Comparative Analysis of Chromosome 8 Copy Number in Paediatric Solid Tumours

Comparative analysis of chromosome 8 copy number in paediatric solid tumours

Nancy Martin La Rotta

A thesis submitted in fulfilment of the requirements for the degree of Doctor of Philosophy

Molecular Oncology Group, Children’s Cancer Research Unit, Kids Research Institute, The Children’s Hospital at Westmead Discipline of Paediatrics and Child Health, Sydney Medical School

2017

Table of contents

Statement of originality viii

Acknowledgments ix

Abstract xi

Publications arising from this thesis, by the author xv

List of abbreviations xvi

List of Tables xix

List of Figures xxiii

Chapter 1 Literature review 1

1. Introduction 2

1.1 General introduction to childhood cancer 2

1.1.1 Classification of tumours in children 2

1.1.2 Solid tumours of childhood 6

1.1.3 Neuroblastoma 6

1.1.4 Rhabdomyosarcoma 7

1.1.5 Ewing Sarcoma 9

1.1.6 Osteosarcoma 10

1.2 Cytogenetic changes in childhood cancers - relevance to prognosis 11

1.2.1 Introduction to cytogenetic changes in cancer 11

1.2.2. Cytogenetic changes in neuroblastoma 12

1.2.3. Cytogenetic changes in rhabdomyosarcomas 15

1.2.4. Cytogenetic changes in Ewing sarcoma 19

1.2.5. Cytogenetic changes in osteosarcoma 22

1.2.6 Cytogenetic changes relevant to cancer therapy 25

1.3 Technical approaches to detect copy number changes in cancer 26

1.3.1 Classical cytogenetics 26

1.3.2 Comparative Genomic Hybridization (CGH) and array CGH 30

1.3.2.1 Comparative Genomic Hybridization (CGH) 30

1.3.2.2 Array Comparative Genomic Hybridization (aCGH) 31

1.3.3 Multiplex Ligation-dependent Probe Amplification (MLPA) and Quantitative

Polymerase Chain Reaction (qPCR) 34

1.3.3.1 Multiplex ligation-dependent probe amplification (MLPA) 35

1.3.3.2 MLPA and Cancer 37

1.3.3.3 Quantitative Polymerase Chain Reaction (qPCR) 40

1.4 Chromosome 8q gain in adult cancers 43

1.4.1 Breast cancer 43

1.4.2 Prostate cancer 44

1.4.3 Colorectal cancer 45

1.4.4 Human cancer cell lines 45

1.5 Chromosome 8 gain in childhood cancers 46

1.5.1 Neuroblastoma 46

1.5.2 Rhabdomyosarcoma 49

1.5.3 Ewing Sarcoma 51

1.5.4 Osteosarcoma 53

1.6 Statement of the problem, hypothesis and aims of this research 55

1.6.1 Hypotheses 55

Chapter 2 Materials and Methods 57

2.1 Materials 58

2.2 Breast cell lines 59

2.3 Paediatric solid tumour cohort 59

2.4 Genomic DNA extraction from frozen tissues and cultured cells 62

2.5 Spectrophotometric and electrophoretic analysis of genomic DNA 63

2.6 Multiplex Ligation-dependent Probe Amplification (MLPA) 63

2.6.1 MLPA reactions (day 1) 64

2.6.2 MLPA reactions (day 2) 65

2.6.3 MLPA Fragment separation and peak pattern evaluation 66

2.6.4. Data analysis 67

iii

2.7 Array Comparative Genomic Hybridization (aCGH) 69

2.7.1 Array description 71

2.7.2 Sample preparation 71

2.7.3 Restriction digestion of test and reference DNA samples 71

2.7.4 Labeling of test and reference DNA samples 72

2.7.5 Clean-up of labelled DNA samples 73

2.7.6 Hybridization 73

2.7.7 Post-hybridization washing and scanning 74

2.7.8 Data analysis 74

2.8 Quantitative real-time Polymerase Chain Reaction (qPCR) 76

2.9 Statistical Analyses 77

Chapter 3 Analysis of chromosome 8 copy number in paediatric solid tumours using MLPA 78

3.1 Introduction 79

3.2 Aims for chapter 3 83

3.3 Results 84

3.3.1 Use of known copy number changes in 6 breast cell lines, as positive

controls for this paediatric tumour study. 87

3.3.2 Copy number changes on chromosome 8p in the paediatric solid tumour

cohort 95

3.3.3 Copy number changes on chromosome 8q and at reference genes in the

paediatric tumour cohort. 98

3.3.4 Additional MLPA results 99

3.4 Discussion 114

3.4.1 Previous research examining chromosome 8 copy number using MLPA 115

3.4.2 Six Breast cell lines: included in this study as known positive controls 116

3.4.3 Chromosome 8p copy number in paediatric solid tumours 116

3.4.4 Chromosome 8q copy number in paediatric solid tumours 119

3.4.5 Technical considerations 121

Chapter 4 Analysis of copy number in paediatric solid tumours using array

Comparative Genomic Hybridization 123

4.1 Introduction 124

4.2 Aims for aCGH 130

4.3 Results 131

4.3.1 Chromosome 8 copy number changes in breast cell lines 132

4.3.2 Compound plots indicating common copy number changes in paediatric

solid tumours 137

4.3.3 Copy number changes in genes examined by MLPA on chromosome 8 in

31 solid tumours 145

4.3.4 Statistical Analysis for loci examined by MLPA 145

4.4 Discussion 155

4.4.1 Chromosome 8 copy number changes in breast cell lines 155

4.4.2 Common regions of copy number change in paediatric solid tumours 156

4.4.3 Copy number changes in chromosome 8 genes examined by MLPA in the

solid tumour cohort 158

4.4.4 Limitations 160

4.4.5 Summary and conclusions 160

Chapter 5 Comparison of 3 techniques for the assessment of DNA copy number as applied to solid paediatric tumour samples and breast cell lines 162

5.1 Introduction 163

5.2 Aims 165

5.3 Results 166

5.3.1 Comparison of copy number results obtained using MLPA or aCGH across

all measured chromosome 8 loci in 6 breast cell lines 167

5.3.2 Comparison of MLPA and aCGH copy number results obtained for the

paediatric solid tumour cohort, according to tumour type 172

5.3.3 Selection of genes for qPCR 183

5.3.4 Overview of qPCR 186

5.3.5 qPCR results 190

5.4 Discussion 192

5.5 Summary and conclusions 195

Chapter 6 General Discussion 196

6.1 Introduction 197

6.2 MLPA as a stand-alone technique for measuring copy number changes 197

6.3 aCGH results of other genes genome-wide 200

6.4 Summary of aCGH chromosome 8 results obtained for 6 breast cell lines and the solid tumour cohort for genes examined by MLPA 202

Final conclusions 206

References 207

Appendices 238

Appendix 1. Statistical analysis 239

Appendix 2. Copy number changes in cell lines 240

Appendix 3. Copy number changes in solid tumours 243

Appendix 4. Copy number status using aCGH 259

Appendix 5. Copy number categories in cell lines 263

Appendix 6. Copy number categories in solid tumours 264

vii

Statement of originality

The work described in this thesis was performed by Nancy Martin La Rotta, except where due acknowledgement has been made.

I declare that no part of this work has been submitted previously for the purpose of obtaining any other degree at the University of Sydney or at any other institution.

Nancy Martin La Rotta

December 2016

viii

Acknowledgments

I would like to thank God who gave me the opportunity to accomplish one of my deepest yearnings of my heart, to pursue my PhD, and who provided me with the best supervisors that I ever imagine, A. Prof Jennifer Byrne and associate supervisor

Dr. Greg Peters.

I owe my more sincere gratitude and respect to my primary supervisor Jennifer

Byrne, a great scientist and wise person in any possible way. She with their dedication and guidance always helped me to resolve my fears and uncertainties, to see problems from different angles and thus took me to solve them in absolutely impressive ways.

I greatly appreciate the support of the staff from Kids Research Institute (KRI), present and past members, particularly all members of Cancer Research Unit. Dr.

Yuyan Chen who always gave me her wise advise, and also to admin members,

Vanita De’Souza and Janette Clarkson, thank you very much.

I would like to express my immense gratitude to my associate supervisor Dr. Greg

Peters for the support, his critical feedback and words of wisdom. Likewise, I would like to thank the availability and generosity from the scientists in the Cytogenetics and FISH Laboratories at The Children's Hospital at Westmead, especially Ms.

Dorothy Hung, Mr. Artur Darmanian and Mr. Luke St. Heaps.

Last but not less I would like to thank my lovely and supportive husband that was at my side from beginning to end, without his support I would not have been able to achieve what I accomplished. Finally, I want to thank my Mum who travelled from overseas few times to provide me with endless support and encouragement.

Finally, I want to dedicate this work to my three kids for whom I wanted to set an example of persistence and achievement of goals in life, in the hope that they will go far beyond.

This work was supported by the NHMRC scholarship and The Kids Cancer project

Abstract

Chromosomal imbalances are of major significance in cancer initiation and progression. Gains of chromosome 8 in human cancer have been described by a number of authors and are amongst the most common cytogenetic events that occur. However in childhood solid tumours, few studies have examined copy number changes of specific genes on chromosome 8.

Somatic genomic events have been for many years an important component of diagnosis, prognosis and therapy selection in paediatric oncology. Different techniques have been designed to identify gene dosage or copy number changes in a quantitative fashion. In this project we used Multiplex Ligation-dependent Probe

Amplification (MLPA), array Comparative Genomic Hybridization (aCGH) and quantitative PCR (qPCR).

MLPA was used as a screening tool to assess chromosome 8 copy number changes in a cohort of 31 paediatric solid tumours - osteosarcoma n=5, Ewing sarcoma n=8, rhabdomyosarcoma n=8, and neuroblastoma n=10, and 6 breast cell lines (5 breast cancer and a non-tumorigenic cell line MCF-10A), some of which were expected to show defined copy number changes on chromosome 8. The SALSA MLPA KIT

P014-A1 Chromosome 8 kit has 32 probes mapping to 30 chromosome 8 genes, and 9 reference probes mapping to other chromosomes. Using MLPA, osteosarcomas and neuroblastoma showed the greatest number of copy number changes at chromosome 8 loci with copy number gain/loss at 21/30 loci and 29/30 loci, respectively. Ewing sarcoma and rhabdomyosarcomas showed fewer

chromosome 8 copy number changes at 9/30 loci and 11/30 loci, respectively. In the breast cell lines, MLPA results confirmed as expected copy number loss at 8p loci in

5/6 breast cell lines, and copy number gains at MYC in 5/6 cell lines, followed by

TPD52 in 4/6 cell lines. As an additional finding, we obtained some discrepant results within the 4 MLPA experiments conducted for both cell lines and the solid tumour cohort. MLPA results could not be consistently reproduced for EIF3E and

EXT1.

To obtain a broader view of copy number changes in this tumour cohort, we analysed genomic DNA samples using array Comparative Genomic Hybridization

(aCGH). We used a customized 60-mer-oligonucleotide high-density Agilent eArray

SurePrint G3 2x400K with 52,639 probes along chromosome 8, 39,654 probes in 95 genes known to be important in cancer, and evenly spaced whole-genome coverage

(253,502 probes) for the remainder of the human genome.

aCGH chromosome 8 results showed that osteosarcoma samples had the most frequent copy number changes, between 9 and 21 per sample, congruent with the hallmarks of osteosarcoma that include an unusually high frequency of genome rearrangements. In Ewing sarcoma, a gain of whole chromosome 8 was detected in

37% (3/8) of samples, consistent with this being one of the most prominent features of this solid tumour. Rhabdomyosarcomas and neuroblastoma showed infrequent gain or loss at chromosome 8 loci. From the list of cancer genes with lower oligonucleotide probe spacing, 5/5 osteosarcoma samples showed copy number loss at RB1, TP53 and APC, and 8/10 neuroblastoma samples showed copy number gain at MYCN, findings that are in agreement with the literature. From results generated

xii

from genome-wide probes, a chromosomal region located at chromosome 22 band q11.22 showed copy number gain in all 4 solid tumour types, and in 21/31 samples analysed. This region contained the PRAME gene or preferentially expressed antigen in melanoma, which has been proposed as a promising target for immunotherapy in some hematologic malignancies, but has not specifically been studied in osteosarcoma, rhabdomyosarcoma or Ewing sarcoma. In addition, the aCGH data presented in this thesis revealed some genes that had not previously been linked in literature with paediatric solid tumours, including KCNK9, that showed copy number gain in all osteosarcoma samples, and RECQL4 that showed copy number gain in most rhabdomyosarcomas. Further studies in these genes may prove to be beneficial for the understanding of tumourigenesis and progress in paediatric solid tumours.

Comparison of MLPA versus aCGH copy number results was made based on the results for the 30 genes included in the chromosome 8 MLPA kit. Clear discrepancies in copy number measured by MLPA and aCGH were obtained. To quantify differences and propose genes to be used for copy number validation using a third technique, MLPA and aCGH results were compared. A list of genes was generated, from the ones that presented results with the greatest degree of disagreement to genes with least disagreement. From this list, 6 genes were chosen to be validated with a third technique namely qPCR, including MYC that was selected as a control because it presented the least disagreement between MLPA and aCGH results. Due to the small amount of DNA remaining for many samples after having applied both MLPA and aCGH techniques, 7 tumour samples (one osteosarcoma, 2 Ewing sarcomas, 2 rhabdomyosarcomas, and 2 neuroblastomas)

xiii

and 2 cell lines were selected to perform qPCR analyses. Copy number results obtained using MLPA versus qPCR showed agreement percentages ranging from

11.1-67%, and 78% of agreement for MYC. In contrast, qPCR versus aCGH showed

100% agreement for all but one gene, EXT1, which showed 89% agreement (in 8/9 samples). These results demonstrate that aCGH copy number results could be more frequently validated by qPCR than could MLPA results.

In summary, copy number results produced by aCGH were more consistently in agreement with those produced by qPCR, making qPCR a more reliable technique for diagnostic purposes than MLPA. To the best of our knowledge, this is the first time that such a technical comparison has been performed for any MLPA kit targeting human chromosome 8.

xiv

Publications arising from this thesis

Nancy Martin La Rotta, Le Myo Thwe, Gregory B. Peters, Jennifer A. Byrne.

Comparative analysis of chromosome 8 copy number in paediatric solid tumours.

(manuscript in preparation).

Publications from the author during PhD candidature

Jennifer A. Byrne, Yuyan Chen, Nancy Martin La Rotta, Gregory B. Peters (2012).

Challenges in identifying novel amplification targets in human cancers: chromosome

8q21 as a case study. Genes Cancer 3: 87-101.

List of abbreviations

aCGH Array Comparative Genomic Hybridization

AGRF Australian Genome Research Facility

AIHW Australian Institute of Health and Welfare

CGH Comparative Genomic Hybridization

ALK Anaplastic Lymphoma Kinase

ARMS Alveolar Rhabdomyosarcoma

ARMSp Alveolar Rhabdomyosarcoma translocation positive

ARMSn Alveolar Rhabdomyosarcoma translocation negative

BAC Bacterial Artificial Chromosome bp base pair

BSA Bovine Serum Albumin cDNA Complementary Deoxyribonucleic acid

CRC Colorectal cancer

CTC Circulating tumour cells

Cy3 Cyanine 3

Cy5 Cyanine 5

DNA Deoxyribonucleic acid dNTPs Deoxyribonucleotides

DQ Dosage quotient

EDTA Ethylenediaminetetraacetic acid

ERMS Embryonal RMS

xvi

ES Ewing sarcoma

FBS Foetal bovine serum

FISH Fluorescence In Situ Hybridization

Hg18 Human Genome Assembly 18

Hg19 Human Genome Assembly 19

LOH Loss of heterozygosity

HNPP Hereditary Neuropathy with liability to Pressure Palsies

ICCC-3 International Classification of Childhood Cancer

ISH In situ hybridization kb Kilobase

LOH Loss of heterozygosity

Mb Megabase mCGH Metaphase Comparative Genomic Hybridization min Minute(s)

MLPA Multiplex Ligation-dependent Probe Amplification

NB Neuroblastoma

NCBI National Center for Biotechnology Information

OS Osteosarcoma

PCR Polymerase Chain Reaction

PNET Primitive Neuroectodermal tumour qPCR Quantitative Polymerase Chain Reaction

RMS Rhabdomyosarcoma

RPMI Roswell Park Memorial Institute

RT-PCR Reverse Transcription PCR

SIOP International Society of Paediatric Oncology

xvii

SIOPEN Society of Paediatric Oncology European Neuroblastoma

Network

STR Short Tandem Repeats

STAC Significance Testing for Aberrant Copy number

TBE Tris borate EDTA

TCR T- Cell Receptor

TE Tris EDTA

UCSC University of California Santa Cruz

UV Ultraviolet

xviii

List of Tables

Table 1.1 International Classification of Childhood Cancer (ICCC-3) 4

Table 1.2 Genes targeted by cytogenetic changes in neuroblastoma 14

Table 1.3 Genes targeted in embryonal rhabdomyosarcoma 17

Table 1.4 Genes targeted in alveolar rhabdomyosarcoma 18

Table 1.5 Genes targeted in Ewing sarcoma 21

Table 1.6 Genes targeted in osteosarcoma 24

Table 1.7 Advantages and limitations of classical cytogenetic

analysis 29

Table 1.8 Advantages and limitations of aCGH 33

Table 1.9 Advantages and limitations of MLPA 39

Table 1.10 Advantages and limitations of qPCR 42

Table 2.1 Reagents and suppliers 58

Table 2.2 Summary of paediatric tumour cohort 61

Table 2.3 Coordinates of genes examined using MLPA according to

the hg18 genome assembly 68

Table 2.4 Sequences of PCR primers used in this study 77

Table 3.1 Literature describing chromosome 8 showing copy

number gain and/or loss in breast cell lines used in this

study 89

Table 3.2 Chromosome 8p copy number status in 6 breast cell lines 90

Table 3.3 Chromosome 8q copy number status in 6 breast cell lines 92

Table 3.4 Copy number status at 9 reference genes 6 in breast cell

lines 94

xix

Table 3.5 Chromosome 8p copy number status in solid tumour

cohort 96

Table 3.6 Chromosome 8p copy number status by gene and solid

tumour type 97

Table 3.7 Chromosome 8q copy number changes in osteosarcoma

n=5 100

Table 3.8 Chromosome 8q copy number changes in Ewing

sarcoma n=8 101

Table 3.9 Chromosome 8q copy number changes in

rhabdomyosarcoma n=8 102

Table 3.10 Chromosome 8q copy number changes in neuroblastoma

n=10 103

Table 3.11 Chromosome 8q copy number status in tumour cohort 104

Table 3.12 Copy number status at 9 reference genes in 31 tumour

samples 107

Table 3.13 Copy number results obtained using MLPA in a cohort of

31 paediatric solid tumours and 6 breast cell lines 118

Table 4.1 Cancer genes with lower oligonucleotide probe spacing

according to Agilent eArray data 127

Table 4.2 Customized Agilent eArray SurePrint G3 2X400K array 129

Table 4.3 Chromosome 8 copy number status of genes previously

examined using MLPA in breast cell lines (n=6) 136

Table 4.4 Common copy number changes as significant peaks in

osteosarcoma n=5 139

Table 4.5 Common copy number changes as significant peaks in

Ewing sarcoma n=8 140

Table 4.6 Common copy number changes as significant peaks in

ERMS n=6 141

Table 4.7 Common copy number changes as significant peaks in

neuroblastoma n=10 142

Table 4.8 Common copy number changes as significant peaks in

ARMS n=2 143

Table 4.9 Cancer genes with lower oligonucleotide probe spacing

with recurrent gain/loss in each type of tumour 144

Table 4.10 Copy number changes at chromosome 8 MLPA genes in

osteosarcoma n=5 146

Table 4.11 Copy number changes at chromosome 8 MLPA genes in

Ewing sarcoma n=8 148

Table 4.12 Copy number changes at chromosome 8 MLPA genes in

rhabdomyosarcoma n=8 150

Table 4.13 Copy number changes at chromosome 8 MLPA genes in

neuroblastoma n=10 152

Table 4.14 Statistical analysis using Fisher’s exact test for each type of tumour 154

Table 5.1 Agreement and disagreement between MLPA and aCGH

copy number results in breast cell lines (n=6) 170

Table 5.2 Summary of agreement between MLPA and aCGH

results obtained in 6 breast at chromosome 8p and 8q

loci 171

xxi

Table 5.3 Agreement and disagreement between MLPA and aCGH

copy number results in osteosarcoma (n=5) 174

Table 5.4 Agreement and disagreement between MLPA and aCGH

copy number results in Ewing sarcoma (n=8) 176

Table 5.5 Agreement and disagreement between MLPA and aCGH

copy number results in rhabdomyosarcoma (n=8) 178

Table 5.6 Agreement and disagreement between MLPA and aCGH

copy number results in neuroblastoma (n=10) 180

Table 5.7 Summary of the agreement between MLPA and aCGH

results obtained in the overall solid tumour cohort (n=31) 182

Table 5.8 Summary of agreement/disagreement between copy

number results obtained using MLPA or aCGH for

chromosome 8 loci 184

Table 5.9 Ranked scores for the degree of disagreement between

copy number results obtained using MLPA or aCGH for

chromosome 8 loci 185

Table 5.10 Copy number results obtained using qPCR and resulting

copy number categories 189

Table 5.11 Comparison of copy number categories resulting from

MLPA versus qPCR, and qPCR versus aCGH 191

xxii

List of Figures

Figure 1.1 Resolution capability of different methods to reveal

different sized genomic imbalances 28

Figure 1.2 Outline of Multiplex Ligation-dependent Probe

Amplification technique 36

Figure 1.3 Summary of patterns of total gains and losses

detected in 23 localized resectable neuroblastoma

samples 48

Figure 1.4 Frequency of specific amplifications and gains in

ARMSp versus ARMSn and ERMS 50

Figure 1.5 Chromosomal locations of copy number changes in

Ewing sarcoma samples (n=31) 52

Figure 1.6 Genetic aberrations in human osteosarcomas samples 54

Figure 2.1 Illustration of general principles of the aCGH 70

procedure

Figure 3.1 Number of instances of the terms “MLPA”, and “MLPA”

and “cancer” in the research literature 81

Figure 3.2 Location of MLPA probe set on chromosome 8 using

the UCSC human genome browser 86

Figure 3.3 MLPA probe ratios obtained for the SK-BR-3 cell line 91

Figure 3.4 MLPA probe ratios obtained for the

Rhabdomyosarcoma sample F29R 109

Figure 3.5 MLPA probe ratios obtained for the MCF-10A cell line 111

xxiii

Figure 3.6 MLPA probe ratios obtained for the Ewing sarcoma

sample F15E 113

Figure 4.1 Genomic copy number changes in breast cell lines

(n=6) 134

Figure 4.2 Chromosome 8 copy number gain/loss measured in

breast cell lines (n=6) 135

Figure 4.3 Chromosome 8 copy number alterations measured in in

osteosarcoma (n=5) 147

Figure 4.4 Chromosome 8 copy number alterations measured in

Ewing sarcoma (n=8) 149

Figure 4.5 Chromosome 8 copy number alterations measured in

rhabdomyosarcoma (n=8) 151

Figure 4.6 Chromosome 8 copy number alterations measured in

neuroblastoma (n=10) 153

Figure 5.1 Comparison of copy number predictions obtained

using MLPA and aCGH at chromosome 8p and 8q loci

in MCF-10A cells 170

Figure 5.2 Comparison of copy number predictions obtained

using MLPA and aCGH at chromosome 8p and 8q loci

in SK-BR-3 cells 171

Figure 5.3 Comparison of copy number predictions obtained

using MLPA and aCGH at chromosome 8p and 8q loci

in osteosarcoma M11O case 175

xxiv

Figure 5.4 Comparison of copy number predictions obtained

using MLPA and aCGH at chromosome 8p and 8q loci

in Ewing sarcoma M22E case 177

Figure 5.5 Comparison of copy number predictions obtained

using MLPA and aCGH at chromosome 8p and 8q loci

in ERMS F29R case 179

Figure 5.6 Comparison of copy number predictions obtained

using MLPA and aCGH at chromosome 8p and 8q loci

in neuroblastoma M39N case 181

Figure 5.7 Relative RECQL4 copy number measured using qPCR 188

xxv

Chapter 1 Literature review

1. Introduction

Chromosomal imbalances are of major significance in cancer initiation and progression. Gains of chromosome 8, as part of these chromosomal imbalances, have been described by a number of authors, and are amongst the most common cytogenetic events that occur in cancer. Nevertheless few studies have examined copy number changes of specific genes on chromosome 8 in childhood solid tumours. This thesis presents a literature review with a general introduction to childhood cancer.

A clear association between DNA copy number changes and prognosis has been found in a variety of tumours. Consequently, this review will explore specific cytogenetic changes in childhood cancers, focusing upon the four types of tumours examined in this project, namely neuroblastoma, rhabdomyosarcoma, Ewing sarcoma and osteosarcoma, and their relevance to treatment and prognosis. This will be followed by a discussion of technical approaches to detect copy number changes in cancer. Finally, the introduction will describe chromosome 8q gain in adult cancers, and chromosome 8 gains in childhood cancers.

1.1 General introduction to childhood cancer

1.1.1 Classification of tumours in children

The current standard for childhood cancer classification, termed “International

Classification of Childhood Cancer Third edition” (ICCC-3) provides 12 groups of

childhood cancers (Table 1.1). This coding system is based on morphology rather than the primary site of origin, as in adults (Steliarova-Foucher et al., 2005).

In the United States, cancer is the fourth most common cause of death among individuals aged 1 to 19 years old, with an increased incidence from 1.51 cases per million per year in 2002 to 1.66 cases per million per year in 2006 (Haddy et al.,

2010).

In Australia, paediatric cancer was the second most common cause of death in

2004-2008 after injuries, according to the Australian Institute of Health and Welfare

(2009). From 1983-2006, 13,925 childhood cancers were diagnosed in Australia: leukaemia (34%) was the most common childhood cancer, followed by cancers of the central nervous system (23%), and lymphomas (10%). Similar distributions were also reported in the United States (Linabery et al., 2008), and Europe (Baade et al.,

2010; Kaatsch, 2010).

Survival rates for childhood cancers have improved dramatically since 1960s, when the overall 5-year survival was estimated at 28%, due in part to more developed cytotoxic chemotherapies and radiotherapies. In general, the overall cure rate for children with cancer now stands at approximately 80% (Andam et al., 2008; Pui et al., 2011).

Table 1.1 International Classification of Childhood Cancer (ICCC-3) Steliarova-Foucher et al. (2005)

I II III IV

Lymphomas Central nervous system and Neuroblastoma and other peripheral Leukemias reticuloendothelial neoplasm miscellaneous nervous cell tumours

Lymphoid leukemias Hodgkin lymphoma Ependymomas and Neuroblastoma and choroid plexus tumour ganglioneuroblastoma

Acute myeloid leukemias Non-Hodgkin lymphoma Astrocytomas Other peripheral diseases nervous cell tumours

Chronic myeloproliferative Burkitt lymphoma Intracranial and intraspinal

embryonal tumours

Myelodysplastic syndrome Miscellaneous Other gliomas and other myeloproliferative disease lymphoreticular neoplasm

Unspecified and other Unspecified lymphomas Other specified intracranial

specified leukemias and intraspinal neoplasm

Unspecified intracranial

and intraspinal neoplasm

V VI VII VIII Retinoblastoma Renal Hepatic tumours Malignant bone tumours tumours

Nephroblastoma Hepatoblastoma Osteosarcomas

and other non epithelial Hepatic carcinoma

renal tumours

Renal carcinomas Unspecified malignant Chondrosarcomas

hepatic tumours

Unspecified malignant Ewing tumour and related

renal tumours sarcomas of bone

Other specified malignant

bone tumours

Unspecified malignant

bone tumours

Table 1.1 International Classification of Childhood Cancer (ICCC-3) Steliarova-Foucher et al. (2005) (cont.)

IX X XI XII Other malignant epithelial Other and unspecified Soft tissue and Germ cell tumours neoplasm malignant other extraosseous trophoblastic tumours and malignant melanomas neoplasm sarcomas and neoplasm of gonads

Other specified malignant Rhabdomyosarcomas Intracranial and intraspinal Adrenocortical carcinomas tumours germ cell tumours

Fibrosarcomas, peripheral Malignant extracranial and Thyroid carcinomas Other unspecified malignant extragonadal germ cell nerve sheath tumours tumours tumours

and other fibrous neoplasm Kaposi sarcoma Malignant gonadal germ Nasopharyngeal carcinomas cell tumours

Other specified Gonadal carcinomas Malignant melanomas soft tissue sarcomas

Unspecified soft Other and unspecified Skin carcinomas tissue sarcomas malignant gonadal tumours Other and unspecified carcinomas

1.1.2 Solid tumours of childhood

There are several differences between solid tumours in children and adults in terms of their origin and histological subtypes, etiologic characteristics, their responses to treatment and outcomes (Raymond et al., 2001). In fact, some solid tumours in children never develop in adults, such as neuroblastoma, Wilm’s tumour, rhabdomyosarcoma, some osteosarcomas, some Ewing sarcoma and retinoblastoma (Kline et al., 2003). During prenatal and postnatal periods of development, tissue growth and differentiation take place, which are regulated by cell division and apoptosis. From this point of view, childhood solid tumours are intimately linked to processes such as organogenesis, tissue growth and maturation

(Scotting et al., 2005).

1.1.3 Neuroblastoma

In Australia, there were on average 36 cases of neuroblastoma (NB) diagnosed per year between 1983-2006 (Baade et al., 2010). NB is one of the most frequent extracranial solid tumours in children, the majority of which are diagnosed before the age of five years. NB is believed to originate from neural crest cells of the sympathetic nervous system (Aravindan et al., 2013). The main morphological features of NB are undifferentiated small round blue cells, with hyperchromatic nuclei and scant cytoplasm, pseudo-rosettes surrounding eosinophilic material in the interstitial space, and microcalcifications, which occur in up to 50% of tumours

(Pearson, 2004).

NB has a variable clinical course, ranging from spontaneous remission to rapid

disease progression and death (Messahel et al., 2005; Pui et al., 2011). Spontaneous remission of even large tumours has been well documented and seems to be a common event in infants, in which case NB is described as NB stage 4s (Hero et al.,

2008; Schilling et al., 2002; Woods et al., 2002).

Shimada. et al. (1999) classified NB according to histopathological features as favourable or unfavourable, depending on the grade of cell differentiation. This system has since been modified several times, and today the main rationale is to find a system that clearly differentiates between embryonic NB, which will regress, and those NB that show aggressive clinical behaviour (Cohn et al., 2009; Pearson, 2004).

1.1.4 Rhabdomyosarcoma

In the decade from 1997-2006, an average of 16.5 cases of rhabdomyosarcoma

(RMS) were diagnosed in Australia per year (Baade et al., 2010). The incidence of

RMS in Australia between 2003-2007 was 107 (5.3%) cases in children (0-14 years), and 36 (1.7%) cases in adolescents and young adults (15-29 years), with percentages based on the total number of cancer cases diagnosed in patients 0-29 years, over the 5 year period according to the Australian Institute of Health and

Welfare (AIHW, 2011). The AIHW reported 102 new cases of rhabdomyosarcoma diagnosed between 1983-2007 in patients aged 15 - 19 years, which showed a 5 year survival of 49%. This was the second lowest survival observed for all cancers in adolescents and young adults for that period (AIHW, 2011).

RMS is the most common soft tissue sarcoma of childhood and adolescents and is identified on the basis of histopathology. It exhibits features of skeletal muscle and is

thought to derive from mesenchymal cell precursors that have failed to differentiate, or have otherwise developed aberrantly along the skeletal muscle axis (Meyer et al.,

2004). Originally defined under light microscopy by morphology, RMS are composed of small, round, blue cells that often exhibit cross-striations (Diao et al., 2014;

Raymond et al., 2001).

The two major variants of rhabdomyosarcomas are embryonal RMS (ERMS), whose name comes from the morphological appearance of fetal muscle, and alveolar RMS

(ARMS), because it resembles pulmonary alveoli morphology (Davicioni et al.,

2009). The ERMS variant often occurs in the head, neck, or trunk, and is associated with genomic instability reflecting highly variable karyotypes. In general, ERMS shows less aggressive behaviour and better prognosis than ARMS (Missiaglia et al.,

2012).

ARMS is diagnosed in up to 40% of RMS patients, often in the extremities, and shows characteristic translocations such as t(2;13)(q35;q14) and t(1;13)(p36;q14).

These translocations form a fusion gene between the 5’ end of either PAX3 or PAX7 and the 3’ end of FOXO1. The ARMS subgroup exhibiting translocations has been termed ARMSp (translocation positive) (Williamson et al., 2010). However 45% of

ARMS cases do not exhibit these characteristic translocations, and have been classified as a third group, called ARMSn for “Alveolar rhabdomyosarcoma negative”

(Davicioni et al., 2009). A previous study by Wachtel et al. (2004) also described a third RMS group, based on different protein expression profiles. Here ARMSp showed consistent high expression of AP2β and P-cadherin proteins, ERMS expressed endothelial growth factor and fibrillin-2, while ARMSn, although

presenting with the histopathology of ARMS also expressed some levels of ERMS proteins. This study described ARMSn as an intermediate between ARMSp and

ERMS (Wachtel et al., 2004). Williamson et al. (2010) also described the presence of a third group, ARMSn, although some results showed that ARMSn and ERMS shared whole-chromosome copy number gains of chromosome 8, and high levels of expression of genes on this chromosome (Williamson et al., 2010). Stratification of patients with RMS uses a combination of clinical parameters with the presence or absence of PAX3/FOXO1 chimeric genes (Missiaglia et al., 2012).

1.1.5 Ewing Sarcoma

Ewing sarcoma (ES) is the second most common primary bone cancer after osteosarcoma that affects adolescent and young adults (Toomey et al., 2010).

Baade et al. (2010) described the incidence of ES to be 12.5 cases per year in

Australia between 1997-2006. The most recent study by the Cancer Institute of NSW reported 25 new cases of Ewing sarcoma (16 males and 9 females) between 2004-

2008 in children aged 0-14 years (Tracey et al., 2010) .

James Ewing originally described ES in 1921. It is a highly aggressive bone and soft tissue associated malignancy, the most common sites being the lower extremities

(45%), followed by the pelvis (20%), upper extremities (13%), axial skeleton and ribs

(13%), and face (2%) (Burchill, 2003). ES’s aggressive biological behavior is reflected by the presence of metastases in around 25% of patients at diagnosis

(Surdez et al., 2012). Morphologically, ES displays both mesenchymal stem cell and neural crest stem cell features, contains a tumour cell subpopulation with cancer stem cell properties, and is thought to arise from primary mesenchymal stem cells

(Suva et al., 2009). Sixty percent of patients with localized ES are treated with intensive chemotherapy, followed by surgery and/or radiotherapy. Metastatic ES has a poor prognosis, with 80% of these patients dying earlier than 5 years after diagnosis (Spraker et al., 2012).

1.1.6 Osteosarcoma

Osteosarcoma (OS) is the most common type of bone cancer in children, with an average number of 12.1 cases diagnosed per year in Australia between 1997-2006

(Baade et al., 2010). Between 2003-2007, 62 OS cases were diagnosed in children

0-14 years old, whereas 88 cases were diagnosed in adolescents or young adults

15-29 years old, representing 3% and 4.2% of cancers diagnosed, respectively

(AIHW, 2011).

The defining characteristic of OS is the production of osteoid, which is the unmineralized organic portion of the bone matrix that forms prior to the maturation of bone tissue, in malignant mesenchymal neoplastic cells (Ottaviano et al., 2010). OS can be divided into two broad groups, namely those that appear within the bone or intramedullary OS, which are highly malignant and most frequently occur during adolescence, and those that arise in the surface of bone, are classified as being less aggressive and contain cells that are either well or moderately differentiated (Rytting et al., 2000). Intramedullary OS in patients younger than 20 years of age originate from the metaphyseal region of long bones of arms and legs and the knees and shoulders in areas of rapid growth (Tan et al., 2009). OS is associated with progressive lung metastases (Longhi et al., 2006). An unusually high frequency of genome rearrangements and the presence of a complex karyotype are also

hallmarks of OS (Man et al., 2004). The most important predicting factor outcome remains the histologic response, or “regression grade”, to neoadjuvant chemotherapy, that aims to eradicate micrometastases and reduce tumour burden.

This remains the gold standard for predicting outcome in patients with non- metastatic OS at the time of surgery (Bacci et al., 2002; Smida et al., 2010).

1.2 Cytogenetic changes in childhood cancers - relevance to prognosis

1.2.1 Introduction to cytogenetic changes in cancer

Events that could initiate tumourigenesis have been widely studied. These include the acquisition of somatic copy number aberrations (gains and losses in the number of gene copies) that involve amplification of oncogenes and/or loss of tumour suppressor genes (Fridlyand et al., 2006), point mutations, and structural rearrangements of the genes themselves. Such changes could reflect one or more mechanisms that work to maintain genomic integrity and/or to regulate cell cycle progression (Mattison et al., 2009; Snijders et al., 2003).

The concept of gene amplification refers to an increase of at least five copies of a

DNA segment that is less than 20 megabases in length, while gene gain refers to a lower copy number increase (Myllykangas et al., 2007). Gene gain results from ploidy changes or unequal translocations, is directed towards intact chromosomes, entire chromosome arms, or larger chromosomal regions and usually involves only one intra-chromosomal break-point (Myllykangas et al., 2007).

Amplification is one of the major mechanisms activating oncogenes in the course of cancer progression (Myllykangas et al., 2008). On the other hand, additional common events in cancer are the absence, loss or silencing of functional tumour suppressor genes. This occurs through chromosomal or intra-chromosomal deletions, inactivation of the remaining copy of the tumour suppressor gene through point mutations or by shutting down the gene promoter region by hypermethylation, leaving no functional tumour suppressor gene to protect the organism (Herman et al., 2003; Mao et al., 2007).

1.2.2. Cytogenetic changes in neuroblastoma

The DNA content of NB is usually divided into two categories. Near diploid/tetraploid cases suggest a less aggressive tumour with a fundamental defect in mitosis that produce whole chromosome gains or losses (Brodeur, 2003). In contrast, hyperdiploid tumours are more malignant, with a fundamental defect in genomic stability, and show chromosomal rearrangements, and unbalanced translocations

(Caren et al., 2010). Amplification of the MYCN oncogene located on the distal short arm of chromosome 2 (2p24.3) is a prognostic marker for NB that was identified more than 20 years ago (Schwab et al., 2003). This oncogene is amplified in 20 -

30% of all NB (Goto et al., 2001) and MYCN amplification is strongly associated with advanced disease stage and poor outcome. Tumours with MYCN copy number higher than 10 are classified as a high risk (Castel et al., 2007). Amplification of

MYCN has also been described in RMS and ES (Williamson et al., 2005).

A member of the tyrosine kinase family, Anaplastic Lymphoma Kinase (ALK), located at 2p23-24, is amplified in about 8% of all sporadic NB (Janoueix-Lerosey et al.,

2008). This gene has a role in NB pathogenesis, along with deletion of KIF1B located at 1p36.2 and loss of CADM1 located at 11q23.3 (De Preter et al., 2006)

(Kumps et al., 2013).

Two different groups of aggressive NB have been identified, these being NB with

MYCN amplification, and NB with chromosome 11q deletion (Cohn et al., 2009).

Allelic loss of 11q is present in 35 - 45% of primary NB and is rarely observed in

MYCN-amplified tumours (Attiyeh et al., 2005). Many other chromosomal abnormalities such as chromosome 1p, 3p and 11q deletions have also been identified as statistically significant prognostic factors in NB (Janoueix-Lerosey et al.,

2009). Genes targeted by cytogenetic changes in NB are summarized in Table 1.2.

Table 1.2 Genes targeted by cytogenetic changes in neuroblastoma

Cytogenetic Event Gene(s) targeted Cytoband(s) Prognostic factor

Gene Amplification MYCN 2p24.3 Poor prognosis Copy number gain NME1, NME2, PPM1D 17q Poor prognosis Copy number gain Whole chromosome Chr 17+ Good prognosis Gene Amplification ALK 2p23-24 Poor prognosis

Gene deletion CHD5, KIF1B 1p36.31 Increased risk of relapse Gene deletion CADM1 11q23.2 Poor prognosis Gene deletion RASSF1A 3p21.31 Poor prognosis Somatic Mutations TP53 17p13.1 Poor prognosis

Adapted from: Ambros et al. (2009), Boeva et al. (2013), Van Maerken et al. (2011)

and Van Roy et al. (2009).

1.2.3. Cytogenetic changes in rhabdomyosarcomas

Genomic analysis shows that ERMS tumours have more structural and copy number variations than ARMS (Williamson et al., 2010). ERMS are typically characterized by loss of heterozygosity (LOH), which is understood as a gross chromosomal event leading to the loss of entire genes and surrounding regions (Joseph et al., 2014).

This can include LOH of the short arm of chromosome 11 (11p15.5) where Kratz et al. (2007) found the same somatic HRAS mutation in 5 sporadic ERMS cases.

ERMS has also been associated with mutations in the RAS/NF1 pathway, which are significantly associated with intermediate and high-risk prognosis (Chen. et al.,

2013). Many other chromosomal gains such as chromosomes 2, 7, 8, 12, 13 and 17 have been associated with ERMS (Davicioni et al., 2009; Rao et al., 2010). However many more studies are needed to establish with certainty the prognostic significance of some of these chromosomal rearrangements (Chen. et al., 2013). Genes targeted by cytogenetic changes in ERMS are summarized in Table 1.3.

Structural abnormalities of chromosomes result from chromosomal breakage and rejoining of the broken ends to form new combinations. In a reciprocal translocation, chromosomal material is exchanged between two or more nonhomologous chromosomes (Sandberg et al., 2000). ARMS are characterized by balanced translocations such as t(2;13)(q35;q14) that produces a fusion gene between PAX3 and FOXO1A, detected in 70% of ARMS cases. The translocation t(1;13)(p36;q14) fuses PAX7 to FOXO1A in an analogous manner, and is detected in 10 - 15% of

ARMS cases. ARMS that do not present with typical translocations are called ARMS

negative (ARMSn), may harbour atypical translocations t(2:2) (p23;q35) or t(2;8)

(q35;q13) leading to PAX3/NCOA1 and PAX3/NCOA2 fusions respectively, being detected in a subset of ARMSn cases (Sumegi et al., 2010). Other gene amplification events in ARMS may target genes such as MYCN (2p24), CDK4

(12q13-q14) and MIR17HG (13q31), which encodes the polycistronic microRNA miR-17-92 cluster (Barr et al., 2009; Reichek et al., 2011) (Table 1.4).

Table 1.3 Genes targeted in embryonal rhabdomyosarcom a oncogeness

Cytogenetic Event Gene(s) Targeted Cytoband(s) Prognostic significance

Copy number gain TTC7A 2p21 Poor prognosis

Copy number gain TNS1 2q35 Poor prognosis

Copy number gain GLI2 2q14.2 Poor prognosis

Copy number gain MOGAT1 2q36.1 Poor prognosis

Copy number gain FGFR4 5q35.2-q35.3 Poor prognosis

Copy number gain CREB3L1, DGKZ 11p11.2 Poor prognosis

Copy number gain GLI1, GEFT, OS9, CDK4 12q13.3-q14.1 Poor prognosis

Copy number loss PPP1R12B 1q32.1 Not yet identified

Copy number loss PTPRG, FHIT 3p14.2 Not yet identified

Copy number loss F11, ANKRD37, UFSP2, 4q35.1-35.2 Not yet identified

Copy number loss CDKN2A/B 9p21.3 Not yet identified

Copy number loss NF1 17q11.2 Poor prognosis

Copy number loss ATXN10 22q13.31 Not yet identified

Somatic Mutations KRAS, NRAS 12p12.1, 1p13.2, Poor prognosis

Somatic Mutations PIK3CA 3q26.3 Not yet identified

Somatic Mutations TP53 17p13.1 Not yet identified

Loss of heterozygosity HRAS 11p15.5 Less aggressive

Adapted from: Chen. et al. (2013), Davicioni et al. (2009), Kratz et al. (2007), Rao et al. (2010), and Shern et al. (2014).

Table 1.4 Genes targeted in alveolar rhabdomyosarcoma

Cytogenetic Event Gene(s) Targeted Cytoband(s) Prognostic significance

Gene amplification MYCN 2p24 Poor survival

Gene amplification CDK4 12q13-q14 Poor survival

Gene amplification miR-17-92 cluster 13q31 Poor survival

Chromosomal translocations PAX3/FOXO1A t(2;13)(q35;q14) Aggressive Less aggressive than PAX3 Chromosomal translocations PAX7/FPOXO1A t(1;13)(p36;q14) translocations

Chromosomal translocations PAX3/NCOA1 t(2;2)(p23;q35) Aggressive

Chromosomal translocations PAX3/NCOA2 t(2;8)(q35; q13) Aggressive

Adapted from: Barr (2012), Cao et al. (2010), Kratz et al. (2007), Paulson et al.

(2011) and Shern et al. (2014).

1.2.4. Cytogenetic changes in Ewing sarcoma

Ewing sarcoma (ES) is another example of a malignancy driven by a fusion oncogene. Characteristic causal translocations have been identified between the

EWS gene on chromosome 22q12 and genes encoding two ETS transcription factors, namely FLI1 in chromosome 11q24 or ERG on chromosome 21q22 (Toomey et al., 2010). The EWS gene encodes a multifunctional protein involved in various cellular processes including gene expression, cell signalling and RNA processing and transport (Oakland et al., 2013; Paronetto et al., 2011). Around 85% of ES cases present with t(11;22)(q24;q12) translocations which generate EWS/FLI1 fusion genes whereas another 5-10% present with t(21;22)(q22;q12) translocations that generate a ERG/EWS fusions (Toomey et al., 2010).

The remaining 5% of ES shows alternate translocations that fuse EWSR1 (or a highly similar gene, TLS/FUS) (Ng et al., 2007; Shing et al., 2003) to genes encoding other ETS family transcription factors including ETV1, ETV4 and FEV (Im et al.,

2000; Jedlicka, 2010). EWS/FLI1 and alternate fusions have been classified as oncogenes based on their ability to transform immortalized murine NIH3T3 cells

(Braunreiter et al., 2006). Jedlicka (2010) and Ludwig (2008) have proposed that tumour development occurs because the fusion between EWSR1 and FLI1 leads to the formation of a potent, highly expressed aberrant transcription factor able to activate an oncogenic program in the affected cells.

Other genomic alterations in ES include recurrent copy number gains of chromosome 1q, 8, 12, and 20 as well losses of 16q and 9p21 (Jahromi et al., 2012)

(Table 1.5).

Table 1.5 Genes targeted in Ewing sarcoma

Cytogenetic Event Gene(s) Targeted Cytoband(s) Prognostic significance

Chromosomal translocation EWSR1-FLI1 t(11;22)(q24;q12) Good prognosis

Chromosomal translocation EWSR1-ERG t(21;22)(q22;q12) Good prognosis

Chromosomal translocation EWSR1-ETV1 t(7;22)(p22;q12) Good prognosis

Chromosomal translocation EWSR1-ETV4 t(17;22)(q12;q12) Good prognosis

Chromosomal translocation EWSR1-FEV t(2;22)(q33;q12) Good prognosis

Chromosomal translocation FUS-ERG t(16;21)(p11;q22) Good prognosis

Chromosomal translocation FUS-FEV t(2;16)(q35;p11) Good prognosis

Chromosomal translocation EWSR1-ZSG inv(22) Good prognosis

Copy number gain RELN 7q22.1-q31.1 Metastasis

Copy number gain MYC 8q11.1-qter Poor prognosis

Trisomy Whole Chr 8+ Poor prognosis

Copy number gain ZNF217, SUMO1P1, BCAS1 20p13-pcen Poor prognosis

Copy number gain CEBPB 20q11.21-qter Poor prognosis

Copy number gain SMARCB1 22q11.1-qter Metastasis

Somatic Mutations TP53 17p13.1 Poor prognosis

Adapted from: Huang et al. (2005), Jahromi et al. (2012), Jain et al. (2010) and

Szuhai et al. (2012).

1.2.5. Cytogenetic changes in osteosarcoma

As osteosarcoma (OS) is genetically unstable, with many translocations, amplifications and deletions, finding a common tumour-specific marker has not been possible (Wang, 2005). Nevertheless, common gains of chromosome arms 6p, 8q,

12q and 17p were highlighted when the results of array comparative genomic hybridization were compared (Man et al., 2004; Sadikovic et al., 2009; Squire et al.,

2003).

A variety of oncogenes have been implicated in OS tumourigenesis. Up to 12% of

OS show MYC amplification, while MYC expression is also associated with a higher risk of relapse (Tang et al., 2008). Another oncogene associated with OS is MDM2, located on chromosome 12q14, and an important negative regulator of p53. MDM2 is amplified in up to 10% of OS (Lopes et al., 2001). In the same genomic region lies

CDK4, which is an important kinase regulating cell-cycle progression, and MDM2 and CDK4 are reported to be often co-amplified (Martin et al., 2012; Smida et al.,

2010).

