UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Molecular targeting of polo-like kinase 1 (PLK1) for the treatment of brain tumours Lee, Cathy 2013

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

Item Metadata


24-ubc_2013_fall_lee_cathy.pdf [ 318.68MB ]
JSON: 24-1.0074094.json
JSON-LD: 24-1.0074094-ld.json
RDF/XML (Pretty): 24-1.0074094-rdf.xml
RDF/JSON: 24-1.0074094-rdf.json
Turtle: 24-1.0074094-turtle.txt
N-Triples: 24-1.0074094-rdf-ntriples.txt
Original Record: 24-1.0074094-source.json
Full Text

Full Text

MOLECULAR TARGETING OF POLO-LIKE KINASE 1 (PLK1) FOR THE TREATMENT OF BRAIN TUMOURS  by  Cathy Lee  BSc., The University of British Columbia, 2004 MSc., The Univerisity of British Columbia, 2008 	   A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF   DOCTOR OF PHILOSOPHY in The Faculty of Graduate and Postdoctoral Studies  (Experimental Medicine)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)    August 2013   Cathy Lee, 2013  ii ABSTRACT Brain cancer is a disease that is difficult to treat and has a high mortality rate.  Two of the major challenges facing current treatments are disease relapse and drug resistance.  Disease relapse may be attributable to brain tumour initiating cells (BTICs), which are multi-potent stem-like cells endowed with the ability to self-renew and generate differentiated cells that form the bulk of a tumour.  BTICs are notably resistant to conventional chemotherapy and radiation and may therefore survive treatments, repopulate new tumours and cause disease relapse.  Failure to respond to Temozolomide (TMZ), a DNA alkylating agent used as a front-line therapy for glioma, is frequently observed in glioma patients with O6-methylguanine-DNA methyltransferase over-expression.  Furthermore, despite being largely ineffective in the pediatric population, TMZ is routinely used in clinic due to the lack of therapeutic agents that cross the blood-brain-barrier.  There is an imperative need to seek alternative therapeutic strategies to address these issues.  In the studies described herein, we identified polo-like kinase 1 (PLK1) as a novel molecular target for the treatment of the most common malignant  brain tumours in adults and children, glioblastoma multiforme (GBM) and medulloblastoma (MB), respectively.  In our studies, PLK1 is highly expressed in GBM and MB cultured cell lines, tumour tissues and patient-derived primary isolates but not in the normal brain cells including astrocytes, neurons and human neural stem cells.  Targeting PLK1 by siRNA or the small molecule inhibitor BI2536 suppressed cell growth and induced cell cycle arrest accompanied by apoptosis in both GBM and MB.  Of note, PLK1 inhibition had the unique capacity of obliterating GBM and MB BTICs, as well as TMZ-resistant GBM cells.  To further the study, we demonstrated that BI2536 delayed disease progression and prolonged the survival of mice bearing orthotopic GBM or MB brain tumours.  In addition, we established the prognostic importance of PLK1 as an independent marker for patient survival in both GBM and MB, lending further support to targeting this kinase in brain cancer treatment.     iii PREFACE Chapter 2: Polo-like kinase 1 (PLK1) inhibition kills glioblastoma multiforme brain tumour cells in part through loss of SOX2 and delays tumour progression in mice. The experiments were mostly designed and conducted by Cathy Lee (80%) who was also involved in the manuscript preparation and writing.  Contribution of co-authors: § Abbas Fotovati assisted with immunofluorescence staining in Figure 2.2B and Supplementary Figure S2.2. § Joanna Triscott assisted with brain tumour sample processing and contributed to Figure 2.2A/E and Supplementary Figure S2.4D. § James Chen provided the results for Supplementary Figure S2.4B. § Ash Singhal and Christopher Dunham provided patient brain tumour specimens for the studies.  John Kerr assisted with administrative work for the study at BC Children’s Hospital.  Rod Rassekh and Christopher Dunham established the IRB approval by writing the ethics proposal (Ethic Certificate Number H09-02812). § Hiroaki Wakimoto provided patient-derived BTIC lines: GBM4, GBM8 and BT74. § Sheila Singh provided the RNA of normal human neural stem cells hNSC101 and hNSC167. § Chris Jones performed the statistical analyses and provided the data in Figure 2.1 and Supplementary Table S2.1. § Maite Verreault performed the animal studies in Figure 2.5D-E.  § Aarthi Jayanthan performed the colony forming assay in Figure 2.2F and Supplementary Figure S2.4F. § Sandra Dunn provided guidance to the project and assisted with writing and editting of the manuscript.  A version of this chapter was published.  Cathy Lee, Abbas Fotovati, Joanna Triscott, James Chen, Chitra Venugopal, Ash Singhal, Christopher Dunham, John M. Kerr, Maite Verreault, Stephen Yip, Hiroaki Wakimoto, Chris Jones, Aarthi Jayanthan, Aru Narendran, Sheila K. Singh, Sandra E. Dunn (2012). Polo-like kinase 1 (PLK1) inhibition kills glioblastoma multiforme brain tumour cells in part through loss of SOX2 and delays tumour progression in mice. Stem Cells 2012 Jun;30(6):1064-75.    iv Chapter 3: Disulfiram, a drug widely used to control alcoholism, supresses the self-renewal of glioblastoma and over-rides resistance to temozolomide.  Cathy Lee contributed to 1/3 of the figures (excluding supplementary data) published in this manuscript.  The list of figures contributed by CL includes: Figure 3.3 B-C, Figure 3.4 B-C, Figure 3.5C-D, Figure 3.5A-B.  Contribution of co-authors: § Joanna Triscott designed and conducted the majority of the experiments in this project. All the disulfiram experiments were performed by JT.  § Kaiji Hu provided technical support to quantify cells in the growth assays using the Cellomics instrument. § Abbas Fotovati contributed to Figure 3.7. § Rachel Berns and Mary Pambid performed the studies for Figure 3.3A and Figure 3.4A. § Stepehen Yip provided the data for Figure 3.2A. § Brian Toyota provided brain tumour specimens aBT001 and aBT003. § Ash Singhal and Christopher Dunham provided patient brain tumour specimens for the studies.  John Kerr assisted with administrative work for the study at BC Children’s Hospital.  Rod Rassekh and Christopher Dunham established the IRB approval by writing the ethics proposal (Ethic Certificate Number H09-02812).  A version of this chapter was published.  Joanna Triscott, Cathy Lee, Kaiji Hu, Abbas Fotovati, Rachel Berns, Mary Pambid, Margaret Luk, Richard E. Cast, Esther Kong, Eric Toyota, Stephen Yip, Brian Toyota, Sandra E. Dunn (2012). Disulfiram, a drug widely used to control alcoholism, supresses the self-renewal of glioblastoma and over-rides resistance to temozolomide. Oncotarget 2012 Oct;(3)10:1112-23.    v Chapter 4: Personalizing the treatment for medulloblastoma: Polo-like kinase 1 (PLK1) as a molecular target for the sonic hedgehog (SHH) subtype.  Cathy Lee designed and performed most of the in vitro experiments in the study and contributed to 60% of the figures (excluding tables and supplementary data) in the manuscript.  CL was also involved in manuscript preparation and editing.  Contribution of co-authors: § Joanna Triscott prepared the patient information in Table 4.1 and 4.2 and performed statistical analyses for all the patient data in Figure 4.1A-C and 4.2A, C, D and Supplementary Figure S4.1-4.4. § Colleen Foster prepared the data in Supplementary Figure S4.5. § Branavan Manoranjan performed the animal studies and contributed to Figure 4.5D. § Mary Pambid provided the results for Figure 4.1D and Supplementary Figure S4.6C-D. § Abbas Fotovati contributed to Figure 4.2D and Supplementary Figure S4.7. § Rachel Berns assisted with various experiments in the study. § Paul Northcott provided data for Supplementary Table S4.1. § Christopher Maxwell provided guidance and support for the Aurora kinase NanoString study. § Ash Singhal and Christopher Dunham provided patient tumour specimens. CD provided the images for Figure 4.2B and Supplementary Figure S4.7.  § Ash Singhal and Christopher Dunham provided patient brain tumour specimens for the studies.  John Kerr assisted with administrative work for the study at BC Children’s Hospital.  Rod Rassekh and Christopher Dunham established the IRB approval by writing the ethics proposal (Ethic Certificate Number H09-02812).  A version of this chapter was prepared as a manuscript.  Cathy Lee*, Joanna Triscott*, Colleen Foster, Branavan Manoranjan, Mary Rose Pambid, Abbas Fotovati, Rachel Berns, Chitra Venugopal, Katherine O’Halloran, Aru Narendran, Paul Northcott, Michael D. Taylor, Sheila K. Singh, Ash Singhal, Rod Rassekh, Christopher A. Maxwell, Christopher Dunham, Sandra E. Dunn. Personalizing the treatment for medulloblastoma: Polo-Like Kinase 1 (PLK1) as a molecular target for the sonic hedgehog (SHH) subtype.  * Authors contributed equally.    vi TABLE OF CONTENTS  ABSTRACT .................................................................................................................................. ii	  PREFACE .................................................................................................................................... iii	  TABLE OF CONTENTS .............................................................................................................. vi	  LIST OF TABLES ......................................................................................................................... x	  LIST OF FIGURES ...................................................................................................................... xi	  LIST OF ABBREVIATIONS ...................................................................................................... xiii	  ACKNOWLEDGEMENTS ....................................................................................................... xviii	  DEDICATION ............................................................................................................................ xix	  CHAPTER 1: INTRODUCTION .................................................................................................... 1	  1.1	   BACKGROUND .............................................................................................................. 1	  1.2	   MALIGNANT GLIOMAS ................................................................................................. 1	  1.2.1	   GLIOBLASTOMA MULTIFORME (GBM): EPIDEMIOLOGY, GENETIC PREDISPOSITION AND SYMPTOMS ................................................................. 1	  1.2.2	   DIFFUSE INTRINSIC PONTINE GLIOMA (DIPG): EPIDEMIOLOGY, GENETIC PREDISPOSITION AND THERAPIES ................................................................. 3	  1.3	   PATHOLOGY ................................................................................................................. 3	  1.4	   CELL OF ORIGIN ........................................................................................................... 4	  1.5	   MOLECULAR ABERRATIONS ...................................................................................... 5	  1.6	   MOLECULAR PROFILING AND CLASSIFICATIONS .................................................. 6	  1.7	   STANDARD TREATMENT ............................................................................................. 7	  1.7.1	   SURGERY ............................................................................................................... 7	  1.7.2	   RADIATION ............................................................................................................. 8	  1.7.3	   CHEMOTHERAPY .................................................................................................. 9	  1.8	   TEMOZOLOMIDE (TMZ) .............................................................................................. 10	  1.8.1	   BACKGROUND .................................................................................................... 10	  1.8.2	   PHARMACOLOGY ............................................................................................... 10	  1.8.3	   MECHANISM OF ACTION .................................................................................... 11	  1.8.4	   MECHANISM OF RESISTANCE .......................................................................... 11	   O6- METHYLGUANINE DNA METHYLTRANSFERASE (MGMT) ............... 11	   ALTERNATIVE MODES FOR TMZ RESISTANCE ...................................... 13	  1.9	   TMZ TREATMENT IN CHILDREN ............................................................................... 14	  1.10	   NOVEL THERAPIES FOR GBM TREATMENT ........................................................... 15	  1.10.1	   MOLECULAR TARGETED THERAPY ............................................................... 15	  1.10.2	   DISULFIRAM (DSF) ............................................................................................ 16	  1.11	   MEDULLOBLASTOMA (MB) ....................................................................................... 17	  1.11.1	   EPIDEMIOLOGY, GENETIC PREDISPOSITION, SYMPTOMS AND PROGNOSIS ....................................................................................................... 17	   vii 1.11.2	   PATHOLOGY ...................................................................................................... 19	  1.11.3	   CELL OF ORIGIN ............................................................................................... 20	  1.11.4	   MOLECULAR ABERRATIONS .......................................................................... 21	   SONIC HEDGEHOG (SHH) PATHWAY ..................................................... 21	   WNT PATHWAY ......................................................................................... 23	   ADDITIONAL MOLECULAR ABERRATIONS IDENTIFIED ...................... 25	  1.11.5	   MOLECULAR PROFILING AND CLASSIFICATIONS ....................................... 26	  1.11.6	   STANDARD TREATMENT ................................................................................. 29	   SURGERY AND RADIATION ..................................................................... 29	   CHEMOTHERAPY ...................................................................................... 30	  1.12	   POLO-LIKE KINASE 1 (PLK1) .................................................................................... 31	  1.12.1	   BACKGROUND .................................................................................................. 31	  1.12.2	   FAMILY MEMBERS ............................................................................................ 32	   PLK1 ............................................................................................................ 32	   PLK2 ............................................................................................................ 33	   PLK3 ............................................................................................................ 33	   PLK4 ............................................................................................................ 34	   PLK5 ............................................................................................................ 34	  1.12.3	   PROTEIN STRUCTURE ..................................................................................... 35	  1.12.4	   BIOLOGY AND FUNCTIONS ............................................................................. 36	  1.12.5	   ROLE IN CANCER PATHOGENESIS ................................................................ 41	  1.12.6	   INHIBITORS AND CLINICAL TRIALS ............................................................... 42	  1.13	   CANCER STEM CELL THEORY ................................................................................. 43	  1.13.1	   DEFINITION OF CANCER STEM CELLS .......................................................... 45	  1.13.2	   IDENTIFICATION AND ISOLATION OF BRAIN CANCER STEM CELLS (BCSCs) .............................................................................................................. 46	   NEUROSPHERE ASSAY ............................................................................ 46	   CELL SURFACE ANTIGEN SORTING ...................................................... 47	   CD133 (PROMININ 1, PROM1) ........................................................... 47	   CD15 (LEWIS x, STAGE-SPECIFIC EMBRYONIC ANTIGEN 1, SSEA-1) ...................................................................................................... 48	  1.13.3	   ALTERNATIVE METHODS FOR BCSC ENRICHMENT .................................... 48	   HOECHST 33342 EXCLUSION .................................................................. 48	   ALDEHYDE DEHYDROGENASE ACTIVITY ASSAY ................................ 49	  1.13.4	   THERAPEUTIC STRATEGIES TARGETING BCSCs ....................................... 50	   PATHWAY-SPECIFIC SMALL MOLECULE INHIBITORS ........................ 50	   DIFFERENTIATION THERAPY .................................................................. 51	   ONCOLYTIC VIRUSES ............................................................................... 52	  1.14	   RATIONALE AND HYPOTHESIS OF THE STUDY .................................................... 52	  CHAPTER 2: POLO-LIKE KINASE 1 (PLK1) INHIBITION KILLS GLIOBLASTOMA MULTIFORME BRAIN TUMOUR CELLS IN PART THROUGH LOSS OF SOX2 AND DELAYS TUMOUR PROGRESSION IN MICE. ......................................................................................... 53	  2.1	   INTRODUCTION ........................................................................................................... 53	  2.2	   RESULTS ..................................................................................................................... 54	  2.3	   DISCUSSION ................................................................................................................ 58	  2.4	   MATERIALS AND METHODS ..................................................................................... 61	   viii 2.5	   FIGURES ...................................................................................................................... 66	  2.6	   SUPPLEMENTARY DATA ........................................................................................... 76	  CHAPTER 3: DISULFIRAM, A DRUG WIDELY USED TO CONTROL ALCOHOLISM, SUPPRESSES THE SELF-RENEWAL OF GLIOBLASTOMA AND OVER-RIDES RESISTANCE TO TEMOZOLOMIDE. ....................................................................................... 83	  3.1	   INTRODUCTION ........................................................................................................... 83	  3.2	   RESULTS ..................................................................................................................... 84	  3.3	   DISCUSSION ................................................................................................................ 86	  3.4	   MATERIALS AND METHODS ..................................................................................... 87	  3.5	   FIGURES ...................................................................................................................... 91	  3.6	   SUPPLEMENTARY DATA ........................................................................................... 98	  CHAPTER 4: PERSONALIZING THE TREATMENT FOR MEDULLOBLASTOMA: POLO-LIKE KINASE 1 (PLK1) AS A MOLECULAR TARGET FOR THE SONIC HEDGEHOG (SHH) SUBTYPE. ................................................................................................................................ 102	  4.1	   INTRODUCTION ......................................................................................................... 102	  4.2	   RESULTS ................................................................................................................... 103	  4.3	   DISCUSSION .............................................................................................................. 106	  4.4	   MATERIALS AND METHODS ................................................................................... 108	  4.5	   TABLES ...................................................................................................................... 113	  4.6	   FIGURES .................................................................................................................... 115	  4.7	   SUPPLEMENTARY DATA ......................................................................................... 123	  CHAPTER 5: DISCUSSION ..................................................................................................... 131	  5.1	   SUMMARY .................................................................................................................. 131	  5.2	   IMPLICATIONS AND FUTURE DIRECTIONS ........................................................... 133	  5.2.1	   PLK1 INHIBITION SUPPRESSES THE SELF-RENEWAL OF GBM BTICs ..... 133	  5.2.2	   PLK1 INHIBITION HELPS OVERCOME TMZ RESISTANCE IN GBM CELLS . 139	  5.2.3	   PLK1 INHIBITION AS A POTENTIAL THERAPEUTIC STRATEGY FOR SHH MB .......................................................................................................................... 141	  5.3	   CONCLUDING REMARKS: PLK1 AS A MOLECULAR TARGET IN GBM AND MB …………………………………………………………………………………………...144	  REFERENCES ......................................................................................................................... 146	  Appendix A: Patient-derived brain tumour cells and their response to PLK1 inhibitor. . 203	  Appendix B: PLK1 inhibition does not affect the phosphorylation of STAT3 at Ser727 and the expression of SOX2. ........................................................................................................ 204	  Appendix C: MAPK inhibition does not affect the growth, survival and SOX2 expression of SF188 cells. ......................................................................................................................... 205	   ix Appendix D: Cell sorting of SF188 based on CD133 expression. ...................................... 206	  Appendix E: CD15-positive and CD15-negative Daoy cells express similar levels of SOX2, musashi, Bmi1 and PLK1. ...................................................................................................... 207	  Appendix F: ALDH inhibition does not significantly affect the growth of SF188 and U251 cells. ......................................................................................................................................... 208	      x LIST OF TABLES Table 1.1 The substrates of PLK1 in G2/M transition and the different stages of mitosis. ............................................................................................................................... 40	  Table S2.1 Correlation between PLK1 and similary expressed genes expressed in primary GBM based on Affmetrix U133. .......................................................................... 76	  Table 4.1 Summary of the pediatric MB patients included in the study cohort. ............... 113	  Table 4.2 Univariate and multivariate analyses of clinical, pathological and biological endpoints. ....................................................................................................... 114	      xi LIST OF FIGURES Figure 1.1 The SHH pathway. .................................................................................................. 23	  Figure 1.2 The WNT pathway. .................................................................................................. 24	  Figure 1.3 The structure of PLKs. ........................................................................................... 36	  Figure 1.4 The regulation of CDK1 by PLK1 in the G2/M transition. .................................... 37	  Figure 1.5 Therapeutic predictions of the cancer stem cell model. ..................................... 44	  Figure 2.1 PLK1 is highly expressed in primary GBM where it is associated with the proliferative subtype and poor survival. ................................................................... 66	  Figure 2.2 PLK1 is over-expressed in BTICs and its inhibition suppresses the self-renewal of these cells. ....................................................................................................... 68	  Figure 2.3 PLK1 inhibition suppresses cell growth and induces apoptosis in brain cancer cells SF188. ............................................................................................................ 70	  Figure 2.4 PLK1 inhibition down-regulates the expression of SOX2, which is required for the growth and survival of GBM cells. ....................................................... 72	  Figure 2.5 BI2536 suppresses the growth of GBM cell line U251 in vitro and tumour formation in vivo. ................................................................................................. 74	  Figure S2.1 PLK1 and neural stem cell markers SOX2, musashi and Bmi1 are co-expressed in dissociated BT74 cells. .............................................................................. 77	  Figure S2.2 PLK1 level is elevated in BTICs compared to normal human neural stem cells and its expression is significantly down-regulated after lineage differentiation. .................................................................................................................... 78	  Figure S2.3 BI2536 suppresses tumoursphere formation of primary brain tumour cells but exerts a minimal effect on the differentiation of primary hematopoietic stem cells isolated from a patient. .......................................................... 79	  Figure S2.4 PLK1 inhibition represses the cell growth of pediatric and adult GBM cell lines SF188 and Gli36. ................................................................................................ 81	  Figure S2.5 PLK1 knockdown by two targeting siRNAs decreases the transcript and protein levels of SOX2 and alters cellular morphology. ......................................... 82	  Figure 3.1 DSF inhibits GBM cell growth and self-renewal. ................................................. 91	  Figure 3.2 Freshly isolated GBM cells are sensitive to DSF yet resistant to TMZ. ............. 92	  Figure 3.3 DSF inhibits the expression of PLK1 in pediatric GBM SF188 cells. ................. 93	  Figure 3.4 DSF inhibits the expression of PLK1 in adult GBM U251 cells. ......................... 94	  Figure 3.5 Targeting PLK1 inhibits growth of drug resistant cells with up-regulated PLK1 protein. ..................................................................................................................... 95	  Figure 3.6 PLK1 inhibitors can be used to over-come TMZ resistance. .............................. 96	  Figure 3.7 aBT001 and aBT003 express high levels of PLK1. .............................................. 97	  Figure S3.1 SF188 cells are TMZ resistant. ............................................................................ 98	  Figure S3.2 Combination treatment of DSF augments TMZ cytotoxicity. ........................... 99	   xii Figure S3.3 High doses of DSF are safe for normal human astrocytes. ........................... 100	  Figure S3.4 DSF inhibits the expression of PLK1. .............................................................. 101	  Figure 4.1 Molecular characteristics of MB dictate outcome and offer potential drug targets. ..................................................................................................................... 115	  Figure 4.2 PLK1 expression correlates with poor patient survival. ................................... 116	  Figure 4.3 The response of primary MB cells to BI2536 is correlated with the expression of PLK1 in the tumours. .............................................................................. 117	  Figure 4.4 PLK1 inhibition suppresses cell growth and induces apoptosis in Daoy cells. .................................................................................................................................. 119	  Figure 4.5 PLK1 inhibition suppresses tumoursphere formation in vitro and delayed disease progression in vivo. ............................................................................ 121	  Figure S4.1 PLK1 is a tumour-specific protein that is highest in the SHH subtype. ........ 124	  Figure S4.2 AURKA expression in MB. ................................................................................. 125	  Figure S4.3 Patient relapse and outcome in children over the age of 3 years old. .......... 126	  Figure S4.4 Only children who received radiation were included to avoid this treatment as a confounding variable. ............................................................................ 127	  Figure S4.5 PLK1 protein levels were prognostic for relapse. ........................................... 128	  Figure S4.6 The response of MB cell lines to BI2536 and chemotherapeutic agents. .............................................................................................................................. 129	  Figure S4.7 Daoy xenografts were similar to large cell anaplastic medulloblastoma. ............................................................................................................ 130	  Figure 5.1 Model for the potential therapeutic benefit of PLK1 inhibitors in brain cancer treatment. ............................................................................................................. 133	      xiii LIST OF ABBREVIATIONS AA   Anaplastic astrocytoma  ABC   Adenosine triphosphate binding-cassette  ALDH   Aldehyde dehydrogenase  ANOVA  Analysis of variance APC   Adenomatous polyposis coli  APC/C   Anaphase promoting complex or cyclosome ARF   ADP ribosylation factor ATCC   American tissue culture collection ATRT   Atypical teratoid/rhabdoid tumour  BAAA   BODIPY aminoacetaldehyde  BBB   Blood-brain-barrier   BCNU   Bis-chloroethylnitrosourea  BCRP   Breast cancer resistant protein   BCSC   Brain cancer stem cell BDNF   Brain-derived neurotrophic factor  BER   Base excision repair  b-FGF   Basic fibroblast growth factor  BMP   Bone morphogenetic protein    BTIC   Brain tumour initiating cell CCG   Children’s Cancer Group CCIC   Colon cancer initiating cell CCNU   N-(2-chloroethyl)-N'-cyclohexyl-N-nitrosourea  CCSG   Children’s cancer study group  CDK   Cyclin-dependent kinase   CNS   Central nervous system   xiv CPBTC  Canadian pediatric brain tumour consortium  CSF   Cerebral spinal fluid CSC   Cancer stem cell  CT   Computerized tomography Cu   Copper DIPG   Diffuse intrinsic pontine glioma DLT   Dose-limiting toxicity    DN   Desmoplastic/nodular  DSH   Dishevelled  DSF   Disulfiram  DSF/Cu  Disulfiram/copper  DTIC   Dacarbazine EGF   Epidermal growth factor EGFR   Epidermal growth factor receptor EMI-1   Early mitotic inhibitor-1  EORTC/NCIC European Organization for the Research and Treatment of Cancer and the National Cancer Institute FNK   FGF-inducible kinase  FSTL   Follistatin-like FTI   Farnesyl transferase inhibitor  FZD   Frizzled GBM   Glioblastoma multiforme  GFAP   Glial fibrillary acidic protein GIC   Glioma-initiating cell GNP   Granule neuron precursor  GSK   Glycogen synthase kinase    GTSE   G2- and S-phase-expressed   xv HA   Human astrocyte HDAC   Histone deacetylase HIF   Hypoxia-inducible factor  HRPC   Hormone refractory prostate cancer    HSP   Heatshock protein   IAP   Inhibitor of apoptosis protein IDH   Isocitrate dehydrogenase INK   Inhibitor of cyclin dependent kinase IGF   Insulin-like growth factor  LCA   Large-cell/anaplastic    MDM   Mitochondrial distribution and morphology MB   Medulloblastoma  MBEN   MB with extensive nodularity  MGMT   O6-methylguanine DNA methyltransferase  MKLP   Mitotic kinesin-like protein    MMR   Mismatch repair  MPF   M-phase-promoting factor  MRI   Magnetic resonance imaging  MTIC   5-(3-methyltriazen-l-yl)imidazole-4-carboxamide  mTOR   Mammalian target of rapamycin  NGF   Nerve growth factor  NLP   Ninein-like protein  N3-meA  N3-methyladenine  NSC   Neural stem cell  NSCLC  Non-small cell lung cancer O6-BG   O6-benzylguanine  xvi oHSV   Oncolytic herpes simplex virus O6-MG   O6-methylguanine  PARP   Poly (ADP-ribose) polymerase  PCR   Polymerase chain reaction PB   Polo-box PBD   Polo-box domain  PDGFA  Platelet-derived growth factor receptor-α  PI3K   Phosphatidylinositol 3-kinase  PK   Pharmacokinetic  PLK   Polo-like kinase PNET   Primitive neuroectodermal tumour  PTCH1  Patched  PTEN   Phosphatase and tensin homology RB   Retinoblastoma  REST   RE1-silencing transcription factor ROS   Reactive oxygen species  RT   Radiation therapy  RTK   Receptor tyrosine kinase   SHH   Sonic hedgehog  SLK   STE20-like kinase  SMO   Smoothened  SOX   Sex determining region Y-box SNK   Serum-inducible kinase SSEA   Stage-specific embryonic antigen     SP   Side population  STK   Serine/threonine kinase   xvii SUFU   Suppressor of fused  SVV-001  Seneca valley virus-001    SVZ   Subventricular zone  S100A4  S100 calcium binding protein A4     TBP   TATA-box binding protein TCF/LEF  T-cell factors/lymphoid enhancer factor TCGA   The cancer genome atlas  TIC   Tumour initiating cell  TMA   Tissue microarray TMZ   Temozolomide TUNEL Terminal deoxynucleotidyl transferase biotin-dUTP nick end labeling    VEGF   Vascular endothelial growth factor  VZ   Ventricular zone WHO   World Health Organization  YB-1   Y-box binding protein-1      xviii ACKNOWLEDGEMENTS I began my research career in 2005 Sandi’s lab where I pursued my MSc. training and graduated in 2007.  At the time I was unsure whether I wanted to pursue a higher degree; therefore, I took a year (2008) working as a research technician to gain new research experience and direction.  During that period of time, I worked with a group of senior scientists on a pediatric cancer project, which I felt deeply engaged with and interested in. The one-year work experience in 2008 was a turning point in my research career as it allowed me to be involved in a project which I truly felt passionate about and committed to.  Therefore, in 2009, I decided to return to the lab to pursue a PhD.  Eight years sounds like a long time but it flew by really quickly in the Dunn lab.  I feel very fortunate to have worked in a lab where everyone is eager to work collaboratively, to foster a scientific environment together as a team, and to provide each other with emotional support as friends.  I would like to first thank my supervisor, Dr. Sandra Dunn, for giving me this precious opportunity to pursue my graduate studies in such a wonderful lab.  It was Sandi who took my hands, led me into the oncology field, and introduced me to the wonder of science, which I fell in love with.  Sandi has helped me tremendously in my training over the years and has inspired me in many different ways.  I am truly grateful for what she has done for me. I also want to thank all my colleagues for always being so kind and helpful assisting me with experiments and everything else.  I would like to express my gratitude, especially to Drs. Anna Stratford and Abbas Fotovati as well as all the students in the Dunn Lab Brain Team: Joanna Triscott, Mary Pambid, James Chen and Rachel Berns.  Anna is my first go-to person when I need technical and emotional support and I am really grateful having an amazing friend like her in the lab.  Abbas reminds me of the "Master Shifu" in Kung-Fu Panda.  He always provides me with wise and philosophical words on life.  Joanna, Mary, James and Rachel are the nicest teammates I could ever have.  They have helped me so much with all the studies in my projects.  I certainly cherish the friendship I have developed with everyone else in the lab: Kristen, Alastair, Kaiji and Brenda.  I will miss everyone very much. Furthermore, I would like to acknowledge our collaborators: Dr. Christopher Dunham, Dr. Stephen Yip and Dr. Sheila Singh for helping us obtain patient tumour specimens and for their contribution to the study.  I also want to thank my supervisory committee: Dr. Catherine Pallen, Dr. Christopher Maxwell and Dr. Michael Cox.  I am very grateful to have a committee that is always so helpful to me.  Thank you all so much for your guidance and support. This research could not be accomplished without the financial assistance from the Michael Smith Foundation for Health Research and the Hannah's Heroes Foundation whom I regard as my research family.  I will always remember the spirit of the Hannah's Heroes.  Thank you Mr. Tore Hatlen, Cheryl, Lisa, Nicole, my favourite grandma Dorothy and all the other members of the Foundation’s Board and Committee. I would like to thank my dear parents.  They are my guardian angels who give me unconditional love and support throughout these years.  Being their daughter is one of the best things in my life.  I love you Mom and Dad. Last but not least, I want to acknowledge my husband, Andy, for his love, patience and full support of my studies despite the “conflict-of-interest.”  I really thank him for always being there for me and for coming to visit me whenever he could.  I am truly blessed to be with such a wonderful man.   xix   DEDICATION   To my family & The Hannah's Heroes Foundation                                                                                                         1 CHAPTER 1: INTRODUCTION 1.1 BACKGROUND Brain cancer is a disease that is difficult to treat and has a high mortality rate.  Since intracranial solid neoplasms are located in the confined area of the brain parenchyma and calvarium, which controls various vital functions of the body, their impact on patients is immediate and devastating.  Every year, approximately 2,800 Canadians are diagnosed with brain cancers [Canadian Cancer Registry database at Statistics Canada].  Brain tumours do not occur frequently in adults and fall outside the list of the top-ten most common cancer types in the National Cancer Institute 2012 study [http://www.cancer.gov/].  However, brain tumours account for a disproportionally high number of cancer-related deaths due to the aggressive nature of the disease and the limitations in current therapies.  We selected two brain cancer models to study in this project: glioblastoma multiforme (GBM) and medulloblastoma (MB), which represent the most common malignant brain tumours in adult and pediatric populations, respectively.  1.2 MALIGNANT GLIOMAS 1.2.1 GLIOBLASTOMA MULTIFORME (GBM): EPIDEMIOLOGY, GENETIC PREDISPOSITION AND SYMPTOMS The annual incidence of malignant gliomas (grade III and IV) is approximately 5 cases per 100,000 people (Louis et al., 2007).  GBM is diagnosed in 60-70% of these patients, with a medium age of 64 years (Fisher et al., 2007; Wen and Kesari, 2008).  There is a gender and ethnicity preponderance of male over female (40% higher) and of Caucasians over non-Caucasians (two-fold higher) (Wen and Kesari, 2008).  GBM is a grade IV astrocytoma based on the 2007 World Health Organization (WHO) classification of tumours of the central nervous system (Louis et al., 2007).  The high grade assigned to GBM reflects the cytologically malignant nature of the tumours, which are mitotically active and necrosis-prone.  The tumour is preferentially located in the cerebral hemispheres (Brandes et al., 2008) and often follows a rapid pre- and post-operative clinical course that is associated with a fatal outcome (Louis et al., 2007).  The median survival of GBM is approximately 14 months in patients receiving radiation and chemotherapy (Stupp et al., 2005). The underlying cause for the majority of GBM remains unclear.  Exposure to ionizing radiation is the only established risk factor (Fisher et al., 2007).  Other proposed risks include head injury, foods containing N-nitroso compounds, cellular phones and electromagnetic fields; however the evidence for these is inconclusive (Fisher et al., 2007).  Familial inheritance accounts for approximately 5% of GBM cases.  Specifically, mutations in genes such as NF1  2 (encodes neurofibromin 1), TP53 (encodes p53) and CTNNB1 (encodes β-catenin) predispose individuals to familial syndromes: neurofibromatosis types 1 and 2, the Li-Fraumeni syndrome, and Turcot syndrome, respectively, as well as brain cancers (Kyritsis et al., 2010).  However, most familial cases suggest genetic associations rather than genetic causes to brain tumours. To date, no primary prevention or screening procedures are advised for brain cancers.  As a result, tumours are normally not detected until patients present with neurological symptoms such as confusion, memory loss, motor weakness, headache, seizure and progressive neurological deficit.  For many patients, brain tumours are diagnosed several months after the initial symptoms are manifested.  Therefore, brain computerized tomography (CT) or magnetic resonance imaging (MRI) scanning performed at a first occurrence of new neurological symptoms or epileptic seizures is crucial for detecting the disease at an earlier stage (Brandes et al., 2008). Unlike most cancers that follow a progressive course of development, GBM is an anomaly and does not always follow this pattern of disease evolution.  Secondary GBM may develop from a lower grade astrocytoma (eg. diffuse WHO grade II astrocytoma or WHO grade III anaplastic astrocytoma), with a traceable history of disease development, generally over a period of 5-10 years (Holland, 2001; Maher et al., 2001).  However, most diagnosed cases are considered the “primary GBM” and occur de novo after a very brief clinical history without signs of a less malignant precursor lesion [reviewed by (Ohgaki and Kleihues, 2012)].  Due to the lack of obvious symptoms prior to the manifestation of a full-blown disease, patients diagnosed are often at the terminal stage when therapies become largely ineffective.   Primary and secondary GBM are characterized by different genetic abnormalities.  Primary GBM is believed to arise directly from the malignant transformation of astrocytes or precursors with genetic aberrations such as INK4A loss (Biernat et al., 1997b), human epidermal growth factor receptor (EGFR) amplification/mutation (Lang et al., 1994; Schlegel et al., 1994), phosphatase and tensin homology (PTEN) loss (Tohma et al., 1998) and MDM2 amplification (Biernat et al., 1997a).  Secondary GBM, on the other hand, may be triggered by an “initiation pathway” (Zhu and Parada, 2002) involving the loss of NF1 (Reilly et al., 2000) and TP53 (Rasheed et al., 1994; von Deimling et al., 1992; Watanabe et al., 1997), IDH1/2 mutations (Yan et al., 2009) or over-expression of PDGF/PDGFR (Guha et al., 1995; Hermanson et al., 1992; Yamaguchi et al., 1994) in astrocytes or precursors.  This leads to the formation of low-grade tumours such as grade I pilocytic astrocytoma and grade II astrocytoma.  A “progression pathway” (Zhu and Parada, 2002) entailing RB1 loss (Burns et al., 1998;  3 Ichimura et al., 1996; Ueki et al., 1996) or CDK4 amplification (Burns et al., 1998; Ichimura et al., 1996) subsequently follows and brings the disease to its final stage. 1.2.2 DIFFUSE INTRINSIC PONTINE GLIOMA (DIPG): EPIDEMIOLOGY, GENETIC PREDISPOSITION AND THERAPIES  While GBM is more common in adult patients, diffuse intrinsic pontine glioma (DIPG) is an aggressive form of glioma that is more prevalent in children with 3-year survival rates of only 5-10%.  Brainstem tumours account for 10-20% of all pediatric tumours and DIPG represents the majority (80%) of brainstem tumours (Recinos et al., 2007).  Most DIPG tumours are histologically classified as grade II fibrillary or grade III-IV high-grade gliomas (anaplastic astrocytomas or GBMs).  The peak incidence of this disease is between 5-10 years of age (Karajannis et al., 2008).    Due to the anatomical location of DIPG, these tumours are often diagnosed in the absence of a biopsy.  The limited access of tissue for study has hindered the ability to decipher the molecular underpinning of this disease.  Research has nevertheless revealed EGFR over-expression (Gilbertson et al., 2003) and p53 mutations (Cheng et al., 2000; Louis et al., 1993) to be common in brainstem gliomas.  Recent molecular analyses identified frequent genetic alterations in platelet-derived growth factor receptor-α (PDGFα), poly (ADP-ribose) polymerase (PARP1) (Paugh et al., 2011; Zarghooni et al., 2010), MET (Paugh et al., 2011), cell cycle regulators such as Aurora B (Buczkowicz et al., 2012) and chromatin modifiers such as histone H3.3 (Schwartzentruber et al., 2012; Khuong-Quang et al., 2012) , suggesting that these proteins may serve as novel molecular targets for the treatment of this disease.  DIPGs are often inoperable due to their intimate association with the brainstem; therefore radiation is the mainstay therapy.  Although the majority of patients respond initially to radiotherapy, disease recurrence is inevitable (Recinos et al., 2007).  The addition of chemotherapy, including high-dose chemotherapy, does not improve patient outcome (Massimino et al., 2008).  DIPG remains one of the most challenging brain cancers to treat; novel therapies are therefore urgently needed. 1.3 PATHOLOGY The WHO classifies gliomas into four prognostic grades based on the histological features of tumours: grade I (pilocytic astrocytoma), grade II (diffuse astrocytoma), grade III (anaplastic astrocytoma) and grade IV (glioblastoma multiforme) (Louis et al., 2007).  Grade III and IV tumours are malignant gliomas, which are invasive, histologically heterogeneous, and tend to grow infiltratively.   4 Anaplastic astrocytomas are diffusely infiltrating brain tumours that display focal or dispersed anaplasia and are characterized by increased cellularity, nuclear atypia and higher mitotic activity.  GBM tumours are distinguished from grade III tumours by having areas of microvascular proliferation and necrosis (Brat and Van Meir, 2004).  Atypical glioblastoma variants include: 1) small-cell glioblastomas which have amplified EGFR gene, 2) giant-cell glioblastomas which are characterized by multi-nucleated giant cells, 3) gliosarcomas which contain a prominent sacomatous element and 4) glioblastomas with oligodendroglial features, which are correlated with a more optimistic prognosis than standard glioblastomas (Louis et al., 2007).  The histological heterogeneity of GBM is attributable to the composition of tumours by both the neoplastic and stromal tissues.  Therefore, examinations of tumours at the molecular level by gene expression profiling or integrated genomic analyses have become a useful tool in categorizing or distinguishing tumours with similar or different biologies. 1.4 CELL OF ORIGIN Most glioma classifications made over 80 years ago proposed that GBM arises from astrocytes (Louis et al., 2001).  Yet, in mature adult brains, differentiated brain cells such as astrocytes are not known to undergo active cell division, which would be a prerequisite for tumour formation.  Glial cells might be subject to neoplastic transformation during reactive proliferation evoked by trauma, but no definitive conclusion can be drawn from brain injury and the development of glial brain tumours from epidemiological studies (Louis, 2006).  Recent studies have suggested that GBM may arise from more “primitive” cells, namely the stem and progenitor cells, in the higher levels of the brain cell hierarchy (Galli et al., 2004; Hemmati et al., 2003; Singh et al., 2003).  The neural stem cells (NSCs), found in discrete regions of the brain, namely the subventricular zone (SVZ) of the lateral ventricles and dentate gyrus of the hippocampus of the adult brain, are proliferative, multi-potent, and migratory (Alvarez-Buylla and Lois, 1995).  These features, along with the longevity of NSCs, make them ideal targets for malignant transformation.  The highly proliferative progenitor cells, when introduced with key genetic mutations, may also be able to seed a new tumour.  In transgenic mouse studies, oncogenic transformation by RAS or AKT provokes brain tumour formation more readily in the nestin-expressing (more primitive) glial progenitor cells than in the GFAP-expressing (more mature) differentiated cells (Holland et al., 2000).  However, conflicting experimental evidence indicates that differentiated astrocytes could revert to a more “primitive” cell with a progenitor  5 phenotype when CDKN2A/p16 was deleted (Bachoo et al., 2002).  This potential plasticity of cells may complicate the identification of the cell-of-origin of GBM. 1.5 MOLECULAR ABERRATIONS A number of molecular aberrations have been identified in GBM, some of which are exclusive to primary or secondary GBM.  The PTEN/EGFR/AKT/mTOR signaling pathway is frequently activated in primary GBM.  EGFR amplification is observed in ~40% of primary GBM (Ekstrand et al., 1992) but rarely in secondary GBM (Watanabe et al., 1996).  In addition to EGFR amplification, a truncated version of the gene named EGFRvIII which is characterized by the deletion of exons 2 to 7, is found almost exclusively in adult patients and is associated with constitutive activation of the receptor.  The EGFRvIII variant, which is found in 50-60% of the patients with amplification of the same gene (Frederick et al., 2000; Sugawa et al., 1990), confers a growth advantage to tumour cells, in part through activation of the phosphatidylinositol 3-kinase (PI3K)/AKT pathway and also down-regulation of p27 (Narita et al., 2002).  The involvement of oncogenic receptor tyrosine kinases (RTKs) in GBM is further demonstrated by the PDGFRA.  Enhanced signaling through this receptor is observed in a significant subset of GBM (Di Rocco et al., 1998; Westermark et al., 1995).  Furthermore, PTEN, the major negative modulator of the PI3K/AKT pathway, is mutated in 15-40% of GBM (Dahia, 2000; Knobbe et al., 2002) and almost exclusively in primary GBM (Tohma et al., 1998).  Together, the over-expression and mutation of EGFR or PDGFR and/or inactivation of PTEN enhances signaling through the PI3K/AKT/mTOR axis, leading to increased cell growth and survival.    Evasion of apoptosis could be achieved by inactivating mutation of TP53 (Newcomb et al., 1993) or amplification of MDM2/4, both of which are commonly seen in GBM patients (Rao et al., 2010; Reifenberger et al., 1994).  Consistent with this, Li-Fraumeni syndrome patients carrying familial mutation in the TP53 gene are more susceptible to various human malignancies including low-grade astrocytic tumours and secondary GBM (Srivastava et al., 1990) (Malkin et al., 1990).  Furthermore, the stringent control at the G1/S-phase of the cell cycle is abrogated in brain tumour cells through mutations in various key proteins.  For example, the genes that encode cyclin-dependent kinase 4 and 6 (CDK4 and 6), which drive cell cycle progression and antagonize the effect of retinoblastoma (RB1), are amplified in GBM (Costello et al., 1997).  In addition, inactivating mutations or promoter methylation of the RB1 gene occurs in GBM patients with higher frequencies in the secondary GBM tumours (Nakamura et al., 2001).  The deletion of the CDKN2A locus decreases the expression of INK4A and ADP ribosylation factor  6 (ARF), which are key positive regulators of RB and p53.  In essence, GBM is a pathological consequence of deregulated cell cycle, accelerated cell proliferation and enhanced cell survival invoked by genetic mutations that are found in the vast majority of human cancers. 1.6 MOLECULAR PROFILING AND CLASSIFICATIONS The diagnosis of brain tumours has long been based on clinicopathological assessment.  This is a valuable approach that allows the determination of tumour type, grade and stage of the disease, which are correlated with clinical outcome (Van Meir et al., 2010).  Recent studies from large-scale gene expression profiling reveal that an additional level of molecular heterogeneity may exist between different types of tumours as well as within the same histopathological category of tumours.  The transcriptional profiles of GBM reflect the detailed, underlying tumour biology, which cannot be discerned by conventional pathological analyses and is believed to outperform conventional morphological analyses in classifying the disease (Godard et al., 2003; Nutt et al., 2003) and in predicting patient outcome and response to treatment (Van Meir et al., 2010). Molecular classification by transcriptional profiling unveiled the associations between glioma biology and neuroglial developmental stages (Brennan et al., 2009; Phillips et al., 2006; Verhaak et al., 2010).  The first proposed classification schemes segregated a cohort of WHO grade III and IV malignant gliomas into three subtypes: proneural, proliferative and mesenchymal, each of which exhibits distinct gene signatures resembling those of fetal and adult brain, hematopoietic stem cells and various soft tissues, respectively.  Interestingly, the proneural subclass is composed mainly of nearly all WHO grade III tumours in the study and a subset of GBMs from younger patients with optimal survival (Hui et al., 2001).  The tumours in the proliferative and mesenchymal subtypes show poor prognosis and are characterized by gene expression profile associated with cell proliferation and angiogenesis, respectively (Phillips et al., 2006). The study from The Cancer Genome Atlas (TCGA) has further corroborated the previous findings.  Five hundred untreated, primary GBM specimens were collected from major brain tumour centres in the United States and were subjected to molecular analyses encompassing gene sequencing, gene copy number alterations, epigenetic methylation and gene expression profiling.  The results from the TCGA study are largely consistent with previous reports, with the addition of one more subclass: neural, and re-definition of the proliferative subclass as classical (Verhaak et al., 2010).  The four well-known subtypes of GBM: classical, mesenchymal,  7 proneural and neural, and the molecular signatures associated with each of them will be discussed below (Verhaak et al., 2010). The “classical” GBM is characterized by highly proliferative cells with chromosome 7 gain, chromosome 10 loss (93%) and EGFR gene amplification.  These tumours respond to conventional radiation and chemotherapies, likely due to the presence of functional p53 in this group of patients.  The classical subtype of GBM have increased expression of the neural precursor and stem cell marker NES, and signaling molecules in the Notch (NOTCH3, JAG1 and LFNG) and SHH (SMO, GAS1 and GLI2) pathways. The “mesenchymal” GBM, with a gene expression profile associated with mesenchyme and angiogenesis, exhibits frequent inactivation of the NF1 (37%), TP53 (32%) and PTEN (32%) genes.  MET, CD44 and genes in the TNF superfamily and NFκB family are often over-expressed in these tumours, which initially respond to aggressive chemotherapies. The “proneural” subclass of GBM shows elevated expression of genes involved in neuronal development, such as PDGFRA, OLIG2, TCF3 and NKX2-2 (oligodendrocytic) and SOX, DCX, DLL3, ASCL1 and TCF4 (proneural).  Younger patients as well as patients having tumours with PDGFRA amplification/mutation, isocitrate dehydrogenase 1 (IDH1) (30%) or TP53 mutations (54%) frequently fall in this category.  The fact that IDH1/2 gene mutation is found in lower-grade gliomas suggests that secondary GBM, which exhibits a progressive course of development, might belong to the “proneural” subclass (Parsons et al., 2008; Verhaak et al., 2010).  Patients in this subclass have slightly better survival than those of the other subclasses and may respond well to inhibitors of the hypoxia-inducible factor (HIF), PI3K and PDGFRA pathways. The “neural” subclass of GBM is not as well understood and has gene signatures similar to those of normal brain tissues, with activation of neuron markers including NEFL, GABRA1, SYT1 and SLC12A5. Molecular profiling and classification of GBM not only provides us with a more in-depth understanding of tumour biology, which is the underlying force that drives specific tumour behaviour, but also guides us to design more personalized therapies that will selectively target the Achilles’ heel of the disease. 