New Approaches and Gene Targets in Molecular Pathogenesis of Colorectal Cancer

New approaches and gene targets in molecular pathogenesis of colorectal cancer

Author Lee, Katherine

Published 2018-11

Thesis Type Thesis (PhD Doctorate)

School School of Medicine

DOI https://doi.org/10.25904/1912/1616

Downloaded from http://hdl.handle.net/10072/384291

Griffith Research Online https://research-repository.griffith.edu.au

New approaches and gene targets in molecular pathogenesis of colorectal cancer

Katherine Ting-Wei Lee Master of Science in Medical and Health Sciences (By Research)

School of Medicine Griffith University

A dissertation submitted in fulfilment of the requirements for the degree of Doctor of Philosophy November 2018

Abstract

Two genes namely GAEC1 (gene amplified in oesophageal carcinoma one) and

FAM134B (family with sequence similarity 143, member B) were first identified in 2001.

Subsequently, further characterisation of their biological function in colorectal adenocarcinoma was done. While knock down of GAEC1 dampened cancer progression by inhibiting cellular functions such as proliferation and increasing apoptosis, knock down of

FAM134B enhanced these cellular functions and drove the progression of colon adenocarcinoma. These results were further supported with in vivo studies by xenotransplanting colon adenocarcinoma cells containing with either knock down GAEC1 or

FAM134B into immunodeficient mice. Results showed that mice xenotransplanted with colon adenocarcinoma cells with knock down GAEC1, lost the ability to form tumour whereas, mice xenotransplanted with colon adenocarcinoma cells with knock down FAM134B¸ formed larger tumour compared to the control mice. Thus, these studies confirm GAEC1 as a transforming oncogene and FAM134B as a tumour suppressor gene. Nonetheless, current literature lacks knowledge on the absolute copy number aberration (CNA) and ploidy number of GAEC1, in depth genetic and epigenetic regulation, functional insights, interacting signalling pathways and molecular partners of GAEC1 and FAM134B.

This study first developed a duplex DNA binding dye chemistry-based droplet digital

PCR (ddPCR) method and compared the efficiency with real-time PCR (qPCR), the current gold standard method for accurate detection of CNA and ploidy determination. Results showed that ddPCR has better reproducibility compared to qPCR. Both assays were comparable at 96.5% of the time. In addition, the run-time and cost per sample was greatly reduce, thus, providing a better option for use in clinical settings.

Following on, this study was the first to attempt identifying mutations of GAEC1. The ddPCR assay was first used to determine the GAEC1 CNA and its ploidy number in patients

2 with colorectal adenocarcinoma. Mutations within the exon sequence of GAEC1 was called using Sanger sequencing and the mRNA and protein expression were determined using qPCR and western blot (confirmed with immunofluorescence assay). To this end, 24% of patients with colorectal adenocarcinoma also have GAEC1 amplification. Additionally, GAEC1 mutations were observed in 10% of patients with colorectal adenocarcinoma including one missense mutation, four loss of heterozygosity and two substitutions. Increased GAEC1 copy and GAEC1 mutations were significantly associated with biological aggressiveness of the adenocarcinoma (P<0.05). Furthermore, GAEC1 mutations were frequently associated with

GAEC1 protein expression and is associated with poorer clinical outcomes, thus, indicating the oncogenic potential of GAEC1.

Subsequently, GAEC1 was overexpressed in colon adenocarcinoma and non-neoplastic colonic epithelial cells to confirm the oncogenic effects of GAEC1. Multiple in vitro assays including cell proliferation wound healing, clonogenic, apoptosis, cell cycle and extracellular flux were performed. In addition, several potential transcriptional factors that interact with

GAEC1 were identified using western blot analysis. Functional studies revealed that GAEC1 overexpression significantly enhance cell proliferation, migration and clonogenic potential

(P<0.05) of colon adenocarcinoma cells of the early stage. Moreover, GAEC1 also affected the mitochondrial respiration of colon adenocarcinoma through reducing cell coupling efficiency.

A further look into the molecular mechanism of GAEC1 showed that when GAEC1 is overexpressed, increase in the expression of pAkt, FOXO3 and MMP9 were noted. These findings indicate that GAEC1 may interact with these protein molecules and may perhaps play a role in regulating gene pathways such as the Akt pathway.

Later, FAM134B overexpression was used to confirm the tumour suppressive effects of

FAM134B on colon adenocarcinoma. Various in vitro assays were done, several potential transcriptional factors that interact with FAM134B were identified using western blot analysis

3 and correlated with immunohistochemistry staining in cancer tissues of patients with colorectal adenocarcinoma. This study found that FAM134B is involved in the apoptosis and mitochondrial function of colon adenocarcinoma. FAM134B overexpression significantly increased the expression of EB1, p53 and MIF. While 70% of patients with colorectal adenocarcinoma showed that low expression of FAM134B was significantly correlated with increased expression of MIF, 30% of these patients showed concurrent expressions of MIF and

FAM134B expression (P<0.05). Low expression of FAM134B coupled with high expression of

MIF is significantly associated with microsatellite instability (P<0.05). Taken together, chapter proves for the first time that FAM134B overexpression affects mitochondrial functionality and potentially interacts with MIF, EB1 and p53 to affect colon adenocarcinoma progression.

Collectively, from clinical studies to molecular studies, this study has provided critical knowledge towards further understanding the oncogenic properties of GAEC1 and tumour suppressive properties of FAM134B in colorectal adenocarcinoma. Both GAEC1 and

FAM134B portrayed significant correlation in patients with colorectal adenocarcinoma.

Likewise, mutations in both GAEC1 and FAM134B has been determined for the first time.

Overexpression of GAEC1 or FAM134B showed significant impact on colon adenocarcinoma in vitro. Consequently, potential interacting partners were identified with the manipulation of the expression of GAEC1 or FAM134B. These findings act as a basis for further studies on validating these genes as potential biomarkers for the detection of colorectal cancer or as potential therapeutic options.

Statement of Originality

This work has not previously been submitted for a degree or diploma in any university. To the best of my knowledge and belief, the thesis contains no material previously published or written by another person except where due reference is made in the thesis itself.

(Signed)______Katherine Ting-Wei Lee

Acknowledgement

The completion of this dissertation would not be possible without the tremendous support of multiple individuals and organisation. I would like to express my utmost gratitude and sincere thanks to all of them.

To begin with, I would like to thank my dedicated supervisors Professor Alfred Lam,

Dr. Vinod Gopalan and Dr. Robert Smith for their continual support, mentoring, guidance and critical input throughout my postgraduate degree. Without their expertise and timely feedback and suggestions, it would not have been possible for me to complete this dissertation. I would also like to thank Dr. Johnny Tang for his kind donation of the GAEC1 plasmid and constructive input during my candidature.

In addition, I would like to thank Griffith University, Griffith Graduate Research

School, Griffith Health and Menzies Health Institute of Queensland (MHIQ) for awarding me the scholarship and support to carry out this research. Special thanks Dr. Cu Tai Lu from Gold

Coast University Hospital for the recruitment of the clinical samples used in this study, Melissa

Leung for keeping the tissue biomarker database well organised and up to date and for organising monthly group meetings and Christmas celebrations. Not to forget, the NH

Diagnostic Laboratory team that helped the preparation of formalin-fixed paraffin-embedded tissue specimens.

I would also like to portray my sincere thanks to Dr. Jelena Vider who have kindly provided her expert advice and training for the use of flow cytometry, immunofluorescence microscope and Seahorse analyser. Her relentless help in obtaining consumables at crucial stages. My heartfelt thanks also go to Dr. Cameron Flegg for his experienced advices in immunofluorescence staining, confocal microscopy and his ever willingness to provide help and advice when requested. Sincere gratitude to Institute of Health and Biomedical Innovation

6 at the Queensland University of Technology for providing the droplet digital system so I could carry out my experiments.

Special thanks to the great support and encouragement that I have received from the whole team at the Cancer Molecular Pathology Group, School of Medicine, Griffith University throughout my candidature. Their ever willingness to lend a helping hand, share their wealth of knowledge and provide mental support has helped me come this far. These wonderful individuals that I would love to thanks includes Anna Chruścik, Dr. Farhadul Islam, Dr. S.M.

Raijul Wahab, Dr. Suja Pillai, Dr. Md. Hakimul Haque, Dr. Nassim Saremi, Tracie Cheng,

Jeremy Martin, Dr. Haleh Vogha, Dr. Afraa Mahdi Jawad Mamoori, Dr. Faeza Ebrahimi,

Faysal-Bin Hamid, Dr. Armin Ariana, Dr. Hamidreza Maroof and Dr. Atiqur Rahman. Sincere thanks to Christine Pickersgill, Dr. Avinash Kundur, Dr. Brittany McCormack, Dr. Paulina

Janeczek, Elísabet Ósk Guðmundsdóttir, Christine M Villahermosa, Chima Enwerem, Dr.

Loribeth Evertz, Dr. Karishma Sachaphibulkij, Dr. Alexandre Rcom H'cheo Gauthier, Fr. Paul

Eloagu and Kate Delaforce for their companionship, advice and strength to go on.

Last, but not least, I would like to express my deepest gratitude towards my loving parents, younger siblings and other family members who have been nothing but supportive throughout my PhD candidature. Their endless support, love and care has been the pillar that has supported the realisation of my dreams and passion. I can never thank my loved ones enough for everything that they have given to me. Thank you all so much for everything you have done for me. Without you all, I would have never made it today.

Table of Content

Abstract 2 Statement of Originality 5 Acknowledgement 6 Table of Content 8 List of Figures 12 List of Tables 15 List of Abbreviations 16 Peer Reviewed Publications during PhD 22 Publications arising from this thesis 22 Publications tangential to this thesis 22 Peer-reviewed conference abstracts 22 Acknowledgement of Papers Included in this Thesis 24 Chapter 1: General Introduction 27 1.1 Background 27 1.2 Rationale for the research 28 1.3 Aims and objectives 29 1.4 Significance of the research 31 1.5 Structure of the dissertation 33 Chapter 2: Literature Review 36 2.1 Cancer 36 2.1.1: Cancer at a glance 36 2.1.2: Statistical data on cancer occurrences 36 2.2: Colorectal Cancer 37 2.2.1: Genetic aetiology of colorectal cancer 37 2.2.2: Molecular carcinogenesis in colorectal cancer 39 2.2.3: Genetic architecture of colorectal cancer 40 2.3 Identification of novel and pathologically significant molecular targes, GAEC1 and FAM134B 43 2.3.1 Gene amplified in oesophageal cancer 1 (GAEC1) 44 2.3.2 Colorectal cancer and GAEC1 45 2.3.3 Family with sequence similarity 134 member B (FAM134B) 48 2.3.4 Colorectal cancer and FAM134B 50

Chapter 3: Materials and methods 53 3. 1 Recruitment of tissue samples (Aims 1, 2 and 4) 53 3.2 Clinicopathological parameters (Aims 1, 2 and 4) 53 3.3 Clinical Management (Aims 1, 2 and 4) 54 3.4 Fresh frozen tissues processing (Aims 1, 2 and 4) 54 3.5 Formalin-fixed paraffin-embedded tissues (FFPE) processing (Aim 2 and 4) 55 3.6 Hematoxylin and eosin staining (Aims 1, 2 and 4) 55 3.7 Cell culture (Aim 2, 3 and 4) 55 3.8 Genomic DNA extraction (Aim 1 and 2) 57 3.9 Droplet digital PCR (ddPCR) analysis (Aim 1) 58 STATEMENT OF CONTRIBUTION TO CO-AUTHORED PUBLISHED PAPER 59 3.9.1 Detailed description of copy number alteration using ddPCR 60 3.10 Real-time PCR (qPCR) (Aim 2, 3 and 4) 81 3.11 Conventional PCR (Aim 2 and 3) 81 3.12 Purification of PCR products (Aim 2) 82 3.13 Sanger sequencing analysis (Aim 2) 82 STATEMENT OF CONTRIBUTION TO CO-AUTHORED PUBLISHED PAPER 84 3.13.1 Detailed description of targeted single gene mutation analysis with Sanger sequencing 85 3.14 Immunofluorescence (Aim 2) 106 3.15 Protein extraction and quantification (Aim 2, 3 and 4) 106 3.15.1 Protein extraction from cells 106 3.15.2 Protein extraction from tissue sample 107 3.16 Western blot analysis (Aim 2, 3 and 4) 107 3.17 Cloning of plasmid into bacteria (Aim 2, 3 and 4) 108 3.18 Plasmid transfection (Aim 2, 3 and 4) 109 3.19 Cell proliferation assay (MTT) (Aim 3 and 4) 110 3.20 Wound healing assay (Aim 3 and 4) 111 3.21 Clonogenic assay (Aim 3 and 4) 111 3.22 Apoptotic assay (Aim 3 and 4) 111 3.23 Cell cycle analysis (Aim 3 and 4) 112 3.24 Extracellular flux assay (Aim 3 and 4) 112 3.25 In silico analysis (Aim 3 and 4) 113 3.26 Statistical analysis (Aim 1, 2, 3 and 4) 113

Chapter 4: Developing an accurate method for the detection of copy number aberration and ploidy determination of GAEC1 in patients with colorectal cancer. 114 4.1 Preface 114 STATEMENT OF CONTRIBUTION TO CO-AUTHORED PUBLISHED PAPER 115 4.2 Detailed description of Droplet digital vs real-time PCRs for accurate quantification of gene copy number aberrations in colorectal cancers 116 4.2.1 Introduction 118 4.2.2 Results 120 4.2.3 Discussion 126 4.2.4 Methods 129 4.2.5 Acknowledgements 133 4.2.6 Authors contribution 133 4.2.7 Competing interest 134 4.2.8 References 134 Chapter 5: Identification of GAEC1 mutations 136 5.1 Preface 136 STATEMENT OF CONTRIBUTION TO CO-AUTHORED PUBLISHED PAPER 137 5.2 Detailed description of GAEC1 mutations and copy number aberration association with biological aggressiveness in colorectal cancer 138 5.2.1 Abstract 138 5.2.2 Introduction 139 5.2.3 Material and methods 140 5.2.4 Results 150 5.2.5 Discussion 164 5.2.6 Acknowledgement 167 5.2.7 Funding 167 5.2.8 Authors Contribution 167 5.2.9 Conflict of Interest 167 5.3 Inclusion of GAEC1 protein sequence in GenBank, NCBI 168 5.3.1 GenBank flat file for GAEC1 169 Chapter 6: Functional role and interacting partners of GAEC1 in colon cancer 170 6.1 Preface 170 STATEMENT OF CONTRIBUTION TO CO-AUTHORED PUBLISHED PAPER 171 6.2 Detailed description of GAEC1 drives colon cancer progression 172 6.2.1 Abstract 173 6.2.2 Keywords 173

6.2.3 Introduction 174 6.2.4 Materials and Methods 175 6.2.5 Results 181 6.2.6 Discussion 189 6.2.7 Conflict of Interest 192 6.2.8 Acknowledgement 192 6.2.9 References 192 Chapter 7: Functional role and interacting partner of FAM134B 195 7.1 Preface 195 STATEMENT OF CONTRIBUTION TO CO-AUTHORED PUBLISHED PAPER 197 7.2 Detailed description of Family with sequence similarity 143, member B (FAM134B) overexpression in colon cancers and its tumor suppressor properties via regulation of macrophage migration inhibitory factor (MIF) and other partners 198 7.2.1 Abstract 199 7.2.2 Keywords 199 7.2.3 Highlights 200 7.2.4 Introduction 200 7.2.5 Materials and Methods 202 7.2.6 Results 208 7.2.7 Discussion 218 7.2.8 Conclusion 221 7.2.9 Acknowledgement 221 7.2.10 References 222 7.2.11 Funding 224 7.2.12 Conflict of Interest 225 Chapter 8: Summaries and conclusions 226 Chapter 9: References 230 Appendices 239

List of Figures

Figure 2.1: Anatomy of the lower digestive system [20] ...... 37

Figure 2.2: Genetic alteration and carcinogenesis steps in colorectal cancer [6]...... 39

Figure 2.3: Number of somatic mutations in different human cancers as detected by genome- wide sequencing studies [6] ...... 42

Figure 2.4 Nucleotide with a predicted amino acid sequence of GAEC1...... 45

Figure 3.1: Mycoplasma contamination testing ...... 57

Figure 3.2: Outlook of Primer-BLAST Website for the design of primers ...... 69

Figure 3.3: Internal hybridisation parameters section for design of probes ...... 70

Figure 3.4: Multiple target detection (genes of interest) using EvaGreen-based ddPCR...... 72

Figure 3.5: Reaction preparation containing only one primer set at each well...... 73

Figure 3.6: Example gel image after a 2% agarose gel ran at 70V for 60 minutes...... 96

Figure 3.7: Example of a good quality chromatogram...... 98

Figure 3.8: Overview of the blastn suite webpage...... 98

Figure 3.9: Example of a single nucleotide mutation found through Sanger sequencing...... 99

Figure 4.1: Quantification of GAEC1 plasmid DNA serial dilutions by qPCR...... 120

Figure 4.2: 1D and 2D droplet plots of droplet clusters obtained with the QuantaSoft software version 1.7.4...... 121

Figure 4.3: Correlation between qPCR and ddPCR GAEC1 copy number detection in patients with colorectal cancer...... 123

Figure 4.4: Bland-Altman plot to compare the measurements of two different techniques. . 123

Figure 4.5: Inter-assay CV% between qPCR and ddPCR assays...... 124

Figure 4.6: (A) ddPCR workflow diagram and (B) qPCR workflow diagram for CNA detection.

...... 125

Figure 5.1: Schematic representation of the methodological flow used...... 142

Figure 5.2: Representative GAEC1 gene amplification and copy number aberration (CNA) in patients with colorectal cancer...... 146

Figure 5.3: Survival rate correlated with (a) GAEC1 copy number aberration in colorectal cancer tissues and (b) adjacent non-neoplastic tissue...... 151

Figure 5.4: Heterozygosity at (a) rs181281674 and (b) rs181800740 as observed in GAEC1 found in colorectal cancer tissues...... 153

Figure 5.5: Loss of heterozygosity (LOH) in GAEC1 found in colorectal cancer tissues. .... 154

Figure 5.6:Novel mutations in GAEC1 found in colorectal cancer tissues...... 155

Figure 5.7: CNA of 79 matched samples of patients with colorectal cancer...... 158

Figure 5.8: GAEC1 expression in patients with colorectal cancer...... 163

Figure 5.9: Screenshot of the GAEC1 mRNA, complete cds and microsatellite sequences on

NCBI Nucleotide website...... 169

Figure 5.10: Screenshot of GAEC1 mRNA, complte cds with amino acid sequence ...... 169

Figure 6.1:Overexpression of GAEC1 in colon cancer (SW480 and SW48) and non-neoplastic colonic epithelial cell (FHC)...... 181

Figure 6.2: Effect of GAEC1 overexpression on the growth rate of SW48, SW480 and FHC cells...... 183

Figure 6.3: Effect of GAEC1 overexpression on the wound healing potential of both colon cancer (A) SW48, (B) SW480 and non-neoplastic colon epithelial cells (C) FHC...... 184

Figure 6.4: Impact of GAEC1 overexpression on cell cycle kinetics of colon cancer cells (SW48 and SW480) and non-neoplastic colon epithelial cells (FHC) in (A) histogram and (B) bar graph...... 185

Figure 6.5: Outcome of GAEC1 ectopic expression on apoptosis of colon cancer cells in (A) dot plots and (B) graph...... 186

Figure 6.7: Overexpression of GAEC1 induced the expression of (A) pAkt, (B) FOXO-3 and

Figure 7.1: Overexpression of FAM134B in colon cancer cells (SW48 and SW480) and non- neoplastic colon epithelial cells (FHC)...... 208

Figure 7.2：Outcome of FAM134B ectopic expression on apoptosis of colon cancer cells in

(A) dot plots and (B) graph...... 211

Figure 7.3: Impact of FAM134B overexpression on cell cycle kinetics of colon cancer cells

(SW48 and SW480) and non-neoplastic colon epithelial cells (FHC) in (A) histogram and (B) bar graph...... 212

Figure 7.4: Influence of FAM134B overexpression on mitochondrial respiration of colon cancer.

...... 214

Figure 7.5: Overexpression of FAM134B successfully changed the protein expression of colon cancer cells...... 215

Figure 7.6: Association between MIF and FAM134B in patients with colorectal cancer and its representative expression of MIF in colorectal adenocarcinoma at x25 magnification...... 216

List of Tables

Table 2.1: Novel interacting partners of FAM134B protein [88] ...... 52

Table 3.1: Amplification or deletion of genes involved in oesophageal adenocarcinoma ...... 61

Table 3.2: Gene mutations found in oesophageal adenocarcinoma discovered through sequencing...... 87

Table 3.3: Clusters of genes which were implicated with oesophageal carcinoma...... 91

Table 3.4: Current main purposes of Sanger sequencing ...... 91

Table 4.1: Cost comparison between qPCR assays and ddPCR assays for CNA detection using

DNA binding dye chemistry...... 125

Table 5.1: Mutations detected in the exon region of GAEC1 in Australian patients with colorectal cancer ...... 152

Table 5.2: Frequent polymorphisms detected in GAEC1 in colorectal cancer...... 152

Table 5.3: Statistical correlation between GAEC1 mutation and clinicopathological features of

Australian patients with colorectal cancer ...... 156

Table 5.4: Statistical correlation between GAEC1 variant Chr7: g.101939277 and clinicopathological features of Australian patients with colorectal cancer ...... 157

Table 5.6: Changes in GAEC1 mRNA, protein or copy number aberration with or without

GAEC1 mutation...... 166

Table 6.1: Primer design for quantitative real time PCR (qPCR) analysis ...... 177

Table 7.1: Statistical correlation between low FAM134B mRNA expression in the presence of

MIF and the clinicopathological features of Australian patients with colorectal cancer ...... 217

List of Abbreviations

AGRF Australian Genome Research Facility

AJCC American Joint Committee on Cancer

Akt protein kinase B

APC adenomatous polyposis coli

ARE Adenylate-uridylate-rich element

ATCC American type culture collection

ATP adenine triphosphate

BDT Big Dye Terminator

BLAST Basic Local Alignment Search Tool

BRAF B-Raf or v-Raf murine sarcoma viral oncogene homolog B

BRCA-1 Breast cancer 1, early onset

BSA bovine serum albumin

CAMSAP2 Calmodulin Regulated Spectrin Associated Protein Family Member 2

CAPN10 calpain 10

CAP1 adenylyl cyclase-associated protein 1

CCND2 δ-catenin

CDK cyclin-dependent kinase

CDKN2A cyclin-dependent kinase inhibitor 2A

CNA copy number aberration

CO2 carbon dioxide

CRC colorectal cancer

CTNNB1 β-catenin

CV coefficient of variation

CyPB cyclophilin B

DAPI 4,6-diamidino-2-phenylindole ddPCR droplet digital polymerase chain reaction

DDBJ DNA Data Bank of Japan

DMEM Dulbecco's Modified Eagle Medium

DMSO dimethyl sulfoxide

DNA deoxyribonucleic acid dPCR digital polymerase chain reaction ddPCR droplet digital PCR

EAC oesophageal adenocarcinoma

EB1 microtubule plus-end tracking of end-binding protein 1

ECM extracellular matrix

EDTA ethylenediaminetetraacetic acid

EGFR epidermal growth factor receptor, also known as HER2

ENA European Nucleotide Archive

ER endoplasmic reticulum

ESCC esophageal squamous cell carcinoma

EYA4 eye absent 4

FAP familial adenomatous polyposis

FAM134B family with sequence similarity 143, member B

FBS foetal bovine serum

FFPE formalin fixed paraffin-embedded tissues

FISH fluorescence in situ hybridisation

FITC fluorescein isothiocyanate

FOXO3 forkhead box O3

GABARAP gamma-aminobutyric acid receptor-associated protein

GAEC1 gene amplified in oesophageal cancer 1

GAPDH glyceraldehyde 3-phosphate dehydrogenase gDNA genomic deoxyribonucleic acid

GMO genetic modified organism

GWAS genome wide association studies

HBB haemoglobin beta

HDR higher degree by research

HEPES 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid

HER2 human epidermal growth factor 2, also known as EGFR

HLA human leukocyte antigen

HPV human papillomavirus

HRAS transforming protein p21

HuR Hu-antigen 1

H&E haemotoxylin & eosin

ICGC International Cancer Genome Consortium

KDELR2 KDEL Endoplasmic Reticulum Protein Retention Receptor 2

KRAS Kirsten rat sarcoma viral oncogene homolog

LB buffer lithium borate buffer

LB media Luria-Bertani media

LC3 microtubules-associated protein 1A/1B-light chain 3

LIR GABARAP-interacting region

LC-MS/MS Liquid chromatography coupled with tandem mass spectrometry

LoA limit of agreement

LOH loss of heterozygosity

MDM2 murine double minute-2

MIQE Minimum Information for Publication of Quantitative Real-Time PCR

Experiments

MMPs matrix metalloproteinases mRNA messenger ribonucleic acid

MSI microsatellite instability

MTT 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide mtDNA mitochondria DNA

NCBI National Centre for Biotechnology Information

NGS next generation sequencing

NO2 nitrogen

NMD non-sense mediated mRNA decay

NTC non-template control

OXPHOS mitochondrial oxidative phosphorylation

PBS phosphate buffered saline

PBST phosphate buffered saline with Tween 20

PCR polymerase chain reaction

PE plating efficiency

PHLPP PH domain and leucine rich repeat protein phosphatase

PI propidium iodide

PIK3CA Phosphatidylinositol-4,5-Bisphosphate 3-Kinase Catalytic Subunit Alpha

PROVEAN Protein Variation Effect Analyser

PVDF polyvinylidene flouride qPCR real-time PCR

RACE 5’ and 3’-rapid amplification of cDNA ends

RETREG1 reticulophagy regulator 1

RHD reticulon-homology domain

RNA ribonucleic acid

ROS reactive oxygen species

RPMI Roswell Park Memorial Institute

SELENBP1 Selenium Binding Protein 1

SDS sodium dodecyl sulphate

SF survival fraction

SH2B2 SH2 B adaptor protein 2

SIPK salicylate-induced protein kinase

SMAD4 SMAD family member 4

SMYLE myomegalin-like EB1 binding protein

SNP single nucleotide polymorphisms

SOB super optimal broth

SOC super optimal broth with catabolite repression media

SRPK1 Serine/threonine-protein kinase

TAE Tris-acetate-EDTA

TNM tumour, lymph node and metastases

TP53 tumour protein p53

UCSC University of California Santa Cruz

UPR unfolded protein response

UTR untranslated region

VSIG10L V-Set and Immunoglobulin Domain Containing 10 Like

WHO World Health Organization

WIPK wound induced protein kinase

ZNF217 Zinc Finger Protein 217

5p chromosome 5, p arm

7q chromosome 7, q arm

7q22.1 chromosome 7, q arm, band 22, sub-band 1

Peer Reviewed Publications during PhD

Publications arising from this thesis

1. Lee KT, Vider J, Tang JC, Gopalan V, Lam AK. GAEC1 drives colon cancer progression. Molecular carcinogenesis. 2019. 2. Lee KTW, Gopalan V, Lam AK. Somatic DNA Copy-Number Alterations Detection for oesophageal Adenocarcinoma Using Digital Polymerase Chain Reaction. In: Lam AK, editor. oesophageal Adenocarcinoma: Methods and Protocols. New York, NY: Springer New York; 2018. p. 195-212. 3. Lee KTW, Smith RA, Gopalan V, Lam AK. Targeted Single Gene Mutation in oesophageal Adenocarcinoma. In: Lam AK, editor. oesophageal Adenocarcinoma: Methods and Protocols. New York, NY: Springer New York; 2018. p. 213-29. 4. Lee KT, Gopalan V, Islam F, Wahab R, Mamoori A, Lu CT, et al. GAEC1 mutations and copy number aberration is associated with biological aggressiveness of colorectal cancer. Eur J Cell Biol. 2018.

Publications tangential to this thesis

1. Lee KT, Gopalan V, Lam AK. Roles of long-non-coding RNAs in cancer therapy through the PI3K/Akt signalling pathway. Histology and histopathology. 2019:18081. 2. Wahab R, Gopalan V, Islam F, Mamoori A, Lee KT, Lu CT, et al. Expression of GAEC1 mRNA and protein and its association with clinical and pathological parameters of patients with colorectal adenocarcinoma. Exp Mol Pathol. 2018;104(1):71-5. 3. Mamoori A, Wahab R, Islam F, Lee K, Vider J, Lu CT, et al. Clinical and biological significance of miR-193a-3p targeted KRAS in colorectal cancer pathogenesis. Hum Pathol. 2018;71:145-56. 4. Islam F, Gopalan V, Wahab R, Lee KT, Haque MH, Mamoori A, et al. Novel FAM134B mutations and their clinicopathological significance in colorectal cancer. Hum Genet. 2017;136(3):321-37. 5. Lee KT, Tan JK, Lam AK, Gan SY. MicroRNAs serving as potential biomarkers and therapeutic targets in nasopharyngeal carcinoma: A critical review. Critical reviews in oncology/hematology. 2016;103:1-9.

Peer-reviewed conference abstracts

1. Lee KT, Vinod G, Lam AK. Identifying novel mutations sites in GAEC1 oncogene in patients with colorectal carcinoma. International Student Research Forum (ISRF) 2018 2. Islam F, Vinod G, Wahab R, Lee KT, Mamoori A, Lu C, Smith RA, Lam AK. Mutational status, expression and functional behaviors of FAM134B in colorectal cancer. AACR Annual Meeting 2017. Cancer Research 77(13):3420. 3. Mamoori A, Wahab R, Islam F, Lee KT, Smith RA, Vinod G, Lam AK. Downregulation of miR-193a and its correlation with clinical and pathological behavior of colorectal cancer. AACR Annual Meeting 2017. Cancer Research 77(13 Supplement):465-465.

4. Lee KT, Smith RA, Islam F, Wahab R, Vinod G, Lam AK. Identification of novel GAEC1 (gene amplified in oesophageal cancer 1) mutations in colorectal cancer. 3rd EACR Conference on Cancer Genomics 2017. 5. Lee KT, Chruscik A, Vinod G, Lam AK. Comparative analysis of GAEC1 copy number variation using real-time PCR and droplet digital PCR. 3rd EACR Conference on Cancer Genomics 2017. 6. Islam F, Vinod G, Lee KT, Vider J, Wahab R, Ebrahimi F, Lu C, Kasem K, Lam AK. Inhibition of miR-186-5p suppress colon cancer growth: an anti-miRNA mediated upregulation of tumour suppressor FAM134B. 3rd EACR Conference on Cancer Genomics 2017. 7. Lee KT, Vinod G, Lam AK. Multiplexing DNA Binding Chemistry Based Assay with the Use of Digital PCR for GAEC1 Copy Number Variation Identification in Colorectal Cancer. ASMR National Scientific Conference: Next Generation Healthcare: Merging Biology and Technology 2016 8. Lee TW, Tan EL, Ng CC, Gan SY. Effects of transforming growth factor β1 and latent membrane protein 1 on miRNA expression of nasopharyngeal carcinoma. 37th Annual Conference of the Malaysian Society for Biochemistry & Molecular Biology. 9. Lee TW, Tan EL, Ng CC, Gan SY. Expression Profiling of MicroRNA in Nasopharyngeal Carcinoma in Response to Interleukin-6 and Latent Membrane Protein One. 5th International Symposium on Nasopharyngeal Carcinoma 2011.

Acknowledgement of Papers Included in this Thesis

Included in this thesis are papers in Chapters 3, 4, 5, 6 and 7 which are co-authored with other researchers. My contribution to each co-authored paper is outlined at the front of the relevant chapter. The bibliographic details/ status for these papers including all authors, are:

Chapter 3:

Book Chapter 1 Title: Somatic DNA Copy-Number Alterations Detection for oesophageal Adenocarcinoma Using Digital Polymerase Chain Reaction

Authors: Katherine Ting-Wei Lee, Vinod Gopalan, Alfred King-yin Lam

Name of the book: oesophageal Adenocarcinoma: Methods and Protocols, Methods in Molecular Biology, vol. 1756

Status: Published

Book Chapter 2

Title: Targeted single gene mutation with Sanger sequencing in oesophageal Adenocarcinoma

Authors: Katherine Ting-Wei Lee, Vinod Gopalan, Alfred King-yin Lam

Name of the book: oesophageal Adenocarcinoma: Methods and Protocols, Methods in Molecular Biology, vol. 1756

Status: Published

Chapter 4:

Title: Droplet digital vs real-time PCRs for accurate quantification of gene copy number aberrations in colorectal cancers

Authors: Katherine Ting-Wei Lee, Anna Chruścik, Vinod Gopalan, Alfred King-yin Lam

Name of the journal: Scientific Report

Status: Submitted

Chapter 5:

Title: GAEC1 mutations and copy number aberration is associated with biological aggressiveness of colorectal cancer

Authors: Katherine Ting-Wei Lee, Vinod Gopalan, Farhadul Islam, Riajul Wahab, Afraa Mamoori, Cu-tai Lu, Robert Anthony Smith, Alfred King-yin Lam

Name of the journal: European Journal of Cell Biology

Status: Published

Chapter 6:

Title: Overexpression of GAEC1 drives colon cancer progression

Authors: Katherine Ting-Wei Lee, Jelena Vider, Johnny Cheuk-on Tang, Vinod Gopalan, Alfred King-yin Lam

Name of the journal: Cancer Letters

Status: Published

Chapter 7:

Title: FAM134B overexpression in colon cancers and its tumour suppressor properties via regulation of MIF and other partners

Authors: Katherine Ting-Wei Lee, Farhadul Islam, Anna Chruścik, Jelena Vider, Jeremy Martin1, Vinod Gopalan, Alfred King-yin Lam

Name of the journal: Cancer Letters

Status: Submitted

Appropriate acknowledgements of those who contributed to the research but did not qualify as authors are included in each paper.

(Signed) ______(Date)___ 22/11/2018__

Katherine Ting- Wei Lee

(Countersigned) ______(Date)____ 22/11/2018_

Principal Supervisor 1: Prof. Alfred Lam

(Countersigned) ______(Date)___ 22/11/2018___

Principal Supervisor 2: Dr. Vinod Gopalan

Chapter 1: General Introduction

1.1 Background

Colorectal cancer (CRC) or cancer of the large bowel remains one of the most common

cancer types globally [1]. The development of colorectal cancer is a slow one, taking up to

decades to fully mature. Therefore, unlike most other neoplasms, the various developmental

stages of colorectal cancer can be acquired for analysis. This has made colorectal cancer a

sought after system to study genetic alterations for the development of cancer [2].

The development of cancer is multi-factorial and often require the accumulation of

multiple mutations to occur before progressing into full-blown cancer [2]. Majority of the

colorectal cancers begin with benign polyps which progresses to malignant lesions through

acquiring a set of mutations [2, 3]. These initiated cells will then go through a stage known as

promotion, where cells obtain further mutations such as Kirsten rat sarcoma viral oncogene

homolog (KRAS) and adenomatous polyposis coli (APC) mutations [4-6]. This is followed by

the progression stage where initiated cells are irreversibly converted into malignant cells

through various gene mutations such as B-Raf or v-Raf murine sarcoma viral oncogene

homolog B (BRAF), Phosphatidylinositol-4,5-Bisphosphate 3-Kinase Catalytic Subunit Alpha

(PIK3CA), SMAD family member 4 (SMAD4) and tumour protein p53 (TP53) [5-9].

Interestingly, clinical observations of these mutation developments can be used as a form of

measurement indication for the different stages of colorectal cancer. Therefore, understanding

the genetic aetiology of colorectal cancer can further assist the prediction of an individual’s

risk in developing colorectal cancer and therefore, help develop preventive strategies and

disease management [10].

Approximately 20 years ago, our group identified two novel genes of interest through

comparative genomic hybridisation analysis to compare the difference between non-neoplastic

oesophageal mucosae and oesophageal squamous cell carcinoma [11]. These genes were

individually located on chromosome 7q and 5p respectively and were found to play a

substantial role in the pathogenesis of oesophageal squamous cell carcinoma [11-13]. These

genes were subsequently termed gene amplified in oesophageal cancer 1 (GAEC1) and family

with sequence similarity 143, member B (FAM134B).

Later, both GAEC1 and FAM134B were identified in colorectal cancer and was found

to be associated with various pathological features including but not limited to tumour size,

perforation and survival outcome [13, 14]. Through some functional studies, we have identified

that while GAEC1 functions as an oncogene, FAM134B acts as a tumour suppressor gene.

Nonetheless, knowledge of the molecular mechanisms of both GAEC1 and FAM134B is still

lacking. Understanding the effects of overexpression of GAEC1 and FAM134B can help

confirm the oncogenic or tumour suppressive function of each gene respectively and rule out

possible duality function in these genes. The opposite spectrum in the functionality of both

genes in colorectal cancer makes them a valuable pair of genes to study as they could help

unveil the imperative role that they play in colorectal cancer biology. This will not only fill the

gap of knowledge regarding the functionality of these genes, but also potentially help identify

and develop better therapeutic strategies, effective biomarker and disease management of

colorectal cancer.

1.2 Rationale for the research

Colorectal cancer is the second most common cancer in Australia with an estimate that

1 in 16 females and 1 in 11 males will be diagnosed with colorectal cancer before the age of

85 [1]. Colorectal cancer is the sloth of cancers and takes approximately 17 years for a large

benign tumour to evolve into advanced cancer. However, upon becoming malignant, colorectal

cancer only requires 2 years to achieve metastasis [10]. The survival rate for patients with

earlier stages (stage I-III) of colorectal cancer is relatively high (between 90%-100%).

Unfortunately, upon metastasis, the survival rate drops below 20% [15]. Therefore, early

detection and disease intervention of colorectal cancer is crucial. To this end, researches in the

field of developing biomarkers and strategies for early detection and/or diagnosis of colorectal

cancer is of utmost importance.

1.3 Aims and objectives

The main objective of this research work is directed toward developing and targeting

new biomarkers for early detection of patients with colorectal cancer. The main focus of this

study will be to better understand the molecular function of GAEC1 and FAM134B in the

pathogenesis of colorectal cancer. This research will develop a new method for accurate

detection of copy numbers and ploidy determinations, investigate the mutation status of

GAEC1 and its clinicopathological association of patients with colorectal cancer. The research

will also identify the association of the expression of GAEC1 and FAM134B with the

pathological features of Australian patients with colorectal cancer. To understand the molecular

mechanisms of both GAEC1 and FAM134B, the mutation profile of GAEC1 and FAM134B

will be evaluated. In addition, possible cancer pathway associations of GAEC1 and FAM134B

in colon cancer cell lines will also be evaluated. To do so, the specific aims for this research

entails:

Aim 1: Accurate quantification via the detection of copy number aberration and ploidy

determination of GAEC1 in patients with colorectal cancer.

Hypothesis: Droplet digital PCR (ddPCR) produces more reproducible results

compared to real-time PCR (qPCR) for the detection of GAEC1 copies and ploidy

determination.

Key objective:

Develop a droplet digital PCR duplex DNA binding dye chemistry-based method for the detection of GAEC1 copies and ploidy determination of GAEC1.

Aim 2: Detection of GAEC1 mutations, pathogenic implications and its association with copy number aberration and clinical characteristics of patients with colorectal cancer

Hypothesis: GAEC1 copies and expression changes in colorectal cancer tissues may be caused by mutations

Key objectives:

1. Identification of mutation sites of GAEC1 in colorectal cancer

2. Investigate the association of GAEC1 mutation with GAEC1 copy number

aberration

3. Investigate the association of GAEC1 mutation and GAEC1 expression.

Aim 3: Investigate the oncogenic and cell transformation properties of GAEC1 in vitro in colon cancers.

Hypothesis: GAEC1 is involved in cancer cellular processes through interacting with other gene product within the cancer molecular pathway.

Key objectives:

1. Overexpression of GAEC1 in colon cancer and normal colon cells.

2. Investigate the impact of GAEC1 gene on cellular behaviours such as cell

proliferation, migration, colony formation, cell cycle and apoptosis in colon cancer

and normal colon cells.

3. Identification of potential interacting partners of GAEC1 gene in colon cancer to

identify the potential cancer molecular pathway that GAEC1 is involved.

Aim 4: Similar to GAEC1, investigate the copy number change, mutation status and the

cell restoration roles of FAM134B gene in colon cancers.

Hypothesis: FAM134B is involved in cancer cellular processes through interacting with

other gene product within the cancer molecular pathway.

Key objectives:

1. Overexpression of FAM134B in colon cancer and normal colon cells.

2. Investigate the impact of FAM134B gene on cellular behaviours such as cell

proliferation, migration, colony formation, cell cycle and apoptosis in colon cancer

and normal colon cells.

3. Identification of potential interacting partners of FAM134B gene in colon cancer to

identify the potential cancer molecular pathway that FAM134B is involved.

1.4 Significance of the research

The survival rate of patients with early stages of colorectal cancer is rather favourable,

having approximately 90% five-year survival rate. Unfortunately, this value dwindles

significantly once metastasis occurs, reducing the five-year survival rate to only about 20%

[15]. Therefore, early detection and treatment of colorectal cancer are extremely crucial to aid

in combating this disease. The proposed research project focuses on exploring the molecular

role, functional characteristics and clinical significance of the oncogene GAEC1 and tumour

suppressor gene FAM134B in colorectal cancer pathogenesis. Both these genes are postulated

to play a role in the early pathogenesis of colorectal cancer [15, 16].

The current focus of patient management and point-of-care is to minimise unnecessary

invasiveness to patients. This is made possible with the utilisation of non-invasive biological

specimens such as blood, urine, sputum, stool and so on. However, these samples often consist

of only minute biological materials (DNA, RNA and/or protein) and are often present with

31 inhibitors and contaminants. Therefore, it is essential to develop a method that is highly sensitive and can accurately detect biomarkers within these specimens. To do so, this study first focuses on developing a new cost-effective method that can accurately identify the copy number changes of GAEC1 in patients with colorectal cancer. This method will then be compared to the current gold standard method to evaluate the accuracy and sensitivity of the method. Comparison of time efficiency as well as cost effective will also be analysed. This is because, a short sample to results turnaround time as well as providing a lower cost alternative to current detection methods are crucial for the current healthcare settings. Furthermore, the principal of this method is not only limited to GAEC1 but can also be applied on FAM134B and any other genes.

Next is on identifying the mutation profile of these genes. These mutations that are discovered can then be compared with the changes of GAEC1 and FAM134B expression and/or copy number at different pathological stages of colorectal cancer to identify its’ correlation.

Furthermore, the association between mutations of either of these genes and clinicopathological features could be determined. This will help in better understanding the roles of both GAEC1 and FAM134B in the genetic aetiology of colorectal cancer to help accurate prediction of individual’s risk and enhance personalised disease management.

Moving forth, this research will exogenously express GAEC1 or FAM134B (separately) in both colon cancer cells and non-neoplastic colonic epithelial cells to better understand the molecular mechanisms of these genes. This is done by comparing the changes observed between the original cells and cells ectopically expressed with either GAEC1 or FAM134B.

Various cell-based assays would be performed to compare the effects of these genes on the physiological state of the cells. Furthermore, Western blot analysis was performed to identify associated proteins were affected in accordance with the changes observed through the functional assays performed.

1.5 Structure of the dissertation

This dissertation includes nine (9) chapters. Chapter 3 includes two method book

chapters that have been published. Chapter 5 is a published peer-reviewed journal article.

Chapter 4, 6 and 7 include research articles that have been submitted to peer-reviewed journals.

Chapter 8 consists of the discussion and future perspective of this research. The overview

content of each chapter outlines are as follow:

Chapter 1: This chapter includes a general introduction explaining the principles of the current

research, covering the background of the study, rational of the research, aims and

objective and, the significance of the research.

Chapter 2: This chapter compiles the available literature leading up to the importance of

working on the current dissertation. The literature covers the current status of

colorectal cancer in the Australian population and the essential reasons for

developing new biomarkers and treatment strategies especially for early stages of

colorectal cancer. A brief overview and history of both GAEC1 and FAM134B are

also highlighted in this chapter, including currently available findings on their

functionality and pathological correlation in colorectal cancer.

Chapter 3: This chapter assembles the materials and methods used throughout this research.

This includes the procedures of tissue sample recruitment, which have had ethical

approval, genetically modified organism project approval, to the various

experimental protocols used.

Chapter 4: The current chapter developed a novel method for copy number aberration

detection of GAEC1 in patients with colorectal cancer. A duplex reaction utilising

DNA binding dye chemistry was utilised in this particular methodology allowing

for multiplexing capabilities that were not possible in the past before the invention

of droplet digital PCR (ddPCR). The assay developed was then compared with the

gold standard qPCR assay to identify the efficiency, reproducibility and the

effectiveness of the system. In addition, the cost effectiveness and overall time

required were compared to identify the best method for future copy number

aberration studies.

Chapter 5: In this chapter, the identification of GAEC1 mutation profile in colorectal cancer

and their clinical significance are presented. Both cancer and matched non-

neoplastic tissues were sequenced using Sanger sequencing to identify novel

mutations. The mutation profile obtained was the crossed compared with the

GAEC1 expression and copy number aberration of the individual patient to identify

any correlation. The clinical association of GAEC1 mutations identified in patients

with colorectal cancer was also analysed.

Chapter 6: In vitro functional role of GAEC1 in colon cancer cells following overexpression

of the gene are described in this chapter. Functional assays such as cell

proliferation, clonogenic potential, wound healing, cell cycle analysis, apoptosis

and extracellular flux assays were performed to understand the effects of GAEC1

in colon cancer pathogenesis. In addition, western blot assays and

immunofluorescence assays were also performed to identify possible proteins that

are interacting with GAEC1 to cause these physiological changes.

Chapter 7: In vitro functional role of FAM134B in colon cancer cells following overexpression

of the gene are described in this chapter. Functional assays such as cell

proliferation, clonogenic potential, wound healing, cell cycle analysis, apoptosis

and extracellular flux assays were performed to understand the effects of FAM134B

in colon cancer pathogenesis. In addition, western blot assays and

immunofluorescence assays were also performed to identify possible proteins that

are interacting with FAM134B to cause these physiological changes.

Chapter 8: Summary, conclusions and future perspective of the entire research are presented

in this chapter.

Chapter 9: All referencing (excluding those in a publication or submission) are listed here.

Chapter 2: Literature Review

2.1 Cancer

2.1.1: Cancer at a glance

Cancer is a term used to describe the uncontrolled proliferation of cells that can metastasise and invade other parts of the body. In other words, they are cells that have gone rogue and fail to respond to signals that regulate cellular growth and death. Cancer is a multifactorial disease. Both internal (mutations, hormones and immune conditions) and external factors (tobacco, radiation and infectious organisms) can play a role in initiating or promoting carcinogenesis [17].

