF, or post-PV MF patients, were published in 2010 (153 patients) [299]. Two randomized trials (COMFORT-1 and -2) were conducted with a similar patient cohort (528 patients) [300, 301, 409]. From ruxolitinib trial initiation to now, approximately five years have passed. If JAK2 inhibitor-resistant mutations in patients are possible, should they have been observed by now? In total, approximately 1,000 patients have been treated with JAK2 inhibitors for primary or secondary myelofibrosis [409], bringing to total close to the trial-phase numbers of IM. If the molecular mechanisms of JAK2-based disease is similar to BCR-ABL in CML, we would most likely have observed resistance. However, such a comparison is difficult. Ruxolitinib does not induce a near-complete cessation of symptoms, as does IM. For example, phase III ruxolitinib trial data suggests 29 - 42% of patients experienced a spleen response, 14% anemia response, a 44% thrombocytosis response, and an unremarkable reduction in JAK2 allele burden [300,301,409]. In contrast, IM induced a hematologic response in 95% of patients, and a cessation of disease progression in 89% [367]. Relapse in IM-treated CML may be much more obvious than relapse in ruxolitinib-treated MF. The genetic heterogeneity of MPNs may also contribute to lack of mutation-based relapse in patients. How much of the primary disease is JAK2 V617F dependent? If it is not the primary lesion in disease development, should we expect to see inhibitor-resistant mutations at all?

An important parameter of targeted therapy is how a drug modulates the activity of the mutated compared to wild-type protein. As is the case with JAK2, drug success Chapter 4. Discussion 130

depends on selective inhibition of the mutant if the wild-type target is essential. In IM- treated CML patients, inhibition of ABL, c-kit, and PDGF-R in adult tissues appear have relatively little impact on quality of life. However, in the case of JAK2-targeted disease, a V617F-specific drug may alleviate some of the most significant hurdles faced by the current class of drugs. Given what we have recently learned about the dual- specificity ‘pseudo’ kinase domain, is such a goal even feasible? The V617F mutation is thought to alleviate the auto-inhibitory function of the dual-specificity kinase domain, giving the kinase domain a higher level of basal activity. How could such an interaction be targeted with a drug? This question is perhaps the most significant moving forward with JAK2-specific drugs.

Examining Chromosome 9q34 Deletion as a Marker

for Genomic Instability in Chronic Myeloid Leukemia

Summary of rationale and results

Chronic myeloid leukemia is the result of the t(9;22)(q34;q11) chromosomal translocation. At the time of translocation, 10 to 20% of patients undergo a loss of genomic DNA at the 9q34 locus, immediately centromeric (5’) of the ABL breakpoint. Initial reports suggest that loss in this region is significantly predictive of worse survival on IFN-α, as compared with patients exhibiting a reciprocal translocation. Four hypotheses have been put forth to explain this finding. Previous studies have demonstrated that BCR-ABL expression does not correlate with deletion status, and thus does not explain the survival difference. Early studies claim that a stable ABL-BCR protein product not detectable, but a recent publication demonstrates that two isoforms are present in CML and B-ALL primary patient samples. Expression of ABL-BCR will need to be evaluated as a possible predictor of patient outcome. Herein we evaluated the remaining two hypotheses. Chapter 4. Discussion 131

First, we examined genomic copy number of nine diagnostic CML samples. Three 9q34+ and six 9q34– were queried with the Affymetrix SNP 6.0 array comparative genomic hybridization system. Results suggest that no sample contains a genomic aberration on the scale of the 9q34 losses (120 kilobases). Microdeletions were examined in comparison to two published studies, and we confirmed small deletions in the retinoblastoma 1 tumour suppressor. No other aberrations from similar studies were confirmed in our sample cohort. The fourth hypothesis suggests hemizygous loss of one or more tumour suppressors at 9q34 is responsible for a reduced survival time. PRDM12 seemed the likely candidate at study outset as it is a member of a family of genes known to have roles in cancer development. However, expression data from human patient samples and murine hematopoietic progenitors suggested PRDM12 is not expressed at a level com- patible with a role in disease progression. EXOSC2 is expressed in primary samples and at a stage in development that supports a role in disease. If the tumour suppressor hypothesis is correct, EXOSC2 is the most likely candidate.

Outstanding issues and thoughts pertaining to our study

Genomic instability has been documented in many types of human cancer and is in- variably a character of late-stage disease [415–417]. Mutations are acquired in single- or double-strand break repair pathways and result in inadequately maintained genomic DNA, gradual accumulation of mutations, and disease progression. In the case of 9q34– CML, if the genomic instability hypothesis is correct, such a mutation would have to occur before the t(9;22)(q34;q11) chromosomal translocation. DNA repair mutations in the non-homologous end-joining (NHEJ), microhomology-mediated end joining (MMEJ), or homologous repair (HR) may lead to a greater frequency of the Philadelphia translocation as a primary event, and perhaps loss of breakpoint genomic material at the time of translocation.

At the time of our aCGH experiment, the question of whether 9q34– patients suﬀered Chapter 4. Discussion 132

from genomic instability was still outstanding [131]. Since then, three groups have undertaken similar studies and did not report a difference in genomic aberrations between 9q34+ and 9q34– [142, 384, 385]. Disease progression is characterized by an unstable genome, but no difference between the two patient cohorts is noted. Our aCGH results give weak support to the existing body of evidence, however, excessive noise in our nine samples prevented examination of micro-deletions and relegated these results to confirmatory. We were able to finely map the 9q34 and 22q11 regions among our samples, which display heterogeneity in deletion size. We can confidently state that none of our nine samples bear aberrations on the scale of the 9q34 and 22q11 loss. Finally, we were able to support a recent publication documenting loss of the retinoblastoma 1 gene in chronic phase CML [142]. Considering our results and those published, it appears genome stability does not differ between 9q34+ and 9q34– patients. The remaining hypothesis posits deletion of at tumour suppressor at 9q34 results in the differential survival time.

Our results indicate that PRDM12 is not a candidate for the tumour suppressor hypothesis. Expression data from three sources (BioGPS, our patient sample qPCR, and murine hematopoietic progenitors expression query) suggest the expression of PRDM12 in the mammalian hematopoietic compartment is almost undetectable. Lack of protein expression in a panel of CML cell lines was conﬁrmed by a previous student in the Barber laboratory (Stefanie Turley, MSc thesis [418]). Such results strongly suggest that PRDM12 does not have a role in blood cell growth and diﬀerentiation. Consequently, we will discuss how low EXOSC2 expression may play a role in CML disease development and progression.

EXOSC2 (also known as hRrp4) is a component of the eukaryotic exosome, a nine- or ten-subunit exoribonuclease complex [163] that exists in both the cytoplasm and nucleus [419]. With the help of accessory proteins, snRNA, snoRNA, and rRNA are targeted for 3’ processing by the 3’ - 5’ exonuclease domains associated with the core particle [147, 161, 162]. In contrast, AU-rich element mRNAs are targeted for processive degradation Chapter 4. Discussion 133

[180]. Initial studies describing the structure and function of the eukaryotic exosome were completed in S. cerevisiae [134, 154, 420, 421]. Mutational studies demonstrated each exosome subunit to be essential to life (Table 1.1). Biochemical characterization of the mammalian exosome demonstrated the predicted similarity in composition and function [179]. Our conceptual understanding of the exosome was greatly improved with the publishing of the human exosome crystal structure, which described a hollow cylinder composed of six unique subunits, and a cap of three subunits [147]. The six proteins that make up the core have modified RNase PH domains, which lack catalytic activity and play a role in substrate specificity [157, 158]. EXOSC2 is one of three cap proteins, all of which contain either KH or S1 RNA-binding domains. We suggest here that the loss of EXOSC2 would reduce the number of functional exosome complexes, and such a reduction would have specific and testable consequences on cellular function. To test this hypothesis, one could induce RNAi-based expression modulation of exosome components, followed by a qPCR query of exosome substrate mRNA. Most significant to disease development, we would predict a stabilization of AU-rich elements and therefore suggest examining expression of GM-CSF, Ras, and IL-3 [368]. The functional consequence of a reduction of exosomes on non-coding RNA species is currently unknown. Each member of the S. cerevisiae exosome complex is essential to growth in haploid yeast models [134]. Comparatively, the importance of the human exosome is demonstrated by its lack of roles in human disease. The first publication documenting a role of a core exosome component in human disease was released this year, describing the role of EXOSC3 in pontocerebellar hypoplasia and spinal motor neuron degeneration [156]. The molecular consequence of those mutations has yet to be described. Concerning CML, can hematopoietic cells recover from a loss of half their functional exosomes? Will the hypothesized stabilization of certain growth-promoting RNA species outweigh the lack of non-coding RNA maturation? Further molecular and growth studies will need to be conducted to answer these questions. Chapter 4. Discussion 134

The consequence of 9q34 deletions on transcript stability can also be addressed in a more high-throughput and clinically relevant manner. 9q34+ and 9q34– patient samples could be isolated and subjected to custom expression microarrays, with the choice of microarray targets selected as follows. The AU-rich element mRNA transcriptome has been documented and curated in the AU-rich element database (ARED) [175–177]. We ran a simple sorting program that compared the ∼2,700 transcripts in ARED with genes upregulated in CD34+ CML stem and progenitor cells (GEO Profile GDS2342). The result is 140 transcripts that were significantly upregulated in CD34+ CML cells and also contain AU-rich elements, some of which will be described here. VEGFA is a gene that plays a role in internal and external autocrine loops and regulates migration and survival in acute leukemia cells [422]. Aberrant promoter demethylation results in increased PAX4 expression in hematologic malignancies, and studies in cell lines confirmed that over-expression results in increased growth [423]. CDC6 is over-expressed in a variety of human cancers [424,425], and leads to silencing of the INK4/ARF locus via recruitment of histone deacetylases [426]. Therefore, it would be of particular interest to query these and other genes with documented roles in CML to determine if there is a general stability increase in the 9q34– patient cohort.

A second bioinformatic approach is to retroactively examine expression of the 140 transcripts (isolated from our ARED / CML microarray comparison) among a large database of CML patient expression microarrays. If our mRNA stability hypothesis is correct, unsupervised clustering according to expression level of those 140 transcripts may result in two groups – one containing 10 - 20% of patients (representing 9q34–), and the second 80 - 90% (9q34+). Examination of EXOSC2 expression may support the clustering (low EXOSC2 expression in 10 - 20%, high in the 80 - 90%), allowing us a post-experimental prediction of 9q34 deletion status. Unfortunately, at the time of this writing no such database of CML expression array samples was available.

Should a role for the exosome-based degradation of significant mRNAs be suggested Chapter 4. Discussion 135 by clinical experiments, in vivo experiments will be required to validate the findings. All exosome components are essential in yeast and the importance of a functional human particle is evident from the single human disease caused by a mutation in EXOSC3 [156]. Germline knockouts would be of limited value due to the potential of embryonic lethality. Since we are interested in haploinsufficiency, we suggest a generating an Exosc2 transgenic with loxP sites [427, 428] flanking Exosc2. Mating a loxP-Exosc2 mouse with a partner expressing Cre under the vav1 [322] or Mx1 [323, 324] promoter would result in a reduced Exosc2 expression in the hematopoietic lineages. It is likely that the Mx1 promoter would be more informative because hemizygosity can be induced in adulthood through pI:pC administration, rather than be present throughout gestation and development where it may negatively impact development. Bone marrow from these transgenic lines can be subjected to the BMT model discussed earlier (section 1.3.4). Exosc2 +/– bone marrow marrow cells infected with a BCR-ABL retrovirus and transplanted would be an extremely interesting model to study the 9q34 deletion. If Exosc2 is important in the murine BMT model, we suggest that mice transplanted with Exosc2 +/– marrow expressing BCR-ABL would die with a shorter disease latency than mice with wild-type marrow transduced with BCR-ABL. A final question to consider is why 9q34 deletions are predictive of progression-free survival in patients treated with IFN-α, but that difference is lost when treated with small-molecule inhibitors. Unfortunately, the mechanism of IFN-α-mediated suppression of the leukemic clone was not well understood at the time of IM development [429], and the tremendous success of IM therapy has reduced the incentive and motivation to explore this topic further. One hypothesis is that IFN-α acts as an anti-proliferative signal and suppress expansion of the leukemic clone through signal transduction and gene transcription. A thorough comparison of type-I interferon-inducible genes with the ARED database may help to address this question. Chapter 5

Concluding Remarks

136 Chapter 5. Concluding Remarks 137

Hematopoiesis is the development and maturation of blood cells. The Barber laboratory studies this process both from a fundamental and disease perspective. In this thesis, we have examined two distinct perspectives of myeloproliferative neoplasm progression. Chapter Two used an in vitro screen to identify mutations of JAK2 that are insensitive to the small-molecule inhibitor JAK Inhibitor-I. We identified a panel of mutations and demonstrated an increase in growth and JAK2 downstream signalling in the presence of high concentrations of inhibitor. We developed and utilized a novel kinase assay that fuses the JAK2 activation loop to a GST substrate, which allowed us to directly query the level of inhibitor resistance conferred by specific mutations. G935R was identified as a JAK2 mutation that conferred significant resistance to JAK Inhibitor-I, and was corroborated by several other groups at the time of this writing. Chapter Three examines the underlying cause of rapid disease progression in cohort a cohort of CML patients that incurred a deletion at the 9q34 locus, one of the two BCR-ABL translocation breakpoints. Results suggest that neither cohort experience large-scale aberrations, and may not have differing genomic stability. Considering the function and temporal expression of EXOSC2, one of the genes that map to 9q34, we suggest it may play a significant role in disease progression. Collectively, the data presented here furthers our knowledge of myeloproliferative neoplasms in both a clinical and fundamental manner. We will continue to investigate the the role of JAK2 mutants in disease progression through mouse models, and suggest that the results herein presented help steer the course of next-generation JAK2 inhibitor design. References

[1] H. Wu, X. Liu, R. Jaenisch, and H.F. Lodish. Generation of committed erythroid bfu-e and cfu-e progenitors does not require erythropoietin or the erythropoietin receptor. Cell, 83(1):59–67, 1995.

[2] O. Livnah, E.A. Stura, S.A. Middleton, D.L. Johnson, L.K. Jolliﬀe, and I.A. Wilson. Crystallographic evidence for preformed dimers of erythropoietin receptor before ligand activation. Science, 283(5404):987, 1999.

[3] LO Jacobson, E. Goldwasser, W. Fried, and L. PlzaK. Role of the kidney in erythropoiesis. Nature, 179:633 – 634, 1957.

[4] A.P. Maxwell, T.R.J. Lappin, C.F. Johnston, J.M. Bridges, and M.G. McGeown. Erythro- poietin production in kidney tubular cells. British journal of haematology, 74(4):535–539, 1990.

[5] R.F. Chapman, J. Stray-Gundersen, and B.D. Levine. Individual variation in response to altitude training. Journal of applied physiology, 85(4):1448–1456, 1998.

[6] I. Remy, I.A. Wilson, and S.W. Michnick. Erythropoietin receptor activation by a ligand- induced conformation change. Science, 283(5404):990, 1999.

[7] S. Pelletier, S. Gingras, M. Funakoshi-Tago, S. Howell, and J.N. Ihle. Two domains of the erythropoietin receptor are suﬃcient for jak2 binding/activation and function. Molecular and cellular biology, 26(22):8527–8538, 2006.

[8] N. Jiang, T.C. He, A. Miyajima, and D.M. Wojchowski. The box1 domain of the erythropoietin receptor speciﬁes janus kinase 2 activation and functions mitogenically within an interleukin 2 β-receptor chimera. Journal of Biological Chemistry, 271(28):16472–16476, 1996.

[9] B.A. Witthuhn, F.W. Quelle, O. Silvennoinen, T. Yi, B. Tang, O. Miura, and J.N. Ihle. Jak2 associates with the erythropoietin receptor and is tyrosine phosphorylated and activated following stimulation with erythropoietin. Cell, 74(2):227–236, 1993.

[10] AF Wilks, AG Harpur, RR Kurban, SJ Ralph, G. Z¨urcher, and A. Ziemiecki. Two novel protein-tyrosine kinases, each with a second phosphotransferase-related catalytic domain, deﬁne a new class of protein kinase. Molecular and cellular biology, 11(4):2057–2065, 1991.

[11] AG Harpur, AC Andres, A. Ziemiecki, RR Aston, AF Wilks, et al. Jak2, a third member of the jak family of protein tyrosine kinases. Oncogene, 7(7):1347, 1992.

138 References 139

[12] L.J. Huang, S.N. Constantinescu, and H.F. Lodish. The n-terminal domain of janus kinase 2 is required for golgi processing and cell surface expression of erythropoietin receptor. Molecular cell, 8(6):1327–1338, 2001.

[13] D. Kampa and J. Burnside. Computational and functional analysis of the putative sh2 domain in janus kinases. Biochemical and biophysical research communications, 278(1):175– 182, 2000.

[14] S.P. Gorantla, T.N. Dechow, R. Grundler, A.L. Illert, C.M. zum B¨uschenfelde, M. Kremer, C. Peschel, and J. Duyster. Oncogenic jak2v617f requires an intact sh2-like domain for constitutive activation and induction of a myeloproliferative disease in mice. Blood, 116(22):4600–4611, 2010.

