HER Family Receptor Activation and Dimerisation in Colorectal Cancer and Cancer-Derived Exosomes

HER family receptor activation and dimerisation in colorectal cancer and cancer-derived exosomes

Fangfei Gao

A thesis submitted to the University College London for the degree of Doctor of Philosophy

December 2019

Department of Oncology

UCL Cancer Institute

University College London

72 Huntley Street, London, WC1E 6DD, UK

DECLARATION

I, Fangfei Gao confirm that the work presented in this thesis is my own. Where information has been derived from other sources, I confirm that this has been indicated in this thesis.

ABSTRACT

The EGFR pathway is upregulated in several human cancers including colorectal and head and neck cancers. The anti-EGFR antibodies cetuximab and panitumumab are currently widely used in the management of metastatic colorectal cancer with limited duration of response and almost inevitable development of resistance. HER receptors are able to form alternative dimers and can therefore compensate the loss of function of one receptor during targeted therapies. Usage of pan-HER inhibitor such as AZD8931 has been trialed and may represent one way of overcoming resistance to anti-EGFR therapy. However the effects of anti-HER therapy in patients cannot be easily measured without tissue biopsy. Analysis of exosomes offers a potential platform for the longitudinal monitoring of HER signalling in the form of a liquid biopsy by examining their protein and miRNA contents.

Foster Resonance Energy Transfer (FRET) assay using fluorescence lifetime imaging microscopy (FLIM) was used for quantitative analysis of HER dimerisation in colorectal cancer cells. Treatment with cetuximab resulted in increase in HER3 phosphorylation and HER2-HER3 heterodimerisation in the sensitive cell line LIM1215, but not in resistant line DLD1 (KRAS WT). Treatment with AZD8931 resulted in reduced phosphorylation in HER1, HER2 and HER3 in both cell lines, and similar HER dimerisation changes as with treatment with cetuximab. HER3 activation and HER2- HER3 dimer rewiring upon anti-HER therapy could contribute to treatment resistance in colorectal cancer cells.

Cell-derived and circulating exosomes were isolated using differential ultracentrifugation, and characterised using Nanosight and immunostaining. Reverse phase protein array provided an alternative high-throughput platform to dot blot for exosomal protein analysis but requires further optimisation. Exosomal HER protein quantity reflected HER expression in vitro. In a cohort of patients with advanced colorectal cancer patients undergoing first line systemic therapy, circulating exosomal HER3 changes predicted treatment failure prior to radiological progression. In another cohort of patients receiving first line AZD8931, early changes in exosomal HER2-HER3 dimer may be predictive of response to anti-HER therapy. Furthermore, levels of exosomal miR-21 could be used for monitoring of treatment response in patients with colorectal cancer.

IMPACT STATEMENT

Colorectal cancer is the third most common cancer in the world. Despite recent therapeutic advances the disease still causes more than 650,000 deaths worldwide each year. While at early stages surgery is curative, the prognosis for patients with advanced disease and metastatic spread remains bleak with five-year survival rates of 5%. The standard first line systemic therapy for patients with metastatic colorectal cancer without RAS mutation consist of combination chemotherapy with anti-EGFR therapy such as cetuximab or panitumumab. Resistance to anti-EGFR therapy develop due to a number of mechanisms, one of which is through oncogenic receptor signal rewiring. The selection of patients who are likely to benefit from treatment with anti- HER therapy remains controversial. CEA is the only circulating biomarker used in clinic for monitoring colorectal cancer burden but has low sensitivity and specificity. There is increasing demand for better prognostic and predictive biomarkers that can be obtained non-invasively for monitoring treatment response.

In this project, we demonstrated the successful use of FRET/FLIM to examine HER dimer interaction in colorectal cancer cells and circulating exosomes, and found HER2- HER3 heterodimer can be modulated upon anti-HER treatment, suggesting the importance of HER dimerisation status in early tumour evolution. We optimised techniques for proteomic and genetic analysis of cell-derived and circulating exosomes, and identified several exosomal protein and miRNA biomarkers correlating to HER signalling that could be predictive of treatment failure.

The identification of exosomal protein, miRNA and dimer biomarkers that are predictive of treatment response will allow better stratification of patients to therapy and early switch to alternative treatment prior to later radiological detection of cancer progression. The validation of these biomarkers will enable longitudinal monitoring of patients through therapy in the form of liquid biopsies rather than current practice of invasive tissue biopsies.

The techniques we optimised to analyse HER dimerisation and exosomal contents could be adopted by the research community to address other research questions such as protein interactions within the exosomes, whether certain proteins affect an outcome in a certain group of patients. As exosomes are implicated in many physiological and pathological processes including cancer, cardiac diseases, neurodegenerative disorders, the potential application could be arched through different specialities. Refinement of our methods on exosome works could lead to

4 commercialisation of blood tests that give hints to various disease status or prognosis stratification.

ACKNOWLEDGEMENTS

I would like to thank my supervisors Prof. Daniel Hochhauser and Prof. Tony Ng for their guidance and support during this long journey and for giving me the opportunity to carry out my PhD studies in their laboratory. I wish to show appreciation to Prof. John Hartley for his encouragement along the way and identifying me as part of the group. This study would not been possible without the generous funds from the Comprehensive Cancer Imaging Centre and Cancer Research UK.

I would like to thank all members of the CRUK Drug-DNA interactions research group at UCL Cancer Institute, and the Dimbleby department and Randall Division at KCL for their endless help and support, with special thanks to Dr Paul Barber, Dr Gregory Weitsman, Dr Giovanna Alfano, Dr Juanjo Garcia Gomez, Dr Jose Miguel Vicencio Bustamante, Dr James Monypenny, Dr Valenti Gomez, Dr Sylwia Jones, Dr Arola Fortian Bernabeu, Dr Arman Esfandiari, Dr Oana Coban, Dr Miguel Esteras, and Dr Valeria Santoro, for their scientific support and patience. Thanks to Dr Shih-hsin Chen, Dr Myria Galazi, Avigayil Chalk, Simon Corbett, Jo Clancy, Dr Ruhe Chowdhury, Dr Kenrick Ng for their friendship and advice sharing this PhD journey.

Thanks to Prof Neil Carragher and Mr Kenny McCleod from Edinburgh University for their help with setting up the Reverse Phrase Protein Array.

Finally, thanks to my family, my husband, my parents, and my baby girl, for their unconditional love and everlasting support.

Contents

DECLARATION 2

ABSTRACT 3

IMPACT STATEMENT 4

ACKNOWLEDGEMENTS 6

TABLE OF CONTENTS 7

LIST OF FIGURES 12

LIST OF TABLES 16

LIST OF ABBREVIATIONS 17

Chapter 1...... 22 Introduction ...... 22 1.1 Colorectal cancer overview ...... 23 1.2 Pathogenesis of colorectal cancer and molecular subtypes ...... 24 1.3 Chemotherapy and combination regimens in colorectal cancer ...... 26 1.3.1 5-Fluorouracil ...... 26 1.3.2 Oxaliplatin ...... 27 1.3.3 Irinotecan ...... 28 1.3.4 Combination chemotherapy and sequencing ...... 29 1.3.5 Other cytotoxic agents ...... 29 1.4 Biological targeted agents in colorectal cancer ...... 30 1.4.1 Anti-EGFR therapies ...... 30 1.4.2 Anti-VEGF strategies ...... 31 1.5 The Human Epidermal Growth Factor Receptor Family ...... 31 1.6 The epidermal growth factor receptor (EGFR) ...... 35 1.7 HER receptors as biological targets in colorectal cancer ...... 37 1.7.1 Cetuximab ...... 38 1.7.2 Panitumumab ...... 40 1.7.3 Clinical activities of cetuximab and panitumumab in mCRC ...... 40

1.8 Mechanisms of resistance to current anti-EGFR therapy ...... 41 1.8.1 RAS status in predicting response to anti-EGFR agents ...... 41 1.8.2 BRAF mutation ...... 43 1.8.3 Impact of primary tumour location on prognosis and choice of first line biological agent ...... 43 1.8.4 PIK3CA mutations ...... 44 1.8.5 MET overexpression and amplification ...... 45 1.8.6 EGFR mutations ...... 46 1.8.7 HER2 amplification and signalling ...... 48 1.8.8 Role of HER3 in anti-EGFR resistance ...... 48 1.9 HER dimers as biomarker to monitor tumour response to therapy ...... 50 1.10 Using FRET-FLIM technology to examine HER dimerisation ...... 51 1.11 Overcoming resistance to anti-EGFR therapy ...... 53 1.12 Detection of emerging resistance biomarkers using liquid biopsy ...... 54 1.13 Exosomes and carcinogenesis ...... 55 1.13.1 Exosome biogenesis and cargo sorting is regulated ...... 55 1.13.2 Exosomal protein cargo reflects their origin and is functionally specific ...... 57 1.13.3 Function of tumour-derived exosomes in cancer pathogenesis ...... 58 1.13.4 S100A8/A9 secreted via exosomes from activated MDSCs can promote tumour growth and metastasis ...... 59 1.13.5 Shuttling of exosomal miRNAs can affect target cell gene expression and signalling ...... 60 1.13.6 Exosome as a potential biomarker in colorectal cancer ...... 61 1.14 Evaluation of exosome isolation methods ...... 64 1.15 Aims and objectives...... 65 Chapter 2...... 66 Materials and Methods ...... 66 2.1 Materials ...... 67 2.1.1 Colorectal cancer cell lines and culture conditions ...... 67 2.1.2 Reagents for cell culture ...... 67 2.1.3 Cell stimulation ...... 67 2.1.4 Chemotherapy drugs and inhibitors ...... 67 2.1.5 Antibodies ...... 68 2.1.5.1 Primary antibodies used for immunoblotting and dot blots ...... 68

2.1.5.3 Secondary antibodies ...... 69 2.2 Methods ...... 70 2.2.1 Cell line maintenance ...... 70 2.2.2 Storage and retrieval from liquid nitrogen ...... 70 2.2.3 Assessing growth inhibition and cell viability - MTT assay ...... 70 2.2.4 Genotyping of cell lines for KRAS mutations ...... 71 2.2.5 Cell treatments ...... 72 2.2.6 Immunoblotting analysis ...... 72 2.2.7 Densitometry analysis ...... 74 2.2.8 Exosome extraction and analysis ...... 74 2.2.9 Preparation of cells for imaging ...... 79 2.2.10 Preparation of cell pellets ...... 79 2.2.11 Preparation of FFPE tissue for FLIM imaging ...... 80 2.2.12 Single photon fluorescence lifetime imaging ...... 80 2.2.13 FRET/FLIM Analysis ...... 81 Chapter 3...... 82 HER family receptor activation and dimerisation upon ligand stimulation and drug treatments in colorectal cancer cell lines ...... 82 3.1 Introduction ...... 83 3.2 Aims...... 85 3.3 Results ...... 86 3.3.1 Characteristics of colorectal cancer cell lines ...... 86 3.3.2 Genotyping of colorectal cancer cell lines for KRAS ...... 87 3.3.3 Effect of ligand stimulation on HER activation and signalling in colorectal cancer cell lines ...... 87 3.3.4 Time required for HER activation upon ligand treatment ...... 91 3.3.5 Effect of ligand stimulation on HER dimerisation ...... 92 3.3.6. Effect of anti-HER therapies on HER dimerisation in colorectal cancer cell lines ...... 97 3.4 Discussion ...... 105 3.4.1. Effect of ligand stimulation on HER activation in colorectal cancer cells ...... 105 3.4.2. Quantitative analysis of HER dimers in colorectal cancer cells using FRET/FLIM ...... 106 3.4.3. Effect of Cetuximab on HER activation and dimerisation in colorectal cancer cells ...... 107

3.4.4. Effect of AZD8931 on HER activation and dimerisation in colorectal cancer cells ...... 109 Chapter 4...... 111 Exosomes as biomarker for monitoring HER signaling in colorectal cancer cells ...... 111 4.1 Introduction ...... 112 4.2 Aims...... 116 4.3 Results ...... 117 4.3.1. Optimisation of exosome isolation and quantification from cultured colorectal cell lines ...... 117 4.3.2. Detection of exosomes by immunoblotting ...... 122 4.3.3. Exosomal HER expression in different colorectal cancer cell lines 125 4.3.4. Effect of cetuximab and AZD8931 on exosomal HER excretion in colorectal cancer cell lines ...... 128 4.4 Discussion ...... 130 4.4.1 Isolation and characterisation of colorectal cancer cells secreted exosomes ...... 130 4.4.2. Exosomal HER protein level as marker of HER expression in colorectal cancer cell lines ...... 132 4.4.3. Effect of cetuximab and AZD8931 on exosomal HER protein secretion from colorectal cancer cell lines ...... 133 4.4.4. Challenges with cell-based exosome work ...... 134 Chapter 5...... 136 Role of circulating exosomes in HER signalling in colorectal cancer patients undergoing systemic treatment...... 136 5.1 Introduction ...... 137 5.2 Aims...... 141 5.3 Results ...... 142 5.3.1 Colorectal cancer patient cohort and recruitment ...... 142 5.3.2 Optimisation of exosome isolation from patient blood ...... 144 5.3.3. Exosomal HER protein content in serum-derived exosomes ...... 144 5.3.4. High throughput proteomic analysis of exosomes using reverse phase protein array (RPPA) ...... 147 5.3.5. Analysis of miRNAs in circulating exosomes ...... 153 5.3.6. Exosomal proteins and miRNAs involved in HER signalling as predictive biomarkers in colorectal cancer patients undergoing treatment ...... 155

5.3.7. Effect of treatment with AZD8931 on plasma exosomal proteins in patients ...... 164 5.4 Discussion ...... 170 5.4.1. Exosomes can be reliably isolated and characterised from blood of patients with colorectal cancer ...... 170 5.4.2. RPPA (Reverse Phase Protein Array) as an alternative method for exosomal protein analysis ...... 171 5.4.3. Detection of circulating exosomal HER proteins in colorectal cancer patients ...... 173 5.4.4. The potential role of circulating exosomal proteins as predictive biomarker in colorectal cancer patients undergoing systemic treatment 174 5.4.5. Effect of AZD8931 on circulating exosomal proteins in colorectal cancer patients ...... 175 5.4.6. Exosomal miRNA analysis by ddPCR and normalisation strategy .. 176 5.4.7. Exosomal miRNAs as markers of HER signalling in colorectal cancer ...... 178 Chapter 6...... 180 Conclusions ...... 180 References ...... 186

LIST OF FIGURES

Figure 1.1. Chemical structure of 5-Fluorouracil……………………………………27 Figure 1.2. Chemical structure of oxaliplatin………………………………………...28 Figure 1.3. Chemical structure of irinotecan and SN-38…………………………...28 Figure 1.4. HER receptors and their ligands...... 32 Figure 1.5. HER family dimerisation……………………………………………...……34 Figure 1.6. Signalling heterogeneity in the HER signalling network…………….35 Figure 1.7. Structure and development of cetuximab……………………………...39 Figure 1.8. Expanded RAS analysis…………………………………………………...42 Figure 1.9. Differences in clinical factors and molecular markers of right- versus left-sided colorectal cancer……………………………………………………44 Figure 1.10. Structural analysis of EGFR extracellular domain mutants……….46 Figure 1.11. Model for mechanism of action of Sym004…………………………..47 Figure 1.12. Chemical structure of AZD8931………………………...………………50 Figure 1.13. Jablonski representation of Forster resonance energy transfer…52 Figure 1.14. Exosome biogenesis occurs within MVBs of endosomal system..55 Figure 1.15. Protein composition of a canonical exosome………………………..58 Figure 1.16. Exosomes as circulating biomarker...………..………………………..62 Figure 3.1. Genotyping of colorectal cancer cell lines for KRAS exon 2….……87 Figure 3.2. Effect of ligand stimulation on HER activation among 4 colorectal cancer cell lines……………………………………………………….…………………..88 Figure 3.3. Phosphorylated to Total HER protein ratio in SW48 upon ligand stimulation with EGF or NRG1……………………………………………..…………..89 Figure 3.4. Phosphorylated to Total HER protein ratio in DLD1 (KRAS WT) upon ligand stimulation with EGF or NRG1……………………………………….………...89 Figure 3.5. Phosphorylated to Total HER protein ratio in HCT116 (KRAS WT) upon ligand stimulation with EGF or NRG1……………………………..…….……..90 Figure 3.6. Phosphorylated to Total HER protein ratio in DiFi upon ligand stimulation with EGF or NRG1……………………………………………..…...……...90 Figure 3.7. Effect of ligand stimulation on downstream signalling in colorectal cancer cell lines…………………………………………………………………..…….…91 Figure 3.8. Ligand stimulation time-course in DLD1 (KRAS WT)…………..…….92 Figure 3.9. FRET efficiency between HER1:2 upon EGF stimulation in DLD1 (KRAS WT) fixed cells……………………………..……………………………………..93 Figure 3.10. FRET efficiency between HER2:3 upon NRG1 stimulation in DLD1 (KRAS WT) fixed cells……………………………………………………….…………...94

Figure 3.11. FRET efficiency between HER1:2 upon EGF stimulation in DLD1 (KRAS WT) cell pellet…………………………………………………………………….95 Figure 3.12. FRET efficiency between HER1:3 upon EGF stimulation in DLD1 (KRAS WT) cell pellet…………………………………………………………………….96 Figure 3.13. FRET efficiency between HER2:3 upon EGF stimulation in DLD1 (KRAS WT) cell pellet………………………………………………………………….…97 Figure 3.14. Effect of ligand stimulation on HER activation in DLD1 (WT) and LIM1215………………………………………..…………………………………………...98 Figure 3.15. Effect of cetuximab and AZD8931 on HER activation in DLD1 (WT) and LIM1215……………………………………………………………………………….99 Figure 3.16. FRET/FLIM analysis of HER2:3 dimerisation in LIM1215…………100 Figure 3.17. FRET/FLIM analysis of HER1:3 dimerisation in LIM1215…………101 Figure 3.18. FRET/FLIM analysis of HER1:3 dimerisation in DLD1 (WT)……...102 Figure 3.19. FRET/FLIM analysis of HER2:3 dimerisation in DLD1 (WT)……...103 Figure 3.20. FRET/FLIM analysis of HER1:3 dimerisation in DLD1 G13D……..104 Figure 3.21. FRET/FLIM analysis of HER2:3 dimerisation in DLD1 G13D……..104 Figure 4.1. Optimised exosome isolation protocol from cultured cells……….117 Figure 4.2. FBS-derived exosomes can be largely removed through ultracentrifugation or using commercial exo-free FBS………………….….…....119 Figure 4.3. Analysis of exosomes using Nanosight LM10-HS…………………..120 Figure 4.4. Immunoelectromicroscopy of exosomes……………………………..121 Figure 4.5. Analysis of exosomal protein concentration….………..……..……..122 Figure 4.6. Detection of exosomes by immunoblotting for surface markers using dot blot…………..………………………………………………………………...123 Figure 4.7. Verifying specificities of anti-HER1, anti-HER2 and anti-HER3 antibodies using siRNA knockdown…………………………….…………………..124 Figure 4.8. Verifying specificities of anti-pHER1, anti-pHER2 and anti-pHER3 antibodies for dot blot through whole length western blots……..……………..125 Figure 4.9 Validation of dot-blot antibodies by target protein knockdown…..125 Figure 4.10. HER composition in DLD1 (KRAS WT) cell and cell-derived exosomes under unstimulated condition…………………………………………..126 Figure 4.11. Heterogeneity of HER proteins in CRC cell lines and cell-derived exosomes……………………………………………………………………………...... 127 Figure 4.12. Effect of 24hr treatment with Cetuximab and AZD8931 on exosomal proteins in DLD-1 (KRAS WT) and LIM1215…………………………………….… 129 Figure 5.1. Signaling pathways modulated by miR-21…………………………...140 Figure 5.2. Exosome isolation from patient blood……………….………………..144

Figure 5.3. Heterogeneous production of exosomes among four metastatic colorectal cancer patients……………………………………………………………..145 Figure 5.4. Dot blot analysis of exosomes derived from 2 new colorectal cancer patients: P5 with metastatic and P10 with localised colorectal cancer……….146 Figure 5.5: Dot blot analysis of exosomes derived from 8 newly diagnosed colorectal cancer patients………………………….……………………………….…147 Figure 5.6. Summary of RPPA output for pilot 4 study…………………....……..149 Figure 5.7. Sample microarray layout………………………..……………………...150 Figure 5.8. 1% Tween found to be the best blocking condition for best dot morphology and reduce exosome clumping……………………………………….152 Figure 5.9. Exosomal miRNA analysis workflow and droplet digital PCR…....153 Figure 5.10. Expression levels of miR-21-5p and miR-30c-5p transcripts in circulating exosomes were analysed using droplets digital PCR…………..….154 Figure 5.11. Exosomal miRNA, protein and circulating CEA in CRC05….……156 Figure 5.12. Exosomal miRNA, protein and circulating CEA in CRC07……….157 Figure 5.13. Exosomal miRNA, protein and circulating CEA in CRC08……….158 Figure 5.14. Exosomal miRNA, protein and circulating CEA in CRC09……….158 Figure 5.15. Exosomal miRNA, protein and circulating CEA in CRC10……….159 Figure 5.16. Exosomal miRNA, protein and circulating CEA in CRC11……….160 Figure 5.17. Exosomal miRNA, protein and circulating CEA in CRC12……….160 Figure 5.18. Spearman’s correlation of changes in exosomal protein signals from pre-treatment to 3 months………………………….…………….…………….161 Figure 5.19. Boxplot showing changes in exosomal HER3 at 3 months for each category of patients according to radiological response at 6 months..…..….162 Figure 5.20. Boxplot showing changes in miR21 at 3 months for each category of patients according to radiological response at 3 months..……………...…..163 Figure 5.21. Scatter plot showing non-significant correlation between change in HER3 and miR30c at 3 months…………………………..………………………..163 Figure 5.22. Effect of AZD8931 with FOLFIRI on exosomal proteins in a cohort of metastatic colorectal cancer patients……………….………..…………………165 Figure 5.23. Changes in exosomal HER2:HER3 FRET efficiency in colorectal cancer patients 5-7 days after first dose of AZD8931…..……………………….166 Figure 5.24. Changes in exosomal HER1:HER3 FRET efficiency in colorectal cancer patients 5-7 days after first dose of AZD8931……………………………167 Figure 5.25. Heatmap of Spearman’s correlation and pair plots between circulating exosomal protein changes after first dose of AZD8931 in patients with advanced colorectal cancer……………………………………………………168

Figure 5.26. Heatmap of Spearman’s correlation and pair plots between circulating exosomal proteins at baseline in patients with advanced colorectal cancer..………………………………………………………………………………..…169 Figure 5.27. Reverse Phase Protein Array Workflow……………………………172

LIST OF TABLES

Table 1.1. Systemic agents used in management of colorectal cancer……….23 Table 1.2. Colorectal cancer consensus gene expression-based subtypes….26 Table 1.3. Summary of FDA approved EGFR targeting agents………………….38 Table 2.1. Primary antibodies used for immunoblotting and dot blots………...68 Table 2.2. Directly labelled antibodies employed in FRET/FLIM imaging….…..69 Table 3.1. Summary of methods for assessing protein interactions………..….84 Table 3.2 Characteristics of colorectal cancer cell lines………………….………86 Table 4.1 Comparing the advantages and disadvantages of CTCs, ctDNA and exosomes as circulating biomarkers………………………………………………..113 Table 4.2 Comparison of exosome isolation techniques………………………...114 Table 5.1. miRNAs that sustain or repress EGFR signalling…………..………..139 Table 5.2. Colorectal cancer patient recruited from oncology clinic…………..142 Table 5.3. Colorectal cancer patient recruited from nuclear medicine clinic...143 Table 5.4. Patient sample cohort for RPPA pilot study 4…………………….…..148 Table 5.5. Antibody performance for pilot study 4………………………………..151 Table 5.6. Cohort of treatment-naive advanced colorectal cancer patients recruited for exosome biomarker study…………………………………………….155 Table 5.7. Characteristics and clinical outcomes of a cohort of patients with advanced colorectal cancer treated with first line AZD8931 and FOLFIRI…....164

LIST OF ABBREVIATIONS

ADAM A Disintegrin AND Metallo-Protease ADCC Antibody-Dependent Cell-Mediated Cytotoxicity AGO2 Argonaute 2 ALIX ALG-2 interacting protein X AKT Protein Kinase B (PKB) AP-1 Activator protein 1 Apaf-1 Apoptotic protease activating factor 1 APC Adenomatous Polyposis Coli AREG Amphiregulin BCA Bicinchoninic acid BCL B-cell lymphoma BRAF V-raf murine sarcoma viral oncogene homolog B1 BSA Bovine Serum Albumin BTC Betacelluline CAV1 Caveolin-1 CDC Complement dependent cytotoxicity CEA Carcino-embryonic antigen CHMP4 Charged multivesicular body protein 4 CIMP CpG Island methylator phenotype CIN Chromosomal instability c-MET Tyrosine-protein kinase MET CMS1-4 Consensus molecular subtypes 1-4 CRC Colorectal cancer CSF1R Colony stimulating factor 1 receptor CT Computer Tomography ctDNA Circulating tumour DNA CTLA-4 Cytotoxic T-lymphocyte-associated protein 4 Cy5 Cyanine 5 ddPCR Digital Droplet Polymerase Chain Reaction dNTP Deoxynucleoside triphosphate DMSO Dimethysulphoxide dTTP Deoxythymidine triphosphate ECD Extracellular domain EDTA Ethylenediaminetetraacetic acid EGF Epidermal growth factor

EFNB2 Ephrin B2 EGN Epigen EGFR Epidermal growth factor receptor ELISA Enzyme-linked immunosorbent assay EpCam Epithelial cell adhesion molecule ERK Extracellular receptor kinase EPR Epiregulin EREG Epiregulin ERK Extracellular Signal-Regulated Kinases ESCRT Endosomal sorting complex required for transport machinery EV Extracellular vesicle FAP Familial Adenomatous Polyposis FBS Fetal Bovine Serum FFPE Formalin fixed paraffin-embedded FGFR Fibroblast growth factor receptor FLIM Fluorescence Lifetime Microscopy FLOT1/2 Flotillin 1/2 FOLFOX Folinic acid, Fluorouracil, Oxaliplatin FOLFIRI Folinic acid, Fluorouracil, Irinotecan FOLFIRINOX Folinic acid, Fluorouracil, Oxaliplatin, Irinotecan FoxO3a forkhead box O3 FRET Forster Resonance Energy Transfer 5-FU Fluorouracil GAPDH Glyceraldehyde 3-phosphate dehydrogenase GTP Guanosine TriPhosphate HB-EGF Heparin-Binding EGF HER1 Human epidermal growth factor receptor 1 HER2 Human epidermal growth factor receptor 2 HER3 Human epidermal growth factor receptor 3 HER4 Human epidermal growth factor receptor 4 HGF Hepatocyte growth factor HMBOX1 Homeobox Telomere-Binding Protein 1 HNPCC Hereditary Non-Polyposis Colon Cancer HNSCC Head and Neck Squamous Cell Carcinoma HRP Horseradish Peroxidase HSC70 Heat shock cognate 70 HSP90 Heat shock protein 90

IHC Immunohistochemistry ILV Intraluminal vesicles IRS Insulin receptor substrate JAG1 Jagged1 JAK Janus kinase KRAS Kirsten rat sarcoma viral oncogene homolog LAMP Lysosomal-associated membrane protein MARK Mitogen-activated protein kinase mAbs monoclonal antibodies MBV Multivesicular body mCRC metastatic colorectal cancer MCH Major Histocompatibility Complex MDSC myeloid-derived suppressor cells MGMT O-6-Methyguanine-DNA Methyltransferase MHC Major histocompatibility complex MLH MutL Homolog 1 MMR Mismatch repair mRNA messenger RNA MSH MutS Homolog MSI Microsatellite instability MTA1 Metastasis Associated 1 mTOR mammalian target of rapamycin MTT Methylthiazolydiphenyl-tetrazolium MVB multivesicular bodies NaOH Sodium Hydroxide NF-kB Nuclear Factor kB NK Natural Killer NRAS Neuroblastoma rat sarcoma viral oncogene homolog NRG Neuregulin (heregulin) NSCLC Non-small cell lung carcinoma NTA Nanosight tracking analysis ORR Overall response rate OS Overall Survival PBS Phosphate buffered saline PCR Polymerase Chain Reaction PD-1 Programmed cell death protein 1 PDCD4 Programmed cell death 4

PDGFR Platelet-derived growth factor receptor PET Positron Emission Tomography PFA Paraformaldehyde PFS Progression-free survival pHER Phosphorylated human epidermal receptor PI3K Phosphoinositide 3-kinase PIK3CA Phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha PKC Protein kinase C PROM1 Prominin 1 PTEN Phosphate and tensin homologue phosphatase qRT-PCR Quantitative reverse transcription PCR RAGE Receptor for advanced glycation endproducts RAP1 Ras-related protein 1 Rb Retinoblastoma RHOB Ras homolog gene family member B RISC RNA-induced silencing complex RPPA Reverse Phase Protein Array RR Response rate RTK Receptor Tyrosine Kinases S100A8/A9 Calprotectin shRNA Short hairpin ribonucleic acid siRNA Small interfering ribonucleic acid SCNA Somatic copy number alteractions SDS Sodium Dodecyl Sulfate SN-38 7-ethyl-10-hydroxycamptothecin (Irinotecan active metabolite) SNARE Soluble NSF attachment protein receptor Src V-Src Avian Sarcoma Viral Oncogene Homolog SST Serum-separating tube STAT Signal Transducer and Activator of Transcription TAE Tris-acetate-EDTA TBST Tris-buffered saline with Tween TGF Transforming growth factor TSG101 Tumour susceptibility gene 101 TK Tyrosine kinase TKIs Tyrosine Kinase Inhibitors TLR4 Toll-like receptor 4 TNIK TRAF2 and NCK interacting kinase

TNFa Tumour necrosis factor alpha TOP I Topoisomerase I TP53 Tumour protein 53 TSG101 tumour susceptibility gene 101 UBC9 Ubiquitin carrier protein 9 VEGF Vascular Endothelial Growth Factor WT Wildtype

Chapter 1

Introduction

1.1 Colorectal cancer overview Colorectal cancer is the third most common cancer in men and second in women worldwide (Ferlay et al., 2015). The incidence of colorectal cancer is more common in the developed countries including the United Kingdom. Worldwide, 1.4 million patients are diagnosed every year of whom 694,000 die from this disease (Ferlay et al., 2015). The incidence rate of colorectal cancer has been increasing since early 1990s. In 2015, 41,804 new cases of colorectal cancer were diagnosed in the UK. The 5-year survival for early stage colorectal cancer is approximately 90%, whereas for locally advanced or metastatic cases are 71% and 13%, respectively (Siegel et al, 2012). The majority of colorectal cancer arise from sequential somatic mutations, approximately 5% can be attributed to hereditable genetic alterations (Burn et al., 2013).

As with any other form of malignancy, the treatment of colorectal cancer is dependent on the stage of the disease. Surgical resection of the tumour and surrounding lymph nodes is the main therapeutic option for stage I to III patients, with or without adjuvant chemotherapy post operatively to reduce the risk of recurrence. For stage IV patients, surgery may be curative in highly selected patients with resectable bowel disease and resectable isolated hepatic or pulmonary metastases. Unresectable liver lesions may be managed locally with transarterial chemo-embolisation or radiofrequency ablation if symptomatic. For majority of patients with stage IV disease, surgery is often palliative to prevent bowel obstruction. Combination systemic therapy including chemotherapy and biologically targeted therapies such as EGFR inhibitors are the mainstay treatment modality for patients with metastatic disease. Table 1.1 shows the overview of systemic agents currently used in the treatment of advanced colorectal cancer.

Table 1.1: Systemic agents used in management of colorectal cancer. Chemotherapy EGFR targeted agents VEGF targeted agents

Fluoropyrimidines Cetuximab Bevacizumab

Oxaliplatin Panitumumab Afibercept

Irinotecan Regorafenib

TAS-102 Ramucirumab (Trifluridine/Tipiracil)

Unfortunately response to anti-EGFR therapy in patients with metastatic colorectal cancer are variable and of limited duration, due to intrinsic or acquired resistance mechanisms described in Chapter 1.8. One mechanism of resistance is activation of alternative signalling pathways through formation of alternative dimers within the HER family. Currently resistance to anti-HER therapy cannot be easily measured resulting in late detection of measurable radiological disease progression. The ability to monitor HER signalling and resistance longitudinally and non-invasively in clinical setting in the form of liquid biopsy will allow better monitoring of treatment response and stratification of patients to personalised anti-cancer therapy.

1.2 Pathogenesis of colorectal cancer and molecular subtypes Colorectal cancer develops through accumulation of genetic and epigenetic alterations leading to the transformation of normal colon mucosa to adenoma, then to carcinoma (Vogelstein et al., 2013). This is a heterogenous process in terms of the order and grade of genetic alterations (Manne et al., 2012). Three molecular pathways have been described either acting independently or in combination in colorectal cancer to result in genomic instability, including chromosomal instability (CIN), microsatellite instability (MSI) and CpG island methylator phenotype (CIMP).

Chromosomal instability (CIN) is the most frequent type of genomic instability, found in 75- 85% of colorectal cancers, and can be caused by inactivation of several tumour suppressor genes including APC or TP53 (Markowitz & Bertagnolli, 2009). Aneuploidy and loss of heterozygosity are the major players in CIN tumours, which not only constitute most of the sporadic tumours, but also involve Familial Adenomatous Polyposis (FAP) cases associated with germline mutations in the APC gene (Smith et al., 2002; Fearon, 2011)

The microsatellite instability (MIS) pathway involves the inactivation of genes in the DNA mismatch repair (MMR) family either by aberrant DNA methylation or somatic mutation (Boland & Goel, 2010), and accounts up for 15% of colorectal cancers. The inactivation of genes in the MMR machinery can be inherited or acquired. Germline mutations in DNA mismatch repair (MMR) genes including MLH1, MSH2, MSH6 and PMS2 can lead to hereditary non-polyposis colorectal cancer (HNPCC) or Lynch syndrome (Brenner et al, 2014). The sporadic cases are often resulted from methylation of MLH1 gene, associated with serrated neoplasia pathway and BRAF V600E mutation (Wang et al., 2003; Fearon, 2011).

The CpG island methylator phenotype (CIMP) pathway is characterised by promoter hypermethylation of various tumour suppressor genes, most importantly MGMT and MLH1, preventing them from undergoing transcription and therefore effectively inactivating them (Toyota et al., 1999). This hypermethylation is often associated with BRAF mutation and microsatellite instability (Inoue et al., 2015; Weisenberger et al., 2006).

Recent advances in genomic testing allowed better classification of colorectal cancer by the Colorectal Cancer Subtyping Consortium. Researchers used six previously published classification systems to data from more than 4000 predominately stage II or III colorectal cancer samples, which analysed 30 different gene expression sets across multiple platforms and sample preparation methods (Sadanandam et al., 2013; Guinney et al., 2015). Four subtypes were identified (CMS1-4) covering approximately 87% of all CRC cases, leaving 13% molecularly uncharacterised (Table 1.). CMS1 (microsatellite instability immune, 14%) tumours are characterised as hypermutated, microsatellite unstable, and with strong immune activation. CMS1 tumours had lower chromosomal instability and DNA methylation, but more frequent BRAF mutation compared to other subtypes. These tumours are often right-sided, once relapsed are associated with worse survival, and do not respond well to anti-EGFR therapies. CMS (Canonical, 37%) tumours are characterised as epithelial, chromosomally unstable, and with marked WNT and MYC signalling activation. They are often left-sided, associated with a higher survival rate after relapse and better response to anti-EGFR therapies. CMS3 (metabolic, 13%) subtypes are epithelial tumours with evident metabolic dysregulation. They have intermediate level of somatic copy number alterations and DNA methylation. KRAS is most commonly seen in this subtype. CMS4 (mesenchymal, 23%) tumours have prominent TGFb activation, stromal invasion and angiogenesis. They are characterised by low methylation, chromosomal instability, and very high levels of somatic copy number alterations. There are frequent mutations of APC, KRAS, PIK3CA and TP53. Patients are often diagnosed at advanced stages, are anti-EGFR resistant, and has the worst survival out of the four subtypes.

Table 1.2: Colorectal cancer consensus gene expression-based subtypes (Adapted from (Guinney et al., 2015). SCNA: somatic copy number alterations. CIMP: CpG island methylator phenotype.

CMS1 CMS2 CMS3 CMS4

MSI immune Canonical Metabolic Mesenchymal

14% 37% 13% 23%

MSI, CIMP high, SCNA high Mixed MSI status, SCNA SCNA high hypermutation low, CIMP low

BRAF mutations KRAS mutations

Immune infiltration WNT and MYC Metabolic deregulation Stromal infiltration, and activation activation TGFb activation, angiogenesis

Worse survival after Worse relapse-free and relapse overall survival

1.3 Chemotherapy and combination regimens in colorectal cancer Chemotherapy is the main treatment for patients with advanced colorectal cancer (DeSantis et al., 2014). Fluoropyrimidine (5-fluorouracil or capecitabine), oxaliplatin, and irinotecan and are the main chemotherapeutic agents used for the treatment of colorectal cancer patients.

1.3.1 5-Fluorouracil 5-Fluorouracil (5-FU) is an analogue of uracil with a fluorine atom in place of the hydrogen at C-5 position (Longley et al., 2003). It is an antimetabolite with the action of misincorporation of fluoronucleotides in DNA or RNA during their synthesis and inhibition of the thymidylate synthase enzyme (Longley et al., 2003). Intracellular 5-FU is converted into three active metabolites: fluorodeoxyuridine monophosphate (FdUMP), fluorodeoxyuridine triphosphate (FdUTP) and fluorouridine triphosphate (FUTP). FdUMP and FdUTP mainly cause DNA disruption and TS inhibition, and FUTP incorporates into cellular RNA resulting in disruption in RNA processing. 5-FU is

26 typically administered with the biomodulator leucovorin (LV), a reduced form of folate that stabilises the binding between 5-FU and thymidylate synthase (de Gramont et al., 1997; Petrelli et al., 1989). Capecitabine is an oral pro-drug of 5-FU that is absorbed intact through the gastrointestinal mucosa and undergoes a three-step enzymatic conversion to fluorouracil (Pentheroudakis & Twelves, 2002).

Figure 1.1: Chemical structure of 5-Fluorouracil

Deficiency of dihydropyrimidine dehydrogenase (DPD), the exclusive enzymatic step for catabolism and deactivation of fluoropyrimidines, can lead to severe toxicities with 5-FU. 5-FU-mediated cytotoxicity may be dependent on intact DNA MMR function, as 5-FU based chemotherapy does not prolong survival in patients with MMR-deficient colorectal cancer (Meyers et al., 2005; Tajima et al., 2004).

To date, fluorouracil continues to form the backbone of almost all chemotherapeutic regimens for patients with metastatic colorectal cancer and has extended median survival from approximately 6 to 12 months (Thirion et al., 2004).

1.3.2 Oxaliplatin Oxaliplatin is a platinum derivative which acts by covalently bind DNA to induce DNA damage. The platinum compounds become activated in the cell through a chemical reaction where a leaving group present in the platinum agent gets displaced by water molecules (aquation), resulting in an activated and positively charged compound which interact with DNA, RNA and proteins in the cell to form DNA adducts, intra- and inter- strand crosslinks and DNA-protein crosslinks (Rabik & Dolan, 2007). Oxaliplatin has a 1,2-diaminocycloexhane (DACH) carrier which forms bulky DNA adducts and lead to strand distortions which impair recognition from MMR proteins (Raymond et al., 1998).

Figure 1.2: Chemical structure of oxaliplatin

In patients with metastatic colorectal cancer, single-agent oxaliplatin has limited efficacy, but is more efficacious clinically when combined with 5-FU-leucovorin, possibly due to oxaliplatin-induced downregulation of thymidylate synthetase (Raymond et al., 2002; de Gramont et al., 2000).

1.3.3 Irinotecan Irinotecan is a semi-synthetic derivative of the natural alkaloid camptothecin that is activated by plasma or cellular carboxylesterases to active metabolite SN-38. SN-38 inhibits topoisomerase I, an enzyme that catalyses breakage and rejoining of DNA strands during DNA replication, thereby causes DNA fragmentation and programmed cell death (Gilbert et al., 2012). SN-38 traps TOP1-DNA cleavage complexes by docking between the DNA at the cleavage site and the TOP1 enzyme (Tomicic & Kaina, 2013). Collision with advancing replication forks results in the generation of double strand breaks (Pommier, 2006).

Irinotecan SN38

Figure 1.3: Chemical structure of irinotecan and SN-38

SN-38 is metabolised predominately in the liver where it is inactivated by glucuronidation and excreted through the biliary system. A polymorphism in the uridine

28 diphosphate glucuronosyltransferase isoform 1A1 (UGTA1A) gene leads to decreased activation of SN-38 and increased treatment-related toxicity (Ando et al., 2005).

Irinotecan increased tumour shrinkage and improved survival by at least 2 months when used in combination with 5-FU-leucovorin in patients with advanced colorectal cancer, compared with 5-FU-leucovorin alone (Douillard et al., 2000; Saltz et al., 2000).

1.3.4 Combination chemotherapy and sequencing The typical first-line chemotherapy for the treatment of advanced colorectal cancer comprises of fluoropyrimidine (5-fluorouracil or capecitabine) combined with irinotecan or oxaliplatin, which gives higher response rates and better progression-free and overall survival data compared with fluoropyrimidine alone (Grothey et al., 2004; de Gramont et al., 2000; Douillard et al., 2000). FOLFOX and FOLFIRI have been shown to have comparable survival outcomes in the GERCOR and GOIM trials, but different interactions with biologicals and toxicity profiles (Colucci et al., 2005; Tournigand et al., 2004). The triplet combination chemotherapy regimen FOLFOXIRI has been demonstrated to be superior to FOLFIRI in an Italian study (Cremolini et al., 2015).

Second-line chemotherapy are offered to patients with good performance status and organ function. A retrospective analysis indicates exposure to all three cytotoxics in various sequences results in the longest survival (Grothey et al., 2004). Therefore, patients who progressed on irinotecan-based regimen should be offered second-line oxaliplatin-containing combination, and vice versa.

1.3.5 Other cytotoxic agents TAS102 (Trifluridine/Tipiracil) is an oral combination of thymidine-based nucleic acid analogue, trifluridine, and a thymidine phosphorylase inhibitor, tipiracil hydrochloride. Trifluridine is the active cytotoxic component, where its triphosphate form is incorporated into DNA during DNA synthesis inhibiting tumour cell growth (Matsushita et al., 1999; Emura et al., 2004; SUZUKI et al., 2011). Trifluridine is incorporated into DNA by phosphorylation by thymidylate kinase to TF-TMP, which then covalently binds to tyrosine 146 of the active site of thymidylate synthase inhibiting the enzyme’s activity. This depletes the cell of dTTP and causes accumulation of deoxyuridine monophosphate which increases the likelihood that uracil gets misincorporated into the DNA. Also, subsequent phosphorylations of TF-TMP cause increased level of TF-

TTP within the cell, resulting in incorporation into DNA that leads to DNA strand breaks. Tipiracil hydrochloride is a potent inhibitor of thymidine phosphorylase and prevents rapid degradation of trifluridine (Fukushima et al., 2000).

TAS-102 is a newer cytotoxic agent and its clinical activity in patients with treatment- refractory colorectal cancer was confirmed in the phase III RECOURSE trial, where a small but significant survival benefit was observed for TAS102-treated patients compared to placebo who had received at least two prior lines of chemotherapy and prior biological therapy (Mayer et al., 2015; Van Cutsem et al., 2018).

1.4 Biological targeted agents in colorectal cancer Anti-epidermal growth factor receptor monoclonal antibodies cetuximab and panitumumab, and anti-vascular endothelial growth factor bevacizumab have been shown to improve clinical outcome of patients with metastatic colorectal cancer when combined with chemotherapy doublet in the first-line setting (Maughan et al., 2011; Tveit et al., 2012; Heinemann et al., 2014; Schwartzberg et al., 2014; Tol et al., 2009a).

Recent evidence from randomised phase III trials and meta-analyses suggested metastatic colorectal cancer patients with RAS WT tumours may experience improved survival outcomes when chemotherapy is combined with cetuximab compared to bevacizumab in the first-line setting (Arnold et al., 2017; Holch et al., 2017; Brulé et al., 2015). Patients with right-sided tumours appear to derive less significant benefits from available therapies than do those tumours originating from left side of colon, in terms of overall response rate (ORR), progression-free survival (PFS) and overall survival (OS).

1.4.1 Anti-EGFR therapies The two main anti-EGFR therapeutic agents demonstrating effectiveness and survival benefits in the treatment of metastatic colorectal cancer are monoclonal antibodies Cetuximab and Panitumumab, either used alone or in combination with cytotoxics, as further described in Section 1.7.

1.4.2 Anti-VEGF strategies Bevacizumab is a humanised anti-VEGF-A monoclonal antibody (Ferrara, 2004). It increases the activity of the cytotoxic regimen used in colorectal cancer. It has been shown to increase survival, PFS and RR in first line treatment in combination with 5- FU/LV/irinotecan, and in combination with 5-FU/LV or capecitabine alone (Cunningham et al., 2013; Kabbinavar et al., 2008; Hurwitz et al., 2004). It has also been shown to improve PFS in combination oxaliplatin-based chemotherapy in first line treatment of metastatic colorectal cancer (Saltz et al., 2008). The combination of FOLFOXIRI plus bevacizumab has shown superior PFS and RR than FOLRIFI plus bevacizumab (Cremolini et al., 2015).

Aflibercept is a recombinant fusion protein that blocks the activity of VEGF-A, VEGF- B and placenta growth factor, improves survival, PFS and RR when combined with FOLFIRI in second line in oxaliplatin pre-treated patients, whether or not the patients were pre-treated with bevacizumab (Van Cutsem et al., 2012).

Regorafenib is an oral multikinase inhibitor, including kinases involved in tumour angiogenesis (VEGFR-1, -2, -3), oncogenesis (KIT, RET, RAF-1, BRAF), metastasis (VEGFR3, PDGFR, FGFR), and tumour immunity (CSF1R) (Wilhelm et al., 2011). It has shown efficacy in patients pre-treated with all other options in the phase III CORRECT trial, where it statistically prolonged OS compared with placebo (Grothey et al., 2013). The survival benefit of regorafenib in treatment refractory metastatic colorectal cancer patients was confirmed in the Asian study CONCUR (Li et al., 2014).

Ramucirumab is a fully humanised IgG-1 monoclonal antibody that binds to extracellular VEGF-binding domain of VEGFR-2 and blocks VEGF ligands from binding this site and activating receptor. It has been approved for use in second-line setting in combination with FOLFIRI for patients progressed on first-line bevacizumab-, oxaliplatin- and fluoropyrimidine-containing regimens based on results from the RAISE trial (Tabernero et al., 2015).

1.5 The Human Epidermal Growth Factor Receptor Family Human receptor tyrosine kinases are a class of cell surface growth factor receptors with intrinsic tyrosine kinase activity regulating essential cellular functions such as survival, growth, differentiation, migration and cell death (Gschwind et al., 2004). So far 20 subfamilies of 58 human RTKs have been identified (Lemmon & Schlessinger,

2010). The human epidermal receptor (HER) family consists of four members: EGFR (HER-1, ErbB-1), HER-2 (ErbB-2), HER-3 (ErbB-3), and HER-4 (ErbB-4).

Figure 1.4: HER receptors and their ligands.

The HER receptors are transmembrane glycoproteins with shared structure homology of an extracellular ligand-binding region, a single pass alpha-helical transmembrane domain and an intracellular domain. The extracellular region consists of four domains arranged as tandem repeat of a two-domain unit consisting of approximately 190 amino acid leucine-rich domain, followed by approximately 120 amino acid cysteine- rich domain (Cho & Leahy, 2002; Lemmon, 2009). The intracellular domain contains 3 domains: 50 amino acid juxtamembrane region, 250 amino acid tyrosine kinase domain, and 229 amino acid carboxyl terminal tail.

With the exception of HER2, all other HER family members are competent for growth factor-ligand stimulation. Epidermal growth factor (EGF), transforming growth factor alpha (TGFa), amphiregulin (ARG) and epigen (EGN) are specific ligands for EGFR. Betacelluline (BTC), heparin binding EGF (HB-EGF) and epiregulin (EPR) can bind and activate both EGFR and HER4. Neuregulin sub-family NRG 1 and 2 bind both HER3 and HER4, and NRG3 and 4 are HER4 specific (Guo et al., 2003; Singh & Harris, 2005; Singh et al., 2016a).

Ligand binding leads to receptor dimerisation, which is essential to HER family activation. This can be homo- or hetero-dimeric. EGFR and HER4 undergo homodimeristaion and are capable of autophosphorylation. HER1:HER4 heterodimers are recognised and may influence cellular behaviour such as migration but appear to form less frequently.

HER2 lacks an identifiable ligand and its activation occurs by heterodimerisation with another member of ligand-activated HER family receptor, or ligand-independent homodimerisation in HER2 overexpressing cells (Yarden & Sliwkowski, 2001). In contrast to the other HER receptors where in the inactivated state the dimerisation arm of the receptor remain hidden, the dimerisation arm of the HER2 is permanently partially exposed, permitting its dimerisation in the absence of ligand-binding activity (Jorissen et al., 2003). Heterodimers containing HER2 have particularly high ligand binding and signalling potency compared with those without as the heterodimers have relatively slow rate of ligand dissociation and HER2 is a slowly internalising receptor (Baulida et al., 1996; Zaczek et al., 2005).

HER3 has a weakly active tyrosine kinase domain and HER3 homodimers are completely inactive therefore must heterodimerise to become activated. In general heterodimeric pairings lead to more tumorigenic outcomes than their homodimeric counterparts and is thought to be particularly true for those that include HER2 or HER3 partners. The HER2:HER3 dimer is the most oncogenic HER pair, through activation of the PI3K/Akt pathway (Way & Lin, 2005).

An asymmetric activator-receiver model where each of a pair of HER receptors is dedicated to either activating or being activated in the process of phosphorylation and signal transduction, appears to be the basis of activation for the HER family in general. In their inactive, unbound state, the default HER family structure is to adopt a closed, tethered conformation. Ligand binding to extracellular domain promotes a conformational change to an extended structure that reveals the dimerisation arm. This engages another ligand-bound receptor molecule, overcoming the resting auto- inhibited state with subsequent activation of the kinase domain and propagation of downstream signalling.

Figure 1.5: HER family dimerisation. a. The HER family consists of 4 receptors. HER1 (ErbB1/EGFR), HER3 (ErbB3) and HER4 (ErbB4) have been crystallised as a closed, ‘tethered’ structure in the inactive state, by crosslinkage between different domains, as illustrated. HER2 (ErbB2) has no known ligand and may exist in an open or extended state. b. Ligand binding induces a conformational change, in which the receptor adopts the open structure. This exposes domain II for dimerisation with another HER receptor, as well as bringing the intracellular tyrosine kinase domain into close proximity, for transphosphorylation and activation. The kinase domain binding is asymmetric with the amino-terminal (N-terminal) lobe of one TK interacting with the carboxyl-terminal (c- terminal) lobe of the other. Reproduced from (Baselga & Swain, 2009)

Signal processing by the HER family receptors is mediated further by activation of downstream intracellular signalling pathways including Mitogen Activated Protein Kinase (MAPK), PI3K/Akt, lipid kinases, the Src kinase and STAT transcription factors pathway that control cell growth, proliferation, survival, angiogenesis and differentiation (Yarden & Sliwkowski, 2001; Wheeler et al., 2010).

Figure 1.6: Signalling heterogeneity in the HER signalling network.