2.1.2: Statistical data on cancer occurrences

According to GLOBOCAN 2012, 14.1 million new cancer cases were diagnosed in 2012 with 8.2 million cancer deaths and a staggering 32.6 million people living with cancer around the globe. Contrary to popular belief, the mortality rates were 15% higher in men and 8% in women in more developed regions than in less developed regions [18].

In Australia, cancer is one of the leading causes of death with mortalities in 2014 accounting for an approximate of three of every ten deaths [1]. Cancer costs account for more than $4.5 billion in direct health system costs (6.9%). This represents a significant concern for

Australia. To address this issue, a total of $378 million (22 % of all health research expenditure) was spent on cancer research in just 2000-2001 [19]. The top five cancer occurrences in

Australia are lung cancer, followed by colorectal cancer, prostate cancer, breast cancer and pancreatic cancer. In total, these five cancers account for nearly half (47%) of all cancers diagnosed in Australia [1].

2.2: Colorectal Cancer

The human digestive system encompasses the oesophagus, stomach, small intestine and large intestine (Figure 2.1). The large intestine is approximately six feet long and includes the colon (large bowel), rectum and anal canal. The large intestine helps remove waste and toxins from the body [20]. Colorectal cancer affects the large intestine and/or the rectum.

In Australia, bowel cancer (14,962 cases) ranks the second most common cancer with a mortality of 4,071 cases [1]. It is estimated that before the age of 75, 1 in 30 females and 1 in

21 males will be diagnosed with colorectal cancer. Before the age of 85, 1 in 16 females and 1 in 11 males are at risk of being diagnosed with colorectal cancer [1].

Figure 2.1: Anatomy of the lower digestive system [20]

2.2.1: Genetic aetiology of colorectal cancer

The aetiological factors underlying colorectal cancer are complex and multifactorial.

These include dietary, lifestyle, genetic hereditary and somatic mutations. While some exposures significantly increase the chances of colorectal cancer, some exposures had equivocal outcomes [5, 21]. In terms of lifestyle factors, people with high occupational or

37 physical activity tend to have a lower risk of developing colorectal cancer [5, 21-23]. A

Western dietary pattern (a diet rich in red meat and/or processed meat or fats) on the other hand confers an increased risk of developing colorectal cancer [5, 21, 23].

Significant findings have been made through identifying specific gene mutations in both inherited colorectal cancer and sporadic colorectal cancer. Of the two, sporadic colorectal cancer makes up the bulk of colorectal cancer cases accounting for approximately 75% of total colorectal cancer [24]. While hereditary colorectal cancer cases comprise of only a minor portion of all colorectal cases, molecular basis discoveries made by investigation of inherited colorectal cancer have also unearthed specific factors and mechanisms that contribute to sporadic colorectal cancer development [5]. These molecular defects generally encompass either modification that leads into loss of function of tumour suppressor genes or high function of oncogenes [5, 25]. Types of molecular alterations of cellular genes into oncogenic variant alleles include single-nucleotide polymorphism, deletion, insertion and substitution [5, 6].

However, a single mutation in the colorectal cell does not mean that the person will ultimately develop colorectal cancer. Multiple mutations are required for a pre-neoplastic cell to progress into a full-blown cancer cell, a process known as carcinogenesis: a process that is well studied in colorectal cancer. Reason being, unlike most other neoplasms, colorectal tumour at the various cancer developmental stage (from very small benign legions to large metastatic neoplasia) can be acquired for analysis. Moreover, both intrinsic and extrinsic factors contribute to the development of colorectal cancer thus allowing in-depth research on both inherited and somatic modifications. Therefore, colorectal cancer is one of the most ideal systems to study genetic alterations in the development of cancer [2].

2.2.2: Molecular carcinogenesis in colorectal cancer

Numerous clinical and histopathological data suggest that the majority of colorectal cancer begin as benign polyps (adenomas) [2, 3]. Polyps are localised lesions situated above the immediate mucosa [5]. Overtime, these adenomatous polyps progress to malignant lesions by attaining a set of mutations (Figure 2.2), a process known as carcinogenesis [6]. In the past, carcinogenesis was defined as an initiation by mutagens followed by promoters which stimulate the growth of abnormal cells [26]. Presently, oncology defines carcinogenesis as a multistage process of build-up gene defects that determine the characteristic traits of a cancer cell such as self-sufficiency in growth signals, infinite replicative potential, sustained angiogenesis, tissue invasion and metastasis [27].

Figure 2.2: Genetic alteration and carcinogenesis steps in colorectal cancer [6].

According to Poirier, carcinogenesis is broken down into three main stages which include initiation, promotion and tumour cell progression [28]. Initiation takes place when cells are exposed to carcinogens, inflammatory agents or tumour promotors that causes irreversible gene alteration or mutation to the cell [7-9]. Initial mutations, also known as “gatekeeping” mutations provide superior growth capabilities to the said cells, allowing them to outgrow their surrounding epithelial cells and thus form microscopic clone [6]. Adenomatous polyposis coli

(APC) tumour suppressor or germline mutations in APC (familial adenomatous polyposis or

FAP) is the most common gene mutated during the initiation of colorectal cancer. Mutated

APC increases production of CTNNB1 (β-catenin), a key transcriptional regulator that drives

Wnt signalling output [29-31]. The cell then becomes stem-cell-like and can proliferate symmetrically, creating a benign and dysplasia adenoma [32].

However, a person with initiated cells (benign polyps) may not necessarily acquire colorectal cancer before death [5]. Initiated cells will first go through a process known as promotion where cells are exposed to non-genotoxic carcinogens which allows them to attain further mutations or “passenger” mutations such as Kirsten rat sarcoma viral oncogene homolog (KRAS) gene mutation. KRAS mutation contributes to colorectal adenoma expansion and apoptosis evasion but is not essential during adenoma initiation [5-9]. The combination of both, APC and KRAS mutations allow for larger adenomas at early stages of colorectal tumorigenesis but with minimal malignant potential (e.g., aberrant crypt foci and hyperplastic polyps) [4-6]. Tumour promotion is a reversible process, which increases the frequency and accelerates the earlier appearance of tumours in carcinogen-initiated individuals [7-9].

Subsequently, during the progression stage, promoter-dependent, premalignant, initiated cell is irreversibly converted to a malignant and tumour promoter independent cell through a series of other gene mutations such as BRAF, PIK3CA, SMAD4 and TP53 [5-9].

These mutations allow the malignant neoplasm to invade through the basal membrane into the epithelial tissue underlying the colorectal epithelium with further metastases to the lymph nodes and distant organ such as the liver.

2.2.3: Genetic architecture of colorectal cancer

In colorectal cancer, mutations accumulate throughout the years. This is probably why a 90-year-old patient has nearly double the number of mutations as a morphologically identical colorectal cancer as in a 45-year-old patient. These mutations are usually “passenger mutations”

40 which does not drive neoplastic process but accumulates in numbers of mutations over the years. Hence, colorectal cancer has one of the highest numbers of microsatellite instability detected by genome-wide sequencing studies (Figure 2.3) in comparison with other cancer types [6]. Interestingly, the number of mutations is a form of measurement indication of the stage of colorectal cancer. Through clinical observations and analysis of common mutations,

Jones and colleagues concluded that it takes approximately 17 years for a large benign tumour to evolve into an advanced cancer and less than 2 years for the same cells to achieve metastasis

[10]. Understanding the genetic aetiology of colorectal cancer can, therefore, assist in the accurate prediction of an individual’s risk and help improve preventative strategies and disease management.

In recent years, a significant amount of effort has been made to identify the genetic architecture of colorectal cancer [33]. It is now understood that 12-35 % of colorectal cancer are attributed to inherited susceptibility [34-36]. Of these inherited cases, approximately 25 %

(3-5% of all colorectal cancer) are categorised as rare high penetrance mutations syndromes

[37]. While high penetrance syndrome represents only a small proportion of all colorectal cancer cases, it indicates an imminent risk of developing colorectal cancer. Therefore, comprehending high penetrance syndrome can lead to better disease management for patients and their relatives [5, 33]. In addition, the study of the molecular basis of familial cases has helped identify specific factors and mechanisms that contribute to sporadic colorectal cancer development.

Figure 2.3: Number of somatic mutations in different human cancers as detected by genome-wide sequencing studies [6]

Likewise, when looking at the bulk of colorectal cancer cases, common variant/ polymorphism weakly affects the sporadic development of colorectal cancer in individuals with or without genealogy linked to colorectal cancer. While common variants are categorised as low penetrance mutations, their attribution towards colorectal cancer should not be taken lightly given their high prevalence in the population. Moreover, as different individuals have varied risk alleles, understanding common variants can further enhance personalised colorectal cancer risk predictions [33]. In addition, common variants may alter the risk of colorectal cancer in individuals with hereditary syndromes [33, 38]. Hence, identifying the genetic architecture of colorectal cancer will assist in value-added treatment and deterrence of the disease [33].

As yet, genome-wide association studies (GWAS) have discovered more than 40 independent loci that are involved with low penetrant variants [33]. While intensive efforts have been made to identify the genetic tell-tale of colorectal cancer, it is estimated that a wide number of novel genetic risk factors remain unearthed [6, 33]. This is due to the technical limitations of GWAS. As nature holds it, some regions of the genomes are more challenging to amplify, capture or explicitly mapped to the genome allowing room for misidentification or

“missed detection” of certain mutations [39-41]. Additionally, the coverage/ read depth

(number of times a specific nucleotide in a given gene will be observed in a sequence data) varies considerably. Hence, certain region would lack representation due to chance factor [42,

43]. Finally, yet importantly, the primary tumour is often heterogeneous in nature, containing both neoplastic and normal cells. Presence of these stromal cells increases background DNA content which can mask or insipid rare mutation signal thus reducing the probability of detecting the mutation [44].

2.3 Identification of novel and pathologically significant molecular targets, GAEC1 and FAM134B

Approximately 20 years ago, our group performed a comparative genomic hybridisation analysis to compare the difference between oesophageal squamous cell carcinoma (ESCC) with its corresponding non-neoplastic oesophageal mucosae [11]. Results exhibited a strong signal of genomic amplification at both chromosome 7q and 5p indicating that these regions played a substantial role in the pathogenesis of ESCC [11-13]. These genes were later on known as a gene amplified in oesophageal cancer 1 (GAEC1) [13] and family with sequence similarity 143, member B (FAM134B) [16]. Six years down the road, these genes were further characterised, and their functions began to unfold.

2.3.1 Gene amplified in oesophageal cancer 1 (GAEC1)

Further analysis into GAEC1 sequence showed exact homology to the expressed sequence tag at 7q22 within the 7q22.1 locus [13]. Utilising 5’ and 3’-rapid amplification of cDNA ends (RACE) technique, GAEC1 was identified as a 2052bp long single-exon gene that consists of 327 bp coding sequence, 1207 bp of 3’ and 518 bp of 5’ non-coding sequence

(Figure 2.4). The gene is located within the first intron region of SH2 B adaptor protein 2

(SH2B2) gene in 7q22, thus fitting the criteria as a nested gene. GAEC1 encodes a novel protein of 109 amino acids (~15 kDa). Additionally, the non-coding 3’ end of GAEC1 mRNA contains several putative polyadenylation signals and a binding site for the Adenylate-uridylate-rich element (AU-rich element; ARE)-binding protein [13].

Interestingly, GAEC1 is frequently overexpressed in a variety of normal tissues, particularly the spleen, small intestine and peripheral blood [13] which are all abundant in immune cells [45-47], suggesting that GAEC1 may play a role in the immune system. While

GAEC1 was expressed in multiple normal tissues; densitometry analysis suggests that GAEC1 band intensity level was higher in 60% of ESCC cell lines and 35% of primary tumours as compared to their normal counterparts (NE1 cell line and paired non-tumour specimen, respectively). Overexpression of GAEC1 in vitro and in vivo showed capabilities to form foci colonies comparable to HRAS. Injection of GAEC1 in 3T3 cells into athymic nude mice successfully formed undifferentiated sarcoma thus proving its capability as a transforming oncogene in ESCC [13]. Further functional studies showed that suppression of GAEC1 in

ESCC cell line KYSE150 inhibited cell proliferation, increased apoptotic population and severely downregulated Calpain 10 (CAPN10). At the tissue level, elevated expression of

CAPN10 is significantly associated with longer survival outcome in patients with ESCC [48].

Interestingly, CAPN10 is associated with type 2 diabetes mellitus and was reported as a link between the disease and colorectal cancer susceptibility [49].

Figure 2.4 Nucleotide with a predicted amino acid sequence of GAEC1. The 5′ and 3′ non-coding sequences ranged from −518 to −1 and 330 to 1534 respectively. The translation initiation codon (ATG) is underlined and italicized, and the stop codon (TGA) is underlined. GAEC1 contains a tract of six GTG trinucleotides repeats in its coding DNA and a track of four CAC trinucleotides in its 3′ non-coding region (shown in bold letters). The Kozak sequence is shaded in grey, the putative polyadenylation signals are boxed, and the binding site of ARE-binding protein for RNA degradation is double-underlined [13].

2.3.2 Colorectal cancer and GAEC1

2.3.2.1 Biomarker potential of GAEC1

The first study on clinical properties of GAEC1 in patients with colorectal cancer was reported in 2010. This study focused on the somatic DNA copy number alteration of GAEC1 in colorectal cancer. CNA is a segment of genomic DNA which varies in its copies compared to a universal reference genome, either through (copy number gains or losses) [50-54]. CNA

45 is an imperative element of genetic variation which has shown to affect a greater portion of the genome in comparison to single nucleotide polymorphisms (SNP) and can cause a wide range of diseases [55]. Since the basis of CNA is in identifying structural duplications and deletions, it makes sense to utilise CNA detection as a cancer risk indicator, a disease well known for its genomic instability and structural dynamism [55].

Identification of GAEC1 copies allowed significant discrimination between the different pathological stages of colorectal tissues (non-neoplastic colorectal tissues, adenomas and colorectal adenocarcinomas) [56]. More importantly, GAEC1 copies altered significantly between the normal population and patients with colorectal cancer [56]. GAEC1 was even able to distinguish between the different ethnicity of patients with colorectal cancer. In particular,

GAEC1 is highly amplified in the Japanese population in comparison with the Australian population [57]. Furthermore, GAEC1 copies also differed in relation to the site of colorectal cancer [56]. These strengthen the potential for GAEC1 as a biomarker for colorectal cancer.

Variation of GAEC1 may encompass changes occurring in more than one gene [51].

This means that not only will a change in CNA of GAEC1 can affect the gene itself; it may also affect other genes. While some CNA has little or no effect on phenotypes, others may give rise to phenotypic variation and influence susceptibility to disease [53, 58]. Even if a CNA presents itself as an apparent non-pathogenic CNA, it could still predispose a region to chromosomal rearrangements, ultimately giving rise to disease [53, 58].

The expression of GAEC1 on both transcriptional and translational level was then scrutinised to identify its association with clinical and pathological parameter of patients with colorectal cancer. In this regard, Wahab and colleagues found that as the stages of colorectal cancer increased, the immunohistochemical staining intensity of GAEC1 in the clinicopathological tissues increased [59], a clear indication of GAEC1 biomarker potential.

However, to utilise GAEC1 as a biomarker in the clinical settings, a more accurate and/or absolute CNA quantification is required.

23.2.2 Oncogenic properties of GAEC1

Subsequent functional studies of GAEC1 showed that silencing of GAEC1 in colon cancer cells significantly reduced colony formation and impeded the ability of colon cancer cells to migrate. Silencing of GAEC1 also induced apoptosis in colon cancer cells [60].

Furthermore, subcutaneous injection of SW480 colon cancer cells with or without silencing of

GAEC1 in severely combined immunodeficiency mice revealed that tumour formation ceased upon silencing of GAEC1 in comparison to the wild-type SW480 [60]. Reduced expression of

GAEC1 was associated with reducing oncogenic properties; this further strengthens the previous notion of GAEC1 as a transforming oncogene in both ESCC and colorectal cancer.

2.3.2.3 Molecular mechanism uncertainties of GAEC1

Nonetheless, there is still a huge gap of knowledge in the exact manner GAEC1 affects colorectal cancer. To-date, there are no studies on the molecular mechanism by which GAEC1 affects colorectal cancer. What are the pathways or genes that GAEC1 interact with to drive cancer? Are there any epigenetic modifications or mutations which allows GAEC1 expression changes at the different pathological stages and cancer sites? A mutational analysis done by

Law and colleagues indicated that there was no mutation, alternative splicing, polymorphism or somatic mutations of GAEC1 at the coding region in ESCC [13]. Even so, many studies have proven that, mutations occurring outside the coding region have the potential to affect gene expression.

The truth is, only a small portion of the DNA sequence is used to make proteins. The other parts of the sequence contain regulatory elements that specify how, when and where the proteins

47 are made. Any protein binding site within this region is subjected to a mutation that potentially changes the identity or relative positioning of nucleotides involved in the DNA-protein interaction.

These mutations can consequently inactivate promoters, regulatory sequences or terminators thus affecting the gene expression [61]. For example, mutation at the 5’ regulatory region of a gene can alter the pattern of gene expression while mutation at the introns, exons, or untranslated regions of a gene can affect RNA processing [62].

In addition, the ARE-binding protein site at the non-coding 3’ site of GAEC1 may prove a worthy site to look for possible mutation sites. AREs plays an essential role in the regulation of gene expression by rapidly degrading mRNA [63, 64] or stabilising a mRNA such as Hu-antigen 1

(HuR) [65]. Therefore, expending the search for mutation to include the non-coding region of

GAEC1 may help further identify possible sites of mutations that may affect GAEC1 expression.

2.3.3 Family with sequence similarity 134 member B (FAM134B)

Family with sequence similarity 143, member B (FAM134B) is also known as JK-1 or reticulophagy regulator 1 (RETREG1) [16]. This gene belongs to a family of three genes,

FAM134A, FAM134B and FAM134C [66]. FAM134B (accession no. NM_019000) is a gene containing ~3MB sequence, located at 5p15.1, downstream of δ-catenin (CCND2) which consist of nine exons and produces two variants [67]. Human FAM134B cDNA variant 1

(accession no. NM_001034850.1) encodes a protein of 497 amino acids (~54kDa) containing a conserved (DDFELL) microtubules-associated protein 1A/1B-light chain 3 (LC3), gamma- aminobutyric acid receptor-associated protein (GABARAP)-interacting region (LIR), C- terminal coiled-coil domain and two long hydrophobic segments harbouring reticulon- homology domain (RHD) [66, 67]. Whereas, FAM134B variant two (accession no.

NM_0190000.3) varies at the 5’ untranslated region (UTR) and encodes a smaller variant protein of 359 amino acids (~39kDa) compared to variant one [14, 67].

2.3.3.1 FAM134B association with various human diseases

Since 2001, studies have found various health conditions associated with FAM134B ranging including maintaining the homeostasis of normal physiological state of the cells through regulating endoplasmic reticulum (ER) turnover by lysosomal degradation [68, 69] and functioning as a potential stem cell marker [70]. Interestingly, FAM134B also plays a vital role in various diseases [66, 67, 71-77]. For example, FAM134B inhibits Ebola, Dengue, West

Nile and Zika viral replication [71, 72]. Mutation of FAM134B caused hereditary sensory and autonomic neuropathy type II disease [66, 67, 73-76]. FAM134B and tumour necrosis factor receptor superfamily, member 19 showed strong synergistic epistasis in influencing the susceptibility of vascular dementia [77]. In addition, changes in FAM134B expression has also been linked to several diseases [14, 16, 78, 79]. For example, overexpression of FAM134B is associated with higher risk of developing allergic rhinitis [78] and later stages of oesophageal cancer [14]. Whereas, in colorectal cancer and breast cancer, FAM134B seem to inhibit cancer growth [16, 79].

Since FAM134B is involved in ER turn over, any physiological changes of the

FAM134B protein may lead to ER-stress, which can subsequently activate unfolded protein response (UPR) in cancer [80]. Interestingly ER-stress or UPR may facilitate cancer growth or induce cancer cell death. On the one hand, ER-stress or UPR could promote better adaptation of cancer cells to the tumour microenvironment and enhance cancer growth [16]. On the other hand, ER-stress or UPR could result in cell cytotoxicity and induce p53- mediated apoptosis of cancer cells, subsequently leading to cell death [81, 82]. Thus, whether or not FAM134B functions as an oncogenic or a tumour suppressive gene is largely dependent on the cellular context or tumour microenvironment [16].

2.3.4 Colorectal cancer and FAM134B

In 2014, Kasem and colleagues identified the expression of the FAM134B protein in the nuclei of colorectal cancer adenocarcinoma cells using immunohistochemistry [79].

However, FAM134B protein expression seemed to undergo cytosol-nucleus transport in vitro and was found at the cytoplasm instead [83]. The reason behind this intra-compartment translocation of FAM134B protein remains unknown.

Deletion of FAM134B was frequently observed in patients with colorectal adenocarcinoma [84]. In addition, reduced expression of FAM134B mRNA and protein were correlated with various clinicopathological features such as increased tumour size, the presence of lymphovascular invasion, recurrence [84, 85] and poorer pathological outcomes [83, 85].

Interestingly, reduced expression of FAM134B was also observed in colon cancer cells derived from later pathological stages compared to earlier stages or non-neoplastic colonic epithelial cells [83]. These findings indicated the tumour suppressive role of FAM134B in colorectal cancer.

2.3.4.1 Tumour suppressive roles of FAM134B

Moving forward, further functional studies were carried out to further scrutinise the role of FAM134B in colorectal cancer. Kasem and colleagues (2014) found that silencing of

FAM134B in colon cancer enhanced cell proliferation, colony formation, migration rate and wound healing capacity of cancer cells in vitro [79, 83]. Furthermore, inhibition of FAM134B also altered the cell cycle of colon cancer cells by increasing the population of cells within the synthesis phase in vitro [83]. More importantly, xenotransplantation of FAM134B knock down colon cancer cells into immunodeficient mice resulted in a larger and higher histological grade carcinoma compared to xenotransplantation of the control cancer cells [83].

2.3.4.2 Understanding the molecular mechanisms of FAM134B

More in-depth functional analysis into FAM134B has helped unveil the mechanism by which FAM134B exercised its tumour suppressive function in colorectal cancer [86, 87]. In a recent study, methylation of FAM134B [86] was more frequently observed in patients with colon adenocarcinoma when compared to adenomas and non-neoplastic mucosae. The same was observed in vitro where colon cancer cells exhibited higher methylation of FAM134B in comparison to normal epithelial cells. Methylation of FAM134B was inversely correlated with

FAM134B expression and copy number and was clinically associated with biological aggressiveness of colon adenocarcinoma. Whereas, demethylation of FAM134B in vitro, restored the expression of FAM134B [86].

Islam and colleagues (2017) found that miR-186-5p which functioned as an oncogenic miRNA in colon cancer could repress the expression of FAM134B in colon cancer to drive cancer progression [87]. On the other hand, anti-miR-186-5p mediated upregulation of

FAM134B impeded cancer cell growth [87]. This highlights the potential of FAM134B as a novel therapeutic target for colorectal cancer initiation and progression.

In addition, the interacting partners of FAM134B were identified for the first time [88].

Liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS) was used to analyse anti-FAM134B co-immune precipitation of FAM134B interacting complex to identify potential interactors of FAM134B in vitro. A total of 28 novel binding partners were identified

(Table 2.1). Further immunoassays confirmed the physical interaction of FAM134B with

CAP1, EB1, CYPB and KDELR2 in colon cancer cells. Knock down of FAM134B significantly up-regulated EB1, a gene which inactivates the APC (adenomatous polyposis coli) gene and activates the β-catenin through the Wnt/β-catenin signalling pathway in colorectal carcinogenesis [88].

Table 2.1: Novel interacting partners of FAM134B protein [88] Protein name Gene name Adenylyl cyclase‐associated protein 1 CAP1 40S ribosomal protein S28 RPS28 ER lumen protein‐retaining receptor 2 KDELR2 Microtubule‐associated protein 4 MAP4 Microtubule‐associated protein RP/EB family MAPRE1/EB1 member 1 Ferritin heavy chain FTH1 Peptidyl‐prolyl cis‐trans isomerase B PPIB/CYPB Complement component 1Q subcomponent‐ C1QBP binding protein Proteasome activator complex subunit 1 PSME1 4F2 cell‐surface antigen heavy chain SLC3A2 T‐complex protein 1 subunit zeta CCT6A Serine hydroxy‐methyl transferase SHMT1 Calnexin CANX 40S ribosomal protein S20 RPS20 40S ribosomal protein S5 RPS5 Peroxiredoxin‐5 PRDX5 Neutral alpha‐glucosidase AB GANAB 26S proteasome non‐ATPase regulatory subunit 6 PSMD6 Translationally‐controlled tumour protein TPT1 Eukaryotic translation initiation factor 3 subunit H EIF3H Tubulin alpha‐1C chain TUBA1C Bifunctional purine biosynthesis protein PURH ATIC Replication protein A 14 kDa subunit RPA3 Aminoacyl tRNA synthase complex‐interacting AIMP2 multifunctional protein 2 Glutaredoxin‐1 GLRX Dolichol‐phosphate mannosyltransferase subunit 1 DPM1 Cytoplasmic dynein 1 intermediate chain 2 DYNC1I2 Proteasome subunit alpha type‐1 PSMA1

Chapter 3: Materials and methods

3. 1 Recruitment of tissue samples (Aims 1, 2 and 4)

Cancer tissues and matched non-cancer tissues near the surgical resection margin were

collected from the same patient who underwent resection of colorectal carcinomas at a hospital

in Queensland between the years 2012–2014 in a Queensland-based hospital. Patients were

recruited on an ongoing basis without bias. These samples were snap frozen in liquid nitrogen

and stored at −80°C until use. Tissues that lacked adequate cancer mass or were lost in clinical

follow-up were excluded from further analysis. Ethical approval was obtained for the use of

these tissues from the Griffith University Human Research Ethics Committee (GU Ref No:

MSC/17/10/HREC).

3.2 Clinicopathological parameters (Aims 1, 2 and 4)

Each resection specimen was fixed in formalin and macroscopically examined. The size

(maximum dimension in mm) and site of the cancer was documented. Proximal cancers were

cancer located in the caecum, ascending colon and transverse colon whereas distal cancers were

defined as the cancers found in the region of descending colon, sigmoid colon and rectum.

The specimens were then dissected for paraffin block preparation. Histological sections

from the formalin fixed paraffin-embedded tissues (FFPE) were microtomed and stained with

haematoxylin and eosin (H&E) staining and viewed under a light microscope by a pathologist

for reviewing and grading. Colorectal carcinomas were graded and typed according to the

World Health Organization (WHO) criteria [89]. The adenocarcinoma was graded into well-

differentiated (grade 1), moderately differentiated (grade 2) and poorly differentiated (grade 3).

The colorectal carcinomas were then staged according to the 8th edition of Cancer staging

Manual of American Joint Committee on Cancer (AJCC) based on tumour, lymph node and

metastases (TNM) classification [90]. The cancer tissues were tested for microsatellite

53 instability (MSI) markers (by immunochemistry) and tested for BRAF mutation status (via

Sanger sequencing) where MSI proved positive according to the clinical guidelines. The clinical status and prognosis were obtained with the help of a surgeon and pathologist. Specific details of the tissues used are noted in the relevant sections of each chapter.

3.3 Clinical Management (Aims 1, 2 and 4)

Clinical management was executed according to standardised multi-disciplinary protocol. All patients’ status was discussed in a weekly multi-disciplinary team management meeting for the duration of their course of management. Post-adjuvant therapy option was decided based on pathological staging. The follow-up period was determined as the interim between the date of surgical resection and the date of death or end date of the study. The actuarial survival rate of the patients was assessed from the date of surgery to the date of death or last follow-up. Only cancer-related deaths were accepted as the endpoint in the statistical analysis [91]. Persistence or recurrence of the disease was also documented.

3.4 Fresh frozen tissues processing (Aims 1, 2 and 4)

For each tissue, additional sections were also taken and stained with hematoxylin and eosin (H&E) to ensure the pathological specificity of the samples used. A total of 125 matched fresh frozen tissue samples were sliced at 7μm thick sections using a cryostat system (Leica

Biosystems, Nussloch, Germany) and placed on a negatively charged slide. Tissues are then immersed in Quick dip fixation for at least 30s before cleaning with water by dipping them in distilled water 10 times in preparation for H&E staining. Only tissues that are comprised of at least 70% of cancer were used for further analysis.

3.5 Formalin-fixed paraffin-embedded tissues (FFPE) processing (Aim 2 and 4)

For each tissue, additional sections were also taken and stained with hematoxylin and eosin (H&E) to ensure the pathological specificity of the samples used. Samples were sliced at

7 μm thick sections using a microtome system (Leica Biosystems, Nussloch, Germany) and placed on a negatively charged slide. These slides were heated at 65°C for 35min to ensure the sections are fixed. After that, slides were dipped in xylene for 10min twice, followed by dipping in 100% ethanol for 10min twice and then 95°C for 5min. Slides were then rinsed under running tap water for 5min. Only tissues that are comprised of at least 70% of cancer were accepted as samples for further analysis.

3.6 Hematoxylin and eosin staining (Aims 1, 2 and 4)

Briefly, before staining with hematoxylin for at least 30s and washing off the stain by dipping into distilled water for another 10 times. Then, slides were stained with Scott’s bluing solution (Appendix D) for 10s before again washing of the stains by dipping the slides for 10 times in distilled water. Next, slides were dipped 5 times in 70% ethanol followed with soaking for 20s in Eosin. Slides were then dipped in 100% ethanol for 30 dips before dipping it in xylene 20 times and then mounting the slides with mounting media. Slides prepared were then checked by a practising pathologist to identify the morphology of the tissue.

3.7 Cell culture (Aim 2, 3 and 4)

Three colon cancer cell lines SW480 (ATCC® CCL-228™; Duke’s class B/ stage II, colonic adenocarcinoma), SW48 (ATCC® CCL-231™; Duke’s class C/ stage III, colonic adenocarcinoma) and HCT 116 (ATCC® CCL-247™; Duke’s class D/ stage V, colonic

55 carcinoma) and one non-neoplastic colon epithelial cell line FHC (ATCC® CRL-1831™) were obtained from the American type culture collection (ATCC) and maintained accordingly.

SW480, SW48 and HCT 116 cells were cultured in RPMI (Roswell Park Memorial

Institute) 1640 medium supplemented with 10% foetal bovine serum (FBS) and 1% penicillin/streptomycin and incubated in a 37°C incubator supplemented with 5% CO2. FHC cells were cultured in Dulbecco's Modified Eagle Medium (DMEM): F-12 media supplemented with extra buffer HEPES (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid) for a final concentration of 25mM, 10ng/mL cholera toxin, 0.005mg/mL insulin, 0.005mg/mL of transferrin, 100ng/mL transferrin, 10% FBS and 1% penicillin/streptomycin and incubated in a 37°C incubator supplemented with 5% CO2.

All cells were routinely checked for mycoplasma contamination using e-MycoTM mycoplasma PCR detection kit. Briefly, 5 x 104 cells were harvested and washed with phosphate buffered saline (PBS) and pelleted twice. Then, cells were re-suspended in 100µL of PBS and heated at 95°C for 10mins before vortexing for 10s and subsequently centrifuge at

13,000rpm for 2mins. Quality and concentration of DNA was measured with the A260/A280 ratio as well as the A260/A230 ratio using NanoDrop Spectrophotometer (Thermo Fisher

Scientific, Waltham, Massachusetts, United States). 30ng of genomic DNA (gDNA) were used as a template and added into the e-MycoTM. Mycoplasma PCR Detection Kit (ver2.0) tube and then topped up with DNase free water to 20µL. PCR was performed on the thermal cycle using the following condition: Initial denaturation (94°C for 1min); denaturation (94°C for 30s), annealing (60°C for 20s) and extension (72°C for 1min) for 35 cycles; and, final extension

(72°C for 5mins). PCR product was resolved with a 2% agarose gel electrophoresis (Figure

3.1).

Figure 3.1: Mycoplasma contamination testing Lanes in order from left to right are a 100kb ladder, non-template control (NTC), mycoplasma positive cells, mycoplasma negative cells and a positive control. The pale band at the bottom represents the end of the front. Mycoplasma positive detection has a band size of ~270kb and internal control band is ~170kb. Electrophoresis was ran on a 2% gel for 90min at 60V.

3.8 Genomic DNA extraction (Aim 1 and 2)

Matched samples which showed possible signs of mutations were run through a 2% gel electrophoresis and its band was excised and cleaned up using the NucleoSpin® Gel and PCR

Clean-up kit according to the manufacturer’s protocol. Specifically, 400μL of Buffer NTI was added to 200mg of an agarose gel containing the excised band and incubated for 10min at 50 °C, vortexing at 2min intervals until completely dissolved. Next, 550μL of the sample was loaded onto the NucloSpin® Gel and PCR Clean-up Column and centrifuged at 11,000g for 30s. Flow- through was discarded and any remaining sample was loaded on the column and the same process was repeated. The column was then washed with 550μL Buffer NT3 and centrifuged at 11,000g for 30s. Flow-through was discarded, and the washing step was repeated. Next, the column was centrifuged at 11,000g for 1min to remove any remaining Buffer NT3. Care is ensured that column and flow-through do not come in contact. Then, columns were left to incubate at 70°C for 5min before loading 15μL DNase RNase free water onto the membrane

57 of the column for 3min and then centrifuging at 11,000g for 1min to elute DNA. Quality and concentration of DNA were measured with A260/A280 ratio using NanoDrop

Spectrophotometer (Thermo Fisher Scientific, Waltham, Massachusetts, United States). The extracts were stored at 4°C until used.

3.9 Droplet digital PCR (ddPCR) analysis (Aim 1)

ddPCR was used to examine the copy number aberration (CNA) of GAEC1 in colorectal cancer tissues and adjacent non-neoplastic mucosae. All reactions were prepared at a 1.2x of the total reaction to account for potential reaction loss. Samples were diluted to 30ng/µL. ddPCR was performed in a total volume of 22 µL reaction mixture containing 12µL QX200 ddPCR EvaGreen Supermix (Bio-Rad Laboratories, Hercules, California, USA), relevant volumes of primer set and DNase free water control. QX200™ AutoDG™ Droplet Digital™

PCR System (Bio-Rad Laboratories, Hercules, California, USA) was used to generate the PCR droplets and the C1000 thermal cycler (Bio-Rad Laboratories, Hercules, California, USA) was used to amplify the reactions to the endpoint. PCR products were then transferred onto a

QX200 Droplet Reader (Bio-Rad Laboratories, Hercules, California, USA) analysed by

QuantaSoft software version 1.7.4 (Bio-Rad Laboratories, Hercules, California, USA) and

QuantaSoftTM Analysis Pro Software 1.0.596 (Bio-Rad Laboratories, Hercules, California,

USA). Specific details of the protocol and primer pairs used in the present study are described in detail in Chapter 4 and 5.

STATEMENT OF CONTRIBUTION TO CO-AUTHORED PUBLISHED PAPER

Chapters 3 includes two co-authored papers. The bibliographic details of the first co-authored paper, including all authors, are:

Title: Somatic DNA Copy-Number Alterations Detection for oesophageal Adenocarcinoma Using Digital Polymerase Chain Reaction

Authors: Katherine Ting-Wei Lee, Vinod Gopalan, Alfred King-yin Lam

Name of the book: oesophageal Adenocarcinoma: Methods and Protocols, Methods in Molecular Biology, vol. 1756

Status: Published

Publication link: https://doi.org/10.1007/978-1-4939-7734-5_18

My contribution to the paper involved: I drafted the manuscript, designed, optimised and ran the entire ddPCR for CNA detection and performed the entire workflow for sample preparation. I processed the tissue samples and harvested the DNA samples. I was involved in the statistical analysis of this manuscript.

(Signed) ______(Date)___ 22/11/2018__ Katherine Ting-Wei Lee

(Countersigned) ______(Date)___ 22/11/2018__ Corresponding author of paper and principal supervisor: Prof. Alfred King-Yin Lam

(Countersigned) ______(Date)__ 22/11/2018___ Corresponding author of paper and principal supervisor: Dr. Vinod Gopalan

3.9.1 Detailed description of copy number alteration using ddPCR

Somatic DNA Copy-Number Alterations Detection for oesophageal Adenocarcinoma Using Polymerase Chain Reaction

Katherine T. W. Lee, Vinod Gopalan, Alfred K. Lam

3.9.1.1 Abstract

Somatic copy-number alterations are commonly found in cancer and play key roles in activating oncogenes and deactivating tumour suppressor genes. Digital polymerase chain reaction is an effective way to detect the changes in copy number. In oesophageal adenocarcinoma, detection of somatic copy-number alterations could predict the prognosis of patients as well as the response to therapy. This chapter will review the methods involved in digital polymerase chain reaction for the research or potential clinical applications in oesophageal adenocarcinoma.

Key Words: Digital PCR, oesophageal adenocarcinoma, somatic copy-number alterations

3.9.1.2 Introduction

Reports have indicated that genomic mutations are frequent in oesophageal adenocarcinoma [1]. Among these mutations, somatic copy-number alterations involve a greater fragment of the genome in cancer compared to any other mutations [2]. Somatic copy- number alterations are changes in the copy number of a genome fragment within the somatic tissue. Somatic copy-number alterations are exceptionally common in cancer [3] and play imperative roles in activating oncogenes [4, 5] and deactivating tumour suppressor genes [5-

8]. Unravelling how somatic copy-number alterations affect the biological phenotypic has enabled further progresses in cancer diagnostics and therapeutics [9, 10]. Table 3.1 summarises the changes in copy number noted in patients with oesophageal adenocarcinoma.

Table 3.1: Amplification or deletion of genes involved in oesophageal adenocarcinoma Amplification/ Gene Ref. Deletion Amplification Anaplastic lymphoma receptor tyrosine kinase (ALK) [11] AKT serine/threonine kinase 1 (AKT1) [12] ATM serine/threonine kinase (ATM) [13] ATP binding cassette subfamily B member 1 (ABCB1, [13] MDR) ATP binding cassette subfamily C member 4 (ABCC4) [14] B-Cell CLL/Lymphoma 9 (BCL9) [15] B lymphoma Mo-MLV insertion region 1 homolog [16] (Bmi-1) Calneuron 1 (CALN1) [11] Cell division protein kinase 6 (CDK6) [17] Cerebellin 4 precursor (CBLN4) [11] Cortactin (CTTN, EMS1) [14] Cyclin D1 (CCND1) [12, 14] Cyclin E (CCNE1) [12, 18] c-Myc (Myc) [11, 12, 15, 19, 20] Cutaneous T-cell lymphoma-associated antigen 1 [15] (CTAGE1) DEAH-box helicase 15 (DHX15, DDX15) [13] Dual specificity tyrosine phosphorylation regulated [12, 21] kinase 2 (DYRK2) Ephrin type-B receptor 4 is a protein (EPHB4) [22] Epidermal growth factor receptor (EGFR) [12, 17, 19, 20, 23] Family with sequence similarity 84 member B [11] (FAM84B) Fibroblast growth factor receptor 2 (FGFR2) [24] Fibroblast growth factor 3 (FGF3) [14] GATA binding protein 4 (GATA4) [11] Granulysin (GNLY) [13] Growth factor receptor-bound protein 7 gene (GRB7) [25] Forkhead Box A1 (FOXA1, HNF3alpha) [12] Human epidermal receptor growth factor 2 (ERBB2/ [19, 23, 25- HER2) 28] Immediate early response 3 (IER3) [14] Insulin like growth factor 1 receptor (IGF1R) [12-14] KRAS proto-oncogene, GTPase (KRAS) [12] Kruppel like factor 12 (KLF12) [19] Laminin subunit alpha 3 (LAMA3) [13] Leucine rich repeat containing 37 member A3 [11] (LRRC37A3) Mediator of DNA damage checkpoint 1 (MDC1, [14] KIAA0170) MET proto-oncogene (MET) [17] Murine double minute -2 (MDM2) [29]

Table 3.1 (continued) Amplification/ Gene Ref. Deletion Amplification MutS homolog 2 (MSH2) [13] (continued) MYB proto-oncogene like 2 (MYBL2) [13, 14] Myotubularin related protein 9 (MTMR9) [23] Nei like DNA glycosylase 2 (NEIL2) [23] NME/NM23 nucleoside diphosphate kinase 1 (NME1) [13] Nuclear receptor coactivator 3 (NCOA3, AIB1) [12, 14] Phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic [12] subunit alpha (PIK3CA) Protein tyrosine phosphatase, non-receptor type 1 [13, 14] (PTPN1) RAD51 recombinase (RAD51) [30] Replication factor C subunit 3 (RFC3) [31] Retinoic acid receptor alpha (RARA) [32] Sarcospan (SSPN) [11] Serpin Family E Member 1 (SERPINE1) [13] Small nuclear ribonucleoprotein polypeptide N [13] (SNRPN) Solute carrier family 25 member 29 (SLC25A29) [11] SRY-box 1 (SOX1) [11] SRY-box 7 (SOX7) [11] Telomerase RNA component (TERC) [13] The G-protein coupled bile acid receptor (TGR5) [33] TNF receptor superfamily member 6b (TNFRSF6B) [13] Topoisomerase (DNA) II alpha (TOP2A) [32] Transglutaminase 2 (TGM2) [34] Vitamin D receptor (VDR) [35] V-Ki-ras2 Kirsten rat sarcoma viral oncogene homolog [17, 23] (KRAS) Williams-Beuren syndrome chromosome region 17 [11] (WBSCR17) Wilms tumour 1 (WT1) [23] Zinc finger protein 217 (ZNF217) [13-15] Zinc finger protein 536 (ZNF536) [11] Deletion AKT serine/threonine kinase 3 (AKT3) [13] Cadherin 3 (CDH3) [36] Cadherin 1 (CDH1) [36] Cadherin 20 (CDH20) [11] Checkpoint with forkhead-associated domain and [37] RING-finger domain (CHFR) Coagulation Factor III, Tissue Factor (F3) [14] Cyclin-dependent kinase Inhibitor 2A (CDKN2A) [11, 13, 15, 17, 19, 36, 38-40] Cyclin-dependent kinase Inhibitor 2B (CDKN2B) [15, 17, 19] Cyclin dependent kinase 4 (SAS, CDK4) [13] Cystatin S (CST4) [11] Dystrophin (DMD) [15]

Table 3.1 (continued) Amplification/ Gene Ref. Deletion Deletion Focadhesin (FOCAD, KIAA1797) [11] (continued) Fragile histidine triad protein (FHIT) [11, 13, 15, 17, 19] Galanin Receptor 1 (GALR1) [15] Integrin Subunit Alpha V (ITGAV) [36] Macrophage Migration Inhibitory Factor [14] (Glycosylation-Inhibiting Factor) (MIF) Muscleblind Like Splicing Regulator 1 (MBNL1) [40] Platelet-derived growth factor subunit B (PDGFB) [13] Protein Tyrosine Phosphatase, Receptor Type D [19] (PTPRD) Phosphodiesterase 4D (PDE4D) [15] Runt-related transcription factor 1 (RUNX1) [17, 36] SMAD Family Member 4 (SMAD4) [15, 17, 19, 36] Selenium-binding protein 1 (SELENBP1) [41] The Ras-association domain family member 1 (RASSFI) [13] WW domain-containing oxidoreductase (WWOX) [15, 17] 3-Hydroxyacyl-CoA dehydratase 4 (HACD4, [11] PTPLAD2)

In patients with oesophageal adenocarcinoma that underwent neoadjuvant chemotherapy, those that showed a reduction in DNA copy-number entropy was found to have an improved survival despite the lack of clear morphological evidence in response to the therapy [42]. This indicates that copy number changes might function as a more sensitive tumour response marker in comparison to the changes in morphological tumour phenotype.

For example, patients whom responded to neoadjuvant cisplatin-based chemotherapy and were tested to have lower EGFR copy number had a significantly better disease-free survival than non-responders. In addition, non-responder patients were more often associated with amplified

EGFR copy. These results indicate that EGFR copy number may act as an independent adverse prognostic factor for neoadjuvant cisplatin-based chemotherapy in patients with oesophageal adenocarcinoma [43].

Identification of copy number changes such as the amplification of mitochondria DNA

(mtDNA) in the peripheral blood leukocytes [44] or tissue [45] was associated with increased susceptibility to oesophageal adenocarcinoma. Meanwhile, murine double minute-2 (MDM2) can be used as a pre-screening tool before following up with fluorescence in situ hybridisation

(FISH) for identification of oesophageal adenocarcinoma [29].

Somatic copy-number alterations can also be used to predict outcomes of oesophageal adenocarcinoma. For example, lower level of HER-2 was correlated to better outcome[46] while a relatively lower increased in ZNF217 (Zinc Finger Protein 217) copy number was associated with poorer prognosis and a higher copy number was correlated with improved outcome [46]. Combined amplification of cyclin-dependent kinase 4 and 6 (CDK4 and CDK6) functioned as a better predictor of survival outcome than individually [47].

Specific somatic copy-number alterations analysis has also been a major help in considering target therapy options for oesophageal adenocarcinoma. For example, 15% of the patients with oesophageal adenocarcinoma that are found to have an over-amplification of

HER-2 has a higher survival rate when treated with anti-HER-2 targeted therapy (trastuzumab)

[48, 49]. Therefore, further identifying the altered genomic target can help clinicians in choosing the best treatment method.

On the other hand, Li and colleagues found that the somatic genomes of patients with

Barrett’s oesophagus that did not progress to oesophageal adenocarcinoma were exceptionally stable throughout the observation period. However, genomic alterations were progressively observed in patients with Barrett’s oesophagus up to 48 months before oesophageal adenocarcinoma development, particularly on chromosome 18q [50].

Overall, the values of detecting the copy number change highlight the importance of somatic copy-number alterations analysis. In the past, copy number alterations have been mostly analysed with the use of real-time PCR (qPCR). However, with recent advances in

64 technology, several new methods such as next-generation sequencing and digital PCR (dPCR) have been developed to identify accurately the somatic copy-number alterations. In this chapter, we will be focusing on the use of dPCR to accurately quantify somatic copy-number alteration in patients with oesophageal adenocarcinoma.

While the concept of this technology came from two decades ago [51, 52], this method allows for absolute quantification of targets contained in a sample without the prerequisite of huge technical replicates. dPCR workflow is relatively similar to that of quantitative real-time

PCR (qPCR) wherein the reaction mixture containing sample nucleic acid, primers and/or probes plus the master-mix are prepared accordingly. Unlike qPCR though, dPCR first partitions samples into tens of thousands of minute size reaction chamber prior to thermal cycling to the endpoint and counting the positive signals versus the negative signals [51-54].

Currently, there are two ways of partitioning the samples, one through physical partitioning chambers [54] and another through droplet generation via a water-oil emulsion concept [55].

In the following section, we describe the emulsion-based method of dPCR known as droplet digital PCR (ddPCR).

