Identification and Characterisation of Lgr4/Β-Catenin Signalling in Acute Myeloid Leukaemic Stem Cells

Identification and Characterisation of Lgr4/β-catenin Signalling in Acute Myeloid Leukaemic Stem Cells

Hangyu Yi

Bachelor of Medical Science (Honours Class 1)

Submitted in fulfilment of the requirements for the degree of

Doctor of Philosophy

School of Women’s and Children’s Health

Faculty of Medicine

University of New South Wales

February 2016

PLEASE TYPE THE UNIVERSITY OF NEW SOUTH WALES Thesis/Dissertation Sheet

Surname or Family name: Yi

First name: Hangyu Other name/s:

Abbreviation for degree as given in the University calendar: PhD

School: School of Women’s and Children’s Health Faculty: Medicine

Title: Identification and characterisation of Lgr4/β- catenin signalling in acute myeloid leukaemic stem cells

Abstract 350 words maximum: (PLEASE TYPE) Acute myeloid leukaemia (AML) is a deadly form of leukaemia resulting in the highest number of leukaemia-associated deaths. The high mortality rate is due to frequent relapse caused by the persistence of drug-resistant leukaemic stem cells (LSCs). We have previously demonstrated an essential role for β-catenin signalling in regulating LSCs in AML.

Leucine-rich repeat containing G protein-coupled receptor 4 (Lgr4) has recently been identified as the receptor for R-spondin (Rspo) proteins to activate β-catenin. This study showed that Rspo2/Rspo3 cooperating with Wnt3a potently potentiated β-catenin activation in haematopoietic stem cell (HSC)-derived pre-LSCs. Overexpression of Lgr4 augmented activation of Wnt/β-catenin signalling and promoted leukaemogenesis in vivo. Inhibition of Lgr4 reduced β-catenin activity, completely abolished Rspo3/Wnt3a/β-catenin signalling and prevented leukaemia development. A microarray experiment of 104 AML patient samples showed that high Lgr4 expression was associated with poor outcomes of AML patients. Altogether, these findings provide strong evidence demonstrating Lgr4 to be a critical regulator of β-catenin signalling in AML.

Gene expression profiling identified Rgs1 (regulator of G protein signalling 1) to be a major component of Lgr4 signalling. Rgs1 has been shown to bind directly to G protein subunits Gαq and Gαi (Moratz et al., J Immunol, 2000), and treatment of pre-LSCs with Gαq and Gαi inhibitors indicated Gαq to be a key downstream effector of Lgr4/Rgs1 signalling. Further functional studies showed that Gαq knockdown reduced β-catenin expression, attenuated the effect of Wnt3a/Rspo3-potentiated β-catenin activation and impaired LSC self-renewal, recapitulating the role of Lgr4 in LSC regulation. These data support the view that Gαq is an integral component of Lgr4 signalling in LSCs.

Gene expression analysis also showed that Lgr4 knockdown increased expression of Gadd45a (growth arrest and DNA damage- inducible gene) and repressed several mitochondrial associated genes. Functional studies showed that Gadd45a deletion significantly increased in vivo LSC self-renewal and enhanced AML progression. Blockade of Lgr4 signalling inhibited mitochondrial energy metabolism, on which LSCs rely for survival.

Collectively, this study has identified a novel Wnt3a/Rspo2/Rspo3-Lgr4Gαqβ-catenin signalling pathway governing LSCs and interference with components of this pathway may represent a promising therapeutic approach for eradicating LSCs in AML.

Declaration relating to disposition of project thesis/dissertation

I hereby grant to the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or in part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all property rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation.

I also authorise University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstracts International (this is applicable to doctoral theses only).

…………………………………………………… ……………………………………..……… ……….……………………...… Signature Witness Date

The University recognises that there may be exceptional circumstances requiring restrictions on copying or conditions on use. Requests for restriction for a period of up to 2 years must be made in writing. Requests for a longer period of restriction may be considered in exceptional circumstances and require the approval of the Dean of Graduate Research.

FOR OFFICE USE ONLY Date of completion of requirements for Award:

THIS SHEET IS TO BE GLUED TO THE INSIDE FRONT COVER OF THE THESIS

Originality statement

‘I hereby declare that this submission is my own work and to the best of my knowledge it contains no materials previously published or written by another person, or substantial proportions of material which have been accepted for the award of any other degree or diploma at UNSW or any other educational institution, except where due acknowledgement is made in the thesis. Any contribution made to the research by others, with whom I have worked at UNSW or elsewhere, is explicitly acknowledged in the thesis. I also declare that the intellectual content of this thesis is the product of my own work, except to the extent that assistance from others in the project's design and conception or in style, presentation and linguistic expression is acknowledged.’

Signed ………………………………………

Date ………………………………………

‘I hereby grant the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all proprietary rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation.

I also authorise University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstract International (this is applicable to doctoral theses only).

I have either used no substantial portions of copyright material in my thesis or I have obtained permission to use copyright material; where permission has not been granted I have applied/will apply for a partial restriction of the digital copy of my thesis or dissertation.'

Signed ………………………………………

Date ………………………………………

Authenticity statement

‘I certify that the Library deposit digital copy is a direct equivalent of the final officially approved version of my thesis. No emendation of content has occurred and if there are any minor variations in formatting, they are the result of the conversion to digital format.’

Signed ………………………………………

Date ………………………………………

Abstract

Acute myeloid leukaemia (AML) is a deadly form of leukaemia resulting in the highest number of leukaemia-associated deaths. The high mortality rate is due to frequent relapse caused by the persistence of drug-resistant leukaemic stem cells (LSCs). We have previously demonstrated an essential role for β-catenin signalling in regulating LSCs in AML.

iii

Gene expression analysis also showed that Lgr4 knockdown increased expression of Gadd45a (growth arrest and DNA damage-inducible gene) and repressed several mitochondrial associated genes. Functional studies showed that Gadd45a deletion significantly increased in vivo LSC self-renewal and enhanced AML progression. Blockade of Lgr4 signalling inhibited mitochondrial energy metabolism, on which LSCs rely for survival.

Collectively, this study has identified a novel Wnt3a/Rspo2/Rspo3-Lgr4Gαqβ- catenin signalling pathway governing LSCs and interference with components of this pathway may represent a promising therapeutic approach for eradicating LSCs in AML.

1 Introduction ...... 1 1.1 Acute myeloid leukaemia ...... 2 1.1.1 Heterogeneity of AML ...... 4 1.1.1.1 Favourable genetic features ...... 8 1.1.1.2 Unfavourable genetic features ...... 9 1.1.1.3 MLL-rearranged AML ...... 10 1.1.2 Treatment of AML ...... 13 1.1.2.1 Current treatment regimens ...... 13 1.1.2.2 Targeted therapies for AML ...... 15 1.2 Leukaemic stem cells ...... 15 1.2.1 Haematopoietic stem cells (HSCs) and Normal haematopoiesis ...... 16 1.2.2 Regulation of HSCs ...... 20 1.2.2.1 Notch signalling ...... 20 1.2.2.2 Hedgehog signalling ...... 20 1.2.2.3 Wnt signalling ...... 21 1.2.2.4 Other signalling pathways ...... 22 1.2.3 Leukaemic haematopoiesis ...... 23 1.2.4 Heterogeneity in LSCs ...... 27 1.2.4.1 Phenotypic heterogeneity of LSCs ...... 27 1.2.4.2 Molecular heterogeneity of LSCs in AML ...... 28 1.2.5 Regulation of LSCs ...... 29 1.2.5.1 Notch signalling ...... 30 1.2.5.2 Hedgehog signalling ...... 30 1.2.5.3 Wnt/β-catenin signalling ...... 31 1.2.6 Targeting LSCs ...... 33 1.3 G protein-coupled receptors ...... 36 1.3.1 G proteins ...... 39 1.3.2 Regulator of G protein signalling (Rgs) family ...... 40 1.3.3 Leucine-rich repeat-containing G-protein coupled receptor 4 (Lgr4) ...... 40

1.4 Thesis hypothesis and aims ...... 43 2 Materials and Methods ...... 45 2.1 Materials ...... 46 2.2 Methods ...... 57 2.2.1 Maintenance of cells ...... 57 2.2.1.1 Cell culture ...... 57 2.2.1.2 Generating Wnt3a conditioned medium ...... 57 2.2.1.3 Freezing and thawing cells ...... 58 2.2.2 Stable cell line production ...... 59 2.2.2.1 Bacterial transformation and plasmid preparation ...... 59 2.2.2.2 Transient virus production in 293T cells ...... 59 2.2.2.3 Viral transduction ...... 60 2.2.3 Flow cytometric analysis ...... 61 2.2.3.1 Flow cytometry and FACS analysis ...... 61 2.2.3.2 Haematopoietic cell isolation by FACS ...... 61 2.2.4 Gene expression analysis ...... 61 2.2.4.1 RNA isolation ...... 61 2.2.4.2 cDNA preparation ...... 62 2.2.4.3 Quantitative real-time PCR ...... 62 2.2.4.4 Gene expression microarray ...... 63 2.2.5 Protein analysis ...... 64 2.2.5.1 Isolation of total cellular protein ...... 64 2.2.5.2 Protein gel electrophoresis ...... 64 2.2.5.3 Western blotting ...... 65 2.2.6 Functionality assays ...... 65 2.2.6.1 Apoptosis assay ...... 65 2.2.6.2 Cytospin and Wright Giemsa Staining ...... 66 2.2.6.3 Colony-forming assay ...... 66 2.2.6.4 Adenosine-5’-triphosphate (ATP) assay ...... 67

2.2.6.5 Oxygen consumption assay ...... 68 2.2.6.6 Detection of mitochondrial reactive oxygen species ...... 68 2.2.7 In vivo studies ...... 69 2.2.7.1 Animals ...... 69 2.2.7.2 In vivo transplantation ...... 69 2.2.7.3 Isolation of bone marrow and spleen cells ...... 69 2.2.7.4 In vivo BrdU staining ...... 70 2.2.8 Statistical analysis ...... 70 3 The role of Lgr4 in AML development ...... 72 3.1 Introduction ...... 73 3.2 Lgr4 is significantly overexpressed in human MLL-AML patients ...... 74 3.3 High Lgr4 expression is associated with poor patient survival ...... 76 3.4 Lgr4 is significantly upregulated in murine MLL-AML LSCs compared to normal HSCs ...... 79 3.5 Ligand Rspo or Wnt3a alone is not capable of enhancing β-catenin expression in MLL-AML ...... 80 3.6 Rspo2 and Rspo3 synergise with Wnt3a to enhance β-catenin activation in pre- LSCs 83 3.7 Co-treatment of Wnt3a and Rspo3 enhances the proliferation ability of KLSMLL pre-LSCs ...... 85 3.8 Lgr4 knockdown in KLSMLL pre-LSCs suppresses β-catenin expression .... 87 3.9 Lgr4 knockdown impairs cell proliferation of KLSMLL pre-LSCs in vitro ... 89 3.10 Ectopic overexpression of β-catenin rescues the Lgr4-deficient phenotype 91 3.11 Lgr4 knockdown in KLSMLL pre-LSCs impairs AML initiation and maintenance in vivo ...... 93 3.12 Lgr4 knockdown in GMPMLL pre-LSCs suppresses β-catenin expression and impairs cell proliferation in vitro ...... 97 3.13 Lgr4 knockdown in GMPMLL pre-LSCs impairs AML initiation and maintenance ...... 100 3.14 Lgr4 overexpression enhances the proliferative potential of KLSA9M pre- LSC in vitro ...... 103

vii

3.15 Lgr4 overexpression in KLSA9M pre-LSCs exacerbates AML in vivo .... 106 3.16 Lgr4 overexpression cannot transform KLS cells ...... 112 3.17 Lgr4 overexpression enhances cell proliferation of GMPA9M in vitro .... 115 3.18 Summary ...... 118 4 Characterising crucial components of the Lgr4 signalling cascade ...... 119 4.1 Introduction ...... 120 4.2 Identification of downstream Lgr4 signalling components by gene expression profiling ...... 120 4.3 Validation of downstream target genes identified by microarray ...... 123 4.4 Gene set enrichment analysis and its validation ...... 126 4.5 Rgs1 is a major downstream target of Lgr4 ...... 129 4.6 Gαq is a downstream target of Rgs1 and its inhibition impairs cell proliferation of MLL-AML in vitro ...... 131 4.7 Gαq inhibition enhances differentiation, apoptosis, and suppresses β-catenin expression of MLL-AML ...... 134 4.8 shRNA-mediated knockdown of Gαq in pre-LSCs suppresses β-catenin expression and impairs cell proliferation in vitro ...... 137 4.9 Inhibition of Gαq in KLSMLL pre-LSCs reverses the effect of Wnt3a/Rspo3 on cell proliferation and β-catenin activation ...... 143 4.10 shRNA-mediated knockdown of Lgr4 or Gαq abolishes the effect of Wnt3a/Rspo3 on β-catenin activation ...... 146 4.11 Summary ...... 148 5 Lgr4/β-catenin signalling pathway regulates AML LSC activity by altering mitochondrial function ...... 149 5.1 Introduction ...... 150 5.2 Identifying common downstream target genes of Lgr4 and Gαq ...... 150 5.3 Gadd45a knockout enhances cell proliferation of MLL-AF9 transformed c- Kit+ bone marrow cells in vitro ...... 152 5.4 Gadd45a knockout enhances MLL leukaemogenesis in vivo ...... 155 5.5 Mitochondrial genes are regulated by both Lgr4 and Gαq ...... 157 5.6 Lgr4 signalling regulates mitochondrial OXPHOS in pre-LSCs ...... 159

viii

5.7 Lgr4 signalling regulates pre-LSC ATP production ...... 162 5.8 Lgr4 signalling regulates ROS production of pre-LSCs ...... 163 5.9 ROS production induced by inhibiting mitochondrial OXPHOS complexes in pre-LSCs ...... 166 5.10 Lgr4 protects pre-LSCs from ROS induced damage ...... 169 5.11 Summary ...... 174 6 Discussion and future directions ...... 175 6.1 The interplay between Wnt3a, Rspo2/3 and Lgr4 ...... 176 6.2 Lgr4 is crucial for β-catenin activation in LSCs and AML disease reconstitution ...... 178 6.3 Gαq is the G protein messenger relaying signals between Lgr4 and β-catenin 181 6.4 Tumour suppressor Gadd45a affects LSC development ...... 183 6.5 Lgr4 as a master regulator of mitochondrial energy generation and ROS production in pre-LSCs ...... 185 6.6 Conclusions and future perspectives ...... 187 7 References ...... 193

List of Figures

Figure 1.1 Age-standardised mortality rate of AML in Australia...... 3 Figure 1.2 Molecular heterogeneity of AML...... 5 Figure 1.3 Major MLL fusion partners in paediatric and adult AML...... 11 Figure 1.4 Schematic diagram of haematopoietic hierarchy...... 18 Figure 1.5 Schematic diagram of normal and leukaemic haematopoiesis...... 25 Figure 1.6 Schematic diagram of conventional chemotherapy versus CSC targeted therapy...... 34 Figure 3.1 Lgr4 and β-catenin are upregulated in human MLL-AML patients...... 72 Figure 3.2 High Lgr4 expression is associated with poor clinical outcome in AML patients...... 75 Figure 3.3 Lgr4 and β-catenin are upregulated in mouse MLL-AML...... 77 Figure 3.4 Rspo1-4 or Wnt3 alone cannot activate β-catenin in KLSMLL pre-LSCs....79 Figure 3.5 Rspo2 and Rspo3 synergise with Wnt3a to activate β-catenin in KLSMLL pre-LSCs...... 82 Figure 3.6 Co-treatment of Wnt3a and Rspo3 enhances colony formation of KLSMLL pre-LSCs...... 84 Figure 3.7 Lgr4 knockdown in KLSMLL pre-LSCs suppresses β-catenin expression...86 Figure 3.8 Lgr4 knockdown suppresses colony formation of KLSMLL pre-LSCs...... 88 Figure 3.9 Ectopic overexpression of β-catenin rescues the Lgr4-deficient phenotype..90 Figure 3.10 Lgr4 knockdown impairs KLSMLL induced AML initiation and maintenance...... 93 Figure 3.11 Knockdown of Lgr4 in GMPMLL pre-LSCs suppresses β-catenin expression and colony formation...... 96 Figure 3.12 Lgr4 knockdown impairs GMPMLL induced AML initiation and maintenance...... 99 Figure 3.13 Lgr4 overexpression enhances proliferative potential of KLSA9M pre-LSC in vitro...... 102

Figure 3.14 Lgr4 overexpression in KLSA9M pre-LSCs exacerbates AML in vivo....105 Figure 3.15 Overexpression of Lgr4 enhances in vivo proliferation of KLSA9M pre- LSCs...... 107 Figure 3.16 Overexpression of Lgr4 enhances in vivo proliferation of KLSA9M LSCs...... 108 Figure 3.17 Lgr4 alone cannot transform KLS cells...... 111 Figure 3.18 Lgr4 overexpression enhances proliferation of GMPA9M in vitro...... 114 Figure 4.1 Identification of downstream Lgr4 signalling components by gene expression profiling...... 119 Figure 4.2 Validation of microarray data using qRT-PCR...... 122 Figure 4.3 Lgr4 mediated enrichment of Wnt/β-catenin gene signature...... 125 Figure 4.4 Validation of Rgs1 protein expressions...... 128 Figure 4.5 Gαq inhibitor but not Gαi inhibitor impairs colony formation of KLSMLL pre-LSCs...... 130 Figure 4.6 Inhibition of Gαq in KLSMLL pre-LSCs enhances differentiation and apoptosis, and suppresses β-catenin expression...... 133 Figure 4.7 Knockdown of Gαq in KLSMLL pre-LSCs suppresses β-catenin expression and impairs colony formation...... 137 Figure 4.8 Knockdown of Gαq in GMPMLL pre-LSCs suppresses β-catenin expression and impairs colony formation...... 139 Figure 4.9 Inhibition of Gαq reverses the phenotypes caused by co-incubation of Wnt3a and Rspo3...... 142 Figure 4.10 Knockdown of Lgr4 or Gαq abolishes the effect of Wnt3a/Rspo3 on β- catenin activation...... 145 Figure 5.1 Common target genes shared by both Lgr4 and Gαq...... 149 Figure 5.2 Gadd45a knockout enhances colony forming ability of MLL-AF9 transduced c-Kit+ bone marrow cells...... 151 Figure 5.3 Gadd45a knockout enhances MLL leukaemogenesis in vivo...... 154 Figure 5.4 Lgr4 upregulates mitochondrial DNAs...... 156 Figure 5.5 Lgr4 signalling regulates mitochondrial OXPHOS in pre-LSCs...... 158

Figure 5.6 Lgr4 or Gαq knockdown impairs mitochondrial ATP production...... 160 Figure 5.7 Lgr4 regulates mitochondrial ROS production...... 162 Figure 5.8 Gadd45a regulates mitochondrial ROS production...... 163 Figure 5.9 Antimycin A and rotenone increase ROS production of KLSA9M pre- LSCs...... 165 Figure 5.10 Lgr4 protects KLSA9M pre-LSCs from antimycin A induced ROS damage...... 168 Figure 5.11 Lgr4 protects KLSA9M pre-LSCs from rotenone induced ROS damage...... 170 Figure 6.1 Schematic diagram of Lgr4 signalling network in LSCs...... 188

xii

List of Tables

Table 1.1 The French-American-British (FAB) classification of AML...... 6 Table 1.2 The World Health Organisation (WHO) classification of AML...... 7 Table 1.3 European Leukaemia Net AML classifications...... 14 Table 1.4 Wnt/β-catenin signalling in CSCs...... 32 Table 1.5 Selected GPCRs that are involved in cancer...... 38 Table 1.6 Lgr4 tissue expression patterns in mice...... 42 Table 2.1 Purchased Reagents...... 45 Table 2.2 Prepared Reagents...... 48 Table 2.3 Cells...... 49 Table 2.4 Vectors...... 50 Table 2.5 FACS antibodies...... 51 Table 2.6 Western blotting primary antibodies...... 52 Table 2.7 Western blotting secondary antibodies...... 53 Table 2.8 Primers...... 54

xiii

Acknowledgements

I would like to take this opportunity to thank my supervisor Dr Jenny Wang for all her support and guidance over the past four years. Jenny has always been there to help no matter how small the matter was. I would not be here without her support.

I would also like to thank my joint supervisor Professor Maria Kavallaris. The lung cancer project not working out was rather unfortunate but Maria has given me valuable advice on helping me finish my PhD smoothly. I would like to thank my panel review members Professor Murray Norris and Dr Michelle Henderson for their constructive feedback in my annual review and presentations. I would like to thank Dr Amanda Philp for her advice and making this PhD journey a smooth one.

Furthermore, I would like to express my gratitude to everyone in the Cancer Stem Cell and Biology group at the Children’s Cancer Institute Australia. Thank you Dr Jennifer Lynch and Dr Katerina Bendak for meticulously looking over my thesis and all the help you’ve given me. Thank you Halina Leung, Estrella Gonzales and Dr Philipp Dietrich. We all started the journey at the same time and shared so many moments, you have made my PhD a much more enjoyable and memorable one. Thank you Ashley Yang, Kathryn Mathews, Benson Ton, Kimberley Anderson, Brendon Martinez, Hannah McCalmont and Florida Voli for all the help, advice and moments you’ve shared with me, I wish you all the very best.

Thanks to the School of Women’s and Children’s Health at UNSW and the Children’s Cancer Institute Australia for giving me the opportunity to conduct research at the highest calibre. I want to thank the Australian Postgraduate Award and the CCIA Postgraduate Top up scheme for funding my scholarships.

xiv

I would like to thank all my friends and family, especially my mother for her love and support. I would not be in this beautiful country without you. Finally, I would like to thank my beautiful wife Crystal. Thank you for being the most loving and understanding person, countless late nights and weekends, joys and sorrows, sweets and bitters, you have always being there with me.

Abbreviations

°C Degrees Celsius 7-AAD 7-Aminoactinomycin D A9M Hoxa9 and Meis1a Akt Protein binase B ALL Acute lymphoblastic leukaemia AML Acute myeloid leukaemia APC Allophycocyanin, flow cytometry; adenomatous polyposis coli, Wnt signalling APML Acute promyelocytic leukaemia ATP Adenosine-5’-triphosphate Bcl2 B-cell lymphoma 2 Bcl-Abl Break cluster region-abelson murine leukaemia viral oncogene homolog 1 BRCA1 Breast cancer type 1 susceptibility protein BrdU Bromodeoxyuridine CBF Core-binding factor CBP Cyclic AMP response element-binding protein CD Cluster of differentiation CEBPA CCAAT/enhancer binding protein CK1α Casein kinase 1α CLL Chronic lymphocytic leukaemia CLP Common lymphoid progenitor CML Chronic myeloid leukaemia CMP Common myeloid leukaemia CSC Cancer stem cell Cox Cyclooxygenase Csnk1e Casein kinase 1 epsilon

xvi

Cytb Cytochrome b reductase Dkk Dickkopf DMEM Dulbecco’s Modified Eagle’s media DMSO Dimethyl sulphoxide DNA Deoxyribonucleic acid dNTP Deoxynucleoside triphosphate Dsh Dishevelled Dsh-PDZ Dsh PDZ domain DTT Dithiothreitol ENL Eleven nineteen leukaemia ETO Eight twenty one EV Empty vector EVI1 Ecotropic virus integration site 1 FAB French-American-British FACS Fluorescence activated cell sorting Fc Fragment crystallisable FCS Fetal calf serum FLT3 Fms-like tyrosine kinase 3 FoxO Forkhead box O FSB First strand buffer Fzd Frizzled g Gram(s), mass; gravitational force, centrifugation steps Gadd45 Growth arrest and DNA damage-inducible 45 proteins GDP Guanosine diphosphate GLI GLI family zinc finger proteins GM-CSF Granulocyte macrophage colony-stimulating factor GMP Granulocyte macrophage progenitor GP GP-antagonist 2A GPCR G protein-coupled receptor GSEA Gene set enrichment analysis

xvii

GSK3 Glycogen synthase kinase-3-β GTP Guanosine triphosphate h Hour(s) HEK293 Human embryonic kidney 293 HEPES 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid HSC Haematopoietic stem cell Homeotic Hox IFN Interferon IL Interleukin IL-3Rα Interleukin 3 receptor alpha IMDM Iscove's Modified Dulbecco's Medium Irf2 Interferon regulatory factor 2 ITD Internal tandem duplication IWP Inhibitor of Wnt production IWR Inhibitor of Wnt response KLS c-Kit+Lin-Sca-1+ KOH Potassium hydroxide L Liter(s) LB Luria broth Lcn2 Lipocalin 2 LEF Lymphoid enhancer-binding factor Lgr Leucine repeat-containing G-protein coupled receptor Lin Haematopoietic lineage markers LRP Low density lipoprotein receptor-related protein LSC Leukaemia stem cell LT-HSC Long-term HSC m Milli when in front of a unit; meter when used as a unit M-MLV RT Moloney Murine Leukaemia Virus Reverse Transcriptase MEP Megakaryocyte/erythrocyte progenitors min Minute(s)

xviii

MLL Mixed lineage leukaemia MLL-AF9 MLL acute lymphoblastic leukaemia 1 fused gene from chromosome 9 MMP-9 Matrix metallopeptidase 9 MPP Multipotent progenitor mtDNA Mitochondrial DNA ND2 NADH dehydrogenase 2 ND4l NADH dehydrogenase 4l NFκB Nuclear factor-kappa B NK Natural killer NOD Non-obese diabetic NPM1 Nucleophosmin 1 OCR Oxygen consumption rate OXPHOS Oxidative phosphorylation PAR1 Proteinase-activated receptor 1 PBS Phosphate buffered saline PCA Perchloric acid PGE1 Prostaglandin E receptor 1 (ptger1) PI Propidium iodide PI3K Phosphoinositide 3-kinase PML Promyelocytic leukaemia PTCH1 Patched homolog 1 PTEN Phosphatase and tensin homolog PTX Pertussis toxin PVDF Polyvinylidene difluoride qRT-PCR Quantitative real-time polymerase chain reaction RARA Retinoic acid receptor alpha Rgs Regulator of G protein signalling Rnf43 Ring finger 43 ROS Reactive oxygen species Rspo Roof plate-specific spondin

xix

RUNX1 Runt-related transcription factor 1 Sca-1 Stem cell antigen-1 SCF Stem cell factor Scr Scrambled SCID Severe combined immunodeficient SDS Sodium dodecyl sulphate SEM Standard error of the mean SMO Smoothened Shh Sonic hedgehog shRNA Short hairpin RNA SL-IC SCID leukaemia-initiating cell Smo Smoothened TANK Tankyrase TBS Tris buffered saline TCF T cell-specific transcription factor Thy-1 Thymocyte differentiation antigen 1 TNF-α Tumour necrosis factor alpha Txndc15 Thioredoxin domain-containing protein 15 WBC White blood cell Wnt Wingless-related integration site WT Wild-type Znrf3 Zinc and ring finger 3 μ Micro

1 Introduction

1.1 Acute myeloid leukaemia Leukaemia is a cancer of the white blood cells and are characterised by high numbers of abnormal white blood cells or blast cells. It can be classified into four main types: acute myeloid leukaemia (AML), acute lymphoblastic leukaemia (ALL), chronic lymphocytic leukaemia (CLL) and chronic myeloid leukaemia (CML). Leukaemia is the 11th most common cancer worldwide; however, it is the most common type of cancer in children. In 2012, more than 352,000 new cases were diagnosed globally while around 265,000 deaths were reported at the same time [1]. Australia has one of the highest leukaemia incidence rates in the world. It is estimated that more than 3200 people will be diagnosed with leukaemia this year. Although there was a steady increase in the five- year survival rate for leukaemia in Australia over the past few decades (from 37% in 1982 to 49% in 2004), variations in survival among different types of leukaemia were large [2]. For example, the overall five-year survival rate for CLL was relatively high (about 73%); in contrast, AML has an overall five-year survival rate of only 24% [2].

AML is an aggressive acute leukaemia characterised by the rapid growth of abnormal immature myeloid cells in the bone marrow and peripheral blood. The accumulation of leukaemic cells causes a decrease in normal while blood cells, red blood cells and platelets, which lead to infection and bleeding in addition to organ infiltration. Treatment for AML depends on a number of factors including subtype, stage of the disease and severity of symptoms. Chemotherapy, radiotherapy and stem cell transplantation remain the most effective treatments to date. However, these therapies frequently fail to achieve long term remission and the majority of AML patients experience relapse in which the AML cells become drug-resistant [3]. In Australia, there were 812 reported deaths from AML in 2011, accounting for 1.9 percent of all cancer death. The incidence of AML has remained stable since 1982 (3.2 cases per 100,000 people in 1982 vs. 3.7 cases per 100,000 people in 2010). However, the mortality rate has not improved over the past few decades (Figure 1.1). Therefore, there is urgent need for new therapies to treat AML.

Figure 1.1 Age-standardised mortality rate of AML in Australia

Age-standardised mortality rate of AML in Australia by gender from 1968 to 2012. Data retrieved from Australian Cancer Incidence and Mortality (ACIM) books: Acute myeloid leukaemia. Australian Institute of Health and Welfare.

1.1.1 Heterogeneity of AML AML is a heterogeneous disease caused by a range of pathogenic mechanisms. This can be manifested by the morphology of the cells and the degree of differentiation. A widely used classification system is the French-American-British (FAB) system which divides AML into nine subtypes ([4]; M0 – M7, Table 1.1). They are classified based on the type of cell from which the leukaemia was developed and the degree of differentiation. This is achieved by morphological examination and histochemical staining. In addition to the phenotypic variations, the heterogeneity of AML is also manifested by the molecular heterogeneity associated with the disease. To date, over 200 chromosomal aberrations and genetic mutations were identified, of which the most common aberrations include Fms-like tyrosine kinase 3 internal tandem duplication (FLT3 ITD), promyelocytic leukaemia/retinoic acid receptor-α (PML/RARA) and nucleophosmin 1 (NPM1) ([5,6], Figure 1.2). Due to the phenotypic and molecular heterogeneity of AML, the World Health Organization (WHO) has developed a more comprehensive and clinically useful classification of AML, incorporating different genetic aberrations ([7], Table 1.2). The clinical outcome and genetics of AML are tightly linked such that cytogenetic and molecular genetic analyses such as chromosomal translocations have been routinely performed upon diagnosis in order to tailor treatment regimens among patients with different subtypes [8-11]. Tailoring treatment regimens according to cytogenetic differences has been reported to improve clinical outcome significantly [12-14], highlighting the importance of defining genetic subtypes of AML.

Figure 1.2 Molecular heterogeneity of AML

Pie chart illustrating the relative frequencies of common recurrent genetic abnormalities in AML. The most common mutations include FLT3 ITD, PML/RARA and NPM1. Adapted from [15].

Table 1.1 The French-American-British (FAB) classification of AML

Subtype Name M0 Acute myeloblastic leukaemia, minimally differentiated M1 Acute myeloblastic leukaemia, without maturation M2 Acute myeloblastic leukaemia, with granulocytic maturation M3 Promyelocytic, or acute promyelocytic leukaemia M4 Acute myelomonocytic leukaemia M4eo Myelomonocytic together with bone marrow eosinophilia M5 Acute monoblastic leukaemia (M5a) or acute monocytic leukaemia (M5b) M6 Acute erythroid leukaemia M7 Acute megakaryoblastic leukaemia Table adapted from [4]

Table 1.2 The World Health Organisation (WHO) classification of AML

AML with recurrent genetic abnormalities AML with t(8;21)(q22;q22) RUNX1-RUNX1T1 (CBFA-ETO) AML with inv(16)(p13q22) or t(16;16)(p13;q22) CBFB-MYH11 APL with t(15;17)(q22;q11–12) PML-RARA AML with t(9;11)(p22;q23) MLLT3-MLL and other balanced translocations of 11q23 (MLL) AML with t(6;9)(p23;q34) DEK-NUP214 AML with inv(3)(q21q26.2) or t(3;3)(q21;q26.2) RPN1-EVI1 AML (megakaryoblastic) with t(1;22)(p13;q13) RBM15-MKL1 AML with mutated NPM1 AML with mutated CEBPA Acute myeloid leukaemia with myelodysplasia-related changes Therapy-related myeloid neoplasms AML, not otherwise specified AML with minimal differentiation AML without maturation AML with maturation Acute myelomonocytic leukaemia Acute monoblastic/monocytic leukaemia Acute erythroid leukaemia Acute megakaryoblastic leukaemia Acute basophilic leukaemia Acute panmyelosis with myelofibrosis Myeloid sarcoma Myeloid proliferations related to Down syndrome (+21) Transient abnormal myelopoiesis Myeloid leukaemia associated with Down syndrome Blastic plasmacytoid dendritic cell neoplasms Table adapted from [15]

1.1.1.1 Favourable genetic features Acute promyelocytic leukaemia (APML) is a subtype of AML characterised by translocation of chromosome 15 and 17, which in the majority of cases create a fusion oncoprotein Promyelocytic Leukaemia/Retinoic Acid Receptor Alpha (PML/RARA). It has a favourable prognosis with patients frequently achieving long-term remission. This high overall survival rate is due primarily to its responsiveness to arsenic trioxide and all-trans retinoic acid, which target PML/RARA [16]. Therefore, upon diagnosis of APML, it is crucial to confirm or exclude the PML/RARA rearrangement in order to provide patients with the appropriate treatment.

AML with translocations involving the core-binding factor (CBF) is another subtype of leukaemia with a favourable prognosis. Unlike APML, there is currently no specific targeted therapy for CBF-rearranged AML, yet patients respond well to current standard therapy. However, CBF-rearranged AMLs with a mutated KIT gene exhibit a higher relapse rate and decreased overall survival [17].

The Nucleophosmin 1 (NPM1) gene is one of the most commonly mutated genes in AML. It encodes a nucleolar phosphoprotein and the mutation causes aberrant activation which inactivates the tumour suppressor p53/ARF pathway, resulting in tumour growth [18]. The majority of patients with NPM1 mutation have normal karyotype and it does not occur together with APML or CBF-rearrangement [19]. Nevertheless, emerging data suggest that NPM1 mutation cooperates with other mutations to influence the outcome of treatment. For example, patients with NPM1 mutations but lack FLT3 mutations have a favourable prognosis that is similar to CBF- rearranged AML, while concurrent FLT3 mutations result in poor prognosis [20].

AML with mutations in the CCAAT/enhancer-binding protein alpha gene (CEBPA) accounts for about 5-10% of all AML cases. They are mostly found in patients with normal karyotype and are associated with a favourable prognosis [21]. Most patients with CEBPA mutations are mutually exclusive with NPM1 mutations. Concurrent FLT3 mutations are rare therefore it is unclear whether FLT3 mutations affect patient

prognosis [22]. The leukaemic blast cells of CEBPA mutated AML have a distinct immunophenotype [high CD7 (Cluster of differentiation 7) and CD34 expression] and gene expression profile [23].

1.1.1.2 Unfavourable genetic features FLT3 is a receptor tyrosine kinase found on the surface of many haematopoietic cells including HSCs. Its signalling was found to be important for the normal development and survival of HSCs [24]. The expression of FLT3 is normally lost upon differentiation whereas AML cells retain high expression of FLT3 for their survival and proliferation [25-27]. FLT3 mutations are the most common genetic abnormality in AML, among which the FLT3 ITD is the most common mutation type, occurring in 15-35% of AML patients [28]. FLT3 ITD has been shown to promote haematopoietic cell proliferation and block myeloid differentiation [29,30]. FLT3 mutations in general have been associated with poor clinical outcome while the presence of FLT3 ITD correlates with an even higher risk of relapse and lower overall survival [31,32].

The ecotropic virus integration site 1 (EVI1) is a nuclear transcription factor involved in embryogenesis, cell cycle and differentiation [33]. It was also found to regulate the development and proliferation of HSCs [34]. Recent genome wide analysis revealed that in AML cells EVI1 bind directly to genes responsible for terminal myeloid differentiation and numerous genes involved in cell cycle and apoptosis [35]. In about 10% of AML cases EVI1 was found to be aberrantly upregulated, and patients had low complete remission rate and very poor prognosis [36,37].

The Runt-related transcription factor 1 (RUNX1) gene, also known as AML1, is crucial for the development of haematopoietic stem cells (HSCs) in the embryo [38]. It also plays a crucial role in haematopoiesis by regulating various haematopoietic genes including granulocyte macrophage colony-stimulating factor (GM-CSF), interleukin-3 (IL-3), FLT3 and B-cell lymphoma 2 (Bcl2) [39,40]. RUNX1 mutations were mainly found in M0, M1 and M3 subtypes of AML, and were frequently found together with

cytogenetic aberrations, the most frequent being trisomy 13 [41]. AML patients with RUNX1 mutations had much shorter overall and event-free survival compared to RUNX1 wild-type, and multivariate analysis revealed that RUNX1 is a strong adverse prognostic factor independent of other prognostically relevant factors for AML survival [41].

1.1.1.3 MLL-rearranged AML The Mixed Lineage Leukaemia (MLL) gene located on chromosome 11, band q23, is one of the most common translocated genes in hematopoietic malignancies. MLL translocation is the most common chromosomal aberration found in paediatric AML with over 70% of paediatric AML patients retaining this translocation. It also occurs frequently in adult AML with a frequency of approximately 10% [42]. MLL encodes a histone methyltransferase that is essential in regulating gene expression during early development and haematopoiesis by chromatin modifications. However, MLL rearrangement disrupts its normal methyltransferase activity and results in aberrant expression of oncogenes such as the homeotic (Hox) genes [42,43].

The most characteristic feature of MLL-rearranged AML is its heterogeneity. More than 50 different MLL fusion partners have been identified. Nevertheless, only a small subset accounts for the majority of AML cases. The most frequent MLL translocation is t(9;11)(p22;q23) or MLL-AF9, accounting for more than one third of all MLL- rearranged AML; other frequently observed MLL translocations include t(10;11)(p12;q23) or MLL-AF10, t(6;11)(q27;q23) or MLL-AF6, t(11;19)(q23;p13.1) or MLL-ELL and t(11;19)(q23;p13.3) or MLL-ENL ([44-46]; Figure 1.3). Patients with MLL translocations generally have a poor clinical outcome although prognosis varies across different studies and depends on which translocation partners are involved. For example, AML with MLL-AF9 translocation has been reported to have a 5-year overall survival of 63% [47] while another international study reported a 5-year overall survival of 40% [48]. Although optimized intensive treatment regimens have improved the

clinical outcome for patients with MLL translocation, 30% to 50% of patients still die from the disease within 5 years [49,50].