A range of ligands activate the HER receptors (green boxes). There are 10 potential dimerisation partnerships between the 4 HER family members which results in various possibilities for multi-functionality and signalling specificity (green circles). Each HER dimer pair in the cell surface membrane can give rise to multiple unique downstream signals with MARK and PI3K-AKT being of particular importance (red circles/blue squares). A range of phenotypical outcomes is determined through the control of these signalling mediators by the respective HER receptor pairing (yellow squares). Reproduced from (Yarden & Sliwkowski, 2001).

1.6 The epidermal growth factor receptor (EGFR) EGFR is the most important HER family receptor in colorectal cancer and has been established as one of the most clinically relevant oncogenes in this tumour type. It is a glycoprotein of 170kDa, encoded by the EGFR gene located on the short arm of chromosome 7p12 consisting of approximately 200,000 base pairs containing 30 exons (Reiter et al., 2001). It is present on all epithelial and stromal cells, and is expressed on many glial and smooth muscle cells as well.

Like the other HER molecules, structurally EGFR consists of three regions: an extracellular domain, a transmembrane domain and an intracellular domain (Ogiso et al., 2002). The extracellular region of EGFR is further divided into four domains:

Domains I, II, III and IV, with DI and DIII contributing to ligand binding, and cysteine rich domains II and IV required for receptor dimerisation (Ferguson, 2008). The transmembrane domain of the receptor contributes to formation of dimer-stabilising component of the activated state of EGFR (Hubbard, 2009). The intracellular domain is the enzymatic core of the receptor where tyrosine residues are found which can become phosphorylated by kinase activity, and act as docking sites to facilitate binding of signal mediators (Zhang et al., 2006).

Activating ligands of EGFR include EGF, transforming growth factor α (TGF-α), amphiregulin (AREG), betacellulin (BTC), epigen (EGN), epiregulin (EPR), heparin- binding EGF-like growth factor (HB-EGF) (Guo et al., 2003; Singh et al., 2016b). Each contains an EGF-like domain that is responsible for receptor binding and activation, with a characteristic pattern of six spatially conserved cysteines. They also have a transmembrane domain that links the extracellular amino-terminus domain with a cytoplasmic carboxy-terminus tail. EGFR ligands are released from membrane bound proligand forms by membrane bound metalloprotease enzymes of the ADAM family. ADAM 17 is a key enzyme regulating release of EGFR ligands including EGF, amphiregulin, and heparin-binding EGF (Sahin et al., 2004; Adrain & Freeman, 2014).

Binding of an activating ligand results in dimerisation of EGFR, either with itself (homodimerization) or with another member of the family (heterodimeristaion). Monomeric EGFR normally exists in a tethered closed and auto-inhibited state in which the dimerisation arm (domain IV) is not exposed. Ligand binding to extracellular domains I and III stabilises the receptor in the open conformation required for dimerisation and activation (Dawson et al., 2005). EGFR dimerisation occurs via asymmetric mechanism of interaction of receptor monomers which takes place when the C-lobe of one monomer (activator or donor) and the N-lobe of another monomer (receiver or acceptor) are in contact (Zhang et al., 2006). The activated kinase can then phosphorylate tyrosine residues on the receiver monomer which can switch position and activate each other (Jura et al., 2009b). The phosphorylation of intracellular tyrosine residues within the carboxyl tails lead to recruitment of Src- Homology (SH2) and phosphor-tyrosine binding (PTB) domains leading to activation of downstream signalling pathways (Lemmon et al., 2014; Lemmon & Schlessinger, 2010). The two main primary signalling pathways activated by EGFR are Ras-Raf- MAPK pathway which lead to cell cycle progression and proliferation, and PTEN/PI3K/AKT pathway which leads to protein synthesis, cell growth, survival and mobility (Yarden & Sliwkowski, 2001; Oda et al., 2005; Lemmon et al., 2014).

Each ligand appears to activate the EGFR in the same way and induce EGFR internalisation and trafficking to early endosomes (Roepstorff et al., 2009). However the activation of EGFR by different ligands has been shown to cause distinct downstream biological activities, due to differential ligand induced dimer pairing and endocytic sorting or recycling (Singh & Harris, 2005; Freed et al., 2017; Wilson et al., 2012).

EGFR was the first tyrosine-kinase receptor to be linked to human cancers including colorectal cancer (Hynes & Lane, 2005). EGFR malignant activation is mediated by both ligand dependent and independent mechanisms. The former comprises of autocrine production of ligands in EGFR overexpressing tumours including colorectal, lung, and head and neck (Gschwind et al., 2001). The latter includes receptor overexpression via gene amplification, acquisition of activating mutations (e.g. EGFRvIII), and escape from endocytic degradation (Grandal et al., 2007; Greenall & Johns, 2016). The aberrant activation of EGFR in cancer leads to a range of cellular responses including cell proliferation, apoptosis, cell motility, modulation of epithelial- mesenchymal transition (EMT), upregulation of matrix metalloproteinases and vascular endothelial growth factor-A (Sigismund et al., 2018; Wee & Wang, 2017).

1.7 HER receptors as biological targets in colorectal cancer EGFR is the best studied receptor among the HER family and has become an important therapeutic target. The currently available EGFR-targeted agents are classified in two categories: monoclonal antibodies (mAbs) which targeting the extracellular ligand binding domain of the receptor, and small tyrosine kinase inhibitors (TKIs) which targeting the intracellular ATP binding domain in the cytoplasmic tail. Gefitinib and erlotinib are ATP-competitive and reversible inhibitors of EGFR and effective in the treatment of NSCLC patients with EGFR mutation, but has not demonstrated clinical efficacy in colorectal cancer. AZD8931 is a pan-HER small molecule tyrosine kinase inhibitor shown pre-clinical efficacy in colorectal cancer and is further described below.

Table 1.3: Summary of FDA approved EGFR targeting agents.

EGFR Mode of action FDA approved Inhibitor

Cetuximab EGFR monoclonal Ab RAS WT colorectal cancer

Metastatic head and neck cancer

Panitumumab EGFR monoclonal Ab RAS WT colorectal cancer

Gefitinib EGFR tyrosine kinase Non-small cell lung cancer inhibitor

Erlotinib EGFR tyrosine kinase First-line treatment of metastatic non-small cell lung inhibitor cancer with EGFR exon 19 deletions or exon 21 L858R substitution

Maintenance treatment of non-small cell lung cancer

Afatinib EGFR and HER2 First-line treatment of metastatic non-small cell lung tyrosine kinase cancer with EGFR exon 19 deletions or exon 21 inhibitor L858R substitution

Lapatinib EGFR and HER2 Hormone receptor positive metastatic breast cancer tyrosine kinase that overexpresses HER2 inhibitor Along with capecitabine for HER2 overexpressing breast cancer

1.7.1 Cetuximab Cetuximab (IMC-C225, Erbitux [Merck KGaA, Darmstadt, Germany]) is a 152kDa IgG1 human-mouse chimeric moAb that binds selectively to domain III of the extracellular segment of EGFR and directly blocks activating ligand binding, leading to EGFR down regulation following receptor internalization and degradation (Li et al., 2005; Wong, 2005; Vincenzi et al., 2008). Studies on the interaction of cetuximab with EGFR

38 indicated that the fragment antigen binding region (Fab) binds the extracellular ligand binding domain III of EGFR in a region that overlaps partially with the growth factor binding site (Li et al., 2005). Although ligands can still bind to domain I of the extracellular ligand binding domain, EGFR-bound cetuximab sterically blocks domain I from occupying the position observed in the ligand-bound structures (Peipp, Dechant & Valerius, 2008). Cetuximab is able to stabilise EGFR in the tethered closed conformation, inhibiting receptor dimerisation and inducing direct inhibition of its tyrosine kinase activity. The antitumour activity of cetuximab relies also on inhibition of cell cycle progression (Wu et al., 1995), inhibition of angiogenesis (Perrotte et al., 1999), invasion and metastasis (Perrotte et al., 1999), inhibition of DNA repair (Dittmann et al., 2005), and increase in the levels and activation of pro-apoptotic molecules (Liu et al., 2001). Cetuximab has also shown anti-tumour activity in vivo by mediating antibody dependent cellular cytotoxicity (ADCC) and complement dependent cytotoxicity (CDC) (Correale et al., 2010).

Figure 1.7: Structure and development of cetuximab. Reproduced from (Brand et al., 2011)

Cetuximab has been approved by the FDA (Food and Drugs Administration) for the treatment of KRAS WT metastatic colorectal cancer as single agent or in combination with irinotecan-based chemotherapy. Cetuximab is also used for the treatment of recurrent locoregional and metastatic head and neck squamous cell carcinoma (HNSCC) in combination with platinum-based chemotherapy as a first-line treatment.

Unfortunately, treatment response to cetuximab is limited and with poor outcomes in a subset of patients due to the presence of primary or acquired resistance (Bertotti & Sassi, 2015).

1.7.2 Panitumumab Panitumumab (ABX-EGF, Vectibix) is a 147kDa fully humanised immunoglobulin G2 antibody. The anti-tumour activity of panitumumab is exerted by the inhibition of EGF- induced EGFR phosphorylation, the induction of apoptosis, cell cycle arrest and inhibition of angiogenesis (Keating, 2010).

1.7.3 Clinical activities of cetuximab and panitumumab in mCRC The anti-EGFR antibodies cetuximab and panitumumab have been shown to be active in all lines of treatment either as a single agent or in combination with chemotherapy in patients without RAS mutation. The expanded RAS (exons 3 and 4 of KRAS, and exons 2-4 of NRAS) analysis was shown to be superior to the KRAS (exon 2) analysis in predicting efficacy in the expanded RAS wild-type patients and a potential detrimental effect in patients harbouring any RAS mutation in their tumour (Douillard et al., 2013; Bokemeyer et al., 2014).

Cetuximab improved survival of chemorefractory patients compared to best supportive care (Jonker et al., 2007). Panitumumab improved PFS compared with BSA in chemorefractory metastatic RAS WT colorectal cancer (Van Cutsem et al., 2007). Both antibodies have comparable clinical activity as single agents as shown in a phase III head-to-head comparison trial (Price et al., 2014).

Both cetuximab and panitumumab increased the activity of cytotoxic doublet in first- line treatment of RAS WT patients. Survival, PFS and RR benefits have been demonstrated for the combination of FOLFIRI/cetuximab compared with FOLRIFI alone in first-line (Van Cutsem et al., 2009; Maughan et al., 2011; Tveit et al., 2012). Improved RR and PFS have been demonstrated for first-line combination of FOLFOX and cetuximab compared with FOLFOX (Bokemeyer et al., 2009; Maughan et al., 2011; Tveit et al., 2012). Panitumumab also increased ORR, PFS and OS when combined with FOLFOX in first-line treatment of RAS WT metastatic colorectal cancer (Douillard et al., 2010, 2013).

In the second-line setting, addition of cetuximab or panitumumab with irinotecan-based cytotoxics have shown improved RR and PFS, but not OS, compared to chemotherapy alone, probably due to cross-over to anti-EGFR antibodies in later lines of treatment (Sobrero et al., 2008; Peeters et al., 2010; Seymour et al., 2013).

However, only approximately 20% of patients with metastatic colorectal cancer are sensitive to anti-EGFR therapy (Meyerhardt & Mayer, 2005), and even those who initially respond to therapy eventually develop resistance to it. A number of mechanisms are involved in the development of resistance to anti-EGFR therapy.

1.8 Mechanisms of resistance to current anti-EGFR therapy Cancer cells acquire genetic alterations during multi-step process of cancer development, as well as in response to drug treatment in order to adapt to new environment and escape inhibitory effects of the targeted agents. Cetuximab and panitumumab bind to the extracellular domain of EGFR to prevent activation of receptor tyrosine kinase and multiple downstream signal transduction cascades that relate to cell survival, proliferation, metastasis and angiogenesis (Yarden & Sliwkowski, 2001). The RAS-RAF-MARK, PI3K-PTEN-AKT, and JAK/STAT pathways have all been implicated in resistance against EGFR inhibition (Ciardiello & Tortora, 2008). Mutations in KRAS, NRAS, BRAF and PIK3CA genes can lead to constitutive activation of EGFR and intracellular signalling (Hsu et al., 2016; De Roock et al., 2010). RAS mutation accounts for majority of primary resistance to cetuximab or panitumumab in colorectal cancer (Bardelli & Siena, 2010). Acquired resistance to therapy include activation and increased expression of other RTKs, deregulation of EGFR internalisation and degradation, constitutive activation of downstream pathways, KRAS gene alterations, development of EGFR mutations and proto-oncogene MET amplifications (Wheeler et al., 2008; Brand et al., 2011; Yonesaka et al., 2011; Montagut et al., 2012; Bardelli et al., 2013).

1.8.1 RAS status in predicting response to anti-EGFR agents RAS is a family of three small GTPases (KRAS, NRAS and HRAS) that work as downstream effectors within the mitogen-activated protein kinase (MAPK) pathway, coupling EGFR with intracellular signalling cascades (Bos, 1989). Activating mutations in RAS are common in colorectal cancer, and detected in about 50% of patients. The incidence of KRAS mutations in colorectal cancer is 30-40%, most of which occur in exon 2 (codon 12 and 13), exon 3 (codon 61) and exon 4 (codon 117 and 146) (Bardelli & Siena, 2010; Roock et al., 2011). NRAS mutations found in about 3-5% of cases (Fernandez-Medarde & Santos, 2011). Activating KRAS mutations alter the intrinsic GTPase property of the RAS protein sustaining a constantly active GTP-bound state,

41 leads to constitutive MAPK pathway activation that results in mitogenic and antiapoptotic signalling (De Roock et al., 2010).

For the initial mutation analyses in the anti-EGFR trials (cetuximab and panitumumab), tumours were only screened for KRAS mutations in codon 12 and 13 of exon 2 of KRAS gene as these were first to be identified to be associated with resistance to anti- EGFR therapy in metastatic colorectal cancer (Allegra et al., 2009). Subsequent analyses suggested patients harbouring activating mutations in exons 3 and 4 of KRAS or exons 2, 3 and 4 of NRAS gene also confer resistance to EGFR-targeted antibodies (Heinemann et al., 2014; Douillard et al., 2013; Bokemeyer et al., 2014; Van Cutsem et al., 2015). The predictive value of extended RAS analysis was first demonstrated in the PRIME trial (Douillard et al., 2013). The expanded RAS analysis including the detection of mutations in exons 3 and 4 of KRAS gene as well as mutations in the NRAS (exon 2, 3, 4) gene is superior to the KRAS (exon 2) analysis in predicting efficacy in RAS WT patients and potential detrimental effect in patients harbouring any RAS mutations.

Figure 1.8: Expanded RAS analysis.

RAS mutation causes the majority of primary resistance to anti-EGFR therapy. Approximately 50% of acquired resistance cases also occur due to secondary KRAS mutations (Misale et al., 2012; Diaz Jr et al., 2012). Misale et al and Diaz et al showed this acquired alterations in KRAS could be an expansion of pre-existing resistant clones under the pressure of anti-EGFR therapies, given the existence of inter- and intra-tumour heterogeneity.

1.8.2 BRAF mutation The BRAF gene is located on human chromosome 7 and encodes the BRAF protein, serine/threonine-protein kinase B-Raf. It is a member of the Raf family protein kinases involved in cellular signal transduction downstream of KRAS. Activating BRAF mutations mostly occur in codon 600 known as the V600E mutation, found in around 10% of sporadic colorectal cancer cases (Tol et al., 2009b). The BRAF V600E mutation is generally mutually exclusive with RAS mutations. BRAF mutant tumours are associated with poorer survival, right sided tumours and microsatellite instability (Di Nicolantonio et al., 2008). Despite been a well-established negative prognostic marker, there are conflicting results from studies to indicate the predictive value of BRAF status to anti-EGFR agents. Several studies have showed that patients harbouring BRAF mutations do not benefit from anti-EGFR treatments in second or later lines of therapy (Seymour et al., 2013; Laurent-Puig et al., 2009; Di Nicolantonio et al., 2008). But data from first line setting are less clear. BRAF inhibitor vemurafenib has been tested in previously treated patients with BRAF mutated metastatic colorectal cancers in a Phase II trial showing only modest clinical activity (5% of RR) in comparison with the impressive results obtained in BRAF-mutated melanoma (RR 48-67%) (Kopetz et al., 2015). In cell line models, resistance to BRAF-targeted approach seems to be caused by either persistent activation of the EGFR signalling pathway or the activation of other pathways such as PIK3Ca/AKT (Mao et al., 2013; Prahallad et al., 2012). Several clinical trials have assessed the effects of BRAF inhibitors in combination with MEK, EGFR and PI3K inhibitors in colorectal cancer (Corcoran et al., 2015; Yaeger et al., 2015).

1.8.3 Impact of primary tumour location on prognosis and choice of first line biological agent Emerging evidence suggests left- and right- sided colorectal cancers have distinct clinical and molecular characteristics. Approximately two-thirds of colorectal cancers occur on the left side, defined as those distal to the splenic flexure, and remaining one- third affecting the right side (Meza et al., 2010). Right-sided colorectal cancers tend to affect older patients, females, with disease more advanced, poorly differentiated, and of mucinous pathology (Benedix et al., 2010; Powell et al., 2012). These tumours are more likely to be associated with microsatellite instability, RAS/BRAF mutations, and CpG island methylator-phenotype (CIMP)-high status, and worse clinical outcomes in terms of PFS and OS (Missiaglia et al., 2014; Holch et al., 2017). In contrast, left-sided

43 tumours tend to carry chromosomal instability and EGFR amplification, and associated with lower risk of death (Petrelli et al., 2017).

Increasing evidence suggests primary tumour localisation should be considered when selecting a first-line biological agent. Analysis of CRYSTAL and PRIME trials found significantly improved OS and PFS with first line anti-EGFR in the RAS WT left-sided group, but not the RAS WT right-sided cohort (Douillard et al., 2013; Ciardiello et al., 2014). Additionally, pooled analysis of CALGB/SWOG 80405, FIRE-3, and PEAK studies on the impact of both anti-EGFR and anti-VEGF agents found left-sided tumours had significantly improved OS and trend towards better PFS when treated with anti-EGFR therapy, whereas right-sided cancers had significantly improved PFS and trend towards improved OS when treated with anti-VEGF agents (Holch et al., 2017). Further evidence for differential responses to EGFR-targeted treatment based on tumour location is provided by the Phase II AIO VOLFI trial, which reported higher response rate of left-sided metastatic colorectal cancers compared to right treated with FOLFOXIFI and panitumumab (Geissler et al., 2019). Cetuximab response signatures such as amplification of the HER family and stronger EGFR signalling were observed more frequently in left-sided KRAS WT metastatic colorectal cancers.

Figure 1.9: Differences in clinical factors and molecular markers of right- versus left- sided colorectal cancer.

1.8.4 PIK3CA mutations The PIK3Ca/PTEN signalling pathway is downstream of EGFR and can be blocked by EGFR inhibitors (Ciardiello & Tortora, 2008). Molecular alterations in this pathway can

44 lead to activation of downstream signalling through EGFR-independent mechanisms. Activating mutations of PIK3CA, mostly occurring in exons 9 (E542K, E545K) and 20 (H1047R), have been found in 10-20% of colorectal cancers, and can coexist with both RAS and BRAF mutations (Barault et al., 2008; Nosho et al., 2008). They predicted lack of benefit from cetuximab treatment in preclinical models (Karakas et al., 2006; Huang et al., 2007; Jhawer et al., 2008). A large retrospective analysis of colorectal cancers patients treated with cetuximab found only PIK3CA exon 20 mutations predicted lack of response to cetuximab in KRAS WT patient, and exon 9 mutations and KRAS mutations were associated suggesting a secondary role of PIK3CA exon 9 mutations on cetuximab efficacy (De Roock et al., 2010). Subsequently two meta- analyses described negative predictive role of PIK3CA exon 20 mutations in ORR, PFS and OS in WT KRAS metastatic colorectal cancers treated with anti-EGFR therapies (Huang et al., 2014; Mao et al., 2012).

1.8.5 MET overexpression and amplification The MET oncogene encodes tyrosine kinase receptor which is activated by its ligand, Hepatocyte growth factor (HGF), and leads to cell proliferation and survival through activation of the PI3K/AKT, RAC1/cell division control protein 42 (CDC42), RAP1 and RAS/MAPK signalling pathways (Gherardi et al., 2012). In metastatic colorectal cancers MET is amplified in approximately 2% of cases and associated with development of distant metastases and correlated with poor outcomes (Liu et al., 2015; Di Renzo et al., 1995). EGFR-MET interactions with subsequent activation of the MET pathway induced by the overexpression of TGF-α have been proposed to account for acquired resistance to cetuximab in colorectal cancers cells (Troiani et al., 2013). Liska et al showed HGF-mediated MET activation could rescue colorectal cancers cells from cetuximab-induced apoptosis or cell cycle arrest by restoring signalling through the AKT and MAPK pathways (Liska et al., 2011). Effects of cetuximab could be restored by MET inhibition and silencing. In vivo study showed emergence of EMT amplification correlated with acquired resistance to cetuximab and panitumumab in colorectal cancers, and possibly arises from expansion of pre-existing MET amplified clones under pressure of anti-EGFR therapy (Bardelli et al., 2013). In a randomised phase II clinical trial of chemo-refractory KRAS WT metastatic colorectal cancers the combination of anti-HGF moAb and panitumumab led to higher RR and trend for better outcome in patients with MET-overexpressed tumours (Van Cutsem et al., 2014).

1.8.6 EGFR mutations EGFR mutations in colorectal cancers represent a mechanism of acquired resistance and occur in 7-16% of patients treated with cetuximab and 1% treated with panitumumab (Price et al., 2015; Montagut et al., 2012). Montagut et al discovered a point mutation in ECD of EGFR (S492R) in DiFi colon cancer cell line which does not allow binding of cetuximab to receptor but does allow binding of panitumumab (Montagut et al., 2012). Arena et al discovered several other acquired mutations in the EGFR ECD including R451C and K467T in metastatic colorectal cancers patients, and S464L, G465R and I491M in colorectal cancers cell lines made resistant to cetuximab (Arena et al., 2015). Functionally, all of these mutations prevent binding of cetuximab, and only R451C and K467T mutations are permissive for interaction with panitumumab.

Figure 1.10: Structural analysis of EGFR extracellular domain mutants. Adapted from (Arena et al., 2015). (A). General overview of EGFR extracellular domain bound to the antigen- binding fragment of cetuximab (Cetux) as crystallized by Li et al. Domain III contains the 6 mutations S464L, S465R, K467T, I491M, S492R and R451C. Mutations R451C occurs at the interface of contact with domain IV, whereas the rest are located at the surface recognised by cetuximab (B).

One strategy to overcome the resistance to anti-EGFR moAbs mediated by EGFR mutations is to create moAbs that bind to different epitopes located in this region, one example been Sym004.

1.8.6.1 Sym004 as a novel anti-EGFR agent Sym004 is a one to one mixture of two chimeric IgG1 antibodies 992 and 1024 (futuximab and modotuximab), which have been shown to work in synergy via binding of two distinct non-overlapping epitopes on the extracellular domain III of EGFR with high affinities (Wong et al., 1992). These antibodies bind simultaneously to ECD of EGFR, inducing internalisation of EGFR followed by receptor degradation in cancer cell lines, thus inhibits ligand-induced phosphorylation of downstream signalling, and has been shown to be a potent inhibitor of tumour growth in different xenograft models. Sym004 potently inhibited cell proliferation in vitro in the presence of EGF suggesting it is less sensitive to ligand-dependent resistance, and shown superiority than cetuximab in EGFR-dependent tumour xenografts and has the potential to treat tumours with acquired resistance to other EGFR targeted agents (Iida et al., 2013; Pedersen et al., 2010).

Figure 1.11: Model for mechanism of action of Sym004. Adapted from (Pedersen et al., 2010)

In a Phase I trial in where 42 patients with metastatic colorectal cancers and acquired resistance to anti-EGFR inhibitors, 44% had tumour shrinkage, 67% gained overall

47 disease control (Dienstmann et al., 2015). However results from another Phase II study did not show OS or PFS benefit when used in an unselected group of metastatic colorectal cancers patients with acquired EGFR resistance vs standard of care (capecitabine/5FU or best supportive care) (Montagut et al., 2018). Authors suggested coexistence of other genetic alterations related to anti-EGFR resistance untargeted by Sym004 could account for this lack of clinical benefit. Subpopulation analyse of OS in the triple-negative cases (RAS, BRAF and EGFR ECD mutations negative in ctDNA) showed clinically meaningful increase in median OS in Sym004 treatment arms. Sym004 is currently been tested in Phase 1b/2a trial in combination with FOLFIRI in patients with metastatic colorectal cancer progressing after first line therapy.

1.8.7 HER2 amplification and signalling It has been suggested primary or acquired resistance to cetuximab could be due to either HER2 gene amplification or through heregulin up-regulation (a HER3 ligand) leading to persistent extracellular signal-regulated kinase signalling and resistance to cetuximab (Bardelli & Jänne, 2012; Yonesaka et al., 2011). Inhibition of HER2 or disruption of HER2:HER3 heterodimerisation restored cetuximab sensitivity in vitro and in vivo. Authors suggested combination of HER2 inhibitor with anti-EGFR therapy is a potential therapeutic strategy to treat patients with cetuximab-resistant cancers. In another study, the response of 85 patient-derived metastatic colorectal cancer xenografts to cetuximab closely mirrored that of patients, and the cetuximab-resistant subset exhibited HER2 amplification (Bertotti et al., 2011). In a further paper, inhibition of HER2 by lapatinib or disruption of HER2:3 heterodimerisation by pertuzumab restored cetuximab sensitivity in vitro and in vivo (Blackwell et al., 2010). IgG1 molecule MEHD7945A with dual EGFR and HER3 inhibition has shown enhanced efficacy in multiple tumour models compared with nonspecific anti-HER antibodies (Schaefer et al., 2011).

1.8.8 Role of HER3 in anti-EGFR resistance HER3 has been described to have a role as potential biomarker of resistance to anti- EGFR treatments. In a cohort of metastatic colorectal cancers patients treated in second or third line therapy with irinotecan and cetuximab, HER3 overexpression has been associated with shorter PFS and OS (Scartozzi et al., 2011), and predicted lack of efficacy of panitumumab (Seligmann et al., 2015). HER3 has a central role in driving oncogenic signals in tumours (Baselga & Swain, 2009), and preclinical and clinical

48 data have led to the hypothesis that HER3 is an escape pathway to EGFR blockade through a compensatory shift to HER3 signalling, predominately through the PI3K/AKT pathway (Scartozzi et al., 2011; Sergina et al., 2007; Wheeler et al., 2008). HER3 has been found mutated in approximately 11% of colorectal cancers patients (Jaiswal et al., 2013). The significance of HER3 in pathogenesis of colorectal cancer has been shown in an ApcMin/+ mouse model where genetic ablation of HER3 resulted in attenuation of tumorigenesis and this could also be demonstrated in human colorectal cancer cell lines (Lee et al., 2009). AZD8931 treatment of ApcMin/+ mouse model caused a reduction in tumour burden in the small bowel and colon (Alferez et al., 2010). Experiments with breast cancer cells in vitro have demonstrated that initial resistance to anti-proliferative effects of EGFR and HER2 inhibitors may result from reactivation of HER3 (Sergina et al., 2007). A lapatinib-induced HER2:HER3 heterodimer has been demonstrated by FRET/FLIM (Claus et al., 2018). This effect of HER3 escape could similarly occur in colorectal cancer reducing efficacy of anti-EGFR therapy in colorectal cancer.

The potential role of HER2 and HER3 in anti-EGFR resistance means a combination therapy approach involving both EGFR inhibitor and either an HER2 or HER3 inhibitor could potentially overcome this resistance.

1.8.8.1 AZD8931 as a pan-HER inhibitor AZD8931 is a novel oral tyrosine kinase inhibitor with equipotent activity against HER1, HER2 and HER3 (Hickinson et al., 2010). This is the first HER targeted agent developed to have convincingly demonstrated simultaneous inhibition of HER1, HER2 and HER3 receptor activation clinically. In vitro AZD8931 showed reversible inhibition against all three receptor phosphorylation in cells. Studies have shown that AZD8931 was significantly more potent in inhibiting cell growth in vitro and tumour growth in vivo across different cell line models as compared with gefitinib or lapatinib (Yonesaka et al., 2011). In vitro studies indicated that AZD8931 is much more potent in inhibiting activated HER2/HER3 heterodimers, formed at the presence of ligand, than any of the other agents.

In the multicentre randomised Phase II/III FOCUS4-D, where 32 patients with metastatic colorectal cancers WT for BRAF, PIK3Ca, KRAS and NRAS after first line induction therapy were randomised 1:1 to either AZD8931 vs placebo, there was no PFS benefit of AZD8931 compared to placebo (Adams et al., 2018). PANTHER is a Phase I/II trial investigating the effect of using AZD8931 in combination with FOLFIRI

49 chemotherapy in first line setting to determine the importance of schedule and activity in patients with metastatic colorectal cancer.

Figure 1.12: Chemical structure of AZD8931.

1.9 HER dimers as biomarker to monitor tumour response to therapy The development of acquired drug resistance in cancers, due to activation of alternative signalling pathways (or ‘rewiring’) over time is a common problem for targeted therapies.

It has been observed that tumour adaptation under therapeutic pressure activates alternative HER heterodimers, such as EGFR:HER2 heterodimer identified in metastatic colorectal cancer patient-derived tumour xenotransplants (Bertotti et al., 2011), and HER2:HER3 heterodimer which is driving some breast and ovarian tumours (Mills & Yarden, 2010). Such heterodimers might be stabilised by secreted growth factors (e.g., neuregulin-1), or by overexpression of the respective receptors under the selection pressure of anti-HER therapies (Ritter et al., 2007).

From early studies conducted mainly in heavily pretreated chemotherapy-refractory patients and chemotherapy-naïve patients with metastatic colorectal cancer, only 10- 20% of patients clinically benefited from anti-EGFR moAbs (Saltz et al., 2004). The levels of EGFR expression, measured in formalin fixed paraffin-embedded (FFPE) tissues by immunohistochemistry (IHC), did not correlate with clinical response to cetuximab or panitumumab (Chung et al., 2005).

Reported expression levels of HER receptors in colorectal cancer have been highly variable, documented expression for EGFR ranges from 8% to 100%, HER2 1-89%, HER3 16-89%, and HER4 11-81% (Khelwatty et al., 2013; Ooi et al., 2004;

Cunningham et al., 2006; Lee et al., 2002; Ljuslinder et al., 2009; Maurer et al., 1998). No clear association have been found between the expression of HER proteins and prognosis (Kluftinger et al., 1992; Scartozzi et al., 2011). Factors such as the use of different antibodies, differences in antigen retrieval techniques, scoring systems, different patient populations and sample size could contribute to the wide variation in the reported expression of HER proteins (Modjtahedi & Essapen, 2009; Khelwatty et al., 2013).

Therefore, individual levels of HER receptor expression does not contain information that pertains to the HER protein subnetwork connectivity, e.g., dimer formation, which underpins the robustness or plasticity of the tumour proteome and a key part of the resistance machinery (Fruhwirth et al., 2011). The formation of HER dimer may be important both as tumour proliferative driver (Holbro & Beerli, 2003), and as mechanism of tumour escape under pressure of HER1 inhibition.

1.10 Using FRET-FLIM technology to examine HER dimerisation Fluorescence microscopy methods provide a convenient approach to study protein conformation and dynamic interactions between HER network members at the plasma membrane of living cells.

Fluorophores are attached to the proteins of interest by direct conjugation of antibodies to examine protein-protein interactions in cancer cell lines and tissues. A fluorophore is a molecule containing a fluorescent group. The fluorescent group absorbs energy at a specific wavelength to achieve a higher excitation state, before it emits energy at a longer, specific wavelength to revert to ground state. The emission of light is fluorescence and the fluorescent lifetime is the average time that a fluorophore remains in the excited state. The fluorescent lifetime varies according to the fluorophore and its environment, but is consistent for a specific fluorophore in a specific setting.

Foster resonance energy transfer (FRET) is the process of non-radiative energy transfer from one fluorophore (donor) to another (acceptor) via intermolecular Van der Waals (dipole-dipole) interactions whilst in close proximity of within 10nm (Stryer & Haugland, 1967). FRET occurs when emission spectrum of the donor fluorophore overlaps with the absorption spectrum of the acceptor, and the dipole orientations of the fluorophores are parallel. The FRET efficiency is relative to the inverse sixth power of the distance R between the two molecules.

FRET eff = 1/[1 + (R/R0)6]

The Foster radius R0 is the distance at which FRET eff is half its maximum value and depends on the spectral characteristics of the fluorophore.

FRET observations can be made either in steady-state, by detecting the quenching of the fluorescence emitted by the donor as energy is transferred to the acceptor, or time- resolved, by measuring the shortening of fluorescence lifetime of the donor.

Fluorescence Lifetime Imaging Microscopy (FLIM) produces an image based on the differences in the excited state decay rate from a fluorescent sample. The fluorescence lifetime is defined as the average time that a molecule remains in an excited state prior to returning to the ground state by emitting a photon. As the fluorescence lifetime does not depend on concentration, absorption by the sample, sample thickness, photo-bleaching and/or excitation intensity it is more robust than intensity-based methods.

Figure 1.13: Jablonski representation of Forster resonance energy transfer (FRET). Figure adapted from Principles of Fluorescence Spectroscopy, Lakowicz, J.R. As a fluorophore absorbs light, it is excited from the ground state (S0) to a higher vibrational level (S1=first electronic state). At each energy level it exists in a number of vibrational energy levels (horizontal levels). When a molecule returns to S0 from the lowest energy vibrational state of S1, fluorescence occurs. The donor fluorophore (Alexa546) is excited by a light source at a specific wavelength. This fluorescent lifetime decay produces the control lifetime, tau. When Alexa546 is in close proximity to the acceptor, Cy5, energy from the excited Alexa546 is transferred to Cy5, and FRET occurs, thus reducing the control lifetime.

Alexa546 and Cy5 for example form a good fluorophore pair. When Alexa546 is excited by short pulse of light (laser source) the intensity of fluorescence signal is measured as a function of time. The fluorescence lifetime of Alexa546 is then calculated by fitting an exponential model to the decay of signal (Figure 1.3).

FRET/FLIM technology has significantly increased our understanding of molecular pathways involved in cell signalling. In particular, these assays allow study of protein- protein interactions between receptors such as the receptor tyrosine kinase family. The main advantage of using donor FLIM over other FRET assays, which are based on intensity measurements of sensitised acceptor emission is it does not require corrections for donor emission crosstalk in the acceptor channel, direct acceptor excitation at the donor excitation wavelength, or knowledge of detection-correction (which involves quantum yields, donor and acceptor detector efficiencies and varies with conditions such as pH, optical alignment, etc) (Festy et al., 2007; Bastiaens & Squire, 1999; Wouters et al., 2001).

In order to investigate these vital and HER-network based biological mechanisms that correlated with drug resistance, FRET/FLIM technique can be used to measure the HER dimers in cells and FFPE tissues, pre- and post- drug treatment (Bublil et al., 2010; Ng et al., 1999; Tarcic et al., 2009; Bosch-Vilaró et al., 2017). This approach of combining FLIM with single molecule super resolution imaging, which covers the full protein interaction distance scale (from <10nm to 250nm), will bring considerable insights in terms of drug response compared to current proximity ligation assay, reporting on distance scale of the order of -40nm rather than protein dimer separation of <10nm (Söderberg et al., 2006).

1.11 Overcoming resistance to anti-EGFR therapy Aberrant biomarkers including RAS/BRAF/PIK3CA mutations, MET amplification, HER2 amplification, HER3 upregulation can all result in resistance to anti-EGFR therapy through constitutive activation of EGFR downstream signalling. Therefore knockdown or inhibition of the resistance pathways through simultaneous or sequential targeting of the aberrant biomarkers could be an effective way to restore sensitivity to EGFR inhibition. Sorafenib, a potent inhibitor of V600E BRAF has shown activity in combination with cetuximab in V600E BRAF-mutant colorectal cancer cells (Di Nicolantonio et al., 2008). Another way is to target essential downstream effectors of

EGFR such as mitrogen-activated protein kinase (MEK or MAP2K) and mammalian target of rapamycin (mTOR), which are downstream of BRAF and PI3K. Dual EGFR and MEK and/or mTOR inhibitors showed improved response in tumour models harbouring aberrant biomarkers (Queralt et al., 2016; Solit et al., 2006; Zhang et al., 2009).

Another promising approach is immune check-point blockade with anti-CTLA4 (ipilimumab) or PD-1 (pembrolizumab, nivolumab). Mismatch repair status has been found to be a useful biomarker in predicting clinical benefit of immune checkpoint blockade with pembrolizumab. A higher response rate is seen in patients with microstallite instability high (MSI-High) tumours (Le et al., 2015). Further, preclinical and clinical evidence suggested immune system contributes to the therapeutic effect of moAbs in vivo (Perez et al., 2015), which lead to investigation of combination of immune modulators with cetuximab as first line therapy of KRAS WT metastatic colorectal cancers.

1.12 Detection of emerging resistance biomarkers using liquid biopsy Early detection of resistance cell clones to ani-EGFR moAbs can be helpful in guiding personalised anti-cancer therapy. Classical tumour biopsy is invasive and not representative of tumour heterogeneity. Liquid biopsies detecting ctDNA and CTC in blood samples have demonstrated useful tools for monitoring of emergence of drug resistance during course of treatment. Researchers have demonstrated detection of mutations predictive of EGFR moAbs resistance up to 10 months prior to radiological progression (Bettegowda et al., 2014; Misale et al., 2012; Murtaza et al., 2013). In the last decade tumour-derived exosomes have been found to play important function in carcinogenesis and could be used as potential biomarker to monitoring for disease resistance and treatment response.

1.13 Exosomes and carcinogenesis 1.13.1 Exosome biogenesis and cargo sorting is regulated Exosomes are 40-100nm vesicles of endosomal origin derived from most eukaryotic cells and many cancer cells, and present in body fluid including plasma, urine, pleural effusions and ascites (Mathivanan & Fahner, 2012). They were first described as vesicles secreted by sheep reticulocytes that release excess transferrin receptor in the extracellular space during cell maturation (Harding et al., 1983; Pan & Johnstone, 1983), and thus considered for more than a decade as organelles that expel obsolete cellular components in the extracellular space. Exosomes originate from inward budding of the endosomal membrane in multivesicular bodies (MVB) and are released in the extracellular milieu upon fusion with the plasma membrane (Figure 1.6). Sorting of membrane proteins into this pathway usually occurs by ubiquitination of cytoplasmic domains and recognition of this sequence by the endosomal sorting complex required for transport machinery (ESCRT), which is recycled. They are characterised by a lipid bilayer, a cup-shaped morphology under electron microscope, flotation density of 1.10- 1.21g/ml in sucrose gradient, and characteristic surface marker proteins reflecting their endosomal origin. The exosomal cargo compose of lipid rafts, cytoskeleton proteins, membrane trafficking proteins, transmembrane molecules, signalling molecules, mRNA and miRNAs (Taylor & Gercel-Taylor, 2008).

Figure 1.14: Exosome biogenesis occurs within MVBs of endosomal system. Following endocytosis into early endosomes (EE), the cargo is packaged into ILVs within MVBs upon inward budding of the membrane. MVBs can then fuse with lysosomes resulting in degradation of the cargo, or alternatively, the MVBs can fuse with the plasma membrane, resulting in release of the ILVs as exosomes, a process which is regulated by Rab GTPases. Adapted from (Bellingham et al., 2012).

Two main models have been proposed to explain exosome formation and cargo sorting. In the first model, lipids play a key role in endosomal membrane deformation and exosome cargo selection. Exosomes are enriched in sphingolipids (especially ceramide and gangliosides) and cholesterol (Wubbolts et al., 2003; Laulagnier et al., 2004). The activity of neutral sphingomyelinase 2 (nSMase2) which processes sphingolipids to release ceramide, is necessary for exosome formation and cargo sorting (Trajkovic et al., 2008; Kosaka et al., 2010; Mittelbrunn, 2011). Also, clustering of lipids and proteins within the endosomal membrane may function as sorting signal for exosomes and trigger vesiculation of endosome membrane (Vidal et al., 1997; Stoorvogel et al., 2002). In the second model, the Endosomal Sorting Complex Required for Transport (ESCRT) pathway regulates exosome biogenesis and cargo selection (Niel et al., 2006; Baietti et al., 2012b; Putz et al., 2012). The ESCRT system, consisting of ESCRT-0, -I, -II, -III protein complexes, recognises and retains ubiquitylated cargo in the endosome membrane, deforms the endosome membrane and then catalyses the abscission of endosomal invaginations to form vesicles that contain sorted material (Raiborg & Stenmark, 2009; Babst, 2011). Exosomes secreted from different cell types are enriched in several ESCRT proteins including tumour susceptibility gene 101 (TSG101), and charged multivesicular body protein 4 (CHMP4) (Mathivanan et al., 2010). It appears that both modes of exosome formation and cargo sorting may cooperate to yield lipid-driven and ESCRT-regulated exosome production (Babst, 2011). Once formed, exosome secretion in the extracellular space is aided by Rab proteins and SNARE (soluble NSF attachment protein receptor) proteins which guide MVB docking and fusion with the plasma membrane respectively (Beckett et al., 2013; Bobrie et al., 2011; Gross et al., 2012; Peinado et al., 2012). Exosomes are released constitutively, although their secretion can be stimulated by cellular stress or high rates of proliferation (Stoorvogel et al., 2002).

During exosome formation cytosolic proteins are incorporated into their lumen and thus extracellular domains of the transmembrane proteins are oriented toward the extracellular environment following release into the extracellular milieu. This membrane protein orientation has been demonstrated by labelling whole mounts of exosomes for the extracellular domains of membrane protein such as MHC class II molecules for immunoelectron microscopy.

There is little consensus on the mode of exosome interaction with recipient cells and exosomes of different origins may have different modes of interaction with the recipient

56 cells (Bobrie et al., 2011). Proposed mechanisms include entry through endocytosis, ligand-receptor binding or through direct fusion with plasma membrane of the target cell (Mulcahy et al., 2014).

1.13.2 Exosomal protein cargo reflects their origin and is functionally specific Exosomes secreted by different cell types contain commonly present core proteins involved in exosome formation and secretion, and cell-type specific proteins which mediate specific functions upon exosome capture by recipient cells. Exosomes are rich in a set of common plasma membrane proteins including the tetraspanins (CD9, CD63, CD81, CD82), cytosolic proteins (actin, tubulin, heat shock proteins (HSP90, HSC70), flotillin), proteins involved in membrane transport and fusion (annexins, Rab protein family), lysosomal markers – LAMP1, LAMP2, and endosomal proteins from the Golgi apparatus, the endoplasmic reticulum and endocytic pathway which reflect their origin from the endosomal membrane.

Cell type specific proteins include antigen presenting molecules (MHC class I and II proteins) and signal molecules (such as growth factors and receptors) (Théry et al., 2009; Zöller, 2009). Exosome-bound oncogenic tyrosine kinase receptors, such as EGFRvIII and HGF receptor mediate tumour cell-regulated angiogenesis and premetastatic niche formation respectively (Al-Nedawi et al., 2009; Peinado et al., 2012). Full length EGFR has been identified in exosomes isolated from pancreatic cell lines (Adamczyk et al., 2011b) and brain tumours (Graner et al., 2009); and EGFR ligands, such as amphiregulin and TGFα, have been identified in exosomes from breast and colorectal cell lines (Higginbotham et al., 2011).

Overall, the protein composition of exosomes is dynamic and heterogeneous, and in response to environment these changes lead to changes in their extracellular function. The list of oncogenic protein cargo of exosomes is continuously expanding on ExoCarta, which is a database of exosomal proteins and RNA and lipids (Mathivanan & Simpson, 2009). The function and composition of cancer cell-derived exosomes are likely to be different from non-cancerous cells making them potentially suitable cancer biomarkers.

Figure 1.15: Protein composition of a canonical exosome. This schematic showing composition of a typical exosome based on data from 15 proteomic analyses carried out on exosomes purified from cultured cells and biological fluids. Proteins found in at least 30% of different exosomes are listed and asterisk indicates proteins present in at least 50% of exosomes. EEF, eukaryotic translation elongation factor; ERM, ezrin, radixin and moesin; GAPDH, glyceraldehyde 3-phosphate dehydrogenase-activating protein; HSP, heat shock protein; MFGE8, milk fat globule EGF factor 8 protein; MVB, multivesicular body; MVP, major vault protein; RAP1B, RAS related protein 1B; Rho GDI, Rho GDP dissociation inhibitor; TSG101, tumour susceptibility gene 101. Reproduced from (Théry et al., 2009).

1.13.3 Function of tumour-derived exosomes in cancer pathogenesis In the last decade, there has been increased interest and research in exosome function and it has becoming increasingly clear exosomes are potent components of cell-to-cell communication in diverse physiological and pathological processes (Stoorvogel et al., 2002; Théry et al., 2009; Yang & Robbins, 2011). Exosomes secreted by malignant cells modulate tumour-induced immune suppression, angiogenesis, stromal remodelling, tumour cell invasion and formation of premetastatic niche (Al-Nedawi et al., 2009; Théry et al., 2009; Taylor & Gercel-Taylor, 2011; Peinado et al., 2012).

Exosomes can modulate target cells through cell-to-cell interactions, with the transfer of both protein and RNA, including mRNA and miRNA, to neighbouring cells. This information transfer can change the pattern of gene expression of the target cells to induce cell signalling.

1.13.4 S100A8/A9 secreted via exosomes from activated MDSCs can promote tumour growth and metastasis S100A8/A9 belongs to a family of more than 20 low molecular weight intracellular EF- hand motif calcium binding proteins (Marenholz et al., 2004). They are expressed predominately by myeloid cells, including granulocytes, monocytes, monocytic myeloid-derived suppressor cells (MDSCs) and immature cells of myeloid lineage. Although the proteins are products of distinct genes, they are often co-expressed and function mainly as heterodimer of S100A8/A9 (calprotectin) (Roth et al., 2003). In addition to expression within inflammatory milieu, strong upregulation of these proteins has been observed in many tumours including gastric, colon, pancreatic, bladder, ovarian, breast and skin (Gebhardt et al., 2006). S100A8/9 exhibit concentration- dependent dichotomy of function in tumours. At high concentrations (80-100ug/mL) they exert apoptotic effects on tumour cells (Ghavami & Kerkhoff, 2004), whereas at low concentrations (<25ug/mL) promote tumour cell growth (Ghavami & Rashedi, 2008; Turovskaya et al., 2008) and migration (Hermani & Servi, 2006; Hiratsuka et al., 2006; Moon et al., 2008; Ang et al., 2010).

S100A8/9 expression is downregulated during normal differentiation of myeloid precursors to macrophages and dendritic cells. However in cancer, Cheng and colleagues showed tumour-derived factors promote sustained STAT3-dependent upregulation of S100A9 in myeloid precursors which results in inhibition of differentiation to dendritic cells and accumulation of MDSCs (Cheng et al., 2008). S100A8/9 is not only synthesized and secreted by MDSCs, they also bind to carboxylated N-glycans on RAGE of the MDSC cell surface, and activate intracellular signalling that promote their migration (Sinha et al., 2008), suggesting these proteins support an autocrine feedback loop that sustains accumulation of MDSCs and maintenance of immune suppression within tumour microenvironment (Ostrand- Rosenberg, 2008).

S100A8 and A9 heterodimers and homodimers also have paracrine functions through interactions with RAGE and TLR4 on tumour cells (Ichikawa et al., 2011; Källberg et al., 2012; Yin et al., 2013). In colon tumour cells, RAGE binding activates MAPK and

NF-kB signalling pathways, and upregulates expression of genes associated with tumour growth and invasion (Ichikawa et al., 2011; Yin et al., 2013). Extracellular S100A8/9 also contributes to formation of pre-metastatic niche. The release of TNFa, TGFb and VEGFA from the primary tumour promotes their expression in pre- metastatic lung endothelium and lung-associated myeloid cells (Hiratsuka et al., 2006). Furthermore, extracellular expression of S100A8/A9 induces expression of serum amyloid A3 (SAA3) which potentiates its own secretion via TLR4-mediated NF-kB signalling cascade that recruits CD11b+ myeloid cells to pre-metastatic lung which produce a pro-inflammatory milieu that mobilises circulating tumour cells and promotes pulmonary metastasis (Hiratsuka & Watanabe, 2008). SAA is also involved in an intercellular network including hepatocytes that forms the basis of a pro-metastatic niche in the liver (Lee et al., 2019).

Importantly, it has been shown S100A8/A9 is released via exosomes by activated MDSCs (Burke & Choksawangkarn, 2014), and inhibits the differentiation of myeloid precursors to mature, pro-inflammatory macrophages and promotes further accumulation of MDSCs to the pre-metastatic lung, thus initiating an immunosuppressive microenvironment (Cheng et al., 2008). In this microenvironment, NK cell function is impaired, accompanied by significant increase in metastatic spread.

1.13.5 Shuttling of exosomal miRNAs can affect target cell gene expression and signalling MicroRNAs (miRNA) are a class of small, non-coding RNA molecules that are involved in post-transcriptional modification of gene expression. First identified in Caenorhabditis elegans (Lagos-Quintana et al., 2001), miRNA have been shown to possess oncogenic or tumour suppressive properties based on how each miRNA affects gene expression (Ahmed, 2007). miRNAs are highly conserved across many species and implicated in a variety of biological functions including development, proliferation, differentiation, apoptosis, stress response and T cell maturation. miRNA genes represent only 1% of the genome in some species, but at least 30% of the genome is regulated by at least one miRNA.

Most of miRNAs are transcribed by the polymerase II machinery into pri-miRNA transcripts which are several kilobases long, 5’ capped and contain a poly-A tail. They are then cleaved by a combination of RNase III-like enzyme named Drosha and RNA- binding protein DGCR8 into smaller pre-miRNA transcripts which are 70-100nt in size and contain the mature miRNA hairpin. The pre-miRNA is transported out of the

60 nucleus to cytoplasm by exportin -5 and cleaved by RNase II enzyme Dicer to form mature miRNA that is approximately 22nt long. This is incorporated into the protein complex RISC containing Argonaute 2 (AGO2) and several other proteins. Only one strand of RNA guide the RISC complex to its target messenger RNA whilst the other strand is degraded. Upon binding to 3’ UTR of target mRNA, over 90% result in imperfect pairing that promotes translational repression, and the rest of perfect complementarity results in target been degraded. miRNAs dysregulation has been implicated in a range of diseases, from developmental and metabolic diseases to cancer, by affecting pathways involved in development, differentiation, transcriptional and post-transcriptional gene silencing, and stem cell and germ line maintenance. The presence of microRNA in exosomes was first discovered by Valadi et al and referred to as ‘exosomal shuttle RNA’ (esRNA). Exosomes offer protection from degradation by RNase enzymes (Mitchell et al., 2008), thereby permitting miRNA manufactured in one cell to be taken up by a neighbouring cell with presumed effects on gene expression in target cell (Valadi et al., 2007; Kosaka et al., 2010). Distinct exosomal miRNA profiles have been associated with different cancer types including breast (Corcoran et al., 2011), lung (Rosell et al., 2009), ovarian (Taylor & Gercel-Taylor, 2008), and glioblastoma (Skog et al., 2008).

1.13.6 Exosome as a potential biomarker in colorectal cancer Exosomes accumulate in the plasma, ascites, and pleural effusions of cancer patients, and can be isolated non-invasively using liquid biopsy, making their utility as diagnostic biomarkers very promising (Andre et al., 2002; Peinado et al., 2012).

Figure 1.16: Exosomes as circulating biomarker.

It has been suggested that increased tumour exosomes and exosomal proteins in patient serum correlate with disease progression (Taylor & Gercel-Taylor, 2008; Peinado et al., 2012). In a small study involving 91 patients with colorectal cancer, high levels of serum exosomes correlated with high levels of CEA, histological poor differentiation, and shorter overall survival than patients with low serum exosomes (Silva et al., 2012).

