Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The influence of the immune system on acute lymphoblastic leukemia progression Fidanza, Mario 2017

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2017_september_fidanza_mario.pdf [ 7.2MB ]
Metadata
JSON: 24-1.0348676.json
JSON-LD: 24-1.0348676-ld.json
RDF/XML (Pretty): 24-1.0348676-rdf.xml
RDF/JSON: 24-1.0348676-rdf.json
Turtle: 24-1.0348676-turtle.txt
N-Triples: 24-1.0348676-rdf-ntriples.txt
Original Record: 24-1.0348676-source.json
Full Text
24-1.0348676-fulltext.txt
Citation
24-1.0348676.ris

Full Text

	  THE INFLUENCE OF THE IMMUNE SYSTEM ON ACUTE LYMPHOBLASTIC LEUKEMIA PROGRESSION by  Mario Fidanza  B.Sc., Simon Fraser University, 2012  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Experimental Medicine)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  June 2017  ©Mario Fidanza, 2017  	   ii Abstract  The immune system has been proposed to have an impact on the etiology of B-cell ALL; however, a mechanistic explanation of this influence remains elusive.  Epidemiological studies have uncovered a paradox, such that surrogates of infection exposure have consistently been associated with reduced ALL risk, while documented infections have a more variable influence, with both positive and negative risk associations being uncovered.  Despite these contradictory findings, timing of infection exposure has been identified as a critical variable.  A mechanistic explanation for the variable influences of infection or the importance of timing remain poorly defined.  In this study, I use the Eµ-RET and E2A-PBX1 mouse models to assess how the basal and stimulated immune system is capable of influencing disease progression.   My results indicate that in the absence of infection, resting IFN-γ is capable of influencing disease progression in the Eμ-RET mouse.  By modulating SOCS-1 expression levels, IFN-γ restricted the IL-7-driven proliferation of leukemia-initiating cells (LICs) and caused a significant delay in disease progression.  Furthermore, TLR ligand-mediated immune modulation inhibited disease progression by inducing the depletion of LICs through a mechanism that involved the direct activity of type-1 and type-2 interferon.  Importantly, the relevance of this mechanism in humans was validated through the use of both human immune-effector and leukemia cells.  Finally, I demonstrate that quantitative and qualitative differences in neonatal immune responses, in particular the increased capacity for IL-17A production by γδ T-cells, confer significant infection-induced protection from ALL progression in neonatal mice. 	   iii Mechanistically these findings represent several firsts in the investigation of how the immune system influences B-cell ALL progression.  They identify both type-1 and type-2 IFN and IL-17A as important inhibitory factors, and demonstrate the significant impact of immune modulation during the pre-leukemic phase on subsequent disease risk and progression.  Furthermore, the results presented here provide a mechanistic explanation for the importance of timing in the association between infection exposure and ALL risk.  Collectively, my results indicate that in addition to the educational influence suggested by the “delayed infection” hypothesis, early-life infection exposure may also have an active inhibitory impact on the progression of B-cell ALL. 	  	  	  	  	  	  	  	  	  	  	  	  	  	  	   iv Lay summary 	  	  Leukemia, which is cancer of the blood, is the most common childhood cancer.  There is evidence that infection exposure can influence the development of leukemia; for example, early-life daycare attendance, which is thought to increase infection exposure, is associated with a reduced risk of developing leukemia.  However, despite this, and other similar associations being recognized, the mechanisms through which infection exposure and immune responses impact leukemia progression remain largely unknown.  In this thesis, I assess how the immune system, in both its resting and activated state, is capable of influencing the development of leukemia.  The work described here outlines several previously unrecognized ways that the immune system is capable of impacting leukemia development.  Furthermore, it underscores the critical nature of several variables such as, timing of exposure and functionality of the host immune system in determining the impact of infections on leukemia risk and development.         	   v Preface 	  	   The original hypothesis of this thesis, that early-life immune modulation has the capacity to directly inhibit B-cell ALL progression, was formulated by Dr. Gregor Reid.  I expanded and formulated the final hypotheses with the assistance of Dr. Reid.  A version of Chapter 3 has been published: M. Fidanza, A.E. Seif, S. Jo, A. Kariminia, N. Rolf, L.M. Sly, S.A. Grupp and G.S.D. Reid.  IFN-γ directly inhibits murine B-cell precursor leukemia-initiating cell proliferation early in life. 2017. Eur. J. Immunol.  Dr. Reid and I formulated the hypothesis and designed the experiments.  I performed the experiments.  Dr. Reid and I conducted the data analysis and wrote the manuscript.  Dr. Seif assisted with statistical analyses.  A version of Chapter 4 has been published: M. Fidanza, A.E. Seif, A. DeMicco, N. Rolf, S. Jo, B. Yin, Y. Li, D.M. Barrett, J. Duque-Afonso, M.L. Cleary, C.H. Bassing, S.A. Grupp and G.S.D. Reid.  Inhibition of precursor B-cell malignancy progression by toll-like receptor ligand-induced immune responses. 2016. Leukemia. 30(10): 2116-2119.  Dr. Reid and I formulated the hypothesis and designed the experiments.  Dr. Seif, Dr. Rolf and I performed the experiments.  Dr. Reid and I conducted the data analysis and wrote the manuscript. Dr. Seif assisted with statistical analyses  For chapter 5, Dr. Reid and I formulated the hypothesis and designed the experiments.  I performed the experiments.  Dr. Reid and I conducted the data analysis and wrote the manuscript.  Sheka Aloyouni and Dr. Tobias Kollmann provided the attenuated L.monocytogenes. Sehyun Cho and Dr. Soren Gantt provided the murine-gammaherpesvirus-68 and Karen Simmons and Dr. Soren Gantt provided the murine-cytomegalovirus.   	   vi Table of contents Abstract .............................................................................................................................. ii Lay summary .................................................................................................................... iv Preface ................................................................................................................................ v Table of contents...............................................................................................................vi List of figures .................................................................................................................... xi List of abbreviations ...................................................................................................... xiv Acknowledgements ........................................................................................................ xix Chapter 1: Introduction ................................................................................................... 1 1.1 Acute lymphoblastic leukemia: clinical perspectives .................................................... 1 1.1.1 Overview ..................................................................................................................... 1 1.1.2 Risk stratification in B-cell ALL ................................................................................. 2 1.1.4 Treatment of childhood B-cell ALL ........................................................................... 4 1.1.5 Clinical challenges of relapse ...................................................................................... 6 1.2 Leukemogenesis in childhood B-cell ALL ...................................................................... 7 1.2.1 B-cell development ..................................................................................................... 7 1.2.2 Genetic sub-types of B-cell ALL .............................................................................. 10 1.2.3 Clonal evolution ........................................................................................................ 15 1.2.4 In utero origin of B-cell ALL .................................................................................... 16 1.2.5 Clonal evolution of relapse ....................................................................................... 20 1.3 Etiology of childhood B-cell ALL ................................................................................. 23 1.3.1 Overview ................................................................................................................... 23 1.3.2 The population mixing hypothesis ............................................................................ 25 	   vii 1.3.3 The delayed infection hypothesis .............................................................................. 26 1.3.4 Epidemiological support for the delayed infection hypothesis ................................. 27 1.3.5 Experimental support for the delayed infection hypothesis ...................................... 30 1.3.6 Additional hypotheses ............................................................................................... 31 1.3.7 Experimental models of ALL .................................................................................... 34 1.4 Immune ontogeny ........................................................................................................... 37 1.4.1 Overview ................................................................................................................... 37 1.4.2 Soluble factors ........................................................................................................... 38 1.4.3 Toll-like receptors ..................................................................................................... 39 1.4.4 T-Helper immune responses ..................................................................................... 40 1.4.5 Anti-viral responses .................................................................................................. 45 1.4.6 Modeling immune responses to common early life infections ................................. 46 1.5 Hypotheses ...................................................................................................................... 49 Chapter 2: Materials and methods ................................................................................ 52 2.1 Mice .............................................................................................................................. 52 2.2 Cells and cell culture .................................................................................................... 55 2.3 LIC burden assessment ................................................................................................ 55 2.4 LIC flow sorting ........................................................................................................... 56 2.5 Survival studies ............................................................................................................ 56 2.6 In vivo proliferation assay ............................................................................................ 57 2.7 Suppressor of Cytokine Signaling 1 (SOCS1) expression analysis ............................. 57 2.8 IFN-γ sensitivity assay ................................................................................................. 58 2.9 LIC adoptive transfer into NSG mice .......................................................................... 58 2.10 In vivo IFN-γ sensitivity ............................................................................................. 59 2.11 IL-7 dependence assay ............................................................................................... 59 	   viii 2.12 SOCS-1 RNA induction assay ................................................................................... 60 2.13 Direct effect of TLR ligands on LICs ........................................................................ 60 2.14 The contribution of immune-mediated effects of TLR ligands on LICs ................... 60 2.15 Myeloid effector cell assay ........................................................................................ 61 2.16 Normal and abnormal B-cell precursor assay ............................................................ 61 2.17 Conditioned supernatant assay ................................................................................... 62 2.18 Human leukemia cell assay ........................................................................................ 62 2.19 In vivo TLR ligand administration ............................................................................. 63 2.20 Infections .................................................................................................................... 63 2.21 Infection responses in Eµ-RET mice ......................................................................... 64 2.22 Adoptive transfer infection studies ............................................................................ 64 2.23 Antibody-based cytokine blockade and cell depletion .............................................. 65 2.24 Cytokine analysis ....................................................................................................... 65 2.25 Statistical methods ..................................................................................................... 66 2.26 Services ...................................................................................................................... 66 Chapter 3: Basal IFN-γ production delays disease onset in the Eµ-RET mouse ...... 67 3.1 Overview and rationale ................................................................................................ 67 3.2 IFN-γ reduces leukemia-initiating cell burden and delays disease onset ..................... 69 3.3 Normal B-cell precursors are also affected by basal IFN-γ ......................................... 73 3.4 IFN-γ induces SOCS-1 expression ............................................................................... 74 3.5 LIC expansion is directly inhibited by IFN-γ .............................................................. 76 3.6 IFN-γ-mediated inhibitory activity is restricted to the pre-leukemic phase ................. 79 3.7 Mechanisms of IFN-γ insensitivity in leukemic cells .................................................. 82 3.8 Discussion .................................................................................................................... 85 	   ix Chapter 4: Inhibition of precursor B-cell malignancy progression by toll-like receptor mediated immune responses ........................................................................... 92 4.1 Overview and rationale ................................................................................................ 92 4.2 Stimulation with specific TLR agonists results in LIC depletion in vitro ................... 94 4.3 The indirect effects of TLR agonists on LICs .............................................................. 95 4.4 The indirect effect of TLR activation on LICs is mediated through a lymphocyte-independent mechanism ......................................................................................................... 97 4.5 TLR ligand-induced LIC depletion is mediated by type-1 and type-2 interferon ........ 99 4.6 The inhibitory effect of IFN-γ and IFN-α is mediated through a direct mechanism ........ 101 4.7 TLR mediated inhibition may be relevant in humans ................................................ 103 4.8 The TLR mediated inhibitory mechanism is relevant in vivo .................................... 105 4.9 Discussion .................................................................................................................. 109 Chapter 5: The IL-12/IL-23 immune response pathway determines the outcome of infection on acute lymphoblastic leukemia progression ............................................ 114 5.1 Overview and rationale .............................................................................................. 114 5.2 Neonatal administration of Lm significantly alters disease progression .................... 116 5.3 Lm-induced LIC depletion requires IL-23 and IL-17A .............................................. 125 5.4 Production of IL-17A by γδ TCR+ T-cells is required for Lm-induced LIC depletion .... 132 5.5 Infection-dependent LIC depletion is not specific to Lm ........................................... 139 5.6 Disruption of the IL-12/IL-23 signaling axis during infection induces LIC expansion ... 142 5.7 Discussion .................................................................................................................. 148 Chapter 6: General discussion and perspectives........................................................ 155 6.1 The current model of infectious ALL etiology .......................................................... 155 6.2 The impact of the resting immune system on B-cell ALL ......................................... 156 6.3 The impact of immune modulation on B-cell ALL ................................................... 158 	   x 6.4 Timing is critical ........................................................................................................ 160 6.5 The generation of pro-leukemic infection responses ................................................. 162 6.6 Closing perspectives ................................................................................................... 163 REFERENCES .............................................................................................................. 166               	  	  	  	  	  	  	   xi List of figures  Figure 1.1.  Distribution of genetic subtypes in childhood B-cell ALL...........................14 Figure 1.2.  In utero initiation of childhood B-cell ALL..................................................19 Figure 1.3.  The clonal origins of relapse..........................................................................23 Figure 2.1.  Flow cytometric analysis of LICs in the Eµ-RET mouse model...................54 Figure 2.2.  Flow cytometric analysis of E2A-PBX1 leukemia cell burden.....................54 Figure 3.1.  LIC burden is higher in IFNγko Eµ-RET mice..............................................71 Figure 3.2.  IFN-γ restricts LIC proliferation in vivo.........................................................72 Figure 3.3.  IFN-γ influences disease kinetics in the Eµ-RET mouse...............................73 Figure 3.4.  Normal B-cell precursors are also affected by basal IFN-γ...........................74 Figure 3.5.  Basal IFN-γ induces higher resting expression of SOCS-1 in LICs..............76 Figure 3.6.  Wild type LICs are less sensitive to exogenous IFN-γ...................................78 Figure 3.7.  IFN-γ causes durable changes in SOCS-1 expression and proliferation........79 Figure 3.8.  In vivo IFN-γ did not impact leukemic cell outgrowth...................................81 Figure 3.9.  In vivo IFN-γ significantly inhibits LIC outgrowth........................................82 Figure 3.10.  Mechanisms of reduced IFN-γ sensitivity in leukemic cells........................84 Figure 3.11.  Alternative mechanisms of reduced IFN-γ sensitivity in leukemic cells.....85 Figure 4.1.  Direct effect of TLR ligation on LICs in vitro...............................................95 Figure 4.2.  The indirect effects of TLR activation on LIC survival.................................97 Figure 4.3.  TLR ligand induced LIC depletion is mediated by myeloid effector  cells....................................................................................................................................99 Figure 4.4.  TLR-mediated LIC depletion is dependent upon IFN-γ and IFN-α.............101 	   xii Figure 4.5.  The inhibitory effect of IFN-γ and IFN-α is mediated through direct and overlapping mechanisms..................................................................................................103 Figure 4.6.  TLR-mediated inhibition may be relevant in human B-cell ALL................105 Figure 4.7.  CpG administration induced significant depletion of LICs..........................107 Figure 4.8.  CpG induced LIC depletion is IFN-γ dependent..........................................108 Figure 4.9.  CpG induced immune responses significantly delay disease onset..............109 Figure 5.1.  Neonatal but not adult Lm infection induces significant LIC depletion.......120 Figure 5.2.  Neonatal Lm infection significantly delays disease onset in Eµ-RET  mice..................................................................................................................................121 Figure 5.3.  Differential responses to Lm are not related to inoculum size.....................122 Figure 5.4.  Adult mouse-derived 289 cells are responsive to Lm induced inhibitory mechanisms......................................................................................................................123 Figure 5.5.  Neonatal but not adult Lm infection induces depletion of E2A-PBX1+ leukemia cells...................................................................................................................124 Figure 5.6.  Neonatal Lm alters disease progression in an E2A-PBX1 leukemia adoptive      transfer model..................................................................................................................125 Figure 5.7.  Stat-4 signaling is required for Lm induced LIC depletion..........................128 Figure 5.8.  IL-23 is required for Lm-induced LIC depletion..........................................129 Figure 5.9.  Lm infection induces IL-17A production in neonatal mice..........................130 Figure 5.10.  IL-17A signaling is required for Lm-induced LIC depletion.....................131 Figure 5.11.  Lymphocytes are required for Lm-induced depletion of LICs...................135 Figure 5.12.  Lm-induced LIC depletion is not mediated by CD4+ or CD8+ T-cells.....136 Figure 5.13.  NKT-cells are not involved in the Lm-induced depletion of LICs............137 	   xiii Figure 5.14.  γδ T-cells are required for infection-induced depletion of LICs................138 Figure 5.15.  γδ T-cells are the significant producers of IL-17A during the immune response to Lm.................................................................................................................139 Figure 5.16.  Neonatal but not adult MCMV infection induces LIC depletion...............141 Figure 5.17.  Neonatal but not adult MHV-68 infection induces LIC depletion.............142 Figure 5.18.  Neutralization of basal IL-12/IL-23p40 has no effect on LIC burden.......144 Figure 5.19.  Dysfunctional immune responses to Lm induce LIC expansion................145 Figure 5.20.  Dysfunctional immune responses to MCMV induce LIC expansion.........146 Figure 5.21.  Dysfunctional immune responses to MHV-68 do not induce LIC                              expansion.........................................................................................................................147 Figure 5.22.  Dysfunctional adult immune responses induce LIC expansion.................148            	   xiv List of abbreviations  ABL1  Abelson Murine Leukemia Viral Oncogene Homolog 1 ACTH  Adrenocorticotropic Hormone AID  Activation-induced Cytidine Deaminase ALL  Acute Lymphoblastic Leukemia ATP  Adenosine Tri-Phosphate B220  Protein Tyrosine Phosphatase Receptor Type C BCP  B-cell Precursors BCR  B-cell Receptor BM  Bone Marrow BP-1  Glutamyl Aminopeptidase cAMP  Cyclic Adenosine Mono-Phosphate CAR  Chimeric Antigen Receptor CD-4  Cluster of Differentiation 4 CD-8  Cluster of Differentiation 8 CD-19  Cluster of Differentiation-19 (B Lymphocyte Antigen) CD-24  Cluster of Differentiation-24 CD28  Cluster of Differentiation-28 CD43  Cluster of Differentiation-43 CD80  Cluster of Differentiation-80 CD86  Cluster of Differentiation-86 CFU  Colony Forming Units CLPs  Common Lymphoid Progenitors  	   xv CMV  Cytomegalovirus CNS  Central Nervous System CRE  Cre-Recombinase DNA  Deoxyribonucleic Acid E2A  E2A Immunoglobulin Enhancer-binding Factors E12/E47 EBF  Early B-cell Factor EBV  Epstein-Barr Virus EFS  Event Free Survival ETV6  Transcription Factor ETV6 FACS  Fluorescence Activated Cell Sorting FBS  Fetal Bovine Serum FcR  Fc Receptor GCs  Glucocorticoids HH  High-hyperdiploidy HSCs  Hematopoietic Stem Cells HSPs  Heat Shock Proteins IFN-α  Interferon-alpha IFN-γ  Interferon-gamma Ig  Immunoglobulin IL-1β  Interleukin-1-beta IL-6  Interleukin-6 IL-7  Interleukin-7 IL-10  Interleukin-10 	   xvi IL-17  Interleukin-17 IL-12  Interleukin-12 IL-23  Interleukin-23 IP  Intraperitoneal IRF-3  Interferon Regulatory Factor 3 IRF-7  Interferon Regulatory Factor 7 IV  Intravenous JAK  Janus Kinase LIC  Leukemia Initiating Cells Lm  Listeria monocytogenes MAPK  Mitogen-activated Protein Kinase MB-1  B-cell Antigen Receptor Complex-associated protein alpha chain mCMV Murine Cytomegalovirus MHC  Major Histocompatibility Complex MHV-68 Murine-gammaherpesvirus-68 MLL-4 Mixed Lineage Leukemia-4 MRD  Minimum Residual Disease MX-1  Interferon-induced GTP-binding Protein MX-1 MYC  V-Myc Avian Myelocytomatosis Viral Oncogene Homolog MyD88 Myeloid Differentiation Primary Response Protein 88 NEAA  Non-Essential Amino Acids NF-κB  Nuclear Factor kappa-light-chain-enhancer of Activated B-cells NK  Natural-Killer Cells 	   xvii NKT  Natural-Killer T-cells NSG  NOD-scid/IL-2Rγ null PAMPs Pathogen Associated Molecular Patterns Pax5  B-cell Specific Activator Protein PBMC  Peripheral Blood Mononuclear Cell PBS  Phosphate Buffered Saline PBX-1  Pre-B-cell Leukemia Transcription Factor 1 PCR  Polymerase Chain Reaction PFU  Plaque Forming Units PRR  Pattern Recognition Receptor q-RTPCR Quantitative Reverse Transcription Polymerase Chain Reaction RAG  Recombination Activating Gene RBC  Red Blood Cell RET  Rearranged During Transfection RNA  Ribonucleic Acid ROS  Reactive Oxygen Species RUNX1 Runt Related Transcription Factor 1 SPF  Specific Pathogen Free SPL  Spleen STAT  Signal Transducer and Activator of Transcription TAC  Tris Ammonium Chloride TCR  T-cell Receptor  TGFβ  Transforming Growth Factor Beta 	   xviii Th-1  Type-1 T Helper Cell Th-2  Type-2 T Helper Cell Th-17  Type-17 T Helper Cell TIR  Toll/IL-1R TLR  Toll-like Receptor TNF-α  Tumor Necrosis Factor Alpha TRIF  TIR-domain-containing Adaptor-inducing Interferon-beta WBC  White Blood Cell γδ T-cell Gamma-Delta T-cell                	   xix Acknowledgements  I owe significant thanks to Dr. Gregor Reid, who supported my research over the last five years.  He has provided tremendous support and valuable mentorship while demonstrating nearly endless patience. I thank my supervisory committee members: Dr. Tobias Kollmann, Dr. Laura Sly and Dr. Christopher Maxwell, who have spent their valuable time attending my committee meetings, offering their advice and guidance and evaluating my progress. I offer my gratitude to my fellow students in the UBC Experimental Medicine program.  I also owe a tremendous debt to Dr. Nina Rolf, Dr. Arnawaz Bashir and Dr. Abbas Fotovati who were always willing to provide advice and assistance. Thanks to the Michael Cuccione Foundation, who provided a year of graduate studentship support. Thanks to the BC Children’s Hospital Research Institute, who provided two years of graduate studentship support. I thank my girlfriend, Annie DeBrincat and my brother, Nicolas Fidanza, for always lending a supportive ear. I owe a tremendous debt of gratitude to my parents, Mariano and Pia Fidanza, who have supported me through my many years of education, without them this wouldn’t have been possible. 	   1 Chapter 1: Introduction 	  1.1 Acute lymphoblastic leukemia: clinical perspectives 	  1.1.1 Overview 	  Acute lymphoblastic leukemia (ALL) is a malignant disorder that is characterized by the uncontrolled proliferation of clonal lymphoid progenitor cells that exhibit arrested maturation (Bhojwani et al. 2015; Cortes & Kantarjian 1995; Longo et al. 2015; Pui et al. 2008).  ALL is the most common malignancy in children and represents the second leading cause of cancer related death in people below the age of twenty (Linabery & Ross 2008; Smith et al. 2010) .  Common symptoms associated with ALL include frequent bruising and/or bleeding, anemia-induced fatigue, bone and joint pain, and frequent infections that are the result of blast accumulation compromising normal bone-marrow function (Bernbeck et al. 2009; Rao et al. 2013).   ALL can be broadly sub-divided into B-cell ALL and T-cell ALL based on the lineage of the lymphoid progenitor cell affected.  T-cell ALL accounts for approximately 15% of childhood ALL cases, and 25% of adult ALL cases (Chiaretti & Foà 2009; Harrison 2011).  T-cell ALL is characteristically more common in males than females, and while primarily considered an adolescent disease, incidence of T-cell ALL is nearly uniformly distributed over the first twenty years of life (Van Vlierberghe & Ferrando 2012).  While a genetically heterogeneous disease, constitutive activation of NOTCH1 is the single most common oncogenic lesion in T-cell ALL, found in more than 60% of patients (Pear & Aster 2004; Van Vlierberghe & Ferrando 2012; Weng et al. 2004).  T-cell ALL has historically been associated with a poor prognosis; however, because of advancements in treatment strategies, including the introduction of intensified 	   2 chemotherapy, cure rates are approaching 75% in adolescents and 50% in adults with T-cell ALL (Van Vlierberghe & Ferrando 2012).  Despite these advancements however, the prognosis for patients with primary resistant or relapsed disease remains poor (Van Vlierberghe & Ferrando 2012).  Conversely, B-cell ALL accounts for approximately 80% of childhood ALL cases.  Unlike, T-cell ALL, it shows a marked incidence peak in children between the ages of two and five (Inaba et al. 2013; Longo et al. 2015; Pui et al. 2008).  Although significantly less frequent, B-cell ALL also occurs in both infants (under 1 year of age) and adults (Downing & Shannon 2002).  Infant ALL is considered to be a biologically distinct disease from childhood B-cell ALL and is associated with its own characteristic genetic lesions, most frequently MLL (mixed-lineage leukemia gene) rearrangements (Carroll et al. 2003; Sanjuan-pla et al. 2015).  Adult and childhood B-cell ALL share common genetic lesions albeit with different relative frequencies, a characteristic that is likely associated with the differential clinical outcomes for children and adults. (Faderl et al. 2010; Jabbour et al. 2005; Pui et al. 2008;).  While current cure rates for childhood B-cell ALL are approaching 90%, outcomes for patients with infant or adult ALL remain significantly worse (Longo et al. 2015; Pui et al. 2008).  This remainder of thesis will focus on childhood B-cell precursor (BCP) ALL. 1.1.2 Risk stratification in B-cell ALL 	   Comprehensive risk-stratification protocols have had a significant impact on the progressive improvements in childhood B-cell ALL cure rates (Faderl et al. 2010; Pui et al. 2008; Zhou et al. 2012).  In current treatment strategies, the intensity of first-line treatment regimens is stratified based on prognostic indicators and is aimed at tailoring 	   3 treatments to improve cure-rates while limiting treatment related toxicity and morbidity (Vrooman & Silverman 2016; Yeoh et al. 2002).  Currently there are four defined risk groups for childhood ALL; very high risk with a five year event-free-survival (EFS) under 45%, high risk, standard risk and finally low risk, which is defined by a five year EFS over 85% (Schultz et al. 2007).    Risk stratification is based on independent prognostic indicators that are conclusively evaluated to determine risk parameters and therefore the appropriate treatment strategy.  In addition to underlying genetic subtype, there are several prognostic indicators that are evaluated during risk stratification.  Age at diagnosis is one of these indicators.  In general ALL occurring in children between one and ten years of age is associated with a better outcome than when occurring in children under one year old (infant ALL) or over the age of ten (Vrooman & Silverman 2016; Yeoh et al. 2002).  Another prognostic indicator is white blood cell (WBC) count at diagnosis.  Lower WBC counts, typically classified as being below 50,000/mm3, are associated with a better prognosis than high presenting WBC counts (Vrooman & Silverman 2016).   Response to treatment is gauged by remission status.  Remission can be defined either morphologically or molecularly.  Morphological remission is classified by the restoration of normal hematopoiesis and a blast count below 5% in the bone marrow as assessed by light microscopy (Foroni et al. 1999).  However, morphological assessments of the bone marrow are relatively insensitive, and this 5% threshold represents the low detection limit for this method.  Assessment of blast count using flow cytometry to detect aberrant immunophenotypes or polymerase-chain reaction to detect diagnostic Ig rearrangements are at least 100-fold more sensitive and are capable of detecting leukemic 	   4 blasts at a frequency of 0.01% (Campana 2010; Campana & Coustan-Smith 1999; Foroni et al. 1999). The absence of detectable blasts using these quantification methods qualifies as a molecular remission.  Patients who achieve morphological but not molecular remission are minimum residual disease (MRD) positive (Pui & Campana 2000). Several properties associated with a patient’s early response to initial treatments, in the context of remission status, are important prognostic variables (Panzer-Grumayer et al. 2000).  For example it has been demonstrated that patients who do not achieve morphological remission during the first month of treatment (induction failure) have a poor prognosis (Pieters & Carroll 2010; Pui & Evans 2006; Yeoh et al. 2002). Additionally, the presence of MRD is another useful measure of early treatment responses and is a strong prognostic marker of long-term disease outcome (Bassan et al. 2009; Campana 2010; Conter et al. 2010; Sawada et al. 2009).  MRD represents such a significant prognostic factor that treatment intensity is adjusted based on MRD evaluation at several different points along the treatment regimen (Campana 2010; Pieters & Carroll 2010; Vrooman & Silverman 2016).  Finally, there are also emerging indicators, such as absolute lymphocyte count after early stage treatment, which are now being recognized as valuable prognostic markers, despite a current lack of mechanistic insight (De Angulo et al. 2008). 1.1.4 Treatment of childhood B-cell ALL 	   Over the last forty years survival rates have increased dramatically, and overall survival for childhood B-cell ALL now approaches 90% (Pui & Evans 2006).  This improvement can be attributed to the combined effects of a number of different influences, such as the identification of new drugs and combinational chemotherapies, prophylactic treatment of sanctuary sights such as the central nervous system (CNS) and 	   5 more developed risk-stratification protocols (Hunger & Mullighan 2015; Inaba et al. 2013; Pui & Evans 2006).    Specific treatment protocols may vary from patient to patient based on prognostic markers and risk stratification; however in nearly all cases, the structure of the overall treatment regimen is similar and involves four key phases: induction, consolidation in concert with CNS guided therapy, re-induction, and maintenance (Cooper & Brown 2015; Pieters & Carroll 2010;).  The primary goal of the induction phase of treatment is to induce a complete (morphological and molecular) remission in the bone marrow.  The intensity of this phase is often guided by risk-stratification but is always characterized by the concurrent administration of at least three drugs, a glucocorticoid, vincristine and L-asparaginase (Cooper & Brown 2015; Pui & Evans 2006).  The incorporation of a fourth drug, an anthracycline, is often included for high-risk patients.  This phase of treatment typically lasts four to six weeks, and the first MRD evaluation occurs after the induction phase, and may impact the subsequent treatment regimen (Campana 2010; Cooper & Brown 2015).  Once remission is achieved, the consolidation phase of treatment begins.  This phase is designed to reduce systemic MRD burden and prevent sanctuary site relapse.  This phase often involves the intrathecal administration of drugs such as methotrexate or 6-mercaptopurine to prevent CNS relapse.  MRD status is usually assessed again at the end of the consolidation phase (Cooper & Brown 2015; Pieters & Carroll 2010).  The re-induction phase basically represents a repetition of the induction phase treatment, and is administered during the first few months of remission (Cooper & Brown 2015; Pui & Evans 2006).  Incorporation of this phase has been found to significantly reduce the risk of later relapse.  Finally, the maintenance phase of treatment 	   6 lasts for approximately two years and involves administration of lower doses of methotrexate and 6-mercaptopurine, sometimes in combination with other drugs (Rabin & Poplack 2011; Toyoda et al. 2000).  As males have a higher risk of relapse than females, they are often kept on maintenance therapy for a significantly longer duration (Rabin & Poplack 2011).   While these treatment regimens achieve cure in approximately 90% of all children with B-cell ALL, primary refractory and relapsed disease still represent significant clinical challenges and account for a large proportion of all cancer-related childhood deaths (Bhojwani et al. 2012; Hunger & Mullighan 2015).  However, new treatment strategies have had some success in these patients.  For example, in the case of BCR-ABL1+ ALL, which was previously associated with a poor prognosis, the use of targeted protein tyrosine kinase inhibitors such as imatinib and dasatinib have shown to be very effective (Fielding 2011).  Additionally, new immunotherapies, such as the bi-specific CD19/CD3 antibody blinatumomab and autologous chimeric-antigen receptor (CAR) T-cell therapy, have had some success as treatments for relapsed or refractory B-cell ALL (Grupp et al. 2013; Maude et al. 2015; Nagorsen et al. 2012; Topp et al. 2011). 1.1.5 Clinical challenges of relapse 	   Relapse occurs in 15-20% of B-cell ALL patients and cure rates after relapse are much lower, generally ranging between 15-50% (Hunger & Mullighan 2015).  Relapsed disease has its own set of prognostic markers that are associated with treatment outcome.  For example, time to relapse is a significant indicator, with a shorter duration being associated with a poorer outcome (Cooper & Brown 2015; Raetz & Bhatla 2012).  Additionally, immunophenotype and location of relapse are also strong prognostic 	   7 indicators (Bhojwani et al. 2015).  Allogeneic stem cell transplant is used as a therapy much more frequently for relapsed ALL, and is employed in nearly 50% of relapsed patients depending on specific suitability based on the assessment of several treatment response indicators, such as MRD status (Bhojwani et al. 2015; Locatelli et al. 2012). Leukemic cells at relapse are typically genetically and phenotypically distinct from those at first presentation.  It has been demonstrated that on average, leukemic cells at relapse harbour significantly more genetic alterations than those present at initial presentation (Goto 2015; Roberts et al. 2014;).  These mutations appear to preferentially occur in genes involved in B-cell development, and may be responsible for inducing a more “stem cell-like” state in relapse clones (Mullighan et al. 2011; Roberts et al. 2014).  Additionally, leukemia blasts at relapse typically display enhanced drug resistance relative to those at original diagnosis.  This is likely due to the emergence of drug resistant sub-clones that were either present in low frequency at diagnosis or that developed successively over the course of initial therapy (Goto 2015; Styczynski et al. 2007).  