[15] D. Ungureanu, J. Wu, T. Pekkala, Y. Niranjan, C. Young, O.N. Jensen, C.F. Xu, T.A. Neubert, R.C. Skoda, S.R. Hubbard, et al. The pseudokinase domain of jak2 is a dual- speciﬁcity protein kinase that negatively regulates cytokine signaling. Nature Structural and Molecular Biology, 18(9):971–976, 2011.

[16] J. Feng, B.A. Witthuhn, T. Matsuda, F. Kohlhuber, I.M. Kerr, and J.N. Ihle. Activation of jak2 catalytic activity requires phosphorylation of y1007 in the kinase activation loop. Molecular and cellular biology, 17(5):2497–2501, 1997.

[17] JE Damen, H. Wakao, A. Miyajima, J. Krosl, RK Humphries, RL Cutler, and G. Krystal. Tyrosine 343 in the erythropoietin receptor positively regulates erythropoietin-induced cell proliferation and stat5 activation. The EMBO journal, 14(22):5557, 1995.

[18] F.W. Quelle, D. Wang, T. Nosaka, W.E. Thierfelder, D. Stravopodis, Y. Weinstein, and J.N. Ihle. Erythropoietin induces activation of stat5 through association with speciﬁc tyrosines on the receptor that are not required for a mitogenic response. Molecular and cellular biology, 16(4):1622–1631, 1996.

[19] S. Gobert, S. Chretien, F. Gouilleux, O. Muller, C. Pallard, I. Dusanter-Fourt, B. Groner, C. Lacombe, S. Gisselbrecht, and P. Mayeux. Identiﬁcation of tyrosine residues within the intracellular domain of the erythropoietin receptor crucial for stat5 activation. The EMBO journal, 15(10):2434, 1996.

[20] E. Parganas, D. Wang, D. Stravopodis, D.J. Topham, J.C. Marine, S. Teglund, E.F. Vanin, S. Bodner, O.R. Colamonici, J.M. van Deursen, et al. Jak2 is essential for signaling through a variety of cytokine receptors. Cell, 93(3):385–395, 1998.

[21] O. Silvennoinen, B.A. Witthuhn, F.W. Quelle, J.L. Cleveland, T. Yi, and J.N. Ihle. Structure of the murine jak2 protein-tyrosine kinase and its role in interleukin 3 signal transduction. Proceedings of the National Academy of Sciences, 90(18):8429, 1993.

[22] S.E. Doyle and J.C. Gasson. Characterization of the role of the human granulocyte- macrophage colony-stimulating factor receptor α subunit in the activation of jak2 and stat5. Blood, 92(3):867–876, 1998.

[23] S.T. Sawyer and K. Penta. Association of jak2 and stat5 with erythropoietin receptors. Journal of Biological Chemistry, 271(50):32430, 1996. References 140

[24] M. Sattler, MA Durstin, DA Frank, K. Okuda, K. Kaushansky, R. Salgia, and JD Griﬃn. The thrombopoietin receptor c-mpl activates jak2 and tyk2 tyrosine kinases. Experimental Hematology, 23(9):1040, 1995.

[25] J.G. Drachman, J.D. Griﬃn, and K. Kaushansky. The c-mpl ligand (thrombopoietin) stimulates tyrosine phosphorylation of jak2, shc, and c-mpl. Journal of Biological Chem- istry, 270(10):4979, 1995.

[26] K. Igarashi, G. Garotta, L. Ozmen, A. Ziemiecki, A.F. Wilks, A.G. Harpur, A.C. Larner, and D.S. Finbloom. Interferon-gamma induces tyrosine phosphorylation of interferon- gamma receptor and regulated association of protein tyrosine kinases, jak1 and jak2, with its receptor. Journal of Biological Chemistry, 269(20):14333–14336, 1994.

[27] M. Narazaki, B.A. Witthuhn, K. Yoshida, O. Silvennoinen, K. Yasukawa, J.N. Ihle, T. Kishimoto, and T. Taga. Activation of jak2 kinase mediated by the interleukin 6 signal transducer gp130. Proceedings of the National Academy of Sciences, 91(6):2285, 1994.

[28] P.C. Heinrich, I. Behrmann, G. M¨uller-Newen, F. Schaper, and L. Graeve. Interleukin- 6-type cytokine signalling through the gp130/jak/stat pathway. Biochemical Journal, 334(Pt 2):297, 1998.

[29] A. Oda, K. Sawada, B.J. Druker, K. Ozaki, H. Takano, K. Koizumi, Y. Fukada, M. Handa, T. Koike, and Y. Ikeda. Erythropoietin induces tyrosine phosphorylation of jak2, stat5a, and stat5b in primary cultured human erythroid precursors. Blood, 92(2):443–451, 1998.

[30] D.L. Barber, B.K. Beattie, J.M. Mason, M.H.H. Nguyen, M. Yoakim, B.G. Neel, A.D. D’Andrea, and D.A. Frank. A common epitope is shared by activated signal transducer and activator of transcription-5 (stat5) and the phosphorylated erythropoietin receptor: implications for the docking model of stat activation. Blood, 97(8):2230–2237, 2001.

[31] K. Kirito, K. Nakajima, T. Watanabe, M. Uchida, M. Tanaka, K. Ozawa, and N. Ko- matsu. Identiﬁcation of the human erythropoietin receptor region required for stat1 and stat3 activation. Blood, 99(1):102–110, 2002.

[32] C.M. Horvath, Z. Wen, and JE Darnell. A stat protein domain that determines dna sequence recognition suggests a novel dna-binding domain. Genes & development, 9(8):984– 994, 1995.

[33] X. Xu, Y.L. Sun, and T. Hoey. Cooperative dna binding and sequence-selective recognition conferred by the stat amino-terminal domain. Science, 273(5276):794–797, 1996.

[34] A. Halupa, M.L. Bailey, K. Huang, N.N. Iscove, D.E. Levy, and D.L. Barber. A novel role for stat1 in regulating murine erythropoiesis: deletion of stat1 results in overall reduction of erythroid progenitors and alters their distribution. Blood, 105(2):552–561, 2005.

[35] T. Tauchi, G.S. Feng, R. Shen, M. Hoatlin, G.C. Bagby Jr, D. Kabat, L. Lu, and H.E. Broxmeyer. Involvement of sh2-containing phosphotyrosine phosphatase syp in erythropoietin receptor signal transduction pathways. Journal of Biological Chemistry, 270(10):5631–5635, 1995. References 141

[36] R.L. Cutler, L. Liu, JE Damen, and G. Krystal. Multiple cytokines induce the tyrosine phosphorylation of shc and its association with grb2 in hemopoietic cells. Journal of Biological Chemistry, 268(29):21463–21465, 1993.

[37] T. Tauchi, J.E. Damen, K. Toyama, G.S. Feng, H.E. Broxmeyer, and G. Krystal. Tyrosine 425 within the activated erythropoietin receptor binds syp, reduces the erythropoietin required for syp tyrosine phosphorylation, and promotes mitogenesis. Blood, 87(11):4495– 4501, 1996.

[38] J.M. Mason, B.K. Beattie, Q. Liu, D.J. Dumont, and D.L. Barber. The sh2 inositol 5- phosphatase ship1 is recruited in an sh2-dependent manner to the erythropoietin receptor. Journal of Biological Chemistry, 275(6):4398, 2000.

[39] Y. Miura, O. Miura, J.N. Ihle, and N. Aoki. Activation of the mitogen-activated protein kinase pathway by the erythropoietin receptor. Journal of Biological Chemistry, 269(47):29962–29969, 1994.

[40] X. Sui, S.B. Krantz, M. You, and Z. Zhao. Synergistic activation of map kinase (erk1/2) by erythropoietin and stem cell factor is essential for expanded erythropoiesis. Blood, 92(4):1142–1149, 1998.

[41] J.E. Damen, R.L. Cutler, H. Jiao, T. Yi, and G. Krystal. Phosphorylation of tyrosine 503 in the erythropoietin receptor (epr) is essential for binding the p85 subunit of phosphatidylinositol (pi) 3-kinase and for epr-associated pi 3-kinase activity. Journal of Biological Chemistry, 270(40):23402–23408, 1995.

[42] A. Wickrema, S. Uddin, A. Sharma, F. Chen, Y. Alsayed, S. Ahmad, S.T. Sawyer, G. Krystal, T. Yi, K. Nishada, et al. Engagement of gab1 and gab2 in erythropoietin signaling. Journal of Biological Chemistry, 274(35):24469–24474, 1999.

[43] J.E. Damen, AL Mui, L. Puil, T. Pawson, and G. Krystal. Phosphatidylinositol 3-kinase associates, via its src homology 2 domains, with the activated erythropoietin receptor. Blood, 81(12):3204–3210, 1993.

[44] E.A. Sivertsen, M.E. Hystad, K.B. Gutzkow, G. Døsen, E.B. Smeland, H.K. Blomhoﬀ, and J.H. Myklebust. Pi3k/akt-dependent epo-induced signalling and target genes in human early erythroid progenitor cells. British journal of haematology, 135(1):117–128, 2006.

[45] TP McDonald, MB Cottrell, RE Clift, WC Cullen, FK Lin, et al. High doses of recombinant erythropoietin stimulate platelet production in mice. Experimental hematology, 15(6):719, 1987.

[46] T. Ishibashi, JA Koziol, and SA Burstein. Human recombinant erythropoietin promotes diﬀerentiation of murine megakaryocytes in vitro. Journal of Clinical Investigation, 79(1):286, 1987.

[47] M.V. Berridge, J.K. Fraser, J.M. Carter, and F.K. Lin. Eﬀects of recombinant human erythropoietin on megakaryocytes and on platelet production in the rat. Blood, 72(3):970– 977, 1988. References 142

[48] M.W. Kieran, A.C. Perkins, S.H. Orkin, and L.I. Zon. Thrombopoietin rescues in vitro erythroid colony formation from mouse embryos lacking the erythropoietin receptor. Pro- ceedings of the National Academy of Sciences, 93(17):9126, 1996.

[49] I. Vigon, J.P. Mornon, L. Cocault, M.T. Mitjavila, P. Tambourin, S. Gisselbrecht, and M. Souyri. Molecular cloning and characterization of mpl, the human homolog of the v-mpl oncogene: identiﬁcation of a member of the hematopoietic growth factor receptor superfamily. Proceedings of the National Academy of Sciences, 89(12):5640, 1992.

[50] SI Lok, K. Kaushansky, R.D. Holly, J.L. Kuijper, C.E. Lofton-Day, P.J. Oort, F.J. Grant, M.D. Heipel, S.K. Burkhead, J.M. Kramer, et al. Cloning and expression of murine thrombopoietin cdna and stimulation of platelet production in vivo. Nature, 369:565 – 568, 1994.

[51] M.D. Feese, T. Tamada, Y. Kato, Y. Maeda, M. Hirose, Y. Matsukura, H. Shigematsu, T. Muto, A. Matsumoto, H. Watarai, et al. Structure of the receptor-binding domain of human thrombopoietin determined by complexation with a neutralizing antibody frag- ment. Proceedings of the National Academy of Sciences of the United States of America, 101(7):1816, 2004.

[52] WS Alexander, D. Metcalf, and AR Dunn. Point mutations within a dimer interface homology domain of c-mpl induce constitutive receptor activity and tumorigenicity. The EMBO journal, 14(22):5569, 1995.

[53] A.L. Gurney, S.C. Wong, W.J. Henzel, and F.J. de Sauvage. Distinct regions of c-mpl cytoplasmic domain are coupled to the jak-stat signal transduction pathway and shc phosphorylation. Proceedings of the National Academy of Sciences, 92(12):5292, 1995.

[54] PC Hendrie, K. Miyazawa, YC Yang, CD Langefeld, HE Broxmeyer, et al. Mast cell growth factor (c-kit ligand) enhances cytokine stimulation of proliferation of the human factor-dependent cell line, m07e. Experimental hematology, 19(10):1031, 1991.

[55] PJ Tortolani, JA Johnston, CM Bacon, DW McVicar, A. Shimosaka, D. Linnekin, DL Longo, and JJ O’shea. Thrombopoietin induces tyrosine phosphorylation and activation of the janus kinase, jak2. Blood, 85(12):3444–3451, 1995.

[56] Y. Miyakawa, A. Oda, BJ Druker, H. Miyazaki, M. Handa, H. Ohashi, and Y. Ikeda. Thrombopoietin induces tyrosine phosphorylation of stat3 and stat5 in human blood platelets. Blood, 87(2):439–446, 1996.

[57] M. Yamada, N. Komatsu, K. Okada, T. Kato, H. Miyazaki, and Y. Miura. Thrombopoi- etin induces tyrosine phosphorylation and activation of mitogen-activated protein kinases in a human thrombopoietin-dependent cell line. Biochemical and biophysical research communications, 217(1):230–237, 1995.

[58] Y. Miyakawa, P. Rojnuckarin, T. Habib, and K. Kaushansky. Thrombopoietin induces phosphoinositol 3-kinase activation through shp2, gab, and insulin receptor substrate proteins in baf3 cells and primary murine megakaryocytes. Journal of Biological Chemistry, 276(4):2494–2502, 2001. References 143

[59] K. Kirito, T. Watanabe, K. Sawada, H. Endo, K. Ozawa, and N. Komatsu. Throm- bopoietin regulates bcl-xl gene expression through stat5 and phosphatidylinositol 3-kinase activation pathways. Journal of Biological Chemistry, 277(10):8329–8337, 2002.

[60] I. Matsumura, J. Ishikawa, K. Nakajima, K. Oritani, Y. Tomiyama, J. Miyagawa, T. Kato, H. Miyazaki, Y. Matsuzawa, and Y. Kanakura. Thrombopoietin-induced diﬀerentiation of a human megakaryoblastic leukemia cell line, cmk, involves transcriptional activation of p21 (waf1/cip1) by stat5. Molecular and cellular biology, 17(5):2933–2943, 1997.

[61] J.B. Konopka, S.M. Watanabe, and O.N. Witte. An alteration of the human c-abl protein in k562 leukemia cells unmasks associated tyrosine kinase activity. Cell, 37(3):1035–1042, 1984.

[62] P. Peeters, S.D. Raynaud, J. Cools, I. Wlodarska, J. Grosgeorge, P. Philip, F. Monpoux, L. Van Rompaey, M. Baens, H. Van den Berghe, et al. Fusion of tel, the ets-variant gene 6 (etv6), to the receptor-associated kinase jak2 as a result of t (9; 12) in a lymphoid and t (9; 15; 12) in a myeloid leukemia. Blood, 90(7):2535, 1997.

[63] V. Lacronique, A. Boureux, V. Della Valle, H. Poirel, C.T. Quang, M. Mauchauﬀ´e, C. Berthou, M. Lessard, R. Berger, J. Ghysdael, et al. A tel-jak2 fusion protein with constitutive kinase activity in human leukemia. Science, 278(5341):1309, 1997.

[64] H. Kiyoi, T. Naoe, S. Yokota, M. Nakao, S. Minami, K. Kuriyama, A. Takeshita, K. Saito, S. Hasegawa, S. Shimodaira, et al. Internal tandem duplication of ﬂt3 associated with leukocytosis in acute promyelocytic leukemia. leukemia study group of the ministry of health and welfare (kohseisho). Leukemia, 11(9):1447, 1997.

[65] H. Kiyoi, M. Towatari, S. Yokota, M. Hamaguchi, R. Ohno, H. Saito, and T. Naoe. Internal tandem duplication of the ﬂt3 gene is a novel modality of elongation mutation which causes constitutive activation of the product. Leukemia, 12(9):1333–1337, 1998.

[66] Y. Yamamoto, H. Kiyoi, Y. Nakano, R. Suzuki, Y. Kodera, S. Miyawaki, N. Asou, K. Kuriyama, F. Yagasaki, C. Shimazaki, et al. Activating mutation of d835 within the activation loop of ﬂt3 in human hematologic malignancies. Blood, 97(8):2434, 2001.

[67] R.L. Levine, A. Pardanani, A. Teﬀeri, and D.G. Gilliland. Role of jak2 in the pathogenesis and therapy of myeloproliferative disorders. Nature Reviews Cancer, 7(9):673–683, 2007.

[68] J. Cools, D.J. DeAngelo, J. Gotlib, E.H. Stover, R.D. Legare, J. Cortes, J. Kutok, J. Clark, I. Galinsky, J.D. Griﬃn, et al. A tyrosine kinase created by fusion of the pdgfra and ﬁp1l1 genes as a therapeutic target of imatinib in idiopathic hypereosinophilic syndrome. New England Journal of Medicine, 348(13):1201–1214, 2003.

[69] C. Graux, J. Cools, C. Melotte, H. Quentmeier, A. Ferrando, R. Levine, JR Vermeesch, M. Stul, B. Dutta, N. Boeckx, et al. Fusion of nup214 to abl1 on ampliﬁed episomes in t-cell acute lymphoblastic leukemia. Nature genetics, 36(10):1084–1089, 2004.