1.7 STANDARD TREATMENT 1.7.1 SURGERY Surgery is the first therapeutic modality for GBM because it leads to bulk reduction of tumours and decompression of the brain, thereby alleviating the symptoms of intracranial  8 hypertension.  The extent of surgical resection of tumours is correlated with patient survival.  Chang et al. reported in the RTOG/ECOG studies that the 18-month survival of patients is longer in those who received total resection (34%) as opposed to partial resection (25%) or biopsy (15%) alone (Chang et al., 1983).  The result was subsequently confirmed by Simpson’s retrospective review and recent studies, suggesting a longer median survival for complete surgical excision (11.3 months) compared with biopsy (6.6 months) (Simpson et al., 1993).  Maximal resection is an independent variable and one of the most important factors associated with longer survival times in GBM patients (Allahdini et al., 2010; Salvati et al., 2012). The extent of surgical resection is determined by the age and performance status of patients as well as the site and size of tumours.  If a gross total surgical resection is not achievable, a maximal safe surgery attempting to debulk the tumour will be performed with the hope of improving quality of life of patients and extending their survival.  The fact that GBM tumours grow infiltratively renders complete resection virtually unachievable and disease progression almost inevitable.  Despite the intrinsic challenges in surgical removal of these tumours, novel techniques such as the use of fluorescent prodrug 5-aminolevulinic acid allows surgeons to visualize tumours more easily, leading to an improvement in macroscopic total resection of tumours (Friedman et al., 2009). Other surgical techniques are performed to assist diagnosis rather than treating the disease. For example, cytoreductive surgery permits the obtainment of a tissue sample that can be subject to histopathological examination for diagnostic purposes.  A stereotactic biopsy is done for histological confirmation of diagnosis when craniotomy is not possible (Brandes et al., 2008). 1.7.2   RADIATION Radiation therapy (RT) is a mainstay of therapy for GBM patients after surgery.  Clinical reports provide unequivocal support for the use of adjuvant radiation in the treatment of malignant gliomas.  These landmark studies resulted from two multi-institutional phase III randomized trials which compared conventional fractionated adjuvant RT to best supportive care after surgery.  Patients who received RT in combination with surgery showed a significant longer survival (9 months vs. 3.5 months and 10.5 months vs. 5.2 months from the two studies) than those who did not receive RT (Kristiansen et al., 1981; Walker et al., 1978).  The value of RT in the management of malignant glioma was therefore established and has become an integral part of brain cancer treatments.  Post-operative external-beam radiotherapy is a standard adjuvant treatment for GBM and is given within 6 weeks of surgery.  Highly  9 sophisticated and computer-assisted dosimetry delivers 60 Gy in 30 fractions to the target volume, specified as a 2-3cm ring of tissue surrounding the perimeter of the contrast enhanced lesion, over a total of 6 weeks (Brandes et al., 2008). Hyperfractionated RT, which divides radiation into small doses given more than once a day to achieve a higher total dose, does not significantly improve overall survival compared to standard external treatment (Laperriere et al., 1998) and remains investigational.  Hypofractionated RT is different from hyperfractionated RT in that the total dose of radiation is partitioned into large doses and treatments are provided less than once a day.  While hypofractionated, intensity-modulated radiotherapy with concurrent and adjuvant temozolomide (TMZ) does not offer additional survival benefit compared to current standard-of-care therapy in newly diagnosed GBM patients (Reddy et al., 2012), it prolongs the survival of elderly patients (>70 years of age) compared to those receiving standard RT in a recent phase II clinicial trial (Malmstrom et al., 2012). Stereotactic radiotherapy, which is also known as radiosurgery, involves the use of X-rays generated by cobalt sources (gamma knife) or a linear accelerator to transfer a large and highly focused dose to the tumour with a minimal dose distribution to neighbouring normal tissue (Brandes et al., 2008). The extreme accuracy of radiation delivery allows specific targeting of tumour cells while sparing the adjacent normal cells.  Stereotactic surgery is not a component of the standard treatment in newly diagnosed GBM but has been explored for its use in recurrent disease (Lederman et al., 2000; Minniti et al., 2013). 1.7.3 CHEMOTHERAPY Several randomized clinical trials have been conducted since the late 1970’s to evaluate the effects of adjuvant chemotherapy on the survival of brain tumour patients.  Most treatment protocols followed a nitrosourea-based regimen, including mostly N-(2-chloroethyl)-N'-cyclohexyl-N-nitrosourea (CCNU, or lomustine) and bis-chloroethylnitrosourea (BCNU, or carmustine).  The role of chemotherapy could not be clearly defined by the marginally significant results from early studies (Eyre et al., 1983; Shapiro et al., 1992; Walker et al., 1980), which may be due to the size and heterogeneity of the patient population or caveats in statistical analyses.  A meta-analysis performed by Fine et al. attempted to increase the sample size by incorporating the results from 16 randomized clinical trials involving more than 3000 patients, multiple chemotherapeutic agents and dosing schedules (Fine et al., 1993).  In this study, the combination of RT and adjuvant chemotherapy showed a relative increase of 23.4% in 1-year survival and 52.4% in 2-year survival compared to RT alone.  Yet, some prognostic factors were  10 not comparable in the two groups.  For example, the RT/chemotherapy combination treatment group was composed of a greater number of younger patients with better performance status.  Conclusive evidence for the efficacy of chemotherapy did not emerge until a phase III randomized trial published in 2005, which convincingly established the role of a DNA alkylating agent, TMZ, in the treatment of newly diagnosed malignant gliomas.  1.8 TEMOZOLOMIDE (TMZ) 1.8.1 BACKGROUND A report published by the European Organization for the Research and Treatment of Cancer and the National Cancer Institute of Canada Clinical Trials Group (EORTC/NCIC) in 2005 led to a revolutionary change in the therapy of newly diagnosed GBM.  This well-conducted, multi-institutional study by Stupp et al. involved 573 patients from 85 centres in 15 countries (Stupp et al., 2005).  Patients with newly diagnosed, histologically confirmed GBM were randomly assigned to receive RT alone or concurrent RT and TMZ, followed by 6 cycles of adjuvant TMZ, and the overall survival of patients was assessed.  The combination of RT and TMZ improved the median survival (14.6 months) of patients compared to RT only (12.1 months).  The benefit of TMZ is better appreciated when the 2-year survival rates are compared between the two treatment groups.  The RT plus TMZ combination doubled the survival rate (26.5%) compared to that of RT alone (10.4%).  Therefore, the addition of TMZ to RT offered survival benefit with tolerable toxicity in newly diagnosed GBM patients.  This finding provided convincing evidence that supports the use of chemotherapy, specifically TMZ, in the treatment of brain cancer.  TMZ has since then become part of the front-line treatment for newly diagnosed malignant gliomas. 1.8.2 PHARMACOLOGY  TMZ, which was developed in the 1980’s as part of a rational drug development initiative (Newlands et al., 1997), belongs to the imidazoletetrazine class of compounds.  Mitozolomide, the first compound of this class, showed excellent anti-tumour activity in pre-clinical models of astrocytomas, GBM and ependymomas (Stevens et al., 1987; Stevens et al., 1984).  However, early clinical trials were aborted due to the variable pharmacokinetics as well as severe side-effects, such as unpredictable and long-lasting hematological toxicities and even sudden death associated with the treatments.  Dacarbazine (DTIC), another compound of the imidazotetrazine class used in Hodgkin’s disease, is a prodrug that needs metabolic activation by the liver.  The use of DTIC as a therapeutic agent is limited by its inter-patient variability of pharmacokinetic parameters (Breithaupt et al., 1982).  TMZ was subsequently developed as an alternative drug  11 that is structurally related to DTIC with measurable in vitro activity and a more desirable toxicity profile (Moore et al., 1998). TMZ is a prodrug that undergoes spontaneous hydrolysis and is converted into an active metabolite 5-(3-methyltriazen-l-yl)imidazole-4-carboxamide (MTIC) at physiologic pH without the need of enzymatic demethylation in the liver (Stupp et al., 2001).  TMZ is stable in the acidic gastric environment and thus shows 100% oral bioavailability (Stevens et al., 1987).  The compound exhibits linear pharmacokinetics with a plasma half-life of 1.8hrs.  Maximum plasma concentrations are reached approximately 1hr after drug administration and range from 1.53-7.22g/mL.  The drug is excreted mainly through the kidney (Nicholson et al., 1998).  The fact that TMZ demonstrates good penetration in the brain tissues as well as anti-tumour activities in preclinical studies makes it particularly attractive for the treatment of brain tumours (Patel et al., 2003). 1.8.3 MECHANISM OF ACTION MTIC, the active metabolite generated from the TMZ prodrug, is responsible for DNA alkylation on the N7 of guanine, O3 of adenine and O6 of guanine on DNA.  The major site of alkylation occurs on the N7 of guanine (65-80%).  O6-methylguanine (O6-MG), despite being the minority of adducts formed (8%), is largely accountable for the cytotoxic effect of the drug (Hickman et al., 1985; Tentori and Graziani, 2002).  The adduct “tricks” the DNA synthesis apparatus into incorporating thymine instead of cytosine opposite to the O6-MG.  This incorrect DNA pair is subsequently recognized by the DNA mismatch repair pathway, which assumes repeated rounds of futile attempts to fix the error, leading to a state of chronic strand break production and activation of apoptotic pathways (Barone et al., 2006).  The cytotoxicity of TMZ therefore depends on a functional DNA mismatch repair system (D'Atri et al., 1998). 1.8.4 MECHANISM OF RESISTANCE O6- METHYLGUANINE DNA METHYLTRANSFERASE (MGMT)  TMZ was shown to be an effective therapy in the EORTC/NCIC study published by Stupp et al. in 2005.  However, similar to all the other drugs used in cancer treatments, TMZ does not elicit therapeutic responses in every GBM patient.  In a companion analysis, Hegi et al. stratified patients from the EORTC/NCIC trial based on the promoter methylation of O6-methylguanine DNA methyltransferase (MGMT) (Hegi et al., 2005). MGMT is a DNA repair protein which guards the cellular genome against the mutagenic effects of alkylating agents (Kaina et al., 2007).  MGMT removes alkyl adducts generated by TMZ on the O6-position of guanine and covalently transfers the alkyl group to its internal active  12 site represented by a cysteine residue (Marchesi et al., 2007; Pegg, 2000).  Therefore, MGMT counteracts the cytotoxic effects of TMZ and promotes TMZ resistance. In the study by Hegi et al., the MGMT promoter was methylated in 45% of 206 assessable cases.  Interestingly, the efficacy of TMZ was found to be tightly correlated with the methylation status of MGMT promoter.  Specifically, patients with tumour promoter hypermethylation (ie. MGMT gene silencing) experienced improvement in median survival (21.7 months vs. 15.3 months) and 2-year survival (48% vs.10%) compared to patients with hypomethylated tumour MGMT.  It was therefore concluded that GBM patients containing a methylated MGMT promoter benefit from TMZ while those who had a non-methylated MGMT promoter do not (Hegi et al., 2005).  This result confirmed the previous findings suggesting that MGMT expression or promoter methylation is correlated with response to chloroethylnitrosoureas and methylating agents (Belanich et al., 1996; Chen et al., 1999; Jaeckle et al., 1998).  Furthermore, other studies have shown that MGMT expression is an independent prognostic marker for GBM.  The Southwest Oncology Group reported that MGMT expression levels in newly diagnosed malignant astrocytoma were inversely correlated with tumour response and survival in patients treated with BCNU (Jaeckle et al., 1998).  In a retrospective study, Esteller et al. demonstrated that methylated MGMT promoter was associated with better clinical outcome in 47 glioma patients treated with whole-brain RT and alkylating agents (Esteller et al., 2000).  Furthermore, a recent study found a concordance between MGMT methylation status in tissue and in serum and indicated that serum MGMT promoter methylation was an independent factor for longer overall and progression-free survival (Balana et al., 2011).  Patients having secondary GBM tumours harbouring both IDH1 mutations and MGMT promoter methylation show better response to TMZ treatment compared to other patients (SongTao et al., 2012).  All together, these studies provide an important insight into the significance of MGMT in regulating disease progression and response to therapy.   Given the crucial role of MGMT in TMZ resistance, effort has been made to mitigate the function of this DNA repair protein.  For example, O6-benzylguanine (O6-BG) is a potent agent that inactivates MGMT stoichiometrically (Pegg, 1990).  O6-BG is a low-molecular weight inhibitor of MGMT that depletes the enzyme and increases the sensitivity of tumour cells to alkylating agents in vitro and in vivo (Dolan et al., 1998; Friedman, 2000; Rabik et al., 2006).  However, the adverse effects of the drug on the bone marrow limit its clinical use.  A novel and potent MGMT inhibitor Lomeguatrib was recently developed with more favourable toxicity profile.  The result from a phase I study of 38 patients indicated that an MGMT-depleting dose of  13 lomeguatrib was achievable in combination with standard TMZ dosing without provoking intolerable bone marrow toxicities (Ranson et al., 2006). ALTERNATIVE MODES FOR TMZ RESISTANCE  The mismatch repair (MMR) system is a DNA repair pathway responsible for correcting replication errors such as base-base mismatches and insertion/deletion loops (Jiricny, 2006).  Interestingly, the cytotoxic effects of TMZ have been shown to depend on functional MMR; triazene compounds induce apoptosis in MMR-proficient but not in MMR-deficient tumour cells (Meyers et al., 2004; O'Brien and Brown, 2006; Stojic et al., 2004).  This observation may be explained by the “futile cycling model” (Bignami et al., 2000) described as follows.  The MMR system attempts to process the O6-meG:T and O6-meG:C by removing the mismatched base in the newly synthesized strand.  However, the O6-methylguanine is not removed and remains in the template strand that continues generating errors in the next round of DNA replication.  These reiterated futile attempts at repair subsequently result in DNA strand breaks, chronic checkpoint activation and eventually apoptosis, mitotic catastrophe or senescence (D'Atri et al., 1998; Kaina et al., 1997). Therefore, TMZ cytotoxicity requires a functional MMR system. Another mechanism of TMZ resistance involves the base excision repair (BER) pathway, which protects cells against lesions from cellular metabolism (eg. methylation or oxidation of DNA bases) and against base modification from physical or chemical agents.  Under most circumstances, tumour cells have a fully functional BER system, allowing the cells to repair the N3-methyladenine (N3-meA), a major adduct formed by TMZ treatment.  N3-meA is not particularly toxic to tumour cells unless BER is disabled (Horton and Wilson, 2007).  Therefore, cancer cells could be rendered sensitive to TMZ by inactivation of BER.  In light of this, inhibitors to PARP, a component of the BER system, have been tested.  The PARP inhibitors have been shown to augment the anti-tumour effect of TMZ in glioma, melanoma, colorectal and breast cancers, lymphoma and leukemia (Tentori et al., 1997; Tentori et al., 1999).   In addition to the components of the BER system, novel mediators of TMZ resistance, such as miRNAs (Ujifuku et al., 2010), aldehyde dehydrogenase 1A1 (ALDH1A1) (Schafer et al., 2012), FOXM1 (Zhang et al., 2012) and the DNA repair protein ALKBH2 (Johannessen et al., 2012) have been recently identified and may serve as molecular targets for the treatment of recurrent GBM. Our laboratory demonstrated that TMZ resistance may be modulated in part by the expression of Y-box binding protein-1 (YB-1), which is an oncogenic protein that drives the transcription of genes associated with stem cell properties, cell growth and drug resistance  14 (Finkbeiner et al., 2009; Fotovati et al., 2011; Stratford et al., 2007; To et al., 2010).  YB-1 level is elevated in GBM tumours compared to normal brain tissue.  Knock-down of YB-1 sensitized pediatric GBM cells to TMZ (Gao et al., 2009).  Studies are underway to understand the interplay between YB-1 and other cell cycle proteins in regulating the response to TMZ. 1.9 TMZ TREATMENT IN CHILDREN  A study recently published by the Children’s Oncology Group evaluated the efficacy of TMZ in children with high-grade astrocytomas and examined the relationship between MGMT expression and patient survival (Cohen et al., 2011).  A total of 107 patients with anaplastic astrocytoma (AA), GBM or gliosarcoma were enrolled and underwent concurrent chemoradiotherapy with TMZ, followed by adjuvant chemotherapy with TMZ. The 3-year overall survival rate was 22 ± 5%.  TMZ given during RT and as adjuvant therapy did not improve event-free survival in both AA and GBM compared to the results from a previous study (CCG-945) that did not include TMZ in the treatment regimen.  In conclusion, TMZ failed to improve outcome in children with high-grade astrocytomas and over-expression of MGMT is adversely associated with survival (Cohen et al., 2011).   After years of investigation, results from several independent groups led to the same conclusion, indicating the ineffectiveness of TMZ in the pediatric populations.  Two phase II studies revealed only limited overall objective response to TMZ in children and adolescents with recurrent CNS tumours (Nicholson et al., 2007; Ruggiero et al., 2006).  A multi-institutional study involving 31 pediatric patients with newly diagnosed high-grade gliomas also did not find any significant benefit in the addition of TMZ to RT (Broniscer et al., 2006).  In line with this, the Children’s Oncology Group phase II study in newly diagnosed high-grade glioma patients showed a disappointingly high rate of treatment failures and low 3-year event-free survival (7-13%) in children treated with TMZ (Cohen et al., 2011).  Consistent with these studies across the globe, a recent survey by the Canadian Pediatric Brain Tumour Consortium (CPBTC) found the same trend in 16 participating pediatric oncology centres of the CPBTC.  One hundred and thirty seven children with brain tumours were treated with TMZ between January 2000 and March 2006.  The results suggested that although some children treated with TMZ for recurrent CNS tumours achieved a response, the overall survival rate (23%) was dismal considering 59% of the surviving patients had low-grade tumours which are typically associated with a better prognosis (Bartels et al., 2011).    15 1.10 NOVEL THERAPIES FOR GBM TREATMENT 1.10.1 MOLECULAR TARGETED THERAPY The identification of genetic aberrations and deregulated signaling pathways in GBM has spurred the evaluation of molecular targeted therapies (Van Meir et al., 2010).  A successful example of this is the approval of bevacizumab, a humanized monoclonal antibody against vascular endothelial growth factor (VEGF), for the treatment of recurrent GBM.  A number of clinical studies demonstrated a superior response rate and increased 6-month progression-free survival of bevacizumab treatment compared to temozolomide (Guiu et al., 2008; Nghiemphu et al., 2009; Pope et al., 2006; Zuniga et al., 2009).  The addition of bevacizumab to standard RT and chemotherapy has also shown moderate efficacy in patients with progressive high-grade gliomas (Friedman et al., 2009).  Despite the possible prolongation in PFS, disease progression is almost inevitable and in some instances, tumours become more aggressive and untreatable after bevacizumab treatment (Iwamoto et al., 2009; Narayana et al., 2009).  EGFR was believed to be an attractive molecular target in GBM (Van Meir et al., 2010) due to its elevated expression in approximately 40% of primary GBMs, especially in the “classical” subtype.  Furthermore, increased EGFR signaling stimulates cell proliferation, invasion, angiogenesis and inhibits apoptosis.  However, EGFR small molecule inhibitors gefitinib and erlotinib, albeit being well tolerated in patients, fail to provide survival benefits (Van Meir et al., 2010).  In addition, neither the EGFR/HER-2 inhibitor (lapatinib) (Thiessen et al., 2010) nor the monoclonal antibody against EGFR (cetuximab) was shown to be efficacious (Neyns et al., 2009).  A recent study by Vivanco et al. suggested that EGFR inhibitors (eg. erlotinib) targeting the active kinase conformation often found in lung cancers showed a lack of efficacy in GBM, which typically harbour mutations in the extracellular domain of EGFR (Vivanco et al., 2012).  The difference in therapeutic response to EGFR inhibitors may therefore be attributed to the distinct conformations of mutant EGFR seen in these two types of cancers.  Attempts have also been made to target other receptor tyrosine kinases (RTKs) such as PDGFR, which is commonly over-expressed in the “proneural” subtype of GBM.  The PDGFR inhibitor imatinib mesylate demonstrated remarkable anti-tumour activities in an orthotopic glioma model (Kilic et al., 2000) but showed rare responses and no increase in PFS in clinical trials (Wen et al., 2006).  The combination of erlotinib with temozolomide (TMZ) and RT was assessed in a recent phase II trial in patients with newly diagnosed GBM and was shown to be inefficacious with an unacceptable toxicity (Peereboom et al., 2010).  An alternative therapeutic strategy is to target the intracellular signaling pathways.  The PI3K/AKT/mTOR pathway is often deregulated in GBM due to over-expression of RTKs such as  16 EGFR, PDGFR and MET, as well as loss of PTEN tumour suppressor function (Van Meir et al., 2010).  mTOR inhibitors such as rapamycin, CCI-779 and RAD001 have been evaluated in clinical trials and demonstrated minimal activity against malignant gliomas as single agents (Chang et al., 2005; Galanis et al., 2005).  A recent phase I clinical trial suggested that RAD001 in combination with RT/TMZ and adjuvant TMZ was well tolerated (Sarkaria et al., 2011), warranting further investigation in future studies.   Another signaling pathway activated downstream of EGFR and PDGFR is the RAS/RAF/MAPK pathway.  The localization of RAS to the intracellular surface of the cell membrane is a prerequisite for its activation, a process that can be interrupted by the farnesyl transferase inhibitors (FTIs).  Although the FTIs have shown promising activity in glioma models, the efficacy on patients with recurrent malignant gliomas was uncertain in a phase II trial (Cloughesy et al., 2006).  A RAS/MAPK signaling inhibitor, TLN-4601, has also recently been shown ineffective as a monotherapy in patients with GBM at first progression (Mason et al., 2012). 1.10.2 DISULFIRAM (DSF) Disulfiram is a small molecule inhibitor that has been used as a front-line treatment for alcoholism for more than 60 years (Johansson, 1992).  The drug has been increasingly recognized for its potential role in cancer treatment due to recent studies that demonstrated its strong anti-neoplastic activity in vitro and in vivo (Chen et al., 2006; Conticello et al., 2012; Iljin et al., 2009).  We have recently published a study showing the growth inhibitory and cytotoxic effect of disulfiram (DSF) on TMZ-resistant GBM cell lines, tumour initiating cells (TICs) and primary cells isolated from patient.  DSF inhibited the proliferation and/or self-renewal of GBM cells at an IC90 of 100nM, a concentration at which had no effects on normal human astrocytes.  In addition, PLK1 expression was suppressed upon DSF treatment (Triscott et al., 2012).  Coincident with our manuscript, another report by Liu et al. also demonstrated the efficacy of disulfiram/copper (DSF/Cu) on GBM cells in vitro (Liu et al., 2012).  Liu et al. suggested that DSF was cytotoxic to GBM cells in a Cu-dependent manner and that the combination of DSF/Cu and gemcitabine produced a synergistic effect in eliminating the cancer cells.  The authors further examined the impact of DSF/Cu at the molecular level and showed that the drug treatment induced the production of reactive oxygen species (ROS), activated the JNK and p38 pathways and inhibited NFκB activity in GBM cell lines.  The intrinsic apoptotic pathway may be triggered through the regulation of the BCL2 family of proteins.  Finally, DSF/Cu also targeted the stem-like GBM cells (Liu et al., 2012). Consistent with these results, Yip et al. reported that  17 DSF/Cu is highly toxic to breast cancer cells with an IC50 of 200-500nM and enhanced the effect of paclitaxel.  Furthermore, the drug inhibited the mammosphere formation of the “cancer stem-like” cell populations (ALDH1+/CD44high/CD24low) (Yip et al., 2011a).  Together, these results suggest that DSF effectively eliminates cancer cells with the potential to target the cancer stem cells. The anti-cancer properties of DSF are not yet fully understood and several mechanisms have been proposed.  Recent experimental evidence indicates that the cytotoxicity of DSF may depend on the presence of Cu (Cen et al., 2004b; Chen et al., 2006), which induces the generation of ROS through redox reactions in human cells (Barceloux, 1999).  However, the restricted transport of Cu into tumour cells due to regulation by the trans-membrane Cu transporter Ctr1 has limited the clinical efficacy of copper in cancer treatment.  This problem could be circumvented by the addition of DSF, which is a bivalent metal ion chelator that could complex with Cu and facilitate its transport across the cell membrane.  Indeed, DSF/Cu is a more potent ROS inducer compared to Cu (Burkitt et al., 1998).  A separate study suggested that DSF is a proteosome inhibitor.  Under physiological condition, DSF is rapidly converted to a reduced metabolite, diethyldithiocarbamate, which forms bis(diethyldithiocarbamate)-copper(II) complexes in the presence of copper ions.  This copper (II) complex may be accountable for the selective anti-tumour activity of DSF by inhibiting the proteosome (Skrott and Cvek, 2012).  In another study, DSF acts as a DNMT1 inhibitor that reduces global (5me)C content, thereby re-activating epigenetically silenced genes and inhibiting growth of prostate cancer cells (Lin et al., 2011a).   DSF is also an ALDH inhibitor (Lipsky et al., 2001).  However, the suppression of this protein is unlikely the primary mechanism for cytotoxicity in our model.  According to our study, the dose required to suppress GBM cell growth is much higher than that needed to inhibit ALDH activities.  In addition, blocking ALDH function with a pan inhibitor (DEAB) or targeting specific isoforms (ALDH1A1 and ALDH1A3) did not significantly affect GBM growth or survival, suggesting that ALDH inhibition is not the key determinant of cytotoxicity of DSF [(Triscott et al., 2012).  Further studies will be required to determine the mechanism of action of this drug on brain cancer cells. 1.11 MEDULLOBLASTOMA (MB) 1.11.1 EPIDEMIOLOGY, GENETIC PREDISPOSITION, SYMPTOMS AND PROGNOSIS As a group, brain tumours are the second most commonly diagnosed cancer and the leading cause of solid-tumour death in children.  Every year, approximately 8.5 per 100,000 new  18 pediatric cases of brain cancers are reported, with the majority of them (20-30%) being medulloblastoma (MB) (Kieran et al., 2010).  MB is a malignant CNS tumour, which belongs to the primitive neuroectodermal (PNETs) class of brain tumours.   The PNET family of tumours, which include medulloblastoma, supratentorial PNETs and atypical teratoid/rhabdoid tumours (ATRTs), are identified as a group of undifferentiated embryonal tumours of neuroepithelial origin (Samkari et al., 2012).  These tumours were originally grouped together as PNETs based on their histologic appearance, which is reminiscent of cells found largely during the embryonic development.  The tumour cells are characterized as densely packed, small round cells with prominent nuclei and scant cytoplasm (Kleihues et al., 2002).  As implied by their “embryonal” designation, PNETs such as MB, frequently occur in early childhood.  Indeed, MB is the most common malignant brain tumour in children, accounting for approximately 7-8% of all intracranial tumours and 18% of all pediatric brain cancers (Kaatsch et al., 2001), with a male to female ratio of ~1.5:1 (McNeil et al., 2002).  Patients are mostly diagnosed at the age of 6-7 years old (Moschovi et al., 1998; Sussman et al., 1990) with a median survival of 10 years (Lefkowitz et al., 1988; McNeil et al., 2002; Packer et al., 1999).  Approximately 10-15% of MBs are diagnosed in infancy (Crawford et al., 2007).   There are no known environmental causes for MB.  A small proportion (1-2%) of MB patients have Gorlin syndrome, which predisposes individuals to developing nevoid basal cell carcinomas diagnosed by characteristic dermatological and skeletal features (Rood et al., 2004).  MB is seen in up to 40% of Turcot syndrome patients with familial adenomatous polyposis due to genetic mutations on chromosome 5q21 (Hamilton et al., 1995).   MB occurs in the posterior fossa.  In most cases, the tumours arise in the cerebellar vermis and project into the fourth ventricle.  Larger tumours that involve the cerebellar hemispheres are found in older children and young adults.  Leptomeningeal dissemination of tumour cells along the neuroaxis can occur as disease progresses.  Although extraneural metastasis of MB could occur in the bone and throughout the CNS, systemic metastasis is infrequent (Albright et al., 1996; Berger et al., 1991). The clinical presentation of MB is age-dependent.  In younger children and infants, behavioural changes, lethargy, vomiting or increasing head circumference may be detected.  Head tilt, truncal ataxia, gait disturbances and lethargy are seen in older children.  Adult MB frequently arises in the cerebellar hemisphere, causing unilateral syndromes of cerebellar dysfunction such as dysmetria and sometimes hearing loss.   The clinical staging of MB guides clinicians when deciding treatment options after surgery (Albright et al., 1996).  Patients can be stratified into the standard-risk and high-risk  19 groups on the basis of age, extent of surgical resection and tumour spread (Albright et al., 1996; Laurent et al., 1985).  Specifically, patients older than 3 years of age, with a residual tumour <1.5cm3 post-surgery and no sign of disease dissemination are considered “standard-risk”, while those having the reverse clinical presentations are regarded as “high-risk” patients that have increased chance of relapse and should receive more aggressive treatments.  The 5-year survivorship is >70% and 16-70% in standard-risk and high-risk patients, respectively (Sirachainan et al., 2011; Taillandier et al., 2011; Packer et al., 2006).  A number of novel risk stratification or prognostic markers have recently been proposed.  For example, histone deacetylase 5 and histone deacetylase 9 (HDAC5 and HDAC 9) are significantly up-regulated in the high-risk compared to the low-risk MBs where high expression is associated with poor survival (Milde et al., 2010).  Furthermore, in the study by Remke et al., follistatin-like 5 (FSTL5) negativity delineates a large group of “good-prognosis” patients, who would typically be classified in the “non-WNT/SHH” group that is associated with intermediate or high-risk (Remke et al., 2011).  FSTL5 may therefore serve as an important marker that aids accurate prognostication of the non-WNT/non-SHH tumours. 1.11.2 PATHOLOGY Classic MB tumours are characterized by sheets of small, round, blue and densely packed cells when stained with hematoxylin.  MB tumours are positive for vimentin and synaptophysin staining (Burger et al., 1987; Coffin et al., 1990).  These tumours also feature oval shaped cells that may display significant pleiomorphism.  The undifferentiated cells show scant cytoplasm and loosely packed chromatin.  Mitosis is seen in up to 80% of tumours, as demonstrated by positivity in Ki67/MIB1 staining (Crawford et al., 2007); cytological indication of ongoing apoptosis is also common in tumour cells.  These pathological characteristics define the most commonly found subtype comprising 73% of all pediatric MBs (Ellison, 2010).  Differentiation may be observed along astrocytic, neuronal or mesenchymal lineage.  Neuronal differentiation is characterized by Homer-Wright rosettes, or less commonly pseudorosettes, which are formed by elongated cells surrounding acellular areas. The histologic variants of MB include desmoplastic/nodular (DN) MB, MB with extensive nodularity (MBEN), large-cell/anaplastic (LCA) MB, medullomyoblastoma and melanotic MB (Kleihues et al., 2002).  The desmoplastic MB is composed of densely packed, highly proliferative, mitotically active, reticulin-rich areas that surround reticulin-free nodules (Katsetos et al., 1989).  These tumours often have contrasting areas of high and low cellularity, forming a  20 marble-like histological pattern.  Cells within the nodules frequently show cytologic signs of neuronal differentiation while the surrounding cells are anaplastic and mitotically active.      The large-cell MB is composed of lobular sheets of round cells with abundant cytoplasm, prominent nucleoli and pleomorphic nuclei (Giangaspero et al., 1992).  The large-cell MBs are rarely pure and are usually infused with an anaplastic phenotype.  Therefore, it is sometimes referred to as the LCA MB.  Anaplastic and large cell MB can be distinguished histologically.  Anaplastic MB shows cytoplasmic and nuclear pleomorphism, whereas large-cell MB has round monomorphic cells with a single nucleolus.  This subclass is the most aggressive variant with a propensity to metastasize and accounts for 4% of the cases (Crawford et al., 2007).  In addition, large necrotic areas with high mitotic rates and apoptosis are commonly seen.   Medullomyoblastoma shows areas of focal myogenic differentiation.  Melanotic MB is characterized by undifferentiated cells containing melanin.   All MB variants are aggressive and malignant WHO grade IV brain tumours; however, substantial differences in clinical outcome and biological behaviour are associated with each pathological feature.  Infants with the MBEN and DN subclasses show similar clinical outcome and are categorized as low-risk (de Haas et al., 2008) whereas patients with the LCA tumours have the worst prognosis.  1.11.3 CELL OF ORIGIN Cerebellar development initiates during embryonic growth in several regions of the brain, including the upper rhombic lip and the ventricular zone (VZ), and reaches maturation months after birth (Wang and Zoghbi, 2001).  These specific regions contain cells that are highly proliferative, therefore they are particularly vulnerable to neoplastic transformation when oncogenic mutations are introduced.   The VZ consists of stem and progenitor cells that line the fourth ventricle.  These cells generate Purkinje neurons, interneurons and glial cells.  During development of the CNS, granule neuron progenitors (GNPs) divide in the external granule layer on the surface of the cerebellum and subsequently exit the cell cycle and migrate inward to the internal granule layer where they turn into post-mitotic granule neurons (Sotelo, 2004).  The GNPs are stimulated to divide in response to the SHH ligands produced by Purkinje neurons (Dahmane and Ruiz i Altaba, 1999; Wallace, 1999; Wechsler-Reya and Scott, 1999).  Therefore, enhanced SHH pathway activation due to over-expression of the ligands or mutational activation of the pathway components may accelerate the expansion of GNPs, leading to tumour formation.  This speculation was supported by convincing lines of evidence from two independent studies.              21 Schüller et al. demonstrated that cerebellar GNPs derive from multi-lineage embryonic CNS progenitors and that SHH pathway activation generated MBs, but not other brain tumours, in early and late-stage CNS progenitors. The acquisition of “granule cell lineage identity” was proposed to be a key factor in the development of SHH pathway-driven MBs (Schuller et al., 2008).  In a separate study by Yang et al., the authors showed that SHH pathway activation in neuronal progenitors led to MB formation by 3 months of age in animals.  The tumours initiated in stem cells, however, have a much shorter latency and develop more rapidly and aggressively with all animals succumbing to the disease in 3-4 weeks.  Intriguingly, in the stem cell populations, SHH pathway activation promotes proliferation and only causes tumour formation after neuronal lineage commitment (Yang et al., 2008).  Together these results suggest the importance of neuronal lineage identity in MBs that develop from the cerebellar stem and progenitor cells. While the SHH-driven MB is associated with abnormal expansion of progenitor cells in the external granular layer of the cerebellum, experimental evidence suggests that the WNT-driven tumours may derive from a completely different anatomical area of the brain.  A recent study by Gibson et al. showed that the genes marking the human WNT subtype MB were found to be more frequently expressed in the lower rhombic lip and embryonic dorsal brainstem.  The authors subsequently demonstrated that the WNT MB may derive from cells of the dorsal brain stem with activating mutations in CTNNB1.  While these mutations in CTNNB1 have minimal impact on cerebellar GNPs, they could cause aberrant accumulation of proliferating Zic1(+) precursor cells in the embryonic dorsal brainstem.  Furthermore, concurrent mutations of CTNNB1 and Tp53 drive the formation of MBs that are reminiscent of the human WNT-subtype MB (Gibson et al., 2010).  1.11.4 MOLECULAR ABERRATIONS SONIC HEDGEHOG (SHH) PATHWAY There is mounting evidence suggesting the involvement of the sonic hedgehog (SHH) pathway in the pathogenesis of MB (Entz-Werle et al., 2008).  When SHH, the ligand, binds to the receptor Patched (PTCH1), the inhibition of the downstream effector Smoothened (SMO) is relieved.  Subsequently, the GLI family transcription factors are released from the inhibitory protein complex including Suppressor of Fused (SUFU) and translocate into the nucleus to up-regulate the expression of target genes involved in cell cycle progression, adhesion, signal transduction and apoptosis (Yoon et al., 2002) (Figure 1.1).  Interestingly, the GLI transcription factor has recently been shown to directly bind to and up-regulate the transcription of BMI, a key  22 stem cell regulator gene (Wang et al., 2012).   The mitogenic effects of the SHH pathway is recognized by its role in driving the proliferation of GNPs in the cerebellum.  Deregulation of SHH pathway may therefore unleash the mitogenic power that enables GNPs to proliferate uncontrollably, resulting in tumour formation.   In the clinical setting, inactivating mutations of PTCH1 and SUFU and activating mutations of SMO are found in ~15% of sporadic MB (Pietsch et al., 1997; Raffel et al., 1997; Reifenberger et al., 1998; Taylor et al., 2002).  Furthermore, individuals with familial inheritance of mutations in PTCH1, the cause of Gorlin syndrome, are susceptible to developing various tumour types, including MB (Hahn et al., 1996). A number of animal models have been established to emulate MB tumourigenesis by disrupting different components of the SHH pathway.  For example, approximately 15% of mice heterozygous for Ptch mutations develop CNS tumours that are histologically indistinguishable from human MB in the posterior fossa by 10-months of age (Goodrich et al., 1997; Zurawel et al., 2000).  The Smo/Smo mice, developed by Olson’s group, carry constitutive activation of SMO and show high incidence and early onset of tumour formation with leptomeningeal spread.  Furthermore, Smo/Smo MBs could be serially transplanted in vivo, demonstrating the aggressive nature of these tumour cells (Hatton et al., 2008).  The early onset and high tumour incidence of Smo/Smo mice makes it an ideal model for preclinical investigations.  Together, these results suggest that deregulation of the SHH pathway is sufficient to drive tumourigenesis in vivo.  23  Figure 1.1 The SHH pathway. When the ligand SHH binds to the receptor Patched (PTCH1), the inhibition of the downstream effector Smoothened (SMO) is alleviated.  Subsequently, the GLI family transcription factors are released from the inhibitory protein complex including Suppressor of Fused (SUFU) and enter the nucleus to drive the expression of target genes. STK36 and SUFU are involved in the regulation of GLI. WNT PATHWAY Another signaling cascade that is implicated in the development of MB is the WNT pathway.  The canonical WNT signaling pathway plays a significant role in modulating NSC proliferation and defining the midbrain-hindbrain boundary from which the cerebellum develops (Thomas and Capecchi, 1990) during embryogenesis. The importance of the WNT pathway in MB pathogenesis was recognized when researchers found an association between Turcot syndrome and susceptibility to brain tumours.  Individuals with this familial syndrome carry germline mutations in adenomatous polyposis coli (APC, a component of the WNT pathway) and have an increased risk of developing certain MBs or other neuroepithelial tumours (Huse and Holland, 2010).  Mutations in the WNT pathway, specifically in APC, AXIN1/2 and CTNNB1 (encodes β-catenin), are found in approximately 15-20% of all MBs (Ellison, 2010). Positive immunostaining in β-catenin is seen in ~18% of MB patients (Eberhart et al., 2000).   HH PTCH SMO GLI GLI SUFU STK36 Crosstalk with !-catenin in the WNT pathway !-Cat Mutation of the gene is associated with the development of MB Adapted from Zhou et al. Nature Reviews Drug Discovery (2009)    24 The canonical WNT pathway is initiated by the binding of WNT ligands to the family of cell surface Frizzled (FZD) receptors, which leads to the activation of dishevelled (DSH) and release of β-catenin from the inhibitory protein complex composed of APC, AXIN proteins and glycogen synthase kinase 3 (GSK3).  In the absence of WNT stimulation, CK1α and GSK3 phosphorylate β-catenin leading to its ubiquitination and degradation.  The inhibition of the APC/AXIN/GSK3 complex therefore increases the intracellular concentration of β-catenin, which subsequently translocates into the nucleus to act as a transcriptional co-activator by interacting with a family of transcription factors T-cell factors/lymphoid enhancer factors (TCF/LEFs).  This results in the transcription of target genes such as MYC, cyclin D (CCND1) and RE1-silencing transcription factor (REST), which are involved in cell proliferation, differentiation and evasion of apoptosis (Eberhart et al., 2000; Rossi et al., 2008) (Figure 1.2).  Studies suggest that crosstalk may occur between the WNT and SHH pathways.  For example, GSK3, an inhibitory molecule in the WNT pathway, may interact and phosphorylate GLI in the SHH pathway, leading to β-TrCT-mediated degradation of GLI proteins (Pan et al., 2006).  In addition, SUFU, the inhibitor in the SHH pathway, can mediate the nuclear export of β-catenin, resulting in the degradation of β-catenin in the cytoplasm (Marino, 2005).    DKK WNT SFRP FZ LRP5-6 APC !-Cat !-Cat Axin GSK! CK1" TCF- LEF Adapted from Zhou et al. Nature Reviews Drug Discovery (2009)   DSH Crosstalk with GLI in the SHH pathway GLI Mutation of the gene is associated with the development of MB  25 Figure 1.2 The WNT pathway. The WNT ligands bind to the family of cell surface Frizzled (FZD) receptors and LRP5-6, resulting in the activation of dishevelled (DSH) and release of β-catenin from the inhibitory protein complex consisting of APC, AXIN proteins and glycogen synthase kinase 3 (GSK3).  In the absence of WNT stimulation, CK1α and GSK3 phosphorylate β-catenin leading to its ubiquitination and degradation.  The inhibition of the APC/AXIN/GSK3 complex therefore causes the accumulation of β-catenin, which subsequently translocates into the nucleus to interact with transcription factors T-cell factors/lymphoid enhancer factors (TCF/LEFs) to drive the expression of target genes.  SFRP and DKK are endogenous secreted antagonists of the WNT pathway. ADDITIONAL MOLECULAR ABERRATIONS IDENTIFIED MYC and MYCN amplification are frequently identified in a subset of MB with large cell/anaplastic histological features (Aldosari et al., 2002; Tomlinson et al., 1994).  Indeed, the over-expression of Myc and MycN expedites MB tumour formation in mice, a characteristic shared in common with other known oncoproteins such as AKT, BCL2 and insulin-like growth factor 2 (IGF2) (Browd et al., 2006; McCall et al., 2007; Rao et al., 2003; Rao et al., 2004; Yang et al., 2008).   Two articles were published back-to-back in “Cell ” in February 2012 reporting the animal models of MYC-driven MB.  The study by Pei et al. demonstrated that cerebellar stem cells expressing Myc and mutant Trp53 (p53) generate aggressive tumours following orthotopic transplantation (Pei et al., 2012).  The tumours are composed of large, pleiomorphic cells and resemble human MYC-driven MB at the molecular level.  Interestingly, antagonists of PI3K/mTOR signaling but not SHH signaling suppressed the growth of these tumour cells.  This result was corroborated by the study of Kawauchi et al. who created a mouse model of MYC-subgroup of MB by transducing Trp53-null cerebellar progenitor cells with Myc (Kawauchi et al., 2012).  These transgenic mice bore tumours with molecular characteristics (eg. NPR3 expression) and gene signatures seen in the human MYC-driven MB.  Of note, the Myc-engineered MBs contained a large number of tumour-propagating cells with high proliferative potential.  These tumour cells are resistant to SHH signaling inhibitors, suggesting that the MYC-driven MB is distinct from the SHH-driven MB.  These animal models present valuable tools for studying a specific subgroup of aggressive MB. The ERBB family of RTKs, IGF1R, and PDGFR are also implicated in MB pathogenesis and are associated with poor prognosis (MacDonald et al., 2001; Rossi et al., 2008).  Over-expression of ERBB2, seen in 28% of MB, may promote tumourigenesis by activating the MAPK and AKT pathways and by increasing the expression of pro-metastatic genes such as MEK5, S100 calcium binding protein A4 (S100A4) and CCL5 (Calabrese et al., 2003; Hernan et al.,  26 2003).  ERBB2 protein over-expression is associated with poor clinical outcome (Gajjar et al., 2004; Gilbertson et al., 1995; Gilbertson et al., 1997). Another genetic lesion frequently seen (30-50% of cases) in MB is isochromosome 17q:i(17)(q10), which involves the loss of chromosome 17p and gain of 17q (Biegel et al., 1997; Bigner et al., 1988; Nicholson et al., 2000).  The commonly deleted region of 17q13.2-13.3 encompasses the loci of several tumour suppressor genes including TP53, the loss of which accelerates malignant transformation.  Indeed, Li-Fraumeni patients with germline mutations in TP53 are more prone to develop MB (Malkin et al., 1990).  Despite being a rare genetic alteration in sporadic MB, TP53 mutations seem to be associated with poor clinical outcome (Cogen and McDonald, 1996; Tomlinson et al., 1994).  1.11.5 MOLECULAR PROFILING AND CLASSIFICATIONS  The advent of microarray technology in the 1990’s enabled researchers to profile a large numbers of primary tumour samples by comprehensive oligonucleotide expression array.  Pomeroy et al. made the first attempt to subclassify MB by subjecting the gene expression profiles of 99 primary MBs to hierarchical clustering analyses.  The study revealed that classic MB is molecularly distinct from desmoplastic MB, with the latter carrying genetic signatures associated with SHH receptor PTCH and other components of the SHH pathway (Pomeroy et al., 2002).  Sporadic desmoplastic MBs are therefore similar to familial desmoplastic MBs (due to Gorlin syndrome) in that they both carry genetic aberrations in the SHH pathway.  Significantly, this study also showed that MB, ATRT (characterized by mutations in SMARCB1/INI1/hSNF5), and supratentorial PNETs, although being historically categorized under the “PNET family” due to histopathological resemblance, are in fact molecularly distinct disease entities. Advances in the field of array technology in recent years further facilitated the integrative genomic approaches in identifying molecular subgroups of MB.  The initial study by Thompson et al. demonstrated five distinct groups of MBs based on unsupervised clustering of the gene expression profiles of 46 tumours.  The two most well-characterized subgroups are the WNT and SHH MBs. The WNT tumours often exhibit monosomy 6 and frequent genetic aberrations in the WNT pathways, such as CTNNB1 mutation.  The SHH tumours, on the other hand, show an up-regulation of the SHH pathway (Thompson et al., 2006).  This result was subsequently confirmed by three independent groups: Northcott et al., Cho et al., and Kool et al., who also identified the WNT and SHH subgroups as well as the non-WNT/SHH subgroup in each of their own patient cohorts (Cho et al., 2011; Kool et al., 2008; Northcott et al., 2011b).  These three  27 studies were also consistent with previous findings suggesting that the WNT subgroup of tumours have a favourable prognosis (Clifford et al., 2006; Pfister et al., 2009).  Furthermore, in multiple cohorts, molecular classification of the disease seems to be more specific and accurate at predicting patient outcome than histology and Chang staging (Ellison et al., 2005; Northcott et al., 2011b; Pfaff et al., 2010).  The presence of metastases is no longer a powerful predictor of outcome when the molecular subgroup of the disease is taken into consideration.  For instance, patients with WNT MBs have a >90% overall survival regardless of metastatic status (M0 or M+).  Conversely, patients with group 3 MB do poorly, with a 5-year survival of <30%, regardless of metastatic disease.  Therefore, these molecular profiling studies have led to the development of new prediction models that may be more reliable than existing clinical models for prognostication (Ramaswamy et al., 2011).  The current consensus in the field is that there are four molecular subgroups of MB, each of which carries distinct gene signatures and is associated with different biological behaviours and clinical outcome (Cho et al., 2011; Kool et al., 2008; Northcott et al., 2011b; Thompson et al., 2006). As mentioned previously, the WNT MBs are characterized by genetic mutations in the WNT pathway.  Immunohistochemical staining of nuclear β-catenin and DKK1, mutation of CTNNB1 and detection of monosomy 6 are examples of classic markers of WNT tumours, which are correlated with very good long-term prognosis (>90% survival rate) compared to other subgroups.  Patients who succumb to the disease often die from complications of therapies or secondary neoplasms rather than recurrent WNT tumours (Ellison et al., 2011b).  The gender ratio of WNT MBs is ~1:1 male: female.  The tumours typically show a classic MB histology; they are rarely metastatic and can occur at all ages but very rarely in infants (Taylor et al., 2012). The SHH MBs have been largely identified on the basis of transcriptional profiling (Cho et al., 2011; Kool et al., 2008; Northcott et al., 2009; Northcott et al., 2011b; Rutkowski et al., 2010b; Zhou et al., 2010) or immunohistochemical staining of SFRP1 (Al-Halabi et al., 2011; Northcott et al., 2011b; Zhou et al., 2010) and GAB1 (Ellison et al., 2011a).  Patients with this subgroup of MBs often harbour somatic mutations in PTCH, SMO and SUFU or amplifications of GLI1 and GLI2, which are components of the SHH pathway (McManamy et al., 2007; Northcott et al., 2011a; Wetmore et al., 2001).  In line with this, individuals with germline mutations of SUFU are predisposed to infantile MB (Brugieres et al., 2010; Packer, 2011; Taylor et al., 2002; Wetmore et al., 2001), which often belongs to the SHH subgroup.  This molecular subgroup of MB curiously shows a bimodal age distribution; it is most frequent in both infants (0-3 years) and adults (>16 years) but is rare in children (3-16 years).  The gender ratio is about  28 1:1.  The majority of nodular/desmoplastic MBs belongs to the SHH subgroup.  However, classic and large-cell/anaplastic histology are also seen.  The prognosis of SHH MB, similar to that of group 4 tumours, is intermediate between that of WNT (most favourable) and group 3 (least favourable) tumours (Taylor et al., 2012).  However, whether this finding is true for just pediatric tumours is not clear.  A recent study showed that a new subtype of SHH MB is characterized by CXCR4.  The CXCR4-SHH tumours are more common in youngest patients, and are associated with desmoplastic histology as well as an increased expression of MATH1 and cyclin D1 (Sengupta et al., 2012). The group 3 MBs, which often carry a classic or large-cell/anaplastic histology, are the most aggressive tumours among all the subgroups.  These tumours can be recognized by immunohistochemical positivity of NPR3 though transcriptional profiling remains the most reliable method to assign subgroup affiliation.  Group 3 MBs over-express several genes that were initially identified through their function in retinal development; yet the role of these genes in the pathogenesis of this subgroup of tumours remains enigmatic (Cho et al., 2011; Kool et al., 2008; Northcott et al., 2011b).  One of the most characteristic molecular aberrations of group 3 tumours is MYC over-expression.  Although both WNT subgroup and group 3 tumours show elevated levels of MYC expression, MYC amplification is exclusive to group 3 MB (Cho et al., 2011; Kool et al., 2008; Northcott et al., 2011b).  As a result, Hatten et al. suggested that group 3 could be called the MYC group (Hatten and Roussel, 2011).  Group 3 MBs are more common in males than in females and occur more frequently in infants and children.  The disease is rarely seen in adults and is often metastatic with poor prognosis (Cho et al., 2011; Kool et al., 2008; Northcott et al., 2011b). The group 4 tumours account for a big proportion of (>30%) MBs; however, it is the least understood molecular subgroup.  These tumours are currently identified through transcriptional profiling and perhaps through immunohistochemical staining of KCNA1 as well, although the latter requires further validation (Cho et al., 2011).  Isochromosome 17q (66% of cases) and loss of the X chromosome in female patients are the most common cytogenetic changes observed in group 4 MB (Kool et al., 2012; Taylor et al., 2012).  