1.2 Leukemogenesis in childhood B-cell ALL 	  1.2.1 B-cell development 	   B-cells are a class of lymphocytes that arise from hematopoietic stem cells (HSCs) in the bone marrow through a process called hematopoiesis.  Once the development process is complete, differentiated B-cells exit the bone marrow and migrate to the lymph nodes where they compete for access to follicular dendritic cells in order to receive necessary pro-survival signals (Young et al. 1994).  There are multiple intermediate stages between hematopoietic stem cells and differentiated B-cells; the fate 	   8 of the cells at each of these stages is determined by a strictly coordinated interplay between numerous transcription factors and extrinsic cues provided by the bone marrow microenvironment (Burrows & Cooper 1997).    To begin this process HSCs differentiate into common lymphoid progenitors (CLPs) in which B-cell-specific commitment occurs.  Differentiation of CLPs into B-cell precursors is induced by expression of the transcription factors E2A and EBF (Busslinger et al. 2000; O’Riordan & Grosschedl 1999).  These transcription factors induce PAX5, which is required for the restriction of CLPs to the B-cell lineage (Nutt et al. 1999; Pridans et al. 2008; Revilla-I-Domingo et al. 2012).  The very early stage of B-cell development, referred to as pre-pro-B-cells, have not begun rearrangement of either the heavy or light chains of the BCR (Geier & Schlissel 2006).  IL-7 serves as an extrinsic signal that specifies B-cell fate and promotes the transition from CLP to pre-pro-B-cell.  IL-7 binds to the IL-7 receptor (IL-7R), which signals though the JAK-STAT signaling pathway to promote proliferation and survival of early B-cell progenitors.  IL-7 signaling is an absolute requirement for murine B-cell development, and is required for the survival and proliferation of several early B-cell precursor stages (Ceredig & Rolink 2012; Jiang et al. 2005).  The specific role of IL-7 in human B-cell development however, remains poorly understood.  Further differentiation of B-cells is triggered by the rearrangement of the Ig heavy-chain through a process known as V(D)J recombination.  This process is mediated by the recombination-activating genes (RAG1 and RAG2) which are responsible for creating double-stranded breaks in DNA segments and then re-joining segments by non-homologous end joining (Grawunder et al. 1998; Li et al. 1997; Mombaerts et al. 1992; 	   9 Schatz et al. 1989).  This process begins in early pro-B-cells when the DH and JH segments are recombined. This is followed by the joining of the VH segment to the rearranged DHJH segments to complete the production of the heavy chain variable region; this final rearrangement is performed during the late pro-B-cell stage (Schatz et al. 1989).  The recombined heavy chain then pairs with the surrogate light chain, which is composed of VpreB (VPREB1), λ5 (IGLL1) and the accessory signaling molecules Ig-α and Ig-β to form the pre-B-cell receptor (pre-BCR) (Geier & Schlissel 2006; Herzog et al. 2009).  Expression of the pre-BCR on the cell surface marks the transition from late pro-B to large pre-B stage of B-cell differentiation (Hardy & Hayakawa 2001; Herzog et al. 2009). Successful formation of the pre-BCR represents the major checkpoint in B-cell differentiation.  The large pre-B-cell stage is characterized by a down regulation of RAG1 and RAG2 expression and extensive IL-7-driven proliferation.  As this proliferative period commences, cells migrate through the bone marrow, IL-7R expression is down regulated, and RAG expression is re-induced (Hardy & Hayakawa 2001).  These changes mark the transition to the small pre-B-cell stage in which Ig light-chain (IgI) rearrangement begins.  Successful rearrangement of the IgI locus leads to expression of the complete B-cell receptor (BCR) on the surface of the cell and completes the process of B-cell differentiation in the bone marrow (Clark et al. 2014; Prak & Weigert 1995).  Successful formation of the pre-BCR marks the major checkpoint in B-cell development (Hendriks & Middendorp 2004).  The default fate of immature B-cells at the pro-B-cell stage is apoptosis; autonomous signaling through the pre-BCR, which indicates successful heavy chain rearrangement, rescues cells from this fate and facilitates 	   10 the proliferation and survival of B-cell precursors (Mårtensson et al. 2010).  In B-cell precursors in which heavy-chain rearrangement was unproductive, the pre-BCR will not be assembled and cells will undergo apoptosis in a process termed negative clonal selection (Mårtensson et al. 2010; Rickert 2013).  In virtually all cases, B-cell ALL exhibits maturational arrest at the late-pro-B-cell or pre-B-cell stage (Nahar & Müschen 2009).  This observation suggests that subversion of the pre-BCR checkpoint may be required for leukemogenesis.  This is typically accomplished through disruption of normal transcription-factor networks or through the production of aberrant tyrosine kinases that promote growth and proliferation (Buchner et al. 2015; Nahar & Müschen 2009).  While arrest at the pre-BCR checkpoint is nearly universal, B-cell ALL is a genetically heterogeneous disease, suggesting that there are multiple pathways to achieve checkpoint disruption (Nahar & Müschen 2009; Pang et al. 2014). 1.2.2 Genetic sub-types of B-cell ALL 	   Recurrent chromosomal abnormalities, primarily translocations and aneuploidy, are a hallmark of B-cell ALL (Armstrong & Look 2005; Pui et al. 2008; Teitell & Pandolfi 2009).  These chromosomal exchanges result in the illegitimate recombination of genes that can cause either the deregulation of oncogene expression, or the production of chimeric fusion-gene products with properties that are different than their wild type constituent proteins (Zelent et al. 2004).  Generally the result of these translocations and gene fusions is the production of constitutively active kinases or transcription factors with altered or impaired function (Holmfeldt et al. 2013; Pui & Evans 2006).  These alterations are present in both children and adults, but with differing frequencies in each age group.  The non-random genetic patterns in B-cell ALL allows for the separation of 	   11 leukemias into several cytogenetic subgroups that demonstrate distinct pathobiological features (Chan 2002; Onciu 2009).  In B-cell ALL, there are three commonly found balanced chromosomal translocations that generate fusion genes, namely ETV6-RUNX1 (TEL-AML1), E2A-PBX1, and BCR-ABL1.  ETV6-RUNX1 and E2A-PBX1 encode oncogenic transcription factors, while the BCR-ABL1 fusion gene encodes a constitutively active tyrosine kinase (Mullighan 2012; Pui et al. 2004).   The t(12;21)(p13;q22) translocation which results in the production of the ETV6-RUNX1 fusion gene is the single most commonly occurring translocation in B-cell ALL (Figure 1.1) (Armstrong & Look 2005; Jamil et al. 2000; Shurtleff et al. 1995).  While standard karyotyping is usually insufficient to detect this translocation, more advanced molecular techniques have indicated that the ETV6-RUNX1 translocation is present in upwards of 25% of B-cell ALL patients (Jamil et al. 2000).  The ETV6-RUNX1 fusion generally contains the helix-loop-helix domain of ETV6 fused to the DNA binding and transactivation domains of RUNX1 (Jamil et al. 2000; Morrow et al. 2007).  While the exact molecular mechanism of ETV6-RUNX1-driven leukemogenesis is not completely understood, both genes are critically involved in B-cell development: ETV6 is a potent transcriptional repressor that has a role in early lymphoid progenitor commitment (De Braekeleer et al. 2012; Lopez et al. 1999), while RUNX1 serves as the DNA binding subunit of a transcription factor referred to as core-binding factor, which serves a critical role in hematopoiesis (Pui et al. 2004).  The ETV6-RUNX1 fusion protein is thought to drive leukemogenesis by altering the self-renewal and differentiation of B-cell progenitors (Morrow et al. 2004; Morrow et al. 2007).  In general, leukemias carrying the 	   12 t(12;21)(p13;q22) translocation are associated with a very good prognosis (Holmfeldt et al. 2013; Maloney et al. 1999; Zelent et al. 2004; Zuna et al. 1999).  The t(1;19)(q23;p13) chromosomal translocation leads to the production of the E2A-PBX1 fusion gene.  This translocation is present in 5-15% of B-cell ALL cases and is associated with a favorable overall outcome, but a higher risk of CNS relapse (Jeha et al. 2009).  The E2A-PBX1 fusion protein promotes leukemogensis through the induction of aberrant expression of WNT16, which can lead to autocrine stimulation of the WNT/β-catenin signalling pathway in these cells (McWhirter et al. 1999).  Signaling through this pathway is thought promote survival and proliferation, circumventing the requirement for successful completion of the pre-BCR checkpoint (Nahar & Müschen 2009; Rickert 2013).  The t(9;22)(q34;q11) chromosomal translocation results in the production of the “Philadelphia” chromosome and expression of the constitutively active chimeric tyrosine-kinase BCR-ABL1 (Heisterkamp et al. 1985; Rowley 1973).  While being found in upwards of 50% of adult B-cell ALL cases, this translocation occurs in approximately 5% of childhood B-cell ALL cases (Mancini et al. 2005; Mullighan 2012).  This BCR-ABL1 fusion gene promotes leukemogenesis by mimicking the pro-survival signal conferred by autonomous pre-BCR signaling, primarily through the phosphorylation of BTK (Feldhahn et al. 2005; Klein et al. 2004).  Of these three geBnetic subtypes, the BCR-ABL1 fusion gene is associated with the worst outcome (Schultz et al. 2007).  The main cause of the unfavorable prognosis associated with this translocation is increased genomic instability induced by the aberrant expression of activation induced cytidine deaminase (AID) in this subtype (Feldhahn et al. 2007). 	   13  In addition to reciprocal chromosomal translocations, aneuploidy is also common in B-cell ALL.  Chromosomal gains occur most frequently, and high-hyperdiploidy (HH) is the most common genetic subtype of childhood B-cell ALL.  This subtype is defined by the presence of 51-67 chromosomes and accounts for approximately 30% of all childhood B-cell ALL cases (Ito et al. 1999; Mullighan 2012).  This subtype, like the t(12;21)(p13;q22) translocation is associated with a very good prognosis (Mullighan 2012; Paulsson & Johansson 2009).  The characteristic feature of HH ALL is chromosomal gains, most commonly trisomies and in some cases tetrasomies.  These duplications can affect any chromosome, but more than 70% of HH ALL cases feature gains in chromosomes X, 4, 6, 10, 14, 17, 18 or 21 (Paulsson et al. 2010).  The fact that most other B-cell leukemia subtypes are characterized by specific fusion genes or mutational patterns has caused some to speculate that hyperdiploidy is secondary to a cryptic genetic event, and that in itself does not have any leukemogenic effect (Onodera et al. 1992; Paulsson et al. 2015; Paulsson et al. 2010; Talamo et al. 2010).  However, significant molecular investigation has so far failed to identify any such primary genetic lesion (Paulsson & Johansson 2009).  Furthermore, HH cases were found to harbour very few mutations in coding sequences.  Two groups of mutations were identified as putative driver mutations, those occurring in genes involved in the RAS signalling pathway, and those occurring in histone modification genes (Paulsson et al. 2015).  However, the fact that these types of mutations are also found in other B-cell ALL subtypes, and that they are not ubiquitous in all hyperdiploid cases, suggests that it is unlikely that these mutations underlie the clinical features of this genetic subtype (Paulsson et al. 2015).  These findings indicate that aneuploidy is the only genetic event common to hyperdiploid 	   14 ALL and is in fact the primary driver event in these cases (Paulsson et al. 2015; Panzer-Grümayer et al. 2002).  The favourable prognosis for HH ALL is associated with high sensitivity to methotrexate treatment; however, aside from high intracellular accumulation of polyglutamates, the molecular rationale for this sensitivity is poorly understood (Chan 2002).  Figure 1.1.  Distribution of genetic subtypes in B-cell ALL Frequency of the major genetic subtypes of childhood B-cell ALL.  As indicated reoccurring chromosomal translocations account for approximately 40% of all cases.  Genetic subtype is commonly used as a prognostic marker during risk-stratification protocols.     BCR8ABL1)5%)ETV68RUNX1)25%)E2A8PBX1)10%)Hyperdiploid)29%)BCR8ABL1)Like)10%)Hypodiploid)1%)Other)20%)	   15 1.2.3 Clonal evolution 	   The development of cancer is an evolutionary process that is driven by the sequential accumulation and selection of mutations that promote the expansion of sub-clones that possess advantageous phenotypes (Nowell 1976).  In B-cell ALL, the phenotypic changes that provide a selective advantage are a block in differentiation and/or an increased capacity for proliferation or survival.  The chromosomal aberrations that are commonly associated with B-cell ALL are thought to represent initiating or “first hit” mutational events that are required, but insufficient, to drive leukemogenesis.  This concept is supported by modeling experiments in mice (Fischer et al. 2005; Tsuzuki et al. 2004) and human cells (Hong et al. 2008).  Cooperating mutations, often in genes involved in B-cell differentiation or cell-cycle control, are therefore also necessary for the development of fully malignant disease, suggesting that B-cell ALL may fit a minimum two-step model of pathogenesis (Greaves & Maley 2012; Zelent et al. 2004).  The model for the mutational development of B-cell ALL is, therefore, that chromosomal translocations or spontaneous aneuploidy occur as initiating events that are sufficient to spawn clonal pre-leukemic progeny (Greaves & Wiemels 2003).  These abnormal cells then independently incur spontaneous secondary or cooperating mutations and are subject to clonal selection.  This cell population remains clinically silent until accumulating mutations that facilitate overt transformation and malignant outgrowth that manifests in clinical symptoms.    	   16 1.2.4 In utero origin of B-cell ALL 	   Clinically evident symptoms of B-cell ALL manifest only a few weeks prior to diagnosis (Saha et al. 1993).  The development of leukemia, therefore, likely involves a clinically silent phase that occurs between initiation and overt clinical presentation (Wiemels et al. 1999).  Considering the young age of most patients at presentation, as well as the latency expected during clonal evolution, it was proposed that the initiating events that drive ALL development might occur in utero, with subsequent “second-hit” acquisition occurring in a subset of patients after birth (Greaves & Wiemels 2003).  Early evidence in support of this hypothesis came from twin studies.  These studies assessed concordance rates and genetic profiles in monozygotic twins, where one or both twins developed B-cell ALL.  In 1964, MacMahon and Levy performed the first assessment of ALL concordance rates in monozygotic twins.  Based on their results, they postulated that the disease involved a prenatal initiating event.  Further twin studies were valuable in lending support to this hypothesis.  For example, while the concordance rates for infant ALL, a biologically and clinically distinct disease from childhood ALL, are remarkably high, the concordance rates for childhood B-cell ALL are comparatively low, in the neighbourhood of 10% (Greaves et al. 2003; Teuffel et al. 2004).  The high concordance rate and markedly short latency in infant leukemia suggests that at the time of birth, MLL fusion-driven leukemogenesis is far enough along that it has become inevitable.  Conversely, the low concordance rates and relatively long latency period associated with childhood B-cell ALL are indicative of an in utero initiating event that generates a population of cells that still require subsequent genetic modification to develop into overt leukemia.  Since these initial twin studies were carried out 	   17 advancements in molecular genetics have enabled more sophisticated analysis in subsequent studies.  It has been demonstrated that concordant monozygotic twins with ETV6-RUNX1+ B-cell ALL often have identical ETV6-RUNX1 fusions, indicative of a single initiating event that established a pre-leukemic cell population in one twin that then spread to the other via blood exchange within a single, monochorionic placenta (Ford et al. 1998; Greaves et al. 2003; Maia et al. 2003; Wiemels et al. 1999).  In one case involving a pair of twins with an unusually large gap in age of diagnosis (9 years), researchers were able to demonstrate that while the later diagnosed twin had morphologically normal bone marrow at the time of the earlier twin’s diagnosis, cells harbouring an identical ETV6-RUNX1 fusion were still present (Ford et al. 1998).  This finding once again points to a prenatal origin of the original pre-leukemic clone, but also suggests that the subsequent latency period can be both protracted and extremely variable.  This is likely indicative of the requirement for secondary genetic alterations which occur sporadically after birth (Greaves & Wiemels 2003).    The postulates derived from the study of concordant twins were further validated by the retrospective analysis of Guthrie cards.  Small amounts of blood are collected from nearly all newborn children.  These blood spots can be used as sources of relatively intact DNA that can be interrogated by PCR and other genetic techniques to uncover the presence of specific genetic abnormalities (McEwen & Reilly 1994).  Retrospective analysis of Guthrie cards from B-cell ALL patients has demonstrated that nearly 70% of patients had detectable pre-leukemic cells present at birth.  Cells bearing hyperdiploid chromosomal gains (Panzer-Grümayer et al. 2002; Taub et al. 2002;), the MLL-AF4 fusion (Gale et al. 1997), the ETV6-RUNX1 fusion (Hjalgrim et al. 2002; Wiemels et al. 	   18 1999), as well as diagnostic clone specific IgH rearrangements (Fasching et al. 2000; Taub et al. 2002) have all been identified on a large proportion of Guthrie spots from children who would later be diagnosed with B-cell ALL.  Twin-studies have also supported this temporal model of B-cell ALL pathogenesis; such that B-cell ALL development involves an in utero initiating event followed by postnatal second-hit acquisition and clonal evolution. Analysis of concordant ETV6-RUNX1+ B-cell ALL in monozygotic twins demonstrated that while the t(12;21)(p13;q22) translocation was identical between twin pairs, all other mutations were discordant (Bateman et al. 2010).  These findings identify the ETV6-RUNX1 fusion as the only common event that has a prenatal origin, and demonstrates that the incurrence of secondary cooperating mutations drives divergent pathways of clonal evolution (Bateman et al. 2010).  These findings have been further corroborated by whole genome sequencing analysis of concordant twin pairs (Ma et al. 2013).  The exclusive discordance of secondary mutations indicates that they are of postnatal origin; if they had been acquired in utero a larger proportion would be shared between twins as a result of inter-fetal transfer (Bateman et al. 2010).   Interestingly, a large-scale analysis of cord blood samples from children who did not develop leukemia indicated that ETV6-RUNX1+ cells were present in approximately 1% of all newborn children, a rate nearly one-hundred fold higher than the risk of developing overt leukemia bearing this particular fusion gene (Mori et al. 2002).  This finding suggests that the presence of pre-leukemic cells at birth does not guarantee the eventual development of overt leukemia, and confirms that this transition depends on the acquisition of subsequent post-natal genetic aberrations (Anderson et al. 2011).  The fact 	   19 that only a small proportion of children that carry ETV6-RUNX1+ pre-leukemic cells go on to develop leukemia indicates that while in utero initiating events occur relatively commonly, acquisition of the postnatal cooperating mutations that are required to drive leukemogenesis is a relatively rare occurrence (Figure 1.2).  Collectively, these observations underscore the importance of understanding the postnatal factors that influence the survival or proliferation of pre-leukemic cells as they may have a significant impact on subsequent disease risk and progression.   Figure 1.2.  In utero initiation of childhood B-cell ALL A two-step model for the development of childhood B-cell ALL; in utero initiating events such as balanced chromosomal translocations or aneuploidy give rise to clinically silent pre-leukemic cells.  Postnatal “second-hit” acquisition in these abnormal cells drives leukemogenesis in a small subset of susceptible children.  Adapted from Greaves M., 2006. Infection, immune responses and the aetiology of childhood leukaemia. Nature Reviews Cancer, 6(3), pp.193–203. 	  	  First-Hit!Frequency: Common!Clinically Silent!Pre-Leukemic Phase!Birth! 2-15 Years!Of Age!Second-Hit!Frequency: Rare !Co-operating Mutations!In B-cell Development or!Cell-cycle Regulation Genes!Overt Leukemia!Clinical Symptoms!Chromosomal)Transloca-ons)Or)Aneuploidy)Detectable!Pre-leukemic !cells!	   20 1.2.5 Clonal evolution of relapse 	  The pathways that drive the clonal evolution of pre-leukemia to leukemia at diagnosis may also be relevant in the pathogenesis of relapsed disease.  Patient sequencing analysis has demonstrated that relapse clones can have four distinct clonal origins.  Firstly, they can represent a de novo leukemic clone that arose completely independently of diagnostic clone.  They may also be direct descendants of the ancestral clone that acquired additional genetic lesions throughout the course of treatment.  The third possibility is that the relapse clone is identical to the diagnostic clone; in this situation, initial treatment failed to eliminate all diagnostic blasts, which were then capable of re-expansion without requiring additional genetic alteration.  Finally, relapse clones may evolve independently from an ancestral “pre-leukemic” clone that pre-dates and also gave rise to the original diagnostic clone.  Genomic analyses of matched samples from ALL patients at diagnosis and relapse provide useful data on the origin and evolution of relapse clones in the different subtypes of B-cell ALL (Choi et al. 2007; Mullighan et al. 2008; Van Delft et al. 2011).  Mullighan and colleagues performed one of the largest studies of this type in 2008, which permitted an approximation of the frequency of each of these four relapse origins.  They demonstrated that only a very small proportion of relapse clones represent a de novo secondary leukemia (<10%).  A similarly small proportion of relapse clones were identical to matched diagnostic clones.  In a significant proportion (~80%) of cases, there is evidence of a clonal association between original diagnostic and relapse clones; approximately 30% of relapse clones evolved directly from the original diagnostic clone, and approximately 50% show evidence of being derived from a common ancestral clone (Figure 1.3) (Mullighan et al. 	   21 2008).  A number of additional studies in both hyperdiploid and ETV6-RUNX1+ ALL also support the conclusion that in the majority of patients, relapse evolves from a shared ancestral clone (Choi et al. 2007; Panzer-Grumayer et al. 2005; Peham et al. 2004; Van Delft et al. 2011).   Overall, this led to development of a model for the most common pathway of B-cell ALL progression in children, in which an initiating event results in the production of a pool of clonal pre-leukemic cells.  Clones within this population may then independently acquire subsequent genetic lesions and eventually evolve into the diagnostic clone.   While front-line combinational chemotherapies usually eliminate the dominant clone at diagnosis, they fail to eradicate the entire ancestral abnormal cell reservoir.  Secondary transformative events occurring within this ancestral pool then give rise to a new leukemic clone that presents as relapse (Ford et al. 2001).  This has been demonstrated in studies assessing the clonal evolution of ETV6-RUNX1+ B-cell ALL relapse (Ford et al. 2001; Zuna et al. 2004).  The initial ETV6-RUNX1 producing fusion event generates a pool of ETV6-RUNX1+ pre-leukemic cells, subsequent genetic events, most commonly deletion of the un-rearranged ETV6 allele, facilitate progression towards overt leukemia.  Genetic analysis of ETV6-RUNX1+ relapse patients has indicated that, in the majority of patients, while the ETV6-RUNX1 fusion junction is identical between the diagnostic and the relapse clone, the nature of cooperating mutational events, such as wild type ETV6 allele loss, or other copy-number alterations, is unique (Ford et al. 2001; Mullighan et al. 2008; Van Delft et al. 2011).  Latency of relapse arising from an ancestral pre-leukemic clone can be both protracted and highly variable.  This is demonstrated by cases of very late relapse (3-20 years), an unusual clinical feature of B-	   22 cell precursor ALL (Ford et al. 2015).  Such cases demonstrate that covert ancestral pre-leukemic cells can survive for significant periods of time in patients who have achieved clinical remission.  As relapse represents the fundamental clinical hurdle remaining in the treatment of BCP ALL, the fact that it likely represents the oncogenic evolution of distinct ancestral pre-leukemic clones indicates that developing an understanding of the factors that influence the survival and persistence of this pre-leukemic cell population, both prior to diagnosis and during remission, may be paramount to further advancement in the treatment of ALL.        	   23  Figure 1.3.  The clonal origins of relapse A model describing the clonal origins of relapse in childhood B-cell ALL.  From Mullighan, C.G. et al., 2008. Genomic analysis of the clonal origins of relapsed acute lymphoblastic leukemia. Science, 322(5906), pp.1377–1380.  Reprinted with permission from AAAS.  1.3 Etiology of childhood B-cell ALL 	  1.3.1 Overview 	   A growing body of investigation, particularly genome-wide associated studies, have suggested that a number of inherited genetic variations are important determinants of both ALL risk and patient responses to particular therapeutic intervention (Moriyama et al. 2015; Papaemmanuil et al. 2009).  Prior to these studies however, genetic predisposition was not considered to have a significant impact on B-cell ALL risk Figure 2.Clonal relationship of diagnosis and relapse samples in ALL. The majority of relapse caseshave a clear relationship to the presenting diagnosis leukemic clone, either arising through theacquisition of additional genetic lesions, or more commonly, arisin  from a ancestral (pre-diagnosis) clone. In the latter scenario, the relapse clone retains some but not all of the lesionsfound in the diagnostic sample, while acquiring new lesions. Lesion specific backtrackingstudies revealed that in most cases the relapse clone exists as a minor sub-clone within thediagnostic sample prior to the initiation of therapy. In only a minority of ALL cases does therelapse clone represent the emergence of a genetically distinct and thus unrelated secondleukemia.Mullighan et al. Page 8Science. Author manuscript; available in PMC 2009 September 17.NIH-PA Author ManuscriptNIH-PA Author ManuscriptNIH-PA Author Manuscript	   24 (Taylor and Birch 1996) and most research focused on identifying environmental risk factors for ALL.   The topic of how environmental exposures impact ALL risk remains controversial (Inaba et al. 2013).  Large-scale epidemiological studies have uncovered dozens of candidate exposures that may increase ALL risk; these include exposure to electric fields, parental smoking and alcohol consumption, and exposures to pesticides and other chemicals or toxins.  However, the majority of these studies lack reproducibility and have, therefore, been discounted or their candidate exposures have been determined to have little influence.  One exception is ionising radiation which is one of the only established casual exposures for childhood ALL (Inaba et al. 2013; Mitchell et al. 2010).  Evidence of this influence is derived from the impact of the 1945 atomic bombs in Japan (Hsu et al. 2013; Preston et al. 1994) and from the elevated risk caused by X-ray exposure during pregnancy (Doll & Wakeford 1997).  However, exposures to high doses of ionising radiation are no longer relevant and it has been suggested that any proposed causal exposure should explain the marked peak of childhood B-cell ALL incidence between the ages of two and five.  One candidate that satisfies this criterion is exposure to infection.   Infection was the first suggested environmental driver of childhood ALL, and it remains the strongest candidate.  Originally proposed nearly a century ago, the primary rationale for the suggestion that infection exposures may be causal for ALL was that the age distribution for several common infectious diseases such as measles, chicken-pox and croup, is very similar to that of childhood ALL (Greaves 2006; Wiemels 2012).  There was early debate as to whether this influence was associated with a specific infectious 	   25 agent (Kellett 1937), or was instead the result of an abnormal immune response to a range of infections (Poynton et al. 1922).  While these postulates were then largely ignored for the next several decades, the discovery of causal infectious agents in other cancers, such as Epstein-Barr virus for Burkitt lymphoma, renewed interest in the field.  This manifested in several exhaustive searches for a specific causative infectious agent for childhood ALL; to date, no such agent has been identified (Bartenhagen et al. 2017; MacKenzie et al. 2006), leaving an adverse immune response to common infections as the more plausible explanation.  Several immune-related models have been proposed 1.3.2 The population mixing hypothesis 	   Leo Kinlen proposed the first model for how infection exposure influences ALL in the early 1980s (Kinlen 1988).  The goal of his model was to explain sporadic incidences of ALL clustering that occurred throughout the United Kingdom.  The first of these clusters occurred near nuclear processing plants in the remote towns of Sellafield England and Dounreay Scotland.  The original explanation for these clusters involved environmental radioactive contamination in the area or germ-line transmission of mutations from exposed parents who worked at the plants.  However, Kinlen argued that the evidence supporting the role of radiation exposure was insufficient and instead proposed a “population mixing” hypothesis.  This model is based on two principle assumptions; one, that people living in rural or geographically isolated communities may escape typical exposure patterns to common, wide-spread infectious agents; and two, that any mass immigration of “outsiders” to such areas may result in epidemics of common infections.  He postulated that in rural areas such as Sellafield and Dounreay, the absence of typical infection exposures created a population that was susceptible to adverse 	   26 immune responses as a result of a lack of acquired immunity.  A large number of immigrants, coming into these communities to work at the nearby plants, brought with them a suite of common infectious agents, which were quickly disseminated throughout the native population.  The natives in these areas were prone to adverse immune responses to these common infections and these adverse reactions, through some unknown mechanism, drove leukemogenesis in susceptible children.  Additionally, Kinlen postulated that it was likely an abnormal response to a specific unidentified infectious agent, most likely a virus, within this group of newly introduced pathogens, that was responsible for the increased ALL incidence.    Kinlen found support for his hypothesis by looking at the leukemia incidence rates in other selected isolated communities that experienced significant immigration.  These communities lacked the confounding variable of nearby nuclear processing plants and still demonstrated an average 3-fold increase in leukemia-associated deaths, relative to national averages, lending support to Kinlen’s hypothesis; however no specific infectious agent has ever been identified (Kinlen & Petridou 1995; Kinlen 2015). 1.3.3 The delayed infection hypothesis 	   While there is observational support for Kinlen’s hypothesis, it only specifically addresses the rare subset of clustered leukemia cases.  The more widely applicable “delayed infection” hypothesis was proposed by Greaves in 1988.  Greaves’ hypothesis is conceptually similar to Kinlen’s, but it aims to explain the broader impact of infection on ALL.  While Kinlen favoured a defective response to a specific pathogen as an explanation, Greaves’ hypothesis did not require a specific causal pathogen, but instead suggested the explanation had more to do with broad immune response abnormalities. 	   27 His model postulates that leukemia is the result of an abnormal immune response that develops as a consequence of a lack of early-life exposures to common, non-specific infections.     Many consider Greaves’ hypothesis to be a theoretical extension of the hygiene hypothesis, which has been used to explain the high incidence of autoimmune and allergic disorders in the developed world.  It has been suggested that these two hypotheses reflect the same underlying mismatch between the evolutionary genetics modulating immune development and the pattern of immune exposures in the developed world (Greaves 2006).  Later iterations of the “delayed infection” hypothesis also included a proposed mechanistic explanation for how abnormal immune responses may drive leukemogeneis.  In this model, a dysregulated T-cell response to infection results in excessive inflammation and a cytokine release profile that is capable of either suppressing hematopoiesis or inducing apoptosis in immune cells.  It is possible that existing pre-leukemic cells possess an intrinsic survival and or proliferative advantage that then allows them to expand in the suppressed environment.  This proliferative or regenerative wave of pre-leukemic cells may then result in the incurrence of second-hit mutations which occur either by chance, or as a consequence of the genotoxic cytokine milieu, resulting in the generation of overt leukemic cells which will emerge from this cell population (Greaves 2006).   1.3.4 Epidemiological support for the delayed infection hypothesis While the underlying mechanism is yet to be identified, there is significant evidence that supports the broader concepts of Greaves’ hypothesis, most of which has come from epidemiological studies evaluating how infection exposure affects ALL risk.  	   28 While the broad results of these studies appear discordant, with both increased and decreased risk being reported, some common patterns have been uncovered.  A consistent protective effect has been observed when surrogates of infection exposure, such as early-life daycare attendance, were analyzed (Chan et al. 2002; ; Gilham et al. 2005; Infante-Rivard et al. 2000; Ma et al. 2002; Ma et al. 2005; Ma et al. 2010; Perrillat et al. 2002; Urayama et al. 2011).  It is well appreciated that playgroup or daycare attendance in early-life increases the likelihood of being exposed to common infections, both bacterial and viral.  The two largest studies of this kind, the United Kingdom Childhood Cancer Study (UKCCS) and the Northern California Childhood Leukemia Study (NCCLS), together analyzed over four thousand ALL patients and eight thousand healthy controls.  Both of these studies indicated a significant protective influence of daycare attendance that began within the first six months of life (Ma et al. 2010; Rudant et al. 2015; Urayama et al. 2010).  The results of these studies also suggest that the extent of protection is quantitatively related to the total childhood daycare hours, which is a combinatorial measure of time spent in daycare and the number of other children interacted with at daycare (Ma et al. 2002; Urayama et al. 2010; Urayama et al. 2011).   While these data provide intriguing support for Greaves’ delayed infection hypothesis, daycare attendance is at best an indirect measure of infection exposure.  Additional studies that evaluated the specific infection histories of ALL patients, either using surveys or practitioner records, were conducted to get a more accurate measure of how infection history impacts ALL risk.  Unlike the relatively consistent results from studies investigating daycare attendance, the studies evaluating specific infection histories have produced discordant conclusions, with both increased and decreased risk 	   29 being reported in association with certain infections.  A history of roseola (Chan et al. 2002), gastro-intestinal infections (Jourdan-Da Silva et al. 2004), and ear-infections (Urayama et al. 2010) have been associated with protection from ALL.  On the other hand, several positive associations have also been uncovered.  For example, it has been reported that having any documented infection during the first year of life is associated with increased ALL risk ( Crouch et al. 2012; Dockerty et al. 1999), and that, on average, ALL patients experience significantly more early-life infections than healthy controls (Crouch et al. 2012; Roman et al. 2007).  Additionally, specific infections such as influenza have been associated with increased ALL risk (Chan et al. 2002; Dockerty et al. 1999).  To complicate matters further, there have also been reports that infection history has no impact on ALL incidence (Cardwell et al. 2008).   The conflicting conclusions drawn from the infection studies underscore the complicated nature of this association.  In a meta-review of several large-scale epidemiological studies, Rudant and colleagues suggest that the fact that association with ALL is more consistent for daycare attendance than for infection history is likely indicative of two coexisting mechanisms: The first being that exposure to infectious agents, even those that are weakly symptomatic and do not require physician visits, protect against ALL via immune stimulation; and the second that infections may be more symptomatic in those children that will develop ALL as a dysregulated immune context already exists at the time of early exposure (Rudant et al. 2015).  Under these conditions, proxies of exposure to infections that remain asymptomatic may be inversely correlated with ALL risk, while the impact of documented, symptomatic infections would be significantly more variable, and likely depend on both the nature of the infectious agent 	   30 and the intensity of the resulting symptoms (Rudant et al. 2015).  Evidence in support of this conclusion is derived from the fact that polymorphisms in cytokine genes, which can affect both resting and induced expression levels, have been correlated with ALL development (Chang et al. 2010; Cloppenborg et al. 2005; Han et al. 2010; Winkler et al. 2015).  Additionally, reduced levels in neonatal blood of IL-10, a cytokine intricately involved in determining the intensity and duration of immune responses, has been associated with an increased risk of ALL development (Chang et al. 2011). 1.3.5 Experimental support for the delayed infection hypothesis 	   Considering the complexity of analyzing this type of association with retrospective human studies, relevant and applicable research models represent an important resource for uncovering the specific nature of the association between infection exposure and ALL risk.  While limited, some experimental evidence supporting the postulates outlined in Greaves’ delayed infection hypothesis does exist.   From a proof of concept perspective, Martin-Lorenzo and colleagues have implicated infection exposure as a causal driver of leukemia in a Pax5+/- mouse model.  They demonstrate a striking difference in leukemia incidence between mice that were housed in a specific pathogen-free (SPF) environment for life, versus those that were housed in an SPF environment for the first two months of life and then moved to a conventional facility.  