[70] S. Pakakasama, S. Kajanachumpol, S. Kanjanapongkul, N. Sirachainan, A. Meekaewkun- chorn, V. Ningsanond, and S. Hongeng. Simple multiplex rt-pcr for identifying common fusion transcripts in childhood acute leukemia. International Journal of Laboratory Hematology, 30(4):286–291, 2008. References 144

[71] S. Verstovsek, H. Lin, H. Kantarjian, G. Saglio, D. De Micheli, F. Pane, G. Garcia- Manero, M. Intrieri, B. Rotoli, F. Salvatore, et al. Neutrophilic-chronic myeloid leukemia. Cancer, 94(9):2416–2425, 2002.

[72] R. Kurzrock, M. Shtalrid, M. Talpaz, W.S. Kloetzer, and J.U. Gutterman. Expression of c-abl in philadelphia-positive acute myelogenous leukemia. Blood, 70(5):1584–1588, 1987.

[73] R. Virchow. Weisses blut. Frorieps Notizen, 36:151–156, 1845.

[74] J.H. Bennett. Case of Hypertrophy of the Spleen and Liver: In which Death Took Place from Suppuration of the Blood... Stark and Comp., 1845.

[75] D. Craigie. Case of Disease of the Spleen: In which Death Took Place in Consequence of the Presence of Purulent Matter in the Blood. J. Stark, 1845.

[76] S. Wong and O.N. Witte. The bcr-abl story: bench to bedside and back. Annu. Rev. Immunol., 22:247–306, 2004.

[77] E. Neumann. Uber myelogene leukamie. Berliner Klinische Wochenschrift, 15:69, 1878.

[78] P. Ehrlich. Farbenanalytische Untersuchungen zur Histologie und Klinik des Blutes: Gesammelte Mittheilungen, volume 1. A. Hirschwald, 1891.

[79] PC Nowell. A minute chromosome in human chronic granulocytic leukemia. Science, 132:1497, 1960.

[80] JD Rowly. A new consistent chromosomal abnormality in chronic myelogenous leukemia identiﬁed by quinacrine ﬂuorescence and giemsa staining. Nature, 243:290–293, 1973.

[81] J.D. Rowley. A new consistent chromosomal adnormality in chronic myelogeneus leukaemia identiﬁed by quinacrine ﬂuorescence and giemsa staining. Landmarks in medical genetics: Classic papers with commentaries, 243(51):104, 2004.

[82] A. De Klein, A.G. van Kessel, G. Grosveld, C.R. Bartram, A. Hagemeijer, D. Bootsma, N.K. Spurr, N. Heisterkamp, J. Groﬀen, and J.R. Stephenson. A cellular oncogene is translocated to the philadelphia chromosome in chronic myelocytic leukaemia. Nature, 300:765–767, 1982.

[83] AJ Muller, JC Young, AM Pendergast, M. Pondel, NR Landau, DR Littman, and ON Witte. Bcr ﬁrst exon sequences speciﬁcally activate the bcr/abl tyrosine kinase oncogene of philadelphia chromosome-positive human leukemias. Molecular and cellular biology, 11(4):1785–1792, 1991.

[84] C.L. Sawyers. Chronic myeloid leukemia. New England Journal of Medicine, 340(17):1330–1340, 1999.

[85] F. Pane, F. Frigeri, M. Sindona, L. Luciano, F. Ferrara, R. Cimino, G. Meloni, G. Saglio, F. Salvatore, and B. Rotoli. Neutrophilic-chronic myeloid leukemia: a distinct disease with a speciﬁc molecular marker (bcr/abl with c3/a2 junction)[see comments]. Blood, 88(7):2410–2414, 1996. References 145

[86] S.L. Chissoe, A. Bodenteich, Y.F. Wang, Y.P. Wang, D. Burian, S.W. Clifton, J. Crabtree, A. Freeman, K. Iyer, L. Jian, et al. Sequence and analysis of the human abl gene, the bcr gene, and regions involved in the philadelphia chromosomal translocation. Genomics, 27(1):67–82, 1995.

[87] S. Faderl, M. Talpaz, Z. Estrov, S. OBrien, R. Kurzrock, and H.M. Kantarjian. The biology of chronic myeloid leukemia. New England Journal of Medicine, 341(3):164–172, 1999.

[88] E. Laurent, M. Talpaz, H. Kantarjian, and R. Kurzrock. The bcr gene and philadelphia chromosome-positive leukemogenesis. Cancer research, 61(6):2343, 2001.

[89] A.T. Chinwalla, L.L. Cook, K.D. Delehaunty, G.A. Fewell, L.A. Fulton, R.S. Fulton, T.A. Graves, L.D.W. Hillier, E.R. Mardis, J.D. McPherson, et al. Initial sequencing and comparative analysis of the mouse genome. Nature, 420(6915):520–562, 2002.

[90] J.R. McWhirter, D.L. Galasso, and JY Wang. A coiled-coil oligomerization domain of bcr is essential for the transforming function of bcr-abl oncoproteins. Molecular and cellular biology, 13(12):7587–7595, 1993.

[91] X. Zhao, S. Ghaﬀari, H. Lodish, V.N. Malashkevich, and P.S. Kim. Structure of the bcr- abl oncoprotein oligomerization domain. Nature Structural & Molecular Biology, 9(2):117– 120, 2002.

[92] Y. Maru and O.N. Witte. The bcr gene encodes a novel serine/threonine kinase activity within a single exon. Cell, 67(3):459–468, 1991.

[93] G.W. Reuther, H. Fu, L.D. Cripe, R.J. Collier, and A.M. Pendergast. Association of the protein kinases c-bcr and bcr-abl with proteins of the 14-3-3 family. Science, 266(5182):129–133, 1994.

[94] S. Braselmann and F. McCormick. Bcr and raf form a complex in vivo via 14-3-3 proteins. The EMBO journal, 14(19):4839, 1995.

[95] S.J. Clokie, K.Y. Cheung, S. Mackie, R. Marquez, A.H. Peden, and A. Aitken. Bcr kinase phosphorylates 14-3-3 tau on residue 233. Febs Journal, 272(15):3767–3776, 2005.

[96] D. Ron, M. Zannini, M. Lewis, RB Wickner, LT Hunt, G. Graziani, SR Tronick, SA Aaronson, A. Eva, et al. A region of proto-dbl essential for its transforming activity shows sequence similarity to a yeast cell cycle gene, cdc24, and the human breakpoint cluster gene, bcr. The New biologist, 3(4):372, 1991.

[97] D. Diekmann, S. Brill, M.D. Garrett, N. Totty, J. Hsuan, C. Monfries, C. Hall, L. Lim, and A. Hall. Bcr encodes a gtpase-activating protein for p21rac. Nature, 351(6325):400–402, 1991.

[98] K.J. Simpson, L.M. Selfors, J. Bui, A. Reynolds, D. Leake, A. Khvorova, and J.S. Brugge. Identiﬁcation of genes that regulate epithelial cell migration using an sirna screening approach. Nature cell biology, 10(9):1027–1038, 2008.

[99] J.W. Voncken, H. Schaick, V. Kaartinen, K. Deemer, T. Coates, B. Landing, P. Patten- gale, O. Dorseuil, G.M. Bokoch, J. Groﬀen, et al. Increased neutrophil respiratory burst in bcr-null mutants. Cell, 80(5):719–728, 1995. References 146

[100] S.K. Shore, R.V. Tantravahi, E.P. Reddy, et al. Transforming pathways activated by the v-abl tyrosine kinase. Oncogene, 21(56):8568, 2002.

[101] V.L. Tybulewicz, C.E. Crawford, P.K. Jackson, R.T. Bronson, R.C. Mulligan, et al. Neonatal lethality and lymphopenia in mice with a homozygous disruption of the c-abl proto-oncogene. Cell, 65(7):1153, 1991.

[102] Y.Q. Gosser, J. Zheng, M. Overduin, B.J. Mayer, and D. Cowburn. The solution structure of abl sh3, and its relationship to sh2 in the sh (32) construct. Structure, 3(10):1075–1086, 1995.

[103] B. Nagar, O. Hantschel, M.A. Young, K. Scheﬀzek, D. Veach, W. Bornmann, B. Clarkson, G. Superti-Furga, and J. Kuriyan. Structural basis for the autoinhibition of c-abl tyrosine kinase. Cell, 112(6):859–871, 2003.

[104] O. Hantschel and G. Superti-Furga. Regulation of the c-abl and bcr–abl tyrosine kinases. Nature Reviews Molecular Cell Biology, 5(1):33–44, 2004.

[105] A.M. Pendergast, L.A. Quilliam, L.D. Cripe, C.H. Bassing, Z. Dai, N. Li, A. Batzer, K.M. Rabun, C.J. Der, J. Schlessinger, et al. Bcr-abl-induced oncogenesis is mediated by direct interaction with the sh2 domain of the grb-2 adaptor protein. Cell, 75(1):175–185, 1993.

[106] S. Maignan, J.P. Guilloteau, N. Fromage, B. Arnoux, J. Becquart, and A. Ducruix. Crystal structure of the mammalian grb2 adaptor. Science, 268(5208):291–293, 1995.

[107] L. Puil, J. Liu, G. Gish, G. Mbamalu, D. Bowtell, PG Pelicci, R. Arlinghaus, and T. Paw- son. Bcr-abl oncoproteins bind directly to activators of the ras signalling pathway. The EMBO journal, 13(4):764, 1994.

[108] T. Tauchi, H.S. Boswell, D. Leibowitz, and H.E. Broxmeyer. Coupling between p210bcr- abl and shc and grb2 adaptor proteins in hematopoietic cells permits growth factor receptor-independent link to ras activation pathway. The Journal of experimental medicine, 179(1):167–175, 1994.

[109] C.L. Sawyers, J. McLaughlin, and O.N. Witte. Genetic requirement for ras in the transformation of ﬁbroblasts and hematopoietic cells by the bcr-abl oncogene. The Journal of experimental medicine, 181(1):307–313, 1995.

[110] J. Downward et al. Targeting ras signalling pathways in cancer therapy. Nature Reviews Cancer, 3(1):11–22, 2003.

[111] M. Sattler, M.G. Mohi, Y.B. Pride, L.R. Quinnan, N.A. Malouf, K. Podar, F. Gesbert, H. Iwasaki, S. Li, R.A. Van Etten, et al. Critical role for gab2 in transformation by bcr/abl. Cancer cell, 1(5):479–492, 2002.

[112] T. Skorski, A. Bellacosa, M. Nieborowska-Skorska, M. Majewski, R. Martinez, J.K. Choi, R. Trotta, P. Wlodarski, D. Perrotti, T.O. Chan, et al. Transformation of hematopoietic cells by bcr/abl requires activation of a pi-3k/akt-dependent pathway. The EMBO Journal, 16(20):6151–6161, 1997. References 147

[113] P. Crespo, K.E. Schuebel, A.A. Ostrom, J.S. Gutkind, and X.R. Bustelo. Phosphotyrosine-dependent activation of rac-1 gdp/gtp exchange by the vav proto- oncogene product. Nature, 385:169–172, 1997.

[114] F. Bassermann, T. Jahn, C. Miething, P. Seipel, R.Y. Bai, S. Coutinho, V.L. Tybulewicz, C. Peschel, and J. Duyster. Association of bcr-abl with the proto-oncogene vav is implicated in activation of the rac-1 pathway. Journal of Biological Chemistry, 277(14):12437– 12445, 2002.

[115] T. Daubon, J. Chasseriau, A. El Ali, J. Rivet, A. Kitzis, B. Constantin, and N. Bourmeyster. Diﬀerential motility of p190bcr-abl-and p210bcr-abl-expressing cells: respective roles of vav and bcr-abl gefs. Oncogene, 27(19):2673–2685, 2007.

[116] M.W.N. Deininger, J.M. Goldman, and J.V. Melo. The molecular biology of chronic myeloid leukemia. Blood, 96(10):3343–3356, 2000.

[117] R. Ren. Mechanisms of bcr–abl in the pathogenesis of chronic myelogenous leukaemia. Nature Reviews Cancer, 5(3):172–183, 2005.

[118] T. O’Hare, M.S. Zabriskie, A.M. Eiring, and M.W. Deininger. Pushing the limits of targeted therapy in chronic myeloid leukaemia. Nature Reviews Cancer, 12(8):513–526, 2012.

[119] G.Q. Daley, R.A. Van Etten, and D. Baltimore. Induction of chronic myelogenous leukemia in mice by the p210bcr/abl gene of the philadelphia chromosome. Science, 247(4944):824–830, 1990.

[120] W.S. Pear, J.P. Miller, L. Xu, J.C. Pui, B. Soﬀer, R.C. Quackenbush, A.M. Pendergast, R. Bronson, J.C. Aster, M.L. Scott, et al. Eﬃcient and rapid induction of a chronic myelogenous leukemia-like myeloproliferative disease in mice receiving p210 bcr/abl- transduced bone marrow. Blood, 92(10):3780–3792, 1998.

[121] S. Li, R.L. Ilaria, R.P. Million, G.Q. Daley, and R.A. Van Etten. The p190, p210, and p230 forms of the bcr/abl oncogene induce a similar chronic myeloid leukemia–like syndrome in mice but have diﬀerent lymphoid leukemogenic activity. The Journal of experimental medicine, 189(9):1399, 1999.

[122] R.P. Million and R.A. Van Etten. The grb2 binding site is required for the induction of chronic myeloid leukemia-like disease in mice by the bcr/abl tyrosine kinase. Blood, 96(2):664–670, 2000.

[123] Y. He, J.A. Wertheim, L. Xu, J.P. Miller, F.G. Karnell, J.K. Choi, R. Ren, and W.S. Pear. The coiled-coil domain and tyr177 of bcr are required to induce a murine chronic myelogenous leukemia–like disease by bcr/abl. Blood, 99(8):2957–2968, 2002.

[124] K.M. Smith, R. Yacobi, and R.A. Van Etten. Autoinhibition of bcr-abl through its sh3 domain. Molecular cell, 12(1):27–37, 2003.

[125] Y. Maru, O.N. Witte, and M. Shibuya. Deletion of the abl sh3 domain reactivates de- oligomerized bcr-abl for growth factor independence. FEBS letters, 379(3):244–246, 1996. References 148

[126] PB Sinclair, EP Nacheva, M. Leversha, N. Telford, J. Chang, A. Reid, A. Bench, K. Champion, B. Huntly, and AR Green. Large deletions at the t (9; 22) breakpoint are common and may identify a poor-prognosis subgroup of patients with chronic myeloid leukemia. Blood, 95(3):738, 2000.

[127] J. De La Fuente, K. Merx, E.J. Steer, M. Muller, R.M. Szydlo, O. Maywald, U. Berger, R. Hehlmann, J.M. Goldman, N.C.P. Cross, et al. Abl-bcr expression does not correlate with deletions on the derivative chromosome 9 or survival in chronic myeloid leukemia. Blood, 98(9):2879, 2001.

[128] B.J.P. Huntly, A.J. Bench, E. Delabesse, A.G. Reid, J. Li, M.A. Scott, L. Campbell, J. Byrne, E. Pinto, A. Brizard, et al. Derivative chromosome 9 deletions in chronic myeloid leukemia: poor prognosis is not associated with loss of abl-bcr expression, elevated bcr-abl levels, or karyotypic instability. Blood, 99(12):4547, 2002.

[129] M. Fournier, S. Lacrosse, M. Jamar, V. Bours, and C. Herens. Deletions of the 3’bcr and 5’abl regions in patients with philadelphia-positive chronic myeloid leukemia: a one-step process occurring in about 10% of the cases without any evidence of genetic instability in the target cells. Cancer Genetics and Cytogenetics, 160(2):184–187, 2005.

[130] X. Zheng, C. Oancea, R. Henschler, M.A.S. Moore, and M. Ruthardt. Reciprocal t (9; 22) abl/bcr fusion proteins: leukemogenic potential and eﬀects on b cell commitment. PloS one, 4(10):e7661, 2009.

[131] B.J.P. Huntly, A. Bench, and A.R. Green. Double jeopardy from a single translocation: deletions of the derivative chromosome 9 in chronic myeloid leukemia. Blood, 102(4):1160, 2003.

[132] E. Kolomietz, P. Marrano, K. Yee, B. Thai, I. Braude, A. Kolomietz, K. Chun, S. Minkin, S. Kamel-Reid, M. Minden, et al. Quantitative pcr identiﬁes a minimal deleted region of 120 kb extending from the philadelphia chromosome abl translocation breakpoint in chronic myeloid leukemia with poor outcome. Leukemia, 17(7):1313–1323, 2003.

[133] AG Reid and EP Nacheva. A potential role for prdm12 in the pathogenesis of chronic myeloid leukaemia with derivative chromosome 9 deletion. Leukemia, 18(1):178–180, 2003.

[134] P. Mitchell, E. Petfalski, A. Shevchenko, M. Mann, and D. Tollervey. The exosome: a conserved eukaryotic rna processing complex containing multiple 3’ - 5’ exoribonucleases. Cell, 91(4):457–466, 1997.

[135] E. Kolomietz, J. Al-Maghrabi, S. Brennan, J. Karaskova, S. Minkin, J. Lipton, and J.A. Squire. Primary chromosomal rearrangements of leukemia are frequently accompanied by extensive submicroscopic deletions and may lead to altered prognosis. Blood, 97(11):3581, 2001.

