1

MicroRNAs are novel biomarkers for the detection of colorectal

neoplasia and high risk Dukes’ B cancers

Thesis submitted for the degree of Doctorate of Philosophy at

the University of Leicester

by

Mr Muhammad Imran Aslam

Department of Cancer Studies and Molecular Medicine

University of Leicester, UK

May 2016

2

1 Abstract

MicroRNAs are novel biomarkers for the detection of colorectal neoplasia and high

risk Dukes’ B cancers

Aims: This study aimed to identify which circulating miRNAs can be used for the early

detection of colorectal cancers (CRCs) and to assess the utility of tissue miRNAs

combined with common gene mutations, to predict the development of metastasis in

patients with Dukes’ B CRCs.

Methods: microRNA (miRNA) expression profiling was performed for total RNA

extracted from plasma samples (colonoscopy negative controls=11, adenomas=9,

carcinomas=12) and formalin-fixed paraffin embedded (FFPE) matched paired

cancerous with adjacent normal tissue (n=20, 5 cases from each group of Dukes’ A,

Dukes’ B with metastasis during 5 year follow up, Dukes’ B without metastasis during 5

year follow up, and Dukes’ C) using Taqman® MicroRNA Array, Megaplex™ RT and pre-

amplification primers Human Pool A v.2.1 and Pool B v.2.0. Discriminatory miRNAs

identified from plasma and tissue expression profiles were validated further on cohorts

of plasma (n=190) and FFPE tissues (n=72). Three common gene mutations (KRAS, BRAF

and PIK3CA) were analysed in DNA extracted from FFPE cancer tissue. miRNA

expression analysis was applied to circulating exosomes to quantify CRC-related

exosomal miRNAs.

Results: Receiver operating characteristics analysis showed miR-135b was associated

with an area under the curve value of 0.82 (95% CI: 0.71-0.92), with 80% sensitivity and

84% specificity for the detection of adenomas and carcinomas. miR-135b was also

detectable in immunoaffinity-isolated plasma exosomes from patients with CRC. No

significant differences were noted for mutation status and the development of

metastasis. Expression levels of miR-135b and miR-15b were significantly associated

with Dukes’ B cancers tissue and the development of metastasis.

Conclusions: miR-135b is a novel diagnostic and prognostic marker. Its expression levels

in blood and tissue can be used for the early detection of CRCs and to predict the

development of metastasis in Dukes’ B cancers.

3

2 Acknowledgements

I would like to express my deepest gratitude to my supervisors, Dr Howard Pringle, Mr

John Jameson and Mr Baljit Singh, for their excellent guidance, caring and patience, and

for providing me with an excellent atmosphere for doing research. I would specially like

to thank Dr Howard Pringle who has shown the attitude and the substance of a genius:

he continually and persuasively conveyed a spirit of adventure in regard to research and

scholarship, and an excitement in regard to teaching. Without his supervision and

constant help this dissertation would not have been possible. He let me experience the

research of biomarkers in colorectal cancers and practical issues beyond the textbooks,

patiently corrected my writing and financially supported my research. I would never

have been able to finish my dissertation without the guidance of my supervisors, help

from friends and support from my family.

I would like to thank Bowel Diseases Research Foundation, Midland Gastroenterology

Society and East Midlands Business Development Agency for their financial support for

work on CRC related miRNAs. I would also like to thank, National Bowel Cancer

Screening Project audit committee, staff at the Department of Cancer Studies and

Molecular Medicine, Clinical Sciences Unit at Glenfield Hospital, Research Laboratory

for Kidney transplant and operating theatres at Leicester General Hospital. I would like

to pay my special thanks to Mr. Jaganathan Venkatesh, Dr. Karen Page, Dr. Ankur

Karmokar, Dr. Christopher Boes, Dr. Shona Potter, Mrs Angie Gillies, Mrs Linda Potter,

Stefan Hyman and Professor Chris Binns for their technical support during my studies

at University of Leicester.

Words cannot express how grateful I am to my father and mother for all of sacrifices

that they have made on my behalf. Their prayers for me are what have sustained me

thus far. I would also like to thank my brothers and sisters who supported me while I

was writing, and encouraged me to strive towards my goals. Finally, I would like express

appreciation for my beloved wife Rabia Aslam, who spent sleepless nights with my two

lovely daughters Zoha and Eman while I was stuck in the laboratory running my

experiments. She was always my support in the moments when there was no one to

answer my queries. She was always there cheering me up and stood by me through the

good times and bad.

4

3 Contents

Abstract ............................................................................................................................ 2

Acknowledgements .......................................................................................................... 3

Contents……………………………………………………………………………………………………………………..4

List of Tables ................................................................................................................... 10

List of Figures .................................................................................................................. 12

List of Abbreviations …………………………………………………………………………………………………14

1 Introduction ............................................................................................................ 20

1.1 Incidence of colorectal cancer ......................................................................... 20

1.2 Risk of CRC ....................................................................................................... 20

1.2.1 Non-modifiable risk factors ...................................................................... 22

1.2.2 Modifiable risk factors .............................................................................. 22

1.3 CRC staging ...................................................................................................... 22

1.3.1 The TNM staging system .......................................................................... 23

1.3.2 Dukes’ stage ............................................................................................. 25

1.3.3 Survivals from CRC in the UK .................................................................... 25

1.3.4 Biomarkers for high risk features in Dukes’ B cancers ............................. 29

1.4 CRC development and progression ................................................................. 30

1.4.1 Chromosome instability pathway (CIS) .................................................... 31

1.4.2 K-ras and BRAF mutations ........................................................................ 32

1.4.3 Microsatellite instability pathway and mismatch repair ......................... 33

1.4.4 Cell surface receptors ............................................................................... 34

1.5 What are miRNAs? ........................................................................................... 36

1.6 miRNA biogenesis in human cells .................................................................... 38

1.7 Mechanism of action and cellular function of miRNAs ................................... 40

1.8 Methods of miRNA analysis and quantification .............................................. 40

1.9 Role of miRNAs in CRC development .............................................................. 43

1.10 The need for biomarkers for the detection of colorectal neoplasia ........... 48

1.11 Utility of circulating satellite miRNAs for CRC detection and tumour-derived miRNAs in body fluids ...................................................................... 50

1.11.1 Circulating miRNAs in CRC patients ......................................................... 50

1.11.2 Plasma miRNAs as biomarkers for the detection and screening of CRC . 51

1.12 Circulating exosomal miRNAs for CRC ......................................................... 52

1.13 Aims and objectives of this study ................................................................ 56

5

2 Methods ................................................................................................................. 58

2.1 Ethical permission ............................................................................................ 58

2.2 Recruitment ..................................................................................................... 59

2.2.1 Inclusion criteria ....................................................................................... 60

2.2.2 Exclusion Criteria ...................................................................................... 60

2.3 Blood sample collection ................................................................................... 60

2.4 Processing of whole blood samples ................................................................ 62

2.5 Extraction of RNA from 1 ml plasma ............................................................... 63

2.6 Plasma Total RNA quantification ..................................................................... 63

2.6.1 RNA concentration with RNA Clean & Concentrator™-100 ..................... 64

2.6.2 RNA concentration with SpeedVac® concentrators ................................. 64

2.7 Tissue collection, preparation and storage of fresh frozen tissue .................. 64

2.8 Extraction of total RNA from snap-frozen tissue ............................................. 65

2.9 Formalin-fixed paraffin-embedded tissue sample collection ......................... 66

2.9.1 RNA Extraction from FFPE Tissue ............................................................. 69

2.9.2 DNA Extraction from FFPE Tissue ............................................................. 69

2.9.3 Mutation analysis of FFPE tissue .............................................................. 69

2.10 miRNA expression profiling .......................................................................... 73

2.10.1 Chemistry overview for miRNA expression profiling ................................ 73

2.10.2 Expression Profiling for plasma samples: ................................................. 81

2.10.3 miRNA Expression Profiling for High Risk Dukes’ B .................................. 88

2.10.4 Tissue miRNA expression profiling with TaqMan® miRNA Arrays ........... 88

2.11 QRT-PCR for validation cohorts ................................................................... 89

2.12 Isolation of exosomes from harvested cell line culture media by ultracentrifugation ..................................................................................................... 91

2.12.1 Isolation of exosomes from plasma samples by ultracentrifugation ....... 91

2.12.2 Transmission electron microscopy ........................................................... 92

2.12.3 Dynamic Light Scattering ......................................................................... 92

2.12.4 Immunoprecipitation: antibody coupling ................................................. 93

2.12.5 Flow cytometry to assess coupling of antibody with beads ..................... 94

2.12.6 Immunoprecipitation of exosomes with antibody-coupled Dynabeads ... 95

2.12.7 Immunoprecipitation of cell line exosomes with antibody-coupled Dynabeads .............................................................................................................. 95

2.12.8 Extraction of total RNA from exosomes isolated from harvested media and plasma ............................................................................................................ 95

6

2.12.9 Extraction of RNA from immunoprecipitated exosomes and supernatants........................................................................................................... 96

2.12.10 Immunoprecipitation of plasma exosomes with GPA33-coupled antibody. ………………………………………………………………………………………………………………………….97

2.12.11 miRNA expression analysis for exosomal RNA ..................................... 97

2.12.12 Cell sorting for stem cell-related miRNAs. ............................................ 97

2.13 Statistical Analysis: ....................................................................................... 98

2.13.1 Analysis of expression profiling to identify target miRNAs ...................... 98

2.13.2 Validation of diagnostic plasma miRNAs ................................................. 99

2.13.3 Analysis of tissue miRNAs for high risk Dukes’ B cancer ........................ 100

2.13.4 Software used for analysis ..................................................................... 100

3 Results .................................................................................................................. 102

3.1 Summary of results ........................................................................................ 102

3.2 Concentrating RNA samples .......................................................................... 104

3.2.1 RNA Clean & Concentrator™-100 and SpeedVac® concentrator ........... 104

3.2.2 Different elution volumes of total RNA extracted with miRvanaTM RNA isolation kit ................................................................................................... 106

3.3 Discovery phase: expression profiling of plasma miRNAs to identify discriminatory miRNAs for the detection of adenomas and carcinomas ................ 108

3.3.1 Total RNA concentrations for use in the miRNA array ........................... 108

3.3.2 Expression profiling array for plasma miRNAs ....................................... 110

3.3.3 Validation of Novel Circulating miRNAs for CRC Detection: Initial Validation Phase ......................................................................................... 122

3.3.4 Validation of miR-135b in an independent cohort ................................. 137

3.4 Discussion ...................................................................................................... 143

3.4.1 Role of miR-135b in colorectal neoplasia initiation and progression .... 143

3.4.2 Specificity of miR-135b for CRCs ............................................................ 145

3.4.3 miR-34a in circulation............................................................................. 146

3.4.4 miR-431 .................................................................................................. 147

3.4.5 Other studies investigating role of miRNAs for CRC screening .............. 147

3.4.6 Strengths and weaknesses of study ....................................................... 151

3.4.7 Clinical application of plasma miRNA based detection of colorectal neoplasia .............................................................................................................. 153

3.4.8 Conclusion .............................................................................................. 155

4 Results: Analysis of exosomal miRNAs ................................................................ 157

4.1 Summary of Results ....................................................................................... 157

4.2 Identification of exosomes on Transmission Electron Microscopy ............... 159

7

4.3 Dynamic Light Scattering to assess the size of exosomes isolated by ultracentrifugation ................................................................................................... 160

4.4 Flow cytometry (FACS) for detection of exosomes isolated by ultracentrifugation ................................................................................................... 161

4.5 Comparison of total RNA and miRNA concentrations for different volumes of plasma used for ultracentrifugation ...................................................... 163

4.6 Comparison of exosomal miRNAs with source miRNAs ................................ 165

4.7 Assessment of CD133 and CD326 (EpCAM/ESA) antibody coupling with beads by FACS .................................................................................................. 167

4.8 Isolation of exosomes with antibody coupled Dynabeads with FACS .......... 171

4.9 Analysis of exosomal miRNAs isolated through immune isolation and FACS ………………………………………………………………………………………………………………….173

4.10 Direct immune isolation of exosomes by antibody coupled Dynabeads ...... 175

4.11 Analysis of plasma exosomal miRNAs from immunoprecipitated exosomes using CD133, CD326 and GPA33 coupled Dynabeads.............................................. 178

4.12 Discussion ................................................................................................... 181

4.12.1 Selection of miRNAs and antibodies for this feasibility study ............... 181

4.12.2 Specificity of immunoaffinity isolated exosomal miRNAs for CRCs ....... 182

4.12.3 Literature review of exosomal miR-21 ................................................... 182

4.12.4 Use of CD326 for circulating exosomal miRNA analysis in non CRCSs... 183

4.12.5 Total plasma vs exosomal miRNAs ......................................................... 184

4.12.6 Role of exosomal miRNAs in cancer development and progression ..... 185

4.12.7 Limitations of the study.......................................................................... 188

4.12.8 Conclusion .............................................................................................. 188

5 Results: Tissue miRNAs ......................................................................................... 191

5.1 Summary of results ........................................................................................ 191

5.2 miRNA expression signature for different Dukes’ stages .............................. 193

5.3 Confirmation of Expression Profiling data with QRT-PCR ............................ 196

5.4 Validation of selected miRNAs on second cohort ......................................... 198

5.4.1 Tumour versus normal tissue miRNAs .................................................... 198

5.4.2 Dukes’ stage ‘low risk B’ versus ‘high risk B’ .......................................... 200

5.4.3 Mutation Analysis for cancer tissue ....................................................... 204

5.4.4 Correlation with other clinico-pathological variables ............................ 207

5.5 Discussion ...................................................................................................... 210

5.5.1 miR-135b and APC. ................................................................................. 210

5.5.2 miR-135b in metastatic cancers ............................................................. 210

5.5.3 Role of miR-15b in CRCs ......................................................................... 211

8

5.5.4 miR-21 as a prognostic marker .............................................................. 211

5.5.5 Combinatorial approach for prognostic tissue miRNAs in CRC .............. 213

5.5.6 miR-34a and miR-125a-5p ...................................................................... 214

5.5.7 miR-708 and miR-182 ............................................................................. 215

5.5.8 BRAF mutations in right colonic cancers ................................................ 215

5.5.9 Role of miRNAs in CRC metastasis ......................................................... 215

5.5.10 Other potential prognostic markers for CRCs ........................................ 216

5.5.11 Limitation of study ................................................................................. 217

5.5.12 Conclusion .............................................................................................. 219

6 Final Discussion .................................................................................................... 221

6.1 Discussion ...................................................................................................... 221

6.2 Conclusion ..................................................................................................... 225

7 Appendices ........................................................................................................... 228

7.1 Appendix I: Patient Information Sheet for colorectal disease progression .. 228

7.2 Appendix II: Tissue bank patient information sheet ..................................... 232

7.3 Appendix III: Consent form ............................................................................ 235

7.4 Appendix IV: Consent form for Colorectal Tissue Bank ................................. 237

7.5 Appendix V: NBSCP Screening Approval ........................................................ 239

7.6 Appendix VI: Plasma Array Participants Characteristics ............................... 241

7.7 Appendix VII: Table of Z-Score calculated for both adenoma and carcinoma expressions from miRNA data .................................................................................. 242

7.8 Appendix VIII: Clinicopathological characteristics of individual participants in control, adenoma and cancer groups .................................................................. 243

7.9 Appendix IX: Subgroup of controls for initial validation cohort. ................... 246

7.10 Appendix X: Initial Validation Cohort: Concentrations of Total RNA as detected on Nanodrop ND-1000 Spectrophotometer ............................................. 247

7.11 Appendix XI: CT values based expression levels of different miRNAs analysed for initial validation cohort ....................................................................... 250

7.12 Appendix XII: ∆CT values for expression levels of different miRNAs analysed for initial validation cohort ....................................................................... 252

7.13 Appendix XIII: ∆∆CT values for expression levels of different miRNAs analysed for initial validation cohort ....................................................................... 254

7.14 Appendix XIV: ROC analysis for the detection of adenoma and carcinoma by using miR-191 ...................................................................................................... 255

7.15 Appendix XV: Patient Characteristics for the final validation cohort ............ 256

7.16 Appendix XVI: Total RNA concentrations for final validation cohort ............ 258

7.17 Appendix XVII: Table of miR-135b for Normal, Adenoma and Carcinoma ... 259

9

7.18 Appendix XVIII: List of publications, grants, awards and presentations ....... 262

8 References ............................................................................................................ 265

10

4 List of Tables

Table 1: Comparison of TNM and Dukes’ staging systems. ........................................... 27

Table 2: Different types of RNAs. ................................................................................... 37

Table 3: Comparison of different detection systems for miRNAs. ................................ 42

Table 4: Summary of commonly expressed miRNAs in CRC tissue in comparison to adjacent healthy colonic mucosa. .................................................................................. 47

Table 5: Screening programmes for different cancers, their accuracy of detection and percentage reduction in cancer-related mortality from cancer research UK................ 49

Table 6: Isolation and characterisation of CRC cell line exosomes. ............................... 55

Table 7: Ethical permission with details of study title, principle investigator, ethics committee and relevant consent form versions. ........................................................... 58

Table 8: Patient demographics and clinicopathological variables. ................................ 68

Table 9: Primers and probes used for mutation analysis. .............................................. 71

Table 10: Mutation analysis by RT-PCR. ......................................................................... 72

Table 11: Thermal profiles for PCRs for mutation analysis. ........................................... 72

Table 12: Reagent concentrations for reverse transcription, pre-amplification and Array used for miRNA Taqman miRNA Array ................................................................. 79

Table 13: Thermal profiles for different PCR reactions. ................................................ 80

Table 14: Characteristics of participants for plasma miRNA expression profiling ......... 81

Table 15: Reagents for RT+ and RT- master mix. ........................................................... 83

Table 16: -RT reaction Master Mix for reverse transcription reaction. ......................... 83

Table 17: Master mix for preamplification reaction. ..................................................... 85

Table 18: Master mix for TaqMan® Array RT-PCR. ........................................................ 86

Table 19: Characteristics of cancer patients used for expression analysis by using freshly frozen tissue. ...................................................................................................... 89

Table 20: Characteristics of participants used for first validation of plasma miRNA panel. .............................................................................................................................. 90

Table 21: Total RNA concentrations measured with Nanodrop ND-1000 Spectrophotometer. ..................................................................................................... 109

Table 22: Z-scores calculated from miRNA array data for both adenoma and carcinoma samples. ........................................................................................................................ 112

Table 23: Complimentary miRNAs and their detection rates for neoplasia. ............... 113

Table 24: Summary of median performance scores and errors for 6 discriminatory miRNAs identified with Bioinformatics analysis. ......................................................... 114

Table 25: P-values for inter-group differences in expression levels of miRNA for adenoma or carcinoma groups in comparison to the control group. .......................... 115

Table 26: Panel of 29 miRNAs selected for further validation. .................................... 121

Table 27: ROC analysis for detection of the collective group of adenoma and carcinoma. .................................................................................................................... 123

Table 28: ROC analysis for the detection of adenoma. ................................................ 124

Table 29: ROC analysis for the detection of carcinomas. ............................................ 125

Table 30: Inter-group comparison of miR-135b, miR-431 and miR-34a. ..................... 126

Table 31: Median expression levels of miR-135b with associated p-values for differences in expression levels in groups with adenoma, carcinoma, and adenoma and carcinoma. ............................................................................................................. 137

11

Table 32: Summary of detection rates of adenomas and carcinomas based on CT values of miR-135b. ...................................................................................................... 139

Table 33: Summary of detection rates of adenoma and carcinoma based on ΔCT values of miR-135b. ................................................................................................................. 141

Table 34: Summary of in vitro studies investigating the role of miR-135b in non-CRCs. ...................................................................................................................................... 145

Table 35: Comparison of the sensitivity and specificity of different miRNAs for their utility as biomarkers for detection of adenocarcinoma and adenoma*. .................... 149

Table 36: Faecal miRNAs for CRC detection and screening. ........................................ 151

Table 37: Total RNA concentration for exosomes from plasma and HT29 harvested media. ........................................................................................................................... 163

Table 38: Sample characteristics and RNA concentrations for CD133 and CD326 bound exosomes isolated by FACS. ......................................................................................... 173

Table 39: Expression levels of different miRNAs for exosomes isolated with CD133 and CD326 bound Dynabead and FACS. ............................................................................. 174

Table 40: Comparative concentrations of RNA extracted from exosomes. ................ 175

Table 41: Total RNA concentrations of immunoaffinity isolated exosomes. .............. 178

Table 42: Summary of PCR-based validation of selected miRNAs on samples used for array. ............................................................................................................................ 197

Table 43: Linear regression analysis of miRNA expressions for Dukes’ stages. ........... 203

Table 44: Frequency table for different positive mutations detected for BRAF, K-ras and PIK3CA mutations. ................................................................................................. 205

Table 45: P-values for inter-group and intra-group comparisons between the three mutated genes. ............................................................................................................ 206

Table 46: Summary of studies dealing with potential role of different miRNAs for identification of high risk CRCs. .................................................................................... 214

12

5 List of Figures

Figure 1: Risk of CRC development in the UK population. ............................................. 21 Figure 2: TNM staging for CRC. ...................................................................................... 24 Figure 3: The Dukes’ staging system for CRC. ................................................................ 26 Figure 4: Comparison of five year survival rates for CRCs based on cancer stage ....... 27 Figure 5: Age-adjusted one, five and ten year survivals for different cancers. ........... 28 Figure 6: Biogenesis, processing and function of miRNAs. ............................................ 39 Figure 7: The mechanism of biogenesis and function of oncomiRNAs. ......................... 45 Figure 8: The proposed mechanism of biogenesis and function of tumour suppressor miRNAs ........................................................................................................ 46 Figure 9: Formation and release of exosomes by human cells. ..................................... 53 Figure 10: Reverse transcription of miRNA. ................................................................... 74 Figure 11: Preamplification with miRNA specific forward and reverse primers. .......... 74 Figure 12: RT-PCR for miRNA expression study. ............................................................ 76 Figure 13: MicroRNA expression profiling protocol flow chart. .................................... 78 Figure 14: miRNA array for expression profiling. ........................................................... 87 Figure 15: Total RNA concentrations and expression levels (CT) of SnRNA RNU6B and miR-21. .................................................................................................................. 105 Figure 16: Concentrations of RNA extracted with miRvanaTM RNA isolation kit. ........ 106 Figure 17: CT values of SnRNA RNU6B and miR-21 in different elution volumes of RNA extracted with miRvanaTM RNA isolation kit. ................................................... 107 Figure 18: Hierarchical cluster analysis. ....................................................................... 110 Figure 19: Principle component analysis for discriminatory miRNAs .......................... 111 Figure 20: ∆CT-based expression analysis of plasma miR-135b detected in expression profiling. ..................................................................................................... 116 Figure 21: ∆CT-based expression analysis of plasma miR-192* detected in expression profiling. ..................................................................................................... 117 Figure 22: ∆CT-based expression analysis of plasma miR-502-5P detected in expression profiling. ..................................................................................................... 118 Figure 23: ∆CT-based expression analysis of plasma miR-564 detected in expression profiling. ..................................................................................................... 119 Figure 24: Expression levels of miRNA and their matched miRNA* in plasma and tumours ................................................................................................................. 120 Figure 25: Comparative expression of miR-135b for different groups of controls and neoplasia. .............................................................................................................. 127 Figure 26: Comparative expression of miR-431 for different groups of normal and neoplasia. .............................................................................................................. 128 Figure 27: Comparative expression of miR-34a in different groups of controls and neoplasia……….………………………………………………………………………………………………………..129 Figure 28: ∆CT-based expression analysis of plasma miR-135b detected in the validation cohort……………………………………………………………………………………………………..131 Figure 29: ∆CT-based expression analysis of plasma miR-431 detected in validation cohort. …………………………………………………………………………………………………………………….132 Figure 30: ∆CT-based expression analysis of plasma miR-34a detected validation cohort……………………………………………………………………………………………………………………..133

13

Figure 31: Logistic regression probability values for ROC analysis of panel of miRNAs….. ............................................................................................. ……………….. 135 Figure 32: Logistic regression probability values for ROC analysis of selected miRNAs………………………………………………………………………………………………………………… 136 Figure 33: CT-based expression analysis of plasma miR-135b detected in an independent cohort ………………………………………………………………………………………. 138 Figure 34: ∆CT-based expression analysis of plasma miR-135b detected in an independent cohort…………………………………………………………………………………………….. 140 Figure 35: Electron micrographs of exosomes isolated from plasma by combination of filtration (100 nm) and ultracentrifugation without sucrose gradient……………… 159 Figure 36: The size distribution of exosomes………………………………………………………….160 Figure 37: A comparison of events detected for exosomes coupled to fluorescence stained exosomes surface antibodies (CD44 & CD326). .............................................. 162 Figure 38: Comparison of total RNA concentration for volume of plasma used to isolate exosomes isolated by filtration with 100 nm filter and ultracentrifugation… 164 Figure 39: Comparison of miRNA expression levels for different volumes of plasma used to isolate exosomes. ............................................................................................ 165 Figure 40: Comparison of miRNA expression for plasma exosomes, whole plasma, HT29 cell line harvested exosomes and HT29 cell lysate. ............................................ 166 Figure 41: Relative expression levels (∆CT) of miRNAs isolated from plasma and HT29 cell line culture. ................................................................................................... 166 Figure 42: FACS analysis of unstained beads. No fluorescence activity was detected by FACS ......................................................................................................................... 168 Figure 43: FACS analysis for CD326 (EpCAM/ESA)-APC coupled Dynabeads. ............. 169 Figure 44: FACS analysis for CD326 (EpCAM/ESA)-APC coupled Dynabeads. ............. 170 Figure 45: Comparison of FACS analysis for CD133-PE coupled beads bound exosomes to control CD133-PE coupled beads stored in Aldefluor solution at 4°C. .. 172 Figure 46: Expression levels of exosomal miRNAs from plasma and cell line exosomes isolated by immunoprecipitation with GPA33 coupled Dynabeads. .......... 175 Figure 47: FACS analysis for GPA33 coupled Dynabead bound exosomes isolated from HT29 cell line culture harvested media. .............................................................. 177 Figure 48: RNU6B normalised expression levels for different miRNAs of exosomes isolated by different types of antibody coupled Dynabeads. ...................................... 179 Figure 49: Diagnostic utility of exosomal miR-21 isolated with CD326 coupled Dynabeads. ..................................................................................................... 180 Figure 50: Significance analysis for microarrays (SAM). .............................................. 194 Figure 51: Hierarchical cluster analysis of expression profiling data for different Dukes’ stages. ............................................................................................................... 195 Figure 52: Comparison of RNU6B normalized miRNA expression (ΔCT) in cancerous and adjacent normal tissue. ......................................................................................... 199 Figure 53: Comparison of miRNA expression levels of different stages. ..................... 201 Figure 54: Plot of miR-135b expression for paired adjacent normal and cancerous tissue ............................................................................................................................ 202 Figure 55: Plot of miR-34a expressions and tumour T stages as identified on pathological examination. ............................................................................................ 208 Figure 56: Comparison of miRNA expression levels for tumours with histological evidence of EMVI and Mutant KRAS gene. .................................................................. 209

14

6 List of Abbreviations

ACFs Aberrant crypt foci

ACPGBI Association of Coloproctology of Great Britain and Ireland

AFAP Attenuated FAP

ANOVA One way analysis of variance

APC Adenomatous polyposis coli

ASB-PCR Allele specific binding PCR

AUC Area under the curve

BAX Bcl-2 associated X protein

Bcl2 B-cell lymphoma 2

Bcl-XL B-cell lymphoma-extra large

BER Base excision repair

BLAST Basic local alignment search tool

BNIP2 BCL2/adenovirus E1B 19 kDa protein-interacting protein 2

BRAF v-Raf murine sarcoma viral oncogene homologue B1

CA19.9 carbohydrate antigen 19-9

Cadherins Calcium dependant adherins

CAGR Cancer associated genomic regions

Cdc25a Cell division cycle 25 homolog A

Cdc42 Cell division control protein 42 homolog

cDNA Complementary deoxyribonucleic acid

CEA Carcinoembryonic antigen

cfDNA Circulating free DNA

CIMP CpG islands methylation phenotype

CIS Chromosome instability

c-myc myelocytomatosis viral oncogene homolog

COX2 Cytochrome oxidase subunit 2

CRC Colorectal cancer

CSMM Cancer studies and molecular medicine

CT Threshold cycle

CT Computerised tomography

15

D2O Sucrosedeuterium oxide

DLS Dynamic light scattering

DNA Deoxyribonucleic acid

DNMT3A DNA methyltransferase 3A

DSB Double strand break repair system

E2F Transcription factor E2F

E2F1 Transcription factor E2F1

E2F3 Transcription factor E2F3

ECM Extracellular matrix

EDTA Ethylenediamine tetra acetic acid

EGF Epidermal growth factor

EGFR Epidermal growth factor receptor

EMVI Extramural vascular invasion

EpCAM Epithelial cell adhesion molecule

ER-β Estrogen receptor beta

EU European union

EVI5 Ecotropic viral integration site 5 protein homolog

FACS Fluorescence-activated cell sorting

FAM 6-Carboxyfluorescein

FAP Familial adenomatous polyposis

FDA Food and drug administration

FFPE Formalin fixed paraffin embedded

FIH-1 Hypoxia-inducible factor 1-alpha inhibitor

FITs Faecal immunochemical test

FLI1 Friend leukemia integration 1 transcription factor

FOBT Faecal occult blood test

FOXO1 Forkhead box protein O1

GBM Glioblastoma multiforme

GPA33 Glycoprotein 33

GSK3-β Glycogen synthase kinase 3 beta

GTP Guanosine triphosphatase

16

H&E Haematoxylin and Eosin

HIF-1 Hypoxia-inducible factor 1

HIV Human immunodeficiency virus

HNPCC Hereditary non-polyposis colon cancer

HoxA7 Homeobox protein A7

HoxB8 Homeobox protein B8

HoxC8 Homeobox protein C8

HoxD8 Homeobox protein D8

HSF1 Heat shock factor protein 1

HSPGs Heparan sulfate proteoglycans

IARC International agency for research on cancer

IBD Inflammatory bowel disease

ICAMs Intercellular adhesion molecule

IHC Immunohistochemistry

IL-4 Interleukin -4

IMS Industrial methylated spirit

KRAS Kirsten Rat sarcoma viral oncogene homologue

LEMD1 LEM domain-containing 1

LFA-1 Lymphocyte function-associated antigen

LNA Locked nucleic acid

LNNMC Lymph node negative metastatic cancer

LOH Loss of heterozygosity

MAP MutY Homolog associated polyposis

MAPK Mitogen-activated protein kinases

Mcl1 Induced myeloid leukaemia cell differentiation protein Mcl-1

MDM2 Mouse double minute 2 homolog

MeV Multi-experiment viewer

MGB Minor groove binder

MHCII Major histocompatibility complex

MID1 Mouse double minute 1 homolog

MIRNA Micro ribonucleic acid

17

miRNA MicroRNA

MLH1 Mutt homologue 1

MMR Mismatch repair

mRNA Messenger RNA

MSH2 Mismatch repair protein 2

MSI Microsatellite Instability

MSI-H Microsatellite instability high

MSI-L Microsatellite instability low

MSS Microsatellite stable

MTCH2 Mitochondrial carrier homolog 2

MUTYH MutY homolog

MVB Multivesicular body

NBCSP NHS bowel cancer screening programme

NER Nucleotide excision repair system

NFIB Nuclear factor 1 B-type

NFQ Non-fluorescent quencher

NF-κB Nuclear factor kappa-light-chain enhancer of activated B cells

NTC No template control

PBS Phosphate-buffered saline

PDCD4 Programmed cell death protein 4

PI3K Phosphatidylinositol 3-kinase

PIK3CA Phosphatidylinositol-4,5-bisphosphate 3-kinase gene

piRNA Piwi-interacting RNA

piwi P-element induced wimpy testis

PMS2 Mismatch repair endonuclease 2

pri-miRNA Primary transcripts of miRNA

PTEN Phosphotase and tensin homolog

QRTPCR Quantitative reverse transcriptase polymerase chain reaction

RB Retinoblastoma

RECK Reversion-inducing-cysteine-rich protein with kazal motifs

RhoA Ras homolog gene family, member A

18

RhoB Ras homolog gene family member B

RISC RNA induces silencing complex

RNP Ribonucleoprotein

ROC Receiver operating characteristics

rRNA Ribosomal RNA

RT Reverse transcription

RT-PCR Real time polymerase chain reactions

SAM Significance analysis for microarrays

SDS Sodium dodecyl sulphate

siRNA Small interfering RNA

SIRT1 Sirtuin 1

SMAD 7 Mothers against decapentaplegic homolog 7

SNPs Single nucleotide polymorphisms

snRNA Small nucleolar RNA

STAT3 Signal transducer and activator of transcription 3

TCF T-cell factor

TE Tris EDTA

TEMS Transanal endoscopic mucosal surgery

TGF-α Transforming growth factor alpha

TGF-β Transforming growth factor beta

TIMP3 Metalloproteinase inhibitor 3

TLR Toll-like receptors

TME Transmission electron microscopy

TNM Tumour, nodal & metastasis

TPM1 Tropomyosin 1

tRNAs Transfer RNA

TSG TNF-stimulated gene

TYMS Thymidylate synthetase

UTR Untranslated region

VEGFR Vascular endothelial growth factor

Wnt Wnt-signalling pathway

19

Chapter 1: Introduction

miRNAs are novel biomarkers of

colorectal cancer

20

1 Introduction

1.1 Incidence of colorectal cancer

Colorectal cancer (CRC) is the third most common neoplasm worldwide. According to

the International Agency for Research on Cancer (IARC), approximately 1.24 million new

cases of CRC were detected worldwide in 2008 (Ferlay et al, 2008). It is the third most

common cancer in men (10.0% of the total) and the second most common in women

(9.4% of the total) worldwide. IARC data have shown that more than half of all CRC cases

occur in the more developed regions of the world, including Europe, America and Japan

(Ferlay et al, 2008).

In the European Union (EU27) alone 334,000 new cases of CRC were detected in 2008,

whereas in the UK approximately 38,000 people were diagnosed with CRC (National UK

Statistics). The incidence of CRC is on rise in Europe, particularly in Southern and Eastern

Europe, where rates were originally lower than in western part of Europe (Coleman et

al, 1993 & Bray et al, 2004). Contrary to the current trend in Europe, the incidence rate

of CRC in the USA has fallen in the last two decades (National Cancer Institute –

Surveillance, Epidemiology & End Result data - 2006).

Epidemiological studies have identified that a rapid trend of ‘westernization’, involving

a change in diet and lifestyle, has resulted in increased incidence rates of CRC in

developing countries (Marchand et al, 1999, Flood et al, 2000, Boyle et al, 2008, & Ferlay

et al, 2010). The occurrence of CRC is strongly related to age, with nearly 80% of cases

arising in people who are 60 years or older, although there has also been a recent

increase in CRC incidence in people younger than 60 years.

1.2 Risk of CRC

The lifetime risk for developing CRC in men is 1 in 14, whereas in women it is 1 in 18.4

(National Statistics, UK). People under the age of 64 have significantly lower risk (men

1.58% and women 1.1%) of CRC development. The risk increases significantly after the

age of 64 years, with a lifetime risk of 7.18% for men and 5.43% for women (Figure 1).

21

Figure 1: Risk of CRC development in the UK population. Bar graph compares the risk

of CRC for different genders in UK (Cancer Research UK).

CRC can be subdivided into hereditary (<5%), familial (20-25%) and sporadic (75%)

disease. Several factors related to increasing or decreasing risk of CRC are divided into

modifiable and non-modifiable risk factors.

22

1.2.1 Non-modifiable risk factors

Non-modifiable risk factors include personal or family history of CRC or adenomatous

polyps, and a personal history of chronic inflammatory bowel disease. Typically, CRC

cases present sporadically, with a family history of CRC occurring in 25% of patients

(Migliore et al, 2011). However, inherited mutations in major CRC genes occur in 5-6%

of patients, and in the rest the familial forms and gene-environment interactions lead

to disease (Jasperson et al, 2010). Some inherited conditions which predispose an

individual to CRC development are familial adenomatous polyposis (FAP), attenuated

FAP (AFAP), MutY Homolog (MUTYH)-associated polyposis (MAP), and Lynch

syndrome/HNPCC (hereditary nonpolyposis CRC). Rare syndromes include

hamartomatous polyposis conditions (such as Peutz-Jeghers syndrome and juvenile

polyposis syndrome) and hyperplastic polyposis (Aaltonen et al, 1993, Hemminki et al.

1997 & Migliore et al, 2011).

1.2.2 Modifiable risk factors

Environmental factors (such as physical inactivity, obesity, high consumption of red or

processed meats, smoking and moderate-to-heavy alcohol consumption) are known as

modifiable risk factors (American Cancer Society, 2011). However, environmental

factors are suggested to be the risk factors which play a major role in the aetiology of

this disease (Boyle and Leon, 2002).

1.3 CRC staging

Colorectal tumours are of epithelial origin and pathologically classified into three major

categories, namely hyperplastic polyps, neoplastic polyps (premalignant adenomas)

and cancers. The cancer category represents 95% of all colorectal tumours. CRC is

staged for its extent into the bowel wall and its spread to lymph nodes and distant

organs. Two commonly used systems for staging of CRC are the TNM and Dukes’

classifications.

23

1.3.1 The TNM staging system

The TNM system is the universally accepted classification system for the anatomic

extent of CRC spread (Sobin and Fleming, 1997) (Figure 2). The TNM staging system

describes the local extent of the primary tumour (T), its spread to regional lymph nodes

(N) and spread to distant body organs (M).

1.3.1.1 T stage

There are four stages of tumour size in bowel cancer:

T1: the tumour is limited to the mucosal layer

T2: the tumour has spread to the muscularis layer

T3: the tumour has spread to the serosa layer

T4: the tumour has spread beyond the serosal layer and into another part of the

bowel or other nearby organs or structures

1.3.1.2 N stage

N stage describes spread of cancer cells to adjacent lymph nodes:

N0: there are no lymph nodes containing cancer cells

N1: one to three lymph nodes close to the bowel contain cancer cells

N2: there are cancer cells in four or more nearby lymph nodes

1.3.1.3 M stage

M stage describes the spread of the cancer to distant organs (metastasis):

M0: the cancer has not spread to other organs

M1: the cancer has spread to other parts of the body (most commonly, to liver

and lungs)

24

Figure 2: TNM staging for CRC. Illustration of TNM staging showing extent and spread

of cancer within bowel wall (T), to lymph nodes (N) and to distant organs (M).

25

1.3.2 Dukes’ stage

For many years, the Dukes classification system (Dukes and Bussey, 1958) was the gold

standard for tumour staging used by pathologists worldwide. The Dukes' staging system

is divided into four groups: A, B, C and D. Dukes' A is an early bowel cancer and Dukes'

D is advanced (Figure 3).

Dukes’ A: cancer limited to mucosal layer

Dukes’ B: cancer spread to muscularis layer

Dukes’ C: cancer spread to adjacent lymph nodes

Dukes’ D: cancer spread to distant body organs (e.g. liver or lungs)

It is also common practice to use both TNM and Dukes’ staging in the UK to stage CRC.

The Dukes’ stage, which is based on histological examination of the resected specimen,

is more commonly used to guide post-operative treatment (Table 1).

1.3.3 Survivals from CRC in the UK

The survival and prognosis of patients suffering from CRC depends on the stage of the

tumour at time of detection. Five year survival significantly reduces from 95% for

localized early cancerous lesions to <10% for advanced metastatic cancers (Figure 4).

One year survival rates for colonic and rectal cancers are 75% and 80%, respectively.

Five and ten year survivals for both colonic and rectal cancers are similar and have been

reported as 60% (Cancer Research UK). Age-adjusted one, five and ten years survivals

are 76%, 59% and 57%, respectively (Figure 5, Cancer Research UK).

26

Figure 3: The Dukes’ staging system for CRC. Histological classification of CRCs based

on the spread of cancer to bowel wall, lymph nodes and distant organs.

27

Dukes Stage Spread of CRC TNM Stage

A Submucosa T1, N0, M0

Muscularis propria T2, N0, M0

B Beyond muscularis propria T3, N0, M0

Adjacent organs T4, N0, M0

C 1-3 lymph node metastasis T1-4, N1, M0

≥ 4 lymph nodes metastasis T1-4, N2, M0

D Distant organ metastasis T1-4, N0-2, M1

Table 1: Comparison of TNM and Dukes’ staging systems.

Figure 4: Comparison of five year survival rates for CRCs based on cancer stage. Bar

chart shows overall survivals for Dukes’ B stage are 80%, Dukes’ C 63% and Dukes’ D are

7% (Cancer Research UK)

28

Figure 5: Age-adjusted one, five and ten year survivals for different cancers. Survivals

for bowel cancer are weighted average derived from data for colon and rectum cancers

(Cancer Research UK). Bowel cancer is ranked 11th for its overall survival in comparison

to other cancers. Age-adjusted one, five and ten year survival are 76%, 59% and 57%,

respectively.

29

1.3.3.1 High risk Dukes’ B CRCs

Approximately 25% of patients with CRC are diagnosed with Dukes’ B or Stage II disease

(localized to the primary tumour site, no lymph node or distant metastases) and have a

five-year survival of 75-80% when surgically resected (Cancer Research, UK). Despite

surgical resection being highly effective for patients with Dukes’ B or Stage II disease, a

significant proportion (20–25%) of these patients develop fatal metastatic disease. This

has led to uncertainty about the role of post-operative chemotherapy for Dukes’ B or

stage II disease, as has been highlighted in the recently published third edition of

“Guidelines for the Management of CRC 2007” from the Association of Coloproctology

of Great Britain and Ireland (ACPGBI).

1.3.3.2 Identification of high-risk features in Dukes’ B cancers

A number of poor risk features can be identified in Dukes’ B cancers, such as serosal

involvement (T4), perforated or obstructed tumours, poorly differentiated or mucinous

histology and perineural or extramural vascular invasion (Guidelines for the

Management of CRC 2007” from the Association of Coloproctology of Great Britain and

Ireland). A combination of these features may confer a poorer prognosis in a node-

negative tumour. This has led to oncologists often referring to Dukes’ B cancers as ‘good

B’s’ and ‘bad B’s’. Chemotherapy is frequently offered to patients with ‘bad B’s’ (Dukes’

B carcinoma and the presence of adverse risk features). However, patients with Dukes’

B without poor risk histological features may still develop metastatic disease (O` Connor

et al, 2011).

1.3.4 Biomarkers for high risk features in Dukes’ B cancers

It is not possible to accurately predict the development of metastasis in Dukes’ B

tumours based solely on histological examination of resected cancer specimens.

Therefore, molecular markers are required to identify high risk Dukes’ B patients who

may benefit from post-operative chemotherapy. The progression of CRC results from

the sequential accumulation of genetic alterations in oncogenic and tumour suppressor

genes in colonic epithelium. Common somatic mutations found in CRCs are Kirsten rat

30

sarcoma viral oncogene homolog (K-ras), B-Raf proto-oncogene (BRAF) and

phosphatidylinositol-4,5-bisphosphate 3-kinase gene (PIK3CA). Although the frequency

of common somatic mutations (K-ras 30-40%, BRAF 15%, PIK3CA ~15% of CRCs) are

independent of tumour stage, mutation-specific gene expression profiles in other

cancers have successfully identified aggressive carcinomas (De Roock et al, 2010).

Furthermore, studies have also reported accelerated metastatic progression in patients

with the K-ras mutation. However, no study so far has looked at molecular biomarkers

in Dukes’ B tumour tissue to identify markers that may predict the development of

metastases after a curative resection of Dukes’ B tumours.

1.4 CRC development and progression

The prevailing view is to see CRC development as a multistep carcinogenesis arising

from multiple mutations in oncogenes and tumour suppressor genes, which in turn

cause up- and down-stream effects and numerous changes in mitogenic signalling

pathways (Khare et al, 2012 & Pancione et al, 2012). Proliferation studies have shown

that the rapid turnover and immense number of mitoses in the colon results in tens of

thousands of mutations occurring in the normal colonic mucosa each day (Barnes and

Lindahl, 2004). Very efficient genomic repair systems (called caretakers) such as the

mismatch repair system (MMR), the base excision repair system (BER), the nucleotide

excision repair system (NER), and the double strand break repair system (DSB)

continuously scan the genome for replication errors and mutations, and in many

instances also repair the genome. If mutations are too large or extensive to repair, the

cell is directed to apoptosis through a complex signal pathway, eventually shutting

down mitochondrial function (Wang and El-Deiry, 2003). Nevertheless, genomic repair

systems are not perfect, and occasionally mutations slip through the control systems.

The development of CRC requires a long exposure (decades) to carcinogens and an

accumulation of several mutations in key oncogenes and tumour suppressor genes. In

the case of tumour suppressor genes, both alleles must be silenced. In contrast, only

one oncogene is required for accelerated gene function. Therefore, for sporadic CRC it

takes several decades to acquire two hits on the two loci on one of the key tumour

31

suppressor genes. In patients with a strong family history of bowel cancer, their pre-

existing inherited mutation means that they only need to acquire one hit to knockout

the gene (Blanpain et al, 2013). The two major and well-established genetic pathways

leading to CRC are the chromosome instability pathway (CIS), representing and

characterising sporadic CRC, and the microsatellite instability pathway (MSI), which is

the principle pathway of hereditary non-polyposis colon cancer (HNPCC).

1.4.1 Chromosome instability pathway (CIS)

Sporadic CRCs develop through an adenoma-carcinoma pathway characterised by an

accumulation of mutations in key genes (Fearon and Vogelstein, 1990). However, it

does not explain why the majority of adenomas never acquire an invasive potential to

subsequently develop into cancer. This leads to an assumption that invasive potential

is acquired through multiple environmental factors and luminal events to culminate in

a certain lethal combination. The adenoma-carcinoma pathway is initiated by

inactivation of adenomatous polyposis coli (APC) (5q21) in the normal epithelium,

resulting in an accumulation of β-catenin that subsequently increases during stepwise

development. The next genetic event involves hypomethylation and occurs in

hyperplastic polyps. K-ras mutations are identified in slightly larger adenomas following

the loss of the 18q-arm during the transition to late adenomas. Several TNF-stimulated

genes (TSGs), such as Mothers against decapentaplegic homolog 7 (SMAD7), which are

involved in transforming growth factor beta (TGF-β) and WNT-signalling (Broderick and

Carvajal-Carmona, 2007) have been suggested as target genes for 18q loss. Ultimately,

loss of the 17p-arm includes the tumour protein 53 (p53) gene in the final progression

from late adenoma to carcinoma (Fearon and Vogelstein, 1990).

1.4.1.1 Adenomatous polyposis coli mutation (APC)

The APC gene, known as “the gatekeeper gene”, is involved in intercellular

communication, cell orientation, transcription and proliferation. The APC mutation is

found in both sporadic CRC and in FAP. Individuals with FAP carry an inherited mutation

in one APC allele. A second hit in the remaining allele usually inactivates the gene within

32

the first 30 years of life, resulting in hundreds to thousands of adenomas and

subsequent carcinomas. The APC mutation is seldom found in aberrant crypt foci (ACFs)

but occurs increasingly in adenomas and carcinomas (Jass et al, 2002 & Worthley et al,

2007). The APC protein modulates intercellular communication between colonocytes

through calcium-dependant adherins (cadherins). APC binds to the cytoplasmic domain

of the cadherin molecule together with two other molecules, β-catenin and glycogen

synthase kinase 3 beta (GSK3-β). The binding of β-catenin to the cadherin complex

secures low levels of free β-catenin in the cytoplasm. Loss of APC leads to β-catenin

translocation into the nucleus and upregulation of signalling through the WNT-

pathway, which accelerates proliferation and impairs differentiation and apoptosis.

Also, the loss of functional APC appears to interfere with normal mitosis, since APC-

deficient cells do not adequately detect replication errors during metaphase and

continue into anaphase, ultimately contributing to CIS (Draviam et al, 2006).

Furthermore, the APC mutation increases cytochrome oxidase 2 (COX2) activity, which

occurs with a simultaneous upregulation in epidermal growth factor (EGF) activity.

1.4.1.2 p53, the "Guardian of the Genome"

p53 is an important gene for maintaining genome stability. A p53 replication error or

mutation stops or slows down the cell cycle in G1/S phase to allow the repair of DNA

damage. If the DNA damage is too extensive to be repaired, p53 induces apoptosis

through the caspase pathway by shutting down mitochondrial function (Amaral et al,

2010). In unstressed cells, p53 is kept at a low level by continuous degradation. The

mutation in p53 is crucial for carcinogenesis to transform from a non-invasive to an

invasive disease. p53 mutations are found in adenomas (5%), malignant polyps (50%)

and invasive CRC (75%) with increasing frequency correlating with the extent of

malignancy (Suppiah and Greenman, 2013 & Bahnassy et al, 2014).

1.4.2 K-ras and BRAF mutations

The K-ras mutation is found in 30-50% of CRCs and provides the colonocytes with a

growth advantage, as guanosine triphosphatase (GTP) activity is lost with K-ras

33

mutation. This increases levels of GTP to result in constant signalling through

downstream pathways. The K-ras gene product (K-ras protein) is responsible for the

transduction of mitogenic signals from the (EGF) receptor (EGFR) on the cell surface to

the cell nucleus (Dobre et al, 2013). A primary K-ras mutation generally leads to a self-

limiting hyperplastic or borderline lesion and may be implicated in the serrated pathway

(Bettington et al, 2013 & Leggett and Whitehall et al, 2010), through which serrated

adenomas and carcinomas may also develop. Alone, the K-ras mutation is neither

sufficient nor necessary to drive the malignant transformation; this would require

additional “drivers” (Moon et al, 2014). K-ras mutations are frequently found in up to

95% of early dysplasias including in ACFs and also in hyperplastic polyps (Otori et al,

1997, Alrawi et al, 2006 & Feng et al, 2011). The sequence in which the K-ras mutation

occurs in relation to the APC mutation is also important. If a K-ras mutation occurs after

an APC mutation, the dysplastic lesion often progresses to cancer (Vogelstein and

Kinzler, 2004). BRAF is another downstream effector molecule of the K-ras pathway.

Wild-type BRAF CRCs are typically microsatellite stable tumours displaying CIS. Studies

have shown that BRAF mutations, also known as V600E, appear to be a valid indicator

of poor prognosis in CIS/microsatellite stable CRC (Bond et al, 2014).

1.4.3 Microsatellite instability pathway and mismatch repair

MSI results from a failure of the mismatch repair system (MMR) to correct base errors

and maintain genomic stability. Consequently, cells with abnormally functioning MMR

accumulate errors rather than correcting those (Wimmer et al, 2014). In humans, nine

genes with MMR function have been identified. Five out of nine MMR genes are

involved in HNPCC. These five genes and the frequency in which they are mutated are

mutt homologue 1 (MLH1, 49%), mismatch repair protein homolog 2 (MSH2, 38%),

mismatch repair protein homolog 6 (MSH6, 9%), mismatch repair endonuclease 2

(PMS2, 2%) and mismatch repair endonuclease 1 (PMS1, 0.3%) (Carethers, 2014). CRCs

can be divided into microsatellite instability high (MSI-H) if two or more MMR genes are

mutated, and microsatellite instability low (MSI-L) if only one mutation is found, or

microsatellite stable (MSS) (Poulogiannis et al, 2010). At least two mechanisms can

34

result in a defective MMR. An MMR gene mutation can result in a malfunctioning gene

product (protein), as occurs in HNPCC.

Alternatively, a silenced or under-produced MMR gene product can be caused by

hypermethylation, which can be seen in sporadic CRC (usually by silencing of MLH1).

Hypermethylation of a gene often leads to under-expression or “silencing”, and is hence

referred to as an epigenetic event (Carethers, 2014).

99.5% of human DNA is identical, but it is the pattern of microsatellites that makes each

individual’s DNA profile unique enough to become a DNA fingerprint (Jeffreys et al,

1985). A microsatellite is a non-coding stretch of DNA in which short sequences of

nucleotides are repeated many times. The repeated sequence is naturally occurring and

often simple, consisting of two to four nucleotides and can be repeated 3 to 100 times.

Hundreds of thousands of microsatellites are scattered throughout the genome

(Jeffreys et al, 1985). With loss of function of MMR, the lengths of microsatellites are

not replicated reliably, meaning that base mismatches are not corrected and new

microsatellite fragments of different lengths may be created. MSI and defective MMR

also increase the risk of strand slippage. When the polymerase complex reaches a

nucleotide repeat, the enzyme is temporarily released from the template strand and

the strand slippage occurs. The new strand detaches from the template strand and pairs

again with a repeat further upstream. MSI and strand slippage increase the risk of

mutations in nearby coding areas (Viguera et al, 2001).

1.4.4 Cell surface receptors

Cells have thousands of surface receptors. Those of importance to CRC are the growth

factor receptors. A growth factor receptor consists of a minimum of one, but commonly

several, proteins which are products of different proto-oncogenes (Heinemann et al,

2009). A cell surface receptor has an extracellular, transmembrane and intracellular

domain. The extracellular domain is a stereo-chemical site that only binds to specific

molecules known as ligands (Heinemann et al, 2009 & Deller et al, 2000). The

transmembrane domain is merely an ion-channel, whereas the intracellular domain,

which often utilises the actions of tyrosine kinases, triggers an intracellular signalling

35

cascade to mediate downstream cellular effects. Currently, at least two cell surface

receptors are considered important in the treatment of CRC: the EGFR and the vascular

EGF receptor (VEGFR).

1.4.4.1 EGFR

EGFR is expressed on the cell surface and its downstream signalling to the nucleus is

triggered when an appropriate ligand binds to the receptor. The main ligands of EGFR

are EGF and transforming growth factor alpha (TGF-α). Signalling from EGFR is

transmitted to the nucleus via the signal transducer proteins known as SMAD (Lo et al,

2001). The effects of EGFR signalling protect the cell from apoptosis, facilitate invasion

and promote angiogenesis through the activation of the mitogen-activated protein

kinases (MAPK) pathway (Oda et al, 2005). Studies have shown that EGFR protein is

overexpressed in 20-80% of CRCs, caused in part by gene amplification and also, but

rarely, by gene mutation (Di Fiore et al, 2010). The most important effector molecule in

the EGFR pathway is K-ras. A K-ras mutation leads to constant signalling through this

pathway. Additionally, the APC mutation increases COX2 activity in CRC. Increased

levels of pro-inflammatory cytokines contribute to proliferation and antagonise GSK-

3β. As a result, EGF activity is upregulated alongside COX2 upregulation. Interestingly,

one of the three domains of the COX enzyme is identical to EGF. Whether this is

responsible for increased EGF activity is unknown.

1.4.4.2 VEGFR

A systematic review concluded that angiogenesis is closely regulated by a range of pro-

and anti-angiogenic factors in normal conditions (Adams et al, 2007). Within solid

tumours hypoxic areas develop due to insufficient blood supply. Part of the hypoxic cell

response involves the induction of the transcription factor hypoxia-inducible factor 1

(HIF-1), which directly upregulates VEGF to promote new blood vessel formation

(Carmeliet et al, 1998 and Dewhirst, 2009). Increased VEGFR signalling has been

demonstrated in CRC (Tebbutt et al, 2010).

36

The sequential progression of colorectal neoplasia from adenoma to carcinoma

highlights that opportunities exist to improve cancer-specific survival by altering the

natural course of disease development. Such interventions could potentially be

chemotherapy preventive for high risk individuals, facilitate the early detection of

colorectal neoplasia, allow chemotherapy to down-stage the cancer prior to surgical

resection, and be beneficial therapy for palliation of symptoms in advanced stage

cancer.

Recent advances in proteomics and genomics provide a vast amount of information

about the role of micro-molecules in several cancer-related pathways. These advances

have focused on the detection of micro-molecules released from tumour cells and their

role in different cancers. The discovery of tumour-specific microRNAs (miRNAs) has

opened a new era of biomarker research that holds great potential for future cancer

detection strategies.

1.5 What are miRNAs?

miRNAs are single-stranded, evolutionarily conserved, small (17–25 ribonucleotides)

non-coding RNA molecules (Lee et al, 1993). miRNAs act as negative regulators of target

genes by directing specific messenger RNA (mRNA) cleavage or translational inhibition

by mediating the activity of RNA induced silencing complex (RISC) (Bartel et al, 2004 &

2009). So far, approximately 1400 mature human miRNAs have been described in the

Sanger miRBase, version 17, an international registry and database for miRNA

nomenclature, targets, functions and their implications in different diseases. In the

database, each mature miRNA in human and non-human species is assigned a unique

identifier number for universal standardization. For example, human miRNA 21 is

referred to as hsa-miR-21. Table 2 summarizes the different types of RNAs and their

size, mode of action and function in the human cells.

37

Types of RNA Size Mode of action Function

miRNA

17-25 Directs RISC Translational inhibition of

mRNA and gene

expression

mRNA 900-1500 Conveys genetic information

from DNA to the ribosomes

Directs and induces

protein synthesis

Small interfering

RNA (siRNA)

20-25

RNA interference and related

pathways

Interference of gene

expression

Piwi-interacting

RNA (piRNA)

26-31 RNA-protein complex

formation with piwi proteins

Transcriptional gene

silencing of

retrotransposons and

other genetic elements in

germline cells

Small nucleolar

RNA (snRNA)

70-200 Act as ribonucleoprotein (RNP)

complexes to guide the

enzymatic modification of

target RNAs at sites

determined by RNA:RNA

antisense interactions

Chemical modifications of

other RNAs, e.g.

methylation,

pseudouridylation

Transfer RNA

(tRNAs)

73 to 93

Transfers a specific active

amino acid to a growing

polypeptide chain at the

ribosomal site of protein

synthesis

Amino acid carriers and

protein synthesis during

translation

Ribosomal RNA

(rRNA)

120-5050 Decode mRNA into amino

acids

Protein synthesis in

ribosomes

Long non-coding

RNA (lncRNA)

>200 Binds to complementary RNA

and affect RNA processing

Pre and post transcription

regulation of genes

Table 2: Different types of RNAs.

Table shows sizes (number of nucleotides), mode of action and functions of different

RNAs in the human cells.

38

1.6 miRNA biogenesis in human cells

miRNAs are mostly transcribed from intragenic or intergenic regions by RNA

polymerase II into primary transcripts (pri-miRNAs) of variable length (1-3 kb). In the

nucleus, pri-miRNA transcripts are processed further by the nuclear ribo-nuclease

enzyme Drosha. This results in hairpin intermediate of about 70–100 nucleotides, called

pre-miRNA. Pre-miRNA is then transported out of the nucleus by a transporting protein,

exportin-5. Once in the cytoplasm, pre-miRNA is processed by another ribonuclease

enzyme, Dicer, into mature double-stranded miRNA. The two strands of miRNA (known

as a miRNA/miRNA* complex) are separated by Dicer processing. After strand

separation, the mature miRNA strand (miRNA-, also called the guide strand) is

incorporated into a complex with RISC, at which point the passenger strand, denoted

with a star (miRNA*), is degraded (Hammond et al, 2000, Lee et al, 2003, Bohnsack et

al, 2004 & Thimmaiah et al, 2005). This miRNA/RISC complex is responsible for miRNA

function (Figure 6). If on miRNA cloning or array the passenger strand is found at low

frequency (less than 15% of the guide strand), it is named miR*. However, if both

passenger and guide strand are equal in distribution, these two strands are named 5p

or 3p depending on their location within either 5' or 3' segments of the miRNA molecule

respectively. In this case, both strands have the potential to be incorporated into a

complex with RISC and subsequently fulfil a biological role, and in many cases miRNA*

strands are conserved and play an important function in cell homeostasis. However, the

functional role of the miRNA* strand has only recently been focused on in miRNA

studies. Well-conserved miRNA* strands may prove to be important links in cancer

regulation networks and thus deserve further study (Stark et al, 2007, Okamura et al,

2008, Zhou et al, 2010 & Guo et al, 2010).

39

Figure 6: Biogenesis, processing and function of miRNAs.

Figure illustrates the biogenesis of miRNAs in the nucleus, their transport into the

cytoplasm and their processing by Drosha and Dicer enzymes. It also shows the

involvement of RISC and miRNAs in different pathways of translational inhibition or

activation.