Higginbotham et al have shown human colorectal cancer cells release exosomes containing full-length, signalling competent EGFR ligands including AREG, TGF-α and HB-EGF. Exosomes from DLD-1 colon cancer cells with mutant KRAS allele exhibited both higher AREG levels and greater invasive potential than exosomes from isogenic KRAS wildtype cell line. The authors suggested that EGFR ligand signalling via exosomes may contribute to cancer field effect and priming of metastatic niche (Higginbotham et al., 2011). Another study showed mutant KRAS status dramatically affects composition of the exosome proteome by containing higher levels of cancer- associated proteins including KRAS, EGFR, src family kinases and integrins. Exosomes derived from KRAS mutant colorectal cancer cells can transfer mutant KRAS to cells expressing only wild-type KRAS, leading to enhanced three-dimensional growth of these wild-type cells (Beckler & Higginbotham, 2013).

It has also been shown that cetuximab treatment significantly altered the exosomal cargo of a responsive cell line (Caco-2) by increasing abundance of miRNAs and proteins activating proliferation and inflammation and reduced miRNAs and proteins related to immune suppression (Ragusa et al., 2014).

Proteomic profiling of exosomes from primary (SW480) and metastatic (SW620) human isogenic colorectal cancer cell lines showed selective enrichment of metastatic factors MET, S100A8/A9, YNC, signal transduction molecules SRC, JAG1, EFNB2, TNIK, and lipid raft and associated components FLOT1/2, CAV1 and PROM1, in metastatic colorectal cancer cell exosomes relative to primary colorectal cancer cell exosomes (Ji et al., 2013).

In colorectal cancer, many miRNAs show aberrant expression patterns (Schetter et al., 2008; Schepeler et al., 2008), and some have been shown to be involved in tumourigenesis by targeting tumour-associated genes (Asangani et al., 2008b; Sayed et al., 2008; Wang et al., 2009; Huang et al., 2011). Studies have reported role of circulating exosomal miRNAs as biomarker of colon cancer. In one paper, microarray analysis of miRNAs in serum samples of 88 colorectal cancer patients were analysed and found increased levels of miRNAs let-7a, -1229, -1246, -150, -21, -223 and -23a in colorectal cancer patients compared with healthy controls, and were significantly downregulated after surgical resection of tumours (Ogata-Kawata et al., 2014). This change was also confirmed in five colon cancer cell lines compared with normal colon cell line. In particular, serum miR21 has been identified as a diagnostic and poor prognostic biomarker in colorectal cancer (Toiyama et al., 2013). Similarly, miR-204- 5p has been identified as a tumour suppressor through directly inhibiting RAB22A, a member of the RAS oncogene family, and downregulated in colorectal cancers associated with poor prognosis (Yin et al., 2014). Functional analyses showed reintroduction of miR-204-5p into colorectal cancer cells markedly suppressed cell proliferation and invasion both in vitro and in vivo, and increased sensitivity of colorectal cancer cells to oxaliplatin, 5-fluorouracil, and cisplatin.

Though the analysis of exosomes for cancer biomarkers has revealed many possible candidates (particularly miRNAs), none have demonstrated enough promise to be implemented clinically. Predictive biomarkers to drive personalised cancer treatment and increase therapeutic index and overall survival are potentially of great benefit. Exosomal analysis of HER proteins and downstream signalling molecules, HER dimers, S100A8/A9 and immunosuppressive miRNAs all hold promise as feasible candidate biomarkers and may be important clinically in the management of colorectal cancer. Further investigation of the dynamic expressional profile of exosomal contents throughout tumour development and treatment, combined with improved isolation methods, are needed before the clinical implementation of exosomes as biomarker for either diagnosis or prognosis of colorectal cancer.

1.14 Evaluation of exosome isolation methods Despite increasing interest recently in exosome research, there is still no consensus in exosome isolation process. One important challenge is the lack of standard methods to obtain highly pure and well-characterised exosome populations. A robust method with the ability to minimise co-isolating protein aggregates and other membranous particles is a prerequisite to identify consistent and biologically relevant exosome- specific functions and biomarkers. The most widely used technique employed is sequential centrifugation at 300g and 10,000g to remove cells, debris and large vesicles, followed by a two hour 100,000g ultracentrifugation to pellet exosomes. Some groups further purify exosomes through ultracentrifugation on a sucrose gradient or filtration through a 0.2um filter.

Recent expansion in exosome research attracted the release of commercial exosome isolation kits such as ‘Exoquick’ which precipitates exosomes in a single step from small volumes of cultured media or serum. A recent study suggested this procedure is not specific for exosomes derived from endosomes due to particles recovered bearing high levels of CD9 but low amounts of CD63 (Bobrie et al., 2012).

Other methods for exosome purification include immunoaffinity techniques, which use beads or other solid substrates coated with an antibody that recognise an exosomal marker such as CD63. Such techniques are useful in isolating exosomes for analysis of their composition, but alterations of exosomal surface and structure during acid elution processes limit their use for functional studies. Similarly, flotation of vesicles on sucrose gradient results in incomplete separation.

When isolating exosomes attention is needed to minimise possible contaminating vesicles. This could be done by culturing cells in either serum-free medium or medium with serum depleted of small vesicles; by avoiding over-confluency of cells as apoptotic bodies and other membrane vesicles of non-endocytic origin will accumulate when there are more than 5% dead cells; and by completing exosome isolation protocol in single setting as interval freezing of cell supernatant or biological fluid may allow large membrane vesicles to break into smaller exosome-sized vesicles prior to final pelleting (Chaput & Théry, 2011).

1.15 Aims and objectives The aim of this thesis is to investigate HER family receptor activation and dimerisation in colorectal cancer and cancer-derived exosomes upon anti-cancer treatment. Three working hypotheses form the basis of this project:

1. HER family receptor activation and dimer rewiring upon treatment pressure contribute to resistance to anti-HER therapy in colorectal cancer. 2. Proteins and miRNAs in exosomes can be analysed non-invasively and used to monitor response to anti-HER therapy in colorectal cancer. 3. Longitudinal monitoring of circulating exosomal HER protein and miRNA markers in patients with colorectal cancer can help predict treatment failure in advance of biochemical and imaging progression.

The following objectives will be addressed:

1. Chapter three investigates HER family receptor activation and dimerisation upon ligand stimulation and drug treatments in colorectal cancer cell lines. 2. Chapter four investigates the role of exosomes for monitoring HER signalling in colorectal cancer cell lines. 3. Chapter five examines the role of circulating exosomes as biomarker in monitoring HER signalling in colorectal cancer patients undergoing systemic treatment.

Chapter 2

Materials and Methods

2.1 Materials 2.1.1 Colorectal cancer cell lines and culture conditions SW48 human colorectal cancer cells were provided by Merck-Serono (Darmstadt)

Human colorectal cancer cells DLD1, HCT116 KRAS wild type and KRAS G13D isogenic cell lines were kindly provided by Professor Bert Vogelstein (John Hopkins University, USA).

DiFi human colorectal cancer cells were kindly provided by Professor Alberto Bardelli (University of Torino, School of Medicine).

LIM1215 human colorectal cancer cells were obtained from the Ludwig Institute for Cancer Research, New York, NY, USA.

2.1.2 Reagents for cell culture SW48, DLD1 and HCT116 isogenic cell lines were grown in McCoy’s 5A modified media (Sigma-Aldrich). DiFi cells were grown in HAM’s F12 (PAA). LIM1215 cells were grown in RPMI-1640 (Sigma-Aldrich).

Cell culture media was supplemented with 10% heat inactivated Fetal Bovine Serum (FBA) (Gibco), 2mM of L-Glutamine (Sigma), and 2mM of penicillin-streptomycin (PAA).

All cell lines were regularly tested for mycoplasma contamination.

Trypsin/EDTA (0.25% trypsin, 0.02% EDTA) (PAA) was used to mobilise cells for passages.

2.1.3 Cell stimulation EGF: recombinant human epidermal growth factor, 50ng/ml (PeproTech, USA)

NRG1: neuregulin/recombinant human heregulin – b1, 50-100ng/ml (PeproTech, USA)

2.1.4 Chemotherapy drugs and inhibitors Cetuximab (5mg/ml) was obtained from Merck-Serono KGaA (Darmstadt)

Oxaliplatin (150mg/ml) was obtained from the London Oncology Clinic (LOC, London)

SN-38 (the active metabolite of irinotecan 7-ethyl-10-hydroxycamptothecin) was obtained from Sigma-Aldrich.

Sym004 (5mg/ml) was obtained from Symphogen.

AZD8931 (5mg/ml) was obtained from Selleck Chemicals.

2.1.5 Antibodies 2.1.5.1 Primary antibodies used for immunoblotting and dot blots

Table 2.1. Primary antibodies used for immunoblotting and dot blots

Primary Dilution Species Dilution buffer Supplier antibodies

EGFR 1:1000 Rabbit TTBS-0.5% BSA Cell Signalling p-EGFR (Tyr1148) 1:1000 Rabbit TTBS-0.5% BSA Cell Signalling

HER2 1:1000 Rabbit TTBS-0.5% BSA Cell Signalling

p-HER2 (Tyr1248) 1:1000 Rabbit TTBS-0.5% BSA Cell Signalling

HER3 1:1000 Rabbit TTBS-0.5% BSA Cell Signalling

p-HER3 1:1000 Rabbit TTBS-0.5% BSA Cell Signalling

HER4 1:1000 Rabbit TTBS-0.5% BSA Cell Signalling

p-HER4 1:1000 Rabbit TTBS-0.5% BSA Cell Signalling

AKT 1:1000 Rabbit TTBS-0.5% BSA Cell Signalling

p-AKT (Ser473) 1:1000 Rabbit TTBS-0.5% BSA Cell Signalling

ERK1/2 1:1000 Rabbit TTBS-0.5% BSA Cell Signalling

p-ERK1/2 1:1000 Rabbit TTBS-0.5% BSA Cell Signalling

Calnexin 1:1000 Rabbit TTBS-0.5% BSA Cell Signalling

Tubulin 1:1000 Rabbit TTBS-0.5% BSA Cell Signalling

CD63 1:1000 Mouse TTBS-0.5% BSA Genetex

TSG101 1:1000 Mouse TTBS-0.5% BSA Genetex

EpCam 1:1000 Mouse TTBS-0.5% BSA Sigma Aldrich

S100A9 1:1000 Rabbit TTBS-0.5% BSA Prof Thomas Vogl, University of Muenster (0.55ug/ml) ALIX 1:1000 Mouse TTBS-0.5% BSA Cell Signalling

2.1.5.3 Secondary antibodies Polyclonal goat anti-rabbit/mouse HRP (1:2000, Cell Signalling) was used for immunoblotting.

2.1.5.4 Antibody labelling for FRET/FLIM

For FRET, donor antibodies were kindly labelled by Dr Gregory Weitsman. Alexa Fluor 546 was used for donor antibodies and Cyanine 5 (Cy5) dyes (GE Healthcare) for acceptors, both prepared in N,N-Dimethylformamide according to the manufacturers instructions. The antibody of interest and dye were combined in a Bicine buffer (pH 8.6, 1M) and size-exclusion chromatography used to remove excess dye (Zeba TM Spin Desalting microcolumns, Thermo Scientific). A 10µl aliquot of the labelled antibody was used to determine labelling reaction: absorption was measured at 280nm and 558nm for Alexa Fluor 546 and 280nm and 650nm for Cy5 with the following formulas used to generate the dye:protein (D/P) ratios:

(Alexa 546 D/P) = [1.64 (A558)] / [A280 – (0.12 × A558)]

(Cy5 D/P) = [0.68 (A650)] / [A280 – (0.05 × A650)]

Target dye : protein was aimed for 3:1 for acceptor fluorophore (Cy5)-labelled protein, and 1:1 for donor fluorophore (Alexa Fluor 546).

Table 2.2. Directly labelled antibodies employed in FRET/FLIM imaging

Target Clone Manufacturer Dye D:P Concentration

HER1 Ab15 Thermo Cy5 3.5 10µg/ml Scientific HER2 Ab17 Thermo Cy5 3.5 10µg/ml Scientific HER3 B9A11 Monogram A546 0.7 5µg/ml Biosciences

2.2 Methods 2.2.1 Cell line maintenance Cells were cultured in tissue culture dishes (Thermo Scientific) at 37°C in a humidified atmosphere of 5% CO2 and procedures were carried out in sterile conditions in Class II biological safety cabinets. Cells were routinely passaged by removing the media and gently washed once with sterile PBS. Following PBS removal, cells were detached by incubation with 2-5ml of Trypsin-EDTA for 3-5min at 37°C. After detachment of cells, equal volume of complete media was added to the cells in order to inactivate trypsin. Cells were then collected and centrifuged at 1000rpm for 5min. The resulting cell pellets were resuspended in fresh media before reseeding into culture dishes. Cells were passaged for a maximum of two months before being discarded and fresh cultures were set-up.

2.2.2 Storage and retrieval from liquid nitrogen Cells were frozen in liquid nitrogen for long-term storage. Cells plated in 10cm tissue cultured dishes while exponentially growing, were trypsinised and collected by centrifugation at 1000rpm for 5min. The cell pellets were then resuspended in 1ml of freezing media [media supplemented with 20% FBS and 10% dimethylsulphoxide (DMSO, Sigma-Aldrich)] per culture dish, aliquoted in crytotubes (Thermo scientific) and stored at -80°C for 24-72hrs prior to long-term storage in liquid nitrogen tanks.

In order to retrieve cells from frozen stocks, the cryotubes were quickly warmed-up in a 37°C waterbath, cells supplemented with 5ml of fresh media and centrifuged for 5min at 1000rpm to remove the DMSO. The resulting pellet was then gently resuspended in 10ml of fresh media and cells seeded in a 10cm tissue culture dish and incubated at 37°C.

2.2.3 Assessing growth inhibition and cell viability - MTT assay Methylthiazolydiphenyl-tetrazolium bromide (MTT) assay (Mosmann, 1983, Carmichael et al, 1987) was used to measure cell proliferation either in normal media or in presence of increasing concentrations of chemotherapy or inhibitor drugs. The assay is based on the ability of viable cells with active mitochondria to reduce the yellow soluble tetrazolium salt (MTT) to form a blue formazan precipitate. Experiments were conducted in flat-bottom 96 well plates (Corning). 200µl of 2.5-5× 10⁴ cells/ml were seeded per well and left to adhere overnight. Cells were exposed to increasing

70 concentrations of cetuximab, oxaliplatin, SN-38 or Sym004 to determine the IC50 values for each cell line used. Drug dilutions were prepared in complete culture media. For each concentration used, 3-5 technical repeats were performed and untreated control samples were incubated with complete media supplemented with vehicle DMSO for DMSO-diluted drugs. Following 72hr of drug exposure, 20ul of MTT-PBS (5mg/ml) solution were added to each well and plates were left for a further 4h. At the end of the incubation, the cell culture media was removed by gentle aspiration and formazan crystals were resuspended in 200µl of DMSO. The quantity of formazan is measured by absorbance at 540nm on Varioskan Flash Multimode Reader (Thermo Scientific), and this correlated with amount of viable cells. Percentage proliferation was calculated by averaging the absorbance values obtained from each of the technical repeats and by normalising the values of drug treated samples to the absorbance values obtained from the untreated control samples (% proliferation = OD of treated/OD of untreated ×100).

2.2.4 Genotyping of cell lines for KRAS mutations 5×10⁶ cultured colorectal cancer cells were centrifuged at 300g (or 190rpm) for 5min to form a cell pellet, which was then resuspended in 200µl of PBS. Cellular DNA was extracted using the DNeasy Blood and Tissue Kit (Qiagen) and its given protocol. 20ul of proteinase K and 200µl of Buffer AL were added and sample incubated at 56°C for 10min. 200µl of 96% ethanol was added and content transferred into DNeasy mini spin columns placed in 2ml collection tube and centrifuged at 6000g (or maximum of 8000rpm) for 1min. The flow through was discarded along with collection tube, and remaining spin column was placed into a new 2ml collection tube and 500µl Buffer AW1 added and centrifuged for another 1min at 6000g. The residual spin column is placed into another 2ml collection tube with 500µl buffer AW2 added and centrifuged for 3min at 20000g (or 14000rpm). Finally the same spin column was placed into a new 1.5ml microcentrifuge tube and DNA eluded by adding 200µl of Buffer AE to the centre of spin column membrane, incubating for 1min at room temperature, and centrifuging at 6000g for 1min to obtain the flow through containing cellular DNA.

Eluded cellular DNA was quantified using Nanodrop 3300 (Thermo Scientific) and samples diluted to 100ng/ml for PCR.

PCR of DNA was done using the GoTaq PCR Core System (Promega). 100ng of template DNA was mixed with GoTaq DNA Polymerase (5units/µl), MgCl2 (25mM), forward and reverse primers (10µM each), PCR nucleotide mix (10mM) and 5×

Colourless GoTaq Flexi Buffer in a microcentrifuge tube, and underwent PCR reaction as below:

Initial denaturation 95°C 2min 1×

Denaturation 95°C 1min 1×

Annealing 42-65°C 1min 30×

Extension 72°C 1min/Kb

Final extension 72°C 5min 1×

Soak 4°C Indefinite 1×

Analysis of PCR products was done by agarose gel electrophoresis. 1% ethidium bromide-stained agarose gel were made using TAE buffer. 10µl of samples were supplemented with 2µl of loading buffer and run alongside 10µl of ladder in the TAE buffer filled tank at 80V until midway, and imaged under UV light to ensure presence of good quantity of PCR products.

PCR products (1ng/µl/100bp) were sent for KRAS sequencing at Source Bioscience.

2.2.5 Cell treatments Cell stimulation was performed with EGF (100ng/ml) or NRG1 (or Heregulin-b2, 50ng/ml) for HER experiments for desired period at 37°C, after starvation of cells for 4h. Inhibition was performed using humanised monoclonal antibody cetuximab (30µg/ml), or Sym004 (30µg/ml), or AZD8931 for desired duration at 37°C. Experiments always included non-stimulated/inhibited controls.

2.2.6 Immunoblotting analysis 2.2.6.1 Total protein extraction 2.5-5×10⁵ cells/well were seeded in 6 well plates and left to adhere overnight. Cells were stimulated with ligands or treated with drugs for the desired time before total cellular proteins extracted on the next day. Cell culture media was removed on ice, and cells were gently washed twice with ice-cold sterile PBS (Sigma-Aldrich). 100µl of CelLytic M cell lysis reagent (Sigma-Aldrich) supplemented with 1× phosphatase and

72 protease inhibitors (Roche) were added per well and left for 5min on ice. Cell scrapers (VWR) were used to detach the cells from the culture plates which were then collected in 1.5ml Eppendorf tubes (Starlab) and left on ice for 30mins to lyse. Cell extracts were centrifuged at 13,000rpm for 20min at 4°C and the whole cell protein extract (supernatant) was transferred in a fresh tube and stored at -80°C until further use.

2.2.6.2 Protein quantification (BCA) Proteins were quantified using colorimetric RC-DC protein assay from Bio-Rad laboratories. 1:5 dilution of each protein sample was prepared in order to quantify each sample in triplicate (35µl). Reagents S and A were mixed at 1:50 ratio and 25µl of this mix was added per well on a 96 well plate. A Bovine Serum Albumin (BSA) (Sigma- Aldrich) standard curve was constructed to calculate the protein concentration of samples under investigation. 10µl of each of the BSA standards (100µg/ml, 200µg/ml, 400µg/ml, 600µg/ml, 800µg/ml, 1000µg/ml) and the unknown sample replicates were mixed with 200µl of reagent B in corresponding wells, and the content of the plate was mixed on a shaking platform for 10min at room temperature. Absorbance (OD) was measured at 750nm with the Varioskan Flash Multimode Reader (Thermo Scientific). The average absorbance measurement of a blank sample was subtracted from all the measurements of both standard and unknown samples. A standard curve was prepared by plotting the average absorbance value of each BSA standard against its concentration. The standard curve generated can be used to calculate the concentration of each unknown sample.

2.2.6.3 Immunoblotting 30-50µg of protein were heat-denatured for 5min at 95°C in sample buffer containing 100mM Tris-Cl (pH 6.8), 4% SDS, 10% 2-mercaptoethanol, 20% glycerol, and 0.02% bromophenol blue (Life Technologies) and resolved on a 7% Tris-acetate, 10% or 4- 12% Bis-Tris NuPAGE gels (Novex, Pre-cast gels, Life Technologies). 10× running NuPAGE (60.5g Tris-Base, Sigma-Aldrich, 89.5g Tricine, Sigma-Aldrich, and 10g SDS, Sigma-Aldrich), or MOPS (52.3g MOPS, Sigma-Aldrich, 30.3g Tris-Base, Sigma- Aldrich, 5g SDS, Sigma-Aldrich, and 1.5g EDTA, Sigma-Aldrich) running buffers were diluted to 1× with distilled water and used for protein electrophoresis on Tris-acetate and Bis-tris gels respectively at 170V. The protein marker ‘Precision-Plus Protein Kaleidoscope’ (Bio-Rad) was used as a molecular weight protein standard. Proteins were subsequently transferred to polyvinylidene difluoride membranes (Immobilon-P

73 transfer membrane, Millipore) activated by immersion in methanol (VWR) for 1min. The transfer buffer was prepared with a 1:10 dilution of 10x Tris-Glycine buffer (30.3g Tris- Base, Sigma-Aldrich), 144.1g Glycine, Sigma-Aldrich) in distilled water and 20% methanol. Proteins were transferred at 35V for 2.5h at room temperature using the XCell II blot module (Life Technologies) and membranes were subsequently blocked for 1hr at room temperature in blocking buffer containing 5% BSA (Sigma-Aldrich) in 1× TBS, and 0.1% Tween-20 (Sigma-Adrich). All primary antibodies were incubated overnight at 4°C. The antibody dilutions were kept at 4°C and supplemented with 0.02% Sodium Azide (Severn Biotech). Anti-rabbit or mouse IgG, HRP-linked antibodies (Cell Signalling Technologies) were used to detect the primary antibodies’ binding at 1:1000-1:5000 dilutions. Membranes were washed three times in 1×TBS-0.1% Tween- 20 buffer for 10min following incubation with primary and secondary antibodies. The antibody binding to the protein of interest was detected by enhanced chemiluminescence (ECL system, Amersham) on autoradiography film (Kodak X- Omat). Reprobing of membranes with different antibodies specific for proteins of similar molecular weight was performed by stripping the original bound antibody from the membrane with a 15min incubation with the Restore TM PLUS western blot stripping buffer (Thermo Scientific). Alternatively, protein lysates were re-analysed by western blotting in the exact same conditions.

2.2.7 Densitometry analysis The ImageJ software was used to compare the intensity of bands obtained from immunoblotting analysis. Blots were scanned, bands of interest were selected and intensity (pixel density) was measured. The band intensity obtained with the antibody of interest was normalised to the intensity of the respective loading control. In some cases, the values obtained from the treated samples were then normalised to the untreated control which was assigned a value of 1, and presented as fold increase to control sample.

2.2.8 Exosome extraction and analysis 2.2.8.1 Exosome extraction from cell lines Colorectal cancer cell lines were grown in complete media until the cell density reached a confluence of >70%. They were washed thoroughly three times with PBS before subsequent incubation in FBS-free media with or without ligand/drug for 24h. The

74 media was collected and exosomes were enriched using sequential centrifugation as follows: first cell media was centrifuged at 300g for 10mins to remove debris, then 10,000g for 60min at 4°C to remove apoptotic bodies and microvesicles, followed by centrifugation at 100,000g for 120min at 4°C to precipitate exosomes and protein aggregates which was washed in 1ml of PBS. At this point, an aliquot of exosome solution was taken for nanoparticle tracking analysis or electron microscopy and stored at ≤20°C. The resuspended exosomes were centrifuged additionally at 100,000g for 120min at 4°C to precipitate exosomes which was resuspended in PBS, or lysed for western blot analysis, or stored at ≤20°C.

2.2.8.2 Serum separation from whole blood 10-20ml of whole blood was collected in SST Vacutainer and allowed to stand for at least 30mins but no more than 24h at 4°C. Whole blood was centrifuged at 1000g for 10mins at room temperature and the separated serum was aspirated into labelled cryovials. Serum was stored at -80°C for subsequent exosome extraction.

2.2.8.3 Exosome isolation from patient serum Serum was diluted with an equal volume of PBS and centrifuged for 30 min at 2000g, 4°C. The supernatant was carefully transferred to new centrifuge tubes, centrifuged for 45 min at 12,000g, 4°C and transferred again to ultracentrifuge tubes (Beckman Coulter Polyallomer tubes with cap assemblies 453952-AB). The volume of supernatant was supplemented with PBS to achieve a total of 8-10ml. Ultracentrifugation was performed for 2h at 100,000g, 4°C (using a Beckman Coulter type 70.1Ti rotor). The supernatant was discarded and exosome pellets were resuspended in 1ml PBS. The suspension was further diluted with the addition of another 7-9ml of PBS, filtered through a 0.22μm filter (BIOFIL Syringe filter) before final ultracentrifugation for 90min at 100,000g, 4°C. The supernatant was discarded and the exosome pellet was resuspended in 50-200μl PBS. Exosomes were kept at 4°C for immediate use and the unused samples were kept at −80°C for long-term storage.

2.2.8.4 Nanosight quantification of exosomes Concentration and size distribution of exosome preparations were measured using the Nanosight LM10-HS (Nanosight Ltd, Amesbury, UK). Properly diluted exosome suspensions were loaded into the instrument for analysis. Five recordings of 30 seconds each were captured, analysed and the data from at least 5,000 individual particle tracks were analysed using the NTA 3.1 software per sample.

2.2.8.5 Exosome protein quantification (microBCA) Exosomes extracted from cells or serum were lysed in 0.1M NaOH for 10m at room temperature and protein concentration was determined using microBCA protein assay kit according to manufacturer’s instructions (Thermo Scientific).

2.2.8.6 Dot blot analysis of exosomes Exosomes isolated from cultured media or patient serum suspended in PBS was adjusted by protein mass (as per microBCA analysis), or particle number (as per nanosight analysis), and loaded onto nitrocellulose membrane and allowed to air-dry over 20mins. Dried membranes were then blocked in 5% milk or BSA in TBS-0.1% Tween-20 for 60mins prior to incubation with primary antibody diluted in 5% BSA overnight. Next morning membranes are washed thrice with TBST, prior to incubation with secondary anti-rabbit or mouse IgG, HPR-linked antibodies (Cell Signalling Technologies) at 1:2000 diluted in 5% milk or BSA. After incubation with secondary antibody for 60min, membranes are washed thrice with TBST prior to detection by enhanced chemiluminescence (ELC prime, Amersham) using LAS4000 Camera system.

Again, densitometry calculation of each dot was analysed using ImageJ software. The dot intensity of exosomes obtained with the antibody of interest was normalised to the intensity of the loading control CD63 or ALIX.

As dot blots do not allow protein separation by size, the specificity of each primary antibody used in dot blot was tested by running full western blot analysis against protein of interest to ensure no significant non-specific bands were detected.

2.2.8.7 Proteomic analysis of exosomes using reverse phase protein array Isolated exosomes were suspended in PBS and submitted to Edinburgh University for RPPA. Linear values from RPPA results were used for bar graphs and calculation of protein content averages in exosomes.

2.2.8.8 Total RNA extraction from exosomes RNA of exosomes was extracted using the TRIzol™ Plus RNA Purification Kit (Thermo Fisher, UK) according to the manufacturer’ instructions. Final RNA was eluded in 60µl of RNA-free water and stored at -80°C before further use.

2.2.8.9 Gene expression analysis Quantification of gene expression in circulating exosomes was performed by ddPCR (Bio-Rad QX100 system). Normalisation of the RNA, between cycles therapy, was performed using the expression levels of the housekeeping gene 18S (Assay ID, Hs99999901_s1). For each sample equal volume of RNA was used as template and cDNA synthesis performed using the SuperScript® VILO™ MasterMix (Thermo Fisher, UK) according to the below conditions.

2.2.8.9 First strand cDNA synthesis Super Script Vilo master mix Invitrogen 11755-050

RT+ Reaction mix X1

RNA 8µl

Vilo master mix 2µl

Incubation

25°C 10min

42°C 1h

85°C 10min

4°C ∞

MicroRNAs were reverse-transcribed individually using the TaqMan™ MicroRNA Reverse Transcription Kit (Thermo Fisher, UK). For each sample, the normalised amount of RNA was reverse transcribed in a 15µl reaction as miRNAs cDNA synthesis.

RT+ Reaction mix X1

RNA X µl

10✕ Reverse Transcription 1.5µl Buffe

100mM dNTPs (with dTTP) 0.15µl

RNase Inhibitor, 20 U/µL 0.19µl

MultiScribe™ Reverse 1µl Transcriptase, 50 U/µL

Nuclease-free Water Xµl

Incubation

16°C 30min

42°C 1h

85°C 10min

4°C ∞

Reactions was performed on a Bio-Rad thermocycler QX100.

2.2.8.10 miRNA quantification using ddPCR 1.5 µl (18S) or 7.5 µl (miRNAs) of cDNA was added to a 20µl reaction containing 12.5µl 2X ddPCR Supermix for Probes (Bio-Rad) and 1µl 20X TaqMan miRNA PCR primer probe set; each reaction was carried out in triplicate. Thermo cycling conditions were

78 as follows: 95°C for 10min, then 50 cycles of 95°C for 10s and 61°C for 30s and a final inactivation step at 98°C for 12min. PCR products were analysed using the QuantaSoft™ Software (Bio-Rad), with the thresholds for detection set manually based on the results from negative control wells. Droplets were scored as positive or negative based on relative fluorescence intensity in FAM or VIC/HEX channels.

2.2.9 Preparation of cells for imaging Cells were plated in 96 well plates under investigated treatment conditions. After stimulation or inhibition with reagents cells were washed in PBS, and fixed with 4% PFA for 15min, and washed 3 times again in PBS. Cells are then treated with 0.3% triton diluted in PBS for 10min to permeabilise cell membrane, washed with PBS, followed by 0.1% freshly made Sodium borohydride (in PBS at 1mg/ml, Sigma Aldrich, SIBC2310V) for 15min to reduce background autofluorescence. After washing 3 times in PBS cells are blocked with 2% filtered BSA for 30min followed by incubation with conjugated antibody diluted in blocking buffer overnight at 4°C. On the second day conjugated antibody is washed away with PBS and mounting media added at 30µl per well and left for 1 hour prior to imaging.

2.2.10 Preparation of cell pellets In order to translate the dual-antibody based HER dimer assay from application in vitro, on mammalian cell lines, to human tissues, either fresh frozen or formalin-fixed paraffin embedded (FFPE), a block of cell pellets were made in order to test the fluorophore- conjugated antibodies prior to application in human tissue. DLD1 (KRAS WT) cells were used to make cell pellets.

Two 15cm dishes were seeded per pellet, for 24-36 hours, until fully confluent. Pellets were made in control (untreated) cells, as well as stimulated with EGF or neuregulin for 15mins after serum starvation for 4 hours. At fixation, media was aspirated off, and cells washed twice with 20ml of PBS. Sterile cell scrapers were used to scrape off the cells from the plates. 10ml of PBS was added to each dish and mixed in the dish to form a cell suspension, which was transferred to a 50ml centrifuge tube. Cells were spun down for 5min at 200g to form a pellet. Supernatant was then aspirated off, and the cell pellet resuspended in 15ml of neutral buffered formalin. The cells were left to rotate with formalin for 1 hour at room temperature.

The fixed cells were then taken to the tissue bank for further processing. The cells were spun down once again at 200g for 5min to re-form a pellet, and supernatant was decanted. The bottom of the Falcon tube was cut off using a scalpel just above the cell pellet. The pellet was then carefully removed from the Falcon tube and placed in speciwrap containing a few drops of formalin. The speciwrap was folded to fit inside a biopsy cage, and then placed in a cassette with lid. It was processed overnight, as standard, dehydrated in a series of alcohols and xylene, followed by embedding in molten paraffin wax, and held at no more than 60°C until the next morning.

2.2.11 Preparation of FFPE tissue for FLIM imaging FFPE tissue or cell pellets were baked at 56°C overnight, followed by normalisation room temperature. Dewaxing was carried out by immersing slides in xylene for 7.5mins twice, followed by absolute ethanol twice for 5mins each, 70% ethanol for 5mins and washing in distilled water for 3-5min. Antigen retrieval was carried out by heating within Coplin jars in a water bath at 95-98°C for 40min or in the Ventana Antigen retrieval system on a hot plate to 110°C for 25min, with DAKO Hercep Epitope retrieval solution in both cases.

Sample borders were demarcated with a wax pen, and samples washed 3 times with TBS, followed by incubation with 0.1% Sodium borohydride in TBS, to quench autofluorescence, followed by a further 3 washes with TBS. slides were blocked with filtered 1% BSA in TBS, followed by incubation with the fluorophore-conjugated antibody of interest, diluted in filtered 1% BSA in TBS, overnight at 4°C. Slides were then washed 3 times with TBS, for 5mins each time, followed by distilled water. Coverslips were applied with Mowiol medium containing DABCO and left for at least 6 hours at room temperature prior to transfer to -20°C for storage.

2.2.12 Single photon fluorescence lifetime imaging The single photon imaging platform used was built in-house and based on a TE2000 fluorescence microscope with a Nikon 20x air objective (Nikon, Kingston upon Thames, UK), and a CCD camera for detection of fluorescence. Fluorescence excitation is provided by a 553nm emitting laser diode which generates optical pulses with a duration of 40ps at a repetition rate of 80MHz. fluorescence lifetime imaging capability was incorporated with addition of a PMT detector (PMH-100, Becker and Hickl) and time-correlated single photon counting (TCSPC) electronics (SPC830, Becker & Hickl).

Single photon, time-resolved fluorescence is detected at 593+/-10nm using a photomultiplier tube (PMH-100, Becker and Hickl) with a 200ps time resolution and 20× (NA0.5) objective. Automated operation is achieved by incorporating a motorized microscope stage (Marzhauser GmbH, Wetzler, Gemany), a closed-loop objective lens mount with a 500µm range of travel (Piezosystem Jena GmbH, Jena, Germany) and a motorized filter cube selector.

Points for each coverslip or slide were defined manually, logging x,y,z coordinates which were used for imaging using automation. Points were focused using the high precision computer controlled objective stage (Piezo-system). Imaging was then performed sequentially in the Cy3 cube for Alexa-546 detection and the Cy5 cube for Cyanine 5 staining. Epifluorescence images were saved for each point in both channels and then the microscope switched to laser-scanning mode to acquire the single photo data and save time-resolved images. FLIM images have 256 × 256 pixel resolutions and ADC was set to 64 time channels.

2.2.13 FRET/FLIM Analysis The FLIM images obtained with our lifetime microscope were batch analysed by a PC workstation running in-house exponential fitting TRI2 software (v2.7.8.9, CRUK/MRC Oxford Institute for Radiation Oncology, Oxford), based on the Levenburg-Marquardt algorithm with instrumental response iterative reconvolution capabilities. This program outputs files where all the fitting parameters are recorded for each images pixel into Excel format. These files are then analysed to produce a distribution of lifetime, and an average lifetime, from which FRET efficiency can be calculated according to FRET eff = 1–(DA/D), where D and DA are the average lifetime of Alexa546 in the matching D and DA images respectively. Data from cell preparations was fitted with a single exponential decay analysis, whereas data from FFPE cell pellet or tissue was fitted with tri-exponential decay analysis.

Chapter 3

HER family receptor activation and dimerisation upon ligand stimulation and drug treatments in colorectal cancer cell lines

3.1 Introduction The HER pathways are known to be hyperactivated by various mechanisms in cancer. Expression levels of HER receptors in colorectal cancer have been documented to be highly variable and unlikely to directly represent prognostic significance (Kluftinger et al., 1992; Scartozzi et al., 2011). EGFR has been the best-studied receptor and is currently an important therapeutic target in colorectal cancer. Monoclonal antibodies targeting extracellular ligand binding domains of the receptor such as cetuximab and panitumumab are currently widely used in the management of metastatic colorectal cancer. However duration of response to such therapy is limited and development of resistance is almost inevitable. Pan-HER inhibitor such as AZD8931 has been trialed and may represent one way of overcoming resistance to anti-EGFR therapy. Possible roles of heterodimerisation of HER members, autocrine production of EGFR ligands, and upregulation of HER3 ligand heregulin have been suggested to facilitate tumourigenesis and development of resistance to therapy in colorectal cancer, and should be further explored in future studies (Yonesaka et al., 2011; Hobor et al., 2014).

The expression level of HER family is unreliable as a predictive marker for targeted therapies in cancer. HER receptors are able to form alternative dimers and can therefore compensate the loss of function of one receptor during targeted therapies. The HER dimerisation status maybe is more important than HER receptor expression in determining sensitivity or resistance to therapy (Weitsman et al., 2016; Waterhouse et al., 2011). Therefore, the ability to assess the dimerisation receptor pairs within tumours could be useful as a prognostic or predictive biomarker for targeted therapies in colorectal cancer.

In cell lines co-immunoprecipitation can be used to make an assessment of HER dimerisation. However, this method is not ideal as the two HER proteins assessed may not have interacted with each other directly but through a tertiary protein. Assessment of HER dimers in tumours is also difficult as the conventional method of IHC can only assess expression levels and intercellular co-localisation, which does not provide a true indication of their dimerisation status. The same limitation also applies to intensity- based fluorescence/confocal microscopy, which can reveal intracellular co-localisation but not demonstrate definitively the existence of HER dimers.

Assays have been developed to assess protein interaction based on ligation of proteins, including Vera Tag and Proximity Ligation Assays (PLA), which utilise DNA ligation. However, ligation assays can give positive results at a distance between substrates of

30-40nm, which is not confirmative of dimerisation only co-localisation (Söderberg et al., 2006; Baier et al., 2005; Skovsen et al., 2005).

Foster Resonance Energy Transfer (FRET) assays using fluorescence lifetime imaging microscopy (FLIM) have been developed to analyse the interaction between pairs of molecules (Kelleher et al., 2009). This is a gold standard technique for measuring protein proximity within <10nm range (Ng et al., 1999). Fluorescence Lifetime Imaging (FLIM) is well suited to the analysis of interaction within HER family receptors in cells and FFPE tissues. Using FLIM, it is possible to measure FRET to quantify interactions between HER receptors at nanometer scale to establish the potential role that crosstalk might play in colorectal cancer. Such data could help characterize or predict response to anti-HER therapy.

Table 3.1: Summary of methods for assessing protein interactions.

3.2 Aims 1. To investigate the effect of ligand stimulation on HER activation and dimerisation in colorectal cancer cell lines.

2. To examine the effect of drug treatments with cetuximab and AZD8931 on HER activation and dimerisation in colorectal cancer cell lines.

3.3 Results 3.3.1 Characteristics of colorectal cancer cell lines Five colorectal cancer cell lines were used to investigate the different effects of ligand stimulation and drug treatment on HER activation and signalling. These include SW48, DiFi, LIM1215, DLD1 and HCT116 KRAS wild type and KRAS G13D isogenic cell lines. The DLD1 and HCT116 KRAS G13D and KRAS wild type isogenic cell lines were generated by Vogelstein et al., from the knock-out of either the KRAS wild type allele or the KRAS G13D allele from the parental heterozygous G13D cell lines. Table 3.1 outlines the main biological characteristics of the cell lines used in this project in terms of their origin, EGFR and KRAS status, and other alterations of important cancer- related genes.

Table 3.2: Characteristics of colorectal cancer cell lines. Information for each cell line was obtained from the COSMIC database and the Broad Institute Cancer Cell Line Encyclopedia. DLD-1* and HCT-116* are the KRAS wild type isogenic cell lines.

Cell line Source KRAS status EGFR status Mutations in other genes

SW48 Large intestine, Wild type G719S PIK3CA (G914R) adenocarcinoma

DLD-1 * Large intestine, Wild type Wild type PIK3CA (E545K), adenocarcinoma TP53 (S241F)

HCT-116 * Large intestine, Wild type Wild type PIK3CA (H1047R) carcinoma

DiFi Large intestine, Wild type Amplification TP53 (K12R) adenocarcinoma (FAP)

LIM1215 Large intestine, Wild type Wild type B-catenin adenocarcinoma (HPNCC)

3.3.2 Genotyping of colorectal cancer cell lines for KRAS SW48, DiFi, LIM1215, DLD1 and HCT116 KRAS wild type isogenic lines were genotyped for KRAS exon 2 using methods described previously to ensure they were truly wild type for KRAS prior to their use. This result is illustrated in Figure 3.1.

Figure 3.1: Genotyping of colorectal cancer cell lines for KRAS exon 2.

3.3.3 Effect of ligand stimulation on HER activation and signalling in colorectal cancer cell lines SW48, DiFi, DLD1 and HCT116 KRAS wild type isogenic lines were used to study effect of ligand stimulation on HER family receptor phosphorylation and downstream signalling. There were clear heterogeneity observed in basal HER receptor levels among the four cell lines studied. DiFi had the highest level of HER1 and it was derived from a patient with Familial Adenomatous Polyposis known to overexpress HER1. Ligand stimulation induced differential HER transphosphorylation pattern between cell lines. The degree of HER activation as expressed by fold increase of phosphorylated to total protein ratio, and phosphorylated protein to loading control (Tubulin) were calculated using densitometric analysis and presented in Figures 3.2-3.6. EGF induced significant phosphorylation of HER1 and HER2 in SW48, DLD1 (KRAS WT) and DiFi, but not in HCT116; whereas NRG1 caused significant phosphorylation in HER3 in SW48, DLD1 (KRAS WT) and HCT116 (KRAS WT). As receptor phosphorylation is usually preceded by auto or hetero-dimerisation of the receptor, it is probable that EGF induces HER1-HER1 homodimer or HER1:HER2 heterodimer formation in SW48, DLD1 (KRAS WT) and DiFi, whereas NRG induces HER2:HER3 dimerisation in SW48, DLD1 (KRAS WT) and HCT116 (KRAS WT).

Figure 3.2: Effect of ligand stimulation on HER activation among 4 colorectal cancer cell lines. SW48, DLD1 (KRAS WT), HCT116 (KRAS WT) and DiFi were plated on 6 well plates and starved of FBS for 4 hours prior to stimulation with either EGF (100ng/ml) or NRG1 (50ng/ml) for 15mins. Cell lysates were collected, protein quantified and analysed for total and phosphorylated HER proteins using immunoblotting. Tubulin or Calnexin were used as protein loading control. These shown results are representative of three biological repeats.

Figure 3.3: Phosphorylated to Total HER protein ratio in SW48 upon ligand stimulation with EGF or NRG1. Cells were starved of FBS for 4 hours prior to addition of EGF (100ng/ml) or NRG1 (50ng/ml) for 15mins. Densitometric analyses of the immunoblots presented in Figure 3.2 are shown. The Image J software was used for quantification of western blot bands. Fold increase of phosphorylated HER1, HER2 and HER3 were calculated by normalisation to the respective total HER proteins. This graph represents results derived from 3 biological repeats.

Figure 3.4: Phosphorylated to Total HER protein ratio in DLD1 (KRAS WT) upon ligand stimulation with EGF or NRG1.

Figure 3.5: Phosphorylated to Total HER protein ratio in HCT116 (KRAS WT) upon ligand stimulation with EGF or NRG1.

Figure 3.6: Phosphorylated to Total HER protein ratio in DiFi upon ligand stimulation with EGF or NRG1.

Despite differential effect of ligand on HER receptor activation and possibly dimerisation between different cell lines, the effect on downstream signalling is uniformly observed in all: stimulation with EGF induced more significant phosphorylation of ERK whereas stimulation with NRG1 induced a more significant activation on AKT (Figure 3.7).

Figure 3.7: Effect of ligand stimulation on downstream signalling in colorectal cancer cell lines. Plated cells were serum-starved for 4 hours before treated with EGF or NRG1 for 15min prior to protein extraction. Shown above are representative blots of at least 3 biological repeats for each protein of interest.

3.3.4 Time required for HER activation upon ligand treatment EGF was added over time course of 60mins in order to determine time required for maximum activation of HER receptors, and by deduction, receptor dimerisation (Figure 3.8). It was observed that both EGF and NRG1 caused maximum phosphorylation of HER1 at 30mins but only 2mins of treatment resulted in maximum phosphorylation of HER2 and HER3. The effect of NRG1 on HER3 activation sustained over 60min period whereas the effect of ligand on other receptors caused a more transient effect.

Figure 3.8: Ligand stimulation time-course in DLD1 (KRAS WT). Cells were plated on 6 well plates and starved of FBS for 4 hours prior to stimulation with EGF (100ng/ml) or NRG1 (50nmg/ml) at different time intervals over 60mins. Cell lysates were analysed for total and phosphorylated HER proteins. Tubulin was used as protein loading control.

3.3.5 Effect of ligand stimulation on HER dimerisation The effect of ligand stimulation on HER dimerisation in colorectal cancer cells was examined using FRET/FLIM experiments. Cells were seeded on glass coverslips inside 24-well plates, treated with ligands of interest, and fixed with 4% PFA. They were then stained with antibodies against HER receptors of interest, conjugated to a fluorophore, either donor alone, or donor in the presence of the acceptor antibody.

In DLD1 (KRAS WT), there was increased FRET efficiency or dimerisation between HER1 and HER2 upon EGF stimulation (p=0.014, Figure 3.9), and trend toward increased FRET efficiency or dimerisation between HER2 and HER3 upon NRG1 stimulation (p=0.19, Figure 3.10). Attempts were made testing the effect of ligand stimulation on other colorectal cell lines by growing SW48, DiFi and HCT116 on

92 coverslips and 96-well plates, untreated or treated with agents to increase cell adhesion. Unfortunately there were poor cell morphology and significant loss of cells after various buffering and washing stages required for staining.

Previously colleagues in our lab have successfully used FRET/FLIM to quantify HER dimerisation in breast cancer cells and FFPE tissues (Bublil et al., 2010; Weitsman et al., 2016; Nuciforo et al., 2015). Results in this section demonstrate for the first time the successful use of FRET/FLIM to quantify HER dimerisation in colorectal cancer cells.

Figure 3.9: FRET efficiency between HER1:HER2 upon EGF stimulation in DLD1 (KRAS WT) fixed cells. Cells were plated on coverslips inside 24-well plates, serum starved for 4 hours prior to addition of EGF (100ng/ml) for 15min, prior to fixation with 4% PFA, and staining with IgG anti-HER2-Cy5 (acceptor) and IgG anti-HER1-Alexa546 (donor) overnight, at 4°C. D: Donor alone; DA: Donor with acceptor. Cumulative FRET-FLIM histograms show average FRET efficiency from 5 technical repeats each from 2 independent experiments.

Figure 3.10: FRET efficiency between HER2:HER3 upon NRG1 stimulation in DLD1 (KRAS WT) fixed cells. Cells were plated on coverslips inside 24-well plates, serum starved for 4 hours prior to addition of NRG1 (50ng/ml) for 15min, prior to fixation with 4% PFA, and staining with IgG anti-HER2-Cy5 (acceptor) and IgG anti-HER3-Alexa546 (donor) overnight, at 4°C. D: Donor alone; DA: Donor with acceptor. Cumulative FRET-FLIM histograms show average FRET efficiency from 5 technical repeats each from 2 independent experiments.

One intention of the project was to examine changes in HER dimerisation using FFPE tissues of colorectal cancer patients undergoing anti-cancer treatment. Hence the effect of ligand stimulation in FFPE cell pellets was examined, again using DLD1 (KRAS WT). Treatment with EGF induced maximum HER1:HER2 dimerisation at 2mins (p=0.005, Figure 3.11). There was no change in HER1:HER3 dimerisation (Figure 3.12), but reduced HER2:HER3 dimerisation (p=0.001, Figure 3.13), with 5min of EGF stimulation, in DLD1 (KRAS WT).

These results further validate the use of FRET/FLIM technique in measuring HER dimerisation in colorectal cancer FFPE cell pellets.

Figure 3.11: FRET efficiency between HER1:HER2 upon EGF stimulation in DLD1 (KRAS WT) cell pellet. Cells were serum starved for 4 hours prior to addition of EGF (100ng/ml) for 2 and 5min, and processed to form FFPE cell pellets, and stained with IgG anti-HER2-Cy5 (acceptor) and IgG anti-HER1-Alexa546 (donor) overnight, at 4°C. D: Donor alone; DA: Donor with acceptor. Cumulative FRET-FLIM histograms show average FRET efficiency from 5 technical repeats each from 2 independent experiments.

Figure 3.12: FRET efficiency between HER1:HER3 upon EGF stimulation in DLD1 (KRAS WT) cell pellet. Cells were serum starved for 4 hours prior to addition of EGF (100ng/ml) for 15min, and processed to form cell pellets, and stained with IgG anti-HER1-Cy5 (acceptor) and IgG anti-HER3-Alexa546 (donor) overnight, at 4°C. D: Donor alone; DA: Donor with acceptor. Cumulative FRET-FLIM histograms show average FRET efficiency from 5 technical repeats each from 2 independent experiments.

Figure 3.13: FRET efficiency between HER2:HER3 upon EGF stimulation in DLD1 (KRAS WT) cell pellet. Cells were serum starved for 4 hours prior to addition of EGF (100ng/ml) for 15min, and processed to form cell pellets, and stained with IgG anti-HER2-Cy5 (acceptor) and IgG anti-HER3-Alexa546 (donor) overnight, at 4°C. D: Donor alone; DA: Donor with acceptor. Cumulative FRET-FLIM histograms show average FRET efficiency each from 5 technical repeats from 2 independent experiments.

3.3.6. Effect of anti-HER therapies on HER dimerisation in colorectal cancer cell lines Having validated the use of FRET/FLIM on quantification of HER dimerisation in colorectal cancer cells with ligand stimulation, this section examines the effect of anti- EGFR monoclonal antibody cetuximab and small molecule pan-HER inhibitor AZD8931 on HER dimerisation on colorectal cancer cells.

Ligand stimulation with EGF and NRG1 for 15min induced comparable transphosphorylation on DLD1 (KRAS WT) and LIM1215 (Figure 3.14).

Figure 3.14. Effect of ligand stimulation on HER activation in DLD1* (KRAS WT) and LIM1215. Cells were plated on 6-well plates and starved of FBS for 4 hours prior to stimulation with either EGF (100ng/ml) or NRG1 (50ng/ml) for 15min. Cell lysates were collected, protein quantified and analysed for total and phosphorylated HER proteins using immunoblotting. GAPDH was used as protein loading control.

In DLD1 (KRAS WT), treatment with both cetuximab and AZD8931 for 24h caused significant decrease in pHER1, pHER2 and pHER3, with AZD8931 causing more suppressive effect on HER phosphorylation than cetuximab (Figure 3.15). In LIM1215, treatment with cetuximab caused significant decrease in pHER2 but an increase in pHER3. Treatment with AZD8931 on the other hand caused significant decrease in pHER1, pHER2 and pHER3 (Figure 3.15).

Figure 3.15. Effect of cetuximab and AZD8931 on HER activation in DLD1 (KRAS WT) and LIM1215. Cells were plated in 15cm dishes and treated with either cetuximab or AZD8931 in exo-free FBS for 24h. Cell lysates were collected, protein quantified and analysed for total and phosphorylated HER proteins using immunoblotting. GAPDH was used as protein loading control. This result is representative of 3 biological repeats.

The effect of cetuximab and AZD8931 on HER dimerisation of these cell lines were further examined using FRET/FLIM. The pseudocolour lifetime map for cells treated with drug of interest is compared to that of untreated cells, showing either a decrease or increase in fluorescence lifetime (expressed in nanoseconds, ns), which represents energy transfer from donor to acceptor fluorophore due to the nanoscale proximity of both fluorophores.

In LIM1215, there was increased FRET efficiency or dimerisation between HER2 and HER3, but no change in FRET efficiency or dimerisation between HER1 and HER3, with both cetuximab and AZD8931 (Figures 3.16-3.17). The Alexa546 and Cy5 intensity images, as well as lifetime images of donor alone and donor in the presence of acceptor, under untreated versus drug-treated conditions are shown. Graphs show the distribution of pixelwise FRET efficiencies for each condition. Rightward shifting of the distribution curve with drug treatment from control indicates increased FRET efficiency or dimerisation of receptors, whereas a leftward shift indicates reduced FRET efficiency or dimerisation of receptors.