[136] B.J.P. Huntly, A.G. Reid, A.J. Bench, L.J. Campbell, N. Telford, P. Shepherd, J. Szer, H.M. Prince, P. Turner, C. Grace, et al. Deletions of the derivative chromosome 9 occur at the time of the philadelphia translocation and provide a powerful and independent prognostic indicator in chronic myeloid leukemia. Blood, 98(6):1732, 2001. References 149

[137] A.G. Reid, B.J.P. Huntly, C. Grace, A.R. Green, and E.P. Nacheva. Survival implications of molecular heterogeneity in variant philadelphia-positive chronic myeloid leukaemia. British journal of Haematology, 121(3):419–427, 2003.

[138] B.J.P. Huntly, F. Guilhot, A.G. Reid, G. Vassiliou, E. Hennig, C. Franke, J. Byrne, A. Brizard, D. Niederwieser, J. Freeman-Edward, et al. Imatinib improves but may not fully reverse the poor prognosis of patients with cml with derivative chromosome 9 deletions. Blood, 102(6):2205, 2003.

[139] A. Quintas-Cardama, H. Kantarjian, M. Talpaz, S. O’Brien, G. Garcia-Manero, S. Ver- stovsek, M.B. Rios, K. Hayes, A. Glassman, B.N. Bekele, et al. Imatinib mesylate therapy may overcome the poor prognostic signiﬁcance of deletions of derivative chromosome 9 in patients with chronic myelogenous leukemia. Blood, 105(6):2281, 2005.

[140] G. Calabrese, D. Fantasia, R. Di Gianﬁlippo, L. Stuppia, R. Di Lorenzo, and C. PALKA. Fluorescence in situ hybridization analysis of minimal residual disease and the relevance of the der (9) deletion in imatinib-treated patients with chronic myeloid leukemia. Haema- tologica, 91(7):994–995, 2006.

[141] F. Ghaith, S. Abdou, A. El-Bendary, D. Shahin, M. Eid, WA Megeed, I. EL-SHEIKH, W. Farrag, and S. Yousuf. Prognostic relevance of 9q34 deletion and the suppressor of cytokine signalling-1 in cml patients. International Journal of Laboratory Hematology, 32(1p2):103–112, 2010.

[142] J. Huh, C.W. Jung, J.W. Kim, H.J. Kim, S.H. Kim, M.G. Shin, Y.K. Kim, H. Joon Kim, J.S. Suh, J.H. Moon, et al. Genome-wide high density single-nucleotide polymorphism array-based karyotyping improves detection of clonal aberrations including der (9) deletion, but does not predict treatment outcomes after imatinib therapy in chronic myeloid leukemia. Annals of Hematology, 90(11):1255–64, 2011.

[143] C.J. Wilusz and J. Wilusz. Bringing the role of mrna decay in the control of gene expression into focus. TRENDS in Genetics, 20(10):491–497, 2004.

[144] M.J. Moore. From birth to death: the complex lives of eukaryotic mrnas. Science’s STKE, 309(5740):1514, 2005.

[145] F.W. Larimer, C.L. Hsu, M.K. Maupin, and A. Stevens. Characterization of the xrn1 gene encoding a 5’ - 3’ exoribonuclease: sequence data and analysis of disparate protein and mrna levels of gene-disrupted yeast cells. Gene, 120(1):51–57, 1992.

[146] J. Houseley and D. Tollervey. The many pathways of rna degradation. Cell, 136(4):763– 776, 2009.

[147] Q. Liu, J.C. Greimann, and C.D. Lima. Reconstitution, activities, and structure of the eukaryotic rna exosome. Cell, 127(6):1223–1237, 2006.

[148] J.R. Anderson, D. Mukherjee, K. Muthukumaraswamy, K.C.M. Moraes, C.J. Wilusz, and J. Wilusz. Sequence-speciﬁc rna binding mediated by the rnase ph domain of components of the exosome. Rna, 12(10):1810–1816, 2006.

[149] N.V. Grishin. Kh domain: one motif, two folds. Nucleic acids research, 29(3):638–643, 2001. References 150

[150] M.F. Garc´ıa-Mayoral, D. Hollingworth, L. Masino, I. D´ıaz-Moreno,G. Kelly, R. Gherzi, C.F. Chou, C.Y. Chen, and A. Ramos. The structure of the c-terminal kh domains of ksrp reveals a noncanonical motif important for mrna degradation. Structure, 15(4):485–498, 2007.

[151] T. Suryanarayana and A.R. Subramanian. Function of the repeating homologous sequences in nucleic acid binding domain of ribosomal protein s1. Biochemistry, 23(6):1047– 1051, 1984.

[152] M. Bycroft, TJ Hubbard, M. Proctor, SM Freund, A.G. Murzin, et al. The solution structure of the s1 rna binding domain: a member of an ancient nucleic acid-binding fold. Cell, 88(2):235, 1997.

[153] T. Kadowaki, R. Schneiter, M. Hitomi, and AM Tartakoﬀ. Mutations in nucleolar proteins lead to nucleolar accumulation of polya+ rna in saccharomyces cerevisiae. Molecular biology of the cell, 6(9):1103, 1995.

[154] C. Allmang, E. Petfalski, A. Podtelejnikov, M. Mann, D. Tollervey, and P. Mitchell. The yeast exosome and human pm–scl are related complexes of 3’ - 5’ exonucleases. Genes & development, 13(16):2148–2158, 1999.

[155] R. Brouwer, C. Allmang, R. Raijmakers, Y. Van Aarssen, W.V. Egberts, E. Petfalski, W.J. Van Venrooij, D. Tollervey, and G.J.M. Pruijn. Three novel components of the human exosome. Journal of Biological Chemistry, 276(9):6177–6184, 2001.

[156] J. Wan, M. Yourshaw, H. Mamsa, S. Rudnik-Sch¨oneborn, M.P. Menezes, J.E. Hong, D.W. Leong, J. Senderek, M.S. Salman, D. Chitayat, et al. Mutations in the rna exosome component gene exosc3 cause pontocerebellar hypoplasia and spinal motor neuron degeneration. Nature genetics, 44(6):704–708, 2012.

[157] A. Dziembowski, E. Lorentzen, E. Conti, and B. S´eraphin. A single subunit, dis3, is essentially responsible for yeast exosome core activity. Nature structural & molecular biology, 14(1):15–22, 2006.

[158] E. Lorentzen, J. Basquin, R. Tomecki, A. Dziembowski, and E. Conti. Structure of the active subunit of the yeast exosome core, rrp44: diverse modes of substrate recruitment in the rnase ii nuclease family. Molecular cell, 29(6):717–728, 2008.

[159] C. Schneider, J.T. Anderson, and D. Tollervey. The exosome subunit rrp44 plays a direct role in rna substrate recognition. Molecular cell, 27(2):324–331, 2007.

[160] P. Mitchell, E. Petfalski, and D. Tollervey. The 3’end of yeast 5.8 s rrna is generated by an exonuclease processing mechanism. Genes & development, 10(4):502–513, 1996.

[161] C. Allmang, J. Kufel, G. Chanfreau, P. Mitchell, E. Petfalski, and D. Tollervey. Functions of the exosome in rrna, snorna and snrna synthesis. The EMBO journal, 18(19):5399– 5410, 1999.

[162] C. Allmang, P. Mitchell, E. Petfalski, and D. Tollervey. Degradation of ribosomal rna precursors by the exosome. Nucleic acids research, 28(8):1684–1691, 2000.

[163] J. Houseley, J. LaCava, and D. Tollervey. Rna-quality control by the exosome. Nature Reviews Molecular Cell Biology, 7(7):529–539, 2006. References 151

[164] R.F. Doolittle, D.F. Feng, S. Tsang, G. Cho, and E. Little. Determining divergence times of the major kingdoms of living organisms with a protein clock. Science, 271(5248):470– 477, 1996.

[165] AJ Carpousis et al. The escherichia coli rna degradosome: structure, function and relationship to other ribonucleolytic multienyzme complexes. Biochemical Society Transac- tions, 30(2):1, 2002.

[166] A. Van Hoof, R.R. Staples, R.E. Baker, and R. Parker. Function of the ski4p (csl4p) and ski7p proteins in 3?-to-5? degradation of mrna. Molecular and cellular biology, 20(21):8230–8243, 2000.

[167] M.W. Briggs, K.T.D. Burkard, and J.S. Butler. Rrp6p, the yeast homologue of the human pm-scl 100-kda autoantigen, is essential for eﬃcient 5.8 s rrna 3? end formation. Journal of Biological Chemistry, 273(21):13255–13263, 1998.

[168] L. Kuai, F. Fang, J.S. Butler, and F. Sherman. Polyadenylation of rrna in saccharomyces cerevisiae. Proceedings of the National Academy of Sciences of the United States of Amer- ica, 101(23):8581, 2004.

[169] N. Kinoshita, M. Goebl, and M. Yanagida. The ﬁssion yeast dis3+ gene encodes a 110-kda essential protein implicated in mitotic control. Molecular and cellular biology, 11(12):5839–5847, 1991.

[170] E.V. Koonin, Y.I. Wolf, and L. Aravind. Prediction of the archaeal exosome and its connections with the proteasome and the translation and transcription machineries by a comparative-genomic approach. Genome research, 11(2):240–252, 2001.

[171] G. Schilders, E. van Dijk, R. Raijmakers, and G.J.M. Pruijn. Cell and molecular biology of the exosome: how to make or break an rna. International review of cytology, 251:159–208, 2006.

[172] D. Caput, B. Beutler, K. Hartog, R. Thayer, S. Brown-Shimer, and A. Cerami. Identiﬁ- cation of a common nucleotide sequence in the 3’-untranslated region of mrna molecules specifying inﬂammatory mediators. Proceedings of the National Academy of Sciences, 83(6):1670, 1986.

[173] G. Shaw and R. Kamen. A conserved au sequence from the 3’untranslated region of gm-csf mrna mediates selective mrna degradation. Cell, 46(5):659, 1986.

[174] M. Akashi, G. Shaw, M. Hachiya, E. Elstner, G. Suzuki, and P. Koeﬄer. Number and location of auuua motifs: role in regulating transiently expressed rnas. Blood, 83(11):3182– 3187, 1994.

[175] T. Bakheet, M. Frevel, B.R.G. Williams, W. Greer, and K.S.A. Khabar. Ared: human au-rich element-containing mrna database reveals an unexpectedly diverse functional repertoire of encoded proteins. Nucleic acids research, 29(1):246–254, 2001.

[176] T. Bakheet, B.R.G. Williams, and K.S.A. Khabar. Ared 2.0: an update of au-rich element mrna database. Nucleic acids research, 31(1):421–423, 2003.

[177] T. Bakheet, B.R.G. Williams, and K.S.A. Khabar. Ared 3.0: the large and diverse au-rich transcriptome. Nucleic acids research, 34(suppl 1):D111–D114, 2006. References 152

[178] A. Bevilacqua, M.C. Ceriani, S. Capaccioli, and A. Nicolin. Post-transcriptional regulation of gene expression by degradation of messenger rnas. Journal of cellular physiology, 195(3):356–372, 2003.

[179] C.Y. Chen, R. Gherzi, S.E. Ong, E.L. Chan, R. Raijmakers, G.J.M. Pruijn, G. Stoecklin, C. Moroni, M. Mann, and M. Karin. Au binding proteins recruit the exosome to degrade are-containing mrnas. Cell, 107(4):451–464, 2001.

[180] D. Mukherjee, M. Gao, J.P. O’Connor, R. Raijmakers, G. Pruijn, C.S. Lutz, and J. Wilusz. The mammalian exosome mediates the eﬃcient degradation of mrnas that contain au-rich elements. The EMBO journal, 21(1):165–174, 2002.

[181] I. Schneider. Cell lines derived from late embryonic stages of drosophila melanogaster. Journal of embryology and experimental morphology, 27(2):353–365, 1972.

[182] D.L. Kiss and E.D. Andrulis. Genome-wide analysis reveals distinct substrate speciﬁcities of rrp6, dis3, and core exosome subunits. RNA, 16(4):781–791, 2010.

[183] T. Hohenauer and A.W. Moore. The prdm family: expanding roles in stem cells and development. Development, 139(13):2267–2282, 2012.

[184] C.K. Fog, G.G. Galli, and A.H. Lund. Prdm proteins: Important players in diﬀerentiation and disease. BioEssays, 2012.

[185] S. Huang, G. Shao, and L. Liu. The pr domain of the rb-binding zinc ﬁnger protein riz1 is a protein binding interface and is related to the set domain functioning in chromatin- mediated gene expression. Journal of Biological Chemistry, 273(26):15933–15939, 1998.

[186] E. Kinameri, T. Inoue, J. Aruga, I. Imayoshi, R. Kageyama, T. Shimogori, and A.W. Moore. Prdm proto-oncogene transcription factor family expression and interaction with the notch-hes pathway in mouse neurogenesis. PLoS One, 3(12):e3859, 2008.

[187] C.A. Turner, D.H. Mack, M.M. Davis, et al. Blimp-1, a novel zinc ﬁnger-containing protein that can drive the maturation of b lymphocytes into immunoglobulin-secreting cells. Cell, 77(2):297–306, 1994.

[188] I.M. Buyse, G. Shao, and S. Huang. The retinoblastoma protein binds to riz, a zinc- ﬁnger protein that shares an epitope with the adenovirus e1a protein. Proceedings of the National Academy of Sciences, 92(10):4467, 1995.

[189] M. Trubia, F. Albano, F. Cavazzini, GR Cambrin, G. Quarta, F. Fabbiano, F. Ciambelli, D. Magro, JM Hernandez, M. Mancini, et al. Characterization of a recurrent translocation t (2; 3)(p15–22; q26) occurring in acute myeloid leukaemia. Leukemia, 20(1):48–54, 2005.

[190] N. Mochizuki, S. Shimizu, T. Nagasawa, H. Tanaka, M. Taniwaki, J. Yokota, and K. Mor- ishita. A novel gene, mel1, mapped to 1p36. 3 is highly homologous to the mds1/evi1 gene and is transcriptionally activated in t (1; 3)(p36; q21)-positive leukemia cells. Blood, 96(9):3209–3214, 2000.

[191] H. Lissauer. Zwei f¨allevon leucaemie. Berl. Klin. Wochenschr, 40:403, 1865.

[192] CG Geary. The story of chronic myeloid leukaemia. British journal of haematology, 110(1):2–11, 2000. References 153

[193] N. Senn. Case of splenomedullary leukaemia successfully treated by the use of roentgen ray. Med Record, 63:281, 1903.

[194] W.J. HOFFMAN and L.F. CRAVER. Chronic myelogenous leukemia. Journal of the American Medical Association, 97(12):836–840, 1931.

[195] DA Galton. Myleran in chronic myeloid leukaemia; results of treatment. Lancet, 264(6753):208, 1953.

[196] B. Stearns, K.A. Losee, and J. Bernstein. Hydroxyurea. a new type of potential antitumor agent1. Journal of medicinal chemistry, 6(2):201–201, 1963.

[197] BJ Kennedy and J.W. Yarbro. Metabolic and therapeutic eﬀects of hydroxyurea in chronic myeloid leukemia. JAMA: the journal of the American Medical Association, 195(12):1038, 1966.

[198] M. Talpaz, KB McCredie, GM Mavligit, and JU Gutterman. Leukocyte interferon- induced myeloid cytoreduction in chronic myelogenous leukemia. Blood, 62(3):689–692, 1983.

[199] B.J. Druker. Translation of the philadelphia chromosome into therapy for cml. Blood, 112(13):4808–4817, 2008.

[200] E. Buchdunger, J. Zimmermann, H. Mett, T. Meyer, M. M¨uller,U. Regenass, and N.B. Lydon. Selective inhibition of the platelet-derived growth factor signal transduction pathway by a protein-tyrosine kinase inhibitor of the 2-phenylaminopyrimidine class. Proceed- ings of the National Academy of Sciences, 92(7):2558, 1995.

[201] B.J. Druker, N.B. Lydon, et al. Lessons learned from the development of an ab1 tyrosine kinase inhibitor for chronic myelogenous leukemia. Journal of Clinical Investigation, 105(1):3–8, 2000.

[202] M. Deininger, E. Buchdunger, and B.J. Druker. The development of imatinib as a therapeutic agent for chronic myeloid leukemia. Blood, 105(7):2640–2653, 2005.

[203] B.J. Druker, S. Tamura, E. Buchdunger, S. Ohno, G.M. Segal, S. Fanning, J. Zimmer- mann, and N.B. Lydon. Eﬀects of a selective inhibitor of the abl tyrosine kinase on the growth of bcr-abl positive cells. Nature Medicine, 2(5):561–566, 1996.

[204] B.J. Druker, M. Talpaz, D.J. Resta, B. Peng, E. Buchdunger, J.M. Ford, N.B. Lydon, H. Kantarjian, R. Capdeville, S. Ohno-Jones, et al. Eﬃcacy and safety of a speciﬁc inhibitor of the bcr-abl tyrosine kinase in chronic myeloid leukemia. New England Journal of Medicine, 344(14):1031–1037, 2001.

