Tumour-Stroma Signalling in Cancer Cell Motility and Metastasis

Valbona Luga

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy,

Department of Molecular Genetics, University of Toronto

Valbona Luga

Doctor of Philosophy

Department of Molecular Genetics University of Toronto

2013

Abstract

The tumour-associated stroma, consisting of fibroblasts, inflammatory cells, vasculature and extracellular matrix proteins, plays a critical role in tumour growth, but how it regulates cancer cell migration and metastasis is poorly understood. The Wnt-planar cell polarity (PCP) pathway regulates convergent extension movements in vertebrate development. However, it is unclear whether this pathway also functions in cancer cell migration. In addition, the factors that mobilize long-range signalling of Wnt morphogens, which are tightly associated with the plasma membrane, have yet to be completely characterized. Here, I show that fibroblasts secrete membrane microvesicles of endocytic origin, termed exosomes, which promote tumour cell protrusive activity, motility and metastasis via the exosome component Cd81. In addition, I demonstrate that fibroblast exosomes activate autocrine Wnt-PCP signalling in breast cancer cells as detected by the association of Wnt with Fzd receptors and the asymmetric distribution of

Fzd-Dvl and Vangl-Pk complexes in exosome-stimulated cancer cell protrusive structures.

Moreover, I show that Pk expression in breast cancer cells is essential for fibroblast-stimulated cancer cell metastasis. Lastly, I reveal that trafficking in cancer cells promotes tethering of autocrine Wnt11 to fibroblast exosomes. These studies further our understanding of the role of ii the tumour-associated stroma in cancer metastasis and bring us closer to a more targeted approach for the treatment of cancer spread.

iii

Acknowledgements

My journey as a PhD student has been an amazing learning experience. This would not have been possible without a number of people who were there for me to share both the challenges and rewards of my scientific endeavours.

Jeff, it has been an honour and privilege to be your student. I am deeply grateful to you for believing in me and helping me realize that science is my call in life. Your passion, brilliance and mastery of science will always be an inspiration to me.

Lil, thank you for your support and advice. You are an important role model to me and I am very grateful for your mentorship.

Thanks to my committee members, Dr. Igor Jurisica, Dr. David Kaplan and Dr. Rama Khokha, for their suggestions and support.

Thank you to all the past and current members of the Wrana and Attisano labs. I spent the last nine years of my life mainly in the lab and many of you became my friends and family. John, thank you for your mentorship during my first year in the lab when I barely knew about science. Christine, thank you for teaching me as a young student, aka your “little One”, how to properly design and carry out experiments. You are a remarkable scientist and I am particularly proud to be your friend. Miriam, I will always be thankful to you for your support and encouragment during my search for the “Luga factors”. Your wisdom and advice have been greatly helpful to me throughout my PhD career. I deeply cherish our friendship and will sorely miss our laughs. Abi, thanks for being a good friend and I will miss our chats by the phosphoimager. Liang, thank you for your help; it was a pleasure working with you. Alicia, thank you for your help and friendship. To everyone else, thank you for your support and friendship. I will miss all of you.

To my family, thank you for your unconditional love and helping me become the person I am today. Dad, you taught me to dream big and believe in myself. Rovi, your intelligence, talent and search for perfection have been deeply inspirational to me. Koli, your bigheartedness and sense of justice are your best virtues and I look up to you for that. Calvin, Jason, Brandon, Laetitia and Delora, I thank you for bringing pure joy to my life. Mom, your love and devotion

iv to your children is unparalleled. You taught me early in my life to work hard and be thorough; these lessons served me well during my PhD. I dedicate this thesis to you.

Finally, Rohit, thank you for being by my side for the last eleven years. Your wisdom, patience and love have been crucial for my professional and personal growth. Words cannot express my gratitude and love for you and I look forward to our life journey together.

Table of Contents

Abstract ...... ii

Acknowledgements ...... iv

Table of Contents ...... vi

List of Figures and Tables ...... ix

List of Abbreviations and Symbols ...... x

Chapter 1 : General Introduction ...... 1 1.1. Cell Motility and Cancer ...... 2 1.1.1. Overview of cancer hallmarks ...... 2 1.1.2. Autonomous cancer hallmarks ...... 2 1.1.3. Neovascularization and suppression of the immune system ...... 5 1.1.4. Apical-basolateral cell polarity as a barrier to tumour cell invasiveness ...... 8 1.1.5. Tumour cell migration ...... 12 1.1.6. Metastasis – colonization of distant organs ...... 22 1.1.7. The tumour microenvironment ...... 27 1.1.8. Fibroblasts as components of the tumour microenvironment ...... 30 1.1.9. Identification and function of cancer stem-like cells ...... 32 1.2. Noncanonical Wnt-Planar Cell Polarity Pathway ...... 33 1.2.1. Planar cell polarity and its manifestations ...... 33 1.2.2. Molecular PCP components in the fly ...... 36 1.2.3. Models of PCP signalling in the fly ...... 37 1.2.4. PCP signalling in vertebrates ...... 39 1.2.5. Wnt structure and post-translational modifications ...... 44 1.2.6. Wnt signalling pathways ...... 46 1.2.7. Role of PCP in cell motility and cancer ...... 49 1.3. Exosomes as Potent Mediators of Intercellular Communication ...... 50 1.3.1. Modes of intercellular communication ...... 50 1.3.2. Introduction to exosomes ...... 50 1.3.3. Exosome biogenesis ...... 51 1.3.4. Exosome components ...... 54 1.3.5. Exosome functions ...... 55 1.3.6. Exosomes as biomarkers for disease ...... 57 1.3.7. Exosomes as therapeutic tools ...... 57 vi

1.3.8. Tetraspanins – masters of multiple biological functions ...... 58 1.3.9. Cd81 – an exemplary tetraspanin ...... 63 1.4. Overview of Thesis ...... 67

Chapter 2 : Fibroblast-secreted Exosomes Regulate Breast Cancer Cell Motility and Metastasis ...... 69 2.1. Introduction ...... 70 2.2. Materials and Methods ...... 71 2.2.1. Cell Culture ...... 71 2.2.2. Conditioned Media Preparation ...... 72 2.2.3. Single Cell Motility Assay ...... 72 2.2.4. 3D Matrigel Cultures ...... 73 2.2.5. Orthotopic Mouse Model of Breast Cancer Metastasis ...... 73 2.2.6. Immunohistochemistry ...... 73 2.2.7. Protein Chromatography ...... 74 2.2.8. Mass Spectrometry ...... 75 2.2.9. Exosome Isolation By Differential Ultracentrifugation ...... 76 2.2.10. Electron Microscopy ...... 76 2.2.11. Exosome Purification by Sucrose Gradient Ultracentrifugation ...... 77 2.2.12. Methyl-β-Cyclodextrin (MβCD) Treatment ...... 77 2.2.13. Exosome Immuno-Magnetic Extraction ...... 77 2.2.14. Immunoblotting ...... 78 2.2.15. RNA Extraction and Real-Time Quantitative PCR (RT-qPCR) ...... 78 2.2.16. PCR Primer Sets ...... 79 2.2.17. Cell Transfection ...... 79 2.2.18. shRNA Sequences ...... 80 2.2.19. Cell Lysis ...... 80 2.2.20. Analysis of Human Stroma Gene Expression Data ...... 81 2.3. Results ...... 81 2.3.1. L cell-secreted Factors Stimulate Breast Cancer Cell Protrusive Activity and Motility ...... 81 2.3.2. L cell-secreted Factors Stimulate Breast Cancer Cell Invasiveness and Metastatic Potential 86 2.3.3. Identification and Characterization of L cell-secreted Exosomes ...... 90 2.3.4. Functional Characterization of L cell-secreted Exosomes ...... 104 2.3.5. Cd81 Functions As a Component of L cell-secreted Exosomes to Promote BCC Motility .... 107 2.3.6. Cd81 Is Necessary For L cell-stimulation of BCC Metastasis ...... 110 2.3.7. Human CAFs Secrete Cd81-positive Exosomes That Stimulate BCC Motility ...... 113 2.4. Discussion ...... 115

vii

Chapter 3 : Fibroblast Exosomes Mobilize the Wnt-PCP Pathway to Promote Breast Cancer Cell Motility and Metastasis ...... 118 3.1. Introduction ...... 119 3.2. Materials and Methods ...... 120 3.2.1. Cell Culture ...... 120 3.2.2. Conditioned Media Preparation ...... 120 3.2.3. Single Cell Motility Assay ...... 121 3.2.4. Cell Transfection ...... 121 3.2.5. shRNA Sequences ...... 121 3.2.6. Cell Lysis And Immunoblotting ...... 122 3.2.7. RNA Extraction and Real-Time Quantitative PCR (RT-qPCR) ...... 122 3.2.8. PCR Primers ...... 123 3.2.9. Immunofluorescence ...... 124 3.2.10. L cell / MDA-MB-231 Orthotopic Mouse Model of Breast Cancer Metastasis ...... 124 3.2.11. DNA Constructs ...... 125 3.2.12. Exosome Isolation by Differential Ultracentrifugation ...... 125 3.2.13. Exosome Internalization Assay: Immunoprecipitation ...... 125 3.2.14. Exosome Internalization Assay: Immunofluorescence and Dynasore Treatment ...... 126 3.2.15. Electron Microscopy ...... 126 3.2.16. Exosome Immuno-Magnetic Extraction ...... 126 3.3. Results ...... 127 3.3.1. Core PCP Components Mediate Fibroblast-Induced BCC Protrusive Activity and Motility ... 127 3.3.2. Pk1 Expression Is Necessary For L cell-Stimulated BCC Metastasis ...... 131 3.3.3. Core PCP Components Distribute Asymmetrically in ACM-treated BCCs ...... 134 3.3.4. Autocrine Wnts Mediate Fibroblast-Induced BCC Protrusive Activity and Motility ...... 141 3.3.5. Fibroblast Exosomes Mobilize Autocrine Wnt-PCP Signalling in BCCs ...... 147 3.4. Discussion ...... 157

Chapter 4 : Thesis Summary and Future Directions ...... 160 4.1. Overview ...... 161 4.2. Unanswered Questions and Future Directions ...... 163

References ...... 168

viii

List of Figures and Tables

Figure 1.1 The Tumour-associated Stroma Influences Tumour Growth ...... 7 Figure 1.2 Apical-Basal Polarity in Epithelial Cells ...... 11 Figure 1.3 Remodelling of the Actin Cytoskeleton ...... 14 Figure 1.4 Cell Migration As a Multistep Process ...... 18 Figure 1.5 The Metastatic Cascade ...... 26 Figure 1.6 PCP in Drosophila and Vertebrates ...... 42 Figure 1.7 Wnt Signalling Pathways ...... 48 Figure 1.8 Schematic Structure of Cd81 As a Prototypical Tetraspanin ...... 61 Figure 1.9 Cd81 and Tetraspanin-Enriched Microdomains (TEMs) ...... 64 Figure 2.1 L cell-secreted Factors Stimulate BCC Protrusive Activity and Motility ...... 83 Figure 2.2 ACM Promotes Protrusive Activity and Motility in Several BCCs ...... 85 Figure 2.3 L cell-secreted Factors Promote MDA-MB-231 Invasiveness and Metastasis ...... 88 Figure 2.4 Fractionation of ACM by Protein Chromatography ...... 91 Table 2-1 Proteins in Fraction Q3 Identified by Mass Spectrometry ...... 93 Figure 2.5 The Active Fractions of ACM Contain Exosomes ...... 102 Figure 2.6 L cell-secreted Exosomes Activate MDA-MB-231 Cell Motility ...... 105 Figure 2.7 Cd81 Functions As a Component of Exosomes ...... 108 Figure 2.8 Cd81 Is Necessary For L cell-stimulated BCC Metastasis ...... 111 Figure 2.9 Human CAFs Secrete Cd81-positive Exosomes That Stimulate BCC Motility ...... 114 Figure 3.1 Core PCP Components Mediate ACM Activity in BCCs ...... 129 Figure 3.2 Prickle1 Is Required For L cell-stimulated MDA-MB-231 Cell Metastasis ...... 132 Figure 3.3 Localization of Fzd6-Dvl1 Complexes in ACM-treated MDA-MB-231 Cells ...... 136 Figure 3.4 Pk1 and Vangl1 Localize Asymmetrically in ACM-treated MDA-MB-231 Cells ...... 137 Figure 3.5 Localization of Fzd7-Dvl3 Complexes in ACM-stimulated SUM-159PT Cells ...... 139 Figure 3.6 Fzd6-Dvl1 Complexes Localize Asymmetrically in Exosome-stimulated BCCs ...... 140 Figure 3.7 Porcupine Expression Is Necessary in BCCs ...... 143 Figure 3.8 BCC-produced Wnt11 Is Necessary For ACM Activity ...... 145 Figure 3.9 BCCs Are Deficient For Cd81-positive Exosomes ...... 148 Figure 3.10 MDA-MB-231 Cells Internalize L cell-secreted Exosomes ...... 150 Figure 3.11 Wnt11 Associates With L cell Exosomes Upon Trafficking In BCCs ...... 154 Figure 3.12 A Model For Fibroblast Exosomes Stimulating Wnt-PCP Signalling In BCCs ...... 156

List of Abbreviations and Symbols

α alpha β beta γ gamma µ micro

AB apicobasal polarity ACM active conditioned media ADAM a disintegrin and metalloproteinase AP anteroposterior APC adenomatous polyposis coli aPKC atypical protein kinase C Arp2/3 actin related proteins 2 and 3

BCC breast cancer cell BMP bone morphogenetic protein

CAF cancer-associated fibroblast CCL CC chemokine ligand Cd81 cluster of differentiation 81 Cdc42 cell division control protein 42 homolog CE convergent extension Celsr cadherin EGF LAG seven-pass G-type receptor Ck1 casein kinase 1 CRB crumbs CRD cysteine-rich domain CSC cancer stem-like cells CTD carboxy-terminal domain

Dgo diego DLG discs-large homologue DNA deoxyribonucleic acid Dsh dishevelled (fruit fly) DV dorsoventral Dvl dishevelled (vertebrates)

ECM extracellular matrix EGF epidermal growth factor EGFR epidermal growth factor receptor ERM ezrin-radixin-moesin EM electron microscopy EMT epithelial-to-mesenchymal transition x

ESCRT Endosomal Sorting Complex Required for Transport EWI glutamate-tryptophan-isoleucine

F-actin filamentous actin FAK focal adhesion kinase FGF fibroblast growth factor Fmi flamingo Fz frizzled receptor (fruit fly) Fzd frizzled receptor (vertebrate)

GAP GTPase activating protein GEF guanine nucleotide exchange factor GSK3β glycogen synthase kinase 3 beta GTPase guanosine triphosphatase

HGF hepatocyte growth factor

ICAM intercellular-adhesion molecule IGF insulin growth factor Igsf8 immunoglobulin superfamily member 8 IL interleukin

JNK c-Jun N-terminal kinase

LEL large extracellular loop LR left-right LRP lipoprotein receptor-related protein

MET mesenchymal-to-epithelial transition MHC major histocompatibility complex miRNA microRNA ML mediolateral MLC myosin light-chain MLCK myosin light-chain kinase MMP matrix metalloproteinase MS mass spectrometry MT microtubule MVB multivesicular body Myc myelocytomatosis viral oncogene homolog

NK natural killer cell NTD amino-terminal domain

PAK Gβγ/p21-activated kinase 1 PAR partitioning defective PCP planar cell polarity PDGF platelet derived growth factor PDGFRβ platelet derived growth factor receptor beta PI3K phosphoinositide 3-kinase Pk prickle PKC protein kinase C PTEN phosphatase and tensin homolog Ptgfrn prostaglandin F2 receptor negative regulator p53 tumour protein p53

Rab member Ras oncogene family Ras rat sarcoma viral oncogene homolog Rac1 Ras-related C3 botulinum toxin substrate 1 RET receptor tyrosine kinase RNA ribonucleic acid RhoA ras homolog family member A ROCK Rho kinase Ryk Receptor-like tyrosine kinase

SCRIB scribble homologue SDF stromal cell-derived factor Sf Smurf siRNA small interfering RNA SMA smooth muscle actin Smad mothers against decapentaplegic homolog Smurf Smad ubiquination regulatory factor Src sarcoma viral oncogene homolog

TEM tetraspanin-enriched membrane TGFβ transforming growth factor beta TGFβRI transforming growth factor beta receptor I TGFβRII transforming growth factor beta receptor II TIAM-1 T-cell lymphoma invasion and metastasis-1 TNFα tumour necrosis factor α

Vang van Gogh Vangl vang-like VEGF vascular endothelial growth factor

WASP Wiskott-Aldrich syndrome protein WAVE WASP family verprolin-homology protein Wg wingless Wls wntless xii

Chapter 1 : General Introduction

1.1. Cell Motility and Cancer

1.1.1. Overview of cancer hallmarks

Cancers arise from cells with genomic aberrations, which provide the genetic diversity that is necessary to drive malignancy (Hanahan and Weinberg, 2011; Talmadge and Fidler, 2010). All cancers share universal characteristics, or hallmarks, that provide self-sufficiency in growth signals, insensitivity to anti-growth signals, resistance to cell death, limitless replicative potential, reprogramming of energy metabolism, induced angiogenesis, evasion of attack from the immune system and activation of invasion and metastasis (Hanahan and Weinberg, 2011).

Importantly, these hallmarks are interconnected as, for example, major oncogenes and tumour suppressor genes that influence cell proliferation also affect cell survival, cell metabolism, cell motility, angiogenesis and evasion of the immune system (Hanahan and Weinberg, 2011;

Koppenol et al., 2011; Mantovani et al., 2008). Also, a very important conceptual progress in the field of cancer research has been the recognition that cancer biology is tightly regulated by the surrounding stroma (Bissell and Hines, 2011).