None of the life-long SPF-housed mice developed ALL, while 22% of those that were moved to a conventional facility after the second month developed disease (Martin-Lorenzo et al. 2015).  This result supports the notion, that a lack of early exposure to infection, followed by subsequent infection exposure later in life, may increase ALL risk. 	   31  In terms of infection exposure being able to drive the pre-leukemia to leukemia transition, Markus Muschen’s group demonstrated that repeated rounds of toll-like receptor (TLR) 4 ligation, in concert with growth factor starvation, was capable of driving leukemogenesis via the induction and off-target activity of AID and RAG1-RAG2 (Swaminathan et al. 2015).  Additionally, in regards to pre-leukemic cells possessing an intrinsic growth or survival advantage in the suppressive inflammatory environment established during an immune response, Greaves has demonstrated that while ETV6-RUNX1+ BaF3 cells demonstrated slower basal proliferation rates than their wild type (ETV6-RUNX1-) counterparts, the presence of the TEL-AML1 fusion imparted significant resistance to the anti-proliferative effects of TGF-β (Ford et al. 2009).  Transduction of ETV6-RUNX1 also confers a growth advantage on human CD34+ cells in the presence of TGF-β, validating the potential relevance of this observation to human ALL.    Overall, there is significant epidemiological and limited experimental evidence supporting Greaves’ delayed infection hypothesis, and it remains the most widely accepted and applied model for ALL etiology developed to date.  While the association between early-life infection exposure and ALL risk remains incompletely understood, it is generally accepted that the immune system and its responses play a role.  Developing a more definitive understanding of the mechanism(s) behind this association will be required for full validation of this hypothesis. 1.3.6 Additional hypotheses 	   There are two additional hypotheses that attempt to develop a mechanistic model for the association between infection exposure and ALL risk.  While these models 	   32 currently lack experimental validation, they both represent intriguing candidate explanations that warrant further exploration.  The first is “the adrenal hypothesis” proposed by Schmiegelow and colleagues in 2008.  This hypothesis incorporates a number of additional observations.  Firstly, that glucocorticosteroids (GCs) are among the most effective anti-leukemic treatments available; secondly, that adrenocorticotropic hormone (ACTH) can increase cortisol levels, which leads to morphological remission of ALL; and finally, that during infection-induced stress ACTH may induce cortisol release that results in GC levels that are nearly equal to those used in GC-based therapies (Schmiegelow et al. 2008).  With these observations in mind, Schmiegelow asserts that early infections affect ALL risk by inducing direct anti-leukemic effects of cortisol and by altering of the T-helper type 1 (Th1)/T-helper type 2 (Th2) balance.  Repeated early-life infections would, therefore, serve to serially deplete the pre-leukemic cell reservoir, thereby reducing the risk of subsequent ALL development.  Schmiegelow’s hypothesis also provides an alternative explanation for the much lower incidence of leukemia in the developing world.  While Greaves, in line with the hygiene hypothesis, suggests that repeated infection exposure serves to “educate” the developing immune system so as to prevent dysregulated responses later in life, Schmiegelow suggests that repeated early-life infection exposure will serve to alter the Th1/Th2 balance during subsequent immune responses.  He postulates that in developing countries, high infection burden may force the immune system to adapt in order to avoid the development of overactive Th1-based immune responses, which are particularly dangerous for malnourished children.  The immune system will accomplish this by increased cortisol induction during an immune response.  The production of cortisol will favour the production of anti-inflammatory 	   33 cytokines such as IL-4 and IL-10, favouring a Th2 response, while also serving to directly eliminate pre-leukemic cells.  As excessive Th1-driven inflammatory responses are also dangerous for healthy children, this model could also be relevant in the developed world.  While experimental support for Schmiegelow’s hypothesis is currently lacking, in animal models it has been demonstrated that stimulation of the early-life immune system with bacterial endotoxins can induce permanent changes in the nature of subsequent immune responses; one of these changes is higher, and longer-sustained plasma cortisone levels (Shanks et al. 2000).  A detraction from the adrenal hypothesis is that it proposes that infection severity should be negatively correlated with ALL risk; that is, a history of severe infections, akin to those that would be required to permanently alter subsequent immune responses, should reduce ALL risk.  The epidemiological studies involving practitioner records, however, do not support this assertion.  Another mechanistic hypothesis is Richardson’s “infective lymphoid recovery” hypothesis.  This model seeks to directly explain the apparently paradoxical conclusions obtained from epidemiological studies assessing the influence of infection exposure on ALL risk.  This hypothesis postulates that hygienic conditions in infancy inhibit early immune development, causing reduced efficacy of the secondary immune response to later infections.  This results in the over-production of heat-shock proteins (HSPs) in response to later infectious fevers.  HSPs then induce increased production of Th1 inflammatory cytokines and generate an anti-apoptotic, pro-mutagenic environment.  The glucocorticoid release that follows Th1 cytokine production causes acute thymic involution, a decline an anti-tumor immunosurviellance, and maturation arrest of B-cell precursors (Richardson 2011).  Recurrent infections lead to progressively degraded stress 	   34 responses and higher production of hemostatic cytokines, such as IL-7, as the body tries to reconstitute atrophied lymphoid tissues, stimulating the proliferation of pre-leukemic B-cell precursors in the previously established pro-mutagenic cytokine milieu, thus promoting evolution to overt leukemia.  In the context of conclusions drawn from the existing epidemiological data, the infective lymphoid recovery hypothesis postulates that broad exposure to early-life infections, as might be conferred by daycare attendance, prevent leukemia by promoting appropriate immune priming.  Conversely, in children with inadequate immune priming, repeated infections later in life drive second-hit acquisition in pre-leukemic cells through the induction of genotoxic and proliferative stresses.  While potentially explaining the seemingly contradictory influences of surrogates of infection exposure and documented early-life infection, this hypothesis is currently lacking experimental support.       1.3.7 Experimental models of ALL 	   There are multiple murine models utilized for B-cell ALL research.  Each of these models is accompanied by a unique set of advantages and disadvantages.  Xenograft models, which are commonly utilized, involve adoptively transferring human hematopoietic and leukemia cells into immune-deficient NOD-SCID and NOD-scid/IL-2Rγ null (NSG) mouse strains (Lee et al. 2007; Meyer & Debatin 2011).  While these represent valuable platforms for drug screening, they are unsuitable for evaluating how the immune system influences ALL progression, primarily because the mouse strains used have severe immune defects which are required to achieve inter-species cell engraftment.  Another method used to create ALL mouse models involves the transduction of ALL-associated genes into murine B-cell precursors or appropriate 	   35 murine cell lines, followed by subsequent adoptive transfer of these transformed cells into recipient mice.  While this method permits the analysis of the ALL progression in the context of an intact immune system, adoptive transfer models of an in situ disease do not accurately model microenvironment interaction and normal tumor progression (Aparicio et al. 2015; Richmond & Su 2008).  Multiple efforts have been made to create a transgenic mouse that carries a BCP- ALL-associated transgene and exhibits a two-stage model of leukemogenesis with varying degrees of success.  For example, the results of efforts to make a transgenic mouse model driven by the ETV6-RUNX1 fusion gene have been less than ideal, as these mice demonstrate remarkably low disease penetrance with a variable phenotype that also includes T-cell leukemias (Fischer et al. 2005; Schindler et al. 2009; Tsuzuki et al. 2004; Vandel Weyden et al. 2011).  Similarly, a commonly utilized mouse model is the Eµ-MYC mouse, which was designed to model Burkitt’s lymphoma, expresses the MYC oncogene under the control of the µ-heavy chain enhancer (Sheppard et al. 1998).  While these mice have been extensively utilized, the disease presentation in this model tends to be quite variable in terms of B-cell stage affected, with a mature B-cell lymphoma being the most common presentation, rendering it less than ideal for the study of factors that influence the development of childhood ALL which exhibits nearly universal arrest at B-cell precursor stages. For my research I have utilized both the Eµ-RET and E2A-PBX1 transgenic mouse models.  The Eµ-RET mouse carries a RET/RFP fusion transgene under the control of the µ-heavy chain enhancer.  This fusion protein carries the tyrosine kinase domain of RET and the amino terminal end of the transcriptional activator RFP (Iwamoto 	   36 et al. 1991; Wasserman et al. 1998; Zeng et al. 1998).  These mice have high penetrance of a late-pro B-cell stage ALL with a peak incidence between 4 and 7 months of age.  The Eµ-RET mouse exhibits a distinct two-stage pathology.  Expression of the transgene begins in utero and leads to the expansion of an abnormal pre-leukemic cell population that is detectable in the fetal liver.  As these mice age, pre-leukemic cells become disseminated throughout the body and are readily detectable in the blood, spleen and bone marrow, months prior to overt disease onset. This feature, which mimics the in utero origins of human pre-leukemic cells, allows for the analysis of factors that influence pre-leukemic cell survival in the presence of an intact immune system.  However, Despite being expressed in early B-cell development, RET is not associated with human childhood B-cell ALL.  Mutations in RET are instead associated with human papillary carcinoma and multiple endocrine neoplasia (Mulligan et al. 1993). Cleary and colleagues recently developed the E2A-PBX1 mouse model (Duque-Afonso et al. 2015).  These mice carry the E2A-PBX1 fusion gene, which is found in 5-15% of childhood B-cell ALL cases, under the control of three distinct cre-inducible promoters, CD19, Mb1 and Mx1.  These mice exhibit ALL penetrance ranging from 5-50%, based on the promoter used to drive E2A-PBX1 expression.  This model also exhibits a two-stage pathology, and overt transformation is commonly associated with spontaneous loss of Pax5.  This model has a definable pre-leukemic phase during which factors that influence the survival of these cells can be studied. Also, leukemia in this model is driven by a transgene that is relevant to human childhood ALL.  However, this model exhibits a slight variability in phenotypic presentation, as two different, although highly similar, sub-types of B-cell ALL are induced.   	   37 1.4 Immune ontogeny 	  1.4.1 Overview 	   There is a large amount of evidence indicating that components of the immune system have an influence on the etiology of childhood B-cell ALL.  Given the sharp peak incidence of B-cell ALL between the ages of two and five, it is important to consider the immune system in the context in which it exists both leading up to and during this particular time period.  It is well established that there are significant quantitative and qualitative differences in the early-life immune system compared to the adult immune system (Cuenca et al. 2013; Kollmann et al. 2012; Marchant & Kollmann 2015).  Because of increased early susceptibility to various infections, the neonatal immune system was originally considered to be incompetent or dysfunctional.   Recent studies have demonstrated that this is not the case and that while the early immune system displays deficits in some response pathways, it is equally, if not more proficient than the adult in others (Kollmann et al. 2012; Vanden-Eijnden et al. 2006). These differences reflect the unique immune challenges associated with different life stages.  The most important challenge during the early-life period is mediating the transition from the sterile intra-uterine environment to the pathogen-rich outside world; particularly avoiding excessive inflammation while maintaining mucosal barrier integrity during microbial colonization.  The majority of the studies that assess the properties of the early-life immune system have done so by comparing the neonatal or early-life immune responses (0-2 years) to those of the adult.  The subsequent sections will outline several key differences between the early-life and adult immune system. 	   38 1.4.2 Soluble factors 	   There are several quantitative and qualitative differences in plasma composition during early-life versus adulthood; these differences have an impact on immune responses in these settings.  For example, it has been shown that neonatal plasma has significantly higher basal concentration of several anti-inflammatory cytokines, such as IL-10 and TGFβ, compared to adult plasma (Dowling et al., 2015).  Additionally, there are also several other immunosuppressive factors that are expressed at much higher concentrations in early-life, including proteins, lipids, purines and sugars.  Higher expression of these immunosuppressive factors is thought to be involved in facilitating early-life microbial colonization, as well as preventing excessive pro-inflammatory responses, which may be particularly harmful to infants.  One of the most important of these factors is adenosine, which is found in significantly higher concentrations in plasma early in life (Levy et al. 2006).  Through the action of adenylate cyclase, high adenosine results in the intracellular accumulation of cyclic adenosine monophosphate (c-AMP).  By inhibiting the phosphorylation of the p38 mitogen-activated protein kinase (MAPK), c-AMP acts as a potent inhibitor of inflammatory cytokine production.  Adenosine, therefore, inhibits the production of cytokines that require MAPK signaling for induction, while having little impact on the production of those that do not.  By inhibiting the production of cytokines such as IL-12, TNFα, type 1 and type 2 interferons and IL-1β, while having no influence on the production of IL-6 and IL-23, adenosine inhibits Th1 inflammatory responses and skews towards Th2 and Th17 based responses, driving further production of cytokines such as IL-10, IL-23 and IL-6 (Levy et al. 2006; Haskó et al. 2000).  The high adenosine in early-life plasma is likely a reflection of higher 	   39 concentrations of phosphatases that are capable of converting adenosine-triphosphate (ATP) into adenosine.  ATP is a molecular danger signal that is recognized by pattern-recognition receptors (PRRs) to induce inflammatory responses.  Higher early-life expression of enzymes such as CD73 and alkaline phosphatase results in enhanced processing of ATP into adenosine, thereby preventing induction of inflammatory responses by ATP and increasing concentrations of anti-inflammatory adenosine.   1.4.3 Toll-like receptors 	  Toll-like receptors (TLRs) have a crucial role in innate defense against invading pathogens (Kumar et al. 2009).  They are capable of detecting a broad range of pathogen associated molecular patterns (PAMPs) and are critically involved in the initiation of both innate and adaptive immune responses (Blasius & Beutler 2010). TLRs are type-1 integral membrane glycoproteins.  The N-terminal domains of these proteins contain leucine-rich-repeat motifs that are responsible for binding ligands, while the intracellular domain, referred to as the toll/IL1-R (TIR) domain, has three homologous domains that recruit signaling proteins to propagate the TLR signal (Medzhitov 2001; Takeda et al. 2003).  There are nine distinct proteins in the TLR family, each with unique ligand specificity, localized to the cell membrane or the endoplasmic reticulum (Akira 2003).  Upon ligand binding, TLRs dimerize and undergo a conformational change that is required for the recruitment of downstream signaling molecules (Aderem & Ulevitch 2000; Takeda et al. 2003).  The majority of TLRs signal through an adaptor protein called myeloid differentiation primary-response protein 88 (MyD88) that is responsible for the recruitment of subsequent signaling proteins.  A small subset of TLRs, namely TLR-3 and TLR-4, are capable of signal propagation in the absence of MyD88 by instead 	   40 utilizing an adaptor protein called TRIF; this MyD88-independent signaling cascade seems to result in increased production of type-1 interferon (Kumar et al. 2009; Takeda et al. 2003; Medzhitov 2001).  It has been demonstrated that TLR function is well developed in newborns (Strunk et al. 2011).  The expression of TLRs and the majority of their downstream signaling components are relatively stable during the first five years of life and occur at levels comparable to that of adults (Dasari et al. 2011; Reece et al. 2011). While expression is similar, TLR-mediated production of innate effector molecules such as oxygen radicals is markedly reduced in early-life (Bae et al. 2011; Corbett et al. 2010). These oxygen radicals are often associated with the induction of inflammatory responses and their reduced TLR-dependent production during early-life may be a reflection of the bias away from potentially harmful responses.  Additionally, the pattern of cytokine release induced by TLR ligation is markedly different in early-life compared to adulthood (Corbett et al. 2010; Kollmann et al. 2012; Marchant & Kollmann 2015).  Broadly, while adult responses are dominated by the production of Th1-skewing cytokines such as IL-12p70, early-life responses are characterized by the production of anti-inflammatory cytokines, such as IL-10, and Th17-inducing cytokines, such as IL-23 and IL-6 (Kollman et al. 2012).  1.4.4 T-Helper immune responses 	  In order to be effective, immune responses must be tailored to the specific nature of the invading pathogen.  T-helper lymphocytes are responsible for ensuring that the most appropriate immune response is induced.  Three separate signals are required to induce the activation and differentiation of naïve CD4+ T helper cells.  Following 	   41 recognition of a pathogen, often by TLR ligation, macrophages and dendritic cells phagocytose invading microbes.  Theses microbes are digested and processed in the phagolysosome, and components of them (antigens) presented on the cell surface in the context of major-histocompatibility complex (MHC) molecules.  Recognition of the antigen:MHC complex by T-cell receptors on the surface of naïve CD4+ T-cells provides the first activation signal.  Engagement of TLRs on the surface of the dendritic cell or macrophage also results in the expression of co-stimulatory molecules, such as CD80/86 on the cell surface.  Binding of the CD80/86 by CD28 on the T-cell surface serves as the second activation signal.  The third signal required for T-helper cell activation and differentiation is provided by the cytokine milieu present at the time the first two activation signals are received.  As TLRs recognize specific PAMPs, the combination of TLRs engaged in response to a particular pathogen provides some information about the potential identity of that particular pathogen.  Macrophages and dendritic cells interpret these TLR signals, and produce and release particular cytokines based on the information they impart, in order to induce the most appropriate immune response.  This cytokine profile is a primary determinant of the type of T-helper cell differentiation, which in turn determines the specific set of cytokines they will subsequently produce (Abbas et al. 1996; O’Garra & Arai 2000).  There are three primary types of T-helper (Th) cells, Th1, Th2 and Th17.   Th1 immune responses are induced by intracellular pathogens.  Several cytokines can contribute to induction of a Th1 response, but the role of IL-12p70 is primary.  Once induced, Th1 CD4+ cells are characterized by the production of IL-2, TNFα and interferon-gamma (IFNγ).  IFNγ, the primary Th1 cytokine, is a pleiotropic cytokine that 	   42 plays a critical role in both innate and adaptive immune response.  Natural killer (NK) cells, CD8+ T-cells and Th1 CD4+ T-cells are the most significant producers of IFNγ, however several other cells have also been demonstrated to be capable of producing it under specific circumstances.  IFNγ exerts its effects by binding to its receptor, a heterodimer composed of the IFNγR1 and R2 chains that is expressed on the surface of a number of different immune cell subsets (Farrar & Schreiber 1993).  The IFNγ receptor signals through the JAK-STAT signaling pathway and is capable of inducing a broad range of cellular responses (Schroder et al. 2004).  Mice deficient in IFNγ display significant impairment in several aspects of both innate and adaptive immune responses and exhibit increased susceptibility to infection with intracellular bacteria and viruses.  The production of cytokines that support Th1 differentiation, such as IL-12p70, is deficient during the neonatal period (Debock & Flamand 2014; Kollmann et al. 2012).  This deficit can be attributed to intrinsic expression abnormalities and extrinsic suppression, and is likely a reflection of the need to avoid potentially harmful inflammatory responses during microbial colonization (Kollmann et al. 2012). Th2 immune responses are generated in response to extracellular pathogens such as helminthes and nematodes.  The primary cytokine responsible for the induction of Th2 differentiation is IL-4 (O’Garra & Arai 2000).  Subsequently, Th2 CD4+ T-cells are characterized by the production of IL-4, IL-6, and IL-13, which are required for optimal antibody production.  The inappropriate or over production of Th2 cytokines has been associated with the development of atopy and allergic asthma (Romagnani 2004; Schuh et al. 2003).  In contrast to Th1 responses, the neonatal immune system is proficient in 	   43 the production of cytokines that induce Th2 differentiation and Th2 immune responses reach maturity early in life (Debock & Flamand 2014). Another T helper subset is Th17 cells, which are characterized by their production of IL-17.  Initial differentiation of Th17 cells is induced by IL-6, TGFβ and IL-21, which stimulate surface expression of the IL-23 receptor (Boniface et al. 2008; Miossec et al. 2009).  IL-23 is a hetero-dimeric cytokine consisting of the IL-12p40 subunit and the IL-23p19 subunit (Watford et al. 2004).  Signaling through the IL-23 receptor is required for complete Th17 differentiation and induction of IL-17 production (Iwakura & Ishigame 2006).  Mice that lack IL-23p19 demonstrate a near complete lack of Th17 cells and demonstrate a profound defect in IL-17 production (Cua et al. 2003; Langrish et al. 2005; Stockinger & Veldhoen 2007).  The primary functions of Th17 cells appear to be maintaining mucosal barrier integrity as well as providing defence against the pathogens not well controlled by either Th1 or Th2 responses (Korn et al. 2009).  Through the production of IL-17, Th17 cells are potent inducers of tissue inflammation and inappropriate or excessive activation of these cells has been associated with a number of autoimmune and inflammatory diseases (Boniface et al. 2008; Langrish et al. 2005).  IL-17 promotes granulocyte and neutrophil accumulation and activation at infection sites, while also promoting mucosal barrier integrity by stimulating tight junction formation and inducing mucin secretion (Bradley Forlow et al. 2001; Chen et al. 2003).   CD4+ Th17 cells are not the only producers of IL-17.  γδ TCR+ T-cells have been shown to be important early producers of IL-17 in response to infection, sometimes in levels that exceed that of CD4+ Th17 cells (Cua & Tato 2010).  While the nature of ligands recognized by the γδ TCR remains poorly understood, this cell population has 	   44 been demonstrated to respond to a range of microbial products as well as stressed epithelial cells to produce significant amounts of IL-17 (Chien et al. 1996; Haas et al. 1993; Kabelitz et al. 2005).  A subset of γδ TCR+ T-cells have also been shown to express functional TLRs on their surface (Martin et al. 2009).  Engagement of these TLRs has been demonstrated to induce the rapid production and secretion of IL-17 in a response that is potentiated by concurrent reception of IL-23.  Interestingly, the development of these IL-17 producing γδ TCR+ T-cells appears to be restricted to a functional embryonic wave (Haas et al. 2012).  This finding may explain the observation that the total number IL-17 producing γδ TCR+ T-cells peaks near birth and then slowly decreases with age.  Clearly the pattern of cytokines produced by dendritic cells and macrophages after pathogen detection plays a significant role in determining the nature of the subsequent immune response.  Cytokine induction patterns shift rather dramatically based on age.  It has been demonstrated that relative to adults, neonates produce significantly less IL-12p70 in response to infections.  Conversely however, production of IL-6 and IL-23 from neonatal macrophages and dendritic cells dramatically exceeds that of adults (Kollmann et al. 2012; Vanden Eijnden et al. 2006).  This cytokine pattern leads to the increased number and function of Th17 cells and, conversely, low numbers of Th1 cells in early-life (Debock & Flamand 2014).  These patterns shift throughout life as IL-23 levels slowly decrease over the first two years, while production of pro-inflammatory and Th1-supporting cytokines gradually increase (Corbett et al. 2010; Kollmann et al. 2012).  One of the last cytokines to reach adult level production after TLR activation is IL-12p70.  	   45  The mechanism of how cytokine production profiles change after TLR stimulation remains poorly understood; however, some observations have provided insight.  For example, while the early-life expression of TLRs and their downstream signaling components is similar to that of adults (Dasari et al. 2011), the function of a subset of these signaling components seems to be impaired in early-life.  IL-12p70 and IL-23 are heterodimeric cytokines that share a common subunit (IL-12p40).  The nuclear translocation of IRF5 in concert with the activation of NFκB and MAPK are critical for the production of IL-12p40 and IL-23p19 (Kollmann et al. 2012).  While these pathways are intact and functional in the neonate, the IRF3-induced production of IL-12p35 is defective (Kollmann et al. 2012; Vanden Eijnden et al. 2006).  While the early steps of IRF3 activation, including phosphorylation and nuclear translocation, occur at similar levels in neonates and adults, subsequent IRF3 signaling is impaired in neonates.  Neonatal IRF3 demonstrates a reduced capacity for co-activator association and subsequent DNA binding, resulting in reduced expression of IL-12p35 (Kollmann et al. 2012; Vanden Eijnden et al. 2006).  Maintained production of IL-12p40 and IL-23p19 in conjunction with reduced production of IL-12p35 leads to increased IL-23 and impaired production of IL-12p70.  The cause of this deficit is not completely understood but it has been proposed that it may be related to age-dependent epigenetic patterning (Kollmann et al. 2012). 1.4.5 Anti-viral responses 	   Antiviral responses are characterized by the TLR-induced production of type-1 interferon.  Despite being strikingly deficient around the time of birth, this anti-viral response is one of the first to mature (Danis et al. 2008).  The defect in type-1 interferon 	   46 production at birth seems to be associated with a failure of IRF7 to translocate into the nucleus and induce expression of these proteins.  However, through an unknown mechanism this defect is resolved and the production of type-1 interferon reaches adult levels within a few weeks of birth.  Type-1 interferons signal through a heterodimeric receptor called the interferon-alpha receptor (IFNAR) that is composed of two subunits IFNAR1 and IFNAR2, and is expressed in most tissues (Hervas-Stubbs et al. 2011).  Type-1 interferon signals primarily through the JAK-STAT signaling pathway and have a broad range of effects.  They induce the up-regulation of MHC class 1 (MHC-1) expression and induce anti-proliferative responses in virally infected cells. Furthermore type-1 interferon has been demonstrated to influence monocyte/macrophage differentiation and function (Bogdan et al. 2004), as well as promote the production of IL-12p70 (Gautier et al. 2005).  In general the effects of type-1 interferon work in concert to induce the development of an “anti-viral” state in the host.  Their rapid maturation after birth is likely indicative of the importance of these cytokines in early-life protection from viral invasion. 1.4.6 Modeling immune responses to common early life infections 	   Studying immune ontogeny is a difficult task in humans.  Constraints in appropriate human sample collection have required most researchers to use mouse models to study neonatal immunity.   Mice are, however, limited by species-specific characteristics that complicate the question of translational relevance.  It has been demonstrated that mice express divergent Toll-like receptors that can respond differently to certain agonists (Dowling & Levy 2014; Marchant & Kollmann 2015).  Mice also exhibit significant differences in leukocyte subset proportions as well as an absence of 	   47 leukocyte defensins and a more limited Fc receptor repertoire (Mestas & Hughes 2015).  Using mice as a model system also forces compromises in terms of route of infection exposure and inoculum size.  There are also developmental differences as it is difficult to approximate and correlate stages of immune development between mice and humans; for example mouse hematopoiesis continues to occur in the spleen well into adulthood, while the spleen stops being a site of hematopoiesis before birth in humans.  Despite these differences and complications, the murine and human immune systems are remarkably similar and mice remain an indispensable tool for studying immune ontogeny and modeling early-life immune responses.  Models have been developed to study immune responses to a multitude of human infections that commonly occur during early life.  For example, attenuated and non-attenuated Listeria monocytogenes (Lm) has been used extensively as a model organism to assess early-life immune responses to intracellular pathogens (Cossart 2011; Pamer 2004).  Lm is a food-borne gram-positive intracellular bacteria that causes listeriosis in humans (Farber & Peterkin 1991).  Lm is taken up by phagocytic cells, but escapes the phagosome through the action of a bacterial protein called listeriolysin-O which forms pores in the phagosome membrane.  Once in the cytosol, Lm replicates and is then capable of spreading to neighbouring cells by inducing the re-organization of actin filaments (Hamon et al. 2006).  Lm is also capable of entering non-phagocytic cells through an active process.  Two proteins on the surface of Lm called internalin-A and B bind to E-cadherin on the surface of epithelial cells at tight junctions.  This binding induces cytoskeletal re-arrangements that facilitate the uptake of Lm, and it is through this mechanism that Lm is capable of crossing into the lumen of the gut (Hamon et al. 	   48 2006).  Adult responses to Lm are primarily Th1 skewed and involve significant production of IFNγ (Pamer 2004).  While several different effector cell subsets are involved in the response to Lm, it has been shown that Lm exposure is capable of inducing significant production of IL-17 from γδ TCR+ T-cells (Laird et al. 2013).  Several features of Lm make it a valuable model for studying immune responses to intracellular bacteria.  For example, it is easy to culture, relatively safe to handle, and its life cycle is well defined; also its genome has been sequenced and several different attenuated strains are commercially available (Hamon et al. 2006; Pamer 2004).  Other very commonly encountered early-life infectious diseases have been extensively modeled in mice.  Epstein-Barr virus (EBV) is a ubiquitous herpesvirus that is responsible for causing mononucleosis, and less frequently B-cell lymphoproliferative diseases such as Burkitt lymphoma in humans.  EBV is spread through bodily fluids and is commonly encountered early in life; by the age of four the frequency of seropositivity for EBV-specific antibodies is near 100% in developing countries and between 25-50% in developed countries (Cohen 2000; Jenson 2011; Carvalho-Queiroz et al. 2016).  In the majority of cases, EBV acquired during childhood is subclinical, while EBV acquired later in life is associated with more severe symptoms.  In mice, EBV is commonly modeled using the murine-gammaherpes-68 virus (MHV-68).  Like EBV, this virus infects B-cells, induces virus-driven activation and proliferation of B-cells and has an acute phase followed by life-long latent phase; MHV-68 and EBV have also been demonstrated to elicit similar immune responses (Olivadoti et al. 2007).    Another infectious agent commonly encountered in early-life that has been extensively modeled in mice is cytomegalovirus (CMV).  It has been demonstrated that 	   49 between 30 and 70% of children aged between 4 and 10 in the United States are seropositive for CMV-specific antibodies (Staras et al. 2008).  CMV infection is largely asymptomatic; however congenital CMV infection (passed from mother to fetus) can be extremely serious and often results in neurological defects and deficits (Leung et al. 2003; Ross & Boppana 2005) .  Like EBV, CMV is also characterized an acute phase followed by a life-long latent phase.  Because of the strict species specificity of CMV, it is modeled in mice using murine-cytomegalovirus (m-CMV), which is closely related and possesses several shared characteristics.  Control of m-CMV infection in the mouse is mediated primarily by interferon and CD8+ cytotoxic T-cells (Krmpotic et al. 2003). 1.5 Hypotheses 	   The association between infection exposure and ALL risk remains controversial.  While multiple models have been developed to explain the somewhat discordant conclusions derived from epidemiological studies, each of these models suffers from a lack of experimentally identified mechanism.  Only by gaining greater understanding of the molecular pathways involved will the nature of the associations proposed from these epidemiological studies be revealed.  The Eµ-RET and E2A-PBX1 transgenic mouse models represent two relevant systems from which such mechanistic explanations may be derived.  The overall hypothesis of this work is that both the resting and activated immune system is capable of influencing leukemia progression by modulating the survival, proliferation and evolution of pre-leukemic cells.  More specific hypotheses are outlined in the proceeding chapters as follows.  The work described in Chapter 3 was designed to evaluate the hypothesis that defects in the immune system increase ALL risk even in the absence of infection exposure.  	   50 This work was designed to explain the association between polymorphisms in cytokine genes, particularly interferon-gamma (IFN-γ),	  and increased ALL risk.  I found that basal IFN-γ was capable of significantly altering disease kinetics by directly inhibiting the proliferation of pre-leukemic cells.  The majority of the existing epidemiological data indicates that infection exposures impact ALL risk.  After identifying IFN-γ as a modulator of pre-leukemic cell proliferation and survival, we attempted to determine the influence of IFN-γ induction on disease progression within our mouse models; this work is described in chapter 4.  It was designed to test the hypothesis that immune modulation with toll-like receptor (TLR) ligands will protect against ALL development by negatively impacting the survival and/or proliferation of pre-leukemic cells.  I found that stimulation with the TLR ligands CpG or R848 was capable of significant pre-leukemic cell killing via the production of both type-1 and type-2 interferon.  Furthermore, administration of CpG in vivo resulted in a marked reduction in pre-leukemic cell burden that was associated with a significant delay in disease onset in the Eµ-RET mouse model.  While existing epidemiological studies have produced conflicting results, one consistent finding is that the timing of infection exposure represents a critical variable in this proposed association.  Therefore, the work described in chapter 5 was designed to interrogate the hypothesis that neonatal infection exposures will have a particularly significant inhibitory impact on disease progression in the Eµ-RET and E2A-PBX1 mouse models.  I found that neonatal exposures to a variety of live and attenuated pathogens induced the depletion of both pre-leukemic and leukemic cells, while adult exposures to the same infections had no impact.  This impact was associated with the IL-	   51 23-induced production of IL-17 by γδ T-cells, a response that occurred in neonatal but not adult mice.  This represents the first mechanistic explanation for the observed importance of timing in the association between infection exposure and ALL risk.                              	  	   52 Chapter 2: Materials and methods 	  2.1 Mice 	   IFNγ-/-, STAT4-/-, STAT6-/- and Rag1-/- mice (all on BALB/c background) were obtained from Jackson Laboratories.  These knockout mice were then crossed with wild type Eµ-RET mice, originally provided by Dr. Stephan A. Grupp (University of Pennsylvania), to generate stable gene-knockout colonies carrying a single copy of the Eµ-RET transgene.  Hemizygous Eµ-RET mice on an otherwise wild type BALB/c background were maintained by in-house breeding of Eµ-RET sires with BALB/c dams obtained from Jackson Laboratories.  A wild type C57BL/6 colony was maintained by in-house breeding from Charles River stock.  The NOD-scid/IL2Rγ null (NSG) mice used in the adoptive transfer studies were also obtained for Jackson Laboratories.  All mice were housed at the BC Children’s Hospital Research Institute animal care facility under specific pathogen-free conditions.  All experiments were performed in accordance with Canadian Council on Animal Care guidelines under a University of British Columbia Animal Care Committee-approved protocol (A15-0187). The Eµ-RET mouse represents a stable and reproducible two-stage model of late pro-B-cell malignancy.  In this model a fetal-derived abnormal B-cell population with temporal similarities to human in utero-generated pre-leukemic cells, eventually progresses to overt disease after an extended latency of 3-12 months.  This leukemia initiating cell (LIC) population is defined by a characteristic B220int/BP-1hi phenotype; these cells are readily detectable in the spleen, blood and bone marrow of transgenic mice months prior to disease onset, and are completely absent in wild type BALB/c mice (Figure 2.1).  A surrogate measure of LIC burden can be obtained by quantifying the 	   53 number of abnormal LIC present in the spleen of transgene positive mice at various time points.  The E2A-PBX1 mouse model also represents a two-stage model of B-cell precursor malignancy that is characterized by a prolonged and variable pre-leukemic phase (Duque-Afonso et al. 2015).  Bone marrow derived from pre-leukemic and leukemic E2A-PBX1 mice was used in both in vitro and in vivo experiments throughout this thesis to validate the results derived from the Eµ-RET mouse model.  The E2A-PBX1 transgene carries a green-fluorescence protein (GFP) tag, and so E2A-PBX1+ LICs can be identified by both their B-cell precursor phenotype as well as GFP expression.  