Over-expression of genes involved in neuronal development and differentiation has been reported in this subgroup of tumours although the relevance to disease pathogenesis has yet to be ascertained (Cho et al., 2011; Kool et al., 2008; Northcott et al., 2011b).  Two recent genomic studies have not only confirmed the genetic mutations previously reported, but have also identified novel genetic alterations in distinct subgroups of MBs.  Significantly, tetraploidy was found to be a frequent early event in group 3 and 4 tumours (Jones et al., 2012), which are also frequently subjected to  29 genetic aberrations (for example, KDM6A and ZMYM3, which are the regulators of H3K27 and H3K4 trimethylation in group 3 and 4 tumours) targeting specific components of the epigenetic machinery (Robinson et al., 2012).  The improved understanding of the biology of the group 3 and 4 tumours may guide the design of therapies to these two subgroups of diseases for which pathway-specific inhibitors are yet available.  1.11.6 STANDARD TREATMENT SURGERY AND RADIATION Surgery remains the mainstay of MB treatment and offers survival benefits, particularly to children with localized disease (Grill et al., 2005; Zeltzer et al., 1999).  Surgical challenges may result from the proximity of tumours with the fourth ventricle or even the brainstem, and the associated risk of morbidity is substantially increased if complete resection is attempted (Crawford et al., 2007).  Therefore, involvement of the floor of the fourth ventricle sometimes precludes removing the entire tumour.  Surgical morbidities can be infectious or non-infectious on top of direct neurological consequences of surgical manipulation (Crawford et al., 2007).  It was reported that approximately 50% of MB patients experience long-term speech and language apraxia after surgery, known as cerebellar mutism (Robertson et al., 2006).  Despite the associated risks and morbidity, surgery, with the goal of debulking tumour mass, relieving hydrocephalus and obtaining a histological diagnosis, remains the front-line treatment due to its benefit in children with newly diagnosed MB (Klesse and Bowers, 2010). Radiation therapy is the most effective adjuvant to surgical resection in the management of MB.  However, RT is delayed or not given to children less than 3 years of age because of the detrimental effects that radiation have on the developing brain. For children who receive the therapy, radiation is directed to the primary tumour site as well as the craniospinal axis, with the aim to decimate any residual tumour cells not detected in the cerebral spinal fluid (CSF) or by MRI.  The previous protocol suggested 36 Gy of photon radiation for both average-risk and high-risk groups with a posterior fossa boost to a total dose of 54.0-55.8 Gy.  Though RT increased the 5-year EFS to approximately 60% (Crawford et al., 2007), the deleterious side-effects associated with the treatment, including neurocognitive decline, mobility impairment and endocrinopathies, are so profound that a reduction in the dose of radiation has been considered for average-risk patients (Deutsch et al., 1996; Gajjar et al., 2006; Packer et al., 1999; Thomas et al., 2000).  A prospective, randomized trial comparing standard-dose (36.0 Gy) and reduced-dose (23.4 Gy) RT in average-risk MB patients without adjuvant chemotherapy showed early treatment failures in the reduced-dose arm of the treatment (Deutsch et al., 1996).  As a result,  30 a number of subsequent studies attempted to incorporate chemotherapy during and after RT and showed an improvement in patient survival.  For example, the study by Packer et al. demonstrated that concurrent RT and vincristine, followed by adjuvant chemotherapy led to a promising 5-year EFS of ~80% (Packer et al., 1999; Packer et a., 2006).  Together, these successful clinical experiences established the foundation for reduced-dose RT in the current treatment of average-risk MB. CHEMOTHERAPY Although chemotherapy has been used in MB treatment for nearly five decades, the benefit of its use was not confirmed until the last twenty years (Evans et al., 1990; Tait et al., 1990; Taylor et al., 2003).  Cisplatin, vincristine, lomustine, cyclophosphamide, CCNU and oral etoposide are some of the most commonly used chemotherapeutic agents in MB treatment.  The primary goal for the use of chemotherapy is to augment, delay or even avoid RT.  For patients who have non-disseminated MB after gross total resection, post-surgery chemotherapy treatment alone can be curative (Grill et al., 2005).  However, under most circumstances, chemotherapeutic agents are given in combination with RT to provide maximal benefits to patients.  The drugs can be given concurrently with RT or administered during the “maintenance phase”, which is after surgery and RT. The Children’s Cancer Study Group (CCSG)-942 protocol involved a randomized controlled trial comparing the efficacy of high-dose RT (36 Gy) plus chemotherapy with RT alone.  Results suggested that the addition of chemotherapy prolonged the survival of high-risk but not average-risk MB patients (Evans et al., 1990).  A subsequent report by the Children’s Cancer Group (CCG) showed that RT dose-reduction from 36 to 23.4 Gy, when combined with chemotherapy, was able to maintain the rates of event-free and overall survival (Packer et al., 1999).  In Europe, the use of high-dose intrathecal and intravenous methotrexate, despite being efficacious, was associated with neurological toxicities and adverse effects on neurocognition (Nelson and Frank, 1981).  Current treatment approaches include the use of myeloablative doses of chemotherapy followed by autologous stem cell rescue, yielding a promising 5-year survival of 85% and 70% for average-risk and high-risk patients, respectively, with tolerable toxicities (Packer et al., 2006; Gajjar et al., 2006).  Chemotherapy is now considered the standard-of-care for children in both risk groups (Gottardo and Gajjar, 2006). The outcome for infants treated with chemotherapy is unfortunately much less optimistic than older children, with a 5-year progression-free survival of ~30%.  Furthermore, an increase in the chemotherapy doses paired with autologous stem cell rescue has not shown much  31 improvement in outcome.  In contrast to older children, the survival rates for children <3 years of age with MB treated by surgery, RT and chemotherapy was dismal (25-45%) until the last decade (Zeltzer et al., 1999).  It appears that the disease is more fatal and aggressive in younger children and this observation may be explained in part by increased surgical risks, under-treatment and a potentially “more aggressive” biology of tumours in these patients (Rutkowski et al., 2010a).  Radiation is not considered or is delayed in infants and young children <3 years of age to avoid RT-induced cognitive deficits and other severe long-term sequelae.  As a result, chemotherapy plays an integral part to the management of MB in this specific subgroup of patients.   High-dose chemotherapy, which has been assessed as part of the multi-modal treatment strategy in early childhood MB, is used in combination with conventional chemotherapy to avoid or delay RT, or as part of salvage chemotherapy when disease recurs (Rutkowski et al., 2010a).  For example, the Head Start I study, designed to avoid RT, consists of a myeloablative consolidation chemotherapy supplemented with autologous bone marrow transplant after surgery and conventional induction chemotherapy.  The optimal objective response rate (86%) and a 2-year overall survival rate of 62% (Mason et al., 1998) resulted in the continuation of the study, with a slight modification to the protocol (Head Start II), which adds high-dose systemic methotrexate to the induction regimen (Chi et al., 2004).  The Head Start II treatment showed a 5-year overall survival of 70+/-10% in children less than 3-years of age with non-metastatic MB at diagnosis (Dhall et al., 2008).  The current Head Start III study is investigating an induction regimen modified by the addition of oral etoposide and temozolomide (Rutkowski et al., 2010a).  The long-term side-effects of chemotherapy and radiation adversely affect the quality-of-life of the survivors of MB.  Therefore, novel therapeutic strategies that are less harmful to patients are desperately needed.  From a genome-wide siRNA library screen, we identified polo-like kinase 1 (PLK1) as a potential molecular target for the treatment of pediatric malignancies.  We demonstrated that the inhibition of this kinase remarkably suppressed the cell growth and induced apoptosis of a panel of pediatric cancer cell lines in 72 hours (Hu et al., 2009).  The initial promising result encouraged us to pursue this target further in brain cancers. 1.12 POLO-LIKE KINASE 1 (PLK1) 1.12.1 BACKGROUND Cancer is a disease characterized by accelerated cell proliferation.  Various therapeutic strategies have therefore been designed to target molecules that control cell division.  For example, mutations of CDKs have been identified and actively pursued as potential molecular  32 targets for cancer treatments (Fischer and Lane, 2000).  Another new class of cell cycle regulators, namely the family of polo-like kinases (PLK), has recently emerged and is recognized for pivotal roles in processes such as regulation of checkpoint progression and cell division during mitosis.   The first identified PLK family member was Polo, which is a serine/threonine kinase in Drosophila melanogaster (Fenton and Glover, 1993; Glover et al., 1998).  The PLK proteins were originally named after the phenotype in flies.  In Drosophila larval neuroblasts, mutation in the polo gene was found to result in the formation of abnormal spindle poles and induce aberrant mitoses (Barr et al., 2004; Golsteyn et al., 1994).  Five mammalian PLK family members: PLK1, PLK2 (SNK), PLK3 (FNK or PRK), PLK4 (SAK) and the most recently discovered PLK5, were subsequently identified (Clay et al., 1993; Fode et al., 1994; Hamanaka et al., 1994; Holtrich et al., 1994; Holtrich et al., 2000; Karn et al., 1997; Lake and Jelinek, 1993; Ouyang et al., 1997; Simmons et al., 1992).  The most well-characterized member of the human PLK family is PLK1, the activity and expression of which are integral to the precise regulation of cell division (Mundt et al., 1997; Smith et al., 1997).  There is mounting evidence suggesting that PLK1 belongs to a family of disease-relevant protein kinases which can be inhibited by different drugs and that the molecular targeting of PLK1 represents a promising approach for the development of novel anti-cancer therapies (Strebhardt and Ullrich, 2006).  1.12.2 FAMILY MEMBERS The PLK family is comprised of five members: PLK1, 2, 3, 4 and 5 and will be referred collectively as “PLKs.”  The members of the PLK family share commonalities in their overall structures and functions.  PLK1, 2, 3 and 5 have two conserved polo-box motifs except PLK4, which only has single polo-box (PB) motif.  The polo-box domains (PBDs) of PLK1, 2 and 3 preferentially bind pSer or pThr peptides (Elia et al., 2003a; Elia et al., 2003b) and mediate protein-protein interactions (Lee et al., 1998; Logarinho and Sunkel, 1998).  Importantly, through binding specific substrates via their PBDs, PLKs are able to localize precisely to various mitotic structures, to phosphorylate the downstream molecules and to carry out various functions at different stages of the cell cycle.  PLKs are involved in regulation of G1/S transition, mitotic entry, cytokinesis and cellular response to DNA damage (Golsteyn et al., 1994; Lane and Nigg, 1996; Lee et al., 1995; Moutinho-Santos et al., 1999; Song et al., 2000). PLK1 PLK1 is the best-characterized PLK family member.  The expression of this kinase is cell cycle-regulated; it rises in G2 and peaks in M phase where it regulates much of the machinery  33 involved in mitosis (Barr et al., 2004; Golsteyn et al., 1994).  PLK1 plays multiple roles in the regulation of cell cycle progression including centrosome maturation, bipolar spindle formation (Casenghi et al., 2005; Feng et al., 2006; Lane and Nigg, 1996; Rapley et al., 2005; Sumara et al., 2004; van Vugt et al., 2004b; Yarm, 2002), mitotic entry (Roshak et al., 2000), metaphase-to-anaphase transition (Golan et al., 2002; Hansen et al., 2004; Moshe et al., 2004) and cytokinesis (Neef et al., 2003; Niiya et al., 2006; Seong et al., 2002; Zhou et al., 2003).  PLK1 is an oncogenic kinase that confers growth and survival advantage in cancer cells.  Cells treated with PLK1 inhibitors undergo pro-metaphase arrest (Liu and Erikson, 2003) and subsequent mitotic catastrophe as a result of abnormal spindle formation and chronic spindle checkpoint activation (Lenart et al., 2007; Steegmaier et al., 2007).  In multiple cancer models, PLK1 inhibition specifically eliminates the malignant cells while leaving the non-malignant cells unharmed (Cogswell et al., 2000; Grinshtein et al., 2011; Liu et al., 2006b; Renner et al., 2009).  Furthermore, PLK1 is highly expressed in cancer cells but not in their normal cell counterparts (Holtrich et al., 1994; Lee et al., 2012; Yuan et al., 1997), rendering this kinase a particularly attractive molecular target for cancer therapeutics. PLK2 PLK2, which is also known as serum-inducible kinase (SNK), is expressed primarily in early G1 where it controls the entry into S phase (Liby et al., 2001; Ma et al., 2003).  Unlike PLK1 which negatively regulates the activity and stability of p53 (Ando et al., 2004b) and is inhibited by DNA damage in G2/M (Jang et al., 2007), PLK2 is a direct transcriptional target of p53 and is activated upon DNA damage (Burns et al., 2003).  PLK2 functions collaboratively with CHK1 and CHK2 to invoke S-phase arrest (Matthew et al., 2007).  In contrast to the prevalent over-expression of PLK1 in tumourigenesis, epigenetic silencing of PLK2 due to CpG methylation in the transcriptional regulatory elements of this gene is frequently seen in B-cell malignancies (Smith et al., 2006).  PLK2 is implicated in centriole duplication (Warnke et al., 2004) as well as regulation of synaptic plasticity in the nervous system (Ang et al., 2008; Inglis et al., 2009; Seeburg et al., 2008). PLK3 PLK3, also designated as FGF-inducible kinase (FNK or PRK), is considered to be a tumour suppressor and a regulator of cellular response to DNA damage and angiogenesis (Bahassi el et al., 2002; Donohue et al., 1995; Xie et al., 2001).  PLK3 expression is constant throughout the cell cycle but is induced by DNA damage, which results in ATM-dependent phosphorylation, and the activation of PLK3 (Bahassi el et al., 2002; Xie et al., 2001; Xie et al.,  34 2002).  The activated PLK3 facilitates the phosphorylation of CHK2 by ATM, leading to checkpoint activation (Bahassi el et al., 2002).  PLK3 further contributes to the DNA checkpoint response by phosphorylating p53 at Ser20, thereby increasing the stability of p53 (Xie et al., 2001).  PLK3 is down-regulated in cancers of the lung, head and neck, uterus and bladder (Ando et al., 2004b; Dai et al., 2000; Li et al., 1996) but up-regulated in some other cancers (Ando et al., 2004b; Weichert et al., 2004a; Weichert et al., 2005). PLK4 PLK4 (SAK) is required for late mitotic progression, cell survival and post-gastrulation embryonic development (Hudson et al., 2001).  PLK4 plays a role in M phase progression (Seeburg and Sheng, 2008); it localizes to the centrosomes and the cleavage furrow during cytokinesis as seen with PLK1 and shares the function with PLK2 in centriole duplication (Bettencourt-Dias et al., 2005).  Mice heterozygous for Plk4 show haploinsufficiency with increased susceptibility to cancer development (Ko et al., 2005).  In line with this, approximately 50% of human hepatocellular carcinomas have loss of heterozygosity at the PLK4 locus (Rosario et al., 2010).  However, over-expression of this kinase also results in multi-nucleated cells (Fode et al., 1996) and may be associated with the development of colon cancers (Macmillan et al., 2001).  A stringent regulation of the expression of PLK4 therefore seems to be crucial for maintaining genomic stability. PLK5 PLK5 is the most recently discovered PLK family member.  In the first report published by Andrysik et al., PLK5 was found to share more sequence homology with PLK2 and PLK3 than with PLK1 and PLK4 in mice (Andrysik et al., 2010).  In addition, mouse Plk5 is a DNA damage inducible gene and the protein localizes mainly to the nucleolus.  Moreover, ectopic expression of this kinase results in G1 arrest, decreased DNA synthesis and apoptosis.  De Cárcer et al. further characterized PLK5 and demonstrated its tumour suppressor role in brain cancer pathogenesis (de Carcer et al., 2011).  Interestingly, unlike PLK1 that accelerates cell cycle progression, PLK5 exerts the opposite effects and is down-regulated in proliferating cells.  This kinase is predominantly expressed in mouse and human brains and was shown to regulate the formation of neuritic processes when the brain-derived neurotrophic factor (BDNF)/nerve growth factor (NGF)-RAS pathway is activated in neurons. Consistent with its tumour suppressor role, over-expression of PLK5 induces apoptosis of brain tumour cells.  The human PLK5 gene is largely silenced by promoter hypermethylation in astrocytoma and GBM.  35 1.12.3 PROTEIN STRUCTURE Human PLKs consist of a highly conserved N-terminal serine/threonine kinase domain of 252 amino acids and a unique C-terminal duplicated polo-box motif (together named the polo-box domain or PBD), each of which is 60-70 amino acids in length (Figure 1.3).  This structural feature is shared among all the PLKs except PLK4, which has only one polo-box (PB) (Leung et al., 2002) and PLK5, which lacks the kinase domain (de Carcer et al., 2011).  The following discussion will focus specifically on PLK1. The kinase domain of human PLK1 was modeled (Lowery et al., 2005) using the SWISS-MODEL server (Guex and Peitsch, 1997) based on the crystal structure of Aurora A (Bayliss et al., 2003; Nowakowski et al., 2002), Aurora B (Cheetham et al., 2002) and AKT (Yang et al., 2002).  According to the available homology-model structure, the N-terminal catalytic domain of PLK1 includes almost all the invariant residues found in most serine/threonine protein kinases (Hanks and Quinn, 1991; Hanks et al., 1988).  The consensus motif in the ATP-binding domain of PLK1 is Gly-X-Gly-X-X-Ala instead of the canonical consensus sequence: Gly-X-Gly-X-X-Gly seen in typical protein kinases (Hanks et al., 1988).  The activation of PLK1 may be regulated by human STE20-like kinase (SLK) and serine/threonine kinase 10 (STK10), which phosphorylate Thr-210 within the T-loop to stimulate PLK1 kinase activity (Ellinger-Ziegelbauer et al., 2000; Walter et al., 2003).  Recent studies suggest that Aurora A and Bora are additional upstream regulators that cooperatively activate PLK1 and control mitotic entry (Seki et al., 2008). The structure analysis of the PBD was first derived from the crystal structure of the murine protein Sak, residues 845-919 (Leung et al., 2002).  The PBD of human PLK1 is comprised of a polo-box that crystallizes as an intermolecular dimer containing two α-helices and two six-stranded β-sheets.  The polo-cap (Pc) region, a special structure in the PBD, holds both polo-boxes in the correct orientation (Cheng et al., 2003).  Results from oriented peptide libraries indicate that the PBD contains a phosphopeptide binding motif (Elia et al., 2003a; Yaffe and Cantley, 2000), which is a key structure regulating the interaction of PLK1 with other proteins (Lee et al., 2008b).  Specifically, the substrates of PLK1 must be phosphorylated by other kinases prior to binding to PBD.  The mutual inhibitory interaction between the kinase domain and PBD reduces the kinase activity by 3-fold and phosphopeptide binding capacity by 10-fold (Elia et al., 2003b; Jang et al., 2002).  The binding of optimal phosphopeptide to the full-length PLK1 not only targets the kinase to its substrates but also stimulates its kinase activity (Chopra et al., 2010).  CDKs, MAP kinases and other mitotic kinases are some of the proteins that have been shown to phosphorylate PLK1 docking proteins or downstream substrates,  36 thereby generating these “primed” pSer/pThr proteins that can be subsequently recognized by the PBD of PLK1.  The interaction between PLK1 and its phosphorylated substrates allows specific subcellular localization of PLK1, a prerequisite for this kinase to carry out pleiotropic functions in mitosis (Lee et al., 2008b).  The optimal sequence motif recognized by the PBD is Ser-[pSer/pThr]-[Pro/X] (Elia et al., 2003a).    Figure 1.3 The structure of PLKs. The PLK proteins are composed of two parts: the N-terminal kinase domain and the C-terminal polo-box domain (PBD), which contains 2 polo-boxes, PBs (except PLK4 which has only one PB).  The two domains are joined by a linker region and a small region called the polo-box cap (Pc) that may be part of the PBD.  Trp414, His538 and Lys540 are three key residues implicated in phospho-peptide binding.  The T-loop (Thr210) and the D-box motif are involved in the activation and degradation of PLK1, respectively.  1.12.4 BIOLOGY AND FUNCTIONS  The interactions between PLK1 and its upstream regulators as well as its substrates control its activation and subcellular localization (Chopra et al., 2010).  Recent studies demonstrate that Bora, a G2/M-expressed protein, facilitates Aurora A to phosphorylate the Thr210 residue in the T-loop of PLK1, thereby activating the kinase and promoting mitotic entry (Macurek et al., 2008; Seki et al., 2008). The transition from G2 to mitosis is tightly regulated by M-phase-promoting factor (MPF), which consists of CDK1 and its cognate cyclins.  The CDK/cyclin complexes are kept inactive by the inhibitory phosphorylation mediated by WEE1 on Tyr15 of CDK1 to prevent premature entry into mitosis.  PLK1 drives mitotic progression by relieving the constraint imposed by WEE1 (Chopra et al., 2010).  Specifically, PLK1 activates CDC25C, which in turn dephosphorylates the inhibitory phosphate on Tyr15 of CDK1 and thus activates CDK/cyclin complexes.  In addition, Kinase Domain Polo-Box Domain (PBD) Pc PB1 PB2 Thr210 Arg337 Trp414 His538 Lys540 303 407 506 603 Phospho-selectivity D-box T-loop Adapted from Barr et al., Nature Reviews Molecular Cell Biology (2004)    37 PLK1 directly phosphorylates WEE1, resulting in its degradation and a decreased inhibitory phosphorylation on CDK1.  The biological effects of CDC25C phosphorylation by PLK1 remain to be clarified.  While early reports indicate that phosphorylation controls CDC25 activity (Kumagai and Dunphy, 1996), recent studies focus on the effects on CDC25 subcellular localization (Toyoshima-Morimoto et al., 2002).  Further experimental evidence suggests that PLK1 may phosphorylate cyclin B just outside the NES at Ser133, thereby activating CDK1/cyclin B at the centrosome and promoting nuclear import of the complex (Jackman et al., 2003; Yuan et al., 2002).  The diverse effects that PLK1 has on various mitotic regulators converge to form a positive feedback that irreversibly drives mitotic entry (Figure 1.4).    Figure 1.4 The regulation of CDK1 by PLK1 in the G2/M transition. The CDK1/cyclinB complexes are kept inactive due to the inhibitory phosphorylation by WEE1 on Tyr15 and MYT1 on Thr 14 of CDK1 to prevent premature entry into mitosis.  PLK1 drives the G2/M transition by inhibiting WEE1 and MYT1.  PLK1 activates CDC25C, which subsequently dephosphorylates the inhibitory phosphate on Tyr15 and Thr14 of CDK1 and thus activates the CDK1/cyclinB complexes.  CDK1, when activated, can phosphorylate CDC25C (activation) and WEE1/MYT1 (inhibition), resulting in a positive feedback loop.  The CDK1/cyclinB complexes might also activate PLK1 (dashed arrow), which phosphorylates cyclin B.  Furthermore, PLK1-mediated phosphorylation of WEE1 results in its degradation.  PLK1 plays an essential role in mitotic spindle formation.  The establishment of bipolar spindles may be one of the most important functions of PLK1 in the cell cycle because the most pronounced biological effect of PLK1 inhibition is defects in spindle morphology, which leads to subsequent mitotic arrest (Kraft et al., 2003; Lee et al., 1995; Liu and Erikson, 2002; van Vugt Thr14-P Tyr15-P CDK1 Cyclin B PLK1 CDC25 WEE1 MYT1 Thr14 Tyr15 CDK1 Cyclin B P Low  CDK  activity High  CDK  activity Adapted from Barr et al. Nature Reviews Molecular Cell Biology (2004)  38 and Medema, 2005).  Indeed, the characteristic phenotype of dysfunctional PLKs is a severe mitotic spindle defect, which is seen in various organisms, from yeast to mammals, in loss-of-function studies (Kitada et al., 1993; Llamazares et al., 1991; Sunkel and Glover, 1988).  In Drosophila, Polo mutants showed a high percentage of monopolar spindles or bipolar spindles with one of the poles having abnormal morphology (Sunkel and Glover, 1988).  An observation was similarly made in human cells treated with a targeting antibody that interferes with PLK1 function (Lane and Nigg, 1996).   The loss of bipolar spindle formation can be attributed to impaired centrosome maturation and separation.  PLK1 was found to regulate spindle morphology by recruiting  γ-tubulin to the centrosome (Lane and Nigg, 1996).  In addition, PLK1 phosphorylates Ninein-like protein (NLP) and OP18/Stathmin to displace these negative regulators of microtubule nucleation, thus allowing tubulin-nucleation to occur (Casenghi et al., 2003).  PLK1 further contributes to spindle formation by positively regulating microtubule-stabilizing proteins ASP and TCTP (do Carmo Avides et al., 2001; Yarm, 2002).  PLK1 phosphorylation of ASP is required for the aster-forming capacity of centrosomes and the establishment of robust spindles (Elia et al., 2003a).  A study by van Vugt et al. showed similar mitotic defects in PLK1-deficient cells and monastrol (a drug that inhibits centrosome maturation)-treated cells.  Interestingly, the inactivation of spindle checkpoint signaling permits both PLK1-deficient cells and monastrol-treated cells, carrying a monopolar spindle phenotype, to progress through mitosis without disrupting midbody formation and cleavage furrow ingression.  This result argues that centrosome defects are the primary aberrations that arise after PLK1 inactivation whereas other phenotypes may occur as a manifestation of the impaired spindle function (van Vugt et al., 2004b). As a cell progresses through prophase and metaphase, duplicated chromosomes become tethered to the spindle poles and are aligned along the equatorial plate.  The anaphase promoting complex or cyclosome (APC)/C is a multi-subunit ubiquitin ligase that targets mitotic proteins for proteosomal degradation.  In anaphase, APC/C ubiquitinates securin that helps hold the paired sister chromatids together, thereby allowing the separation of duplicated chromosomes.  To ensure accurate chromosome segregation in anaphase, APC/C activity is under strict control by the spindle checkpoint (Chopra et al., 2010) and its ancillary proteins CDC20 and CDH1 (Peters, 2002).  The early mitotic inhibitor-1 (EMI-1) binds to and inhibits the regulatory subunit (CDC20) of APC/C (Reimann et al., 2001b).  In this way, APC/C is kept inactive by EMI-1 to allow accumulation of cyclins A and B during interphase and to prevent premature activation of APC/C (Reimann et al., 2001a; Reimann et al., 2001b).  When  39 duplicated chromosomes have attained bipolar attachment and are properly aligned at the metaphase plate, the inhibitory control on CDC20 by EMI1 must be relieved to allow APC/C-mediated ubiquitination of securin to proceed.  A two-step model was proposed: CDK1 primes EMI1 for PLK1-mediated phosphorylation and subsequent degradation (Margottin-Goguet et al., 2003; Moshe et al., 2004).  Therefore, PLK1 regulates APC/C activity indirectly through EMI1.  Experimental evidence suggests that PLK1 may also directly control APC/C activity by phosphorylating its subunits, leading to an enhanced ubiquitin ligase activity (Golan et al., 2002; Kotani et al., 1998; Kraft et al., 2003).  The role of PLK1 in cytokinesis is not very well defined.  The difficulty in pinpointing a specific role for PLK1 in cytokinesis lies in the inability to analyze this process in the absence of PLK1 without affecting the previous mitotic processes (van Vugt and Medema, 2005).  Although PLK1 has been shown to interact with mitotic kinesin-like protein-1/-2 (MKLP1/CHO1 and MKLP2) (Adams et al., 1998; Lee et al., 1995; Liu et al., 2004; Neef et al., 2003), NIR2 (Litvak et al., 2004), NUDC (Zhou et al., 2003) and localize to the central spindle, its exact function at this subcellular location is not well characterized.  In the study by van Vugt et al., midbody formation and furrow ingression occurred in PLK1-depleted cells, suggesting that this kinase may be dispensable for these two processes to occur in cytokinesis (van Vugt et al., 2004b).  While the removal of the PLK1 phosphorylation sites in MKLP1/CHO1 and MKLP2 has no effects on furrow ingression, it prevents abscission, indicating an absolute requirement of PLK1 function in the final step of cytokinesis (Liu et al., 2004; Neef et al., 2003).  Additional studies, however, suggest that it is the efficient elimination of PLK1 but not its activity that is required for mitotic exit.  For example, disrupting the D-box motif in PLK1 prevents its degradation and delays mitotic exit (Lindon and Pines, 2004).  In addition, inactivation of PLK1 may increase microtubule bundling at the central spindle, leading to a local decrease in microtubule concentration, a process thought to precede cytokinesis (Dechant and Glotzer, 2003).  To summarize, proper progression of cytokinesis requires the correct localization of PLK1 to the central spindle as well as the timely degradation of this kinase (van Vugt and Medema, 2005).  The substrates of PLK1 in G2/M transition and the various steps of mitosis are summarized in Table 1.1.    40  Table 1.1 The substrates of PLK1 in G2/M transition and the different stages of mitosis.  The DNA damage checkpoint is a complex network that prevents CDK/cyclin activation and arrests cell cycle progression in order to provide time for cells to repair damaged DNA (van Vugt and Medema, 2005).  This process entails the activation of ATM and ATR, which relays the signal to downstream checkpoint mediators including p53, CHK1, CHK2, histone H2AX and components of the DNA repair mechanisms (Shiloh, 2003).  Consequently, CDC25 family members, CDK1/cyclin B complex and PLK1 (Smits et al., 2000) are inactivated and WEE1 function is enhanced.  PLK1, a direct target of the G2 DNA damage checkpoint, is inhibited by functional ATM or ATR (van Vugt et al., 2001b).  All together, these actions ensure that all the proteins and mechanisms that drive mitotic entry are completely inactivated so that the cell has sufficient time to repair DNA damage.   Not only is PLK1 a target for the DNA damage checkpoint, it is also a key regulator that resumes cell cycle progression after the damaged DNA is repaired.  The first evidence for this role came from studies in S. cerevisiae in which yeast PLK- CDC5 was shown to mediate checkpoint adaptation (Toczyski et al., 1997), a process that permits a cell with irreparable G2/M DNA damage to re-enter the cell cycle (Sandell and Zakian, 1993).  Checkpoint adaptation confers survival advantage to yeasts (Galgoczy and Toczyski, 2001; Lee et al., 2001) but the effects on the human genome could be detrimental.  Therefore, checkpoint adaptation is not seen in mammalian cells, which restart the cell cycle only after DNA damage is completely repaired.  PLK1 was shown to be a critical factor in this “checkpoint recovery” process through regulation of CDK1 phosphorylation (van Vugt et al., 2004a).  In addition, the G2 checkpoint needs to be silenced in the recovery process and this is mediated by PLK1 phosphorylation and degradation of Claspin, which is a co-activator of CHK1.  The destruction of Claspin inactivates CHK1, thereby allowing the activation of CDC25B/C (Mailand et al., 2006; Mamely et al., 2006; Peschiaroli et al., 2006).  During checkpoint activation, the inhibition of CDK1/cyclin B is relieved by PLK1-mediated phosphorylation and degradation of WEE1 (van Vugt et al., 2004a).  A recent PLK1 Substrates ReferencesG2/M Transition CDC25C Qian et al., 2001; Toyoshima-Morimoto et al., 2002Cyclin B  Toyoshima-Morimoto et al., 2001; Qian et al., 2001MYT1 Nakajima et al., 2003WEE1 Watanabe et al., 2004M-PhaseMicrotubule Dynamics Stathmin/Op18 Budde et al., 2004TCTP Cucchi et al., 2010; Johnson et al., 2008Microtubule Nucleation/Anchoring NLP Casenghi et al., 2003; Rapley et al., 2005Anaphase/Telophase/Cytokinesis MKLP1/2  Lee et al., 1995; Adams et al., 1998; Neef et al., 2003NUDC Zhou et al., 2003 41 study by Liu et al. demonstrated a link between PLK1 and p53 in checkpoint recovery (Liu et al., 2010).  The G2- and S-phase-expressed 1 (GTSE1) protein is a negative regulator of p53 and antagonizes its action by exporting it out of the nucleus and inducing its degradation (Monte et al., 2003; Monte et al., 2004).  During checkpoint recovery, PLK1 phosphorylates GTSE1 at Ser435, activates its nuclear import signal and encourages its nuclear accumulation.  In the later stages of DNA damage-induced cell cycle arrest, GTSE1 begins to accumulate in the nucleus (Monte et al., 2003; Monte et al., 2004), suggesting that it may be involved in checkpoint recovery by down-regulating p53 (Bahassi el, 2011).  Interestingly, although PLK1 is activated in G2-phase in an unperturbed cell cycle and in checkpoint recovery, it is only essential in the latter process (Kelm et al., 2002; Yamashiro et al., 2008). 1.12.5 ROLE IN CANCER PATHOGENESIS PLK1 plays a key role in various stages of mitotic progression and its expression is strongly associated with cell proliferation.  In addition to promoting cell division, PLK1 over-expression may lead to chromosomal instability and aneuploidy, a hallmark of cancer (Degenhardt and Lampkin, 2010).  Experiments have shown that high PLK1 expression allows cells to over-ride mitotic checkpoints, leading to premature cell division and subsequent aneuploidy (Kops et al., 2005).  The oncogenic property of PLK1 was further established in an experiment where over-expression of this kinase transformed NIH3T3 cells, allowing them to grow in anchorage-independent conditions in soft agar and to form tumours in nude mice (Smith et al., 1997). Although most experimental evidence suggests that PLK1 is involved in tumourigenesis, the genetic evidence for PLK1 being an oncogene is not very strong (Degenhardt and Lampkin, 2010).  The PLK1 gene is rarely amplified.  The missense mutations reported thus far in cell lines disrupt PLK1-heatshock protein 90 (HSP90) interaction, resulting in reduced stability of PLK1 (Simizu and Osada, 2000). Nevertheless, the fact that PLK1 is over-expressed in a wide range of human malignancies and its inhibition selectively eliminates tumour cells but not normal cells makes it an attractive molecular target for cancer therapy.  The expression of PLK1 is below the limit of detection in most adult tissues except in organs, such as the thymus, spleen and testis, where active cell proliferation is observed (Golsteyn et al., 1994; Hamanaka et al., 1994; Holtrich et al., 1994).  However, its expression is elevated in prostate tumours, ovarian carcinomas, hepatoblastomas, melanomas, breast and brain tumours where high levels are correlated with poor prognosis (Kneisel et al., 2002; Lee et al., 2012; Maire et al., 2013; Weichert et al., 2004a;  42 Weichert et al., 2004b; Yamada et al., 2004).  Early studies showed that PLK1-depletion by synthetic dsRNAs and short hairpin RNAs blocked the proliferation of cultured tumour cells (Liu and Erikson, 2003; Spankuch-Schmitt et al., 2002) and reduced the growth of human tumour cells in xenograft models (Spankuch et al., 2004).  Significantly, PLK1 inhibition preferentially eliminates cancer cells while leaving normal cells unharmed, providing a potential therapeutic window in which we could target this kinase safely (Cogswell et al., 2000; Liu et al., 2006b; Xie et al., 2005).  Various small molecule inhibitors of PLK1 have recently been developed and evaluated in phase I/II clinical trials for their efficacies on a number of human malignancies.  Three of the inhibitors used in my studies- BI2536, BI6727 and GSK461364 will be discussed in the following paragraphs.  1.12.6 INHIBITORS AND CLINICAL TRIALS BI2536 (Boehringer Ingelheim, Ingelheim, Germany), a dihydropteridinone derivative, is a first-in-class prototype ATP-competitive inhibitor to PLK1.  This compound inhibits the enzymatic activity of PLK1 with an IC50 of 0.8nM and shows >10,000-fold selectivity to PLK1 against 63 other tyrosine and serine/threonine kinases examined (Steegmaier et al., 2007).  In cell-based assays, BI2536 suppresses the cell proliferation of >20 cancer cell lines irrespective of the tissue of origin and oncogenome status, with EC50 values in the range of 2-25nM.  BI2536-treated tumour cells display perturbation of spindle assembly, pro-metaphase arrest, and immunostained positively for phospho-histone H3 (Steegmaier et al., 2007).  In addition, apoptosis is induced and is evidenced by PARP cleavage and positive staining in terminal deoxynucleotidyl transferase biotin-dUTP nick end labeling (TUNEL) assays (Mok et al., 2012).  In human carcinoma xenograft models, the intravenous administration of BI2536 once or twice weekly at well-tolerated doses results in marked inhibition of tumour growth and/or tumour regression (Steegmaier et al., 2007).  A phase I dose-escalation study showed that BI2536 was well-tolerated under a maximum tolerated dose of once-weekly 200mg/cycle or twice-weekly 100mg/cycle.  The compound exhibits a favourable safety profile and measurable anti-tumour activity with neutropenia and thrombocytopenia being the dose-limiting toxicities (DLTs) (Mross et al., 2008).  The promising results from phase I studies prompted further studies in phase II trials, which enrolled patients with non-small cell lung cancer (NSCLC), advanced pancreatic cancer and hormone refractory prostate cancer (HRPC) to be treated with BI2536 as a single agent.  In the NSCLC study, only 4.2% of the patients showed a partial response.  Dose-dependent and reversible neutropenia occurred in 37% of the patients and was the main side-effect (Sebastian et al., 2010a).  In the metastatic HRPC study, BI2536 was well tolerated with  43 some indications of anti-tumour activity by prostate-specific antigen (PSA) reduction and radiologically stable disease.  The modest efficacy of BI2536 as a single agent from this and other recent phase II studies (Schoffski et al., 2010a; Sebastian et al., 2010a) suggests that combination therapies should be considered. BI6727 is the second-in-class dihydropteridinone derivative developed by Boehringer Ingelheim.  This second generation compound is similar to BI2536 in many aspects.  For example, BI6727 also potently and selectively inhibits PLK1 by targeting its ATP-binding pocket and induces prometaphase arrest and subsequent apoptosis (Rudolph et al., 2009).  Furthermore, BI6727 inhibits the proliferation of a broad range of tumour cell lines in vitro with EC50 values of 11-37nM and demonstrates efficacy in xenograft models of multiple human cancers (Rudolph et al., 2009).  This compound has distinct pharmacokinetic (PK) properties characterized by sustained tumour exposure, a long terminal half-life, a large volume of distribution and good oral bioavailability (Rudolph et al., 2009).  The first-in-man trial of BI6727 by Schöffski et al. demonstrated a favourable PK profile of the drug with manageable hematological toxicities.  Stable disease at best response was observed in 40% of patients with encouraging anti-tumour activity (Schoffski et al., 2012).     GSK461364 (GlaxoSmithKline, Middlesex, UK) is a selective i.v. thiophene amide, ATP-competitive inhibitor of PLK1 (Schoffski, 2009).  Different concentrations of GSK461364 induce different phenotypes in cells: misaligned chromosomes and mitotic arrest with severely perturbed mitotic spindles are observed at 10nM and 250nM, respectively.  Delayed entry into mitosis and prometaphase arrest are seen at concentrations >300nM (Chopra et al., 2010).  The compound shows 400-fold greater potency for PLK1 than for PLK2 and PLK3 and inhibits the proliferation of >120 cell lines with IC50 <50nM in >83% of the cell lines tested.  In xenograft models, complete tumour growth inhibition or delay was observed in animals treated with GSK461364 (Schoffski, 2009).  The results from a phase I trial in patients with solid malignancies showed a best response of prolonged stable disease of more than 16 weeks in 15% of the patients.  The DLTs are neutropenia and thrombocytopenia.  Due to the high incidence (20%) of venous thrombotic emboli, further clinical evaluation of GSK461364 may require a coadministration of prophylactic anticoagulation (Olmos et al., 2011). 1.13 CANCER STEM CELL THEORY The concept of “no-neuron dogma”, which implies the absence of brain stem cells in adulthood, was challenged in the 1960’s when the generation of new, functional brain cells was described in the adult mammalian CNS (Altman, 1966; Altman and Das, 1965; Dacey and  44 Wallace, 1974).  In the late 1980’s, Nottebohm’s group demonstrated the functional relevance of adult neurogenesis in songbirds (Alvarez-Buylla et al., 1988; Goldman and Nottebohm, 1983).  Subsequently, Reynolds and Weiss reported the isolation of neural stem cells (NSCs) from adult mouse brain (Reynolds and Weiss, 1992), followed by a series of in vitro experiments highlighting the presence of human NSCs in adult brain, specifically the dentate gyrus of the hippocampus (Roy et al., 2000).  The first tantalizing evidence of the existence of NSCs in vivo in adult human brain was described in 2001 (Palmer et al., 2001).  The major implication of these studies is the existence of mitotically active stem and progenitor cells within a discrete region of the mature brain.  These long-lived cells have high potential in self-renewal, and over time, hold the greatest opportunity of accumulating genetic defects that transform normal cells (Al-Hajj et al., 2004; Vescovi et al., 2006).  The “cancer stem cell theory” proposes an hierarchical organization in tumours in which only a specific subset of cells can self-renew, proliferate extensively, establish and maintain a tumour clone while the rest of the variably differentiated cells cannot (Reya et al., 2001), a concept that differs from the traditional view of cancer development.  Furthermore, cancer stem cells have been shown to be resistant to conventional therapies and may be one of the causes for disease relapse (Figure 1.5).  The clonal evolution theory, on the contrary, indicates that all cells in a tumour have similar tumourigenic potential that is activated asynchronously and at low frequency.  These cells acquire genetic mutations, each of which confers additional proliferative and survival advantage.  The malignant progression towards a disease state proceeds in a manner that resembles Darwinian evolution (Reya et al., 2001).     Therapeutic Predictions of Cancer Stem Cell Model Treatments that kill CSCs may remove the “seeds” of a tumour. The “seeds” of cancer are eliminated-> complete cure may be achieved. Chemotherapeutic agents kill most of the cells in a tumour but NOT the CSCs. Cancer stem cells survive and repopulate to form new tumors-> disease recurs! Cancer Stem Cells  45 Figure 1.5 Therapeutic predictions of cancer stem cells model. Cancer stem cells, which are resistant to chemo- and radiation therapy, may survive the treatments and repopulate a new tumour.  Therapeutic strategies that target the CSCs may eliminate the “seeds” of a tumour, leading to a complete cure of the disease. 1.13.1 DEFINITION OF CANCER STEM CELLS The fact that brain tumours tend to recur after surgery may be attributable to the diffuse and infiltrative nature of the tumours and the presence of cancer stem cells (CSCs), which are also known as tumour initiating cells (TICs).  I realize that there may be nuances in the definitions of CSCs and TICs; however these two terms will be used interchangeably in this review for convenience.  The discovery of CSCs was first made by John Dick’s group through the isolation of a small population of cells in acute myeloid leukaemia that were capable of initiating leukaemia after transplantation.  These cells self-renewed, bore molecular features of normal haematopoietic stem cells, and were named “leukaemia initiating cells” (Bonnet and Dick, 1997; Hope et al., 2004; Lapidot et al., 1994).  A wealth of literature was subsequently published that reported the isolation of CSCs from breast (Al-Hajj et al., 2003), skin (Fang et al., 2005), pancreas (Li et al., 2007), colon (O'Brien et al., 2007), prostate (Patrawala et al., 2006) and brain tumours (Galli et al., 2004; Hemmati et al., 2003; Ignatova et al., 2002; Singh et al., 2003; Singh et al., 2004). CSCs or TICs are defined by the characteristics outlined by Vescovi and colleagues (Vescovi et al., 2006).  First, CSCs must be able to self-renew extensively, demonstrated either ex vivo by sequential clonogenic and population kinetic analyses (Galli et al., 2004; Gritti et al., 1996) or in vivo by serial, orthotopic transplantation (Galli et al., 2004; Singh et al., 2004).  Although transient amplifying progenitors are highly proliferative, they show limited self-renewal and cannot be propagated for an extended period of time.  Equally important, stem cells need to be able to generate progenies several orders of magnitude more abundant than the starting population (Reynolds and Rietze, 2005).  Furthermore, the term “tumour initiating cells” literally implicates the ability of these cells to instigate tumour formation when implanted in animals.  The resulting tumours recapitulate the histopathological features of the parental tumours from where the cells derived.  CSCs generate tumourigenic as well as non-tumourigenic cells and they are capable of undergoing multi-lineage differentiation though the latter may not be an absolute requirement for their identification.  Unlike normal stem cells, CSCs harbour karyotypic or genetic alterations in addition to aberrant differentiation properties (Vescovi et al., 2006).   46 Despite the phenotypic and functional similarities between normal stem cells and CSCs, it is important to be aware of the fundamental differences between these cells.  “While normal stem cells are known for the vigilance with which their proliferation is controlled and for the care with which their genomic integrity is maintained, CSCs lack such properties,” suggested in a CSC review by Shackleton (Shackleton, 2010).  The perennial nature and high proliferative potential of stem cells has made it tempting to speculate that CSCs may originate from malignant transformation of normal stem cells, yet thus far the cell-of-origin remains elusive.  Studies have shown that genetic alterations by way of tumour suppressor gene ablation or oncogene activation increase the frequency of tumour formation in primitive nestin-expressing cells but not in the more differentiated glial fibrillary acidic protein (GFAP)-expressing astrocytes (Holland et al., 2000; Holland et al., 1998).  However, conflicting experimental results indicate that differentiated astrocytes and NSCs may be equally permissive to transformation when key genetic alterations are introduced.  In particular, mature astrocytes and neural progenitor cells with EGFR over-expression and p16Ink4a deficiency generate tumours at similar frequencies and intriguingly, the combination of these genetic lesions seem to “re-program” astrocytes and enable them to acquire a “stem-like cell” phenotype (Bachoo et al., 2002).  This result was recently confirmed by Friedmann-Morvinski et al., demonstrating that astrocytes and even mature neurons could be “de-differentiated” and subsequently give rise to malignant gliomas by the transduction of oncogenic lentiviral vectors.  Microarray analysis further revealed that the tumours of astrocytic and neuronal origin match the mesenchymal GBM subtype (Friedmann-Morvinski et al., 2012).  Therefore, all together, these experimental results indicate that glioma may develop from CSCs, progenitors, or the “de-differentiated” mature glial or neuronal cell population.  1.13.2 IDENTIFICATION AND ISOLATION OF BRAIN CANCER STEM CELLS (BCSCs) NEUROSPHERE ASSAY The evidence of the existence of stem-like cells in brain cancer was first reported by Steindler and colleagues who obtained sphere-forming cells from post-operative glioblastoma multiforme (GBM) specimens (Laywell et al., 2002).  Subsequently, Hemmati et al. and Galli et al. grew primary tumour cells isolated from patient specimens in a neurosphere assay with conditions permissive for neural stem cell growth.  The tumourspheres formed expressed NSC markers such as musashi, BMI1 and SOX2 and were capable of aberrant multi-lineage differentiation (Ignatova et al., 2002).  In addition, cells dissociated from these spheres were  47 tumourigenic in animals and formed highly infiltrative lesions characteristic of GBM (Galli et al., 2004; Hemmati et al., 2003).  In neurosphere assays, cells dissociated from primary brain tumour specimens were plated in neurobasal medium supplemented with the growth factors epidermal growth factor (EGF) and basic fibroblast growth factor (b-FGF).  The variably differentiated cells cannot survive and therefore only the neural stem and progenitor cells proliferate clonally to form free-floating spheres in the liquid culture.  Tumourspheres can be serially passaged and cells from the spheres can be induced to undergo lineage differentiation upon growth factor withdrawal or addition of serum (Reynolds and Weiss, 1992).  It should be noted that there is no “one-to-one relationship” between neurospheres and NSCs.  Estimating stem cell frequency based on sphere forming frequency in the neurosphere assay provides an invalid readout (Reynolds and Rietze, 2005).  According to Reynolds et al., sphere-forming frequency approximates progenitor cell activity more closely than stem cell activity.  Most of the spheres in this assay are likely derived from progenitor cells but not from stem cells, therefore serial passaging is crucial to the enrichment of the NSC population (Marshall et al., 2007; Reynolds and Rietze, 2005). CELL SURFACE ANTIGEN SORTING CD133 (PROMININ 1, PROM1)  In the study of Singh et al., brain tumour initiating cells (BTICs) were enriched by a cell sorting paradigm based on the expression of CD133 (prominin 1), a cell surface glycoprotein.  The CD133+ cells exhibit molecular characteristics of NSCs in vitro and are much more tumourigenic than the CD133- cells, with only 100 CD133+ cells required to form tumours that are phenocopies of the parental tumours.  On the contrary, 105 CD133- cells engraft but do not produce tumours in immunocompromised mice (Singh et al., 2004). CD133+ cells have subsequently been found to be resistant to chemo- and radiation therapies.  The CD133+ cell population is enriched in tumours from relapsed patients compared to their initial diagnosis (Liu et al., 2006a).  Transcript levels of drug resistant proteins such as breast cancer resistant proteins (BCRP), MGMT and anti-apoptotic proteins: BCL-2, BCL-xL, MCL-1, and the inhibitor of apoptosis protein (IAP) family are elevated in CD133+ cells which are notably resistant to carboplatin, etoposide, paclitaxel and TMZ (Liu et al., 2006a).  CD133+ cells survive radiotherapy by preferentially activating the DNA repair pathway, and radio-resistance could be reversed by the treatment with CHK1/2 inhibitors (Bao et al., 2006).  Recent studies suggest that CD133- cells may also be tumourigenic and that CD133+ and CD133- cells may represent CSCs from distinct cells-of-origin.  While CD133+ cells  48 resemble fetal NSCs, display proneural signature genes and grow as neurospheres, CD133- cells resemble adult NSCs, show a mesenchymal transcriptional profile and grow semi-adherently (Beier et al., 2007a; Lottaz et al., 2010).  Interestingly, CD133low GBM (having ≤ 3% CD133+ cells in the tumour) show more invasive, proliferative and angiogenic growth than CD133high tumours (having ≥ 3% CD133+ cells in the tumour) (Joo et al., 2008).  CD133- cells are capable of initiating tumour formation and giving rise to tumours that contain CD133+ cells. Interestingly, in vivo passaging of these tumours leads to up-regulation of CD133 expression (Wang et al., 2008).  Despite the uncertainties in the robustness of CD133 as a marker for BCSC identification, experimental evidence has nevertheless lent strong support to the clinical importance of this subset of cells.  The presence of >2% CD133+ cells with high Ki67 labelling is an independent prognostic marker for poor survival in GBM patients (Pallini et al., 2008).  Furthermore, the co-expression of nestin and CD133 independently predicts poor clinical outcome of glioma patients (Zhang et al., 2008). CD15 (LEWIS x, STAGE-SPECIFIC EMBRYONIC ANTIGEN 1, SSEA-1) In addition to CD133, another cell surface antigen, CD15 (SSEA-1), has recently been studied in the isolation of BCSCs.  CD15 is a carbohydrate epitope expressed on normal neutrophils (Skubitz and Snook, 1987), stem and progenitor cells in the adult/embryonic nervous systems, (Allendoerfer et al., 1995; Capela and Temple, 2002; Capela and Temple, 2006) as well as cells from various solid tumours (Fox et al., 1983; McCarthy et al., 1985).  CD15 cell sorting enriches for proliferative and tumourigenic cells in vitro and in vivo in GBM and MB (Read et al., 2009; Son et al., 2009; Ward et al., 2009).  However, the fact that CD15 is also expressed in a subset of granular neuronal progenitors may argue against it being a marker exclusive to NSCs.  Furthermore, in the study of Read et al., CD15+ cells do not form neurospheres and show no evidence of multi-lineage differentiation despite being capable of medulloblastoma tumour formation in animals.  Therefore, CD15+ cells might be progenitor-like cells with a unique capacity for tumour propagation (Read et al., 2009).  1.13.3 ALTERNATIVE METHODS FOR BCSC ENRICHMENT HOECHST 33342 EXCLUSION Stem cells and CSCs express elevated levels of multi-functional drug efflux proteins such as the adenosine triphosphate binding-cassette (ABC) transporters on their cell surface.  These cells may thus be protected from certain chemotherapeutic agents by effluxing the drugs using the ABC transporters (Chaudhary and Roninson, 1991; Dean, 2009; Hu et al., 2008; Hu et  49 al., 2010; Moitra et al., 2011).  Interestingly, Hoechst 33342, a fluorescent DNA binding dye, is also one of the substrates of the ABC drug efflux proteins.  The unique biological property of stem cells and CSCs in extruding chemotherapeutic agents as well as Hoechst 33342 is exploited for the isolation of this sub-population of cells.  In this dye exclusion assay, the cells are stained with Hoechst 33342 and subjected to flow cytometry.  Stem cells and CSCs, due to their enhanced capacity in effluxing the dye, will show low Hoechst staining intensity that segregates these cells from the majority of the Hoechst-stained cells (express lower levels of drug efflux proteins) on a flow cytometry dot plot.  Therefore, these Hoechst-effluxing cells are also known as “side-population” (SP) cells (Alvi et al., 2003; Hirschmann-Jax et al., 2005; Hirschmann-Jax et al., 2004; Ho et al., 2007; Kim et al., 2002; Scharenberg et al., 2002; Szotek et al., 2006).  This technique has been used to successfully identify CSCs in brain cancers (Shen et al., 2008); however it does not provide a pure population of CSCs as normal stem cells are also included and CSCs may not always be in the side population.  