Figure 3.16: FRET/FLIM analysis of HER2:HER3 dimerisation in LIM1215 fixed cells. Cells were plated on coverslips inside 24-well plates, serum starved for 3 hours prior to addition of cetuximab (15µg/ml or 100nM) or AZD8931 (10µM) for 1 hour, prior to fixation with 4% PFA, and staining with IgG anti-HER2-Cy5 (acceptor) and IgG anti-HER3-Alexa546 (donor) overnight, at 4°C. D: Donor alone; DA: Donor with acceptor. Averaged FRET values were calculated on a pixel-by-pixel basis, and the distribution of pixelwise FRET efficiencies plotted for each condition tested. Using paired T-test, p=0.043 for cetuximab vs control, p=0.003 for AZD8931 vs control. Results representative of 5 technical repeats each from 2 independent experiments.

100

Figure 3.17: FRET/FLIM analysis of HER1:HER3 dimerisation in LIM1215 fixed cells. Cells were plated on coverslips inside 24 well plates, serum starved for 3 hours prior to addition of cetuximab (15µg/ml or 100nM) or AZD8931 (10µM) for 1 hour, prior to fixation with 4% PFA, and staining with IgG anti-HER1-Cy5 (acceptor) and IgG anti-HER3-Alexa546 (donor) overnight, at 4°C. p=0.67 for cetuximab vs control, p=0.93 for AZD8931 vs control. Results representative of 5 technical repeats each from 2 independent experiments.

101

In DLD1 (KRAS WT), there was reduced FRET efficiency or dimerisation of HER2:HER3 and HER1:HER3, with cetuximab and AZD8931, under the same conditions. In Figure 3.18, treatment with cetuximab or AZD8931 in DLD1 (KRAS WT) resulted in significant leftward shifting of the FRET efficiency distribution curves for HER1:HER3 dimerisation (p=0.020 for cetuximab vs control, p=0.027 for AZD8931 vs control), indicating reduced HER1:HER3 interactions upon drug treatment.

Figure 3.18: FRET/FLIM analysis of HER1:HER3 dimerisation in DLD1 (KRAS WT) fixed cells. Cells were plated on coverslips inside 24 well plates, serum starved for 3 hours prior to addition of cetuximab (15µg/ml or 100nM) or AZD8931 (10µM) for 1 hour, prior to fixation with 4% PFA, and staining with IgG anti-HER1-Cy5 (acceptor) and IgG anti-HER3-Alexa546 (donor) overnight, at 4°C. p=0.020 for cetuximab vs control, p=0.027 for AZD8931 vs control. Results are representative of 5 technical repeats each from 2 independent experiments.

In Figure 3.19, treatment with cetuximab and AZD8931 in DLD1 (KRAS WT) again resulted in leftward shifting of the FRET efficiency distribution curves for HER2:HER3 dimerisation (p=0.077 for cetuximab vs control, p=0.026 for AZD8931 vs control), indicating reduced HER2:HER3 interactions upon at least AZD8931 treatment.

102

Figure 3.19: FRET/FLIM analysis of HER2:HER3 dimerisation in DLD1 (KRAS WT) fixed cells. Cells were plated on coverslips inside 24 well plates, serum starved for 3 hours prior to addition of cetuximab (15µg/ml or 100nM) or AZD8931 (10µM) for 1 hour, prior to fixation with 4% PFA, and staining with IgG anti-HER2-Cy5 (acceptor) and IgG anti-HER3-Alexa546 (donor) overnight, at 4°C. p=0.077 for cetuximab vs control, p=0.026 for AZD8931 vs control. Results are representative of 5 technical repeats each from 2 independent experiments.

In order to examine whether KRAS status had any influence on HER dimerisation upon anti-HER therapy, the same set of experiments was carried out in DLD1 KRAS G13D cell line. Figure 3.20 shows significant reduction of HER1:HER3 FRET efficiencies with both cetuximab and AZD8931 (p=0.011 for cetuximab vs control, p=0.003 for AZD8931 vs control). Figure 3.21 shows significant reduction of HER2:HER3 FRET efficiencies with cetuximab and AZD8931 (p=0.003 for cetuximab vs control, p=0.003 for AZD8931 vs control). These results indicate KRAS status does not affect HER dimerisation with either agent

103

Figure 3.20: FRET/FLIM analysis of HER1:HER3 dimerisation in DLD1 KRAS G13D fixed cells. Cells were plated on coverslips inside 24 well plates, serum starved for 3 hours prior to addition of cetuximab (15µg/ml or 100nM) or AZD8931 (10µM) for 1 hour, prior to fixation with 4% PFA, and staining with IgG anti-HER1-Cy5 (acceptor) and IgG anti-HER3-Alexa546 (donor) overnight, at 4°C. p=0.011 for cetuximab vs control, p=0.003 for AZD8931 vs control. Results are representative of 5 technical repeats each from 2 independent experiments.

Figure 3.21: FRET/FLIM analysis of HER2:HER3 dimerisation in DLD1 KRAS G13D fixed cells. Cells were plated on coverslips inside 24 well plates, serum starved for 3 hours prior to addition of cetuximab (15µg/ml or 100nM) or AZD8931 (10µM) for 1 hour, prior to fixation with 4% PFA, and staining with IgG anti-HER2-Cy5 (acceptor) and IgG anti-HER3-Alexa546 (donor) overnight, at 4°C. p=0.003 for cetuximab vs control, p=0.003 for AZD8931 vs control. Results are representative of 5 technical repeats each from 2 independent experiments.

104

3.4 Discussion 3.4.1. Effect of ligand stimulation on HER activation in colorectal cancer cells Results in this chapter demonstrated differential basal level of expression of HER protein among different colorectal cell lines. This is in accordance with variable expression of HER family receptors in colorectal cancer patients (Khelwatty et al., 2013; Ooi et al., 2004; Cunningham et al., 2006; Lee et al., 2002; Ljuslinder et al., 2009; Maurer et al., 1998). Aside from tumour heterogeneity, factors including the use of different antibodies in immunohistochemistry, different patient populations and sample size are thought to contribute to this wide variation in the reported expression of HER family members (Khelwatty et al., 2013; Modjtahedi & Essapen, 2009). No clear association has been found between expression of HER proteins and prognosis in patients (Kluftinger et al., 1992; Scartozzi et al., 2011).

Ligand stimulation with EGF and NRG1 resulted in differential transphosphorylation and thus by postulation dimerisation patterns of HER family receptors in different colorectal cancer cell lines examined. There are seven ligands described against HER1: EGF, TGFα, AREG, EREG, BTC, HB-EGF, and EPI, out of which EGF, TGFα and AREG are specific ligands for HER1 (Singh et al., 2016b). While HER2 itself has no known ligand, HER3 binds the growth factor neuregulin (NRG or heregulin) to induce heterodimerisation and signalling. Each ligand activates HER in the same way, through ligand binding, receptor dimerisation, receptor transautophosphorylation, and the recruitment of signalling proteins or adaptors. Prior to ligand binding, domain II of HER1 is folded into domain IV in a tethered conformation that auto-inhibits dimerisation. EGF binding to HER1 monomers at domains I and III promotes domain rearrangement to expose the dimerisation arm in domain II, leading to a stabilised open conformation. Dimerisation of domain II is followed by rearrangements in the transmembrane domain that also leads to rearrangements in the juxtamembrane segment, which then forms interactions with the kinase domain that are important for the formation and stability of HER1 dimer. The C-terminus of the activating kinase inserts into the active site of the receiving kinase, so that this allosteric interaction can activate the receiving kinase, resulting in transautophosphorylation (Yarden & Schlessinger, 1987).

The effect on downstream activation of AKT and ERK was uniform across all cell lines tested, EGF induced predominately ERK phosphorylation whereas NRG1 induced predominately AKT phosphorylation. This finding is consistent with current literature that distinct downstream biological activities are described by activation of different ligands, due to ligand induction of different preferred dimer pairs within the HER family.

105

In this case, although both EGF and NRG1 stimulate ERK1/2, EGF acts primarily through HER1 and HER2 heterodimers to preferentially stimulate PKC, whereas NRG1 acts primarily through HER2 and HER3 heterodimers to preferentially stimulate AKT (Sweeney et al., 2001).

3.4.2. Quantitative analysis of HER dimers in colorectal cancer cells using FRET/FLIM The importance of analysing HER dimer over individual HER protein expression, and the emerging role of HER dimer as biomarker has been well studied in breast cancer. The HER2:HER3 heterodimer is known to be central to oncogenic proliferation in breast cancer, driving cellular proliferation via the PI3K/AKT and MAPK pathways (Aertgeerts et al., 2011; Jura et al., 2009a; Zhang et al., 2012). HER2:HER3 dimer quantification did not correlate with HER2 protein expression by IHC, or with mRNA levels (Spears et al., 2012). However HER2:HER3 dimer predicted breast cancer metastatic relapse independently of HER2 expression up to 10 years post-surgery (Weitsman et al., 2016). The complexity of a network of interacting HER family receptors, which undergo multiple homo- and hetero-dimerisation interactions, is likely to result in this lack of correlation between HER2:HER3 dimer and total abundances of HER2 or HER3. Similarly, understanding the mechanism of HER activation and signalling, examining the HER receptor dimerisation patterns may be more important than expression level per se in colorectal cancer prognosis and prediction of response to targeted therapies.

Among all the techniques FRET is regarded as the gold standard technique to assess protein interaction since the distance between determinants is <10nm (Kelleher et al., 2009; Fruhwirth et al., 2011). Among many ways of determining FRET, methods based on the use of FLIM are the most sensitive in their ability to determine small changes in interaction distance and can thus be used as the basis for highly sensitive assays.

Results in section 3.3.5 demonstrated for the first time the feasibility of using FRET/FLIM to quantify HER dimerisation in colorectal cancer cells and FFPE cell pellets. EGF induced statistically increased HER1:HER2 dimerisation (p=0.014) and NRG1 induced a trend towards increased HER2:HER3 dimerisation (p=0.19) in DLD1 (KRAS WT) fixed cells at 15min. Similarly, EGF caused increased HER1:HER2 dimerisation (p=0.05), no change in HER1:HER3 dimerisation, and reduced HER2:HER3 dimerisation (p=0.001) at 2min in DLD1 (KRAS WT) FFPE cell pellets.

106

Time-course experiments showed EGF caused increased HER1 phosphorylation from 2min to maximum at 30mins, as well as increased HER2 and HER3 phosphorylation at 2min in DLD1 (KRAS WT). NRG1 caused maximum HER3 and HER2 phosphorylation at 2min. As HER dimerisation precedes receptor phosphorylation, the duration of ligand treatment is likely to have huge influence on the degree of HER dimer formation, as seen in the effect of EGF on HER1:HER2 dimerisation comparing 2min vs 5min. In the case of NRG1, 2min of ligand stimulation or less is likely to give maximal HER2:HER3 dimerisation, explaining the non-significant increase in HER2:HER3 interaction at 15min. Should there be more time available for this project, time-course experiments to investigate the effect of ligand on HER dimer formation will find the best duration needed for interaction. Nonetheless, overall results in 3.3.5 demonstrate FRET/FLIM to be an effective and reliable technique in studying HER dimerisation in colorectal cancer cells.

The validation of FRET/FLIM for quantitative analysis of HER dimers in colorectal FFPE cell pellets makes it a useful technique for examining HER dimers in FFPE tissues from colorectal cancer patients in clinic. Indeed, HER2:HER3 dimer quantification by FRET/FLIM predicted breast cancer metastatic relapse and resistance to cetuximab in colorectal cancer when examined by others in our group (Weitsman et al., 2016; Bosch-Vilaró et al., 2017).

Variations in the results of FRET efficiency obtained comparing cell pellets to fixed cells treated under the same conditions could be due to the presence of interfering endogenous and preparation-induced fluorescence emissions (Berezin & Achilefu, 2011; Tadrous et al., 2003). To overcome this, a modified analysis algorithm specifically derived to circumvent the problem of contaminating fluorescence signals in FFPE tissues was used in analysing HER dimers in cell pellets, with the help of Dr Paul Barber.

3.4.3. Effect of Cetuximab on HER activation and dimerisation in colorectal cancer cells EGFR inhibitor cetuximab is clinically indicated for treatment of metastatic colorectal cancer patients without KRAS mutation. However its response rate in KRAS, BRAF and NRAS wild-type tumours accounts for only 41%, leaving a high percentage of non- responders (De Roock et al., 2010). Early clinical studies did not confirm a correlation between EGFR expression level by IHC and clinical response to EGFR inhibitor therapy. Chung and colleagues showed that several colorectal cancer patients who

107 received cetuximab exhibited a major objective response despite the absence of measurable EGFR (Chung et al., 2005). These suggest that EGFR expression alone do not serve as robust predictors for response to cetuximab therapy.

Different mechanisms have been postulated for development of resistance against anti-EGFR therapy, one of which maybe through HER dimer rewiring. HER2 and HER3 have been frequently implicated in the development of treatment resistance to cetuximab. Cetuximab is known to block ligand binding to EGFR, and prevents EGFR homodimerisation and EGFR:HER2 heterodimerisations (Patel et al., 2009). Upregulation of ligands against HER3 has been demonstrated upon anti-EGFR therapy through autocrine secretion. Cetuximab treatment induced HER3 activation and HER2:HER3 dimerisation in HNSCC cell lines (Wang et al., 2017). Inhibition of HER2 by lapatinib or disruption of HER2-HER3 heterodimerisation by pertuzumab restored cetuximab sensitivity in vitro and in vivo (Yonesaka et al., 2011). Thus it was important to explore in vitro effect of ligand stimulation and drug treatment among different colorectal cancer cell lines to see there is also differential receptor activation and dimer rewiring.

Results in 3.3.6 showed treatment with cetuximab for 24h resulted in significant increase in HER3 phosphorylation and decrease in HER2 phosphorylation in LIM1215, but not in DLD1 (KRAS WT). DLD1 (KRAS WT) is a cell line that is known to be more resistant to cetuximab, and harbors both PIK3Ca and TP53 mutations. There is also differential response in terms of HER dimerisation of these two cell lines upon cetuximab treatment using FRET/FLIM. In DLD1 (both KRAS WT and G13D lines), treatment with cetuximab at one hour caused reduction in HER2:HER3 and HER1:HER3 heterodimers formed. KRAS mutational status did not influence on the level of dimers formed. In LIM1215, treatment with cetuximab caused increased HER2:HER3 dimerisation but no change in HER1:HER3 dimerisation. Overall, this suggests in a cetuximab-sensitive colorectal cancer cell line such as LIM1215, cetuximab treatment led to feedback induction of HER3 phosphorylation, coupled with an induction of HER2-HER3 heterodimers, suggesting HER2 allows the kinase- impaired HER3 receptor to gain signalling capacity.

Previous work in our group has shown HER2:HER3 crosstalk was required for endogenous feedback HER3 phosphorylation upon cetuximab treatment in colorectal cancer cells, and co-treatment with cetuximab and lapatinib can limit this effect (Bosch- Vilaró et al., 2017). HER3-mediated resistance to EGFR inhibition with gefitinib has been already described in lung cancer cells (Engelman et al., 2007). Similarly,

108 increased HER3 abundance and HER heterodimers after EGFR-targeted therapy has been shown for breast cancer patients (Tao et al., 2014). Extent of HER2:HER3 dimer formation in tumour blocks predicted likelihood of metastatic relapse after surgery independently of HER2 expression (Weitsman et al., 2016). In the CALGB 80203 study, where metastatic colorectal cancer patients were randomly allocated to receive either FOLFIRI or FOLFOX with or without cetuximab, KRAS wild-type patients with high levels of HER3 did not benefit from cetuximab and were associated with therapy resistance (Cushman et al., 2015). Most recently, subset analysis from the COIN trial (Maughan et al., 2011) showed that in patients treated with chemotherapy with cetuximab, those who had increased HER2:HER3 FRET efficiency in their tumour blocks were associated with superior progression-free survival (Barber et al, unpublished results).

Overall, results suggest feedback activation of HER3 or new HER dimer formation could form the basis for reduced cetuximab efficacy in colorectal cancer patients, and dynamic monitoring of changes in HER dimer formation might be useful to predict efficacy of cetuximab or other anti-HER therapies/combinations.

3.4.4. Effect of AZD8931 on HER activation and dimerisation in colorectal cancer cells AZD8931 is a novel oral small-molecule inhibitor with equipotent activity against HER1, HER2 and HER3 (Hickinson et al., 2010). In vitro AZD8931 showed reversible inhibition against all three receptor phosphorylation in cells, and indicated that AZD8931 is much more potent in inhibiting activated HER2:HER3 heterodimers, formed at the presence of ligand, than lapatinib or gefitinib.

Results in 3.3.6 showed treatment with AZD8931 for 24h caused reduced phosphorylation in HER1, HER2 and HER3 in both cell lines, effect more pronounced in LIM1215, a cell line that is more sensitive to anti-HER1 therapy. Again there is differential induction/suppression of HER dimerisation of these two cell lines upon AZD8931 treatment. In DLD1 (both KRAS WT and G13D lines), treatment with AZD8931 at one hour caused reduction in HER2:HER3 and HER1:HER3 heterodimers formed, similar to effect seen with cetuximab. KRAS mutational status did not influence on the level of dimers formed. In LIM1215, treatment with AZD8931 increased HER2:HER3 dimerisation but did not affect levels of HER1:HER3 dimerisation, again a similar observation to effect seen with cetuximab.

109

The mechanism for dimerisaton shift in HER2:HER3 heterodimer with AZD8931 treatment without a simultaneous increase in HER3 phosphorylation in LIM1215 needs further exploration. It may be this observed increase in HER2:HER3 dimerisation at one hour is a more transient process as AZD8931 is a multi-HER inhibitor, and therefore does not lead to sustained upregulation of HER3. Time-course studies may help further determine the sustainability of dimers formed. AZD8931 is known to cause reversible inhibition of HER1 (IC50, 4nmol/l), HER2 (IC50, 3nmol/l), and HER3 (IC50, 4nmol/l) phosphorylation in cells (Mu et al., 2014). At any given concentration it may have caused differential levels of inhibition of HER family and resultant dimer pairing.

110

Chapter 4

Exosomes as biomarker for monitoring HER signaling in colorectal cancer cells

111

4.1 Introduction The central role of HER signalling in the pathogenesis of colorectal cancer makes HER receptors and dimers attractive biomarkers for analysis. The effect of anti-HER therapy in patients cannot be easily measured or monitored without tissue biopsy. Classical tumour biopsy is invasive and not representative of tumour heterogeneity. Liquid biopsies detecting ctDNA and CTC in blood samples have demonstrated useful tools for monitoring of emergence of drug resistance during the course of treatment. Researchers have demonstrated detection of mutations predictive of EGFR moAbs resistance up to 10 months prior to radiological progression (Bettegowda et al., 2014; Misale et al., 2012; Murtaza et al., 2013).

In the last decade tumour-derived exosomes have been found to play important function in carcinogenesis and could be used as potential biomarker to monitoring for disease resistance and treatment response. Exosomes are 40–150 nm extracellular vesicles (EVs), 1.13–1.19 g/mL in density, identified with classic ‘cup’ morphology (Théry et al., 2002). They are released from diverse cell types under normal and pathological conditions, and are abundant in biofluids and cellular supernatants. The composition of exosomes are cell-type specific and can be distinct from the originated cells due to selective sorting of cargo into exosomes. These proteins included constitutive components of exosomes such as tetraspanins (CD9, CD63, CD81, CD82, CD151), (Mathivanan et al., 2010), cell type-specific molecules such as histocompatibility complex (MHC) class-I and class-II, adhesion molecules, proteases, heat shock proteins, the ESCRT components TSG101 and ALIX (Théry et al., 2002). In addition to proteins and lipids, exosomes contain large amounts of nucleic acids, such as mRNA, microRNA, non-coding RNA and DNA, which are protected from degradation due to the double lipid membrane (Mathivanan & Simpson, 2009). Due to their presence and stability in most body fluids and resemblance of their contents to parental cells, exosomes have great potential as liquid biopsy specimens for various diseases. In particular, exosomes are known to contain combinatorial mixtures of tumour cell signatures, including oncoproteins and nucleic acids that are reflective of genetic or signalling alterations in cancer cells of origin, therefore serve as biomarkers for early detection of cancer (Kosaka et al., 2019; Nedaeinia et al., 2017).

112

Table 4.1: Comparing the advantages and disadvantages of CTCs, ctDNA and exosomes as circulating biomarkers.

Despite significant efforts made in exosome research in recent years, progress has been held back by challenges such as insufficient separation methods, difficulties in characterisation, and lack of specific biomarkers (Li et al., 2019). Currently there is a lack of standardisation and optimisation for exosome purification protocols (Li et al., 2017). Exosome characterisation is challenged by the heterogeneity of EV isolates, resulting in a mixed size distribution and difficulties in profiling of cargo contents (Théry et al., 2006; Gardiner et al., 2016).

113

Table 4.2: Comparison of exosome isolation techniques.

Exosomes have been shown to be involved in initiation, growth, progression and drug- resistance of cancer involving interactions with the microenvironment by transferring oncogenic proteins and nucleic acids (Thakur et al., 2014; Demory Beckler et al., 2013; Al-Nedawi et al., 2008). Exosome uptake induced upregulation of angiogenesis related genes and resulted in enhanced endothelial cell proliferation, migration, and sprouting (Ribeiro et al., 2013). Cancer exosomes are responsible for stromal activation (Khalyfa

114 et al., 2016), induction of the angiogenic switch (Wang et al., 2016), increased vascular permeability (Zomer et al., 2015). Exosomes contributed to metastasis by aiding in the epithelial-to-mesenchymal transition and formation of pre-metastatic niche (Costa- Silva et al., 2015; Peinado et al., 2012). Exosomes have been shown to induce immune escape, by promoting apoptosis of cytotoxic T-cells, expansion and function of regulatory T-cells (Tregs), induction of M2 polarization of macrophages, inhibition of cytotoxicity of natural killer (NK) cells and differentiation of dendritic cells (DC), expansion and activation of myeloid-derived suppressor cells (MDSC) and mobilization of neutrophils (Chornoguz et al., 2011; Taylor & Gercel-Taylor, 2011). Finally, exosomes protected cancer cells from the cytotoxic effects of chemotherapy drugs and transfer chemoresistance properties to nearby cells (Chen et al., 2014; Boelens et al., 2014; Hu et al., 2015; Federici et al., 2014; Yin et al., 2012).

The proteomic of exosomes has been profiled in many cancer cell lines including colorectal cancer (Lee et al., 2018; Demory Beckler et al., 2013; Shiromizu et al., 2017). Full length EGFR has been isolated in exosomes derived from pancreatic cell lines (Adamczyk et al., 2011a) and brain tumours (Graner et al., 2009). EGFR ligands, such as amphiregulin and TGFα, have been identified in exosomes from breast and colorectal cell lines (Higginbotham et al., 2011). Exosomes secreted by HER2- overexpressing breast cancer cells express full-length HER2 on their surface, which can bind and sequester trastuzumab and lower its therapeutic efficacy (Ciravolo et al., 2012).

It has been suggested that increased tumour exosomes and exosomal proteins in patient serum correlate with disease progression (Taylor & Gercel-Taylor, 2008; Peinado et al., 2012). It has also been shown that cetuximab treatment significantly altered the exosomal cargo of colorectal cancer cell line Caco-2 by increasing abundance of miRNAs and proteins activating proliferation and inflammation and reduced miRNAs and proteins related to immune suppression (Ragusa et al., 2014). Proteomic profiling of exosomes from primary (SW480) and metastatic (SW620) human isogenic colorectal cancer cell lines showed selective enrichment of metastatic factors MET, S100A8/A9, YNC, signal transduction molecules SRC, JAG1, EFNB2, TNIK, and lipid raft and associated components FLOT1/2, CAV1 and PROM1, in metastatic colorectal cancer cell exosomes relative to primary colorectal cancer cell exosomes (Ji et al., 2013).

Though the analysis of exosomes for cancer biomarkers has revealed many possible candidates, none have demonstrated enough promise to be implemented clinically. To

115 date, limited research has been done looking at the effect of drug treatment on exosomal proteins, including those involved in HER signalling. It is important to examine whether signature changes in cancer cell-derived exosomes mirrored cellular changes upon drug treatment in colorectal cancer, and to determine their suitability as circulating biomarker in patients with colorectal cancer.

4.2 Aims 1. To optimise exosome isolation and quantification from cultured colorectal cancer cell lines.

2. To optimise exosomal protein analysis by immunoblotting.

3. To examine the effect of cetuximab and AZD8931 on HER activation and signalling in colorectal cancer cell-derived exosomes.

116

4.3 Results 4.3.1. Optimisation of exosome isolation from cultured colorectal cell lines Exosomes were isolated from cultured cells through differential ultracentrifugation, as illustrated in Figure 4.1. At least eight dishes of cultured colorectal cancer cells were used to generate sufficient quantity of exosomes for protein quantification and immunoblotting. Cells were grown in complete media until the cell density reached a confluence of >70%, before incubation with FBS-free media for 24h. The media was collected and centrifuged at variable incremental speeds: 300g for 10min to remove cell debris, followed by 2000g to remove apoptotic bodies, then 12,200g for 45min at 4°C to remove microvesicles. The collected supernatant was filtered and subjected to ultracentrifugation at 100,000g for 120min at 4°C twice to precipitate the final exosome pellet. Exosome pellet was stored at -80°C in various constitutes depending on subsequent planned use. It can be stored as dry pellet for later miRNA or dimer FRET/FLIM analysis; resuspended in 100-200µl of PBS for particle number or protein quantification and dot blot analysis, or lysed for western blot analysis. It was observed that frequent thaw and refreezing of exosomes could affect exosome structure and protein quantity.

Figure 4.1: Optimised exosome isolation protocol from cultured cells. Exosomes were isolated from cultured cell media using differential ultracentrifugation and suspended in native states in PBS and stored at -80°C until further use.

Under normal conditions, 10% fetal bovine serum is added into normal media to facilitate cell proliferation. However fetal bovine serum also contains abundant

117 exosomes which need to be eliminated from cultured media under experimental conditions where cell-derived exosomes were examined. In order to confirm this we tested the amount of exosomes present in McCoy’s 5A media supplemented with various concentrations and constitutes of FBS, including starvation media (i.e., media without FBS), media supplemented with 1% FBS and 10% FBS (normal media), normal media treated with 4h ultracentrifugation in attempt to remove any FBS-derived exosomes, and media supplemented with 10% commercially purchased exo-free FBS (Exo-FBS TM, Cambridge Bioscience). Exosomes were isolated from 15ml of each testing media using protocol described in Figure 4.1, and the particle number of pellet exosomes analysed using Nanoparticle tracking analysis described in Figure 4.3. Exosomes were also isolated from 2 plates of 15ml of McCoy’s 5A cell culture media (after 24h starvation) from DLD1 (KRAS WT), as well as 1ml serum from a colorectal cancer patient, and used as positive control.

Figure 4.2 illustrates the relative amount of exosomes generated from each condition. No particles were isolated from starvation media. There were abundant number of particles isolated from normal media, at least double the amount to those generated from single 15cm dish of colorectal cancer cells, and approximately 10-15 times that generated from media supplemented with exo-free FBS. The variation in particle numbers generated from two preparations of cell culture plates of DLD1 (KRAS WT) could be explained by technical variations and challenges involved in multistage differential centrifugation protocol further explored in Section 4.4.4. Interestingly, there were comparable amounts of particles present in normal media treated with 4h ultracentrifugation and media supplemented with 10% commercial exo-free FBS, suggesting FBS-derived exosomes can be largely removed through 4h of in-house ultracentrifugation.

In cell-line work, at least 24 hours is needed for exosomes release and accumulation in the media prior to isolation, yet prolonged serum starvation could itself lead to protein degradation or activation of survival pathways. In-house ultracentrifugation of normal media in open top reusable ultracentrifuge tubes renders the treated media unsterile and introduce potential source of contamination. Therefore, commercial exo-free FBS was subsequently used as replacement where exosomes were removed through ultracentrifugation under sterile manufacturing conditions without other biological alterations.

118

Figure 4.2: FBS-derived exosomes can be largely removed through ultracentrifugation in house or using commercial exo-free FBS. 15ml of McCoy’s 5A media of different FBS composition underwent exosome extraction protocol and number of particles in final preparation analysed using Nanoparticle tracking analysis. This was compared to exosome isolated from two preparations of 15ml of cultured media using DLD1 (KRAS WT), as well as exosomes isolated from 1ml of patient serum. SM: starvation media; UC: ultracentrifugation; NM: normal media; FBS: fetal bovine serum.

4.3.2. Identification and quantification of exosomes Quantification of exosome concentration and size range were done using Nanosight LM10-HS (Nanosight Ltd, Figure 4.3). Nanoparticle tracking analysis (NTA) is based on the principle that the rate of Brownian movement of nanoparticles in solution is linear to their size. In this method, a 405nm laser light is directed at a fixed angle to the vesicle suspension, and the scattered light is captured using a microscope and high-sensitivity camera. By tracking the movement of individual nanoparticles over time, the software calculates their diameter. The data was processed by the Nanoparticle Tracking Analysis (NTA) 3.1 software, which uses statistical methods to calculate the size and concentration of the vesicles. One advantage of this method is the ability to measure size distribution in polydisperse samples. Figure 4.3 gives an example of NTA analysis of exosome preparation from DLD1 (KRAS WT) cell culture media demonstrating homogenous size range of particles with mode of 103nm and concentration of 4.03E8/ml from five automated runs.

119

Figure 4.3: Analysis of exosomes for particle concentration using Nanosight LM10-HS. A representative image showing appearance of exosomes under nanosight and an example of graph analysis of exosomes isolated from DLD1 (KRAS WT) demonstrating size distribution and concentration range from 5 automated runs using the software.

Since exosomes are nano-sized particles the best way to visualise them is with electron microscopy (Pearson et al, 2015). Standard transmission electron microscopy (TEM) is a powerful imaging technique that enables the investigation of biological structures at a nanoscale, achieved by using electron beams to illuminate the samples to obtain a very high magnification. A number of sample processing steps, including fixation and dehydration, are required prior to analysis. In collaboration with Dr Mark Turmaine (UCL), exosomes stained with gold-labelled anti-CD63 were visualised using immunoelectromicroscopy (Figure 4.4). This method allows detection and direct imaging of transmembrane protein CD63, which bound to a selective secondary antibody labelled with gold particles. We observed classical concave ‘cup-shaped vesicles’ within the expected size range (50-150nm) of exosomes, isolated from both DLD1 KRAS WT cell culture and patient serum. This confirms successful isolation of exosomes from colorectal cancer cells and patient serum through the protocol of differential centrifugation. Although the vesicles are believed to be spherical in solution, they have the appearance of collapsed vesicles after drying. However the overall quality of staining using gold-labelled anti-CD63 was variable and not every single exosome observed by TEM expressed this protein.

120

Figure 4.4: Immunoelectromicroscopy of exosomes. Exosomes from DLD1 (*KRAS WT) and serum of a patient were stained with gold-labelled anti- CD63, showing the cup-shaped morphology. The gold particles (black dots) bound to the exosome membrane indicates presence of tetraspanin CD63. The scale bar is 100nm.

Another way of quantifying exosomes is by protein mass, as variations on exosomal particle sizes mean there is no linear correlation between exosome number and protein mass. Exosomal protein quantification is important for standardisation in proteomic analysis of exosomes, and was measured using a modified microBCA protein assay after exosome preparations lysed in 0.1M NaOH for 10mins at room temperature. However, the threshold of microBCA assay was detected to be at 25µg/ml, as observed by plateauing of absorbance values at lower concentrations (Figure 4.5). This caused a significant challenge for subsequent works with both cell-line and circulating blood derived exosomes as often the resultant exosome pellet generated per treatment condition (despite maximising input volumes) had protein mass lower than threshold detection value for protein quantification, or that if used for protein quantification would result in insufficient residual quantity for further studies.

121

Figure 4.5: Analysis of exosomal protein concentration using microBCA assay. The threshold of microBCA assay was found to be at 25µg/ml.

4.3.3. Detection of exosomes by immunoblotting CD63 and EpCam are well-known proteins constitutively expressed on the surface of exosomes derived from colorectal cancer cells, and are used for detection and quantification of exosomes in addition to Nanosight and protein quantification methods above (Figure 4.6).

Exosomes were isolated from PBS or McCoy’s 5A media of various constitutions as described in Figure 4.2, and stained against CD63 and EpCam using dot blot analysis. Two preparations of exosomes isolated each from 1ml of patient serum were used as positive control. Exosomes were also extracted from cell culture media from single 15cm dish of DLD1 (KRAS WT). There was good correlation of staining against CD63 and EpCam in the samples tested confirming both to be suitable exosomal markers for protein analysis. The tested serum-derived exosomes were abundant in protein quantity. Unfortunately this preparation of DLD1-derived exosomes were low in protein quantity giving very low signal against CD63 and no signal against EpCam. There is some cross-reactivity of anti-human secondary mouse antibodies to FBS-derived exosomes present in 10% McCoy’s 5A media (normal media).

122

Figure 4.6: Detection of exosomes by immunoblotting for surface markers using dot blot. Exosomes were isolated from 15ml of PBS or 15ml of McCoy’s 5A media of various constitutions, including starvation media (i.e., media without FBS), media supplemented with 1% FBS and 10% FBS (normal media), media treated with 4h ultracentrifugation, and media supplemented with 10% commercial exo-free FBS (Exo-FBS TM, Cambridge Bioscience). CD63 and EpCam were used as surface marker against exosomes. Anti-human secondary antibodies were used. Two preparations of exosomes isolated each from 1ml of patient serum were used as positive control. Exosomes were also extracted from 15mls of cell culture media from a 15cm dish of DLD1 KRAS WT (*DLD1 exo).

Dot blot was used in preference to western blot analysis for studying exosomal protein contents as it is more sensitive and requires significantly less protein quantity for loading (around 0.5µg to 1.5µg of protein per single dot). Typically, 8 plates of cultured colorectal cancer cells generated 8-10µg of exosomal proteins. As dot blots do not allow protein separation by size, the antibodies used against exosome surface markers and other proteins of interest in subsequent sections were tested through either siRNA knockdown or running full length western blots to ensure their specificity prior to use in dot blot analysis.

In figure 4.7, protein lysates from untreated and siHER1/siHER2/siHER3 knockdown lines of DLD1 (KRAS WT) and LIM1215 were run on whole length western blots and dot blots, and immunoblotted against anti-HER antibody (Cell signalling) of interest. There was clear and significant reduction of respective HER protein in the siRNA cell lines.

123

Figure 4.7: Verifying specificity of anti-HER1, anti-HER2 and anti-HER3 antibodies using siRNA knockdown. a) 15µg of protein lysate per lane from untreated or siRNA knockdown cell lines was used for whole length western blots. b) 1µg of protein lysate from untreated or siRNA knockdown cell lines was suspended in 5µl PBS and loaded per dot onto nitrocellulose membrane.

The specificity for anti-phospho-HER antibodies were tested using whole length western blots against four different cell lines untreated and stimulated with ligands, demonstrating clear variations in protein quantity and cleanliness of bands generated.

124

Figure 4.8: Verifying specificity of anti-pHER1, anti-pHER2 and anti-pHER3 antibodies for dot blot through whole length western blots. 15µg of protein lysate per lane from untreated cell lines was used for whole length western blots. Cell lines analysed include SW48, DLD1 (KRAS WT), HCT116 and DiFi.

The specificities of antibodies against exosomal surface makers ALIX, EpCam, and cMet were validated using shRNA knockdown of target proteins, by Dr. James Monypenny of Pitmilly (Figure 4.9).

Figure 4.9: Validation of dotblotting antibodies by target protein knockdown. shRNA knockdown of H1975 cell line for ALIX, EpCAM and cMet, and dot blot of exosomes confirmed respective loss of target protein signals.

4.3.4. Exosomal HER expression in different colorectal cancer cell lines The expression level of HER proteins in DLD1 (KRAS WT) cell line derived exosomes were analysed and compared to cell lysate in unstimulated condition, using both western and dot blot analyses, as shown in Figure 4.10. CD63, EpCam and ALIX were

125 clearly more enriched in exosomes compared to cells. Calnexin was used as a negative control for cell-derived exosomes. HER proteins were abundantly present in DLD1-derived exosomes. The levels of HER1 and HER3 were higher in cell lysates whereas HER2 showed a higher concentration in exosomes.

Dot blot analysis of cellular and exosomal proteins were performed after normalising by protein amount. There is overall good correlation between the two methods of immunoblotting, confirming dot blot as a suitable alternative to western blot for exosomal protein analysis, and with the added advantage of requiring much smaller quantity of protein. Additionally, dot blot analysis of exosomal proteins of interest were tested to be best performed using native exosomes suspended in PBS whereas addition of standard lysis buffer was required for western blot.

Figure 4.10: HER proteins expression in DLD1 (KRAS WT) cell lysates and cell-derived exosomes under unstimulated condition. 8 plates of DLD1 (KRAS WT) were grown in 10% exo-free FBS for 24h and both cell lysates (Cell) and exosomes extracted from cultured media (Exo) were prepared. Cellular proteins were quantified using standard BCA assay. Exosomal proteins were quantified using microBCA assay previously described. (A). 8µg of protein was loaded per lane using western blot for both cell lysates and exosomes. (B). 1µg of protein was loaded per dot (5µl volume) for both cell lysates and exosomes for comparison. Calnexin was used as a negative control for cell-derived exosomes.

The composition of HER proteins in exosomes derived from four different unstimulated colorectal cancer cell lines were compared in Figure 4.11. Nanoparticle tracking analysis showed good quality and concentration of exosomes extracted from SW48,

126

DLD1 (KRAS WT), HCT116 (KRAS WT) and DiFi. Exosomes were normalised against nanosight particle number as illustrated. Densitometry quantification of cellular and exosomal proteins was analysed using ImageJ software.

Exosomes extracted from DiFi expressed the highest amount of HER1, which correlated to findings in cell lysates. Exosomes from SW48 expressed the highest amount of HER2, whereas in cell lysates DLD1 (KRAS WT) expressed the highest level of HER2 on immunoblotting. Exosomes from DLD1 (KRAS WT) expressed the highest amount of HER3, as with its cell lysate. Overall, HCT116 (KRAS WT) secreted the lowest amount of all 3 HER proteins out of 4 cell lines tested, this again correlated with the expression levels in its cell lysates.

Figure 4.11: Heterogeneity of HER proteins in CRC cell lines and cell-derived exosomes. (A) SW48, DLD1 (*KRAS WT), HCT116 (*KRAS WT) and DiFi cells were grown in 10% exo- free FBS for 24h and exosomes extracted from cultured media, resuspended in PBS, and quantified by nanosight analysis. The dot blot was loaded according to particle number as protein amount was below detection threshold by microBCA analysis. 7.2×10⁸ particles were loaded per dot (7.5µl volume) for SW48, DLD1 (*KRAS WT), HCT116 (*KRAS WT), and 2.9×10⁸ particles per dot loaded for DiFi. (B) Nanosight blots for quality control of exosome preparations extracted from each cell line. (C) Quantification of dot blot experiment shown in (A). The normalisation was done by number of particles. (D) Quantification of western blot experiment shown in Figure 3.2.

127

4.3.5. Effect of cetuximab and AZD8931 on exosomal HER excretion in colorectal cancer cell lines Having established that HER proteins are expressed and can be reliably detected in colorectal cancer cell-derived exosomes, we wanted to analyse the effect of treatment with cetuximab and AZD8931 on the level of HER proteins expression in cells lysates compared to exosomal preparations.

DLD1 (KRAS WT) and LIM1215 cells were grown to 70% confluence, before changed to fresh media supplemented with 10% exo-free FBS and treated with either cetuximab or AZD8931 for 24h. After drug treatment, exosome preparations from culture media and cell lysates were prepared for protein analysis. Exosomes were first analysed for quality control and quantified using Nanoparticle tracking analysis, and 5µl of exosome preparation in PBS were loaded per dot. Again, densitometry quantification of cellular and exosomal proteins was analysed using ImageJ software. Cellular HER proteins were normalised against GAPDH. Exosomal HER proteins were normalised against CD63 as loading control.

We observed a differential expression of HER proteins in secreted exosomes compare to cell lysates in both cell lines. Cetuximab caused significant increase in HER3 in LIM1215 cell lysates but not exosomes. In DLD1 (KRAS WT), AZD8931 caused significant exosomal HER2 and HER3 secretion but not in cell lysates (Figure 4.12).

128

Figure 4.12: Effect of 24hr treatment with Cetuximab or AZD8931 on exosomal proteins in DLD1 (KRAS WT) and LIM1215. Plated cells were treated with cetuximab (100nM) or AZD8931 (10µM) in media supplemented with exo-free FBS for 24 hours, prior to extraction of cellular proteins or exosomes from media for immunoblotting. Exosomes were first analysed for quality control and quantified using Nanosight, and 5µl of exosome preparation in PBS were loaded per dot. GAPDH for cell lysates and CD63 for exosomes were used as loading control. (A) Western blot and dot blot analysis of HER protein in cell lysate and exosomes derived from DLD1 (WT). (B & C) Relative changes in cellular and exosomal HER proteins with cetuximab and AZD8931 in DLD1 (WT). (D) Western blot and dot blot analysis of HER protein in cell lysate and exosomes derived from LIM1215. (E & F) Relative changes in cellular and exosomal HER proteins with cetuximab and AZD8931 in LIM1215. Shown results are representative of three biological repeats.

129

4.4 Discussion 4.4.1 Isolation and characterisation of colorectal cancer cells secreted exosomes Results in this chapter showed colorectal cancer cell lines including DLD1 (KRAS WT), LIM1215, DiFi, SW48 and HCT116 (KRAS WT) secrete exosomes that can be isolated through differential ultracentrifugation.

In this method, particles of different sizes were separated by the use of series of differential centrifugal forces and times (Livshits et al., 2015). Cells, cellular debris, and large extracellular vesicles (EVs) are separated first with low centrifugal forces (<15,000g). Exosomes are then collected using strong centrifugal forces (100,000- 200,000g) and long centrifugal times. Density gradient ultracentrifugation is an alternative and more stringent form of ultracentrifugation, based on isolation by sucrose-gradient solution in a centrifuge tube, with a gradual density decrease from bottom to top, so that different EVs stay at different layers of gradient (Iwai et al., 2016; Tauro et al., 2012). This method allows preservation of biological structure and properties of exosomes due to the cushioning effect of the sucrose density gradient solution rather than exposure to huge centrifugal force with standard ultracentrifugation. Though accepted as the gold standard of exosome enrichment, ultracentrifugation requires long run times, large sample volumes and expensive equipments. Other techniques used for exosome isolation include size-based isolation techniques, coprecipitation, and immunoaffinity enrichment, each with their advantages and disadvantages (Böing et al., 2014; Zeringer et al., 2015; Zarovni et al., 2015).

10% fetal Bovine Serum (FBS) supplemented cell culture media contained exosomes that can be removed by 4 hours of ultracentrifugation at 100,000g, or by supplementation with alternative commercially available exo-free FBS. The latter was chosen for subsequent experiments in this project due to the constraints of maximum holding volume of 132ml per ultracentrifugation run and inability to maintain sterility of the culture media using reusable open-topped ultracentrifugation tubes.

The quality and concentration of the exosome pellet were examined using Nanosight LM10-HS analysis. The Nanosight technology relies on laser light scattering microscopy on Brownian motion of the particles providing size-based particle count as well as respective concentration (Griffiths et al., 2010). Nanoparticle Tracking Analysis provides a relatively simple and reproducible method for analysing exosomes. However a main drawback of the technique is that in heterogenous samples, it fails to discriminate between exosomes and other similar sized particles, such as clusters of

130 exosomes, cellular debris or protein aggregates, which can create false readings and compromise the reliability of the analysis. Therefore this method is often combined with sample imaging by electron microscopy (Wu et al., 2015b), and in this project, electromicroscopy with immunostaining against exosomal marker CD63 (Viens et al., 2008). Although we and other researchers described the morphology of exosomes as ‘cup-shaped’ on TEM, it seems to be an artefact generated by fixation and contrasting steps (Théry et al., 2006), which is also associated with shrinking of vesicles (van der Pol et al., 2014). Studies using scanning electron microscopy and cryo-electron microscopy revealed exosomes have round/spherical shape (Raposo & Stoorvogel, 2013).

Colorectal cancer cells grown on 15cm dishes to full confluence produced exosomes that were abundant in number (1E10) but low in protein mass. Exosomes are constantly produced by cells at steady rate however a minimum of 24h was required to accumulate sufficient exosomes per dish for protein analysis. Exosomal protein concentration can be analysed using microBCA assay, but required a threshold protein concentration of 25µg/ml.

Colorectal cancer cell-derived exosomes can be characterised by the presence of surface markers CD63 and ALIX, which were also used as normalisers for exosomal protein analysis where quantification of limited amount of protein was not possible. Exosome proteins mainly originate from cytoplasm and cell membrane. Calnexin, excluded in EV components, was not found in our cell-derived exosome samples, indicating the absence of contamination with intracellular proteins during exosome purification. Exosomes contain multiple proteins that not only reflect the exosome’s origin but also affect the homeostatic equilibrium of the organism and disease development. CD63 is a tetraspanin that is known to be enriched in the membrane of exosomes and considered one of the classical markers of exosomes (Théry et al., 1999). ALIX is involved in the budding of membranes into endosome and exosomes. In association with endosome sorting complex required for transport (ESCRT), it also helps sort proteins into the endosomes to be released as exosomes. It binds to syntenin and syndecan to control the numbers of exosomes formed (Baietti et al., 2012a). Currently, the presence of known exosomal proteins have been analysed by mass spectrometry, ELISA and western blotting. However most of these methods require a series of pre-treatment measure, such as exosome lysis and protein isolation, and many tedious sample purification procedures, all of which result in difficulties in accurate quantitative analysis of proteins and restricting their application in exosome biomarker screening (Witwer et al., 2013).

131

In this project, western blot analysis of exosomal proteins was shown to be effective but limited by low abundance of protein. Dot blot analysis was developed and used in preference due to its sensitivity and ability to test multiple markers using a small amount of protein. However, the principle of western blots separate proteins by their size and electronic charges with sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), dot blots omit this step and probe the target protein by the specificity of the recognised antigen (Heinicke et al., 1992). Therefore, it was important to use highly specific antibodies in dot blots analysis of exosomal proteins to minimise non-specific bindings, and this required antibody validation to ensure the reliability and reproducibility of data obtained.

The specificities of antibodies used against HER1, HER2, and HER3 were confirmed through siRNA knockdown of target proteins. The specificities of antibodies against pHER1, pHER2 and pHER3 were confirmed using full-length western blots. The specificities of antibodies against exosomal surface makers CD63, ALIX, EpCam, and c-Met were validated using shRNA knockdown of target proteins, by Dr. James Monypenny of Pitmilly.

4.4.2. Exosomal HER protein level as marker of HER expression in colorectal cancer cell lines It has been well documented that cancer exosomes selectively package proteins and nucleic acids in order to exert desired effect such as distant priming of metastatic niche (Valadi et al., 2007; Colombo et al., 2014; Hoshino et al., 2015). Their composition is cell type dependent and can be influenced by different cellular conditions or treatments (Hessvik & Llorente, 2018). Also, exosomes released by a cell line are heterogeneous in contents (Kowal et al., 2016).

Results in 4.3.3 showed under unstimulated conditions, DLD1 (KRAS WT)-secreted exosomes were selectively enriched in CD63, ALIX and EpCam compared to cell lysates. DLD1 (KRAS WT) packaged HER1, HER2 and HER3 in their secreted exosomes to different extends compared to the cell lysates.

When comparing exosomes secreted from four different colorectal cancer cell lines, there was correlation between HER expressions in cells compared to their respective exosomes under unstimulated conditions. DiFi, a cell line with the highest expression level of HER1, also secreted exosomes with the highest level of HER1. HER3 expression was highest in DLD1 (KRAS WT) cell lysates as well as DLD1 (KRAS WT)-

132 derived exosomes. Similarly, the cell line with the lowest HER expressions, HCT116 (KRAS WT), secreted exosomes with the lowest amount of HER proteins. Our results showed for the first time that exosomal HER protein levels correlate with cellular protein levels in colorectal cancer cells at baseline, and could serve as useful biomarker for measuring HER expression in cells.

4.4.3. Effect of cetuximab and AZD8931 on exosomal HER protein secretion from colorectal cancer cell lines In Chapter 3 we saw treatment with cetuximab in sensitive cell line LIM1215 caused feedback upregulation of phosphorylated HER3 at 24h and increased HER2-HER3 dimer formation at 1hr. This effect was not seen with cetuximab-resistant cell line DLD1 (KRAS WT). Treatment with AZD8931 led to downregulation of all phosphorylated HER proteins in both cells, with more pronounced effect seen with LIM1215. This Chapter aimed to examine how these changes following drug treatment is translated into exosomes which may help gain insight into their role as potential biomarkers in monitoring patients through treatment in clinic.

Results in 4.3.4 showed treatment with cetuximab did not significantly alter exosomal HER protein secretions in either DLD1 (KRAS WT) or LIM1215. Treatment with AZD8931 caused a significant increase in exosomal HER2 and HER3 secretion from DLD1 (KRAS WT), and a trend increase in exosomal HER2 secretion from LIM1215. These changes in exosomes did not correlate with protein changes seen in cell lysates. Comparing these results to findings in literature, van Dommelen et al reported reduced EGFR and pEGFR in extracellular vesicles released from A-431 cells (epidermoid carcinoma) after cetuximab treatment, which mirrored parental cellular changes (van Dommelen et al., 2016). Ragusa et al found treatment with cetuximab significantly altered cargo of two colorectal cancer cell-derived exosomes, which did not precisely mirror those in source cells (Ragusa et al., 2014). In Caco2 (cetuximab-sensitive), cetuximab increased abundance of exosome contents activating proliferation and inflammation, and reduced those involved in immunosuppression. In HCT116 (cetuximab-resistant), cetuximab increased exosome contents related to cell to cell signalling, cell growth and proliferation. At steady state, around 10% of miRNAs expressed were different by exosomes and their source cells for both cell lines, and cetuximab treatment significantly accentuated this difference. The predominant biomolecular function of exosomes is likely dependent on their specific molecular phenotype as on that of their source cells. However environmental factors can impact

133 the behaviour of tumour-derived exosomes. In response to drug treatment, colorectal cancer cells specifically alter exosomal profiles relating to drug resistance and tumour cell survival in vitro, reflecting in vivo phenomena occurring in patients.

A further step of interest would be to examine the effect of drug treatments on exosomal HER heterodimers, to see whether there is correlation with effect seen on cells. Coban et al demonstrated that it is possible to apply FRET/FLIM technology to interrogate the HER heterodimers on circulating exosomes from lung cancer xenografts (unpublished data). Galazi et al has also developed a protocol for immunostaining and FLIM of patient-derived exosomes to characterise HER heterodimers from clinical samples of patients with NSCLC, and found significant reduction of HER1-HER3 dimer in a proportion of patients following treatment with Cisplatin (Ng et al., 2018).

4.4.4. Challenges with cell-based exosome work Exosome work is overall challenging as the isolation and enrichment of exosomes is an elaborate and labour-intensive task, and technical variations may impact interpretation of results. A large volume of cell media is required to generate sufficient quantity of exosomes for protein analysis using differential centrifugation. Cells were grown in large number of dishes over two days to reach sufficient confluence before addition of drug of interest in exo-free FBS. It was difficult to control for equal cell numbers per plate prior to drug treatment per condition despite homogenous plating of cells. A minimum of further 24h was needed to allow sufficient accumulation of exosomes prior to extraction. Thereon a full exosome extraction protocol takes place over two days, plus the analysis of exosomes for quality control and concentration using Nanosight prior to protein staining. Sometimes different treatment conditions were tested on consecutive weeks due to limitations on maximum holding volumes in the ultracentrifuge per run. Other published methods for exosome isolation including various commercial exosome-extraction kits (Total exosome isolation Kit, Thermo Fisher Scientific, TF; or miRCURY Exosome Isolation Kit, Exiqon, EX) were tested in our lab and generated exosome preprations of inferior quality with broader particle size range and increased contamination tested by Nanosight analysis.