Figure 1.3 Major MLL fusion partners in paediatric and adult AML

Pie charts illustrating the percentages of the most frequent MLL fusion partners in paediatric AML and adult AML. The most common MLL fusion partners in paediatric AML are AF9 (33%), AF10 (17%), AF6 (14%), ELL (11%) and EPS15 (3%). The most common MLL fusion partners in adult AML are AF9 (32%), ENL (14%), ELL (9%), AF10 (5%) and AF6 (5%). Adapted from [49].

1.1.2 Treatment of AML

1.1.2.1 Current treatment regimens The current standard of care for the treatment of AML includes induction chemotherapy, followed by post-remission therapy which includes consolidation chemotherapy, autologous stem cell transplantation or allogeneic stem cell transplantation. Initial treatment decisions are based on clinical presentations and typically consist of 7 days of continuous intravenous infusion of cytarabine and 3 days of anthracycline, which was developed four decades ago [51]. Modifications are based on the dose of the two drugs and the type of anthracycline according to clinical features of each individual patient. However, the efficacy of this treatment is far from optimal and decades of clinical trials have failed to identify a superior treatment combination [3].

Following intensive induction chemotherapy, usually 40% - 75% of patients will achieve a first complete remission. However, virtually all patients will relapse without additional treatment [52]. Post-remission therapy aims to eliminate undetectable leukaemic cells that survived induction chemotherapy and hence is essential to prevent relapse. Resistance to chemotherapy is largely dependent on molecular genetic and cytogenetic aberrations and The European Leukaemia Net has stratified patients into four risk groups to guide treatment selections ([53]; Table 1.3). Patients in the favourable genetic group are predicted to have a much higher chance of maintaining long-term remission after intensive consolidation chemotherapy. In contrast, patients in the adverse group are rarely cured with intensive consolidation chemotherapy and thus should be offered stem cell transplantation and/or enrolled in clinical trials [54]. However, cure rate using conventional therapies are now reaching a plateau and better therapies are urgently needed.

Table 1.3 European Leukaemia Net AML classifications

Genetic Group Subsets Favourable t(8;21)(q22;q22); RUNX1-RUNX1T1 inv(16)(p13.1q22) or t(16;16)(p13.1;q22); CBFB-MYH11 Mutated NPM1 without FLT3-ITD (normal karyotype) Mutated CEBPA (normal karyotype) Intermediate I Mutated NPM1 and FLT3-ITD (normal karyotype) Wild-type NPM1 and FLT3-ITD (normal karyotype) Wild-type NPM1 without FLT3-ITD (normal karyotype) Intermediate II t(9;11)(p22;q23); MLLT3-MLL Cytogenetic abnormalities other than favourable or adverse Adverse inv(3)(q21q26.2) or t(3;3)(q21;q26.2); RPN1-EVI1 t(6;9)(p23;q34); DEK-NUP214 t(v;11)(v;q23); MLL rearranged -5 or del(5q); -7; abnormal(17p); complex karyotype Adapted from [54]

1.1.2.2 Targeted therapies for AML The identification of specific molecular aberrations in cancer has led the way for the development of targeted therapies with better efficacy and fewer side effects. The most prominent example is the identification of Break cluster region-abelson murine leukaemia viral oncogene homolog 1 (Bcl-Abl) kinase inhibitor Imatinib that specifically targets the activated Bcl-Abl kinase in the treatment of CML [55]. In the AML subtype APML, targeted therapies such as arsenic trioxide and all-trans retinoic acid are used to target the PML/RARA fusion oncoprotein that is characteristic of APML [16]. However, the development of targeted therapies in other types of AML has been slow.

FLT3 mutations are the most common genetic aberrations in AML and confer a poor prognosis [28]. FLT3 mutations result in constitutive activation of the surface receptor, therefore, small molecule inhibitor targeting represents an attractive approach to inhibit FLT3 activation. A tremendous effort has been focused on identifying small molecule inhibitors of FLT3 over the past decade. Several drug candidates have demonstrated promising pre-clinical results and are currently in various stages of clinical trials to determine their true therapeutic efficacy [56]. For example, the second generation FLT3 inhibitor quizartinib showed promising results in Phase I study with over 50% response rate in FLT3 ITD AML patients, and the preliminary results from Phase II study showed an overall response rate of 68%, with a number of long-term survivors [56].

1.2 Leukaemic stem cells Leukaemic stem cells (LSCs) are leukaemic cells that possess the characteristics of normal stem cells such as self-renewal and quiescence. AML LSCs were the first cancer stem cells (CSCs) discovered and have been fundamental in driving the field of CSC research forward. The LSC model helps to explain one of the biggest problems in AML: the high relapse rate. More than 50% of patients will relapse after initial treatment, after which the cells become resistant to conventional chemotherapy and the majority of

patients succumb to the disease [57]. It is now believed that LSCs are responsible for such high relapse rate in AML due to the fact that conventional chemotherapies only target rapidly proliferating blast cells whereas LSCs are quiescent most of the time, and are protected by their microenvironment [58,59]. They ultimately repopulate in the bone marrow causing relapse. The characterisation of AML LSCs was made possible due to our understanding of normal haematopoietic stem cells (HSCs) and the techniques used in dissecting HSC biology. A thorough understanding of the properties of normal HSCs is essential for developing LSC targeted therapies with minimal toxicity.

1.2.1 Haematopoietic stem cells (HSCs) and Normal haematopoiesis HSCs are multipotent stem cells which maintain blood formation. They are mainly found in the bone marrow. They can give rise to all blood cell types such as neutrophils and T-cells. The first evidence suggesting the existence of HSCs came from the 1960s when researchers from several different laboratories demonstrated formation of clonogenic mixed colonies in the spleen after transplantation of bone marrow cells into irradiated mice, and these colony-forming cells can occasionally reconstitute the haematopoietic system after secondary transplantation [60-62]. Extensive research on HSCs ensued trying to purify and characterise these cells.

HSCs are functionally defined by their two unique capabilities: self-renewal, which is the ability to give rise to identical daughter cells without differentiation, and multipotency, the ability to differentiate into all types of mature blood cells. One of the first experiments that demonstrated these properties was performed by Becker et al. where they demonstrated that a very small population of bone marrow cells could generate multiple types of myeloerythroid cells and self-replicate [62]. More direct evidence of HSCs was later obtained by tracking progeny in mice [63]. Early attempts to purify HSCs and the multipotent progenitors was carried out using density gradient centrifugation and elutriation according to their physical properties such as cell sizes

[64]. With the advent of fluorescence-activated cell sorting (FACS), rare HSCs can be purified based on expression of a panel of cell surface markers. HSCs in mouse bone marrow can be separated using Thy-1lo, haematopoietic lineage marker-/lo (Lin-/lo), c- Kit+ and Sca-1+ [65]. However, there is no human homolog for Sca-1. Human HSCs are separated based on CD34 expression [66].

Unlike the progenitor cells, HSCs are inactive and remain dormant for a long period of time unless environmental cues such as stress which trigger them to emerge from dormancy and proliferate. Studies have shown that as few as a single HSC is able to reconstitute the haematopoietic system [67]. Adult HSCs only comprise 0.02% of the total bone marrow cells; however, it is the self-renewal ability which equips them with long term reconstitution capacity.

During normal haematopoiesis, HSCs generate both lymphoid cells (T-cells, B-cells, and natural killer cells) and myeloid cells (macrophages, granulocytes, megakaryocytes and erythrocytes). It is a stepwise process which involves multiple intermediate progenitors (Figure 1.4). In 1994, Morrison and Weissman separated distinct subpopulations from Thy-1loLin-/loc-Kit+Sca-1+ bone marrow which were highly enriched for HSCs. They identified long-term HSCs which were capable of long-term haematopoietic reconstitution, and short-term HSC, which could only transiently reconstitute the haematopoietic system [68]. Later, Weissman and colleagues identified multipotent progenitors (MPPs), common lymphoid progenitors (CLP) and common myeloid progenitors (CMPs) [69-71]. CLPs give rise to more mature progenitors which ultimately generate lymphoid cells such as T-cells, B-cells and natural killer (NK) cells. Similarly, CMPs generate all myeloid cells including megakaryocytes and erythrocytes (Figure 1.4).

Figure 1.4 Schematic diagram of haematopoietic hierarchy

Haematopoietic stem cells (HSCs) are functionally defined as either long-term (LT- HSCs) or short-term repopulating stem cells (ST-HSCs) by their ability to produce long-term or transient haematopoietic reconstitution. HSCs give rise to multipotent progenitors (MPPs) which can then differentiate into common lymphoid progenitors (CLPs) and common myeloid progenitors (CMPs). CLPs differentiate into cells of lymphoid lineage including T-cells, B-cells and NK cells. CMPs give rise to either granulocyte macrophage progenitors (GMPs), which differentiate into granulocytes and macrophages, or Megakaryocyte-erythroid progenitors (MEPs), which differentiate into megakaryocytes and erythrocytes. Adapted from [72].

1.2.2 Regulation of HSCs The self-renewal capacity of HSCs underpins their long-term multilineage reconstitution potential. Studies to elucidate the underlying mechanisms were carried out extensively. A number of signalling pathways responsible for their self-renewal capacity have been proposed and demonstrated to regulate HSC self-renewal, although there were some contradictory findings from different laboratories.

1.2.2.1 Notch signalling The Notch signalling pathway is one of the most highly conserved signalling pathways present in all mammals [73]. Early studies have shown that Notch was expressed in human CD34+ haematopoietic cells, which were undifferentiated blood cells containing HSCs as well as progenitor cells [74]. Notch 1 activation inhibited granulocytic differentiation and allowed expansion of undifferentiated cells [74]. Therefore, Notch may serve to retain the self-renewal ability of HSCs. Several years later, it was shown that HSCs can be immortalised by constitutive activation of Notch1 signalling, which enabled them to passage long term in vitro and undergo multilineage reconstitution in vivo [75]. In addition, the Notch ligand Jagged-1 was demonstrated to increase the in vitro expansion capacity as well as in vivo engraftment and survival of HSCs [76]. In contrast, studies using transgenic mice that carry dominant-negative form of Mastermind-like1 (a global inhibitor of Notch signalling) or lack CSL/RBPJ (transcription factors required for Notch signalling) showed that the frequency of long- term HSCs were not affected, indicating that the maintenance of adult HSCs is independent of Notch signalling [77].

1.2.2.2 Hedgehog signalling The Hedgehog signalling pathway is a highly conserved pathway that plays a vital role in animal development [78]. However, its role in haematopoiesis, especially in regulating HSCs is not entirely clear. An initial study by Bhardwaj et al. showed that

human HSC enriched CD34+CD38-Lin- cells express Protein Patched Homolog 1 (PTCH1) and Smoothened (SMO), which are the two cell surface receptors of the Hedgehog signalling axis, as well as some of their downstream targets such as GLI1, GLI2 and GLI3 [79]. In addition, antibodies against Hedgehog inhibited the proliferation of these cells in vitro. HSCs treated with the Sonic hedgehog (Shh) ligand in culture for one week was able to engraft in Non-obese diabetic/Severe combined immunodeficient (NOD/SCID) mice, whereas control treated HSCs could not engraft, suggesting these cells underwent differentiation without Shh signalling [79]. Furthermore, Zhao et al. observed that in SMO deficient mice, although the frequency of HSCs was unchanged, there was a clear defect in HSC function during primary and secondary transplantation [80]. In contrast, a study by Dierks et al. examined the role of SMO by using HSCs isolated from SMOnull mice and found no difference in engraftment compared to normal HSCs [81]. Two additional studies observed similar findings that the loss of SMO had no significant effect on HSC functions such as homing or engraftment, and inhibition of SMO by a small molecule inhibitor had no effect on haematopoiesis [82,83].

1.2.2.3 Wnt signalling The Wnt signalling pathway is another extensively studied pathway that is involved in HSC functioning. In particular, the canonical Wnt pathway (or Wnt/β-catenin pathway) has been shown to regulate HSC self-renewal [84]. Overexpression of β-catenin in HSCs increased cell growth and decreased cell differentiation for many weeks of in vitro cell culture, whereas ectopic expression of Axin (a component of the β-catenin destruction complex) led to impaired HSC proliferation both in vitro and in vivo [84]. Purified Wnt3a (a canonical Wnt ligand) was shown to maintain HSC self-renewal, reduce their differentiation in vitro, and increase in vivo reconstitution [85]. A study using Wnt3a knockout embryos have found that Wnt3a deficiency resulted in a decrease in HSC number in the foetal liver, a major haematopoietic organ in the developing embryo, and severe reduction of in vivo reconstitution capacity of HSCs [86]. Another

research group using loxP-β-catenin and Vav-cre transgenic mice demonstrated that the deletion of β-catenin did not cause impairment of initial HSC establishment, but resulted in reduction of long-term maintenance [87]. In contrast, a study done by Cobas et al. using floxed β-catenin mice with IFN-inducible Mx promoter showed that the β- catenin deficient HSCs were able to generate all blood lineages, and the development of T and B cells were normal in the absence of β-catenin [88]. One possibility could be that the absence of β-catenin was compensated by its close relative γ-catenin, whose overexpression could enhance the self-renewal of HSCs [89]. However, a study by Jeannet et al. has shown that HSCs with combined absence of β-catenin and γ-catenin could still maintain their long-term reconstitution capacity [90].

The conflicting findings may be due to different experimental designs and approaches such as the time of which β-catenin is deleted and how it is deleted. It may also be due to functional redundancies of unknown proteins or other developmental pathways such as the Notch and Hedgehog pathway described above. An increasing amount of studies focused on the native microenvironment of HSCs since the HSC niche has been shown to be a key factor in regulating the fate of HSCs. It has been shown that the HSC niche provided HSCs with signals such as Wnt, Shh and Notch ligands [91]. Therefore it is likely that cross talk exists among these signalling pathways, and studies are being conducted to further dissect their distinct roles in HSC self-renewal and differentiation.

1.2.2.4 Other signalling pathways Apart from the three major developmental signalling pathways described above, other pathways have also been implicated in the regulation of HSCs. The Janus kinase/Signal Transducer and Activator of Transcription (JAK/STAT) pathway, which can be activated by stress signals such as inflammation, has been shown to be involved in regulating HSCs [92]. Essers et al. demonstrated that interferon alpha, one of the cytokines that activate JAK/STAT, could stimulate dormant HSCs to exit the G0 phase and proliferate in vivo [92]. Another group observed that type 1 interferons caused

proliferation and exhaustion of HSCs, while Interferon regulatory factor 2 (Irf2), a negative regulator of interferon, is crucial for maintaining the stem cell properties of HSCs [93]. They found that the number of total HSCs was significantly less in Irf2-/- mice compared to Irf2+/- mice, and the Irf2-/- HSCs have impaired reconstitution capacity.

The Protein kinase B (Akt)/Forkhead box O (FoxO) pathway is another pathway recently been identified as an essential regulator of HSC self-renewal. Members of this pathway such as Phosphatase and tensin homolog (PTEN), FoxO1, 2, and 3 have been shown to promote HSC self-renewal [94,95]. The loss of PTEN in HSCs resulted in impairment of self-renewal, and HSCs lacking FoxO had similar defects as well as hyper-proliferation and increased reaction oxygen species (ROS) [94,95]. In addition, other mechanisms such as telomere length and telomerase activity have been proposed to control HSCs self-renewal [96]. Together, this evidence suggests that HSCs are possibly maintained by multiple mechanisms. The regulation of HSC quiescence, self- renewal and differentiation is likely the result of an intricate balance among the different signalling pathways, and further research is needed to carefully dissect their relationships.

1.2.3 Leukaemic haematopoiesis The first formal identification of LSCs in AML was made recently by John Dick and colleagues, who have demonstrated that bone marrow cells with a phenotype of Lin- CD34+CD38- were enriched for LSCs by xenograft transplantation studies using NOD/SCID mice. These Lin-CD34+CD38- cells were capable of passing on the disease to recipient mice thereby demonstrating their self-renewal capacity. Lin-CD34+CD38+ or CD34- populations were not capable of transferring disease to recipient mice [97,98]. Furthermore, these LSCs could differentiate into mature blast cells and recapitulate the disease phenotype [98].

The development of AML is generally a multistep process where cooperating mutations are required for malignant transformation [9,99]. Experimental evidence generated from mouse models of AML has shown that AML arises from transformed HSCs and committed progenitors with accumulated genetic and epigenetic aberrations [100-102]. The initial transforming event(s) is thought to generate preleukaemic stem cells (pre- LSCs), which acquire further oncogenic hits to become fully transformed LSCs (Figure 1.5). Therefore, pre-LSCs consist of self-renewing HSCs or progenitors (which acquired self-renewing capacity through the initial oncogenic hits) that can give rise to leukaemia in vivo upon gradual accumulation of additional oncogenic hits, whereas LSCs are self-renewing leukaemia-initiating cells that generate short onset, fully penetrant leukaemia upon transplantation in vivo. For example, pre-LSCs could be generated from HSCs or progenitors such as granulocyte macrophage progenitors (GMPs) by MLL-AF9 translocation, the most common chromosomal translocation in paediatric AML [100,103]. These pre-LSCs expressing MLL-AF9 fusion oncogene possessed enhanced self-renewing ability and blocked differentiation partly via regulating its key target genes such as Meis1a, and subsequent genetic/epigenetic changes gave rise to fully developed LSCs upon transplantation in vivo [100]. Recent studies have shown that aberrant β-catenin signalling is a crucial step for HSCs or progenitors to develop into AML LSCs, highlighting the therapeutic potential of targeting β-catenin signalling [104,105].

Figure 1.5 Schematic diagram of normal and leukaemic haematopoiesis

(A) In normal haematopoiesis, self-renewing HSCs give rise to progenitor cells which ultimately differentiate into mature cells of the haematopoietic system. (B) In leukaemic haematopoiesis, HSCs or progenitor cells can be transformed into pre-LSCs through oncogenic hits. Pre-LSCs can develop into LSCs following second oncogenic hits and form the tumour bulk as leukaemic blasts. Adapted from [106].

1.2.4 Heterogeneity in LSCs LSCs were once thought of as a single population of leukaemia initiating stem cells; however, with more sophisticated techniques, increasing evidence suggests that LSCs display significant levels of heterogeneity. Therefore, a deeper understanding of their heterogeneous properties underlies better therapeutic targeting of LSCs in the future.

1.2.4.1 Phenotypic heterogeneity of LSCs Initial studies by John Dick and colleagues used NOD/SCID mice for xenograft transplantation and demonstrated that LSCs were defined as the Lin-CD34+CD38- population [97,98]. Since then, more refined xenograft transplant models were -/- -/- developed such as NOD/SCID/β2m (β2 microglobulin null) and NOD/SCID/IL2Rγ (interleukin 2 receptor γ chain null) mice. These new mouse models increased the engraftment efficiency and expanded the identification of LSCs to include more mature population (Lin-CD34+CD38+) in some AML patients [107], although the frequency of LSCs was still rare, up to 1 in 1 million cells. This suggests that heterogeneity exits within the LSC population itself.

The majority of AML patients have high CD34 expression, however, in some cases the percentage of CD34+ cells are very low such as those with NPM1 mutations. Contrary to the established LSC markers, in one-half of patients with NPM1 mutations LSCs were exclusively found in the CD34- population, whereas in the remaining cases LSCs were found in both CD34+ and CD34- populations [108]. Furthermore, more recent studies have shown that, in a mouse model of AML, LSC activity existed in three immunophenotypically distinct populations (Lin−Kit+, Gr1+Kit+ and Lym+Kit+), each of which corresponded to a distinct lineage on the haematopoietic hierarchy [109]. These cells recapitulated the original tumour phenotype in vivo and possessed a common signalling network including the Mitogen-Activated Protein Kinase (MAPK) pathway and Phosphoinositide 3-Kinase (PI3K)/Akt pathway [109]. In another study using human xenograft models, it was found that two separate populations demonstrated LSC

activity. Both populations were hierarchically organised, one resembled GMPs and the other mirrored lymphoid-primed multipotent progenitors [110].

One of the major challenges to design LSC-specific therapies is that LSCs are more similar to normal HSCs than their own differentiated progeny [97]. However, a number of surface markers including T-cell immunoglobulin domain and mucin domain 3 (TIM3), CD47, CD25 and CD32 have been identified that are more highly expressed on LSCs compared to normal HSCs. TIM3 is a T helper type 1 cell specific surface protein that negatively regulates T helper type 1 immunity and regulates macrophage activation [111,112]. It was shown that normal HSCs mainly resided in TIM3-negative population whereas LSCs were found mainly in TIM3-positive population. This differential expression of TIM3 also led to the separation between LSCs and HSCs in many AML samples [113]. CD47 is a transmembrane protein involved in a range of cellular processes such as apoptosis and proliferation. A recent study found that it was highly expressed in AML LSCs compared to HSCs and was a poor prognostic factor in three independent cohorts of AML patients [114]. By blocking CD47 using a monoclonal antibody, in vivo engraftment was significantly impaired, and anti-CD47 antibody treatment in human AML xenograft mice reduced the number of AML LSCs [114]. Despite the identification of markers that are highly expressed in LSCs, there is a large degree of variation among patients with regards to the percentage of LSCs that express these markers, including the more recently identified CD25 and CD32, which were found to be expressed in about 30% of LSCs [115]. Therefore this heterogeneity of surface marker expression adds another challenge to the development of LSC-targeted therapies.

1.2.4.2 Molecular heterogeneity of LSCs in AML LSCs and HSCs are strikingly similar with regard to their stem cell properties such as self-renewal and quiescence. Molecular characterisation of LSCs could shed light on the molecular drivers of their aggressiveness. Gene expression studies in both AML patient

samples and mouse models provided invaluable insight into the underlying molecular drivers of LSCs. Studies in leukaemia mouse models established by MLL fusion oncogenes revealed that LSCs resided in the c-Kit+ population [101]. These cells had a gene expression profile more similar to embryonic stem cells rather than adult HSCs [116]. Another study by Krivtsov et al. demonstrated that committed progenitors such as GMPs could be transformed into LSCs by MLL-AF9 and the resulting LSCs retained similar gene expression profiles to normal GMPs [100]. They also noticed only a subset of genes normally expressed in HSCs (the self-renewal-associated signature) is activated in these LSCs. These include Hoxa9, Hoxa10 and (Myeloid Ecotropic Viral Integration Site 1) Meis1. Clusters of Hox genes have been reported to be essential for haematopoiesis as well as HSC self-renewal [117]. Their dysregulations have long been recognised in many cases of AML and were linked to the pathogenesis of AML [118- 121]. Meis1 is also highly expressed in AML with MLL translocations and has been shown to play a crucial role in LSC self-renewal and proliferation [122]. More interestingly, this self-renewal-associated signature from MLL-AF9 transformed cells resembles that of the LSCs derived from loss of CEBPA p42 isoform [102], a completely different mutation from MLL-AF9, inferring a common mechanism for AML initiation.

1.2.5 Regulation of LSCs As discussed earlier, signalling pathways such as the Notch, Hedgehog and Wnt pathways have been shown to regulate self-renewal in HSCs. Unfortunately, the same pathways have been found to be dysregulated in cancers, suggesting functional overlap between normal and cancer stem cells. However, these pathways or genes are often aberrantly expressed or mutated, therefore, allowing a therapeutic window whereby targeting these pathways could destroy LSCs while sparing normal HSCs.

1.2.5.1 Notch signalling The Notch signalling pathway has been shown to be activated in several types of CSCs including breast, pancreatic and lung CSCs [123-125], and targeting of Notch signalling showed promising anti-tumour effect [126]. However, Notch signalling is silenced in human AML samples as well as in mouse AML LSCs, and studies have shown that inhibition of Notch did not affect the growth or survival of AML cells [127,128]. More interestingly, activation of Notch signalling suppressed AML leukaemogenesis and induced differentiation and cell death of AML LSCs, suggesting a tumour suppressive role of Notch signalling in AML [128]. Thus, targeting Notch signalling may not be a suitable approach for treating AML.

1.2.5.2 Hedgehog signalling The Hedgehog signalling pathway has also been implicated in various types of cancers. In multiple myeloma (MM), the Hedgehog pathway has been shown to be highly expressed in MM stem cells, and is required for maintaining the undifferentiated state of MM stem cells [129]. A recent study in CML showed that knockout of SMO as well as pharmacological inhibition of SMO significantly impaired CML development and depleted CML LSCs [80], justifying the therapeutic potential of targeting Hedgehog signalling. In AML with FLT3 ITD, it was shown that GLI2, an important effector of the Hedgehog pathway, was highly expressed compared to AML with wild-type FLT3 [130]. Activation of Hedgehog signalling exacerbated AML, and combined inhibition of FLT3 and Hedgehog signalling inhibited leukaemic growth [130]. Conversely, in an AML mouse model driven by MLL-AF9 translocation, SMO deficiency did not affect MLL-AF9 mediated leukaemogenesis, suggesting Hedgehog signalling is dispensable for this AML subtype.

1.2.5.3 Wnt/β-catenin signalling The Wnt/β-catenin pathway has been shown to be associated with a wide range of cellular processes such as self-renewal and proliferation, of which aberrant activation contributes to carcinogenesis. Hence dysregulation of Wnt signalling has been associated with several types of cancers and CSCs (Table 1.4). For example, high Wnt activity functionally defines the colon CSC population and increased Wnt signalling restores the CSC phenotype in more differentiated tumour cells [131]. Our laboratory has previously demonstrated that β-catenin signalling is essential for the conversion and establishment of LSCs in AML induced either by MLL-AF9 or by co-expression of key MLL targets Hoxa9 and Meis1a (A9M) [104,105,132]. Given that β-catenin activity has been shown to be dispensable for adult HSC survival and proliferation [88,90], targeting β-catenin signalling represents promising therapeutic value. However, β-catenin is very difficult to target pharmacologically due to its structure and nuclear localisation [133], and other Wnt components are not detected in AML LSCs [104], suggesting alternative regulators of β-catenin signalling. Hence there is great interest in identifying tractable signalling components that drive constitutive pathway activation in leukaemogenesis.

Table 1.4 Wnt/β-catenin signalling in CSCs

Cancer type Key findings References AML Wnt/β-catenin signalling is essential for the [104,105] establishment of MLL-LSCs. Colon cancer High Wnt activity functionally designates colon [131] CSCs and myofibroblast-secreted factors regulate CSC phenotype through β-catenin activation. Lung cancer Wnt/β-catenin and KrasG12D enhance lung [134] tumourigenesis and switch to a highly proliferative embryonic distal progenitor phenotype. Breast cancer Periostin regulates mammary CSCs maintenance [135] through recruitment of Wnt ligands. Brain cancer Foxm1, whose expression is increased by Wnt [136] ligands, promotes β-catenin signalling and is required for glioma formation. Prostate cancer Wnt signalling regulates the self-renewal of prostate [137] cancer cells with stem cell characteristics. Skin cancer Cutaneous cancer stem cell maintenance is dependent [138] on β-catenin signalling. Table adapted from [139]

1.2.6 Targeting LSCs Frequent disease relapse and chemoresistance are the main reasons for current treatment failure in the majority of cancers. It is postulated that CSCs are responsible for this treatment failure because they are intrinsically drug resistant and quiescent ([140-142]; Figure 1.6). Conventional chemotherapies target rapidly proliferating cancer cells and due to their quiescent cell cycle status CSCs readily evade destruction by such treatment. Several studies have reported that CSCs exhibited enhanced survival after standard therapy compared to other cancer cells within the tumour. Bao et al. demonstrated that CD133+ glioblastoma stem cells were less sensitive to ionizing radiation, the most effective therapy for grade IV glioblastoma, compared to CD133- glioblastoma cells [143]. This is due to more effective activation of DNA damage checkpoint in response to radiation [143]. Additionally, Breast CSCs have been reported to resist chemotherapy and radiation therapy due to their low level of ROS and higher expression of ROS scavenging systems, and depletion of ROS scavengers in these cells resulted in radiosensitisation [144,145]. In AML, the persistence of LSCs underlies the high frequency of relapse and poor overall survival as described in previous sections. This is further supported by the observation that AML patients with higher LSC gene expression signature have poorer clinical outcome [146-149]. Therefore, targeting LSCs may hold promising therapeutic potential for treating AML.

Antibody based cancer cell targeting has been an area of extensive research over the past decade. A prominent example is gemtuzumab ozogamicin, clinically available for treating AML for many years. It consists of an anti-CD33 antibody conjugated to the cytotoxic agent calicheamicin, therefore targeting leukaemic blast cells which express the CD33 marker [150]. Interestingly, CD33 was not thought to be present in LSCs. However, more recent studies have shown that some LSCs do express CD33 [151,152], hence explaining its efficacy against relapsed AML.

Interleukin 3 receptor alpha (IL-3Rα or CD123) is an aberrantly expressed surface marker in AML LSCs [153]. CD34+CD123+ cells from AML patients are capable of reconstituting the disease in NOD/SCID mice and have abnormal nuclear factor-kappa

Figure 1.6 Schematic diagram of conventional chemotherapy versus CSC targeted therapy

(A) Conventional chemotherapy targets the bulk of the tumour (green) but spares CSCs (yellow). Persistent CSCs can then regrow the tumour causing disease relapse. (B) CSC targeted therapy can eradicate CSCs which are responsible for the tumour maintenance, leading to tumour involution.

B (NFκB) activity [153,154]. In addition, AML patients with higher CD123 expression have a lower complete remission rate and poor clinical outcome [155]. A more recent study has demonstrated that CD123-neutralizing antibody (7G3) dramatically reduced AML LSC engraftment, improved mouse survival and significantly reduced AML burden in pre-established disease [156]. CSL362, a CD123 neutralizing antibody derived from 7G3, is under Phase I clinical trial and preliminary results showed that it is safe and well tolerated [157]. A Phase II clinical trial is currently been planned. In addition, a number of other surface markers have shown promising results against LSCs. A study by Jin et al. demonstrated that targeting of CD44 by a monoclonal antibody could eradicate LSCs in human AML xenograft mouse models through interference with the LSC niche [158]. An anti-TIM3 antibody has been shown to block engraftment of AML and eliminate LSCs in xenograft transplantations while sparing normal HSCs [159].

In addition to antibody based targeting, small molecule inhibitors have also been exploited to selectively target LSCs. Dimethylamino-parthenolide (DMAPT) is a synthetic analog of parthenolide (PTL), which occurs naturally in the feverfew plants. DMAPT and PTL can both selectively eradicate AML LSCs through inhibition of NFκB activity, induction of p53 and ROS production, while sparing normal HSCs [160,161]. PTL has poor pharmacologic properties; nevertheless, DMAPT was developed based on PTL to improve solubility and bioavailability while maintaining its efficacy against LSCs [161]. DMAPT is currently being evaluated in a Phase I-II clinical trial for treating AML.

Fenretinide is a synthetic retinoid but lacks the functional group for retinoid receptor activity. It has been used clinically as a chemopreventive agent for several cancers including breast and lung cancer [162-164]. A recent study evaluating the effect of fenretinide in AML showed that fenretinide can preferentially eradicate LSCs in primary AML samples while sparing normal counterparts [165]. It significantly impaired engraftment of AML LSCs but not normal HSCs in vivo. Mechanistic studies revealed that fenretinide induces production of ROS and suppresses NFκB and Wnt

signalling [165]. Fenretinide is now in Phase I clinical trials to determine the safety and dosing in AML patients.

The cure rate for AML using conventional chemotherapy has now reached a plateau. Despite intensive research on different combinations of chemotherapeutics, mortality rate has not significantly improved. Our current knowledge of LSCs is still limited; however, the LSC model unveils enormous opportunities for the development of better therapies that selectively target the “Achilles heel” of AML. A deeper understanding of the similarities and dissimilarities of LSCs to their normal counterpart underlies the success of future LSC targeted therapies. Hence more research is urgently needed in order to achieve better clinical outcome for AML patients.

1.3 G protein-coupled receptors G protein-coupled receptors (GPCRs) are membrane-spanning cell surface receptors consisting of seven transmembrane elements and an intracellular and extracellular terminus. A GPCR is activated by binding of its ligand to the extracelluar terminus which activates G proteins that are bound to the intracellular compartment of the receptor. This then triggers a series of signalling cascades that ultimately lead to regulation of cellular processes [166]. With about 900 members, GPCRs form the largest family of receptors that mediate signal transduction [166]. Because of the large and diverse family, GPCRs are involved in a wide variety of physiological processes such as immune modulation, mood regulation and inflammation [167-169]. Moreover, aberrant GPCR signalling has been shown to be associated with a wide range of human diseases including cardiovascular diseases and cancer [170]. Drugs that target dysfunctioning GPCRs have been successful in the clinic. In fact, GPCRs represents the largest family of validated pharmacological targets for treating human diseases [170]. However, only a small fraction of these GPCRs are targeted by current drugs. There is enormous therapeutic potential on the remaining GPCRs and huge efforts are being made towards understanding and targeting these receptors.

The association of GPCRs and cancer was pioneered by Young and colleagues who discovered that the MAS (Marker assisted selection) oncogene encodes a GPCR in 1986 [171]. Since then, an explosion of research effort ensued and a large number of GPCRs were identified and their involvement in various cancers characterised (Table 1.5). Recent studies using human xenograft mouse models of AML have revealed a number of LSC specific GPCRs [115,146]. Analysis of gene expression profiles of human LSCs compared to normal HSCs identified LSC-specific signatures, which include GPCRs such as GPR126, GPR34 and GPR84 [115,146]. Our laboratory has recently showed that GPR84 is highly expressed in β-catenin-driven AML LSCs and suppression of GPR84 significantly impaired LSC maintenance in fully developed AML through inhibition of β-catenin signalling [132]. Since GPCRs are highly tractable drug targets, inhibition of LSC-specific GPCR signalling may represent promising therapeutic value.

Table 1.5 Selected GPCRs that are involved in cancer

Receptor Cancer Lysophosphatidic acid receptor Colon cancer [172]; ovarian cancer [173]; breast cancer [174] Sphingosine 1-phosphate receptor Glioma [175]; ovarian cancer [174] Protease activated receptor 1 Breast cancer [176]; skin cancer [177]; leukaemia [178] Prostaglandin E2 receptor Colon cancer [179]; prostate cancer [180]; breast cancer [181] Bradykinin receptors Chondrosarcoma [182]; prostate cancer [183] Angiotensin II type 1 receptor Gastric cancer [184]; prostate cancer [185] G protein coupled oestrogen receptor Breast cancer [186]; ovarian cancer [187]; thyroid cancer [188] Smoothened receptor Multiple myeloma [129]; colon cancer [189] Frizzled receptor Colon cancer [190]; lung cancer [191]; leukaemia [192] Table adapted from [170]

1.3.1 G proteins G proteins, or guanine nucleotide-binding proteins, are signalling transduction proteins involved in relaying external cues to the cells. There are two classes of G proteins: monomeric small GTPases and heterotrimeric G proteins. The monomeric small G proteins are functionally independent enzymes that hydrolyse guanosine triphosphate (GTP) to guanosine diphosphate (GDP). The heterotrimeric G proteins are immediate downstream partners of GPCRs that are involved in transmitting signals from GPCRs to downstream effectors, and are composed of α, β and γ subunits. The Gα subunit has a GTPase domain, and is activated by the receptor through exchange of GDP to GTP.

There are four main families of Gα proteins: Gαi, Gαs, Gαq and Gα12/13 family, each of which has distinct downstream signal transduction mechanisms [193]. Upon receptor activation, the Gα subunit is dissociated from the receptor and β and γ subunits then transmit the signal downstream. The β and γ subunits are closely bound and are also released upon receptor activation to interact with downstream effectors [194].

G proteins are crucial in signal transduction of GPCRs and abnormal G protein signalling was identified in several cancers. Recent large scale deep-sequencing analysis revealed that mutations in G proteins were present in a wide variety of cancer types [195]. For example, GNAS (which encodes for Gαs protein) mutations were found in pituitary tumours (28%), pancreatic cancers (12%), thyroid adenomas (5%), colon cancers (4%), parathyroid cancers (3%) and ovarian cancers (3%) [195]. GNAQ (which encodes for Gαq protein) mutations were identified in a large number of uveal melanoma, tumours arising from the meninges and blue naevi of the skin [195]. Consistent with the role of aberrant GPCR signalling in cancer development, activated Gα proteins have been shown to facilitate tumourigenesis. For instance, activating mutations of Gαq or Gα13 were shown to induce malignant transformation [196,197]. Although multiple GPCRs were identified that are involved in AML leukaemogenesis, our knowledge on the role of G proteins in AML is still limited. G proteins are essential players in GPCR signalling, hence further characterisation of GPCRs and G protein

signalling in AML may provide valuable information on developing more effective therapeutics against AML.

1.3.2 Regulator of G protein signalling (Rgs) family Regulator of G protein signalling (Rgs) family consists of a group of evolutionally conserved proteins that mediate a wide variety of signals through regulating G protein signalling. Each Rgs protein contains an Rgs domain which acts as GTPase activating proteins to accelerate the hydrolysis of GTP to GDP, resulting in deactivation of Gα subunits and premature termination of downstream signalling [198-200]. The Rgs proteins can also act as antagonists by preventing activated Gα subunits from interacting with downstream effectors [201,202]. Since GPCR and G protein signalling are associated with many different cancers as described in the previous sections, it is not surprising that Rgs proteins are also involved in diverse types of cancer. Multiple Rgs proteins were differentially expressed in cancers such as prostate cancer [203,204] and breast cancer [205,206]. In AML with FLT3 ITD mutation, Rgs2 expression was decreased and overexpression of Rgs2 inhibited AML cell proliferation, counteracted differentiation block induced by FLT3 ITD and impaired leukaemic transformation [207]. Therefore, cancer cells could be targeted by manipulating Rgs expression and further studies are required to better understand the role of Rgs proteins in cancer.

1.3.3 Leucine-rich repeat-containing G-protein coupled receptor 4 (Lgr4) The leucine-rich repeat-containing G-protein coupled receptor 4 (Lgr4 or GPR48) belongs to a class of GPCRs characterised by a large extracellular domain containing multiple leucine-rich repeats [208]. Lgr4 has two other close homologues, Lgr5 and Lgr6, that share over 50% sequence identity. Phylogenetic analysis showed that Lgr4 is highly conserved throughout evolution [209]. Gene knockout studies revealed the involvement of Lgr4 in embryonic development and tissue homeostasis (Table 1.6). More recent studies have shown that Lgr4 plays a crucial role in the survival and self-

renewal of colon stem cells [210,211]. In addition, Lgr4 is a key regulator of prostate stem cells and mammary stem cells, and is involved in prostate and mammary gland development [212,213].