In addition, the SP cells may contain differentiated, non stem-like cells that are counted as part of the SP due to an elevated expression of drug efflux proteins (Dean et al., 2005). ALDEHYDE DEHYDROGENASE ACTIVITY ASSAY Aldehyde dehydrogenases are a group of enzymes that catalyze the oxidation of aldehydes.  In 1996, Storms et al. demonstrated the isolation of hematopoietic progenitor cells on the basis of ALDH activity.  In the assay they developed, a fluorescent substrate for ALDH, named BODIPY aminoacetaldehyde (BAAA), was examined for its potential for isolating primitive hematopoietic cells.  Results indicate that a subset of cells with low orthogonal light scattering (lo) and bright fluorescent intensity (br) [SSCloALDHbr] are enriched for hematopoietic progenitor cells (Storms et al., 1999).  Subsequently, the identification of putative NSCs in the central nervous system was reported by Corti et al. who showed that SSCloALDHbr cells obtained from murine adult and embryonic neurospheres cells were capable of self-renewal and multi-potent differentiation in vitro and vivo (Corti et al., 2006).  The enhanced functional activity of ALDH is not only observed in the progenitor cells in the hematopoietic and central nervous system (Christ et al., 2007; Fallon et al., 2003; Hess et al., 2006) but also in human malignancies such as cancers of the lung (Jiang et al., 2009; Liang and Shi, 2011), breast (Charafe-Jauffret et al., 2009), colon (Chu et al., 2009), liver (Ma et al., 2008), brain (Quemener et al., 1990; Rasper et al., 2010), and head and neck squamous cells (Chen et al., 2009), with an association with disease progression (Su et al., 2010), metastasis (Charafe-Jauffret et al., 2010; Marcato et al., 2010; Sun and Wang, 2010) and poor clinical outcome (Charafe-Jauffret et  50 al., 2010; Nalwoga et al., 2010; Yoshioka et al., 2011).  Together, these results indicate the clinical importance of ALDH-high cells in the pathogenesis of cancers.  It is likely that a combination of cell surface markers and other techniques will be required to purify CSCs as elegantly exemplified in the hematopoietic system, though challenges arise because of a minimal overlap in the cell population isolated using different cell surface markers (Al-Hajj et al., 2003; Dalerba et al., 2007; Visvader and Lindeman, 2008).  Moreover, the fact that brain cancer is a complex disease in which a heterogeneous population of cells contribute to disease progression would refute the idea of solely targeting the CSC population for effective treatments (Huse and Holland, 2010).  An effective therapeutic strategy against brain tumours would likely involve targeting most if not all the cells in a tumour. 1.13.4 THERAPEUTIC STRATEGIES TARGETING BCSCs PATHWAY-SPECIFIC SMALL MOLECULE INHIBITORS Cancer is a disease often characterized by aberrant proliferation and abnormal differentiation.  If CSCs were indeed the roots of disease recurrence, it would be reasonable to target aberrant self-renewal pathways for the treatment of brain cancers.  Cyclopamine, an inhibitor of smoothened, provides a   proof-of-principle for targeting SHH pathway in the treatment of MB (Berman et al., 2002; Taipale et al., 2000).  GDC-0449 and LDE-225 are the examples of hedgehog pathway inhibitors that are currently being evaluated in phase I/II clinical trials in solid tumours and MB respectively (Amin et al., 2010; Morrison et al., 2011; Von Hoff et al., 2009).  However, treatment efficacy can be hampered by drug resistance that develops as a result of activating mutations in SMO, chromosomal amplification of GLI2, or cross-talk with the EGFR pathway (Dijkgraaf et al., 2011; Mimeault and Batra, 2010; Yauch et al., 2009).  Research is underway to examine resistant mechanisms in these hedgehog inhibitor-refractory diseases.  In one study, analyses of gene expression signatures indicated a role of the PI3K pathway in mediating drug resistance in MB.  The combinatorial treatment of PI3K inhibitor NVP-BKM120 or the dual PI3K/mTOR (mammalian target of rapamycin) inhibitor NVP-BEZ235 with SMO antagonist markedly delays the development of drug resistance (Buonamici et al., 2010).  In another study, simultaneous inhibition of Notch and SHH pathways eliminates GBM-derived cells much more effectively than monotherapy does (Schreck et al., 2010).  Rational drug combinations targeting multiple developmental pathways may thus offer more significant benefit.  Resveratrol is an inhibitor of the WNT pathway; a recent study by Yang et al. demonstrated the efficacy of resveratrol in suppressing tumourigenicity and enhanced radiosensitivity of primary GBM TICs by inhibiting the STAT3 signaling axis (Yang et al., 2011).   51 DIFFERENTIATION THERAPY Aggressive brain tumours often include a large proportion of immature, primitive cells that appear to be “locked” in a state of de-differentiation, which causes them to proliferate uncontrollably.  An alternative therapeutic strategy is therefore to promote the differentiation of these CSCs.  By treating GBM CSCs with bone morphogenetic protein 4 (BMP4), Piccirillo et al. were able to decrease proliferation and to induce differentiation of the CSCs in vitro.  BMP4 treatment increased tumour latency and prolonged animal survival.  In vivo serial passaging was unachievable due to the depletion of CSCs by BMP4 (Piccirillo et al., 2006).  Consistent with this, BMP7 released from endogenous neural precursor cells can act as paracrine tumour suppressors to inhibit the proliferation, self-renewal and tumour-initiation of stem-like GBM cells (Chirasani et al., 2010).  However, if BMPs were to be used in clinical setting, it would be important to consider the BMP receptor status in patients as Lee et al. noticed a subset of GBM patients were refractory to BMP4 treatment and curiously showed a paradoxical increase in tumour cell proliferation.  These BMP4-resistant cells were subsequently found to express low levels of BMP receptor 1B as a result of epigenetic silencing and could be re-sensitized to BMP4-induced differentiation by ectopic restoration of BMP receptor (Lee et al., 2008c).   Our laboratory has studied the role of a transcription factor named Y-box binding protein 1 (YB-1) in brain cancers.  Silencing YB-1 by siRNA not only decreases cell proliferation, clonogenicity and invasion in vitro but also delays the onset of tumour formation in vivo.  Moreover, YB-1 knock-down enhances the apoptotic effect of temozolomide (TMZ) in adult and pediatric GBM (Gao et al., 2009).  Further investigation revealed that YB-1 is a critical factor regulating the differentiation of NSCs.  YB-1 expression is high in normal NSCs and dramatically decreases when the cells undergo lineage differentiation.  However, in GBM tumours where cells do not undergo a normal process of differentiation, the expression of YB-1 is elevated, corresponding to a higher level of cellular proliferation as evidenced by Ki67 staining.  Silencing YB-1 expression suppresses the expression of the neural stem cell markers SOX2, musashi and BMI1 and intriguingly, forces GBM cells to acquire a more differentiated, astrocytic morphology, accompanied by an increased expression of the differentiation marker GFAP.  Interestingly, YB-1 can be inhibited by BMP4 (Fotovati et al., 2011).  Although thus far there is a lack of small molecule inhibitor targeting YB-1, research is currently underway to develop small molecule inhibitors that target the upstream regulator (RSK, ribosomal S6 kinase) of YB-1.  52 ONCOLYTIC VIRUSES Additional methods have been designed to target CSCs.  With the use of oncolytic herpes simplex virus (oHSV), Wakimoto et al. were able to inhibit the self-renewal and killed GBM-derived CSCs in vitro and in vivo (Wakimoto et al., 2009).  A recent study of novel virus-gene therapy involves oncolytic virus carrying exogenous endostatin-angiostatin fusion gene (vae) which not only infects and kills glioma stem cells but also inhibits the proliferation of human brain microvascular endothelial cells in the CSC niche (Zhu et al., 2011).  The combination of oHSV and PI3K/AKT inhibitors synergistically induces apoptosis of GBM CSCs but not human astrocytes and prolongs survival of animals.  This therapeutic strategy might also effectively target medulloblastoma CSCs that reside in the perivascular niche and exhibit radio-resistance due to enhanced PI3K/AKT signaling (Hambardzumyan et al., 2008).  In another study, Seneca Valley Virus-001 (SVV-001) administered systemically passes the blood-brain-barrier (BBB) and kills the primary xenograft medulloblastoma cells, infects the CD133+ cells and eliminates tumour cells capable of neurosphere formation.  SVV-001 targets both tumourigenic and non-tumourigenic cells and may be well-suited for medulloblastoma treatment (Yu et al., 2011).   Brain cancer stem cells appear to be a Janus-faced entity looking into the past and future of a perverted path of cell development and encompassing characteristics of both.  On-going research into their developmental intricacies and molecular signatures will no doubt help to unravel the biological underpinnings of these cells in brain cancer.  Ultimately, it will lead to the development of precise diagnostic tools and more efficacious, targeted therapeutic agents.  1.14 RATIONALE AND HYPOTHESIS OF THE STUDY There is a growing interest in targeting mitotic kinases for the treatment of cancers. Knowing the pleiotropic roles of PLK1 in regulating cell cycle progression and the utter dependence of malignant cells on this kinase for proliferation and survival, we questioned if the brain cancer stem-like cells and TMZ-resistant GBM cells would be susceptible to the inhibition of PLK1.  Furthermore, we wanted to investigate this oncogenic kinase in SHH MB in pre-clinical studies to explore the possibility of PLK1 molecular targeted therapy in future personalized treatment.  HYPOTHESIS: PLK1 inhibition suppresses the expansion of BTICs and TMZ-resistant GBM cells and may serve as a molecular target for the treatment of SHH MB.   53 CHAPTER 2: POLO-LIKE KINASE 1 (PLK1) INHIBITION KILLS GLIOBLASTOMA MULTIFORME BRAIN TUMOUR CELLS IN PART THROUGH LOSS OF SOX2 AND DELAYS TUMOUR PROGRESSION IN MICE.  2.1 INTRODUCTION Glioblastoma multiforme (GBM) is among the most aggressive human tumours, and despite recent treatment advances, remains refractory to current therapies with few long-term survivors.  There are three molecular subtypes of GBM- proliferative, proneural and mesenchymal according to the molecular characterization of high-grade gliomas by Philips et al. (Phillips et al., 2006).  Patients with the proliferative subtype of the tumours are particularly prone to die from the disease.  Although surgery is the most effective way for debulking tumours, complete resection is often difficult to achieve due to the location and infiltrative nature of GBM.  In addition, adjuvant chemotherapy and radiation therapy often fail to provide long-term disease control, and disease relapses/progression is expected in GBM patients (McKinney, 2004).  The cancer stem cell hypothesis postulates a hierarchical organization in tumours in which only a small subset of cells are capable of tumour formation.  Similar to “stem cells”, tumour initiating cells (TICs) exhibit self-renewal and multi-lineage differentiation capabilities; moreover, TICs form tumours upon serial transplantation in immuno-compromised mice (Al-Hajj et al., 2003; Bonnet and Dick, 1997; O'Brien et al., 2007; Singh et al., 2003).  Brain tumour initiating cells (BTICs) have been found to be resistant to both chemo- and radiation therapies due to elevated expression of drug efflux proteins, enhanced DNA repair activity and evasion of apoptosis (Bao et al., 2006; Blazek et al., 2007; Eramo et al., 2006; Hussein et al., 2011; Liu et al., 2006a).  Therefore, BTICs may play a role in tumour relapse. Our lab has a long-standing interest in identifying molecular targets for the treatment of brain cancers (Faury et al., 2007; Fotovati et al., 2011; Gao et al., 2009).  We have recently performed a large-scale, genome-wide siRNA library screen of 691 human kinases, aiming to identify novel molecular targets for the treatment of childhood malignancies, including pediatric brain cancers.  Results from the siRNA screen indicated that polo-like kinase 1 (PLK1) was one of the lead targets, which when inhibited, resulted in 80-90% growth suppression in 72hrs of a panel of pediatric cancer cells including GBM (Hu et al., 2009).  PLK1 is a serine/threonine kinase that plays important roles in centrosome maturation, bipolar spindle formation (Casenghi et al., 2005; Feng et al., 2006; Lane and Nigg, 1996; Rapley et al., 2005; Sumara et al., 2004; van Vugt et al., 2004b; Yarm, 2002), mitotic entry (Roshak et al., 2000), metaphase-to-anaphase transition (Golan et al., 2002; Hansen et al., 2004; Moshe et al., 2004) and  54 cytokinesis (Neef et al., 2003; Niiya et al., 2006; Seong et al., 2002; Zhou et al., 2003) in M phase of the cell cycle.  This disease-relevant kinase is believed to be a promising therapeutic target for cancer treatments.  The facts that PLK1 is differentially expressed in cancer and normal cells (Ando et al., 2004b; Holtrich et al., 1994; Knecht et al., 1999; Tokumitsu et al., 1999; Wikman et al., 2002; Wolf et al., 2000) and that malignant cells show exclusive dependency on PLK1 for growth and survival (Cogswell et al., 2000; Grinshtein et al., 2011; Liu et al., 2006b; Renner et al., 2009) suggest that a therapeutic window may exist for targeting this protein.  BI2536 is the first-in-class dihydropteridinone derivative, an ATP-competitive inhibitor that exhibits >10,000x specificity for PLK1 compared to a panel of 63 kinases examined and has excellent in vivo anti-tumour activity (Steegmaier et al., 2007).  The cytotoxic effect of PLK1 inhibition can be attributed to its anti-mitotic effects.  Cells treated with PLK1 inhibitor are noted for the dumbbell-like chromatin structure, suggestive of 4N DNA cells arrested in pro-metaphase (Liu and Erikson, 2003).  Disruption of bipolar spindle formation and chronic spindle checkpoint activation ultimately results in mitotic catastrophe (Lenart et al., 2007; Steegmaier et al., 2007).  The remarkable cytotoxic activity of PLK1 inhibitors is a highly desirable feature in cancer treatments. The promising results we have obtained from our initial kinase siRNA library screen prompted us to examine the functional role of PLK1 in brain cancers.  Specifically, we questioned whether PLK1 would be a potential molecular target for the treatment of GBM.  Therefore, in this study we characterized the effects of PLK1 inhibition on adult and pediatric GBM cell lines as well as TICs isolated from patient GBM tumour specimens, as these cells are known to escape current therapy and may account for tumour relapse and recurrence.  Here, we show that biological and pharmacological inhibition of this kinase leads to remarkable growth suppression, accompanied by apoptosis and loss of SOX2 expression in adult and/or pediatric GBM cells.  In particular, treatment with the PLK1 small molecule inhibitor BI2536 impairs the ability of BTICs to self-renew and form tumourspheres in vitro.  Furthermore, compared to the control, mice with intracranial GBM tumour survived significantly longer when treated with the PLK1 inhibitor.  Together, these results suggest the potential therapeutic value of PLK1 inhibitor in the treatment of GBM.   2.2 RESULTS To establish the rationale for targeting PLK1 in GBM, we investigated its expression in primary tumours and more specifically in BTICs.  Initially, we questioned whether PLK1 was  55 highly expressed in primary GBM by evaluating publicly available expression microarray data from 467 cases.  In this cohort, 400 cases were adult and 67 were pediatric GBM.  PLK1 was notably expressed at high levels in the proliferative subclass (Figure 2.1A-B) which has previously been associated with more aggressive disease and poor patient survival (Phillips et al., 2006).  Consistent with these results, PLK1 was co-expressed with several genes associated with cell cycle progression, cytokinesis and DNA replication (Figure 2.1A and Supplementary Table S2.1).  In these patients, PLK1 expression was associated with less favourable overall survival (log rank test, P=0.042, data not shown).  To independently confirm this trend, we analyzed PLK1 mRNA in 343 cases of gliomas from the UCSC database.  PLK1 was up-regulated in 49% of the patients (171/343) where high levels were associated with poor survival (Figure 2.1C, log rank test, P=6.47x10-5).   To investigate this further in patient samples, we obtained patient-derived BTICs from adult and pediatric brain tumour specimens and examined the transcript level of PLK1 by RT-PCR.  Primary adult BTIC lines- GBM4, GBM8, BT74, L0, L1, L2 were well characterized for their stem cell properties in vitro and vivo (Piccirillo et al., 2006; Wakimoto et al., 2009).  Interestingly, PLK1 mRNA was expressed 110-470 times higher than normal human astrocytes.  Likewise, PLK1 was highly expressed in BT241 which was derived from an adult GBM specimen, BT005 which was isolated from a primary pediatric GBM and SF188, which is a cell line established from an 8yr-old child with GBM (Trent et al., 1986) (Figure 2.2A).  Furthermore, we compared PLK1 level in BTICs and normal human neural stem cells (hNSCs) isolated from fetal brains.  PLK1 was found to be 2.37-fold to 10.23-fold more abundant in GBM4, GBM8 and BT74 compared to human neural stem cells isolated from the two different samples (Supplementary Figure S2.2A-B) and its expression was down-regulated significantly (~59-fold to 83-fold) after lineage differentiation, as demonstrated by the low level of PLK1 in human astrocytes (HA) compared to hNSC101 and hNSC167 (Supplementary Figure S2.2C-D).  In order to confirm that PLK1 is expressed in tumourspheres formed by BTICs, we immunostained BT74 with PLK1 along with known NSC markers SOX2 (Brazel et al., 2005; Komitova and Eriksson, 2004), nestin (Fukuda et al., 2003; Gage, 2002) and musashi (Kaneko et al., 2000) and found that they were coordinately expressed (Figure 2.2B).  BT74 tumourspheres were also dissociated into single cells, which were plated in monolayer and immunostained with PLK1, SOX2, musashi and Bmi1 antibodies.  We were able to confirm that PLK1 and these stem cell markers were co-expressed in single cells (Supplementary Figure S2.1). We then questioned whether PLK1 was essential for sustaining the growth of BTICs to provide further support for targeting this kinase in patients.  Primary adult BTIC lines- GBM4,  56 GBM8, BT74, L0, L1 and L2 were treated with the PLK1 small molecule inhibitor BI2536 (5 or 10nM) which is currently in several clinical trials (Schoffski, 2009) for other malignancies but not brain tumours.  PLK1 inhibition blocked tumoursphere formation in all the BTIC lines examined (Figure 2.2C and Supplementary Figure S2.3A).  BT74, which was not very responsive to PLK1 inhibition initially (Figure 2.2C), showed a decrease in secondary and tertiary sphere formation, indicated as P1 and P2, upon serial passaging (Figure 2.2D).  Of note, primary sphere formation was completely abolished in GBM4, GBM8, L0 and L2 because these BTICs were not viable after drug treatment (trypan blue staining, data not shown).  In addition, we obtained a primary pediatric malignant glioma tumour specimen from the British Columbia Children’s Hospital referred to as BT005.  It too expressed PLK1 182 times higher than normal human astrocytes (Figure 2.2A).  These cells were characterized for their stem cell properties by RT-PCR and showed elevated levels of SOX2, musashi and Bmi1 compared to normal human astrocytes (Supplementary Figure S2.3B).  In addition, these cells responded very well to BI2536 in that the drug inhibited tumoursphere growth by >90% (Figure 2.2E).  On the contrary, the cells isolated from primary pediatric GBM tumour BT011, which curiously expressed negligible levels of PLK1 (Supplementary Figure S2.3D), were insensitive to PLK1 inhibition (Supplementary Figure S2.3E), suggesting the importance of target specificity to effectively eliminate these cells.  The control and BI2536-treated BT011 cells were dissociated and re-plated in fresh drug-containing medium every 6 days and no suppression in tumoursphere formation was observed even after the second serial passaging, as shown in Supplementary Figure S2.3E.  Together, these data suggest that children with GBM may benefit from being treated with PLK1 inhibitor; however, the potential for side-effects to normal stem cells needs to be considered.  In this realm, we performed a colony-forming assay on primary hematopoietic stem cells from two pediatric patients and assessed cellular growth and differentiation.  The effect of BI2536 on the formation of multi-lineage colonies from CD34+ bone marrow cells was insignificant at doses that inhibited the cancer stem cells: ie. 5-10nM (Figure 2.2F and Supplementary Figure S2.3F).  To further our understanding of the potential of PLK inhibitors to block the growth of pediatric GBM, we treated SF188 cells with siRNA as well as BI2536.  Following PLK1 knockdown, the growth of SF188 cells was suppressed by ~75-80% in 72hrs (crystal violet stained cells and bar graph in Figure 2.3A; Supplementary Figure S2.4A).  Likewise, BI2536 suppressed the growth of these cells in a dose-dependent manner (crystal violet stained cells and growth curve in Figure 2.3B).  The IC90 after 72hrs was ~5nM (Figure 2.3B).  BI2536 treatment reduced the phosphorylation of CDC25CSer198, a known PLK1 substrate (Figure 2.3B) and this corresponded with a large increase in G2 fraction, suggestive of a G2/M arrest (Figure  57 2.3C).  The loss of PLK1 expression via siRNA or kinase activity by way of BI2536 treatment led to apoptosis as shown by increased Annexin V staining (Figure 2.3D), PARP cleavage and phosphorylation of histone H2AX (Figure 2.3E).  PLK1 inhibition by siRNA or small molecule inhibitor also suppressed the growth of Gli36 (Supplementary Figure S2.4B), another GBM cell line (provided by Dr. David Louis) that expresses abundant human epidermal growth factor receptor (EGFR).  The necessity for the PLK1 pathway was specific for cancer cells as its blockade with siRNA or BI2536 had no effect on the growth of normal human astrocytes (Figure 2.3F). Having demonstrated that PLK1 inhibition suppressed cell growth and induced apoptosis of SF188 cells in monolayer, next we questioned whether it may also affect the expansion of putative cancer stem-like cells in neurosphere cultures.  Our preliminary results indicated that SF188 cells plated at low cell seeding density were capable of proliferating clonally to form tumourspheres >50µm within one week in neurosphere culture condition (data not shown).  This suggests to us that there might exist a population of cancer stem-like cells in this cell line.  Similar to the results we obtained from the patient-derived primary BTICs, the putative cancer stem-like cells in SF188 were susceptible to PLK1 inhibition as the tumoursphere formation was significantly inhibited upon BI2536 or siRNA treatment.  The inhibitor and siRNA reduced the total number as well as the size of the tumourspheres formed (Figure 2.4A and Supplementary Figure S2.3C).  Consistent with the inhibition of stem cell properties, PLK1 knockdown (Figure 2.4B) or BI2536 treatment (Figure 2.4C) reduced the mRNA and protein expression of SOX2 and musashi.  In addition, PLK1 knockdown with two different PLK1-targeting siRNAs consistently decreased SOX2 expression (Supplementary Figure S2.5A).  Furthermore, corresponding to the reduction in the expression of stem cell markers, SF188 cells that survived 6 days after siRNA or BI2536 treatment intriguingly underwent dramatic cellular morphological changes that rendered these cells to assume a stellate appearance, bearing structural similarity to astrocytes (Figure 2.4D and Supplementary Figure S2.5B for additional images) and this was accompanied by an increase in the transcript level of glial fibrillary acidic protein (GFAP) (Figure 2.4E).  Next, we questioned whether the loss of cell proliferation, survival and stem-like properties of PLK1-inhibited cells was associated with the decrease in SOX2.  Therefore, we silenced the expression of SOX2 and examined the biological effects on GBM cells.  SOX2 knockdown reduced cell growth (Figure 2.4F) and induced cell death (Figure 2.4G) in SF188 cells in 72hrs.  Phosphorylation of H2AX, an early marker of apoptosis, was observed at 48hrs (Figure 2.4H) while cleavage of caspase 3 (Figure 2.4I) occurred at 72hrs in siSOX2 cells.   Finally, we tested the in vivo efficacy of BI2536 in mice against the well-established  58 U251 model that was previously characterized by our group (Verreault et al., 2011).  First we conducted a time-course study to establish whether or not U251 (adult GBM) cells were sensitive to PLK1 inhibition. BI2536 blocked U251 growth in a dose-dependent manner after 72 hrs in vitro (Figure 2.5A).  To further support these results, PLK1 was inhibited with siRNA where the loss of PLK1 expression led to 90% growth suppression (Figure 2.5B).  This was correlated with a loss of SOX2 expression (Figure 2.5C).  We therefore established that U251 cells were sensitive to PLK1 inhibition with siRNA as well as BI2536, warranting further examination in mice. U251 cells were introduced via intracranial injection and tumours were allowed to form before the animals were randomized into three groups (control, vehicle-treated or BI2536-treated).  The test group of mice were treated with BI2536 (50mg/kg) once weekly for four weeks.  BI2536 prolonged the survival of animals compared to those that were untreated or given vehicle control agent (Figure 2.5D).  Health status and body weight of the mice were monitored throughout the study.  The PLK1 inhibitor did not cause adverse side-effects to the animals as they appeared healthy during the course of the treatment (day 21-42) and maintained normal body weight (Figure 2.5E).  We concluded that BI2536 was well-tolerated and that it had the desired effect of significantly delaying brain tumour progression.   2.3 DISCUSSION In this study, we conducted a comprehensive investigation on the potential of PLK1 inhibitors in GBM.  We approached this by integrating microarray data from 467 patients and identified PLK1 to be over-expressed in the proliferative subtype of GBM, which is associated with poor prognosis (Phillips et al., 2006).  Subsequently, we examined the effect of PLK1 inhibition on GBM in vitro and in vivo.  To address this question, we have used multiple models, including cell lines (SF188, U251, Gli36 and immortalized human astrocytes HA), cancer stem-like cells derived from patients (GBM4, GBM8, BT74, L0, L1 and L2), primary cells isolated from adult and pediatric GBM specimens (BT241, BT005 and BT011) and hematopoietic stem cells derived from 2 patients.  We reported that pharmacological or genetic inhibition of PLK1 remarkably suppressed growth, induced cell cycle arrest and apoptosis in these aggressive brain tumour cells and shut down the growth of BTICs.  These changes were associated with a loss of SOX2, the inhibition of which partially phenocopied the growth suppression and apoptosis observed following PLK1 suppression.  SOX2 [(sex determining region Y)-box 2] is a transcription factor that is required for the self-renewal of embryonic stem cells (Gubbay et al., 1990) as well as primitive neural cells (Bylund et al., 2003; Graham et al., 2003; Gubbay et al., 1990; Pevny and Lovell-Badge, 1997;  59 Wood and Episkopou, 1999).  Recent experimental evidence indicates that SOX2 is not only a marker of neural stem cells but is also functionally involved in the pathogenesis of brain cancers (Alonso et al., 2011; Annovazzi et al., 2011; Schmitz et al., 2007; Sun and Zhang, 2011).  In addition, SOX2 was demonstrated to confer proliferative advantage and to maintain tumourigenicity of glioma-intiating cells (Gangemi et al., 2009; Ikushima et al., 2011a).  Together these results led us to question whether the growth suppression and apoptosis from PLK1 inhibition would be mediated in part through the down-regulation of SOX2.  For the first time, we showed that SOX2 expression is regulated by PLK1.  In addition, SOX2 knockdown partially phenocopied the effect of PLK1 inhibition; however the magnitude of growth suppression and cell death was not as significant as what was observed in PLK1 inhibition and this is likely due to the pivotal and pleiotropic roles that PLK1 plays in mitosis (Feng et al., 2006; Hansen et al., 2004; Lane and Nigg, 1996; Moshe et al., 2004; Rapley et al., 2005; Roshak et al., 2000; Seong et al., 2002; van Vugt et al., 2004b; Yarm, 2002; Zhou et al., 2003).  However, these results have nevertheless provided a rational and probable explanation for the loss of proliferative capacity and stemness properties of BTICs treated with BI2536 in tumoursphere assays.  Research is currently underway to understand the mechanism by which PLK1 modulates SOX2 expression, the link between these two proteins and their functional cooperation in regulating the proliferation and survival of GBM cells.  Finally, results from our in vivo xenograft experiment indicated that systemic administration of BI2536 prolonged the survival of animals with orthotopic GBM tumours.  Together these results suggest the potential of targeting PLK1 in brain cancer treatments. PLK1 is an emerging new molecular target for therapy.  This is recognized by the fact that several PLK1 inhibitors have been developed and are currently under investigations in phase I or II clinical trials.  BI2536 passed phase I clinical trials and demonstrated measurable anti-tumour activity with favourable toxicity profiles in advanced solid tumours (Hofheinz et al., 2010; Mross et al., 2008).  BI6727 is a second-in-class, ATP-competitive PLK1 inhibitor tested in non-Hodgkin’s lymphoma, acute myeloid leukemia, small cell lung cancer and solid tumours (Lens et al., 2010).  The results from the phase I trial are promising showing disease stabilization in 32% of the patients with advanced or metastatic solid tumours (Schoffski, 2009).  Additional PLK1 inhibitors such as GSK461364 and ON01910 have been evaluated in phase I clinical trials and demonstrated therapeutic efficacy in patients with advanced solid tumours (Jimeno et al., 2008; Olmos et al., 2011).  PLK1 inhibitors may be an alternative treatment to tumours refractory to standard anti-mitotic agents due to mutations in tubulin proteins (Hari et al., 2006) or over-expression of drug efflux proteins such as p-glycoproteins (Greenberger et al.,  60 1988).  In addition, unlike standard anti-mitotic agents such as vinca alkaloids and taxanes, PLK1 inhibition does not cause chemotherapy-induced peripheral neuropathy (CIPN), a pathological condition associated with motor and sensory deficits.  The major side-effect and dose limiting toxicity of PLK1 inhibitors is neutropenia (Mross et al., 2008) which is reversible and clinically manageable.  Our finding has shown that at the concentration of 10nM of BI2536, where an inhibition of >70% of tumour cell growth was achieved, no significant inhibitory effect was observed on hematopoietic colony formation.  In addition to testing PLK1 inhibitor on hematopoietic stem cells, we believe that it is equally important to evaluate the effects of BI2536 on hNSCs in order to assess the safety of the inhibitor.  In our study, we have shown that PLK1 transcripts were 2.37-10.23 times more abundant in the BTICs compared to normal human NSCs.  This result suggests to us that there may exist a therapeutic window for the clinical use of PLK1 inhibitors at a dose range that would show effective anti-tumour activity with minimal hematopoietic and neuro- toxicities.  However, additional correlative and toxicity studies in animal models and early phase clinical trials are needed to confirm this observation.  Intriguingly, pediatric GBM cells SF188 treated with PLK1 siRNA or inhibitor may have undergone differentiation as evidenced by a dramatic alteration in cellular morphology and increased level of GFAP.  It is believed that cancers may arise from aberrations in the lineage differentiation of stem and progenitor cells due to genetic or epigenetic abnormalities; therefore, these cells are trapped in a perpetual state of self-renewal and proliferation that prevent them from becoming terminally differentiated.  In this study we demonstrated that PLK1 inhibition not only induced apoptosis of pediatric GBM cells but may have also forced these cells down the differentiation path that will decelerate their growth and possibly render them sensitive to conventional anti-cancer therapies.  Given the well-known difficulty in treating glial tumours, our results argued for the possibility of treating this patient population with PLK1 inhibitors.  This could therefore shed light on the treatment of GBM, a rare type of brain tumour found in children where new therapies are desperately needed because unlike adults patients, children afflicted with malignant glioma do not respond well to Temozolomide (TMZ) (Barone et al., 2006; Bartels et al., 2011; Lashford et al., 2002), a front-line therapy for this disease.  Disease relapse is one of the major road-blocks to successful treatments in brain cancers.  BTICs reportedly express elevated levels of MGMT, BCL-2 and BCL-xL transcripts (Liu et al., 2006a) and have an enhanced ability to repair DNA following radiation (Bao et al., 2006).  Given these features, it is not surprising that BTICs survive radiation and chemotherapy.  The fact that BTICs are resistant to conventional anti-cancer therapies and are enriched in tumours of relapsed patients suggests that they play a clinically important role in disease  61 recurrence (Dirks, 2006).  We demonstrated that PLK1 inhibition not only affected the cell division of rapidly proliferating SF188 cells but also the self-renewal of putative cancer stem-like cells from patient samples.  BTIC lines GBM4, GBM8, L0 and L2 could not be serially passaged because the cells died upon BI2536 treatment.  In addition, even those that were capable of primary sphere formation such as BT74, self-renewal was greatly impaired when BI2536 treatment was continued.  Recently, BT74 cells have been shown to be resistant to chemotherapy and molecular-targeting agents (Cheema et al., 2011; Kanai et al., 2011).  Thus, our results are particularly exciting as it suggests that PLK1 inhibitor may potentially eliminate the cells that make up the bulk of the tumours as well as deplete the notoriously chemotherapy and radiation-resistant BTICs.  Our results are consistent with a study recently published by Grinshtein et al. who demonstrated that PLK1 inhibition suppressed the expansion of TICs isolated from bone marrow metastases of neuroblastoma patients (Grinshtein et al., 2011).  As a result, PLK1 is a unique molecular target because its inhibition kills a range of cancer cells. To conclude, we have identified PLK1 as a factor critical to the survival of brain cancer cells and BTICs.  Thus, PLK1 is positioned as a promising molecular target that has the capacity to overcome therapy resistance imposed by BTICs.  2.4 MATERIALS AND METHODS Bioinformatic Analysis Publicly available gene expression data for a series of pediatric (Paugh et al., 2010) and adult (Lee et al., 2008c) primary GBM samples profiled on Affymetrix U133 platforms were downloaded and integrated in R 2.11 (http://www.r-project.org/).  CEL files were read using the Affymetrix package of BioConductor 2.6, and normalized for each study using the robust multichip average (rma) algorithm (Irizarry et al., 2003).  Common probesets across the respective studies and platforms were extracted, and reduced from probeset to gene-level data (taking the maximum value across all samples), with unannotated probesets excluded.  The data were row (gene)- and column (sample)- centred (Lusa et al., 2007; Perou et al., 2010), giving a working dataset of 467 samples and 13,169 unique known genes.  Samples were clustered into “proneural”, “proliferative” or “mesenchymal” subclasses based upon the relative expression of 323 overlapping genes from the Phillips 2006 classifier (Phillips et al., 2006) using Ward’s hierarchical clustering.  PLK1 levels were investigated for subclass-specific expression correlations using ANOVA (Analysis of Variance), and visualized by boxplots.  Association with overall survival was performed using the log-rank test.  Highly correlated genes were determined by calculating Pearson’s correlation coefficients.   62  Cell Culture  SF188 and U251 cells were obtained from the American Tissue Culture Collection (ATCC).  GBM4, GBM8 and BT74 primary BTICs were characterized by Wakimoto et al. describing the ability of these cells to form tumourspheres upon serial passaging in vitro and secondary tumours in vivo.  Furthermore, the cells are capable of multi-lineage differentiation upon induction (Pandita et al., 2004; Wakimoto et al., 2009).  L0, L1 and L2 primary BTICs were isolated from GBM patients as previously described (Piccirillo et al., 2006).  The cells grow as tumourspheres, express markers of NSCs and retain high self-renewal and differentiation potential.  Similar methods were used to isolate primary GBM BTICs BT241 (Singh et al., 2003).  Primary brain tumour cells were isolated from BT005 and BT011 according to the protocol previously described by us (Lenkiewicz et al., 2009).  The cDNA of human neural stem cells (hNSC101 and hNSC167) was synthesized using RNA extracted from fetal brains (no genetic defects detected) obtained from two abortion cases.  All the primary cells and BTICs were obtained through informed consent in abidance with the respective Institutional Review Board guidelines.   Transfection and Immunofluorescence Staining  PLK1 transfections were performed using Lipofectamine RNAiMAX [Invitrogen] as previously described (Hu et al., 2009) using control oligo (sequence: UUC UCC GAA CGU GUC ACG U, Qiagen) and PLK1 siRNA oligo (siRNA#1 sense sequence: CGG GCA AGA UUG UGC CUA A dTdT.  siRNA#2 sense sequence: ACG GCA GCG UGC AGA UCA A dTdT, Dharmacon).  SOX2 siRNA was purchased from Sigma (SASI_Hs01_00050572).  Primary antibodies used for Western blotting studies include anti-PLK1 [Sigma-Aldrich], anti-P-H2AXS139 antibody [Abcam], anti-caspase 3 [Cell Signaling Technology], anti-poly(ADP-ribose) polymerase [Cell Signaling Technology], anti-P-CDC25CSer198 [Cell Signaling Technology], anti-SOX2 [Millipore and Cell Signaling Technology], anti-musashi [Abcam] and anti-Bmi1 [Abcam], anti-vinculin [Upstate], anti-tubuin [Cell Signaling Technology] and anti-pan-actin [Cell Signaling Technology].  Immunofluorescence staining was performed on the BT74 tumourspheres according to the procedure previously described by us (Fotovati et al., 2011).  To examine PLK1 and neural stem cell expression in BT74 single cells, the tumourspheres were dissociated using solution A, B and C in the Neurocult Chemical Dissociation Kit (STEMCELL Technologies), fixed in 1:1 mixture of cold acetone and methanol on microscope slides and immediately incubated in -20oC for 20 minutes which allowed attachment of the cells to the glass slides.  The cells were  63 subsequently washed 2x with PBS and incubated with PLK1 [Sigma, 1:300], SOX2 [Cell Signaling Technology, 1:100], musashi [Abcam, 1:100] and Bmi1 [Abcam, 1:100] antibodies overnight at 4oC.  The following day, the cells were washed 3x with PBS and incubated with mouse and rabbit secondary antibodies conjugated to Alexa Fluor 488 and Alexa Fluor 546 at room temperature for 1hr.  After the secondary antibody incubation, the cells were washed 3 times, 5 minutes each and Hoechst dye (diluted to 1µg/ml) was added before the cells were visualized under the confocal microscope Olympus Fluoview FV10i.  Real-Time Quantitative Reverse-Transcription PCR  RNA was extracted from the cells using the Qiagen RNeasy Kit following the manufacturer’s protocol. Synthesis of cDNA and real-time PCR experiments were conducted using FAM-labeled Taqman Assay-on-Demand probes according to the method previously described by us (Wu et al., 2006).  TATA-box binding protein (TBP) or 18s mRNA was used as house keeping genes for data normalization.   Tumoursphere Assay  SF188, patient-derived BTICs: GBM4, GBM8, BT74, L0, L1, L2, primary brain tumour cells: BT241, BT005 and BT011 were dissociated into single cells and plated in neurobasal medium (1x104 cells per well in 6-well plates) supplemented with 20ng/ml human recombinant EGF, 20ng/ml human recombinant basic FGF [Stem Cell Technology] and 2µg/ml heparin on ultra low-attachment, coated culture plates [Corning].  Tumourspheres >50µm were quantified and photomicrographs were taken 6 days after culture.   Colony Forming Cell Assay Reagents specific for this assay were purchased from STEMCELL Technologies and the assay was performed according to manufacturer’s instructions.  Briefly, normal CD34+ cells were obtained from residual cells of bone marrow transplant and used following informed consent.  BI2536 was diluted (0.1, 1, 2, 10 and 100nM) and added to six separate tubes of MethoCult [STEMCELL Technologies], a methylcellulose matrix containing recombinant human cytokines stem cell factor, granulocyte macrophage colony-stimulating factor, interleukin-3, granulocyte colony stimulating factor and erythropoietin.  Following the addition of CD34+ cells at a final concentration of 5x103 cells per tube, the mixtures were vortexed and allowed to stand for 5 minutes.  DMSO was used as vehicle control for BI2536.  MethoCult mixtures were then dispensed into 35mm dishes using 5ml syringes and blunt end needles at a volume of 1.1 ml  64 per dish.  The medium was evenly distributed across the surface of each dish by gentle tilting and rotation.  The dishes were then placed in a 150mm tissue culture dish along with a 35 mm dish containing sterile water to maintain humidity.  The tissue culture dish was placed in a 37oC humidified incubator containing 5% CO2 for 12 days.  The number of myeloid and erythroid derived colonies in both the treated and control dishes were counted as described by the assay protocol.   Cell Cycle Analysis Cells were harvested by trypsinization, washed once with cold PBS and fixed in 70% ethanol overnight.  The cells were washed once with cold PBS prior to the addition of staining buffer which was composed of 40µg/ml propidium iodide and 200µg/ml RNase A in cold PBS.  The cells were incubated at room temperature, in the dark, for 30 minutes and 100µl of cold PBS was added directly to the cell suspension when the cells were ready to be analyzed by flow cytometry.   Annexin V Staining and Quantification of Cell Growth by Hoechst Staining  SF188 cells were treated with 5nM BI2536 for 48hrs and stained with Annexin V [Promega] as previously described (Lee et al., 2008a).  To evaluate the effect of PLK1 inhibition on cell growth, SF188 cells were plated (5,000 cell/well) in 96-well plates, treated with BI2536 or PLK1 siRNA for 72hrs and stained with Hoechst dye (1µg/ml) in 100µl of PBS containing 2% paraformaldehyde.  The stained cells were kept at room temperature, in the dark, on a rocking platform for 30 minutes.  The plates were analyzed and the images were taken on the ArrayScan VTI Reader [Cellomics].   BI2536 Efficacy Test in Orthotopic Xenograft Model U251 human glioma cells were obtained from ATCC.  Intracranial tumours were established using the protocol we previously published (Verreault et al., 2011).  In brief, on day 0, 7.5 x104 cells were implanted intracerebrally using a stereotaxic injection frame.  After the orthotopic tumours established, BI2536 (50mg/ml diluted in 0.1N HCl) was delivered intravenously into the Rag2M mice (7-10 weeks old, 8 mice per treatment group) once a week for four weeks.  The health status and body weight of animals were monitored closely during the course of the experiment in accordance with the protocol approved by the British Columbia Cancer Research Centre. The mice were observed until they displayed obvious signs of neurological deficits and appeared unwell.   65  Statistical Analysis All quantitative data presented were analyzed as mean value ±	  standard error.  For the microarray and animal studies, log-rank analysis was performed on the Kaplan Meier curve to determine statistical significance of the results.  The number of samples used and the respective P-values are listed in the figure legends.  The level of significance for the in vitro cell growth/death data was determined by Student’s two-tailed T-test (*P-value<0.05; **P-value<0.01).      66 2.5 FIGURES                                 Fig. 1. PLK1 is highly expressed in primary GBM where it is associated with the proliferative?subtype and poor survival. ?C. 0.00?0.05?0.10?0.15?0.20?0.25?0.30?0.35?0.40?0.45?0.50?0.55?0.60?0.65?0.70?0.75?0.80?0.85?0.90?0.95?1.00?0? 500? 1000? 1500? 2000? 2500? 3000? 3500? 4000? 4500? 5000? 5500?6000? 6500?7000?7500?Probability of Survival?Days in Study?All Glioma?PLK1 High?PLK1 Intermediate?PLK1 Low?P=6.47x10-5 ?n=343?B. Proneural?Proliferative?Mesenchymal?Relative Difference in PLK1  Transcript Level (fold) A. Proneural? Proliferative? Mesenchymal?P<0.001? 67 Figure 2.1 PLK1 is highly expressed in primary GBM where it is associated with the proliferative subtype and poor survival. (A) Heatmap shows the relative expression values for the most highly correlated genes with PLK1 in 467 pediatric and adult GBM, classified into subgroups according to the schema of Phillips et al. 2006 (Phillips et al., 2006) (red = high expression, blue = low expression). Gene expression patterns across GBM subtypes indicates that PLK1 is co-expressed along with several genes involved in cell proliferation and division. Genes (horizontal axis) are listed in descending order of Pearson’s correlation coefficient. (B) Boxplots (dark red = proliferative subgroup, dark blue = proneural subgroup, dark green = mesenchymal subgroup) show mRNA expression values for PLK1 in 467 pediatric and adult GBM classified into subgroups according to the schema of Phillips et al. 2006 (Phillips et al., 2006). PLK1 is significantly over-expressed in the proliferative subclass as compared with the other two [P<0.001, ANOVA (Analysis of Variance)]. (C) Kaplan-Meier curve shows the overall survival of patients with glioma (blue line) from the USCS online database for which clinical outcome data was available (n=343). PLK1 is highly expressed in 171/343 (49%) of the cases. Patients with high expression (red line) have significantly shorter survival then those with intermediate (yellow line) or low (green line) level of PLK1 (log-rank test, P=6.7x10-5).     68     A. 203?191?437?113?138? 130?369?182?470?1?Adult GBM?Pediatric GBM?0?100?200?300?400?500?HA?GBM4?GBM8?BT74?GBM-L0?GBM-L1?GBM-L2?BT241?BT005?SF188?Relative PLK1  Transcript Level? (fold difference)?0?10?20?30?40?50?DMSO? 1 nM? 2 nM? 10 nM?100 nM?Total Number of Colonies?Total?Myeloid?Erythroid?BI2536?F. DAPI?PLK1?Musashi? Merged, All?     PLK1?Musashi  ?PLK1?SOX2? SOX2?  PLK1? DAPI? Merged, All?Nestin? PLK1?Nestin?PLK1? DAPI? Merged, All?B. ? 1     ? 2     ?10     ?1 0 (nM)    ?0?50?100?150?200?DMSO? 10nM BI2536?Total Number of Tumourspheres ?(per well)?E. DMSO? BI2536?P0?P1?C.  ?   (nM)?GBM4?GBM8?DMSO? ?5? 10?BT74?BI2536 ?0?20?40?60?80?100?DMSO?BI2536-5nM?BI2536-10nM?** ** ** ** * GBM4?GBM8? BT74?Total Number of Tumourspheres?(per well)?D. P1? P2?DMSO?BI2536 ?P0?* * 0?10?20?30?40?50?60?70?80?90?P0? P1? P2?Total Number of Tumourspheres? (per well)? DMSO?BI2536(10nM)? 69 Figure 2.2 PLK1 is over-expressed in BTICs and its inhibition suppresses the self-renewal of these cells. (A) Real-time PCR was performed on the transcripts isolated from a panel of BTIC cultures- GBM4, GBM8, BT74, L0, L1, L2, primary brain cancer cells- BT241 and BT005 (isolated from adult and pediatric malignant glioma, respectively), brain cancer cell line- SF188, and human astrocyte (HA) cell line to compare the level of PLK1. PLK1 transcript is between 110-470 fold more abundant in primary BTICs and cancer cell line than in the immortalized human astrocytes.  The experiment was performed in triplicates on two separate occasions. (B) Whole-mount immunofluorescence staining of BT74 tumourspheres was performed to characterize PLK1 along with neural stem cell markers nestin, SOX2, and musashi. Z-stack images (60x magnification) of the tumourspheres were taken on Fluoview FV10i (Olympus, Japan).  The IF study was repeated by two individuals on two separate occasions. (C) Primary BTICs: GBM4, GBM8 and BT74 (1x104 cells per well in 6-well plates) were treated with 5 or 10nM BI2536 for 6 days. The total number of tumourspheres (>50µm) in each well was counted and photomicrographs of the spheres were taken (scale bar=500µm). The treatments were performed in duplicates on three separate occasions. (D) Self-renewal capability of BT74 was examined in tumoursphere assays. BT74 cells (1x104 cells per well in 6-well plates) were treated with 10nM of BI2536 for 6 days. The tumourspheres formed were dissociated and re-plated in fresh medium containing the inhibitor for additional 6 days. The procedure was repeated until re-plating was unachievable due to low cell number. P0, P1 and P2 indicate primary, secondary and tertiary tumoursphere formation, respectively. Scale bar=500µm. (E) Single cells were isolated from post-surgical pediatric malignant glioma BT005 according to the protocol we previously established (Singh et al., 2003) and cultured in neurosphere-supportive culture condition which allowed the cells to form tumourspheres in a week. The cells were subsequently dissociated and treated with 10nM BI2536. The number of tumourspheres was enumerated and photomicrographs (scale bar=200µm) were taken after 6 days in culture.  The experiment was performed in duplicates on one occasion due to the scarcity of the material. (F) Analysis of the effect of BI2536 on in vitro hematopoietic colony formation was performed on normal bone marrow derived CD34+ cells isolated from one patient. The cells were incubated with DMSO or increasing concentrations of BI2536 in methylcellulose cultures that contained cytokines to stimulate hematopoiesis. After 12 days in culture, myeloid and erythroid colonies were enumerated by counting under an inverted microscope based on morphology. Data presented here is from one of the two independent experiments.    70   Figure 2.3 PLK1 inhibition suppresses cell growth and induces apoptosis in brain cancer cells SF188. (A) SF188 cells were treated with 5nM of PLK1 siRNA for 72hrs, stained with Hoechst and quantified. The growth of the cells was suppressed by ~75-80% after PLK1 siRNA treatment, a result confirmed by crystal violet staining of the cells. PLK1 knockdown by siRNA decreased its Fig. 3. PLK1 inhibition suppresses cell growth and induces apoptosis in brain cancer ?cells SF188. ?C. DMSO? BI2536?%G1=51.1?%S=33.7?%G2=11.7?%G1=11.6?%S=44.1?%G2=35.3?F. 0.00?0.20?0.40?0.60?0.80?1.00?1.20?!"#$%"&'"&()"'*(+,-./0#1'*(+,-./.2#1'*(+,-./32#1'4156'78309:/3;0#1'78309:/0#1'78309:/.2#1'5    0   20?PLK1 siRNA?Ctl?  2.5   5   0 (nM)?BI2536?Ctl?Relative Cell Growth (fold change)?D. Control ? siPLK1? DMSO? BI2536?7.49?13.7? 28.3 ?5.34?12.6 ?17.3 ?30 ?49.7?7AAD?Annexin V-PE?PLK1?Vinculin?Control ?siPLK1?A. Control ?  ?siPLK1?0?0.2?0.4?0.6?0.8?1?1.2?Control ?siPLK1?* Relative Cell Growth? (fold change)?DMSO?BI2536 ?PARP ?PARP (cleaved)?Control ?siPLK1?P-H2AXSer139?Actin?E. P-CDC25CSer198?Actin?DMSO?1? 2.5?5?BI2536 (nM)?DMSO?BI2536?0?1?2?3?4?5?6?7?8?0h? 24h? 48h? 72h?Relative Cell Growth ?(fold change)?* * * DMSO?0.5nM?1nM?5nM?10nM?50nM?100nM?BI2536?B. (Time)? 71 protein level, as shown by immunoblotting.  The Western blotting was repeated on three separate occasions. (B) SF188 cells were treated with BI2536 (0.5-100nM) for 24, 48 and 72hrs. At the end of each time-point, the cells were stained with Hoechst and quantified. BI2536 inhibited the growth of the cells in 72hrs. The cells were additionally stained with crystal violet at 72hrs and growth suppression was confirmed. The cell growth assay was performed in triplicates on two separate occasions. (C) SF188 cells were treated with 5nM BI2536 for 24hrs, fixed in cold 70% ethanol, stained with propidium iodide and subjected to flow cytometry for analysis of cell cycle profile. G2/M cell cycle arrest was observed after PLK1 inhibition. The study was repeated on two separate occasions. (D) SF188 cells were treated with 5nM of BI2536 for 48hrs. Apoptosis was measured by Annexin V-PE/7AAD staining. Induction of apoptosis in PLK1-inhibited cells was demonstrated by an increase in the percentage of Annexin V-PE-positive cells. This experiment was repeated twice. (E) SF188 cells were treated with 5nM of PLK1 siRNA or inhibitor and the total proteins were extracted for immunoblotting. Apoptosis was confirmed at the molecular level by PARP cleavage and phosphorylation of H2AX at Ser139. Represented result (n=3 for the study) is presented. (F) Immortalized human astrocytes HA were treated with either PLK1 siRNA (5-20nM) or BI2536 (2.5-10nM) and cell growth in monolayer was assessed 72hrs after treatments.      72  Figure 2.4 PLK1 inhibition down-regulates the expression of SOX2, which is required for the growth and survival of GBM cells. (A) SF188 cells (1x104 cells per well in 6-well plates) were plated in neurobasal growth medium containing 5 or 10nM BI2536 for 6 days. The total number of tumourspheres (>50µm) in each well was counted and photomicrographs of the spheres were taken (scale bar=500µm). The treatments were performed in duplicates on three separate occasions. (B) The transcript and A. DMSO? 5??10     (nM)?020?40?60?80?100?120?DMSO?Total Number of Tumourspheres ?(per well) ? ?5       10 (nM)?BI2536?* * BI2536?Fig. 4. PLK1 inhibition down-regulates the expression of SOX2, which is required for the?growth and survival of GBM cells. ?D. Control? siPLK1?DMSO? BI2536?SOX2 P-H2AXSer139 Tubulin Control  siSOX2  48hrs H. SOX2 Tubulin 72hrs Control  siSOX2  Cleaved caspase 3 Caspase 3 I. 0?0.2?0.4?0.6?0.8?1?1.2?Sox2? Musashi? Bmi1?DMSO?BI2536?* * SOX2?SOX2?Musashi?Bmi1?Tubulin?DMSO ?BI2536??C. Relative Transcript Level?(fold difference)?Musashi?Bmi1?B. Control?siPLK1?Relative Transcript Level?(fold difference)?SOX2?Actin?Control ?siPLK1??Musashi?Bmi1?PLK1?** ** * 0?0.2?0.4?0.6?0.8?1?1.2?Sox2? Musashi? Bmi1?Musashi?Bmi1?SOX2?E. 0?0.5?1?1.5?2?2.5?3?3.5?PLK1? Sox2? GFAP?Control oligo?siPLK1#1?siPLK1#2?** ** *   *   *   **   OX2?Relative Transcript Level?(fold difference)?F. 0?0.5?1?1.5?Control?siSox2 ?* si OX2?Relative Cell Growth?(fold change)?0?1?2?3?4?5?Control ?siSox2?* si OX2?G. Relative Cell Death?(fold change)? 73 protein expression of neural stem cell markers SOX2, musashi and Bmi1 were measured by RT-PCR and immunoblotting 36hrs and 48hrs respectively PLK1 siRNA treatment in SF188 cells. The immunoblotting was repeated three times. The RT-PCR experiments were performed in triplicates on two separate occasions. (C) The transcript and protein expression of SOX2, musashi and Bmi1 were measured by RT-PCR and immunoblotting 36hrs and 48hrs respectively after BI2536 treatment in SF188 cells. The immunoblotting was repeated three times. The RT-PCR experiments were performed in triplicates on two separate occasions.  (D) SF188 cells were treated with 5nM of PLK1 siRNA or BI2536 for 6 days and photomicrographs were taken on the cells that remained after the treatment. Representative photomicrographs of the cells that underwent dramatic cellular morphological alterations are shown. Scale bar=280µm. The study was repeated on two separate occasions. (E) Total RNA from the cells treated with 5nM PLK1 siRNA #1 and #2 for 36hrs was extracted and subjected to RT-PCR to quantify the transcripts of PLK1, SOX2 and GFAP. (F) SF188 cells were treated with 100nM SOX2 siRNA for 72hrs. The cells were stained with Hoechst and quantified. The number of viable cells was enumerated and the relative fold difference in cell growth is shown in the bar graph. The experiment was done in triplicates and repeated three times. (G) The number of non-viable cells after 100nM, 72hrs siSOX2 treatment was enumerated based on enhanced Hoechst staining due to chromatin condensation and the relative fold difference in cell death is shown in the bar chart. The experiment was done in triplicates and repeated three times. (H) Proteins extracted from the cells treated with 100nM control or SOX2 siRNA for 48hrs were subjected to immunoblotting to examine the phosphoryation of H2AX at Ser139, which is a marker of early apoptosis. (I) SF188 cells were treated with 100nM control or SOX2 siRNA for 72hrs and immunoblotting was performed on the total protein lysates. Increased caspase 3 cleavage was observed in the siSOX2-treated cells compared to the control cells. The Western blotting for Figure (H) and (I) was repeated twice.     74   Figure 2.5 BI2536 suppresses the growth of GBM cell line U251 in vitro and tumour formation in vivo. (A) U251 cells were treated with increasing concentrations of BI2536 (0.5-100nM) in a 72-hour time-course study. BI2536 (5nM) inhibited the growth of U251 cells in 72hrs. The experiment was performed in triplicates on two separate occasions. (B) U251 cells were treated with 5nM PLK1 siRNA for 72hrs, stained with Hoechst and quantified. PLK1 knockdown decreased U251 Fig. 5. BI2536 suppresses the growth of GBM cell line U251 in vitro and tumour formation ?in vivo. ?A. 0?1?2?3?4?5?6?7?0h? 24h? 48h? 72h?Relative Cell Growth (fold change)?* * DMSO?0.5nM?1nM?5nM?10nM?50nM?100nM?BI2536?(Time)?D. * 0?10?20?30?40?50?60?70?80?90?100?!"# $"# %"# &"# '"# ("#Animal Survival (%)?(Days)?Vehicle ?BI2536 ?Control ?P=0.012?E. -30.00?-25.00?-20.00?-15.00?-10.00?-5.00?0.00?5.00?10.00?15.00?0? 20? 40? 60? 80?Control?Vehicle?Test Agent?(Days) ?Weight Loss Relative to Study Initiation (%)?BI2536 ?C. 0?0.2?0.4?0.6?0.8?1?1.2?PLK1? SOX2?Control oligo?siPLK1#1?* * Relative Transcript Level?(fold difference)?B. -0.5?0?0.5?1?1.5?Control? siPLK1?Relative Cell Growth ?(fold change)?*  75 cell growth by ~90%. The experiment was repeated on three separate occasions. (C) Total RNA was extracted from U251 cells treated with BI2536 for 36hrs. The transcripts were quantified by RT-PCR. The RT-PCR experiment was performed in triplicates and the experiment was repeated on two separate occasions. (D) Orthotopic GBM tumours were established by intracranial injection of U251 cells into Rag2m mice (n=8 per treatment group), which were subsequently treated with 50mg/kg BI2536 i.v. weekly for four weeks. Log-rank test was performed on the Kaplan Meier curves of the experimental animals. The difference in the survival of vehicle control and BI2536-treated mice was statistically significant with a P-value of 0.012. (E) The body weight of the untreated, vehicle control- treated and BI2536-treated animals was measured regularly for 60-80 days. The mice that were treated with BI2536 showed significantly less body weight loss compared to the control mice during the course of the experiment.     76 2.6 SUPPLEMENTARY DATA  Table S2.1 Correlation between PLK1 and similary expressed genes expressed in primary GBM based on Affmetrix U133. The genes that are most highly correlated with PLK1 by mRNA expression from Affymetrix U133 microarray profiling 467 pediatric and adult glioblastomas are listed in descending order according to their Pearson’s correlation coefficients.   !"##$%&%'()*+,-).$%,Supplementary Table. 1. Correlation between PLK1 and similarly expressed genes ?expressed in primary GBM based on Affymetrix  U133 microarrays.? 77  Figure S2.1 PLK1 and neural stem cell markers SOX2, musashi and Bmi1 are co-expressed in dissociated BT74 cells. BT74 tumourspheres were dissociated into single cells and fixed in 1:1 mixture of acetone and methanol directly on glass slides. The cells were stained with antibodies against PLK1, SOX2, musashi and Bmi and the images (60x, scale bar=100µm) were obtained by confocal microscopy on Fluoview FV10i (Olympus, Japan). PLK1 was co-expressed with SOX2, musashi and Bmi1 in the nuclei of the majority of the cells examined.  !"##$%&%'()*+,-./"*%0,Supplementary Fig. 1. PLK1 and neural stem cell markers SOX2, musashi and Bmi1 are ?co-expressed in dissociated BT74 cells.?Hoechst?Hoechst? Hoechst?PLK1? PLK1? PLK1?SOX2? Bmi1?Musashi?PLK1?SOX2?PLK1?Musashi?PLK1?Bmi1?Merged, All? Merged, All? Merged, All? 78    Figure S2.2 PLK1 level is elevated in BTICs compared to normal human neural stem cells and its expression is significantly down-regulated after lineage differentiation. (A) RT-PCR was performed to examine PLK1 level in hNSC167 and GBM4, GBM8 and BT74. The transcript level of PLK1 is 4.1-fold, 6.41-fold, 10.23-fold higher in GBM4, GBM8 and BT74, respectively, compared to hNSC167. (B) PLK1 level in hNSC101 and BTICs was examined by RT-PCR. The transcript level of PLK1 is 2.37-fold, 3.71-fold, 5.92-fold higher in GBM4, GBM8 and BT74, respectively, compared to hNSC101. The results from Figure (A) and (B) were reproducible in the experiments conducted on two separate occasions. (C) RT-PCR was performed to examine PLK1 level in hNSC167 and HA. The HA express ~59-fold less PLK1 transcripts compared to the neural stem cells. (D) PLK1 transcript level in hNSC101 and HA was quantified by RT-PCR. PLK1 level is ~83-fold lower in HA compared to the neural stem cells.     !"##$%&%'()*+,-./"*%0,Supplementary Fig. 2. PLK1 level is elevated in BTICs compared to normal human ?neural stem cells and its expression is significantly down-regulated after lineage?differentiation.?A. 0?2?4?6?8?10?12?Relative PLK1 Level ?(fold difference)?* * * 1?4.1?6.41?10.23?hNSC167?GBM4? GBM8? BT74?0?1?2?3?4?5?6?7?Relative PLK1 Level? (fold difference)?***1?2.37?3.71?5.92?hNSC101? GBM4? GBM8? BT74?B. C. 1?0?0.5?1?1.5?hNSC_167? HA?Relative PLK1 Level ?(fold difference)?0.017?* hNSC ?0?0.5?1?1.5?hNSC_101? HA?Relative PLK1 Level ?(fold difference)?1?0.012?* D. hNSC ? 79  Figure S2.3 BI2536 suppresses tumoursphere formation of primary brain tumour cells but exerts a minimal effect on the differentiation of primary hematopoietic stem cells isolated from a patient. (A) Primary BTICs: L0, L1 and L2 (1x104 cells per well in 6-well plates) were treated with 5 or 10nM BI2536 for 6 days. The total number of tumourspheres (>50µm) in each well was counted and photomicrographs of the spheres were taken (scale bar=500µm). (B) Total RNA extraction 0?10?20?30?40?50?60?DMSO?0.1 nM? 1 nM? 2 nM? 10 nM?100 nM?Number of Colonies?Total?Myeloid?Erythroid?BI2536?Patient No. 2 ?!"##$%&%'()*+,-./"*%0,Supplementary Fig. 3. BI2536 suppresses tumoursphere formation of primary brain ?tumour cells but exerts a minimal effect on the differentiation of primary ?hematopoietic stem cells isolated from a patient. ?1,123,4,423,HA? BT011? Relative PLK1 Transcript Level?(fold difference) ?1?0.0335?D. 0?10?20?30?40?DMSO? BI2536?Total Number of  BT011 Tumourspheres ?(per well)?E. 0?50?100?150?200?250?L0? L1? L2?DMSO?BI2536-5nM?BI2536-10nM?DMSO? ?5? 10   (nM)?L0?L1?L2?BI2536 ?Total  Number of Tumourspheres?(per well)?* * ** ** * A. F. Relative level (fold difference) of SOX2, musashi ?and Bmi1 transcripts in human astrocytes (HA) ?and BT005 ?0?20?40?60?80?100?120?140?160?Control? SiPLK1?Control? siPLK1?Total Number of Tumourspheres?(per well)?B. * si ?C.  80 was performed on the immortalized human astrocytes HA and tumourspheres of BT005, followed by reverse transcription and qPCR. The total cDNA level of neural stem cell markers SOX2, musashi and Bmi1 was quantified and compared (in relative fold difference) between the two cell cultures. (C) SF188 cells were transfected with 5nM PLK1 siRNA for 24hrs and re-plated in neurobasal medium supplemented with EGF and FGF, a condition that supports the growth and expansion of stem and progenitor cell population. The number of tumourspheres was enumerated 6 days after culturing and representative photomicrographs (scale bar=200µm) are shown to compare the colonies in the control and siPLK1 treatment. (D) Single cells were isolated from a post-surgical pediatric GBM specimen (referred to as BT011) and grown as tumourspheres for two weeks. The RNA from the tumourspheres was isolated and RT-PCR was performed to quantify the level of PLK1 in BT011 and HA. (E) BT011 cells were plated in BI2536-containing medium for 6 days and the cells were serially passaged two more times before the tumourspheres were quantified. PLK1 inhibitor did not suppress the self-renewal and/or proliferation of these primary brain tumour cells, which expressed negligible level of PLK1. (F) Analysis of the effect of BI2536 on in vitro hematopoietic colony formation was performed on normal bone marrow derived CD34+ cells isolated from a second patient. The cells from this patient were incubated with DMSO or increasing concentrations of BI2536 (0.1-100nM) in methylcellulose cultures that contained cytokines to stimulate hematopoiesis. After 12 days in culture, myeloid and erythroid colonies were enumerated by counting under an inverted microscope based on morphology. All the studies presented in this figure were performed in replicates on one occasion due to the scarcity of the patient specimens.    81   Figure S2.4 PLK1 inhibition represses the cell growth of pediatric and adult GBM cell lines SF188 and Gli36. (A) SF188 cells were treated with two different siRNAs targeting PLK1. The growth of the cells was assessed in 72hrs by Hoechst staining and quantification on the Cellomics high-content screening instrument. PLK1 siRNA #1 and #2 inhibited cell growth to a similar extent in 72hrs. (B) The effect of PLK1 inhibition by siRNA or small molecular inhibitor was evaluated in Gli36 (adult GBM cell line) 72hrs after treatment. Gli36 cells were also very sensitive to PLK1 inhibition as 5nM of siRNA or BI2536 suppressed cell growth by ~80-90%.    !"##$%&%'()*+,-./"*%0,Supplementary Fig. 4. PLK1 inhibition represses the cell growth of pediatric and adult ?GBM cell lines SF188 and Gli36. ?A. 01.0?2.0?3.0?4.0?5.0?6.0?  0? 24? 48? 72?Relative Cell Growth ?(fold change)?(hrs)?Control ?siPLK1#1?siPLK1#2 ?B. BI2536?(nM)?0?0.2?0.4?0.6?0.8?1?1.2?DMSO?2.5 ??? 5?  ?? 10?? 20?Relative Cell Growth?(fold change)?*?*? *? 0?0.2?0.4?0.6?0.8?1?1.2?Control ??siPLK1?Relative Cell Growth?(fold change)?* * * * * * *  82    Figure S2.5 PLK1 knockdown by two targeting siRNAs decreases the transcript and protein levels of SOX2 and alters cellular morphology. (A) SF188 cells were treated with 5nM of PLK1 siRNA #1 or #2. Total RNA and proteins were extracted for RT-PCR and immunoblotting (36hrs and 48hrs respectively) to examine the expression of SOX2. The study was repeated in triplicates on two separate occasions. (B) SF188 cells were treated with 5nM of PLK1 siRNA or BI2536 for 6 days and photomicrographs were taken on the cells that remained after the treatment. Additional representative photomicrographs of the cells that underwent dramatic cellular morphological alterations are shown (scale bar=280µm). This experiment was repeated twice. !"##$%&%'()*+,-./"*%0,Supplementary Fig. 5. PLK1 knockdown by two targeting siRNAs decreases the transcript?and protein levels of SOX2 and alters cellular morphology.  ?SOX2?Actin?PLK1?Control?siPLK1#1?siPLK1#2?0?0.2?0.4?0.6?0.8?1?1.2?Control? siPLK1#1?siPLK1#2?PLK1?SOX2?*? *?A. Relative Transcript Level? (fold difference)?B. Control? siPLK1 ? DMSO? BI2536? 83 CHAPTER 3: DISULFIRAM, A DRUG WIDELY USED TO CONTROL ALCOHOLISM, SUPPRESSES THE SELF-RENEWAL OF GLIOBLASTOMA AND OVER-RIDES RESISTANCE TO TEMOZOLOMIDE.  3.1 INTRODUCTION Glioblastoma (GBM) is an aggressive type of brain tumour with limited treatment options.  Due to the location and infiltrative nature of GBM tumours, surgical resection and radiation is often ineffective thus recurrence is especially common.  Under the current treatment regime of temozolomide (TMZ) and radiation the median expected survival following resection is only 14 months (Rock et al., 2012).  Resistance to the alkylating agent TMZ is common in GBM.  Tumours expressing O6-methylguanine methyltransferase (MGMT) avoid growth inhibition by enzymatically removing the methyl groups added to DNA by TMZ (Gerstner et al., 2009; Hegi et al., 2005).  However, even MGMT silenced cases acquire TMZ resistance.  For example, chronic exposure to TMZ has been shown to generate mutations in mismatch repair genes and offers an additional route of treatment resistance (Yip et al., 2009).  There are few options available to overcome GBM growth and recurrence. Tumour re-growth and relapse is a major problem in treating GBM.  Brain tumour initiating cells, or BTICs, are a subpopulation of undifferentiated and highly drug resistant cells that may elude common surgical resection and chemotherapeutic treatments (Bao et al., 2006; Deleyrolle et al., 2011; Galli et al., 2004; Liu et al., 2006a; Singh et al., 2004).  These cells are characteristically very tumourigenic with greater drug export systems compared to normal cells.  They have the ability to self-renew, generate lineage specific progenies through differentiation and can initiate tumour formation when implanted into animals (Deleyrolle and Reynolds, 2009; Dell’Albani, 2008). Propagation and in vitro assessment of the BTIC population is done using neurosphere tissue culture conditions (Deleyrolle, 2011; Galli et al., 2004; Park and Rich, 2009; Reynolds and Weiss, 1992; Singh et al., 2004). The self-renewing properties of BTIC cells allow them to be serial passaged using these growth conditions and continually forming new spheroid cell clusters.  With the potential to evade current treatment protocols, there must be alternative methods developed for targeting the BTIC subpopulation in order to prevent GBM relapse (Park and Rich, 2009).  Polo-like kinase 1 (PLK1) is a key serine/threonine kinase involved in many essential cell cycle functions, such as: mitotic entry, centrosome maturation, cell cycle progression and cytokinesis (Arnaud et al., 1998; Golsteyn et al., 1995; Lane and Nigg, 1996; Mundt et al., 1997; van Vugt et al., 2004b).  Our group has demonstrated PLK1 to be a promising therapeutic target for brain tumours as it is very highly over-expressed in cancer compared to normal tissue (Hu et  84 al., 2009; Lee et al., 2012) As well, patients with GBM tumours that express high levels of PLK1 have a much greater probability of dying from the disease (Lee et al., 2012).  Recently we have shown that PLK1 inhibition delayed tumour growth in an orthotopic brain tumour model and also demonstrated PLK1 to be essential for sustaining the growth of BTICs as tumourspheres (Lee et al., 2012).  Although chemical inhibitors for PLK1 are being examined for clinical use (Garuti et al., 2012; Wasch et al., 2010), the long and expensive process of drug development prompts the question of whether currently approved off-patent drugs may have undiscovered anti-cancer potential. Disulfiram (DSF) has been safely used for the treatment of alcohol abuse for over sixty years.  Originally we identified DSF in a screen for drugs that inhibit tumour-initiating cells using the Prestwick Library (unpublished data).  DSF was attractive to us because it is a small molecule and as such it crosses the blood brain barrier (Eneanya et al., 1981; Maj et al., 1970; Oskarsson, 1984).  In a position paper by Kast et al., DSF was proposed for the treatment of GBM (Kast and Belda-Iniesta, 2009); therefore, we hypothesize that DSF will target BTICs.  This study provides in vitro evidence that DSF is an effective treatment for GBM and suggests it augments cytotoxicity of the currently used chemotherapeutic agent, TMZ.  The data presented here suggests new uses for the clinically safe compound, DSF, as an alternative treatment for cancer patients.   3.2 RESULTS SF188 cells are pediatric GBM cells that are unaffected by TMZ (5-15 µM) at physiologically achievable concentrations (5-15 µM) based on cell growth assays (Supplementary Figure S3.1A).  These classically TMZ resistant cells were sensitive to DSF given that 500 nM suppress growth in monolayer by ~100% in 72 hours (Figure 3.1A).  The ability of these cells to self-renew was also completely inhibited (Figure 3.1B).  BT74 cells are primary adult BTICs, which are also refractory to TMZ (Kanai et al., 2011).  However they are sensitive to DSF in neurosphere self-renewal assays (Figure 3.1C).  Likewise, GBM4 BTIC cells are sensitive to DSF in self-renewal assays (Figure 3.1D).  Examples of the impact on BT74 and GBM4 neurosphere formation are illustrated (Figure 3.1E-F).  Next we asked whether the combination of TMZ and DSF would have added benefit.  Low doses of DSF (50 nM) or TMZ (10 µM) had no effect as a single agents; however, together they inhibited growth by ~50% (Supplemental Figure S3.2A-B). Freshly isolated GBM cells were obtained from two adult patients; they are referred to as aBT001 and aBT003.  Sequencing of IDH1 and IDH2 was negative for mutation in either tumour  85 sample (data not shown).  Both cases had unmethylated MGMT (Figure 3.2A) suggesting that they may be refractory to TMZ.   As expected, TMZ did not inhibit the growth of aBT001 in monolayer (Figure 3.2B).   However, these cells were sensitive to DSF, where 500 nM inhibited growth by 87% after three days (Figure 3.2C).  The aBT003 cells were also refractory to TMZ in neurosphere assays (Figure 3.2D).  Conversely, DSF inhibited their self-renewal capacity by 95-98% (Figure 3.2D-E).  It is noteworthy that while DSF inhibited the growth of GBM cells it had no effect on the proliferation of normal human astrocytes (>10 µM, Supplementary Figure S3.3).  We recently reported that highly proliferative GBM express PLK1 and that these cells depend on this kinase for survival (Lee et al., 2012), yet its role in the context of TMZ resistance was not addressed.  Further, because DSF had such an impressive negative effect on the growth of GBM we questioned the mechanism and assessed the impact on PLK1.  Notably, DSF inhibited PLK1 expression in SF188 cells when the cells were exposed to 500 nM of the drug for 72 hrs (Figure 3.3A).  Likewise, DSF also inhibited PLK1 in these cells with 250 nM (Supplementary Figure S3.4A).  Inhibiting PLK1 with siRNA blocked the growth of these cells and induced apoptosis (Figure 3.3B-C).  Likewise, DSF (500 nM) inhibited PLK1 protein expression in U251 adult GBM cells and it suppressed their growth (Figure 3.4A-B).  Lower doses of DSF (250 nM) also inhibited PLK1 protein levels in the U251 cells (Supplementary Figure S3.4B).  Similarly, 200 nM DSF suppressed the growth of U251 cells by 80% while 500 nM completely eliminated the cells (Figure 3.4B).   Inhibiting PLK1 with the small molecule BI-2536 or with siRNA blocked the U251 cell growth (Figure 3.4C-D).  To test whether PLK1 inhibition could effectively target a truly TMZ resistant cell population, the partially TMZ sensitive U251 cells were treated with 10µM TMZ every 2 days for a week.  Interestingly, treatment of U251 cells with 10µM TMZ resulted in a dramatic increase of PLK1 protein compared to untreated and DMSO treated cells (Supplementary Figure S3.5A).  Although not responsive to TMZ, the residual cells demonstrated near complete growth inhibition with increasing concentrations of BI2536 (5-100nM) (Supplemental Figure 3.5B).  These findings suggest PLK1 to be a potential driver of TMZ resistance that can be overcome through therapeutic intervention.  The TMZ resistant BTIC cells BT74 and B241 were also sensitive to PLK1 inhibition (Figure 3.5A-B).  There was no additional benefit from combining BI-2536 and TMZ (Figure 3.5A-B).  As previously mentioned freshly isolated aBT001 and aBT003 GBM cells were sensitive to DSF therefore we addressed whether they also expressed PLK1, which they did (Figure 3.6A-B).  Given that the PLK1 target was expressed, aBT001 cells were treated for 72 hrs with increasing amounts of BI-2636 that inhibited their growth by up to 80% and induced  86 apoptosis (Figure 3.6C).  Thus, PLK1 inhibition phenocopied the effect of DSF in blocking the growth of refractory GBM cells.  3.3 DISCUSSION In the present study, we demonstrate the efficacy of DSF completely suppressing GBM cell growth in vitro.  More importantly, we observed the same degree of inhibition in brain tumour initiating cells or BTICs.  Concentrations as low as 100nM DSF suppressed cell growth in monolayer and this corresponded to an inhibition of BTIC self-renewal using neurosphere assays.  Importantly, DSF was highly effective in cells that are refractory to TMZ.  Many challenges exist in the treatment of GBM, one of which is the immense problem of TMZ resistance.  In some patients, the invasive tumour cells respond initially but most patients eventually relapse.  For other patients, their tumours are resistant from the start of treatment.  TMZ is associated with low long-term survival rate and is ineffective at targeting the BTIC subpopulation that potentially repopulates the tumour (Burkhardt et al., 2011; Gao et al., 2009; Rock et al., 2012).  Of importance, we showed that DSF was highly effective in situations where cells have developed TMZ resistance. The combination of TMZ and DSF was helpful in some instances but not others.  Several other studies showed that DSF was additive to standard cancer chemotherapy agents: paclitaxel (Yip et al., 2011a), gemcitabine (Guo et al., 2010a), docetaxel (Budman and Calabro, 2002), cyclophosphamide (Magni et al., 1996), and 5-fluorouracil (Wang et al., 2003).  This data also agrees with an in vitro study considering DSF toxicity to breast cancer stem cells (Yip et al., 2011b).  Similarly, we show enhanced cytotoxicity to TMZ and investigate the use of DSF to target the BTIC sub-population in BT-74 cells. DSF is being evaluated for treatment of malignancies.  An ongoing phase I clinical trial is currently investigating the use of DSF to treat malignancies that have metastasized to the liver (ClinicalTrials.gov Identifier: NCT00742911), and another phase II trial is considering its use in combination with cisplatin for treating metastatic small cell lung carcinoma (ClinicalTrials.gov Identifier: NCT00312819).  Verma et al. (1990) has also conducted phase II clinical trials of DSF as a method to decrease nephrotoxicity of cisplatin in a randomized study of cisplatin sensitive malignancies (Verma et al., 1990).  They reported no benefit to toxicity by combining the two drugs; however, this study had a markedly high patient drop-out rate making the data analysis inconsistent between groups (Verma et al., 1990).  It may be speculated that this study had issues due to the use of extremely high concentrations of DSF (> 3200mg).  This level far exceeds the minimal concentrations required to elicit a response as suggested by our in vitro  87 data, and is significantly higher than the dose of 250mg/day used to treat alcoholism (Suh et al., 2006; Verma et al., 1990).  At this point there has been no reported clinical experimentation in the use of DSF for the treatment of solid brain tumours.  DSF has only been used in clinical trials involving cancers that have metastasized and we believe that DSF may have great benefit if used to treat primary tumours.  Not only is DSF an inexpensive and easily administered drug, it is able to cross the blood-brain barrier, which is a major limitation in brain therapeutic design (Faiman et al., 1980; Maj et al., 1970; Oskarsson, 1984). The mechanism of action for DSF was somewhat elusive for us in that we suspected that it killed cells through aldehyde dehydrogenase (ALDH) inhibition (Lipsky et al., 2001).  However, this was not the case.  We noted that while DSF inhibits ALDH activity it was at doses that were higher then that which was required to suppress the growth of GBM cells.  In addition, blocking ALDH activity with a pan inhibitor (DEAB) did not suppress growth to the degree as DSF.  ALDH1A1 and ALDH1A3 were also individually silenced with siRNA however this resulted in little or no growth suppression (data not shown).  Loss of ALDH did not induce apoptosis in GBM cells.  These studies redirected our attention toward a pathway that may cause considerable cell death and to this end we address the possible link with PLK1.  We knew that GBM cells were absolutely dependent upon PLK1 because we recently reported that inhibiting it suppressed the growth of GBM cells by as much as 100% and this was associated with cell death (Lee et al., 2012).  Serendipitously we noted that DSF also inhibited PLK1 expression in both SF188 and U251 cells.  This novel mechanism would have broad reaching implication given that PLK1 is central to the growth of many types of cancer (Lee et al., 2012; McInnes and Wyatt, 2011).  This is the first demonstration of an off-patent drug that inhibits its expression and as such it opens up several new lines of investigation.  In the present study we demonstrated the efficacy of DSF in suppressing refractory GBM growth and self-renewal at low concentrations.  Coupled with these findings, DSF has been used safely in humans for over half a century and therefore believe it has excellent potential to be repositioned for the treatment of GBM.   3.4 MATERIALS AND METHODS Cell Culture  The ATCC supplied pediatric GBM SF188, adult GBM U251, and normal human astrocyte cells.  GBM4 and BT74 GBM cells were obtained from collaboration with Wakimoto et al., who characterized primary, and serial passaged, spheres both in vito and in vivo (Pandita et al., 2004; Wakimoto et al., 2009; Wakimoto et al., 2012).  Primary patient BTIC cells aBT001  88 and aBT003 were isolated using methods previously described (Fotovati et al., 2011; Piccirillo et al., 2006).  All primary samples were acquired in abidance with the guidelines of the Institutional Review Board, and along with patient consent.  SF188, U251, and aBT001 cells were grown in monolayer using Minimum Essential Medium/Earle’s Balanced Salt Solutions (MEM/EBSS) [Hyclone, Logan UT, USA] and Dulbecco’s Modified Eagle Medium (DMEM)/High Glucose (Hyclone), respectively, supplemented with 10% fetal bovine serum.  BT74, aBT003, and sphere assays were grown non-adherently using NeuralBasal medium with Neurocult supplement and growth factors, EGF (20ng/ml), FGF (20ng/mL) and heparin (2mg/mL).  Normal human astrocytes were grown in Astrocyte medium (ScienCell cat. 1801) on adherent plants coated with poly-L-lysine (ScienCell cat. 0413).  Drug Treatment and Growth Assay Growth assays were conducted by plating 1000 cells/well in 96-well plates with a range of concentrations of TMZ or DSF (formulated to include copper).  All treatments were done in triplicate, and plates incubated at 37°C in a 5% CO2 incubator.  After 72 hours, cells are fixed with 2% paraformaldehyde in 100µl PBS, and stained with Hoechst 33342 dye (2µg/ml) at room temperature for 30 minutes before a wash with 100µl PBS.  Plate analysis and image capture was done using an ArrayScan VTI Reader (Thermal Fisher) (Hu et al., 2009).  Dimethyl sulfoxide (DMSO) was used to reconstitute TMZ and DSF and was used as a solvent control.  Ethanol was used as a control for DEAB.  For analysis of effect of BI2536 on U251 TMZ resistant cells, U251 cells were treated with 10 µM of TMZ (or DMSO) every 2 days for a total of 7 days.  The cells were harvested for protein extraction and Western Blot was run to examine PLK1 expression.  The cells were then re-plated and treated with increasing concentrations of BI2536 (5-100nM) for 5 days before the cells were stained with Hoechst dye and quantified on Cellomics.  The effect of PLK1 inhibition was investigated using siRNA or BI-2536 as previously described (Lee et al., 2012).  Neurosphere Assay  BT74, GBM4 and SF188 BTIC cells were enriched for using a neurosphere suspension assay (Note: BT74 and GBM4 cells are always maintained as spheres as were the primary isolates described below).  Approximately 10 000 cells/well were plated into a low adherent 6 well dish using neurobasal medium supplemented with human recombinant EGF (20ng/ml), human recombinant FGF (20ng/ml) and heparin (2µg/ml) [Stem Cell Technologies].  Freshly isolated BTICs were obtained from adult patients under informed consent according to the BC  89 Cancer Agency guidelines.  Tumour cells were isolated as previously described by us (Lee et al., 2012).  Neurospheres were grown for 5-6 days following plating.  Spheres >30µm were counted and photographed using an Aniovert 40CFL microscope and AxioCam MRc camera.  NeuroCult Chemical Dissociation kit (Stem Cell Technologies, cat. #05707) was used to passage cells, which are counted and replated as single cells.  All drug treatments of TMZ and DSF were done at the time of plating, and repeated during serial passaging.   PLK1 Regulation and Expression SF188 or U251 cells were treated with DSF (100-500 nM) for 24 hrs, proteins were harvested and levels of PLK1 were evaluated by immunoblotting.  RNA for gene expression analysis was isolated using Qiagen RNeasy Mini Kit (Cat. #74106).  Transcript expression was determined using qRT-PCR with PLK1 Assay on Demand (Applied Biosystems, cat. #4331182).  PLK1 expression was silenced using siRNA as previously described (Lee et al., 2012).  Tumour cell growth following siPLK1 transfection was evaluated compared to scramble control RNA in SF188 cells.  PLK1 was also inhibited with BI-2536 and growth was assessed in SF188, BT74, and BT241 all of which are TMZ resistant.    Real-Time Quantitative Reverse-Transcription PCR  RNA was extracted from the cells using the Qiagen RNeasy Kit following the manufacturer’s protocol. Synthesis of cDNA and real-time PCR experiments were conducted using FAM-labeled Taqman Assay-on-Demand probes according to the method previously described by us (Wu et al., 2006).  TATA-box binding protein (TBP) or 18s mRNA was used as house keeping genes for data normalization.   Immunohistochemisty Primary adult GBM cases (aBT001 and aBT003) were formalin-fixed, paraffin embedded, sectioned and immunostained for PLK1. PLK1 protein expression was evaluated using LSBio antibody (diluted 1:200, PLK1 rabbit anti-Human polyclonal Antibody LS-B4225- LSBio LifeSpan Bioscience, Seattle, WA).  The secondary antibody was universal detection kit from DAKO LSAB2 System-HRP (DAKO, Carpinteria, CA).  The MGMT status of these tumours was determined by PCR as previously described (Gerstner et al., 2009).     90 Statistical Analysis Experimental data was collected from multiple experiments and reported as the treatment mean ± standard error.  Significance was calculated using the student’s t-test, where *p<0.05, and **p<0.01.    91 3.5 FIGURES  Figure 3.1 DSF inhibits GBM cell growth and self-renewal. (A-B) SF188 cells were treated with 50, 100 or 500 nM DSF and tumour growth was assessed in monolayer or in serial neurosphere assays. The studies of Figure (A) and (B) were repeated 3 and 2 times, respectively. (C-D) Adult GBM BT74 and GBM4 cells were treated with 50-500 nM DSF and self-renewal was assessed in neurosphere assays. Microscopy that demonstrates the effect of DSF treatment on BT74 (E) and GBM4 (F) neurosphere growth. Scale bar = 200 µm. The studies of Figure (E) and (F) were repeated twice and once, respectively.   A. Figure 1. E. C. 85?134?34? 30?0?50?100?150?DMSO? 50nM DSF?100nM DSF?500nM DSF?Average Number of Spheres (per well)? SF188?Passage 1? Passage 2?263?209?27? 5?0?100?200?300?DMSO? 50nM DSF?100nM DSF?500nM DSF?Average Number of Spheres (per well)? BT74?Passage 1? Passage 2?BT74!"#$$%!DMSO?100nM DSF?50nM DSF?500nM DSF?216?169?29? 1?!"#!!"$!!"%!!"DMSO? 50nM DSF?100nM DSF?500nM DSF?Average Number of Spheres (per well) GBM4?Passage 1? Passage 2?GBM4 Cells?DMSO? 50nM DSF?100nM DSF? 500nM DSF?100%?85%? 67%?1%?0%?50%?100%?150%?DMSO? 50nM DSF?100nM DSF?500nM DSF?Relative Cell Growth? SF188?B. D. F.  92  Figure 3.2 Freshly isolated GBM cells are sensitive to DSF yet resistant to TMZ.  (A) Primary GBM cells referred to aBT001 and aBT003 were isolated from adult patients with GBM.  DNA was isolated from the tumours and subjected to MGMT analysis by PCR.  In both cases the MGMT promoter was not methylated indicating that the protein would be expressed (M = methylated, UM = unmethylated). (B-C) The growth of aBT001 was unaffected by TMZ however DSF suppressed their growth by as much as 92% with a single treatment. Cell growth was assessed after 72 hrs. The studies of Figure (B) and (C) were repeated 2 and 3 times, respectively. (D-E) TMZ was ineffective at suppressing self-renewal when aBT003 cells were exposed to the drug.  However DSF suppressed self-renewal by 95-98% based on a single exposure. Scale bar = 200 µm. The experiment was performed once.  A. B. 92?105?5? 2?0?20?40?60?80?100?120?DMSO? 10uM TMZ?100nM DSF?500nM DSF?Total Number of Spheres ?(per well)?aBT003?DMSO? 10uM TMZ?100nM DSF? 500nM DSF?aBT003 Cells?82%?72%? 76%? 69%?13%? 8%?0%?20%?40%?60%?80%?100%?120%?DMSO?10nM?50nM?100nM?200nM?500nM?1uM?Relative Cell Growth ?aBT001 ?DSF? Solvent Control?!"100%?105%?85%?90%?95%?100%?105%?110%?115%?DMSO? 10uM TMZ?Relative Cell Growth ?aBT001 ?D. Ladder?M? UM?+ Control?M? UM?- Control?M?M?UM? UM?aBT003?aBT001?C. E.  93   Figure 3.3 DSF inhibits the expression of PLK1 in pediatric GBM SF188 cells. (A) SF188 cells were exposed to DSF for 24 hrs using DMSO as a solvent control. Protein and transcript levels of PLK1 were assessed using immunoblotting or qRT-PCR. The immunoblotting was repeated 3 times and qRT-PCR experiment was performed twice. (B-C) PLK1 inhibition with siRNA inhibits SF188 cell growth and induced apoptosis based on PARP and caspase 3 cleavage. Scramble RNA oligo was transfected as a control.  The efficacy of PLK1 inhibition on SF188 growth is exemplified in combination with 10 µM TMZ. The experiment for Figure (B) was performed in triplicates. The immunoblotting for Figure (C) was repeated twice.  A. DMSO?TMZ? 10   50   100 (μM) ?!"!"!" !"0.00?0.20?0.40?0.60?0.80?1.00?1.20?Relative Cell Growth ?(fold change)?Control siPLK1?PARP ?PLK1?Actin?DMSO?Control ? siPLK1?Caspase 3 ?DMSO?PARP (cleaved)?Caspase 3 (cleaved)?10?50?100?10?50?100?TMZ? TMZ?(μM)?B. C. PLK1?Actin?DMSO? 100nM ? 500nM ?SF188 Cells?DSF?SF18 8  CellsPLK1 Gene Expression (Fold Difference)D MSO 100nM DSF 500nM DSF0. DSF500nM DSFTreatment!!" 94  Figure 3.4 DSF inhibits the expression of PLK1 in adult GBM U251 cells. (A) U251 cells were exposed to DSF for 24 hrs using DMSO as a solvent control. Protein and transcript levels of PLK1 were assessed using immunoblotting or qRT-PCR. The immunoblotting was repeated 3 times and qRT-PCR experiment was performed twice. (B-C) Inhibiting PLK1 with BI-2536 or siRNA inhibits their growth. The efficacy of PLK1 inhibition on U251 growth is exemplified in combination with 10 µM TMZ (n=2). (D) Likewise, DSF inhibits the growth of U251 cells in a dose-dependent manner using DMSO as a solvent control. The study was repeated 3 times. A. Figure 4. B. C. 100?0.00?0.20?0.40?0.60?0.80?1.00?1.20?1.40?Relative Cell Growth ?(fold change)?DMSO?BI2536-2.5nM?BI2536-5nM?*?*? *?**? **?*?**?DMSO? 10? 50?TMZ?(μM)?0.00?0.20?0.40?0.60?0.80?1.00?1.20?1.40?Relative Cell Growth ?(fold change)? Control Oligo?siPLK1?*? *?*? *?DMSO?10? 50? 100?TMZ?(μM)?100nM ?PLK1?Actin?DMSO?500nM ?U251 Cells?DSF? U251 CellsPLK1 Gene Expression (Fold Difference)DMSO 100nM DSF 500nM DSF0. DSF500nM DSFT r e a t m e n t!!"110%? 100%?82%?19%?0%?0%?20%?40%?60%?80%?100%?120%?140%?DMSO? 10nM? 50nM? 100nM? 200nM? 500nM?Relative Cell Growth ?U251 Cells?DSF? Solvent Control?!!"D.  95  Figure 3.5 Targeting PLK1 inhibits growth of drug resistant cells with up-regulated PLK1 protein. (A) U251 cells have lower PLK1 transcript expression than TMZ resistant SF188 cells. U251 cells were treated with 10 µM TMZ every 2 days for a total of 7 days and (B) partial TMZ sensitivity is demonstrated in a growth assay. The experiments for Figure (A) and (B) were performed 3 and 2 times, respectively. (C) Immunoblot demonstrating an increase in PLK1 protein levels in TMZ treated U251 cells compared to untreated and DMSO after 24 hours. Actin is used as a loading control protein. The study was repeated on two separate occasions. (D) The surviving TMZ resistant cells were re-plated and treated with increasing concentrations of BI-2536 for 5 days. This experiment was repeated in triplicates on two separate occasions.   A. Figure 5. B. PLK1 ExpressionCell LineRelative PLK1 Transcript ExpressionSF188 U2510. U251 cells T reatmentRelative Percent Cell GrowthUntreated DMSO TMZ050100150200100%?112%?28.5%?BI2536 treatment of U251 cells surviving TMZConcentration of BI2536Relative Percent Cell GrowthDMSO5nM 10nM 20nM50nM100nM050100150100%?27.7%?10.4%?8.6%? 6.6%? 4.7%?D. PLK1?Actin?Untreated? DMSO? TMZ?U251 Cells? 96   Figure 3.6 PLK1 inhibitors can be used to over-come TMZ resistance. (A) BT74 cells are resistant to TMZ yet sensitive to PLK1 inhibition with BI-2536. (B) BT241 cells are a second example to which the cells are TMZ resistant yet sensitive to PLK1 inhibition. Both models are maintained as primary isolates and only cultured as neurospheres. The combination of TMZ and BI-2536 did not further improve self-renewal inhibition. Scale bar = 500 µm. Results from Figure (A) and (B) were reproducible from two separate experiments performed in duplicates.   P1 P2 !" !"##BI2536?TMZ? ## !"!"!" !"##BI2536?TMZ? ## !"!"P0 P1 P2 P0: Primary sphere formation?P1: Secondary sphere formation?P2: Tertiary sphere formation?BT74?BT241?P0 P0: Primary sphere formation?P1: Secondary sphere formation?P2: Tertiary sphere formation?0?20?40?60?80?100?120?Total Number of Neurospheres?(per well)?P0?P1?P2?$" $$ $"! !## #BI2536?TMZ? # !!"0?20?40?60?80?100?120?P0?P1?P2?! !##BI2536?TMZ? ## !!"A. B. Total Number of Neurospheres 
(per well)?$ $$$" $$" 97  Figure 3.7 aBT001 and aBT003 express high levels of PLK1. (A) PLK1 levels were assessed in aBT001 and aBT003 by immunostaining. Both cases express high levels of PLK1.  (B-C) The PLK1 inhibitor BI-2536 suppressed the growth of aBT001 and induced cell death. The study of Figure (C) was repeated twice.   !""#!""#aBT001?B.?A.?500 uM?500 uM?100 uM?100 uM?Figure 7. 0%?20%?40%?60%?80%?100%?120%?Untreated?DMSO?5nM BI-2536?10nM BI-2536?25nM BI-2536?Relative Percent Cell Growth?Control? BI2536?0?0.5?1?1.5?2?2.5?3?3.5?Untreated?DMSO?5nM BI-2536?10nM BI-2536?25nM BI-2536?Percent Apoptotic Cells?Control? BI2536?C. aBT001?Plk1 Expression?Plk1 Expression?aBT003? aBT003? 98 3.6 SUPPLEMENTARY DATA    Figure S3.1 SF188 cells are TMZ resistant. SF188 cells plated in triplicate wells of a 72 hrs monolayer growth assays are resistant to TMZ at concentrations of 1, 5, 10 and 15 µM. The experiment was performed 3 times.   Supplemental Figure 1. 105%? 102%? 100%?86%?0%?50%?100%?150%?DMSO? 1uM ? 5uM ? 10uM ? 15uM ?Relative Cell Growth ?TMZ ? Solvent Control ?SF188 Cells? 99  Figure S3.2 Combination treatment of DSF augments TMZ cytotoxicity. (A) SF188 cells treated with DSF (50 nM-1 µM), in combination with 10 µM TMZ, in a 72 hrs 96 well plate monolayer growth assay. Cells were plated in triplicate and calculated relative to DMSO control treatment growth [**p< 0.005]. (B) BT74 neurosphere assay testing 10 µM TMZ alone, and in combination with DSF (10-200 nM). BT74 spheres >30 µM were counted following 5-6 days of non-adherent growth in neurobasal medium supplemented with growth factors, then chemically dissociated to serial passage and grown for an additional 5-6 days. Morphology of BT74 spheres are shown following 6 days of drug treatment. Scale bar = 200 µm. The studies of Figure (A) and (B) were performed once.   A.?Supplemental Figure 2. 137?122?141?69?4? 2?0?20?40?60?80?100?120?140?160?DMSO?10uM TMZ?10nM DSF 50nM DSF 100nM DSF 200nM DSF Total Number of Spheres ?(per well)?BT74?50nM DSF +TMZ?200nM DSF +TMZ?DMSO? 10uM TMZ?10nM DSF +TMZ?100nM DSF +TMZ?80%?27%? 33%?10%?0%?20%?40%?60%?80%?100%?120%?DMSO? 50nM DSF? 100nM DSF?200nM DSF?1uM DSF?Relative Cell Growth ?TMZ + DSF?DSF + 10uM TMZ? Solvent Control?B.?!!" 100  Figure S3.3 High doses of DSF are safe for normal human astrocytes.  DSF was given to normal human astrocytes at a concentration range of 5 nM-10 µM and cell proliferation was assessed 72 hrs later. DSF had no effect on the growth of normal cells. The experiment was performed twice.   126%?105%? 115%? 99%?124%?89%?118%?0%?50%?100%?150%?DMSO?5nM?10nM?100nM?500nM?1uM?5uM?10uM?Relative Cell Growth ?Normal Human Astrocytes?DSF? Solvent Control?Supplemental Figure 3.  101  Figure S3.4 DSF inhibits the expression of PLK1. (A) SF188 or (B) U251 cells were treated with DSF and the proteins were harvested 72 hrs later. In both cell lines, 250 nM DSF suppressed PLK1 protein levels based on immunoblotting. The experiments for Figure (A) and (B) were repeated three times.   Supplemental Figure 4. PLK1?DMSO? 100nM? 250nM ?SF188 Cells?Tubulin?DSF?PLK1?DMSO? 250nM? 500nM ?U251 Cells?Tubulin?DSF?A.?B.? 102 CHAPTER 4: PERSONALIZING THE TREATMENT FOR MEDULLOBLASTOMA: POLO-LIKE KINASE 1 (PLK1) AS A MOLECULAR TARGET FOR THE SONIC HEDGEHOG (SHH) SUBTYPE.  4.1 INTRODUCTION Medulloblastoma (MB) is the most common malignant pediatric brain tumour.  The current treatment for MB entails maximal safe resection, whole brain and spinal cord radiation for children over the age of 3, and aggressive chemotherapy including stem cell transplantation in younger children yielding five-year survival rates of 60-80%.  The advances in medical treatments have improved patient survival considerably from 5% in the 1960’s (Agerlin et al., 1999) to >70% for the standard-risk disease in recent years (Sirachainan et al., 2011; Taillandier et al., 2011).  Yet the five-year survival rate for the high-risk disease is still dismal (16-70%) (Sirachainan et al., 2011) and almost all survivors will inevitably suffer from various adverse, life-long sequelae from treatment.  These undesirable side effects are attributable to the detrimental impacts that surgical procedures, radiation and chemotherapy have on the developing brain (Mabbott et al., 2005).  Therefore, it is imperative to search for novel therapeutics that could improve the cure rate, as well as, the quality of life for patients. Medulloblastoma can be divided into four different subgroups or subtypes that include WNT, SHH, Group 3 and Group 4.  These subtypes were originally described based on differences in gene expression using cDNA microarrays (Northcott et al., 2011b) and then later by immunohistochemistry (IHC) (Ellison et al., 2011a; Northcott et al., 2011b).  Microarrays are problematic because they require fresh snap-frozen tissues whereas IHC is hindered by the subjectivity of scoring and different IHC staining across laboratories.  More recently, mRNA based assays have been developed using the Nanostring nCounter system to avoid some of the problems associated with prior methods of MB sub-classification (Northcott et al., 2011b).  The advantages of NanoString technology are: 1) multiple genes are used to distinguish MB subtypes, 2) it is highly quantitative, and 3) it does not require an amplification step allowing for low abundance genes to be detected from archival formalin-fixed paraffin embedded (FFPE) tissues.  The MB subtypes differ not only in genetic signatures but also with respect to response to clinical therapy (Kool et al., 2012).  In studies of MB where adult and pediatric patients were evaluated collectively, the WNT molecular subtype is associated with the best prognosis while the Group 3 tumours fare the poorest (SHH and Group 4 tumours correlate with intermediate outcome).  Extensive insights into the biology of the SHH pathway have spearheaded significant progress into the development of related targeted therapies, notably to Smoothened for which  103 there are several open clinical trials (Lauth et al., 2007).  There are already reports of acquired resistance due to point mutations in SMO (Yauch et al., 2009), amplification in GLI2 (Dijkgraaf et al., 2011) and signaling through the PI3K pathway (Buonamici et al., 2010).  It is therefore possible that other signal transduction pathways may provide alternative approaches to the management of MB.  We provide clinical and pre-clinical evidence suggesting that polo-like kinase 1 is a provocative molecular target for pediatric MB. 4.2 RESULTS SHH MB shows the poor outcome in the BCCH patient cohort To evaluate MB tumours, we obtained 74 patient specimens from 1986 to January 2012 for which the clinical follow-up was available (Table 4.1).  RNA was extracted from the patient tumour tissue and subjected to analysis using the nCounter system.  In parallel, the same patient samples were also used to create a companion tissue microarray (TMA) for protein analysis.  Tumour samples were assigned using subtype classification (Supplementary Table S4.1), and univariate Kaplan-Meier survival analysis demonstrated that patients with SHH tumours had high probabilities of relapse (Figure 4.1A) and death (Figure 4.1B).  Patients with Group 3 tumours also did very poorly as described in a study that included adult and pediatric patients (Northcott et al., 2011a).  Children with WNT or Group 4 tumours relapsed less frequently and lived longer (Figure 4.1A-B).  A heatmap of the patients gene expression illustrates how each of the subtypes differed (Figure 4.1C).    Drug screen identifies PLK1 inhibitors to be a cytotoxic agent in Daoy cells This data prompted us to address whether there may be drugs in the “pipeline” that would be beneficial for combating SHH type tumours.  Therefore, we subtyped four MB cell lines (ONS76, UW228, UW426 and Daoy) and each of them clustered into the SHH group (Figure 4.1C and Supplementary Table S4.1).  Subsequently, we conducted a drug library screen in Daoy cells against 129 drugs, most of which are in clinical trials.  The cells were treated with 1 or 10µM of the compounds and viability of the cells was evaluated 72 hrs later.  In order to narrow down the list of 129 small molecules, the compounds had to meet specific criteria: 1) ≥70% growth inhibition, 2) activity at 1µM and more so with 10 uM, 3) potential to cross the blood-brain-barrier, 4) currently in clinical trials, and 5) novelty.  Initially, 11/129 compounds partially fulfilled these criteria.  However, some compounds had shown toxicity in clinical trials or were previously studied.  With this knowledge, we chose to pursue PLK1 because of its relative  104 novelty as a target and the 90% growth inhibition efficacy demonstrated by the PLK1 inhibitors group: BI6727, BI2536, and GSK461364 (Figure 4.1D).   PLK1 expression is an independent prognostic marker predicting poor survival in MB In patients, PLK1 levels were evaluated in our cohort and compared to normal cerebellum.  PLK1 mRNA was higher in the vast majority of MB compared to normal cerebellum using the nCounter system (P<0.001) (Figure 4.2A, Supplementary Figure S4.1A) and by IHC (Figure 4.2B).  PLK1 also tracked with other genes associated with proliferation in SHH such as HMMR (data not shown) and Aurora A, which was also found to be over-expressed in MB specimens compared to normal cerebellum (Supplementary Figure S4.2A-C).  This is consistent with PLK1, Aurora A and HMMR being associated with tumour initiation in the PTCH1+/- mouse model (Read et al., 2009).  The mRNA expression of PLK1 was associated with higher rates of relapse and poor survival as shown in a Kaplan-Meier univariate analysis (Figure 4.2C-D).  Similarly, multivariate survival analysis was done using Cox regression proportional hazards method and the results are displayed in Table 4.2.  The variables identified as independent factors affecting patient survival include: presence of metastasis (HR= 3.488; 95% CI 1.093-11.126; P=0.035) and expression of PLK1 (HR=3.673; 95% CI 1.426-9.460; P=0.007).  Age and radiation treatment were significantly associated with survival in a univariate Log-Rank test, but did not show significance in the multivariate analysis.  As well, clinical characteristics such as: sex, extent of resection, chemotherapy and chemotherapy plus radiation combination, were not significant variables associated with survival in the Log-Rank test  (Table 4.2).  Subsequently, we characterized PLK1 between the MB subgroups where it was highest in SHH group tumours (Supplementary Figure S4.1C-D).  One of the reasons children with SHH tumours may do so poorly is that many of them are younger and are treated with radiation sparing infant protocols at our institutions (Fouladi et al., 2005).  We analyzed the patients in our cohort and stratified the SHH subtype by eliminating those that were under the age of 3 years old.  However, patients with the SHH subtype still had the worst outcomes (Supplementary Figure S4.3A-B).  Another consideration would be that those patients who did not receive radiation may do worse and therefore this could be a confounding variable (Fouladi et al., 2005).  As our centre would consider radiation sparing protocols in children over the age of three, the data was re-analyzed to include only those that received radiation and again the SHH subtype was consistently the most aggressive (Supplementary Figure S4.4A-B).  When children under the age of 3 years old and those that did not receive radiation were eliminated from the analysis, PLK1 remained significantly associated with relapse and poor survival (Supplementary Figure S4.3C-D, S4.4C- 105 D). By immunostaining a companion TMA of the same cases that were evaluated by NanoString, PLK1 protein was associated with relapse and poor survival in univariate analyses (Supplementary Figure S4.5A-B).    Patient-derived MB cells expressing high level of PLK1 are susceptible to PLK1 inhibition Primary patient-derived MB cells were obtained from surgical specimens to further explore the possibility of inhibiting PLK1.  BTX001 and BT014 were grown as tumourspheres.  BTX001 was classified as a SHH tumour (Supplementary Table S4.1, sample name “DUNN-102A”), and PLK1 mRNA was higher in BTX001 compared to normal cerebellar tissue (CB) or human neural stem cells by qRT-PCR (Figure 4.3A).  Importantly, BI2536 hindered self-renewal upon serial passaging (Figure 4.3B).  In contrast, BT014 cells express a low level of PLK1 (Figure 4.3C) and they were not responsive to BI2536 (Figure 4.3D).  As an in vitro evaluation for the safety of BI2536, we performed a counter screen on hNSCs to demonstrate that the PLK1 inhibitor had negligible effects on the growth of these cells (Figure 4.3E).  We also examined the expression of PLK1 mRNA in additional freshly isolated primary specimens (BTX001, BT006, BT007, BT274, BT008) and found the levels were higher than normal human astrocytes, which is another predominant cell type found in the brain (Figure 4.3F).  The expression of PLK1 in the Daoy model was also comparable to the primary isolates and thus much higher then the HA cells.  By immunostaining for PLK1 in normal adult mouse brain, we noted that it was not expressed in the white matter, mature neurons, or Purkinje cells (data not shown).    PLK1 inhibition suppresses cell growth, induces G2/M arrest and apoptosis in Daoy cells To better characterize the biological effects of PLK1 suppression on MB cells, we first silenced the expression of this protein in Daoy cells.  Coinciding with the result from the drug library screen, PLK1 inhibition with siRNA also suppressed the cell growth of Daoy cells by ~90% in 72hrs (Figure 4.4A).  BI2536 repressed Daoy cell growth in a dose-dependent manner in 72hrs (Figure 4.4B).  Likewise, PLK1 protein and mRNA expression was detected in two additional MB cell lines ONS76 and UW426 (Supplementary Figure S4.6A-B), which were equally sensitive to PLK1 inhibition using BI2536 (Supplementary Figure S4.6C-D).  As expected, BI2536 suppressed phosphorylation of the known PLK1 substrate CDC25C (Figure 4.4B).  PLK1-targeting siRNA or BI2536 eliminated the majority of the cells in 72hrs, as shown by crystal violet staining (Figure 4.4C).  BI2536 treatment at 24hrs halted the cell cycle, causing G2/M arrest (Figure 4.4D).  At a later time-point (48hrs), PLK1 inhibition induced apoptosis as  106 shown by an increase in Annexin V staining (Figure 4.4E), PARP cleavage and phosphorylation of histone H2AX (Figure 4.4F).    Inhibition of PLK1 hinders the self-renewal and/or proliferation of Daoy tumoursphere cells and prolongs the survival of mice bearing MB tumours Along with these effects, PLK1 was detected in Daoy tumourspheres along with established neural stem cell markers SOX2 (Komitova and Eriksson, 2004), musashi (Kaneko et al., 2000) and Bmi1 (Molofsky et al., 2003), prompting us to address whether BI2536 may also block self-renewal (Figure 4.5A).  Indeed self-renewal was completely abolished with 10nM BI2536 (Figure 4.5B).  Having demonstrated that Daoy (Figure 4.5C) and two additional MB cell lines (Supplementary Figure S4.6E-F) responded to chemotherapeutic agents (vincristine, cisplatin and etoposide) that are used for MB treatment, we took the study a step further using an animal model to compare the in vivo efficacy of BI2536 and conventional chemotherapeutic agents used in intensive chemotherapeutic protocols such as the infant Head Start protocols.  BI2536 prolonged the survival of animals and the effect was comparable to the standard-of-care drug therapies (Figure 4.5D).  It is important to note that the morphology of the Daoy xenografts in our NOD-CB17-SCID mouse mice was similar to primary pediatric large cell anaplastic medulloblastoma (Supplementary Figure S4.7A).  The tumours had prominent nuclei, cell wrapping, an open chromatin pattern, multiple mitotic figures and pleomorphism (Supplementary Figure S4.7B-C).  This is consistent with a study reported by Shu et.al. (Shu et al., 2006) where the Daoy cells were injected intracranially into Rag2 Scid mice.  In addition, we noted that the tumour cells invaded into the white matter, disseminated throughout the cerebellum and there was leptomeningial spread (Supplementary Figure S4.7D-F).  While SHH tumours are more often thought of as being desmoplastic nodular in appearance, large cell anaplastic tumours are also not uncommon (Northcott et al., 2012a). 4.3 DISCUSSION In this study, we discovered that children diagnosed with MB were more likely to relapse and die if their tumour was classified in the SHH subtype.  This is in contrast to previously published studies (Kool et al., 2012; Northcott et al., 2011b) and we suspect that this could be due to differences in the way patients are treated, such as avoiding radiation in younger patients and/or the composition of the patient cohort.  In our study, we only included pediatric MB whereas in other reports the analyses included adult and pediatric tumours (Northcott et al., 2011a).  The differences may lie in the age of the patients as it is known that the SHH group is  107 more often diagnosed in infants (Kool et al., 2012).  Further, the SHH subtype in children and adults are associated with different outcomes (Northcott et al., 2012a). PLK1 is an oncogenic kinase that confers a growth and survival advantage in cancer cells through its central role in mitosis (Elez et al., 2000).  In multiple cancer models, PLK1 inhibition specifically eliminates the malignant cells while leaving the non-malignant cells unharmed (Cogswell et al., 2000; Liu et al., 2006b).  Furthermore, PLK1 is highly expressed in cancer cells but not in their normal cell counterparts (Holtrich et al., 1994; Tokumitsu et al., 1999) rendering this kinase a particularly attractive molecular target for cancer therapeutics.  