There were also challenges with handling of pelleted exosomes. Exosomes are nanoparticles of 50-150nm in diameter hence invisible to eye after pelleting inside the ultracentrifuge tubes. The removal of supernatant and resuspension of exosomes is a delicate skill that requires extreme caution to ensure consistency. Further, it is known

134 that repeated thaw and refreezing of exosomal preparations can affect exosomal quality (Muller et al., 2014), and this was minimised as much as possible by aliquoting of final preparations into separate tubes for different uses. Additionally, exosomes tend to clump and form aggregates upon freezing, therefore it was important after defrosting on ice to vortex adequately to allow disaggregation of clumps and even loading of samples. Despite these measures variable results were seen between technical repeats of protein analysis. Sonication of samples prior to loading may be a better way to ensure homogeneity of dot blots (Hosseini-Beheshti et al., 2012).

135

Chapter 5

The role of circulating exosomes in HER signalling in colorectal cancer patients undergoing systemic treatment

136

5.1 Introduction Colorectal cancer is the third commonly diagnosed malignancy and one of the most frequent causes of cancer death worldwide (Ferlay et al., 2015). Despite advances in screening and diagnosis, treatment for patients with metastatic colorectal cancer remains poor, with 5-year survival rates of 13% (Siegel et al, 2012). The mainstay treatment for patients with metastatic colorectal cancer are combination chemotherapy involving 5-FU, oxaliplatin and irinotecan, with biologically targeted therapies such as anti-EGFR or anti-VEGF inhibitors. Unfortunately responses to all therapies are variable and of limited duration, due to intrinsic or acquired resistance mechanisms described in Chapter 1.8.

Currently there are few existing biomarkers for the diagnosis and monitoring of patients with colorectal cancer through treatment. The use of serum biomarkers such as CEA and CA19-9 are limited by their low sensitivity and specificity (Ludwig & Weinstein, 2005). Imaging modalities such as Magnetic Resonance Imaging (MRI) and Computed Tomography (CT) fail to detect early tumour dissemination or micro-metastases. Although colonoscopy is still the most effective method to diagnose early stage colorectal cancer, this approach has poor patient compliance and does not provide real-time monitoring of tumour progression and therapeutic response, especially in patients with metastatic disease (Sun et al., 2008). Therefore, it is necessary to identify better bio-markers that can be used for the early diagnosis, detection of recurrence, and monitoring of response through treatment. The advent of liquid biopsy as an alternate, quick, convenient, and minimally invasive method has shown tremendous potential to help clinicians achieve early diagnosis of cancer and guide decision- making during the course of treatment. It is expected to be an informative or easily accessible tool to provide comprehensive information of tumours beyond the invasive tissue biopsies.

Additionally, the ability to monitor HER signalling and resistance to anti-EGFR therapy longitudinally and non-invasively in clinical setting will allow better monitoring of treatment response and stratification of patients to personalised anti-cancer therapy. There is an urgent need to develop reliable tools for patient stratification to ensure only those likely to benefit are given a specific treatment. Predictive biomarkers to drive personalised cancer treatment and increase therapeutic index and overall survival are potentially of great benefit. The proteomic and genomic biomarkers in tumour-derived exosomes coupled with established methodologies for isolating exosomes from body fluids has generated much promise in using exosomes as circulating biomarker.

137

Though the analysis of exosomes for cancer biomarkers has revealed many possible candidates, none has demonstrated enough promise to be implemented clinically. It has been suggested tumour exosomes and exosomal proteins correlated with tumour grade and disease progression in colorectal cancer, and EGFR signaling via exosomes may contribute to cancer field effect and priming of metastatic niche (Higginbotham et al., 2011). microRNAs are a group of non-protein encoding RNAs that are 19-25nt in length that exert important regulatory role of cell activity at the transcriptional level. miRNAs have been shown to be promising diagnostic, prognostic and predictive biomarkers, based on their ability to distinguish normal vs malignant phenotypes, and their stability in comparison to proteins and other nucleic acids in circulation and fixed tissues. Circulating exosomal miRNAs are increasingly attractive compared to free miRNAs as they are secreted in abundance from malignant tissue compared to normal tissue, and shown to be representative of parental tumour in miRNA profile, as well as been highly protected from degradation through encapsulation.

Abundant evidence indicates that the de-regulation of miRNA expression and activity via genomic alterations, miRNA gene methylation, aberrant transcription or processing are intimately involved in cancer initiation, maintenance and progression (Ling et al., 2011). Beyond the functions they exert in the cells that produce them, miRNAs can be secreted and transferred to other cells enclosed inside extracellular vesicles including exosomes, making them potentially useful biomarkers. Table 5.1 shows a list of microRNAs that have been implicated in HER signalling.

138

Table 5.1: microRNAs that sustain or repress EGFR signalling. Adapted from (Gomez et al., 2013).

miRNA MiR Regulator Targets

miR-21 EGFR, HER2, c- PTEN, SPRTY, PDCD4 MET, AP-1

Let-7 EGFR, c-myc, LIN28 Ras, c-myc

miR-7 EGFR EGFR, IRS1, IRS2, RAF-1

miR-34 p53 c-MET

miR-221/222 EGFR, c-MET PTEN, Apaf-1

miR-30b/c EGFR, c-MET BIM

miR-125a/b HER2, HER3 MAPK, PI3K

miR-149 HER3 HER3

Of particular interest, miR-21 is one of the miRNAs that are frequently overexpressed in colorectal cancer, and a well-known onco-miRNA implicated in several other cancers. Studies have shown an association between elevated levels of miR-21 and downregulation of tumour suppressor genes including phosphatase and tensin homlog (PTEN), programmed cell death 4 (PDCD4), tissue inhibitor of metalloproteinase 3 (TIMP3) and RHOB (Peacock et al., 2014; Asangani et al., 2008a; XIONG et al., 2013; Sayed et al., 2008; Meng et al., 2007). miRNA-21 expression has been shown to positively associate with the diagnosis, staging, and prognosis of colorectal cancer (Schetter et al., 2008; Shibuya et al., 2010; Toiyama et al., 2013). It has also been shown that high levels of tumoral miR-21 expression are associated with poor response to chemotherapy in colorectal cancer patients (Li et al., 2016). Figure 5.1 shows signaling pathways that are been modulated by miR-21.

139

Figure 5.1. Signaling pathways modulated by miR-21. Reproduced from (Gomez et al., 2013).

The role of miR-30c in colorectal cancer is less clear. Studies have shown its involvement in many malignancies acting as tumour suppressor or tumour promoter, by targeting the expression of KRAS, UBC9, MTA1 and HMBOX1. Previous studies demonstrated aberrant expression of miR-30c was related to progression of multiple cancers including breast cancer, non-small cell lung cancer and gastric cancer (Mu & Su, 2012; Gu et al., 2013; Tanic et al., 2012). It is downregulated in squamous cell carcinoma of vulva, and suppressed prostate cancer survival by targeting the oncoprotein alternative splicing factor (ASF)/splicing factor 2 (SF2) (Huang et al., 2017). Although miR-30c has been widely regarded as a tumour suppressor in a variety of cancers, its functions in colorectal cancer remain unknown and the mechanisms underlying the tumour suppressive role not well described.

Overall, exosomal analysis of HER proteins and downstream signalling molecules, HER dimers, S100A8/A9 and immunosuppressive miRNAs all hold promise as feasible candidate biomarkers and may be important clinically in the management of colorectal cancer. Further investigation of the dynamic expressional profile of exosomal contents throughout tumour development and treatment, combined with improved isolation

140 methods, are needed before the clinical implementation of exosomes as biomarker for either diagnosis or prognosis of colorectal cancer.

5.2 Aims 1. Optimisation of exosome isolation from patient blood.

2. Identification and characterisation of circulating exosomes by immunoblotting.

3. High throughput proteomic analysis of exosomes using reverse phase protein array (RPPA).

4. Exosomal miRNA analysis by digital droplet polymerase chain reaction (ddPCR).

5. Exosomal proteins and miRNAs as markers of HER signalling in colorectal cancer patients undergoing anti-cancer treatment.

141

5.3 Results

5.3.1 Colorectal cancer patient cohort and recruitment Patients with newly diagnosed, advanced colorectal cancer suitable for systemic treatment were identified and recruited at the UCLH oncology clinic (Table 5.2). Patients consented onto the UCL Biobank Study. Clinical information including basic patient demographics (age and sex), location of the primary tumour, cancer stage, RAS/RAF status, treatment intent and regime were collected and stored securely in locked cupboards within clinical research facility. Each patient was assigned an unique ID on recruitment (CRC01-12) and their details kept anonymous throughout the study in accordance with Good Clinical Practice guidelines.

Table 5.2: Colorectal cancer patients recruited from oncology clinic. Newly diagnosed patients suitable for systemic treatment were recruited for exosome study. *CRC01, *CRC03, *CRC04 were patients recruited after initiation of systemic treatment and their bloods were used mainly for assay optimisation at the initial stages.

Patient Sex Primary Stage RAS/RAF Treatment Treatment ID intent

CRC01* M colon T3N1M1 KRAS MT FOLFIRINOX Palliative

CRC02 F Colon T2N1M1 BRAF FOLFIRINOX Neoadjuvant pVal600Glu

CRC03* F Sigmoid T3N1M1 WT FOLFIRI Palliative CRC04* M Sigmoid T4N2M1 WT FOLFIRI + Palliative cetuximab CRC05 M Rectosigmoid T4N2M0 WT FOLFIRI + Neoadjuvant AZD8931 CRC06 F Rectal ypT4aN2M0 WT FOLFIRI + Adjuvant R2 excision Cetuximab CRC07 M Sigmoid pT4bN2M1 WT FOLFIRINOX Neoadjuvant #1 Raltrexed Oxaliplatin #2 CRC08 F Caecum T3N2M1 KRAS MT FOLFOX Neoadjuvant (liver) CRC09 F Rectal T3NM1 WT FOLFOX Palliative CRC10 M Colon M1 WT FOLFOX Palliative CRC11 F Colon M1 WT FOLFIRI Palliative CRC12 M Rectal T3N2M1 WT FOLFIRINOX Palliative

142

Bloods were collected for exosome isolation and analysed prior to first and second cycle of systemic treatment, and at 3 months since start of therapy. 7 of the 12 patients on recruitment were treatment-naïve about to commence on first line systemic therapy, 1 patient (CRC06) was post-surgery for adjuvant chemotherapy due to incomplete resection, and 3 (CRC01, CRC03, CRC04) were recruited after initiation of systemic treatment hence their bloods were used mainly for assay optimisation during the initial stages. No bloods were obtained for CRC02 at 3 months due to logistical difficulties.

A separate cohort of patients with newly diagnosed colorectal cancer of all stages referred from across the region for PET-CT imaging were identified and recruited at the UCLH nuclear medicine department (Table 5.3). Bloods were collected upon cannulation but prior to administration of radionucleotide contrast and imaging. Only single point pre-treatment bloods were collected for this group of patients (P1-10).

Table 5.3: Colorectal cancer patients recruited from nuclear medicine clinic. Newly diagnosed patients coming for PET-CT pre-treatment were recruited prior to any treatment.

Patient ID Sex Primary Stage

P1 M Colon M1

P2 M Rectal M1

P3 M Rectal N+M0

P4 M Colon M0

P5 F Caecal M1

P6 M Sigmoid M0 P7 M Sigmoid M1

P8 M Rectosigmoid N+M0

P9 M Colon N+M0

P10 F Colon N+M0

143

5.3.2 Optimisation of exosome isolation from patient blood Comparable to exosome isolation from cultured cells, exosomes were isolated from patient blood through differential ultracentrifugation (Figure 5.2). 500µl to 2000µl of serum or plasma were used to purify exosome pellets which were suspended in 50- 200µl of PBS, and stored at -80°C until further analysis. Particle concentration was analysed using Nanosight LM10-HS (Nanosight Ltd), and protein quantification done using microBCA protein assay as previously described. There was no significant difference in the quantity or quality of exosomes obtained from serum or plasma.

Figure 5.2: Exosome isolation from patient blood. 10-15ml of clotted blood were taken in SST tubes and left standing at room temperature for 30-60min prior to serum separation. Serum were stored as 1ml aliqouts in -80°C or underwent differential centrifugation for exosome isolation.

5.3.3. Exosomal HER protein content in serum-derived exosomes Immunoblotting for exosomal surface markers CD63 and/or ALIX were used for detection and quantification of serum-derived exosomes. Immunoblotting for EpCam and CD45 delineate exosome population derived from epithelial cells and immune cells,

144 respectively. Primary antibodies used for dot blot analyses were tested for their specificity through either siRNA knockdown against protein of interest or running full- length western blots prior to use.

Using dot blot analysis, serum-derived exosomal proteins were examined. Figure 5.3 shows heterogeneous production of serum-derived exosomes from four different patients with metastatic colorectal cancer. There is differential staining against CD63, which accounts for all exosomes isolated. When adjusted to CD63, the relative proportion of epithelial cell-derived (EpCam) vs immune-cell-derived (CD45) exosome is similar among the 4 patients. The exosomal HER protein content was also heterogeneous. Of interest, CRC02, the patient who was treatment naïve had significantly more HER protein per CD63 compared to patients who were already in the middle of their treatment.

Figure 5.3: Heterogeneous production of exosomes among four metastatic colorectal cancer patients. CRC02 was treatment-naïve whereas the other three patients were in the middle of their systemic treatment. Exosomes were isolated from 2ml of serum from each patient and suspended in 400µl of PBS. (A) 5µl/dot was loaded onto nitrocellulose membrane and incubated against protein of interest. (B, C) Relative protein concentration was plotted after adjusting to CD63 for each sample. Staining against EpCam and CD45 delineate exosome populations derived from epithelial cells and immune cells, respectively.

145

Figure 5.4 shows dot blot analyses of exosomal proteins from two newly diagnosed colorectal cancer patients. Exosomes were isolated simultaneously from 2mls of serum from each patient. Though P5, patient with metastatic disease at diagnosis, had fewer total CD63 compared to P10, a patient with localised disease, P5 had higher concentration of HER proteins when adjusted to CD63 compared to P10.

Figure 5.4: Dot blot analysis of exosomes derived from 2 new colorectal cancer patients: P5 with metastatic and P10 with localised colorectal cancer. (A) Exosomes were isolated from 2ml of serum and suspended in 100µl of PBS. 5µl/dot was loaded onto nitrocellulose membrane and stained with antibodies of interest. (B) Densitometric calculations of dot blots after proteins adjusted to CD63 and relative protein concentration plotted.

Unfortunately, a similar trend was not observed when a group of 8 patients were analysed, 4 of which had metastatic and the other 4 had localised colorectal cancer (Figure 5.5).

146

Figure 5.5: Dot blot analysis of exosomes derived from 8 newly diagnosed colorectal cancer patients. Comparison between 4 patients with localised disease (Local) vs 4 patients with metastatic disease (Met). Exosomes quality was examined by Nanosight analysis. 5µl/dot was loaded. Exosomal protein contents were adjusted to CD63 and relative protein concentration plotted.

5.3.4. High throughput proteomic analysis of exosomes using reverse phase protein array (RPPA) One limitation of proteomic analysis of exosomes using dot blots is the small quantify of proteins obtained from either cell cultures in vitro or limited amount of clinical samples in vivo. RPPA is an established platform for high throughput quantitative analysis of protein content using small volumes of cell lysates and clinical samples. It involves using nitrocellulose coated slides and a similar working protocol to dot blot. Rather than using secondary rabbit or mouse antibodies conjugated to HRP, RPPA employs secondary antibodies directly conjugated to a fluorophore, which is read as fluorescent signal correlating to the amount of protein present. We performed the quantification of exosomal protein expression using a platform based on RPPA in collaboration with Professor Neil Carrighar and Dr Kenneth McCleod from Edinburgh University. They optimised a list of more than 200 primary antibodies against epitopes of interest.

147

In the first pilot study exosome samples from DLD1 (KRAS WT) were tested under native, lysed, reduced or denatured states, or a combination of above, against eighteen different primary antibodies, and found native proteins performed better overall. Signal analysis from Fastgreen stain was found to be a good indicator of protein loading. In the second study, exosomes from cell cultures were compared with exosomes from patient serum, and signal to background noise ratio was seen to be significantly lower in exosomes derived from patient serum compared to those derived from cell line. In a third study, nine primary antibodies of our interest were tested using different blocking and buffering conditions we found that 5% BSA was the best buffer for most of them. In addition, cross-adsorbed secondary antibodies were deployed and sample loading doubled which enabled good signal to noise ratio for at least six of the nine antibodies tested. In the fourth pilot study, 12 samples of patient serum-derived exosomes were tested as shown on Table 5.4.

Table 5.4: Patient sample cohort for RPPA pilot study 4. RPPA analysis of 12 exosome samples derived from 9 patients with newly diagnosed colorectal cancer.

Sample Patient Primary Stage Treatment Cycle Number ID

1 + 2 CRC02 Colon T2N1M1 FOLFIRINOX Pre #1 Pre #2

3 + 4 CRC05 Rectosigmoid T4N2M0 FOLFIRI + Pre #1 AZD8931 Pre #2

5 + 6 CRC07 Sigmoid pT4bN2M1 FOLFIRINOX Pre #1 Pre #2

7 P1 Colon M1 Pre-treat

8 P2 Rectal M1 Pre-treat

9 P3 Rectal N+M0 Pre-treat

10 P4 Colon M0 Pre-treat

11 P5 Caecal M1 Pre-treat

12 P6 Sigmoid M0 Pre-treat

148

In this study, 30 primary antibodies of interest were tested as shown in Figure 5.6. 4 dilution series of protein sample were printed with 2 replicates for each sample. Rabbit and mouse secondary cross-adsorbed antibodies were stained alone for subtraction of background signal. Fastgreen stain was used to normalise each specific antibody- protein signal to total protein across the sample series. Upon staining with dyes like Fast Green FCF, signal obtained reflect the total amount of protein per spot immobilised on each slide per print run.

Figure 5.6: Summary of RPPA output for pilot study 4. 30 primary antibodies against proteins of interest were analysed for 12 samples (indicated by 12 rows per antibody stained) using the array. 4 dilution series of protein sample were printed (left to right) with 2 replicates (left to right) for each sample. Rabbit and mouse secondary cross-adsorbed antibodies were printed to indicate background signal obtained.

An example of microarray layout is shown in Figure 5.7. Eight different protein samples were printed across different rows. Two replicates of four dilution series for each sample at 100%, 75%, 50% and 25% were printed per row with clear reduction of fluorescent signals, from left to right.

149

Figure 5.7: Sample microarray layout. 4 dilution series of protein sample were printed (left to right) with 2 replicates (left to right) for each sample.

The performance of tested primary antibodies is summarised in Table 5.5. Overall there was good signal-to-noise above background values for 14 of the antibodies, lower but still above background for another 7 antibodies, and 9 antibodies gave signals close to background levels.

150

Table 5.5: Antibody performance for pilot study 4. 30 primary antibodies were selected and tested for this study including 10 (highlighted blue) provided by our lab.

Target Supplier Cat # Species Block Comments MEK1/2 Cell Signaling Technologies 9122 rabbit BSA close to background MEK1/2 P Ser217/221 Cell Signaling Technologies 9154 rabbit BSA close to background Stat1 P Tyr701 Cell Signaling Technologies 9171 rabbit BSA close to background Stat5 Invitrogen (Biosource) 44-368G rabbit BSA Well above background Stat5 P Tyr694 Cell Signaling Technologies 9351 rabbit BSA close to background Akt Cell Signaling Technologies 9272 rabbit BSA close to background Akt P Ser473 Cell Signaling Technologies 4060 rabbit BSA above background p44/42 MAPK (ERK1/2) Cell Signaling Technologies 9102 rabbit BSA above background p44/42 MAPK (ERK1/2) P Thr202/Thr185,Tyr204/Tyr187Cell Signaling Technologies 4370 rabbit BSA Well above background Bcl-2 Epitomics 1017-1 rabbit BSA above background Bcl-x Epitomics 1018 rabbit BSA Well above background Bid Epitomics 1008 rabbit BSA close to background p90 S6 kinase (Rsk1-3) Santa Cruz sc-231 rabbit BSA close to background Met Cell Signaling Technologies 4560 rabbit BSA Well above background Met P Tyr1349 Signalway 11238 rabbit BSA Well above background Her1 Cell Signaling Technologies d38b1 rabbit BSA Well above background Her2 Cell Signaling Technologies 29d8 rabbit BSA Well above background Her3 Cell Signaling Technologies 1b2e rabbit BSA Well above background pHer1 tyr1148 Cell Signaling Technologies rabbit BSA Well above background pHer2 tyr1248 Cell Signaling Technologies rabbit BSA above background pHer3 tyr1289 Cell Signaling Technologies rabbit BSA above background calnexin Cell Signaling Technologies rabbit BSA above background * (negative control) Stat3 Cell Signaling Technologies 1264 rabbit BSA close to background PTEN Cell Signaling Technologies 9552 rabbit BSA Well above background PTEN P Cell Signaling Technologies 9554 rabbit BSA close to background Stat1 Cell Signaling Technologies 9176 mouse BSA Well above background Stat3 P Tyr705 Cell Signaling Technologies 9138 mouse BSA Well above background CD63 genetex gtx28219 mouse BSA Well above background EpCam sigma sab4700423 mouse BSA Well above background S100 a8/9 homemade mouse milk above background

Analysis of results across 12 samples show signal for exosome marker CD63 follows similar trend overall to EpCam which correlates to exosomes of epithelial origin. There seem to be a correlation between CD63 concentration and exosome number analysed using Nanosight in only 5 of the 12 samples. Unfortunately 2 of the 12 samples, both from a single patient, did not generate significant signal for any of the proteins tested. One possible explanation for this failure maybe that as exosomes were suspended in native states in PBS and often tend to form aggregates, loading sample of such small volume (picolitre) during the array may miss these particles if sample was not well dispersed prior to pipetting. Calnexin, a good negative control for cell-derived exosomes, it was not a good negative control for patient-derived exosomes probably due to cross-reactivity to other non-specific proteins. HER proteins showed good signal

151 above the background in most of the samples, as well as other proteins involved in downstream signalling pathways such as c-Met and ERK.

Following on from above, a final optimisation study was carried out to further reduce background signal and optimise dot morphology. Exosomes extracted from DLD1 (KRAS WT) was tested against 16 primary antibodies of interest under 6 different printing conditions. Blocking with 1% tween showed the largest decrease in exosome clumping and promoted the best dot morphology (Figure 5.8).

Figure 5.8: 1% Tween is the best blocking condition to improve dot morphology and reduce exosome clumping. Calnexin and S100 were used as negative control as they are absent from DLD1 secreted exosomes, showing very low background signals. Mouse antibodies were very clean under all conditions, whereas Rabbit antibodies gave higher background signals. 1% tween gave the best spot morphology, allowing the exosomes to dis- aggregate across the whole feature area. The minimum sample volume required per print slide (16 antibodies including secondary mouse and rabbit controls) was 20µl.

152

5.3.5. Analysis of miRNAs in circulating exosomes Exosomes from colorectal cancer cells or patient blood were isolated using differential ultracentrifugation previously described, and stored as dry pellet in -80°C until ready for RNA extraction. Quantification of exosomal RNA by standard bioanalyser method was challenged by small amount of RNA available. Therefore normalisation of RNA was done using housekeeping gene 18S ribosomal RNA (rRNA) which was quantitatively analysed using ddPCR. Exosomal miRNAs of interest were subsequently analysed using ddPCR against specific probes previously optimised by Dr. Giovanna Alfano. Figure 5.9 illustrates the workflow of exosomal miRNA analysis and the basic principle of ddPCR.

Figure 5.9: Exosomal miRNA analysis workflow and droplet digital PCR (ddPCR).

An example of expression levels of circulating exosomal miR-21-5p and miR-30c-5p transcripts as analysed using ddPCR in CRC10 through treatment is shown in Figure 5.10. After amplification, PCR products were read in the droplet reader and data analysed using the QuantaSoft™ Software (Bio-Rad). The thresholds for detection (green line) were manually set based on results obtained from negative control wells. Blue dots represents PCR positive droplets obtained and green line shows the

153 threshold above which positive droplets were analysed. 18S ribosomal RNA was used as control and normaliser for quantification of specific miRNA of interest, and is clearly present in abundant quantities in this patient. At a glance, the figure shows a decrease of miR-21-5p and miR-30c-5p transcripts through treatment, in patient CRC10. However after normalisation against 18S a different trend of change was observed as shown in section 5.3.6 (Figure 5.15).

Figure 5.10: Expression levels of miR-21-5p and miR-30c-5p transcripts in circulating exosomes were analysed using droplets digital PCR (ddPCR) assays. For each patient quantification of miRNAs was performed pre-cycle1 (Pre#1), pre-cycle2 (Pre#2) and at 3 months (3m) post treatment. Blue dots represents PCR positive droplets obtained and green line shows the threshold above which positive droplets are analysed.

154

5.3.6. Exosomal proteins and miRNAs involved in HER signalling as predictive biomarkers in colorectal cancer patients undergoing treatment Serial bloods from 7 patients with advanced colorectal cancer undergone first-line systemic therapy recruited from the UCLH oncology clinic were analysed for exosomal protein markers CD63, ALIX, EpCam, HER receptors, c-Met and S100A9; as well as miR-21-5p and miR-30c-5p. Serum CEA level taken from hospital biochemistry testing where available and radiological response at 3 and 6 months measured by contrast CT chest/abdomen/pelvis were also collated, and data presented for each patient as shown in Figures 5.10-5.16.

Table 5.6: Cohort of treatment-naive advanced colorectal cancer patients recruited for exosome biomarker study. SD: stable disease. PR: partial response. PD: progressive disease.

Patient Sex Stage RAS/RAF Treatment Radiological Radiological ID response at 3m response at 6m

CRC05 M T4N2M0 WT FOLFIRI + SD SD AZD8931

CRC07 M T4N2M1 WT FOLFIRINOX #1 PR PR Raltrexed Oxali #2

CRC08 F T3N2M1 KRAS MT FOLFOX PR PR

CRC09 F T3N1M1 WT FOLFOX N/A N/A

CRC10 M T3N2M1 WT FOLFOX PR PD

CRC11 F T4N1M1 WT FOLFIRI + PR PD Panitumumab from #5 CRC12 M T3N2M1 WT FOLFIRINOX SD PD

CRC05 was treated with first-line FOLRIFI and AZD8931, and bloods were collected prior to first, second and third cycle of treatment, as well as at 3 months (Figure 5.10). 1ml of serum per time point was used to isolate exosomes at one time, and resuspended in 50ul of PBS for proteins analysis, or directly used for RNA extraction and miRNA analysis. 5ul of exosomes in PBS was diluted in 995ul of PBS for nanosight analysis of particle concentration and quality. Positive staining against exosomal surface markers CD63, ALIX and EpCam on dot blotting further confirmed successful

155 exosome isolation from each preparation (Figure 5.11A). Densitometry calculations of relative protein changes through the course of the treatment was plotted, after normalisation with CD63 or ALIX as loading control (Figure 5.11E & F). When adjusted to ALIX, we observed a reduction of exosomal HER1, HER2, HER3, c-MET and S100A9 after 1 cycle of therapy. After 3m, there was continued reduction of HER1, HER2 and HER3, but c-MET and S100A9 increased back to near baseline levels. Analysis of miRNA after normalisation to S18 showed increased miR-21-5p after 1 cycle of treatment which reduced at 3m but the level still higher than baseline (Figure 5.11B). miR-30c-5p also increased after 1 cycle of treatment but reduced to half the baseline amount at 3m (Figure 5.11C). Interestingly this patient was a non-secretor of CEA the level of which remained non-elevated throughout (Figure 5.11D). Radiologically the patient had stable disease after 3m of treatment compared to baseline.

Figure 5.11: Exosomal miRNA, protein and circulating CEA in CRC05. This patient was treated with first-line FOLFIRI and AZD8931. Bloods were collected prior to first, second, and third cycle of treatment, as well as at 3 months after initiation of therapy. 5µl of exosomes suspended in PBS was loaded per dot, and relative quantification of proteins adjusted to CD63 or ALIX. (A) Dot blot of exosomal proteins through treatment. (B) Fold change of miR-21-5p compared to baseline. (C) Fold change of miR-30c-5p compared to baseline. (D) CEA levels through treatment. (E & F) Relative fold change of exosomal proteins compared to baseline, after normalisation to CD63 or ALIX.

156

CRC07 was treated with FOLFIRINOX for one cycle followed by a change to oxaliplatin and raltrexed from cycle 2 due to treatment toxicity, and bloods were collected prior to first, second cycle of treatment and at 3 months (Figure 5.12). There were reduction in exosomal HER proteins and S100A9 after cycle 1 and 3m of treatment. There was also reduction of miR-21-5p after cycle 1 and at 3m. miR-30c-5p increased after cycle 1 but reduced to below baseline at 3m. This patient had elevated CEA at baseline which reduced to near normal range after cycle 1 and slightly increased at 3m. Radiologically the patient had partial response after 3m of treatment compared to baseline.

Figure 5.12: Exosomal miRNA, protein and circulating CEA in CRC07. This patient was treated with FOLFIRINOX for one cycle followed by a change to oxaliplatin and raltrexed from cycle 2 due to treatment toxicity following cycle one. Bloods were collected prior to first, second cycle of treatment, and at 3 months after initiation of therapy. 5µl of exosomes suspended in PBS was loaded per dot, and relative quantification of proteins adjusted to CD63 or ALIX. Data for both protein and miRNA plotted as fold changes compared to pre-treatment levels.

CRC08 was treated with first-line FOLFOX, and bloods were collected prior to first, second cycle of treatment and at 3 months (Figure 5.13). There was initial increase followed by reduction in exosomal HER2, and reduction in exosomal HER3 at 3m. miR- 21-5p level increased after cycle 1 but reduced to half the baseline at 3m. miR-30c-5p level was unchanged after cycle 1 but became undetectable at 3m. CEA as not tested in this patient and radiologically the patient had partial response at 3 months.

157

Figure 5.13: Exosomal miRNA and protein in CRC08. This patient was treated with first-line FOLFOX. Bloods were collected prior to first, second cycle of treatment, and at 3 months after initiation of therapy.

CRC09 was treated with first-line FOLFOX, and bloods were collected prior to first, second cycle of treatment and at 3 months (Figure 5.14). There was initial reduction followed by increase in exosomal HER3 and c-MET, and increased S100A9 at 3m. miR-21-5p level increased after cycle 1 but reduced to still above baseline at 3m. miR- 30c-5p level increased after cycle 1 but reduced to below baseline at 3m. CEA as not tested and unfortunately this patient was admitted to hospital and died of sepsis prior 3m restaging scan.

Figure 5.14: Exosomal miRNA and protein in CRC09. This patient was treated with first-line FOLFOX. Bloods were collected prior to first, second cycle of treatment, and at 3 months after initiation of therapy.

158

CRC10 was treated with first-line FOLFOX, and bloods were collected prior to first, second cycle of treatment and at 3 months (Figure 5.15). miR-21-5p level increased after cycle 1 but reduced to below baseline at 3m. miR-30c-5p level was unchanged after cycle 1 but reduced at 3m. CEA initially increased after cycle 1 but reduced to below baseline at 3m. Radiologically the patient had partial response at 3 months.

Figure 5.15: Exosomal miRNA, protein and circulating CEA in CRC10. This patient was treated with first-line FOLFOX. Bloods were collected prior to first, second cycle of treatment, and at 3 months after initiation of therapy.

CRC11 was treated with first-line FOLFIRI with addition of panitumumab from cycle 5 (2.5 months after initiation of therapy), and bloods were collected prior to first, second cycle of treatment and at 3 months (Figure 5.16). There was increase in exosomal HER1, HER2, HER3, c-MET and S100A9 after 3 months of treatment. Both miR-21- 5p and miR-30c-5p levels reduced after cycle 1 and increased slightly to still below baseline at 3m. CEA level reduced throughout this treatment period. Radiologically the patient had partial response at 3 months.

159

Figure 5.16: Exosomal miRNA, protein and circulating CEA in CRC11. This patient was treated with first-line FOLFIRI with addition of panitumumab from cycle 5 (2.5 months after initiation of treatment). Bloods were collected prior to first, second cycle of treatment, and at 3 months after initiation of therapy.

CRC12 was treated with first-line FOLFIRINOX, and bloods were collected prior to first, second cycle of treatment and at 3 months (Figure 5.17). There was increase in exosomal HER3 after 1 cycle and at 3m of treatment. miR-21-5p increased after 1 cycle but reduced to below baseline at 3m. miR-30c-5p level steadily reduced after cycle 1 and at 3m. CEA level was unchanged at 3m. Radiologically the patient had stable disease at 3 months.

Figure 5.17: Exosomal miRNA, protein and circulating CEA in CRC12. This patient was treated with first-line FOLFIRINOX. Bloods were collected prior to first, second cycle of treatment, and at 3 months after initiation of therapy.

160

CEA is an existing and only cancer marker used in clinical setting for patients with colorectal cancer, where it is useful for assessing response to therapy in those patients who are secretors of CEA. It is a poor diagnostic marker due to poor specificity as elevation in levels can be observed in other malignant and non-malignant conditions.

Overall, exosomal proteins normalised by CD63 strongly correlated with exosomal proteins normalised by ALIX. After adjustment by Bonferroni correction for multiplicity, statistically significant positive correlations were observed in changes of exosomal proteins at 3 months between HER1 and HER2, HER1 and HER3 (Figure 5.18). When data were explored using a less stringent criteria with p < 0.05, there was also positive correlations observed between HER2 and HER3 (p=0.012). Only 3 of 7 patients had documented elevated CEA levels pre and post-treatment hence the correlation of exosomal markers with circulating CEA could not be confidently analysed.

Figure 5.18. Spearman’s correlation of changes in exosomal protein signals from pre- treatment to 3 months. The signals were normalised by the intensity of ALIX. Significance level was adjusted by Bonferroni correction.

161

Further analysis showed patients who had radiological progression of disease at 6 months had statistically higher increase of exosomal HER3 at 3 months compared to baseline compared to patients who had response to therapy (Pearson’s correlation=0.82, p=0.046, Figure 5.19).

Figure 5.19. Boxplot showing changes in exosomal HER3 at 3 months for each category of patients according to radiological response at 6 months. Boxplot shows the median (the middle thick line), interquartile range (IQR, the difference between 75th and 25th percentile, the box), and 1.5x IQR above and below the box (the vertical lines). Pearson’s product-moment correlation=0.82, p=0.046.

Additionally, patients who responded to treatment at 3 months had greater reduction in exosomal miR21 compared to patients who had stable disease, though statistically non-significant (Spearman’s correlation=0.58, p=0.23, Figure 5.20).

162

Figure 5.20. Boxplot showing changes in miR21 at 3 months for each category of patients according to radiological response at 3 months. Boxplot shows the median (the middle thick line), interquartile range (IQR, the difference between 75th and 25th percentile, the box), and 1.5x IQR above and below the box (the vertical lines). Pearson’s product- moment correlation=0.58, p=0.23.

No significant correlation was found between changes in exosomal HER3 and miR30c at 3 months (Figure 5.21).

Figure 5.21. Scatter plot showing non significant correlation between change in HER3 and miR30c at 3 months. Pearson’s product-moment correlation=-0.34, p=0.46.

163

5.3.7. Effect of treatment with AZD8931 on plasma exosomal proteins in patients In Chapter 3, the effect of treatment with AZD8931 on HER activation and dimerisation in colorectal cancer cells was examined. Here the effect of AZD8931 on HER signalling and dimerisation in colorectal cancer patients is explored within the PANTHER trial.

Exosomes from the serial plasma of 13 patients (T01-T13) with metastatic colorectal cancer where they were treated with first line AZD8931 and FOLFIRI chemotherapy were extracted and analysed (Table 5.7).

Table 5.7. Characteristics and clinical outcomes of a cohort of patients with advanced colorectal cancer treated with first line AZD8931 and FOLFIRI.

Progression Radiological Trial Age at Radiological response Sex free survival response at 12 number diagnosis at 24 weeks (months) weeks

Progressive T01 Male 66 2.96 N/A Disease T02 Male 77 8.64 Partial Response Partial Response

T03 Male 61 10.22 Stable Disease Partial Response

T04 Male 47 5.55 Stable Disease Progressive Disease

Not Evaluable Not completed as off T05 Male 73 . Patient too ill treatment

T06 Male 65 8.97 Partial Response Partial Response

Not completed as off T07 Female 56 4.53 Stable Disease treatment

T08 Male 62 8.97 Stable Disease Stable Disease

T09 Female 62 8.71 Partial Response Partial Response

T10 Male 50 . Stable Disease Stable Disease

Not completed as off T11 Male 48 6.41 Stable Disease treatment

T12 Male 43 . N/A N/A

Not completed as off T13 Male 66 4.93 Stable Disease treatment

164

Bloods were obtained at 4 different time points: immediately pre-treatment, 8-10h post first dose of AZD8931, immediately pre-first full cycle of combination therapy (5-7 days post first AZD8931 dose), and 8-10h post first cycle. Nanosight analysis of exosome pellet proved good quality and quantity of exosomes obtained despite previous freeze and thaw of plasma, and prolonged storage at -80°C for over one year in some samples. Two haemolysed samples were identified and excluded from analysis.

Exosomal proteins were immunoblotted against HER proteins and S100A9, with CD63 and ALIX used as loading control for quantification (Figure 5.22). There was heterogeneous production of exosomes among different patients and longitudinally through treatment course for each patient.

Figure 5.22: Effect of AZD8931 with FOLFIRI on exosomal proteins in a cohort of metastatic colorectal cancer patients. 200-500µl of plasma were used to extract exosomes at 4 different time points per patient, before and after first dose of AZD8931; and 5-7 days later, before and after first cycle of chemotherapy combined with AZD8931. Exosome pellets were resuspended in 20-50µl of PBS. Exosome numbers and quality were examined using Nanosight analysis. 5µl/dot were used for protein analysis. There was insufficient plasma for dot blotting at T11.3 for all protein markers, and T13.3 for ALIX. There were no samples received for T12.3 or T12.4.

165

Exosomal HER1:HER3 and HER2:HER3 dimers were analysed using FRET/FLIM by Dr Gregory Weitsman using an optimised protocol. FRET efficiencies for each dimer pair was calculated for each patient at baseline and prior to first full cycle of treatment (5-7 days after the first dose of AZD8931), and the changes in FRET efficiencies were plotted.

Figure 5.23 illustrates the changes in exosomal HER2:HER3 FRET efficiency in these patients. No results were obtained for T10 and T13 due to analytical difficulties. No post-treatment bloods were collected for T12. Four patients demonstrated increased HER2:HER3 dimerisation in circulating exosomes after treatment with AZD8931 as shown by positive changes in FRET efficiency (T1, T4, T7, T11). Six patients demonstrated reduced HER2:HER3 dimerisation in circulating exosomes as shown by negative changes in FRET efficiency (T2, T3, T5, T6, T8, T9).

Changes in exosomal HER2:HER3 FRET efficiency (%) 20 15 10 5 0 -5 -10 -15 -20 T1 T2 T3 T4 T5 T6 T7 T8 T9 T10 T11 T12 T13

Figure 5.23: Changes in exosomal HER2:HER3 FRET efficiency in colorectal cancer patients 5-7 days after first dose of AZD8931. No results obtained for T10 and T13 due to analytical difficulties. There was no post-treatment bloods for T12. Positive changes in FRET efficiency indicate increased dimerisation whereas negative changes indicate reduced dimerisation.

Similarly, Figure 5.24 demonstrates the changes in exosomal HER1:HER3 FRET efficiency in these patients. No post-treatment bloods were collected for T12. Nine patients showed increased HER1:HER3 dimerisation in circulating exosomes after treatment with AZD8931 as shown by positive changes in FRET efficiency (T3, T4, T5, T6, T7, T9, T10, T11, T13). Three patients showed reduced HER1:HER3 dimerisation

166 in circulating exosomes as shown by negative changes in FRET efficiency (T1, T2, T8)).

Changes in exosomal HER1:HER3 FRET efficiency (%) 35 30 25 20 15 10 5 0 -5 -10 -15 T1 T2 T3 T4 T5 T6 T7 T8 T9 T10 T11 T12 T13

Figure 5.24: Changes in exosomal HER1:HER3 FRET efficiency in colorectal cancer patients 5-7 days after first dose of AZD8931. There was no post treatment bloods for T12.

Multivariate analysis using programme R followed by multiple correlation correction showed a significant negative correlation between changes in exosomal S100A9 and HER2:HER3 dimer quantity after first dose of AZD8931 (Spearman’s correlation = - 0.97, p=0.0012). There was trend toward positive correlation between HER2:HER3 dimer quantity after first dose of AZD8931 with radiological response at 12 weeks (Figure 5.25, Spearman’s correlation =0.71, p=0.11).

However, there was no significant correlation between pre-treatment exosomal HER proteins levels or HER2:HER3 or HER1:HER3 dimerisation levels with radiological response at 12 weeks (Figure 5.26).

167

Figure 5.25. Heatmap of Spearman’s correlation and pair plots between circulating exosomal protein changes after first dose of AZD8931 in patients with advanced colorectal cancer. The signals were normalised by the intensity of ALIX. Individual p values shown in the heatmap have not been corrected by multiplicity. Significance level after adjustment by Bonferroni correction is p < 0.002. Boxes with colour red indicate positive correlation and colour blue indicate negative correlation. There was significant negative correlation between changes in exosomal S100A9 and HER2:HER3 dimers. There was trend toward positive correlation between HER2:HER3 dimer changes with radiological response at 12 weeks.

168

Figure 5.26. Heatmap of Spearman’s correlation and pair plots between circulating exosomal proteins at baseline in patients with advanced colorectal cancer. The signals were normalised by the intensity of ALIX. Significance level after adjustment by Bonferroni correction is p < 0.002. None of the pair plots were statistically significant.

169

5.4 Discussion 5.4.1. Exosomes can be reliably isolated and characterised from blood of patients with colorectal cancer This chapter demonstrated exosomes can be purified from circulating blood, either serum or plasma of patients, via differential ultracentrifugation. Higher yields of exosomal protein were obtained when sample was diluted in 1:1 or 1:2 PBS to reduce viscosity at the initial stage, or by double the duration of the ultracentrifugation at 100,000g. 500µl of serum/plasma produced comparable number of exosomes to those produced from 8 plates of 15cm dish cells. Another 500µl was required to produce sufficient exosomes for miRNA analysis and FRET analysis of HER dimers. The quality of exosomes produced as analysed by Nanosight and immunoelectron microscopy was similar to those produced from cell culture, except for an increased tendency for background contamination. This is due to more heterogeneous components of serum and plasma compared to cell media. In the case of Nanosight analysis, this might have limitations in precisely capturing exosomes alone excluding noise created by lipoproteins, protein aggregates, and other biological vesicles from serum (Vestad et al., 2017). Another criticism of Nanosight technology is the evidence of operator handling bias that may influence the accuracy and reproducibility of measurements (Gerritzen et al., 2017). One could minimise background contamination by careful removal of supernatants at each stage of differential centrifugation, followed by gentle washing of the exosome pellet prior to the final ultracentrifugation step.

Serum-derived exosomes from colorectal cancer patients can be characterised by the presence of surface markers CD63 and ALIX, which can be used as loading control and to normalise during exosomal protein quantification. There was good correlation between exosomal proteins normalised by CD63 and ALIX for the cohort of clinic patient samples, but less so for the Panther patient samples. This may be due to differences in the quality of bloods collected and stored of these 2 cohorts. For the clinic patients, blood samples were collected by myself, serum isolated and stored at - 80°C within 2 hours until further exosome extraction. Whereas for the Panther patients, blood samples were collected by trial nurses primarily for pharmacokinetic and pharmacodynamics studies in the context of Phase I trial, and plasma had been thaw and refrozen a couple of time during processing before the residual volume sent to us for exosome study.

Overall, circulating exosome production varied greatly between patients with colorectal cancer and within the same patient through course of treatment, as quantified by

170

Nanosight analysis and dot blot experiments against loading control proteins. There was no direct correlation between cancer stage and exosome production.

5.4.2. RPPA (Reverse Phase Protein Array) as an alternative method for exosomal protein analysis Analysis of exosomal protein using conventional immunoblotting has been challenging due to small quantity of protein obtained using differential ultracentrifugation. Dot blot analysis of exosomal protein has proven to be more sensitive requiring only 0.5-1.5µg protein/dot against validated antibodies through either siRNA or shRNA knockdown. Nevertheless, dot blot can be demanding and labour-intensive when reproducibility needs to be tested by doing multiple runs, or comparison needs to be made across numerous samples.

RPPA is a platform that can potentially overcome this demand due to its ability to analyse multiple samples on a single array using picolitre range of protein sample per dot (Boellner & Becker, 2015; Hennessy et al., 2010; Tibes et al., 2006; Macleod et al., 2017). The technology of Reverse Phase Protein Array (RPPA), a miniaturized dot blot, was described in 2001 and is based on the simultaneous measurement of a single protein in multiple samples (Figure 1.6). RPPA has proven to be useful for efficient tissue protein quantification and for analysis of signalling cascades even on the basis of very low amount of protein sample, enabling quantification of biomarkers in tumours at early time points (Malinowsky et al., 2010; Wilson et al., 2010). RPPA is increasingly being used to determine deregulated signalling networks in cancer tissues. RPPA is one of the three most common methods to validate mass spectrometry discovered biomarkers (Honda et al., 2013). RPPA has also been used to analyse heterogeneity of protein levels within a tumour and in primary tumour and lymph node metastases of the same patient (Silvestri et al., 2013; Malinowsky et al., 2012).

The basic workflow of RPPA involve 5 steps: sample preparation, antibody validation, spotting, signal detection and data analysis (Figure 5.27). Reliability of RPPA results largely depends on the quality of antibodies used. Usually proteins are immobilised on nitrocellulose-coated glass using solid pin-based contact printing, although other printing technologies (piezoelectric or inkjet spotting) and substrates have been described (Akbani et al., 2014; Ressine et al., 2007). In addition, the Zeptosens RPPA platform based on Planar Wave Guide (PWG) technology permits highly sensitive quantitative protein profiling. The immobilised proteins are detected with antibodies whose specificity for the antigen of interest has been validated as above. The presence

171 of antigen-antibody binding can be detected via near-infrared dyes, chemiluminecence, chromogenic reporter, or planter waveguide technologies. The generated signal is proportional to the antibody bound to the spotted proteins and may be quantified to estimate relative protein concentration. It is possible to detect proteins in the fg/mL range with linearity in the pg/mL range (Rapkiewicz et al., 2007).

Figure 5.27: Reverse Phase Protein Array Workflow. A) Printing of samples in a neatly organised array format onto nitrocellulose-coated glass slides; B&C) Incubation with a target-specific primary antibody to detect protein of interest followed by signal detection of primary antibody through fluorescence, chemiluminescence or colorimetric methods; D) Target intensities are quantified after scanning and analysing signal intensities of individual spots; E) Data processing and quality control can be performed with R-Package RPPanalyser. Adapted from (Wachter et al., 2015).

Dilution series can be carried out to confirm linearity of the signal output and protein stain performed to calculate protein loading. Results shown in section 5.3.4 have demonstrated potential feasibility of using RPPA in detecting native proteins in both cell-derived and patient-derived exosomes. Further optimisation is needed to demonstrate reproducibility of results and to reduce background signal (secondary antibody alone control). The majority of work done so far using RPPA have been on cell lysates or tumour FFPE or frozen tissues. Camacho et al showed promise in the use of RPPA in analysing exosomal proteins in patients with glioblastoma (Camacho et al., 2013). Chen et al used RPPA to analyse exosomal proteins in primary and metastatic melanoma cell lines (Chen et al., 2018). We have great hopes that this method could be put into more broad use in future research of exosomes in cancer patients.

172

5.4.3. Detection of circulating exosomal HER proteins in colorectal cancer patients Currently CEA (carcinoembryonic antigen) is the only circulating blood marker used in detection and monitoring of colorectal cancer patients in clinical setting. However CEA has low specificity and sensitivity for cancer diagnosis, is secreted in a variety of non- malignant conditions, and not elevated in level in a proportion of colorectal cancer patients (Duffy, 2001; Marwan G. Fakih, 2006). There is a clear need for improved biomarkers with higher specificity for cancer diagnosis and treatment response.

Results in Chapter 3 suggested HER dimer rewiring as a resistance mechanism against anti-EGFR therapy, and the potential role of monitoring HER3 and HER dimer in response to anti-HER therapies. Cellular HER expression and dimerisation can be quantified using methods described previously, but in the case of patients, HER expression is currently evaluated by performing a tissue biopsy that is highly invasive, and impractical for longitudinal monitoring of colorectal cancer patients with advanced stage and many with poor performance status. Therefore the ability to accurately quantify HER proteins and HER dimers in circulating exosomes would allow non- invasive monitoring through the use of liquid biopsies.

Results in this chapter showed HER proteins have been consistently detected from both colorectal cancer cell- and patient serum-derived exosomes. Using CD63 as the exosomal ‘loading control’, HER proteins were higher in a treatment-naïve patient compared to three other patients already in middle of their treatment, indicating their potential role in predicting treatment response. HER proteins were also seen to be relatively higher in a patient with metastatic colorectal cancer compared to another with localised cancer prior to any anti-cancer treatment, indicating their possible correlation with tumour burden. However, testing a group of four patients with localised vs another four with advanced colorectal cancer did not show significant correlation between cancer stage and exosomal HER protein levels.

Possible limitations could include small sample size and that the samples were processed in two batches due to limitation of space in the centrifuge. Limited available clinical material, particularly in the context of a trial, restricts the ability to generate replicate results. Another difficulty in interpreting and comparing results of exosomal proteins between different patients is the heterogeneity of exosome population as they are secreted from various cells rather than only colorectal cancer cells. Many confounding factors including patient demographics, co-morbidities, medications, and presence of any acute infectious or inflammatory conditions could all affect the levels

173 of analysed proteins. Therefore, the ability to distinguish tumour-specific exosomes by proteomic analysis of surface proteins (surfaceome profiling) would allow more stringent interpretation of the obtained results (Castillo et al., 2018).

5.4.4. The potential role of circulating exosomal proteins as predictive biomarker in colorectal cancer patients undergoing systemic treatment Data generated from a cohort of seven patients with advanced colorectal cancer undergoing first line systemic treatment showed statistically significant correlation between changes in exosomal HER1 and HER2, and HER1 and HER3 after 3 months. There was also trend towards positive correlation between HER2 and HER3.

Changes in exosomal HER3 at 3 months compared to pre-treatment negatively correlated with radiological response at 6 months: patient who had increased HER3 at 3 months developed progressive disease at 6 months despite treatment, whereas patients who had reduced HER3 at 3 months tended to continue responding to treatment at 6 months. It is worth noting only two out of the seven patients received anti-HER therapy, and all patients were treated with combination chemotherapy. Increased HER3 expression has been associated with many solid malignancies including colorectal cancer, and plays a critical role in HER2 mediated tumorigenesis (Beji et al., 2012). Activation of HER3 signalling has been implicated as an important cause of treatment failure to EGFR therapy (Vlacich & Coffey, 2011). Results in Chapter 3 showed feedback activation of HER3 and increased HER2:HER3 dimer formation could be a mechanism for reduced cetuximab efficacy in colorectal cancer cells. Yonesaka et al demonstrated secondary resistance to cetuximab in a subset of colorectal cancer patients is caused by upregulated circulating NRG that induced HER3 activation (Yonesaka et al., 2011). Dual targeting of HER3 and HER1 with MEHD7945A is capable of overcoming acquired resistance to cetuximab (Huang et al., 2013). Activation of HER3 signalling has also been shown to induce resistance to several chemotherapeutic agents including doxorubicin, docetaxel, 5-flurouracil and etoposide (Knuefermann et al., 2003; Bezler et al., 2012). Increased HER3 expression caused paclitaxel resistance in breast cancer cells via AKT-mediated upregulation of Survivin (Wang et al., 2010).