At least three tumour suppressor genes have been associated with osteosarcoma, namely RB1, P16INK4A and TP53. RB1, in its hypophosphorylated state, acts as a tumour suppressor by binding to and inactivating E2F3, resulting in cell cycle arrest

(Kansara et al., 2014). Approximately 70% of sporadic osteosarcoma cases have

RB1 mutations, and RB1 inactivation has been implicated as a poor prognostic factor in OS (Marina et al., 2004). P16INK4A has been associated with the RB1 signalling pathway, and mutations in this gene mimic the RB1 mutation phenotype (Kansara et

al., 2007). P16INK4A mutations also serve as a poor prognostic factor in paediatric

OS (Mohseny et al., 2009). TP53 is frequently mutated in patients with OS (Wagner et al., 2011). TP53 is a transcription factor that regulates cell cycle progression through apoptotic and DNA repair mechanisms (Sandberg et al., 2003). Studies have frequently reported point mutations, gene rearrangements, and allelic loss at the TP53 locus in OS (Tang et al., 2008) (Table 1.6).

Table 1.6 Genes targeted in osteosarcoma

Prognostic Cytogenetic Event Gene(s) Targeted Cytoband(s) significance

HIST2H2BE, HIST2H4B, FLG, Copy number gain ARNT 1q21.1-q21.2 Metastases

Copy number gain RUNX2 6p12.3-p21.1 Metastases

Gene amplification CCND3-PTK7, RUNX2 6p12-p21 Poor prognosis

Gene amplification VEGFA 6p21-p23 Poor prognosis

Gene amplification FGFR1 8p11.2 Metastases

Gene amplification MYC 8q24.21 Poor prognosis

Copy number gain RECQL4 8q24.4 Poor prognosis

Copy number gain FGF19 11q13 Poor prognosis

Copy number gain EGFR 7p11.2 Poor prognosis

Gene amplification VEGFA 6p21.1 Poor prognosis

Gene amplification MDM2 12q13 Metastases

Gene amplification CDK4 12q14 Poor prognosis

Gene amplification E2F3 6p22 Poor prognosis

Gene deletion LSAMP 3q13.31 Poor prognosis

CDKN2A (also known as P16 and Gene deletion INK4A) 9p21 Poor prognosis

Loss of heterozygosity PCDH15, ZWINT1 10q21.1 Poor prognosis

Loss of heterozygosity PTEN 10q23.3 Poor prognosis

Gene deletion WEE1, ST5, LMO1 11p15.1-15.4 Poor prognosis Mutation- Loss of 13q14.13- heterozygosity RB1 13q14.2 Poor prognosis Mutation- Loss of heterozygosity TP53 17p13 Metastases

Adapted from: Kansara et al. (2007), Kansara et al. (2014) Martin et al. (2012),

Sadikovic et al. (2009), Sadikovic et al. (2013), Smida et al. (2010) and Wagner et al. (2011).

1.2.6 Cytogenetic changes relevant to cancer therapy

Many genes affected by the cytogenetic changes characterizing different types of tumours have been key to developing different targeted therapies. These include targeted agents that inhibit molecular pathways crucial for tumour growth and maintenance, and cancer immunotherapy that attempts to stimulate a host immune response to trigger long-lived tumour destruction (Vanneman et al., 2012). Targeted therapies and cytotoxic agents also modulate immune responses, which raises the possibility that these treatment strategies might be effectively combined with immunotherapy to improve clinical outcomes (Mellman et al., 2011; Vanneman et al.,

2012).

Small-molecule drugs have been heavily explored and used as targeted therapies.

The upper molecular weight limit of approximately 900 Daltons offers the advantage of compounds reaching intracellular sites easily. These small molecules can inhibit a specific function of a multifunctional protein, or disrupt protein-protein interactions

(Arkin et al., 2004).

There are many targeted agents in clinical trials, such as the small molecule Nutlin-3

(Van Maerken et al., 2006). Preclinical data have shown potential therapeutic use of nutlin-3 in various paediatric tumours, including NB, RMS, ES and OS (Smith et al.,

2012). Nutlin-3 belongs to the class of selective inhibitors of the p53-MDM2 interaction that is currently most advanced in clinical development (Saha et al.,

2013).

1.3 Technical approaches to detect copy number changes in cancer

1.3.1 Classical cytogenetics

Classical cytogenetic analysis is based on chromosome banding patterns, most commonly produced by the Giemsa stain, that produces G banding. The number of alternating light and dark bands detectable with G-banding in the haploid genome varies with the level of chromosomal contraction in each metaphase cell, with a range of 350–550 bands per haploid set (Gersen, 2005). Since each band represents approximately 5-10 Mb of DNA, each band could potentially contain hundreds of genes (Kirchhoff et al., 2001) (Fig. 1.1). Despite classical cytogenetics not being considered a high-resolution technique, a significant strength is that it provides a global assessment of both numerical and structural abnormalities in a single assay (Table 1.7).

The numerical abnormalities manifest as changes in complete sets of chromosomes, as triploid (3 N) or tetraploid (4 N) complements, or in the number of individual chromosomes, such as loss (monosomy) or gain (e.g. trisomy) of a single chromosome. However standard cytogenetic analysis is also a labour-intensive, comparatively costly, low-throughput technique reliant on highly trained technical staff, and is associated with a sample culture failure rate of 10 – 40% (Gersen,

2005). There are additional concerns that the process of in vitro culturing may select for chromosomal abnormalities that are more viable under these conditions

(Caramins et al., 2011).

Mitelman et al. (2004) reported that although solid tumours represent most malignant

diseases, they could be under investigated due to difficulties in obtaining high-quality metaphase spreads compared with leukaemias, where metaphase chromosomes are readily obtained, and as a consequence, had been more thoroughly investigated using classical cytogenetics (Ozkinay et al., 1979). To overcome this metaphase quality problem, Bauman et al. (1980) first used Fluorescence In Situ Hybridization

(FISH) to detect cryptic/submicroscopic abnormalities, leading to the rise of molecular cytogenetics, based on in situ hybridization (ISH) techniques (Bridge,

2008).

FISH was initially applied to metaphase chromosome spreads (Van Prooijen-Knegt et al., 1982) and since then, many applications have been found, including revealing trisomies, monosomies, fusion genes, specific genes, and translocations, among other applications (Lee et al., 2010). An important advantage of FISH, as opposed to conventional cytogenetics, is that it can be performed on nondividing (interphase) cells from fresh or aged samples (e.g., touch imprint cytological preparations, cytospin preparations, blood smears), sections of paraffin-embedded tissue, and disaggregated cells retrieved from fresh, frozen, or paraffin-embedded material.

Importantly, this procedure can provide results when the tissue is insufficient or unsatisfactory for cytogenetic analysis, when conventional cytogenetic analysis has failed, or when cryptic rearrangements are present (Bridge, 2008).

Resolution capability

Figure 1.1. Resolution capability of different methods to reveal different sized genomic imbalances. Karyotyping can detect large chromosome rearrangements with low resolution (5-10 Mb). FISH provides higher resolution (100 kb) for detecting submicroscopic genomic imbalances. Multiplex Ligation-dependent Probe

Amplification (MLPA) and reverse transcription PCR (RT-PCR) complement FISH and Sanger DNA sequencing, but these methods present limited throughput, low information and the requirement of choosing candidate target genes before the test.

High-resolution microarrays (Oligo microarrays, BAC array CGH) provide the most efficient means to identify genomic imbalances widely, with a resolution less than 500 bp. Figure reproduced from Shen et al. (2009)

Table 1.7 Advantages and limitations of classical cytogenetic analysis

Advantages

• Gives global information in a single assay, includes primary and secondary anomalies • Knowledge of anticipated anomaly or historical diagnosis not necessary • Variants undetectable by interphase FISH or RT-PCR may be uncovered • Sensitive and specific • Can be performed on fine-needle aspirates • Provides direction for molecular studies of pathogenetically important genes

Limitations

• Requires fresh tissue • Although direct preparations can be performed, cell culture required (1-10 days) • May encounter complex karyotypes with suboptimal morphology • Submicroscopic or cryptic rearrangements may result in false-negative results • Normal karyotypes may be observed following therapy-induced tumour necrosis or overgrowth of normal supporting stromal cells • Difficulties with bone tumours and release of cells from bone matrix

Adapted from Bridge (2008), Gersen (2005).

1.3.2 Comparative Genomic Hybridization (CGH) and array CGH

1.3.2.1 Comparative Genomic Hybridization (CGH)

A. Kallioniemi et al. (1992) developed and were among the first to use Comparative

Genomic Hybridization (CGH) in cell lines and primary bladder tumours. This technique, as originally developed, characterizes gross copy number changes, based on the same principles of FISH. However instead of hybridizing a chromosome band, or part of a fusion gene, fluorescently labelled DNA is applied throughout the entire genome (Holcomb et al., 2011). CGH uses two different DNAs, namely a reference DNA (usually normal DNA, typically labeled with a green fluorochrome) and a test DNA (tumour DNA, typically labeled red). The DNAs are co- hybridized to a normal metaphase chromosome spread. The resulting fluorescence ratios of test to reference are then quantified by digital image analysis and used to identify genomic losses or gains in the test sample (Morozova et al., 2008).

The resolution of CGH is relatively low (5-10 Mb) (Holcomb et al., 2011). CGH also required at least two weeks of meticulous work in the laboratory and the ability to accurately karyotype the human genome (Mantripragada et al., 2004). Limitations include being unsuitable to detect reciprocal translocations, Robertsonian translocations, inversions, and ploidy (Carter, 2007). On the other hand, CGH does not require previous knowledge of chromosomal imbalances to start the analysis and obtain satisfactory results (Morozova et al., 2008)

1.3.2.2 Array Comparative Genomic Hybridization (aCGH)

To overcome the problems of low resolution conventional CGH, Kallioniemi et al.

(2008), using similar CGH principles, replaced metaphase chromosomes with arrayed DNA probes attached to a glass slide as targets for hybridization. This improved the performance of the technique and opened it to a range of possibilities and platforms for the array CGH (aCGH) age (Kallioniemi, 2008).

Different aCGH platforms have been developed, the first using bacterial artificial chromosomes (BAC), each containing 100-200 kb of genomic sequence. Due to this large target size, signals were more intense, resulting in better sensitivity to detect copy number changes (Pinkel et al., 2005). However BACs allowed only low genomic resolution, typically 1.5 Mb (Albertson et al., 2000). In 2004, a new version of BAC array, called tiling arrays, was able to cover a whole genome using overlapping sequences, achieving high resolution analysis of a test genome

(Krzywinski et al., 2004). However, producing and maintaining approximately 30,000 genomic clones (the number required for a full human genomic tiling path) was demanding and time consuming (Holcomb et al., 2011).

Since 2004, other probe substrates have been used in aCGH including small insert genomic clones (1.5-2.5 kb), cDNA’s (0.5-2 kb), PCR products (0.1-1.5 kb) and oligonucleotides (25-85 bp) (Carter, 2007). Also the completion of the human genome project had an impact on the availability of probes for aCGH; as probe availability was previously been restricted by available knowledge (Zhang et al.,

2009). The spacing and lengths of the DNA probes determine the genomic resolution

of different aCGH platforms. Oligonucleotide arrays can provide higher resolution

(usually 5-50 kb), but show lower sensitivity, and can fail to reliably detect low-copy number changes due to poorer signals (De Lellis et al., 2007).

One of the first uses for CGH was in cancer, and similarly aCGH has been used in different types of cancers such breast cancer, glioblastoma, rhabdomyosarcoma, ovarian cancer, gastric cancer, prostate cancer (Kallioniemi, 2008; Shinawi et al.,

2008; Watson et al., 2007), haematological cancers (Schwaenen et al., 2004), and bladder tumours (Shaffer et al., 2004). Similarly, aCGH has been applied to human genetic disorders, such as Cri du Chat syndrome caused by partial deletion of chromosome 5p (Levy et al., 2002), Prader-Willi syndrome, and Angelman syndrome

(Buiting, 2010). In both syndromes, the majority of cases are the result of a 3–5 Mb deletion of the Prader-Willi syndrome, and Angelman syndrome critical region on chromosome 15q. These small aberrations cannot be detected using cytogenetics or conventional CGH, but can be readily detected using aCGH (L'Hermine et al., 2003)

(Table 1.8).

Table 1.8 Advantages and limitations of aCGH

Advantages • High reproducibility and precise mapping of aberrations • No need for cell culture, so turn around time is shorter than with cytogenetic methods • Requires only few micrograms of DNA • aCGH can reveal unsuspected genomic imbalances • aCGH detects genomic duplications that cannot be identified by metaphase or interphase FISH analysis

Limitations

• Unable to identify balanced rearrangements such as translocations and inversions • Unable to detect polyploidy • Costly equipment and reagents

Adapted from: Ballif et al. (2006), Shinawi et al. (2008) and Holcomb et al. (2011).

1.3.3 Multiplex Ligation-dependent Probe Amplification (MLPA) and

Quantitative Polymerase Chain Reaction (qPCR)

Basic principles for conventional Polymerase Chain Reaction (PCR) will be mentioned because MLPA and quantitative PCR (qPCR) are based on this technique. PCR amplifies a single copy or a few copies of a piece of DNA across several orders of magnitude, generating thousand to millions of copies of a particular

DNA sequence (Bartlett et al., 2003). The method relies on the thermal cycling, consisting of cycles of repeated heating and cooling of the reaction for DNA melting and enzymatic replication of the DNA. The PCR uses a DNA template (target DNA), at least one pair of specific primers (short fragments of DNA, that contain sequences complementary to the target region), deoxyribonucleotides (dNTPs), a suitable buffer solution and a thermo-stable DNA polymerase.

In the first step, the two strands of the DNA double helix are separated (or

“denatured”) at a high temperature of 94-98°C. In the second step, the temperature is lowered to 50-65°C to allow PCR primer annealing for selective target DNA amplification and this temperature cycle is repeated many times. In the third step, the temperature rises 72°C and the DNA polymerase synthesizes a new DNA strand.

The extension time depends both on the DNA polymerase used, and the length of the DNA fragment to be amplified. The final step is usually performed at 70-74°C for

5-15 minutes after the last PCR cycle to ensure that any remaining single-stranded

DNA is fully extended (Carr et al., 2012). Thus, PCR presents three stages, namely exponential amplification, in which the amount of product doubles every cycle, the levelling off stage, where the reaction slows as the DNA polymerase loses activity

and as consumption of reagents such as dNTPs and primers causes them to become limiting, and finally the plateau stage, where no more product accumulates due to exhaustion of reagents and enzyme (Pavlov et al., 2004).

1.3.3.1 Multiplex ligation-dependent probe amplification (MLPA)

Multiplex ligation-dependent probe amplification (MLPA) is a technique for detecting copy number changes of human genomic DNA sequences using DNA samples derived from blood, amniotic fluid, or tumours (Buffart et al., 2009; Kluwe et al., 2005;

Slater et al., 2003). MLPA is a variant of multiplex PCR, in that MLPA amplifies multiple target genes at the same time. The MLPA reaction can be divided in four main steps as outlined in Fig 1.2.

The first step involves DNA denaturation and probe hybridization, each probe being composed of two “half probes” forward and reverse primer sequences, respectively.

Both halves with a target-specific sequence flanked by a universal primer; one of these half probes also has a variable length stuffer fragment in between to generate the size differences necessary in the probe to allow electrophoretic resolution. In the second step, a specific ligase enzyme ligates the two parts of each MLPA probe to each other. This is followed by PCR amplification using the universal primer pair.

Finally, DNA fragments are separated according to the unique length of each amplicon, followed by quantification according to peak height or area, allowing the calculation of relative amounts of target DNA sequences in the input DNA sample

(Kozlowski et al., 2008; Langerak et al., 2005; Patrinos et al., 2009; Stuppia et al.,

2012) (Figure 1.2).

Figure 1.2 Outline of Multiplex Ligation-dependent Probe Amplification technique. The MLPA reaction can be divided in four main steps: (1) DNA denaturation and probe hybridization; (2) ligation reaction, (3) PCR amplification, and

(4) DNA fragment separation and data analysis. Reproduced from MRC-HOLLAND

MLPA®. https://www.mlpa.com.

The MLPA technique has been applied to neuromuscular disorders, as they are generally caused by deletions or duplications of specific genes (Uwineza et al.,

2013). MLPA has been applied to the diagnosis of spinal muscular atrophy (SMA),

Charcot Marie Tooth, and hereditary neuropathy with liability to pressure palsies

(HNPP) (Stuppia et al., 2012). MLPA has been demonstrated to be a powerful tool for the study of disorders that are caused by deletions or duplications of specific genes (Table 1.9).

1.3.3.2 MLPA and Cancer

MLPA has been used to analyze germline deletions/duplications in genes related to hereditary cancer, for example BRCA1 and BRCA2 in ovarian and breast cancers,

APC in familial adenomatous polyposis, and mismatch repair genes in hereditary non-polyposis colorectal cancer (Chan et al., 2006; Montagna et al., 2003; Nielsen et al., 2007). This technique has also been applied to assess amplification of the HER-

2/neu gene encoding a transmembrane tyrosine kinase, which represents a prognostic marker and a therapeutic target for breast cancer (Purnomosari et al.,

2006).

MLPA has been used to study paediatric cancers such as leukaemia. Veronese et al.

(2013) compared the results obtained using MLPA and interphase FISH (iFISH) in

77 chronic lymphocytic leukaemia (CLL) patients, based on characteristic chromosome abnormalities of this leukaemia. They used a commercial MLPA kit for

CLL, and concluded that MLPA could be used as first-line diagnostic test to detect recurrent aberrations in CLL, however this technique needs to be complemented with iFISH and other molecular techniques. MLPA has been used for diagnosis of

melanocytic tumours since the histopathology sometimes presents problems

(McGinnis et al., 2002). Takata et al. (2005) analysed paraffin-embedded tissue sections of 24 primary melanomas, 14 Spitz nevi and 17 common melanocytic nevi, using commercially available MLPA kits that showed copy number changes in 76 genes spanning almost all chromosome arms. Results from this study concluded that

MLPA could be used as an adjunctive diagnostic tool for melanocytic tumours

(Takata et al., 2005). This technique has also been used in Wilm’s tumour (Hollink et al., 2009), Li-Fraumeni syndrome (Ruijs et al., 2010), medulloblastoma (Jackson et al., 2009), ependymoma (Evans et al., 2007), retinoblastoma (Quelin et al., 2009) and gliomas (Jeuken et al., 2007).

MLPA may be used for detecting common genetic alterations in neuroblastoma, and new kits have been tuned by the manufacturer to fit specific needs. In cooperation with SIOP Europe Neuroblastoma Biology MLPA Committee, a new kit was developed for risk evaluation in neuroblastoma based on 310 neuroblastomas and 8 neuroblastoma cell lines, and results were validated by FISH, BAC array-CGH and

SNP-array. This study resulted in the redesign of the corresponding MLPA kit and a formulation of interpretation standards that were much more successful, making this technique cost-effective, reliable and robust for this type of cancer (Ambros et al.,

2011). In terms of the scope of the current project, it is worth highlighting that MLPA has not been previously used in a mixed tumour cohort including the types of solid tumours under study.

Table 1.9 Advantages and limitations of MLPA

Advantages

• Semi-quantitative technique that can discriminate sequences that differ by only a single nucleotide • Reported to require only 20 ng genomic DNA • High throughput technique, as results can be available within 24 hours when loading a maximum number of 96 samples (depending on thermocycler capacity) • Can examine more than 40 genes per reaction • Less expensive per sample than FISH, requires less manpower and is not observer dependent • Works well on fragmented DNA • Could be used as an initial screening tool in research and diagnostic settings

Limitations

• Only relative changes in copy number are detected. It is not possible to distinguish polyploidy from diploidy or haploidy. • Does not detect balanced translocations or inversions • Not suitable for the detection of unknown point mutations • More sensitive to contaminants (e.g., small remnants of phenol) and DNA degradation than conventional PCR • In tumour samples, it is difficult to detect gene deletions if the sample from which the DNA was derived contained less than 50% cancer cells. • Results may be compromised by the presence of unreliable MLPA probes and unsuited reference material

Adapted from: Bravaccini et al. (2012), Homig-Holzel et al. (2012), Nicholson et al.

(2013), Schouten et al. (2002) and Stuppia et al. (2012).

1.3.3.3 Quantitative Polymerase Chain Reaction (qPCR)

In conventional PCR, amplicons are detected by end-point analysis, whereas in real- time PCR, accumulation is measured as the reaction progresses or in real time, also known as quantitative PCR (qPCR) (Vandesompele et al., 2002). Detection is made possible by including a fluorescent molecule that reports an increase in the amount of DNA, with a proportional increase in fluorescent signal as the PCR progresses.

PCR amplicon accumulation can be measured by a variety of fluorescent detection chemistries, such as the generic double-stranded DNA binding dye SYBR Green I, or sequence-specific probes such TaqMan, adjacent hybridization probes, or molecular beacons.

The excitation and emission maxima of SYBR Green I, allow the use of this dye with any real-time thermocycler (Zipper et al., 2004), also it requires the use of melting curve analysis. Detection takes place during the extension step of real-time PCR.

The fluorescent signal is directly proportional to the DNA concentration, and the linear correlation between PCR product and fluorescent intensity is used to calculate the amount of template present at the beginning of the reaction.

Absolute quantification requires the addition of external standards in every reaction to determine the absolute amount of the target nucleic acid of interest. Quantities obtained from a qPCR experiment must be normalized in such a way that the data become biologically meaningful. This is done by using a normalizer or housekeeping gene to correct for intersample variations between different experiments, and to adjust or standardize the obtained target quantity to the unit amount of sample

(number of cells or micrograms of nucleic acid) used as a PCR template (Sails et al.,

2009) (Table 1.10).

qPCR has been used in different fields. In microbiology, qPCR is used to analyze food safety, food spoilage and fermentation and for microbial risk assessment of water quality (Gadkar et al., 2013). qPCR is applied for diagnosing infectious disease

(Sails et al., 2009), such as the Swine flu outbreak in humans (Poon et al., 2009).

qPCR remains the method of choice for validating results obtained from whole- genome screening techniques such as aCGH (Benes et al., 2010), and is also an important technique for the analysis of circulating tumour cells (CTC). These cells are released from the primary tumour and enter the blood circulation, from where they can migrate to distant organs, implant and produce metastasis. Analysis of CTC includes an initial enrichment step with use of specific markers to identify the cells and distinguish them from leukocytes. qPCR offers a significant improvement over conventional reverse transcription PCR, as the quantification of markers present in both tumour and non-tumour cells allows the discrimination between authentic CTCs and contaminant cells (Alunni-Fabbroni et al., 2010; Xi et al., 2007).

Table 1.10 Advantages and limitations of qPCR

Advantages

• Simple and straightforward technique • No need for post-PCR manipulations • High throughput screening capacity and degree of automation • Can be performed with minimal amounts of sample (25 ng of DNA) in a few hours time • High technical sensitivity (<5 copies) and a high precision (<2% standard deviation)

Limitations

• The amplicon size cannot be monitored without opening the system • Requires a reliable source of template of known concentration for absolute quantification • Costly compared to MLPA and conventional PCR • Standards must be amplified in parallel with the sample every time the experiment is performed, which increases the cost and time per reaction

Adapted from: Bartlett et al. (2003), Bustin et al. (2009), Dhanasekaran et al. (2010),

Wang et al. (2013a) and Zipper et al. (2004).

1.4 Chromosome 8q gain in adult cancers

The gain of the long arm of chromosome 8 is one of the most common cytogenetic events that occur in cancer. The entire chromosomal arm is often gained, which may reflect that chromosome 8q harbours multiple oncogenes. Proposed chromosome 8q amplification targets include MYC, MAL2, WWP1, and TPD52 (Saramaki et al.,

2007), but there are likely others to be identified. As will be described, it is not fully understood whether the same or different genes are targeted in different tumours.

1.4.1 Breast cancer

Several studies have demonstrated that chromosome 8 is very frequently altered in breast cancer. Rummukainen et al. (2001) reported chromosome 8q gains that often included either the whole q arm or the telomeric region, whereas chromosome 8p showed either diploid copy number or loss. Naylor et al. (2005) also underlined the gain of 8q24 from 139.3-144.8 Mb in 79% (37/47) primary breast tumours and 83%

(15/18) breast cancer cell lines. This emphasized a number of genes of potential interest such as MYC, PTK2 and KCNK9. Two years later, using a high resolution

BAC array, Rodriguez et al. (2007) revealed many novel regions of copy number gain/loss in the SK-BR-3 breast cancer cell line, with amplification of LACTB2,

ZFHX4, MMP16, and DECR1, all located on chromosome 8, as well as copy number loss at TP53INP1, ZFPM2 and COL22A1. These new findings were validated using

FISH probes (Rodriguez et al., 2007). One of the most commonly amplified genes in breast cancer is the oncogene MYC located at 8q24, and its role as a driver in the

8q24 amplicon is well established (Kao et al., 2009; Rodriguez et al., 2007). MYC is amplified in approximately 5% of breast cancers and is associated with high grade,

advanced stage, and poor patient survival (Chen et al., 2008; Rodriguez-Pinilla et al.,

2007).

Other chromosome 8q amplicons have been associated with breast cancer.

Choschzick et al. (2010) used a tissue microarray that contained 2000 breast cancers to study the significance of proximal chromosome 8q21 amplification in terms of breast cancer phenotype and patient prognosis. TMEM70 amplification was found to be significantly associated with reduced overall survival (Choschzick et al.,

2010). Other specific amplicons at chromosome 8q24.3, 8p11.2, and 17q21.33- q25.1 were associated with early relapse in estrogen receptor-positive breast cancers, despite adjuvant hormonal therapy (Bilal et al., 2012).

The chromosomal region 8p11-p12 has been found frequently amplified in 10%-15% breast carcinoma cases (Ethier, 2003). Turner et al. (2010) confirmed that overexpression of FGFR1 located at 8p11-p12 was highly correlated with copy number gain. FGFR1 amplification was also a frequent event in the proliferative luminal B subtype, and ER-positive breast cancers, suggesting that FGFR1 overexpression may contribute to the poor prognosis of breast cancer (Turner et al.,

2010).

1.4.2 Prostate cancer

Aberrations of chromosome 8 are one of the most frequent cytogenetic changes in prostate cancer (Hughes et al., 2006), with deletions of 8p and gains of 8q being among the most frequent genomic alterations (El Gammal et al., 2010). Using FISH,

El Gammal et al. (2010) analysed 1,954 prostate cancers in tissue microarray

format, and concluded that whereas 8p deletions and 8q gains were relatively rare in early stage prostate cancer, 8q gains markedly increased from primary to metastatic cancers (El Gammal et al., 2010). Monoallelic deletions on chromosome 8p were associated with more advanced invasive and aggressive prostate cancers (Bethel et al., 2006).

1.4.3 Colorectal cancer

One of the most common cancers is colorectal cancer (CRC) and in the majority of patients the tumour has already metastasized to lymph nodes and/or the liver at the time of diagnosis (Ghadimi et al., 2003). Chromosomal abnormalities occur in a non- random pattern along the pathway from adenoma to carcinoma and then to advanced lesions and the formation of metastases (Berg et al., 2010). Using data from 244 CRC cases previously analysed by CGH Li et al. (2011) applied a branching tree and multiple distance-based tree models to identify the six most common gains of chromosomal regions. Chromosome 8q gain was present in 48.3% of cases, and loss of chromosome 8p12-p23 in 50.9% of cases was also among the nine most common losses (Li et al., 2011).

1.4.4 Human cancer cell lines

By combining high-resolution aCGH analysis of 598 human cancer cell lines with insertion sites from 1,005 mouse tumours, Mattison et al. (2010) created a cross- species oncogenomic data-set to reveal candidate tumour suppressor genes and oncogenes mutated in both human and mouse tumours. In this study, 27 genes were recurrently amplified in human tumours, with 6 genes located on chromosome 8,

namely MYC, SLA, TRP53INP1, PTP4A3, PAG1, and TPD52 (Mattison et al., 2010).

Some of these genes are already well-known oncogenes, or genes implicated in cellular processes including growth, and tumourigenesis. This study again highlighted that chromosome 8 has an important role in tumourigenesis and cancer progression (Mattison et al., 2010).

1.5 Chromosome 8 gain in childhood cancers

A clear association between DNA copy number aberrations and prognosis has been found in a variety of childhood tumours (section 1.2). However, chromosome 8q gain has been poorly characterized in childhood tumours, relative to common adult cancers.

1.5.1 Neuroblastoma

The MYC family has three members: MYC, MYCN and MYCL, and all have been implicated in different cancer types (Prochownik et al., 2007).In human neuroblastoma (NB), a significant body of literature describes basic, translational, and clinical data on MYCN whereas little is known and reported about the role of

MYC on chromosome 8q24.21 (Wang et al., 2013b). Some MYC transcription factors, including MYCN, also induce apoptosis, a function that should inhibit tumour formation and could be involved in spontaneous tumour regression, a phenomenon that is present in some NB (Adhikary et al., 2005).

Interestingly, neuroblastoma-derived cell lines that lacked amplified MYCN, expressed MYC at abnormal levels (Westermann et al., 2008). Similarly, MYC

protein expression was almost exclusively found in MYCN-non-amplified tumours, and this was associated with poor prognosis for the undifferentiated subtype (Wang et al., 2013b). Although chromosome 8 aberrations are not the most common in NB,

Pezzolo et al. (2009) reported 6/23 NB cases with chromosome 8q24.3 gain using aCGH (Fig 1.3). Similarly, Ambros et al. (2011) reported, among other chromosome aberrations, infrequent additional segmental aberrations in chromosome 8 in 4/10 neuroblastic tumours: copy number gain in chromosome 8p in 2 cases, and copy number gain and loss of chromosome 8q in the other 2 cases (Ambros et al., 2011).

Figure 1.3 Summary of patterns of total gains and losses detected in 23 localized resectable neuroblastoma samples. Summary of patterns of total gains

(right, green lines), and losses (left, red lines), with numbers of cases on top of each line in 23 neuroblastomas using array-CGH. Reproduced from Pezzolo et al. (2009).

1.5.2 Rhabdomyosarcoma

Gain of chromosome 8 is one of the most common chromosomal alterations in

ERMS and ARMSn (Fig 1.4) (Williamson et al., 2010). Increased expression of chromosome 8 genes was reported in ERMS tumours with chromosome 8 gains relative to those tumours without gain (Chen. et al., 2013). ARMSn cases may harbour atypical t(2;8)(q35;q13) translocations producing PAX3/NCOA2 fusion genes (Sumegi et al., 2010), which have been postulated to play a dual role in RMS tumourigenesis by inhibiting myogenic differentiation and promoting cell proliferation

(Yoshida et al., 2013). Using hierarchical clustering analysis of gene expression in

160 RMS, two main groups of RMS were found. The group that expressed the chimeric transcription factor PAX3/FOXO1 or PAX7/FOXO1 were mainly ARMS. It was noted that genes with increased expression in tumours lacking PAX/FOXO1 translocations were mostly localized on chromosome 8q, specifically at chromosome

8q24 (Davicioni et al., 2009).

Figure 1.4 Frequency of specific amplifications and gains in ARMSp versus

ARMSn and ERMS. A. Plots of frequency of gains (red), losses (green), and amplifications (black) by subtype ARMSp are represented by PAX3/FOXO1 and

PAX7/FOXO1 fusion positive cases, whereas ERMS and ARMSn involve more whole-chromosome changes B. Frequency of specific changes listed by RMS subtype. Reproduced from Williamson et al. (2010).

1.5.3 Ewing Sarcoma

Gain of chromosome 8 is one of the most frequent chromosome alterations in Ewing sarcoma (Toomey et al., 2010). For example, in the study of Savola et al. (2009), gain of chromosome 8q was the most frequent copy number change, occurring in

67% (21/31) cases (Fig 1.5). Trisomies of chromosome 8, 12 and gain of chromosome 1q have been associated with poor clinical outcome in Ewing sarcoma

(Amiel et al., 2003; Roberts et al., 2008). MYC was proposed as a candidate driver linked to poor prognosis, as trisomy 8 can be full or partial, and sporadic cases can show focal MYC gain (Jahromi et al., 2012). Full or partial trisomy 8 is associated with poor clinical outcome and this significance increased when focal MYC gains were considered (Jahromi et al., 2012; Mackintosh et al., 2010).

Figure 1.5 Chromosomal locations of copy number changes in Ewing sarcoma samples (n=31). The ideogram shows the summary of gains and losses of DNA copy number and their frequencies in Ewing sarcomas (n=31) analysed by aCGH.

Gains (green) are shown on the right of each chromosome and losses (red) on the left (numbers refer to the percentages of gain or loss per band). Reproduced from

Savola et al. (2009).

1.5.4 Osteosarcoma

Chromosome 8q gain was identified as the most frequent aberration in 47 high-grade osteosarcoma samples (Ozaki et al., 2002). Copy number gain at 8q24 harbouring

MYC is one of the most frequent genomic alterations that has been also associated with adverse overall survival (Mirabello et al., 2010; Smida et al., 2010).

Other regions on chromosome 8q including 8q21.3-8q23, and 8q21 commonly present copy number gain, suggesting that other oncogenes located within these bands could have roles in osteosarcoma pathogenesis (Fig 1.6) (Batanian et al.,

2002; Bayani et al., 2003; Kresse et al., 2009). Copy number loss on 8p also has been associated with sporadic osteosarcoma (Atiye et al., 2005; Squire et al., 2003).

An integrative approach involving integrative whole-genome analysis of DNA copy number, promoter methylation, and gene expression analysis in 10 primary osteosarcomas showed the region with the most significant copy number loss was chromosome 8p21.2-p21.3 (Sadikovic et al., 2009).

Figure 1.6 Genetic aberrations in human osteosarcomas samples

The X axis numbered with 1-22 denotes chromosome numbers, whereas the Y axis shows the fractions of tumours with gain or loss. Red denotes significantly recurrent amplifications and green denotes significantly recurrent deletions. Grey represents nonsignificant of aberrations. A. Genetic aberrations in 20 osteosarcomas measured using aCGH B. Genetic aberrations in 36 Norwegian osteosarcomas measured using BAC arrays. Reproduced from Du et al. (2014).

1.6 Statement of the problem, hypothesis and aims of this research

Human chromosome 8 contains a number of important genes that are amplified in adult cancers. Specifically, the long arm of chromosome 8 is often gained, reflecting the multiple oncogenes it harbours, including MYC, WWP1 and TPD52, although others may remain to be identified. In contrast, chromosome 8 gain has been comparatively poorly characterized in paediatric solid tumours.

In this research, we will explore and determine the significance of copy-number changes on chromosome 8 in 4 different types of paediatric solid tumours, namely osteosarcoma, Ewing sarcoma, rhabdomyosarcoma and neuroblastoma. The study will be based on a total of 31 cases obtained from the Tumour Bank, Children’s

Hospital at Westmead.

1.6.1 Hypotheses

(1) Critical chromosome 8 amplification target genes contribute to the formation

and progression of paediatric solid tumours.

Aim: To produce and mine genomic profiles for a cohort of human cancer

cell lines, and paediatric solid tumours diagnosed at Children’s Hospital at

Westmead using the MLPA Salsa kit for chromosome 8, and high-density

array comparative genomic hybridization.

(2) Different methods of detecting copy number changes should produce

comparable results.

Aim: To compare results obtained using MLPA and high-density aCGH. A

validating third technique, quantitative PCR, will be used to examine

agreements or disagreements between the techniques.

In summary, this project has the capacity to identify novel amplified genomic regions, particularly in childhood cancers that have been under-investigated relative to other types, and to compare MLPA and aCGH as research and diagnostic tools.

Chapter 2

Materials and Methods

2.1 Materials

Table 2.1 Reagents and suppliers

Reagent Supplier Application

Agarose Bio-Rad DNA quality verification DNase/RNase-free distilled water Invitrogen aCGH, qPCR, MLPA

DNeasy Blood & Tissue Kit Qiagen DNA extraction

Ethanol (95% to 100% molecular biology grade) Promega DNA extraction, aCGH Proteinase K Qiagen DNA extraction

QuantiFast SYBR Green Kit Qiagen qPCR

Male genomic DNA Promega MLPA

SALSA MLPA Probemix P014-B1 Chromosome 8 MRC-Holland MLPA

Genomic DNA Enzymatic labelling kit Agilent aCGH

Rsa I BioLabs aCGH

Alu I BioLabs aCGH

Agilent Oligo aCGH hybridization kit Agilent aCGH

Agilent Oligo aCGH wash buffer 1 and 2 set Agilent aCGH

Hybridization chamber Agilent aCGH

Stabilization and Drying solution Agilent aCGH

1x TE (pH 8.0), Molecular grade Promega aCGH

Amicon Ultra-0.5, ultracel-30 Membrane, 30kDa. Millipore aCGH

SurePrint G3 CGH Customized microarrays 2X400K Agilent aCGH

Hybridization chamber, stainless Agilent aCGH

Human Cot-1 DNA Invitrogen aCGH

Male and female human reference DNA Coriell Institute aCGH

2.2 Breast cell lines

This project used 5 breast cancer cell lines AU-565, HCC2157, MCF-7,

MDA-MB-231, and SK-BR-3 and a breast epithelial cell line MCF-10A, a near normal breast cell line that was spontaneously immortalized and does not show characteristics of invasiveness or tumour formation (Marella et al., 2009). All cell lines were already in the laboratory from previous projects, and all of them had been previously short tandem repeat-typed to ensure they were correct and uncontaminated. In general the cell lines were stored in liquid nitrogen and thawed by placing sample tubes in a room temperature water bath until half of each tube was thawed. Thereafter cells were transferred to a new tube containing 500µL RPMI media, centrifuged for 1 min at 2000 rpm at RT, and the supernatant was discarded.

Cells were plated in 10 cm tissue culture dishes and cultured in RPMI media (Life

Technologies) supplemented with 10% foetal bovine serum (FBS) (Life

Technologies) in a humidified incubator at 37˚C with 5% CO2 until 80% of cells were confluent. Cells were then harvested by scraping the tissue culture dish for genomic

DNA extraction.

2.3 Paediatric solid tumour cohort

Tumour tissues was removed through surgery from patients by qualified staff and stored in the Tumour Bank at the Children’s Hospital at Westmead. Parental or guardian consent was obtained for the use of tissue samples in unspecified future research. Four different types of solid tumours were received, namely osteosarcoma

(5 samples), Ewing sarcoma (8 samples), rhabdomyosarcoma (8 samples) and neuroblastoma (10 samples), for a total of 31 tumour tissue samples. All samples

were from different patients; 26 tumour samples were taken at diagnosis and 5 tumour samples were post chemotherapy (Table 2.2). Tumour tissue specimens were then stored at -80˚C in the Tumour Bank at the Children’s Hospital at

Westmead. Frozen tumour tissues were used for genomic DNA extraction.

Patients were diagnosed between 1986 and 2009. Specimens were de-identified with a laboratory number to protect patient identity (Table 2.2). Dr Lesley Thwe and

Ms Amanda Rush obtained relevant clinical data through their access to patient databases and medical records of The Children’s Hospital at Westmead.

This project was approved by the Human Research Ethics Committee, Sydney

Children’s Hospitals Network and the Children’s Hospital Tumour Bank at

Westmead.

Table 2.2 Summary of paediatric tumour cohort

Sample Patient Sample Diagnosis Age at diagnosis ID Gender status

F 9-O F Osteosarcoma 13 years - 6 months Diagnosis F 10-O F Osteosarcoma 9 years -11 months Diagnosis M 11-O M Osteosarcoma 15 years Diagnosis F 12-O F Osteosarcoma 4 years - 9 months Diagnosis M 46-O M Osteosarcoma 10 years - 11 months Diagnosis M 13-E M Ewing Sarcoma 2 years - 5 months Diagnosis M 14-E M Ewing Sarcoma 10 years -1 month Diagnosis F 15-E F Ewing Sarcoma 11 years - 2 months Diagnosis M 16-E M Ewing Sarcoma; PNET 11 years - 2 months Diagnosis F 17-E F Ewing Sarcoma 10 years - 9 months Diagnosis F 18-E F Ewing Sarcoma 5 years - 3 months Diagnosis 13 years -11 Diagnosis M 22-E M Ewing Sarcoma months F 24-E F Ewing Sarcoma 7 years - 10 months Diagnosis M 25-R M Rhabdomyosarcoma; Embryonal 7 years - 10 months Diagnosis F 26-R F Rhabdomyosarcoma; Embryonal 1 year - 7 months Diagnosis M 27-R M Rhabdomyosarcoma; Alveolar 13 years - 3 months Diagnosis F 28-R F Rhabdomyosarcoma; Embryonal 10 years - 8 months Diagnosis F 29-R F Rhabdomyosarcoma; Embryonal 2 years - 7 months Diagnosis M 31-R M Rhabdomyosarcoma; Embryonal 4 years - 4 months Diagnosis M 32-R M Rhabdomyosarcoma; Alveolar 13 years - 8 months Diagnosis M 35-R M Rhabdomyosarcoma; Embryonal 3 years - 1 month Post-chemo M 36-N M Neuroblastoma 4 years - 1 month Diagnosis F 37-N F Neuroblastoma 1 year - 2 months Diagnosis F 38-N F Neuroblastoma 7 months Diagnosis M 39-N M Neuroblastoma 1 year - 11 months Diagnosis F 40-N F Ganglioneuroblastoma 3 years - 3 months Post-chemo M 41-N M Neuroblastoma 6 years - 9 months Post-chemo F 42-N F Neuroblastoma 1year - 9 months Post-chemo M 43-N M Neuroblastoma 1year - 6 months Diagnosis F 44-N F Neuroblastoma 2 months Diagnosis F 45-N F Neuroblastoma 1year - 2months Post-chemo

2.4 Genomic DNA extraction from frozen tissues and cultured cells

The DNeasy blood & Tissue (QIAGEN) kit was used to extract genomic DNA from frozen tissue and cultured cells, with the primary difference being that tissue required a longer incubation after adding ALT tissue lysis buffer and proteinase K. Frozen tissue was cut into small pieces of approximately 0.2-0.4 cm length, placed in a 1.5 ml microcentrifuge tube, and then 180µL of buffer ALT and 20µL ready-to-use proteinase K solution [40 mAU/mg solution] were added. Samples were vortexed occasionally during incubation at 56˚C, and then placed at 56˚C overnight in a rotating oven until the tissue was completely lysed. Cultured cells were harvested via trypsinisation, washed with 200µL PBS, and then 20µL ready-to-use proteinase K solution and 180µL Lysis buffer (AL) were added, mixed thoroughly by vortexing, and samples were incubated at 56˚C for 10 min.

After incubation at 56˚C, 200µL 100% ethanol was added to each tube, and mixed by vortexing. Each sample was immediately loaded onto a DNeasy Mini spin column, which was placed in a 1.5mL collection tube. Tubes and columns were centrifuged at

6000xg for 1 min at RT, the flow-through was discarded and the DNeasy Mini spin column was placed in a fresh collection tube. Spin columns were washed with 500µL

AW1 Buffer and then centrifuged at 6000xg for 1 min. The spin column was placed in a fresh collection tube and a second wash of 500µl AW2 Buffer was added to the spin column. To elute DNA, the spin column was placed in fresh collection tube,

200µL AE Buffer was added directly onto the DNeasy membrane and incubated for 1 minute at RT. Spin columns were then centrifuged at 6000xg for 1 minute at RT and

the eluted genomic DNA was then stored at ‐20°C until required for quantification and later use.

2.5 Spectrophotometric and electrophoretic analysis of genomic DNA

DNA samples were quantified with a ND-1000 UV-VIS NanoDrop spectrophotometer

(Thermo Scientific), with A260/A280 ratios used as an indicator of DNA purity and

A260/A230 ratios used to detect other contaminations such as EDTA, carbohydrates, phenol, and TRIzol reagent that absorb at 230nm (NanoDrop software v3.1.2)

(Thermo-scientific, 2011) . Accepted parameters were 50-300 ng/µL for DNA concentration, 1.70 - 2.0 for A260/A280 ratios, and 1.90 - 3.0 for A260/A230 ratios.

Furthermore, DNA quality (high molecular weight) was assessed by performing electrophoresis with 10 ng DNA on a 0.8% (w/v) agarose gel in 1X TBE.

Agarose gels were run at 180 V for 40 minutes using a Bio-Rad Powerpac 300 electrophoresis power supply. The gel was stained with ethidium bromide by adding

50 µg ethidium bromide to a 100 ml agarose gel. The DNA was visualised using a

UV transilluminator (Advanced Technocracy Inc.) and photographed with a UVP

PhotoDoc-It DigiCam 70.

2.6 Multiplex Ligation-dependent Probe Amplification (MLPA)

The SALSA MLPA KIT P014-A1 Chromosome 8 kit (MRC Holland, Amsterdam) was used, which contains 32 probes in 30 genes along chromosome 8, and 9 control probes located in 6 genes on other chromosomes (chromosomes 3, 5, 11, 12, 13 and 17) (Table 2.3). MLPA products were analysed using an ABI-prism 3100 DNA

sequencer, and raw data were analysed using PeakScanner Software v1.0 (Applied

Biosystems). For copy number quantification, peak area values were exported to an

Excel template file (Microsoft 7) developed by Sydney Genome Diagnostics at

Children’s Hospital at Westmead.

The manufacturer states that probe ratios below 0.7 indicate heterozygous deletion

(copy number change from two to one allele), and that probe ratios above 1.3 indicate allele duplication (copy number change from two to three alleles) or greater

(MRC-Holland, 2011). Thus, the presence of two copies of a target sequence is indicated by probe ratios of 0.7-1.3 (Schouten et al., 2002). The quality of each

MLPA reaction was evaluated by the results obtained for a normal control. In this project we used a commercially available human male genomic DNA pool

(Promega), similar to Buffart et al. (2009). This was run in triplicate with each run to provide information about amplification efficiency and whether the correct amount of

DNA had been used for each MLPA reaction (Caceres et al., 2011).

2.6.1 MLPA reactions (day 1)

Five microliters (10-20 ng/µL) genomic DNA extracted and quantified according to sections 2.4 and 2.5, was added to a 0.2 mL PCR tube, placed in a 9600 Perkin-

Elmer thermocycler and denaturation was performed at 98°C for 5 min. The normal male DNA reference sample was included in triplicate, as per MRC-Holland recommendations. These reference samples were distributed randomly over the sample plate to account for placement variation. Samples were allowed to cool to

25°C before proceeding with the following steps.

Hybridisation master mix for each sample was prepared using 1.5 µL MLPA buffer and 1.5 µL probemix in a clean PCR tube. The sample was mixed by pipetting, and 3

µL hybridization master mix was added to each sample at 25°C, and mixed by pipetting. Samples were incubated at 95°C for 1 min and then the hybridization program was run at 60°C for 16-20 hours.

The next day, ligase master mix was prepared for each reaction using 3 µL Ligase-

65 buffer A, 3 µL Ligase-65 buffer B and 25 µL dH2O in a clean tube, and mixed by pipetting. Finally, 1 µL Ligase-65 was added to the master mix, and then 32 µL ligase master mix was added to each sample tube at 54°C and mixed by pipetting.

Samples were incubated at 54°C for 15 min to allow ligation. Heat inactivation of ligase-65 was performed at 98°C for 5 min. Samples were allowed to cool to 15°C before proceeding with the next steps.

2.6.2 MLPA reactions (day 2)

PCR buffer mix was prepared using 4 µL SALSA PCR buffer and 26 µL dH2O, which were mixed briefly by vortexing. A total of 30 µL PCR buffer mix was transferred to each sample tube containing ligation product. A polymerase master mix was prepared using 2 µL SALSA PCR primers, 2 µL SALSA enzyme dilution buffer, 5.5

µL dH2O and 0.5 µL SALSA polymerase, mixed by pipetting and stored on ice until required. Sample tubes containing PCR buffer mix and ligation product were placed in the 9600 thermocycler at 60°C, and 10 µL polymerase master mix was added to each sample tube and mixed by pipetting. The thermocycler program was as follows:

30 cycles of (95°C for 30 sec, 60°C for 30 sec and 72°C 60 sec) followed by an

incubation at 72°C for 20 min. Samples were cooled to 15°C and stored at 4°C until required.

2.6.3 MLPA Fragment separation and peak pattern evaluation

The Australian Genome Research Facility (AGRF) at Westmead processed all samples using capillary electrophoresis fragmentation, a process that was conducted with the ABI-prism 3100 sequencer equipment and according to the manufacturer’s specific conditions for the SALSA MLPA Chromosome 8 kit.

Visual peak patterns were evaluated using the PeakScanner Software v1.0 (Applied

Biosystems). The kit contained four DNA quality control fragments (Q-fragments)

(64,70,76, and 82 bp) and one ligation-dependent control fragment at 92 bp. These fragments already contained both MLPA primer sequences, meaning that they did not need to be hybridised or ligated to be amplified as they act as a “quality control reaction”, to provide a warning when insufficient sample DNA is present (MRC-

Holland, 2011).

In consequence, they are present in very small quantities and their amplification should be easily outcompeted by MLPA probes when the latter are present in normal amounts. Thus when the peaks of the Q-fragments are low or invisible, this signifies that sufficient DNA sample was present and the ligation reaction was successful. In contrast, when the peaks of all four Q-fragments were higher than one third of the height of the 92 bp ligation-dependent control fragment and the MLPA probes, this was an indication of insufficient DNA or that the ligation reaction had failed (MRC-

Holland, 2011).