1.1.2. Autonomous cancer hallmarks

Tumour growth is facilitated by a series of distinct and complementary events that occur within cancer cells (Hanahan and Weinberg, 2011; Talmadge and Fidler, 2010). Continuous proliferative signalling is thought to be one of the earliest hallmarks acquired by cancer cells

(Hanahan and Weinberg, 2011; Talmadge and Fidler, 2010). Normally, tissues maintain their characteristic architecture and function by tightly regulating the production of growth-promoting signals, consisting typically of growth factors, which instruct growth and division in

neighbouring cells in a temporally and spatially organized fashion (Lemmon and Schlessinger,

2010; Witsch et al., 2010). Cancer cells acquire the ability to sustain proliferative signalling independently of the cues provided by their neighbouring cells. They do so via several mechanisms, including elevated self-production of growth factors (GFs) (such as epidermal GF

(EGF) and insulin-like GF (IGF) (Witsch et al., 2010), co-option of normal cells in the tumour microenvironment to produce growth signals that in turn act on cancer cells (Kalluri and

Zeisberg, 2006), upregulation of expression of growth factor receptors (such as EGF receptor

(EGFR)) (Lemmon and Schlessinger, 2010), or acquisition of oncogenic mutations that lead to constitutive activation of growth factor receptors (such EGFR, platelet-derived GF receptor

(PDGFR), fibroblast GF receptor (FGFR)) (Lemmon and Schlessinger, 2010) or downstream effectors (such as rat sarcoma viral oncogene homolog (Ras) and myelocytomatosis viral oncogene homolog (Myc)) (Dang, 2012; Karnoub and Weinberg, 2008) that in turn signal uncontrolled cancer cell proliferation (Hanahan and Weinberg, 2011). Tissue homeostasis is also maintained by growth inhibitory signals, including transforming growth factor beta (TGFβ)

(Massague, 2012) and phosphatase and tensin homolog (PTEN) (Song et al., 2012), and cellular programs, such as cell senescence (Collado and Serrano, 2010; Nardella et al., 2011) and contact inhibition (Martin-Belmonte and Perez-Moreno, 2012), which tumour cells must disable in order to proliferate excessively. TGFβ signals through the TGFβ type 1 and type 2 serine- threonine kinase receptors (TβRI and TβRII respectively), which activate the receptor-regulated

Smad proteins (Smad2/3) that in a complex with common partner Smad (Smad4) mediate

TGFβ-induced expression of several cell cycle inhibitory factors and inhibition of cell proliferation factors (Attisano and Wrana, 2002; Massague, 2012). Indeed, TβRII and Smad4 are commonly inactivated through mutation and loss of heterozygosity in various cancers

(Attisano and Wrana, 2002; Massague, 2012). Other tumour suppressor genes, such as merlin 3

and liver kinase B1 (LKB1), which regulate contact inhibition of cell proliferation, are also commonly inactivated in many cancers (Li et al., 2012a; Martin-Belmonte and Perez-Moreno,

2012). Next, cancer cells overcome senescence, which is a cellular program that is normally activated once the cells have undergone a limited number of successive cell growth and division cycles (Collado and Serrano, 2010; Nardella et al., 2011) and is controlled by tumour protein p53 (p53) and retinoblastoma-associated (RB) proteins (Polager and Ginsberg, 2009). Cellular senescence prevents genomic instability due to end-to-end fusion of chromosomes with eroded telomeres (Collado and Serrano, 2010; Harley, 2008; Nardella et al., 2011). Initially, cancer cells use the unprotected ends of chromosomes as a means to generate genomic instability, which leads to activation of oncogenes and suppression of tumour-suppressor genes (Campbell,

2012; Harley, 2008). Then, cancer cells acquire the ability to express telomerase, a specialized

DNA polymerase that adds telomere repeat segments to the ends of telomeric DNA, and consequently attain replicative immortality (Campbell, 2012; Hanahan and Weinberg, 2011;

Harley, 2008). Furthermore, cancer cells must resist programmed cell death, which is another line of defence against hyperproliferation and genomic instability. Cancer cells manipulate potent cellular programs, including apoptosis, autophagy and necrosis, which are carried out by a myriad of regulatory (p53, B-cell CLL/Lymphoma (Bcl) proteins, Beclin-1) and effector

(caspases 8) components, in order to attenuate cell death and ensure their survival (Giansanti et al., 2011; Kreuzaler and Watson, 2012). Finally, cancer cells also reprogram their glucose metabolism to carry out aerobic glycolysis, a phenomenon known as the Warburg effect, which is thought to divert intermediates for nucleotide and amino acid biosynthesis that in turn act as building blocks for cell growth and proliferation (Koppenol et al., 2011).

1.1.3. Neovascularization and suppression of the immune system

Tumour cells must also interact with normal cells to direct neovascularization and inflammation, which support tumour growth (Folkman, 2007; Hanahan and Weinberg, 2011; Mantovani et al.,

2008) (Figure 1.1A). Tumours demand a constant supply of nutrients, O2/CO2 gas exchange as well as removal of metabolic waste. To fulfill this requirement, tumour cells stimulate new blood vessel formation via angiogenesis (Folkman, 2007), a process whereby endothelial cells are recruited and assemble new blood vessels branching from existing ones, or vasculogenesis, a process whereby bone-marrow derived cells produce endothelial cells that assemble into new vessels (Lyden et al., 2001). Tumour cells switch on this processes by producing angiogenic factors, such as the vascular endothelial growth factors (VEGFs), fibroblast growth factors

(FGFs) and angiopoietin 2 (ANGPT2), which act on endothelial cells or bone marrow-derived cells to instruct new blood vessel growth (Folkman, 2007) (Figure 1.1B). Cancer cells also inhibit the production of anti-angiogenic factors, such as endostatin and thrombospondin 1

(TSP-1), that impede the induction of angiogenesis (Folkman, 2007). Tumour blood vessels have aberrant characteristics, in that they sprout and convolute excessively and have erratic blood flow, and are conducive to malignancy due to deficiency in pericytes, which normally prevent the endothelial wall from being leaky (Hanahan and Weinberg, 2011). Cancer cells also secrete VEGF-C and VEGF-D to stimulate lymphangiogenesis, the process of new lymphatic vessel formation from lymphatic endothelial cells (Tammela and Alitalo, 2010). The lymphatic system plays a critical role in tumour dissemination and metastasis and the density of lymphatic vessels correlates with the incidence of lymph node metastasis and poor prognosis (Tammela and Alitalo, 2010).

Cancer cells also participate in a complex dialogue with cells of the immune system, which acts to both abolish and stimulate tumourigenesis (Mantovani et al., 2008). Studies have shown that the adaptive immune system carries out surveillance and attacks nascent tumours (Mantovani et al., 2008). However, tumours cells acquire the ability to evade the adaptive immune system by several mechanisms, including the secretion of immunosuppressive factors, such as interleukin-

10 (IL-10) and TGFβ, and recruitment of inflammatory cells, including macrophages with an

M2 phenotype, neutrophils and mast cells, as well as distinct subclasses of cells of the adaptive immune system, such as T helper 2 (TH2) cells and certain B lymphocytes that suppress the activity of cytotoxic lymphocytes (Mantovani et al., 2008) (Figure 1.1A). Activation of oncogenes, such as Ras, Myc and receptor tyrosine kinase (RET), and inactivation of tumour suppressor genes, such as TGFβRII, in tumour cells, leads to induction of a transcriptional program that drives the expression of several chemokines (such as CC-chemokine ligand 2

(CCL2) and CCL20), cytokines (such as IL-1β, IL-6, IL-23 and tumour necrosis factor α

(TNFα)) and prostaglandins that recruit inflammatory cells to the tumour microenvironment

(Mantovani et al., 2008) (Figure 1.1B). Importantly, inflammatory cells secrete chemicals, such as reactive oxygen and nitrogen species, which can damage DNA and accelerate genetic aberrations that induce malignancy (Mantovani et al., 2008). In fact, clinical studies indicate that inflammation arising from microbial infections or autoimmune diseases increase the risk of developing cancer and that therapy with non-steroidal anti-inflammatory agents decrease the incidence of mortality from different tumours (Mantovani et al., 2008). In addition, inflammatory cells produce growth factors, chemokines, cytokines and metalloproteases that amplify the inflammatory response, induce angiogenesis and lymphangiogenesis, remodel the

ECM and promote tumour cell survival, migration, invasion and metastasis (Mantovani et al.,

2008). 6

A Epithelial Cell Cancer Cell (CC)

Basement Cancer-Associated Membrane Fibroblast (CAFs) Inflammatory Cells (ICs)

Extracellular Bone Marrow Derived Matrix (ECM) Stem & Progenitor Cells (BMDCs)

Endothelial Lymphatic Cell (EC) Endothelial Motile & Cell (LEC) Invasive Cancer Pericyte Cell (MCC) B

TGF MMPs HGF ROS FGF2 TGF CSF EGF EGF SDF-1 TNF Tenascin C IL-1 VEGF-C CCL5 VEGF-D CCL21 VEGF FGF ANGPT2

Figure 1.1 The Tumour-associated Stroma Influences Tumour Growth

(A) Tumour progression is supported by several stromal cells, including cancer-associated fibroblasts (CAFs), bone marrow-derived cells (BMDCs) and inflammatory cells (ICs), and vasculature cells (blood and lymphatic). (B) Cancer cells (CCs) communicate with stromal cells by numerous secreted factors. For instance, CCs secrete factors that activate CAFs, which in turn remodel ECM and promote CC invasiveness. Moreover, CCs secrete several factors that induce angiogenesis (by recruiting endothelial cells (ECs) or BMDCs) or lymphangiogenesis (by recruiting lymphatic ECs (LECs)) within the tumour. The new vasculature provides 7

nutrients and removes metabolic waste. Further, CCs recruit ICs, including macrophages, neutrophils and mast cells, which promote tumour growth, CC invasiveness and activate angiogenesis. ICs also secrete factors that inhibit immune cells, such as cytotoxic T cells and dendritic cells (not shown), which destroy tumour cells.

1.1.4. Apical-basolateral cell polarity as a barrier to tumour cell invasiveness

Although tumour growth can be detrimental to patient health, it is metastasis, the process of spreading and growth of tumours in secondary organs, that accounts for approximately 90% of cancer-associated mortality (Talmadge and Fidler, 2010). In order to disseminate in distant tissues, cancer cells must first detach from the primary tumour and invade the surrounding environment. In the case of epithelial cells, acquisition of an invasive phenotype involves loss of apical-basolateral (AB) cell polarity and cell-cell adhesions, alteration in cell shape, and increased cell motility (Martin-Belmonte and Perez-Moreno, 2012; Thiery et al., 2009) (Figure

1.2). Normal epithelium exists as a single layer of cells attached to each other via cell-cell junctions, including tight junctions, adherens junctions and desmosomes, which form within lateral membranes (Martin-Belmonte and Perez-Moreno, 2012; Thiery et al., 2009) (Figure 1.2).

Also, epithelial cells attach to the underlying basement membrane via integrins in their basal membrane and face the luminal space or the external environment via their apical membrane.

Epithelial cells exhibit apical-basolateral polarity, which is characterized by asymmetric distribution of cellular components, such as proteins and lipids, and organelles along the AB axis (Martin-Belmonte and Perez-Moreno, 2012) (Figure 1.2). Crosstalk between cell adhesions

(tight junctions, which are comprised of transmembrane proteins occludin, claudin and junctional adhesion molecule (JAM), and cytoplasmic proteins zonula occludens (ZO) 1-3, and

adherens junctions, which are comprised of two adhesive complexes: cadherin-catenins (p-120 catenin, α-catenin and β-catenin) and nectin-afadin) and three triprotein polarity complexes (the partitioning-defective 3 (PAR3)-PAR6-atypical protein kinase C (aPKC) complex, the transmembrane protein crumbs (CRB)-protein associated with lin-seven (PALS)-PALS1- associated tight-junction protein (PATJ) complex and the scribble homologue (SCRIB)-lethal

(2) giant larvae homologue (LGL)-discs-large homologue (DLG) complex) is critical for the establishment and maintenance of AB polarity (Martin-Belmonte and Perez-Moreno, 2012;

Thiery et al., 2009) (Figure 1.2). The PAR6 polarity complex promotes the establishment of the apical-basal membrane border as it is a critical regulator of tight-junctions (Ozdamar et al.,

2005), the CRB complex is required for the establishment of the apical membrane and the

SCRIB complex defines the basolateral plasma domain (Martin-Belmonte and Perez-Moreno,

2012). In addition, cell adhesion proteins and polarity complexes connect to microtubules and the actin cytoskeleton to regulate cell shape and polarity (Iden and Collard, 2008; Martin-

Belmonte and Perez-Moreno, 2012; Thiery et al., 2009). Moreover, cell-cell adhesion proteins and polarity complexes crosstalk with signalling pathways such as TGFβ, Wnt, Hedgehog and

Hippo pathways and mechanistic target of rapamycin serine/threonine kinase (mTOR)- dependent energy metabolism to control cell polarity, adhesion, growth, proliferation, differentiation and cell death (Martin-Belmonte and Perez-Moreno, 2012; Thiery et al., 2009).

AB polarity is a critical barrier to the formation of carcinoma, which is cancer of epithelial origin that accounts for the most prevalent form of malignancy in humans. Misregulation of expression and/or function of the protein components that comprise cell adhesions and polarity complexes, results in breakage of intercellular adhesions, loss of polarized characteristics of

epithelial cells, hyperproliferation and acquisition of migratory and invasive potential (Martin-

Belmonte and Perez-Moreno, 2012; Thiery et al., 2009). Loss of AB cell polarity is closely associated with epithelial-to-mesenchymal transition (EMT), a developmentally regulatory program whereby epithelial cells lose their differentiated characteristics and acquire mesenchymal characteristics through activation of a transcriptional program, which represses the expression of several polarity proteins and induces the expression of other proteins necessary for the mesenchymal phenotype (Thiery et al., 2009). TGFβ signalling through

Smad-dependent and Smad-independent pathways plays a pivotal role in AB cell polarity and

EMT (Barrios-Rodiles et al., 2005; Heldin et al., 2012; Ozdamar et al., 2005; Varelas et al.,

2010) (Figure 1.2). Occludin interacts with and restricts the localization of TβRI at tight junctions, where the receptor associates constitutively with Par6 (Barrios-Rodiles et al., 2005).

Upon TGFβ binding to the TβRI/TβRII receptor complex, TβRII phosphorylates Par6, which then associates with the Smad ubiquitination regulator factor-1 (Smurf1) that in turn targets the guanosine triphosphatase RhoA for degradation, thereby leading to loss of tight junctions

(Ozdamar et al., 2005). In addition, TGFβ signalling through the Smad pathway, directly activates expression of EMT-inducing transcription factors, such as Snail and Twist, which repress the expression of several proteins required for AB polarity, including E-catherin, Par3 and Crumbs (Heldin et al., 2012; Thiery et al., 2009). Moreover, TGFβ influences the activity of several other signalling pathways, including Notch, Wnt and integrin signalling, which also trigger EMT (Heldin et al., 2012; Thiery et al., 2009). Importantly, TGFβ signalling is controlled by the Hippo pathway members, WW domain containing transcription regulator 1

(Taz) and Yes-associated protein (Yap), which sequester Smad2/3 and are coupled to the

Crumbs polarity complex, thus suppressing TGFβ signalling at high cell densities (Varelas et al., 2010). Loss of AB cell polarity and cell-cell adhesions and acquisition of a mesenchymal 10

phenotype, including enhanced migratory potential and upregulation of ECM-remodelling metalloproteinases, allows cancer cells to detach from the primary tumour and invade the surrounding tissue (Thiery et al., 2009).

Luminal Space

Apical F-actin PATJ CRB PALS JAM Microtubule Claudin ZO3 Golgi Occludin ZO2 Tight ZO1 TIAM1 Rac1 junction PAR6 aPKC RhoA TRI PAR3 Cdc42 Nectin Afadin

Lateral Adherens Nucleus E-catherin -cat -actinin -cat junction p120 Desmocollin Desmoglein DLG Desmosomes Integrins SCRIB LGL Connexin Gap + junction Basal

Basement Membrane

Figure 1.2 Apical-Basal Polarity in Epithelial Cells

Epithelial cells are polarized along the apical-basal (AB) axis as detected by the orientation of cellular organelles (including the nucleus, Golgi and microtubules (MTs)), and asymmetrical distribution of lipids (not shown) and proteins (such as cell junction components). The epithelium is held together by cell-cell adhesions, comprised of tight junctions (green), adherens junctions (aquamarine), desmosomes (blue) and gap junctions (grey), and is attached to the basement membrane via cell-matrix adhesions mediated by integrins (yellow). Tight junctions are localized at the border of apical and basolateral domains and hinder the movement of proteins and lipids between these membranes. Both tight junctions and adherens junctions

together with the polarity complexes: CRB (components indicated in deep blue), PAR6 (components indicated in deep purple) and SCRIB (components indicated in deep orange) establish and maintain AB polarity. The polarity complexes function by antagonizing each other: the CRB and the PAR6 complexes establish the apical-lateral border while the SCRIB complex establishes the lateral-basal border. Also, tight junction and adherens junction components and polarity proteins interact with MTs (p120 catenin and PAR6; not shown) and/or the actin cytoskeleton components (shown in red) to regulate cell shape and polarity. TGFβ receptor type I (TβRI) interacts with occludin and PAR6 and regulates tight junction maintenance in response to TGFβ signalling. Figure adapted from (Martin-Belmonte and Perez- Moreno, 2012; Thiery and Sleeman, 2006).

1.1.5. Tumour cell migration

Acquisition of a migratory phenotype is critical for tumour cell invasive and metastatic capacity

(Friedl and Alexander, 2011; Nurnberg et al., 2011). Cell migration is facilitated primarily by the actin cytoskeleton, which consists of polymerized actin, also known as filamentous actin (F- actin), and several actin-binding proteins that regulate the rate and organization of actin polymerization (Campellone and Welch, 2010; Pollard, 2007; Ridley et al., 2003) (Figure 1.3).

Actin filaments are structurally polarized and grow rapidly at the plus end and slowly at the minus end (Campellone and Welch, 2010; Ridley et al., 2003). Actin nucleation factors, such as formins and the actin-related protein 2/3 (Arp2/3) complex stimulate actin filament elongation and branching respectively, while capping proteins, such as gelsolin, halt filament elongation by binding to the plus end of F-actin (Campellone and Welch, 2010; Pollard, 2007). Other actin- binding proteins regulate the pool of available actin by either binding actin monomers (such as profilin) or disassembling older filaments (such as members of the actin depolymerization factor

(ADF)/cofilin family) (Campellone and Welch, 2010; Pollard, 2007; Ridley et al., 2003).

Furthermore, cortactin proteins stabilize new branches while filamin A and α-actinin stabilize the whole cytoskeleton by cross-linking F-actin (Pollard, 2007; Ridley et al., 2003). In response to extracellular stimuli, the Rho (Ras homologous) family of small guanosine triphosphate

(GTP)-binding proteins, including Cdc42, Rac and Rho, act as the main regulators of the actin cytoskeleton (Campellone and Welch, 2010; Iden and Collard, 2008; Ridley et al., 2003) (Figure

1.3). Rho family proteins are GTPases that are activated by guanine nucleotide exchange factors (GEFs) and inactivated by GTPase activating proteins (GAPs) (Iden and Collard, 2008).