[205] M E Gorre, M Mohammed, K Ellwood, N Hsu, R Paquette, P N Rao, and C L Sawyers. Clinical resistance to sti-571 cancer therapy caused by bcr-abl gene mutation or ampliﬁ- cation. Science, 293(5531):876–80, Aug 2001.

[206] S. Branford, Z. Rudzki, S. Walsh, A. Grigg, C. Arthur, K. Taylor, R. Herrmann, K.P. Lynch, and T.P. Hughes. High frequency of point mutations clustered within the adenosine triphosphate–binding region of bcr/abl in patients with chronic myeloid leukemia or ph-positive acute lymphoblastic leukemia who develop imatinib (sti571) resistance. Blood, 99(9):3472–3476, 2002. References 154

[207] T. O’Hare, D.K. Walters, E.P. Stoﬀregen, T. Jia, P.W. Manley, J. Mestan, S.W. Cowan- Jacob, F.Y. Lee, M.C. Heinrich, M.W.N. Deininger, et al. In vitro activity of bcr-abl inhibitors amn107 and bms-354825 against clinically relevant imatinib-resistant abl kinase domain mutants. Cancer Research, 65(11):4500–4505, 2005.

[208] M. Azam, R.R. Latek, and G.Q. Daley. Mechanisms of autoinhibition and sti- 571/imatinib resistance revealed by mutagenesis of bcr-abl. Cell, 112(6):831–843, 2003.

[209] H.A. Bradeen, C.A. Eide, T. O’Hare, K.J. Johnson, S.G. Willis, F.Y. Lee, B.J. Druker, and M.W. Deininger. Comparison of imatinib mesylate, dasatinib (bms-354825), and nilotinib (amn107) in an n-ethyl-n-nitrosourea (enu)-based mutagenesis screen: high eﬃcacy of drug combinations. Blood, 108(7):2332, 2006.

[210] A. Ray, S.W. Cowan-Jacob, P.W. Manley, J. Mestan, and J.D. Griﬃn. Identiﬁcation of bcr-abl point mutations conferring resistance to the abl kinase inhibitor amn107 (nilotinib) by a random mutagenesis study. Blood, 109(11):5011–5015, 2007.

[211] E. Weisberg, P.W. Manley, W. Breitenstein, J. Br¨uggen,S.W. Cowan-Jacob, A. Ray, B. Huntly, D. Fabbro, G. Fendrich, E. Hall-Meyers, et al. Characterization of amn107, a selective inhibitor of native and mutant bcr-abl. Cancer cell, 7(2):129–141, 2005.

[212] H.M. Kantarjian, F. Giles, N. Gattermann, K. Bhalla, G. Alimena, F. Palandri, G.J. Ossenkoppele, F.E. Nicolini, S.G. O’Brien, M. Litzow, et al. Nilotinib (formerly amn107), a highly selective bcr-abl tyrosine kinase inhibitor, is eﬀective in patients with philadelphia chromosome–positive chronic myelogenous leukemia in chronic phase following imatinib resistance and intolerance. Blood, 110(10):3540–3546, 2007.

[213] T. Schindler, W. Bornmann, P. Pellicena, W.T. Miller, B. Clarkson, and J. Kuriyan. Structural mechanism for sti-571 inhibition of abelson tyrosine kinase. Science’s STKE, 289(5486):1938, 2000.

[214] N.P. Shah, J.M. Nicoll, B. Nagar, M.E. Gorre, R.L. Paquette, J. Kuriyan, and C.L. Sawyers. Multiple bcr-abl kinase domain mutations confer polyclonal resistance to the tyrosine kinase inhibitor imatinib (sti571) in chronic phase and blast crisis chronic myeloid leukemia. Cancer Cell, 2(2):117–125, 2002.

[215] J.F. Dorsey, R. Jove, A.J. Kraker, and J. Wu. The pyrido [2, 3-d] pyrimidine derivative pd180970 inhibits p210bcr-abl tyrosine kinase and induces apoptosis of k562 leukemic cells. Cancer research, 60(12):3127, 2000.

[216] B. Nagar, W.G. Bornmann, P. Pellicena, T. Schindler, D.R. Veach, W.T. Miller, B. Clark- son, and J. Kuriyan. Crystal structures of the kinase domain of c-abl in complex with the small molecule inhibitors pd173955 and imatinib (sti-571). Cancer research, 62(15):4236, 2002.

[217] P. La Ros´ee,A.S. Corbin, E.P. Stoﬀregen, M.W. Deininger, and B.J. Druker. Activity of the bcr-abl kinase inhibitor pd180970 against clinically relevant bcr-abl isoforms that cause resistance to imatinib mesylate (gleevec, sti571). Cancer research, 62(24):7149, 2002. References 155

[218] J. Cortes, P. Rousselot, D.W. Kim, E. Ritchie, N. Hamerschlak, S. Coutre, A. Hochhaus, F. Guilhot, G. Saglio, J. Apperley, et al. Dasatinib induces complete hematologic and cytogenetic responses in patients with imatinib-resistant or-intolerant chronic myeloid leukemia in blast crisis. Blood, 109(8):3207, 2007.

[219] F. Guilhot, J. Apperley, D.W. Kim, E.O. Bullorsky, M. Baccarani, G.J. Roboz, S. Amadori, C.A. de Souza, J.H. Lipton, A. Hochhaus, et al. Dasatinib induces signiﬁcant hematologic and cytogenetic responses in patients with imatinib-resistant or-intolerant chronic myeloid leukemia in accelerated phase. Blood, 109(10):4143, 2007.

[220] T. O’Hare, C.A. Eide, J.W. Tyner, A.S. Corbin, M.J. Wong, S. Buchanan, K. Holme, K.A. Jessen, C. Tang, H.A. Lewis, et al. Sgx393 inhibits the cml mutant bcr-ablt315i and preempts in vitro resistance when combined with nilotinib or dasatinib. Proceedings of the National Academy of Sciences, 105(14):5507, 2008.

[221] W.W. Chan, S.C. Wise, M.D. Kaufman, Y.M. Ahn, C.L. Ensinger, T. Haack, M.M. Hood, J. Jones, J.W. Lord, W.P. Lu, et al. Conformational control inhibition of the bcr-abl1 tyrosine kinase, including the gatekeeper t315i mutant, by the switch-control inhibitor dcc-2036. Cancer cell, 19(4):556–568, 2011.

[222] J. Zhang, F.J. Adri´an,W. Jahnke, S.W. Cowan-Jacob, A.G. Li, R.E. Iacob, T. Sim, J. Powers, C. Dierks, F. Sun, et al. Targeting bcr–abl by combining allosteric with atp- binding-site inhibitors. Nature, 463(7280):501–506, 2010.

[223] W. Dameshek. Editorial: some speculations on the myeloproliferative syndromes. Blood, 6(4):372, 1951.

[224] NI Berlin et al. Diagnosis and classiﬁcation of the polycythemias. In Seminars in hematology, volume 12, page 339, 1975.

[225] B.J. An´ıa,V.J. Suman, J.L. Sobell, M.B. Codd, M.N. Silverstein, and L.J. Melton III. Trends in the incidence of polycythemia vera among olmsted county, minnesota residents, 1935–1989. American journal of hematology, 47(2):89–93, 1994.

[226] A. Teﬀeri, E. Rumi, G. Finazzi, et al. Survival and prognosis among 1263 patients with polycythemia vera: an international study. In Blood (ASH Annual Meeting Abstracts), number 21 in 118, page 277, 2011.

[227] G. Finazzi, M. Ruggeri, F. Rodeghiero, and T. Barbui. Second malignancies in patients with essential thrombocythaemia treated with busulphan and hydroxyurea: long-term follow-up of a randomized clinical trial. British journal of haematology, 110(3):577–583, 2000.

[228] R.A. Mesa, M.N. Silverstein, S.J. Jacobsen, P.C. Wollan, and A. Teﬀeri. Population-based incidence and survival ﬁgures in essential thrombocythemia and agnogenic myeloid metaplasia: An olmsted county study, 1976–1995. American journal of hematology, 61(1):10– 15, 1999.

[229] T. Barbui, J. Thiele, F. Passamonti, E. Rumi, E. Boveri, M. Ruggeri, F. Rodeghiero, E.S.G. d’Amore, M.L. Randi, I. Bertozzi, et al. Survival and disease progression in essential thrombocythemia are signiﬁcantly inﬂuenced by accurate morphologic diagnosis: an international study. Journal of Clinical Oncology, 29(23):3179–3184, 2011. References 156

[230] A. Teﬀeri. How i treat myeloﬁbrosis. Blood, 117(13):3494–3504, 2011.

[231] JF Prchal and AA Axelrad. Bone-marrow responses in polycythemia vera. New England Journal of Medicine, 290(24):1382–1382, 1974.

[232] CH Dai, SB Krantz, EN Dessypris, R.T.J. Means, ST Horn, and HS Gilbert. Poly- cythemia vera. ii. hypersensitivity of bone marrow erythroid, granulocyte-macrophage, and megakaryocyte progenitor cells to interleukin-3 and granulocyte-macrophage colony- stimulating factor. Blood, 80(4):891–899, 1992.

[233] Y. Li, G. Hetet, A.M. Maurer, Y. Chait, D. Dhermy, and J. Briere. Spontaneous megakaryocyte colony formation in myeloproliferative disorders is not neutralizable by antibodies against il3, il6 and gm-csf. British journal of haematology, 87(3):471–476, 1994.

[234] A.F. Wilks and A.G. Harpur. Cytokine signal transduction and the jak family of protein tyrosine kinases. Bioessays, 16(5):313–320, 1994.

[235] H. Luo, WP Hanratty, and CR Dearolf. An amino acid substitution in the drosophila hoptum-l jak kinase causes leukemia-like hematopoietic defects. The EMBO journal, 14(7):1412, 1995.

[236] H. Luo, P. Rose, D. Barber, W.P. Hanratty, S. Lee, T.M. Roberts, A.D. D’Andrea, and C.R. Dearolf. Mutation in the jak kinase jh2 domain hyperactivates drosophila and mammalian jak-stat pathways. Molecular and cellular biology, 17(3):1562–1571, 1997.

[237] E.J. Baxter, L.M. Scott, P.J. Campbell, C. East, N. Fourouclas, S. Swanton, G.S. Vassil- iou, A.J. Bench, E.M. Boyd, N. Curtin, et al. Acquired mutation of the tyrosine kinase jak2 in human myeloproliferative disorders. The Lancet, 365(9464):1054–1061, 2005.

[238] C. James, V. Ugo, J.P. Le Cou´edic,J. Staerk, F. Delhommeau, C. Lacout, L. Gar¸con, H. Raslova, R. Berger, A. Bennaceur-Griscelli, et al. A unique clonal jak 2 mutation leading to constitutive signalling causes polycythaemia vera. Nature, 434(7037):1144– 1148, 2005.

[239] R. Kralovics, F. Passamonti, A.S. Buser, S.S. Teo, R. Tiedt, J.R. Passweg, A. Tichelli, M. Cazzola, and R.C. Skoda. A gain-of-function mutation of jak2 in myeloproliferative disorders. New England Journal of Medicine, 352(17):1779–1790, 2005.

[240] R.L. Levine, M. Wadleigh, J. Cools, B.L. Ebert, G. Wernig, B.J.P. Huntly, T.J. Boggon, I. Wlodarska, J.J. Clark, S. Moore, et al. Activating mutation in the tyrosine kinase jak2 in polycythemia vera, essential thrombocythemia, and myeloid metaplasia with myeloﬁ- brosis. Cancer Cell, 7(4):387–397, 2005.

[241] V. Ugo, C. Marzac, I. Teyssandier, F. Larbret, Y. L´ecluse,N. Debili, W. Vainchenker, and N. Casadevall. Multiple signaling pathways are involved in erythropoietin-independent diﬀerentiation of erythroid progenitors in polycythemia vera. Experimental hematology, 32(2):179–187, 2004.

[242] B. Mohit and K. Fan. Hybrid cell line from a cloned immunoglobulin-producing mouse myeloma and a nonproducing mouse lymphoma. Science, 171(3966):75–77, 1971. References 157

[243] Z. Chen, M. Notohamiprodjo, X.Y. Guan, E. Paietta, S. Blackwell, K. Stout, A. Turner, K. Richkind, J.M. Trent, A. Lamb, et al. Gain of 9p in the pathogenesis of polycythemia vera. Genes, Chromosomes and Cancer, 22(4):321–324, 1998.

[244] N.B. Westwood, A.M. Gruszka-Westwood, S. Atkinson, and T.C. Pearson. Polycythemia vera: analysis of dna from blood granulocytes using comparative genomic hybridization. haematologica, 86(5):464–469, 2001.

[245] R. Kralovics, Y. Guan, and J.T. Prchal. Acquired uniparental disomy of chromosome 9p is a frequent stem cell defect in polycythemia vera. Experimental hematology, 30(3):229– 236, 2002.

[246] P. Martin and T. Papayannopoulou. Hel cells: a new human erythroleukemia cell line with spontaneous and induced globin expression. Science, 216(4551):1233–1235, 1982.

[247] E. Klein, F. V´anky, H. Ben-Bassat, H. Neumann, P. Ralph, J. Zeuthen, and A. Polliack. Properties of the k562 cell line, derived from a patient with chronic myeloid leukemia. International Journal of Cancer, 18(4):421–431, 1976.

[248] A. Teﬀeri and A. Pardanani. Jak inhibitors in myeloproliferative neoplasms: rationale, current data and perspective. Blood reviews, 2011.

[249] L.M. Scott, W. Tong, R.L. Levine, M.A. Scott, P.A. Beer, M.R. Stratton, P.A. Futreal, W.N. Erber, M.F. McMullin, C.N. Harrison, et al. Jak2 exon 12 mutations in polycythemia vera and idiopathic erythrocytosis. New England Journal of Medicine, 356(5):459–468, 2007.

[250] D. Pietra, S. Li, A. Brisci, F. Passamonti, E. Rumi, A. Theocharides, M. Ferrari, H. Gisslinger, R. Kralovics, L. Cremonesi, et al. Somatic mutations of jak2 exon 12 in patients with jak2 (v617f)-negative myeloproliferative disorders. Blood, 111(3):1686– 1689, 2008.

[251] L.M. Scott. The jak2 exon 12 mutations: A comprehensive review. American journal of hematology, 2011.

[252] Y. Pikman, B.H. Lee, T. Mercher, E. McDowell, B.L. Ebert, M. Gozo, A. Cuker, G. Wernig, S. Moore, I. Galinsky, et al. Mplw515l is a novel somatic activating mutation in myeloﬁbrosis with myeloid metaplasia. PLoS medicine, 3(7):e270, 2006.

[253] J.S. Greenberger, M.A. Sakakeeny, R.K. Humphries, C.J. Eaves, and R.J. Eckner. Demon- stration of permanent factor-dependent multipotential (erythroid/neutrophil/basophil) hematopoietic progenitor cell lines. Proceedings of the National Academy of Sciences, 80(10):2931, 1983.

[254] C.E. Rudd. Lnk adaptor: novel negative regulator of b cell lymphopoiesis. Science’s STKE, 2001(85):pe1, 2001.

[255] S. Gery, S. Gueller, K. Chumakova, N. Kawamata, L. Liu, and H.P. Koeﬄer. Adaptor protein lnk negatively regulates the mutant mpl, mplw515l associated with myeloproliferative disorders. Blood, 110(9):3360–3364, 2007. References 158

[256] A. Bersenev, C. Wu, J. Balcerek, and W. Tong. Lnk controls mouse hematopoietic stem cell self-renewal and quiescence through direct interactions with jak2. The Journal of clinical investigation, 118(8):2832, 2008.

[257] S. Takaki, H. Morita, Y. Tezuka, and K. Takatsu. Enhanced hematopoiesis by hematopoietic progenitor cells lacking intracellular adaptor protein, lnk. The Journal of experimental medicine, 195(2):151, 2002.

[258] L. Velazquez, A.M. Cheng, H.E. Fleming, C. Furlonger, S. Vesely, A. Bernstein, C.J. Paige, and T. Pawson. Cytokine signaling and hematopoietic homeostasis are disrupted in lnk-deﬁcient mice. The Journal of experimental medicine, 195(12):1599–1611, 2002.

[259] W. Tong and H.F. Lodish. Lnk inhibits tpo–mpl signaling and tpo-mediated megakary- ocytopoiesis. The Journal of experimental medicine, 200(5):569, 2004.

[260] S.T. Oh, E.F. Simonds, C. Jones, M.B. Hale, Y. Goltsev, K.D. Gibbs Jr, J.D. Merker, J.L. Zehnder, G.P. Nolan, and J. Gotlib. Novel mutations in the inhibitory adaptor protein lnk drive jak-stat signaling in patients with myeloproliferative neoplasms. Blood, 116(6):988–992, 2010.

[261] F. Delhommeau, S. Dupont, C. James, A. Masse, JP le Couedic, VD Valle, et al. Tet2 is a novel tumor suppressor gene inactivated in myeloproliferative neoplasms: identiﬁcation of a pre-jak2 v617f event. Blood, 112(11), 2008.

[262] F. Delhommeau, S. Dupont, V.D. Valle, C. James, S. Trannoy, A. Mass´e,O. Kosmider, J.P. Le Couedic, F. Robert, A. Alberdi, et al. Mutation in tet2 in myeloid cancers. New England Journal of Medicine, 360(22):2289–2301, 2009.