2.6.4. Data analysis

All data that passed the peak pattern evaluation were subjected to analysis. Intra- sample normalization of electrophoresis results was conducted to correct for common sources of results variability such as pipetting accuracy, bleaching of the fluorescent label and variations in capillarity. The intra-normalised probe ratio of a sample was divided by the average intra-normalised probe ratio of all (n=3) reference samples. This step provides a probe ratio for each probe for each sample

(MRC-Holland, 2011).

Table 2.3 Coordinates of genes examined using MLPA according to the hg18 genome assembly. (UCSC Genome Browser Mar. 2006 NCBI36) MLPA Probes Chromosome Gene Cyto Exon Product Gene Coordinates Size Band N0 Size (bp)

Chromosome 8

1 DLGAP 1,484,199-1,644,049 159,851 bp 8p23.3 9 166 2 MFHAS 8,679,409-8,788,541 109,133 bp 8p23.1 1 481 3 MSRA 9,990,476-10,323,805 333,330 bp 8p23.1 3 364 4 GATA4 11,599,126-11,603,846 4,721 bp 8p23.1 7 418 5 CTSB 11,737,443-11,763,055 25,613 bp 8p22 9 136 6 TUSC3 15,442,101-15,666,366 224,266 bp 8p22 5 184 7 ChGn 19,341,621-19,504,550 162,930 bp 8p21.3 6 454 8 FGFR1 38,396,208-38,444,520 48,313 bp 8p11.2 4 373 9 PRKDC 48,848,222-49,035,296 187,075 bp 8q11.21 7 400 10 MOS 57,188,055-57,189,095 1,041 bp 8q11.23 1 247 11 CHD7 61,753,893-61,942,021 188,129 bp 8q12.2 15 256 12 MYBL 67,636,968-67,687,729 50,762 bp 8q13.1 14 177 13 NCOA2 71,186,821-71,478,574 291,754 bp 8q13.3 3 337 14 TPD52 80,993,650-81,155,565 161,916 bp 8q21.13 9 202 15 E2F5 86,287,162-86,314,006 26,844 bp 8q21.2 6 274 16 RAD54B 95,459,972-95,518,308 58,337 bp 8q22.1 2 292 17 ST7 105,578,376-105,670,344 91,969 bp 8q22.3 5 355 18 EIF3E 109,283,148-109,330,135 46,988 bp 8q23.1 7 142 19 EIF3H 117,726,236-117,847,675 121,440 bp 8q24.11 8 427 20 EXT1 118,880,783-119,193,239 312,457 bp 8q24.11 1 328 21 RNF139 125,556,189-125,570,040 13,852 bp 8q24.13 2 409 22 MYC 128,817,498-128,822,855 5,358 bp 8q24.21 3 238 23 MYC 128,817,498-128,822,855 5,358 bp 8q24.21 3 157 24 DDEF1 131,268,664-131,509,230 240,567 bp 8q24.22 27 301 25 KCNQ3 133,210,438-133,562,186 351,749 bp 8q24.22 13 436 26 SLA 134,130,006-134,156,557 26,552 bp 8q24.22 9 220 27 KHDR 136,538,898-136,729,030 190,133 bp 8q24.22 4 382 28 KCNK9 140,699,431-140,784,481 85,051 bp 8q24.3 1 346 29 PTK2 141,737,683-142,080,514 342,832 bp 8q24.3 26 319 30 PTP4A 142,501,189-142,510,802 9,614 bp 8q24.3 3 265 31 RECQL4 145,707,479-145,714,008 6,530 bp 8q24.3 17 463 32 RECQL4 145,707,479-145,714,008 6,530 bp 8q24.3 13 211

Reference probes

1 VHL 10,18,319-10,168,746 10,428bp 3p25.3 2 391 2 IL4 132,037,272-132,046,267 8,996bp 5q31.1 1 130 3 EHF 34,599,244-34,639,657 40,414bp 11p12 8 472 4 HIPK3 33,235,744-33,332,515 96,772bp 11p13 17 193 5 CTTN 69,922,260-69,960,338 38,079bp 11q13 16 229 6 CD27 6,424,312-6,431,145 6,834bp 12p13 6 444 7 ABCC4 94,470,084-94,751,688 281,605bp 13q32 23 148 8 ARHGEF7 110,565,625-110,745,543 179,919bp 13q34 24 310 9 PMP22 15,073,821-15,104,818 30,998bp 17p12 4 283

2.7 Array Comparative Genomic Hybridization (aCGH)

aCGH is a molecular cytogenetic technique designed for detecting relative differences in copy number between test (sample to be interrogated) and reference genomes (Morozova et al., 2008). After hybridization, the amount of fluorescent signal can be measured and a ratio between test and reference DNA can be calculated based on the ratio of the signal intensity (Fig 2.1).

Figure 2.1 Illustration of general principles of aCGH. Reproduced from Shinawi et al. (2008).

(a) Patient or test and control or reference DNA are labeled with fluorescent dyes

and applied to the microarray. Test and reference DNA compete to attach, or

hybridize to the microarray.

(b) The microarray scanner measures the fluorescent signal from the slides and

these are converted into image files

fluorescent intensities

(d) Computer software analyses the data and generates a copy number plot.

2.7.1 Array description

The aCGH was performed using a customized design from the Agilent eArray

SurePrint G3 2x400K. This array contained a total of 420,288 distinct biological 60- mer oligonucleotide probes. High resolution was provided for chromosome 8, with median probe spacing of 1.97 kb, and in approximately 100 genes important in oncology, with a median probe spacing of 250 bp (Table 4.2). The rest of the human genome was interrogated with even whole-genome coverage of 253,502 probes, and a 11.8 kb median probe spacing. Probe sequences and gene annotations were based on the NCBI Build 37 of the human genome and UCSC version hg19 released in February 2009.

2.7.2 Sample preparation

The SurePrint G3 2x400K array must be hybridised with a mapped reference DNA.

For this purpose we used commercially available HAPMAP reference DNA from The

Coriell Institute for Medical Research, catalogue numbers NA12891 (European male) and NA12878 (European female), as test and reference DNA samples must be sex- matched. All DNA samples were assessed for concentration and purity as described in section 2.5.

2.7.3 Restriction digestion of test and reference DNA samples

DNA samples have to be cut into small pieces (approximately 0.3-2.0 kb), to be able to hybridize to the oligonucleotide probes present on the microarray. Restriction digestion was performed using the Genomic DNA enzymatic labelling kit (Agilent).

The manufacturer’s recommendations were to use a digestion master mix that contained 2 µg of each test genomic DNA and corresponding reference genomic

DNA sample, 5U Alu I and 5U Rsa I, 2.5 µL of 10x NE Buffer 4 (New England

BioLabs), 0.2 µL 10 µg/ µL BSA (New England BioLabs), and 1.3 µL water. Five µL master mix was added to 20 µL genomic DNA for a final volume of 25 µL for each reaction. After mixing, the samples were centrifuged for 1 min at 8,000xg

(Spectrafuge 24D Labnet). Tubes were incubated at 37˚C for 2.5 to 3 h to allow digestion of the double stranded DNA into smaller fragments.

2.7.4 Labeling of test and reference DNA samples

After the DNA was digested, test and reference DNA samples were labeled in order to be displayed when the test or reference DNA samples hybridize with oligonucleotide-probes on the microarray, using an Agilent Genomic DNA Labeling

Kit (Agilent). From the kit 5 µL random primers were added to each 25 µL sample, denatured at 95˚C for 5 min and incubated at 4˚C for 1 min on the thermocycler

(Eppendorf). Tubes were then removed from the thermocycler and briefly centrifuged. A labelling master mix was made using the reagents provided in the kit,

5 µL dNTPs, 10 µL Buffer, 3 µL 1 mM dUTP (Cy5 test samples, Cy3 for reference), 1

µL water and 1 µL KLENOW enzyme in a final reaction volume of 20 µL dispensed into each corresponding PCR tube. All test and reference samples were incubated at

37˚C for 16 - 20 hours overnight.

2.7.5 Clean-up of labelled DNA samples

The individually labeled test and reference DNA samples were purified using Amicon ultra columns (Millipore), by adding 25 µL test or reference DNA, and 410 µL 1xTE buffer, and centrifuging columns at 8,000xg for 10 min at RT. Supernatants were discarded and this step was repeated. Columns were inverted and placed in new collection tubes, centrifuged at 4,000xg for 1 min at RT, and then the eluted samples were retained.

2.7.6 Hybridization

Following purification, the appropriate Cy3-labeled test and Cy5-labeled reference samples were mixed together and combined with 60.5 µL 1x TE buffer (Promega),

25 µL 1.0 mg/mL Cot-1 DNA, 26 µL 10x Blocking Agent (Agilent), and 130 µL 2x

Hi-RPM Buffer (Agilent). Tubes were centrifuged at 8,500xg for 1 min at RT, then placed in the thermocycler (Eppendorf) for denaturation at 95 ˚C for 5 min and then incubated at 37˚C for 30 min.

The hybridization mixture was slowly dispensed onto a 2x400K microarray slide and assembled in an Agilent SureHyb chamber. The assembled slide chamber was placed in a rotator rack in a hybridization oven at 65˚C set to constant 20 rpm rotation for 18 -24 h.

2.7.7 Post-hybridization washing and scanning

A Coplin jar was filled with Oligo aCGH wash buffer 1 (Agilent) at RT. The hybridization chamber was disassembled and the array gasket was transferred to be completely immersed in the Coplin jar. Each microarray was removed and placed into the slide rack in Oligo aCGH wash buffer 1 at RT. This was incubated on a shaking incubator for 5 min at 37°C, with constant agitation at 60 rpm. Slides were then transferred into a slide rack containing Oligo aCGH wash buffer 2 (Agilent) pre- warmed at 37°C, and then shaken for 1 min at 37˚C at 60 rpm. The slide rack was removed and drained very carefully, and the edges of each slide were wiped to remove excess wash solution. After washing, each slide as loaded into individual slide holders and covered with a plastic coverslip. Microarray slides were loaded into

Agilent microarray scanner bundle G2565BA with 5-µm resolution immediately after washing, to minimize the impact of environmental oxidants on fluorescent signals.

The number of slots as selected and each slide was given a laboratory identification number. Slides were scanned into image (.tif) files and translated into log ratios, to allow the identification of the copy number aberrations.

2.7.8 Data analysis

Captured images were transformed to data with Feature Extraction version 10.10

(Agilent), and then imported into Agilent CytoGenomics Software version 2.7.2 for analysis, as well as Nexus copy number version 7.5 software (BioDiscovery Santa

Clara, CA).

Nexus Copy Number, v 7.5 (BioDiscovery) is a microarray analytical software for the analysis of genomic microarray datasets. The algorithms contained within Nexus perform the normalization, segmentation, and identification of corresponding copy number events from the raw data of all genomic files within a single project, using build 37 (human genome 19) of the genome as the common scaffold for all tumour profiles.

Agilent CytoGenomics v 2.7.2 was used primarily to visualise first results from aCGH, although it was not versatile enough for our analysis. Nexus v7.5, on the other hand, offered better user experience, and plotting functionality and frequency of copy number alteration across the genome not available in CytoGenomics.

In Nexus, for the rank segmentation algorithm we used the following parameters. A mean log2 ratio of 0.25 - 0.75 was classified as copy number gain, > 0.75 was classified as high-level copy number gain, < - 0.25 was classified as hemizygous loss, and < - 0.75 was classified as homozygous loss. The categories high copy gain and gain were combined, as well as the categories loss and homozygous loss. The threshold setting to make a positive call was at least 5 consecutive probes located in the gene or region of interest.

2.8 Quantitative real-time Polymerase Chain Reaction (qPCR)

Quantitative PCR was used to validate the results of MLPA and aCGH. Due to the small amount of DNA that remained after applied the other techniques, we were unable to perform qPCR for all samples and all genes, so a representative group was selected. Six genes were amplified, RECQL4, PTP4A3, KCNK9, EXT1, TPD52, and MYC, in 7 samples (one osteosarcoma, 2 Ewing sarcomas, 2 rhabdomyosarcomas, 2 neuroblastomas and 2 cell lines).

The sequences of PCR primers used in this study are shown in Table 2.4. VAV2, located on chromosome 9, has been previously validated for use in qPCR by Mosse et al. (2007) and since then has been similarly employed by other groups (Lau et al.,

2013) (Bower et al., 2012). VAV2 was also disomic in all tumours analysed by aCGH using Nexus copy number software. The relative copy number of target genes was therefore determined by normalization to VAV2 copy number using the relative standard curve method with serial dilutions (1:10, 1:20, 1:40, 1:80,1:160, 1:320) of

DNA from the AU 565 cell line.

Quantitative PCR was performed in a Rotor-Gene Q (Qiagen) using 1 x Quanti-

FAST SYBR Green master mix (Qiagen), 5 µL 1ng/µL DNA template and corresponding primers for each gene (Table 2.4), in a final reaction volume of 25 µL.

Each sample was measured in triplicate to compare gene copy number to that of the

VAV2 endogenous control that was performed to standardize the amount of DNA sample added to a reaction. The amount of target and endogenous reference is determined from the appropriate standard curve. Then, the target amount was divided by the endogenous control amount to obtain a normalized target value.

Again, the calibrator in our project was a genomic female DNA control (Promega) or

1x sample. Each of the normalized target values is divided by the calibrator normalized target value to generate the copy number.

The PCR conditions were as follows: one cycle of 5 min at 95˚C, followed by 35 cycles of (15 sec at 95°C, 30 sec at 61°C and 20 sec at 72°C) for all sets of primers.

Relative DNA copy numbers were calculated using Rotor-Gene 6 (Version 6.1) software (Qiagen) by normalizing raw data obtained for each gene against the VAV2 gene as a control. A single peak in the appropriate melting curve was used to determine that there was only one PCR product amplified per PCR primer set.

Table 2.4 Sequences of PCR primers used in this study.

PRIMER Gene Size Sense primer Antisense Primer No Name bp 5’è3’ 5’è3’ 1 VAV2 103 TGGGCATGACTGAAGATGAC ATCTGCCCTCACCTTCTCAA

2 TPD52 115 CAGTTTAGAGCCCAG GGA AA CGATCATCCAACGTA GCATG

3 MYC 162 AGGACTATCCTGCTGCCAAG TTAGCTCGTTCCTCCTCTGG CACTTGTGGAACAATGGTAGG CCTTAGAAAAGAGGGGAATAG 4 EXT1 175 A AAAC 5 KNCK9 150 TCTGTCCCTCATCGTCTGC TCCTCGCTGCTGATGTTGTA CCTCGATTCCATTATCATTTA 6 RECQL4 129 GACACCAGCTCTATCCATAC CTGC 7 PTP4A3 106 CATGTCTGCTTCCCTCCGTA CCTCACAGAACTTGGCCTTC

2.9 Statistical Analyses

Statistically significant differences between expected copy number change events for each category and actual 8p and 8q events for the same was assessed using

Fisher’s exact test for each type of tumour. A p-value of 0.05 or less was considered statistically significant. Fisher’s Exact test was performed using the 5 by 2 calculator from “SISA Five by 2 exact” (Uitenbroek, 1997), using the exact option.

Chapter 3

Analysis of chromosome 8 copy number in

paediatric solid tumours using MLPA

3.1 Introduction

Most normal individuals contain a diploid copy number for autosomal genes or, in terms of dosage, a normal expected dosage quotient (DQ) of 1.0 (Lake et al., 2012;

Wallace, 2003). In contrast, a distinctive characteristic of cancer is alterations in copy number from total or partial loss to the gain of one copy or many gene copies.

Moreover, the presence of deletions or duplications of specific genes has been associated with specific types of cancer (Lee et al., 2010). However, normal individuals can also present copy number variations that are considered benign and will not directly cause disease. Some known benign variants are summarized in the

Database of Genomic Variants (Macdonald et al., 2014).

Myllykangas et al. (2007) studied different models of gene amplification specificity and applied the cell type lineage model, which presumes that the evolution of molecular changes in cancer cells is a consequence of specific properties of the precursor stem cells. They concluded that gene amplification appears to be characteristic of solid tissue tumourigenesis rather than specific translocations, even though amplifications are also present in haematological malignancies but at lower frequencies (Myllykangas et al., 2007). The DNA amplification mechanism implies that two independent double-strand breaks have to occur to allow the intermediate

DNA to amplify (Myllykangas et al., 2007). The same research group defined the concept of amplification versus copy number gain. Amplification is the main mechanism in oncogene activation and refers to an increase of at least five copies of a DNA segment that is less than 20 megabases in length, whereas gain is a low

copy number increase (less than 5), which results from ploidy changes or unequal translocations (Myllykangas et al., 2007).

Chromosome 8 is one of the most frequently gained chromosomes in different cancer types (Myllykangas et al., 2007 & 2008). Using machine-learning techniques,

Myllykangas et al., aimed to classify human cancers based on their amplification patterns. Copy number amplification was studied across 4400 tumour cases representing 82 cancer types. Among other results, they found that amplifications were selected according to the anatomical locations and biological background of the cancers. In fact, 14 different types of cancers showed amplification in 4 different genes located on chromosome 8, demonstrating amplification hot spots at FGFR1

(8p11-12), MYC (8q24.1), PTK2 (8q24.3), EIF3S3 (8q24.11), and GATA4 (8p23.1- p22).

Different techniques have been designed to identify gene dosage or copy number changes in a quantitative fashion, such as Fluorescence In Situ Hybridization (FISH), array Comparative Genomic Hybridization (aCGH), quantitative PCR (qPCR) and

Multiplex Ligation-dependent Probe Amplification (MLPA), among others. MLPA has remained stable since 2010, indicating that this technique is still widely used, including cancer. According to HighWire http://highwire.stanford.edu the word

“MLPA” appeared in 81 scientific articles since the creation of the technique in 2002 until 2005. Searching for “MLPA” and “cancer” identified 51 papers in the same period. From 2006 until 2010, “MLPA” appeared in 669 articles, with 386 articles containing both “MLPA” and “cancer” (Fig 3.1). These data show the increasing

interest in MLPA within the scientific community, and highlight its relevance in cancer research, which is our focus in this study.

200

180

160

140

120

100

80 Number of papers 60

0 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 MLPA 7 13 17 44 72 104 133 175 185 173 187 180 171 MLPA & Cancer 4 9 10 28 41 76 75 87 107 100 106 104 88 Year of publication

Figure 3.1: Number of instances of the term “MLPA”, (blue bars) and “MLPA” (red bars) and “cancer” in the research literature per year, according to the HighWire search engine. Exact numbers of publications are also listed below the X axis.

MLPA is a valuable tool to support and complement classic pathology. It presents many advantages, including the ability to work with very small amounts of DNA (20 -

200 ng), being compatible with DNA extracted from paraffin-embedded as well as fresh-frozen material and having a low cost per reaction (Homig-Holzel et al., 2012).

Furthermore, results can be obtained within 24 hours, and it is also suitable for cancer risk screening, where large numbers of samples have to be analysed independently of an acute indication (Homig-Holzel et al., 2012).

In our project, this technique was used as a screening tool to assess chromosome 8 copy number changes in a cohort of 31 paediatric solid tumours. As the SALSA

MLPA KIT P014-A1 Chromosome 8 kit has not been tested before in this setting, we also determined copy number changes in 6 cell lines, some of which were expected to show defined copy number changes on chromosome 8.

3.2 Aims for chapter 3

1. To assess chromosome 8 copy number changes in a cohort of 31 paediatric

solid tumours and 6 cell lines, using the SALSA MLPA KIT P014-A1

Chromosome 8 kit (MRC Holland, Amsterdam).

2. To identify recurrent regions of copy number gain and loss, both within the

overall cohort and according to tumour type.

3.3 Results

MLPA is a multiplex PCR assay that in this project utilized 32 probes mapping to chromosome 8 loci (Fig 3.2, Table 2.3), and 9 reference probes mapping to other chromosomes. One of the differences between PCR and MLPA is the use of normal controls, usually around three, but depending on the number of total samples could be more (MRC-Holland, 2011). These controls provide not only a reference for performance of the MLPA reaction, but will also be used to normalize the copy number ratio, which is the quotient between the actual dosage per gene versus the expected dosage, calculated based on the normal controls. In this study, the overall dosage numbers were determined by the PeakScanner software v 1.0 (Applied

Biosystems).

As a normal control we used a male genomic DNA sample (Promega). The MLPA kit manufacturers recommended running the normal control in triplicate, within each experiment (MRC-Holland, 2011). MLPA copy number results are presented as means ± standard error, as calculated from 4 independent experiments. For MYC and RECQL4 genes, two probes were used in each gene (Table 2.3), and the results obtained for the two probe sets were averaged.

Copy numbers from raw data were converted to categories, for comparison with results from other techniques. A sample ratio above 1.3 thus considered a gain

(coded as 2), while sample ratios between 0.7 and 1.3 were classed as normal range or disomic (coded as 1). Sample ratios below 0.7 were classed as losses (coded as

0) (MRC-Holland, 2011).

Table 2.2 summarizes the paediatric tumour cohort used for both the MLPA and array Comparative Genomic Hybridization (aCGH) studies of Chapters 3 and 4.

85 UCSC Genome Browser on Human Mar. 2006 NCBI36/hg18 Assembly Chromosome 8: 1,484,199-145,739,091

Chr 8 (p23.3-q24.3) 1. DLGAP2 8p23.3 2. MFHAS 8p23.1 3. MSRA 8p23.1 4. GATA4 8p23.1 5. CTSB 8p22 6. TUSC3 8p22 7. ChGn 8p21.3 8. FGFR 8p11.2

9. PRKDC 8q11.21 10. MOS 8q11.23 11. CHD7 8q12.2 12. MYBL 8q13.1 13. NCOA2 8q13.3 14. TPD52 8q21.13 15. E2F5 8q21.2 16. RAD54B 8q22.1 17. ST7 8q22.3 18. EIF3E 8q23.1 19. EIF3H 8q24.11 20. EXT1 8q24.11 21. RNF139 8q24.1 22. MYC 8q24.21 23. MYC 8q24.21 24. DDEF1 8q24.22 25. KCNQ3 8q24.22 26. SLA 8q24.22 27. KHDRBS3 8q24.22 28. KCNK9 8q24.3 29. PTK2 8q24.3 30. PTP4A 8q24.3 31. RECQL4 8q24.3 32. RECQL4 8q24.3 Figure 3.2 Location of MLPA probe set on chromosome 8 using the UCSC human genome browser. Genes for which MLPA probes were included in the P014-A1 chromosome 8 kit are shown at left, with vertical lines showing the location of each gene with respect to the ideogram above. The vertical line located at the centromere, separates probes located 8pter from 8qter.

3.3.1 Use of known copy number changes in 6 breast cell lines, as positive controls for this paediatric tumour study.

Cell lines are a renewable and reliable resource, of proven utility in the study of any particular tissue type. Ideally, such cell lines will comprise a clonal cell population, of uniform genotype, ultimate for comparison with any uncharacteristic “test” cells, or tissue. Furthermore, breast cancer cell lines retain significant molecular features common to clinically recognised breast cancers (Riaz et al., 2013). In this project, we worked with 6 breast cell lines, with previously described patterns of chromosome 8 gains and losses (Table 3.1).

As expected (given the previously known findings in Table 3.1), 5 of these 6 breast cell lines showed copy number loss at 8p loci using the MLPA analysis performed within the present study (Table 3.2). One example (cell line SK-BR-3) is represented in figure 3.3, showing focal copy number loss from the region 8p23.3-p11.2. On the other hand, cell line HCC2157 was exceptional, as its TUSC3 copy number lies within the disomic range, MCF-10A was the only cell line found to have no loss within 8p, being disomic for all chromosome 8 genes tested (Table 3.2).

Also as expected (Table 3.1), copy number gains were frequently identified at chromosome 8q loci, among the 6 breast cell lines examined. Gains of MYC were detected in 5/6 (83%) cell lines, and gain of TPD52 in 4/6 (66.7%) (Table 3.3). Within

8q, two genes MOS and E2F5 were predominantly disomic, while KCNK9 showed copy number loss in 4/6 cell lines tested or 66.7% (Table 3.3).

Most reference loci were disomic in most cell lines. In particular CTTN and HIPK3 were disomic in 5/6 cell lines (Table 3.4), although other loci did show gain or loss,

(e.g. EHF, CD27, ABCC4, and PMP22). As reference probe in the kit are used to normalise the results for each sample, it is reasonable to expect some copy number changes at reference loci in breast cell lines, as for example AU565, MDA-MB-231,

MCF-7, and SK-BR-3 cell lines, have all been reported to show copy number loss at

PMP22 (Saito et al., 2009).

Table 3.1 Literature describing chromosome 8 showing copy number gain and/or loss in breast cell lines used in this study

Cell lines AU565 HCC2157 MCF-7 MCF-10A MDA-MB-231 SK-BR-3 Authors

Rummukainen et al. (2001) Loss 8p CGH Gain 8q23-qter Amplification MYC

Schouten et al. (2002) Amplification MYC MLPA (2 different probes) Loss 8p Loss 8p Jonsson et al. (2007) Gain entire Chr BAC-array CGH Gain 8q 8 Amplification MYC 8q23.3-q24.13

Rodriguez et al. (2007) Loss 8p BAC array -FISH Gain 8q (TPD52 & MYC) Kao et al. (2009) Loss 8p Loss 8p Disomic 8p Loss 8p Loss 8p aCGH Gain 8q Gain MYC Gain MYC Gain MYC Gain MYC

Marella et al. (2009) aCGH Gain entire Chr 8

Saito et al. (2009) Loss 8p Loss 8p Loss 8p Loss 8p aCGH Gain 8q Gain 8q Gain 8q Gain 8q Roslan et al. (2013) High copy Gain TPD52 Gain TPD52 High copy number QPCR number TPD52 TPD52 Deletion of entire Chr -8 Rondon-Lagos et al. (2014) +8, der(8) t(8;21)(p23;?) G banding - M-FISH t(8;21)(q24;?), dup(8); der(8)t(8;15)(p11;?) t(8;17)(q24;?)

Table 3.2 Chromosome 8p copy number status in 6 breast cell lines studied in the current project

Categories: 2=Gain; 1=Disomic; 0=Loss

Samples DLGAP2 MFHAS MSRA GATA4 CTSB TUSC3 ChGn FGFR1 AU 565 0 0 0 0 0 0 0 0 HCC2157 0 0 0 0 0 1 0 0 MCF-7 0 0 0 0 0 0 0 0 MCF-10A 1 1 1 1 1 1 1 1 MDA-MB-231 0 0 0 0 0 0 0 0 SK-BR-3 0 0 0 0 0 0 0 0

DLGAP2 MFHAS MSRA GATA4 CTSB TUSC3 ChGn FGFR1

Gain 0(0%) 0(0%) 0(0%) 0(0%) 0(0%) 0(0%) 0(0%) 0(0%) Disomy 1(17%) 1(17%) 1(17%) 1(17%) 1(17%) 2(33%) 1(17%) 1(17%) Loss 5(83%) 5(83%) 5(83%) 5(83%) 5(83%) 4(66.6%) 5(83%) 5(83%)

4.5

3.5

2.5

Proberatio 2

1.5

0.5

0 CTSB 8p22CTSB ST7 8q22.3ST7 E2F58q21.2 SLA 8q24.22 SLA PTK2 8q24.3PTK2 ChGn 8p21.3 FGFR 8p11.2 FGFR 8p11.2 MYBL 8q13.1 MYBL TUSC38p22 CHD7 8q12.2 MYC8q24.21 MYC8q24.21 EIF3E 8q23.1 EIF3E MOS 8q11.23 8q11.23 MOS MSRA 8p23.1 MSRA NCOA 8q13.3 NCOA EXT1 8q24.11 8q24.11 EXT1 GATA4 8p23.1 GATA4 PTP4A 8q24.3 PTP4A RAD54 8q21.3 EIF3H 8q24.11 8q24.11 EIF3H KCNK9 8q24.3 MFHAS 8p23.1MFHAS TPD528q21.13 DDEF1 8q24.21 PRKDC 8q11.21 PRKDC 8q11.21 DLGAP2 8p23.3 KCNQ3 8q24.22 RECQL4 8q24.3 RECQL4 8q24.3 RNF139 8q24.13 KHDRBS3 8q24.22 Ref. probe1463Ref. 17p12 Ref. probe 0980 11p12 probeRef. 0980 11p12 probeRef. 0976 11p13 probeRef. 0434 11q13 Ref. probeRef. 0678 12p13 probeRef. 0798 13q32 probeRef. 0871 13q34 Ref. probe 1160 3p25.3 probeRef. 1160 Ref. probeRef. 0797 5q31.1

Figure 3.3 MLPA probe ratios obtained for the SK-BR-3 cell line studied in the current project. Results shown are mean probe ratio ± SE from 4 independent experiments. On the

X-axis, brown columns represent genes located on chromosome 8p, blue columns represent genes located on chromosome 8q, and red columns represent reference probes.

Probe ratios are shown on the Y-axis. Probe ratios between 0.7 and 1.3 were considered in the normal range or disomic, indicated on the graph as a rectangular dashed region. Probe ratios below 0.7 were considered copy number loss, and probe ratios above 1.3 were considered copy number gain.

Table 3.3 Chromosome 8q copy number status in 6 breast cell lines studied in the current project

Categories: 2=Gain; 1=Disomic; 0=Loss

RAD54 Samples PRKDC MOS CHD7 MYBL NCOA2 TPD52 E2F5 B AU565 0 1 1 1 1 2 1 0 HCC2157 2 2 2 2 2 2 2 2 MCF-7 1 1 1 1 0 1 1 2 MCF-10A 1 1 1 1 1 1 1 1 MDA-MB-231 2 1 2 2 1 2 1 1 SK-BR-3 1 1 1 1 1 2 1 0

RAD54 PRKDC MOS CHD7 MYBL NCOA2 TPD52 E2F5 B

Gain 2(33.3%) 1(16.7%) 2(33.3%) 2(33.3%) 1(16.7%) 4(66.7%) 1(16.7%) 2(33.3%) Disomy 3(50.0%) 5(83.3%) 4(66.7%) 4(66.7%) 4(66.7%) 2(33.3%) 5(83.3%) 2(33.3%) Loss 1(16.7%) 0(0.0%) 0(0.0%) 0(0.0%) 1(16.7%) 0(0.0%) 0(0.0%) 2(33.3%)

Categories: 2=Gain; 1=Disomic; 0=Loss

Samples ST7 EIF3E EIF3H EXT1 RNF139 MYC DDEF1

AU565 0 0 2 2 2 2 1 HCC2157 1 2 2 1 1 2 1 MCF-7 2 2 1 1 2 2 2 MCF-10A 1 1 1 1 1 2 1 MDA-MB-231 1 1 1 1 0 1 2 SK-BR-3 0 0 2 2 2 2 1

ST7 EIF3E EIF3H EXT1 RNF139 MYC DDEF1

Gain 1(16.7%) 2(33.3%) 3(50.0%) 2(33.3%) 3(50.0%) 5(83.3%) 2(33.3%) Disomy 3(50.0%) 2(33.3%) 3(50.0%) 4(66.7%) 2(33.3%) 1(16.7%) 4(66.7%) Loss 2(33.3%) 2(33.3%) 0(0%) 0(0%) 1(16.7%) 0(0%) 0(0%)

Table 3.3 Chromosome 8q copy number status in 6 breast cell lines studied in the current project (cont.)

Categories: 2=Gain; 1=Disomic; 0=Loss

Samples KCNQ3 SLA KHDRBS3 KCNK9 PTK2 PTP4A RECQL4 AU565 0 0 0 0 0 0 1 HCC2157 1 1 1 1 1 1 2 MCF-7 1 1 1 0 1 1 1 MCF-10A 1 2 1 1 1 1 1 MDA-MB-231 1 1 1 0 1 1 1 SK-BR-3 0 1 0 0 0 0 0

KCNQ3 SLA KHDRBS3 KCNK9 PTK2 PTP4A RECQL4

Gain 0(0%) 1(16.7%) 0(0%) 0(0%) 0(0%) 0(0%) 1(16.7%) Disomy 4(66.7%) 4(66.7%) 4(66.7%) 2(33.3%) 4(66.7%) 4(66.7%) 4(66.7%) Loss 2(33.3%) 1(16.7%) 2(33.3%) 4(66.7%) 2(33.3%) 2(33.3%) 1(16.7%)

Table 3.4 Copy number status at 9 reference genes in 6 breast cell lines studied in the current project

Categories: 2=Gain; 1=Disomic; 0=Loss

Ref 3p Ref 5q Ref 11p Ref 11p Ref 11q Ref 12p Ref 13q Ref 13q Ref 17p Gene VHL IL4 EHF HIPK3 CTTN CD27 ABCC4 ARHGEF7 PMP22

Samples AU565 0 0 0 1 1 0 0 1 0

HCC2157 1 1 1 0 1 0 0 0 0

MCF-7 1 1 0 1 1 1 0 0 1

MCF-10A 1 2 1 1 1 2 1 1 1

MDA-MB-231 1 1 1 1 1 1 1 1 2

SK-BR-3 0 0 0 1 0 0 1 1 0

Ref 3p Ref 5q Ref 11p Ref 11p Ref 11q Ref 12p Ref 13q Ref 13q Ref 17p Gene VHL IL4 EHF HIPK3 CTTN CD27 ABCC4 ARHGEF7 PMP22

Gain 0(0%) 1(16.7%) 0(0%) 0(0%) 0(0%) 1(16.7%) 0(0%) 0(0%) 1(16.7%) Disomy 4(66.7%) 3(50.0%) 3(50.0%) 5(83.3%) 5(83.3%) 2(33.3%) 3(50.0%) 4(66.7%) 2(33.3%) Loss 2(33.3%) 22(33.3%) 3(50.0%) 1(16.7%) 1(16.7%) 3(50.0%) 3(50.0%) 2(33.3%) 3(50.0%)

3.3.2 Copy number changes on chromosome 8p in the paediatric solid tumour cohort

Among the four classes of paediatric tumour we have tested (Table 3.5), osteosarcomas show more frequent chromosome 8p changes. Two of the five osteosarcoma samples show loss for at least half of the loci examined, while one shows copy number gain at 3 loci on 8p (Tables 3.5, 3.6). By contrast, most Ewing sarcomas and rhabdomyosarcomas show no evidence of copy number gain or loss at any of the 8p loci tested (Tables 3.5, 3.6). The neuroblastoma cohort showed frequent copy number loss at ChGn (5/10, or 50% cases) followed by MSRA (4/10,

40% cases) (Tables 3.5, 3.6).

Table 3.5 Chromosome 8p copy number status in solid tumour cohort

Categories: 2=Gain; 1=Disomic; 0=Loss

Gene DLGAP2 MFHAS MSRA GATA4 CTSB TUSC3 ChGn FGFR1 Samples F9Oa 1 1 1 1 1 1 1 1 F10O 1 1 1 1 1 1 1 1 M11O 1 1 2 1 1 2 2 1 F12O 0 1 0 0 0 1 0 0 M46O 0 0 0 0 1 1 1 1 M13E 1 1 1 1 1 1 1 1 M14E 1 1 1 1 1 2 1 1 F15E 1 1 1 1 1 1 1 1 M16E 1 1 1 1 1 1 1 1 F17E 1 1 1 1 1 2 1 1 F18E 1 2 1 1 0 1 1 1 M22E 1 1 1 1 1 2 1 1 F24E 1 1 1 1 1 1 1 1 M25R 1 1 1 1 1 1 1 1 F26R 1 1 1 1 1 1 1 1 M27R 1 1 1 1 1 1 1 1 F28R 1 1 1 1 1 2 1 1 F29R 1 1 1 1 1 1 1 1 M31R 1 1 1 1 1 2 1 1 M32R 1 1 1 1 1 1 1 1 M35R 1 1 1 1 1 2 1 1 M36N 1 1 1 1 1 1 0 1 F37N 1 2 0 1 2 1 0 2 F38N 2 1 0 1 1 1 0 1 M39N 1 2 0 1 1 1 0 1 F40N 1 1 1 1 1 1 1 1 M41N 1 1 1 1 1 2 1 1 F42N 1 1 1 1 1 2 1 1 M43N 1 1 1 1 1 1 1 1 F44N 1 1 1 1 1 1 1 1 F45N 2 0 0 0 2 2 0 0

DLGAP2 MFHAS MSRA GATA4 CTSB TUSC3 ChGn FGFR1

Gain 2 (6.5%) 3 (9.7%) 1 (3.2%) 0 (0%) 2 (6.5%) 10(32.3%) 1 (3.2%) 1 (3.2%) Disomy 27(87.1%) 26(83.9%) 24(77.4%) 28(90.3%) 27(87.1%) 21(67.7%) 24(77.4%) 28(90.3%) Loss 2 (6.4%) 2 (6.4%) 6 (19.4%) 3 (9.7%) 2 (6.4%) 0 (0%) 6 (19.4%) 2 (6.5%) a Sample identifiers ending in “O” are osteosarcomas, “E” are Ewing sarcomas, “R” are rhabdomyosarcomas, “N” are neuroblastomas.

Table 3.6 Chromosome 8p copy number status by gene and solid tumour type

Osteosarcoma n=5

Gene DLGAP2 MFHAS MSRA GATA4 CTSB TUSC3 ChGn FGFR1 Samples Gain 0(0%) 0(0%) 1(20.0%) 0(0%) 0(0%) 1(20.0%) 1(20.0%) 0(0%) Disomy 3(60.0%) 4(80.0%) 2(40.0%) 3(60.0%) 4(80.0%) 4(80.0%) 3(60.0%) 4(80.0%) Loss 2(40.0%) 1(20.0%) 2(40.0%) 2(40.0%) 1(20.0%) 0(0%) 1(20.0%) 1(20.0%)

Ewing sarcoma n=8

Gene DLGAP2 MFHAS MSRA GATA4 CTSB TUSC3 ChGn FGFR1 Samples Gain 0(0%) 1(12.5%) 0(0%) 0(0%) 0(0%) 3(37.5%) 0(0%) 0(0%) Disomy 8(100.0%) 7(87.5%) 8(100.0%) 8(100.0%) 7(87.5%) 5(62.5%) 8(100.0%) 8(100.0%) Loss 0(0%) 0(0%) 0(0%) 0(0%) 1(12.5%) 0(0%) 0(0%) 0(0%)

Rhabdomyosarcoma n=8

Gene DLGAP2 MFHAS MSRA GATA4 CTSB TUSC3 ChGn FGFR1 Samples Gain 0(0%) 0(0%) 0(0%) 0(0%) 0(0%) 3(37.5%) 0(0%) 0(0%) Disomy 8(100.0%) 8(100.0%) 8(100.0%) 8(100.0%) 8(100.0%) 5(62.5%) 8(100.0%) 8(100.0%) Loss 0(0%) 0(0%) 0(0%) 0(0%) 0(0%) 0(0%) 0(0%) 0(0%)

Neuroblastoma n=10

Gene DLGAP2 MFHAS MSRA GATA4 CTSB TUSC3 ChGn FGFR1 Samples Gain 2(20.0%) 2(20.0%) 0(0.0%) 0(0.0%) 2(20.0%) 3(30.0%) 0(0.0%) 1(10.0%) Disomy 8(80.0%) 7(70.0%) 6(60.0%) 9(90.0%) 8(80.0%) 7(70.0%) 5(50.0%) 8(80.0%) Loss 0(0.0%) 1(10.0%) 4(40.0%) 1(10.0%) 0(0.0%) 0(0.0%) 5(50.0%) 1(10.0%)

3.3.3 Copy number changes on chromosome 8q and at reference genes in the paediatric tumour cohort.

In the osteosarcoma group, six chromosome 8q genes were gained in 2/5 (40%) of cases, these being MYBL, TPD52, EIF3E, EIF3H, EXT1 and SLA. Only KCNK9 showed copy number loss in 2/5 (40%) cases (Tables 3.7, 3.11). Ewing sarcoma samples were predominantly disomic at chromosome 8q loci. Gains were found in

EXT1 and SLA in 2/8 (25%) cases, and copy number losses were measured at

KCNK9 2/8 (25%) cases (Tables 3.8, 3.11). Rhabdomyosarcoma samples showed gain at TPD52 (3/8, 37.5% cases), RAD54 (2/8, 25% cases), whereas losses were reported at EXT1 (3/8, 37.5% cases) and PTP4A and KCNK9 (2/8, 25% cases)

(Tables 3.9, 3.11). In neuroblastoma samples losses were more frequent, RECQL4 was gained in 4/10 (40%) cases, whereas TPD52 was gained 3/10 (30%) cases,

KHDRBS3 locus showed reduced copy number in 3/10 (30%) cases, and RAD54,

ST7, EIF3E, KCNQ3, KCNK9 and PTP4A all showed copy number loss in 2/10

(20%) cases (Tables 3.10, 3.11). Overall, TPD52 was predicted to be gained most frequently, in 9/31 (29%) of the overall cohort followed by SLA in 6/31 (19.4%) cases (Table 3.11).

Fisher's exact test was used to assess differences in the numbers per copy number category between 8p loci and 8q loci for each type of tumour. P-values <0.05 were considered statistically significant. Only for osteosarcoma samples did this statistical test show significant differences between 8p and 8q loci. Here, the count of copy number losses at 8p was significantly higher than expected; conversely, the number of copy number gains at 8q was higher than expected (Appendix 1). For the other

three types of tumours, Fisher's exact test did not reach statistical significance

(Appendix 1).

Reference probes showed disomic copy number in most samples examined (18/31,

58% of cases, Table 3.12). The reference probes for ARHGEF7 on chromosome

13q24 showed disomic copy number in 31/31 (100%) cases of the paediatric solid tumour cohort. However some cases showed obvious copy number changes at reference probe loci (Fig 3.4).

3.3.4 Additional MLPA results

Some discrepant results were observed within the 4 MLPA runs for both cell lines and the solid tumour cohort. Discrepant results were obtained at a number of loci, including at EIF3E located at 8q23.1 and EXT1 located at 8q24.11 (Figs 3.5 and

3.6).

Table 3.7 Chromosome 8q copy number changes in osteosarcomas n=5

Gene Gain Disomy Loss

PRKDC 0(0.0%) 5(100.0%) 0(0.0%) MOS 0(0.0%) 4(80.0%) 1(20.0%) CHD7 1(20.0%) 4(80.0%) 0(0.0%) MYBL 2(40.0%) 3(60.0%) 0(0.0%) NCOA2 1(20.0%) 4(80.0%) 0(0.0%) TPD52 2(40.0%) 2(40.0%) 1(20.0%) E2F5 0(0.0%) 4(80.0%) 1(20.0%) RAD54B 1(20.0%) 4(80.0%) 0(0.0%) ST7 1(20.0%) 3(60.0%) 1(20.0%) EIF3E 2(40.0%) 3(60.0%) 0(0.0%) EIF3H 2(40.0%) 3(60.0%) 0(0.0%) EXT1 2(40.0%) 3(60.0%) 0(0.0%) RNF139 1(20.0%) 4(80.0%) 0(0.0%) MYC 1(20.0%) 4(80.0%) 0(0.0%) DDEF1 0(0.0%) 5(100.0%) 0(0.0%) KCNQ3 0(0.0%) 5(100.0%) 0(0.0%) SLA 2(40.0%) 3(60.0%) 0(0.0%) KHDRBS3 1(20.0%) 4(80.0%) 0(0.0%) KCNK9 0(0.0%) 3(60.0%) 2(40.0%) PTK2 0(0.0%) 4(80.0%) 1(20.0%) PTP4A 0(0.0%) 4(80.0%) 1(20.0%) RECQL4 0(0.0%) 4(80.0%) 1(20.0%)

100

Table 3.8 Chromosome 8q copy number changes in Ewing sarcomas n=8

Gene Gain Disomy Loss

PRKDC 0(0.0%) 8(100.0%) 0(0.0%) MOS 1(12.5%) 7(87.5%) 0(0.0%) CHD7 1(12.5%) 7(87.5%) 0(0.0%) MYBL 1(12.5%) 7(87.5%) 0(0.0%) NCOA2 0(0.0%) 8(100.0%) 0(0.0%) TPD52 3(37.5%) 5(62.5%) 0(0.0%) E2F5 0(0.0%) 8(100.0%) 0(0.0%) RAD54B 2(25.0%) 6(75.0%) 0(0.0%) ST7 0(0.0%) 8(100.0%) 0(0.0%) EIF3E 1(12.5%) 6(75.0%) 1(12.5%) EIF3H 0(0.0%) 8(100.0%) 0(0.0%) EXT1 2(25.0%) 6(75.0%) 0(0.0%) RNF139 0(0.0%) 8(100.0%) 0(0.0%) MYC 0(0.0%) 8(100.0%) 0(0.0%) DDEF1 0(0.0%) 8(100.0%) 0(0.0%) KCNQ3 0(0.0%) 8(100.0%) 0(0.0%) SLA 2(25.0%) 6(75.0%) 0(0.0%) KHDRBS3 0(0.0%) 8(100.0%) 0(0.0%) KCNK9 0(0.0%) 6(75.0%) 2(25.0%) PTK2 0(0.0%) 8(100.0%) 0(0.0%) PTP4A 0(0.0%) 8(100.0%) 0(0.0%) RECQL4 0(0.0%) 8(100.0%) 0(0.0%)

101

Table 3.9 Chromosome 8q copy number changes in rhabdomyosarcomas n=8

Gene Gain Disomy Loss

PRKDC 0(0.0%) 8(100.0%) 0(0.0%) MOS 1(12.5%) 7(87.5%) 0(0.0%) CHD7 1(12.5%) 7(87.5%) 0(0.0%) MYBL 1(12.5%) 7(87.5%) 0(0.0%) NCOA2 0(0.0%) 8(100.0%) 0(0.0%) TPD52 3(37.5%) 5(62.5%) 0(0.0%) E2F5 0(0.0%) 8(100.0%) 0(0.0%) RAD54B 2(25.0%) 6(75.0%) 0(0.0%) ST7 0(0.0%) 8(100.0%) 0(0.0%) EIF3E 0(0.0%) 8(100.0%) 0(0.0%) EIF3H 0(0.0%) 8(100.0%) 0(0.0%) EXT1 0(0.0%) 5(62.5%) 3(37.5%) RNF139 0(0.0%) 8(100.0%) 0(0.0%) MYC 1(12.5%) 7(87.5%) 0(0.0%) DDEF1 0(0.0%) 8(100.0%) 0(0.0%) KCNQ3 0(0.0%) 8(100.0%) 0(0.0%) SLA 1(12.5%) 7(87.5%) 0(0.0%) KHDRBS3 0(0.0%) 8(100.0%) 0(0.0%) KCNK9 0(0.0%) 6(75.0%) 2(25.0%) PTK2 0(0.0%) 8(100.0%) 0(0.0%) PTP4A 0(0.0%) 6(75.0%) 2(25.0%) RECQL4 0(0.0%) 8(100.0%) 0(0.0%)

102

Table 3.10 Chromosome 8q copy number changes in neuroblastomas n=10

Gene Gain Disomy Loss

PRKDC 0(0.0%) 7(70.0%) 3(30.0%) MOS 1(10.0%) 9(90.0%) 0(0.0%) CHD7 0(0.0%) 9(90.0%) 1(10.0%) MYBL 1(10.0%) 6(60.0%) 3(30.0%) NCOA2 1(10.0%) 9(90.0%) 0(0.0%) TPD52 3(30.0%) 5(50.0%) 2(20.0%) E2F5 0(0.0%) 8(80.0%) 2(20.0%) RAD54B 0(0.0%) 9(90.0%) 1(10.0%) ST7 0(0.0%) 7(70.0%) 3(30.0%) EIF3E 1(10.0%) 9(90.0%) 0(0.0%) EIF3H 0(0.0%) 6(60.0%) 4(40.0%) EXT1 0(0.0%) 8(80.0%) 2(20.0%) RNF139 0(0.0%) 8(80.0%) 2(20.0%) MYC 0(0.0%) 10(100.0)% 0(0.0%) DDEF1 0(0.0%) 9(90.0%) 1(10.0%) KCNQ3 0(0.0%) 8(80.0%) 2(20.0%) SLA 1(10.0%) 9(90.0%) 0(0.0%) KHDRBS3 0(0.0%) 7(70.0%) 3(30.0%) KCNK9 0(0.0%) 8(80.0%) 2(20.0%) PTK2 0(0.0%) 9(90.0%) 1(10.0%) PTP4A 0(0.0%) 8(80.0%) 2(20.0%) RECQL4 4(40.0%) 5(50.0%) 1(10.0%)

103

Table 3.11 Chromosome 8q copy number status in tumour cohort

Categories: 2=Gain; 1=Disomic; 0=Loss

Gene PRKDC MOS CHD7 MYBL NCOA2 TPD52 E2F5 RAD54B Samples F9Oa 1 1 1 1 1 2 1 1 F10O 1 0 1 2 1 2 1 1 M11O 1 1 1 1 1 1 1 1 F12O 1 1 2 2 2 0 0 1 M46O 1 1 1 1 1 1 1 2 M13E 1 1 1 1 1 1 1 1 M14E 1 1 1 1 1 1 1 1 F15E 1 1 1 1 1 1 1 1 M16E 1 1 1 1 1 1 1 1 F17E 1 1 2 1 1 2 1 1 F18E 1 1 1 1 1 1 1 1 M22E 1 1 1 1 1 1 1 1 F24E 1 1 1 1 1 1 1 1 M25R 1 1 1 1 1 1 1 1 F26R 1 1 1 1 1 1 1 1 M27R 1 1 1 1 1 1 1 1 F28R 1 1 1 1 1 1 1 1 F29R 1 1 1 1 1 2 1 2 M31R 1 1 2 1 1 2 1 1 M32R 1 1 1 1 1 1 1 1 M35R 1 2 1 2 1 2 1 2 M36N 1 1 1 1 1 1 1 1 F37N 0 1 1 0 1 0 0 1 F38N 1 1 1 0 1 1 1 1 M39N 0 1 0 0 1 0 0 0 F40N 1 1 1 1 1 1 1 1 M41N 1 1 1 1 1 2 1 1 F42N 1 1 1 1 1 2 1 1 M43N 1 1 1 1 1 1 1 1 F44N 1 1 1 1 1 1 1 1 F45N 0 2 1 2 0 2 1 1

PRKDC MOS CHD7 MYBL NCOA2 TPD52 E2F5 RAD54B

Gain 0(0%) 2(6.5%) 3(9.7%) 4(12.9%) 1(3.2%) 9(29%) 0(0%) 3(9.7%) Disomy 28(90.3%) 28(90.3%) 27(87.1%) 24(77.4%) 27(87.1%) 19(61.3%) 28(90.3%) 27(87.1%) Loss 3(9.7%) 1(3.2%) 1(3.2%) 3(9.7%) 1(3.2%) 3(9.7%) 3(9.7%) 1(3.2%) a Sample identifiers ending in “O” are osteosarcomas, “E” are Ewing sarcomas, “R” are rhabdomyosarcomas, “N” are neuroblastomas.