Rho GTPases and their GEFs and GAPs regulate several actin nucleators and actin-nucleation promoting factors (Campellone and Welch, 2010; Iden and Collard, 2008). For instance, Cdc42 and Rac, regulate the activity of Wiskott-Aldrich syndrome protein (WASP)/WASP family verprolin-homology protein (WAVE) proteins respectively, which in turn function to promote

Arp2/3-dependent actin polymerization (Campellone and Welch, 2010). Moreover, different

GTPases, including RhoA, RhoB, RhoC and RhoD, bind and activate several formins

(Campellone and Welch, 2010). Importantly, the expression of several actin assembly factors is misregulated in cancer and correlates with advanced disease (Nurnberg et al., 2011).

Figure 1.3 Remodelling of the Actin Cytoskeleton

Extracellular stimuli (1) induce activation of Rho GTPases (2), which activate actin nucleation promoting factors, such as WAVE/WASP proteins (3) that in turn bind and activate the ARP2/3 protein complex (4). Consequently, the Arp2/3 complex binds on the side of an existing actin filament (F-actin) and stimulates the assembly of new actin filaments, which are elongated at the plus (+) end by profilin (5). The force generated during actin polymerization pushes the membrane forward (6). Capping proteins terminate elongation (7). As actin hydrolyzes ATP at the negative (-) end, ADF/cofilin severs and depolymerizes ADP-actin filaments (8). Next, profilin binds monomeric actin and catalyzes the exchange of ADP for ATP (9), thus regenerating the pool of available actin. Formins remain attached to the positive end of actin filaments and promote filament elongation by profilin while preventing capping (10). Figure adapted from (Pollard, 2007).

Cell motility has been best studied and defined in two-dimensional (2D) surfaces using in vitro assays (Friedl and Wolf, 2009; Ridley et al., 2003). This research has revealed that cell motility is a multi-step and cyclical process that includes 1) generation of cell polarity along the front- rear axis of movement, 2) extension of cell protrusions in the direction of movement, 3) stabilization of protrusions at the leading edge by cell-matrix adhesions, 4) acto-myosin- mediated cell contraction that generates the necessary force for cell movement, and 5) disassembly of cell adhesions rearward of the leading edge and rear-end withdrawal (Friedl and

Wolf, 2009; Ridley et al., 2003) (Figure 1.4). First, a front-rear cell polarization is required for cell motility and involves the organization of the nucleus, the Golgi apparatus, the microtubule

(MT) cytoskeleton (including the centrosome), and cell protrusive activity toward the leading front of movement (Etienne-Manneville, 2008; Iden and Collard, 2008; Ridley et al., 2003).

This allows growth of MTs into the lamella where MT-delivered, Golgi-derived vesicles provide the necessary cellular components for protrusive activity (Ridley et al., 2003). Cdc42 is considered to be a master regulator of polarity in motile cells (Etienne-Manneville, 2008; Iden and Collard, 2008; Ridley et al., 2003) (Figure 1.4). In response to extracellular signals, several downstream pathways, including the Gβγ/p21-activated kinase 1 (PAK1)/PAK-associated GEF

(PIXα) pathway and the phosphoinositide 3-kinase (PI3K) pathway, stimulate the accumulation and activation of Cdc42 in a gradient toward the leading front of migrating cells (Iden and

Collard, 2008; Li et al., 2003; Ridley et al., 2003). Importantly, the Par6/Par3/aPKC polarity complex, which controls AB polarity, also plays a critical role in regulating front-rear polarity in migrating cells (Etienne-Manneville, 2008; Iden and Collard, 2008; Ridley et al., 2003) (Figure

1.4). For instance, activated Cdc42 binds Par6 and promotes activation of aPKC at the leading front (Etienne-Manneville and Hall, 2001), where aPKC phosphorylates and thus inactivates glycogen synthase kinase 3 beta (GSK3β) (Etienne-Manneville and Hall, 2003). This promotes 15

the association of adenomatous polyposis coli (APC) protein with DLG, which capture the MTs at the leading edge and thus allow targeted vesicle delivery to the leading front (Etienne-

Manneville and Hall, 2003). Moreover, the Cdc42/Par6/aPKC complex cooperates with the noncanonical Wnt5a pathway to regulate polarized reorganization of MTs in migratory cells

(Schlessinger et al., 2007). Furthermore, Par3 interacts with the Rac activator T-cell lymphoma invasion and metastasis-1 (TIAM-1) to regulate Rac-dependent actin polymerization at cell protrusions (Iden and Collard, 2008) (Figure 1.4). Next, the assembly of new actin filaments in the leading front of cell protrusions provides the force that pushes the cell membrane forward.

The actin polymerization machinery that drives cell protrusions is regulated primarily by Rac and Rho, which act in negative feedback loops to regulate each other’s activity (Iden and

Collard, 2008; Ridley et al., 2003). Here the Par6 polarity complex also plays a critical role

(Etienne-Manneville, 2008; Iden and Collard, 2008). For example, at the leading front, Par6 recruits the E3 ubiquitin ligase Smurf1, which ubiquitinates and promotes degradation of RhoA at the tips of cell protrusions (Wang et al., 2003) (Figure 1.4). Thus, the Par6-nucleated complexes have a central role in regulating front-rear cell polarity and protrusive activity during cell migration.

Stabilization of newly formed cell protrusions to ECM is necessary for efficient migration

(Friedl and Alexander, 2011; Ridley et al., 2003). Integrins, which are heterodimeric membrane receptors consisting of α and β chains, serve as the main linking proteins for the actin cytoskeleton and ECM (Friedl and Alexander, 2011; Ridley et al., 2003) (Figure 1.4). Upon binding to ECM proteins, such as collagen, fibronectin and laminin, integrins cluster and form focal complexes that are linked to the actin cytoskeleton via several adaptor proteins, including

talin, paxillin and vinculin (Friedl and Alexander, 2011; Mitra et al., 2005). In mesenchymal cells, focal complexes rearward of the leading edge evolve into larger and more stable focal adhesions by the concerted action of Rho/Rho kinase (ROCK) signalling and integrin-associated signalling complexes that include the focal adhesion kinase (FAK) and the sarcoma viral oncogene homolog (Src) kinase (Mitra et al., 2005). Focal adhesions are connected to thick bundles of F-actin, known as stress fibers, which undergo acto-myosin-dependent contraction and subsequently strengthen the focal adhesions. Myosin generates contractile force by moving parallel actin filaments relative to one another (Olson and Sahai, 2009). Myosin activity is positively regulated by the phosphorylation of the myosin light-chain (MLC) by a number of kinases, including the Rho kinase (ROCK), the MLC kinase (MLCK), and PAK, and is negatively regulated by the myosin phosphatase (protein phosphatase-1M (PP1M)), which is phosphorylated and inhibited by ROCK (Olson and Sahai, 2009; Ridley et al., 2003). Cell migration proceeds as cells undergo acto-myosin contraction and move over cell adhesions, which serve as traction sites (Ridley et al., 2003). Finally, the cycle of cell migration is completed as focal adhesions in the trailing edge are disassembled by the combined action of extracellular proteases (matrix metalloproteinases (MMPs) and a disintegrin and metalloproteinases (ADAMs)), intracellular proteases (calpain), kinases (including FAK and

Src), and Rho GTPases (Friedl and Alexander, 2011; Mitra et al., 2005; Ridley et al., 2003).

Front-Rear Axis

Golgi MT Actin-rich Vesicle Cell Protrusion MTOC

Nucleus F-Actin

Focal Stress Adhesion A Fibers Myosin

Focal Complex

A Extracellular ECM B Stimuli protein PM PM Receptor Integrins Src Src Paxillin Paxillin Cdc42 TalinFAK TalinFAK P PAR6 Vinculin Vinculin p190 aPKC PAR3 -actininP RhoGEF Smurf TIAM1 Rac1 P RhoA MLC RhoA GSK3 Actin remodelling ROCK MLCK APC P P DLG MLC Actomyosin MLC MLC contraction PP1M MTOC/Golgi orientation ROCK

Figure 1.4 Cell Migration As a Multistep Process

Migratory cells exhibit intrinsic polarity, characterized by the orientation of the nucleus, the Golgi and the microtubule (MT) organizing center (MTOC) along the front-rear axis. MTs provide cellular components, contained within vesicles, at the tip of protrusions. The Cdc42- PAR6 complex regulates cell polarity by controlling both organelle orientation and actin remodelling at cell protrusions (box A). Upon ligand binding, integrin clustering and signalling at the tip of cell protrusions leads to formation of focal contacts, which consist of a number of integrin-associated adaptor proteins and kinases that connect the cell membrane to the actin 18

cytoskeleton and regulate actin dynamics (box B). The focal adhesion kinase (FAK) is a key regulator of focal adhesions. On one hand, phosphorylation of p190RhoGEF by FAK stimulates RhoA activation and thus increases myosin-mediated contraction of stress fibers, which leads to maturation of focal complexes into focal adhesions. On the other hand, phosphorylation of α- actinin by FAK leads to the release of actin stress fibers and disassembly of focal adhesions. Figure adapted from (Iden and Collard, 2008; Mitra et al., 2005; Ridley et al., 2003).

Cell migration in vivo can occur in 2D surfaces, which are found in the walls of blood and lymph vessels and the pleura covering the lungs and thorax wall, or in 3D matrices of connective tissue where several sides of a migrating cell are in contact with other cells or the

ECM (Friedl and Alexander, 2011). Depending on the tumour type and the surrounding environment, cancer cells can adopt different forms of migration: single cell migration, which is enabled by loss of cell-cell adhesions, and collective migration, which occurs when intercellular adhesions are retained (Friedl and Alexander, 2011; Nurnberg et al., 2011). There are two types of single cell migration, mesenchymal and amoeboid, and several forms of collective migration, including clusters and strands (Friedl and Alexander, 2011). In the mesenchymal mode of migration, cells adopt a spindle-shaped morphology, form cell-matrix adhesions that mediate high traction forces, and proteolytically remodel the ECM (Friedl and Alexander, 2011). In the amoeboid mode of migration, cells exhibit a spherical morphology, lack focal adhesions, and do not proteolyse the matrix but instead squeeze through ECM gaps (Friedl and Alexander, 2011).

Collective cell movement requires cell-cell adhesion and coordination between the cells, which move coherently and proteolytically dissolve the ECM (Friedl and Alexander, 2011). Often, the cells at the leading front of cell clusters exhibit a mesenchymal phenotype and generate matrix tracks along which the rest of the cells follow. Alternatively, cancer cells recruit other cell

types, such as macrophages or fibroblasts, to produce metalloproteases that digest and remodel the matrix and thus promote collective cancer cell migration and invasion (Friedl and

Alexander, 2011). Evidence suggests that cancer cells migrate as single cells toward the blood circulatory system and as clusters of cells toward the lymphatic system (Giampieri et al., 2009).

Importantly, depending on the microenvironmental conditions, tumour cells can switch their migratory phenotype from one form to another (Friedl and Alexander, 2011). The cellular programs underlying this plasticity are currently unclear, although, it is known that TGFβ signalling promotes single cell motility at the invasive front of the primary tumour and in its absence tumour cells move collectively (Giampieri et al., 2009).

Different modes of cell migration are enabled by different types of cell membrane protrusive structures. Cancer cells form several types of protrusive structures, including filopodia, lamellopodia, pseudopodia, invadopodia and membrane blebs (Friedl and Alexander, 2011;

Nurnberg et al., 2011; Olson and Sahai, 2009). Filopodia are thin, finger-like protrusions formed by parallel bundles of actin filaments that are elongated by enabled/vasodilator- stimulated phosphoprotein (Ena/VASP) family of proteins, which prevent capping and branching of F-actin, and are held together by fascin (Mattila and Lappalainen, 2008; Olson and

Sahai, 2009). Filopodia act primarily as sensors for the extracellular environment (Mattila and

Lappalainen, 2008; Olson and Sahai, 2009). Moreover, filopodia enable mesenchymal cell movement and Rac-dependent amoeboid cell movement (Friedl and Alexander, 2011; Mattila and Lappalainen, 2008). Lamellopodia are thin, sheet-like protrusions formed by branched F- actin networks generated by the WAVE/WASP-Arp2/3 actin polymerization machinery (Mattila and Lappalainen, 2008; Olson and Sahai, 2009). Lamellopodia drive mesenchymal cell motility

on flat 2D substrates (Olson and Sahai, 2009). Pseudopodia are round, lamellopodia-like protrusions of variable size that are formed by a branched network of actin filaments that is remodelled by the WAVE/WASP-Arp2/3 nucleation machinery (Nurnberg et al., 2011; Olson and Sahai, 2009). Pseudopodia facilitate both mesenchymal and amoeboid migration in 3D matrices (Olson and Sahai, 2009). Invadopodia are similar to pseudopodia, except they have

ECM proteolytic activity and their presence correlates with the invasive capacity of tumour cells

(Nurnberg et al., 2011; Olson and Sahai, 2009). Migratory membrane blebs, which are distinct from apoptosis-associated blebs, form by bulging out of the plasma membrane that is devoid of cortical actin (a dense filamentous actin network just under the plasma membrane) (Nurnberg et al., 2011). Blebs are stabilized by newly polymerized cortical actin, in a process that requires formins and ezrin-radixin-moesin (ERM) proteins (Nurnberg et al., 2011; Olson and Sahai,

2009). The actomyosin machinery, regulated by RhoA/ROCK signalling, mediates contraction of cortical actin and thus plays a critical role in bleb-dependent movement (Nurnberg et al.,

2011; Olson and Sahai, 2009). Membrane blebbing enables rapid amoeboid cell migration in

3D environments.

Intensive research in the last few decades has shown that several chemokines (stromal cell- derived factor 1 (SDF-1), CCL5, CCL21), growth factors (HGF, EGF, FGF) and morphogens

(Wnt, TGFβ), released by tumour cells or activated stromal cells, are critical regulators of tumour cell migration (Friedl and Alexander, 2011). These factors engage several downstream signalling networks, including PI3K, Src, and c-Jun N-terminal kinase (JNK) pathways, to regulate Rho GTPase function and actin cytoskeleton dynamics (Friedl and Alexander, 2011).

Currently, we are capable of intersecting these pathways in order to inhibit tumour cell

migration and invasive capacity. Next, our challenge is to thoroughly understand how the tumour-associated stroma regulates these signalling pathways in cancer cells.

1.1.6. Metastasis – colonization of distant organs

Metastatic tumour formation depends on a series of discrete processes starting with tumour cells invading the adjacent tissue (local invasion), then entering the blood and lymphatic vessels

(intravasation), and being carried away to distant sites where cancer cells escape the circulatory systems (extravasate) and form small nodules (micrometastases) that eventually grow into fully developed macroscopic colonies (known as secondary tumours or metastases) (Hanahan and

Weinberg, 2011; Talmadge and Fidler, 2010) (Figure 1.5). Although metastasis is the main cause of death for cancer patients, the cellular processes and molecular mechanisms that regulate cancer cell intravasation, survival in the circulatory system, extravasation, specificity of metastastic sites and growth in secondary organs remain the least understood aspects of malignancy. During these processes, cancer cells are exposed to different environments and need to respond both physically and biochemically in order to survive (Talmadge and Fidler,

2010; Wirtz et al., 2011). Consistent with this, only a few cells within the primary tumour have the ability to overcome all the challenges encountered during the metastatic process (Talmadge and Fidler, 2010). During intravasation and extravasation, cancer cells squeeze through endothelial cell junctions and thus need to resist severe cell deformations (Wirtz et al., 2011).

This requires increased elasticity of the cytoplasm and the interphase nucleus that is driven by cytoskeletal remodelling, chromatin organization and nuclear envelope interactions, via linkers of the nucleus and cytoskeleton (LINC) proteins, SUN domain-containing proteins and Klarsicht homology (KASH) domain-containing proteins, with the cytoskeleton (Wirtz et al., 2011).

Once in the circulatory system, cancer cells must overcome shear stress, collisions with other cell types and immunologic surveillance (Talmadge and Fidler, 2010; Wirtz et al., 2011). Shear stress arises from movement of adjacent layers of fluid of different viscosity with different velocities (Wirtz et al., 2011). Activated platelets and leukocytes cluster with cancer cells and protect the latter from shear stress and immunologic surveillance (Gay and Felding-Habermann,

2011; Wirtz et al., 2011) (Figure 1.5). For extravasation to occur, circulating tumour cells have to be either trapped in small vessels, whose diameter is less than that of tumour cells, or adhere to the vessel walls. Adhesion of circulating tumour cells depends on shear flow and is mediated by receptor-ligand interactions between cancer cells and activated endothelium (Gay and

Felding-Habermann, 2011; Wirtz et al., 2011). Activated platelets and leukocytes also facilitate extravasation by mediating tumour cell adhesion to blood vessels, initially via selectins, which are cell adhesion molecules that bind cancer cell-expressed glycoconjugate ligands, including

Cd44 and mucins, and then via integrins, including αIIbβ3 (in platelets), β2 and α4β1 (in leukocytes) and αvβ3 (in cancer cells) (Gay and Felding-Habermann, 2011; Wirtz et al., 2011).

Moreover, it is thought that activated platelets release a multitude of growth factors, including

VEGF, PDGF, TGFβ, and EGF, which act on endothelial cells to induce vascular permeability, thus promoting tumour cell extravasation and invasion into the new microenvironment (Gay and

Felding-Habermann, 2011).

The location of metastatic sites has been a controversial topic in the field of cancer research

(Talmadge and Fidler, 2010). In 1889, Stephen Paget recognized that cancer cells show a specific affinity or tropism for particular organs (Paget, 1989). For instance, breast cancer cells metastasize typically to lungs, bones, liver and less often brain tissue while prostate cancer cells

almost exclusively metastasize to bone and liver but not lungs (Paget, 1989; Talmadge and

Fidler, 2010). As a result, Paget first proposed the “seed and soil” hypothesis, which upholds that organ specificity of metastasis corresponds to innate characteristics of the tumour cell, the

“seed”, as well as a particular organ microenvironment, the “soil”, which has to fit the requirements of cancer cells (Paget, 1989). This was later challenged by an alternative hypothesis, which states that the location of a metastatic site depends on the pattern of blood flow (Talmadge and Fidler, 2010). However, both the pattern of blood flow and the local microenvironment may have a complementary role in influencing the site at which secondary metastases grow (Talmadge and Fidler, 2010).

Once at the secondary organ, single tumour cells form micrometastatic lesions that eventually grow into macrometastatic tumours, in a process known as colonization (Talmadge and Fidler,

2010). However, colonization is thought to be a rate-limiting step in the metastatic cascade because many patients have a number of microscopic colonies that do not progress to macrometastases (Bissell and Hines, 2011; Talmadge and Fidler, 2010). It has been suggested that EMT, which is associated with cell cycle arrest in disseminating tumour cells, may prevent efficient growth of metastases (Brabletz, 2012) (Figure 1.5). This was recently supported by the finding that in secondary tissues, cancer cells undergo a mesenchymal-to-epithelial transition

(MET), which facilitates growth of metastases by promoting stem-cell like properties (Ocana et al., 2012), proliferation (Tsai et al., 2012) and alteration of the secretome (Korpal et al., 2011).