For in vitro assays, E2A-PBX1+ LIC were FACS purified from bone marrow samples derived from pre-leukemic mice based on their B-cell precursor phenotype (B220int/BP-1hi) and GFP expression.  For in vivo experiments, bone marrow derived from overtly leukemic E2A-PBX1 mice was adoptively transferred into C57BL/6 recipient mice.  In vivo E2A-PBX1 leukemia burden analysis was performed by quantifying the number of leukemic cells in the spleen of recipient mice using flow cytometry (Figure 2.2).     	   54  Figure 2.1.  Flow cytometric analysis of LICs in the Eµ-RET mouse model Eµ-RET mice exhibit a prolonged and variable pre-leukemic phase.  LICs are generated in utero and disseminate throughout the body after birth.  These cells are defined by their characteristic B220int/BP-1hi phenotype and can be readily detected in the bone marrow, spleen and blood of transgene-positive mice.  We use absolute number of splenic LICs as a surrogate measure of systemic LIC burden in the Eµ-RET mouse.  Numbers indicate LIC proportion as percent of splenocytes.   Figure 2.2.  Flow cytometric analysis of E2A-PBX1 leukemia cell burden In all in vivo experiments cells from overtly leukemic E2A-PBX1 mice were adoptively transferred into wild type C57BL/6 recipients.  Splenic leukemia burden in recipient mice was assessed based on the B-cell precursor cell phenotype and GFP expression of E2A-PBX1 leukemia cells.  Numbers indicate leukemia cell proportion as a percent of total splenocytes.  	   55 2.2 Cells and cell culture 	   Leukemic and pre-leukemic E2A-PBX1 cells were derived from bone marrow samples isolated from pre-leukemic and overtly leukemic E2A-PBX1 mice (Duque-Afonso et al. 2015).  These bone marrow samples were generously provided by Dr. Michael Cleary (Stanford University).  Pre-leukemic and leukemic Eµ-RET-cells were derived directly from the spleen and/or bone marrow of Eµ-RET mice via fluorescence activated cell sorting (FACS).  Primary human ALL cells were obtained from the BC Children’s Hospital Biobank.  The human RS4;11 cell line (B-cell precursor leukemia derived from the bone marrow of a 32 year old female leukemia patient) was obtained from ATCC (Manassas, VA) (#CRL-1873).  Finally, the Eµ-RET derived 289 leukemia cell line was provided by Dr. Stephan Grupp.  All cells were cultured at 37 degrees Celsius with 5% CO2 in complete RPMI supplemented with 20% fetal bovine serum (FBS), non-essential amino acids (NEAA), HEPES buffer, 2-mercaptoethanol, L-glutamine and 250pg/mL IL-7.  Mouse IL-7 was obtained from Sigma Life Sciences (Catalog #I4892).   2.3 LIC burden assessment 	   Wild type and IFNγ-/- Eμ-RET mice were sacrificed at 14 days of age and whole splenocytes and bone marrow were extracted from these mice.  RBC lysis was performed using a single 6-minute incubation in Tris-Ammonium Chloride (TAC).  Cells were then washed in phosphate buffered saline (PBS) and re-suspended in 1mL of PBS with 2% fetal bovine serum (FBS).  5% of total splenocytes, and 10% of total bone marrow cells were stained for evaluation by flow flow cytometry using the following antibodies B220/CD45R (Clone: RA3-6B2; BioLegend), CD43 (Clone: S11; BioLegend), BP-1/Ly-	   56 5 (Clone: 6C3; BioLegend), 7-amino-actinomycin D.   LICs were identified based on their the characteristic B220int/CD43int/BP-1hi phenotype (see figure 2.1).  An absolute count of LICs in both the spleen and bone marrow was calculated using CountBright beads (Invitrogen; Carlsbad, CA).  The protocol described here was also used to assess mature B and T-cell counts in the spleen of Eμ-RET and wild type BALB/c mice.  In this case B-cells were defined by a B220hi/IgM+ phenotype and T-cells were defined by a CD3+/CD4 or CD8+ phenotype (All antibodies were purchased from BioLegend.  IgM, Clone: RMM-1; CD8a, Clone: 53-6.7; CD4, Clone: RM4-5; CD3, Clone: 17A2).  Unless otherwise stated, this general protocol was used in all subsequently described assessments of ex vivo LIC burden.  2.4 LIC flow sorting 	   Splenocytes were harvested from wild type and IFN-γ-/- Eμ-RET mice.  Splenocytes were processed (as above) and stained with B220, BP-1, and 7AAD for 30 minutes.  Cells were then washed with 10mL of PBS+2% FBS and filtered through a 50μM filter.  LICs (B220int/BP-1hi) were purified by fluorescence activated cell sorting using the FACS Aria.   2.5 Survival studies 	   Wild type and IFNγ-/- Eμ-RET mice were followed to disease onset.  Disease progression was monitored by assessment of LIC counts in peripheral blood.  Disease onset was defined by the presence of palpable lymph nodes or a peripheral blood WBC>15,000/μL.  	   57 2.6 In vivo proliferation assay 	   Wild type and IFN-γ-/- Eµ-RET mice were injected with 1mg of 5-bromodeoxyuridine (BrdU) on days 8 and 9 of age.  Spleen and bone marrow from these mice were harvested at day 10 and 11 and proliferation was assessed by examining the proportion of BrdU+ cells in the BP-1+ LIC population.  Flow cytometric analysis of BrdU incorporation was performed using the BD Pharminogen FITC BrdU Flow Kit obtained from BD Biosciences (San Jose, CA).  Results are normalized to the average proportion of BrdU+ cells in each respective wild type LIC compartment. 2.7 Suppressor of Cytokine Signaling 1 (SOCS1) expression analysis 	   SOCS-1 RNA and protein expression were assessed in LICs derived from wild type and IFN-γ-/- Eµ-RET mice.  LICs were purified from the spleen and bone marrow of 2-4 week old mice using FACS.  For RNA expression analysis, RNA was purified using the Arcturus PicoPure RNA isolation kit (Applied Biosystems; Foster City, CA) and first-strand cDNA synthesis was performed using qScript cDNA Supermix kit (QuantaBio; Beverly, MA). Quantitative RT-PCR was performed on a BioRad CFX96 thermo-cycler using the iTaq Universal SYBR Green Supermix Kit (BioRad; Hercules, CA) and SOCS-1 expression levels calculated as relative transcript abundance (RTA) normalized to β-actin using the delta-delta CT analysis method. The primers used were: ActbF: TTCTTTGCAGCTCCTTCGTT; ActbR: ATGGAGGGGAATACAGCC; SOCS1F: GTGGTTGTGGAGGGTGAGAT; SOCS1R: CCCAGACACAAGCTGCCA.  For SOCS-1 protein analysis, whole splenocytes were extracted from wild type and IFN-γ-/- Eµ-RET mice.  Cells were fixed in BD cytofix buffer for 15 minutes on ice.  Cells were then permeabilized in BD PermII buffer for 10 minutes at room temperature.  	   58 Permeabilized cells were then incubated with a polyclonal rabbit anti-SOCS1 antibody (Abcam, Cat# ab135718) for twelve hours.  After primary labelling, antibody-bound cells were incubated with donkey anti-rabbit IgG secondary antibody conjugated to the BV421 flurochrome (BioLegend; Clone: Poly4064) for 40 minutes on ice.  SOCS-1 expression in LICs was then assessed by flow cytometry. 2.8 IFN-γ sensitivity assay 	   LIC and leukemic cells were purified from wild type and IFNγ-/- Eµ-RET mice.  LICs were FACS-purified while leukemic cells were derived directly from the spleens of moribund mice.  5x104 LICs or leukemic cells were cultured in RPMI with 250pg/mL of IL-7 for 72 hours in the presence of IFN-γ at concentrations ranging from 0 to 10 units/mL.  Sensitivity to IFN-γ was assessed by calculating the viable cell number after 72 hours via flow cytometry using 7-AAD and countrbight beads.  Recombinant murine IFN-γ was obtained from PeproTech (Rocky Hill, NJ) (Catalog #315-05).   2.9 LIC adoptive transfer into NSG mice 	  3x105 FACS purified LICs derived from wild type and IFNγ-/- Eμ-RET were adoptively transferred via tail-vein injection into NSG recipient mice.  All recipient mice were pre-treated with 200μg of IFN-γ (BioXcell; Clone: R4-6A2) neutralizing antibody on day 0 and IFN-γ neutralization was maintained by additional 100μg injections of the neutralizing antibody every 3 days.  LIC burden, BrdU incorporation and SOCS-1 RNA expression were assessed 14 days after adoptive transfer.   	   59 2.10 In vivo IFN-γ sensitivity 	    To assess the in vivo sensitivity of leukemic cells to IFN-γ, leukemic cells were harvested from the spleens of moribund IFN-γ-/- Eµ-RET mice.  104 leukemic cells were adoptively transferred into wild type and IFN-γ-/- BALB/c mice via intravenous injection.  For burden analysis, recipient mice were euthanized two weeks later and splenic leukemia burden was assessed and compared between the two recipient mice genotypes.  For survival analysis, recipient mice were followed to disease onset.  Disease onset was defined by the presence of palpable lymph nodes or a white blood cell count (WBC) exceeding 15,000/uL.  To assess the in vivo sensitivity of LICs to IFN-γ, LICs were purified from the spleens of four-week old IFN-γ-/- Eµ-RET mice by FACS.  2x105 LICs were adoptively transferred into wild type and IFN-γ-/- BALB/c mice.  Recipient mice were euthanized 45 days after adoptive transfer and LIC burden in the spleen and bone marrow of recipient mice was assessed using flow cytometry. 2.11 IL-7 dependence assay 	   FACS purified LICs and leukemic cells were obtained from healthy and moribund wild type Eµ-RET mice, respectively. 5x104 LICs or leukemic cells were cultured in combination with 5x105 stromal cells obtained from the bone marrow of NSG mice in RPMI supplemented with 250pg/mL of IL-7.  Cells were cultured with and without IL-7R blocking antibody (10ug/mL; BioXcell; Clone: A7R34) for 72 hours.  After 72 hours viable cell number was assessed using flow cytometry.  IL-7 dependence was determined by calculating the total number of viable cells in culture after 72 hours and comparing it 	   60 to the original number of cells plated to generate a ratio (cells recovered/5x104 cells plated) 2.12 SOCS-1 RNA induction assay 	   FACS purified LICs and leukemic cells were isolated from wild type and IFN-γ-/- Eµ-RET mice.  5x105 cells were incubated in 200µL of serum derived from wild type and IFN-γ-/- BALB/c mice for 2 hours.  Post incubation, RNA was extracted using the Arcturus PicoPure RNA isolation kit and qPCR analysis was performed following the protocol previously described in section 2.5. 2.13 Direct effect of TLR ligands on LICs 	   The following TLR ligands were obtained from InvivoGen (San Diego, CA):  ODN 1826, a Class B CpG oligonucleotide (henceforth referred to as CpG) (Catalog #tlrl-1826):  R848, Imidazoquinoline compound (henceforth referred to as R848) (Catalog #tlrl-r848); and Poly(I:C)/LyoVecTM (henceforth referred to as pIC) (Catalog #tlrlpiclv).  5x104 FACS purified LICs derived from wild type Eµ-RET mice were cultured for 48 hours with or without the addition of TLR ligands.  The concentration of TLR ligands used was as follows: CpG 10μg/mL, R848 10μg/mL, pIC 50μg/mL.  Direct impact of TLR activation on LIC survival was assessed via flow cytometry using CountBright beads.  2.14 The contribution of immune-mediated effects of TLR ligands on LICs 	   To assess the indirect effects of TLR ligands on LIC survival, whole splenocytes were harvested from wild type Eµ-RET mice.  2x105 bulk splenocytes were cultured with 	   61 or without TLR ligands (at the above indicated concentration) for 48 hours.  The indirect effect of TLR ligation on LICs was determined by assessing the viable number of LICs in the culture after 48 hours by flow cytometry using CountBright beads. 2.15 Myeloid effector cell assay 	   5x104 FACs purified LICs derived from Eμ-RET and E2A-PBX1 mice were cultured with 5x105 bone marrow effector cells derived from NSG mice with or without TLR ligands (at the above indicated concentration) for 48 hours.  The myeloid effector cell-mediated indirect effect of TLR activation on LICs were determined by assessing the viable number of LICs in culture after 48 hours by flow cytometry using CountBright beads. 2.16 Normal and abnormal B-cell precursor assay 	   3x105 bone marrow cells derived from wild type BALB/c or Eμ-RET mice were cultured with or without TLR ligands (at the above indicated concentrations) for 48 hours.  To assess the impact of type 1 and 2 interferon to the depletion mechanism, 25μg/mL of IFN-γ-neutralization and IFNαR1-blocking (BioLegend; Clone: MAR1-5A3) antibodies were added to the culture.  The effect of TLR ligation on both LIC and normal B-cell precursors (BCP) was determined by assessing the viable number of LICs and BCPs in culture after 48 hours by flow cytometry using CountBright beads.  Normal B-cell precursors in BALB/c mice were defined by a B220int/CD43int/CD24+/BP-1hi phenotype.   	   62 2.17 Conditioned supernatant assay 	   To assess the impact of cytokine-mediated activity on LICs, 3x105 bone-marrow effector cells derived from BALB/c mice were cultured with and without TLR ligands (at the above indicated concentrations) for 24 hours.  After incubation, cultures were centrifuged and filtered (0.2μM) to remove all cells.  The supernatants from these cultures were then pre-treated with 25μg/mL of IFN-γ and/or IFNαR1 antibodies for 20 minutes at 370 Celcius.  5x104 FACS purified LICs derived from Eμ-RET mice were cultured in this pre-treated supernatant for 48 hours.  The direct effect of TLR ligand-induced type 1 and type 2 IFN production on LICs was determined by assessing the viable number of LICs in culture after 48 hours by flow cytometry using CountBright beads. 2.18 Human leukemia cell assay 	   PBMCs were isolated from whole blood samples from healthy adult volunteers using ficoll.  3x105 human PBMCs were cultured with and without TLR ligands (at the above indicated concentrations) for 24 hours.  After incubation cultures were centrifuged and filtered (0.2μM) to remove PBMCs.  Supernatants were pre-treated with 25μg/mL IFN-γ and/or IFNαR1 antibodies for 20 minutes at 370 C.  105 RS4;11 or primary human leukemia cells were cultured in the pre-treated supernatant for 48 hours.  The impact of TLR-induced IFN production on these cells was determined by assessing the viable number of leukemic cells in culture after 48 hours using flow cytometry with CountBright beads.  	   63 2.19 In vivo TLR ligand administration 	   4-6-week old wild type and IFN-γ-/- Eμ-RET mice received 100μg of CpG via intraperitoneal injection once a week over a four week treatment period.  One week after the final treatment mice were sacrificed and LIC burden in the spleen and bone marrow was assessed by flow cytometry.  For survival analysis, treated and untreated (PBS) mice were followed to disease onset.  Disease progression was monitored by LIC count in peripheral blood and disease onset was defined by the presence of palpable lymph nodes or a WBC>15,000/μL in peripheral blood. 2.20 Infections 	   Three infection models were used in this thesis: Listeria monocytogenes (Lm), murine-cytomegalovirus (MCMV) and murine-gammaherpesvirus-68 (MHV-68).  For both MCMV and MHV-68, live, non-attenuated virus strains were used.  For Lm, a live ΔtrpS;ActA attenuated strain was used.  The trpS attenuation introduces a metabolic defect that renders this Lm strain incapable of producing tryptophan.  Additionally, the ActA attenuation in the strain eliminates the capacity for cell-to-cell spreading.  Furthermore, this ActA attenuation ensures that this Lm strain is only capable of infecting phagocytic cells, as this mutation eliminates the bacterium’s ability to infect non-phagocytic cells. All mice were treated at day 6 (neonatal) or day 34 (adult) of age.  For Lm infection, neonatal mice received 1x103 or 1x104 cfu of trpS and ActA Lm, while adult mice received 1x104 cfu.  Lm stocks were obtained from Dr. Tobias Kollman (University of Biritsh Columbia).  For MCMV infection, neonatal mice received 1x102 pfu while adults received 1x103 pfu; MCMV stocks were provided by Dr. Soren Gantt (University 	   64 of British Columbia).  For MHV-68 infection, neonatal mice received 4x102 pfu while adults received 4x103 pfu; MHV-68 stocks were provided by Dr. Soren Gantt (University of British Columbia). Lm and MCMV infections were administered via intra-peritoneal (i.p.) injection while MHV-68 was administered intra-nasally without anesthesia.  Mock-infected mice received matched volumes of saline administered intra-nasally or via i.p injection.   2.21 Infection responses in Eµ-RET mice 	   For evaluation of LIC populations, spleen and bone marrow were harvested from Eμ-RET mice at 2 and 6 weeks of age (8 days post-infection). After red blood cell lysis and washing, LIC numbers were calculated by flow cytometry with CountBright beads (Invitrogen, Carlsbad, CA).  For survival analyses, infected or mock infected (saline) mice were followed to disease onset.  Disease onset was defined by the presence of palpable lymph nodes or WBC counts of >15,000/ml. The B-cell precursor phenotype (CD43int/B220int/BP-1+) of the expanded cell population was confirmed in all cases by flow cytometry. 2.22 Adoptive transfer infection studies 	   5x103 289 cell line and primary E2A-PBX1 leukemic cells, suspended in total volume of 30μL PBS, were adoptively transferred into two-day-old neonatal mice via superficial temporal vein injection, using G30 syringes with visualization of vascular anatomy by trans-illumination (performed by Dr. Abbas Fotovati). Recipient mice were then infected or mock infected at day 6 of age and sacrificed 8 days later; adoptively transferred cell burden was then assessed using flow cytometry with CountBright beads (289=B220int/BP-1hi; E2A-PBX1= BP-1hi/GFP+). Mice that received E2A-PBX1 leukemic 	   65 cells were also followed for time to disease onset. Disease was defined by the presence of palpable nodes or a peripheral blood WBC count exceeding 15,000/μL. For adult mice (30 days old), 5x103 289 or E2A-PBX-1 cells were adoptively transferred via tail-vein injection and mice were infected 4 days later. All adult recipients were sacrificed at day 42 of age and splenic LIC burden was assessed using flow cytometry. All graphs are a representation of the proportion of cells in infected mice relative to the average for mock-infected mice receiving the same leukemia cells. 2.23 Antibody-based cytokine blockade and cell depletion 	   Cytokine neutralization antibodies (IL-12p75, IL-23p19, IL-17A) were administered concurrently with infection in 200μg doses via i.p injection.  Cell depletion antibodies (CD8a, CD4, TCR γ/δ) were administered two days prior to infection.  Burden was assessed eight days after infection by flow cytometry using CountBright beads.  Antibodies used for in vivo cell depletion or cytokine neutralization were obtained from either BioXcell: CD4 (Clone: GK1.5), CD8 (Clone: 53-6.72), IL-12/23p40 (Clone: C17.8), IL-17A (Clone: 17F3), γδ TCR+ (UC7-13D5), IL-12p75 (Clone: R2-9A5); or BioLegend: IL-23p19 (Clone: MMp19B2). 2.24 Cytokine analysis 	  Serum was collected from 7 and 35-day-old mice 18 hours after mock/infection. Serum samples were analyzed using the MSD “U-Plex” platform (Meso Scale Diagnostics, Rockville, MD).  Graphs represent the fold induction and pg/mL of IL-17A.   	   66 2.25 Statistical methods 	  Kaplan-Meier curves were generated for spontaneous and leukemic cell adoptive transfer survival studies and were analyzed by log-rank tests.  Analyses of all in vitro cell survival assays and all in vivo burden analyses with Eμ-RET LICs/leukemic cells, 289 cell line and E2A-PBX1+ pre-leukemic/leukemic cells in both the spontaneous and adoptive transfer experiments were performed using Mann-Whitney test or a one-way ANOVA with Bonferroni post-hoc analysis in any case where more than two groups were being compared.  Two-way ANOVAs were used any time more than one variable was involved (eg in vitro IFN-γ sensitivity assay). Statistical analyses were performed using Prism 5 for Mac OS X (GraphPad, San Diego, CA).  Specific n values for each experiment are listed in figure legends.   All graphs represent at least two independent repeats.   2.26 Services 	   Flow cytometry work was conducted at the BC Children’s Hospital Research Institute (BCHRI) Flow Core Facility.   Data acquisition was performed on the FACSCalibur, the LSRFortessa X-20, and the LSR-II.  Fluorescence activated cell sorting was conducted on the FACSAria by Lixin Xu.  All post-acquisition analysis was performed using FlowJo software (Tree Star)     	   67 Chapter 3: Basal IFN-γ production delays disease onset in the Eµ-RET mouse 	  3.1 Overview and rationale 	   Polymorphisms in several cytokine genes, including IFN-γ, have been associated with increased risk or reduced latency of childhood B-cell ALL (Chang et al. 2010; Cloppenborg et al. 2005; Han et al. 2010; Wiegering et al. 2014).  However, the mechanism underlying these associations remains unknown.  While it has been proposed that these polymorphisms may represent underlying immune response defects that promote dysregulated immune responses that may be pro-leukemic (Rudant et al. 2015), experimental support for this hypothesis is currently lacking.  Polymorphisms in cytokine genes can influence the production of the cytokines encoded by these genes both basally and in response to infection exposure.  It is therefore possible that basal or abnormal BCP cell-induced cytokine production may also contribute to anti-ALL mechanisms in the absence of infection.  While this possibility has not been explored experimentally, the association between low IL-10 protein levels at birth and increased B-cell ALL risk (Chang et al. 2011) may support this hypothesis.  This association suggests that in the absence of infection, early-life immune development may impact subsequent disease progression.  The mechanism of this influence similarly remains unknown; however, several observations suggest that factors that influence the survival or expansion of pre-leukemic cells may have a significant impact on B-cell ALL development.  IFN-γ has also been identified as a critical mediator of anti-tumor immune surveillance mechanisms in spontaneous and chemical-induced solid tumor models (Bui et al. 2012; Dunn et al. 2005; Kaplan et al. 1998; Shankaran et al. 2001).  While IFN-γ 	   68 has been demonstrated to act through innate and adaptive immune mechanisms to prevent cancer progression, a common observation in these models is that malignant cells derived from IFN-γ-/- mice are more immunogenic than their wild type counterparts when adoptively transferred into wild type secondary hosts.  This observation has been attributed to a lack of IFN-γ mediated immune editing in the original host.  The immunoediting model suggests that immune effector mechanisms drive the selection of tumor cell variants with reduced immunogenicity (Dunn et al. 2002; Dunn et al. 2006; Kim et al. 2007; Schreiber et al. 2011).  Malignant cells that developed in IFN-γ-/- mice were never exposed to the inhibitory influence of IFN-γ, and therefore, resistance to IFN-γ-mediated inhibitory immune mechanisms was never selected for; when these cells are adoptively transferred into wild type recipients, they remain sensitive to IFN-γ and are often either rejected or demonstrate inhibited growth.  While the role of IFN-γ in immune surveillance against solid tumors is consistent with the immunoediting model, its role in hematological malignancies does not conform to this model and remains poorly understood.  It has been demonstrated that lymphoma incidence was higher in IFN-γ-/- deficient mice as compared to their wild type counterparts.  However, unlike the results with solid tumors, these lymphoma cells did not display any indication of increased immunogenicity when transplanted into wild type secondary recipients (Street et al. 2002).  These findings indicate that while IFN-γ is capable of impacting the progression of hematological malignancies, the mechanism of this influence may be distinct from those involved in solid tumor models    These observations formed the basis for my investigation into the role of basal IFN-γ in leukemia progression in the Eµ-RET mouse.  IFN-γ has been demonstrated to be 	   69 a potent inhibitor of B-cell precursor cell proliferation through a direct mechanism that involves the induction of a regulatory protein called suppressor of cytokine signaling-1 (SOCS-1) (Corfe et al. 2011).  IFN-γ may, therefore, be a factor that is relevant to the survival and/or proliferation of maturation-arrested LICs or overt leukemic cells.  This raises the possibility that in contrast to its immune-mediated impact on solid tumors, IFN-γ may act through a direct mechanism to inhibit the development or progression of B-cell leukemia that demonstrates maturational arrest at an IFN-γ-sensitive developmental stage.  As the subsequently detailed work will illustrate, I found that IFN-γ is capable of delaying disease onset by restricting the proliferation of LICs via induction of elevated basal expression of SOCS-1.  Additionally, my results demonstrate that overt malignant transformation is associated with reduced sensitivity to basal levels of IFN-γ in vivo. 3.2 IFN-γ reduces leukemia-initiating cell burden and delays disease onset 	   Detection of ETV6-RUNX1+ pre-leukemic blasts on the Guthrie cards of approximately 1% of all newborn children, while the incidence of this particular leukemia sub-type is approximately 100-fold lower, suggests that the development of leukemia is not an inevitable outcome (Mori et al. 2002).  Furthermore, it has been observed that the pre-leukemic cell “load” at birth may be associated with time to disease onset (Hjalgrim et al. 2002; Taub et al. 2002).  These observations indicate that factors that influence the survival and/or expansion of the pre-leukemic cell population may have a significant impact on both ALL risk and progression.  To determine if basal IFN-γ was capable of influencing disease progression in the Eµ-RET mouse, I assessed whether the presence of IFN-γ had an influence on early-life LIC burden by comparing the splenic LIC burden in two-week old wild type and IFN-γ-/- 	   70 Eµ-RET mice.  The absolute number of LICs in the spleen of IFN-γ-/- Eµ-RET mice significantly exceeded that of their wild type counterparts (Figure 3.1).  As IFN-γ has been demonstrated to be a potent inhibitor of B-cell precursor proliferation, I assessed whether this differential LIC burden was associated with an IFN-γ-dependent effect on LIC proliferation in the spleen and bone marrow of wild type Eµ-RET mice.  This was accomplished by comparing the in vivo incorporation of BrdU in LICs in wild type and IFN-γ-/- Eµ-RET mice.  The proportion of actively proliferating LICs was significantly greater in both the spleen and bone marrow of IFN-γ-/- Eµ-RET mice as compared to wild type LICs (Figure 3.2).  Finally, in order to assess the impact of this IFN-γ influence during early LIC population growth on overall disease progression, I compared disease-free survival between the two genotypes.  IFN-γ-/- Eµ-RET had significantly reduced disease latency compared to their wild type counterparts (118 vs. 148 Days) (Figure 3.3).              	   71          Figure 3.1.  LIC burden is higher in IFNγko Eµ-RET mice Splenic LIC burden in two-week old wild type and IFNγko Eµ-RET was compared using flow cytometry based on the characteristic B220int/CD43int/BP-1hi phenotype of LICs.  N= 10 mice from each genotype.  Mann-Whitney, bars represent mean±S.D.; **p<0.01.              	   72          Figure 3.2.  IFNγ restricts LIC proliferation in vivo In vivo proliferation of LICs in the spleen and bone marrow of wild type and IFNγko Eµ-RET mice was assessed by measuring BrdU incorporation over a 48-hour period.  The percentage of BrdU+ LICs in each mouse was normalized to the wild type average from each cell compartment.  N= 8 wild type and 7 IFN-γko Eµ-RET mice.  Mann-Whitney, bars represent mean±S.D.; *p<0.05, ****p<0.0001.             	   73         Figure 3.3.  IFN-γ influences disease kinetics in the Eµ-RET mouse Disease-free survival compared in wild type (n=70) and IFN-γko Eµ-RET mice (n=58).  Overt disease was defined by the presence of palpable lymph nodes or a peripheral blood WBC>15,000/µL. Median survival = 118 days (IFN-γko) and 148 days (wild type). Log-rank.  3.3 Normal B-cell precursors are also affected by basal IFN-γ 	   In the spleen, cells with the Eµ-RET LIC phenotype are a distinct population with no counterpart in transgene-negative mice (See Figure 2.1).  However, in the bone marrow it is difficult to phenotypically discriminate between abnormal LICs and normal B-cell precursors at the same developmental stage.  To assess whether the observed influence of IFN-γ was unique to LICs or shared by normal B-cell precursors, I compared splenic B-cell numbers in wild type and IFN-γ-/- Eµ-RET mice.  I used mature B-cell numbers as a surrogate measure of B-cell precursor proliferation.  In Eµ-RET mice, the absence of IFN-γ had no effect on splenic T-cell (CD3+) numbers; however, the number of mature B-cells (B220 high/IgM+) was significantly higher in IFN-γ-/- than in wild type Eµ-RET mice (Figure 3.4A).  To determine if this influence was related to abnormally elevated IFN-γ production in response to the presence of LICs, I assessed the 	   74 impact of IFN-γ on splenic B-cell numbers in wild type and IFN-γ-/- BALB/c mice.  Consistent with our findings in the Eµ-RET mouse, IFN-γ-/- BALB/c mice had a significantly higher number of mature B-cells than their wild type counterparts (Figure 3.4B).  These results indicate the anti-proliferative influence of IFN-γ is relevant in both abnormal LICs and normal B-cell precursors.  Furthermore, it confirms the assertion that the observed influence is related to basal IFN-γ production rather than elevated production in response to the presence of LICs.          A.               B.  Figure 3.4.  Normal B-cell precursors are also affected by basal IFN-γ Splenic T and B-cell numbers were compared in IFNγko Eµ-RET mice using flow cytometry. (n=6 wild type and 7 IFNγko Eµ-RET mice)(A).  T-cells were defined by the expression of CD3, while mature B-cells were defined by the combined expression of B220 and IgM.  Mature B-cell numbers were also assessed in the spleens of wild type and IFNγko BALB/c mice (n=5 wild type and 6 IFNγko) (B).  Mann-Whitney, bars represent mean±S.D.; **p<0.01.   3.4 IFN-γ induces SOCS-1 expression 	   IFN-γ has previously been demonstrated to potently inhibit the proliferation of B-cell precursors (Garvy & Riley 1994).  SOCS-1, which is induced by IFN-γ and subsequently inhibits JAK/STAT signaling (Corfe et al. 2011), was found to be the 	   75 primary mediator of this inhibitory effect. SOCS-1 has two overlapping activities in B-cell precursors: first, it works as a negative feedback mechanism to oppose IFN-γ signaling; and secondly, it inhibits IL-7 signaling, which is also mediated by the JAK/STAT pathway and is necessary for the survival and proliferation of B-cell precursors (Corfe et al. 2011; Corfe & Paige 2012).  This mechanism was found to be relevant in vivo, as SOCS-1-/- mice demonstrate significant early-life B-cell compartment defects.  B-cell counts in these mice could be restored when IFN-γ signaling was blocked; indicating that a lack of SOCS-1-mediated control of IFN-γ signaling was responsible (Starr et al. 1998).    To assess if a related mechanism was involved in the influence of basal IFN-γ on LICs in the Eµ-RET mouse, I measured ex vivo SOCS-1 RNA and protein expression in LICs derived from wild type and IFN-γ-/- Eµ-RET mice. Both SOCS-1 RNA (Figure 3.5A) and protein (Figure 3.5B) expression was significantly higher in LICs derived from wild type Eµ-RET mice as compared to their counterparts in IFN-γ-/- Eµ-RET mice.                   	   76 A.               B.  Figure 3.5.  Basal IFN-γ induces higher resting expression of SOCS-1 in LICs SOCS-1 RNA and protein expression were measured in LICs derived from wild type and IFNγko Eµ-RET.  SOCS-1 RNA expression was assessed using qRT-PCR.  SOCS-1 RNA expression was normalized to that of β-actin (n=4 mice from each genotype) (A).  SOCS-1 protein expression was assessed using intracellular flow cytometry (n=6 mice from each genotype (B).  Mann-Whitney, bars represent mean±S.D.; **p<0.01.   3.5 LIC expansion is directly inhibited by IFN-γ 	   The higher expression of SOCS-1 induced by basal IFN-γ in wild type LICs indicates that direct effects on LIC proliferation may underlie the reduced LIC burden observed in wild type Eµ-RET mice relative to their IFN-γ-/- counterparts. To determine whether wild type and IFN-γ-/- derived LICs exhibit differential sensitivity to the direct effects of IFN-γ, I assessed the impact of IFN-γ on these cells in vitro.  Consistent with a lack of prior exposure to such direct inhibitory activity, purified LICs derived from IFN-γ-/- Eµ-RET mice were more sensitive to the suppressive effects of exogenously administered IFN-γ than their wild type counterparts (Figure 3.6).   However, given the previously demonstrated role of IFN-γ in the immune editing of cancer cells (Bui et al. 2012; Dunn et al. 2005; Shankaran et al. 2001), it is also possible that indirect, immune-mediated influences may contribute to the reduced LIC 	   77 burden and delayed disease onset observed in vivo.  To test this and to evaluate the role of IFN-γ-induced changes in intrinsic LIC behavior in vivo in the absence lymphocytes and other IFN-γ-mediated immune editing mechanisms, I adoptively transferred equal numbers of FACS-purified LICs derived from wild type and IFN-γ-/- Eµ-RET mice into NSG mice.  I maintained an IFN-γ-deficient environment in these mice through the use of IFN-γ neutralizing antibodies and assessed LIC burden, proliferation and SOCS-1 expression after two weeks.  Mice that received LICs derived from IFN-γ-/- Eµ-RET donors had significantly higher splenic burden than those that received wild type derived LICs (Figure 3.7A).  This expansion was associated with increased LIC proliferation in the bone marrow of recipient mice (Fig 3.7B) and reduced SOCS-1 expression (Figure 3.7C).  These results demonstrate a direct impact of IFN-γ on LIC survival, and indicate that the presence of IFN-γ may influence LIC proliferation in vivo through the establishment of a durable SOCS-1 rheostat.            	   78                          Figure 3.6.  Wild type LICs are less sensitive to exogenous IFN-γ Sensitivity to exogenously administered IFN-γ was measured in vitro.  FACS-purified LICS derived from wild type and IFNγko Eµ-RET mice were cultured for 72 hours with the indicated concentration of IFN-γ.  Calculating the number of viable LICs in culture after 72 hours via flow cytometry using CountBright beads assessed sensitivity to IFNγ.  N=6 mice from each genotype, pooled from 3 independent experiments.  Two-way ANOVA, bars represent mean±S.D.                       	   79    A.        B.                   C.              Figure 3.7. IFN-γ causes durable changes in SOCS-1 expression and proliferation An equal number of wild type and IFNγko Eµ-RET derived LICs were adoptively transferred into NSG recipient mice.  Two weeks after LIC adoptive transfer into NSG mice, splenic burden was assessed in recipient mice using flow cytometry (A).  LIC proliferation was assessed by measuring BrdU incorporation over a 48-hour period (B).  SOCS-1 RNA expression was assessed in adoptively transferred LICs by qRT-PCR.  SOCS-1 RNA expression was normalized to that of β-actin (C).  Mann Whitney, bars represent mean±S.D; *p<0.05, **p<0.01.  3.6 IFN-γ-mediated inhibitory activity is restricted to the pre-leukemic phase 	   IFN-γ has been demonstrated to be a key mediator of immune surveillance in a number of solid tumor models (Bui et al. 2012; Dunn et al. 2005; Kaplan et al. 1998).  A principle observation in these models is that malignant cells derived from IFN-γ-deficient mice are generally more immunogenic than their wild type counterparts.  This increased 	   80 immunogenicity is revealed by their rejection or significantly slower growth when transplanted into wild type recipients as a result of first-time exposure to IFN-γ.  Assessing the immunogenicity of malignant cells arising in immune deficient hosts in an adoptive transfer model has become the definitive immunoediting experiment.  However, the results from such experiments with hematological malignancies are less clear.  It has been demonstrated that while lymphoma incidence was increased in IFN-γ-deficient mice, the malignant cells arising in these mice did not exhibit any indication of increased immunogenicity (Street et al. 2002). To assess whether the IFN-γ-mediated inhibitory effect observed in this study was consistent with the immunoediting model, I adoptively transferred overt leukemia cells derived from IFN-γ-/- Eµ-RET mice into wild type and IFN-γ-/- BALB/c recipients, and compared the expansion of these cells in each setting.  Early leukemic burden was assessed in these recipient mice 14 days after adoptive transfer (Figure 3.8A).  No difference in early leukemic cell burden was observed between wild type and IFNγ-/- recipients, suggesting that leukemic cells derived from IFN-γ-/- Eµ-RET mice did not exhibit increased immunogenicity in wild type mice as a result of a lack of IFN-γ-mediated immune editing in the primary host.  This finding was confirmed by comparing disease progression in wild type and IFN-γ-/- recipient mice.  No significant difference in disease kinetics was observed between wild type and IFN-γ-/- recipient mice (Figure 3.8B).  These findings do not conform with the postulates of the immunoediting model and support the previous assertion that the IFN-γ-dependent delay in disease kinetics was not mediated by an indirect immune surveillance mechanism.  These findings are in stark contrast to those obtained when a similar experiment was performed using LICs derived from IFN-γ-/- Eµ-RET mice.  These LICs were adoptively 	   81 transferred into wild type and IFN-γ-/- BALB/c recipients, and splenic LIC burden in recipient mice was assessed 45 days later.  In contrast to the results obtained with overt leukemic cells, IFN-γ-/- recipients had significantly higher LIC burden than their wild type counterparts (Figure 3.9).  Collectively, these results indicate that LICs but not overt leukemic cells are sensitive to basal IFN-γ in vivo.                 A      B.               Figure 3.8. In vivo IFN-γ did not impact leukemic cell outgrowth Leukemic cells derived from IFNγko Eµ-RET mice were adoptively transferred into wild type and IFNγko BALB/c recipients; splenic leukemic cell burden was assessed after two weeks by flow cytometry using CountBright beads.  Cell counts were normalized to the wild type average in each experiment (A).  Survival of IFNγko leukemia recipient mice was assessed (n=6 different leukemias) (B). N= 6 IFNγko leukemia samples.  Bars indicate mean±S.D.; **p<0.01. Log-Rank.             IFNγ+/+ IFNγko0.00.51.01.52.02.5Recipient GenotypeProportion of cellsns	   82                  Figure 3.9.  In vivo IFN-γ significantly inhibits LIC outgrowth FACS purified LICs derived from IFNγko Eµ-RET mice were adoptively transferred into wild type and IFNγko BALB/c recipient mice.  Splenic LIC burden in these mice was assessed 45 days after adoptive transfer by flow cytometry using CountBright beads.  Cell counts were normalized to the wild type average in each experiment.  N=4 pooled IFNγko LIC samples adoptively transferred into 9 mice from each genotype.  Mann-Whitney, bars represent mean±S.D.; **p<0.01.   3.7 Mechanisms of IFN-γ insensitivity in leukemic cells 	   Basal IFN-γ levels are capable of inhibiting the proliferation of LICs in vivo.  However, our results indicate that overt transformation, even in the absence of previous exposure to IFN-γ, renders leukemic cells insensitive to endogenous IFN-γ in vivo.  Our model in LICs suggests that IFN-γ is capable of modulating LIC proliferation by inducing expression of SOCS-1, which inhibits IL-7 signaling.  IL-7 is a required growth factor for both LICs and normal pro- and pre-B-cells; in the absence of IL-7 signaling, proliferation in these cells will be arrested and apoptosis will eventually be initiated (Allman & Miller 2003; Corfe et al. 2011; Corfe & Paige 2012; Fleming & Paige 2002;).  As self-sufficiency in growth factor signaling is a hallmark of cancer (Hanahan & IFNγ+/+ IFNγko051015Recipient GenotypeProportion of cells**	   83 Weinberg 2000), I hypothesized that overt leukemic cells may have acquired IL-7 independence while LICs maintain their dependence on IL-7 for both survival and proliferation.  If this is the case, overt transformation may indirectly reduce IFN-γ sensitivity by rendering leukemic cells less sensitive to the inhibitory action of SOCS-1.  