40

1.7 Mechanism of action and cellular function of miRNAs

The specificity of miRNA targeting is defined by Watson–Crick complementarities

between positions two to eight of the 5’ end of miRNA sequence with the corresponding

positions of the 3′ untranslated region (UTR) of their target mRNAs. When miRNA and

its target mRNA sequence show perfect complementarities, RISC induces mRNA

degradation. Should an imperfect miRNA–mRNA target pairing occur, translation into

the protein is blocked (Bartel et al, 2004 & 2009). Regardless of which of these two

events occur, the net result is a decrease in the production of the proteins encoded by

the mRNA targets.

Each miRNA has the potential to target a large number of genes; on average

approximately 500 genes are targeted by each miRNA family. Conversely, an estimated

60% of mRNAs have one or more evolutionarily conserved sequences that are predicted

to interact with miRNAs (Friedman et al, 2009). miRNAs have been shown to bind to the

open reading frame or to the 5′ UTR of the target genes and, in some cases, activate

rather than to inhibit gene expression (Ørom et al, 2008). It has also reported that

miRNAs can bind to ribonucleoproteins in a seed sequence in a RISC-independent

manner and then interfere with their RNA-binding functions, resulting in decoy activity

(Eiring et al, 2010). miRNAs can also regulate gene expression at the transcriptional level

by binding directly to the DNA (Khraiwesh et al, 2010), as illustrated in Figure 6.

1.8 Methods of miRNA analysis and quantification

Numerous approaches have been developed to analyse and quantify miRNA. A

commonly adopted strategy is to perform mass scale expression profiling to identify

signatures of miRNAs within small cohorts of patients, and thus identify the most

significantly deregulated miRNAs. Expression profiling is usually followed by validation

of the selected miRNAs on an independent cohort using hybridization-microarrays, real

time polymerase chain reactions (RT-PCR) and deep-sequencing (Meyer et al, 2010).

Most of these approaches are developed alongside northern blotting, the gold

standard. Each approach has its own advantages and disadvantages such as throughput,

sensitivity, ease of use and cost. RT-PCR can detect miniscule concentrations of RNA

41

with superior sensitivity and is both cost and time-effective (Chen et al, 2005).

Microarray-based techniques have the advantages of being relatively cost-effective,

quick and simple to use (Pradervand et al, 2010). Ultra-high throughput miRNA

sequencing allows de novo detection and relative quantification of miRNAs, but is

expensive and time-consuming during data generation and analysis (Wang et al, 2007).

A key component of miRNA detection and quantification that must be addressed is the

selection of endogenous controls for relative quantification. In RT-PCR-based detection

systems, several small nuclear and small nucleolar RNAs such as RNU6B are

recommended for normalising miRNA expression signatures and profiles in tissues, cell

lines and body fluid samples. However, RNU6B is unstable in high temperatures and

rapidly degrades, resulting in poor reproducibility and repeatability of experiments.

Accordingly, many researchers have selected invariant and the most stable miRNAs as

endogenous controls instead (Meyer et al, 2010). To overcome this problem of

normalization in RT-PCR and other detection systems, researchers have used different

statistical strategies including global mean expression, quantile, scaling and normalizing

factors. Of note, some normalization methods have been challenged whereas others

have been adapted to suit the specific nature of miRNA profiling experiments. At

present, there is no gold standard normalization strategy for any of the methods of

detection. Table 3 provides a comparison between different detection systems for their

practical application, throughput and cost and time expenditure.

42

Detection system

miRNA RT-PCR

expression profiling

miRNA-Array miRNA-

Sequencing

Method PCR Hybridization Deep

sequencing

Initial RNA

concentration

10 ng 100 ng 250 ng

Time required < 24 hours 24-48 hours > 1 week

Cost Low-medium for pool

profiling. Even lower

for custom designed

individual assays.

Low-medium for

Pool Profiling

High

Throughput Medium-high High Ultra-high

Utility Relative and absolute

quantification of

miRNAs

Relative and

absolute

quantification of

miRNAs

Relative

quantification

of known

miRNAs.

Identification

of novel

miRNA

sequences.

Table 3: Comparison of different detection systems for miRNAs.

Different detection systems with their practical application, throughput, cost and time

expenditure.

43

1.9 Role of miRNAs in CRC development

miRNAs play important roles in colorectal tumour biology including oncogenesis,

progression, invasion, metastasis and angiogenesis (Esquela-Kerscher et al, 2006;

Huang et al, 2008; Liu et al, 2011 & Lee et al, 2007). The initiation and progression of

colorectal neoplasia result from sequential accumulation of genetic alterations in

oncogenic and tumour suppressor genes in the colonic epithelium (Fearon et al, 1990).

miRNAs interfere with genetic mutations and are involved in different hallmarks of

cancer development, progression, invasion and metastasis.

Slaby and colleagues summarized the role of different miRNAs in CRC development and

emphasized the importance of APC, TP53 mutations and WNT signalling (Slaby et al,

2009). The initiation of colonic neoplasia is strongly linked to inactivation of APC gene

and activation of WNT signalling. APC inactivation has been found in more than 60% of

colonic tumours and such inactivation is associated with up-regulation of miR-135a/b

in colonic epithelial cells (Fearon et al, 1990, Segditsas et al, 2006, Nagel et al, 2008).

An accumulation of any further somatic mutations leads to further dysregulation of

miRNAs and the activation of additional downstream pathways. For example, let-7,

miR-18a* and miR-143 are strongly linked to K-ras knockdown and activation of the

EGFR-MAPK pathway (Akao et al, 2006; Chen et al, 2009 & Tsang et al, 2009). In

contrast, miR-21 and miR-126 are associated with either the augmentation or

inactivation of the phosphatidylinositol-3-kinase (PI3K) pathway, respectively

(Krichevsky et al, 2009 & Guo et al, 2008). Activation of these downstream pathways

results in autonomous tumour cell growth, survival and angiogenesis. Loss of p53 is a

critical step in the transformation of adenoma to adenocarcinoma, as indicated by the

finding that nearly 50-70% of colonic adenocarcinomas are found to exhibit p53

mutations (Fearon et al, 1990). miR-34a has been identified as a direct downstream

target of p53 and experimental replacement of miR-34a achieves p53-induced effects

of apoptosis and cell cycle arrest (Chang et al, 2007). A commonly up-regulated miR-17-

92 cluster (containing miR-17, miR-18a, miR-19a, miR-20a, miR-19b and miR-92a) also

drives the progression of adenoma to adenocarcinoma by up-regulation of c-myc

(Diosdado et al, 2009).

44

Another cancer pathway, known as the second serrated neoplasia pathway, has

recently been acknowledged, and for the most part is considered independent of APC

and TP53 mutations. It involves distinct molecular features of somatic BRAF mutation

concordance with high CpG islands methylation phenotype (CIMP-H) and MSI-H

associated with MLH1 methylation (Spring et al, 2006, Casey et al, 2005). Involvement

of miRNAs in the latter pathway is slowing emerging and would require further

functional studies to find a precise role for miRNAs.

Functional and mechanistic studies of miRNAs have shown that the replacement or

knockdown of distinct miRNAs in vitro results in distinct cytogenetic abnormalities

leading to either tumour cell proliferation or apoptosis (Huang et al, 2008).

Consequently, it is believed that deregulation of miRNA genes that target mRNAs of

tumour suppressor genes or oncogenes contributes to tumour formation by inducing

cell proliferation, invasion, angiogenesis and decreased apoptosis (Esquela-Kerscher et

al, 2006).

This has led to the belief that over-expressed miRNAs in tumour cells function by

inhibiting different tumour suppressor genes, whereas miRNAs that are often silenced

in tumour cells down-regulate the expression of oncogenes in normal tissue.

Amplifications, translocations, pleomorphisms or mutations in miRNAs transcribing

genes results in over production miRNAs (Figure 7). In contrast, mutations, deletions,

promoter methylations or any other abnormalities in the miRNA biogenesis results in

silencing of miRNAs in tumour cells (Esquela-Kerscher et al, 2006), as shown in Figure

8. Different studies have identified dysregulated miRNAs in CRC tissue specimens. Table

4 shows summarises some dysregulated miRNAs in colorectal tumour tissue compared

to adjacent normal colonic mucosa.

45

Figure 7: The mechanism of biogenesis and function of oncomiRNAs. Amplifications,

translocations, mutations or pleomorphisms in miRNA transcribing genes results in the

over-production of pri-miRNA and pre-miRNAs in the nucleus. Further processing by

Dicer results in higher levels of mature miRNAs in the cytoplasm. These overexpressed

miRNAs target tumour suppressor mRNAs in the cytoplasm and lead to the down-

regulation of mRNAs.

46

Figure 8: The proposed mechanism of biogenesis and function of tumour suppressor

miRNAs. Promotor hypermethylation/deactylation, homozygous/heterozygous

deletions, mutations or pleomorphisms in miRNA transcription genes result in under-

production or complete loss of pri-miRNAs. Defects in miRNA-processing machinery,

such as ineffective processing by Drosha/Dicer or defective pairing with RISC, can result

into inefficient levels of mature miRNAs in the cytoplasm. Low levels of tumour

suppressor miRNAs result in over expression of oncogenic mRNAs (Ras, Bcl2, Mcl1).

47

Studies Down-regulated miRNAs in CRC tissue

Up-regulated miRNAs in CRC tissue

Michael et al, 2003

let-7, miR-16, miR-24, miR-26a, miR-102, miR-143, miR-145, miR-200b

Volinia et al, 2006

let-7a-1, miR-9-3, miR-23b, miR-138, miR-218

miR-16, miR-17-5p, miR-20a, miR-21, miR-29b ,miR-141, miR-195, miR-199a

Xi et al, 2006 let-7b, let-7 g , miR-26a , miR-30a-3p, miR-132, miR-181a, miR-181b, miR-296, miR-320, miR-372

miR-10a, miR-15b ,miR-23a, miR-25, miR-27a, miR-27b, miR-30c, miR-107, miR-125a, miR-191, miR-200c, miR-339

Bandrés. et al, 2006

miR-133b, miR-145 miR-31, miR-96, miR-135b, miR-183

Akao et al, 2006

miR-143, miR-145, let -7

Nakajima et al, 2006

let-7 g, miR-181b, miR-200c

Lanza et al, 2007

miR-17-5p, miR-20, miR-25, miR-92, miR-93-1, miR-106a

Rossi et al, 2007

miR-200b, miR-210 , miR-224

miR-19a, miR-20, miR-21, miR-23a, miR-25, miR-27a, miR-27b, miR-29a, miR-30e, miR-124b, miR-132, miR-133a, miR-135b, miR-141, miR-147, miR-151, miR-152, miR-182, miR-185

Slaby et al, 2007

miR-31, miR-143, miR-145 miR-21

Monzo et al, 2008

miR-145 miR-17-5p ,miR-21, miR-30c, miR-106a, miR-107, miR-191, miR-221

Schepeler et al, 2008

miR-101, miR-145, miR-455, miR-484

miR-20a, miR-92, miR-510, miR-513

Schetter et al, 2008

miR-20a, miR-21, miR-106a, miR-181b, miR-203

Arndt et al, 2009

miR-1, miR-10b, miR-30a-3p, miR-30a-5p, miR-30c, miR-125a, miR-133a, miR-139, miR-143, miR-145, miR-195, miR-378*, miR-422a, miR-422b, miR-497

miR-17-5p, miR-18a, miR-19a, miR-19b, miR-20a, miR-21, miR-25, miR-29a, miR-29b, miR-31, miR-34a, miR-93, miR-95, miR-96, miR-106a, miR-106b, miR-130b, miR-181b, miR-182, miR-183, miR-203, miR-224

Slattery et al, 2011

miR-143, miR-145, miR-192, miR-215

miR-21, miR-21*, miR-183, miR-92a, miR-17, miR-18a, miR-19a, miR-34a

Table 4: Summary of commonly expressed miRNAs in CRC tissue in comparison to adjacent healthy colonic mucosa.

48

1.10 The need for biomarkers for the detection of colorectal neoplasia

Early detection and treatment is central for CRC management (Zlobec et al, 2008).

Therefore, recent emphasis has shifted towards bowel cancer screening (Walsh et al,

2003). A Cochrane systematic review of CRC screening using the faecal occult blood test

(FOBT) has shown that inviting men and women aged 45 to 74 for bowel cancer

screening using the FOBT could reduce the mortality rate of bowel cancer by 16.3%

(Hardcastle et al, 1996, Kronborg et al, 1996, UK CRC Screening Pilot Group, 2004,

Perkins et al, 2008). Based on the final evaluation report of the pilot and a formal

options appraisal, a national screening programme for bowel cancer was introduced in

UK. The programme in includes screening of men and women aged 60 to 70. It has been

estimated that by 2025, around 2,400 lives could be saved every year by the FOBT

testing element of the NHS Bowel Cancer Screening Programme (NBCSP). Other CRC

detection modalities available to clinicians for the detection of colorectal neoplasia

include flexible sigmoidoscopy, colonoscopy, barium enema, CT colonography, blood

and stool based biomarkers. A main stay of bowel cancer screening is the detection of

precancerous lesions (adenomas), and its cost effective detection potentially

determines the use of above mentioned detection modalities as a primary CRC

screening test. A comparative study of different screening modalities has shown the

sensitivity of detecting advanced adenomas of 20% for FOBT and 100% for colonoscopy

(Graser et al, 2008).

Despite its superior sensitivity for detection of both adenoma and carcinoma, the use

of colonoscopy for screening, however, is limited to high-risk individuals and for those

with a positive FOBT. Nnoaham and Lines analysed the cost projection and volume of a

screening programme (FOBT plus colonoscopy) for a hypothetical population of 500,000

individuals aged between 60 and 74 years in the South Central Strategic Health

Authority in the UK (Nnoaham and Lines, 2008). The projected cost of initial screening

by FOBT was £394,157 for the detection of 34 cancers. This cost, and the cost of

additional surveillance colonoscopies, would increase yearly. Table 5 compares

different aspects of the well-established screening programmes for breast and cervical

cancers to the recently established CRC screening programme in the UK. The significant

difference to note is the low sensitivity of detection of the FOBT.

49

Age of

population

screened

(years)

Reduction in

cancer-

related

mortality

Sensitivity

(%)

Specificity

(%)

Mammography for

breast cancer

(Blanks et al, 2000)

50–70 35 68–90 82–97

Cervical smear test

for cervical cancer

(Coste et al, 2003)

25–64 95 72 94

Faecal occult blood

test for bowel

cancer

(Hewitson et al,

2007)

60–69 16 40 (non-

hydrated)

*60 (hydrated)

> 90 (non-

hydrated),

< 90

(hydrated)

Table 5: Screening programmes for different cancers, their accuracy of detection and

percentage reduction in cancer-related mortality from cancer research UK.

Mortality from cervical cancer is comparatively low because of a screening test with a

high sensitivity (72%) and specificity (94%). The sensitivity of the faecal occult blood test

is only 40% with non-hydrated specimens. *The higher sensitivity (60%) with hydrated

specimens compromises specificity.

50

Currently, symptomatic CRC is primarily diagnosed through colonoscopy, a procedure

that is expensive, requires bowel preparation, sedation, and may be associated with

medical complications. Many patients delay or completely avoid having a colonoscopy

because of its invasive and unpleasant nature. Alternative, less-invasive diagnostic

methods such as FOBT are limited by low sensitivity. Zhu and colleagues have reported

that the Guaiac FOBT has 54% sensitivity and 80% specificity, and faecal immuno-

chemical test (FIT) has 67% sensitivity and 85% specificity for a cohort of symptomatic

patients (Zhu, et al 2011). Other potential diagnostic biomarkers include the

carcinoembryonic antigen (CEA) in blood (Duffy et al, 2001 & Fakih et al, 2006), but this

also has limited sensitivity and specificity (36-74% and 87%, respectively). Hence, a

more accurate and precise blood test that can diagnose CRC earlier is desirable.

1.11 Utility of circulating satellite miRNAs for CRC detection and tumour-derived

miRNAs in body fluids

miRNAs are present in body fluids, especially blood, and are potentially useful clinical

biomarkers. Recent studies have shown that tumour-derived miRNAs are present in

human serum in remarkably stable form and are protected from endogenous

ribonuclease activity (Mitchell et al, 2008). These tumour-derived miRNAs are present

in circulating blood at levels sufficient to be used as measurable biomarkers for the

detection of tumours (Mitchell et al, 2008 & Gilad et al, 2008). The levels of plasma and

serum miRNAs correlate strongly, suggesting that either plasma or serum can be used

for investigating these blood-based biomarkers. The detection of placental miRNAs in

maternal serum at levels that increase with gestational age reveals their potential to

act as biomarkers for diverse physiological and pathological conditions (Chim et al,

2008, Hromadníková et al, 2010, Kotlabova et al, 2011).

1.11.1 Circulating miRNAs in CRC patients

Given that aberrantly expressed miRNAs in CRC tissue are transported into blood,

circulating miRNAs can potentially serve as non-invasive markers for CRC detection. In

2008, Chen and colleagues used a high-throughput sequencing technique and

compared the miRNA expression profiles of patients with CRC and healthy controls

51

(Chen et al, 2008). miRNA expression profiles of CRC and healthy controls were

significantly different. However, more than 75% of the aberrantly expressed miRNAs

detected in the serum of CRC patients were also present in the serum of patients with

lung cancer. A similar trend was also observed in another study in which expression

profiles generated from plasma of breast cancer patients were compared with CRC and

other solid organ cancers (Heneghan et al, 2010). Identification and quantification of

cancer-related circulating miRNAs are associated with challenges in terms of sample

preparation, experimental design, pre-analytic variation, selection of diagnostic

miRNAs, data normalization and data analysis. Many of these obstacles have recently

been addressed and now act as a guide to help overcome these issues (Meyer et al,

2010 & Kroh et al, 2010).

1.11.2 Plasma miRNAs as biomarkers for the detection and screening of CRC

The universally applicable, simple, low cost and sensitive technique of RT-PCR for the

detection and quantification of miRNA (Chen et al, 2005) in the blood may be a more

efficient and reliable method of screening individuals with non-cancer diseases and

cancers. Once the accuracy of detection has been established, an assessment of cost

and acceptability by the population in terms of use as a mass screening technique is be

required. A model of RT–PCR for serum RNA analysis suggests that, when used for a

pool of 84 patients, costs was less than £1 per clinical specimen (Rouet et al, 2007). The

RT-PCR-based detection technique is used in hospitals for various viral screenings (such

as hepatitis C and human immunodeficiency virus) and genetic testing for genotyping.

With the addition of RNA extraction and modifications in the RT-PCR assay protocols,

the technique can be adapted for miRNA study. The time taken for a single RT-PCR run

is 4–6 hours (Chen et al, 2005) and multiple samples can be processed together (Rouet

et al, 2007). Thus, the results should be available within 24–48 hours. miRNAs can be

introduced into various aspects of the national bowel cancer screening programme. The

best option might be a yearly miRNA blood test in primary care, with colonoscopic

assessment for those with a positive result. At present, approximately 50% of

individuals with a positive FOBT subjected to colonoscopic examination have normal

52

findings. The replacement of the FOBT with the more sensitive miRNA test should

reduce the rate of negative colonoscopies.

Though the analysis of circulating miRNAs in CRC patients has identified several

diagnostic miRNAs, their diagnostic accuracy is still questionable. This has been due to

overlapping miRNA expression with other cancers, non-cancerous conditions and

variability of individual miRNA expression with stage and grade of tumour. It is possible

that common carcinogenesis-related miRNAs are shared by different types of tumours

and investigators are detecting cancer-related but not tissue specific miRNAs.

Another explanation of the findings is that miRNAs released into the circulation

originate from immune cells taking part in a systemic immune response to the tumour,

causing abnormal proliferation of colonic cells (Dong et al, 2011). This might also explain

the finding of commonly deregulated miRNAs in patients with CRC and ulcerative colitis

(Pekow et al, 2011). Furthermore, studies have focused on measuring circulating levels

of either single miRNAs or a subset of known miRNAs. Due to reasons mentioned above,

a single miRNA-based detection strategy would be ineffective, whereas a CRC tissue-

specific expression signature generated from plasma or serum of patients with CRC and

adenoma could be more informative and accurate.

1.12 Circulating exosomal miRNAs for CRC

The recent discovery of exosome-mediated transport of cancer-related miRNAs into the

circulation has shifted the focus of miRNA studies towards the isolation of tissue-

specific circulating exosomes and their encompassed miRNAs. Exosomes are

membrane-bound small vesicles (20-100 nm in diameter) and are released via

endocytosis by a variety of cells in both healthy and disease conditions (Théry et al,

2002, Keller et al, 2006). Exosomes correspond to internal vesicles of multivesicular

bodies (MVBs) and are released into the extracellular environment upon fusion of MVBs

with the plasma membrane, (Théry et al, 2002, Cocucci et al, 2009). Figure 9 shows the

formation and release of exosomes from human cell.

53

Since exosome formation involves two inward budding processes, exosomes maintain

the same topological arrangement as the cell, with membrane proteins on the outside

and some cytosol on the inside. Exosomes contain cytoplasmic proteins, miRNAs and

mRNA transcripts (Valadi et al, 2007). The topical orientation of the exosomal

membrane may help identify their source and could be achieved using surface antigen-

directed antibodies such as anti-MHCII. One drawback of this isolation method is that

unless all exosomes contain the specific surface antigen used for the isolation, only a

fraction of exosomes will be isolated.

Figure 9: Formation and release of exosomes by human cells.

Exosomes are formed by double invagination of cell membrane and are released by exocytosis.

54

Circulating exosomes can also be isolated based on their size, density and surface

proteins. A commonly used method of purifying exosomes involves removal of cells and

debris by either a filtration process or by a series of centrifugations (differential

centrifugation), followed by a final high-speed centrifugation (ultracentrifugation) to

pellet the exosomes. Exosomes have a specific density and can be purified by floatation

onto a sucrose density gradient or by sucrosedeuterium oxide (D2O) cushions. Another

purification method is based on exosome size and utilizes chromatography. The size and

characterisation of exosomes is performed using transmission electron microscopy,

immune-electron microscopy, flow cytometry and dynamic light scattering. There is,

however, a growing need for a fast and reliable method that yields a highly purified

exosome fraction.

Based on this immunoaffinity strategy, several groups have isolated exosomes from the

blood of patients with different cancers and have performed miRNA expression profiles

on total RNA isolated from these purified and probably tumour-specific exosomes

(Taylor et al, 2008, Logozzi et al, 2009, Rabinowits et al, 2009). Patients with cancer have

relatively higher quantities of exosomes and miRNAs in the circulation Rabinowits et al,

2009). For instance, the analysis of miRNAs extracted from circulating exosomes in

patients with ovarian cancer has been proven to be equivalent to ovarian tissue biopsies

(Taylor et al, 2008). By using a similar approach of isolation and analysis, exosomal

miRNAs in CRC can be evaluated for their diagnostic accuracy and may provide a

breakthrough in diagnostic modality. Table 6 summarizes the exosome isolation and

characterisation methods used by different groups to analyse exosomes specific to CRC

cells and methods of isolation of circulating exosomes for miRNA analysis of other

cancers (Simpson et al, 2009). There is, however, a growing need for a fast and reliable

method that yields a highly purified exosome fraction.

55

Studies CRC cell

lines

Isolation method Characterisation and

validation of exosomes

Huber et al,

2005

SW403

1869col

CRC28462

Differential

centrifugation

Transmission electron

microscopy

Immune electron microscopy

Fluorescence-activated cell

sorting (FACS)

Western Blotting

Mathivanan

et al,2010

LIM1215 Filtration,

diafiltration (5K)

ultracentrifugation

Immuoaffinity

Transmission electron

microscopy

Immune electron microscopy

Western blotting

Choi et al,

2007

HT29 Differential

centrifugation

Diafiltration(100k)

Density gradient

Transmission electron

microscopy

Western blotting

van Nigel et

al, 2001

HT29-19A

T84-

DRB1*0401/

CIITA

Differential

centrifugation

Density gradient

Transmission electron

microscopy

Immune electron microscopy

Western blotting

Isolation and characterisation of circulating exosomes for miRNA analysis

Studies Cancer type Isolation method Specific method/ technique

Logozzi et

al, 2009

Malignant

melanoma

Ultracentrifugation

and filtration

400 x g 20 min isolate plasma

1,200 x g 20 min

10,000 x g 30 min and filter

through 0.22 μm filter

1,00,000 x g 60 min

Rabinowits,

et al, 2009

Lung cancer Immunoaffinity

Ultracentrifugation

anti-EpCAM-coated

immunobeads

Taylor et

al, 2008

Ovarian

cancer

Immunoaffinity

Ultracentrifugation

anti-EpCAM antibody-coated

immunobeads

Table 6: Isolation and characterisation of CRC cell line exosomes.

Table shows different isolation techniques used in previous studies for exosomal

isolation in vitro. Table also shows isolation techniques for circulating exosomes in

patients with different cancers.

56

1.13 Aims and objectives of this study

This study aimed to:

1) Identify which circulating miRNAs can be used for the detection of colorectal

neoplasia.

The objectives were to:

a) Develop and compare miRNA expression profiles for RNA isolated from plasma

of participants with colorectal adenomas, carcinomas and participants without

any significant colonic pathology on colonoscopic examination.

b) Apply discriminatory miRNAs to a larger cohort to assess their diagnostic

accuracy for the detection of adenoma and carcinoma.

2) Examine the feasibility of using plasma exosomal miRNAs for the detection of CRCs.

The objective was to:

a) Evaluate different exosomes isolation techniques to isolate and analyse

exosomal miRNAs in the plasma of patients with CRC.

3) Examine the utility of tissue miRNAs combined with common gene mutations in

CRC as biomarkers to predict the development of metastasis in patients with high

risk Dukes’ B cancers.

The objectives were to:

a) Extract total RNA and DNA from formalin-fixed paraffin-embedded (FFPE)

cancer and adjacent normal tissue specimens.

b) Identify specific tissue miRNAs associated with different stages of CRC.

c) Screen for common gene mutations (K-ras, BRAF, PIK3CA) in primary CRC tissue.

d) Combine data to predict the development of metastasis in patients with Dukes’

B cancers.

57

Chapter 2: Methods

58

2 Methods

2.1 Ethical permission

Ethical permission was obtained from the National Patient Safety Agency and National

Research Ethics Committee for Leicestershire, Northamptonshire and Rutland (Table 7).

The ethics committee gave favourable opinions for ‘markers of tumour progression in

CRC’ and the development of a ‘colorectal tissue bank’. The ethics committee gave a

favourable opinion to collect blood and human tissue samples from the potential

participants for ‘markers of tumour progression in CRC’, and for storage of leftover

tissue from the study in the ‘colorectal tissue bank’ at the Department of Cancer Studies

and Molecular Medicine (CSMM) at the Robert Kilpatrick Clinical Sciences Building at

the University of Leicester. The ethics committee permitted the recruitment of

participants for the ‘colorectal tissue bank’. A validated and committee-approved

participant information (PI) sheet and consent form was used for ‘markers of tumour

progression in CRC’ and ‘colorectal tissue bank’ (Appendix I-IV).

PI Reference Study Title Ethics Committee PI and consent

versions

JH

Pringle

05/Q2502/27 Colorectal

tissue bank

Leicestershire,

Northamptonshire and

Rutland Research Ethics

Version 7.0

Version 8.0

JH

Pringle

05/Q2502/28 Markers of

tumour

progression

in CRC

Leicestershire,

Northamptonshire and

Rutland Research Ethics

Version 6.0

version 7.0

Table 7: Ethical permission with details of study title, principle investigator, ethics committee and relevant consent form versions.

59

2.2 Recruitment

For the plasma miRNA study, recruitment commenced in October, 2008, and finished

in December, 2011. A total of 265 participants agreed and contributed towards sample

collection. Participants were recruited from University Hospitals of Leicester NHS Trust.

Permission was obtained from Research Committee for National Bowel Cancer

Screening Programme (NBCSP) to include asymptomatic participants attending

screening colonoscopy at Glenfield General Hospital screening hub (Appendix V).

Symptomatic patients attending day case endoscopy unit for the diagnostic

colonoscopy for their bowel symptoms were also invited.

Patients with disease (CRC and polyps) were recruited from the Department of

Colorectal Surgery at Leicester General Hospital, University Hospitals of Leicester NHS

Trust. Potential participants were identified through colorectal specialist clinics, CRC

multidisciplinary team meetings and pre-assessment clinics. All consecutive patients

under the care of two consultant colorectal surgeons (Mr. John Jameson and Mr. Baljit

Singh) with histological diagnosis of CRC and undergoing curative resection of cancer

with laparoscopic/open or Transanal Microscopic Surgery (TEMS) at Leicester General

Hospital, University Hospital of Leicester NHS Trust, were invited to participate in the

study. Patients attending follow up outpatient clinic appointments after colonoscopy

examination were also invited to participate in the study.

Participants with normal colonoscopy or findings of diverticular disease, haemorrhoids

or mild colitis were used as controls. Participants with diagnosis of cancer, polyps of any

type (except with hyperplastic polyps) were used as the diseased group. All potential

participants were invited via a postal invite letter which included approved participant

information sheets.

Written consent was obtained from each participant in outpatient/endoscopy/pre-

assessment clinics on the day of procedure. Participants with language requirements

were explained the patient information in their native language (Hindi, Punjabi, Urdu)

and given extra time for decision making to participate. When necessary, official

translators and language lines were used to provide an explanation of the study and to

obtain consent for participation.

60

Blood specimens were taken 24 hours prior to surgery from patients undergoing

surgical resection of CRCs and large polyps. Blood samples were collected from

participants 1-2 hours prior to colonoscopy procedure in the endoscopy unit. Blood

samples were collected from participants in outpatient surgical clinic on the day of their

follow-up appointment.

2.2.1 Inclusion criteria

I. Patients aged 25-90 years

II. Patients undergoing surgical resection of colorectal neoplasia

III. Patients undergoing surgical resection of bowel diverticular disease

IV. Asymptomatic healthy controls without any bowel symptoms

V. Patients undergoing colonoscopic examination of large bowel for:

Family history of CRC or IBD

Surveillance after CRC and polyp resection

Surveillance for dysplasia in the background of IBD

Surveillance colonoscopy for FAP/HNPCC

Symptoms of bowel disease

Positive FOBT

2.2.2 Exclusion Criteria

I. Synchronous carcinoma of other body organ

II. Age >90 and <25 years

III. Pregnancy

IV. HIV

V. Hepatitis C with or without anti-viral treatment

VI. Patients with needle phobia

2.3 Blood sample collection

The blood sampling procedure was explained to all potential participants. Every

participant was informed that the procedure involved wiping an area of skin with

alcohol to sterilize and then puncturing a suitable vein with a 21 gauge needle.

Participants were explained that 10-15 ml whole blood would be drawn. Participants

61

uncomfortable with the procedure and/or allergic to alcohol swipes were not asked to

donate blood. Patient details were checked against details on the patient consent form

and blood sampling tubes were labelled with study number, participant unique

identification number, hospital number, and date and time of collection.

All blood samples were collected by the investigator (Mr Aslam) who had essential

training and prior experience in blood sampling from NHS patients. The investigator

wore latex/vinyl-free gloves in case of latex allergy throughout the procedure.

Participants were also informed of the local trust protocol for accidental needle stick

injury. During the blood sampling procedure, research participants were reminded that

they could ask any questions about the procedure and even withdraw their consent in

cases of difficult sampling. For the majority of patients the medial cubital vein was

sampled from, although other veins on the forearm and back of the hand were also

used in some patients. The skin superficial to the identified vein was cleaned with an

alcohol wipe and a tourniquet was applied 5-10 cm above the intended site of venous

puncture. Blood samples were collected using a closed, vacutainer system. The needle

attached to the vacutainer was unsheathed and inserted through the skin and into the

vein at an angle of 15-30°.

Two 7.5 ml blood sampling tubes containing ethylenediamine tetra acetic acid (EDTA)

as a preservative were used to collect 10-15 ml whole blood. Upon completion of blood

collection the tourniquet was removed, needle withdrawn and sterile cotton applied at

the site of venous puncture and secured with a Micropore (Micropore Technologies,

Derbyshire, UK). The participant was asked to apply pressure for up to five minutes to

stop the bleeding and reduce the risk of bruising. The vacutainer and needle were

disposed in the sharp incarceration bin and blood sampling tubes were gently inverted

to mix anticoagulant with the whole blood. Participants were thanked for their

participation and advised about reporting any complications arising from the blood

sampling procedure.

After completion of blood sampling, all non-sharp clinical waste was disposed of in the

clinical waste bin. Strict hand hygiene and aseptic techniques were used. Blood sample

tubes containing samples were stored at room temperature and processed within 2

62

hours of sampling time. Where blood samples were drawn from two or more

participants, extra caution was taken for sample labelling and storing the blood tubes

in different plastic carrier packs.

2.4 Processing of whole blood samples

Whole blood was separated into plasma and cellular fractions within 2 hours of blood

sampling. Whole blood samples in EDTA were processed at Clinical Sciences Buildings

of three different sites of University Hospitals of Leicester NHS Trust. The Clinical

Sciences Buildings are within five minutes walking distance from the clinical areas

where blood sampling was performed.

EDTA blood sampling tubes were placed in a Jouan centrifuge. The tubes were

centrifuged at 850 x g for 10 min at 10°C. The supernatant plasma layer from two EDTA

tubes for each participant was removed and decanted into a 15 ml BD Falcon 352096

polypropylene conical centrifuge tube (BD Bioscience, Bedford, USA). Isolated plasma

in Falcon tubes was centrifuged again at 850 x g for 10 min at 10°C. Supernatant plasma

was removed and aliquoted into 1 ml samples in multiple 1.5 ml eppendorfs. Buffy coat

was carefully aspirated and decanted into 1.5 ml eppendorfs. Packed red blood cells

(RBC) in the EDTA tube were removed and aliquoted in 1-2 1.5 ml eppendorfs.

Each eppendorf was labelled with a study number, participant’s unique identification

number and date of collection. Plasma, buffy coat and packed RBCs in eppendorf were

stored at -80°C for later use. Sample and aliquot numbers were recorded on a

spreadsheet of patient samples. All samples were placed in -80°C freezer in Room 304

of the Robert Kilpatrick Sciences Building at the University of Leicester. Freezer

temperature was checked on daily basis and recorded in the laboratory freezer recoding

data. When aliquots were removed the spreadsheet was updated accordingly.

63

2.5 Extraction of RNA from 1 ml plasma

1 ml plasma in 1.5 ml eppendorfs was thawed at room temperature. 100 μl 5 N acetic

acid (1 ml glacial acetic acid + 2.48 ml sterile ultrapure water) was added to 1 ml plasma.

The plasma was vortexed and 1.1 ml was transferred into a 15 ml BD Falcon centrifuge tube.

3750 μl of T9424 TRI Reagent® Solution (Sigma-Aldrich, USA) was added to the mixture in

the Falcon tube, vortexed and incubated at room temperature for 5 min. 1 ml of chloroform

was then added and the sample was vortexed and incubated at room temp for 3 min. The

sample in the Falcon tube was centrifuged at 4000 rpm for 15 min at 4⁰C. The aqueous

phase was transferred to a clean 15 ml Falcon tube. The volume of the aqueous phase

was measured and 1.25 times the volume of 100% ethanol (Hayman Speciality Products,

Essex, England) was added and vortexed.

The above volume was applied to a mirVana miRNA isolation column (Ambion®-

AM156), as per the manufacturer’s protocol. The column was centrifuged at 6500 rpm

for 15 sec. Flow through from the column was discarded. The column was filled with

700 μl of Wash Solution 1 (mirVana™ miRNA Isolation Kit), centrifuged at 6500 rpm for

15 sec and flow through was discarded. The column was washed twice with 500 μl of

Wash Solution 2/3 at 6500 rpm for 15 sec.

After discarding the flow through from the second wash, the column was transferred to

a clean eppendorf and centrifuged at 6500 rpm for 1 min to dry the membrane. The

column was transferred to a clean labelled eppendorf and 50 μl of pre-heated RNase-

free water (95⁰C) was added to the column. The column in the eppendorf was

centrifuged at 13000 rpm for 30 sec and total RNA was transferred to an eppendorf

before being stored at -20⁰C.

2.6 Plasma Total RNA quantification

A Nanodrop ND-1000 Spectrophotometer (Thermo Scientific, USA) was used to

measure the total RNA concentration in each sample. The pedestal was wiped with a

dry tissue paper before the start of measurement. The spectrophotometer was

initialised as per the user guide. 1.2 µl of RNAase-free water was used as a blank to

achieve a reading of 0.0 ng/l. For sample RNA quantification, the sample was mixed well

prior to measurement, pedestal and top were cleaned, 1.2 µl of sample was applied to

64

pedestal and the lid was closed. Sample ID was typed into the software. The instrument

provided a total RNA concentration in ng/l and 260/280 value for each sample. Sample

values were recorded with corresponding 260/280 values. Pedestal and top were wiped

with a dry tissue between each sample measurement. When finished, pedestal and top

were wiped with ultrapure water and dried with a tissue. The instrument was switched

off and log user was completed after each use.

2.6.1 RNA concentration with RNA Clean & Concentrator™-100

10 ng/µl total RNA in 50 µl volume was added to 100 µl RNA Binding buffer solution.

150 µl of ethanol was added and then the solution was mixed and applied to a Zymo-

Spin™ IC Column, as per manufacturer’s protocol. RNA was eluted in 15 µl

DNase/RNase-free water. RNA concentration was measured as described above and

stored at -20°C.

2.6.2 RNA concentration with SpeedVac® concentrators

10 ng/µl total RNA in 50 µl volume in 200 µl capacity eppendorf was applied to a Savant

SpeedVac DNA 110 Concentrator for 8 hours at 40°C, 1600 rpm and a vacuum of 5

Kilopascal (37.5 torr). Concentrated volume was removed to a clean eppendorf. RNA

concentration was measured as described above and stored at -20°C.

2.7 Tissue collection, preparation and storage of fresh frozen tissue

Postgraduate student, Mr. Aslam, was trained by consultant histopathologist, Dr. Kevin

West, and a pathology technician for sample collection, specimen dissection, tissue

selection and taking biopsies of freshly frozen cancer and adjacent normal tissues. On

the day of specimen collection, theatre staff, clinicians and laboratory technicians were

liaised with to coordinate tissue sampling.

Clinical details matching the consent form were confirmed with documentation in the

theatre log. Specimen pots were labelled with a participant identification number and

65

'H' Reference Number, allocated from the specimen receipt book. Resected cancer

specimens were dissected within 15 min of extraction from the participant. Each

specimen was opened through its entire length, avoiding cutting through the tumour.

In case of circumferential tumour, specimens were opened proximally and distally

sufficiently to see the corresponding edges of tumour. Two to three tissue biopsies of

2-3 mm in size were taken from the edge and centre of tumour tissue using a scalpel

and small forceps. Two to three full thickness colonic tissue biopsies were taken at least

5 cm away from the edge of tumour tissue. Biopsies were placed in labelled 15 ml BD

Falcon tubes, stored on ice, transported to the laboratory and processed in the

extraction hood.

Approximately 25 ml of isopentane was incubated in a small beaker by standing in liquid

nitrogen until isopentane became syrup-like with a layer of solid at the bottom.

Cancerous and normal tissue biopsy samples were placed and orientated onto a piece

of cork no larger than 0.4 × 0.4 × 0.3 cm. Embedding fluid was used to secure tissue

onto the cork pieces. Biopsy tissue on cork was placed in the isopentane syrup for 30

sec to allow snap-freezing. Each piece of tissue/cork was removed from the isopentane

using forceps and inserted into cryotubes. Lids were secured on cryotubes before

placing into liquid nitrogen. Cryotubes with tissue were transported and stored in an

allocated liquid nitrogen tissue storage container. Specimen logbooks and tissue

storage logs were updated with participant identification numbers, study numbers,

serial numbers and tissue types. All surfaces and instruments were cleaned with

detergent solution. Waste tissues and gloves were disposed in the orange bin.

2.8 Extraction of total RNA from snap-frozen tissue

Cryotubes containing freshly frozen cancerous and adjacent normal tissue were taken

out from liquid nitrogen storage container. Tissue for further processing was selected

and leftover tissue was returned to storage in liquid nitrogen. Freshly frozen tissue was

cut into 5 μm sections. Five sections were used for RNA extraction and one section was

used for haematoxylin and eosin (H&E) staining.

Frozen tissue sections on slides were stained by fixing sections in 95% industrial

methylated spirits (IMS) for 30 sec, rinsing in tap water for 30 sec, staining in

66

haematoxylin for 5 min, rinsing in running tap water for 5 min, staining in eosin for 20

sec and then rinsing in running tap water. Slides were dehydrated, cleared, dried and

mounted with DPX mountant and a glass coverslip. Five sections of each cancerous and

healthy adjacent tissue specimen were immersed in 1 ml of T9424 TRI Reagent®

Solution (Sigma-Aldrich, USA) in a 1.5 ml eppendorf. The resulting lysate was incubated

for 1 min, vortexed and stored at -20°C.

1 ml TRI Reagent/lysate was used for RNA extraction and the remainder was stored in

1 ml aliquots at -20°C. TRI-Reagent/lysate was incubated at room temperature for 5 min

and 200 ml chloroform was added and shaken vigorously to mix. The mixture was

incubated for 2-3 min at room temperature and centrifuged at 4000 rpm for 15 min at

4°C. The mixture was separated into an upper aqueous phase containing RNA, an

interphase, and a lower phenol red-chloroform phase containing DNA.

The upper aqueous phase containing RNA was transferred to a clean tube, without

disturbing the interface. 500 μl TRI Reagent was added to the aqueous phase. 100 μl

chloroform was added to mixture, vortexed, incubated at room temperature for 3 min

and centrifuged at 4000 rpm for 15 min at 4⁰C. The supernatant aqueous phase was

transfer to a clean 15 ml Falcon tube. The volume of aqueous phase was measured and

1.25 times the volume of 100% ethanol (Hayman Speciality Products, Essex, England)

was added and vortexed. RNA was extracted as described in the method above.

2.9 Formalin-fixed paraffin-embedded tissue sample collection

Patients who underwent curative resection for Dukes’ B CRC and subsequently

developed metastatic disease at some stage during the following five years (‘high risk

Bs’) were identified from the University Hospitals of Leicester NHS Trust CRC database,

a prospectively collected database of outcomes of patients treated for CRC in University

Hospitals of Leicester over the last 11 years. FFPE cancerous and adjacent normal tissue

specimen blocks were obtained from Department of Histopathology at University

Hospitals of Leicester NHS Trust. Histological diagnosis and Dukes’ stage was

reconfirmed by a consultant histopathologist. Archived FFPE cancerous and adjacent

normal tissue for case-matched controls (without metastases at five year follow up),

67

Dukes’ A, Dukes’ B ‘low risk B’ and Dukes’ C were obtained from the Department of

Pathology.

All patients with Dukes’ C CRC received postoperative chemotherapy and no patient

with ‘low risk B’ tumour had post-operative chemotherapy. In total, matched pairs of

surgically removed cancer and adjacent normal tissues for 100 patients were obtained.

Ethical permission was obtained from the Local Research Ethics Committee

(05/Q2502/28- ‘Markers of tumour progression in colorectal cancer’). Table 8 shows

the basic demographics and clinicopathological variables for all specimens. The table

summarises patient demographics and clinicopathological variables including tumour

type, stage, location, background, serosal involvement and extramural vascular

invasion.

68

Characteristics

Number Characteristics Number

Gender Male 60 Age Mean 71.6

Female 40 Range 37-96

Dukes' stage A 13 Tumour stage T1 5

B 50 T2 7

C1 31 T3 54

C2 6 T4 34

Tumour

location

Left 52 Extramural

vascular invasion

Present 57

Right 48 Absent 40

Tumour

background

Polyp 22 Serosal

involvement

Present 20

Diverticulitis 13 Absent 77

Ulcerated 3 Tumour type Mucinous 18

None 56 Adenocarcinoma 79

Table 8: Patient demographics and clinicopathological variables.

Patient demographics and clinicopathological variables including tumour type, stage,

location, background, serosal involvement and extramural vascular invasion.

69

2.9.1 RNA Extraction from FFPE Tissue

FFPE tissue blocks were cut into 5 µm sections with a microtome. H&E stained slides

were prepared for each tissue specimen. Marked out H&E stained sections for

corresponding tissue were used as a guide to manually micro-dissect the sections to

obtain mucosa or cancerous cells from unstained, de-waxed and rehydrated FFPE

sections for each specimen. Micro-dissected tissue was dissolved in 500 µl 0.05 Tris

buffer (pH 7.65) (TrisBase, Fisher Scientific Ltd, UK) and digested with 5 µl of proteinase-

K (10 mg/ml) (Roche Ltd, Hertfordshire, UK) by incubating overnight at 56°C. TRI-

Reagent/Chloroform and RNeasy® mini kit (Qiagen, UK) were used as per the

manufacturer’s protocol to extract total RNA. The concentration and integrity of RNA

samples were measured with the Agilent 2100 Bioanalyzer (Agilent Technologies, Santa

Clara, CA). RNA samples were stored at -20°C.

2.9.2 DNA Extraction from FFPE Tissue

Tissue from micro-dissected sections (as described above) were dissolved in 300 µl 0.05

Tris buffer pH 8 / 0.1% SDS (TrisBase, Fisher Scientific Ltd, UK) and digested with 25 µl

of proteinase K (10 mg/ml) (Roche Ltd, Hertfordshire, UK) for 72 h at 56°C. After two

repeat steps of phase isolation with 400 µl of phenol/chloroform/IAA, DNA was

precipitated with 1 M NaCl and ice cold 100% ethanol by incubating overnight at -20°C.

Precipitated DNA was washed with 500 µl 70% ethanol and sodium chloride and

resuspended in 50 µl of 1X Tris EDTA Buffer (EDTA Disodium salt AnalaR NORMAPUR,

VWR®, Leicester, UK). DNA concentration was measured with the Agilent 2100

Bioanalyzer and DNA samples were stored at -20°C.

2.9.3 Mutation analysis of FFPE tissue

Mutational analysis was performed using allele-specific binding PCR (ASB-PCR) to

identify the mutational status of PIK3CA, BRAF and K-ras oncogenes. Oligonucleotide

primers and probes were designed individually and ordered from Sigma-Genosys, UK.

All primer and probe sequences were generated using a combination of the Primer_3

software (Applied Biosystems) followed by sequence confirmation using NCBI

70

Nucleotide Basic Local Alignment Search Tool (BLAST). The mutations detected by

primer and probe sequences are summarized in Table 9.

Tumour DNA (10 ng), Taqman® Genotyping Mastermix (5 µl) (Applied Biosystems, UK)

mutant and wild-type probe (0.2 µl) (Applied Biosystems, UK) and 0.6 µl corresponding

forward and reverse primers (Sigma®, UK) for each gene of interest were used for each

10 µl PCR.

The reporter for the mutant probes was FAM-MGB, and for wild type was VIC-MGB.

PCR reactions were conducted using 7500 Fast Real-Time System (Applied Biosystems,

UK). DNA mutation analysis were performed under supervision by an M.Sc student Mr.

J Venkatesh. Probes, positive and negative control cell lines and annealing temperatures

used for mutation analysis are summarized in Table 10. The standard PCR profile was

carried out under the cycling conditions described in Table 11.

71

Table 9: Primers and probes used for mutation analysis.

Primers and probes sequences for wild type and mutant gene sequences of BRAF

(V600E), PIK3CA (A3140G, G1633A) and K-ras (KRAS122, KRAS123, KRAS133) genes.

Gene Mutation Primer type Primer sequence

BRAF V600E Forward primer 5'-TCATGAAGACCTCACAGTAAAAATAGGT-3'

Reverse primer 5'-ATCCAGACAACTGTTCAAACTGATG-3'

Wild type probe 5'-TAGCTACAGTGAAATC-3'

Mutant probe 5'-TAGCTACAGAGAAATC-3'

PIK3CA A3140G Forward primer 5'CAAGAGGCTTTGGAGTATTTCATG-3'

Reverse primer 5'-TGTTTAATTGTGTGGAAGATCCAATC-3'

Wild type probe 5'-ATGCACGTCATGGTG-3'

Mutant probe 5'-ATGCACATCATGGTGG-3'

G1633A Forward primer 5'-GCAATTTCTACACGAGATCCTCTCT-3'

Reverse primer 5'-CATTTTAGCACTTACCTGTGACTCCAT-3'

Wild type probe 5'-TGAAATCACTAAGCAGGA-3'

Mutant probe 5'-ATCACTGAGCAGGAGAA-3'

K-ras Forward primer 5'-AGGCCTGCTGAAAATGACTGA-3'

Reverse primer 5'-TGTATCGTCAAGGCAAGGCACTCTTGC-3'

Wild type probe 5'-CTACGCCACCAGCTC-3'

KRAS122 Mutant probe 5'-TACGCCADAGCTC-3'

KRAS123 Mutant probe 5'-TACGCCADCAGCTC-3'

KRAS133 Mutant probe 5'-CTACGTCACCAGCTC-3'

72

Table 10: Mutation analysis by RT-PCR.

Summary of mutant regions, probe annealing temperature, positive and wild type cell

lines controls for mutation analysis of BRAF, PIK3CA and K-ras genes.

Table 11: Thermal profiles for PCRs for mutation analysis.

Tabulated thermal profiles for different reactions including reverse transcription pre-

amplification, miRNA expression profiling, RT-PCR for validation cohort and RT-PCR for

mutational analysis.

Gene Mutation

region

Probe annealing

temperature (°C)

Positive control Wild type control

PIK3CA G1633A 60 MCF7 HCT116

A3140G 62 HCT116 MCF7

BRAF V600E 60 Skmel28 Tonsil

K-ras KRAS122 63 H358 H460

KRAS123 61 SW480 Tonsil

KRAS133 64 HCT116 Tonsil

Reaction Temperature

(°C)

Time Number of

cycles

Mutational analysis 50 2 min 1

95 10 min 1

95 15 sec

40 Probe annealing

temperature

(see Table 4)

1 min

73

2.10 miRNA expression profiling

miRNA expression analyses were performed using QRT-PCR TaqMan® miRNA Megaplex

QRT-PCR protocol for both miRNA expression profiling (700 miRNAs) and to validate

selected panel of miRNAs for both plasma and tissue analysis.

2.10.1 Chemistry overview for miRNA expression profiling

Relative quantitation using Megaplex™ Pools was accomplished using reverse

transcription, pre-amplification and real-time PCR steps.

2.10.1.1 Reverse transcription

In this step, cDNA was reverse transcribed from total RNA samples. The reverse

transcription used specific miRNA primers and reagents from the Megaplex Primers and

the TaqMan® MicroRNA Reverse Transcription Kit (Figure 10).

2.10.1.2 Pre-amplification

In the pre-amplification step, the reverse transcription product was uniformly

amplified from cDNA templates using the Megaplex PreAmp Primers and the

TaqMan® PreAmp Master Mix (Figure 11).

74

Figure 10: Reverse transcription of miRNA.

Illustration shows miRNA strand with RT loop primers leading to synthesis of first cDNA

strand with megaplex primers

Figure 11: Preamplification with miRNA specific forward and reverse primers.

Figure shows that miRNA-specific cDNA template was duplicated with forward primer

in first cycle and each cDNA strand was doubled with each further reaction.

75

2.10.1.3 Real-time PCR

During RT-PCR each TaqMan MGB probe annealed specifically to its complementary

sequence between the forward and reverse primer sites. The TaqMan MGB probes

contained a reporter dye (FAM™ dye) linked to the 5′ end of the probe, a minor groove

binder (MGB) at the 3´ end of the probe, which allowed the design of shorter probes

with greater specificity, and a non-fluorescent quencher (NFQ) at the 3′ end of the

probe. When the oligonucleotide probe was in intact, the proximity of the quencher

dye to the reporter dye quenched the reporter dye signal. AmpliTaq Gold® DNA

Polymerase extended the primers bound to the cDNA template. AmpliTaq Gold®

enzyme (containing 5´ nuclease activity) cleaved the probes that were hybridized to the

target sequence. When the hybridized probes were cleaved by AmpliTaq Gold® enzyme,

the quencher was separated from the reporter dye, increasing the fluorescence of the

reporter dye. Therefore, the fluorescence signal generated by PCR amplification

indicated the gene expression level (Figure 12).

76

Figure 12: RT-PCR for miRNA expression study.

Biochemistry overview of RT-PCR reactions using TaqMan MGB probe annealed

specifically to its complementary sequence between the forward and reverse primer

sites.

77

2.10.1.4 miRNA Expression Profiling experiment with TaqMan® miRNA Arrays

miRNA expression profiles for RNA extracted from plasma, freshly frozen tissues and

FFPE tissues were developed according to Applied Biosystems® user protocol by using

the following products:

Megaplex™ RT Primers (Human pool A v2.1 and Pool B v2.0)

Megaplex™ PreAmp Primers (Human pool A v2.1 and Pool B v2.0)

TaqMan® MicroRNA Array (Human card A v2.1 & Card B v2.0)

7900HT Fast Real Time PCR system (Applied Biosystems)

The manufacturer had designed Megaplex™ Pools (A & B) to detect and quantitate up

to 380 human miRNAs per pool using TaqMan® MicroRNA Array Card A v2.1 & Card B

v2.0 on 7900HT Fast Real Time PCR system. Megaplex Primer Pools consist of

matching RT primers and PreAmp pools. Megaplex™ RT Primers were a set of two

predefined pools (Pool A and Pool B) of up to 380 stem-looped reverse transcription

(RT) primers per pool that enable the simultaneous synthesis of cDNA for mature

miRNAs. Megaplex™ PreAmp Primers were a set of two pools (Pool A and Pool B) of

gene-specific forward and reverse primers intended for use with very small quantities

of starting material. The primers enabled the unbiased pre-amplification of the miRNA

cDNA target by PCR prior to loading the TaqMan® MicroRNA Array. TaqMan® MicroRNA

Arrays were a set of two 384-well microfluidic cards (Array A and Array B) containing

dried TaqMan primers and probes. The array enabled quantitation of gene expression

levels of up to 380 miRNAs and controls. This was accomplished by loading the cDNA

product (with or without pre-amplification) onto the array for PCR amplification and

real time analysis. For each sample (and optional no template control), one Megaplex

RT reaction was run, followed by one pre-amplification reaction per array. This provided

sufficient sample to be loaded in one TaqMan MicroRNA array. For a full miRNA profiles,

two Megaplex RT reactions (Pool A and B), two pre-amplification reactions (Pool A and

B), and two TaqMan MicroRNA Arrays (Array A and B) were run for each sample. Figure

13 shows the protocol for Megaplex QRT-PCR used for miRNA expression profiling,

Table 12 shows the concentrations of reagents used for each reaction and Table 13

shows thermal profiles for each reaction.

78

Figure 13: MicroRNA expression profiling protocol flow chart.

TaqMan® MicroRNA Arrays and protocol as per guide from manufacturer.

79

Table 12: Reagent concentrations for reverse transcription, pre-amplification and miRNA Taqman miRNA Array.

Reaction Reagents Concentration

(µl)

Reverse transcription Total Plasma RNA 3.00

20X Taqman® Megaplex RT Primers 0.80

100 mM dNTP 0.25

10X RT buffer 0.80

MultiScribeTM Reverse Transcriptase

(50 units/µl)

1.50

RNase Inhibitor (20 units/µl) 0.125

MgCl2 1.03

Total reaction volume 7.50

Pre-amplification cDNA for Human Pool A or B 5.00

Taqman® Megaplex™ PreAmp Primer

Pool

2.50

Taqman® PreAmp MasterMix 12.5

RNAase/ DNAase free water 5.00

Total reaction volume 25.00

PCR check prior to array

1:20 diluted Pre-amplified cDNA 2.00

TaqMan® Universal PCR Master Mix

No AmpErase® UNG

5.00

RNAase/ DNAase free water 2.50

Taqman MicroRNA PCR Probe 0.50

Total reaction volume 10.00

Taqman MicroRNA

Array

Pooled 1:4 diluted pre-amplified

cDNA

9.00

TaqMan® Universal PCR Master Mix

No AmpErase® UNG

450.00

RNAase/ DNAase free water 441.00

Total reaction volume 900.00

80

Table 13: Thermal profiles for different PCR reactions.

Tabulated thermal profiles for different reactions including reverse transcription pre-

amplification, miRNA expression profiling, RT-PCR for validation cohort and PCR for

mutational analysis.

Reaction Temperature (°C) Time Number of cycles

Reverse

Transcription

16 2 Minutes

40

42 1 Minute

50 1 second

85 5 Minutes 1

4 ∞ Storage

Pre-

Amplification

95 10 Minutes 1

55 2 Minutes 1

72 2 Minutes 1

95 15 Seconds

12 60 4 Minutes

99.9 10 Minutes 1

4 ∞ Storage

MicroRNA

expression

profiling

50 2 minutes 1

94.5 10 minutes 1

97 30seconds

40 59.7 1 Minute

qRT PCR 95 20 Seconds 1

95 1 Second

40 60 20 Seconds

81

2.10.2 Expression Profiling for plasma samples:

2.10.2.1 Participant groups and characteristics

Participants were grouped into CRC, benign adenoma and participants without

neoplasia, confirmed on colonoscopy. Patients were matched for age. Patients with CRC

were included to represent different sites and stages. Participants with benign polyps

included polyps with low grade dysplasia, moderate grade dysplasia and high grade

dysplasia. Patients who underwent colonoscopy for bowel symptoms and showed no

neoplastic disease at the time of colonoscopic examination were used as controls. Table

14 shows the characteristics of participants with neoplastic large bowel disease and

controls without any significant bowel disease seen on the colonoscopy. 3 µl RNA

extracted from plasma samples was for used in miRNA expression profiling by using

Taqman™ Array card. Table 14 compares the characteristics of participants in each

group of controls, adenomas and carcinomas. Appendix VI shows the characteristics for

participants used for plasma miRNA expression profiling.

Participant Characteristics Normal Adenoma Carcinoma

Number (n) 11 9 12

Gender M/F 6/5 5/4 7/5

Median Age (Years) 65 64 65.5

Dukes Stage N/A N/A A = (2)

B= (5)

C = (4)

D = (1)

Grade of Dysplasia N/A High= 2, Low =7

Location N/A Right =5, Left =4 Right = 5

Left = 7

Pre-operative

Chemotherapy/

Radiotherapy

N/A N/A 1

Previous Polyp / Carcinoma

Excision

3 1 1

Significant Background

Benign Disease

3 0 0

Table 14: Characteristics of participants for plasma miRNA expression profiling.

82

miRNA expression profiling was performed in two batches. The first run was performed

for 16 samples and second run was performed for another set of 16 samples following

the same protocol. Here is a description of the first run of 16 samples used for miRNA

expression profiling.

2.10.2.2 Reverse transcription reaction for cDNA of plasma miRNAs

TaqMan® MicroRNA Reverse Transcription Kit and the Megaplex™ RT Primers (mixture

volume of 4.5 µl) were used to synthesize single-stranded cDNA from plasma total RNA

samples (3 µl). The quantitative reverse transcription (QRT) reaction had a final volume

of 7.5 μl. Megaplex RT reactions were prepared and run on separate two occasions for

pool A and pool B.

Megaplex RT Primers, TaqMan® MicroRNA Reverse Transcription Kit component, MgCl2

and RNA samples were thawed on ice. A master mix was prepared for 18 reactions (to

account for pipetting loss) by combining the reagents (listed in Table 15) in a 1.5 ml

micro-centrifuge tube. A negative control (-RT) master mix solution was prepared for

three reactions by replacing the MultiScribe™ Reverse Transcriptase (50 U/μl) with

RNAase-free water (Table 16).

83

+RT reaction mix

components

Volume for

one sample

(μl)

Volume for 16 samples

(μl) (prepared for 18

reactions)

Megaplex™ RT Primers (10X) 0.80 14.4

dNTPs with dTTP (100 mM) 0.20 3.6

MultiScribe™ Reverse

Transcriptase (50 U/μl)

1.50 27

10X RT Buffer 0.80 14.4

MgCl2 (25 mM) 0.90 16.2

RNase Inhibitor (20 U/μl) 0.10 1.80

Nuclease-free water 0.20 3.60

Total 4.50 81

Table 15: Reagents for RT+ and RT- master mix.