Various small molecule inhibitors to this kinase have been designed and evaluated in phase I/II clinical trials including BI2536 (Mross et al., 2008), BI6727 (Rudolph et al., 2009) and GSK461364 (Olmos et al., 2011).  None of these trials have specifically addressed the possibility that PLK1 inhibitors may be beneficial for the treatment of brain tumours.  However to this point, we published that PLK1 inhibitors could be used to target glioblastoma (Lee et al., 2012) and may be additionally beneficial in overcoming temozolomide resistance (Triscott et al., 2012). Targeted therapies directed at specific MB subtypes is the next step in translating this paradigm shift in molecular classification schemes.  We illustrate that PLK1 is a promising drug target for MB because 1) it is highly expressed in tumours relative to normal brain tissues and 2) there are small molecule inhibitors that suppress primary SHH MB cells and cell lines in vitro and in vivo.  It is forseeable that PLK1 inhibition could also provide benefit to other types of MB as it is also expressed in Group 3 tumours (Supplementary Figure S4.1B-C).  By cDNA microarray, Harris et al. examined PLK1 levels in 16 pediatric MB specimens and found no association between PLK1 expression and patient outcome.  The authors also examined mRNA by microarray in a cohort of 120 pediatric and adult MB patients where PLK1 is higher in tumours compared to normal tissues, but this was not associated with any particular tumour subtype (Harris et al., 2012).  Presumably, the possible association with patient outcomes was not reported likely because the clinical data were not available.  In conclusion, patients that have tumours expressing very high levels of PLK1 are considered to be at elevated risk for relapse and death.  Often times, these tumours fall in the SHH subtype and therefore patients may benefit from inhibitors to that pathway.  Since there are several PLK1 inhibitors in clinical trials for adult malignancies, we propose that these drugs may also provide benefit for selected MB patients.  In the future, it may be desirable to personalize the treatment of MB by selecting patients with high PLK1 using accurate, sensitive methods such as that of the NanoString nCounter technology, where subgroup affiliation of  108 tumours can be assigned rapidly and reproducibly (Northcott et al., 2012b).  Further, we anticipate that PLK1 inhibitors may have fewer detrimental side-effects as PLK1 is not expressed at high levels in normal brain tissue.  Therefore, it could be a great improvement to many of the chemotherapies currently being used that can often cause long-term adverse effects (Dhall et al., 2008).  These pre-clinical studies pave the way for improving the treatment of MB through PLK1 inhibition.  4.4 MATERIALS AND METHODS Medulloblastoma TMA and nCounter Analyses We obtained 74 tumour blocks from patients diagnosed with MB between 1986 and 2012.  Samples included both primary and relapse specimens.  Medical charts were reviewed and pertinent clinicopathologic data was recorded (RR, KO and CF).  Patient ages ranged from 3 months to 16.8 years old.  A review of this cohort is shown in Table 1.   NanoString nCounter Gene Expression Profiling RNA was extracted from three 20µm scrolls of FFPE tissue using Qiagen RNeasy FFPE kit (Valencia, CA, USA).  Exactly 250ng of RNA was run for each patient sample and RNA quality was assessed using Nanodrop spectrometry.  A 2100 Bioanalyzer (Agilent Technologies, Mississauga, ON, Canada) was used to spot check RNA quality in random samples.  For the MB cell lines, they were grown as tumourspheres, RNA was isolated and subsequent analysis was performed in a similar manner to the FFPE tissues.  Analysis using the nCounter Gene Expression system was conducted at the Centre for Translational and Applied Genomics (CTAG) (BC Cancer Agency, Vancouver, BC, Canada).  A custom codeset synthesized by NanoString Technologies (Seattle, WA, USA) was designed which included 22 MB specific subtyping gene probes (Northcott et al., 2012a) plus other genes of interest which specifically included PLK1 (NM_005030.3) and AURKA (NM_003600.2).  The recommendations outlined by NanoString Technologies were all followed regarding mRNA sample preparation, hybridization, detection and scanning, and data normalization.   Gene Expression Data Analysis Gene expression data was analyzed for MB subgroup assignment using R statistical programming version 2.15.2 for Windows, and R package pamr (PAM) version 2.23 as previously described (Northcott et al., 2011b).  Cutoffs for PLK1 expression were assigned based on z score deviation from the mean expression of the cohort. NanoString output scores  109 below 400 were considered low, 400 to 580 medium, and above 580 considered high expressors.  Heatmaps were generated using unsupervised hierarchical clustering with average linkage using Cluster version 3.0 and Treeview version 1.60.  Real-Time Quantitative Reverse-Transcription PCR  RNA was extracted from the cells using the Qiagen RNeasy Kit following the manufacturer’s protocol. Synthesis of cDNA and real-time PCR experiments were conducted using FAM-labeled Taqman Assay-on-Demand probes according to the method previously described by us (Wu et al., 2006).  TATA-box binding protein (TBP) or 18s mRNA was used as house keeping genes for data normalization.   Immunohistochemistry The original histologic slides were reviewed on all cases to confirm the diagnosis of MB.  Atypical teratoid rhabdoid tumours (ATRTs) were excluded from further analysis via INI-1 immunohistochemistry.  A tissue microarray (TMA) was subsequently constructed from the original FFPE blocks; triplicate 1mm cores were extracted from every tumour resection performed on the patients.  A total of 83 patients were included, with patient outcomes data available for 74 of the 83 cases.  Antibodies to NRP3, KCNA1 and SFRP1 were used to subtype the Group 3, Group 4 and SHH tumour subtypes as previously described (Northcott et al., 2011b).  ß -catenin was used for the identification of the WNT subgroup.  For these proposed subgroup specific antibodies, the Ventana Benchmark XT autostainer was employed; vendors, clones, dilutions and protocols are listed as follows: ß-catenin- BD (ab610154) antibody, dilution 1:100, protocol cc1; SFRP1- Abcam (ab4193) antibody, dilution 1:200, protocol cc1; NPR3- Sigma (HPA031065) antibody, dilution 1:30, protease protocol; KCNA1- Abcam (ab32433) antibody, dilution 1:200, protocol cc1. For PLK1 IHC staining, TMA slides were incubated with CitriSolvTM for de-paraffinization.  After de-paraffinization, the slides were dehydrated in 100% ethanol and rehydrated using an ethanol gradient.  The slides were steam-heated in 0.1 mol/L citrate buffer for 20 minutes for antigen retrieval.  Endogenous peroxidase activity was quenched by incubating with 3% v/v H2O2, at room temperature.  The slides were rinsed by 3 × 5-minute washes in PBS containing 0.2% (v/v) Triton X-100.  The sections were treated with DAKO protein block for 10 minutes and incubated overnight with anti- rabbit polyclonal anti-human PLK antibody, Lifespan Bioscience, diluted (1:200) in PBS containing 0.3% (v/v) Triton X-100 and 0.1% (w/v) bovine serum albumin.  All slides were counterstained with hematoxylin for 40 seconds and rinsed with tap water.  The  110 slides were mounted with Cytoseal XYL and coverslips.  Slides were viewed with the Nikon Eclipse 50i light microscope and were photographed using a Digital-Sight DS-Fi1 camera (Nikon, Japan).  All IHC stains (with the exception of ß-catenin) were scored semi-quantitatively using a 4 point scale (0-3) that took into consideration the intensity and the diffusivity of the staining.  Subsequently, these scores were binarized into “low” (0,1) or “high” (2,3) categories for survival analysis.  ß-catenin was scored according to the presence of nuclear staining only; cases were considered to be either nuclear positive or negative.  Drug Library Screen The small-molecule targeted therapeutic agents used in the screening analysis were synthesized, purity checked and purchased from Chemietek (Minneapolis, MN, USA).  The stock compounds were diluted to working stocks of 20µM and stored at -20°C.  The Daoy cell line (pediatric MB) was seeded (3,000 cells/well) into 96-well plates (BD; Becton Dickinson, Franklin Lakes, NJ, USA) overnight.  The cells were treated with 1µM and 10µM for 72 hrs. Cells were then fixed in 2% paraformaldehyde and stained with Hoechst 33342 nuclear dye (1µg/mL) (Sigma-Aldrich, Oakville, ON, Canada).  They were stored in the dark at 4°C in phosphate-buffered solution (PBS) until analysis with the ArrayScan high-content screening system (HCS; Thermo Fisher Scientific, Pittsburgh, PA, USA).   Cell Culture  Daoy cells (no C-MYC amplification detected; Keles et al., 1995) were obtained from ATCC (Manassas, VA, USA).  Primary brain tumour cells were isolated from BTX001, BT006, BT007, BT008, BT014 and BT274 according to the protocol we previously described (Lenkiewicz et al., 2009).  They were cultured as tumourspheres using Neurocult media (Stem Cell technologies, Vancouver, BC, Canada) (Lee et al., 2012).  All primary cells were obtained through informed consent in abidance with the respective local research ethics board guidelines.  Tumoursphere Assays  A single-cell suspension of Daoy and patient-derived primary MB cells (BTX001 and BT014) were plated in Neurocult medium (1x104 cells per well in 6-well plates) supplemented with 20ng/mL human recombinant EGF, 20ng/mL human recombinant, basic FGF and 2µg/mL heparin (Stem Cell Technologies) on ultra low-binding culture plates (Corning, Massachusetts, USA).  Tumourspheres >50µm were quantified and photomicrographs were taken after 6 days  111 of culture.   Transfection and Immunofluorescence Staining  Small interfering RNA transfections were performed using Lipofectamine RNAiMAX (Invitrogen, Burlington, ON, Canada) as previously described (Hu et al., 2009) using control oligo (sequence: UUC UCC GAA CGU GUC ACG U, Qiagen) and PLK1 siRNA oligo (siRNA sense sequence: CGG GCA AGA UUG UGC CUA A dTdT; Dharmacon, Lafayette, CO, USA).  Primary antibodies used for Western blotting studies include anti-PLK1 (Sigma-Aldrich), anti-P-H2AXS139 antibody (Abcam, Boston, MA, USA), anti-poly(ADP-ribose) polymerase (Cell Signaling Technology), anti-P-CDC25CSer198 (Cell Signaling Technology, Massachusetts, USA) and anti-pan-actin (Cell Signaling Technology).  Immunofluorescence staining was performed on the sections of Daoy tumourspheres according to the procedure we previously described (Fotovati et al., 2011).  Cell Cycle Analysis Medulloblastoma cells were harvested by trypsinization, washed once with cold PBS and fixed in 70% ethanol overnight.  The cells were washed once with cold PBS prior to the addition of staining buffer which was composed of 40µg/mL propidium iodide and 200µg/mL RNase A in cold PBS.  The cells were incubated at room temperature, in the dark, for 30min and 100µl of cold PBS was added directly to the cell suspension when the cells were ready to be analyzed by flow cytometry.   Annexin V Staining and Quantification of Cell Growth by Hoechst Staining  Daoy cells were treated with 2.5nM BI2536 for 48hrs and stained with Annexin V (Promega, Wisconsin, USA) as previously described (Lee et al., 2008a).  To evaluate the effect of PLK1 inhibition on cell growth, Daoy or human neural stem cells (H9, hESC-derived, GIBCO, Burlington, ON, Canada) were plated (3,000 cell/well) in 96-well plates, treated with BI2536 or PLK1 siRNA for 72hrs and stained with Hoechst dye (1µg/mL) in 100µLof PBS containing 2% paraformaldehyde.  The stained cells were kept at room temperature, in the dark, on a rocking platform for 30 minutes.   The plates were analyzed and the images were taken on the ArrayScan VTI Reader (Thermal Fisher, Cellomics, Massachusetts, USA).   112 In vivo Evaluation of BI2536 Compared to Chemotherapy BTIC xenografts from the Daoy MB cell line were injected into the right frontal lobe of NOD-CB17-SCID mouse brains according to Research Ethics Board-approved protocols (n=18).  Mice were injected with biological replicates consisting of 106 single-cell suspensions.  Mice were randomly divided into three treatment groups (n=6) following engraftment of orthotopic tumours: control (0.1N HCl), chemotherapy (vincristine, cisplatin, cyclophosphamide), and treatment (BI2536). Intraperitoneal injections were performed for delivery of all study group agents.  The control group consisted of one weekly injection of 0.1N HCl for four weeks.  The chemotherapy group consisted of one weekly injection of vincristine (1.05µg/kg) and cisplatin (2.5mg/kg) on day 1 followed by cyclophosphamide (0.0352mg/kg) on day 2 for three weeks.  The treatment group consisted of one weekly injection of BI2536 (50mg/mL diluted in 0.1N HCl) for four weeks.  The mice were observed until they displayed obvious signs of neurological deficits and appeared unwell.  Tumours were removed at the end of the study, formalin-fixed, parafin embedded and the presence of tumours was confirmed by hematoxylin and eosin staining.  Statistical Analysis All quantitative data presented were analyzed as mean value ±	  standard error.  For the TMA, clinical survival analysis and animal studies, log-rank analysis was performed on the Kaplan-Meier curve to determine statistical significance of the results.  Multivariate survival analysis was conducted using Cox regression proportional hazards and a 95% confidence interval.  All survival analysis and Spearmans Rank correlation test were done using SPSS version 20.0 statistical software (IBM, Chicago, IL, USA).  The number of samples used and the respective P-values are listed in the figure legends.  The level of significance for the in vitro cell growth/death data was determined by Student’s two-tailed T-test and difference in PLK1 expression between subtypes was assessed using one-way ANOVA (*P value<0.05; **P value<0.01).   113 4.5 TABLES  Table 4.1 Summary of the pediatric MB patients included in the study cohort. All of the children were referred to the British Columbia Children’s Hospital between 1986-2011.   Characteristic: ? ™ WNT ? SH H ? Group 3? Group 4 ? ? Number of Patients:μ 74μ 9 (12%)μ 21 (29%)μ 18 (24%)μ 26 (35%)μAge ? ™ ™ ? <3 yearsμ 20 (27%)μ 0 (0%)μ 12 (57%)μ 7 (39%)μ 1 (4%)μ ? Between 3 to 8 yearsμ 33 (45%)μ 1 (11%)μ 8 (38%)μ 8 (44%)μ 15 (58%)μ ? Between 8 to 12 yearsμ 12 (16%)μ 4 (44%)μ 0 (0%)μ 3 (17%)μ 6 (23%)μ ? >12-18 years μ 9 (12%)μ 4 (44%)μ 1 (5%)μ 0 (0%)μ 4 (15%)μ ? ™ ™ ? Average (years)μ 6.2μ 11.0μ 3.7μ 4.6μ 7.8μ ? Range (years)μ 0.3 to 16.8 μ6.1 to 14.8μ0.3 to 16.8μ1.6 to 10.2μ2.5 to 15.4μ ? ™ ™Sex ? ™ ™ ? Femaleμ 27 (37%)μ 7 (78%)μ 9 (43%)μ 7 (39%)μ 4 (15%)μ ? Maleμ 46 (63%)μ 2 (22%)μ 12 (57%)μ 11 (61%)μ 22 (85%)μ ? ™ ™Metastasis? ™ ™ ? M0μ 39 (53%)μ 5 (56%)μ 12 (57%)μ 8 (44%)μ 14 (54%)μ ? M1μ 14 (19%)μ 1 (11%)μ 3 (14%)μ 5 (28%)μ 5 (19%)μ ? M2μ 4 (5%)μ 0 (0%)μ 2 (10%)μ 1 (6%)μ 1 (4%)μ ? M3μ 10 (14%)μ 1 (11%)μ 1 (5%)μ 3 (16%)μ 5 (19%)μ ? Unknownμ 7 (9%)μ 2 (22%)μ 3 (14%)μ 1 (6%)μ 1 (4%)μ ? ™ ™Extent of Resection ? ™ ™ ? Gross Total Resectionμ 49 (66%)μ 6 (67%)μ 15 (71%)μ 10 (56%)μ 18 (69%)μ ? Subtotal or lessμ 22 (30%)μ 3 (33%)μ 5 (24%)μ 6 (33%)μ 8 (31%)μ ? Unknownμ 3 (4%)μ 0 (0%)μ 1 (5%)μ 2 (11%)μ 0 (0%)μ ? ™ ™Treatment ? ™ ™ ? Chemotherapy Onlyμ 16 (22%)μ 0 (0%)μ 9 (43%)μ 3 (16%)μ 4 (15%)μ ? Radiation Onlyμ 10 (14%)μ 4 (44%)μ 3 (14%)μ 1 (6%)μ 19 (73%)μ ?Both Chemo and Radiationμ 44 (59%)μ 5 (56%)μ 7 (33%)μ 13 (72%)μ 2 (8%)μ ? No Treatmentμ 4 (5%)μ 0 (0%)μ 2 (10%)μ 1 (6%)μ 1 (4%)μTable 1. Summary of the pediatric MB patients included in the study cohort. μ 114   Table 4.2 Univariate and multivariate analyses of clinical, pathological and biological endpoints.   Table 2. Univariate and multivariate analyses of clinical,?pathological and biological endpoints.? ?Variable? No. (n=73)?Log-Rank test (p value)?No. (n=64)?Hazard ratio (95% confidence interval)?Cox regression analysis (p value)?Age?  ?  ?  ?  ?  ?< 3 years? 20? 0.011? 15? 0.422 (0.122 to 1.463)? 0.174 (n.s)? ≥ 3 years? 53?  ? 49?  ? ?  ?  ?  ?  ?Sex?  ?  ?  ?  ?Male? 46? 0.831 (n.s)? 40? 1.211 (0.450 to 3.255)? 0.705 (n.s)?Female? 27?  ? 24?  ? ?  ?  ?  ?  ?Metastasis?  ?  ?  ?  ?Present? 28? 0.043? 25? 3.488 (1.093 to 11.126)? 0.035?Not Present? 39?  ? 39?  ? ?  ?  ?  ?  ?Extent of Resection?  ?  ?  ?  ?Gross Total Resection? 49? 0.601 (n.s)? 45? 0.882 (0.317 to 2.455)? 0.809 (n.s)?Subtotal  Resection or less? 21?  ? 19?  ? ?  ?  ?  ?  ?Radiation?  ?  ?  ?  ?Yes? 53? 0.021? 48? 3.519 (0.345 to 35.888)? 0.288 (n.s)?No? 18?  ? 16?  ? ?  ?  ?  ?  ?Chemotherapy?  ?  ?  ?  ?Yes? 60? 0.965 (n.s)? 53? 3.325 (0.542 to 20.390)? 0.194 (n.s)?No? 13?  ? 11?  ? ?  ?  ?  ?  ?Combined Treatment?  ?  ?  ?  ?Chemo & Radiation? 44? 0.429 (n.s)? 39? 0.661 (0.056 to 7.782)? 0.742 (n.s)?Single or No Treatment? 29?  ? 25?  ? ?  ?  ?  ?  ?PLK1 Transcript?  ?  ?  ?  ?High? 30? 0.017? 26? 3.673 (1.426 to 9.460)? 0.007?Low? 43?  ? 38?  ?  ?n.s = not significant  115 4.6 FIGURES   Figure 4.1 Molecular characteristics of MB dictate outcome and offer potential drug targets. (A-B) Patients with the SHH subtype of MB had the highest rates of relapse and the worst chance of overall survival, respectively. (C) The patients as well as four MB cell lines were subtyped into the different catagories based on gene expression and represented using a heatmap (Red= high expression, Green= low expression).  (D) A library of 129 small molecule inhibitors was screened against Daoy cells in a 72hr growth assay. A. SHH?Group 4?WNT?Time to LFU (years)25 . 02 0 . 015 . 01 0 . 05 . 0. 0Cum Survival1 . 00 . 80 . 60 . 40 . 20 . 0Medulloblastoma Subtype and Survival Group 4 -censoredGroup 3 -censoredSHH-censoredWNT -censoredGroup 4Group 3SHH WNTMedulloblastoma SubtypePage 1Time in years?Log rank P=0.010?Breslow P=.004?0.0? 5.0? 10.0? 15.0? 20.0? 25.0?1.0?0. ?0. ?0. ?0.2?0. ?Group 3?Cumulative Survival?0?20?40?60?80?100?120?DMSO?1uM LBH-589 (Panobinostat)?10uM LBH-589 (Panobinostat)?1uM AZD1152?10uM AZD1152?1uM BI 2536?10uM BI 2536?1uM BI 6727 (Volasertib)?10uM BI 6727 (Volasertib)?1uM GSK461364?10uM GSK461364?1uM ARQ-197 (Tivatinib)?10uM ARQ-197 (Tivatinib)?1uM AZD8055?10uM AZD8055?1uM INK128?10uM INK128?Cell Viability (%)?PLK1 inhibitors?HDAC?PLK1?c-MET?mTOR?AURK?C. B. Figure 1?!"#$%&'$()*+ !"#$!$#$%"#$%$#$"#$#$,)-.(."/"0'%-1%2$/(3*$41)$$%56)7"7(/%#$$#&$#'$#($#!$#$Medulloblastoma Subtype on Relapse-free SurvivalPage 1SHH?Group 4?WNT?Group 3?0.0? 5.0? 10.0? 15.0? 20.0? 25.0?Time in years?1.0?0.8?0.6?0.4?0.2?0.0?Relapse-free Survival?Log rank P=0.048?Breslow P=0.021?D. D. SHH WNT GROUP 3 GROUP 4 ONS76 UW228 UW426 DAOY -3   0  3 116   Figure 4.2 PLK1 expression correlates with poor patient survival. (A) NanoString nCounter analysis of 74 patient samples (blue bars) shows PLK1 mRNA is overexpressed in MB compared to normal cerebellum (red line).  (B) Representative examples of PLK1 protein expression in normal cerebellum compared to MB (20x magnification).  (C-D) High PLK1 mRNA expression predicts probability of patient relapse (P=0.004) and poor overall survival (P=0.003).  PLK in Tumour Tissue ?PLK in Normal Brain ?PLK1 Expression in MB compared to Normal Brain?A. PLK1 Gene Expression?0?500?1000?1500?2000?Medulloblastoma Patients?B. !"#$%&'$()*+!"#$!$#$%"#$%$#$"#$#$,)-.(."/"0'%-1%)$#("2"23%1)$$%-1%)$/(4*$%#$$#&$#'$#($#!$#$,567%894)$**"-2%,)$:";0*%,(0"$20%<$/(*4$%Page 1Low PLK1?Medium PLK1?High PLK1?Log rank P = 0.004?Breslow=0.001 ?PLK1 Expression Predicts Patient Relapse?25.0?Time in years?0.0? 5.0? 10.0? 15.0? 20.0?1.0?0.8?0.6?0.4?0.2?0.0?Probably for Relapse-Free Survival?Normal Cerebellum?Medulloblastoma?!"#$%&'$()*+!"#$!$#$%"#$%$#$"#$#$,-).".(/%#$$#&$#'$#($#!$#$0123%456)$**"78%0)$9":;*%0(;"$8;%,-).".(/Page 1High PLK1?Log rank= 0.003?Breslow P=0.003?Cumulative Survival?Medium PLK1?Low PLK1?PLK1 Expression Predicts Patient Survival?D. C. Figure 2 Time in years?0.0? 5.0? 10.0? 15.0? 20.0?1.0?0.8?0.6?0.4?0.2?0.0?25.0? 117  Figure 4.3 The response of primary MB cells to BI2536 is correlated with the expression of PLK1 in the tumours. (A) PLK1 mRNA from the cerebellum tissue (CB), human neural stem cells (hNSC) and a patient-derived MB sample BTX001 (molecularly classified as SHH, please refer to Supplementary Table S4.1, sample “DUNN-102A.”) (B) BI2536 (10nM) inhibited the self-renewal of BTX001 (6 day exposure) (scale bar=140µm). (C) PLK1 mRNA in the CB, hNSC and Figure 3 B. DMSO? BI2536?P1?P0?0?10?20?30?40?50?* P0?P1?Total Number of ?Tumourspheres ?(per well)?DMSO? BI2536?Relative PLK1 Transcript Level ?(fold difference)?A. 0?1?2?3?4?5?6?7?8?CB? hNSC? BTX001?0?50?100?150?Number of Tumourspheres? (per well)? P0?P1?DMSO   10       20     (nM) BI2536 DMSO? 10nM? 20nM?BI2536 D. P1?C. 0?1?2?3?4?5?CB? hNSC? BT014?Relative PLK1 Transcript Level ?(fold difference)?F. 0?50?100?150?200?250?DMSO? 1? 5? 10?BI2536?(nM)?Total Cell Number? (per random field counted)?E. 0?50?100?150?200?HA?BTX001?BT006?BT007?BT274?Daoy?BT008?Relative PLK1 Transcript Level (fold difference)?Primary Medulloblastoma ?Supratentorial PNET ?SHH SHH  118 BT014, (a patient-derived MB with low PLK1 levels). (D) BT014 cells were not responsive to 10nM BI2536 (6 day treatment) (scale bar=140µm). (E) hNSCs were treated with 1, 5 and 10nM of BI2536 for 72hrs, stained with Hoechst dye. BI2536, even at the highest dose tested (10nM), had a negligible impact on their growth. (F) PLK1 mRNA levels in freshly isolated patient-derived MB samples (BTX001, BT006, BT007, BT008, BT274), relative to normal human astrocytes (HA).  All mRNA levels were quantified by qRT-PCR. BT006 and BT007 were molecularly classified as SHH (Supplementary Table S4.1). All these studies in Figure 4.3 were performed in replicates on one occasion due to the scarcity of study material.  119    Figure 4.4 PLK1 inhibition suppresses cell growth and induces apoptosis in Daoy cells. (A) Daoy cells were treated with 5nM of PLK1 siRNA for 72hrs, knockdown was confirmed by immunoblotting (top) and cell growth was assessed by Hoechst staining (bottom). The Western blotting was repeated on three separate occasions.  (B) Daoy cells were treated with 0.5-100nM BI2536 for 24, 48 and 72hrs then growth was assessed by Hoechst staining. The cell growth Figure 4?DMSO ? BI2536 ?  Control ? ? siPLK1 ?B. A. DMSO ? BI2536 ?%G1=26.7 ?%S=53 ?%G2=18.5 ?%G1=7.16 ?%S=25.6 ?%G2=53.2 ?0?1?2?3?4?5?6?0h? 24h? 48h? 72h? (hrs)?*  * *  * *  Relative Growth (fold change)?D MSO 0.5nM 1nM 5nM 10nM 50nM 100nM BI2536 DMSO?1? 2.5?5?P-CDC25C Ser198?Actin?BI2536 (nM) ?19.1?4.14?54.5?32?14.4?3.2?30.6?16.8?7AAD ?Annexin V-PE ?PLK1 ?Vinculin ?Control? ? siPLK1 ?0?0.2?0.4?0.6?0.8?1?1.2?Control Oligo?siPLK1 ?Relative Cell Growth ?     (fold change)  ?* C. D. E. Control? siPLK1 ? DMSO ? BI2536 ?P-H2AX Ser139?Actin?PARP ?PARP (cleaved)?Control?siPLK1?DMSO?BI2536?F. 90 kDa ?16 kDa ?45 kDa ? 120 assay was performed in triplicates on two separate occasions. The drug-treated cells were also subjected to immunoblotting to show a decrease in the phosphorylation of CDC25C, which is a known downstream substrate of PLK1.  (C) Daoy cells were treated with PLK1 siRNA (5nM) or BI2536 (2.5nM) for 72hrs, and the viable cells were visualized by crystal violet staining. (D) Daoy cells were treated with 5nM BI2536 for 24hrs, subjected to flow cytometry for analysis of cell cycle profile with propidium iodide. The study was repeated on two separate occasions.  (E-F) Daoy cells were treated with 5nM BI2536 for 48hrs. Apoptosis was measured by Annexin V-PE/7AAD staining (n=2) or by immunoblotting (n=3) for PARP cleavage and P-H2AXSe139.   121   Figure 4.5 PLK1 inhibition suppresses tumoursphere formation in vitro and delayed disease progression in vivo. (A) Daoy tumourspheres were stained for PLK1, SOX2, musashi and Bmi1. The nuclei were visualized by Hoechst staining. (B) BI2536 inhibited the self-renewal of Daoy tumourspheres A. DMSO?010?20?30?40?50?60?DMSO? ?5? ?10   (nM)?Total Number of Tumourspheres ?(per well)?BI2536?*?*?5?? 10    (nM)?BI2536?B. Hoechst? PLK1? SOX2? SOX2?PLK1?Merged, All?Musashi? Musashi?Bmi1? Bmi1?Hoechst? PLK1?PLK1?Merged, All?Hoechst? PLK1?PLK1?Merged, All?C. Fraction Survival?BI2536?Chemotherapy?Control?P=0.0285?(Days)?0.0?0.8?0.2?1.0?0.6?0.4?0 20? 40? 60? 80?0?0.2?0.4?0.6?0.8?1?1.2?Relative Cell Growth (fold change)?0 ? 1 ? 2 ? 0 ? 50 ?100 ? 0 ?0.25 ?0.5 ?Vincristine ?(nM)?Cisplatin ?(ng/ml)?Etoposide ?(μM)?D.  122 when treated with 5 or 10nM BI2536 for 6 days. The experiment was performed in duplicates on two separate occasions. (C) Daoy cells were treated with vincristine, cisplatin or etoposide for 72hrs and stained with Hoechst for quantification. The study was conducted in triplicates and repeated twice. (D) Daoy cells were transplanted intracranially into mice and treated with BI2536 (n=6) or chemotherapy (n=6) according to a Headstart schedule and compared to control (n=6). BI2536-treated mice lived longer then controls (P=0.0142, Log rank) and was comparable to the benefit observed from chemotherapy (P=0.0336). No significant difference in survival was observed between BI2536- and chemotherapy-treated mice (P=0.4205).    123 4.7 SUPPLEMENTARY DATA  Table S4.1 PAM class prediction validation for MB subgroup assignment Sample Labels Predicted Class WNT SHH Group 3 Group 4 DUNN-101A SHH 1.0E-09 1.0E+00 4.2E-07 3.7E-07 DUNN-102A SHH 5.3E-07 1.0E+00 6.2E-06 4.9E-06 DUNN-107A Group4 4.7E-07 7.2E-06 2.4E-01 7.6E-01 DUNN-109A Group3 2.1E-05 1.2E-04 9.9E-01 8.9E-03 DUNN-113A Group4 2.9E-04 1.2E-03 2.3E-02 9.8E-01 DUNN-114A Group4 1.1E-07 1.8E-05 8.5E-03 9.9E-01 DUNN-115A SHH 4.0E-08 1.0E+00 9.0E-06 3.3E-06 DUNN-117A Group3 1.5E-07 8.1E-07 8.6E-01 1.4E-01 DUNN-118A Group4 3.3E-06 1.0E-05 9.2E-02 9.1E-01 DUNN-120A Group4 5.2E-04 1.7E-03 1.1E-02 9.9E-01 DUNN-121A Group3 6.6E-08 3.9E-05 1.0E+00 2.1E-03 DUNN-122A Group3 3.0E-05 6.8E-06 9.9E-01 5.7E-03 DUNN-123A Group4 4.6E-08 4.7E-06 5.0E-02 9.5E-01 DUNN-124A SHH 4.3E-10 1.0E+00 6.7E-07 3.3E-07 DUNN-125A Group4 3.8E-07 4.0E-05 5.4E-03 9.9E-01 DUNN-126A Group4 4.5E-05 1.5E-03 6.1E-02 9.4E-01 DUNN-127A Group3 2.2E-05 3.1E-04 9.8E-01 1.8E-02 DUNN-130A Group3 4.0E-05 3.3E-04 7.9E-01 2.1E-01 DUNN-131A Group4 1.3E-07 2.3E-04 1.7E-01 8.3E-01 DUNN-132A Group4 4.8E-08 1.5E-05 8.7E-04 1.0E+00 DUNN-133A Group3 1.1E-04 2.4E-05 9.9E-01 1.1E-02 DUNN-135A Group3 2.8E-03 1.9E-04 9.4E-01 5.8E-02 DUNN-138A SHH 2.9E-08 1.0E+00 3.3E-06 3.2E-06 DUNN-139A WNT 1.0E+00 8.9E-06 3.5E-05 3.9E-06 DUNN-140A Group4 4.6E-05 5.4E-05 4.8E-01 5.2E-01 DUNN-142A WNT 1.0E+00 2.5E-07 9.8E-07 2.3E-07 DUNN-143A Group4 9.6E-09 1.8E-05 1.9E-03 1.0E+00 DUNN-145A Group4 7.5E-04 3.1E-04 1.4E-01 8.6E-01 DUNN-146A SHH 1.8E-05 6.1E-01 3.5E-01 3.5E-02 DUNN-147A Group3 5.5E-05 1.7E-04 1.0E+00 2.4E-03 DUNN-149A WNT 1.0E+00 4.9E-05 5.0E-05 2.1E-04 DUNN-150A Group3 6.2E-06 5.4E-05 5.3E-01 4.7E-01 DUNN-151A Group4 5.6E-08 3.9E-05 2.0E-03 1.0E+00 DUNN-152A Group4 2.8E-03 4.2E-04 8.7E-02 9.1E-01 DUNN-153A Group4 4.1E-06 1.3E-04 1.4E-02 9.9E-01 DUNN-154A Group4 3.4E-06 2.5E-05 1.0E-01 9.0E-01 DUNN-155A Group3 8.6E-07 1.2E-06 9.9E-01 1.0E-02 DUNN-156A Group4 2.5E-03 8.0E-04 9.4E-02 9.0E-01 DUNN-157A SHH 1.1E-07 1.0E+00 8.0E-06 8.6E-06 DUNN-184A WNT 1.0E+00 2.2E-09 8.9E-13 2.2E-11 DUNN-185A SHH 2.3E-14 1.0E+00 7.2E-09 7.3E-08 DUNN-186A SHH 1.4E-12 1.0E+00 3.6E-07 8.3E-06 DUNN-104A Group4 1.8E-15 3.0E-11 2.6E-05 1.0E+00 DUNN-105A Group4 6.8E-11 1.2E-06 5.5E-02 9.4E-01 DUNN-106A SHH 1.7E-14 1.0E+00 5.2E-08 1.6E-05 DUNN-108A Group4 5.3E-12 2.8E-08 2.1E-03 1.0E+00 DUNN-109A Group3 2.3E-08 1.0E-07 1.0E+00 4.8E-04 DUNN-110A SHH 6.8E-13 1.0E+00 3.6E-07 1.2E-06 DUNN-112A SHH 4.8E-09 1.0E+00 1.6E-05 8.5E-04 DUNN-136A Group4 5.2E-09 5.3E-06 4.9E-02 9.5E-01 DUNN-158A Group4 6.8E-14 1.3E-09 1.9E-04 1.0E+00 DUNN-159A Group4 1.1E-14 6.1E-11 2.5E-03 1.0E+00 DUNN-160A Group3 1.9E-12 1.6E-10 1.0E+00 2.6E-04 DUNN-161A SHH 3.1E-12 1.0E+00 1.2E-08 3.2E-08 DUNN-162A Group3 7.8E-12 1.0E-10 8.8E-01 1.2E-01 DUNN-163A Group3 2.5E-10 7.3E-09 1.0E+00 2.2E-03 DUNN-164A WNT 1.0E+00 1.1E-10 2.1E-09 5.2E-11 DUNN-165A Group4 7.0E-13 1.8E-09 7.9E-02 9.2E-01 DUNN-166A WNT 1.0E+00 5.0E-12 3.2E-11 1.2E-11 DUNN-167A SHH 2.5E-02 9.7E-01 9.7E-04 4.3E-03 DUNN-168A SHH 2.8E-14 1.0E+00 1.4E-07 6.2E-06 DUNN-169A Group3 3.2E-13 1.6E-11 1.0E+00 1.4E-03 DUNN-170A SHH 5.5E-10 1.0E+00 8.5E-07 1.6E-05 DUNN-171A SHH 3.0E-15 1.0E+00 2.7E-09 9.9E-08 DUNN-173A Group3 5.0E-10 1.9E-08 9.9E-01 1.0E-02 DUNN-174A Group3 3.2E-11 1.1E-09 1.0E+00 8.0E-04 DUNN-175A WNT 1.0E+00 1.8E-10 2.8E-09 6.6E-11 DUNN-177A SHH 6.4E-04 1.0E+00 1.4E-04 2.9E-03 DUNN-178A WNT 1.0E+00 2.6E-11 3.9E-09 5.7E-10 DUNN-179A WNT 1.0E+00 1.0E-07 3.1E-10 1.3E-09 DUNN-180A SHH 7.8E-15 1.0E+00 1.0E-10 2.8E-09 DUNN-181A SHH 9.4E-14 1.0E+00 1.6E-10 4.3E-08 DUNN-182A Group4 3.8E-12 1.7E-07 2.2E-04 1.0E+00 DUNN-183A SHH 1.8E-04 9.9E-01 3.7E-04 8.9E-03 Daoy SHH 1.2E-05 1.0E+00 1.4E-05 6.4E-05 ONS76 SHH 1.1E-08 1.0E+00 1.0E-04 2.7E-04 UW228 SHH 2.3E-08 1.0E+00 1.5E-04 5.0E-04 UW426 SHH 3.7E-09 1.0E+00 2.4E-05 2.1E-05 Supplementary Table 1? 124   Figure S4.1 PLK1 is a tumour-specific protein that is highest in the SHH subtype.   (A) The average of all MB patient samples were significantly higher then the normal cerebellum.  (B) Across the four MB subtypes, PLK mRNA was highest in the SHH type. As well, (C) SHH classified tumours had a greater frequency of high PLK1 and fewer low expressing tumours compared to the distribution of cases in other subgroups.  Supplementary Figure 1?A. Average Gene Expression?Normal ?Cerebellum?Medulloblastoma?0?200?400?600?800? P<0.001   PLK1 Transcript Expression ?(Gene Count)?Medulloblastoma Subtype?0?500?1000?1500?2000?P=0.0485?WNT? SHH? Group 3?Group 4?B. C. 010203040SubtypeNumber of patientsLow PLK1Medium PLK1WNTSHH Group 3 Group 4High PLK1 125  Figure S4.2 AURKA expression in MB. (A) AURKA and PLK1 have a significant positive correlation in gene expression (R=0.74, P=7E-14). (B-C) NanoString nCounter analysis of 74 patient samples (green bars) shows mRNA is over-expressed in MB compared to normal cerebellum (red line). A. B. PLK1 Gene ExpressionAURKA Gene Expression0 500 1000 1500 20000200400600800Spearmans Correlation = 0.740 p = 7E-14  Average Gene ExpressionNormal Cerebellum Medulloblastoma  0100200300400P<0.001  C. AURKA Gene Expression?Medulloblastoma Patients?200?400?600? 126    Figure S4.3 Patient relapse and outcome in children over the age of 3 years old.  (A-B) The SHH subtype was associated with increased rates of relapse and poor survival. (C-D) PLK1 mRNA was also associated with more rapid rates of relapse and poor survival.  Outcome for patients >3 years of age (N=53) ?!"#$%&'%()*%+,$-./0!"#$!$#$%"#$%$#$"#$#$1.'2-2"3"&,%'4%5$3-6/$74.$$%89.:":-3%#$$#&$#'$#($#!$#$Level s  o f  P L K1 in f luen ce rel a p se - f ree surv iv a l  in p a t ient s  > 3  ye a r s  o f  a gePage 1Time to LFU (years)!"#$!$#$%"#$%$#$"#$#$!"#$%$&'%#$$#&$#'$#($#!$#$Levels of PLK1 influence survival in patients >3 years of agePage 1Log Rank = 0.044 Breslow= 0.012 Log Rank = 0.005 Breslow= 0.002 Low PLK1?Medium PLK1?High PLK1?Low PLK1?Medium PLK1?High PLK1?Time to LFU (years) of Relapse-free Survival1. of relapse-free survival in patients >3 years of aged based on subtypePage 1Log Rank = 0.039 Breslow= 0.020 Group 4 Group 3 WNT SHH Time to LFU (years) of survival in patients >3 years of age based on subtypePage 1Log Rank = 0.014 Breslow= 0.006 Group 4 Group 3 WNT SHH A. B. C. D.  127  Figure S4.4 Only children who received radiation were included to avoid this treatment as a confounding variable. (A-B) The SHH subtype was associated with early relapse and worst overall survival. (C-D) PLK1 was also significantly linked to poor patient outcomes where relapse was more immediate as was death.   Outcome considering only patients that received radiation treatment (N=53) ?!"#$%&'%()*%+,$-./0!"#$!$#$%"#$%$#$"#$#$Probability of Relapse-free Survival%#$$#&$#'$#($#!$#$($1$2/%'3%4(56%"7328$79$%.$2-:/$;3.$$%/8.1"1-2%'3%:-&"$7&/%.$9$"1"7<%.-="-&"'7Page 1!"#$%&'%()*%+,$-./0!"#$!$#$%"#$%$#$"#$#$Survival%#$$#&$#'$#($#!$#$Levels of P L K 1  influence overall surival of patients re ceivin g  radiationPage 1Log Rank = 0.003 Breslow= 0.001 Log Rank = 0.001 Breslow= 0.001 Low PLK1?Medium PLK1 ?High PLK1?Low PLK1?Medium PLK1 ?High PLK1?Time to LFU (years) of Relapse-free Survival1. of relapse-free survival in only patients receiving radiationPage 1Log Rank = 0.036 Breslow= 0.017 Group 4 Group 3 WNT SHH Time to LFU (years) of survival in only patients receiving radiationPage 1Log Rank = 0.042 Breslow= 0.018 Group 4 Group 3 WNT SHH A. B. C. D.  128  Figure S4.5 PLK1 protein levels were prognostic for relapse. (A) and poor survival (B) based on Log-Rank and Breslow analyses.   Log Rank P= 0.018?Breslow P= 0.005?Time from Diagnosis to Death (years)?Cumulative Overall Survival?0.0?0.2?0.4?0.6?0.8?1.0?0.0? 5.0? 10.0? 15.0? 20.0? 25.0?PLK1 low?PLK1 high?B. Log Rank P= 0.030?Breslow P= 0.023?Time from Diagnosis to Death (years)?Cumulative Event-Free Survival?0.0?0.2?0.4?0.6?0.8?1.0?0.0? 5.0? 10.0? 15.0? 20.0? 25.0?PLK1 low?PLK1 high?A.  129    Figure S4.6 The response of MB cell lines to BI2536 and chemotherapeutic agents. (A-B) PLK1 protein and mRNA were assessed in Daoy, ONS76 and UW426 MB cell lines. (C-D) The ONS76 and UW426 cells were sensitive to BI2536 as it inhibited growth in a time and dose-dependent manner. (E-F) The ONS76 and UW426 cells responded to chemotherapeutic agents- vincristine, cisplatin and etoposide in 72hr cell growth assays.   PLK1?Actin?ONS76?UW426?A.  Daoy?0?0.5?1?1.5?2?Daoy? ONS76? UW426?Relative PLK1 Transcript ?Level (fold difference)?B. ONS76?C. 0?100?200?300?400?500?600?0? 20? 40? 60? 80?DMSO?1nM?2.5nM?5nM?7.5nM?10nM?Total Number of Cells ?(per random field counted)?Time in Hours?UW426?0?100?200?300?400?500?600?0? 50? 100?DMSO?1nM?2.5nM?5nM?7.5nM?10nM?Total Number of Cells ?(per random field counted)?D. Time in Hours?0?0.2?0.4?0.6?0.8?1?1.2?Relative Cell Growth (fold change)?0 ?1 ?2 ?0 ?0.5 ?1 ?0 ?200 ?500 ?Vincristine ?(nM)?Cisplatin ?(ng/ml)?Etoposide ?(μM)?0?0.2?0.4?0.6?0.8?1?1.2?Relative Cell Growth (fold change)?0 ?1 ?2 ?0 ?0.5 ?1 ?0 ?300 ?500 ?Vincristine ?(nM)?Cisplatin ?(ng/ml)?Etoposide ?(μM)?ONS76?E. UW426?F.  130   Figure S4.7 Daoy xenografts were similar to large cell anaplastic medulloblastoma. For example, large cells (A), cell wrapping (B), and high mitotic activity was evident (C).  The tumour cells were highly invasive into the white matter (D), cerebellum (E) and along the leptomeninges (F). The magnifications for the images are as follows: Figure (A), (B), (D) and (F) are 400x and Figure (C) and (E) are 200x.   A. B. Cell “w ra p ping ”?Classic large cells ?C. D. Invasive cells undergoing ?mitosis ?High mitotic index ?F. E. Tumour in subarachnoid space ?Over cerebellum ?Leptomeningeal spread ? 131 CHAPTER 5: DISCUSSION 5.1 SUMMARY  The objective of the studies presented herein was to investigate PLK1 as a molecular target for the treatment of brain tumours.  The fact that there is currently a lack of effective molecular targeted agent against brain tumours in general prompted us to seek a novel therapy.  The idea of targeting PLK1 stemmed from a previously conducted genome-wide siRNA library screen in which we demonstrated this kinase as one of the lead targets in a panel of pediatric cancer cell lines including rhabdomyosarcoma, Ewing’s sarcoma, neuroblastoma and GBM (Hu et al., 2009).  Silencing PLK1 expression resulted in remarkable growth suppression (~80%) and induced apoptosis in the cancer cells in 72hrs.  The initial promising results in a GBM cell line encouraged us to examine this further in brain cancer where we addressed PLK1 as a potential molecular target in a broad range of GBM and MB models.  The body of work presented in Chapters 2, 3 and 4 of this thesis consistently showed that the inhibition of PLK1 by siRNA or a small molecule inhibitor eliminated GBM and MB cells in vitro and that BI2536 prolonged the survival of animals bearing these tumours.  Specifically, we provided pre-clinical, in vitro evidence suggesting that molecular targeting of PLK1 may help overcome TMZ resistance (Chapter 3) and prevent the expansion of BTICs (Chapter 2), which are believed to be one of the causes of disease relapse.  The rationale for targeting this kinase in a clinical setting was further supported by the differential expression of PLK1 in normal and malignant brain tissues as well as the unequivocal correlation between high PLK1 expression and poor prognosis in both GBM and MB (Chapter 2 and 4).   In Chapter 2, we examined the biological effects of PLK1 inhibition on GBM cell lines, patient-derived BTIC cells as well as minimally-cultured primary brain tumour cells.  We first showed that PLK1 expression was elevated in BTICs and brain cancer cell lines compared to normal human astrocytes and human neural stem cells.  Furthermore, high PLK1 expression was correlated with poor patient survival.  Next, we demonstrated that PLK1 inhibition by way of siRNA or small molecule inhibitor led to G2/M cell cycle arrest, suppression of cell growth and induction of apoptosis in the pediatric GBM cell line SF188.  Intriguingly, the cells treated with PLK1-targeting siRNA or BI2536 assumed a stellate, astrocytic cellular morphology with a concomitant increase in GFAP transcript levels and showed a reduction in the expression of neural stem cell markers SOX2 and musashi.  In line with this, adult primary GBM cells treated with BI2536 either died or became restricted in their capacity to divide clonally and form tumourspheres.  Finally, BI2536 prolonged the survival of mice bearing orthotopic GBM tumours.  132 In Chapter 3, we sought to identify off-patent compounds that could be re-purposed for the treatment of GBM.  For the first time, disulfiram, a small molecule inhibitor used to control alcoholism for the past 60 years, was demonstrated to inhibit the self-renewal of TMZ-resistant pediatric GBM cell line SF188 and/or primary adult BTIC cultures BT74, GBM4, aBT001 and aBT003.  The point of intersect between this study and the PLK1 work in the thesis is that disulfiram treatment not only suppressed the expansion of TMZ-resistant cells but also curiously decreased the total expression of PLK1.  We subsequently showed that TMZ-resistant BTICs were also sensitive to PLK1 inhibition.  Specifically, BI2536 treatment decreased the size and number of tumourspheres formed.  Therefore, disulfiram might inhibit the growth of TMZ-resistant GBM cells in part through down-regulation of PLK1 and it has the potential to be re-purposed for the treatment of TMZ-refractory GBM. In Chapter 4, we explored the idea of targeting PLK1 as a potential therapeutic strategy for the treatment of SHH MB, a specific subgroup of MB commonly seen in infants and young children.  Using the state-of-art Nanostring technology, we quantitatively analyzed the transcript levels of PLK1 in archival MB tissues collected from 1986-2012 at the BC Children’s Hospital and found that PLK1 was expressed at high levels compared to the normal cerebellum in almost all the samples examined.  In addition, high-PLK1 expression was more frequently found in the SHH MB compared to the other subgroups and high levels were significantly associated with poor overall and relapse-free survival.  Inhibition of PLK1 by siRNA or BI2536 suppressed cell growth and induced cell death in Daoy cells, which is the model of SHH MB in the study.  The observations were subsequently confirmed in primary, patient-derived MB cells in which the response to BI2536 was found to be associated with the amount of PLK1 in cells.  Finally, using the Daoy xenograft model, we demonstrated that the mice treated with BI2536 had an equivalent length of survival time compared with those treated with chemotherapy and that the Daoy tumours showed representative histologic features of human MB with characteristic leptomeningeal spread.  Summarizing the results from Chapter 2, 3, 4 and the studies by others, we proposed a model (Figure 5.1) to explain why PLK1 inhibition may be helpful in brain cancer treatment.  133  Figure 5.1 Model explaining the potential therapeutic benefit of PLK1 inhibitors in brain cancer treatment. (A) PLK1 inhibitors may target the rapidly-proliferating progenitors as well as BTICs, which may help debulk tumours and prevent disease relapse, respectively. (B) TMZ-resistant cells were shown to be sensitive to BI2536.  Therefore, PLK1 inhibition may be an alternative treatment option for TMZ-refractory brain tumours. (C) PLK1 inhibitors have been shown to work synergistically with radiation in eliminating MB brain cancer cells (Harris et al., 2012).  5.2 IMPLICATIONS AND FUTURE DIRECTIONS 5.2.1 PLK1 INHIBITION SUPPRESSES THE SELF-RENEWAL OF GBM BTICs  In Chapter 2, we demonstrated that PLK1 inhibition suppressed the expansion of patient-derived adult BTICs.  GBM4 and GBM8 cells died and no primary tumourspheres were formed following BI2536 treatment.  In contrast, BT74 cells required serial drug treatment to deplete the tumoursphere formation (Figure 2.2D).  This may be explained in part by the highly elevated level of PLK1 in BT74 compared to GBM4 and GBM8.  Therefore, a higher drug concentration may be needed to stochiometrically inhibit the higher amount of PLK1 present in BT74 cells.  In addition, it is possible that rather than inducing cell death, BI2536 may have targeted the self-renewal and asymmetric division of BT74, which led to a reduction of the total number, as well as the size, of the tumourspheres after serial passaging.   Brain tumour initiating cells or “cancer stem cells” (BTICs/CSCs) Progenitor cells Differentiated cells PLK1 inhibitors PLK1 inhibition may eliminate  1) the fast-proliferating progenitors [debulk tumour]  2) BTICs/CSCs [prevent disease recurrence] PLK1 inhibitors PLK1 inhibitors + Radiation TMZ-resistant cells TMZ-sensitive cells PLK1 inhibition may target brain cancer cells that are refractory to TMZ [overcome drug resistance]  (A) (B) (C) PLK1 inhibitors may work synergistically with radiation in killing brain cancer cells  [enhance therapeutic effects of RT]   134  Molecular characterizations of GBM4, GBM8 and BT74 (Cheema et al., 2011) revealed that GBM4 and BT74 are p53-mutated while GBM8 has wild-type p53.  Ras and p53 mutations have been reported to be the predictors of sensitivity to PLK1 inhibitors (Degenhardt et al., 2010; Luo et al., 2009).  However, the lack of correlation between p53 status and sensitivity to BI2536 in our study suggests that additional factors may play a role.  We have collected over 20 pediatric brain tumours in the past 2.5 years and tested the effects of BI2536 or BI6727 on any sample we were able to culture (6 of out 24 samples; Appendix A).  The majority of the primary brain tumour cultures we examined were sensitive to PLK1 inhibition although a few of them (eg. BT011- pediatric GBM; BT014- pediatric MB; BT022- relapsed pediatric GBM) were not.  Interestingly, the cells that were completely resistant to BI2536 or BI6727 were found to express negligible level of PLK1.  Therefore, based on our experience, the responsiveness of brain cancer cells to PLK1 inhibitors appears to be correlated to the amount of PLK1 in cells.  The result has multiple implications.  It suggests that the biological effects observed after BI2536 treatment can be unequivocally attributed to PLK1 inhibition and are not a consequence of non-specific cytotoxicity.  The dramatic growth suppressive and apoptotic effects in PLK1-inhibited cells indicate that we may have targeted the “Achilles’ heel” of brain tumour cells, which show “oncogene addiction” to this kinase.  Applying the inhibitor to tumour cells with very low abundance of PLK1 would be analogous to shooting arrows into thin air, where the “target” is largely absent.  Therefore, the modest efficacy of PLK1 inhibitors in recent clinical trials (Mross et al., 2012; Schoffski et al., 2010b; Sebastian et al., 2010b) may be explained by the lack of patient stratification in the design of those studies.  As a result, the future clinical usage of PLK1 inhibitor will require routine assessments of PLK1 levels in each tumour tissue prior to the treatments.  The “low-PLK” in our study is loosely defined as a PLK1 level less than or equivalent to that measured in the normal brain tissue counterpart, which is known to express minuscule amount of the protein (Cheng et al., 2012; Winkles and Alberts, 2005).  We postulate that patients having tumours with such low level of PLK would be refractory to the PLK1 inhibitors regardless of p53 and Ras mutational status.  However, genetic analyses on p53 and Ras should be performed on tumours that express higher levels of PLK1 as the results may still provide predictive value to the responsiveness to the therapy.  We showed that PLK1 inhibition decreased the expression of neural stem cell markers such as SOX2 (Figure 2.4B-C).  We chose to study SOX2 further because this protein is not only a stem cell marker but may also be functionally involved in the pathogenesis of cancers.  Studies indicate that SOX2 may play a role in the self-renewal, proliferation, migration/invasion and survival of cancer cells.  Silencing SOX2 halted the proliferation and hindered the  135 tumourigenic capacity of GBM tumour-initiating cells (TICs) in mice (Gangemi et al., 2009).  Similarly, this stem cell marker is also detected in lung and breast tumours where over-expression maintains stemness and confers tumourigenicity of TICs (Leis et al., 2012; Nakatsugawa et al., 2011).  Interestingly, SOX2 may mediate epithelial-to-mesenchymal transition (EMT), cell migration and invasion in part through matrix metalloproteinase MMP2 (Han et al., 2012; Xu et al., 2013).  Furthermore, SOX2 has been shown to up-regulate survivin expression (Lin et al., 2012) to enhance the survival of neural stem cells (Feng et al., 2013).  The apoptotic effect we observed in SOX2-silenced SF188 cells might therefore be a consequence of reduced survivin expression. The wide-range of biological effects that SOX2 induces in cancer cells motivated us to investigate the regulation of its expression by PLK1.  PLK1-mediated phosphorylation of peptidyl-prolyl cis/trans isomerase 1 (Pin1) was shown to increase the protein stability of Pin1 by inhibiting its ubiquitination (Eckerdt et al., 2005).  In a separate study, Pin1 was reported to phosphorylate and activate STAT3 (Lufei et al., 2007).  We were interested in STAT3 as a potential regulator of SOX2 expression because this transcription factor is required for the proliferation and maintenance of multipotency in GBM stem cells (Sherry et al., 2009).  Moreover, STAT3 was shown to regulate SOX2 expression in neural precursor cells and glioblastoma stem cells (Foshay and Gallicano, 2008; Guryanova et al., 2011).  Piecing the information together, we proposed a hypothetical signaling pathway- PLK1/Pin1/STAT3/SOX2 in which PLK1 increases the protein stability of Pin1, which in turn activates STAT3, leading to the expression of SOX2.  However, we did not observe an alteration in the phosphorylation of STAT3 at Ser727 after PLK1 inhibition (Appendix B).  SOX2 expression was also unaffected by STAT3 knockdown.  Therefore, STAT3 is not the major regulator of SOX2 expression in our model.  We further examined the involvement of the MAPK pathway because we observed a dramatic decrease in the phosphorylation of ERK1/2 after PLK1 inhibition.  In a recent study, SOX2 was found to be a target gene of the EGFR/Src/AKT signaling in the stem-like side-population cells of NSCLC (Singh et al., 2012).  Thus we questioned whether transcription factors activated by the MAPK pathway would control SOX2 expression.  The MAPK pathway inhibitor PD98059 decreased SOX2 expression by ~40% at 72hrs in our initial study.  However, this result was not reproducible.  Furthermore, the fact that PD98059 treatment did not phenocopy the effects of SOX2 inhibition (growth suppression and apoptosis) would refute the idea that MAPK pathway is a direct upstream regulator of SOX2 in our brain cancer model (Appendix C).  136 Although the mechanism of SOX2 gene regulation remains unclear in our model, other studies provide informative clues for future directions.  A recent report suggests that OCT4 interacts with SOX4 to form an OCT4/SOX4 complex that activates the enhancer activity of SOX2 gene in glioma-initiating cells (Ikushima et al., 2011b).  In addition, cyclopamine (an inhibitor targeting SMO in the SHH pathway) treatment reduces the expression of SOX2 and Bmi1, inhibits tumoursphere formation and induces differentiation of glioma-initiating cells, suggesting that the SHH pathway may act upstream of SOX2 and Bmi1 (Gopinath et al., 2013).  Although the SHH pathway is best known for its involvement in MB pathogenesis, it is also recognized for its role in the proliferation and tumourigenesis of glioma TICs (Bar et al., 2007; Takezaki et al., 2011; Xu et al., 2008).  It would therefore be interesting to find out if PLK1 down-regulates SOX2 and Bmi1 through the SHH pathway in our GBM models.  An alternative route of regulation may occur at the level of translation.  A recent study by Jain et al. indicates that retinoic acid treatment leads to CBP/p300-mediated acetylation of p53.  The stabilized p53 subsequently activates the expression of miR-34a and miR-145 which in turn repress stem cell factors such as SOX2, OCT4, KLF4 and LIN28A.  In this way, p53 acts as a “barrier to de-differentiation”, preventing “backsliding to pluripotency” in human embryonic stem cells (Jain et al., 2012).  Given the established reciprocal inhibitory interactions between p53 and PLK1 (Ando et al., 2004a; McKenzie et al., 2010), we speculate that PLK1 inhibition by siRNA or BI2536 may alleviate its suppression on p53 function, thereby allowing p53 to up-regulate the miRNAs that reduces SOX2 expression.  The interesting interplay between PLK1 and p53 appears to provide a plausible explanation to the differentiation-like phenotype of PLK1-inhibited GBM cells seen in our study.  Consistent with this, we showed that PLK1 level is much lower in normal human astrocytes compared to hNSCs, suggesting that PLK1 expression may be associated with the “proliferative state” of stem and progenitor cells.  The expression of this kinase may be down-regulated after lineage differentiation.  However, the fact that SF188 cells carry non-functional, mutated p53 (Chen et al., 1995) suggests that other proteins carrying similar functions to p53 may be involved.  Additional studies will be required to decipher the mechanism by which PLK1 regulates SOX2 expression. We noticed that SOX2 did not always co-localize with PLK1 in the immunofluorescence staining of BT74 tumourspheres (Figure 2.2B).  A few SOX2-high cells (red fluorescence) show a very low level of green fluorescence representing PLK1.  This brings up the question of whether or not PLK1 is expressed in the “cancer stem cells.”  To answer this question, we will need to be able to distinguish stem cells from the other cells.  Although the neurosphere assay provides a convenient way to propagate BTICs and measure their self-renewal, it is argued that  137 the expansion of cells might be an artifact from the culturing medium, which specifically selects the cells (not necessarily stem-like or progenitor cells) that thrive in such growth conditions (Kelly et al., 2009).  Furthermore, sphere-forming frequency approximates progenitor cell activity more closely than stem cell activity (Reynolds and Rietze, 2005).  Therefore, the BTICs we grew in neurosphere culture were likely a heterogeneous population of stem cells, progenitors and differentiated cells.             Carboxyfluorescein diacetate N-succinimidyl ester (CSFE) labeling was shown to effectively enrich for hematopoietic stem cells (HSCs) and this is based on the premise that stem cells divide slowly.  CSFE labels cells by covalently binding to intracellular proteins and becomes equally distributed to daughter cells upon cell division (Lyons and Parish, 1994; Weston and Parish, 1990).  The HSCs are mostly slow-dividing or quiescent and as such, would retain CSFE more so than the rapidly-proliferating cells do.  The idea of using this simple label-retaining technique to track stem cells seems appealing.  Yet, we believe that cancer stem-like cells may not be all slow-dividing or quiescent.  We isolated CD49high/CD24low putative breast cancer stem-like cells and found that these cells proliferated approximately 3 times faster than the non stem-like CD49low/CD24high cells.  In addition, GBM4, GBM8 and BT74, the long-term passageable BTIC cultures that must comprise a substantial number of cancer stem cells show a doubling time almost comparable to that of brain cancer cell lines SF188 and U251.  The 2 to10-fold higher level of PLK1 in the BTICs compared to the normal human neural stem cells implies that self-renewal and/or proliferation is probably accelerated in the cancer stem-like cells relative to the normal cell counterpart.  Therefore, label-retention technique based on low cell-division rate may not be applicable to the identification of cancer stem cells in brain tumours.     Cell surface antigen sorting is another technique commonly used to isolate putative CSCs.  A recent study by Venugopal et al. suggests that Bmi1 marks the intermediate precursors during the differentiation of BTICs.  Specifically, they proposed that the combination of CD133 and Bmi1 expression may allow the distinction of stem cells, progenitor cells and fully differentiated cells, which are CD133+/Bmi1intermediate, CD133-/Bmi1high and CD133-/Bmilow, respectively (Venugopal et al., 2012).  If this holds true, it would be interesting to immunostain the CD133+/Bmi1intermediate cells (putative stem cells) and CD133-/Bmi1high cells (putative progenitor cells) for PLK1 and SOX2 to find out if these proteins co-localize in stem and/or progenitor cells.  We believe that PLK1 may be expressed in the CSCs, progenitors and differentiated cells in brain tumours because the majority of the cells from the BT74 or Daoy tumourspheres are PLK1-positive in the immunofluorescence study.  Since PLK1 expression is highly correlated with the proliferative rate of a cell, we speculate that PLK1 level is probably  138 highest in the progenitor cells.  The red (SOX2-labeled) cells in the BT74 tumourspheres might be the CSCs that have high levels of stem cell marker and low-intermediate levels of PLK, while the yellow (SOX2-labeled and PLK1-labeled) cells might be the progenitors that express high levels of SOX2 and PLK1.  The green (PLK1-labeled) cells might therefore be the more differentiated cells that have lost the expression of the stem cell marker SOX2.  This hypothesis could be confirmed only when the methods for precise isolation of CSCs and progenitors are developed.  The recent development of non-invasive molecular imaging technology permits the tracking of tumour stem-like cells in vivo and emerges as a powerful technique for monitoring the fate of these cells and their response to therapy.  The clinical application of this technology could be significant as it would allow physicians to follow the history and development of CSCs, which might be one of the roots of disease recurrence.  The technique entails immunostaining of cell-surface CSC markers such as CD133, CD44, CD166 and CD34 followed by PET, MRI or optical imaging (Xia et al., 2012).  Although the identity of CSC remains elusive, a study by Gilbertson’s group convincingly demonstrated that prominin 1 (CD133) positivity marked intestinal stem cells, which became susceptible to neoplastic transformation upon aberrant WNT signaling activation (Zhu et al., 2008).  In an attempt to isolate putative CSCs in our studies, we have tried prospectively sorting SF188 and Daoy cells for CD133 and CD15 expression, respectively.  In SF188 cells, 0.15% of the total cell population was CD133+ (Appendix D).  