Importantly, EMT and MET are regulated by reciprocal interactions between EMT-inducing transcription factors, such as the zinc finger E-box binding homeobox (ZEB) family members and Snail, and MET-inducing microRNAs, including miR-200 and miR-34 families (Brabletz,

2012; Thiery et al., 2009). It is currently unknown what signalling molecules induce MET in metastatic cells, however, a TGFβ-related protein, the bone morphogenetic protein (BMP), stimulates expression of miR-200 family members to drive MET during somatic cell reprogramming (Samavarchi-Tehrani et al., 2010). In addition, BMP7 induces MET in renal fibroblasts and in prostate and breast cancer cells in vitro (Brabletz, 2012). So far, the role of

BMP7 in cancer cell metastasis has been investigated by overexpressing the ligand in cancer cells or pre-treating cells with the ligand before injection in mice (Buijs et al., 2012). Thus, in the future it will be important to determine whether BMP specifically induces metastatic colonization in secondary organs.

Primary Tumour Motility, Invasion and Intravasation

Extravasation Arrest and Adherence Circulation

Platelets & Leukocytes

Activation of Stroma Colonization

MET

Figure 1.5 The Metastatic Cascade

Metastasis is a multistep process that initiates as cancer cells (CCs) acquire a migratory and invasive phenotype through epithelial to mesenchymal transition (labelling is similar to Figure 1.1). Then, CCs detach from the primary tumour, migrate along ECM fibres, invade the local tissue and enter nearby blood/lymphatic vessels (intravasation). In the blood circulatory system, CCs aggregate with platelets and leukocytes, which protect CCs from shear force and immune attack and aid CC adhesion to the vessel walls at distant organs. There, CCs squeeze through the endothelium (extravasate) and migrate into the local tissue. Next, CCs undergo mesenchymal to epithelial transition (MET), which promotes CC proliferation. Moreover, metastatic CCs and the primary tumour CCs secrete signals that modify the microenvironment 26

in the secondary site by remodelling the ECM, activating local fibroblasts, and inducing angiogenesis and inflammation, which support secondary tumour growth (colonization).

1.1.7. The tumour microenvironment

Tumours are considered as aberrant organs composed of cancer cells and multiple other cell types, including fibroblasts, adipocytes, inflammatory cells, and vascular cells, which together with the extracellular matrix (ECM) components, such as proteoglycans, hyaluronic acid and fibrous proteins, constitute the tumour-associated stroma or microenvironment (Bissell and

Hines, 2011; Egeblad et al., 2010) (Figure 1.1). Stromal cells communicate reciprocally with tumour cells and each other via growth factors, cytokines, metabolites and ECM proteins to support tumour initiation, growth and metastasis (Bissell and Hines, 2011; Egeblad et al., 2010).

Evidence suggests that, in specific contexts, the microenvironment restrains and may revert the malignant phenotype in a dominant fashion over the genotype of tumour cells (Bissell and

Hines, 2011). Moreover, tumour malignancy correlates with stage-specific stromal activation, which in later stages is more supportive of tumour progression than in earlier stages (Bissell and

Hines, 2011; Egeblad et al., 2010).

In 1986, Harold Dvorak recognized that “all solid tumours, regardless of their site of origin, require stroma if they are to grow beyond a minimal size of 1 to 2 mm” (Dvorak, 1986).

Moreover, Dvorak proposed that: “tumours behave in the body like wounds and, in fact, induce their stroma by activating the host’s wound healing response” (Dvorak, 1986). During a wound

response, tissue injury results in haemorrhage or leakage of plasma, which clots rapidly and forms a gel, consisting of fibrin and fibronectin, that traps platelets and other blood cells

(Dvorak, 1986). Platelets normally take up and sequester a multitude of endogenous angiogenesis-regulatory proteins, such VEGF, FGF2, PDGF and TGFβ, in their alpha granules and secrete these molecules in a wound response (Folkman, 2007). These molecules then can act as mitogens and chemoattractants for connective tissue cells, including fibroblasts, inflammatory cells and endothelial cells that communicate in concert with each other to rebuild the damaged tissue. Dvorak hypothesized that similarly to a wound response, tumour-associated stroma becomes “reactive” upon leakage of plasma proteins from hyperpermeable blood vessels that initiate a cascade of events that result in replication and migration of inflammatory cells and fibroblasts and the formation of new blood vessels (Dvorak, 1986). Moreover, he realized that differently from wound healing, fibroblast proliferation and inflammation is persistent in the tumour-associated stroma and thus tumours can be considered as wounds that do not heal

(Dvorak, 1986).

In the last 25 years, research has shown that indeed cancer cells co-opt and manipulate the normal wound healing response in order to drive tumour growth both in the primary and secondary sites (Bissell and Hines, 2011; Folkman, 2007; Mantovani et al., 2008). As mentioned above, tumour cells secrete factors that stimulate recruitment of endothelial cells or bone marrow derived cells, which form new blood vessels (Folkman, 2007; Lyden et al., 2001), and inflammatory cells, such as macrophages and myeloid progenitors, which support tumour cell proliferation, evasion of apoptosis, angiogenesis and metastasis (Egeblad et al., 2010;

Mantovani et al., 2008) (Figure 1.2). Other stromal components are also involved in tumour

growth and metastasis (Figure 1.2). For example, mesenchymal stem cells (MSCs) interact directly with cancer cells by secreting Ccl5, a chemokine that binds its cognate receptor in cancer cells and promotes their invasive behaviour (Karnoub et al., 2007). Lymphatic endothelial cells produce cytokines, such as Ccl21, that guide tumour cells toward the lymphatic system (Tammela and Alitalo, 2010). Adipocytes are also emerging as important components of the tumour-associated stroma as they secrete collagen VI, adiponectin and IL-6 that promote tumour cells survival, growth and invasive capacity (Dirat et al., 2011; Iyengar et al., 2005;

Landskroner-Eiger et al., 2009).

The process of metastatic colonization is also tightly linked to the function of the microenvironment (Egeblad et al., 2010; Talmadge and Fidler, 2010). Accumulating evidence suggests that primary tumour cells prepare the microenvironment that they will eventually colonize (also known as the premetastatic microenvironment or niche), in order to support metastatic cell survival and growth by inducing angiogenesis, endothelial permeability and remodelling of ECM in the premetastatic niche (Egeblad et al., 2010; Gao et al., 2008; Kaplan et al., 2005; Peinado et al., 2012; Psaila and Lyden, 2009). Both signalling molecules (VEGF and lysyl oxidase) and microvesicles (including exosomes) secreted by cancer cells have been shown to mediate the communication with the premetastatic niche (Egeblad et al., 2010; Kaplan et al., 2005; Peinado et al., 2012; Psaila and Lyden, 2009). Considering the complexity of the cross talk between tumour and stromal cells, it is imperative that these interactions are fully understood in order to better target tumours in patients.

1.1.8. Fibroblasts as components of the tumour microenvironment

Fibroblasts constitute, in many cases of carcinoma, the most abundant cell population of the tumour-associated stroma (Kalluri and Zeisberg, 2006). Fibroblasts are highly heterogeneous cells, which are broadly defined as non-vascular, non-epithelial and non-inflammatory cells that are the primary components of the connective tissue (Kalluri and Zeisberg, 2006; Rasanen and

Vaheri, 2010). Fibroblasts are embedded within the matrix of the connective tissue and regulate

ECM turnover by synthesizing ECM components, including type I, III and V collagens, fibronectin, and MMPs (Egeblad et al., 2010; Kalluri and Zeisberg, 2006). Fibroblasts also secrete type IV collagen and laminins, which constitute the basement membrane that plays a critical role in the maintenance of epithelial AB polarity. Moreover, fibroblasts secrete growth factors that help maintain epithelial homeostasis (Egeblad et al., 2010; Kalluri and Zeisberg,

2006). Fibroblasts also function in wound repair by invading lesions and producing ECM that functions as scaffold for other cells and facilitates contractions of healing wounds (Egeblad et al., 2010; Kalluri and Zeisberg, 2006; Rasanen and Vaheri, 2010). Fibroblasts from wounds or scar tissue exhibit an activated phenotype, characterized by enhanced secretion of ECM proteins

(including type I collagen and tenascin C), ECM-degrading proteases (including MMP2, MMP3 and MMP9) and growth factors (including HGF, IGF, NGF, Wnt1, EGF and FGF2), and they have a higher proliferation rate than their normal counterparts (Kalluri and Zeisberg, 2006;

Rasanen and Vaheri, 2010). Similarly, fibroblasts in close proximity to cancer cells are also activated and are defined as cancer associated fibroblasts (CAFs). In wound healing, fibrosis and cancer, TGFβ (secreted by injured epithelial cells or cancer cells), intercellular-adhesion molecule 1 (ICAM1) and vascular-cell adhesion molecule 1 (VCAM1) (secreted by leukocytes), reactive oxygen species and complement factor C1 have been found to signal the transition of normal fibroblasts to their active phenotype (Bissell and Hines, 2011; Kalluri and Zeisberg, 30

2006). In contrast to wound healing, however, fibroblasts in fibrotic tissue and cancer are perpetually activated and the signals responsible for this effect are currently unknown.

Cancer-associated fibroblasts can be divided into at least two distinct subtypes 1) reprogrammed variants of connective tissue-derived fibroblasts and 2) myofibroblasts, which express α-smooth muscle actin (SMA) (Bissell and Hines, 2011; Egeblad et al., 2010). Myofibroblasts are rarely found in healthy epithelial tissue and their presence is more common in wounds and fibrotic tissue (Bissell and Hines, 2011; Egeblad et al., 2010; Kalluri and Zeisberg, 2006). The origins of CAFs is currently unclear, however, evidence suggests that they originate from local residing fibroblasts, from adipocytes, from pericytes, from bone marrow derived mesenchymal cells, or by epithelial- or endothelial- to mesenchymal transitions (Egeblad et al., 2010; Rasanen and

Vaheri, 2010; Thiery et al., 2009). So far, CAFs have been studied as a singular cell population and thus the contributions of each subtype of cells to cancer remain to be resolved. Several

CAF-expressed-markers, including PDGFRβ, carbonic anhydrase IX (CAIX), and α-SMA correlate with tumour recurrence, metastasis and/or poor prognosis (Rasanen and Vaheri, 2010).

CAFs are intimately involved in the initiation and progression of cancer (Kalluri and Zeisberg,

2006; Rasanen and Vaheri, 2010). They secrete several factors, including HGF and SDF-1, to promote epithelial cell neoplastic transformation and cancer cell proliferation (Bhowmick et al.,

2004; Orimo et al., 2005). In addition, CAFs interact with other stromal components via paracrine loops to stimulate tumour-enhancing inflammation, angiogenesis and lymphangiogenesis (Erez et al., 2010; Kalluri and Zeisberg, 2006; Orimo et al., 2005; Tammela and Alitalo, 2010). Also, both protease- and mechanotransduction force-dependent ECM remodelling by CAFs, is essential for tumour cell invasion and metastasis (Gaggioli et al., 2007;

Goetz et al., 2011; Kalluri and Zeisberg, 2006; Sternlicht et al., 1999). Although CAFs are predominant constituents of the tumour microenvironment and are key determinants of malignant progression, the paracrine signals that mediate direct cross talk between CAFs and tumour cells in metastasis are poorly understood.

1.1.9. Identification and function of cancer stem-like cells

In the last two decades, it has been shown that a subclass of cells within liquid (Lapidot et al.,

1994) or solid tumours (Al-Hajj et al., 2003), termed cancer stem-like cells (CSCs), have self- renewing capacity and even express markers and share transcriptional profiles with normal stem cells in the tissue of origin (Hanahan and Weinberg, 2011). Accordingly, CSCs are able to efficiently give rise to new tumours upon inoculation into recipient host mice (Hanahan and

Weinberg, 2011; Visvader, 2011). The origins of CSCs remain unclear and might be different from one tumour type to another. CSCs may rise from oncogenically transformed normal stem cells or from their partially differentiated progenitor cells, as observed for instance in intestinal and prostate cancer (Visvader, 2011). Alternatively, CSCs may rise from differentiated cells that acquire self-renewal capabilities, possibly through the involvement of EMT (Mani et al.,

2008; Morel et al., 2008), although the newly-discovered EMT activator, the paired-related homeobox transcription factor 1 (Prrx1), suppresses stemness traits (Ocana et al., 2012). This field of research is still in its early stages and a lot of questions remain with regards to the stem cell-like properties of CSCs and their importance in tumourigenesis (Clevers, 2011).

1.2. Noncanonical Wnt-Planar Cell Polarity Pathway

1.2.1. Planar cell polarity and its manifestations

Planar cell polarity (PCP), also known as tissue polarity, refers to organization of cell behaviour at the tissue level in a plane perpendicular to the axis of apical-basal cell polarity (Bayly and

Axelrod, 2011; Gray et al., 2011; Wallingford, 2012; Wu and Mlodzik, 2009). PCP requires the establishment of asymmetrical organization of protein complexes or subcellular structures within cells or asymmetric orientation of a group of cells with the alignment of these asymmetries across a tissue (Gray et al., 2011; Wu and Mlodzik, 2009). Moreover, PCP is coordinated at the tissue level with regards to the anteroposterior (AP), dorsoventral (DV) and left-right (LR) body axes (Gray et al., 2011).

Planar cell polarity (PCP) was first studied in the fruit fly, Drosophila melanogaster, where PCP is manifested in the AP orientation of hair bristles in the thorax and abdomen, the DV arrangement of ommatidia in the compound eye, the proximal-distal (PD) organization of leg cells and leg segments and the PD orientation of wing hairs from hexagonal epithelial cells

(Bayly and Axelrod, 2011; Wu and Mlodzik, 2009; Zallen, 2007) (Figure 1.6 A-B). Here, PCP directs polarity by employing a range of cellular mechanisms in different tissues under different biological contexts. For instance, in the fly wing, every epithelial cell reorganizes its actin cytoskeleton to generate a distally pointed actin-based hair (cytochrome) at its distal vertex

(Bayly and Axelrod, 2011; Wu and Mlodzik, 2009; Zallen, 2007) (Figure 1.6A). In the compound eye of Drosophila, a group of cells that form a unit eye, termed ommatidia, make a coordinated decision, which requires transcriptional activation and intercellular communication

(Zallen, 2007). In the ommatidia, the R3 and R4 photoreceptor cells differentiate in an 33

asymmetric fashion and the entire structure rotates as a unit in opposite directions in the dorsal and ventral halves of the eye (Bayly and Axelrod, 2011; Wu and Mlodzik, 2009; Zallen, 2007)

(Figure 1.6B). PCP can also influence the orientation of bristles, which consist of a four-cell cluster including a shaft, socket, sheath and neuron that arise from a single sensory organ precursor cell (Shulman et al., 1998; Zallen, 2007). Asymmetric and oriented division of the latter into its daughter cells is regulated by PCP signalling, thus indicating that the polarity of the precursor cell influences the organization of its descendants (Shulman et al., 1998; Zallen,

2007). PCP also regulates the morphogenesis of the fly dorsal thorax epithelium by regulating cell junction tension, which defines the pattern of local tissue contraction and shape of the epithelium mainly via oriented cell rearrangements (Bosveld et al., 2012). Thus, regulation of the actin cytoskeleton, oriented cell division and group-cell rotation are some of the cellular mechanisms by which PCP signalling regulates tissue polarity in the fly (Zallen, 2007).

Similarly, in vertebrates, PCP directs polarity in a multitude of different tissues and in different biological contexts (Figure 1.6). Some examples include, convergent and extension (CE) movements during gastrulation and neurulation, the orientation of hair bundles in inner ear sensory cells, the orientation of body hair, directional beating and assembly of cilia and migration of facial branchiomotor neurons (Gray et al., 2011; Wallingford, 2012). During gastrulation, CE movements of dorsal mesodermal cells allow the extension of the embryonic tissue along the AP axis and narrowing of the embryo along the mediolateral (ML) axis (Gray et al., 2011; Wallingford, 2012) (Figure 1.6C). At first, cells align, elongate and extend protrusions in the ML direction. Then, they form stable cell-cell contacts (intercalations) between anterior and posterior neighbouring cells within a plane (mediolateral intercalations)

and in the more shallow and deeper cell layers (radial intercalations), which allow the coordinated and enhanced movement of the cells toward the dorsal midline (Gray et al., 2011;

Wallingford, 2012). Similarly, during neurulation, CE movements of neuro-ectodermal cells allow the fusion of the neural folds to form the neural tube (Gray et al., 2011; Wang et al.,

2006). During both gastrulation and neurulation, PCP signalling regulates and coordinates polarized cell behaviour, such as cell polarity, protrusive activity, directional migration as well as oriented cell division (Gong et al., 2004; Gray et al., 2011). In addition, PCP signalling in vertebrates is essential for establishing left-right embryonic asymmetry by orienting the cilia in the AP direction and thus regulating the leftward cilia-driven fluid flow, which is required for left-right asymmetric gene expression in the mouse ventral node, zebrafish Kupffer’s vesicle and Xenopus gastrocoel roof plate (Antic et al., 2010; Borovina et al., 2010; Gray et al., 2011).

Here, the cells display planar polarization of the cilia, which protrude from near posterior portion of each cells’ apical surface and are tilted in the posterior direction in a PCP pathway- dependent manner (Gray et al., 2011). Moreover, in mice, PCP signalling controls the ML alignment of the kinocilium (microtubule-based cellular protrusion) and the adjacent stereocilia

(actin-based cellular protrusions), which make up the sensory hair bundles in the inner ear cells

(Gray et al., 2011; Wang et al., 2006) (Figure 1.6D). PCP signalling is also critical for the global AP alignment of body hair by regulating cell shape changes, cytoskeletal polarization and polarized gene expression in nascent hair follicles (Devenport and Fuchs, 2008; Gray et al.,

2011; Guo et al., 2004). Disruption of PCP signalling results in dysmorphogenesis of a wildly diverse array of tissues (Wallingford, 2012). As a result, aberrant PCP signalling leads to deafness, heart defects, polycystic kidney disease, ciliopathy and craniorachischisis (an open neural tube) (Gray et al., 2011; Simons and Mlodzik, 2008; Wu et al., 2011; Yang, 2012).