-RT reaction mix

components

Volume for

one sample

(μl)

Volume for two samples

(μl) (prepared for 3

reactions)

Megaplex™ RT Primers (10X) 0.80 2.4

dNTPs with dTTP (100 mM) 0.20 0.60

10X RT Buffer 0.80 2.4

MgCl2 (25 mM) 0.90 2.7

RNase Inhibitor (20 U/μl) 0.10 0.30

Nuclease-free water 1.70 5.10

Total 4.50 13.5

Table 16: -RT reaction Master Mix for reverse transcription reaction.

84

Micro-centrifuge tube was inverted six times to mix, and then tubes were centrifuged

briefly. All MicroAmp® 8-Tube Strips were labelled with sample identification number.

4.5 μl RT reaction master mix (16 RT+ Reactions, 2 RT- Reactions) and one no template

control (NTC) master mix was pipetted into each of the labelled MicroAmp® 8-Tube

Strips tubes. 3 μl total plasma RNA was added into each corresponding labelled tube

containing RT reaction mix. 3 μl RNAase-free water was added to a tube labelled NTC.

MicroAmp® 8-Tube Strips tubes were sealed with MicroAmp® Optical Strip Caps. Tubes

were inverted six times to mix the solutions and were spun briefly. Tubes were

incubated on ice for 5 min. RT reaction was set up at Veriti® 96-Well Thermal Cycler

(Applied Biosystems, USA) for reaction volume of 7.5 μl and thermal-cycling conditions

as shown in the Table 1313. MicroAmp® 8-Tube strips tubes were loaded in the thermal

cycler and reaction was run. cDNA was stored at -20°C.

2.10.2.3 Megaplex pre-amplification reactions for plasma miRNAs

This step was used to pre-amplify specific cDNA targets to increase the quantity of

desired cDNA for microRNA expression analysis using TaqMan® MicroRNA Arrays. The

pre-amplification reaction had a final volume of 25 μl and contained 2.5 μl RT product

and 22.5 μl PreAmp reaction mix. The following steps were used in a run of Megaplex™

preamplification reactions. Megaplex™ PreAmp Primers were thawed on ice and mixed

by inverting six times and spinning briefly. TaqMan® PreAmp Master Mix (2X) was mixed

by swirling the Master Mix vial. A pre-amplification reaction master mix for 20 reactions

(to account for pipetting loss) was prepared by combining the reagents in a 1.5 ml

eppendorf (Table 17).

85

PreAmp reaction mix

components

Volume for

one Sample

(μl)

Volume for 19 Samples -

Prepare for 20 Reactions (μl)

TaqMan® PreAmp Master Mix

(2x)

12.5 250

Megaplex™ PreAmp Primers

(10x)

2.5 50

Nuclease-free water 7.5 150

Total 22.5 450

Table 17: Master mix for preamplification reaction.

2.5 μl of the pre-amplification master mix was pipetted into labelled MicroAmp® 8-Tube

Strips tubes (16 RT+ Reactions, 2 RT- Reactions) and one NTC. Strips tubes were mixed

and centrifuged briefly. In each of the labelled MicroAmp® 8-Tube Strips, 2.5 μL of

corresponding Megaplex RT reaction product was pipetted. MicroAmp® Optical Strip

Caps were used to seal the tubes and then invert six times to mix the solution and were

spun briefly. Tubes were incubated on ice for 5 min. Pre-amplification reaction was set

up on Veriti® 96-Well Thermal Cycler (Applied Biosystems, USA) for reaction volume of

25 μl, 9700Std ramp speed and the thermal-cycling conditions described in Table 7.

Prepared tubes were loaded and reaction was run. After removing the 8-tube strips

from the thermal cycler, tubes were centrifuged briefly. 75 μl 0.1X TE buffer pH 8.0 was

added to each tube and then sealed, inverted six times to mix and spun briefly. Diluted

pre-amplified product was stored at -20°C.

2.10.2.4 Assessment of pre-amplified cDNA with RT-PCR in 7500 Fast Real-Time

System

The pre-amplified cDNA for individual samples was assessed by RT-PCR in 7500 Fast

Real-Time System. This was to ensure that reverse transcription and pre-amplification

had been successful. 2.0 µl 1:20 diluted pre-amplified cDNA was used per reaction.

TaqMan® hsa-miR-21 assay (Catalogue no 4427975, Assay ID 002438) and SnRNA

86

RNU6B (Catalogue no 715680, Assay ID 001973) were used to assess the pre-amplified

product through RT-PCR reaction on StepOne 7500HT Fast Real Time PCR system

(Applied Biosystems).

2.10.2.5 TaqMan® miRNA Array and Real-Time PCR

During this step, DNA polymerase from the TaqMan® Universal PCR Master Mix and

TaqMan® MicroRNA Array amplify the target cDNA using sequence-specific primers and

probes. The following steps were used in a run of TaqMan® Array Real-Time PCR:

Stored and labelled PreAmp product was thawed on ice, mixed by inverting tube six

times and then centrifuged the tube briefly. TaqMan Universal PCR Master Mix was

mixed by swirling the vial and following were combined the in a 1.5 ml Eppendorf

(Table 18).

Component Volume for one array (μl)

TaqMan® Universal PCR Master Mix, No AmpErase®

UNG, 2X

450

Diluted PreAmp product 9

Nuclease-free water 441

Total 900

Table 18: Master mix for TaqMan® Array RT-PCR.

‡ Includes 12.5% excess for volume loss from pipetting.

87

The eppendorf was inverted six times to mix and then centrifuged briefly. The miRNA

Array card was loaded and run as per guide from the manufacturer (Applied Biosystems

TaqMan® Array User Bulletin -PN 4371129).100 μl PCR mix was dispensed into the

TaqMan MicroRNA Array card. Array card was sealed, centrifuged twice at 2000 rpm for

1 min at 4C° and then loaded on the 7900HT Fast Real Time PCR system (Applied

Biosystems) according to user bulletin –PN 4371129. Array was loaded and run using

the 384 well TaqMan Low Density Array default thermal-cycling conditions (Table 13).

Figure Figure 1414 shows the final step in miRNA Array reaction.

Figure 14: miRNA array for expression profiling.

Figure shows miRNA Array card with Megaplex Primers, 7900HT Fast Real Time PCR

system and amplification curves for miRNAs.

88

2.10.2.6 Data extraction from 7900HT Fast Real-Time PCR System:

miRNA Array data from RT-PCR on 7900HT Fast Real-Time PCR System was analysed for

Comparative CT (RQ) according to manufacturer protocol (Applied Biosystems 7900HT

Fast Real-Time PCR System Relative Quantitation Using Comparative CT Getting Started

Guide, PN 4364016). Raw data from 7900HT Fast Real Time PCR system was exported

into Microsoft Office Excel. Each array data was named and saved as a separate Excel

sheet.

2.10.3 miRNA Expression Profiling for High Risk Dukes’ B

20 cases of age and gender matched paired tumour and adjacent normal tissues, five

from each of Dukes’ Stage ‘A’, ‘low risk B’, ‘high risk B’ and ‘C’ were used for the

MicroRNA expression profiling as a training cohort . 100 ng RNA was reverse transcribed

to cDNA using Taqman® Megaplex™ RT Primers Human Pool A v2.1 & Pool B v2.0

(Applied Biosystems), according to manufacturers’ protocol. 5 µl cDNA for each sample

was pre-amplified using Taqman® Megaplex™ PreAmp Primers Human ‘Pool A v2.1’ and

‘Pool B v2.0’. Pre-amplified cDNA (25 µl) was diluted to 100 µl by adding 75 µl of 0.1X

Tris EDTA Buffer. After dilution, 20 µl of pre-amplified cDNA from each of the 5 samples

in a group was pooled to 100 µl. This volume was used for miRNA expression profiling

by using TaqMan® miRNA array.

2.10.4 Tissue miRNA expression profiling with TaqMan® miRNA Arrays

Freshly frozen cancerous and adjacent normal healthy colonic tissue from five

participants was used to develop tissue expression signature (Table 19). 100 ng/3 µl

total RNA extracted from freshly frozen tissue was used to run Megaplex reverse

transcription, Megaplex pre-amplification, TaqMan® MicroRNA Array Real Time PCR

reactions according to according to manufacturer’s - user protocol, as described above.

Pre-amplified cDNA (25 µl) was diluted to 100 µl by adding 75 µl of 0.1X Tris EDTA Buffer.

After dilution, 20 µl of pre-amplified cDNA from each of the five cancer samples and

five normal tissue samples were pooled to 100 µl. This volume was used for miRNA

expression profiling by using TaqMan® miRNA array.

89

Table 19: Characteristics of cancer patients used for expression analysis by using freshly frozen tissue.

2.11 QRT-PCR for validation cohorts

QRT-PCR was used to analyse a panel of miRNAs in the validation cohorts and for

exosomal miRNA analysis. For validation of miRNAs in plasma, total RNA was reverse

transcribed and pre-amplified using Megaplex™ RT and Pre-amplification Primers Pool

A & Pool B on Veriti® thermal cycler (Applied Biosystems, UK). Megaplex™ RT and Pre-

amplification Primers Pool A was used for the validation of miRNAs specific to high risk

Dukes’ B cancer tissue and final validation of a single plasma miRNA for the detection

of adenoma and carcinoma. 4.5 µl 1:20 diluted Pre-amplified cDNA was used in 10 µl

reaction volume to run real time PCR on 7500 fast Real-Time system (Applied

Biosystems). For the validation of a panel of miRNAs selected from plasma miRNA

expression profiling, a cohort of 94 participants (Table 20) was used. (Appendix VIII & IX

shows the subcategories of normal and detailed characteristics of participants.)

H Number Gender Age Diagnosis/Stage

H45/09 M 67

Sigmoid colon adenocarcinoma,

Duke's A pT2,pN0,rM0

H56/09 M 75

Sigmoid colon adenocarcinoma,

Duke's A pT2,pN0,rM0

H164/09 M 70

Ascending colon adenocarcinoma,

Dukes B, pT3 pN0, rM0

H175/09 M 58

Sigmoid adnenocarcinoma,

Dukes' C2, pT4, pN2, rM0,

H176/09 F 48

Ascending colon adenocarcinoma,

Dukes B, pT3 pN0, ?rM1

90

Table 20: Characteristics of participants used for first validation of plasma miRNA panel.

Controls Adenoma Carcinoma

Number (n) 32 28 34

Gender M:F 17:15 16:12 26:08

Median Age (Years) 62 68 67

Dukes Stage

N/A

N/A

A = (06)

B = (13)

C = (12)

D = (03)

Grade of Dysplasia N/A High = 5

Low = 21

N/A

Location N/A Right = 10

Left = 17

Right = 12

Left = 22

Pre-operative Chemotherapy/

Radiotherapy

N/A N/A 3

Previous Polyp / Carcinoma

Excision

8 3 2

Significant Background Benign

Disease

9 0 0

91

2.12 Isolation of exosomes from harvested cell line culture media by

ultracentrifugation

Extraction hood and incubator were cleaned prior to cell line culture. Incubator settings

were set up for the standard cell line culture requirements of temperature and CO2.

Settings were maintained and checked every 24 hours. 450 ml McCoy's 5A culture

media was mixed with 45 ml 10% foetal bovine serum (FBS) and 5 ml ampicillin in 50 ml

centrifuge tubes. McCoy's 5A media was warmed to 37°C for 30 min in a water bath.

HT29 cells stored in liquid nitrogen were defrosted in water bath at 37°C.

Defrosted cells were added to 10 ml McCoy's 5A media and centrifuged at 1000 rpm for

5 min. Supernatant was discarded and cell pellet was resuspended in 10 ml media. 10

ml media was dispensed into three 25 cc cell culture flasks and 3 ml of HT29 cells were

added to each flask. Flasks were incubated at 37°C with 95% air, 5% CO2. Cultures were

washed with PBS and media was replaced after 48 hours of culture. After 72 hours, 75-

80% confluence was achieved. Flasks were washed again and media was replaced with

serum-free media.

Media was harvested from flasks after 96 hours and centrifuged at 1000 x g and 10,000

x g to remove cells and cell debris, respectively. 10 ml harvested media was filtered

through 300 μm and 100 μm filters. 5 ml harvested media was stored for

immunoprecipitation. 5 ml filtered harvested media was centrifuged at 140,000 x g for

90 min at 4°C to pellet exosomes. 4.5 ml supernatant was removed from centrifuged

tube and discarded. A replacement volume of 4.5 ml PBS was added to the tube prior

and then centrifuged again at 140,000 x g for 90 min at 4°C to pellet the exosomes. 4.9

μl supernatant was discarded and 100 μl of leftover supernatant was stored at -80°C,

and is herein referred to as an exosome pellet.

2.12.1 Isolation of exosomes from plasma samples by ultracentrifugation

1 ml plasma stored in 1.5 ml eppendorf at -80°C was defrosted and centrifuged at

10,000 x g at 4°C for 10 min, followed by 30,000 x g at 4°C for 30 min. Supernatant was

transferred to another 1.5 ml eppendorf and filtered through 300 μm and 100 μm

filters. Filtrate was prepared for ultracentrifugation by adding PBS to bring the final

92

volume to 5 ml. Filtered and diluted plasma was centrifuged at 140,000 x g for 90 min

at 4°C to pellet the exosomes. 4.5 ml supernatant was removed from tube and

discarded. A replacement volume of 4.5 ml PBS was added to the tube and then

centrifuged again at 140,000 x g for 90 min at 4°C to pellet the exosomes. 4.9 μl of

supernatant was discarded and 100 μl of exosome pellet was stored at -80°C. Exosome

pellets were separated and stored at -80°C for further analysis. The presence of

exosomes was confirmed by flow cytometry, dynamic light scattering and electron

microscopy.

2.12.2 Transmission electron microscopy

Transmission electron microscopy was performed at the Department of Imaging and

Electron Microscopy at the University of Leicester. 100 μl exosome pellets isolated from

plasma and cell line culture harvested media were thawed to room temperature. 10 μl

exosomes were added to 490 μl PBS. 500 μl PBS was used as a control solution. 10 µl

exosome solution in PBS and control PBS solution was placed on Parafilm and a formvar

carbon-coated nickel grid was mounted on each drop for 30 to 60 min.

The grid was positioned with the coated side facing the drop containing exosomes or

control solution. Three drops of PBS, each 30 μl, were placed on the Parafilm to wash

the grid by sequentially positioning the grid on top of the droplets of PBS with absorbing

paper in between. Two samples were fixed with 2% paraformaldehyde on the Parafilm

and grids were placed on top of the drop for 10 min. Samples were washed with 10 µl

PBS and fixed with 2.5% glutaraldehyde in PBS for 30 min on ice. After rinsing, the pellet

was sequentially stained with osmium tetroxide and uranyl acetate, and then

dehydrated and embedded in Polybed 812. Tissue was sectioned on a Reichert-Jung

Ultracut, viewed on a Zeiss 902 electron microscope and recorded with Kodak E.M. film.

All electron microscopy reagents were purchased from Electron Microscopy Sciences.

2.12.3 Dynamic Light Scattering

Dynamic light scattering was performed at the Department of Physics at the University

of Leicester. The sizes of particles in exosomes eluted in PBS and exosome-free PBS

(control) were measured. PBS was filtered through a 100 µm filter. 10 µl plasma and

93

harvested media exosomes were eluted in 490 µl filtered PBS. Particle size distribution

was determined by a Nicomp 370 submicron particle sizer (Agilent Technologies Inc,

Particle Sizing Systems Division, Santa Barbara, CA, USA). Nicomp 370 instrument was

initiated in line with the departmental protocol. The samples were transferred to a

cuvette for size measurement using the following settings of dynamic light scattering

for two solutions: PBS viscosity 1.05cP, temperature 20°C and refraction index 1.33. All

measurements for buffer and samples were taken at room temperature. The Zetasizer

Nano software adjusted the channel width for each sample based on the fluctuation

rate of scattered light, and the final size distribution for each sample was calculated

using the number-weight Gaussian setting within the software.

2.12.4 Immunoprecipitation: antibody coupling

Invitrogen Antibody Coupling Kit with Dynabeads® M-270 Epoxy beads was used to

couple three different antibodies as per manufacturer protocol from Invitrogen. The

following antibodies were used for coupling to beads:

CD133 – PE conjugated (Miltenyi Biotec, MACS, UK) antibody (50 µg/ml)

CD326 (EpCAM) – APC conjugated (Miltenyi Biotec, MACS, UK) antibody (50 µg/ml)

Moisture on unused beads deactivates the reactive groups necessary for covalent

antibody coupling. To avoid condensation on unused beads, beads were stored at room

temperature prior to opening the bottle.

Magnet was disinfected to prevent accidental sample contamination. 5 mg Dynabeads®

M-270 Epoxy beads were washed with 1 ml of C1 solution by pipetting and vortexing in

an eppendorf. This eppendorf was labelled as ‘Antibody +ve’. In a second eppendorf

labelled as ‘Control’, 5 mg Dynabeads® M-270 Epoxy beads were washed with 1 ml C1

solution by pipetting and vortexing. Two eppendorfs were placed on the magnet for 1

min and beads were allowed to collect at the side of the tube. Supernatants from both

eppendorfs were removed. To the ‘Antibody +ve’ eppendorf, 100 µl antibody at 5

µg/100 µl, 150 µl C1 solution and 250 µl C2 solution were added and mixed by gentle

vortexing. To the ‘Control’ eppendorf, 250 µl C1 solution and 250 µl C2 solution were

added and mixed by gentle vortexing. Mixtures were incubated on a roller for 16 hours

94

at 4°C to ensure the fluid and beads were mixed well. After overnight incubation, each

eppendorf was placed on the magnet for 1 min and then supernatant was removed.

Beads were washed with 800 µl HB solution and 800 µl LB solution in two steps.

Beads in both ‘Antibody +ve’ and ‘control’ eppendorfs were washed twice with 800 µl

SB solution twice for 15 sec. Beads were washed with SB Wash for 1 min and incubated

on a roller at room temperature for 15 min. Eppendorfs were placed in the magnet for

1 min and then supernatant was removed. Control and antibody-coupled beads were

resuspended in 450 μl SB solution and 50 µl Aldefluor buffer solution. Resuspended

beads were stored at 4°C until required.

2.12.5 Flow cytometry to assess coupling of antibody with beads

Antibody-coupled Dynabeads suspended in 50 µl Aldefluor buffer and unstained non-

reactive Dynabeads (control) in Aldefluor buffer were used for flow cytometry

analysis:

a. CD133 – PE coupled Dynabeads

b. CD326 (EpCAM) – APC coupled Dynabeads

c. ‘Control’ unstained uncoupled Dynabeads

All FACS analysis and sorting was done using the BDFACS Aria II SORP. An 85 µm nozzle

was used for sorting or analysis of cells. The flow chamber, deflection plates and

sample loading area were all cleaned with 70% IMS prior to sorting. For triple colour

experiments, appropriate compensation was set up and applied to the sorting

experiment to avoid any crossover of fluorescence signals from respective

fluorochromes. The sorted beads were collected in a tube holder consisting of four

tubes filled with sample collection solution (PBS). The sorting gates were set up to

exclude any doublets. The minimum number of events recorded for any analysis was

10,000, provided there were enough cells for analysis.

95

2.12.6 Immunoprecipitation of plasma exosomes with antibody-coupled Dynabeads

3 ml plasma sample stored at -80°C (1 ml plasma per 1.5 ml eppendorf) was removed

from the freezer, defrosted at room temperature and centrifuged at 10,000 x g at 4°C

for 10 min, and 30,000 x g at 4°C for 30 min. The resultant supernatant was filtered

through a 100 µl filter. 1 ml filtrate was separated into two 1.5 ml eppendorfs. 100 µl

CD133 antibody-coupled Dynabeads and 100 µl CD326 (EpCAM) antibody-coupled

Dynabeads were added to 1 ml filtered plasma. 200 µl ‘Control’ unstained uncoupled

Dynabeads were added to 1 ml of filtered plasma. Plasma and Dynabeads solutions

were mixed by pipetting, and incubated at 4°C on a roller for 20 min. Each eppendorf

was placed on magnet to isolate Dynabeads. Supernatants were removed and stored at

-80°C. Precipitated Dynabeads were washed with 1 ml PBS and two washes with 500 µl

Aldefluor. Exosomes bound to Dynabeads were eluted in 100 µl of Aldefluor and stored

at -80°C.

2.12.7 Immunoprecipitation of cell line exosomes with antibody-coupled Dynabeads

5 ml harvested culture media from HT29 cultures was defrosted at room temperature.

2 ml media was transferred to two eppendorfs and centrifuged at 10,000 x g at 4°C for

10 min, and then 30,000 x g at 4°C for 30 min. Supernatants were filtered through 100

µl filters and then separated into two eppendorfs. 100 µl CD133 antibody-coupled

Dynabeads and 100 µl CD326 (EpCAM) antibody-coupled Dynabeads were added to 1

ml filtered media. 200 µl ‘Control’ unstained uncoupled Dynabeads were added to 1 ml

filtered media. Harvested media and Dynabeads solutions were mixed by pipetting and

incubated at 4°C on a roller for 20 min. Each eppendorf was placed on a magnet to

isolate Dynabeads. Supernatants were isolated and stored at -80°C. Precipitated

Dynabeads were washed twice with 1 ml PBS and exosomes bound to Dynabeads were

eluted in 100 µl PBS and stored at -80°C.

2.12.8 Extraction of total RNA from exosomes isolated from harvested media and

plasma

Exosomes previously isolated by ultracentrifugation were defrosted at room

temperature and 100 µl was transferred to an eppendorf. 10 μl 5 N acetic acid was

added to 100 µl exosomes and vortexed. 375 μl of Tri-reagent solution was added,

96

vortexed and incubated at room temperature for 5 min. Next, 100 µl chloroform was

added, solution was vortexed and incubated at room temperature for 3 min. The

solution was centrifuged at 4000 rpm for 15 min at 4⁰C and then the supernatant

aqueous layer was transferred to a clean 1.5 ml eppendorf, and the volume of the

aqueous layer was measured. 100% ethanol with 1.25 times the volume of aqueous

layer was added and vortexed. Total RNA extraction was completed as previously

described using a mirVana column. RNA was eluted in 50 μl of pre-heated RNAase-free

water. Total RNA concentration was measured and RNA was stored at -20⁰C.

2.12.9 Extraction of RNA from immunoprecipitated exosomes and supernatants

100 µl exosomes immunoprecipitated with antibody-coupled Dynabeads was mixed

with 10 μl 5 N acetic acid. 375 μl Tri-reagent solution was added, vortexed and

incubated at room temperature for 5 min. 100 µl of chloroform was then added,

solution was vortexed and incubated at room temperature for 3 min. Beads were

isolated by magnet and supernatant was removed and centrifuged at 4000 rpm for 15

min at 4⁰C. The aqueous layer was transferred to a clean 1.5 ml eppendorf and the

volume of the aqueous layer was measured. 100% ethanol with 1.25 times the volume

of aqueous layer was added and vortexed.

Total RNA was extracted using the previously described RNA isolation and miRvana

Total RNA extraction methods. RNA was eluted in 50 μl of pre-heated RNAase-free

water. Total RNA concentration was measured and RNA was stored at -20⁰C. 100 μl 5 N

acetic acid was added to 1 ml volume of supernatant plasma in 1.5 ml eppendorf. The

solution was vortexed and transferred into 15 ml conical centrifuge tubes (BD

Bioscience, Bedford, USA).

3750 μl of T9424 TRI Reagent® Soultion (Sigma-Aldrich, USA) was added to the Falcon

tube, vortexed and incubated at room temperature for 5 min. 1 ml chloroform was also

added, vortexed and incubated at room temperature for 3 min. Solution in the Falcon

tube was centrifuged at 4000 rpm for 15 min at 4⁰C. The aqueous phase was transfer

to a clean 15 ml BD Falcon tube. The volume of aqueous phase was measured and 1.25

97

times the volume of 100% ethanol (Hayman Speciality Products, Essex, England) was

added and vortexed. Total RNA was extracted and RNA was eluted and stored as

described above.

2.12.10 Immunoprecipitation of plasma exosomes with GPA33-coupled

antibody

GPA33 is a cell surface A33 antigen, also known as glycoprotein A33. GPA33 is expressed

in normal gastrointestinal epithelium and in 95% of colon cancers. The predicted

mature protein has a 213 amino acid extracellular region, a single transmembrane

domain and a 62 amino acid intracellular tail. Like CD326 (EpCAM), GPA33 can be a

potential cell surface protein marker for CRC-specific exosomes. GPA33 was evaluated

for its potential use in immunoprecipitation of plasma exosomes in patients with CRC.

GPA33 antibody coupling with Dynabeads was achieved using the antibody bead

coupling protocol described above with 5 µl of GPA33 antibody (1 µg/µl). Antibody-

coupled beads were stored in 500 μl SB solution at 4°C. The final concentration of

GPA33-coupled Dynabeads was 1 mg/100 µl.

2.12.11 miRNA expression analysis for exosomal RNA

Exosomal RNAs were quantified by Megaplex cDNA synthesis with reverse

transcription, pre-amplification and RT-PCR in line with methods described earlier.

Exosomal RNA from plasma and harvested media from HT29 cultures were assessed for

the presence of common miRNAs (miR-21, miR-192*, miR-369, miR-589) and SnRNA

RNU6B. Plasma samples from patients with CRC, adenoma and healthy controls were

evaluated for detection of miR-21, miR-135b and SnRNA RNU6B. Megaplex™ RT and

Pre-amplification Primers Pool A was used for the analysis of exosomal miRNAs and final

validation of a single plasma miRNA for the detection of adenoma and carcinoma.

2.12.12 Cell sorting for stem cell-related miRNAs.

The freshly collected and transported cancer and adjacent normal tissues were put into

5 ml media 199 and cut into pieces with sterile scissors until the suspension could easily

98

be pipetted through a 10 ml pipette. After mincing was complete, the volume of media

199 was made up to 10 ml and the required amount of collagenase type IV

(Worthington Chemicals) added to obtain a working concentration of 2000 U/ml.

The tissue suspension was incubated in a 37°C agitator for 60-90 min (the incubation

time needs to be optimised for each tissue type), pipetting cells every 15 min. Following

collagenase digestion, the cell suspension was filtered through a 100 µm filter to discard

any large lumps of undigested tissues. The cell suspension was then centrifuged at 350

x g for 5 min, supernatant removed and cells washed with HBSS for 5 min.

The cell suspension was filtered again through a 100 µm filter and washed twice with

HBSS. After the final wash, the cell suspension was filtered through a 40 µm filter. The

filtrate constituted a suspension of single cells. The following protocol is appropriate for

single or multiple staining using CD133 and ESA. The single cell suspension was washed

with 500 µl Aldeflour assay buffer and centrifuged at 1300 rpm (2017 x g) for 3 min. The

supernatant was discarded and the pellet resuspended in 100 µl Aldeflour assay buffer.

Following this, dual staining was carried out by addition of CD133 and CD326

fluorescence conjugated antibodies at a dilution of 1:10 in the 100 µl cell suspension.

The cells were incubated with the antibodies for 30 min at 4°C in dark. After incubation,

cells were washed and re-suspended in 500 µl of Aldeflour assay buffer and the

resulting cell suspension was analysed via flow cytometry as described above.

2.13 Statistical Analysis:

2.13.1 Analysis of expression profiling to identify target miRNAs

I. Raw data from 7900HT Fast Real Time PCR system was exported into Microsoft

Office Excel files. PCR based quantification values (CT) were plotted using

Applied Biosystems software. Only well-expressed miRNAs (CT <35) were

considered for further analysis. MicroRNA expression levels (CT) in array were

normalised to expression levels of endogenous controls (MammU6, snRNA

RNU6B, geNorm, global mean of expression profiling or most stable miRNA in

expression profiling) and relative expression levels (ΔCT) were calculated.

99

Normalisation calculations were carried out in Microsoft Office Excel (Microsoft

Software Corporation, Washington, USA). All fold changes were calculated as

power to base 2.

II. Comparative expression levels (ΔΔCT) for different groups were compared to

identify target miRNAs for:

a) Plasma samples of CRC, polyps and healthy controls with normal

colonoscopy

b) FFPE tissue miRNAs for high risk and low risk Dukes’ B CRC

c) Plasma miRNA expression profiles were compared to tissue

miRNA expression profiles to validate the source of plasma

miRNAs identified in the discovery panel

III. Principal component analysis (PCA), hierarchy cluster analysis (HCA),

Significance analysis for microarrays (SAM), Bioinformatics analysis and Z-Score

calculations were performed for MammU6 normalized data for individual

sample and for the average expression of each miRNA in different groups.

IV. Intergroup comparisons were also performed on profiling data using the Mann

Whitney-U-test, one-way analysis of variance (ANOVA) with Bonferroni

corrections and the Student’s t-test.

2.13.2 Validation of diagnostic plasma miRNAs

I. Receiver Operating Characteristic (ROC) analysis was applied to generate ROC

curves for expression of target miRNAs. Log rank analyses were performed to

compare the ROC curves.

II. Area under the curve (AUC) was used to evaluate the sensitivity and specificity

of studied target miRNAs for their use in detection of CRCs, adenomas and

controls.

III. Logistic regression analyses were performed on each panel of miRNAs to

calculate the probability value for disease status. These probability values were

further used to generate ROC curve for diagnostic panel.

100

2.13.3 Analysis of tissue miRNAs for high risk Dukes’ B cancer

I. For tissue miRNAs regression analysis were used to study for their association

with patient age, gender, tumour size, tumour location, differentiation grade or

dysplasia and Dukes’ stage, as demonstrated by preoperative imaging and the

histopathology diagnosis.

II. Comparative expression levels (ΔΔCT) for tumour to normal adjacent tissue were

calculated and intergroup comparisons were performed using the Mann-

Whitney-U-test, student’s-test, One way analysis of variance (ANOVA) with

Bonferroni test, and Wilcoxon signed-rank test. Fischer’s exact test was carried

out to identify significant frequency of mutations.

2.13.4 Software used for analysis

Statistical analysis were carried out with Microsoft Office Excel v.2010 & 2013

(Microsoft, Redmond, USA) Multi Experiment Viewer v4.4 (BioITeam, Texas, USA), SPSS

18.0 (IBM, Newyork, USA), Graphpad Pprism 5 (Graphpad Software Inc, San Diego, USA)

softwares. P<0.05 was considered significant. Expression graphs were created in

Graphpad Prism 5 (Graphpad Software Inc, San Diego, California).

101

Chapter 3: Results

Analysis of plasma miRNAs for the

detection of adenomas and carcinomas

102

3 Results

3.1 Summary of results

The current strategy for bowel cancer screening is based on the utility of FOBT.

Inaccuracy associated with FOBT has led to many misdiagnoses and unnecessary

invasive investigations. Recent studies have shown that tumour-derived miRNAs are

present in the blood at levels sufficient for their use as measurable biomarkers of

different tumours. The aim of this study was to identify which circulating miRNAs could

be used for the early detection of colorectal neoplasia. The objective was to develop

and compare miRNA expression profiles of RNA isolated from the plasma of participants

with colorectal adenomas and carcinomas with participants without any significant

colonic pathology on colonoscopic examination. The other objective was to apply

discriminatory miRNAs identified from expression profiling in a larger cohort to assess

their diagnostic accuracy for the detection of adenomas and carcinomas. miRNA

expression profiling was performed on 32 participants (controls=11, adenomas=9,

carcinomas=12) using Taqman® MicroRNA Array, Megaplex™ RT and pre-amplification

primers Human Pool A v.2.1 and Pool B v.2.0. MammU6 normalized profiling data was

analysed by Z-scores, principal component analysis, hierarchy cluster analysis,

bioinformatics and student’s t-test. A panel of 29 discriminatory miRNAs identified

from profiling data (miR-16, miR-23b, miR-34a, miR-92a, miR-95, miR-135b, miR-181c,

miR-181c*, miR-182, miR-182*, miR-191, miR-192*, miR-195, miR-200a*, miR-203,

miR-205, miR-369-5p, miR-410, miR-431, miR-486-3P, miR-486-5p miR-487b, miR-502-

5p, miR-532-5p, miR-564, miR-566, miR-589, miR-592 & miR-624*) was validated on an

initial cohort of 94 symptomatic participants (controls=32, adenoma=28,

carcinoma=34). Receiver operating characteristics (ROC) and logistic regression analysis

were performed to assess the diagnostic accuracy of individual miRNAs and different

groups/panels of miRNAs. A panel of 18 miRNAs (miR-135b, miR-34a, miR-431, miR-16,

miR-369-5p, miR-23b, miR-191, miR-21, miR-589, miR-487b, miR-95, miR-484, miR-195,

miR-181C*, miR-410, miR-532-5p, miR-192*, miR-203) was associated with an area

under the curve (AUC) value of 0.92 (95% CI: 0.85-1.00), with 92% sensitivity and 88%

specificity for the detection of adenomas and carcinomas. A panel of 3 miRNAs (miR-

135b, miR-431 and miR-34a) had AUC value of 0.92 (95% CI: 0.84 - 0.99) with 91%

103

sensitivity and 88% specificity. The individual values of AUCs for the detection of both

adenomas and carcinomas for miR-135b, miR-431 and miR-34a were 0.77, 0.78 and

0.79, respectively. Based on the highest diagnostic accuracy and ease of detection miR-

135b was selected and further validated on an independent final cohort of participants

(asymptomatic controls=25, adenomas =30, carcinoma=41). When validated in an

independent final cohort, miR-135b was associated with an AUC value of 0.82 (95% CI:

0.71-0.92), with 80% sensitivity and 84% specificity for the detection of adenomas and

carcinomas. This study has thus succeeded in identifying plasma miRNAs for the early

detection of CRC.

104

3.2 Concentrating RNA samples

Since RNA extraction using the miRvanaTM RNA isolation kit yielded low concentrations

of RNA, further investigations were conducted to increase the yield of total RNA from

samples.

3.2.1 RNA Clean & Concentrator™-100 and SpeedVac® concentrator

The concentration of total RNA extracted with miRvanaTM from freshly frozen CRC tissue

of 4 patients was 10 ng/µl. A comparison between the RNA Clean & Concentrator™-100

and the SpeedVac® concentrator was conducted to concentrate the final amount of

total RNA per sample. 50 µl total RNA was used for each concentration experiment. The

average concentration of RNA in four samples was increased to 25 ng/µl using the RNA

Clean & Concentrator™-100, and to 45 ng/µl with the SpeedVac® concentrator (Figure

1Figure 155).

The comparative analysis for SnRNA RNU6B expression showed 4-fold increase in

expression of RNA concentrated with the RNA Clean & Concentrator™-100, and 8-fold

increase with the SpeedVac® concentrator. miR-21 expression analysis showed a 2-fold

reduction in expression levels of RNA concentrated with RNA Clean & Concentrator™-

100 and a 16-fold reduction with the SpeedVac® concentrator.

105

Figure 15: Total RNA concentrations and expression levels (CT) of SnRNA RNU6B and miR-21.

3 µl RNA was used in Taqman® RT-PCR with MegaplexTM pool A RT and preamplification

primers. Taqman® SnRNA RNU6B and Taqman® miR-21 assay were used to detect

expression levels (CT). The matched pair Student’s t-test was used to compare the yield

of RNA amplified by two concentration kits with each other and the concentration of

RNA in the original samples. In the graph, higher CT values indicate lower expression. A

significant increase in SnRNA RNU6B expression was detected in samples concentrated

with the RNA Clean & Concentrator™-100 (p=0.041) and with the SpeedVac®

concentrator (p<0.001) in comparison to the original samples. miR-21 levels were

significantly reduced in samples concentrated by the Clean & Concentrator™-100

(p=0.001) and SpeedVac® concentrator (p<0.001).

106

3.2.2 Different elution volumes of total RNA extracted with miRvanaTM RNA

isolation kit

Total RNA concentrations and expression levels of SnRNA RNU6B and miR-21 were

eluted in different volumes following extraction with the miRvanaTM RNA isolation kit.

3 µl RNA was used for Taqman® RT-PCR, as described above. Increases in the

concentration of total RNA were seen with reductions of elution volumes, which ranged

from 50 µl to 12.5 µls (Figure 16)

Figure 16: Concentrations of RNA extracted with miRvanaTM RNA isolation kit.

RNA concentrations of samples extracted from 1 ml of plasma were 3.52, 8.28 and

10.34 ng/μl in elution volumes of 50, 25 and 12.5 µl, respectively. 13.17 ng/µl RNA in

25µl of elution volume was extracted from 2 ml plasma using one filter cartridge.

107

Expression levels of SnRNA RNU6B increased with a reduction in elution volume, but

miR-21 expression levels reduced with a reduction in elution volume. This indicated that

miRNAs were lost when a smaller volume of eluent was used to elute RNA from the

glass fibre filter cartridge. The concentration of both miRNAs increased by 2-fold when

extracted from 2 ml plasma in comparison to 1 ml plasma with a reduced elution volume

of 25 µl (Figure 17) However, this required large volumes of filtrate – as much as 7 ml –

to be applied to the filtrate cartridge. Therefore, for this study, RNA extracted from 1

ml plasma was eluted in 50 µl water.

Figure 17: CT values of SnRNA RNU6B and miR-21 in different elution volumes of RNA extracted with miRvanaTM RNA isolation kit.

3µl RNA was used in Taqman® RT-PCR with MegaplexTM pool A RT and preamplification

primers. Taqman® SnRNA RNU6B and the Taqman® miR-21 assay were used to detect

expression levels. Higher CT values indicate lower expression. An increase in expression

levels of SnRNA RNU6B was seen with a concomitant reduction in elution volume.

Conversely, decreased expression levels of miR-21 were seen with an increase in elution

volume. Unlike RNA extracted from 1 ml plasma, which was eluted in 12.5, 25 or 50 µl,

an increase in expression of both SnRNA RNU6B and miR-21 was detected in RNA

extracted from 2 ml plasma used when eluted in 25 µl.

108

3.3 Discovery phase: expression profiling of plasma miRNAs to identify

discriminatory miRNAs for the detection of adenomas and carcinomas

3.3.1 Total RNA concentrations for use in the miRNA array

1 ml plasma taken from 32 participants (n = 11 controls, n = 9 adenoma, n = 12

carcinoma) was used for RNA extraction for the Taqman® MicroRNA expression

profiling experiment. The median concentrations of RNA extracted from control,

adenoma and carcinoma samples were 6.35, 4.5 and 6.65 ng/µl, respectively. Analysis

of RNA by spectrophotometry reported median values of the 260/280 ratio of 1.46, 1.41

and 1.48 for control, adenoma and carcinoma groups, respectively (Table 21).

109

Group Sample number RNA concentration (ng/µl) 260/280 value

Control

H333/08 7.2 1.84

H341/08 8.2 1.65

H398/08 6.4 1.24

H400/08 6.3 1.43

H400A/08 4.7 1.61

H515/08 8.5 1.17

H520/08/A 9.2 1.21

H57/09 2.4 1.49

H60/09 2.1 1.28

H174/09 5.9 1.63

H271/09 26.8 1.53

H274/09 13 1.84

Adenoma H339/08 4.4 1.46

H510/08 4.5 2.02

H516/08 11.9 0.91

H518/08 4.8 1.18

H58/09 6 1.49

H169/09 5.4 1.4

H170/09 2.6 1.57

H177/09 3.6 1.41

H180A/09 2.1 1.02

Carcinoma

H333/08 7.2 1.21

H399/08 7.7 1

H513/08 11.3 1.18

H43/09 5.2 1.14

H45/09 4.7 2.23

H56/09 4.2 1.4

H166/09 6.1 1.57

H175/09 5.1 1.23

H176/09 2.9 2.29

H276/09 12.8 1.56

H281/09 24.1 1.66

H287/09 16 1.69

Table 21: Total RNA concentrations measured with Nanodrop ND-1000 Spectrophotometer.

Concentration and 260/280 ratios of RNA extracted from individual samples of control,

adenoma and carcinoma groups. Sample numbers are indicated by H/number; RNA

concentrations are expressed as ng/µl.

110

3.3.2 Expression profiling array for plasma miRNAs

miRNA expression profiling data was analysed using different statistical tests to identify

discriminatory patterns of miRNA expressions in adenoma and carcinoma samples in

comparison with control samples.

3.3.2.1 Hierarchal cluster analysis

MammU6 normalised (∆CT) values were used for clustering and classification of cases

based on miRNA expression levels in the plasma samples of the three groups. Nearest

neighbouring based on miRNA clustering grouped six cases (n = 3 adenoma, n = 3

carcinoma) amongst controls. Similarly, three cases of controls were spread out in the

cluster of adenoma and carcinoma samples. Clustering of cases based on HCA is shown

in Figure 18.

Figure 18: Hierarchical cluster analysis.

Clustering analysis for proximity matrix based on agglomeration schedule and clustering

method of nearest neighbouring with euclidean distance measurement.

111

3.3.2.2 Principle component analysis

Principle component analysis identified a panel of 20 diagnostic miRNAs with

discriminatory (p<0.001) expression levels for carcinoma and adenoma (Figure 19). PCA

identified miR-502-5P as a stand out miRNA to distinguish adenomas and carcinomas

from normal controls.

Figure 19: Principle component analysis for discriminatory miRNAs

2 dimensional figure showing the spread of different discriminatory 20 miRNAs (miR-

16, miR-21, miR-23b, miR-34a, miR-92a, miR-135b, miR-139-5p, miR-182, miR-192*,

miR-193a-5p, miR-200a*, miR-410, miR-431, miR-487b, miR-449a, miR-564, miR-589,

miR-592, miR-624*, miR-645)

112

3.3.2.3 Calculation of Z-scores

Average and standard deviation of MammU6 normalised expression levels (∆CT) for

control samples was used to calculate Z-scores for the expression level of miRNAs in

adenoma and carcinoma samples. The following equation was used to compute the Z-

scores:

𝑍 𝑆𝑐𝑜𝑟𝑒

=Expression levels for a case of adenoma or carcinoma − Average expression levels of normals

Standard deviation in expression levels of normals

Z-scores values of >1.96 and < -1.50 were deemed significant to identify the differences

in expression levels. Number of participants with Z-score >1.96 or <-1.50 were counted

for both adenoma and carcinoma groups. A total pick up rate (%) was calculated based

on the percentage of participants with Z-scores of >1.96 or <-1.50 (Table 22, Appendix

VII). miR-502-5P expression was associated with the highest number of cases with

significant Z-scores (n = 13), detecting 75% cases of carcinoma and 44% cases of

adenoma.

miRNA Cases with Z-score >1.96 or < -1.50 Pick up rates %

Adenoma

(n)

Carcinoma

(n)

Total (n)

Adenoma Carcinoma Total

miR-502-5p 4 9 13 44 75 62

miR-192* 2 7 9 22 58 43

miR-564 5 4 9 56 33 43

miR-410 3 5 8 33 42 38

miR-431 2 5 7 22 42 33

miR-200a* 2 5 7 22 42 33

miR-135b 4 3 7 44 25 33

miR-182 3 5 8 33 42 38

miR-186 3 5 8 33 42 38

miR-16 5 1 6 44 8.3 28

miR-34a 4 2 6 44 17 28

miR-23b 1 4 5 11 33 24

miR-203 4 0 4 44 0 19

Table 22: Z-scores calculated from miRNA array data for both adenoma and carcinoma samples.

113

A combination of different miRNAs was used to identify complementary miRNAs for the

detection of more cases. For example, a combination of miR-502-5p, miR-192*, miR-

410 and miR-139-5p showed 100% identification rates for adenoma and carcinoma.

Table 23 shows different combinations of complimentary miRNAs and their pick up

rates based on Z-scores.

Complementary miRNA

combinations

Z-score-based detection rates for neoplasia

Adenoma Carcinoma Total

miR-502-5p, miR-192*, miR-410 and

miR-139 5P 100 100 100

miR-502-5p and miR-139 5P 89 83 86

miR-502-5p, miR192* and miR-564 89 83 86

miR-502-5p, miR-192* and miR-410 67 92 81

miR-502-5p, miR-192* and miR-

135b 78 83 81

miR-502-5p- and miR-192* 67 83 76

miR-502-5p, and miR-564 78 67 76

miR-502-5p and miR-410 56 83 71

miR-502-5p, and miR-135b 56 67 62

Table 23: Complimentary miRNAs and their detection rates for neoplasia.

114

3.3.2.4 Identification of diagnostic miRNAs based on bioinformatics analysis

Six significant miRNAs were identified based on the bioinformatics analysis performed

by Dr Graham Ball of the Department of Bioinformatics at the University of Nottingham.

Initially, a panel of 20 miRNAs were identified based on median performance score to

discriminate controls from adenomas and carcinomas. This panel was further refined in

a test performance and a set of six discriminatory miRNAs (miR-502-5p, miR-369-5p,

miR-193a-5p, miR-589, miR-92a and miR-449a) was identified. miR-502-5p was

associated with the highest performance score in training (0.85) and test performance

(0.80) (Table 24)

miRNA Median training

performance

Average

training error

Median test

performance

Average test

error

miR-502-5p 0.85 0.066 0.80 0.068

miR-369-5p 0.71 0.091 0.80 0.097

miR-193a-5p 0.71 0.099 0.80 0.097

miR-589 0.71 0.096 0.80 0.097

miR-449a 0.71 0.800 0.75 0.099

miR-92a 0.64 0.800 0.50 0.105

Table 24: Summary of median performance scores and errors for 6 discriminatory miRNAs identified with Bioinformatics analysis.

3.3.2.5 Inter-group comparisons of expression levels using the Student’s t-test.

Highly significant miRNAs identified from non-matched paired Student’s t–test of

MammU6 normalised profiling data have been shown in table 25. This table shows p-

values for differences in expression levels of selected miRNAs in adenoma and

carcinoma groups.in comparison with the control group. Inter-group comparisons

showed significantly higher expression levels of miR-502-5p for carcinoma (p<0.0001)

and adenoma (p=0.0349) groups in comparison with control group. Expression levels

(∆CT) of miR-135b, miR-455-5p and miR-564 were significantly higher in the plasma of

patients with adenomas only in comparison with controls. In contrast, expression levels

of miR-182*, miR-369-5p, miR-431, miR-487b, miR-589 and miR-645 were significantly

different for carcinomas only.

115

Table 25: P-values for inter-group differences in expression levels of selected miRNA for adenoma or carcinoma groups in comparison to the control group.

3.3.2.6 ROC analysis for selected example miRNAs

ROC analysis for miR-135b, miR-192*, miR-502-5P and miR-645 were applied to assess

their diagnostic accuracy for adenomas, carcinomas and a combined group of both

adenomas and carcinomas. For the detection of both adenoma and carcinoma, the area

under the curve (AUC) values were 0.73, 0.60, 0.87 and 0.71 for miR-135b, miR-192*,

miR-502-5P and miR-645, respectively. For adenoma only, AUC values were 0.83, 0.58,

0.78 and 0.82 for miR-135b, miR-192*, miR-502-5p and miR-645, respectively. For

carcinoma only, AUCs were 0.65, 0.61, 0.93 and 0.64 for miR-135b, miR-192*, miR-502-

5p and miR-645, respectively. Relative expression levels of miR-135b, miR-192*, miR-

502-5p and miR-645 for the different groups are shown in Figures 20 to 23, respectively.

Figures also show ROC curves illustrating the diagnostic accuracy of each miRNA for

adenomas, carcinomas and a combined group of adenomas and carcinomas.

miRNA Non matched paired Student’s t-test (p-values)

Adenoma Carcinoma Adenoma + Carcinoma

miR-502-5P 0.0349 <0.0001 0.0006

miR-589 0.6292 0.0017 0.0360

miR-564 0.0064 0.2837 0.0510

miR-135b 0.0489 0.1476 0.0563

miR-455-5p 0.0149 0.3029 0.0819

miR-431 0.4670 0.0378 0.0880

miR-369-5p 0.8684 0.0091 0.1006

miR-645 0.5419 0.0077 0.1304

miR-487b 0.4299 0.0719 0.1378

miR-182* 0.5305 0.0489 0.8811

116

N o rm a l A d e n o m a C a rc in o m a A d e n o m a + C a rc in o m a

0

2 0

4 0

6 0

m iR -1 3 5 b

Ra

nk

s

0 5 0 1 0 0

0

5 0

1 0 0

m iR -1 3 5 b N o rm a l v s A d e n o m a + C a rc in o m a

1 0 0 % - S p e c if ic ity %

Se

ns

itiv

ity

%

0 5 0 1 0 0

0

5 0

1 0 0

m iR -1 3 5 b N o rm a l v s A d e n o m a

1 0 0 % - S p e c if ic ity %

Se

ns

itiv

ity

%

0 5 0 1 0 0

0

5 0

1 0 0

m iR -1 3 5 b N o rm a l v s C a rc in o m a

1 0 0 % - S p e c if ic ity %

Se

ns

itiv

ity

%

a

c

b

d

Figure 20: ∆CT-based expression analysis of plasma miR-135b detected in expression profiling.

MammU6 normalised expression levels (∆CT) of miR-135b were compared (Mann-

Whitney-U-Test) and displayed in box plots. The diagnostic accuracy for the detection

of adenomas, carcinomas and the combined group of adenomas and carcinomas was

evaluated by ROC analysis. (a) Significantly higher expression levels were detected for

the combined group of patients with adenomas and carcinomas in comparison to

controls (median difference = 2.53, 95%CI: 0.08 – 5.3, p=0.038). (b) ROC analysis for the

detection of both adenoma and carcinoma show an AUC value of 0.73 (95% CI: 0.54 -

0.92) with 62% sensitivity and 80% specificity at a likelihood ratio of 3.01. (c) ROC

analysis for the detection of adenomas shows an AUC value of 0.83 (95% CI: 0.64 – 1.01).

(d) ROC analysis for the detection of carcinomas shows an AUC value of 0.65 (95% CI:

0.42 - 0.89).

117

N o rm a l A d e n o m a C a rc in o m a A d e n o m a + C a rc in o m a

0

2 0

4 0

6 0

m iR -1 9 2 *

Ra

nk

s

0 5 0 1 0 0

0

5 0

1 0 0

m iR -1 9 2 * N o rm a l v s A d e n o m a + C a rc in o m a

1 0 0 % - S p e c if ic ity %

Se

ns

itiv

ity

%

0 5 0 1 0 0

0

5 0

1 0 0

m iR -1 9 2 * N o rm a l v s A d e n o m a

1 0 0 % - S p e c if ic ity %

Se

ns

itiv

ity

%

0 5 0 1 0 0

0

5 0

1 0 0

m iR -1 9 2 * N o rm a l v s C a rc in o m a

1 0 0 % - S p e c if ic ity %

Se

ns

itiv

ity

%

a

c

b

d

Figure 21: ∆CT-based expression analysis of plasma miR-192* detected in expression profiling.

MammU6 normalised expression levels (∆CT) of miR-192* were compared (Mann-

Whitney-U-Test) and plotted in scatter graphs. (a) Box plot shows higher expression

levels were detected for a combined group of patients with adenomas and carcinomas

(median difference = 1.87, 95% CI: 0.99 – 4.49, p=0.36) in comparison to controls. (b)

ROC curve for the detection of both adenoma and carcinoma shows an AUC value of

0.60 (95% CI: 0.40 - 0.80) with 52% sensitivity and 80% specificity at a likelihood ratio of

2.61. (c) ROC curve for the detection of adenomas shows an AUC value of 0.58 (95% CI:

0.31 – 0.86). (d) ROC curve for the detection of carcinomas shows an AUC value of 0.61

(95% CI: 0.35 – 0.88).

118

N o rm a l A d e n o m a C a rc in o m a A d e n o m a + C a rc in o m a

0

2 0

4 0

6 0

m iR -5 0 2 -5 P

Ra

nk

s

0 5 0 1 0 0

0

5 0

1 0 0

m iR -5 0 2 -5 P N o rm a l v s A d e n o m a + C a rc in o m a

1 0 0 % - S p e c if ic ity %

Se

ns

itiv

ity

%

0 5 0 1 0 0

0

5 0

1 0 0

m iR -5 0 2 -5 P N o rm a l v s A d e n o m a

1 0 0 % - S p e c if ic ity %

Se

ns

itiv

ity

%

0 5 0 1 0 0

0

5 0

1 0 0

m iR -5 0 2 -5 P N o rm a l v s C a rc in o m a

1 0 0 % - S p e c if ic ity %

Se

ns

itiv

ity

%

a b

c d

Figure 22: ∆CT based expression analysis of plasma miR-502-5p detected in expression profiling.

MammU6 normalised expression levels (∆CT) of miR-502-5p were compared (Mann-

Whitney-U-Test) and plotted in scatter graphs. (a) Significantly higher expression levels

were detected for a combined group of patients with adenomas and carcinomas

(median difference = 3.79, 95% CI: 2.02 – 5.38, p=0.0005, ANOVA with Bonferroni

correction) in comparison to controls. (b) ROC curve for the detection of both adenoma

and carcinoma shows an AUC value of 0.87 (95% CI: 0.74 - 0.99) with 86% sensitivity

and 80% specificity at a likelihood ratio of 4.28. (c) ROC curve for the detection of

adenomas shows an AUC value of 0.78 (95% CI: 0.55 – 1.02). (d) ROC curve for the

detection of carcinomas shows an AUC value of 0.93 (95% CI: 0.82 – 1.04).

119

N o rm a l A d e n o m a C a rc in o m a A d e n o m a + C a rc in o m a

0

2 0

4 0

6 0

m iR -5 6 4

Ra

nk

s

0 5 0 1 0 0

0

5 0

1 0 0

m iR -5 6 4 N o rm a l v s A d e n o m a + C a rc in o m a

1 0 0 % - S p e c if ic ity %

Se

ns

itiv

ity

%

0 5 0 1 0 0

0

5 0

1 0 0

m iR -5 6 4 N o rm a l v s A d e n o m a

1 0 0 % - S p e c if ic ity %

Se

ns

itiv

ity

%

0 5 0 1 0 0

0

5 0

1 0 0

m iR -5 6 4 N o rm a l v s C a rc in o m a

1 0 0 % - S p e c if ic ity %

Se

ns

itiv

ity

%

a

c

b

d

Figure 23: ∆CT-based expression analysis of plasma miR-564 detected in expression profiling.

MammU6 normalised expression levels (∆CT) of miR-564 were compared (Mann-

Whitney-U-Test) and plotted in scatter graphs. (a) Significantly higher expression levels

were detected for a combined group of patients with adenomas and carcinomas

(median difference = 2.30, 95% CI: 0.05 – 4.00, p=0.05) in comparison to controls. (b)

ROC curve for the detection of both adenomas and carcinomas shows an AUC value of

0.71 (95% CI: 0.54 - 0.90) with 52% sensitivity and 80% specificity at a likelihood ratio of

2.61. (c) ROC curve for the detection of adenomas shows an AUC value of 0.82 (95% CI:

0.63 – 1.01). (d) ROC curve for the detection of carcinomas shows an AUC value of 0.64

(95% CI: 0.40 – 0.88).

120

3.3.2.7 Upregulation of passenger strand miRNA* in plasma and tissue

Expression profiling on 32 plasma samples and pooled cDNA from cancerous and

adjacent normal tissue from five patients with CRC was compared to assess the

presence of diagnostic miRNAs in circulation. One novel finding was the upregulation of

passenger strand miRNA* in the cancer tissue in comparison to their counterpart

mature strand miRNA. The comparison of MammU6 normalised miRNA* and miRNA

expressions (∆∆CT) for cancer tissue and the corresponding plasma levels showed

significantly ∆∆CT values for miR-624* and miR-182* in cancer tissue and plasma

samples from CRC patients. No such change in expression levels was noted for adenoma

samples in comparison with control samples. Figure 24 shows the ∆∆CT values for

passenger strand miRNA* in comparison to their counterpart mature miRNA in CRC

tissue and matching plasma samples

Figure 24: Expression levels of miRNA and their matched miRNA* in plasma and tumours

Figure shows upregulation of miRNA* in comparison to their counterpart miRNAs in

tissue and plasma samples taken from patients with cancer.

121

3.3.2.8 Selection of miRNAs for the validation cohort

Based on different statistical analysis, discriminatory patterns for passenger miRNA*

and their counterpart mature miRNAs, specific pattern of expressions in adenomas, and

a literature search of significant tissue-based miRNAs in CRC, a panel of 29 miRNAs was

selected for further validation (Table 26).

miRNA PCA Z-Score

t-test

Bioinfor-matics

Star vs. nonstar

Specific patterns

Literature

1 miR-16

2 miR-21

3 miR-23b

4 miR-34a

5 miR-92a

6 miR-95

7 miR-135b

8 miR-181c*

9 miR-182

10 miR-182*

11 miR-191

12 miR-192*

13 miR-195

14 miR-200a*

15 miR-203

16 miR-205

17 miR-369-5p

18 miR-410

19 miR-431

20 miR-486-3p

21 miR-486-5p

22 miR-487b

23 miR-502-5p

24 miR-532-5p

25 miR-564

26 miR-566

27 miR-589

28 miR-592

29 miR-624*

Table 26: Panel of 29 miRNAs selected for further validation.

Table summarises the miRNAs and method of selection for further validation

122

3.3.3 Validation of Novel Circulating miRNAs for CRC Detection: Initial Validation

Phase

A panel of 29 miRNAs was applied to a cohort of 94 participants (n = 32 controls, n = 28

adenoma, n = 34 carcinoma). All participants in the control group had no adenoma or

carcinoma at the time of blood sample collection. However, for the purpose of analysis,

participants with significant benign bowel disease (n = 9) and a history of neoplastic

disease (adenoma/carcinoma excised, n = 8) were excluded from the analysis. Appendix

X shows the total RNA concentrations for all the participants in the initial validation

cohort. Appendix XI - XIII show CT and ∆CT based analysis for validation cohort. Analysis

were also performed for miR-191 based normalized expressions and data is shown in

appendix XIV.

3.3.3.1 ROC analysis for the validation cohort

SnRNA RNU6B normalized expression levels (∆CT) of miRNAs were used to assess AUC

for the detection of adenomas, carcinomas and combined group of adenomas and

carcinomas. Tables 27 to 29 show AUC for different miRNAs for the detection of the

combined group of adenomas and carcinomas

Based on higher values of AUC and p-values, ROC curves for miR-135b, miR-34a and

miR-431 were selected for graphic presentation. Expressions levels for sub-categories

of controls (significant past history of neoplasia and significant benign bowel disease)

and participants with adenoma and carcinoma were compared by ANOVA with

Bonferroni corrections.

123

miRNAs

AUC

Standard Error p-value

95% CI for AUC - combined group

Lower bound Higher bound

miR-135b 0.793 0.054 <0.001 0.686 0.899

miR-34a 0.820 0.058 <0.001 0.706 0.935

miR-431 0.809 0.051 <0.001 0.710 0.908

miR-16 0.757 0.071 0.001 0.619 0.896

miR-369-5P 0.743 0.064 0.002 0.618 0.868

miR-23b 0.741 0.064 0.003 0.616 0.866

miR-191 0.731 0.064 0.004 0.606 0.855

miR-21 0.725 0.065 0.005 0.599 0.852

miR-589 0.722 0.064 0.006 0.596 0.848

miR-487b 0.719 0.065 0.006 0.592 0.847

miR-95 0.713 0.069 0.008 0.579 0.847

miR-484 0.707 0.068 0.010 0.574 0.840

miR-195 0.704 0.075 0.011 0.557 0.851

miR-181C* 0.698 0.065 0.014 0.571 0.825

miR-410 0.691 0.069 0.017 0.556 0.827

miR-532-5P 0.675 0.070 0.029 0.537 0.813

miR-192* 0.674 0.064 0.030 0.548 0.799

miR-203 0.665 0.074 0.040 0.520 0.809

miR-486-5P 0.646 0.081 0.069 0.487 0.805

miR-566 0.638 0.073 0.086 0.495 0.780

miR-502-5P 0.636 0.070 0.091 0.499 0.772

miR-624* 0.613 0.069 0.159 0.477 0.749

miR-205 0.600 0.078 0.215 0.447 0.753

miR-566 0.582 0.078 0.307 0.429 0.736

miR-200a* 0.563 0.068 0.429 0.430 0.696

miR-92a 0.550 0.081 0.533 0.391 0.709

miR-592 0.473 0.072 0.733 0.332 0.613

miR-182* 0.476 0.069 0.763 0.341 0.611

miR-486-3P 0.511 0.090 0.893 0.334 0.688

Table 27: ROC analysis for detection of the combined group of adenoma and

carcinoma.