The endogenous ligands of Lgr4 remained elusive for more than a decade since the discovery of this receptor [208]. The failure to identify its ligands, together with lack of understanding of its signal transduction cascades, greatly hampered further research into its biological function and significance in human disease. However, in 2011 several independent groups identified that Roof plate-specific spondin 1 – 4 (Rspo1-4), a group of secreted proteins that are known to potentiate β-catenin expression in normal Wnt/β- catenin signalling and act as stem cell growth factors, function as ligands of Lgr4 [210,214-216]. They have shown that Rspo1-4 can bind to Lgr4 and activate β-catenin signalling in HEK293T cells. Rspo1-4 could also synergise with Wnt ligands such Wnt3a to potentiate β-catenin signalling in these cells [210,214-216]. Aberrant Wnt/β- catenin signalling is characteristic of several cancer types such as colon cancer and therefore a number of studies interrogated the role of Lgr4 in tumourigenesis. For instance, a systematic analysis of over 70 pairs of primary human colon cancer samples revealed frequent gain-of-function gene fusions of Rspo2/Rspo3 together with Wnt target genes [217]. In addition, all of the Rspo fusion tumours expressed Lgr4 at high levels, suggesting a possible role of Lgr4 in colon cancer [217]. Another study by Wu et al. showed that Lgr4 is a poor prognostic factor of colorectal cancer, and it promoted tumour metastasis and activated β-catenin signalling in colorectal cancer, suggesting that Lgr4/β-catenin signalling is an important contributor to colon cancer [218].

As described earlier, β-catenin signalling plays a crucial role in the development of AML LSCs [104,105,132]. Despite the important roles Rspo-Lgr4 plays in normal stem cells and colon cancer [210,211,217,218], their involvement in leukaemogenesis has not yet been investigated.

Table 1.6 Lgr4 tissue expression patterns in mice

Tissue Loss-of-function phenotype References Adrenal gland Unknown [219] Bladder Unknown [219] Bone Block in osteoblast differentiation and bone [219,220] remodelling Eye impaired iris/cornea development [219,221] Male reproductive Impaired epididymis development [222] tract Intestine Impaired epithelial proliferation and block in [210,211] Paneth cell differentiation Gall bladder Impaired embryonic development [219,223] Kidney Impaired kidney morphogenesis/glomerulus [224] formation Skin Impaired keratinocyte migration (eyelid) and [225,226] impaired hair follicle formation Tongue Unknown [219] Trachea Unknown [219] Cartilage Unknown [219] Teeth Unknown [219] Female reproductive Impaired uterus development/reduced fertility [227,228] tract Mammary gland Impaired mammary gland branching/elongation [229] Liver Impaired erythropoiesis [230] Table adapted from [231]

1.4 Thesis hypothesis and aims AML is the most deadly form of leukaemia resulting in the highest number of leukaemia-associated deaths worldwide. The stagnant clinical outcome of AML is mainly due to high rate of relapse, after which cancer cells become resistant to current therapies. It is believed that the persistence of LSCs is the main culprit due to their intrinsic quiescence [232]. Recent studies have identified that β-catenin signalling is crucial for the development of LSCs [104,105]. However, β-catenin is very difficult to target due to its structure [133], and other Wnt components were not detected in LSCs [104], suggesting alternative regulators of β-catenin signalling. Previous studies have shown that Lgr4 is important in colon stem cell maintenance and is associated with β- catenin-driven colon cancer [210,211,217,218]. Therefore, Lgr4 may play an important role in the development of LSCs in AML (Figure 1.7).

Hypothesis: Lgr4 sustains aberratn β-catenin signalling in AML LSC, and plays a crucial role in AML initiation and progression.

Aim 1: Investigate the role of Lgr4 in AML development.

Aim 2: Characterise crucial components of the Lgr4 signalling cascade.

Aim 3: Identify mechanisms of Lgr4 signalling regulation

Figure 1.7 Schematic diagram of Lgr4/Wnt signalling in 293T cells

Activation of β-catenin can be achieved by stimulating Lgr4 with its ligand RSPO or through the canonical Wnt pathway. In addition, Lgr4 synergises with the Wnt pathway to further activate β-catenin signalling. Rgs can bind to and inhibit G protein signalling. Despite Lgr4 being a GPCR, it is not know which G protein Lgr4 couples to, and little is known on the ability of G protein to activate β-catenin signalling. However, the recent finding that β-catenin signalling is crucial for AML LSC development might suggest a functional role of Lgr4 in this particular disease.

2 Materials and Methods

2.1 Materials

Table 2.1 Purchased Reagents

Reagent Supplier 7-Aminoactinomycin D (7AAD) BD Australia and New Zealand Annexin V APC BD Australia and New Zealand Antimycin A Sigma-Aldrich Pty Ltd, Australia APC BrdU flow kit BD Australia and New Zealand ATP assay kit Abcam, UK BD Pharm LyseTM BD Australia and New Zealand Bicinchoninic acid (BCA) protein assay kit Thermo Fisher Scientific Australia Deoxynucleoside triphosphate (dNTP) Invitrogen Australia DH5α competent cells Thermo Fisher Scientific Australia Dimethyl sulphoxide (DMSO) Sigma-Aldrich Pty Ltd, Australia Dithiothreiol (DTT) Invitrogen Australia DNase BD Australia and New Zealand DNase-free RNase Invitrogen Australia DNase-free RNase-free sterile water Invitrogen Australia DNase I Invitrogen Australia Dulbecco’s Modified Eagle’s Medium (DMEM) Invitrogen Australia Ethanol Sigma-Aldrich Pty Ltd, Australia Fetal calf serum (FCS) Invitrogen Australia First strand buffer (FSB) Invitrogen Australia GP-antagonist 2A (GP) Merck Millipore Corporation, USA 7-Aminoactinomycin D (7AAD) BD Australia and New Zealand

Hexadimethrine bromide (polybrene) Sigma-Aldrich Pty Ltd, Australia IC261 Sigma-Aldrich Pty Ltd, Australia Iscove's Modified Dulbecco's Medium (IMDM) Invitrogen Australia Lipofectamine 2000 Invitrogen Australia Luria broth (LB) Sigma-Aldrich Pty Ltd, Australia Methanol Univar Australia Pty Ltd MethoCult M3234 (M3234) StemCell Technologies, Canada MitoSOXTM Red Thermo Fisher Scientific Australia MitoXpress® Xtra kit Luxcel Biosciences, Ireland Moloney Murine Leukaemia Virus Reverse Invitrogen Australia Transcriptase Non-fat dairy milk Coles Australia NuPAGE® LDS Sample Buffer (4X) Thermo Fisher Scientific Australia NuPAGE® Sample Reducing Agent (10X) Thermo Fisher Scientific Australia Opti-MEM I reduced serum media Invitrogen Australia Penicillin (10 mg/mL), Streptomycin (10 Invitrogen Australia mg/mL), L-glutamine (29.2 mg/mL) [PSG] Pertussis toxin (PTX) Sigma-Aldrich Pty Ltd, Australia Perchloric acid (PCA) Sigma-Aldrich Pty Ltd, Australia Phosphate buffer Sigma-Aldrich Pty Ltd, Australia Phosphate buffered saline (PBS) Invitrogen Australia Polyvinylidene difluoride (PVDF) Merck Millipore Corporation, USA Ponceau S Sigma-Aldrich Pty Ltd, Australia Potassium hydroxide (KOH) Sigma-Aldrich Pty Ltd, Australia Precision Plus ProteinTM Dual Colour Standard Bio-Rad Laboratories, Inc., USA Protease Inhibitor Cocktail Sigma-Aldrich Pty Ltd, Australia Puromycin Sigma-Aldrich Pty Ltd, Australia

QIAprep Spin Miniprep kit QIAGEN Pty Ltd, Australia Random primers Invitrogen Australia Recombinant mouse stem cell factor (SCF) Peprotech Inc., USA Recombinant mouse interleukin-3 (IL-3) Peprotech Inc., USA Recombinant mouse interleukin-6 (IL-6) Peprotech Inc., USA Recombinant mouse R-spondin 1 (Rspo1) Peprotech Inc., USA Recombinant mouse R-spondin 2 (Rspo2) Peprotech Inc., USA Recombinant mouse R-spondin 3 (Rspo3) Peprotech Inc., USA Recombinant mouse R-spondin 4 (Rspo4) Peprotech Inc., USA RNAsin Invitrogen Australia Rotenone Sigma-Aldrich Pty Ltd, Australia SCH79797 Tocris Bioscience, UK Sodium dodecyl sulphate (SDS) Bio-Rad Laboratories, Inc., USA SuperSignal HRP substrate Thermo Fisher Scientific Australia Super HR-G 30 autoradiography film Fujifilm corporation Japan SYBR green Thermo Fisher Scientific Australia Tris buffered saline (20X, TBS) VWR International, Australia Trypan Blue, 0.4% Invitrogen Australia Trypsin Ethylenediaminetetraacetic acid solution Invitrogen Australia Tween 20 Sigma-Aldrich Pty Ltd, Australia

Table 2.2 Prepared Reagents

Name Constituents

10X Running buffer 29 g Tris, 144 g glycine, up to 1 L with milliQ H2O

10X Transfer buffer 30.3 g Tris, 144 g glycine, up to 1 L with milliQ H2O Annexin V binding 10 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid buffer (pH 7.4) (HEPES), 140 mM NaCl, 2.5 mM CaCl2 Giemsa stain 1 g of Giemsa stain powder in 66 mL glycerol and 66 mL methanol Wright stain 3 g/L Wright stain powder in methanol RIPA buffer 1.5 mL 5M NaCl, 0.5 mL NP-40, 5 mL 5% Na deoxycholate, 250 uL 20% SDS, 2.5 mL 1M Tris-HCl pH7.5, 40.25 mL H2O TBST 50 mL TBS (20X), 500 uL Tween 20, up to 1 L with milliQ H2O

Table 2.3 Cells

Name Cell type Exogenous genes Growth media 293T Human embryonal kidney SV40 large T DMEM/10%FCS antigen GMPA9M Lin− (CD3, CD4, CD8a, HoxA9/Meis1a- M3234 CD19, B220, Gr-1, Ter119, GFP IL7R)c-Kit+Sca-1- CD34+FcRγ+ mouse bone marrow cells GMPMLL Lin−c-Kit+Sca-1-CD34+FcRγ+ MLL-AF9-GFP M3234 mouse bone marrow cells KLSA9M c-Kit+Lin−Sca-1+ mouse bone HoxA9/Meis1a- M3234 marrow cells GFP KLSMLL c-Kit+Lin−Sca-1+ mouse bone MLL-AF9-GFP M3234 marrow cells L Cells Mouse fibroblast None DMEM/10%FCS (ATCC® CRL2648 ™) L-Wnt3a Mouse fibroblast Wnt3a expressing DMEM/10%FCS Cells construct (ATCC® CRL2648 ™) MLL- Gadd45a-/- c-Kit+ mouse bone MLL-AF9-GFP M3234 Gadd45a-/- marrow cells MLL-WT Wild-type (WT) c-Kit+ mouse MLL-AF9-GFP M3234 bone marrow cells

Table 2.4 Vectors

Name Vector Supplier Empty vector (EV) pMSCV-Neo OriGene Technologies, Inc., USA Gαq shRNA#1 psi-LVRU6MH GeneCopoeia, Inc., USA Gαq shRNA#2 psi-LVRU6MH GeneCopoeia, Inc., USA Gαq shRNA#3 psi-LVRU6MH GeneCopoeia, Inc., USA Lgr4 cDNA pMSCV-Neo OriGene Technologies, Inc., USA Lgr4 shRNA#1 psi-LVRU6MH GeneCopoeia, Inc., USA Lgr4 shRNA#2 psi-LVRU6MH GeneCopoeia, Inc., USA Lgr4 shRNA#3 psi-LVRU6MH GeneCopoeia, Inc., USA Lgr4 shRNA#4 psi-LVRU6MH GeneCopoeia, Inc., USA Scrambled shRNA psi-LVRU6MH GeneCopoeia, Inc., USA β-cat* pMSCV-Neo Clontech Laboratories USA

Table 2.5 FACS antibodies

Antibody Supplier APC-anti-BrdU BD Australia and New Zealand APC-anti-CD117 (c-Kit) BD Australia and New Zealand Biotin-CD19 BD Australia and New Zealand Biotin-CD3 BD Australia and New Zealand Biotin-CD4 BD Australia and New Zealand Biotin-CD8a BD Australia and New Zealand Biotin-IL-7R BD Australia and New Zealand Biotin-Ter119 BD Australia and New Zealand BV421-Sca-1 BD Australia and New Zealand PerCP-Cy5.5-Streptavidin BD Australia and New Zealand

Table 2.6 Western blotting primary antibodies

Antibodies MW (kD) Dilution Buffer Supplier Mouse polyclonal anti- 92 1:500 10% NFDM BD Australia β-catenin antibody and New Zealand Rabbit polyclonal anti- 42 1:2000 0.5% NFDM Sigma-Aldrich Actin antibody Pty Ltd, Australia Rabbit polyclonal anti- 48 1:1000 5% NFDM Abcam, UK Csnk1e antibody Rabbit polyclonal anti- 31 1:500 5% NFDM Santa Cruz Cytochrome b Biotechnology, reductase antibody USA Rabbit polyclonal anti- 104 1:500 5% NFDM Abcam, UK Lgr4 antibody Rabbit polyclonal anti- 100 1:500 5% NFDM Abcam, UK Lgr5 antibody Rabbit polyclonal anti- 94 1:1000 5% NFDM Cell signalling phospho-β-catenin Technology, antibody USA Rabbit polyclonal anti- 24 1:1000 1% NFDM Abcam, UK Rgs1 antibody All primary antibodies were probed at 4°C overnight. MW: molecular weight; NDFM: non-fat dairy milk

Table 2.7 Western blotting secondary antibodies

Antibodies Dilution Buffer Supplier Goat anti-rabbit 1:10,000 0.5% NFDM Thermo Fisher Scientific Australia Goat anti-mouse 1:500 10% NFDM Thermo Fisher Scientific Australia All secondary antibodies were probed at room temperature for 1 hour. NDFM: non-fat dairy milk

Table 2.8 Primers

Gene Sequence Asb2 Forward 5'- CAT GGA CCA GGT CCC CTT CG-3' Reverse 5'- CTT CAA GGC CTC TTCA TCC CCT TC-3' Ddit3 Forward 5'- CCC CAG GAA ACG AAG AGG AAG AAT-3' Reverse 5'- CCT GGG CCA TAG AAC TCT GAC TGG-3' F2r Forward 5'- CGT CCC TAT GAG CCA GCC A -3' Reverse 5'- CAC CGT AGC ATC TGT CCT CTC TG -3' Gadd45a Forward 5'- ATG ACT TTG GAG GAA TTC TCG GCT G-3' Reverse 5'- ACG TTG AGC AGC TTG GCC-3' Gadd45g Forward 5'- CAC AGC CAG GAT GCA GGG-3' Reverse 5'- TCT TCA TCG GCA GCC AGC AC-3' Gbp2 Forward 5'- GAA GAT GTT GAG AAG GGT GAC AAC CAG-3' Reverse 5'- CTC CGT CAC AAT AGT GCA GCT GG-3' Hoxa5 Forward 5'- CCA CAT CAG CAG CAG AGA GG-3' Reverse 5'- GGG TCA GGT AGC GGT TGA AG-3' Irf5 Forward 5'- GCT CCC ACA GAG AGC CAA CC-3' Reverse 5'- CTT TGG GTA AGG AAT AGG GTG CGT TG-3' Klf4 Forward 5'- CCG ACT AAC CGT TGG CGT GAG-3' Reverse 5'- GGT CTC CCT CCG GGC TA-3' Lcn2 Forward 5'- TCC TGG TCA GGG ACC AGG A-3' Reverse 5'- GTG GTG GCC ACT TGC ACA TTG-3' Mmp12 Forward 5'- TTG CAT TTG GAG CTC ACG GAG AC-3' Reverse 5'- AAG AGG TTT GTG CCT TGA AAA CTT TTA G-3' Pdgfa Forward 5'- GAC TCC GTA GGG GCT GAG AGA-3' Reverse 5'- CGT AAA TGA CCG TCC TGG TCT TGC-3' Ptgir Forward 5'- TGC CTC TCA TGA TCC GAG GCT-3' Reverse 5'- CAA CCA GAA CTT GAG GCG TTG GA-3'

Rgs1 Forward 5'- CAA GTC CAA AGA CAT ACT TTC TGC TGA A-3' Reverse 5'- TTA TAG TCC TCA CAA GCC AAC CAG A-3' Timp2 Forward 5'- ATG CAG ACG TAG TGA TCA GAG CCA-3' Reverse 5'- GTC AGG TCC TTT GAA CAT CTT TAT CTG CT-3'

2.2 Methods

2.2.1 Maintenance of cells

2.2.1.1 Cell culture HEK 293T cells were cultured in T75 tissue culture flasks in 12 mL of Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% (v/v) fetal calf serum (FCS), 100 U/mL penicillin, 100 μg/mL streptomycin and 29.2 mg/mL L-glutamine (PSG). Murine L cells (ATCC® CRL2648™) and L-Wnt3a cells (ATCC® CRL2647™) were generously provided by Dr Sylvie Shen (Children’s Cancer Institute Australia, University of New South Wales, Sydney) and were cultured as above. All cells were grown at 37°C in a humidified environment with 5% CO2.

Murine LSCs and pre-LSCs were cultured in 35 mm culture dishes in 1.1 mL of methylcellulose (MethoCult™ M3234) enriched with 19% Iscove's Modified Dulbecco's Medium (IMDM) supplemented with PSG and 50 ng/mL interleukin-3 (IL- 3). Murine HSCs were cultured as above with additional supplements of 50 ng/mL stem cell factor (SCF) and 50 ng/mL interleukin-6 (IL-6). All cells were grown at 37°C in a humidified environment with 5% CO2.

Cell counts were performed by adding equal volume of 0.4% sterile-filtered trypan blue solution followed by counting on a haemocytometer under a light microscope.

2.2.1.2 Generating Wnt3a conditioned medium L-Wnt3a cells or L cells (as control) were harvested once they reached confluence and were split 1:10 in 10 mL DMEM/10%FCS in a T75 flask. After 4 days incubation at

37°C/5% CO2, the culture medium was removed and sterile filtered to produce the first batch of medium. Another 10 mL fresh culture medium was added and cells were incubated for another 3 days at 37°C/5% CO2. The culture medium was again removed and sterile filtered to produce the second batch of medium. Cells were discarded and the

Wnt3a conditioned medium or control conditioned medium was made by combining the first and second batch of media. For use in treating pre-LSCs, the Wnt3a conditioned medium or control conditioned medium was mixed with culture medium at a ratio of 1:10 (v/v).

2.2.1.3 Freezing and thawing cells For cell freezing, cells were centrifuged at 500 × g for 5 minutes at room temperature. The cell pellet was then resuspended in ice-cold freezing media consisting of 90% FCS and 10% dimethyl sulphoxide (DMSO). The cell suspension was transferred to cryovials at 1 million cells per vial, placed into a cell freezing container and stored at - 80°C for 24 h. Cells were then transferred to a vapour phase liquid nitrogen tank for long-term storage.

For thawing cell lines, cryovials were agitated in a 37°C water bath until completely thawed and then added to 10 mL of thawing media consisting of 80% DMEM, 20% FCS and PSG. After centrifuging at 500 × g for 5 minutes at room temperature, the cell pellet was resuspended in 12 mL of thawing media and transferred into a T75 flask.

After 24 hours incubation at 37°C/5% CO2, the media was replaced with DMEM/10% FCS supplemented with PSG.

For thawing LSCs and pre-LSCs, cyovials were agitated in a 37°C water bath until completely thawed and cells were washed with 14 mL IMDM. Cells were then resuspended in 50 µL IMDM and transferred into 1.1 mL M3234/19% IMDM supplemented with PSG and 50 ng/mL IL-3 in a 35 mm tissue culture dish.

2.2.2 Stable cell line production

2.2.2.1 Bacterial transformation and plasmid preparation ShRNA/over-expression plasmids were transformed into chemically competent DH5α Escherichia coli cells as per manufacturer’s instructions. Briefly, DH5α cells were incubated with 10 ng (1 µL) plasmid for 30 minutes on ice. The mixture was then incubated at 42°C for 40 seconds and immediately immersed in ice for 2 minutes. A 500 µL volume of pre-warmed Luria broth (LB) was added per sample and the suspension was incubated at 37°C with shaking for 1 hour. The samples were then transferred into a 15 mL tube containing LB supplemented with 100 mg/mL ampicillin and incubated at 37°C for 16 hours prior plasmid preparation.

Bacterial pellets were harvested by centrifugation at 3000 × g for 15 minutes at room temperature. Plasmids were subsequently extracted from the bacterial pellets using the QIAprep Spin Miniprep kit (QIAGEN) as per manufacturer’s instructions. Isolated plasmids were then sequenced by the Australian Genome Research Facility (NSW, Australia).

2.2.2.2 Transient virus production in 293T cells HEK 293T cells were transiently transfected with shRNA/cDNA plasmids using Lipofectamine 2000 as per manufacturer’s instructions. Briefly, cells were seeded at a density of 0.5 × 106 cells per well in a 6-well plate in DMEM/10% FCS and incubated at 37°C/5% CO2 for 24 hours. A 6 µL volume of lipofectamine 2000 was added to 244 µL of Opti-MEM I reduced serum medium and incubated for 10 minutes at room temperature. Plasmid mixtures for lentiviral shRNA (1 µg of shRNA plasmid, 0.75 µg psPAX2 packaging plasmid and 0.25 µg pMD2.G envelope plasmid) or retroviral cDNA (1 µg of cDNA plasmid and 1 µg of psi-eco packaging vector) were diluted in Opti-MEM I reduced serum medium to a final volume of 250 µL. The plasmid solution was combined with the lipofectamine 2000 solution and the resulting mixture was incubated at room temperature for 30 minutes. The mixture was then added drop wise to

each well containing 500 µL DMEM. After 16 hours incubation at 37°C/5% CO2, media was replaced with 3 mL DMEM/10%FCS. After 24 hours incubation at 37°C/5%

CO2, virus-containing media (VCM) was collected and filtered through a 0.45 µm low- protein binding sterile filter. VCM was then concentrated via centrifugation at 3000 × g at 4°C for 24 hours. Again, 3 mL of fresh DMEM/10%FCS was added to each well and cells were incubated at 37°C/5% CO2 for a further 24 hours. At the 48 hour time point VCM was again harvested and concentrated as described above. Concentrated VCM was stored at -80°C for later use.

2.2.2.3 Viral transduction To produce stable knockdown or over-expression of the gene of interest LSCs and pre- LSCs were subjected to two rounds of viral transduction. LSCs and pre-LSCs were seeded at a density of 5 × 104 cells in a U-bottomed 96-well plate and incubated with concentrated VCM supplemented with 8 μg/mL polybrene and 50 ng/mL IL-3 (50 ng/mL SCF and 50 ng/mL IL-6 was added when transducing normal HSCs). Cells were centrifuged for 1.5 hours at 500 × g at 30°C and then incubated for 4 hours at 37°C/5%

CO2. The media was replaced with fresh concentrated VCM supplemented with 8 μg/mL polybrene and 50 ng/mL IL-3. Cells were centrifuged for 30 minutes at 500 × g at 30°C and then incubated for 16 hours at 37°C/5% CO2. Following the second round of transduction, cells were washed twice with IMDM and plated into 1.1 mL M3234/19% IMDM supplemented with PSG and 50 ng/mL IL-3.

Transduced cells were selected via antibiotics for 5 days (G418 for cDNA vectors; puromycin for shRNA vectors) or by fluorescence-activated cell sorting (FACS).

2.2.3 Flow cytometric analysis

2.2.3.1 Flow cytometry and FACS analysis Cells (0.5 – 1 × 106) were washed with PBS and resuspended in 500 µL PBS. Samples were then analysed using an LSRFortessa™ cell analyzer (BD) or FACSCanto™ II flow cytometer (BD). According to the samples being examined, cells were analysed for green fluorescent protein (GFP) (B-530/30), Allophycocyanin (APC) (R-670/14), Phycoerythrin (PE) (B-585/15) and/or PerCP-Cy5.5 (B-710/50) fluorescence. FACS was performed using a BD Influx™ high-speed cell sorter (BD). Cells were sorted under sterile conditions into 5 mL tube containing 2.5 mL PBS and 1 mL FCS. Compensation settings were determined using unstained and single antibody-stained cells. Data were analysed using FlowJo software version 7.6.5 (TreeStar).

2.2.3.2 Haematopoietic cell isolation by FACS Mouse bone marrow cells (30 – 50 × 106) were washed with PBS and resuspended in 1 mL PBS. A 20 µL volume of each biotin conjugated lineage antibody (CD3, CD4, CD8a, CD19, B220, Gr-1, Ter119 and IL-7R) were added and the cells were incubated at 4°C for 30 minutes. Cells were washed with PBS and 20 µL each of PerCP- streptavidin, FITC anti-mouse Sca-1, APC anti-mouse c-Kit, PE anti-mouse CD34 and PE-Cy7 anti-mouse FcγRII/III were added. After 20 minutes incubation at 4°C, cells were washed with PBS prior to cell sorting. HSCs were sorted from Lin- Sca-1+ c-kit+, and myeloid progenitor GMPs were isolated from Lin− Sca-1− c-kit+ CD34+ FcγRII/IIIhi.

2.2.4 Gene expression analysis

2.2.4.1 RNA isolation Total RNA was extracted using a RNeasy Mini kit (QIAGEN), as per manufacturer’s instructions. A cell pellet containing 1 × 106 cells was lysed using 350 µL buffer RLT. A 350 µL volume of 70% ethanol was added and mixed well by pipetting. The resulting

mixture was transferred to an RNeas spin column placed in a 2 mL collection tube and centrifuged at 16000 × g for 30 seconds. The flow-through was discarded. A 700 µL volume of buffer RW1 was then added to the RNeasy spin column and was spun at 16000 × g for 30 seconds. The flow-through was discarded. The same procedure was followed twice using 500 µL buffer RPE each time and the flow-through was discarded. The RNeasy spin column was then placed in a new 1.5 mL collection tube and 30 µL RNase-free water was added in the spin column. The column was centrifuged at 16000 × g for 1 minute to elute the RNA. Total RNA was quantitated using a NanoDrop 1000 Spectrophotometer (Thermo Fisher Scientific) and stored at -80°C.

2.2.4.2 cDNA preparation One μg of RNA was mixed with 0.5 µL of DNase I, 2 µL of first strand buffer (FSB) and nuclease-free water to make up 10 µL final solution. The solution was heated to 37ºC for 30 minutes and 75ºC for 5 minutes. The solution was transferred back on ice and mixed with 10 µL of cDNA mastermix containing 2 µL FSB, 1 µL Moloney Murine Leukaemia Virus Reverse Transcriptase (M-MLV RT), 0.1 μg/μL random primers, 0.5 µL RNAsin ribonuclease inhibitor, 1 µL dithiothreitol (DTT), 0.5 mM dNTPs and 3.5 µL nuclease-free water. The resulting mixture was heated at 37ºC for 1 hour prior to dilution with 30 µL nuclease-free water.

2.2.4.3 Quantitative real-time PCR Following reverse transcription, cDNA was analysed for target gene expression by quantitative real-time polymerase chain reaction (qRT-PCR) using QuantiTect SYBR® Green PCR Kits (QIAGEN) in 96-well format at a final volume of 25 μL per well. Each reaction contained 12.5 μL SYBR Green, 150 nM forward and reverse primer (the primers used for all reactions are listed in Table 2.8), 100 ng cDNA and 8.5 μL RNase- free water. The plate was centrifuged at 300 × g for 1 minute and amplified using the ABI 7900HT Real Time PCR machine (Applied Biosystems) using the default protocol.

Briefly, samples were heated to 50ºC for 2 minutes, then 95ºC for 10 minutes. Then samples were amplified for 40 cycles at 95ºC for 15 seconds and 60ºC for 1 minute. Control samples containing no template cDNA and no PCR reagent mix were run alongside sample reactions.

Data from qRT-PCR were analysed using SDS2.3 software (Applied Biosystems). The efficiency, reproducibility and dynamic range of the assay was determined by constructing a standard curve using serial dilution of a known template every time a new primer pair was used. Primer pairs were used only if efficiency of the assay was above 90%, the slope of the curve around 3.0 and the threshold cycle (Ct) values for all technical replicates were similar. The presence of non-specific products were identified by constructing melting curves for each samples at 0.1ºC intervals between 60ºC and 95ºC. For each gene under analysis, the Ct was manually set in the logarithmic amplification phase. Relative expression levels were determined using the comparative Ct (ΔΔCt) method [233]. Fold change (FC) values were calculated by the formula of FC = 2-(ΔΔCt).

2.2.4.4 Gene expression microarray Total RNA was extracted as described in Section 2.2.15. Triplicates were used per sample. Samples were analysed using Illumina mouse WG-6 v2.0 expression arrays. Quality control, labelling, hybridization and scanning were performed by the Ramaciotti Centre for Gene Function Analysis (University of New South Wales), with each sample hybridized to an individual array. Microarray results were analysed using GenePattern software (Broad Institute, MIT) under the expertise and guidance of Dr Bing Liu (Kid’s Cancer Alliance, UNSW) and Dr Amit Sinha (Harvard Medical School).

2.2.5 Protein analysis

2.2.5.1 Isolation of total cellular protein Total cellular protein was extracted by resuspending a cell pellet containing 2 × 106 cells in 30 μL ice-cold Radioimmunoprecipitation assay (RIPA) buffer (described in Table 2.2) supplemented with protease inhibitor cocktail and incubated on ice for 30 minutes. Supernatant was collected after centrifuging at 4000 × g at 4ºC for 15 minutes. Protein lysate was aliquotted and stored at -80ºC. Protein concentration was measured using the BCA protein assay kit (Thermo Fisher Scientific) as per manufacturer’s protocol. Briefly, 2.5 μL of protein lysate was diluted 1:10 in RIPA buffer in a 96-well plate. Protein concentrations are estimated by reference to the absorbance readings obtained for a series of standard protein dilutions, which are assayed alongside the unknown samples. Bovine serum albumin (BSA) protein standards from 0.125 mg/mL to 2 mg/mL were prepared. After addition of 200 μL bicinchoninic acid (BCA) solution to each sample including the standards, the plate was incubated at 37ºC for 30 minutes in the dark and the absorbance was read at 570 nm on a Microplate Reader III (Bio-Rad Laboratories). To estimate protein concentrations, the average 570 nm measurements for the blank replicates were subtracted from all standard and unknown sample readings. A standard curve was then prepared by plotting the average 570 nm measurements for each BSA standard versus its concentration in μg/ml and this standard curve was used to determine the protein concentration of all unknown samples.

2.2.5.2 Protein gel electrophoresis Each protein sample (40 μg) was mixed with NuPAGE® LDS Sample Buffer (Thermo Fisher Scientific), NuPAGE® Sample Reducing Agent (Thermo Fisher Scientific) and milliQ water to make up a final volume of 15 μL. Protein samples were then heat inactivated at 95ºC for 5 minutes. Samples and 8 μL of Precision Plus Protein Kaleidoscope standard (Bio-Rad Laboratories) were then loaded onto a 10% precast

polyacrylamide gel. Electrophoresis was performed at constant 100 V for 90 minutes in 1X running buffer (described in Table 2.2).

2.2.5.3 Western blotting Following gel electrophoresis, polyvinylidene difluoride membranes (PVDF) were used to capture the proteins as they were transferred via wet electro-blotting using the Mini- Protean 3 wet transfer unit (Bio-Rad Laboratories). Proteins were transferred from the gel to the PVDF membrane in ice-cold transfer buffer (described in Table 2.2) at constant 200 mA for 2 hours. Uniform protein transfer was confirmed by Ponceau S staining. The membrane was then blocked with 10% non-fat dietary milk (NFDM) in TBS containing 0.05% Tween 20 (TBST) for 1 hour at room temperature. The membrane was subsequently incubated with primary antibody with gentle agitation according to the conditions detailed in Table 2.6. The membrane was then washed three times for 10 minutes in TBST followed by incubation with secondary antibody (detailed in Table 2.7) diluted in TBST for 1 hour at room temperature and finally the membrane was washed three times for 10 minutes in TBST. The horseradish peroxidise activity of the secondary antibodies were detected using the SuperSignal West Dura Chemiluminescent Substrate according to the manufacturer’s instructions and visualised on Super HR-G 30 autoradiography film.

2.2.6 Functionality assays

2.2.6.1 Apoptosis assay To analyse the proportion of dead cells and cells undergoing apoptotic cell death, cells were double stained with annexin V and 7-amino-actinomycin D (7-AAD). Annexin V binds to phosphatidylserine which is exposed on the outer leaflet of the plasma membrane during early apoptosis. 7-AAD stains for late stage apoptotic cells by

binding to DNA due to the passage of 7-AAD into the nucleus, whereas 7-AAD is excluded by intact cells.

Cells (0.5 – 1 × 106) were harvested and washed twice with PBS. Subsequently, cells were washed twice with annexin V binding buffer, resuspended in 100 µL annexin V binding buffer containing 2 µL annexin V APC, and incubated at room temperature in the dark for 10 minutes. A 400 µL volume of annexin V binding buffer containing 4 µL 7-AAD was then added to the sample. Following 15 minutes incubation at 4°C protected from light, cells were analysed by flow cytometry with an emission wavelength of 670±14 nm for annexin V and 710±50 nm for 7-AAD.

2.2.6.2 Cytospin and Wright Giemsa Staining Cells were washed and resuspended in 200 μL PBS at a density of 5 x 104 cells/mL. The cell suspension was centrifuged onto a microscope slide at 500 × g for 10 minutes at room temperature using a cytofunnel. Slides were then air dried and fixed with 100% methanol for 2 minutes.

Wright and Giemsa stains are histological stains that facilitate the differentiation of blood cell types. The main components are eosin (red) and methylene blue dyes. For Wright Giemsa staining, slides were stained with Wright stain for 30 minutes, followed by wright stain diluted 1:10 in phosphate buffer for 1 hour, Giemsa stain diluted 1:10 in phosphate buffer for 30 minutes. The slides were subsequently washed with phosphate buffer and submerged in the buffer for 5 minutes. The slides were then air dried and mounted before examining by light microscopy.

2.2.6.3 Colony-forming assay LSCs and pre-LSCs were seeded at 1000 cells per 35-mm dish in M3234/19% IMDM supplemented with 50 ng/mL IL-3. After 6 days incubation at 37°C/5% CO2, the total number of colonies were counted using an inverted light microscope. Cells were then

harvested for cell counting using a haemocytometer. For colony-forming assays with inhibitor and/or ligand treatment, the same procedures were followed as above with addition of the inhibitor and/or ligand at the time of seeding.

For HSCs, cells were seeded at 20,000 cells per 35-mm dish in M3234/19% IMDM supplemented with 50 ng/mL IL-3, 50 ng/mL SCF and 50 ng/mL IL-6. Colony and cell count were performed as described above.

2.2.6.4 Adenosine-5’-triphosphate (ATP) assay Cellular adenosine-5’-triphosphate (ATP) concentrations were measured using fluorometric ATP assay kit (Abcam) as per manufacturer’s protocol. For sample preparation, 1 × 106 cells were homogenized in ATP assay buffer supplied in the kit and supernatant was collected following centrifugation at 13000 × g for 2 minutes. Samples were then deproteinised by adding perchloric acid (PCA) to a final concentration of 1 M followed by centrifugation at 13000 × g for 2 minutes. Supernatant was collected and neutralised by adding potassium hydroxide (KOH) to precipitate excess PCA. The pH was adjusted to the range of 6.5 – 8. Samples were then centrifuged at 13000 × g for 15 minutes and supernatant was collected for immediate use in ATP assay.

For ATP assay, 50 μL samples were mixed with 50 μL reaction mixture supplied in the kit using 96-well format. ATP concentration standards from 0.2 nmol/well to 1 nmol/well were prepared alongside the unknown samples. Plates were incubated at room temperature for 30 minutes protected from light. Fluorescent signal was measured on a SpectraMax M5 microplate reader (Molecular Devices) using wavelength of 535 nm for excitation and 587 nm for emission. ATP concentrations were calculated by reference to the ATP concentration standards.

2.2.6.5 Oxygen consumption assay The oxygen consumption rate (OCR) was measured using the fluorescent oxygen probe MitoXpress® Xtra (Luxcel Biosciences). MitoXpress® Xtra is quenched by oxygen, through molecular collision, and thus the amount of fluorescence signal is inversely proportional to the amount of extracellular oxygen in the sample. Cells were seeded in 96-well plate at a density of 0.5 × 105 cells per well in 150 μL IMDM/15%FCS supplemented with 50 ng/mL IL-3. Subsequently, 10 μL reconstituted MitoXpress® reagent was added to each well except wells for use as blank control. 10 μL IMDM/15%FCS was added to the blank control wells. Each well was then sealed by adding two drops of pre-warmed HS mineral oil. The plate was immediately read on a SpectraMax M5 microplate reader (Molecular Devices) using wavelength of 380 nm for excitation and 650 nm for emission. Fluorescence intensity was measured every 5 minutes for 120 minutes at 37°C. To calculate the OCR, linear regression was applied to determine the slope of intensity versus time.

2.2.6.6 Detection of mitochondrial reactive oxygen species The production of mitochondrial reactive oxygen species (ROS) was detected using the fluorogenic probe MitoSOXTM Red. MitoSOXTM Red is a chemical that selectively targets mitochondria and is rapidly oxidised by superoxide. The oxidation of MitoSOXTM Red produces red fluorescence which can be measured by flow cytometric analysis. MitoSOXTM Red stock solution was made up to 2.5 mM stock with DMSO, and a concentration of 2.5 μM was used for staining. Cells were washed twice with PBS and resuspended in 100 µL MitoSOXTM Red solution. Following 30 minutes incubation at 37°C in the dark, cells were washed with PBS, then analysed by flow cytometry with an emission wavelength of 585±15 nm.

2.2.7 In vivo studies

2.2.7.1 Animals Six- to 8-week-old female C57BL/6 mice (Australian BioResources, Mossvale, NSW, Australia) were used as bone marrow donors and recipients. Six mice were placed in ventirack cages and housed in a pathogen-free environment for one week prior to initiation of experiments. All in vivo experimental studies used in this thesis were approved by the Animal Care and Ethics Committee of the University of New South Wales (ethics number ACEC 11/124B; NLRD 14-06).