1.2.2. Molecular PCP components in the fly

The molecular components involved in establishing PCP were first discovered in Drosophila genetic screens that used the ommatidia, sensory bristles or wing hairs as model systems (Bayly and Axelrod, 2011; Lawrence et al., 2007; Wu and Mlodzik, 2009; Zallen, 2007). These studies allowed the discovery of a number of genes involved in PCP that were divided in two distinct groups, or modules, the core Frizzled/PCP module and the Fat/Dachsous module, according to the morphological manifestations observed upon their mutation or misexpression in the fly tissue (Bayly and Axelrod, 2011; Lawrence et al., 2007; Wu and Mlodzik, 2009; Zallen, 2007).

The core Frizzled/PCP module, consists of a seven-pass transmembrane protein, Frizzled (Fz,

Fzd in vertebrates); a four-pass transmembrane protein, van Gogh/Strabismus (Vang/Stbm, van

Gogh-like (Vangl) in vertebrates); a seven-pass transmembrane atypical catherin,

Flamingo/Starry Night (Fmi/Stan, the cadherin EGF LAG seven-pass G-type receptor (Celsr) in vertebrates) and the cytoplasmic proteins: Dishevelled (Dsh, Dvl in vertebrates), Prickle (Pk) and Diego (Dgo, Diversin/Inversin (Div/Inv) in vertebrates/mouse respectively) (Bayly and

Axelrod, 2011; Wu and Mlodzik, 2009; Zallen, 2007). Importantly, mutations of Dsh, Pk, Fmi or Dgo result in PCP defects within the mutant cells, indicating that these factors are only involved in conducting PCP signalling in a cell autonomous manner (Wu and Mlodzik, 2009).

In contrast, mutations of Fz and Vang, in addition to PCP defects within mutant cells, result in reversal of the polarity of adjacent, wild type cells, indicating that these factors act both in a cell autonomous and non-autonomous manner to coordinate and propagate polarity at a more global level (Lawrence et al., 2007; Wu and Mlodzik, 2009; Zallen, 2007). The expression of Fmi, which does not have a non-autonomous effect on cellular polarity, is required for the non- autonomous effects of Fz (Wu and Mlodzik, 2009). The second set of PCP genes, includes Fat

(Ft, Fat in vertebrates) and Dachsous (Ds, Dchs in vertebrates), two large atypical catherins that 36

can interact heterophilically across cell boundaries, and the transmembrane, Golgi protein, Four- jointed (Fj, Fjx in vertebrates), a kinase that acts in a graded fashion to modulate Ft/Ds activity

(Bayly and Axelrod, 2011; Blair, 2012; Goodrich and Strutt, 2011). Interference with normal

Ft, Ds and Fj expression results in randomization of PCP within mutant cells and reversal of the polarity of the neighbouring cells, indicating that these genes act in both autonomous and non- autonomous manners to regulate PCP (Lawrence et al., 2007; Zallen, 2007). Moreover, mutations in the Ft/Ds group alter the shape of the wings and legs and disturb tissue growth through regulation of the Hippo pathway (Lawrence et al., 2007; Sopko and McNeill, 2009). In addition to the two main PCP modules that include genes with a predominant PCP function, several genes, such as Rho GTPases, ROCK, aPKC and Scribble, have been identified that are required for polarization only in specific tissues and are referred to as “the effector genes”

(Bayly and Axelrod, 2011; Wu and Mlodzik, 2009). Moreover, recent genetic analyses in the fly have revealed novel regulators of PCP whose function has yet to be studied in detail (Weber et al., 2012).

1.2.3. Models of PCP signalling in the fly

Although great progress has been made in identifying the fly PCP factors, it is still unclear how planar cell polarity signalling is initiated, conducted within the polarized cells and propagated across the tissue (Bayly and Axelrod, 2011; Blair, 2012; Lawrence et al., 2007; Wu and

Mlodzik, 2009). Indeed, the initial signal that biases the asymmetrical subcellular localization of core Fz/PCP molecules and directs planar polarity at the tissue level remains unknown. Fz and Vang double mutant clones display phenotypes similar to single Fz mutant clones, suggesting that PCP signalling is instructed by Fz activity (Wu and Mlodzik, 2009). Moreover,

non-autonomous Fz signalling is required before core PCP molecules become asymmetrically localized, suggesting that Fz non-autonomous signalling has an instructive role in the global polarization of tissues (Wu and Mlodzik, 2009). Fz is the receptor of the Wnt morphogen, which could signal PCP from a distant source and thus instruct graded Fz activity. However, in the fly, there is no direct evidence that Wnt ligands act to trigger PCP. Genetic studies suggest that the fly Wnt4 (dWnt4) ligand can alter PCP in the eye and wing but this may be due to indirect effects on the expression and function of other genes (Wu and Mlodzik, 2009). The

Ft/Ds module was initially proposed to act upstream of the Fz/PCP core module to regulate tissue polarity in a global manner. However, the current consensus is that the Ft/Ds pathway acts in parallel to the Fz/PCP pathway and may influence core protein polarity only in particular contexts and tissues (Bayly and Axelrod, 2011; Blair, 2012; Gray et al., 2011; Lawrence et al.,

2007). As a result, the search continues for the mystery factor X that first signals PCP in the fly.

The molecular mechanisms employed by the core Fz/PCP module in conducting tissue polarity are better understood. Fz/PCP signalling is mediated by complex interactions between the core components that function in regulatory loops rather than a linear pathway (Wu and Mlodzik,

2009). This is accompanied by an underlying asymmetric distribution of core PCP proteins along the PCP axis, which is necessary for PCP signalling (Bayly and Axelrod, 2011; Wu and

Mlodzik, 2009) (Figure 1.6). For instance, in the fly wing cells, Vangl and Pk are enriched at the proximal and Fz, Dsh, and Dgo at the distal apical cell surface (Wu and Mlodzik, 2009).

Before PCP is established, the core PCP components are localized uniformly around the apical- lateral cell cortex in epithelial cells (Wu and Mlodzik, 2009). Frizzled recruitment and activity at the cell surface initiates apical recruitment of other core PCP components that interact with

each other to maintain their apical membrane association in a positive feedback loop (Wu and

Mlodzik, 2009). Polarization of the core components along the apical-basal axis is a prerequisite for their subsequent asymmetrical localization along the PCP axis (Wu and

Mlodzik, 2009). Once at the apical surface, Fmi-Fz complexes are endocytosed and transported distally in vesicles on microtubules aligned along the proximal-distal axis parallel to the cell surface (Shimada et al., 2006). Then, Fz, which is enriched at the distal cell surface, binds and retains Vang on the proximal surface of the neighbouring cell (Wu and Mlodzik, 2009).

Moreover, Pk antagonizes Fz-Dsh interaction while Dgo interacts with Pk and Vang and may counteract Pk activity in order to allow the Fz-Dsh complexes to form at the distal surface (Wu and Mlodzik, 2009). Fmi localizes to both proximal and distal surfaces, interacts homophilically in trans between neighbouring cells and is thought to stabilize the Fz/Dsh/Dgo and Vang/Pk complexes through interaction with Fz and Vang (Wu and Mlodzik, 2009). Once segregated in opposite sides of the cell, the Fz-Dsh-Dgo complex acts positively to promote

PCP signalling through activation of Dsh and its effectors, while the Vang-Pk complex functions to limit Dsh activation by feedback loop interactions (Wu and Mlodzik, 2009). Thus, heterophilic core protein interactions at cell boundaries result in subcellular asymmetry that is critical for conducting the PCP signal across the tissue.

1.2.4. PCP signalling in vertebrates

The core Fz/PCP genes are evolutionary conserved and required for PCP establishment in vertebrates (Gray et al., 2011; Wallingford, 2012; Wu and Mlodzik, 2009). The number of vertebrate core PCP homologs is considerably higher than in the fly, including ten Frizzled

(Fzd1-10), three Dvl (Dvl1-3), two Vangl (Vangl1-2), two Pk (Pk1-2) and three Celsr (Celsr1-3)

genes. Even though functional redundancy between the different homologs is common, single and compound mutants of several core PCP genes manifest PCP defects in vertebrates (Gray et al., 2011). Moreover, the phenomenon of Fzd and Vangl nonautonomy observed in the fly also occurs in vertebrates. For instance, morphological and molecular polarization of hair follicles in the mouse epidermis require non-autonomous functions of Vangl2 (Devenport and Fuchs,

2008). In addition, asymmetric localization of the core PCP molecules is conserved in vertebrates (Wallingford, 2012) (Figure 1.6 C-E). For instance, in the mouse node, overexpressed Pk2 and Vangl1 localize to anterior cell edges (Antic et al., 2010), while overexpressed Dvl localizes to the posterior cell edges (Hashimoto et al., 2010). Asymmetric distribution of core PCP proteins is also observed in dynamically remodelled tissue, where Pk accumulates in puncta in mesodermal and neuroectodermal cells (Ciruna et al., 2006; Narimatsu et al., 2009; Yin et al., 2008) (Figure 1.6C). The polarized subcellular localization of core PCP molecules in vertebrates suggests that the PCP signalling mechanisms are at least at some extent evolutionary conserved (Gray et al., 2011; Wu and Mlodzik, 2009). Indeed, Vangl2 and Celsr1 are dependent on one another for their proper asymmetric distribution along the AP axis in the developing epidermis (Devenport and Fuchs, 2008). Moreover, Fzd6 association with the membrane dependents on both Celsr and Vangl2 (Devenport and Fuchs, 2008). Furthermore, membrane association and asymmetric localization of Dvl and Pk depends on Fzd and Vangl expression (Ciruna et al., 2006; Takeuchi et al., 2003; Veeman et al., 2003; Yin et al., 2008).

Similar to the fly system, Pk inhibition of Frizzled-mediated membrane association of Dvl is also conserved in vertebrates (Carreira-Barbosa et al., 2003; Jiang et al., 2005). However, the core PCP molecules are not always asymmetrically distributed with the same partners as observed in the fly (Gray et al., 2011; Wallingford, 2012). For instance, in inner ear cells, Dvl accumulates at the lateral cell edges Wang, 2005 #433}, whereas Fzd3/6 accumulates with

Vangl2 and Pk along the medial cells edges (Montcouquiol et al., 2006; Narimatsu et al., 2009;

Wang et al., 2006) (Figure 1.6D). These findings suggest that in vertebrates other mechanisms are implicated in establishing PCP. Consistent with this, genetic studies indicate that a number of novel proteins regulate vertebrate PCP by employing several different molecular mechanisms

(Bayly and Axelrod, 2011; Gray et al., 2011; Wallingford, 2012). For instance, the receptor tyrosine kinase Ror2 binds Wnt5a and regulates PCP in the developing limb bud by inducing

Vangl2 phosphorylation and enhancing Vangl2 asymmetric localization in chondrocytes (Gao et al., 2011). The receptor-like tyrosine kinase (Ryk), a Wnt binding protein, regulates PCP by interacting with Vangl2 and promoting its stability (Andre et al., 2012). Vertebrate protein tyrosine kinase 7 (PTK7) regulates vertebrate CE movements (Yen et al., 2009) by recruiting

Dvl to the plasma membrane through Fzd (Shnitsar and Borchers, 2008) and receptor of activated protein kinase C 1 (RACK1) (Wehner et al., 2011). In addition, Smurf ubiquitin ligases control PCP and CE in mice by physically interacting with Dvl that is bound to Par6 and targeting Pk1 for ubiquitin-mediated degradation (Narimatsu et al., 2009). Moreover, XGAP, a

GTPase-activating protein for ADP ribosylation factors (ArfGAP), interacts with the

Par6/aPKC/14-3-3ε polarity complex to direct its localization to mediolateral ends of dorsal mesoderm cells and restrict their protrusive activity in CE movements in Xenopus (Hyodo-

Miura et al., 2006). The number of genes involved in vertebrate PCP signalling is rapidly increasing and future studies will be necessary to understand their function in different biological contexts (Wallingford, 2012).

A Wing hairs B Ommatidia P D D

Equator

5 Actin- R4 6 V based 8 7 hair R3 2 1

C D E Ependymal Convergent extension Cochlear hair cells multiciliated cells C R A M L L M L kinocilium

stereocilium

P cilia

basal bodies

Figure 1.6 PCP in Drosophila and Vertebrates

(A) The fly wing epithelium is polarized in the proximal (P) to distal (D) axis. Each cell extends an actin-based protrusion (hair) in the distal vertex (top panel). Fzd (dark green) and Dvl (light green) localize at the distal apical membrane while Vangl (orange) and Pk (yellow) localize at the proximal apical membrane (bottom panel). (B) In the developing fly eye, the ommatidia, which consist of a cluster of cells (including R3 and R4 photoreceptor cells), are organized along the equator (top panel). Fzd localizes to the R3 side of the R3/R4 membrane border while Vangl localizes to the R4 cell membrane (bottom panel). (C) During CE movements in vertebrates, cells extend protrusions in their medial (M) and lateral (L) ends, intercalate and move towards the medial axis thus redistributing along the anterior (A) to posterior (P) axis (top panel). Dvl and Pk puncta localize posteriorly and anteriorly respectively

(bottom panel). (D) In the mouse cochlea, hair cells are polarized along the ML axis, as revealed by the orientation of stereocilia (actin-based protrusions) and kinocilium (tubulin-based protrusion) (top panel). Dvl localizes laterally while Vangl, Pk and Fzd localize along the medial membrane (bottom panel). (E) In the mouse brain, ependymal cells are polarized along the caudal (C) to rostral (R) axis as revealed by the orientation of basal bodies and cilia (top panel). The asymmetrical distribution of Vangl is shown (bottom panel). Figure was adapted from (Goodrich and Strutt, 2011; Wallingford, 2012).

Unlike in the fly system, genetic studies in different vertebrate model systems have revealed that

Wnt morphogens play an important role in providing global regulation of tissue polarity

(Wallingford, 2012) and might even trigger the initial asymmetry that is propagated across the tissue (Goldstein et al., 2006; Gros et al., 2009). For instance, in zebrafish, silberblick (an orthologue of human Wnt11) (Heisenberg et al., 2000) and pipetail (an orthologue of human

Wnt5b) (Kilian et al., 2003) are specifically required for CE movements. Similarly, in Xenopus laevis, overexpressed Wnt11 (Tada and Smith, 2000) and Wnt5a (Moon et al., 1993;

Wallingford et al., 2001b) interfere with CE movements. Moreover, in mice, Wnt ligands regulate PCP in CE movements (Qian et al., 2007), primitive muscle fibre orientation (Gros et al., 2009), kidney tubule morphogenesis (Karner et al., 2009) and cartilage extension in mouse limb buds (Gao et al., 2011). Importantly, knypek, a heparan sulphate proteoglycan, potentiates

Wnt11 signalling and promotes polarized cell behaviours during gastrulation movements in zebrafish (Topczewski et al., 2001). Even though it is clear that Wnts are critical in regulating

PCP in vertebrates, the molecular mechanisms involved are poorly understood.

1.2.5. Wnt structure and post-translational modifications

Wnts belong to a family of secreted signalling molecules (19 members in mammals) that regulate a myriad of cell biological processes, including cell proliferation, differentiation, survival, polarity, migration and stem-cell self-renewal, during embryonic development, tissue homeostasis and disease (Willert and Nusse, 2012). The name “Wnt” was derived from the

Drosophila Wnt homologue, wingless (wg), and mouse Int1, an oncogene often activated upon insertional mutagenesis by the mouse mammary tumour virus (MMTV) in mammary carcinoma

(Rijsewijk et al., 1987). All Wnt proteins are characterized by the presence of 22 or 24 cysteine residues, most of which are thought to form intramolecular disulphide bonds that are critical for

Wnt structure and function (Willert and Nusse, 2012). More than thirty years after the discovery of Wnts, the crystal structure of Xenopus Wnt8 (XWnt8) in complex with the cysteine-rich domain (CRD) of mouse Frizzled8 was recently resolved (Janda et al., 2012). This crystal indicated that XWnt8 has a two-domain structure, analogous to a “hand” with “thumb” and “index finger” grabbing the Fzd8 CRD (Janda et al., 2012). The “thumb” contains the amino-terminal domain (NTD), comprised of seven alpha helices held together by five disulphide bonds, while the “index finger” contains the carboxy-terminal domain (CTD), composed of two beta sheets held together by six disulphide bonds (Janda et al., 2012). The contact point between the XWnt8 NTD (thumb) and the Fzd8-CRD is mediated by a palmitoleic acid lipid group that protrudes from serine 187 (a conserved site required for Wnt palmitoylation and function (Takada et al., 2006)) into a deep groove in the CRD of the receptor (Janda et al.,

2012). The contact point between the CTD of XWnt8 and receptor CRD is mediated by hydrophobic amino acids, which are also conserved among different Wnts (Janda et al., 2012).

Wnts are subject to posttranslational modifications, including glycosylation and palmitoylation, which drive their tight association with the plasma membrane and ECM and are required for

Wnt activity (Coudreuse and Korswagen, 2007; Port and Basler, 2010; Willert and Nusse,

2012). In particular, Wnts are palmitoylated in a manner that is dependent on Porcupine, a multi-transmembrane protein that resides in the endoplasmic reticulum (ER) and is essential for

Wnt secretion and function (Herr and Basler, 2012; Kadowaki et al., 1996; Takada et al., 2006).

In addition, some Wnt ligands, for instance Wnt5a and Wnt11, can also be sulfonated by tyrosyl protein sulfotransferase-1 (TPST-1) and this modification enhances their signalling activity (Cha et al., 2009). Moreover, glycophosphatidyinositol (gpi)-like anchors can also be added to Wnts, for instance Wnt1 and Wnt3a, in a manner that is dependent on the ER-resident glycoprotein

Oto, which regulates retention of Wnts in the ER and Wnt signalling (Zoltewicz et al., 2009).

From the ER, Wnt transport to the Golgi is facilitated by two p24 protein family members,

Éclair and Emp24, which are involved in cargo selection from the ER to Golgi (Port et al.,

2011), and from there to the plasma membrane by the seven-pass transmembrane protein

Wntless (Wls), whose function is essential for Wnt secretion (Banziger et al., 2006; Bartscherer et al., 2006; Goodman et al., 2006; Herr and Basler, 2012). Once at the plasma membrane,

Wnts, which are highly hydrophobic, depend on several factors to function as long range signalling molecules. In the fly, heparan sulphate proteoglycans (HSPGs) of the glypican family regulate Wg gradient formation and signalling by a restricted diffusion mechanism (Baeg et al., 2001; Han et al., 2005). In addition, high-density lipoproteins particles, which consist of a central core of triglycerides and cholesterol enveloped by an outer phospholipid monolayer that is scaffolded by apoliproteins, act to solubilize Wg and are required for long-range Wg signalling (Panakova et al., 2005). Recently, Secreted Wg-interacting molecule (Swim), a putative member of the Lipocalin family of extracellular transport proteins, has also been shown

to carry Wg and promote its long-range signalling (Mulligan et al., 2012). Finally, exosome- like vesicles secreted at the Drosophila neuromuscular junction carry Wg in association with

Wntless (Korkut et al., 2009) and mammalian Wnts are secreted on exosomes (Beckett et al.,

2012; Gross et al., 2012; Ratajczak et al., 2006). However, more investigation is required to establish the significance of these mechanisms in different Wnt-dependent processes.