3.9.1.3 Materials

Prepare all reagents using ultrapure water and analytical grade chemicals. Ensure all plastic wares and reagents are DNase-free. Unless otherwise indicated, prepare and store all reagents at room temperature. Follow the rules and regulations for waste disposal. Appropriate personal protective equipment must be worn during experiments to safeguard oneself against laboratory dangers.

3.9.1.3.1 Tissue collection and DNA extraction

1. Snap-frozen cancer tissue and adjacent non-neoplastic tissues from patients with

oesophageal adenocarcinoma.

2. 100% undenatured ethanol.

3. DNase-free water.

4. Micropipettes (10µL, 20µL, 200µL and 1000µL).

5. DNase-free filtered tips (10µL, 20µL, 200µL and 1000µL).

6. Microcentrifuge tubes.

7. Microscope slides with a cover slip.

8. Haematoxylin and eosin (H&E) staining reagents and holders.

9. Microscope.

10. Cryotome machine and relevant consumables (blades, cryocassettes and etc.).

11. Microcentrifuge machine.

12. Vortexer.

13. Nanospectrophotometer (e.g.: NanoDrop Spectrophotometer).

14. DNA extraction kit (e.g.: Tissue & Blood extraction Kit from Qiagen).

3.9.1.3.2 ddPCR primer optimisation

1. DNase-free water.

2. Micropipettes (10µL, 20µL, 200µL and 1000µL).

3. DNase-free filtered tips (10µL, 20µL, 200µL and 1000µL).

4. Electrophoresis gel tank system and agarose gel caster with well combs.

5. 1x Tris-acetate-EDTA (TAE) buffer.

6. 10x loading dye.

7. 100bp DNA ladder.

8. DNase-free water.

9. 5% agarose gel

10. Microcentrifuge tubes.

11. Polymerase chain reaction (PCR) tubes.

12. 96-well PCR plates.

13. Pierceable foil heat seal.

14. Primers.

15. ddPCR supermix (e.g.: Bio-Rad ddPCR™ Supermix for Probes).

16. Restriction enzyme.

17. PCR hood.

18. Conventional PCR machine with gradient enabled function.

19. Droplet generator (E.g.: QX200 Droplet Generator).

20. Droplet reader (E.g.: QX200 Droplet Reader).

21. Microfluidic cartridge.

22. Gaskets.

23. Generation oil (E.g.: Bio-Rad Droplet Generation Oil for Probe).

24. Microcentrifuge machine.

25. Computer

26. QuantaSoft software.

3.9.1.4 Method

3.9.1.4.1 Genomic DNA extraction

1. Oesophageal cancer tissues are cryotomed into 5 slices of 5µM sections.

2. Place a section onto a microscope slide for haematoxylin and eosin (H&E) staining to

ensure pathological confirmation (see NOTE 1).

3. Extract genomic DNA from an oesophageal tissue sample using commercially

available tissue extraction kit according to the manufacturer’s guideline (see NOTE 2

& 3).

4. Measure the A260/A280 ratio and the A260/A230 ratio of the resulting elution using a

nanospectrophotometer (see NOTE 4).

5. Ensure that the A260/A280 ratio is approximately 1.8 and the A260/A230 ratio is between

2.0-2.2 (see NOTE 5).

6. Store extracted DNA in -20°C (see NOTE 6).

3.9.1.4.2 Droplet digital PCR (ddPCR) primer design

1. Both primer sets for the target gene and control gene are designed to be approximately

80bp and 60bp individually. Primers can be designed using a variety of available tools

ranging from paid software to freeware such as the NCBI (National Centre for

Biotechnology Information) primer design tool

(https://www.ncbi.nlm.nih.gov/tools/primer-blast/) which uses Primer3 and BLAST to

help design the primers (Figure 3.2).

2. Briefly, enter the accession, gi or FASTA sequence of your gene of interest at the

PCR template section (Figure 3.2).

3. Next, correct the PCR product size to 60bp or 80bp at the Primer Parameters section

(Figure 3.2).

Figure 3.2: Outlook of Primer-BLAST Website for the design of primers

4. Click on Advanced parameters at the bottom of Get Primers button (Figure 3.2) and

scroll to internal hybridization oligo parameters section (Figure 3.3) and tick on the

Pick internal hybridization oligo to design your probes.

5. Fill up any other additional criteria of your preference and then hit Get Primers

(Figure 3.3).

Figure 3.3: Internal hybridisation parameters section for design of probes

3.9.1.4.3 Annealing temperature optimisation for ddPCR primers on conventional PCR

1. Optimise the annealing temperature of the primer sets designed by running a

conventional PCR with gradient temperature.

2. PCR hood should be exposed to ultra violet lights for 25 minutes prior to using it.

3. Prepare all PCR materials within the PCR hood to reduce chances of contamination.

4. Dissolve 5g of agarose powder in 100mL of 1x TAE buffer inside a 250mL conical

flask (see NOTE 7).

5. Put the mixture in the microwave oven to heat for 2 minutes (see NOTE 8).

CAUTION! Conical flask may be hot when removing from the microwave oven.

Ensure that protective heat gloves are worn when handling hot flask.

6. Run a 5% agarose gel at 60V for 150 minutes to determine primer specificity and

identify the best annealing temperature (see NOTE 9).

7. Non-template control must be used in all assays to ensure that there is no cross

contaminations (see NOTE 10).

3.9.1.4.4 ddPCR gradient optimisation

1. Run a second set of gradient ddPCR to select the final annealing temperature.

2. Prepare reaction mixture according to the manufacturer’s protocol with consideration

of pipetting error (see NOTE 11).

3. For example, according to the ddPCR supermix protocol, a reaction mix containing

11µL 2x ddPCR supermix for probes, 1µL of 1µM control gene primer set, 1µL of

1µM target gene primer set, 1µL of 2U/µL restriction enzyme which does not cut the

amplicon of both the target gene and control gene (see NOTE 12), 30ng of template

and DNase free water topped up to a final volume of 22µL.

4. Transfer 20µL reaction volume into the sample section of a microfluidic cartridge and

subsequently with 70µL of generation oil to the oil section.

5. Close the cartridge with a gasket and then place it into a droplet generator.

6. Carefully transfer all of the droplets into a 96-well plate (see NOTE 13) and seal it

with a pierceable foil heat seal.

7. Amplify the sample to end point with a gradient enabled PCR machine using the

following cycling conditions (see NOTE 14): 3 minutes of initial denaturation at

95°C, followed by 40 cycles of 10 seconds denaturation at 95°C, 10 seconds of

annealing at said optimised temperature and 20 seconds of extension at 72°C. Then, a

final extension at 72°C for 10 minutes and infinite holding at 12°C (see NOTE 15).

8. Transfer the PCR plate containing the amplicons onto a droplet reader where target

DNA molecules will be aligned and singulated through a detection channel within the

system to calculate the number of targets in an absolute manner according to

fluorescence amplitude.

9. Analysis can then be performed with the given software that comes with the system

such as QuantaSoft software version 1.7.4.

10. Assign threshold manually for each target as shown in Figure 3.4 according to

manufacturer’s protocol.

11. The temperature that gives the most distinct separation between the four clusters (as

seen in Figure 3.4) is chosen (see NOTE 16).

Figure 3.4: Multiple target detection (genes of interest) using EvaGreen-based ddPCR. (a) One dimensional droplet plot of resulting duplex assay optimised with different primer concentration. The different target genes show different amplitudes. (b) Two dimensional droplet plots of duplex assay. In all cases, manual thresholds were drawn to assign clusters. The different target genes form different clusters.

3.9.1.4.5 ddPCR primer concentration optimisation

1. In order to further distinguish the clusters between one another, the primer

concentration of each target has to be manipulated.

2. Target 1 primer concentration can be tested at a range of 25nM-50nM while Target 2

primer concentration can be tested at a range of 100nM-150nM (see NOTE 17).

3. Repeat step 2-9 of Section 3.4 but instead of using a gradient temperature, use the

optimised temperature instead.

4. Primer combination that gives the most distinct separation between the four clusters

(as seen in Figure 3.4) is chosen (see NOTE 16).

3.9.1.4.6 Target cluster determination

1. In order to identify the different clusters, a single-plex assay containing Target 1 and

Target 2 will be prepared separately and ran on another round of ddPCR in parallel

with the optimised duplex assay.

2. Three reactions of a single sample will be prepared to contain primer set for only

target 1, containing primer set for only target 2 and containing primer set for both

targets.

3. The resulting 1D plot image should look like Figure 3.5.

4. Once the targets are identified, the assay can be run at 96 reactions per run.

Figure 3.5: Reaction preparation containing only one primer set at each well. They are used to identify the amplification cluster of each gene. Well one (left column) contains both primer sets for the duplex detection; well two (middle column) and well three (right column) contains two different primer sets to enable identification of the different clusters within the first well. In all cases, manual thresholds are drawn to assign clusters.

3.9.1.5 Notes

1. Ensure that there is no folding of tissue when placing the section onto the slide. Also,

ensure that the section is a 5µM thick section. This is to ensure proper staining and

visualisation can be done.

2. Ensure that all buffers are topped up with the appropriate solvents required.

3. Use DNase-free water to elute DNA instead of the provided elution buffer. Add the

recommended amount of DNase-free water according to the manufacturer’s protocol

and incubate the column on the bench for 5min before eluting the DNA.

4. Use DNase-free water as blank.

5. Samples with low A260/A280 ratio may have protein contamination while higher

A260/A280 ratio may have RNA contamination. Low A260/A230 ratio may indicate

residual (such as phenol) carryover while high A260/A230 ratio may indicate a dirty

blank used. Read the manufacturer’s troubleshooting guide to help solve this issue.

6. Storing at -20°C prevents evaporation of sample over time and ensures sample

integrity.

7. Using conical flask helps prevent overflowing of agarose liquid when heating it in the

microwave oven.

8. 5% agarose gel is very concentrated and viscous and can solidify in a very short time.

After the initial 2min microwave oven heating, repeat the heating at 1min interval

until all agarose powder is fully dissolved. Floating pieces of gel observed in the flask

is an indication that the agarose gel is not completely melted. Once the gel is fully

dissolved, quickly but carefully transfer the solution onto the agarose gel cast. Ensure

that all bubbles are removed from the gel and leave to solidify.

9. Routinely check the electrophoresis run to ensure that the band does not run over the

gel.

10. Non-template control is prepared by replacing the template with DNase-free water.

11. Prepare 22µL of reaction mix instead of 20µL in consideration of pipetting error and

evaporation.

12. A restriction enzyme is added into the ddPCR reaction mixture to help reduce noise

such as non-specific amplification and thus creating a much cleaner data.

13. Slant the cartridge at approximately 30° and place the micropipette tip vertically into

the well and slowly draw up the droplets.

14. Ramp rate for the entire cycle is set to 2°C/s to ensure even heat distribution

throughout the droplets.

REMEMBER! Set the sample volume to 40µL (this is an approximate value for the

total volume of droplets produced from the emulsion of the generator oil and reaction

mixture) and the heated lid to 105°C.

15. The lowest most of droplet clusters are the non-target containing negative droplets,

while the next two bands of increased fluorescence amplitude were comprised of

droplets that possessed only one target type. The final population of droplets contain

both target sequences and consequently had the greatest fluorescence signal.

16. Note that primer concentration range is written just as a suggested guideline and

further optimisation may be required by increasing the range. However, as a starting

point, the suggested primer concentration range may be used.

3.9.1.6 References

1. Nones K, Waddell N, Wayte N, Patch AM, Bailey P, Newell F, et al. Genomic catastrophes frequently arise in oesophageal adenocarcinoma and drive tumorigenesis. Nat Commun. 2014;5:5224.

2. Zack TI, Schumacher SE, Carter SL, Cherniack AD, Saksena G, Tabak B, et al. Pan-cancer patterns of somatic copy number alteration. Nature Genetics. 2013;45:1134.

3. Beroukhim R, Mermel CH, Porter D, Wei G, Raychaudhuri S, Donovan J, et al. The landscape of somatic copy-number alteration across human cancers. Nature. 2010;463(7283):899-905.

4. Baudis M. Genomic imbalances in 5918 malignant epithelial tumors: an explorative meta- analysis of chromosomal CGH data. BMC Cancer. 2007;7(1):226.

5. The Cancer Genome Atlas Research N. Comprehensive genomic characterization of squamous cell lung cancers. Nature. 2012;489:519.

6. Stephens PJ, McBride DJ, Lin M-L, Varela I, Pleasance ED, Simpson JT, et al. Complex landscapes of somatic rearrangement in human breast cancer genomes. Nature. 2009;462:1005.

7. Weir BA, Woo MS, Getz G, Perner S, Ding L, Beroukhim R, et al. Characterizing the cancer genome in lung adenocarcinoma. Nature. 2007;450:893.

8. Xue W, Kitzing T, Roessler S, Zuber J, Krasnitz A, Schultz N, et al. A cluster of cooperating tumor-suppressor gene candidates in chromosomal deletions. Proceedings of the National Academy of Sciences. 2012;109(21):8212.

9. T. CDV, Agnes A, Wei-Li L, Jing S, Emily OD, D. LR, et al. MET tyrosine kinase receptor expression and amplification as prognostic biomarkers of survival in gastroesophageal adenocarcinoma. Cancer. 2017;123(6):1061-70.

10. Kim SY, Ahn T, Bang H, Ham JS, Kim J, Kim ST, et al. Acquired resistance to LY2874455 in FGFR2-amplified gastric cancer through an emergence of novel FGFR2- ACSL5 fusion. Oncotarget. 2017;8(9):15014-22.

11. Wiech T, Nikolopoulos E, Weis R, Langer R, Bartholome K, Timmer J, et al. Genome- wide analysis of genetic alterations in Barrett's adenocarcinoma using single nucleotide polymorphism arrays. Laboratory investigation; a journal of technical methods and pathology. 2009;89(4):385-97.

12. Miller CT, Moy JR, Lin L, Schipper M, Normolle D, Brenner DE, et al. Gene amplification in oesophageal adenocarcinomas and Barrett's with high-grade dysplasia. Clinical cancer research : an official journal of the American Association for Cancer Research. 2003;9(13):4819-25.

13. Albrecht B, Hausmann M, Zitzelsberger H, Stein H, Siewert JR, Hopt U, et al. Array- based comparative genomic hybridization for the detection of DNA sequence copy number changes in Barrett's adenocarcinoma. The Journal of pathology. 2004;203(3):780-8.

14. Pasello G, Agata S, Bonaldi L, Corradin A, Montagna M, Zamarchi R, et al. DNA copy number alterations correlate with survival of oesophageal adenocarcinoma patients. Modern pathology : an official journal of the United States and Canadian Academy of Pathology, Inc. 2009;22(1):58-65.

15. Nancarrow DJ, Handoko HY, Smithers BM, Gotley DC, Drew PA, Watson DI, et al. Genome-wide copy number analysis in oesophageal adenocarcinoma using high-density single-nucleotide polymorphism arrays. Cancer Res. 2008;68(11):4163-72.

16. Choy B, Bandla S, Xia Y, Tan D, Pennathur A, Luketich JD, et al. Clinicopathologic characteristics of high expression of Bmi-1 in oesophageal adenocarcinoma and squamous cell carcinoma. BMC gastroenterology. 2012;12:146.

17. Davison JM, Yee M, Krill-Burger JM, Lyons-Weiler MA, Kelly LA, Sciulli CM, et al. The degree of segmental aneuploidy measured by total copy number abnormalities predicts survival and recurrence in superficial gastroesophageal adenocarcinoma. PloS one. 2014;9(1):e79079.

18. Lin L, Prescott MS, Zhu Z, Singh P, Chun SY, Kuick RD, et al. Identification and characterization of a 19q12 amplicon in oesophageal adenocarcinomas reveals cyclin E as the best candidate gene for this amplicon. Cancer Res. 2000;60(24):7021-7.

19. Frankel A, Armour N, Nancarrow D, Krause L, Hayward N, Lampe G, et al. Genome- wide analysis of oesophageal adenocarcinoma yields specific copy number aberrations that correlate with prognosis. Genes, chromosomes & cancer. 2014;53(4):324-38.

20. Rygiel AM, Milano F, Ten Kate FJ, Schaap A, Wang KK, Peppelenbosch MP, et al. Gains and amplifications of c-myc, EGFR, and 20.q13 loci in the no dysplasia-dysplasia- adenocarcinoma sequence of Barrett's esophagus. Cancer Epidemiol Biomarkers Prev. 2008;17(6):1380-5.

21. Miller CT, Aggarwal S, Lin TK, Dagenais SL, Contreras JI, Orringer MB, et al. Amplification and overexpression of the dual-specificity tyrosine-(Y)-phosphorylation regulated kinase 2 (DYRK2) gene in oesophageal and lung adenocarcinomas. Cancer Res. 2003;63(14):4136-43.

22. Hasina R, Mollberg N, Kawada I, Mutreja K, Kanade G, Yala S, et al. Critical role for the receptor tyrosine kinase EPHB4 in oesophageal cancers. Cancer Res. 2013;73(1):184- 94.

23. Hjortland GO, Meza-Zepeda LA, Beiske K, Ree AH, Tveito S, Hoifodt H, et al. Genome wide single cell analysis of chemotherapy resistant metastatic cells in a case of gastroesophageal adenocarcinoma. BMC cancer. 2011;11:455.

24. Tokunaga R, Imamura Y, Nakamura K, Ishimoto T, Nakagawa S, Miyake K, et al. Fibroblast growth factor receptor 2 expression, but not its genetic amplification, is associated with tumor growth and worse survival in esophagogastric junction adenocarcinoma. Oncotarget. 2016;7(15):19748-61.

25. Walch A, Specht K, Braselmann H, Stein H, Siewert JR, Hopt U, et al. Coamplification and coexpression of GRB7 and ERBB2 is found in high grade intraepithelial neoplasia and in invasive Barrett's carcinoma. Int J Cancer. 2004;112(5):747-53. 77

26. Kumarasinghe MP, de Boer WB, Khor TS, Ooi EM, Jene N, Jayasinghe S, et al. HER2 status in gastric/gastro-oesophageal junctional cancers: should determination of gene amplification by SISH use HER2 copy number or HER2: CEP17 ratio? Pathology. 2014;46(3):184-7.

27. Lange T, Nentwich MF, Luth M, Yekebas E, Schumacher U. Trastuzumab has anti- metastatic and anti-angiogenic activity in a spontaneous metastasis xenograft model of oesophageal adenocarcinoma. Cancer Lett. 2011;308(1):54-61.

28. Safran H, Dipetrillo T, Akerman P, Ng T, Evans D, Steinhoff M, et al. Phase I/II study of trastuzumab, paclitaxel, cisplatin and radiation for locally advanced, HER2 overexpressing, oesophageal adenocarcinoma. International journal of radiation oncology, biology, physics. 2007;67(2):405-9.

29. Michalk M, Meinrath J, Kunstlinger H, Koitzsch U, Drebber U, Merkelbach-Bruse S, et al. MDM2 gene amplification in oesophageal carcinoma. Oncol Rep. 2016;35(4):2223-7.

30. Pal J, Bertheau R, Buon L, Qazi A, Batchu RB, Bandyopadhyay S, et al. Genomic evolution in Barrett's adenocarcinoma cells: critical roles of elevated hsRAD51, homologous recombination and Alu sequences in the genome. Oncogene. 2011;30(33):3585-98.

31. Lockwood WW, Thu KL, Lin L, Pikor LA, Chari R, Lam WL, et al. Integrative genomics identified RFC3 as an amplified candidate oncogene in oesophageal adenocarcinoma. Clinical cancer research : an official journal of the American Association for Cancer Research. 2012;18(7):1936-46.

32. Akagi T, Ito T, Kato M, Jin Z, Cheng Y, Kan T, et al. Chromosomal abnormalities and novel disease-related regions in progression from Barrett's esophagus to oesophageal adenocarcinoma. Int J Cancer. 2009;125(10):2349-59.

33. Pang C, LaLonde A, Godfrey TE, Que J, Sun J, Wu TT, et al. Bile salt receptor TGR5 is highly expressed in oesophageal adenocarcinoma and precancerous lesions with significantly worse overall survival and gender differences. Clinical and experimental gastroenterology. 2017;10:29-37.

34. Leicht DT, Kausar T, Wang Z, Ferrer-Torres D, Wang TD, Thomas DG, et al. TGM2: a cell surface marker in oesophageal adenocarcinomas. Journal of thoracic oncology : official publication of the International Association for the Study of Lung Cancer. 2014;9(6):872-81.

35. Zhou Z, Xia Y, Bandla S, Zakharov V, Wu S, Peters J, et al. Vitamin D receptor is highly expressed in precancerous lesions and oesophageal adenocarcinoma with significant sex difference. Hum Pathol. 2014;45(8):1744-51.

36. Boonstra JJ, van Marion R, Douben HJ, Lanchbury JS, Timms KM, Abkevich V, et al. Mapping of homozygous deletions in verified oesophageal adenocarcinoma cell lines and xenografts. Genes, chromosomes & cancer. 2012;51(3):272-82.

37. Soutto M, Peng D, Razvi M, Ruemmele P, Hartmann A, Roessner A, et al. Epigenetic and genetic silencing of CHFR in oesophageal adenocarcinomas. Cancer. 2010;116(17):4033-42. 78

38. Bandla S, Peters JH, Ruff D, Chen SM, Li CY, Song K, et al. Comparison of cancer- associated genetic abnormalities in columnar-lined esophagus tissues with and without goblet cells. Annals of surgery. 2014;260(1):72-80.

39. Zhang K, Wu X, Wang J, Lopez J, Zhou W, Yang L, et al. Circulating miRNA profile in oesophageal adenocarcinoma. American journal of cancer research. 2016;6(11):2713-21.

40. Goh XY, Rees JR, Paterson AL, Chin SF, Marioni JC, Save V, et al. Integrative analysis of array-comparative genomic hybridisation and matched gene expression profiling data reveals novel genes with prognostic significance in oesophageal adenocarcinoma. Gut. 2011;60(10):1317-26.

41. Silvers AL, Lin L, Bass AJ, Chen G, Wang Z, Thomas DG, et al. Decreased selenium- binding protein 1 in oesophageal adenocarcinoma results from posttranscriptional and epigenetic regulation and affects chemosensitivity. Clinical cancer research : an official journal of the American Association for Cancer Research. 2010;16(7):2009-21.

42. Obulkasim A, Ylstra B, van Essen HF, Benner C, Stenning S, Langley R, et al. Reduced genomic tumor heterogeneity after neoadjuvant chemotherapy is related to favorable outcome in patients with oesophageal adenocarcinoma. Oncotarget. 2016;7(28):44084-95.

43. Aichler M, Motschmann M, Jutting U, Luber B, Becker K, Ott K, et al. Epidermal growth factor receptor (EGFR) is an independent adverse prognostic factor in oesophageal adenocarcinoma patients treated with cisplatin-based neoadjuvant chemotherapy. Oncotarget. 2014;5(16):6620-32.

44. Xu E, Sun W, Gu J, Chow WH, Ajani JA, Wu X. Association of mitochondrial DNA copy number in peripheral blood leukocytes with risk of oesophageal adenocarcinoma. Carcinogenesis. 2013;34(11):2521-4.

45. Lee S, Han MJ, Lee KS, Back SC, Hwang D, Kim HY, et al. Frequent occurrence of mitochondrial DNA mutations in Barrett's metaplasia without the presence of dysplasia. PloS one. 2012;7(5):e37571.

46. Geppert CI, Rummele P, Sarbia M, Langer R, Feith M, Morrison L, et al. Multi-colour FISH in oesophageal adenocarcinoma-predictors of prognosis independent of stage and grade. British journal of cancer. 2014;110(12):2985-95.

47. Ismail A, Bandla S, Reveiller M, Toia L, Zhou Z, Gooding WE, et al. Early G(1) cyclin- dependent kinases as prognostic markers and potential therapeutic targets in oesophageal adenocarcinoma. Clinical cancer research : an official journal of the American Association for Cancer Research. 2011;17(13):4513-22.

48. Bang YJ, Van Cutsem E, Feyereislova A, Chung HC, Shen L, Sawaki A, et al. Trastuzumab in combination with chemotherapy versus chemotherapy alone for treatment of HER2-positive advanced gastric or gastro-oesophageal junction cancer (ToGA): a phase 3, open-label, randomised controlled trial. Lancet (London, England). 2010;376(9742):687- 97.

49. Almhanna K, Meredith KL, Hoffe SE, Shridhar R, Coppola D. Targeting the human epidermal growth factor receptor 2 in oesophageal cancer. Cancer control : journal of the Moffitt Cancer Center. 2013;20(2):111-6. 79

50. Li X, Galipeau PC, Paulson TG, Sanchez CA, Arnaudo J, Liu K, et al. Temporal and spatial evolution of somatic chromosomal alterations: a case-cohort study of Barrett's esophagus. Cancer Prev Res (Phila). 2014;7(1):114-27.

51. Sykes PJ, Neoh SH, Brisco MJ, Hughes E, Condon J, Morley AA. Quantitation of targets for PCR by use of limiting dilution. Biotechniques. 1992;13(3):444-9.

52. Vogelstein B, Kinzler KW. Digital PCR. Proc Natl Acad Sci U S A. 1999;96(16):9236- 41.

53. Fan HC, Quake SR. Detection of aneuploidy with digital polymerase chain reaction. Anal Chem. 2007;79(19):7576-9.

54. Baker M. Digital PCR hits its stride. Nat Meth. 2012;9(6):541-4.

55. McDermott GP, Do D, Litterst CM, Maar D, Hindson CM, Steenblock ER, et al. Multiplexed target detection using DNA-binding dye chemistry in droplet digital PCR. Anal Chem. 2013;85(23):11619-27.

3.10 Real-time PCR (qPCR) (Aim 2, 3 and 4)

Total RNA is first reverse transcribed into cDNA using SuperscriptTM II First Strand

Synthesis Kit (Qiagen, Hilden, Germany) following manufacturer’s protocol. A primer set was designed to cover the entire coding sequence of GAEC1 with some overlapping for thorough mutation scanning using National Center for Biotechnology Information (NCBI) available tool. qPCR reaction was prepared in a total of 20µL final volume consisting of 10µL 2x SensifastTM

SYBR No-ROX mix (Bioline, London, UK), and the final concentration of 0.5µM of primer pair and 60ng of template. Gradient PCR was then performed, and the PCR product was ran on a 2% gel electrophoresis to ensure specificity of the PCR reaction. Once optimised, samples were run on the Quantistudio Flex 6 qPCR system (Thermo Fisher Scientific, Waltham,

Massachusetts, USA) followed by a melt curve analysis with the default system settings to ensure amplification specificity. Specific details of the protocol and primer pairs used in the present study are described in detail in Chapter 5.

3.11 Conventional PCR (Aim 2 and 3)

A primer set was designed to cover the entire coding sequence of GAEC1 with some overlapping for thorough mutation scanning using NCBI available tool. Gradient PCR was then performed, and the PCR product was ran on a 2% gel electrophoresis to ensure specificity of the PCR reaction. PCR was performed using the PrimeSTAR® Max DNA polymerase (TaKaRa

Bio Inc, Shiga Prefecture, Japan) with a final volume of 20µL. Specific details of the protocol and primer pairs used in the present study are described in detail in Chapter 5 and 6.

3.12 Purification of PCR products (Aim 2)

After each matched sample were amplified with primers that flanked the entire coding region of GAEC1, the PCR product was run through a 2% gel electrophoresis and its band was excised and cleaned up using the NucleoSpin® Gel and PCR Clean-up kit according to the manufacturer’s protocol. Specifically, 400μL of Buffer NTI was added to 200mg of an agarose gel containing the excised band and incubated for 10min at 50°C, vortexing at 2min intervals until completely dissolved. Next, 550μL of the sample was loaded onto the NucloSpin® Gel and PCR Clean-up Column and centrifuged at 11,000g for 30s. Flow-through was discarded and any remaining sample was loaded on the column and the same process was repeated. The column was then washed with 550μL Buffer NT3 and centrifuged at 11,000g for 30s. Flow- through was discarded and the washing step was repeated. Next, the column was centrifuged at 11,000g for 1min to remove any remaining Buffer NT3. Care is ensured that column and flow-through do not come in contact. Then, columns were left to incubate at 70°C for 5min before loading 15μL DNase RNase free water onto the membrane of the column for 3min and then centrifuging at 11,000g for 1min to elute DNA. Quality and concentration of DNA were measured with A260/A280 ratio using NanoDrop Spectrophotometer (Thermo Fisher Scientific,

Waltham, Massachusetts, United States). The extracts were stored at 4°C until used.

3.13 Sanger sequencing analysis (Aim 2)

Sanger sequence was used to confirm the mutations in GAEC1 sequence in colorectal cancer. Briefly, samples were amplified with a conventional PCR, purified and then sent to

Australian Genome Research Facility (AGRF) where the samples were subjected to Big Dye

Terminator (BDT) labelling, purification and sequencing on the AB 3730xl Capillary sequencer (Applied Biosystems, CA, USA). The quality of sequencing results was then checked to ensure that the sequence signal intensity ranged between 700 to 6000. Sequences

82 were then analysed using Chromas and Sequence Scanner to obtain the FASTA format and chromatogram. A detailed description can be found at Section 3.12.1.

STATEMENT OF CONTRIBUTION TO CO-AUTHORED PUBLISHED PAPER

Chapters 3 includes two co-authored papers. The bibliographic details of the second co-authored paper, including all authors, are:

Title: Targeted single gene mutation with Sanger sequencing in oesophageal Adenocarcinoma

Authors: Katherine Ting-Wei Lee, Vinod Gopalan, Alfred King-yin Lam

Name of the book: oesophageal Adenocarcinoma: Methods and Protocols, Methods in Molecular Biology, vol. 1756

Status: Published

Publication link: https://doi.org/10.1007/978-1-4939-7734-5_19

My contribution to the paper involved: I drafted the manuscript, designed, optimised and ran the entire workflow for sample preparation for Sanger sequencing. I processed the tissue samples and harvested the DNA samples. I was involved in the Sanger sequencing analysis and statistical analysis of this manuscript.

(Signed) ______(Date)___ 22/11/2018__ Katherine Ting-Wei Lee

(Countersigned) ______(Date)___ 22/11/2018__ Corresponding author of paper and principal supervisor: Prof. Alfred King-Yin Lam

(Countersigned) ______(Date)__ 22/11/2018___ Corresponding author of paper and principal supervisor: Dr. Vinod Gopalan

3.13.1 Detailed description of targeted single gene mutation analysis with Sanger sequencing

Targeted single gene mutation with Sanger sequencing in oesophageal Adenocarcinoma

Katherine T. W. Lee, Vinod Gopalan, Alfred K. Lam

3.13.1.1 Abstract

Esophageal adenocarcinoma is heterogeneous and studies have reviewed many important mutations that contribute to the pathogenesis of cancer. These discoveries have helped paved the way into identifying new gene markers or gene targets to develop a novel molecular directed therapy for better patient outcomes in oesophageal adenocarcinoma.

Despite the recent bloom in next-generation sequencing, Sanger sequencing still represents the gold standard method for the study of the driver genes in oesophageal adenocarcinoma. This chapter focuses on the sequencing techniques in the identification of single gene mutations.

Key Words: oesophageal adenocarcinoma, Sanger sequencing, targeted gene, mutation

3.13.1.2 Introduction

Throughout the last decade or so, a major effort has been placed into identifying novel mutations in oesophageal adenocarcinoma (EAC) in hopes of improving patient outcome.

Many researchers found that oesophageal adenocarcinoma is highly mutated and highly heterogeneous [1-3] with a huge amount of new tumour suppressor genes [2, 3], oncogenes [1], single nucleotide polymorphisms and copy number alterations [1, 4, 5]. These discoveries have helped in understanding the pathogenesis of oesophageal adenocarcinoma and identifying clinically relevant genomic alteration. As a result, a new avenue for molecular directed therapy is now within reach [5].

Genomic alteration could potentially be involved in the pathogenesis of oesophageal adenocarcinoma (Table 3.2). Patients with V-Set and Immunoglobulin Domain Containing 10

Like (VSIG10L) mutations (Table 3.2) showed increased susceptibility to familial Barrett’s oesophagus or oesophageal adenocarcinoma [6]. Methylation of Eye absent 4 (EYA4) was most frequently observed in patients with oesophageal adenocarcinoma, followed by Barrett’s oesophagus and subsequently, the normal oesophageal tissue (Table 3.2). Similarly, methylation of cyclin-dependent kinase inhibitor 2A (CDKN2A) was observed in a stage- dependent in oesophageal adenocarcinoma [7], from none in normal epithelial tissue to 82% in oesophageal adenocarcinoma (Table 3.2). Furthermore, loss of heterozygosity (LOH) of

CDKN2A (Table 3.2) was observed in more than 50% of both oesophageal adenocarcinoma and premalignant lesions [7].

Meanwhile, Selenium Binding Protein 1 (SELENBP1) and SMAD family member 4

(SMAD4) mutation provided a clearer genetic distinction between oesophageal adenocarcinoma and high-grade dysplasia. SELENBP1 was reduced in oesophageal adenocarcinoma as compared to Barrett’s oesophagus [8] while SMAD4 mutation was only observed in oesophageal adenocarcinoma and not in high-grade dysplasia [2, 5].

Table 3.2: Gene mutations found in oesophageal adenocarcinoma discovered through sequencing

Gene Implications in oesophageal adenocarcinoma (EAC) Ref ABCB1 Heterozygote mutation of ATP Binding Cassette Subfamily B Member 1 [9] (ABCB1) in MFD1 oesophageal adenocarcinoma cell line on chromosome 7 at position 87179790 and 87150168 ARID1A Mutations of AT-rich interaction domain 1A (ARID1A) were detected in [10] 15% of oesophageal adenocarcinoma. Knockdown of AIRD1A enhanced cell growth, proliferation and invasion; and, was associated with nuclear accumulation of p53. CCND1 Amplification of cyclin D1 (CCND1) in approximately 13% of patients [5] with oesophageal adenocarcinoma. CDH11 Overexpression of cadherin 11 (CDH11) in oesophageal adenocarcinoma [11] compared to normal tissues. CDKN2A Substitution, gene homozygous deletion or truncation of cyclin- [5] dependent kinase inhibitor 2A (CDKN2A) gene were observed in approximately in 32% of patients with oesophageal adenocarcinoma. Methylation of CDKN2A was observed in 82% of oesophageal [7] adenocarcinoma cases, 30% in premalignant lesions but not in normal oesophageal squamous cell epithelial. Methylation of CDKN2A results in p16 inactivation. LOH at CDKN2A was found in 68% of oesophageal adenocarcinoma and 55% of premalignant lesions. CLDN3 Claudin 3 (CLDN3) was overexpressed in oesophageal adenocarcinoma [11] compared to normal tissues. E-cadherin Lower expression was associated with less preferable pathological [12] parameters and survival for patients with oesophageal adenocarcinoma. However, mutations were uncommon. EGFR Substitution, amplification or truncation of epidermal growth factor [5] receptor (EGFR) was observed in approximatelyn14% of patients with oesophageal adenocarcinoma. ERBB2 Amplification, substitution or truncation of receptor tyrosine-protein [5] kinase (ERBB2) was observed in approximately 23% of patients with oesophageal adenocarcinoma. EYA4 Eye absent 4 (EYA4) gene methylation was observed in 83% of [13] oesophageal adenocarcinoma and 77% of BE but only 3% in normal oesophageal tissue and represents a potential candidate marker FGF3/4/19 Amplification of fibroblast growth factor 3/4/19 (FGF3/4/19) was [5] observed in approximately 12% of patients with oesophageal adenocarcinoma. ICAM1 Intercellular Adhesion Molecule 1 (ICAM1) was overexpressed in [11] oesophageal adenocarcinoma compared to normal tissues. KRAS Amplification or substitution of V-Ki-ras2 Kirsten rat sarcoma viral [5] oncogene homolog (KRAS) in approximately 23% of patients with oesophageal adenocarcinoma. KRAS amplification due to breakage-fusion-bridge [1]

Table 3.2 continued Gene Implications in oesophageal adenocarcinoma (EAC) Ref MDM2 Mouse double minute 2 homolog (MDM2) amplification was observed [14] in 4% of patients with oesophageal adenocarcinoma. The MDM2 mutation was due to chromothripsis-derived double-minute [1] chromosome formation or breakage-fusion-bridge MYC MYC (c-Myc) amplification in approximately 16% of patients with [5] oesophageal adenocarcinoma Amplification due to chromothripsis-derived double-minute [1] chromosome formation RFC3 Amplification of replication factor C (Activator 1) 3 (RFC3) occurred [1] due to breakage-fusion-bridge SEMA5A Homozygote mutation of semaphorin 5A (SEMA5A) in MFD1 [9] oesophageal adenocarcinoma cell line on chromosome 5 at position 9190519 SELENBP1 Expression reduced when Barrett’s oesophagus progresses to [8] oesophageal adenocarcinoma. Overexpression of Selenium Binding Protein 1 (SELENBP1) in Flo-1 cells heightened apoptosis, augmented cellular senescence and boosted cisplatin cytotoxicity, indicating its possible significance as a predictor of response towards chemoprevention or chemosensitisation with certain types of selenium in oesophageal adenocarcinoma. SMAD4 SMAD family member 4 (SMAD4) mutated in 13% of patients with [2, 5] oesophageal adenocarcinoma. The antiproliferative response was re-established with reactivation of [15] transient transfection of SMAD4. SPG20 Spartin (SPG20) is mutated in 7% of oesophageal adenocarcinoma [3] TLR4 Toll-like receptor 4 (TLR4) is mutated in 6% of oesophageal [3] adenocarcinoma TP53 Tumour protein p53 (TP53) substitution or truncation in approximately [5] 83% of patients with oesophageal adenocarcinoma. TP53-mutant cells go through genome doubling, proceeded by [16] attainment of oncogenic amplification Mutation of TP53 was indicative of high-grade dysplasia in BE or [2, oesophageal adenocarcinoma occurrences, occurring during the 17] transition from low-grade to high-grade dysplasia. Patients without TP53 mutation did not develop high-grade dysplasia or [18] metaplasia. Homozygote mutant in MFD-1 cell line on chromosome 17 at position [9] 7577120 TP53 mutation was not found in human papillomavirus (HPV) -positive [19, oesophageal adenocarcinoma but found in some HPV-negative 20] oesophageal adenocarcinoma cases. 47% of patients with oesophageal adenocarcinoma oesophageal [21] adenocarcinoma had TP53 mutation and had reduced 5-year overall survival TP53 mutations and deletion do not occur commonly. However, [22] accumulation of the protein is fairly common, suggesting an alternative mechanism responsible for the common TP53 immunochemistry detection. 88

Table 3.2 continued Gene Implications in esophageal adenocarcinoma (EAC) Ref TP53 DNA sequencing detected more TP53 mutation than [23] immunohistochemical staining (75% by sequencing vs 58% by immunohistochemistry). TP53 mutation was not detected in the remaining 25% of oesophageal adenocarcinoma indicating that it is insufficient to function by itself as a single biomarker for BE monitoring but could play a role in a panel of biomarkers due to its high specificity. 33-73% of patients with oesophageal adenocarcinoma had TP53 [14, mutation while 0-4% of patients with BE had no TP53 mutation 18, suggesting that TP53 mutation presents much later in the progression 24- from BE to oesophageal adenocarcinoma and is not a good prognostic 26] marker for risk of progression from BE to oesophageal adenocarcinoma. In most cases, TP53 mutation was associated with poor tumour [18, differentiation, advanced tumour/nodes/metastasis (TNM) stages and 21, lymph node metastasis except for one study which claimed that the 25] mutation was found in well-differentiated carcinomas. VSIG10L V-Set and Immunoglobulin Domain Containing 10 Like (VSIG10L) is a [6] candidate for familial BE and subsequently oesophageal adenocarcinoma susceptibility gene.

Discovering gene mutations and understanding the link between these mutations and the pathogenesis of the disease can lead to the identification of potential biomarkers for oesophageal adenocarcinoma risk assessment or detection. Tumour protein p53 (TP53) mutation (Table 3.2) a typical example of good risk stratification candidate marker for determining the key point of therapeutic intervention in patients with oesophageal adenocarcinoma [2, 16-18]. Patients without TP53 mutations did not show high-grade dysplasia or metaplasia [18] while patients with TP53 mutations either had oesophageal adenocarcinoma or had transitioned from low-grade dysplasia to high-grade dysplasia Barrett’s oesophagus [2, 17].

However, TP53 mutations are not detected in approximately 25% of oesophageal adenocarcinoma [22, 23] suggesting its’ insufficiency as a sole biomarker but could work better as part of a panel of biomarkers [24]. For example, combined mutations of TP53, CDKN2A

[16, 27, 28] and APC [27] could provide a better prediction for oesophageal adenocarcinoma or cancer surveillance in patients with Barrett’s oesophagus (Table 3.3). Similarly,

89 overexpression of KRAS and suppression of SMAD4 were proposed to support the progression of oesophageal adenocarcinoma [1].

Besides being able to pinpoint the pathogenesis of oesophageal adenocarcinoma, gene mutations have also been associated with several pathological factors and assisted in the prediction of patient survival outcomes. For example, TP53 mutation (Table 3.2) was associated with poor tumour differentiation, tumour/nodes/metastasis (TNM) stages and lymph node metastasis [21, 25] and human papillomavirus (HPV) -negative oesophageal adenocarcinoma [19, 20]. Meanwhile, low expression of E-cadherin (Table 3.2) was associated with undesirable clinicopathological features and reduced survival outcome [12] although mutation of this gene was rare [12].

With such promising outcomes observed with discovering gene mutations, it is therefore essential to continue the search and identification of gene mutations in oesophageal.

This can be done through a variety of method; however, we will be focusing on the Sanger sequencing method, the gold standard method [39] for mutation detection in the clinical settings. Besides mutation detection, Sanger sequencing is also useful for several applications

(Table 3.4) including human leukocyte antigen (HLA) typing [30], multiple region sequencing

[31], next-generation sequencing (NGS) result validation [32, 33], plasmid sequences, inserts and/or mutations confirmation [34], single disease-causing genetic variant identification [35] and targeted genomic regions with huge sample size [36-38].

Table 3.3: Clusters of genes which were implicated with oesophageal carcinoma. Gene Implications in esophageal adenocarcinoma (EAC) Ref Clusters ELMO1, Dimerisation partners and intracellular mediators of Rho family GTPase, [3] DOCK2 RAC1 a gene previously indicated with other cancer. SMAD4, Amplification of oncogene KRAS and deletion of SMAD4 tumour suppressor [1] KRAS gene may facilitate oesophageal adenocarcinoma progression. TP53, Loss of CDKN2A followed by TP53 inactivation and aneuploidy led to [16] CDKN2A, oesophageal adenocarcinoma C progression from Barrett’s oesophagus [27] APC TP53, CDKN2A and APC combined mutations maybe be useful for dysplasia [28] and cancer surveillance in patients with Barrett’s oesophagus Combination of LOH and DNA abnormalities in 17p and 9p provides better oesophageal adenocarcinoma risk prediction than any single TP53, CDKN2A or DNA abnormalities alone. SKI, Deleted in oesophageal adenocarcinoma and could function as a novel [29] PRKZ biomarker for oesophageal adenocarcinoma.

Table 3.4: Current main purposes of Sanger sequencing Applications Example Ref Human leukocyte antigen Speeding up the process of identifying a match in [30] (HLA) typing unrelated donors for hematopoietic stem cell transplantation Multiple region sequencing Detection of HtrA Serine Peptidase 2 (HTRA2) and [31] anoctamin 3 (ANO3) mutation in patients with essential tremor Next-generation NGS confirmation of mutations found in patients with [32] sequencing (NGS) result osteogenesis imperfecta. validation Confirmation of potential causative variants for Leber [33] congenital amaurosis. Plasmid sequences, inserts Deletion of dystrosphin was confirmed with Sanger [34] and/or mutations sequencing in Duchenne muscular dystrophy patients confirmation Single disease-causing Detection of breast cancer type 2 susceptibility protein [35] genetic variant (BRCA2) mutation (c.1313delAAGA) in Moroccan identification population as first line screening for patients with breast and ovarian cancer. Targeted genomic regions Detection of multidrug- and extensively drug-resistant [36] with huge sample size tuberculosis in sputum specimen. Germline testing of BRCA1/2 in blood samples of [37] patients with high-grade serous ovarian cancer. Detection of telomerase reverse transcriptase (TERT) [38] in patients with thyroid carcinoma

3.13.1.3 Material

Prepare all reagents using ultrapure water and molecular grade chemicals. Ensure all plastic

wares and reagents are DNase-free. Store all reagents at room temperature unless indicated.

Try to reduce the amount of freeze-thaw for each reagent by dividing them in aliquots and

storing them. Appropriate personal protective equipment must be worn during experiments to

safeguard oneself against laboratory dangers. Follow the rules and regulations for waste

disposal.

3.13.1.3.1 Tissue collection and DNA extraction

Refer to Chapter 3, section 3.8.1.3.1.

3.13.1.3.2 PCR run and clean up

1. DNase-free water.

2. Micropipettes (10µL, 20µL, 200µL and 1000µL).

3. DNase-free filtered tips (10µL, 20µL, 200µL and 1000µL)

4. Electrophoresis gel tank system and agarose gel caster with well comb

5. 1x Tris-acetate-EDTA (TAE) buffer

6. 6x loading dye

7. 100bp DNA ladder

8. Agarose powder

9. Microcentrifuge tubes

10. Polymerase chain reaction (PCR) tubes

11. Primers.

12. PCR mastermix (preferably high fidelity mastermix).

13. PCR hood

14. 96-well PCR plate

15. Thermal cycler

16. Microcentrifuge machine

17. Gel imager/ transilluminator

18. Gel cutter

19. Gel clean up kit

20. Dry bath

21. Nanospectrophotometer

3.13.1.3.3 Sanger sequencing

1. Sanger sequencing reaction kit.

2. DNase-free water.

3. Micropipettes (10µL, 20µL, 200µL and 1000µL).

4. DNase-free filtered tips (10µL, 20µL, 200µL and 1000µL)

5. Microcentrifuge machine

6. Thermal cycler

7. Vortexer

8. Reagent reservoir

3.13.1.4 Method

3.13.1.4.1 Genomic DNA extraction from clinical samples

Refer to Chapter 3, section 3.8.1.4.1.

3.13.1.4.2 Targeted gene PCR

1. All samples are diluted with DNase-free water only.

2. Prepare all PCR materials within the PCR hood to reduce chances of contamination

(see NOTE 1).

3. Thaw all PCR required ingredients on ice.

4. Briefly tap and spin down each content to ensure homogeneity.

5. Prepare a master reaction mixture (see NOTE 2) according to the manufacturer’s

protocol with consideration of pipetting error (see NOTE 3).

6. A minimal of two technical replicates per sample (see NOTE 4) is required.

7. For example, for three biological samples, a total of eight master reaction mixtures

will be prepared (see NOTE 5).

8. Then for each reaction, aliquot a single reaction of the master reaction mixture into a

PCR tube and add the template (see NOTE 6) to it.