Table shows AUC values for each miRNA with standard error, and 95% confidence

intervals for AUC. miRNAs are ranked based on statistical significance (p-value).

124

Table 28: ROC analysis for the detection of adenoma. AUC values for each miRNA with standard error and 95% confidence intervals. miRNAs

are ranked by their statistical significance (p-value). miR-135b, miR34a, miR-431, miR-

369-5P, miR-16 and miR-21 were statistically significant (P<0.05).

miRNAs AUC Standard

Error p-value

95% CI for AUC - adenoma group

Lower bound Higher bound

miR-34a 0.771 0.083 0.006 0.608 0.934

miR-135b 0.743 0.085 0.013 0.576 0.910

miR-431 0.721 0.085 0.023 0.555 0.888

miR-369-5P 0.715 0.088 0.028 0.543 0.887

miR-16 0.715 0.089 0.028 0.541 0.889

miR-21 0.697 0.089 0.044 0.522 0.871

miR-195 0.678 0.092 0.068 0.498 0.858

miR-566 0.678 0.092 0.068 0.497 0.859

miR-487b 0.659 0.092 0.103 0.480 0.839

miR-23b 0.656 0.092 0.110 0.475 0.837

miR-502-5P 0.653 0.095 0.117 0.467 0.840

miR-181C* 0.641 0.094 0.149 0.457 0.825

miR-191 0.638 0.093 0.159 0.456 0.820

miR-410 0.638 0.093 0.159 0.456 0.820

miR-589 0.635 0.095 0.168 0.448 0.822

miR-532-5P 0.635 0.094 0.168 0.450 0.819

miR-484 0.632 0.095 0.178 0.445 0.818

miR-486-5P 0.625 0.095 0.199 0.440 0.811

miR-95 0.619 0.095 0.222 0.433 0.805

miR-203 0.607 0.097 0.274 0.417 0.796

miR-192* 0.591 0.097 0.350 0.401 0.781

miR-624* 0.579 0.097 0.419 0.388 0.770

miR-200a* 0.542 0.099 0.669 0.348 0.736

miR-205 0.539 0.098 0.692 0.347 0.731

miR-182* 0.480 0.102 0.837 0.280 0.680

miR-566 0.520 0.098 0.837 0.327 0.713

miR-592 0.489 0.099 0.912 0.296 0.683

miR-486-3P 0.495 0.100 0.962 0.299 0.692

miR-92a 0.502 0.099 0.987 0.307 0.690

125

Table 29: ROC analysis for the detection of carcinomas.

AUC values for each miRNA with standard error and 95% confidence intervals. miRNAs

are ranked by their statistical significance (p-value). miR-135b, miR34a, miR-431, miR-

191 and miR-23b were statistically most significant (p≤0.001).

miRNAs AUC Standard

Error p-value

95% CI for AUC carcinoma group

Lower bound Higher bound

miR-135b 0.812 0.060 <0.001 0.694 0.931

miR-34a 0.843 0.060 <0.001 0.725 0.960

miR-431 0.849 0.051 <0.001 0.749 0.949

miR-191 0.771 0.067 0.001 0.640 0.902

miR-23b 0.777 0.067 0.001 0.647 0.908

miR-16 0.774 0.072 0.001 0.633 0.916

miR-589 0.760 0.068 0.002 0.626 0.894

miR-369-5p 0.750 0.068 0.003 0.617 0.883

miR-95 0.757 0.071 0.003 0.618 0.896

miR-484 0.738 0.072 0.005 0.597 0.878

miR-487b 0.742 0.068 0.005 0.609 0.876

miR-21 0.733 0.070 0.006 0.595 0.870

miR-181C* 0.719 0.072 0.010 0.578 0.860

miR-195 0.709 0.078 0.014 0.555 0.863

miR-410 0.711 0.072 0.014 0.569 0.853

miR-192* 0.707 0.073 0.015 0.565 0.850

miR-203 0.688 0.076 0.027 0.538 0.838

miR-532-5P 0.688 0.076 0.027 0.540 0.837

miR-486-5P 0.649 0.086 0.082 0.481 0.816

miR-205 0.626 0.084 0.139 0.462 0.790

miR-624* 0.622 0.076 0.154 0.473 0.771

miR-502-5P 0.620 0.078 0.160 0.466 0.774

miR-566 0.614 0.079 0.183 0.458 0.769

miR-566 0.612 0.082 0.189 0.451 0.773

miR-200a* 0.568 0.077 0.429 0.416 0.719

miR-92a 0.563 0.086 0.462 0.394 0.731

miR-592 0.459 0.078 0.635 0.306 0.613

miR-182* 0.463 0.079 0.662 0.308 0.618

miR-486-3P 0.512 0.094 0.889 0.327 0.697

126

3.3.3.2 Comparison of expressions in different groups

SNRNA RNU6B normalized expression of each miRNA (miR-135b, miR-431, miR-34a)

was used to calculate the average expression of that miRNA. This average expression

was then used to compare the difference in expression levels between each case and

compared with the control group. Inter-group comparisons by ANOVA with Bonferroni

corrections showed statistically significant differences in expression levels (Table 30)

Comparisons

Mean difference

95% CI p-value

miR-135b

Controls versus past neoplasia 4.349 1.055 to 7.643 0.0037

Controls versus significant benign disease 2.577 -0.4845 to 5.639 0.1479

Controls versus adenoma 2.467 -0.09803 to 5.031 0.0658

Controls versus carcinoma 3.233 0.9722 to 5.494 0.0014

Controls versus adenoma and carcinoma 2.968 0.8364 to 5.100 0.0019

miR-431

Controls versus past neoplasia -2.666 -7.280 to 1.949 0.6693

Controls versus significant benign disease

-3.793 -8.254 to 0.6681 0.1402

Controls versus adenoma -2.603 -6.255 to 1.049 0.3246

Controls versus carcinoma -4.798 -8.051 to -1.545 0.0009

Controls versus adenoma and carcinoma

-4.030 -7.105 to -0.9541 0.0041

miR-34a

Controls versus past neoplasia 4.045 0.6538 to 7.437 0.0044

Controls versus significant benign disease

4.080 0.8015 to 7.359 0.0043

Controls versus adenoma 3.038 0.3539 to 5.722 0.0044

Controls versus carcinoma 3.968 1.577 to 6.359 0.0001

Controls versus adenoma and carcinoma

3.643 1.382 to 5.903 0.0002

Table 30: Inter-group comparison of miR-135b, miR-431 and miR-34a.

Table shows comparison of expression levels normalized to average of expression for

controls, mean differences in expression levels with 95% CI and p-values.

127

Significantly higher expression levels of miR-135b were noted for controls with a past

history of neoplasia, participants with carcinoma and a cumulative group of participants

with adenoma and carcinoma (ANOVA with Bonferroni correction, p<0.0034) (Figure

25).Inter-group comparisons of miR-431 expression levels showed significantly higher

expression in a carcinoma group and a combined group of adenomas and carcinomas.

No statistically significant differences were noted for patients with adenoma, significant

benign disease or a history of neoplasia. There was increase in expression levels for

these groups but the differences did not reach statistical significance (Figure 26).

ANOVA showed a significantly higher level of miR-34a expression in all groups in

comparison with controls (p<0.0008) (Figure 27).

N o rm a l P a s t N e o p la s ia S ig n if ic a n t B e n ig n d is e a s e A d e n o m a C a rc in o m a A d e n o m a + C a rc in o m a

-1 0

-5

0

5

R e la t iv e e x p re s s io n o f c a s e s n o rm a lis e d to a v e ra g e e x p re s s io n o f m iR -1 3 5 b in n o rm a ls

C a s e s

De

lta

CT

va

lue

s

Figure 25: Comparative expression of miR-135b for different groups of controls and

neoplasia.

Significantly higher expression levels were noted for controls with a past history of

neoplasia, participants with carcinoma and a cumulative group of participants with

adenoma and carcinoma (ANOVA with Bonferroni correction, p<0.0034).

128

N o rm a l P a s t N e o p la s ia S ig n if ic a n t B e n ig n d is e a s e A d e n o m a C a rc in o m a A d e n o m a + C a rc in o m a

-1 0

-5

0

5

1 0

R e la t iv e e x p re s s io n o f c a s e s n o rm a lis e d to a v e ra g e e x p re s s io n o f m iR -4 3 1 in n o rm a ls

C a s e s

De

lta

CT

va

lue

s

Figure 26: Comparative expression of miR-431 for different groups of normal and

neoplasia.

Significantly higher expression levels were noted for carcinoma and a cumulative group

of participants with adenoma and carcinoma (ANOVA with Bonferroni correction,

p<0.0059).

129

N o r m a l P a s t N e o p la s ia S ig n if ic a n t B e n ig n d is e a s e Ad e n o m a C a r c in o m a Ad e n o m a + C a r c in o m a

-1 0

-5

0

5

R e la t iv e e x p re s s io n o f c a s e s n o rm a lis e d to a v e ra g e e x p re s s io n o f m iR -3 4 a in n o rm a ls

C a s e s

De

lta

C

T v

alu

es

Figure 27: Comparative expression of miR-34a in different groups of controls and

neoplasia.

Significantly higher expression levels were noted for carcinoma and a cumulative group

of participants with adenoma and carcinoma (ANOVA with Bonferroni correction,

p=0.0008).

130

3.3.3.3 ROC analysis for miR-34a, miR-135b and miR-431

For three selected miRNAs, the ROC analyses and curves are displayed with AUC,

sensitivities, specificities and likelihood ratios in figures 28-30. miR-135b ∆CT values

were compared (Mann Whitney U Test) for different groups and plotted in scatter

graphs. The diagnostic accuracy for the detection of adenoma, carcinoma and

combined group of adenoma and carcinoma was evaluated with ROC analysis.

131

0 5 0 1 0 0

0

5 0

1 0 0

m iR -1 3 5 b N o rm a l v s A d e n o m a + C a rc in o m a

1 0 0 % - S p e c if ic ity %

Se

ns

itiv

ity

%

0 5 0 1 0 0

0

5 0

1 0 0

m iR -1 3 5 b N o rm a l v s A d e n o m a

1 0 0 % - S p e c if ic ity %

Se

ns

itiv

ity

%

0 5 0 1 0 0

0

5 0

1 0 0

m iR -1 3 5 b N o rm a l v s C a rc in o m a

1 0 0 % - S p e c if ic ity %

Se

ns

itiv

ity

%

N o rm a l A d e n o m a + C a rc in o m a

0

2 0

4 0

6 0

8 0

Ra

nk

s

m iR -1 3 5 b N o rm a l v s A d e n o m a + C a rc in o m a

N o rm a l A d e n o m a

0

1 0

2 0

3 0

4 0

m iR -1 3 5 b N o rm a l v s A d e n o m a

Ra

nk

s

N o rm a l C a rc in o m a

0

2 0

4 0

6 0

m iR -1 3 5 b N o rm a l v s C a rc in o m a

Ra

nk

s

a

b d

c e

f

Figure 28: ∆CT-based expression analysis of plasma miR-135b detected in the

validation cohort.

miR-135b ∆CT values were compared (Mann-Whitney-U-Test) for different groups and

plotted in scatter graphs. (a) Significantly higher expression levels were detected for a

combined group of patients with adenomas and carcinomas (median difference = 3.52,

95% CI: 1.28 – 4.97, p=0.0004) in comparison to controls. (b) ROC analysis for the

detection of both adenomas and carcinomas show AUC value of 0.77 (95% CI: 0.66 to

0.88) with 74% sensitivity and 88% specificity (likelihood ratio = 6.33). (c) Significantly

higher expression levels were detected in the adenoma group in comparison to controls

(median difference = 2.77, 95% CI: 0.33 - 4.99, p=0.0275). (d) ROC analysis for the

detection of adenomas shows an AUC value of 0.71 (95% CI: 0.53 - 0.89) with 63%

sensitivity and 88% specificity (likelihood ratio = 5.36). (e) Significantly higher expression

levels were detected in the carcinoma-only group in comparison with controls (median

difference = 3.90, 95% CI: 1.37 - 5.92, p=0.0002). (f) ROC analysis for the detection of

carcinoma shows an AUC value of 0.80 (95% CI: 0.68 - 0.92) with 80% sensitivity and

88% specificity (likelihood ratio = 6.84).

132

0 5 0 1 0 0

0

5 0

1 0 0

m iR -4 3 1 N o rm a l v s A d e n o m a + C a rc in o m a

1 0 0 % - S p e c if ic ity %

Se

ns

itiv

ity

%

0 5 0 1 0 0

0

5 0

1 0 0

m iR -4 3 1 N o rm a l v s A d e n o m a

1 0 0 % - S p e c if ic ity %

Se

ns

itiv

ity

%

0 5 0 1 0 0

0

5 0

1 0 0

m iR -4 3 1 N o rm a l v s C a rc in o m a

1 0 0 % - S p e c if ic ity %

Se

ns

itiv

ity

%

N o rm a l A d e n o m a + C a rc in o m a

0

2 0

4 0

6 0

8 0

1 0 0

Ra

nk

s

m iR -4 3 1 N o rm a l v s A d e n o m a + C a rc in o m a

N o rm a l A d e n o m a

0

1 0

2 0

3 0

4 0

m iR -4 3 1 N o rm a l v s A d e n o m a

Ra

nk

s

N o rm a l C a rc in o m a

0

2 0

4 0

6 0

m iR -4 3 1 N o rm a l v s C a rc in o m a

Ra

nk

s

a

b

c

d

e

f

Figure 29: ∆CT-based expression analysis of plasma miR-431 detected in validation

cohort.

(a) Significantly higher expression levels were detected in the combined group of

patients with adenomas and carcinomas (median difference = 3.78, 95% CI: 1.81 – 6.21,

p=0.0003) in comparison to controls. (b) ROC analysis for the detection of both

adenomas and carcinomas shows an AUC value of 0.78 (95% CI: 0.68 to 0.88) with 60%

sensitivity and 88% specificity (likelihood ratio = 4.95). (c) Significantly higher expression

levels were detected in adenomas in comparison to controls (median difference = 1.84,

95% CI: 0.14 – 5.60, p=0.033). (d) ROC analysis for the detection of adenomas shows an

AUC value of 0.70 (95% CI: 0.53 - 0.86) with 43% sensitivity and 88% specificity

(likelihood ratio = 3.64). (e) Significantly higher expression levels for the carcinoma only

group in comparison with controls (median difference = 4.46, 95% CI: 2.65 – 6.95,

p<0.0001). (f) ROC analysis for the detection of carcinoma shows an AUC value of 0.82

(95% CI: 0.71 - 0.93) with 61% sensitivity and 88% specificity (likelihood ratio = 5.23).

133

0 5 0 1 0 0

0

5 0

1 0 0

m iR -3 4 a N o rm a l v s A d e n o m a + C a rc in o m a

1 0 0 % - S p e c if ic ity %

Se

ns

itiv

ity

%

0 5 0 1 0 0

0

5 0

1 0 0

m iR -3 4 a N o rm a l v s A d e n o m a

1 0 0 % - S p e c if ic ity %

Se

ns

itiv

ity

%

0 5 0 1 0 0

0

5 0

1 0 0

m iR -3 4 a N o rm a l v s C a rc in o m a

1 0 0 % - S p e c if ic ity %

Se

ns

itiv

ity

%

N o rm a l A d e n o m a + C a rc in o m a

0

2 0

4 0

6 0

8 0

1 0 0

Ra

nk

s

m iR -3 4 a N o rm a l v s A d e n o m a + C a rc in o m a

N o rm a l A d e n o m a

0

1 0

2 0

3 0

4 0

m iR -3 4 a N o rm a l v s A d e n o m aR

an

ks

N o rm a l C a rc in o m a

0

2 0

4 0

6 0

m iR -3 4 a N o rm a l v s C a rc in o m a

Ra

nk

s

a

b d

c

f

e

Figure 30: ∆CT-based expression analysis of plasma miR-34a detected validation

cohort.

a) Significantly higher expression levels were detected in combined group of patients

with adenomas and carcinomas (median difference = 4.45, 95% CI: 2.14 – 5.55,

p=0.0001) in comparison to controls. (b) ROC analysis for the detection of both

adenomas and carcinomas shows an AUC value of 0.79 (95% CI: 0.67 to 0.91) with 61%

sensitivity and 88% specificity (likelihood ratio = 5.24). (c) Significantly higher expression

levels were detected in adenomas in comparison to controls (median difference = 4.54,

95% CI: 1.26 – 5.60, p=0.0095). (d) ROC analysis for the detection of adenomas shows

an AUC value of 0.74 (95% CI: 0.58 to 0.91) with 62% sensitivity and 88% specificity

(likelihood ratio = 5.26). (e) Significantly higher expression levels for carcinoma only

group in comparison with controls (median difference = 4.36, 95% CI: 2.22 - 5.83,

p<0.0001). (f) ROC analysis for the detection of carcinoma shows an AUC value of 0.82

(95% CI: 0.70 - 0.94) with 64% sensitivity and 88% specificity (likelihood ratio = 5.44).

134

3.3.3.4 Logistic regression combined with ROC analysis

SnRNA RNU6B normalised expression levels for significant miRNAs (p<0.05, ROC

analysis) were used in a logistic regression model to calculate a probability value for the

presence of adenomas and carcinomas. Probability values ranged from 0.00 to 1.00,

where 1.00 indicated the presence of adenoma or carcinoma and 0.00 was indicative of

normal health status. Probability values generated from logistic regression analyses

were used for ROC analysis to assess AUC, sensitivity, specificity and likelihood ratios.

ROC analysis for the detection of both adenomas and carcinomas with this panel of 18

significant miRNAs (miR-135b, miR-34a, miR-431, miR-16, miR-369-5p, miR-23b, miR-

191, miR-21, miR-589, miR-487b, miR-95, miR-484, miR-195, miR-181C*, miR-410, miR-

532-5p, miR-192*, miR-203) show an AUC value of 0.92 (95% CI: 0.85 - 1.00) with 92%

sensitivity and 88% specificity (likelihood ratio= 7.77) as shown in figure 31. Probability

values were also calculated for a panel of the three most significant miRNAs (miR-34a,

miR-135b and miR-431). ROC curves for this panel with ROC analysis for the detection

of both adenomas and carcinomas showed an AUC value of 0.92 (95% CI: 0.84 - 0.99)

with 91% sensitivity and 88% specificity (likelihood ratio= 7.77) (Figure 32).

135

0 5 0 1 0 0

0

5 0

1 0 0

N o rm a l v s A d e n o m a + C a rc in o m a

1 0 0 % - S p e c if ic ity %

Se

ns

itiv

ity

%

0 5 0 1 0 0

0

5 0

1 0 0

N o rm a l V s A d e n o m a

1 0 0 % - S p e c if ic ity %S

en

sit

ivit

y%

0 5 0 1 0 0

0

5 0

1 0 0

N o rm a l v s C a rc in o m a

1 0 0 % - S p e c if ic ity %

Se

ns

itiv

ity

%

N o rm a l A d e n o m a + C a rc in o m a

0 .0

0 .5

1 .0

N o rm a l v s . A d e n o m a + C a rc in o m a

C a s e s

Pro

ba

bil

ity

N o rm a l A d e n o m a

0 .0

0 .5

1 .0

N o rm a l v s A d e n o m a

C a s e s

Pro

ba

bil

ity

N o rm a l C a rc in o m a

0 .0

0 .5

1 .0

N o rm a l v s C a rc in o m a

C a s e s

Pro

ba

bil

ity

a

b

c

d

e

f

Figure 31: Logistic regression combined with ROC analysis for a panel of 18 miRNAs.

(a) A median probability value of 0.67 (95%CI: 0.53 – 0.69, p<0.0001) for the combined

group of patients with adenomas and carcinomas was detected. (b) ROC analysis for the

detection of both adenomas and carcinomas shows an AUC value of 0.92 (95% CI: 0.85

- 1.00) with 92% sensitivity and 88% specificity (likelihood ratio= 7.77). (c) A median

probability value of 0.67 (95% CI: 0.43 – 0.69, p<0.0001) was calculated for the

detection of adenomas. (d) ROC analysis for the detection of adenomas shows an AUC

value of 0.90 (95% CI: 0.80 - 1.00) with 90% sensitivity and 88% specificity (likelihood

ratio= 7.60). (e) A median probability value of 0.68 (95%CI: 0.54 – 0.70, Mann-Whitney-

U-Test, p<0.0001) was calculated for the detection of adenomas. (f) ROC analysis for

the detection of carcinoma shows AUC value of 0.93 (95% CI: 0.86 - 1.00) with 92.5%

sensitivity and 88% specificity (likelihood ratio= 7.86).

136

0 5 0 1 0 0

0

5 0

1 0 0

N o rm a l v s A d e n o m a + C a rc in o m a

1 0 0 % - S p e c if ic ity %

Se

ns

itiv

ity

%

0 5 0 1 0 0

0

5 0

1 0 0

N o rm a l V s A d e n o m a

1 0 0 % - S p e c if ic ity %

Se

ns

itiv

ity

%

0 5 0 1 0 0

0

5 0

1 0 0

N o rm a l v s C a rc in o m a

1 0 0 % - S p e c if ic ity %

Se

ns

itiv

ity

%

N o rm a l A d e n o m a + C a rc in o m a

0 .0

0 .5

1 .0

N o rm a l v s . A d e n o m a + C a rc in o m a

C a s e s

Pro

ba

bil

ity

N o rm a l A d e n o m a

0 .0

0 .5

1 .0

N o rm a l v s A d e n o m a

C a s e s

Pro

ba

bil

ity

N o rm a l C a rc in o m a

0 .0

0 .5

1 .0

N o rm a l v s C a rc in o m a

C a s e s

Pro

ba

bil

ity

a

b

c

d

e

f

Figure 32: Logistic regression combined with ROC analysis for a panel of 3 miRNAs

(miR-34a, miR-135b & miR-431).

(a) A median probability value of 0.56 (95% CI: 0.45 – 0.72, p<0.0001) was calculated

for a combined group of patients with adenomas and carcinomas. (b) ROC analysis for

the detection of both adenomas and carcinomas shows an AUC value of 0.92 (95% CI:

0.84 - 0.99) with 91% sensitivity and 88% specificity (cut-off value = 7.77). (c) A median

probability value of 0.54 (95% CI: 0.34 – 0.65, p<0.0001) was calculated for the

detection of adenomas. (d) ROC analysis for the detection of adenomas shows an AUC

value of 0.90 (95% CI: 0.79 - 1.01) with 90% sensitivity and 88% specificity (likelihood

ratio= 7.65). (e) A median probability value of 0.57 (95% CI: 0.45 – 0.65, p<0.0001) was

calculated for the detection of adenomas. (f) ROC analysis for the detection of

carcinoma shows an AUC value of 0.93 (95% CI: 0.85 - 1.00) with 90% sensitivity and

88% specificity (likelihood ratio= 7.65).

137

3.3.4 Validation of miR-135b in an independent cohort

Based on superior diagnostic accuracy, miR-135b was validated in a second

independent cohort of 30 adenomas, 41 carcinoma and 25 controls (Appendix XV &

XVI). RT-PCR was conducted in the same way as conducted in the previous validation

work. The non-matched paired Student’s t-test was performed to assess the differences

in expression levels of miR-135b for participants with adenoma and carcinoma.

Expression levels of miR-135b were not detectable in the majority of control cases

(median CT values = 38.13) (Appendix XVII). However, the presence of detectable levels

of miR-135b correlated with the presence of both adenomas and carcinomas. Analysis

of miR-135b expression levels normalised with snRNA RNU6B (∆CT) showed

upregulation in expression in participants with adenoma in comparison with controls,

but this did not reach statistical significance (p=0.109). However, the difference in

expression levels (∆CT) for carcinomas remained statistically significant in comparison

with expression levels (∆CT) for controls. Table 31 summarises the median expression

levels of miR-135b (CT and SnRNA RNU6B normalised ∆CT) for adenoma, carcinoma and

control groups.

Table 31: Median expression levels of miR-135b with associated p-values for differences in expression levels in groups with adenoma, carcinoma, and adenoma and carcinoma.

Expression levels (CT) and snRNA RNU6B normalised (∆CT) for miR-135b were compared

(Mann-Whitney-U-Test) for different groups and plotted in box plots. The diagnostic

accuracy for the detection of adenoma, carcinoma and combined group of adenoma &

carcinoma was evaluated with ROC analysis (Figure 33 & Figure 34). Tabulated ROC

analysis with AUC, sensitivity, specificity, cut off and likelihood ratios are given in Table

32 and Table 33.

miR-135b Median expression levels

of miR-135b

p-values

(non-matched paired Student’s t-test )

Control

Adenoma

Carcinoma

Control

vs.

Adenoma

Control

vs.

Carcinoma

Control Vs.

Adenoma +

Carcinoma

CT 38.13 34.07 29.49 0.0025 p<0.0001 P<0.0001

∆CT -6.74 -4.85 -0.73 0.1093 P<0.001 p<0.0001

138

0 5 0 1 0 0

0

5 0

1 0 0

N o rm a l v s . A d e n o m a + C a rc in o m a

1 0 0 % - S p e c if ic ity %

Se

ns

itiv

ity

%

0 5 0 1 0 0

0

5 0

1 0 0

N o rm a l v s . A d e n o m a

1 0 0 % - S p e c if ic ity %

Se

ns

itiv

ity

%

0 5 0 1 0 0

0

5 0

1 0 0

R O C c u rv e : R O C o f N o rm a l v s . C a rc in o m a

1 0 0 % - S p e c if ic ity %

Se

ns

itiv

ity

%

N o rm a l A d e n o m a + c a rc in o m a

2 0

2 5

3 0

3 5

4 0

N o rm a l v s A d e n o m a + C a rc in o m a

C a s e s

CT

va

lue

s

N o rm a l A d e n o m a

2 0

2 5

3 0

3 5

4 0

N o rm a l v s . A d e n o m a

C a s e s

CT

va

lue

s

N o rm a l C a rc in o m a

2 0

2 5

3 0

3 5

4 0

N o rm a l v s . C a rc in o m a

C a s e s

CT

va

lue

s

a

b

c

d

e

fS e n s itiv ity % Id e n t ity %

S e n s itiv ity %

Figure 33: CT-based expression analysis of miR-135b in an independent cohort.

(a) Significantly higher expression levels were detected for combined group of patients

with adenomas and carcinomas (median difference = 10.9, 95% CI: 5.12 – 10.52,

p<0.0001) in comparison to controls. (b) ROC analysis for the detection of both

adenomas and carcinomas shows an AUC value of 0.82 (95% CI: 0.71 to 0.92) with 80%

sensitivity and 84% specificity (likelihood ratio= 5.018). (c) Significantly higher

expression levels were detected in adenomas in comparison to controls (median

difference = 5.45, 95% CI: 0.00 - 7.18, p=0.0007). (d) ROC analysis for the detection of

adenomas shows an AUC value of 0.74 (95% CI: 0.6 to 0.88) with 70% sensitivity and

84% specificity (likelihood ratio= 4.37). (e) Significantly higher expression levels for

carcinoma-only group in comparison with controls (median difference =10.72, 95% CI:

10.00 - 11.88, p<0.0001). (f) ROC analysis for the detection of carcinomas shows an AUC

value of 0.87 (95% CI: 0.77 - 0.97) with 90% sensitivity and 84% specificity (likelihood

ratio  = 5.64).

139

Table 32: Summary of detection rates of adenomas and carcinomas based on CT values of miR-135b.

Sensitivity, specificity, 95% confidence intervals and likelihood ratios are given with

cut-off values.

Cutoff Sensitivity 95% CI Specificity 95 %CI Likelihood ratio

< 25.86 7.042 2.326 to 15.67 96 79.65 to 99.90 1.761

< 26.01 8.451 3.165 to 17.49 96 79.65 to 99.90 2.113

< 26.28 9.859 4.057 to 19.26 96 79.65 to 99.90 2.465

< 26.50 11.27 4.992 to 21.00 96 79.65 to 99.90 2.817

< 26.96 18.31 10.13 to 29.27 96 79.65 to 99.90 4.577

< 27.14 19.72 11.22 to 30.86 96 79.65 to 99.90 4.93

< 27.20 21.13 12.33 to 32.44 96 79.65 to 99.90 5.282

< 27.27 22.54 13.46 to 34.00 96 79.65 to 99.90 5.634

< 27.77 29.58 19.33 to 41.59 96 79.65 to 99.90 7.394

< 27.96 29.58 19.33 to 41.59 92 73.97 to 99.02 3.697

< 28.15 30.99 20.54 to 43.08 92 73.97 to 99.02 3.873

< 28.94 39.44 28.03 to 51.75 92 73.97 to 99.02 4.93

< 29.07 40.85 29.32 to 53.16 92 73.97 to 99.02 5.106

< 29.17 42.25 30.61 to 54.56 92 73.97 to 99.02 5.282

< 29.23 42.25 30.61 to 54.56 88 68.78 to 97.45 3.521

< 29.38 43.66 31.92 to 55.95 88 68.78 to 97.45 3.638

< 29.56 45.07 33.23 to 57.34 88 68.78 to 97.45 3.756

< 29.72 47.89 35.88 to 60.08 88 68.78 to 97.45 3.991

< 29.82 49.3 37.22 to 61.44 88 68.78 to 97.45 4.108

< 29.94 50.7 38.56 to 62.78 88 68.78 to 97.45 4.225

< 29.98 52.11 39.92 to 64.12 88 68.78 to 97.45 4.343

< 30.08 53.52 41.29 to 65.45 88 68.78 to 97.45 4.46

< 30.18 54.93 42.66 to 66.77 88 68.78 to 97.45 4.577

< 30.28 56.34 44.05 to 68.09 88 68.78 to 97.45 4.695

< 30.55 56.34 44.05 to 68.09 84 63.92 to 95.46 3.521

< 30.73 57.75 45.44 to 69.39 84 63.92 to 95.46 3.609

< 30.96 60.56 48.25 to 71.97 84 63.92 to 95.46 3.785

< 31.66 64.79 52.54 to 75.76 84 63.92 to 95.46 4.049

< 32.87 67.61 55.45 to 78.24 84 63.92 to 95.46 4.225

< 33.36 70.42 58.41 to 80.67 84 63.92 to 95.46 4.401

< 34.00 71.83 59.90 to 81.87 84 63.92 to 95.46 4.489

< 34.55 73.24 61.41 to 83.06 84 63.92 to 95.46 4.577

< 34.75 74.65 62.92 to 84.23 84 63.92 to 95.46 4.665

< 35.84 77.46 66.00 to 86.54 84 63.92 to 95.46 4.842

< 36.23 78.87 67.56 to 87.67 84 63.92 to 95.46 4.93

< 36.64 80.28 69.14 to 88.78 84 63.92 to 95.46 5.018

< 38.51 81.69 70.73 to 89.87 84 63.92 to 95.46 5.106

140

0 5 0 1 0 0

0

5 0

1 0 0

N o rm a l v s A d e n o m a + C a rc in o m a

1 0 0 % - S p e c if ic ity %

Se

ns

itiv

ity

%

0 5 0 1 0 0

0

5 0

1 0 0

N o rm a l v s . A d e n o m a

1 0 0 % - S p e c if ic ity %

Se

ns

itiv

ity

%

0 5 0 1 0 0

0

5 0

1 0 0

N o rm a l v s . C a rc in o m a

1 0 0 % - S p e c if ic ity %

Se

ns

itiv

ity

%

N o rm a l A d e n o m a + c a rc in o m a

-5

0

5

1 0

1 5

N o rm a l v s A d e n o m a + C a rc in o m a

C a s e s

De

lta

CT

va

lue

s

N o rm a l A d e n o m a

-5

0

5

1 0

1 5

N o rm a l v s . A d e n o m a

C a s e s

De

lta

CT

va

lue

s

a

b

c

d

e

f

N o rm a l C a rc in o m a

-5

0

5

1 0

1 5

N o rm a l v s . C a rc in o m a

C a s e s

De

lta

CT

va

lue

s

Id e n t ity %S e n s itiv ity %

Figure 34: ∆CT-based expression analysis of plasma miR-135b detected in an

independent cohort.

(a) Scatter plot shows significantly higher expression levels were detected for a

combined group of patients with adenomas and carcinomas (median difference = 7.34,

95% CI: 2.99 - 7.93, p=0.0008) in comparison to controls. (b) ROC analysis for the

detection of both adenoma and carcinoma shows an AUC value of 0.72 (95% CI: 0.60 -

0.85) with 80% sensitivity and 84% specificity (likelihood ratio = 4.01). (c) No statistically

significant differences in expression levels for adenomas and controls (median

difference = 3.62, 95% CI: 0.65 - 5.175, p=0.0007). (d) ROC analysis for the detection of

adenomas shows an AUC value of 0.62 (95% CI: 0.46 - 0.78) with 66% sensitivity and

80% specificity (likelihood ratio = 3.33). (e) Significantly higher expression levels for

carcinoma-only group in comparison with controls (median difference = 8.55, 95% CI:

6.68 - 8.93, p<0.0001). (f) ROC analysis for the detection of carcinoma shows an AUC

value of 0.80 (95% CI: 0.67 - 0.93) with 87% sensitivity and 84% specificity (likelihood

ratio = 5.48).

141

Table 33: Summary of detection rates of adenoma and carcinoma based on ΔCT values of miR-135b.

Sensitivity, specificity, 95% confidence intervals and likelihood ratios are given with

cut-off values.

Cutoff Sensitivity 95% CI Specificity 95% CI Likelihood ratio

< -2.930 1.408 0.0356 to 7.599 96 79.65 to 99.90 0.3521

< -2.690 2.817 0.3430 to 9.808 92 73.97 to 99.02 0.3521

< -2.065 7.042 2.326 to 15.67 88 68.78 to 97.45 0.5869

< -1.680 9.859 4.057 to 19.26 88 68.78 to 97.45 0.8216

< -1.405 11.27 4.992 to 21.00 88 68.78 to 97.45 0.939

< -1.045 19.72 11.22 to 30.86 88 68.78 to 97.45 1.643

< -0.980 22.54 13.46 to 34.00 88 68.78 to 97.45 1.878

< -0.515 30.99 20.54 to 43.08 88 68.78 to 97.45 2.582

< -0.480 33.8 23.00 to 46.01 88 68.78 to 97.45 2.817

< -0.290 35.21 24.24 to 47.46 84 63.92 to 95.46 2.201

< 0.125 40.85 29.32 to 53.16 84 63.92 to 95.46 2.553

< 0.310 43.66 31.92 to 55.95 84 63.92 to 95.46 2.729

< 0.660 49.3 37.22 to 61.44 84 63.92 to 95.46 3.081

< 0.915 50.7 38.56 to 62.78 84 63.92 to 95.46 3.169

< 1.020 53.52 41.29 to 65.45 84 63.92 to 95.46 3.345

< 1.245 56.34 44.05 to 68.09 84 63.92 to 95.46 3.521

< 1.970 60.56 48.25 to 71.97 84 63.92 to 95.46 3.785

< 3.525 67.61 55.45 to 78.24 84 63.92 to 95.46 4.225

< 3.975 70.42 58.41 to 80.67 84 63.92 to 95.46 4.401

< 5.210 74.65 62.92 to 84.23 84 63.92 to 95.46 4.665

< 5.710 76.06 64.46 to 85.39 84 63.92 to 95.46 4.754

< 6.600 78.87 67.56 to 87.67 80 59.30 to 93.17 3.944

< 6.845 80.28 69.14 to 88.78 80 59.30 to 93.17 4.014

< 6.915 80.28 69.14 to 88.78 76 54.87 to 90.64 3.345

< 7.215 80.28 69.14 to 88.78 68 46.50 to 85.05 2.509

< 7.305 80.28 69.14 to 88.78 64 42.52 to 82.03 2.23

< 7.595 80.28 69.14 to 88.78 60 38.67 to 78.87 2.007

< 7.840 80.28 69.14 to 88.78 56 34.93 to 75.60 1.825

< 8.075 81.69 70.73 to 89.87 52 31.31 to 72.20 1.702

< 8.280 81.69 70.73 to 89.87 48 27.80 to 68.69 1.571

< 8.400 83.1 72.34 to 90.95 44 24.40 to 65.07 1.484

< 8.415 83.1 72.34 to 90.95 36 17.97 to 57.48 1.298

< 8.645 85.92 75.62 to 93.03 28 12.07 to 49.39 1.193

< 8.685 87.32 77.30 to 94.04 28 12.07 to 49.39 1.213

< 9.450 88.73 79.00 to 95.01 20 6.831 to 40.70 1.109

< 9.650 90.14 80.74 to 95.94 16 4.538 to 36.08 1.073

< 9.915 91.55 82.51 to 96.84 12 2.547 to 31.22 1.04

142

In this feasibility study miRNA expression profiles of more than 700 mature miRNAs in

blood samples collected from symptomatic patients (n = 32) diagnosed with CRC,

adenoma or neoplasia-free healthy controls were established. In this discovery phase,

a panel of 29 miRNAs was selected and applied to a cohort of symptomatic patients (n

= 94) undergoing surgical or endoscopic resection of CRC or adenoma compared with

colorectal neoplasia-free controls. For a set of three miRNAs (miR-34a, miR-135b and

miR-431), ROC analysis showed 91% sensitivity and 88% specificity for the detection of

adenomas and carcinomas. In this set, miR-135b had superior detection rates for

adenomas and carcinomas (Figure 28), with 74% sensitivity and 88% specificity to detect

both adenomas and carcinomas. Hence, miR-135b was validated further on an

independent cohort of patients (n = 96) with adenomas, carcinomas and asymptomatic

controls undergoing colonoscopy after a positive FOBT or healthy controls with no

bowel symptoms. For this cohort, ROC analysis showed 80% sensitivity and 84%

specificity in detection of adenomas and carcinomas. The detection rates were far

superior for carcinomas, with 90% sensitivity and 84% specificity being achieved. The

diagnostic accuracy of this method should undergo further evaluation in a larger cohort

of asymptomatic patients before introducing it as novel biomarker for CRC screening.

143

3.4 Discussion

This study has shown that miR-135b is a potential novel biomarker for the early

detection of colorectal neoplasia. This is the first study that has shown the upregulation

of miR-135b in the blood of patients with adenomas and carcinomas. Its upregulation

in colorectal adenomas and carcinoma tissue has previously been reported in different

studies (Akao et al, 2006, Bandrés. et al, 2006, Rossi et al, 2007 & Nagel et al, 2008).

However, this is the first study evaluating the diagnostic utility of miR-135b in an

independent cohort of patients with colorectal neoplasia.

3.4.1 Role of miR-135b in colorectal neoplasia initiation and progression

As miR-135b is strongly linked to in CRCs, its use in the detection of adenomas and

carcinomas based on a blood assay is supported by its role in tumour initiation and

progression.Studies evaluating the role of miR-135b in tumour initiation have found

higher levels of miR-135b in adenoma and carcinoma tissue and have associated it with

the loss of APC expression (Nagel et al, 2008 & Aslam et al, 2015). It is well-established

that bi-allelic mutations in the APC gene are primary initiating events in the

transformation of adenoma to carcinoma through deregulation of the β-catenin

pathway (Fearon and Vogelstein, 1990, Segditsas and Tomlinson, 2006). These

mutations are typically the result of premature stop codons in the APC transcript and

are detected in more than two thirds of CRC tissues (Nagel et al, 2008). Elevated

expression of miR-135b with loss of APC suggests that miR-135b is a direct target of APC

(Nagel et al, 2008, Maragkakis et al, 2009 & Aslam et al, 2015). It is well-established that

APC loss and associated lack of β-catenin degradation results in free cytoplasmic β-

catenin that enters the nucleus and activates Wnt-regulated genes (Segditsas and

Tomlinson, 2006). This Wnt activation results in interaction with the T-cell factor (TCF)

family of transcription factors and a concomitant recruitment of coactivators (Survivin,

c-Myc and Cyclin D1). Consequently, there is lack of apoptosis and increased

proliferation of abnormal cells resulting in autonomous growth and the formation of an

adenoma. Association of miR-135b with APC is further supported by a confirmation of

a miR-135b binding site at the APC 3′ untranslated region. This suggests that miR-135b

upregulation leads to a more penetrant effect of APC derangement (Nagel et al, 2008)

and causes an increase in the size of the adenoma.

144

In the last decade, miR-135b has generated interest in multiple functional studies due

to its role in epigenetics. miR-135b is located on 1q32.1, which shows a DNA copy

number gain in CRCs (Nagel et al, 2008). miR-135b is encoded in the first intron of

LEMD1 gene that is aberrantly expressed in colonic cancers (Yuki et al, 2004). In vitro

investigations have shown a demethylation-associated increase in expressions of

LEMD1 and miR-135b, suggesting there is an epigenetic control of miR-135b in cancer

cells. Furthermore, investigators have predicted that the transcription factor NF-κB has

a potential binding site on the miR-135b promoter region. After demethylation, TNFα,

a known potentiator of NF-κB transcription, increased expression of miR-135b,

suggesting a potential link between inflammatory signals and miR-135b transcription

(Lin et al, 2013). miR-135b upregulation may be the common initiating pathway

between malignancy derived from somatic mutations and those derived from

inflammatory backgrounds where Wnt derangement is typically a later event. miR-135b

expressions are elevated in tumours generated in mouse models of APC mutations as

well as inflammatory bowel disease (Necela et al, 2013). Furthermore, the expression

of miR-135b is inversely correlated with serum E2 level and ER-β mRNA expression in

cancer tissues of CRC patients (He et al, 2012).

In a recently published study, there are significantly higher expressions of miR-135b in

tumours with mutant K-ras gene expression. These findings might suggest that in cases

where somatic mutations are present, miR-135b upregulation may suppress any

residual β-catenin regulation-driven tumour suppressive activity. This would explain a

progressive increase in miR-135b-driven proliferation in adenomas, where an early

event of miR-135b upregulation by loss of APC is further amplified by a second hit

mutation, leading to further proliferation in adenomas with dysplasia (Nagel et al,

2008).

Studies also have shown that miR-135b expression are controlled by STAT3 regulation

(Yu et al, 2009, Yue et al, 2009). STAT3 is an activator of anti-apoptotic pathways and

proliferative behaviours in different cancers (Corvinus et al, 2005 & Yu et al, 2009).

STAT3 is activated by epidermal growth factor receptor (EGFR) and consequently

multiple anti-EGFR drugs have been used to treat CRCs. It has been shown that higher

expression levels of miR-135b in CRCs are associated with increased activity of STAT3,

145

resulting in inhibition of tumour suppressors (Yu et al, 2009 & Yue et al, 2009). A screen

for miRNAs deregulated in CSCs has identified miR-135b as highly expressed in the stem

cell-enriched cell population (Lin et al, 2011 & Sanchez-Diaz et al, 2013). Mechanistic

studies have shown implications for miR-135b being an important driver in maintaining

cell oncogenicity in the self-renewing behaviours of CSCs (Lin et al, 2011, Sanchez-Diaz

et al, 2013).

3.4.2 Specificity of miR-135b for CRCs

miR-135b is not specific to CRC. It is upregulated in other cancers such as lung, breast,

liver, bone and blood cancer, and in vitro studies have confirmed its role in cancer cell

proliferation, invasion, migration, metastasis and drug resistance (Li et al, 2014, Pei et

al, 2015, Umezu et al, 2014, Xio et al, 2014 & Jung et al, 2014) ( Table 34).

Studies Cancer type Significant findings Target Mechanism

Li et al, 2014

Hepatocellular HSF1/miR-135b/RECK&EVI5 axis

HSF1, RECK, EVI5

Cell migration and invasion in vitro, metastasis in vivo

Pei et al, 2015

Osteosarcoma Upregulation of miR-135b in osteosarcoma tissue

FOXO1 Proliferation and invasion in vitro

Umezu et al, 2014

Multiple myeloma

Upregulation of miR-135b in exosomes from hypoxia resistant multiple myeloma cell line

FIH-1 Promotes angiogenesis

Xio et al, 2014

Glioblastoma Multiforme

miR-135b-GSK3β axis in radio resistant human GBM cell line U87R

GSK3β Reversal of radio-resistance by restoration of GSK3β

Jung et al, 2014

Hepatocellular Gα12gep inhibits FOXO1 resulting in the inhibition of FOXO1 de novo synthesis by miR-135b

MDM2 Knockdown of JunB (or c-Jun) decreased miR-135b levels, thereby increasing FOXO1.

Arigoni, et al, 2013

Breast Cancer Lung Cancer

miR-135b coordinates progression of ErbB2-driven mammary carcinomas

MID1 MTCH2

Stemness and anchorage-independent growth in vitro

Table 34: Summary of in vitro studies investigating the role of miR-135b in non-CRCs.

146

Upregulation of miR-135b in non-CRCs raises the issue about the specificity of a miR-

135b-based blood assay for CRCs. miR-135b levels in circulation have been found to be

raised in circulation in patients with bladder cancer, endometrial cancer, in maternal

circulation of pregnant women (Chim et al, 2008, Scheffer et al, 2014, Tsukamoto et al,

2014). In this respect, the most significant study has used circulating miR-135b for the

early detection and follow-up of endometrioid endometrial carcinoma (Tsukamoto et

al, 2014). This study has reported AUC of 0.97 for the detection of endometrioid

endometrial carcinoma. If used as screening test for CRC, plasma miR-135b might yield

a significant number of false positive results leading to patients with inaccurate

modality of investigations. A recent study by Faltejskova and colleagues has questioned

the utility of circulating miR-135b in the detection of CRCs (Faltejskova et al, 2012). In

that study, total RNA extracted from the serum of 100 CRC patients and 30 healthy

controls was analysed by QRT-PCR to quantify miR-17-3p, miR-29a, miR-92a and miR-

135b. Comparative analysis of miR-16-based normalised expression of miR-17-3p, miR-

29a and miR-92a showed no significant differences for patient with cancers and healthy

controls. The levels of miR-135b in serum were too low to be quantified accurately in

that study. This did not dismiss the case of miR-135b as potential biomarker for CRCs.

This suggests that work done in this study has been associated with a superior methods

of RNA isolation, pre-amplification and detection of miR-135b.

3.4.3 miR-34a in circulation

miR-34 is a well-recognised tumour suppressor miRNA. Its expression has been found

to be downregulated in the one third of patients of CRC in comparison to healthy

controls. This downregulation of miR-34a expression has been linked to upregulation of

the E2F pathway and the downregulation of the p53/ p21 pathway (Tazawa et al, 2007).

miR-34a is a direct transcriptional target of p53; the promoter region of miR-34a

contains a canonical p53 binding site (Yamakuchi & Lowenstein et al, 2009). p53-driven

tumour suppression involves the coordinated activation of multiple transcriptional

targets including miR-34a, leading to inhibition of cell proliferation and activation of

apoptosis. In contrast to the above, miR-34a indirectly deacetylates p53 via activation

of SIRT1 and decreases p53 transcriptional activation. Activation of SIRT1 mediates the

survival of cells during periods of severe stress through the inhibition of apoptosis

147

(Tazawa et al, 2007). Perhaps, in response to stress in adjacent normal tissue of colonic

cancer there is higher production of miR-34a from adjacent normal tissue, resulting in

a high influx of miR-34a into circulation.

In a study by Eichelser and colleagues, higher levels of miR-34a were associated with

AUC of 0.63 for its utility in the of breast cancer (Eichelser et al, 2013). In another study,

serum levels of miR-34a were significantly elevated in patients with breast cancer

compared to healthy controls (Roth et al, 2010) and its high levels correlated with

advanced tumour stages. In contrast to these two studies, other studies have reported

significantly reduced levels of miR-34a in patients with lymphoma, breast and colon

cancers (Nugent et al, 2012 & Fang et al, 2012).

3.4.4 miR-431

miR-431 is another miRNA deemed significant in this study. Its role in CRC is not well

studied, but a recent publication investigating the identification of a relevant group of

miRNAs in cancer using fuzzy mutual information has identified miR-431 as a primary

miRNA linked to lung cancer, targeting 14 genes (Pal et al, 2015). Two other studies

(Tanaka et al, 2014, Tanaka et al, 2014) have explored the role of the miR-431 cancer

pathway in medulloblastoma and have suggested a strong link of miR-431 in MAPK

signalling pathway.

3.4.5 Other studies investigating role of miRNAs for CRC screening

Various studies have been conducted in the last decade to investigate the role of

circulating miRNAs for the detection of colorectal neoplasia (Table 35). These studies

used either whole plasma or total RNA extracted from a defined amount of plasma

samples collected from healthy controls and diseased patients. RT-PCR based detection

systems were applied to detect selected circulating miRNAs. The selection of miRNAs

was based either on results of plasma miRNA expression profiling experiments

performed on relatively small cohorts of healthy and diseased patients or highly

upregulated miRNAs in CRC tissues. Different studies have reported different sets of

miRNAs and varying degrees of diagnostic accuracy. Pu and colleagues have used whole

plasma and bypassed the total RNA extraction step in RT-PCR. They have reported

higher levels of miR-221 in the plasma samples of patients with CRCs and an AUC value

148

of 0.606 with 86% sensitivity and 41% specificity to detect CRCs (Pu et al, 2010). Ng and

colleagues have studied miR-17-3p and miR-92 and have reported 89% sensitivity and

70% specificity of miR-92 for the detection of CRCs (Ng et al, 2009). In another study, a

panel of 12 miRNAs including miR-29a and miR-92a has been found to be associated

with a combined AUC of 0.883 with 83% sensitivity and 84.7% specificity to detect both

adenomas and carcinomas (Huang et al, 2010). However, their detection rates were

significantly lower for adenomas only, with sensitivities of approximately 60% and

specificities of approximately 80%. A study by Cheng and colleagues has investigated

the diagnostic and staging potential of circulating miR-141. In addition to its sensitivity

of 66.7% and specificity of 80.8% to discriminate controls from CRC patients, the higher

levels of plasma miR-141 predicted poor survival and was an independent prognostic

factor for advanced CRC (Cheng et al, 2011). Other researchers used a panel of miRNAs

to improve the diagnostic accuracy of circulating miRNAs for CRC and adenoma

detection. Kanaan and colleagues have performed a study evaluating 380 plasma

miRNAs and found that a panel of 8 miRNAs (miR-15b, miR-17, miR-142-3p, miR-532-

3p, miR-195, miR-331, miR-532 and miR-652) was associated with 91% sensitivity and

57% specificity to detect carcinomas and 88% sensitivity and 64% specificity to detect

adenomas.

By following the strategy of using miRNA panels, a panel of six miRNAs (miR-18a, miR-

19a, miR-19b, miR-15b, miR-29a and miR-335) was applied to a cohort twice the size of

cohort used by Kanann and colleagues (Giraldez et al, 2013). The results of this panel

have been very encouraging, with sensitivity and specificity for the detection of

adenomas and carcinomas reported as approximately 80%. Giraldez and colleague also

found significantly upregulated levels of miR-18a in patients with advanced adenomas.

In a recent study (Wang et al. 2014), researchers compared the diagnostic accuracy of

a panel of serum-based miRNAs with CEA and CA19-9. Researchers reported

sensitivities of 23, 35 and 93% for CEA, CA19.9 and the evaluated panel of miRNAs. The

specificity of detecting CRCs for panel was 91%.

149

Tissue type

Studies Participant number

Target miRNAs Diagnostic accuracy

Sensitivity% Specificity%

Whole plasma

Pu et al, 2010

CRC (103) Controls (37)

miR-221 86 41

Plasma /serum RNA

Ng et al, 2009

CRC (90) Controls (40)

miR-17-3p 64 70

miR-92 89 70

Huang et al, 2010

CRC (100) Adenomas* (37) Controls (59)

miR-29 69 62.2*

89.1 84.7*

miR-92a 84 64.9*

71 81.4*

Cheng et al, 2011

CRC I-IV (102) Controls (48)

miR-141 66.7 80.8

Kanaan et al, 2012

CRC (50) Controls (50)

miR-21 90 90

Kanaan, al, 2013

CRC (45) Adenoma* (16) Control (26)

miRNA panel: miR-15b, miR-17, miR-142, miR-195, miR-331,miR-532,

91 88*

57 64*

Giráldez et al, 2013.

CRC (63) Adenoma* (60) Control (73)

miRNA panel: miR-15b, miR-18a, miR-19a, miR-19b, miR-29a,miR-335

78 80*

79 80*

Wang et al, 2014

CRC (83) Control (59)

miRNA panel: let-7g, miR-21, miR-31, miR-92a, miR-203

93 91

Table 35: Comparison of the sensitivity and specificity of different miRNAs for their utility as biomarkers for detection of adenocarcinoma and adenoma*.

RT-PCR based quantification of miRNAs has been the preferred method of study in

majority of the above mentioned studies.

Colonic epithelium contains the most dynamic population of human cells. Highly

differentiated colonocytes are continuously shed into the colon of healthy individuals

and patients with CRC (Brittan et al, 2004 & Loktionov et al, 2007). Exfoliated

colonocytes from the healthy colon and neoplastic lesions carry important genetic and

epigenetic information that could be utilized for subsequent testing, such as detection

of mutant genes or dysregulated mRNAs, proteins and miRNAs (Loktionov et al, 2009).

It is proposed that even small neoplastic loci can alter the rate of colonic cell exfoliation

150

and may aid early detection of these lesions (Loktionov et al, 2007). The effectiveness

of an exfoliated colonocyte-based detection system requires efficient isolation of

colonocytes while minimizing the amount of background faecal debris. Strategies such

as density gradient centrifugation and/or immunoaffinity on either homogenized stool

samples or scrapings from the stool surface are used to try to maximise retrieval of

colonocytes (Loktionov et al, 2007). Nonetheless, cell yields are often still very low with

conspicuous background debris, making cell identification difficult and time consuming

(Deuter et al, 1995). Such preparations would be unsuitable for high-throughput

population screening programs (White et al, 2009). Furthermore, colonocytes shed

from a proximal colonic region travel further and are more exposed to cytolytic agents,

making them less likely to be preserved for sampling. If this does prove to be a common

problem, stool miRNA markers for right-sided CRC will be less effective. There is

evidence that this is indeed the case (Koga et al, 2010). In this study, immunomagnetic

beads were conjugated with EpCAM monoclonal antibodies to isolate colonocytes from

stool. Despite the selection of two highly upregulated miRNAs in CRC cells, the

sensitivity of detection was approximately 70% (Table 36).

However, the detection rate for left-sided colonic and rectal tumours was significantly

higher, suggesting the potential utility of exfoliated colonocyte-based miRNA assays as

an alternative to flexible sigmoidoscopy. Profound deregulation of apoptosis is known

to be a characteristic feature of cancer. As a result of apoptosis, tumour-specific

proteins and genetic information in the form of DNA, RNA and miRNA are released into

the colon lumen (Ahlquist et al, 2010). The stool environment is much more complex

and hostile than plasma, and human RNAs are rapidly degraded and constitute <1% of

total stool RNA (Ahlquist et al, 2010). In contrast with the rapid degradation of mRNA,

human miRNAs are packaged into microvesicles and are well protected from

degradation. The available data indicates that stool miRNA analysis can distinguish

between patients with adenomas and carcinomas from healthy controls (Link et al,

2010). The detection of miRNAs from stool specimens requires efficient protocols for

stool preparation, miRNA extraction and quantitative analysis (Ahmed et al, 2009). The

utility of stool miRNAs as biomarkers is still in its infancy; further studies of stool miRNA

are needed on larger cohorts to validate its diagnostic accuracy.

151

Table 36: Faecal miRNAs for CRC detection and screening.

RT-PCR based quantification of miRNAs has been the preferred method of study in

majority of the above studies.

3.4.6 Strengths and weaknesses of study

RNA concentration experiments suggest it is extremely difficult to enrich and

concentrate miRNAs due to their size. However, recent advances in RNA isolation and

purification technologies have made detection of small concentrations of miRNAs

possible. A comparative study of five commercial kits reported comparable quantities

of miRNA were extracted (Brunet-Vega et al, 2015). However, this did change

significantly with or without the spike of exogenous control RNA or miRNA. In this study,

concentration columns and the heat vacuum effectively concentrated large RNAs but

might have facilitated miRNA degradation during heating steps. Heating-associated

degradation has raised concerns about the optimum storage temperature for RNA

samples used in miRNA analysis. In light of this, total RNA should be stored at -80°C.

It is an intriguing finding that miR-135b was not sufficiently significant to allow

discrimination of both CRC and adenoma cases in the expression profiling analysis.

Interpenetration of the array data is complex and requires normalisation techniques,

and previous studies have focused on identifying novel miRNAs for the early detection

Tissue type Studies Participant

numbers

Target

miRNAs

Diagnostic accuracy

Sensitivity

%

Specificity %

Exfoliated

colonocytes

Koga et

al, 2010

CRC (197)

Control (119)

miR-17-92

69.5

81.5

miR-135 46.2

95

Faeces

Wu et

al, 2014

CRC (104)

Adenoma (169)

Control (109)

miR-135b 78 (CRC)

65

(Adenoma)

68

Link et

al, 2010

CRC (10)

Adenoma (9)

Control (10)

miR-21

miR-106

Distinguished adenoma and

carcinoma from healthy

controls P<0.05

152

of colorectal neoplasia. Because miR-135b is the most studied miRNA in CRC initiation

and progression, its role should have been thoroughly investigated in early detection of

CRCs. Until the role of a miRNAs is more established, the focus of biomarker studies

should be on miRNAs with a clear role in cancer pathway. However, this strategy might

hinder the inquisitive nature of clinical translational studies for biomarker

development, since this would prevent the pursuit and discovery of novel markers and

their potential roles.

An alternative diagnosis of patients with bowel symptoms rests on a wide spectrum of

benign diseases, varying from minor haemorrhoids to life-impairing inflammatory

bowel diseases. It is important to detect the presence of significant benign diseases too

since their existence will eventually require confirmation by colonoscopy. This study

found discriminatory levels of circulating miRNAs in patients with significant benign

disease. However, future studies should assess a cohort of patients with minor and non-

significant bowel diseases such as haemorrhoids, non-complicated diverticular diseases

and irritable bowel syndrome. A miRNA-based blood test associated with false positive

cases for such a cohort would make the screening test unsuitable. Recently, a panel of

circulating miRNAs have been shown to significantly differentiate colonic adenomas

and carcinomas from haemorrhoids and other benign diseases (Verma et al, 2015).

Another limitation of this study is the characterisation of adenomas based on the size

and grade of adenomas. It is probably difficult to detect small, low grade adenomas on

blood tests, and the natural history of colonic neoplasia progression suggests that it

might take years for such adenomas to develop into cancers. Detection of diseases such

as complicated diverticular disease and inflammatory bowel disease is also of a

paramount significance. Patients with this spectrum of disease would usually require

further assessment and frequent treatments. miR-135b has also been found elevated

in the plasma of patients with ulcerative colitis, and a panel of miRNAs has been found

to have the potential to monitor the severity of ulcerative colitis and detect the

presence of colitis associated dysplasia (Patel et al, 2015).

Significantly, discriminatory levels of miR-135b, miR-34a and miR-431 in patients with

previous colorectal neoplasia might be attributable to the presence of cell-free cancer-

153

specific miRNAs present in circulation years after cancer resection in the absence of any

detectable cancer disease. It is not surprising to find cancer-related genetic information

in the blood circulation years after the treatment of primary cancer with no evidence

of disease recurrence. Genomic analysis of circulating free DNA (cfDNA) has detected a

specific genetic profile mirroring primary tumours up to 12 years after cancer resection,

despite no other evidence of cancer recurrence (Shaw et al, 2012). Another explanation

of high levels of miR-135b, miR-34a and miR-431 in patients with previous history of

colorectal neoplasia is field change in the rest of the colon, rendering such patients as

high-risk for the development of metachronous colorectal neoplasia. Patients who have

undergone CRC excision routinely undergo strict surveillance over an extended time

period to monitor the development of further neoplasia. Circulating miRNAs with the

potential to discriminate patients with and without the presence of metachronous

neoplasia may be of potential use as surveillance markers.

Numerous variables affect the analytical measurements of circulating miRNAs.