2.2.7.2 In vivo transplantation Prior to transplantation, mice were sublethally irradiated at a single dose of 680 cGy using X-RAD 320 irradiator (Precision X-Ray, North Branford, USA). Mice were then heated using an infrared lamp to increase vein size. Primary transplantation was performed by intravenous injection of 1 × 106 pre-LSCs into the lateral tail vein of sublethally irradiated mice. Secondary transplantation was performed by intravenous injection of 1 × 105 GFP+ leukemic cells from primary recipients into sublethally irradiated mice. Mice were monitored daily for one week and fortnightly thereafter. Conditions under which mice were euthanised include hunched posture, lethargy, hind leg paralysis and >20% weight loss.

2.2.7.3 Isolation of bone marrow and spleen cells

Mice were euthanised by CO2 inhalation. An incision was made through the mid abdominal skin and the spleen was harvested and placed in phosphate buffered saline (PBS). The spleen cells were harvested by using mortar and pestle. The hind limbs were dissected from the body at the hip joint followed by removing of all muscles and connective tissue. Both femurs were snapped in half and bone marrow was flushed out with PBS using a syringe. Bone marrow and spleen cells were washed with PBS and

incubated with red cell lysis buffer (BD Pharm LyseTM) pre-diluted 10 times with milliQ H2O. Cells were then washed with PBS and passed through a 40 μM cell strainer to obtain single-cell suspensions for immediate flow cytometric analysis and cell freezing.

2.2.7.4 In vivo BrdU staining In vivo cell proliferation was assessed by bromodeoxyuridine (BrdU) staining. Mice were inoculated with 100 μL 10 mg/mL BrdU solution by intra peritoneal (IP) injection. After two hours BrdU has successfully penetrated into the bone marrow and mice were then euthanised. Bone marrow cells were harvested as described in Section 2.2.7.3 and APC BrdU flow kit was used according to manufacturer’s instructions. Briefly, cells were resuspended in 100 μL BD Cytofix/Cytoperm Buffer per sample and incubated for 20 minutes at room temperature. Samples were washed with 1 mL BD Perm/Wash Buffer and centrifuged at 500 × g for 8 minutes at room temperature. Cells were then resuspended in 100 μL BD Cytoperm Permeabilization Buffer Plus and incubated on ice for 10 minutes. Samples were then washed with 1 mL BD Perm/Wash Buffer, centrifuged at 500 × g for 8 minutes at room temperature and re-fixed in 100 μL BD Cytofix/Cytoperm Buffer for 5 minutes. An additional wash with 1 mL BD Perm/Wash Buffer was performed prior to adding 100 μL of 300 μg/mL DNase. Cells were incubated at 37°C for 1 hour. After washing with 1 mL BD Perm/Wash Buffer, samples were incubated in 50 μL BD Perm/Wash Buffer containing 1 μL APC-anti-BrdU antibody for 20 minutes before analysing by flow cytometry with an emission wavelength of 670±14 nM.

2.2.8 Statistical analysis Statistical analysis was performed using GraphPad Prism 5 software (GraphPad Software). Mouse survival was graphed as survival curves by Kaplan–Meier analysis. Log-rank test was used to determine whether the difference in survival curves were

statistically significant (p<0.05). Gene set enrichment analysis (GSEA) was performed to compare our microarray data to previously published Wnt/β-catenin target genes. L- CalcTM software (StemCell Technologies) was used to assess in vivo limiting dilution. For comparing two groups, two-sided Student's t-test was used and for comparing three or more groups, one-way ANOVA was used where a p value of less than 0.05 was considered statistically significant. In all cases error bars are representative of the standard error of the mean (SEM) of three biological repeat experiments unless otherwise stated.

3 The role of Lgr4 in AML development

3.1 Introduction AML is one of the most common leukaemias both in children and adults. The 5-year survival rate in Australia is only 24% [2]. The poor survival is mainly attributed to high relapse rate. We have previously demonstrated that the Wnt/β-catenin signalling pathway plays a critical role in the establishment of AML LSCs [104], which are believed to be responsible for AML relapse [104,105,234]. Our previous studies have shown that aberrant activation of β-catenin signalling is essential for the transformation of normal HSCs and GMPs into LSCs in MLL-AF9 and Hoxa9/Meis1a-induced AML [104]. However, the aberrant activation of β-catenin is not accompanied by upregulation or mutations of previously known Wnt pathway components [104], suggesting alternative upstream regulators of β-catenin.

Recent studies have demonstrated that R-spondins (Rspo1-4) are potent enhancers of Wnt/β-catenin signalling in embryonic development and stem cell growth [235-239]. Recurrent gene fusions involving Rspo2 and Rspo3 have been recently discovered in approximately 10% of human colon cancers, and they are mutually exclusive with mutations in APC [217], which is a key component of the destruction complex that degrades β-catenin. These findings suggest a possible role of Rspo in Wnt/β-catenin activation in cancer.

Rspos have been recently identified as ligands of Lgr4 [210,214-216]. Lgr4 has been reported to play a crucial role in the survival and self-renewal of normal colon stem cells [210,211]. Additionally, Lgr4 is a poor prognostic factor that promotes metastasis and activates β-catenin signalling in colon cancer [218]. The observation that Lgr4 is often expressed at high levels in cancers containing Rspo fusions suggests that aberrant Rspo-Lgr4 signalling may serve as a driving mechanism in colon tumourigenesis. The role of Rspo-Lgr4 in leukaemogenesis has not yet been investigated. Given the crucial role that Rpso plays in enhancing Wnt signals, we sought to determine if Rspo is a positive mediator of β-catenin signalling in leukaemia. This chapter aims to examine the role of Rspo-Lgr4 in LSCs and AML development.

3.2 Lgr4 is significantly overexpressed in human MLL-AML patients To study the role of Lgr4 in MLL-AF9 translocated AML (MLL-AML), We first examined the gene expression profiles of Lgr4 in HemaExplorer (http://servers.binf.ku.dk/hemaexplorer), a curated database of human mRNA expression profiles in both normal and malignant haematopoiesis [240]. Lgr4 expression was significantly upregulated in MLL-AML cells compared to normal HSCs (n=8 for normal HSCs and n=38 for MLL-AML; p<0.0001; Figure 3.1A). The expression of β-catenin was also compared in these cells and the β-catenin mRNA levels were significantly increased in MLL-AML cells (p<0.0001; Figure 3.1B). The increased expression of β-catenin in MLL-AML is consistent with previous findings that Wnt/β-catenin signalling is a pathway essential for AML development [104,241,242].

Figure 3.1 Lgr4 and β-catenin are upregulated in human MLL-AML patients

Gene expression data were obtained from the HemaExplorer database (http://servers.binf.ku.dk/hemaexplorer, [240]). Relative mRNA expression of (A) Lgr4 and (B) β-catenin was compared between human HSCs (n=8) and MLL-AML cells (n=38).

3.3 High Lgr4 expression is associated with poor patient survival To further address the role of Lgr4 in AML, a microarray experiment was performed in collaboration with Prof Richard D'Andrea (University of South Australia) to analyse the gene expression profiles of bone marrow samples from a mixed cohort of 104 AML patients. Cox regression analysis of event-free survival, defined as the duration from start of the treatment to disease progression or death, showed that high Lgr4 expression was associated with an unfavourable outcome in AML patients with a median follow-up of 142 days (p=0.00844; hazard ratio [HR], 1.89907; 95% confidence interval [CI], 1.17 - 3.084, Figure 3.2A). In addition, multivariate analysis for this group of patients showed that Lgr4 is predictive of event-free survival independently of other predictors such as age and FLT3/NPM1 risk classification. Furthermore, high Lgr4 expression is associated with poor overall survival in patients with abnormal or complex karyotype (p=0.006). This finding indicates the potential clinical value of Lgr4 in AML. In contrast, expression of Lgr5, a close homolog of Lgr4, showed no significant association with clinical outcome in our patient cohort (p=0.85783, Figure 3.2B).

Figure 3.2 High Lgr4 expression is associated with poor clinical outcome in AML patients

(A) Cox regression analysis of event-free survival shows an inferior outcome for patients with high Lgr4 expression (shown in red) in comparison with patients with low Lgr4 expression (shown in green) in a total cohort of 104 AML patients. Median follow-up is 142 days for Lgr4 high group and 283 days for Lgr4 low group. (B) Cox regression analysis of event-free survival shows no significant difference in patient outcome between high Lgr5 expression and low Lgr5 expression. Lgr4 expression levels in all patient samples were ranked and the upper-quartile was defined as Lgr4- high and the rest was defined as Lgr4-low. The exact cut-off value is 6.808097 in log2 scale, and 112.0576 in raw scale.

Microarray data were kindly provided by Prof Richard D'Andrea and Cox regression analysis was performed by Dr Bing Liu.

3.4 Lgr4 is significantly upregulated in murine MLL-AML LSCs compared to normal HSCs To investigate the expression of Lgr4 in our well-characterised mouse AML model [100,104,105], western blot analysis was performed on mouse LSCs derived from MLL-AF9-transduced c-Kit+Lin−Sca-1+ (KLSMLL) and MLL-AF9-transduced c- Kit+Lin−Sca-1-CD34+FcRγ+ (GMPMLL) vs mouse HSC-enriched c-Kit+Lin−Sca-1+ (KLS) cells as described previously [100,104]. Figure 3.3 shows that Lgr4 expression was substantially increased in LSCs compared to KLS cells. Furthermore, the level of β-catenin was concordantly upregulated in these LSCs. This indicates a positive correlation between Lgr4 and β-catenin expression in MLL-AML.

Figure 3.3 Lgr4 and β-catenin are upregulated in mouse MLL-AML

Mouse LSCs derived from MLL-AF9-transduced c-Kit+Lin−Sca-1+ (KLSMLL) and MLL-AF9-transduced c-Kit+Lin−Sca-1-CD34+FcRγ+ (GMPMLL) and mouse HSCs enriched c-Kit+Lin−Sca-1+ (KLS) cells were harvested for lysates one week after transduction. Lysates were analysed by western blotting for expression of Lgr4 and β- catenin. Actin was used as loading control.

3.5 Ligand Rspo or Wnt3a alone is not capable of enhancing β-catenin expression in MLL-AML Rspo1-4 are a family of secreted proteins that potentiate Wnt/β-catenin signalling in normal cell regulation [237]. They play important roles in embryonic development and are potent stem cell growth factors [243]. The mechanism of Rspo signalling is poorly understood to date; however, it was recently discovered that Rspo function as ligands of the Lgr4 receptor to activate β-catenin signalling in HEK293 cells [210,214-216]. These studies showed that Rspo1-4 directly bind to Lgr4 and activate β-catenin signalling. The activity of the Rspo ligands has not been investigated in leukaemia to date. We demonstrated that Lgr4 was highly expressed in AML LSCs, we therefore sought to determine if Rspo ligands play a role in the activation of β-catenin in MLL driven leukaemogenesis.

An interesting study by Carmon et al. has demonstrated that Rspo1 and Rspo2 have higher binding affinity and potency for Lgr4 than Rspo3. Rspo4 has the lowest ligand binding affinity for Lgr4 [214]. Therefore, we began our study by investigating the activity of Rspo ligands in KLSMLL pre-LSCs. Cells were individually incubated with Rspo1-4 (50 – 400 ng/ml) for 24 hours followed by western blot analysis of β-catenin expression. As illustrated in Figure 3.4A-D, none of the Rspo ligands enhanced β- catenin expression in KLSMLL pre-LSCs.

Previous studies have demonstrated that the maximum level of β-catenin activation, caused by Rspo ligands, is dependent on the concentration of Wnt3a [214]. We then tested whether Wnt3a alone could enhance β-catenin expression in KLSMLL pre-LSCs. Cells were stimulated with Wnt3a conditioned medium for 24 hours, as described in Section 2.2.1.2. The cells were then harvested and interrogated by western blot analysis for β-catenin expression. An increase in β-catenin expression was not observed in KLSMLL pre-LSCs (Figure 3.4E).

A Rspo1 0 so 100 200 400 ng/ml

~ - eaten in

Actin

B Rspo2 0 so 100 200 400 ng/ml

~ - caten i n

Actin c Rspo3 0 so 100 200 400 ng/ml

~-cate n i n

Actin

D Rspo4 0 so 100 200 400 ng/ml

~-caten i n I ' '• 'A •' ',.. ,Il

Actin

E Control Wnt3a

~ - caten i n

Actin

Control Wnt3a

Figure 3.4 Rspo1-4 or Wnt3 alone cannot activate β-catenin in KLSMLL pre-LSCs

(A-D) KLSMLL pre-LSCs were incubated with Rspo1, 2, 3 or 4 (50 – 400 ng/mL) independently in IMDM/15%FCS medium for 24 hours. Lysates were analysed by western blotting for expression of β-catenin. (E) KLSMLL pre-LSCs were incubated in Wnt3a conditioned medium for 24 hours. Lysates were analysed by western blotting for expression of β-catenin. Actin was used as loading control. Graph represents 3 independent experiments analysed by ImageJ.

3.6 Rspo2 and Rspo3 synergise with Wnt3a to enhance β-catenin activation in pre-LSCs Although Rspo ligands and Wnt3a failed to activate β-catenin independently in KLSMLL pre-LSCs, we hypothesised that these proteins may be needed in combination in order to enhance β-catenin expression. The combination of Wnt3a and Rspo has been shown to enhance β-catenin activity more than 10-fold compared to Wnt3a or Rspo alone in HEK293 cells [210], indicating a synergistic mechanism of action. To examine whether Wnt3a and Rspo synergise in KLSMLL pre-LSCs to activate β-catenin, cells were incubated with Rspo1-4 in combination with Wnt3a conditioned medium for 24 hours followed by western blot analysis of β-catenin expression. Interestingly, β-catenin expression was exclusively upregulated when Wnt3a was combined with Rspo2 or Rspo3, but not with Rspo1 or Rspo4 (Figure 3.5A and B). The level of the inactive form of β-catenin (phospho-β-catenin) was not changed (Figure 3.5A), suggesting an increase in the level of active β-catenin expression. This finding suggests that the synergistic action of Rspo2/Rspo3 with Wnt3a might play a crucial role in enhancing β-catenin expression in pre-LSCs thereby promoting AML disease progression.

Figure 3.5 Rspo2 and Rspo3 synergise with Wnt3a to activate β-catenin in KLSMLL pre-LSCs

(A) KLSMLL pre-LSCs were incubated with Rspo1, 2 or 3 in combination with Wnt3a conditioned medium for 24 hours. Lysates were analysed by western blotting for expression of β-catenin and phospho-β-catenin. (B) KLSMLL pre-LSCs were incubated with Rspo4 in combination with Wnt3a conditioned medium for 24 hours. Lysates were analysed by western blotting for expression of β-catenin.

3.7 Co-treatment of Wnt3a and Rspo3 enhances the proliferation ability of KLSMLL pre-LSCs As described in the previous section, Rspo2/Rspo3 synergize with Wnt3a to activate β- catenin in KLSMLL pre-LSCs. Since β-catenin has been shown to be a driver of LSC function [104,105], we sought to examine whether the increase of β-catenin caused by co-treatment of Wnt3a and Rspo3 could lead to enhanced proliferation in KLSMLL pre- LSCs. To address this, colony forming assays were carried out in the presence of Wnt3a alone, Rspo3 alone, or a combination of Wnt3a and Rspo3. Colony forming assay is an in vitro cell proliferation assay based on the ability of a single cell to grow into a colony. With respect to leukaemic cells, it essentially tests every cell in a population for its ability to divide indefinitely. Briefly, KLSMLL pre-LSCs were seeded at a density of 1000 cells per culture dish and incubated with Rspo3, Wnt3a or their combination in methylcellulose medium for 6 days. Figure 3.6A and B show that single treatment of Wnt3a or Rspo3 did not enhance the colony forming or cell proliferation ability, compared to the control treated cells, consistent with our findings that single treatment of Wnt3a or Rspo3 cannot activate β-catenin. In contrast, incubating cells with both Wnt3a and Rspo3 induced a significant increase in both cell number (p<0.0001) and the number of colonies formed (p=0.0292) after six days incubation (Figure 3.6). This indicates that co-treatment of Wnt3a and Rspo3 is required to enhance the proliferative abilities of KLSMLL pre-LSCs mediated by β-catenin activation.

Figure 3.6 Co-treatment of Wnt3a and Rspo3 enhances colony formation of KLSMLL pre-LSCs

KLSMLL pre-LSCs were seeded at a density of 1000 cells per culture dish and incubated with Rspo3, Wnt3a or their combination in methylcellulose medium. The number of colonies (A) and the number of cells (B) were counted after 6 days incubation. Results from three independent experiments were graphed as mean ± standard error of the mean (SEM).

3.8 Lgr4 knockdown in KLSMLL pre-LSCs suppresses β-catenin expression To examine the functional effect of Lgr4 in AML LSCs, we first performed shRNA- mediated stable knockdown of Lgr4 (described in Section 2.2.2). Briefly, KLSMLL pre-LSCs were transduced with either one of four different shRNAs against Lgr4 or a non-targeting scrambled shRNA (Scr). Following lentiviral transductions of the shRNAs, the knockdown efficiency of Lgr4 shRNAs was confirmed by western blot. Figure 3.7A shows that Lgr4 expression was substantially reduced with shRNA#3 and shRNA#4 compared to Scr control cells. Lgr4 shRNA#1 and #2 had no effects on Lgr4 expression and were not used for any further experimentation. As expected, cells transduced with Lgr4 shRNA#3 and #4 displayed significant lower β-catenin levels compared to Scr control (Figure 3.7A lane 1, 4 and 5). Additionally, cells transduced with Lgr4 shRNA#1 or #2, which did not have Lgr4 knockdown, showed no discernible difference in β-catenin expression (Figure 3.7A land 1, 2 and 3). This suggests that Lgr4 positively regulates β-catenin expression.

As described in Section 1.3.3, Lgr5 is a close homologue of Lgr4 that shares the same ligands (Rspo1-4) and previous studies have demonstrated both Lgr4 and Lgr5 provide Wnt signals in intestinal stem cells [210,214,215], suggesting a possible functional redundancy between these two receptors. Hence we investigated whether Lgr4 knockdown is counteracted by increased levels of Lgr5. Interestingly, Figure 3.7B shows that Lgr4 knockdown by shRNA#3 or shRNA#4 did not induce upregulation of Lgr5 and the expression of Lgr5 was similar compared to control. Therefore, Lgr5 does not compensate for the loss of Lgr4 in KLSMLL pre-LSCs.

Figure 3.7 Lgr4 knockdown in KLSMLL pre-LSCs suppresses β-catenin expression

(A) KLSMLL pre-LSCs were transduced with four different shRNAs against Lgr4 and a Scr control. Cells were harvested two weeks after transduction. Lysates were analysed by western blotting for expression of Lgr4 and β-catenin. (B) KLSMLL pre-LSCs expressing Lgr4 shRNA#3, #4 or Scr control were harvested when appropriate level of confluence was achieved. Lysates were analysed by western blotting for expression of Lgr4 and Lgr5.

3.9 Lgr4 knockdown impairs cell proliferation of KLSMLL pre-LSCs in vitro Since β-catenin plays a key role in driving cell proliferation of LSCs in AML [104,105,244], and we showed that Lgr4 knockdown reduced expression levels of β- catenin, we next examined whether ablation of Lgr4 causes any functional abnormality, in line with a β-catenin deficient phenotype. Colony forming assays were carried out to examine the cell proliferation potential of Lgr4 deficient KLSMLL pre-LSCs. Figure 3.8A and B show that Lgr4 knockdown induced by shRNA#3 or #4 significantly decreased the colony numbers (p=0.0005 for shRNA#3; p=0.0006 for shRNA#4) and cell numbers (p=0.0008 for shRNA#3; p=0.0009 for shRNA#4) compared to Scr. The impaired phenotype was also reflected by the smaller sizes of the colonies in Lgr4 shRNA#3 and #4 compared to Scr (Figure 3.8C). Hence Lgr4 knockdown impaired the cell proliferation ability of KLSMLL pre-LSCs in vitro.

Figure 3.8 Lgr4 knockdown suppresses colony formation of KLSMLL pre-LSCs

KLSMLL pre-LSCs expressing Lgr4 shRNA#3, #4 or Scr control were seeded at a density of 1000 cells per culture dish and incubated in methylcellulose medium. The number of colonies (A) and the number of cells (B) were counted after 6 days incubation. Results from three independent experiments were graphed as mean ± SEM. (C) Colony morphology of KLSMLL pre-LSCs expressing Lgr4 shRNA#3, #4 or Scr control was analysed by bright-field microscopy (10x magnification).

3.10 Ectopic overexpression of β-catenin rescues the Lgr4-deficient phenotype To further investigate the functional relationship between Lgr4 and β-catenin, we sought to determine whether the Lgr4-deficient phenotype could be rescued by overexpressing a construct encoding a constitutively active form of β-catenin (β-cat*) in Lgr4 deficient cells. Accordingly, KLSMLL pre-LSCs expressing Lgr4 shRNA#3 or #4 were transduced with β-cat* and western blot analysis was conducted. Re-expression of β-catenin was observed in both Lgr4 knockdown cell lines compared to previous figures (Figure 3.9A and Figure 3.7). Furthermore, the expression levels of β-catenin in the β- cat* rescued cells were similar to that of empty vector (EV) (Figure 3.9A). Subsequent colony forming assays of the β-cat* rescued cells showed that β-cat* expression in Lgr4 deficient cells enhanced their colony formation, as there was no significant difference in the β-cat* rescued cells compared to Scr+EV cells (Figure 3.9B). Thus, ectopic overexpression of β-cat* rescued the Lgr4-deficient phenotype of KLSMLL pre-LSCs in vitro.

Figure 3.9 Ectopic overexpression of β-catenin rescues the Lgr4-deficient phenotype

(A) KLSMLL pre-LSCs expressing Lgr4 shRNA#3 or #4 were transduced with constitutively active β-catenin (β-cat*). Lysate were analysed by western blotting for expression β-catenin. (B) KLSMLL pre-LSCs expressing Scr+EV, Lgr4 shRNA#3+β- cat*, or Lgr4 shRNA#4+β-cat* were seeded at a density of 1000 cells per culture dish and incubated in methylcellulose medium. The number of colonies was counted after 6 days incubation. Results from three independent experiments were graphed as mean ± SEM.

3.11 Lgr4 knockdown in KLSMLL pre-LSCs impairs AML initiation and maintenance in vivo Our previous studies have demonstrated that β-catenin is required for the development of AML in vivo [104]. Given that Lgr4 is a positive regulator of β-catenin, we sought to determine whether Lgr4 is essential for leukaemogenesis in vivo. To this end, we evaluated the role of Lgr4 in leukaemia initiation by inoculating KLSMLL pre-LSCs expressing Lgr4 shRNA#3 and #4 into sublethally irradiated C57BL/6 mice. Cells were injected into the lateral tail vein and the mice were monitored and sacrificed upon displaying symptoms of leukaemia (described in detail in Section 2.2.7). To determine the organ-invasive capacity or aggressiveness of pre-LSCs expressing Lgr4 shRNA, spleen weight was measured in addition to bone marrow and spleen cells being harvested for flow cytometric analysis at time of sacrifice. Bone marrow and spleen are two major sites of infiltration during disease progression and the percentage of leukaemic cells present in these sites is indicative of the extent of disease burden. The pre-LSCs carry a GFP marker cassette and therefore the extent of disease burden can be measured by analysing the percentage of GFP present in the bone marrow and spleen using flow cytometric analysis.

Strikingly, only 2 out of 13 mice that transplanted with KLSMLL pre-LSCs expressing Lgr4 shRNA#3 and 4 out of 13 mice that transplanted with KLSMLL pre-LSCs expressing Lgr4 shRNA#4 developed AML while all control mice succumbed to short latency primary AML (Figure 3.10A). At time of sacrifice, flow cytometric analysis revealed that there were less GFP+ leukaemic cells in the bone marrow (Figure 3.10B) and spleen (Figure 3.10C) of mice inoculated with KLSMLL pre-LSCs expressing Lgr4 shRNA#3 or #4 compared to Scr control mice. Due to the extremely low frequency of disease onset in the Lgr4 shRNA groups, tissue could only be harvested from two mice in each of these groups. The sample size is too small for any valid statistical conclusion on infiltration; nevertheless, the decrease in GFP+ leukaemic cells was apparent. Figure 3.10D shows that mice inoculated with KLSMLL pre-LSCs expressing Lgr4 shRNA#3 or #4 give rise to lower spleen weight upon sacrifice. No statistical test was carried out

for shRNA#3 due to small sample size of sick mice. Together, these results indicate that Lgr4 plays an important role in AML initiation.

Another important factor in determining AML burden is disease maintenance. AML maintenance is the ability of the already developed LSCs, after primary passage in mice, to give rise to AML in secondary transplantations. To investigate whether Lgr4 is involved in regulating AML maintenance, GFP+ leukaemic cells, harvested from primary recipients, were flow sorted and inoculated into secondary recipient mice. Remarkably, Lgr4 deficient leukaemic cells completely lost their ability to develop secondary AML, whereas Scr control cells gave rise to AML with a shorter latency in secondary recipients than in primary recipients (Figure 3.10E). Therefore, Lgr4 is essential in AML maintenance in vivo.

Figure 3.10 Lgr4 knockdown impairs KLSMLL induced AML initiation and maintenance

(A) Kaplan–Meier survival analysis of mice transplanted with KLSMLL pre-LSCs expressing Scr, Lgr4 shRNA#3 or #4. At time of sacrifice, the percentage of GFP+ leukaemic cells in the (B) bone marrow and (C) spleen were analysed by flow cytometry. (D) Spleen weight was measured upon sacrifice. (E) Kaplan–Meier survival analysis of mice transplanted with sorted GFP+ mouse bone marrow cells derived from (A). Each data point represents one mouse. Horizontal lines represent the median.

3.12 Lgr4 knockdown in GMPMLL pre-LSCs suppresses β-catenin expression and impairs cell proliferation in vitro The previous section has demonstrated that Lgr4 plays a crucial role in KLSMLL pre- LSCs driven AML initiation and maintenance. AML is a heterogeneous disease and identifying the cell of origin in AML has become one of the most discussed topics in AML stem cell biology. Previous reports have shown that the MLL-AF9 fusion oncogene can also impart self-renewal and leukaemogenic capacity on haematopoietic progenitor cells such as GMPs [100,245]. Therefore, we sought to investigate whether Lgr4 also plays a role in GMPMLL-driven leukaemogenesis.

To examine whether Lgr4 regulates β-catenin in GMPMLL pre-LSCs, we performed shRNA-mediated stable knockdown of Lgr4 and examined the effect of Lgr4 knockdown on the expression of β-catenin. Figure 3.11A shows that only GMPMLL pre-LSCs expressing Lgr4 shRNA#3 exhibited a reduction of Lgr4 and β-catenin expression. The reduction of β-catenin in concert with Lgr4 knockdown by shRNA#3 suggests that Lgr4 also plays a role in β-catenin regulation in GMPMLL pre-LSCs.

Since Lgr4 deficiency in KLSMLL pre-LSCs impairs cell proliferative potential in vitro, we sought to determine whether Lgr4 knockdown by shRNA#3 in GMPMLL pre- LSCs could also affect cell proliferation. Colony forming ability was assessed and as shown in Figure 3.11 knockdown of Lgr4 significantly decreased the number of colonies formed (p=0.0013; Figure 3.11B) and cell number (p=0.044; Figure 3.11C) in GMPMLL pre-LSCs. Together these results indicate that Lgr4 regulates β-catenin in GMPMLL pre-LSCs and Lgr4 knockdown impairs cell proliferation of GMPMLL pre- LSCs.

Figure 3.11 Knockdown of Lgr4 in GMPMLL pre-LSCs suppresses β-catenin expression and colony formation

(A) GMPMLL pre-LSCs were transduced with Lgr4 shRNA#3, #4 or Scr control. Cells were harvested two weeks after transduction. Lysates were analysed by western blotting for expression of Lgr4 and β-catenin. (B and C) GMPMLL pre-LSCs expressing Lgr4 shRNA#3, or Scr control were seeded at a density of 1000 cells per culture dish and incubated in methylcellulose medium. The number of colonies (B) and the number of cells (C) were counted after 6 days. Results from three independent experiments were graphed as mean ± SEM.

3.13 Lgr4 knockdown in GMPMLL pre-LSCs impairs AML initiation and maintenance In addition to HSC enriched KLS cells (described in Section 3.4), more matured progenitor cells, GMPs, are able to be transformed into pre-LSCs, which then give rise to AML [100,245]. As outlined above, Lgr4 deficiency causes impaired GMPMLL pre- LSCs function in vitro. To investigate whether Lgr4 plays a role in the leukaemogenesis of GMPMLL pre-LSCs in vivo, we inoculated GMPMLL pre-LSCs expressing Lgr4 shRNA#3 or Scr into sublethally irradiated C57BL/6 mice. Figure 3.12A shows that only 3 out of 9 mice that transplanted with GMPMLL pre-LSCs expressing Lgr4 shRNA#3 developed AML, while 7 out of 9 mice that transplanted with Scr control cells developed AML with much shorter latency.

At time of sacrifice, flow cytometric analysis revealed that there were less GFP+ leukaemic cells in the bone marrow of mice inoculated with GMPMLL pre-LSCs expressing Lgr4 shRNA#3 compared to Scr control mice (Figure 3.12B). Due to the extremely low frequency of disease onset in the Lgr4 shRNA#3 group, tissue could only be harvested from two mice that eventually developed AML and the sample size was too small for any valid statistical conclusion. Flow cytometric analysis of spleen cells showed no observable difference between shRNA#3 and Scr transduced cells (Figure 3.12C). Spleen weight from Lgr4 shRNA#3 group showed no statistical difference compared to Scr control (p=0.5256; Figure 3.12D), indicating that the aggressiveness of GMPMLL pre-LSCs is less reliant on Lgr4 compared to KLSMLL pre-LSCs.

To further evaluate the role of Lgr4 in AML maintenance, GFP+ leukaemic cells, harvested from primary recipients, were flow sorted and inoculated into secondary recipient mice. Figure 3.12E shows that Lgr4 deficiency in GMPMLL pre-LSCs significantly delayed the onset of AML in secondary transplantation (p=0.027). Together these results indicate that Lgr4 plays an important role in AML initiation and maintenance in GMPMLL pre-LSCs.

100

101

Figure 3.12 Lgr4 knockdown impairs GMPMLL induced AML initiation and maintenance

(A) Kaplan–Meier survival analysis of mice transplanted with GMPMLL pre-LSCs expressing Scr or Lgr4 shRNA#3. At time of sacrifice, the percentage of GFP+ leukaemic cells in the (B) bone marrow and (C) spleen were analysed by flow cytometry. (D) Spleen weight was measured upon sacrifice. (E) Kaplan–Meier survival analysis of mice transplanted with sorted GFP+ mouse bone marrow cells derived from (A). Each data point represents one mouse. Horizontal lines represent the median.

102

3.14 Lgr4 overexpression enhances the proliferative potential of KLSA9M pre-LSC in vitro Previous studies have shown that high levels of β-catenin expression correlate with more aggressive disease progression and significantly poor clinical outcomes in AML [246,247]. In murine models of AML, MLL-AF9 transduced KLS cells (KLSMLL pre- LSCs) generate a more aggressive type of AML in vivo compared to Hoxa9/Meis1a transduced KLS cells (KLSA9M pre-LSCs) [104,245]. However, the levels of β-catenin expression have not been directly compared between these two cell types to date. To investigate whether the more aggressive KLSMLL pre-LSCs have higher β-catenin expression than KLSA9M pre-LSCs, western blot analysis was conducted using these two cells types. Western blot analysis showed that KLSMLL pre-LSCs had higher β- catenin expression compared to KLSA9M pre-LSCs (Figure 3.13A). We were interested to see whether the differences in β-catenin expression correlate with differential Lgr4 expression. Western blot analysis of KLSMLL and KLSA9M pre- LSCs revealed that Lgr4 is more highly expressed in KLSMLL pre-LSCs, correlating with the observed increase in β-catenin expression (Figure 3.13A).

Given that Lgr4 knockdown significantly impaired initiation and maintenance of MLL- AML, this led us to hypothesise that Lgr4 may be the key component of β-catenin signalling that drives AML aggressiveness. We next investigated whether overexpression of Lgr4 could enhance leukaemogenesis. KLSA9M pre-LSCs were used to overexpress Lgr4, because they have lower endogenous levels of Lgr4 and β-catenin compared to KLSMLL pre-LSCs. Briefly, KLSA9M pre-LSCs were retrovirally transduced with a Lgr4 cDNA construct or EV (described in detail in Section 2.2.2) and overexpression was confirmed by western blot analysis (Figure 3.13B). As expected, western blot analysis also revealed an upregulation of β-catenin in KLSA9M pre-LSCs expressing Lgr4 cDNA (Figure 3.13B), confirming that Lgr4 is a positive regulator of β-catenin. Furthermore, colony forming assays revealed that KLSA9M pre-LSCs expressing Lgr4 cDNA generated significantly more colonies (p=0.0224; Figure 3.13C)

103

and number of cells (p=0.0016; Figure 3.13D) compared to EV control. These results further support Lgr4 as a key player in leukaemic development.

104

Figure 3.13 Lgr4 overexpression enhances proliferative potential of KLSA9M pre- LSCs in vitro

(A) KLSA9M and KLSMLL pre-LSCs were harvest when appropriate level of confluence was achieved. Lysates were analysed by western blotting for expression of Lgr4 and β-catenin. (B) KLSA9M pre-LSCs were transduced with Lgr4 cDNA or empty vector (EV) control. Cells were harvested two weeks after transduction. Lysates were analysed by western blotting for expression of Lgr4 and β-catenin. (C and D) KLSA9M pre-LSCs expressing Lgr4 cDNA or EV control were seeded at a density of 1000 cells per culture dish and incubated in methylcellulose medium. The number of colonies (C) and the number of cells (D were counted after 6 days incubation. Results from three independent experiments were graphed as mean ± SEM.

105

3.15 Lgr4 overexpression in KLSA9M pre-LSCs exacerbates AML in vivo Next, we examined whether Lgr4 overexpression in KLSA9M pre-LSCs could lead to a more aggressive disease phenotype in vivo. To this end, KLSA9M pre-LSCs expressing Lgr4 cDNA or EV control were inoculated into sublethally irradiated C56BL/6 mice. AML disease onset was substantially shortened upon Lgr4 overexpression (p<0.0001; Figure 3.14A). KLSA9M pre-LSCs transduced with Lgr4 cDNA showed substantially more aggressive AML compared to EV. Interestingly, the time of disease onset was comparable to AML derived from KLSMLL pre-LSCs (Figure 3.14A). Flow cytometric analysis of bone marrow cells derived from mice upon sacrifice, showed that there was a significant increase of GFP+ leukaemic cells in the Lgr4 cDNA group compared to EV (p=0.0457; Figure 3.14B). In contrast, flow cytometric analysis of spleen cells revealed no significant difference between Lgr4 cDNA group and EV p=0.0894; Figure 3.14C). This is possibly due to the already high spleen infiltration in the EV group; however, a slightly higher average GFP percentage was observed in the Lgr4 cDNA group. These results together indicate that Lgr4 can significantly enhance the aggressiveness of KLSA9M pre-LSCs in vivo and further support Lgr4 as an essential player in AML development.

To investigate the effect of Lgr4 overexpression on the proliferative capacity of pre- LSCs in vivo, mouse bone marrow cells from Lgr4 cDNA and EV groups were harvested 30 days post-inoculation. Flow cytometric analysis revealed that there was a substantial increase in GFP+ population in the bone marrow of mice transplanted with KLSA9M pre-LSCs expressing Lgr4 cDNA compared to EV (p<0.0001; Figure 3.15A and B). This indicates that Lgr4 overexpressing cells have enhanced proliferation rates in the bone marrow thus leading to increased disease burden in mice. To examine whether Lgr4 also has an effect on fully developed LSCs, GFP+ bone marrow cells harvested from primary recipients were transplanted into secondary recipients. Secondary recipient mice were sacrificed 8 days after inoculation and flow cytometric analysis of bone marrow cells was performed. A significant increase of GFP+ leukaemic

106

cells was observed in bone marrow derived from the Lgr4 cDNA group compared to EV (p=0.0174; Figure 3.16A and B). Additionally, Figure 3.16C shows that the cells transduced with Lgr4 cDNA still retained Lgr4 and β-catenin overexpression as detected by western blot analysis following bone marrow harvest from secondary recipient mice.

Together these results indicate that increased levels of Lgr4 confer a growth advantage to KLSA9M pre-LSCs and LSCs, and significantly shorten AML onset.

107

Figure 3.14 Lgr4 overexpression in KLSA9M pre-LSCs exacerbates AML in vivo

(A) Kaplan–Meier survival analysis of mice transplanted with KLSA9M pre-LSCs expressing Lgr4 cDNA or EV control. At time of sacrifice, the percentage of GFP+ leukaemic cells in the (B) bone marrow and (C) spleen were analysed by flow cytometry. (D) Spleen weight was measured upon sacrifice. Each data point represents one mouse. Horizontal lines represent the median.

108

Figure 3.15 Overexpression of Lgr4 enhances in vivo proliferation of KLSA9M pre-LSCs Bone marrow cells of primary recipient mice were harvested 30 days after inoculation of KLSA9M pre-LSCs expressing Lgr4 cDNA or EV control. Cells were analysed by flow cytometry. (A) Representative flow scatter plots and (B) dot plot showing engraftment of KLSA9M pre-LSCs expressing Lgr4 cDNA or EV in mouse bone marrow. Each data point represents one mouse. Horizontal lines represent the median.

109

110

Figure 3.16 Overexpression of Lgr4 enhances in vivo proliferation of KLSA9M LSCs Bone marrow cells of secondary recipient mice were harvested 8 days after inoculation of KLSA9M LSCs expressing Lgr4 cDNA or EV control. Cells were analysed by flow cytometry. (A) Representative flow scatter plots and (B) dot plot showing engraftment of KLSA9M pre-LSCs expressing Lgr4 cDNA or EV in mouse bone marrow. (C) KLSA9M LSCs expressing Lgr4 cDNA or EV control were analysed by western blotting for expression of Lgr4 and β-catenin. Each data point represents one mouse. Horizontal lines represent the median.