1.2.6. Wnt signalling pathways

Wnts trigger several signalling cascades that are broadly categorized as canonical (Wnt-β- catenin pathway) and noncanonical pathways (including the Wnt-PCP pathway) (Angers and

Moon, 2009; Gao and Chen, 2010; Willert and Nusse, 2012) (Figure 1.7). Importantly, the signalling pathways activated by Wnts depend on the expressed repertory of receptors and downstream effectors within a particular cell type rather than Wnt intrinsic properties (Anastas and Moon, 2013; Willert and Nusse, 2012). For instance, both Wnt5a and Wnt11, which regulate a number of PCP processes, can also activate canonical signalling (Willert and Nusse,

2012). The Wnt-β-catenin signalling pathway ultimately leads to the regulation of a β-catenin- dependent transcriptional program (Angers and Moon, 2009) (Figure 1.7A). In the absence of

Wnt stimulation, cytoplasmic β-catenin is recruited into a “destruction complex”, consisting of adenomatous polyposis coli (APC) and Axin, which stimulate phosphorylation of β-catenin by glycogen synthase kinase 3 beta (GSK3β) and casein kinase 1 (Ck1). This leads to ubiquitylation and proteasomal degradation of β-catenin. The binding of Wnts to Frizzled, which act as G protein coupled receptors (GPCRs) (Angers and Moon, 2009), and lipoprotein receptor-related proteins 5 (LRP5) or LRP6 coreceptors leads to Dvl activation, which polymerizes and promotes clustering of Wnt-Fzd-LRP5/6 complexes (Gao and Chen, 2010). 46

This leads to phosphorylation of LRP5/6 by GSK3β and Ck1 and recruitment of Axin to the plasma membrane. Consequently, β-catenin is released from the destruction complex and translocates to the nucleus where it acts as a co-activator for the T-cell factor (TCF) family of transcription factors and activates transcription of Wnt target genes (Angers and Moon, 2009).

The noncanonical Wnt pathways, include at least two pathways, the Wnt-PCP and the Wnt-Ca+2 pathways, which are not only independent of β-catenin function but also antagonize Wnt-β- catenin signalling (Angers and Moon, 2009). In vertebrates, accumulating evidence shows that the Wnt-PCP pathway regulates the stability of core PCP proteins and their asymmetric distribution (Andre et al., 2012; Gao et al., 2011; Narimatsu et al., 2009). For instance, upon

Wnt binding to Frizzled receptors, Dvl is phosphorylated and recruits Smurf ubiquitin ligases, which target Pk for degradation (Narimatsu et al., 2009) (Figure 1.7B). Moreover, the Wnt-PCP pathway is directly connected to the actin cytoskeleton (Figure 1.7B). For example, Wnt binding to Fzd receptors promotes Dvl interaction with dishevelled associated activator of morphogenesis 1 (Daam1), a formin protein, and the Rho GEF (WGEF, weak-similarity GEF) that in turn activate the RhoA/ROCK pathway (Gao and Chen, 2010). Upon Wnt stimulation,

Dvl also binds Dvl-associating protein with a high frequency of leucine residues (Daple), which promotes the association of Dvl with aPKC and consequently Rac activation (Ishida-Takagishi et al., 2012). In addition, Wnt association with Ryk and Ror2 receptors triggers activation of

Src and JNK respectively (Angers and Moon, 2009; Gao and Chen, 2010; Nomachi et al., 2008).

The other noncanonical Wnt pathway, the Wnt-Ca+2 pathway, is less understood but may be part of a large signalling network with the Wnt-PCP pathway since regulation of intracellular calcium is critical for CE movements in vertebrates (Veeman et al., 2003; Wallingford et al.,

2001a). Activation of the Wnt-Ca+2 pathway, through Fzd and Ror2 receptors, triggers the release of intracellular calcium and subsequent activation of calcium-sensitive proteins such as calmodulin-dependent protein kinase II (CAMKII) and PKC (Angers and Moon, 2009) (Figure

1.7C).

LRP5/6 Ryk Ror2 A B Vangl Vangl C Fzd Fzd Fzd Wnt Wnt ECM Wnt Wnt Wnt PM P Axin G Dvl GG APC GSK3 P PAR6 P Daple Dvl Dvl Dvl Dvl CK1 FLNa Cdc42 WGEFDaam Src PLC P Daple PAR6 Dvl Dvl APC Pk JNK Axin Smurf aPKC GSK3 RhoA -cat IP3 + DAG CK1 Rac1 Cytosol ROCK [Ca+2] PKC

-cat TCF CAMK Nucleus

Figure 1.7 Wnt Signalling Pathways

(A) Overview of the canonical Wnt-β-catenin pathway. In the absence of Wnt, cytosolic β- catenin (β-cat) exists in a complex with adenomatous polyposis coli (APC) and Axin and is phosphorylated by glycogen synthase kinase 3 beta (GSK3β) and casein kinase 1 (Ck1). This leads to ubiquitylation and proteasomal degradation of β-cat. Upon Wnt binding to Frizzled (Fzd) receptor and LDL receptor-related proteins 5 (LRP5) or LRP6 coreceptor, Dvl is activated and inhibits β-cat degradation. As a result, β-cat accumulates in the nucleus where it activates transcription of Wnt target genes in cooperation with the T-cell factor (TCF) transcription factor. (B) Overview of Wnt-PCP pathways. Wnt-PCP signalling regulates the stability of core PCP proteins. For instance, upon Wnt stimulation, phosphorylated Dvl in complex with PAR6

interacts with Smurf, which ubiquitinates and targets Pk for degradation. Moreover, upon ligand binding, Ryk and Ror2 receptors interact with Vangl and regulate Vangl protein stability. Wnt-PCP signalling also regulates the actin cytoskeleton. For example, Wnt binding to Fzd leads to the association of Dvl with Dishevelled associated activator of morphogenesis (Daam) and weak-similarity GEF (WGEF) that in turn activate the RhoA/ROCK pathway. Dvl also interacts with Dvl-associating protein with a high frequency of leucine residues (Daple), which promotes association of Dvl with aPKC and Rac activation. Moreover, upon Wnt5a binding, Ror2 regulates the actin cytoskeleton by binding the actin cross-linking protein Filamin A (FLNa) and activating the Cdc42/c-Jun N-terminal kinase (JNK) pathway. (C) The Wnt-Ca+2 pathway. Wnt binding to Fzd receptors triggers ativation of heterotrimeric G proteins (Gα, Gβ, and Gγ) and subsequent activation of phospholipase C (PLC), which generates second messengers diacylglycerol (DAG) and inositol-1,4,5-triphosphate (IP3) that trigger a Ca+2 flux and activation of PKC and calmodulin-dependent protein kinase (CAMK). Figure was adapted from (Angers and Moon, 2009; Gao and Chen, 2010).

1.2.7. Role of PCP in cell motility and cancer

Core PCP genes have been found to be aberrantly expressed in a number of cancers (Wang,

2009). Moreover, several core PCP components, including Vangl1, Frizzled7 and Wnt5a regulate tumour cell growth, migration and metastasis (Wang, 2009). However, it is unknown whether these components function in the context of the core PCP pathway and what molecular mechanisms underlie the potential role of Wnt-PCP signalling in cancer (Anastas and Moon,

2013). PCP signalling has a prominent role in regulating cytoskeletal dynamics that underlie different cell behaviours, including mediolaterally polarized lamellopodia, mediolateral alignment of cellular long axes and large-scale, polarized cell intercalations, which drive CE movements during development and are conserved from ascidians to mice (Wallingford, 2012).

Thus, one outstanding question is whether core PCP signalling regulates cancer cell motility and metastasis in an analogous manner to its function during embryonic development.

1.3. Exosomes as Potent Mediators of Intercellular Communication

1.3.1. Modes of intercellular communication

Cell-cell communication involves more than soluble factors, such as chemokines, cytokines, growth factors and extracellular matrix proteins. Recent studies show that cells also use other prominent forms of communication that involve membrane microvesicles, tunnelling nanotubules, filopodial bridges (or cytonemes) and other specialized cell-to-cell contacts

(Abounit and Zurzolo, 2012; Bobrie et al., 2011; Sanders et al., 2013; Sherer and Mothes, 2008).

Membrane microvesicles contain biologically active molecules, including nucleic acids, lipids and proteins that are horizontally transferred and have a complex and broad array of activities in recipient cells. Exosomes are a specific subtype of secreted microvesicles that are distinguished by several characteristics such as site of origin, size, density and molecular content.

1.3.2. Introduction to exosomes

Exosomes are small membrane vesicles of endosomal origin that are secreted by many different cell types, including cancer cells, and are present in body fluids such as plasma, breast milk and ascites (Mathivanan et al., 2012). Exosomes have specific structural and morphological characteristics that include a lipid bilayer, a typical size of 30-100 nm in diameter, a cup-shaped morphology in electron microscopy, a flotation density of 1.10-1.21 g/ml in sucrose gradient and characteristic marker proteins that reflect their endosomal origin. Moreover, exosomes 50

contain a broad array of proteins, nucleic acids and lipids and act as potent intercellular signalling components (Ratajczak et al., 2006; Taylor and Gercel-Taylor, 2008). Exosomes 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 and

Johnstone, 1983). Consequently, for more than a decade, exosomes were considered as organelles that expel obsolete cellular components in the extracellular space. Then, B cell- secreted exosomes were shown to present antigens to T cells and act as important regulators of immunologic processes (Raposo et al., 1996). In the last decade, there has been increased interest and research in exosome function and a number of studies have demonstrated that exosomes are potent components of cell-to-cell communication in diverse physiological and pathological processes that are evolutionary conserved (Stoorvogel et al., 2002; Thery et al.,

2009; Yang and Robbins, 2011).

1.3.3. Exosome biogenesis

The mechanisms involved in exosome formation, cargo selection, secretion and internalization by recipient cells are poorly understood. Exosomes originate upon inward budding of the limiting endosomal membrane in multivesicular bodies (MVBs) and are released in the extracellular milieu upon fusion of MVBs with the plasma membrane. The machinery that drives MVB formation is thought to be directly relevant to exosome production (Babst, 2011;

Stoorvogel et al., 2002). Two main but mutually nonexclusive 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 and cholesterol and the activity of neutral sphingomyelinase 2 (nSMase2), which

processes sphingolipids to release ceramide, is necessary for exosome formation and cargo sorting (Kosaka et al., 2010; Mittelbrunn et al., 2011; Trajkovic et al., 2008). Ceramide may induce negative curvature of the endosome membrane and/or coalescence of lipid rafts, which are enriched in sphingolipids and ceramide, into larger membrane domains (Babst, 2011).

Moreover, clustering of lipids and proteins within the endosomal membrane may function as a sorting signal for exosomes and trigger vesiculation of the endosome membrane (Fang et al.,

2007; Stoorvogel et al., 2002; Vidal et al., 1997). In the second model, the Endosomal Sorting

Complex Required for Transport (ESCRT) pathway regulates exosome biogenesis and cargo selection (Baietti et al., 2012; Putz et al., 2012; van Niel et al., 2006). The ESCRT system, consisting of ESCRT-0, -I, -II and -III protein complexes, recognizes and retains ubiquitylated cargo in the endosome membrane, deforms the endosome membrane and then catalyses the abscission of the endosomal invaginations to form vesicles that contain sorted material (Babst,

2011; Henne et al., 2011; Raiborg and Stenmark, 2009). This system may be necessary for targeting of PTEN in exosomes, since PTEN secretion as an exosome component requires its ubiquitination by Nedd4-1 and adaptor proteins Ndfip1 (Putz et al., 2012). Moreover, 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). Recently, the cell surface transmembrane protein, Syndecan, and its cytoplasmic partner, Syntenin, were shown to interact with programmed cell death 6 interacting protein (PDCD6IP), also known as ALIX, which associates with TSG101 and CHMP4

(Odorizzi, 2006), to drive exosome membrane budding and cargo sorting (Baietti et al., 2012).

Thus, based on the current data, it seems that both modes of exosome formation and cargo sorting may in fact cooperate to yield lipid-driven and ESCRT-regulated exosome production

(Babst, 2011).

Once exosomes are formed, their secretion in the extracellular space is aided by Rab proteins

(such as Rab11, Rab27 and Rab35) and SNARE (soluble NSF attachment protein receptor) proteins which guide MVB docking and fusion with the plasma membrane respectively (Beckett et al., 2012; Bobrie et al., 2011; Gross et al., 2012; Peinado et al., 2012). Exosomes are released constitutively, although their secretion can be stimulated by increased cytosolic Ca+2, cellular stress or high rates of proliferation (Stoorvogel et al., 2002). After secretion, exosomes can function in the neighbouring cells or disseminate systemically (Peinado et al., 2012). However, very little is known about how exosomes deliver their cargo in target cells and whether they are transmitted across the target cells. Exosomes bearing preformed major histocompatibility complex (MHC) peptides can be captured by dendritic cells via ICAM1 ligand and integrin leukocyte function-associated antigen 1 (LFA1)-receptor interactions and be presented on the surface of the recipient cells without being internalized (Segura et al., 2007). Exosomes can fuse with the plasma membrane of the recipient cells in a process that depends on microenvironmental pH (Parolini et al., 2009). Moreover, exosomes can be internalized by the recipient cells by endocytosis upon ligand-receptor interactions (Morelli et al., 2004) or macropinocytosis in a receptor independent manner (Fitzner et al., 2011). There is little consensus on the mode of exosome interaction with recipient cells and it has been proposed that exosomes from different origins have different modes of interactions with the recipient cells

(Bobrie et al., 2011).

1.3.4. Exosome components

Proteomic analysis of exosomes secreted by different cell types has shown that they contain commonly present and cell type-specific proteins. Exosomes are rich in a set of common plasma membrane proteins (such as tetraspanins Cd9, Cd63, Cd81 and Cd82), cytosolic proteins

(including actin, tubulin, heat shock proteins and flotillin) as well as endosomal proteins from the Golgi apparatus, the endoplasmic reticulum and the endocytic pathway (such as Rab proteins) which reflect their origin from the endosomal membrane. In addition, exosomes contain distinct proteins, including antigen presenting molecules (MHC class I and MHC class

II proteins) and signalling molecules (including a variety of growth factors and receptors) that vary depending on the cell type from which they are secreted (Simons and Raposo, 2009; Thery et al., 2009; Zoller, 2009). Whereas common exosome proteins are mainly involved in exosome formation and secretion, cell-type specific molecules mediate specific functions upon exosome capture by recipient cells. For instance, 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., 2009a; Peinado et al., 2012).

Nucleic acids, including messenger RNA (mRNA), microRNA (miRNA), mitochondrial RNA

(mtRNA), retrotransposon RNA and single stranded DNA fragments have also been isolated in exosomes (Balaj et al., 2011; Mittelbrunn et al., 2011; Pegtel et al., 2010; Ratajczak et al., 2006;

Skog et al., 2008; Taylor and Gercel-Taylor, 2008; Valadi et al., 2007). Their presence is characteristic of the cell type of origin and they function to regulate gene expression in target cells (Mittelbrunn et al., 2011; Skog et al., 2008; Valadi et al., 2007). Importantly, specific nucleic acid cargo are enriched in exosomes when compared to their cell of origin (Balaj et al.,

2011; Hunter et al., 2008; Mittelbrunn et al., 2011; Skog et al., 2008; Taylor and Gercel-Taylor,

2008), suggesting that a targeting mechanism must exist that selectively packages nucleic acids in exosomes. Currently, an intensive effort is placed on the study of exosomes with regards to their function as vehicles of intercellular exchange of genetic material.

Lipids are also critical components of exosomes that function to regulate exosome biogenesis.

Generally, exosomes are enriched in sphingolipids (especially ceramide and gangliosides) and cholesterol (Laulagnier et al., 2004; Wubbolts et al., 2003), which are thought to facilitate exosome production and secretion. The exosome membrane is more rigid than the plasma membrane and both lipid and protein organization may be involved (Laulagnier et al., 2004).

Currently, it is unclear whether lipids function as signalling components of exosomes in recipient cells.

1.3.5. Exosome functions

Exosomes act as vehicles of intercellular exchange of multiple molecular components and as a result are involved in a multitude of physiological and pathological processes. Exosomes act as potent modulators of the immune system (Stoorvogel et al., 2002; Thery et al., 2009). Antigen presenting cells such as dendritic cells (DCs) or B cells release exosomes containing preformed peptide-MHC complexes that stimulate T cell activation (Raposo et al., 1996; Stoorvogel et al.,

2002; Zitvogel et al., 1998). In addition, the human placenta releases exosomes bearing MHC proteins and NKG2D ligands that induce maternal-fetal tolerance by inactivating T cells and natural killer cells respectively (Hedlund et al., 2009). Exosomes also facilitate transfer of

miRNAs from endothelial cells to smooth muscle cells in the vasculature to regulate endothelial cell response to shear stress and atherosclerosis formation (Hergenreider et al., 2012). As mentioned earlier, exosomes also carry Wnt morphogens across the fly neuromuscular junction

(Korkut et al., 2009) and in the wing imaginal disc (Gross et al., 2012).

Exosomes are also involved in a number of pathological processes, including cancer. Exosomes secreted by malignant cells modulate tumour-induced immune suppression, angiogenesis, stromal remodelling, tumour cell invasion and formation of the premetastatic niche (Al-Nedawi et al., 2009b; Peinado et al., 2012; Taylor and Gercel-Taylor, 2011; Thery et al., 2009).

Tumour-secreted exosomes suppress the immune system by inducing apoptosis of activated cytotoxic T cells (Kim et al., 2005), inhibiting NK function and impairing normal monocyte differentiation into dendritic cells (Valenti et al., 2006; Yu et al., 2007). Tumour exosomes can also induce angiogenesis and thus manipulate their environment to favour tumour growth and invasion (Sheldon et al., 2010; Skog et al., 2008). In addition, tumour cell-derived exosomes promote tumour metastasis by conditioning the premetastatic niche. For instance, exosomes from pancreatic adenocarcinoma cells promote metastasis by modulating the premetastatic niche in lungs and lymph nodes (Jung et al., 2009). Also, melanoma-derived exosomes promote tumour growth and metastasis by horizontally transferring the oncogenic HGF receptor to bone marrow derived cells (BMDCs), which in turn help prepare the premetastatic niche (Hood et al.,

2011; Peinado et al., 2012). Exosomes secreted by the tumour-associated stroma also play a critical role in tumourigenesis. For instance, exosomes secreted by activated platelets induce the expression of several proangiogenic factors and invasive factors in lung cancer cells and thus promote angiogenesis and metastasis (Janowska-Wieczorek et al., 2005). Moreover, exosomes

from tumour-associated macrophages (TAMs) transport oncogenic miRNAs into BCCs that promote their invasiveness (Yang et al., 2011).