However, for the purpose of the blood assay, one advantage in the utility of miR-135b

was its ability to detect the presence of adenomas and carcinomas based on the

presence or absence of detectable miR-135b in the blood samples used for analysis.

Diagnostic miRNAs identified in other studies based their analysis on a differential

change in expression levels in diseased from healthy controls. Such an analysis makes

the experimental controls very difficult to develop and standardise. A miRNA blood

assay based on the presence or absence of detectable miRNA in circulation will remove

additional steps of normalisation with endogenous/exogenous controls, and will allow

easy interpretation of the findings.

3.4.7 Clinical application of plasma miRNA based detection of colorectal neoplasia

The accuracy of circulating miRNA-based detection is far superior to stool-based

detection modalities and may be comparable with endoscopic modalities if conducted

with a panel of miRNAs. Furthermore, the ability of circulating miRNAs to detect

adenomas highlights the potential utility of circulating miRNAs in bowel cancer

screening. Therefore, in addition to a stand-alone blood test for the detection of

symptomatic CRCs presenting through outpatient clinics, miRNA-based blood assays

154

can be used as primary screening tools in bowel cancer screening programmes. With its

higher sensitivity and specificity, this may prove to be cost-effective and help reduce

the need for unnecessary colonic investigations. This study investigated the role of

circulating miRNAs for their utility in CRC screening, and hence it is essential to have an

update on current in use modalities for bowel cancer screening. This can provide insight

into the use of miRNA-based blood tests as stand-alone screening tests or an adjuvant

to existing modalities. A key requirement of a biomarker for early detection of CRC is

high sensitivity and specificity. Colonoscopies remain the gold standard for screening

but are associated with high costs and invasiveness, so has limited use as a primary

screening test in most countries, including the UK. Many studies have now successfully

investigated and reported the potential of miRNAs as diagnostic biomarkers for CRC

screening, but more work is required to produce a highly efficient, robust and

standardized method for miRNA detection in clinical practices. Until the development

of a non-invasive biomarker began to be investigated, studies have focused on the

effective use of currently available stool-based diagnostic modalities of FOBT or FITs. A

multi-target stool DNA test, which combines mutant and methylated DNA markers and

a faecal immunochemical test, recently performed favourably in a large cross-sectional

validation study and has been approved by the US Food and Drug Administration (FDA)

for the screening of asymptomatic, average-risk individuals (Dickinson et al, 2015). Such

stool or blood DNA-based testing are also being considered as screening modalities in

the UK.

Recently, 60% of UK screening centres have participated in the ‘Bowel Scope’

programme with single-stop flexible sigmoidoscopy for participants aged 55 to 64 years.

‘Bowel scope’ screening is an alternative and complementary bowel screening

methodology to FOBT. Evidence shows men and women aged 55 to 64 years attending

a one-off bowel scope screening test for bowel cancer can reduce cancer-related

mortality and incidence of CRC (Atkins et al, 2010 & Holme et al, 2014). Flexible

sigmoidoscopies do not require sedation and are a more simple procedure to perform

but provide no screening of lesions in the proximal colon. Because proximal advanced

adenomas can often exist in the absence of distal lesions, flexible sigmoidoscopies may

yield misleading test results and give patients a false sense of security. Perhaps when

155

used as an adjuvant to ‘Bowel Scope’ miRNA-based blood tests, this will not only

increase the detection rates for right-sided bowel lesions, but will also alleviate the risk

of missing any significant bowel pathology. Although individual miRNA markers have

shown promise in detecting both early adenomas and carcinomas, effective screening

tests will likely employ a panel of miRNAs to improve reliability and consistency.

However, this will likely be more expensive than measuring single miRNAs, and will

require further studies to ensure that the optimal panel is selected to maximize

sensitivity and specificity.

3.4.8 Conclusion

This study has met its aim: to identify circulating miRNAs that can be used for the

detection of colorectal neoplasia. Plasma miRNA expression profiles for different

participants were successfully developed and discriminatory miRNAs were applied to a

larger cohort to assess their diagnostic accuracy for the detection of adenomas and

carcinomas. miR-135b was identified as a plasma miRNA with superior diagnostic

accuracy and detectability for both adenomas and carcinomas. Megaplex™ RT and pre-

amplification primers pools have allowed the detection and quantification of multiple

miRNAs even when present at low levels. The selection of asymptomatic controls with

positive FOBT in NBCSP and symptomatic controls with benign bowel disease attending

endoscopy unit has ensured controls with wide spectrum of benign diseases were used.

Low detection rates for adenomas and an inability to discriminate benign from

neoplastic disease have been the limitations of this study. In addition, although

individual miRNAs such as miR-135b have shown promise in detecting both early

adenomas and carcinomas, effective bowel cancer screening tests based on blood

miRNAs will likely employ a panel of miRNAs to improve reliability and consistency. A

panel of miRNAs for bowel cancer screening will probably be more expensive than a

single miRNA, and will require further studies to ensure the optimal panel is selected,

and sensitivity and specificity are maximised. An additional validation study in

conjunction with national bowel cancer screening programme should be designed and

conducted to validate the diagnostic accuracy of selected miRNAs in larger cohorts

before miRNA-based blood tests are incorporated into cancer screening programmes.

156

Chapter 4: Results

Analysis of exosomal miRNAs for the

detection of colorectal neoplasia

157

4 Results: Analysis of exosomal miRNAs

4.1 Summary of Results

Previous studies have shown that miRNAs extracted from circulating exosomes of

patients with non-gastrointestinal cancers are similar to those extracted from tissue

biopsies. Using a similar approach, exosomal miRNAs in CRC can be evaluated for their

diagnostic accuracy and may provide a breakthrough for diagnostic modality. The aim

of this study was to examine the feasibility of using plasma exosomal miRNAs for the

detection of CRCs. The objective was to evaluate different techniques to isolate and

analyse exosomal miRNAs from the plasma of patients with CRCs. Ultracentrifugation

and immunoaffinity isolation techniques were assessed for the isolation of exosomes

from the plasma of patients with CRCs, healthy controls, and from media harvested

from CRC cell line cultures. CD133, CD326 and GPA33 antibodies were assessed for their

suitability for antibody-coupled Dynabead-based immunoaffinity isolation of plasma

exosomes. Size distribution of exosomes was assessed by transmission electron

microscopy, dynamic light scattering and flow cytometery. miRNA expression levels in

total RNA extracted from exosomes was analysed by Megaplex™ QRT-PCR. Selected

miRNAs were analysed and their expression levels were compared between healthy

controls and patients with adenomas and carcinomas. miRNA expression analysis for

exosomes isolated by ultracentrifugation showed an enrichment of miRNAs (miR-21,

miR-192*, miR-369, miR-502-5P and miR-589 miRNAs) in exosomes representing the

miRNA patterns of their source sample. Analysis of miRNAs (miR-21, miR-31, miR-135b,

miR-192*, miR-502) in CD326 immunoaffinity-isolated plasma exosomes showed

detectable expression levels of miR-135b in 83% of cases with adenomas and

carcinomas. 50% of cases with adenomas and carcinomas showed detectable

expression levels for miR-135b in plasma exosomes isolated via the GPA33

immunoaffinity method. Intergroup comparison of expression levels of miR-21 in

plasma exosomes isolated by CD326 immunoaffinity showed significantly higher

expressions in the combined adenoma and carcinoma group in comparison with healthy

controls. ROC analysis of CD326 immunoaffinity-isolated exosomal miR-21 showed an

AUC value of 0.97 (95% CI= 0.87 - 1.07; p=0.028) for the detection of adenoma and

carcinoma. A similar trend of higher expression levels of miR-31 in CD133

158

immunoaffinity-isolated exosomes from patients with adenoma and carcinoma in

comparison to healthy controls was observed. Thus, this study successfully developed

a protocol for the isolation of circulating exosomes by immunoaffinity, and also

successfully analysed exosomal miRNAs in plasma from patients with CRCs.

159

4.2 Identification of exosomes on Transmission Electron Microscopy

Exosomes isolated by a combination method of filtration and ultracentrifugation of

HT29 colorectal cancer cell line culture harvested media, and plasma from a patient

with colorectal cancer (H276/09) and a healthy control (H400/08), were evaluated with

Transmission Electron Microscopy (Figure 35). The particle size varied from 40-80 nm.

However, a significant number of particles ranging from 81-100 nm were also seen on

TME. As a 100 nm filter was used to filter plasma and harvested media, the

ultracentrifugation technique of exosome precipitation precipitated all microvesicles

less than 100 nm.

Figure 35: Electron micrographs of exosomes isolated from plasma by combination of filtration (100 nm) and ultracentrifugation without sucrose gradient.

The TME figure shows a typical morphology of multiple particles sizes (40-100

nm).Particles of 40-60 nm are non-immune stained exosomes fixed with 2%

gluteraldehyde, as seen on transmission electron microscopy.

160

4.3 Dynamic Light Scattering to assess the size of exosomes isolated by

ultracentrifugation

Exosomes from two plasma samples (H400/08 & H276/09) and HT29 colorectal cell line

culture harvested media were diluted with PBS solution. PBS was used as control

solution with a viscosity of 1.05cP, temperature of 20°C & refraction index of 1.33. All

the measurements for the buffer and samples were taken at room temperature of 37°C.

The plot of size distribution to volume of sample showed a peak of particle size at 21.4

nm and made up 20.8% of total particles (Figure 36). For PBS, this made up to 80% of

same-sized particles. The fraction of particles ranging from 40-60 nm was nearly

undetectable in control PBS. Most DLS measurements were carried out at 20°C using

Malvern© Zeta Nanosizer software. This instrument operates at 4 mW He-Ne laser

power, scattering angle of 173° and wavelength of 633 nm.

Figure 36: The size distribution of exosomes (b) as measured on Malvern© Zetasizer

Nano (a) by Dynamic Light Scattering. The Zetasizer Nano software reported the size

and volume distribution of nanoparticles in a sample. The peak intensity was seen at

21.4 nm and constituted 20.8% of the particles representing nonspecific particles.

161

4.4 Flow cytometry (FACS) for detection of exosomes isolated by

ultracentrifugation

Exosomes isolated from the plasma of a patient with CRC (H276/09) and a colonoscopy

negative healthy control (H400/08) were only detectable on coupling of vesicles with

fluorescence conjugated surface antibodies for exosomes surface ligands. One of the

limitations of FACS was an inability to detect unstained particles smaller than 300 nm.

The FACS showed no detectability for unstained and uncoupled exosomes. There was a

4.6% detection rate for fluorescent antibody-coupled exosomes, whereas no events

were detectable for unstained exosomes preparations (Figure 37).

162

Figure 37: A comparison of events detected for exosomes coupled to fluorescence

stained exosomes surface antibodies (CD44 & CD326). The FACS showed no

detectability for unstained and uncoupled exosomes. 10 µl of exosome preparation

from H400/08 and H276/09 was diluted with 90µl of PBS containing with CD44–PE

conjugated antibody and CD326 (EpCAM)–APC conjugated antibody. One of the

limitations of FACS was an inability to detect unstained particles smaller than 300 nm.

163

4.5 Comparison of total RNA and miRNA concentrations for different volumes of

plasma used for ultracentrifugation

Total RNA concentration for exosomes isolated by ultracentrifugation of plasma

samples was low for H400/08 and H276/09. A significant enrichment of RNA extracted

from cell line harvested media was noticed due to large volume of harvested media

directly used for exosome isolation (Table 37).

Sample Volume of

specimen

RNA concentration

(ng/µl)

260/280

H400/08 – Healthy control 1 ml of Plasma 1.53 0.32

H276/09 - Cancer 1 ml of Plasma 2.32 0.29

HT29 harvested media 10 ml 10.5 1.44

HT29 culture cells 1 ml of stored

lysate

908 1.65

Table 37: Total RNA concentration for exosomes from plasma and HT29 harvested media.

Table shows exosomal RNA concentrations extracted from 1 ml plasma samples from

healthy controls and patient with CRC. Table also shows RNA concentrations for HT29

cell line culture harvested media and RNA concentration for HT29 lysate.

164

A comparison was made between the total concentration of RNA extracted from 1 and

3 ml plasma and the corresponding concentration of miR-21 and snRNA RNU6B. The

concentration of RNA extracted from plasma increased once plasma volume was

increased to 3mls. (Figure 38) An approximate 2 to 4 fold increase in expression levels

of miR-21 was found when the volume of plasma was increased from 1 to 3 ml (Figure

39).

Figure 38: Comparison of total RNA concentration for volume of plasma used to

isolate exosomes isolated by filtration with 100 nm filter and ultracentrifugation. The

concentration of total RNA increased with the increase in volume (H274/09: 3.88 to

9.44 ng/µl, H287/09: 4.6 to 12.4 ng/µl).

0

2

4

6

8

10

12

14

H274/09 H287/09Tota

l R

NA

con

cen

tra

tio

n n

g/µ

l

Volume of plasma used for isolation of microvesicles

Total RNA concentration for different volumes of Plasma used

for Ultracentrifugation

RNA from 1ml of Plasma RNA from 3ml of Plasma

165

Figure 39: Comparison of miRNA expression levels for different volumes of plasma

used to isolate exosomes isolated by filtration with 100 nm filter and

ultracentrifugation. An approximate 2 to 4 fold increase in expression levels of miR-21

was found when the volume of plasma was increased from 1 to 3 ml.

4.6 Comparison of exosomal miRNAs with source miRNAs

Plasma exosomes were isolated from 3 ml of plasma sample collected from a patient

with colorectal cancer. RNA extracted from exosomes isolated from 10mls of harvested

media and 1 ml lysate of cell line lysate were used for comparisons. Megaplex QRT-PCR

protocol was used for the quantification of expressions (CT) of miR-21, miR-192*, miR-

369, miR-502-5P and miR-589 miRNAs. SnRNA RNU6B was used as a control gene for

the relative expression (∆CT) of each miRNA. CT values based expression analysis

showed miR-21, miR-192*, miR-502-5P were detectable in HT29 cell line lysate, HT 29

exosomes, plasma samples and exosomes isolated from plasma samples. miR-369 was

not detectable in any of the samples and miR-589 was not detectable in plasma

exosomes and plasma despite its enrichment in cell lines exosomes. Though higher

levels of expression for miR-21(∆CT = -4) were identified from 3 ml plasma, the protocol

for exosome isolation was long and was inconvenient. Figure 40 and 41 show CT & ∆CT

based expression analysis of different miRNAs studied for different samples.

15

17

19

21

23

25

27

29

31

H274/09 - miR-21 H287/09 - miR-21 H274/09 - SnRNA

RNU6B

H287/09 - SnRNA

RNU6B

Exp

ress

ion

lev

els

(CT

) v

alu

es o

n Q

RT

-PC

R

miR-21 and SnRNA RNU6B for 1ml and 3mls of plasma

miRNA expression levels (CT) for different volumes of plasma

RNA from 1ml of Plasma RNA from 3ml of Plasma

166

Figure 40: Comparison of miRNA expression for plasma exosomes, whole plasma,

HT29 cell line harvested exosomes and HT29 cell lysate.

Bar graph shows expression levels of miR-21, miR-192*, miR-369, miR-502-5P, miR-589

and snRNA RNU6B.

Figure 41: Relative expression levels (∆CT) of miRNAs isolated from plasma and HT29

cell line culture.

Bar graph shows SnRNA RNU6B normalised expressions (∆CT) for miR-21, miR-192* and

miR-502-5P.

0

5

10

15

20

25

30

35

40

45

miR-21 SnRNA

RNU6B

miR-192* miR-369 miR-502-5P miR-589

Exp

ress

ion

Lev

els

of

miR

NA

s (C

T) V

alu

es

Megaplex QRT-PCR relative quantification of selected miRNAS

Comparison of exosomal miRNAs with Source of Isolation

H274/09 - Exosomes H274/09 - Whole Plasma HT29 - Exosomes HT29 - Cell Lysate

-5

0

5

10

15

20

25

30

miR-21 miR-192* miR-502-5P

Sn

RN

A R

NU

6B

ba

sed

rela

tiv

e ex

pre

ssio

ns

(∆CT)

Expression levels for miRNAs in exosomes and their isolation source

Relative Expression levels of Exosomal miRNAs

H274/09 - Exosomes H274/09 - Whole Plasma

HT29 - Exosomes HT29 - Cell Lysate

167

4.7 Assessment of CD133 and CD326 (EpCAM/ESA) antibody coupling with beads by

FACS

Beads coated with fluorescently labelled antibodies specific for CD133 and CD326, and

unstained beads were analysed to assess the coupling of beads with antibody by the

locally available FACS facility in the department of Cancer Studies and Molecular

Medicine. For unstained beads, no fluorescence was detected for events counted

through FACS (Figure 42). For antibody coupled fluorescence conjugated CD133-PE

(65%) and CD326-APC beads, 43% of total events were associated with detection of

specific fluorescence. This suggested that approximately 50% beads were coupled with

antibody (Figure 43 and Figure 44).

168

Figure 42: FACS analysis of unstained beads. No fluorescence activity was detected by

FACS on P3 gate. The number and size of the particles were identical to other

measurements for fluorescence conjugated antibody coupled beads.

169

Figure 43: FACS analysis for CD326 (EpCAM/ESA)-APC coupled Dynabeads.

FACS analysis show 43% of beads were associated with fluorescence, suggesting

effective coupling with antibody.

170

Figure 44: FACS analysis for CD326 (EpCAM/ESA)-APC coupled Dynabeads.

FACS analysis show 65.5% of the beads were associated with fluorescence, suggesting

effective coupling with antibody.

171

4.8 Isolation of exosomes with antibody coupled Dynabeads with FACS

Plasma exosomes were isolated from 4 samples by immune-coupling of exosomes with

antibody coupled beads. The fluorescence conjugated to these antibodies was used to

isolate exosomes attached to antibody coupled beads. For FACS analysis, the number

of events representing bound exosomes with antibody coupled beads was compared

with events associated with control unbound antibody coupled beads. Figure 45 shows

the comparison between events for CD133-PE coupled beads bound to exosomes and

CD133-PE coupled beads only.

172

Figure 45: Comparison of FACS analysis for CD133-PE coupled beads bound exosomes

to control CD133-PE coupled beads stored in Aldefluor solution at 4°C.

For isolation of plasma exosomes from plasma (H59/09, H179/09, H260/10, H275/10

and H276/10), 1 ml of plasma was filtered through a 100 nm filter. 45 µl of ED133-PE

and CD326-APC antibody coupled beads was added to filtrate and incubated for 2 hours

at 4°C on a roller. Beads were washed using a magnet with PBS and analysed by FACS.

173

4.9 Analysis of exosomal miRNAs isolated through immune isolation and FACS

Plasma samples from 5 participants representing a wide spectrum of colorectal disease

were analysed for the detection of miR-21, miR-31, miR-135b, miR-192*, and miR-502.

Total RNA concentrations for CD133 and CD326 bead bound exosomes are shown in

table 38. QRT-PCR showed that miR-192* and miR-502 were not detected in both types

of exosomes isolated from all 5 samples (Table 39). miR-135b was only detected in

CD326 exosomes isolated from plasma of patients with Ulcerative colitis. miR-21 was

detected in all samples for both CD133 and CD326 bound exosomes. The differential

expression (∆CT) of miR-21 in exosomes isolated with CD326 bound exosomes could

differentiate between caecal cancer and rectal polyp cases from 3 other cases.

Expression levels of miR-21 in CD133 bound exosomes isolated from plasma of patients

with caecal cancer were 2 fold higher in comparison to other controls. A trend similar

to miR-21 was present for miR-31 and could differentiate caecal carcinoma from

control, but it failed to differentiate a rectal adenoma case.

Sample Specimen characteristic Total RNA concentration

CD133-exosomes CD326-exosomes

H59/09 Normal Colonoscopy 0.96 1.22

H179/09 Normal Colonoscopy 1.02 1.34

H260/10 Large Rectal Polyp 1.26 1.38

H276/10 Pan Ulcerative Colitis with

Dysplasia

1.48 1.58

H275/10 Caecal Carcinoma 1.09 1.22

Table 38: Sample characteristics and RNA concentrations for CD133 and CD326 bound exosomes isolated by FACS.

Total RNA extracted through a Tri-Reagent and mirVana column was quantified using a

spectrophotometer.

174

Sample miR-21 miR-31 miR-135b miR-192* miR-502 RNU6B

Exosomes Isolated with CD133-PE coupled Dynabeads

H59/09 36 (4) 40 (8) 40 40 40 32

H179/09 35 (3) 40 (8) 40 40 40 32

H260/10 34 (4) 40 (10) 40 40 40 30

H276/10 28 (4) 34 (10) 40 40 40 24

H275/10 32 (2) 36 (6) 40 40 40 30

Exosomes Isolated with CD326-APC coupled Dynabeads

H59/09 34 (3) 38 (7) 40 40 40 31

H179/09 35 (3) 40 (8) 40 40 40 32

H260/10 32 (1) 34 (3) 40 40 40 31

H276/10 27 (5) 31 (11) 34 40 40 22

H275/10 29 (1) 32 (1) 40 40 40 28

Table 39: Expression levels of different miRNAs for exosomes isolated with CD133 and CD326 bound Dynabead and FACS.

Table shows expression levels for each miRNA (CT) and snRNA RNU6B normalised

expression levels (∆CT) of miR-21 and miR-31. CT values of 40 means no detectable

expression levels for miR-135b, miR-192* and miR-502.

175

4.10 Direct immune isolation of exosomes by antibody coupled Dynabeads

HT29 colorectal cancer cell line culture harvested media and a plasma sample from a

patient with caecal cancer (H275/09) was used for direct immunoprecipitation of

exosomes using a bench magnet. Total RNA extracted from 2 samples for each type of

antibody bound beads is given in table 40, and Figure 46 shows expression levels of

miRNAs isolated with GPA33 antibody coupled beads.

Sample Type RNA concentration (ng/µl)

CD133 CD326 GPA33

HT29 Harvested Media 3.6 4.81 5.10

H275/10 1.12 2.29 1.78

Table 40: Comparative concentrations of RNA extracted from exosomes.

RNA concentration are shown for exosomes isolated through direct

immuneoprecipitation by CD133, CD326 and GPA33 antibody coupled Dynabeads.

Figure 46: Expression levels of exosomal miRNAs from plasma and cell line exosomes

isolated by immunoprecipitation with GPA33 coupled Dynabeads. Expression levels

(CT) of miR-135b, miR-31 miR-31 are plotted to show the differences of exosomes

isolated from HT29 cell line culture harvested media and plasma from a patient with

colonic cancer.

15

20

25

30

35

40

miR-21 miR-31 miR-135b SnRNA RNU6B

Exp

ress

ion

Lev

els

of

miR

NA

s (C

T)

Expression levels of miRNAs for plasma and Cell line harvested media

Exosomal miRNAs isolated with GPA33 coupled Dynaeneads

HT29 Harvested Media H275/10 Plasma

176

GPA33 coupled beads were assessed for their coupling efficiency, and the isolation of

exosomes was similar to the method used for the assessment of CD133-PE and CD326-

APC coupled beads. FACS assessment of antibody coupling was achieved by detecting

the fluorescence of anti-goat antigen conjugated with APC. Figure 47 shows the FACS

analysis for GPA33 coupled Dynabeads assessed for bound fluorescence and number of

events detected over 6 seconds.

177

Figure 47: FACS analysis for GPA33 coupled Dynabead bound exosomes isolated from

HT29 cell line culture harvested media.

GPA33 coupled beads were conjugated with Aldefluor-488 anti-goat antibody for FACS

analysis.

178

4.11 Analysis of plasma exosomal miRNAs from immunoprecipitated exosomes using

CD133, CD326 and GPA33 coupled Dynabeads

Plasma samples from nine participants (n = 3 controls, n = 3 adenomas and n = 3

carcinomas) were analysed for exosomal miRNAs isolated by immunoprecipitation

using antibody coupled Dynabeads. A total of 3 ml (1 ml of plasma for each set of

antibody coupled beads: CD133, CD326 and GPA33) was used for the direct isolation of

exosomes. RNA isolated was quantified by spectrophotometry. The concentrations of

total RNA isolated from each sample for different antibodies are shown in Table 41.

Sample

Number

Sample ID RNA concentration (ng/µl)

CD133 CD326 (EpCAM) GPA33

Control 1 H335/08 1.08 1.26 0.76

Control 2 H339/08 0.88 1.02 0.92

Control 3 H515/08 1.23 1.43 1.06

Adenoma 1 H334/08 0.92 0.84 0.84

Adenoma 2 H170/09 1.02 1.08 0.88

Adenoma 3 H273/10 1.25 1.56 1.32

Carcinoma 1 H660/10 1.65 1.52 1.43

Carcinoma 2 H265/10 0.82 1.34 1.12

Carcinoma 3 H268/10 1.13 1.32 1.04

Table 41: Total RNA concentrations of immunoaffinity isolated exosomes.

RNA concentrations for CD133, CD326 and GPA33 immunoaffinity isolated exosomes

are compared in the table.

miR-182 was not detectable for exosomes isolated with the 3 different types of

antibodies. miR-135b was not detected for all cases of exosomes isolated with CD133

antibody coupled beads (Figure 48). Exosomes isolated with CD326 and GPA33

coupled beads showed no expression of miR-135b in healthy controls. However, miR-

135B expression was detected for 5 out of 6 adenoma and carcinoma cases for CD326,

and 3 out of 6 adenoma and carcinoma cases for GPA33. Intergroup comparison of

controls and the adenoma/carcinoma combined group showed significantly higher

179

relative expressions of CD326 isoltaed exosomal miR-21 for the adenoma/carcinoma

group (2 tailed student t-test, p=0.03, 95% CI: 0.24 – 5.4). ROC analysis for miR-21

expressions showed AUC of 0.97 for adenoma and carcinoma (Figure 49).

Figure 48: RNU6B normalised expression levels for different miRNAs of exosomes

isolated by different types of antibody coupled Dynabeads.

The differential expression of miR-21 in CD326 isolated exosomes from controls and

groups of adenoma and carcinoma showed significantly higher expression levels (2

tailed student t-test, p=0.03, 95% CI: 0.24 – 5.4). A similar trend of miR-31 expression

was seen for exosomes isolated with CD133 coupled Dynabeads (2 tailed student t-test,

p=0.038, 95% CI: 0.11 – 3.2). No statistically significant differences in expression levels

(∆CT) of different miRNAs were seen for GPA33 bound exosomes. The expression of

miR-135b (CT) was detectable in exosomes isolated with GPA33 coupled beads from 2

cancer patients and 1 adenoma case.

180

Figure 49: Diagnostic utility of exosomal miR-21 isolated with CD326 coupled

Dynabeads. The box plot shows differences in mean expression levels for controls and

participants with adenoma and carcinoma. ROC analysis show AUC of 0.97 (95% CI=

0.87 to 1.07 and p=0.028).

181

4.12 Discussion

It is feasible to develop a protocol and detect exosomal miRNAs from plasma samples

of patients with colorectal neoplasia. The antibody coupled Dynabeads-based immune-

affinity isolation method is an efficient way of isolating exosomes circulating in the

blood of patients with adenomas and carcinomas. Other methods of filtration and

ultracentrifugation are time consuming, and the isolation of exosomes may not be

cancer-specific, hence fail to override the existing protocol of plasma miRNAs analysis

by direct extraction of RNA from whole plasma. A recent review (Sato-Kuwabara et al,

2015) on exosomal miRNAs in the literature has summarised exosomal miRNAs isolation

techniques and significant miRNAs for cell lines and body fluid related to different

cancers. Most of these reports summarised in this review were based on exosomal

isolation by ultracentrifugation for a variety of cancer cell lines including leukaemia,

melanoma, glioblastoma and cancers of solid organs including breast, lungs, prostate,

ovary and gastrointestinal tract. From the summary, it is clear that ultracentrifugation

is a preferred method of isolation of exosomes harvested from cell line culture.

Whereas, for body fluids, immunoaffinity isolation yields tumour-related miRNAs for

their utility in diagnosis of different cancers (Sato-Kuwabara et al, 2015).

4.12.1 Selection of miRNAs and antibodies for this feasibility study

This feasibility study for the analysis of exosomal miRNAs used a known panel of cancer

related miRNAs (miR-21, miR-31, miR-135b and miR-182). Expression levels of these

miRNAs were found to be upregulated in plasma samples of a significant number of

patients with colorectal neoplasia in our previous validation cohort using whole plasma

for total RNA isolation. The selection of CD326 (EpCAM) and CD133 antibody for

Dynabead coupling was based on their utility in isolation of cancer stem cells from

cancer tissue specimens taken at the time of resectional surgery. CD326 (EpCAM) and

CD133 surface markers remain 2 potent markers for CRC stem cells isolation (Fanali et

al, 2014) and CD326 (EpCAM) has recently been studied for its role in cancer initiation

and progression (Failli et al, 2014). A previous study has shown that the isolation of CRC

specific highly purified exosomes is possible by using GPA33 antibody for

immunoaffinity capture (Mathivanan et al, 2010). GPA33 is exquisitely restricted to

epithelial cells lining the entire gastrointestinal tract. Its expression has been observed

182

in over 95% of CRCs and associated metastases (Garinchesa et al, 1996). Its use has been

investigated as a therapeutic target for the treatment of locally advanced and

metastatic colorectal cancers (Ciprotti et al, 2014 & Infante et al, 2013).

4.12.2 Specificity of immunoaffinity isolated exosomal miRNAs for CRCs

The use of CD326 (EpCAM), CD133 and GPA33 antibodies coupled with Dynabeads was

studied for the isolation of exosomes from HT29 cell line culture harvested media and

plasma samples from patients with CRC. Detection of commonly upregulated miRNAs

in CRC tissue (miR-21, miR-31 & miR-135b) in exosomes isolated from culture harvested

media and plasma samples of patients with CRC was suggestive of cancer-related

miRNAs in exosomes and their transport to body fluids.

Though detection of circulating exosomal miR-135b and miR-31 was not statistically

significant, miR-135b remains a primary focus of early detection due to its role in

adenoma initiation and progression.

4.12.3 Literature review of exosomal miR-21

Detection of miR-21 has been found significant in this feasibility study but is

upregulated in a wide range of solid tumours including CRCs (Volinia et al, 2006). It has

been secreted in plasma exosomes from patients affected by different cancer types,

such as ovarian, lung and colon carcinomas and pancreatic cancers (Wang et al, 2012,

Cappellesso et al, 2014, Leidinger et al, 2014, Ogata-Kawata et al, 2014 & Que et al,

2014).

Upregulation of miR-21 has been shown to promote cellular proliferation, survival,

invasion and migration in different cancer cell lines (Lu et al, 2008), while its knockdown

decreased tumour cell survival in vitro and tumour growth in vivo in a murine xenograft

model, accompanied by enhanced apoptosis (Yan et al, 2011). Interestingly, miR-21 and

-29a secreted by tumour cells via exosomes have been shown to bind to toll-like

receptors (TLR) on immune cells, leading to TLR-mediated (nuclear factor kappa-light-

chain enhancer of activated B cells) NF-κB activation and secretion of pro-metastatic

inflammatory cytokines that may ultimately lead to tumour growth and metastasis

(Fabbri et al, 2013).

183

miR-21 has been found to be a potential diagnostic exosomal miRNA, and has also been

identified in circulation for exosomes in glioblasmatoma (Skog et al, 2008), and ovarian

(Taylor et al, 2008), lung (Rabinowits et al, 2009) and oesophageal (Tanaka et al, 2013)

cancers. In another study (Ogata-Kawata et al, 2014), Ogata-Kawata and colleagues

performed a microarray-based profiling of exosomal miRNAs isolated by

ultracentrifugation of 750 µl of serum from patients with CRC and harvested media

collected from colon cancer cell lines. In this study, exosomal miRNA expression

profiling identified a panel of 8 upregulated miRNAs (let-7a, miR-21, miR-23a, miR-150,

miR-1224-5p, miR-1229 and miR-1246) in the plasma of CRC patients in comparison to

healthy controls. Expression levels of these exosomal miRNAs are significantly reduced

in the circulation once patients have gone through resectional surgery. Their analysis

for miR-21 showed a true positive detection rate at 61.4% and false positive detection

rates lower than 10%. They also noted that a combined usage of 8 miRNAs did not

provide more diagnostic power.

4.12.4 Use of CD326 (EpCAM) for circulating exosomal miRNA analysis in non CRCSs

Magnet-activated cell sorting (MACS) for CD326 (EpCAM) was used to isolate plasma

exosomes by immunoaffinity for both ovarian and lung cancers (Taylor et al, 2008 and

Rabinowits et al, 2009). EpCAM-positive exosomes were detectable in patients with

benign ovarian disease, ovarian cancer and healthy controls. miRNA expression analysis

for exosomes isolated by immunoaffinity identified a panel of miRNAs including miR-

21, miR-141, miR-200a, miR-200c, miR-200b, miR-203, miR-205 and miR-214. While

exosomal miRNAs from ovarian cancer patients exhibited similar profiles, these profiles

were significantly distinct from benign diseases and normal controls.

In a second study (Rabinowits et al, 2009), researchers measured total exosomal and

miRNA concentrations for EpCAM immunoaffinity isolated exosomes in plasma of

patients with stage I-IV lung adenocarcinoma and healthy controls. The mean exosome

concentration was 2.85 mg/ml (95% CI, 1.94–3.76) for the lung adenocarcinoma group

versus 0.77 mg/ml (95% CI, 0.68–0.86) for the control group (P < 0.001). The mean

miRNA concentration was 158.6 ng/ml (95% CI, 145.7–171.5) for the lung

adenocarcinoma group versus 68.1 ng/ml (95% CI, 57.2–78.9) for the control group (P

184

< 0.001). A panel of exosomal miRNAs (miR-17-3p, miR-21, miR-106a, miR-146, miR-

155, miR-191, miR-192, miR203, miR-205, miR-210, miR-212, and miR-214) was found

to be elevated in patients with lung cancer in comparison to healthy controls. An

elevated level of miR-21 in EpCAM immunoaffinity isolated exosomes suggests that its

use as diagnostic biomarker may not be suitable as a screening test for a CRC.

4.12.5 Total plasma vs exosomal miRNAs

Several other reports have suggested the use of circulating miRNAs in whole plasma or

serum, including miR-141, miR-21, miR-221, miR-29, and miR92a, as diagnostic

biomarkers of CRC (Ng et al, 2009, Huang et al, 2010, Pu et al, 2010, Cheng et al, 2011

& Toiyama et al, 2013). These miRNAs were shown to display sensitivities and

specificities comparable to those of the colon cancer markers currently in use. Notably,

miR141, miR-221, and miR-21 were reported to be upregulated in cancer tissues

(Scheffer et al, 2012, Kawaguchi et al, 2013 & Si et al, 2013); the total miRNA set from

whole plasma or serum may include endogenous cellular miRNAs derived from broken

or circulating tumour cells (Swarup et al, 2007 & Madhavan et al, 2012), which may

explain why miR-141, miR-221, and miR-21 are present at high levels in whole serum or

plasma samples from cancer patients. In addition, serum/plasma miRNAs, which are not

associated with vesicles, were reported to show differential stability in response to

treatment by RNase (Koberle et al, 2013), suggesting that exosomal miRNAs are more

preferable as biological specimens for developing diagnostic biomarkers because of

their stability in serum/plasma. In fact, the results presented here demonstrate that

exosomal miRNA levels also reflect cancer-bearing status and pathological changes of

patients. Exosomes are actively released from cancer cells and their specific

constituents depend on the cell from which they originate. Indeed, cancer cells may use

an as yet unidentified mechanism to embed a subset of miRNAs into exosomes.

In fact, exosomal miRNAs are more stable and resistant to degradation than cellular

miRNAs (Hu et al, 2012). In addition, exosomes can deliver multiple messages

simultaneously, which make them an attractive way of exchanging specific subsets of

mRNA, miRNA, or proteins between donor and recipient cells, also at a distance. Their

specific mechanism of distant communication and their mode of transport in circulation

185

are still unclear. Exosomes display exquisite target cell selectivity in vitro and in vivo,

based, at least in part, on target cell ligand interactions with exosomal tetraspanin-

associated receptors. Maintenance of internalization complexes and reuse of these

complexes for exosome uptake appear to be a common theme. Importantly, the

engagement of protein complexes in internalization-prone membrane domains

provides an explanation for the target cell selectivity, which is unlikely to rely exclusively

on single adhesion molecules, which are frequently expressed on many cells.

The binding of exosomes to the surface of recipient cells is mediated by classical

adhesion molecules involved in cell–cell interactions, such as integrins and ICAMs

including CD326 (EpCAM). However, other molecular pairs more specific to the

exosome membrane, such as TIM-binding phosphatidylserines, carbohydrate/ lectin

receptors and heparan sulfateproteoglycans (HSPGs), could be involved as well (Thuma

et al, 2014). This opens a new search for surface markers to identify cancer specific

markers to isolate these exosomes by immunoaffinity. GPA33 is expressed in >95% of

CRCs and holds immense potential for the isolation of CRC specific exosomes in

circulation.

Accumulating evidence from the literature supports the idea that exosomal miRNAs can

act as regulators of gene expression in distant cells. Particularly in cancer, exosomes

have multiple functions including the promotion of local and systemic processes that

lead to cell growth and dissemination, or impairment of the immune system response.

miRNAs can act either as tumour suppressors or oncogenes (oncomiRs), depending on

target genes and cancer types. Furthermore, a particular miRNA can exploit both

tumour-suppressive and oncogenic functions depending on the cellular context of its

target genes in different cancers (Chen et al, 2012).

4.12.6 Role of exosomal miRNAs in cancer development and progression

Exosomal miRNAs play important roles in tumour development, progression and

metastasis. Their role in the alteration of the immune system in patients with cancer

has been studied in detail. It has been postulated that they confer suppressive effects

on anti-tumour immune responses (Taylor et al, 2011 & Filipazzi et al, 2012). For

example, inhibition of MHC class I gene transcription by overly expressed miR-9 in many

186

cancers leads to a failure to recognise tumour cells by the host’s immune system (Gao

et al, 2013). Similarly, miR-222 down-regulates the expression of intracellular cell

adhesion molecule 1 (ICAM-1) on surfaces of tumour cells, and binding of ICAM-1 to

lymphocyte function-associated antigen (LFA-1) is essential for optimal activation of

cytotoxic T cells, which in turn mediate tumour cell lysis (Ueda et al, 2009).

In some cases, exosomal pathways might discard tumour-suppressor miRNAs or enrich

tumour-promoting miRNAs to enhance their metastatic potential. For example,

enrichment of let-7 miRNA family in exosomes derived from metastatic gastric cancer

cells in comparison with non-metastatic parental cells leads to their aggressive

behaviour (Ohshima et al, 2010). Similarly, metastatic cells from bladder carcinoma

secrete exosomes associated with higher levels of miRNAs associated with invasion,

angiogenesis, and metastasis (Ostenfeld et al, 2014).

Recently, IL-4-activated macrophages have been found to regulate the invasiveness of

breast cancer cells through exosome-mediated delivery of the miR-223, highlighting a

novel communication mechanism between tumour-associated macrophages and

cancer cells (Yang et al, 2011). The pivotal role of exosomes in tumour progression has

recently been highlighted by the discovery that breast cancer exosomes cause cell-

independent miRNA biogenesis and stimulate non-tumorigenic epithelial cells to form

tumour colonies at a distance (Melo et al, 2014).

It has been reported that melanoma-released exosomes can modify distant lymph

nodes to facilitate melanoma growth and metastasis (Hood et al, 2011). Exosomal

mRNAs and miRNAs derived from tumour cells were recovered in lymph node stroma

and lung fibroblasts (Rana et al, 2013). In addition to stromal modification, cancer cell–

derived exosomes modulate protease-induced degradation of extracellular matrix

(ECM) proteins including collagens, laminin, and fibronectin. This in turn leads to loss of

cell adhesion, an increase in cell motility, and invasiveness (Mu et al, 2013).

Exosomal miRNAs, such as miR-105, play important role in the promotion of cancer cell

migration by targeting tight junction proteins at endothelial monolayers in metastatic

breast cancer cells. In vitro experiments have shown that over expression of miR-105 in

non-metastatic cancer cells can induce metastasis and vascular permeability in distant

187

organs, whereas inhibition of miR-105 in highly metastatic tumours can reverse these

effects (Zhu et al, 2014).

Exosomal miRNAs from the miR-200 family, responsible for epithelial-mesenchymal

transition, can transfer the molecules from highly metastatic tumour cells to weakly

metastatic cells by internalisation of exosomes (Epstein et al, 2014 & Le et al, 2014). It

has also been identified that miRNA-enriched exosomes released by CD105 cancer stem

cells from renal carcinomas can modify the tumour microenvironment by triggering

angiogenesis (Grange et al, 2011). Specific exosomal miRNAs, such as those of the miR-

17-92 cluster, have an important role in neoplasia-to-endothelial cell communication

for regulating endothelial gene expression during tumour angiogenesis in leukaemia

cells (Umezu et al, 2013).

It has also been shown that tumour-secreted miR-9 encapsulated into exosomes

promotes endothelial cell migration and tumour angiogenesis participating in

intercellular communication and function (Zhuang et al, 2012). Moreover, exosome

angiogenic miR-210, known to be increased in the serum of cancer patients with

malignant breast cancer, regulates the metastatic ability of cancer cells through

suppression of specific target genes, resulting in enhanced angiogenesis (Kosaka et al,

2013). These findings suggest that the horizontal transfer of exosomal miRNAs from

cancer cells can dictate the micro-environment needed for cancer progression.

The exosomal miRNA profiling from blood of cancer patients and healthy controls has

often revealed important differences in relation to tumour progression, highlighting a

possible use of these miRNAs as disease prognostic biomarkers (Ye et al, 2014). In

addition, many tumours displaying drug resistance show alterations in the expression

of miRNAs. The up- or down-regulation of miRNAs affects the expression of several

target proteins associated with drug sensitivity (Migliore et al, 2013).

Moreover, different studies indicate that exosomes act as vehicles for the exchange of

genetic cargo between heterogeneous populations of tumour cells, generating a way of

transmitting drug resistance (Corcoran et al, 2012 and O'Brien et al, 2013). Recently

Chen and colleagues reported that exosomes from breast cancer cells are capable of

delivering a subset of miRNAs associated with drug resistance (miR-100, miR-222 and

188

miR-30a) to drug sensitive cells (Chen et al, 2014). In contrast, exosomes vesicles can

be used to alter drug resistance by delivering miRNAs essential for drug sensitivity

(Munoz et al, 2013).

4.12.7 Limitations of the study

The limitations of this feasibility study include small sample size and low total RNA

concentrations extracted from immunoaffinity isolated exosomes. Such a small

concentration of RNA is associated with loss of some important diagnostic miRNAs

expressed at low levels, such as miR-135b and miR-182. This could potentially explain

why only 3 out of 5 adenomas and carcinoma with miR-135b were detected.

GPA33 antibody was used without its prior validation on human CRC tissues by

immunostaining and comparison with controls. There is a possibility that miRNAs or

exosomes might have attached to active beads directly rather through an antigen-

antibody reaction for immune isolation. Therefore, assessment and measurement of

surface marker proteins may be used in future to identify whether a direct reaction with

beads has occurred.

4.12.8 Conclusion

This study has met its aim: to examine the feasibility of using plasma exosomal miRNAs

for the detection of CRCs. In this study, different isolation techniques for circulating

exosomes were successfully evaluated and exosomal miRNAs for patients with CRC

were analysed. Unique surface membrane properties and enrichment with cancer-

related miRNAs has permitted immunoaffinity isolation of exosomes in plasma samples

and detection of cancer-related miRNAs by QRT-PCR. In future studies, CD133, CD326

and GPA33 antibody-coupled Dynabeads should be used for the immunoaffinity

isolation of plasma exosomes. Low RNA concentrations extracted from immunoaffinity-

isolated exosomes may have caused a loss of important diagnostic miRNAs. Therefore,

the development of miRNA expression signature for participants with adenomas,

carcinomas and healthy controls might help identify cancer-related miRNAs encased in

circulating exosomes. Since the total RNA yield from immunoaffinity-isolated exosomes

is low, the use of Megaplex™ RT and pre-amplification primers should be continued for

QRT-PCR-based detection. Once adenoma and carcinoma-related exosomal miRNAs are

189

identified for expression profiling, their diagnostic accuracy should be validated in an

independent cohort of participants with adenomas and carcinomas, compared with

healthy controls.

190

Chapter 5: Results

miRNA expression profiling-based

identification of high risk Dukes’ stage B

CRCs

191

5 Results: Tissue miRNAs

5.1 Summary of results

Despite surgical resection being highly effective for patients with Dukes’ stage B CRCs,

approximately 25% of these patients develop metastatic disease during the follow up

period. This has led to uncertainty about the suitability of post-operative chemotherapy

for Dukes’ B CRCs. It is not possible to accurately predict the development of metastasis

in this group of CRCs by histological examination of resected cancer specimens alone.

Therefore, molecular markers are required to identify which high risk Dukes’ B cancer

patients may benefit from post-operative chemotherapy.

The aim of this study was to examine the utility of tissue miRNAs combined with

common gene mutations in CRCs as biomarkers for the prediction of metastasis

development in patients with high risk Dukes’ B cancers. The objectives were: (i) to

identify specific tissue miRNAs associated with different stages of CRCs, (ii) screen for

common gene mutations (KRAS, BRAF, PIK3CA) in formalin-fixed paraffin-embedded

(FFPE) CRC tissue, and (iii) combine all data to predict the development of metastasis in

patients with Dukes’ B cancers.

Patients who underwent curative bowel resection for Dukes’ A, Dukes’ C or ‘low-risk B’

(patients without distant metastases at 5-year follow-up) were compared with case-

matched ‘high-risk B’ (patients who developed metastatic disease at some stage during

the following 5 years). miRNAs from tumour and adjacent normal tissue and common

gene mutations (KRAS, BRAF, PIK3CA) in primary cancer tissue were analysed to identify

prognostic tissue markers for the development of metastasis in patients with Dukes’ B

CRCs. Twenty age- and gender-matched paired tumour and adjacent normal tissues,

five from Dukes A, ‘low-risk B’, ‘high-risk B’ and Dukes’ C groups were used for miRNA

expression profiling, described as a training cohort, using Taqman Megaplex™ RT and

pre-amplification primers Human Pool A v.2.1 and Pool B v.2.0. A panel of

discriminatory miRNAs for ‘high risk B’ (miR-15b, miR-21, miR-32, miR-34a, miR-125a-

5p, miR-135b, miR-182, miR-184, miR-302b, miR-330-3p, miR-330-5p, miR-381, miR-

483, mR-508, miR-708) was selected and confirmed with QRT-PCR on individual array

samples. A panel significantly discriminating 8 miRNAs was further analysed in a

192

validation cohort (n=72). Expression levels of miR-15b and miR-135b were significantly

downregulated (p < 0.001) in ‘high-risk B’ cancers compared with Dukes A, ‘low-risk B’

and Dukes’ C without metastasis. 97 cancer tissue specimens were analysed for

mutation status. The KRAS oncogene was found to be mutated in 28 cases, whereas

BRAF mutations were detected in 19 cases, and PIK3CA mutations were detected in 9

cases. No significant differences of mutation status and the development of metastasis

were observed. These results suggest that expression analysis of miR-15b and miR-135b

in FFPE tissue specimens can predict the development of metastasis in Dukes’ B cancers,

but by this requires further validation in a larger cohort.

193

5.2 miRNA expression signature for different Dukes’ stages

Not all miRNAs in the array were expressed in colorectal tissue. miRNAs with expression

levels of (CT) < 35 and present in at least one pooled sample of normal or cancerous

tissue were included for further analysis. Significance analysis for microarrays (SAM) of

RNU6B normalised expression profiling data showed significantly different miRNAs for

‘low risk B’ and “high risk B” tumours (Figure 50). These differentially expressed miRNAs

were selected and compared further for paired pools of adjacent normal and cancerous

tissue to identify deregulated miRNAs in tumour tissue (paired Student’s t-test, p<0.05).

The miRNA expression signature derived from Multiexperiment Viewer v4.4 (MeV)

software shows miRNAs are deregulated in tumour tissue and Dukes’ stage B (Figure

51). miRNAs which were deregulated in tumour tissue in comparison with adjacent

normal tissue and significantly different from “high risk B” tumours were selected from

TaqMan® MicroRNA Array Card A v2.1 for further validation, and included miR-15b,

miR-21, miR-32, miR-125a-5p, miR-135b, miR-182, miR-302b, miR-330-3p, miR-330-

5p,miR-381, miR-483, mR-508 and miR-708. miR-34a was selected due to its strong

relation with p53, and miR-184 and RNU6B were selected as endogenous controls for

normalization purposes.

194

Figure 50: Significance analysis for microarrays (SAM).

SAM measured the strength of the relationship between miRNA expression in ‘high risk

B’ and ‘low risk B’, and in Dukes’ C and A. Repeated permutations of RNU6B normalised

data were used to identify miRNAs which were significantly different for ‘high risk B’.

195

Figure 51: Hierarchical cluster analysis of expression profiling data for different Dukes’ stages.

A hierarchical cluster dendrogram of 41 differentially expressed miRNAs in either

tumour or normal tissue, determined using the Multiexperiment Viewer v4.4 software

using miRNA expression profiles derived from TaqMan® MicroRNA Array cards. 20 cases

of CRC were profiled representing Dukes’ A, Dukes’ ‘low-risk B’ (represented as Good

B), Dukes’ ‘high-risk B’ (represented as Bad B) and Dukes’ C. and miRNAs with CT < 35

that were present in the majority of samples were analysed statistically.

196

5.3 Confirmation of Expression Profiling data with QRT-PCR

Further confirmation of selected miRNAs from 20 tumour and paired micro-dissected

adjacent normal tissues used in Taqman® MicroRNA Array, and showed that miR-21,

miR-34a, miR-135b and miR-708 were significantly up regulated in cancer tissue

compared to adjacent normal tissue (paired Student’s t-test, p≤0.05). Levels of miR-

135b in tumour tissue were 1525-fold higher compared to adjacent normal tissue

(Student’s t-test, p<0.05). The average expression levels of miR-15b, miR-32, miR-125a-

5p, miR-302b, miR-182 and miR-708 in cancer tissue were higher than that detected in

normal tissue. There were no significant differences in the expression levels of miR-330-

3p or miR-330-5p in tumour and adjacent normal tissue. The comparison of ‘high risk B’

with ‘low risk B’ tumour tissue (analysed by Mann-Whitney U test, Z-Score and a direct

comparison of CT values of both groups) identified that miR-15b, miR-21, miR-34a, miR-

125a-5p, miR-135b, miR-182, miR-508 and miR-708 were differentially expressed

between the two groups (Mann-Whitney U test, p≤0.10). Analysis of variance (ANOVA)

for RNU6B normalized expression levels of miRNAs in tumour tissue relative to adjacent

normal tissue (ΔΔCT) for 4 tumour groups showed no significant difference in

expressions of miR-32, miR-302b, miR-330-3p, miR-330-5p, miR-387 and miR-483.

Expression levels of miR-15b and miR-125a-5p were significantly lower, whereas

expression levels of miR-21, miR-135b, miR-182 and miR-708 were significantly higher

in ‘high risk B’ tumours in comparison to ‘low risk B’ tumours (Mann-Whitney U test,

p<0.05). Table 42 summarizes the PCR validation of selected miRNAs identified from

array with 20 paired tumours and normal tissue samples. Based on the p-value obtained

(Mann-Whitney U test, p≤0.05) seven miRNAs (miR-15b, miR-21, miR-34a, miR-135b,

miR-125a-5p, miR-182 and miR-708) were selected for further validation on an

independent cohort of 72 paired tumour and adjacent normal tissue.

197

miRNA Expression in

tumour

tissue

Tumour vs.

normal

(fold change)

p-value Low risk B vs.

high risk B

(fold change)

p-value

miR-135b Up regulated 1525.29 0.0002 5.29 0.04

miR-21 Up regulated 10.65 0.01 1.53 0.05

miR-708 Up regulated 7.04 0.01 2.33 0.02

miR-32 Up regulated 5.11 0.28 0.02 0.26

miR-15b Up regulated 5.01 0.33 0.01 0.07

miR-302b Up regulated 4.43 0.33 2.71 0.86

miR-508 Up regulated 4.15 0.40 0.18 0.09

miR-483 Up regulated 4.00 0.34 3.53 0.92

miR-381 Up regulated 3.75 0.43 0.32 0.40

miR-125a-5p Up regulated 2.59 0.45 0.04 0.10

miR-34a Up regulated 2.50 0.05 0.78 0.10

miR-184 Up regulated 1.21 0.90 0.58 0.61

miR-330-3p No Change 0.48 0.12 0.43 0.30

miR-330-5p No Change 0.39 0.14 0.26 0.56

Table 42: Summary of PCR-based confirmation of selected miRNAs on samples used for array.

Changes in expression levels of miRNAs in tumour and normal tissue are expressed as

fold changed generated from ∆∆CT values with corresponding p-values (paired

Student’s t-test). Expression levels of each miRNA, both direct CT values and RNU6B

normalized ∆CT were compared between ‘high risk B’ and ‘low risk B’ tumours.

198

5.4 Validation of selected miRNAs on second cohort

5.4.1 Tumour versus normal tissue miRNAs

Comparison of RNU6B normalized miRNA expression (ΔCT) in cancerous and adjacent

normal tissue using non-matched pair analysis and the Student’s t-test showed a

significant upregulation of miR-21, miR-34a, miR-135b, miR-182 and miR-708 in tumour

tissue (p<0.05). Matched pair analysis of miRNAs in cancer and their adjacent normal

tissue showed down-regulation of miR-125a-5p and upregulation of miR-15b, miR-21,

miR-34a, miR-135b, miR-182 and miR-708 in tumour tissue (Wilcoxon matched pairs

signed rank test (Figure 52).

199

Figure 52: Comparison of RNU6B normalized miRNA expression (ΔCT) in cancerous and adjacent normal tissue.

Matched pairs of cancerous and adjacent normal tissue were compared with Wilcoxon

matched pairs signed rank test and showed upregulation of miR-21, miR-34a, miR-135b,

miR-182 and miR-708.

200

5.4.2 Dukes’ stage ‘low risk B’ versus ‘high risk B’

RNU6B normalized expressions (∆CT) of 8 selected miRNAs in tumour tissue of different

Dukes’ stages were compared using ANOVA with Bonferroni’s correction. For ‘low risk

B’ and ‘high risk B’ tumours there was no significant difference in expression levels (∆CT)

of miR-21, miR-34a, miR-125a, miR-182, miR-508 and miR-708 in tumour tissues.

Expression levels of miR-135b were significantly lower in ‘high risk B’ tumours compared

to Dukes’ A and Dukes’ C (p=0.0001) and ‘low risk B’ (p=0.0003) tumours. Similarly,

expression levels of miR-15b were significantly lower in ‘high risk B’ tumour tissue

compared with Dukes’ stage A, C and ‘low risk B’ tumour tissue (p<0.001; Figure 53).

201

Figure 53: Comparison of miRNA expression levels of different stages.

Expression levels (∆CT) of miR-15b and miR-135b were significantly lower for ‘high risk

B’ tumours (p<0.001) in comparison with ‘low risk B’ and Dukes’ C tumours (ANOVA

with Bonferroni’s correction).a) Expression levels of miR-15b in ‘high risk B’ in

comparison to ‘low risk B’ tumour tissue (mean difference in ∆CT = 3.29, 95% CI = 0.37-

6.03) and Dukes’ stage C tumours (mean difference in ∆CT = 3.95, 95% CI = 1.11-6.79).

b) Expression levels of miR-135b in ‘high risk B’ tumour tissue in comparison to ‘low risk

B’ tumour (mean difference in ∆CT = 7.36, 95% CI = 3.47-11.25) and Dukes’ stage C

tumours (mean difference in ∆CT = 8.23, 95% CI = 4.31-12.16).

202

The expression levels of miR-135b in tumour and adjacent normal tissue showed

relatively lower expressions in comparison to Dukes’ C and “low-risk” Dukes’ B (Figure

54).

Figure 54: Plot of miR-135b expression for paired adjacent normal and cancerous tissue.

Non-normalized expression levels of miR-135b are plotted for both tumour and paired

normal tissue. Both ‘low risk B’ and Dukes’ C tumour have two distinct subpopulation

of tumour. One group has no detectable expression of miR-135b in micro-dissected

adjacent normal tissue similar to ‘high risk B’ tumours. However, expression level of

miR-135b in ‘high risk B’ tumours are much lower (Wilcoxon matched pairs signed rank

test, p<0.001). In other subgroups of ‘low risk B’ and Dukes’ C tumours, miR-135b is well

expressed in adjacent normal tissue too.

203

Trend analysis of ΔCT of tumour miRNA expression for Dukes’’ stage A, B and C revealed

a linear trend for expression levels of miR-15b, miR-125a-5p and miR-135b from Dukes’

stage A to Dukes’’ C (p<0.05) (Table 43). Further analysis for left sided tumours showed

that miR-15b, miR-21 and miR-34a had linear trend of progression for their levels of

expression with increasing cancer stage.

miRNAs All tumours Left-sided tumours

R square P- value R square P- value

miR-15b 0.26 0.027 0.41 0.0167

miR-21 0.02 0.9572 0.23 0.04

miR-34a 0.10 0.3837 0.36 0.0398

miR-125a-5p 0.28 0.0189 0.30 0.0985

miR-135b 0.23 0.0439 0.24 0.1738

miR-182 0.16 0.1863 0.30 0.0901

miR-708 0.14 0.2435 0.30 0.0948

Table 43: Linear regression analysis of miRNA expressions for Dukes’ stages.

Expression levels of miRNAs were analysed for increasing expressions for higher tumour

stages and correlations were expressed as R square and p-values.

204

5.4.3 Mutation Analysis for cancer tissue

97 cancer tissue specimens were analysed for their K-ras, BRAF and PIK3CA mutation

status, whereas 3 cancer tissue specimens could not be analysed due to poor quality

DNA. The K-ras oncogene was mutated in 28 cases, whereas 19 and 9 occurred in BRAF

and PIK3CA, respectively. The K-ras mutation was more prevalent in left-sided tumours

(n=16, 32.5%) and the BRAF mutation was more prevalent in right-sided cancers (n=17,

35.5%). 43 (56.5%) cancer specimens without a background of polyp had mutated genes

(BRAF 15, KRAS 21 PIK3CA 7) whereas 13 (62%) specimens with cancer polyps were

found to have mutated genes (BRAF 4, KRAS 7, PIK3CA 2). 8 (20%) of cancer specimens

from the ‘low risk B’ group had a mutated BRAF gene, and only 1 (5%) of cancer

specimen from the ‘high risk B’ group had a mutated BRAF gene. For frequency of

mutated K-ras genes, in the ‘low risk B’ group 12 (31%) specimens exhibited mutated

status, and in the ‘high risk B’ group 6 (40%) specimens had mutated status. The

frequency of different mutated genes detected in different groups of patients is

tabulated in Table 44.

Mutation comparisons for all 3 oncogenes revealed no significant difference between

‘low risk B’ and “high- risk B” cancer tissue. K-ras mutations were significantly higher in

Dukes’ stage A compared to Dukes’ stage C (p=0.011). BRAF and K-ras were mutually

exclusive, except in two subjects where complex mutation was observed. Similarly, in

two subjects mutations were observed in both K-ras codon 12 and codon 13. The BRAF

mutation was significantly higher in right-sided cancers (p=0.0001) and cancerous tissue

resected from female patients (p=0.02). No significant difference was observed with

tumour background. The comparative analysis of mutation status with significance in

differences (p-values) for clinical variables is shown in Table 45.

205

Clinical

parameter

Subgroups

Total number

of samples (n)

Frequency of mutations

detected (n)

Dukes’

Stage

BRAF K-ras PIK3CA

A 7 1 5 2

Low risk B 39 8 12 3

High risk B 15 1 6 1

C 36 3 5 3

Gender Male 59 5 16 3

Female 38 14 12 6

Location Left sided 49 2 16 5

Right sided 48 17 12 4

Background No polyps 76 15 21 7

Polyps 21 4 7 2

T-Stage T1 + T2 12 2 6 2

T3 51 8 12 3

T4 34 9 10 4

Table 44: Frequency table for different positive mutations detected for BRAF, K-ras and PIK3CA mutations.