174

5.4.5. Effect of AZD8931 on circulating exosomal proteins in colorectal cancer patients Data from colorectal cancer patients treated with first line AZD8931 did not show any correlation between individual exosomal HER protein pairs, either at 8-10 hours or 5- 7 days post treatment. Changes in none of the exosomal proteins at these early stage predicted treatment response at 12 weeks. This is similar finding to the previous clinic cohort of patients, where we also did not observe any association in early exosomal protein changes (comparing pre-treatment cycle 2 to pre-treatment), with clinical outcomes. Kinetic studies of exosome release are not documented in the literature. For cell-line based work investigators usually choose a specific collection time that varies from a few hours to several days. An increasing release of exosomes during the first 24h has been observed in PC-3 cells (Llorente et al., 2007). Longer treatment duration is likely required for sustained activation of alternative HER signalling.

However interestingly, early changes in exosomal HER2:HER3 dimer quantity after AZD8931 treatment negatively correlated with changes in exosomal S100A9, and suggested radiological response at 12 weeks. Increased S100A9 positive myeloid- derived suppressor cells in the blood of non-small cell lung cancer patients with mutant EGFR treated with tyrosine kinase inhibitors was associated with shorter progression- free survival (Feng et al., 2018). In Chapter three we demonstrated HER2:HER3 dimer increase with 1hr of AZD8931 in sensitive cell line LIM1215 but not in resistance line DLD1. Furthermore, our analysis of the phase-3 MRC COIN trial showed in patients with advanced colorectal cancer who did not benefit from first line cetuximab in addition to oxaliplatin-based chemotherapy, a higher level of tumoral HER2:HER3 interaction measured using FRET/FLIM correlated with a better overall survival (Barber et al, unpublished work). One hypothesis is that that low affinity EGFR ligands (AREG and EREG) which are positively predictive of anti-EGFR treatment outcome, may induce prolonged and more widespread signalling, involving the formation of non-EGFR heterodimers such as HER2:HER3 (Freed et al., 2017; Komurasaki et al., 1997). Taken together this result suggests the role of HER2:HER3 dimer as potential predictive biomarker in colorectal cancer patients undergoing anti-HER therapy. A larger prospective trial is required to validate this finding.

In the PANTHER analysis, no correlation was observed between pre-treatment exosomal HER proteins or HER dimer quantity with radiological response at 12 weeks.

One potential limitation of validation of analysis from this set of data is, as mentioned previously, the suboptimal quality of blood used for exosome extraction, as they were

175 residual plasma samples following intended pharmacokinetic and pharmacodynamics studies. Plasma is potentially more challenging type of sample compared to serum for exosome extraction as it contains high levels of clotting factors (Panteleev et al., 2017). Nonetheless, the quality of exosomes extracted was validated using Nanosight and staining against surface markers CD63 and ALIX.

There are two advantages of studying exosomal proteins in this cohort of patients compared to the previous cohort from oncology clinic. Firstly patients are treated homogenously with AZD8931, and bloods were collected at the same time intervals relative to treatment for each patient. Secondly pairs of bloods were collected immediately pre- and post- treatment (8-10hrs apart) allowing examination of effect of drug treatment on circulating exosomes without presence of other confounding factors which would occur during a longer time interval. In theory, 10h is unlikely sufficient time interval to allow cellular protein changes to translate into changes in exosomes. Hence it was also important to look at treatment effect at longer time interval of 5-7 days to look for any more sustained effect of AZD8931 therapy on exosomal proteins.

5.4.6. Exosomal miRNA analysis by ddPCR and normalisation strategy The standard analysis of miRNA from exosomes involves exosome isolation, total RNA purification, cDNA conversion and miRNA detection. Many techniques have been described for recovery of exosomal miRNAs with varying yields including Northern blot, microarray chip, in-situ hybridisation and qRT-PCR (Deo et al., 2006; Lee et al., 1993; Li et al., 2009; Reza et al., 2018). In clinical settings, limited blood volumes often restrict exosome concentration and miRNA profiling. Finding a reliable and sensitive method for quantification of circulating miRNA is key to their use as potential cancer biomarkers.

Droplet digital PCR has been used for quantification of miRNA of interest and it is a recently developed alternative technique to qRT-PCR (Campomenosi et al., 2016). The principle is based on partitioning of the sample into thousands of micro-reactions of defined volume (Pinheiro et al., 2012). After the PCR reaction, the analysis of each droplet either positive or negative for the nucleic acid of interest, allows the estimation of the number of molecules in the reaction under the assumption of a Poisson distribution. Results are expressed as target copies per microlitre of reaction (Huggett et al., 2013). Also this technique is known to have increased precision and sensitivity in detecting low target copies, and is insensitive to potential PCR inhibitors (Rački et

176 al., 2014; Jarry et al., 2014). These are important points to take into consideration in studies with low input of clinical material and exosome concentration.

Conventionally, RNA concentration can be quantified using NanoDrop by measuring the absorbance at 260nm. Quantification and quality assessment of small RNA including miRNA fraction can be done using the ‘Small RNA assay’ (Agilent Technologies). The low sample input volume means the yield of RNA is insufficient to accurately measure by NanoDrop. One prerequisite for validating ddPCR data is proper normalisation with respect to stably expressed endogenous reference genes. However, genes that meet all of the criteria of a control gene for exosomal miRNAs have not yet been identified (Vandesompele et al., 2002; Andersen et al., 2004; Shimamoto et al., 2013). Different studies use different normalisation strategies to report miRNA expression, and this leads to ambiguous data interpretation and misleading biological interpretation of data. Reference genes such as small nuclear/nucleolar RNAs, have been used for miRNA quantification as they share similar properties such as RNA stability and size, and are abundantly expressed (Song et al., 2012; Roth et al., 2010). However though these transcripts display constant expression in single analyses, their expression levels can change under different experimental conditions (Mahdipour et al., 2015). Therefore different normalisers need to be established for different sample types, and a combination of several normalisers might be more helpful than a single universal normaliser (Schwarzenbach et al., 2015).

18S ribosomal RNA has been widely used as control for qRT-PCR analyses because of its invariant expression across tissues, cells, and experimental treatments (Kuchipudi et al., 2012). It is also highly expressed in most cell types. 18S rRNA has been used for miRNA normalisation in literature but nowhere does it suggest that it is the best endogenous marker (Schwarzenbach et al., 2015). Dr Alfano from our lab has reliably demonstrated abundant expression of 18S rRNA from exosomes extracted from lung cancer cell lines and blood of patients with lung, and head & neck cancers using ddPCR. We were not able to get amplification of other candidate miRNA normalisers including RNU6N and miR-16 from patient-derived exosomes, and it was difficult to identify a suitable miRNA that would not be affected by drug treatments. Based on these 18S rRNA was used as the normaliser for miRNA analysis of exosomes in this project, and has been shown to be consistently amplified in abundance in circulating exosomes of colorectal cancer patients.

177

5.4.7. Exosomal miRNAs as markers of HER signalling in colorectal cancer So far we have looked at biomarkers involved in HER signalling in terms of exosomal proteins and HER dimers. Recent research efforts have been trying to understand the miRNA-mediated regulation of HER signalling networks in cancers. Candidate miRNAs shown to be involved in the regulation of HER signalling were shown in Table 5.1.

Exosomal miR-21-5p and miR-30c-5p were examined as possible biomarkers for HER signalling in Section 5.3.6, and have been reproducibly detected in seven patients with advanced colorectal cancer undergoing first line systemic treatment using ddPCR.

There was trend toward negative correlation between changes in miR-21-5p from pre- treatment to three months with treatment response at 3 months. This finding supports the pro-oncogenic role of miR-21 described in literature (Peacock et al., 2014; Asangani et al., 2008a; XIONG et al., 2013; Sayed et al., 2008; Meng et al., 2007). It also validates the use of ddPCR for the clinical detection of circulating miR-21 from colorectal cancer patients, with 18S rRNA as normaliser despite previously described limitations. The sensitivities of CEA and miR-21 for detecting colorectal cancer has been documented to be 30.7% and 61.4%, respectively, and if CEA and miR-21 were combined, the sensitivity increased to 72.7% (Ogata-kawata et al., 2014). The accuracy of exosomal miR-21 for diagnosis was likely to be better than that of circulating miR-21 in various cancers (Shen et al., 2011; Wu et al., 2015a). As miR-21 has been described in many solid cancers besides colorectal cancer, exosomal miR- 21 alone is unlikely to be useful diagnostically, but could be a prognostic and predictive adjunct for response to therapy, coupled with other markers. Seike et al demonstrated EGFR signalling pathway upregulates miR-21 expression in lung cancer cells, with correlation between levels of miR-21 and phosphorylated EGFR (Seike et al., 2009). Barker et al suggested the presence of positive regulatory loop between miR-21 and HER signalling, whereby miR-21 promotes HER signalling (in part via suppression of PTEN) which in turn upregulates expression of miR-21, making miR-21 an attractive target to augment the therapeutic inhibition of HER signalling in cancer (Barker et al., 2010).

No significant correlations were found between changes in exosomal miR-30c with miR-20 or HER3 at 3 months. miR-30c has been implicated to act as either onco- miRNA or tumour suppressive miRNA in different cancers. Zhao et al showed miR-30c overexpression inhibited colorectal cancer proliferation partially via the regulation of BCL9, which is a driver for Wnt pathway activation (Zhao et al., 2019). Zhang et al

178 described inhibition of ADAM19 by miR30c contributed to the anti-tumour effect in colorectal cancer (Zhang et al., 2015). Furthermore, miR30c has been found to target the forkhead transcription factor FoxO3a (Jeon et al., 2015), which plays a role in HER3 reactivation in cancer cells treated with EGFR-targeted agents (Chakrabarty et al., 2012). A larger prospective study is required to validate the function of miR-30c in colorectal cancer.

179

Chapter 6

Conclusions and future work

180

The aim of this project was to investigate HER family receptor activation and dimerisation in colorectal cancer cells and cancer cell-derived exosomes upon anti- cancer treatment. The following objectives were pursued:

• Examine the effect of ligand stimulation and drug treatment on HER family receptor activation through phosphorylation and dimerisation using FRET/FLIM (Chapter three). • Understand the role of cell-derived exosomes for monitoring HER signalling in colorectal cancer cell lines (Chapter four). • Investigate the role of circulating exosomal protein and miRNA biomarkers in predicting treatment response and failure in colorectal cancer patients (Chapter five).

Results in Chapter three demonstrated differential basal level of expression of HER protein among different colorectal cell lines. Ligand stimulation with EGF and NRG1 resulted in differential transphosphorylation and thus by postulation dimerisation patterns of HER family receptors in different colorectal cancer cell lines examined. The effect on downstream activation of AKT and ERK was uniform across all cell lines, EGF induced predominately ERK phosphorylation whereas NRG1 induced predominately AKT phosphorylation. FRET/FLIM was shown to be an effective and reliable technique for quantitative analysis of HER dimerisation in colorectal cancer cells and FFPE cell pellets, as tested by ligand stimulation. Treatment with cetuximab resulted in significant increase in HER3 phosphorylation and HER2:HER3 heterodimerisation in sensitive cell line LIM1215, but not in resistance line DLD1 (KRAS WT). Treatment with AZD8931 resulted in reduced phosphorylation in HER1, HER2 and HER3 in both cell lines, and similar HER dimerisation changes as with treatment with cetuximab. Studies have frequently implicated the activation of HER2 and HER3 as a mechanism for treatment resistance to cetuximab. HER2:HER3 dimerisation has been found to be elevated upon cetuximab treatment in both colorectal and HNSCC cell lines. The findings in this Chapter positively support the central role of HER dimer rewiring in anti-HER resistance, and validate the use of FRET/FLIM in quantitative analysis of HER dimers in colorectal cancer. Dynamic monitoring of changes in HER dimer formation might be useful to predict efficacy of cetuximab and other anti-HER therapies.

Results in Chapter four showed colorectal cancer cell lines including DLD1 (KRAS WT), LIM1215, DiFi, SW48 and HCT116 (KRAS WT) secrete exosomes which can be isolated through differential ultracentrifugation. Exosomes were visualised using

181 electromicroscopy with immunostaining against exosomal marker CD63. The quality and concentration of the exosome pellet were examined using Nanosight LM10-HS analysis. Colorectal cancer cell-derived exosomes can be characterised by the presence of surface markers CD63 and ALIX, which were also used as normalisers for exosomal protein analysis. Dot blot was developed and used in preference to western blot due to its sensitivity and ability to test multiple markers using a small quantity of protein. Under unstimulated conditions, colorectal cancer cell-derived exosomes were selectively enriched in CD63, ALIX and EpCam, and packaged HER proteins to different extends compared to parental cells. Exosomal HER protein levels correlated with cellular protein levels in colorectal cancer cells at baseline, and could serve as useful biomarker for measuring HER expression in cells. Examining the effect of anti- HER therapy on exosomal secretion, treatment with cetuximab did not significantly alter exosomal HER protein content in either DLD1 (KRAS WT) or LIM1215. Treatment with AZD8931 caused a significant increase in exosomal HER2 and HER3 secretion from DLD1 (KRAS WT), and a trend increase in exosomal HER2 secretion from LIM1215. These changes in exosomes did not correlate with protein changes seen in cell lysates. Studies have demonstrated in response to drug treatment, colorectal cancer cells specifically alter exosomal profiles relating to drug resistance and tumour cell survival in vitro, reflecting in vivo phenomena occurring in patients. Analysis of other exosomal protein markers including phosphor-HER and downstream signalling proteins, as well as quantification of HER dimers using FRET/FLIM will help further evaluate role of exosomes as biomarkers for monitoring HER signalling in colorectal cancer.

Results in Chapter five demonstrated exosomes can be purified from circulating blood, both serum and plasma of patients, via differential ultracentrifugation. Circulating exosomes from colorectal cancer patients can be characterised by the presence of surface markers CD63 and ALIX, which can be used as loading control and to measure exosomal protein yield. Reverse Phase Protein Array is an alternative platform to dot blot in quantifying exosomal proteins in multiple samples on a single array using picolitre range of protein sample per dot. Further optimisation is needed to demonstrate reproducibility of results and reduce background signal. HER proteins have been consistently detected from circulating exosomes of colorectal cancer patients, and may have potential role in predicting treatment response. No significant correlation was demonstrated between cancer stage and exosomal HER levels in a cohort of eight patients. In a cohort of seven patients with advanced colorectal cancer undergoing first line systemic treatment, there were statistically significant correlation between

182 changes in exosomal HER1 and HER2, and HER1 and HER3 after 3 months. There was also trend towards positive correlation between HER2 and HER3. Pre-treatment to 3 months change in exosomal HER3 negatively correlated with radiological response at 6 months. In another cohort of colorectal cancer patients treated with one dose of AZD8931, early changes in exosomal HER proteins (from 8h to 7 days) did not correlate with each other or with radiological response at 12 weeks. However, after 5-7 days post AZD8931, changes in exosomal HER2:HER3 dimer quantity negatively correlated with S100A9, and was associated with trend towards positive radiological response at 12 weeks. Furthermore, exosomal miR-21-5p and miR-30c-5p were examined as possible biomarkers for HER signalling, and were reproducibly detected in seven patients with advanced colorectal cancer undergoing first line systemic treatment using ddPCR. There was trend toward negative correlation between changes in miR-21-5p from pre-treatment to three months with treatment response at 3 months, supporting its proto-oncogenic role as suggested in literature. No significant correlations were found between changes in exosomal miR-30c with miR-20 or HER3 at 3 months. Further studies are required to evaluate whether miR-30c plays a tumour- suppressive or oncogenic role in colorectal cancer.

In conclusion, results in this thesis show HER3 activation and HER2:HER3 dimer rewiring upon anti-HER1 therapy could contribute to treatment resistance in colorectal cancer cells. Exosomal HER protein quantity reflected HER expression in vitro and can be used to monitor treatment response to systemic therapy in vivo. Increase in exosomal HER3 predicted treatment failure prior to radiological progression. Early changes in exosomal HER2:HER3 dimer quantity may be an important predictive biomarker in colorectal cancer patients receiving anti-HER therapy. Exosomal miR-21 could be an additional biomarker for HER signalling and monitoring of treatment response in patients with colorectal cancer.

Future prospects It has been long recognised that rewiring of signalling cascades leads to resistance development against cancer therapies. It is important to understand the underlying mechanisms of signalling rewiring that regulate drug resistance in order to improve the prognosis and efficacy of treatments offered for colorectal cancer patients.

In line with results shown in this project, increased HER3 abundance and HER heterodimers after EGFR-targeted therapy has been shown in breast cancer patients,

183 and lapatinib treatment normalised HER3 phosphorylation and HER2:HER3 dimer formation. HER3-mediated resistance to EGFR inhibition with gefitinib has also been described in lung cancer cells. Sensitivity to EGFR inhibitors in colorectal cancer patients could be improved by co-inhibition of HER3, and combinational anti-HER therapies should be trialled in such patients for potential of enhanced response rate and duration of response.

Exosome has been demonstrated a promising tool for future cancer research. The identification of exosomal protein, miRNA and dimer biomarkers that are predictive of treatment response will allow better stratification of patients to therapy and early switch to alternative treatment prior to later radiological detection of cancer progression. The validation of these biomarkers will enable longitudinal monitoring of patients through therapy in the form of liquid biopsies rather than current practice of invasive tissue biopsies.

The ability to measure receptor dimerisation status in circulating exosomes using FRET/FLIM will help to design combinational regimes to counteract any observed oncogenic rewiring in individual patients thereby suppressing any rewired resistance. Paired HER dimer analysis within tumour tissue and circulating exosomes in colorectal cancer patients undergoing treatment will help further establish role of dimers as predictive biomarker for anti-cancer therapies. Prospective clinical studies of larger sample size are needed to validate the role of candidate miRNAs involved in HER signalling.

In this project we looked mainly at the effects of cetuximab and AZD8931 on HER dimerisation. Effect of other anti-HER therapies such as Sym004 as well as cytotoxic chemotherapy agents such as oxaliplatin and irinotecan used for treatment of advanced colorectal cancer could be explored.

Furthermore, there are still areas for improvement of techniques that have been developed in this project, including further optimisation of exosome isolation method that is time efficient without compromised quality, identification and purification of cancer-specific exosomes in circulating blood, further optimisation of proteomic analysis of exosomal protein including RPPA, and identification of better normalisation marker for exosomal miRNA analysis.

184 outcome in a certain group of patients. As exosomes are implicated in many physiological and pathological processes including cancer, cardiac diseases, neurodegenerative disorders, the potential application could be arched through different specialities. Refinement of our methods on exosome works could lead to commercialisation of blood tests that give hints to various disease status or prognosis stratification.

185

References

Adamczyk, K., Klein-Scory, S. and Tehrani, M. (2011a) Characterization of soluble and exosomal forms of the EGFR released from pancreatic cancer cells. Life sciences. 89: 304–312.

Adamczyk, K.A., Klein-Scory, S., Tehrani, M.M., Warnken, U., Schmiegel, W., Schn??lzer, M. and Schwarte-Waldhoff, I. (2011b) Characterization of soluble and exosomal forms of the EGFR released from pancreatic cancer cells. Life Sciences. 89: 304–312.

Adams, R., Brown, E., Brown, L., Butler, R., Falk, S., Fisher, D., Kaplan, R., Quirke, P., Richman, S., Samuel, L., Seligmann, J., Seymour, M., Shiu, K.K., Wasan, H., Wilson, R. and Maughan, T. (2018) Inhibition of EGFR, HER2, and HER3 signalling in patients with colorectal cancer wild-type for BRAF, PIK3CA, KRAS, and NRAS (FOCUS4-D): a phase 2–3 randomised trial. The Lancet Gastroenterology & Hepatology. 3: 162–171.

Adrain, C. and Freeman, M. (2014) Regulation of Receptor Tyrosine Kinase Ligand Processing. Cold Spring Harbor Perspectives in Biology. 6: a008995–a008995.

Aertgeerts, K., Skene, R., Yano, J., Sang, B.C., Zou, H., Snell, G., Jennings, A., Iwamoto, K., Habuka, N., Hirokawa, A., Ishikawa, T., Tanaka, T., Miki, H., Ohta, Y. and Sogabe, S. (2011) Structural analysis of the mechanism of inhibition and allosteric activation of the kinase domain of HER2 protein. Journal of Biological Chemistry. 286: 18756–18765.

Ahmed, F. (2007) Role of miRNA in carcinogenesis and biomarker selection: a methodological view. Expert review of molecular diagnostics. 7: 569–603.

Akbani, R., Becker, K.-F., Carragher, N., Goldstein, T., de Koning, L., Korf, U., Liotta, L., Mills, G.B., Nishizuka, S.S., Pawlak, M., Petricoin, E.F., Pollard, H.B., Serrels, B. and Zhu, J. (2014) Realizing the promise of reverse phase protein arrays for clinical, translational, and basic research: a workshop report: the RPPA (Reverse Phase Protein Array) society. Molecular & cellular proteomics : MCP. 13: 1625–1643.

Al-Nedawi, K., Meehan, B., Micallef, J., Lhotak, V., May, L., Guha, A. and Rak, J. (2008) Intercellular transfer of the oncogenic receptor EGFRvIII by microvesicles derived from tumour cells. Nature Cell Biology. 10: 619–624.

186

Al-Nedawi, K., Meehan, B. and Rak, J. (2009) Microvesicles: messengers and mediators of tumor progression. Cell cycle. 8: 2014–2018.

Alferez, D., Logié, A., Seedhar, P., Poulsom, R., Klinowska, T. and Wilkinson, R.W. (2010) 471 Inhibiting signalling by erbB receptor tyrosine kinases with AZD8931, a potent reversible small molecule inhibitor, reduces intestinal adenoma formation in the ApcMin/+ mouse model. European Journal of Cancer Supplements. 8: 151.

Allegra, C.J., Jessup, J.M., Somerfield, M.R., Hamilton, S.R., Hammond, E.H., Hayes, D.F., McAllister, P.K., Morton, R.F. and Schilsky, R.L. (2009) American Society of Clinical Oncology Provisional Clinical Opinion: Testing for KRAS Gene Mutations in Patients With Metastatic Colorectal Carcinoma to Predict Response to Anti–Epidermal Growth Factor Receptor Monoclonal Antibody Therapy. Journal of Clinical Oncology. 27: 2091–2096.

Andersen, C.L., Jensen, J.L. and Ørntoft, T.F. (2004) Normalization of Real-Time Quantitative Reverse Transcription-PCR Data: A Model-Based Variance Estimation Approach to Identify Genes Suited for Normalization, Applied to Bladder and Colon Cancer Data Sets. Cancer Research. 64: 5245–5250.

Ando, M., Hasegawa, Y. and Ando, Y. (2005) Pharmacogenetics of irinotecan: A promoter polymorphism of UGT1A1 gene and severe adverse reactions to irinotecan. Investigational New Drugs. 23: 539–545.

Andre, F., Schartz, N.E.C., Movassagh, M., Flament, C., Pautier, P., Morice, P., Pomel, C., Lhomme, C., Escudier, B., Le Chevalier, T., Tursz, T., Amigorena, S., Raposo, G., Angevin, E. and Zitvogel, L. (2002) Malignant effusions and immunogenic tumour-derived exosomes. Lancet (London, England). 360: 295–305.

Ang, C., Nedjadi, T. and Sheikh, A. (2010) Smad4 loss is associated with fewer S100A8-positive monocytes in colorectal tumors and attenuated response to S100A8 in colorectal and pancreatic cancer cells. Carcinogenesis. 31: 1541–1551.

Arena, S., Bellosillo, B., Siravegna, G., Martinez, A., Canadas, I., Lazzari, L., Ferruz, N., Russo, M., Misale, S., Gonzalez, I., Iglesias, M., Gavilan, E., Corti, G., Hobor, S., Crisafulli, G., Salido, M., Sanchez, J., Dalmases, A., Bellmunt, J., et al. (2015) Emergence of Multiple EGFR Extracellular Mutations during Cetuximab Treatment in Colorectal Cancer. Clinical Cancer Research. 21: 2157–2166.

Arnold, D., Lueza, B., Douillard, J.-Y., Peeters, M., Lenz, H.-J., Venook, A.,

187

Heinemann, V., Van Cutsem, E., Pignon, J.-P., Tabernero, J., Cervantes, A. and Ciardiello, F. (2017) Prognostic and predictive value of primary tumour side in patients with RAS wild-type metastatic colorectal cancer treated with chemotherapy and EGFR directed antibodies in six randomized trials†. Annals of Oncology. 28: 1713–1729.

Asangani, I., Rasheed, S. and Nikolova, D. (2008a) MicroRNA-21 (miR-21) post- transcriptionally downregulates tumor suppressor Pdcd4 and stimulates invasion, intravasation and metastasis in colorectal cancer. Oncogene. 27: 2128–2136.

Asangani, I.A., Rasheed, S.A.K., Nikolova, D.A., Leupold, J.H., Colburn, N.H., Post, S. and Allgayer, H. (2008b) MicroRNA-21 (miR-21) post-transcriptionally downregulates tumor suppressor Pdcd4 and stimulates invasion, intravasation and metastasis in colorectal cancer. Oncogene. 27: 2128–2136.

Babst, M. (2011) MVB vesicle formation: ESCRT-dependent, ESCRT-independent and everything in between. Current opinion in cell biology. 23: 4452–4457.

Baier, J., Maier, M., Engl, R., Landthaler, M. and Bäumler, W. (2005) Time-Resolved Investigations of Singlet Oxygen Luminescence in Water, in Phosphatidylcholine, and in Aqueous Suspensions of Phosphatidylcholine or HT29 Cells. The Journal of Physical Chemistry B. 109: 3041–3046.

Baietti, M., Zhang, Z. and Mortier, E. (2012a) Syndecan-syntenin-ALIX regulates the biogenesis of exosomes. Nature cell biology. 14: 677–685.

Baietti, M.F., Zhang, Z., Mortier, E., Melchior, A., Degeest, G., Geeraerts, A., Ivarsson, Y., Depoortere, F., Coomans, C., Vermeiren, E., Zimmermann, P. and David, G. (2012b) Syndecan–syntenin–ALIX regulates the biogenesis of exosomes. Nature Cell Biology. 14: 677–685.

Barault, L., Veyrie, N., Jooste, V., Lecorre, D., Chapusot, C., Ferraz, J.-M., Lièvre, A., Cortet, M., Bouvier, A.-M., Rat, P., Roignot, P., Faivre, J., Laurent-Puig, P. and Piard, F. (2008) Mutations in the RAS-MAPK, PI(3)K (phosphatidylinositol-3-OH kinase) signaling network correlate with poor survival in a population-based series of colon cancers. International Journal of Cancer. 122: 2255–2259.

Bardelli, A., Corso, S., Bertotti, A., Hobor, S., Valtorta, E., Siravegna, G., Sartore- Bianchi, A., Scala, E., Cassingena, A., Zecchin, D., Apicella, M., Migliardi, G., Galimi, F., Lauricella, C., Zanon, C., Perera, T., Veronese, S., Corti, G., Amatu, A., et al.

188

(2013) Amplification of the MET receptor drives resistance to anti-EGFR therapies in colorectal cancer. Cancer Discovery. 3: 658–673.

Bardelli, A. and Jänne, P. (2012) The road to resistance: EGFR mutation and cetuximab. Nature medicine. 18: 199–200.

Bardelli, A. and Siena, S. (2010) Molecular mechanisms of resistance to cetuximab and panitumumab in colorectal cancer Journal of Clinical Oncology. 28 (7) pp. 1254– 1261.

Barker, A., Giles, K.M., Epis, M.R., Zhang, P.M., Kalinowski, F. and Leedman, P.J. (2010) Regulation of ErbB receptor signalling in cancer cells by microRNA. Current Opinion in Pharmacology. 10: 655–661.

Baselga, J. and Swain, S.M. (2009) Novel anticancer targets: revisiting ERBB2 and discovering ERBB3. Nature reviews. Cancer. 9: 463–475.

Bastiaens, P.I. and Squire, A. (1999) Fluorescence lifetime imaging microscopy: spatial resolution of biochemical processes in the cell. Trends in cell biology. 9: 48– 52.

Baulida, J., Kraus, M.H., Alimandi, M., Di Fiore, P.P. and Carpenter, G. (1996) All ErbB receptors other than the epidermal growth factor receptor are endocytosis impaired. The Journal of biological chemistry. 271: 5251–5257.

Beckett, K., Monier, S., Palmer, L. and Alexandre, C. (2013) Drosophila S2 Cells Secrete Wingless on Exosome‐Like Vesicles but the Wingless Gradient Forms Independently of Exosomes. Traffic. 14: 82–96.

Beckler, M. and Higginbotham, J. (2013) Proteomic analysis of exosomes from mutant KRAS colon cancer cells identifies intercellular transfer of mutant KRAS. Molecular & Cellular proteomics. 12. 343-355

Beji, A., Horst, D., Engel, J., Kirchner, T. and Ullrich, A. (2012) Toward the Prognostic Significance and Therapeutic Potential of HER3 Receptor Tyrosine Kinase in Human Colon Cancer. Clinical Cancer Research. 18: 956–968.

Bellingham, S.A., Guo, B.B., Coleman, B.M. and Hill, A.F. (2012) Exosomes: Vehicles for the Transfer of Toxic Proteins Associated with Neurodegenerative Diseases. Frontiers in Physiology. 3

Benedix, F., Kube, R., Meyer, F., Schmidt, U., Gastinger, I., Lippert, H. and

189

Colon/Rectum Carcinomas (Primary Tumor) Study Group (2010) Comparison of 17,641 Patients With Right- and Left-Sided Colon Cancer: Differences in Epidemiology, Perioperative Course, Histology, and Survival. Diseases of the Colon & Rectum. 53: 57–64.

Berezin, M.Y. and Achilefu, S. (2011) Fluorescence Lifetime Measurements and Biological Imaging. Chemical Reviews. 110: 2641–2684.

Bertotti, A., Migliardi, G., Galimi, F., Sassi, F., Torti, D., Isella, C., Corà, D., Di Nicolantonio, F., Buscarino, M., Petti, C., Ribero, D., Russolillo, N., Muratore, A., Massucco, P., Pisacane, A., Molinaro, L., Valtorta, E., Sartore-Bianchi, A., Risio, M., et al. (2011) A molecularly annotated platform of patient-derived xenografts ("xenopatients") identifies HER2 as an effective therapeutic target in cetuximab-resistant colorectal cancer. Cancer discovery. 1: 508–523.

Bertotti, A. and Sassi, F. (2015) Molecular Pathways: Sensitivity and Resistance to Anti-EGFR Antibodies. Clinical cancer research : an official journal of the American Association for Cancer Research. 21: 3377–3383.

Bettegowda, C., Sausen, M., Leary, R.J., Kinde, I., Wang, Y., Agrawal, N., Bartlett, B.R., Wang, H., Luber, B., Alani, R.M., Antonarakis, E.S., Azad, N.S., Bardelli, A., Brem, H., Cameron, J.L., Lee, C.C., Fecher, L.A., Gallia, G.L., Gibbs, P., et al. (2014) Detection of Circulating Tumor DNA in Early- and Late-Stage Human Malignancies. Science Translational Medicine. 6: 224ra24-224ra24.

Bezler, M., Hengstler, J.G. and Ullrich, A. (2012) Inhibition of doxorubicin-induced HER3-PI3K-AKT signalling enhances apoptosis of ovarian cancer cells. Molecular Oncology. 6: 516–529.

Blackwell, K.L., Burstein, H.J., Storniolo, A.M., Rugo, H., Sledge, G., Koehler, M., Ellis, C., Casey, M., Vukelja, S., Bischoff, J., Baselga, J. and O’Shaughnessy, J. (2010) Randomized study of Lapatinib alone or in combination with trastuzumab in women with ErbB2-positive, trastuzumab-refractory metastatic breast cancer. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 28: 1124–1130.

Bobrie, A., Colombo, M., Krumeich, S., Raposo, G. and Théry, C. (2012) Diverse subpopulations of vesicles secreted by different intracellular mechanisms are present in exosome preparations obtained by differential ultracentrifugation Journal of Extracellular Vesicles. 1 pp. 1–11.

190

Bobrie, A., Colombo, M., Raposo, G. and Théry, C. (2011) Exosome secretion: molecular mechanisms and roles in immune responses. Traffic. 12: 1659–1668.

Boelens, M.C., Wu, T.J., Nabet, B.Y., Xu, B., Qiu, Y., Yoon, T., Azzam, D.J., Twyman-Saint Victor, C., Wiemann, B.Z., Ishwaran, H., Ter Brugge, P.J., Jonkers, J., Slingerland, J. and Minn, A.J. (2014) Exosome transfer from stromal to breast cancer cells regulates therapy resistance pathways. Cell. 159: 499–513.

Boellner, S. and Becker, K. (2015) Reverse Phase Protein Arrays—Quantitative Assessment of Multiple Biomarkers in Biopsies for Clinical Use. Microarrays. 98–114.

Böing, A.N., van der Pol, E., Grootemaat, A.E., Coumans, F.A.W., Sturk, A. and Nieuwland, R. (2014) Single-step isolation of extracellular vesicles by size-exclusion chromatography. Journal of Extracellular Vesicles. 3: 23430.

Bokemeyer, C., Bondarenko, I., Makhson, A., Hartmann, J.T., Aparicio, J., de Braud, F., Donea, S., Ludwig, H., Schuch, G., Stroh, C., Loos, A.H., Zubel, A. and Koralewski, P. (2009) Fluorouracil, leucovorin, and oxaliplatin with and without cetuximab in the first-line treatment of metastatic colorectal cancer. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 27: 663–671.

Bokemeyer, C., Kohne, C.-H., Ciardiello, F., Lenz, H.-J., Heinemann, V., Klinkhardt, U., Beier, F., Duecker, K. and Tejpar, S. (2014) Treatment outcome according to tumor RAS mutation status in OPUS study patients with metastatic colorectal cancer (mCRC) randomized to FOLFOX4 with/without cetuximab. Journal of Clinical Oncology. 32: 3505–3505.

Boland, C.R. and Goel, A. (2010) Microsatellite Instability in Colorectal Cancer. Gastroenterology. 138: 2073-2087.e3.

Bos, J.L. (1989) ras oncogenes in human cancer: a review. Cancer research. 49: 4682–4689.

Bosch-Vilaró, A., Jacobs, B., Pomella, V., Asbagh, L.A., Kirkland, R., Michel, J., Singh, S., Liu, X., Kim, P., Weitsman, G., Barber, P.R., Vojnovic, B., Ng, T. and Tejpar, S. (2017) Feedback activation of HER3 attenuates response to EGFR inhibitors in colon cancer cells. Oncotarget. 8: 4277.

Brand, T.M., Iida, M. and Wheeler, D.L. (2011) Molecular mechanisms of resistance to the EGFR monoclonal antibody cetuximab Cancer Biology and Therapy. 11 (9) pp. 777–792.

191

Brenner, H., Kloor, M. and Pox, C.P. (2014) Colorectal cancer. The Lancet. 383: 1490–1502.

Brulé, S.Y., Jonker, D.J., Karapetis, C.S., O’Callaghan, C.J., Moore, M.J., Wong, R., Tebbutt, N.C., Underhill, C., Yip, D., Zalcberg, J.R., Tu, D. and Goodwin, R.A. (2015) Location of colon cancer (right-sided versus left-sided) as a prognostic factor and a predictor of benefit from cetuximab in NCIC CO.17. European Journal of Cancer. 51: 1405–1414.

Bublil, E., Pines, G., Patel, G. and Fruhwirth, G. (2010) Kinase-mediated quasi- dimers of EGFR. The FASEB Journal. 24: 4744–4755.

Burke, M. and Choksawangkarn, W. (2014) Exosomes from myeloid-derived suppressor cells carry biologically active proteins. Journal of proteome research. 13: 836–843.

Burn, J., Mathers, J. and Bishop, D.T. (2013) Genetics, inheritance and strategies for prevention in populations at high risk of colorectal cancer (CRC). Recent results in cancer research. Fortschritte der Krebsforschung. Progrès dans les recherches sur le cancer. 191: 157–183.

Camacho, L., Guerrero, P. and Marchetti, D. (2013) MicroRNA and Protein Profiling of Brain Metastasis Competent Cell-Derived Exosomes. PLoS ONE. 8:.

Campomenosi, P., Gini, E., Noonan, D.M., Poli, A., D’Antona, P., Rotolo, N., Dominioni, L. and Imperatori, A. (2016) A comparison between quantitative PCR and droplet digital PCR technologies for circulating microRNA quantification in human lung cancer. BMC Biotechnology. 16: 60.

Castillo, J., Bernard, V., San Lucas, F.A., Allenson, K., Capello, M., Kim, D.U., Gascoyne, P., Mulu, F.C., Stephens, B.M., Huang, J., Wang, H., Momin, A.A., Jacamo, R.O., Katz, M., Wolff, R., Javle, M., Varadhachary, G., Wistuba, I.I., Hanash, S., et al. (2018) Surfaceome profiling enables isolation of cancer-specific exosomal cargo in liquid biopsies from pancreatic cancer patients. Annals of Oncology. 29: 223–229.

Chakrabarty, A., Sanchez, V., Kuba, M.G., Rinehart, C. and Arteaga, C.L. (2012) Feedback upregulation of HER3 (ErbB3) expression and activity attenuates antitumor effect of PI3K inhibitors. Proceedings of the National Academy of Sciences. 109: 2718–2723.

192

Chaput, N. and Théry, C. (2011) Exosomes: immune properties and potential clinical implementations. Seminars in immunopathology. 33: 419–440.

Chen, G., Huang, A.C., Zhang, W., Zhang, G., Wu, M., Xu, W., Yu, Z., Yang, J., Wang, B., Sun, H., Xia, H., Man, Q., Zhong, W., Antelo, L.F., Wu, B., Xiong, X., Liu, X., Guan, L., Li, T., et al. (2018) Exosomal PD-L1 contributes to immunosuppression and is associated with anti-PD-1 response. Nature. 560: 382–386.

Chen, W. xian, Liu, X. min, Lv, M. meng, Chen, L., Zhao, J. hua, Zhong, S. liang, Ji, M. hua, Hu, Q., Luo, Z., Wu, J. zhong and Tang, J. hai (2014) Exosomes from drug- resistant breast cancer cells transmit chemoresistance by a horizontal transfer of microRNAs. PloS one. 9: e95240.

Cheng, P., Corzo, C. and Luetteke, N. (2008) Inhibition of dendritic cell differentiation and accumulation of myeloid-derived suppressor cells in cancer is regulated by S100A9 protein. The Journal of experimental medicine. 205: 2235–2249.

Cho, H. and Leahy, D. (2002) Structure of the extracellular region of HER3 reveals an interdomain tether. Science. 297: 1330–1333.

Chornoguz, O., Grmai, L., Sinha, P., Artemenko, K.A., Zubarev, R.A. and Ostrand- Rosenberg, S. (2011) Proteomic pathway analysis reveals inflammation increases myeloid-derived suppressor cell resistance to apoptosis. Molecular and Cellular Proteomics. 10:.

Chung, K.Y., Shia, J., Kemeny, N.E., Shah, M., Schwartz, G.K., Tse, A., Hamilton, A., Pan, D., Schrag, D., Schwartz, L., Klimstra, D.S., Fridman, D., Kelsen, D.P. and Saltz, L.B. (2005) Cetuximab shows activity in colorectal cancer patients with tumors that do not express the epidermal growth factor receptor by immunohistochemistry. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 23: 1803–1810.

Ciardiello, F., Lenz, H.-J., Kohne, C.-H., Heinemann, V., Tejpar, S., Melezinek, I., Beier, F., Stroh, C. and Van Cutsem, E. (2014) Treatment outcome according to tumor RAS mutation status in CRYSTAL study patients with metastatic colorectal cancer (mCRC) randomized to FOLFIRI with/without cetuximab. Journal of Clinical Oncology. 32: 3506–3506.

Ciardiello, F. and Tortora, G. (2008) Drug therapy: EGFR antagonists in cancer treatment New England Journal of Medicine. 358 (11) p. 1160.

193

Ciravolo, V., Huber, V., Ghedini, G.C., Venturelli, E., Bianchi, F., Campiglio, M., Morelli, D., Villa, A., Mina, P. Della, Menard, S., Filipazzi, P., Rivoltini, L., Tagliabue, E. and Pupa, S.M. (2012) Potential role of HER2-overexpressing exosomes in countering trastuzumab-based therapy. Journal of Cellular Physiology. 227: 658– 667.

Claus, J., Patel, G., Autore, F., Colomba, A., Weitsman, G., Soliman, T.N., Roberts, S., Zanetti-Domingues, L.C., Hirsch, M., Collu, F., George, R., Ortiz-Zapater, E., Barber, P.R., Vojnovic, B., Yarden, Y., Martin-Fernandez, M.L., Cameron, A., Fraternali, F., Ng, T., et al. (2018) Inhibitor-induced HER2-HER3 heterodimerisation promotes proliferation through a novel dimer interface. eLife. 7:

Colombo, M., Raposo, G. and Théry, C. (2014) Biogenesis, Secretion, and Intercellular Interactions of Exosomes and Other Extracellular Vesicles. Annual Review of Cell and Developmental Biology. 30: 255–289.

Colucci, G., Gebbia, V., Paoletti, G., Giuliani, F., Caruso, M., Gebbia, N., Cartenì, G., Agostara, B., Pezzella, G., Manzione, L., Borsellino, N., Misino, A., Romito, S., Durini, E., Cordio, S., Di Seri, M., Lopez, M. and Maiello, E. (2005) Phase III randomized trial of FOLFIRI versus FOLFOX4 in the treatment of advanced colorectal cancer: A Multicenter Study of the Gruppo Oncologico Dell’Italia Meridionale. Journal of Clinical Oncology. 23: 4866–4875.

Corcoran, C., Friel, A. and Duffy, M. (2011) Intracellular and extracellular microRNAs in breast cancer. Clinical chemistry. 57: 18–32.

Corcoran, R.B., Atreya, C.E., Falchook, G.S., Kwak, E.L., Ryan, D.P., Bendell, J.C., Hamid, O., Messersmith, W.A., Daud, A., Kurzrock, R., Pierobon, M., Sun, P., Cunningham, E., Little, S., Orford, K., Motwani, M., Bai, Y., Patel, K., Venook, A.P., et al. (2015) Combined BRAF and MEK Inhibition With Dabrafenib and Trametinib in BRAF V600–Mutant Colorectal Cancer. Journal of Clinical Oncology. 33: 4023–4031.

Correale, P., Marra, M., Remondo, C., Migali, C., Misso, G., Arcuri, F.P., Del Vecchio, M.T., Carducci, A., Loiacono, L., Tassone, P., Abbruzzese, A., Tagliaferri, P. and Caraglia, M. (2010) Cytotoxic drugs up-regulate epidermal growth factor receptor (EGFR) expression in colon cancer cells and enhance their susceptibility to EGFR-targeted antibody-dependent cell-mediated-cytotoxicity (ADCC). European journal of cancer (Oxford, England : 1990). 46: 1703–1711.

Costa-Silva, B., Aiello, N.M., Ocean, A.J., Singh, S., Zhang, H., Thakur, B.K., Becker,

194

A., Hoshino, A., Mark, M.T., Molina, H., Xiang, J., Zhang, T., Theilen, T.-M., García- Santos, G., Williams, C., Ararso, Y., Huang, Y., Rodrigues, G., Shen, T.-L., et al. (2015) Pancreatic cancer exosomes initiate pre-metastatic niche formation in the liver. Nature cell biology. 17: 1–7.

Cremolini, C., Loupakis, F., Antoniotti, C., Lupi, C., Sensi, E., Lonardi, S., Mezi, S., Tomasello, G., Ronzoni, M., Zaniboni, A., Tonini, G., Carlomagno, C., Allegrini, G., Chiara, S., D’Amico, M., Granetto, C., Cazzaniga, M., Boni, L., Fontanini, G., et al. (2015) FOLFOXIRI plus bevacizumab versus FOLFIRI plus bevacizumab as first-line treatment of patients with metastatic colorectal cancer: updated overall survival and molecular subgroup analyses of the open-label, phase 3 TRIBE study. The Lancet Oncology. 16: 1306–1315.

Cunningham, D., Lang, I., Marcuello, E., Lorusso, V., Ocvirk, J., Shin, D.B., Jonker, D., Osborne, S., Andre, N., Waterkamp, D., Saunders, M.P. and AVEX study investigators (2013) Bevacizumab plus capecitabine versus capecitabine alone in elderly patients with previously untreated metastatic colorectal cancer (AVEX): an open-label, randomised phase 3 trial. The Lancet. Oncology. 14: 1077–1085.

Cunningham, M.P., Essapen, S., Thomas, H., Green, M., Lovell, D.P., Topham, C., Marks, C. and Modjtahedi, H. (2006) Coexpression of the IGF-IR, EGFR and HER-2 is common in colorectal cancer patients. International journal of oncology. 28: 329– 335.

Cushman, S.M., Jiang, C., Hatch, A.J., Shterev, I., Sibley, A.B., Niedzwiecki, D., Venook, A.P., Owzar, K., Hurwitz, H.I. and Nixon, A.B. (2015) Gene expression markers of efficacy and resistance to cetuximab treatment in metastatic colorectal cancer: results from CALGB 80203 (Alliance). Clinical cancer research : an official journal of the American Association for Cancer Research. 21: 1078–1086.

Van Cutsem, E., Eng, C., Nowara, E., wieboda-Sadlej, A., Tebbutt, N.C., Mitchell, E., Davidenko, I., Stephenson, J., Elez, E., Prenen, H., Deng, H., Tang, R., McCaffery, I., Oliner, K.S., Chen, L., Gansert, J., Loh, E., Smethurst, D. and Tabernero, J. (2014) Randomized Phase Ib/II Trial of Rilotumumab or Ganitumab with Panitumumab versus Panitumumab Alone in Patients with Wild-type KRAS Metastatic Colorectal Cancer. Clinical Cancer Research. 20: 4240–4250.

Van Cutsem, E., Köhne, C.H., Hitre, E., Zaluski, J., Chien, C.R.C., Makhson, A., D’Haens, G., Pintér, T., Lim, R., Bodoky, G., Roh, J.K., Folprecht, G., Ruff, P., Stroh,

195

C., Tejpar, S., Schlichting, M., Nippgen, J. and Rougier, P. (2009) Cetuximab and chemotherapy as initial treatment for metastatic colorectal cancer. New England Journal of Medicine. 360: 1408–1417.

Van Cutsem, E., Lenz, H.J., Köhne, C.H., Heinemann, V., Tejpar, S., Melezínek, I., Beier, F., Stroh, C., Rougier, P., Han Van Krieken, J. and Ciardiello, F. (2015) Fluorouracil, leucovorin, and irinotecan plus cetuximab treatment and RAS mutations in colorectal cancer. Journal of Clinical Oncology. 33: 692–700.

Van Cutsem, E., Mayer, R.J., Laurent, S., Winkler, R., Grávalos, C., Benavides, M., Longo-Munoz, F., Portales, F., Ciardiello, F., Siena, S., Yamaguchi, K., Muro, K., Denda, T., Tsuji, Y., Makris, L., Loehrer, P., Lenz, H.-J., Ohtsu, A. and RECOURSE Study Group (2018) The subgroups of the phase III RECOURSE trial of trifluridine/tipiracil (TAS-102) versus placebo with best supportive care in patients with metastatic colorectal cancer. European Journal of Cancer. 90: 63–72.

Van Cutsem, E., Peeters, M., Siena, S., Humblet, Y., Hendlisz, A., Neyns, B., Canon, J.-L., Van Laethem, J.-L., Maurel, J., Richardson, G., Wolf, M. and Amado, R.G. (2007) Open-Label Phase III Trial of Panitumumab Plus Best Supportive Care Compared With Best Supportive Care Alone in Patients With Chemotherapy- Refractory Metastatic Colorectal Cancer. Journal of Clinical Oncology. 25: 1658– 1664.

Van Cutsem, E., Tabernero, J., Lakomy, R., Prenen, H., Prausová, J., Macarulla, T., Ruff, P., van Hazel, G.A., Moiseyenko, V., Ferry, D., McKendrick, J., Polikoff, J., Tellier, A., Castan, R. and Allegra, C. (2012) Addition of Aflibercept to Fluorouracil, Leucovorin, and Irinotecan Improves Survival in a Phase III Randomized Trial in Patients With Metastatic Colorectal Cancer Previously Treated With an Oxaliplatin- Based Regimen. Journal of Clinical Oncology. 30: 3499–3506.

Dawson, J.P., Berger, M.B., Lin, C.-C., Schlessinger, J., Lemmon, M.A. and Ferguson, K.M. (2005) Epidermal Growth Factor Receptor Dimerization and Activation Require Ligand-Induced Conformational Changes in the Dimer Interface. Molecular and Cellular Biology. 25: 7734–7742.

Demory Beckler, M., Higginbotham, J.N., Franklin, J.L., Ham, A.-J., Halvey, P.J., Imasuen, I.E., Whitwell, C., Li, M., Liebler, D.C. and Coffey, R.J. (2013) Proteomic analysis of exosomes from mutant KRAS colon cancer cells identifies intercellular transfer of mutant KRAS. Molecular & cellular proteomics : MCP. 12: 343–355.

196

Deo, M., Yu, J.-Y., Chung, K.-H., Tippens, M. and Turner, D.L. (2006) Detection of mammalian microRNA expression by in situ hybridization with RNA oligonucleotides. Developmental Dynamics. 235: 2538–2548.

DeSantis, C.E., Lin, C.C., Mariotto, A.B., Siegel, R.L., Stein, K.D., Kramer, J.L., Alteri, R., Robbins, A.S. and Jemal, A. (2014) Cancer treatment and survivorship statistics, 2014. CA: A Cancer Journal for Clinicians. 64: 252–271.

Diaz Jr, L.A., Williams, R.T., Wu, J., Kinde, I., Hecht, J.R., Berlin, J., Allen, B., Bozic, I., Reiter, J.G., Nowak, M.A., Kinzler, K.W., Oliner, K.S. and Vogelstein, B. (2012) The molecular evolution of acquired resistance to targeted EGFR blockade in colorectal cancers. Nature. 486: 537–540.

Dienstmann, R., Patnaik, A., Garcia-Carbonero, R., Cervantes, A., Benavent, M., Rosello, S., Tops, B.B.J., van der Post, R.S., Argiles, G., Skartved, N.J.O., Hansen, U.H., Hald, R., Pedersen, M.W., Kragh, M., Horak, I.D., Braun, S., Van Cutsem, E., Tolcher, A.W. and Tabernero, J. (2015) Safety and Activity of the First-in-Class Sym004 Anti-EGFR Antibody Mixture in Patients with Refractory Colorectal Cancer. Cancer Discovery. 5: 598–609.

Dittmann, K., Mayer, C., Fehrenbacher, B., Schaller, M., Raju, U., Milas, L., Chen, D.J., Kehlbach, R. and Rodemann, H.P. (2005) Radiation-induced epidermal growth factor receptor nuclear import is linked to activation of DNA-dependent protein kinase. The Journal of biological chemistry. 280: 31182–31189. van Dommelen, S.M., van der Meel, R., van Solinge, W.W., Coimbra, M., Vader, P. and Schiffelers, R.M. (2016) Cetuximab treatment alters the content of extracellular vesicles released from tumor cells. Nanomedicine. 11: 881–890.