111

3.16 Lgr4 overexpression cannot transform KLS cells So far we have shown that Lgr4 plays a critical role in LSC self-renewal and proliferation, and its expression drives the aggressiveness of AML. Next, we aimed to investigate whether overexpression of Lgr4 alone can transform or enhance the self- renewal ability of HSC enriched KLS cells. Following transduction of Lgr4 cDNA or EV into KLS cells, colony forming assays were conducted for four consecutive weeks. As shown in Figure 3.17A and B, Lgr4 cDNA did not cause a significant difference in colony numbers or cell numbers compared to EV transduced KLS cells. In addition, Figure 3.17C shows that the size and shape of the colonies in Lgr4 cDNA or EV transduced KLS cells were very similar. Furthermore, both Lgr4 cDNA and EV transduced KLS cells stopped proliferating after four weeks of in vitro culture, suggesting that Lgr4 itself cannot sustain self-renewal and inhibit differentiation of HSCs in vitro.

112

113

Figure 3.17 Lgr4 alone cannot transform KLS cells KLS cells were transduced with Lgr4 cDNA or EV control. Cells were then incubated at a density of 20,000 per culture dish in methylcellulose medium. The number of colonies (A) and the number of cells (B) were counted after 6 days incubation. This was repeated for four consecutive weeks. Results from three independent experiments were graphed as mean ± SEM. (C) Colonies formed by KLS cells expressing Lgr4 cDNA or EV were analysed using bright-field microscopy (10x magnification).

114

3.17 Lgr4 overexpression enhances cell proliferation of GMPA9M in vitro Previous studies have shown that Hoxa9/Meis1a transduced GMPs (GMPA9M) cannot cause leukaemia, possibly due to low levels of β-catenin expression [104]. Nonetheless, overexpression of β-catenin in GMPA9M gave rise to AML similar to KLSA9M [104]. To further investigate the leukaemia-enhancing effect of Lgr4, we examined whether overexpression of Lgr4 can increase β-catenin levels in GMPA9M and consequently drive AML development. To this end, the endogenous levels of β-catenin expression were compared. Western blot analysis of GMPA9M and KLSA9M pre-LSCs confirmed the higher β-catenin expression in KLSA9M pre-LSCs (Figure 3.18A). Additionally, western blot analysis revealed higher Lgr4 expression in KLSA9M pre-LSCs compared to GMPA9M (Figure 3.18A). To overexpress Lgr4 in GMPA9M, we retrovirally transduced a Lgr4 cDNA construct into GMPA9M and western blot analysis confirmed upregulation of both Lgr4 and β-catenin compared to EV control (Figure 3.18B). To assess the functional effect of Lgr4 overexpression in GMPA9M, colony forming assays were conducted. As shown in Figure 3.18, colony number (p=0.005; Figure 3.18C) and cell number (p=0.0007; Figure 3.18D) were significantly increased in GMPA9M expressing Lgr4 cDNA compared to EV. This result indicates an increase in proliferative potential in vitro, consistent with the observation in KLSA9M pre-LSCs (Section 3.9).

To examine whether reactivation of β-catenin by overexpressing Lgr4 can transform GMPA9M into leukaemogenic cells, we inoculated GMPA9M expressing Lgr4 cDNA or EV into sublethally irradiated C57BL/6 mice. Interestingly, mouse survival analysis revealed no significant difference between Lgr4 cDNA and EV group (p=0.2209; Figure 3.18E). This could be due to levels of β-catenin expression not being sufficient enough for leukaemic transformation in these cells. Possibly a higher level of β-catenin expression is required for cells to become leukaemogenic, β-catenin expression in GMPA9M expressing Lgr4 cDNA is low as evident in Figure 3.18B.

115

116

Figure 3.18 Lgr4 overexpression enhances proliferation of GMPA9M in vitro (A) GMPA9M and KLSA9M pre-LSCs were harvest when appropriate level of confluence was achieved. Lysates were analysed by western blotting for expression of Lgr4 and β-catenin. (B) GMPA9M pre-LSCs were transduced with Lgr4 cDNA or EV control. Cells were harvested two weeks after transduction. Lysates were analysed by western blotting for expression of Lgr4 and β-catenin. (C and D) GMPA9M pre-LSCs expressing Lgr4 cDNA or EV control were seeded at a density of 1000 cells per culture dish and incubated in methylcellulose medium for 6 days. Results from three independent experiments were graphed as mean ± SEM. (E) Kaplan–Meier survival analysis of mice transplanted with GMPA9M expressing Lgr4 cDNA or EV control.

117

3.18 Summary Lgr4 is a G protein coupled receptor that has recently been shown to upregulate and activate β-catenin signalling and is upregulated in colon cancer [210,214,215,217,218], which is largely driven by aberrant expression of Wnt/β-catenin [248]. Despite the critical role that Wnt/β-catenin plays in AML LSCs [104], the role of Lgr4 in AML has not been investigated to date. Here, we show that Lgr4 is upregulated in human MLL- AML compared to normal HSCs and high Lgr4 expression correlates with poor clinical outcome. High Lgr4 expression was also observed specifically in the LSC compartment. We also showed that the specific combination of Rspo2/Rspo3 with Wnt3a could activate Lgr4 to potentiate β-catenin signalling, leading to enhanced cell proliferation. shRNA-mediated knockdown of Lgr4 in pre-LSCs substantially reduced β-catenin expression levels, indicating Lgr4 is a key regulator of β-catenin expression in AML. Furthermore, Lgr4 deficiency significantly impaired LSC development and leukaemogenesis in vivo, while Lgr4 overexpression substantially enhanced LSC proliferation and shortened AML disease onset in vivo. In contrast, Lgr4 alone was insufficient to transform normal HSCs, suggesting additional oncogenic hits are required to cooperate with Lgr4 for malignant transformation.

To date, drugs developed to target G-protein coupled receptors comprise a vast majority of the currently available treatments in the market. Targeting the Lgr4/β-catenin signalling axis by inhibiting Lgr4 provides promising therapeutic value. However, no compound has been developed to inhibit Lgr4 to date. Therefore, we aim to characterise important components of the Lgr4/β-catenin signalling pathway in order to identify alternative mechanisms of pharmacologically inhibiting this critical signalling cascade.

118

4 Characterising crucial components of the Lgr4 signalling cascade

119

4.1 Introduction AML is largely a stem cell disease and studies have demonstrated the central role of Wnt/β-catenin signalling in the establishment of AML LSCs [104,105,232]. As previously described, β-catenin is difficult to target due to its nuclear localisation and structure [133]. Furthermore, β-catenin is the only Wnt member upregulated in LSCs [104]. We therefore aim to identify druggable upstream regulators of β-catenin as a means of abolishing LSC activity in AML. While we identified the GPCR Lgr4 as a promising candidate, there are currently no commercially available small molecular inhibitors targeting Lgr4. De novo development of a small molecule inhibitor of a specific target is a lengthy process and is beyond the expertise of our lab. Nonetheless, we aimed to increase our understanding of LSC biology and further delineate the Lgr4 pathway to identify important signalling nodes for potential therapeutic applications within the Lgr4/β-catenin signalling axis itself.

4.2 Identification of downstream Lgr4 signalling components by gene expression profiling As described in Chapter 3, activation of Lgr4 in KLSA9M pre-LSCs produced a substantially more aggressive AML in vivo similar to the rate of disease progression induced by KLSMLL pre-LSCs. In order to identify important signalling molecules regulated by Lgr4, we performed a microarray experiment on KLSA9M LSCs expressing Lgr4 cDNA or EV control (described in Section 2.2.4).

Microarray analysis revealed that overexpression of Lgr4 significantly downregulated a total of 107 genes and upregulated 56 genes in KLSA9M LSC compared to EV control (Figure 4.1). A cut-off value of 3-fold up or down-regulation and a statistically significant p-value of less than 0.01 across three biological repeat arrays were used as the selection criteria for candidate target identification.

120

121

Figure 4.1 Identification of downstream Lgr4 signalling components by gene expression profiling Microarray gene expression profiling was performed using 3 biological replicates of KLSA9M LSCs expressing Lgr4 cDNA or EV control. Heatmap showing the most differentially expressed genes following overexpression of Lgr4 in KLSA9M LSCs compared to EV control. 107 genes were significantly downregulated by Lgr4 and 56 genes were significantly upregulated by Lgr4 (fold change > 3 and p<0.01).

122

4.3 Validation of downstream target genes identified by microarray Independent qRT-PCR was performed to validate the top 14 candidate genes and 12 out of 14 were confirmed to be significantly downregulated (Figure 4.2A) or upregulated (Figure 4.2B). Interestingly, these top candidate genes were all shown to be involved in the pathogenesis of various types of cancer [177,249-259]. For example, among the genes downregulated by Lgr4 is Lcn2 (lipocalin 2) and Gadd45g (Growth arrest and DNA-damage-inducible protein gamma). Loss of Lcn2, a small extracellular protein, has been noted to increase tumour multiplicity in intestinal cancer [260]. Lcn2 has also been shown to inhibit tumour angiogenesis in breast cancer cell lines [261]. Gadd45 proteins are cellular stress sensors that are downregulated in various types of cancer [254] and are positively regulated by tumour suppressors such as p53 and BRAC1 [262,263].

123

124

Figure 4.2 Validation of microarray data using qRT-PCR The top downregulated (A) and upregulated (B) genes identified by microarray gene expression profiling were validated by qRT-PCR. Results from three independent experiments were graphed as mean ± SEM.

125

4.4 Gene set enrichment analysis and its validation To further investigate the role of Lgr4 in mediating Wnt/β-catenin signalling, gene set enrichment analysis (GSEA) was performed to compare our microarray data to previously published Wnt/β-catenin target genes. GSEA is a computational method to examine whether a pre-defined set of genes (such as genes in a common pathway) show significantly different expression levels between two samples [264]. Interestingly, GSEA analysis revealed that known Wnt/β-catenin target genes including Csnk1e, Apc and Ctnnb1 were highly enriched (p<0.001) in Lgr4 overexpressing KLSA9M LSCs (Figure 4.3A).

To assess whether Lgr4-regulated genes involved in Wnt/β-catenin signalling activation have a functional role in MLL pre-LSCs, we confirmed the altered protein level of one of the top target genes, Csnk1e (Casein kinase 1 epsilon), which has been reported to stabilize β-catenin and to inhibit granulocytic differentiation in haematopoiesis [265]. Western blot analysis showed that Lgr4 depletion significantly reduced Csnk1e protein expression in KLSMLL pre-LSCs (Figure 4.3B). Furthermore, treatment of KLSMLL pre-LSCs with IC261 (4 µM), a specific Csnk1e inhibitor [266], led to substantial reduction in colony formation (Figure 4.3C), consistent with Lgr4 potentiation of Wnt/β-catenin signalling and leukaemogenesis, in part via regulation of Csnk1e.

126

A Enrichment plot: KEGG_WNT_SIGNALING_PATHWAY

0.30 NES = 1.34 0.25 !;! p <0.001 "'u 0.20 FOR = 0.21 i5v O. IS ~ E 0.10 u E o.os ~ 0.00 "'~ w ~.OS -0.10

B c KLSMLL pre-LSC - IC261 treatment KLSMLL pre-LSC p<0.0001

Scr shlgr4#3 shlgr4#4 120 Csnk1e

Actin ~ 80 '2 0 0 u ~ 40

0 OMSO

127

Figure 4.3 Wnt/β-catenin gene signature regulated by Lgr4 (A) Gene set enrichment analysis (GSEA) of KLSA9M LSC expressing Lgr4 cDNA or EV. Red rectangle represents the subset of genes highly upregulated in KLSA9M LSC expressing Lgr4 cDNA compared to EV control. (B) KLSMLL pre-LSCs expressing Lgr4shRNA#3, #4 or Scr control were harvested when appropriate level of confluence was achieved. Lysates were analysed by western blotting for expression of Csnk1e. (C) KLSMLL pre-LSCs were seeded at a density of 1000 cells per culture dish and incubated with IC261 (3 µM and 4 µM) or DMSO as control in methylcellulose medium. The number of colonies was counted after 6 days incubation. Results from three independent experiments were graphed as mean ± SEM.

128

4.5 Rgs1 is a major downstream target of Lgr4 Lgr4 belongs to the GPCR family. This family of proteins is known for coupling with G proteins in order to transduce cell signalling as described in the introduction [267]. We therefore sought to identify and focus in depth further experimentation on any G proteins that were identified from our microarray data. Interestingly, we identified Regulator of G protein signalling 1 (Rgs1) among the top 10 downregulated genes by Lgr4 overexpression in KLSA9M LSCs compared to EV control (Figure 4.1).

Rgs1 is a member of the regulator of G protein signalling family, which consists of more than 20 Rgs proteins. As described previously, G proteins consist of α, β and γ subunits. Rgs proteins negatively regulate G proteins via the GTPase activating protein activity of their Rgs domain to accelerate the hydrolysis of GTP to GDP, resulting in deactivation of Gα subunits and premature termination of downstream signalling [198,199]. The binding of Rgs proteins to Gα subunits prevents G proteins from binding to their effectors, therefore attenuating downstream signalling [200]. Rgs proteins equip cells with the ability to control the duration and magnitude of GPCR signalling, one of the most important mechanisms by which cells transduce extracellular signals to intracellular signalling pathways. Dysregulation of this mechanism has been implicated in a variety of cancers including AML [249].

To validate the regulation of Rgs1 expression by Lgr4 as evidenced by the microarray, western blot analysis was performed. As shown in Figure 4.4A, expression of Rgs1 was decreased in Lgr4 overexpressing KLSA9M LSCs compared to EV control. In addition, Rgs1 expression was increased in Lgr4 deficient KLSMLL pre-LSCs compared to Scr (Figure 4.4B). Hence Rgs1 is a bona fide downstream target of Lgr4.

129

Figure 4.4 Validation of Rgs1 protein expressions (A) KLSA9M LSCs expressing Lgr4 cDNA or EV control were analysed by western blotting for expression of Rgs1. (B) KLSMLL pre-LSCs expressing Lgr4shRNA#3, #4 or Scr control were analysed by western blotting for expression of Rgs1.

130

4.6 Gαq is a downstream target of Rgs1 and its inhibition impairs cell proliferation of MLL-AML in vitro Previous studies of Rgs proteins have demonstrated that Rgs1 can interact with and regulate the activity of the G alpha protein family members Gαi and Gαq to suppress downstream signalling [268,269]. To investigate if these Gα proteins have a role in regulating AML LSC activity we used a small molecule inhibitor approach. KLSMLL pre-LSCs were treated with pertussis toxin (PTX; Gαi inhibitor) and GP-antagonist 2A (GP; Gαq inhibitor) to examine the effect of Gαi or Gαq inhibition.

Pertussis toxin (PTX) is a toxin produced by the notorious whooping cough causing bacterium Bordetella pertussis [270]. It is widely used as a Gαi inhibitor and its mechanism of action is to catalyse the ADP-ribosylation of Gαi thereby preventing the G protein from binding to its cognate G protein-coupled receptors [271,272]. To test if Gαi plays a role in the proliferation of KLSMLL pre-LSCs, colony forming assays were carried out in the absence or presence of PTX. Previous studies have demonstrated that the concentration of PTX needed for half-maximal effect on ADP-ribosylation is approximately 0.4 µg/mL, with the majority of studies using concentrations of less than 1 µg/mL to inhibit Gαi [273-276]. However, colony forming assays revealed no significant difference either in colony number or cell number in the presence of 4 µg/mL or 8 µg/mL of PTX (Figure 4.5A and B). These results suggest that Gαi may not be involved in the regulation of KLSMLL pre-LSCs.

GP-antagonist 2A (GP) is an 11 amino acid peptide that is designed based on the neuropeptide substance P [277]. It selectively inhibits Gαq signalling by interacting with the G protein but not the receptor [277,278]. Its main application is in neurobiology where studies have used concentrations ranging from 10 – 50 µM to inhibit Gαq [279-281]. Its application has not been tested in any cancer model to date. To investigate if inhibition of Gαq is capable of perturbing the proliferation ability of KLSMLL pre-LSCs, colony forming assays were conducted in the presence of 20 or 40 µM GP. Figure 4.5C and D show that GP significantly reduced both colony number and cell number at concentrations of 20 and 40 µM respectively, with the effect being more

131

prominent at 40 µM. This indicates Gαq inhibition by GP impairs the proliferation ability of KLSMLL pre-LSCs, suggesting that Gαq but not Gαi is involved in regulating LSC activity in MLL-AML.

132

Figure 4.5 Gαq inhibitor but not Gαi inhibitor impairs colony formation of KLSMLL pre-LSCs KLSMLL pre-LSCs were seeded at a density of 1000 cells per culture dish and incubated with Gαi inhibitor pertussis toxin (4 µg/mL and 8 µg/mL) in methylcellulose medium. The number of colonies (A) and the number of cells (B) were counted after 6 days incubation. KLSMLL pre-LSCs were seeded at a density of 1000 cells per culture dish and incubated with Gαq inhibitor GP antagonist 2A (GP) (20 µM and 40 µM) in methylcellulose medium. The number of colonies (C) and the number of cells (D) were counted after 6 days incubation. Results from three independent experiments were graphed as mean ± SEM.

133

4.7 Gαq inhibition enhances differentiation, apoptosis, and suppresses β- catenin expression of MLL-AML In the previous section, we have shown that Gαq inhibition by GP treatment impairs colony forming ability of KLSMLL pre-LSCs. In this section, we examined the phenotypic alterations in LSC behaviour induced by GP treatment. Following GP treatment of KLSMLL pre-LSCs, it was observed that colonies in the GP treated group were smaller in size compared to the control. More interestingly, GP treated colonies displayed an unusual, jagged and irregular colony morphology (Figure 4.6A), indicative of alterations in the differentiation status of cells. To investigate the differentiation status and morphology of the cells in more detail, Wright and Giemsa staining was performed on control and GP treated cells following five days treatment. As expected, it was observed that some cells in the GP treated group displayed a multi-nucleated cell morphology in contrast to the large mononuclear appearance of control treated cells, indicating a more mature differentiation status induced by Gαq inhibition (Figure 4.6B).

Next, we sought to investigate if GP inhibits pre-LSC activity by inducing apoptosis. Annexin V and 7-AAD double staining was performed following 5 days treatment with GP. As shown in Figure 4.6C, a slight increase in the double-positive apoptotic populations in GP treated KLSMLL pre-LSCs compared to the control was observed.

As described in Chapter 3, Lgr4 is a key regulator of β-catenin in AML LSCs and plays an important role in regulation of LSC self-renewal and proliferation. Previous studies have shown that the expression of β-catenin can be regulated by Gαq [282,283]. Since inhibition of Gαq by GP was shown in this study to impair the proliferative ability of KLSMLL pre-LSCs, we hypothesised that Gαq might regulate β-catenin expression. As expected, western blot analysis of KLSMLL pre-LSCs treated with GP showed a marked decrease of β-catenin expression compared to control cells (Figure 4.6D). These results together indicate that Gαq may be an important component of Lgr4/β-catenin signalling and pharmaceutical inhibition of Gαq impairs pre-LSCs by blocking differentiation and increasing apoptosis.

134

135

Figure 4.6 Inhibition of Gαq in KLSMLL pre-LSCs enhances differentiation, apoptosis, and suppresses β-catenin expression KLSMLL pre-LSCs were incubated with Gαq inhibitor (GP, 40 µM) in methylcellulose medium for 5 days. (A) Colonies formed by GP or control treated cells were analysed using bright-field microscopy (10x magnification). (B) Wright-Giemsa staining was performed on GP or control treated cells and cell morphology was analysed using bright-field microscopy (100x magnification). (C) Representative flow scatter plots of GP or control treated cells following double staining with Annexin V APC and 7-AAD. (D) Lysate from GP and control treated cells were analysed by western blotting for expression of β-catenin.

136

4.8 shRNA-mediated knockdown of Gαq in pre-LSCs suppresses β-catenin expression and impairs cell proliferation in vitro In the previous section we have demonstrated that pharmacological inhibition of Gαq reduces β-catenin expression in KLSMLL pre-LSCs. To further validate the role of Gαq in β-catenin regulation, we performed Gαq shRNA-mediated stable knockdown to compare the resulting phenotypic alterations and to eliminate false positive results induced by possible off-target effects of GP. KLSMLL pre-LSCs were transduced with either one of three different shRNAs against Gαq or a non-targeting Scr control. Following transduction, western blot analysis was conducted to assess the Gαq knockdown efficiency. As shown in Figure 4.7A, each of the Gαq shRNAs markedly reduced Gαq expression compared to Scr control. In addition, β-catenin expression was examined and western blot analysis revealed a significant reduction in the level of β- catenin in Gαq knockdown cells compared to Scr control. These results are consistent with the reduction in β-catenin expression by GP treatment (Figure 4.6D). Together these results confirm that Gαq is a bona fide upstream regulator of β-catenin in MLL- AML LSCs.

To investigate whether knockdown of Gαq impairs cell proliferation of KLSMLL pre- LSCs, colony forming assays were conducted on KLSMLL pre-LSCs expressing Gαq shRNA or Scr control. Consistent with GP treatment, colony forming assays showed that Gαq knockdown by shRNA significantly reduced the number of colonies formed (Figure 4.7B) and cell number (Figure 4.7C) of KLSMLL pre-LSCs, suggesting impaired proliferative ability. Furthermore, flow cytometric analysis following Annexin V and 7-AAD apoptosis analysis revealed that Gαq-deficient KLSMLL pre-LSCs displayed an increase in apoptotic cell death (Figure 4.7D), again consistent with GP mediated inhibition (Figure 4.6C).

As described in Chapter 3, Lgr4 is involved in β-catenin activation of both KLSMLL and GMPMLL pre-LSCs and contributes to their leukaemogenesis. After establishing a role for Gαq in KLSMLL pre-LSCs, we sought to determine whether Gαq plays a role in GMPMLL pre-LSCs. Three different shRNAs against Gαq and a non-targeting Scr

137

control were transduced into GMPMLL pre-LSCs. Subsequent western blot analysis confirmed Gαq knockdown by all three shRNA constructs compared to Scr control (Figure 4.8A). To examine whether knockdown of Gαq in GMPMLL pre-LSCs could downregulate β-catenin expression as seen in KLSMLL pre-LSCs, western blot analysis was conducted and as shown in Figure 4.8A GMPMLL pre-LSCs expressing Gαq shRNA exhibit lower β-catenin expression compared to Scr control. This indicates that Gαq is an upstream regulator of β-catenin expression in GMPMLL pre-LSCs.

To investigate whether knockdown of Gαq impairs cell proliferation of GMPMLL pre- LSCs, colony forming assays were performed. Consistent with the results observed in KLSMLL pre-LSCs, colony forming assays of Gαq-deficient GMPMLL pre-LSCs showed a significant reduction of the number of colonies formed (Figure 4.8B) and cell number (Figure 4.8C) compared to Scr control. In addition, flow cytometric analysis following Annexin V and 7-AAD apoptosis analysis revealed that Gαq-deficient GMPMLL pre-LSCs had an increase in the percentage of apoptotic cells compared to Scr control (Figure 4.8D), which is in line with previous observations on KLSMLL pre- LSCs as described in Figure 4.7D.

Together these results indicate that Gαq is an upstream regulator of β-catenin expression in MLL-AML pre-LSCs. Additionally, Gαq plays an important role in regulating apoptosis and proliferation ability of pre-LSCs in vitro.

138

139

Figure 4.7 Knockdown of Gαq in KLSMLL pre-LSCs suppresses β-catenin expression and impairs colony formation (A) KLSMLL pre-LSCs were transduced with three different Gαq shRNA or Scr control. Cells were harvested two weeks after transduction. Lysates were analysed by western blotting for expression of Gαq and β-catenin. (B and C) KLSMLL pre-LSCs expressing Gαq shRNA#1, #2, #3 or Scr control were seeded at a density of 1000 cells per culture dish and incubated in methylcellulose medium for 6 days. Results from three independent experiments were graphed as mean ± SEM. (D) Representative flow scatter plots of KLSMLL pre-LSCs expressing Scr or Gαq shRNA#1 following double staining with Annexin V APC and 7-AAD.

140

141

Figure 4.8 Knockdown of Gαq in GMPMLL pre-LSCs suppresses β-catenin expression and impairs colony formation (A) GMPMLL pre-LSCs were transduced with three different Gαq shRNA or Scr control. Cells were harvested two weeks after transduction. Lysates were analysed by western blotting for expression of Gαq and β-catenin. (B and C) GMPMLL pre-LSCs expressing Gαq shRNA#1, #2, #3 or Scr control were seeded at a density of 1000 cells per culture dish and incubated in methylcellulose medium for 6 days. Results from three independent experiments were graphed as mean ± SEM. (D) Representative flow scatter plots of GMPMLL pre-LSCs expressing Scr or Gαq shRNA#1 following double staining with Annexin V APC and 7-AAD.

142

4.9 Inhibition of Gαq in KLSMLL pre-LSCs reverses the effect of Wnt3a/Rspo3 on cell proliferation and β-catenin activation In previous sections, we have demonstrated that Gαq inhibition by GP causes downregulation of β-catenin and impairs the proliferation ability of KLSMLL pre-LSCs in vitro. In addition, knockdown of Gαq by shRNAs induced phenotypic changes consistent with GP treatment. Furthermore, we showed that Rspo3 in combination with Wnt3a increased β-catenin expression and proliferative ability in KLSMLL pre-LSCs. To further investigate whether Gαq is functionally involved in the Wnt3a/Rspo3/Lgr4/β-catenin signalling axis to regulate pre-LSC proliferation, colony forming assays were conducted in the presence of Wnt3a alone, Rspo3 alone, Wnt3a/Rspo3 or Wnt3a/Rspo3/GP. Figure 4.9A and B show that GP treatment significantly reversed the enhancement of colony formation (p=0.0003) and cell proliferation (p=0.0002) induced by co-treatment of Wnt3a and Rspo3. Consistent with the functional result, subsequent western blot analysis of the treated cells showed that GP treatment attenuated the effect of Wnt3a/Rspo3-potentiated β-catenin activation (Figure 4.9C). These results suggest that Gαq is part of the Wnt3a/Rspo3/Lgr4/β- catenin signalling axis. Together this study identifies a previously uncharacterised Wnt3a/Rspo3 driven Lgr4-Gαq-β-catenin signalling pathway that governs the proliferative ability of KLSMLL pre-LSCs in AML.

143

144

Figure 4.9 Inhibition of Gαq reverses the phenotypes induced by co-incubation of Wnt3a and Rspo3 KLSMLL pre-LSCs were seeded at a density of 1000 cells per culture dish and incubated with Wnt3a, Rspo3, Wnt3a/Rspo3 or Wnt3a/Rspo3/GP in methylcellulose. The number of colonies (A) and the number of cells (B) were counted after 6 days incubation. Results from three independent experiments were graphed as mean ± SEM. (C) KLSMLL pre-LSCs were incubated with Wnt3a, Rspo3, Wnt3a/Rspo3 or Wnt3a/Rspo3/GP in methylcellulose for 6 days. Lysates were analysed by western blotting for expression of β-catenin.

145

4.10 shRNA-mediated knockdown of Lgr4 or Gαq abolishes the effect of Wnt3a/Rspo3 on β-catenin activation To further validate the newly identified Wnt3a/Rspo3-Lgr4-Gαq-β-catenin signalling axis, we hypothesised that knockdown of Lgr4 or Gαq would impair the activation of β- catenin caused by co-treatment of Wnt3a and Rspo3. To address this, KLSMLL pre- LSCs expressing Lgr4 shRNA#4 or Scr were incubated with Wnt3a conditioned media together with 50 ng/ml of Rspo3 for 24 hours. Consistent with our findings in Section 3.3, β-catenin expression was significantly upregulated in KLSMLL pre-LSCs expressing Scr upon Wnt3a/Rspo3 stimulation (Figure 4.10A lane 1 and 3). In contrast, when interrogating KLSMLL pre-LSCs expressing Lgr4 shRNA#4, co-treatment of Wnt3a and Rspo3 was not able to significantly upregulate β-catenin expression (Figure 4.10A lane 3 and lane 6). This indicates that activation of β-catenin by Wnt3a/Rspo3 requires Lgr4.

To further validate the involvement of Gαq in the Wnt3a/Rspo3-Lgr4-Gαq-β-catenin signalling axis, we incubated KLSMLL pre-LSCs expressing Gαq shRNA#1 or Scr with Wnt3a conditioned media together with 50 ng/mL of Rspo3 for 24 hours. It was expected that knockdown of Gαq would impair the activation of β-catenin by Wnt3a/Rspo3 if Gαq is part of the Wnt3a/Rspo3/Lgr4/β-catenin axis. Indeed, subsequent western blot analysis showed that Gαq knockdown completely abolished the activation of β-catenin by Wnt3a/Rspo3 (Figure 4.10B). This indicates that Gαq is an integral member of the Lgr4/β-catenin signalling axis that drives LSC function in AML. Therefore, our findings strongly suggest that Gαq is an important player of Wnt3a/Rspo3/Lgr4/β-catenin signalling, regulating proliferation and apoptosis of KLSMLL pre-LSCs.

146

Figure 4.10 Knockdown of Lgr4 or Gαq abolishes the effect of Wnt3a/Rspo3 on β- catenin activation (A) KLSMLL pre-LSCs expressing Scr or Lgr4 shRNA#4 were incubated with Wnt3a or Wnt3a/Rspo3 in IMDM/15%FCS for 24 hours. Lysates were analysed by western blotting for expression of β-catenin. (B) KLSMLL pre-LSCs expressing Scr or Gαq shRNA#1 were incubated with Wnt3a or Wnt3a/Rspo3 in IMDM/15%FCS for 24 hours. Lysates were analysed by western blotting for expression of β-catenin.

147

4.11 Summary Lgr4 belongs to the G protein-coupled receptor family that signals via heterotrimeric G proteins. In Chapter 3, we demonstrated the importance of Lgr4 in AML LSC development and leukaemogenesis. However, due to the fact that a pharmacological inhibitor for Lgr4 is not commercially available at present, we aimed to further delineate the Lgr4 pathway and find potential druggable targets within this signalling network. Our microarray expression analysis comparing Lgr4 cDNA to EV control transduced KLSA9M LSCs revealed that the Wnt pathway was activated and the microarray data were confirmed by functional validation of an identified Wnt pathway component, Csnk1e. Inhibition of Csnk1e by its selective inhibitor significantly reduced colony forming ability of KLSMLL pre-LSCs, consistent with Lgr4-mediated potentiation of Wnt/β-catenin signalling and leukaemogenesis. Our microarray data also revealed that Rgs1 is the most differentially expressed gene regulated by Lgr4 and is the only G protein-related gene among the top candidates in the microarray data. Rgs1 is a known inhibitor of Gαq and Gαi signalling [268,269]. To examine the role of Gαq and Gαi inhibition in KLSMLL pre-LSCs, colony forming assays were performed in the presence of the Gαi inhibitor PTX or the Gαq inhibitor GP. We showed that only Gαq was involved in regulating the proliferation of KLSMLL pre-LSCs. Further characterisation studies using GP treatment and Gαq shRNA mediated knockdown confirmed Gαq as a key regulator in KLSMLL pre-LSC proliferation and apoptosis, as well as a modulator of β-catenin expression. Furthermore, knockdown of Lgr4/Gαq completely abolished Wnt3a/Rspo3-enhanced β-catenin activation, suggesting that Gαq is absolutely required for the activation of β-catenin expression in the Wnt3a/Rspo3/Lgr4/β-catenin signalling cascade. Together, this study identifies a previously uncharacterised Wnt3a/Rspo3-Lgr4-Gαq-β-catenin signalling pathway that governs the proliferative potential of MLL-AML pre-LSCs and LSCs.

148

5 Lgr4/β-catenin signalling pathway regulates AML LSC activity by altering mitochondrial function

149

5.1 Introduction As described in the previous two chapters, Lgr4 plays a crucial role in the activation of β-catenin signalling and development of AML LSCs. Its ligands Rspo2/3 synergise with Wnt3a to activate β-catenin and regulate the proliferative ability of pre-LSCs. We have identified a novel Wnt3a/Rspo3-Lgr4-Gαq-β-catenin signalling pathway in pre-LSCs and LSCs. In this chapter, we aimed to further understand this signalling pathway by investigating the mechanisms of how this signalling axis governs LSC function.

5.2 Identifying common downstream target genes of Lgr4 and Gαq To further investigate how Lgr4/Gαq/β-catenin signalling axis governs LSC function, we sought to identify common target genes shared by both Lgr4 and Gαq. To this end, microarray studies were conducted using KLSMLL LSCs expressing Scr, Lgr4 shRNA or Gαq shRNA. The data presented in Figure 5.1 of this section was obtained by Dr Jennifer Lynch.

Microarray expression profiling was performed using 3 biological replicates of KLSMLL LSCs expressing Scr, Lgr4 shRNA or Gαq shRNA. A total of 31 genes were identified as common targets regulated by both Lgr4 and Gαq (Figure 5.1). The highest differentially expressed gene was the Growth arrest and DNA-damage-inducible protein alpha (Gadd45a), a stress sensor that modulates the response to cellular stress such as treatment with DNA-damaging agents [254]. Interestingly, Gadd45a is one of the three isoforms of Gadd45, of which the other two isoforms (Gadd45g and Gadd45b) have been identified among the top 10 downregulated genes in our previous microarray comparing Lgr4 overexpressing KLSA9M LSCs with EV (Figure 4.1). This striking finding prompted us to hypothesise that Gadd45a may be a key downstream effector of Lgr4/Gαq and play an important role in AML development.

150

Figure 5.1 Common target genes shared by both Lgr4 and Gαq Microarray gene expression profiling was performed using 3 biological replicates of KLSMLL LSCs expressing Lgr4 shRNA, Gαq shRNA or Scr control. Heatmap showing the most differentially expressed genes following Lgr4 or Gαq knockdown. A total of 31 common target genes are regulated by both Lgr4 and Gαq. (fold change > 1.6 and p<0.05). This microarray experiment was conducted by Dr Jennifer Lynch.

151

5.3 Gadd45a knockout enhances cell proliferation of MLL-AF9 transformed c-Kit+ bone marrow cells in vitro Gadd45a is a growth arrest and DNA damage response mediator that has been shown to regulate cell survival, DNA repair and apoptosis [284-286]. Studies have shown that Gadd45a expression is suppressed in a number of cancers including AML [287-290]. As described in the previous section, Gadd45a was identified as the most highly differentially expressed common target gene upregulated by depletion of Lgr4 and Gαq; hence we sought to explore the functional role of Gadd45a in MLL-rearranged AML. To address this, Gadd45-/- and WT c-Kit+ bone marrow cells (a kind gift from Prof Richard D’Andrea, University of South Australia), which contain haematopoietic stem and progenitor cells, were retrovirally transduced with the MLL-AF9-GFP oncogene. After transduction, GFP positive cells were sorted using FACS to enrich for MLL-AF9 containing cells and subsequently qRT-PCR was conducted to assess the expression levels of Gadd45a. As shown in Figure 5.2A, qRT-PCR confirmed the loss of Gadd45a transcript compared to WT control. To investigate the effect of Gadd45a-/- on cell proliferation in vitro, colony forming assays were conducted on MLL-AF9 transformed Gadd45-/- c-kit+ bone marrow cells (MLL-Gadd45-/-) and MLL-AF9 transformed WT cells (MLL-WT). Figure 5.2B and C show that MLL-Gadd45a-/- cells gave rise to significantly higher number of colonies (p=0.0136) as well as total number of cells (p=0.0142) compared to MLL-WT control, suggesting an enhancement of cell proliferation in vitro. Therefore our findings are in line with previous publications which suggest a tumour suppressor role of Gadd45a [254].

152

153

Figure 5.2 Gadd45a knockout enhances colony forming ability of MLL-AF9 + transduced c-Kit bone marrow cells (A) Relative mRNA expression of Gadd45a was analysed by qRT-PCR in MLLAF9 + transduced WT c-Kit mouse bone marrow cells (MLL-WT) and MLLAF9 transduced -/- + -/- Gadd45a c-Kit mouse bone marrow cells (MLL-Gadd45a ). (B and C) MLL-WT -/- and MLL-Gadd45a cells were seeded at a density of 1000 cells per culture dish and incubated in methylcellulose medium. The number of colonies (B) and the number of cells (C) were counted after 6 days incubation. Results from three independent experiments were graphed as mean ± SEM.

154

5.4 Gadd45a knockout enhances MLL leukaemogenesis in vivo Recent studies have shown that AML patients with poor survival express lower levels of Gadd45a or hypermethylation of the Gadd45a promoter [289,291]. In line with this, we have shown that MLL-Gadd45a-/- cells are significantly more proliferative in vitro compared to MLL-WT control. Next, we investigated whether Gadd45a plays a role in MLL leukaemogenesis in vivo using our mouse model of MLL-AML. To this end, MLL-WT and MLL-Gadd45a-/- cells were inoculated into sublethally irradiated C57BL/6 mice. Gadd45a knockout significantly shortened the onset of AML compared to MLL-WT control (Figure 5.3A; p=0.0374). This indicates that loss of Gadd45a plays a significant role in AML initiation and enhances leukaemogenesis in vivo.

To investigate the effect of Gadd45a deletion on established AML and the self-renewal potential of MLL LSCs, we performed in vivo serial transplantations. Sorted GFP+ leukaemic cells harvested from primary transplantation were serially transplanted into secondary and tertiary recipient mice. MLL-Gadd45a-/- cells exhibited gradually increased ability to reduce disease latency in secondary and tertiary recipients, as compared to WT control cells (secondary, p=0.0204, Figure 5.3B; tertiary, p=0.0009, Figure 5.3C), consistent with a progressive increase in in vivo self-renewal. In addition, in vivo limiting dilution transplantation assays were performed to examine the LSC frequency. Sorted GFP+ leukemic cells from the bone marrow of secondary recipients were serially diluted to 1000 and 100 cells and transplanted into tertiary recipient mice. As shown in Figure 5.3D, Gadd45a deletion resulted in a greater than 10-fold increase in the frequency of long-term repopulating leukaemia initiating cells compared to WT control (1/142, 95% CI: 1/428 to 1/47; and 1/1650, 95% CI: 1/5117 to 1/532, p=0.00114), indicating a stemness-promoting role for Gadd45a loss in MLL AML. These results collectively show that loss of Gadd45a promotes MLL leukaemogenesis in vivo by increasing the LSC frequency and highlight a key role of Gadd45a as a downstream mediator of Lgr4 activity.