To assess this, overt leukemic cells and LICs were cultured for 72 hours with or without IL-7R blocking antibody; cell survival and/or proliferation in the absence of IL-7 signaling would be indicative of self-sufficiency in growth factor signaling.  By calculating the number of viable cells recovered after 72 hours in the presence of IL-7R blocking antibody and dividing it by the number of cells initially plated (5x104), I obtained a ratio that was indicative of IL-7 independence.  While LICs demonstrate complete dependence on IL-7 for survival, leukemic cells showed markedly increased growth-factor independence, with a dependence ratio of at least 1.0 obtained for all leukemia samples tested (Figure 3.10).    An alternative mechanistic explanation for the acquired IFN-γ insensitivity of overt leukemic cells is that while LICs upregulate SOCS-1 expression in response to IFN-γ, overt leukemic cells may have lost the ability to do so.  The rationale for such a loss is that SOCS-1 is induced by a number of different cytokines, some of which may also have a negative impact on the survival and/or expansion of pre-leukemic or leukemic cells.  SOCS-1 induction may therefore represent a functional hurdle to disease progression.  To assess whether this mechanism may be a factor in the Eµ-RET mouse, I measured SOCS-1 RNA induction in LICs and leukemic cells after incubation in serum obtained from untreated wild type mice.  LICs and leukemic cells derived from both wild type and IFN-γ-/- Eµ-RET mice were incubated in wild type and IFN-γ-/- serum for two 	   84 hours.  Incubation in serum derived from wild type mice is the most accurate way to assess the impact of basal concentrations of IFN-γ on SOCS-1 RNA expression.  While incubation in wild type serum induced a two-fold induction of SOCS-1 RNA in LICs derived from IFN-γ-/- deficient mice (Figure 3.11A), overtly leukemic cells from neither wild type nor IFN-γ-/- mice demonstrated any IFN-γ-induced increase of SOCS-1 expression (Figure 3.11B).  These results indicate that while basal concentrations of IFN-γ are capable of inducing SOCS-1 RNA expression in LICs, similar induction is not observed in overtly leukemic cells.  This impaired induction of SOCS-1 by basal concentrations of IFN-γ in leukemic cells may underlie the observed differential sensitivity of LICs and leukemic cells to IFN-γ in vivo.                                   Figure 3.10. Mechanisms of reduced IFN-γ sensitivity in leukemic cells LICs and overly leukemic cells were cultured for 72 hours in the presence of IL-7R blocking antibody.  A ratio was obtained by dividing the number of viable cells recovered after 72 hours by the number of cells initially plated (5x104).  Mann-Whitney, bars represent mean±S.D.; ****p>0.0001 	   85  A.             B.  Figure 3.11.  Alternative mechanisms of reduced IFN-γ sensitivity in leukemic cells SOCS-1 RNA expression in LICs (A) and overt leukemic cells (B) derived from wild type and IFNγko Eµ-RET mice, was assessed after two hour incubation in serum derived from wild type and IFNγko donors.  SOCS-1 RNA expression was normalized to that of β-actin.  N=3 wild type leukemias, 4 IFN-γ-ko leukemias, and 4 LIC samples from each genotype pooled from 3 independent experiments.  Bars represent mean±S.D.  3.8 Discussion 	   Polymorphisms in several cytokine genes have been associated with an increased risk or reduced latency of B-cell ALL (Chang et al. 2010; Cloppenborg et al. 2005; Han et al. 2010; Wiegering et al. 2014).  The mechanism underlying this association remains unknown, however it has been suggested that these polymorphisms may promote the development of abnormal, potentially pro-leukemic immune responses (Rudant et al. 2015).  These polymorphisms may alter the production of cytokines during an immune response and such changes may significantly alter the balance or nature of immune responses.  However, while this hypothesis may be correct, polymorphisms in cytokine genes likely also effect the basal production of cytokines in the absence of an infection.  It is therefore possible that basal or malignant cell-induced cytokine production may contriubte to anti-ALL mechanisms.  While this mechanism has not been investigated 	   86 experimentally, some observations suggest that it may be relevant.  For example, it has been demonstrated that lower IL-10 protein levels at birth are associated risk of B-cell ALL (Chang et al. 2011).  These low IL-10 levels have also been interpreted as evidence of dysregulated immune function that may facilitate the development of abnormal immune responses later in life.  However, it is also possible that this association could be interpreted as an indication that early-life immune development, in the absence of infection, may impact ALL development.  It is therefore possible that defective cytokine expression may impact ALL risk through multiple mechanisms. The results outlined in this chapter reveal a direct effect of IFN-γ activity on early-life expansion of LICs that is responsible for the earlier leukemia onset in IFN-γ-/- Eµ-RET mice.  Several of my findings, specifically lower SOCS-1 expression in LICs derived from IFN-γ-/- Eµ-RET mice and the up-regulation of SOCS-1 RNA after incubation of IFN-γ-/- derived LICs in wild type serum, indicate that basal IFN-γ levels are sufficient for the IFN-γ-mediated inhibitory activity directly induced in LICs.  My results suggest that IFN-γ is capable of delaying disease kinetics in the Eµ-RET mouse model by inhibiting the early-life proliferation of LICs, thereby restricting the size of the pre-leukemic cell pool.  While our results do not directly demonstrate that higher LIC burden is causally associated with earlier disease onset, an association between early LIC burden and disease latency has been uncovered in humans.  Several studies have indicated that pre-leukemic cell “load” at birth, as estimated from analysis of Guthrie cards, is correlated with age of diagnosis (Fasching et al. 2000; Gale et al. 1997; Taub et al. 2002;). Additionally, evaluation of ETV6-RUNX1 translocation positive cell burden at different ages suggested that LIC burden peaks near birth and then gradually 	   87 declines, making the neonatal period critical for establishing the LIC burden that may contribute significantly to subsequent ALL progression (Lausten-Thomsen et al. 2010).  Additional support for our hypothesis comes from the finding that low IFN-γ producer genotypes are associated with an earlier age of diagnosis in ALL patients (Cloppenborg et al. 2005).  Although the mechanism of this association has not been defined, our results provide a model that may also be relevant in humans.  The observation that ETV6-RUNX1-transduced B-cell precursors have slower basal growth rates relative to non-transduced cells, but had a significant growth advantage in the presence of TGFβ in vitro, revealed the importance of immune context when evaluating the growth properties of LICs (Ford et al. 2009).  Our in vivo results demonstrate a similar context dependency in our model.  In the absence of IFN-γ, SOCS-1 expression is inhibitory to B-cell precursors and thus detrimental to LICs by restricting IL-7-driven proliferation; however, in the presence of IFN-γ, SOCS-1 is absolutely required for LIC survival.  This is demonstrated by the fact that SOCS1-/- mice have profound deficits in mature B-cell numbers; a characteristic that has been attributed to increased apoptosis in B-cell precursors that were incapable of inducing SOCS-1-mediated negative-feedback mechanisms to limit IFN-γ signaling (Starr et al. 1998).  This duality is supported by findings in our model: while LIC proliferation and expansion are elevated in IFN-γ-/- mice, LICs isolated from IFN-γ-/- Eµ-RET mice display increased sensitivity to IFN-γ both in vitro and in vivo, which may be a consequence of their reduced basal SOCS-1 expression.  In contrast to the TGFβ-mediated growth advantage conferred by TEL-AML1 expression, Eµ-RET derived LICs appear to be as sensitive as 	   88 normal B-cell precursors to the inhibitory effects of IFN-γ, potentially reflecting the commonality of this pathway to normal and abnormal B-cell precursors. IFN-γ has been implicated in the immune-mediated protection against a number of different solid cancers that have been modeled extensively in mice (Bui et al. 2012; Dunn et al. 2002; Kaplan et al. 1998; Shankaran et al. 2001; Street et al. 2002).  This impact was attributed to tumor-specific immunosurviellance mediated by both the innate and adaptive immune arms, and malignant cells derived from IFN-γ-/- mice often exhibit increased immunogenicity upon adoptive transfer into wild type secondary hosts.  While the role for IFN-γ in the immune-mediated inhibition of solid-tumors is well established, its role in hematological malignancies remains poorly defined.  An increased incidence of spontaneous lymphomas has been reported in the absence of IFN-γ; however, in contrast to the findings in other immune-deficient models, IFN-γ-deficient tumors exhibited no difference in outgrowth in wild type or IFN-γ-deficient secondary recipients (Street et al. 2001).  This finding suggests that IFN-γ mediated protection from hematological malignancies may involve alternative mechanisms than those involved in solid tumor development.  I similarly failed to detect any indication of increased immunogenicity in leukemia cells from IFN-γ-/- Eµ-RET mice in adoptive transfer studies.  This finding indicates that the inhibitory action of IFN-γ described here does not fit the immunoediting model, and that direct rather than indirect immune-mediated inhibitory mechanisms are responsible for the observed influence.   While leukemic cells derived from IFN-γ-/- Eµ-RET mice displayed similar outgrowth in both wild type and IFN-γ-deficient secondary recipients, outgrowth of IFN-γ-/- LICs was significantly inhibited by the presence of IFN-γ in vivo.  This finding 	   89 suggests that, even without previous exposure to IFN-γ, overt transformation results in IFN-γ insensitivity.  My results have uncovered two potential mechanisms for this acquired insensitivity.  Firstly, that overt transformation is accompanied by self-sufficiency in growth factor signaling, rendering leukemic cells less reliant on IL-7 signaling for proliferation and survival, and therefore less sensitive than LICs to the inhibitory effects of IFN-γ-induced SOCS-1 expression.  And second, that while LICs up-regulate SOCS-1 expression in response to basal IFN-γ, overt leukemic cells have selectively lost the ability to initiate a similar response.  The contribution of these two mechanisms to the acquired insensitivity of overt leukemic cells to IFN-γ in vivo is currently unknown, but our results indicate that one or both mechanisms are likely at play in the Eµ-RET mouse.  While our results indicate that basal IFN-γ influences disease progression in the Eµ-RET mouse through a direct mechanism that results in the restriction of LIC proliferation, they do not rule out the existence of a subsequent, infection-driven IFN-γ mediated inhibitory mechanism.  The existence of such an inhibitory influence has been suggested by findings from epidemiological studies assessing the influence of early-life infection exposure on ALL risk.  It is also hypothesized that defects in these normal protective response pathways, such as polymorphisms that influence cytokine production, may instead result in dysregulated responses that are more severe and potentially pro-leukemic.  In line with this hypothesis, there is a growing body of experimental evidence in support of a role of infection responses in the etiology of B-cell ALL.  For example, it has been demonstrated that several rounds of TLR-4 stimulation in combination with growth factor starvation has the potential to drive leukemogenesis through the induction 	   90 and inappropriate action of AID and RAG1/2 (Swaminathan et al. 2015).  Additionally, non-specific infection exposure late in life after a significant absence of early-life immune priming increased ALL incidence in a heterozygous Pax5 deletion mouse model (Martin-Lorenzo et al. 2015).  Together with the results presented in this chapter, this growing body of experimental evidence is refining the model of how the immune system may influence ALL risk and progression.  In this model, the influence of cytokines on ALL kinetics may be exerted at several different time-points both in the context of immune responses and also basal production by the resting immune system.  It is possible that a child with a high IFN-γ producer genotype is likely to have both elevated basal production during the neonatal period, as well as higher induced levels during an immune response, compared to a low IFN-γ producer.  Based on my results it is therefore possible to speculate that infection responses in a high producer may potentiate the protective effect imparted by basal IFN-γ, while a low-producer will have deficient basal cytokine production and may be predisposed to the development of abnormal responses that may be both more severe and pro-leukemic.  The influence of cytokine levels then, whether protective or otherwise, may be compounded throughout childhood.  This study only specifically addresses the role of basal IFN-γ; however, considering the established role of SOCS proteins in B-cell development, it is possible that similar mechanisms are applicable to other cytokines.  Several cytokines, such as TGFβ, known to impact the survival or proliferation of B-cell precursors have been shown to induce expression of SOCS proteins, and in some context have been associated with B-cell ALL in humans (Qin et al. 2009; Wiegering et al. 2014; Yoshimura et al. 	   91 2007).  It is therefore possible that a similar mechanism may be applicable to these cytokines, potentially broadening the relevance of the model defined here.  Taken together the results described here provide a model for the influence of IFN-γ on disease progression in the Eµ-RET mouse model.  They highlight the impact of basal IFN-γ on the LIC population and demonstrate how overt transformation can indirectly alter IFN-γ sensitivity.  While the relevance of this mechanism in humans is not assessed in this work, IFN-γ has been demonstrated to modulate the survival of normal human B-cell precursors.  The impact of this inhibition in vivo or how the underlying genetic alterations that drive B-cell ALL humans may alter pre-leukemic cell sensitivity to such an inhibitory influence remains unknown.  These results do however provide a potential explanation for the mechanisms through which polymorphisms in cytokine genes may impact B-cell ALL risk in humans.              	   92 Chapter 4: Inhibition of precursor B-cell malignancy progression by toll-like receptor mediated immune responses  4.1 Overview and rationale 	   While the conclusions derived from epidemiological studies assessing the influence of early-life infection exposure on ALL risk are somewhat discordant, a consistent protective effect has been uncovered when surrogates of infection exposure, such as daycare attendance are evaluated (Ma et al. 2010; Rudant et al. 2015; Urayama et al. 2010).  Currently, the mechanism of this influence remains undefined.  The prevailing hypothesis used to explain this association is Greaves’ delayed infection hypothesis (Greaves 1988; Greaves 2006).  This model postulates that leukemia is the result of an abnormal immune response that occurs as a consequence of a lack of early-life immune priming.  In this scenario, early-life infection exposure is a passive protective influence that serves to “educate” the developing immune system and prevent the subsequent development of abnormal, pro-leukemic immune responses later in life.  While this model may very well be applicable in humans, it also remains plausible that early-life infection exposure may play a more active role in inhibiting ALL progression.  While experimental evidence for such an influence remains limited, the results presented in the previous chapter illustrate that basal IFN-γ is capable of directly inhibiting LIC proliferation and causing a corresponding delay in disease onset.   Based on these findings, I hypothesized that infection exposures that induce IFN-γ would have a dramatic effect on both LIC survival and disease kinetics in the Eµ-RET mouse. As no specific pathogens have been definitively demonstrated to have an impact on B-cell ALL risk either directly or indirectly, we decided to utilize non-specific TLR-	   93 mediated responses to assess the impact of non-specific infection exposure on ALL risk.  TLRs are important early sensors of infection that are capable of inducing broad changes in the immune environment and are involved in the induction of both innate and adaptive immune responses (Blasius & Beutler 2010; Takeda et al. 2003; Underhill & Ozinsky 2002). TLRs sense invading microorganisms via the recognition of microbe associated molecules referred to as pathogen associated molecular patterns (PAMPs).  There are 10 distinct TLRs in humans, each of them recognizes and is activated by a small but specific collection of PAMPs (Moresco et al. 2011; Underhill & Ozinsky 2002).  TLRs demonstrate differential cellular localization based on the type of PAMPs they recognize.  Several TLRs, particularly those that recognize bacterial associated molecules, are localized to the cell surface, while those that recognize nucleic acids (TLR3, 7, 8 and 9) are intracellular, primarily localized to the endosomal compartment (Blasius & Beutler 2010; Takeda et al. 2003).  These nucleic sensing TLRs are important sensors of viruses and their function is strictly regulated to prevent inappropriate activation in response to host genetic material.  Activation of these TLRs has been demonstrated to induce significant production of type-1 and type-2 interferon (Blasius & Beutler 2010; Moresco et al. 2011).  Ligands for intracellular TLRs, therefore, represent a useful tool to mimic the broad innate immune responses induced by intracellular pathogens. As the subsequently detailed work illustrates, I found that immune modulation using certain TLR agonists induced the significant depletion of LIC both in vitro and in vivo.  The mechanism of this influence involved the lymphocyte independent production of both type-1 and type-1 interferon.  The applicability of this mechanism was validated in vitro using pre-leukemic cells derived from the E2A-PBX-1 mouse model, a human 	   94 leukemia cell line (RS4;11), and finally primary human leukemia cells.  The results presented here demonstrate that an alternative mechanistic explanation to the observed association between infection surrogates and reduced ALL risk may exist.  Rather than simply serving to educate the immune system against the development of abnormal, pro-leukemic immune responses later in life, my results indicate that infection exposure may have an active role in preventing ALL development or progression.  4.2 Stimulation with specific TLR agonists results in LIC depletion in vitro 	   As TLRs are expressed on a wide range of immune effector cells, including B-cell precursors, I first assessed whether activation of endosomal TLRs had any direct effects on LIC survival and/or proliferation.  I used 3 different TLR agonists for this investigation: CpG oligodeoxynucleotide 1826 (CpG), which is an unmethylated olignonucleotide with two non-palindromic CpG motifs that is recognized by TLR-9; R848, which is a guanosine derivative recognized by TLR-7 and 8; and poly I:C (pIC), which is a mismatched double stranded RNA that is recognized by TLR-3.  Administration of these three TLR agonists had differential direct effects on purified LICs.  While CpG and R848 demonstrated a modest LIC depletion after a 48 hour incubation, poly I:C induced significant LIC expansion (Figure 4.1).  This result indicates that the direct effect of TLR activation on LICs depends upon the specific identity of the TLR that is activated.   	   95  Figure 4.1.  Direct effect of TLR ligation on LICs in vitro The direct effects of TLR ligation on LICs was assessed by incubating purified LICs derived from Eµ-RET mice with CpG, R848 or pIC for 48 hours.  Impact on LIC viability was assessed by comparing the number of viable cells recovered after 48 hours of TLR agonist incubation to that of cells cultured in the absence of TLR ligands.  N=5 pooled LIC samples.  Mann-Whitney, bars represent mean±S.D.; **p<0.01, ***p<0.001.   4.3 The indirect effects of TLR agonists on LICs 	   While administration of TLR agonists had both modest and differential direct effects on LIC survival, this is not representative of the type of response that would be induced in either humans or mice.  TLRs are expressed on a variety of immune effector cell subsets and activation of these TLRs by ligand binding can drive broad changes in the immune environment.  As the relevant immune responses in humans occur in the context of an intact immune system, it is more physiologically relevant to assess the influence of TLR activation on LIC survival in the presence of additional immune effector cell subsets.  The presence of additional responding cell subsets will result in the production of cytokines that may have a direct or indirect influence on LICs.   	   96  To assess the impact of TLR activation on LIC survival in the context of a broader immune response, whole splenocytes derived from healthy 4-6 week old Eµ-RET mice were incubated with or without TLR ligands.  By assessing the number of viable LICs in these cultures after a 48-hour incubation I was able to determine the indirect, immune response-mediated effects of TLR activation on LIC survival.  While CpG and R848 demonstrated only modest direct LIC depletion, the depletion effect of these ligands increased dramatically in the presence of other immune effector cells (Figure 4.2).  Furthermore, the presence of other immune effector cells converted the direct stimulatory effect of pIC into a net negative effect on LIC survival.  These results indicate that immune responses induced by TLR agonists may potentiate or even oppose the direct impact of TLR activation on LIC survival.  These findings reiterate the importance of assessing the impact of immune modulation on LICs in the correct immune context.      	   97  Figure 4.2.  The indirect effects of TLR activation on LIC survival The indirect effects of TLR activation on LIC survival were assessed by culturing bulk splenocytes derived from Eµ-RET mice with or without TLR ligands for 48 hours.  TLR ligand mediated impact on LIC survival was assessed by comparing the number of viable cells recovered to in treated cultures from those cultured without TLR agonists.  N= 10 mice in 5 independent experiments.  Mann-Whitney, bars indicate mean±S.D.; *p<0.05, ***p<0.001.   4.4 The indirect effect of TLR activation on LICs is mediated through a lymphocyte-independent mechanism   TLRs are expressed by a number of different immune cell subsets (Takeda et al. 2003).  Recognition of nucleic acid-related molecules by endosomal TLRs results in characteristic production of type-1 and type-2 interferon.  Effector cells in the myeloid compartment, especially plasmacytoid dendritic cells, have been identified as key mediators of this response (Blasius & Beutler 2010).  To assess which cell types were required to facilitate the indirect TLR-mediated depletion of LICs in vitro, I restricted the effector cell subsets in the assay to those found in the bone marrow of NSG mice.  These mice lack all cells in the lymphoid compartment while having a largely intact myeloid compartment.  Assessing the indirect effect of TLRs on LICs in this context facilitated the assessment of the broad class of effector cells required for the observed LIC 	   98 depletion.  After a 48-hour co-incubation of purified LICs from Eµ-RET mice with NSG BM derived effector cells in the presence of TLR ligands, all three TLR ligands tested resulted in a significant depletion of LICs (Figure 4.3A).  Importantly, the degree of depletion is equal to that observed in bulk splenocyte cultures, indicating that effector cells in the myeloid compartment are sufficient to induce the full extent of LIC depletion, and that the mechanism of this influence does not involve cells from the lymphoid compartment.    While the Eµ-RET mouse model represents a two-stage model of B-cell precursor malignancy with variable latency, it is not driven by a transgene that is relevant in human B-cell ALL.  As described previously, different ALL genetic subtypes have distinct pathobiological features and the underlying genetic aberrations have a significant impact on risk stratification and overall outcome.  These findings suggest that the specific genetic changes may have an impact on the response of pre-leukemic cells to extrinsic inhibitory factors.  Therefore, to evaluate TLR-induced effects on LICs that carry a human B-cell ALL-relevant genetic aberration we utilized the E2A-PBX1 mouse model.  LICs were purified from the bone marrow of E2A-PBX1 Cd19-Cre mice and were cultured for 48 hours with effector cells derived from NSG mice, with or without TLR ligands.  Consistent with the results derived from the Eµ-RET mouse model, CpG, R848 and pIC were capable of inducing significant depletion of E2A-PBX1+ LICs through an indirect mechanism that is mediated by myeloid effector cells (Figure 4.3B).  This result indicates that the observed TLR-mediated inhibitory effect is not an artifact that is specific to the Eµ-RET mouse model, and is instead also relevant in LICs that carry a human-relevant transgene. 	   99  A.               B.  Figure 4.3.  TLR ligand-induced LIC depletion is mediated by myeloid effector cells Purified LICs derived from Eµ-RET (A) and E2A-PBX1 (B) mice were co-cultured with bone-marrow effector cells from NSG mice.  The number of viable LICs was assessed after 48 hours by flow cytometry, and compared to LICs cultured without TLR ligands.  N= 6 pooled LIC samples from Eµ-RET mice and 4 LIC samples from E2A-PBX1 mice.  Mann-Whitney, bars represent mean±S.D.; **p<0.01, ***p<0.001.   4.5 TLR ligand-induced LIC depletion is mediated by type-1 and type-2 interferon   Recognition of foreign genetic material by nucleic acid-sensing endosomal TLRs results in the significant induction of both type-1 and type-2 interferon by cells in the myeloid compartment (Blasius & Beutler 2010).  I have previously demonstrated that IFN-γ is capable of inhibiting the proliferation of Eµ-RET-derived LICs both in vitro and in vivo.  To assess whether the TLR-mediated depletion observed is dependent on the production of interferon by effector cells in the myeloid compartment, I attempted to inhibit LIC depletion by concurrent administration of IFN-γ and IFN-α blocking antibodies.  Additionally, in order to rule out the impact of specific underlying genetic abnormalities on this response, I assessed the effect of TLR activation on normal late-pro B-cells derived from wild type BALB/c mice.  In order to accomplish this, TLR agonists 	   100 were administered to bone marrow cell cultures from Eµ-RET (Figure 4.4A) and BALB/c (Figure 4.4B) mice, with or without neutralizing antibodies targeting IFN-γ and IFN-αR1.  In both settings, CpG and R848 induced significant LIC and normal B-cell precursor depletion.  This depletion was significantly abrogated by the presence of blocking antibodies.  Interestingly, administration of pIC in combination with IFN-γ and IFN-α neutralizing antibodies caused an expansion of both LICs and normal B-cell precursors that was similar in magnitude to the direct effect of pIC on purified LICs.  These results indicate that the mechanism of TLR agonist-mediated depletion of LICs depends primarily upon the production and signaling of type-1 and type-2 interferon; in the presence of IFN-γ and IFN-α blockade, LIC depletion was similar in magnitude to that achieved by direct TLR stimulation of LICs.  Additionally, these results indicate that the TLR-induced activity is not specific to LICs and instead represents a stage-specific response of B-cell precursors.  While specific B-cell ALL associated genetic lesions may impart insensitivity to this mechanism, the finding that it is maintained in normal B-cell precursors provides another indication that it may be relevant in humans.         	   101  A.          B.   Figure 4.4.  TLR-mediated LIC depletion is dependent upon IFN-γ and IFNα The effect or TLR activation on Eµ-RET LICs and normal late pro-B-cells was assessed.  Bone marrow cells derived from Eµ-RET (A) and BALB/c (B) mice were cultured for 48 hours with or without TLR agonists and IFN-γ and IFN-αR1 blocking antibodies.  Number of viable cells after 48 hours was assessed using flow cytometry (B220int/BP-1hi for LIC and B220int/CD43int/BP-1+/CD24+ for normal late pro B-cells).  N= 6 Eµ-RET and 6 BALB/c mice in 3 independent experiments.  One-way ANOVA, bars indicate mean±S.D.; *p<0.05, **p<0.01, ***p<0.001.   4.6 The inhibitory effect of IFN-γ and IFN-α is mediated through a direct mechanism   My results indicated that TLR-mediated LIC depletion was dependent upon the production and signaling of IFN-γ and IFN-α.  This finding established two potential scenarios: first, that the inhibitory effect was mediated by the direct action of these cytokines on LICs; or second, that after production, these cytokines act indirectly on LICs through the activation of other immune effector cell subsets.  While these two scenarios are not mutually exclusive, we have previously demonstrated that IFN-γ is capable of directly inhibiting the proliferation of Eµ-RET LICs.  However, that study assessed the impact of basal IFN-γ on LICs, and it remains possible that IFN-γ produced during an immune response may impact LICs through an alternative or additional 	   102 mechanism.  To address this issue, I cultured purified LICs in supernatant derived from TLR agonist (CpG or R848)-stimulated bone marrow cultures.  By excluding effector cell subsets from the culture, I was able to assess the direct impact of TLR-induced IFN-γ and IFN-α on Eµ-RET derived LICs.  Incubation of LICs in TLR-conditioned supernatant achieved a depletion of similar magnitude to that observed in the presence of immune effector cells.  Additionally, for both CpG (Figure 4.5A) and R848 (Figure 4.5B) the blockade of both type 1 and 2 IFN was necessary for maximal inhibition of LIC depletion, suggesting that these cytokines act directly, through overlapping mechanisms to reduce LIC numbers.  Again, the depletion achieved in the presence of IFN-blocking antibodies was similar in magnitude to that resulting from the direct action of TLR ligands, as assessed by using supernatant from non-stimulated cells to which TLR agonists were added prior to incubation with LICs.  This finding indicates that direct IFN activity accounts for most, if not all, of the immune-mediated TLR agonist-induced LIC depletion.          	   103  A.          B.  Figure 4.5.  The inhibitory effect of IFN-γ and IFN-α is mediated through direct and overlapping mechanisms Purified LICs derived from Eµ-RET mice were cultured for 48 hours in supernatant derived from TLR stimulated BALB/c bone marrow cultures.  The individual effects of type-1 and type-2 IFN on LIC depletion was assessed by the addition blocking antibodies to the supernatant prior to LIC incubation.  The direct effect of TLR ligands on LICs was assessed by adding TLR agonists to supernatant derived from unstimulated bone-marrow cultures immediately prior to LIC incubation.  N=6 pooled LIC samples in 3 independent experiments.  One-way ANOVA, bars represent mean±S.D.; ***p<0.001.   4.7 TLR mediated inhibition may be relevant in humans 	   I have demonstrated the capacity of immune modulation by specific TLR ligands to inhibit LIC survival in vitro.  This mechanism is primarily dependent upon the direct action of IFN-γ and IFN-α on LICs, and has been validated with two genetic sub-types of LICs and normal BCP cells.  However, while mouse models are useful tools to investigate B-cell ALL development, when assessing the impact of immune modulation on ALL progression it is important to recognize that there are several differences between the human and murine immune systems (Mestas & Hughes 2015).  In order to assess the relevance of this mechanism to human B-cell ALL, it is important to utilize a system that 	   104 involves both human leukemia cells and a human immune response.  In order to accomplish this we isolated PBMCs from healthy human donor blood and stimulated them with CpG or R848.  We then cultured the human B-cell ALL line RS4;11 and a primary human B-cell ALL sample in supernatants harvested from the stimulated PBMCs.  Supernatants from both CpG- and R848-stimulated PBMCs caused a significant depletion of RS4;11 cells (Figure 4.6A).  Consistent with the observations from LICs derived from the Eµ-RET and E2A-PBX1 mouse models, this depletion was partially abrogated by the administration of IFN-γ and IFN-αR1 blocking antibodies.  Interestingly, in the primary human B-cell ALL sample, while R848 induced an IFN-γ- and IFN-α-mediated depletion, CpG had very little effect on cell survival (Figure 4.6B).  Moreover, supernatants from CpG stimulated PBMCs treated with IFN-neutralizing antibodies actually caused a modest expansion of leukemic cells.  These results indicate that in at least some capacity, a similar TLR-induced inhibitory mechanism may be relevant in humans.  Furthermore, this human mechanism is also mediated by the direct effects of IFN-γ and IFN-α.  And finally, while significantly more variable than the response observed in pre-leukemic cells, overt leukemic cells, even immortalized cell lines, demonstrate some sensitivity to TLR mediated inhibitory mechanisms.  Overall, these results indicate that while variable, a similar infection-induced inhibitory mechanism may be relevant in humans.     	   105    A.        B.  Figure 4.6.  TLR-mediated inhibition may be relevant in human B-cell ALL PBMCs were isolated from human blood samples and stimulated with CpG and R848 for 24 hours.  The human B-cell ALL cell line RS4;11 (A) and a primary human B-cell ALL sample (B) were incubated in supernatants derived from the stimulated PBMC cultures.  Viable cell number was assessed by flow cytometry after 48 hours, and the contribution of IFNs to the observed effect was evaluated by the addition of IFN-γ and IFN-α neutralizing antibodies.  N=1 primary human ALL sample, graphs represent cumulative results from 2 independent experiments.  One-way ANOVA, bars represent mean±S.D.; ***p<0.001.   4.8 The TLR mediated inhibitory mechanism is relevant in vivo 	   While informative, in vitro studies fail to capture the complex interactions associated with specific microenvironments on immune activity and outcome.  It is impossible to account for the complexities and regulation of a developing immune response in vitro.  It was, therefore, critical that we evaluate the impact of TLR agonist induced immune responses on ALL progression in vivo.  To do this we administered CpG to Eµ-RET mice via intraperitoneal injection and assessed both LIC burden and disease kinetics in treated versus untreated mice.  Consistent with the results acquired in our in vitro assays, administration of CpG caused a significant depletion of LICs in both the bone marrow and spleen of treated 	   106 mice, relative to their untreated counterparts (Figure 4.7).  Our in vitro results indicated that TLR-mediated inhibition was dependent on the production and signaling of both type-1 and type-2 interferon.  Consistent with this finding, the in vivo TLR agonist-induced depletion of LICs was abrogated in the absence of IFN-γ, as was demonstrated by the failure to achieve a TLR ligand-induced inhibitory effect in CpG-treated IFN-γ-/- Eµ-RET mice (Figure 4.8).  Finally, this CpG-induced LIC depletion in wild type Eµ-RET mice was associated with a significant delay in disease onset (Figure 4.9).  Additionally, the results from this survival study indicate that the TLR-induced protection extends beyond an acute LIC-depletion response, as nearly 25% of treated mice achieved long-term disease-free survival, while 100% of untreated mice succumbed to disease within 250 days.  These results further validate the relevance of our in vitro findings and demonstrate that immune responses induced by TLR agonists are capable significantly altering ALL progression.           	   107  Figure 4.7.  CpG administration induced significant depletion of LICs Administration of CpG to Eµ-RET mice induced significant depletion of LICs relative to untreated controls.  100µg of CpG was administered once-a-week over a 4-week treatment period.  LIC burden in treated and untreated mice was subsequently assessed using flow cytometry.  N=9 mice in each treatment group.  Mann-Whitney, bars represent mean±S.D.; **p<0.01, ***p<0.001.                          	   108   Figure 4.8.  CpG-induced LIC depletion is IFN-γ dependent Wild type and IFN-γ-/- Eµ-RET mice were treated weekly with 100µg of CpG for a 4 week treatment period.  After this treatment period mice were sacrificed and splenic LIC burden was assessed using flow cytometry.  N= 10 wild type and 7 IFN-γ-/- Eµ-RET mice in each treatment group.  One-way ANOVA, bars represent mean±S.D.; **p<0.01                      	   109  Figure 4.9.  CpG-induced immune responses significantly delay disease onset Wild type Eµ-RET mice were treated with 100µg of CpG every two weeks for 8 weeks and then followed to disease onset.  Disease progression was monitored via flow cytometric assessment of peripheral blood blast burden.  Disease was defined by the presence of palpable lymph nodes of a WBC>15,000/µL.  N= 27 CpG treated and 51 PBS treated mice.  Median survival = 152 days (PBS) and 216 days (CpG).  ****p<0.0001. Log-Rank.   4.9 Discussion 	   It is now well established that ALL is a two-phase disease with a distinct pre-leukemic stage that is initiated in utero.  The finding that nearly 1% of all newborns carry the ETV6-RUNX1 t(12;21)(p13;q22) chromosomal translocation, a frequency nearly 100-fold higher than the incidence of this particular leukemic sub-type, indicates that the evolution of pre-leukemia to overt leukemia is not inevitable (Mori et al. 2002).  It is therefore possible that the pre-leukemic disease phase represents a unique window during which immune modulation may significantly alter disease risk and/or progression.  Observational support for this notion comes from the finding that pre-leukemic cell “load” at birth is associated with age of diagnosis (Gale et al. 1997; Hjalgrim et al. 2002; 	   110 Taub et al. 2002).  Early-life influences that negatively affect the survival or expansion of pre-leukemic cells may therefore have a significant protective effect. I have previously described how basal IFN-γ is capable of directly restricting LIC proliferation and delaying disease progression in the Eµ-RET mouse model.  We, therefore, hypothesized that immune modulation that induces IFN-γ production would further alter disease kinetics by negatively impacting the survival and/or proliferation of LICs.  The results presented in this chapter demonstrate the sensitivity of leukemia-initiating B-cell precursors from two different transgenic mouse models of ALL to changes in the immune environment, and include the first description of the impact of in vivo immune modulation on disease development.  We demonstrate that specific TLR activation is capable of inducing significant depletion of LICs derived from the Eµ-RET and E2A-PBX1 Cd19-Cre mouse models in vitro.  Our results demonstrate that this depletion is mediated by the direct, overlapping effects of IFN-γ and IFN-α on LICs.  This impact of TLR activation was maintained in vivo and treatment with the TLR-9 agonist CpG resulted in a significant depletion of LICs in the spleen and bone marrow of treated mice, which was associated with a significant delay in disease onset.  This depletion response was largely absent in CpG-treated IFN-γ-/- Eµ-RET mice, indicating that our in vitro defined mechanism was also relevant in vivo.  Importantly, we also demonstrate that this TLR-mediated inhibitory mechanism may be relevant in humans.  