We did not pursue this further because a number of studies showed that CD133- cells were also endowed with stem-like properties, raising the question on the robustness of CD133 as a CSC marker (Beier et al., 2007b; Ogden et al., 2008; Sun et al., 2009).  In the MB model Daoy cells, we obtained ~20% CD15+ cells but observed no difference in the transcript levels of PLK1, the neural stem cell markers SOX2, musashi and Bmi1 in the CD15+ and CD15- cells (Appendix E).  It was subsequently reported that CD15 expression may be more closely associated with neural progenitor cells and not with stem cells (Read et al., 2009).  Indeed, mounting evidence has suggested that single markers may not be sufficient to precisely isolate CSCs as many of these markers are probably expressed in both stem and progenitor cell population.  Because of the lack of unambiguous CSC markers, it appears that the neurosphere culture conditions coupled with in vitro functional assays and in vivo CSC serial transplantation have become a major alternative for such studies (Azari et al., 2011; Pollard et al., 2009).   Despite the fact that CSCs or TICs have been implicated in mediating chemo- and radiation resistance, increasing experimental evidence suggests that progenitors and differentiated cells may revert to stem-like cells through the acquisition of genetic mutations  139 (Jandial et al., 2011).  A recent study by Friedmann-Morvinski et al. showed that astrocytes and even mature neurons could be “de-differentiated” and subsequently gave rise to malignant gliomas when introduced with specific oncogenic mutations (Friedmann-Morvinski et al., 2012).  Furthermore, genetic heterogeneity and the clonal evolution theory could very well be another mechanism by which drug and radiation resistance arises.  Therefore, a real “cure” to cancer would likely involve the elimination of most if not all the cancer cells as well as the supporting cells in the surrounding niche of a tumour.  5.2.2 PLK1 INHIBITION HELPS OVERCOME TMZ RESISTANCE IN GBM CELLS  In Chapter 3, we showed that disulfiram, DSF, killed TMZ-resistant brain cancer cells.  Mounting experimental evidence suggests that DSF may be a potential anti-cancer therapy although the mechanism of action on tumour cells is not fully understood (discussed in the Introduction: 1.10.2).  It was proposed that DSF potentiates the cytotoxicity of copper by functioning as a bivalent metal chelator to facilitate the transport of copper across the cell membrane.  In our study, copper was added in the formulation of DSF.  Although copper was not included in the trial by Dufour et al. in breast cancer (Dufour et al., 1993), ditiocarb (the metabolite of DSF) may chelate copper in serum and in cancer cells to exert its effects (Liu et al., 2013a).  For those who have investigated the biological effects of DSF from the perspective of “cancer stem cells” or “tumour initiating cells”, the inhibition of aldehyde dehydrogenase, ALDH, may explain the cytotoxicity of DSF.  However, in our study, we were not able to replicate the effects of DSF by silencing ALDH1A1 and ALDH1A3 or inhibiting the proteins with a pan-ALDH inhibitor DEAB (Appendix F), suggesting that the small molecule inhibitor may target multiple isoforms of ALDH and/or other pathways to exert the apoptotic effects (Triscott et al., 2012).   An interesting observation we made in our study was that DSF decreased PLK1 expression (Figure 3.3A and 3.4A).  We hypothesized that DSF may act as a proteasome inhibitor (Skrott and Cvek, 2012) to increase the half-life of proteins that target PLK1 for degradation.  Alternatively, the drug may inhibit the transcription and/or translation machinery regulating PLK1 expression in a mechanism that is not yet understood.  In this Chapter, we also showed that BI2536 induced apoptosis in the TMZ-resistant cell line SF188 and suppressed the self-renewal of primary brain cancer cells BT74 and BT241 (Figure 3.6).  This is of particular significance as a subset of adult GBM patients and the majority of pediatric patients are relatively unresponsive to TMZ, which is the mainstay of treatment for this type of brain tumour.  Our finding was corroborated by a study showing that  140 the PLK1 small molecule inhibitor, ON01910, effectively eliminated multi-drug resistant (MDR)-positive tumour cells that are notably resistant to various chemotherapeutic drugs (Gumireddy et al., 2005).  Furthermore, PLK1 inhibition may also help override resistance to molecular targeted therapy such as imatinib.  Gleixner et al. demonstrated that BI2536 inhibited the proliferation of imatinib-resistant chronic myeloid leukemic cells that carry the T315 mutation of BCR/ABL with reasonable IC50 values (1-50nmol/L).  The efficacy of PLK1 inhibitors in these models is likely due to the targeting of a “weak spot” that tumour cells have yet to develop drug resistance against.  To this end, we have provided convincing pre-clinical results indicating that PLK1 inhibitors may benefit patients who are refractory to the standard-of-care.    Thus far, we still do not understand the mechanism by which PLK1 inhibition helps overcome drug resistance.  Our preliminary studies showed that PLK1 was co-localized with YB-1 at the spindle poles of TMZ-resistant GBM cells.  Research is underway to investigate if these two proteins function cooperatively to drive the progression through mitosis in the presence of DNA damage from TMZ treatment.  Recently, my colleague is assessing whether over-expression of PLK1 in TMZ-sensitive brain cancer cells mediates drug resistance.  In Figure 3.5, we showed an increased expression of PLK1 after prolonged exposure to TMZ.  It is possible that the tumour cells up-regulate PLK1 as a survival mechanism.  Another plausible explanation is that long-term TMZ treatment may result in the gradual elimination of cells expressing a lower level of PLK1, resulting in an “enrichment” of tumour cells that have high PLK1 expression at the end of the experiment.  The latter hypothesis seems to be favoured as recent studies suggest that the genetic heterogeneity within tumours may drive clonal expansion of drug-resistant cells in a manner reminiscent of the Darwinian evolution, thereby contributing to therapeutic failure (Chen et al., 2012; Diaz et al., 2012; Gerlinger and Swanton, 2010).  We proposed that DSF may kill GBM cells in part by decreasing PLK1 expression.  However, the down-regulation of PLK1 by DSF is only one of the many consequences of the drug treatment.  The over-expression of PLK1 is therefore not expected to fully rescue the cells from the cytotoxicity of DSF.  The fact that DSF targets multiple proteins and/or pathways and exerts remarkable potency in pre-clinical studies may reflect the “dirty (lack of high specificity in a drug) is good” concept in cancer treatment.  Furthermore, the minimal impact DSF has on normal human astrocytes further supports the utility of this drug in GBM treatment.   In the phase II clinical trial of ditiocarb by Dufour et al., the overall survival was significantly higher (81%) in the ditiocarb group compared to the placebo group (55%) (Dufour et al., 1993).  The result from this trial was not followed up until recently when new promising  141 pre-clinical results of DSF have been reported in melanoma (Cen et al., 2004a), osteosarcoma (Cho et al., 2007), breast (Yip et al., 2011b; Zhang et al., 2010), prostate (Lin et al., 2011b) and colon (Guo et al., 2010b) cancers.  These exciting results have spurred public interest in repurposing this inexpensive and reasonably safe drug for cancer treatments.  The study by Liu et al. corroborated our finding (Triscott et al., 2012), suggesting that DSF eliminates GBM cancer stem-like cells (Liu et al., 2012).  As a result of this novel discovery, the Swedish Ben and Catherine Ivy Centre for Advanced Brain Tumour Treatment received a $2.5 million grant for a project that will support pre-clinical trials to further validate the use of DSF in brain cancer treatment.  The impact of these studies is profound- requests for DSF treatment have been made by brain cancer patients who fail to respond to conventional therapies.  We believe that early phase clinical trials are urgently needed to evaluate the efficacy of this drug in brain tumours.  To conclude, we have identified DSF as a safe and readily accessible drug that holds great promise to be re-purposed for the treatment of TMZ-refractory GBM.   5.2.3 PLK1 INHIBITION AS A POTENTIAL THERAPEUTIC STRATEGY FOR SHH MB  In Chapter 4, we showed that a high PLK1 level is correlated with poor overall and relapse-free survival in MB patients (Figure 4.2C-D).  High expression of PLK1 has been associated with poor prognosis or survival in a number of cancers such as hepatoblastoma (Yamada et al., 2004), gastric carcinoma (Kanaji et al., 2006), malignant glioma (Cheng et al., 2012; Lee et al., 2012), breast cancer (King et al., 2012) and gallbladder cancer (Wang et al., 2013).  A study by Hansford et al. showed that the bone marrow metastases from neuroblastoma patients contained an enriched population of TICs, which is a marker for minimal residual disease in high-risk patients (Hansford et al., 2007).  Interestingly, molecular characterization of these cells revealed that they expressed an elevated level of PLK1 (Grinshtein et al., 2011).  Together, these results suggest that PLK1 expression may facilitate the invasion and/or metastasis of cancer cells or confer growth/survival advantage in the cells (such as TICs) that have the propensity to metastasize.  It is also plausible that PLK1 expression becomes up-regulated to promote clonal expansion in cells at the secondary site. A recent report by Glinsky predicts that cancer cells with a stem cell-like expression profile of a “death-from-cancer signature” would acquire a metastasis-enabling, anoikis-resistant, aneuploid phenotype with abnormal cell cycle control.  PLK1 was proposed to be one of the “death-from-cancer signature” genes in the study (Glinsky, 2006).  Consistent with this, we found PLK1 expression to be closely associated with relapse-free survival in brain tumours and the prognostic power of expression levels of this kinase even may supersede the well- 142 established prognostic factors such as age and metastasis.  As an interesting side note, my colleague analyzed the expression of nestin, OCT4 and CD15 in the four subgroups of MBs and did not find a correlation between the level of stem cell markers and patient survival.  In addition, these markers are not expressed at a higher level in Group 3 MB, which has the worst prognosis.  Therefore, a “stem cell-like” expression profile may not always confer a less optimistic prognosis and survival in patients.  Why is PLK1 expression correlated with poor survival?  We do not know the exact answer but have made a few speculations.  PLK1-high patients often experience shorter progression-free survival, suggesting that the disease recurs quickly after the treatments, such as radiotherapy.  Harris et al. has recently demonstrated that PLK1 inhibition significantly sensitized MB cells to radiation (Harris et al., 2012), a result also seen in the head and neck cancer (Gerster et al., 2010).  The inhibition of PLK1 might potentiate the DNA damage response by amplifying the response of the ATM/ATR pathway (van Vugt et al., 2001a), leading to the radio-sensitizing effect.  Thus far there is no study that directly shows the involvement of PLK1 in mediating radiation resistance.  It would be interesting to over-express PLK1 in RT-sensitive cells and find out if an increased level of this kinase drives RT resistance.  Local invasion is one of the major obstacles in brain cancer treatment.  The shorter relapse-free survival in PLK1-high patients may thus be due to the invasiveness of tumours.  The role of PLK1 in this process is not well understood.  However, a study by Rizki et al. showed that PLK1 mediated invasion through regulating the activity of vimentin and cell surface level of ß1-integrin in basal-like breast cancer (Rizki et al., 2007).  It would therefore be of interest to examine whether modulating PLK1 levels promotes invasion in brain tumours.  A recent study by Barrett et al. suggests that self-renewal does not predict tumourigenic potential in mice.  They showed that high Id1 expression identified tumour cells with high self-renewal capacity.  Intriguingly, the Id1low cells (progenitors-like cells) developed tumours more rapidly and with higher penetrance than the Id1high cells (stem-like cells).  Moreover, suppressing self-renewal through the deletion of Id1 had a modest effect on the survival of animals while silencing Olig2 in the Id1low cells offered survival benefit (Barrett et al., 2012).  Because PLK1 expression is correlated with proliferation rate, we speculate that the Id1low cells may have higher levels of PLK1 than the Id1high cells.  Therefore, these Id1low/PLK1high progenitors-like cells would instigate tumour formation more rapidly than the stem-like cells.  Translating this into the clinical setting would mean that the PLK1-high tumours may be endowed with the capacity to generate detectable tumours with shorter latency, which would be linked to aggressive disease and poor prognosis.  143 To date, we still do not know whether PLK1 is “associated” with or is “causative” of poor prognosis and survival.  Nevertheless, results from our pre-clinical studies support the molecular targeting of this kinase because of the impressive anti-growth and anti-survival effects it has on brain cancer cells.  This is further facilitated by the development of a number of small molecule inhibitors that are highly specific to PLK1 (Gumireddy et al., 2005; Olmos et al., 2011; Steegmaier et al., 2007).  An integral question to pose is whether or not the inhibitors cross the blood-brain-barrier (BBB).  We have shown that BI2536 increased the life expectancy of mice bearing orthotopic GBM and MB tumours.  It is most likely that the inhibitor has crossed the BBB and reached the brain to exert this therapeutic effect.  However, in order to obtain a definitive answer, we will need to perform a pharmacokinetic study to quantitatively measure the concentration of the inhibitor in the brain after administration. In Chapter 4, we showed that the animals treated with chemotherapeutic agents and BI2536 had equivalent length of survival time (Figure 4.5D).  We questioned if the combination treatment would be more efficacious.  Our preliminary in vitro studies suggest that there was no additive or synergistic effect when BI2536 was combined with etoposide or cisplatin in Daoy cells; the latter result was confirmed by Guerreiro et al. (Guerreiro et al., 2011).  At first glance, the combination of vincristine and BI2536 appeared to suppress cell growth more effectively than single agents did.  However, when my colleague included the proper solvent controls into the experiments, we did not observe an additive or synergistic effect with the drug combination.  Methanol, the solvent for vincristine, seemed to have non-specific toxicity on cells (data not published).   Although we did not show a benefit of combining BI2536 with vincristine, etoposide or cisplatin in Daoy cells, we believe that the future clinical application of PLK1 inhibitors would most likely involve the combination with chemotherapeutic agents, which has already shown improved efficacy compared to single agents (Grinshtein et al., 2011; Gumireddy et al., 2005; Maire et al., 2013).  For example, Maire et al. demonstrated that combining BI2536 and doxorubicin and cyclophosphamide led to a more rapid response and prevented relapse in xenograft models established from biopsies of triple-negative breast cancer patients (Maire et al., 2013).  In the study by Grinshtein et al., the combination of BI2536 and irinotecan suppressed neuroblastoma tumour growth and prolonged the survival of animals more effectively than single agents did (Grinshtein et al., 2011).  Further mechanistic studies in the biology of PLK1 and animal experiments will help identify beneficial drug combinations.  144 5.3 CONCLUDING REMARKS: PLK1 AS A MOLECULAR TARGET IN GBM AND MB  In Chapters 2, 3 and 4, we presented experimental evidence suggesting that PLK1 may be a good molecular target for brain cancer treatment because of the following reasons: 1) the kinase is over-expressed in MB and GBM tissues but not in the normal brain tissue counterparts; 2) PLK1 expression is highly correlated with poor patient survival; 3) the inhibition of this kinase significantly suppressed cell growth and induced apoptosis in tumour cells but not in normal astrocytes or neural stem cells; 4) PLK1 inhibitors reduce TMZ-resistant brain cancer cells.   In our studies, we showed that PLK1 inhibition repressed the self-renewal of BTICs.  We have also demonstrated in breast cancer models that PLK1 inhibitors targeted the CD49f+/CD24- putative cancer stem-like cells and blocked their growth and self-renewal (Hu et al., 2012).  Consistent with this, a recent study indicates that BI6727 (PLK1 inhibitor) but not chemotherapeutic drugs lower the number of colon cancer initiating cells (CCICs) in vivo.  The CD133+ CCICs continued proliferating in the presence of chemotherapy.  In stark contrast, BI6727 eliminated the proliferating CD133+ CCICs; the residual cells surviving BI6727 treatment were quiescent cells that became susceptible to PLK1 inhibition when they entered the next round of cell cycle (Francescangeli et al., 2012).  The unique capacity of the PLK1 inhibitor to repress the self-renewal of cancer stem-like cells distinguishes it from chemotherapeutic agents that have been reported to enrich those cells following treatments (Calcagno et al., 2010; Liu et al., 2006a; Liu et al., 2013b).  However, since the cytotoxic effect of PLK1 inhibition is only observed in dividing cells, it raises the question of the appropriate length of time that the inhibitor should be given to eliminate most if not all the cancer stem-like cells.  The fact that the PLK1 inhibitor is an “anti-mitotic” agent implies that there may be a limit to its efficacy.  Specifically, a few “dormant” cancer stem-like cells may escape PLK1 inhibition and become the seeds that subsequently repopulate a tumour.  Nevertheless, this is a novel and exciting discovery since we have identified a potential therapeutic strategy that may target cells that make up the bulk of the tumour as well as those that are notoriously difficult to eliminate by conventional therapy. BI2536 is an ATP-competitive inhibitor highly specific to PLK1.  The study by Steegmaier et al. demonstrated the excellent selectivity of this inhibitor to PLK1 compared to other PLK family members as well as the additional 63 kinases examined (Steegmaier et al., 2007).  In contrast to the flexibility of mitogenic signal transductions, mitosis is an exquisite process that is strictly regulated to ensure the fidelity of the genome (Plyte and Musacchio, 2007).  Furthermore, PLK2, 3 and 5 are involved in the DNA damage response to prevent oncogenesis  145 (van de Weerdt and Medema, 2006).  The high potency and selectivity of BI2536 may therefore be advantageous since any off-target effects influencing the functions of other PLKs could have deleterious effects on cells.    In our in vivo study, we showed that animals treated with BI2536 had equivalent lengths of survival time compared to those treated with chemotherapy.  Further examination of the brain tissues revealed no qualitative differences between the two groups of animals.  No significant neuro-pathological damage was observed in either chemotherapy or BI2536 treatment.  Although there is no obvious damage to the brain from the chemotherapeutic regime in this study, we wondered if BI2536 may still be an alternative therapeutic option considering that vincristine and cisplatin are known to cause peripheral neuropathy or auditory impairment.  As a future direction, we would be very interested in studying the long-term behavioural changes in animals treated with chemotherapy or PLK1 inhibitors. In conclusion, we have shown that PLK1 is a promising molecular target for brain cancer treatment.  As the safety and efficacies of PLK1 inhibitors have already been evaluated in phase I/II clinical trials in a wide-range of human malignancies such as oesophageal carcinoma, colorectal adenocarcinoma, malignant melanoma, breast cancer, ovarian adenocarcinoma and renal cancer (Hofheinz et al., 2010; Mross et al., 2008; Olmos et al., 2011), we believe that further investigations of PLK1 inhibitors in brain tumours are worth pursuing in view of the promising results from our and others’ pre-clinical studies (Harris et al., 2012; Lee et al., 2012).     146 REFERENCES Adams, R. R., Tavares, A. A., Salzberg, A., Bellen, H. J., and Glover, D. M. (1998). pavarotti encodes a kinesin-like protein required to organize the central spindle and contractile ring for cytokinesis. Genes Dev 12, 1483-1494.  Agerlin, N., Gjerris, F., Brincker, H., Haase, J., Laursen, H., Moller, K. A., Ovesen, N., Reske-Nielsen, E., and Schmidt, K. (1999). Childhood medulloblastoma in Denmark 1960-1984. A population-based retrospective study. Childs Nerv Syst 15, 29-36; discussion 36-27.  Al-Hajj, M., Becker, M. W., Wicha, M., Weissman, I., and Clarke, M. F. (2004). Therapeutic implications of cancer stem cells. Curr Opin Genet Dev 14, 43-47.  Al-Hajj, M., Wicha, M. S., Benito-Hernandez, A., Morrison, S. J., and Clarke, M. F. (2003). Prospective identification of tumorigenic breast cancer cells. Proc Natl Acad Sci U S A 100, 3983-3988.  Al-Halabi, H., Nantel, A., Klekner, A., Guiot, M. C., Albrecht, S., Hauser, P., Garami, M., Bognar, L., Kavan, P., Gerges, N., et al. (2011). Preponderance of sonic hedgehog pathway activation characterizes adult medulloblastoma. Acta Neuropathol 121, 229-239.  Albright, A. L., Wisoff, J. H., Zeltzer, P. M., Boyett, J. M., Rorke, L. B., and Stanley, P. (1996). Effects of medulloblastoma resections on outcome in children: a report from the Children's Cancer Group. Neurosurgery 38, 265-271.  Aldosari, N., Bigner, S. H., Burger, P. C., Becker, L., Kepner, J. L., Friedman, H. S., and McLendon, R. E. (2002). MYCC and MYCN oncogene amplification in  medulloblastoma. A fluorescence in situ hybridization study on paraffin sections from the Children's Oncology Group. Arch Pathol Lab Med 126, 540-544.  Allahdini, F., Amirjamshidi, A., Reza-Zarei, M., and Abdollahi, M. (2010). Evaluating the prognostic factors effective on the outcome of patients with glioblastoma multiformis: does maximal resection of the tumor lengthen the median survival? World Neurosurg 73, 128-134; discussion e116.  Allendoerfer, K. L., Magnani, J. L., and Patterson, P. H. (1995). FORSE-1, an antibody that labels regionally restricted subpopulations of progenitor cells in the embryonic central nervous system, recognizes the Le(x) carbohydrate on a proteoglycan and two glycolipid antigens. Mol Cell Neurosci 6, 381-395.  Alonso, M. M., Diez-Valle, R., Manterola, L., Rubio, A., Liu, D., Cortes-Santiago, N., Urquiza, L., Jauregi, P., de Munain, A. L., Sampron, N., et al. (2011). Genetic and epigenetic modifications of sox2 contribute to the invasive phenotype of malignant gliomas. PLoS One 6, e26740.  Altman, J. (1966). Proliferation and migration of undifferentiated precursor cells in the rat during postnatal gliogenesis. Exp Neurol 16, 263-278.   147 Altman, J., and Das, G. D. (1965). Autoradiographic and histological evidence of postnatal hippocampal neurogenesis in rats. J Comp Neurol 124, 319-335.  Alvarez-Buylla, A., and Lois, C. (1995). Neuronal stem cells in the brain of adult vertebrates. Stem Cells 13, 263-272.  Alvarez-Buylla, A., Theelen, M., and Nottebohm, F. (1988). Birth of projection neurons in the higher vocal center of the canary forebrain before, during, and after song learning. Proc Natl Acad Sci U S A 85, 8722-8726.  Alvi, A. J., Clayton, H., Joshi, C., Enver, T., Ashworth, A., Vivanco, M. M., Dale, T. C., and Smalley, M. J. (2003). Functional and molecular characterisation of mammary side population cells. Breast Cancer Res 5, R1-8.  Amin, S. H., Tibes, R., Kim, J. E., and Hybarger, C. P. (2010). Hedgehog antagonist GDC-0449 is effective in the treatment of advanced basal cell carcinoma. Laryngoscope 120, 2456-2459.  Ando, K., Ozaki, T., Yamamoto, H., Furuya, K., Hosoda, M., Hayashi, S., Fukuzawa, M., and Nakagawara, A. (2004a). Polo-like kinase 1 (Plk1) inhibits p53 function by physical interaction and phosphorylation. The Journal of biological chemistry 279, 25549-25561.  Ando, K., Ozaki, T., Yamamoto, H., Furuya, K., Hosoda, M., Hayashi, S., Fukuzawa, M., and Nakagawara, A. (2004b). Polo-like kinase 1 (Plk1) inhibits p53 function by physical interaction and phosphorylation. J Biol Chem 279, 25549-25561.  Andrysik, Z., Bernstein, W. Z., Deng, L., Myer, D. L., Li, Y. Q., Tischfield, J. A., Stambrook, P. J., and Bahassi el, M. (2010). The novel mouse Polo-like kinase 5 responds to DNA damage and localizes in the nucleolus. Nucleic Acids Res 38, 2931-2943.  Ang, X. L., Seeburg, D. P., Sheng, M., and Harper, J. W. (2008). Regulation of postsynaptic RapGAP SPAR by Polo-like kinase 2 and the SCFbeta-TRCP ubiquitin ligase in hippocampal neurons. J Biol Chem 283, 29424-29432.  Annovazzi, L., Mellai, M., Caldera, V., Valente, G., and Schiffer, D. (2011). SOX2 expression and amplification in gliomas and glioma cell lines. Cancer Genomics Proteomics 8, 139-147.  Arnaud, L., Pines, J., and Nigg, E. A. (1998). GFP tagging reveals human Polo-like kinase 1 at the kinetochore/centromere region of mitotic chromosomes. Chromosoma 107, 424-429.  Azari, H., Louis, S. A., Sharififar, S., Vedam-Mai, V., and Reynolds, B. A. (2011). Neural-colony forming cell assay: an assay to discriminate bona fide neural stem cells from neural progenitor cells. J Vis Exp.  Bachoo, R. M., Maher, E. A., Ligon, K. L., Sharpless, N. E., Chan, S. S., You, M. J., Tang, Y., DeFrances, J., Stover, E., Weissleder, R., et al. (2002). Epidermal growth factor receptor and Ink4a/Arf: convergent mechanisms governing terminal differentiation and transformation along the neural stem cell to astrocyte axis. Cancer Cell 1, 269-277.  148  Bahassi el, M. (2011). Polo-like kinases and DNA damage checkpoint: beyond the traditional mitotic functions. Exp Biol Med (Maywood) 236, 648-657.  Bahassi el, M., Conn, C. W., Myer, D. L., Hennigan, R. F., McGowan, C. H., Sanchez, Y., and Stambrook, P. J. (2002). Mammalian Polo-like kinase 3 (Plk3) is a multifunctional protein involved in stress response pathways. Oncogene 21, 6633-6640.  Balana, C., Carrato, C., Ramirez, J. L., Cardona, A. F., Berdiel, M., Sanchez, J. J., Taron, M., Hostalot, C., Musulen, E., Ariza, A., and Rosell, R. (2011). Tumour and serum MGMT promoter methylation and protein expression in glioblastoma patients. Clin Transl Oncol 13, 677-685.  Bao, S., Wu, Q., McLendon, R. E., Hao, Y., Shi, Q., Hjelmeland, A. B., Dewhirst, M. W., Bigner, D. D., and Rich, J. N. (2006). Glioma stem cells promote radioresistance by preferential activation of the DNA damage response. Nature 444, 756-760.  Bar, E. E., Chaudhry, A., Lin, A., Fan, X., Schreck, K., Matsui, W., Piccirillo, S., Vescovi, A. L., DiMeco, F., Olivi, A., and Eberhart, C. G. (2007). Cyclopamine-mediated hedgehog pathway inhibition depletes stem-like cancer cells in glioblastoma. Stem Cells 25, 2524-2533.  Barceloux, D. G. (1999). Copper. J Toxicol Clin Toxicol 37, 217-230.  Barone, G., Maurizi, P., Tamburrini, G., and Riccardi, R. (2006). Role of temozolomide in pediatric brain tumors. Childs Nerv Syst 22, 652-661.  Barr, F. A., Sillje, H. H., and Nigg, E. A. (2004). Polo-like kinases and the orchestration of cell division. Nat Rev Mol Cell Biol 5, 429-440.  Barrett, L. E., Granot, Z., Coker, C., Iavarone, A., Hambardzumyan, D., Holland, E. C., Nam, H. S., and Benezra, R. (2012). Self-renewal does not predict tumor growth potential in mouse models of high-grade glioma. Cancer Cell 21, 11-24.  Bartels, U., Baruchel, S., Carret, A. S., Crooks, B., Hukin, J., Johnston, D., Silva, M., Strother, D., Wilson, B., Zelcer, S., et al. (2011). The use and effectiveness of temozolomide in children with central nervous system tumours: a survey from the Canadian Paediatric Brain Tumour Consortium. Curr Oncol 18, e19-24.  Bayliss, R., Sardon, T., Vernos, I., and Conti, E. (2003). Structural basis of Aurora-A activation by TPX2 at the mitotic spindle. Mol Cell 12, 851-862.  Beier, D., Hau, P., Proescholdt, M., Lohmeier, A., Wischhusen, J., Oefner, P. J., Aigner, L., Brawanski, A., Bogdahn, U., and Beier, C. P. (2007a). CD133(+) and CD133(-) glioblastoma-derived cancer stem cells show differential growth characteristics and molecular profiles. Cancer Res 67, 4010-4015.  Beier, D., Hau, P., Proescholdt, M., Lohmeier, A., Wischhusen, J., Oefner, P. J., Aigner, L., Brawanski, A., Bogdahn, U., and Beier, C. P. (2007b). CD133(+) and CD133(-)  149 glioblastoma-derived cancer stem cells show differential growth characteristics and molecular profiles. Cancer research 67, 4010-4015.  Belanich, M., Pastor, M., Randall, T., Guerra, D., Kibitel, J., Alas, L., Li, B., Citron, M., Wasserman, P., White, A., et al. (1996). Retrospective study of the correlation between the DNA repair protein alkyltransferase and survival of brain tumor patients treated with carmustine. Cancer Res 56, 783-788.  Berger, M. S., Baumeister, B., Geyer, J. R., Milstein, J., Kanev, P. M., and LeRoux, P. D. (1991). The risks of metastases from shunting in children with primary central nervous system tumors. J Neurosurg 74, 872-877.  Berman, D. M., Karhadkar, S. S., Hallahan, A. R., Pritchard, J. I., Eberhart, C. G., Watkins, D. N., Chen, J. K., Cooper, M. K., Taipale, J., Olson, J. M., and Beachy, P. A. (2002). Medulloblastoma growth inhibition by hedgehog pathway blockade. Science 297, 1559-1561.  Bettencourt-Dias, M., Rodrigues-Martins, A., Carpenter, L., Riparbelli, M., Lehmann, L., Gatt, M. K., Carmo, N., Balloux, F., Callaini, G., and Glover, D. M. (2005). SAK/PLK4 is required for centriole duplication and flagella development. Curr Biol 15, 2199-2207.  Biegel, J. A., Janss, A. J., Raffel, C., Sutton, L., Rorke, L. B., Harper, J. M., and Phillips, P. C. (1997). Prognostic significance of chromosome 17p deletions in childhood primitive neuroectodermal tumors (medulloblastomas) of the central nervous system. Clin Cancer Res 3, 473-478.  Biernat, W., Kleihues, P., Yonekawa, Y., and Ohgaki, H. (1997a). Amplification and overexpression of MDM2 in primary (de novo) glioblastomas. J Neuropathol Exp Neurol 56, 180-185.  Biernat, W., Tohma, Y., Yonekawa, Y., Kleihues, P., and Ohgaki, H. (1997b). Alterations of cell cycle regulatory genes in primary (de novo) and secondary glioblastomas. Acta Neuropathol 94, 303-309.  Bignami, M., O'Driscoll, M., Aquilina, G., and Karran, P. (2000). Unmasking a killer: DNA O(6)-methylguanine and the cytotoxicity of methylating agents. Mutat Res 462, 71-82.  Bigner, S. H., Mark, J., Friedman, H. S., Biegel, J. A., and Bigner, D. D. (1988). Structural chromosomal abnormalities in human medulloblastoma. Cancer Genet Cytogenet 30, 91-101.  Blazek, E. R., Foutch, J. L., and Maki, G. (2007). Daoy medulloblastoma cells that express CD133 are radioresistant relative to CD133- cells, and the CD133+ sector is enlarged by hypoxia. Int J Radiat Oncol Biol Phys 67, 1-5.  Bonnet, D., and Dick, J. E. (1997). Human acute myeloid leukemia is organized as a hierarchy that originates from a primitive hematopoietic cell. Nat Med 3, 730-737.  Brandes, A. A., Tosoni, A., Franceschi, E., Reni, M., Gatta, G., and Vecht, C. (2008). Glioblastoma in adults. Crit Rev Oncol Hematol 67, 139-152.  150 Brat, D. J., and Van Meir, E. G. (2004). Vaso-occlusive and prothrombotic mechanisms associated with tumor hypoxia, necrosis, and accelerated growth in glioblastoma. Lab Invest 84, 397-405.  Brazel, C. Y., Limke, T. L., Osborne, J. K., Miura, T., Cai, J., Pevny, L., and Rao, M. S. (2005). Sox2 expression defines a heterogeneous population of neurosphere-forming cells in the adult murine brain. Aging Cell 4, 197-207.  Breithaupt, H., Dammann, A., and Aigner, K. (1982). Pharmacokinetics of dacarbazine (DTIC) and its metabolite 5-aminoimidazole-4-carboxamide (AIC) following different dose schedules. Cancer Chemother Pharmacol 9, 103-109.  Brennan, C., Momota, H., Hambardzumyan, D., Ozawa, T., Tandon, A., Pedraza, A., and Holland, E. (2009). Glioblastoma subclasses can be defined by activity among signal transduction pathways and associated genomic alterations. PLoS One 4, e7752.  Broniscer, A., Chintagumpala, M., Fouladi, M., Krasin, M. J., Kocak, M., Bowers, D. C., Iacono, L. C., Merchant, T. E., Stewart, C. F., Houghton, P. J., et al. (2006).  Temozolomide after radiotherapy for newly diagnosed high-grade glioma and unfavorable low-grade glioma in children. J Neurooncol 76, 313-319.  Browd, S. R., Kenney, A. M., Gottfried, O. N., Yoon, J. W., Walterhouse, D., Pedone, C. A., and Fults, D. W. (2006). N-myc can substitute for insulin-like growth factor signaling in a mouse model of sonic hedgehog-induced medulloblastoma. Cancer Res 66, 2666-2672.  Brugieres, L., Pierron, G., Chompret, A., Paillerets, B. B., Di Rocco, F., Varlet, P., Pierre-Kahn, A., Caron, O., Grill, J., and Delattre, O. (2010). Incomplete penetrance of the predisposition to medulloblastoma associated with germ-line SUFU mutations. J Med Genet 47, 142-144.  Buczkowicz, P., Zarghooni, M., Bartels, U., Morrison, A., Misuraca, K. L., Chan, T., Bouffet, E., Huang, A., Becher, O., and Hawkins, C. (2012). Aurora Kinase B Is a Potential Therapeutic Target in Pediatric Diffuse Intrinsic Pontine Glioma. Brain pathology.  Budde, P.P., Kumagai, A., Dunphy, W.G., Heald, R. (2001). Regulation of Op18 during spindle assembly in Xenopus egg extracts. J Cell Biol 153(1), 149-58.  Budman, D. R., and Calabro, A. (2002). In vitro search for synergy and antagonism: evaluation of docetaxel combinations in breast cancer cell lines. Breast Cancer Res Treat 74, 41-46.  Buonamici, S., Williams, J., Morrissey, M., Wang, A., Guo, R., Vattay, A., Hsiao, K., Yuan, J., Green, J., Ospina, B., et al. (2010). Interfering with resistance to smoothened antagonists by inhibition of the PI3K pathway in medulloblastoma. Sci Transl Med 2, 51ra70.  Burger, P. C., Grahmann, F. C., Bliestle, A., and Kleihues, P. (1987). Differentiation in the medulloblastoma. A histological and immunohistochemical study. Acta Neuropathol 73, 115-123.   151 Burkhardt, J. K., Shin, B. J., and Boockvar, J. A. (2011). Neural stem cells and glioma stem-like cells respond differently to chemotherapeutic drugs: selectivity at the cellular level. Neurosurgery 68, N21-22.  Burkitt, M. J., Bishop, H. S., Milne, L., Tsang, S. Y., Provan, G. J., Nobel, C. S., Orrenius, S., and Slater, A. F. (1998). Dithiocarbamate toxicity toward thymocytes involves their copper-catalyzed conversion to thiuram disulfides, which oxidize glutathione in a redox cycle without the release of reactive oxygen species. Arch Biochem Biophys 353, 73-84.  Burns, K. L., Ueki, K., Jhung, S. L., Koh, J., and Louis, D. N. (1998). Molecular genetic correlates of p16, cdk4, and pRb immunohistochemistry in glioblastomas. J Neuropathol Exp Neurol 57, 122-130.  Burns, T. F., Fei, P., Scata, K. A., Dicker, D. T., and El-Deiry, W. S. (2003). Silencing of the novel p53 target gene Snk/Plk2 leads to mitotic catastrophe in paclitaxel (taxol)-exposed cells. Mol Cell Biol 23, 5556-5571.  Bylund, M., Andersson, E., Novitch, B. G., and Muhr, J. (2003). Vertebrate neurogenesis is counteracted by Sox1-3 activity. Nat Neurosci 6, 1162-1168.  Calabrese, C., Frank, A., Maclean, K., and Gilbertson, R. (2003). Medulloblastoma sensitivity to 17-allylamino-17-demethoxygeldanamycin requires MEK/ERKM. J Biol Chem 278, 24951-24959.  Calcagno, A. M., Salcido, C. D., Gillet, J. P., Wu, C. P., Fostel, J. M., Mumau, M. D., Gottesman, M. M., Varticovski, L., and Ambudkar, S. V. (2010). Prolonged drug selection of breast cancer cells and enrichment of cancer stem cell characteristics. Journal of the National Cancer Institute 102, 1637-1652.  Capela, A., and Temple, S. (2002). LeX/ssea-1 is expressed by adult mouse CNS stem cells, identifying them as nonependymal. Neuron 35, 865-875.  Capela, A., and Temple, S. (2006). LeX is expressed by principle progenitor cells in the embryonic nervous system, is secreted into their environment and binds Wnt-1. Dev Biol 291, 300-313.  Casenghi, M., Barr, F. A., and Nigg, E. A. (2005). Phosphorylation of Nlp by Plk1 negatively regulates its dynein-dynactin-dependent targeting to the centrosome. J Cell Sci 118, 5101-5108.  Casenghi, M., Meraldi, P., Weinhart, U., Duncan, P. I., Korner, R., and Nigg, E. A. (2003). Polo-like kinase 1 regulates Nlp, a centrosome protein involved in microtubule nucleation. Dev Cell 5, 113-125.  Cen, D., Brayton, D., Shahandeh, B., Meyskens, F. L., Jr., and Farmer, P. J. (2004a). Disulfiram facilitates intracellular Cu uptake and induces apoptosis in human melanoma cells. Journal of medicinal chemistry 47, 6914-6920.  Chang, C. H., Horton, J., Schoenfeld, D., Salazer, O., Perez-Tamayo, R., Kramer, S., Weinstein, A., Nelson, J. S., and Tsukada, Y. (1983). Comparison of postoperative  152 radiotherapy and combined postoperative radiotherapy and chemotherapy in the multidisciplinary management of malignant gliomas. A joint Radiation Therapy Oncology Group and Eastern Cooperative Oncology Group study. Cancer 52, 997-1007.  Chang, S. M., Wen, P., Cloughesy, T., Greenberg, H., Schiff, D., Conrad, C., Fink, K., Robins, H. I., De Angelis, L., Raizer, J., et al. (2005). Phase II study of CCI-779 in patients with recurrent glioblastoma multiforme. Invest New Drugs 23, 357-361.  Charafe-Jauffret, E., Ginestier, C., Iovino, F., Tarpin, C., Diebel, M., Esterni, B., Houvenaeghel, G., Extra, J. M., Bertucci, F., Jacquemier, J., et al. (2010). Aldehyde dehydrogenase 1-positive cancer stem cells mediate metastasis and poor clinical outcome in inflammatory breast cancer. Clin Cancer Res 16, 45-55.  Charafe-Jauffret, E., Ginestier, C., Iovino, F., Wicinski, J., Cervera, N., Finetti, P., Hur, M. H., Diebel, M. E., Monville, F., Dutcher, J., et al. (2009). Breast cancer cell lines contain functional cancer stem cells with metastatic capacity and a distinct molecular signature. Cancer Res 69, 1302-1313.  Chaudhary, P. M., and Roninson, I. B. (1991). Expression and activity of P-glycoprotein, a multidrug efflux pump, in human hematopoietic stem cells. Cell 66, 85-94.  Cheema, T. A., Kanai, R., Kim, G. W., Wakimoto, H., Passer, B. J., Rabkin, S. D., and Martuza, R. L. (2011). Enhanced anti-tumor efficacy of low dose etoposide with oncolytic herpes simplex virus in human glioblastoma stem cell xenografts. Clin Cancer Res. 17(23):7383-93.   Cheetham, G. M., Knegtel, R. M., Coll, J. T., Renwick, S. B., Swenson, L., Weber, P., Lippke, J. A., and Austen, D. A. (2002). Crystal structure of aurora-2, an oncogenic serine/threonine kinase. J Biol Chem 277, 42419-42422.  Chen, D., Cui, Q. C., Yang, H., and Dou, Q. P. (2006). Disulfiram, a clinically used anti-alcoholism drug and copper-binding agent, induces apoptotic cell death in breast cancer cultures and xenografts via inhibition of the proteasome activity. Cancer Res 66, 10425-10433.  Chen, J., Li, Y., Yu, T. S., McKay, R. M., Burns, D. K., Kernie, S. G., and Parada, L. F. (2012). A restricted cell population propagates glioblastoma growth after chemotherapy. Nature 488, 522-526.  Chen, P., Iavarone, A., Fick, J., Edwards, M., Prados, M., and Israel, M. A. (1995). Constitutional p53 mutations associated with brain tumors in young adults. Cancer Genet Cytogenet 82, 106-115.  Chen, Y. C., Chen, Y. W., Hsu, H. S., Tseng, L. M., Huang, P. I., Lu, K. H., Chen, D. T., Tai, L. K., Yung, M. C., Chang, S. C., et al. (2009). Aldehyde dehydrogenase 1 is a putative marker for cancer stem cells in head and neck squamous cancer. Biochem Biophys Res Commun 385, 307-313.  Chen, Z. P., Yarosh, D., Garcia, Y., Tampieri, D., Mohr, G., Malapetsa, A., Langleben, A., and Panasci, L. C. (1999). Relationship between O6-methylguanine-DNA  153 methyltransferase levels and clinical response induced by chloroethylnitrosourea therapy in glioma patients. Can J Neurol Sci 26, 104-109.  Cheng, K. Y., Lowe, E. D., Sinclair, J., Nigg, E. A., and Johnson, L. N. (2003). The crystal structure of the human polo-like kinase-1 polo box domain and its phospho-peptide complex. EMBO J 22, 5757-5768.  Cheng, M. W., Wang, B. C., Weng, Z. Q., and Zhu, X. W. (2012). Clinicopathological significance of Polo-like kinase 1 (PLK1) expression in human malignant glioma. Acta Histochem 114, 503-509.  Cheng, Y., Pang, J. C., Ng, H. K., Ding, M., Zhang, S. F., Zheng, J., Liu, D. G., and Poon, W. S. (2000). Pilocytic astrocytomas do not show most of the genetic changes commonly seen in diffuse astrocytomas. Histopathology 37, 437-444.  Chi, S. N., Gardner, S. L., Levy, A. S., Knopp, E. A., Miller, D. C., Wisoff, J. H., Weiner, H. L., and Finlay, J. L. (2004). Feasibility and response to induction chemotherapy intensified with high-dose methotrexate for young children with newly diagnosed high-risk disseminated medulloblastoma. J Clin Oncol 22, 4881-4887.  Chirasani, S. R., Sternjak, A., Wend, P., Momma, S., Campos, B., Herrmann, I. M., Graf, D., Mitsiadis, T., Herold-Mende, C., Besser, D., et al. (2010). Bone morphogenetic protein-7 release from endogenous neural precursor cells suppresses the tumourigenicity of stem-like glioblastoma cells. Brain 133, 1961-1972.  Cho, H. J., Lee, T. S., Park, J. B., Park, K. K., Choe, J. Y., Sin, D. I., Park, Y. Y., Moon, Y. S., Lee, K. G., Yeo, J. H., et al. (2007). Disulfiram suppresses invasive ability of osteosarcoma cells via the inhibition of MMP-2 and MMP-9 expression. J Biochem Mol Biol 40, 1069-1076.  Cho, Y. J., Tsherniak, A., Tamayo, P., Santagata, S., Ligon, A., Greulich, H., Berhoukim, R., Amani, V., Goumnerova, L., Eberhart, C. G., et al. (2011). Integrative genomic analysis of medulloblastoma identifies a molecular subgroup that drives poor clinical outcome. J Clin Oncol 29, 1424-1430.  Chopra, P., Sethi, G., Dastidar, S. G., and Ray, A. (2010). Polo-like kinase inhibitors: an emerging opportunity for cancer therapeutics. Expert Opin Investig Drugs 19, 27-43.  Christ, O., Lucke, K., Imren, S., Leung, K., Hamilton, M., Eaves, A., Smith, C., and Eaves, C. (2007). Improved purification of hematopoietic stem cells based on their elevated aldehyde dehydrogenase activity. Haematologica 92, 1165-1172.  Chu, P., Clanton, D. J., Snipas, T. S., Lee, J., Mitchell, E., Nguyen, M. L., Hare, E., and Peach, R. J. (2009). Characterization of a subpopulation of colon cancer cells with stem cell-like properties. Int J Cancer 124, 1312-1321.  Clay, F. J., McEwen, S. J., Bertoncello, I., Wilks, A. F., and Dunn, A. R. (1993). Identification and cloning of a protein kinase-encoding mouse gene, Plk, related to the polo gene of Drosophila. Proc Natl Acad Sci U S A 90, 4882-4886.   154 Clement, V., Marino, D., Cudalbu, C., Hamou, M. F., Mlynarik, V., de Tribolet, N., Dietrich, P. Y., Gruetter, R., Hegi, M. E., and Radovanovic, I. (2010). Marker-independent identification of glioma-initiating cells. Nat Methods 7, 224-228.  Clifford, S. C., Lusher, M. E., Lindsey, J. C., Langdon, J. A., Gilbertson, R. J., Straughton, D., and Ellison, D. W. (2006). Wnt/Wingless pathway activation and chromosome 6 loss characterize a distinct molecular sub-group of medulloblastomas associated with a favorable prognosis. Cell Cycle 5, 2666-2670.  Cloughesy, T. F., Wen, P. Y., Robins, H. I., Chang, S. M., Groves, M. D., Fink, K. L., Junck, L., Schiff, D., Abrey, L., Gilbert, M. R., et al. (2006). Phase II trial of tipifarnib in patients with recurrent malignant glioma either receiving or not receiving enzyme-inducing antiepileptic drugs: a North American Brain Tumor Consortium Study. J Clin Oncol 24, 3651-3656.  Coffin, C. M., Braun, J. T., Wick, M. R., and Dehner, L. P. (1990). A clinicopathologic and immunohistochemical analysis of 53 cases of medulloblastoma with emphasis on synaptophysin expression. Mod Pathol 3, 164-170.  Cogen, P. H., and McDonald, J. D. (1996). Tumor suppressor genes and medulloblastoma. J Neurooncol 29, 103-112.  Cogswell, J. P., Brown, C. E., Bisi, J. E., and Neill, S. D. (2000). Dominant-negative polo-like kinase 1 induces mitotic catastrophe independent of cdc25C function. Cell Growth Differ 11, 615-623.  Cohen, K. J., Pollack, I. F., Zhou, T., Buxton, A., Holmes, E. J., Burger, P. C., Brat, D. J., Rosenblum, M. K., Hamilton, R. L., Lavey, R. S., and Heideman, R. L. (2011). Temozolomide in the treatment of high-grade gliomas in children: a report from the Children's Oncology Group. Neuro Oncol 13, 317-323.  Conticello, C., Martinetti, D., Adamo, L., Buccheri, S., Giuffrida, R., Parrinello, N., Lombardo, L., Anastasi, G., Amato, G., Cavalli, M., et al. (2012). Disulfiram, an old drug with new potential therapeutic uses for human hematological malignancies. Int J Cancer 131, 2197-2203.  Corti, S., Locatelli, F., Papadimitriou, D., Donadoni, C., Salani, S., Del Bo, R., Strazzer, S., Bresolin, N., and Comi, G. P. (2006). Identification of a primitive brain-derived neural stem cell population based on aldehyde dehydrogenase activity. Stem Cells 24, 975-985.  Costello, J. F., Plass, C., Arap, W., Chapman, V. M., Held, W. A., Berger, M. S., Su Huang, H. J., and Cavenee, W. K. (1997). Cyclin-dependent kinase 6 (CDK6) amplification in human gliomas identified using two-dimensional separation of genomic DNA. Cancer Res 57, 1250-1254.  Crawford, J. R., MacDonald, T. J., and Packer, R. J. (2007). Medulloblastoma in childhood: new biological advances. Lancet Neurol 6, 1073-1085.  Cucchi, U., Gianellini, L.M., De Ponti, A., Sola, F., Alzani, R., Patton, V., Pezzoni, A., Troiani, S., Saccardo, M.B., Rizzi, S., et al. (2010). Phosphorylation of TCTP as a marker  155 for polo-like kinase-1 activity in vivo. Anticancer Res 30(12), 4973-85. D'Atri, S., Tentori, L., Lacal, P. M., Graziani, G., Pagani, E., Benincasa, E., Zambruno, G., Bonmassar, E., and Jiricny, J. (1998). Involvement of the mismatch repair system in temozolomide-induced apoptosis. Mol Pharmacol 54, 334-341.  Dacey, M. L., and Wallace, R. B. (1974). Postnatal neurogenesis in the feline cerebellum: a structural-functional investigation. Acta Neurobiol Exp (Wars) 34, 253-263.  Dahia, P. L. (2000). PTEN, a unique tumor suppressor gene. Endocr Relat Cancer 7, 115-129.  Dahmane, N., and Ruiz i Altaba, A. (1999). Sonic hedgehog regulates the growth and patterning of the cerebellum. Development 126, 3089-3100.  Dai, W., Li, Y., Ouyang, B., Pan, H., Reissmann, P., Li, J., Wiest, J., Stambrook, P., Gluckman, J. L., Noffsinger, A., and Bejarano, P. (2000). PRK, a cell cycle gene localized to 8p21, is downregulated in head and neck cancer. Genes Chromosomes Cancer 27, 332-336.  Dalerba, P., Dylla, S. J., Park, I. K., Liu, R., Wang, X., Cho, R. W., Hoey, T., Gurney, A., Huang, E. H., Simeone, D. M., et al. (2007). Phenotypic characterization of human colorectal cancer stem cells. Proc Natl Acad Sci U S A 104, 10158-10163.  de Carcer, G., Escobar, B., Higuero, A. M., Garcia, L., Anson, A., Perez, G., Mollejo, M., Manning, G., Melendez, B., Abad-Rodriguez, J., and Malumbres, M. (2011). Plk5, a polo box domain-only protein with specific roles in neuron differentiation and glioblastoma suppression. Mol Cell Biol 31, 1225-1239.  de Haas, T., Hasselt, N., Troost, D., Caron, H., Popovic, M., Zadravec-Zaletel, L., Grajkowska, W., Perek, M., Osterheld, M. C., Ellison, D., et al. (2008). Molecular risk stratification of medulloblastoma patients based on immunohistochemical analysis of MYC, LDHB, and CCNB1 expression. Clin Cancer Res 14, 4154-4160.  Dean, M. (2009). ABC transporters, drug resistance, and cancer stem cells. J Mammary Gland Biol Neoplasia 14, 3-9.  Dean, M., Fojo, T., and Bates, S. (2005). Tumour stem cells and drug resistance. Nat Rev Cancer 5, 275-284.  Dechant, R., and Glotzer, M. (2003). Centrosome separation and central spindle assembly act in redundant pathways that regulate microtubule density and trigger cleavage furrow formation. Dev Cell 4, 333-344.  Degenhardt, Y., Greshock, J., Laquerre, S., Gilmartin, A. G., Jing, J., Richter, M., Zhang, X., Bleam, M., Halsey, W., Hughes, A., et al. (2010). Sensitivity of cancer cells to Plk1 inhibitor GSK461364A is associated with loss of p53 function and chromosome instability. Molecular cancer therapeutics 9, 2079-2089.   156 Degenhardt, Y., and Lampkin, T. (2010). Targeting Polo-like kinase in cancer therapy. Clin Cancer Res 16, 384-389.  Deleyrolle, L. P., Harding, A., Cato, K., Siebzehnrubl, F. A., Rahman, M., Azari, H., Olson, S., Gabrielli, B., Osborne, G., Vescovi, A., and Reynolds, B. A. (2011). Evidence for label-retaining tumour-initiating cells in human glioblastoma. Brain 134, 1331-1343.  Deleyrolle, L. P., and Reynolds, B. A. (2009). Isolation, expansion, and differentiation of adult Mammalian neural stem and progenitor cells using the neurosphere assay. Methods Mol Biol 549, 91-101.  Dell'Albani, P. (2008). Stem cell markers in gliomas. Neurochem Res 33, 2407-2415.  Deutsch, M., Thomas, P. R., Krischer, J., Boyett, J. M., Albright, L., Aronin, P., Langston, J., Allen, J. C., Packer, R. J., Linggood, R., et al. (1996). Results of a prospective randomized trial comparing standard dose neuraxis irradiation (3,600 cGy/20) with reduced neuraxis irradiation (2,340 cGy/13) in patients with low-stage medulloblastoma. A Combined Children's Cancer Group-Pediatric Oncology Group Study. Pediatr Neurosurg 24, 167-176; discussion 176-167.  Dhall, G., Grodman, H., Ji, L., Sands, S., Gardner, S., Dunkel, I. J., McCowage, G. B., Diez, B., Allen, J. C., Gopalan, A., et al. (2008). Outcome of children less than three years old at diagnosis with non-metastatic medulloblastoma treated with chemotherapy on the "Head Start" I and II protocols. Pediatr Blood Cancer 50, 1169-1175.  Di Rocco, F., Carroll, R. S., Zhang, J., and Black, P. M. (1998). Platelet-derived growth factor and its receptor expression in human oligodendrogliomas. Neurosurgery 42, 341-346.  Diaz, L. A., Jr., Williams, R. T., Wu, J., Kinde, I., Hecht, J. R., Berlin, J., Allen, B., Bozic, I., Reiter, J. G., Nowak, M. A., et al. (2012). The molecular evolution of acquired resistance to targeted EGFR blockade in colorectal cancers. Nature 486, 537-540.  Dijkgraaf, G. J., Alicke, B., Weinmann, L., Januario, T., West, K., Modrusan, Z., Burdick, D., Goldsmith, R., Robarge, K., Sutherlin, D., et al. (2011). Small molecule inhibition of GDC-0449 refractory smoothened mutants and downstream mechanisms of drug resistance. Cancer Res 71, 435-444.  Dirks, P. B. (2006). Cancer: stem cells and brain tumours. Nature 444, 687-688. do Carmo Avides, M., Tavares, A., and Glover, D. M. (2001). Polo kinase and Asp are needed to promote the mitotic organizing activity of centrosomes. Nat Cell Biol 3, 421-424.  Dolan, M. E., Roy, S. K., Fasanmade, A. A., Paras, P. R., Schilsky, R. L., and Ratain, M. J. (1998). O6-benzylguanine in humans: metabolic, pharmacokinetic, and pharmacodynamic findings. J Clin Oncol 16, 1803-1810.  Donohue, P. J., Alberts, G. F., Guo, Y., and Winkles, J. A. (1995). Identification by targeted differential display of an immediate early gene encoding a putative serine/threonine kinase. J Biol Chem 270, 10351-10357.  157  Dufour, P., Lang, J. M., Giron, C., Duclos, B., Haehnel, P., Jaeck, D., Jung, J. M., and Oberling, F. (1993). Sodium dithiocarb as adjuvant immunotherapy for high risk breast cancer: a randomized study. Biotherapy 6, 9-12.  Eberhart, C. G., Tihan, T., and Burger, P. C. (2000). Nuclear localization and mutation of beta-catenin in medulloblastomas. J Neuropathol Exp Neurol 59, 333-337.  Eckerdt, F., Yuan, J., Saxena, K., Martin, B., Kappel, S., Lindenau, C., Kramer, A., Naumann, S., Daum, S., Fischer, G., et al. (2005). Polo-like kinase 1-mediated phosphorylation stabilizes Pin1 by inhibiting its ubiquitination in human cells. J Biol Chem 280, 36575-36583.  Ekstrand, A. J., Sugawa, N., James, C. D., and Collins, V. P. (1992). Amplified and rearranged epidermal growth factor receptor genes in human glioblastomas reveal deletions of sequences encoding portions of the N- and/or C-terminal tails. Proc Natl Acad Sci U S A 89, 4309-4313.  Elez, R., Piiper, A., Giannini, C. D., Brendel, M., and Zeuzem, S. (2000). Polo-like kinase1, a new target for antisense tumor therapy. Biochem Biophys Res Commun 269, 352-356.  Elia, A. E., Cantley, L. C., and Yaffe, M. B. (2003a). Proteomic screen finds pSer/pThr-binding domain localizing Plk1 to mitotic substrates. Science 299, 1228-1231.  Elia, A. E., Rellos, P., Haire, L. F., Chao, J. W., Ivins, F. J., Hoepker, K., Mohammad, D., Cantley, L. C., Smerdon, S. J., and Yaffe, M. B. (2003b). The molecular basis for phosphodependent substrate targeting and regulation of Plks by the Polo-box domain. Cell 115, 83-95.  Ellinger-Ziegelbauer, H., Karasuyama, H., Yamada, E., Tsujikawa, K., Todokoro, K., and Nishida, E. (2000). Ste20-like kinase (SLK), a regulatory kinase for polo-like kinase (Plk) during the G2/M transition in somatic cells. Genes Cells 5, 491-498.  Ellison, D. W. (2010). Childhood medulloblastoma: novel approaches to the classification of a heterogeneous disease. Acta Neuropathol 120, 305-316.  Ellison, D. W., Dalton, J., Kocak, M., Nicholson, S. L., Fraga, C., Neale, G., Kenney, A. M., Brat, D. J., Perry, A., Yong, W. H., et al. (2011a). Medulloblastoma: clinicopathological correlates of SHH, WNT, and non-SHH/WNT molecular subgroups. Acta Neuropathol 121, 381-396.  Ellison, D. W., Kocak, M., Dalton, J., Megahed, H., Lusher, M. E., Ryan, S. L., Zhao, W., Nicholson, S. L., Taylor, R. E., Bailey, S., and Clifford, S. C. (2011b). Definition of disease-risk stratification groups in childhood medulloblastoma using combined clinical, pathologic, and molecular variables. J Clin Oncol 29, 1400-1407.  Ellison, D. W., Onilude, O. E., Lindsey, J. C., Lusher, M. E., Weston, C. L., Taylor, R. E., Pearson, A. D., and Clifford, S. C. (2005). beta-Catenin status predicts a favorable outcome in childhood medulloblastoma: the United Kingdom Children's Cancer Study Group Brain Tumour Committee. J Clin Oncol 23, 7951-7957.  158  Eneanya, D. I., Bianchine, J. R., Duran, D. O., and Andresen, B. D. (1981). The actions of metabolic fate of disulfiram. Annu Rev Pharmacol Toxicol 21, 575-596.  Entz-Werle, N., Carli, E. D., Ducassou, S., Legrain, M., Grill, J., and Dufour, C. (2008). Medulloblastoma: what is the role of molecular genetics? Expert Rev Anticancer Ther 8, 1169-1181.  Eramo, A., Ricci-Vitiani, L., Zeuner, A., Pallini, R., Lotti, F., Sette, G., Pilozzi, E., Larocca, L. M., Peschle, C., and De Maria, R. (2006). Chemotherapy resistance of glioblastoma stem cells. Cell Death Differ 13, 1238-1241.  Esteller, M., Garcia-Foncillas, J., Andion, E., Goodman, S. N., Hidalgo, O. F., Vanaclocha, V., Baylin, S. B., and Herman, J. G. (2000). Inactivation of the DNA-repair gene MGMT and the clinical response of gliomas to alkylating agents. N Engl J Med 343, 1350-1354.  Evans, A. E., Jenkin, R. D., Sposto, R., Ortega, J. A., Wilson, C. B., Wara, W., Ertel, I. J., Kramer, S., Chang, C. H., Leikin, S. L., and et al. (1990). The treatment of medulloblastoma. Results of a prospective randomized trial of radiation therapy with and without CCNU, vincristine, and prednisone. J Neurosurg 72, 572-582.  Eyre, H. J., Quagliana, J. M., Eltringham, J. R., Frank, J., O'Bryan, R. M., McDonald, B., and Rivkin, S. E. (1983). Randomized comparisons of radiotherapy and CCNU versus radiotherapy, CCNU plus procarbazine for the treatment of malignant gliomas following surgery. A Southwest Oncology Group Report. J Neurooncol 1, 171-177.  Faiman, M. D., Artman, L., and Haya, K. (1980). Disulfiram distribution and elimination in the rat after oral and intraperitoneal administration. Alcohol Clin Exp Res 4, 412-419.  