104

Table 3.11 Chromosome 8q Copy number status in tumour cohort (cont.)

Categories: 2=Gain; 1=Disomic; 0=Loss

Gene ST7 EIF3E EIF3H EXT1 RNF139 MYC DDEF1 Samples F9Oa 1 1 1 1 1 1 1 F10O 1 1 1 1 1 1 1 M11O 1 1 1 1 1 1 1 F12O 0 2 2 2 2 2 1 M46O 2 2 2 2 1 1 1 M13E 1 1 1 1 1 1 1 M14E 1 2 1 2 1 1 1 F15E 1 1 1 1 1 1 1 M16E 1 1 1 1 1 1 1 F17E 1 1 1 1 1 1 1 F18E 1 0 1 1 1 1 1 M22E 1 1 1 1 1 1 1 F24E 1 1 1 2 1 1 1 M25R 1 1 1 1 1 1 1 F26R 1 1 1 1 1 1 1 M27R 1 1 1 1 1 1 1 F28R 1 1 1 1 1 1 1 F29R 1 1 1 0 1 2 1 M31R 1 1 1 1 1 1 1 M32R 1 1 1 0 1 1 1 M35R 1 1 1 0 1 1 1 M36N 1 1 1 1 1 1 1 F37N 0 1 0 0 0 1 1 F38N 1 1 0 1 1 1 1 M39N 0 1 0 1 1 1 1 F40N 1 1 1 1 1 1 1 M41N 1 1 1 1 1 1 1 F42N 1 1 1 1 1 1 1 M43N 1 1 1 1 1 1 1 F44N 1 1 1 1 1 1 1 F45N 0 2 0 0 0 1 0

ST7 EIF3E EIF3H EXT1 RNF13 MYC DDEF1

Gain 1(3.2%) 4(12.9%) 2(6.5%) 4(12.9%) 1(3.2%) 2(6.5%) 0(0%) Disomy 26(83.9%) 26(83.9%) 25(80.6%) 22(71.0%) 28(90.3%) 29(93.5%) 30(96.8%) Loss 4(12.9%) 1(3.2%) 4(12.9%) 5(16.1%) 2(6.5%) 0(0%) 1(3.2%) a Sample identifiers ending in “O” are osteosarcomas, “E” are Ewing sarcomas, “R” are rhabdomyosarcomas, “N” are neuroblastomas.

105

Table 3.11 Chromosome 8q Copy number status in tumour cohort (cont.)

Categories: 2=Gain; 1=Disomic; 0=Loss

Gene KCNQ3 SLA KHDRBS3 KCNK9 PTK2 PTP4A RECQL4 Samples F9Oa 1 1 1 0 1 1 1 F10O 1 2 1 1 1 1 1 M11O 1 1 1 1 1 1 0 F12O 1 1 1 0 0 0 1 M46O 1 2 2 1 1 1 1 M13E 1 1 1 1 1 1 1 M14E 1 1 1 1 1 1 1 F15E 1 1 1 1 1 1 1 M16E 1 1 1 1 1 1 1 F17E 1 2 1 0 1 1 1 F18E 1 1 1 1 1 1 1 M22E 1 2 1 0 1 1 1 F24E 1 1 1 1 1 1 1 M25R 1 1 1 1 1 1 1 F26R 1 1 1 1 1 1 1 M27R 1 1 1 1 1 1 1 F28R 1 1 1 1 1 1 1 F29R 1 1 1 0 1 0 1 M31R 1 1 1 1 1 1 1 M32R 1 1 1 1 1 1 1 M35R 1 2 1 0 1 0 1 M36N 1 1 1 1 1 1 1 F37N 0 1 0 1 1 1 2 F38N 1 1 1 1 1 1 2 M39N 1 1 0 1 1 1 2 F40N 1 1 1 1 1 0 1 M41N 1 1 1 1 1 1 2 F42N 1 1 1 0 1 1 1 M43N 1 1 1 1 1 1 1 F44N 1 1 1 1 1 1 1 F45N 0 2 0 0 0 0 0

KCNQ3 SLA KHDR KCNK9 PTK2 PTP4A RECQL4

Gain 0(0%) 6(19.4%) 1(3.2%) 0(0%) 0(0%) 0(0%) 4(12.9%) Disomy 29(93.5%) 25(80.6%) 27(87.1%) 23(74.2%) 29(93.5%) 26(83.9%) 25(80.6%) Loss 2(6.5%) 0(0%) 3(9.7%) 8(25.8%) 2(6.5%) 5(16.1%) 2(6.5%) a Sample identifiers ending in “O” are osteosarcomas, “E” are Ewing sarcomas, “R” are rhabdomyosarcomas, “N” are neuroblastomas.

106

Table 3.12 Copy number status at 9 reference genes in 31 tumour samples

Categories: 2=Gain; 1=Disomic; 0=Loss

Ref 3p Ref 5q Ref 11p Ref 11p Ref 11q Ref 12p Ref 13q Ref 13q Ref 17p

Gene VHL IL4 EHF HIPK3 CTTN CD27 ABCC4 ARHGEF7 PMP22

Sample F9Oa 0 1 1 1 1 0 0 1 2 F10O 1 1 1 1 1 0 0 1 2 M11O 0 1 1 1 0 1 0 1 1 F12O 0 0 0 1 1 0 1 1 0 M46O 0 1 0 1 1 1 1 1 1 M13E 1 1 1 1 1 1 1 1 1 M14E 0 1 1 1 1 0 1 1 1 F15E 1 1 0 1 1 1 1 1 1 M16E 1 1 0 1 1 0 1 1 1 F17E 1 1 1 2 1 0 1 1 1 F18E 1 1 1 1 1 1 1 1 1 M22E 1 2 1 2 1 0 1 1 1 F24E 1 1 1 1 1 1 1 1 1 M25R 1 1 1 1 1 0 1 1 1 F26R 1 1 0 1 1 1 1 1 0 M27R 1 2 1 2 1 1 1 1 1 F28R 1 2 1 1 1 1 0 1 1 F29R 0 0 2 1 0 1 1 1 0 M31R 0 1 1 1 0 1 1 1 1 M32R 1 1 1 1 1 1 1 1 1 M35R 1 2 1 2 1 0 1 1 1 M36N 1 1 0 1 1 1 1 1 2 F37N 1 1 1 1 2 2 1 1 2 F38N 0 1 0 1 2 1 1 1 1 M39N 1 1 0 1 2 2 1 1 2 F40N 1 1 1 1 1 1 1 1 1 M41N 0 1 0 1 1 1 1 1 0 F42N 1 2 1 2 1 0 1 1 1 M43N 1 1 1 1 2 1 1 1 1 F44N 1 1 0 0 0 1 1 1 1 F45N 0 2 0 2 1 0 2 1 0

Ref 3p Ref 5q Ref 11p Ref 11p Ref 11q Ref 12p Ref 13q Ref 13q Ref 17p Gene VHL IL4 EHF HIPK3 CTTN CD27 ABCC4 ARHGEF7 PMP22

Gain (%) 0(0) 6(19.4) 1(3.2) 6(19.4) 4(12.9) 2(6.4) 1(3.2) 0(0) 5(16.1) Disomy (%) 21(67.7) 23(74.19) 19(61.3) 24(77.4) 23(74.19) 18(58.1) 26(83.9) 31(100.0) 21(67.7) Loss (%) 10(32.3) 2(6.4) 11(35.5) 1(3.2) 4(12.9) 11(35.4) 4(12.9) 0(0) 5(16.1) a Sample identifiers ending in “O” are osteosarcomas, “E” are Ewing sarcomas, “R” are rhabdomyosarcomas, “N” are neuroblastomas.

107

A. 5 4.5 4 3.5 3 2.5 I 2 Probe ratio 1.5 II

1 III 0.5 IV 0 CTSB 8p22CTSB ST7 8q22.3ST7 E2F58q21.2 SLA 8q24.22 SLA PTK2 8q24.3PTK2 ChGn 8p21.3 FGFR 8p11.2 FGFR 8p11.2 MYBL 8q13.1 MYBL TUSC38p22 CHD7 8q12.2 MOS 8q11.23 8q11.23 MOS MYC8q24.21 MYC8q24.21 EIF3E 8q23.1 EIF3E MSRA 8p23.1 MSRA 8q13.3 NCOA EXT1 8q24.11 8q24.11 EXT1 PTP4A 8q24.3 PTP4A GATA4 8p23.1 GATA4 RAD54 8q21.3 EIF3H 8q24.11 8q24.11 EIF3H KCNK9 8q24.3 MFHAS 8p23.1MFHAS TPD528q21.13 DDEF1 8q24.21 DLGAP2 8p23.3 PRKDC 8q11.21 PRKDC 8q11.21 KCNQ3 8q24.22 RECQL4 8q24.3 RECQL4 8q24.3 RNF139 8q24.13 KHDRBS3 8q24.22 Ref. probe1463Ref. 17p12 Ref. probe 0980 11p12 probeRef. 0980 11p12 probeRef. 0976 11p13 probeRef. 0434 11q13 Ref. probeRef. 0678 12p13 probeRef. 0798 13q32 probeRef. 0871 13q34 Ref. probe 1160 3p25.3 probeRef. 1160 Ref. probeRef. 0797 5q31.1

4.5 4 3.5 3 2.5 2

Probe ratio 1.5 1 0.5 0 CTSB 8p22CTSB ST7 8q22.3ST7 E2F58q21.2 SLA 8q24.22 SLA PTK2 8q24.3PTK2 ChGn 8p21.3 FGFR 8p11.2 FGFR 8p11.2 MYBL 8q13.1 MYBL TUSC38p22 CHD7 8q12.2 MOS 8q11.23 8q11.23 MOS MYC8q24.21 MYC8q24.21 EIF3E 8q23.1 EIF3E MSRA 8p23.1 MSRA 8q13.3 NCOA EXT1 8q24.11 8q24.11 EXT1 PTP4A 8q24.3 PTP4A GATA4 8p23.1 GATA4 RAD54 8q21.3 EIF3H 8q24.11 8q24.11 EIF3H KCNK9 8q24.3 MFHAS 8p23.1MFHAS TPD528q21.13 DDEF1 8q24.21 DLGAP2 8p23.3 PRKDC 8q11.21 PRKDC 8q11.21 KCNQ3 8q24.22 RECQL4 8q24.3 RECQL4 8q24.3 RNF139 8q24.13 KHDRBS3 8q24.22 Ref. probe1463Ref. 17p12 Ref. probe 0980 11p12 probeRef. 0980 11p12 probeRef. 0976 11p13 probeRef. 0434 11q13 Ref. probeRef. 0678 12p13 probeRef. 0798 13q32 probeRef. 0871 13q34 Ref. probe 1160 3p25.3 probeRef. 1160 Ref. probeRef. 0797 5q31.1

108

Figure 3.4 MLPA probe ratios obtained for the Rhabdomyosarcoma sample

F29R. Probe ratios (Y axis) as (A) individual values from 4 independent experiments

(B) mean probe ratios ± SE. On the X axis, brown columns represent genes located on chromosome 8p, blue columns represent genes located on chromosome 8q, and red columns represent reference probes. The orange column highlights high-level of copy number gain at one reference locus. Probe ratios between 0.7 and 1.3 were considered in the normal range or disomic, probe ratios below 0.7 were considered copy number loss, and probe ratios above 1.3 were considered copy number gain.

109

A. 3

2.5

EIF3E EXT1

2 I 1.5 II Probe ratio III 1 IV

0.5

ST7 8q22.3ST7 CTSB8p22 E2F58q21.2 SLA 8q24.22 SLA PTK28q24.3 FGFR 8p11.2 FGFR 8p11.2 ChGn 8p21.3 TUSC38p22 CHD7 8q12.2 MYC8q24.21 MYC8q24.21 MOS 8q11.23 8q11.23 MOS 8q23.1 EIF3E MYBL 8q13.1 MYBL MSRA 8p23.1 MSRA 8q13.3 NCOA EXT1 8q24.11 8q24.11 EXT1 PTP4A 8q24.3 PTP4A GATA4 8p23.1 GATA4 RAD54 8q21.3

8q24.11 EIF3H KCNK9 8q24.3 MFHAS8p23.1 TPD52 8q21.13 DDEF1 8q24.21 DLGAP28p23.3 RECQL48q24.3 RECQL48q24.3 PRKDC 8q11.21 PRKDC 8q11.21 KCNQ3 8q24.22 RNF139 8q24.13 KHDRBS3 8q24.22 Ref.probe1463 17p12 Ref. probe 0980 11p12 0980 Ref.probe 11p12 0976 Ref.probe 11p13 0434 Ref.probe 11q13 Ref. probe 0678 Ref.probe 12p13 0798 Ref.probe 13q32 Ref. probe 0871 Ref.probe 13q34 Ref. probe 1160 3p25.3 1160 Ref.probe Ref. probe 0797 Ref.probe 5q31.1

1.5

probe ratio probe 0.5

110

Figure 3.5 MLPA probe ratios obtained for the MCF-10A cell line. Probe ratios (Y axis) as (A) individual values from 4 independent experiments (B) mean probe ratios

± SE. On the X-axis, brown columns represent genes located on chromosome 8p, blue columns represent genes located on chromosome 8q, while dark blue columns highlight two probes that gave highly variable results (for EIF3E and EXT1). Probe ratios between 0.7 and 1.3 were considered in the normal range or disomic, probe ratios below 0.7 were considered copy number loss, and probe ratios above 1.3 were considered copy number gain.

111

2.5

2 EIF3E EXT1

1.5

1 I

ratio Probe

II 0.5 III

IV 0

CTSB 8p22CTSB 8q22.3ST7 E2F58q21.2 SLA 8q24.22 SLA PTK2 8q24.3PTK2 ChGn 8p21.3 FGFR 8p11.2 FGFR 8p11.2 MYBL 8q13.1 MYBL TUSC38p22 CHD7 8q12.2 MOS 8q11.23 8q11.23 MOS MYC8q24.21 MYC8q24.21 EIF3E 8q23.1 EIF3E MSRA 8p23.1 MSRA 8q13.3 NCOA EXT1 8q24.11 8q24.11 EXT1 PTP4A 8q24.3 PTP4A GATA4 8p23.1 GATA4 RAD54 8q21.3 EIF3H 8q24.11 8q24.11 EIF3H KCNK9 8q24.3 MFHAS 8p23.1MFHAS Ref. probe1463Ref. TPD528q21.13

probeRef. 1160 DDEF1 8q24.21 Ref. probeRef. 0797 probeRef. 0980 probeRef. 0976 probeRef. 0434 probeRef. 0678 probeRef. 0798 probeRef. 0871 DLGAP2 8p23.3 PRKDC 8q11.21 PRKDC 8q11.21 KCNQ3 8q24.22 RECQL4 8q24.3 RECQL4 8q24.3 RNF139 8q24.13

KHDRBS3 8q24.22

2.5

1.5 ratio

Probe 1

0.5

0 CTSB 8p22CTSB ST7 8q22.3ST7 E2F58q21.2 SLA 8q24.22 SLA PTK2 8q24.3PTK2 ChGn 8p21.3 FGFR 8p11.2 FGFR 8p11.2 MYBL 8q13.1 MYBL TUSC38p22 CHD7 8q12.2 MOS 8q11.23 8q11.23 MOS MYC8q24.21 MYC8q24.21 EIF3E 8q23.1 EIF3E MSRA 8p23.1 MSRA 8q13.3 NCOA EXT1 8q24.11 8q24.11 EXT1 PTP4A 8q24.3 PTP4A GATA4 8p23.1 GATA4 RAD54 8q21.3 EIF3H 8q24.11 8q24.11 EIF3H KCNK9 8q24.3 MFHAS 8p23.1MFHAS TPD528q21.13 DDEF1 8q24.21 DLGAP2 8p23.3 PRKDC 8q11.21 PRKDC 8q11.21 KCNQ3 8q24.22 RECQL4 8q24.3 RECQL4 8q24.3 RNF139 8q24.13 KHDRBS3 8q24.22 Ref. probe1463Ref. 17p12 Ref. probe 0980 11p12 probeRef. 0980 11p12 probeRef. 0976 11p13 probeRef. 0434 11q13 Ref. probeRef. 0678 12p13 probeRef. 0798 13q32 probeRef. 0871 13q34 Ref. probe 1160 3p25.3 probeRef. 1160 Ref. probeRef. 0797 5q31.1

112

Figure 3.6 MLPA probe ratios obtained for the Ewing sarcoma sample F15E.

Probe ratios (Y axis) as (A) individual values from 4 independent experiments, and

(B) mean probe ratios ± SE. On the X-axis, brown columns represent genes located on chromosome 8p, blue columns represent genes located on chromosome 8q, and bright blue columns highlight two probes that gave highly variable results. Probe ratios between 0.7 and 1.3 were considered in the normal range or disomic, probe ratios below 0.7 were considered copy number loss, and probe ratios above 1.3 were considered copy number gain.

113

3.4 Discussion

Chromosome 8 contains a number of important genes that are amplified in specific adult cancers such as prostate (Tomlins et al., 2007) and breast (Rodriguez et al.,

2007), and in childhood cancers such as osteosarcoma (Man et al., 2004), Burkitt lymphoma (De Falco et al., 2007) and some leukaemias (Le Beau et al., 2002). In this study we present a detailed analysis of DNA copy number changes at 30 genes across chromosome 8 in four different types of paediatric solid tumours with the aim of associating gene copy number alterations with tumour type.

Although the MLPA technique has increased in use, over recent years (Fig 3.1), it has seldom been applied to a study of paediatric solid tumours. Among the four types of cancers studied in this project, only neuroblastoma has been previously examined using MLPA. Ambros et al. (2011) presented and validated MLPA to better assign patients to certain genetic risk groups, and also to discover genomic regions of interest with possible clinical impact. A neuroblastoma-specific MLPA kit was designed by the SIOP (Europe) Neuroblastoma Biology committee in cooperation with MRC-Holland, testing 310 neuroblastomas and 8 neuroblastoma cell lines in 9 different SIOPEN reference laboratories. The study concluded that the technique was reliable as the results were validated with FISH for MYCN amplification, cost-effective when a large number of samples is analysed in one experiment and the robustness is based on the PCR technique itself (Ambros et al., 2011).

114

3.4.1 Previous research examining chromosome 8 copy number using MLPA

Buffart et al. (2005) used a chromosome 8 specific MLPA kit to measure the DNA copy number status of 25 chromosome 8 genes in primary colorectal cancer and metastatic disease. Two genes located at chromosome 8q23-24, namely MYC and

PTP4A3, were shown to be amplified and overexpressed in primary colorectal cancer and metastases. The majority of primary tumours and their corresponding liver metastases presented similar gains and losses, with significant difference for only TPD52 and EIF3S6, where gain was higher in liver metastases (Buffart et al.,

2005). Subsequently, Buffart et al. (2009) studied DNA copy number changes in 58 individual genes in 63 formalin-fixed and paraffin-embedded gastric adenocarcinomas, using different MLPA kits for chromosome 8, 13 and 20. They reported that 11 MLPA probes located in 8 genes on chromosome 8, showed copy number gains, with frequencies from 25.4% (TPD52) to 73% (MYC) (Buffart et al.,

2009).

Bremmer et al. (2005) used normal tissue, preneoplastic lesions and oral squamous cell carcinoma with a partial kit for chromosome 8 (5 probes in 5 genes located on chromosome 8) as well as other MLPA kits, to develop and evaluate a noninvasive screening test for oral preneoplastic lesions. Here, MLPA was described as a sensitive, reliable, high-throughput and easy to perform technique (Bremmer et al.,

2005).

More recently Rumiato et al. (2011) analysed 59 esophageal cancer samples with 4

MLPA kits, including the chromosome 8 kit. They reported high frequency of copy

115

number loss of four genes of 8p: DLGAP2, MSRA, CTSB and ChGn, in squamous cell carcinoma. Yong et al. (2014) examined 33 oral squamous cell carcinomas using the chromosome 8 MLPA kit. The most frequent copy number losses were found on chromosome 8p, while the most frequent copy number gains were found on chromosome 8q, at EIF3E, MYC, RECQL4 and MYBL1.

3.4.2 Six Breast cell lines: included in this study as known positive controls

We analysed six breast cell lines with chromosome 8 copy number changes that had been reported previously (Table 3.1). Five of 6 cell lines presented copy number loss at most chromosome 8p loci (Table 3.13), except MCF-10. This is a spontaneously immortalized breast cell line, considered epithelial in origin, as it does not show characteristics of invasiveness or tumour formation (Marella et al., 2009).

Of the other five cell lines, SK-BR-3 has previously shown three well defined chromosome 8 regions, namely copy number loss from 8p23.3-p11.1, and two regions with copy number gain at 8q13.1-q21.3 that contains TPD52, and

8q23.2-q24.21, that includes MYC (Choschzick et al., 2010; Rodriguez et al., 2007).

The amplification of MYC in SK-BR-3 cell lines was also previously demonstrated by

Schouten et al. (2002); where they used two different MLPA probe mixtures sets that contained 41 different probes specific for human DNA (Schouten et al., 2002). The fact that our results in breast cell lines are broadly in accordance to those reported in the literature cited suggest that our results are reliable.

3.4.3 Chromosome 8p copy number in paediatric solid tumours

Among these tumours, the osteosarcomas group showed copy number loss at 7/8

116

chromosome 8p loci, whereas Ewing sarcoma and rhabdomyosarcomas had very few evident copy number changes. Neuroblastomas showed copy number losses at

ChGn in 5/10 (50%) cases, followed by MSRA in 4/10 (40%) cases (Table 3.13).

MSRA (Methionine sulfoxide reductase A), located at 8p23.1, was considered by Lei et al. (2007) as a possible candidate metastasis suppressor gene. This gene was disomic in Ewing sarcoma and rhabdomyosarcoma, whereas it exhibited losses in both osteosarcoma (2/5, 40%) and neuroblastoma (4/10, 40%).

ChGn or CSGALNACT1, (Chondroitin sulfate N-acetylgalactosaminyltransferase 1), is a member of the GalNac-transferase gene family, expressed in most biological tissues and cell types. This gene showed copy number loss in 50% (5/10) of neuroblastomas and in 20% (1/5) of osteosarcomas.

To our knowledge, the specific role of these genes in neuroblastoma or osteosarcoma has not been well studied yet. It is the first time that specific genes with copy number losses located in 8p has been associated with neuroblastomas or osteosarcomas, results that could be the basis of further study.

117

Table 3.13 Copy number results obtained using MLPA in a cohort of 31 paediatric solid tumours and 6 breast cell lines

8p (8 genes) Chromosome 8q (22 genes)

Genes

DLGAP2 MFHAS MSRA GATA4 CTSB TUSC3 ChGn FGFR1 PRKDC MOS CHD7 MYBL NCOA2 TPD52 E2F5 RAD54B ST7 EIF3E EIF3H EXT1 RNF139 MYC DDEF1 KCNQ3 SLA KHDRBS3 KCNK9 PTK2 PTP4A RECQL4 Samples

F 9-O 1a 1 1 1 1 1 1 1 1 1 1 1 1 2b 1 1 1 1 1 1 1 1 1 1 1 1 0c 1 1 1 F 10-O 1 1 1 1 1 1 1 1 1 0 1 2 1 2 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 M 11-O 1 1 2 1 1 2 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 F 12-O 0 1 0 0 0 1 0 0 1 1 2 2 2 0 0 1 0 2 2 2 2 2 1 1 1 1 0 0 0 1 M 46-O 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 1 1 1 1 2 2 1 1 1 1 M 13-E 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 M 14-E 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 2 1 2 1 1 1 1 1 1 1 1 1 1 F 15-E 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 M 16-E 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 F 17-E 1 1 1 1 1 2 1 1 1 1 2 1 1 2 1 1 1 1 1 1 1 1 1 1 2 1 0 1 1 1 F 18-E 1 2 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 M 22-E 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 0 1 1 1 F 24-E 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 M 25-R 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 F 26-R 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 M 27-R 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 F 28-R 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 F 29-R 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 2 1 1 1 0 1 2 1 1 1 1 0 1 0 1 M 31-R 1 1 1 1 1 2 1 1 1 1 2 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 M 32-R 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 M 35-R 1 1 1 1 1 2 1 1 1 2 1 2 1 2 1 2 1 1 1 0 1 1 1 1 2 1 0 1 0 1 M 36-N 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 F 37-N 1 2 0 1 2 1 0 2 0 1 1 0 1 0 0 1 0 1 0 0 0 1 1 0 1 0 1 1 1 2 F 38-N 2 1 0 1 1 1 0 1 1 1 1 0 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 2 M 39-N 1 2 0 1 1 1 0 1 0 1 0 0 1 0 0 0 0 1 0 1 1 1 1 1 1 0 1 1 1 2 F 40-N 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 M 41-N 1 1 1 1 1 2 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 F 42-N 1 1 1 1 1 2 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 M 43-N 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 F 44-N 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 F 45-N 2 0 0 0 2 2 0 0 0 2 1 2 0 2 1 1 0 2 0 0 0 1 0 0 2 0 0 0 0 0 AU565 0 0 0 0 0 0 0 0 0 1 1 1 1 2 1 0 0 0 2 2 2 2 1 0 0 0 0 0 0 1 HCC2157 0 0 0 0 0 1 0 0 2 2 2 2 2 2 2 2 1 2 2 1 1 2 1 1 1 1 1 1 1 2 MCF-7 0 0 0 0 0 0 0 0 1 1 1 1 0 1 1 2 2 2 1 1 2 2 2 1 1 1 0 1 1 1 MCF-10A 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 2 1 1 1 1 1 MDA-MB-231 0 0 0 0 0 0 0 0 2 1 2 2 1 2 1 1 1 1 1 1 0 1 2 1 1 1 0 1 1 1 SKBR 3 0 0 0 0 0 0 0 0 1 1 1 1 1 2 1 0 0 0 2 2 2 2 1 0 1 0 0 0 0 0 a1 = diploid copy number b2 = copy number gain c0 = copy number loss

118

3.4.4 Chromosome 8q copy number in paediatric solid tumours

The osteosarcoma cohort showed copy number loss at 8/22 loci and copy number gain at 13/22 loci. This finding, comprising of a moderate number of copy number changes across 8q (Table 3.13) was not in agreement with the literature. Typically,

OS is characterized by the presence of heterogeneous complex chromosomal abnormalities (Ragland et al., 2002), with a high degree of aneuploidy, gene amplification, and multiple unbalanced structural chromosomal rearrangements

(Sandberg et al., 2000). In contrast, only one sample here showed copy number gain

(at MYC) and one other sample showed copy number loss at RECQL4 (Tables 3.11,

3.13).

Instability of chromosome 8q has been described by many laboratories, often involving MYC, in cytoband 8q24.21, which is gained at varying frequencies (Bayani et al., 2003; Smida et al., 2010; Stock et al., 2000). Commonly, other regions of 8q, including 8q23-qter, 8q21.3-8q23, and 8q21 also undergo copy number gains (dos

Santos Aguiar et al., 2007; Kresse et al., 2009; Smida et al., 2010). In addition, copy number gain of RECQL4, mapping within 8q23-qter, has been reported as a frequent event in osteosarcoma (Maire et al., 2009; Nishijo et al., 2004).

In this study, the chromosome 8q locus that showed the most frequent copy number gain was TPD52, followed by SLA. TPD52 or Tumour Protein D52 (8q21.13) has been associated with cellular transformation, proliferation and metastasis and its overexpression has been demonstrated in various adult cancers (Byrne et al., 2010).

In paediatric cancers, TPD52 immunostaining was analysed in a large cohort of

119

Ewing sarcoma familial tumours, and was detected significantly more frequently in disseminated versus localized tumours (Machado et al., 2011). Among the 8q loci we have studied: TPD52 showed copy number gain in all four solid tumours classes tested here: namely, 2/5 (40%) osteosarcomas, 1/8 (25%) Ewing sarcomas, 3/8

(37.5%) rhabdomyosarcomas, and 3/10 (30%) neuroblastomas (Tables 3.11 and

3.13).

SLA also known as Src-like-adaptor (8q24.22) is an adapter protein, that negatively regulates T-cell receptor (TCR) signalling (Dragone et al., 2006). In the present study, SLA showed copy number gain in 2/5 (40%) osteosarcomas, 2/8 (25%) Ewing sarcomas, 1/8 (12.5%) rhabdomyosarcoma, and 1/10 (10%) neuroblastoma (Tables

3.11 and 3.13).

Another result that we did not expect was that in Ewing sarcoma, gain of the entire chromosome 8 was not identified for any sample, and only 5/22 chromosome 8q loci showed copy number gain (Table 3.13). This was also surprising, as it is known that gain of chromosome 8 is one of the most prominent features of this type of tumour

(Savola et al., 2009; Toomey et al., 2010).

In the osteosarcoma samples, when Fisher’s exact test was applied (Appendix 1), a significant difference was shown (p= 0.016) between the expected number of copy number events per category and the actual count at 8p loci and 8q loci. The number of copy number losses at 8p was significantly higher than expected, as also was the number of copy number gains in 8q. In this regards, the OS group is comparable with the breast cancer controls. This matter is discussed further in a later chapter, in

120

relation to the general nature of chromosome mutation, and its effect on simultaneous patterns of gains versus loss.

3.4.5 Technical considerations

The MRC-Holland MLPA® technique guide (Schouten et al., 2002) suggests that

MLPA can be performed using 50 to 100 ng of DNA. However, in our hands 50 ng of

DNA was insufficient (data not shown). Combaret et al. (2012) reported similar problems, and Stuppia et al. (2012) recommended using 100 to 200 ng of genomic

DNA for optimal results.

This problem is exacerbated when DNA is derived from tumour samples containing less than 50% neoplastic cells. In such cases of tumour mosaicism, it can be difficult to detect copy number changes for a specific gene, as the presence of normal cells can mask the detection of abnormal cells (Stern et al., 2004). Similarly, Ambros et al.

(2011) mentioned that discrepant results may be obtained due to low tumour cell content, particularly in post-therapy specimens. Another limitation of MLPA is its inability to detect “balanced” rearrangements, where copy number is preserved at normal levels (Bravaccini et al., 2012). In this project a proportion of Ewing sarcomas, alveolar rhabdomyosarcomas and neuroblastomas may well have had balanced translocations (whereas the same would be much less likely for osteosarcomas and embryonal rhabdomyosarcomas).

It has also been reported that MLPA analysis can be compromised by the presence of unreliable probes (Stuppia et al., 2012). We obtained some evidence of this,

121

shown in Fig 3.5 and Fig 3.6, where it can be seen that the MLPA chromosome 8 kit may include some MLPA probes that lack reproducibility.

While MLPA comprises a relatively rapid and convenient technique, not all of our results conformed to the literature’s expectations. This led us to undertake further copy number analyses, for the same control cell lines and paediatric tumour cohort, using aCGH testing (Chapter 4).

122

Chapter 4

Analysis of copy number in paediatric solid tumours

using array Comparative Genomic Hybridization

123

4.1 Introduction

Array Genomic hybridization (aCGH) was designed to detect DNA copy number alterations such as gene amplifications, aneuploidy, deletions and non-reciprocal translocations, which are characteristic features of cancer genomes (Kallioniemi,

2008). aCGH has been widely used to analyze DNA copy number aberrations on a genome-wide scale in a single experiment (Yatsenko et al., 2009). Whereas MLPA,

(chapter 3) obtains copy number data for a restricted number of genes, aCGH allows us to define rearrangements in a more detailed and comprehensive manner across the whole genome.

Array CGH uses two types of genomic DNA, tumoural and reference DNA, that are differentially labeled with fluorochromes, usually Cy3 and Cy5 respectively. Samples are hybridized to a microarray that contains oligonucleotide probes, which can be customized to particular genes, or designed to analyze all known genes. After hybridization, the amount of fluorescent signal can be measured and a ratio between tumour and reference DNA can be calculated based on the ratio of the signal intensity.

In this project we used a customized design from the Agilent eArray SurePrint G3

2x400K. This array contained a total of 420,288 distinct biological 60-mer oligonucleotide probes. The array provided high resolution on chromosome 8, with a median probe spacing of 1.92 kb for chromosome 8, and high resolution in 95 genes important in cancer, with 250 bp spacing https://earray.chem.agilent.com/suredesign

(Table 4.1). The rest of the human genome was provided with even whole-genome

124

coverage of 253,502 probes with 11.8 kb median probe spacing (Table 4.2). The probe sequences and gene annotations were based on NCBI Build 37 of the human genome, and UCSC version hg19 released in February 2009.

Array CGH has revolutionized cytogenetic diagnostics and has given the clinician a greater appreciation of variability in the clinical presentation of many well-described conditions (Cytrynbaum et al., 2005; Gropman et al., 2007). The introduction of aCGH in clinical practice has effectively eliminated all the technical impediments of traditional cytogenetics and FISH (Table 1.6), and allowed the detection of such conditions with relative, but not complete, independence from the clinician's diagnostic judgment (Bejjani et al., 2008). These advantages explain why the use of aCGH for the diagnosis of constitutional anomalies is progressing faster than the use of expression microarrays for the prediction of clinical outcome (e.g., in cancer), for which a few applications are entering clinical practice (Bejjani et al., 2008).

A major limitation for using array CGH is the high experimental cost. However high quality DNA, as obtained from cell lines or fresh frozen tissue, minimizes hybridization failures and therefore overall costs. As with other clinical diagnostic methods, there are other limitations to aCGH technology. aCGH is not able to identify balanced rearrangements such as translocations, ring chromosomes or inversions, as the copy number is not affected (Holcomb et al., 2011).

As was discussed previously (section 1.4), gains of chromosome 8 are amongst the most common cytogenetic events that occur in cancer. Nevertheless few studies have examined copy number changes of chromosome 8 in childhood solid tumours

125

(section 1.5). The design of these arrays allowed us to examine the status of genes tested by MLPA, copy number changes at important genes involved in cancer, as well as broader cytogenetic changes affecting any region of the genome.

126

Table 4.1 Cancer genes with lower oligonucleotide probe spacing according to Agilent eArray data Name Chromosome Cytogenetic location

JUN Chr 1 1p32.1 LCK Chr 1 1p35.1 MYCL1/MYCL Chr 1 1p34.2 NRAS Chr 1 1p13.2 SKI Chr 1 1p36.33 TAL1 Chr 1 1p33 NTRK1 Chr 1 1q23.1 ABL2 Chr 1 1q25.2 FGR Chr 1 1p36.11 MSH2 Chr 2 2p21 MYCN Chr 2 2p24.3 REL Chr 2 2p16.1 FOSL2 Chr 2 2p24.2 HER4/ERBB4 Chr 2 2q34 MLH1 Chr 3 3p22.2 RAF1 Chr 3 3p25.2 TGFBR2 Chr 3 3p24.1 THRB Chr 3 3p24.2 VHL Chr 3 3p25.3 HEK/EPHA3 Chr 3 3p11.1 KIT Chr 4 4q12 MCC Chr 5 5q22.2 APC Chr 5 5q21 CSF1R Chr 5 5q32 FER Chr 5 5q21.3 MAS1 Chr 6 6q25.3 MYB Chr 6 6q23.3 PIM1 Chr 6 6p21.2 ROS1 Chr 6 6q22.1 WAF1/CDKN1A Chr 6 6p21.2 FYN Chr 6 6q21 MET Chr 7 7q31.2 EGFR/ ERBB-1 Chr 7 7p11.2 WNT2 Chr 7 7q31.2 LYN Chr 8 8q12.1 MOS Chr 8 8q12.1 MYBA/MYBL1 Chr 8 8q13.1 MYC Chr 8 8q24.21 TPD52 Chr 8 8q21.13 INK4B/CDKN2B Chr 9 9p21.3 INK4A/CDKN2A Chr 9 9p21.3 PTC/RET Chr 10 10q11.21 KIP2/CDKN1C Chr 11 11p15.4 CBL Chr 11 11q23.3 CCND1 Chr 11 11q13.3 ERGB2/FLI1 Chr 11 11q24.3 ETS1 Chr 11 11q24.3 FGF3 Chr 11 11q13.3 FRA1/FOSL1 Chr 11 11q13.1 HRAS1 Chr 11 11p15.5 HSTF1/FGF4 Chr 11 11q13.3 WT1 Chr 11 11p13

127

Table 4.1 Cancer genes with lower oligonucleotide probe spacing according to Agilent eArray data (cont.) Name Chromosome Cytogenetic location

KRAS2A Chr 12 12p12.1 WNT1 Chr 12 12q13.12 CDK4 Chr 12 12q14.1 ELK3 Chr 12 12q23.1 ERBB3 Chr 12 12q13.2 HST2/FGF6 Chr 12 12p13.32 RB1 Chr 13 13q14.2 BRCA2 Chr 13 13q13.1 MAX Chr 14 14q23.3 FOS Chr 14 14q24.3 BCL2ALPHA/BCL2A1 Chr 15 15q25.1 FES Chr 15 15q26.1 E-CAD/CDH1 Chr 16 16q22.1 INT2/INTS2 Chr 17 17q23.2 NF1 Chr 17 17q11.2 RASD1 Chr 17 17p11.2 TOP3A Chr 17 17p11.2 TP53/P53 Chr 17 17p13.1 THRA/ ERBA1/ NR1A1 Chr 17 17q21.1 BRCA1/BRAT1 Chr 17 17q21.31 CRK Chr 17 17p13.3 ERBB2 Chr 17 17q12 BCL2BETA/BCL2 Chr 18 18q21.33 DCC Chr 18 18q21.2 DPC4/SMAD4 Chr 18 18q21.2 YES1 Chr 18 18p11.32 JUNB Chr 19 19p13.2 JUND Chr 19 19p13.11 AKT2 Chr 19 19q13.2 BCL3 Chr 19 19q13.32 VAV1 Chr 19 19p13.3 SRC Chr 20 20q11.23 MYBB/MYBL2 Chr 20 20q13.12 RBAP/E2F1 Chr 20 20q11.22 HCK Chr 20 20q11.21 TIAM1 Chr 21 21q22.11 ERG Chr 21 21q22.2 ETS2 Chr 21 21q22.2 NF2 Chr 22 22q12.2 PDGFB Chr 22 22q13.1 BCR Chr 22 22q11.23 DBL/MCF2 Chr X Xq27.1 ELK1 Chr X Xp11.23

128

Table 4.2 Customized Agilent eArray SurePrint G3 2X400K array

Probe location Number of probes Median Probe Spacing (kb)

Chromosome 8 52,639 1.92

Genes important in oncology 39,654 0.25-0.5 n =95

Whole genome 420,288 11.81

Information gathered from booklet provided by Agilent eArray customized array

129

4.2 Aims for aCGH

1. To obtain copy number data for chromosome 8 in our cohort of paediatric

solid tumours and breast cancer cell lines.

2. To allow a broader assessment of copy number gains and losses in the same

cohort in order to identify regions commonly gained or lost.

130

4.3 Results

When employing aCGH, the presence of chromosomal imbalances can be detected and quantified by calculating the ratio of signal intensities of test DNA versus reference DNA. For statistical purposes, these ratios are usually converted to a log2

(test/ref) ratio. The log2 (Cy5/Cy3) scaling has the consequence of centring the ratios at approximately zero, making DNA copy number losses appear negative, and copy number gains positive. To correct for experimental artefacts, aCGH data are normalized so that the ratio for a diploid genome is set to a standard value, usually

1.0 on a linear scale (0.0 on a log2 scale), and threshold values are established for loss, no change, and gain status relative to the diploid reference DNA (Yatsenko et al., 2009). The typical mean normal ratio for aCGH has been established as 0.0

±0.25 on log2 scale, indicating equal copy number within test and reference sample

DNA (Yatsenko et al., 2009). A change in log2 (Cy5/Cy3) ratio in a chromosome region between 0.25 - 0.75 was classified as genomic gain, > 0.75 as high copy number gain, < -0.25 as hemizygous loss, and < -0.75 as homozygous deletion

(Rothenberg et al., 2010). Oligonucleotide microarray data were obtained using

Agilent Feature Extraction software 10.10, and were imported into Agilent

CytoGenomics software V2.7.2 and Nexus Copy number v7.5 for analysis. The threshold setting to make a positive call was at least 5 consecutive probes, and the built-in QC thresholds, e.g. MAPD < 0.3, were used.

According to the characteristics of the design of these arrays (Table 4.2), genome-wide results will be presented first with common regions of genomic gain or loss identified, including genes of known relevance to cancer (Table 4.1). Results for

131

chromosome 8 will then be presented for each tumour type, and finally aCGH results for genes previously explored by MLPA will be described.

4.3.1 Chromosome 8 copy number changes in breast cell lines

A genomic compound plot showing copy number changes for all cell lines, showed that chromosome 8q presented more copy number gains than other chromosomes with the exception of chromosome 1q (Fig 4.1).

Breast cancer cell lines (n=5) showed copy number loss at almost every chromosome 8p locus (Fig 4.2), MCF-10A cells however presented a disomic copy number in chromosome 8p (Fig 4.2B).

Copy number gains were frequently identified at MLPA chromosome 8q loci in all cell lines (Table 4.3). Four loci (EIF3H, EXT1, MYC and PTP4A3) were gained in all (6/6) cell lines, followed by RNF139 which was gained in 83% (5/6) cell lines and 12 loci were gained in 67% (4/6) cell lines. These copy number changes are broadly consistent with those reported previously in these breast cell lines (Table 3.1). All

Individual genome copy number views for each cell line are shown in Appendix 2.

132

133

Figure 4.1 Genomic copy number changes in breast cell lines (n=6). Blue indicates copy number gain events plotted above the baseline. Losses are shown in red below the baseline. A. Compound plot showing frequency of gain/loss on all chromosomes. The Y axis indicates the percentage of the cell lines population showing a copy number gain/loss at specific points along the genome (X axis, Chr 1

è Chr Y). Dark blue and dark red indicate more than two-fold copy gain and homozygous loss, respectively. B. Individual copy number plots for each cell line.

The Y axis represents the copy number events for each cell line. The X axis, shows copy number gain/loss across the genome, the baseline indicates disomic copy number. The compound and individual plots were generated using Nexus Copy number software v 7.5.

134

Figure 4.2 Chromosome 8 copy number gain/loss measured in breast cell lines (n=6). Blue indicates copy number gain events and are plotted above the baseline. Losses are shown in red below the baseline. A. Compound plot showing frequency of gain/loss along chromosome 8, shown as an ideogram above and also according to Mb position on the X axis. The Y axis indicates the percentage of cell lines showing copy number gain/loss at specific points along chromosome 8. Dark blue and dark red indicate more than two-fold copy gain and homozygous loss, respectively. B. Individual copy number plots for each cell line. The Y axis represents the copy number events for each cell line. The X axis, shows copy number gain/loss along chromosome 8 and the baseline indicates disomic copy number. The compound and individual plots were generated using Nexus Copy number software v7.5.

135

Table 4.3 Chromosome 8 copy number status of genes previously examined using MLPA in breast cell lines (n=6)

AU565 HCC2157 MCF-7 MCF-10A MDA-MB-231 SK-BR-3 Gain Disomic Loss

n(%) n(%) n(%) Genes

DLGAP2 0 0 0 1 0 0 0(0) 1(17) 5(83) MFHAS1 0 0 0 1 0 0 0(0) 1(17) 5(83) MSRA 0 0 0 1 0 0 0(0) 1(17) 5(83) GATA4 0 0 0 1 0 0 0(0) 1(17) 5(83) CTSB 0 0 0 1 0 0 0(0) 1(17) 5(83) TUSC3 0 1 0 1 0 0 0(0) 2(33) 5(83) ChGn 0 0 0 1 0 0 0(0) 1(17) 5(83) FGFR1 0 0 0 1 0 0 0(0) 1(17) 5(83) PRKDC 1 2 1 1 2 2 3(50) 3(50) 0(0) MOS 1 2 1 1 2 2 3(50) 3(50) 0(0) CHD7 1 2 1 1 2 2 3(50) 3(50) 0(0) MYBL1 1 2 1 1 2 2 3(50) 3(50) 0(0) NCOA2 2 2 0 1 2 2 4(67) 1(17) 1(17) TPD52 2 2 1 1 2 2 4(67) 2(33) 0(0) E2F5 2 2 1 1 2 2 4(67) 2(33) 0(0) RAD54B 0 2 2 1 2 0 3(50) 1(17) 2(33) ST7 0 2 2 2 2 0 4(67) 0(0) 2(33) EIF3E 0 2 2 2 2 0 4(67) 0(0) 2(33) EIF3H 2 2 2 2 2 2 6(100) 0(0) 0(0) EXT1 2 2 2 2 2 2 6(100) 0(0) 0(0) RNF139 2 2 2 2 0 2 5(83) 0(0) 1(17) MYC 2 2 2 2 2 2 6(100) 0(0) 0(0) DDEF1 1 2 2 2 2 1 4(67) 2(33) 0(0) KCNQ3 0 2 2 2 2 1 4(67) 1(17) 1(17) SLA 1 2 2 2 2 1 4(67) 2(33) 0(0) KHDRB 1 2 2 2 2 1 4(67) 2(33) 0(0) S3 KCNK9 0 2 2 2 2 0 4(67) 0(0) 2(33) PTK2 0 2 2 2 2 0 4(67) 0(0) 2(33) PTP4A3 2 2 2 2 2 2 6(100) 0(0) 0(0) RECQL4 2 2 2 1 1 2 4(67) 2(33) 0(0)

2= copy number gain; 1= disomic; 0= copy number loss

136

4.3.2 Compound plots indicating common copy number changes in paediatric solid tumours

Compound copy number plots were derived to show significant peaks using all data obtained from aCGH experiments. This feature was used to establish the significance of the copy number aberrations in a region. The software applies the Significance

Testing for Aberrant Copy number (STAC), that was developed to test the significance of DNA copy number aberrations across multiple array experiments

(Diskin et al., 2006). The algorithm identifies a set of copy number changes that are stacked on top of each other, such that they would not occur randomly. The cut-off

35% frequency was used, suggested by Nexus software and traditionally used by researchers (Diskin et al., 2006). Therefore common copy number gains and losses present in more than 35% (11/31) of samples, and significant at p=0.05 were considered of potential interest. STAC was then used to indicate common regions of copy number gains or losses in the 4 types of cancer in our study. Tables constructed from the significant peaks and compound plots will be presented according to high frequency, significant p-value and common events observed.

Genes commonly gained across the 4 types of cancers were located at chromosome

22q11.22. In osteosarcomas, copy number gain was shown at MIR650, PRAME and

IGLL5 among other genes, with frequency of 80% and p=0.001 (one-tailed z-test)

(Table 4.4). These same genes presented copy number gain in Ewing sarcoma with a frequency of 62.5% and p=<0.001 (Table 4.5), and in ERMS with frequency of 83.3% and p=0.001 (Table 4.6). Similarly in neuroblastoma, common copy number gain was present at chromosome 22q11.22 including PRAME, BCR and IGLL5, with a

137

frequency of 70% and p value=<0.001 (Table 4.7). ARMS (n=2), no significant peaks were present at 22q11.22, however, two genes with copy number loss were at chromosome 5q31.1 (SAR1B and SEC24A) with frequency of 100% and p=0.002

(Table 4.8).

The customized array also offered cancer genes with lower oligonucleotide probe spacing (n=95) (Table 4.1). From this list, no gene consistently showed copy number gain or loss in all samples. According to tumour type, genes that consistently showed gain/loss were included in Table 4.10. In osteosarcomas, 6/95 cancer genes showed copy number loss, and BRCA2 at chromosome 13q13.1 presented copy number loss in 4/5 samples, and copy number gain in 1/5 samples. No cancer genes from this list presented copy number changes in any samples of Ewing sarcoma. In ERMS, CRK at chromosome 17p13 showed copy number gain in 5/6 samples and loss in 1/6 sample. Only 2/96 genes showed copy number gain in the two ARMS samples namely WNT1 on chromosome 12q13.12 and CDK4 located on chromosome

12q14.19. The 10 neuroblastoma samples did not present any copy number loss at any of the 95 genes, however copy number gain was measured in 8/10 samples at

MYCN at chromosome 2p24.3 (Table 4.9).