Although the responses were significantly more variable, we demonstrate that supernatant from TLR stimulated human PBMCs was capable of inducing significant IFN-dependent depletion of a human B-cell ALL cell line (RS4;11) and a primary human B-cell ALL sample.  While this result cannot be validated in vivo, it serves as a strong 	   111 indication that similar responses may have an impact on human B-cell ALL.  In contrast to our work with mouse cells, the human cells used in this study are fully leukemic rather than pre-leukemic; the variability observed in this assay may, therefore, be reflective of changes that are associated with overt transformation.  It is hypothetically possible that the response of human pre-leukemic cells may be more dramatic and less variable. Our results provide support for the hypothesis that immune modulation during the pre-leukemic phase can significantly alter subsequent disease progression.  Furthermore, they provide a feasible mechanistic explanation for the association of infection exposure and reduced ALL risk.  Infection has been associated with both an increased and a decreased risk of ALL.  The existing experimental support for an association between infection exposure and ALL only addresses the positive association.  For example, it has been demonstrated that ETV6-RUNX1-transduced B-cell precursors have a significant growth advantage over their normal non-transduced counterparts in the presence of TGFβ (Ford et al. 2009).  Additionally, the TLR-4-induced expression and off-target activity of AID and RAG1/2 has been demonstrated to mediate the malignant transformation of ETV6-RUNX1 transduced B-cell precursors (Swaminathan et al. 2015).  And finally, transfer of Pax5+/- mice from a specific pathogen free (SPF) environment to a conventional facility later in life dramatically increased leukemia incidence in this mouse model (Martin-Lorenzo et al. 2015).  The work described here is the first study to demonstrate a negative association between infection exposure and ALL progression.  While it fails to capture to full range of potential immune-related effects, when combined with the growing body of evidence supporting the existence of leukemia-promoting immune-modulation, it does expose the complexity of variables that may contribute to 	   112 these diverging outcomes.  For example, variables such as the specific immune response pathways engaged, the context and extent of direct activity on LICs, and the immune responsiveness of the host may all contribute to the overall effect of infection exposure on LIC survival and subsequent disease risk or progression.  The impact of variables such as these on response outcome is demonstrated in this study.  The ability of IFN-neutralization in the presence of immune effector cell subsets to convert the effect of TLR-3 activation from pre-leukemic cell depletion to a potentially pro-leukemic LIC expansion illustrates how host immune responsiveness is critical in determining the outcome of infection exposure on LICs.  This is further supported by the inability to achieve TLR-mediated LIC depletion in the absence of IFN.  Together these results demonstrate that the impact of immune modulation on ALL progression or risk must be evaluated in the correct context before the relevance of any such mechanism can be determined.  This study identifies type 1 and 2 IFN as in vivo modifiers of LIC survival that may significantly alter the effect of the identified immune-mediated pro-leukemic responses such as the ETV6-RUNX1 growth advantage in the presence of TGFβ or TLR-4 driven transformation.  While these influences were found to be pro-leukemic, their impact was assessed in the absence of an intact and properly functioning immune response.  It is, therefore, possible that concurrently generated inhibitory influences, such as the production of type-1 and 2 IFNs, may override such pro-leukemic activities in the context of a normal immune response, as is demonstrated here with the effect of TLR-3 activation. Infection exposures occurring early in life have been found to have the most significant impact on ALL risk (Ma et al. 2002; Ma et al. 2010; Urayama et al. 2011).  	   113 This study fails to address the variable of timing in the influence of immune modulation on B-cell ALL development.  Additionally, while I demonstrate that CpG-induced immune responses were capable of significantly altering LIC population dynamics and disease progression in the Eµ-RET mouse, the treatment regimen used was intensive, and would hypothetically be akin to relatively severe infection exposure in humans.  This aspect of the study is seemingly at odds with the fact that it is undocumented infections that confer the greatest protective benefit in humans.  This discrepancy may suggest the existence of a specific temporal window of influence in humans, such that infections during the first year of life may confer a more dramatic protective effect than those occurring later, potentially explaining the intensive treatment required to induce a protective response in adult mice in vivo. The results presented here provide the first mechanistic explanation for the negative association between infection exposure and ALL risk.  While they likely fail to recapitulate the full complexity of similar exposures in humans, they identify a relevant response pathway and serve as a proof of principle for the existence of such influences.  They also serve to underscore the need to understand the interaction between inhibitory and pro-leukemic responses, the development of the immune system, and the influence of response abnormalities, such as polymorphisms in cytokine genes, before a unifying explanation for the contrasting influences of infection exposure on B-cell ALL can be proposed.    	   114 Chapter 5: The IL-12/IL-23 immune response pathway determines the outcome of infection on acute lymphoblastic leukemia progression 	  5.1 Overview and rationale 	   While the conclusions drawn from epidemiological studies assessing the association between infection exposure and ALL risk are somewhat discordant and remain controversial, one common finding is that the timing of infection exposure represents a critical variable (Crouch et al. 2012; Jourdan-Da Silva et al. 2004; Ma et al. 2002; Ma et al. 2010; Roman et al. 2007; Urayama et al. 2011).  It has been demonstrated that regardless of the direction of association, infection exposures occurring within the first year of life have the most significant impact on ALL risk.  It is well established that the early-life immune system is quantitatively and qualitatively different from the adult immune system (Corbett et al. 2010; Dasari et al. 2011; Debock & Flamand 2014; Kollmann et al. 2012; Levy 2007).  Based on an increased early-life susceptibility to various infections, the neonatal immune system was originally considered to be incompetent of dysfunctional.  However, a growing body of evidence has demonstrated that this is not the case, and that while the early immune system displays deficits in some response pathways, it is equally, if not more proficient than the adult at mounting others (Kollmann et al. 2012; Vanden Eijnden et al. 2006).  In a broad sense, relative to the adult, the neonatal immune system shows a defect in mounting Th1 responses, and is characterized by Th2 and Th17 immune responses (Corbett et al. 2010; Kollmann et al. 2012).  These differences reflect the unique immune challenges associated with the neonatal life stage, the most important of which is avoiding excessive inflammation while maintaining mucosal barrier integrity during microbial colonization. 	   115 Despite being identified as important, the impact of infection exposure during the neonatal period on ALL risk remains poorly understood.  As my previous work has underscored the importance of host immune responsiveness in the determining the outcome of infection exposure on LICs, it is possible that the quantitative and qualitative differences associated with the neonatal immune system will alter the impact of immune stimulation on disease progression.  As the subsequently described work will demonstrate, I have found that infection exposure during the neonatal period is capable of profoundly altering disease kinetics in both the Eµ-RET and E2A-PBX1 mouse models.  This response outcome is mediated by the IL-23-induced production of IL-17 by γδ TCR+ T-cells, in an immune response that is induced in neonatal but not adult mice.  Furthermore, I demonstrate that this LIC depletion is not specific to a single infection but instead was broadly applicable to 3 distinct models, including commonly encountered childhood infections.   Both increased and decreased risk of ALL have been associated with exposure to early-life infections; interestingly, my results demonstrate that while induction of normal neonatal response patterns are protective, disruption of the IL-12/IL-23 signaling axis during infection resulted in significant LIC expansion in both neonatal and adult mice.  This study provides the first mechanistic insight into the observed importance of timing in the association between infection exposure and ALL risk.  Furthermore, this study identifies a previously unrecognized inhibitory pathway that may be relevant in children and demonstrates that dysfunction of this particular response pathway may provide an explanation for the paradoxical results derived from studies assessing the influence of infections on ALL risk. 	   116 5.2 Neonatal administration of Lm significantly alters disease progression 	   Early-life daycare attendance has been associated with reduced B-cell ALL risk, while the impact of documented infections appears to be more variable with both increased and decreased ALL risk being reported (Roman et al. 2007; Rudant et al. 2015; Urayama et al. 2011).  As asymptomatic, subclinical infections have demonstrated the most consistent impact on ALL risk, it was important to use an infection model that induced a similar response in order to interrogate the mechanism of this influence.  To accomplish this I utilized an ΔTrpS;ActA attenuated strain of Listeria monocytogenes (subsequently referred to as Lm).  Attenuated Lm infections are mild and self-limiting, and have been studied extensively in mice to investigate broad patterns of early-life immune responses to intracellular pathogens (Hamon et al. 2006).  In order to assess the impact of timed Lm infection on ALL progression in the Eµ-RET mouse, I administered a single injection of Lm (104 cfu) at day 6 (neonatal) or day 34 (adult) of life.  Mice were sacrificed 8 days after infection exposure and LIC burden at each time point was assessed and compared to mock-infected (saline) controls.  While adult Lm exposure had no influence on splenic LIC burden, neonatal Lm caused a significant decrease in LIC burden in treated mice relative to untreated controls (Figure 5.1).  Consistent with CpG-induced reduction in LIC burden (Figure 4.9), the LIC depletion induced by neonatal Lm infection resulted in a significant delay in disease onset.  Adult Lm infection, which did not alter LIC numbers, had no observable impact on disease kinetics (Figure 5.2).  Determining appropriate inoculum size for in vivo models is difficult.  While there is no indication that real world infectious inoculum sizes differ between adults and 	   117 children, to rule out that the observed response pattern was simply associated with a higher cfu per gram body weight inoculum size in neonatal mice I reduced the neonatal inoculum size to 103 cfu.  This ten-fold dose reduction more than accounted for the average five-fold weight different between neonatal and adult mice.  Despite adjusting the cfu/gram body-weight ratio, neonatal Lm infection still induced significant depletion of LIC burden in infected versus mock-infected mice (Figure 5.3).  These results indicate that the differential effect of adult and neonatal Lm exposure on LIC burden was not simply related to inoculum size as related to body weight.  After addressing the variable of inoculum size, there were two primary possibilities to explain the observation that Lm infection induced LIC depletion in neonatal but not adult mice.  First, that an intrinsic difference in neonatal LICs causes them to be more sensitive to induced inhibitory mechanisms than adult LICs; and second, that the nature of the immune response induced in neonatal mice was qualitatively and/or quantitatively different from that of the adult.  To address the first possibility, I utilized the adult Eµ-RET mouse-derived 289 leukemia cell line.  289 cells were adoptively transferred into newborn and adult mice that were subsequently infected with Lm and splenic 289 cell burden was assessed eight days after infection.  Consistent with the results obtained in situ with Eµ-RET LICs, neonatal Lm infection induced a significant depletion of 289 cells while adult infection had no impact on leukemic cell burden (Figure 5.4).  This result eliminated the variable of LIC age from the previously observed results and implicated qualitative or quantitative differences in neonatal and adult immune responses as the primary mediator of the differential outcomes observed.  	   118 Furthermore, this result indicates that overt leukemic cells, even those that have been immortalized as a cell line, retain sensitivity the depletion mechanism induced by Lm.  As stated earlier, the primary drawback of the Eµ-RET mouse model is that leukemia in this model is not driven by a transgene that is relevant to human B-cell ALL.  As different genetic subtypes of human B-cell ALL have been observed to have different pathobiological characteristics, it is important to assess the impact of early-life immune modulation on LICs or leukemic cells that harbor genetic aberrations that are relevant to human B-cell ALL.  To address this issue I again utilized E2A-PBX1+ cells to assess outcome with a clinically encountered genetic alteration.  To do this, we adoptively transferred E2A-PBX1+ leukemic cells into newborn and adult C57BL/6 mice.  Recipient mice were then infected with Lm at day 6 or day 34 of life, and E2A-PBX1+ leukemic cell burden was assessed eight days after infection.  Once again, neonatal Lm infection induced significant depletion of leukemic cells, while adult infection had no effect (Figure 5.5).  The initiating event in most B-cell ALL cases has been demonstrated to occur in utero and result in the production of a pre-leukemic cell reservoir from which the leukemic clone may eventually arise (Greaves & Wiemels 2003).  There is, therefore, a finite number of pre-leukemic cells and any inhibitory influence that is capable of eliminating these cells in their entirety would prevent the subsequent development of leukemia.  While neonatal Lm infection induced a near complete depletion of LICs in Eµ-RET mice, it induced a significant delay in disease onset rather than disease prevention.  Failure to prevent disease in the Eµ-RET model results from the fact that constitutive transgene expression drives the consistent replenishment of LICs.  An acute depletion 	   119 event, regardless of how effective, would therefore be unlikely to completely prevent disease occurrence. To assess the impact of neonatal Lm infection on ALL that arises from a finite cell pool, we adoptively transferred E2A-PBX1 leukemic cells into newborn mice that were subsequently infected with Lm at day 6.  We then compared the survival of Lm-infected versus mock-infected recipient mice.  In contrast to the impact observed in the Eµ-RET setting, the depletion activity of neonatal Lm infection achieved significant long-term disease-free survival in the majority of recipient mice, while all of the mock-infected mice developed disease within 150 days (Figure 5.6).  This result indicates that in the context of leukemia that is driven by a finite number of initiating cells, as is thought to be the case in the majority of humans, neonatal Lm infection is capable of inducing disease prevention rather than just kinetic delays.     	   120  Figure 5.1.  Neonatal but not adult Lm infection induces significant LIC depletion The impact of Lm infection on splenic LIC burden was assessed in Eµ-RET mice.  Mice were treated with 104 cfu of Lm via intraperitoneal injection on day 6 (neonatal) or day 34 (adult) of life.  Splenic LIC burden was assessed 8 days after infection using flow cytometry and compared to mock-infected (saline) control mice.  N= 9 adult mice in each treatment group and 12 neonatal mice in each treatment group.  One-way ANOVA, bars represent mean±S.D.; ****p<0.0001.                     	   121  Figure 5.2.  Neonatal Lm infection significantly delays disease onset in Eµ-RET mice The impact of neonatal and adult Lm infection on disease progression was assessed.  Mice were infected with 104 Lm on day 6 or day 34 of life and subsequently followed to disease onset.  Disease progression was monitored by assessment of LIC burden in peripheral blood. Disease onset was defined by the presence of palpable lymph nodes or a peripheral blood WBC>15,000/µL.  Survival in Lm infected mice was compared to that of mock-infected mice that were injected with saline at both day 6 and day 34 of life. N= 24 mock-infected mice, 29 D34 Lm infected mice and 22 D6 Lm infected mice.  Median survival =144 days for mock-infected mice, 150 days for D34 Lm infected mice and 224 days for D6 Lm infected mice. Log-Rank.                   	   122  Figure 5.3.  Differential responses to Lm are not related to inoculum size In order to assess whether inoculum size was responsible for the differential Lm response in neonatal and adult mice, neonatal Eµ-RET mice were treated with 103 cfu and splenic LIC burden in infected mice was assessed 8 days after infection using flow cytometry.  N=12 mock-infected and 14 Lm infected mice.  Mann-Whitney, bars represent mean±S.D.; ****p<0.0001.                       	   123  Figure 5.4. Adult mouse-derived leukemic 289 cells are responsive to Lm	  induced inhibitory mechanisms 289 cells were adoptively transferred into newborn and adult mice.  Recipient mice were infected with Lm on day 6 or day 34 of life and splenic 289 cell burden was assessed 8 days after infection exposure by flow cytometry.  Changes in 289 cell burden were assessed by comparison to the mean 289 burden of mock-infected mice at the corresponding age (eg. 0.5 in day 6 Lm would be indicative of a 289 burden that was equal to half of the mean burden in day 6 mock infected mice).  N=5 adult and 8 neonatal mice in each treatment group.  One-way ANOVA, bars represent mean±S.D.; ****p<0.0001.                  	   124  Figure 5.5.  Neonatal but not adult Lm infection induces depletion of E2A-PBX1+ leukemia cells Overt leukemic cells derived from the E2A-PBX1 mouse model were adoptively transferred into newborn or adult C57BL/6 mice.  Recipient mice were infected with Lm on day 6 or day 34 of life, respectively.  Splenic E2A-PBX1+ leukemic cell burden was assessed 8 days after infection exposure using flow cytometry.  Changes in leukemic cell burden were assessed by comparison to the mean leukemic cell burden of mock-infected mice at each age.  The figure represents the results derived from 3 independent leukemia samples.  One-way ANOVA, bars represent mean±S.D.; ****p<0.0001.                  	   125  Figure 5.6.  Neonatal Lm alters disease progression in an E2A-PBX1 leukemia adoptive transfer model E2A-PBX1+ leukemic cells were adoptively transferred into newborn mice.  Recipient mice were infected with Lm at day 6 of life and followed to disease onset.  Disease progression was monitored by assessing blast burden in peripheral blood.  Disease was defined by the presence of palpable nodes of a peripheral blood WBC>15,000/uL.  The figure represents the pooled results from two independent leukemia samples.  N= 14 mice in each treatment group.  14/14 mock-infected and 2/14 Lm infected mice developed disease. Log-Rank.  5.3 Lm-induced LIC depletion requires IL-23 and IL-17A 	   With inoculum size and LIC age ruled out as the critical variables, the most likely explanation remaining for the differential effects is quantitative and/or qualitative differences in neonatal and adult responses to Lm infection.  Likely to address unique early-life immunological challenges, the neonatal immune system is broadly characterized by a reduced capacity to mount Th1 responses, while being highly proficient at mounting Th17 immune responses (Kollmann et al. 2012).  This Th1-deficiency is caused by an intrinsic defect in IL-12p35 production, while the Th17 	   126 proficiency is facilitated by the increased capacity of neonatal dendritic cells and macrophages to produce IL-23 (Kollmann et al. 2012; Vanden Eijnden et al. 2006).  In addition to the fact that IL-23 production is one of the key differences between neonatal and adult immune response, IL-23 has also been identified as an anti-leukemic factor in human B-cell ALL (Cocco et al. 2010), making it an intriguing candidate as a potential contributor to the Lm-induced LIC and leukemic cell depletion observed.    To determine the relevance of IL-23 signaling to the Lm-induced depletion of LICs, I first assessed impact of Lm infection in Stat4-/- Eµ-RET mice.  Stat4 is required for signal propagation downstream of the IL-23 receptor (Watford et al. 2004); without Stat4, cells in these mice are capable of producing IL-23 but incapable of responding to it.  Indicating a role for IL-23 in the observed Lm-induced LIC depletion, no LIC depletion was induced in neonatal Stat4-/- mice, while Stat6-/- mice, which have intact IL-23 signaling networks, demonstrated significant Lm-induced LIC depletion (Figure 5.7).  While a lack of Lm-induced LIC depletion in Stat4-/- Eµ-RET mice supports a role for IL-23 in this inhibitory mechanism, IL-23 is not the only cytokine that signals through Stat4; IL-12 signaling is also propagated through Stat4.  In order to assess the relative contribution of IL-12 and IL-23 to the Lm-induced depletion of LICs, neonatal Lm infection was administered concurrently with IL-12p75 or IL-23p19 neutralizing antibodies.  Consistent with a prominent role of IL-23 in infection induced LIC depletion, IL-23p19 neutralizing antibody completely abrogated infection-induced LIC depletion (Figure 5.8).  Conversely, neutralization of IL-12 had no impact on LIC depletion.  These results demonstrate that IL-23, and not IL-12, is an important contributor to the 	   127 mechanism of Lm-induced LIC depletion.  As increased relative IL-23 production is a characteristic of neonatal immune responses, this result also indicates that early-life immune system may be uniquely adept at mounting anti-leukemic immune responses.  One of the primary functions of IL-23 is to induce differentiation and promote survival and proliferation of Th17 T-cells, which are characterized by their production of IL-17 (Boniface et al. 2008; Cua et al. 2003; Iwakura & Ishigame 2006).  This is demonstrated by the significant deficit in Th17 cells in IL-23 knock-out mice (Cua & Tato 2010; Langrish et al. 2005; Stockinger & Veldhoen 2007).  To assess whether this function of IL-23 was involved in the Lm-induced LIC depletion, I first assessed the ability of Lm infection to induce IL-17 production in neonatal and adult BALB/c mice.  Measurement of IL-17A concentration in the serum of neonatal and adult mice revealed that while adult Lm infection did not induce production of IL-17A, neonatal infection resulted in a nearly twenty-fold increase in IL-17A concentration (Figure 5.9).  To specifically address whether IL-17A was involved in the infection-induced LIC depletion observed, neonatal Lm infection was administered concurrently with IL-17A neutralizing antibody.  The inhibition of IL-17A signaling completely abrogated the Lm-induced depletion of LICs in Eµ-RET mice and E2A-PBX1+ leukemic cell elimination in adoptive transfer recipients (Figure 5.10).   Collectively these results demonstrate that IL-23 and IL-17 are required for the Lm-induced depletion of Eµ-RET LICs and E2A-PBX1+ leukemic cells.  Furthermore, these results provide a mechanistic explanation for the differential effects observed in neonatal and adult mice.  IL-17A signaling is a key mediator of the Lm-induced effect on LICs; therefore the differential capacity of neonatal and adult mice to produce IL-17 may 	   128 underlie the diverging effects of infection in these two settings.  As IL-23 is known to be an important inducer of IL-17A production, the demonstration that both IL-23 and IL-17A mediated the Lm-induced LIC depletion is likely a reflection of the critical nature of IL-23 as a key upstream mediator of IL-17A production, rather than a direct contributor to LIC depletion.  This is in contrast to the previous demonstration that IL-23 has direct anti-leukemic effects (Cocco et al. 2010).   Figure 5.7.  Stat4 signaling is required for Lm induced LIC depletion Lm-induced LIC depletion was assessed in Stat6-/- and Stat4-/- Eµ-RET mice.  Mice were infected with 104 cfu of Lm on day 6 of life.  Splenic LIC burden was assessed by flow cytometry 8 days after infection and was compared to that of mock-infected control mice from each genotype.  N=10 Stat6-k.o and 5 Stat4-k.o mice in each treatment group.  One-way ANOVA, bars represent mean±S.D.; ****p<0.0001.     	   129  Figure 5.8.  IL-23 is required for Lm-induced LIC depletion The relative contribution of IL-12 and IL-23 to the observed infection-dependent depletion of LICs was assessed.  Lm infection was administered concurrently with 200µg of IL-12p75 or IL-23p19 neutralizing antibody.  The effect on splenic LIC burden was assessed by flow cytometry 8 days after infection.  N=12 mock-infected mice, 14 Lm infected mice, 9 aIL-12p75 treated mice and 8 aIL-23p19 treated mice.  One-way ANOVA, bars represent mean±S.D.; ****p<0.0001.                      	   130   Figure 5.9.  Lm infection induces IL-17A production in neonatal mice Production of IL-17A in response to Lm infection was measured in neonatal and adult mice.  Mice were infected with Lm on day 6 and day 34 of life.  Serum was collected from Lm- and mock-infected mice of each age group 18 hours after treatment and IL-17A was measured by ELISA.  IL-17A concentrations were normalized to the average concentration in mock-infected mice at each time-point.  N=14 Adult mock, 13 Adult Lm, 14 D6 Mock, and 12 D6 Lm treated mice.  One-way ANOVA, bars represent mean±S.D.; ****p<0.0001.                   	   131    A.                B.   Figure 5.10.  IL-17A signaling is required for Lm-induced LIC depletion IL-17A-neutralizing or control (Cont) antibody was administered concurrently with 104 cfu of Lm to neonatal Eµ-RET mice.  Splenic LIC burden was assessed 8 days after infection by flow cytometry (A).  E2A-PBX1+ leukemic cells were adoptively transferred into newborn C57BL/6 mice.  Recipient mice were infected with 104 cfu Lm with or without the concurrent administration of 200µg of IL-17A neutralizing antibody.  Splenic leukemic cell burden was assessed 8 days after infection by flow cytometry (B).  Impact on leukemic cell burden was assessed by comparison to the average burden in mock-infected control mice.  One-way ANOVA, bars represent mean±S.D.; ****p<0.0001.                     	   132 5.4 Production of IL-17A by γδ TCR+ T-cells is required for Lm-induced LIC depletion   Lm infection induced significant IL-17A production in mice only 18 hours after exposure, long before adaptive T helper immune responses develop.  Therefore, in order to evaluate the general requirement for lymphocytes in the Lm-induced depletion of LICs I assessed the ability of infection to deplete LICs in Rag1-/- Eµ-RET mice which lack all recombination-dependent mature lymphocyte subsets.  Lm infection of neonatal Rag1-/- Eµ-RET mice failed to induce any degree of LIC depletion, as determined by comparison to mock-infected control mice (Figure 5.11).  This result suggests that a recombination-independent response to infection-induced IL-23 signaling was insufficient to achieve protection from ALL progression.   While CD4+ Th-17 cells are characterized by the production IL-17, they are not the only effector cell subtype that is capable of producing IL-17.  To assess the contribution of the major CD4+ and CD8+ T-cell subsets to the inhibitory effects induced by Lm infection, I depleted these cells from Eµ-RET mice prior to Lm infection and then assessed the subsequent impact of infection in this context on LIC burden.  Pre-depletion of both CD4+ and CD8+ T-cells had no impact on LIC depletion (Figure 5.12).  This result indicates that, despite being characterized by the ability to produce IL-17A, CD4+ Th-17 cells are not responsible for mediating the LIC depletion induced by Lm infection.    The rapid induction of IL-17A in infected mice in combination with a lack of CD4+ T-cell involvement suggests that an innate lymphoid source of IL-17A is the most likely contributor to the Lm-induced depletion mechanism.  Natural killer T (NKT) cells are an immune effector cell subset that fits this description.  NKT-cells are innate-like T-cells that express a semi-invariant αβ T-cell receptor repertoire and contribute to defense 	   133 against a variety of pathogens (Bendelac et al. 2007; Mattner et al. 2005).  NKT-cells have been demonstrated to be capable of producing significant amounts of IL-17 in a response that is potentiated by reception of IL-23 (Rachitskaya et al. 2008).  However, despite being IL-6 independent, production of IL-17 by NKT-cells is thought to require TCR engagement (Rachitskaya et al. 2008).  The complete repertoire of antigens recognized by NKT-cells is currently unknown and the in situ relevance of these cells in the immune response to Lm is not well understood (Emoto et al. 2010; Ranson et al. 2005).  To determine the contribution on NKT-cells to the infection-induced LIC depletion, I assessed the impact of Lm infection on LIC burden in CD1d -/- Eµ-RET mice, which lack all NKT-cell subsets.  In the absence of NKT-cells, Lm infection induced significant depletion of splenic LICs (Figure 5.13), indicating that this effector cell subtype is not involved in the infection-induced depletion of LICs.  Whether the lack of an NKT-cell contribution is related to an inability of Lm to induce NKT-cell activation in our model, or that NKT-cells are activated but do not produce IL-17 in response to Lm, remains unknown.  γδ T-cells are another immune effector cell subset that is capable of producing IL-17 (Cua & Tato 2010).  These T-cells are characterized by the expression of a unique T-cell receptor composed of one γ-chain and one δ-chain.  These effector cells are considered innate-like lymphocytes and they have been shown to be important early producers of IL-17 during an immune response (Caccamo et al. 2011; Roark et al. 2008).  Interestingly, IL-17 producing γδ TCR+ T-cells have been shown to develop exclusively during a specific temporal window in fetal life, and the frequency of these cells has been observed to peak during the perinatal period and decline gradually through life (Haas et 	   134 al. 2012).  A subset of γδ T-cells have been shown to express TLRs and have the capacity to produce significant amounts of IL-17 through a TCR-independent response pathway that is potentiated by reception of IL-23 concurrently with TLR activation (Martin et al. 2009). Importantly, γδ T-cells have been shown to be activated to rapidly produce significant amounts IL-17 by Lm exposure (Laird et al. 2013).  To assess the relevance of this effector cell subset in the Lm-induced depletion of LICs, I depleted γδ TCR+ T-cells from Eµ-RET mice prior to Lm infection.  In the absence of γδ T-cells, Lm infection failed to induce any significant alteration to the LIC population (Figure 5.14).  This result indicates that γδ T-cells are required for the infection-induced depletion of LICs in the Eµ-RET mice.  To validate that this requirement for γδ T-cells was related to their production of IL-17A, I assessed IL-17A serum concentrations in wild type and γδ T-cell-depleted Lm-infected mice.  IL-17A production 18 hours after Lm exposure was significantly reduced in γδ T-cell-depleted mice as compared to wild type infected mice (Figure 5.15).  Collectively, these results indicate that γδ T-cells are the key producers of IL-17A during early-life immune responses to Lm and as a result, represent a critical effector cell population involved in the Lm-induced depletion of LICs.  As the LIC depletion was completely abrogated in the absence of IL-23, γδ T-cells and IL-17A, our results implicate IL-23-induced production of IL-17A by γδ T-cells as the central response pathway responsible for the depletion of LICs.    	   135  Figure 5.11.  Lymphocytes are required for Lm-induced depletion of LICs To assess the relative contribution of lymphocytes to infection-induced LIC depletion, Rag1-/- Eµ-RET mice were infected with 104 Lm on day 6 of life.  Splenic LIC burden in these mice was measured 8 days after infection and compared to that of mock-infected control mice.  N=7 mice in each treatment group.  Mann-Whitney, bars represent mean±S.D.                      	   136  Figure 5.12.  Lm-induced LIC depletion is not mediated by CD4+ or CD8+ T-cells CD4+ and CD8+ T-cells were depleted from 4-day old Eµ-RET mice with targeted antibodies.  Mice were then infected with 104 cfu of Lm at day 6.  Splenic LIC burden was assessed by flow cytometry 8 days after infection and compared to Lm treated mice that were not antibody-depleted.  One-way ANOVA, bars represent mean±S.D.; ****p<0.0001.                      	   137  Figure 5.13.  NKT-cells are not involved in the Lm-induced depletion of LICs To assess the influence of NKT-cells in the depletion of LICs, CD1d-/- Eµ-RET mice infected with 104 cfu of Lm on day 6 of life.  Splenic LIC burden was assessed by flow cytometry 8 days after infection and compared to mock-infected   CD1d-/- Eµ-RET mice.  N=7 mock-infected and 11 Lm infected mice.  Mann-Whitney, bars represent mean±S.D.; ****p<0.0001.                     	   138  Figure 5.14.  γδ T-cells are required for infection-induced depletion of LICs γδ TCR+ T-cells were depleted from 4-day old Eµ-RET mice.  Mice were then infected with 104 cfu of Lm at day 6.  Splenic LIC burden was assessed by flow cytometry 8 days after infection and compared to Lm treated mice that were not antibody depleted.  One-way ANOVA, bars represent mean±S.D.                       	   139  Figure 5.15.  γδ T-cells are the significant producers of IL-17A during the immune response to Lm. To assess the contribution of γδ T-cells to IL-17A production in response to Lm, γδ T-cells were depleted from 4-day old mice followed by Lm infection of depleted and non-depleted mice at day 6 of age.  Serum was collected from infected mice 18 hours after infection and IL-17A concentrations were analyzed by ELISA.  Absolute IL-17A concentrations were compared in mock-infected, Lm-infected and γδ T-cell-depleted Lm-infected mice.  N=14 Mock, 12 Lm, 8 γ/δ TCR+ T-cell deplete+Lm.  One-way ANOVA, bars represent mean±S.D.; ****p<0.0001.   5.5 Infection-dependent LIC depletion is not specific to Lm 	   Extensive investigation has failed to identify a specific infectious agent that is strongly associated with B-cell ALL risk (Bartenhagen et al. 2017; MacKenzie et al. 2006) and current models of B-cell ALL etiology propose that it is the response to childhood infections in general that underlie the association with leukemia risk.  Lm is an extensively studied infectious agent that is commonly used to model immune responses to intracellular bacteria (Hamon et al. 2006).  However, while attenuated Lm is a suitable model for a mild, self-limiting early-life infection in mice, it is not a frequently 	   140 encountered neonatal pathogen in humans.  To assess the generality of the LIC-depleting response, as well as its potential relevance in children, we modeled two common viral infections of early childhood: murine cytomegalovirus (MCMV) and murine gammaherpesvirus-68 (MHV-68), a mouse model of Epstein-Barr virus (EBV) (Krmpotic et al. 2003; Olivadoti et al. 2007).  Assessment of sero-positivity for antibodies against these pathogens indicates that they are widely disseminated throughout the population (Carvalho-Queiroz et al. 2016; Jenson 2011; Staras et al. 2008).  Additionally, normal responses to the acute phase of both of these viruses are generally asymptomatic in immune competent individuals and, interestingly, early-life daycare attendance has been demonstrated to increase exposure to both of these pathogens (Adler 1986; Hesse et al. 2015; Joseph et al. 2006).  To assess the impact of exposure to these infectious agents on LIC burden we replicated the timing model previously utilized with Lm.  Similar to the influence of Lm, neither intra-peritoneal MCMV infection (Figure 5.16) nor intra-nasal MHV-68 (Figure 5.17) infection had any effect on LIC numbers in adult Eμ-RET mice.  However, when these infection exposures occurred during the neonatal period, each resulted in a significant depletion of LICs.  Furthermore, the LIC depletion induced by both MCMV and MHV-68 was entirely dependent on IL-17A, indicating that the mechanism of action is maintained across all infectious agents tested.  These results demonstrate that the infection-induced depletion of LICs is not limited to Lm and is relevant to other infections.  The finding supports the notion that the impact of early-life infection on ALL risk may be related to broad differences in neonatal response patterns rather than the exposure to any specific pathogen. 	   141  Figure 5.16.  Neonatal but not adult MCMV infection induces LIC depletion Adult Eµ-RET mice were infected on day 34 of life with 103 pfu of MCVC via intraperitoneal infection while neonatal mice were infection on day 6 of life with 102 pfu of MCMV, with or without IL-17A neutralizing antibodies.  Splenic LIC burden was assessed 8 days after infection and compared to mock-infected control mice.  One-way ANOVA, bars represent mean±S.D.; ****p<0.0001.              	   142  Figure 5.17.  Neonatal but not adult MHV-68 infection induces LIC depletion Adult Eµ-RET mice were infected on day 34 of life with 4x103 pfu of MHV-68 via intranasal infection while neonatal mice were infected on day 6 of life with 4x102 pfu of MHV-68, with or without IL-17A neutralizing antibodies.  Splenic LIC burden was assessed 8 days after infection and compared to mock-infected control mice.  One-way ANOVA, bars represent mean±S.D.; ****p<0.0001.   5.6 Disruption of the IL-12/IL-23 signaling axis during infection induces LIC expansion   It has been proposed that a dysfunctional immune response to common infections, which result in more severe symptoms that may require medical intervention, may be pro-leukemic (Greaves 1988; Greaves 2006; Wiemels 2012).  Pre-existing underlying immune response defects, such as low producer polymorphisms in cytokine genes, may cause such dysfunctional responses (Rudant et al. 2015).  Consistent with this hypothesis is the finding that polymorphisms in several cytokine genes such as IFN-γ and IL-12A have been associated with reduced latency and increased risk of B-cell ALL, respectively (Chang et al. 2010; Chang et al. 2011; Cloppenborg et al. 2005; Han et al. 2010; Winkler 	   143 et al. 2015).  Additionally, premature infants are profoundly deficient in IL-12/IL-23p40 production and have demonstrated both a higher risk of severe infection and ALL (Lausten-Thomsen et al. 2010; Lavoie et al. 2010).  Based on these observations, I investigated the effect of IL-12/IL-23p40 blockade on the protective activity induced by neonatal infection. Neutralization of IL-12/IL-23p40 in the absence of infection had no impact on LIC burden in Eµ-RET mice (Figure 5.18).  However, in the context of both Lm (Figure 5.19) and MCMV (Figure 5.20) infection, neutralization of IL-12/IL-23p40 caused a significant expansion of the splenic LIC population in infected mice.  However, IL-12/IL-23p40 neutralization during MHV-68 infection had no significant on splenic LIC burden (Figure 5.21).  This result suggests that a dysregulated response to some, but not all, neonatal infections could increase ALL risk by expanding the pool of abnormal cells available for full transformation. The strong LIC expansion after impaired response to early-life MCMV infection is consistent with an increased risk of ALL in children infected with human CMV in utero (Francis et al. 2016), where IL-12/IL-23p40 expression is lacking (Huygens et al. 2014).  