9. Non-template control must be used in all assays to ensure that there are no cross

contaminations (see NOTE 7).

10. Top up with DNase-free water to a final volume of 20µL.

11. Amplify the sample to endpoint with a thermal cycler according to the manufacturer’s

protocol for the mastermix (see NOTE 8).

12. Store the amplicons in -20°C freezer until further use.

3.13.1.4.3 PCR Clean Up

1. Prepare a 2% agarose gel in TAE buffer (see NOTE 9) inside a conical flask (see

NOTE 10).

CAUTION! Conical flask may be hot when removing from the microwave oven.

Ensure that protective heat gloves are worn when handling hot flask.

2. Add an intercalating dye such as ethidium bromide or Sybr Safe for band visualisation

later.

CAUTION! Intercalating dyes are mutagens. Ensure protective wear such as glove are

worn and disposed of properly after use.

3. Perform electrophoresis at 70V for 60minutes (see NOTE 11).

4. Visualise sample using a gel imager (Figure 3.6) or a UV transilluminator.

CAUTION! When dealing with a UV transilluminator, ensure that protective gear such

as an acrylic shield or a face shield is used to prevent unnecessary exposure to UV.

5. With the help of a gel cutter, slice out the band with the approximate band size and

transfer it into a microcentrifuge tube for PCR clean up.

6. The excised gel containing the band of interest is then purified using a PCR cleanup

kit according to the manufacturer’s protocol.

7. Briefly, dissolve the excised gel with the solution provided at 56℃ for 10minutes,

vortexing occasionally to help the dissolving process. Then transfer the mixture onto

the column provided and spin it at 12,000 rpm. Discard flow-through and wash the

column twice with washing buffer (see NOTE 12). Finally, add 30µL of DNase-free

water onto the membrane of the column (see NOTE 13) and elute the purified gel

product.

8. Measure the concentration, A260/A280 ratio and A260/A230 ratio of the resulting elution

using a nanospectrophotometer.

9. Store the elution in -20°C freezer until further use.

Figure 3.6: Example gel image after a 2% agarose gel ran at 70V for 60 minutes. Well 1 represents a 100bp ladder, well 2-6 represents sample PCR product and well 7 represents the none-template control.

3.13.1.4.4 Sanger sequencing reaction preparation

1. Thaw all required ingredients for Sanger sequencing preparation on ice away from the

light.

2. Briefly tap and spin down each content to ensure homogeneity (see NOTE 14).

3. Prepare run sequencing reaction according to the manufacturer’s protocol for both

forward reaction (with forward primer) and reverse reaction (with reverse primer). (see

NOTE 15)

IMPORTANT! Change pipette tips after each pipetting to prevent cross-contamination

(see NOTE 16).

4. Try to prevent the formation of bubbles as much as possible (see NOTE 17).

5. Place the sample into the thermal cycler and amplify the sample to endpoint according to

the manufacturer’s protocol (see NOTE 18).

6. Store the samples at 4°C until ready for the purification process.

3.13.1.4.5 Purifying the sequencing reaction

1. Thaw all required ingredients for purifying sequencing reaction on ice.

2. Briefly tap and spin down each content to ensure homogeneity.

3.13.1.4.6 Sanger sequencing analysis

1. Check the quality of sequencing results before analysing (see NOTE 19).

2. A few free software is available to analyse Sanger sequencing result. Some examples

include Chromas, Sequence Scanner or MEGA6.

3. Open the chromatogram file with one of this software.

4. A good chromatogram should appear as in Figure 3.7.

5. To align your sequence with the sequences available in National Center for Biotechnology

Information (NCBI) database, copy the FASTA format from the sequence obtained and

paste it in at the “Enter Query Sequence” column at the blastn section in Basic Local

Alignment Search Tool (BLAST)

(https://blast.ncbi.nlm.nih.gov/Blast.cgi?PROGRAM=blastn&PAGE_TYPE=BlastSearch

&LINK_LOC=blasthome).

6. Click on the “align two or more sequence” (Figure 3.8) if you want to compare your

sequence with a template sequence of your own rather NCBI database.

7. Fill up the rest of the form and click BLAST.

8. Mutations can be spotted from a difference in the identity between the query and subject

sequence and can be confirmed via the chromatogram (Figure 3.9Figure 3.9).

Figure 3.7: Example of a good quality chromatogram. Good base calling, high intensity and sharp peak of the chromatogram are an indication of good chromatogram result.

Figure 3.8: Overview of the blastn suite webpage.

Figure 3.9: Example of a single nucleotide mutation found through Sanger sequencing. In the tumour sequence, a change from a C>T nucleotide is observed.

3.13.1.5 Note

1. PCR hood should be exposed to ultraviolet lights for 25 minutes prior to using it.

2. Master reaction mixture is all the required ingredients to run a PCR reaction

excluding the template. For example, for 2x PCR mastermix protocol, prepare 10µL

2x mastermix, 1µL of 10µM gene forward primer and 1µL of 10µM gene reverse

primer.

3. Prepare an extra reaction for every eight reactions in consideration of pipetting error

and evaporation.

4. For every sample three technical replicates are recommended.

5. Three biological samples in duplicates (six reactions in total) with one non-template

control and one extra reaction.

6. It is recommended to use between 50ng to 500ng of the template when working with

genomic DNA.

7. Non-template control is prepared by replacing the template with DNase-free water.

8. Set the lid temperature to 105°C or closest temperature available on the PCR machine

used.

9. For every 100mL of 1x TAE buffer, dissolve 2g of agarose powder. Ensure that

agarose powder is completely dissolved. Once the gel is fully dissolved, quickly but

carefully transfer the solution onto the agarose gel cast. Ensure that all bubbles are

removed from the gel and leave to solidify.

10. Using conical flask helps prevent overflowing of agarose liquid when heating it in the

microwave oven.

11. Routinely check the electrophoresis run to ensure that the band does not run over the

gel.

12. Protocol may vary from manufacturer to manufacturer.

100

13. Usage of DNase-free water rather than the provided elution buffer can help improve

Sanger sequencing result. In addition, the total volume of DNase-free water used

during elution can be reduced if a higher concentration of elute is desired.

14. Optimal primer annealing temperature for Sanger sequencing protocol should be

around 60°C to 65°C.

15. Prepare an extra 10% of additional volume to compensate for pipetting error and

evaporation.

16. As a general guideline, for less than 48 samples, prepare them in PCR tubes. If more

than 48 samples, prepare them in 96-well PCR plate.

17. If bubbles are present at the top of the well, or in small quantities, it will not adversely

affect the reaction.

18. Amplicons that are less than 500bp can be run with a shorter extension time (for

example 2minutes).

19. Sequence signal intensity should range between 700 to 6000. Signal below 700 may

be affected by background and adjacent sequences while signal of over 6000 may

produce overload/ poor reads.

101

3.13.1.6 Reference

1. Nones K, Waddell N, Wayte N, Patch AM, Bailey P, Newell F, et al. Genomic catastrophes frequently arise in oesophageal adenocarcinoma and drive tumorigenesis. Nat Commun. 2014;5:5224.

2. Weaver JM, Ross-Innes CS, Shannon N, Lynch AG, Forshew T, Barbera M, et al. Ordering of mutations in preinvasive disease stages of oesophageal carcinogenesis. Nat Genet. 2014;46(8):837-43.

3. Dulak AM, Stojanov P, Peng S, Lawrence MS, Fox C, Stewart C, et al. Exome and whole- genome sequencing of oesophageal adenocarcinoma identifies recurrent driver events and mutational complexity. Nat Genet. 2013;45(5):478-86.

4. Contino G, Eldridge MD, Secrier M, Bower L, Fels Elliott R, Weaver J, et al. Whole- genome sequencing of nine oesophageal adenocarcinoma cell lines. F1000Res. 2016;5:1336.

5. Wang K, Johnson A, Ali SM, Klempner SJ, Bekaii-Saab T, Vacirca JL, et al. Comprehensive Genomic Profiling of Advanced oesophageal Squamous Cell Carcinomas and oesophageal Adenocarcinomas Reveals Similarities and Differences. Oncologist. 2015;20(10):1132-9.

6. Fecteau RE, Kong J, Kresak A, Brock W, Song Y, Fujioka H, et al. Association Between Germline Mutation in VSIG10L and Familial Barrett Neoplasia. JAMA oncology. 2016;2(10):1333-9.

7. Bian YS, Osterheld MC, Fontolliet C, Bosman FT, Benhattar J. p16 inactivation by methylation of the CDKN2A promoter occurs early during neoplastic progression in Barrett's esophagus. Gastroenterology. 2002;122(4):1113-21.

8. Silvers AL, Lin L, Bass AJ, Chen G, Wang Z, Thomas DG, et al. Decreased selenium- binding protein 1 in oesophageal adenocarcinoma results from posttranscriptional and epigenetic regulation and affects chemosensitivity. Clinical cancer research : an official journal of the American Association for Cancer Research. 2010;16(7):2009-21.

9. Garcia E, Hayden A, Birts C, Britton E, Cowie A, Pickard K, et al. Authentication and characterisation of a new oesophageal adenocarcinoma cell line: MFD-1. Sci Rep. 2016;6:32417.

10. Streppel MM, Lata S, DelaBastide M, Montgomery EA, Wang JS, Canto MI, et al. Next- generation sequencing of endoscopic biopsies identifies ARID1A as a tumor-suppressor gene in Barrett's esophagus. Oncogene. 2014;33(3):347-57.

11. Ferrer-Torres D, Nancarrow DJ, Kuick R, Thomas DG, Nadal E, Lin J, et al. Genomic similarity between gastroesophageal junction and oesophageal Barrett's adenocarcinomas. Oncotarget. 2016;7(34):54867-82.

12. Wijnhoven BP, de Both NJ, van Dekken H, Tilanus HW, Dinjens WN. E-cadherin gene mutations are rare in adenocarcinomas of the oesophagus. British journal of cancer. 1999;80(10):1652-7.

102

13. Zou H, Osborn NK, Harrington JJ, Klatt KK, Molina JR, Burgart LJ, et al. Frequent methylation of eyes absent 4 gene in Barrett's esophagus and oesophageal adenocarcinoma. Cancer Epidemiol Biomarkers Prev. 2005;14(4):830-4.

14. Taniere P, Martel-Planche G, Maurici D, Lombard-Bohas C, Scoazec JY, Montesano R, et al. Molecular and clinical differences between adenocarcinomas of the esophagus and of the gastric cardia. The American journal of pathology. 2001;158(1):33-40.

15. Onwuegbusi BA, Aitchison A, Chin SF, Kranjac T, Mills I, Huang Y, et al. Impaired transforming growth factor beta signalling in Barrett's carcinogenesis due to frequent SMAD4 inactivation. Gut. 2006;55(6):764-74.

16. Stachler MD, Taylor-Weiner A, Peng S, McKenna A, Agoston AT, Odze RD, et al. Paired exome analysis of Barrett's esophagus and adenocarcinoma. Nat Genet. 2015;47(9):1047- 55.

17. Bian YS, Osterheld MC, Bosman FT, Benhattar J, Fontolliet C. p53 gene mutation and protein accumulation during neoplastic progression in Barrett's esophagus. Modern pathology : an official journal of the United States and Canadian Academy of Pathology, Inc. 2001;14(5):397-403.

18. Dolan K, Walker SJ, Gosney J, Field JK, Sutton R. TP53 mutations in malignant and premalignant Barrett's esophagus. Diseases of the esophagus : official journal of the International Society for Diseases of the Esophagus. 2003;16(2):83-9.

19. Rajendra S, Wang B, Merrett N, Sharma P, Humphris J, Lee HC, et al. Genomic analysis of HPV-positive versus HPV-negative oesophageal adenocarcinoma identifies a differential mutational landscape. Journal of medical genetics. 2016;53(4):227-31.

20. Astori G, Merluzzi S, Arzese A, Brosolo P, de Pretis G, Maieron R, et al. Detection of human papillomavirus DNA and p53 gene mutations in oesophageal cancer samples and adjacent normal mucosa. Digestion. 2001;64(1):9-14.

21. Madani K, Zhao R, Lim HJ, Casson AG. Prognostic value of p53 mutations in oesophageal adenocarcinoma: final results of a 15-year prospective study. European journal of cardio- thoracic surgery : official journal of the European Association for Cardio-thoracic Surgery. 2010;37(6):1427-32.

22. Doak SH, Jenkins GJ, Parry EM, Griffiths AP, Shah V, Baxter JN, et al. Characterisation of p53 status at the gene, chromosomal and protein levels in oesophageal adenocarcinoma. British journal of cancer. 2003;89(9):1729-35.

23. Chung SM, Kao J, Hyjek E, Chen YT. p53 in oesophageal adenocarcinoma: a critical reassessment of mutation frequency and identification of 72Arg as the dominant allele. International journal of oncology. 2007;31(6):1351-5.

24. Novotna K, Trkova M, Pazdro A, Smejkal M, Soukupova A, Kodetova D, et al. TP53 gene mutations are rare in nondysplastic Barrett's esophagus. Digestive diseases and sciences. 2006;51(1):110-3.

25. Casson AG, Evans SC, Gillis A, Porter GA, Veugelers P, Darnton SJ, et al. Clinical implications of p53 tumor suppressor gene mutation and protein expression in oesophageal

103

adenocarcinomas: results of a ten-year prospective study. The Journal of thoracic and cardiovascular surgery. 2003;125(5):1121-31.

26. Agrawal N, Jiao Y, Bettegowda C, Hutfless SM, Wang Y, David S, et al. Comparative genomic analysis of oesophageal adenocarcinoma and squamous cell carcinoma. Cancer discovery. 2012;2(10):899-905.

27. Del Portillo A, Lagana SM, Yao Y, Uehara T, Jhala N, Ganguly T, et al. Evaluation of Mutational Testing of Preneoplastic Barrett's Mucosa by Next-Generation Sequencing of Formalin-Fixed, Paraffin-Embedded Endoscopic Samples for Detection of Concurrent Dysplasia and Adenocarcinoma in Barrett's Esophagus. J Mol Diagn. 2015;17(4):412-9.

28. Galipeau PC, Li X, Blount PL, Maley CC, Sanchez CA, Odze RD, et al. NSAIDs modulate CDKN2A, TP53, and DNA content risk for progression to oesophageal adenocarcinoma. PLoS medicine. 2007;4(2):e67.

29. Wang X, Li X, Cheng Y, Sun X, Sun X, Self S, et al. Copy number alterations detected by whole-exome and whole-genome sequencing of oesophageal adenocarcinoma. Human genomics. 2015;9:22.

30. Tu B, Cha N, Yang R, Ng J, Hurley CK. A one-step DNA sequencing strategy to HLA type hematopoietic stem cell donors at recruitment - rethinking typing strategies. Tissue Antigens. 2013;81(3):150-60.

31. Renaud M, Marcel C, Rudolf G, Schaeffer M, Lagha-Boukbiza O, Chanson JB, et al. A step toward essential tremor gene discovery: identification of extreme phenotype and screening of HTRA2 and ANO3. BMC Neurol. 2016;16(1):238.

32. Liu Y, Asan, Ma D, Lv F, Xu X, Wang J, et al. Gene mutation spectrum and genotype- phenotype correlation in a cohort of Chinese osteogenesis imperfecta patients revealed by targeted next generation sequencing. Osteoporos Int. 2017.

33. Jinda W, Taylor TD, Suzuki Y, Thongnoppakhun W, Limwongse C, Lertrit P, et al. Whole Exome Sequencing in Eight Thai Patients With Leber Congenital Amaurosis Reveals Mutations in the CTNNA1 and CYP4V2 Genes. Invest Ophthalmol Vis Sci. 2017;58(4):2413-20.

34. Wang DN, Wang ZQ, Yan L, He J, Lin MT, Chen WJ, et al. Clinical and mutational characteristics of Duchenne muscular dystrophy patients based on a comprehensive database in South China. Neuromuscul Disord. 2017;27(8):715-22.

35. Laarabi FZ, Ratbi I, Elalaoui SC, Mezzouar L, Doubaj Y, Bouguenouch L, et al. High frequency of the recurrent c.1310_1313delAAGA BRCA2 mutation in the North-East of Morocco and implication for hereditary breast-ovarian cancer prevention and control. BMC Res Notes. 2017;10(1):188.

36. Chen J, Peng P, Du Y, Ren Y, Chen L, Rao Y, et al. Early detection of multidrug- and pre- extensively drug-resistant tuberculosis from smear-positive sputum by direct sequencing. BMC Infect Dis. 2017;17(1):300.

37. Dougherty BA, Lai Z, Hodgson DR, Orr MCM, Hawryluk M, Sun J, et al. Biological and clinical evidence for somatic mutations in BRCA1 and BRCA2 as predictive markers for

104

olaparib response in high-grade serous ovarian cancers in the maintenance setting. Oncotarget. 2017;8(27):43653-61.

38. Tiedje V, Ting S, Herold T, Synoracki S, Latteyer S, Moeller LC, et al. NGS based identification of mutational hotspots for targeted therapy in anaplastic thyroid carcinoma. Oncotarget. 2017;8(26):42613-20.

39. Lee SH, Zhou S, Zhou T, Hong G. Sanger Sequencing for BRCA1 c.68_69del, BRCA1 c.5266dup and BRCA2 c.5946del Mutation Screen on Pap Smear Cytology Samples. Int J Mol Sci. 2016;17(2):229.

105

3.14 Immunofluorescence (Aim 2)

Antigen retrieval solution was preheated for 3min at 30% microwave power. FFPE slides that were prepared according to Chapter 3.4 were then placed into the antigen retrieval solution and further heated at 10% power for 3min with the microwave. Slides were then left to cool for 30mins before washing under running tap water for 5min. After that, slides were washed with PBS for 5mins and then incubated with peroxidase blocking reagent for 10mins.

Next, slides were then washed again with PBS for 5mins before blocking with 5% bovine serum albumin (BSA) for 1hr at room temperature and then incubating with GAEC1 primary antibody overnight at 4°C. The next day, slides were washed with PBS 3 times, 5mins each and then incubated with FITC secondary antibody at room temperature for 2hrs. The slides were then washed with PBS 3 times, 5 mins each before adding mounting media containing

4,6-diamidino-2-phenylindole (DAPI) to cover the slides. Specific details of the protocol and antibodies used in the present study are described in detail in Chapter 5.

3.15 Protein extraction and quantification (Aim 2, 3 and 4)

3.15.1 Protein extraction from cells

A maximum of 1x108 cells were harvested, re-suspended in ice cold PBS, and pelleted at 4,000rpm for 5min. The cells were then lysed with 1mL NP40 lysis buffer (Thermo Fisher,

Waltham, Massachusetts, United States) supplemented with 2% of protease inhibitor by leaving for it to stand on ice for 30min, vortexing every 5min. Later, the mixture was centrifuged at 14,000rpm for 20min at 4°C to pellet the cell debris. The supernatant containing the proteins were then transferred into a new micro-centrifuge tube. Concentration of the protein was then determined using Pierce BCA protein assay kit (Thermo Fisher, Waltham,

Massachusetts, United States) according to the manufacturer’s protocol. Proteins were stored at -80°C freezer until further use.

106

3.15.2 Protein extraction from tissue sample

A total of 5mg of tissues were sliced with a sterilise blade on ice and placed in a round bottom micro-centrifuge tube containing liquid NO2 to be snap frozen. A volume of 200µL of lysis NP40 lysis buffer supplemented with 2% of protease inhibitor was then added to the frozen tissue and homogenized with a homogenizer. The blade was rinsed twice 300µL of lysis buffer cocktail twice. The concoction was then vortexed and left to incubate with agitation at

120rpm, 4°C for 2hrs. After the incubation, the mixture was centrifuged at 12,000rpm for

20min at 4°C to pellet the cell debris. Supernatant containing the proteins were then transferred into a new microcentrifuge tube. Concentration of the protein was then determined using Pierce

BCA protein assay kit according to the manufacturer’s protocol. Proteins were stored at -80°C freezer until further use.

3.16 Western blot analysis (Aim 2, 3 and 4)

Total protein (30µg) was prepared in 12µL of 1x LB (lithium borate) buffer and cooked at 70°C for 10min. Proteins were then loaded on the 4-15% Mini-PROTEAN® TGX™ Precast

Gels (Bio-RadLaboratories, Hercules, CA) and ran at 72V for 120mins. After the run, the protein was transferred onto a polyvinylidene fluoride (PVDF) membrane with the Trans-blot®

Turbo™ system (Bio-Rad Laboratories, Hercules, CA) at 25V, 2.5A for 7 min. The membrane was then blocked in 5% non-fat milk for 1hr at room temperature. After that, the membrane was washed with PBST (Appendix D) twice, 5min each before overnight incubation with the relevant primary antibody at relevant dilution ratio at 4°C. The next day, the membrane was washed with PBST 6 times, 5min each. Then, the membrane was incubated in the appropriate secondary antibody at the relevant dilution ratio for 1hr at room temperature. The membrane was washed with PBST 6 times, 5min each before incubating with SuperSignalTM West Pico

PLUS Chemiluminescent Substrate (Thermo Fisher, Waltham, Massachusetts, United States)

107 for 5min and then visualising with the ChemiDoc MP TM Imaging System (Bio-Rad

Laboratories, Hercules, CA).

If membrane was reused, the membrane was first stripped with either a harsh stripping buffer (Appendix D) or a mild stripping buffer (Appendix D) before repeating the blocking, primary antibody incubation, secondary incubation and visualising with chemiluminescent substrate again. Briefly, for mild stripping, the membrane was incubated in mild stripping buffer at room temperature for 10min twice before washing another 4 times with PBST, 5min each.

3.17 Cloning of plasmid into bacteria (Aim 2, 3 and 4)

Bacteria transformation was performed using the NEB 5-alpha competent E. coli (High

Efficiency) kit according to the manufacturer’s protocol. Briefly, Immediately after NEB 5- alpha competent E. coli cells were thawed on ice, cells were mixed gently and 50µL cells were transferred into a microcentrifuge tube on ice. Plasmid DNA of 100ng was added into the mixture and mixed carefully by flicking the tube 5 times and left to incubate on ice for 30min.

Later, cells were heat shocked at 42℃ for exactly 30s before placing on ice for another 5min.

A volume of 950µL super optimal broth (SOB) with catabolite repression (SOC) media

(Appendix E) was then added to the mixture at room temperature and left to incubate for 60min at 37℃ with 250rpm shaking. Once done, mixed cells were diluted 10 times and 100µL was added into the selection plate containing selective plate and incubated overnight at 37℃.

Next day, a starter culture is prepared by transferring a single colony from the selective plate into 5mL of LB media with the appropriate selective antibiotic and incubated at 37℃ for

8hr with 300rpm shaking. After the incubation, 200µL of the starter culture is added to 100mL of LB broth containing appropriate selective antibiotic and incubated for 16h shaking at

300rpm at 37°C.

108

Following from that, bacterial cells are pelleted at 6000g got 15min at 4°C. Plasmids from the bacterial cells are then harvested using the EndoFree Plasmid Maxi Kit (Qiagen,

Hilden, Germany). Briefly, the bacterial pellet is re-suspended in 10mL Buffer P1 containing

LyseBlue. The mixture is vortexed until no cell clumps remain before adding 10mL of Buffer

P2, inverted six times and incubated at room temperature for 5min. A volume of 10mL Buffer

P3 is added to the mixture, inverted six times, and poured into the barrel of the QIAfilter

Cartridge to incubate for 10min at room temperature. After the incubation, the cap from the

QIAfilter Cartridge outlet nozzle was removed to allow the lysate to filter through. Then,

2.5mL Buffer ER is added to the filtered lysate, inverted 10 times, and incubated on ice for

30min.

Meanwhile, QIAGEN-tip 500 is equilibrated using 10mL of Buffer QBT. Filtered lysates then flow through the QIAGEN-tip via gravitational flow. QIAGEN-tip is washed twice with Buffer QC, 30mL each before eluting the DNA with 15mL of Buffer QN. DNA is then precipitated out by adding 10.5mL of isopropanol and centrifuging at 15,000g for 30min at

4°C. Then 5mL of 70% ethanol is added to wash the pellet before centrifuging at 15,000g for another 10min. Pellet is air dried for 10mins and re-dissolved in 300µL of endotoxin-free

Buffer TE. Specific details of the protocol and antibiotic used in the present study are described in detail in Chapter 5 and 6.

3.18 Plasmid transfection (Aim 2, 3 and 4)

A genetically modified organism (GMO) project approval was obtained from Griffith

University Institutional Biosafety Committee (NLRD/006/15). Cells (SW48, SW480 and FHC) were transfected using jetPRIME® Transfection Reagent (Polyplus-transfection® SA,

Strasbourg, France).

109

For a 6-well plate format, a total of 4.5x105 cells were plated into a 6-well plate for 24h in 2mL of media to achieve 70% confluency. To prepare the transfection cocktail, 200µL of jetPRIME buffer containing 2µg of plasmid, is vortexed for 5s, and short spun. A total of 4µL jetPRIME® transfection reagent is then added into the mixture, vortexed for 10s, and short spun again. The transfection cocktail is incubated at room temperature for 10min before adding

200µL into each well dropwise.

For a 96-well plate format, a total of 2x104 cells are plated into a 96-well plate for 24h in 0.1mL media. A total of 100ng of the plasmid in 5µL of jetPRIME buffer and 0.3µL jetPRIME® transfection reagent was used instead for each well.

Four hours later, the media containing transfection cocktail was removed and replenished with fresh media. Cells were harvested the next day for transfection efficiency confirmation and further analysis. Specific details of the protocol and plasmid used in the present study are described in detail in Chapter 4, 5 and 6.

3.19 Cell proliferation assay (MTT) (Aim 3 and 4)

After transfection, cells were left to incubate for 72h. Subsequently, 10µL of 12mM

MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) solution is added into each well and incubated at 37°C for another 4h. After the incubation, 90µL of media was removed and replaced with 100µL of DMSO (dimethyl sulfoxide) and incubated for another

10min at 37°C. The plate was agitated to mix and then read at 540nm wavelength with a microplate reader. Specific details of the protocol and plasmid used in the present study are described in detail in Chapter 5 and 6.

110

3.20 Wound healing assay (Aim 3 and 4)

A scratch was gently made with a P-200 pipette tip on transfected cells from Section

3.17 and media was replenished. The plate is visualised under an inverted microscope and an image is taken as Day 0. Media is replaced every 24h and a fresh is taken thereafter for four days. Wound closure rate was measured with ImageJ 1.52 software. Specific details of the protocol and plasmid used in the present study are described in detail in Chapter 5 and 6.

3.21 Clonogenic assay (Aim 3 and 4)

Transfected cells from Section 3.17 were harvested and 200 cells were plated onto a 6- well plate. Cells were left to incubate in a 5% CO2 supplemented incubator at 37°C while replenishing fresh media routinely until enough cells per colony were achieved (50 cells/ colony) [92, 93]. Next, cells were rinsed with PBS and fixed with acetic acid/methanol 1:7, vol/vol for 5min, before incubating with 0.5% crystal violet solution at room temperature for

2h. After the incubation, the cells were washed with tap water and left to air dry. The number of colonies formed obtained was used to calculate the survival fractions according to Franken and colleagues (2006). Specific details of the protocol and plasmid used in the present study are described in detail in Chapter 5 and 6.

3.22 Apoptotic assay (Aim 3 and 4)

Approximately a million cells were harvested and washed with PBS before re- suspending in Annexin V binding buffer. Cells were later on stained with 2.5µL of Alexa

Fluor® 350 (Invitrogen, Carlsbad, California) and 25ng/mL of PI solution. Cells were cooked at 58℃ for 5min to be used as positive control. A set of unstained cells, cells stained with only propidium iodide (PI), cells stained with only Alexa Fluor® 350, and cells stained with both

111

PI and Alexa Fluor® 350 were used to setup gating parameters. Cells were ran through a flow cytometry (BD Fortessa, BD Biosciences, Franklin Lakes, New Jersey) and analysed with

FlowJoTM 10. Specific details of the protocol and plasmid used in the present study are described in detail in Chapter 5 and 6.

3.23 Cell cycle analysis (Aim 3 and 4)

Approximately a million cells were harvested and fixed in ice cold 70% ethanol for 2h to 1 week. Later, ethanol suspended cells were pelleted at 2000rpm for 5min. Ethanol was completely removed and re-washed with 1mL PBS. Cells were again spun down at 2000rpm for 5min before treating with 500ng/mL RNase A and then staining with 50ng/mL of propidium iodide (PI) solution. Cells were flowed through a flow cytometry (BD Fortessa, BD

Biosciences, Franklin Lakes, New Jersey) and analysed with FlowJoTM 10. Specific details of the protocol and plasmid used in the present study are described in detail in Chapter 5 and 6.

3.24 Extracellular flux assay (Aim 3 and 4)

The bioenergetics of colon cancer cells in response to GAEC1 overexpression was determined using a Seahorse Bioscience XFp Extracellular Flux Analyzer (Searhose

Bioscience, Boston, MA). Approximately 10,000 cells were seeded overnight in a Seahorse

XFp Cell Culture Miniplate. After 24h incubation, cells were transfected with jetPRIME®

Transfection Reagent (Polyplus-transfection® SA, Strasbourg, France) in accordance with the

96-well format as previously written in section 3.17. Four hours later, the media containing transfection cocktail was removed and replenished with fresh media. The next day, media is replaced with Agilent Seahorse XF Base Medium supplemented with 1mM pyruvate, 2mM glutamine and 0.11mM glucose. Cells are left to incubate in a 37°C incubator without CO2 for approximately 40min before running on the analyser. The cells are then automatically stressed

112 with 1.0µM oligomycin, 1.0µM FCCP and 0.5µM rotenone/antimycin. All results were normalised according to the protein concentration from each well. Specific details of the protocol and plasmid used in the present study are described in detail in Chapter 5 and 6.

3.25 In silico analysis (Aim 3 and 4)

Sequence information of GAEC1 was obtained from Law and colleagues [13] and was used when required by a method as input. Due to lack of information on readily available bioinformatics tools such as Mutation Taster [94], results were compared to the databases -

UCSC (University of California Santa Cruz) Genome Browser Database [95] and International

Cancer Genome Consortium (ICGC) Data Portal [96]. Also, PROVEAN (Protein Variation

Effect Analyser) [97] was used to evaluate the sequences of the identified mutations within the coding region. The cut-off value used for PROVEAN was -2.5 for predicting the pathogenic non-pathogenic variants in this study.

3.26 Statistical analysis (Aim 1, 2, 3 and 4)

All statistical analysis was performed using the Statistical Package for Social Sciences for Windows (version 25.0 IBM SPSS Inc., New York, NY, USA) unless otherwise specifically mentioned. The significance level was taken at p<0.05.

113

Chapter 4: Developing an accurate method for the detection of copy number aberration and ploidy determination of GAEC1 in patients with colorectal cancer.

4.1 Preface

Copy number aberration (CNA) has been proven as a key indicator for predicting cancer

risk [98-101]. GAEC1 has shown remarkable findings in the molecular studies of colorectal

cancer [15, 59, 60]. Studies indicated that copy number changes of GAEC1 were significantly

associated with numerous clinicopathological characteristics such as tumour site [56] and

perforation [15, 59] and bio-aggressiveness [15, 59]. Additionally, changes of GAEC1 copy

number was also significantly different in patients with colorectal cancer and the general

population [57]. Therefore, identifying patients with differential copies of GAEC1 can guide

the decisions of clinicians in personalised therapy options and assess potential survival

outcomes for better clinician judgement.

To date, the gold standard method used for accurate quantification of copy number

aberration and ploidy determination is with the use of real-time polymerase chain (PCR) where

the relative rates of increasing fluorescence signal during the exponential amplification of a

target gene are obtained [102-104]. The efficiency and accuracy of these measurements are

subjected to various factors including amplification rate differences between the reference gene

and the target genes, template loading variation and system run-to-run variation [105]. To

address these pitfalls, a method capable of providing results in an absolute quantification

manner is needed.

In 1992, Sykes and colleagues found that absolute quantification of a target within a

sample could be known through limiting dilution and Poisson statistics [106]. In addition,

Weaver and colleagues found out that these systemic errors could be minimised by increasing

the number of technical replicates to achieve the desired precision [106]. Unfortunately, as

finer discrimination was required, the number of replicates needed to be increased drastically

114

[106]. These were the original two ideas that coined the development of digital PCR (dPCR).

This study aims to develop a duplex DNA binding dye chemistry based droplet digital PCR

(ddPCR) method and compare the two platforms (ddPCR and qPCR) in their performance for detecting GAEC1 copies. In addition, the cost and time effectiveness of using either platform is evaluated.

STATEMENT OF CONTRIBUTION TO CO-AUTHORED PUBLISHED PAPER

Chapters 4 includes a co-authored paper. The bibliographic details of the co-authored paper, including all authors, are:

Title: Droplet digital vs real-time PCRs for accurate quantification of gene copy number aberrations in colorectal cancers

Authors: Katherine Ting-Wei Lee, Anna Chruścik, Vinod Gopalan, Alfred King-yin Lam

Name of the journal: Scientific Report

Status: Submitted

Publication link: https://www.nature.com/srep/

My contribution to the paper involved: I drafted the manuscript, designed, optimised and ran the entire ddPCR for CNA detection and performed the entire workflow for sample preparation for Sanger sequencing. I processed the tissue samples and harvested the DNA samples. I planned and performed the optimisation for qPCR assays. I was involved in the Sanger sequencing analysis and statistical analysis of this manuscript.

(Signed) ______(Date)___ 22/11/2018__ Katherine Ting-Wei Lee

(Countersigned) ______(Date)___ 22/11/2018__ Corresponding author of paper and principal supervisor: Prof. Alfred King-Yin Lam

(Countersigned) ______(Date)__ 22/11/2018___ Corresponding author of paper and principal supervisor: Dr. Vinod Gopalan 115

4.2 Detailed description of Droplet digital vs real-time PCRs for accurate quantification of gene copy number aberrations in colorectal cancers

Droplet digital vs real-time PCRs for accurate quantification of gene copy number aberrations in colorectal cancers

Katherine Ting-Wei Lee1, Anna Chruścik1, Vinod Gopalan1*, Alfred King-yin Lam1*

1Cancer Molecular Pathology, School of Medicine, Griffith University, Gold Coast, Queensland, 4222, Australia

Address for correspondence:

Dr Vinod Gopalan School of Medicine and Medical Sciences, Griffith University, Gold Coast Campus, Gold Coast QLD 4222, Australia. E-mail: [email protected] Telephone +61 7 56780717; Fax +61 7 56780708

Professor Alfred K Lam, Head of Pathology, Griffith Medical School, Gold Coast Campus, Gold Coast QLD 4222, Australia. E-mail: [email protected] Telephone +61 7 56780718; Fax +61 7 56780303

116

This study aims to compare the efficiency between droplet digital PCR (ddPCR) and real-time quantitative PCR (qPCR) for accurate detection of copy number aberration (CNA).

Fresh-frozen colorectal cancer tissues were collected from 56 patients with colorectal adenocarcinomas. Genomic DNAs were extracted and CNAs were quantified with both ddPCR and qPCR assays. Our findings indicated that gene (GAEC1) copies were successfully detected in both qPCR and ddPCR assay. Both assays showed moderate correlation with a ddPCR assay having a smaller coefficient variance (0.84% to 4.17%) than qPCR assay (0.06% to 53.04%) indicating that ddPCR has better reproducibility compared to qPCR. Bland-Altman plot showed that the two assays were comparable 96.5% (n=54/56) of the time. Furthermore, the cost and run-time are greatly reduced with the usage of ddPCR with increasing sample reactions. To conclude, with the ability to increase throughput while retaining robustness and reducing cost per sample, the usage of multiplexing Evagreen based assays on ddPCR is a better option for use in clinical diagnostics.

117

4.2.1 Introduction

Copy number aberration (CNA) is a segment of genomic DNA, which varies in its copies when compared to a universal reference genome, either through an increased or reduced frequency of the DNA segment (copy number gains or losses) [1-5]. These variants may encompass changes in more than one gene [2]. As CNA predicts genomic instability and structural dynamism, it has a profound influence on medical genetics leading to the identification of disease-causing genomic alterations in many diseases [5], including cancer [6].

Accurate CNA detection is essential for diagnosis, treatment and predicting prognosis of patients with cancer [7]. For example, Xu and colleagues indicated the effectiveness of CNA in identifying patients at high risk of hepatocellular carcinoma among patients with chronic liver diseases [8]. Furthermore, understanding the difference in CNA between different subtype of cancer such as Breast cancer 1, early onset (BRCA-1) associated ovarian cancer and sporadic cancer can help in providing information for genetic counselling [6, 9].

Initially, there were restrictions in the capacity to define CNA at high resolution [10].

Throughout the last three decades, polymerase chain reaction (PCR) based methods have been the gold standard when it came to detection and quantification of nucleic acids [11-13].

However, these quantification methods were based on relative rates of increasing fluorescence signal during the exponential amplification of the target (gene of interest) and reference gene.

These measurements are subjected to influence by numerous factors including DNA concentration, analytical errors as well as the difference in amplification rates between the target and reference genes [14].

Digital PCR enables large replicates to be performed with just one standard sample reaction volume by partitioning the sample into nanolitre size and amplifying the product to the endpoint. By creating tens of thousands of partitions, the DNA molecule within the sample can be quantified in an absolute manner enabling small differences to be determined. Running

118

PCR to endpoint in digital PCR ensures that all concerns that may arise from analysis of exponential amplification can be removed [15]. Currently, there are two types of digital PCR based technology available in the market. One that uses a droplet-based technology, where samples are partitioned into droplets based on oil-based emulsion, while the other technology is based on a physical partitioning concept whereby samples are physically partitioned on a chip.

With the use of a reference gene, a successful duplex assay with DNA binding chemistry in ddPCR is done through manipulating the primer concentration and/or length of the amplicon, which ultimately gives rise to different strengths of the fluorescent signal. This allows the distinction between the different amplicon products within the same tube. In multiplexing assays which use basic DNA binding dye chemistry, a reduction in the otherwise incurred cost of probe-based assays would be imminent. Furthermore, multiplexing can help save sample-to-results time, the overall cost of sample processing and, in the long run, proof beneficial for a diagnostic setting.

We previously described using a duplex DNA binding chemistry (such as Sybr Green or EvaGreen) assay in ddPCR technology [16]. This allowed us to detect the absolute quantification of GAEC1 copies, a known oncogene in colorectal carcinomas, in addition to determining the ploidy of GAEC1 by using a diploid control reference gene. In this study, we aim to compare the performance of our established duplex DNA binding dye chemistry assay on the ddPCR platform to the standard qPCR assay for identifying GAEC1 copies in colorectal cancer samples. In addition, we would evaluate the cost and time effectiveness of using ddPCR over qPCR as a contributing factor to replacing qPCR with ddPCR as the go-to method for absolute quantification.

119

4.2.2 Results

4.2.2.1 Detection of GAEC1 DNA copies with qPCR assay

To identify GAEC1 copies in colorectal cancer samples using qPCR, a tenfold serial dilution of GAEC1 plasmid DNA was used to create a calibration curve for qPCR assay (Figure

4.1A). The standard curve generated had an efficiency of 95%, with a slope of -3.451 and exhibited good linearity of R2 = 0.996. Furthermore, melt curve analysis (Figure 4.1B) showed only a single peak for all dilutions, further confirming the specificity of our primer design and qPCR assay.

Figure 4.1: Quantification of GAEC1 plasmid DNA serial dilutions by qPCR. (A) Tenfold serial dilution of GAEC1 plasmid DNA standard curve generated using qPCR. The range of plasmid copies used was from 2,000,000 to 200 copies. The slope of the fitted line is -3.451, R2 = 0.996, with the good linearity of 94.883%. (B) Melt curve results produced from the qPCR standard curve assay. The melt curve figure indicates only one clean peak throughout all serial dilutions, indicating the standards were free of contamination and had specific amplification.

120

4.2.2.2 Detection of GAEC1 copies with ddPCR

Figure 4.2: 1D and 2D droplet plots of droplet clusters obtained with the QuantaSoft software version 1.7.4. (A) From the 1D plot, the lowest droplet clusters are the double negative droplet containing non-targets. The next two bands of increased fluorescence amplitude contain droplet positive for either one amplicon type (HBB target or GAEC1 target). The topmost highest fluorescence amplitude contains double positive droplets for both amplicons (HBB and GAEC1 targets) detected. (B) Upon running singleplex assay containing either HBB or GAEC1 primers, the two in between bands were identified. HBB cluster was at the lower cluster whereas GAEC1 was situated at the higher cluster of the two clusters. (C) The same can be observed with the 2D plot where the lowest amplitude for both channel 1 and channel 2 is the negative droplets, followed by either HBB target or GAEC1 target and the combined target at the farthest top right corner.

121

Performing a duplex reaction with ddPCR resulted in the appearance of four distinct droplet populations of independent fluorescence amplitude (Figure 4.2) that were distinguishable in both, 1D and 2D droplet plots of the QuantaSoft software version 1.7.4. The lowest droplet clusters are the non-target containing negative droplets, while the second lowest cluster was the HBB cluster, followed by the GAEC1 cluster and subsequently the final population of droplets contain both HBB and GAEC1 target sequence and consequently had the greatest fluorescence signal.

4.2.2.3 Diagnostic performance comparison between ddPCR and qPCR

In this study, a moderate correlation (R2 = 0.532; P ≤ 0.01) between ddPCR and qPCR assays were observed, with a linear regression slope (Figure 4.3) intercept of 3.287 (R2 = 0.283;

P ≤ 0.001). Both qPCR and ddPCR were able to detect GAEC1 copies in all the colorectal cancers (15,098.12, limit of agreement [LoA]: 2,378.66-32,574.9) as seen in the Bland-Altman analysis (Figure 4.4). To note, GAEC1 copies detection were not completely matched with those by qPCR. Both assays were in concordance in 96.5% (n=54/56) of the time.

The coefficient of variation (CV%) was calculated for each sample to identify reproducibility between runs. The inter-assay CV for the copy number detection revealed a large difference between the two assays (Figure 4.5). The ddPCR assay showed a smaller range of CV% (0.84% to 4.17%) in comparison to qPCR assay (0.06% to 53.04%), indicating a superior reproducibility of copy numbers detection with ddPCR when compared to qPCR [17,

18].

122

Figure 4.3: Correlation between qPCR and ddPCR GAEC1 copy number detection in patients with colorectal cancer. The linear graph showed a positively moderate correlation between qPCR and ddPCR assay for the detection of GAEC1 copies with a linear regression slope intercept of 3.287 (R2 = 0.283; P ≤0.001).

Figure 4.4: Bland-Altman plot to compare the measurements of two different techniques. Horizontal lines represent the mean difference between results gathered from the two techniques and the limits of agreement (LoA), defined by the upper and lower limits of the 95% CI. In the case of the 56 tumour patient samples tested for the GAEC1 aberration, the two methods correlate well with one another 96.5% (n=54/56) of the time.

123

Figure 4.5: Inter-assay CV% between qPCR and ddPCR assays. Histograms indicate the average copy of copy number in each sample. The lighter histogram indicates GAEC1 copies detected by qPCR assay while the darker histogram indicates GAEC1 copies detected by the ddPCR assay. Lines showed the trend of variation of CV for both qPCR and ddPCR assays with repeated tests of all 56 colorectal cancer samples. The lighter line indicates CV% detected by qPCR assay while the darker line indicates CV% detected by the ddPCR assay. The ddPCR assay (0.84% to 4.17%) is much more precise when compared to the qPCR assay (0.06% to 53.04%) for quantification of GAEC1 copies.

4.2.2.4 Cost-effectiveness of ddPCR

As seen in Table 4.1, this study has estimated the overall cost per reaction for running a qPCR assay is relatively cheap (approximately 1 Australian dollar), the cost is greatly increased with replicates and the inability to run duplex assays with SYBR Green dye on qPCR systems. Whereas, while the cost of ddPCR per reaction is higher (approximately 4 Australian dollar), due to the ability to run duplex assays and without the need to prepare multiple reactions for technical replicates (because the approximate 20,000nL size droplets act as technical replicates), the overall cost for an entire run becomes cheaper. The final cost calculated including replicates was approximately $4 for ddPCR and $8 for qPCR (Table 4.1).

124

Table 4.1: Cost comparison between qPCR assays and ddPCR assays for CNA detection using DNA binding dye chemistry. no. of reactions minimal Cost Cost/ required for CNA data replicates/ inclusive of reaction* to be generated** target replicates* qPCR assay for CNA $1.3 2 3 $7.8 ddPCR assay for CNA $4.2 1 1 $4.2 * Cost is just an estimated price in Australian Dollars. ** The number of reactions required to run both the control and target gene assays.

4.2.2.5 Sample to output timeliness

Figure 4.6: (A) ddPCR workflow diagram and (B) qPCR workflow diagram for CNA detection. The total run time for ddPCR for 56 samples from preparation to droplet generation, PCR run and droplet reading took approximately 4 hours with hands on time of approximately 1 hour and 20 minutes for sample preparation. The total run time for qPCR on the other hand took approximately 10 hours with 4 hours hands on time for sample preparation. The time needed for ddPCR to run the entire assay was equivalent to the amount of time needed just for sample preparation in qPCR. These timing were excluding workflow optimisation or standard curve preparation and run time.

125

As noted in Figure 6, qPCR required fewer procedures to follow and setup per run cycle.

However, due to the increased need for replicates and the inability to run a duplex reaction, the overall runtime was less favourable when compared to the ddPCR system. The current study took ~ 4 hours to run 56 samples on a ddPCR system while it took ~10 hours to run 56 samples on a qPCR system. Additionally, the need to generate a quality standard curve with good PCR efficiency and R2 value as suggested by the MIQE (Minimum Information for Publication of

Quantitative Real-Time PCR Experiments) guideline for qPCR [19] can be both time consuming and expensive.

4.2.3 Discussion

This study revealed that ddPCR assay produced more consistent results than those from the qPCR assay in the quantification of GAEC1 CNA and developed a newer, more reliable and cost-effective method to detect GAEC1 copies in the clinical setting. This can help reduce the sample-to-result time and save overall expenditure. Moreover, the data produced from a duplex assay is quantitative with precise CNA integer, rather than expressed as amplification or deletion [20-22].

With ddPCR, copy numbers can be obtained without the need of generating a standard curve [17, 18, 23, 24]. This is because, unlike qPCR, ddPCR digitises its data to achieve absolute quantification. This is done by fractioning the PCR reaction into nanoliter-sized reactions. Targets are individually assessed at single molecular sensitivity by running PCR to end-point and then calculating the fluorescent signals as a binomial event (positive or negative).

With the incorporation of the Poisson algorithm, digital PCR can calculate the absolute number of target copies in the original sample from the ratio of positive total partitions [25, 26].