155

Figure 5.3 Gadd45a knockout enhances MLL leukaemogenesis in vivo (A) Kaplan–Meier survival analysis of mice transplanted with MLL-WT or MLL- -/- Gadd45a cells. (B) Kaplan–Meier survival analysis of secondary transplantation with + FACS sorted GFP mouse bone marrow cells derived from (A). (C) Kaplan–Meier + survival analysis of tertiary transplantation with FACS sorted GFP mouse bone marrow cells derived from (B). (D) In vivo limiting dilution analysis of tertiary transplantation reveals greater than 10-fold increase in LSC frequency upon Gadd45a knockout compared to WT control (1/142, 95% CI: 1/428 to 1/47; and 1/1650, 95% CI: 1/5117 to 1/532, p=0.00114).

156

5.5 Mitochondrial genes are regulated by both Lgr4 and Gαq LSCs in AML depend mainly on mitochondrial oxidative phosphorylation (OXPHOS) for their energy production and survival [292]. Consistently, our microarray data obtained from KLSMLL LSCs expressing Scr, Lgr4 shRNA or Gαq shRNA revealed that mitochondrial DNA (mtDNA)-coded subunits of OXPHOS complexes, including complex III subunit Cytb and complex I subunits Nd2 and Nd4l, are common target genes regulated by both Lgr4 and Gαq (Figure 5.1). Cytb encodes for cytochrome b reductase, Nd2 encodes for NADH dehydrogenase 2 and Nd4l encodes for NADH dehydrogenase 4l, all of which are critical components for mitochondrial OXPHOS [293-295]. To further validate the expression of Cytb, Nd2 and Nd4l, qRT-PCR was performed on KLSMLL LSCs expressing Scr, Lgr4 shRNA or Gαq shRNA. As shown in Figure 5.4, all three mtDNAs were significantly downregulated by Lgr4 or Gαq knockdown.

157

Figure 5.4 Lgr4 upregulates mitochondrial DNAs Relative mRNA expression of Nd4l, Nd2 and Cytb were validated by qRT-PCR in KLSMLL LSCs expressing Scr, shLgr4 or shGαq. Results from three independent experiments were graphed as mean ± SEM.

158

5.6 Lgr4 signalling regulates mitochondrial OXPHOS in pre-LSCs Since mitochondrial complex I and III are enzymes essential for OXPHOS, we next examined the effect of Lgr4 on mitochondrial activity by measuring the oxygen consumption rate (OCR), which is indicative of OXPHOS. Oxygen consumption assay was performed on KLSA9M pre-LSCs expressing Lgr4 cDNA or EV control, as shown in Figure 5.5A and B, overexpression of Lgr4 substantially increased the OCR compared to EV control, indicating enhanced mitochondrial activity. Furthermore, shRNA-mediated knockdown of Lgr4 significantly impaired mitochondrial OXPHOS, as evidenced by decreased OCR compared to Scr control (Figure 5.5C and D).

Since mtDNAs are also downstream targets of Gαq as shown in the microarray data, we sought to determine if Gαq can regulate mitochondrial OXPHOS. Indeed, shRNA- mediated knockdown of Gαq resulted in a significant decrease in the OCR compared to Scr control (Figure 5.5E and F), indicating a Gαq-induced impairment of mitochondrial function.

159

160

Figure 5.5 Lgr4 signalling regulates mitochondrial OXPHOS in pre-LSCs Pre-LSCs were incubated with the fluorescent oxygen probe MitoXpress and fluorescence intensity was measured every 5 minutes for 120 minutes at 37°C. Representative dot plots and bar graphs showing mitochondrial oxygen consumption rate (OCR) in KLSA9M pre-LSCs expressing Lgr4 cDNA vs EV control (A and B); KLSMLL pre-LSCs expressing Scr vs Lgr4 shRNA (C and D) and KLSMLL pre-LSCs expressing Scr vs Gαq shRNA (E and F). OCR was calculated as the slope of the linear regression line. Results from three independent experiments were graphed as mean ± SEM.

161

5.7 Lgr4 signalling regulates pre-LSC ATP production Cellular adenosine-5’-triphosphate (ATP) is the prime source of energy for all cellular processes and is solely produced in mitochondria [296]. To further validate the role of Lgr4 and Gαq on mitochondrial energy generation, the concentrations of ATP were measured in KLSMLL pre-LSCs expressing Scr, Lgr4 shRNA or Gαq shRNA. As shown in Figure 5.6, knockdown of Lgr4 or Gαq significantly reduced the levels of cellular ATP, hence consistent with our observations in OCR. These results collectively show that the Lgr4 signalling pathway regulate pre-LSCs by modulating mitochondrial energy generation.

Figure 5.6 Lgr4 or Gαq knockdown impairs mitochondrial ATP production KLSMLL pre-LSCs expressing Scr, Lgr4 shRNA or Gαq shRNA were harvested when appropriate level of confluence was achieved. ATP assay kit (Abcam) was used to measure the total basal ATP concentration (μM) in each sample. Results from three independent experiments were graphed as mean ± SEM.

162

5.8 Lgr4 signalling regulates ROS production of pre-LSCs Similar to normal HSCs, AML LSCs are characterised by their relatively low levels of ROS (ROSlow) [292,297]. Since we have shown that Lgr4 signalling axis regulates mitochondrial complex I and III subunits, which are the two main sites of ROS production, we hypothesised that Lgr4 may regulate ROS production of pre-LSCs.

To investigate the effect of Lgr4 on ROS production, the ROS sensing fluorogenic probe MitoSOXTM Red was used (described in Section 2.2.6.6). We initially tested KLSA9M pre-LSCs expressing Lgr4 cDNA and EV control. Flow cytometric analysis revealed that Lgr4 overexpression increased the percentage of ROSlow cells compared to EV control (Figure 5.7A), indicating a higher proportion of LSC enriched population. To further validate this finding, KLSMLL pre-LSCs expressing Lgr4 shRNA or Scr control were examined for ROS production and as shown in Figure 5.7B, Lgr4 deficiency reduced the percentage of ROSlow cells compared to Scr control, indicating that Lgr4 depletion induces loss of the LSC-enriched compartment. These results show that Lgr4 regulates ROS production of pre-LSCs.

As described previously, Gadd45a is a common target of Lgr4 and Gαq, and its deletion plays a significant role in promoting MLL-AML progression in vivo by increasing the LSC frequency. Hence we hypothesised that Gadd45a may regulate ROS production to modulate LSC frequency. Western blot analysis firstly showed that Gadd45a deletion caused a substantial increase in Cytb expression (Figure 5.8A), suggesting Gadd45a may be involved in mitochondrial function. Subsequent flow cytometric analysis showed that MLL-Gadd45a-/- cells exhibited higher percentage of ROSlow population compared to MLL-WT cells (Figure 5.8B), indicating a higher LSC enriched population, concordant with our in vivo results. Together these results provide evidence suggesting that the Lgr4 signalling cascade regulates pre-LSC function by modulating ROS production.

163

Figure 5.7 Lgr4 regulates mitochondrial ROS production (A) KLSA9M pre-LSCs expressing Lgr4 cDNA or EV control were harvested when appropriate level of confluence was achieved. Cells were stained with superoxide indicator MitoSOX Red followed by flow cytometric analysis and representative flow low contour plots illustrating the percentage of ROS cells. (B) KLSMLL pre-LSCs expressing Lgr4 shRNA or Scr control were harvested when appropriate level of confluence was achieved. Cells were stained with MitoSOX Red followed by flow cytometric analysis and representative flow contour plots illustrating the percentage of low ROS cells.

164

Figure 5.8 Gadd45a regulates mitochondrial ROS production -/- MLL-WT and MLL-Gadd45a cells were harvested when appropriate level of confluence were achieved. (A) Lysates were analysed by western blotting for expression of Cytb. (B) Cells were stained with superoxide indicator MitoSOX Red followed by flow cytometric analysis and representative flow contour plots illustrating the low percentage of ROS cells.

165

5.9 ROS production induced by inhibiting mitochondrial OXPHOS complexes in pre-LSCs Previous studies have identified inhibitors of mitochondrial complex I and complex III to elucidate the mechanisms of mitochondrial electron transport chain activity [298]. Inhibition of electron transport disrupts the proton gradient across the mitochondrial inner membrane, which increases the production of ROS [298]. Since AML LSCs rely on OXPHOS for their energy generation, we hypothesised that pre-LSCs are vulnerable to inhibition of mitochondrial complex I or complex III. The well characterised mitochondrial inhibitors antimycin A and rotenone were used to inhibit mitochondrial complex III and I, respectively [299-303]. KLSA9M pre-LSCs were treated with 100 nM antimycin A or rotenone for 24 hours followed by flow cytometric analysis of ROS production. Figure 5.9A shows that both inhibitors caused an increase in ROS production compared to control treated cells. The percentage of viable cells was examined by Trypan Blue staining after inhibitor treatment. A significant decrease in the percentage of viable cells was observed after antimycin A or rotenone treatment (Figure 5.9B), indicating that pre-LSCs are susceptible to ROS induced damage.

166

5 to % 1 . 9 ' 4 l'1l'v1 tO 0 0 1 3 10 "" j ( );:!' .. ~ · .... 2 "f';) f:..-'tt - 10 Rotenone 0 0 SOt' SOt' lOOt' I 2500' 20Ct' I I 5 to % nM 8 . Rotenone l'1l'v1 31 100 A to' ~ 100 OOC7 A . 3 / ~ nM O n mycin to ~ = i i p ' .. 100 \.\."- Ant , ' ... : 10" rtirnyc l .. A 0 Contro 0 SOl< 501<-l ~ l501< I 1001< 7 ~ 9 8 10 ~ ~ c "' .. ~ ·s; ~ ~ ~ Q. I I 5 10 B 3% . 15 ' \ to I ) [l l I ed ,- R 3 •'Qt to X •I Cortro .'lJ ~ . toSO 2 T~ ,~:;· to (Mi S o RO 0 SOl< SOl< 1001< ~ l501< I ~ u "' .. .. ~ L "' Vi -o

167

Figure 5.9 Antimycin A and rotenone increase ROS production of KLSA9M pre- LSCs KLSA9M pre-LSCs were treated with mitochondrial complex III inhibitor antimycin A (100 nM) or complex I inhibitor rotenone (100 nM) for 24 hours. (A) Cells were stained with MitoSOX Red followed by flow cytometric analysis and representative flow high contour plots illustrating the percentage of ROS cells. (B) Bar graph illustrating the percentage of viable cells following antimycin A or rotenone treatment. Results from three independent experiments were graphed as mean ± SEM.

168

5.10 Lgr4 protects pre-LSCs from ROS induced damage As described previously, Lgr4 overexpression promotes LSC self-renewal, proliferation, survival, and decreases ROS production. Hence we postulated that overexpression of Lgr4 may reverse the ROShigh phenotype caused by the mitochondrial inhibitors and enhance cell survival. To test this, KLSA9M pre-LSCs expressing Lgr4 cDNA or EV control were treated with 100 nM antimycin A for 24 hours. Flow cytometric analysis revealed that antimycin A caused an increase in ROS production in EV control cells, consistent with previous observation (Figure 5.10A). In contrast, under the same treatment condition, overexpression of Lgr4 in KLSA9M pre-LSCs abolished the ROS increase induced by antimycin A (Figure 5.10B). Furthermore, antimycin A treatment significantly decreased the percentage of viable cells in EV control (Figure 5.10C; p=0.0205), whereas it did not have a significant effect on Lgr4 overexpressing cells (Figure 5.10C; p=0.1198). Additionally, Lgr4 overexpression significantly increased the overall cell viability following antimycin A treatment compared to EV control (Figure 5.10C; p=0.0078). These results suggest that Lgr4 overexpression protects KLSA9M pre-LSCs from antimycin A induced ROS damage. To further test our hypothesis, 100 nM rotenone was used for the treatment under the same experimental conditions as antimycin A. Figure 5.11A and B show that Lgr4 overexpression reduced the percentage of ROShigh cells following rotenone treatment compared to EV control. Although rotenone treatment significantly decreased the overall cell viability of both Lgr4 overexpression and EV cells (Figure 5.11C; p<0.0001 and p=0.0014 respectively), overexpression of Lgr4 resulted in a significant increase in the overall cell viability upon rotenone treatment compared to EV control (Figure 5.11C; p=0.0033). This suggests that Lgr4 can partially reverse the ROS damage induced by rotenone. Taken together, these results suggest that Lgr4 plays an important role in promoting the survival of pre-LSCs by regulating mitochondrial ROS production.

169

170

Figure 5.10 Lgr4 protects KLSA9M pre-LSCs from antimycin A induced ROS damage KLSA9M pre-LSCs expressing Lgr4 cDNA or EV control were treated with 100 nM antimycin A for 24 hours. (A) Representative flow contour plots of KLSA9M pre-LSCs expressing EV control following antimycin A treatment illustrating the percentage of high ROS cells. (B) Representative flow contour plots of KLSA9M pre-LSCs expressing high Lgr4 cDNA following antimycin A treatment illustrating the percentage of ROS cells. (C) Bar graph illustrating the percentage of viable cells following antimycin A treatment. Results from three independent experiments were graphed as mean ± SEM.

171

172

Figure 5.11 Lgr4 protects KLSA9M pre-LSCs from rotenone induced ROS damage KLSA9M pre-LSCs expressing Lgr4 cDNA or EV control were treated with 100 nM rotenone for 24 hours.(A) Representative flow contour plots of KLSA9M pre-LSCs expressing EV control following rotenone treatment illustrating the percentage of high ROS cells. (B) Representative flow contour plots of KLSA9M pre-LSCs expressing high Lgr4 cDNA following rotenone treatment illustrating the percentage of ROS cells. (C) Bar graph illustrating the percentage of viable cells following rotenone treatment. Results from three independent experiments were graphed as mean ± SEM.

173

5.11 Summary In Chapters 3 and 4, we established Lgr4 as a key player in β-catenin-driven AML LSCs, and identified a Wnt3a/Rspo3-Lgr4-Gαq-β-catenin signalling pathway that governs the self-renewal and proliferation of pre-LSCs and LSCs. In this chapter, we attempted to further understand the mechanism of how this signalling pathway governs LSC function. Using microarray expression profiling, we identified Gadd45a as a major target gene regulated by both Lgr4 and Gαq. Deletion of Gadd45a in MLL-mediated AML LSCs significantly enhanced self-renewal and proliferation, and generated more highly aggressive AML in vivo. Furthermore, serial transplantation and limiting dilution assays revealed that Gadd45a deletion significantly increased the frequency of leukaemia initiating cells. Hence Gadd45a plays a crucial tumour suppressive role in MLL-AML. Using microarray expression profiling, mtDNAs were also identified among the common target genes regulated by Lgr4 and Gαq. Bioenergenic analysis revealed that Lgr4 signalling regulates mitochondrial OXPHOS and ATP production in pre-LSC. Additionally, mitochondrial ROS production was regulated by Lgr4 and its downstream effectors. Lastly, pre-LSCs were vulnerable to mitochondrial complex I and complex III inhibition, which increased ROS production and reduced cell viability. However, overexpression of Lgr4 reversed the ROShigh profile and rescued cell death induced by antimycin A or rotenone, suggesting that Lgr4 protects pre-LSCs from ROS damage. Together, these data indicate that Lgr4 and its downstream signalling components regulate LSC activity by modulating the mitochondrial-driven energy metabolism of the cell.

174

6 Discussion and future directions

175

Despite the advances in chemotherapy and treatment options for other types of leukaemia, treatment options for AML have remained unchanged for over 30 years and the prognosis of patients diagnosed with AML is far from optimistic. The main cause is frequent disease relapse due to the inability of conventional treatments to eliminate all leukaemic cells. The persistence of LSCs following treatment is believed to be responsible for relapse due to their quiescence nature and high intrinsic drug-resistance [105,234]. Recent studies have highlighted a crucial role for β-catenin in the development and survival of AML LSCs [104,105]. β-catenin is not absolutely required for adult HSC survival and proliferation thereby presenting a window of opportunity for therapeutic targeting [87,88,90]. However, targeting β-catenin directly via small molecule inhibitors proved to be very challenging due to the extensive binding surface with its partners [133]. Moreover, previous data generated by our laboratory showed that β-catenin upstream Wnt regulators are not present in AML LSCs, suggesting alternative pathways involved in the activation of β-catenin [104]. Here, we identified a Wnt3a/Rspo3-Lgr4-Gαq-β-catenin signalling axis that governs LSC survival and provide evidence to suggest that targeting various aspects of this signalling axis may be of promising therapeutic value for the eradication of AML LSCs.

6.1 The interplay between Wnt3a, Rspo2/3 and Lgr4 The Rspo family is potent enhancers of Wnt/β-catenin signalling and stem cell growth factors [235,237,243]. The recent identification of Rspo1-4 as ligands of the Lgr4 receptor has stimulated immense interest in the role of Lgr4 in human disease [210,214,215]. Although Rspo/Lgr4 has been reported to play an important role in colon and lung cancer [217,218,304], no study has evaluated their importance in AML to date. Here, we showed that AML LSCs express high levels of Lgr4 and high Lgr4 expression is associated with an unfavourable clinical outcome whereas the expression of Lgr5, a close homolog of Lgr4, has no significant association with clinical outcome. It is interesting to note that Rspo1-4 can bind to both Lgr4 and 5 to activate β-catenin in HEK293 cells and that both Lgr4 and 5 are stem cells markers for the β-catenin driven

176

colon and mammary stem cells [210,211,213,305]. Lgr4 and 5 have also been implicated in aberrant Wnt/β-catenin signalling in colon cancer [218,306]. Given the functional similarities between these two receptors, it is interesting that only Lgr4 is associated with AML survival, a disease likewise characterised by aberrant Wnt/β- catenin signalling. This suggests that AML development is driven by a unique gene signature and signalling axes compared to other types of cancer, which is supported by several recent studies [104,115].

Our observation that Rspo1-4 cannot activate β-catenin expression in AML LSCs is unexpected, given that Rspo1-4 are bona fide Lgr4 ligands [210,214,215] and Lgr4 is highly expressed in AML LSCs. One possible explanation is that the signal from Lgr4 to β-catenin is already saturated, as shown by the high level of β-catenin expression in these cells. Another possibility is the requirement for the presence of the Wnt ligand, Wnt3a, as previous studies have demonstrated that the maximum level of β-catenin activation, caused by Rspo, is dependent on the concentration of Wnt3a [214]. Interestingly, our data showed that the expression of β-catenin is substantially increased only when Wnt3a is combined with Rspo2 or Rspo3. This highly synergistic relationship between Wnt3a and Rspo has been reported in numerous previous findings [210,214-216]. This suggests that LSCs may maintain adequate β-catenin signalling for their survival by relying on a very low presence of ligands. This is especially advantageous in a harsh environment where there is limited amount of ligands thereby enabling a drug resistant phenotype. Our findings are also in line with a recent study, where RNA-seq analysis was performed on primary human colon tumours highlighting recurrent Rspo2 and Rspo3 fusions accompanied by high Lgr4 expression [217], suggesting that Rspo2/3-Lgr4 signalling is possibly involved in colon cancer development. Since aberrant β-catenin signalling is also a defining feature in colon CSCs [131], these results together suggest that Rspo2/3-Lgr4 signalling may be important in colon CSCs. Therefore, future studies to investigate the involvement of Rspo2/3-Lgr4 in other β-catenin driven CSCs such as colon CSCs is warranted.

177

Despite the fact that Wnt3a and Rspo synergise to activate β-catenin through Lgr4, the molecular mechanism of their interactions are poorly understood. Recent studies have suggested that the cell-surface transmembrane E3 ubiquitin ligase zinc and ring finger 3 (Znrf3) and its homologue ring finger 43 (Rnf43) are involved in Rspo-mediated Wnt activation [307-309]. Znrf3 and Rnf43 negatively regulate Wnt signalling by interacting with the Wnt receptor complex and promoting degradation of the Wnt receptor Frizzled and LRP6 [307]. Rspo interacts with Znrf3 to promote Lgr4-mediated clearance of Znrf3 thereby enhances Wnt signalling [308,309]. These studies suggest potential interactions among Rspo, Lgr4, Znrf3 and Frizzled, and further investigation is needed to elucidate their involvement in AML LSCs. In addition, our findings that Wnt3a/Rspo2/3-Lgr4 enhances β-catenin and proliferation in AML LSCs also provide a rationale for future development of small molecule or antibody based inhibition of Lgr4 signalling. The crystal structures of Rspo1 binding to extracellular domain of Lgr4 and the complex of Znrf3-Rspo1 were recently solved [310,311], providing us with invaluable insight into their interactions and revealing key binding areas so that antibodies or small molecules could be rationally designed in order to block their interactions.

6.2 Lgr4 is crucial for β-catenin activation in LSCs and AML disease reconstitution Since it was discovered that Rspo is a ligand for Lgr4, which activates β-catenin signalling, several studies have investigated the role of Lgr4 in human diseases such as cancer [304,312,313]. Although this study is the first to evaluate the role of Lgr4 in AML, recent studies have suggested a tumorigenic role for Lgr4 in prostate, lung and colon cancers [217,218,304,312]. Interestingly, prostate, lung and colon cancer have all been shown to be associated with aberrant β-catenin signalling [314]. Hence, our finding that Lgr4 is essential for AML LSCs and promotes AML leukaemogenesis is in agreement with the fact that aberrant β-catenin signalling drives AML [104,105]. A recent study using whole genome sequencing identified a rare nonsense mutation within

178

the Lgr4 gene [313]. Styrkarsdottir et al. found that this nonsense mutation is associated with an increased risk of squamous cell carcinoma of the skin and cancers of the gallbladder and biliary tract, suggesting a tumour suppressor role for Lgr4 [313]. However, these cancers have not been shown to be associated with dysregulation of β- catenin signalling, thus, Lgr4 may signal via different pathways to regulate tumorigenesis in these cancer types. This suggests a context dependent role of Lgr4 in different types of cancer. Studies using Lgr4 knockout mice have shown that loss of Lgr4 results in a wide range of developmental defects that are not associated with β- catenin [222,225,315], suggesting that Lgr4 may also regulates β-catenin-independent developmental pathways in a cellular context manner.

AML is a heterogeneous disease and studies have shown that both HSCs and the more mature GMPs, when transformed by the MLL-AF9 oncogene, can give rise to AML in vivo [100,104,245]. The question of the leukaemia cell-of-origin has led to much discussion about how it accounts for the observed clinical heterogeneity and drug response in AML. In Chapter 3, we have shown that loss of Lgr4 severely impaired the development of LSCs and AML reconstitution derived from KLSMLL pre-LSCs and resulted in subsequent inability of LSCs to develop AML in secondary transplantations, whereas showing less effect on GMPMLL derived LSCs. This indicates that KLSMLL LSCs are more reliant on the Lgr4/β-catenin signalling pathway to drive their aggressiveness. A recent study by Krivtsov et al. demonstrated that, despite KLSMLL- derived and GMPMLL-derived LSCs having a similar immunophenotype, KLSMLL- derived LSCs are significantly more aggressive and chemo-resistant than GMPMLL- derived LSCs, and possessed a distinct gene expression profile that is also reflected in patient samples associated with poor prognosis [245]. Thus, KLSMLL LSCs may possess a gene expression profile that is more dependent on Lgr4 signalling compared to GMPMLL LSCs. For future investigations, it is worth comparing the gene expression profiles of these two LSC types to help explain the cellular origin-dependent effect of Lgr4-mediated signalling.

179

In addition to MLL transformed HSCs and GMPs, we have shown that Lgr4 also plays a significant role in Hoxa9/Meis1a transformed HSCs or KLSA9M. Although derived from HSCs, KLSA9M LSCs are much less aggressive compared to KLSMLL LSCs in part due to differential β-catenin expression [104,245]. Here, we demonstrated that overexpression of Lgr4 enhanced in vivo proliferation of KLSA9M pre-LSCs. Interestingly, examination of the FSC/SSC (forward/side scatter) profiles of these cells revealed a more blast-like phenotype in Lgr4 overexpression cells whereas in EV control the cells were more granulocytic (Suppl. Figure 1). This suggests that Lgr4 overexpression may block differentiation or reverse the mature phenotype of some leukaemic cells. Furthermore, we demonstrated that Lgr4 overexpression in KLSA9M pre-LSCs led to an increase in β-catenin expression. In vivo transplantation revealed a substantial decrease in AML onset and produced leukaemia with latencies slightly more aggressive than those in KLSMLL. This further supports our hypothesis that Lgr4 is essential in AML LSC development and the notion that Lgr4 plays a more prominent role in HSC-derived AML compared to GMP-derived AML. Hence, determining the leukaemia cell-of-origin in AML patients in the future may be of clinical importance in order to tailor the treatment regimen and improve patient response.

To further assess the oncogenic potential of Lgr4, Lgr4 was ectopically overexpressed in HSCs. However, overexpression of Lgr4 alone did not enhance the self-renewal capacity of HSCs in vitro, consistent with the “two-hit hypothesis” model of leukaemogenesis, which indicates that at least two different genetic aberrations are required for malignant transformation. In our mouse model of AML, the MLL-AF9 oncogene serves as the first hit that transform HSCs or GMPs into pre-LSCs. When pre- LSCs are transplanted into mice, they acquire a second hit before transforming into full blown LSCs. This may suggest that Lgr4 alone is incapable of transforming HSCs and may require additional genetic modifications. Future studies would determine whether Wnt3a/Rspo3-Lgr4-Gαq-β-catenin signalling affects the biology of normal HSCs, for instance, the effect of co-treatment of Rspo2/3 and Wnt3a on Lgr4/Gαq in HSCs and the effect of Lgr4 on in vivo self-renewal of HSCs.

180

Previous studies have shown that Hoxa9/Meis1a transduced GMPs or GMPA9M cannot cause leukaemia, possibly due to low levels of β-catenin expression [104]. Nevertheless, overexpression of β-catenin in GMPA9M gives rise to AML in vivo similar to that of KLSA9M [104]. We have shown that Lgr4 overexpression enhanced colony forming ability of GMPA9M. However, when transplanted into mice, Lgr4 overexpression in GMPA9M did not significantly affect mouse survival. This coincides with our discussion above that the effect of Lgr4 signalling depends on the leukaemia cell-of- origin. Furthermore, the level of Lgr4/β-catenin expression may not be sufficient for leukaemic transformation in these cells. Since Wnt/β-catenin is one of the main regulators of HSCs self-renewal [84-86], HSCs will have a baseline level of β-catenin expression that is higher than more mature progenitor cells such as GMPs. This may explain why Hoxa9/Meis1a could transform HSCs but not GMPs; and further activation of β-catenin via Lgr4 in GMPA9M may require additional factors not regulated by Hoxa9/Meis1a. As described previously, β-catenin activation in LSCs requires co- treatment of Rspo2/3 and Wnt3a. Hence, co-expression of Rspo2/3 and Wnt3a along with Lgr4 may be needed to fully transform GMPA9M and is worth investigating further.

6.3 Gαq is the G protein messenger relaying signals between Lgr4 and β- catenin Our microarray data indicated that the negative regulator of G protein signalling Rgs1 is a downstream component of Lgr4 signalling. Previous studies have demonstrated that Rgs2 expression is significantly reduced in AML with FLT3 ITD mutations, which impart a particularly poor prognosis [207]. Rgs2 has been shown to play a similar role in the cellular environment to Rgs1, its closest family member [199,268]. It has also been shown that overexpression of Rgs2 in FLT3 ITD expressing AML cells inhibits cell proliferation and clonal growth, counteracts differentiation block induced by FLT3 ITD, and impairs leukaemic transformation [207]. Therefore, our finding that Lgr4

181

downregulates Rgs1 and enhances leukaemogenesis is in line with the role of Rgs2 in AML with FLT3 ITD mutations.

To further investigate the functional role of Rgs1 in LSCs, we initially conducted shRNA-mediated stable knockdown of Rgs1 in KLSMLL pre-LSCs. However, knockdown was not successful despite using 3 independent Rgs1 shRNAs (results not shown). Since Rgs proteins are highly conserved throughout evolution and Rgs1 can interact with and regulate Gαi and Gαq to suppress downstream signalling [268,269], we postulated that Gαi or Gαq might be involved in regulating MLL-AML. Using inhibitor treatment and shRNA-mediated knockdown, Gαq, but not Gαi, was found to mediate the signalling transduction from Wnt3/Rspo3-Lgr4 to β-catenin and regulate the self-renewal ability of KLSMLL pre-LSCs. The selective Gαq inhibitor GP- antagonist 2A we tested is the only commercially available Gαq inhibitor on the market. It is an 11 amino acid peptide that is designed based on the neuropeptide substance P and it selectively inhibits Gαq signalling by interacting with the G protein but not the receptor [277,278]. Its relative large molecular weight and inefficiency to penetrate cell membrane has limited its use in medical research [279-281]. We have shown that it is capable of inhibiting β-catenin expression and the growth of pre-LSCs. However, the toxicity of this drug was not tested in normal HSCs, hence we cannot exclude the possibility that the drug effect on LSCs might be due to widespread haematopoietic cell toxicity. Hence future treatment of normal cells with these drugs is needed in order to exclude the toxicity for the development of therapeutic strategies. Furthermore, future experiments could be conducted to investigate the functional role of Rgs1 in AML LSCs using other gene manipulation methods such as the CRISPR/Cas9 system, and to examine whether Rgs1 directly interact with Gαq protein by co-immunoprecipitation assay.

Despite Gαq being identified as an important player in governing MLL pre-LSCs, other Gα proteins such as Gαs might also play a role in regulating β-catenin and self-renewal. Since we did not have an effective tool to examine a direct interaction between Rgs1 and Gαq, we also examined whether Gαs, another important class of Gα protein,

182

regulates β-catenin and proliferation of MLL pre-LSCs. Interestingly, inhibition of Gαs by a selective inhibitor in a range of concentrations failed to induce any difference in β- catenin expression or LSC proliferation (results not shown). Therefore, Gαq may exclusively mediate Lgr4/β-catenin signalling in AML LSCs. This is supported by our finding that Gαq knockdown completely abolished the effect of Wnt3a/Rspo3 on β- catenin activation. Interestingly, two of the research groups that first identified Rspo1-4 as ligands of Lgr4 observed that Lgr4 does not signal via G proteins in HEK293T cells [214,216]. However, other groups have shown that a constitutively active form of Lgr4 induced by a point mutation, T755I, can significantly increase intracellular cAMP levels in HEK293T cells, indicating Gαs activation [221,230]. These contradictory results in HEK293T cells along with our findings in AML pre-LSCs, suggest a context dependent mechanism of Lgr4-mediated signalling.

6.4 Tumour suppressor Gadd45a affects LSC development To further understand this Lgr4/Gαq/β-catenin signalling axis, we investigated the mechanisms of how this signalling axis governs LSC function. Our microarray has identified Gadd45a to be common downstream target of Lgr4 and Gαq. It is worth noting that Gadd45a has been reported to be downregulated in human AML [120,289,316]. This study is the first to demonstrate that Gadd45a downregulation is functionally involved in leukaemogenesis. Our data showed that Gadd45a knockout enhances LSC frequency and increases leukaemogenesis. This finding agrees with a recent study which showed Gadd45a regulates HSC stress response in mice and that Gadd45a knockout in HSCs exhibits more DNA damage accumulation, reduces the ability to induce apoptosis and promotes leukaemogenesis [291]. Furthermore, DNA hypermethylation of Gadd45a promoter is present in a large proportion of AML patients and predicts poor overall survival [289], indicating the importance of Gadd45a inhibition in AML leukaemogenesis.

183

The downregulation of Gadd45a and hypermethylation of its promoter suggest a tumour suppressor of Gadd45a in AML LSCs, hence strategies to induce its expression could have therapeutic potential. Several studies have demonstrated the potential therapeutic value of inducing Gadd45a expression. Adenoviral-mediated delivery of the Gadd45a gene resulted in chemosensitivity in pancreatic ductal adenocarcinoma when treated with chemotherapeutics such as etoposide, cisplatin and 5-fluorouracil [317]. Overexpression of Gadd45a in prostate cancer cell lines caused chemosensitisation to docetaxel, a conventional chemotherapeutic agent for multiple cancers [287]. Furthermore, multidrug resistant osteosarcoma has been linked to the failure to induce endogenous Gadd45a expression by chemotherapeutic agents such as paclitaxel and doxorubicin, and ectopic expression of Gadd45a sensitised osteosarcoma cells to these agents [318]. In addition to genetic manipulations to increase Gadd45a expression, several agents have been shown to induce the expression of Gadd45a. For example, decitabine, a DNA demethylation agent used for treating myelodysplastic syndromes, has been shown to demethylate Gadd45a promoter to induce Gadd45a expression and apoptosis of osteosarcoma cells [319]. Interestingly, decitabine has recently been approved by the European Commission for treating high-risk elderly AML patients as a single agent, and more clinical trials are currently underway to evaluate its use in combination with other chemotherapeutics and in more defined subgroup of AML patients. As mentioned above, DNA hypermethylation of Gadd45a is present in a large proportion of AML patients with poor clinical outcome [289]; hence the effect of decitabine could be partly due to its induction of Gadd45a expression. Future studies could investigate the effect of Gadd45a overexpression in LSCs, decitabine treatment on Gadd45a methylation and evaluate if decitabine treatment could have any effect on LSC activity.

184

6.5 Lgr4 as a master regulator of mitochondrial energy generation and ROS production in pre-LSCs To further investigate the mechanisms of how Lgr4 signalling cascade governs AML LSC activity, our microarray data revealed that several mtDNA-coded subunits of OXPHOS complexes are regulated by both Lgr4 and Gαq. Mitochondrial complex I and III are indispensible for the normal function of the respiratory chain that is required for supplying metabolic energy (ATP) [293-295]. In addition, an increasing amount of findings have demonstrated that many oncogenic signalling pathways converge to regulate cancer cell metabolism in order to facilitate their survival and proliferation [320]. This prompted us to further investigate the relationship between Lgr4 and OXPHOS in pre-LSCs. Our data revealed that loss of Lgr4 or Gαq significantly impairs mitochondrial OXPHOS and ATP production. Consistently, a recent study has discovered that LSCs in AML are highly reliant on mitochondrial OXPHOS for their energy production and survival, and are deficient in the ability to efficiently utilise the glycolytic pathway [292]. This is contrary to normal HSCs which efficiently utilise glycolysis instead of OXPHOS [321]. Interestingly, glioma stem cells have also been shown to be reliant on OXPHOS [322], suggesting that dependence on OXPHOS might be a broad characteristic of CSCs. Since OXPHOS is a slower but more efficient way to generate energy compared to glycolysis, LSCs may benefit from this highly efficient mechanism in a nutrient deficient tumour microenvironment.

Mitochondrial complex I and III are the main sites for electron transfer during OXPHOS, hence are prone to electron leakage which results in the production of ROS such as superoxide [323-325]. ROS can then oxidise proteins and cause DNA mutations which may contribute to disease initiation and progression [326]. Although harmful to the cells in excess amount, ROS plays a key role in the intrinsic apoptotic pathway for regulating cell survival [323,325,327]. A recent study by Lagadinou et al. showed that LSCs from AML patients are characterised by relatively low levels of ROS [292]. They also demonstrated that ROSlow population generated more colonies and had significantly higher levels of engraftment in mouse bone marrow after transplantation compared to

185

the ROShigh counterpart [292]. Consistent with their finding, we showed that Lgr4 overexpression or Gadd45a deletion decreased ROS production and enhanced leukaemogenesis in vivo, indicating mitochondrial energy regulation in the stemness- promoting ability of the Lgr4 signalling cascade.

To further investigate the effect of mitochondrial stress in pre-LSCs, we used the mitochondrial complex I inhibitor rotenone and complex III inhibitor antimycin A, both of which have been shown to increase ROS production [299-303]. Rotenone works by inhibiting electron transfer from complex I to ubiquinone thereby accumulating electrons within the mitochondrial matrix which results in cellular oxygen been reduced to radicals [302,303]. Antimycin A works similarly by binding to complex III thereby inhibiting the oxidation of ubiquinone [299-301]. We showed that low concentrations of antimycin A or rotenone increased ROS production and impaired cell survival in pre- LSCs, indicating that these cells are vulnerable to mitochondrial stress such as excessive ROS production. Hence, strategies to increase ROS may be a valuable approach to eliminate LSCs. In fact, several chemotherapeutic agents that are currently used in the clinic induce ROS production. For example, paclitaxel, vincristine and anti- folates interfere with the electron transport chain, resulting in high ROS production [328]. However, electron transport chain is essential for normal cell function, thus targeting it may elicit high toxicity. Cancer cells frequently have increased expression of antioxidant to counteract their relatively high ROS profile due to metabolic and signalling aberrations [329]. Strategies to inhibit antioxidant may tip this balance and result in cancer cell death. Several recent studies have exploited the idea of inhibiting antioxidant and identified a number of small molecule inhibitors that eliminate cancer cells through targeting the glutathione antioxidant system [330-332]. Therefore, future studies could investigate whether pre-LSCs are sensitive to antioxidant inhibition and examine whether these small molecule inhibitors mentioned above can be used to target pre-LSCs.

Lastly, we showed that overexpression of Lgr4 reversed the effect of antimycin A or rotenone on ROS production and cell viability, suggesting that Lgr4 signalling protects

186

pre-LSCs/LSCs from ROS induced damage. The exact mechanisms of how the Lgr4 signalling cascade protects pre-LSCs remains to be investigated. We postulate that Lgr4 possibly utilises antioxidant systems to counterbalance excessive ROS production. Interestingly, our microarray data revealed that Txndc15 (Thioredoxin domain- containing protein 15) is one of the most differentially expressed common target genes regulated by both Lgr4 and Gαq. Txndc15 belongs to the thioredoxin interacting protein which plays an important role in redox regulation and defence against oxidative stress [333-336]. It would be interesting to investigate if Lgr4 can employ the thioredoxin ROS scavenger system to prevent pre-LSCs/LSCs from oxidative damage. Therefore, future studies could evaluate the role of antioxidant system in governing pre-LSC/LSC survival and investigate the therapeutic potential of inhibiting this mechanism.

6.6 Conclusions and future perspectives In summary, this study has identified for the first time that Lgr4 is an important regulator of β-catenin-driven AML LSCs. We demonstrated that LSCs have significantly elevated Lgr4 expression levels compared to normal haematopoietic stem cells and high Lgr4 expression correlated with poor patient survival. We also identified that the specific combination of Wnt3a and Rspo2/3 is capable of enhancing β-catenin activation and self-renewal through Lgr4. Knockdown of Lgr4 markedly reduced β- catenin expression, significantly impaired LSC self-renewal. Most strikingly, knockdown of Lgr4 severely impaired the development and AML in vivo and the loss of Lgr4 resulted in inability of subsequent LSCs to develop AML in secondary transplantations. Conversely, Lgr4 overexpression increased β-catenin, enhanced self- renewal, and substantially shortened AML disease onset. These findings suggest that Lgr4 is a critical regulator of β-catenin and a potential anti-LSC drug target.