138

Table 4.4 Common copy number changes in osteosarcoma n=5

Chromosome Region Number of Event Frequency p-Value Gene Symbols Length (bp) genes %

Chr14q13.1 231169 3 CN Loss 60 0.046 SPTSSA, EAPP, SNX6 Chr14q23.2 249246 3 CN Loss 60 0.046 WDR89, JA429503, SGPP1 Chr14q32.33 461433 12 CN Gain 80 0.018 IGH, DKFZp686O16217, IGHE, IGHG1, IGHD, FLJ00382, MIR4539, AK128652, KIAA0125, ADAM6, BC042994, LINC00226

Chr22q11.22 793423 17 CN Gain 80 0.001 VPREB1, LOC96610, BMS1P20, ZNF280B, ZNF280A, PRAME, LL22NC03-63E9.3, LOC648691, BCR, POM121L1P, DQ597441, GGTLC2, DKFZp667J0810, MIR650, MIR5571, IGLL5, DQ575049

139

Table 4.5 Common copy number changes in Ewing sarcoma n=8

Chromosome Region Number Event Frequency p-Value Gene Symbols Length (bp) of genes %

Chr1q21.2 192043 2 CN Loss 37.5 0.005 NBPF25P, BC062745 Chr3q26.1 90142 1 CN Gain 37.5 <0.001 BC073807 Chr4p14 151263 2 CN Loss 50 0.007 UBE2K, PDS5A Chr4p14 90942 2 CN Loss 50 0.007 MIR4802, RBM47 Chr6q21 312590 4 CN Loss 37.5 0.034 CDK19, BC047513, AMD1, GTF3C6 Chr10q21.3 208948 5 CN Loss 50 0.002 PBLD, HNRNPH3, RUFY2, DNA2, SLC25A16 Chr15q11.1 - q11.2 2231225 44 CN Gain 50 <0.001 DQ576041, DQ571479, BC107108, DQ592463, CHEK2P2, HERC2P3, HERC2P7, DQ582073, GOLGA6L6, DQ594309, DQ595648, DQ600342, DQ582939, DQ578838, GOLGA8CP, DQ572979, JB175342, NBEAP1, DQ573684, DQ595048, MIR3118-2, MIR3118-3, MIR3118-4, POTEB2, POTEB, POTEB3, NF1P2, MIR5701-1, MIR5701-2, MIR5701-3, CT60, LINC01193, LOC646214, CXADRP2, DQ583164, DQ582260, DQ590589, DQ587539, POTEB3, DQ786202, LOC101927079, LOC727924, OR4M2, OR4N4 Chr15q25.2 164959 11 CN Gain 37.5 0.017 CPEB1, DQ601936, CPEB1-AS1, LOC283692, AP3B2, LOC338963, ACTG1P17, LOC283693, SNHG21, SCARNA15, FSD2 Chr17q21.31 624031 9 CN Loss 50 0.009 KANSL1, KANSL1-AS1, LRRC37A, ARL17A, ARL17B, NSFP1, LRRC37A2, ARL17A, NSF Chr19q13.12 239303 10 CN Gain 50 <0.001 ZNF461, LINC01534, ZNF567, ZNF850, AX747375, LOC728485, BC024306, ZNF790-AS1, ZNF790, ZNF345 Chr21p11.2 - p11.1 269522 7 CN Gain 37.5 0.001 TPTE, Mir_548, BAGE3, BAGE2, BAGE4, BAGE5, BAGE Chr22q11.22 324442 11 CN Gain 62.5 <0.001 VPREB1, PRAME, BCR, POM121L1P, DQ597441, GGTLC2, DKFZp667J0810, MIR650, MIR5571, IGLL5, DQ575049

140

Table 4.6 Common copy number changes in ERMS n=6

Chromosome Region Number Event Frequency p-Value Gene Symbols Length (bp) of genes %

Chr3p25.2 2666 1 CN Loss 83.3 <0.001 RAF1 Chr14q32.33 81594 9 CN Gain 66.7 <0.001 IGH, DKFZp686O16217, IGHE, IGHG1, IGHD, FLJ00382, MIR4539, AK128652, KIAA0125 Chr15q15.2 354412 2 CN Loss 83.3 0.032 TTBK2, UBR1 Chr15q15.2 - q15.3 101353 5 CN Loss 83.3 0.032 TGM7, LCMT2, ADAL, ZSCAN29, TUBGCP4 Chr17q25.3 29047 4 CN Gain 66.7 <0.001 ASPSCR1, STRA13, LRRC45, RAC3 6 LOC102723376, DUX4, MIR8078, ROCK1P1, USP14, Chr18p11.32 222088 CN Loss 50.0 0.026 THOC1 10 ZNF461, LINC01534, ZNF567, ZNF850, AX747375, Chr19q13.12 223669 CN Gain 66.7 0.001 LOC728485, BC024306, ZNF790-AS1, ZNF790, ZNF345 Chr22q11.22 813789 11 CN Gain 83.3 0.001 VPREB1, PRAME, BCR, POM121L1P, DQ597441, GGTLC2, DKFZp667J0810, MIR650, MIR5571, IGLL5, DQ575049

141

Table 4.7 Common copy number changes in neuroblastoma n=10

Chromosome Region Number of Event Frequency p-Value Gene Symbols Length (bp) genes %

Chr1p13.2 58755 1 CN Gain 50 0.018 TSPAN2 Chr2p24.3 273543 5 CN Gain 80 <0.001 LOC101926966, AK093525, MYCNUT, MYCNOS, MYCN Chr3q26.1 90142 1 CN Gain 40 <0.001 BC073807 Chr9q31.3 137178 1 CN Loss 40 0.048 SVEP1 Chr10q11.21 30154 1 CN Gain 40 <0.001 RET Chr13q14.2 5807 1 CN Gain 50 0.021 RB1 Chr14q32.33 81594 9 CN Gain 70 <0.001 IGH, DKFZp686O16217, IGHE, IGHG1, IGHD, FLJ00382, MIR4539, AK128652, KIAA0125

Chr15q25.2 164959 11 CN Gain 40 0.002 CPEB1, DQ601936, CPEB1-AS1, LOC283692, AP3B2, LOC338963, ACTG1P17, LOC283693, SNHG21, SCARNA15, FSD2 Chr19p13.12 189489 6 CN Loss 50 0.028 OR7C1, OR7A5, OR7A10, OR7A17, OR7C2, SLC1A6 Chr19q13.12 149777 7 CN Gain 40 <0.001 ZNF850, AX747375, LOC728485, BC024306, ZNF790- AS1, ZNF790, ZNF345 Chr20p13 42334 1 CN Loss 40 0.004 SIRPB1 Chr22q11.22 338246 12 CN Gain 70 <0.001 PRAME, LL22NC03-63E9.3, LOC648691, BCR, POM121L1P, DQ597441, GGTLC2, DKFZp667J0810, MIR650, MIR5571, IGLL5, DQ575049

142

Table 4.8 Common copy number changes in ARMS n=2

Chromosome Region Length Number Event Frequency % p-Value Gene Symbols of genes

Chr3p25.2 5133 1 CN Loss 100 0.001 RAF1 Chr3q26.1 124085 1 CN Gain 100 0.018 BC073807 Chr4p14 230779 2 CN Loss 100 0.029 UBE2K, PDS5A Chr5q21.3 3587 1 CN Gain 100 0.004 FER Chr5q31.1 97963 2 CN Loss 100 0.002 SAR1B, SEC24A Chr5q32 34588 2 CN Gain 100 0.004 HMGXB3, CSF1R Chr8p11.22 161902 2 CN Gain 100 0.005 ADAM5, ADAM3A Chr11p15.4 159414 3 CN Loss 100 0.007 DENND5A, TMEM41B, IPO7 Chr11q11 117974 4 CN Gain 100 0.025 OR4C11, OR4P4, OR4S2, OR4C6 Chr17q21.31 409201 6 CN Loss 100 0.039 KANSL1, KANSL1-AS1, LRRC37A, ARL17A, ARL17B, NSFP1 Chr19q13.12 232489 10 CN Gain 100 0.014 ZNF461, LINC01534, ZNF567, ZNF850, AX747375, LOC728485, BC024306, ZNF790- AS1, ZNF790, ZNF345 Chr20p13 42334 1 CN Loss 100 0.003 SIRPB1 ChrXp11.22 265786 2 CN Loss 100 0.011 PHF8, FAM120C ChrXq28 94845 4 CN Gain 100 0.01 IL9R, WASH1, WASH6P, DDX11L16

143

Table 4.9 Cancer genes with lower oligonucleotide probe spacing with recurrent gain/loss in each type of tumour

Osteosarcomas ERMS ARMS Neuroblastoma n=5 n=6 n=2 n=10

Copy Copy Copy Copy Chr.a Gene Number Chr. Gene Number Chr. Gene Number Chr. Gene Number

3p25.2 RAF Loss (5/5) 17p13.3 CRK Loss (5/6) 12q13.12 WNT1 Gain (2/2) 2p24.3 MYCN Gain (8/10) 5q22.2 APC Loss (5/5) Gain (1/6) 12q14.1 CDK4 Gain (2/2)

5q21.3 FER Loss (5/5) 13q13.1 BRCA2 Loss (4/5) Gain (1/5) 13q14.2 RB1 Loss (5/5)

17p13.1 TP53 Loss (5/5) aChr. = chromosome

144

4.3.3 Copy number changes by aCGH in genes previously examined by MLPA on chromosome 8 in 31 solid tumours

Most osteosarcomas showed no evidence of copy number gain or loss at the chromosome 8p loci examined using MLPA (Fig 4.3). 10/22 chromosome 8q genes were gained in 4/5 (80%) samples, and KCNK9 showed copy number gain in all osteosarcomas (Table 4.10). Gain of the entire chromosome 8 was evident in 3/8

Ewing sarcoma samples (Fig 4.4). The highest copy number gains were found at

CHD7, MYBL1 and PTP4A3 in 4/8 (50%) samples (Table 4.11). Results for rhabdomyosarcomas were not split into ARMS and ERMS samples as in previous section; these tumour types were predominantly disomic at chromosome 8 loci

(Fig 4.5), but RECQL4 showed copy number gain in 7/8 (87%) samples (Table 4.12).

In neuroblastoma samples, few chromosome 8 copy number gains were found (Fig

4.6), and 3 genes DLGAP2, EXT1 and PTP4A3A were gained in 4/10 (40%) samples (Table 4.13). Individual genome copy number results for each paediatric solid tumour are shown in Appendix 3.

4.3.4 Statistical Analysis for loci examined by MLPA

Application of the Fisher’s exact test to aCGH data for MLPA genes showed that for osteosarcoma, the numbers of 8p losses and 8q gains were significantly higher than expected. For the other tumour types, Fisher's exact test did not reach statistical significance (Table 4.14).

145

Table 4.10 Copy number changes by aCGH at chromosome 8 MLPA genes in osteosarcoma n=5

Gene Gain Disomy Loss

DLGAP2 1(20%) 4(80%) 0(0%) MFHAS1 2(40%) 3(60%) 0(0%) MSRA 1(20%) 4(80%) 0(0%) GATA4 1(20%) 4(80%) 0(0%) CTSB 0(0%) 5(100%) 0(0%) TUSC3 1(20%) 4(80%) 0(0%) ChGn 1(20%) 4(80%) 0(0%) FGFR1 1(20%) 3(60%) 1(20%) PRKDC 2(40%) 3(60%) 0(0%) MOS 2(40%) 1(20%) 2(40%) CHD7 3(60%) 1(20%) 1(20%) MYBL1 4(80%) 1(20%) 0(0%) NCOA2 2(40%) 1(20%) 2(40%) TPD52 3(60%) 1(20%) 1(20%) E2F5 1(20%) 3(60%) 1(20%) RAD54B 4(80%) 1(20%) 0(0%) ST7 3(60%) 1(20%) 1(20%) EIF3E 4(80%) 1(20%) 0(0%) EIF3H 4(80%) 1(20%) 0(0%) EXT1 4(80%) 1(20%) 0(0%) RNF139 3(60%) 2(40%) 0(0%) MYC 3(60%) 2(40%) 0(0%) DDEF1 2(40%) 3(60%) 0(0%) KCNQ3 4(80%) 1(20%) 0(0%) SLA 4(80%) 1(20%) 0(0%) KHDRBS3 4(80%) 1(20%) 0(0%) KCNK9 5(100%) 0(0%) 0(0%) PTK2 4(80%) 1(20%) 0(0%) PTP4A3 4(80%) 1(20%) 0(0%) RECQL4 3(60%) 2(40%) 0(0%)

146

Figure 4.3 Chromosome 8 copy number alterations measured in osteosarcoma

(n=5). Blue indicates copy number gain events shown above the baseline. Losses are

shown in red below the baseline. A. Compound plot showing frequency of gain/loss

along chromosome 8, shown as an ideogram above and also according to Mb position

on the X axis. The Y axis indicates the percentage of osteosarcomas showing copy

number change gain/loss at a specific point along chromosome 8. Dark blue and dark

red indicate more than two-fold copy gain and homozygous loss, respectively. B.

Individual copy number plots for each osteosarcoma. The Y axis represents the copy

number events for each osteosarcoma. The X axis shows copy number gain/loss along

chromosome 8 and the baseline indicates disomic copy number. The compound and

individual plots were generated using Nexus Copy number software v 7.5.

147

Table 4.11 Copy number changes by aCGH at chromosome 8 MLPA genes in

Ewing sarcoma n=8

Gene Gain Disomy Loss

DLGAP2 4(50%) 4(50%) 0(0%) MFHAS1 2(25%) 5(63%) 1(13%) MSRA 3(38%) 5(63%) 0(0%) GATA4 2(25%) 6(75%) 0(0%) CTSB 2(25%) 6(75%) 0(0%) TUSC3 3(38%) 5(63%) 0(0%) ChGn 3(38%) 5(63%) 0(0%) FGFR1 2(25%) 6(75%) 0(0%) PRKDC 2(25%) 6(75%) 0(0%) MOS 1(13%) 7(88%) 0(0%) CHD7 4(50%) 4(50%) 0(0%) MYBL1 4(50%) 4(50%) 0(0%) NCOA2 3(38%) 5(63%) 0(0%) TPD52 2(25%) 5(63%) 1(13%) E2F5 2(25%) 6(75%) 0(0%) RAD54B 2(25%) 6(75%) 0(0%) ST7 3(38%) 5(63%) 0(0%) EIF3E 3(38%) 5(63%) 0(0%) EIF3H 3(38%) 5(63%) 0(0%) EXT1 3(38%) 5(63%) 0(0%) RNF139 2(25%) 6(75%) 0(0%) MYC 2(25%) 6(75%) 0(0%) DDEF1 3(38%) 5(63%) 0(0%) KCNQ3 3(38%) 5(63%) 0(0%) SLA 3(38%) 5(63%) 0(0%) KHDRBS3 3(38%) 5(63%) 0(0%) KCNK9 2(25%) 6(75%) 0(0%) PTK2 3(38%) 5(63%) 0(0%) PTP4A3 4(50%) 4(50%) 0(0%) RECQL4 2(25%) 6(75%) 0(0%)

148

Figure 4.4 Chromosome 8 copy number alterations measured in Ewing sarcoma

(n=8). Blue indicates copy number gain events shown above the baseline. Losses are

shown in red below the baseline. A. Compound plot showing frequency of gain/loss

along chromosome 8, shown as an ideogram above and also according to Mb position

on the X axis. The Y axis indicates the percentage of Ewing sarcomas showing copy

number change gain/loss at a specific point along chromosome 8. Dark blue and dark

red indicate more than two-fold copy gain and homozygous loss, respectively. B.

Individual copy number plots for Ewing sarcoma. The Y axis represents the copy

number events for each Ewing sarcoma. The X axis shows copy number gain/loss

along chromosome 8 and the baseline indicates disomic copy number. The compound

and individual plots were generated using Nexus Copy number software v 7.5.

149

Table 4.12 Copy number changes by aCGH at chromosome 8 MLPA genes in rhabdomyosarcoma n=8

Genes Gain Disomy Loss

DLGAP2 2(25%) 6(75%) 0(0%) MFHAS1 0(0%) 8(100%) 0(0%) MSRA 1(13%) 7(87%) 0(0%) GATA4 0(0%) 8(100%) 0(0%) CTSB 0(0%) 8(100%) 1(13%) TUSC3 0(0%) 8(100%) 0(0%) ChGn 0(0%) 8(100%) 0(0%) FGFR1 0(0%) 7(87%) 1(13%) PRKDC 0(0%) 7(87%) 1(13%) MOS 0(0%) 8(100%) 0(0%) CHD7 0(0%) 8(100%) 0(0%) MYBL1 0(0%) 8(100%) 0(0%) NCOA2 1(13%) 6(75%) 1(13%) TPD52 2(25%) 5(62%) 1(13%) E2F5 1(13%) 6(75%) 1(13%) RAD54B 1(13%) 6(75%) 1(13%) ST7 1(13%) 7(87%) 0(0%) EIF3E 1(13%) 6(75%) 1(13%) EIF3H 1(13%) 7(87%) 0(0%) EXT1 1(13%) 7(87%) 0(0%) RNF139 0(0%) 7(87%) 1(13%) MYC 1(13%) 6(74%) 1(13%) DDEF1 0(0%) 7(87%) 1(13%) KCNQ3 2(25%) 6(75%) 0(0%) SLA 2(25%) 6(75%) 0(0%) KHDRBS3 2(25%) 6(75%) 0(0%) KCNK9 0(0%) 8(100%) 0(0%) PTK2 0(0%) 6(75%) 2(25%) PTP4A3 2(25%) 6(75%) 0(0%) RECQL4 7(87%) 1(13%) 0(0%)

150

Figure 4.5 Chromosome 8 copy number alterations measured in

rhabdomyosarcoma (n=8). Two cases (M27R and M32R) were ARMS samples, all

others were ERMS samples. Blue indicates copy number gain events shown above the

baseline. Losses are shown in red below the baseline. A. Compound plot showing

frequency of gain/loss along chromosome 8, shown as an ideogram above and also

according to Mb position on the X axis. The Y axis indicates the percentage of

rhabdomyosarcomas showing copy number change gain/loss at a specific point along

chromosome 8. Dark blue and dark red indicate more than two-fold gain and

homozygous loss, respectively. B. Individual copy number plots for rhabdomyosarcomas.

The Y axis represents the copy number events for each rhabdomyosarcomas. The X axis

shows copy number gain/loss along chromosome 8 and the baseline indicates disomic

copy number. The compound and individual plots were generated using Nexus Copy

number software v 7.5.

151

Table 4.13 Copy number changes by aCGH at chromosome 8 MLPA genes in neuroblastoma n=10

Gene Gain Disomy Loss

DLGAP2 4(40%) 6(60%) 0(0%) MFHAS1 1(10%) 9(90%) 0(0%) MSRA 2(20%) 8(80%) 0(0%) GATA4 1(10%) 9(90%) 0(0%) CTSB 1(10%) 9(90%) 0(0%) TUSC3 1(10%) 9(90%) 0(0%) ChGn 1(10%) 9(90%) 0(0%) FGFR1 2(20%) 8(80%) 0(0%) PRKDC 2(20%) 8(80%) 0(0%) MOS 2(20%) 8(80%) 0(0%) CHD7 1(10%) 9(90%) 0(0%) MYBL1 2(20%) 7(70%) 1(10%) NCOA2 1(10%) 9(90%) 0(0%) TPD52 1(10%) 9(90%) 0(0%) E2F5 2(20%) 8(80%) 0(0%) RAD54B 1(10%) 8(80%) 1(10%) ST7 1(10%) 9(90%) 0(0%) EIF3E 1(10%) 9(90%) 0(0%) EIF3H 1(10%) 9(90%) 0(0%) EXT1 4(40%) 6(60%) 0(0%) RNF13 2(20%) 8(80%) 0(0%) MYC 2(20%) 8(80%) 0(0%) DDEF1 2(20%) 8(80%) 0(0%) KCNQ3 2(20%) 8(80%) 0(0%) SLA 2(20%) 8(80%) 0(0%) KHDRBS3 1(10%) 9(90%) 0(0%) KCNK9 3(30%) 7(70%) 0(0%) PTK2 2(20%) 8(80%) 0(0%) PTP4A3 4(40%) 6(60%) 0(0%) RECQL4 3(30%) 6(60%) 1(10%)

152

Figure 4.6 Chromosome 8 copy number alterations measured in neuroblastoma

(n=10). Blue indicates copy number gain events shown above the baseline. Losses are

shown in red below the baseline. A. Compound plot showing frequency of gain/loss

along chromosome 8, shown as an ideogram above and also according to Mb position

on the X axis. The Y axis indicates the percentage of neuroblastomas showing copy

number change gain/loss at a specific point along chromosome 8. Dark blue and dark

red indicate more than two-fold copy gain and homozygous loss, respectively. B.

Individual copy number plots for neuroblastomas. The Y axis represents the copy

number events for each neuroblastoma. The X axis shows copy number gain/loss along

chromosome 8 and the baseline indicates disomic copy number. The compound and

individual plots were generated using Nexus Copy number software v 7.5.

153

Table 4.14 Statistical analysis of chromosome 8p versus 8q using Fisher’s exact test for each type of tumour p-values ≤ 0.05 were considered statistically significant

Categories: 2=Gain; 1=Disomy; 0=Loss

Osteosarcoma Category 2 Category 1 Category 0 8p 8 31 1 8q 72 30 8 p=1.1E-7

Ewing sarcoma Category 2 Category 1 Category 0 8p 21 42 1 8q 59 116 1 p=0.60

ARMS Category 2 Category 1 Category 0 8p 0 14 2 8q 1 33 10 p=0.4695

ERMS Category 2 Category 1 Category 0 8p 3 45 0 8q 24 107 1 p=0.0678

Neuroblastoma Category 2 Category 1 Category 0 8p 13 67 0 8q 42 175 3 p=0.59

154

4.4 Discussion

The importance of chromosome 8 in different types of cancer was revealed in the study of Myllykangas et al. (2006) where they performed a bibliomics survey using

838 published chromosomal CGH results for 73 distinct neoplasms. These analyses showed that 8p23-q12 and 8q23 are hot spot areas, with a significant overrepresentation of cancer genes. Chromosome 8 was one of the most frequently amplified chromosomes across all tumour types (Myllykangas et al., 2006).

4.4.1 Chromosome 8 copy number changes in breast cell lines

Cancer cell lines have been used to investigate cancer pathobiology as well as to test new therapies (Kao et al., 2009). Cell lines are also used to study genomic alterations assuming that they reflect the genotype and phenotype of the primary tumour. Tsuji et al. (2010) studied the differences and similarities between cell lines and primary tumour tissue using aCGH, and concluded that established cell lines carry cell line-specific DNA copy number aberrations together with recurrent aberrations detected in primary tumour tissues. Structural and numerical alterations of chromosome 8 in breast cell lines have been reported, including partial or complete deletions of 8p and gains of 8q (Rummukainen et al., 2001) (Table 3.2).

In this project, 5 breast cell lines showed copy number loss for almost every MLPA locus on 8p, except TUSC3 in HCC2157 cells. MCF-10A cells have been previously reported to carry 2 copies of chromosome 8p (Marella et al., 2009), which was confirmed in the present study (Fig 4.2). In SK-BR-3 cells, two amplicons have been reported, namely 8q21-23, that contains putative target genes TPD52 and WWP1,

155

and 8q24 that contains MYC (Rodriguez-Pinilla et al., 2007; Ruiz et al., 2006). Copy number gain was also measured at TPD52 and MYC in SK-BR-3 cells using aCGH

(Table 4.3). Kao et al. (2009) studied 52 breast cancer cell lines using aCGH, which include 5/6 cell lines used in the present study, and 25/30 genes examined by

MLPA. In most cases, 104/125 (83%), there was agreement between their findings

(data not shown) and our results (Kao et al., 2009), suggesting that reproducible aCGH results were generated in the present study.

4.4.2 Common regions of copy number change in paediatric solid tumours

Chromosome 22q11.22 was commonly gained in the paediatric solid tumours analysed, and harbours genes such as PRAME, VPREB1, and POM121L1P, among others (Tables 4.4, 4.5, 4.6, and 4.7). PRAME, or preferentially expressed antigen in melanoma, is an attractive potential target for tumour immunotherapy as PRAME is overexpressed in many solid tumours and haematological malignancies whilst showing minimal expression in normal tissues (Bankovic et al., 2010). For example,

PRAME transcripts can be detected by QRT-PCR analysis in approximately 20–30% of patients with chronic leukaemia, but not in normal donors (Paydas et al., 2007).

High expression of PRAME has been associated with poor prognosis in solid tumours such as neuroblastoma and breast cancer (Doolan et al., 2008). Specifically in neuroblastoma patients, high expression of PRAME is linked with advanced tumour stage, older age at diagnosis, and poor clinical outcome (Doolan et al., 2008;

Oberthuer et al., 2004). Tan et al. (2012) reported that PRAME was expressed in five osteosarcoma cell lines and in more than 70% of osteosarcoma patient specimens.

Furthermore, PRAME siRNA knockdown significantly suppressed proliferation,

156

colony formation, and G1 cell cycle arrest in U-2OS cells, thus suggesting that

PRAME plays an important role in cell proliferation and disease progression in osteosarcoma (Tan et al., 2012). Although no literature was found relating Ewing sarcoma or rhabdomyosarcoma with PRAME, the reported studies in other solid tumours indicate this gene could be a relevant target for future studies. The other genes in this region have not been explored in the present study.

The criteria to select genes from the list of genes important in cancer (Table 4.1) were consistency in gain or loss in each tumour type, high frequency, p-value and literature that supported importance and novelty (Table 4.9). Osteosarcomas (n=5) presented copy number loss at 6 cancer genes: RAF (3p25.2), involved in a signalling pathway as an important factor of osteosarcoma growth and metastasis

(Yu et al., 2011), APC (5q21), RB1 (13q14.2) and TP53 (17p13.1) widely known to be involved in osteosarcoma tumourigenesis and progression (Kansara et al., 2014;

Martin et al., 2012), and BRCA2 (13q13.1) and FER (5q21.3). No previous literature describes these last 2 genes in osteosarcoma.

In ERMS (n=6), CRK (17p13.3) was lost in 5/6 cases. Previously implicated in cancers such as lung, breast, gliomas, gastric cancer, sarcomas, ovarian and hematopoietic cancers (Sriram et al., 2010), CRK is a proto-oncogene that encodes a member of an adapter protein family that binds to several tyrosine-phosphorylated proteins (Park et al., 2008); no literature was found relating CRK with ERMS. The genomic region 12q13-q14 that includes WNT1 (12q13.12) and CDK4 (12q14.1) showed copy number gain in alveolar rhabdomyosarcomas (n=2) (Table 4.8).

Amplification of this genomic region has been found in diverse tumours such as lung

157

cancer (Myllykangas et al., 2007), liposarcomas (Shimada et al., 2006), and anaplastic astrocytoma (Backlund et al., 2005). Barr et al. (2009) reported that a

12q13-q14 region was amplified preferentially in PAX3-FOXO1-positive ARMS cases, and was associated with significantly worse progression and overall survival that was independent of gene fusion status.

4.4.3 Copy number changes in chromosome 8 genes examined by MLPA in the solid tumour cohort

The solid tumour showing the most chromosome 8 copy number changes was osteosarcoma, predominantly copy number gains at chromosome 8q. This is consistent with the literature that classifies osteosarcomas as a solid tumour with an unusually high frequency of genome rearrangements, and the presence of a complex karyotype as hallmarks (Kansara et al., 2014; Man et al., 2004).

From the list of genes on chromosome 8 examined by MLPA, only KNCK9 (8q24.3) showed copy number gain in all 5 osteosarcomas. KNCK9, or potassium channel gene subfamily K member 9, has been associated with tumourigenesis of osteosarcoma, after induced fluctuations in extracellular pH significantly reduced the proliferation rate of MG63 human osteoblast-like cells (Li et al., 2013). Not many recent studies correlate copy number gain of this gene with tumourigenesis or progression of this solid tumour, thus KCNK9 may represent a candidate gene for further study.

158

Ewing sarcomas showed trisomy of chromosome 8 in 3/8 cases, which is one of the characteristics of this solid tumour, albeit typically reported in 50-60 % of cases

(Mackintosh et al., 2010).

In rhabdomyosarcoma, RECQL4 (8q24.3) showed copy number gain in 7/8 (83%) cases, representing 6/6 ERMS and 1/2 ARMS. The protein encoded by RECQL4 has been associated with the formation of a member of the RecQ helicases. The helicases bind to the DNA and temporarily unwind the two double DNA helix, participates in cell cycle control/checkpoint and DNA damage repair (Nishijo et al.,

2004). Not much literature was found relating to this gene with rhabdomyosarcoma, although germline mutations in RECQL4 have been associated with osteosarcoma risk (Hicks et al., 2007). Amplification and overexpression of RECQL4 have been reported in colorectal (Buffart et al., 2005), breast (Fang et al., 2013), laryngeal

(Saglam et al., 2007), and cervical cancers (Narayan et al., 2007).

In neuroblastoma, no chromosome 8 gene presented copy number gain or loss for the majority of samples, with most genes showing a disomic pattern in most (9/10) samples. Some authors attribute the flat or disomic pattern of neuroblastoma with a higher amount of Schwann cell stroma (more than 50%), this being composed of normal cells in differentiating/maturing neuroblastomas with diploid DNA content and lack of aberrations (Ambros et al., 2011). However in this project we did not performed any pathology test showing the stromal cell percentage to be sure about this fact.

159

4.4.4 Limitations

The size of the tumour cohort was a limiting factor, since finding genes with copy number change, with statistical significance, was a challenge in such a small population. While we nonetheless identified numerous significant peaks of gain or loss in the tumour types analysed (Tables 4.4 - 4.8), including genes of known significance such as MYCN in neuroblastoma (Table 4.7), the small sample numbers reduced our ability to be confident of the significance of many of these results.

The aCGH technique also presents inherent limitations, as it is unable to detect balanced rearrangements as translocations and inversions. Thus we could have overlooked important genes related to the four cancers that we studied. Such is the case in ARMS where the main characteristic is a reciprocal chromosomal translocation t(2;13) in approximately 70% of the ARMS cases, and a less common t(1;13) variant detected in approximately 10% of ARMS (Wang, 2012). Similarly, over

80% of Ewing sarcomas present with t(11;22) (Jahromi et al., 2012). Finally, the need for specialised equipment, materials and processes to perform aCGH makes this technique still unavailable to many scientists, and expensive for regular clinical application.

4.4.5 Summary and conclusions

aCGH is a complex technique that requires prior basic cytogenetic knowledge and attention to detail. Performing aCGH on well-studied breast cell lines produced results with good concordance with the literature, indicating that the technique performed correctly in our hands. Our custom design that focused attention on 95

160

genes with known importance in cancer also facilitated comparisons between our results and these of previous studies. However, it is an undeniable fact that molecular techniques such as aCGH generate lots of data, which in turn are processed by means of complex algorithms. In this deluge of information, we are confronted with the problem of high dimensional statistics and performing such analyses is not an easy task. We therefore relied on specialised software packages, an area that is also evolving quickly.

In the next chapter of this dissertation, aCGH copy number data for specific chromosome 8 genes will be compared with the results of quantitative PCR, and our previous MLPA results.

161

Chapter 5

Comparison of 3 techniques for the assessment of

DNA copy number as applied to solid paediatric

tumour samples and breast cell lines

162

5.1 Introduction

Throughout this project we have presented literature supporting the notion that, for chromosome 8, copy number variation can range from whole copies of chromosome 8, through whole chromosome arms, to individual genes. In various ways, each has the potential to make some contribution to the development of cancer. Chromosome 8 alterations in childhood cancers include the gain of the entire chromosome 8 in Ewing sarcoma (Toomey et al., 2010), translocation t(2;8) producing the fusion gene PAX3/NCOA2 in rhabdomyosarcomas (Sumegi et al.,

2010), t(8;14), t(2;8), t(8;14;18) and (8;22) translocations in Burkitt lymphoma

(Schmitz et al., 2012), and recurrent focal gains of 8q24 in acute myeloid leukaemia (Kuhn et al., 2012).

Techniques to identify alterations in DNA copy number include MLPA, aCGH, quantitative PCR (qPCR) and next generation sequencing (Carter, 2007). aCGH profiling is currently a gold standard technique for measurement of DNA copy number (Hayes et al., 2013). aCGH targets thousands of genomic sites simultaneously in a single patient, allowing broad coverage of the genome.

However, possible disadvantages include expense, and the considerable amount of time required to analyse the data. MLPA has emerged as an alternative, being a relatively economic and rapid technique that can be applied to many samples simultaneously, with the constraint that the number of genes that can be assayed at once is limited (Stuppia et al., 2012).

163

To complement our research in this area (chapters 3 and 4), we used a third technique, qPCR, to assess chromosome 8 copy number at particular loci in a subset of the samples analysed. This technique presents many advantages, such as its low consumable and instrumentation costs, fast turnaround and assay development time, and high sensitivity among others (Table 1.10). These qPCR experiments were performed to resolve discrepancies in copy number results, as measured by aCGH and MLPA. This process of resolution will be described below.

164

5.2 Aims

1. To compare the results obtained by the MLPA “Salsa kit P014-A1” for

chromosome 8q loci with results obtained for the same loci using aCGH.

2. To perform qPCR for a subset of genes analysed by both MLPA and aCGH,

and compare the results obtained by these 3 techniques.

3. To identify the most reliable technique for assessing chromosome 8 copy

number status in childhood solid tumours.

165

5.3 Results

Comparison of MLPA versus aCGH copy number results was made based on the results for the 30 genes included in the MLPA kit (Table 2.3). The genes MYC and

RECQL4 each included two sets of MLPA probes per gene. In order to compare the results from MLPA with the results from aCGH, results obtained for the two sets of MLPA probes were averaged for each gene.

In the case of aCGH we used a customized 60-mer-oligonucleotide high density array with 52,639 probes mapping throughout chromosome 8, with approximately even spacing. In the initial stage aCGH data were analysed using the Agilent proprietary software “CytoGenomics”. While this could visualize the basic aCGH results, this software was not sufficiently versatile for our final analysis. “Nexus v7.5”, on the other hand, offered better flexibility, and plotting functionality at frequencies not available in CytoGenomics. Both packages aim to detect and quantify copy number by calculating the ratio of signal intensities obtained for test

DNA versus reference DNA. Nexus uses the intensity values to arrive at log2 ratios that are segmented hierarchically into neighbouring regions, using

BioDiscovery’s “Rank Segmentation” algorithm. At the completion of this process, the entire genome can be represented as a series of segments, with each segment having a “cluster value” which is the median log2 ratio value of all the probes in that region.

To facilitate comparison of data obtained via two different numeric scales (MLPA uses a linear scale, whereas aCGH uses a logarithmic scale), copy number data

166

from MLPA and aCGH analyses were converted to 3 categories; gain (coded as

2), normal range (or disomy, coded as 1) and copy number loss (coded as 0).

Category tables (see chapter 3 for MLPA, and Appendix 4 for aCGH) for each technique were used to construct a summary table that compares agreements and disagreements between the results obtained by MLPA and aCGH (Table 5.1).

5.3.1 Comparison of copy number results obtained using MLPA or aCGH across all measured chromosome 8 loci in 6 breast cell lines

When comparing MLPA and aCGH results for loci of chromosome arm 8p and 8q loci for the 6 breast cell lines, all chromosome 8p loci (6 cell lines x 8 genes = 48 data points) showed concordant results using MLPA and aCGH (Table 5.2).

However, for chromosome 8q loci the percentage of concordance was only 53%

(70 concordant results /22 chromosome 8q genes x 6 cell lines) (Table 5.2).

Examples of two cell lines comparing results obtained by MLPA and aCGH are presented in Figures 5.1 and 5.2.

167

Table 5.1 Agreement and disagreement between MLPA and aCGH copy number results in breast cell lines (n=6)

8p (8 genes) Chromosome 8q (22 genes)

Genes è

ST7 SLA MYC MOS E2F5 EXT1 PTK2 ChGn CTSB CHD7 EIF3E MYBL EIF3H

MSRA TPD52 PTP4A FGFR1 DDEF1 TUSC3 GATA4 KCNK9 KCNQ3 NCOA2 PRKDC MFHAS RNF139

RAD54B RECQL4 DLGAP2

Cell lines ê KHDRBS3

AU565 A1 A A A A A A A D2 A A A D A D A A A A A A A A A D D A A D D HCC2157 A A A A A A A A A A A A A A A A D A A D D A D D D D D D D A MCF7 A A A A A A A A A A A A A A A A A A D D A A A D D D D D D D MCF10A A A A A A A A A A A A A A A A A D D D D D A D D A D D D D A MDAMB231 A A A A A A A A A D A A D A D D D D D D A D A D D D D D D A SKBR3 A A A A A A A A D D D D D A D A A A A A A A A D A D A A D D

Agreement % 100 100 100 100 100 100 100 100 67 67 83 83 50 100 50 83 50 67 50 33 67 83 67 17 33 0 33 33 0 50

1A= Agreement,

2D=Disagreement

168

Table 5.2 Summary of agreement between MLPA and aCGH results obtained In 6 breast cell lines at chromosome 8p and 8q loci

Number of genes Agreement Chromosome 8 X 6 cell lines %

8p 48 48 (100%) (8 genes)

8q 132 70 (53%) (22 genes)

8p + 8q 180 118 (66%)

169

MLPA Copy number Categoriesnumber Copy

aCGH 0 ST7 SLA SLA MOS MOS MYC EXT1 PTK2 E2F5 ChGn CTSB CHD7 EIF3E EIF3E MYBL MYBL EIF3H MSRA MSRA TPD52 GATA4 GATA4 PTP4A PTP4A FGFR1 TUSC3 DDEF1 KCNK9 NCOA2 KCNQ3 MFHAS PRKDC RNF139 RAD54B DLGAP2 RECQL4 KHDRBS3

8p genes 8q genes 8pter 8qter

Figure 5.1 Comparison of copy number predictions obtained using MLPA

(orange) and aCGH (blue) at chromosome 8p and 8q loci in MCF-10A cells.

Copy number categories are shown on the Y axis (0 = copy number loss, 1 = disomy, 2 = copy number gain), whereas the loci examined are shown on the X axis.

170

MLPA 1 aCGH Copy number Categoriesnumber Copy 0 ST7 SLA SLA MOS MOS MYC EXT1 PTK2 E2F5 ChGn CTSB CHD7 EIF3E EIF3E MYBL MYBL EIF3H MSRA MSRA TPD52 GATA4 GATA4 FGFR1 TUSC3 DDEF1 PTP4A PTP4A KCNK9 NCOA2 KCNQ3 MFHAS PRKDC RNF139 RAD54B DLGAP2 RECQL4 KHDRBS3

8p genes 8q genes 8pter 8qter

Figure 5.2 Comparison of copy number predictions obtained using MLPA

(orange) and aCGH (blue) at chromosome 8p and 8q loci in SK-BR-3 cells.

Copy number categories are shown on the Y axis (0 = copy number loss, 1 = disomy, 2 = copy number gain), whereas the loci examined are shown on the X axis.

171

5.3.2 Comparison of MLPA and aCGH copy number results obtained for the paediatric solid tumour cohort, according to tumour type

As comparison of MLPA and aCGH results revealed different levels of agreement for chromosome 8p versus 8q loci in the breast cell lines, we performed similar analyses for the paediatric solid tumour types examined in this study. For osteosarcoma samples (n=5) there was 60% (24/40) agreement between MLPA and aCGH results for chromosome 8p loci, and 44% (48/110) agreement for chromosome 8q loci. The total percentage of agreement was 48% (72/150) for all chromosome 8 loci (Tables

5.3 and 5.7). Results for a single (but representative) osteosarcoma case are shown in Figure 5.3.

In Ewing sarcoma (n=8), the percentage agreement was 61% (39/64) for 8p loci and

63% (110/176) for 8q loci, a total of 62% agreement for all chromosome 8 loci

(Tables 5.4 and 5.7). Again, the results for a single Ewing sarcoma case are shown as an example in Figure 5.4.

For rhabdomyosarcoma (n=8) we noted 88% (56/64) agreement between MLPA and aCGH results at chromosome 8p loci, 75% (132/176) agreement for chromosome 8q loci, and 78% (188/240) agreement for all chromosome 8 loci (Tables 5.5 and 5.7).

Results for a ERMS case are shown in Figure 5.5 as an example. Finally, for neuroblastoma (n=10), 58% (46/80) agreement was noted between MLPA and aCGH results for chromosome 8p loci, and 62% (136/220) agreement was noted between MLPA and aCGH results for chromosome for 8q loci, with a total of 61%

172

(182/300) agreement noted for all chromosome 8 loci (Tables 5.6 and 5.7). Results for a single neuroblastoma case are shown in Figure 5.6.

173

Table 5.3 Agreement and disagreement between MLPA and aCGH copy number results in osteosarcoma (n=5)

8p (8 genes) Chromosome 8q (22 genes)

Genes è

ST7

SLA MYC MOS E2F5 EXT1 PTK2 ChGn CTSB CHD7 EIF3E MYBL EIF3H MSRA TPD52 PTP4A FGFR1 DDEF1 TUSC3 GATA4 KCNK9 KCNQ3 NCOA2 PRKDC MFHAS RNF139 RAD54B RECQL4 DLGAP2 Samples ê KHDRBS3 F9O A1 A A A A A A A A D2 D D D A D D D D D D D D D D D D D D D D F10O A A A A A A A D A A D A D D A A A A A A A A A A D A D D D D M11O D D A D A A A D A D A A D D A D D D D D A A A D D D D A A D F12O D D D D D A D D D D A A A A A D A A A A A A A D D D D D D A M46O D D D D A A A A D A D D A D A A A A A A D D D D A A D D D D Agreement % 40 40 60 40 80 100 80 40 60 40 40 60 40 40 80 40 60 60 60 60 60 60 60 20 20 40 0 20 20 20

1A= Agreement

2D=Disagreement

174

1 MLPA

aCGH Copy number Categoriesnumber Copy

0 ST7 SLA SLA MYC MOS MOS E2F5 EXT1 PTK2 ChGn CTSB CHD7 MYBL MYBL EIF3E EIF3E EIF3H MSRA MSRA TPD52 GATA4 GATA4 PTP4A PTP4A FGFR1 TUSC3 DDEF1 KCNK9 NCOA2 KCNQ3 MFHAS PRKDC RNF139 RAD54B DLGAP2 RECQL4 KHDRBS3

8p genes 8q genes 8pter 8qter

Figure 5.3 Comparison of copy number predictions obtained using MLPA

(orange) and aCGH (blue) at chromosome 8p and 8q loci in osteosarcoma

M11O case. Copy number categories shown on the Y axis (0 = copy number loss, 1

= disomy, 2 = copy number gain), whereas the loci examined are shown on the X axis.

175

Table 5.4 Agreement and disagreement between MLPA and aCGH copy number results in Ewing sarcoma (n=8)

8p (8 genes) Chromosome 8q (22 genes)

Genes è

ST7 SLA

MYC MOS E2F5 EXT1 PTK2 ChGn CTSB CHD7 EIF3E MYBL EIF3H MSRA TPD52 PTP4A FGFR1 DDEF1 TUSC3 GATA4 KCNK9 KCNQ3 NCOA2 PRKDC MFHAS RNF139 RAD54B RECQL4 Samples ê DLGAP2 KHDRBS3

M13E D2 A1 A A A A A A A A D D A A A A A A A A A A A A A A A A D D M14E D A D A A A D A A A D D D A A D D A D A A A D D D D A D A A F15E D D D D D D D D D D D D D D D D D D D D D D D D D D D D D A M16E D D D D D D D D D A D D D D D A D D D D D D D D D D D D D A F17E A A A A A D A A A A D A A D A A A A A A A A A A D A D A D A F18E A D A A D A A A A A A A A A A A A D A A A A A A A A A A A A M22E A D A A A D A A A A A A A D A A A A A A A A A A D A D A A A F24E A A A A A A A A A A A A A A A A A A A D A A A A A A A A A D Agreement % 50 50 62.5 75 62.5 50 62.5 75 75 87.5 37.5 50 62.5 50 75 75 62.5 62.5 62.5 62.5 75 75 62.5 62.5 37.5 62.5 50 62.5 50 75

1A= Agreement

2D=Disagreement

176

MLPA

Copy number Categoriesnumber Copy aCGH 0 ST7 SLA SLA MYC MOS MOS E2F5 EXT1 PTK2 ChGn CTSB CHD7 EIF3E EIF3E MYBL MYBL EIF3H MSRA MSRA TPD52 GATA4 GATA4 FGFR1 TUSC3 DDEF1 PTP4A PTP4A KCNK9 NCOA2 KCNQ3 MFHAS PRKDC RNF139 RAD54B DLGAP2 RECQL4 KHDRBS3

8p genes 8q genes 8pter 8qter

Figure 5.4 Comparison of copy number predictions obtained using MLPA

(orange) and aCGH (blue) at chromosome 8p and 8q loci in Ewing sarcoma

M22E case. Copy number categories shown on the Y axis (0 = copy number loss,

1 = disomy, 2 = copy number gain), whereas the loci examined are shown on the X axis.

177

Table 5.5 Agreement and disagreement between MLPA and aCGH copy number results in rhabdomyosarcoma (n=8)

8p (8 genes) Chromosome 8q (22 genes)

Genes è

ST7 SLA MYC

MOS E2F5 EXT1 PTK2 ChGn CTSB CHD7 EIF3E MYBL EIF3H MSRA TPD52 PTP4A FGFR1 DDEF1 TUSC3 GATA4 KCNK9 KCNQ3 NCOA2 PRKDC MFHAS RNF139 RAD54B RECQL4 Samples ê DLGAP2 KHDRBS3

M25R D2 A1 D A A A A A A A A A A A A A A A A A A A A D D D A A A D F26R A A A A A A A A A A A A A A A A A A A A A A A A A A A A D D M27R A A A A D A A D D A A A D D D D A D A A D D D A A A A D A A F28R A A A A A D A A A A A A A A A A A A A A A A A A A A A D A D F29R D A A A A A A A A A A A A A D A D D A D A A A A A A D A D D M31R A A A A A D A A A A D A D A A A A A D D A A A A D D A A A D M32R A A A A A A A A A A A A A A A A A A A D A A A A A A A A A D M35R A A A A A D A A A D A D A D A D A A A D A A A D D A D A D D Agreement % 75 100 87.5 100 87.5 62.5 100 87.5 87.5 87.5 87.5 87.5 75 75 75 75 87.5 75 87.5 50 87.5 87.5 87.5 75 62.5 75 75 75 62.5 12.5

1A= Agreement

2D=Disagreement

178

MLPA

aCGH Copy number Categoriesnumber Copy 0 ST7 SLA SLA MOS MOS MYC E2F5 EXT1 PTK2 ChGn CTSB CHD7 EIF3E EIF3E MYBL MYBL EIF3H MSRA MSRA TPD52 GATA4 GATA4 PTP4A PTP4A FGFR1 TUSC3 DDEF1 KCNK9 NCOA2 KCNQ3 MFHAS PRKDC RNF139 RAD54B DLGAP2 RECQL4 KHDRBS3

8p genes 8q genes 8pter 8qter

Figure 5.5 Comparison of copy number predictions obtained using MLPA

(orange) and aCGH (blue) at chromosome 8p and 8q loci in ERMS F29R case. Copy number categories shown on the Y axis (0 = copy number loss, 1 = disomy, 2 = copy number gain), whereas the loci examined are shown on the X axis.

179

Table 5.6 Agreement and disagreement between MLPA and aCGH copy number results in neuroblastoma (n=10)

8p (8 genes) Chromosome 8q (22 genes)

Genes è

ST7 SLA

MYC MOS E2F5 EXT1 PTK2 ChGn CTSB CHD7 EIF3E MYBL EIF3H MSRA TPD52 PTP4A FGFR1 DDEF1 TUSC3 GATA4 KCNK9 KCNQ3 NCOA2 PRKDC MFHAS RNF139 RAD54B RECQL4 Samples ê DLGAP2 KHDRBS3

1 2 M36N A A A A A A D A A A A A A A A A A A A A A A A A A A A A D A F37N A D D A D A D D D A A D A D D A D A D D D A A D A D A A A D F38N D A D A A A D A A A A D A A A A A A D D A A A A A A A A A D M39N D D D A A A D D D D D D A D D A D A D D A D D D D D A D D D F40N D A A A A A A A A A A D A A A A A A A A A A A A A A D A D D M41N D A A A A D A A A A A A A D A A A A A A A A A A A A D A D D F42N A A A A A D A A A A A A A D A A A A A A A A A A A A D A A A M43N A A A A A A A A A A A A A A A A A A A A A A A A A A A A A D F44N D D D D D D D D D D D D D D D D D D D D D D D D D D D D D D F45N D D D D D D D D D D A D D D A A D D D D D A D D D D D D D D Agreement % 40 60 50 80 70 60 40 60 60 70 80 40 80 40 70 90 60 80 50 50 70 80 70 60 70 60 50 70 40 20

1A= Agreement

2D=Disagreement

180

1 MLPA

8p genes 8q genes 8pter 8qter

Figure 5.6 Comparison of copy number predictions obtained using MLPA

(orange) and aCGH (blue) at chromosome 8p and 8q loci in neuroblastoma

M39N case. Copy number categories shown on the Y axis (0 = copy number loss,

1 = disomy, 2 = copy number gain), whereas the loci examined are shown on the X axis.