This study indicates a moderate correlation between ddPCR and qPCR assay. However, contrary to our findings, the results obtained by Tang and colleagues (2016) showed a strong

126 correlation between the two assays when compared on a Bland-Altman plot [23]. This suggests that the direct comparison of the two assays without the use of a standard curve in both assays for normalisation of samples can result in greater variation between the obtained results. This shows that, if need be, a qPCR assay setup can be easily transferred onto the ddPCR platform with a simple normalisation with readily available standards. However, if a new assay is to be set up, running it directly on a ddPCR platform may produce more reproducible and accurate result.

The smaller range of CV % (0.84% to 4.17%) obtained from ddPCR in comparison to qPCR assay (0.06% to 53.04%) indicates its superior reproducibility of copy numbers detection

[17, 18]. Reason being, ddPCR does not rely on an external standard curve for the quantification of GAEC1 copies, whereas the use of a calibration curve is essential in qPCR assays. This presents a major setback for qPCR as the quality of the data produced is strongly dependent on the accuracy of the standards. The calibration curve is usually generated from a serial dilution of a standard of known concentration. Unfortunately, these standards are often expensive and not readily available [17]. As in this study, GAEC1 is not commercially available. Therefore, the process of generating a standard curve becomes laborious and time- consuming.

In addition, the accuracy of a standard curve relies heavily on many factors including primer efficiencies [18], possible contaminants affecting the DNA polymerase activity [27], external calibration of the thermocycler itself [17, 18], DNA calibrator batch variations [18],

PCR inhibitors [17, 18], poor sample quality [17], variable threshold set point [18] and estimated Cq value cut-off point [18]. With the use of ddPCR, these uncertainties can be removed as assays are run to end-point and the results are digitalised [25, 26] as previously discussed. The ddPCR platform allows researchers to bypass these heavily varying factors and therefore, provides for more accurate and reproducible output.

127

This finding broadly supports the work of other studies in this area, which concluded the better reproducibility and precision of ddPCR over qPCR technology. A study found that when detecting the presence of foodborne pathogens, ddPCR assay showed superiority in all aspects of their analysis including sensitivity, reproducibility and time efficiency [28]. Tang and colleagues (2016) compared qPCR and ddPCR assay efficiency in detecting Hepatitis B virus in clinical samples [23]. Their research found that majority of the Hepatitis B virus detection level within the clinical and non-clinical samples fall within the ddPCR range in comparison to qPCR, with significantly higher reproducibility.

In addition, this study successfully demonstrated the reduced cost and time required when using ddPCR over qPCR. With less usage of plastic wares such as 96-well PCR plates, micropipette tips and etcetera, a greener/cost-effective methodology is established with ddPCR.

Taken together, ddPCR proves to be advantageous in terms of throughput and cost in comparison to qPCR for routine settings [29] such as in clinical diagnostics.

Both ddPCR and qPCR assays successfully detected GAEC1 copies in all samples and were moderately correlated perhaps due to the lack of normalisation of samples with a standard curve in ddPCR. This shows that qPCR assay setups can be easily transferred onto a ddPCR platform with a simple normalisation using readily available standards. However, looking at the long-term run cost and timeliness, if a new assay were to be set up, running it straight on a ddPCR platform would be the better option.

The ddPCR platform allows for multiplexing DNA binding dye chemistry based assays, which cannot be done in the case of qPCR. By duplexing a copy number assay with a diploid control reference gene, the ploidy of the target gene (GAEC1, in this case) can be found.

Furthermore, by detecting both target and reference genes in the same reaction tube, sample preparation to sample preparation variation is removed. Duplexing assays also allow more efficient sample and assay to test ratio, saving overall cost and runtime. Moreover, with less

128 inherent variance from sample to sample topped with the ability to detect complex sample at low concentration eases the need of a large amount of sample collection thus improving overall patient care and cost per sample. The ability to multiplex a DNA binding dye chemistry while retaining the robustness of the test is a strong indication to implement ddPCR in clinical settings.

4.2.4 Methods

4.2.4.1 Recruitment of matched tissue sample

Cancer tissues were collected from patients with colorectal cancer across 2012-2014 from hospitals in Queensland, Australia with their consent. Ethical approval for the use of these tissues was attained from the Griffith University Human Research Ethics Committee (GU

Ref No. MSC/17/10/HREC). All methods were carried out by the relevant guidelines and regulations required by the ethics committee. Briefly, these samples were flash frozen in liquid nitrogen prior to cryotomy at 10 µM sections for pathological confirmation. Tissues samples were excluded from the studies if they lacked appropriate cancer proportion or if patients were lost to clinical follow-up. After reviewing, 56 patients (28 females; 28 males) with colorectal adenocarcinomas were included for the analysis.

4.2.4.2 DNA extraction

Genomic DNA was extracted from the collected colorectal tissue samples using Tissue

& Blood extraction kit (Qiagen, Hilden, Germany) following the manufacturer’s guideline.

Quality and quantity of DNA extracted were then obtained through measurement of A260/A280 ratio and A260/A230 ratio using NanoDrop Spectrophotometer (Thermo Fisher Scientific,

Waltham, MA, USA). DNA samples were stored at -20°C prior to use.

129

4.2.4.3 Droplet digital PCR (ddPCR) assay setup

Even though many published papers compared ddPCR with qPCR by generating a standard curve of reference material [17, 18, 23, 24], this study did not use a standard curve for the ddPCR assay. This is because ddPCR does not require an external standard curve to obtain the copy values [30]. Rather, a reference material as suggested by the digital MIQE (Minimum

Information for Publication of Quantitative Real-Time PCR Experiments) guidelines were used instead [31]. The ddPCR assay setup was prepared according to our previous protocol [16]. A stringent analysis performed by Pinheiro and colleagues (2012) identified that both simplex

(single reaction assays) and duplex assays produced results that coincide with one another [30].

As we have previously optimised a duplex assay for GAEC1 copies detection [16], we implemented the same assay in this study.

The reaction mixture was prepared with 10µL of 2x ddPCR EvaGreen supermix, with a final concentration of 100nM GAEC1 primer set and 50nM HBB primer set, 1µL of 2U/µL

HINDIII restriction enzyme, 1µL of the template and topped up with DNase and RNase free water to a final volume of 20µL. The reaction mix was partitioned to a ~ 20,000 nanoliter-sized droplets with the QX200™ AutoDG™ (Bio-Rad Laboratories). Cycling conditions were set to 5min initial denaturation at 95°C; followed by 40 cycles of 30s denaturation at 95°C, 1 minute annealing at 62°C all at a ramp rate of 2°C/s; signal stabilisation at 4°C for 5 minutes, then at 90°C for 5 minutes; and, an infinite holding at 12°C. PCR products were then transferred onto a QX200 Droplet Reader (Bio-Rad Laboratories) where target DNA molecules are calculated in an absolute manner by singulating each droplet and detecting the fluorescence amplitude in it. The analysis was then performed by QuantaSoft software version 1.7.4 (Bio-

Rad Laboratories) and QuantaSoftTM Analysis Pro Software 1.0.596 (Bio-Rad Laboratories) which incorporated Poisson statistics to call copy number of target sequence without the need of a standard curve [30].

130

4.2.4.4 Preparation of cloned GAEC1 plasmid DNA standards

GAEC1 plasmid DNA was kindly provided by Johnny Tang and has been used in other studies 20. This plasmid was cloned into NEB 5-alpha competent E. coli (New England Biolabs,

Massachusetts, US) according to the manufacturer’s protocol. Selective media containing

100µg/mL ampicillin Luria-Bertani (LB) was used to ensure only appropriate selection of cells containing GAEC1 survived. The next day, a single colony from the bacterial growth was transferred into 5mL of selective media as a selective starter culture. They were incubated for

8 hours at 37°C, shaking at 300 rpm. Then, 500µL of the starter culture was aliquot into 250mL of fresh selective media and left to incubate for 15 hours at 37°C, shaking at 250rpm. Following on, cells were spun down at 6000xg for 15 minutes at 4°C and harvested with Endofree Plasmid

Maxi Kit (Qiagen, Hilden, Germany) according to the manufacturer's protocol.

Quality and quantity of plasmid DNA extracted were then obtained through measurement of A260/A280 ratio and A260/A230 ratio using NanoDrop Spectrophotometer

(Thermo Fisher Scientific, Waltham, MA, USA). Plasmid DNAs were then sequenced using the plasmid vector primers specifically designed for pcDNA3.1 including T7 promoter and

BGH reverse promotors (Supplementary Table 4.1) to confirm successful plasmid cloning as suggested by the manufacturer of pcDNA3.1 (Invitrogen, California, USA).

GAEC1 plasmid DNA standards were constructed with a tenfold dilution series ranging from 2,000,000 copies to 200 copies. The copy numbers of GAEC1 plasmid DNA (5884bp) are calculated based on the following formula [32] where n is the number of base pairs, m is the mass of the DNA and NA is the Avogadro’s constant (6.02 x 1023 bp mol-1), and M is the average molecular weight of a base pair (660 g mol-1).

131

n x M m= NA

n x 660 g mol-1 m= (6.02 x 1023 bp mol-1)

m=n x 1.096 x 10-21g bp-1

Supplementary Table 4.1: Primer design for GAEC1 plasmid Sanger sequencing Target Genes Primers T7 promoter 5’- TAATACGACTCACTATAGGG-3’ BGH reverse primer 5’- TAGAAGGCACAGTCGAGG-3’

4.2.4.5 qPCR assay setup

The qPCR assay was carried out using the Quantstudio Flex 6 (Applied Biosystems,

California, US) in a final volume of 10µL reaction containing 5µL of SensiFAST SYBR Green

No-Rox Master Mix (Bioline, London, UK), 0.4µL of each 10µM GAEC1 forward and reverse primer (Supplementary Table 4.2), 30ng DNA template topped up with DNase free water to

10µL volume. The thermal cycling conditions were set to initial denaturation at 95°C for 3 minutes followed by 40 cycles of denaturation at 95°C for 5 seconds, annealing at 62°C for 10 seconds and extension at 72°C for 20 seconds. Standard curves constructed with GAEC1 plasmid DNA were included in every qPCR to obtain quantitative results rather than raw Cq values.

A non-template control (NTC) was used in every run and samples were measured in duplicates. The output data was analysed using the QuantStudio Real-Time PCR Software

(Invitrogen).

132

Supplementary Table 4.2: Primer design for quantitative real-time PCR (qPCR) analysis and droplet digital PCR (ddPCR) analysis Target Genes Primers Amplicon size (GenBank number) (base pairs) GAEC1 5′-CAGGGAAGAAGCAAGTTCCA-3 62 (AC005088) 5′-CCATCTGACACAGAGAGTGC-3′ HBB 5′-GCCCAGTTTCTATTGGTCTCCT-3′ 82 (NC_000011) 5′-TGGATGAAGTTGGTGGTGAGG-3′

4.2.4.6 Data Analysis

All statistical analysis was carried out using IBM (New York, US) SPSS Statistics version 25 except those explicitly stated. Pearson’s correlation and linear regression were used to evaluate the relationship between qPCR and ddPCR assays. Bland-Altman plot (Figure 4.4) was used to determine the limits of agreement (LoA), while the coefficient of variance (CV)

(standard deviation/mean*100) was used to assess the level of variability between the replicates.

4.2.5 Acknowledgements

The authors would like to thank Dr Johnny Tang for the provision of GAEC1 plasmid. We also like to thank the NH Diagnostic laboratory for the processing of the tissue. The project is supported by the student scholarship from Griffith University and funding from Menzies

Health Institute Queensland.

4.2.6 Authors contribution

KTL prepared Figure 2, 6 and Table 1. AC prepared Figure 3. KTL and AC planned and performed the overall experiment, drafted the manuscript, prepared Figure 1, 4 and 5. AKL and VG supervised and participated in editing the manuscript. All authors read and approved the final manuscript.

133

4.2.7 Competing interest

The authors declare no competing interests.

4.2.8 References

1 Redon R, Ishikawa S, Fitch KR, Feuk L, Perry GH, Andrews TD, et al. Global variation in copy number in the human genome. Nature. 2006;444(7118):444-54. 2. Cook EH, Jr., Scherer SW. Copy-number variations associated with neuropsychiatric conditions. Nature. 2008;455(7215):919-23. 3. Park C, Kim JI, Hong SN, Jung HM, Kim TJ, Lee S, et al. A copy number variation in PKD1L2 is associated with colorectal cancer predisposition in korean population. Int J Cancer. 2017;140(1):86-94. 4. Lupski JR. Structural variation in the human genome. N Engl J Med. 2007;356(11):1169-71. 5. Coughlin CR, 2nd, Scharer GH, Shaikh TH. Clinical impact of copy number variation analysis using high-resolution microarray technologies: advantages, limitations and concerns. Genome Med. 2012;4(10):80. 6. Kuiper RP, Ligtenberg MJ, Hoogerbrugge N, Geurts van Kessel A. Germline copy number variation and cancer risk. Curr Opin Genet Dev. 2010;20(3):282-9. 7. Serrati S, De Summa S, Pilato B, Petriella D, Lacalamita R, Tommasi S, et al. Next- generation sequencing: advances and applications in cancer diagnosis. Onco Targets Ther. 2016;9:7355-65. 8. Xu H, Zhu X, Xu Z, Hu Y, Bo S, Xing T, et al. Non-invasive Analysis of Genomic Copy Number Variation in Patients with Hepatocellular Carcinoma by Next Generation DNA Sequencing. J Cancer. 2015;6(3):247-53. 9. Yoshihara K, Tajima A, Adachi S, Quan J, Sekine M, Kase H, et al. Germline copy number variations in BRCA1-associated ovarian cancer patients. Genes Chromosomes Cancer. 2011;50(3):167-77. 10. Sebat J, Lakshmi B, Troge J, Alexander J, Young J, Lundin P, et al. Large-scale copy number polymorphism in the human genome. Science. 2004;305(5683):525-8. 11. Heid CA, Stevens J, Livak KJ, Williams PM. Real time quantitative PCR. Genome Res. 1996;6(10):986-94. 12. Kubista M, Andrade JM, Bengtsson M, Forootan A, Jonak J, Lind K, et al. The real- time polymerase chain reaction. Mol Aspects Med. 2006;27(2-3):95-125. 13. Dobnik D, Stebih D, Blejec A, Morisset D, Zel J. Multiplex quantification of four DNA targets in one reaction with Bio-Rad droplet digital PCR system for GMO detection. Sci Rep. 2016;6:35451. 14. Karlen Y, McNair A, Perseguers S, Mazza C, Mermod N. Statistical significance of quantitative PCR. BMC Bioinformatics. 2007;8:131. 15. Hindson BJ, Ness KD, Masquelier DA, Belgrader P, Heredia NJ, Makarewicz AJ, et al. High-throughput droplet digital PCR system for absolute quantitation of DNA copy number. Analytical chemistry. 2011;83(22):8604-10. 16. Lee KT, Gopalan V, Islam F, Wahab R, Mamoori A, Lu CT, et al. GAEC1 mutations and copy number aberration is associated with biological aggressiveness of colorectal cancer. European journal of cell biology. 2018;97(3):230-41.

134

17. Maheshwari Y, Selvaraj V, Hajeri S, Yokomi R. Application of droplet digital PCR for quantitative detection of Spiroplasma citri in comparison with real time PCR. PLoS One. 2017;12(9):e0184751. 18. Zhao Y, Xia Q, Yin Y, Wang Z. Comparison of Droplet Digital PCR and Quantitative PCR Assays for Quantitative Detection of Xanthomonas citri Subsp. citri. PLoS One. 2016;11(7):e0159004. 19. Bustin SA, Benes V, Garson JA, Hellemans J, Huggett J, Kubista M, et al. The MIQE guidelines: minimum information for publication of quantitative real-time PCR experiments. Clin Chem. 2009;55(4):611-22. 20. Law FB, Chen YW, Wong KY, Ying J, Tao Q, Langford C, et al. Identification of a novel tumor transforming gene GAEC1 at 7q22 which encodes a nuclear protein and is frequently amplified and overexpressed in oesophageal squamous cell carcinoma. Oncogene. 2007;26(40):5877-88. 21. Gopalan V, Smith RA, Nassiri MR, Yasuda K, Salajegheh A, Kim SY, et al. GAEC1 and colorectal cancer: a study of the relationships between a novel oncogene and clinicopathologic features. Human pathology. 2010;41(7):1009-15. 22. Gopalan V, Yasuda K, Pillai S, Tiang T, Leung M, Lu CT, et al. Gene amplified in oesophageal cancer 1 (GAEC1) amplification in colorectal cancers and its impact on patient's survival. J Clin Pathol. 2013;66(8):721-3. 23. Tang H, Cai Q, Li H, Hu P. Comparison of droplet digital PCR to real-time PCR for quantification of hepatitis B virus DNA. Bioscience, Biotechnology, and Biochemistry. 2016;80(11):2159-64. 24. Pichon M, Gaymard A, Josset L, Valette M, Millat G, Lina B, et al. Characterization of oseltamivir-resistant influenza virus populations in immunosuppressed patients using digital-droplet PCR: Comparison with qPCR and next generation sequencing analysis. Antiviral research. 2017;145:160-7. 25. Miotke L, Lau BT, Rumma RT, Ji HP. High sensitivity detection and quantitation of DNA copy number and single nucleotide variants with single color droplet digital PCR. Anal Chem. 2014;86(5):2618-24. 26. Beck J, Bierau S, Balzer S, Andag R, Kanzow P, Schmitz J, et al. Digital droplet PCR for rapid quantification of donor DNA in the circulation of transplant recipients as a potential universal biomarker of graft injury. Clin Chem. 2013;59(12):1732-41. 27. Taylor SC, Laperriere G, Germain H. Droplet Digital PCR versus qPCR for gene expression analysis with low abundant targets: from variable nonsense to publication quality data. Scientific reports. 2017;7(1):2409. 28. Wang M, Yang J, Gai Z, Huo S, Zhu J, Li J, et al. Comparison between digital PCR and real-time PCR in detection of Salmonella typhimurium in milk. International journal of food microbiology. 2018;266:251-6. 29. Morisset D, Stebih D, Milavec M, Gruden K, Zel J. Quantitative analysis of food and feed samples with droplet digital PCR. PLoS One. 2013;8(5):e62583. 30. Pinheiro LB, Coleman VA, Hindson CM, Herrmann J, Hindson BJ, Bhat S, et al. Evaluation of a droplet digital polymerase chain reaction format for DNA copy number quantification. Analytical chemistry. 2012;84(2):1003-11. 31. Huggett JF, Foy CA, Benes V, Emslie K, Garson JA, Haynes R, et al. The Digital MIQE Guidelines: Minimum Information for Publication of Quantitative Digital PCR Experiments. Clinical Chemistry. 2013;59(6):892. 32. Creating Standard Curves with Genomic DNA or Plasmid DNA Templates for Use in Quantitative PCR: Applied Biosystem; 2003 [Available from: http://www6.appliedbiosystems.com/support/tutorials/pdf/quant_pcr.pdf.

135

Chapter 5: Identification of GAEC1 mutations

5.1 Preface

Studies have shown that GAEC1 copies differ significantly between colorectal adenocarcinoma, colorectal adenoma and non-neoplastic colorectal tissues. Indicating that

GAEC1 copy number is related to the site of cancer [56]. Furthermore, immunohistochemistry staining of colon cancer at different stages showed that GAEC1 expression increased with increasing cancer stage [59]. Changes of GAEC1 in both copy number and expression levels indicate that there must be specific epigenetic modification or mutations at play. Genetic mutations have led to the loss of function of many tumour suppressor genes or the activation of oncogenic properties. Nonetheless, only a single study has been performed to identify

GAEC1 mutations in oesophageal squamous cell carcinoma and no mutations have been reported within the coding region of GAEC1 thus far. This chapter focuses on identifying potential GAEC1 mutation throughout the exon region of the gene in colon cancer and determine its association with clinicopathological factors in patients with colorectal carcinoma.

In addition, the impact of mutations on the expression of GAEC1 mRNA in colon cancer were examined.

136

STATEMENT OF CONTRIBUTION TO CO-AUTHORED PUBLISHED PAPER

Chapters 5 includes a co-authored paper. The bibliographic details of the co-authored paper, including all authors, are:

Title: GAEC1 mutations and copy number aberration is associated with biological aggressiveness of colorectal cancer

Authors: Katherine Ting-Wei Lee, Vinod Gopalan, Farhadul Islam, Riajul Wahab, Afraa Mamoori, Cu-tai Lu, Robert Anthony Smith, Alfred King-yin Lam

Name of the journal: European Journal of Cell Biology

Status: Published

Publication link: https://doi.org/10.1016/j.ejcb.2018.03.002

My contribution to the paper involved: I drafted the manuscript, designed, optimised and ran the entire ddPCR for CNA detection and performed the entire workflow for sample preparation for Sanger sequencing. I processed the tissue samples and harvested the DNA samples. I was involved in the Sanger sequencing analysis and statistical analysis of this manuscript.

(Signed) ______(Date)___ 22/11/2018__ Katherine Ting-Wei Lee

(Countersigned) ______(Date)___ 22/11/2018__ Corresponding author of paper and principal supervisor: Prof. Alfred King-Yin Lam

(Countersigned) ______(Date)__ 22/11/2018___ Corresponding author of paper and principal supervisor: Dr. Vinod Gopalan

137

5.2 Detailed description of GAEC1 mutations and copy number aberration association with biological aggressiveness in colorectal cancer

GAEC1 mutations and copy number aberration is associated with biological aggressiveness of colorectal cancer

Katherine Ting-Wei Lee1, Vinod Gopalan1, Farhadul Islam1,2, Riajul Wahab1, Afraa Mamoori, Cu-tai Lu3, Robert Anthony Smith1,4, Alfred King-yin Lam1*

1Cancer Molecular Pathology, School of Medicine and Griffith Health Institute, Griffith University, Gold Coast, Queensland, 4222, Australia 2Department of Biochemistry of Molecular Biology, University of Rajshahi-6205, Bangladesh 3Department of Surgery, Gold Coast University Hospital, Southport, Queensland, 4215, Australia 4Genomics Research Centre, Institute of Health and Biomedical Innovation, Queensland University of Technology, Kelvin Grove, Queensland 4059, Australia

5.2.1 Abstract

GAEC1 (gene amplified in oesophageal cancer 1) is a transforming oncogene with tumorigenic potential observed in both oesophageal squamous cell carcinoma and colorectal cancer. Nonetheless, there has been a lack of study done on this gene to understand how this gene exerts its oncogenic properties in cancer. This study aims to identify novel mutation sites in GAEC1. To do so, seventy-nine matched colorectal cancers were tested for GAEC1 mutation via Sanger sequencing. The mutations noted were investigated for the correlations with the clinicopathological parameters of the patients with cancer. Additionally, GAEC1 copy number aberration (CNA), mRNA and protein expression were determined with the use of droplet digital (dd) polymerase chain reaction (PCR), real-time PCR and Western blot (confirmed with immunofluorescence analysis). GAEC1 mutation was noted in 8.8% (n=7/79) of the cancer tissues including one missense mutation, four loss of heterozygosity (LOH) and two substitutions. These mutations were significantly associated with cancer perforation (p=0.021).

GAEC1 mutation is frequently associated with increased GAEC1 protein expression.

Nevertheless, GAEC1 mRNA and protein are only weakly associated. Taken together, GAEC1

138 mutation affects GAEC1 expression and is associated with poorer clinical outcomes. This further strengthens the role of GAEC1 as an oncogene.

Keywords colorectal cancer, ddPCR, sequencing, copy number aberration, mutation, polymorphism

5.2.2 Introduction

Gene amplified in oesophageal cancer 1 (GAEC1) was first reported in the molecular pathogenesis of oesophageal squamous cell carcinoma [1]. GAEC1 is located within the first intron of SH2B adaptor protein 2 (SH2B2) gene in 7q22 [2] and consisted of only a single- exon [1]. Thus, it fitted the criteria as a nested gene [3, 4]. The 2052bp long GAEC1 mRNA produces a small nuclear protein of 109 amino acids (~15kDa) [1, 5].

Overexpression of GAEC1 in 3T3 fibroblast cells formed undifferentiated sarcoma in athymic nude mice and increased colony formation, indicating its role as a transforming oncogene [1]. On the other hand, knockdown of GAEC1 fully suppressed xenograft tumour growth in mice [5]. In vitro analysis also showed that knockdown of GAEC1 inhibited cell proliferation, reduced cell migration capacity, decreased clonogenic potentiality and induced apoptosis in both oesophageal squamous cell carcinoma [6] and colon cancer cells [5]. In addition, knockdown of GAEC1 was also associated with reduced expression of Bcl-2 (B-cell lymphoma 2) and KRAS (Kirsten rat sarcoma virus) and augmented expression of p53 [5].

Despite having preliminary data on genetic variations and in vitro/in vivo functional properties, the mutation profile of GAEC1 and their associated pathogenic effects in human cancers are still unknown. Furthermore, no mutations or polymorphisms have been noted in

GAEC1 coding sequence [1]. Thus, for the first time, the current study aims to identify possible mutations and determine the absolute quantification of GAEC1 copy number aberration in patients with colorectal cancer. Due to the nature of GAEC1 that had only one exon and a total

139 sequence length of 2052bp, Sanger sequencing was the preferred method over Next Generation

Sequencing (NGS). Additionally, the correlations of GAEC1 mutations with various clinical and pathological parameters in patients with colorectal cancer were analysed in this study.

5.2.3 Material and methods

5.2.3.1 Recruitment of tissue samples and clinicopathological data

Cancer tissues and matched non-neoplastic mucosae near the surgical resection margin were obtained from the same patient who underwent resection of colorectal carcinomas by a colorectal surgeon (CTL) between the years 2012 to 2014 in a Queensland-based hospital. The patients were recruited on an ongoing basis during the study period. The patients were excluded from the study if the prospectively collected tissues lacked adequate cancer mass or were lost in clinical follow-up. Ethical approval was obtained for the use of these tissues from the Griffith University Human Research Ethics Committee (GU Ref No: MSC/17/10/HREC).

These samples were snap frozen in liquid nitrogen and stored at -80°C until use.

Colorectal carcinomas were graded and typed according to the World Health Organization

(WHO) criteria [7] and staged according to the tumour, lymph node and metastases (TNM) classification [8]. Only adenocarcinomas were included in the study. After examination, 79 patients (40 females; 39 males) with colorectal adenocarcinoma were included for the analysis.

A schematic summary of the exclusion criteria is shown in Figure 5.1.

The mean age of patients within this study was 67 years (ranging from 24 to 91). The site and size (length in mm) of the cancers were recorded. Proximal cancers were cancer located in the caecum, ascending colon and transverse colon whereas distal cancers were defined as the cancers found in the region of descending colon, sigmoid colon and rectum. 45.6% of colorectal cancers were found in the proximal colon (n=36) while 54.4% of colorectal cancers were found in the distal colorectum. Out of the 79 patients, 86.1% (n=68) had no cancer

140 perforation while the remaining 13.9% (n=11) patients had perforation. A total of 2.5% (n=2) patients were diagnosed with stage I carcinomas, 11.4% (n=9) with stage II carcinomas, 60.8%

(n=48) with stage III carcinomas and 25.3% (n=20) with stage IV carcinomas.

5.2.3.2 Clinical Management

Clinical management was by the standardised multi-disciplinary protocol. All the patients were discussed in weekly multi-disciplinary team management meeting during their management. The cancer tissues were tested for microsatellite instability (MSI) markers (by immunochemistry) and tested for BRAF mutation status (via Sanger sequencing) where MSI proved positive according to the clinical guidelines. Pathological staging was used as a determining factor for post adjuvant therapy. The follow-up period was outlined as the interval between the date of surgical resection and the date of death or end date of the study. The actuarial survival rate of the patients was assessed from the date of surgery to the date of death of the last follow-up. Only cancer-related death was accepted as the endpoint in the statistical analysis [9]. Persistence or recurrence of the disease was also noted.

141

Figure 5.1: Schematic representation of the methodological flow used. Inclusion criteria – patients with colorectal carcinomas, underwent surgical resection by the same surgeon; and, the same pathologist examined all clinicopathological details. Exclusion criteria – patients who were lost to follow up or no adequate cancer tissue collected.

5.2.3.3 Cell culture and transfection

Three colon cancer cell lines SW480 (ATCC® CCL-228™), SW48 (ATCC® CCL-

231™) and HCT 116 (ATCC® CCL-247™) and one non-neoplastic colon epithelial cell line

FHC (ATCC® CRL-1831™) were obtained from the American type culture collection (ATCC) and maintained accordingly.

SW480, SW48 and HCT 116 cells were cultured in RPMI (Roswell Park Memorial

Institute) 1640 medium supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin and incubated in a 37°C incubator supplemented with 5% CO2. FHC cells were cultured in Dulbecco's Modified Eagle Medium (DMEM):F-12 media supplemented with HEPES (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid) buffer for a final concentration of 25mM, 10ng/mL cholera toxin, 0.005mg/mL insulin, 0.005mg/mL of transferrin, 100ng/mL transferrin, 10% FBS and 1% penicillin/streptomycin and incubated in

142 a 37°C incubator supplemented with 5% CO2. All cells were routinely checked for mycoplasma contamination. SW48 cells were transiently transfected using jetPRIME® Polyplus transfection reagent (Polyplus-transfection® SA, Illkirch, France) according to the manufacturer’s protocol with GAEC1 plasmid kindly provided by Law and colleague [1].

5.2.3.4 Genomic DNA and RNA extraction

Tissues were sliced to 7µm sections using a cryotome. Then, the sections were stained with haematoxylin & eosin (H&E) to confirm the pathological diagnosis. Tissues were only accepted as samples for further analysis if the cancer comprises 70% of the tissue. The clinical and pathological information of every case was reviewed by the author (AKL). Genomic DNA

(gDNA) and RNA were then extracted from the selected tissues with Tissue & Blood extraction kit and miRNeasy Mini kit (Qiagen, Hilden, Germany) respectively according to the manufacturer’s protocol. gDNA and RNA for cell lines were also harvested using this kit.

Quality and concentration of both gDNA and RNA were measured with A260/A280 ratio as well as A260/A230 ratio using a NanoDrop Spectrophotometer (Thermo Fisher Scientific,

Waltham, Massachusetts, USA). The extracts were stored at 4°C until used.

5.2.3.5 Targeted sequencing

A primer set was designed to cover the entire coding sequence of GAEC1 for thorough mutation scanning using NCBI available tool (Supplementary Table 5.1).

Supplementary Table 5.1: Primer design for conventional PCR assay Target Genes Primers Amplicon size (GenBank number) (base pairs) GAEC1 5′- CCCTGGGGGTCCTAATGGTA -3 745 (AC005088) 5′- GGAGTCGGGAGTTTGGAAGG -3′

143

A gradient PCR was performed on GAEC1 primer set and ran on a 2% gel electrophoresis to ensure specificity of the PCR reaction. The PCR reaction cocktail was prepared using 60ng of gDNA, 10uL of PrimeSTAR® Max DNA polymerase (TaKaRa Bio Inc,

Kusatsu, Shiga Prefecture, Japan), and the final concentration of 0.5µM of GAEC1 primer in a

20µL final volume. The thermal cycling condition was set to 98°C for 10s (initial denaturation), followed by 35 cycles of 98°C for 10s, 60°C for 5s and 72°C for the 30s (denaturing and annealing and extension).

These PCR products were then run through a 2% gel electrophoresis. The bands were excised and cleaned up using the NucleoSpin® Gel and PCR Clean-up kit (MACHEREY-

NAGEL GmbH & Co. KG, Düren, Germany) according to the manufacturer’s protocol. The purified PCR products were then prepared according to guidelines provided by Australian

Genome Research Facility (AGRF) before subjecting them to Sanger sequencing by corresponding forward and reverse primer utilising Big Dye Terminator (BDT) chemistry

Version 3.1 (Applied Biosystems, Foster City, California, USA) via standardised PCR cycling conditions. The analysis was done using a 3730xl capillary sequencer (Applied Biosystems,

Foster City, California, USA) by AGRF. Sequence analysis was performed with Chromas

2.4.3 software.

5.2.3.6 Droplet digital PCR (ddPCR) primer designs

With ddPCR, samples are portioned into approximately 20,000 partitions of nanolitre size droplets and amplified to endpoint using a thermal cycler. DNA molecules within the sample are then counted in an absolute manner by singulating and detecting the fluorescence of each droplet, thus, enabling small differences to be determined [10, 11].

144

ddPCR analysis was performed in a duplex format co-amplifying both the target

GAEC1 (GenBank accession number AC005088) and control gene haemoglobin beta (HBB)

(GenBank accession number NC_000011) genes. Both primer sets for these genes were designed using NCBI (National Center for Biotechnology Information) available tools according to the recommended criteria for duplex reaction set up in EvaGreen based assays for ddPCR [10, 11]. Primer sets were first optimised and confirmed by gel electrophoresis (Figure

5.2a) before optimisation on the ddPCR platform (Figure 5.2b & 5.2c). The list of optimised primer sets is summarised in Supplementary Table 5.2.

Supplementary Table 5.2: Primer design for ddPCR duplex assay Target Genes Primers Amplicon size (GenBank number) (base pairs) GAEC1 5′-CAGGGAAGAAGCAAGTTCCA-3 62 (AC005088) 5′-CCATCTGACACAGAGAGTGC-3′ HBB 5′-GCCCAGTTTCTATTGGTCTCCT-3′ 82 (NC_000011) 5′-TGGATGAAGTTGGTGGTGAGG-3′

145

Figure 5.2: Representative GAEC1 gene amplification and copy number aberration (CNA) in patients with colorectal cancer. (a) Duplex gradient optimisation of GAEC1 (62bp) and HBB (82bp) in 5% agarose gel. Lane 1 contains 100bp DNA ladder; lane 2 and 3 contain PCR product with an annealing temperature of 62.5°C and 64.1°C respectively; lane 4 contains non-template control (NTC). Optimised annealing temperature was determined as 62.5°C. Representative (b) 1D droplet plot and (c) 2D plot of amplified duplex ddPCR product from GAEC1 and HBB. (d) Representative copy number changes between matched tumour and non-tumour tissue in each patient with colorectal cancer. Solid circle symbol (●) represents CNA in colorectal cancer tumour biopsies while diamond symbol (◊) represents CNA in adjacent non-neoplastic mucosae. Error bars indicate the Poisson 95% confidence interval.

146

5.2.3.7 Copy number aberration (CNA) analysis with digital droplet PCR (ddPCR)

Duplex EvaGreen assay mix containing 30ng of gDNA, 2U/µL of HINDIII restriction enzyme, QX200 ddPCR EvaGreen Supermix (Bio-Rad Laboratories, Hercules, California,

USA), and gene assay final concentration of 100nM of GAEC1 primer set and 50nM of HBB primer set in a 22µL final volume was prepared. Water was used as a non-template control.

QX200™ AutoDG™ Droplet Digital™ PCR System (Bio-Rad Laboratories, Hercules,

California, USA) was used to generate approximately 20,000 droplets and then thermally cycled in a conventional thermal cycler. The thermal cycling conditions were set to 95°C for

5min (initial denaturation), followed by 40 cycles of 95°C for 30s and 62°C for 1 min with a ramp rate of 2°C/s (denaturing and annealing), 4°C for 5 min and 90°C for 5 min (signal stabilisation).

PCR products were then transferred onto a QX200 Droplet Reader (Bio-Rad

Laboratories, Hercules, California, USA) analysed by QuantaSoft software version 1.7.4 (Bio-

Rad Laboratories, Hercules, California, USA) and QuantaSoftTM Analysis Pro Software

1.0.596 (Bio-Rad Laboratories, Hercules, California, USA). The calculation was based on

(X/Y)*2 where “X” is the copy of GAEC1 gene (target gene) per microliter, “Y” is the copy of the reference gene (HBB) per microliter and “2” is the CNA of the reference gene per genome.

CNA values of 1±0.49 were scored as one copy whereas CNA value of 2±0.49 was scored as two copies and CNA value of 3±0.49 was scored as three copies. CNA of each cancer sample was then compared with its matched non-neoplastic tissue to identify GAEC1 amplification or deletion.

Supplementary Table 5.3: Primer design for qPCR assay Target Genes Primers Amplicon size (GenBank number) (base pairs) GAEC1 5′-CCTCAGGGAAGAAGCAAGTT-3 121 (AC005088) 5′-TCTTGCATGGTGCCAGTT-3′ GAPDH 5′-TGCACCACCAACTGCTTAGC-3′ 88 (NM_002046) 5′-GGCATGGACTGTGGTCATGAG-3′

147

5.2.3.8 Real-time PCR (qPCR) analysis

Total RNA was reversely transcribed into cDNA with SuperscriptTM II First Strand

Synthesis Kit (Qiagen, Hilden, Germany) according to the manufacturer’s protocol. qPCR analysis was performed with GAPDH (NM_002046) as the control gene and GAEC1 target gene. Both primer sets for qPCR analysis are listed in Supplementary Table 5.3. qPCR cocktail was prepared using 60ng of cDNA, 10uL of 2x SensifastTM SYBR No-ROX mix (Bioline,

London, UK), and the final concentration of 0.5µM of primer pair in a final volume of 20µL.

The thermal cycling condition was set to 95°C for 2min (initial denaturation), followed by 40 cycles of 95°C for 5s, 60°C for 30s followed by a melt curve analysis with the default system settings using the Quantistudio Flex 6 qPCR system (Thermo Fisher Scientific, Waltham,

Massachusetts, USA). The expression level of GAEC1 was analysed using 2 –(delta)(delta)Ct as previously reported [5, 12, 13].

5.2.3.9 Western blot Analysis

NP40 lysis buffer (Thermo Fisher Scientific, Waltham, Massachusetts, USA) enriched with 2% protease inhibitor (Sigma-Aldrich, St. Louis, Missouri, USA) were added into a minimum of 80mg of fresh frozen tissue before homogenization. Homogenate was then left to incubate with agitation for 2h at 4°C before extracting the total protein according to the manufacturer’s protocol for NP40 lysis buffer. Proteins were quantified using the Pierce BCA

Protein Assay Kit (Thermo Fisher Scientific, Waltham, Massachusetts, USA) according to the manufacturer’s protocol. 60µg of protein was loaded onto a 4%-15% gradient SDS-page and transferred onto a polyvinylidene difluoride (PVDF) membrane to detect the protein bands according to our published protocol [5]. Western blot images were then developed using

Chemidoc MP imaging system (Bio-Rad Laboratories, Hercules, California, USA) with

148

Supersignal West Pico Chemiluminescent Substrate (Thermo Fisher Scientific, Waltham,

Massachusetts, USA).

5.2.3.10 Immunofluorescence

Immunofluorescence was then used to identify and the correlation between GAEC1

CNA, mRNA and/or GAEC1 mutation on GAEC1 protein expression. Formalin-fixed paraffin- embedded (FFPE) tissue samples were sectioned into 7µm and at de-waxed by heating at 65°C for 35min and dipping them in xylene for 20min. Samples were then rehydrated with ethanol and water (from 100% ethanol to 95% ethanol) and then incubated in heated antigen retrieval solution for 12min at 10% power in a microwave oven. Samples were then blocked with peroxidase block solution, before blocking with 5% bovine serum albumin (BSA) and then incubating with a custom-made monoclonal GAEC1 antibody (Promab, Biotechnologies,

Richmond, CA, USA) (1:500) overnight at 4°C. After that, the samples were incubated with a secondary antibody labelled with fluorescein isothiocyanate (FITC) (Sigma-Aldrich, St. Louis,

Missouri, USA) at room temperature for 2h before mounting with a glass coverslip. All slides were viewed under the Nikon AIR+ confocal microscope (Nikon Inc, Tokyo, Japan).

5.2.3.11 In silico analysis

Sequence information of GAEC1 was obtained from Law and colleagues [1] and was used when required by a method as input. Due to lack of information on readily available bioinformatics tools such as Mutation Taster [14], results were compared to the databases -

UCSC (University of California Santa Cruz) Genome Browser Database [15] and International

Cancer Genome Consortium (ICGC) Data Portal [16]. Also, PROVEAN (Protein Variation

Effect Analyser) [17] was used to evaluate the sequences of the identified mutations within the

149 coding region. The cut-off value used for PROVEAN was -2.5 for predicting the pathogenic non-pathogenic variants in this study.

5.2.3.12 Statistical analysis

QuantaSoft software version 1.7.4 (Bio-Rad Laboratories, Hercules, California, USA) and QuantaSoftTM Analysis Pro Software 1.0.596 incorporated Poisson statistics were used to call copy number of target sequence without the need of a standard curve [18].

Correlations of GAEC1 copies and mutations with different clinicopathological parameters were performed using chi-square test, likelihood ratio and Fisher’s exact test with

Statistical Package for Social Sciences for Windows (version 25.0 IBM SPSS Inc., New York,

NY, USA). Survival analysis was tested using the Kaplan-Meier method. The significance level was taken at p<0.05.

5.2.4 Results

5.2.4.1 Identification of GAEC1 DNA in colorectal cancer tissues

Both 62bp GAEC1 amplicon and 82bp HBB amplicon (reference gene) were detected in all selected colorectal cancer tissues, and no amplification was observed in non-template control (NTC). Also, successful simultaneous detection of both genes using DNA binding chemistry was achieved with droplet digital PCR (ddPCR) (Figure 5.2d). GAEC1 amplicons

(745bp) were detected in all samples, and no amplification was observed in NTC using conventional PCR (Supplementary Figure 5.1).

150

Supplementary Figure 5.1: Representative amplified GAEC1 PCR product in 2 % agarose gel. Lane 1 contains 100 bp DNA ladder while lane 2-3 represents GAEC1 amplicon that was detected in all samples. Lane 4 includes NTC. Amplicon size of 745bp was detected.

5.2.4.2 Identification of GAEC1 variants in colorectal cancer tissues and cell lines

The median overall follow-up for patients with colorectal cancer is 42 months and the survival rate correlated with the pathological stages of cancer (p<0.001) (Figure 5.3a). Patients with colorectal cancer having GAEC1 mutation had lower survival rates (Figure 5.3b) than those without mutation (p=0.323). Although not statistically significant, there was a quantitative difference between GAEC1 copy number change and patient survival.

Figure 5.3: Survival rate correlated with (a) GAEC1 copy number aberration in colorectal cancer tissues and (b) adjacent non-neoplastic tissue. Increased GAEC1 copy number and presence of mutation reduces the survival rate of patients with colorectal cancer.

151

Table 5.1: Mutations detected in the exon region of GAEC1 in Australian patients with colorectal cancer Sample Stage Mutation Nucleotide change Type of Codon In silico code region mutation change prediction PROVEAN 73 3 Coding Chr7:g.101938714C>A Missense p.Ala2As Deleterious p 76 4 Kozak Chr7:g.101938708C LOH - rs803097 22 2 Kozak Chr7:g.101938708T LOH - rs803097 101 1 5’ UTR Chr7:g.101939150C>T LOH - rs140185601 28 2 5’ UTR Chr7:g.101939070G>A - - rs181281674 105 2 5’UTR Chr7:g.101939127C>T - - rs2242581 116 4 5’ UTR Chr7:g.101939258G>A - - rs181800740 Mutation detected in matched non-cancer tissue samples 23 1 5’ UTR Chr7:g.101939277A>G LOH - Chr7:g.101939237C>T Substitution - 122 2 5’UTR Chr7:g.101939277G>A Substitution - 123 4 5’UTR Chr7:g.101939277G>A Substitution - Nucleotide positions were designated based on human GRCh37/h19; 5’ UTR = 5' untranslated region; 3’ UTR = 3’ untranslated region; WT = wildtype; LOH = loss of heterozygosity; “-” indicates not applicable

Eight variants were noted within GAEC1 in 12.7% (n=10/79) of colorectal cancer

(Table 5.1). Among these, three variants were identified as novel mutations. The five variants were identified by PROVEAN as intron variant within SH2B2 gene: rs803097, rs140185601, rs181281674, rs2242581, rs181800740 (Table 5.1). Allele frequencies for rs803097 was 54.4%

(n=43) with TT, 10.1% (n=8) with CC and 35.4% (n=28) with TC while for rs2245281, the allele frequencies were 1.3% (n=1) with TT, 67.1% (n=53) with CC and 31.6% (n=25) with

TC (Table 5.2). The other three intron variants had only one occurrence of heterozygosity

(Table 5.1).

Table 5.2: Frequent polymorphisms detected in GAEC1 in colorectal cancer. Alleles rs803097 (%, n) rs2242581 (%, n) TT 54.4%, n=43 1.3%, n=1 CC 10.1%, n=8 67.1%, n=53 TC 35.4%, n=28 31.6%, n=25 Total 100.0%, n=79 100.0%, n=79

152

Figure 5.4: Heterozygosity at (a) rs181281674 and (b) rs181800740 as observed in GAEC1 found in colorectal cancer tissues.

Of the three, rs181281674 and rs181800740 were observed in both cancers and matched non-neoplastic mucosal tissues (Figure 5.4). The rs140185601, on the other hand, was found to have a loss of heterozygosity (LOH) at the tumour tissue (Figure 5.5a). LOH was also observed in two cases at the rs803097 and rs2242581 (Figure 5.5b & 5.5c respectively). All cell lines tested (FHC, SW48, SW480 and HCT116) displayed TT allele for rs803097, CC allele for rs2245281, CC allele for rs140185601, GG allele for rs181281674 and GG allele for rs181800740.

Intriguingly, a mutation at Chr7:101939277G>A (Figure 5.6a) was found in matched non-neoplastic tissue. A missense resulting in a change in the second amino acid from alanine to aspartic acid (Figure 5.6b) was predicted by PROVEAN to be deleterious (Table 5.1).

Moreover, a mutation occurring at Chr:7.g101939237C>T (Figure 5.6c) which has never been identified was also observed.

153

Figure 5.5: Loss of heterozygosity (LOH) in GAEC1 found in colorectal cancer tissues. LOH was confirmed with Sanger sequencing. Representation of LOH found at (a) rs140185601 with LOH resulting in a CC allele, (b) rs803097 with LOH resulting in a CC allele.

154

Figure 5.6:Novel mutations in GAEC1 found in colorectal cancer tissues. Mutation sites were confirmed by Sanger sequencing. Representation of mutation (a) Chr:7g.101939277G>A occurring at the matched non-neoplastic mucosal tissue, (b) Chr:7g.101938714C>A occurring at the cancer tissue changing the second amino acid from alanine to aspartic acid; and (c) Chr:7g.101939237C>T.

Together, these mutations were significantly associated with perforation in colorectal cancer (p=0.021) (Table 5.3). GAEC1 mutations that occurred at Chr7:g.101939277 were significantly correlated with patient’s gender (p=0.04) as well as the site of cancer occurrence

(p=0.028). Only females and/or cancer located at the proximal colon had a substitution of

Chr7:g.101939277G>A occurring at their non-neoplastic tissue (Table 5.4).