In addition, a similar LIC expansion was also observed in Lm- and MCMV-infected adult mice in which IL-12/IL-23p40 was neutralized (Figure 5.22).  The observation that dysfunctional immune responses are capable of inducing LIC expansion in adult mice indicates that the lack of LIC-depletion activity in these mice during normal infection responses is not simply the result of a failure of adult mice to respond to these infectious stimuli.  In addition, the expansion in adults indicates that while quantitative and qualitative differences in the adult immune system preclude the generation of anti-leukemic 	   144 responses, the adult immune system retains the capacity to generate potentially pro-leukemic immune responses when specific immune response pathways are disrupted.    Figure 5.18.  Neutralization of basal IL-12/IL-23p40 has no effect on LIC burden Non-infected Eµ-RET mice were treated with 200µg of IL-12/IL23p40 neutralizing antibody.  Splenic LIC burden was assessed 8 days after antibody administration and compared to that of non-treated control mice.  Mann-Whitney, bars represent mean±S.D.                 	   145  Figure 5.19.  Dysfunctional immune responses to Lm induce LIC expansion 6-day old Eµ-RET mice were infected with 104 cfu Lm with the concurrent administration of 200µg of IL-12/IL-23p40 neutralizing antibody.  Splenic LIC burden was assessed 8 days later by flow cytometry and compared to mock-infected mice.  Mann-Whitney, bars represent mean±S.D.; *p<0.05.                      	   146  Figure 5.20. Dysfunctional immune responses to MCMV induce LIC expansion 6-day old Eµ-RET mice were infected with 102 pfu MCMV with the concurrent administration of 200µg of IL-12/IL-23p40 neutralizing antibody.  Splenic LIC burden was assessed 8 days later by flow cytometry and compared to mock-infected mice.  Mann-Whitney, bars represent mean±S.D.; ****p<0.0001.                     	   147  Figure 5.21.  Dysfunctional immune responses to MHV-68 do not induce LIC expansion 6-day old Eµ-RET mice were infected with 4x102 pfu MHV-68 with the concurrent administration of 200µg of IL-12/IL-23p40 neutralizing antibody.  Splenic LIC burden was assessed 8 days later by flow cytometry and compared to mock-infected mice.  Mann-Whitney, bars represent mean±S.D.                      	   148  Figure 5.22.  Dysfunctional adult immune responses induce LIC expansion The impact of IL-12/IL-23p40 neutralization in the context of adult infection exposure on LICs was assessed.  Adult Eµ-RET mice were infected with 104 cfu of Lm or 103 pfu of MCMV via intraperitoneal injection.  200µg of IL-12/IL-23 neutralizing antibody were administered concurrently with the infectious agents and splenic LIC burden was assessed 8 days later by flow cytometry.  One-way ANOVA, bars represent mean±S.D.; ***p<0.001, ****p<0.0001.   5.7 Discussion   A role for infection exposure in the etiology of B-cell ALL is well established.  However, the nature of this influence remains controversial as the results of large-scale epidemiological studies assessing the association between infection exposure and ALL risk have produced contradictory results.  However, despite these discrepancies, timing of infection exposure has consistently been uncovered as a critical variable, with exposures occurring within the first year of life having the most significant influence, regardless of the direction of the association reported (Chan et al. 2002; Crouch et al. 2012; Ma et al. 2002; Ma et al. 2010; Jourdan-Da Silva et al. 2004; Roman et al. 2007; Urayama et al. 	   149 2011). To date, the mechanism behind the influence of timing on the association between infection exposure and B-cell ALL risk remains unknown.  This study demonstrates that quantitative and qualitative differences in the neonatal immune system result in the initiation of protective early-life immune responses that are not mounted by the mature immune system.  The neonatal immune system is significantly different from that of the adult; avoiding excessive inflammation while maintaining mucosal integrity during commensal colonization is the critical post-natal challenge that influences the nature of early-life immune responses (Debock & Flamand 2014; Kollmann et al. 2012).  High circulating concentrations of adenosine in concert with a defect in the ability of IRF3 to bind both CBP and DNA lead to an intrinsic deficit in IL-12p35 production (Haskó et al. 2000; Kollmann et al. 2012).  Conversely, neonatal production of IL-23p19 and IL-12/23p40 is maintained at levels equal to or greater than that of adults (Vanden Eijnden et al. 2006).  These conditions lead to low IL12p75 production and high IL-23 production, skewing neonatal immune responses away from Th1 and toward Th17.  The results of this study indicate that the early-life skewing of immune responses to high IL-23 production generates potent IL-17-driven immune activity that inhibits B-cell ALL progression through the depletion of LICs present in the neonatal environment.  While the results presented here do not completely define the mechanism of infection-induced LIC depletion, they indicate that IL-17A production by γδ T-cells is a necessary component.  The implication of γδ T-cells in the LIC depletion mechanism may serve as an indication that the impact of the neonatal immune system on LIC depletion may extend beyond the increased capacity for IL-23 production.  The 	   150 development of IL-17 producing γδ T-cells is restricted to a narrow temporal window in fetal life (Haas et al. 2012).  These cells develop in the neonatal thymus and then disseminate throughout the body; however, the frequency of these cells peaks during the perinatal period and subsequently declines with age (Haas et al. 2012).  It is, therefore, plausible that the relative impact of these cells during an immune response is highest during this period as well.  γδ T-cells have been demonstrated to express a limited repertoire of TLRs that facilitate TCR-independent T-cell activation and production of IL-17A in a response that is potentiated by concurrent reception of IL-23 (Martin et al. 2009).  It is, therefore, likely that the multiple unique features of the neonatal immune system work together in an overlapping response pathway to facilitate the infection-associated LIC depletion observed in this study.  For instance, relative to adults, the increased capacity of the neonatal immune system to produce IL-23 facilitates increased production of IL-17A by γδ T-cells, a response difference that may be compounded by their relatively increased frequency during early-life.  As IL-23 induces an IL-17-producing phenotype in γδ T-cells, it is also possible that the cytokine milieu induced in neonates supports a more uniform differentiation of γδ T-cells to an IL-17-producing, rather than an IFN-γ producing phenotype, than in adults.   The relative contribution of these features to the increased function of γδ T-cells during the neonatal-specific infection-induced LIC depletion remains to be established.  Notably, ours is not the first study to implicate γδ T-cells in immune-mediated anti-leukemic mechanisms; a survival advantage has been observed in leukemia patients with high γδ T-cell frequency following allogenic stem-cell transplantation (Godder et al. 2007).  While the mechanism of this protection is yet to be defined, it is proposed that γδ 	   151 T-cells may facilitate graft-versus-leukemia responses without causing graft-versus-host disease.  How allogenic stem cell transplant might influence the overall frequency, distribution, or effector function of γδ T-cells remains unknown. Overall, the results of my study indicate that a combination of unique features of the neonatal immune system facilitate the increased production of IL-17, which is responsible for mediating the infection-induced LIC depletion described here.  An important observation in this study is that the protective influence of early-life infection extends to a range of infectious agents across different pathogen classes.  While the epidemiological studies assessing the association between infection exposure and ALL risk have produced paradoxical results, a common pattern has emerged; asymptomatic or mild infections have been associated with protection while those that are more severe and require a physician’s visit have a more variable effect.  The underlying explanation for this paradox remains unknown. It has been suggested that these findings can be explained by the existence of underlying immune defects that result in more severe, potentially pro-leukemic, responses to infections that would be significantly milder in immune competent individuals (Rudant et al. 2015). Alternatively, it is possible that only mild, asymptomatic infections are capable of conferring protection, while more severe infection responses do not. It is, therefore, important to evaluate the impact on B-cell ALL of common early-life infections that are usually either asymptomatic or mild enough to infrequently require a doctor’s visit; CMV and EBV are just such infections (Carvalho-Queiroz et al. 2016; Cohen 2000).  Interestingly, exposure to both of these infectious agents has been demonstrated increase dramatically with daycare attendance (Hesse et al. 2015; Joseph et al. 2006; Staras et al. 2008).  While our results do not 	   152 specifically address the impact of more severe infections on LIC population dynamics, they reveal that mild infections are capable of protecting against leukemia progression by inducing the depletion of pre-leukemic cells.  This response may underlie the protective association between surrogates of infection exposure and reduced ALL risk.   While demonstrating that mild or asymptomatic infections may be responsible for mediating the protective influence observed in humans, this study also demonstrates the capacity of immune responses to switch from protective to pro-leukemic.  I demonstrate that dysregulaton of the IL-12/IL-23 response axis during an immune response completely abrogated LIC- and leukemic cell-depletion and instead induced significant blast expansion.  While the influence of such a defect in human immune responses is currently unknown, this finding may provide an explanation for the increased ALL risk associated with recorded early infections.  There are several observations that support the potential human relevance of my findings.  Premature infants, which have a profound deficit in both IL-12 and IL-23 production also have an increased risk of B-cell ALL (Lausten-Thomsen et al. 2010; Lavoie et al. 2010).  Premature infants are also more susceptible to infection; the increased ALL incidence observed in these children may be related to the pro-leukemic infection responses that occur in the absence of IL-12 and IL-23.   Additionally, an increased risk of ALL has been reported for children that were infected with human CMV in utero (Francis et al. 2016), where IL-12/IL-23p40 expression is lacking (Huygens et al. 2014).  Finally, significant associations between ALL risk and polymorphisms in genes central to these pathways, including IL12A, IL12B, Stat4, and Stat6, have been reported in candidate gene association studies which have the potential to identify variants with more subtle effects on disease risk (Chang et 	   153 al. 2010; Han et al. 2010; Moriyama et al. 2015).  These observations uncover a pattern in that combinatorial defects in IL-12 and IL-23 production or signaling, either genetic or developmental, are associated with an increased risk of ALL.  This pattern may be reflective of an underlying pro-leukemic mechanism that is similar to the abnormal response pattern modeled in this study. The work described here supports the refinement of the working model of the influence of infection exposure on B-cell ALL risk.  Mild or asymptomatic infections, such as CMV or EBV, during the first year of life result in the generation of potent IL-17-based immune responses that inhibit the development of leukemia by eradicating pre-leukemic cells.  As the immune system matures, the capacity for IL-12p35 production increases and immune responses shift from predominantly Th17 to predominantly Th1.  While these immune responses may impart a protective influence through the production of type-1 and type-2 interferon, the efficiency of pre-leukemic cell depletion is reduced, limiting the protective effect of later-life infections.  Conversely, infections in children with dysfunctional immune responses caused by either genetics (eg polymorphisms in cytokine genes) or immune ontogeny (eg premature infants or delayed immune maturation) are more likely to require a physician’s visit and have a more variable effect on ALL risk.  The specific nature of this effect will depend on the infectious agent involved or the nature of the immune response defect, with combinatorial deficits in IL-12 and IL-23 production or signaling having a significant pro-leukemic influence.  Additionally, while not specifically addressed here, it is also possible that more severe, symptomatic infections may be pro-leukemic in both normal children and those with underlying immune response dysregulation.  While observational, this suggestion is 	   154 supported by the association of more severe infections, such as influenza, with a significantly higher risk of B-cell ALL (Dockerty et al. 1999; Petridou et al. 2002).  While this model is consistent with existing experimental and epidemiological data, further assessment in humans is required to validate this model.  Overall, this study provides the first mechanistic explanation for the observed impact of timing of exposure on the association between infection and ALL risk.  Additionally, we demonstrate that while typical immune responses to mild infections are protective, disruption of normal response pathways can transform a protective response into a pro-leukemic response.  Interestingly, while protective responses were only induced in neonatal mice, a dysregulated pro-leukemic response could be generated in both neonatal and adult mice, an observation that further implicates timing as a critical variable.  This study provides a model to explain the apparently contradictory effects of early-life infection on childhood ALL, and identifies a previously under-appreciated protective response pathway that merits further investigation.                  	   155 Chapter 6: General discussion and perspectives 	  6.1 The current model of infectious ALL etiology 	   As has been extensively described in previous chapters, a growing number of epidemiological studies have demonstrated an association between infection exposure and B-cell ALL risk.  The nature of this association however, remains controversial with both increased and decreased risk being reported, and the underlying explanation for these contradictory results is unknown. The goal of my thesis work was to identify the specific immune mechanisms that exerted an influence over ALL progression and to evaluate these pathways for their ability to explain the discordant epidemiologic findings. Throughout, my experimental rationale and design was informed by epidemiologic findings and the postulates of the current prevailing model of the interaction between ALL and the immune system, Greaves’ delayed infection hypothesis (Greaves 1988; Greaves 2006). Broadly, it postulates that a lack of early life infections, as might be experienced in an affluent society, predisposes the immune system to aberrant, pro-leukemic responses upon later life infection exposure.  These aberrant immune responses promote mutation acquisition in pre-leukemic cells that arose in utero and thereby promote leukemogenesis.  My research describes three distinct immune influences on ALL progression that provide strong mechanistic support for many facets of Greaves’ hypothesis and reveal previously unappreciated actively protective mechanisms. Overall, these results suggest refinements to the current model that provide an explanation for the dual effects of infection on ALL.  	   156 6.2 The impact of the resting immune system on B-cell ALL 	   Epidemiological studies have uncovered an association between a number of polymorphisms in cytokine genes and an increased risk or reduced latency of B-cell ALL (Chang et al. 2010; Cloppenborg et al. 2005; Han et al. 2010; Wiegering et al. 2014).  Most explanations for this association are based on the infectious model of ALL and postulate that the polymorphisms alter cytokine production and promote the development of abnormal or deregulated immune responses that are pro-leukemic (Rudant et al. 2015).  While this explanation may have merit, experimental support is currently lacking.  These polymorphisms, however, are also likely to affect basal production of cytokines.  Therefore, we hypothesized that the general cytokine milieu, in the absence of infectious stimulation, would influence leukemia development, and specifically that deficient IFN-γ production may compromise an inhibitory influence.  The work detailed in chapter 3 highlights the existence of an inhibitory mechanism mediated by basal IFN-γ.  I demonstrate that IFN-γ inhibits the early-life proliferation of LICs in the Eµ-RET mouse model, and by doing so restricts LIC pool expansion and delays disease kinetics.  These results are consistent with the finding that ALL patients with a low IFN-γ producer genotype were significantly younger at the time of clinical diagnosis than those with a high expressing genotype (Cloppenborg et al. 2005).  While it has not been demonstrated that this difference is associated with pre-leukemic cell population dynamics, it has been observed that pre-leukemic cell “load” at birth is associated with age at clinical disease onset (Gale et al. 1997; Hjalgrim et al. 2002; Taub et al. 2002).  While indirect, these observations suggest that the mechanism uncovered in the Eµ-RET mouse may have relevance in humans. 	   157  IFN-γ has been implicated as a key mediator of inhibitory cancer immune surveillance mechanisms in a number of solid cancer models (Bui et al. 2012; Dunn et al. 2006; Ikeda et al. 2002; Kaplan et al. 1998; Shankaran et al. 2001).  While IFN-γ acts through both the adaptive and innate immune mechanisms to prevent or delay the development of cancer in these models, one common finding is that malignant cells derived from IFN-γ-/- mice in these models are more immunogenic than their wild type counterparts as assessed by the ability of these cells to cause disease in wild type secondary recipients.  This model has provided an effective framework for understanding the evolving interactions between cancer and the immune system, and the role of IFN-γ-mediated inhibition of solid cancers appears to be consistent with the postulates of this model.  However, the role of IFN-γ in hematological malignancies has not always been consistent with the immunoediting model (Street et al. 2002).  The results presented in chapter 3 demonstrate an editing-independent activity in which IFN-γ mediates a protective influence through a direct mechanism that involves the induced expression of SOCS-1 and subsequent restriction of IL-7-driven proliferation (Corfe et al. 2011).  Consistent with previous reports, this study reveals that leukemic cells derived from IFN-γ−/− Eµ-RET mice did not display any increased immunogenicity in wild type recipients, and indicates that overt transformation is accompanied by a loss of sensitivity to the direct IFN-γ mediated inhibitory mechanism.  Furthermore, they demonstrate that growth factor self-sufficiency and disruption of SOCS-1 induction represent two putative mechanisms for this acquired insensitivity to basal IFNγ.  The results presented in Chapter 3 provide the first model to mechanistically explain the association of polymorphisms in cytokine genes and increased risk or 	   158 accelerated progression of B-cell ALL.  While we only demonstrate the relevance of this mechanism with IFN-γ, considering the established role of SOCS proteins in B-cell development, it is possible that similar mechanisms are applicable to other cytokines as well.  TGFβ, for example, has been demonstrated to impact the survival and proliferation of B-cell precursors, induce production of SOCS proteins, and has been suggested to impact aspects of B-cell ALL development in humans (Qin et al. 2009; Tang et al. 1997; Wiegering et al. 2014; Yoshimura et al. 2007).  It is therefore possible that a similar mechanism may be applicable to such cytokines, potentially broadening the relevance of the model defined here. 6.3 The impact of immune modulation on B-cell ALL 	   Greaves hypothesizes that an abnormal or dysfunctional response to common infections occurring as a result of a lack of early-life immune priming, promotes, through either genotoxic or proliferative stress, the secondary mutation event.  While studies are few, experimental evidence supports such a pro-leukemic influence of infection (Ford et al. 2009; Martin-Lorenzo et al. 2015; Swaminathan et al. 2015). However, early mild infections are consistently protective against ALL and 99% of infants who are positive for ETV6-RUNX1 translocations at birth do not develop ALL. These observations suggest that protective mechanisms may also influence the incidence of disease progression, but no such inhibitory mechanism has been described.  Having demonstrated that IFN-γ is capable of directly impacting the proliferation and survival of LICs, I speculated that infections that induce IFN might confer protection through a more active mechanism. 	   159  The results presented in chapter 4 represent the first mechanistic demonstration of an active protective mechanism induced by infection-based immune modulation.  We demonstrate that through the production of type-1 and type-2 IFN, TLR ligands are capable of inducing the depletion of LICs derived from the Eµ-RET and E2A-PBX1 mouse models.  This induced LIC depletion was associated with a significant delay in disease onset in the Eµ-RET mouse.  Importantly, we also demonstrate the potential relevance of this inhibitory mechanism in humans.  As the immune systems of mice and humans have several differences, it was important to validate this response pathway using human leukemic cells and human immune effector cells.  Furthermore, our results demonstrate that variables such as the immune response pathways engaged, the context and extent of direct activity on LICs or leukemic cells and the immune responsiveness of the host may all have a significant effect on the overall outcome of infection exposure on disease risk or progression.  The significance of these variables is demonstrated by the ability of IFN neutralization in the presence of immune effector cell subsets to convert the effect of TLR-3 activation from pre-leukemic cell depletion to a potential pro-leukemic pre-leukemic cell expansion, as well as the inability to achieve TLR mediated LIC depletion in the absence of IFN.  These results demonstrate that the impact of infection responses on ALL progression or risk must be evaluated in the context of a functioning immune system.    While the observed effect on human leukemia cells was more variable than that observed in mouse LICs, it is important to note that overtly malignant cells are likely very different than their pre-malignant predecessors.  Self-sufficiency in growth factor signaling, resistance to programmed cell death and insensitivity to anti-growth signaling 	   160 are all hallmarks of cancer (Hanahan & Weinberg 2000); any or all of these features may be altered during malignant transformation, and these changes may have a dramatic influence on the sensitivity of overt leukemic cells to immune-mediated inhibitory mechanisms.  I believe that any demonstration of sensitivity to these immune response pathways in overt human leukemia cells should be interpreted as a strength rather than a weakness of this study, despite the observed variability of outcome.    Overall, the results presented in chapter 4 represent the first study that provides a mechanistic explanation for the observed protective effect of infection exposure.  While, they do not preclude the existence of a more passive “educational” influence as postulated by Greaves, the results of this study suggest that certain immune responses are capable of actively protecting against ALL development by inducing the depletion of the pre-leukemic cell reservoir.  Furthermore, they demonstrate the complexity of the interaction between infection exposure and B-cell ALL etiology, and identify host immune responsiveness as a critical variable in determining the outcome of such exposures. 6.4 Timing is critical 	   Regardless of the direction of association with ALL risk, the timing of infection exposure has been identified as a critical variable.  In regards to the effect of daycare attendance on ALL risk, for example, a review of the findings derived from the NCCLS suggests that the magnitude of the protective effect associated with the same number of childhood hours in daycare was stronger for attendance during infancy than for daycare attendance at any other time leading up to diagnosis.  This finding suggests that early-life infection exposure confers a greater protective effect than later occurring infections (Ma 	   161 et al. 2010).  Conversely, it has also been demonstrated that a history of documented infections occurring during infancy is associated with an increased risk of ALL (Crouch et al. 2012; Roman et al. 2007).  These studies clearly implicate timing as a critical component of the association between infection exposure and ALL risk.  However, as of yet, the underlying nature of the impact of timing remains unknown.   The results presented in chapter 5 represent the first mechanistic explanation for the importance of timing in the association between infection and ALL.  My results demonstrate that qualitative and quantitative features of the neonatal immune system facilitate the induction of protective immune responses that are not mounted by the mature immune system.  Once considered incompetent or dysfunctional, it is now understood that while the early immune system displays deficits in some pathways, it is equally if not more proficient than the adult at mounting others, including IL-23 mediated Th17 responses (Debock & Flamand 2014; Kollmann et al. 2012; Vanden Eijnden et al. 2006). This increased capacity for IL-23 production in combination with a higher frequency of γδ T-cells facilitates the rapid production of IL-17A during neonatal immune responses. My results demonstrate that this specific neonatal IL-17A response to a range of infectious agents, including those that model widely disseminated and often asymptomatic infections in humans, is required to facilitate LIC and leukemic cell depletion in both the Eµ-RET and E2A-PBX1 mouse models, respectively.  By demonstrating the unique capacity of the neonatal immune system to mount infection-induced protective responses, which do not occur in adults, this study identifies a novel protective immune response pathway that merits further investigation for its relevance in children. 	   162 6.5 The generation of pro-leukemic infection responses 	   It has been suggested that the association of early-life infection with both increased and decreased ALL risk may be the result of the presence of underlying immune response defects (Rudant et al. 2015).  In children that carry such defects, infections may be more severe and may frequently require medical intervention.  Such defects may also cause a predisposition towards the development of pro-leukemic immune responses that promote second hit acquisition via induced genotoxic or proliferative stress.  While there is experimental evidence that supports a pro-leukemic influence of infection exposure, no current model explains the paradox uncovered by epidemiological studies.  The results described in chapter 5 not only outline a potential infection-induced pro-leukemic immune response pathway but also address the discordant results uncovered by epidemiological studies.  Our results demonstrate that while normal neonatal immune responses to mild infections are capable of inhibiting disease progression by negatively impacting the survival of LICs and leukemic cells, disruption of this pathway can convert the outcome of infection exposure from protective to pro-leukemic.  I demonstrate that the concurrent disruption of both IL-12 and IL-23 signaling, via the neutralization of IL-12/23p40, completely abrogates infection-induced LIC depletion and instead induces considerable expansion of this abnormal population in treated mice.  Interestingly, while infection-induced inhibitory effects are only induced in the neonatal setting, this expansion can be induced in both neonates and adults.  This observation again implicates timing as an important variable in outcome, such that neonatal exposures are protective in competent hosts while being pro-leukemic in the 	   163 context of a dysfunctional immune response, while exposures in the adult have a less efficient protective influence but retain the capacity to induce pro-leukemic responses if normal response pathways are disrupted.  The finding that premature infants have a higher incidence of B-cell ALL (Lausten-Thomsen et al. 2010), and that in utero CMV infection increases ALL risk (Francis et al. 2016), support the potential of infection exposure in the context of combined deficits of IL-12 and IL-23 to be pro-leukemic.  The results presented in chapter 5 represent the first demonstration that the same infection exposures can be both protective and pro-leukemic in vivo.  They provide a potential explanation for the contradictory results derived from epidemiological studies and once again implicate immune responsiveness as a critical variable in determining the outcome of infection exposure on ALL risk.  While this is not the first study to demonstrate the potential of infection exposure to induce pro-leukemic responses, it is the first to identify a specific response pathway that can convert protective influences into pro-leukemic influences.  Additional investigation will be required to uncover the exact mechanism of each of these influences; however, as it stands the work presented in chapter 5 represents a significant advancement in the understanding of how infection exposure influences ALL risk. 6.6 Closing perspectives 	  Collectively, the results presented in this thesis provide the experimental framework for the development of a model to describe the influence of the immune system on B-cell ALL development.  I propose that in the absence of infection, the early-life production of cytokines, such as IFN-γ, by the resting immune system may impact ALL development by modulating the population dynamics of LICs that arise in utero.  	   164 Dysregulation of these inhibitory mechanisms may underlie the association between polymorphisms in cytokine genes and increased ALL risk.  Infection exposures have a distinct but potentially overlapping influence on B-cell ALL development, the effects of which depend on the temporal window during which these exposures occur.  I propose that mild or asymptomatic infections such as CMV or EBV during infancy result in the generation of potent IL-17-based immune responses that inhibit the development of leukemia by eradicating pre-leukemic cells.  As the immune system matures, the capacity for IL-12p35 production increases and the immune responses to these infectious agents shift from being predominantly Th-17 to predominantly Th-1. While these immune responses may also impart a protective influence through the production of type-1 and type-2 interferon, the efficiency of pre-leukemic cell depletion induced by these responses is reduced, and the protective effect of later-life infection exposure is thereby limited.   My results suggest that the immune system influences ALL development through several mechanisms that are both distinct and overlapping.  The collective impact of these influences will depend on several factors such as the immune responsiveness of the host, the infectious agents encountered and the timing of these exposures.  With this in mind, it is possible that irregularities, such as polymorphisms in cytokine genes or delayed immune maturation, may influence the impact of the immune system on ALL development through multiple different mechanisms.  Polymorphisms are likely to influence the production of cytokines both basally, and during and immune response.  Therefore, a child with a low IFN-γ producer genotype, for example, will likely have lower production of basal early-life IFN-γ, as well as lower production during infection 	   165 exposures that occur after their Th1 responses pathways have matured.  Collectively, reduced IFN-γ production would therefore limit the immune-mediated impact on ALL development to infection exposures that occur during the neonatal window and induce the IL-17A mediated depletion of LICs.  Conversely, in a child with a reduced capacity for IL-23 production, neonatal infection exposures may be unlikely to result in significant depletion of LICs, thereby limiting the protective effect of such exposures.  The overall impact of the immune system on ALL development therefore, depends on a number of interrelated variables, and is likely compounded at multiple points throughout early childhood.  Overall, the results presented in this thesis highlight the important influence of the immune system on the development and progression of B-cell acute lymphoblastic leukemia.  Together they provide a mechanistic model for how the basal and activated immune system are capable of impacting disease kinetics, and help to explain a number of previously identified, but poorly understood associations.  They identify both type-1 and type-2 IFN and IL-17A as inhibitory factors, and demonstrate that influences that are capable of altering pre-leukemic cell survival and/or proliferation have a significant impact on subsequent disease risk and progression.  Finally, they provide a mechanistic explanation for the observed impact of timing of infection exposure and demonstrate that the outcome of such exposures can be either protective or pro-leukemic, depending on the immune responsiveness of the host in which they occur.  Mechanistically these findings represent several firsts in the investigation of immune system influences on B-cell ALL.   	   166 REFERENCES  Abbas, A.K., Murphy, K.M. & Sher, A., 1996. Functional diversity of helper T lymphocytes. Nature, 383(6603), pp.787–793. Aderem, A. & Ulevitch, R.J., 2000. Toll-like receptors in the induction of the innate immune response. Nature, 406(6797), pp.782–787.  Adler, S.P., 1986. Molecular epidemiology of cytomegalovirus: evidence for viral transmission to parents from children infected at a day care center. Pediatric infectious disease, 5(3), pp.315–318.  Ajrouche, R. et al., 2015. Childhood acute lymphoblastic leukaemia and indicators of early immune stimulation: the Estelle study (SFCE). British journal of cancer, 112(6), pp.1017–1026. Akira, S., 2003. Mammalian Toll-like receptors. Current Opinion in Immunology, 15(1), pp.5–11. Allman, D. & Miller, J.P., 2003. Common lymphoid progenitors, early B-lineage precursors, and IL-7: characterizing the trophic and instructive signals underlying early B-cell development. Immunologic Research, 27(2–3), pp.131–140.  Anderson, K. et al., 2011. Genetic variegation of clonal architecture and propagating cells in leukemia. Nature, 469, pp.356-361 De Angulo, G. et al., 2008. Absolute lymphocyte count is a novel prognostic indicator in ALL and AML: implications for risk stratification and future studies. Cancer, 112(2), pp.407–415. Aparicio, S., Hidalgo, M. & Kung, A.L., 2015. Examining the utility of patient-derived xenograft mouse models. Nature Reviews Cancer, 15(5), pp.311–316.  	   167 Armstrong, S.A. & Look, A.T., 2005. Molecular genetics of acute lymphoblastic leukemia. Journal of Clinical Oncology, 23(26), pp.6306–6315. Bae, Y.S. et al., 2011. Regulation of reactive oxygen species generation in cell signaling. Molecules and Cells, 32(6), pp.491–509. Bartenhagen, C. et al., 2017. Infection as a cause of childhood leukemia-virus detection employing whole genome sequencing. Hematologica. Available at: http://www.haematologica.org/lookup/doi/10.3324/haematol.2016.155382. Bassan, R. et al., 2009. Improved risk classification for risk-specific therapy based on the molecular study of minimal residual disease (MRD) in adult acute lymphoblastic leukemia (ALL). Blood, 113(18), pp.4153–4162. Bateman, C.M. et al., 2010. Acquisition of genome-wide copy number alterations in monozygotic twins with acute lymphoblastic leukemia. Blood, 115(17), pp.3553–3558. Bendelac, A., Savage, P.B. & Teyton, L., 2007. The biology of NKT-cells. Annual Review of Immunology, 25, pp.297–336. Bernbeck, B. et al., 2009. Symptoms of childhood acute lymphoblastic leukemia: red flags to recognize leukemia in daily practice. Klinische Pädiatrie, 221(6), pp.369–373. Bhojwani, D. et al., 2012. ETV6-RUNX1-positive childhood acute lymphoblastic leukemia: improved outcome with contemporary therapy. Leukemia, 26, pp.265-270 Bhojwani, D., Yang, J.J. & Pui, C.-H., 2015. Biology of Childhood Acute Lymphoblastic Leukemia. Pediatric Clinics of North America, 62(1), pp.47–60. Blasius, A.L. & Beutler, B., 2010. Intracellular Toll-like Receptors. Immunity, 32(3), 	   168 pp.305–315. Bogdan, C., Mattner, J. & Schleicher, U., 2004. The role of type I interferons in non-viral infections. Immunological Reviews, 202, pp.33–48. Boniface, K. et al., 2008. From interleukin-23 to T-helper 17 cells: Human T-helper cell differentiation revisited. Immunological Reviews, 226(1), pp.132–146. Bradley Forlow, S. et al., 2001. Increased granulopoiesis through interleukin-17 and granulocyte colony-stimulating factor in leukocyte adhesion molecule-deficient mice. Blood, 98(12), pp.3309–3314. De Braekeleer, E. et al., 2012. ETV6 fusion genes in hematological malignancies: A review. Leukemia Research, 36(8), pp.945–961. Buchner, M. et al., 2015. Mechanisms of pre-B-cell receptor checkpoint control and its oncogenic subversion in acute lymphoblastic leukemia. Immunological Reviews, 263(1), pp.192–209. Bui, J. et al., 2012. Cancer immunoediting by the innate immune system in the absence of adaptive immunity. Cancer Research, 72(8), pp.526–526. Burrows, P.D. & Cooper, M.D., 1997. B-cell development and differentiation. Current opinion in immunology, 9(2), pp.239–244. Busslinger, M., Nutt, S.L. & Rolink, A.G., 2000. Lineage commitment in lymphopoiesis. Current Opinion in Immunology, 12(2), pp.151–158. Caccamo, N. et al., 2011. Differentiation, phenotype, and function of interleukin-17-producing human vγ9vδ2 T-cells. Blood, 118(1), pp.129–138. Campana, D., 2010. Minimal Residual Disease in Acute Lymphoblastic Leukemia. Hematology, 2010(1), pp.7–12. 	   169 Campana, D. & Coustan-Smith, E., 1999. Detection of minimal residual disease in acute leukemia by flow cytometry. Cytometry, 38(4), pp.139–152. Cardwell, C.R. et al., 2008. Infections in early life and childhood leukaemia risk: a UK case-control study of general practitioner records. British journal of cancer, 99(9), pp.1529–1533.  Carroll, W.L. et al., 2003. Pediatric acute lymphoblastic leukemia. Hematology Am Soc Hematol Educ Program, pp.102–131. Carvalho-Queiroz, C. et al., 2016. Associations between EBV and CMV Seropositivity, Early Exposures, and Gut Microbiota in a Prospective Birth Cohort: A 10-Year Follow-up. Frontiers in Pediatrics, 4, p.93.  Ceredig, R. & Rolink, A.G., 2012. The key role of IL-7 in lymphopoiesis. Seminars in Immunology, 24(3), pp.159–164. Chan, K.W., 2002. Acute lymphoblastic leukemia. Current problems in pediatric and adolescent health care, 32(2), pp.40–49.  Chan, L.C. et al., 2002. Is the timing of exposure to infection a major determinant of acute lymphoblastic leukaemia in Hong Kong? Paediatric and Perinatal Epidemiology, 16(2), pp.154–165. Chang, J.S. et al., 2010. Genetic polymorphisms in adaptive immunity genes and childhood acute lymphoblastic leukemia. Cancer epidemiology, biomarkers & prevention, 19(9), pp.2152–2163. Chang, J.S. et al., 2011. Profound deficit of IL10 at birth in children who develop childhood acute lymphoblastic leukemia. Cancer Epidemiology Biomarkers and Prevention, 20(8), pp.1736–1740. 	   