Frequencies of mutations are given for different clinicopathological features and

subgroups of participants with CRC.

206

Clinical

parameter

Subgroups P-values for inter-group and intra-group

comparisons of mutant genes

BRAF K-ras PIK3CA

Dukes’ Stage

A vs low-risk B 0.701 0.075 0.309

A vs high-risk B 0.641 0.189 0.298

A vs C 0.52 0.01 0.324

Low-risk B vs C 0.649 0.08 0.92

Low-risk B vs high-risk B 0.146 0.546 0.989

High-risk B vs C 0.071 0.084 0.839

Gender

Male vs. Female 0.002 0.644 0.113

Location

Left vs. Right 0.0001 0.342 0.746

Background

No polyps vs. polyps 0.945 0.631 0.966

T -Stage T2 vs T3 0.929 0.041 0.269

T3 vs T4 0.248 0.556 0.371

T4 vs T2 0.47 0.068 0.412

Table 45: P-values for inter-group and intra-group comparisons between the three mutated genes.

P-values are tabulated for BRAF, K-ras and PIK3CA mutant genes with comparisons for

different clinical variables.

207

5.4.4 Correlation with other clinico-pathological variables

No significance was observed in miRNA expressions that correlated with age, gender,

background, tumour location, BRAF and PIK3CA mutation status. Analysis of miR-182

expression in tumour tissue showed higher expressions levels in Dukes’ stage C

(p=0.02), T3 tumour (p=0.006) and left sided tumours (p=0.001). Furthermore, the

expression levels of miR-182 were significantly lower in PIK3CA (p=0.004), BRAF

(p=0.0064) and K-ras (p=0.012) mutant tumours. The expression levels of miR-34a

increased with higher T stages (Figure 55).Linear trend analysis of the expression levels

of miR-34a showed a progressive increase of expression from T2 staged tumours to T3

and T4 staged tumours (p=0.038). Furthermore, in the presence of extramural vascular

invasion (EMVI), expression levels of miR-125a-5p, miR-135b, miR-182 and miR-708

were significantly different (p<0.05) in tumours without EMVI (Figure 56). The Mann-

Whitney-U-Test reported significantly lower expression levels of miR-34a (p=0.028) and

miR-708 (p=0.023) in K-ras mutated tumour tissues when compared with wild-type

tumours (Figure 56). The expression levels of miR-21 in both high risk Dukes’ B and C

tumours were significantly higher especially in left-sided tumours in comparison to their

adjacent normal tissue (p<0.05).

208

Figure 55: Plot of miR-34a expressions and tumour T stages as identified on pathological examination.

There was no significant difference in the expression levels of miR-34a between T1 and

T2 staged tumour tissues. T1 and T2 staged tumours were combined for comparison

with T3 and T4 tumours. One-way analysis of variance with Bonferroni's correction

showed significantly higher expression levels of miR-34a in T3 (mean difference in ∆CT

= -1.949, 95% CI = 3.245 to -0.6523; p=0.001) and T4 (mean difference in ∆CT = -1.933,

95% CI = -3.332 to -0.5345) tumour tissue compared to T1 and T2 tumours. There was

no significant difference in the expression levels of miR-34a between T3 and T4 staged

tumours.

209

Figure 56: Comparison of miRNA expression levels for tumours with histological evidence of EMVI and Mutant KRAS gene.

Expression levels for EMVI positive and negative tumours of (a) miR-125a (median

difference in ∆CT = -1.56, 99% CI = -1.354 to -1.699), (b) miR-135b (median difference

in ∆CT = -2.76, 99%CI = -2.69 to -2.85), (c) miR-182 (median difference in ∆CT = -2.96,

99% CI = -2.75 to-3.14) and (d) miR-708 (median difference in ∆CT = 0.96, 99% CI = 0.7

to 1.21). Expression levels of miR-34a and miR-708 were significantly higher for K-ras

mutated tumours (p<0.05) in comparison with tumours with wild-type K-ras (Mann-

Whitney U test). e) Differential expression of miR-34a (median difference in ∆CT = -

0.885, 99% CI = -0.661 to -1.11) for K-ras mutant and wild-type genes f) Comparison of

expression levels of miR-708 (median difference in ∆CT = -1.13, 99%CI = -0.84 to -1.42)

for mutant and wild-type K-ras.

210

5.5 Discussion

The results of this study show that miR-135b is overexpressed in micro-dissected CRC

tissue in comparison to the matched adjacent normal mucosal tissue. These findings are

in line with the previously published studies (Motoyama et al, 2009, Sarver et al, 2009,

Wang et al, 2010, Bandres et al, 2006). However, no studies have co-related levels of

miR-135b with cancer stage or identified the prognostic significance of miR-135b for

Dukes’ stage. In this study, these results have shown no difference in expression levels

of miR-135b for Dukes’ stage B or C. However, lower expressions levels of miR-135b

significantly differentiate between ‘high-risk’ and ‘low-risk’ B cancers.

5.5.1 miR-135b and APC.

Inactivation of the APC gene is a major initiating event in colorectal carcinogenesis

(Salby et al, 2009), leading to stimulation of the Wnt pathway via free β-catenin

(Segditsas et al, 2006). APC inactivation has been found in more than 60% of colonic

tumours (Fearon &Vogelstein, 1990), and is associated with upregulation of miR-135b

in CRC cells (Nagel et al, 2008). miR-135b levels are upregulated in colorectal adenomas

and carcinomas and correlate with low APC levels (Aslam et al, 2015). These

observations suggest that alteration in the miR-135 family can be one of the early

events in the molecular pathogenesis of CRC. In this study using matched pair analysis,

it was observed that in ‘high-risk’ Dukes’ B cancers the expression levels of miR-135b

were not significantly different for tumour and paired normal adjacent mucosal tissue.

These findings suggest an alternative mechanism of tumour initiation in ‘high-risk’

Dukes’ B tumours may exist. Although the initial validation of array with RT-PCR showed

higher levels of miR-135b in high-risk Dukes’ B tumours, the results were skewed by

relatively higher expressions of one sample in the pooled group. The validation using

the second cohort also identified a similar trend for two cases.

5.5.2 miR-135b in metastatic cancers

For the validation cohort, no differences in miR-15b levels were noted for cancerous

and adjacent normal tissue. A previously published expression profiling study by Xi and

colleagues has shown similar results for miR-15b in CRC (Xi et al, 2006). In that study,

the levels of miR-15b were found to be significantly lower for patients who developed

211

metastatic diseases after surgical resection of Dukes’ B tumour. A prognostic study of

miR-15b in hepatocellular carcinoma has shown a similar trend where lower expression

levels of miR-15b are associated with higher risk of recurrence (Chung et al, 2010). In

another study, miR-15b was found to be down-regulated in CRC tissue but did not

correlate with the presence of lymph node metastasis and disease-free survival (Wang

et al, 2012).

5.5.3 Role of miR-15b in CRCs

The exact role of miR-15b in CRC development and progression is yet not fully

understood. However, the mechanistic studies have shown that the miR-15b correlates

with E2F-regulated genes and appears to be part of the E2F-regulatory network (Ofir et

al, 2011). E2F1, miR-15b and cyclin E constitute a feed-forward loop that modulates E2F

activity and cell-cycle progression, where inhibition of miR-15b expression results in

enhanced E2F1-induced cyclin E and G(1)/S transition (Ofir et al, 2011). In another in

vitro study, knockdown of HPV E7 down-regulated the expression levels of miR-15b in

colorectal cell lines (Myklebust et al, 2011). Bcl-2 has been also shown to be a target of

miR-15b in gastric cancer, however miR-15b knockdown in CRC cell lines did not show

any change in expression levels of Bcl-2 (Davidson et al, 2009).

5.5.4 miR-21 as a prognostic marker

In this study, the results show that miR-21 levels were not significantly different for low-

risk and high-risk Dukes’ B CRCs. However, the expressions levels were higher in high-

risk Dukes’ B and C and especially in left-sided tumours, suggesting that miR-21 is

significantly involved in tumour metastasis. It has already been established that

elevated miR-21 expression leads to reduced apoptosis and increased cell proliferation,

cell migration, intravasation and metastasis by targeting several tumour suppressor

genes such as programmed cell death 4 (PDCD4), phosphatase and tensin homolog

(PTEN), tropomyosin 1 (TPM1), cell division cycle 25 homolog A (Cdc25a), reversion-

inducing-cysteine-rich protein with kazal motifs (RECK), TIMP3, maspin, nuclear factor

1 B-type, and Ras homolog gene family member B (RHOB) (Schetter et al, 2012). The

role of miR-21 in cell migration and intravasation has already led to studies exploring its

role as prognostic marker for high-risk stage II CRCs

212

It has been reported that miR-21 might play a role in intravasation through down-

regulation of the tumour suppressor Pdcd4 (Asangani et al, 2008). Higher expression of

miR-21 has been noted in metastatic CRC compared to non-metastatic CRC, and miR-

21 has been associated with lymph node metastasis (Vickers et al, 2012 & Slaby et al,

2007). Upregulation of miRNA-21 has initiated tumour formation but also increased

cancer invasion, intravasation and metastatic potential (Asangani et al, 2008, Wang et

al, 2009a). Schetter and colleagues have detected miR-21 expression in primary tumour

tissues and identified this as a prognostic marker in stage II CRC (Schetter et al, 2008).

This group subsequently extended their study to evaluate the expression status of 23

inflammatory genes and discovered that a combination of inflammatory risk score and

miR-21 expression was a better predictor of prognosis than either of these when

applied individually (Schetter et al, 2009). It has also been proposed that when used in

combination with other prognostic parameters such as staging, MSI status, genotyping

and mRNA profiling, miR-21 expression may improve risk stratification to help guide the

right treatment strategies and may result in much improved survival rates (Schetter et

al, 2012).

Kjaer-Frifeldt and colleagues have reported results of in situ hybridization using FFPE

specimens from 520 patients with stage II colon cancer, and showed that increased miR-

21 expression levels in tumour tissue correlates significantly with poor disease-free

survival rates (Kjaer-Frifeldt et al, 2012). In another study, array-based profiling of 40

samples of paired stage II CRC and adjacent normal mucosal tissues was followed by RT-

PCR based validation of miRNAs in 138 stage II cancer specimens (Zhang et al, 2013).

This study has identified a panel of 6 miRNAs (miR-20a-5p, miR-21-5p, miR-103a-3p,

miR-106b-5p, miR-143-5p and miR-215) that can be used as prognostic and predictive

markers of disease recurrence in patients with stage II CRC. It has been established that

higher expression levels of miR-21 in CRC tissue are linked to a worse prognosis and

poor therapeutic outcomes. This highlights its potential use as prognostic and predictive

biomarkers for CRCs (Shibuya et al, 2010, Kulda et al, 2010, Nielson et al, 2011, Zhang

et al, 2009).

213

5.5.5 Combinatorial approach for prognostic tissue miRNAs in CRC

It has also been proposed that when used in combination with other prognostic

parameters such as staging, MSI status, genotyping and mRNA profiling, it may improve

risk stratification to help guide the right treatment strategies and may result in

improved survival rates (Schetter et al, 2012).

In a recently published study researchers have analysed miR-320e expression in stage II

and stage III CRCs to predict cancer recurrence and disease-free survival (Perez-

Carbonell et al, 2015). It has been identified that higher levels of miR-320e expression

was associated with poor disease-free survival. Another study has shown that lower

levels of miR-29a expression in CRC tissue specimens is associated with higher risk of

disease recurrence in patients with stage II CRCs (Weissmann-Brenner et al, 2012).

In a prognostic study of miR-203 for patient ethnicity and tumour stage, Bovel and

colleagues have found poor survival for black ethnic patients with stages II CRCs (Bovel

et al, 2013). To help develop a miRNA based prognostic biomarker, Schepeler and

colleagues have analysed 315 miRNAs profiles in stage II CRC tissue specimens

exhibiting features such as microsatellite stability and microsatellite instability

(Schepeler et al, 2008). These authors have identified four miRNAs (miR-142-3p, miR-

212, miR-151, and miR-144) capable of accurately (sensitivity 92%, specificity 81%)

discriminating the microsatellite instability status of patients. In addition, a significant

correlation with recurrence-free survival for expression levels of miR-320 and miR-498

has indicated their potential use as prognostic marker in CRC (Schepeler et al, 2008). In

another study of miR143, lower expression levels of miRNA-143 served as an

independent prognostic biomarker for CRC tissue with wild type K-ras (Pichler et al,

2012). Table 46 summarises the findings of studies for the purpose of identification of

high-risk CRCs based on tissue expressions.

214

Studies miRNA Tissue miRNA

expression

Zhang et al, 2013 miR-20a Up regulation

Schetter et al, 2008, Schetter

et al, 2009, Kjaer-Frifeldt et al,

2012 & Oue et al, 2014

miR-21 Up regulation

Weissmann-Brenner et al,

2012

miR-29a Down regulation

Zhang et al, 2013 miR-20a-5p, miR-21-5p,

miR-103a-3p miR-106b-

5p, miR-143-5p, miR-215

Up regulation

Bovell et al, 2013 miR-203 Up regulation

Schepeler et al, 2008 miR-320, miR-498 Down regulation

Perez-Carbonell et al, 2015 miR-320e Up regulation

Pichler et al, 2012 miR-143 Down regulation

Table 46: Summary of studies dealing with potential role of different miRNAs for identification of high risk CRCs.

5.5.6 miR-34a and miR-125a-5p

In this study, the levels of miR-34a were significantly higher in cancerous tissue in

comparison with adjacent normal tissue. Such over expression of miR-34a in cancer

tissue has also been identified by other researchers (Slattery et al, 2011). The

downregulation of miR-125a-5p has also been reported for CRC and lung cancer (Zhang

et al, 2009 & Jiang et al, 2011). Studies in lung cancer have shown that miR-125a-5p acts

as tumour suppressor by inducing apoptosis mediated by p53 (Jiang et al, 2011) but the

exact role of miR-125a-5p in CRC is still unknown. Overexpression of p53-dependent

miR-34a and down-regulation of miR-125a-5p suggests there are complex interactions

between these two miRNAs in the p53-driven cancer pathway.

215

5.5.7 miR-708 and miR-182

Upregulation of miR-708 in APC knockdown tumours and colitis-associated tumours in

mouse colon highlights its role in tumour initiation and transformation via known

cancer signalling pathways (Necela et al, 2011). miR-182 upregulation in CRC tissue has

been shown by Liu and colleagues (Liu et al, 2013). Other in vitro studies have identified

miR-182 as an inhibitor of apoptosis associated with increased cancer cell survival and

cancer progression (Ma et al, 2011, Cekaite et al, 2012). More importantly, the miR-

183-96-182 gene cluster is located in the 7q32 genomic region and its amplification has

been detected in 26% of primary solid organ tumours and 30% of liver metastases

(Midgley and Kerr D, 1999). The higher expression of miR-182 in T3 tumours in

comparison to T1 and T2, EMVI-positive tumours and Dukes’ stage C tumours suggests

that miR-182 is a driving force in the development of metastasis.

5.5.8 BRAF mutations in right colonic cancers

Currently, there is a widespread belief that the mechanism of tumour spread in the

colon varies based on the tumour location. Even in familial cancers almost 100% of FAP

is diagnosed on the left side and approximately 70% of the other familial syndrome,

HNPCC, is observed on the right side (Brim et al, 2008). BRAF mutations were also found

to significantly higher in right-sided tumours. BRAF mutations have been shown to drive

microsatellite instability (Bufill et al, 1990). Although in this study no miRNA showed

any differences in terms of location, when analysed within the location groups

significant expression changes were observed. This suggests that location-specific

miRNA expression levels could predict prognosis.

5.5.9 Role of miRNAs in CRC metastasis

The development and refinement of a prognostic or predictive marker requires a

detailed understanding of the exact mechanism of cancer cell dissociation from primary

tumour mass, movement through the basement membrane and into the blood and

lymph vessels in CRC. miRNAs are thought to control CRC metastasis at least at the level

of tumour growth, invasion and intravasation (Tokarz and Blasiak, 2012). Different

miRNAs including miR-135b have been associated with essential features of tumour

growth such as cancer cell proliferation and reduced apoptosis. Once they have entered

216

the bloodstream, cancer cells are attacked by the immune system and may result in

cancer cell degradation (Bockhorn et al, 2007). A number of miRNAs have been

discovered to play critical roles in modulation of innate and adaptive immune responses

in initial and advanced stages of cancer (Harris et al, 2008, Lu and Liston, 2009).

Although the exact role of miRNA in cancer cell intravasation is unknown, it seems that

miRNAs help cancer cells to evade recognition by the immune system in the blood and

lymph vessels. The escape of cancer cells from capillaries to invade the parenchyma is

one step prior to the development of cancer cell colonies in a site distant to primary

tumour. Here again, the role of miRNAs in this process is unknown, but it is

hypothesized that these molecules may influence cancer cell extravasation. (Tokarz and

Blasiak, 2012). The final step in cancer metastasis is colonization of distant organs by

cancer cells and the formation of secondary tumours. It is known that circulating cancer

cells from different body organs show an affinity for particular organs (Paget, 1989).

This organ-specific targeting is explained by the “seed and soil” hypothesis; cancer cells

are the “seed” and the specific organ microenvironment is the “soil”. Metastatic

colonization of this microenvironment may depend upon the ability of certain cancer

cells to proliferate and adapt to new conditions, an ability possessed by cancer stem

cells. miRNAs may regulate the pathways required for the cancer cells to obtain this

stem cell-like phenotype and consequently may play a role in CRC metastasis (Dieter et

al, 2011).

5.5.10 Other potential prognostic markers for CRCs

Commercially, there has been an emergence of several new predictive tests including

ColoGuideEx, ColoPrint, OncoDefender-CRC and Oncotype DX. These tests follow the

principle of measuring multiple gene expressions to distinguish between high- and low-

risk tumours (Agesen et al, 2012, Tan and Tan, 2011, Webber et al, 2010, Kelley and

Venook, 2011). The main drawback associated with these new tests is the need for CRC

tissue samples. Consequently, these tests can only be conducted after cancer resection

surgery or endoscopic biopsy. There therefore remains a need for non-invasive blood

or stool-based prognostic and predictive biomarkers and for this purpose miRNAs in

plasma, serum and faecal matter of patients with CRC have been studied (Pu et al, 2010,

Cheng et al, 2011, Kalimutho et al, 2011 & Wang and Gu, 2012a). Researchers have

217

identified higher levels of miR-221 and miR-141 in plasma as an independent prognostic

factor for poor overall survival in CRC patients (Pu et al, 2010, Cheng et al, 2011).

Researchers have also shown that miR-29a and miR-141 are significantly higher in

Dukes’ stage D, and this allows discrimination between metastatic and non-metastatic

tumours (Cheng et al, 2011 & Wang and Gu, 2012a). Other blood or stool-based

potential prognostic or predictive biomarkers include single nucleotide polymorphisms

(SNPs) associated with miRNA. It has been postulated that the presence of SNPs in pri-

, pre- and mature miRNAs can modulate gene expression by altering miRNA function.

SNPs in pre-miR-423 (rs6505162) and pre-miR-608 (rs4919510) variant genotypes of

these SNPs have been significantly associated with recurrence-free survivals in patients

with CRC (Xing et al, 2012).

5.5.11 Limitation of study

There are a few limitations associated with tissue miRNA analysis work. The sample size

of Dukes’ stage A and ‘high-risk’ Dukes’ B tumours used in the validation cohort were

smaller than those of ‘low-risk’ Dukes’ stage B and Dukes’ C tumours. Although almost

an equal number of left- and right-sided cancers were analysed in the study, there was

a certain amount of bias towards the right-sided cancers; the right-sided cancer group

contained more non-metastatic than metastatic cancers. Although clinicopathological

variables data was collected retrospectively by two independent observers, like many

other retrospective studies there remains a chance of bias in the accuracy of data for

clinicopathological variables.

One of the arguable limitations in the method of studying miRNAs is the use of FFPE

tissues. It has been proven that in FFPE tissues, due to tissue sample processing, fixation

and storage, RNA degradation occurs. Hence, RNA quantification poses difficult

challenge for researchers (Macabeo-Ong et al, 2002 & Cronin et al, 2004). In contrast,

due to their stability and small size, miRNAs are better preserved and be readily

extracted from FFPE samples (Hui et al, 2009). Historically, FFPE tissue has been a widely

used archive material for biomarker discovery and validation (Lewis et al, 2001).

Furthermore, the availability of clinicopathological data associated with these samples

is of enormous use for its correlation with disease recurrence, development of

218

metastasis and survival. This therefore greatly enhances our ability to assess miRNAs as

cancer biomarkers. A possible explanation of the discordance of the results of this study

and the heterogeneity of these findings when compared with those of previous similar

studies are technical differences associated with the components of the multistep

process of the isolation and analysis of miRNAs. It is also strikingly important that other

prognostic and predictive marker studies have used different endpoints against which

the miRNA expressions have been tested. Some studies have explored correlations

between primary cancer tissue miRNAs and nodal involvement or microscopic features

of poor outcomes. In contrast, other studies have evaluated correlations of miRNA

expressions with outcome endpoints such as development of metastasis, disease

recurrence, disease-free and overall survival.

During the expression profiling it would have been ideal to run the array on individual

samples rather than pooling the cases together. However, this was done because it was

cost-effective. The use of a standard protocol of cDNA preparation for both the

validation and array cohorts would have been a better way of analysing all samples

together. Additionally, RNU6B is not an ideal housekeeping gene when used alone.

RNU6B levels are so high that they are amplified even when RNA concentrations are

low, thereby nullifying the difference in expression.

Another limitation is the characterisation of normal tissue. Although the normal tissue

used in this study was referred as adjacent normal, the exact location of the tissue with

respect to tumour is not known and it could have possible caused changes, and could

thus explain why no significant difference was observed when data were normalised.

For future studies, it may be preferable to have access to tissues from microscopically

and macroscopically normal colons confirmed on colonoscopic examination and

biopsies. Previously, colonic biopsy tissues were deemed inadequate for miRNA analysis

due to poor yield of total RNA from such samples. Now, with improved protocols and

more efficient RNA extraction kits, small colonic biopsies will be adequate for tissue

miRNA analysis.

219

5.5.12 Conclusion

This study has met its aim: to examine the utility of miRNAs and common gene

mutations in FFPE cancer tissue specimens as biomarkers to predict the development

of metastasis in patients with high risk Dukes’ B cancer. Common gene mutations such

as KRAS, BRAF, PIK3CA, and discriminatory miRNAs identified from expression profiling

were successfully analysed in different stages of CRC. This study has identified that the

expression levels of miR-135b and miR-15b were significantly lower in high risk Dukes’

B cancers in comparison with low risk Dukes’ B cancers. These differential expressions

of miR-135b and miR-15b is useful, and allows these miRNAs to be used as biomarkers

for the prediction of metastasis development in patients with high risk Dukes’ B cancers.

Pooling of cDNAs for expression profiling, identification of normal healthy tissue

adjacent to cancer tissue and a small number of patients in high risk Dukes’ B group

were weaknesses of this study. In future, a large scale evaluation should be performed

to assess miR-15b and miR-135b expression in FFPE cancer and in defined paired normal

tissue specimens. Expression levels of these miRNAs should be correlated with a

prospectively maintained database with details of CRC treatments, follow ups and

outcomes.

220

Chapter 6: Final Discussion

221

6 Final Discussion

6.1 Discussion

This study has met the following aims: (i) to identify circulating miRNAs for the early

detection of colorectal neoplasia; (ii) to analyse exosomal miRNAs in the plasma of

patients with CRC; and (iii) to evaluate the utility of tissue miRNAs combined with

common gene mutations in CRC for the prediction of metastasis development in

patients with high risk Dukes’ B cancers. Strengths, weaknesses and limitations of this

study have been discussed previously in Chapters 3, 4 and 5. In this final chapter, the

discussion mainly focuses on issues surrounding the miRNA detection method and

linked strategies, and how these detection methods can be improved in future studies.

In this study, the preferential utility of plasma samples collected in heparin-free blood

collection tubes ensured that plasma remained a more suitable source of miRNAs for

detection by QRT-PCR (Mitchell et al, 2008). This is primarily due to the fact that the

coagulation process in serum can cause haemolysis, which causes additional RNA to be

released. Studies have also reported significantly higher concentrations of miRNAs in

the serum in comparison to plasma samples collected from the same individual (Wang

et al, 2012). Although extraction of miRNAs from plasma was a relatively

straightforward process, differences in the processing of blood samples to isolate

plasma may have caused haemolysis or higher concentrations of residual blood cells

which, in turn, may have contributed to artificially higher levels of miRNA levels in

samples, as shown in previous studies (Cheng et al, 2013 & Kroh et al, 2010). This is

particularly important in exosomes isolated by immunoaffinity for miRNA analysis

because other blood cells and microvesicles might carry the same surface antigens, and

could subsequently cause their precipitation and an increase in the concentration of

non-specific miRNAs. Thus, quality control mechanisms such as additional

centrifugation, the rejection of samples with platelet counts above a certain threshold

and the measurement of haemolysis have been recommended to control for biases that

might confound the measurement of miRNA levels.

FFPE tissues are widely used archive material for biomarker discovery and validation

(Lewis et al. 2001). These types of samples represent a challenge for mRNA profiling

222

because RNA degradation occurs during fixation (Macabeo-Ong et al, 2002) and storage

(Cronin et al. 2004). However, due to their stability and small size, miRNAs are better

preserved than mRNA and can be readily extracted from FFPE samples (Hui et al. 2009).

Additionally, accurate miRNA detection does not require large tissue specimens and

miRNA expression can be measured easily in biopsy specimens. Given the invasive

nature of fresh or frozen tissue collection and the availability of FFPE tissues, this

method makes measurement of miRNA levels for diagnostic purposes more feasible.

This applies specifically to specimens collected from participants undergoing NBCSP,

where making a correct diagnosis and assessing the completeness of excision of a lesion

are prioritised. Any tissue collected for snap-freezing might lead to these NBCSP

priorities becoming compromised. In this study, microdissection of FFPE tissue has been

used with tissues of interest marked by a histopathologist. Accordingly, for large cohort

studies dealing with FFPE tissue, tissue processing for miRNA analysis will need a

dedicated specialist who can review and mark FFPE tissue samples.

Before miRNAs can be used as biomarkers they must first be isolated from either plasma

or FFPE tissue sections. Total RNA extraction techniques used for blood and tissue

samples were dissimilar in this study, but both required the use of different commercial

RNA extraction kits and Tri-reagent phenol-chloroform for extraction. Several studies

have found variations in the amount and quality of RNA extracted using different kits.

Three different studies comparing three commercial kits, Qiagen’s miRNeasy, Ambion’s

miRVana Kit and Norgen’s Purification Kit, have reported a preference for the Qiagen’s

miRNeasy kit due to a higher quality and yield of miRNAs (Kroh et al, 2010, Monleau et

al, 2014, Li et al, 2012). Although the miRNeasy kit was used for RNA extraction from

FFPE tissues in this study, it has been recommended that switching to this kit for future

plasma sample processing would be useful. However, another determining factor for

selecting the isolation kit is the volume of the donor sample. There has been rapid rise

in the availability of commercial kits for miRNA extraction that allow smaller volumes

of starting material. Their efficacy and cost effectiveness must be questioned before

their use in large scale validation studies in the future. For studies in the near future,

total RNA should be extracted from 1 ml plasma sample and purified by using Tri-

223

reagent and Qiagen’s miRNeasy RNA isolation kit, and RNA quality and quantity should

be assessed with a Thermo Scientific NanoDrop 2000c spectrophotometer.

Relative quantification of miRNAs by QRT-PCR was the method of choice in this study,

with Megaplex QRT-PCR also being used with an additional step of pre-amplification.

Unlike standard QRT-PCR, this pre-amplification step allowed very low concentrations

of miRNAs to be detected. Pre-amplification and Megaplex QRT-PCR also allowed

multiple miRNAs to be studied simultaneously. Disadvantages of this strategy was that

any technical noise introduced during cDNA synthesis could cause up to 100-fold

variations in cDNA yields (Baker, 2010, Benes et al, 2010, Chugh et al, 2012, Stahlberg

et al, 2004 & Vandesompele et al, 2002). Previous studies have relied heavily upon the

use of microarray-based technologies, which often do not include the growing list of

miRNAs, and hybridization-based non-quantitative techniques can yield data that may

fail subsequent validation steps (Ng et al, 2009, Hofsli et al, 2013, Giraldez et al, 2013).

Sequencing technologies such as next-generation sequencing (NGS) for miRNA profiling

in biomarker research now facilitate the measurement of absolute abundance over

dynamic range, which was not possible using conventional microarray technology in the

past. In addition, the use of NSG will aid the identification of novel miRNAs, miRNA

sequence changes and miRNA variants such as isomiRs (Kuchenbauer et al, 2008, Guo

et al, 2011, Llorens et al, 2013, Landgraf et al, 2007, Lee et al, 2010 & Morin et al, 2008).

A recent study has reported the absolute quantities of circulating miRNAs in classical

Hodgkin lymphoma as copy numbers, which were calculated from known copy numbers

of cel-miR-39 spiked-in in plasma (Jones et al, 2014). Jones and colleagues have shown

a high correlation between relative and absolute quantification values for all evaluated

miRNAs. In future studies, this method will allow absolute values of plasma miRNAs to

be quantified by eliminating the technical bias of RNA extraction. Furthermore, Droplet

Digital PCR, a next generation RT-PCR method, provides absolute quantification of

nucleic acids without producing a standard curve, and detects copy number changes of

DNA and RNA targets with high sensitivity in comparison to RT-PCR methods (Day et al,

2013). In future studies, combining the cel-miR-39 spiking approach and new Droplet

Digital PCR technology may allow quantification of miRNA levels to almost absolute

224

values leading to identification of low abundance miRNA in whole plasma and

exosomes.

The biggest obstacle of RT-PCR-based detection of miRNAs was achieving successful

normalisation of miRNA expression with an endogenous control. Previous studies have

used miR-16, snRNA RNU6B, snRNA SNORD43, or synthetic versions of Caenorhabditis

elegans miRNAs such as cel-miR-39, cel-miR-54 and cel-miR-238, as control genes

(Kawamura et al, 2014). The spike-in C. elegans synthetic miRNAs are often used to

account for technical variability during extraction rather than for biological variability,

and thus cannot be used as endogenous controls (Kroh et al, 2010). A study evaluating

seven candidate reference genes (let-7a, miR-16, miR-93, miR-103, miR-192, miR-451,

and RNU6B) concluded that miR-16 and miR-93 are good reference genes, but deemed

RNU6B unsuitable due to its low expression levels in the serum of cancer patients (Song

et al, 2012). In this study, optimum levels of expressions for RNU6B control for both

tissue and plasma miRNAs were found. However, RNU6B expressions were significantly

lower in exosomal miRNA analysis. This might have been due to low yield of total RNA

during extraction. For miR-16, the use of RNU6B as a control gene is inappropriate

because its expression levels can be significantly affected by haemolysis and in disease

(Pritchard et al, 2012 & McDonald et al, 2011).

In this study a uniform volume of plasma samples and an equal number of tissue

sections and RNA were used for normalisation. It is worth noting that some studies have

selected their endogenous control by evaluating comparative expression of several

control candidates in sample populations (Nugent et al, 2012). Of particular concern

regarding the selection of endogenous controls is the observation that two miRNAs

selected in a pilot study as candidate endogenous controls were differentially expressed

in the validation study (Kjersem et al, 2013). Thus, it is clear that selecting a reliable

endogenous control is important for the reliability and accuracy of miRNA

quantification. This is a challenge that is well understood in the field as many studies,

including those studying circulating miRNA biomarkers, recognise this as a possible

limitation. These issues must be dealt with carefully in the discovery phase of future

projects as well as in future clinical settings to ensure miRNA quantification is accurate

(Cortez et al, 2009, Jung et al, 2012).

225

A significant challenge of using miR-135b as a biomarker is its disease specificity. Like

many other tissue and blood miRNAs (Chen et al, 2013, Ng et al, 2009, Ren et al, 2013,

Toiyama et al, 2013 & Wang et al, 2012), miR-135b lacks specificity for CRCs. This study

has made an attempt to address this limitation by directly comparing its expression in

tumour tissue and blood to clarify the interaction and origin of circulating miRNAs.

However, developing miRNA expression profiles limited to CRCs and its relevant

controls did not yield miRNAs specific for CRCs alone. This implies that in order to

increase confidence in the specificity of a miRNA-based biomarker, controls for other

cancer types need to be included in expression profiling studies (Kanaan et al, 2013). In

addition, investigations aimed at understanding the mechanism by which certain

circulating miRNAs become differentially expressed may also help pinpoint miRNAs

which are suitable markers for specific cancer types. This type of investigation would

likely provide more insight into understanding the discrepancies observed between the

levels of particular miRNAs in tumour tissue and corresponding blood samples. (Wang

et al, 2012 & Toiyama et al, 2013). Although recent studies have suggested that miRNA

expression profiles are likely influenced by age, sex and race (Takahashi et al, 2013, Bala

et al, 2012, & Witwer et al, 2012), future studies should also analyse miRNAs for

variations in lifestyle factors, such as diet, smoking and alcohol consumption, all of

which may have a profound effect on miRNA expression.

6.2 Conclusion

This study has successfully investigated the potential use of miRNAs for the early

detection of CRCs and the prediction of metastasis in high risk Dukes’ B cancers.

However, the road to producing a highly efficient, robust and standardized method for

miRNA detection in clinical practices remains a long one. Although individual miRNAs

have shown promise in the early detection of CRCs and Dukes’ B cancers at risk of

developing metastasis, an effective diagnostic, screening or prognostic test will likely

require a panel of miRNAs to improve accuracy, reliability and consistency. Circulating

and tissue miRNA-based biomarkers will eventually benefit patients after thorough

validation by multi-centre prospective studies using the most robust and uniform

methodology, including optimum sample handling, RNA isolation, quantification

protocols, reference standards and statistical analyses.

226

227

Appendices

228

7 Appendices

7.1 Appendix I: Patient information sheet for colorectal disease progression

STUDY TITLE: MARKERS OF COLORECTAL DISEASE

Local Investigators: Dr J H Pringle / Mr. Muhammad Imran Aslam

Department of Cancer Studies and Molecular Medicine

Leicester Royal Infirmary

You are being invited to take part in a research study. Before you decide it is important

for you to understand why the research is being done and what it will involve. Please

take time to read the following information carefully and discuss it with others if you

wish. Ask us if there is anything that is not clear or if you would like more information.

Take time to decide whether or not you wish to take part.

1. Why have I been chosen?

You have been chosen because you have been recently diagnosed with a cancer We are

requesting your agreement to allow us to study a portion cancer tissue that will be

removed as part of your surgical procedure as well as a sample of your blood to be taken

around the time of your procedure

2. What is the purpose of the study?

We are studying changes that occur in the bowel from a range of diseases including

inflammatory conditions and bowel tumours. We are comparing this to changes

observed in patients who have cancers of other body parts i.e lungs, pancreas, stomach,

ovaries, prostate, kidneys, bladder and breast We hope this will aid diagnosis, identify

mechanisms that are important in the development of cancers, detection of cancer and

help understand the factors that make cancers spread. This is with the aim that we can

develop diagnostic tests and new treatments to prevent spread and improve cure rates.

3. Do I have to take part?

229

It is up to you to decide whether or not to take part. If you do decide to take part you

will be given this information sheet to keep and be asked to sign a consent form. If you

decide to take part you are still free to withdraw at any time and without giving a

reason. In practice, withdrawal would mean destruction of any donated tissue samples

or blood samples and, should you also wish, any associated data. A decision to withdraw

at any time, or a decision not to take part, will not affect the standard of care you

receive.

4. What will happen to me if I take part?

The tissue removed during your procedure is always examined by a pathologist to

identify the changes present. If you agree to take part in our studies, following this

routine examination, further small pieces of tissue will be selected for our studies, with

4 to 6 samples collected in total. This tissue would otherwise be discarded, and its

selection will not alter the routine assessment of your tissue. Since not all of the sample

will be used in this study we also request that we can store the sample for further similar

studies (see attached ‘Tissue Bank Information Sheet’). Some tissue will also be used

from archived material in a tissue bank in order to make sure that there is a

representative sample size for each aspect of disease. Use of archived tissue also allows

us to gather samples for which we have long term follow up data that would otherwise

not be possible if all samples were collected prospectively. The blood samples will be

collected in small tubes in the usual way and will be destroyed when the study is

complete and you have not consented for its storage in the Tissue Bank.

5. What are the possible disadvantages and risks of taking part?

If you chose to take part, the study will involve selecting samples of the diseased tissue

that has been removed as a part of your procedure. This will take place following the

examination that is always carried out on surgically removed tissue and will in no way

alter how your tissue will be treated.

Because we also require a blood test, the risks are limited to discomfort at the site from

where the blood is taken.

6. What are the possible benefits of taking part?

There is no benefit to you personally from taking part in this study. However, we hope

that the information we get may allow us to develop new treatments to prevent spread

and improve cure rates for bowel cancer.

230

7. What if new information becomes available?

We will not be performing any tests that have an influence on your care. It is therefore

unlikely that the study will yield any new information that will affect you personally.

8. What if something goes wrong?

The chance of any problem arising because of your inclusion in the study is very small.

However, if you are harmed by taking part in this research project, there are no special

compensation arrangements. If you are harmed due to someone’s negligence, then

you may have grounds for a legal action but you may have to pay for it. Regardless of

this, if you wish to complain, or have any concerns about any aspect of the way you

have been approached or treated during the course of this study, the normal National

Health Service complaints mechanisms would be available to you.’

9. Will my taking part in this study be kept confidential?

All information which is collected about you during the course of the research will be

kept strictly confidential. Any information about you which leaves the hospital will have

your name and address removed so that you cannot be recognised from it.

10. What will happen to the results of the research study?

The results from this study will be presented at scientific meetings and be published in

scientific journals. You will not be identified in any report/publication.

11. Who is organising and funding the research?

This study is a small-scale study that is being supported by the researchers University

departmental funds. The researchers will not receive extra payments for performing

this study.

12. Who has reviewed the study?

All research that involves NHS patients or staff, information from NHS medical records

or uses NHS premises or facilities must be approved by an NHS Research Ethics

231

Committee before it goes ahead. Approval does not guarantee that you will not come

to any harm if you take part. However, approval means that the committee is satisfied

that your rights will be respected, that any risks have been reduced to a minimum and

balanced against possible benefits and that you have been given sufficient information

on which to make an informed decision’.

13. Contact for Further Information

For further information, please contact:

Dr. Howard Pringle,

Department of Cancer Studies and Molecular Medicine,

University of Leicester,

Leicester

LE2 7LX

E-mail: JHP@le.ac.uk

Phone: 0116 252 3227

Thank you for reading this.

Please keep this copy of the Information Sheet to refer to in future. If you agree to take

part in the study, you will also receive a copy of the signed consent form to keep.

232

7.2 Appendix II: Tissue bank patient information sheet

Title of the proposed tissue bank: Colorectal Tissue Bank

Location tissue bank:

RKCSB

Leicester Royal Infirmary

Infirmary Square Leicester

LE2 7LX

What is a tissue bank?

A tissue bank is a collection of tissue and blood samples being stored long term that will

be used to assist in future research into a specific disease or group of diseases or

investigate disease processes and their treatment. Tissue banks are a valuable research

resource and are increasingly being established at local, regional and national level.

Why have I been chosen?

You have been asked to participate with the Colorectal tissue bank because you are

recently diagnosed with a bowel cancer or a cancer of other part of body.As you are

due to undergo an either an investigation or surgical procedure as part of the

management recommended by the consultant surgeon responsible for your care. We

use this opportunity to invite you to donate your tissue and blood samples to colorectal

tissue bank.

What will the tissues in the tissue bank be used for?

The tissues will be used to help increase our understanding of bowel cancer and to more

accurately predict how the different individual tumours will behave. We will be studying

how tumours spread and how this may be blocked. We will study other diseases that

occur in the bowel to understand what factors are unique to cancer, and whether new

treatments could be produced against them.All research that involves NHS patients or

staff, information from NHS medical records or uses NHS patients or staff, information

from NHS medical records or uses NHS premises or facilities must be approved by an

233

NHS Research Ethics Committee before it goes ahead. Approval does not guarantee

that you will not come to any harm if you take part. However, approval means that the

Committee is satisfied that your rights will be respected, that any risks have been

reduced to a minimum and balanced against possible benefits and that you have been

given sufficient information on which to make an informed decision to take part or not.

How much of my tissue will be taken?

During your surgical procedure the diseased tissue will be removed by the surgeon.

Following a routine examination by the pathologist, further small pieces of tissue will

be selected for our tissue bank with 4 to 6 samples collected in total. This tissue would

otherwise be discarded, and its selection will not alter the routine assessment of your

tissue. We will also store any blood samples collected at the time of this procedure.

Will I be contacted again in the future?

If, as a result of any research carried out on your tissue, new information becomes

available which may have an impact on your care, this information will be discussed

with you either by your hospital Consultant or General Practitioner. We will contact you

again to seek permission to use your tissue samples if an unanticipated use emerges in

the future, which is not described in this information sheet.

Who will have access to my tissue and how will confidentiality be maintained?

Access to your tissue samples will be only available through the Colorectal Tissue Bank,

controlled by the University Hospitals of Leicester NHS Trust.

The handling of your tissue samples will be treated with the usual degree of

confidentiality under the data protection act. All samples will be anonymised before

transfer to other research partners - you will not be identified in any way from your

tissue and blood sample. Basic clinical information regarding the reason for your

procedure, you age, sex and the results of the pathological examination of the tissue

will be linked to the samples. This will not involve using your name or address.

Will I receive payment for the tissue that I donate to the tissue bank?

234

You will not receive any payment for the tissue or blood. The tissue is a gift - neither

yourself nor your relatives will benefit from any inventions that result from the use of

the tissue.

What happens if I wish to have my tissue removed from the tissue bank?

If you do not wish your tissues and blood to be held in the tissue bank you withdraw

them without justifying your decision and your future treatment will not be affected.

If you wish to have your tissue removed from the tissue bank please contact:

Mr Muhammad Imran Aslam or Dr. J.H. Pringle

Department of Cancer Studies

Robert Kilpatrick Clinical Sciences Building

Leicester Royal Infirmary

Leicester LE2 7LX

Tel: +44 116 2523227

Or:

Research Office

Directorate of Research & Development

University Hospitals of Leicester NHS Trust,

Leicester General Hospital,

Gwendolen Road,

Leicester LE5 4PW

Tel: 0116 258 4109

235

7.3 Appendix III: Consent form

Centre Number: __________

Study Number: ___________

Patient Identification Number for this trial: _________

CONSENT FORM

Title of Project: MARKERS OF COLORECTAL DISEASE

Name of Researcher / Principal Investigator:

Mr Muhammad Imran Aslam / Dr James Howard Pringle

This form should be read in conjunction with the Patient Information

Leaflet, version no 6 dated 08.12.2009

Please initial box

1. I confirm that I have read and understand the information sheet for the above study and have had the opportunity to ask questions.

2. I understand that I may withdraw my consent to my tissue and blood sample being used at any time without justifying my decision and without affecting my normal care and medical management.

1.1.1.1.1.1.1 Patient name, address, DOB (or

ID label)

236

3. I agree to donate tissue from my procedure and blood samples and allow their use in medical research as described in the Patient Information Leaflet.

4. I understand that the tissue and blood sample are gifts and that I will not benefit from any intellectual property that results from its use.

5. I understand that the tissue or blood sample will not be used to undertake any genetic tests whose results may have adverse consequences on my or my families insurance or employment.

6. I understand that if research using my tissue or blood sample produces information, which has immediate clinical relevance to me, I will be contacted by my hospital consultant or GP to discuss how this may affect my treatment or follow up.

7. I understand that blood samples and associated clinical data may be transferred to commercial / non-commercial research partners of the University Hospitals of Leicester NHS Trust, but that the information will be coded prior to transfer.

8. I agree to take part in the above study.

________________________ ________________ ___________________

Name of Patient Date Signature

_________________________ ________________ ___________

Researcher Date Signature

1 for patient; 1 for researcher; 1 to be kept with hospital notes

237

7.4 Appendix IV: Consent form for colorectal tissue bank

PATIENT CONSENT FORM

Colorectal Tissue Bank

Tissue Bank Custodian: Dr. C Richards and Mr J Jameson

Patient’s details:

Name:

D.O.B.

Address:

Hospital No.

This form should be read in conjunction with the Patient Information

Leaflet, version no 7 dated 08.12.2009 Please initial box

9. I agree to donate the tissue samples as identified below to the Colorectal Tissue Bank and allow their use in medical research as described in the Patient Information Sheet entitled Colorectal Tissue Bank, Version 6 dated 1.10.2008

10. I understand that I may withdraw my consent to my tissue and blood sample being used at any time without justifying my decision and without affecting my normal care and medical management.

11. I understand that members of University Hospitals of Leicester NHS Trust and Leicester University research teams may wish to view relevant sections of my medical records, but that all the information will be treated as confidential.

12. I understand that samples from the tissue bank and associated clinical data may be transferred to non-commercial research partners of the University Hospitals of Leicester NHS Trust and Leicester University, but that the information will be anonymised prior to transfer.

238

13. I understand medical research is covered for mishaps in the same way, as for patients undergoing treatment in the NHS i.e. compensation is only available if negligence occurs.

14. I understand that samples from the tissue bank will not be used to undertake any genetic tests whose results may have adverse consequences on my or my families insurance or employment.

15. I understand that if research using my tissues produces information, which has immediate clinical relevance to me, I will be informed by my hospital consultant or GP and be given an opportunity to discuss the results.

16. I understand that the tissue is a gift and that I will not benefit from any intellectual property that results from the use of the tissue.

17. I would be willing to be contacted again regarding future use of this tissue for purposes not foreseen at the present time.

I have read the patient information leaflet relating to the Colorectal Tissue Bank

and have had the opportunity to ask any questions.

Signature of patient

........................................………………………......Date......................................

(Name in BLOCK LETTERS)

..................................................................................................………………….I

confirm I have explained the purpose of the tissue bank, as detailed in the

Patient Information Sheet, in terms, which in my judgement are suited to the

understanding of the patient.

Signature of individual taking consent

...........................................…………………………Date......................................

(Name in BLOCK LETTERS)

239

7.5 Appendix V: NBSCP screening approval

Bowel Cancer Screening Programme

17 July 2011

Dear Mr Aslam The Bowel Cancer Screening Programme (BCSP) Research Committee met on 29 June

2011 to discuss your research plans: Circulating MicroRNAs are Novel Biomarker

for Colorectal Cancer Screening.

The Research Committee considered this project to be exciting.

The Committee give their support to the project.

The Committee would like you to be aware that as a trial called seAFOod is taking place

at the same centre, careful management should be undertaken so as not to overburden

patients. You should also manage the risk of contamination from this chemoprevention

study.

This letter of support can be used as permission to gain the relevant access to the Screening Programme.

As a condition of our support, the BCSP Research Committee require you to keep us

informed of developments with the project including

• when the research project has started • which (if any) Bowel Cancer Screening Centres are taking part in the

project • when fully recruited • any change of status

Research Committee

Professor John Scholefield Gillian Liddington Chair Administrator Mr. M. I. Aslam mia7@leicester.ac.uk

NCRI Clinical Studies Groups Secretariat Angel Building

407 St John Street London

EC1V 4AD

Tel: 020 3469 8533

gillian.liddington@cancer.org.uk

240

• any significant adverse reactions • when complete and • when written up (including a copy of your findings for the Committee)

The Research Committee requires that you notify them of any incidents that would be

recorded on the National Research Ethics Service (NRES) Breaches Register.

Undertaking research within the Screening Programme further to this letter of support

assumes your agreement to fulfil this obligation. NRES will share information with

BCSP Research Committee regarding any breaches. We wish you well with your

research.

Yours sincerely

Gillian Liddington

On behalf of the NHS Bowel Cancer Screening Programme Research Committee

241

7.6 Appendix VI: Plasma array participants characteristics

H45/09 M 67 Duke's A Sigmoid Adenocarcinoma, pT2,pN0,RM0

H56/09 M 75 Duke's A Sigmoid Adenocarcinoma,pT2,pN0,RM0

H276/09 M 60 Dukes' A Rectal Adenocarcinoma , pT3,pN2,pM1

H43/09 F 73 Dukes' B Caecal Adenocarcinoma,pT3,pN0,RMo

H333/08 F 86 Dukes' B, Rectal Adenocarcinoma rectum ,pT2, pN0,M0

H166/09 M 73 Dukes' B, Rectal Adenocarcinoma, pT3, pN0, pMx,

H281/09 F 77 Duke's B, Ascending Colon Adenocarcinoma,pT3 pN0pMx

H176/09 F 48 Dukes B, Asecending Colon adenocarcinoma, pT3 pN0? pM1

H513/08 M 72 Dukes' C1, Caecal Adenocarcinoma T4N2Mx

H399/08 F 83 Dukes' C1, Rectal Adenocarcinoma,pT3,pN2,RMx

H287/09 M 69 Dukes' C2, Caecal Adenocarcinoma,pT3,pN2,pM1

H175/09 M 58 Dukes' C2, Sigmoid Adnenocarcinoma, pT4, pN2 pMx,

H339/08 M 62 Tubular adenoma with low grade dysplasia

H510/08 F 92 Tubulovillous adenoma low grade dysplasia

H516/08 F 72 Tubular adenoma with low grade dysplasia

H58/09 M 47 Multiple tiny tubular adenomas with low grade dysplasia

H169/09 M 64 Tubulovillous with low grade dysplasia

H170/09 F 56 Tubular Adenoma with low grade dysplasia

H180A/09 F 80 Tubulovillous adenoma with moderate dysplasia

H177/09 M 74 Tubulovillous adenoma with high-grade dysplasia

H518/08 M 58 Tubulovillous adenoma with severe dysplasia

H341/08 M 79 No Polyp or Cancer on Colonoscopy

H400A/08 F 80 No Polyp or Cancer on Colonoscopy

H400/08 M 45 No Polyp or Cancer on Colonoscopy

H515/08 F 86 No Polyp or Cancer on Colonoscopy

H520A /08 M 79 No Polyp or Cancer on Colonoscopy

H396/08 F 66 No Polyp or Cancer on Colonoscopy

H174/09 F 52 No Polyp or Cancer on Colonoscopy

H271/09 F 83 No Polyp or Cancer on Colonoscopy

H274/09 M 63 No Polyp or Cancer on Colonoscopy

H60/09 M 71 No Polyp or Cancer on Colonoscopy

H400A/08 F 80 No Polyp or Cancer on Colonoscopy

242

7.7 Appendix VII: Table of Z-Score calculated for both adenoma and carcinoma

expressions from miRNA data

Detector Adenoma (A) Z-Score Carcinoma (C) Z-Score Total (n) Pick Up Rates %

>

1.96

< -

1.50

(n)

detected

>

1.96

< -

1.50

(n)

detected A + C A C Total

hsa-miR-502-5p 4 0 4 9 0 9 13 44 75 62

hsa-miR-192* 2 0 2 7 0 7 9 22 58 43

hsa-miR-564 5 0 5 4 0 4 9 56 33 43

hsa-miR-410 3 0 3 5 0 5 8 33 42 38

hsa-miR-431 2 0 2 5 0 5 7 22 42 33

hsa-miR-200a* 2 0 2 5 0 5 7 22 42 33

hsa-miR-135b 4 0 4 3 0 3 7 44 25 33

hsa-miR-139-5p 1 5 6 4 1 5 11 67 42 52

hsa-miR-182 1 2 3 4 1 5 8 33 42 38

hsa-miR-186 1 2 3 4 1 5 8 33 42 38

hsa-miR-23b 1 0 1 4 0 4 5 11 33 24

hsa-miR-139-5p 1 5 6 4 1 5 11 67 42 52

hsa-miR-182 1 2 3 4 1 5 8 33 42 38

hsa-miR-221 1 4 5 3 2 5 10 56 42 48

hsa-miR-942 1 1 2 3 2 5 7 22 42 33

hsa-miR-16 4 1 5 0 1 1 6 56 8 29

hsa-miR-203 4 0 4 0 0 0 4 44 0 19

243

7.8 Appendix VIII: Clinicopathological characteristics of individual participants in

control, adenoma and cancer groups

Groups H / number Age Gender Diagnosis/Characteristics

Controls

H334/08/MIA 74 F Normal Colonoscopy and biopsies

H335/08 65 F Normal Colonoscopy and biopsies

H338/08 39 F Normal Colonoscopy and biopsies

H339/08 61 M Normal Colonoscopy

H341/08 80 M Normal Colonoscopy

H340/08 72 M Normal Colonoscopy and biopsies

H391/08 69 M Normal Colonoscopy and biopsies

H396/08 65 F Normal Colonoscopy and biopsies

H400A/08 79 F Normal Colonoscopy

H400/08 44 M Normal Colonoscopy

H515/08 85 F Normal Colonoscopy

H518/08 57 M Normal FOS -Post Total colectomy 2000, iloanal

pouch,

H519/08 66 F Normal Colonoscopy

H520/08/A 78 M Normal Colonoscopy and biopsies

H48/09 58 M Excision of polyp 1/12ago, Now Colonoscopy

normal

H54/09 63 M Normal Colonoscopy

H57/09 69 M Normal Colonoscopy

H59/09 74 F Normal Colonoscopy

H60/09 71 M Normal Colonoscopy

H171/09 58 F Normal Colonoscopy – mild erythema

H174/09 52 F Normal Colonoscopy

H179/09 57 F Normal Colonoscopy

H180/09 54 M Normal Colonoscopy

H271/09 83 F Normal Colonoscopy

H272/09 54 M Normal Colonoscopy

H273/09 48 M Normal Colonoscopy

H274/09 63 M Normal Colonoscopy

H275/09 46 M Normal Colonoscopy, Previous Right

Hemicolectomy

H276/10 23 F Ulcerative Colitis

244

H659/10 72 M Colovesical Fistula

H956/10 74 M Diverticular abscess – follow up colonoscopy

normal

Adenomas

H334/08 56 M Colonic Tubular adenoma with low grade

dysplasia

H393/08 77 F Colonic Hyperplastic polyp

H395/08 72 M Colonic Hyperplastic polyp

H398/08 66 M Tubulovillous Adenoma

H510/08 90 F Tubulovillous Adenoma

H514/08 72 M Sigmoid Tubulovillous adenoma with low grade

dysplasia

H516/08 71 F Caecal Tubular Adenoma with low grade

dysplasia

H520/08/B 75 F Rectal Tubulovillous Adenoma with low grade

dysplasia

H46/09 52 M Rectal Adenoma with low grade dysplasia

H58/09 47 M Multiple tiny tubular Adenomas with low grade

dysplasia

H167/09 86 M Rectal Tubulovillous Adenoma with low garde

dysplasia

H169/09 64 M Tubulovillous Adenoma with low grade dysplasia

H170/09 56 F Tubular Adenoma with low grade dysplasia

H172/09 67 M Sigmoid tubular Adenoma with low grade

dysplasia

H177/09 74 M Tubulovillous adenoma with high-grade dysplasia

H178/09 67 M Descending colonic Adenomas with low grade

dysplasia

H180A/09 80 F Tubulovillous Adenoma with moderate dysplasia

H260/10 79 M Rectal Adenoma with low grade dysplasia

H261/10 88 F Rectal Adenoma with low grade dysplasia

H267/10 78 F Rectal Adenoma with low grade dysplasia

H273/10 57 F Rectal Adenoma with low grade dysplasia

H658/10 56 M Rectal Adenoma with low grade dysplasia

H663/10 60 M Rectal Adenoma with low grade dysplasia

H666/10 63 F Rectal Adenoma with low grade dysplasia

H672/10 82 F Rectal Adenoma with moderate to high grade

dysplasia

H679/10 62 M Sigmoid Adenoma with high grade Dysplasia

H950/10 38 F Sigmoid Adenoma with high grade Dysplasia

H960B/10 68 M Rectal Adenoma with low grade dysplasia

Carcinomas

H513/08 71 M rT4,N2,M0, Caecal Carcinoma

H399/08 82 F rT3,N2,M0, Rectal Carcinoma,

H43/09 73 F rT3,N0,M0 , Caecal carcinoma,

245

H45/09 67 M rT2,N0,M0 , Sigmoid Colon Carcinoma ,

H56/09 74 M rT2,N0,M0 ,.Sigmoid Colon Carcinoma,

H164/09 70 M rT3,N0,M0., Ascending Colon Carcinoma

H166/09 73 M rT3,N0,M0 , Rectal Carcinoma, ,

H175/09 58 M rT4,N2,M0 , Sigmoid Colon Carcinoma, ,

H176/09 47 F rT3 N0?M1 , Ascending Colon Carcinoma,

H276/09 60 M rT3,N2,M1 , Rectal Carcinoma

H279/09 78 M rT3,N0,M0 , Sigmoid Colon Carcinoma

H281/09 77 F rT3,N0,M0 , Caecal/Ascending Colon Carcinoma

H287/09 69 M rT3,N2,M1 , Caecal Carcinoma

H293/09 50 M rT2,N0,M0 , Rectal Carcinoma - short course

radiotherapy

H265/10 87 M rT2,N0,M0 , Caecal Carcinoma

H266/10 65 M rT3,N1/N2,M0 , Recto-Sigmoid Carcinoma

H268/10 64 M rT2,N0,M0 , Ascending Colon Carcinoma

H275/10 76 M rT3,?N1,M0 , Ascending Colon Carcinoma

H277/10 66 M rT3,N1,M0 , Recto-Sigmoid Colon

H278/10 69 F rT3,N1,M0 , Caecal Carcinoma

H279/10 60 M rT3,N1,M0 , Rectal Carcinoma

H288/10 64 M rT2,N0,M0 ., Rectal Carcinoma- post

radiotherapy

H300B/10 82 M rT3,N0,M0 , Caecal/Ascending Colonic carcinoma

H300A/10 69 M rT3,N1,M0 , Sigmoid Colon Carcinoma

H660/10 75 M rT2,N0,M0., Rectal Carcinoma

H662/10 48 F rT2,N1,M0., Rectal Carcinoma, short course

radiotherapy

H664/10 68 M rT3,N1, M0 , Transverse Colon Carcinoma

H673/10 64 F rT3,N1MO , Sigmoid Colon Carcinoma

H674/10 69 M rT3,N1MO , Rectal Carcinoma

H678/10 66 M rT3,N1,M0 , Rectal Carcinoma- long course

radiotherapy

H680/10 82 F rT3,N0,M0 , Recto-Sigmoid carcinoma

H680A/10 71 M rT3,N1,M1 , Ascending Colonic carcinoma

H951/10 63 M rT3,N0,M0., Rectal Carcinoma

H952/10 73 M rT3/T4,N1,M0 , Sigmoid Colon Carcinoma

246

7.9 Appendix IX: Subgroup of controls for initial validation cohort.

Grouping is based on previous significant history of neoplasia, or non-neoplastic disease.

Symptomatic, Normal Colonoscopy ,No Significant Bowel Disease History

Specimen No Patient Characteristics

H339/08 Normal Colonoscopy & Biopsy

H519/08 Normal Colonoscopy & Biopsy

H334/08/MIA Normal Colonoscopy & Biopsy,

H335/08 Normal Colonoscopy & Biopsy

H179/09 Normal Colonoscopy & Biopsy

H338/08 Normal Colonoscopy & Biopsy

H272/09 Normal Colonoscopy & Biopsy

H400/08 Normal Colonoscopy & Biopsy

H515/08 Normal Colonoscopy & Biopsy

H520/08/A Normal Colonoscopy & Biopsy

H57/09 Normal Colonoscopy & Biopsy

H396/08 Normal Colonoscopy & Biopsy,

H400A/08 Normal Colonoscopy & Biopsy

H391/08 Normal Colonoscopy & Biopsy

H518/08 Iloanal Pouch – flexible sigmoidoscopy normal

Symptomatic, Normal Colonoscopy, A Significant Bowel Neoplasia Disease History

H275/09 Previous Right Hemicolectomy +Chemotherapy

H395/08 Severely dysplastic Rectal Adenoma excised 18 months prior to colonoscopy

H54/09 Sigmoid volvulous

H180/09 Right Hemicolectomy 18 years ago, APC mutation +ve

H60/09 Right Hemicolectomy 9 years ago for severely dysplastic polyp.