155

Table 5.3: Statistical correlation between GAEC1 mutation and clinicopathological features of Australian patients with colorectal cancer Characteristics No. of GAEC1 Mutation p value patients No Yes Gender Female 40(50.6%) 35(87.5%) 5(12.5%) 0.226 Male 39(49.4%) 37(94.9%) 2(5.1%) Age <70 39(49.4%) 34(87.2%) 5(12.8%) 0.205 ≥70 40(50.6%) 38(95.0%) 2(5.0%) Site Proximal colon 36(45.6%) 32(88.9%) 4(11.1%) 0.520 Distal colorectal 43(54.4%) 40(93.0%) 3(7.0%) Synchronous cancer No 72(91.1%) 66(91.7%) 6(8.3%) 0.492 Yes 7(8.9%) 6(85.7%) 1(14.3%) Size ≤40mm 44(55.7%) 40(90.9%) 4(9.1%) 0.628 >40mm 35(44.3%) 32(91.4%) 3(8.6%) Stage I and II 43(54.4%) 39(90.7%) 4(9.3%) 0.600 III and IV 36(45.6%) 33(91.7%) 3(8.3%) T-stage T1 2(2.5%) 2(100.0%) 0(0.0%) T2 9(11.4%) 7(77.8%) 2(22.2%) 0.454 T3 48(60.8%) 45(93.8%) 3(6.3%) T4 20(25.3%) 18(90.0%) 2(10.0%) Lymph node metastasis Absent 61(77.2%) 55(90.2%) 6(9.8%) 0.495 Present 18(22.8%) 17(94.4%) 1(5.6%) Distant metastasis Absent 62(78.5%) 57(91.9%) 5(8.1%) 0.470 Present 17(21.5%) 15(88.2%) 2(11.8%) Perforation Absent 68(86.1%) 64(94.1%) 4(5.9%) 0.021 Present 11(13.9%) 8(72.7%) 3(27.3%) MSI Absent 64(81.0%) 58(90.6%) 6(9.4%) 0.602 Present 15(19.0%) 14(93.3%) 1(6.7%) MSI microsatellite instability.

156

Table 5.4: Statistical correlation between GAEC1 variant Chr7: g.101939277 and clinicopathological features of Australian patients with colorectal cancer Characteristics No. of Chr7: g.101939277 patients G A p value Gender Female 40(50.6%) 37(92.5%) 3(7.5%) 0.040 Male 39(49.4%) 39(100.0%) 0(0.0%) Age <70 39(49.4%) 38(97.4%) 1(2.6%) 0.510 ≥70 40(50.6%) 38(95.0%) 2(5.0%) Site Proximal colon 36(45.6%) 33(91.7%) 3(8.3%) 0.028 Distal colorectal 43(54.4%) 43(100.0%) 0(0.0%) Synchronous cancer No 72(91.1%) 69(95.8%) 3(4.2%) 0.451 Yes 7(8.9%) 7(100.0%) 0(0.0%) Size ≤40mm 44(55.7%) 43(97.7%) 1(2.3%) 0.414 >40mm 35(44.3%) 33(94.3%) 2(5.7%) Stage I and II 43(54.4%) 41(95.3%) 2(4.7%) 0.567 III and IV 36(45.6%) 35(97.2%) 1(2.8%) T-stage T1 2(2.5%) 2(100.0%) 0(0.0%) T2 9(11.4%) 8(88.9%) 1(11.1%) 0.457 T3 48(60.8%) 46(95.8%) 2(4.2%) T4 20(25.3%) 20(100.0%) 0(0.0%) Lymph node metastasis Absent 61(77.2%) 58(95.1%) 3(4.9%) 0.208 Present 18(22.8%) 18(100.0%) 0(0.0%) Distant metastasis Absent 62(78.5%) 60(96.8%) 2(3.2%) 0.522 Present 17(21.5%) 16(94.1%) 1(5.9%) Perforation Absent 68(86.1%) 65(95.6%) 3(4.4%) 0.338 Present 11(13.9%) 11(100.0%) 0(0.0%) MSI Absent 64(81.0%) 62(96.9%) 2(3.1%) 0.473 Present 15(19.0%) 14(93.3%) 1(6.7%) Nucleotide positions were designated based on human GRCh37/h19; MSI microsatellite instability

157

5.2.4.3 GAEC1 copy number aberrations

The current study showed that only 21.5% (n=17) colorectal cancer had GAEC1

CNA of 2 while 53.2% (n=42) had a copy of GAEC1 deleted in patients with colorectal cancer (Figure 5.7). Of the remaining 25.3% (n=20) patients, amplification GAEC1 CNA in cancer tissues compared to their matched non-neoplastic tissues was observed in the bulk majority of patients (24.1%, n=19) while deletion was only observed in 1.3% (n=1) of patients with colorectal cancer (Figure 5.7). Changes in GAEC1 CNA (Table 5.5) was not observed in patients with synchronous cancer (p=0.038) nor in patients who portrayed microsatellite instability (p=0.040). Interestingly, lower GAEC1 CNA in the cancer tissues was more frequently associated with cancer occurring at distal region (p=0.018) while an increase in GAEC1 CNA was significantly correlated with later T-stage

(p=0.048) and presence of perforation (p=0.001) (Supplementary Table 5.4). Increased

GAEC1 CNA in matched non-neoplastic tissues, on the other hand, was associated with increased tumour size (p=0.003) (Supplementary Table 5.4).

C N A in c o lo re c ta l c a n c e r

C N A in a d ja c e n t n o rm a l tis s u e

) A

N 3

(

e b

A 2

p 1

o C

3 7 9 0 1 2 3 5 1 1 7 0 1 2 5 6 6 7 8 9 6 7 9 0 3 4 6 2 0 1 2 3 6 7 8 9 1 2 4 5 6 7 9 0 1 2 3 5 6 7 8 9

4 5 0 1 3 4 0 1 2 3 0 1 2 3 2 5 6 7 9 4 5

1 5 7 8 6

1 1 1 1 2 2 2 2 2 2 2 2 3 3 3 3 4 4 4 4 5 5 6 6 6 7 7 1 1 2 7 7 7 7 7 7 8 8 8 8 8 8 8 9 9 9 9 9 9 9 9 9

0 0 1 1 1 1 2 2 2 0 0 0 0 1 1 1 1 1 2 2 2

D D D D D

1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

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 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 D D D D D D D D D D D D D D S a m p le C o d e Figure 5.7: CNA of 79 matched samples of patients with colorectal cancer. CNA ranged between 1-3 copies with 19 cases showing amplification in copies compared to their matched non-neoplastic mucosal tissue, 1 cases showing deletion in copies compared to their matched non-neoplastic mucosal tissue and 59 cases showing no apparent changes between their matched counterparts. Solid circle symbol (●) represents CNA in colorectal cancer tumour biopsies while diamond symbol (◊) represents CNA in adjacent normal tissue. Error bars indicate the Poisson 95% confidence interval. 158

Table 5.5: GAEC1 copy number aberration (between cancer tissues and matched non-neoplastic tissues) and statistical correlation between clinicopathological features of Australian patients with colorectal cancer. Characteristics No. of GAEC1 CNA Changes p value patients Yes No Gender Female 40(50.6%) 11(27.5%) 29(72.5%) 0.424 Male 39(49.4%) 9(23.1%) 30(76.9%) Age <70 39(49.4%) 11(28.2%) 28(71.8%) 0.373 ≥70 40(50.6%) 9(22.5%) 31(77.5%) Site Proximal colon 36(45.6%) 12(33.3%) 24(66.7%) 0.108 Distal colorectal 43(54.4%) 8(18.6%) 35(81.4%) Synchronous cancer No 72(91.1%) 20(27.8%) 52(72.2%) 0.038 Yes 7(8.9%) 0(0.0%) 7(100.0%) Size ≤40mm 44(55.7%) 12(27.3%) 32(72.1%) 0.428 >40mm 35(44.3%) 8(22.9%) 27(77.1%) Stage I and II 43(54.4%) 12(27.9%) 31(72.1%) 0.377 III and IV 36(45.6%) 8(22.2%) 28(77.8%) T-stage T1 2(2.5%) 0(0.0%) 2(100.0%) T2 9(11.4%) 1(11.1%) 8(88.9%) 0.399 T3 48(60.8%) 14(29.2%) 34(70.8%) T4 20(25.3%) 5(25.0%) 15(75.0%) Lymph node metastasis Absent 61(77.2%) 15(24.6%) 46(75.4%) 0.503 Present 18(22.8%) 5(27.8%) 13(72.2%) Distant metastasis Absent 62(78.5%) 16(25.8%) 46(74.2%) 0.561 Present 17(21.5%) 4(23.5%) 13(76.5%) Perforation Absent 68(86.1%) 15(22.1%) 53(77.9%) 0.098 Present 11(13.9%) 5(45.5%) 6(54.5%) MSI Absent 64(81.0%) 19(29.7%) 45(70.3%) 0.040 Present 15(19.0%) 1(6.7%) 14(93.3%) MSI microsatellite instability.

159

Supplementary Table 5.4: GAEC1 copy number in cancer tissue and macthed non-neoplastic tissues and statistical correlation between clinicopathological features of Australian patients with colorectal cancer Characteristics No. of GAEC1 Copy in cancer tissue GAEC1 Copy in Matched Non-neoplastic tissues patients 1 2 3 p value 1 2 p value Gender Female 40(50.6%) 22(55.0%) 17(42.5%) 1(2.5%) 0.995 32(80.0%) 8(20.0%) 0.278 Male 39(49.4%) 21(53.8%) 17(43.6%) 1(2.6%) 28(71.8%) 11(28.2%) Age <70 39(49.4%) 19(48.7%) 18(46.2%) 2(5.1%) 0.177 29(74.4%) 10(25.6%) 0.475 ≥70 40(50.6%) 24(60.0%) 16(40.0%) 0(0.0%) 31(77.5%) 9(22.5%) Site Proximal colon 36(45.6%) 15(41.7%) 21(58.3%) 0(0.00%) 0.018 27(75.0%) 9(25.0%) 0.532 Distal colorectal 43(54.4%) 28(65.1%) 13(30.2%) 2(4.7%) 33(76.7%) 10(23.3%) Synchronous cancer No 72(91.1%) 38(52.8%) 32(44.4%) 2(2.8%) 0.326 55(76.4%) 17(23.6%) 0.539 Yes 7(8.1%) 5(71.4%) 2(28.6%) 0(0.0%) 5(71.4%) 2(28.6%) Size ≤40mm 44(55.7%) 27(61.4%) 17(38.6%) 0(0.0%) 0.087 39(88.6%) 5(11.4%) 0.003 >40mm 35(44.3%) 16(45.7%) 17(48.6%) 2(5.7%) 21(60.0%) 14(40.0%) Stage I and II 43(54.4%) 22(51.2%) 21(48.8%) 0(0.0%) 0.130 34(79.1%) 9(20.9%) 0.327 III and IV 36(45.6%) 21(58.3%) 13(36.1%) 2(5.6%) 26(72.2%) 10(27.8%) T-stage T1 2(2.5%) 2(100.0%) 0(0.0%) 0(0.0%) 2(100.0%) 0(0.0%) T2 9(11.4%) 7(77.8%) 2(22.2%) 0(0.0%) 0.048 8(88.9%) 1(11.1%) 0.404 T3 48(60.8%) 22(45.8%) 26(54.2%) 0(0.0%) 34(70.8%) 14(29.2%) T4 20(25.3%) 12(60.0%) 6(30.0%) 2(2.5%) 16(80.0%) 4(20.0%) Lymph node metastasis Absent 61(77.2%) 35(57.4%) 24(39.3%) 2(3.3%) 0.320 47(77.0%) 14(23.0%) 0.446 Present 18(22.8%) 8(44.4%) 10(55.6%) 0(0.0%) 13(72.2%) 5(27.8%) Distant metastasis Absent 62(78.5%) 35(56.5%) 26(41.9%) 1(1.6%) 0.366 48(77.4%) 14(22.6%) 0.385 Present 17(21.5%) 8(47.1%) 8(47.1%) 1(5.9%) 12(70.6%) 5(29.4%)

160

Supplementary Table 5.4 continued

Characteristics No. of GAEC1 Copy in cancer tissue GAEC1 Copy in Matched Non-neoplastic tissues patients 1 2 3 p value 1 2 p value Perforation Absent 68(86.1%) 39(57.4%) 29(42.6%) 0(0.0%) 0.001 52(76.5%) 16(23.5%) 0.524 Present 11(13.9%) 4(36.4%) 5(45.5%) 2(18.2%) 8(72.7%) 3(27.3%) MSI Absent 64(81.0%) 34(53.1%) 28(43.8%) 2(3.1%) 0.527 50(78.1%) 14(21.9%) 0.267 Present 15(19.0%) 9(60.0%) 6(40.0%) 0(0.0%) 10(66.7%) 5(33.3%) MSI microsatellite instability.

161

5.2.4.4 Effects of mutation in GAEC1 expression

The impact of GAEC1 mutations on their expression changes were examined by Western blot and immunofluorescence. Four colorectal cancer cases positive of

GAEC1 mutation were compared against seven GAEC1 wild-type colorectal cancer cases. GAEC1 transfected SW48 cells were used as a positive control for Western blot.

Patients with non-mutated GAEC1 sequence (T119, T57 and T01; except T40) had remarkably lower GAEC1 expression while patients with mutated GAEC1 sequence

(T22 and T73; except for T23 and T76) had a higher GAEC1 expression (Figure 5.8a

& 5.8b). Coincidentally, patients with GAEC1 variants (rs181281674 and rsrs2242581) showed higher GAEC1 expression. These findings were confirmed by immunofluorescence (Figure 5.8c). However, the GAEC1 protein expression finding was not strongly correlated with the mRNA expression (Figure 5.8d).

162

Figure 5.8: GAEC1 expression in patients with colorectal cancer. (a) Western blot analysis of GAEC1 protein expression with GAPDH as housekeeping protein. (b) Bar graph of GAEC1 band intensity normalised to GAPDH. Positive control used for Western blot is the SW48 cells transfected with GAEC1. Samples with mutations showed increased protein expression (T22, T73 and T23; except T76). (c) Representative images of colorectal cancer tissues with (i) low GAEC1 expression and (ii) high GAEC1 expression. (d) GAEC1 mRNA expression with GAPDH as an internal control gene.

163

5.2.5 Discussion

Despite having limited information on the GAEC1 gene, a search of its sequence region (Chr7:g.101938192-101940242) on the ICGC (international cancer genome consortium) [16] showed numerous (n=26) variants in the GAEC1 sequence.

In the current study, 8.8% (n=7/79) of colorectal cancers were found to have GAEC1 mutation (Table 5.3) and were clinically correlated with poor patient survival (Figure

3.3b). Fascinatingly, both mutations (Table 5.3) and increased of GAEC1 CNA in tumour tissues (Supplementary Table 5.4) were found to significantly (p=0.0021 and p=0.001 respectively) affect perforation in colorectal cancer (associated with clinical aggressiveness of cancer).

In most cases (except for the case of sample 76), increased expression of

GAEC1 (Figure 5.8) was found in patients with GAEC1 mutation. This study noted a change in the amino acid sequence of the GAEC1 protein in a colorectal cancer case with a missense mutation in GAEC1. Based on the computation analysis with

PROVEAN, this amino acid change (p.Ala2Asp) has a deleterious effect (Table 5.2).

Interestingly, this colorectal cancer tissue also showed a remarkable increase in

GAEC1 protein expression (Figure 5.8a & 5.8b), despite having only a single DNA copy (Table 5.6) and a two-fold increase in mRNA expression (Figure 5.8d). This indicates the potential of the mutation at p.Ala2Asp to cause overexpression of the

GAEC1 protein. Coincidentally, this mutation was found in a patient with late stage

(stage 3) colorectal cancer (Table 5.2). Both sample 76 and sample 22 has LOH occurring at rs803097 (Figure 5.5a & 5.5b) conveniently located at the Kozak sequence (Table 5.3). It is not surprising that the LOH to a CC allele in sample 76 resulted in reduced GAEC1 expression while LOH to a TT allele in sample 22 resulted in GAEC1 overexpression (Figure 5.8a & 5.8b). Unexpectedly, sample 40 which had no mutation and had a TT allele at rs803097 (results not shown) also showed

164 overexpression of GAEC1 (Figure 5.8a & 5.8b). These findings further strengthen the oncogenic and growth promoting properties of GAEC1 [1, 5, 6, 12, 13].

Interestingly, the novel mutation at Chr7:g.101939277 (Figure 5.6a) were only identified in matched non-neoplastic tissue and was only observed in female patients

(p=0.04) and in patients with cancer in their proximal colon (p=0.028) (Table 5.4).

While these mutations are found in the non-coding region of GAEC1, non-coding mutations can cause deleterious effects in many diverse ways. For example, they may alter promoters and enhancers whom control transcription [19-21]. Furthermore, a regulatory sequence present at non-coding region can regulate transcription by looping and targeting coding sequence thousands of kilobases away [22]. Changes in these regions may alter various transcription factor-binding sites, which alters this regulatory control. Alteration at the 3’UTR can also regulate gene expression such as through modifications to miRNA target sites [21]. Nonetheless, this study provides preliminary findings in the detection of novel GAEC1 mutations in colorectal cancer.

Further functional studies are needed to confirm the effect they have.

In this study, we noted that the changes in GAEC1 protein expression are only weakly associated with GAEC1 mRNA expression and GAEC1 CNA (Table 5.6). This finding is not surprising as many studies have found a poor correlation between mRNA abundance and protein abundance due to multifactorial conditions that may influence the central dogma of molecular biology continuously [23, 24]. Even so, the presence of apparently non-pathogenic CNA could also predispose a region to chromosomal rearrangements, which ultimately give rise to disease [25, 26]. As seen in the present study, an increase in GAEC1 copies trended towards a reduced survival rate in patients with colorectal cancer (Figure 5.3c & 5.3d). Only 21.5% (n=17/79) of patient with colorectal cancer showed two copies of GAEC1 in matched tissues (both matched cancer tissues and non-neoplastic tissues). Of the remaining 79.5% (n=62/79), 53.2%

165

(n=42/79) had a deletion in GAEC1 in matches tissues while 24.1% (n=19/79) showed

GAEC1 amplification and 1.3% (n=1/79) showed GAEC1 deletion in cancer tissue when compared to matched non-neoplastic tissue. GAEC1 amplification frequencies in colorectal adenocarcinoma determined in this study was comparable to those reported in oesophageal squamous cell carcinoma [1].

To conclude, GAEC1 mutations occur in a small proportion of colorectal carcinoma. These mutations are frequently associated with increased GAEC1 protein expression. We found that an increase in GAEC1 copy and GAEC1 mutations were significantly associated with biological aggressiveness of the cancer in terms of cancer perforation and survival rates of patients. These indicate that GAEC1 mutation as a potential predictive biomarker for colorectal cancer aggression. Together, these findings augment the role of GAEC1 as an oncogene.

Table 5.6: Changes in GAEC1 mRNA, protein or copy number aberration with or without GAEC1 mutation.

GAEC1 Sample Variant mRNA Protein CNA T73 Codon A2D - ↑ 1 Chr7:g.101939237T, T23 ↑ ↓ 2 mutation Chr7:g.101939277A T76 LOH rs803097 C - ↓ 3 T22 LOH rs803097 T ↑ ↑ 2 T28 rs181281674 ↓ ↓ 2 T40 - ↓ ↑ 3 T119 - ↑ ↓ 2 wildtype T57 - ↓ ↓ 1 T105 - - ↑ 1 T01 - - ↓ 1 T37 - - ↑ 2 Increase or decrease in mRNA expression is denoted by a changed of ± 4 folds while protein expression is based on Western blot image. “↑” indicates increase expression whereas “↓” indicates reduced expression and “-” indicates no expression change. CNA indicated in this table is the CNA obtained from tumour tissues.

166

5.2.6 Acknowledgement

The authors would like to thank Dr Johnny Tang for the provision of the GAEC1 plasmid. We also thank the NH Diagnostic laboratory for the processing of the tissue.

5.2.7 Funding

The project was supported by the student scholarship from Griffith University and funding from Menzies Health Institute Queensland.

5.2.8 Authors Contribution

KTWL drafted the manuscript, designed, optimised and ran the entire ddPCR for CNA detection and performed the entire workflow for sample preparation for Sanger sequencing. KTWL, IF and RW processed the tissue samples and harvested the DNA samples. KTWL, RAS and VG were involved in the Sanger sequencing analysis.

KTWL, VG and AKYL were involved in the statistical analysis of this manuscript.

RAS, VG and AKYL supervised and participated in writing and editing the manuscript.

All authors read and approved the final manuscript.

5.2.9 Conflict of Interest

The authors have no conflict of interest in publishing this manuscript.

167

5.3 Inclusion of GAEC1 protein sequence in GenBank, NCBI

One of the biggest issues faced during the data collection for GAEC1 mutation information for the manuscript in Chapter 5 with the title “GAEC1 mutations and copy number aberrations are associated with biological aggressiveness of colorectal cancer” was the lack of bioinformatics data that could be generated. This was largely because most bioinformatics tools available runs their information through a database such as the GenBank in the NCBI database. To date, GAEC1 protein is not included in the

GenBank NCBI database due to lack of publication addressing GAEC1 genetic and functional characteristics.

Therefore, the primary objective of this work was to get GAEC1 and its protein sequence annotated and recognised as a gene target for future studies and to internationalise the use of GAEC1 as a protein/antibody. Annotating GAEC1 gives the gene an accession number and protein identifier. This allows easy compilation of all information regarding GAEC1. Researchers can easily retrieve data from the NCBI database. Furthermore, having GAEC1 included in the database will open up exciting in silico analysis using bioinformatics tool to further predict and direct more advance studies on GAEC1 protein.

To do this, all published information on GAEC1 protein (including information gathered from this thesis work directly and tangentially) are collated and submitted to

GenBank at NCBI. Information regarding the DNA sequence, mRNA sequence and protein sequence are also given. The overall information submitted to GenBank is listed as a GenBank flat file for GAEC1 and has been successfully published in

GenBank, European Nucleotide Archive (ENA) in Europe and the DNA Data Bank of

Japan (DDBJ) .

168

5.3.1 GenBank flat file for GAEC1

Figure 5.9: Screenshot of the GAEC1 mRNA, complete cds and microsatellite sequences on NCBI Nucleotide website. Link: https://www.ncbi.nlm.nih.gov/nuccore/MG720123

Figure 5.10: Screenshot of GAEC1 mRNA, complete cds with amino acid sequence Link: https://www.ncbi.nlm.nih.gov/nuccore/MG720124

169

Chapter 6: Functional role and interacting partners of GAEC1 in colon cancer

6.1 Preface

Oncogenes exercise their effects by interacting with various modulators

involved in key cancer regulatory pathways. Throughout multiple clinical analysis and

functional assay, it is evident that GAEC1 functions as an oncogene and plays a vital

role in the pathogenesis of colon cancer. Nonetheless, studies to-date have focused on

GAEC1 suppression and have noted the effects on how abrogation of GAEC1 can

hamper colon cancer progression. This study aims to investigate the oncogenic/cell

transformation properties of GAEC1 normal human colonic epithelial cells for the first

time. This chapter has focuses on GAEC1 overexpression in an effort to confirm the

previous findings observed in GAEC1 knock down studies and to identify its effects

on key cancer regulatory genes in notorious oncogenic pathways such as Akt pathway.

This can help narrow down future studies of GAEC1 associated key modulator

targeting studies. Even more importantly, this study aims to identify the potential of

GAEC1 overexpression induced carcinogenesis in colon cells (both normal epithelial

and cancer).

170

STATEMENT OF CONTRIBUTION TO CO-AUTHORED PUBLISHED PAPER

Chapters 6 includes a co-authored paper. The bibliographic details of the co-authored paper, including all authors, are:

Title: GAEC1 drives colon cancer progression

Authors: Katherine Ting-Wei Lee, Jelena Vider, Johnny Cheuk-on Tang, Vinod Gopalan, Alfred King-yin Lam

Name of the journal: Molecular Carcinogenesis

Status: Published

Publication link: https://doi.org/10.1002/mc.22998

My contribution to the paper involved: I drafted the manuscript, optimised and performed all the cell culture, plasmid transfection qPCR, western blot and cell-based assays. I was also involved in the data analysis and statistical analysis of this manuscript.

(Signed) ______(Date)___ 22/11/2018__ Katherine Ting-Wei Lee

(Countersigned) ______(Date)___ 22/11/2018__ Corresponding author of paper and principal supervisor: Prof. Alfred King-Yin Lam

(Countersigned) ______(Date)__ 22/11/2018___ Corresponding author of paper and principal supervisor: Dr. Vinod Gopalan

171

6.2 Detailed description of GAEC1 drives colon cancer progression

Katherine Ting-Wei Lee1, Jelena Vider2, Johnny Cheuk-on Tang3, Vinod Gopalan1,2,

Alfred King-yin Lam1*

1Cancer Molecular Pathology, School of Medicine, Griffith University, Gold Coast, Queensland, 4222, Australia

2 School of Medical Science, Menzies Health Institute Queensland, Griffith University, Gold Coast, Queensland, Australia

3 State Key Laboratory of Chemical Biology and Drug Discovery, Department of Applied Biology and Chemical Technology, The Hong Kong Polytechnic University, Hong Kong, China

Running head: GAEC1 drives colon cancer progression

*Address for correspondence: Dr Vinod Gopalan School of Medicine and Medical Sciences, Griffith University, Gold Coast Campus, Gold Coast QLD 4222, Australia. E-mail: [email protected] Telephone +61 7 56780717; Fax +61 7 56780708

Professor Alfred K Lam, Head of Pathology, Griffith Medical School, Gold Coast Campus, Gold Coast QLD 4222, Australia. E-mail: [email protected] Telephone +61 7 56780718; Fax +61 7 56780303

172

6.2.1 Abstract

Gene amplified in esophageal cancer one (GAEC1) expression and copy number changes are frequently associated with the pathogenesis of colorectal carcinomas. The current study aimed to identify the pathway and its transcriptional factors that GAEC1 interacts with colorectal cancer to gain better understanding of the mechanism by which this gene exercises its effect on colorectal cancer. Two colonic adenocarcinoma cell lines (SW48 and SW480) and a non-neoplastic colon epithelial cell line (FHC) were transfected with GAEC1 to assess the oncogenic potential of

GAEC1 overexpression. Multiple in vitro assays including cell proliferation wound healing, clonogenic, apoptosis, cell cycle and extracellular flux were performed.

Western blot was performed to identify potential gene-interaction partners of GAEC1 in vitro. Results showed that overexpression of GAEC1 significantly increased cell proliferation, migration and clonogenic potential (p<0.05) of colonic adenocarcinoma.

Furthermore, GAEC1 portrayed its ability to influence mitochondrial respiration changes. The observations were in tandem with a significant increase in the expression of pAkt, FOXO3 and MMP9. Thus, GAEC1 has a role in regulating gene pathways, potentially in the Akt pathway. This could help in developing of targeted therapies in future.

6.2.2 Keywords

Oncogene; FOXO3; MMP9; pAkt; mitochondrial

173

6.2.3 Introduction

GAEC1 (Gene amplified in esophageal cancer 1), is located at chromosome

7q22.1 [1], within the first intron region of SH2 B adaptor protein 2 (SH2B2) gene [2].

GAEC1 copy changes proved significant in distinguishing between patients with colorectal cancer from the healthy population [3]. Furthermore, GAEC1 was also able to discriminate between the different ethnicity of patients with colorectal cancer [3].

This gene was even capable of discerning between the different sites of cancer within the bowel [4] and was frequently associated with the bio-aggression of cancer [2].

GAEC1 encodes for a protein that is predominantly expressed with the cytoplasm of colon cancer. Immunohistochemical staining intensity of GAEC1 was found to increase in a stage-dependent manner in patients with colorectal cancer [5].

Interestingly, overexpression of GAEC1 was frequently associated with GAEC1 mutation. Approximately 10% of patients with colorectal cancer have mutations within the exon region of GAEC1 [2]. Furthermore, these patients are more likely to have a cancer with perforation of the bowel wall. These findings suggest that GAEC1 plays an essential role in tumorigenesis.

Wahab and colleague (2017) found that suppression of GAEC1 induced antigrowth properties in vitro and in vivo. Colon cancer cells exhibited increased apoptosis, G2/M phase arrest and had a significant reduction in cell proliferation, migration rate and clonogenic potential with the knockout of GAEC1 [6]. Nonetheless, these studies only focused on knock out studies with no overexpression work of

GAEC1 to confirm these properties. Furthermore, the molecular mechanism in which

GAEC1 affects colorectal cancer progression remains unknown.

The matrix metalloproteinases (MMPs) play a vital role in the proteolysis of the extracellular matrix (ECM) molecules [7, 8] and promotes tumour progression by participating in all aspects of carcinogenesis especially in invasion and migration [7,

174

9]. Interestingly, recent findings indicate that FOXO3 induces MMP9 through the Akt pathway [7, 10]. FOXO3 is a tumour suppressor because it induces cell-cycle arrest and apoptosis [11-13]. However, there has been a shift in the paradigm as high level of FOXO3 had been related to poor prognosis [11, 14] with lower overall survival and recurrence-free survival outcomes in patients with cancer [12, 15-17]. Since suppression of GAEC1 significantly impeded the pathogenicity of colon cancer, we hypothesised that GAEC1 might interact with pAkt, FOXO3 and/or MMP9 to exercise its tumorigenic effects.

This study aims to overexpress GAEC1 in both no-neoplastic colonic epithelial and colon cancer cells to confirm the oncogenic potential and transformation capabilities of GAEC1. Furthermore, western blots were performed to identify the possible interaction between GAEC1 and pAkt, FOXO3 and/or MMP9. This will help in further understanding the mechanistic function of this gene by associating the functional assays with commonly affected pathways and target molecules.

6.2.4 Materials and Methods

6.2.4.1 Cell culture

Two colon cancer cell lines SW480 (ATCC CCL228; stage II, colonic adenocarcinoma) and SW48 (ATCC CCL-231; stage III colonic adenocarcinoma), and one non-neoplastic colonic epithelial cell line (FHC) (ATCC CRL-1831) were used in this study. All cells were maintained according to the American Type Culture

Collection (ATCC) guidelines. SW480 and SW48 were maintained with RPMI

(Roswell Park Memorial Institute) medium supplemented with 10% foetal bovine serum (FBS) and 1% penicillin/ streptomycin at 37℃ in an incubator supplemented with 5% CO2. FHC were cultured in Dulbecco’s Modified Eagle Medium (DMEM):

F-12 medium supplemented with 25mM buffer HEPES (4-(2-hydroxyethyl)-1-

175 piperazineethanesulfonic acid, 10ng/mL cholera toxin, 0.005mg/mL insulin,

0.005mg/mL of transferrin, 100ng/mL transferrin, 10% FBS and 1% penicillin/streptomycin.

6.2.4.2 Plasmid transfection

A GMO (genetically modified organism) project approval was granted from A

GMO (genetically modified organism) project approval was granted from Griffith

University Institutional Biosafety Committee (NLRD/006/15). One of the authors

(JCT) contributed the GAEC1 plasmid with the sequence according to our previous publication [1]. Both GAEC1 plasmid and the control plasmid (mock vector) were transiently overexpressed in all cell lines using jetPRIME® Polyplus transfection reagent (Polyplus-transfection® SA, Illkirch, France) according to the manufacturer’s protocol. Both the control plasmid (mock vector) and the GAEC1 plasmid were constructed using the pcDNA3.1 (-) vector. Proteins and mRNAs from post- transfected cells were harvested and analysed via qPCR (quantitative polymerase chain reaction) and Western blotting respectively to confirm overexpression. All the downstream assays following GAEC1 amplification were performed as three independent experiments.

6.2.4.3 Quantitative real time PCR (qPCR) analysis

Total RNA was extracted from all cells using the RNeasy mini kit (Qiagen,

Hilden, Germany) according to the manufacturer’s guideline. Reverse transcription was then performed using the iScriptTM cDNA Synthesis Kit (Bio-Rad Laboratories,

Hercules, CA) and the expression changes of GAEC1 in the different cell lines were examined using Quantistudio Flex 6 qPCR system (Thermo Fisher Scientific,

Waltham, MA). Samples were prepared in a 20µL volume containing 10µL of 2x

176

SensifastTM SYBR No-Rox mix (Bioline, London, UK), 1µL of 10µM primer pair,

1µL of 60ng template topped up with RNase-free water and qPCR was performed according to the protocol previously described [2]. Amplification efficiency of GAEC1

(MG20123) was normalised against the housekeeping gene glyceraldehyde 3- phosphate dehydrogenase (GAPDH) (NM_002046) with primer details as shown in

Table 6.1.

Table 6.1: Primer design for quantitative real time PCR (qPCR) analysis Target Genes Primers Amplicon size (GenBank number) (base pairs) GAEC1 5′-CCTCAGGGAAGAAGCAAGTT-3 121 (MG20123) 5′-TCTTGCATGGTGCCAGTT-3′ GAPDH 5′-TGCACCACCAACTGCTTAGC-3′ 88 (NM_002046) 5′-GGCATGGACTGTGGTCATGAG-3′

6.2.4.4 Western blot analysis

Cells were lysed using NP40 lysis buffer (Thermo Fisher Scientific) enriched with 2% protease inhibitor (Sigma-Aldrich, St. Louis, MO) and the total protein was harvested by the recommended guidelines. Proteins were then quantified using the

Pierce BCA Protein Assay Kit (Thermo Fisher Scientific) and 30µg of protein was loaded onto a Mini-PROTEAN® TGX™ Precast Gels (Bio-Rad Laboratories) to separate and transferred onto a polyvinylidene fluoride (PVDF) membrane with the

Trans-blot® Turbo™ system (Bio-Rad Laboratories). The membrane was then blocked with 5% non-fat milk powder for one hour at room temperature before overnight incubation of primary antibody at 4℃. The primary antibodies used included

GAEC1 (1:500) (Promab, Biotechnologies, Richmond, CA), GAPDH (1:1000) (Santa

Cruz, Dallas, TX, USA), phosphorylated protein kinase B (pAkt) (1:50) (Santa Cruz),

Forkhead box O3 (FOXO3) (1:100) (Santa Cruz) and Matrix metallopeptidase 9

(MMP-9) (1:100) (Santa Cruz). The membrane was then washed three times with

177

PBST and incubated with relevant secondary antibodies (1:10,000) (Abcam,

Cambridge UK) at room temperature for one hour. The blots were then incubated with

SuperSignalTM West Pico PLUS Chemiluminescent Substrate (Thermo Fisher) for five minutes and then visualizing with the ChemiDoc MP TM Imaging System (Bio-Rad

Laboratories). Expression of GAEC1 protein was quantified and normalized to

GAPDH with ImageJ 1.52 software.

6.2.4.5 Cell proliferation assay

Cells were plated in a 96-well plate to 70% confluency, 24h prior to transfection. The media in each well was replaced with fresh media 4h post- transfection and cells were left to incubate for 72 hours in a 5% CO2 supplemented incubator at 37°C. Subsequently, 10µL of MTT (3-(4,5-Dimethylthiazol-2-yl)-2,5-

Diphenyltetrazolium Bromide) (12mM) was added and the cells were incubated for an additional 4 hours. Approximately 90% of the media was then removed and topped up with 100µL of dimethyl sulfoxide (DMSO) and further incubated for another 10min to dissolve the formazan crystal. Samples were then read at 540nm wavelength.

6.2.4.6 Wound healing assay

Cells were plated in a 6-well plate to 70% confluency, 24 hours before transfection. The media in each well were replenished four hours post-transfection.

After 24h, a P-200 tip was used to scratch the centre of the well. Wound areas were imaged every 24 hours, and the media were changed routinely for up to four days. All wound areas were measured and compared using ImageJ 1.52I.

178

6.2.4.7 Clonogenic assay

Approximately 200 cells were seeded into a 6-well plate with 2mL of complete media and maintained in a 5% CO2 supplemented incubator at 37°C. Cells were routinely checked. When cells in the control plates had reached a minimum of 50 cells/colony (minimum scoring) [18,19], the medium was removed, and the cells were rinsed with phosphate buffered saline (PBS). Then, cells were fixed using fixation solution (acetic acid/methanol 1:7, vol/vol) for five minutes, before removing it and incubating with 0.5% crystal violet solution at room temperature for two hours. After the incubation, the cells were washed with tap water and left to air dry. The number of colonies formed was then used to calculate the survival fractions (SF) according to

Franken and colleagues (2006). Briefly, SF is calculated using the plating efficiency

(PE). Different cell lines have different PE. PE is expressed as the number of colonies form over the number of cells seeded in percentage. Whereas, SF is the number of colonies that arise after transfection of cells expressed in PE [18]:

number of colonies formed PE= x 100% number of cells seeded

number of colonies formed after transfection SF= number of cells seeded x PE

6.2.4.8 Cell cycle assay

Approximately a million cells were harvested and fixed in ice cold 70% ethanol for two hours. Later, ethanol suspended cells were pelleted and washed with PBS before treating with 500ng/mL RNase A and then staining with 50ng/mL of propidium iodide (PI) solution. Cells were run through flow cytometry (BD Fortessa, BD

Biosciences, Franklin Lakes, NJ, USA) and analysed with FlowJoTM 10.

179

6.2.4.9 Cell apoptosis assay

Approximately a million cells were harvested and washed with PBS before re- suspending in Annexin V binding buffer. Cells were later on stained with 2.5µL of

Alexa Fluor® 350 (Invitrogen, Carlsbad, CA, USA) and 25ng/mL of PI solution. Cells were run through flow cytometry (BD Fortessa, BD Biosciences) and analysed with

FlowJoTM 10.

6.2.4.10 Extracellular flux assay

Cells were plated in a Seahorse XFp Cell Culture Miniplate to 85% confluency in a 5% CO2 supplemented incubator at 37°C. Subsequently, media is replaced with

Agilent Seahorse XF Base Medium supplemented with 1mM pyruvate, 2mM glutamine and 0.11mM glucose, pH 7.4 37°C. Cells are incubated at 37°C and ambient

CO2 for 40 minutes before the assay. Assays were performed with a Seahorse XFp extracellular flux analyser (Agilent, Santa Clara, CA, USA) according to previously published protocol [20].

6.2.4.11 Data analysis

All statistical analysis was carried out using IBM (New York, NY, USA) SPSS

Statistics version 25. GraphPad Prism 7 was used to plot the graphs and error bars. All flow cytometry analyses were performed with FlowJoTM 10. Cell cycle analysis was performed using Dean-Jett Fox model found within the FlowJoTM 10 software. ImageJ was used to quantitate all bands and qualitative images. The significance level was taken at P<0.05.

180

6.2.5 Results

6.2.5.1 GAEC1 transfection efficiency verification

Figure 6.1:Overexpression of GAEC1 in colon cancer (SW480 and SW48) and non-neoplastic colonic epithelial cell (FHC). (A) Fold changes of GAEC1 observed in SW48, SW480 and FHC transfected with GAEC1 plasmid (SW48+GAEC1, SW480+GAEC1 and FHC+GAEC1) with qPCR analysis. (B) Western blot analysis of GAEC1 protein bands from all cells normalized against GAPDH and (C) quantified using ImageJ. All cells transfected with GAEC1 plasmid showed a significant increase in GAEC1 expression. Immunofluorescence images shown are a representative of each cell line and the results plotted as mean ± standard deviation with * indicating a P<0.05 while ** indicates P<0.01 and *** indicates P<0.001 when compared to both each control and mock vector respectively. Mock vector indicates cells transfected with an empty vector.

Compared to the control cells, all three colon cell lines showed a significant increase (P<0.05) in GAEC1 expression following its transfection (Figure 6.1). The 181 relative expression of GAEC1 mRNA and protein levels were obtained through normalisation with GAPDH housekeeping protein. Only a minimal change of GAEC1 expression is observed in all mock controls (SW48+mock, SW480+mock and FHC+mock).

Overexpression was most significantly observed in SW480 transfected with GAEC1

(more than 3000-fold increase) at the transcription level (Figure 6.1A). At the translation level, all cells overexpressing GAEC1 (SW48+GAEC1, SW480+GAEC1 and

FHC+GAEC1) showed a similar ratio of increment in protein expression when quantified with ImageJ software (Figure 6.1C).

6.2.5.2 Overexpression of GAEC1 induce cell proliferation, increased cell migration and clonogenic potential

Cells overexpressing GAEC1 was subjected to various cell-based assays to understand the biological response of these cells towards GAEC1. To this end, the current study found that overexpression of GAEC1 significantly increased the proliferation (P<0.05) of all cells in comparison to their respective mock control cells.

Increase in proliferation was observed across 72 hours with a similar trend in each cell line (Figure 6.2A). Additionally, all cells showed a significant increase in clonogenic potential (P<0.05) when overexpressing GAEC1 (Figure 6.2B-2C). The survival fraction increased 10% in SW48 (P<0.01) when compared between SW48+mock and

SW48+GAEC1 while the survival fraction of SW480 increased 6% (P<0.05) between

SW480+mock and SW480+GAEC1. Likewise, the survival fraction of FHC only had a small increase of 1% (P<0.05) between FHC+mock and FHC+GAEC1. In consistent with that, ectopic expression of GAEC1 also significantly enhanced the migration capabilities (P<0.001) of the cells as they showed a greater rate in wound healing compared to their mock counterpart (Figure 6.3).

182

Figure 6.2: Effect of GAEC1 overexpression on the growth rate of SW48, SW480 and FHC cells. (A) The rate of proliferation was measured 72 hours post-transfection and normalized against the mock counterpart. (B) Representative image of colony formation in each cell type either transfected with the mock vector of with GAEC1 plasmid. Cells were incubated over one week to ensure colony size formed consist of more than 50 cells each before analysis. (C) A number of colonies formed for each cell-lines either transfected with mock vector of or GAEC1 plasmid. Overexpression of GAEC1 significantly increased the proliferation rate and clonogenic potential of all cells at different pathological stages (FHC – non-neoplastic cells, SW480 – stage II carcinoma and SW48 – stage III carcinoma) in comparison to their respective mock control cells (cells transfected with an empty vector). Increased expression of GAEC1 induced a similar increase of clonogenic potential in SW48 and SW480 followed by FHC. The clonogenic figure shown is a representative of each cell line and the results shown are a mean ± standard deviation performed in triplicates. * indicates P<0.05 while ** indicates P<0.01.

183

Figure 6.3: Effect of GAEC1 overexpression on the wound healing potential of both colon cancer (A) SW48, (B) SW480 and non-neoplastic colon epithelial cells (C) FHC. All cells were observed across four days including day 0 when the scratch was done. In all three instances, ectopic expression of GAEC1 enhanced the wound closure rate in comparison to their mock counterpart. The rate of wound closure was highest in SW48, followed by SW480 and subsequently in FHC. Scratch test image shown are a representative of each cell line and the results shown are a mean ± standard deviation performed in triplicates. *** indicates P<0.001.

6.2.5.3 GAEC1 overexpression induced cell cycle changes

When cells were ectopically expressed with GAEC1, a significant increase of cells was observed in the G1 phase of stage III cancer (SW48+GAEC1 cells) (P<0.01) but not in non-neoplastic colonic epithelial cells (FHC+GAEC1) or stage II cancer

(SW480+GAEC1 cells) (Figure 6.4B). Whilst 39.07 ± 1.78% of the SW48+mock cell population is at S phase, overexpression of GAEC1 decreased the cell population to

28.87 ± 1.01% (P<0.01). This decrease is coupled with an increase in the cell population at G1 phase (from 33.53 ± 1.72% in SW48+mock to 43.03 ± 0.32% in

184

SW48+GAEC1; P<0.01) and G2 phase (from 25.87 ± 1.07% in SW48+mock to 27.93 ±

1.73% in SW48+GAEC1; P<0.05). These results indicate that regulation of the synthesis phase could be altered by the overexpression of GAEC1 in more advanced stage of colon adenocarcinoma (stage III vs stage II or non-neoplastic colonic epithelial cells).

Figure 6.4: Impact of GAEC1 overexpression on cell cycle kinetics of colon cancer cells (SW48 and SW480) and non-neoplastic colon epithelial cells (FHC) in (A) histogram and (B) bar graph. Ectopic expression of GAEC1 had no significant changes to stage II colon carcinoma cells (SW480) and non-neoplastic colonic epithelial cells (FHC). Significant increase of cell population at G1 phase and G2 phase coupled with a significant decrease in S phase were observed in stage III colon carcinoma cells (SW48). Histograms in this figure are a representative for each cell line and the results shown are a mean ± standard deviation performed in triplicates. * indicates P<0.05 while ** indicates P<0.01.

185

6.2.5.4 GAEC1 affects apoptosis in colon cancer cells

In Figure 6.5B, overexpression of GAEC1 resulted in increased the FHC+GAEC1 apoptotic population from 8.61 ± 0.04% to 9.87 ± 0.13% (P<0.001) compared to

FHC+mock. Whereas, overexpression of GAEC1 did not affect the apoptotic population of stage II (SW480+GAEC1) cells. Interestingly, stage III cancer (SW48+GAEC1) cells reacted differently by having reduced apoptotic population from 6.33 ± 0.10% to 4.91

± 0.03% (P<0.001) when compared to SW48+mock.

Figure 6.5: Outcome of GAEC1 ectopic expression on apoptosis of colon cancer cells in (A) dot plots and (B) graph. In the dot plots (A), Q1 represents dead cells, Q2 represents a late apoptotic stage, Q3 represents early apoptotic stage and Q4 represents healthy cells. GAEC1 overexpression significantly reduced the apoptotic population of SW48 cells (stage III colon carcinoma) while significantly increasing the apoptotic population of FHC cells (non-neoplastic colonic epithelial cells). No changes in the apoptotic population were observed in stage II colon carcinoma cells (SW480). Dot plots shown in this figure are representative dot plots for each cell line Results shown are a mean ± standard deviation performed in triplicates. *** indicates P<0.001.

186

6.2.5.5 Bioenergetics changes in colon cancer cells

Figure 6.6: Influence of GAEC1 overexpression on mitochondrial respiration regarding oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) in (A) SW48, (B) SW480 and (C) FHC cells and (D) coupling efficiency changes. GAEC1 significantly reduced the coupling efficiency of SW48+GAEC1 cells compared to SW48+mock cells (D). Results shown are a mean ± standard error performed in triplicates. * indicates P<0.05 whilst ** indicates P<0.0, and *** indicates P<0.001.

187

A mitochondrial stress assay was measured using the Seahorse XFp extracellular flux analyser to measure the impact of GAEC1 overexpression on mitochondrial function and glycolysis of colon cancer cells (Figure 6.6). After 24 hours of transfection, the coupling efficiency of SW48+GAEC1 cells (Figure 6.6D) significantly decrease (P<0.001) when compared to SW48+mock.

6.2.5.6 Downstream interaction of GAEC1 in colon cancer cells

Overexpression of GAEC1 successfully increased the expression of several key regulatory genes, which promotes the progression of cancer including pAkt, FOXO3 and MMP9 (Figure 6.7). A significant increase in pAkt (Figure 6.7A) and MMP9

(Figure 6.7C), were observed SW480+GAEC1 (P<0.05) and FHC+GAEC1 (P<0.01). In addition, an increase in FOXO3 (Figure 6.7B) was also observed in SW480+GAEC1 (not significant) and FHC+GAEC1 (P<0.01).