181

Table 5.7 Summary of agreement between MLPA and aCGH results obtained in the overall solid tumour cohort (n=31)

Osteosarcoma n=5 Ewing sarcoma n=8 Rhabdomyosarcoma n=8 Neuroblastoma n=10

N˚ concordant results Agreement N˚ concordant results Agreement N˚ concordant results Agreement N˚ concordant results Agreement N˚ genes x N˚ tumours (%) N˚ genes x N˚ tumours (%) N˚ genes x N˚ tumours (%) N˚ genes x N˚ tumours (%)

Chr 8

8p 24/40 60% 39/64 61% 56/64 88% 46/80 58% (8 genes)

8q 48/110 44% 110/176 63% 132/176 75% 136/220 62% (22 genes)

8p + 8q 72/150 48% 149/240 62% 188/240 78% 182/300 61%

182

5.3.3 Selection of genes for qPCR

It was clear that there were disagreements between the chromosome 8p and 8q copy number results obtained by MPLA and aCGH techniques (Table 5.7,

Appendixes 5 and 6). To quantify differences and propose genes to be used for copy number validation using a third technique, MLPA and aCGH results were plotted using the software package “MS Excel”, to calculate differences between the copy number results of both techniques.

To clarify which genes presented results with a greater degree of disagreement, we calculated the average distance between copy number categories for each gene. For each patient and gene, if the copy number categories predicted by both techniques were the same, the difference was 0. If one technique predicted disomic copy number and the other predicted either gain or loss, the difference was 1. If one technique predicted copy number gain and the other predicted copy number loss, the difference was 2 (Table 5.8). Mean values across all patients for each gene gave us the mean difference per gene (Table 5.9), where results were then sorted in descending order. RECQL4 presented with the largest difference (0.71), followed by

PTP4A3 (0.61), KCNK9 (0.58), EXT1 and TPD52 (0.48 for both). Accordingly, these genes were chosen for validation with an independent technique, namely quantitative PCR. We also selected MYC as a control because this gene presented the least mean difference between the techniques (0.23) (Table 5.9).

183

Table 5.8 Summary of agreement/disagreement between copy number results obtained using MLPA or aCGH for chromosome 8 loci

8p (8 genes) Chromosome 8q (22 genes)

Genes è

ST7 SLA MYC MOS E2F5 EXT1 PTK2 ChGn CTSB CHD7 EIF3E MYBL EIF3H

MSRA TPD52 PTP4A FGFR1 DDEF1 TUSC3 GATA4 KCNK9 KCNQ3 NCOA2 PRKDC MFHAS RNF139 RAD54B RECQL4 DLGAP2

Samples KHDRBS3 ê a b c F9O 0 0 0 0 0 0 0 0 0 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 F10O 0 0 0 0 0 0 0 1 0 0 1 0 1 1 0 0 0 0 0 0 0 0 0 0 1 0 1 1 1 1 M11O 1 1 0 1 0 0 0 1 0 1 0 0 1 1 0 1 1 1 1 1 0 0 0 1 1 1 1 0 0 1 F12O 1 1 1 1 1 0 1 1 1 1 0 0 0 0 0 1 0 0 0 0 0 0 0 1 1 1 2 2 2 0 M46O 1 1 1 1 0 0 0 0 1 0 1 1 0 1 0 0 0 0 0 0 1 1 1 1 0 0 1 1 1 1 M13E 1 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 M14E 1 0 1 0 0 0 1 0 0 0 1 1 1 0 0 1 1 0 1 0 0 0 1 1 1 1 0 1 0 0 F15E 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 M16E 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 0 F17E 0 0 0 0 0 1 0 0 0 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 1 0 1 0 F18E 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 M22E 0 1 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 0 F24E 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1 M25R 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 0 0 0 1 F26R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 M27R 0 0 0 0 1 0 0 1 1 0 0 0 1 1 1 1 0 1 0 0 1 1 1 0 0 0 0 1 0 0 F28R 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 F29R 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 1 0 1 0 0 0 0 0 0 1 0 1 1 M31R 0 0 0 0 0 1 0 0 0 0 1 0 1 0 0 0 0 0 1 1 0 0 0 0 1 1 0 0 0 1 M32R 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1 M35R 0 0 0 0 0 1 0 0 0 1 0 1 0 1 0 1 0 0 0 1 0 0 0 1 1 0 1 0 2 1 M36N 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 F37N 0 1 1 0 1 0 1 1 1 0 0 1 0 1 1 0 1 0 1 2 2 0 0 1 0 1 0 0 0 1 F38N 1 0 1 0 0 0 1 0 0 0 0 1 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 2 M39N 1 1 2 0 0 0 1 1 2 1 1 2 0 1 2 0 1 0 1 1 0 1 1 1 1 1 0 1 1 1 F40N 1 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 1 M41N 1 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 1 F42N 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 M43N 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 F44N 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 F45N 1 1 1 1 1 1 1 1 1 1 0 1 1 1 0 0 1 1 1 1 1 0 1 1 1 1 1 1 1 1

a Copy number categories concordant between MLPA and aCGH (0). b One technique predicted disomic copy number and the other predicted gain or loss (1). c One technique predicted copy number gain and the other predicted copy number loss ( 2).

184

Table 5.9 Ranked scores for the degree of disagreement between copy number results obtained using MLPA or aCGH for chromosome 8 loci

Mean Rank Gene Difference

1 RECQL4a 0.71 2 PTP4A3 0.61 3 KCNK9 0.58 4 EXT1 0.48 5 DLGAP2 0.48 6 SLA 0.48 7 TPD52 0.48 8 MYBL 0.45 9 KCNQ3 0.42 10 PTK2 0.42 11 KHDRBS3 0.39 12 MSRA 0.39 13 CHD7 0.35 14 EIF3H 0.35 15 MFHAS 0.35 16 TUSC3 0.35 17 ChGn 0.32 18 FGFR1 0.32 19 NCOA2 0.32 20 PRKDC 0.32 21 ST7 0.32 22 DDEF1 0.29 23 E2F5 0.29 24 EIF3E 0.29 25 RNF139 0.29 26 CTSB 0.26 27 MOS 0.26 28 RAD54B 0.26 29 GATA4 0.23 30 MYC 0.23

a Genes shown in bold were selected for qPCR validation.

185

5.3.4 Overview of qPCR

qPCR is a powerful tool used for research and diagnostics, as it can provide an estimate of copy number, relative to control DNA and a control gene, for any particular target gene (Wang et al., 2013a). In the last decade, qPCR has been used to validate copy number status as determined by aCGH results in different cancers

(Ferreira et al., 2008; Lipska et al., 2013; Pasic et al., 2010; Pazzaglia et al., 2009). qPCR offers advantages such as operating in a closed system, therefore no post-

PCR manipulations are needed. All reaction products are detected by the Rotor-

Gene 6 as the reaction progresses in real time. qPCR also offers high throughput screening capacity, and is a highly sensitive technique (Wang et al., 2013a). One of its specific advantages for this particular project is that it is possible to perform qPCR on minimal amounts of genomic DNA (25 to 40 ng). qPCR was therefore used to validate copy number results obtained using MLPA and/or aCGH.

Due to the small amounts of DNA remaining for many samples after having applied the initial techniques, we were unable to perform these qPCR tests for all samples and all genes. A group of genes was selected based on the calculated differences between MLPA and aCGH copy number results (Tables 5.8 and 5.9). Six chromosome 8q genes (RECQL4, PTP4A3, KCNK9, EXT1, TPD52, and MYC) were amplified in 7 tumour samples (one osteosarcoma F10O, 2 Ewing sarcomas M13E and F15E, 2 rhabdomyosarcomas F28R and M32R and 2 neuroblastomas F38N and

F44N) plus 2 cell lines: HCC2157 and SK-BR-3.

186

Copy numbers of target genes were normalized using VAV2, which has been previously used as a qPCR genomic control by Mosse et al. (2007). The disomic state of VAV2 was also predicted in each tumour sample and cell lines using Nexus copy number software (data not shown). This gene was included in triplicate in every assay. Also, a genomic normal female DNA control (Promega) was used as a calibrator. All copy number data were calculated as described in section 2.8. All copy number results for qPCR were converted to copy number categories (as described earlier) so they could be compared directly with MLPA and aCGH results (Table

5.10). Relative copy numbers of 1.4 - 2.6 were considered disomic: below 1.39 was considered copy number loss and above 2.61 was considered copy number gain (as per study of Pazzaglia et al. (2009)). As an example of this categorisation process, qPCR results for RECQL4 are shown in Figure 5.7.

187

12.00

10.00

8.00 Copynumber (qPCR) I

6.00 II RECQL4 III 4.00

2.00

0.00 F10O M13E F15E F28R M32R F38N F44N HCC2157 SK-BR-3 F.C.

Figure 5.7 Relative RECQL4 copy number measured using qPCR. Copy number results are shown on the Y axis as means ± SE from 3 independent experiments

(I, II, III), normalized to the results obtained for the normal female control (F.C.) which was set at 2.0. On the X axis, the 7 samples, 2 cell lines and normal female control (F.C.) are shown.

188

Table 5.10 Copy number results obtained using qPCR and resulting copy number categories

RECQL4 PTP4A3 Copy number Copy number Sample Copy number categorya Sample Copy number category F10O 2.77 2 F10O 2.74 2 M13E 4.71 2 M13E 4.18 2 F15E 1.95 1 F15E 3.79 2 F28R 4.06 2 F28R 2.07 1 M32R 3.07 2 M32R 2.24 1 F38N 1.20 0 F38N 2.44 1 F44N 6.81 2 F44N 3.31 2 HCC2157 2.91 2 HCC2157 3.06 1 SK-BR-3 3.27 2 SK-BR-3 2.98 2

KCNK9 EXT1 Copy number Copy number Sample Copy number category Sample Copy number category F10O 3.46 2 F10O 2.05 1 M13E 1.92 1 M13E 3.22 2 F15E 6.97 2 F15E 3.67 2 F28R 2.36 1 F28R 1.93 1 M32R 2.06 1 M32R 1.89 1 F38N 1.66 1 F38N 2.89 2 F44N 3.28 2 F44N 2.74 2 HCC2157 3.80 2 HCC2157 3.22 2 SK-BR-3 0.17 0 SK-BR-3 3.91 2

TPD52 MYC Copy number Copy number Sample Copy number category Sample Copy number category F10O 2.63 1 F10O 2.23 1 M13E 1.55 1 M13E 2.35 1 F15E 3.16 2 F15E 2.68 2 F28R 2.05 1 F28R 1.80 1 M32R 1.86 1 M32R 1.81 1 F38N 1.40 1 F38N 1.74 1 F44N 2.79 2 F44N 2.93 2 HCC2157 4.07 2 HCC2157 6.45 2 SK-BR-3 18.08 2 SK-BR-3 26.59 2 aCategories: 2=Gain; 1=Disomy; 0=Loss

189

5.3.5 qPCR results

Results for all techniques are presented in Table 5.11, which shows the agreement between copy number results obtained using MLPA versus qPCR, and using aCGH versus qPCR. In terms of the agreement between MLPA and qPCR, the percentage of estimates in agreement, ranged from 11.1% (1/9 samples) for RECQL4, through to 33% (3/9 samples) for PTP4A3 and EXT1, 56% (5/9 samples) for KCNK9, 67%

(6/9 samples) for TPD52 and 78% (7/9 samples) for MYC (Table 5.11). In contrast, qPCR and aCGH showed 100% agreement (9/9 samples) for all genes but one,

EXT1, which showed 89% agreement (8/9 samples). These results demonstrate that aCGH copy number results could be more frequently validated by qPCR than MLPA results (Table 5.11).

190

Table 5.11 Comparison of copy number categories resulting from MLPA versus qPCR, and qPCR versus aCGH

RECQL4 PTP4A3 Samples MLPA qPCR aCGH Samples MLPA qPCR aCGH F10O 1a 2 2 F10O 1 2 2 M13E 1 2 2 M13E 1 2 2 F15E 1 1 1 F15E 1 2 2 F28R 1 2 2 F28R 1 1 1 M32R 1 2 2 M32R 1 1 1 F38N 2 0 0 F38N 1 1 1 F44N 1 2 2 F44N 1 2 2 HCC2157 2 2 2 HCC2157 1 2 2 SK-BR-3 0 2 2 SK-BR-3 0 2 2 Agreement 1/9 9/9 Agreement 3/9 9/9

KCNK9 EXT1 Samples MLPA qPCR aCGH Samples MLPA qPCR aCGH F10O 1 2 2 F10O 1 1 1 M13E 1 1 1 M13E 1 2 1 F15E 1 2 2 F15E 1 2 2 F28R 1 1 1 F28R 1 1 1 M32R 1 1 1 M32R 0 1 1 F38N 1 1 1 F38N 1 2 2 F44N 1 2 2 F44N 1 2 2 HCC2157 1 2 2 HCC2157 1 2 2 SK-BR-3 0 0 0 SK-BR-3 2 2 2 Agreement 5/9 9/9 Agreement 3/9 8/9

TPD52 MYC Samples MLPA qPCR aCGH Samples MLPA qPCR aCGH F10O 2 1 1 F10O 1 1 1 M13E 1 1 1 M13E 1 1 1 F15E 1 2 2 F15E 1 2 2 F28R 1 1 1 F28R 1 1 1 M32R 1 1 1 M32R 1 1 1 F38N 1 1 1 F38N 1 1 1 F44N 1 2 2 F44N 1 2 2 HCC2157 2 2 2 HCC2157 2 2 2 SK-BR-3 2 2 2 SK-BR-3 2 2 2 Agreement 6/9 9/9 Agreement 7/9 9/9

aCategories: 2=Gain; 1=Disomy; 0=Loss

191

5.4 Discussion

The results of this chapter support the use of a multi-platform approach to detect copy number alterations in tumour samples. We first used MLPA, a well-established

PCR-based technique (Homig-Holzel et al., 2012). The second approach, aCGH, is a more elaborate technique with high-resolution and genome-wide scope. Whereas we noted 100% (48/48) agreement between the categorical copy number results obtained at 8p loci using MLPA and aCGH in 6 breast cell lines, the agreement between the two techniques was only 53% (70/132) at chromosome 8q loci (Table

5.2). Furthermore, in the solid tumour cohort, overall percentage agreement was

67% for 8p loci and 61% for chromosome 8q loci (Table 5.7).

Other groups have also reported discrepancies between MLPA and aCGH copy number results. Ambros et al. (2011) tested 310 neuroblastoma samples and 8 neuroblastoma cell lines, where they examined 100 loci across chromosomes 1, 2,

3, 4, 7, 9, 11, 12, 14, and 17 using MLPA and aCGH. The data analysis for MLPA was done using prototype MLPAVizard software and aCGH used CNAG 3.0 software. The results produced by each technique were compared and achieved strong agreement of 99.5%. The 0.5% of divergent results were said to be due to low tumour cell content and, specifically for neuroblastoma, a so-called flat disomic profile that can be caused by less than 60% tumour cell content, in those neuroblastomas with a higher amount of Schwann cell stroma (Ambros et al., 2011).

Combaret et al. (2012) compared these same techniques in 91 neuroblastoma samples, examining chromosomes 1, 2, 3, 4, 7, 9, 11, 12, 14, and 17. Combaret et al. (2012) used Agilent software to analyze the aCGH results, and MLPAVizard

192

software to analyze the results of MLPA. They reported 82% agreement between

MLPA and aCGH, as 5.5% neuroblastoma MLPA results were not interpretable, and

12% cases showed discrepancies between the two techniques. No other technique was used to clarify copy number disagreements, although the authors stated that

MLPA does not always provide conclusive results for genome analysis (Combaret et al., 2012).

In our project, we identified 36% disagreement between MLPA and aCGH in our paediatric tumour cohort, and to our knowledge, this is the first time such a high discrepancy has been reported. We decided to use qPCR to validate the genomic copy number results of MLPA and/or aCGH. Data from the 3 techniques (Table 5.11) showed clear differences. Interestingly RECQL4 was the worst performing gene when MLPA and qPCR were compared, and MYC was the best performing gene

(Table 5.11), supporting our choice of genes for qPCR validation (Table 5.9).

MLPA was especially poor in its prediction of copy number gain, where for example;

MLPA reported copy number gain at RECQL4 in sample F38N (when loss was reported by other two techniques), and at TPD52 in sample F10O (when qPCR and aCGH both reported disomy). On the other hand, MLPA also failed to detect gain in many other samples (Table 5.11). This indicates that MLPA more frequently failed to detect copy number gain (false negative data) than it produced false positive data.

This suggests that using MLPA alone to study chromosome 8q copy number gain may result in false-negative results.

193

Both rhabdomyosarcoma samples presented RECQL4 copy number gain using qPCR and aCGH (Table 5.11). It is known that this gene plays an important role in maintaining genome integrity, as a member of the RecQ helicase family, unwinding complementary strands of DNA for processes such as transcription of RNA, DNA repair, and recombination and replication processes for which single-stranded DNA is required (Davari et al., 2010). Amplification and overexpression of RECQL4 has been reported in breast (Thomassen et al., 2009), laryngeal (Saglam et al., 2007) and cervical cancer (Narayan et al., 2007), which is consistent with our results in 7/8 rhabdomyosarcoma samples, as detected by aCGH (Table 4.12).

194

5.5 Summary and conclusions

When applied to the six breast tumour cell lines tested here, MLPA demonstrated excellent concordance of copy number results for 8p loci. On the other hand, 8q loci showed relatively poor concordance, with 70/132 (53%) copy number category agreements. In the paediatric solid tumour cohort, our results showed similar relatively poor concordance, for both chromosome 8p and 8q loci.

Use of qPCR supported results obtained by aCGH in almost every case (53/54 comparisons) versus much poorer agreement with MLPA (25/54 comparisons).

Results from aCGH appear to be more reliable than those from MLPA and so we urge caution, when applying this particular MLPA kit for copy number studies of human tumour samples.

195

Chapter 6

General Discussion

196

6.1 Introduction

Cancer is the second most common cause of death in children in developed countries (Kaatsch, 2010). Furthermore, childhood solid tumours present genetic, biological and environmental causes that are different from those of adult solid tumours (Scotting et al., 2005).

For many years, the recognition of somatic genomic events has been an important component of diagnosis, prognosis and therapy selection in paediatric oncology

(Janeway et al., 2013). Recent technological advances that facilitate this characterization at the molecular level, such as high-density aCGH and next- generation sequencing, have led to the detection of many gene aberrations that could be identified as drivers for particular types of cancer (Heitzer et al., 2013).

However at the time that this study commenced (2010), few studies had examined copy number changes of specific genes on chromosome 8 in childhood solid tumours. The current project thus aimed to assess chromosome 8 copy number changes in a cohort of 31 solid tumours, using MLPA and aCGH methodologies.

Results presented in this thesis demonstrate that while MLPA (chapter 3) presents numerous advantages, the results obtained using the particular kit employed

(SALSA MLPA KIT P014-A1) differed from those obtained using both aCGH (chapter

4) and qPCR (chapter 5).

6.2 MLPA as a stand-alone technique for measuring copy number changes

MLPA is a technique that received growing interest from the scientific community,

197

over some years, with applications in different fields. One of the most attractive advantages of this technique is that, being based on PCR, quick results can be obtained without substantial cost, as long as the laboratory already has some basic infrastructure in place (Hochstenbach et al., 2005). However, some studies have reported discrepancies between copy number results produced by MLPA and aCGH;

Stuppia et al. (2012) mentioned that the analysis of results obtained using MLPA could be difficult to interpret due to the presence of unreliable probes. An unreliable probe gives different results when an experiment is repeated, while other probes from the same experiment show consistent results, Hence this is not a problem of the experimental procedure. Instead, it is a problem inherent to the probe per se

(Combaret et al., 2012; Niba et al., 2014).

This creates uncertainties about which MLPA probes are reliable and which are not, and which MLPA kits have been sufficiently validated against another techniques.

This is particularly clear from the Ambros et al. (2011) study, where 3 different MLPA kits for neuroblastoma comprising 100 loci in total were analysed by inter-laboratory comparative testing. After the establishment of interpretation guidelines and SNP- array validation, that particular study led to a modification of the original A1 version of the MLPA kits, through inclusion of additional probes on chromosome arms 1q,

3q, 4q, 7p, 9p and 11p, exchanging of the probes on chromosome 2p, and omitting the probes for MYCN. Together these alterations led to the release of the revised

“B1” versions of the kits. It seems therefore that MLPA is a technique that is constantly developing.

According to Conraths et al. (2006), there are basic guidelines that a molecular

198

technique should follow if it is going to be used under routine diagnostic conditions or for epidemiological studies. It is necessary to assess the sensitivity and specificity of the technique, and these parameters can be estimated by comparing the method with an existing reliable method, or recognised “gold standard” technique (Conraths et al., 2006). The P014-A1 kit did not meet this criterion, as in our study at least, this kit failed validation using aCGH, considered now as a gold standard technique for measurement of DNA copy number. To our knowledge, this is the first time any publication has validated the entire set of probes of this particular MLPA kit. .

There is a belief in the scientific community that a technique that has been released to the market for over a decade and has been used in many fields is a ready-to-be- used technique. The previous chromosome 8 kit version (P014-A1) has been released to the market and reported in a number of peer-reviewed MLPA publications (Bremmer et al., 2005; Buffart et al., 2005; Buffart et al., 2009; Rumiato et al., 2011). The Buffart group published the results of using 29 MLPA probes covering 25 genes on chromosome 8 to compare DNA copy number status in colorectal cancer according to the presence/absence of metastatic disease (Buffart et al., 2005), whereas Buffart et al. (2009) used an MLPA probe mix that contained

11 probe sets on chromosome 8, and other probes on other chromosomes to study copy number aberration in gastric cancer. In both publications, aCGH was used as validation technique and no disagreement was reported between the two techniques at any loci (Buffart et al., 2005; Buffart et al., 2009).

More recently, Yong et al. (2014) used the MLPA kit P014-A1 to analyze 33 oral squamous cell carcinoma samples, and reported amplification at the EIF3E locus in

199

9/33 samples. However in our study, EIF3E represented one of 2 clearly unreliable

MLPA probe sets within the kit.

MRC Holland released a new kit for chromosome 8 (P014-B1) in late 2013, with new probes and 12 reference probes in total, implying that this is an improved product relative to P014-A1. However, there is no published analysis that compares the performance of the two kit versions. One of the consequences of launching a new kit to market without notice is that previous results may be automatically obsolete, and it may be difficult to compare results obtained with the two kit versions, if the previous version is withdrawn from sale.

6.3 aCGH results of other genes genome-wide

Using all data obtained from our aCGH experiments, it was possible to observe particular genes that showed significant copy number gain/loss in each tumour type.

To do this we used Nexus software’s “Significant Peak” feature, selecting each tumour type to obtain tables that showed individual genes with significant peaks, presented according to high frequency, significant p-value and commonalities of copy number gain or loss observed (section 4.3.2).

While it was our aim to focus upon chromosome 8, we in fact found, after comparing all specific regions found for each tumour type, a common region of copy number gain at chromosome 22 band q11.22. Ten genes located in this region showed copy number gain including PRAME, or “preferentially expressed antigen in melanoma”, with frequencies from 62.5% in Ewing sarcomas to 83.3% in embryonal rhabdomyosarcoma (Tables 4.4, 4.5, 4.6 and 4.8). High expression of PRAME has

200

been associated with poor prognosis in solid tumours such as breast cancer and neuroblastoma (Oberthuer et al., 2004). In breast cancer, high PRAME expression correlates with increased probability of metastases (Epping et al., 2008). In neuroblastoma patients, high expression of PRAME is associated with advanced tumour stage, older age at diagnosis, and poor clinical outcome (Oberthuer et al.,

2004). In osteosarcoma biopsy samples, PRAME protein was found in more than

70% of patients (Tan et al., 2012). In the same study siRNA knockdown of PRAME in U-2OS osteosarcoma cells significantly suppressed proliferation and colony formation, and promoted G1 cell cycle arrest, suggesting that PRAME plays an important role in cell proliferation and disease progression in osteosarcoma (Tan et al., 2012). This gene is over-expressed in many solid tumours and haematological malignancies, whilst showing minimal expression in normal tissues. It is therefore a promising target for immunotherapy (Yan et al., 2011). Also, PRAME has been postulated as a potential therapeutic target for treatment of leukemias (Gunn et al.,

2009, (Gutierrez-Cosio et al., 2012), neuroblastoma (Oberthuer et al., 2004) and osteosarcoma (Tan et al., 2012). According to the results we found for this particular region, we suggest that PRAME copy number gain may be related to the ove- expression of PRAME reported previously in neuroblastoma and osteosarcoma

(Oberthuer et al. (2004) Tan et al. (2012). Future studies could also examine the expression of this gene in Ewing sarcoma and embryonal rhabdomyosarcoma, as these have not been studied in this context.

201

6.4 Summary of aCGH chromosome 8 results obtained for 6 breast cell lines and the solid tumour cohort for genes examined by MLPA

Chromosomal alterations in chromosome 8 have been frequently observed in many cancers. Gains of chromosome 8q and losses of 8p, described by El Gammal et al.

(2010) as 8p-/8q+, are common in prostate cancer (El Gammal et al., 2010), breast cancer (Choschzick et al., 2010), colorectal cancer (Jasmine et al., 2012), hepatocellular carcinoma (Roessler et al., 2012) and oral squamous cell carcinoma

(Yoshioka et al., 2013) among others, suggesting a possible relationship between loss of tumour suppressor genes located on the chromosome 8p arm and gain of genes that trigger carcinogenic processes located on the chromosome 8q arm (El

Gammal et al., 2010; Roessler et al., 2012). In our project, this same pattern was observed in 5/6 breast cell lines consistent with previous publications (section 3.3.1 and 4.4.1). However for paediatric solid tumour samples, this pattern was not observed.

Of the 4 tumour types examined, osteosarcoma showed more copy number gains/losses overall, which is consistent with the literature which states that osteosarcomas are characterized by chromosomal instability and high copy number gene amplification (Lim et al., 2005). Of the genes included in the MLPA kit, KCNK9 showed copy number gain in all osteosarcomas (5/5) and in 4/6 breast cell lines analysed by aCGH. Mu et al. (2003) reported that KCNK9 was amplified and overexpressed in breast tumours and breast cell lines, concluding that KCNK9 confers resistance to both hypoxia and serum deprivation, suggesting that its amplification and overexpression plays a physiologically important role in human

202

breast cancer. Lately KCNK9 has been found to be expressed in epithelial ovarian cancer, conferring a significant survival advantage in patients with increased expression (Innamaa et al., 2013). These investigations highlight the potential importance of copy number gain of KCNK9 in osteosarcoma, an aspect that could be explored in future studies.

One of the most prominent features of Ewing sarcoma is the recurring trisomy of chromosome 8 (Savola et al., 2009). According to Toomey et al. (2010), gains of the entire chromosome 8 have been reported by nearly every CGH study, and these occur in 23-80% of Ewing sarcoma samples. In our project, gain of whole chromosome 8 occurred in 37% (3/8) of Ewing sarcoma samples tested using aCGH, which was consistent with previous results (Toomey et al., 2010).

Chromosome 8 gain is commonly detected in ERMS (Wang, 2012). However, in our

ERMS subset (n=6) we only observed 16/30 chromosome 8 loci showing copy number gain. This finding excludes the presence of full trisomy 8, in any of our six samples. Setting aside the possibility of analytical errors, this could be due to normal cells contamination or noisy samples leading to intermittent detection of copy number change, or simply because partiality in the small cohort we used.

As the literature would predict, our ARMS subset (n=2), which is characterised by recurrent reciprocal and balanced translocation, did not show any major copy number change. Admittedly, the number of samples was too small to permit any statistical argument here, but nonetheless, our findings are consistent with the literature, for the ARMS tumour sub-category. In comparison with most other tumour

203

types, these are relatively lacking in copy number changes overall.

RECQL4 showed copy number gain in 15/31 samples from the paediatric tumour cohort and 4/6 breast cell lines, as based on aCGH data. Amplification and overexpression of RECQL4 has been reported in cervical (Narayan et al., 2007), laryngeal (Saglam et al., 2007) and breast cancer (Thomassen et al., 2009). In this last study, RECQL4 was found to be a potential metastasis-promoting gene

(Thomassen et al., 2009). Recently Fang et al. (2013) demonstrated that an increased copy number of RECQL4 plays an important role in tumourigenesis of breast cancer. They identified overexpression of RECQL4 due to gene amplification in both breast tumour tissues and breast tumour cell lines. In our project, we used two of the same cell lines used by Fang et al. (2013) and we obtained the same results at the RECQL4 locus, namely: disomic copy number for MDA-MB-231 and copy number gain for MCF-7 cells (see Appendix 5). In this regard, our results are in agreement with those of other studies.

Most literature concerning RECQL4 and childhood cancer associates RECQL4 with osteosarcoma. For instance, Rothmund-Thomson syndrome (OMIM#268400), which is caused by mutations in RECQL4, has been strongly associated with predisposition to osteosarcoma (Wang et al., 2003). In our results, however, this gene showed copy number gain in 7/8 rhabdomyosarcomas, which is a novel finding that could form the basis of future investigations.

In neuroblastoma samples, no chromosome 8 gene copy number gain or loss was consistently identified, across all of our 10 samples. However, when analysing the

204

data for cancer genes on other chromosomes (and hence with lower oligonucleotide probe spacing: see Table 4.1), copy number gain was detected in 8/10 samples at the MYCN locus on chromosome 2, band p24.3 (Table 4.10). MYCN amplification is a marker of critical importance in neuroblastoma, as it signifies aggressiveness and high metastatic potential (Ambros et al., 2011), in a subgroup of 20-30% cases

(Cohn et al., 2004). As 80% of our neuroblastoma subset showed copy number gain in MYCN, and we saw a relatively flat copy number profile across chromosome 8; we may conclude that this is one type of tumour for which chromosome 8 copy number variation is not of prime importance.

Finally: we consider the four genes RECQL4, PTP4A3, KCNK9, and EXT1, which presented copy number gain in most of the tumour cohort samples analysed by aCGH, but along with TPD52 and MYC, showed discrepant MLPA results in some cases. Most (5/6) genes showed 100% agreement between the results of aCGH and qPCR, except for EXT1 (with 89% agreement), supporting our conclusion that qPCR is an excellent, or “gold standard” technique for validation of copy number data obtained by other means.

205

Final conclusions

Copy number changes at chromosome 8 are widely recognized in several cancers.

The data presented in this thesis revealed some genes that had not previously been linked with specific cancers, such as KCNK9 (showing copy number gain in all osteosarcoma samples) and RECQL4 that showed copy number gain in most rhabdomyosarcomas. Also we present for the first time such data concerning the gene PRAME located at chromosome 22q11.22. This gene that showed copy number gain in all 4 solid tumour types examined, and has been proposed as a promising target for immunotherapy in some hematologic malignancies. However, it has not been subject to study in osteosarcoma, rhabdomyosarcoma or Ewing sarcoma. Further studies in these areas may prove to be beneficial.

206

References

207

Adhikary, S., & Eilers, M. (2005). Transcriptional regulation and transformation by Myc proteins. Nat Rev Mol Cell Biol, 6(8), 635-645. doi: 10.1038/nrm1703

AIHW, A. I. o. H. a. W. (2011). Cancer in adolescents and young adults in Australia. Canberra: AIHW: Published by the Australian Institute of Health and Welfare.

Albertson, D. G., Ylstra, B., Segraves, R., Collins, C., Dairkee, S. H., Kowbel, D., Kuo, W. L., Gray, J. W., & Pinkel, D. (2000). Quantitative mapping of amplicon structure by array CGH identifies CYP24 as a candidate oncogene. Nat Genet, 25(2), 144-146. doi: 10.1038/75985

Alunni-Fabbroni, M., & Sandri, M. T. (2010). Circulating tumour cells in clinical practice: Methods of detection and possible characterization. Methods, 50(4), 289- 297. doi: 10.1016/j.ymeth.2010.01.027

Ambros, Ambros, I. M., Brodeur, G. M., Haber, M., Khan, J., Nakagawara, A., Schleiermacher, G., Speleman, F., Spitz, R., London, W. B., Cohn, S. L., Pearson, A. D., & Maris, J. M. (2009). International consensus for neuroblastoma molecular diagnostics: report from the International Neuroblastoma Risk Group (INRG) Biology Committee. Br J Cancer, 100(9), 1471-1482. doi: 10.1038/sj.bjc.6605014

Ambros, Brunner, B., Aigner, G., Bedwell, C., Beiske, K., Benard, J., Bown, N., Combaret, V., Couturier, J., Defferrari, R., Gross, N., Jeison, M., Lunec, J., Marques, B., Martinsson, T., Mazzocco, K., Noguera, R., Schleiermacher, G., Speleman, F., Stallings, R., Tonini, G. P., Tweddle, D. A., Valent, A., Vicha, A., Roy, N. V., Villamon, E., Ziegler, A., Preuner, S., Drobics, M., Ladenstein, R., Amann, G., Schuit, R. J., Potschger, U., & Ambros, P. F. (2011). A multilocus technique for risk evaluation of patients with neuroblastoma. Clin Cancer Res, 17(4), 792-804. doi: 10.1158/1078-0432.CCR-10-0830

Amiel, A., Ohali, A., Fejgin, M., Sardos-Albertini, F., Bouaron, N., Cohen, I. J., Yaniv, I., Zaizov, R., & Avigad, S. (2003). Molecular cytogenetic parameters in Ewing sarcoma. Cancer Genet Cytogenet, 140(2), 107-112.

Andam, R., & Silva, M. (2008). A journey to pediatric chemotherapy competence. J Pediatr Nurs, 23(4), 257-268. doi: 10.1016/j.pedn.2006.12.005

Aravindan, S., Natarajan, M., Herman, T. S., & Aravindan, N. (2013). Radiation- induced TNFalpha cross signaling-dependent nuclear import of NFkappaB favors metastasis in neuroblastoma. Clin Exp Metastasis, 30(6), 807-817. doi: 10.1007/s10585-013-9580-y

Arkin, M. R., & Wells, J. A. (2004). Small-molecule inhibitors of protein-protein interactions: progressing towards the dream. Nat Rev Drug Discov, 3(4), 301-317. doi: 10.1038/nrd1343

208

Atiye, J., Wolf, M., Kaur, S., Monni, O., Bohling, T., Kivioja, A., Tas, E., Serra, M., Tarkkanen, M., & Knuutila, S. (2005). Gene amplifications in osteosarcoma-CGH microarray analysis. Genes Chromosomes Cancer, 42(2), 158-163. doi: 10.1002/gcc.20120

Attiyeh, E. F., London, W. B., Mosse, Y. P., Wang, Q., Winter, C., Khazi, D., McGrady, P. W., Seeger, R. C., Look, A. T., Shimada, H., Brodeur, G. M., Cohn, S. L., Matthay, K. K., Maris, J. M., & Children's Oncology, G. (2005). Chromosome 1p and 11q deletions and outcome in neuroblastoma. N Engl J Med, 353(21), 2243- 2253. doi: 10.1056/NEJMoa052399

Baade, P. D., Youlden, D. R., Valery, P. C., Hassall, T., Ward, L., Green, A. C., & Aitken, J. F. (2010). Trends in incidence of childhood cancer in Australia, 1983-2006. Br J Cancer, 102(3), 620-626. doi: 10.1038/sj.bjc.6605503

Bacci, G., Ferrari, S., Longhi, A., Picci, P., Mercuri, M., Alvegard, T. A., Saeter, G., Donati, D., Manfrini, M., Lari, S., Briccoli, A., Forni, C., & Italian Sarcoma Group/Scandinavian Sarcoma, G. (2002). High dose ifosfamide in combination with high dose methotrexate, adriamycin and cisplatin in the neoadjuvant treatment of extremity osteosarcoma: preliminary results of an Italian Sarcoma Group/Scandinavian Sarcoma Group pilot study. J Chemother, 14(2), 198-206. doi: 10.1179/joc.2002.14.2.198

Backlund, L. M., Nilsson, B. R., Liu, L., Ichimura, K., & Collins, V. P. (2005). Mutations in Rb1 pathway-related genes are associated with poor prognosis in anaplastic astrocytomas. Br J Cancer, 93(1), 124-130. doi: 10.1038/sj.bjc.6602661

Ballif, B. C., Rorem, E. A., Sundin, K., Lincicum, M., Gaskin, S., Coppinger, J., Kashork, C. D., Shaffer, L. G., & Bejjani, B. A. (2006). Detection of low-level mosaicism by array CGH in routine diagnostic specimens. Am J Med Genet A, 140(24), 2757-2767. doi: 10.1002/ajmg.a.31539

Bankovic, J., Stojsic, J., Jovanovic, D., Andjelkovic, T., Milinkovic, V., Ruzdijic, S., & Tanic, N. (2010). Identification of genes associated with non-small-cell lung cancer promotion and progression. Lung Cancer, 67(2), 151-159. doi: 10.1016/j.lungcan.2009.04.010

Barr, F. G. (2012). New Treatments for Rhabdomyosarcoma: the Importance of Target Practice. Clin Cancer Res, 18(3), 595-597. doi: 10.1158/1078-0432.CCR-11- 2832

Barr, F. G., Duan, F., Smith, L. M., Gustafson, D., Pitts, M., Hammond, S., & Gastier- Foster, J. M. (2009). Genomic and clinical analyses of 2p24 and 12q13-q14 amplification in alveolar rhabdomyosarcoma: a report from the Children's Oncology Group. Genes Chromosomes Cancer, 48(8), 661-672. doi: 10.1002/gcc.20673

209

Bartlett, J. M., & Stirling, D. (2003). A short history of the polymerase chain reaction. Methods Mol Biol, 226, 3-6. doi: 10.1385/1-59259-384-4:3

Batanian, J. R., Cavalli, L. R., Aldosari, N. M., Ma, E., Sotelo-Avila, C., Ramos, M. B., Rone, J. D., Thorpe, C. M., & Haddad, B. R. (2002). Evaluation of paediatric osteosarcomas by classic cytogenetic and CGH analyses. Mol Pathol, 55(6), 389- 393.

Bauman, J. G., Wiegant, J., Borst, P., & van Duijn, P. (1980). A new method for fluorescence microscopical localization of specific DNA sequences by in situ hybridization of fluorochromelabelled RNA. Exp Cell Res, 128(2), 485-490.

Bayani, J., Zielenska, M., Pandita, A., Al-Romaih, K., Karaskova, J., Harrison, K., Bridge, J. A., Sorensen, P., Thorner, P., & Squire, J. A. (2003). Spectral karyotyping identifies recurrent complex rearrangements of chromosomes 8, 17, and 20 in osteosarcomas. Genes Chromosomes Cancer, 36(1), 7-16. doi: 10.1002/gcc.10132

Bejjani, B. A., & Shaffer, L. G. (2008). Clinical utility of contemporary molecular cytogenetics. Annu Rev Genomics Hum Genet, 9, 71-86.

Benes, V., & Castoldi, M. (2010). Expression profiling of microRNA using real-time quantitative PCR, how to use it and what is available. Methods, 50(4), 244-249. doi: 10.1016/j.ymeth.2010.01.026

Berg, M., Agesen, T. H., Thiis-Evensen, E., group, I. N.-s., Merok, M. A., Teixeira, M. R., Vatn, M. H., Nesbakken, A., Skotheim, R. I., & Lothe, R. A. (2010). Distinct high resolution genome profiles of early onset and late onset colorectal cancer integrated with gene expression data identify candidate susceptibility loci. Mol Cancer, 9, 100. doi: 10.1186/1476-4598-9-100

Bethel, C. R., Faith, D., Li, X., Guan, B., Hicks, J. L., Lan, F., Jenkins, R. B., Bieberich, C. J., & De Marzo, A. M. (2006). Decreased NKX3.1 protein expression in focal prostatic atrophy, prostatic intraepithelial neoplasia, and adenocarcinoma: association with gleason score and chromosome 8p deletion. Cancer Res, 66(22), 10683-10690. doi: 10.1158/0008-5472.CAN-06-0963

Bilal, E., Vassallo, K., Toppmeyer, D., Barnard, N., Rye, I. H., Almendro, V., Russnes, H., Borresen-Dale, A. L., Levine, A. J., Bhanot, G., & Ganesan, S. (2012). Amplified loci on chromosomes 8 and 17 predict early relapse in ER-positive breast cancers. PLoS One, 7(6), e38575. doi: 10.1371/journal.pone.0038575

Boeva, V., Jouannet, S., Daveau, R., Combaret, V., Pierre-Eugene, C., Cazes, A., Louis-Brennetot, C., Schleiermacher, G., Ferrand, S., Pierron, G., Lermine, A., Rio Frio, T., Raynal, V., Vassal, G., Barillot, E., Delattre, O., & Janoueix-Lerosey, I. (2013). Breakpoint features of genomic rearrangements in neuroblastoma with unbalanced translocations and chromothripsis. PLoS One, 8(8), e72182. doi: 10.1371/journal.pone.0072182

210

Bower, K., Napier, C. E., Cole, S. L., Dagg, R. A., Lau, L. M., Duncan, E. L., Moy, E. L., & Reddel, R. R. (2012). Loss of wild-type ATRX expression in somatic cell hybrids segregates with activation of Alternative Lengthening of Telomeres. PLoS One, 7(11), e50062. doi: 10.1371/journal.pone.0050062

Braunreiter, C. L., Hancock, J. D., Coffin, C. M., Boucher, K. M., & Lessnick, S. L. (2006). Expression of EWS-ETS fusions in NIH3T3 cells reveals significant differences to Ewing's sarcoma. Cell Cycle, 5(23), 2753-2759.

Bravaccini, S., Rengucci, C., Medri, L., Zoli, W., Silvestrini, R., & Amadori, D. (2012). Detection of HER2 and Topo 2 in breast cancers: comparison between MLPA and FISH approaches. J Clin Pathol, 65(2), 183-185. doi: 10.1136/jclinpath-2011-200342

Bremmer, J. F., Braakhuis, B. J., Ruijter-Schippers, H. J., Brink, A., Duarte, H. M., Kuik, D. J., Bloemena, E., Leemans, C. R., van der Waal, I., & Brakenhoff, R. H. (2005). A noninvasive genetic screening test to detect oral preneoplastic lesions. Lab Invest, 85(12), 1481-1488. doi: 10.1038/labinvest.3700342

Bridge, J. A. (2008). Advantages and limitations of cytogenetic, molecular cytogenetic, and molecular diagnostic testing in mesenchymal neoplasms. J Orthop Sci, 13(3), 273-282. doi: 10.1007/s00776-007-1215-1

Brodeur, G. M. (2003). Neuroblastoma: biological insights into a clinical enigma. Nat Rev Cancer, 3(3), 203-216. doi: 10.1038/nrc1014

Buffart, T. E., Coffa, J., Hermsen, M. A., Carvalho, B., van der Sijp, J. R., Ylstra, B., Pals, G., Schouten, J. P., & Meijer, G. A. (2005). DNA copy number changes at 8q11-24 in metastasized colorectal cancer. Cell Oncol, 27(1), 57-65.

Buffart, T. E., van Grieken, N. C., Tijssen, M., Coffa, J., Ylstra, B., Grabsch, H. I., van de Velde, C. J., Carvalho, B., & Meijer, G. A. (2009). High resolution analysis of DNA copy-number aberrations of chromosomes 8, 13, and 20 in gastric cancers. Virchows Arch, 455(3), 213-223. doi: 10.1007/s00428-009-0814-y

Buiting, K. (2010). Prader-Willi syndrome and Angelman syndrome. Am J Med Genet C Semin Med Genet, 154C(3), 365-376. doi: 10.1002/ajmg.c.30273

Burchill, S. A. (2003). Ewing's sarcoma: diagnostic, prognostic, and therapeutic implications of molecular abnormalities. J Clin Pathol, 56(2), 96-102.

Bustin, S. A., Benes, V., Garson, J. A., Hellemans, J., Huggett, J., Kubista, M., Mueller, R., Nolan, T., Pfaffl, M. W., Shipley, G. L., Vandesompele, J., & Wittwer, C. T. (2009). The MIQE guidelines: minimum information for publication of quantitative real-time PCR experiments. Clin Chem, 55(4), 611-622. doi: 10.1373/clinchem.2008.112797

211

Byrne, J. A., Maleki, S., Hardy, J. R., Gloss, B. S., Murali, R., Scurry, J. P., Fanayan, S., Emmanuel, C., Hacker, N. F., Sutherland, R. L., Defazio, A., & O'Brien, P. M. (2010). MAL2 and tumor protein D52 (TPD52) are frequently overexpressed in ovarian carcinoma, but differentially associated with histological subtype and patient outcome. BMC Cancer, 10, 497. doi: 10.1186/1471-2407-10-497

Caceres, A., Armengol, L., Villatoro, S., & Gonzalez, J. R. (2011). MLPAstats: an R GUI package for the integrated analysis of copy number alterations using MLPA data. BMC Bioinformatics, 12, 147. doi: 10.1186/1471-2105-12-147

Cao, L., Yu, Y., Bilke, S., Walker, R. L., Mayeenuddin, L. H., Azorsa, D. O., Yang, F., Pineda, M., Helman, L. J., & Meltzer, P. S. (2010). Genome-wide identification of PAX3-FKHR binding sites in rhabdomyosarcoma reveals candidate target genes important for development and cancer. Cancer Res, 70(16), 6497-6508. doi: 10.1158/0008-5472.CAN-10-0582

Caramins, M. C., Saville, T., Shakeshaft, R., Mullan, G. L., Miller, B., Yip, M. Y., & Buckley, M. F. (2011). A comparison of molecular and cytogenetic techniques for the diagnosis of pregnancy loss. Genet Med, 13(1), 46-51. doi: 10.1097/GIM.0b013e3181faa0d2

Caren, H., Kryh, H., Nethander, M., Sjoberg, R. M., Trager, C., Nilsson, S., Abrahamsson, J., Kogner, P., & Martinsson, T. (2010). High-risk neuroblastoma tumors with 11q-deletion display a poor prognostic, chromosome instability phenotype with later onset. Proc Natl Acad Sci U S A, 107(9), 4323-4328. doi: 10.1073/pnas.0910684107

Carr, A. C., & Moore, S. D. (2012). Robust quantification of polymerase chain reactions using global fitting. PLoS One, 7(5), e37640. doi: 10.1371/journal.pone.0037640

Carter, N. P. (2007). Methods and strategies for analyzing copy number variation using DNA microarrays. Nat Genet, 39(7 Suppl), S16-21. doi: 10.1038/ng2028

Castel, V., Grau, E., Noguera, R., & Martinez, F. (2007). Molecular biology of neuroblastoma. Clin Transl Oncol, 9(8), 478-483.

Chan, T. L., Yuen, S. T., Kong, C. K., Chan, Y. W., Chan, A. S., Ng, W. F., Tsui, W. Y., Lo, M. W., Tam, W. Y., Li, V. S., & Leung, S. Y. (2006). Heritable germline epimutation of MSH2 in a family with hereditary nonpolyposis colorectal cancer. Nat Genet, 38(10), 1178-1183. doi: 10.1038/ng1866

Chen, & Olopade, O. I. (2008). MYC in breast tumor progression. Expert Rev Anticancer Ther, 8(10), 1689-1698. doi: 10.1586/14737140.8.10.1689

Chen., Stewart, E., Shelat, A. A., Qu, C., Bahrami, A., Hatley, M., Wu, G., Bradley, C., McEvoy, J., Pappo, A., Spunt, S., Valentine, M. B., Valentine, V., Krafcik, F.,

212

Lang, W. H., Wierdl, M., Tsurkan, L., Tolleman, V., Federico, S. M., Morton, C., Lu, C., Ding, L., Easton, J., Rusch, M., Nagahawatte, P., Wang, J., Parker, M., Wei, L., Hedlund, E., Finkelstein, D., Edmonson, M., Shurtleff, S., Boggs, K., Mulder, H., Yergeau, D., Skapek, S., Hawkins, D. S., Ramirez, N., Potter, P. M., Sandoval, J. A., Davidoff, A. M., Mardis, E. R., Wilson, R. K., Zhang, J., Downing, J. R., Dyer, M. A., & St. Jude Children's Research Hospital-Washington University Pediatric Cancer Genome, P. (2013). Targeting oxidative stress in embryonal rhabdomyosarcoma. Cancer Cell, 24(6), 710-724. doi: 10.1016/j.ccr.2013.11.002

Choschzick, M., Lassen, P., Lebeau, A., Marx, A. H., Terracciano, L., Heilenkotter, U., Jaenicke, F., Bokemeyer, C., Izbicki, J., Sauter, G., & Simon, R. (2010). Amplification of 8q21 in breast cancer is independent of MYC and associated with poor patient outcome. Mod Pathol, 23(4), 603-610. doi: 10.1038/modpathol.2010.5

Cohn, S. L., Pearson, A. D., London, W. B., Monclair, T., Ambros, P. F., Brodeur, G. M., Faldum, A., Hero, B., Iehara, T., Machin, D., Mosseri, V., Simon, T., Garaventa, A., Castel, V., Matthay, K. K., & Force, I. T. (2009). The International Neuroblastoma Risk Group (INRG) classification system: an INRG Task Force report. J Clin Oncol, 27(2), 289-297. doi: 10.1200/JCO.2008.16.6785

Cohn, S. L., & Tweddle, D. A. (2004). MYCN amplification remains prognostically strong 20 years after its "clinical debut". Eur J Cancer, 40(18), 2639-2642. doi: 10.1016/j.ejca.2004.07.025

Combaret, V., Iacono, I., Brejon, S., Schleiermacher, G., Pierron, G., Couturier, J., Bergeron, C., & Blay, J. Y. (2012). Analysis of genomic alterations in neuroblastoma by multiplex ligation-dependent probe amplification and array comparative genomic hybridization: a comparison of results. Cancer Genet, 205(12), 657-664. doi: 10.1016/j.cancergen.2012.11.002

Conraths, F. J., & Schares, G. (2006). Validation of molecular-diagnostic techniques in the parasitological laboratory. Vet Parasitol, 136(2), 91-98. doi: 10.1016/j.vetpar.2005.12.004

Cytrynbaum, C. S., Smith, A. C., Rubin, T., & Weksberg, R. (2005). Advances in overgrowth syndromes: clinical classification to molecular delineation in Sotos syndrome and Beckwith-Wiedemann syndrome. Curr Opin Pediatr, 17(6), 740-746.