[263] O. Abdel-Wahab, A. Mullally, C. Hedvat, G. Garcia-Manero, J. Patel, M. Wadleigh, S. Malinge, J.J. Yao, O. Kilpivaara, R. Bhat, et al. Genetic characterization of tet1, tet2, and tet3 alterations in myeloid malignancies. Blood, 114(1):144–147, 2009.

[264] A. Teﬀeri, KH Lim, O. Abdel-Wahab, TL Lasho, J. Patel, MM Patnaik, CA Hanson, A. Pardanani, DG Gilliland, and RL Levine. Detection of mutant tet2 in myeloid malignancies other than myeloproliferative neoplasms: Cmml, mds, mds/mpn and aml. Leu- kemia, 23(7):1343–1345, 2009.

[265] L. Couronne, E. Lippert, J. Andrieux, O. Kosmider, I. Radford-Weiss, D. Penther, N. Das- tugue, F. Mugneret, M. Lafage, N. Gachard, et al. Analyses of tet2 mutations in post- myeloproliferative neoplasm acute myeloid leukemias. Leukemia, 24(1):201–203, 2009.

[266] C. Saint-Martin, G. Leroy, F. Delhommeau, G. Panelatti, S. Dupont, C. James, I. Plo, D. Bordessoule, C. Chomienne, A. Delannoy, et al. Analysis of the ten-eleven translocation 2 (tet2) gene in familial myeloproliferative neoplasms. Blood, 114(8):1628–1632, 2009.

[267] A.M. Jankowska, H. Szpurka, R.V. Tiu, H. Makishima, M. Afable, J. Huh, C.L. O’Keefe, R. Ganetzky, M.A. McDevitt, and J.P. Maciejewski. Loss of heterozygosity 4q24 and tet2 mutations associated with myelodysplastic/myeloproliferative neoplasms. Blood, 113(25):6403–6410, 2009.

[268] S.M.C. Langemeijer, R.P. Kuiper, M. Berends, R. Knops, M.G. Aslanyan, M. Massop, E. Stevens-Linders, P. van Hoogen, A.G. van Kessel, R.A.P. Raymakers, et al. Acquired References 159

mutations in tet2 are common in myelodysplastic syndromes. Nature genetics, 41(7):838– 842, 2009. [269] A.M. Mohamedali, A.E. Smith, J. Gaken, N.C. Lea, S.A. Mian, N.B. Westwood, C. Strupp, N. Gattermann, U. Germing, and G.J. Mufti. Novel tet2 mutations associated with upd4q24 in myelodysplastic syndrome. Journal of Clinical Oncology, 27(24):4002– 4006, 2009. [270] O. Kosmider, V. Gelsi-Boyer, C. Racoeur, V. Jooste, N. Vey, B. Quesnel, P. Fenaux, J.N. Bastie, O. Beyne-Rauzy, A. Stamatoulas, et al. Tet2 gene mutation is a frequent and adverse event in chronic myelomonocytic leukemia. Haematologica, 94(12):1676–1681, 2009. [271] A. Teﬀeri, RL Levine, KH Lim, O. Abdel-Wahab, TL Lasho, J. Patel, CM Finke, A. Mul- lally, CY Li, A. Pardanani, et al. Frequent tet2 mutations in systemic mastocytosis: clinical, kitd816v and ﬁp1l1-pdgfra correlates. Leukemia, 23(5):900–904, 2009. [272] R. Ono, T. Taki, T. Taketani, H. Kawaguchi, M. Taniwaki, T. Okamura, K. Kawa, R. Hanada, M. Kobayashi, and Y. Hayashi. Septin6, a human homologue to mouse septin6, is fused to mll in infant acute myeloid leukemia with complex chromosomal abnormalities involving 11q23 and xq24. Cancer research, 62(2):333, 2002. [273] RB Lorsbach, J. Moore, S. Mathew, SC Raimondi, ST Mukatira, and JR Downing. Tet1, a member of a novel protein family, is fused to mll in acute myeloid leukemia containing the t (10; 11)(q22; q23). Leukemia, 17(3):637–641, 2003. [274] P. Borst and R. Sabatini. Base j: discovery, biosynthesis, and possible functions. Annu. Rev. Microbiol., 62:235–251, 2008. [275] M. Tahiliani, K.P. Koh, Y. Shen, W.A. Pastor, H. Bandukwala, Y. Brudno, S. Agar- wal, L.M. Iyer, D.R. Liu, L. Aravind, et al. Conversion of 5-methylcytosine to 5- hydroxymethylcytosine in mammalian dna by mll partner tet1. Science, 324(5929):930– 935, 2009. [276] C.L. Fisher, J. Berger, F. Randazzo, and H.W. Brock. A human homolog of additional sex combs, additional sex combs-like 1, maps to chromosome 20q11. Gene, 306:115–126, 2003. [277] U.H. Park, S.K. Yoon, T. Park, E.J. Kim, and S.J. Um. Additional sex comb-like (asxl) proteins 1 and 2 play opposite roles in adipogenesis via reciprocal regulation of peroxisome proliferator-activated receptor γ. Journal of Biological Chemistry, 286(2):1354–1363, 2011. [278] C.L. Fisher, N. Pineault, C. Brookes, C.D. Helgason, H. Ohta, C. Bodner, J.L. Hess, R.K. Humphries, and H.W. Brock. Loss-of-function additional sex combs like 1 mutations disrupt hematopoiesis but do not cause severe myelodysplasia or leukemia. Blood, 115(1):38–46, 2010. [279] S. Heim and F. Mitelman. Cancer cytogenetics. Wiley Online Library, 1995. [280] T. Ernst, A.J. Chase, J. Score, C.E. Hidalgo-Curtis, C. Bryant, A.V. Jones, K. Waghorn, K. Zoi, F.M. Ross, A. Reiter, et al. Inactivating mutations of the histone methyltransferase gene ezh2 in myeloid disorders. Nature genetics, 42(8):722–726, 2010. References 160

[281] J.A. Simon and C.A. Lange. Roles of the ezh2 histone methyltransferase in cancer epi- genetics. Mutation Research/Fundamental and Molecular Mechanisms of Mutagenesis, 647(1-2):21–29, 2008.

[282] A.P. Bracken and K. Helin. Polycomb group proteins: navigators of lineage pathways led astray in cancer. Nature Reviews Cancer, 9(11):773–784, 2009.

[283] D. O’Carroll, S. Erhardt, M. Pagani, S.C. Barton, M.A. Surani, and T. Jenuwein. The polycomb-group geneezh2 is required for early mouse development. Molecular and cellular biology, 21(13):4330–4336, 2001.

[284] N. Carbuccia, A. Murati, V. Trouplin, M. Brecqueville, J. Adelaide, J. Rey, W. Vainchenker, OA Bernard, M. Chaﬀanet, N. Vey, et al. Mutations of asxl1 gene in myeloproliferative neoplasms. Leukemia, 23(11):2183–2186, 2009.

[285] C. Bellann´e-Chantelot, I. Chaumarel, M. Labopin, F. Bellanger, V. Barbu, C. De Toma, F. Delhommeau, N. Casadevall, W. Vainchenker, G. Thomas, et al. Genetic and clinical implications of the val617phe jak2 mutation in 72 families with myeloproliferative disorders. Blood, 108(1):346–352, 2006.

[286] E. Rumi, F. Passamonti, M.G. Della Porta, C. Elena, L. Arcaini, L. Vanelli, C. Del Curto, D. Pietra, E. Boveri, C. Pascutto, et al. Familial chronic myeloproliferative disorders: clinical phenotype and evidence of disease anticipation. Journal of Clinical Oncology, 25(35):5630–5635, 2007.

[287] O. Landgren, L.R. Goldin, S.Y. Kristinsson, E.A. Helgadottir, J. Samuelsson, and M. Björkholm. Increased risks of polycythemia vera, essential thrombocythemia, and myelofibrosis among 24 577 first-degree relatives of 11 039 patients with myeloproliferative neoplasms in sweden. Blood, 112(6):2199–2204, 2008.

[288] A.V. Jones, A. Chase, R.T. Silver, D. Oscier, K. Zoi, Y.L. Wang, H. Cario, H.L. Pahl, A. Collins, A. Reiter, et al. Jak2 haplotype is a major risk factor for the development of myeloproliferative neoplasms. Nature Genetics, 41(4):446–449, 2009.

[289] O. Kilpivaara, S. Mukherjee, A.M. Schram, M. Wadleigh, A. Mullally, B.L. Ebert, A. Bass, S. Marubayashi, A. Heguy, G. Garcia-Manero, et al. A germline jak2 snp is associated with predisposition to the development of jak2v617f-positive myeloproliferative neoplasms. Na- ture Genetics, 41(4):455–459, 2009.

[290] D. Olcaydu, A. Harutyunyan, R. J¨ager,T. Berg, B. Gisslinger, I. Pabinger, H. Gisslinger, and R. Kralovics. A common jak2 haplotype confers susceptibility to myeloproliferative neoplasms. Nature Genetics, 41(4):450–454, 2009.

[291] P.J. Campbell. Somatic and germline genetics at the jak2 locus. Nature genetics, 41(4):385–386, 2009.

[292] A.V. Jones, P.J. Campbell, P.A. Beer, S. Schnittger, A.M. Vannucchi, K. Zoi, M.J. Percy, M.F. McMullin, L.M. Scott, W. Tapper, et al. The jak2 46/1 haplotype predisposes to mpl-mutated myeloproliferative neoplasms. Blood, 115(22):4517–4523, 2010.

[293] P.A. Beer, F. Delhommeau, J.P. LeCou´edic, M.A. Dawson, E. Chen, D. Bareford, R. Kuˇsec, M.F. McMullin, C.N. Harrison, A.M. Vannucchi, et al. Two routes to References 161

leukemic transformation after a jak2 mutation–positive myeloproliferative neoplasm. Blood, 115(14):2891–2900, 2010.

[294] F.X. Schaub, R. Looser, S. Li, H. Hao-Shen, T. Lehmann, A. Tichelli, and R.C. Skoda. Clonal analysis of tet2 and jak2 mutations suggests that tet2 can be a late event in the progression of myeloproliferative neoplasms. Blood, 115(10):2003, 2010.

[295] A. Theocharides, M. Boissinot, F. Girodon, R. Garand, S.S. Teo, E. Lippert, P. Talmant, A. Tichelli, S. Hermouet, and R.C. Skoda. Leukemic blasts in transformed jak2-v617f– positive myeloproliferative disorders are frequently negative for the jak2-v617f mutation. Blood, 110(1):375–379, 2007.

[296] P.A. Beer, A.V. Jones, A.J. Bench, A. Goday-Fernandez, E.M. Boyd, K.J. Vaghela, W.N. Erber, B. Odeh, C. Wright, M.F. McMullin, et al. Clonal diversity in the myeloproliferative neoplasms: independent origins of genetically distinct clones. British journal of haematology, 144(6):904–908, 2009.

[297] CHM Jamieson, CF Barroga, and WP Vainchenker. Miscreant myeloproliferative disorder stem cells. Leukemia, 22(11):2011–2019, 2008.

[298] A. Quint´as-Cardama,K. Vaddi, P. Liu, T. Manshouri, J. Li, P.A. Scherle, E. Caulder, X. Wen, Y. Li, P. Waeltz, et al. Preclinical characterization of the selective jak1/2 inhibitor incb018424: therapeutic implications for the treatment of myeloproliferative neoplasms. Blood, 115(15):3109–3117, 2010.

[299] S. Verstovsek, H. Kantarjian, R.A. Mesa, A.D. Pardanani, J. Cortes-Franco, D.A. Thomas, Z. Estrov, J.S. Fridman, E.C. Bradley, S. Erickson-Viitanen, et al. Safety and eﬃcacy of incb018424, a jak1 and jak2 inhibitor, in myeloﬁbrosis. New England Journal of Medicine, 363(12):1117–1127, 2010.

[300] S. Verstovsek, RA Mesa, JR Gotlib, RS Levy, V. Gupta, JF DiPersio, et al. Results of comfort-i, a randomized double-blind phase iii trial of jak 1/2 inhibitor incb18424 (424) versus placebo (pb) for patients with myeloﬁbrosis (mf). J Clin Oncol, 29(15):6500, 2011.

[301] S. Verstovsek, R. Mesa, J. Gotlib, RA Levy, V. Gupta, et al. Consistent beneﬁt of ruxolitinib over placebo in spleen volume reduction and symptom improvement across subgroups and overall survival advantage: results from comfort-i. In American Society of Hematology 2011 Annual Meeting (abstract 278), 2011.

[302] G. Wernig, M.G. Kharas, R. Okabe, S.A. Moore, D.S. Leeman, D.E. Cullen, M. Gozo, E.P. McDowell, R.L. Levine, J. Doukas, et al. Eﬃcacy of tg101348, a selective jak2 inhibitor, in treatment of a murine model of jak2v617f-induced polycythemia vera. Cancer cell, 13(4):311–320, 2008.

[303] I. Geron, A.E. Abrahamsson, C.F. Barroga, E. Kavalerchik, J. Gotlib, J.D. Hood, J. Durocher, C.C. Mak, G. Noronha, R.M. Soll, et al. Selective inhibition of jak2-driven erythroid diﬀerentiation of polycythemia vera progenitors. Cancer cell, 13(4):321–330, 2008.

[304] A. Pardanani, J.R. Gotlib, C. Jamieson, J.E. Cortes, M. Talpaz, R.M. Stone, M.H. Silver- man, D.G. Gilliland, J. Shorr, and A. Tefferi. Safety and efficacy of tg101348, a selective jak2 inhibitor, in myelofibrosis. Journal of Clinical Oncology, 29(7):789–796, 2011. References 162

[305] M. Levis, K.F. Tse, B.D. Smith, E. Garrett, and D. Small. A ﬂt3 tyrosine kinase inhibitor is selectively cytotoxic to acute myeloid leukemia blasts harboring ﬂt3 internal tandem duplication mutations. Blood, 98(3):885–887, 2001.

[306] M. Levis, J. Allebach, K.F. Tse, R. Zheng, B.R. Baldwin, B.D. Smith, S. Jones-Bolin, B. Ruggeri, C. Dionne, and D. Small. A ﬂt3-targeted tyrosine kinase inhibitor is cytotoxic to leukemia cells in vitro and in vivo. Blood, 99(11):3885–3891, 2002.

[307] E.O. Hexner, C. Serdikoﬀ, M. Jan, C.R. Swider, C. Robinson, S. Yang, T. Angeles, S.G. Emerson, M. Carroll, B. Ruggeri, et al. Lestaurtinib (cep701) is a jak2 inhibitor that suppresses jak2/stat5 signaling and the proliferation of primary erythroid cells from patients with myeloproliferative disorders. Blood, 111(12):5663–5671, 2008.

[308] F.P.S. Santos, H.M. Kantarjian, N. Jain, T. Manshouri, D.A. Thomas, G. Garcia-Manero, D. Kennedy, Z. Estrov, J. Cortes, and S. Verstovsek. Phase 2 study of cep-701, an orally available jak2 inhibitor, in patients with primary or post-polycythemia vera/essential thrombocythemia myeloﬁbrosis. Blood, 115(6):1131–1136, 2010.

[309] A. Pardanani, T. Lasho, G. Smith, CJ Burns, E. Fantino, and A. Teﬀeri. Cyt387, a selective jak1/jak2 inhibitor: in vitro assessment of kinase selectivity and preclinical studies using cell lines and primary cells from polycythemia vera patients. Leukemia, 23(8):1441–1445, 2009.

[310] J.W. Tyner, T.G. Bumm, J. Deininger, L. Wood, K.J. Aichberger, M.M. Loriaux, B.J. Druker, C.J. Burns, E. Fantino, and M.W. Deininger. Cyt387, a novel jak2 inhibitor, induces hematologic responses and normalizes inﬂammatory cytokines in murine myeloproliferative neoplasms. Blood, 115(25):5232–5240, 2010.

[311] KA Monaghan, T. Khong, CJ Burns, and A. Spencer. The novel jak inhibitor cyt387 suppresses multiple signalling pathways, prevents proliferation and induces apoptosis in phenotypically diverse myeloma cells. Leukemia, 2011.

[312] A.D. William, A.C.H. Lee, S. Blanchard, A. Poulsen, E. Teo, H. Nagaraj, E. Tan, D. Chen, M. Williams, E.T. Sun, et al. Discovery of the macrocycle 11-(2-pyrrolidin-1-yl-ethoxy)- 14, 19-dioxa-5, 7, 26-triaza-tetracyclo [19.3. 1.1 (2, 6). 1 (8, 12)] heptacosa-1 (25), 2 (26), 3, 5, 8, 10, 12 (27), 16, 21, 23-decaene (sb1518), a potent janus kinase 2/fms-liketyrosine kinase-3 (jak2/ﬂt3) inhibitor for the treatment of myeloﬁbrosis and lymphoma. Journal of medicinal chemistry, 2011.