Douillard, J.-Y., Siena, S., Cassidy, J., Tabernero, J., Burkes, R., Barugel, M., Humblet, Y., Bodoky, G., Cunningham, D., Jassem, J., Rivera, F., Kocakova, I., Ruff, P., B asi ska-Morawiec, M., makal, M., Canon, J.-L., Rother, M., Oliner, K.S., Wolf, M., et al. (2010) Randomized, Phase III Trial of Panitumumab With Infusional Fluorouracil, Leucovorin, and Oxaliplatin (FOLFOX4) Versus FOLFOX4 Alone As First-Line Treatment in Patients With Previously Untreated Metastatic Colorectal Cancer: The PRIME Study. Journal of Clinical Oncology. 28: 4697–4705.

Douillard, J.Y., Cunningham, D., Roth, A.D., Navarro, M., James, R.D., Karasek, P., Jandik, P., Iveson, T., Carmichael, J., Alakl, M., Gruia, G., Awad, L. and Rougier, P. (2000) Irinotecan combined with fluorouracil compared with fluorouracil alone as first-

197 line treatment for metastatic colorectal cancer: a multicentre randomised trial. Lancet (London, England). 355: 1041–1047.

Douillard, J.Y., Oliner, K.S., Siena, S., Tabernero, J., Burkes, R., Barugel, M., Humblet, Y., Bodoky, G., Cunningham, D., Jassem, J., Rivera, F., Kocákova, I., Ruff, P., Błasińska-Morawiec, M., Šmakal, M., Canon, J.L., Rother, M., Williams, R., Rong, A., et al. (2013) Panitumumab-FOLFOX4 treatment and RAS mutations in colorectal cancer. New England Journal of Medicine. 369: 1023–1034.

Duffy, M.J. (2001) Carcinoembryonic antigen as a marker for colorectal cancer: is it clinically useful? Clinical chemistry. 47: 624–630.

Emura, T., Suzuki, N., Yamaguchi, M., Ohshimo, H. and Fukushima, M. (2004) A novel combination antimetabolite, TAS-102, exhibits antitumor activity in FU-resistant human cancer cells through a mechanism involving FTD incorporation in DNA. International journal of oncology. 25: 571–578.

Engelman, J.A., Zejnullahu, K., Mitsudomi, T., Song, Y., Hyland, C., Park, J.O., Lindeman, N., Gale, C.-M., Zhao, X., Christensen, J., Kosaka, T., Holmes, A.J., Rogers, A.M., Cappuzzo, F., Mok, T., Lee, C., Johnson, B.E., Cantley, L.C. and Janne, P.A. (2007) MET Amplification Leads to Gefitinib Resistance in Lung Cancer by Activating ERBB3 Signaling. Science. 316: 1039–1043.

Fearon, E.R. (2011) Molecular genetics of colorectal cancer. Annual review of pathology. 6: 479–507.

Federici, C., Petrucci, F., Caimi, S., Cesolini, A., Logozzi, M., Borghi, M., D’Ilio, S., Lugini, L., Violante, N., Azzarito, T., Majorani, C., Brambilla, D. and Fais, S. (2014) Exosome release and low pH belong to a framework of resistance of human melanoma cells to cisplatin. PLoS ONE. 9:.

Feng, P.-H., Yu, C.-T., Chen, K.-Y., Luo, C.-S., Wu, S.M., Liu, C.-Y., Kuo, L.W., Chan, Y.-F., Chen, T.-T., Chang, C.-C., Lee, C.-N., Chuang, H.-C., Lin, C.-F., Han, C.-L., Lee, W.-H. and Lee, K.-Y. (2018) S100A9+ MDSC and TAM-mediated EGFR- TKI resistance in lung adenocarcinoma: the role of RELB. Oncotarget. 9: 7631.

Ferguson, K.M. (2008) Structure-Based View of Epidermal Growth Factor Receptor Regulation. Annual Review of Biophysics. 37: 353–373.

Ferlay, J., Soerjomataram, I., Dikshit, R., Eser, S., Mathers, C., Rebelo, M., Parkin, D.M., Forman, D. and Bray, F. (2015) Cancer incidence and mortality worldwide:

198

Sources, methods and major patterns in GLOBOCAN 2012. International Journal of Cancer. 136: E359–E386.

Fernandez-Medarde, A. and Santos, E. (2011) Ras in Cancer and Developmental Diseases. Genes & Cancer. 2: 344–358.

Ferrara, N. (2004) Vascular Endothelial Growth Factor: Basic Science and Clinical Progress. Endocrine Reviews. 25: 581–611.

Festy, F., Ameer-Beg, S.M., Ng, T. and Suhling, K. (2007) Imaging proteins in vivo using fluorescence lifetime microscopy. Molecular BioSystems. 3: 381.

Freed, D.M., Bessman, N.J., Kiyatkin, A., Salazar-Cavazos, E., Byrne, P.O., Moore, J.O., Valley, C.C., Ferguson, K.M., Leahy, D.J., Lidke, D.S. and Lemmon, M.A. (2017) EGFR Ligands Differentially Stabilize Receptor Dimers to Specify Signaling Kinetics. Cell. 171: 683-695.e18.

Fruhwirth, G.O., Fernandes, L.P., Weitsman, G., Patel, G., Kelleher, M., Lawler, K., Brock, A., Poland, S.P., Matthews, D.R., Kéri, G., Barber, P.R., Vojnovic, B., Ameer- Beg, S.M., Coolen, A.C.C., Fraternali, F. and Ng, T. (2011) How Förster resonance energy transfer imaging improves the understanding of protein interaction networks in cancer biology. Chemphyschem : a European journal of chemical physics and physical chemistry. 12: 442–461.

Fukushima, M., Suzuki, N., Emura, T., Yano, S., Kazuno, H., Tada, Y., Yamada, Y. and Asao, T. (2000) Structure and activity of specific inhibitors of thymidine phosphorylase to potentiate the function of antitumor 2’-deoxyribonucleosides. Biochemical pharmacology. 59: 1227–1236.

Gardiner, C., Vizio, D. Di, Sahoo, S., Théry, C., Witwer, K.W., Wauben, M. and Hill, A.F. (2016) Techniques used for the isolation and characterization of extracellular vesicles: results of a worldwide survey. Journal of Extracellular Vesicles. 5: 32945.

Gebhardt, C., Németh, J., Angel, P. and Hess, J. (2006) S100A8 and S100A9 in inflammation and cancer. Biochemical pharmacology. 72: 1622–1631.

Geissler, M., Tannapfel, A., Reinacher-Schick, A., Martens, U., Ricke, J., Riera- Knorrenschield, J., Kanzler, S., Held, S., Heinemann, V., Seufferlein, T. and Modest, D. (2019) PD-030Final results of the randomized phase II VOLFI trial (AIO- KRK0109): mFOLFOXIRI + Panitumumab versus FOLFOXIRI as first-line treatment in patients with RAS wild-type metastatic colorectal cancer (mCRC). Annals of

199

Oncology. 30:.

Gerritzen, M.J.H., Martens, D.E., Wijffels, R.H. and Stork, M. (2017) High throughput nanoparticle tracking analysis for monitoring outer membrane vesicle production. Journal of extracellular vesicles. 6: 1333883.

Ghavami, S. and Kerkhoff, C. (2004) Mechanism of apoptosis induced by S100A8/A9 in colon cancer cell lines: the role of ROS and the effect of metal ions. Journal of leukocyte biology. 76: 169–175.

Ghavami, S. and Rashedi, I. (2008) S100A8/A9 at low concentration promotes tumor cell growth via RAGE ligation and MAP kinase-dependent pathway. Journal of leukocyte biology. 83: 1484–1492.

Gherardi, E., Birchmeier, W., Birchmeier, C. and Woude, G. Vande (2012) Targeting MET in cancer: rationale and progress. Nature Reviews Cancer. 12: 89–103.

Gilbert, D.C., Chalmers, A.J. and El-Khamisy, S.F. (2012) Topoisomerase I inhibition in colorectal cancer: biomarkers and therapeutic targets. British journal of cancer. 106: 18–24.

Gomez, G.G., Wykosky, J., Zanca, C., Furnari, F.B. and Cavenee, W.K. (2013) Therapeutic resistance in cancer: microRNA regulation of EGFR signaling networks. Cancer biology & medicine. 10: 192–205. de Gramont, A., Bosset, J.F., Milan, C., Rougier, P., Bouché, O., Etienne, P.L., Morvan, F., Louvet, C., Guillot, T., François, E. and Bedenne, L. (1997) Randomized trial comparing monthly low-dose leucovorin and fluorouracil bolus with bimonthly high-dose leucovorin and fluorouracil bolus plus continuous infusion for advanced colorectal cancer: a French intergroup study. Journal of Clinical Oncology. 15: 808– 815. de Gramont, A., Figer, A., Seymour, M., Homerin, M., Hmissi, A., Cassidy, J., Boni, C., Cortes-Funes, H., Cervantes, A., Freyer, G., Papamichael, D., Le Bail, N., Louvet, C., Hendler, D., de Braud, F., Wilson, C., Morvan, F. and Bonetti, A. (2000) Leucovorin and Fluorouracil With or Without Oxaliplatin as First-Line Treatment in Advanced Colorectal Cancer. Journal of Clinical Oncology. 18: 2938–2947.

Grandal, M. V., Zandi, R., Pedersen, M.W., Willumsen, B.M., van Deurs, B. and Poulsen, H.S. (2007) EGFRvIII escapes down-regulation due to impaired internalization and sorting to lysosomes. Carcinogenesis. 28: 1408–1417.

200

Graner, M., Alzate, O. and Dechkovskaia, A. (2009) Proteomic and immunologic analyses of brain tumor exosomes. The FASEB Journal. 23: 1541–1557.

Greenall, S.A. and Johns, T.G. (2016) EGFRvIII: the promiscuous mutation. Cell Death Discovery. 2: 16049.

Griffiths, D., Hole, P., Smith, J., Malloy, A. and Carr, B. (2010) Size and count of nanoparticles by scattering and fluorescence Nanoparticle Tracking Analysis (NTA). In: Nanotechnology 2010: Advanced Materials, CNTs, Particles, Films and Composites - Technical Proceedings of the 2010 NSTI Nanotechnology Conference and Expo, NSTI-Nanotech 2010. 2010 pp. 37–40.

Gross, J., Chaudhary, V., Bartscherer, K. and Boutros, M. (2012) Active Wnt proteins are secreted on exosomes. Nature cell biology. 14: 1036–1045.

Grothey, A., Van Cutsem, E., Sobrero, A., Siena, S., Falcone, A., Ychou, M., Humblet, Y., Bouché, O., Mineur, L., Barone, C., Adenis, A., Tabernero, J., Yoshino, T., Lenz, H.-J., Goldberg, R.M., Sargent, D.J., Cihon, F., Cupit, L., Wagner, A., et al. (2013) Regorafenib monotherapy for previously treated metastatic colorectal cancer (CORRECT): an international, multicentre, randomised, placebo-controlled, phase 3 trial. Lancet (London, England). 381: 303–312.

Grothey, A., Sargent, D., Goldberg, R.M. and Schmoll, H.-J. (2004) Survival of Patients With Advanced Colorectal Cancer Improves With the Availability of Fluorouracil-Leucovorin, Irinotecan, and Oxaliplatin in the Course of Treatment. Journal of Clinical Oncology. 22: 1209–1214.

Gschwind, A., Fischer, O.M. and Ullrich, A. (2004) The discovery of receptor tyrosine kinases: targets for cancer therapy. Nature reviews. Cancer. 4: 361–370.

Gschwind, A., Zwick, E., Prenzel, N., Leserer, M. and Ullrich, A. (2001) Cell communication networks: epidermal growth factor receptor transactivation as the paradigm for interreceptor signal transmission. Oncogene. 20: 1594–1600.

Gu, Y., Zhang, H., Su, D., Mo, M., Song, P., Zhang, F. and Zhang, S. (2013) miR-30b and miR-30c expression predicted response to tyrosine kinase inhibitors as first line treatment in non-small cell lung cancer. Chinese medical journal. 126: 4435–4439.

Guinney, J., Dienstmann, R., Wang, X., de Reyniès, A., Schlicker, A., Soneson, C., Marisa, L., Roepman, P., Nyamundanda, G., Angelino, P., Bot, B.M., Morris, J.S., Simon, I.M., Gerster, S., Fessler, E., De Sousa E Melo, F., Missiaglia, E., Ramay, H.,

201

Barras, D., et al. (2015) The consensus molecular subtypes of colorectal cancer. Nature Medicine. 21: 1350–1356.

Guo, L., Kozlosky, C.J., Ericsson, L.H., Daniel, T.O., Cerretti, D.P. and Johnson, R.S. (2003) Studies of ligand-induced site-specific phosphorylation of epidermal growth factor receptor. Journal of the American Society for Mass Spectrometry. 14: 1022– 1031.

Harding, C., Heuser, J. and Stahl, P. (1983) Receptor-mediated endocytosis of transferrin and recycling of the transferrin receptor in rat reticulocytes. The Journal of cell biology. 97: 329–339.

Heinemann, V., von Weikersthal, L.F., Decker, T., Kiani, A., Vehling-Kaiser, U., Al- Batran, S.-E., Heintges, T., Lerchenmüller, C., Kahl, C., Seipelt, G., Kullmann, F., Stauch, M., Scheithauer, W., Hielscher, J., Scholz, M., Müller, S., Link, H., Niederle, N., Rost, A., et al. (2014) FOLFIRI plus cetuximab versus FOLFIRI plus bevacizumab as first-line treatment for patients with metastatic colorectal cancer (FIRE-3): a randomised, open-label, phase 3 trial. The Lancet Oncology. 15: 1065–1075.

Heinicke, E., Kumar, U. and Munoz, D.G. (1992) Quantitative dot-blot assay for proteins using enhanced chemiluminescence. Journal of immunological methods. 152: 227–236.

Hennessy, B.T., Lu, Y., Gonzalez-Angulo, A.M., Carey, M.S., Myhre, S., Ju, Z., Davies, M.A., Liu, W., Coombes, K., Meric-Bernstam, F., Bedrosian, I., McGahren, M., Agarwal, R., Zhang, F., Overgaard, J., Alsner, J., Neve, R.M., Kuo, W.L., Gray, J.W., et al. (2010) A technical assessment of the utility of reverse phase protein arrays for the study of the functional proteome in non-microdissected human breast cancers. Clinical Proteomics. 6: 129–151.

Hermani, A. and Servi, B. De (2006) S100A8 and S100A9 activate MAP kinase and NF-κB signaling pathways and trigger translocation of RAGE in human prostate cancer cells. Experimental cell research. 312: 184–197.

Hessvik, N.P. and Llorente, A. (2018) Current knowledge on exosome biogenesis and release. Cellular and molecular life sciences : CMLS. 75: 193–208.

Hickinson, D.M., Klinowska, T., Speake, G., Vincent, J., Trigwell, C., Anderton, J., Beck, S., Marshall, G., Davenport, S., Callis, R., Mills, E., Grosios, K., Smith, P., Barlaam, B., Wilkinson, R.W. and Ogilvie, D. (2010) AZD8931, an equipotent,

202 reversible inhibitor of signaling by epidermal growth factor receptor, ERBB2 (HER2), and ERBB3: A unique agent for simultaneous ERBB receptor blockade in cancer. Clinical Cancer Research. 16: 1159–1169.

Higginbotham, J.N., Demory Beckler, M., Gephart, J.D., Franklin, J.L., Bogatcheva, G., Kremers, G.J., Piston, D.W., Ayers, G.D., McConnell, R.E., Tyska, M.J. and Coffey, R.J. (2011) Amphiregulin exosomes increase cancer cell invasion. Current Biology. 21: 779–786.

Hiratsuka, S. and Watanabe, A. (2008) The S100A8–serum amyloid A3–TLR4 paracrine cascade establishes a pre-metastatic phase. Nature cell biology. 10: 1349– 1355.

Hiratsuka, S., Watanabe, A., Aburatani, H. and Maru, Y. (2006) Tumour-mediated upregulation of chemoattractants and recruitment of myeloid cells predetermines lung metastasis. Nature cell biology. 8: 1369–1375.

Hobor, S., Van Emburgh, B.O., Crowley, E., Misale, S., Di Nicolantonio, F. and Bardelli, A. (2014) TGFα and amphiregulin paracrine network promotes resistance to EGFR blockade in colorectal cancer cells. Clinical Cancer Research. 20: 6429–6438.

Holbro, T. and Beerli, R. (2003) The ErbB2/ErbB3 heterodimer functions as an oncogenic unit: ErbB2 requires ErbB3 to drive breast tumor cell proliferation. Proc Natl Acad Sci USA. 100: 8933–8938.

Holch, J.W., Ricard, I., Stintzing, S., Modest, D.P. and Heinemann, V. (2017) The relevance of primary tumour location in patients with metastatic colorectal cancer: A meta-analysis of first-line clinical trials. European Journal of Cancer. 70: 87–98.

Honda, K., Ono, M., Shitashige, M., Masuda, M., Kamita, M., Miura, N. and Yamada, T. (2013) Proteomic approaches to the discovery of cancer biomarkers for early detection and personalized medicine. Japanese journal of clinical oncology. 43: 103– 109.

Hoshino, A., Costa-Silva, B., Shen, T.-L., Rodrigues, G., Hashimoto, A., Tesic Mark, M., Molina, H., Kohsaka, S., Di Giannatale, A., Ceder, S., Singh, S., Williams, C., Soplop, N., Uryu, K., Pharmer, L., King, T., Bojmar, L., Davies, A.E., Ararso, Y., et al. (2015) Tumour exosome integrins determine organotropic metastasis. Nature. 527: 329–335.

Hosseini-Beheshti, E., Pham, S., Adomat, H., Li, N. and Tomlinson Guns, E.S.

203

(2012) Exosomes as biomarker enriched microvesicles: characterization of exosomal proteins derived from a panel of prostate cell lines with distinct AR phenotypes. Molecular & cellular proteomics : MCP. 11: 863–885.

Hsu, H.-C., Thiam, T.K., Lu, Y.-J., Yeh, C.Y., Tsai, W.-S., You, J.F., Hung, H.Y., Tsai, C.-N., Hsu, A., Chen, H.-C., Chen, S.-J. and Yang, T.-S. (2016) Mutations of <i>KRAS/NRAS/BRAF</i> predict cetuximab resistance in metastatic colorectal cancer patients. Oncotarget. 7: 22257–22270.

Hu, Y., Yan, C., Mu, L., Huang, K., Li, X., Tao, D., Wu, Y. and Qin, J. (2015) Fibroblast-derived exosomes contribute to chemoresistance through priming cancer stem cells in colorectal cancer. PLoS ONE. 10:.

Huang, C.-H., Mandelker, D., Schmidt-Kittler, O., Samuels, Y., Velculescu, V.E., Kinzler, K.W., Vogelstein, B., Gabelli, S.B. and Amzel, L.M. (2007) The Structure of a Human p110 /p85 Complex Elucidates the Effects of Oncogenic PI3K Mutations. Science. 318: 1744–1748.

Huang, L., Liu, Z., Deng, D., Tan, A., Liao, M., Mo, Z. and Yang, X. (2014) Systematic review/Meta-analysis Anti-epidermal growth factor receptor monoclonal antibody-based therapy for metastatic colorectal cancer: a meta-analysis of the effect of PIK3CA mutations in KRAS wild-type patients. Archives of Medical Science. 1: 1– 9.

Huang, S., Li, C., Armstrong, E.A., Peet, C.R., Saker, J., Amler, L.C., Sliwkowski, M.X. and Harari, P.M. (2013) Dual Targeting of EGFR and HER3 with MEHD7945A Overcomes Acquired Resistance to EGFR Inhibitors and Radiation. Cancer Research. 73: 824–833.

Huang, Y.-Q., Ling, X.-H., Yuan, R.-Q., Chen, Z.-Y., Yang, S.-B., Huang, H.-X., Zhong, W.-D. and Qiu, S.-P. (2017) miR-30c suppresses prostate cancer survival by targeting the ASF/SF2 splicing factor oncoprotein. Molecular Medicine Reports. 16: 2431–2438.

Huang, Z., Huang, S., Wang, Q., Liang, L. and Ni, S. (2011) MicroRNA-95 promotes cell proliferation and targets sorting Nexin 1 in human colorectal carcinoma. Cancer research. 71: 2582–2589.

Hubbard, S.R. (2009) The Juxtamembrane Region of EGFR Takes Center Stage. Cell. 137: 1181–1183.

204

Huggett, J.F., Foy, C.A., Benes, V., Emslie, K., Garson, J.A., Haynes, R., Hellemans, J., Kubista, M., Mueller, R.D., Nolan, T., Pfaffl, M.W., Shipley, G.L., Vandesompele, J., Wittwer, C.T. and Bustin, S.A. (2013) The Digital MIQE Guidelines: Minimum Information for Publication of Quantitative Digital PCR Experiments. Clinical Chemistry. 59: 892–902.

Hurwitz, H., Fehrenbacher, L., Novotny, W., Cartwright, T., Hainsworth, J., Heim, W., Berlin, J., Baron, A., Griffing, S., Holmgren, E., Ferrara, N., Fyfe, G., Rogers, B., Ross, R. and Kabbinavar, F. (2004) Bevacizumab plus Irinotecan, Fluorouracil, and Leucovorin for Metastatic Colorectal Cancer. New England Journal of Medicine. 350: 2335–2342.

Hynes, N.E. and Lane, H.A. (2005) ERBB receptors and cancer: the complexity of targeted inhibitors. Nature Reviews Cancer. 5: 341–354.

Ichikawa, M., Williams, R., Wang, L., Vogl, T. and Srikrishna, G. (2011) S100A8/A9 activate key genes and pathways in colon tumor progression. Molecular cancer research : MCR. 9: 133–148.

Iida, M., Brand, T.M., Starr, M.M., Li, C., Huppert, E.J., Luthar, N., Pedersen, M.W., Horak, I.D., Kragh, M. and Wheeler, D.L. (2013) Sym004, a Novel EGFR Antibody Mixture, Can Overcome Acquired Resistance to Cetuximab. Neoplasia (New York, N.Y.). 15: 1196–1206.

Inoue, A., Okamoto, K., Fujino, Y., Nakagawa, T., Muguruma, N., Sannomiya, K., Mitsui, Y., Takaoka, T., Kitamura, S., Miyamoto, H., Okahisa, T., Fujimori, T., Imoto, I. and Takayama, T. (2015) B-RAF mutation and accumulated gene methylation in aberrant crypt foci (ACF), sessile serrated adenoma/polyp (SSA/P) and cancer in SSA/P. British Journal of Cancer. 112: 403–412.

Iwai, K., Minamisawa, T., Suga, K., Yajima, Y. and Shiba, K. (2016) Isolation of human salivary extracellular vesicles by iodixanol density gradient ultracentrifugation and their characterizations. Journal of Extracellular Vesicles. 5: 30829.

Jaiswal, B.S., Kljavin, N.M., Stawiski, E.W., Chan, E., Parikh, C., Durinck, S., Chaudhuri, S., Pujara, K., Guillory, J., Edgar, K.A., Janakiraman, V., Scholz, R.-P., Bowman, K.K., Lorenzo, M., Li, H., Wu, J., Yuan, W., Peters, B.A., Kan, Z., et al. (2013) Oncogenic ERBB3 Mutations in Human Cancers. Cancer Cell. 23: 603–617.

Jarry, J., Schadendorf, D., Greenwood, C., Spatz, A. and van Kempen, L.C. (2014)

205

The validity of circulating microRNAs in oncology: five years of challenges and contradictions. Molecular oncology. 8: 819–829.

Jeon, Y.-J., Middleton, J., Kim, T., Laganà, A., Piovan, C., Secchiero, P., Nuovo, G.J., Cui, R., Joshi, P., Romano, G., Di Leva, G., Lee, B.-K., Sun, H.-L., Kim, Y., Fadda, P., Alder, H., Garofalo, M. and Croce, C.M. (2015) A set of NF-κB-regulated microRNAs induces acquired TRAIL resistance in lung cancer. Proceedings of the National Academy of Sciences of the United States of America. 112: E3355-64.

Jhawer, M., Goel, S., Wilson, A.J., Montagna, C., Ling, Y.-H., Byun, D.-S., Nasser, S., Arango, D., Shin, J., Klampfer, L., Augenlicht, L.H., Soler, R.P. and Mariadason, J.M. (2008) PIK3CA Mutation/PTEN Expression Status Predicts Response of Colon Cancer Cells to the Epidermal Growth Factor Receptor Inhibitor Cetuximab. Cancer Research. 68: 1953–1961.

Ji, H., Greening, D., Barnes, T. and Lim, J. (2013) Proteome profiling of exosomes derived from human primary and metastatic colorectal cancer cells reveal differential expression of key metastatic factors and signal. Proteomics. 13: 1672–1686.

Jonker, D.J., O’Callaghan, C.J., Karapetis, C.S., Zalcberg, J.R., Tu, D., Au, H.-J., Berry, S.R., Krahn, M., Price, T., Simes, R.J., Tebbutt, N.C., van Hazel, G., Wierzbicki, R., Langer, C. and Moore, M.J. (2007) Cetuximab for the Treatment of Colorectal Cancer. New England Journal of Medicine. 357: 2040–2048.

Jorissen, R.N., Walker, F., Pouliot, N., Garrett, T.P.J., Ward, C.W. and Burgess, A.W. (2003) Epidermal growth factor receptor: mechanisms of activation and signalling. Experimental cell research. 284: 31–53.

Jura, N., Shan, Y. and Cao, X. (2009a) Structural analysis of the catalytically inactive kinase domain of the human EGF receptor 3. Proc Natl Acad Sci USA. 106: 21608– 21613.

Jura, N., Shan, Y., Cao, X., Shaw, D.E. and Kuriyan, J. (2009b) Structural analysis of the catalytically inactive kinase domain of the human EGF receptor 3. Proceedings of the National Academy of Sciences. 106: 21608–21613.

Kabbinavar, F., Irl, C., Zurlo, A. and Hurwitz, H. (2008) Bevacizumab Improves the Overall and Progression-Free Survival of Patients with Metastatic Colorectal Cancer Treated with 5-Fluorouracil-Based Regimens Irrespective of Baseline Risk. Oncology. 75: 215–223.

206

Källberg, E., Vogl, T., Liberg, D. and Olsson, A. (2012) S100A9 interaction with TLR4 promotes tumor growth. PloS one. 7:.

Karakas, B., Bachman, K.E. and Park, B.H. (2006) Mutation of the PIK3CA oncogene in human cancers. British Journal of Cancer. 94: 455–459.

Keating, G.M. (2010) Panitumumab. Drugs. 70: 1059–1078.

Kelleher, M.T., Fruhwirth, G., Patel, G., Ofo, E., Festy, F., Barber, P.R., Ameer-Beg, S.M., Vojnovic, B., Gillett, C., Coolen, A., Kéri, G., Ellis, P.A. and Ng, T. (2009) The potential of optical proteomic technologies to individualize prognosis and guide rational treatment for cancer patients. Targeted oncology. 4: 235–252.

Khalyfa, A., Almendros, I., Gileles-Hillel, A., Akbarpour, M., Trzepizur, W., Mokhlesi, B., Huang, L., Andrade, J., Farré, R. and Gozal, D. (2016) Circulating exosomes potentiate tumor malignant properties in a mouse model of chronic sleep fragmentation. Oncotarget. 7: 54676–54690.

Khelwatty, S.A., Essapen, S., Seddon, A.M. and Modjtahedi, H. (2013) Prognostic significance and targeting of HER family in colorectal cancer. Frontiers in bioscience (Landmark edition). 18: 394–421.

Kluftinger, A.M., Robinson, B.W., Quenville, N.F., Finley, R.J. and Davis, N.L. (1992) Correlation of epidermal growth factor receptor and c-erbB2 oncogene product to known prognostic indicators of colorectal cancer. Surgical oncology. 1: 97–105.

Knuefermann, C., Lu, Y., Liu, B., Jin, W., Liang, K., Wu, L., Schmidt, M., Mills, G.B., Mendelsohn, J. and Fan, Z. (2003) HER2/PI-3K/Akt activation leads to a multidrug resistance in human breast adenocarcinoma cells. Oncogene. 22: 3205–3212.

Komurasaki, T., Toyoda, H., Uchida, D. and Morimoto, S. (1997) Epiregulin binds to epidermal growth factor receptor and ErbB-4 and induces tryosine phosphorylation of epidermal growth factor receptor, ErbB-2, ErbB-3 and ErbB-4. Oncogene. 15: 2841– 2848.

Kopetz, S., Desai, J., Chan, E., Hecht, J.R., O’Dwyer, P.J., Maru, D., Morris, V., Janku, F., Dasari, A., Chung, W., Issa, J.-P.J., Gibbs, P., James, B., Powis, G., Nolop, K.B., Bhattacharya, S. and Saltz, L. (2015) Phase II Pilot Study of Vemurafenib in Patients With Metastatic BRAF -Mutated Colorectal Cancer. Journal of Clinical Oncology. 33: 4032–4038.

207

Kosaka, N., Iguchi, H. and Yoshioka, Y. (2010) Secretory mechanisms and intercellular transfer of microRNAs in living cells. Journal of Biological Chemistry. 285: 17442–17452.

Kosaka, N., Kogure, A., Yamamoto, T., Urabe, F., Usuba, W., Prieto-Vila, M. and Ochiya, T. (2019) Exploiting the message from cancer: the diagnostic value of extracellular vesicles for clinical applications Experimental and Molecular Medicine. 51 (3).

Kowal, J., Arras, G., Colombo, M., Jouve, M., Morath, J.P., Primdal-Bengtson, B., Dingli, F., Loew, D., Tkach, M. and Théry, C. (2016) Proteomic comparison defines novel markers to characterize heterogeneous populations of extracellular vesicle subtypes. Proceedings of the National Academy of Sciences. 113: E968–E977.

Kuchipudi, S. V, Tellabati, M., Nelli, R.K., White, G.A., Perez, B.B., Sebastian, S., Slomka, M.J., Brookes, S.M., Brown, I.H., Dunham, S.P. and Chang, K.-C. (2012) 18S rRNA is a reliable normalisation gene for real time PCR based on influenza virus infected cells. Virology journal. 9: 230.

Lagos-Quintana, M., Rauhut, R., Lendeckel, W. and Tuschl, T. (2001) Identification of novel genes coding for small expressed RNAs. Science. 294: 853–858.

Laulagnier, K., Motta, C. and Hamdi, S. (2004) Mast cell-and dendritic cell-derived exosomes display a specific lipid composition and an unusual membrane organization. The Biochemical journal. 380: 161–171.

Laurent-Puig, P., Cayre, A., Manceau, G., Buc, E., Bachet, J.-B., Lecomte, T., Rougier, P., Lievre, A., Landi, B., Boige, V., Ducreux, M., Ychou, M., Bibeau, F., Bouché, O., Reid, J., Stone, S. and Penault-Llorca, F. (2009) Analysis of PTEN , BRAF , and EGFR Status in Determining Benefit From Cetuximab Therapy in Wild- Type KRAS Metastatic Colon Cancer. Journal of Clinical Oncology. 27: 5924–5930.

Le, D.T., Uram, J.N., Wang, H., Bartlett, B.R., Kemberling, H., Eyring, A.D., Skora, A.D., Luber, B.S., Azad, N.S., Laheru, D., Biedrzycki, B., Donehower, R.C., Zaheer, A., Fisher, G.A., Crocenzi, T.S., Lee, J.J., Duffy, S.M., Goldberg, R.M., de la Chapelle, A., et al. (2015) PD-1 Blockade in Tumors with Mismatch-Repair Deficiency. New England Journal of Medicine. 372: 2509–2520.

Lee, C.-H., Im, E.-J., Moon, P.-G. and Baek, M.-C. (2018) Discovery of a diagnostic biomarker for colon cancer through proteomic profiling of small extracellular vesicles.

208

BMC Cancer. 18: 1058.

Lee, D., Yu, M., Lee, E., Kim, H., Yang, Y., Kim, K., Pannicia, C., Kurie, J.M. and Threadgill, D.W. (2009) Tumor-specific apoptosis caused by deletion of the ERBB3 pseudo-kinase in mouse intestinal epithelium. The Journal of clinical investigation. 119: 2702–2713.

Lee, J.C., Wang, S.-T., Chow, N.-H. and Yang, H.-B. (2002) Investigation of the prognostic value of coexpressed erbB family members for the survival of colorectal cancer patients after curative surgery. European journal of cancer (Oxford, England : 1990). 38: 1065–1071.

Lee, J.W., Stone, M.L., Porrett, P.M., Thomas, S.K., Komar, C.A., Li, J.H., Delman, D., Graham, K., Gladney, W.L., Hua, X., Black, T.A., Chien, A.L., Majmundar, K.S., Thompson, J.C., Yee, S.S., O’Hara, M.H., Aggarwal, C., Xin, D., Shaked, A., et al. (2019) Hepatocytes direct the formation of a pro-metastatic niche in the liver. Nature. 567: 249–252.

Lee, R.C., Feinbaum, R.L. and Ambros, V. (1993) The C. elegans heterochronic gene lin-4 encodes small RNAs with antisense complementarity to lin-14. Cell. 75: 843–854.

Lemmon, M.A. (2009) Ligand-induced ErbB receptor dimerization Experimental Cell Research. 315 (4) pp. 638–648.

Lemmon, M.A. and Schlessinger, J. (2010) Cell signaling by receptor tyrosine kinases. Cell. 141: 1117–1134.

Lemmon, M.A., Schlessinger, J. and Ferguson, K.M. (2014) The EGFR Family: Not So Prototypical Receptor Tyrosine Kinases. Cold Spring Harbor Perspectives in Biology. 6: a020768–a020768.

Li, C., Zhao, L., Chen, Y., He, T., Chen, X., Mao, J., Li, C., Lyu, J. and Meng, Q.H. (2016) MicroRNA-21 promotes proliferation, migration, and invasion of colorectal cancer, and tumor growth associated with down-regulation of sec23a expression. BMC cancer. 16: 605.

Li, J., Qin, S., Yau, T., Ma, B., Pan, H., Xu, J., Bai, Y., Chi, Y., Wang, L., Yeh, K.H., Bi, F., Cheng, Y., Le, A.T., Lin, J.K., Liu, T., Ma, D., Kappeler, C., Kalmus, J., Xu, R.H., et al. (2014) O-0023 * CONCUR: A RANDOMIZED, DOUBLE-BLIND, PLACEBO-CONTROLLED PHASE 3 STUDY OF REGORAFENIB MONOTHERAPY

209

IN ASIAN PATIENTS WITH PREVIOUSLY TREATED METASTATIC COLORECTAL CANCER (MCRC). Annals of Oncology. 25: ii114–ii115.

Li, J., Yao, B., Huang, H., Wang, Z., Sun, C., Fan, Y., Chang, Q., Li, S., Wang, X. and Xi, J. (2009) Real-Time Polymerase Chain Reaction MicroRNA Detection Based on Enzymatic Stem-Loop Probes Ligation. Analytical Chemistry. 81: 5446–5451.

Li, P., Kaslan, M., Lee, S.H., Yao, J. and Gao, Z. (2017) Progress in Exosome Isolation Techniques. Theranostics. 7: 789.

Li, S., Schmitz, K.R., Jeffrey, P.D., Wiltzius, J.J.W., Kussie, P. and Ferguson, K.M. (2005) Structural basis for inhibition of the epidermal growth factor receptor by cetuximab. Cancer cell. 7: 301–311.

Li, X., Corbett, A.L., Taatizadeh, E., Tasnim, N., Little, J.P., Garnis, C., Daugaard, M., Guns, E., Hoorfar, M. and Li, I.T.S. (2019) Challenges and opportunities in exosome research—Perspectives from biology, engineering, and cancer therapy. APL Bioengineering. 3: 011503.

Ling, H., Zhang, W. and Calin, G.A. (2011) Principles of microRNA involvement in human cancers. Chinese journal of cancer. 30: 739–748.

Liska, D., Chen, C.-T., Bachleitner-Hofmann, T., Christensen, J.G. and Weiser, M.R. (2011) HGF Rescues Colorectal Cancer Cells from EGFR Inhibition via MET Activation. Clinical Cancer Research. 17: 472–482.

Liu, B., Fang, M., Lu, Y., Mendelsohn, J. and Fan, Z. (2001) Fibroblast growth factor and insulin-like growth factor differentially modulate the apoptosis and G1 arrest induced by anti-epidermal growth factor receptor monoclonal antibody. Oncogene. 20: 1913–1922.

Liu, Y., Yu, X.-F., Zou, J. and Luo, Z.-H. (2015) Prognostic value of c-Met in colorectal cancer: A meta-analysis. World Journal of Gastroenterology. 21: 3706.

Livshits, M.A., Khomyakova, E., Evtushenko, E.G., Lazarev, V.N., Kulemin, N.A., Semina, S.E., Generozov, E. V. and Govorun, V.M. (2015) Isolation of exosomes by differential centrifugation: Theoretical analysis of a commonly used protocol. Scientific Reports. 5: 17319.

Ljuslinder, I., Malmer, B., Isaksson-Mettävainio, M., Oberg, A., Henriksson, R., Stenling, R. and Palmqvist, R. (2009) ErbB 1-4 expression alterations in primary

210 colorectal cancers and their corresponding metastases. Anticancer research. 29: 1489–1494.

Llorente, A., van Deurs, B. and Sandvig, K. (2007) Cholesterol regulates prostasome release from secretory lysosomes in PC-3 human prostate cancer cells. European Journal of Cell Biology. 86: 405–415.

Longley, D.B., Harkin, D.P. and Johnston, P.G. (2003) 5-fluorouracil: mechanisms of action and clinical strategies. Nature reviews. Cancer. 3: 330–338.

Ludwig, J.A. and Weinstein, J.N. (2005) Biomarkers in cancer staging, prognosis and treatment selection Nature Reviews Cancer. 5 (11) pp. 845–856.

Macleod, K.G., Serrels, B. and Carragher, N.O. (2017) Reverse Phase Protein Arrays and Drug Discovery. In: Methods in molecular biology (Clifton, N.J.). pp. 153– 169.

Mahdipour, M., van Tol, H.T.A., Stout, T.A.E. and Roelen, B.A.J. (2015) Validating reference microRNAs for normalizing qRT-PCR data in bovine oocytes and preimplantation embryos. BMC Developmental Biology. 15: 25.

Malinowsky, K., Raychaudhuri, M., Buchner, T., Thulke, S., Wolff, C., Höfler, H., Becker, K.-F. and Avril, S. (2012) Common protein biomarkers assessed by reverse phase protein arrays show considerable intratumoral heterogeneity in breast cancer tissues. PloS one. 7: e40285.

Malinowsky, K., Wolff, C., Ergin, B., Berg, D. and Becker, K.F. (2010) Deciphering signaling pathways in clinical tissues for personalized medicine using protein microarrays. Journal of cellular physiology. 225: 364–370.

Manne, U., Shanmugam, C., Katkoori, V.R., Bumpers, H.L. and Grizzle, W.E. (2012) Development and progression of colorectal neoplasia. In: Translational Pathology of Early Cancer. IOS Press. pp. 235–265.

Mao, C., Yang, Z.Y., Hu, X.F., Chen, Q. and Tang, J.L. (2012) PIK3CA exon 20 mutations as a potential biomarker for resistance to anti-EGFR monoclonal antibodies in KRAS wild-type metastatic colorectal cancer: a systematic review and meta-analysis. Annals of Oncology. 23: 1518–1525.

Mao, M., Tian, F., Mariadason, J.M., Tsao, C.C., Lemos, R., Dayyani, F., Gopal, Y.N. V., Jiang, Z.-Q., Wistuba, I.I., Tang, X.M., Bornman, W.G., Bollag, G., Mills, G.B.,

211

Powis, G., Desai, J., Gallick, G.E., Davies, M.A. and Kopetz, S. (2013) Resistance to BRAF Inhibition in BRAF-Mutant Colon Cancer Can Be Overcome with PI3K Inhibition or Demethylating Agents. Clinical Cancer Research. 19: 657–667.

Marenholz, I., Heizmann, C.W. and Fritz, G. (2004) S100 proteins in mouse and man: from evolution to function and pathology (including an update of the nomenclature). Biochemical and biophysical research communications. 322: 1111– 1122.

Markowitz, S.D. and Bertagnolli, M.M. (2009) Molecular origins of cancer: Molecular basis of colorectal cancer. The New England journal of medicine. 361: 2449–2460.

Marwan G. Fakih, M.A.P.M. (2006) CEA Monitoring in Colorectal Cancer.

Mathivanan, S. and Fahner, C. (2012) ExoCarta 2012: database of exosomal proteins, RNA and lipids. Nucleic acids researcu. 40: D1241-4.

Mathivanan, S., Ji, H. and Simpson, R. (2010) Exosomes: extracellular organelles important in intercellular communication. Journal of proteomics. 73: 1907–1920.

Mathivanan, S. and Simpson, R. (2009) ExoCarta: A compendium of exosomal proteins and RNA. Proteomics. 9: 4997–5000.

Matsushita, S., Nitanda, T., Furukawa, T., Sumizawa, T., Tani, A., Nishimoto, K., Akiba, S., Miyadera, K., Fukushima, M., Yamada, Y., Yoshida, H., Kanzaki, T. and Akiyama, S. (1999) The effect of a thymidine phosphorylase inhibitor on angiogenesis and apoptosis in tumors. Cancer research. 59: 1911–1916.

Maughan, T.S., Adams, R.A., Smith, C.G., Meade, A.M., Seymour, M.T., Wilson, R.H., Idziaszczyk, S., Harris, R., Fisher, D., Kenny, S.L., Kay, E., Mitchell, J.K., Madi, A., Jasani, B., James, M.D., Bridgewater, J., Kennedy, M.J., Claes, B., Lambrechts, D., et al. (2011) Addition of cetuximab to oxaliplatin-based first-line combination chemotherapy for treatment of advanced colorectal cancer: results of the randomised phase 3 MRC COIN trial. The Lancet. 377: 2103–2114.

Maurer, C.A., Friess, H., Kretschmann, B., Zimmermann, A., Stauffer, A., Baer, H.U., Korc, M. and Büchler, M.W. (1998) Increased expression of erbB3 in colorectal cancer is associated with concomitant increase in the level of erbB2. Human pathology. 29: 771–777.

Mayer, R.J., Van Cutsem, E., Falcone, A., Yoshino, T., Garcia-Carbonero, R.,

212

Mizunuma, N., Yamazaki, K., Shimada, Y., Tabernero, J., Komatsu, Y., Sobrero, A., Boucher, E., Peeters, M., Tran, B., Lenz, H.-J., Zaniboni, A., Hochster, H., Cleary, J.M., Prenen, H., et al. (2015) Randomized Trial of TAS-102 for Refractory Metastatic Colorectal Cancer. New England Journal of Medicine. 372: 1909–1919.

Meng, F., Henson, R., Wehbe-Janek, H., Ghoshal, K., Jacob, S.T. and Patel, T. (2007) MicroRNA-21 regulates expression of the PTEN tumor suppressor gene in human hepatocellular cancer. Gastroenterology. 133: 647–658.

Meyerhardt, J.A. and Mayer, R.J. (2005) Systemic Therapy for Colorectal Cancer. New England Journal of Medicine. 352: 476–487.

Meyers, M., Wagner, M.W., Mazurek, A., Schmutte, C., Fishel, R. and Boothman, D.A. (2005) DNA mismatch repair-dependent response to fluoropyrimidine-generated damage. The Journal of biological chemistry. 280: 5516–5526.

Meza, R., Jeon, J., Renehan, A.G. and Luebeck, E.G. (2010) Colorectal Cancer Incidence Trends in the United States and United Kingdom: Evidence of Right- to Left-Sided Biological Gradients with Implications for Screening. Cancer Research. 70: 5419–5429.

Mills, G.B. and Yarden, Y. (2010) The rebirth of a phoenix: ovarian cancers are addicted to ErbB-3. Cancer cell. 17: 217–218.

Misale, S., Yaeger, R., Hobor, S., Scala, E., Janakiraman, M., Liska, D., Valtorta, E., Schiavo, R., Buscarino, M., Siravegna, G., Bencardino, K., Cercek, A., Chen, C.-T., Veronese, S., Zanon, C., Sartore-Bianchi, A., Gambacorta, M., Gallicchio, M., Vakiani, E., et al. (2012) Emergence of KRAS mutations and acquired resistance to anti-EGFR therapy in colorectal cancer. Nature. 486: 532–536.

Missiaglia, E., Jacobs, B., D’Ario, G., Di Narzo, A.F., Soneson, C., Budinska, E., Popovici, V., Vecchione, L., Gerster, S., Yan, P., Roth, A.D., Klingbiel, D., Bosman, F.T., Delorenzi, M. and Tejpar, S. (2014) Distal and proximal colon cancers differ in terms of molecular, pathological, and clinical features. Annals of Oncology. 25: 1995– 2001.

Mitchell, J., Court, J. and Mason, M. (2008) Increased exosome production from tumour cell cultures using the Integra CELLine Culture System. Journal of immunological methods. 335: 98–105.

Mittelbrunn, M. (2011) Unidirectional transfer of microRNA-loaded exosomes from T

213 cells to antigen-presenting cells. Nature communications. 2:.

Modjtahedi, H. and Essapen, S. (2009) Epidermal growth factor receptor inhibitors in cancer treatment: advances, challenges and opportunities. Anti-cancer drugs. 20: 851–855.

Montagut, C., Argilés, G., Ciardiello, F., Poulsen, T.T., Dienstmann, R., Kragh, M., Kopetz, S., Lindsted, T., Ding, C., Vidal, J., Clausell-Tormos, J., Siravegna, G., Sánchez-Martín, F.J., Koefoed, K., Pedersen, M.W., Grandal, M.M., Dvorkin, M., Wyrwicz, L., Rovira, A., et al. (2018) Efficacy of Sym004 in Patients With Metastatic Colorectal Cancer With Acquired Resistance to Anti-EGFR Therapy and Molecularly Selected by Circulating Tumor DNA Analyses. JAMA Oncology. 4: e175245.

Montagut, C., Dalmases, A., Bellosillo, B., Crespo, M., Pairet, S., Iglesias, M., Salido, M., Gallen, M., Marsters, S., Tsai, S.P., Minoche, A., Somasekar, S., Serrano, S., Himmelbauer, H., Bellmunt, J., Rovira, A., Settleman, J., Bosch, F. and Albanell, J. (2012) Identification of a mutation in the extracellular domain of the Epidermal Growth Factor Receptor conferring cetuximab resistance in colorectal cancer. Nature Medicine. 18: 221–223.

Moon, A., Yong, H. and Song, J. (2008) Global gene expression profiling unveils S100A8/A9 as candidate markers in H-ras-mediated human breast epithelial cell invasion. Molecular cancer research : MCR. 6: 1544–1553.

Mu, Y. and Su, X. (2012) Polymorphism in pre-miR-30c contributes to gastric cancer risk in a Chinese population. Medical Oncology. 29: 1723–1732.

Mu, Z., Klinowska, T., Dong, X., Foster, E., Womack, C., Fernandez, S. V and Cristofanilli, M. (2014) AZD8931, an equipotent, reversible inhibitor of signaling by epidermal growth factor receptor (EGFR), HER2, and HER3: preclinical activity in HER2 non-amplified inflammatory breast cancer models. Journal of experimental & clinical cancer research. 33: 47.

Mulcahy, L.A., Pink, R.C. and Carter, D.R.F. (2014) Routes and mechanisms of extracellular vesicle uptake. Journal of Extracellular Vesicles. 3: 24641.

Muller, L., Hong, C.-S., Stolz, D.B., Watkins, S.C. and Whiteside, T.L. (2014) Isolation of biologically-active exosomes from human plasma. Journal of immunological methods. 411: 55–65.

Murtaza, M., Dawson, S.-J., Tsui, D.W.Y., Gale, D., Forshew, T., Piskorz, A.M.,

214

Parkinson, C., Chin, S.-F., Kingsbury, Z., Wong, A.S.C., Marass, F., Humphray, S., Hadfield, J., Bentley, D., Chin, T.M., Brenton, J.D., Caldas, C. and Rosenfeld, N. (2013) Non-invasive analysis of acquired resistance to cancer therapy by sequencing of plasma DNA. Nature. 497: 108–112.

Nedaeinia, R., Manian, M., Jazayeri, M.H., Ranjbar, M., Salehi, R., Sharifi, M., Mohaghegh, F., Goli, M., Jahednia, S.H., Avan, A. and Ghayour-Mobarhan, M. (2017) Circulating exosomes and exosomal microRNAs as biomarkers in gastrointestinal cancer Cancer Gene Therapy. 24 (2) pp. 48–56.

Ng, T., Flores-Borja, F., Alfano, G., Barber, P., Weitsman, G., Wong, F., Vicencio, J., Galazi, M., Doyle, J., Greenberg, J., Forster, M.D. and Coolen, A. (2018) The use of exosome and immune profiling to analyze a phase 2 study on the addition of patritumab or placebo to cetuximab and a platinum agent for recurrent / metastatic head and neck cancer (R/M HNSCC) patients. Journal of Clinical Oncology. 36: 6043–6043.

Ng, T., Squire, A., Hansra, G., Bornancin, F., Prevostel, C., Hanby, A., Harris, W., Barnes, D., Schmidt, S., Mellor, H., Bastiaens, P. and Parker, P. (1999) Imaging protein kinase Cαlpha activation in cells. Science. 283: 2085–2089.

Di Nicolantonio, F., Martini, M., Molinari, F., Sartore-Bianchi, A., Arena, S., Saletti, P., De Dosso, S., Mazzucchelli, L., Frattini, M., Siena, S. and Bardelli, A. (2008) Wild- Type BRAF Is Required for Response to Panitumumab or Cetuximab in Metastatic Colorectal Cancer. Journal of Clinical Oncology. 26: 5705–5712.

Niel, G. van, Porto-Carreiro, I., Simoes, S. and Paposo, G. (2006) Exosomes: a common pathway for a specialized function. Journal of biochemistry. 140: 13–21.

Nosho, K., Kawasaki, T., Longtine, J.A., Fuchs, C.S., Ohnishi, M., Suemoto, Y., Kirkner, G.J., Zepf, D., Yan, L. and Ogino, S. (2008) PIK3CA Mutation in Colorectal Cancer: Relationship with Genetic and Epigenetic Alterations. Neoplasia. 10: 534– 541.

Nuciforo, P., Radosevic-Robin, N., Ng, T. and Scaltriti, M. (2015) Quantification of HER family receptors in breast cancer Breast Cancer Research. 17 (1).

Oda, K., Matsuoka, Y., Funahashi, A. and Kitano, H. (2005) A comprehensive pathway map of epidermal growth factor receptor signaling. Molecular systems biology. 1: 2005.0010.

215

Ogata-kawata, H., Izumiya, M., Kurioka, D., Honma, Y. and Yamada, Y. (2014) Circulating Exosomal microRNAs as Biomarkers of Colon Cancer. 9:.

Ogata-Kawata, H., Izumiya, M., Kurioka, D., Honma, Y., Yamada, Y., Furuta, K., Gunji, T., Ohta, H., Okamoto, H., Sonoda, H., Watanabe, M., Nakagama, H., Yokota, J., Kohno, T. and Tsuchiya, N. (2014) Circulating exosomal microRNAs as biomarkers of colon cancer. PLoS ONE. 9:.

Ogiso, H., Ishitani, R., Nureki, O., Fukai, S., Yamanaka, M., Kim, J.-H., Saito, K., Sakamoto, A., Inoue, M., Shirouzu, M. and Yokoyama, S. (2002) Crystal structure of the complex of human epidermal growth factor and receptor extracellular domains. Cell. 110: 775–787.

Ooi, A., Takehana, T., Li, X., Suzuki, S., Kunitomo, K., Iino, H., Fujii, H., Takeda, Y. and Dobashi, Y. (2004) Protein overexpression and gene amplification of HER-2 and EGFR in colorectal cancers: an immunohistochemical and fluorescent in situ hybridization study. Modern pathology : an official journal of the United States and Canadian Academy of Pathology, Inc. 17: 895–904.