1.3.6. Exosomes as biomarkers for disease

Exosomes accumulate in the plasma, ascites and pleural effusions of cancer patients and their utility as diagnostic biomarkers for cancer and other pathological conditions is very promising

(Andre et al., 2002; Peinado et al., 2012). Importantly, increased tumour exosomes in patient sera correlates with disease progression (Taylor and Gercel-Taylor, 2008) and increased amount of proteins in tumour exosomes in patient sera correlates with disease progression (Peinado et al., 2012). Furthermore, exosomes display characteristic tumour proteins and genetic material that can be used as diagnostic biomarkers (Balaj et al., 2011; Peinado et al., 2012; Skog et al.,

2008; Taylor and Gercel-Taylor, 2008). For instance, melanoma-secreted exosomes are enriched in tyrosinase-related protein-2 (TYRP2), very late antigen 4 (VLA-4), heat shock protein 70 (HSP70) and oncogenic HGF receptor, which are predictive of disease progression

(Peinado et al., 2012).

1.3.7. Exosomes as therapeutic tools

Exosomes have been extensively studied as therapeutic and drug delivery tools in different areas of medicine. For instance, exosomes from immature dendritic cells (Dex) bearing donor-MHC antigens induce donor specific allograft tolerance in recipient mice (Li et al., 2012b). Moreover,

Dex modified to express indoleamine 2,3-dioxygenase (IDO), a tryptophan-degrading enzyme that inhibits T cells by depleting them of tryptophan, reduced inflammation and reversed

established rheumatoid arthritis in a murine model (Bianco et al., 2009). The use of self-derived exosomes modified with tissue-specific proteins that target the exosomes to particular cells in the body promises to be an effective way of targeting delivery of nucleic acids and drugs (van den Boorn et al., 2011). Indeed, autologous exosomes modified to express a neuron-targeting cell surface protein were successfully used to deliver siRNA in the mouse brain (Alvarez-Erviti et al., 2011). Exosomes can also be used to deliver drugs to specific target cells with a higher efficiency than when the drugs are administered alone (Sun et al., 2010). Importantly, exosomes are stable in blood serum and cause less immunogenicity than other lipid-based drug-delivery methods (Alvarez-Erviti et al., 2011; Clayton et al., 2003; van den Boorn et al., 2011).

Clinical use of exosomes to target specific tumours has also proven to be effective (Viaud et al.,

2010) and tumour-secreted exosomes bearing tumour-specific antigens can induce an immune response upon transfer to antigen presenting cells such as dendritic cells (Andre et al., 2002;

Wolfers et al., 2001; Zitvogel et al., 1998). In a clinical trial, injection of autologous dendritic cell-derived exosomes in non-small cell lung cancer patients led to tumour regression and long- term stabilization (Morse et al., 2005). As a result, a number of groups are working on developing anticancer exosome-based vaccines and several clinical trials are currently under way (Viaud et al., 2010).

1.3.8. Tetraspanins – masters of multiple biological functions

Tetraspanin protein family members, including Cd9, Cd63, Cd81 and Cd82, are highly enriched in exosomes and are used as exosome markers (Thery et al., 2009; Zoller, 2009). Recent

evidence suggests that tetraspanins function as components of exosomes. For instance, exosomes produced by tetraspanin tetraspanin 8 (Tspn8)-over-expressing cancer cells promote endothelial cell branching and angiogenesis (Gesierich et al., 2006; Nazarenko et al., 2010). It has been proposed that tetraspanins provide a means to target proteins into the MVB pathway independently of the ESCRT system through the incorporation of the target proteins into tetraspanin-enriched membrane (TEMs) domains (Stoorvogel et al., 2002; Wubbolts et al.,

2003). Moreover, tetraspanins may facilitate exosome fusion with the membrane of the recipient cells (Nazarenko et al., 2010). However, the molecular mechanisms that mediate tetraspanin function in the context of exosomes remain largely unknown.

At the cellular level, tetraspanins regulate a plethora of biological functions, including cell morphology, adhesion, motility, invasion, fusion, signalling and protein trafficking (Hemler,

2008; Zoller, 2009). The tetraspanin family consists of at least 33 member proteins (20-30 kDa protein core) that are highly conserved from sponges to mammals and are defined by four transmembrane domains, a conserved cysteine-cysteine-glycine (CCG) motif and four-eight cysteines that form critical disulfide bonds in the second extracellular loop (Hemler, 2008; Levy and Shoham, 2005; Zoller, 2009) (Figure 1.8). The transmembrane domains are flanked by short intracellular amino- and carboxy-terminal tails, a small intracellular loop and a small and a large extracellular loop. The transmembrane domains contain polar residues that stabilize the tertiary structure of the protein and promote association between tetraspanins and other high- affinity protein partners (Hemler, 2008; Levy and Shoham, 2005; Zoller, 2009). The large extracellular loop (LEL) consists of a constant region that is necessary for homodimerization and a variable region that is evolutionarily divergent and is required for interactions with

laterally associated, nontetraspanin proteins and ligands (Hemler, 2008; Levy and Shoham,

2005; Zoller, 2009) (Figure 1.8). Tetraspanins are reversibly palmitoylated at intracellular and juxtamembrane cysteines (Charrin et al., 2009a; Hemler, 2008) possibly through the action of

DHHC2, a thiol-directed protein acyltransferase of the aspartate-histidine-histidine-cysteine

(DHHC) domain-containing protein family (Sharma et al., 2008). This modification is critical for tetraspanin association with other proteins, their function and protection from lysosomal degradation (Charrin et al., 2009a; Hemler, 2008; Sharma et al., 2008). Tetraspanins are found at the plasma membrane, in multiple intracellular compartments, such as early endosomes, late endosomes and lysosomes, and in extracellular vesicles. Some tetraspanins have a tyrosine- based (YXXØ, where X denotes any amino acid and Ø denotes an amino acid with a bulky hydrophobic side chain) sorting motif in their carboxy-terminal region that is thought to localize tetraspanins in specific intracellular compartments (Levy and Shoham, 2005; Zoller, 2009).

Moreover, it has been proposed that tetraspanins without this motif can be sorted through their association with other tetraspanins or associated proteins that contain a sorting motif (Levy and

Shoham, 2005; Zoller, 2009).

LEL Costant Variable

b SEL c G C C a C C e d Extracellular N E C E C Cytosol C C C C COOH NH2

Figure 1.8 Schematic Structure of Cd81 As a Prototypical Tetraspanin

Cd81 consists of 236 amino acids and has four transmembrane (TM) domains (TM1 (13-33), TM2 (64-84), TM3 (90-112) and TM4 (202-224), which contain conserved hydrophilic amino acids (asparagine (N) and glutamic acid (E) (indicated in blue circles)) and are flanked in the intracellular region by short amino (NH2)- and carboxy (COOH)- tails and a small loop, and in the extracellular region by a small extracellular loop (SEL) and a large extracellular loop (LEL). Juxtamembrane cysteines (C) that are palmitoylated in Cd81 (C6, C9, C80, C89, C227 and C228) are indicated in yellow circles. The LEL consists of a constant region, comprised of helices a, b and e, which is conserved among tetraspanins. The LEL variable region is comprised of helices c and d, and contains the conserved cysteine-cysteine-glycine (CCG) motif and cysteines (C156, C157, C175 and C190; indicated in pink circles) that form disulphide bonds (indicated in black lines). Figure adapted from (Delandre et al., 2009; Levy and Shoham, 2005).

In the context of a cell, tetraspanins associate with a multitude of partners including other tetraspanins and nontetraspanin proteins such as immunoglobulin superfamily members, MHC class I and II proteins, integrins, growth factor receptors, G-protein-coupled receptors, several peptidases and cytosolic signal transduction molecules (Hemler, 2008; Levy and Shoham, 2005;

Zoller, 2009) (Figure 1.9). Tetraspanin molecular interactions are classified as type I, II, III according to the strength of detergents that are required to disrupt them (Hemler, 2008; Levy and Shoham, 2005; Zoller, 2009). Direct protein-protein interactions (type I interactions) are rare and include tetraspanin homodimers and heterointeractions with other tetraspanins and immunoglobulin superfamily members (Hemler, 2008; Levy and Shoham, 2005; Zoller, 2009)

(Figure 1.9). The majority of tetraspanin-integrin and tetraspanin-tetraspanin heterointeractions are type II interactions, which are maintained under mild lysis conditions. Palmitoylation of tetraspanins and their associating proteins is critical for this type of interaction (Hemler, 2008;

Levy and Shoham, 2005; Zoller, 2009). Weak type III interactions occur mainly with kinases and are also stabilized by palmitoylation (Hemler, 2008; Levy and Shoham, 2005; Zoller, 2009).

Tetraspanins protrude only 4-5 nm from the plasma membrane and they are best known for their ability to organize themselves and their interacting partners in cholesterol and gangliosides-rich membrane domains (with an area of ~ 0.2µm2) known as tetraspanin-enriched microdomains

(TEMs) (Hemler, 2008). TEMs consists of core complexes of tetraspanins engaged in direct, type I protein-protein interactions that network with other protein complexes via looser type II and III interactions (Charrin et al., 2009a; Hemler, 2005, 2008; Levy and Shoham, 2005; Zoller,

2009). TEMs are thought to provide a scaffold by which tetraspanin-associated proteins are laterally organized and the strength of their response to external stimuli is enhanced (Levy and

Shoham, 2005; Zoller, 2009).

1.3.9. Cd81 – an exemplary tetraspanin

Cd81 is a highly conserved and ubiquitously expressed member of the tetraspanin family that is involved in a myriad of biological processes (Levy and Shoham, 2005; Levy et al., 1998). Cd81 is palmitoylated in six juxtamembrane cysteine residues (Delandre et al., 2009) (Figure 1.8) in a process that depends on the oxidation state of the cell (Clark et al., 2004). Palmitoylation of both Cd81 and its interacting partners is important for their association with each other

(Cherukuri et al., 2004; Clark et al., 2004; Delandre et al., 2009; Montpellier et al., 2011). Cd81 interacts physically and functionally with other tetraspanins such as Cd9, Cd82 and Cd151 as well as nontetraspanin molecules such as EWI proteins, MHC-class II molecules and integrins

(Hemler, 2008; Levy and Shoham, 2005; Levy et al., 1998) (Figure 1.9). Cd81 associates with a high affinity with Cd9, which among tetraspanins is most closely related to Cd81 (Hemler,

2008; Levy and Shoham, 2005). Both Cd81 and Cd9 interact with high affinity and stoichiometry with immunoglobulin (Ig) superfamily member 8 (Igsf8; also known as EWI-2) and prostaglandin F2 receptor negative regulator (Ptgfrn; also known as EWI-F), two members of the Ig protein subfamily characterized by four-six extracellular Ig-domains and a glutamate- tryptophan-isoleucine (EWI) motif, a single transmembrane domain and a short cytoplasmic tail

(Charrin et al., 2009a; Hemler, 2008). Cd81 supports maturation and surface expression of

Igsf8 (Hemler, 2008). Both Igsf8 and Ptgfrn often mediate Cd81 and Cd9 function and link the tetraspanin web to the actin cytoskeleton through their direct association with ERM proteins or

α-actinin (Charrin et al., 2009b; Coffey et al., 2009; Gordon-Alonso et al., 2012; Montpellier et al., 2011; Sala-Valdes et al., 2006). Interestingly, Cd81 and Cd9 also interact with high affinity with GPR56 (Little et al., 2004), a G-protein coupled receptor (GPCR), which belongs to a family of adhesion molecules that includes the core PCP component Celsr (Xu and Hynes,

2007). 63

Cd19 Occludin Cd81 Claudin Cd82 homodimers Cd9 Vangl

Integrins Cholesterol GPR56 Ptgfrn Extracellular Igsf8

Sphingolipid

Cytosol PKC Rac1 G q/11 ERM -actinin G G

Figure 1.9 Cd81 and Tetraspanin-Enriched Microdomains (TEMs)

Tetraspanins act primarily as scaffolding proteins that organize their laterally-interacting partners in TEMs, which are rich in cholesterol and sphingolipids. Cd81 forms homodimers or heterodimers with Cd9 (another tetraspanin most closely related to Cd81) via direct and strong interactions (type I interactions). In addition, Cd81 interacts directly with Cd19 and indirectly with claudin and occludin via Cd9 to regulate hepatitis C virus (HCV) entry in cells. Cd81 also associates with other tetraspanins, including Cd82, in a palmitoylation-dependent manner (type II interaction). Note, Cd82 associates with Vangl1 (Zoller, 2009). Cd81 interacts directly with integrins (α4β1) to regulate B cell adhesion under shear flow condition. Cd81 also associates with protein kinase C (PKC) and may mediate phosphorylation of integrins by PKC (Zhang et al., 2001). In addition, Cd81 forms type I interactions with GPR56, a G-protein coupled receptor (GPCR) that signals via heterotrimeric G proteins (Gα, Gβ, and Gγ). Moreover, Cd81 forms type I interactions with Ig domain containing EWI proteins (Ptgfrn and Igsf8), which interact directly with ERM (ezrin, radixin and moesin) proteins or with α-actinin (Igsf8 only) and thus link the actin cytoskeleton to Cd81 (via type III interactions). Recently, it was found that Rac1 interacts directly with the C-terminal domain of Cd81 to regulate cell protrusive activity. Figure adapted from (Levy and Shoham, 2005). 64

Even though Cd81 functions in a multitude of cellular processes, Cd81 null mice are born in the expected Mendelian ratios, develop normally and are viable (Maecker and Levy, 1997;

Miyazaki et al., 1997; Tsitsikov et al., 1997). However, their fertility is impaired by approximately 40% due to deficiency in sperm-egg fusion (Rubinstein et al., 2006).

Importantly, the fertility of Cd9 knockout mice is also impaired by approximately 95%, while double knockout Cd9 and Cd81 mice are completely infertile (Rubinstein et al., 2006), suggesting redundancy in tetraspanins. Moreover, injection of exogenous Cd81 or Cd9 mRNA in Cd9 deficient oocytes restores the fertilization rate up to 50% and 90% respectively (Kaji et al., 2002). Thus it has been proposed that Cd81 and Cd9 have complementary roles and can compensate for each other in sperm-egg fusion events (Rubinstein et al., 2006).

Cd81 was discovered in 1990 as a target of an antibody that inhibited the proliferation of a lymphoma B cell line; hence, Cd81 is also known as target of antiproliferative antibody 1

(TAPA-1) (Levy and Shoham, 2005). Importantly, Cd81 deficient mice exhibit a weaker antibody response to antigens (Maecker and Levy, 1997; Miyazaki et al., 1997; Tsitsikov et al.,

1997). As a result, Cd81 has been extensively studied for its role in the regulation of the immune system, particularly in B and T cell proliferation and activation (Levy and Shoham,

2005). Cd81 lowers the threshold for B cell activation by associating with Cd19, a signalling molecule that functions together with the complement receptor Cd21 (Levy and Shoham, 2005).

Cd81 acts by translocating the Cd19-Cd21 co-receptor complex that is bridged to the B-cell receptor (BCR) to detergent resistant membranes (DRMs) where downstream cytoplasmic effectors are activated (Levy and Shoham, 2005). In addition, Cd81 is required for Cd19 glycosylation and might chaperone its transport to the plasma membrane and/or stabilize its

expression (Levy and Shoham, 2005). Accordingly, B cells from Cd81 null mice express less

Cd19 on the cell surface (Maecker and Levy, 1997; Miyazaki et al., 1997; Tsitsikov et al.,

1997). On T cells, Cd81 associates with T cell coreceptors, Cd4 and Cd8, and provides a costimulatory signal with the Cd3 subunit of the T cell receptor to support T cell maturation

(Levy and Shoham, 2005). Moreover, Cd81 dynamically concentrates in the immune synapse, a close contact between T cells and antigen presenting cells, such as B cells that is essential for T cell activation (Levy and Shoham, 2005).

Cd81 is also a central regulator of cellular events required for hepatitis C virus (HCV) infection and HCV-induced immune pathogenesis (Levy and Shoham, 2005; Pileri et al., 1998). In hepatocytes, Cd81 aids HCV attachment to cells by interacting with the HCV envelope (E2) glycoprotein via its LEL (Flint et al., 2006; Pileri et al., 1998). Moreover, Cd81 promotes viral entry by triggering signalling cascades that lead to activation of Rho GTPase family members followed by actin-dependent relocalization of E2-Cd81 complexes to cell-cell contacts where

Cd81 colocalizes with other HCV coreceptors, Claudin-1 and Occludin (Brazzoli et al., 2008;

Evans et al., 2007; Ploss et al., 2009). In B cells, it is thought that Cd81 binding of E2 leads to superactivation of the Cd19-Cd21 complex, which engages BCR bound to E2 and signals B cell proliferation that eventually results in lymphoma formation (Levy and Shoham, 2005). Cd81 is also involved in the infectivity of other pathogens, including human T cell leukemia virus-1

(HTLV-1) (Imai and Yoshie, 1993), human immunodeficiency virus (HIV) and influenza virus

(Thali, 2011), human malaria parasite (Plasmodium falciparum) and rodent parasite

(Plasmodium yoelii) (Silvie et al., 2003), although the mechanisms of action are poorly understood.

1.4. Overview of Thesis

Tumour initiation, growth and metastasis is regulated by a multitude of interconnected molecular networks that mediate cross-talk between tumour cells and stromal components.

Hence, tumours are considered as aberrant organs that are supported by several stromal cells.

Fibroblasts are the most abundant cellular component of the tumour-associated stroma and play an important role in tumour progression. However, the role of fibroblast-secreted factors in cancer cell motility and metastasis are poorly understood. Acquisition of a migratory phenotype is critical for cancer cell invasion and metastasis. The Wnt-PCP pathway regulates cell movements during vertebrate development. However, the role of Wnt-PCP signalling in tumour cell migration is currently unclear. Wnt ligands are palmitoylated and thus are insoluble in the extracellular milieu. In order to signal, Wnt molecules depend on other extracellular factors, including heparan sulphate proteoglycans, lipoproteins particles or exosomes, to deliver them to target cells. The relative significance of each mode of Wnt transportation in different biological contexts has yet to be established. Furthermore, exosomes are emerging as potent modes of intracellular communication and their role in different biological contexts has yet to be explored.