188

Figure 6.7: Overexpression of GAEC1 induced the expression of (A) pAkt, (B) FOXO-3 and (C) MMP9. (D) Western blot images of all protein expression changes caused by overexpression of GAEC1. Increase expression of pAkt and MMP9 are observed in SW480+GAEC1 cells and FHC+GAEC1 cells. An increased of FOXO3 protein was observed in FHC+GAEC1 cells. In all instances, overexpression was most significantly observed in FHC+GAEC1 cells, followed by SW480+GAEC1 cells and subsequently SW48+GAEC1 cells. Results shown are a mean ± standard deviation performed in triplicates. * indicates P<0.05 whilst ** indicates P<0.01.

6.2.6 Discussion

This study confirms the in vitro transforming properties of GAEC1 in human epithelial cells for the first time. Results were analogues to the GAEC1 transforming properties noted by Law and colleagues in vivo mouse models injected with fibroblasts cells overexpressing GAEC1 [1]. Overexpression of GAEC1 in non-neoplastic colonic epithelial cells (FHC+GAEC1) and colon cancer cells (SW48+GAEC1 and SW480+GAEC1) enhanced the capabilities of the cells to proliferate, migrate and form new colonies [21,

22]. A cell becomes malignant if it acquires such characteristics, as cell invasion is an initiator for cancer progression and metastasis [21]. These findings correlate well with our previous findings in which GAEC1 knockdown induced antigrowth properties in vitro and in vivo [6].

189

Despite only having a small increase in the survival fraction of FHC cells (1%) compared to SW48 (10%) and SW480 (6%), increased clonogenic capabilities in FHC represents enhanced capabilities to divide indefinitely, a key indicator of improved cell survival capabilities and thus, further highlighting the oncogenic potential of GAEC1.

As apoptosis is a well-known anti-cancer defence mechanism used [23], an initial increase of apoptotic population in FHC+GAEC1 cells when the non-neoplastic cells first undergo transformation into a cancerous cell is common. When apoptosis occurs in pre-cancerous cells, the sub-clones that are more aggressive can potentially repopulate the vacant niches and drive tumour evolution [23]. This can then lead to an increase in cell population as observed in Figure 6.2A.

Interestingly, alteration of cell cycle kinetics was only observed in SW48 cells upon ectopic expression of GAEC1 (Figure 6.4). ). A significant decrease in S phase coupled with an increase in G1 and G2 phase in SW48+GAEC1 implies that GAEC1 might be involved in the regulation of DNA synthesis at later stages of colon cancer progression. Compared to control cells, overexpression of GAEC1 seemed to decrease the cell population at S phase (28.87 ± 1.01%) by retaining 9.5% of the population at the G1 phase and pushing 2.5% of the population to G2 phase. These changes indicate that GAEC1 overexpression may have triggered other protein to coordinate multilayered cell cycle regulation in colon cancer. Such control has been previously reported for example in esophageal squamous cell carcinoma where inhibition or overexpression of miR-424 either inhibited or promoted the G1/S phase transition and

G2/M phase transition through modulating the activities of p21 [24]. However, further studies are needed to understand the mechanism by which GAEC1 affects the cell cycle of colon cancer.

To identify if GAEC1 was involved in the interaction between pAkt, FOXO3 and MMP9, the changes of all three proteins expression levels were observed in

190 response to GAEC1 overexpression. This study noted that MMP9, FOXO3 and pAkt were indeed significantly up-regulated in FHC over-expressed with GAEC1 (Figure

6.7). Akt is reported to inactivate the tumour suppressive properties of FOXO3 protein by promote its phosphorylation and nuclear export [12]. Taken together, these changes suggest that GAEC1 interacts with pAkt, FOXO3 and MMP9 to promote tumour progression in colon cancer. To-date, Akt has been a target for new cancer therapeutic target. Despite multiple developed Akt inhibitors, none of them are approved for oncology use [25] as they fail to achieve satisfactory therapeutic effects or cause adverse effects [26]. These Akt inhibitors commonly targeted the major components within the Akt pathway such as PI3K (phosphoinositide 3-kinase), PDK1 and mTOR

[27]. Perhaps, future studies could target the GAEC1 in the efforts to aid a successful development of Akt inhibitor.

Mitochondrial respiration is vital for energy production on a normal cellular basis. A compile analysis by Zong and colleagues (2009) confirms that mitochondrial biogenesis is frequently upregulated in cancer [28]. In this study, the coupling efficiency of SW48+GAEC1 cells dropped in comparison with SW48+mock cells (Figure

6.6D). This indicates a reduced in ATP production via mitochondria respiration in

SW48+GAEC1 cells. Indeed, cancer cells can undergo metabolic reprogramming and enhance its energy production by utilizing glucose, a scenario known as the Warburg effect [20, 29, 30]. Enhanced glycolysis not only boost ATP production but also generates glycolytic intermediates, which are crucial for aggressive cancers such as colon and stomach cancer [31, 32].

To conclude, the current study confirms the oncogenic properties of GAEC1 via overexpression studies in colon cancer. In addition, the transforming potential of

GAEC1 in non-neoplastic colonic epithelial cells (FHC) was confirmed for the first time. GAEC1 affects mitochondrial respiration by reducing the coupling efficiency in

191 colon adenocarcinoma of the advanced stages (stage III with lymph node metastases).

GAEC1 is involved in the interaction between pAkt, FOXO3 and MMP9 in colon adenocarcinoma and may drive the cancer progression through the positive feedback regulation of Akt/ FOXO-3/MMP-9 signalling pathway.

6.2.7 Conflict of Interest

The authors have no conflict of interest in publishing this manuscript.

6.2.8 Acknowledgement

The authors would like to thank Dr. Cameron Flegg for sharing his expert knowledge in microscopy. We also would like to acknowledge the paper completion assistance from School of Medical Science, Griffith University. Additionally, the project was supported by the HDR scholarship from Griffith University and funding from Menzies

Health Institute Queensland.

6.2.9 References

1. Tang WK, Chui CH, Fatima S, et al. Oncogenic properties of a novel gene JK-1 located in chromosome 5p and its overexpression in human esophageal squamous cell carcinoma. International journal of molecular medicine. 2007;19(6):915-923. 2. Lee KT, Gopalan V, Islam F, et al. GAEC1 mutations and copy number aberration is associated with biological aggressiveness of colorectal cancer. European journal of cell biology. 2018;97(3):230-241. 3. Gopalan V, Yasuda K, Pillai S, et al. Gene amplified in oesophageal cancer 1 (GAEC1) amplification in colorectal cancers and its impact on patient's survival. J Clin Pathol. 2013;66(8):721-723. 4. Gopalan V, Smith RA, Nassiri MR, et al. GAEC1 and colorectal cancer: a study of the relationships between a novel oncogene and clinicopathologic features. Human pathology. 2010;41(7):1009-1015. 5. Wahab R, Gopalan V, Islam F, et al. Expression of GAEC1 mRNA and protein and its association with clinical and pathological parameters of patients with colorectal adenocarcinoma. Experimental and molecular pathology. 2018;104(1):71-75. 6. Islam F, Gopalan V, Wahab R, et al. Novel FAM134B mutations and their clinicopathological significance in colorectal cancer. Human genetics. 2017;136(3):321-337.

192

7. Storz P, Doppler H, Copland JA, et al. FOXO3a promotes tumor cell invasion through the induction of matrix metalloproteinases. Molecular and cellular biology. 2009;29(18):4906-4917. 8. Li H, Zhang K, Liu L-h, et al. A systematic review of matrix metalloproteinase 9 as a biomarker of survival in patients with osteosarcoma. Tumor Biology. 2014;35(6):5487-5491. 9. Egeblad M, Werb Z. New functions for the matrix metalloproteinases in cancer progression. Nature reviews Cancer. 2002;2(3):161-174. 10. Xu K, Pei H, Zhang Z, et al. FoxO3a mediates glioma cell invasion by regulating MMP9 expression. Oncology reports. 2016;36(5):3044-3050. 11. Hornsveld M, Dansen TB, Derksen PW, et al. Re-evaluating the role of FOXOs in cancer. Seminars in cancer biology. 2018;50:90-100. 12. Tenbaum SP, Ordonez-Moran P, Puig I, et al. beta-catenin confers resistance to PI3K and AKT inhibitors and subverts FOXO3a to promote metastasis in colon cancer. Nature medicine. 2012;18(6):892-901. 13. Myatt SS, Lam EW. The emerging roles of forkhead box (Fox) proteins in cancer. Nature reviews Cancer. 2007;7(11):847-859. 14. Qian Z, Ren L, Wu D, et al. Overexpression of FoxO3a is associated with glioblastoma progression and predicts poor patient prognosis. International Journal of Cancer. 2017;140(12):2792-2804. 15. Santamaria CM, Chillon MC, Garcia-Sanz R, et al. High FOXO3a expression is associated with a poorer prognosis in AML with normal cytogenetics. Leukemia research. 2009;33(12):1706-1709. 16. Kumazoe M, Takai M, Bae J, et al. FOXO3 is essential for CD44 expression in pancreatic cancer cells. Oncogene. 2016;36:2643. 17. Chen J, Gomes AR, Monteiro LJ, et al. Constitutively Nuclear FOXO3a Localization Predicts Poor Survival and Promotes Akt Phosphorylation in Breast Cancer. PLOS ONE. 2010;5(8):e12293. 18. Franken NA, Rodermond HM, Stap J, et al. Clonogenic assay of cells in vitro. Nature protocols. 2006;1(5):2315-2319. 19. Yang X. Clonogenic Assay. Bio-protocol. 2012;2(10):e187. 20. Gopalan V, Ebrahimi F, Islam F, et al. Tumour suppressor properties of miR-15a and its regulatory effects on BCL2 and SOX2 proteins in colorectal carcinomas. Experimental cell research. 2018;370(2):245-253. 21. Krakhmal NV, Zavyalova MV, Denisov EV, et al. Cancer Invasion: Patterns and Mechanisms. Acta naturae. 2015;7(2):17-28. 22. Friedl P, Gilmour D. Collective cell migration in morphogenesis, regeneration and cancer. Nature reviews Molecular cell biology. 2009;10(7):445-457. 23. Labi V, Erlacher M. How cell death shapes cancer. Cell death & disease. 2015;6:e1675. 24. Wen J, Hu Y, Liu Q, et al. miR-424 coordinates multilayered regulation of cell cycle progression to promote esophageal squamous cell carcinoma cell proliferation. EBioMedicine. 2018. 25. Nitulescu GM, Margina D, Juzenas P, et al. Akt inhibitors in cancer treatment: The long journey from drug discovery to clinical use (Review). International journal of oncology. 2016;48(3):869-885. 26. Lee KT, Gopalan V, Lam AK. Roles of long-non-coding RNAs in cancer therapy through the PI3K/Akt signalling pathway. Histology and histopathology. 2019:18081. 27. Carnero A. The PKB/AKT pathway in cancer. Current pharmaceutical design. 2010;16(1):34-44.

193

28. Zong W-X, Rabinowitz JD, White E. Mitochondria and Cancer. Molecular cell. 2016;61(5):667-676. 29. Jang M, Kim SS, Lee J. Cancer cell metabolism: implications for therapeutic targets. Experimental & molecular medicine. 2013;45:e45. 30. He Z, Li Z, Zhang X, et al. MiR-422a regulates cellular metabolism and malignancy by targeting pyruvate dehydrogenase kinase 2 in gastric cancer. Cell death & disease. 2018;9(5):505. 31. Abbassi-Ghadi N, Kumar S, Huang J, et al. Metabolomic profiling of oesophago- gastric cancer: a systematic review. European journal of cancer (Oxford, England : 1990). 2013;49(17):3625-3637. 32. Hirayama A, Kami K, Sugimoto M, et al. Quantitative metabolome profiling of colon and stomach cancer microenvironment by capillary electrophoresis time-of- flight mass spectrometry. Cancer Res. 2009;69(11):4918-4925.

194

Chapter 7: Functional role and interacting partner of FAM134B

7.1 Preface

This chapter was done in collaboration with other team members. The section on the copy number change and mutation status were part of the PhD dissertation of another higher degree by research (HDR) candidate and has been published [107]. In summary, FAM134B mutation was identified by Sanger sequencing. Similarly, to

Chapter 5, the mutations of these genes were then tested for any clinicopathological correlations. We found that a total of 46.5% (41/88) patients with colorectal cancer harboured FAM134B mutations. These mutations included 31 potentially pathogenic novel mutations found in both the intronic and coding regions of FAM134B. The bulk of these mutations were single-nucleotide substitutes. Through computational analysis, we found that eight novel frameshift mutations showed potential to cause non-sense mediated mRNA decay (NMD). These mutations were significantly associated with various clinical and pathological features including, tumour size and staging. A detailed description of this study can be found at Appendix G.

For this part of the dissertation, the main focus is to identify the cell restoration roles of FAM134B gene in colon cancer. All FAM134B research done on colon cancer to date focused on knockdown of FAM134B [79, 86, 88]. No studies have been done to confirm the tumour suppressive of FAM134B in colon cancer via overexpressing

FAM134B. Furthermore, FAM134B is known to affect endoplasmic reticulum (ER) stress [68, 108, 109]. In addition, significant associations have been made between endoplasmic reticulum (ER) stress and its effects on mitochondrial function [110]. It is therefore of interest to identify if FAM134B can affect the mitochondrial function of colon cancer. In addition, literature has provided evidence that both FAM134B and

MIF are frequently associated with inflammatory diseases such as allergy rhinitis [111,

195

112]. Since cancer is often associated with chronic inflammation, this study focused on identifying the association of FAM134B and MIF.

196

STATEMENT OF CONTRIBUTION TO CO-AUTHORED PUBLISHED PAPER

Chapters 7 includes a co-authored paper. The bibliographic details of the co-authored paper, including all authors, are:

Title: Family with sequence similarity 143, member B (FAM134B) overexpression in colon cancers and its tumor suppressor properties via regulation of macrophage migration inhibitory factor (MIF) and other partners

Authors: Katherine Ting-Wei Lee, Farhadul Islam, Anna Chruścik, Jeremy Martin, Jelena Vider, Vinod Gopalan, Alfred King-yin Lam

Name of the journal: BBA Molecular Cell Research

Status: Under review

Publication link: https://ees.elsevier.com/bbamcr/default.asp?pg=login.asp%253FloginError%253D1

My contribution to the paper involved: I drafted the manuscript, optimised and performed all the cell culture, plasmid transfection qPCR, western blot and cell based assays. I was also involved in the data analysis and statistical analysis of this manuscript.

(Signed) ______(Date)___ 22/11/2018__ Katherine Ting-Wei Lee

(Countersigned) ______(Date)___ 22/11/2018__ Corresponding author of paper and principal supervisor: Prof. Alfred King-Yin Lam

(Countersigned) ______(Date)__ 22/11/2018___ Corresponding author of paper and principal supervisor: Dr. Vinod Gopalan

197

7.2 Detailed description of Family with sequence similarity 143, member B (FAM134B) overexpression in colon cancers and its tumor suppressor properties via regulation of macrophage migration inhibitory factor (MIF) and other partners

Running title: FAM134B and MIF in colon cancer

Katherine Ting-Wei Lee1, Farhadul Islam1,2, Anna Chruścik1, Jelena Vider3, Jeremy Martin1, Cu-tai Lu4, Vinod Gopalan1, 3, Alfred King-yin Lam1*

1Cancer Molecular Pathology, School of Medicine, Griffith University, Gold Coast, Queensland, 4222, Australia 2 Department of Biochemistry of Molecular Biology, University of Rajshahi, 6205, Bangladesh 3School of Medical Science, Menzies Griffith University, Gold Coast, Queensland, Australia 4Department of Surgery, Gold Coast University Hospital, Gold Coast, Queensland, Australia

Professor Alfred K Lam, Head of Pathology, Griffith Medical School, Gold Coast Campus, Gold Coast QLD 4222, Australia. E-mail: [email protected] Telephone +61 7 56780718; Fax +61 7 56780303

Abbreviations SW48+mock - SW48 transfected with a blank vector (control SW48 cells) SW480+mock - SW480 transfected with a blank vector (control SW480 cells) FHC+mock - FHC transfected with a blank vector (control FHC cells)

SW48+FAM134B - SW48 transfected with a FAM134B vector SW480+FAM134B - SW480 transfected with a FAM134B vector FHC+FAM134B - FHC transfected with a FAM134B vector

198

7.2.1 Abstract

FAM134B (family with sequence similarity 143, member B) is a tumour suppressor associated with Wnt signalling pathway in the pathogenesis of colon cancer.

This study aims to investigate the overexpression-induced properties of FAM134B in colon cancer, examine the potential interaction partners of FAM134B and its impact on mitochondrial function. FAM134B was overexpressed in colon cancer and non- neoplastic colonic epithelial cells. Various cell-based assays including apoptosis, cell cycle, cell proliferation, clonogenic, extracellular flux and wound healing assays were performed. Western blot analysis was used to confirm and identify potential interacting partners of FAM134B in vitro. Immunochemistry and real-time PCR were employed to determine the expression of MIF and FAM134B expression respectively on 63 patients with colorectal carcinoma and the clinicopathological features were compared.

Results showed that FAM134B is involved in apoptosis and mitochondrial function of colon cancer. Overexpression of FAM134B led to increased expression levels of EB1, p53 and MIF suggesting its potential interactions in vitro. In approximately 70% of patients with colorectal cancer, low level of FAM134B exhibited high levels of MIF while the remaining 30% patients showed a concurrent expression of FAM134B and

MIF (P=0.045). High expression of MIF coupled with low expression of FAM134B is associated with microsatellite instability status in colon cancer (P =0.049). In conclusion, loss expression of FAM134B results in more aggressive colorectal cancer.

FAM134B is likely to exert its tumour suppressive function through controlling apoptosis, mitochondrial function via potentially interacting with MIF, p53 and EB1.

7.2.2 Keywords

FAM134B; MIF; p53; EB1; mitochondrial function

199

7.2.3 Highlights

• FAM134B is involved in apoptosis and mitochondrial function of colon

carcinoma.

• FAM134B overexpression increased expression of EB1, p53 and MIF.

• In vitro and pathological experiments showed interaction between FAM134B

and MIF.

• Low FAM134B and high MIF expressions often in MSI stable carcinoma.

7.2.4 Introduction

Family with sequence similarity 143, member B (FAM134B) (accession no.

NM_019000) also known as JK-1 or reticulophagy regulator 1 (RETEG1) belongs to a family of three genes: FAM134A, FAM134B and FAM134C [1, 2]. FAM134B helps maintain homeostasis by regulating the endoplasmic reticulum (ER) turnover

[3, 4]. Changes of FAM134B expression is linked to several diseases including inflammation [5, 6], allergic rhinitis [6] and cancer [7-14], which are highly associated with inflammation. In cancer, FAM134B showed dual roles in different type of cancers either functioning as a tumor suppressor gene [7-13] or an oncogene

[14]. For example, FAM134B functioned as an oncogene in esophageal squamous cell carcinoma [14] but with tumor suppressive features in breast carcinoma [13] and colorectal adenocarcinoma [7-12].

FAM134B mutations frequently occurred in esophageal squamous cell carcinoma [15] and colon adenocarcinoma [11]. Interestingly, these mutations were frequently correlated with bio-aggressiveness of the cancers [11, 15], further strengthening the association of FAM134B in the pathogenesis of these cancers.

Moreover, knockout or hypermethylation of FAM134B were associated with advanced pathological stages of colon cancer [8, 10, 16]. In addition, suppression of

200

FAM134B significantly enhanced the cellular capabilities of colon cancer in terms of cell proliferation, migration and wound healing capabilities [7, 10]. A similar observation occurred in immune-deficient mice xenotransplanted with colon cancer cells with knockout of FAM134B [10].

In the early 2018, interacting partners of FAM134B were identified for the first time, locating the FAM134B within a gene-signaling network in colon cancer.

Physical co-localization of FAM134B with EB1 (Microtubule Plus-end Tracking of

End-binding Protein 1), CAP1 (Adenylyl cyclase-associated protein 1), CyPB

(cyclophilin B) and KDELR2 (KDEL Endoplasmic Reticulum Protein Retention

Receptor 2) were observed [17]. Suppression of FAM134B in colon cancer cells significantly up-regulated EB1 expression and down-regulated KDELR2 expression but did not change the expression of CAP1 and CyPB [17]. Further testing of related proteins indicated that knockout of FAM134B also reduced the expression of APC

(adenomatous polyposis coli) and increased the expression of β-catenin in SW48 cells and noted the involvement of FAM134B in the regulation of WNT/β-catenin pathway via EB1 and APC in colon cancer [17].

Despite having several studies addressing, the biological and molecular function of FAM134B in colon cancer, none has attempted to investigate the overexpression-induced effects of FAM134B colon cancer cells. This study focused on addressing these issues to confirm the tumor suppressor properties and gene interactions of FAM134B following its restoration/overexpression in colon cancer cells.

This study also aims to identify if FAM134B affects the mitochondrial function in colon cancer. In addition, this study will examine the association of FAM134B with the mitochondrial function of colon cancer cells for the first time.

201

7.2.5 Materials and Methods

7.2.5.1 Recruitment of tissue samples and clinicopathological data

Patients with surgical resection of colorectal cancer were recruited in the study with their written consent between 2012 to 2014 at Queensland-based hospitals.

Cancer tissues and matched non-neoplastic mucosae near the surgical resection margin

(as controls) were obtained fresh from these patients. Ethical approval (GU Ref No:

MSC/17/10/HREC) was obtained for the use of these tissues from Griffith University

Human Research Ethical Committee. Samples were snapped frozen and stored in -

80°C until use. The other portion of each of these specimens were sampled by standard pathology protocol. A pathologist (AKL) examined all the pathological features of the colorectal carcinomas. Colorectal carcinomas are graded and typed according to the

World Health Organization (WHO) criteria [18] and staged according to the tumor, lymph node and metastases (TNM) classification [19]. In each case, the pathologist examined the lymphovascular invasion by carcinoma. All the cancer tissues had microsatellite instability (MSI) status studied by immunochemistry [16]. Carcinoma with loss of marker(s) was termed “MSI high” whereas carcinoma positive for all markers was “MSI stable”. After examination, 63 patients with carcinomas were included for analysis.

7.2.5.2 Cell culture

One non-cancer colon epithelial cell line FHC (ATCC CRL-1831) and two colon cancer cell lines SW480 (ATCC CCL228; derived from stage II, colonic adenocarcinoma) and SW48 (ATCC CCL-231; derived from stage III colonic adenocarcinoma) were utilized in this research. All cells were maintained according to our previously published method [20] and routinely tested for mycoplasma contamination.

202

7.2.5.3 Plasmid transfection

A GMO (genetically modified organism) project approval was obtained from

Griffith University Institutional Biosafety Committee (NLRD/006/15). FAM134B

(Myc-DDK-tagged) transcript variant two (NM_019000) were transiently overexpressed in all cell lines using jetPRIME® Polyplus transfection reagent

(Polyplus-transfection® SA, Illkirch, France) according to the manufacturer’s protocol.

Both the control plasmid (blank vector) and the FAM134B plasmid are constructed using the pCMV6-Entry vector (OriGENE, Rockville, MD, USA). Proteins and mRNAs from post-transfected cells were harvested and analyzed via quantitative polymerase chain reaction (qPCR) and Western blot respectively to confirm overexpression at both the transcription and translation level.

7.2.5.4 Quantitative real time PCR (qPCR) analysis

Total RNA was extracted from all cells and tumor tissue using the RNeasy mini kit (Qiagen, Hilden, Germany) according to the manufacturer’s guideline. Reverse transcription was then performed using the iScriptTM cDNA Synthesis Kit (Bio-Rad

Laboratories, Hercules, CA, USA) and the expression changes of FAM134B in the different cell lines were examined using Quantistudio Flex 6 qPCR system (Thermo

Fisher Scientific, Waltham, MA, USA). Samples were prepared using the SensifastTM

SYBR No-Rox mix (Bioline, London, UK) according to the manufacturer’s guideline.

Cycling conditions were set to an initial denaturation of 95°C for 2 minutes followed by 40 cycles of denaturation (95°C) for five seconds and annealing extension at 60°C for 30 seconds followed by a melt curve analysis. Amplification efficiency of

FAM134B was normalised against the control gene GAPDH (NM_002046) with primer details as shown in Supplementary Table 7.1.

203

Supplementary Table 7.1: Primer design for quantitative real time PCR (qPCR) analysis Target Genes Primers Amplicon size (GenBank number) (base pairs) FAM134B 5′-AGACTTTTCAGCTCTTTGTC-3 155 (NM_019000) 5′-GATCATCAGAAGTGTCTGTC-3′ GAPDH 5′-TGCACCACCAACTGCTTAGC-3′ 88 (NM_002046) 5′-GGCATGGACTGTGGTCATGAG-3′

7.2.5.5 Western blot analysis

Total protein was extracted using NP40 lysis buffer (Thermo Fisher Scientific) supplemented with 2% protease inhibitor (Sigma-Aldrich, St. Louis, MO, USA) in accordance with the recommended guidelines. Proteins were then quantified using the

Pierce BCA Protein Assay Kit (Thermo Fisher Scientific). A total of 30µg of protein was loaded onto a 4-15% Mini-PROTEAN® TGX™ Precast Gels (Bio-Rad

Laboratories) for separation and transferred onto a polyvinylidene fluoride (PVDF) membrane. The membrane was then blocked with 5% non-fat milk powder for one hour at room temperature before overnight incubation of primary antibody at 4℃. The primary antibodies used included FAM134B (1:4000) (Promab, Biotechnologies,

Richmond, CA, USA), GAPDH (1:1000) (Santa Cruz, Dallas, TX, USA), p53 (1:1000)

(Cell Signaling Technology, Danvers, MA, USA), MIF (1:200) (Santa Cruz), EB1

(1:100) (Santa Cruz), CyPB (Santa Cruz) and β-catenin (Thermo Fisher Scientific).

The next day, the membrane was washed with PBST (phosphate buffered saline with

Tween 20) and incubated with the relevant secondary antibodies (1: 10,000) (Abcam,

Cambridge, UK) at room temperature for one hour. The blots were then developed using SuperSignalTM West Pico PLUS Chemiluminescent Substrate (Thermo Fisher

Scientific) and visualized with the ChemiDoc MP TM Imaging System (Bio-Rad

Laboratories).

204

7.2.5.6 Immunohistochemistry assay

Immunohistochemistry was used to identify the presence of MIF expression.

Formalin-fixed paraffin-embedded (FFPE) tissue samples were sectioned into 4µm and stained according to our previously published protocol [21].

7.2.5.7 Cell proliferation assay

Transfected cells were left to incubate over 72 hours in a 5% CO2 supplemented incubator at 37°C before adding 10µL of 12mM MTT (3-(4,5-

Dimethylthiazol-2-yl)-2,5-Diphenyltetrazolium Bromide). Then, the cells were incubated for an additional four hours. The crystal formazan formed were dissolved with dimethyl sulfoxide (DMSO). The absorbance of each well was read at 540nm wavelength. All samples were run in three independent replicates and the results were normalized against the control group.

7.2.5.8 Clonogenic assay

A total of 200 cells were seeded into a 6-well plate with 2mL of complete media and maintained in a 5% CO2 supplemented incubator at 37°C. Cells were routinely checked and when cells in the control plates had reached a minimum of 50 cells/colony

(minimum scoring) [22, 23], the medium was aspirated, and the cells were washed with PBS. Then, cells were fixed using fixation solution (acetic acid/methanol 1:7, vol/vol) for five minutes and subsequently incubated with 0.5% crystal violet solution at room temperature for two hours. After the incubation, the cells were washed with tap water and left to air dry. Finally, images of these wells were taken, and the clone formation rates were calculated [22].

205

7.2.5.9 Wound healing assay

A P-200 tip was used to scratch the center of the well. Wound areas were imaged every 24h thereafter and media were changed routinely for up to four days.

All wound areas were measured and compared using ImageJ 1.50I. All samples were run in three independent replicates and the results were normalized against the control group.

7.2.5.10 Cell apoptosis assay

A million cells were pre-washed with PBS before re-suspending in Annexin V binding buffer. Cells were later stained with 2.5µL of Alexa Fluor® 350 (Invitrogen,

Carlsbad, CA, USA) and 25ng/mL of propidium iodide solution. A flow cytometer

(BD Fortessa, BD Biosciences, Frankin Lakes, NJ, USA) was used to read the prepared cells. The analysis was done with FlowJoTM 10. Results shown are representative of three independent experiments.

7.2.5.11 Cell cycle analysis

A million cells were fixed in ice-cold 70% ethanol for at least 2 hours. Next, these cells were pelleted and washed with PBS before staining with 500µL of propidium iodide (PI) solution (10µg/mL) containing 10 µg/mL RNase A. Cells were analyzed with the flow cytometer. All samples were run in three independent replicates and the results were normalized against the control group.

7.2.5.12 Extracellular flux assay

Cells were plated onto a Seahorse XFp Cell Culture Miniplate to 80% confluency prior to performing the assay. On the day of the experiment, media were replaced with Agilent Seahorse XF Base Medium supplemented with 1mM pyruvate,

206

2mM glutamine and 0.11mM glucose, pH 7.4. Cells were then incubated for 40 minutes at 37℃ before transferring onto a Seahorse XFp extracellular flux analyzer

(Agilent, Santa Clara, CA, USA). They were sequentially stressed with 1.0µM oligomycin, 1.0µM FCCP (carbonyl cyanide p-trifluoromethoxy-phenylhydrazone) and 0.5µM rotenone/antimycin. All results were normalized according to the protein concentration in each well. Results shown are representative of three independent experiments.

7.2.5.13 Statistical analysis

All statistical analysis was carried out using IBM (New York, NY, USA) SPSS

Statistics version 25. Correlation of FAM134B mRNA expression and MIF protein expression with different clinicopathological parameters were performed using the chi-square test, likelihood ratio and Fisher’s exact test. GraphPad Prism 7 was used to plot the graphs and error bars. All flow cytometry analyses were performed with

FlowJoTM 10. Cell cycle analysis was performed using Dean-Jett Fox model found within the FlowJoTM 10 software. ImageJ was used to quantitate all bands and qualitative images. The significance level was taken at P<0.05.

207

7.2.6 Results

7.2.6.1 FAM134B overexpression in vitro

Overexpression of FAM134B using the FAM134B (Myc-DDK-tagged) plasmid was confirmed at both the transcriptional and translational level via both real-time PCR

(qPCR) and Western blot analysis (Figure 7.1). All three colon cell lines

(SW48+FAM134B, SW480+FAM134B and FHC+FAM134B) expressed significantly elevated

FAM134B expression of mRNA (Figure 7.1A) and protein (Figure 7.1B and C) levels when normalised against GAPDH control gene compared to their mock counterpart

(SW48+mock, SW480+mock and FHC+mock).

Figure 7.1: Overexpression of FAM134B in colon cancer cells (SW48 and SW480) and non-neoplastic colon epithelial cells (FHC). A significant increase in the expression of FAM134B occurred in all cell lines with both real-time PCR (A) and Western blot analysis (B). Western blot bands obtained were quantified using ImageJ. Results in (A) and (C) were plotted as a mean ± standard deviation with ** indicates P<0.01 and *** indicate P<0.001 when compared to both each control and mock vector respectively. Mock vector indicates cells transfected with an empty vector.

208

7.2.6.2 FAM134B reduces apoptosis and affects cell kinetic changes in colon cancer

Overexpression of FAM134B did not lead to a significant change to cell proliferation, colony formation (Supplementary Figure 7.1) and wound healing rate

(Supplementary Figure 7.2). However, as seen in Figure 7.2 overexpression of

FAM134B significantly reduced the apoptotic rate in cancer cell (SW48+FAM134B) derived from stage III colon cancer from 5.83 ± 0.25% to 3.32 ± 0.1% (P<0.001) and non-neoplastic colonic epithelial cells (FHC+FAM134B) from 9.78 ± 0.07% to 8.51 ± 0.16%

(P<0.001).

Supplementary Figure 7.1: Effect of FAM134B overexpression on the growth potential of SW48, SW480 and FHC cells. The proliferation rate of each cell line was measured 72h post transfection and normalised against their individual mock counterpart. (B) Representative image of colony formation in each cell type either transfected with the mock vector of with FAM134B plasmid are shown. The cells were incubated until colony size formed more than 50 cells each before analysis. (C) A number of colonies formed for each cell-lines either transfected with the mock vector or the FAM134B plasmid. Overexpression of FAM134B did not significantly change the proliferation rate or clonogenic potential of these cell lines. The clonogenic figure shown is a representative of each cell line and the results showed are a mean ± standard deviation performed in triplicates. 209

Supplementary Figure 7.2: Effect of GAEC1 overexpression on the wound healing potential of both colon cancer (A) SW48, (B) SW480 and non-neoplastic colon epithelial cells (C) FHC. All the cells were observed across four days including day 0 when the scratch was done. In all three instances, ectopic expression of FAM134B did not enhance the wound closure rate in comparison to their mock counterpart. Scratch test image shown are a representative of each cell line.

210

Figure 7.2：Outcome of FAM134B ectopic expression on apoptosis of colon cancer cells in (A) dot plots and (B) graph. In (A) the dot plots, Q1 represents dead cells, Q2 represents late apoptotic stage, Q3 represents early apoptotic stage and Q4 represents healthy cells. FAM134B overexpression significantly reduced the apoptotic population of SW48 cells (stage III colon cancer). No changes in apoptotic population occurred in stage II colon cancer (SW480) and non-neoplastic colonic epithelial cells (FHC cells). Dot plots shown in this figure are representative dot plots for each cell line. Results shown are a mean ± standard deviation performed in triplicates. *** indicates P<0.001.

In addition, overexpression of FAM134B significantly changed the cell cycle kinetics of SW48, SW480 and FHC cells. As seen in Figure 7.3, over expression of

FAM134B significantly increased the cell population proportion at G1 phase (from

37.02 ± 0.92 % to 40.67 ± 0.06 %; P<0.05) while reducing the cell population proportion at G2 phase (from 27.93 ± 0.38 % to 25.37 ± 0.57 %; P<0.05) in

SW48+FAM134B cells when compared to its mock counterpart. Conversely, in

FHC+FAM134B cells, accumulation of cells at the G1 phase was significantly reduced

(from 55.17 ± 0.30% to 51.93 ± 0.42%; P<0.01) compared to FHC+mock. Also, a moderate increase in cell proportion at G2 phase (from 10.83 ± 0.25 % to 12.33 ±

211

0.15 %; P<0.05) was noted with these cells. Interestingly, in SW480+FAM134B colon cancer cells, overexpression of FAM134B significantly reduced the cell population proportion at S phase (from 33.35 ± 4.47 % to 24.45 ± 0.17 %; P<0.05). This reduction is parallel with a slight increase in cell population proportion at G1 phase (from 50.43

± 4.8% to 52.80 ± 0.96%) and a significant increase in cell population proportion at

G2 phase (from 18.50 ± 0.45% to 22.93 ± 1.17%) compared to its mock counterpart.

Figure 7.3: Impact of FAM134B overexpression on cell cycle kinetics of colon cancer cells (SW48 and SW480) and non-neoplastic colon epithelial cells (FHC) in (A) histogram and (B) bar graph. Ectopic expression of FAM134B resulted in accumulation of SW48 population at the G1 phase and a reduction in the G2 phase. Whereas, in SW480 cells, a reduction of cell population at the S phase was observed with enforced expression of FAM134B. On the other hand, exogenous expression of FAM134B reduced FHC cell population at the G1 phase and increased cell population at the G2 phase. Histograms in this figure are a representative for each cell line and the results showed are a mean ± standard deviation performed in triplicates. * indicates P<0.05 while ** indicates P<0.01.

212

7.2.6.3 FAM134B affects the bioenergetics changes in colon cancer

A mitochondrial stress assay was performed to determine the effects of

FAM134B on mitochondrial function and glycolysis in colon cancer cells. After 24 hours of transfection, significant bioenergetics changes were observed in cancer cells

(SW48+FAM134B and SW480+FAM134B) but not in non-neoplastic colon epithelial cells

(FHC+FAM134B). Figure 7.4D shows a reduction in non-mitochondrial oxygen consumption (from 4.28 ± 1.23% to 2.40 ± 1.44%; P<0.05), basal respiration (from

5.84 ± 1.28% to 2.56 ± 1.19%; P<0.05) and ATP production (4.97 ± 1.12% to 2.1 ±

0.94%; P<0.05) in SW480+FAM134B cells. On the other hand, overexpression of

FAM134B in SW48+FAM134B cells significantly reduced the extracellular acidification rate (ECAR) (from 13.33 ± 3.15% to 6.48 ± 2.88%; P<0.05) under stressful conditions

(Figure 7.4E).

213

Figure 7.4: Influence of FAM134B overexpression on mitochondrial respiration of colon cancer. Changes after FAM134B overexpression regarding oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) in (A) SW48, (B) SW480 and (C) FHC cells. The changes of (D) non-mitochondrial oxygen consumption rate, basal respiration and ATP production; and, (E) the metabolic potential of stressed OCR and EACR were noted. FAM134B significantly reduced the non-mitochondrial oxygen consumption, basal respiration and ATP production of SW480 cancer cells (D). A reduction of OCR rate and EACR rate were observed when SW48 cells were stressed in the presence of FAM134B. No significant changes were observed in FHC cells. Results shown are a mean ± standard error performed in triplicates. * indicates P<0.05 while ** indicates P<0.0, and *** indicates P<0.001. 214

7.2.6.4 FAM134B promotes the expression of MIF, EB1 and p53

FAM134B overexpression significantly increased the expression level of MIF

(Figure 7.5A) and p53 (Figure 7.5D) across all three cell lines (SW48+FAM134B,

SW480+FAM134B and FHC+FAM134B). EB1 (Figure 7.5B) protein expression was also noted to increase with SW480+FAM134B. Interestingly, as seen in Figure 6A, in patients with colorectal cancer, approximately 70% (n=46/63) of them showed high expression of MIF in the presence of low expression of FAM134B while the other 30% (n=17/63) showed expression of MIF in the presence of high expression of FAM134B (P<0.05).

Expression of MIF was categorized into positive staining (Figure 7.6B) and negative staining (Figure 7.6C). Whereas, no significant changes in the expression of CyPB

(Figure 5C) and β-catenin (Figure 5E) were observed with FAM134B overexpression.

Figure 7.5: Overexpression of FAM134B successfully changed the protein expression of colon cancer cells. Exogenous expression of FAM134B induced the expression of (A) MIF, (B) EB1, and (D) p53. No expression changes were observed for (C) CyPB and (E) β-catenin. Results shown are a mean ± standard deviation performed in triplicates. ** indicates P<0.01 and *** indicates P<0.001. 215

Among the samples collected from 63 patients, 65% (n = 41/63) showed low expression of FAM134B coupled with the expression of MIF (Table 7.1). In these 41 patients (23 mean; 18 women), the expression pattern was often present in MSI stable carcinomas (P=0.049). Other than this, the expression pattern did not correlate with other clinical pathological parameters in this group of patients.

Figure 7.6: Association between MIF and FAM134B in patients with colorectal cancer and its representative expression of MIF in colorectal adenocarcinoma at x25 magnification. (A) Shows the association between MIF and FAM134B mRNA expression in patients with colorectal cancer. (B) Represents moderate to high staining of MIF whereas (C) represents low to negative staining of MIF.

216

Table 7.1: Statistical correlation between low FAM134B mRNA expression in the presence of MIF and the clinicopathological features of Australian patients with colorectal cancer Characteristics No. of Low FAM134B mRNA P value patients expression with MIF expression No Yes Sex Female 18 (43.9%) 2 (11.1%) 16 (88.9%) 0.187 Male 23 (56.1%) 0 (0.0%) 23 (100.0%) Age <70 25 (61.0%) 1 (4.0%) 24 (96.0%) 0.634 ≥70 16 (39.0%) 1 (6.3%) 15 (93.8%) Site Proximal colon 19 (46.3%) 2 (10.5%) 17 (89.5%) 0.209 Distal colorectum 22 (53.7%) 0 (0.0%) 22 (100.0%) Colon cancer type Conventional 37 (90.2%) 1 (2.7%) 36 (97.3%) 0.188 Mucinous 4 (9.8%) 1 (25.0%) 3 (75.0%) Grade Well 4 (9.8%) 0 (0.0%) 4 (100.0%) 0.237 Moderate 32 (78.0%) 1 (3.1%) 31 (96.9%) Poor 5 (12.2%) 1 (20.0%) 4 (80.0%) T-stage T-1 1 (3.2%) 0 (0.0%) 1 (100.0%) 0.446 T-2 5 (12.2%) 1 (20.0%) 4 (80.0%) T-3 24 (58.5%) 1 (4.2%) 23 (95.8%) T-4 11 (26.8%) 0 (0.0%) 11 (28.2%) Lymph node metastasis Absent 23 (56.1%) 2 (8.7%) 21 (91.3%) 0.309 Present 18 (43.9%) 0 (0.0%) 18 (100.0%) Metastasis Absent 29 (70.7%) 2 (6.9%) 27 (93.1%) 0.495 Present 12 (29.3%) 0 (0.0%) 12 (29.3%) Stage I 6 (14.6%) 1 (16.7%) 5 (83.3%) 0.377 II 16 (39.0%) 1 (6.3%) 15 (93.8%) III 7 (17.1%) 0 (0.0%) 7 (100.0%) IV 12 (29.3%) 0 (0.0%) 12 (100.0%) Lymphovascular Invasion 30 (73.2%) 2 (6.7%) 28 (93.3%) 0.530 Absent 11 (26.8%) 0 (0.0%) 11 (100.0%) Present MSI Absent 40 (97.6%) 1 (2.5%) 39 (97.5%) 0.049 Present 1 (2.4%) 1 (100.0%) 0 (0.0%) MSI = microsatellite instability

217

7.2.7 Discussion

Knockout of FAM134B significantly increases the proliferation migration and clonogenic potential of colon cancer cells [7, 9, 10]. FAM134B knockdown was noted to have regulatory effects on β-catenin, EB1 and APC [17]. In concurrence with previous results, upregulation of EB1 expression occurred in cells with reduced

FAM134B overexpression. In addition, both exogenous expression (Figure 7.2) and suppression [10] of FAM134B reduced the rate of apoptosis. Perhaps, both overexpression and suppression of FAM134B are pathogenic towards colon cancer progression and a balance in the gene expression is required for biological functioning of the gene [24, 25]. This interpretation is similar to that of salicylate-induced protein kinase (SIPK) and wound induced protein kinase (WIPK) ability to respond to reactive oxygen species (ROS) in tobacco plants [24]. Furthermore, overexpression of

FAM134B restored the expression of p53 (Figure 7.5D) a key regulator protein that is vital for the suppression of cancer. This suggests that FAM134B deletion plays a more pathogenic role in colon carcinogenesis and restoration of FAM134B could alter the progression of cancer through modulating other genes such as p53.

Conversely, overexpression of FAM134B resulted in an accumulation of cells at different phases of the cell cycle depending on cell type (Figure 7.3B). For example, overexpression of FAM134B resulted in a significant reduction (8.9%) of

SW480+FAM134B cells in the S phase. The reduction is coupled with a 4.3% increase in the cell population at the G2 phase by a 2.4%. This finding agreed with a previous study found where suppression of FAM134B resulted in the accumulation of SW480 cells at S phase [10]. Similarly, the cell population of SW48+FAM134B at the S phase reduced slightly while the cell population at both the G1 and G2 phase increased.

Interestingly, in FHC+FAM134B, this resulted in an opposite effect whereby a slight increase in cell population occurred in both the S phase and G2 phase, and a decrease

218 in the cell population at G1 phase. These results confirm the postulation that

FAM134B is involved in the regulation of DNA synthesis and cell cycle progression

[10]. FAM134B is likely to do so by inducing G1 cell cycle arrest in colon cancer but allows non-neoplastic colon epithelial cells to pass through the G1/S phase checkpoint.

Exogenous suppression of FAM134B resulted in overexpression of EB1 [17].

This study also found overexpression of EB1 in SW480+FAM134B (Figure 7.5BEB1 affects the length of Calmodulin Regulated Spectrin Associated Protein Family

Member 2 (CAMSAP2)-decorated stretches at non-centrosomal microtubules minus ends [26]. When EB1 is silenced, the length of CAMSAP2- decorated stretches at non- centrosomal microtubules minus ends in these cells are reduced. This causes the microtubules to detach from the Golgi membranes, making the Golgi complex more compact. Subsequently, the area occupied by Golgi stacks reduces [26]. An over compacted Golgi can lead to defect in vesicular trafficking. Given FAM134B’s cis-

Golgi localization, we postulate that FAM134B could have done so by affecting the expression of expression binding 1 (EB1) protein which is a known regulator of microtubule formation [26, 27]. Any changes to the physiological expression of

FAM134B will, therefore, disrupt the homeostasis and upset the cell cycle. Indeed,

Bouguenina (2017) reported that a defect in the interaction between EB1 and SMYLE

(myomegalin-like EB1 binding protein) results in an abnormal orientation of the mitotic spindle, thus, triggering G1 cell cycle arrest [27].

Cancer cells often elevate their mitochondrial biogenesis or modify their metabolic activities to enable tumor growth [28, 29]. Understanding the changes in metabolic phenotype can therefore, act as a useful guide for more effective management of patients with cancer [28]. This study showed no significant changes in mitochondrial respiration in non-neoplastic colonic epithelial cells (FHC+FAM134B).

Whereas, both colon cancer cells (SW48+FAM134B and SW480+FAM134B) showed a

219 reduction in non-mitochondrial oxygen consumption rate, basal respiration rate and

ATP production. Colon cancer utilizes mitochondrial oxidative phosphorylation

(OXPHOS) to produce ATP and drive cancer progression [30]. The reductions observed in SW48+FAM134B and SW480+FAM134B are indications that the overall mitochondrial respiration capacity has dropped. These preliminary results indicate that

FAM134B overexpression negatively affects the metabolic function of colon cancer cells but not in non-neoplastic cells. Perhaps, FAM134B exercises its tumor suppressive effects by way of controlling the bioenergetics of colon cancer. This is not surprising as FAM134B affects the endoplasmic reticulum (ER) stress [3, 31, 32] and one of the ER stress response includes changes in mitochondrial function [33].