Davari, P., Hebert, J. L., Albertson, D. G., Huey, B., Roy, R., Mancianti, M. L., Horvai, A. E., McDaniel, L. D., Schultz, R. A., & Epstein, E. H., Jr. (2010). Loss of Blm enhances basal cell carcinoma and rhabdomyosarcoma tumorigenesis in Ptch1+/- mice. Carcinogenesis, 31(6), 968-973. doi: 10.1093/carcin/bgp309

Davicioni, E., Anderson, M. J., Finckenstein, F. G., Lynch, J. C., Qualman, S. J., Shimada, H., Schofield, D. E., Buckley, J. D., Meyer, W. H., Sorensen, P. H., & Triche, T. J. (2009). Molecular classification of rhabdomyosarcoma--genotypic and phenotypic determinants of diagnosis: a report from the Children's Oncology Group. Am J Pathol, 174(2), 550-564. doi: 10.2353/ajpath.2009.080631

213

De Falco, G., Leucci, E., Lenze, D., Piccaluga, P. P., Claudio, P. P., Onnis, A., Cerino, G., Nyagol, J., Mwanda, W., Bellan, C., Hummel, M., Pileri, S., Tosi, P., Stein, H., Giordano, A., & Leoncini, L. (2007). Gene-expression analysis identifies novel RBL2/p130 target genes in endemic Burkitt lymphoma cell lines and primary tumors. Blood, 110(4), 1301-1307. doi: 10.1182/blood-2006-12-064865

De Lellis, L., Curia, M. C., Aceto, G. M., Toracchio, S., Colucci, G., Russo, A., Mariani-Costantini, R., & Cama, A. (2007). Analysis of extended genomic rearrangements in oncological research. Ann Oncol, 18 Suppl 6, vi173-178. doi: 10.1093/annonc/mdm251

De Preter, K., Vandesompele, J., Heimann, P., Yigit, N., Beckman, S., Schramm, A., Eggert, A., Stallings, R. L., Benoit, Y., Renard, M., De Paepe, A., Laureys, G., Pahlman, S., & Speleman, F. (2006). Human fetal neuroblast and neuroblastoma transcriptome analysis confirms neuroblast origin and highlights neuroblastoma candidate genes. Genome Biol, 7(9), R84. doi: 10.1186/gb-2006-7-9-r84

Dhanasekaran, S., Doherty, T. M., Kenneth, J., & Group, T. B. T. S. (2010). Comparison of different standards for real-time PCR-based absolute quantification. J Immunol Methods, 354(1-2), 34-39. doi: 10.1016/j.jim.2010.01.004

Diao, Y., Guo, X., Jiang, L., Wang, G., Zhang, C., Wan, J., Jin, Y., & Wu, Z. (2014). miR-203, a tumor suppressor frequently down-regulated by promoter hypermethylation in rhabdomyosarcoma. J Biol Chem, 289(1), 529-539. doi: 10.1074/jbc.M113.494716

Diskin, S. J., Eck, T., Greshock, J., Mosse, Y. P., Naylor, T., Stoeckert, C. J., Jr., Weber, B. L., Maris, J. M., & Grant, G. R. (2006). STAC: A method for testing the significance of DNA copy number aberrations across multiple array-CGH experiments. Genome Res, 16(9), 1149-1158. doi: 10.1101/gr.5076506

Doolan, P., Clynes, M., Kennedy, S., Mehta, J. P., Crown, J., & O'Driscoll, L. (2008). Prevalence and prognostic and predictive relevance of PRAME in breast cancer. Breast Cancer Res Treat, 109(2), 359-365. doi: 10.1007/s10549-007-9643-3 dos Santos Aguiar, S., de Jesus Girotto Zambaldi, L., dos Santos, A. M., Pinto, W., Jr., & Brandalise, S. R. (2007). Comparative genomic hybridization analysis of abnormalities in chromosome 21 in childhood osteosarcoma. Cancer Genet Cytogenet, 175(1), 35-40. doi: 10.1016/j.cancergencyto.2007.01.007

Dragone, L. L., Myers, M. D., White, C., Gadwal, S., Sosinowski, T., Gu, H., & Weiss, A. (2006). Src-like adaptor protein (SLAP) regulates B cell receptor levels in a c-Cbl-dependent manner. Proc Natl Acad Sci U S A, 103(48), 18202-18207. doi: 10.1073/pnas.0608965103

214

Du, X., Yang, J., Yang, D., Tian, W., & Zhu, Z. (2014). The genetic basis for inactivation of Wnt pathway in human osteosarcoma. BMC Cancer, 14, 450. doi: 10.1186/1471-2407-14-450

El Gammal, A. T., Bruchmann, M., Zustin, J., Isbarn, H., Hellwinkel, O. J., Kollermann, J., Sauter, G., Simon, R., Wilczak, W., Schwarz, J., Bokemeyer, C., Brummendorf, T. H., Izbicki, J. R., Yekebas, E., Fisch, M., Huland, H., Graefen, M., & Schlomm, T. (2010). Chromosome 8p deletions and 8q gains are associated with tumor progression and poor prognosis in prostate cancer. Clin Cancer Res, 16(1), 56-64. doi: 10.1158/1078-0432.CCR-09-1423

Epping, M. T., Hart, A. A., Glas, A. M., Krijgsman, O., & Bernards, R. (2008). PRAME expression and clinical outcome of breast cancer. Br J Cancer, 99(3), 398- 403. doi: 10.1038/sj.bjc.6604494

Ethier, S. P. (2003). Identifying and validating causal genetic alterations in human breast cancer. Breast Cancer Res Treat, 78(3), 285-287.

Evans, D. G., Ramsden, R. T., Shenton, A., Gokhale, C., Bowers, N. L., Huson, S. M., Pichert, G., & Wallace, A. (2007). Mosaicism in neurofibromatosis type 2: an update of risk based on uni/bilaterality of vestibular schwannoma at presentation and sensitive mutation analysis including multiple ligation-dependent probe amplification. J Med Genet, 44(7), 424-428. doi: 10.1136/jmg.2006.047753

Fang, H., Nie, L., Chi, Z., Liu, J., Guo, D., Lu, X., Hei, T. K., Balajee, A. S., & Zhao, Y. (2013). RecQL4 helicase amplification is involved in human breast tumorigenesis. PLoS One, 8(7), e69600. doi: 10.1371/journal.pone.0069600

Ferreira, B. I., Alonso, J., Carrillo, J., Acquadro, F., Largo, C., Suela, J., Teixeira, M. R., Cerveira, N., Molares, A., Gomez-Lopez, G., Pestana, A., Sastre, A., Garcia- Miguel, P., & Cigudosa, J. C. (2008). Array CGH and gene-expression profiling reveals distinct genomic instability patterns associated with DNA repair and cell- cycle checkpoint pathways in Ewing's sarcoma. Oncogene, 27(14), 2084-2090. doi: 10.1038/sj.onc.1210845

Fridlyand, J., Snijders, A. M., Ylstra, B., Li, H., Olshen, A., Segraves, R., Dairkee, S., Tokuyasu, T., Ljung, B. M., Jain, A. N., McLennan, J., Ziegler, J., Chin, K., Devries, S., Feiler, H., Gray, J. W., Waldman, F., Pinkel, D., & Albertson, D. G. (2006). Breast tumor copy number aberration phenotypes and genomic instability. BMC Cancer, 6, 96. doi: 10.1186/1471-2407-6-96

Gadkar, V. Y., & Filion, M. (2013). Quantitative Real-time Polymerase Chain Reaction for Tracking Microbial Gene Expression in Complex Environmental Matrices. Curr Issues Mol Biol, 15(2), 45-58.

Gersen, S. L. K., M.B (Ed.). (2005). The principles of clinical cytogenetics (2nd ed. ed.). Totowa, NJ: Humana Press.

215

Ghadimi, B. M., Grade, M., Liersch, T., Langer, C., Siemer, A., Fuzesi, L., & Becker, H. (2003). Gain of chromosome 8q23-24 is a predictive marker for lymph node positivity in colorectal cancer. Clin Cancer Res, 9(5), 1808-1814.

Goto, S., Umehara, S., Gerbing, R. B., Stram, D. O., Brodeur, G. M., Seeger, R. C., Lukens, J. N., Matthay, K. K., & Shimada, H. (2001). Histopathology (International Neuroblastoma Pathology Classification) and MYCN status in patients with peripheral neuroblastic tumors: a report from the Children's Cancer Group. Cancer, 92(10), 2699-2708.

Gropman, A. L., Elsea, S., Duncan, W. C., Jr., & Smith, A. C. (2007). New developments in Smith-Magenis syndrome (del 17p11.2). Curr Opin Neurol, 20(2), 125-134. doi: 10.1097/WCO.0b013e3280895dba

Gutierrez-Cosio, S., de la Rica, L., Ballestar, E., Santamaria, C., Sanchez-Abarca, L. I., Caballero-Velazquez, T., Blanco, B., Calderon, C., Herrero-Sanchez, C., Carrancio, S., Ciudad, L., Canizo, C., San Miguel, J. F., & Perez-Simon, J. A. (2012). Epigenetic regulation of PRAME in acute myeloid leukemia is different compared to CD34+ cells from healthy donors: effect of 5-AZA treatment. Leuk Res, 36(7), 895- 899. doi: 10.1016/j.leukres.2012.02.030

Haddy, R. I., & Haddy, T. B. (2010). Lifetime follow-up care after childhood cancer. J Am Board Fam Med, 23(5), 647-654. doi: 10.3122/jabfm.2010.05.100031

Hayes, J. L., Tzika, A., Thygesen, H., Berri, S., Wood, H. M., Hewitt, S., Pendlebury, M., Coates, A., Willoughby, L., Watson, C. M., Rabbitts, P., Roberts, P., & Taylor, G. R. (2013). Diagnosis of copy number variation by Illumina next generation sequencing is comparable in performance to oligonucleotide array comparative genomic hybridisation. Genomics, 102(3), 174-181. doi: 10.1016/j.ygeno.2013.04.006

Heitzer, E., Auer, M., Gasch, C., Pichler, M., Ulz, P., Hoffmann, E. M., Lax, S., Waldispuehl-Geigl, J., Mauermann, O., Lackner, C., Hofler, G., Eisner, F., Sill, H., Samonigg, H., Pantel, K., Riethdorf, S., Bauernhofer, T., Geigl, J. B., & Speicher, M. R. (2013). Complex tumor genomes inferred from single circulating tumor cells by array-CGH and next-generation sequencing. Cancer Res, 73(10), 2965-2975. doi: 10.1158/0008-5472.CAN-12-4140

Herman, J. G., & Baylin, S. B. (2003). Gene silencing in cancer in association with promoter hypermethylation. N Engl J Med, 349(21), 2042-2054. doi: 10.1056/NEJMra023075

Hero, B., Simon, T., Spitz, R., Ernestus, K., Gnekow, A. K., Scheel-Walter, H. G., Schwabe, D., Schilling, F. H., Benz-Bohm, G., & Berthold, F. (2008). Localized infant neuroblastomas often show spontaneous regression: results of the prospective trials NB95-S and NB97. J Clin Oncol, 26(9), 1504-1510. doi: 10.1200/JCO.2007.12.3349

216

Hicks, M. J., Roth, J. R., Kozinetz, C. A., & Wang, L. L. (2007). Clinicopathologic features of osteosarcoma in patients with Rothmund-Thomson syndrome. J Clin Oncol, 25(4), 370-375. doi: 10.1200/JCO.2006.08.4558

Hochstenbach, R., Meijer, J., van de Brug, J., Vossebeld-Hoff, I., Jansen, R., van der Luijt, R. B., Sinke, R. J., Page-Christiaens, G. C., Ploos van Amstel, J. K., & de Pater, J. M. (2005). Rapid detection of chromosomal aneuploidies in uncultured amniocytes by multiplex ligation-dependent probe amplification (MLPA). Prenat Diagn, 25(11), 1032-1039. doi: 10.1002/pd.1247

Holcomb, I. N., & Trask, B. J. (2011). Comparative genomic hybridization to detect variation in the copy number of large DNA segments. Cold Spring Harb Protoc, 2011(11), 1323-1333. doi: 10.1101/pdb.top066589

Hollink, I. H., van den Heuvel-Eibrink, M. M., Zimmermann, M., Balgobind, B. V., Arentsen-Peters, S. T., Alders, M., Willasch, A., Kaspers, G. J., Trka, J., Baruchel, A., de Graaf, S. S., Creutzig, U., Pieters, R., Reinhardt, D., & Zwaan, C. M. (2009). Clinical relevance of Wilms tumor 1 gene mutations in childhood acute myeloid leukemia. Blood, 113(23), 5951-5960. doi: 10.1182/blood-2008-09-177949

Homig-Holzel, C., & Savola, S. (2012). Multiplex ligation-dependent probe amplification (MLPA) in tumor diagnostics and prognostics. Diagn Mol Pathol, 21(4), 189-206. doi: 10.1097/PDM.0b013e3182595516

Huang, H. Y., Illei, P. B., Zhao, Z., Mazumdar, M., Huvos, A. G., Healey, J. H., Wexler, L. H., Gorlick, R., Meyers, P., & Ladanyi, M. (2005). Ewing sarcomas with p53 mutation or p16/p14ARF homozygous deletion: a highly lethal subset associated with poor chemoresponse. J Clin Oncol, 23(3), 548-558. doi: 10.1200/JCO.2005.02.081

Hughes, S., Yoshimoto, M., Beheshti, B., Houlston, R. S., Squire, J. A., & Evans, A. (2006). The use of whole genome amplification to study chromosomal changes in prostate cancer: insights into genome-wide signature of preneoplasia associated with cancer progression. BMC Genomics, 7, 65. doi: 10.1186/1471-2164-7-65

Im, Y. H., Kim, H. T., Lee, C., Poulin, D., Welford, S., Sorensen, P. H., Denny, C. T., & Kim, S. J. (2000). EWS-FLI1, EWS-ERG, and EWS-ETV1 oncoproteins of Ewing tumor family all suppress transcription of transforming growth factor beta type II receptor gene. Cancer Res, 60(6), 1536-1540.

Innamaa, A., Jackson, L., Asher, V., Van Shalkwyk, G., Warren, A., Hay, D., Bali, A., Sowter, H., & Khan, R. (2013). Expression and prognostic significance of the oncogenic K2P potassium channel KCNK9 (TASK-3) in ovarian carcinoma. Anticancer Res, 33(4), 1401-1408.

Jackson, E. M., Sievert, A. J., Gai, X., Hakonarson, H., Judkins, A. R., Tooke, L., Perin, J. C., Xie, H., Shaikh, T. H., & Biegel, J. A. (2009). Genomic analysis using

217

high-density single nucleotide polymorphism-based oligonucleotide arrays and multiplex ligation-dependent probe amplification provides a comprehensive analysis of INI1/SMARCB1 in malignant rhabdoid tumors. Clin Cancer Res, 15(6), 1923-1930. doi: 10.1158/1078-0432.CCR-08-2091

Jahromi, M. S., Putnam, A. R., Druzgal, C., Wright, J., Spraker-Perlman, H., Kinsey, M., Zhou, H., Boucher, K. M., Randall, R. L., Jones, K. B., Lucas, D., Rosenberg, A., Thomas, D., Lessnick, S. L., & Schiffman, J. D. (2012). Molecular inversion probe analysis detects novel copy number alterations in Ewing sarcoma. Cancer Genet, 205(7-8), 391-404. doi: 10.1016/j.cancergen.2012.05.012

Jain, S., Xu, R., Prieto, V. G., & Lee, P. (2010). Molecular classification of soft tissue sarcomas and its clinical applications. Int J Clin Exp Pathol, 3(4), 416-428.

Janeway, K. A., Place, A. E., Kieran, M. W., & Harris, M. H. (2013). Future of clinical genomics in pediatric oncology. J Clin Oncol, 31(15), 1893-1903. doi: 10.1200/JCO.2012.46.8470

Janoueix-Lerosey, I., Lequin, D., Brugieres, L., Ribeiro, A., de Pontual, L., Combaret, V., Raynal, V., Puisieux, A., Schleiermacher, G., Pierron, G., Valteau-Couanet, D., Frebourg, T., Michon, J., Lyonnet, S., Amiel, J., & Delattre, O. (2008). Somatic and germline activating mutations of the ALK kinase receptor in neuroblastoma. Nature, 455(7215), 967-970. doi: 10.1038/nature07398

Janoueix-Lerosey, I., Schleiermacher, G., Michels, E., Mosseri, V., Ribeiro, A., Lequin, D., Vermeulen, J., Couturier, J., Peuchmaur, M., Valent, A., Plantaz, D., Rubie, H., Valteau-Couanet, D., Thomas, C., Combaret, V., Rousseau, R., Eggert, A., Michon, J., Speleman, F., & Delattre, O. (2009). Overall genomic pattern is a predictor of outcome in neuroblastoma. J Clin Oncol, 27(7), 1026-1033. doi: 10.1200/JCO.2008.16.0630

Jasmine, F., Rahaman, R., Dodsworth, C., Roy, S., Paul, R., Raza, M., Paul-Brutus, R., Kamal, M., Ahsan, H., & Kibriya, M. G. (2012). A genome-wide study of cytogenetic changes in colorectal cancer using SNP microarrays: opportunities for future personalized treatment. PLoS One, 7(2), e31968. doi: 10.1371/journal.pone.0031968

Jedlicka, p., et al. (2010). Ewing Sarcoma, an enigmatic malignancy of likely progenitor cell origin, driven by transcription factor oncogenic fusions. Int J Clin Exp Pathol, 3(4), 338-347.

Jeuken, J., van den Broecke, C., Gijsen, S., Boots-Sprenger, S., & Wesseling, P. (2007). RAS/RAF pathway activation in gliomas: the result of copy number gains rather than activating mutations. Acta Neuropathol, 114(2), 121-133. doi: 10.1007/s00401-007-0239-0

218

Jonsson, G., Staaf, J., Olsson, E., Heidenblad, M., Vallon-Christersson, J., Osoegawa, K., de Jong, P., Oredsson, S., Ringner, M., Hoglund, M., & Borg, A. (2007). High-resolution genomic profiles of breast cancer cell lines assessed by tiling BAC array comparative genomic hybridization. Genes Chromosomes Cancer, 46(6), 543-558. doi: 10.1002/gcc.20438

Joseph, C. G., Darrah, E., Shah, A. A., Skora, A. D., Casciola-Rosen, L. A., Wigley, F. M., Boin, F., Fava, A., Thoburn, C., Kinde, I., Jiao, Y., Papadopoulos, N., Kinzler, K. W., Vogelstein, B., & Rosen, A. (2014). Association of the autoimmune disease scleroderma with an immunologic response to cancer. Science, 343(6167), 152-157. doi: 10.1126/science.1246886

Kaatsch, P. (2010). Epidemiology of childhood cancer. Cancer Treat Rev, 36(4), 277-285. doi: 10.1016/j.ctrv.2010.02.003

Kallioniemi. (2008). CGH microarrays and cancer. Curr Opin Biotechnol, 19(1), 36- 40. doi: 10.1016/j.copbio.2007.11.004

Kallioniemi, A., Kallioniemi, O. P., Sudar, D., Rutovitz, D., Gray, J. W., Waldman, F., & Pinkel, D. (1992). Comparative genomic hybridization for molecular cytogenetic analysis of solid tumors. Science, 258(5083), 818-821.

Kansara, M., Teng, M. W., Smyth, M. J., & Thomas, D. M. (2014). Translational biology of osteosarcoma. Nat Rev Cancer, 14(11), 722-735. doi: 10.1038/nrc3838

Kansara, M., & Thomas, D. M. (2007). Molecular pathogenesis of osteosarcoma. DNA Cell Biol, 26(1), 1-18. doi: 10.1089/dna.2006.0505

Kao, J., Salari, K., Bocanegra, M., Choi, Y. L., Girard, L., Gandhi, J., Kwei, K. A., Hernandez-Boussard, T., Wang, P., Gazdar, A. F., Minna, J. D., & Pollack, J. R. (2009). Molecular profiling of breast cancer cell lines defines relevant tumor models and provides a resource for cancer gene discovery. PLoS One, 4(7), e6146. doi: 10.1371/journal.pone.0006146

Kirchhoff, M., Rose, H., & Lundsteen, C. (2001). High resolution comparative genomic hybridisation in clinical cytogenetics. J Med Genet, 38(11), 740-744.

Kline, N. E., & Sevier, N. (2003). Solid tumors in children. J Pediatr Nurs, 18(2), 96- 102. doi: 10.1053/jpdn.2003.12

Kluwe, L., Nygren, A. O., Errami, A., Heinrich, B., Matthies, C., Tatagiba, M., & Mautner, V. (2005). Screening for large mutations of the NF2 gene. Genes Chromosomes Cancer, 42(4), 384-391. doi: 10.1002/gcc.20138

Kozlowski, P., Jasinska, A. J., & Kwiatkowski, D. J. (2008). New applications and developments in the use of multiplex ligation-dependent probe amplification. Electrophoresis, 29(23), 4627-4636. doi: 10.1002/elps.200800126

219

Kratz, C. P., Steinemann, D., Niemeyer, C. M., Schlegelberger, B., Koscielniak, E., Kontny, U., & Zenker, M. (2007). Uniparental disomy at chromosome 11p15.5 followed by HRAS mutations in embryonal rhabdomyosarcoma: lessons from Costello syndrome. Hum Mol Genet, 16(4), 374-379. doi: 10.1093/hmg/ddl458

Kresse, S. H., Ohnstad, H. O., Paulsen, E. B., Bjerkehagen, B., Szuhai, K., Serra, M., Schaefer, K. L., Myklebost, O., & Meza-Zepeda, L. A. (2009). LSAMP, a novel candidate tumor suppressor gene in human osteosarcomas, identified by array comparative genomic hybridization. Genes Chromosomes Cancer, 48(8), 679-693. doi: 10.1002/gcc.20675

Krzywinski, M., Bosdet, I., Smailus, D., Chiu, R., Mathewson, C., Wye, N., Barber, S., Brown-John, M., Chan, S., Chand, S., Cloutier, A., Girn, N., Lee, D., Masson, A., Mayo, M., Olson, T., Pandoh, P., Prabhu, A. L., Schoenmakers, E., Tsai, M., Albertson, D., Lam, W., Choy, C. O., Osoegawa, K., Zhao, S., de Jong, P. J., Schein, J., Jones, S., & Marra, M. A. (2004). A set of BAC clones spanning the human genome. Nucleic Acids Res, 32(12), 3651-3660. doi: 10.1093/nar/gkh700

Kuhn, M. W., Radtke, I., Bullinger, L., Goorha, S., Cheng, J., Edelmann, J., Gohlke, J., Su, X., Paschka, P., Pounds, S., Krauter, J., Ganser, A., Quessar, A., Ribeiro, R., Gaidzik, V. I., Shurtleff, S., Kronke, J., Holzmann, K., Ma, J., Schlenk, R. F., Rubnitz, J. E., Dohner, K., Dohner, H., & Downing, J. R. (2012). High-resolution genomic profiling of adult and pediatric core-binding factor acute myeloid leukemia reveals new recurrent genomic alterations. Blood, 119(10), e67-75. doi: 10.1182/blood-2011- 09-380444

Kumps, C., Fieuw, A., Mestdagh, P., Menten, B., Lefever, S., Pattyn, F., De Brouwer, S., Sante, T., Schulte, J. H., Schramm, A., Van Roy, N., Van Maerken, T., Noguera, R., Combaret, V., Devalck, C., Westermann, F., Laureys, G., Eggert, A., Vandesompele, J., De Preter, K., & Speleman, F. (2013). Focal DNA copy number changes in neuroblastoma target MYCN regulated genes. PLoS One, 8(1), e52321. doi: 10.1371/journal.pone.0052321

L'Hermine, A. C., Aboura, A., Brisset, S., Cuisset, L., Castaigne, V., Labrune, P., Frydman, R., & Tachdjian, G. (2003). Fetal phenotype of Prader-Willi syndrome due to maternal disomy for chromosome 15. Prenat Diagn, 23(11), 938-943. doi: 10.1002/pd.732

Lake, S. L., Kalirai, H., Dopierala, J., Damato, B. E., & Coupland, S. E. (2012). Comparison of formalin-fixed and snap-frozen samples analyzed by multiplex ligation-dependent probe amplification for prognostic testing in uveal melanoma. Invest Ophthalmol Vis Sci, 53(6), 2647-2652. doi: 10.1167/iovs.12-9584

Langerak, P., Nygren, A. O., Schouten, J. P., & Jacobs, H. (2005). Rapid and quantitative detection of homologous and non-homologous recombination events using three oligonucleotide MLPA. Nucleic Acids Res, 33(22), e188. doi: 10.1093/nar/gni187

220

Lau, L. M., Dagg, R. A., Henson, J. D., Au, A. Y., Royds, J. A., & Reddel, R. R. (2013). Detection of alternative lengthening of telomeres by telomere quantitative PCR. Nucleic Acids Res, 41(2), e34. doi: 10.1093/nar/gks781

Le Beau, M. M., Bitts, S., Davis, E. M., & Kogan, S. C. (2002). Recurring chromosomal abnormalities in leukemia in PML-RARA transgenic mice parallel human acute promyelocytic leukemia. Blood, 99(8), 2985-2991.

Lee, C., & Scherer, S. W. (2010). The clinical context of copy number variation in the human genome. Expert Rev Mol Med, 12, e8. doi: 10.1017/S1462399410001390

Lei, K. F., Wang, Y. F., Zhu, X. Q., Lu, P. C., Sun, B. S., Jia, H. L., Ren, N., Ye, Q. H., Sun, H. C., Wang, L., Tang, Z. Y., & Qin, L. X. (2007). Identification of MSRA gene on chromosome 8p as a candidate metastasis suppressor for human hepatitis B virus-positive hepatocellular carcinoma. BMC Cancer, 7, 172. doi: 10.1186/1471- 2407-7-172

Levy, B., Dunn, T. M., Kern, J. H., Hirschhorn, K., & Kardon, N. B. (2002). Delineation of the dup5q phenotype by molecular cytogenetic analysis in a patient with dup5q/del 5p (cri du chat). Am J Med Genet, 108(3), 192-197.

Li, X., Chen, J., Lu, B., Peng, S., Desper, R., & Lai, M. (2011). -8p12-23 and +20q are predictors of subtypes and metastatic pathways in colorectal cancer: construction of tree models using comparative genomic hybridization data. OMICS, 15(1-2), 37-47. doi: 10.1089/omi.2010.0101

Lim, G., Karaskova, J., Beheshti, B., Vukovic, B., Bayani, J., Selvarajah, S., Watson, S. K., Lam, W. L., Zielenska, M., & Squire, J. A. (2005). An integrated mBAND and submegabase resolution tiling set (SMRT) CGH array analysis of focal amplification, microdeletions, and ladder structures consistent with breakage-fusion-bridge cycle events in osteosarcoma. Genes Chromosomes Cancer, 42(4), 392-403. doi: 10.1002/gcc.20157

Linabery, A. M., & Ross, J. A. (2008). Trends in childhood cancer incidence in the U.S. (1992-2004). Cancer, 112(2), 416-432. doi: 10.1002/cncr.23169

Lipska, B. S., Koczkowska, M., Wierzba, J., Ploszynska, A., Iliszko, M., Izycka- Swieszewska, E., Adamkiewicz-Drozynska, E., & Limon, J. (2013). On the significance of germline cytogenetic rearrangements at MYCN locus in neuroblastoma. Mol Cytogenet, 6(1), 43. doi: 10.1186/1755-8166-6-43

Longhi, A., Errani, C., De Paolis, M., Mercuri, M., & Bacci, G. (2006). Primary bone osteosarcoma in the pediatric age: state of the art. Cancer Treat Rev, 32(6), 423- 436. doi: 10.1016/j.ctrv.2006.05.005

221

Lopes, M. A., Nikitakis, N. G., Ord, R. A., & Sauk, J., Jr. (2001). Amplification and protein expression of chromosome 12q13-15 genes in osteosarcomas of the jaws. Oral Oncol, 37(7), 566-571.

Ludwig, J. A. (2008). Ewing sarcoma: historical perspectives, current state-of-the-art, and opportunities for targeted therapy in the future. Curr Opin Oncol, 20(4), 412-418. doi: 10.1097/CCO.0b013e328303ba1d

Macdonald, J. R., Ziman, R., Yuen, R. K., Feuk, L., & Scherer, S. W. (2014). The Database of Genomic Variants: a curated collection of structural variation in the human genome. Nucleic Acids Res, 42(1), D986-992. doi: 10.1093/nar/gkt958

Machado, I., Lopez-Guerrero, J. A., Calabuig-Farinas, S., Hardy, J. R., Scotlandi, K., Picci, P., Byrne, J. A., & Llombart-Bosch, A. (2011). Clinical significance of tumor protein D52 immunostaining in a large series of Ewing's sarcoma family of tumors. Pediatr Dev Pathol, 14(3), 255-256. doi: 10.2350/11-01-0956-LET.1

Mackintosh, C., Madoz-Gurpide, J., Ordonez, J. L., Osuna, D., & Herrero-Martin, D. (2010). The molecular pathogenesis of Ewing's sarcoma. Cancer Biol Ther, 9(9), 655-667.

Maire, G., Yoshimoto, M., Chilton-MacNeill, S., Thorner, P. S., Zielenska, M., & Squire, J. A. (2009). Recurrent RECQL4 imbalance and increased gene expression levels are associated with structural chromosomal instability in sporadic osteosarcoma. Neoplasia, 11(3), 260-268, 263p following 268.

Man, T. K., Lu, X. Y., Jaeweon, K., Perlaky, L., Harris, C. P., Shah, S., Ladanyi, M., Gorlick, R., Lau, C. C., & Rao, P. H. (2004). Genome-wide array comparative genomic hybridization analysis reveals distinct amplifications in osteosarcoma. BMC Cancer, 4, 45. doi: 10.1186/1471-2407-4-45

Mantripragada, K. K., Buckley, P. G., de Stahl, T. D., & Dumanski, J. P. (2004). Genomic microarrays in the spotlight. Drug Discov Today, 9(24 Suppl), S45-52.

Mao, X., Young, B. D., & Lu, Y. J. (2007). The application of single nucleotide polymorphism microarrays in cancer research. Curr Genomics, 8(4), 219-228.

Marella, N. V., Malyavantham, K. S., Wang, J., Matsui, S., Liang, P., & Berezney, R. (2009). Cytogenetic and cDNA microarray expression analysis of MCF10 human breast cancer progression cell lines. Cancer Res, 69(14), 5946-5953. doi: 10.1158/0008-5472.CAN-09-0420

Marina, N., Gebhardt, M., Teot, L., & Gorlick, R. (2004). Biology and therapeutic advances for pediatric osteosarcoma. Oncologist, 9(4), 422-441.

Martin, J. W., Squire, J. A., & Zielenska, M. (2012). The genetics of osteosarcoma. Sarcoma, 2012, 627254. doi: 10.1155/2012/627254

222

Mattison, J., Kool, J., Uren, A. G., de Ridder, J., Wessels, L., Jonkers, J., Bignell, G. R., Butler, A., Rust, A. G., Brosch, M., Wilson, C. H., van der Weyden, L., Largaespada, D. A., Stratton, M. R., Futreal, P. A., van Lohuizen, M., Berns, A., Collier, L. S., Hubbard, T., & Adams, D. J. (2010). Novel candidate cancer genes identified by a large-scale cross-species comparative oncogenomics approach. Cancer Res, 70(3), 883-895. doi: 10.1158/0008-5472.CAN-09-1737

Mattison, J., van der Weyden, L., Hubbard, T., & Adams, D. J. (2009). Cancer gene discovery in mouse and man. Biochim Biophys Acta, 1796(2), 140-161. doi: 10.1016/j.bbcan.2009.03.001

McGinnis, K. S., Lessin, S. R., Elder, D. E., Guerry, D. t., Schuchter, L., Ming, M., & Elenitsas, R. (2002). Pathology review of cases presenting to a multidisciplinary pigmented lesion clinic. Arch Dermatol, 138(5), 617-621.

Mellman, I., Coukos, G., & Dranoff, G. (2011). Cancer immunotherapy comes of age. Nature, 480(7378), 480-489. doi: 10.1038/nature10673

Messahel, B., Hing, S., Nash, R., Jeffrey, I., & Pritchard-Jones, K. (2005). Clinical features of molecular pathology of solid tumours in childhood. Lancet Oncol, 6(6), 421-430. doi: 10.1016/S1470-2045(05)70209-6

Meyer, W. H., & Spunt, S. L. (2004). Soft tissue sarcomas of childhood. Cancer Treat Rev, 30(3), 269-280. doi: 10.1016/j.ctrv.2003.11.001

Mirabello, L., Berndt, S. I., Seratti, G. F., Burdett, L., Yeager, M., Chowdhury, S., Teshome, K., Uzoka, A., Douglass, C., Hayes, R. B., Hoover, R. N., Savage, S. A., & National Osteosarcoma Etiology Study, G. (2010). Genetic variation at chromosome 8q24 in osteosarcoma cases and controls. Carcinogenesis, 31(8), 1400-1404. doi: 10.1093/carcin/bgq117

Missiaglia, E., Williamson, D., Chisholm, J., Wirapati, P., Pierron, G., Petel, F., Concordet, J. P., Thway, K., Oberlin, O., Pritchard-Jones, K., Delattre, O., Delorenzi, M., & Shipley, J. (2012). PAX3/FOXO1 fusion gene status is the key prognostic molecular marker in rhabdomyosarcoma and significantly improves current risk stratification. J Clin Oncol, 30(14), 1670-1677. doi: 10.1200/JCO.2011.38.5591

Mitelman, F., Johansson, B., & Mertens, F. (2004). Fusion genes and rearranged genes as a linear function of chromosome aberrations in cancer. Nat Genet, 36(4), 331-334. doi: 10.1038/ng1335

Mohseny, A. B., Szuhai, K., Romeo, S., Buddingh, E. P., Briaire-de Bruijn, I., de Jong, D., van Pel, M., Cleton-Jansen, A. M., & Hogendoorn, P. C. (2009). Osteosarcoma originates from mesenchymal stem cells in consequence of aneuploidization and genomic loss of Cdkn2. J Pathol, 219(3), 294-305. doi: 10.1002/path.2603

223

Montagna, M., Dalla Palma, M., Menin, C., Agata, S., De Nicolo, A., Chieco-Bianchi, L., & D'Andrea, E. (2003). Genomic rearrangements account for more than one-third of the BRCA1 mutations in northern Italian breast/ovarian cancer families. Hum Mol Genet, 12(9), 1055-1061.

Morozova, O., & Marra, M. A. (2008). From cytogenetics to next-generation sequencing technologies: advances in the detection of genome rearrangements in tumors. Biochem Cell Biol, 86(2), 81-91. doi: 10.1139/O08-003

Mosse, Y. P., Diskin, S. J., Wasserman, N., Rinaldi, K., Attiyeh, E. F., Cole, K., Jagannathan, J., Bhambhani, K., Winter, C., & Maris, J. M. (2007). Neuroblastomas have distinct genomic DNA profiles that predict clinical phenotype and regional gene expression. Genes Chromosomes Cancer, 46(10), 936-949. doi: 10.1002/gcc.20477

MRC-Holland. (2011). SALSA MLPA kit P014-A1 Chromosome 8 Version 30; 26-01- 2011 (pp. 1-13).

Mu, D., Chen, L., Zhang, X., See, L. H., Koch, C. M., Yen, C., Tong, J. J., Spiegel, L., Nguyen, K. C., Servoss, A., Peng, Y., Pei, L., Marks, J. R., Lowe, S., Hoey, T., Jan, L. Y., McCombie, W. R., Wigler, M. H., & Powers, S. (2003). Genomic amplification and oncogenic properties of the KCNK9 potassium channel gene. Cancer Cell, 3(3), 297-302.

Myllykangas, S., Bohling, T., & Knuutila, S. (2007). Specificity, selection and significance of gene amplifications in cancer. Semin Cancer Biol, 17(1), 42-55. doi: 10.1016/j.semcancer.2006.10.005

Myllykangas, S., Himberg, J., Bohling, T., Nagy, B., Hollmen, J., & Knuutila, S. (2006). DNA copy number amplification profiling of human neoplasms. Oncogene, 25(55), 7324-7332. doi: 10.1038/sj.onc.1209717

Myllykangas, S., Tikka, J., Bohling, T., Knuutila, S., & Hollmen, J. (2008). Classification of human cancers based on DNA copy number amplification modeling. BMC Med Genomics, 1, 15. doi: 10.1186/1755-8794-1-15

Narayan, G., Bourdon, V., Chaganti, S., Arias-Pulido, H., Nandula, S. V., Rao, P. H., Gissmann, L., Durst, M., Schneider, A., Pothuri, B., Mansukhani, M., Basso, K., Chaganti, R. S., & Murty, V. V. (2007). Gene dosage alterations revealed by cDNA microarray analysis in cervical cancer: identification of candidate amplified and overexpressed genes. Genes Chromosomes Cancer, 46(4), 373-384. doi: 10.1002/gcc.20418

Naylor, T. L., Greshock, J., Wang, Y., Colligon, T., Yu, Q. C., Clemmer, V., Zaks, T. Z., & Weber, B. L. (2005). High resolution genomic analysis of sporadic breast cancer using array-based comparative genomic hybridization. Breast Cancer Res, 7(6), R1186-1198. doi: 10.1186/bcr1356

224

Ng, T. L., O'Sullivan, M. J., Pallen, C. J., Hayes, M., Clarkson, P. W., Winstanley, M., Sorensen, P. H., Nielsen, T. O., & Horsman, D. E. (2007). Ewing sarcoma with novel translocation t(2;16) producing an in-frame fusion of FUS and FEV. J Mol Diagn, 9(4), 459-463. doi: 10.2353/jmoldx.2007.070009

Niba, E. T., Tran, V. K., Tuan-Pham le, A., Vu, D. C., Nguyen, N. K., Ta, V. T., Tran, T. H., Lee, T., Takeshima, Y., & Matsuo, M. (2014). Validation of ambiguous MLPA results by targeted next-generation sequencing discloses a nonsense mutation in the DMD gene. Clin Chim Acta, 436, 155-159. doi: 10.1016/j.cca.2014.05.018

Nicholson, J. M., & Cimini, D. (2013). Cancer karyotypes: survival of the fittest. Front Oncol, 3, 148. doi: 10.3389/fonc.2013.00148

Nielsen, M., Hes, F. J., Nagengast, F. M., Weiss, M. M., Mathus-Vliegen, E. M., Morreau, H., Breuning, M. H., Wijnen, J. T., Tops, C. M., & Vasen, H. F. (2007). Germline mutations in APC and MUTYH are responsible for the majority of families with attenuated familial adenomatous polyposis. Clin Genet, 71(5), 427-433. doi: 10.1111/j.1399-0004.2007.00766.x

Nishijo, K., Nakayama, T., Aoyama, T., Okamoto, T., Ishibe, T., Yasura, K., Shima, Y., Shibata, K. R., Tsuboyama, T., Nakamura, T., & Toguchida, J. (2004). Mutation analysis of the RECQL4 gene in sporadic osteosarcomas. Int J Cancer, 111(3), 367- 372. doi: 10.1002/ijc.20269

Oakland, T. E., Haselton, K. J., & Randall, G. (2013). EWSR1 binds the hepatitis C virus cis-acting replication element and is required for efficient viral replication. J Virol, 87(12), 6625-6634. doi: 10.1128/JVI.01006-12

Oberthuer, A., Hero, B., Spitz, R., Berthold, F., & Fischer, M. (2004). The tumor- associated antigen PRAME is universally expressed in high-stage neuroblastoma and associated with poor outcome. Clin Cancer Res, 10(13), 4307-4313. doi: 10.1158/1078-0432.CCR-03-0813

OMIM#268400. ROTHMUND-THOMSON SYNDROME. from http://www.omim.org/entry/268400

Ottaviano, L., Schaefer, K. L., Gajewski, M., Huckenbeck, W., Baldus, S., Rogel, U., Mackintosh, C., de Alava, E., Myklebost, O., Kresse, S. H., Meza-Zepeda, L. A., Serra, M., Cleton-Jansen, A. M., Hogendoorn, P. C., Buerger, H., Aigner, T., Gabbert, H. E., & Poremba, C. (2010). Molecular characterization of commonly used cell lines for bone tumor research: a trans-European EuroBoNet effort. Genes Chromosomes Cancer, 49(1), 40-51. doi: 10.1002/gcc.20717

Ozaki, T., Schaefer, K. L., Wai, D., Buerger, H., Flege, S., Lindner, N., Kevric, M., Diallo, R., Bankfalvi, A., Brinkschmidt, C., Juergens, H., Winkelmann, W., Dockhorn- Dworniczak, B., Bielack, S. S., & Poremba, C. (2002). Genetic imbalances revealed

225

by comparative genomic hybridization in osteosarcomas. Int J Cancer, 102(4), 355- 365. doi: 10.1002/ijc.10709

Ozkinay, C., & Mitelman, F. (1979). A simple trypsin-Giemsa technique producing simultaneous G- and C-banding in human chromosomes. Hereditas, 90(1), 1-4.

Park, T. J., & Curran, T. (2008). Crk and Crk-like play essential overlapping roles downstream of disabled-1 in the Reelin pathway. J Neurosci, 28(50), 13551-13562. doi: 10.1523/JNEUROSCI.4323-08.2008

Paronetto, M. P., Minana, B., & Valcarcel, J. (2011). The Ewing sarcoma protein regulates DNA damage-induced alternative splicing. Mol Cell, 43(3), 353-368. doi: 10.1016/j.molcel.2011.05.035

Pasic, I., Shlien, A., Durbin, A. D., Stavropoulos, D. J., Baskin, B., Ray, P. N., Novokmet, A., & Malkin, D. (2010). Recurrent focal copy-number changes and loss of heterozygosity implicate two noncoding RNAs and one tumor suppressor gene at chromosome 3q13.31 in osteosarcoma. Cancer Res, 70(1), 160-171. doi: 10.1158/0008-5472.CAN-09-1902

Patrinos, G. P., & Petricoin, E. F. (2009). A new scientific journal linked to a genetic database: towards a novel publication modality. Hum Genomics Proteomics, 2009. doi: 10.4061/2009/597478

Paulson, V., Chandler, G., Rakheja, D., Galindo, R. L., Wilson, K., Amatruda, J. F., & Cameron, S. (2011). High-resolution array CGH identifies common mechanisms that drive embryonal rhabdomyosarcoma pathogenesis. Genes Chromosomes Cancer, 50(6), 397-408. doi: 10.1002/gcc.20864

Pavlov, A. R., Pavlova, N. V., Kozyavkin, S. A., & Slesarev, A. I. (2004). Recent developments in the optimization of thermostable DNA polymerases for efficient applications. Trends Biotechnol, 22(5), 253-260. doi: 10.1016/j.tibtech.2004.02.011

Paydas, S., Tanriverdi, K., Yavuz, S., & Seydaoglu, G. (2007). PRAME mRNA levels in cases with chronic leukemia: Clinical importance and review of the literature. Leuk Res, 31(3), 365-369. doi: 10.1016/j.leukres.2006.06.022

Pazzaglia, L., Chiechi, A., Conti, A., Gamberi, G., Magagnoli, G., Novello, C., Morandi, L., Picci, P., Mercuri, M., & Benassi, M. S. (2009). Genetic and molecular alterations in rhabdomyosarcoma: mRNA overexpression of MCL1 and MAP2K4 genes. Histol Histopathol, 24(1), 61-67.

Pearson, P. R. (2004). Neuroblastoma.

Pezzolo, A., Rossi, E., Gimelli, S., Parodi, F., Negri, F., Conte, M., Pistorio, A., Sementa, A., Pistoia, V., Zuffardi, O., & Gambini, C. (2009). Presence of 1q gain and absence of 7p gain are new predictors of local or metastatic relapse in localized

226

resectable neuroblastoma. Neuro Oncol, 11(2), 192-200. doi: 10.1215/15228517- 2008-086

Pinkel, D., & Albertson, D. G. (2005). Array comparative genomic hybridization and its applications in cancer. Nat Genet, 37 Suppl, S11-17. doi: 10.1038/ng1569

Poon, L. L., Chan, K. H., Smith, G. J., Leung, C. S., Guan, Y., Yuen, K. Y., & Peiris, J. S. (2009). Molecular detection of a novel human influenza (H1N1) of pandemic potential by conventional and real-time quantitative RT-PCR assays. Clin Chem, 55(8), 1555-1558. doi: 10.1373/clinchem.2009.130229

Prochownik, E. V., & Li, Y. (2007). The ever expanding role for c-Myc in promoting genomic instability. Cell Cycle, 6(9), 1024-1029.

Pui, C. H., Gajjar, A. J., Kane, J. R., Qaddoumi, I. A., & Pappo, A. S. (2011). Challenging issues in pediatric oncology. Nat Rev Clin Oncol, 8(9), 540-549. doi: 10.1038/nrclinonc.2011.95

Purnomosari, D., Aryandono, T., Setiaji, K., Nugraha, S. B., Pals, G., & van Diest, P. J. (2006). Comparison of multiplex ligation dependent probe amplification to immunohistochemistry for assessing HER-2/neu amplification in invasive breast cancer. Biotech Histochem, 81(2-3), 79-85. doi: 10.1080/10520290600822198

Quelin, C., Bendavid, C., Dubourg, C., de la Rochebrochard, C., Lucas, J., Henry, C., Jaillard, S., Loget, P., Loeuillet, L., Lacombe, D., Rival, J. M., David, V., Odent, S., & Pasquier, L. (2009). Twelve new patients with 13q deletion syndrome: genotype-phenotype analyses in progress. Eur J Med Genet, 52(1), 41-46. doi: 10.1016/j.ejmg.2008.10.002

Ragland, B. D., Bell, W. C., Lopez, R. R., & Siegal, G. P. (2002). Cytogenetics and molecular biology of osteosarcoma. Lab Invest, 82(4), 365-373.

Rao, P. K., Missiaglia, E., Shields, L., Hyde, G., Yuan, B., Shepherd, C. J., Shipley, J., & Lodish, H. F. (2010). Distinct roles for miR-1 and miR-133a in the proliferation and differentiation of rhabdomyosarcoma cells. FASEB J, 24(9), 3427-3437. doi: 10.1096/fj.09-150698

Raymond, & Robert T. Osteen, T. G. s. e. (2001). Clinical oncology. The American Cancer Society, Inc., Atlanta, Georgia 30329: Emily Pualwan.