H174/09 Polyp excised 5 years ago

H48/09 Excision of polyp 1 month ago, now completion colonoscopy normal.

H271/09 APER for rectal Cancer 5 Years ago

Symptomatic, Normal Colonoscopy, Current/Previous Significant Bowel Disease

H341/08 Diverticulitis, diverticulosis, AAA repair

H273/09 Colonoscopy after diverticular abscess

H393/08 History of colitis -? Crohn’s colitis

H59/09 Previous collageneous colitis 15 years ago

H276/10 Ulcerative Colitis

H659/10 Colovesical Fistula

H171/09 Collageneous collitis + strong FH of CRC

H340/08 Severe Diverticular disease

H956/10 Diverticular Disease with abscess

247

7.10 Appendix X: Initial Validation Cohort: Concentrations of Total RNA as detected

on Nanodrop ND-1000 Spectrophotometer

Serial Number H / Number Total RNA concentration 260/280 Peak

1 H334/08 4.9 1.15 270

2 H393/08 8.6 1.06 275

3 H395/08 9.8 1.24 270

4 H398/08 6.4 1.05 275

5 H510/08 4.5 2.02 270

6 H514/08 9.2 0.92 270

7 H516/08 11.9 0.91 275

8 H520/08/B 7.2 1.18 270

9 H46/09 5 1.04 270

10 H58/09 6 1.49 265

11 H167/09 3.5 1.47 270

12 H169/09 5.4 1.4 270

13 H170/09 2.6 1.57 270

14 H172/09 15.9 1.4 265

15 H177/09 3.6 1.41 270

16 H178/09 13.9 1.29 270

17 H180A/09 2.1 1.02 270

18 H260/10 13.5 1.51 270

19 H261/10 25.9 1.47 270

20 H267/10 4.6 1.55 275

21 H273/10 14.1 1.53 270

22 H658/10 3.9 2.18 275

23 H663/10 3 0.95 275

24 H666/10 5.6 1.42 270

25 H672/10 26.1 1.39 270

25 H679/10 21.8 1.44 270

27 H950/10 5.6 1.4 265

28 H960B/10 15.9 1.32 270

29 H334/08/MIA 2.3 1.52 270

30 H335/08 7 1.15 270

31 H338/08 6.5 1.24 270

32 H339/08 4.4 1.46 270

33 H341/08 8.2 1.84 265

34 H340/08 4.9 1.65 265

35 H391/08 4.6 1.22 270

36 H396/08 15.2 0.73 275

37 H400A/08 4.7 0.93 270

38 H400/08 6.3 1.24 270

39 H515/08 8.5 1.1 265

40 H518/08 4.8 1.18 270

41 H519/08 3.3 0.99 275

42 H520/08/A 9.2 1.17 270

248

Serial Number H / Number Total RNA concentration 260:280 Peak

43 H48/09 3.1 1.72 270

44 H54/09 25.1 1.31 270

45 H57/09 2.4 1.21 270

46 H59/09 17.1 1.15 270

47 H60/09 2.1 1.49 270

48 H171/09 16 1.47 270

49 H174/09 5.9 1.28 270

50 H179/09 16.2 1.2 265

51 H180/09 2.7 1.66 270

52 H271/09 26.8 1.63 270

53 H272/09 22.7 1.3 270

54 H273/09 24.3 1.37 270

55 H274/09 13 1.53 270

56 H275/09 13.5 1.31 270

57 H276/10 12.6 1.32 270

58 H659/10 4.2 2.37 275

59 H665/10 3.1 1.37 260

60 H956/10 5 1.47 265

61 H513/08 11.3 1.18 270

62 H399/08 7.7 1 275

63 H43/09 5.2 1.14 270

64 H45/09 4.7 2.23 270

65 H56/09 4.2 1.4 270

66 H164/09 8.9 1.65 270

67 H166/09 6.1 1.57 270

68 H175/09 5.1 1.23 270

69 H176/09 2.9 2.29 265

70 H276/09 12.8 1.56 270

71 H279/09 1.9 1.98 265

72 H281/09 24.1 1.66 270

73 H287/09 16 1.69 270

74 H293/09 11.6 1.55 265

75 H265/10 7.4 1.04 270

76 H266/10 13.8 1.54 270

77 H268/10 3.5 1.39 275

78 H275/10 12.7 1.63 270

79 H277/10 11.7 1.54 270

80 H278/10 6.9 0.94 260

81 H279/10 2.8 1.98 270

82 H288/10 3.2 1.67 270

83 H300B/10 11.2 1.57 270

84 H300A/10 25.1 1.7 270

85 H660/10 3 2.23 275

86 H662/10 5.4 1.35 270

87 H664/10 4.9 1.31 270

249

Serial Number H / Number Total RNA concentration 260:280 Peak

88 H673/10 3.6 1.71 270

89 H674/10 3.9 1.44 270

90 H678/10 4.5 0.97 265

91 H680/10 3.1 2.62 270

92 H680A/10 4.6 1.75 260

93 H951/10 5 0.78 270

94 H952/10 28.1 1.2 270

250

7.11 Appendix XI: CT values based expression levels of different miRNAs analysed

for initial validation cohort

Average expression levels and standard deviations are given for control, adenoma and

carcinoma groups. 2 tailed non matched pairs student’s t test is used for calculation of

p-values.

miRNAs

Normal/ Controls Average

STD Adenoma Average

STD Carcinoma Average

STD

Normal vs. Carcinoma

Normal vs. Adenoma

T-Test T-Test

(p-value) (p-value)

Megaplex Pool A miRNA miRNA expressions

miR-135b 39.59 1.70 36.81 3.15 34.98 3.47 <0.0001 0.0030

miR-191 19.04 2.61 17.46 2.86 15.65 2.45 0.0001 0.0956

miR-369-5p 37.30 3.70 34.48 4.17 32.05 4.63 0.0001 0.0415

miR-502-5P 30.57 2.18 29.52 3.56 29.12 3.82 0.0780 0.2970

miR-203 30.99 2.81 29.93 3.51 28.32 2.68 0.0024 0.3297

miR-34a 35.68 3.09 32.07 4.02 30.11 3.10 <0.0001 0.0054

miR-17 23.10 4.10 20.06 3.40 17.92 3.03 0.0001 0.0234

miR-484 21.38 2.64 20.00 2.91 18.21 2.54 0.0002 0.1526

miR-95 36.82 3.79 34.70 4.71 32.02 3.69 0.0001 0.1514

miR-195 27.78 3.18 25.66 2.80 24.56 2.75 0.0012 0.0450

miR-205 37.31 3.46 36.37 4.32 34.44 3.83 0.0092 0.4781

miR-23b 32.04 5.24 29.00 5.25 26.23 4.74 0.0005 0.0958

miR-21 21.85 2.69 19.65 2.97 18.40 2.56 0.0001 0.0283

miR-566 38.63 2.64 36.41 3.52 35.68 4.16 0.0025 0.0418

miR-486-5P 28.76 4.39 26.56 3.63 25.19 3.33 0.0061 0.1161

miR-589 34.85 3.77 33.26 4.63 30.47 3.39 0.0003 0.2737

miR-431 30.29 4.44 27.04 4.53 23.75 4.35 <0.0001 0.0396

miR-92a 19.27 2.12 18.65 2.22 17.56 1.88 0.0079 0.3983

miR-410 30.08 3.76 27.98 3.07 26.09 2.81 0.0006 0.0806

miR-487b 32.15 4.08 29.55 4.17 27.23 3.48 0.0002 0.0711

miR-16 26.23 4.39 23.15 3.79 21.40 3.43 0.0005 0.0340

miR-486-3P 25.64 3.12 25.09 3.25 23.83 2.77 0.0482 0.6086

251

miR-532-5P 29.97 3.38 28.10 3.73 26.66 3.47 0.0021 0.1295

RNU6B 23.36 1.08 22.84 1.27 22.20 1.34 0.0016 0.2030

Sample Name

Normal/ Controls Average

STD Adenoma Average

STD Carcinoma Average

STD

Normal vs.

Carcinoma

Normal vs.

Adenoma

T-Test T-Test

(p-value) (p-value)

Megaplex Pool B miRNA expressions

miR-192* 39.23 2.23 37.54 3.80 35.42 3.80 <0.0001 0.1181

miR-200a* 39.90 0.40 39.05 1.92 37.78 2.85 <0.0001 0.0798

miR-624* 34.33 2.11 33.37 2.76 32.28 2.61 0.0036 0.2564

miR-181C* 34.36 4.83 31.92 5.96 29.80 5.06 0.0031 0.1917

miR-182* 39.95 0.19 39.07 2.20 38.57 2.48 0.0013 0.1086

miR-592 38.84 2.00 38.53 2.73 37.67 3.23 0.1060 0.7034

miR-566 32.53 3.33 31.72 3.45 30.53 2.46 0.0357 0.4876

RNU6B 23.17 1.00 22.66 1.17 22.27 1.34 0.0083 0.1797

252

7.12 Appendix XII: ∆CT values for expression levels of different miRNAs analysed for

initial validation cohort

Average expression levels and standard deviations are given for control, adenoma and

carcinoma groups. 2 tailed non matched pairs student’s t test is used for calculation of

p-values.

miRNAs

Normal/ Controls Average

STD Adenoma Average

STD Carcinoma Average

STD

Normal vs.

Carcinoma

Normal vs.

Adenoma

T-Test T-Test

(p-value) (p-value)

Megaplex Pool A miRNA miRNA expressions

miR-135b 18.33 -2.78 -4.61 0.00 0.00 3.47 <0.0001 0.0030

miR-191 -2.22 -1.59 -3.39 0.00 0.10 2.45 0.0001 0.0956

miR-369-5p 16.04 -2.82 -5.25 0.00 0.04 4.63 0.0001 0.0415

miR-502-5P 9.31 -1.05 -1.45 0.08 0.30 3.82 0.0780 0.2970

miR-203 9.73 -1.06 -2.67 0.00 0.33 2.68 0.0024 0.3297

miR-34a 14.42 -3.61 -5.57 0.00 0.01 3.10 <0.0001 0.0054

miR-17 1.84 -3.04 -5.18 0.00 0.02 3.03 0.0001 0.0234

miR-484 0.12 -1.37 -3.17 0.00 0.15 2.54 0.0002 0.1526

miR-95 15.56 -2.12 -4.80 0.00 0.15 3.69 0.0001 0.1514

miR-195 6.52 -2.12 -3.22 0.00 0.04 2.75 0.0012 0.0450

miR-205 16.05 -0.95 -2.88 0.01 0.48 3.83 0.0092 0.4781

miR-23b 10.78 -3.04 -5.81 0.00 0.10 4.74 0.0005 0.0958

miR-21 0.59 -2.20 -3.44 0.00 0.03 2.56 0.0001 0.0283

miR-566 17.37 -2.22 -2.95 0.00 0.04 4.16 0.0025 0.0418

miR-486-5P 7.50 -2.21 -3.57 0.01 0.12 3.33 0.0061 0.1161

miR-589 13.59 -1.59 -4.38 0.00 0.27 3.39 0.0003 0.2737

miR-431 9.03 -3.25 -6.54 0.00 0.04 4.35 <0.0001 0.0396

miR-92a -1.99 -0.63 -1.71 0.01 0.40 1.88 0.0079 0.3983

miR-410 8.82 -2.10 -3.98 0.00 0.08 2.81 0.0006 0.0806

miR-487b 10.89 -2.60 -4.92 0.00 0.07 3.48 0.0002 0.0711

miR-16 4.97 -3.08 -4.82 0.00 0.03 3.43 0.0005 0.0340

253

miR-486-3P 4.39 -0.56 -1.81 0.05 0.61 2.77 0.0482 0.6086

miR-532-5P 8.71 -1.87 -3.31 0.00 0.13 3.47 0.0021 0.1295

Sample Name

Normal/ Controls Average

STD Adenoma Average

STD Carcinoma

Average STD

Normal vs.

Carcinoma

Normal vs.

Adenoma

T-Test T-Test

(p-value) (p-value)

Megaplex Pool B miRNA expressions

miR-192* 18.03 -0.52 -1.15 0.00 0.20 1.34 0.0016 0.2030

miR-200a* 18.71 -1.69 -3.80 0.00 0.12 3.80 <0.0001 0.1181

miR-624* 13.14 -0.86 -2.12 0.00 0.08 2.85 <0.0001 0.0798

miR-181C* 13.16 -0.96 -2.05 0.00 0.26 2.61 0.0036 0.2564

miR-182* 18.76 -2.44 -4.56 0.00 0.19 5.06 0.0031 0.1917

miR-592 17.64 -0.88 -1.38 0.00 0.11 2.48 0.0013 0.1086

miR-566 11.33 -0.31 -1.17 0.11 0.70 3.23 0.1060 0.7034

254

7.13 Appendix XIII: ∆∆CT values for expression levels of different miRNAs analysed

for initial validation cohort

Average expression levels and standard deviations are given for control, adenoma and

carcinoma groups. 2 tailed non matched pairs student’s t test is used for calculation of

p-values.

Sample Name Adenoma vs. Normal ∆∆CT

Cancer vs. Normal ∆∆CT

Normal vs. Tumour T-Test

(p-value)

Normal vs. Adenoma T-Test

(p-value)

miR-135b -2.78 -4.61 <0.0001 0.0030

miR-191 -1.59 -3.39 0.0001 0.0956

miR-369-5p -2.82 -5.25 0.0001 0.0415

miR-502-5P -1.05 -1.45 0.0780 0.2970

miR-203 -1.06 -2.67 0.0024 0.3297

miR-34a -3.61 -5.57 <0.0001 0.0054

miR-17 -3.04 -5.18 0.0001 0.0234

miR-484 -1.37 -3.17 0.0002 0.1526

miR-95 -2.12 -4.80 0.0001 0.1514

miR-195 -2.12 -3.22 0.0012 0.0450

miR-205 -0.95 -2.88 0.0092 0.4781

miR-23b -3.04 -5.81 0.0005 0.0958

miR-21 -2.20 -3.44 0.0001 0.0283

miR-566 -2.22 -2.95 0.0025 0.0418

miR-486-5P -2.21 -3.57 0.0061 0.1161

miR-589 -1.59 -4.38 0.0003 0.2737

miR-431 -3.25 -6.54 <0.0001 0.0396

miR-92a -0.63 -1.71 0.0079 0.3983

miR-410 -2.10 -3.98 0.0006 0.0806

miR-487b -2.60 -4.92 0.0002 0.0711

miR-16 -3.08 -4.82 0.0005 0.0340

miR-486-3P -0.56 -1.81 0.0482 0.6086

miR-532-5P -1.87 -3.31 0.0021 0.1295

miR-192* -0.52 -1.15 0.0016 0.2030

miR-200a* -1.69 -3.80 <0.0001 0.1181

miR-624* -0.86 -2.12 <0.0001 0.0798

miR-181C* -0.96 -2.05 0.0036 0.2564

miR-182* -2.44 -4.56 0.0031 0.1917

miR-592 -0.88 -1.38 0.0013 0.1086

miR-566 -0.31 -1.17 0.1060 0.7034

255

7.14 Appendix XIV: ROC analysis for the detection of adenoma and carcinoma by

using miR-191

miRNAs for

adenoma and

carcinoma AUC

Standard

Error

p-

value

95% Confidence Interval for

AUC

Lower bound Higher bound

miR-34a 0.774 0.060 0.000 0.658 0.891

miR-23b 0.731 0.061 0.003 0.611 0.851

miR-431 0.728 0.059 0.003 0.613 0.844

miR-16 0.706 0.074 0.008 0.561 0.851

miR-486-3P 0.295 0.062 0.009 0.174 0.416

miR-369-5p 0.698 0.066 0.011 0.569 0.827

miR-182* 0.322 0.065 0.023 0.194 0.450

miR-487b 0.674 0.079 0.026 0.519 0.830

miR-592 0.334 0.065 0.034 0.207 0.462

miR-200a* 0.355 0.072 0.064 0.214 0.497

miR-484 0.356 0.068 0.066 0.223 0.489

miR-95 0.629 0.073 0.098 0.487 0.772

miR-624* 0.383 0.072 0.134 0.242 0.523

miR-21 0.615 0.075 0.142 0.467 0.762

miR-181C* 0.615 0.079 0.142 0.460 0.770

miR-135b 0.612 0.073 0.150 0.469 0.756

miR-410 0.604 0.085 0.182 0.437 0.771

miR-203 0.403 0.082 0.214 0.241 0.564

miR-566 0.409 0.080 0.242 0.252 0.565

miR-502-5P 0.410 0.076 0.250 0.262 0.559

miR-192* 0.570 0.069 0.373 0.434 0.705

miR-532-5P 0.548 0.073 0.540 0.405 0.691

miR-589 0.547 0.092 0.546 0.367 0.728

miR-205 0.455 0.080 0.567 0.299 0.611

miR-486-5P 0.534 0.085 0.661 0.367 0.702

miR-566 0.516 0.074 0.841 0.372 0.660

miR-195 0.504 0.080 0.955 0.348 0.661

miR-92a 0.183 0.044 0.000 0.097 0.269

256

7.15 Appendix XV: Patient Characteristics for the final validation cohort

Sample No Diagnosis Sample No Diagnosis Sample No Diagnosis

H1103/10 Gall stones

H1110B/10

Adenoma

H278/09 Hepatic flexure carcinoma

H1104/10 Gall stones

H1110E/10 Adenoma

H286/09 Hepatic flexure carcinoma

H1109F/10 Gall stones

H1110G/10

Adenoma

H298/09 Hepatic flexure carcinoma

H1110F/10 Gall stones

H1110I/10 Adenoma

H957/10 Hepatic flexure carcinoma

H1138C/10 Gall stones

H1110K/10 Adenoma

H280/09 Transverse colonic carcinoma

H1138D/10 Gall stones

H1110L/10 Adenoma

H296/09 Transverse colonic carcinoma

H1138G/10 Gall stones

H1113/10 Adenoma

H669/10 Transverse colonic carcinoma

H47/09 Normal colonoscopy

H1114/10 Adenoma

H168/09 Splenic flexure carcinoma

H1107/10 Normal colonoscopy

H1114C/10 Adenoma

H959/10 Desceding colonic carcinoma

H1109C/10 Normal colonoscopy

H1114H/10 Adenoma

H264/10

Sigmoid colonic carcinoma

H1109J/10 Normal colonoscopy

H1114I/10 Adenoma

H282/09

Sigmoid colonic carcinoma

H1110C/10 Normal colonoscopy

H1115/10 Adenoma

H283/09

Sigmoid colonic carcinoma

H1110J/10 Normal colonoscopy

H1116/10 Adenoma

H297/09

Sigmoid colonic carcinoma

H1112/10 Normal colonoscopy

H1117/10 Adenoma

H668/10

Sigmoid colonic carcinoma

H1114B/10 Normal colonoscopy

H1119/10 Adenoma

H671/10

Sigmoid colonic carcinoma

H1114F/10 Normal colonoscopy

H1120/10 Adenoma

H675/10

Sigmoid colonic carcinoma

H1118/10 Normal colonoscopy

H1138/10 Adenoma

H954/10

Sigmoid colonic carcinoma

H1138A/10 Normal colonoscopy

H1138B/10 Adenoma

H1111/10 Rectosigmoid carcinoma

H1138E/10 Normal colonoscopy

H1138I/10 Adenoma

H263/10 Rectosigmoid carcinoma

H1138F/10 Normal colonoscopy

H1138J/10 Adenoma

H284/09 Rectosigmoid carcinoma

257

H1138L/10 Normal colonoscopy

H1139/10 Adenoma

H159/09 Rectal carcinoma

H1140/10 Normal colonoscopy

H1140A/10 Adenoma

H165/09 Rectal carcinoma

H519/08/A Normal colonoscopy

H1140B/10 Adenoma

H262/10 Rectal carcinoma

H520/08 Normal colonoscopy

H277/09 Caecal carcinoma

H270/10 Rectal carcinoma

H161/09 Normal colonoscopy

H269/10 Caecal carcinoma

H280/10 Rectal carcinoma

H1106/10 Adenoma

H274/10 Caecal carcinoma

H290/09 Rectal carcinoma

H1109A/10 Adenoma

H289/09 Caecal carcinoma

H292/09 Rectal carcinoma

H1109E/10 Adenoma

H677/10 Caecal carcinoma

H299/09 Rectal carcinoma

H1109I/10 Adenoma

H958/10 Caecal carcinoma

H670/10 Rectal carcinoma

H1109L/10 Adenoma

H676/10 Ascending colonic carcinoma

H960/10 Rectal carcinoma

H1110/10 Adenoma

H1106/10 Ascending colonic carcinoma

H960A/10 Rectal carcinoma

H1110A/10 Adenoma H1107/10 Hepatic flexure carcinoma

H960C/10 Rectal carcinoma

258

7.16 Appendix XVI: Total RNA concentrations for final validation cohort

Sample No Concentration ng/µl

Sample No Concentration ng/µl

Sample No Concentration ng/µl

H1103/10 3.8 H1110B/10 13.5 H278/09 5.9

H1104/10 10.4 H1110E/10 8.4 H286/09 7.4

H1109F/10 2.6 H1110G/10 4.8 H298/09 4.6

H1110F/10 12.9 H1110I/10 12.3 H957/10 3.6

H1138C/10 13.6 H1110K/10 14.2 H280/09 13.8

H1138D/10 6.9 H1110L/10 4.8 H296/09 4.1

H1138G/10 4.1 H1113/10 3.9 H669/10 12.3

H47/09 13.5 H1114/10 16.8 H168/09 11.7

H1107/10 15.9 H1114C/10 12.8 H959/10 2.8

H1109C/10 4.6 H1114H/10 18.1 H264/10 3.2

H1109J/10 4.2 H1114I/10 8.8 H282/09 11.6

H1110C/10 3.9 H1115/10 6.4 H283/09 15.1

H1110J/10 4.5 H1116/10 8.9 H297/09 3.9

H1112/10 5.2 H1117/10 4.8 H668/10 4.2

H1114B/10 11.2 H1119/10 14.6 H671/10 3.0

H1114F/10 6.3 H1120/10 8.4 H675/10 5.4

H1118/10 6.8 H1138/10 10.6 H954/10 3.1

H1138A/10 6.5 H1138B/10 14.4 H1111/10 6.6

H1138E/10 11.2 H1138I/10 8.6 H263/10 4.8

H1138F/10 6.4 H1138J/10 9.6 H284/09 4.6

H1138L/10 12.3 H1139/10 14.8 H159/09 14.2

H1140/10 16.4 H1140A/10 12.8 H165/09 15.9

H519/08/A 12.8 H1140B/10 8.6 H262/10 22.3

H520/08 10.4 H277/09 5.0 H270/10 15.8

H161/09 8.6 H269/10 28.1 H280/10 6.2

H1106/10 12.3 H274/10 5.2 H290/09 13.2

H1109A/10 16.6 H289/09 15.9 H292/09 7.1

H1109E/10 8.4 H677/10 12.6 H299/09 4.8

H1109I/10 6.4 H958/10 4.6 H670/10 13.5

H1109L/10 2.8 H676/10 5.6 H960/10 6.7

H1110/10 8.2 H1106/10 3.8 H960A/10 8.7

H1110A/10 5.6 H1107/10 13.5 H960C/10 11.2

259

7.17 Appendix XVII: Table of miR-135b expressions for Normal, Adenoma and Carcinoma

miR-135b expressions (CT), endogenous control snRNA RNU6B expression (CT) and normalised expression of miR-135b (∆CT)

Groups Sample ID CT miR-135b CT SnRNA RNU6B ∆CT miR-135b

Normal

H1103/10 40 33.02 6.98

H1104/10 40 31.59 8.41

H1109F/10 40 33.81 6.19

H1110F/10 40 29.93 10.07

H1138C/10 40 33.15 6.85

H1138D/10 40 31.58 8.42

H1138G/10 40 30.17 9.83

H47/09 40 31.45 8.55

H1107/10 40 32.61 7.39

H1109C/10 40 30.57 9.43

H1109J/10 40 32.2 7.8

H1110C/10 40 32.12 7.88

H1110J/10 25.84 29.44 -3.6

H1112/10 40 32.79 7.21

H1114B/10 40 30.53 9.47

H1114F/10 40 31.59 8.41

H1118/10 40 32.78 7.22

H1138A/10 40 31.82 8.18

H1138E/10 40 30.03 9.97

H1138F/10 40 31.61 8.39

H1138L/10 29.17 29.64 -0.47

H1140/10 30.37 32.97 -2.61

H519/08/A 40 31.28 8.72

H520/08 40 27.42 12.58

H161/09 27.79 30.56 -2.77

Adenoma

H1106/10 40 30.14 9.86

H1109A/10 31.18 28.79 2.39

H1109E/10 28.46 27.47 0.99

H1109I/10 31.6 29.62 1.98

H1109L/10 32.82 29.31 3.51

260

H1110/10 40 30.53 9.47

H1110A/10 34.61 29.38 5.23

H1106/10 25.03 26.57 -1.54

H1110B/10 40 29.56 10.44

H1110E/10 30.74 29.45 1.29

H1110G/10 40 28.76 11.24

H1110I/10 35.47 31.92 3.54

H1110K/10 40 28.37 11.63

H1110L/10 31.72 28.5 3.21

H1113/10 40 31.36 8.64

H1114/10 36.26 31.36 4.9

H1114C/10 40 31.35 8.65

H1114H/10 40 31.62 8.38

H1114I/10 28.28 29.21 -0.93

H1115/10 40 30.61 9.39

H1116/10 30.19 30.22 -0.03

H1117/10 37.01 30.17 6.84

H1119/10 30.73 31.83 -1.1

H1120/10 26.79 27.61 -0.82

H1138/10 36.2 24.55 11.64

H1138B/10 32.92 27.73 5.19

H1138I/10 25.53 25.45 0.08

H1138J/10 34.49 30.28 4.21

H1139/10 27.12 26.28 0.84

H1140A/10 34.88 28.52 6.36

Carcinoma

H277/09 29.96 28.45 1.5

H269/10 40 31.57 8.43

H274/10 25.87 25.7 0.17

H289/09 26.72 29.59 -2.87

H677/10 28.12 28.34 -0.21

H958/10 26.42 27.26 -0.84

H676/10 26.74 27.77 -1.03

H1106/10 27.16 27.68 -0.52

261

H1107/10 29.91 29.6 0.3

H278/09 27.67 27.19 0.48

H286/09 40 32.03 7.97

H298/09 26.14 29.13 -2.99

H957/10 30 29.55 0.46

H280/09 29.64 30.01 -0.37

H296/09 25.59 26.33 -0.74

H669/10 27.45 27.94 -0.49

H168/09 29.71 27.75 1.96

H959/10 31.17 30.76 0.41

H264/10 27.3 29.55 -2.25

H282/09 25.03 27.13 -2.09

H283/09 26.58 27.67 -1.08

H297/09 27.24 27.34 -0.1

H668/10 28.91 29.94 -1.03

H671/10 27.35 29.58 -2.22

H675/10 29.16 30.34 -1.17

H954/10 24.23 26.27 -2.04

H1111/10 27.74 26.75 0.99

H263/10 29.48 29.99 -0.51

H284/09 28.18 29.24 -1.06

H159/09 40 28.99 11.01

H165/09 33.51 29.77 3.74

H262/10 29.73 27.61 2.12

H270/10 28.79 28.46 0.32

H280/10 29.28 28.08 1.2

H290/09 30.72 29.68 1.05

H292/09 33.21 26.98 6.23

H299/09 28.97 30.07 -1.1

H670/10 26.78 28.6 -1.82

H960/10 28.32 29.59 -1.27

H960A/10 40 29.92 10.08

H960C/10 30.16 30.82 -0.67

262

7.18 Appendix XVIII: List of publications, grants, awards and presentations

Publications

1. Aslam MI, Venkatesh J, Jameson JS, West K, Pringle JH, Singh B. Identification of high-risk Dukes B colorectal cancer by microRNA expression profiling: a preliminary study. Colorectal Dis. 2015 Jul;17(7):578-88

2. Verma AM, Patel M, Aslam MI, Jameson J, Pringle JH, Wurm P, Singh B. Circulating plasma microRNAs as a screening method for detection of colorectal adenomas. Lancet. 2015 Feb 26;385 Suppl 1:S100.

3. Patel M, Verma A, Aslam I, Pringle H, Singh B. Novel plasma microRNA biomarkers for the identification of colitis-associated carcinoma. Lancet. 2015 Feb 26;385 Suppl 1:S78.

4. Aslam MI, Hussein S, West K, Singh B, Jameson JS, Pringle JH. MicroRNAs associated with initiation and progression of colonic polyp: a feasibility study. Int J Surg. 2015 Jan;13:272-9

5. Saldanha G, Potter L, Shendge P, Osborne J, Nicholson S, Yii N, Varma S, Aslam MI, El shaw S, Papadogeorgakis E, Pringle JH. Plasma microRNA-21 is associated with tumor burden in cutaneous melanoma. J Invest Dermatol. 2013 May;133(5):1381-4.

6. Aslam MI, Patel M, Singh B, Jameson JS, Pringle JH. MicroRNA manipulation in colorectal cancer cells: from laboratory to clinical application. J Transl Med. 2012 Jun 20;10:128

7. Aslam MI, Patel M, Singh B, Pringle JH, Jameson JS. MicroRNAs are Novel Biomarkers for Detection of Colorectal Cancer. 2012. InTech Book Chapter - "Biomarker". ISBN 979953-307-612-5.

8. Aslam MI, Patel M, Singh B, Pringle JH, Jameson JS. MicroRNA based blood assay for early detection of colorectal cancer. British Journal of Surgery 2011; 98(S2): 6–39 (abstract publication).

9. Aslam MI, Taylor K, Pringle JH, Jameson JS. MicroRNAs are novel biomarkers of colorectal cancer. Br J Surg. 2009 Jul;96(7):702-10.

Grants

1. Project Grant (£29,500) for “Novel Biomarker based staging of colorectal cancer”, Bowel Diseases Research Foundation, 2012-2014

2. Project Grant (£30,000) for “MiRNAs based blood assay for early detection of colorectal cancer”, Bowel Diseases Research Foundation, 2010-2012

3. Innovation Fellowship Grant (£16,000), “MiRNAs based blood assay for early detection of colorectal cancer”, East Midlands Business Development Agency, 2010-2011

4. Small Study Grant (£1000) for “Tissue and blood microRNAs for detection of ulcerative colitis associated dysplasia and carcinoma”, Midlands Gastroenterology Society, 2012

5. Small Study Grants (£1000) “MicroRNAs for sclerosing cholangitis”, Midlands Gastroenterology Society, 2011

263

Awards

1. Winner of NHS Innovation Award (£4,000), “Plasma miRNAs based detection of colorectal cancer”, Medipex Medical Diagnostics & Devices category Awards 2011

2. Best Oral Presentation Award (£100), “Detection of colorectal cancer based on miRNA blood test”, Midlands Gastroenterology Society meeting 2011.

Presentations

1. Aslam MI, Hussein S, Patel M, West K, Jameson JS, Pringle JH,Singh B. hsa-miR-135b is Associated with the Initiation and Progression of Colorectal Neoplasia. American College of Surgeons, Clinical Congress 2014, San Francisco, USA.

2. Aslam MI, Patel M, Taylor K, Potter L, West K, Pringle JH, Jameson JS, Singh B. MicroRNA expression profiling based identification of high risk Dukes’ stage B colorectal cancer. The American society of colon and rectal surgeons USA. Annual scientific meeting. June 2-6 2012

3. Aslam MI, Patel M, Singh B, Pringle JH, Jameson JS. MicroRNA expression profiling based identification of high risk Dukes’ stage B colorectal cancers. Travelling surgical society of Great Britain and Ireland. Trainees’ Paper Prize 2012, Leicester Royal Infirmary, 21 Sep, 2012

4. Aslam MI, Patel M, Singh B, Pringle JH, Jameson JS. Blood based assay for early detection of colorectal adenoma and carcinoma. Society of academic & research surgery university of nottingham 4-5 January 2012 Oral presentation in a plenary session for the Patey Prize

5. Aslam MI, Patel M, Singh B, Pringle JH, Jameson JS. Identification of novel MicroRNAs associated with colorectal cancer. ASGBI 2011

6. Aslam MI, Patel M, Singh B, Pringle JH, Jameson JS. MicroRNAs based blood assay for early detection of colorectal cancer. Association of Coloproctology- Trent Chapter Meeting. 1st November 201

7. Aslam MI, Patel M, Singh B, Pringle JH, Jameson JS. Utility of circulating small RNAs for colorectal cancer detection. Midland Gastroenterology Society. Annual Meeting. 17 June 2011. Worcester.

8. Aslam MI, Patel M, Singh B, Pringle JH, Jameson JS. “Plasma miRNAs based detection of colorectal cancer”. SET for Britain Exhibition of Posters by early-stage and early-career research scientists, engineers and technologists. Westminster, UK Parliament, March 2011

9. Aslam MI, Singh B, Pringle JH, Jameson JS. MicroRNA based blood assay for early detection of colorectal cancer. Society for academic and research surgery (SARS), Dublin. 5-6 January 2011

10. Aslam MI, Singh B, Pringle JH, Jameson JS. MicroRNAs are novel biomarkers of colorectal cancer. MicroRNAs-Europe, University of Oxford. 1-2 November 2010

11. Aslam MI, Pringle JH, Jameson JS. Circulating novel biomarkers for colorectal cancer screening. East Midlands Surgical Society- annual meeting. Leicester General Hospital. 16 May 2010

264

References

265

8 References

1. Aaltonen LA, Peltomäki P, Leach FS, Sistonen P, Pylkkänen L, Mecklin JP, Järvinen H, Powell SM, Jen J, Hamilton SR. Clues to the pathogenesis of familial colorectal cancer. Science. 1993 May 7;260(5109):812-6.

2. Adams RH, Alitalo K. Molecular regulation of angiogenesis and lymphangiogenesis. Nat Rev Mol Cell Biol 2007; 8: 464-478

3. Agesen TH, Sveen A, Merok MA, Lind GE, Nesbakken A, Skotheim RI, Lothe RA ColoGuideEx: a robust gene classifier specific for stage II colorectal cancer prognosis. Gut. 2012 Nov;61(11):1560-7

4. Ahlquist DA. Molecular detection of colorectal neoplasia. Gastroenterology.2010;138: 2127– 2139

5. Ahmed FE, Jeffries CD, Vos PW, Flake G, Nuovo GJ, Sinar DR, Naziri W, Marcuard SP. Diagnostic microRNA markers for screening sporadic human colon cancer and active ulcerative colitis in stool and tissue. Cancer Genomics Proteomics. 2009;6(5):281-95.

6. Akao Y, Nakagawa Y, Naoe T. MicroRNAs 143 and 145 are possible common oncomicroRNAs in human cancers. Oncol Rep. 2006;16(4):845-50

7. Akao Y, Nakagawa Y, Naoe T. let-7 microRNA functions as a potential growth suppressor in human colon cancer cells. Biol Pharm Bull. 2006;29(5):903-6.

8. Akao Y, Noguchi S, Iio A, et al: Dysregulation of microRNA-34a expression causes drug-resistance to 5-FU in human colon cancer DLD-1 cells. Cancer Lett 2011, 300(2):197–204.

9. Alrawi SJ, Schiff M, Carroll RE, Dayton M, Gibbs JF, Kulavlat M, Tan D, Berman K, Stoler DL, Anderson GR. Aberrant crypt foci. Anticancer Res 2006; 26: 107-119

10. Amaral JD, Xavier JM, Steer CJ, Rodrigues CM. The role of p53 in apoptosis. Discov Med 2010; 9: 145-152

11. American Cancer Society. Colorectal cancer risk factors. http://www.cancer.org/cancer/colonandrectumcancer/detailedguide/colorectal-cancer-risk-factors (Accessed on 30 January, 2016)

12. Arigoni M, Barutello G, Riccardo F, Ercole E, Cantarella D, Orso F, Conti L, Lanzardo S, Taverna D, Merighi I, Calogero RA, Cavallo F, Quaglino E. miR-135b coordinates progression of ErbB2-driven mammary carcinomas through suppression of MID1 and MTCH2. Am J Pathol. 2013 Jun;182(6):2058-70.

13. Arndt GM, Dossey L, Cullen LM, Lai A, Druker R, Eisbacher M, Zhang C, Tran N, Fan H, Retzlaff K, Bittner A, Raponi M. Characterization of global microRNA expression reveals oncogenic potential of miR-145 in metastatic colorectal cancer. BMC Cancer. 2009 Oct 20;9:374.

266

14. Asangani IA, Rasheed SA, Nikolova DA, et al: MicroRNA-21 (miR-21) post-transcriptionally downregulates tumor suppressor Pdcd4 and stimulates invasion, intravasation and metastasis in colorectal cancer. Oncogene 2008, 27(15):2128–2136.

15. Atkin W, Edwards R, Kralj-Hans I. Once-only flexible sigmoidoscopy screening in prevention of colorectal cancer: a multicentre randomised controlled trial. The Lancet, 2010, 375(9726):1624-1633

16. Bahnassy AA, Zekri AR, Salem SE, Abou-Bakr AA, Sakr MA, Abdel-Samiaa AG, Al-Bradei M. Differential expression of p53 family proteins in colorectal adenomas and carcinomas: Prognostic and predictive values. Histol Histopathol 2014; 29: 207-216

17. Baker, M. MicroRNA profiling: Separating signal from noise. Nat. Publ. Group 2010, 7, 687–692.

18. Bala S, Petrasek J, Mundkur S, et al. Circulating microRNAs in exosomes indicate hepatocyte injury and inflammation in alcoholic, drug-induced, and inflammatory liver diseases. Hepatology. 2012; 56:1946–1957

19. Bandrés E, Cubedo E, Agirre X, Malumbres R, Zárate R, Ramirez N, Abajo A, Navarro A, Moreno I, Monzó M, García-Foncillas J. Identification by Real-time PCR of 13 mature microRNAs differentially expressed in colorectal cancer and non-tumoral tissues. Mol Cancer. 2006.19;5:29

20. Barnes DE, Lindahl T. Repair and genetic consequences of endogenous DNA base damage in mammalian cells. Annu Rev Genet. 2004;38:445-76.

21. Bartel DP. MicroRNAs: genomics, biogenesis, mechanism, and function. Cell. 2004 23;116(2):281-97

22. Bartel DP. MicroRNAs: target recognition and regulatory functions.Cell: 2009;136, 215–233.

23. Benes, V.; Castoldi, M. Expression profiling of microRNA using real-time quantitative PCR, how to use it and what is available. Methods 2010, 50, 244–249.

24. Bettington M, Walker N, Clouston A, Brown I, Leggett B, Whitehall V. The serrated pathway to colorectal carcinoma: current concepts and challenges. Histopathology 2013; 62: 367-386.

25. Blanks RG, Moss SM, McGahan CE, Quinn MJ, Babb PJ. Effect of NHS breast screening programme on mortality from breast cancer in England and Wales, 1990–8: comparison of observed with predicted mortality. BMJ 2000; 321: 665–669.

26. Blanpain C. Tracing the cellular origin of cancer. Nat Cell Biol 2013; 15: 126-134

27. Bockhorn M, Jain RK, Munn LL. Active versus passive mechanisms in metastasis: do cancer cells crawl into vessels, or are they pushed? Lancet Oncol 2007;8: 444–448.

28. Bohnsack M T, Czaplinski K & Gorlich D. Exportin 5 is a RanGTP-dependent dsRNAbinding protein that mediates nuclear export of pre-miRNAs. RNA: 2004(10);185– 191

267

29. Bond CE, Nancarrow DJ, Wockner LF, Wallace L, Montgomery GW, Leggett BA, Whitehall VL. Microsatellite stable colorectal cancers stratified by the BRAF V600E mutation show distinct patterns of chromosomal instability. PLoS One 2014; 9: e91739

30. Boni V, Bitarte N, Cristobal I, et al: miR-192/miR-215 influence 5-fluorouracil resistance through cell cycle-mediated mechanisms complementary to its post-transcriptional thymidilate synthase regulation. Mol Cancer Ther 2010, 9(8):2265–2275.

31. Bovell LC, Shanmugam C, Putcha BD. The prognostic value of microRNAs varies with patient race/ethnicity and stage of CRC. Clinical cancer research an official journal of the American Association for Cancer Research. 2013; 19:3955–396

32. Boyle P, Langman J. ABC of colorectal cancer Epidemiology.BMJ.2000; 321:805-808.

33. Boyle P, Leon ME. Epidemiology of colorectal cancer. Br Med Bull. 2002;64:1-25.

34. Bray F, Atkin W. International cancer patterns in men: geographical and temporal variations in cancer risk and the role of gender. JMHG.2004;1(1):38-46

35. Brim H, Mokarram P, Naghibalhossaini F et al. Impact of BRAF, MLH1 on the incidence of microsatellite instability high colorectal cancer in populations based study. Mol Cancer 2008; 7: 68.

36. Brittan M, Wright NA. Stem cell in gastrointestinal structure and neoplastic development. Gut 2004;53:899–910

37. Broderick P, Carvajal-Carmona L, Pittman AM, Webb E, Howarth K, Rowan A, Lubbe S, Spain S, Sullivan K, Fielding S, Jaeger E, Vijayakrishnan J, Kemp Z, Gorman M, Chandler I, Papaemmanuil E, Penegar S, Wood W, Sellick G, Qureshi M, Teixeira A, Domingo E, Barclay E, Martin L, Sieber O; CORGI Consortium, Kerr D, Gray R, Peto J, Cazier JB, Tomlinson I, Houlston RS. A genome-wide association study shows that common alleles of SMAD7 influence colorectal cancer risk. Nat Genet. 2007 Nov;39(11):1315-7.

38. Brunet-Vega A, Pericay C, Quílez ME, Ramírez-Lázaro MJ, Calvet X, Lario S Variability in microRNA recovery from plasma: Comparison of five commercial kits. Anal Biochem. 2015 Nov 1;488:28-35.

39. Bufill JA. Colorectal cancer: evidence for distinct genetic categories based on proximal or distal tumor location. Ann Intern Med 1990; 113: 779–88.

40. Cancer Research UK. Statistics and Outlook for Bowel Cancer – Prognosis by Stage.http://cancerhelp.cancerresearchuk.org/type/bowel- cancer/treatment/statistics -and-outlook-for-bowel-cancer#2 (accessed October 2011)

41. Cappellesso R, Tinazzi A, Giurici T, Simonato F, Guzzardo V, Ventura L. Programmed cell death 4 and microRNA 21 inverse expression is maintained in cells and exosomes from ovarian serous carcinoma effusions. Cancer Cytopathol. 2014;122:685–93.

268

42. Carethers JM. Differentiating Lynch-like from Lynch syndrome. Gastroenterology 2014; 146: 602-604.

43. Carina Roth, Brigitte Rack, Volkmar Müller, Wolfgang Janni, Klaus Pantel, Heidi Schwarzenbach. Circulating microRNAs as blood-based markers for patients with primary and metastatic breast cancer. Breast Cancer Res. 2010; 12(6):R90

44. Carmeliet P, Dor Y, Herbert JM, Fukumura D, Brusselmans K, Dewerchin M, Neeman M, Bono F, Abramovitch R, Maxwell P, Koch CJ, Ratcliffe P, Moons L, Jain RK, Collen D, Keshert E. Role of HIF-1alpha in hypoxia-mediated apoptosis, cell proliferation and tumour angiogenesis. Nature 1998; 394: 485-490

45. Casey G, Lindor NM, Papadopoulos N, Thibodeau SN, Moskow J, Steelman S, Buzin CH, Sommer SS, Collins CE, Butz M, Aronson M, Gallinger S, Barker MA, Young JP, Jass JR, Hopper JL, Diep A, Bapat B, Salem M, Seminara D, Haile R; Colon Cancer Family Registry.Conversion analysis for mutation detection in MLH1 and MSH2 in patients with colorectal cancer. JAMA. 2005.16;293(7):799-809

46. Cazzoli R, Buttitta F, Di Nicola M, Malatesta S, Marchetti A, Rom WN, Pass HI. microRNAs derived from circulating exosomes as noninvasive biomarkers for screening and diagnosing lung cancer. J Thorac Oncol. 2013;8:1156–1162

47. Cekaite L, Rantala JK, Bruun J et al. MiR-9, -31, and -182 deregulation promote proliferation and tumor cell survival in colon cancer. Neoplasia 2012; 14: 868–79.

48. Chai H, Liu M, Tian R, Li X, Tang H. miR-20a targets BNIP2 and contributes chemotherapeutic resistance in colorectal adenocarcinoma SW480 and SW620 cell lines. Acta Biochim Biophys Sin (Shanghai) 2011, 43(3):217–225.

49. Chan M, Liaw CS, Ji SM, et al. Identification of circulating microRNA signatures for breast cancer detection. Clinical cancer research: an official journal of the American Association for Cancer Research. 2013; 19:4477–4487.

50. Chang KH, Miller N, Kheirelseid EA, Ingoldsby H, Hennessy E, Curran CE, Curran S, Smith MJ, Regan M, McAnena OJ, Kerin MJ. MicroRNA-21 and PDCD4 expression in colorectal cancer. Eur J Surg Oncol. 2011 Jul;37(7):597-603.

51. Chen C, Ridzon DA, Broomer AJ, Zhou Z, Lee DH, Nguyen JT et al. Real-time quantification of microRNAs by stem–loop RT–PCR. Nucleic Acids Res 2005; 33: 179

52. Chen C, Ridzon DA, Broomer AJ, Zhou Z, Lee DH, Nguyen JT, Barbisin M, Xu NL, Mahuvakar VR, Andersen MR, Lao KQ, Livak KJ, Guegler KJ.Real-time quantification of microRNAs by stem-loop RT-PCR. Nucleic Acids Res.2005; 33:e179

53. Chen PS, Su JL, Hung MC. Dysregulation of microRNAs in cancer. J Biomed Sci. 2012;19:90

54. Chen WX, Liu XM, Lv MM, Chen L, Zhao JH, Zhong SL, et al. Exosomes from drug-resistant breast cancer cells transmit chemoresistance by a horizontal transfer of microRNAs. PLoSOne. 2014;9:e95240.

269

55. Chen X, Ba Y, Ma L, Cai X, Yin Y, Wang K, Guo J, Zhang Y, Chen J, Guo X, Li Q, Li X, Wang W, Zhang Y, Wang J, Jiang X, Xiang Y, Xu C, Zheng P, Zhang J, Li R, Zhang H, Shang X, Gong T, Ning G, Wang J, Zen K, Zhang J, Zhang CY. Characterization of microRNAs in serum: a novel class of biomarkers for diagnosis of cancer and other diseases. Cell Res. 2008;18(10):997-1006.

56. Chendrimada TP, Gregory RI, Kumaraswamy E, Norman J, Cooch N, Nishikura K, Shiekhattar R. TRBP recruits the Dicer complex to Ago2 for microRNA processing and gene silencing. Nature. 2005;436(7051):740-4.

57. Cheng H, Zhang L, Cogdell DE, Zheng H, Schetter AJ, Nykter M, Harris CC, Chen K, Hamilton SR, Zhang W. Circulating plasma MiR-141 is a novel biomarker for metastatic colon cancer and predicts poor prognosis. PLoS One. 2011.17;6(3):e17745

58. Cheng HH, Yi HS, Kim Y, Kroh EM, Chien JW, Eaton KD, Goodman MT, Tait JF, Tewari M, Pritchard CC. Plasma processing conditions substantially influence circulating microRNA biomarker levels. PLoS One. 2013 Jun 7;8(6):e64795.

59. Chim SS, Shing TK, Hung EC, Leung TY, Lau TK, Chiu RW, Lo YM. Detection and characterization of placental microRNAs in maternal plasma. Clin Chem. 2008 Mar; 54(3):482-90.

60. Chim SS, Shing TK, Hung EC, Leung TY, Lau TK, Chiu RW, Lo YMDetection and characterization of placental microRNAs in maternal plasma. Clin Chem. 2008 Mar;54(3):482-90.

61. Choi DS, Lee JM, Park GW, Lim HW, Bang JY, Kim YK, Kwon KH, Kwon HJ, Kim KP, Gho YS. Proteomic analysis of microvesicles derived from human colorectal cancer cells. J Proteome Res. 2007;6(12):4646-55

62. Chugh, P.; Dittmer, D.P. Potential pitfalls in microRNA profiling. Wiley Interdiscip. Rev. RNA 2012, 3, 601–616.

63. Chung GE, Yoon JH, Myung SJ et al. High expression of microRNA-15b predicts a low risk of tumor recurrence following curative resection of hepatocellular carcinoma. Oncol Rep 2010; 23: 113–9.

64. Ciprotti M, Chong G, Gan HK, Chan A, Murone C, MacGregor D, Lee FT, Johns TG, Heath JK, Ernst M, Burgess AW, Scott AM.. Quantitative intratumoural microdistribution and kinetics of (131)I-huA33 antibody in patients with colorectal carcinoma. EJNMMI Res. 2014 May 30;4:22.

65. Circulating microRNAs as biomarkers of colorectal cancer: results from a genome-wide profiling and validation study. Clin Gastroenterol Hepatol. 2013 Jun;11(6):681-8.e3.

66. Cocucci E, Racchetti G, Meldolesi J. Shedding microvesicles: artefacts no more. Trends Cell Biol. 2009;19(2):43-51

67. Coleman MP, Estève J, Damiecki P, Arslan A, Renard H. Trends in cancer incidence and mortality. IARC Sci Publ. 1993; (121):1-806.

270

68. Cooper GM. The Cell: A Molecular Approach. 2nd ed. Sunderland (MA): Sinauer Associates, 2000.

69. Corcoran C, Rani S, O'Brien K, O'Neill A, Prencipe M, Sheikh R, et al. Docetaxel-resistance in prostate cancer: evaluating associated phenotypic changes and potential for resistance transfer via exosomes. PLoSOne. 2012;7:e50999.

70. Cortez MA, Calin GA. MicroRNA identification in plasma and serum: a new tool to diagnose andmonitor diseases. Expert opinion on biological therapy. 2009; 9:703–11.

71. Corvinus FM, Orth C, Moriggl R, Tsareva SA, Wagner S, Pfitzner EB, Baus D, Kaufmann R, Huber LA, Zatloukal K, Beug H, Ohlschläger P, Schütz A, Halbhuber KJ, Friedrich K. Persistent STAT3 activation in colon cancer is associated with enhanced cell proliferation and tumor growth. Neoplasia. 2005 Jun;7(6):545-55.

72. Coste J, Cochand-Priollet B, de Cremoux P, Le Gales C, Cartier I, Molinie V. Cross sectional study of conventional cervical smear, monolayer cytology, and human papillomavirus DNA testing for cervical cancer screening. BMJ 2003; 326: 733.

73. Cronin M, Pho M, Dutta D, Stephans JC, Shak S, Kiefer MC, Esteban JM, Baker JB. 2004. Measurement of gene expression in archival paraffin-embedded tissues: development and performance of a 92-gene reverse transcriptase-polymerase chain reaction assay. Am J Pathol 164:35- 42.

74. Davidson LA, Wang N, Shah MS, Lupton JR, Ivanov I, Chapkin RS. n-3 Polyunsaturated fatty acids modulate carcinogen-directed non-coding microRNA signatures in rat colon. Carcinogenesis 2009; 30: 2077–84.

75. Day E, Dear PH, McCaughan F. Digital PCR strategies in the development and analysis of molecular biomarkers for personalized medicine. Methods. 2013; 59:101–107

76. De Roock W, Claes B, Bernasconi D, De Schutter J, Biesmans B, Fountzilas G, Kalogeras KT, Kotoula V, Papamichael D, Laurent-Puig P, Penault-Llorca F, Rougier P, Vincenzi B, Santini D, Tonini G, Cappuzzo F, Frattini M, Molinari F, Saletti P, De Dosso S, Martini M, Bardelli A, Siena S, Sartore-Bianchi A, Tabernero J, Macarulla T, Di Fiore F, Gangloff AO, Ciardiello F, Pfeiffer P, Qvortrup C, Hansen TP, Van Cutsem E, Piessevaux H, Lambrechts D, Delorenzi M, Tejpar S. Effects of KRAS, BRAF, NRAS, and PIK3CA mutations on the efficacy of cetuximab plus chemotherapy in chemotherapy-refractory metastatic colorectal cancer: a retrospective consortium analysis. Lancet Oncol. 2010 Aug;11(8):753-62.

77. Deller MC, Yvonne Jones E. Cell surface receptors. Curr Opin Struct Biol 2000; 10: 213-219.

78. Deuter R, Pietsch S, Hertel S, Muller O. A method for preparation of fecal DNA suitable for PCR. Nucleic Acids Res 1995;23:3800–1

79. Dewhirst MW. Relationships between cycling hypoxia, HIF-1, angiogenesis and oxidative stress. Radiat Res. 2009 Dec;172(6):653-65.

271

80. Di Fiore F, Sesboüé R, Michel P, Sabourin JC, Frebourg T. Molecular determinants of anti-EGFR sensitivity and resistance in metastatic colorectal cancer. Br J Cancer 2010; 103: 1765-1772

81. Dias N, Stein CA: Antisense oligonucleotides: basic concepts and mechanisms. Mol Cancer Ther 2002, 1(5):347–355.

82. Dickinson BT, Kisiel J, Ahlquist DA, Grady WM. Molecular markers for colorectal cancer screening. Gut. 2015 Sep;64(9):1485-94.

83. Dieter SM, Ball CR, Hoffmann CM, Nowrouzi A, Herbst F, Zavidij O, Abel U, Arens A, Weichert W, Brand K, Koch M, Weitz J, Schmidt M, von Kalle C, Glimm H. Distinct types of tumor-initiating cells form human colon cancer tumors and metastases. Cell Stem Cell.2011; 9: 357–365.

84. Diosdado B, van de Wiel MA, Terhaar Sive Droste JS, Mongera S, Postma C, Meijerink WJ, Carvalho B, Meijer GA. MiR-17-92 cluster is associated with 13q gain and c-myc expression during colorectal adenoma to adenocarcinoma progression. Br J Cancer. 2009 Aug 18;101(4):707-14.

85. Dobre M, Comănescu M, Arsene D, Iosif C, Bussolati G. K-ras gene mutation status in colorectal cancer: comparative analysis of pyrosequencing and PCR-RFLP. Rom J Morphol Embryol 2013; 54: 567-574

86. Dong Y, Wu WK, Wu CW, Sung JJ, Yu J, Ng SS. MicroRNA dysregulation in colorectal cancer: a clinical perspective. Br J Cancer. 2011.15;104(6):893-8

87. Draviam VM, Shapiro I, Aldridge B, Sorger PK. Misorientation and reduced stretching of aligned sister kinetochores promote chromosome missegregation in EB1- or APC-depleted cells. EMBO J 2006; 25: 2814-2827

88. Duffy MJ. Carcinoembryonic antigen as a marker for colorectal cancer: is it clinically useful? Clin Chem. 2001;47:624–630.

89. Dukes CE, Bussey HJ. The spread of rectal cancer and its effect on prognosis. Br J Cancer. 1958 Sep;12(3):309-20.

90. E.R. Fearon, B. Vogelstein, A genetic model for colorectal tumorigenesis, Cell 61 (1990) 759e767.

91. Eichelser C, Flesch-Janys D, Chang-Claude J, Pantel K, Schwarzenbach H. Deregulated serum concentrations of circulating cell-free microRNAs miR-17, miR-34a, miR-155, and miR-373 in human breast cancer development and progression. Clin Chem. 2013; 59, 1489-96.

92. Eiring AM, Harb JG, Neviani P, Garton C, Oaks JJ, Spizzo R, Liu S, Schwind S, Santhanam R, Hickey CJ, Becker H, Chandler JC, Andino R, Cortes J, Hokland P, Huettner CS, Bhatia R, Roy DC, Liebhaber SA, Caligiuri MA, Marcucci G, Garzon R, Croce CM, Calin GA, Perrotti D. miR-328 functions as an RNA decoy to modulate hnRNP E2 regulation of mRNA translation in leukemic blasts. Cell. 2010.5;140(5):652-65.

272

93. Epstein DM. Special delivery: microRNA-200-containing extracellular vesicles provide metastatic message to distal tumor cells. J Clin Invest. 2014;124:5107–8.

94. Esquela-Kerscher A, Slack FJ. Oncomirs - microRNAs with a role in cancer. Nat Rev Cancer. 2006;6(4):259-69

95. Fabbri M, Paone A, Calore F, Galli R, Croce CM. A new role for microRNAs, as ligands of Toll-like receptors. RNABiol. 2013;10:169–74.

96. Failli, Alessandra, Annalisa Legitimo, Francesca Migheli, Fabio Coppedè, John C. Mathers, Roberto Spisni, Paolo Miccoli, Lucia Migliore, Rita Consolini. Efficacy and Feasibility of the Epithelial Cell Adhesion Molecule (EpCAM) Immunomagnetic Cell Sorter for Studies of DNA Methylation in Colorectal CancerInt J Mol Sci. 2014 January; 15(1): 44–57.

97. Fakih MG, Padmanabhan A. CEA monitoring in colorectal cancer. What you should know. Oncology. 2006;20:579–587.

98. Faltejskova P, Bocanek O, Sachlova M, Svoboda M, Kiss I, Vyzula R, Slaby O.Circulating miR-17-3p, miR-29a, miR-92a and miR-135b in serum: Evidence against their usage as biomarkers in colorectal cancer. Cancer Biomark. 2012;12(4):199-204

99. Faltejskova P, Bocanek O, Sachlova M, Svoboda M, Kiss I, Vyzula R, Slaby O. Circulating miR-17-3p, miR-29a, miR-92a and miR-135b in serum: Evidence against their usage as biomarkers in colorectal cancer. Cancer Biomark. 2012;12(4):199-204

100. Fanali, C, Donatella L, Marisa F, Maddalena C, Valerio C, Achille C, Alessandro S. Cancer stem cells in colorectal cancer from pathogenesis to therapy: Controversies and perspectives World J Gastroenterol. 2014 January 28; 20(4): 923–942.

101. Fang C, Zhu DX, Dong HJ, Zhou ZJ, Wang YH, Liu L, Fan L, Miao KR, Liu P, Xu W, Li JY. Serum microRNAs are promising novel biomarkers for diffuse large B cell lymphoma. Ann Hematol 2012; 91, 553-9.

102. Fearon ER, Vogelstein B. A genetic model for colorectal tumorigenesis. Cell 1990; 61: 759–67.

103. Feng Y, Bommer GT, Zhao J, Green M, Sands E, Zhai Y, Brown K, Burberry A, Cho KR, Fearon ER. Mutant KRAS promotes hyperplasia and alters differentiation in the colon epithelium but does not expand the presumptive stem cell pool. Gastroenterology 2011; 141: 1003-1013.e1-10

104. Ferlay J, Parkin DM, Steliarova-Foucher E. Estimates of cancer incidence and mortality in Europe in 2008. Eur J Cancer. 2010;46(4):765-81.

105. Filipazzi P, Burdek M, Villa A, Rivoltini L, Huber V. Recent advances on the role of tumor exosomes in immunosuppression and disease progression. Semin Cancer Biol. 2012;22:342–9

273

106. Flood DM, Weiss NS, Cook LS, Emerson JC, Schwartz SM, Potter JD. Colorectal cancer incidence in Asian migrants to the United States and their descendants. Cancer Causes Control. 2000;11(5):403-11

107. Friedman RC, Farh KK, Burge CB, Bartel DP. Most mammalian mRNAs are conserved targets of microRNAs. Genome Res. 2009;19(1):92-105.

108. Gao F, Zhao ZL, Zhao WT, Fan QR, Wang SC, Li J, et al. miR-9 modulates the expression of interferon-regulated genes and MHC class I molecules in human nasopharyngeal carcinoma cells. Biochem Biophys Res Commun. 2013;431:610–6.

109. Gao HJ. Initial study of microRNA expression profiles of colonic cancer without lymph node metastasis. J Dig Dis 2010; 11: 50–4.

110. Garinchesa P, Sakamoto J, Welt S, Real F, Rettig W. Old L. Organ-specific expression of the colon cancer antigen A33, a cell surface target for antibody-based therapy. Int. J. Oncol 1996; 9, 465-471. 10.