Due to the similarity between macrophage migration inhibitory factor (MIF) and FAM134B on their effect towards the human health such as with immunology [34], inflammation [35], cancer [36-39] and allergy rhinitis [35, 40], further immunological analysis is done to validate the association between FAM134B and MIF expressions in patients with colorectal adenocarcinoma. This study found that overexpression of

FAM134B mRNA drove the expression of MIF (Figure 7.5A) in vitro. Similarly, while approximately 70% of patients with colorectal cancer showed positive for MIF expression showed have low expression of FAM134B mRNA, the remaining 30% of patients with high MIF expression also have higher expression of FAM134B. The relative higher proportion of patients having low expression of FAM134B with MIF expression could be due to the limitation of in vitro models in terms of physiological relevance and the tumor microenvironment changes [41]. Previous studies indicated that lower expression of FAM134B [10] and higher expression of MIF [42] was significantly associated with advanced pathological stages. In addition, low expression of FAM134B coupled with increased expression of MIF was significantly associated with carcinomas that were microsatellite instability (MSI) stable. This shows that these

220 set of patients (low FAM134B and high MIF expression) are less predisposed to hypermutable phenotype due to loss of DNA mismatched repair capability. However, the number of MSI high carcinomas in this study are small. More investigations may further associate the connection between the clinical and pathological features with increased expression of MIF and low expression of FAM134B.

7.2.8 Conclusion

In conclusion, this study showed that FAM134B could exercise its tumor suppressive function through controlling apoptosis, mitochondrial function.

Furthermore, a potential interaction with MIF, p53 and EB1 was noted. Both in vitro and clinical validation of MIF protein and FAM134B further suggest their downstream interaction in colorectal cancer. Nonetheless, further studies are required to understand the mechanism between the FAM134B, EB1, p53 and MIF axis to identify how these genes interact in colon cancer and inflammation.

7.2.9 Acknowledgement

The authors would like to thank Melissa Leung for managing the tissue biomarker database and microscopic pictures of the carcinoma, as well as staff in NH Diagnostics for the preparation of formalin-fixed paraffin-embedded tissue specimens.

221

7.2.10 References

[1] I. Kurth, T. Pamminger, J.C. Hennings, D. Soehendra, A.K. Huebner, A. Rotthier, J. Baets, J. Senderek, H. Topaloglu, S.A. Farrell, G. Nurnberg, P. Nurnberg, P. De Jonghe, A. Gal, C. Kaether, V. Timmerman, C.A. Hubner, Mutations in FAM134B, encoding a newly identified Golgi protein, cause severe sensory and autonomic neuropathy, Nat Genet, 41 (2009) 1179-1181. [2] F. Islam, V. Gopalan, A.K. Lam, RETREG1 (FAM134B): A new player in human diseases: 15 years after the discovery in cancer, Journal of cellular physiology, 233 (2018) 4479-4489. [3] A. Khaminets, T. Heinrich, M. Mari, P. Grumati, A.K. Huebner, M. Akutsu, L. Liebmann, A. Stolz, S. Nietzsche, N. Koch, M. Mauthe, I. Katona, B. Qualmann, J. Weis, F. Reggiori, I. Kurth, C.A. Hubner, I. Dikic, Regulation of endoplasmic reticulum turnover by selective autophagy, Nature, 522 (2015) 354-358. [4] D.C. Rubinsztein, Cell biology: Receptors for selective recycling, Nature, 522 (2015) 291-292. [5] Y. Peng, S.L. Shapiro, V.C. Banduseela, I.A. Dieterich, K.J. Hewitt, E.H. Bresnick, G. Kong, J. Zhang, K.L. Schueler, M.P. Keller, A.D. Attie, T.A. Hacker, R. Sullivan, E. Kielar-Grevstad, S.I. Arriola Apelo, D.W. Lamming, R.M. Anderson, L. Puglielli, Increased transport of acetyl-CoA into the endoplasmic reticulum causes a progeria-like phenotype, Aging cell, 17 (2018) e12820. [6] R. Melchiotti, K.J. Puan, A.K. Andiappan, T.Y. Poh, M. Starke, L. Zhuang, K. Petsch, T.S. Lai, F.T. Chew, A. Larbi, Y. Wang de, M. Poidinger, O. Rotzschke, Genetic analysis of an allergic rhinitis cohort reveals an intercellular epistasis between FAM134B and CD39, BMC medical genetics, 15 (2014) 73. [7] K. Kasem, E. Sullivan, V. Gopalan, A. Salajegheh, R.A. Smith, A.K. Lam, JK1 (FAM134B) represses cell migration in colon cancer: a functional study of a novel gene, Experimental and molecular pathology, 97 (2014) 99-104. [8] K. Kasem, V. Gopalan, A. Salajegheh, C.T. Lu, R.A. Smith, A.K. Lam, JK1 (FAM134B) gene and colorectal cancer: a pilot study on the gene copy number alterations and correlations with clinicopathological parameters, Experimental and molecular pathology, 97 (2014) 31-36. [9] K. Kasem, V. Gopalan, A. Salajegheh, C.T. Lu, R.A. Smith, A.K. Lam, The roles of JK-1 (FAM134B) expressions in colorectal cancer, Experimental cell research, 326 (2014) 166-173. [10] F. Islam, V. Gopalan, R. Wahab, R.A. Smith, B. Qiao, A.K. Lam, Stage dependent expression and tumor suppressive function of FAM134B (JK1) in colon cancer, Molecular carcinogenesis, 56 (2017) 238-249. [11] R. Wahab, V. Gopalan, F. Islam, R.A. Smith, B. Qiao, A.K. Lam, Cellular expression, in-vitro and in-vivo confirmation of GAEC1 oncogenic properties in colon cancer, European journal of cell biology, 96 (2017) 487-495. [12] F. Islam, V. Gopalan, J. Vider, R. Wahab, F. Ebrahimi, C.T. Lu, K. Kasem, A.K.Y. Lam, MicroRNA-186-5p overexpression modulates colon cancer growth by repressing the expression of the FAM134B tumour inhibitor, Experimental cell research, 357 (2017) 260-270. [13] X. Dai, T. Hua, T. Hong, Integrated diagnostic network construction reveals a 4- gene panel and 5 cancer hallmarks driving breast cancer heterogeneity, Scientific reports, 7 (2017) 6827. [14] F.B. Law, Y.W. Chen, K.Y. Wong, J. Ying, Q. Tao, C. Langford, P.Y. Lee, S. Law, R.W. Cheung, C.H. Chui, S.W. Tsao, K.Y. Lam, J. Wong, G. Srivastava, J.C. Tang, Identification of a novel tumor transforming gene GAEC1 at 7q22

222

which encodes a nuclear protein and is frequently amplified and overexpressed in esophageal squamous cell carcinoma, Oncogene, 26 (2007) 5877-5888. [15] M.H. Haque, V. Gopalan, K.W. Chan, M.J. Shiddiky, R.A. Smith, A.K. Lam, Identification of Novel FAM134B (JK1) Mutations in Oesophageal Squamous Cell Carcinoma, Scientific reports, 6 (2016) 29173. [16] F. Islam, V. Gopalan, S. Pillai, C.T. Lu, K. Kasem, A.K. Lam, Promoter hypermethylation inactivate tumor suppressor FAM134B and is associated with poor prognosis in colorectal cancer, Genes, chromosomes & cancer, 57 (2018) 240-251. [17] F. Islam, S. Chaousis, R. Wahab, V. Gopalan, A.K. Lam, Protein interactions of FAM134B with EB1 and APC/beta-catenin in vitro in colon carcinoma, Molecular carcinogenesis, 57 (2018) 1480-1491. [18] S.R. Hamilton, F.T. Bosman, P. Boffetta, M. Ilas, H. Morreau, S.I. Nakamura, P. Quirke, E. Riboli, L.H. Sobin, Carcinoma of the colon and rectum. WHO Classification of Tumours of the Digestive System. In: Bosman FT, Carneiro F, Hruban RH, Theise ND (eds) WHO classification of tumors of the digestive system, IARC Press, Lyon, 2010, 2010, pp. 134-146. [19] J.J. Milburn, R.M. Goldberg, E.A. Asare, A.B. Benson III, Colon and rectum. In AJCC Cancer staging Manual, Amin MB, Edge S, Greene F, Byrd DR, Brookland RK, Washington MK, Gershenwald JE, Compton CC, Hess KR, Sullivan DC, Jessup JM, Brierley JD, Gaspar LE, Schilsky RL, Balch CM, Winchester DP, Asare EA, Madera M, Gress DM, Meyer LR. (Eds.) Eight edition, Springer, 2017; Chapter 20, 2017, pp. 251-274. [20] K.T. Lee, V. Gopalan, F. Islam, R. Wahab, A. Mamoori, C.T. Lu, R.A. Smith, A.K. Lam, GAEC1 mutations and copy number aberration is associated with biological aggressiveness of colorectal cancer, European journal of cell biology, 97 (2018) 230-241. [21] A. Mamoori, V. Gopalan, C.T. Lu, T.C. Chua, D.L. Morris, R.A. Smith, A.K. Lam, Expression pattern of miR-451 and its target MIF (macrophage migration inhibitory factor) in colorectal cancer, J Clin Pathol, 70 (2017) 308-312. [22] N.A. Franken, H.M. Rodermond, J. Stap, J. Haveman, C. van Bree, Clonogenic assay of cells in vitro, Nature protocols, 1 (2006) 2315-2319. [23] X. Yang, Clonogenic Assay, Bio-protocol, 2 (2012) e187. [24] M.A. Samuel, B.E. Ellis, Double jeopardy: both overexpression and suppression of a redox-activated plant mitogen-activated protein kinase render tobacco plants ozone sensitive, The Plant cell, 14 (2002) 2059-2069. [25] D.R. Kelley, D.J. Skinner, C.S. Gasser, Roles of polarity determinants in ovule development, The Plant journal : for cell and molecular biology, 57 (2009) 1054- 1064. [26] C. Yang, J. Wu, C. de Heus, I. Grigoriev, N. Liv, Y. Yao, I. Smal, E. Meijering, J. Klumperman, R.Z. Qi, A. Akhmanova, EB1 and EB3 regulate microtubule minus end organization and Golgi morphology, The Journal of cell biology, 216 (2017) 3179-3198. [27] H. Bouguenina, D. Salaun, A. Mangon, L. Muller, E. Baudelet, L. Camoin, T. Tachibana, S. Cianferani, S. Audebert, P. Verdier-Pinard, A. Badache, EB1- binding-myomegalin protein complex promotes centrosomal microtubules functions, Proceedings of the National Academy of Sciences of the United States of America, 114 (2017) E10687-e10696. [28] M.G. Vander Heiden, R.J. DeBerardinis, Understanding the Intersections between Metabolism and Cancer Biology, Cell, 168 (2017) 657-669. [29] W.-X. Zong, J.D. Rabinowitz, E. White, Mitochondria and Cancer, Molecular cell, 61 (2016) 667-676. 223

[30] I.S. Song, Y.J. Jeong, J. Han, Mitochondrial metabolism in cancer stem cells: a therapeutic target for colon cancer, BMB reports, 48 (2015) 539-540. [31] I. Fregno, M. Molinari, Endoplasmic reticulum turnover: ER-phagy and other flavors in selective and non-selective ER clearance, F1000Research, 7 (2018) 454. [32] S. Song, J. Tan, Y. Miao, Q. Zhang, Crosstalk of ER stress-mediated autophagy and ER-phagy: Involvement of UPR and the core autophagy machinery, Journal of cellular physiology, 233 (2018) 3867-3874. [33] D. Senft, Z.A. Ronai, UPR, autophagy, and mitochondria crosstalk underlies the ER stress response, Trends in biochemical sciences, 40 (2015) 141-148. [34] M. Mittelbronn, M. Platten, P. Zeiner, Y. Dombrowski, B. Frank, C. Zachskorn, P.N. Harter, M. Weller, J. Wischhusen, Macrophage migration inhibitory factor (MIF) expression in human malignant gliomas contributes to immune escape and tumour progression, Acta neuropathologica, 122 (2011) 353-365. [35] T. Kok, A.A. Wasiel, R.H. Cool, B.N. Melgert, G.J. Poelarends, F.J. Dekker, Small-molecule inhibitors of macrophage migration inhibitory factor (MIF) as an emerging class of therapeutics for immune disorders, Drug discovery today, 23 (2018) 1910-1918. [36] R. De, S. Sarkar, S. Mazumder, S. Debsharma, A.A. Siddiqui, S.J. Saha, C. Banerjee, S. Nag, D. Saha, S. Pramanik, U. Bandyopadhyay, Macrophage migration inhibitory factor regulates mitochondrial dynamics and cell growth of human cancer cell lines through CD74-NFkappaB signaling, The Journal of biological chemistry, (2018). [37] F. Kong, X. Deng, X. Kong, Y. Du, L. Li, H. Zhu, Y. Wang, D. Xie, S. Guha, Z. Li, M. Guan, K. Xie, ZFPM2-AS1, a novel lncRNA, attenuates the p53 pathway and promotes gastric carcinogenesis by stabilizing MIF, Oncogene, 37 (2018) 5982-5996. [38] M. Dave, A. Islam, R.V. Jensen, A. Rostagno, J. Ghiso, A.R. Amin, Proteomic Analysis Shows Constitutive Secretion of MIF and p53-associated Activity of COX-2(-/-) Lung Fibroblasts, Genomics, proteomics & bioinformatics, 15 (2017) 339-351. [39] E. Verjans, E. Noetzel, N. Bektas, A.K. Schütz, H. Lue, B. Lennartz, A. Hartmann, E. Dahl, J. Bernhagen, Dual role of macrophage migration inhibitory factor (MIF) in human breast cancer, BMC cancer, 9 (2009) 230-230. [40] A. Iorio, R. Polimanti, S. Piacentini, G.M. Liumbruno, D. Manfellotto, M. Fuciarelli, Deletion polymorphism of GSTT1 gene as protective marker for allergic rhinitis, The clinical respiratory journal, 9 (2015) 481-486. [41] M.E. Katt, A.L. Placone, A.D. Wong, Z.S. Xu, P.C. Searson, In Vitro Tumor Models: Advantages, Disadvantages, Variables, and Selecting the Right Platform, Frontiers in bioengineering and biotechnology, 4 (2016) 12-12. [42] X.X. He, K. Chen, J. Yang, X.Y. Li, H.Y. Gan, C.Y. Liu, T.R. Coleman, Y. Al- Abed, Macrophage migration inhibitory factor promotes colorectal cancer, Molecular medicine (Cambridge, Mass.), 15 (2009) 1-10.

7.2.11 Funding

The project was supported by the HDR scholarship from Griffith University and funding from Menzies Health Institute Queensland. We also would like to

224 acknowledge the paper completion assistance from School of Medical Science,

Griffith University.

7.2.12 Conflict of Interest

The authors have no conflict of interest in publishing this manuscript.

225

Chapter 8: Summaries and conclusions

Early detection and treatment of colorectal cancer has proven vital as once metastasis occurs, the five-year survival rate becomes unfavourable. Therefore, a deeper understanding of the fundamental molecular mechanism of this disease would provide better treatment and management options for patients with colorectal cancer.

The current study unveils the roles of oncogene GAEC1 and tumour suppressor

FAM134B in the pathogenesis of colorectal adenocarcinoma. The molecular role, functional characteristics and clinical significance is further explored. Tissue samples were obtained from patients with colorectal adenocarcinoma and their clinicopathological significance were identified. A new method which provides a more accurate and repaid detection for the copy number aberration and ploidy determination for GAEC1 was developed. Novel mutations of GAEC1 within these patients were identified and any clinicopathological correlations were noted. In addition, both

GAEC1 and FAM134B expression-manipulation experiments were also conducted in vitro models to unveil the role they have in colon adenocarcinoma cells. These findings have provided a wealth of knowledge to the gap of information regarding GAEC1 and

FAM134B molecular and functional roles in the carcinogenesis of the colon and rectum.

In addition, there is a deeper understanding on the intrinsic molecular network underlining colorectal carcinogenesis. This could help open up new horizon to future studies and molecular therapy targeting in colorectal adenocarcinoma research.

This study developed a multiplexing DNA binding dye chemistry based assay

(e.g.: Sybr green) on the ddPCR platform which allows for a duplex detection of copy number and ploidy determination of GAEC1 from clinical samples. This enables the detection of two targets (reference and gene of interest) within the same well without the need of probe based assays that are costlier. Running both reference and target gene within the same reaction tube reduces variations that may arise from sample

226 preparation. Furthermore, duplex assays allow for better sample and assay to test ratio, reducing overall cost and runtime. Inherent variance reduces from sample to sample with ddPCR assay having a smaller coefficient variance (0.84% to 4.17%) than qPCR assay (0.06% to 53.04%) resulting in better reproducibility in ddPCR compared to qPCR. The method also requires less sample, lower consumables and run-time thereby improving overall patient care and reducing cost per sample.

Using this technology, the current study found that 24% of patients showed

GAEC1 amplification, 1.3% of patients showed deletion while 74.7% of patients showed not changes between adenocarcinoma tissues and their matched non- neoplastic mucosal tissue. The genomic DNA from these tissues were subsequently run through Sanger sequencing to identify for GAEC1 mutations. From this, approximately 10% of the adenocarcinoma tissues showed positive for GAEC1 mutation including one missense mutation. Four loss of heterozygosity (LOH) and two substitution. Interestingly, GAEC1 mutation were frequently associated with increased

GAEC1 protein expression. Increased GAEC1 copies and GAEC1 mutations were significantly associated with the biological aggressiveness of colon adenocarcinoma in terms of perforation and survival rate of these patients, suggesting the oncogenic potential of GAEC1.

Moving forth, the research focused on overexpressing GAEC1 in colon adenocarcinoma in vitro to confirm GAEC1 oncogenic effects and identify potential transcriptional factors that GAEC1 interacts with in colorectal adenocarcinoma to gain better understanding of the mechanism of action of this gene. It was found that overexpression of GAEC1 successfully induced cancer like properties in non- neoplastic epithelial cells (FHC). Furthermore, overexpression of GAEC1 significantly increased the cell proliferation, migration and clonogenic potential in colon adenocarcinomas (SW48 and SW480) confirming its oncogenic potential. This chapter

227 also found that GAEC1 was able to influence mitochondrial respiration changes by reducing coupling efficiency in advanced stage colon adenocarcinoma. In addition, overexpression of GAEC1 was correlated with an increase in pAkt, FOXO3 and

MMP9 in vitro. These findings open up a new horizon to the studies of the molecular mechanisms of GAEC1 in colon adenocarcinoma.

A similar study was then conducted on FAM134B to confirm the tumour suppressive effects of FAM134B on colon adenocarcinoma. Likewise, previous studies of FAM134B in colon adenocarcinoma have focused mainly on knock down studies.

In this study, the expression of FAM134B was enforced instead. Results showed

FAM134B is likely to exercise its tumour suppressive function through controlling apoptosis, mitochondrial function. Overexpression of FAM134B was coupled with higher expression of EB1, p53 and MIF in colon cancer and non-neoplastic colonic epithelial cells. Approximately 70% of patients expressing low levels of FAM134B expression also displayed high MIF expression, implying that FAM134B interacts with

MIF to affect colon cancer progression. Interestingly, these population of patients were significantly associated with microsatellite instability. Nonetheless, further studies are required to understand the mechanism between the FAM134B, EB1, p53 and MIF axis to identify how these genes interact in colon adenocarcinoma and/or inflammation. In addition, it is well known that SW480 harbours mutated p53 however, studies done by

Konduru and colleagues has proven that p53 mutations may or may not have a direct impact on its expression changes [113]. Nonetheless, further analysis of p53 mutation status upon overexpression of FAM134B using immunoprecipitation or mass spectrometry needs to be done to further understand the effects of enforced p53 expression by FAM134B overexpression.

In conclusion, the present study confirms that GAEC1 can acts as an oncogene and FAM134B can act as a tumour suppressor in colorectal adenocarcinoma by

228 affecting cell growth and progression. This study is the first to ectopically overexpress

GAEC1 and FAM134B in colon adenocarcinoma in vitro. Furthermore, this is the first study ever to identify mutation in the GAEC1 sequence and have successfully annotated the gene sequence and protein sequence onto GenBank. Proper annotation of GAEC1 allows for easier compilation of information regarding the gene and opens up exciting possibilities for in silico analysis using bioinformatics tools. The increased in GAEC1 expression and presence of GAEC1 mutations were significantly associated to the biological aggressiveness of colorectal adenocarcinomas. In addition, this study noted the relevance of both GAEC1 and FAM134B towards the mitochondrial activity in colon adenocarcinoma. Moreover, the study identified potential interacting partner of GAEC1 and FAM134B in colon adenocarcinoma. Therefore, the findings presented in this study has enriched the current understanding of both GAEC1 and FAM134B in the pathogenesis of colorectal adenocarcinoma. With the development of droplet digital PCR for absolute quantification of GAEC1 copy number aberration and ploidy determination, GenBank annotation, mutation discovery, functional findings such as cell transformation properties and oncogenic/tumour suppressive properties, and novel gene interactions discovered, this study would help give a better understanding of the genetic aetiology of colorectal adenocarcinoma. This will help provide new insights and opportunities for the development of novel treatment and therapeutic options to enhance patient management.

229

Chapter 9: References

1. AIHW, AACR. Cancer in Australia 2017. Canberra: AIHW; 2017. p. 204.

2. Fearon ER, Vogelstein B. A genetic model for colorectal tumorigenesis. Cell. 1990;61(5):759-67.

3. Peate I. Nursing & Health Survival Guide: Cancer Care: Taylor & Francis; 2014.

4. Vogelstein B, Fearon ER, Hamilton SR, Kern SE, Preisinger AC, Leppert M, et al. Genetic alterations during colorectal-tumor development. N Engl J Med. 1988;319(9):525-32.

5. Fearon ER. Molecular genetics of colorectal cancer. Annu Rev Pathol. 2011;6:479-507.

6. Vogelstein B, Papadopoulos N, Velculescu VE, Zhou S, Diaz LA, Jr., Kinzler KW. Cancer genome landscapes. Science. 2013;339(6127):1546-58.

7. Lupulescu AP. Control of precancer cell transformation into cancer cells: its relevance to cancer prevention. Cancer Detect Prev. 1996;20(6):634-7.

8. Trosko JE. Human adult stem cells as the target cells for the initiation of carcinogenesis and for the generation of "cancer stem cells". Int J Stem Cells. 2008;1(1):8-26.

9. Irigaray P, Belpomme D. Basic properties and molecular mechanisms of exogenous chemical carcinogens. Carcinogenesis. 2010;31(2):135-48.

10. Jones S, Chen WD, Parmigiani G, Diehl F, Beerenwinkel N, Antal T, et al. Comparative lesion sequencing provides insights into tumor evolution. Proceedings of the National Academy of Sciences of the United States of America. 2008;105(11):4283-8.

11. Tang JC, Lam KY, Law S, Wong J, Srivastava G. Detection of genetic alterations in esophageal squamous cell carcinomas and adjacent normal epithelia by comparative DNA fingerprinting using inter-simple sequence repeat PCR. Clinical cancer research : an official journal of the American Association for Cancer Research. 2001;7(6):1539-45.

12. Kwong D, Lam A, Guan X, Law S, Tai A, Wong J, et al. Chromosomal aberrations in esophageal squamous cell carcinoma among Chinese: gain of 12p predicts poor prognosis after surgery. Human pathology. 2004;35(3):309-16.

13. Law FB, Chen YW, Wong KY, Ying J, Tao Q, Langford C, et al. Identification of a novel tumor transforming gene GAEC1 at 7q22 which encodes a nuclear protein and is frequently amplified and overexpressed in esophageal squamous cell carcinoma. Oncogene. 2007;26(40):5877-88.

14. Tang WK, Chui CH, Fatima S, Kok SH, Pak KC, Ou TM, et al. Oncogenic properties of a novel gene JK-1 located in chromosome 5p and its overexpression in

230 human esophageal squamous cell carcinoma. International journal of molecular medicine. 2007;19(6):915-23.

15. Lee KT, Gopalan V, Islam F, Wahab R, Mamoori A, Lu CT, et al. GAEC1 mutations and copy number aberration is associated with biological aggressiveness of colorectal cancer. European journal of cell biology. 2018;97(3):230-41.

16. Islam F, Gopalan V, Lam AK. RETREG1 (FAM134B): A new player in human diseases: 15 years after the discovery in cancer. Journal of cellular physiology. 2018;233(6):4479-89.

17. Anand P, Kunnumakkara AB, Sundaram C, Harikumar KB, Tharakan ST, Lai OS, et al. Cancer is a preventable disease that requires major lifestyle changes. Pharm Res. 2008;25(9):2097-116.

18. Ferlay J, Soerjomataram I, Dikshit R, Eser S, Mathers C, Rebelo M, et al. Cancer incidence and mortality worldwide: sources, methods and major patterns in GLOBOCAN 2012. Int J Cancer. 2015;136(5):E359-86.

19. AIHW. Health system expenditure on disease and injury in Australia, 2004-05. Canberra: AIHW; 2010.

20. Institute NC. Colon Cancer Treatment (PDQ®) 2015 [updated 22/7/2015. Available from: http://www.cancer.gov/types/colorectal/patient/colon-treatment- pdq#section/all.

21. Boyle P, Leon ME. Epidemiology of colorectal cancer. Br Med Bull. 2002;64:1-25.

22. Newton RU, Galvao DA. Exercise in prevention and management of cancer. Curr Treat Options Oncol. 2008;9(2-3):135-46.

23. Huxley RR, Ansary-Moghaddam A, Clifton P, Czernichow S, Parr CL, Woodward M. The impact of dietary and lifestyle risk factors on risk of colorectal cancer: a quantitative overview of the epidemiological evidence. Int J Cancer. 2009;125(1):171-80.

24. Taylor DP, Burt RW, Williams MS, Haug PJ, Cannon-Albright LA. Population-based family history-specific risks for colorectal cancer: a constellation approach. Gastroenterology. 2010;138(3):877-85.

25. Hisamuddin IM, Yang VW. Genetics of colorectal cancer. MedGenMed. 2004;6(3):13.

26. Berenblum I, Shubik P. A new, quantitative, approach to the study of the stages of chemical cartinogenesis in the mouse's skin. Br J Cancer. 1947;1(4):383-91.

27. Lichtenstein AV. On evolutionary origin of cancer. Cancer Cell Int. 2005;5(1):5.

28. Poirier LA. Stages in carcinogenesis: alteration by diet. Am J Clin Nutr. 1987;45(1 Suppl):185-91.

231

29. Lau T, Chan E, Callow M, Waaler J, Boggs J, Blake RA, et al. A novel tankyrase small-molecule inhibitor suppresses APC mutation-driven colorectal tumor growth. Cancer Res. 2013;73(10):3132-44.

30. Brannon AR, Vakiani E, Sylvester BE, Scott SN, McDermott G, Shah RH, et al. Comparative sequencing analysis reveals high genomic concordance between matched primary and metastatic colorectal cancer lesions. Genome Biol. 2014;15(8):454.

31. Dow LE, O'Rourke KP, Simon J, Tschaharganeh DF, van Es JH, Clevers H, et al. Apc Restoration Promotes Cellular Differentiation and Reestablishes Crypt Homeostasis in Colorectal Cancer. Cell. 2015;161(7):1539-52.

32. Cheung AF, Carter AM, Kostova KK, Woodruff JF, Crowley D, Bronson RT, et al. Complete deletion of Apc results in severe polyposis in mice. Oncogene. 2010;29(12):1857-64.

33. Peters U, Bien S, Zubair N. Genetic architecture of colorectal cancer. Gut. 2015.

34. Lichtenstein P, Holm NV, Verkasalo PK, Iliadou A, Kaprio J, Koskenvuo M, et al. Environmental and heritable factors in the causation of cancer--analyses of cohorts of twins from Sweden, Denmark, and Finland. N Engl J Med. 2000;343(2):78- 85.

35. Czene K, Lichtenstein P, Hemminki K. Environmental and heritable causes of cancer among 9.6 million individuals in the Swedish Family-Cancer Database. Int J Cancer. 2002;99(2):260-6.

36. Kerber RA, Neklason DW, Samowitz WS, Burt RW. Frequency of familial colon cancer and hereditary nonpolyposis colorectal cancer (Lynch syndrome) in a large population database. Fam Cancer. 2005;4(3):239-44.

37. Jasperson KW, Tuohy TM, Neklason DW, Burt RW. Hereditary and familial colon cancer. Gastroenterology. 2010;138(6):2044-58.

38. Talseth-Palmer BA, Wijnen JT, Grice DM, Scott RJ. Genetic modifiers of cancer risk in Lynch syndrome: a review. Fam Cancer. 2013;12(2):207-16.

39. Forster M, Forster P, Elsharawy A, Hemmrich G, Kreck B, Wittig M, et al. From next-generation sequencing alignments to accurate comparison and validation of single-nucleotide variants: the pibase software. Nucleic Acids Res. 2013;41(1):e16.

40. Bottcher R, Amberg R, Ruzius FP, Guryev V, Verhaegh WF, Beyerlein P, et al. Using a priori knowledge to align sequencing reads to their exact genomic position. Nucleic Acids Res. 2012;40(16):e125.

41. Ding J, Bashashati A, Roth A, Oloumi A, Tse K, Zeng T, et al. Feature-based classifiers for somatic mutation detection in tumour-normal paired sequencing data. Bioinformatics. 2012;28(2):167-75.

42. Mertes F, Elsharawy A, Sauer S, van Helvoort JM, van der Zaag PJ, Franke A, et al. Targeted enrichment of genomic DNA regions for next-generation sequencing. Brief Funct Genomics. 2011;10(6):374-86.

232

43. Harismendy O, Ng PC, Strausberg RL, Wang X, Stockwell TB, Beeson KY, et al. Evaluation of next generation sequencing platforms for population targeted sequencing studies. Genome Biol. 2009;10(3):R32.

44. Gundry M, Vijg J. Direct mutation analysis by high-throughput sequencing: from germline to low-abundant, somatic variants. Mutat Res. 2012;729(1-2):1-15.

45. Orfi E, Szebeni J. The immune system of the gut and potential adverse effects of oral nanocarriers on its function. Advanced drug delivery reviews. 2016;106(Pt B):402-9.

46. Bronte V, Pittet MJ. The spleen in local and systemic regulation of immunity. Immunity. 2013;39(5):806-18.

47. Song F, Qian Y, Peng X, Li X, Xing P, Ye D, et al. The frontline of immune response in peripheral blood. PLoS ONE. 2017;12(8):e0182294.

48. Chan D, Tsoi MY, Liu CD, Chan SH, Law SY, Chan KW, et al. Oncogene GAEC1 regulates CAPN10 expression which predicts survival in esophageal squamous cell carcinoma. World journal of gastroenterology. 2013;19(18):2772-80.

49. Frances CP, Conde MC, Saez ME, Diez SF, Rey CM, Ramirez-Armengol JA, et al. Identification of a protective haplogenotype within CAPN10 gene influencing colorectal cancer susceptibility. Journal of gastroenterology and hepatology. 2007;22(12):2298-302.

50. Redon R, Ishikawa S, Fitch KR, Feuk L, Perry GH, Andrews TD, et al. Global variation in copy number in the human genome. Nature. 2006;444(7118):444-54.

51. Cook EH, Jr., Scherer SW. Copy-number variations associated with neuropsychiatric conditions. Nature. 2008;455(7215):919-23.

52. Park C, Kim JI, Hong SN, Jung HM, Kim TJ, Lee S, et al. A copy number variation in PKD1L2 is associated with colorectal cancer predisposition in korean population. Int J Cancer. 2017;140(1):86-94.

53. Lupski JR. Structural variation in the human genome. N Engl J Med. 2007;356(11):1169-71.

54. Coughlin CR, 2nd, Scharer GH, Shaikh TH. Clinical impact of copy number variation analysis using high-resolution microarray technologies: advantages, limitations and concerns. Genome Med. 2012;4(10):80.

55. Shlien A, Malkin D. Copy number variations and cancer. Genome Med. 2009;1(6):62.

56. Gopalan V, Smith RA, Nassiri MR, Yasuda K, Salajegheh A, Kim SY, et al. GAEC1 and colorectal cancer: a study of the relationships between a novel oncogene and clinicopathologic features. Human pathology. 2010;41(7):1009-15.

57. Gopalan V, Yasuda K, Pillai S, Tiang T, Leung M, Lu CT, et al. Gene amplified in oesophageal cancer 1 (GAEC1) amplification in colorectal cancers and its impact on patient's survival. J Clin Pathol. 2013;66(8):721-3.

233

58. Iafrate AJ, Feuk L, Rivera MN, Listewnik ML, Donahoe PK, Qi Y, et al. Detection of large-scale variation in the human genome. Nat Genet. 2004;36(9):949- 51.

59. Wahab R, Gopalan V, Islam F, Mamoori A, Lee KT, Lu CT, et al. Expression of GAEC1 mRNA and protein and its association with clinical and pathological parameters of patients with colorectal adenocarcinoma. Experimental and molecular pathology. 2018;104(1):71-5.

60. Wahab R, Gopalan V, Islam F, Smith RA, Qiao B, Lam AK. Cellular expression, in-vitro and in-vivo confirmation of GAEC1 oncogenic properties in colon cancer. European journal of cell biology. 2017;96(6):487-95.

61. Brown TA. Genome 3: Garland Science; 2007. 713 p.

62. Govindan R, Morgensztern D, Washington University (Saint Louis Mo.). School of Medicine. The Washington manual of oncology. Third edition. ed. Philadelphia: Wolters Kluwer Health; 2015. xiii, 495 pages p.

63. Spasic M, Friedel CC, Schott J, Kreth J, Leppek K, Hofmann S, et al. Genome- wide assessment of AU-rich elements by the AREScore algorithm. PLoS Genet. 2012;8(1):e1002433.

64. Chen CY, Shyu AB. AU-rich elements: characterization and importance in mRNA degradation. Trends Biochem Sci. 1995;20(11):465-70.

65. Degese MS, Tanos T, Naipauer J, Gingerich T, Chiappe D, Echeverria P, et al. An interplay between the p38 MAPK pathway and AUBPs regulates c-fos mRNA stability during mitogenic stimulation. Biochem J. 2015;467(1):77-90.

66. Kurth I, Pamminger T, Hennings JC, Soehendra D, Huebner AK, Rotthier A, et al. Mutations in FAM134B, encoding a newly identified Golgi protein, cause severe sensory and autonomic neuropathy. Nat Genet. 2009;41(11):1179-81.

67. Kurth I. Hereditary Sensory and Autonomic Neuropathy Type II. In: Adam MP, Ardinger HH, Pagon RA, Wallace SE, Bean LJH, Stephens K, et al., editors. GeneReviews((R)). Seattle (WA): University of Washington, Seattle University of Washington, Seattle. GeneReviews is a registered trademark of the University of Washington, Seattle. All rights reserved.; 1993.

68. Khaminets A, Heinrich T, Mari M, Grumati P, Huebner AK, Akutsu M, et al. Regulation of endoplasmic reticulum turnover by selective autophagy. Nature. 2015;522(7556):354-8.

69. Rubinsztein DC. Cell biology: Receptors for selective recycling. Nature. 2015;522(7556):291-2.

70. Scheubert L, Schmidt R, Repsilber D, Lustrek M, Fuellen G. Learning biomarkers of pluripotent stem cells in mouse. DNA research : an international journal for rapid publication of reports on genes and genomes. 2011;18(4):233-51.

71. Lennemann NJ, Coyne CB. Dengue and Zika viruses subvert reticulophagy by NS2B3-mediated cleavage of FAM134B. Autophagy. 2017;13(2):322-32.

234

72. Chiramel AI, Dougherty JD, Nair V, Robertson SJ, Best SM. FAM134B, the Selective Autophagy Receptor for Endoplasmic Reticulum Turnover, Inhibits Replication of Ebola Virus Strains Makona and Mayinga. The Journal of infectious diseases. 2016;214(suppl 3):S319-s25.

73. Ilgaz Aydinlar E, Rolfs A, Serteser M, Parman Y. Mutation in FAM134B causing hereditary sensory neuropathy with spasticity in a Turkish family. Muscle & nerve. 2014;49(5):774-5.

74. Murphy SM, Davidson GL, Brandner S, Houlden H, Reilly MM. Mutation in FAM134B causing severe hereditary sensory neuropathy. Journal of neurology, neurosurgery, and psychiatry. 2012;83(1):119-20.

75. Riviere JB, Ramalingam S, Lavastre V, Shekarabi M, Holbert S, Lafontaine J, et al. KIF1A, an axonal transporter of synaptic vesicles, is mutated in hereditary sensory and autonomic neuropathy type 2. American journal of human genetics. 2011;89(2):219-30.

76. Davidson G, Murphy S, Polke J, Laura M, Salih M, Muntoni F, et al. Frequency of mutations in the genes associated with hereditary sensory and autonomic neuropathy in a UK cohort. Journal of neurology. 2012;259(8):1673-85.

77. Kong M, Kim Y, Lee C. A strong synergistic epistasis between FAM134B and TNFRSF19 on the susceptibility to vascular dementia. Psychiatric genetics. 2011;21(1):37-41.

78. Melchiotti R, Puan KJ, Andiappan AK, Poh TY, Starke M, Zhuang L, et al. Genetic analysis of an allergic rhinitis cohort reveals an intercellular epistasis between FAM134B and CD39. BMC medical genetics. 2014;15:73.

79. Kasem K, Sullivan E, Gopalan V, Salajegheh A, Smith RA, Lam AK. JK1 (FAM134B) represses cell migration in colon cancer: a functional study of a novel gene. Experimental and molecular pathology. 2014;97(1):99-104.

80. Song S, Tan J, Miao Y, Zhang Q. Crosstalk of ER stress-mediated autophagy and ER-phagy: Involvement of UPR and the core autophagy machinery. Journal of cellular physiology. 2018;233(5):3867-74.

81. Rao RV, Niazi K, Mollahan P, Mao X, Crippen D, Poksay KS, et al. Coupling endoplasmic reticulum stress to the cell-death program: a novel HSP90-independent role for the small chaperone protein p23. Cell death and differentiation. 2006;13(3):415-25.

82. Yadav RK, Chae SW, Kim HR, Chae HJ. Endoplasmic reticulum stress and cancer. Journal of cancer prevention. 2014;19(2):75-88.

83. Islam F, Gopalan V, Wahab R, Smith RA, Qiao B, Lam AK. Stage dependent expression and tumor suppressive function of FAM134B (JK1) in colon cancer. Molecular carcinogenesis. 2017;56(1):238-49.

84. Kasem K, Gopalan V, Salajegheh A, Lu CT, Smith RA, Lam AK. JK1 (FAM134B) gene and colorectal cancer: a pilot study on the gene copy number

235 alterations and correlations with clinicopathological parameters. Experimental and molecular pathology. 2014;97(1):31-6.

85. Kasem K, Gopalan V, Salajegheh A, Lu CT, Smith RA, Lam AK. The roles of JK-1 (FAM134B) expressions in colorectal cancer. Experimental cell research. 2014;326(1):166-73.

86. Islam F, Gopalan V, Pillai S, Lu CT, Kasem K, Lam AK. Promoter hypermethylation inactivate tumor suppressor FAM134B and is associated with poor prognosis in colorectal cancer. Genes, chromosomes & cancer. 2018;57(5):240-51.

87. Islam F, Gopalan V, Vider J, Wahab R, Ebrahimi F, Lu CT, et al. MicroRNA- 186-5p overexpression modulates colon cancer growth by repressing the expression of the FAM134B tumour inhibitor. Experimental cell research. 2017;357(2):260-70.

88. Islam F, Chaousis S, Wahab R, Gopalan V, Lam AK. Protein interactions of FAM134B with EB1 and APC/beta-catenin in vitro in colon carcinoma. Molecular carcinogenesis. 2018;57(11):1480-91.

89. Hamilton SR, Bosman FT, Boffetta P, Ilas M, Morreau H, Nakamura SI, et al. Carcinoma of the colon and rectum. WHO Classification of Tumours of the Digestive System. In: Bosman FT, Carneiro F, Hruban RH, Theise ND (eds) WHO classification of tumors of the digestive system. 4 ed: IARC Press, Lyon, 2010; 2010. p. 134-46.

90. Milburn JJ, Goldberg RM, Asare EA, Benson III AB. Colon and rectum. In AJCC Cancer staging Manual, Amin MB, Edge S, Greene F, Byrd DR, Brookland RK, Washington MK, Gershenwald JE, Compton CC, Hess KR, Sullivan DC, Jessup JM, Brierley JD, Gaspar LE, Schilsky RL, Balch CM, Winchester DP, Asare EA, Madera M, Gress DM, Meyer LR. (Eds.) Eight edition. Springer, 2017; Chapter 20; 2017. p. 251-74.

91. Lam AK, Ong K, Giv MJ, Ho YH. p16 expression in colorectal adenocarcinoma: marker of aggressiveness and morphological types. Pathology. 2008;40(6):580-5.

92. Franken NA, Rodermond HM, Stap J, Haveman J, van Bree C. Clonogenic assay of cells in vitro. Nature protocols. 2006;1(5):2315-9.

93. Yang X. Clonogenic Assay. Bio-protocol. 2012;2(10):e187.

94. Schwarz JM, Cooper DN, Schuelke M, Seelow D. MutationTaster2: mutation prediction for the deep-sequencing age. Nature methods. 2014;11(4):361-2.

95. Rosenbloom KR, Armstrong J, Barber GP, Casper J, Clawson H, Diekhans M, et al. The UCSC Genome Browser database: 2015 update. Nucleic Acids Res. 2015;43(Database issue):D670-81.

96. Zhang J, Baran J, Cros A, Guberman JM, Haider S, Hsu J, et al. International Cancer Genome Consortium Data Portal--a one-stop shop for cancer genomics data. Database : the journal of biological databases and curation. 2011;2011:bar026.

97. Choi Y, Chan AP. PROVEAN web server: a tool to predict the functional effect of amino acid substitutions and indels. Bioinformatics. 2015;31(16):2745-7.

236

98. Kuiper RP, Ligtenberg MJ, Hoogerbrugge N, Geurts van Kessel A. Germline copy number variation and cancer risk. Current opinion in genetics & development. 2010;20(3):282-9.

99. Serrati S, De Summa S, Pilato B, Petriella D, Lacalamita R, Tommasi S, et al. Next-generation sequencing: advances and applications in cancer diagnosis. OncoTargets and therapy. 2016;9:7355-65.

100. Xu H, Zhu X, Xu Z, Hu Y, Bo S, Xing T, et al. Non-invasive Analysis of Genomic Copy Number Variation in Patients with Hepatocellular Carcinoma by Next Generation DNA Sequencing. Journal of Cancer. 2015;6(3):247-53.

101. Yoshihara K, Tajima A, Adachi S, Quan J, Sekine M, Kase H, et al. Germline copy number variations in BRCA1-associated ovarian cancer patients. Genes, chromosomes & cancer. 2011;50(3):167-77.

102. Heid CA, Stevens J, Livak KJ, Williams PM. Real time quantitative PCR. Genome research. 1996;6(10):986-94.

103. Kubista M, Andrade JM, Bengtsson M, Forootan A, Jonak J, Lind K, et al. The real-time polymerase chain reaction. Molecular aspects of medicine. 2006;27(2-3):95- 125.

104. Dobnik D, Stebih D, Blejec A, Morisset D, Zel J. Multiplex quantification of four DNA targets in one reaction with Bio-Rad droplet digital PCR system for GMO detection. Scientific reports. 2016;6:35451.

105. Karlen Y, McNair A, Perseguers S, Mazza C, Mermod N. Statistical significance of quantitative PCR. BMC Bioinformatics. 2007;8:131.

106. Weaver S, Dube S, Mir A, Qin J, Sun G, Ramakrishnan R, et al. Taking qPCR to a higher level: Analysis of CNV reveals the power of high throughput qPCR to enhance quantitative resolution. Methods. 2010;50(4):271-6.

107. Islam F, Gopalan V, Wahab R, Lee KT, Haque MH, Mamoori A, et al. Novel FAM134B mutations and their clinicopathological significance in colorectal cancer. Human genetics. 2017;136(3):321-37.

108. Fregno I, Molinari M. Endoplasmic reticulum turnover: ER-phagy and other flavors in selective and non-selective ER clearance. F1000Research. 2018;7:454.

109. Song S, Tan J, Miao Y, Zhang Q. Crosstalk of ER stress-mediated autophagy and ER-phagy: Involvement of UPR and the core autophagy machinery. Journal of cellular physiology. 2018;233(5):3867-74.

110. Senft D, Ronai ZA. UPR, autophagy, and mitochondria crosstalk underlies the ER stress response. Trends in biochemical sciences. 2015;40(3):141-8.

111. Kok T, Wasiel AA, Cool RH, Melgert BN, Poelarends GJ, Dekker FJ. Small- molecule inhibitors of macrophage migration inhibitory factor (MIF) as an emerging class of therapeutics for immune disorders. Drug discovery today. 2018;23(11):1910- 8.

237

112. Iorio A, Polimanti R, Piacentini S, Liumbruno GM, Manfellotto D, Fuciarelli M. Deletion polymorphism of GSTT1 gene as protective marker for allergic rhinitis. The clinical respiratory journal. 2015;9(4):481-6.

113. Koduru PRK, Raju K, Vadmal V, Menezes G, Shah S, Susin M, et al. Correlation Between Mutation in P53, p53 Expression, Cytogenetics, Histologic Type, and Survival in Patients With B-Cell Non-Hodgkin's Lymphoma. Blood. 1997;90(10):4078-91.

238

Appendices

Appendix A

239

240

241

Appendix B

242

243

244

Appendix C

245

Appendix D

246

247

248

249

Appendix E - Reagents and buffers preparation

PBST (1L) 500µL Tween 20 in 1L of PBS

Scott’s Bluing Solution (3L) Magnesium sulphate (MgSO4) 30.0g Sodium bicarbonate (NaCO3) 2.0g Distilled water 3.0L

Stripping buffer • Mild 100mL 10x TGS buffer 700mL water 10mL Tween20 Adjust pH to 2.2 Bring volume up to 1L with dH2O

• Harsh 20 mL SDS 10% 12.5 mL Tris HCl, pH 6.8, 0.5 M 67.5 mL distilled water 0.8 mL ß-mercaptoethanol

Appendix F – Media preparation

SOB (50mL) 2% Vegetable peptone (or Tryptone) = 1g 0.5% Yeast extract = 0.25g 10mM NaCl = 0.029g (mw = 58.44 g/mol) 2.5mM KCl = 0.008g (mw = 74.5513g/mol) 10mM MgCl2 = 0.048g (mw = 95.211g/mol) 10mM MgSO4 = 0.06g (mw = 120.366g/mol) Dissolved in 50mL of water (autoclaved before used)

SOC (50mL) SOB + 20mM Glucose 20mM Glucose = 0.18g (mw = 180.1559g/mol) (autoclaved before used)

LB Agar (100mL) LB broth ingredient = 2.5g 1.5% agar = 1.5g Dissolved in 100mL of water (autoclaved before used)

Add appropriate antibiotics for preferred selectivity • 100µg/mL Ampicillin (stock = 100mg/mL) = 100uL • 25µg/mL Kanamycin 250

Appendix G

251

252

253

254

255

256

257

258

259

260

261

262

263

264

265

266

267