Microarray gene expression profiling revealed that Rgs1 is a major downstream effector of Lgr4. We determined that the G protein, Gαq, is a critical component of Rgs1/Lgr4- mediated regulation of LSC activity including self-renewal and proliferation ability.

187

Lgr4/Gαq knockdown abolished the effect of Wnt3a/Rspo3 potentiated β-catenin activation, implicating a Wnt3a/Rspo3-Lgr4-Gαq-β-catenin signalling cascade in AML LSCs.

Our microarray data also revealed that Lgr4 signalling regulates LSCs via inhibition of Gadd45a and several mitochondrial associated genes including complex III subunit Cytb and complex I subunits Nd2 and Nd4l. Loss of Gadd45a significantly enhanced MLL-AML progression in primary transplantations, exhibited gradually increased ability to reduce disease latency in secondary and tertiary transplantations and resulted in a 10-fould increase in LSC frequency in the bone marrow. Hence Gadd45a plays a crucial tumour suppressor role in MLL-AML. Through oxidative phosphorylation, mitochondria play an essential role in the supply of metabolic energy (ATP) to the cell. Consistent with our microarray data, bioenergetic analysis showed that Lgr4 knockdown significantly reduced the rate of oxygen consumption and the concentration of basal ATP. Additionally, mitochondrial ROS production was modulated by Lgr4 and its downstream effectors. Lastly, by using mitochondrial complex inhibitors we demonstrated that Lgr4 protects LSCs against oxidative stress. These data indicate that Lgr4 and its downstream signalling components regulate LSC activity possibly by modulating the mitochondrial-driven energy metabolism of the cell.

As discussed previously, there are multiple aspects of this study that need further investigation in order to better delineate the Lgr4 signalling pathway and to fully explore its therapeutic potential. Our microarray analysis revealed that Rgs1 is a major downstream effector of Lgr4. However, due to technical limitation, we could not establish a direct interaction between Lgr4 and Rgs1. Hence it’s not clear whether Rgs1 is a direct downstream target of Lgr4 or there may be other proteins involved. Similarly, the inhibitory action of Rgs1 on Gαq is based on previous studies and further characterisation is needed to confirm their relationship in AML LSCs. We have mainly focused on using pre-LSCs in elucidating the mechanisms of Lgr4 signalling as these cells are the starting points for multistep pathogenesis of AML and development of LSCs. Studies have also demonstrated a strong association between the persistence of

188

pre-LSCs and patient relapse [337]. Nonetheless, further experimentations should also be conducted on LSCs to examine if they function the same way as their precursors. Furthermore, the exact mechanisms of how Lgr4 regulate ROS production still remain to be evaluated. We propose that Lgr4 may employ the thioredoxin ROS scavenger system to prevent pre-LSCs from oxidative damage as discussed above, and further characterisation of this mechanism may unveil new avenues for therapeutic applications. Lastly, all of our functional studies were conducted using murine cells. Despite the fact that our AML mouse model is very well characterised and highly recapitulates human AML [100,104], it is imperative to test our findings in other systems such as human xenograft mouse models and primary human LSCs.

Collectively, our evidence suggests a complex Lgr4 signalling system utilised by AML LSCs for their survival and proliferation, as illustrated in Figure 6.1. This Lgr4 signalling network provides a platform for further in depth characterisation of AML LSCs and interference with components of this pathway may represent a promising therapeutic approach for eradicating LSCs in AML.

189

190

Figure 6.1 Schematic diagram of Lgr4 signalling system in LSCs Lgr4 signalling activation in LSCs requires the presence of both RSPO2/3 and Wnt3a. Lgr4 activation results in inhibition of Rgs1, which may activate β-catenin signalling through Gαq. Activation of β-catenin leads to increased cell proliferation and reduced apoptosis/differentiation. Lgr4 and Gαq both decrease Gadd45a expression, which leads to enhanced cell proliferation and LSC frequency. Furthermore, Lgr4 signalling enhances mitochondrial energy generation, reduces ROS production and protects cells from ROS induced damage, resulting in improved cell survival.

191

Suppl. Figure 1 Overexpression of Lgr4 induces a more immature phenotype in KLSA9M

Representative flow scatter plots of bone marrow cells from primary recipient mice harvested 30 days after inoculation of KLSA9M pre-LSCs expressing Lgr4 cDNA or EV control.

192

7 References

193

1. World Health Organisation (2014) World Cancer Report 2014; Stewart BW, Wild CP, editors. 2. Australian Institute of Health and Welfare (2014) Cancer in Australia: an overview 2014. 3. Shipley JL, Butera JN (2009) Acute myelogenous leukemia. Exp Hematol 37: 649- 658. 4. Bloomfield CD, Brunning RD (1985) FAB M7: acute megakaryoblastic leukemia-- beyond morphology. Ann Intern Med 103: 450-452. 5. Renneville A, Roumier C, Biggio V, Nibourel O, Boissel N, et al. (2008) Cooperating gene mutations in acute myeloid leukemia: a review of the literature. Leukemia 22: 915-931. 6. Mrozek K, Heerema NA, Bloomfield CD (2004) Cytogenetics in acute leukemia. Blood Rev 18: 115-136. 7. Falini B, Tiacci E, Martelli MP, Ascani S, Pileri SA (2010) New classification of acute myeloid leukemia and precursor-related neoplasms: changes and unsolved issues. Discov Med 10: 281-292. 8. Grimwade D, Walker H, Oliver F, Wheatley K, Harrison C, et al. (1998) The importance of diagnostic cytogenetics on outcome in AML: analysis of 1,612 patients entered into the MRC AML 10 trial. The Medical Research Council Adult and Children's Leukaemia Working Parties. Blood 92: 2322-2333. 9. Frohling S, Scholl C, Gilliland DG, Levine RL (2005) Genetics of myeloid malignancies: Pathogenetic and clinical implications. Journal of Clinical Oncology 23: 6285-6295. 10. Mrozek K, Bloomfield CD (2006) Chromosome aberrations, gene mutations and expression changes, and prognosis in adult acute myeloid leukemia. Hematology Am Soc Hematol Educ Program: 169-177. 11. Scholl S, Fricke HJ, Sayer HG, Hoffken K (2009) Clinical implications of molecular genetic aberrations in acute myeloid leukemia. J Cancer Res Clin Oncol 135: 491-505. 12. Byrd JC, Ruppert AS, Mrozek K, Carroll AJ, Edwards CG, et al. (2004) Repetitive cycles of high-dose cytarabine benefit patients with acute myeloid leukemia and inv(16)(p13q22) or t(16;16)(p13;q22): results from CALGB 8461. J Clin Oncol 22: 1087-1094. 13. Byrd JC, Dodge RK, Carroll A, Baer MR, Edwards C, et al. (1999) Patients with t(8;21)(q22;q22) and acute myeloid leukemia have superior failure-free and overall survival when repetitive cycles of high-dose cytarabine are administered. J Clin Oncol 17: 3767-3775. 14. Tallman MS, Andersen JW, Schiffer CA, Appelbaum FR, Feusner JH, et al. (1997) All-trans-retinoic acid in acute promyelocytic leukemia. New England Journal of Medicine 337: 1021-1028. 15. Gulley ML, Shea TC, Fedoriw Y (2010) Genetic tests to evaluate prognosis and predict therapeutic response in acute myeloid leukemia. J Mol Diagn 12: 3-16. 16. de The H, Chen Z (2010) Acute promyelocytic leukaemia: novel insights into the mechanisms of cure. Nat Rev Cancer 10: 775-783.

194

17. Patel JP, Gonen M, Figueroa ME, Fernandez H, Sun Z, et al. (2012) Prognostic relevance of integrated genetic profiling in acute myeloid leukemia. N Engl J Med 366: 1079-1089. 18. Falini B, Martelli MP (2011) NPM1-mutated AML: targeting by disassembling. Blood 118: 2936-2938. 19. Falini B, Nicoletti I, Martelli MF, Mecucci C (2007) Acute myeloid leukemia carrying cytoplasmic/mutated nucleophosmin (NPMc+ AML): biologic and clinical features. Blood 109: 874-885. 20. Schlenk RF, Dohner K, Krauter J, Frohling S, Corbacioglu A, et al. (2008) Mutations and treatment outcome in cytogenetically normal acute myeloid leukemia. N Engl J Med 358: 1909-1918. 21. Ghanem H, Tank N, Tabbara IA (2012) Prognostic implications of genetic aberrations in acute myelogenous leukemia with normal cytogenetics. Am J Hematol 87: 69-77. 22. Green CL, Koo KK, Hills RK, Burnett AK, Linch DC, et al. (2010) Prognostic significance of CEBPA mutations in a large cohort of younger adult patients with acute myeloid leukemia: impact of double CEBPA mutations and the interaction with FLT3 and NPM1 mutations. J Clin Oncol 28: 2739-2747. 23. Lin LI, Chen CY, Lin DT, Tsay W, Tang JL, et al. (2005) Characterization of CEBPA mutations in acute myeloid leukemia: most patients with CEBPA mutations have biallelic mutations and show a distinct immunophenotype of the leukemic cells. Clin Cancer Res 11: 1372-1379. 24. Kikushige Y, Yoshimoto G, Miyamoto T, Iino T, Mori Y, et al. (2008) Human Flt3 is expressed at the hematopoietic stem cell and the granulocyte/macrophage progenitor stages to maintain cell survival. J Immunol 180: 7358-7367. 25. Piacibello W, Fubini L, Sanavio F, Brizzi MF, Severino A, et al. (1995) Effects of human FLT3 ligand on myeloid leukemia cell growth: heterogeneity in response and synergy with other hematopoietic growth factors. Blood 86: 4105-4114. 26. Drexler HG, Quentmeier H (2004) FLT3: receptor and ligand. Growth Factors 22: 71-73. 27. Gotze KS, Ramirez M, Tabor K, Small D, Matthews W, et al. (1998) Flt3high and Flt3low CD34+ progenitor cells isolated from human bone marrow are functionally distinct. Blood 91: 1947-1958. 28. Stirewalt DL, Radich JP (2003) The role of FLT3 in haematopoietic malignancies. Nat Rev Cancer 3: 650-665. 29. Kiyoi H, Towatari M, Yokota S, Hamaguchi M, Ohno R, et al. (1998) Internal tandem duplication of the FLT3 gene is a novel modality of elongation mutation which causes constitutive activation of the product. Leukemia 12: 1333-1337. 30. Hayakawa F, Towatari M, Kiyoi H, Tanimoto M, Kitamura T, et al. (2000) Tandem-duplicated Flt3 constitutively activates STAT5 and MAP kinase and introduces autonomous cell growth in IL-3-dependent cell lines. Oncogene 19: 624-631.

195

31. Ozeki K, Kiyoi H, Hirose Y, Iwai M, Ninomiya M, et al. (2004) Biologic and clinical significance of the FLT3 transcript level in acute myeloid leukemia. Blood 103: 1901-1908. 32. Abu-Duhier FM, Goodeve AC, Wilson GA, Gari MA, Peake IR, et al. (2000) FLT3 internal tandem duplication mutations in adult acute myeloid leukaemia define a high-risk group. Br J Haematol 111: 190-195. 33. Buonamici S, Chakraborty S, Senyuk V, Nucifora G (2003) The role of EVI1 in normal and leukemic cells. Blood Cells Mol Dis 31: 206-212. 34. Yuasa H, Oike Y, Iwama A, Nishikata I, Sugiyama D, et al. (2005) Oncogenic transcription factor Evi1 regulates hematopoietic stem cell proliferation through GATA-2 expression. EMBO J 24: 1976-1987. 35. Glass C, Wuertzer C, Cui X, Bi Y, Davuluri R, et al. (2013) Global Identification of EVI1 Target Genes in Acute Myeloid Leukemia. PLoS One 8: e67134. 36. Groschel S, Lugthart S, Schlenk RF, Valk PJ, Eiwen K, et al. (2010) High EVI1 expression predicts outcome in younger adult patients with acute myeloid leukemia and is associated with distinct cytogenetic abnormalities. J Clin Oncol 28: 2101-2107. 37. Balgobind BV, Lugthart S, Hollink IH, Arentsen-Peters ST, van Wering ER, et al. (2010) EVI1 overexpression in distinct subtypes of pediatric acute myeloid leukemia. Leukemia 24: 942-949. 38. Okuda T, van Deursen J, Hiebert SW, Grosveld G, Downing JR (1996) AML1, the target of multiple chromosomal translocations in human leukemia, is essential for normal fetal liver hematopoiesis. Cell 84: 321-330. 39. Ito Y (2004) Oncogenic potential of the RUNX gene family: 'overview'. Oncogene 23: 4198-4208. 40. Michaud J, Scott HS, Escher R (2003) AML1 interconnected pathways of leukemogenesis. Cancer Invest 21: 105-136. 41. Schnittger S, Dicker F, Kern W, Wendland N, Sundermann J, et al. (2011) RUNX1 mutations are frequent in de novo AML with noncomplex karyotype and confer an unfavorable prognosis. Blood 117: 2348-2357. 42. Krivtsov AV, Armstrong SA (2007) MLL translocations, histone modifications and leukaemia stem-cell development. Nat Rev Cancer 7: 823-833. 43. de Boer J, Walf-Vorderwulbecke V, Williams O (2013) In focus: MLL-rearranged leukemia. Leukemia 27: 1224-1228. 44. Schoch C, Schnittger S, Klaus M, Kern W, Hiddemann W, et al. (2003) AML with 11q23/MLL abnormalities as defined by the WHO classification: incidence, partner chromosomes, FAB subtype, age distribution, and prognostic impact in an unselected series of 1897 cytogenetically analyzed AML cases. Blood 102: 2395-2402. 45. Raimondi SC, Chang MN, Ravindranath Y, Behm FG, Gresik MV, et al. (1999) Chromosomal abnormalities in 478 children with acute myeloid leukemia: clinical characteristics and treatment outcome in a cooperative pediatric oncology group study-POG 8821. Blood 94: 3707-3716.

196

46. Mrozek K, Heinonen K, Lawrence D, Carroll AJ, Koduru PR, et al. (1997) Adult patients with de novo acute myeloid leukemia and t(9; 11)(p22; q23) have a superior outcome to patients with other translocations involving band 11q23: a cancer and leukemia group B study. Blood 90: 4532-4538. 47. Balgobind BV, Raimondi SC, Harbott J, Zimmermann M, Alonzo TA, et al. (2009) Novel prognostic subgroups in childhood 11q23/MLL-rearranged acute myeloid leukemia: results of an international retrospective study. Blood 114: 2489-2496. 48. Grimwade D, Hills RK, Moorman AV, Walker H, Chatters S, et al. (2010) Refinement of cytogenetic classification in acute myeloid leukemia: determination of prognostic significance of rare recurring chromosomal abnormalities among 5876 younger adult patients treated in the United Kingdom Medical Research Council trials. Blood 116: 354-365. 49. Balgobind BV, Zwaan CM, Pieters R, Van den Heuvel-Eibrink MM (2011) The heterogeneity of pediatric MLL-rearranged acute myeloid leukemia. Leukemia 25: 1239-1248. 50. Rollig C, Bornhauser M, Thiede C, Taube F, Kramer M, et al. (2011) Long-term prognosis of acute myeloid leukemia according to the new genetic risk classification of the European LeukemiaNet recommendations: evaluation of the proposed reporting system. J Clin Oncol 29: 2758-2765. 51. Yates JW, Wallace HJ, Jr., Ellison RR, Holland JF (1973) Cytosine arabinoside (NSC-63878) and daunorubicin (NSC-83142) therapy in acute nonlymphocytic leukemia. Cancer Chemother Rep 57: 485-488. 52. Cassileth PA, Harrington DP, Hines JD, Oken MM, Mazza JJ, et al. (1988) Maintenance chemotherapy prolongs remission duration in adult acute nonlymphocytic leukemia. J Clin Oncol 6: 583-587. 53. Dohner H, Estey EH, Amadori S, Appelbaum FR, Buchner T, et al. (2010) Diagnosis and management of acute myeloid leukemia in adults: recommendations from an international expert panel, on behalf of the European LeukemiaNet. Blood 115: 453-474. 54. Roboz GJ (2012) Current treatment of acute myeloid leukemia. Curr Opin Oncol 24: 711-719. 55. Eck MJ, Manley PW (2009) The interplay of structural information and functional studies in kinase drug design: insights from BCR-Abl. Curr Opin Cell Biol 21: 288-295. 56. Grunwald MR, Levis MJ (2013) FLT3 inhibitors for acute myeloid leukemia: a review of their efficacy and mechanisms of resistance. Int J Hematol 97: 683- 694. 57. Rubnitz JE, Gibson B, Smith FO (2010) Acute myeloid leukemia. Hematol Oncol Clin North Am 24: 35-63. 58. Guan Y, Gerhard B, Hogge DE (2003) Detection, isolation, and stimulation of quiescent primitive leukemic progenitor cells from patients with acute myeloid leukemia (AML). Blood 101: 3142-3149.

197

59. Ishikawa F, Yoshida S, Saito Y, Hijikata A, Kitamura H, et al. (2007) Chemotherapy-resistant human AML stem cells home to and engraft within the bone-marrow endosteal region. Nature Biotechnology 25: 1315-1321. 60. Till JE, Mc CE (1961) A direct measurement of the radiation sensitivity of normal mouse bone marrow cells. Radiat Res 14: 213-222. 61. Wu AM, Till JE, Siminovitch L, McCulloch EA (1967) A cytological study of the capacity for differentiation of normal hemopoietic colony-forming cells. J Cell Physiol 69: 177-184. 62. Becker AJ, Mc CE, Till JE (1963) Cytological demonstration of the clonal nature of spleen colonies derived from transplanted mouse marrow cells. Nature 197: 452- 454. 63. Jordan CT, Lemischka IR (1990) Clonal and systemic analysis of long-term hematopoiesis in the mouse. Genes Dev 4: 220-232. 64. Visser JW, Van Bekkum DW (1990) Purification of pluripotent hemopoietic stem cells: past and present. Exp Hematol 18: 248-256. 65. Spangrude GJ, Heimfeld S, Weissman IL (1988) Purification and characterization of mouse hematopoietic stem cells. Science 241: 58-62. 66. Baum CM, Weissman IL, Tsukamoto AS, Buckle AM, Peault B (1992) Isolation of a candidate human hematopoietic stem-cell population. Proc Natl Acad Sci U S A 89: 2804-2808. 67. Osawa M, Hanada K, Hamada H, Nakauchi H (1996) Long-term lymphohematopoietic reconstitution by a single CD34-low/negative hematopoietic stem cell. Science 273: 242-245. 68. Morrison SJ, Weissman IL (1994) The long-term repopulating subset of hematopoietic stem cells is deterministic and isolatable by phenotype. Immunity 1: 661-673. 69. Morrison SJ, Wandycz AM, Hemmati HD, Wright DE, Weissman IL (1997) Identification of a lineage of multipotent hematopoietic progenitors. Development 124: 1929-1939. 70. Kondo M, Weissman IL, Akashi K (1997) Identification of clonogenic common lymphoid progenitors in mouse bone marrow. Cell 91: 661-672. 71. Akashi K, Traver D, Miyamoto T, Weissman IL (2000) A clonogenic common myeloid progenitor that gives rise to all myeloid lineages. Nature 404: 193-197. 72. Kondo M, Wagers AJ, Manz MG, Prohaska SS, Scherer DC, et al. (2003) Biology of hematopoietic stem cells and progenitors: implications for clinical application. Annu Rev Immunol 21: 759-806. 73. Andersson ER, Sandberg R, Lendahl U (2011) Notch signaling: simplicity in design, versatility in function. Development 138: 3593-3612. 74. Milner LA, Kopan R, Martin DI, Bernstein ID (1994) A human homologue of the Drosophila developmental gene, Notch, is expressed in CD34+ hematopoietic precursors. Blood 83: 2057-2062. 75. Varnum-Finney B, Xu L, Brashem-Stein C, Nourigat C, Flowers D, et al. (2000) Pluripotent, cytokine-dependent, hematopoietic stem cells are immortalized by constitutive Notch1 signaling. Nat Med 6: 1278-1281.

198

76. Karanu FN, Murdoch B, Gallacher L, Wu DM, Koremoto M, et al. (2000) The notch ligand jagged-1 represents a novel growth factor of human hematopoietic stem cells. J Exp Med 192: 1365-1372. 77. Maillard I, Koch U, Dumortier A, Shestova O, Xu L, et al. (2008) Canonical notch signaling is dispensable for the maintenance of adult hematopoietic stem cells. Cell Stem Cell 2: 356-366. 78. Briscoe J, Therond PP (2013) The mechanisms of Hedgehog signalling and its roles in development and disease. Nat Rev Mol Cell Biol 14: 416-429. 79. Bhardwaj G, Murdoch B, Wu D, Baker DP, Williams KP, et al. (2001) Sonic hedgehog induces the proliferation of primitive human hematopoietic cells via BMP regulation. Nat Immunol 2: 172-180. 80. Zhao C, Chen A, Jamieson CH, Fereshteh M, Abrahamsson A, et al. (2009) Hedgehog signalling is essential for maintenance of cancer stem cells in myeloid leukaemia. Nature 458: 776-779. 81. Dierks C, Beigi R, Guo GR, Zirlik K, Stegert MR, et al. (2008) Expansion of Bcr- Abl-positive leukemic stem cells is dependent on Hedgehog pathway activation. Cancer Cell 14: 238-249. 82. Hofmann I, Stover EH, Cullen DE, Mao J, Morgan KJ, et al. (2009) Hedgehog signaling is dispensable for adult murine hematopoietic stem cell function and hematopoiesis. Cell Stem Cell 4: 559-567. 83. Gao J, Graves S, Koch U, Liu S, Jankovic V, et al. (2009) Hedgehog signaling is dispensable for adult hematopoietic stem cell function. Cell Stem Cell 4: 548- 558. 84. Reya T, Duncan AW, Ailles L, Domen J, Scherer DC, et al. (2003) A role for Wnt signalling in self-renewal of haematopoietic stem cells. Nature 423: 409-414. 85. Willert K, Brown JD, Danenberg E, Duncan AW, Weissman IL, et al. (2003) Wnt proteins are lipid-modified and can act as stem cell growth factors. Nature 423: 448-452. 86. Luis TC, Weerkamp F, Naber BA, Baert MR, de Haas EF, et al. (2009) Wnt3a deficiency irreversibly impairs hematopoietic stem cell self-renewal and leads to defects in progenitor cell differentiation. Blood 113: 546-554. 87. Zhao C, Blum J, Chen A, Kwon HY, Jung SH, et al. (2007) Loss of beta-catenin impairs the renewal of normal and CML stem cells in vivo. Cancer Cell 12: 528- 541. 88. Cobas M, Wilson A, Ernst B, Mancini SJ, MacDonald HR, et al. (2004) Beta- catenin is dispensable for hematopoiesis and lymphopoiesis. J Exp Med 199: 221-229. 89. Zheng X, Beissert T, Kukoc-Zivojnov N, Puccetti E, Altschmied J, et al. (2004) Gamma-catenin contributes to leukemogenesis induced by AML-associated translocation products by increasing the self-renewal of very primitive progenitor cells. Blood 103: 3535-3543. 90. Jeannet G, Scheller M, Scarpellino L, Duboux S, Gardiol N, et al. (2008) Long- term, multilineage hematopoiesis occurs in the combined absence of beta- catenin and gamma-catenin. Blood 111: 142-149.

199

91. Zhang J, Niu C, Ye L, Huang H, He X, et al. (2003) Identification of the haematopoietic stem cell niche and control of the niche size. Nature 425: 836- 841. 92. Essers MA, Offner S, Blanco-Bose WE, Waibler Z, Kalinke U, et al. (2009) IFNalpha activates dormant haematopoietic stem cells in vivo. Nature 458: 904- 908. 93. Sato T, Onai N, Yoshihara H, Arai F, Suda T, et al. (2009) Interferon regulatory factor-2 protects quiescent hematopoietic stem cells from type I interferon- dependent exhaustion. Nat Med 15: 696-700. 94. Miyamoto K, Araki KY, Naka K, Arai F, Takubo K, et al. (2007) Foxo3a is essential for maintenance of the hematopoietic stem cell pool. Cell Stem Cell 1: 101-112. 95. Tothova Z, Kollipara R, Huntly BJ, Lee BH, Castrillon DH, et al. (2007) FoxOs are critical mediators of hematopoietic stem cell resistance to physiologic oxidative stress. Cell 128: 325-339. 96. Morrison SJ, Prowse KR, Ho P, Weissman IL (1996) Telomerase activity in hematopoietic cells is associated with self-renewal potential. Immunity 5: 207- 216. 97. Lapidot T, Sirard C, Vormoor J, Murdoch B, Hoang T, et al. (1994) A cell initiating human acute myeloid leukaemia after transplantation into SCID mice. Nature 367: 645-648. 98. Bonnet D, Dick JE (1997) Human acute myeloid leukemia is organized as a hierarchy that originates from a primitive hematopoietic cell. Nat Med 3: 730- 737. 99. Kelly LM, Gilliland DG (2002) Genetics of myeloid leukemias. Annu Rev Genomics Hum Genet 3: 179-198. 100. Krivtsov AV, Twomey D, Feng Z, Stubbs MC, Wang Y, et al. (2006) Transformation from committed progenitor to leukaemia stem cell initiated by MLL-AF9. Nature 442: 818-822. 101. Somervaille TC, Cleary ML (2006) Identification and characterization of leukemia stem cells in murine MLL-AF9 acute myeloid leukemia. Cancer Cell 10: 257- 268. 102. Kirstetter P, Schuster MB, Bereshchenko O, Moore S, Dvinge H, et al. (2008) Modeling of C/EBPalpha mutant acute myeloid leukemia reveals a common expression signature of committed myeloid leukemia-initiating cells. Cancer Cell 13: 299-310. 103. Meyer C, Hofmann J, Burmeister T, Groger D, Park TS, et al. (2013) The MLL recombinome of acute leukemias in 2013. Leukemia 27: 2165-2176. 104. Wang Y, Krivtsov AV, Sinha AU, North TE, Goessling W, et al. (2010) The Wnt/beta-catenin pathway is required for the development of leukemia stem cells in AML. Science 327: 1650-1653. 105. Yeung J, Esposito MT, Gandillet A, Zeisig BB, Griessinger E, et al. (2010) beta- Catenin mediates the establishment and drug resistance of MLL leukemic stem cells. Cancer Cell 18: 606-618.

200

106. Reya T, Morrison SJ, Clarke MF, Weissman IL (2001) Stem cells, cancer, and cancer stem cells. Nature 414: 105-111. 107. Taussig DC, Miraki-Moud F, Anjos-Afonso F, Pearce DJ, Allen K, et al. (2008) Anti-CD38 antibody-mediated clearance of human repopulating cells masks the heterogeneity of leukemia-initiating cells. Blood 112: 568-575. 108. Taussig DC, Vargaftig J, Miraki-Moud F, Griessinger E, Sharrock K, et al. (2010) Leukemia-initiating cells from some acute myeloid leukemia patients with mutated nucleophosmin reside in the CD34(-) fraction. Blood 115: 1976-1984. 109. Gibbs KD, Jr., Jager A, Crespo O, Goltsev Y, Trejo A, et al. (2012) Decoupling of tumor-initiating activity from stable immunophenotype in HoxA9-Meis1-driven AML. Cell Stem Cell 10: 210-217. 110. Goardon N, Marchi E, Atzberger A, Quek L, Schuh A, et al. (2011) Coexistence of LMPP-like and GMP-like leukemia stem cells in acute myeloid leukemia. Cancer Cell 19: 138-152. 111. Monney L, Sabatos CA, Gaglia JL, Ryu A, Waldner H, et al. (2002) Th1-specific cell surface protein Tim-3 regulates macrophage activation and severity of an autoimmune disease. Nature 415: 536-541. 112. Zhu C, Anderson AC, Schubart A, Xiong H, Imitola J, et al. (2005) The Tim-3 ligand galectin-9 negatively regulates T helper type 1 immunity. Nat Immunol 6: 1245-1252. 113. Jan M, Chao MP, Cha AC, Alizadeh AA, Gentles AJ, et al. (2011) Prospective separation of normal and leukemic stem cells based on differential expression of TIM3, a human acute myeloid leukemia stem cell marker. Proc Natl Acad Sci U S A 108: 5009-5014. 114. Majeti R, Chao MP, Alizadeh AA, Pang WW, Jaiswal S, et al. (2009) CD47 is an adverse prognostic factor and therapeutic antibody target on human acute myeloid leukemia stem cells. Cell 138: 286-299. 115. Saito Y, Kitamura H, Hijikata A, Tomizawa-Murasawa M, Tanaka S, et al. (2010) Identification of therapeutic targets for quiescent, chemotherapy-resistant human leukemia stem cells. Sci Transl Med 2: 17ra19. 116. Somervaille TC, Matheny CJ, Spencer GJ, Iwasaki M, Rinn JL, et al. (2009) Hierarchical maintenance of MLL myeloid leukemia stem cells employs a transcriptional program shared with embryonic rather than adult stem cells. Cell Stem Cell 4: 129-140. 117. Owens BM, Hawley RG (2002) HOX and non-HOX homeobox genes in leukemic hematopoiesis. Stem Cells 20: 364-379. 118. Bullinger L, Dohner K, Bair E, Frohling S, Schlenk RF, et al. (2004) Use of gene- expression profiling to identify prognostic subclasses in adult acute myeloid leukemia. N Engl J Med 350: 1605-1616. 119. Armstrong SA, Staunton JE, Silverman LB, Pieters R, den Boer ML, et al. (2002) MLL translocations specify a distinct gene expression profile that distinguishes a unique leukemia. Nat Genet 30: 41-47.

201

120. Valk PJ, Verhaak RG, Beijen MA, Erpelinck CA, Barjesteh van Waalwijk van Doorn-Khosrovani S, et al. (2004) Prognostically useful gene-expression profiles in acute myeloid leukemia. N Engl J Med 350: 1617-1628. 121. Ayton PM, Cleary ML (2003) Transformation of myeloid progenitors by MLL oncoproteins is dependent on Hoxa7 and Hoxa9. Genes Dev 17: 2298-2307. 122. Wong P, Iwasaki M, Somervaille TC, So CW, Cleary ML (2007) Meis1 is an essential and rate-limiting regulator of MLL leukemia stem cell potential. Genes Dev 21: 2762-2774. 123. Harrison H, Farnie G, Howell SJ, Rock RE, Stylianou S, et al. (2010) Regulation of breast cancer stem cell activity by signaling through the Notch4 receptor. Cancer Res 70: 709-718. 124. Abel EV, Kim EJ, Wu J, Hynes M, Bednar F, et al. (2014) The Notch pathway is important in maintaining the cancer stem cell population in pancreatic cancer. PLoS One 9: e91983. 125. Hassan KA, Wang L, Korkaya H, Chen G, Maillard I, et al. (2011) Notch pathway activity identifies functional lung cancer stem cells and correlates with worse survival in patients with lung adenocarcinoma. Cancer Research 71: 102nd Annual Meeting of the American Association for Cancer Research. 126. Pannuti A, Foreman K, Rizzo P, Osipo C, Golde T, et al. (2010) Targeting Notch to target cancer stem cells. Clin Cancer Res 16: 3141-3152. 127. Tohda S (2014) NOTCH signaling roles in acute myeloid leukemia cell growth and interaction with other stemness-related signals. Anticancer Res 34: 6259-6264. 128. Lobry C, Ntziachristos P, Ndiaye-Lobry D, Oh P, Cimmino L, et al. (2013) Notch pathway activation targets AML-initiating cell homeostasis and differentiation. J Exp Med 210: 301-319. 129. Peacock CD, Wang Q, Gesell GS, Corcoran-Schwartz IM, Jones E, et al. (2007) Hedgehog signaling maintains a tumor stem cell compartment in multiple myeloma. Proc Natl Acad Sci U S A 104: 4048-4053. 130. Lim Y, Gondek L, Li L, Wang Q, Ma H, et al. (2015) Integration of Hedgehog and mutant FLT3 signaling in myeloid leukemia. Sci Transl Med 7: 291ra296. 131. Vermeulen L, De Sousa EMF, van der Heijden M, Cameron K, de Jong JH, et al. (2010) Wnt activity defines colon cancer stem cells and is regulated by the microenvironment. Nat Cell Biol 12: 468-476. 132. Dietrich PA, Yang C, Leung HH, Lynch JR, Gonzales E, et al. (2014) GPR84 sustains aberrant beta-catenin signaling in leukemic stem cells for maintenance of MLL leukemogenesis. Blood 124: 3284-3294. 133. Polakis P (2012) Drugging Wnt signalling in cancer. EMBO J 31: 2737-2746. 134. Pacheco-Pinedo EC, Durham AC, Stewart KM, Goss AM, Lu MM, et al. (2011) Wnt/beta-catenin signaling accelerates mouse lung tumorigenesis by imposing an embryonic distal progenitor phenotype on lung epithelium. Journal of Clinical Investigation 121: 1935-1945. 135. Malanchi I, Santamaria-Martínez A, Susanto E, Peng H, Lehr H, et al. (2012) Interactions between cancer stem cells and their niche govern metastatic colonization. Nature 481: 85-89.

202

136. Zhang N, Wei P, Gong A, Chiu WT, Lee HT, et al. (2011) FoxM1 promotes beta- catenin nuclear localization and controls Wnt target-gene expression and glioma tumorigenesis. Cancer Cell 20: 427-442. 137. Bisson I, Prowse DM (2009) WNT signaling regulates self-renewal and differentiation of prostate cancer cells with stem cell characteristics. Cell Res 19: 683-697. 138. Malanchi I, Peinado H, Kassen D, Hussenet T, Metzger D, et al. (2008) Cutaneous cancer stem cell maintenance is dependent on beta-catenin signalling. Nature 452: 650-653. 139. Holland JD, Klaus A, Garratt AN, Birchmeier W (2013) Wnt signaling in stem and cancer stem cells. Curr Opin Cell Biol 25: 254-264. 140. Dean M, Fojo T, Bates S (2005) Tumour stem cells and drug resistance. Nat Rev Cancer 5: 275-284. 141. Aguirre-Ghiso JA (2007) Models, mechanisms and clinical evidence for cancer dormancy. Nat Rev Cancer 7: 834-846. 142. Visvader JE, Lindeman GJ (2008) Cancer stem cells in solid tumours: accumulating evidence and unresolved questions. Nat Rev Cancer 8: 755-768. 143. Bao S, Wu Q, McLendon RE, Hao Y, Shi Q, et al. (2006) Glioma stem cells promote radioresistance by preferential activation of the DNA damage response. Nature 444: 756-760. 144. Li X, Lewis MT, Huang J, Gutierrez C, Osborne CK, et al. (2008) Intrinsic resistance of tumorigenic breast cancer cells to chemotherapy. J Natl Cancer Inst 100: 672-679. 145. Diehn M, Cho RW, Lobo NA, Kalisky T, Dorie MJ, et al. (2009) Association of reactive oxygen species levels and radioresistance in cancer stem cells. Nature 458: 780-783. 146. Eppert K, Takenaka K, Lechman ER, Waldron L, Nilsson B, et al. (2011) Stem cell gene expression programs influence clinical outcome in human leukemia. Nat Med 17: 1086-1093. 147. Gentles AJ, Plevritis SK, Majeti R, Alizadeh AA (2010) Association of a leukemic stem cell gene expression signature with clinical outcomes in acute myeloid leukemia. JAMA 304: 2706-2715. 148. Pearce DJ, Taussig D, Zibara K, Smith LL, Ridler CM, et al. (2006) AML engraftment in the NOD/SCID assay reflects the outcome of AML: implications for our understanding of the heterogeneity of AML. Blood 107: 1166-1173. 149. van Rhenen A, Feller N, Kelder A, Westra AH, Rombouts E, et al. (2005) High stem cell frequency in acute myeloid leukemia at diagnosis predicts high minimal residual disease and poor survival. Clin Cancer Res 11: 6520-6527. 150. Hamann PR, Hinman LM, Hollander I, Beyer CF, Lindh D, et al. (2002) Gemtuzumab ozogamicin, a potent and selective anti-CD33 antibody- calicheamicin conjugate for treatment of acute myeloid leukemia. Bioconjug Chem 13: 47-58. 151. Pearce DJ, Taussig DC, Bonnet D (2006) Implications of the expression of myeloid markers on normal and leukemic stem cells. Cell Cycle 5: 271-273.

203

152. Taussig DC, Pearce DJ, Simpson C, Rohatiner AZ, Lister TA, et al. (2005) Hematopoietic stem cells express multiple myeloid markers: implications for the origin and targeted therapy of acute myeloid leukemia. Blood 106: 4086-4092. 153. Jordan CT, Upchurch D, Szilvassy SJ, Guzman ML, Howard DS, et al. (2000) The interleukin-3 receptor alpha chain is a unique marker for human acute myelogenous leukemia stem cells. Leukemia 14: 1777-1784. 154. Guzman ML, Neering SJ, Upchurch D, Grimes B, Howard DS, et al. (2001) Nuclear factor-kappaB is constitutively activated in primitive human acute myelogenous leukemia cells. Blood 98: 2301-2307. 155. Testa U, Riccioni R, Militi S, Coccia E, Stellacci E, et al. (2002) Elevated expression of IL-3Ralpha in acute myelogenous leukemia is associated with enhanced blast proliferation, increased cellularity, and poor prognosis. Blood 100: 2980-2988. 156. Jin L, Lee EM, Ramshaw HS, Busfield SJ, Peoppl AG, et al. (2009) Monoclonal antibody-mediated targeting of CD123, IL-3 receptor alpha chain, eliminates human acute myeloid leukemic stem cells. Cell Stem Cell 5: 31-42. 157. Smith BD, Roboz GJ, Walter RB, Altman JK, Ferguson A, et al. (2014) First-in Man, Phase 1 Study of CSL362 (Anti-IL3Rα / Anti-CD123 Monoclonal Antibody) in Patients with CD123+ Acute Myeloid Leukemia (AML) in CR at High Risk for Early Relapse. Blood 122. 158. Jin L, Hope KJ, Zhai Q, Smadja-Joffe F, Dick JE (2006) Targeting of CD44 eradicates human acute myeloid leukemic stem cells. Nat Med 12: 1167-1174. 159. Kikushige Y, Shima T, Takayanagi S, Urata S, Miyamoto T, et al. (2010) TIM-3 is a promising target to selectively kill acute myeloid leukemia stem cells. Cell Stem Cell 7: 708-717. 160. Guzman ML, Rossi RM, Karnischky L, Li X, Peterson DR, et al. (2005) The sesquiterpene lactone parthenolide induces apoptosis of human acute myelogenous leukemia stem and progenitor cells. Blood 105: 4163-4169. 161. Guzman ML, Rossi RM, Neelakantan S, Li X, Corbett CA, et al. (2007) An orally bioavailable parthenolide analog selectively eradicates acute myelogenous leukemia stem and progenitor cells. Blood 110: 4427-4435. 162. Malone W, Perloff M, Crowell J, Sigman C, Higley H (2003) Fenretinide: a prototype cancer prevention drug. Expert Opin Investig Drugs 12: 1829-1842. 163. Schneider BJ, Worden FP, Gadgeel SM, Parchment RE, Hodges CM, et al. (2009) Phase II trial of fenretinide (NSC 374551) in patients with recurrent small cell lung cancer. Invest New Drugs 27: 571-578. 164. Veronesi U, Mariani L, Decensi A, Formelli F, Camerini T, et al. (2006) Fifteen- year results of a randomized phase III trial of fenretinide to prevent second breast cancer. Ann Oncol 17: 1065-1071. 165. Zhang H, Mi JQ, Fang H, Wang Z, Wang C, et al. (2013) Preferential eradication of acute myelogenous leukemia stem cells by fenretinide. Proc Natl Acad Sci U S A 110: 5606-5611. 166. Pierce KL, Premont RT, Lefkowitz RJ (2002) Seven-transmembrane receptors. Nat Rev Mol Cell Biol 3: 639-650.