Reichek, J. L., Duan, F., Smith, L. M., Gustafson, D. M., O'Connor, R. S., Zhang, C., Dunlevy, M. J., Gastier-Foster, J. M., & Barr, F. G. (2011). Genomic and clinical analysis of amplification of the 13q31 chromosomal region in alveolar rhabdomyosarcoma: a report from the Children's Oncology Group. Clin Cancer Res, 17(6), 1463-1473. doi: 10.1158/1078-0432.CCR-10-0091

227

Riaz, M., van Jaarsveld, M. T., Hollestelle, A., Prager-van der Smissen, W. J., Heine, A. A., Boersma, A. W., Liu, J., Helmijr, J., Ozturk, B., Smid, M., Wiemer, E. A., Foekens, J. A., & Martens, J. W. (2013). miRNA expression profiling of 51 human breast cancer cell lines reveals subtype and driver mutation-specific miRNAs. Breast Cancer Res, 15(2), R33. doi: 10.1186/bcr3415

Roberts, P., Burchill, S. A., Brownhill, S., Cullinane, C. J., Johnston, C., Griffiths, M. J., McMullan, D. J., Bown, N. P., Morris, S. P., & Lewis, I. J. (2008). Ploidy and karyotype complexity are powerful prognostic indicators in the Ewing's sarcoma family of tumors: a study by the United Kingdom Cancer Cytogenetics and the Children's Cancer and Leukaemia Group. Genes Chromosomes Cancer, 47(3), 207- 220. doi: 10.1002/gcc.20523

Rodriguez, V., Chen, Y., Elkahloun, A., Dutra, A., Pak, E., & Chandrasekharappa, S. (2007). Chromosome 8 BAC array comparative genomic hybridization and expression analysis identify amplification and overexpression of TRMT12 in breast cancer. Genes Chromosomes Cancer, 46(7), 694-707. doi: 10.1002/gcc.20454

Rodriguez-Pinilla, S. M., Jones, R. L., Lambros, M. B., Arriola, E., Savage, K., James, M., Pinder, S. E., & Reis-Filho, J. S. (2007). MYC amplification in breast cancer: a chromogenic in situ hybridisation study. J Clin Pathol, 60(9), 1017-1023. doi: 10.1136/jcp.2006.043869

Roessler, S., Long, E. L., Budhu, A., Chen, Y., Zhao, X., Ji, J., Walker, R., Jia, H. L., Ye, Q. H., Qin, L. X., Tang, Z. Y., He, P., Hunter, K. W., Thorgeirsson, S. S., Meltzer, P. S., & Wang, X. W. (2012). Integrative genomic identification of genes on 8p associated with hepatocellular carcinoma progression and patient survival. Gastroenterology, 142(4), 957-966 e912. doi: 10.1053/j.gastro.2011.12.039

Rondon-Lagos, M., Verdun Di Cantogno, L., Marchio, C., Rangel, N., Payan-Gomez, C., Gugliotta, P., Botta, C., Bussolati, G., Ramirez-Clavijo, S. R., Pasini, B., & Sapino, A. (2014). Differences and homologies of chromosomal alterations within and between breast cancer cell lines: a clustering analysis. Mol Cytogenet, 7(1), 8. doi: 10.1186/1755-8166-7-8

Roslan, N., Bieche, I., Bright, R. K., Lidereau, R., Chen, Y., & Byrne, J. A. (2013). TPD52 Represents a Survival Factor in ERBB2-Amplified Breast Cancer Cells. Mol Carcinog. doi: 10.1002/mc.22038

Rothenberg, S. M., & Settleman, J. (2010). Discovering tumor suppressor genes through genome-wide copy number analysis. Curr Genomics, 11(5), 297-310. doi: 10.2174/138920210791616734

Ruijs, M. W., Verhoef, S., Rookus, M. A., Pruntel, R., van der Hout, A. H., Hogervorst, F. B., Kluijt, I., Sijmons, R. H., Aalfs, C. M., Wagner, A., Ausems, M. G., Hoogerbrugge, N., van Asperen, C. J., Gomez Garcia, E. B., Meijers-Heijboer, H., Ten Kate, L. P., Menko, F. H., & van 't Veer, L. J. (2010). TP53 germline mutation testing in 180 families suspected of Li-Fraumeni syndrome: mutation detection rate

228

and relative frequency of cancers in different familial phenotypes. J Med Genet, 47(6), 421-428. doi: 10.1136/jmg.2009.073429

Ruiz, C., Seibt, S., Al Kuraya, K., Siraj, A. K., Mirlacher, M., Schraml, P., Maurer, R., Spichtin, H., Torhorst, J., Popovska, S., Simon, R., & Sauter, G. (2006). Tissue microarrays for comparing molecular features with proliferation activity in breast cancer. Int J Cancer, 118(9), 2190-2194. doi: 10.1002/ijc.21581

Rumiato, E., Pasello, G., Montagna, M., Scaini, M. C., De Salvo, G. L., Parenti, A., Cagol, M., Ruol, A., Ancona, E., Amadori, A., & Saggioro, D. (2011). DNA copy number profile discriminates between esophageal adenocarcinoma and squamous cell carcinoma and represents an independent prognostic parameter in esophageal adenocarcinoma. Cancer Lett, 310(1), 84-93. doi: 10.1016/j.canlet.2011.06.017

Rummukainen, J., Kytola, S., Karhu, R., Farnebo, F., Larsson, C., & Isola, J. J. (2001). Aberrations of chromosome 8 in 16 breast cancer cell lines by comparative genomic hybridization, fluorescence in situ hybridization, and spectral karyotyping. Cancer Genet Cytogenet, 126(1), 1-7.

Rytting, M., Pearson, P., Raymond, A. K., Ayala, A., Murray, J., Yasko, A. W., Johnson, M., & Jaffe, N. (2000). Osteosarcoma in preadolescent patients. Clin Orthop Relat Res(373), 39-50.

Sadikovic, B., Park, P. C., Selvarajah, S., & Zielenska, M. (2013). Array comparative genomic hybridization in osteosarcoma. Methods Mol Biol, 973, 227-247. doi: 10.1007/978-1-62703-281-0_15

Sadikovic, B., Yoshimoto, M., Chilton-MacNeill, S., Thorner, P., Squire, J. A., & Zielenska, M. (2009). Identification of interactive networks of gene expression associated with osteosarcoma oncogenesis by integrated molecular profiling. Hum Mol Genet, 18(11), 1962-1975. doi: 10.1093/hmg/ddp117

Saglam, O., Shah, V., & Worsham, M. J. (2007). Molecular differentiation of early and late stage laryngeal squamous cell carcinoma: an exploratory analysis. Diagn Mol Pathol, 16(4), 218-221. doi: 10.1097/PDM.0b013e3180d0aab5

Saha, M. N., Qiu, L., & Chang, H. (2013). Targeting p53 by small molecules in hematological malignancies. J Hematol Oncol, 6, 23. doi: 10.1186/1756-8722-6-23

Sails, A. D., Saunders, D., Airs, S., Roberts, D., Eltringham, G., & Magee, J. G. (2009). Evaluation of the Cepheid respiratory syncytial virus and influenza virus A/B real-time PCR analyte specific reagent. J Virol Methods, 162(1-2), 88-90. doi: 10.1016/j.jviromet.2009.07.020

Saito, S., Morita, K., & Hirano, T. (2009). High frequency of common DNA copy number abnormalities detected by bacterial artificial chromosome array comparative

229

genomic hybridization in 24 breast cancer cell lines. Hum Cell, 22(1), 1-10. doi: 10.1111/j.1749-0774.2008.00061.x

Sandberg, A. A., & Bridge, J. A. (2000). Updates on cytogenetics and molecular genetics of bone and soft tissue tumors: Ewing sarcoma and peripheral primitive neuroectodermal tumors. Cancer Genet Cytogenet, 123(1), 1-26.

Sandberg, A. A., & Bridge, J. A. (2003). Updates on the cytogenetics and molecular genetics of bone and soft tissue tumors: osteosarcoma and related tumors. Cancer Genet Cytogenet, 145(1), 1-30.

Saramaki, O., & Visakorpi, T. (2007). Chromosomal aberrations in prostate cancer. Front Biosci, 12, 3287-3301.

Savola, S., Klami, A., Tripathi, A., Niini, T., Serra, M., Picci, P., Kaski, S., Zambelli, D., Scotlandi, K., & Knuutila, S. (2009). Combined use of expression and CGH arrays pinpoints novel candidate genes in Ewing sarcoma family of tumors. BMC Cancer, 9, 17. doi: 10.1186/1471-2407-9-17

Schilling, F. H., Spix, C., Berthold, F., Erttmann, R., Fehse, N., Hero, B., Klein, G., Sander, J., Schwarz, K., Treuner, J., Zorn, U., & Michaelis, J. (2002). Neuroblastoma screening at one year of age. N Engl J Med, 346(14), 1047-1053. doi: 10.1056/NEJMoa012277

Schmitz, R., Young, R. M., Ceribelli, M., Jhavar, S., Xiao, W., Zhang, M., Wright, G., Shaffer, A. L., Hodson, D. J., Buras, E., Liu, X., Powell, J., Yang, Y., Xu, W., Zhao, H., Kohlhammer, H., Rosenwald, A., Kluin, P., Muller-Hermelink, H. K., Ott, G., Gascoyne, R. D., Connors, J. M., Rimsza, L. M., Campo, E., Jaffe, E. S., Delabie, J., Smeland, E. B., Ogwang, M. D., Reynolds, S. J., Fisher, R. I., Braziel, R. M., Tubbs, R. R., Cook, J. R., Weisenburger, D. D., Chan, W. C., Pittaluga, S., Wilson, W., Waldmann, T. A., Rowe, M., Mbulaiteye, S. M., Rickinson, A. B., & Staudt, L. M. (2012). Burkitt lymphoma pathogenesis and therapeutic targets from structural and functional genomics. Nature, 490(7418), 116-120. doi: 10.1038/nature11378

Schouten, J. P., McElgunn, C. J., Waaijer, R., Zwijnenburg, D., Diepvens, F., & Pals, G. (2002). Relative quantification of 40 nucleic acid sequences by multiplex ligation- dependent probe amplification. Nucleic Acids Res, 30(12), e57.

Schwab, M., Westermann, F., Hero, B., & Berthold, F. (2003). Neuroblastoma: biology and molecular and chromosomal pathology. Lancet Oncol, 4(8), 472-480.

Schwaenen, C., Nessling, M., Wessendorf, S., Salvi, T., Wrobel, G., Radlwimmer, B., Kestler, H. A., Haslinger, C., Stilgenbauer, S., Dohner, H., Bentz, M., & Lichter, P. (2004). Automated array-based genomic profiling in chronic lymphocytic leukemia: development of a clinical tool and discovery of recurrent genomic alterations. Proc Natl Acad Sci U S A, 101(4), 1039-1044. doi: 10.1073/pnas.0304717101

230

Scotting, P. J., Walker, D. A., & Perilongo, G. (2005). Childhood solid tumours: a developmental disorder. Nat Rev Cancer, 5(6), 481-488. doi: 10.1038/nrc1633

Shaffer, L. G., & Bejjani, B. A. (2004). A cytogeneticist's perspective on genomic microarrays. Hum Reprod Update, 10(3), 221-226. doi: 10.1093/humupd/dmh022

Shen, Y., & Wu, B. L. (2009). Microarray-based genomic DNA profiling technologies in clinical molecular diagnostics. Clin Chem, 55(4), 659-669. doi: 10.1373/clinchem.2008.112821

Shern, J. F., Chen, L., Chmielecki, J., Wei, J. S., Patidar, R., Rosenberg, M., Ambrogio, L., Auclair, D., Wang, J., Song, Y. K., Tolman, C., Hurd, L., Liao, H., Zhang, S., Bogen, D., Brohl, A. S., Sindiri, S., Catchpoole, D., Badgett, T., Getz, G., Mora, J., Anderson, J. R., Skapek, S. X., Barr, F. G., Meyerson, M., Hawkins, D. S., & Khan, J. (2014). Comprehensive genomic analysis of rhabdomyosarcoma reveals a landscape of alterations affecting a common genetic axis in fusion-positive and fusion-negative tumors. Cancer Discov, 4(2), 216-231. doi: 10.1158/2159-8290.CD- 13-0639

Shimada, S., Ishizawa, T., Ishizawa, K., Matsumura, T., Hasegawa, T., & Hirose, T. (2006). The value of MDM2 and CDK4 amplification levels using real-time polymerase chain reaction for the differential diagnosis of liposarcomas and their histologic mimickers. Hum Pathol, 37(9), 1123-1129. doi: 10.1016/j.humpath.2006.04.010

Shimada., Ambros, I. M., Dehner, L. P., Hata, J., Joshi, V. V., Roald, B., Stram, D. O., Gerbing, R. B., Lukens, J. N., Matthay, K. K., & Castleberry, R. P. (1999). The International Neuroblastoma Pathology Classification (the Shimada system). Cancer, 86(2), 364-372.

Shinawi, M., & Cheung, S. W. (2008). The array CGH and its clinical applications. Drug Discov Today, 13(17-18), 760-770. doi: 10.1016/j.drudis.2008.06.007

Shing, D. C., McMullan, D. J., Roberts, P., Smith, K., Chin, S. F., Nicholson, J., Tillman, R. M., Ramani, P., Cullinane, C., & Coleman, N. (2003). FUS/ERG gene fusions in Ewing's tumors. Cancer Res, 63(15), 4568-4576.

Slater, H. R., Bruno, D. L., Ren, H., Pertile, M., Schouten, J. P., & Choo, K. H. (2003). Rapid, high throughput prenatal detection of aneuploidy using a novel quantitative method (MLPA). J Med Genet, 40(12), 907-912.

Smida, J., Baumhoer, D., Rosemann, M., Walch, A., Bielack, S., Poremba, C., Remberger, K., Korsching, E., Scheurlen, W., Dierkes, C., Burdach, S., Jundt, G., Atkinson, M. J., & Nathrath, M. (2010). Genomic alterations and allelic imbalances are strong prognostic predictors in osteosarcoma. Clin Cancer Res, 16(16), 4256- 4267. doi: 10.1158/1078-0432.CCR-10-0284

231

Smith, M. A., Gorlick, R., Kolb, E. A., Lock, R., Carol, H., Maris, J. M., Keir, S. T., Morton, C. L., Reynolds, C. P., Kang, M. H., Arts, J., Bashir, T., Janicot, M., Kurmasheva, R. T., & Houghton, P. J. (2012). Initial testing of JNJ-26854165 (Serdemetan) by the pediatric preclinical testing program. Pediatr Blood Cancer, 59(2), 329-332. doi: 10.1002/pbc.23319

Snijders, A. M., Fridlyand, J., Mans, D. A., Segraves, R., Jain, A. N., Pinkel, D., & Albertson, D. G. (2003). Shaping of tumor and drug-resistant genomes by instability and selection. Oncogene, 22(28), 4370-4379. doi: 10.1038/sj.onc.1206482

Spraker, H. L., Price, S. L., Chaturvedi, A., Schiffman, J. D., Jones, K. B., Lessnick, S. L., Beckerle, M., & Randall, R. L. (2012). The clone wars - revenge of the metastatic rogue state: the sarcoma paradigm. Front Oncol, 2, 2. doi: 10.3389/fonc.2012.00002

Squire, J. A., Pei, J., Marrano, P., Beheshti, B., Bayani, J., Lim, G., Moldovan, L., & Zielenska, M. (2003). High-resolution mapping of amplifications and deletions in pediatric osteosarcoma by use of CGH analysis of cDNA microarrays. Genes Chromosomes Cancer, 38(3), 215-225. doi: 10.1002/gcc.10273

Sriram, G., & Birge, R. B. (2010). Emerging roles for crk in human cancer. Genes Cancer, 1(11), 1132-1139. doi: 10.1177/1947601910397188

Steliarova-Foucher, E., Stiller, C., Lacour, B., & Kaatsch, P. (2005). International Classification of Childhood Cancer, third edition. Cancer, 103(7), 1457-1467. doi: 10.1002/cncr.20910

Stern, R. F., Roberts, R. G., Mann, K., Yau, S. C., Berg, J., & Ogilvie, C. M. (2004). Multiplex ligation-dependent probe amplification using a completely synthetic probe set. Biotechniques, 37(3), 399-405.

Stock, C., Kager, L., Fink, F. M., Gadner, H., & Ambros, P. F. (2000). Chromosomal regions involved in the pathogenesis of osteosarcomas. Genes Chromosomes Cancer, 28(3), 329-336.

Stuppia, L., Antonucci, I., Palka, G., & Gatta, V. (2012). Use of the MLPA Assay in the Molecular Diagnosis of Gene Copy Number Alterations in Human Genetic Diseases. Int J Mol Sci, 13(3), 3245-3276. doi: 10.3390/ijms13033245

Sumegi, J., Streblow, R., Frayer, R. W., Dal Cin, P., Rosenberg, A., Meloni-Ehrig, A., & Bridge, J. A. (2010). Recurrent t(2;2) and t(2;8) translocations in rhabdomyosarcoma without the canonical PAX-FOXO1 fuse PAX3 to members of the nuclear receptor transcriptional coactivator family. Genes Chromosomes Cancer, 49(3), 224-236. doi: 10.1002/gcc.20731

Surdez, D., Benetkiewicz, M., Perrin, V., Han, Z. Y., Pierron, G., Ballet, S., Lamoureux, F., Redini, F., Decouvelaere, A. V., Daudigeos-Dubus, E., Geoerger, B.,

232

de Pinieux, G., Delattre, O., & Tirode, F. (2012). Targeting the EWSR1-FLI1 oncogene-induced protein kinase PKC-beta abolishes ewing sarcoma growth. Cancer Res, 72(17), 4494-4503. doi: 10.1158/0008-5472.CAN-12-0371

Suva, M. L., Riggi, N., Stehle, J. C., Baumer, K., Tercier, S., Joseph, J. M., Suva, D., Clement, V., Provero, P., Cironi, L., Osterheld, M. C., Guillou, L., & Stamenkovic, I. (2009). Identification of cancer stem cells in Ewing's sarcoma. Cancer Res, 69(5), 1776-1781. doi: 10.1158/0008-5472.CAN-08-2242

Szuhai, K., Cleton-Jansen, A. M., Hogendoorn, P. C., & Bovee, J. V. (2012). Molecular pathology and its diagnostic use in bone tumors. Cancer Genet, 205(5), 193-204. doi: 10.1016/j.cancergen.2012.04.001

Takata, M., Suzuki, T., Ansai, S., Kimura, T., Shirasaki, F., Hatta, N., & Saida, T. (2005). Genome profiling of melanocytic tumors using multiplex ligation-dependent probe amplification (MLPA): Its usefulness as an adjunctive diagnostic tool for melanocytic tumors. J Dermatol Sci, 40(1), 51-57. doi: 10.1016/j.jdermsci.2005.06.009

Tan, Choong, P. F., & Dass, C. R. (2009). Osteosarcoma: Conventional treatment vs. gene therapy. Cancer Biol Ther, 8(2), 106-117.

Tan, Zou, C., Yong, B., Han, J., Zhang, L., Su, Q., Yin, J., Wang, J., Huang, G., Peng, T., & Shen, J. (2012). Expression and prognostic relevance of PRAME in primary osteosarcoma. Biochem Biophys Res Commun, 419(4), 801-808. doi: 10.1016/j.bbrc.2012.02.110

Tang, N., Song, W. X., Luo, J., Haydon, R. C., & He, T. C. (2008). Osteosarcoma development and stem cell differentiation. Clin Orthop Relat Res, 466(9), 2114-2130. doi: 10.1007/s11999-008-0335-z

Thermo-scientific. (2011). T042-Technical bulletin. In T. F. Scientific (Ed.).

Thomassen, M., Tan, Q., & Kruse, T. A. (2009). Gene expression meta-analysis identifies chromosomal regions and candidate genes involved in breast cancer metastasis. Breast Cancer Res Treat, 113(2), 239-249. doi: 10.1007/s10549-008- 9927-2

Tomlins, S. A., Mehra, R., Rhodes, D. R., Cao, X., Wang, L., Dhanasekaran, S. M., Kalyana-Sundaram, S., Wei, J. T., Rubin, M. A., Pienta, K. J., Shah, R. B., & Chinnaiyan, A. M. (2007). Integrative molecular concept modeling of prostate cancer progression. Nat Genet, 39(1), 41-51. doi: 10.1038/ng1935

Toomey, E. C., Schiffman, J. D., & Lessnick, S. L. (2010). Recent advances in the molecular pathogenesis of Ewing's sarcoma. Oncogene, 29(32), 4504-4516. doi: 10.1038/onc.2010.205

233

Tracey et al. (2010). Cancer In NSW: Incidence and Mortality Report 2008. Sydney: Cancer Institute NSW, August 2010. © New South Wales Department of Health, August 2010.

Tsuji, K., Kawauchi, S., Saito, S., Furuya, T., Ikemoto, K., Nakao, M., Yamamoto, S., Oka, M., Hirano, T., & Sasaki, K. (2010). Breast cancer cell lines carry cell line- specific genomic alterations that are distinct from aberrations in breast cancer tissues: comparison of the CGH profiles between cancer cell lines and primary cancer tissues. BMC Cancer, 10, 15. doi: 10.1186/1471-2407-10-15

Turner, N., Pearson, A., Sharpe, R., Lambros, M., Geyer, F., Lopez-Garcia, M. A., Natrajan, R., Marchio, C., Iorns, E., Mackay, A., Gillett, C., Grigoriadis, A., Tutt, A., Reis-Filho, J. S., & Ashworth, A. (2010). FGFR1 amplification drives endocrine therapy resistance and is a therapeutic target in breast cancer. Cancer Res, 70(5), 2085-2094. doi: 10.1158/0008-5472.CAN-09-3746

Uitenbroek, D. G. (1997). SISA - Five by two exact. 2015, from http://www.quantitativeskills.com/sisa/statistics/fiveby2.htm

Uwineza, A., Hitayezu, J., Murorunkwere, S., Ndinkabandi, J., Kalala Malu, C. K., Caberg, J. H., Dideberg, V., Bours, V., & Mutesa, L. (2013). Genetic diagnosis of Duchenne and Becker muscular dystrophy using multiplex ligation-dependent probe amplification in Rwandan patients. J Trop Pediatr. doi: 10.1093/tropej/fmt090

Van Maerken, T., Rihani, A., Dreidax, D., De Clercq, S., Yigit, N., Marine, J. C., Westermann, F., De Paepe, A., Vandesompele, J., & Speleman, F. (2011). Functional analysis of the p53 pathway in neuroblastoma cells using the small- molecule MDM2 antagonist nutlin-3. Mol Cancer Ther, 10(6), 983-993. doi: 10.1158/1535-7163.MCT-10-1090

Van Maerken, T., Speleman, F., Vermeulen, J., Lambertz, I., De Clercq, S., De Smet, E., Yigit, N., Coppens, V., Philippe, J., De Paepe, A., Marine, J. C., & Vandesompele, J. (2006). Small-molecule MDM2 antagonists as a new therapy concept for neuroblastoma. Cancer Res, 66(19), 9646-9655. doi: 10.1158/0008- 5472.CAN-06-0792

Van Prooijen-Knegt, A. C., Van Hoek, J. F., Bauman, J. G., Van Duijn, P., Wool, I. G., & Van der Ploeg, M. (1982). In situ hybridization of DNA sequences in human metaphase chromosomes visualized by an indirect fluorescent immunocytochemical procedure. Exp Cell Res, 141(2), 397-407.

Van Roy, N., De Preter, K., Hoebeeck, J., Van Maerken, T., Pattyn, F., Mestdagh, P., Vermeulen, J., Vandesompele, J., & Speleman, F. (2009). The emerging molecular pathogenesis of neuroblastoma: implications for improved risk assessment and targeted therapy. Genome Med, 1(7), 74. doi: 10.1186/gm74

234

Vandesompele, J., De Preter, K., Pattyn, F., Poppe, B., Van Roy, N., De Paepe, A., & Speleman, F. (2002). Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes. Genome Biol, 3(7), RESEARCH0034.

Vanneman, M., & Dranoff, G. (2012). Combining immunotherapy and targeted therapies in cancer treatment. Nat Rev Cancer, 12(4), 237-251. doi: 10.1038/nrc3237

Veronese, L., Tournilhac, O., Combes, P., Prie, N., Pierre-Eymard, E., Guieze, R., Veyrat-Masson, R., Bay, J. O., Vago, P., & Tchirkov, A. (2013). Contribution of MLPA to routine diagnostic testing of recurrent genomic aberrations in chronic lymphocytic leukemia. Cancer Genet, 206(1-2), 19-25. doi: 10.1016/j.cancergen.2012.12.002

Wachtel, M., Dettling, M., Koscielniak, E., Stegmaier, S., Treuner, J., Simon- Klingenstein, K., Buhlmann, P., Niggli, F. K., & Schafer, B. W. (2004). Gene expression signatures identify rhabdomyosarcoma subtypes and detect a novel t(2;2)(q35;p23) translocation fusing PAX3 to NCOA1. Cancer Res, 64(16), 5539- 5545. doi: 10.1158/0008-5472.CAN-04-0844

Wagner, E. R., Luther, G., Zhu, G., Luo, Q., Shi, Q., Kim, S. H., Gao, J. L., Huang, E., Gao, Y., Yang, K., Wang, L., Teven, C., Luo, X., Liu, X., Li, M., Hu, N., Su, Y., Bi, Y., He, B. C., Tang, N., Luo, J., Chen, L., Zuo, G., Rames, R., Haydon, R. C., Luu, H. H., & He, T. C. (2011). Defective osteogenic differentiation in the development of osteosarcoma. Sarcoma, 2011, 325238. doi: 10.1155/2011/325238

Wallace, A. (2003). Analysis MLPA Dosage Data. from http://www.ngrl.org.uk/Manchester/sites/default/files/publications/MLPA/Gene Dosage/Analysing_Dosage_Data_BSHG_2003.pps

Wang. (2005). Biology of osteogenic sarcoma. Cancer J, 11(4), 294-305.

Wang. (2012). Childhood rhabdomyosarcoma: recent advances and prospective views. J Dent Res, 91(4), 341-350. doi: 10.1177/0022034511421490

Wang, Carter, J., & Lench, N. (2013a). Evaluation of real-time quantitative PCR as a standard cytogenetic diagnostic tool for confirmation of microarray (aCGH) results and determination of inheritance. Genet Test Mol Biomarkers, 17(11), 821-825. doi: 10.1089/gtmb.2013.0284

Wang, Gannavarapu, A., Kozinetz, C. A., Levy, M. L., Lewis, R. A., Chintagumpala, M. M., Ruiz-Maldanado, R., Contreras-Ruiz, J., Cunniff, C., Erickson, R. P., Lev, D., Rogers, M., Zackai, E. H., & Plon, S. E. (2003). Association between osteosarcoma and deleterious mutations in the RECQL4 gene in Rothmund-Thomson syndrome. J Natl Cancer Inst, 95(9), 669-674.

235

Wang, Suganuma, R., Ikegaki, N., Tang, X., Naranjo, A., McGrady, P., London, W. B., Hogarty, M. D., Gastier-Foster, J. M., Look, A. T., Park, J. R., Maris, J. M., Cohn, S. L., Seeger, R. C., & Shimada, H. (2013b). Neuroblastoma of undifferentiated subtype, prognostic significance of prominent nucleolar formation, and MYC/MYCN protein expression: a report from the Children's Oncology Group. Cancer, 119(20), 3718-3726. doi: 10.1002/cncr.28251

Watson, S. K., deLeeuw, R. J., Horsman, D. E., Squire, J. A., & Lam, W. L. (2007). Cytogenetically balanced translocations are associated with focal copy number alterations. Hum Genet, 120(6), 795-805. doi: 10.1007/s00439-006-0251-9

Welfare, A. I. o. H. a. (2009). A Picture of Australia's Children 2009.

Westermann, F., Muth, D., Benner, A., Bauer, T., Henrich, K. O., Oberthuer, A., Brors, B., Beissbarth, T., Vandesompele, J., Pattyn, F., Hero, B., Konig, R., Fischer, M., & Schwab, M. (2008). Distinct transcriptional MYCN/c-MYC activities are associated with spontaneous regression or malignant progression in neuroblastomas. Genome Biol, 9(10), R150. doi: 10.1186/gb-2008-9-10-r150

Williamson, D., Lu, Y. J., Gordon, T., Sciot, R., Kelsey, A., Fisher, C., Poremba, C., Anderson, J., Pritchard-Jones, K., & Shipley, J. (2005). Relationship between MYCN copy number and expression in rhabdomyosarcomas and correlation with adverse prognosis in the alveolar subtype. J Clin Oncol, 23(4), 880-888. doi: 10.1200/JCO.2005.11.078

Williamson, D., Missiaglia, E., de Reynies, A., Pierron, G., Thuille, B., Palenzuela, G., Thway, K., Orbach, D., Lae, M., Freneaux, P., Pritchard-Jones, K., Oberlin, O., Shipley, J., & Delattre, O. (2010). Fusion gene-negative alveolar rhabdomyosarcoma is clinically and molecularly indistinguishable from embryonal rhabdomyosarcoma. J Clin Oncol, 28(13), 2151-2158. doi: 10.1200/JCO.2009.26.3814

Woods, W. G., Gao, R. N., Shuster, J. J., Robison, L. L., Bernstein, M., Weitzman, S., Bunin, G., Levy, I., Brossard, J., Dougherty, G., Tuchman, M., & Lemieux, B. (2002). Screening of infants and mortality due to neuroblastoma. N Engl J Med, 346(14), 1041-1046. doi: 10.1056/NEJMoa012387

Xi, L., Nicastri, D. G., El-Hefnawy, T., Hughes, S. J., Luketich, J. D., & Godfrey, T. E. (2007). Optimal markers for real-time quantitative reverse transcription PCR detection of circulating tumor cells from melanoma, breast, colon, esophageal, head and neck, and lung cancers. Clin Chem, 53(7), 1206-1215. doi: 10.1373/clinchem.2006.081828

Yan, M., Himoudi, N., Basu, B. P., Wallace, R., Poon, E., Adams, S., Hasan, F., Xue, S. A., Wilson, N., Dalgleish, A., Williams, O., & Anderson, J. (2011). Increased PRAME antigen-specific killing of malignant cell lines by low avidity CTL clones, following treatment with 5-Aza-2'-Deoxycytidine. Cancer Immunol Immunother, 60(9), 1243-1255. doi: 10.1007/s00262-011-1024-4

236

Yatsenko, S. A., Shaw, C. A., Ou, Z., Pursley, A. N., Patel, A., Bi, W., Cheung, S. W., Lupski, J. R., Chinault, A. C., & Beaudet, A. L. (2009). Microarray-based comparative genomic hybridization using sex-matched reference DNA provides greater sensitivity for detection of sex chromosome imbalances than array- comparative genomic hybridization with sex-mismatched reference DNA. J Mol Diagn, 11(3), 226-237. doi: 10.2353/jmoldx.2009.080064

Yong, Z. W., Zaini, Z. M., Kallarakkal, T. G., Karen-Ng, L. P., Rahman, Z. A., Ismail, S. M., Sharifah, N. A., Mustafa, W. M., Abraham, M. T., Tay, K. K., & Zain, R. B. (2014). Genetic alterations of chromosome 8 genes in oral cancer. Sci Rep, 4, 6073. doi: 10.1038/srep06073

Yoshida, H., Miyachi, M., Sakamoto, K., Ouchi, K., Yagyu, S., Kikuchi, K., Kuwahara, Y., Tsuchiya, K., Imamura, T., Iehara, T., Kakazu, N., Hojo, H., & Hosoi, H. (2013). PAX3-NCOA2 fusion gene has a dual role in promoting the proliferation and inhibiting the myogenic differentiation of rhabdomyosarcoma cells. Oncogene. doi: 10.1038/onc.2013.491

Yoshioka, S., Tsukamoto, Y., Hijiya, N., Nakada, C., Uchida, T., Matsuura, K., Takeuchi, I., Seto, M., Kawano, K., & Moriyama, M. (2013). Genomic profiling of oral squamous cell carcinoma by array-based comparative genomic hybridization. PLoS One, 8(2), e56165. doi: 10.1371/journal.pone.0056165

Yu, Y., Luk, F., Yang, J. L., & Walsh, W. R. (2011). Ras/Raf/MEK/ERK pathway is associated with lung metastasis of osteosarcoma in an orthotopic mouse model. Anticancer Res, 31(4), 1147-1152.

Zhang, F., Gu, W., Hurles, M. E., & Lupski, J. R. (2009). Copy number variation in human health, disease, and evolution. Annu Rev Genomics Hum Genet, 10, 451- 481. doi: 10.1146/annurev.genom.9.081307.164217

Zipper, H., Brunner, H., Bernhagen, J., & Vitzthum, F. (2004). Investigations on DNA intercalation and surface binding by SYBR Green I, its structure determination and methodological implications. Nucleic Acids Res, 32(12), e103. doi: 10.1093/nar/gnh101

237

Appendices

238

Appendix 1. Statistical analysis

Statistical analysis using Fisher’s exact test to assess differences in the number of samples per copy number category at 8p versus 8q loci p-values <0.05 were considered statistically significant Categories: 2=Gain; 1=Disomic; 0=Loss

Osteosarcoma Category 2 Category 1 Category 0 8p 3 27 10 8q 19 82 9

p=0.01652 Ewing sarcoma Category 2 Category 1 Category 0 8p 4 59 1 8q 7 166 3 Fisher test p=0.70163

Rhabdomyosarcoma Cat 2 Cat 1 Cat 0 8p 3 61 0 8q 10 159 7

p=0.32322

ARMS Category 2 Category 1 Category 0 8p 0 16 0 8q 0 43 1

p=0.26667

ERMS Category 2 Category 1 Category 0 8p 3 45 0 8q 10 116 6

p=0.39302

Neuroblastoma Category 2 Category 1 Category 0 8p 10 58 12 8q 11 173 36 p=0.09000

239

Appendix 2. Copy number changes in cell lines

Genome copy number changes measured in 6 breast cell lines. Log2 copy number at individual probes plotted on same scale around the vertical baseline (left of baseline, copy number loss, right of baseline, copy number gain). Shown to the left of each ideogram, sections of which are coloured either red (copy number loss) or blue (copy number gain). All Genome views were generated using Nexus copy number v 7.5.

Cell line AU565

Cell line HCC2157

240

Cell line MCF-7

Cell line MCF10-A

241

Cell line MDA-MB-231

Cell line SK-BR-3

242

Appendix 3. Copy number changes in solid tumours

Genome copy number changes measured in paediatric solid tumours. Log2 copy number at individual probes plotted vertically on same scale (left of baseline, copy number loss, right of baseline, copy number gain). Shown to the left of each ideogram, sections of which are coloured either red (copy number loss) or blue (copy number gain). All Genome views were generated using Nexus copy number v 7.5.

Sample: F9O

Sample: F10O

243

Sample: M11O

Sample: F12O

244

Sample: M46O

245

Sample: M13E

Sample: M14E

246

Sample: F15E

Sample: M16E

247

Sample: F17E

Sample: F18E

248

Sample: M22E

Sample: F24E

249

Sample: M27R (ARMS)

Sample: M32R (ARMS)

250

Sample: M25R (ERMS)

Sample: F26R (ERMS)

251

Sample: F28R (ERMS)

Sample: F29R (ERMS)

252

Sample: M31R (ERMS)

Sample: M35R (ERMS)

253

Sample: M36N

Sample: F37N

254

Sample: F38N

Sample: M39N

255

Sample: F40N

Sample: M41N

256

Sample: F42N

Sample: M43N

257

Sample: F44N

Sample: F45N

258

Appendix 4. Copy number status using aCGH

Copy number status for chromosome 8p MLPA genes in solid tumour cohort using

aCGH

Genes DLGAP2 MFHAS1 MSRA GATA4 CTSB TUSC3 ChGn FGFR1 Samples F9O 1 1 1 1 1 1 1 1 F10O 1 1 1 1 1 1 1 0 M11O 2 2 2 2 1 2 2 2 F12O 1 2 1 1 1 1 1 1 M46O 1 1 1 1 1 1 1 1 M13E 2 1 1 1 1 1 1 1 M14E 2 1 2 1 1 2 2 1 F15E 2 2 2 2 2 2 2 2 M16E 2 2 2 2 2 2 2 2 F17E 1 1 1 1 1 1 1 1 F18E 1 1 1 1 1 1 1 1 M22E 1 0 1 1 1 1 1 1 F24E 1 1 1 1 1 1 1 1 M25R 2 1 2 1 1 1 1 1 F26R 1 1 1 1 1 1 1 1 M27R 1 1 1 1 0 1 1 0 F28R 1 1 1 1 1 1 1 1 F29R 2 1 1 1 1 1 1 1 M31R 1 1 1 1 1 1 1 1 M32R 1 1 1 1 1 1 1 1 M35R 1 1 1 1 1 1 1 1 M36N 1 1 1 1 1 1 1 1 F37N 1 1 1 1 1 1 1 1 F38N 1 1 1 1 1 1 1 1 M39N 2 1 2 1 1 1 1 2 F40N 2 1 1 1 1 1 1 1 M41N 2 1 1 1 1 1 1 1 F42N 1 1 1 1 1 1 1 1 M43N 1 1 1 1 1 1 1 1 F44N 2 2 2 2 2 2 2 2 F45N 1 1 1 1 1 1 1 1

DLGAP2 MFHAS1 MSRA GATA4 CTSB TUSC3 ChGn FGFR1

Gain 11(35.5%) 5(16.1%) 7(22.6%) 4(12.9%) 3(9.7%) 5(16.1%) 5(16.1%) 5(16.1%) Disomy 20(64.5%) 25(80.6%) 24(77.4%) 27(87.1%) 27(87.1%) 26(83.9%) 26(83.9%) 24(77.4%) Loss 0(0.0%) 1(3.3%) 0(0.0%) 0(0.0%) 1(3.3%) 0(0.0%) 0(0.0%) 2(6.5%)

Categories 2: Gain; 1=Disomy; 0=Loss

259

Copy number status for chromosome 8q MLPA genes in solid tumour cohort using

aCGH

Genes PRKDC MOS CHD7 MYBL1 NCOA2 TPD52 E2F5 RAD54B Samples F9O 1 2 2 2 2 2 2 2 F10O 1 0 0 2 0 1 1 1 M11O 1 0 1 1 0 2 1 2 F12O 2 2 2 2 2 0 0 2 M46O 2 1 2 2 1 2 1 2 M13E 1 1 2 2 1 1 1 1 M14E 1 1 2 2 2 1 1 2 F15E 2 2 2 2 2 2 2 2 M16E 2 1 2 2 2 2 2 1 F17E 1 1 1 1 1 1 1 1 F18E 1 1 1 1 1 1 1 1 M22E 1 1 1 1 1 0 1 1 F24E 1 1 1 1 1 1 1 1 M25R 1 1 1 1 1 1 1 1 F26R 1 1 1 1 1 1 1 1 M27R 0 1 1 1 0 0 0 0 F28R 1 1 1 1 1 1 1 1 F29R 1 1 1 1 1 2 2 2 M31R 1 1 1 1 2 2 1 1 M32R 1 1 1 1 1 1 1 1 M35R 1 1 1 1 1 1 1 1 M36N 1 1 1 1 1 1 1 1 F37N 1 1 1 1 1 1 1 1 F38N 1 1 1 1 1 1 1 1 M39N 2 2 1 2 1 1 2 0 F40N 1 1 1 0 1 1 1 1 M41N 1 1 1 1 1 1 1 1 F42N 1 1 1 1 1 1 1 1 M43N 1 1 1 1 1 1 1 1 F44N 2 2 2 2 2 2 2 2 F45N 1 1 1 1 1 1 1 1 PRKDC MOS CHD7 MYBL1 NCOA2 TPD52 E2F5 RAD54B Gain 6(19.4%) 5(16.1%) 8(25.8%) 10(32.3%) 7(22.6%) 8(25.8%) 6(19.4%) 8(25.8%) Disomy 24(77.4%) 24(77.4%) 22(71.0%) 20(64.5%) 21(67.7%) 20(64.5%) 23(74.2%) 21(67.7%) Loss 1(3.2%) 2(6.5%) 1(3.2%) 1(3.2%) 3(9.7%) 3(9.7%) 2(6.4%) 2(6.4%)

Categories: 2=Gain; 1=Disomy; 0=Loss

260

Copy number status for chromosome 8q MLPA genes in solid tumour cohort using aCGH (cont.)

Genes ST7 EIF3E EIF3H EXT1 RNF139 MYC DDEF1 KCNQ3 Samples F9O 2 2 2 2 2 2 2 2 F10O 1 1 1 1 1 1 1 1 M11O 2 2 2 2 1 1 1 2 F12O 0 2 2 2 2 2 1 2 M46O 2 2 2 2 2 2 2 2 M13E 1 1 1 1 1 1 1 1 M14E 2 2 2 2 1 1 2 2 F15E 2 2 2 2 2 2 2 2 M16E 2 2 2 2 2 2 2 2 F17E 1 1 1 1 1 1 1 1 F18E 1 1 1 1 1 1 1 1 M22E 1 1 1 1 1 1 1 1 F24E 1 1 1 1 1 1 1 1 M25R 1 1 1 1 1 1 1 2 F26R 1 1 1 1 1 1 1 1 M27R 1 0 1 1 0 0 0 1 F28R 1 1 1 1 1 1 1 1 F29R 2 2 1 1 1 2 1 1 M31R 1 1 2 2 1 1 1 1 M32R 1 1 1 1 1 1 1 1 M35R 1 1 1 1 1 1 1 2 M36N 1 1 1 1 1 1 1 1 F37N 1 1 1 2 2 1 1 1 F38N 1 1 1 2 1 1 1 1 M39N 1 1 1 2 1 2 2 2 F40N 1 1 1 1 1 1 1 1 M41N 1 1 1 1 1 1 1 1 F42N 1 1 1 1 1 1 1 1 M43N 1 1 1 1 1 1 1 1 F44N 2 2 2 2 2 2 2 2 F45N 1 1 1 1 1 1 1 1

ST7 EIF3E EIF3H EXT1 RNF139 MYC DDEF1 KCNQ3 Gain 8(25.8%) 9(29.0%) 9(29.0%) 12(38.7%) 7(22.6%) 8(25.8%) 7(22.6%) 11(35.5%) Disomy 22(71.0%) 21(67.7%) 22(71.0%) 19(61.3%) 23(74.2%) 22(71.0%) 23(74.2%) 20(64.5%) Loss 1(3.2%) 1(3.2%) 0(0.0%) 0(0.0%) 1(3.2%) 1(3.2%) 1(3.2%) 0(0.0%)

Categories: 2=Gain; 1=Disomy; 0=Loss

261

Copy number status for chromosome 8q MLPA genes in solid tumour cohort

using aCGH (cont.)

Genes SLA KHDRBS3 KCNK9 PTK2 PTP4A3 RECQL4 Samples F9O 2 2 2 2 2 2 F10O 1 1 2 2 2 2 M11O 2 2 2 1 1 1 F12O 2 2 2 2 2 1 M46O 2 2 2 2 2 2 M13E 1 1 1 1 2 2 M14E 2 2 1 2 1 1 F15E 2 2 2 2 2 1 M16E 2 2 2 2 2 1 F17E 1 1 1 1 2 1 F18E 1 1 1 1 1 1 M22E 1 1 1 1 1 1 F24E 1 1 1 1 1 2 M25R 2 2 1 1 1 2 F26R 1 1 1 1 2 2 M27R 1 1 1 0 1 1 F28R 1 1 1 0 1 2 F29R 1 1 1 1 1 2 M31R 2 2 1 1 1 2 M32R 1 1 1 1 1 2 M35R 1 1 1 1 2 2 M36N 1 1 1 1 2 1 F37N 1 1 1 1 1 1 F38N 1 1 1 1 1 0 M39N 2 1 1 2 2 1 F40N 1 1 2 1 1 2 M41N 1 1 2 1 2 1 F42N 1 1 1 1 1 1 M43N 1 1 1 1 1 2 F44N 2 2 2 2 2 2 F45N 1 1 1 1 1 1

SLA KHDRBS3 KCNK9 PTK2 PTP4A3 RECQL4 Gain 11(35.5%) 10(32.3%) 10(32.3%) 9(29.0%) 14(45.2%) 15(48.4%) Disomy 20(64.5%) 21(67.7%) 21(67.7%) 20(64.5%) 17(54.8%) 15(48.4%) Loss 0(0.0%) 0(0.0%) 0(0.0%) 2(6.5%) 0(0.0%) 1(3.2%)

Categories: 2=Gain; 1=Disomy; 0=Loss

262

Appendix 5. Copy number categories in cell lines

Copy number categories measured at chromosome 8p and 8q loci using MLPA or aCGH in breast cell lines

8p (8 genes) Chromosome 8q (22 genes)

ST7 SLA MYC MOS E2F5 EXT1 PTK2 ChGn CTSB CHD7 EIF3E MYBL EIF3H MSRA TPD52 PTP4A FGFR1 DDEF1 Genes TUSC3 GATA4 KCNK9 KCNQ3 NCOA2 PRKDC MFHAS RNF139 RAD54B RECQL4 DLGAP2 è KHDRBS3

Samples aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH ê MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA AU565 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 2 2 2 1 2 0 0 0 0 0 0 2 2 2 2 2 2 2 2 1 1 0 0 0 1 0 1 0 0 0 0 0 2 1 2 HCC2157 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 1 2 2 2 2 2 1 2 1 2 2 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 2 2 MCF-7 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 0 0 1 1 1 1 2 2 2 2 2 2 1 2 1 2 2 2 2 2 2 2 1 2 1 2 1 2 0 2 1 2 1 2 1 2 MCF-10A 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 2 1 2 1 2 1 2 2 2 1 2 1 2 2 2 1 2 1 2 1 2 1 2 1 1

MDA-MB-231 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 2 1 2 2 2 2 2 1 2 2 2 1 2 1 2 1 2 1 2 1 2 1 2 0 0 1 2 2 2 1 2 1 2 1 2 0 2 1 2 1 2 1 1 SK-BR-3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 2 1 2 1 2 1 2 1 2 2 2 1 2 0 0 0 0 0 0 2 2 2 2 2 2 2 2 1 1 0 1 1 1 0 1 0 0 0 0 0 2 0 2

Categories: 2=Gain; 1=Disomy; 0=Loss

263

Appendix 6. Copy number categories in solid tumours

Copy number categories measured at chromosome 8p and 8q loci using MLPA or aCGH in solid tumour cohort

8p (8 genes) Chromosome 8q (22 genes)

ST7 SLA MYC MOS E2F5 EXT1 PTK2 ChGn CTSB CHD7 EIF3E MYBL EIF3H

Genes MSRA TPD52 PTP4A FGFR1 DDEF1 TUSC3 GATA4 KCNK9 KCNQ3 NCOA2 PRKDC MFHAS RNF139 RAD54B RECQL4 DLGAP2

è KHDRBS3

Sample aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH

ê MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA

F9O 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 2 1 2 1 2 2 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 0 2 1 2 1 2 1 2 F10O 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 0 0 1 0 2 2 1 0 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 2 1 2 1 2 1 2 M11O 1 2 1 2 2 2 1 2 1 1 2 2 2 2 1 2 1 1 1 0 1 1 1 1 1 0 1 2 1 1 1 2 1 2 1 2 1 2 1 2 1 1 1 1 1 1 1 2 1 2 1 2 1 2 1 1 1 1 0 1 F12O 0 1 1 2 0 1 0 1 0 1 1 1 0 1 0 1 1 2 1 2 2 2 2 2 2 2 0 0 0 0 1 2 0 0 2 2 2 2 2 2 2 2 2 2 1 1 1 2 1 2 1 2 0 2 0 2 0 2 1 1 M46O 0 1 0 1 0 1 0 1 1 1 1 1 1 1 1 1 1 2 1 1 1 2 1 2 1 1 1 2 1 1 2 2 2 2 2 2 2 2 2 2 1 2 1 2 1 2 1 2 2 2 2 2 1 2 1 2 1 2 1 2 M13E 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 2 M14E 1 2 1 1 1 2 1 1 1 1 2 2 1 2 1 1 1 1 1 1 1 2 1 2 1 2 1 1 1 1 1 2 1 2 2 2 1 2 2 2 1 1 1 1 1 2 1 2 1 2 1 2 1 1 1 2 1 1 1 1 F15E 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 1 M16E 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 1 1 2 1 2 1 2 1 2 1 2 1 1 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 1 F17E 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 0 1 1 1 1 2 1 1 F18E 1 1 2 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 M22E 1 1 1 0 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 0 1 1 1 1 1 1 1 F24E 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 Categories: 2=Gain; 1=Disomy; 0=Loss

264

Copy number measured at chromosome 8p and 8q loci using MLPA or aCGH in solid tumour cohort (Cont.)

8p (8 genes) Chromosome 8q (22 genes)

ST7 SLA MYC MOS E2F5 EXT1 PTK2 ChGn CTSB CHD7 EIF3E MYBL EIF3H

Genes MSRA TPD52 PTP4A FGFR1 DDEF1 TUSC3 GATA4 KCNK9 KCNQ3 NCOA2 PRKDC MFHAS RNF139 RAD54B RECQL4 DLGAP2

è KHDRBS3

Samples aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH aCGH

ê MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA MLPA

M25R 1 2 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 2 1 2 1 1 1 1 1 1 1 2 F26R 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 2 M27R 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 0 1 0 1 1 1 1 1 1 1 0 1 0 1 0 1 0 1 1 1 0 1 1 1 1 1 0 1 0 1 0 1 1 1 1 1 1 1 1 1 0 1 1 1 1 F28R 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 2 F29R 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 1 2 2 2 1 2 1 2 1 1 0 1 1 1 2 2 1 1 1 1 1 1 1 1 0 1 1 1 0 1 1 2 M31R 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 2 1 1 1 1 2 2 2 1 1 1 1 1 1 1 1 1 2 1 2 1 1 1 1 1 1 1 1 1 2 1 2 1 1 1 1 1 1 1 2 M32R 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 M35R 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 2 1 1 1 2 1 1 1 2 1 1 1 2 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 2 2 1 1 1 0 1 1 1 0 2 1 2 M36N 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 F37N 1 1 2 1 0 1 1 1 2 1 1 1 0 1 2 1 0 1 1 1 1 1 0 1 1 1 0 1 0 1 1 1 0 1 1 1 0 1 0 2 0 2 1 1 1 1 0 1 1 1 0 1 1 1 1 1 1 1 2 1 F38N 2 1 1 1 0 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 0 M39N 1 2 2 1 0 2 1 1 1 1 1 1 0 1 1 2 0 2 1 2 0 1 0 2 1 1 0 1 0 2 0 0 0 1 1 1 0 1 1 2 1 1 1 2 1 2 1 2 1 2 0 1 1 1 1 2 1 2 2 1 F40N 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 0 1 1 2 M4N 1 2 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 2 2 1 F42N 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 M43N 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 F44N 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 F45N 2 1 0 1 0 1 0 1 2 1 2 1 0 1 0 1 0 1 2 1 1 1 2 1 0 1 2 1 1 1 1 1 0 1 2 1 0 1 0 1 0 1 1 1 0 1 0 1 2 1 0 1 0 1 0 1 0 1 0 1 Categories: 2=Gain; 1=Disomy; 0=Loss

265