170 Chen, Y. et al., 2003. Stimulation of airway mucin gene expression by interleukin (IL)-17 through IL-6 paracrine/autocrine loop. Journal of Biological Chemistry, 278(19), pp.17036–17043. Chiaretti, S. & Foà, R., 2009. T-cell acute lymphoblastic leukemia. Haematologica, 94(2), pp.160–162. Chien, Y.H., Jores, R. & Crowley, M.P., 1996. Recognition by gamma/delta T-cells. Annual review of immunology, 14, pp.511–32.  Choi, S. et al., 2007. Relapse in children with acute lymphoblastic leukemia involving selection of a preexisting drug-resistant subclone. Blood, 110(2), pp.632–639. Clark, M.R. et al., 2014. Orchestrating B-cell lymphopoiesis through interplay of IL-7 receptor and pre-B-cell receptor signalling. Nature reviews. Immunology, 14(2), pp.69–80.  Cloppenborg, T. et al., 2005. Immunosurveillance of childhood ALL: polymorphic interferon-gamma alleles are associated with age at diagnosis and clinical risk groups. Leukemia, 19(1), pp.44–8.  Cocco, C. et al., 2010. Interleukin-23 acts as antitumor agent on childhood B-acute lymphoblastic leukemia cells. Blood, 116(19), pp.3887–3898. Cohen, J.I., 2000. Epstein-Barr virus infection. The New England journal of medicine, 343(7), pp.481–92.  Conter, V. et al., 2010. Molecular response to treatment redefines all prognostic factors in children and adolescents with B-cell precursor acute lymphoblastic leukemia: results in 3184 patients of the AIEOP-BFM ALL 2000 study. Blood, 115(16), pp.3206–3214. 	   171 Cooper, S.L. & Brown, P.A., 2015. Treatment of pediatric acute lymphoblastic leukemia. Pediatric Clinics of North America, 62(1), pp.61–73. Corbett, N.P. et al., 2010. Ontogeny of toll-like receptor mediated cytokine responses of human blood mononuclear cells. PLoS ONE, 5(11). Corfe, S.A. & Paige, C.J., 2012. The many roles of IL-7 in B-cell development; Mediator of survival, proliferation and differentiation. Seminars in Immunology, 24(3), pp.198–208. Corfe, S. a, Rottapel, R. & Paige, C.J., 2011. Modulation of IL-7 thresholds by SOCS proteins in developing B lineage cells. Journal of Immunology, 187(7), pp.3499–3510.  Cortes, J. & Kantarjian, H., 1995. Acute Lymphoblastic Leukemia. Cancer, 76(12), pp.2393–2417. Cossart, P., 2011. Illuminating the landscape of host-pathogen interactions with the bacterium Listeria monocytogenes. Proc Natl Acad Sci U S A, 108(49), pp.19484–19491.  Crouch, S. et al., 2012. Infectious illness in children subsequently diagnosed with acute lymphoblastic leukemia: Modeling the trends from birth to diagnosis. American Journal of Epidemiology, 176(5), pp.402–408. Cua, D.J. et al., 2003. Interleukin-23 rather than interleukin-12 is the critical cytokine for autoimmune inflammation of the brain. Nature, 421(6924), pp.744–748. Cua, D.J. & Tato, C.M., 2010. Innate IL-17-producing cells: the sentinels of the immune system. Nat Rev Immunol, 10(7), pp.479–489.  Cuenca, A.G. et al., 2013. Role of innate immunity in neonatal infection. American 	   172 Journal of Perinatology, 30(2), pp.105–112. Danis, B. et al., 2008. Interferon regulatory factor 7-mediated responses are defective in cord blood plasmacytoid dendritic cells. European Journal of Immunology, 38(2), pp.507–517. Dasari, P., Zola, H. & Nicholson, I.C., 2011. Expression of Toll-like receptors by neonatal leukocytes. Pediatric Allergy and Immunology, 22(2), pp.221–228. Debock, I. & Flamand, V., 2014. Unbalanced neonatal CD4+ T-cell immunity. Frontiers in Immunology, 5, pp. 393. Van Delft, F.W. et al., 2011. Clonal origins of relapse in ETV6-RUNX1 acute lymphoblastic leukemia. Blood, 117(23), pp.6247–6254. Dockerty, J.D. et al., 1999. Infections, vaccinations, and the risk of childhood leukaemia. British journal of cancer, 80(9), pp.1483–1489.  Doll, R. & Wakeford, R., 1997. Risk of childhood cancer from fetal irradiation. British Journal of Radiology, 70, pp.130–139. Dowling, D.J. & Levy, O., 2014. Ontogeny of early life immunity. Trends in Immunology, 35(7), pp.299–310. Downing, J.R. & Shannon, K.M., 2002. Acute leukemia: A pediatric perspective. Cancer Cell, 2(6), pp.437–445. Dunn, G.P. et al., 2002. Cancer immunoediting: from immunosurveillance to tumor escape. Nature immunology, 3(11), pp.991–998. Dunn, G.P. et al., 2005. Interferon-gamma and cancer immunoediting. Immunologic research, 32(1–3), pp.231–245. Dunn, G.P., Koebel, C.M. & Schreiber, R.D., 2006. Interferons, immunity and cancer 	   173 immunoediting. Nature Reviews Immunology, 6(11), pp.836–848. Duque-Afonso, J. et al., 2015. Comparative genomics reveals multistep pathogenesis of E2A-PBX1 acute lymphoblastic leukemia. Journal of Clinical Investigation, 125(9), pp.3667–3680. Vanden Eijnden, S. et al., 2006. Preferential production of the IL-12(p40)/IL-23(p19) heterodimer by dendritic cells from human newborns. European Journal of Immunology, 36(1), pp.21–26. Emoto, M. et al., 2010. α-galactosylceramide promotes killing of Listeria monocytogenes within the macrophage phagosome through invariant NKT-cell activation. Infection and Immunity, 78(6), pp.2667–2676. Faderl, S. et al., 2010. Adult acute lymphoblastic leukemia. Cancer, 116(1), pp.1165–1176.  Farber, J.M. & Peterkin, P.I., 1991. Listeria monocytogenes, a food-borne pathogen. Microbiological reviews, 55(3), pp.476–511.  Farrar, M.A. & Schreiber, R.D., 1993. The molecular cell biology of interferon-g and its receptor. Annu Rev Immunol, 11, pp.571–611.  Fasching, K. et al., 2000. Presence of clone-specific antigen receptor gene rearrangements at birth indicates an in utero origin of diverse types of early childhood acute lymphoblastic leukemia. Blood, 95(8), pp.2722–2724. Feldhahn, N. et al., 2007. Activation-induced cytidine deaminase acts as a mutator in BCR-ABL1-transformed acute lymphoblastic leukemia cells. The Journal of experimental medicine, 204(5), pp.1157–1166. Feldhahn, N. et al., 2005. Deficiency of Bruton’s tyrosine kinase in B-cell precursor 	   174 leukemia cells. Proc Natl Acad Sci U S A, 102(37), pp.13266–13271.  Fielding, A.K., 2011. Current treatment of Philadelphia chromosome-positive acute lymphoblastic leukemia. Hematology / the Education Program of the American Society of Hematology. American Society of Hematology. Education Program, 2011, pp.231–237. Fischer, M. et al., 2005. Defining the oncogenic function of the TEL/AML1 (ETV6/RUNX1) fusion protein in a mouse model. Oncogene, 24(51), pp.7579–7591. Fleming, H.E. & Paige, C.J., 2002. Cooperation between IL-7 and the pre-B-cell receptor: a key to B-cell selection. Semin Immunol, 14(6), pp.423–430.  Ford, A.M. et al., 2001. Origins of “late” relapse in childhood acute lymphoblastic leukemia with TEL-AML1 fusion genes. Blood, 98(3), pp.558–564. Ford, A.M. et al., 2009. The TEL-AML1 leukemai fusion gene dysregulates the TGF-beta pathway in early B lineage progenitor cells. J. Clin. Invest., 119(4), pp.826-836. Ford, A.M. et al., 1998. Fetal origins of the TEL-AML1 fusion gene in identical twins with leukemia. Proc Natl Acad Sci U S A, 95(8), pp.4584–4588. Ford, A.M. et al., 2015. Protracted dormancy of pre-leukemic stem cells. Leukemia, 29, pp.2202-2207. Foroni, L. et al., 1999. Investigation of minimal residual disease in childhood and adult acute lymphoblastic leukaemia by molecular analysis. British journal of haematology, 105(1), pp.7–24. Francis, S.S. et al., 2016. In utero cytomegalovirus infection and development of childhood acute lymphoblastic leukemia. Blood. E-published ahead of print. 	   175 Gale, K.B. et al., 1997. Backtracking leukemia to birth: identification of clonotypic gene fusion sequences in neonatal blood spots. Proc Natl Acad Sci U S A, 94(25), pp.13950–13954.  Garvy, B.A. & Riley, R.L., 1994. IFN-gamma abrogates IL-7-dependent proliferation in pre-B-cells, coinciding with onset of apoptosis. Immunology, 81(3), pp.381–388.  Gautier, G. et al., 2005. A type I interferon autocrine-paracrine loop is involved in Toll-like receptor-induced interleukin-12p70 secretion by dendritic cells. The Journal of experimental medicine, 201(9), pp.1435–1446. Geier, J.K. & Schlissel, M.S., 2006. Pre-BCR signals and the control of Ig gene rearrangements. Seminars in Immunology, 18(1), pp.31–39. Gilham, C. et al., 2005. Day care in infancy and risk of childhood acute lymphoblastic leukaemia: findings from UK case-control study. BMJ, 330(7503), p.1294.  Godder, K.T. et al., 2007. Long term disease-free survival in acute leukemia patients recovering with increased gammadelta T-cells after partially mismatched related donor bone marrow transplantation. Bone marrow transplantation, 39(12), pp.751–757. Goto, H., 2015. Childhood relapsed acute lymphoblastic leukemia: Its biology and recent progress of treatment. Pediatrics International, 57(6), pp.1059–1066. Grawunder, U. et al., 1998. DNA Ligase IV Is Essential for V(D)J Recombination and DNA Double-Strand Break Repair in Human Precursor Lymphocytes. Molecular Cell, 2(4), pp.477–484.  Greaves, M., 2006. Infection, immune responses and the aetiology of childhood leukaemia. Nature Reviews Cancer, 6(3), pp.193–203.  	   176 Greaves, M. & Maley, C.C., 2012. Clonal evolution in cancer. Nature, 481(7381), pp.306–313. Greaves, M.F. et al., 2003. Leukemia in twins: Lessons in natural history. Blood, 102(7), pp.2321–2333. Greaves, M.F., 1988. Speculations on the cause of childhood acute lymphoblastic leukemia. Leukemia, 2(2), pp.120–125.  Greaves, M.F. & Wiemels, J., 2003. Origins of chromosome translocations in childhood leukaemia. Nat Rev Cancer, 3(9), pp.639–649.  Grupp, S.A. et al., 2013. Chimeric Antigen Receptor–Modified T-cells for Acute Lymphoid Leukemia. New England Journal of Medicine, 368(16), pp.1509–1518.  Haas, J.D. et al., 2012. Development of Interleukin-17-Producing γδ T-cells Is Restricted to a Functional Embryonic Wave. Immunity, 37(1), pp.48–59. Haas, W., Pereira, P. & Tonegawa, S., 1993. Gamma/delta cells. Annual review of immunology, 11, pp.637–685.  Hamon, M., Bierne, H. & Cossart, P., 2006. Listeria monocytogenes: a multifaceted model. Nature reviews. Microbiology, 4(6), pp.423–434. Han, S. et al., 2010. Polymorphisms in innate immunity genes and risk of childhood leukemia. Human Immunology, 71(7), pp.727–730.  Hanahan, D. & Weinberg, R.A., 2000. The hallmarks of cancer. Cell, 100(1), pp.57–70.  Hardy, R.R. & Hayakawa, K., 2001. B-cell development pathways. Annual review of immunology, 19, pp.595–621.  Harrison, C.J., 2011. Acute lymphoblastic leukemia. Clinics in Laboratory Medicine, 31(4), pp.631–647. 	   177 Haskó, G. et al., 2000. Adenosine inhibits IL-12 and TNF-[alpha] production via adenosine A2a receptor-dependent and independent mechanisms. The FASEB journal, 14(13), pp.2065–2074. Heisterkamp, N. et al., 1985. Structural organization of the bcr gene and its role in the Ph’ translocation. Nature, 315, pp.758–761. Heizmann, B., Kastner, P. & Chan, S., 2013. Ikaros is absolutely required for pre-B-cell differentiation by attenuating IL-7 signals. The Journal of experimental medicine, 210(13), pp.2823–32.  Hendriks, R.W. & Middendorp, S., 2004. The pre-BCR checkpoint as a cell-autonomous proliferation switch. Trends in Immunology, 25(5), pp.249–256. Hervas-Stubbs, S. et al., 2011. Direct effects of type I interferons on cells of the immune system. Clinical Cancer Research, 17(9), pp.2619–2627. Herzog, S., Reth, M. & Jumaa, H., 2009. Regulation of B-cell proliferation and differentiation by pre-B-cell receptor signalling. Nature reviews. Immunology, 9(3), pp.195–205. Hesse, J. et al., 2015. Prevalence of antibodies to Epstein-Barr virus (EBV) in childhood and adolescence in Denmark. Scandinavian journal of infectious diseases, 15(4), pp.335–338. Hjalgrim, L.L. et al., 2002. Presence of clone-specific markers at birth in children with acute lymphoblastic leukaemia. British journal of cancer, 87(9), pp.994–999. Holmfeldt, L. et al., 2013. The genomic landscape of hypodiploid acute lymphoblastic leukemia. Nature genetics, 45(3), pp.242–252.  Hong, D. et al., 2008. Initiating and cancer-propagating cells in TEL-AML1-associated 	   178 childhood leukemia. Science, 319(5861), pp.336–339.  Hsu, W.L. et al., 2013. The incidence of leukemia, lymphoma and multiple myeloma among atomic bomb survivors: 1950-2001. Radiation Research, 179(3), pp.361–382.  Hunger, S.P. & Mullighan, C.G., 2015. Acute Lymphoblastic Leukemia in Children. The New England journal of medicine, 373(16), pp.1541–52. Huygens, A. et al., 2014. Immunity to cytomegalovirus in early life. Frontiers in Immunology, 5, pp.552-558. Ikeda, H., Old, L.J. & Schreiber, R.D., 2002. The roles of IFNy in protection against tumor development and cancer immunoediting. Cytokine and Growth Factor Reviews, 13(2), pp.95–109. Inaba, H., Greaves, M. & Mullighan, C.G., 2013. Acute lymphoblastic leukaemia. Lancet, 381(9881), pp.1943–1955. Infante-Rivard, C., Fortier, I. & Olson, E., 2000. Markers of infection, breast-feeding and childhood acute lymphoblastic leukaemia. British Journal of Cancer, 83(11), pp.1559-1563. Ito, C. et al., 1999. Hyperdiploid acute lymphoblastic leukemia with 51 to 65 chromosomes: a distinct biological entity with a marked propensity to undergo apoptosis. Blood, 93(1), pp.315–320. Iwakura, Y. & Ishigame, H., 2006. The IL-23 / IL-17 axis in inflammation. The Journal of Clinical Investigation, 116(5), pp.1218–1222. Iwamoto, T. et al., 1991.  Preferential development of pre-B lymphomas with drastically down-regulated N-myc in the Emu-ret transgenic mice. Eur. J. Immunol, 21, pp. 	   179 1809. Jabbour, E.J., Faderl, S. & Kantarjian, H.M., 2005. Adult Acute Lymphoblastic Leukemia. Mayo Clinic Proceedings, 80(11), pp.1517–1527.  Jamil, A. et al., 2000. TEL/AML-1 fusion gene. Cancer Genetics and Cytogenetics, 122(2), pp.73–78. Jamil, A. et al., 2000. TEL/AML-1 fusion gene. its frequency and prognostic significance in childhood acute lymphoblastic leukemia. Cancer genetics and cytogenetics, 122(2), pp.73–78. Jeha, S. et al., 2009. Increased risk for CNS relapse in pre-B-cell leukemia with the t(1;19)/TCF3-PBX1. Leukemia, 23(8), pp.1406–1409.  Jenson, H.B., 2011. Epstein-Barr Virus. Pediatrics in Review, 32(9), pp.375–384.  Jiang, Q. et al., 2005. Cell biology of IL-7, a key lymphotrophin. Cytokine and Growth Factor Reviews, 16(4–5), pp.513–533. Joseph, S.A. et al., 2006. Cytomegalovirus as an occupational risk in daycare educators. Paediatrics and Child Health, 11(7), pp.401–407. Jourdan-Da Silva, N. et al., 2004. Infectious diseases in the first year of life, perinatal characteristics and childhood acute leukaemia. British Journal of Cancer, 90(1), pp.139-145. Kabelitz, D. et al., 2005. Epithelial defence by gamma delta T-cells. International archives of allergy and immunology, 137(1), pp.73–81.  Kaplan, D.H. et al., 1998. Demonstration of an interferon γ-dependent tumor surveillance system in immunocompetent mice. Immunology, 95(June), pp.7556–7561. Kellett, C.E., 1937. Acute myeloid leukaemia in one of identical twins. Archives of 	   180 disease in childhood, 12(70), pp.239–252.  Kim, R., Emi, M. & Tanabe, K., 2007. Cancer immunoediting from immune surveillance to immune escape. Immunology, 121(1), pp.1–14.  Kinlen, L., 1988. Evidence for an infective cause of childhood leukaemia: Comparison of a Scottish new town with nuclear reprocessing sites in Britain. The Lancet, 332(8624), pp.1323–1327. Kinlen, L., 2015. Population mixing and childhood leukaemia. European Journal of Epidemiology, 30(12), p.1331. Kinlen, L.J. & Petridou, E., 1995. Childhood leukemia and rural population movements: Greece, Italy, and other countries. Cancer Causes and Control, 6(5), pp.445–450. Klein, F. et al., 2004. The BCR-ABL1 kinase bypasses selection for the expression of a pre-B-cell receptor in pre-B acute lymphoblastic leukemia cells. J Exp Med, 199(5), pp.673–685.  Kollmann, T.R. et al., 2012. Innate Immune Sensing by Toll-like Receptors in Newborns and the Elderly. Immunity, 37(5), pp.771–783. Korn, T. et al., 2009. IL-17 and Th17 Cells. Annual review of immunology, 27(1), pp.485–517. Krmpotic, A. et al., 2003. Pathogenesis of murine cytomegalovirus infection. Microbes and Infection, 5(13), pp.1263–1277. Kumar, H., Kawai, T. & Akira, S., 2009. Toll-like receptors and innate immunity. Biochemical and Biophysical Research Communications, 388(4), pp.621–625. Laird, R.M. et al., 2013. γδ T-cells Acquire Effector Fates in the Thymus and Differentiate into Cytokine-Producing Effectors in a Listeria Model of Infection 	   181 Independently of CD28 Costimulation. PLoS ONE, 8(5), pp.1-15. Langrish, C.L. et al., 2005. IL-23 drives a pathogenic T-cell population that induces autoimmune inflammation. The Journal of experimental medicine, 201(2), pp.233–240.  Lausten-Thomsen, U. et al., 2010. Increased risk of ALL among premature infants is not explained by increased prevalence of pre-leukemic cell clones. Blood Cells, Molecules, and Diseases, 44(3), pp.188–190. Lavoie, P.M.M. et al., 2010. Profound Lack of Interleukin (IL)–12/IL‐23p40 in Neonates Born Early in Gestation Is Associated with an Increased Risk of Sepsis. The Journal of Infectious Diseases, 202(11), pp.1754–1763.  Lee, E.M., Bachmann, P.S. & Lock, R.B., 2007. Xenograft models for the preclinical evaluation of new therapies in acute leukemia. Leukemia & lymphoma, 48(April), pp.659–668. Leung, A.K.C., Sauve, R.S. & Davies, H.D., 2003. Congenital cytomegalovirus infection. Journal of the National Medical Association, 95(3), pp.213–218.  Levy, O., 2007. Innate immunity of the newborn: basic mechanisms and clinical correlates. Nat Rev Immunol, 7(5), pp.379–390. Levy, O. et al., 2006. The adenosine system selectively inhibits TLR-mediated TNF-alpha production in the human newborn. Journal of immunology, 177(3), pp.1956–1966.  Li, W., Swanson, P.C. & Desiderio, S., 1997. RAG-1 and RAG-2-dependent assembly of functional complexes with V(D)J recombination substrates in solution. Molecular and cellular biology, 17(12), pp.6932–6939. 	   182 Linabery, A.M. & Ross, J.A., 2008. Trends in childhood cancer incidence in the U.S. (1992-2004). Cancer, 112(2), pp.416–432. Locatelli, F. et al., 2012. How i treat relapsed childhood acute lymphoblastic leukemia. Blood, 120(14), pp.2807–2816. Longo, D.L., Hunger, S.P. & Mullighan, C.G., 2015. Acute Lymphoblastic Leukemia in Children. New England Journal of Medicine, 373(16), pp.1541–1552.  Lopez, R.G. et al., 1999. TEL is a sequence-specific transcriptional repressor. Journal of Biological Chemistry, 274(42), pp.30132–30138. Ma, X. et al., 2002. Daycare attendance and risk of childhood acute lymphoblastic leukaemia. British Journal of Cancer, 86(9), pp.1419–1424.  Ma, X. et al., 2005. Ethnic difference in daycare attendance, early infections, and risk of childhood acute lymphoblastic leukemia. Cancer Epidemiology Biomarkers and Prevention, 14(8), pp.1928–1934. Ma, X. et al., 2010. Infection and pediatric acute lymphoblastic leukemia. Blood cells, Molecules and Disease, 42(2), pp.117-120. Ma, Y. et al., 2013. Developmental timing of mutations revealed by whole-genome sequencing of twins with acute lymphoblastic leukemia. Proc Natl Acad Sci U S A, 110(18), pp.7429-7433. MacKenzie, J. et al., 2006. The putative role of transforming viruses in childhood acute lymphoblastic leukemia. Haematologica, 91, pp.240-243. Maia, A.T. et al., 2003. Prenatal origin of hyperdiploid acute lymphoblastic leukemia in identical twins. Leukemia, 17(11), pp.2202–2206.  Maloney, K. et al., 1999. TEL-AML1 fusion identifies a subset of children with standard 	   183 risk acute lymphoblastic leukemia who have an excellent prognosis when treated with therapy that includes a single delayed intensification. Leukemia, 13(11), pp.1708–1712.  Mancini, M. et al., 2005. A comprehensive genetic classification of adult acute lymphoblastic leukemia (ALL): Analysis of the GIMEMA 0496 protocol. Blood, 105(9), pp.3434–3441. Mandel, E.M. & Grosschedl, R., 2010. Transcription control of early B-cell differentiation. Current Opinion in Immunology, 22(2), pp.161–167. Marchant, A. & Kollmann, T.R., 2015. Understanding the ontogeny of the immune system to promote immune-mediated health for life. Frontiers in Immunology, 6. Mårtensson, I.L. et al., 2010. The pre-B-cell receptor checkpoint. FEBS letters, 584(12), pp.2572–2579.  Martin-Lorenzo, A. et al., 2015. Infection exposure is a causal factor in B-precursor acute lymphoblastic leukemia as a result of Pax5 inherited susceptibility. Cancer discovery, 5(12), pp.1328-1343. Martin, B. et al., 2009. Interleukin-17-Producing  γδ T-cells Selectively Expand in Response to Pathogen Products and Environmental Signals. Immunity, 31(2), pp.321–330. Mattner, J. et al., 2005. Exogenous and endogenous glycolipid antigens activate NKT-cells during microbial infections. Nature, 434(7032), pp.525–529.  Maude, S.L. et al., 2015. CD19-targeted chimeric antigen receptor T-cell therapy for acute lymphoblastic leukemia. Blood, 125(26), pp.4017–4023. McEwen, J.E. & Reilly, P.R., 1994. Stored Guthrie cards as DNA “banks”. American 	   184 journal of human genetics, 55(1), pp.196–200.  McWhirter, J.R. et al., 1999. Oncogenic homeodomain transcription factor E2A-Pbx1 activates a novel WNT gene in pre-B acute lymphoblastoid leukemia. Proc Natl Acad Sci U S A, 96(20), pp.11464–11469.  Medzhitov, R., 2001. Toll-like receptors and innate immunity. Nature reviews. Immunology, 1(2), pp.135–145. Mestas, J. & Hughes, C.C.W., 2015. Mouse and Human Immunology Of Mice and Not Men: Differences between Of Mice and Not Men: Differences between Mouse and Human Immunology. J Immunol References, 172, pp.2731–2738.  Meyer, L.H. & Debatin, K.-M., 2011. Diversity of human leukemia xenograft mouse models: implications for disease biology. Cancer research, 71(23), pp.7141–7144.  Miossec, P., Korn, T. & Kuchroo, V.K., 2009. Interleukin-17 and type 17 helper T-cells. The New England journal of medicine, 361(9), pp.888–898. Mitchell, C. et al., 2010. Long-term follow-up of the United Kingdom medical research council protocols for childhood acute lymphoblastic leukaemia, 1980-2001. Leukemia, 24(2), pp.406–418.  Mombaerts, P. et al., 1992. RAG-1-deficient mice have no mature B and T lymphocytes. Cell, 68(5), pp.869–877. Moresco, E.M.Y., LaVine, D. & Beutler, B., 2011. Toll-like receptors. Current Biology, 21(13), pp.488-493. Mori, H. et al., 2002. Chromosome translocations and covert leukemic clones are generated during normal fetal development. Proc Natl Acad Sci U S A, 99(12), pp.8242–8247.  	   185 Moriyama, T. et al., 2015. Inherited genetic variation in childhood acute lymphoblastic leukemia. Blood, 125(26), pp.3398-3995. Morrow, M. et al., 2007. TEL-AML1 preleukemic activity requires the DNA binding domain of AML1 and the dimerization and corepressor binding domains of TEL. Oncogene, 26(30), pp.4404–4414. Morrow, M. et al., 2004. TEL-AML1 promotes development of specific hematopoietic lineages consistent with preleukemic activity. Blood, 103(10), pp.3890–3896. Mulligan, L.M. et al., 1993. Germ-line mutations of the RET proto-oncogene in multiple endocrine neoplasia type 2A. Nature, 363(6428), pp.458–460.  Mullighan, C.G. et al., 2011. CREBBP mutations in relapsed acute lymphoblastic leukaemia. Nature, 471(7337), pp.235–239.  Mullighan, C.G. et al., 2008. Genomic analysis of the clonal origins of relapsed acute lymphoblastic leukemia. Science, 322(5906), pp.1377–1380.  Mullighan, C.G., 2012. The molecular genetic makeup of acute lymphoblastic leukemia. Hematology, 2012, pp.389–396.  Nagorsen, D. et al., 2012. Blinatumomab: A historical perspective. Pharmacology and Therapeutics, 136(3), pp.334–342. Nahar, R. & Müschen, M., 2009. Pre-B-cell receptor signaling in acute lymphoblastic leukemia. Cell Cycle, 8(23), pp.3874–3877.  Nowell, P.C., 1976. The clonal evolution of tumor cell populations. Science, 194(4260), pp.23–28. Nutt, S.L. et al., 1999. Independent regulation of the two Pax5 alleles during B-cell development. Nature genetics, 21(4), pp.390–395. 	   186 O’Garra, A. & Arai, N., 2000. The molecular basis of T helper 1 and T helper 2 cell differentiation. Trends in Cell Biology, 10(12), pp.542–550. O’Riordan, M. & Grosschedl, R., 1999. Coordinate regulation of B-cell differentiation by the transcription factors EBF and E2A. Immunity, 11(1), pp.21–31. Olivadoti, M. et al., 2007. Murine gammaherpesvirus 68: A model for the study of Epstein-Barr virus infections and related diseases. Comparative Medicine, 57(1), pp.44–50. Onciu, M., 2009. Acute Lymphoblastic Leukemia. Hematology/Oncology Clinics of North America, 23(4), pp.655–674. Onodera, N., McCabe, N.R. & Rubin, C.M., 1992. Formation of a hyperdiploid karyotype in childhood acute lymphoblastic leukemia. Blood, 80(1), pp.203–8. Pamer, E.G., 2004. Immune responses to Listeria monocytogenes. Nat Rev Immunol, 4(10), pp.812–823.  Pang, S.H.M., Carotta, S. & Nutt, S.L., 2014. Transcriptional control of pre-B-cell development and leukemia prevention. Current Topics in Microbiology and Immunology, 381, pp.189–213. Panzer-Grumayer, E.R. et al., 2005. Immunogenotype changes prevail in relapses of young children with TEL-AML1-positive acute lymphoblastic leukemia and derive mainly from clonal selection. Clinical Cancer Research, 11(21), pp.7720–7727. Panzer-Grumayer, E.R. et al., 2000. Rapid molecular response during early induction chemotherapy predicts a good outcome in childhood acute lymphoblastic leukemia. Blood, 95(3), pp.790–794.  Papaemmanuil, E. et al., 2009. Loci on 7p12.2, 10q21.2, 14q11.2 are associated with risk 	   187 of childhood acute lymphoblastic leukemia. Nature Genetics, 41, pp.1006-1010. Papaemmanuil, E. et al., 2014. RAG-mediated recombination is the predominant driver of oncogenic rearrangement in ETV6-RUNX1 acute lymphoblastic leukemia. Nature Genetics, 46, pp.116-125. Paulsson, K. et al., 2010. Genetic landscape of high hyperdiploid childhood acute lymphoblastic leukemia. Proc Natl Acad Sci U S A, 107(50), pp.21719–21724. Paulsson, K. et al., 2015. The genomic landscape of high hyperdiploid childhood acute lymphoblastic leukemia. Nature Genetics, 47(6), pp.672–677.  Paulsson, K. & Johansson, B., 2009. High hyperdiploid childhood acute lymphoblastic leukemia. Genes, chromosomes & cancer, 48(8), pp.637–660. Pear, W.S. & Aster, J.C., 2004. T-cell acute lymphoblastic leukemia/lymphoma: a human cancer commonly associated with aberrant NOTCH1 signaling. Current opinion in hematology, 11(6), pp.426–433. Peham, M. et al., 2004. Clonal Variation of the Immunogenotype in Relapsed ETV6/RUNX1-Positive Acute Lymphoblastic Leukemia Indicates Subclone Formation during Early Stages of Leukemia Development. Genes Chromosomes and Cancer, 39(2), pp.156–160. Perrillat, F. et al., 2002. Day-care, early common infections and childhood acute leukaemia: a multicentre French case-control study. British journal of cancer, 86(7), pp.1064–1069. Pieters, R. & Carroll, W.L., 2010. Biology and Treatment of Acute Lymphoblastic Leukemia. Hematology/Oncology Clinics of North America, 24(1), pp.1–18. Poynton, F.J., Thursfield, H., and Paterson, D., 1922. The severe blood diseases of 	   188 childhood: a series of observations from the Hospital for Sick Children, Great Ormond Street. British J Child Dis, pp.128–144. Prak, E.L. & Weigert, M., 1995. Light chain replacement: a new model for antibody gene rearrangement. The Journal of experimental medicine, 182(2), pp.541–548.  Preston, D.L. et al., 1994. Cancer incidence in atomic bomb survivors. Part III. Leukemia, lymphoma and multiple myeloma, 1950-1987. Radiation research, 137(2), pp.68–97. Pridans, C. et al., 2008. Identification of Pax5 Target Genes in Early B-cell Differentiation. The Journal of Immunology, 180(3), pp.1719–1728.  Pui, C.-H. & Evans, W.E., 2006. Treatment of acute lymphoblastic leukemia. The New England journal of medicine, 354(2), pp.166–178. Pui, C.-H., Relling, M. V. & Downing, J.R., 2004. Acute Lymphoblastic Leukemia. New England Journal of Medicine, 350(15), pp.1535–1548.  Pui, C.-H., Robison, L.L. & Look, A.T., 2008. Acute lymphoblastic leukaemia. Lancet, 371(9617), pp.1030–1043. Pui, C.H. & Campana, D., 2000. New definition of remission in childhood acute lymphoblastic leukemia. Leukemia, 14(5), pp.783–785. Pui, C.H., Robison, L.L. & Look, A.T., 2008. Acute lymphoblastic leukaemia. Lancet, 371(9617), pp.1030–1043.  Qin, H. et al., 2009. TGF-beta promotes Th17 cell development through inhibition of SOCS3. Journal of Immunology, 183(1), pp.97–105.  Rabin, K.R. & Poplack, D.G., 2011. Management strategies in acute lymphoblastic leukemia. Oncology, 25(4), pp.328–335. 	   189 Rachitskaya, A. V et al., 2008. Cutting edge: NKT-cells constitutively express IL-23 receptor and RORgammat and rapidly produce IL-17 upon receptor ligation in an IL-6-independent fashion. Journal of Immunology, 180(8), pp.5167–5171.  Raetz, E. a & Bhatla, T., 2012. Where do we stand in the treatment of relapsed acute lymphoblastic leukemia? Hematology Am Soc Hematol Educ Program, 2012, pp.129–136. Ranson, T. et al., 2005. Invariant V alpha 14+ NKT-cells participate in the early response to enteric Listeria monocytogenes infection. Journal of immunology (Baltimore, Md.  : 1950), 175(2), pp.1137–1144. Rao, D.G. et al., 2013. Acute Lymphoblastic Leukemia in Children. J Indian Aca Oral Med Radiol, 25(3), pp.178–183. Reece, P. et al., 2011. Maternal allergy modulates cord blood hematopoietic progenitor Toll-like receptor expression and function. Journal of Allergy and Clinical Immunology, 127(2), pp.447–453. Renate Panzer-Grümayer, E. et al., 2002. Nondisjunction of chromosomes leading to hyperdiploid childhood B-cell precursor acute lymphoblastic leukemia is an early event during leukemogenesis. Blood, 100(1), pp.347–349. Revilla-I-Domingo, R. et al., 2012. The B-cell identity factor Pax5 regulates distinct transcriptional programmes in early and late B lymphopoiesis. The EMBO journal, 31(14), pp.3130–3146.  Richardson, R.B., 2011. Promotional etiology for common childhood acute lymphoblastic leukemia: The infective lymphoid recovery hypothesis. Leukemia Research, 35(11), pp.1425–1431. 	   190 Richmond, A. & Su, Y., 2008. Mouse xenograft models vs GEM models for human cancer therapeutics. Dis Model Mech, 1(2­3), pp.78–82.  Rickert, R.C., 2013. New insights into pre-BCR and BCR signalling with relevance to B-cell malignancies. Nature reviews. Immunology, 13(8), pp.578–591.  Roark, C.L. et al., 2008. gammadelta T-cells: an important source of IL-17. Current opinion in immunology, 20(3), pp.353–357.  Roberts, K.G. et al., 2014. Targetable kinase-activating lesions in Ph-like acute lymphoblastic leukemia. The New England journal of medicine, 371(11), pp.1005–1015.  Romagnani, S., 2004. Immunologic influences on allergy and the TH1/TH2 balance. Journal of Allergy and Clinical Immunology, 113(3), pp.395–400. Roman, E. et al., 2007. Childhood acute lymphoblastic leukemia and infections in the first year of life: a report from the United Kingdom Childhood Cancer Study. Am.J.Epidemiol., 165, pp.496–504. Ross, S.A. & Boppana, S.B., 2005. Congenital cytomegalovirus infection: Outcome and diagnosis. Seminars in Pediatric Infectious Diseases, 16(1), pp.44–49. Rowley, J.D., 1973. Letter: A new consistent chromosomal abnormality in chronic myelogenous leukaemia identified by quinacrine fluorescence and Giemsa staining. Nature, 243(5405), pp.290–293.  Rudant, J. et al., 2015. Childhood Acute Lymphoblastic Leukemia and Indicators of Early Immune Stimulation: A Childhood Leukemia International Consortium Study. American Journal of Epidemiology, 181(8), pp.549–562. Saha, V. et al., 1993. Determinants of symptom interval in childhood cancer. Archives of 	   191 Disease in Childhood, 68(6), pp.771–774.  Sanjuan-pla, A. et al., 2015. Revisiting the biology of infant t(4;11)/MLL-AF41 B-cell acute lymphoblastic leukemia. Blood, 126(25), pp.2676–2686. Sawada, A. et al., 2009. Significance of four MRD markers in MRD-based treatment strategy for childhood acute lymphoblastic leukemia. Leukemia Research, 33(12), pp.1710–1713. Schatz, D.G., Oettinger, M.A. & Baltimore, D., 1989. The V(D)J recombination activating gene, RAG-1. Cell, 59(6), pp.1035–1048. Schindler, J.W. et al., 2009. TEL-AML1 corrupts hematopoietic stem cells to persist in the bone marrow and initiate leukemia. Cell Stem Cell, 5, pp.43-53. Schmiegelow, K. et al., 2008. Etiology of common childhood acute lymphoblastic leukemia: the adrenal hypothesis. Leukemia, 22(12), pp.2137–2141.  Schreiber, R.D., Old, L.J. & Smyth, M.J., 2011. Cancer Immunoediting: Integrating Immunity’s Roles in Cancer Suppression and Promotion. Science, 331(6024), pp.1565–1570.  Schroder, K. et al., 2004. Interferon- y  : an overview of signals , mechanisms and functions. Journal of leukocyte Biology, 75(2), pp.163–189. Schuh, J.M. et al., 2003. Chemokines and cytokines: Axis and allies in asthma and allergy. Cytokine and Growth Factor Reviews, 14(6), pp.503–510. Schultz, K.R. et al., 2007. Risk- and response-based classification of childhood B-precursor acute lymphoblastic leukemia: A combined analysis of prognostic markers from the Pediatric Oncology Group (POG) and Children’s Cancer Group (CCG). Blood, 109(3), pp.926–935. 	   192 Shankaran, V. et al., 2001. IFNγ and lymphocytes prevent primary tumour development and shape tumour immunogenicity. Nature, 410(6832), pp.1107–1111. Shanks, N. et al., 2000. Early-life exposure to endotoxin alters hypothalamic-pituitary-adrenal function and predisposition to inflammation. Proc Natl Acad Sci U S A, 97(10), pp.5645–5650.  Sheppard, R.D. et al., 1998. Transgenic N-myc mouse model for indolent B-cell lymphoma: tumor characterization and analysis of genetic alterations in spontaneous and retrovirally accelerated tumors. Oncogene, 17(16), pp.2073–2085.  Shurtleff, S.A. et al., 1995. TEL/AML1 fusion resulting from a cryptic t(12;21) is the most common genetic lesion in pediatric ALL and defines a subgroup of patients with an excellent prognosis. Leukemia, 9(12), pp.1985–1989. Smith, M.A. et al., 2010. Outcomes for children and adolescents with cancer: Challenges for the twenty-first century. Journal of Clinical Oncology, 28(15), pp.2625–2634. Staras, S.A.S. et al., 2008. Cytomegalovirus seroprevalence and childhood sources of infection: A population-based study among pre-adolescents in the United States. Journal of Clinical Virology, 43(3), pp.266–271. Starr, R. et al., 1998. Liver degeneration and lymphoid deficiencies in mice lacking suppressor of cytokine signaling-1. Proc Natl Acad Sci U S A, 95(24), pp.14395–14399. Stockinger, B. & Veldhoen, M., 2007. Differentiation and function of Th17 T-cells. Current Opinion in Immunology, 19(3), pp.281–286. Street, S.E.A. et al., 2002. Suppression of lymphoma and epithelial malignancies effected by interferon gamma. The Journal of experimental medicine, 196(1), pp.129–34.  	   193 Street, S.E., Cretney, E. & Smyth, M.J., 2001. Perforin and interferon-gamma activities independently control tumor initiation, growth, and metastasis. Blood, 97(1), pp.192–197.  Strunk, T. et al., 2011. Innate immunity in human newborn infants: prematurity means more than immaturity. The journal of maternal-fetal & neonatal medicine, 24(1), pp.25–31.  Styczynski, J. et al., 2007. Predictive value of multidrug resistance proteins and cellular drug resistance in childhood relapsed acute lymphoblastic leukemia. Journal of Cancer Research and Clinical Oncology, 133(11), pp.875–893. Swaminathan, S. et al., 2015. Mechanisms of clonal evolution in childhood acute lymphoblastic leukemia. Nature immunology, 16(7), pp.766–774.  Takeda, K., Kaisho, T. & Akira, S., 2003. Toll-Like Receptors. Annual Review of Immunology, 21(1), pp.335–376. Talamo, A. et al., 2010. Clonal heterogeneity and chromosomal instability at disease presentation in high hyperdiploid acute lymphoblastic leukemia. Cancer Genetics and Cytogenetics, 203(2), pp.209–214. Tang, J. et al., 1997. TGF-beta down-regulates stromal IL-7 secretion and inhibits proliferation of human B-cell precursors. Journal of Immunology, 159(1), pp.117–125.  Taub, J.W. et al., 2002. High frequency of leukemic clones in newborn screening blood samples of children with B-precursor acute lymphoblastic leukemia. Blood, 99(8), pp.2992–2996. Taylor, G.M. & Birch, J.M., 1996. The hereditary basis of human leukemia. Leukemia, 6, 	   194 pp.210–245. Teitell, M.A. & Pandolfi, P.P., 2009. Molecular genetics of acute lymphoblastic leukemia. Annual review of pathology, 4, pp.175–198.  Teuffel, O. et al., 2004. Prenatal origin of separate evolution of leukemia in identical twins. Leukemia, 18(10), pp.1624–1629.  Topp, M.S. et al., 2011. Targeted Therapy With the T-Cell-Engaging Antibody Blinatumomab of Chemotherapy-Refractory Minimal Residual Disease in B-Lineage Acute Lymphoblastic Leukemia Patients Results in High Response Rate and Prolonged Leukemia-Free Survival. Journal of clinical oncology, 29(18), pp.2493–2498.  Toyoda, Y. et al., 2000. Six months of maintenance chemotherapy after intensified treatment for acute lymphoblastic leukemia of childhood. Journal of Clinical Oncology, 18(7), pp.1508–1516. Tsuzuki, S. et al., 2004. Modeling first-hit functions of the t(12;21) TEL-AML1 translocation in mice. Proc Natl Acad Sci U S A, 101(22), pp.8443–8448.  Underhill, D.M. & Ozinsky, A., 2002. Toll-like receptors: Key mediators of microbe detection. Current Opinion in Immunology, 14(1), pp.103–110. Urayama, K.Y. et al., 2010. A meta-analysis of the association between day-care attendance and childhood acute lymphoblastic leukaemia. International Journal of Epidemiology, 39(3), pp.718–732. Urayama, K.Y. et al., 2011. Early life exposure to infections and risk of childhood acute lymphoblastic leukemia. International journal of cancer, 128(7), pp.1632–1643.  Van der Weyden, L. et al., 2011.  Modeling the evolution of ETV6-RUNX1-induced B-	   195 cell precursor acute lymphoblastic leukemia in mice. Blood, 118, pp.1041-1051. Van Vlierberghe, P. & Ferrando, A., 2012. The molecular es basis of T-cell acute lymphoblastic leukemia. The Journal of Clinical Investigation, 122(10), pp.3398–3406.  Vrooman, L.M. & Silverman, L.B., 2016. Treatment of Childhood Acute Lymphoblastic Leukemia: Prognostic Factors and Clinical Advances. Current Hematologic Malignancy Reports, 11(5), pp.385–394. Wasserman, R., Zeng, X.X. & Hardy, R.R., 1998. The evolution of B precursor leukemia in the Emu-ret mouse. Blood, 92, pp.273–282. Watford, W.T. et al., 2004. Signaling by IL-12 and IL-23 and the immunoregulatory roles of STAT4. Immunological Reviews, 202, pp.139–156. Weng, A.P. et al., 2004. Activating mutations of NOTCH1 in human T-cell acute lymphoblastic leukemia. Science, 306(5694), pp.269–271. Wiegering, V. et al., 2014. TGFβ and IL10 Have an Impact on Risk Group and Prognosis in Childhood ALL. Blood, 124(21), p.5337. Wiemels, J., 2012. Perspectives on the causes of childhood leukemia. Chemico-Biological Interactions, 196(3), pp.59–67. Wiemels, J.L. et al., 1999. Prenatal origin of acute lymphoblastic leukaemia in children. Lancet, 354(9189), pp.1499–1503. Winkler, B. et al., 2015. TGFb and IL10 Have an Impact on Risk Group and Prognosis in Childhood ALL. Peditric Blood and Cancer, 62, pp.72–79. Yeoh, E. et al., 2002. Classification, subtype discovery, and prediction of outcome in pediatric acute lymphoblastic leukemia by gene expression profiling. Cancer cell, 	   196 1(2), pp.133–143. Yoshimura, A., Naka, T. & Kubo, M., 2007. SOCS proteins, cytokine signalling and immune regulation. Nature Reviews Immunology, 7(6), pp.454–465. Young, F. et al., 1994. Influence of immunoglobulin heavy- and light-chain expression on B-cell differentiation. Genes and Development, 8(9), pp.1043–1057. Zelent, A., Greaves, M. & Enver, T., 2004. Role of the TEL-AML1 fusion gene in the molecular pathogenesis of childhood acute lymphoblastic leukaemia. Oncogene, 23(24), p.4275–4283. Zeng, X.X. et al., 1998. The fetal origin of B-precursor leukemia in the E-mu-ret mouse. Blood, 92(10), pp.3529–3536.  Zhou, Y. et al., 2012. Advances in the molecular pathobiology of B-lymphoblastic leukemia. Human Pathology, 43(9), pp.1347–1362. Zuna, J. et al., 1999. TEL/AML1 positivity in childhood ALL: average or better prognosis? Czech Paediatric Haematology Working Group. Leukemia, 13(1), pp.22–24.  Zuna, J. et al., 2004. TEL deletion analysis supports a novel view of relapse in childhood acute lymphoblastic leukemia. Clinical Cancer Research, 10(16), pp.5355–5360.  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0348676/manifest

Comment

Related Items