204

167. Lattin J, Zidar DA, Schroder K, Kellie S, Hume DA, et al. (2007) G-protein- coupled receptor expression, function, and signaling in macrophages. J Leukoc Biol 82: 16-32. 168. Sun L, Ye RD (2012) Role of G protein-coupled receptors in inflammation. Acta Pharmacol Sin 33: 342-350. 169. Catapano LA, Manji HK (2007) G protein-coupled receptors in major psychiatric disorders. Biochim Biophys Acta 1768: 976-993. 170. Lappano R, Maggiolini M (2011) G protein-coupled receptors: novel targets for drug discovery in cancer. Nat Rev Drug Discov 10: 47-60. 171. Young D, Waitches G, Birchmeier C, Fasano O, Wigler M (1986) Isolation and characterization of a new cellular oncogene encoding a protein with multiple potential transmembrane domains. Cell 45: 711-719. 172. Contos JJ, Ishii I, Chun J (2000) Lysophosphatidic acid receptors. Mol Pharmacol 58: 1188-1196. 173. Bian D, Su S, Mahanivong C, Cheng RK, Han Q, et al. (2004) Lysophosphatidic Acid Stimulates Ovarian Cancer Cell Migration via a Ras-MEK Kinase 1 Pathway. Cancer Res 64: 4209-4217. 174. Hu L, Roth JM, Brooks P, Luty J, Karpatkin S (2008) Thrombin up-regulates cathepsin D which enhances angiogenesis, growth, and metastasis. Cancer Res 68: 4666-4673. 175. Malchinkhuu E, Sato K, Horiuchi Y, Mogi C, Ohwada S, et al. (2005) Role of p38 mitogen-activated kinase and c-Jun terminal kinase in migration response to lysophosphatidic acid and sphingosine-1-phosphate in glioma cells. Oncogene 24: 6676-6688. 176. Boire A, Covic L, Agarwal A, Jacques S, Sherifi S, et al. (2005) PAR1 is a matrix metalloprotease-1 receptor that promotes invasion and tumorigenesis of breast cancer cells. Cell 120: 303-313. 177. Massi D, Naldini A, Ardinghi C, Carraro F, Franchi A, et al. (2005) Expression of protease-activated receptors 1 and 2 in melanocytic nevi and malignant melanoma. Hum Pathol 36: 676-685. 178. Veiga Cde S, Carneiro-Lobo TC, Coelho CJ, Carvalho SM, Maia RC, et al. (2011) Increased expression of protease-activated receptor 1 (PAR-1) in human leukemias. Blood Cells Mol Dis 46: 230-234. 179. Castellone MD, Teramoto H, Williams BO, Druey KM, Gutkind JS (2005) Prostaglandin E2 promotes colon cancer cell growth through a Gs-axin-beta- catenin signaling axis. Science 310: 1504-1510. 180. Wang X, Klein RD (2007) Prostaglandin E2 induces vascular endothelial growth factor secretion in prostate cancer cells through EP2 receptor-mediated cAMP pathway. Mol Carcinog 46: 912-923. 181. Chen EP, Smyth EM (2011) COX-2 and PGE2-dependent immunomodulation in breast cancer. Prostaglandins Other Lipid Mediat 96: 14-20. 182. Wang CY, Cusack JC, Jr., Liu R, Baldwin AS, Jr. (1999) Control of inducible chemoresistance: enhanced anti-tumor therapy through increased apoptosis by inhibition of NF-kappaB. Nat Med 5: 412-417.

205

183. Taub JS, Guo R, Leeb-Lundberg LM, Madden JF, Daaka Y (2003) Bradykinin receptor subtype 1 expression and function in prostate cancer. Cancer Res 63: 2037-2041. 184. Kinoshita J, Fushida S, Harada S, Yagi Y, Fujita H, et al. (2009) Local angiotensin II-generation in human gastric cancer: correlation with tumor progression through the activation of ERK1/2, NF-kappaB and survivin. Int J Oncol 34: 1573-1582. 185. Uemura H, Ishiguro H, Nakaigawa N, Nagashima Y, Miyoshi Y, et al. (2003) Angiotensin II receptor blocker shows antiproliferative activity in prostate cancer cells: a possibility of tyrosine kinase inhibitor of growth factor. Mol Cancer Ther 2: 1139-1147. 186. Ariazi EA, Brailoiu E, Yerrum S, Shupp HA, Slifker MJ, et al. (2010) The G protein-coupled receptor GPR30 inhibits proliferation of estrogen receptor- positive breast cancer cells. Cancer Res 70: 1184-1194. 187. Wang C, Lv X, He C, Hua G, Tsai MY, et al. (2013) The G-protein-coupled estrogen receptor agonist G-1 suppresses proliferation of ovarian cancer cells by blocking tubulin polymerization. Cell Death Dis 4: e869. 188. Kavanagh DO, McIlroy M, Myers E, Bane F, Crotty TB, et al. (2010) The role of oestrogen receptor {alpha} in human thyroid cancer: contributions from coregulatory proteins and the tyrosine kinase receptor HER2. Endocr Relat Cancer 17: 255-264. 189. Varnat F, Duquet A, Malerba M, Zbinden M, Mas C, et al. (2009) Human colon cancer epithelial cells harbour active HEDGEHOG-GLI signalling that is essential for tumour growth, recurrence, metastasis and stem cell survival and expansion. EMBO Mol Med 1: 338-351. 190. Holcombe RF, Marsh JL, Waterman ML, Lin F, Milovanovic T, et al. (2002) Expression of Wnt ligands and Frizzled receptors in colonic mucosa and in colon carcinoma. Mol Pathol 55: 220-226. 191. Wang HQ, Xu ML, Ma J, Zhang Y, Xie CH (2012) Frizzled-8 as a putative therapeutic target in human lung cancer. Biochem Biophys Res Commun 417: 62-66. 192. Tickenbrock L, Hehn S, Sargin B, Choudhary C, Baumer N, et al. (2008) Activation of Wnt signalling in acute myeloid leukemia by induction of Frizzled-4. Int J Oncol 33: 1215-1221. 193. Oldham WM, Hamm HE (2008) Heterotrimeric G protein activation by G-protein- coupled receptors. Nat Rev Mol Cell Biol 9: 60-71. 194. Cabrera-Vera TM, Vanhauwe J, Thomas TO, Medkova M, Preininger A, et al. (2003) Insights into G protein structure, function, and regulation. Endocrine Reviews 24: 765-781. 195. O'Hayre M, Vazquez-Prado J, Kufareva I, Stawiski EW, Handel TM, et al. (2013) The emerging mutational landscape of G proteins and G-protein-coupled receptors in cancer. Nat Rev Cancer 13: 412-424.

206

196. Kalinec G, Nazarali AJ, Hermouet S, Xu N, Gutkind JS (1992) Mutated alpha subunit of the Gq protein induces malignant transformation in NIH 3T3 cells. Mol Cell Biol 12: 4687-4693. 197. Xu N, Voyno-Yasenetskaya T, Gutkind JS (1994) Potent transforming activity of the G13 alpha subunit defines a novel family of oncogenes. Biochem Biophys Res Commun 201: 603-609. 198. Kehrl JH (1998) Heterotrimeric G protein signaling: roles in immune function and fine-tuning by RGS proteins. Immunity 8: 1-10. 199. De Vries L, Zheng B, Fischer T, Elenko E, Farquhar MG (2000) The regulator of G protein signaling family. Annu Rev Pharmacol Toxicol 40: 235-271. 200. Hepler JR, Berman DM, Gilman AG, Kozasa T (1997) RGS4 and GAIP are GTPase-activating proteins for Gq alpha and block activation of phospholipase C beta by gamma-thio-GTP-Gq alpha. Proc Natl Acad Sci U S A 94: 428-432. 201. Roy AA, Baragli A, Bernstein LS, Hepler JR, Hebert TE, et al. (2006) RGS2 interacts with Gs and adenylyl cyclase in living cells. Cell Signal 18: 336-348. 202. Schoeber JP, Topala CN, Wang X, Diepens RJ, Lambers TT, et al. (2006) RGS2 inhibits the epithelial Ca2+ channel TRPV6. J Biol Chem 281: 29669-29674. 203. Silva AP, Salim AC, Bulgarelli A, de Souza JE, Osorio E, et al. (2003) Identification of 9 novel transcripts and two RGSL genes within the hereditary prostate cancer region (HPC1) at 1q25. Gene 310: 49-57. 204. Sood R, Bonner TI, Makalowska I, Stephan DA, Robbins CM, et al. (2001) Cloning and characterization of 13 novel transcripts and the human RGS8 gene from the 1q25 region encompassing the hereditary prostate cancer (HPC1) locus. Genomics 73: 211-222. 205. Wiechec E, Overgaard J, Hansen LL (2008) A fragile site within the HPC1 region at 1q25.3 affecting RGS16, RGSL1, and RGSL2 in human breast carcinomas. Genes Chromosomes Cancer 47: 766-780. 206. Smalley MJ, Iravani M, Leao M, Grigoriadis A, Kendrick H, et al. (2007) Regulator of G-protein signalling 2 mRNA is differentially expressed in mammary epithelial subpopulations and over-expressed in the majority of breast cancers. Breast Cancer Res 9: R85. 207. Schwable J, Choudhary C, Thiede C, Tickenbrock L, Sargin B, et al. (2005) RGS2 is an important target gene of Flt3-ITD mutations in AML and functions in myeloid differentiation and leukemic transformation. Blood 105: 2107-2114. 208. Hsu SY, Liang SG, Hsueh AJ (1998) Characterization of two LGR genes homologous to gonadotropin and thyrotropin receptors with extracellular leucine-rich repeats and a G protein-coupled, seven-transmembrane region. Mol Endocrinol 12: 1830-1845. 209. Luo CW, Hsueh AJ (2006) Genomic analyses of the evolution of LGR genes. Chang Gung Med J 29: 2-8. 210. de Lau W, Barker N, Low TY, Koo BK, Li VS, et al. (2011) Lgr5 homologues associate with Wnt receptors and mediate R-spondin signalling. Nature 476: 293-297.

207

211. Mustata RC, Van Loy T, Lefort A, Libert F, Strollo S, et al. (2011) Lgr4 is required for Paneth cell differentiation and maintenance of intestinal stem cells ex vivo. EMBO Rep 12: 558-564. 212. Luo W, Rodriguez M, Valdez JM, Zhu X, Tan K, et al. (2013) Lgr4 is a key regulator of prostate development and prostate stem cell differentiation. Stem Cells 31: 2492-2505. 213. Wang Y, Dong J, Li D, Lai L, Siwko S, et al. (2013) Lgr4 regulates mammary gland development and stem cell activity through the pluripotency transcription factor Sox2. Stem Cells 31: 1921-1931. 214. Carmon KS, Gong X, Lin Q, Thomas A, Liu Q (2011) R-spondins function as ligands of the orphan receptors LGR4 and LGR5 to regulate Wnt/beta-catenin signaling. Proc Natl Acad Sci U S A 108: 11452-11457. 215. Glinka A, Dolde C, Kirsch N, Huang YL, Kazanskaya O, et al. (2011) LGR4 and LGR5 are R-spondin receptors mediating Wnt/beta-catenin and Wnt/PCP signalling. EMBO Rep 12: 1055-1061. 216. Ruffner H, Sprunger J, Charlat O, Leighton-Davies J, Grosshans B, et al. (2012) R- Spondin Potentiates Wnt/beta-Catenin Signaling through Orphan Receptors LGR4 and LGR5. PLoS One 7: e40976. 217. Seshagiri S, Stawiski EW, Durinck S, Modrusan Z, Storm EE, et al. (2012) Recurrent R-spondin fusions in colon cancer. Nature 488: 660-664. 218. Wu J, Xie N, Xie K, Zeng J, Cheng L, et al. (2013) GPR48, a poor prognostic factor, promotes tumor metastasis and activates beta-catenin/TCF signaling in colorectal cancer. Carcinogenesis 34: 2861-2869. 219. Van Schoore G, Mendive F, Pochet R, Vassart G (2005) Expression pattern of the orphan receptor LGR4/GPR48 gene in the mouse. Histochem Cell Biol 124: 35- 50. 220. Luo J, Zhou W, Zhou X, Li D, Weng J, et al. (2009) Regulation of bone formation and remodeling by G-protein-coupled receptor 48. Development 136: 2747- 2756. 221. Weng J, Luo J, Cheng X, Jin C, Zhou X, et al. (2008) Deletion of G protein- coupled receptor 48 leads to ocular anterior segment dysgenesis (ASD) through down-regulation of Pitx2. Proc Natl Acad Sci U S A 105: 6081-6086. 222. Mendive F, Laurent P, Van Schoore G, Skarnes W, Pochet R, et al. (2006) Defective postnatal development of the male reproductive tract in LGR4 knockout mice. Dev Biol 290: 421-434. 223. Yamashita R, Takegawa Y, Sakumoto M, Nakahara M, Kawazu H, et al. (2009) Defective development of the gall bladder and cystic duct in Lgr4- hypomorphic mice. Dev Dyn 238: 993-1000. 224. Kato S, Matsubara M, Matsuo T, Mohri Y, Kazama I, et al. (2006) Leucine-rich repeat-containing G protein-coupled receptor-4 (LGR4, Gpr48) is essential for renal development in mice. Nephron Exp Nephrol 104: e63-75. 225. Kato S, Mohri Y, Matsuo T, Ogawa E, Umezawa A, et al. (2007) Eye-open at birth phenotype with reduced keratinocyte motility in LGR4 null mice. FEBS Lett 581: 4685-4690.

208

226. Mohri Y, Kato S, Umezawa A, Okuyama R, Nishimori K (2008) Impaired hair placode formation with reduced expression of hair follicle-related genes in mice lacking Lgr4. Dev Dyn 237: 2235-2242. 227. Mazerbourg S, Bouley DM, Sudo S, Klein CA, Zhang JV, et al. (2004) Leucine- rich repeat-containing, G protein-coupled receptor 4 null mice exhibit intrauterine growth retardation associated with embryonic and perinatal lethality. Mol Endocrinol 18: 2241-2254. 228. Mohri Y, Umezu T, Hidema S, Tomisawa H, Akamatsu A, et al. (2010) Reduced fertility with impairment of early-stage embryos observed in mice lacking Lgr4 in epithelial tissues. Fertil Steril 94: 2878-2881. 229. Oyama K, Mohri Y, Sone M, Nawa A, Nishimori K (2011) Conditional knockout of Lgr4 leads to impaired ductal elongation and branching morphogenesis in mouse mammary glands. Sex Dev 5: 205-212. 230. Song H, Luo J, Luo W, Weng J, Wang Z, et al. (2008) Inactivation of G-protein- coupled receptor 48 (Gpr48/Lgr4) impairs definitive erythropoiesis at midgestation through down-regulation of the ATF4 signaling pathway. J Biol Chem 283: 36687-36697. 231. Barker N, Tan S, Clevers H (2013) Lgr proteins in epithelial stem cell biology. Development 140: 2484-2494. 232. Dick JE (2008) Stem cell concepts renew cancer research. Blood 112: 4793-4807. 233. Livak KJ, Schmittgen TD (2001) Analysis of relative gene expression data using real-time quantitative PCR and the 2(T)(-Delta Delta C) method. Methods 25: 402-408. 234. Shlush LI, Zandi S, Mitchell A, Chen WC, Brandwein JM, et al. (2014) Identification of pre-leukaemic haematopoietic stem cells in acute leukaemia. Nature 506: 328-333. 235. Kim KA, Kakitani M, Zhao J, Oshima T, Tang T, et al. (2005) Mitogenic influence of human R-spondin1 on the intestinal epithelium. Science 309: 1256-1259. 236. Kazanskaya O, Glinka A, del Barco Barrantes I, Stannek P, Niehrs C, et al. (2004) R-Spondin2 is a secreted activator of Wnt/beta-catenin signaling and is required for Xenopus myogenesis. Dev Cell 7: 525-534. 237. Kim KA, Wagle M, Tran K, Zhan X, Dixon MA, et al. (2008) R-Spondin family members regulate the Wnt pathway by a common mechanism. Mol Biol Cell 19: 2588-2596. 238. Zhou WJ, Geng ZH, Spence JR, Geng JG (2013) Induction of intestinal stem cells by R-spondin 1 and Slit2 augments chemoradioprotection. Nature 501: 107-+. 239. Cai CG, Yu QC, Jiang WM, Liu W, Song WQ, et al. (2014) R-spondin1 is a novel hormone mediator for mammary stem cell self-renewal. Genes & Development 28: 2205-2218. 240. Bagger FO, Rapin N, Theilgaard-Monch K, Kaczkowski B, Thoren LA, et al. (2013) HemaExplorer: a database of mRNA expression profiles in normal and malignant haematopoiesis. Nucleic Acids Res 41: D1034-1039.

209

241. Majeti R, Becker MW, Tian Q, Lee TL, Yan X, et al. (2009) Dysregulated gene expression networks in human acute myelogenous leukemia stem cells. Proc Natl Acad Sci U S A 106: 3396-3401. 242. Simon M, Grandage VL, Linch DC, Khwaja A (2005) Constitutive activation of the Wnt/beta-catenin signalling pathway in acute myeloid leukaemia. Oncogene 24: 2410-2420. 243. de Lau WB, Snel B, Clevers HC (2012) The R-spondin protein family. Genome Biol 13: 242. 244. Zhang Y, Wang J, Wheat J, Chen X, Jin S, et al. (2013) AML1-ETO mediates hematopoietic self-renewal and leukemogenesis through a COX/beta-catenin signaling pathway. Blood 121: 4906-4916. 245. Krivtsov AV, Figueroa ME, Sinha AU, Stubbs MC, Feng Z, et al. (2013) Cell of origin determines clinically relevant subtypes of MLL-rearranged AML. Leukemia 27: 852-860. 246. Ysebaert L, Chicanne G, Demur C, De Toni F, Prade-Houdellier N, et al. (2006) Expression of beta-catenin by acute myeloid leukemia cells predicts enhanced clonogenic capacities and poor prognosis. Leukemia 20: 1211-1216. 247. Xu J, Suzuki M, Niwa Y, Hiraga J, Nagasaka T, et al. (2008) Clinical significance of nuclear non-phosphorylated beta-catenin in acute myeloid leukaemia and myelodysplastic syndrome. Br J Haematol 140: 394-401. 248. Schepers A, Clevers H (2012) Wnt signaling, stem cells, and cancer of the gastrointestinal tract. Cold Spring Harb Perspect Biol 4: a007989. 249. Hurst JH, Hooks SB (2009) Regulator of G-protein signaling (RGS) proteins in cancer biology. Biochem Pharmacol 78: 1289-1297. 250. Yang J, Moses MA (2009) Lipocalin 2: a multifaceted modulator of human cancer. Cell Cycle 8: 2347-2352. 251. Tetreault MP, Yang Y, Katz JP (2013) Kruppel-like factors in cancer. Nat Rev Cancer 13: 701-713. 252. Bi X, Hameed M, Mirani N, Pimenta EM, Anari J, et al. (2011) Loss of interferon regulatory factor 5 (IRF5) expression in human ductal carcinoma correlates with disease stage and contributes to metastasis. Breast Cancer Res 13: R111. 253. Goransson M, Andersson MK, Forni C, Stahlberg A, Andersson C, et al. (2009) The myxoid liposarcoma FUS-DDIT3 fusion oncoprotein deregulates NF- kappaB target genes by interaction with NFKBIZ. Oncogene 28: 270-278. 254. Tamura RE, de Vasconcellos JF, Sarkar D, Libermann TA, Fisher PB, et al. (2012) GADD45 proteins: central players in tumorigenesis. Curr Mol Med 12: 634-651. 255. Hernandez NA, Correa E, Avila EP, Vela TA, Perez VM (2009) PAR1 is selectively over expressed in high grade breast cancer patients: a cohort study. J Transl Med 7: 47. 256. Godoy P, Cadenas C, Hellwig B, Marchan R, Stewart J, et al. (2014) Interferon- inducible guanylate binding protein (GBP2) is associated with better prognosis in breast cancer and indicates an efficient T cell response. Breast Cancer 21: 491-499.

210

257. Kerkela E, Ala-aho R, Klemi P, Grenman S, Shapiro SD, et al. (2002) Metalloelastase (MMP-12) expression by tumour cells in squamous cell carcinoma of the vulva correlates with invasiveness, while that by macrophages predicts better outcome. J Pathol 198: 258-269. 258. Bourboulia D, Han H, Jensen-Taubman S, Gavil N, Isaac B, et al. (2013) TIMP-2 modulates cancer cell transcriptional profile and enhances E-cadherin/beta- catenin complex expression in A549 lung cancer cells. Oncotarget 4: 163-173. 259. Sahraei M, Roy LD, Curry JM, Teresa TL, Nath S, et al. (2012) MUC1 regulates PDGFA expression during pancreatic cancer progression. Oncogene 31: 4935- 4945. 260. Reilly PT, Teo WL, Low MJ, Amoyo-Brion AA, Dominguez-Brauer C, et al. (2013) Lipocalin 2 performs contrasting, location-dependent roles in APCmin tumor initiation and progression. Oncogene 32: 1233-1239. 261. Venkatesha S, Hanai J, Seth P, Karumanchi SA, Sukhatme VP (2006) Lipocalin 2 antagonizes the proangiogenic action of ras in transformed cells. Mol Cancer Res 4: 821-829. 262. Kastan MB, Zhan Q, el-Deiry WS, Carrier F, Jacks T, et al. (1992) A mammalian cell cycle checkpoint pathway utilizing p53 and GADD45 is defective in ataxia- telangiectasia. Cell 71: 587-597. 263. Harkin DP, Bean JM, Miklos D, Song YH, Truong VB, et al. (1999) Induction of GADD45 and JNK/SAPK-dependent apoptosis following inducible expression of BRCA1. Cell 97: 575-586. 264. Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, et al. (2005) Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A 102: 15545-15550. 265. Okamura A, Iwata N, Nagata A, Tamekane A, Shimoyama M, et al. (2004) Involvement of casein kinase Iepsilon in cytokine-induced granulocytic differentiation. Blood 103: 2997-3004. 266. Yang WS, Stockwell BR (2008) Inhibition of casein kinase 1-epsilon induces cancer-cell-selective, PERIOD2-dependent growth arrest. Genome Biol 9: R92. 267. Katritch V, Cherezov V, Stevens RC (2013) Structure-function of the G protein- coupled receptor superfamily. Annu Rev Pharmacol Toxicol 53: 531-556. 268. Soundararajan M, Willard FS, Kimple AJ, Turnbull AP, Ball LJ, et al. (2008) Structural diversity in the RGS domain and its interaction with heterotrimeric G protein alpha-subunits. Proc Natl Acad Sci U S A 105: 6457-6462. 269. Moratz C, Kang VH, Druey KM, Shi CS, Scheschonka A, et al. (2000) Regulator of G protein signaling 1 (RGS1) markedly impairs Gi alpha signaling responses of B lymphocytes. J Immunol 164: 1829-1838. 270. Pittman M (1979) Pertussis toxin: the cause of the harmful effects and prolonged immunity of whooping cough. A hypothesis. Rev Infect Dis 1: 401-412. 271. Hsia JA, Moss J, Hewlett EL, Vaughan M (1984) ADP-ribosylation of adenylate cyclase by pertussis toxin. Effects on inhibitory agonist binding. J Biol Chem 259: 1086-1090.

211

272. Burns DL (1988) Subunit structure and enzymic activity of pertussis toxin. Microbiol Sci 5: 285-287. 273. Katada T, Tamura M, Ui M (1983) The A protomer of islet-activating protein, pertussis toxin, as an active peptide catalyzing ADP-ribosylation of a membrane protein. Arch Biochem Biophys 224: 290-298. 274. Gunther EC, von Bartheld CS, Goodman LJ, Johnson JE, Bothwell M (2000) The G-protein inhibitor, pertussis toxin, inhibits the secretion of brain-derived neurotrophic factor. Neuroscience 100: 569-579. 275. Lee HY, Kim SD, Shim JW, Kim HJ, Yun J, et al. (2010) A pertussis toxin sensitive G-protein-independent pathway is involved in serum amyloid A- induced formyl peptide receptor 2-mediated CCL2 production. Exp Mol Med 42: 302-309. 276. Ayoub MA, Damian M, Gespach C, Ferrandis E, Lavergne O, et al. (2009) Inhibition of heterotrimeric G protein signaling by a small molecule acting on Galpha subunit. J Biol Chem 284: 29136-29145. 277. Mukai H, Munekata E, Higashijima T (1992) G protein antagonists. A novel hydrophobic peptide competes with receptor for G protein binding. J Biol Chem 267: 16237-16243. 278. Folkers K, Hakanson R, Horig J, Xu JC, Leander S (1984) Biological evaluation of substance P antagonists. Br J Pharmacol 83: 449-456. 279. Hunt RA, Bhat GJ, Baker KM (1999) Angiotensin II-stimulated induction of sis- inducing factor is mediated by pertussis toxin-insensitive G(q) proteins in cardiac myocytes. Hypertension 34: 603-608. 280. Tsunoda Y, Owyang C (1995) The regulatory site of functional GTP binding protein coupled to the high affinity cholecystokinin receptor and phospholipase A2 pathway is on the G beta subunit of Gq protein in pancreatic acini. Biochem Biophys Res Commun 211: 648-655. 281. Croxton TL, Lande B, Hirshman CA (1998) Role of G proteins in agonist-induced Ca2+ sensitization of tracheal smooth muscle. Am J Physiol 275: L748-755. 282. Liu X, Rubin JS, Kimmel AR (2005) Rapid, Wnt-induced changes in GSK3beta associations that regulate beta-catenin stabilization are mediated by Galpha proteins. Curr Biol 15: 1989-1997. 283. Armstrong VJ, Muzylak M, Sunters A, Zaman G, Saxon LK, et al. (2007) Wnt/beta-catenin signaling is a component of osteoblastic bone cell early responses to load-bearing and requires estrogen receptor alpha. J Biol Chem 282: 20715-20727. 284. Zhan Q (2005) Gadd45a, a p53- and BRCA1-regulated stress protein, in cellular response to DNA damage. Mutat Res 569: 133-143. 285. Liebermann DA, Hoffman B (2002) Myeloid differentiation (MyD)/growth arrest DNA damage (GADD) genes in tumor suppression, immunity and inflammation. Leukemia 16: 527-541. 286. Hollander MC, Fornace AJ, Jr. (2002) Genomic instability, centrosome amplification, cell cycle checkpoints and Gadd45a. Oncogene 21: 6228-6233.

212

287. Ramachandran K, Gopisetty G, Gordian E, Navarro L, Hader C, et al. (2009) Methylation-mediated repression of GADD45alpha in prostate cancer and its role as a potential therapeutic target. Cancer Res 69: 1527-1535. 288. Wang W, Huper G, Guo Y, Murphy SK, Olson JA, Jr., et al. (2005) Analysis of methylation-sensitive transcriptome identifies GADD45a as a frequently methylated gene in breast cancer. Oncogene 24: 2705-2714. 289. Perugini M, Iarossi DG, Kok CH, Cummings N, Diakiw SM, et al. (2013) GADD45A methylation predicts poor overall survival in acute myeloid leukemia and is associated with IDH1/2 and DNMT3A mutations. Leukemia 27: 1588-1592. 290. Guo W, Dong Z, Guo Y, Chen Z, Kuang G, et al. (2013) Methylation-mediated repression of GADD45A and GADD45G expression in gastric cardia adenocarcinoma. Int J Cancer 133: 2043-2053. 291. Chen Y, Ma X, Zhang M, Wang X, Wang C, et al. (2014) Gadd45a regulates hematopoietic stem cell stress responses in mice. Blood 123: 851-862. 292. Lagadinou ED, Sach A, Callahan K, Rossi RM, Neering SJ, et al. (2013) BCL-2 inhibition targets oxidative phosphorylation and selectively eradicates quiescent human leukemia stem cells. Cell Stem Cell 12: 329-341. 293. Hirst J (2013) Mitochondrial complex I. Annu Rev Biochem 82: 551-575. 294. Esposti MD, De Vries S, Crimi M, Ghelli A, Patarnello T, et al. (1993) Mitochondrial cytochrome b: evolution and structure of the protein. Biochim Biophys Acta 1143: 243-271. 295. Anderson S, Bankier AT, Barrell BG, de Bruijn MH, Coulson AR, et al. (1981) Sequence and organization of the human mitochondrial genome. Nature 290: 457-465. 296. Knowles JR (1980) Enzyme-catalyzed phosphoryl transfer reactions. Annu Rev Biochem 49: 877-919. 297. Jang YY, Sharkis SJ (2007) A low level of reactive oxygen species selects for primitive hematopoietic stem cells that may reside in the low-oxygenic niche. Blood 110: 3056-3063. 298. Balaban RS, Nemoto S, Finkel T (2005) Mitochondria, oxidants, and aging. Cell 120: 483-495. 299. Alexandre A, Lehninger AL (1984) Bypasses of the Antimycin a Block of Mitochondrial Electron-Transport in Relation to Ubisemiquinone Function. Biochimica Et Biophysica Acta 767: 120-129. 300. Campo ML, Kinnally KW, Tedeschi H (1992) The Effect of Antimycin-a on Mouse-Liver Inner Mitochondrial-Membrane Channel Activity. Journal of Biological Chemistry 267: 8123-8127. 301. Xia D, Yu CA, Kim H, Xian JZ, Kachurin AM, et al. (1997) Crystal structure of the cytochrome bc(1) complex from bovine heart mitochondria (vol 277, pg 60, 1997). Science 278: 2037-2037. 302. Li N, Ragheb K, Lawler G, Sturgis J, Rajwa B, et al. (2003) Mitochondrial complex I inhibitor rotenone induces apoptosis through enhancing mitochondrial reactive oxygen species production. J Biol Chem 278: 8516-8525.

213

303. Lambert AJ, Brand MD (2004) Inhibitors of the quinone-binding site allow rapid superoxide production from mitochondrial NADH:ubiquinone oxidoreductase (complex I). J Biol Chem 279: 39414-39420. 304. Gong X, Yi J, Carmon KS, Crumbley CA, Xiong W, et al. (2014) Aberrant RSPO3-LGR4 signaling in Keap1-deficient lung adenocarcinomas promotes tumor aggressiveness. Oncogene. 305. Sato T, Vries RG, Snippert HJ, van de Wetering M, Barker N, et al. (2009) Single Lgr5 stem cells build crypt-villus structures in vitro without a mesenchymal niche. Nature 459: 262-265. 306. Barker N, Ridgway RA, van Es JH, van de Wetering M, Begthel H, et al. (2009) Crypt stem cells as the cells-of-origin of intestinal cancer. Nature 457: 608-611. 307. Koo BK, Spit M, Jordens I, Low TY, Stange DE, et al. (2012) Tumour suppressor RNF43 is a stem-cell E3 ligase that induces endocytosis of Wnt receptors. Nature 488: 665-669. 308. Hao HX, Xie Y, Zhang Y, Charlat O, Oster E, et al. (2012) ZNRF3 promotes Wnt receptor turnover in an R-spondin-sensitive manner. Nature 485: 195-200. 309. Xie Y, Zamponi R, Charlat O, Ramones M, Swalley S, et al. (2013) Interaction with both ZNRF3 and LGR4 is required for the signalling activity of R-spondin. EMBO Rep 14: 1120-1126. 310. Xu K, Xu Y, Rajashankar KR, Robev D, Nikolov DB (2013) Crystal structures of Lgr4 and its complex with R-spondin1. Structure 21: 1683-1689. 311. Peng WC, de Lau W, Madoori PK, Forneris F, Granneman JC, et al. (2013) Structures of Wnt-antagonist ZNRF3 and its complex with R-spondin 1 and implications for signaling. PLoS One 8: e83110. 312. Liang F, Yue J, Wang J, Zhang L, Fan R, et al. (2015) GPCR48/LGR4 promotes tumorigenesis of prostate cancer via PI3K/Akt signaling pathway. Med Oncol 32: 49. 313. Styrkarsdottir U, Thorleifsson G, Sulem P, Gudbjartsson DF, Sigurdsson A, et al. (2013) Nonsense mutation in the LGR4 gene is associated with several human diseases and other traits. Nature 497: 517-520. 314. Clevers H, Nusse R (2012) Wnt/beta-catenin signaling and disease. Cell 149: 1192-1205. 315. Mohri Y, Oyama K, Akamatsu A, Kato S, Nishimori K (2011) Lgr4-Deficient Mice Showed Premature Differentiation of Ureteric Bud With Reduced Expression of Wnt Effector Lef1 and Gata3. Developmental Dynamics 240: 1626-1634. 316. Perugini M, Kok CH, Brown AL, Wilkinson CR, Salerno DG, et al. (2009) Repression of Gadd45alpha by activated FLT3 and GM-CSF receptor mutants contributes to growth, survival and blocked differentiation. Leukemia 23: 729- 738. 317. Li Y, Qian H, Li X, Wang H, Yu J, et al. (2009) Adenoviral-mediated gene transfer of Gadd45a results in suppression by inducing apoptosis and cell cycle arrest in pancreatic cancer cell. J Gene Med 11: 3-13.

214

318. Yang C, Yang S, Wood KB, Hornicek FJ, Schwab JH, et al. (2009) Multidrug resistant osteosarcoma cell lines exhibit deficiency of GADD45alpha expression. Apoptosis 14: 124-133. 319. Al-Romaih K, Sadikovic B, Yoshimoto M, Wang Y, Zielenska M, et al. (2008) Decitabine-induced demethylation of 5' CpG island in GADD45A leads to apoptosis in osteosarcoma cells. Neoplasia 10: 471-480. 320. Cairns RA, Harris IS, Mak TW (2011) Regulation of cancer cell metabolism. Nat Rev Cancer 11: 85-95. 321. Simsek T, Kocabas F, Zheng J, Deberardinis RJ, Mahmoud AI, et al. (2010) The distinct metabolic profile of hematopoietic stem cells reflects their location in a hypoxic niche. Cell Stem Cell 7: 380-390. 322. Vlashi E, Lagadec C, Vergnes L, Matsutani T, Masui K, et al. (2011) Metabolic state of glioma stem cells and nontumorigenic cells. Proc Natl Acad Sci U S A 108: 16062-16067. 323. Turrens JF (1997) Superoxide production by the mitochondrial respiratory chain. Biosci Rep 17: 3-8. 324. Acin-Perez R, Bayona-Bafaluy MP, Fernandez-Silva P, Moreno-Loshuertos R, Perez-Martos A, et al. (2004) Respiratory complex III is required to maintain complex I in mammalian mitochondria. Mol Cell 13: 805-815. 325. Murphy MP (2009) How mitochondria produce reactive oxygen species. Biochem J 417: 1-13. 326. Waris G, Ahsan H (2006) Reactive oxygen species: role in the development of cancer and various chronic conditions. J Carcinog 5: 14. 327. Sinha K, Das J, Pal PB, Sil PC (2013) Oxidative stress: the mitochondria- dependent and mitochondria-independent pathways of apoptosis. Arch Toxicol 87: 1157-1180. 328. Kaufmann SH, Earnshaw WC (2000) Induction of apoptosis by cancer chemotherapy. Exp Cell Res 256: 42-49. 329. Gorrini C, Harris IS, Mak TW (2013) Modulation of oxidative stress as an anticancer strategy. Nat Rev Drug Discov 12: 931-947. 330. Raj L, Ide T, Gurkar AU, Foley M, Schenone M, et al. (2011) Selective killing of cancer cells by a small molecule targeting the stress response to ROS. Nature 475: 231-234. 331. Yang WS, SriRamaratnam R, Welsch ME, Shimada K, Skouta R, et al. (2014) Regulation of ferroptotic cancer cell death by GPX4. Cell 156: 317-331. 332. Trachootham D, Zhou Y, Zhang H, Demizu Y, Chen Z, et al. (2006) Selective killing of oncogenically transformed cells through a ROS-mediated mechanism by beta-phenylethyl isothiocyanate. Cancer Cell 10: 241-252. 333. Junn E, Han SH, Im JY, Yang Y, Cho EW, et al. (2000) Vitamin D3 up-regulated protein 1 mediates oxidative stress via suppressing the thioredoxin function. J Immunol 164: 6287-6295. 334. Nishiyama A, Masutani H, Nakamura H, Nishinaka Y, Yodoi J (2001) Redox regulation by thioredoxin and thioredoxin-binding proteins. IUBMB Life 52: 29- 33.

215

335. Yoshioka J, Schreiter ER, Lee RT (2006) Role of thioredoxin in cell growth through interactions with signaling molecules. Antioxid Redox Signal 8: 2143- 2151. 336. Chung JW, Jeon JH, Yoon SR, Choi I (2006) Vitamin D3 upregulated protein 1 (VDUP1) is a regulator for redox signaling and stress-mediated diseases. J Dermatol 33: 662-669. 337. Pandolfi A, Barreyro L, Steidl U (2013) Concise review: preleukemic stem cells: molecular biology and clinical implications of the precursors to leukemia stem cells. Stem Cells Transl Med 2: 143-150.

216