Ostrand-Rosenberg, S. (2008) Cancer and complement. Nature biotechnology. 26: 1348–1349.

Pan, B. and Johnstone, R. (1983) Fate of the transferrin receptor during maturation of sheep reticulocytes in vitro: selective externalization of the receptor. Cell. 33: 967– 978.

Panteleev, M.A., Abaeva, A.A., Balandina, A.N., Belyaev, A. V., Nechipurenko, D.Y., Obydennyi, S.I., Sveshnikova, A.N., Shibeko, A.M. and Ataullakhanov, F.I. (2017) Extracellular vesicles of blood plasma: content, origin, and properties. Biochemistry (Moscow), Supplement Series A: Membrane and Cell Biology. 11: 187–192.

Patel, D., Bassi, R., Hooper, A., Prewett, M., Hicklin, D.J. and Kang, X. (2009) Anti- epidermal growth factor receptor monoclonal antibody cetuximab inhibits EGFR/HER-2 heterodimerization and activation. International journal of oncology. 34: 25–32.

Peacock, O., Lee, A.C., Cameron, F., Tarbox, R., Vafadar-Isfahani, N., Tufarelli, C. and Lund, J.N. (2014) Inflammation and MiR-21 Pathways Functionally Interact to Downregulate PDCD4 in Colorectal Cancer Kalpana Ghoshal (ed.). PLoS ONE. 9: e110267.

216

Pedersen, M.W., Jacobsen, H.J., Koefoed, K., Hey, A., Pyke, C., Haurum, J.S. and Kragh, M. (2010) Sym004: A novel synergistic anti-epidermal growth factor receptor antibody mixture with superior anticancer efficacy. Cancer Research. 70: 588–597.

Peeters, M., Price, T.J., Cervantes, A., Sobrero, A.F., Ducreux, M., Hotko, Y., André, T., Chan, E., Lordick, F., Punt, C.J.A., Strickland, A.H., Wilson, G., Ciuleanu, T.-E., Roman, L., Van Cutsem, E., Tzekova, V., Collins, S., Oliner, K.S., Rong, A., et al. (2010) Randomized Phase III Study of Panitumumab With Fluorouracil, Leucovorin, and Irinotecan (FOLFIRI) Compared With FOLFIRI Alone As Second-Line Treatment in Patients With Metastatic Colorectal Cancer. Journal of Clinical Oncology. 28: 4706–4713.

Peinado, H., Alečković, M., Lavotshkin, S., Matei, I., Costa-Silva, B., Moreno-Bueno, G., Hergueta-Redondo, M., Williams, C., García-Santos, G., Ghajar, C.M., Nitadori- Hoshino, A., Hoffman, C., Badal, K., Garcia, B.A., Callahan, M.K., Yuan, J., Martins, V.R., Skog, J., Kaplan, R.N., et al. (2012) Melanoma exosomes educate bone marrow progenitor cells toward a pro-metastatic phenotype through MET. Nature medicine. 18: 883–891.

Peipp, M., Dechant, M. and Valerius, T. (2008) Effector mechanisms of therapeutic antibodies against ErbB receptors. Current opinion in immunology. 20: 436–443.

Pentheroudakis, G. and Twelves, C. (2002) The rational development of capecitabine from the laboratory to the clinic Anticancer Research. 22 (6 B) pp. 3589–3596.

Perez, E.A., Thompson, E.A., Ballman, K. V., Anderson, S.K., Asmann, Y.W., Kalari, K.R., Eckel-Passow, J.E., Dueck, A.C., Tenner, K.S., Jen, J., Fan, J.-B., Geiger, X.J., McCullough, A.E., Chen, B., Jenkins, R.B., Sledge, G.W., Winer, E.P., Gralow, J.R. and Reinholz, M.M. (2015) Genomic Analysis Reveals That Immune Function Genes Are Strongly Linked to Clinical Outcome in the North Central Cancer Treatment Group N9831 Adjuvant Trastuzumab Trial. Journal of Clinical Oncology. 33: 701– 708.

Perrotte, P., Matsumoto, T., Inoue, K., Kuniyasu, H., Eve, B.Y., Hicklin, D.J., Radinsky, R. and Dinney, C.P. (1999) Anti-epidermal growth factor receptor antibody C225 inhibits angiogenesis in human transitional cell carcinoma growing orthotopically in nude mice. Clinical cancer research : an official journal of the American Association for Cancer Research. 5: 257–265.

Petrelli, F., Tomasello, G., Borgonovo, K., Ghidini, M., Turati, L., Dallera, P.,

217

Passalacqua, R., Sgroi, G. and Barni, S. (2017) Prognostic Survival Associated With Left-Sided vs Right-Sided Colon Cancer. JAMA Oncology. 3: 211.

Petrelli, N., Douglass, H.O., Herrera, L., Russell, D., Stablein, D.M., Bruckner, H.W., Mayer, R.J., Schinella, R., Green, M.D. and Muggia, F.M. (1989) The modulation of fluorouracil with leucovorin in metastatic colorectal carcinoma: a prospective randomized phase III trial. Gastrointestinal Tumor Study Group. Journal of Clinical Oncology. 7: 1419–1426.

Pinheiro, L.B., Coleman, V.A., Hindson, C.M., Herrmann, J., Hindson, B.J., Bhat, S. and Emslie, K.R. (2012) Evaluation of a Droplet Digital Polymerase Chain Reaction Format for DNA Copy Number Quantification. Analytical Chemistry. 84: 1003–1011. van der Pol, E., Coumans, F.A.W., Grootemaat, A.E., Gardiner, C., Sargent, I.L., Harrison, P., Sturk, A., van Leeuwen, T.G. and Nieuwland, R. (2014) Particle size distribution of exosomes and microvesicles determined by transmission electron microscopy, flow cytometry, nanoparticle tracking analysis, and resistive pulse sensing. Journal of thrombosis and haemostasis : JTH. 12: 1182–1192.

Pommier, Y. (2006) Topoisomerase I inhibitors: camptothecins and beyond. Nature Reviews Cancer. 6: 789–802.

Powell, A.G.M.T., Wallace, R., McKee, R.F., Anderson, J.H., Going, J.J., Edwards, J. and Horgan, P.G. (2012) The relationship between tumour site, clinicopathological characteristics and cancer-specific survival in patients undergoing surgery for colorectal cancer. Colorectal Disease. 14: 1493–1499.

Prahallad, A., Sun, C., Huang, S., Di Nicolantonio, F., Salazar, R., Zecchin, D., Beijersbergen, R.L., Bardelli, A. and Bernards, R. (2012) Unresponsiveness of colon cancer to BRAF(V600E) inhibition through feedback activation of EGFR. Nature. 483: 100–103.

Price, T.J., Newhall, K., Peeters, M., Kim, T.W., Li, J., Cascinu, S., Ruff, P., Suresh, A.V., Thomas, A.L., Tjulandin, S., Boedigheimer, M., Zhang, K., Sidhu, R. and Murugappan, S. (2015) Prevalence and outcomes of patients (pts) with EGFR S492R ectodomain mutations in ASPECCT: Panitumumab (pmab) vs cetuximab (cmab) in pts with chemorefractory wild-type KRAS exon 2 metastatic colorectal cancer (mCRC). Journal of Clinical Oncology. 33: e14623–e14623.

Price, T.J., Peeters, M., Kim, T.W., Li, J., Cascinu, S., Ruff, P., Suresh, A.S.,

218

Thomas, A., Tjulandin, S., Zhang, K., Murugappan, S. and Sidhu, R. (2014) Panitumumab versus cetuximab in patients with chemotherapy-refractory wild-type KRAS exon 2 metastatic colorectal cancer (ASPECCT): a randomised, multicentre, open-label, non-inferiority phase 3 study. The Lancet Oncology. 15: 569–579.

Putz, U., Howitt, J., Doan, A., Goh, C., Low, L., Silke, J. and SS Tan (2012) The tumor suppressor PTEN is exported in exosomes and has phosphatase activity in recipient cells. Science signalling. 5: ra70.

Queralt, B., Cuyàs, E., Bosch-Barrera, J., Massaguer, A., de Llorens, R., Martin- Castillo, B., Brunet, J., Salazar, R. and Menendez, J.A. (2016) Synthetic lethal interaction of cetuximab with MEK1/2 inhibition in NRAS-mutant metastatic colorectal cancer. Oncotarget. 7: 82185–82199.

Rabik, C.A. and Dolan, M.E. (2007) Molecular mechanisms of resistance and toxicity associated with platinating agents. Cancer treatment reviews. 33: 9–23.

Rački, N., Dreo, T., Gutierrez-Aguirre, I., Blejec, A. and Ravnikar, M. (2014) Reverse transcriptase droplet digital PCR shows high resilience to PCR inhibitors from plant, soil and water samples. Plant Methods. 10: 42.

Ragusa, M., Statello, L., Maugeri, M., Barbagallo, C., Passanisi, R., Alhamdani, M.S., Li Destri, G., Cappellani, A., Barbagallo, D., Scalia, M., Valadi, H., Hoheisel, J.D., Di Pietro, C. and Purrello, M. (2014) Highly skewed distribution of miRNAs and proteins between colorectal cancer cells and their exosomes following Cetuximab treatment: biomolecular, genetic and translational implications. Oncoscience. 1: 132–157.

Raiborg, C. and Stenmark, H. (2009) The ESCRT machinery in endosomal sorting of ubiquitylated membrane proteins. Nature. 458: 445–452.

Rapkiewicz, A., Espina, V., Zujewski, J.A., Lebowitz, P.F., Filie, A., Wulfkuhle, J., Camphausen, K., Petricoin, E.F., Liotta, L.A. and Abati, A. (2007) The needle in the haystack: application of breast fine-needle aspirate samples to quantitative protein microarray technology. Cancer. 111: 173–184.

Raposo, G. and Stoorvogel, W. (2013) Extracellular vesicles: Exosomes, microvesicles, and friends Journal of Cell Biology. 200 (4) pp. 373–383.

Raymond, E., Chaney, S.G., Taamma, A. and Cvitkovic, E. (1998) Oxaliplatin: a review of preclinical and clinical studies. Annals of oncology : official journal of the European Society for Medical Oncology. 9: 1053–1071.

219

Raymond, E., Faivre, S., Chaney, S., Woynarowski, J. and Cvitkovic, E. (2002) Cellular and molecular pharmacology of oxaliplatin Molecular Cancer Therapeutics. 1 (3) pp. 227–235.

Reiter, J.L., Threadgill, D.W., Eley, G.D., Strunk, K.E., Danielsen, A.J., Schehl Sinclair, C., Pearsall, R.S., Green, P.J., Yee, D., Lampland, A.L., Balasubramaniam, S., Crossley, T.D., Magnuson, T.R., James, C.D. and Maihle, N.J. (2001) Comparative Genomic Sequence Analysis and Isolation of Human and Mouse Alternative EGFR Transcripts Encoding Truncated Receptor Isoforms. Genomics. 71: 1–20.

Di Renzo, M.F., Olivero, M., Giacomini, A., Porte, H., Chastre, E., Mirossay, L., Nordlinger, B., Bretti, S., Bottardi, S. and Giordano, S. (1995) Overexpression and amplification of the met/HGF receptor gene during the progression of colorectal cancer. Clinical cancer research : an official journal of the American Association for Cancer Research. 1: 147–154.

Ressine, A., Corin, I., Järås, K., Guanti, G., Simone, C., Marko-Varga, G. and Laurell, T. (2007) Porous silicon surfaces: a candidate substrate for reverse protein arrays in cancer biomarker detection. Electrophoresis. 28: 4407–4415.

Reza, A.M.M.T., Cho, S.-K., Choi, Y.-J., Hong, K. and Kim, J.-H. (2018) Microarray profiling of miRNA and mRNA expression in Rag2 knockout and wild-type mouse spleens. Scientific Data. 5: 170199.

Ribeiro, M.F., Zhu, H., Millard, R.W. and Fan, G.C. (2013) Exosomes function in pro- and anti-angiogenesis Current Angiogenesis. 2 (1) pp. 54–59.

Ritter, C.A., Perez-Torres, M., Rinehart, C., Guix, M., Dugger, T., Engelman, J.A. and Arteaga, C.L. (2007) Human breast cancer cells selected for resistance to trastuzumab in vivo overexpress epidermal growth factor receptor and ErbB ligands and remain dependent on the ErbB receptor network. Clinical cancer research : an official journal of the American Association for Cancer Research. 13: 4909–4919.

Roepstorff, K., Grandal, M.V., Henriksen, L., Knudsen, S.L.J., Lerdrup, M., Grøvdal, L., Willumsen, B.M. and van Deurs, B. (2009) Differential effects of EGFR ligands on endocytic sorting of the receptor. Traffic (Copenhagen, Denmark). 10: 1115–1127.

De Roock, W., Claes, B., Bernasconi, D., De Schutter, J., Biesmans, B., Fountzilas, G., Kalogeras, K.T., Kotoula, V., Papamichael, D., Laurent-Puig, P., Penault-Llorca,

220

F., Rougier, P., Vincenzi, B., Santini, D., Tonini, G., Cappuzzo, F., Frattini, M., Molinari, F., Saletti, P., et al. (2010) 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. The Lancet Oncology. 11: 753–762.

Roock, W. De, Vriendt, V. De, Normanno, N., Ciardiello, F. and Tejpar, S. (2011) KRAS, BRAF, PIK3CA, and PTEN mutations: implications for targeted therapies in metastatic colorectal cancer. The Lancet Oncology. 12: 594–603.

Rosell, R., Wei, J. and Taron, M. (2009) Circulating MicroRNA Signatures of Tumor- Derived Exosomes for Early Diagnosis of Non–Small-Cell Lung Cancer. Clinical lung cancer. 10: 8–9.

Roth, C., Rack, B., Müller, V., Janni, W., Pantel, K. and Schwarzenbach, H. (2010) Circulating microRNAs as blood-based markers for patients with primary and metastatic breast cancer. Breast Cancer Research. 12: R90.

Roth, J., Vogl, T., Sorg, C. and Sunderkötter, C. (2003) Phagocyte-specific S100 proteins: a novel group of proinflammatory molecules. Trends in Immunology. 24: 155–158.

Sadanandam, A., Lyssiotis, C.A., Homicsko, K., Collisson, E.A., Gibb, W.J., Wullschleger, S., Ostos, L.C.G., Lannon, W.A., Grotzinger, C., Del Rio, M., Lhermitte, B., Olshen, A.B., Wiedenmann, B., Cantley, L.C., Gray, J.W. and Hanahan, D. (2013) A colorectal cancer classification system that associates cellular phenotype and responses to therapy. Nature Medicine. 19: 619–625.

Sahin, U., Weskamp, G., Kelly, K., Zhou, H.-M., Higashiyama, S., Peschon, J., Hartmann, D., Saftig, P. and Blobel, C.P. (2004) Distinct roles for ADAM10 and ADAM17 in ectodomain shedding of six EGFR ligands. The Journal of Cell Biology. 164: 769–779.

Saltz, L.B., Clarke, S., Díaz-Rubio, E., Scheithauer, W., Figer, A., Wong, R., Koski, S., Lichinitser, M., Yang, T.-S., Rivera, F., Couture, F., Sirzén, F. and Cassidy, J. (2008) Bevacizumab in Combination With Oxaliplatin-Based Chemotherapy As First- Line Therapy in Metastatic Colorectal Cancer: A Randomized Phase III Study. Journal of Clinical Oncology. 26: 2013–2019.

Saltz, L.B., Cox, J. V., Blanke, C., Rosen, L.S., Fehrenbacher, L., Moore, M.J.,

221

Maroun, J.A., Ackland, S.P., Locker, P.K., Pirotta, N., Elfring, G.L. and Miller, L.L. (2000) Irinotecan plus Fluorouracil and Leucovorin for Metastatic Colorectal Cancer. New England Journal of Medicine. 343: 905–914.

Saltz, L.B., Meropol, N.J., Loehrer, P.J., Needle, M.N., Kopit, J. and Mayer, R.J. (2004) Phase II trial of cetuximab in patients with refractory colorectal cancer that expresses the epidermal growth factor receptor. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 22: 1201–1208.

Sayed, D., Rane, S., Lypowy, J., He, M., Chen, I.-Y., Vashistha‡, H., Yan, L., Malhotra, A., Vatner, D. and Abdellatif, M. (2008) MicroRNA-21 targets Sprouty2 and promotes cellular outgrowths. Molecular Biology of the Cell. 19: 3272–3282.

Scartozzi, M., Mandolesi, A., Giampieri, R., Bittoni, A., Pierantoni, C., Zaniboni, A., Galizia, E., Giustini, L., Silva, R.R., Bisonni, R., Berardi, R., Biscotti, T., Biagetti, S., Bearzi, I. and Cascinu, S. (2011) The Role of HER-3 Expression in the Prediction of Clinical Outcome for Advanced Colorectal Cancer Patients Receiving Irinotecan and Cetuximab. The Oncologist. 16: 53–60.

Schaefer, G., Haber, L., Crocker, L.M., Shia, S., Shao, L., Dowbenko, D., Totpal, K., Wong, A., Lee, C. V., Stawicki, S., Clark, R., Fields, C., Lewis Phillips, G.D., Prell, R.A., Danilenko, D.M., Franke, Y., Stephan, J.P., Hwang, J., Wu, Y., et al. (2011) A Two-in-One Antibody against HER3 and EGFR Has Superior Inhibitory Activity Compared with Monospecific Antibodies. Cancer Cell. 20: 472–486.

Schepeler, T., Reinert, J.T., Ostenfeld, M.S., Christensen, L.L., Silahtaroglu, A.N., Dyrskjøt, L., Wiuf, C., Sørensen, F.J., Kruhøffer, M., Laurberg, S., Kauppinen, S., Ørntoft, T.F. and Andersen, C.L. (2008) Diagnostic and prognostic microRNAs in stage II colon cancer. Cancer research. 68: 6416–6424.

Schetter, A.J., Leung, S.Y., Sohn, J.J., Zanetti, K.A., Bowman, E.D., Yanaihara, N., Yuen, S.T., Chan, T.L., Kwong, D.L.W., Au, G.K.H., Liu, C.-G., Calin, G.A., Croce, C.M. and Harris, C.C. (2008) MicroRNA expression profiles associated with prognosis and therapeutic outcome in colon adenocarcinoma. Jama. 299: 425–436.

Schwartzberg, L.S., Rivera, F., Karthaus, M., Fasola, G., Canon, J.-L., Hecht, J.R., Yu, H., Oliner, K.S. and Go, W.Y. (2014) PEAK: A Randomized, Multicenter Phase II Study of Panitumumab Plus Modified Fluorouracil, Leucovorin, and Oxaliplatin (mFOLFOX6) or Bevacizumab Plus mFOLFOX6 in Patients With Previously Untreated, Unresectable, Wild-Type KRAS Exon 2 Metastatic Colorectal Cancer.

222

Journal of Clinical Oncology. 32: 2240–2247.

Schwarzenbach, H., da Silva, A.M., Calin, G. and Pantel, K. (2015) Data Normalization Strategies for MicroRNA Quantification. Clinical Chemistry. 61: 1333– 1342.

Seike, M., Goto, A., Okano, T., Bowman, E.D., Schetter, A.J., Horikawa, I., Mathe, E.A., Jen, J., Yang, P., Sugimura, H., Gemma, A., Kudoh, S., Croce, C.M. and Harris, C.C. (2009) MiR-21 is an EGFR-regulated anti-apoptotic factor in lung cancer in never-smokers. Proceedings of the National Academy of Sciences. 106: 12085– 12090.

Seligmann, J.F., Elliott, F., Jacobs, B., Hatch, A.J., Sibley, A., Richman, S., Jiang, C., Barrett, J., Owzar, K., Quirke, P., Hurwitz, H., Seymour, M.T. and Nixon, A.B. (2015) HER3 as a biomarker of prognosis and panitumumab (Pan) benefit in RAS -wt advanced colorectal cancer (aCRC). Journal of Clinical Oncology. 33: 3583–3583.

Sergina, N. V, Rausch, M., Wang, D., Blair, J., Hann, B., Shokat, K.M. and Moasser, M.M. (2007) Escape from HER-family tyrosine kinase inhibitor therapy by the kinase- inactive HER3. Nature. 445: 437–441.

Seymour, M.T., Brown, S.R., Middleton, G., Maughan, T., Richman, S., Gwyther, S., Lowe, C., Seligmann, J.F., Wadsley, J., Maisey, N., Chau, I., Hill, M., Dawson, L., Falk, S., O’Callaghan, A., Benstead, K., Chambers, P., Oliver, A., Marshall, H., et al. (2013) Panitumumab and irinotecan versus irinotecan alone for patients with KRAS wild-type, fluorouracil-resistant advanced colorectal cancer (PICCOLO): a prospectively stratified randomised trial. The Lancet Oncology. 14: 749–759.

Shen, J., Todd, N.W., Zhang, H., Yu, L., Lingxiao, X., Mei, Y., Guarnera, M., Liao, J., Chou, A., Lu, C.L., Jiang, Z., Fang, H., Katz, R.L. and Jiang, F. (2011) Plasma microRNAs as potential biomarkers for non-small-cell lung cancer. Laboratory Investigation. 91: 579–587.

Shibuya, H., Iinuma, H., Shimada, R., Horiuchi, A. and Watanabe, T. (2010) Clinicopathological and Prognostic Value of MicroRNA-21 and MicroRNA-155 in Colorectal Cancer. Oncology. 79: 313–320.

Shimamoto, Y., Kitamura, H., Niimi, K., Yoshikawa, Y., Hoshi, F., Ishizuka, M. and Takahashi, E. (2013) Selection of suitable reference genes for mRNA quantification studies using common marmoset tissues. Molecular Biology Reports. 40: 6747–

223

6755.

Shiromizu, T., Kume, H., Ishida, M., Adachi, J., Kano, M., Matsubara, H. and Tomonaga, T. (2017) Quantitation of putative colorectal cancer biomarker candidates in serum extracellular vesicles by targeted proteomics. Scientific Reports. 7: 12782.

Sigismund, S., Avanzato, D. and Lanzetti, L. (2018) Emerging functions of the EGFR in cancer. Molecular oncology. 12: 3–20.

Silva, J., Garcia, V., Rodriguez, M., Compte, M., Cisneros, E., Veguillas, P., Garcia, J.M., Dominguez, G., Campos-Martin, Y., Cuevas, J., Peña, C., Herrera, M., Diaz, R., Mohammed, N. and Bonilla, F. (2012) Analysis of exosome release and its prognostic value in human colorectal cancer. Genes, Chromosomes and Cancer. 51: 409–418.

Silvestri, A., Calvert, V., Belluco, C., Lipsky, M., De Maria, R., Deng, J., Colombatti, A., De Marchi, F., Nitti, D., Mammano, E., Liotta, L., Petricoin, E. and Pierobon, M. (2013) Protein pathway activation mapping of colorectal metastatic progression reveals metastasis-specific network alterations. Clinical & experimental metastasis. 30: 309–316.

Singh, A.B. and Harris, R.C. (2005) Autocrine, paracrine and juxtacrine signaling by EGFR ligands. Cellular signalling. 17: 1183–1193.

Singh, B., Carpenter, G. and Coffey, R.J. (2016a) EGF receptor ligands: recent advances. F1000Research. 5: 2270.

Singh, B., Carpenter, G. and Coffey, R.J. (2016b) EGF receptor ligands: Recent advances [version 1; referees: 3 approved] F1000Research. 5.

Sinha, P., Okoro, C., Foell, D., Freeze, H.H., Ostrand-Rosenberg, S. and Srikrishna, G. (2008) Proinflammatory S100 proteins regulate the accumulation of myeloid- derived suppressor cells. The Journal of immunology. 181: 4666–4675.

Skog, J., Würdinger, T., Rijn, S. van, Meijer, D.H., Gainche, L., Curry, W.T., Carter, B.S., Krichevsky, A.M. and Breakefield, X.O. (2008) Glioblastoma microvesicles transport RNA and proteins that promote tumour growth and provide diagnostic biomarkers. Nature cell biology. 10: 1470–1476.

Skovsen, E., Snyder, J.W., Lambert, J.D.C. and Ogilby, P.R. (2005) Lifetime and Diffusion of Singlet Oxygen in a Cell. The Journal of Physical Chemistry B. 109:

224

8570–8573.

Smith, G., Carey, F.A., Beattie, J., Wilkie, M.J. V, Lightfoot, T.J., Coxhead, J., Garner, R.C., Steele, R.J.C. and Wolf, C.R. (2002) Mutations in APC, Kirsten-ras, and p53--alternative genetic pathways to colorectal cancer. Proceedings of the National Academy of Sciences of the United States of America. 99: 9433–9438.

Sobrero, A.F., Maurel, J., Fehrenbacher, L., Scheithauer, W., Abubakr, Y.A., Lutz, M.P., Vega-Villegas, M.E., Eng, C., Steinhauer, E.U., Prausova, J., Lenz, H.-J., Borg, C., Middleton, G., Kröning, H., Luppi, G., Kisker, O., Zubel, A., Langer, C., Kopit, J., et al. (2008) EPIC: Phase III Trial of Cetuximab Plus Irinotecan After Fluoropyrimidine and Oxaliplatin Failure in Patients With Metastatic Colorectal Cancer. Journal of Clinical Oncology. 26: 2311–2319.

Söderberg, O., Gullberg, M., Jarvius, M., Ridderstråle, K., Leuchowius, K.-J., Jarvius, J., Wester, K., Hydbring, P., Bahram, F., Larsson, L.-G. and Landegren, U. (2006) Direct observation of individual endogenous protein complexes in situ by proximity ligation. Nature methods. 3: 995–1000.

Solit, D.B., Garraway, L.A., Pratilas, C.A., Sawai, A., Getz, G., Basso, A., Ye, Q., Lobo, J.M., She, Y., Osman, I., Golub, T.R., Sebolt-Leopold, J., Sellers, W.R. and Rosen, N. (2006) BRAF mutation predicts sensitivity to MEK inhibition. Nature. 439: 358–362.

Song, J., Bai, Z., Han, W., Zhang, J., Meng, H., Bi, J., Ma, X., Han, S. and Zhang, Z. (2012) Identification of Suitable Reference Genes for qPCR Analysis of Serum microRNA in Gastric Cancer Patients. Digestive Diseases and Sciences. 57: 897– 904.

Spears, M., Taylor, K.J., Munro, A.F., Cunningham, C.A., Mallon, E.A., Twelves, C.J., Cameron, D.A., Thomas, J. and Bartlett, J.M.S. (2012) In situ detection of HER2:HER2 and HER2:HER3 protein–protein interactions demonstrates prognostic significance in early breast cancer. Breast Cancer Research and Treatment. 132: 463–470.

Stoorvogel, W., Kleijmeer, M., Geuze, H. and Raposo, G. (2002) The biogenesis and functions of exosomes. Traffic. 3: 321–330.

Stryer, L. and Haugland, R.P. (1967) Energy transfer: a spectroscopic ruler. Proceedings of the National Academy of Sciences of the United States of America.

225

58: 719–726.

Sun, L., Wu, H. and Guan, Y.S. (2008) Colonography by CT, MRI and PET/CT combined with conventional colonoscopy in colorectal cancer screening and staging World Journal of Gastroenterology. 14 (6) pp. 853–863.

SUZUKI, N., NAKAGAWA, F., NUKATSUKA, M. and FUKUSHIMA, M. (2011) Trifluorothymidine exhibits potent antitumor activity via the induction of DNA double- strand breaks. Experimental and Therapeutic Medicine. 2: 393–397.

Sweeney, C., Fambrough, D., Huard, C., Diamonti, A.J., Lander, E.S., Cantley, L.C. and Carraway, K.L. (2001) Growth Factor-specific Signaling Pathway Stimulation and Gene Expression Mediated by ErbB Receptors. Journal of Biological Chemistry. 276: 22685–22698.

Tabernero, J., Yoshino, T., Cohn, A.L., Obermannova, R., Bodoky, G., Garcia- Carbonero, R., Ciuleanu, T.-E., Portnoy, D.C., Van Cutsem, E., Grothey, A., Prausová, J., Garcia-Alfonso, P., Yamazaki, K., Clingan, P.R., Lonardi, S., Kim, T.W., Simms, L., Chang, S.-C., Nasroulah, F., et al. (2015) Ramucirumab versus placebo in combination with second-line FOLFIRI in patients with metastatic colorectal carcinoma that progressed during or after first-line therapy with bevacizumab, oxaliplatin, and a fluoropyrimidine (RAISE): a randomised, double-blind, multicentre, phase 3 study. The Lancet. Oncology. 16: 499–508.

Tadrous, P.J., Siegel, J., French, P.M., Shousha, S., Lalani, E.-N. and Stamp, G.W. (2003) Fluorescence lifetime imaging of unstained tissues: early results in human breast cancer. The Journal of Pathology. 199: 309–317.

Tajima, A., Hess, M.T., Cabrera, B.L., Kolodner, R.D. and Carethers, J.M. (2004) The mismatch repair complex hMutSα recognizes 5-fluorouracil-modified DNA: Implications for chemosensitivity and resistance. Gastroenterology. 127: 1678–1684.

Tanic, M., Yanowsky, K., Rodriguez-Antona, C., Andrés, R., Márquez-Rodas, I., Osorio, A., Benitez, J. and Martinez-Delgado, B. (2012) Deregulated miRNAs in Hereditary Breast Cancer Revealed a Role for miR-30c in Regulating KRAS Oncogene Justin L. Mott (ed.). PLoS ONE. 7: e38847.

Tao, J.J., Castel, P., Radosevic-Robin, N., Elkabets, M., Auricchio, N., Aceto, N., Weitsman, G., Barber, P., Vojnovic, B., Ellis, H., Morse, N., Viola-Villegas, N.T., Bosch, A., Juric, D., Hazra, S., Singh, S., Kim, P., Bergamaschi, A., Maheswaran, S.,

226 et al. (2014) Antagonism of EGFR and HER3 Enhances the Response to Inhibitors of the PI3K-Akt Pathway in Triple-Negative Breast Cancer. Science Signaling. 7: ra29– ra29.

Tarcic, G., Boguslavsky, S.K., Wakim, J., Kiuchi, T., Liu, A., Reinitz, F., Nathanson, D., Takahashi, T., Mischel, P.S., Ng, T. and Yarden, Y. (2009) An unbiased screen identifies DEP-1 tumor suppressor as a phosphatase controlling EGFR endocytosis. Current Biology. 19: 1788–1798.

Tauro, B.J., Greening, D.W., Mathias, R.A., Ji, H., Mathivanan, S., Scott, A.M. and Simpson, R.J. (2012) Comparison of ultracentrifugation, density gradient separation, and immunoaffinity capture methods for isolating human colon cancer cell line LIM1863-derived exosomes. Methods. 56: 293–304.

Taylor, D. and Gercel-Taylor, C. (2011) Exosomes/microvesicles: mediators of cancer-associated immunosuppressive microenvironments. Seminars in immunopathology. 33: 441–454.

Taylor, D. and Gercel-Taylor, C. (2008) MicroRNA signatures of tumor-derived exosomes as diagnostic biomarkers of ovarian cancer. Gynecologic oncology. 110: 13–21.

Thakur, B.K., Zhang, H., Becker, A., Matei, I., Huang, Y., Costa-Silva, B., Zheng, Y., Hoshino, A., Brazier, H., Xiang, J., Williams, C., Rodriguez-Barrueco, R., Silva, J.M., Zhang, W., Hearn, S., Elemento, O., Paknejad, N., Manova-Todorova, K., Welte, K., et al. (2014) Double-stranded DNA in exosomes: A novel biomarker in cancer detection Cell Research. 24 (6) pp. 766–769.

Théry, C., Amigorena, S., Raposo, G. and Clayton, A. (2006) Isolation and Characterization of Exosomes from Cell Culture Supernatants and Biological Fluids. Current Protocols in Cell Biology. 30: 3.22.1-3.22.29.

Théry, C., Ostrowski, M. and Segura, E. (2009) Membrane vesicles as conveyors of immune responses. Nature reviews. Immunology. 9: 581–593.

Théry, C., Regnault, A. and Garin, J. (1999) Molecular characterization of dendritic cell-derived exosomes Selective accumulation of the heat shock protein hsc73. The Journal of cell ….

Théry, C., Zitvogel, L. and Amigorena, S. (2002) Exosomes: Composition, biogenesis and function Nature Reviews Immunology. 2 (8) pp. 569–579.

227

Thirion, P., Michiels, S., Pignon, J.P., Buyse, M., Braud, A.C., Carlson, R.W., O’Connell, M., Sargent, P., Piedbois, P. and Meta-Analysis Group in Cancer (2004) Modulation of Fluorouracil by Leucovorin in Patients With Advanced Colorectal Cancer: An Updated Meta-Analysis. Journal of Clinical Oncology. 22: 3766–3775.

Tibes, R., Qiu, Y., Lu, Y., Hennessy, B., Andreeff, M., Mills, G.B. and Kornblau, S.M. (2006) Reverse phase protein array: validation of a novel proteomic technology and utility for analysis of primary leukemia specimens and hematopoietic stem cells. Molecular cancer therapeutics. 5: 2512–2521.

Toiyama, Y., Takahashi, M., Hur, K., Nagasaka, T., Tanaka, K., Inoue, Y., Kusunoki, M., Boland, C.R. and Goel, A. (2013) Serum miR-21 as a diagnostic and prognostic biomarker in colorectal cancer. Journal of the National Cancer Institute. 105: 849– 859.

Tol, J., Koopman, M., Cats, A., Rodenburg, C.J., Creemers, G.J.M., Schrama, J.G., Erdkamp, F.L.G., Vos, A.H., van Groeningen, C.J., Sinnige, H.A.M., Richel, D.J., Voest, E.E., Dijkstra, J.R., Vink-Börger, M.E., Antonini, N.F., Mol, L., van Krieken, J.H.J.M., Dalesio, O. and Punt, C.J.A. (2009a) Chemotherapy, Bevacizumab, and Cetuximab in Metastatic Colorectal Cancer. New England Journal of Medicine. 360: 563–572.

Tol, J., Nagtegaal, I.D. and Punt, C.J.A. (2009b) BRAF Mutation in Metastatic Colorectal Cancer. New England Journal of Medicine. 361: 98–99.

Tomicic, M.T. and Kaina, B. (2013) Topoisomerase degradation, DSB repair, p53 and IAPs in cancer cell resistance to camptothecin-like topoisomerase I inhibitors. Biochimica et Biophysica Acta (BBA) - Reviews on Cancer. 1835: 11–27.

Tournigand, C., André, T., Achille, E., Lledo, G., Flesh, M., Mery-Mignard, D., Quinaux, E., Couteau, C., Buyse, M., Ganem, G., Landi, B., Colin, P., Louvet, C. and de Gramont, A. (2004) FOLFIRI Followed by FOLFOX6 or the Reverse Sequence in Advanced Colorectal Cancer: A Randomized GERCOR Study. Journal of Clinical Oncology. 22: 229–237.

Toyota, M., Ahuja, N., Ohe-Toyota, M., Herman, J.G., Baylin, S.B. and Issa, J.P. (1999) CpG island methylator phenotype in colorectal cancer. Proceedings of the National Academy of Sciences of the United States of America. 96: 8681–8686.

Trajkovic, K., Hsu, C., Chiantia, S., Rajendran, L., Wenzel, D., Wieland, F., Schwille,

228

P., Brugger, B. and Simons, M. (2008) Ceramide triggers budding of exosome vesicles into multivesicular endosomes. Science. 319: 1244–1247.

Troiani, T., Martinelli, E., Napolitano, S., Vitagliano, D., Ciuffreda, L.P., Costantino, S., Morgillo, F., Capasso, A., Sforza, V., Nappi, A., De Palma, R., D’Aiuto, E., Berrino, L., Bianco, R. and Ciardiello, F. (2013) Increased TGF- as a Mechanism of Acquired Resistance to the Anti-EGFR Inhibitor Cetuximab through EGFR-MET Interaction and Activation of MET Signaling in Colon Cancer Cells. Clinical Cancer Research. 19: 6751–6765.

Turovskaya, O., Foell, D., Sinha, P., Vogl, T., Newlin, R., Nayak, J., Nguyen, M., Olsson, A., Nawroth, P.P., Bierhaus, A., Varki, N., Kronenberg, M., Freeze, H.H. and Srikrishna, G. (2008) RAGE, carboxylated glycans and S100A8/A9 play essential roles in colitis-associated carcinogenesis. Carcinogenesis. 29: 2035–2043.

Tveit, K.M., Guren, T., Glimelius, B., Pfeiffer, P., Sorbye, H., Pyrhonen, S., Sigurdsson, F., Kure, E., Ikdahl, T., Skovlund, E., Fokstuen, T., Hansen, F., Hofsli, E., Birkemeyer, E., Johnsson, A., Starkhammar, H., Yilmaz, M.K., Keldsen, N., Erdal, A.B., et al. (2012) Phase III Trial of Cetuximab With Continuous or Intermittent Fluorouracil, Leucovorin, and Oxaliplatin (Nordic FLOX) Versus FLOX Alone in First- Line Treatment of Metastatic Colorectal Cancer: The NORDIC-VII Study. Journal of Clinical Oncology. 30: 1755–1762.

Valadi, H., Ekström, K., Bossios, A., Sjöstrand, M., Lee, J.J. and Lötvall, J.O. (2007) Exosome-mediated transfer of mRNAs and microRNAs is a novel mechanism of genetic exchange between cells. Nature Cell Biology. 9: 654–659.

Vandesompele, J., De Preter, K., Pattyn, F., Poppe, B., Van Roy, N., De Paepe, A. and Speleman, F. (2002) Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes. Genome Biology. 3: research0034.1.

Vestad, B., Llorente, A., Neurauter, A., Phuyal, S., Kierulf, B., Kierulf, P., Skotland, T., Sandvig, K., Haug, K.B.F. and Øvstebø, R. (2017) Size and concentration analyses of extracellular vesicles by nanoparticle tracking analysis: a variation study. Journal of Extracellular Vesicles. 6: 1344087.

Vidal, M., Mangeat, P. and Hoekstra, D. (1997) Aggregation reroutes molecules from a recycling to a vesicle-mediated secretion pathway during reticulocyte maturation. Journal of Cell Science. 110: 1867–1877.

229

Viens, A., Harper, F., Pichard, E., Comisso, M., Pierron, G. and Ogryzko, V. (2008) Use of protein biotinylation in vivo for immunoelectron microscopic localization of a specific protein isoform. The journal of histochemistry and cytochemistry : official journal of the Histochemistry Society. 56: 911–919.

Vincenzi, B., Schiavon, G., Silletta, M., Santini, D. and Tonini, G. (2008) The biological properties of cetuximab. Critical reviews in oncology/hematology. 68: 93– 106.

Vlacich, G. and Coffey, R.J. (2011) Resistance to EGFR-Targeted Therapy: A Family Affair. Cancer Cell. 20: 423–425.

Vogelstein, B., Papadopoulos, N., Velculescu, V.E., Zhou, S., Diaz, L.A. and Kinzler, K.W. (2013) Cancer genome landscapes Science. 340 (6127) pp. 1546–1558.

Wachter, A., Bernhardt, S., Beissbarth, T. and Korf, U. (2015) Analysis of Reverse Phase Protein Array Data: From Experimental Design towards Targeted Biomarker Discovery. Microarrays. 4: 520–539.

Walker, J. and Quirke, P. (2002) Prognosis and response to therapy in colorectal cancer. European Journal of Cancer. 38: 880–886.

Wang, D., Qian, G., Zhang, H., Magliocca, K.R., Nannapaneni, S., Amin, A.R.M.R., Rossi, M., Patel, M., El-Deiry, M., Wadsworth, J.T., Chen, Z., Khuri, F.R., Shin, D.M., Saba, N.F. and Chen, Z.G. (2017) HER3 Targeting Sensitizes HNSCC to Cetuximab by Reducing HER3 Activity and HER2/HER3 Dimerization: Evidence from Cell Line and Patient-Derived Xenograft Models. Clinical cancer research : an official journal of the American Association for Cancer Research. 23: 677–686.

Wang, J., De Veirman, K., Faict, S., Frassanito, M.A., Ribatti, D., Vacca, A. and Menu, E. (2016) Multiple myeloma exosomes establish a favourable bone marrow microenvironment with enhanced angiogenesis and immunosuppression. Journal of Pathology. 239: 162–173.

Wang, L., Cunningham, J.M., Winters, J.L., Guenther, J.C., French, A.J., Boardman, L.A., Burgart, L.J., McDonnell, S.K., Schaid, D.J. and Thibodeau, S.N. (2003) BRAF mutations in colon cancer are not likely attributable to defective DNA mismatch repair. Cancer research. 63: 5209–5212.

Wang, P., Zou, F., Zhang, X., Li, H., Dulak, A., Tomko, R.J., Lazo, J., Wang, Z., Zhang, L. and Yu, J. (2009) microRNA-21 negatively regulates Cdc25A and cell cycle

230 progression in colon cancer cells. Cancer research. 69: 8157–8165.

Wang, S., Huang, X., Lee, C.-K. and Liu, B. (2010) Elevated expression of erbB3 confers paclitaxel resistance in erbB2-overexpressing breast cancer cells via upregulation of Survivin. Oncogene. 29: 4225–4236.

Waterhouse, B.R., Gijsen, M., Barber, P.R., Tullis, I., Vojnovic, B. and Kong, A. (2011) Assessment of EGFR/HER2 dimerization by FRET-FLIM utilizing Alexa- conjugated secondary antibodies in relation to targeted therapies in cancers. Oncotarget. 2: 728–736.

Way, T.-D. and Lin, J.-K. (2005) Role of HER2/HER3 co-receptor in breast carcinogenesis. Future Oncology. 1: 841–849.

Wee, P. and Wang, Z. (2017) Epidermal Growth Factor Receptor Cell Proliferation Signaling Pathways. Cancers. 9:.

Weisenberger, D.J., Siegmund, K.D., Campan, M., Young, J., Long, T.I., Faasse, M.A., Kang, G.H., Widschwendter, M., Weener, D., Buchanan, D., Koh, H., Simms, L., Barker, M., Leggett, B., Levine, J., Kim, M., French, A.J., Thibodeau, S.N., Jass, J., et al. (2006) CpG island methylator phenotype underlies sporadic microsatellite instability and is tightly associated with BRAF mutation in colorectal cancer. Nature genetics. 38: 787–793.

Weitsman, G., Barber, P.R., Nguyen, L.K., Lawler, K., Patel, G., Woodman, N., Kelleher, M.T., Pinder, S.E., Rowley, M., Ellis, P.A., Purushotham, A.D., Coolen, A.C., Kholodenko, B.N., Vojnovic, B., Gillett, C. and Ng, T. (2016) HER2-HER3 dimer quantification by FLIM-FRET predicts breast cancer metastatic relapse independently of HER2 IHC status. Oncotarget. 7: 51012–51026.

Wheeler, D.L., Dunn, E.F. and Harari, P.M. (2010) Understanding resistance to EGFR inhibitors-impact on future treatment strategies. Nature reviews. Clinical oncology. 7: 493–507.

Wheeler, D.L., Huang, S., Kruser, T.J., Nechrebecki, M.M., Armstrong, E.A., Benavente, S., Gondi, V., Hsu, K.-T. and Harari, P.M. (2008) Mechanisms of acquired resistance to cetuximab: role of HER (ErbB) family members. Oncogene. 27: 3944–3956.

Wilhelm, S.M., Dumas, J., Adnane, L., Lynch, M., Carter, C.A., Schütz, G., Thierauch, K.-H. and Zopf, D. (2011) Regorafenib (BAY 73-4506): A new oral

231 multikinase inhibitor of angiogenic, stromal and oncogenic receptor tyrosine kinases with potent preclinical antitumor activity. International Journal of Cancer. 129: 245– 255.

Wilson, B., Liotta, L.A. and Petricoin, E. (2010) Monitoring proteins and protein networks using reverse phase protein arrays. Disease markers. 28: 225–232.

Wilson, K.J., Mill, C., Lambert, S., Buchman, J., Wilson, T.R., Hernandez-Gordillo, V., Gallo, R.M., Ades, L.M.C., Settleman, J. and Riese, D.J. (2012) EGFR ligands exhibit functional differences in models of paracrine and autocrine signaling. Growth Factors. 30: 107–116.

Witwer, K.W., Buzás, E.I., Bemis, L.T., Bora, A., Lässer, C., Lötvall, J., Nolte-‘t Hoen, E.N., Piper, M.G., Sivaraman, S., Skog, J., Théry, C., Wauben, M.H. and Hochberg, F. (2013) Standardization of sample collection, isolation and analysis methods in extracellular vesicle research. Journal of Extracellular Vesicles. 2: 20360.

Wong, A.J., Ruppert, J.M., Bigner, S.H., Grzeschik, C.H., Humphrey, P.A., Bigner, D.S. and Vogelstein, B. (1992) Structural alterations of the epidermal growth factor receptor gene in human gliomas. Proceedings of the National Academy of Sciences of the United States of America. 89: 2965–2969.

Wong, S.-F. (2005) Cetuximab: an epidermal growth factor receptor monoclonal antibody for the treatment of colorectal cancer. Clinical therapeutics. 27: 684–694.

Wouters, F.S., Verveer, P.J. and Bastiaens, P.I. (2001) Imaging biochemistry inside cells. Trends in cell biology. 11: 203–211.

Wu, K., Li, L. and Li, S. (2015a) Circulating microRNA-21 as a biomarker for the detection of various carcinomas: an updated meta-analysis based on 36 studies. Tumor Biology. 36: 1973–1981.

Wu, X., Fan, Z., Masui, H., Rosen, N. and Mendelsohn, J. (1995) Apoptosis induced by an anti-epidermal growth factor receptor monoclonal antibody in a human colorectal carcinoma cell line and its delay by insulin. The Journal of clinical investigation. 95: 1897–1905.

Wu, Y., Deng, W. and Klinke, D.J. (2015b) Exosomes: Improved methods to characterize their morphology, RNA content, and surface protein biomarkers. Analyst. 140: 6631–6642.

232

Wubbolts, R., Rachel Leckie, Veenhuizen, P., Schwarzmann, G., Möbius, W., Joerg, H., Slot, J.-W., Geuze, H.J. and Stoorvogel, W. (2003) Proteomic and biochemical analyses of human B cell-derived exosomes Potential implications for their function and multivesicular body formation. Journal of Biological Chemistry. 278: 10963– 10972.

XIONG, B., CHENG, Y., MA, L. and ZHANG, C. (2013) MiR-21 regulates biological behavior through the PTEN/PI-3 K/Akt signaling pathway in human colorectal cancer cells. International Journal of Oncology. 42: 219–228.

Yaeger, R., Cercek, A., O’Reilly, E.M., Reidy, D.L., Kemeny, N., Wolinsky, T., Capanu, M., Gollub, M.J., Rosen, N., Berger, M.F., Lacouture, M.E., Vakiani, E. and Saltz, L.B. (2015) Pilot Trial of Combined BRAF and EGFR Inhibition in BRAF-Mutant Metastatic Colorectal Cancer Patients. Clinical Cancer Research. 21: 1313–1320.

Yang, C. and Robbins, P. (2011) The roles of tumor-derived exosomes in cancer pathogenesis. Clinical and Developmental immunology. 842849:.

Yarden, Y. and Schlessinger, J. (1987) Self-phosphorylation of epidermal growth factor receptor: evidence for a model of intermolecular allosteric activation. Biochemistry. 26: 1434–1442.

Yarden, Y. and Sliwkowski, M.X. (2001) Untangling the ErbB signalling network. Nature reviews. Molecular cell biology. 2: 127–137.

Yin, C., Li, H., Zhang, B., Liu, Y., Lu, G. and Lu, S. (2013) RAGE-binding S100A8/A9 promotes the migration and invasion of human breast cancer cells through actin polymerization and epithelial–mesenchymal transition. Breast cancer research and Treatment. 142: 297–309.

Yin, J., Yan, X., Yao, X., Zhang, Y., Shan, Y., Mao, N., Yang, Y. and Pan, L. (2012) Secretion of annexin A3 from ovarian cancer cells and its association with platinum resistance in ovarian cancer patients. Journal of Cellular and Molecular Medicine. 16: 337–348.

Yin, Y., Zhang, B., Wang, W., Fei, B., Quan, C., Zhang, J., Song, M., Bian, Z., Wang, Q., Ni, S., Hu, Y., Mao, Y., Zhou, L., Wang, Y., Yu, J., Du, X., Hua, D. and Huang, Z. (2014) MiR-204-5p inhibits proliferation and invasion and enhances chemotherapeutic sensitivity of colorectal cancer cells by downregulating RAB22A. Clinical Cancer Research. 20: 6187–6699.

233

Yonesaka, K., Zejnullahu, K., Okamoto, I., Satoh, T., Cappuzzo, F., Souglakos, J., Ercan, D., Rogers, A., Roncalli, M., Takeda, M., Fujisaka, Y., Philips, J., Shimizu, T., Maenishi, O., Cho, Y., Sun, J., Destro, A., Taira, K., Takeda, K., et al. (2011) Activation of ERBB2 signaling causes resistance to the EGFR-directed therapeutic antibody cetuximab. Science translational medicine. 3: 99ra86.

Zaczek, A., Brandt, B. and Bielawski, K.P. (2005) The diverse signaling network of EGFR, HER2, HER3 and HER4 tyrosine kinase receptors and the consequences for therapeutic approaches. Histology and histopathology. 20: 1005–1015.

Zarovni, N., Corrado, A., Guazzi, P., Zocco, D., Lari, E., Radano, G., Muhhina, J., Fondelli, C., Gavrilova, J. and Chiesi, A. (2015) Integrated isolation and quantitative analysis of exosome shuttled proteins and nucleic acids using immunocapture approaches. Methods. 87: 46–58.

Zeringer, E., Barta, T., Li, M. and Vlassov, A. V. (2015) Strategies for Isolation of Exosomes. Cold Spring Harbor Protocols. 2015: pdb.top074476.

Zhang, Q., Park, E., Kani, K. and Landgraf, R. (2012) Functional isolation of activated and unilaterally phosphorylated heterodimers of ERBB2 and ERBB3 as scaffolds in ligand-dependent signaling. Proceedings of the National Academy of Sciences. 109: 13237–13242.

Zhang, Q., Yu, L., Qin, D., Huang, R., Jiang, X., Zou, C., Tang, Q., Chen, Y., Wang, G., Wang, X. and Gao, X. (2015) Role of microRNA-30c Targeting ADAM19 in Colorectal Cancer Klaus Roemer (ed.). PLOS ONE. 10: e0120698.

Zhang, X., Gureasko, J., Shen, K., Cole, P.A. and Kuriyan, J. (2006) An allosteric mechanism for activation of the kinase domain of epidermal growth factor receptor. Cell. 125: 1137–1149.

Zhang, Y.-J., Tian, X.-Q., Sun, D.-F., Zhao, S.-L., Xiong, H. and Fang, J.-Y. (2009) Combined Inhibition of MEK and mTOR Signaling Inhibits Initiation and Progression of Colorectal Cancer. Cancer Investigation. 27: 273–285.

Zhao, D.W., Li, M.M., Han, J.P., Wang, Y., Jiang, L.X. and Chang, H.L. (2019) MiR- 30c exerts tumor suppressive functions in colorectal carcinoma by directly targeting BCL9. European Review for Medical and Pharmacological Sciences. 23: 3335–3343.

Zöller, M. (2009) Tetraspanins: push and pull in suppressing and promoting metastasis. Nature Reviews Cancer. 9: 40–55.

234

Zomer, A., Maynard, C., Verweij, F.J., Kamermans, A., Schäfer, R., Beerling, E., Schiffelers, R.M., de Wit, E., Berenguer, J., Ellenbroek, S.I.J., Wurdinger, T., Pegtel, D.M. and van Rheenen, J. (2015) In Vivo Imaging Reveals Extracellular Vesicle- Mediated Phenocopying of Metastatic Behavior. Cell. 161: 1046–1057.

235