Fallon, P., Gentry, T., Balber, A. E., Boulware, D., Janssen, W. E., Smilee, R., Storms, R. W., and Smith, C. (2003). Mobilized peripheral blood SSCloALDHbr cells have the phenotypic and functional properties of primitive haematopoietic cells and their number correlates with engraftment following autologous transplantation. Br J Haematol 122, 99-108.  Fang, D., Nguyen, T. K., Leishear, K., Finko, R., Kulp, A. N., Hotz, S., Van Belle, P. A., Xu, X., Elder, D. E., and Herlyn, M. (2005). A tumorigenic subpopulation with stem cell properties in melanomas. Cancer Res 65, 9328-9337.  Faury, D., Nantel, A., Dunn, S. E., Guiot, M. C., Haque, T., Hauser, P., Garami, M., Bognar, L., Hanzely, Z., Liberski, P. P., et al. (2007). Molecular profiling identifies prognostic subgroups of pediatric glioblastoma and shows increased YB-1 expression in tumors. J Clin Oncol 25, 1196-1208.  Feng, R., Zhou, S., Liu, Y., Song, D., Luan, Z., Dai, X., Li, Y., Tang, N., Wen, J., and Li, L. (2013). Sox2 protects neural stem cells from apoptosis via up-regulating survivin expression. Biochem J 450(3), 459-68.  Feng, Y., Yuan, J. H., Maloid, S. C., Fisher, R., Copeland, T. D., Longo, D. L., Conrads, T. P., Veenstra, T. D., Ferris, A., Hughes, S., et al. (2006). Polo-like kinase 1-mediated  159 phosphorylation of the GTP-binding protein Ran is important for bipolar spindle formation. Biochem Biophys Res Commun 349, 144-152.  Fenton, B., and Glover, D. M. (1993). A conserved mitotic kinase active at late anaphase-telophase in syncytial Drosophila embryos. Nature 363, 637-640.  Fine, H. A., Dear, K. B., Loeffler, J. S., Black, P. M., and Canellos, G. P. (1993). Meta-analysis of radiation therapy with and without adjuvant chemotherapy for malignant gliomas in adults. Cancer 71, 2585-2597.  Finkbeiner, M. R., Astanehe, A., To, K., Fotovati, A., Davies, A. H., Zhao, Y., Jiang, H., Stratford, A. L., Shadeo, A., Boccaccio, C., et al. (2009). Profiling YB-1 target genes uncovers a new mechanism for MET receptor regulation in normal and malignant human mammary cells. Oncogene 28, 1421-1431.  Fischer, P. M., and Lane, D. P. (2000). Inhibitors of cyclin-dependent kinases as anti-cancer therapeutics. Curr Med Chem 7, 1213-1245.  Fisher, J. L., Schwartzbaum, J. A., Wrensch, M., and Wiemels, J. L. (2007). Epidemiology of brain tumors. Neurol Clin 25, 867-890, vii.  Fode, C., Binkert, C., and Dennis, J. W. (1996). Constitutive expression of murine Sak-a suppresses cell growth and induces multinucleation. Mol Cell Biol 16, 4665-4672.  Fode, C., Motro, B., Yousefi, S., Heffernan, M., and Dennis, J. W. (1994). Sak, a murine protein-serine/threonine kinase that is related to the Drosophila polo kinase and involved in cell proliferation. Proc Natl Acad Sci U S A 91, 6388-6392.  Foshay, K. M., and Gallicano, G. I. (2008). Regulation of Sox2 by STAT3 initiates commitment to the neural precursor cell fate. Stem Cells Dev 17, 269-278.  Fotovati, A., Abu-Ali, S., Wang, P. S., Deleyrolle, L. P., Lee, C., Triscott, J., Chen, J. Y., Franciosi, S., Nakamura, Y., Sugita, Y., et al. (2011). YB-1 bridges neural stem cells and brain-tumor initiating cells via its roles in differentiation and cell growth. Cancer Res 71(16), 5569-78.  Fouladi, M., Gilger, E., Kocak, M., Wallace, D., Buchanan, G., Reeves, C., Robbins, N., Merchant, T., Kun, L. E., Khan, R., et al. (2005). Intellectual and functional outcome of children 3 years old or younger who have CNS malignancies. J Clin Oncol 23, 7152-7160.  Fox, N., Damjanov, I., Knowles, B. B., and Solter, D. (1983). Immunohistochemical localization of the mouse stage-specific embryonic antigen 1 in human tissues and tumors. Cancer Res 43, 669-678.  Francescangeli, F., Patrizii, M., Signore, M., Federici, G., Di Franco, S., Pagliuca, A., Baiocchi, M., Biffoni, M., Ricci Vitiani, L., Todaro, M., et al. (2012). Proliferation state and polo-like kinase1 dependence of tumorigenic colon cancer cells. Stem Cells 30, 1819-1830.   160 Frederick, L., Wang, X. Y., Eley, G., and James, C. D. (2000). Diversity and frequency of epidermal growth factor receptor mutations in human glioblastomas. Cancer Res 60, 1383-1387.  Friedman, H. S. (2000). Can O6-alkylguanine-DNA alkyltransferase depletion enhance alkylator activity in the clinic? Clin Cancer Res 6, 2967-2968.  Friedman, H. S., Prados, M. D., Wen, P. Y., Mikkelsen, T., Schiff, D., Abrey, L. E., Yung, W. K., Paleologos, N., Nicholas, M. K., Jensen, R., et al. (2009). Bevacizumab alone and in combination with irinotecan in recurrent glioblastoma. J Clin Oncol 27, 4733-4740.  Friedmann-Morvinski, D., Bushong, E. A., Ke, E., Soda, Y., Marumoto, T., Singer, O., Ellisman, M. H., and Verma, I. M. (2012). Dedifferentiation of neurons and astrocytes by oncogenes can induce gliomas in mice. Science 338, 1080-1084.  Fukuda, S., Kato, F., Tozuka, Y., Yamaguchi, M., Miyamoto, Y., and Hisatsune, T. (2003). Two distinct subpopulations of nestin-positive cells in adult mouse dentate gyrus. J Neurosci 23, 9357-9366.  Gage, F. H. (2002). Neurogenesis in the adult brain. J Neurosci 22, 612-613.  Gajjar, A., Chintagumpala, M., Ashley, D., Kellie, S., Kun, L. E., Merchant, T. E., Woo, S., Wheeler, G., Ahern, V., Krasin, M. J., et al. (2006). Risk-adapted craniospinal radiotherapy followed by high-dose chemotherapy and stem-cell rescue in children with newly diagnosed medulloblastoma (St Jude Medulloblastoma-96): long-term results from a prospective, multicentre trial. Lancet Oncol 7, 813-820.  Gajjar, A., Hernan, R., Kocak, M., Fuller, C., Lee, Y., McKinnon, P. J., Wallace, D., Lau, C., Chintagumpala, M., Ashley, D. M., et al. (2004). Clinical, histopathologic, and molecular markers of prognosis: toward a new disease risk stratification system for medulloblastoma. J Clin Oncol 22, 984-993.  Galanis, E., Buckner, J. C., Maurer, M. J., Kreisberg, J. I., Ballman, K., Boni, J., Peralba, J. M., Jenkins, R. B., Dakhil, S. R., Morton, R. F., et al. (2005). Phase II trial of temsirolimus (CCI-779) in recurrent glioblastoma multiforme: a North Central Cancer Treatment Group Study. J Clin Oncol 23, 5294-5304.  Galgoczy, D. J., and Toczyski, D. P. (2001). Checkpoint adaptation precedes spontaneous and damage-induced genomic instability in yeast. Mol Cell Biol 21, 1710-1718.  Galli, R., Binda, E., Orfanelli, U., Cipelletti, B., Gritti, A., De Vitis, S., Fiocco, R., Foroni, C., Dimeco, F., and Vescovi, A. (2004). Isolation and characterization of tumorigenic, stem-like neural precursors from human glioblastoma. Cancer Res 64, 7011-7021.  Gangemi, R. M., Griffero, F., Marubbi, D., Perera, M., Capra, M. C., Malatesta, P., Ravetti, G. L., Zona, G. L., Daga, A., and Corte, G. (2009). SOX2 silencing in glioblastoma tumor-initiating cells causes stop of proliferation and loss of tumorigenicity. Stem Cells 27, 40-48.   161 Gao, Y., Fotovati, A., Lee, C., Wang, M., Cote, G., Guns, E., Toyota, B., Faury, D., Jabado, N., and Dunn, S. E. (2009). Inhibition of Y-box binding protein-1 slows the growth of glioblastoma multiforme and sensitizes to temozolomide independent O6-methylguanine-DNA methyltransferase. Mol Cancer Ther 8, 3276-3284.  Garuti, L., Roberti, M., and Bottegoni, G. (2012). Polo-like kinases inhibitors. Curr Med Chem 19, 3937-3948.  Gerlinger, M., and Swanton, C. (2010). How Darwinian models inform therapeutic failure initiated by clonal heterogeneity in cancer medicine. British journal of cancer 103, 1139-1143.  Gerster, K., Shi, W., Ng, B., Yue, S., Ito, E., Waldron, J., Gilbert, R., and Liu, F. F. (2010). Targeting polo-like kinase 1 enhances radiation efficacy for head-and-neck squamous cell carcinoma. Int J Radiat Oncol Biol Phys 77, 253-260.  Gerstner, E. R., Yip, S., Wang, D. L., Louis, D. N., Iafrate, A. J., and Batchelor, T. T. (2009). Mgmt methylation is a prognostic biomarker in elderly patients with newly diagnosed glioblastoma. Neurology 73, 1509-1510.  Giangaspero, F., Rigobello, L., Badiali, M., Loda, M., Andreini, L., Basso, G., Zorzi, F., and Montaldi, A. (1992). Large-cell medulloblastomas. A distinct variant with highly aggressive behavior. Am J Surg Pathol 16, 687-693.  Gibson, P., Tong, Y., Robinson, G., Thompson, M. C., Currle, D. S., Eden, C., Kranenburg, T. A., Hogg, T., Poppleton, H., Martin, J., et al. (2010). Subtypes of medulloblastoma have distinct developmental origins. Nature 468, 1095-1099.  Gilbertson, R. J., Hill, D. A., Hernan, R., Kocak, M., Geyer, R., Olson, J., Gajjar, A., Rush, L., Hamilton, R. L., Finkelstein, S. D., and Pollack, I. F. (2003). ERBB1 is amplified and overexpressed in high-grade diffusely infiltrative pediatric brain stem glioma. Clinical cancer research : an official journal of the American Association for Cancer Research 9, 3620-3624.  Gilbertson, R. J., Pearson, A. D., Perry, R. H., Jaros, E., and Kelly, P. J. (1995). Prognostic significance of the c-erbB-2 oncogene product in childhood medulloblastoma. Br J Cancer 71, 473-477.  Gilbertson, R. J., Perry, R. H., Kelly, P. J., Pearson, A. D., and Lunec, J. (1997). Prognostic significance of HER2 and HER4 coexpression in childhood medulloblastoma. Cancer Res 57, 3272-3280.  Glinsky, G. V. (2006). Genomic models of metastatic cancer: functional analysis of death-from-cancer signature genes reveals aneuploid, anoikis-resistant, metastasis-enabling phenotype with altered cell cycle control and activated Polycomb Group (PcG) protein chromatin silencing pathway. Cell Cycle 5, 1208-1216.  Glover, D. M., Hagan, I. M., and Tavares, A. A. (1998). Polo-like kinases: a team that plays throughout mitosis. Genes Dev 12, 3777-3787.   162 Godard, S., Getz, G., Delorenzi, M., Farmer, P., Kobayashi, H., Desbaillets, I., Nozaki, M., Diserens, A. C., Hamou, M. F., Dietrich, P. Y., et al. (2003). Classification of human astrocytic gliomas on the basis of gene expression: a correlated group of genes with angiogenic activity emerges as a strong predictor of subtypes. Cancer Res 63, 6613-6625.  Golan, A., Yudkovsky, Y., and Hershko, A. (2002). The cyclin-ubiquitin ligase activity of cyclosome/APC is jointly activated by protein kinases Cdk1-cyclin B and Plk. J Biol Chem 277, 15552-15557.  Goldman, S. A., and Nottebohm, F. (1983). Neuronal production, migration, and differentiation in a vocal control nucleus of the adult female canary brain. Proc Natl Acad Sci U S A 80, 2390-2394.  Golsteyn, R. M., Mundt, K. E., Fry, A. M., and Nigg, E. A. (1995). Cell cycle regulation of the activity and subcellular localization of Plk1, a human protein kinase implicated in mitotic spindle function. J Cell Biol 129, 1617-1628.  Golsteyn, R. M., Schultz, S. J., Bartek, J., Ziemiecki, A., Ried, T., and Nigg, E. A. (1994). Cell cycle analysis and chromosomal localization of human Plk1, a putative homologue of the mitotic kinases Drosophila polo and Saccharomyces cerevisiae Cdc5. J Cell Sci 107 ( Pt 6), 1509-1517.  Goodrich, L. V., Milenkovic, L., Higgins, K. M., and Scott, M. P. (1997). Altered neural cell fates and medulloblastoma in mouse patched mutants. Science 277, 1109-1113.  Gopinath, S., Malla, R., Alapati, K., Gorantla, B., Gujrati, M., Dinh, D. H., and Rao, J. S. (2013). Cathepsin B and uPAR regulate self-renewal of glioma-initiating cells through GLI-regulated Sox2 and Bmi1 expression. Carcinogenesis.  Gottardo, N. G., and Gajjar, A. (2006). Current therapy for medulloblastoma. Curr Treat Options Neurol 8, 319-334.  Graham, V., Khudyakov, J., Ellis, P., and Pevny, L. (2003). SOX2 functions to maintain neural progenitor identity. Neuron 39, 749-765.  Greenberger, L. M., Lothstein, L., Williams, S. S., and Horwitz, S. B. (1988). Distinct P-glycoprotein precursors are overproduced in independently isolated drug-resistant cell lines. Proc Natl Acad Sci U S A 85, 3762-3766.  Grill, J., Sainte-Rose, C., Jouvet, A., Gentet, J. C., Lejars, O., Frappaz, D., Doz, F., Rialland, X., Pichon, F., Bertozzi, A. I., et al. (2005). Treatment of medulloblastoma with postoperative chemotherapy alone: an SFOP prospective trial in young children. Lancet Oncol 6, 573-580.  Grinshtein, N., Datti, A., Fujitani, M., Uehling, D., Prakesch, M., Isaac, M., Irwin, M. S., Wrana, J. L., Al-Awar, R., and Kaplan, D. R. (2011). Small molecule kinase inhibitor screen identifies polo-like kinase 1 as a target for neuroblastoma tumor-initiating cells. Cancer Res 71, 1385-1395.   163 Gritti, A., Parati, E. A., Cova, L., Frolichsthal, P., Galli, R., Wanke, E., Faravelli, L., Morassutti, D. J., Roisen, F., Nickel, D. D., and Vescovi, A. L. (1996). Multipotential stem cells from the adult mouse brain proliferate and self-renew in response to basic fibroblast growth factor. J Neurosci 16, 1091-1100.  Gubbay, J., Collignon, J., Koopman, P., Capel, B., Economou, A., Munsterberg, A., Vivian, N., Goodfellow, P., and Lovell-Badge, R. (1990). A gene mapping to the sex-determining region of the mouse Y chromosome is a member of a novel family of embryonically expressed genes. Nature 346, 245-250.  Guerreiro, A. S., Fattet, S., Kulesza, D. W., Atamer, A., Elsing, A. N., Shalaby, T., Jackson, S. P., Schoenwaelder, S. M., Grotzer, M. A., Delattre, O., and Arcaro, A. (2011). A sensitized RNA interference screen identifies a novel role for the PI3K p110gamma isoform in medulloblastoma cell proliferation and chemoresistance. Mol Cancer Res 9, 925-935.  Guex, N., and Peitsch, M. C. (1997). SWISS-MODEL and the Swiss-PdbViewer: an environment for comparative protein modeling. Electrophoresis 18, 2714-2723.  Guha, A., Dashner, K., Black, P. M., Wagner, J. A., and Stiles, C. D. (1995). Expression of PDGF and PDGF receptors in human astrocytoma operation specimens supports the existence of an autocrine loop. Int J Cancer 60, 168-173.  Guiu, S., Taillibert, S., Chinot, O., Taillandier, L., Honnorat, J., Dietrich, P. Y., Maire, J. P., Guillamo, J. S., Guiu, B., Catry-Thomas, I., et al. (2008). [Bevacizumab/irinotecan. An active treatment for recurrent high grade gliomas: preliminary results of an ANOCEF Multicenter Study]. Rev Neurol (Paris) 164, 588-594.  Gumireddy, K., Reddy, M. V., Cosenza, S. C., Boominathan, R., Baker, S. J., Papathi, N., Jiang, J., Holland, J., and Reddy, E. P. (2005). ON01910, a non-ATP-competitive small molecule inhibitor of Plk1, is a potent anticancer agent. Cancer Cell 7, 275-286.  Guo, X., Xu, B., Pandey, S., Goessl, E., Brown, J., Armesilla, A. L., Darling, J. L., and Wang, W. (2010a). Disulfiram/copper complex inhibiting NFkappaB activity and potentiating cytotoxic effect of gemcitabine on colon and breast cancer cell lines. Cancer Lett 290, 104-113.  Guo, X., Xu, B., Pandey, S., Goessl, E., Brown, J., Armesilla, A. L., Darling, J. L., and Wang, W. (2010b). Disulfiram/copper complex inhibiting NFkappaB activity and potentiating cytotoxic effect of gemcitabine on colon and breast cancer cell lines. Cancer letters 290, 104-113.  Guryanova, O. A., Wu, Q., Cheng, L., Lathia, J. D., Huang, Z., Yang, J., MacSwords, J., Eyler, C. E., McLendon, R. E., Heddleston, J. M., et al. (2011). Nonreceptor tyrosine kinase BMX maintains self-renewal and tumorigenic potential of glioblastoma stem cells by activating STAT3. Cancer Cell 19, 498-511.  Hahn, H., Wicking, C., Zaphiropoulous, P. G., Gailani, M. R., Shanley, S., Chidambaram, A., Vorechovsky, I., Holmberg, E., Unden, A. B., Gillies, S., et al. (1996). Mutations of the  164 human homolog of Drosophila patched in the nevoid basal cell carcinoma syndrome. Cell 85, 841-851.  Hamanaka, R., Maloid, S., Smith, M. R., O'Connell, C. D., Longo, D. L., and Ferris, D. K. (1994). Cloning and characterization of human and murine homologues of the Drosophila polo serine-threonine kinase. Cell Growth Differ 5, 249-257.  Hambardzumyan, D., Becher, O. J., Rosenblum, M. K., Pandolfi, P. P., Manova-Todorova, K., and Holland, E. C. (2008). PI3K pathway regulates survival of cancer stem cells residing in the perivascular niche following radiation in medulloblastoma in vivo. Genes Dev 22, 436-448.  Hamilton, S. R., Liu, B., Parsons, R. E., Papadopoulos, N., Jen, J., Powell, S. M., Krush, A. J., Berk, T., Cohen, Z., Tetu, B., and et al. (1995). The molecular basis of Turcot's syndrome. N Engl J Med 332, 839-847.  Han, X., Fang, X., Lou, X., Hua, D., Ding, W., Foltz, G., Hood, L., Yuan, Y., and Lin, B. (2012). Silencing SOX2 induced mesenchymal-epithelial transition and its expression predicts liver and lymph node metastasis of CRC patients. PLoS One 7, e41335.  Hanks, S. K., and Quinn, A. M. (1991). Protein kinase catalytic domain sequence database: identification of conserved features of primary structure and classification of family members. Methods Enzymol 200, 38-62.  Hanks, S. K., Quinn, A. M., and Hunter, T. (1988). The protein kinase family: conserved features and deduced phylogeny of the catalytic domains. Science 241, 42-52.  Hansen, D. V., Loktev, A. V., Ban, K. H., and Jackson, P. K. (2004). Plk1 regulates activation of the anaphase promoting complex by phosphorylating and triggering SCFbetaTrCP-dependent destruction of the APC Inhibitor Emi1. Mol Biol Cell 15, 5623-5634.  Hansford, L. M., McKee, A. E., Zhang, L., George, R. E., Gerstle, J. T., Thorner, P. S., Smith, K. M., Look, A. T., Yeger, H., Miller, F. D., et al. (2007). Neuroblastoma cells isolated from bone marrow metastases contain a naturally enriched tumor-initiating cell. Cancer Res 67, 11234-11243.  Hari, M., Loganzo, F., Annable, T., Tan, X., Musto, S., Morilla, D. B., Nettles, J. H., Snyder, J. P., and Greenberger, L. M. (2006). Paclitaxel-resistant cells have a mutation in the paclitaxel-binding region of beta-tubulin (Asp26Glu) and less stable microtubules. Mol Cancer Ther 5, 270-278.  Harris, P. S., Venkataraman, S., Alimova, I., Birks, D. K., Donson, A. M., Knipstein, J., Dubuc, A., Taylor, M. D., Handler, M. H., Foreman, N. K., and Vibhakar, R. (2012). Polo-like kinase 1 (PLK1) inhibition suppresses cell growth and enhances radiation sensitivity in medulloblastoma cells. BMC Cancer 12, 80.  Hatten, M. E., and Roussel, M. F. (2011). Development and cancer of the cerebellum. Trends Neurosci 34, 134-142.   165 Hatton, B. A., Villavicencio, E. H., Tsuchiya, K. D., Pritchard, J. I., Ditzler, S., Pullar, B., Hansen, S., Knoblaugh, S. E., Lee, D., Eberhart, C. G., et al. (2008). The Smo/Smo model: hedgehog-induced medulloblastoma with 90% incidence and leptomeningeal spread. Cancer Res 68, 1768-1776.  Hegi, M. E., Diserens, A. C., Gorlia, T., Hamou, M. F., de Tribolet, N., Weller, M., Kros, J. M., Hainfellner, J. A., Mason, W., Mariani, L., et al. (2005). MGMT gene silencing and benefit from temozolomide in glioblastoma. N Engl J Med 352, 997-1003.  Hemmati, H. D., Nakano, I., Lazareff, J. A., Masterman-Smith, M., Geschwind, D. H., Bronner-Fraser, M., and Kornblum, H. I. (2003). Cancerous stem cells can arise from pediatric brain tumors. Proc Natl Acad Sci U S A 100, 15178-15183.  Hermanson, M., Funa, K., Hartman, M., Claesson-Welsh, L., Heldin, C. H., Westermark, B., and Nister, M. (1992). Platelet-derived growth factor and its receptors in human glioma tissue: expression of messenger RNA and protein suggests the presence of autocrine and paracrine loops. Cancer Res 52, 3213-3219.  Hernan, R., Fasheh, R., Calabrese, C., Frank, A. J., Maclean, K. H., Allard, D., Barraclough, R., and Gilbertson, R. J. (2003). ERBB2 up-regulates S100A4 and several other prometastatic genes in medulloblastoma. Cancer Res 63, 140-148.  Hess, D. A., Wirthlin, L., Craft, T. P., Herrbrich, P. E., Hohm, S. A., Lahey, R., Eades, W. C., Creer, M. H., and Nolta, J. A. (2006). Selection based on CD133 and high aldehyde dehydrogenase activity isolates long-term reconstituting human hematopoietic stem cells. Blood 107, 2162-2169.  Hickman, J. A., Stevens, M. F., Gibson, N. W., Langdon, S. P., Fizames, C., Lavelle, F., Atassi, G., Lunt, E., and Tilson, R. M. (1985). Experimental antitumor activity against murine tumor model systems of 8-carbamoyl-3-(2-chloroethyl)imidazo[5,1-d]-1,2,3,5-tetrazin-4(3 H)-one (mitozolomide), a novel broad-spectrum agent. Cancer Res 45, 3008-3013.  Hirschmann-Jax, C., Foster, A. E., Wulf, G. G., Goodell, M. A., and Brenner, M. K. (2005). A distinct "side population" of cells in human tumor cells: implications for tumor biology and therapy. Cell Cycle 4, 203-205.  Hirschmann-Jax, C., Foster, A. E., Wulf, G. G., Nuchtern, J. G., Jax, T. W., Gobel, U., Goodell, M. A., and Brenner, M. K. (2004). A distinct "side population" of cells with high drug efflux capacity in human tumor cells. Proc Natl Acad Sci U S A 101, 14228-14233.  Ho, M. M., Ng, A. V., Lam, S., and Hung, J. Y. (2007). Side population in human lung cancer cell lines and tumors is enriched with stem-like cancer cells. Cancer Res 67, 4827-4833.  Hofheinz, R. D., Al-Batran, S. E., Hochhaus, A., Jager, E., Reichardt, V. L., Fritsch, H., Trommeshauser, D., and Munzert, G. (2010). An open-label, phase I study of the polo-like kinase-1 inhibitor, BI 2536, in patients with advanced solid tumors. Clin Cancer Res 16, 4666-4674.   166 Holland, E. C. (2001). Gliomagenesis: genetic alterations and mouse models. Nat Rev Genet 2, 120-129.  Holland, E. C., Celestino, J., Dai, C., Schaefer, L., Sawaya, R. E., and Fuller, G. N. (2000). Combined activation of Ras and Akt in neural progenitors induces glioblastoma formation in mice. Nat Genet 25, 55-57.  Holland, E. C., Hively, W. P., DePinho, R. A., and Varmus, H. E. (1998). A constitutively active epidermal growth factor receptor cooperates with disruption of G1 cell-cycle arrest pathways to induce glioma-like lesions in mice. Genes Dev 12, 3675-3685.  Holtrich, U., Wolf, G., Brauninger, A., Karn, T., Bohme, B., Rubsamen-Waigmann, H., and Strebhardt, K. (1994). Induction and down-regulation of PLK, a human serine/threonine kinase expressed in proliferating cells and tumors. Proc Natl Acad Sci U S A 91, 1736-1740.  Holtrich, U., Wolf, G., Yuan, J., Bereiter-Hahn, J., Karn, T., Weiler, M., Kauselmann, G., Rehli, M., Andreesen, R., Kaufmann, M., et al. (2000). Adhesion induced expression of the serine/threonine kinase Fnk in human macrophages. Oncogene 19, 4832-4839.  Hope, K. J., Jin, L., and Dick, J. E. (2004). Acute myeloid leukemia originates from a hierarchy of leukemic stem cell classes that differ in self-renewal capacity. Nat Immunol 5, 738-743.  Horton, J. K., and Wilson, S. H. (2007). Hypersensitivity phenotypes associated with genetic and synthetic inhibitor-induced base excision repair deficiency. DNA Repair (Amst) 6, 530-543.  Hu, C., Li, H., Li, J., Zhu, Z., Yin, S., Hao, X., Yao, M., Zheng, S., and Gu, J. (2008). Analysis of ABCG2 expression and side population identifies intrinsic drug efflux in the HCC cell line MHCC-97L and its modulation by Akt signaling. Carcinogenesis 29, 2289-2297.  Hu, K., Law, J. H., Fotovati, A., and Dunn, S. E. (2012). Small interfering RNA library screen identified polo-like kinase-1 (PLK1) as a potential therapeutic target for breast cancer that uniquely eliminates tumor-initiating cells. Breast cancer research : BCR 14, R22.  Hu, K., Lee, C., Qiu, D., Fotovati, A., Davies, A., Abu-Ali, S., Wai, D., Lawlor, E. R., Triche, T. J., Pallen, C. J., and Dunn, S. E. (2009). Small interfering RNA library screen of human kinases and phosphatases identifies polo-like kinase 1 as a promising new target for the treatment of pediatric rhabdomyosarcomas. Mol Cancer Ther 8, 3024-3035.  Hu, L., McArthur, C., and Jaffe, R. B. (2010). Ovarian cancer stem-like side-population cells are tumourigenic and chemoresistant. Br J Cancer 102, 1276-1283.  Hudson, J. W., Kozarova, A., Cheung, P., Macmillan, J. C., Swallow, C. J., Cross, J. C., and Dennis, J. W. (2001). Late mitotic failure in mice lacking Sak, a polo-like kinase. Curr Biol 11, 441-446.   167 Hui, A. B., Lo, K. W., Yin, X. L., Poon, W. S., and Ng, H. K. (2001). Detection of multiple gene amplifications in glioblastoma multiforme using array-based comparative genomic hybridization. Lab Invest 81, 717-723.  Huse, J. T., and Holland, E. C. (2010). Targeting brain cancer: advances in the molecular pathology of malignant glioma and medulloblastoma. Nat Rev Cancer 10, 319-331.  Hussein, D., Punjaruk, W., Storer, L. C., Shaw, L., Ottoman, R., Peet, A., Miller, S., Bandopadhyay, G., Heath, R., Kumari, R., et al. (2011). Pediatric brain tumor cancer stem cells: cell cycle dynamics, DNA repair, and etoposide extrusion. Neuro Oncol 13, 70-83.  Ichimura, K., Schmidt, E. E., Goike, H. M., and Collins, V. P. (1996). Human glioblastomas with no alterations of the CDKN2A (p16INK4A, MTS1) and CDK4 genes have frequent mutations of the retinoblastoma gene. Oncogene 13, 1065-1072.  Ignatova, T. N., Kukekov, V. G., Laywell, E. D., Suslov, O. N., Vrionis, F. D., and Steindler, D. A. (2002). Human cortical glial tumors contain neural stem-like cells expressing astroglial and neuronal markers in vitro. Glia 39, 193-206.  Ikushima, H., Todo, T., Ino, Y., Takahashi, M., Saito, N., Miyazawa, K., and Miyazono, K. (2011a). Glioma-initiating Cells Retain Their Tumorigenicity through Integration of the Sox Axis and Oct4. J Biol Chem.  Ikushima, H., Todo, T., Ino, Y., Takahashi, M., Saito, N., Miyazawa, K., and Miyazono, K. (2011b). Glioma-initiating cells retain their tumorigenicity through integration of the Sox axis and Oct4 protein. J Biol Chem 286, 41434-41441.  Iljin, K., Ketola, K., Vainio, P., Halonen, P., Kohonen, P., Fey, V., Grafstrom, R. C., Perala, M., and Kallioniemi, O. (2009). High-throughput cell-based screening of 4910 known drugs and drug-like small molecules identifies disulfiram as an inhibitor of prostate cancer cell growth. Clin Cancer Res 15, 6070-6078.  Inglis, K. J., Chereau, D., Brigham, E. F., Chiou, S. S., Schobel, S., Frigon, N. L., Yu, M., Caccavello, R. J., Nelson, S., Motter, R., et al. (2009). Polo-like kinase 2 (PLK2) phosphorylates alpha-synuclein at serine 129 in central nervous system. J Biol Chem 284, 2598-2602.  Irizarry, R. A., Hobbs, B., Collin, F., Beazer-Barclay, Y. D., Antonellis, K. J., Scherf, U., and Speed, T. P. (2003). Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics 4, 249-264.  Iwamoto, F. M., Abrey, L. E., Beal, K., Gutin, P. H., Rosenblum, M. K., Reuter, V. E., DeAngelis, L. M., and Lassman, A. B. (2009). Patterns of relapse and prognosis after bevacizumab failure in recurrent glioblastoma. Neurology 73, 1200-1206.  Jackman, M., Lindon, C., Nigg, E. A., and Pines, J. (2003). Active cyclin B1-Cdk1 first appears on centrosomes in prophase. Nat Cell Biol 5, 143-148.  Jaeckle, K. A., Eyre, H. J., Townsend, J. J., Schulman, S., Knudson, H. M., Belanich, M., Yarosh, D. B., Bearman, S. I., Giroux, D. J., and Schold, S. C. (1998). Correlation of tumor  168 O6 methylguanine-DNA methyltransferase levels with survival of malignant astrocytoma patients treated with bis-chloroethylnitrosourea: a Southwest Oncology Group study. J Clin Oncol 16, 3310-3315.  Jain, A. K., Allton, K., Iacovino, M., Mahen, E., Milczarek, R. J., Zwaka, T. P., Kyba, M., and Barton, M. C. (2012). p53 regulates cell cycle and microRNAs to promote differentiation of human embryonic stem cells. PLoS Biol 10, e1001268.  Jandial, R., Waters, D. J., and Chen, M. Y. (2011). Cancer stem cells can arise from differentiated neoplastic cells. Neurosurgery 69, N22.  Jang, Y. J., Ji, J. H., Choi, Y. C., Ryu, C. J., and Ko, S. Y. (2007). Regulation of Polo-like kinase 1 by DNA damage in mitosis. Inhibition of mitotic PLK-1 by protein phosphatase 2A. J Biol Chem 282, 2473-2482.  Jang, Y. J., Lin, C. Y., Ma, S., and Erikson, R. L. (2002). Functional studies on the role of the C-terminal domain of mammalian polo-like kinase. Proc Natl Acad Sci U S A 99, 1984-1989.  Jiang, F., Qiu, Q., Khanna, A., Todd, N. W., Deepak, J., Xing, L., Wang, H., Liu, Z., Su, Y., Stass, S. A., and Katz, R. L. (2009). Aldehyde dehydrogenase 1 is a tumor stem cell-associated marker in lung cancer. Mol Cancer Res 7, 330-338.  Jimeno, A., Li, J., Messersmith, W. A., Laheru, D., Rudek, M. A., Maniar, M., Hidalgo, M., Baker, S. D., and Donehower, R. C. (2008). Phase I study of ON 01910.Na, a novel modulator of the Polo-like kinase 1 pathway, in adult patients with solid tumors. J Clin Oncol 26, 5504-5510.  Jiricny, J. (2006). The multifaceted mismatch-repair system. Nat Rev Mol Cell Biol 7, 335-346.  Johannessen, T. C., Prestegarden, L., Grudic, A., Hegi, M. E., Tysnes, B. B., and Bjerkvig, R. (2012). The DNA repair protein ALKBH2 mediates temozolomide resistance in human glioblastoma cells. Neuro Oncol.  Johansson, B. (1992). A review of the pharmacokinetics and pharmacodynamics of disulfiram and its metabolites. Acta Psychiatr Scand Suppl 369, 15-26.  Johnson, T.M., Antrobus, R., Johnson, L.N. (2008). Plk1 activation by Ste20-like kinase (Slk) phosphorylation and polo-box phosphopeptide binding assayed with the substrate translationally controlled tumor protein (TCTP). Biochemistry 47(12), 3688-96.  Jones, D. T., Jager, N., Kool, M., Zichner, T., Hutter, B., Sultan, M., Cho, Y. J., Pugh, T. J., Hovestadt, V., Stutz, A. M., et al. (2012). Dissecting the genomic complexity underlying medulloblastoma. Nature 488, 100-105.  Joo, K. M., Kim, S. Y., Jin, X., Song, S. Y., Kong, D. S., Lee, J. I., Jeon, J. W., Kim, M. H., Kang, B. G., Jung, Y., et al. (2008). Clinical and biological implications of CD133-positive and CD133-negative cells in glioblastomas. Lab Invest 88, 808-815.   169 Kaatsch, P., Rickert, C. H., Kuhl, J., Schuz, J., and Michaelis, J. (2001). Population-based epidemiologic data on brain tumors in German children. Cancer 92, 3155-3164.  Kaina, B., Christmann, M., Naumann, S., and Roos, W. P. (2007). MGMT: key node in the battle against genotoxicity, carcinogenicity and apoptosis induced by alkylating agents. DNA Repair (Amst) 6, 1079-1099.  Kaina, B., Ziouta, A., Ochs, K., and Coquerelle, T. (1997). Chromosomal instability, reproductive cell death and apoptosis induced by O6-methylguanine in Mex-, Mex+ and methylation-tolerant mismatch repair compromised cells: facts and models. Mutat Res 381, 227-241.  Kanai, R., Wakimoto, H., Martuza, R. L., and Rabkin, S. D. (2011). A novel oncolytic herpes simplex virus that synergizes with phosphoinositide 3-kinase/Akt pathway inhibitors to target glioblastoma stem cells. Clin Cancer Res 17, 3686-3696.  Kanaji, S., Saito, H., Tsujitani, S., Matsumoto, S., Tatebe, S., Kondo, A., Ozaki, M., Ito, H., and Ikeguchi, M. (2006). Expression of polo-like kinase 1 (PLK1) protein predicts the survival of patients with gastric carcinoma. Oncology 70, 126-133.  Kaneko, Y., Sakakibara, S., Imai, T., Suzuki, A., Nakamura, Y., Sawamoto, K., Ogawa, Y., Toyama, Y., Miyata, T., and Okano, H. (2000). Musashi1: an evolutionally conserved marker for CNS progenitor cells including neural stem cells. Dev Neurosci 22, 139-153.  Karajannis, M., Allen, J. C., and Newcomb, E. W. (2008). Treatment of pediatric brain tumors. J Cell Physiol 217, 584-589.  Karn, T., Holtrich, U., Wolf, G., Hock, B., Strebhardt, K., and Rubsamenwaigmann, H. (1997). Human SAK related to the PLK/polo family of cell cycle kinases shows high mRNA expression in testis. Oncol Rep 4, 505-510.  Kast, R. E., and Belda-Iniesta, C. (2009). Suppressing glioblastoma stem cell function by aldehyde dehydrogenase inhibition with chloramphenicol or disulfiram as a new treatment adjunct: an hypothesis. Curr Stem Cell Res Ther 4, 314-317.  Katsetos, C. D., Herman, M. M., Frankfurter, A., Gass, P., Collins, V. P., Walker, C. C., Rosemberg, S., Barnard, R. O., and Rubinstein, L. J. (1989). Cerebellar desmoplastic medulloblastomas. A further immunohistochemical characterization of the reticulin-free pale islands. Arch Pathol Lab Med 113, 1019-1029.  Kawauchi, D., Robinson, G., Uziel, T., Gibson, P., Rehg, J., Gao, C., Finkelstein, D., Qu, C., Pounds, S., Ellison, D. W., et al. (2012). A mouse model of the most aggressive subgroup of human medulloblastoma. Cancer Cell 21, 168-180.  Kelly, J. J., Stechishin, O., Chojnacki, A., Lun, X., Sun, B., Senger, D. L., Forsyth, P., Auer, R. N., Dunn, J. F., Cairncross, J. G., et al. (2009). Proliferation of human glioblastoma stem cells occurs independently of exogenous mitogens. Stem Cells 27, 1722-1733.   170 Keles, G.E., Berger, M.S., Srinivasan, J., Kolstoe, D.D., Bobola, M.S., Silber, J.R. (1995). Establishment and characterization of four human medulloblastoma-derived cell lines. Oncology Research 7, 493-503. Kelm, O., Wind, M., Lehmann, W. D., and Nigg, E. A. (2002). Cell cycle-regulated phosphorylation of the Xenopus polo-like kinase Plx1. J Biol Chem 277, 25247-25256.  Khuong-Quang, D.A., Buczkowicz, P., Rakopoulos, P., Liu, X.Y., Fontebasso, A.M., Bouffet, E., Bartels, U., Albrecht, S., Schwartzentruber, J., Letourneau, L., et al. (2012). K27M mutation in histone H3.3 defines clinically and biologically distinct subgroups of pediatric diffuse intrinsic pontine gliomas. Acta Neuropathol 124(3), 439-47. Kieran, M. W., Walker, D., Frappaz, D., and Prados, M. (2010). Brain tumors: from childhood through adolescence into adulthood. J Clin Oncol 28, 4783-4789.  Kilic, T., Alberta, J. A., Zdunek, P. R., Acar, M., Iannarelli, P., O'Reilly, T., Buchdunger, E., Black, P. M., and Stiles, C. D. (2000). Intracranial inhibition of platelet-derived growth factor-mediated glioblastoma cell growth by an orally active kinase inhibitor of the 2-phenylaminopyrimidine class. Cancer Res 60, 5143-5150.  Kim, M., Turnquist, H., Jackson, J., Sgagias, M., Yan, Y., Gong, M., Dean, M., Sharp, J. G., and Cowan, K. (2002). The multidrug resistance transporter ABCG2 (breast cancer resistance protein 1) effluxes Hoechst 33342 and is overexpressed in hematopoietic stem cells. Clin Cancer Res 8, 22-28.  King, S. I., Purdie, C. A., Bray, S. E., Quinlan, P. R., Jordan, L. B., Thompson, A. M., and Meek, D. W. (2012). Immunohistochemical detection of Polo-like kinase-1 (PLK1) in primary breast cancer is associated with TP53 mutation and poor clinical outcom. Breast Cancer Res 14, R40.  Kitada, K., Johnson, A. L., Johnston, L. H., and Sugino, A. (1993). A multicopy suppressor gene of the Saccharomyces cerevisiae G1 cell cycle mutant gene dbf4 encodes a protein kinase and is identified as CDC5. Mol Cell Biol 13, 4445-4457.  Kleihues, P., Louis, D. N., Scheithauer, B. W., Rorke, L. B., Reifenberger, G., Burger, P. C., and Cavenee, W. K. (2002). The WHO classification of tumors of the nervous system. J Neuropathol Exp Neurol 61, 215-225; discussion 226-219.  Klesse, L. J., and Bowers, D. C. (2010). Childhood medulloblastoma: current status of biology and treatment. CNS Drugs 24, 285-301.  Knecht, R., Elez, R., Oechler, M., Solbach, C., von Ilberg, C., and Strebhardt, K. (1999). Prognostic significance of polo-like kinase (PLK) expression in squamous cell carcinomas of the head and neck. Cancer Res 59, 2794-2797.  Kneisel, L., Strebhardt, K., Bernd, A., Wolter, M., Binder, A., and Kaufmann, R. (2002). Expression of polo-like kinase (PLK1) in thin melanomas: a novel marker of metastatic disease. J Cutan Pathol 29, 354-358.  Knobbe, C. B., Merlo, A., and Reifenberger, G. (2002). Pten signaling in gliomas. Neuro Oncol 4, 196-211.  171  Ko, M. A., Rosario, C. O., Hudson, J. W., Kulkarni, S., Pollett, A., Dennis, J. W., and Swallow, C. J. (2005). Plk4 haploinsufficiency causes mitotic infidelity and carcinogenesis. Nat Genet 37, 883-888.  Komitova, M., and Eriksson, P. S. (2004). Sox-2 is expressed by neural progenitors and astroglia in the adult rat brain. Neurosci Lett 369, 24-27.  Kool, M., Korshunov, A., Remke, M., Jones, D. T., Schlanstein, M., Northcott, P. A., Cho, Y. J., Koster, J., Schouten-van Meeteren, A., van Vuurden, D., et al. (2012). Molecular subgroups of medulloblastoma: an international meta-analysis of transcriptome, genetic aberrations, and clinical data of WNT, SHH, Group 3, and Group 4 medulloblastomas. Acta Neuropathol 123, 473-484.  Kool, M., Koster, J., Bunt, J., Hasselt, N. E., Lakeman, A., van Sluis, P., Troost, D., Meeteren, N. S., Caron, H. N., Cloos, J., et al. (2008). Integrated genomics identifies five medulloblastoma subtypes with distinct genetic profiles, pathway signatures and clinicopathological features. PLoS One 3, e3088.  Kops, G. J., Weaver, B. A., and Cleveland, D. W. (2005). On the road to cancer: aneuploidy and the mitotic checkpoint. Nat Rev Cancer 5, 773-785.  Kotani, S., Tugendreich, S., Fujii, M., Jorgensen, P. M., Watanabe, N., Hoog, C., Hieter, P., and Todokoro, K. (1998). PKA and MPF-activated polo-like kinase regulate anaphase-promoting complex activity and mitosis progression. Mol Cell 1, 371-380.  Kraft, C., Herzog, F., Gieffers, C., Mechtler, K., Hagting, A., Pines, J., and Peters, J. M. (2003). Mitotic regulation of the human anaphase-promoting complex by phosphorylation. EMBO J 22, 6598-6609.  Kristiansen, K., Hagen, S., Kollevold, T., Torvik, A., Holme, I., Nesbakken, R., Hatlevoll, R., Lindgren, M., Brun, A., Lindgren, S., et al. (1981). Combined modality therapy of operated astrocytomas grade III and IV. Confirmation of the value of postoperative irradiation and lack of potentiation of bleomycin on survival time: a prospective multicenter trial of the Scandinavian Glioblastoma Study Group. Cancer 47, 649-652.  Kumagai, A., and Dunphy, W. G. (1996). Purification and molecular cloning of Plx1, a Cdc25-regulatory kinase from Xenopus egg extracts. Science 273, 1377-1380.  Kyritsis, A. P., Bondy, M. L., Rao, J. S., and Sioka, C. (2010). Inherited predisposition to glioma. Neuro Oncol 12, 104-113.  Lake, R. J., and Jelinek, W. R. (1993). Cell cycle- and terminal differentiation-associated regulation of the mouse mRNA encoding a conserved mitotic protein kinase. Mol Cell Biol 13, 7793-7801.  Lane, H. A., and Nigg, E. A. (1996). Antibody microinjection reveals an essential role for human polo-like kinase 1 (Plk1) in the functional maturation of mitotic centrosomes. J Cell Biol 135, 1701-1713.   172 Lang, F. F., Miller, D. C., Koslow, M., and Newcomb, E. W. (1994). Pathways leading to glioblastoma multiforme: a molecular analysis of genetic alterations in 65 astrocytic tumors. J Neurosurg 81, 427-436.  Laperriere, N. J., Leung, P. M., McKenzie, S., Milosevic, M., Wong, S., Glen, J., Pintilie, M., and Bernstein, M. (1998). Randomized study of brachytherapy in the initial management of patients with malignant astrocytoma. Int J Radiat Oncol Biol Phys 41, 1005-1011.  Lapidot, T., Sirard, C., Vormoor, J., Murdoch, B., Hoang, T., Caceres-Cortes, J., Minden, M., Paterson, B., Caligiuri, M. A., and Dick, J. E. (1994). A cell initiating human acute myeloid leukaemia after transplantation into SCID mice. Nature 367, 645-648.  Lashford, L. S., Thiesse, P., Jouvet, A., Jaspan, T., Couanet, D., Griffiths, P. D., Doz, F., Ironside, J., Robson, K., Hobson, R., et al. (2002). Temozolomide in malignant gliomas of childhood: a United Kingdom Children's Cancer Study Group and French Society for Pediatric Oncology Intergroup Study. J Clin Oncol 20, 4684-4691.  Laurent, J. P., Chang, C. H., and Cohen, M. E. (1985). A classification system for primitive neuroectodermal tumors (medulloblastoma) of the posterior fossa. Cancer 56, 1807-1809.  Lauth, M., Bergstrom, A., Shimokawa, T., and Toftgard, R. (2007). Inhibition of GLI-mediated transcription and tumor cell growth by small-molecule antagonists. Proc Natl Acad Sci U S A 104, 8455-8460.  Laywell, E. D., Kukekov, V. G., Suslov, O., Zheng, T., and Steindler, D. A. (2002). Production and analysis of neurospheres from acutely dissociated and postmortem CNS specimens. Methods Mol Biol 198, 15-27.  Lederman, G., Wronski, M., Arbit, E., Odaimi, M., Wertheim, S., Lombardi, E., Wrzolek, M. (2000). Treatment of recurrnet glioblastoma multiforme using fractionated stereotactic radiosurgery and concurrent paclitaxel. Am J Clin Oncol 23(2):155-9. Lee, C., Dhillon, J., Wang, M. Y., Gao, Y., Hu, K., Park, E., Astanehe, A., Hung, M. C., Eirew, P., Eaves, C. J., and Dunn, S. E. (2008a). Targeting YB-1 in HER-2 overexpressing breast cancer cells induces apoptosis via the mTOR/STAT3 pathway and suppresses tumor growth in mice. Cancer Res 68, 8661-8666.  Lee, C., Fotovati, A., Triscott, J., Chen, J., Venugopal, C., Singhal, A., Dunham, C., Kerr, J. M., Verreault, M., Yip, S., et al. (2012). Polo-like kinase 1 inhibition kills glioblastoma multiforme brain tumor cells in part through loss of SOX2 and delays tumor progression in mice. Stem Cells 30, 1064-1075.  Lee, K. S., Grenfell, T. Z., Yarm, F. R., and Erikson, R. L. (1998). Mutation of the polo-box disrupts localization and mitotic functions of the mammalian polo kinase Plk. Proc Natl Acad Sci U S A 95, 9301-9306.  Lee, K. S., Park, J. E., Kang, Y. H., Zimmerman, W., Soung, N. K., Seong, Y. S., Kwak, S. J., and Erikson, R. L. (2008b). Mechanisms of mammalian polo-like kinase 1 (Plk1) localization: self- versus non-self-priming. Cell Cycle 7, 141-145.   173 Lee, K. S., Yuan, Y. L., Kuriyama, R., and Erikson, R. L. (1995). Plk is an M-phase-specific protein kinase and interacts with a kinesin-like protein, CHO1/MKLP-1. Mol Cell Biol 15, 7143-7151.  Lee, S. E., Pellicioli, A., Malkova, A., Foiani, M., and Haber, J. E. (2001). The Saccharomyces recombination protein Tid1p is required for adaptation from G2/M arrest induced by a double-strand break. Curr Biol 11, 1053-1057.  Lee, Y., Scheck, A. C., Cloughesy, T. F., Lai, A., Dong, J., Farooqi, H. K., Liau, L. M., Horvath, S., Mischel, P. S., and Nelson, S. F. (2008c). Gene expression analysis of glioblastomas identifies the major molecular basis for the prognostic benefit of younger age. BMC Med Genomics 1, 52.  Lefkowitz, I. B., Packer, R. J., Ryan, S. G., Shah, N., Alavi, J., Rorke, L. B., Sutton, L. N., and Schut, L. (1988). Late recurrence of primitive neuroectodermal tumor/medulloblastoma. Cancer 62, 826-830.  Leis, O., Eguiara, A., Lopez-Arribillaga, E., Alberdi, M. J., Hernandez-Garcia, S., Elorriaga, K., Pandiella, A., Rezola, R., and Martin, A. G. (2012). Sox2 expression in breast tumours and activation in breast cancer stem cells. Oncogene 31, 1354-1365.  Lenart, P., Petronczki, M., Steegmaier, M., Di Fiore, B., Lipp, J. J., Hoffmann, M., Rettig, W. J., Kraut, N., and Peters, J. M. (2007). The small-molecule inhibitor BI 2536 reveals novel insights into mitotic roles of polo-like kinase 1. Curr Biol 17, 304-315.  Lenkiewicz, M., Li, N., and Singh, S. K. (2009). Culture and isolation of brain tumor initiating cells. Curr Protoc Stem Cell Biol Chapter 3, Unit3 3.  Lens, S. M., Voest, E. E., and Medema, R. H. (2010). Shared and separate functions of polo-like kinases and aurora kinases in cancer. Nat Rev Cancer 10, 825-841.  Leung, G. C., Hudson, J. W., Kozarova, A., Davidson, A., Dennis, J. W., and Sicheri, F. (2002). The Sak polo-box comprises a structural domain sufficient for mitotic subcellular localization. Nat Struct Biol 9, 719-724.  Li, B., Ouyang, B., Pan, H., Reissmann, P. T., Slamon, D. J., Arceci, R., Lu, L., and Dai, W. (1996). Prk, a cytokine-inducible human protein serine/threonine kinase whose expression appears to be down-regulated in lung carcinomas. J Biol Chem 271, 19402-19408.  Li, C., Heidt, D. G., Dalerba, P., Burant, C. F., Zhang, L., Adsay, V., Wicha, M., Clarke, M. F., and Simeone, D. M. (2007). Identification of pancreatic cancer stem cells. Cancer Res 67, 1030-1037.  Liang, D., and Shi, Y. (2011). Aldehyde dehydrogenase-1 is a specific marker for stem cells in human lung adenocarcinoma. Med Oncol 29(2):633-9.   Liby, K., Wu, H., Ouyang, B., Wu, S., Chen, J., and Dai, W. (2001). Identification of the human homologue of the early-growth response gene Snk, encoding a serum-inducible kinase. DNA Seq 11, 527-533.  174  Lin, F., Lin, P., Zhao, D., Chen, Y., Xiao, L., Qin, W., Li, D., Chen, H., Zhao, B., Zou, H., et al. (2012). Sox2 targets cyclinE, p27 and survivin to regulate androgen-independent human prostate cancer cell proliferation and apoptosis. Cell Prolif 45, 207-216.  Lin, J., Haffner, M. C., Zhang, Y., Lee, B. H., Brennen, W. N., Britton, J., Kachhap, S. K., Shim, J. S., Liu, J. O., Nelson, W. G., et al. (2011a). Disulfiram is a DNA demethylating agent and inhibits prostate cancer cell growth. Prostate 71, 333-343.  Lindon, C., and Pines, J. (2004). Ordered proteolysis in anaphase inactivates Plk1 to contribute to proper mitotic exit in human cells. J Cell Biol 164, 233-241.  Lipsky, J. J., Shen, M. L., and Naylor, S. (2001). In vivo inhibition of aldehyde dehydrogenase by disulfiram. Chem Biol Interact 130-132, 93-102.  Litvak, V., Argov, R., Dahan, N., Ramachandran, S., Amarilio, R., Shainskaya, A., and Lev, S. (2004). Mitotic phosphorylation of the peripheral Golgi protein Nir2 by Cdk1 provides a docking mechanism for Plk1 and affects cytokinesis completion. Mol Cell 14, 319-330.  Liu, G., Yuan, X., Zeng, Z., Tunici, P., Ng, H., Abdulkadir, I. R., Lu, L., Irvin, D., Black, K. L., and Yu, J. S. (2006a). Analysis of gene expression and chemoresistance of CD133+ cancer stem cells in glioblastoma. Mol Cancer 5, 67.  Liu, P., Brown, S., Channathodiyil, P., Kannappan, V., Armesilla, A. L., Darling, J. L., and Wang, W. (2013a). Reply: Cytotoxic effect of disulfiram/copper on human glioblastoma cell lines and ALDH-positive cancer-stem-like cells. British J of Cancer 108(4):994.   Liu, P., Brown, S., Goktug, T., Channathodiyil, P., Kannappan, V., Hugnot, J. P., Guichet, P. O., Bian, X., Armesilla, A. L., Darling, J. L., and Wang, W. (2012). Cytotoxic effect of disulfiram/copper on human glioblastoma cell lines and ALDH-positive cancer-stem-like cells. Br J Cancer 107, 1488-1497.  Liu, X., and Erikson, R. L. (2002). Activation of Cdc2/cyclin B and inhibition of centrosome amplification in cells depleted of Plk1 by siRNA. Proc Natl Acad Sci U S A 99, 8672-8676.  Liu, X., and Erikson, R. L. (2003). Polo-like kinase (Plk)1 depletion induces apoptosis in cancer cells. Proc Natl Acad Sci U S A 100, 5789-5794.  Liu, X., Lei, M., and Erikson, R. L. (2006b). Normal cells, but not cancer cells, survive severe Plk1 depletion. Mol Cell Biol 26, 2093-2108.  Liu, X., Zhou, T., Kuriyama, R., and Erikson, R. L. (2004). Molecular interactions of Polo-like-kinase 1 with the mitotic kinesin-like protein CHO1/MKLP-1. J Cell Sci 117, 3233-3246.  Liu, X. S., Li, H., Song, B., and Liu, X. (2010). Polo-like kinase 1 phosphorylation of G2 and S-phase-expressed 1 protein is essential for p53 inactivation during G2 checkpoint recovery. EMBO Rep 11, 626-632.   175 Liu, Y. P., Yang, C. J., Huang, M. S., Yeh, C. T., Wu, A. T., Lee, Y. C., Lai, T. C., Lee, C. H., Hsiao, Y. W., Lu, J., et al. (2013b). Cisplatin selects for multidrug-resistant CD133+ cells in lung adenocarcinoma by activating notch signaling. Cancer Res 73, 406-416.  Llamazares, S., Moreira, A., Tavares, A., Girdham, C., Spruce, B. A., Gonzalez, C., Karess, R. E., Glover, D. M., and Sunkel, C. E. (1991). polo encodes a protein kinase homolog required for mitosis in Drosophila. Genes Dev 5, 2153-2165.  Logarinho, E., and Sunkel, C. E. (1998). The Drosophila POLO kinase localises to multiple compartments of the mitotic apparatus and is required for the phosphorylation of MPM2 reactive epitopes. J Cell Sci 111 ( Pt 19), 2897-2909.  Lottaz, C., Beier, D., Meyer, K., Kumar, P., Hermann, A., Schwarz, J., Junker, M., Oefner, P. J., Bogdahn, U., Wischhusen, J., et al. (2010). Transcriptional profiles of CD133+ and CD133- glioblastoma-derived cancer stem cell lines suggest different cells of origin. Cancer Res 70, 2030-2040.  Louis, D. N. (2006). Molecular pathology of malignant gliomas. Annu Rev Pathol 1, 97-117.  Louis, D. N., Holland, E. C., and Cairncross, J. G. (2001). Glioma classification: a molecular reappraisal. Am J Pathol 159, 779-786.  Louis, D. N., Ohgaki, H., Wiestler, O. D., Cavenee, W. K., Burger, P. C., Jouvet, A., Scheithauer, B. W., and Kleihues, P. (2007). The 2007 WHO classification of tumours of the central nervous system. Acta Neuropathol 114, 97-109.  Louis, D. N., Rubio, M. P., Correa, K. M., Gusella, J. F., and von Deimling, A. (1993). Molecular genetics of pediatric brain stem gliomas. Application of PCR techniques to small and archival brain tumor specimens. J Neuropathol and Experimental Neurol 52, 507-515.  Lowery, D. M., Lim, D., and Yaffe, M. B. (2005). Structure and function of Polo-like kinases. Oncogene 24, 248-259.  Lufei, C., Koh, T. H., Uchida, T., and Cao, X. (2007). Pin1 is required for the Ser727 phosphorylation-dependent Stat3 activity. Oncogene 26, 7656-7664.  Luo, J., Emanuele, M. J., Li, D., Creighton, C. J., Schlabach, M. R., Westbrook, T. F., Wong, K. K., and Elledge, S. J. (2009). A genome-wide RNAi screen identifies multiple synthetic lethal interactions with the Ras oncogene. Cell 137, 835-848.  Lusa, L., McShane, L. M., Reid, J. F., De Cecco, L., Ambrogi, F., Biganzoli, E., Gariboldi, M., and Pierotti, M. A. (2007). Challenges in projecting clustering results across gene expression-profiling datasets. J Natl Cancer Inst 99, 1715-1723.  Lyons, A. B., and Parish, C. R. (1994). Determination of lymphocyte division by flow cytometry. J Immunol Methods 171, 131-137.   176 Ma, S., Chan, K. W., Lee, T. K., Tang, K. H., Wo, J. Y., Zheng, B. J., and Guan, X. Y. (2008). Aldehyde dehydrogenase discriminates the CD133 liver cancer stem cell populations. Mol Cancer Res 6, 1146-1153.  Ma, S., Liu, M. A., Yuan, Y. L., and Erikson, R. L. (2003). The serum-inducible protein kinase Snk is a G1 phase polo-like kinase that is inhibited by the calcium- and integrin-binding protein CIB. Mol Cancer Res 1, 376-384.  Mabbott, D. J., Spiegler, B. J., Greenberg, M. L., Rutka, J. T., Hyder, D. J., and Bouffet, E. (2005). Serial evaluation of academic and behavioral outcome after treatment with cranial radiation in childhood. J Clin Oncol 23, 2256-2263.  MacDonald, T. J., Brown, K. M., LaFleur, B., Peterson, K., Lawlor, C., Chen, Y., Packer, R. J., Cogen, P., and Stephan, D. A. (2001). Expression profiling of medulloblastoma: PDGFRA and the RAS/MAPK pathway as therapeutic targets for metastatic disease. Nat Genet 29, 143-152.  Macmillan, J. C., Hudson, J. W., Bull, S., Dennis, J. W., and Swallow, C. J. (2001). Comparative expression of the mitotic regulators SAK and PLK in colorectal cancer. Ann Surg Oncol 8, 729-740.  Macurek, L., Lindqvist, A., Lim, D., Lampson, M. A., Klompmaker, R., Freire, R., Clouin, C., Taylor, S. S., Yaffe, M. B., and Medema, R. H. (2008). Polo-like kinase-1 is activated by aurora A to promote checkpoint recovery. Nature 455, 119-123.  Magni, M., Shammah, S., Schiro, R., Mellado, W., Dalla-Favera, R., and Gianni, A. M. (1996). Induction of cyclophosphamide-resistance by aldehyde-dehydrogenase gene transfer. Blood 87, 1097-1103.  Maher, E. A., Furnari, F. B., Bachoo, R. M., Rowitch, D. H., Louis, D. N., Cavenee, W. K., and DePinho, R. A. (2001). Malignant glioma: genetics and biology of a grave matter. Genes Dev 15, 1311-1333.  Mailand, N., Bekker-Jensen, S., Bartek, J., and Lukas, J. (2006). Destruction of Claspin by SCFbetaTrCP restrains Chk1 activation and facilitates recovery from genotoxic stress. Mol Cell 23, 307-318.  Maire, V., Nemati, F., Richardson, M., Vincent-Salomon, A., Tesson, B., Rigaill, G., Gravier, E., Marty-Prouvost, B., De Koning, L., Lang, G., et al. (2013). Polo-like Kinase 1: A Potential Therapeutic Option in Combination with Conventional Chemotherapy for the Management of Patients with Triple-Negative Breast Cancer. Cancer Res 73, 813-823.  Maj, J., Grabowska, M., and Kweik, J. (1970). The effect of disulfiram, diethyldithiocarbamate and dimethyldithiocarbamate on serotonin and 5-hydroxyindole-3-acetic acid brain levels in rats. Biochem Pharmacol 19, 2517-2519.  Malkin, D., Li, F. P., Strong, L. C., Fraumeni, J. F., Jr., Nelson, C. E., Kim, D. H., Kassel, J., Gryka, M. A., Bischoff, F. Z., Tainsky, M. A., and et al. (1990). Germ line p53 mutations in a familial syndrome of breast cancer, sarcomas, and other neoplasms. Science 250, 1233-1238.  177  Malmstrom, A., Gronberg, B. H., Marosi, C., Stupp, R., Frappaz, D., Schultz, H., Abacioglu, U., Tavelin, B., Lhermitte, B., Hegi, M. E., et al. (2012). Temozolomide versus standard 6-week radiotherapy versus hypofractionated radiotherapy in patients older than 60 years with glioblastoma: the Nordic randomised, phase 3 trial. Lancet Oncol 13, 916-926.  Mamely, I., van Vugt, M. A., Smits, V. A., Semple, J. I., Lemmens, B., Perrakis, A., Medema, R. H., and Freire, R. (2006). Polo-like kinase-1 controls proteasome-dependent degradation of Claspin during checkpoint recovery. Curr Biol 16, 1950-1955.  Marcato, P., Dean, C. A., Pan, D., Araslanova, R., Gillis, M., Joshi, M., Helyer, L., Pan, L., Leidal, A., Gujar, S., et al. (2010). Aldehyde Dehydrogenase Activity of Breast Cancer Stem Cells is Primarily Due to Isoform ALDH1A3 and Its Expression is Predictive of Metastasis. Stem Cells 29(1):32-45.  Marchesi, F., Turriziani, M., Tortorelli, G., Avvisati, G., Torino, F., and De Vecchis, L. (2007). Triazene compounds: mechanism of action and related DNA repair systems. Pharmacol Res 56, 275-287.  Margottin-Goguet, F., Hsu, J. Y., Loktev, A., Hsieh, H. M., Reimann, J. D., and Jackson, P. K. (2003). Prophase destruction of Emi1 by the SCF(betaTrCP/Slimb) ubiquitin ligase activates the anaphase promoting complex to allow progression beyond prometaphase. Dev Cell 4, 813-826.  Marino, S. (2005). Medulloblastoma: developmental mechanisms out of control. Trends Mol Med 11, 17-22.  Marshall, G. P., 2nd, Reynolds, B. A., and Laywell, E. D. (2007). Using the neurosphere assay to quantify neural stem cells in vivo. Curr Pharm Biotechnol 8, 141-145.  Mason, W. P., Belanger, K., Nicholas, G., Vallieres, I., Mathieu, D., Kavan, P., Desjardins, A., Omuro, A., and Reymond, D. (2012). A phase II study of the Ras-MAPK signaling pathway inhibitor TLN-4601 in patients with glioblastoma at first progression. J Neurooncol 107, 343-349.  Mason, W. P., Grovas, A., Halpern, S., Dunkel, I. J., Garvin, J., Heller, G., Rosenblum, M., Gardner, S., Lyden, D., Sands, S., et al. (1998). Intensive chemotherapy and bone marrow rescue for young children with newly diagnosed malignant brain tumors. J Clin Oncol 16, 210-221.  Massimino, M., Spreafico, F., Biassoni, V., Simonetti, F., Riva, D., Trecate, G., Giombini, S., Poggi, G., Pecori, E., Pignoli, E., et al. (2008). Diffuse pontine gliomas in children: changing strategies, changing results? A mono-institutional 20-year experience. J Neurooncol 87, 355-361.  Matthew, E. M., Yen, T. J., Dicker, D. T., Dorsey, J. F., Yang, W., Navaraj, A., and El-Deiry, W. S. (2007). Replication stress, defective S-phase checkpoint and increased death in Plk2-deficient human cancer cells. Cell Cycle 6, 2571-2578.   178 McCall, T. D., Pedone, C. A., and Fults, D. W. (2007). Apoptosis suppression by somatic cell transfer of Bcl-2 promotes Sonic hedgehog-dependent medulloblastoma formation in mice. Cancer Res 67, 5179-5185.  McCarthy, N. C., Simpson, J. R., Coghill, G., and Kerr, M. A. (1985). Expression in normal adult, fetal, and neoplastic tissues of a carbohydrate differentiation antigen recognised by antigranulocyte mouse monoclonal antibodies. J Clin Pathol 38, 521-529.  McInnes, C., and Wyatt, M. D. (2011). PLK1 as an oncology target: current status and future potential. Drug Discov Today 16, 619-625.  McKenzie, L., King, S., Marcar, L., Nicol, S., Dias, S. S., Schumm, K., Robertson, P., Bourdon, J. C., Perkins, N., Fuller-Pace, F., and Meek, D. W. (2010). p53-dependent repression of polo-like kinase-1 (PLK1). Cell Cycle 9, 4200-4212.  McKinney, P. A. (2004). Brain tumours: incidence, survival, and aetiology. J Neurol Neurosurg Psychiatry 75 Suppl 2, ii12-17.  McManamy, C. S., Pears, J., Weston, C. L., Hanzely, Z., Ironside, J. W., Taylor, R. E., Grundy, R. G., Clifford, S. C., and Ellison, D. W. (2007). Nodule formation and desmoplasia in medulloblastomas-defining the nodular/desmoplastic variant and its biological behavior. Brain Pathol 17, 151-164.  McNeil, D. E., Cote, T. R., Clegg, L., and Rorke, L. B. (2002). Incidence and trends in pediatric malignancies medulloblastoma/primitive neuroectodermal tumor: a SEER update. Surveillance Epidemiology and End Results. Med Pediatr Oncol 39, 190-194.  Meyers, M., Hwang, A., Wagner, M. W., and Booth