[313] S. Hart, KC Goh, V. Novotny-Diermayr, CY Hu, H. Hentze, YC Tan, B. Madan, C. Ama- lini, YK Loh, LC Ong, et al. Sb1518, a novel macrocyclic pyrimidine-based jak2 inhibitor for the treatment of myeloid and lymphoid malignancies. Leukemia, 25(11):1751–1759, 2011.

[314] K. Uozumi, M. Otsuka, N. Ohno, T. Moriyama, S. Suzuki, S. Shimotakahara, I. Mat- sumura, S. Hanada, and T. Arima. Establishment and characterization of a new human megakaryoblastic cell line (set-2) that spontaneously matures to megakaryocytes and produces platelet-like particles. Leukemia, 14(1):142–152, 2000.

[315] A. Poulsen, A. William, S. Blanchard, A. Lee, H. Nagaraj, H. Wang, E. Teo, E. Tan, K.C. Goh, and B. Dymock. Structure-based design of oxygen-linked macrocyclic kinase References 163

inhibitors: discovery of sb1518 and sb1578, potent inhibitors of janus kinase 2 (jak2) and fms-like tyrosine kinase-3 (flt3). Journal of computer-aided molecular design, pages 1–14, 2012. [316] S. Verstovsek, O. Odenike, B. Scott, Z. Estrov, J. Cortes, DA Thomas, J. Wood, K. Ethi- rajulu, A. Lowe, HJ Zhu, et al. Phase i dose-escalation trial of sb1518, a novel jak2/flt3 inhibitor, in acute and chronic myeloid diseases, including primary or post-essential thrombocythemia/polycythemia vera myelofibrosis. Blood, 114(22):3905, 2009. [317] G. Wernig, T. Mercher, R. Okabe, R.L. Levine, B.H. Lee, and D.G. Gilliland. Expression of jak2v617f causes a polycythemia vera–like disease with associated myelofibrosis in a murine bone marrow transplant model. Blood, 107(11):4274–4281, 2006. [318] C. Lacout, D.F. Pisani, M. Tulliez, F.M. Gachelin, W. Vainchenker, J. Villeval, et al. Jak2 v617f expression in murine hematopoietic cells leads to mpd mimicking human pv with secondary myelofibrosis. Blood, 108(5):1652, 2006. [319] T.G.P. Bumm, C. Elsea, A.S. Corbin, M. Loriaux, D. Sherbenou, L. Wood, J. Deininger, R.T. Silver, B.J. Druker, and M.W.N. Deininger. Characterization of murine jak2v617f- positive myeloproliferative disease. Cancer research, 66(23):11156–11165, 2006. [320] V.M. Zaleskas, D.S. Krause, K. Lazarides, N. Patel, Y. Hu, S. Li, and R.A. Van Etten. Molecular pathogenesis and therapy of polycythemia induced in mice by jak2 v617f. PLoS One, 1(1):e18, 2006. [321] R. Tiedt, H. Hao-Shen, M.A. Sobas, R. Looser, S. Dirnhofer, J. Schwaller, and R.C. Skoda. Ratio of mutant jak2-v617f to wild-type jak2 determines the mpd phenotypes in transgenic mice. Blood, 111(8):3931–3940, 2008. [322] P. Georgiades, S. Ogilvy, H. Duval, D.R. Licence, D.S. Charnock-Jones, S.K. Smith, et al. Vavcre transgenic mice: a tool for mutagenesis in hematopoietic and endothelial lineages. Genesis, 34(4):251–256, 2002. [323] P. Staeheli, P. Danielson, O. Haller, and JG Sutcliffe. Transcriptional activation of the mouse mx gene by type i interferon. Molecular and cellular biology, 6(12):4770–4774, 1986. [324] H. Hug, M. Costas, P. Staeheli, M. Aebi, and C. Weissmann. Organization of the murine mx gene and characterization of its interferon-and virus-inducible promoter. Molecular and cellular biology, 8(8):3065–3079, 1988. [325] H. Akada, D. Yan, H. Zou, S. Fiering, R.E. Hutchison, and M.G. Mohi. Conditional expression of heterozygous or homozygous jak2v617f from its endogenous promoter induces a polycythemia vera–like disease. Blood, 115(17):3589–3597, 2010. [326] A. Mullally, S.W. Lane, B. Ball, C. Megerdichian, R. Okabe, F. Al-Shahrour, M. Paktinat, J.E. Haydu, E. Housman, A.M. Lord, et al. Physiological jak2v617fexpression causes a lethal myeloproliferative neoplasm with differential effects on hematopoietic stem and progenitor cells. Cancer cell, 17(6):584–596, 2010. [327] A. Tefferi and W. Vainchenker. Myeloproliferative neoplasms: molecular pathophysiology, essential clinical understanding, and treatment strategies. Journal of Clinical Oncology, 29(5):573–582, 2011. References 164

[328] R.L. Levine and D.G. Gilliland. Myeloproliferative disorders. Blood, 112(6):2190–2198, 2008.

[329] X. Lu, R. Levine, W. Tong, G. Wernig, Y. Pikman, S. Zarnegar, D.G. Gilliland, and H. Lodish. Expression of a homodimeric type i cytokine receptor is required for jak2v617f- mediated transformation. Proceedings of the National Academy of Sciences of the United States of America, 102(52):18962–18967, 2005.

[330] J.D. Rowley. A new consistent chromosomal abnormality in chronic myelogenous leukaemia identiﬁed by quinacrine ﬂuorescence and giemsa staining. Nature, 243:290– 293, 1973.

[331] B.J. Druker, C.L. Sawyers, H. Kantarjian, D.J. Resta, S.F. Reese, J.M. Ford, R. Capdev- ille, and M. Talpaz. Activity of a speciﬁc inhibitor of the bcr-abl tyrosine kinase in the blast crisis of chronic myeloid leukemia and acute lymphoblastic leukemia with the philadelphia chromosome. New England Journal of Medicine, 344(14):1038–1042, 2001.

[332] E. Buchdunger, J. Zimmermann, H. Mett, T. Meyer, M. Muller, B.J. Druker, and N.B. Lydon. Inhibition of the abl protein-tyrosine kinase in vitro and in vivo by a 2- phenylaminopyrimidine derivative. Cancer Research, 56(1):100–104, 1996.

[333] C.L. Sawyers, A. Hochhaus, E. Feldman, J.M. Goldman, C.B. Miller, O.G. Ottmann, C.A. Schiﬀer, M. Talpaz, F. Guilhot, M.W.N. Deininger, et al. Imatinib induces hematologic and cytogenetic responses in patients with chronic myelogenous leukemia in myeloid blast crisis: results of a phase ii study. Blood, 99(10):3530–3539, 2002.

[334] P. le Coutre, E. Tassi, M. Varella-Garcia, R. Barni, L. Mologni, G. Cabrita, E. Marchesi, R. Supino, and C. Gambacorti-Passerini. Induction of resistance to the abelson inhibitor sti571 in human leukemic cells through gene ampliﬁcation. Blood, 95(5):1758–1766, 2000.

[335] F.X. Mahon, M.W.N. Deininger, B. Schultheis, J. Chabrol, J. Reiﬀers, J.M. Goldman, and J.V. Melo. Selection and characterization of bcr-abl positive cell lines with diﬀerential sensitivity to the tyrosine kinase inhibitor sti571: diverse mechanisms of resistance. Blood, 96(3):1070–1079, 2000.

[336] M. Talpaz, N.P. Shah, H. Kantarjian, N. Donato, J. Nicoll, R. Paquette, J. Cortes, S. O’Brien, C. Nicaise, E. Bleickardt, et al. Dasatinib in imatinib-resistant philadelphia chromosome–positive leukemias. New England Journal of Medicine, 354(24):2531–2541, 2006.

[337] H.M. Kantarjian, F.J. Giles, K.N. Bhalla, J. Pinilla-Ibarz, R.A. Larson, N. Gattermann, O.G. Ottmann, A. Hochhaus, J.P. Radich, G. Saglio, et al. Nilotinib is eﬀective in patients with chronic myeloid leukemia in chronic phase after imatinib resistance or intolerance: 24-month follow-up results. Blood, 117(4):1141–1145, 2011.

[338] N.P. Shah, C. Tran, F.Y. Lee, P. Chen, D. Norris, and C.L. Sawyers. Overriding imatinib resistance with a novel abl kinase inhibitor. Science, 305(5682):399, 2004.

[339] E. Weisberg, X. Deng, HG Choi, R. Barrett, S. Adamia, A. Ray, D. Moreno, AL Kung, N. Gray, and JD Griffin. Beneficial effects of combining a type ii atp competitive inhibitor with an allosteric competitive inhibitor of bcr-abl for the treatment of imatinib-sensitive and imatinib-resistant cml. Leukemia, 24(7):1375–1378, 2010. References 165

[340] A. Pardanani and A. Teﬀeri. Targeting myeloproliferative neoplasms with jak inhibitors. Current Opinion in Hematology, 18(2):105–110, 2011.

[341] N.L. Warner, MA Moore, D. Metcalf, et al. A transplantable myelomonocytic leukemia in balb-c mice: cytology, karyotype, and muramidase content. Journal of the National Cancer Institute, 43(4):963, 1969.

[342] D. Metcalf, MA Moore, and NL Warner. Colony formation in vitro by myelomonocytic leukemic cells. Journal of the National Cancer Institute, 43(4):983, 1969.

[343] S. Ymer, WQJ Tucker, CJ Sanderson, AJ Hapel, HD Campbell, and IG Young. Constitu- tive synthesis of interleukin-3 by leukaemia cell line wehi-3b is due to retroviral insertion near the gene. Nature, 317:255–258, 1985.

[344] S. Swift, J. Lorens, P. Achacoso, and G.P. Nolan. Rapid production of retroviruses for eﬃcient gene delivery to mammalian cells using 293t cell–based systems. Current protocols in immunology, 10:14–20, 1999.

[345] P.W. Laird, A. Zijderveld, K. Linders, M.A. Rudnicki, R. Jaenisch, and A. Berns. Sim- pliﬁed mammalian dna isolation procedure. Nucleic Acids Research, 19(15):4293, 1991.

[346] I.S. Lucet, E. Fantino, M. Styles, R. Bamert, O. Patel, S.E. Broughton, M. Walter, C.J. Burns, H. Treutlein, A.F. Wilks, et al. The structural basis of janus kinase 2 inhibition by a potent and speciﬁc pan-janus kinase inhibitor. Blood, 107(1):176, 2006.

[347] D.J. George, C.A. Dionne, J. Jani, T. Angeles, C. Murakata, J. Lamb, and J.T. Isaacs. Sustained in vivo regression of dunning h rat prostate cancers treated with combinations of androgen ablation and trk tyrosine kinase inhibitors, cep-751 (kt-6587) or cep-701 (kt-5555). Cancer research, 59(10):2395, 1999.

[348] C. Haan, D.C. Kroy, S. Wuller, U. Sommer, T. Nocker, C. Rolvering, I. Behrmann, P.C. Heinrich, and S. Haan. An unusual insertion in jak2 is crucial for kinase activity and diﬀerentially aﬀects cytokine responses. The Journal of Immunology, 182(5):2969, 2009.

[349] Z. Li, S. Xing, S. Wang, W.T. Ho, and Z.J. Zhao. Characterization of a highly effective protein substrate for analysis of jak2v617f activity. Experimental Hematology, 35(11):1624–1632, 2007.

[350] A. Pradhan, Q.T. Lambert, L.N. Griner, and G.W. Reuther. Activation of jak2-v617f by components of heterodimeric cytokine receptors. Journal of Biological Chemistry, 285(22):16651, 2010.

[351] W.K. Hofmann, M. Komor, B. Wassmann, L.C. Jones, H. Gschaidmeier, D. Hoelzer, H.P. Koeﬄer, and O.G. Ottmann. Presence of the bcr-abl mutation glu255lys prior to sti571 (imatinib) treatment in patients with ph+ acute lymphoblastic leukemia. Blood, 102(2):659, 2003.

[352] S.G. Willis, T. Lange, S. Demehri, S. Otto, L. Crossman, D. Niederwieser, E.P. Stoﬀregen, S. McWeeney, I. Kovacs, B. Park, et al. High-sensitivity detection of bcr-abl kinase domain mutations in imatinib-naive patients: correlation with clonal cytogenetic evolution but not response to therapy. Blood, 106(6):2128, 2005. References 166

[353] B.L. Stein, D.M. Williams, C. O’Keefe, O. Rogers, R.G. Ingersoll, J.L. Spivak, A. Verma, J.P. Maciejewski, M.A. McDevitt, and A.R. Moliterno. Disruption of the asxl1 gene is frequent in primary, post-essential thrombocytosis and post-polycythemia vera myeloﬁ- brosis, but not essential thrombocytosis or polycythemia vera: analysis of molecular genetics and clinical phenotypes. Haematologica, page Epub ahead of print, 2011.

[354] T. Hornakova, L. Springuel, J. Devreux, A. Dusa, S.N. Constantinescu, L. Knoops, and J.C. Renauld. Oncogenic jak1 and jak2-activating mutations resistant to atp-competitive inhibitors. Haematologica, 96(6):845, 2011.

[355] A. Deshpande, MM Reddy, GOM Schade, A. Ray, TK Chowdary, JD Griﬃn, and M. Sat- tler. Kinase domain mutations confer resistance to novel inhibitors targeting jak2v617f in myeloproliferative neoplasms. Leukemia, 2011.

[356] O. Weigert, A.A. Lane, L. Bird, N. Kopp, B. Chapuy, D. van Bodegom, A.V. Toms, S. Marubayashi, A.L. Christie, M. McKeown, et al. Genetic resistance to jak2 enzymatic inhibitors is overcome by hsp90 inhibition. The Journal of Experimental Medicine, 2012.

[357] F.H. Epstein, R. Kurzrock, J.U. Gutterman, and M. Talpaz. The molecular genetics of philadelphia chromosome–positive leukemias. New England Journal of Medicine, 319(15):990–998, 1988.

[358] H.M. Kantarjian, M. Shtalrid, R. Kurzrock, M. Buck, W.T. Dalton, A. Lemaistre, S.A. Stass, K.B. McCredie, J. Gutterman, E.J. Freireich, et al. Signiﬁcance and correlations of molecular analysis results in patients with philadelphia chromosome-negative chronic myelogenous leukemia and chronic myelomonocytic leukemia. The American journal of medicine, 85(5):639–644, 1988.

[359] F. Rassool, P. Martiat, A. Taj, I. Klisak, J. Goldman, et al. Interstitial insertion of varying amounts of abl-containing genetic material into chromosome 22 in ph-negative cml. Leukemia: oﬃcial journal of the Leukemia Society of America, Leukemia Research Fund, UK, 4(4):273, 1990.

[360] N. Heisterkamp, G. Jenster, J. Ten Hoeve, D. Zovich, P.K. Pattengale, and J. Groﬀen. Acute leukaemia in bcr/abl transgenic mice. Nature, pages 251–253, 1990.

[361] A.G. Elefanty, I.K. Hariharan, and S. Cory. bcr-abl, the hallmark of chronic myeloid leukaemia in man, induces multiple haemopoietic neoplasms in mice. The EMBO journal, 9(4):1069, 1990.

[362] J.E. Sokal, S.S. Leong, and G.A. Gomez. Preferential inhibition by cytarabine of cfu-gm from patients with chronic granulocytic leukemia. Cancer, 59(1):197–202, 1987.

[363] J.E. Sokal, J.P. Gockerman, and S.H. Bigner. Evidence for a selective antileukemic eﬀect of cytosine arabinoside in chronic granulocytic leukemia. Leukemia research, 12(6):453– 458, 1988.

[364] M.J. Robertson, R. Tantravahi, J.D. Griﬃn, G.P. Canellos, and S.A. Cannistra. Hema- tologic remission and cytogenetic improvement after treatment of stable-phase chronic myelogenous leukemia with continuous infusion of low-dose cytarabine. American journal of hematology, 43(2):95–102, 1993. References 167

[365] M. Talpaz, H.M. Kantarjian, K. McCredie, J.M. Trujillo, M.J. Keating, and J.U. Gut- terman. Hematologic remission and cytogenetic improvement induced by recombinant human interferon alphaa in chronic myelogenous leukemia. New England Journal of Medicine, 314(17):1065–1069, 1986.

[366] M.H. Cohen, G. Williams, J.R. Johnson, J. Duan, J. Gobburu, A. Rahman, K. Benson, J. Leighton, S.K. Kim, R. Wood, et al. Approval summary for imatinib mesylate capsules in the treatment of chronic myelogenous leukemia. Clinical cancer research, 8(5):935–942, 2002.

[367] H. Kantarjian, C. Sawyers, A. Hochhaus, F. Guilhot, C. Schiﬀer, C. Gambacorti-Passerini, D. Niederwieser, D. Resta, R. Capdeville, U. Zoellner, et al. Hematologic and cytogenetic responses to imatinib mesylate in chronic myelogenous leukemia. New England Journal of Medicine, 346(9):645–652, 2002.

[368] C.Y.A. Chen and A.B. Shyu. Au-rich elements: characterization and importance in mrna degradation. Trends in biochemical sciences, 20(11):465–470, 1995.

[369] W. Tam, M. Gomez, A. Chadburn, J.W. Lee, W.C. Chan, and D.M. Knowles. Mutational analysis of prdm1 indicates a tumor-suppressor role in diﬀuse large b-cell lymphomas. Blood, 107(10):4090–4100, 2006.