111. Gilad S, Meiri E, Yogev Y, Benjamin S, Lebanony D, Yerushalmi N, Benjamin H, Kushnir M, Cholakh H, Melamed N, Bentwich Z, Hod M, Goren Y, Chajut A. Serum microRNAs are promising novel biomarkers. PLoS One. 2008.5; 3(9):e3148

112. Giraldez MD, Lozano JJ, Ramirez G. Circulating microRNAs as biomarkers of colorectal cancer: results from a genome-wide profiling and validation study. Clinical gastroenterology and hepatology: the official clinical practice journal of the American Gastroenterological Association. 2013; 11:681–688. e683

113. Grange C, Tapparo M, Collino F, Vitillo L, Damasco C, Deregibus MC, et al. Exosomes released from human renal cancer stem cells stimulate angiogenesis and formation of lung premetastatic niche. Cancer Res. 2011;71:5346–56.

114. Graser A, Stieber P, Nagel D, Schafer C, Horst D, Becker CR et al. Comparison of CT colonography, colonoscopy, sigmoidoscopy and faecal occult blood tests for the detection of advanced adenoma in an average risk population. Gut 2008; 58: 241–248.

115. Graser A, Stieber P, Nagel D, Schafer C, Horst D, Becker CR, Nikolaou K, Lottes A, Geisbüsch S, Kramer H, Wagner AC, Diepolder H, Schirra J, Roth HJ, Seidel D, Göke B, Reiser MF, Kolligs FT. Comparison of CT colonography, colonoscopy, sigmoidoscopy and faecal occult blood tests for the detection of advanced adenoma in an average risk population. Gut. 2009; 58(2):241-8.

116. Guidelines for Management of Colorectal Cancer. Association of Coloproctology of Great Britain and Ireland (ACPGBI), 3rd edn. 2007. http://www.acpgbi.org.uk/ resources/guidelines (accessed September 2014)

117. Guo L, Lu Z. The fate of miRNA* strand through evolutionary analysis: implication for degradation as merely carrier strand or potential regulatory molecule? PLoS One. 2010.30; 5(6):e11387

274

118. Guo L, Yang Q, Lu J, et al. A comprehensive survey of miRNA repertoire and 3′ addition events in the placentas of patients with pre-eclampsia from high-throughput sequencing. PloS one. 2011; 6:e21072.

119. Hammond SM, Bernstein E, Beach D, Hannon GJ. An RNA-directed nuclease mediates posttranscriptional gene silencing in Drosophila cells. Nature. 2000.16; 404(6775):293-6.

120. Hardcastle JD, Chamberlain JO, Robinson MH, Moss SM, Amar SS, Balfour TW, James PD, Mangham CM. Randomised controlled trial of faecal-occult-blood screening for colorectal cancer. Lancet. 1996 Nov 30;348(9040):1472-7.

121. Harris TA, Yamakuchi M, Ferlito M, Mendell JT, Lowenstein CJ (2008) MicroRNA-126 regulates endothelial expression of vascular cell adhesion molecule 1. Proc Natl Acad Sci USA 105: 1516–1521.

122. He YQ, Sheng JQ, Ling XL, Fu L, Jin P, Yen L, Rao J. Estradiol regulates miR-135b and mismatch repair gene expressions via estrogen receptor-β in colorectal cells. Exp Mol Med. 2012 Dec 31;44(12):723-32.

123. Heinemann V, Stintzing S, Kirchner T, Boeck S, Jung A. Clinical relevance of EGFR- and KRAS-status in colorectal cancer patients treated with monoclonal antibodies directed against the EGFR. Cancer Treat Rev 2009; 35: 262-271

124. Heneghan HM, Miller N, Kelly R, Newell J, Kerin MJ. Systemic miRNA-195 differentiates breast cancer from other malignancies and is a potential biomarker for detecting noninvasive and early stage disease. Oncologist. 2010; 15(7):673-82

125. Hewitson P, Glasziou P, Irwig L, Towler B, Watson E. Screening for colorectal cancer using the faecal occult blood test Hemoccult. Cochrane Database Syst Rev.2007; 24(1):CD001216.

126. Hofsli E, Sjursen W, Prestvik WS, et al. Identification of serum microRNA profiles in colon cancer. British journal of cancer. 2013; 108:1712–1719

127. Holme O, Løberg M, Kalager M, Bretthauer M, Hernán M, Aas E, Eide T, Skovlund E, Schneede J, Tveit K, & Hoff G. Effect of Flexible Sigmoidoscopy Screening on Colorectal Cancer Incidence and Mortality. JAMA 2014; 312 (6)

128. Hood JL, San RS, Wickline SA. Exosomes released by melanoma cells prepare sentinel lymph nodes for tumor metastasis. Cancer Res. 2011;71:3792–801.

129. Hromadníková I, Kotlabová K, Jirásek JE, Doucha J. Detection of placenta-specific microRNAs in maternal circulation. Ceska Gynekol. 2010 May; 75(3):252-6.

130. Hu G, Chen D, Li X, Yang K, Wang H, Wu W. miR-133b regulates the MET proto-oncogene and inhibits the growth of colorectal cancer cells in vitro and in vivo. Cancer Biol Ther 2010, 10(2):190–197.

131. Hu G, Drescher KM, Chen XM. Exosomal miRNAs: Biological Properties and Therapeutic Potential. Front Genet. 2012;3:56.

275

132. Huang Q, Gumireddy K, Schrier M, le Sage C, Nagel R, Nair S, Egan DA, Li A, Huang G, Klein-Szanto AJ, Gimotty PA, Katsaros D, Coukos G, Zhang L, Puré E, Agami R. The microRNAs miR-373 and miR-520c promote tumour invasion and metastasis. Nat Cell Biol. 2008; 10(2):202-10.

133. Huang Z, Huang D, Ni S, Peng Z, Sheng W, Du X. Plasma microRNAs are promising novel biomarkers for early detection of colorectal cancer. Int J Cancer. 2010.1; 127(1):11826

134. Huang Z, Huang S, Wang Q, et al: MicroRNA-95 promotes cell proliferation and targets sorting Nexin 1 in human colorectal carcinoma. Cancer Res 2011, 71(7):2582–2589.

135. Huber V, Fais S, Iero M, Lugini L, Canese P, Squarcina P, Zaccheddu A, Colone M, Arancia G, Gentile M, Seregni E, Valenti R, Ballabio G, Belli F, Leo E, Parmiani G, Rivoltini. Human colorectal cancer cells induce T-cell death through release of proapoptotic microvesicles: role in immune escape. Gastroenterology. 2005; 128(7):1796-804

136. Hui AB, Shi W, Boutros PC, Miller N, Pintilie M, Fyles T, McCready D, Wong D, Gerster K, Waldron L, Jurisica I, Penn LZ, Liu FF. 2009. Robust global micro-RNA profiling with formalinfixed paraffin-embedded breast cancer tissues. Lab Invest 89:597-606

137. Identification of N10-substituted phenoxazines as potent and specific inhibitors of Akt signaling. J Biol Chem. 2005 Sep 9;280(36):31924-35.

138. Infante JR, Bendell JC, Goff LW, Jones SF, Chan E, Sudo T, Burris HA, Berlin JD. Safety, pharmacokinetics and pharmacodynamics of the anti-A33 fully-human monoclonal antibody, KRN330, in patients with advanced colorectal cancer. Eur J Cancer. 2013 Apr;49(6):1169-75.

139. Shaw JA, Page K, Blighe K, Hava N, Guttery D, Ward B, Brown J, Ruangpratheep C, Stebbing J, Payne R, Palmieri C, Cleator S, Walker RA, Coombes RC. Genomic analysis of circulating cell-free DNA infers breast cancer dormancy. Genome Res. 2012 Feb;22(2):220-31.

140. Jasperson KW, Tuohy TM, Neklason DW, Burt RW. Hereditary and familial colon cancer.Gastroenterology. 2010 Jun;138(6):2044-58.

141. Jass JR. Pathogenesis of colorectal cancer. Surg Clin North Am 2002; 82: 891-904

142. Jeffreys AJ, Wilson V, Thein SL. Hypervariable ‘minisatellite’ regions in human DNA. Nature 1985; 314: 67-73

143. Jiang L, Huang Q, Chang J, Wang E, Qiu X. MicroRNA HSA-miR-125a-5p induces apoptosis by activating p53 in lung cancer cells. Exp Lung Res 2011; 37: 387–98.

144. Jones K, Nourse JP, Keane C, Bhatnagar A, Gandhi MK. Plasma microRNA are disease response biomarkers in classical Hodgkin lymphoma. Clinical cancer research : an official journal of the American Association for Cancer Research. 2014; 20:253–264

276

145. Jung EJ, Santarpia L, Kim J, Esteva FJ, Moretti E, Buzdar AU. Plasma microRNA 210 levelscorrelate with sensitivity to trastuzumab and tumor presence in breast cancer patients. Cancer.2012; 118(10):2603–14.

146. Jung HS, Seo YR, Yang YM, Koo JH, An J, Lee SJ, Kim KM, Kim SG. Gα12gep oncogene inhibits FOXO1 in hepatocellular carcinoma as a consequence of miR-135b and miR-194 dysregulation. Cell Signal. 2014 Jul;26(7):1456-65.

147. Kalimutho M, Del Vecchio Blanco G, Di Cecilia S, Sileri P, Cretella M, Pallone F, Federici G, Bernardini S. Differential expression of miR-144* as a novel fecal-based diagnostic marker for colorectal cancer. J Gastroenterol 2011;46: 1391–1402.

148. Kanaan Z, Rai SN, Eichenberger MR, Roberts H, Keskey B, Pan J, Galandiuk S.Plasma miR-21: a potential diagnostic marker of colorectal cancer. Ann Surg. 2012 Sep;256(3):544-51

149. Kanaan Z, Roberts H, Eichenberger MR, Billeter A, Ocheretner G, Pan J. A plasma microRNA panel for detection of colorectal adenomas: a step toward more precise screening for colorectal cancer. Annals of surgery. 2013; 258(3):400–8.

150. Kawaguchi T, Komatsu S, Ichikawa D, Morimura R, Tsujiura M. Clinical impact of circulating miR-221 in plasma of patients with pancreatic cancer. Br J Cancer 2013; 108: 361–369.

151. Kawamura M, Toiyama Y, Tanaka K, Inoue Y, Mohri Y, Kusunoki M. Can circulating microRNAs become the test of choice for colorectal cancer? Curr. Colorectal Cancer Rep. 2014, 10, 403–410

152. Keller S, Sanderson MP, Stoeck A. Altevogt P. Exosomes: from biogenesis and secretion to biological function. Immunol Lett. 2006; 107: 102–108

153. Kelley RK, Venook AP. Prognostic and predictive markers in stage II colon cancer: is there a role for gene expression profiling? Clin Colorectal Cancer 2011; 10: 73–80.

154. Khare S, Verma M. Epigenetics of colon cancer. Methods Mol Biol 2012; 863: 177-185

155. Khraiwesh B, Arif MA, Seumel GI, Ossowski S, Weigel D, Reski R, Frank W. Transcriptional control of gene expression by microRNAs. Cell. 2010.8; 140(1):111-22.

156. Kjaer-Frifeldt S, Hansen TF, Nielsen BS. The prognostic importance of miR-21 in stage II colon cancer: a population-based study. British journal of cancer. 2012; 107:1169–1174.

157. Kjersem JB, Ikdahl T, Lingjaerde OC, Guren T, Tveit KM, Kure EH. Plasma microRNAs predicting clinical outcome in metastatic colorectal cancer patients receiving first-line oxaliplatin-based treatment. Mol Oncol. 2014 Feb;8(1):59-67.

158. Köberle V, Pleli T, Schmithals C, Augusto Alonso E, Haupenthal J, Bönig H, Peveling-Oberhag J, Biondi RM, Zeuzem S, Kronenberger B, Waidmann O, Piiper A.

277

Differential stability of cell-free circulating microRNAs: implications for their utilization as biomarkers. PLoS One. 2013 Sep 20;8(9):e75184.

159. Koga Y, Yasunaga M, Takahashi A, Kuroda J, Moriya Y, Akasu T, Fujita S, Yamamoto S, Baba H, Matsumura Y MicroRNA expression profiling of exfoliated colonocytes isolated from feces for colorectal cancer screening. Cancer Prev Res (Phila). 2010; 3(11):1435-42

160. Kosaka N, Iguchi H, Hagiwara K, Yoshioka Y, Takeshita F, Ochiya T. Neutral sphingomyelinase 2 (nSMase2)-dependent exosomal transfer of angiogenic microRNAs regulate cancer cell metastasis. J Biol Chem. 2013;288:10849–59.

161. Kotlabova K, Doucha J, Hromadnikova I. Placental-specific microRNA in maternal circulation--identification of appropriate pregnancy-associated microRNAs with diagnostic potential. J Reprod Immunol. 2011 May;89(2):185-91.

162. Krichevsky AM1, Gabriely G. miR-21: a small multi-faceted RNA. J Cell Mol Med. 2009 Jan;13(1):39-53.

163. Kroh EM, Parkin RK, Mitchell PS, Tewari M. Analysis of circulating microRNA biomarkers in plasma and serum using quantitative reverse transcription-PCR (qRTPCR).Methods. 2010; 50(4):298-301.

164. Kronborg O, Fenger C, Olsen J, Jørgensen OD, Søndergaard O. Randomised study of screening for colorectal cancer with faecal-occult-blood test. Lancet. 1996 Nov 30;348(9040):1467-71.

165. Kuchenbauer F, Morin RD, Argiropoulos B. In-depth characterization of the microRNA transcriptome in a leukemia progression model. Genome research. 2008; 18:1787–1797.

166. Kulda V, Pesta M, Topolcan O. Relevance of miR-21 and miR-143 expression in tissue samples of colorectal carcinoma and its liver metastases. Cancer Genet Cytogenet 2010; 200: 154–60.

167. Landgraf P, Rusu M, Sheridan R. A mammalian microRNA expression atlas based on small RNA library sequencing. Cell. 2007; 129:1401–1414.

168. Lanza G, Ferracin M, Gafà R, Veronese A, Spizzo R, Pichiorri F, Liu CG, Calin GA, Croce CM, Negrini M mRNA/microRNA gene expression profile in microsatellite unstable colorectal cancer. Mol Cancer. 2007. 23; 6:54

169. Le MT, Hamar P, Guo C, Basar E, Perdigao-Henriques R, Balaj L, et al. miR200-containing extracellular vesicles promote breast cancer cell metastasis. J Clin Invest. 2014;124:5109–28.

170. Lee DY, Deng Z, Wang CH, Yang BB. MicroRNA-378 promotes cell survival, tumor growth, and angiogenesis by targeting SuFu and Fus-1 expression. Proc Natl Acad Sci U S A. 2007.18; 104(51):20350-5

278

171. Lee LW, Zhang S, Etheridge A. Complexity of the microRNA repertoire revealed by nextgeneration sequencing. Rna. 2010; 16:2170–2180.

172. Lee RC, Feinbaum RL, Ambros V. The C. elegans heterochronic gene lin-4 encodes small RNAs with antisense complementarity to lin-14. Cell. 1993.3; 75(5):843-54.

173. Lee Y, Ahn C, Han J, Choi H, Kim J, Yim J, Lee J, Provost P, Rådmark O, Kim S, Kim VN. The nuclear RNase III Drosha initiates microRNA processing. Nature. 2003.25; 425(6956):415-9.

174. Leggett B, Whitehall V. Role of the serrated pathway in colorectal cancer pathogenesis. Gastroenterology 2010; 138: 2088-2100

175. Leggett B, Whitehall V. Role of the serrated pathway in colorectal cancer pathogenesis. Gastroenterology. 2010 Jun;138(6):2088-100.

176. Leidinger P, Backes C, Dahmke IN, Galata V, Huwer H, Stehle I, et al. What makes a blood cell based miRNA expression pattern disease specific? - A miRNome analysis of blood cell subsets in lung cancer patients and healthy controls. Oncotarget. 2014;5:9484–97.

177. Li LN, Zhang HD, Zhi R. Down-regulation of some miRNAs by degrading their precursors contributes to anti-cancer effect of mistletoe lectin-I. Br J Pharmacol 2011, 162(2):349–364.

178. Li Y, Xu D, Bao C, Zhang Y, Chen D, Zhao F, Ding J, Liang L, Wang Q, Liu L, Li J, Yao M, Huang S, He X. MicroRNA-135b, a HSF1 target, promotes tumor invasion and metastasis by regulating RECK and EVI5 in hepatocellular carcinoma. Oncotarget. 2015 Feb 10;6(4):2421-33.

179. Li YC, Korol AB, Fahima T, Beiles A, Nevo E. Microsatellites: genomic distribution, putative functions and mutational mechanisms: a review. Mol Ecol 2002; 11: 2453-2465

180. Li Y, Kowdley KV. Method for microRNA isolation from clinical serum samples. Anal. Biochem. 2012, 431, 69–75

181. Libby G, Brewster DH, McClements PL, Carey FA, Black RJ, Birrell J, Fraser CG, Steele RJ. The impact of population-based faecal occult blood test screening on colorectal cancer mortality: a matched cohort study. Br J Cancer. 2012 Jul 10;107(2):255-9.

182. Lin L, Fuchs J, Li C, Olson V, Bekaii-Saab T, Lin J. STAT3 signaling pathway is necessary for cell survival and tumorsphere forming capacity in ALDH⁺/CD133⁺ stem cell-like human colon cancer cells. Biochem Biophys Res Commun. 2011 Dec 16;416(3-4):246-51

183. Link A, Balaguer F, Shen Y, Nagasaka T, Lozano JJ, Boland CR, Goel A. Fecal MicroRNAs as novel biomarkers for colon cancer screening. Cancer Epidemiol Biomarkers Prev. 2010;19(7):1766-74

279

184. Liu H, Du L, Wen Z et al. Up-regulation of miR-182 expression in colorectal cancer tissues and its prognostic value. Int J Colorectal Dis 2013; 28: 697–703.

185. Liu L, Chen L, Xu Y, Li R, Du X. microRNA-195 promotes apoptosis and suppresses tumorigenicity of human colorectal cancer cells. Biochem Biophys Res Commun 2010, 400(2):236–240.

186. Liu M, Lang N, Chen X, Tang Q, Liu S, Huang J, Zheng Y, Bi F. miR-185 targets RhoA and Cdc42 expression and inhibits the proliferation potential of human colorectal cells. Cancer Lett 2011, 301(2):151–160.

187. Liu M, Lang N, Qiu M, Xu F, Li Q, Tang Q, Chen J, Chen X, Zhang S, Liu Z, Zhou J, Zhu Y, Deng Y, Zheng Y, Bi F. miR-137 targets Cdc42 expression, induces cell cycle G1 arrest and inhibits invasion in colorectal cancer cells. Int J Cancer. 2011.15; 128(6):1269-79.

188. Llorens F, Banez-Coronel M, Pantano L, del Río JA, Ferrer I, Estivill X, Martí E. A highly expressed miR-101 isomiR is a functional silencing small RNA. BMC genomics. 2013; 14:104.

189. Lo RS, Wotton D, Massagué J. Epidermal growth factor signaling via Ras controls the Smad transcriptional corepressor TGIF. EMBO J 2001; 20: 128-136.

190. Logozzi M, De Milito A, Lugini L, Borghi M, Calabrò L, Spada M, Perdicchio M, Marino ML, Federici C, Iessi E, Brambilla D, Venturi G, Lozupone F, Santinami M, Huber V, Maio M, Rivoltini L, Fais S. High levels of exosomes expressing CD63 and caveolin-1 in plasma of melanoma patients. PLoS One. 2009; 4(4):e5219

191. Loktionov A, Bandaletova T, Llewelyn AH, Dion C, Lywood HG, Lywood RC, Rockall TA, Stebbing JF, Broughton M, Caffarey S, Marks CG. Colorectal cancer detection by measuring DNA from exfoliated colonocytes obtained by direct contact with rectal mucosa. Int J Oncol. 2009; 34(2):301-11.

192. Loktionov A. Cell exfoliation in the human colon: myth, reality and implications for colorectal cancer screening. Int J Cancer. 2007.1; 120(11):2281-9.

193. Lu LF, Liston A. MicroRNA in the immune system, microRNA as an immune system. Immunology 2009; 127: 291–298.

194. Lu Z, Liu M, Stribinskis V, Klinge CM, Ramos KS, Colburn NH, et al. MicroRNA-21 promotes cell transformation by targeting the programmed cell death 4 gene. Oncogene. 2008;27:4373–9.

195. Maragkakis M, Alexiou P, Papadopoulos GL, Reczko M, Dalamagas T, Giannopoulos G, Goumas G, Koukis E, Kourtis K, Simossis VA, Sethupathy P, Vergoulis T, Koziris N, Sellis T, Tsanakas P, Hatzigeorgiou AG. Accurate microRNA target prediction correlates with protein repression levels. BMC Bioinformatics. 2009 Sep 18;10:295.

196. Ma Q, Yang L, Wang C, Yu YY, Zhou B, Zhou ZG. Differential expression of colon cancer microRNA in microarry study. Sichuan Da Xue Xue Bao Yi Xue Ban 2011; May;42(3):344-8

280

197. Macabeo-Ong M, Ginzinger DG, Dekker N, McMillan A, Regezi JA, Wong DT, Jordan RC. 2002. Effect of duration of fixation on quantitative reverse transcription polymerase chain reaction analyses. Mod Pathol 15:979-987.

198. Madhavan D, Zucknick M, Wallwiener M, Cuk K, Modugno C, et al. (2012) Circulating miRNAs as surrogate markers for circulating tumor cells and prognostic markers in metastatic breast cancer. Clin Cancer Res 18: 5972–5982

199. Marchand L. Combined influence of genetic and dietary factors on colorectal cancer incidence in Japanese Americans. J Natl Cancer Inst Monogr, 1999. 26: p. 101-105.

200. Marsh DJ, Roth S, Lunetta KL, Hemminki A, Dahia PL, Sistonen P, Zheng Z, Caron S, van Orsouw NJ, Bodmer WF, Cottrell SE, Dunlop MG, Eccles D, Hodgson SV, Järvinen H, Kellokumpu I, Markie D, Neale K, Phillips R, Rozen P, Syngal S, Vijg J, Tomlinson IP, Aaltonen LA, Eng C. Exclusion of PTEN and 10q22-24 as the susceptibility locus for juvenile polyposis syndrome. Cancer Res. 1997 Nov 15;57(22):5017-21.

201. Mathivanan S, Lim JW, Tauro BJ, Ji H, Moritz RL, Simpson RJ. Proteomics analysis of A33 immunoaffinity-purified exosomes released from the human colon tumor cell line LIM1215 reveals a tissue-specific protein signature. Mol Cell Proteomics. 2010 Feb;9(2):197-208

202. McDonald JS, Milosevic D, Reddi HV, Grebe SK, Algeciras-Schimnich A. Analysis of circulating microRNA: preanalytical and analytical challenges. Clin Chem. 2011 Jun;57(6):833-40

203. Melo SA, Sugimoto H, O'Connell T, Noritoshi K, Villanueva A, Vidal A, et al. Cancer Exosomes Perform Cell-Independent MicroRNA Biogenesis and Promote Tumorigenesis. Cancer Cell. 2014;26:707–21

204. Meyer SU, Pfaffl MW, Ulbrich SE. Normalization strategies for microRNA profiling experiments: a 'normal' way to a hidden layer of complexity? Biotechnol Lett. 2010 ;32(12):1777-88

205. Michael MZ, O' Connor SM, van Holst Pellekaan NG, Young GP, James RJ. Reduced accumulation of specific microRNAs in colorectal neoplasia. Mol Cancer Res. 2003 ;1(12):882-91.

206. Midgley R, Kerr D. Colorectal cancer. Lancet 1999; 353: 391–9.

207. Migliore C, Giordano S. Resistance to targeted therapies: a role for microRNAs? Trends Mol Med. 2013;19:633–42.

208. Mitchell PS, Parkin RK, Kroh EM, Fritz BR, Wyman SK, Pogosova-Agadjanyan EL, Peterson A, Noteboom J, O’Briant KC, Allen A, Lin DW,Urban N, Drescher CW, Knudsen BS, Stirewalt DL, Gentleman R,Vessella RL, Nelson PS, Martin DB, Tewari M. Circulating microRNAs as stable blood-based markers for cancer detection. Proc Natl Acad Sci USA 2008; 105: 10513–10518

281

209. Monleau M, Bonnel S, Gostan T, Blanchard D, Courgnaud V, Lecellier CH. Comparison of different extraction techniques to profile microRNAs from human sera and peripheral blood mononuclear cells. BMC Genomics. 2014 May 23;15:395

210. Monzo M, Navarro A, Bandres E, Artells R, Moreno I, Gel B, Ibeas R, Moreno J, Martinez F, Diaz T, Martinez A, Balagué O, Garcia-Foncillas J. Overlapping expression of microRNAs in human embryonic colon and colorectal cancer. Cell Res. 2008 Aug;18(8):823-33.

211. Moon BS, Jeong WJ, Park J, Kim TI, Min do S, Choi KY. Role of oncogenic K-Ras in cancer stem cell activation by aberrant Wnt/β-catenin signaling. J Natl Cancer Inst 2014; 106:373

212. Morin RD, O'Connor MD, Griffith M, Kuchenbauer F, Delaney A, Prabhu AL, Zhao Y, McDonald H, Zeng T, Hirst M, Eaves CJ, Marra MA. Application of massively parallel sequencing to microRNA profiling and discovery in human embryonic stem cells. Genome Res. 2008 Apr;18(4):610-21

213. Motoyama K, Inoue H, Takatsuno Y. Over- and under-expressed microRNAs in human colorectal cancer. Int J Oncol 2009; 34: 1069–75.

214. Mu W, Rana S, Zoller M. Host matrix modulation by tumor exosomes promotes motility and invasiveness. Neoplasia. 2013;15:875–87.

215. Mudduluru G, Ceppi P, Kumarswamy R. Regulation of Axl receptor tyrosine kinase expression by miR-34a and miR-199a/b in solid cancer. Oncogene 2011, 30(25):2888-99

216. Munoz JL, Bliss SA, Greco SJ, Ramkissoon SH, Ligon KL, Rameshwar P.Delivery of Functional Anti-miR-9 by Mesenchymal Stem Cell-derived Exosomes to Glioblastoma Multiforme Cells Conferred Chemosensitivity. Mol Ther Nucleic Acids. 2013;2:e126.

217. Myklebust MP, Bruland O, Fluge Ø, Skarstein A, Balteskard L, Dahl O. MicroRNA-15b is induced with E2F-controlled genes in HPV-related cancer. Br J Cancer 2011; 105: 1719–25.

218. Nagel R, le Sage C, B et al. Regulation of the adenomatous polyposis coli gene by the miR-135 family in colorectal cancer. Cancer Res 2008; 68: 5795–802.

219. Nakajima G, Hayashi K, Xi Y, Kudo K, Uchida K, Takasaki K, Yamamoto M, Ju J. Noncoding MicroRNAs hsa-let-7g and hsa-miR-181b are Associated with hemoresponse to S-1 in Colon Cancer. Cancer Genomics Proteomics. 2006;3(5):317324

220. National Statistics UK. Colorectal Cancer England 2009. http://www.ons.gov.uk/ons/rel/cancer-unit/bowel-cancer-in-england/2009/sum-colorectal.html. (Accessed on 30 January, 2016)

221. NCI-SEER, 2006. Incidence of colon and rectum cancer. http://seer.cancer.gov/archive/csr/1975_2006/results_merged/sect_06_colon_rectum.pdf. (Accessed on 30 January, 2016).

282

222. Necela BM, Carr JM, Asmann YW, Thompson EA. Differential expression of microRNAs in tumors from chronically inflamed or genetic (APC(Min/+)) models of colon cancer. PLoS ONE. 2011;6:e18501.

223. Ng EK, Chong WW, Jin H, Lam EK, Shin VY, Yu J, Poon TC, Ng SS, Sung JJ. Differential expression of microRNAs in plasma of patients with colorectal cancer: a potential marker for colorectal cancer screening. Gut. 2009;58(10):1375-81

224. Ng EK, Chong WW, Jin H, Lam EK, Shin VY, Yu J, Poon TC, Ng SS, Sung JJ. Differential expression of microRNAs in plasma of patients with colorectal cancer: a potential marker for colorectal cancer screening. Gut. 2009;58(10):1375-81

225. Ng EK, Tsang WP, Ng SS. MicroRNA-143 targets DNA methyltransferases 3A in colorectal cancer. Br J Cancer 2009, 101(4):699–706.

226. Nielsen BS, Jørgensen S, Fog JU. High levels of microRNA-21 in the stroma of colorectal cancers predict short disease-free survival in stage II colon cancer patients. Clin Exp Metastasis 2011; 28: 27–38.

227. Nnoaham KE, Lines C. Modelling future capacity needs and spending on colonoscopy in the English bowel cancer screening programme. Gut 2008; 57: 1238–1245.

228. Nugent M, Miller N, Kerin MJ. Circulating miR-34a levels are reduced in colorectal cancer. Journal of surgical oncology. 2012; 106(8):947–52.

229. O'Brien K, Rani S, Corcoran C, Wallace R, Hughes L, Friel AM. Exosomes from triple-negative breast cancer cells can transfer phenotypic traits representing their cells of origin to secondary cells. Eur J Cancer.2013;49:1845–59.

230. O'Connor ES, Greenblatt DY, LoConte NK, Gangnon RE, Liou JI, Heise CP, Smith MA. Adjuvant chemotherapy for stage II colon cancer with poor prognostic features. J Clin Oncol. 2011 Sep 1;29(25):3381-8.

231. Oda K, Matsuoka Y, Funahashi A, Kitano H. A comprehensive pathway map of epidermal growth factor receptor signaling. Mol Syst Biol 2005; 1: 2005.0010

232. Ofir M, Hacohen D, Ginsberg D. MiR-15 and miR-16 are direct transcriptional targets of E2F1 that limit E2Finduced proliferation by targeting cyclin E. Mol Cancer Res 2011; 9: 440–7.

233. Ogata-Kawata H, Izumiya M, Kurioka D, Honma Y, Yamada Y, Furuta K. Circulating exosomal microRNAs as biomarkers of colon cancerPLoS One. 2014 Apr 4;9(4):e92921.

234. Ohshima K, Inoue K, Fujiwara A, Hatakeyama K, Kanto K, Watanabe Y, et al. Let-7 microRNA family is selectively secreted into the extracellular environment via exosomes in a metastatic gastric cancer cell line. PLoSOne. 2010;5:e13247.

283

235. Okamura K, Phillips MD, Tyler DM, Duan H, Chou YT, Lai EC. The regulatory activity of microRNA* species has substantial influence on microRNA and 3' UTR evolution. Nat Struct Mol Biol. 2008;15(4):354-63

236. Ørom UA, Nielsen FN, & Lund AH. MicroRNA-10a binds the 5′UTR of ribosomal protein mRNAs and enhances their translation. Mol. Cell: 2008(4); 460–471

237. Ostenfeld MS, Jeppesen DK, Laurberg JR, Boysen AT, Bramsen JB, PrimdalBengtson B, et al. Cellular Disposal of miR23b by RAB27-Dependent Exosome Release Is Linked to Acquisition of Metastatic Properties. Cancer Res. 2014;74:5758–71.

238. Otori K, Oda Y, Sugiyama K, Hasebe T, Mukai K, Fujii T, Tajiri H, Yoshida S, Fukushima S, Esumi H. High frequency of K-ras mutations in human colorectal hyperplastic polyps. Gut 1997; 40: 660-663

239. Oue N, Anami K, Schetter AJ, Schetter AJ, Moehler M, Okayama H, Khan MA, Bowman ED, Mueller A, Schad A, Shimomura M, Hinoi T, Aoyagi K, Sasaki H, Okajima M, Ohdan H, Galle PR, Yasui W, Harris CC. High miR-21 expression from FFPE tissues is associated with poor survival and response to adjuvant chemotherapy in colon cancer. International journal of cancer 2014; 134:1926–1934

240. Paget S. The distribution of secondary growths in cancer of the breast. Cancer Metastasis Rev 1998; 8: 98–101.

241. Pal JK, Ray SS, Pal SK. Identifying relevant group of miRNAs in cancer using fuzzy mutual information. Med Biol Eng Comput. 2015 Aug 12

242. Pancione M, Remo A, Colantuoni V. Genetic and epigenetic events generate multiple pathways in colorectal cancer progression. Patholog Res Int 2012; 2012: 509348

243. Pei H, Jin Z, Chen S, Sun X, Yu J, Guo W. MiR-135b promotes proliferation and invasion of osteosarcoma cells via targeting FOXO1.Mol Cell Biochem. 2015 Feb;400(1-2):245-52

244. Pekow JR, Dougherty U, Mustafi R, Zhu H, Kocherginsky M, Rubin DT, Hanauer SB, Hart J, Chang EB, Fichera A, Joseph LJ, Bissonnette M. miR-143 and miR-145 are downregulated in ulcerative colitis: putative regulators of inflammation and protooncogenes. Inflamm Bowel Dis. 2012 Jan;18(1):94-100.

245. Perez-Carbonell L1, Sinicrope FA2, Alberts SR2, Oberg AL3, Balaguer F4, Castells A4, Boland CR1, Goel A1. MiR-320e is a novel prognostic biomarker in CRC. Br J Cancer. 2015 Jun 30;113(1):83-90

246. Perkins A, Nicholls K, Shaw T, Liu G, Molokhia E. Attitudes toward colorectal cancer screening in the digital age: a survey of practices and attitudes among screening-eligible Alabamians. South Med J. 2013 Aug;106(8):462-7.

247. Petersen VC, Baxter KJ, Love SB et al. Identification of objective pathological prognostic determinants and models of prognosis in Dukes’ B colon cancer. Gut 2002; 51: 65– 9.

284

248. Pichler M, Winter E, Stotz M, Eberhard K, Samonigg H, Lax S, Hoefler G. Down-regulation of KRAS-interacting miRNA-143 predicts poor prognosis but not response to EGFR-targeted agents in colorectal cancer. Br J Cancer 2012; 106: 1826–1832.

249. Pinar G, Ayhan A. Carcinomas associated with Lynch syndrome: a family history. Int Surg. 2011 Oct-Dec;96(4):286-90.

250. Poulogiannis G, Frayling IM, Arends MJ. DNA mismatch repair deficiency in sporadic colorectal cancer and Lynch syndrome. Histopathology 2010; 56: 167-179

251. Pradervand S, Weber J, Lemoine F, Consales F, Paillusson A, Dupasquier M, Thomas J, Richter H, Kaessmann H, Beaudoing E, Hagenbuchle O, Harshman K. Concordance among digital gene expression, microarrays, and qPCR when measuring differential expression of microRNAs. Biotechniques.2010;48:219–222

252. Pritchard CC, Kroh E, Wood B, Arroyo JD, Dougherty KJ, Miyaji MM, Tait JF, Tewari M. Blood cell origin of circulating microRNAs: a cautionary note for cancer biomarker studies. Cancer Prev Res (Phila). 2012 Mar;5(3):492-7.

253. Pu XX, Huang GL, Guo HQ, Guo CC, Li H, Ye S, Ling S, Jiang L, Tian Y, Lin TY. Circulating miR-221 directly amplified from plasma is a potential diagnostic and prognostic marker of colorectal cancer and is correlated with p53 expression. J Gastroenterol Hepatol. 2010 Oct;25(10):1674-80.

254. Que R, Ding G, Chen J, Cao L. Analysis of serum exosomal microRNAs and clinicopathologic features of patients with pancreatic adenocarcinoma. World J Surg Oncol. 2013;11:219.

255. Rabinowits G, Gerçel-Taylor C, Day JM, Taylor DD, Kloecker GH. Exosomal microRNA: a diagnostic marker for lung cancer. Clin Lung Cancer. 2009;10(1):42-6

256. Rana S, Malinowska K, Zoller M. Exosomal tumor microRNA modulates premetastatic organ cells. Neoplasia. 2013;15:281–95.

257. Ren J, Zhang J, Xu N, Han G, Geng Q, Song J, Li S, Zhao J, Chen H. Signature of circulating microRNAs as potential biomarkers in vulnerable coronary artery disease. PLoS One. 2013 Dec 5;8(12):e80738.

258. Rossi L, Bonmassar E, Faraoni I. Modification of miR gene expression pattern in human colon cancer cells following exposure to 5-fluorouracil in vitro. Pharmacol Res. 2007;56(3):248-53

259. Rouet F, Rouzioux C. HIV-1 viral load testing cost in developing countries: what’s new? Expert Rev Mol Diagn 2007; 7: 703–707.

260. Salby O, Svoboda M, Michalek J, Vyzula R. MicroRNAs in colorectal cancer: translation of molecular biology into clinical application. Mol Cancer 2009; 14: 8–102

261. Sanchez-Diaz PC, Hsiao TH, Chang JC, Yue D, Tan MC, Chen HI, Tomlinson GE, Huang Y, Chen Y, Hung JY. De-regulated microRNAs in pediatric cancer stem cells target

285

pathways involved in cell proliferation, cell cycle and development. PLoS One. 2013 Apr 17;8(4):e61622

262. Sarver AL, French AJ, Borralho PM, Thayanithy V, Oberg AL, Silverstein KA, Morlan BW, Riska SM, Boardman LA, Cunningham JM, Subramanian S, Wang L, Smyrk TC, Rodrigues CM, Thibodeau SN, Steer CJ. Human colon cancer profiles show differential microRNA expression depending on mismatch repair status and are characteristic of undifferentiated proliferative states. BMC Cancer. 2009 Nov 18;9:401

263. Sato-Kuwabara Y, Melo SA, Soares FA, Calin GA. The fusion of two worlds: non-coding RNAs and extracellular vesicles--diagnostic and therapeutic implications (Review). Int J Oncol. 2015 Jan;46(1):17-27.

264. Scheffer AR, Holdenrieder S, Kristiansen G, von Ruecker A, Müller SC, Ellinger J. Circulating microRNAs in serum: novel biomarkers for patients with bladder cancer? World J Urol. 2014 Apr;32(2):353-8.

265. Schepeler T, Reinert JT, Ostenfeld MS, Christensen LL, Silahtaroglu AN, Dyrskjøt L, Wiuf C, Sørensen FJ, Kruhøffer M, Laurberg S, Kauppinen S, Ørntoft TF, Andersen CL. Diagnostic and prognostic microRNAs in stage II colon cancer. Cancer Res. 2008.1;68(15):6416-24.

266. Schetter AJ, Leung SY, Sohn JJ, Zanetti KA, Bowman ED, Yanaihara N, Yuen ST, Chan TL, Kwong DL, Au GK, Liu CG, Calin GA, Croce CM, Harris CC. MicroRNA expression profiles associated with prognosis and therapeutic outcome in colon adenocarcinoma. JAMA. 2008.30;299(4):425-36

267. Schetter AJ, Nguyen GH, Bowman ED, Mathé EA, Yuen ST, Hawkes JE, Croce CM, Leung SY, Harris CC. Association of inflammation-related and microRNA gene expression with cancer-specific mortality of colon adenocarcinoma. Clin Cancer Res. 2009 Sep 15;15(18):5878-87.

268. Schetter AJ, Okayama H, Harris CC. The role of microRNAs in colorectal cancer. Cancer J 2012; 18: 244– 52.

269. Schimanski CC, Frerichs K, Rahman F, Berger M, Lang H, Galle PR, Moehler M, Gockel I. High miR-196a levels promote the oncogenic phenotype of colorectal cancer cells. World J Gastroenterol. 2009 May 7;15(17):2089-96.

270. Segditsas S, Tomlinson I. Colorectal cancer and genetic alterations in the Wnt pathway. Oncogene. 2006.4;25(57):7531-7

271. Shibuya H, Iinuma H, Shimada R, Horiuchi A, Watanabe T. Clinicopathological and prognostic value of microRNA21 and microRNA-155 in colorectal cancer. Oncology 2010; 79: 313–20.

272. Si H, Sun X, Chen Y, Cao Y, Chen S, Wang H, Hu C. Circulating microRNA-92a and microRNA-21 as novel minimally invasive biomarkers for primary breast cancer. J Cancer Res Clin Oncol. 2013 Feb;139(2):223-9.

286

273. Simpson RJ, Lim JW, Moritz RL, Mathivanan S. Exosomes: proteomic insights and diagnostic potential. Expert Rev Proteomics. 2009;6(3):267-83

274. Skog J, Würdinger T, van Rijn S, Meijer DH, Gainche L, Sena-Esteves M, Curry WT Jr, Carter BS, Krichevsky AM, Breakefield XO. Glioblastoma microvesicles transport RNA and proteins that promote tumour growth and provide diagnostic biomarkers. Nat Cell Biol. 2008;10:1470–1476.

275. Slaby O, Svoboda M, Fabian P, Smerdova T, Knoflickova D, Bednarikova M, Nenutil R, Vyzula R.Altered expression of miR-21, miR-31, miR-143 and miR-145 is related to clinicopathologic features of colorectal cancer. Oncology. 2007;72(5-6):397-402

276. Slattery ML, Wolff E, Hoffman MD, Pellatt DF, Milash B, Wolff RK MicroRNAs and colon and rectal cancer: differential expression by tumor location and subtype. Genes Chromosomes Cancer. 2011;50(3):196-206

277. Sobin LH, Fleming ID. TNM Classification of Malignant Tumors, fifth edition. Union Internationale Contre le Cancer and the American Joint Committee on Cancer. Cancer. 1997 Nov 1;80(9):1803-4.

278. Song J, Bai Z, Han W, Zhang J, Meng H, Bi J, Ma X, Han S, Zhang Z. Identification of suitable reference genes for qPCR analysis of serum microRNA in gastric cancer patients. Dig Dis Sci. 2012 Apr;57(4):897-904.

279. Spring KJ, Zhao ZZ, Karamatic R, Walsh MD, Whitehall VL, Pike T, Simms LA, Young J, James M, Montgomery GW, Appleyard M, Hewett D, Togashi K, Jass JR, Leggett BA. High prevalence of sessile serrated adenomas with BRAF mutations: a prospective study of patients undergoing colonoscopy. Gastroenterology.2006; 131(5):1400-7.

280. Stahlberg A, Kubista M, Pfaffl M. Comparison of reverse transcriptases in gene expression analysis. Clin. Chem. 2004, 50, 1678–1680.

281. Stark A, Kheradpour P, Parts L, Brennecke J, Hodges E, Hannon GJ, Kellis M. Systematic discovery and characterization of fly microRNAs using 12 Drosophila genomes. Genome Res. 2007;17(12):1865-79

282. Suppiah A, Greenman J. Clinical utility of anti-p53 autoantibody: systematic review and focus on colorectal cancer. World J Gastroenterol 2013; 19: 4651-4670

283. Swarup V, Rajeswari MR. Circulating (cell-free) nucleic acids—a promising, non-invasive tool for early detection of several human diseases. FEBS Lett 2007;581: 795–799.

284. Takahashi K, Yokota S, Tatsumi N, Fukami T, Yokoi T, Nakajima M. Cigarette smoking substantially alters plasma microRNA profiles in healthy subjects. Toxicology and applied pharmacology. 2013; 272:154–160.

285. Tan IB, Tan P. Genetics: an 18-gene signature (ColoPrint®) for colon cancer prognosis. Nat Rev Clin Oncol 2011;8: 131–133.

287

286. Tanaka T, Arai M, Jiang X, Sugaya S, Kanda T, Fujii K, Kita K, Sugita K, Imazeki F, Miyashita T, Kaneda A, Yokosuka O Downregulation of microRNA-431 by human interferon-β inhibits viability of medulloblastoma and glioblastoma cells via upregulation of SOCS6. Int J Oncol. 2014 May;44(5):1685-90.

287. Tanaka T, Sugaya S, Kita K, Arai M, Kanda T, Fujii K, Imazeki F, Sugita K, Yokosuka O, Suzuki N. Inhibition of cell viability by human IFN-β is mediated by microRNA-431. Int J Oncol. 2012 May;40(5):1470-6.

288. Tanaka Y, Kamohara H, Kinoshita K, Kurashige J, Ishimoto T, Iwatsuki M, Watanabe M, Baba H. Clinical impact of serum exosomal microRNA-21 as a clinical biomarker in human esophageal squamous cell carcinoma. Cancer. 2013 Mar 15;119(6):1159-67.

289. Taylor DD, Gercel-Taylor C. Exosomes/exosomes: mediators of cancerassociated immunosuppressive microenvironments. Semin Immunopathol. 2011;33:441–54

290. Taylor DD, Gercel-Taylor C. MicroRNA signatures of tumor-derived exosomes as diagnostic biomarkers of ovarian cancer. Gynecol Oncol. 2008;110(1):13-21

291. Tazawa H, Tsuchiya N, Izumiya M, Nakagama H. Tumor-suppressive miR-34a induces senescence-like growth arrest through modulation of the E2F pathway in human colon cancer cells. Proc Natl Acad Sci USA 2007; 104: 15472–15477

292. Tebbutt NC, Wilson K, Gebski VJ, Cummins MM, Zannino D, van Hazel GA, Robinson B, Broad A, Ganju V, Ackland SP, Forgeson G, Cunningham D, Saunders MP, Stockler MR, Chua Y, Zalcberg JR, Simes RJ, Price TJ. Capecitabine, bevacizumab, and mitomycin in first-line treatment of metastatic colorectal cancer: results of the Australasian Gastrointestinal Trials Group Randomized Phase III MAX Study. J Clin Oncol 2010; 28: 3191-3198

293. Théry C, Zitvogel L, Amigorena S. Exosomes: composition, biogenesis and function. Nat Rev Immunol. 2002; 2: 569–579

294. Thimmaiah KN, Easton JB, Germain GS, Morton CL, Kamath S, Buolamwini JK, Houghton PJ, Thuma F, Zoller M. Outsmart tumor exosomes to steal the cancer initiating cell its niche. Semin Cancer Biol. 2014;28:39–50.

295. Toiyama Y, Takahashi M, Hur K, Nagasaka T, Tanaka K, Inoue Y, et al. Serum miR-21 as adiagnostic and prognostic biomarker in colorectal cancer. Journal of the National Cancer Institute.2013; 105(12):849–59

296. Tsang WP, Ng EK, Ng SS, Jin H, Yu J, Sung JJ, Kwok TT. Oncofetal H19-derived miR-675 regulates tumor suppressor RB in human colorectal cancer. Carcinogenesis. 2010 Mar;31(3):350-8.

297. Tsukamoto O, Miura K, Mishima H, Abe S, Kaneuchi M, Higashijima A, Miura S, Kinoshita A, Yoshiura K, Masuzaki H. Identification of endometrioid endometrial carcinoma-associated microRNAs in tissue and plasma. Gynecol Oncol. 2014 Mar;132(3):715-21.

288

298. Ueda R, Kohanbash G, Sasaki K, Fujita M, Zhu X, Kastenhuber ER, McDonald HA, Potter DM, Hamilton RL, Lotze MT, Khan SA, Sobol RW, Okada H. Dicerregulated microRNAs 222 and 339 promote resistance of cancer cells to cytotoxic T-lymphocytes by down-regulation of ICAM-1. Proc Natl Acad Sci U S A. 2009;106:10746–51.

299. Umezu T, Ohyashiki K, Kuroda M, Ohyashiki JH. Leukemia cell to endothelial cell communication via exosomal miRNAs. Oncogene. 2013;32:2747–55.

300. Umezu T, Tadokoro H, Azuma K, Yoshizawa S, Ohyashiki K, Ohyashiki JH. Exosomal miR-135b shed from hypoxic multiple myeloma cells enhances angiogenesis by targeting factor-inhibiting HIF-1. Blood. 2014 Dec 11;124(25):3748-57.

301. Valadi H, Ekström K, Bossios A, Sjöstrand M, Lee JJ, Lötvall JO. Exosome-mediated transfer of mRNAs and microRNAs is a novel mechanism of genetic exchange between cells. Nat Cell Biol. 2007; 9(6):654-9

302. Valeri N, Gasparini P, Braconi C, et al: MicroRNA-21 induces resistance to 5-fluorouracil by down-regulating human DNA MutS homolog 2 (hMSH2). Proc Natl Acad Sci USA 2010, 107(49):21098–21103.

303. Valeri N, Braconi C, Gasparini P, Murgia C, Lampis A, Paulus-Hock V, Hart JR, Ueno L, Grivennikov SI, Lovat F, Paone A, Cascione L, Sumani KM, Veronese A, Fabbri M, Carasi S, Alder H, Lanza G, Gafa' R, Moyer MP, Ridgway RA, Cordero J, Nuovo GJ, Frankel WL, Rugge M, Fassan M, Groden J, Vogt PK, Karin M, Sansom OJ, Croce CM. MicroRNA-135b promotes cancer progression by acting as a downstream effector of oncogenic pathways in colon cancer. Cancer Cell. 2014 Apr 14;25(4):469-83.

304. van Niel G, Raposo G, Candalh C, Boussac M, Hershberg R, Cerf-Bensussan N, Heyman M. Intestinal epithelial cells secrete exosome-like vesicles. Gastroenterology. 2001 ;121(2):337-49

305. Vandesompele J, De Preter K, Pattyn F, Poppe B, Van Roy N, De Paepe A, Speleman F. Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes. Genome Biol. 2002 Jun 18;3(7):RESEARCH0034.

306. Verma AM, Patel M, Aslam MI, Jameson J, Pringle JH, Wurm P, Singh B. Circulating plasma microRNAs as a screening method for detection of colorectal adenomas. Lancet. 2015 Feb 26;385 Suppl 1:S100.

307. Vickers MM, Bar J, Gorn-Hondermann I, Yarom N, Daneshmand M, Hanson JE, Addison CL, Asmis TR, Jonker DJ, Maroun J, Lorimer IA, Goss GD, Dimitroulakos J. Stage-dependent differential expression of microRNAs in colorectal cancer: potential role as markers of metastatic disease. Clin Exp Metastasis 2012; 29: 123–132.

308. Viguera E, Canceill D, Ehrlich SD. Replication slippage involves DNA polymerase pausing and dissociation. EMBO J 2001; 20: 2587-2595

309. Vogelstein B, Fearon ER, Hamilton SR, Kern SE, Preisinger AC, Leppert M, Nakamura Y, White R, Smits AM, Bos JL. Genetic alterations during colorectal-tumor development. N Engl J Med. 1988; 319(9):525-32

289

310. Vogelstein B, Kinzler KW. Cancer genes and the pathways they control. Nat Med 2004; 10: 789-799

311. Volinia S, Calin GA, Liu CG, Ambs S, Cimmino A, Petrocca F, Visone R, Iorio M, Roldo C, Ferracin M, Prueitt RL, Yanaihara N, Lanza G, Scarpa A, Vecchione A, Negrini M, Harris CC, Croce CM. A microRNA expression signature of human solid tumors defines cancer gene targets.Proc Natl Acad Sci U S A. 2006.14;103(7):2257-61

312. Walsh JM, Terdiman JP. Colorectal cancer screening: scientific review. JAMA. 2003;289:1288–1296.

313. Wang B, Zhang Q. The expression and clinical significance of circulating microRNA-21 in serum of five solid tumors. J Cancer Res Clin Oncol. 2012;138:1659–66.

314. Wang CJ, Stratmann J, Zhou ZG, Sun XF. Suppression of microRNA-31 increases sensitivity to 5-FU at an early stage, and affects cell migration and invasion in HCT-116 colon cancer cells. BMC Cancer. 2010 Nov 9;10:616.

315. Wang H, Ach RA, Curry B. Direct and sensitive miRNA profiling from low-input total RNA. RNA. 2007;13:151–159.

316. Wang J, Huang SK, Zhao M, Yang M, Zhong JL, Gu YY, Peng H, Che YQ, Huang CZ. Identification of a circulating microRNA signature for colorectal cancer detection. PLoS One. 2014 Apr 7;9(4):e87451.

317. Wang K, Zhang S, Marzolf B, Troisch P, Brightman A, Hu Z, Hood LE, Galas DJ. Circulating microRNAs, potential biomarkers for drug-induced liver injury. Proc Natl Acad Sci USA 2009;106: 4402– 4407.

318. Wang LG, Gu J.Serum microRNA-29a is a promising novel marker for early detection of colorectal liver metastasis. Cancer Epidemiol 2012a: 36: e61–7.

319. Wang Q, Huang Z, Ni S, Xiao X, Xu Q, Wang L, Huang D, Tan C, Sheng W, Du X. Plasma miR-601 and miR-760 are novelbiomarkers for the early detection of colorectal cancer. PloS one. 2012; 7(9):e44398.

320. Wang S, El-Deiry WS. Requirement of p53 targets in chemosensitization of colonic carcinoma to death ligand therapy. Proc Natl Acad Sci USA 2003; 100: 15095-15100

321. Wang X, Wang J, Ma H, Zhang J, Zhou X. Downregulation of miR-195 correlates with lymph node metastasis and poor prognosis in colorectal cancer. Med Oncol 2012; 29: 919–27.

322. Wang YX, Zhang XY, Zhang BF, Yang CQ, Chen XM, Initial study of microRNA expression profiles of colonic cancer without lymph node metastasis. J Dig Dis. 2010 Feb;11(1):50-4.

323. Wang K, Yuan Y, Cho JH, McClarty S, Baxter D, Galas DJ. Comparing the MicroRNA spectrum between serum and plasma. PLoS One. 2012;7(7):e41561.

290

324. Webber EM, Lin JS, Evelyn P Whitlock. Oncotype DX tumor gene expression profiling in stage II colon cancer. Application: prognostic, risk prediction. PLoS Curr. 2010 Sep 2;2. pii: RRN1177.

325. Weissmann-Brenner A, Kushnir M, Lithwick Yanai G, Aharonov R, Gibori H, Purim O, Kundel Y, Morgenstern S, Halperin M, Niv Y, Brenner B. Tumor microRNA-29a expression and the risk of recurrence in stage II colon cancer. Int J Oncol. 2012 Jun;40(6):2097-103.

326. White V, Scarpini C, Barbosa-Morais NL, Ikelle E, Carter S, Laskey RA, Miller R, Coleman N. Isolation of stool-derived mucus provides a high yield of colonocytes suitable for early detection of colorectal carcinoma. Cancer Epidemiol Biomarkers Prev. 2009 ;18(7):2006-13

327. Wimmer K, Kratz CP, Vasen HF, Caron O, Colas C, EntzWerle N, Gerdes AM, Goldberg Y, Ilencikova D, Muleris M, Duval A, Lavoine N, Ruiz-Ponte C, Slavc I, Burkhardt B, Brugieres L. Diagnostic criteria for constitutional mismatch repair deficiency syndrome: suggestions of the European consortium ‘care for CMMRD’ (C4CMMRD). J Med Genet 2014; 51: 355-365.

328. Witwer KW. XenomiRs and miRNA homeostasis in health and disease: evidence that diet and dietary miRNAs directly and indirectly influence circulating miRNA profiles. RNA Biol. 2012 Sep;9(9):1147-54.

329. Worthley DL, Whitehall VL, Spring KJ, Leggett BA. Colorectal carcinogenesis: road maps to cancer. World J Gastroenterol 2007; 13: 3784-3791

330. Wu CW, Ng SC, Dong Y, Tian L, Ng SS, Leung WW, Law WT, Yau TO, Chan FK, Sung JJ, Yu J. Identification of microRNA-135b in stool as a potential noninvasive biomarker for colorectal cancer and adenoma. Clin Cancer Res. 2014 Jun 1;20(11):2994-3002

331. Xi Y, Formentini A, Chien M, Weir DB, Russo JJ, Ju J, Kornmann M, Ju J. Prognostic Values of microRNAs in Colorectal Cancer. Biomark Insights. 2006;2:113-121

332. Xiao S, Yang Z, Lv R, Zhao J, Wu M, Liao Y, Liu Q. miR-135b contributes to the radioresistance by targeting GSK3β in human glioblastoma multiforme cells. PLoS One. 2014 Sep 29;9(9):e108810.

333. Xing J, Wan S, Zhou F, Qu F, Li B, Myers RE, Fu X, Palazzo JP, He X, Chen Z, Yang H. Genetic polymorphisms in pre-microRNA genes as prognostic markers of colorectal cancer. Cancer Epidemiol Biomarkers Prev 2012; 21: 217–227.

334. Yamakuchi M and Lowenstein CJ. MiR-34, SIRT1 and p53: the feedback loop. Cell Cycle 2009; 8: 712–715

335. Yan LX, Wu QN, Zhang Y, Li YY, Liao DZ, Hou JH, Fu J, Zeng MS, Yun JP, Wu QL, Zeng YX, Shao JY. Knockdown of miR-21 in human breast cancer cell lines inhibits proliferation, in vitro migration and in vivo tumor growth. Breast Cancer Res. 2011 Jan 10;13(1):R2.

291

336. Yang M, Chen J, Su F, Yu B, Su F, Lin L, Liu Y, Huang JD, Song E.Microvesicles secreted by macrophages shuttle invasion-potentiating microRNAs into breast cancer cells. Mol Cancer. 2011 Sep 22;10:117.

337. Ye SB, Li ZL, Luo DH, Huang BJ, Chen YS, Zhang XS, Cui J, Zeng YX, Li J. Tumor-derived exosomes promote tumor progression and T-cell dysfunction through the regulation of enriched exosomal microRNAs in human nasopharyngeal carcinoma. Oncotarget. 2014 Jul 30;5(14):5439-52.

338. Yu H, Pardoll D, Jove R. STATs in cancer inflammation and immunity: a leading role for STAT3. Nat Rev Cancer. 2009 Nov;9(11):798-809.

339. Yuki D, Lin YM, Fujii Y, Nakamura Y, Furukawa Y. Isolation of LEM domain-containing 1, a novel testis-specific gene expressed in colorectal cancers. Oncol. Rep. 2004;12:275–280.

340. Zhang J, Guo H, Zhang H, Wang H, Qian G, Fan X, Hoffman AR, Hu JF, Ge S. Putative tumor suppressor miR-145 inhibits colon cancer cell growth by targeting oncogene Friend leukemia virus integration 1 gene. Cancer. 2011 Jan 1;117(1):86-95.

341. Zhang JX, Song W, Chen ZH, Wei JH, Liao YJ, Lei J, Hu M, Chen GZ, Liao B, Lu J, Zhao HW, Chen W, He YL, Wang HY, Xie D, Luo JH. Prognostic and predictive value of a microRNA signature in stage II colon cancer: a microRNA expression analysis. Lancet Oncol. 2013 Dec;14(13):1295-306.

342. Zhang Y, Zhou ZG, Wang L, Zhang P, Wang MJ, Cui CF, Guan JT, Chen KL, Zhan L. Clinicopathological significance of microRNA-21 and miR-125 expression in colorectal cancer. Zhonghua Wei Chang Wai Ke Za Zhi. 2009 Nov;12(6):623-6.

343. Zhou H, Huang X, Cui H, Luo X, Tang Y, Chen S, Wu L, Shen N. miR-155 and its star-form partner miR-155* cooperatively regulate type I interferon production by human plasmacytoid dendritic cells. Blood. 2010.23;116(26):5885-94.

344. Zhou W, Fong MY, Min Y, Somlo G, Liu L, Palomares MR, Yu Y, Chow A, O'Connor ST, Chin AR, Yen Y, Wang Y, Marcusson EG, Chu P, Wu J, Wu X, Li AX, Li Z, Gao H, Ren X, Boldin MP, Lin PC, Wang SE. Cancer-secreted miR-105 destroys vascular endothelial barriers to promote metastasis. Cancer Cell. 2014 Apr 14;25(4):501-15.

345. Zhu MM, Xu XT, Nie F, Tong JL, Xiao SD, Ran ZH. Comparison of immunochemical and guaiac-based fecal occult blood test in screening and surveillance for advanced colorectal neoplasms: a meta-analysis. J Dig Dis. 2010 Jun;11(3):148-60.

346. Zhuang G, Wu X, Jiang Z, Kasman I, Yao J, Guan Y, Oeh J, Modrusan Z, Bais C, Sampath D, Ferrara N. Tumour-secreted miR-9 promotes endothelial cell migration and angiogenesis by activating the JAK-STAT pathway. EMBO J. 2012 Aug 29;31(17):3513-23.

347. Zlobec I, Lugli A. Prognostic and predictive factors in colorectal cancer. J Clin

Pathol. 2008;61:561–569.