[370] L. Pasqualucci, M. Compagno, J. Houldsworth, S. Monti, A. Grunn, S.V. Nandula, J.C. Aster, V.V. Murty, M.A. Shipp, and R. Dalla-Favera. Inactivation of the prdm1/blimp1 gene in diﬀuse large b cell lymphoma. The Journal of experimental medicine, 203(2):311– 317, 2006.

[371] G. Steele-Perkins, W. Fang, X.H. Yang, M. Van Gele, T. Carling, J. Gu, I.M. Buyse, J.A. Fletcher, J. Liu, R. Bronson, et al. Tumor formation and inactivation of riz1, an rb-binding member of a nuclear protein–methyltransferase superfamily. Genes & development, 15(17):2250, 2001.

[372] S. Fears, C. Mathieu, N. Zeleznik-Le, S. Huang, J.D. Rowley, and G. Nucifora. Intergenic splicing of mds1 and evi1 occurs in normal tissues as well as in myeloid leukemia and produces a new member of the pr domain family. Proceedings of the National Academy of Sciences, 93(4):1642, 1996.

[373] U. Modlich, A. Schambach, MH Brugman, DC Wicke, S. Knoess, Z. Li, T. Maetzig, C. Rudolph, B. Schlegelberger, and C. Baum. Leukemia induction after a single retroviral vector insertion in evi1 or prdm16. Leukemia, 22(8):1519–1528, 2008.

[374] S. Kreil, M. Pﬁrrmann, C. Haferlach, K. Waghorn, A. Chase, R. Hehlmann, A. Reiter, A. Hochhaus, and N.C.P. Cross. Heterogeneous prognostic impact of derivative chromosome 9 deletions in chronic myelogenous leukemia. Blood, 110(4):1283–1290, 2007.

[375] FL Graham, J. Smiley, WC Russell, and R. Nairn. Characteristics of a human cell line transformed by dna from human adenovirus type 5. Journal of General Virology, 36(1):59–72, 1977.

[376] K. Kuriyama, R.P. Gale, M. Tomonaga, S. Ikeda, E. Yao, I. Klisak, K. Whelan, H. Yakir, M. Ichimaru, and R.S. Sparkes. Cml-t1: a cell line derived from t-lymphocyte acute phase of chronic myelogenous leukemia. Blood, 74(4):1381–1387, 1989. References 168

[377] A. Keating, J.W. Singer, P.D. Killen, G.E. Striker, A.C. Salo, J. Sanders, E.D. Thomas, D. Thorning, and P.J. Fialkow. Donor origin of the in vitro haematopoietic microenvi- ronment after marrow transplantation in man. Nature, pages 280–283, 1982.

[378] W.H. Raskind, C.M. Disteche, A. Keating, and J.W. Singer. Correlation between cytogenetic and molecular ﬁndings in human chronic myelogenous leukemia lines em-2 and em-3. Cancer genetics and cytogenetics, 25(2):271–284, 1987.

[379] D. Seigneurin, P. Champelovier, G. Mouchiroud, R. Berthier, D. Leroux, M. Prenant, J. McGregor, J. Starck, F. Morle, and C. Micouin. Human chronic myeloid leukemic cell line with positive philadelphia chromosome exhibits megakaryocytic and erythroid characteristics. Experimental hematology, 15(8):822, 1987.

[380] M. Okabe, Y. Kunieda, S. Nakane, M. Kurosawa, T. Itaya, W.R. Vogler, M. Shoju, and T. Miyazaki. Establishment and characterization of a new ph1-positive chronic myeloid leukemia cell line mc3 with trilineage phenotype and an altered p53 gene. Leukemia & lymphoma, 16(5-6):493–503, 1995.

[381] M. Ogura, Y. Morishima, R. Ohno, Y. Kato, N. Hirabayashi, H. Nagura, and H. Saito. Establishment of a novel human megakaryoblastic leukemia cell line, meg-01, with positive philadelphia chromosome. Blood, 66(6):1384–1392, 1985.

[382] G. Brady, F. Billia, J. Knox, T. Hoang, I.R. Kirsch, E.B. Voura, R.G. Hawley, R. Cum- ming, M. Buchwald, K. Siminovitch, et al. Analysis of gene expression in a complex diﬀerentiation hierarchy by global ampliﬁcation of cdna from single cells. Current Biol- ogy, 5(8):909–922, 1995.

[383] F. Billia, M. Barbara, J. McEwen, M. Trevisan, and N.N. Iscove. Resolution of pluripo- tential intermediates in murine hematopoietic differentiation by global complementary dna amplification from single cells: confirmation of assignments by expression profiling of cytokine receptor transcripts. Blood, 97(8):2257–2268, 2001.

[384] C.G. Mullighan, C.B. Miller, I. Radtke, L.A. Phillips, J. Dalton, J. Ma, D. White, T.P. Hughes, M.M. Le Beau, C.H. Pui, et al. Bcr–abl1 lymphoblastic leukaemia is characterized by the deletion of ikaros. Nature, 453(7191):110–114, 2008.

[385] V.S. Nadarajan, C.L. Phan, C.H. Ang, K.L. Liang, G.G. Gan, P.C. Bee, and Z. Zakaria. Identiﬁcation of copy number alterations by array comparative genomic hybridization in patients with late chronic or accelerated phase chronic myeloid leukemia treated with imatinib mesylate. International journal of hematology, 93(4):465–473, 2011.

[386] S.H. Friend, RA Bernards, S. Rogelj, R.A. Weinberg, J.M. Rapaport, D.M. Albert, and T.P. Dryja. Identiﬁcation of a human dna segment having properties of the gene that predisposes to retinoblastoma and osteosarcoma. Nature, 323(6089):643–646, 1986.

[387] A.D. Goddard, H. Balakier, M. Canton, J. Dunn, J. Squire, E. Reyes, A. Becker, RA Phillips, and BL Gallie. Infrequent genomic rearrangement and normal expression of the putative rb1 gene in retinoblastoma tumors. Molecular and cellular biology, 8(5):2082– 2088, 1988. References 169

[388] K. Mihara, X.R. Cao, A. Yen, S. Chandler, B. Driscoll, AL Murphree, A. T’ang, and YK Fung. Cell cycle-dependent regulation of phosphorylation of the human retinoblastoma gene product. Science, 246(4935):1300, 1989.

[389] PA Hamel, RA Phillips, M. Muncaster, and BL Gallie. Speculations on the roles of rb1 in tissue-speciﬁc diﬀerentiation, tumor initiation, and tumor progression. The FASEB journal, 7(10):846–854, 1993.

[390] T.P. Hopp, K.S. Prickett, V.L. Price, R.T. Libby, C.J. March, D.P. Cerretti, D.L. Urdal, and P.J. Conlon. A short polypeptide marker sequence useful for recombinant protein identiﬁcation and puriﬁcation. Biotechnology, 6(1205):5, 1988.

[391] A. Einhauer and A. Jungbauer. The ﬂag peptide, a versatile fusion tag for the puriﬁcation of recombinant proteins. Journal of biochemical and biophysical methods, 49(1):455–465, 2001.

[392] A.G. Reid, P.S. Tarpey, and E.P. Nacheva. High-resolution analysis of acquired genomic imbalances in bone marrow samples from chronic myeloid leukemia patients by use of multiple short dna probes. Genes, Chromosomes and Cancer, 37(3):282–290, 2003.

[393] G.W. Dewald, W.A. Wyatt, and R.T. Silver. Atypical bcr and abl d-ﬁsh patterns in chronic myeloid leukemia and their possible role in therapy. Leukemia & lymphoma, 34(5-6):481–491, 1999.

[394] N. Cambier, R. Chopra, A. Strasser, D. Metcalf, A.G. Elefanty, et al. Bcr-abl activates pathways mediating cytokine independence and protection against apoptosis in murine hematopoietic cells in a dose-dependent manner. Oncogene, 16(3):335, 1998.

[395] A. Gaiger, T. Henn, E. Horth, K. Geissler, G. Mitterbauer, T. Maier-Dobersberger, H. Greinix, C. Mannhalter, OA Haas, and K. Lechner. Increase of bcr-abl chimeric mrna expression in tumor cells of patients with chronic myeloid leukemia precedes disease progression. Blood, 86(6):2371–2378, 1995.

[396] M. Towatari, K. Adachi, H. Kato, and H. Saito. Absence of the human retinoblastoma gene product in the megakaryoblastic crisis of chronic myelogenous leukemia. Blood, 78(9):2178–2181, 1991.

[397] G. Gaidano, A. Serra, A. Guerrasio, G. Rege-Cambrin, U. Mazza, and G. Saglio. Genetic analysis of p53 and rb1 tumor-suppressor genes in blast crisis of chronic myeloid leukemia. Annals of hematology, 68(1):3–7, 1994.

[398]Z. A.N.´ Beck, A. Kiss, F.D. Tóth,J. Szabó,A. Bácsi,E.E.B.´ Balogh, A.´ Borbély, B. Telek, E. Kovács, E.´ Oláh,et al. Alterations of p53 and rb genes and the evolution of the accelerated phase of chronic myeloid leukemia. Leukemia & Lymphoma, 38(5-6):587–597, 2000.

[399] C.H.M. Jamieson, L.E. Ailles, S.J. Dylla, M. Muijtjens, C. Jones, J.L. Zehnder, J. Gotlib, K. Li, M.G. Manz, A. Keating, et al. Granulocyte–macrophage progenitors as candidate leukemic stem cells in blast-crisis cml. New England Journal of Medicine, 351(7):657–667, 2004. References 170

[400] Y. Minami, S.A. Stuart, T. Ikawa, Y. Jiang, A. Banno, I.C. Hunton, D.J. Young, T. Naoe, C. Murre, C.H.M. Jamieson, et al. Bcr-abl-transformed gmp as myeloid leukemic stem cells. Proceedings of the National Academy of Sciences, 105(46):17967, 2008.

[401] G. Stoecklin, S. Hahn, and C. Moroni. Functional hierarchy of auuua motifs in mediating rapid interleukin-3 mrna decay. Journal of Biological Chemistry, 269(46):28591–28597, 1994.

[402] H.D. Preisler, A.J. Kinniburgh, G. Wei-Dong, and S. Khan. Expression of the protoonco- genes c-myc, c-fos, and c-fms in acute myelocytic leukemia at diagnosis and in remission. Cancer research, 47(3):874, 1987.

[403] B.A. Jennings and K.I. Mills. c-myc locus ampliﬁcation and the acquisition of trisomy 8 in the evolution of chronic myeloid leukaemia. Leukemia research, 22(10):899–903, 1998.

[404] X. Jiang, A. Lopez, T. Holyoake, A. Eaves, and C. Eaves. Autocrine production and action of il-3 and granulocyte colony-stimulating factor in chronic myeloid leukemia. Proceedings of the National Academy of Sciences, 96(22):12804, 1999.

[405] M.J. Mauro. Deﬁning and managing imatinib resistance. ASH Education Program Book, 2006(1):219–225, 2006.

[406] C. Walz, W. Ahmed, K. Lazarides, M. Betancur, N. Patel, L. Hennighausen, V.M. Za- leskas, and R.A. Van Etten. Essential role for stat5a/b in myeloproliferative neoplasms induced by bcr-abl1 and jak2 v617f in mice. Blood, 2012.

[407] M.R. Marit, M. Chohan, N. Matthew, K. Huang, D.A. Kuntz, D.R. Rose, and D.L. Barber. Random mutagenesis reveals residues of jak2 critical in evading inhibition by a tyrosine kinase inhibitor. PLoS ONE, 7(8):e43437, 2012.

[408] N. Bhagwat, P. Koppikar, O. Kilpivaara, et al. Heterodimeric jak-stat activation as a mechanism of persistence to jak2 inhibitor therapy. In Blood (ASH Annual Meeting Abstracts), number 122 in 118, 2011.

[409] A. Teﬀeri. Jak inhibitors for myeloproliferative neoplasms: clarifying facts from myths. Blood, 2012.

[410] A. Teﬀeri. Novel mutations and their functional and clinical relevance in myeloproliferative neoplasms: Jak2, mpl, tet2, asxl1, cbl, idh and ikzf1. Leukemia, 24(6):1128–1138, 2010.

[411] A. Tefferi, TL Lasho, O. Abdel-Wahab, P. Guglielmelli, J. Patel, D. Caramazza, L. Pieri, CM Finke, O. Kilpivaara, M. Wadleigh, et al. Idh1 and idh2 mutation studies in 1473 patients with chronic-, fibrotic-or blast-phase essential thrombocythemia, polycythemia vera or myelofibrosis. Leukemia, 24(7):1302–1309, 2010.

[412] T. Barbui, G. Barosi, G. Birgegard, F. Cervantes, G. Finazzi, M. Griesshammer, C. Har- rison, H.C. Hasselbalch, R. Hehlmann, R. Hoﬀman, et al. Philadelphia-negative classical myeloproliferative neoplasms: critical concepts and management recommendations from european leukemianet. Journal of Clinical Oncology, 29(6):761–770, 2011. References 171

[413] F. Cervantes, B. Dupriez, A. Pereira, F. Passamonti, J.T. Reilly, E. Morra, A.M. Van- nucchi, R.A. Mesa, J.L. Demory, G. Barosi, et al. New prognostic scoring system for primary myeloﬁbrosis based on a study of the international working group for myeloﬁ- brosis research and treatment. Blood, 113(13):2895–2901, 2009.

[414] R. Capdeville, E. Buchdunger, J. Zimmermann, A. Matter, et al. Glivec (sti571, imatinib), a rationally developed, targeted anticancer drug. Nature Reviews Drug Discovery, 1(7):493–502, 2002.

[415] L. Hartwell. Defects in a cell cycle checkpoint may be responsible for the genomic instability of cancer cells. Cell, 71(4):543, 1992.

[416] Z. Storchova and D. Pellman. From polyploidy to aneuploidy, genome instability and cancer. Nature Reviews Molecular Cell Biology, 5(1):45–54, 2004.

[417] N.J. Ganem, Z. Storchova, and D. Pellman. Tetraploidy, aneuploidy and cancer. Current opinion in genetics & development, 17(2):157–162, 2007.

[418] S.A Turley. Examining the role of prdm12 in 9q34-deleted cml. Master’s thesis, University of Toronto, Department of Medical Biophysics, 2006.

[419] A.C. Graham, D.L. Kiss, and E.D. Andrulis. Diﬀerential distribution of exosome subunits at the nuclear lamina and in cytoplasmic foci. Molecular biology of the cell, 17(3):1399– 1409, 2006.

[420] J.S.J. Anderson and R. Parker. The 3’ - 5’ degradation of yeast mrnas is a general mechanism for mrna turnover that requires the ski2 devh box protein and 3’ - 5’ exonucleases of the exosome complex. The EMBO journal, 17(5):1497–1506, 1998.

[421] P. Mitchell and D. Tollervey. An nmd pathway in yeast involving accelerated deadenyla- tion and exosome-mediated 3’ - 5’ degradation. Molecular cell, 11(5):1405–1413, 2003.

[422] S.C.R. Santos and S. Dias. Internal and external autocrine vegf/kdr loops regulate survival of subsets of acute leukemia through distinct signaling pathways. Blood, 103(10):3883– 3889, 2004.

[423] Y. Li, H. Nagai, T. Ohno, H. Ohashi, T. Murohara, H. Saito, and T. Kinoshita. Aberrant dna demethylation in promoter region and aberrant expression of mrna of pax4 gene in hematologic malignancies. Leukemia research, 30(12):1547–1553, 2006.

[424] N. Murphy, M. Ring, C. Heﬀron, B. King, AG Killalea, C. Hughes, CM Martin, E. McGuinness, O. Sheils, and JJ OLeary. p16ink4a, cdc6, and mcm5: predictive biomark- ers in cervical preinvasive neoplasia and cervical cancer. Journal of clinical pathology, 58(5):525–534, 2005.

[425] E. Lau, T. Tsuji, L. Guo, S.H. Lu, and W. Jiang. The role of pre-replicative complex (pre- rc) components in oncogenesis. The FASEB journal: oﬃcial publication of the Federation of American Societies for Experimental Biology, 21(14):3786, 2007.

[426] S. Gonzalez, P. Klatt, S. Delgado, E. Conde, F. Lopez-Rios, M. Sanchez-Cespedes, J. Mendez, F. Antequera, and M. Serrano. Oncogenic activity of cdc6 through repression of the ink4/arf locus. Nature, 440(7084):702–706, 2006. References 172

[427] P.C. Orban, D. Chui, and J.D. Marth. Tissue-and site-speciﬁc dna recombination in transgenic mice. Proceedings of the National Academy of Sciences, 89(15):6861, 1992.

[428] R. Kuhn, F. Schwenk, M. Aguet, and K. Rajewsky. Inducible gene targeting in mice. Science, 269(5229):1427–1429, 1995.

[429] H. Kantarjian, J.V. Melo, S. Tura, S. Giralt, and M. Talpaz. Chronic myelogenous leukemia: disease biology and current and future therapeutic strategies. ASH Education Program Book, 2000(1):90–109, 2000.