In my major PhD work presented in Chapters 2 and 3, I describe my findings that fibroblast cells, including CAFs, secrete exosomes, which promote breast cancer cell protrusive activity, motility and metastasis. Moreover, I show that tetraspanin Cd81 is required for fibroblast- induced breast cancer cell metastasis. In Chapter 3, I present evidence that the Wnt-PCP pathway mediates fibroblast-induced breast cancer cell protrusive activity, motility and metastasis. In addition, I show that core PCP components are asymmetrically distributed in exosome-induced breast cancer cell protrusions. Finally, I demonstrate that fibroblast exosomes

are loaded with Wnt ligands upon trafficking in breast cancer cells. In Chapter 4, I summarize my findings and consider their significance and implications.

Chapter 2 : Fibroblast-secreted Exosomes Regulate Breast Cancer Cell Motility and Metastasis

A version of this chapter was published by:

Valbona Luga, Liang Zhang, Alicia M. Viloria-Petit, Abiodun A. Ogunjimi, Mohammad R. Inanlou, Elaine Chiu, Abdel Nasser Hosein, Marguerite Buchanan, Mark Basik, and Jeffrey L. Wrana

in Cell 2012 Dec 21;151(7):1542-56.

Attributions:

The mouse experiments in Figures 2.3C-F and 2.8 were performed together with Dr. Liang Zhang and Dr. Alicia M. Viloria-Petit. Media fractionation experiments in Figure 2.4 were performed together with Dr. Abiodun A. Ogunjimi. The mass spectrometry experiments in Figures 2.5C and 2.6C and the microarray data analysis in Figure 3.9A were performed by Dr. Liang Zhang. All the remaining experiments were performed by Valbona Luga.

2.1. Introduction

The tumour-associated stroma, which consist of several cellular components and extracellular matrix proteins, plays a critical role in supporting tumour growth and metastasis (Hanahan and

Weinberg, 2011; Kalluri and Zeisberg, 2006). Stromal cells cross-talk with tumour cells via complex and reciprocal signalling networks that are mediated by growth factors, chemokines and other ECM proteins. Fibroblasts are the most abundant cellular component of the tumour stroma and participate in a number of processes, such as tumour-stimulated angiogenesis, inflammation and ECM-remodelling, which promote tumourigenesis (Erez et al., 2010; Gaggioli et al., 2007; Goetz et al., 2011; Orimo et al., 2005). Cancer-associated fibroblasts respond to tumour signals by secreting ECM proteins, proteases and growth factors that in turn promote tumour cell transformation, proliferation and metastasis (Bhowmick et al., 2004; Gaggioli et al.,

2007; Goetz et al., 2011; Orimo et al., 2005). Even though CAFs are critical constituents of the tumour stroma, the paracrine signals that mediate communication between CAFs and cancer cells in metastasis remain poorly understood.

Exosomes are membrane microvesicles secreted by many different cell types and have specific functions in different biological contexts, including cancer (Al-Nedawi et al., 2009b; Taylor and

Gercel-Taylor, 2011; Thery et al., 2009). Exosomes originate as intraluminal vesicles in large multivesicular bodies (MVBs) and are released in the extracellular milieu upon exocytic fusion of MVBs with the plasma membrane. Exosomes have the same topology as a cell and contain a broad array of biologically active constituents, including growth factors, oncoproteins, microRNAs and sphingolipids (Simons and Raposo, 2009; Thery et al., 2009; Zoller, 2009).

Several studies have shown that tumour cells secrete exosomes that mediate tumour-induced

immune suppression, angiogenesis, stromal remodelling, tumour cell invasion and formation of the premetastatic niche (Al-Nedawi et al., 2009b; Peinado et al., 2012; Taylor and Gercel-

Taylor, 2011; Thery et al., 2009). Although, fibroblast cells can secrete exosomes, it is not known whether CAFs secrete exosomes that regulate cancer cell metastasis. Here, I show that exosomes secreted from fibroblasts, including human primary CAFs, stimulate BCC protrusive activity, motility and metastasis through tetraspanin Cd81.

2.2. Materials and Methods

2.2.1. Cell Culture

MDA-MB-231 cells were provided by Dr. Robert Kerbel (Sunnybrook Health Sciences Centre,

Toronto, Canada) and were cultured in RPMI-1640 medium (Gibco) with 5% fetal bovine serum

(FBS) (HyClone). L cells were purchased from ATCC (CRL-2648) and were cultured in

Dulbecco’s modified Eagle’s medium (DMEM) (Gibco) with 10 % FBS. SUM-159PT cells were purchased from Asterand and were cultured in Ham’s F12 medium (Gibco) with 5 % FBS,

5 mg/ml insulin (Sigma-Aldrich), 1 mg/ml hydrocortisone (Sigma-Aldrich) and 10 mM Hepes.

MDA-MB-468 cells were purchased from ATCC (HTB-132) and were cultured in DMEM/F-12

(1:1) medium (Hyclone) with 10 % FBS. T-47D cells were purchased from ATCC (HTB-133) and were cultured in DMEM with 10 % FBS and 5 mg/ml insulin. EMT-6 cells were cultured in DMEM with 10 % FBS. EpRas and EpH4 cell were a gift from Dr. Martin Oft (Oft et al.,

1996) and were cultured in DMEM with 10 % FBS. Human CAFs from invasive breast carcinoma tissue (Hosein et al., 2010) were provided by Dr. Mark Basik (Lady Davis Institute,

Montreal, Canada) and were cultured in DMEM with 10 % FBS. Cells were incubated at 37 °C in a humidified atmosphere containing 5 % CO2.

2.2.2. Conditioned Media Preparation

Cells were grown to confluency and then were rinsed twice and incubated with DMEM supplemented with 50 U/ml penicillin/streptomycin (Gibco), 0.25 mg/ml amphotericin B

(Gibco) and either 0 % FBS (for L cells) or 0.2 % FBS (for human CAFs). Three days after starvation, the conditioned media was collected, filtered through a 0.2 mm pore size filter

(Nalgene) and stored at 4 °C for up to eight weeks.

2.2.3. Single Cell Motility Assay

Cells were suspended at very low density in assay media prepared from mixing equal volumes of RPMI supplemented with 5 % FBS with either DMEM or conditioned media. The cell suspension was transferred onto a 12-well tissue culture dish and imaged within one hour of cell seeding. Phase-contrast time-lapse movies were obtained using a Leica microscope (DMIRE2) equipped with a 10 x N-Plan objective lens, an Orca-ER camera (Hamamatsu) and a MS-2000 xyz automated stage (ASI). Images were collected at a rate of 6 images/hour for 18 hours using

Volocity software (PerkinElmer). The cells were incubated at 37 °C in a humidified atmosphere containing 5 % CO2 (Pathology Devices, Inc.). The speed of each cell was determined manually by tracing the cell periphery every hour for 18 hrs using Volocity. Statistical significance was calculated using two-tailed unpaired t test using Prism (GraphPad Software, Inc.).

2.2.4. 3D Matrigel Cultures

The cells were suspended in media prepared as for the motility assay (described above) and supplemented with 2 % growth factor reduced Matrigel (356231, BD BioSciences). Then, the cell suspension was transferred with a precooled pipette tip to a 4-well chambered coverglass

(155383, Nunc), previously covered with 100 % Matrigel that was allowed to solidify at 37 °C for 40 minutes. The media was changed every day and images were acquired three days after treatment.

2.2.5. Orthotopic Mouse Model of Breast Cancer Metastasis

The mouse experiments were performed according to protocols approved by the animal facility at Toronto Centre for Phenogenomics (TCP) (Toronto, Canada). Ten-eleven weeks old female

C.B-17 SCID mice (Charles River, Canada) were anesthetized with isofluorane and injected with analgesic (Metacam). Then, 1x104 L cells and 3x106 MDA-MB-231 or SUM-159PT cells, suspended in 50 µl of phosphate buffer saline (PBS), were injected into the right fourth inguinal mammary fat pad (mfp) of mice. Four weeks post injection, the mice were sacrificed under sedation and tumours and lungs were collected for analysis.

2.2.6. Immunohistochemistry

After collection, the lungs were immediately fixed in 10 % buffered formalin phosphate (SF100-

4, Fisher Scientific) overnight. Then, the samples were embedded in paraffin and sectioned

coronally at levels 100 mm apart by staff at the pathology laboratory at TCP. The paraffin- embedded sections were then processed for immunohistochemistry by Dr. Liang Zhang as follows. The samples were incubated in 3 % H2O2 in PBS for 30 min in order to inactivate endogenous peroxidases and then transferred in 10 mM sodium citrate buffer (pH 6.0) and heated using high power in a microwave oven for 12 min to allow antigen retrieval. Anti- human vimentin antibody (M0725, Dako, 1:50), Dako Envision+ (K4001, Dako) and diaminobenzidine (D6190, Sigma-Aldrich) were used sequentially to develop immunohistochemistry. Subsequently, the samples were counterstained with hematoxylin

(GHS332, Sigma-Aldrich). The samples were examined and bright field images were obtained using a Leica DMR upright microscope equipped with a colour CCD camera (Micropublisher

5.0 RTV, QImaging) and Openlab software (Improvision). Two consecutive sections at each level of lung tissue were examined for human-vimentin-positive lesions and only those observed at the same area on two consecutive sections were counted. The area of the lesions was determined using ImageJ (v1.45s) and the average size of the two consecutive lesions was reported. MDA-MB-231 colonies with an area equal to or larger than 500 mm2 were considered as metastases. SUM-159PT metastatic foci of any size were counted.

2.2.7. Protein Chromatography

200 ml of ACM was concentrated to a final volume of 4 ml by using Vivaspin concentrators (10

MWCO; GE Healthcare) and spinning the media using a tabletop centrifuge (Beckman) at 3,000 rpm at 4 °C. The concentrated sample was loaded on a SuperdexTM 200 size exclusion column

(HiLoad 16/60, Pharmacia Biotech) and eluted with PBS without Ca+2/Mg+2 in ten 10 ml fractions at a flow rate of 0.5 ml/min using a fast protein liquid chromatography system

(Amersham). One ml aliquot of each fraction was concentrated using Vivaspin concentrators, resuspended in DMEM, filter-sterilized and tested for activity using the single cell motility assay. Subsequently, eight ml of the active fraction (S6) was concentrated to 500 ml with a

Vivaspin concentrator and then resuspend in 10 ml of buffer solution (40 mM NaCl, 15 mM

Na3PO4 buffer, pH 7.3). The sample was loaded onto a 1 ml Mono Q cation exchange column

(HiTrapTM Q XL, GE Healthcare) and eluted with an increasing concentration of sodium chloride buffer (1.5 M NaCl, 15 mM Na3PO4 buffer, pH 7.3) in four 1 ml fractions. 0.8 ml of each fraction was washed several times with DMEM using Vivaspin concentrators and tested for activity using the single cell motility assay. An aliquot of each fraction was also resolved by

SDS-PAGE (12 %) followed by staining with GelCode® Blue (24590, Thermo Scientific) or silver nitrate (Chevallet et al., 2006).

2.2.8. Mass Spectrometry

Gel bands containing proteins resolved by SDS-PAGE and stained with GelCode® Blue (24590,

Thermo Scientific) were cut and minced in eppendorf tubes using clean blades. Then, the samples were equilibrated for 10 min and washed twice on ice with 100 mM NH4HCO3 (40867,

Sigma-Aldrich) buffer, pH 8.0. Following dehydration with 95 % ethanol on ice for 20 min, the samples were reduced with 5 mM dithiothreitol (11583786001, Roche) at 60 °C for 45 min in

NH4HCO3 buffer. The samples were dehydrated again and then alkylated with 50 mM iodoacetamide (I1149, Sigma-Aldrich) for 45 min at room temperature in the dark. Following another round of dehydration, the samples were digested with trypsin (V5111, Promega) at 37

°C overnight. The reaction was stopped by adding 5 % formic acid and the resulting peptides were separated by liquid chromatography (Agilent 1100 series) and identified by tandem mass

spectrometry using LTQ-XL or LTQ-Orbitrap (Thermo Scientific). The data was analysed using ProHits (Liu et al., 2010).

2.2.9. Exosome Isolation By Differential Ultracentrifugation

Exosomes were purified by sequential ultracentrifugation of the conditioned media at 2,000 x g for 30 min (to remove any floating cells), 10,000 x g for 40 min (to remove any cellular debris) and 100,000 x g for 2-14 hrs (to isolate exosomes) using polyallomer tubes (Beckman), a SW 28 swinging bucket (Beckman) and a refrigerated ultracentrifuge. The 100,000 x g pellet was washed once with PBS and spun down at 100,000 x g for 2 hrs to obtain purified exosomes.

2.2.10. Electron Microscopy

Exosomes purified as described above were fixed overnight at 4 °C in 2 % paraformaldehyde in

PBS, whole mounted on Formvar-carbon coated grids, fixed for 5 min with 1 % glutaraldehyde in PBS, stained for 5 min with uranyl-oxalate solution, pH 7, and then embedded for 10 min on ice in methyl cellulose. Exosomes were labelled with biotinylated hamster anti-Cd81 (NBP1-

28138, Novus Biologicals), biotinylated goat anti-Igsf8 (BAF3117, R&D) or biotinylated sheep anti-Ptgfrn (BAF4495, R&D) antibodies for 1 hr at RT and then with Streptavidin-Gold (S9059,

Sigma-Aldrich) for 30 min at RT. Biotinylated hamster IgG isotype control (NBP120-18472,

Novus Biologicals) was used as control. Images were captured using a transmission electron microscope (TecnaiTM 20, FEI) at 200 kV.

2.2.11. Exosome Purification by Sucrose Gradient Ultracentrifugation

For further purification, exosomes isolated by differential ultracentrifugation (as described above) were floated in a sucrose gradient, prepared as previously described (Thery et al., 2006), at 210,000 x g for 20 hrs at 4 °C. One ml fractions were collected from the top of the tube and the density was measured with a refractometer. Two fractions with a density ranging from 1.15-

1.19 g/ml were pooled, diluted in 20 mM Hepes, pH 7.4, and centrifuged at 100,000 x g for 2 hrs at 4 °C. The exosome pellet was subjected to MS as described above.

2.2.12. Methyl-β-Cyclodextrin (MβCD) Treatment

MβCD (C4555, Sigma-Aldrich) was dissolved in ACM by manually and gently mixing the media. The solution was incubated at 37 °C for 4 hrs and then rinsed several times with DMEM using Vivaspin concentrators. Subsequently, the media was filter-sterilized and tested for activity using the single cell motility assay or ultracentrifuged in order to isolate exosomes as described above.

2.2.13. Exosome Immuno-Magnetic Extraction

ACM was incubated with biotinylated Cd81, Igsf8 or Ptgfrn antibodies and IgG isotype control

(described above) for 20 hrs at 4 °C rotating at 1 rpm using a Sample Mixer (Invitrogen). The biotinylated complexes were captured with Streptavidin T1 beads (656-01, Invitrogen) after incubation for 2 hrs at 4 °C rotating at 1 rpm. The exosome-depleted media was filter-sterilized and tested for activity using the single cell motility assay. The immunocomplexes were processed for MS as described above. 77

2.2.14. Immunoblotting

Protein samples were subjected to SDS-PAGE followed by transfer onto nitrocellulose membrane (162-0115, BioRad), probing with primary antibodies (as indicated below) and HRP- linked secondary antibodies (Santa Cruz), and detection with a SuperSignal chemiluminescence reagent (34095, Thermo Scientific). The antibodies were used as follows: mouse anti-Cd81 (sc-

166029, Santa Cruz, 1:1000), rabbit anti-Igsf8 (ab74890, Abcam, 1:1000), rabbit anti-Ptgfrn

(ab97567, Abcam, 1:1000), mouse anti-Flotillin1 (610820, BD, 1:1000), mouse anti-αTubulin

(T6199, Sigma-Aldrich, 1:10000) and mouse anti-βActin (A1978, Sigma-Aldrich, 1:10000).

2.2.15. RNA Extraction and Real-Time Quantitative PCR (RT-qPCR)

Total RNA was isolated from cell lysate using the RNeasy mini kit (Qiagen). Any genomic

DNA in the RNA preparation was digested with RNase-free DNase (Qiagen) following manufacturer’s instructions. Total RNA was then reverse transcribed with SuperScriptIII reverse transcriptase (18080-044, Invitrogen) using oligo(dT)18 primers (Thermo Scientific) following manufacturers’ instructions. qPCR was performed using the SYBR® Green PCR

Master Mix (Applied Biosystems) and an ABI 7500 apparatus. The primer sets are listed below.

Gene expression was normalized to the expression of endogenous GAPDH or HPRT. Relative

-ΔΔC changes in gene expression were quantified using the 2 T method (Livak and Schmittgen,

2001).

2.2.16. PCR Primer Sets

Gene Species RefSeq ID Primer Sequence 5'-3' Amplicon Size (bp)

Cd63 Mouse NM_001042580 TTTGCTCTACGTTCTCCTGC 81

AAGACAACCTGAACCGCTAC

Cd81 Mouse NM_133655 ATGTTTGTAGGCTTCCTGGG 143

GCGATCTGGTCTTTGTTTACG

Cd82 Mouse NM_007656 TGTATCGGTGCTGTCAATGAG 142

CTGTGTTCCCCATCTCCTTC

Igsf8 Mouse NM_080419 GATACCCTATTTGTGCCCCTG 148

AGACAGTCAACACCTGCAAG

Porcn Mouse NM_016913 GCTATCTTCCACCTGTCCTAC 148

TCCAGCATCCAAAAGTGACC

Ptgfrn Mouse NM_011197 TCCCACATTTAACGCCTCTG 148

AGAGTGTACCGCAAACCAG

Hprt Mouse NM_013556 CACAGGACTAGAACACCTGC 250

GCTGGTGAAAAGGACCTCT

2.2.17. Cell Transfection

L cells were transfected with 40 nM siGENOME SMARTpool siRNAs (Dharmacon) using

LipofectamineTM RNAiMAX (Invitrogen) transfection reagent according to manufacturers’ instructions. siCONTROL Non-Targeting siRNA Pool#2 (D-001206-14; Dharmacon) was used as control siRNA. L cells were transfected with cDNA or shRNA expression vectors using

LipofectamineTM LTX (Invitrogen) according to manufacturer’s instructions. shRNAs for

mouse Cd81 (RHS4430-98485409 and RHS4430-101132069: Thermo Scientific) were purchased from Open Biosystems and the sequences are listed below. TRC Lentiviral eGFP shRNA (RHS4459; Thermo Scientific) was used as control shRNA. L cells were selected with

10 mg/ml puromycin.

2.2.18. shRNA Sequences

Gene Item Cat. # Clone ID Hairpin Sequence

Cd81 RHS4430- V2LHS_148 TGCTGTTGACAGTGAGCGCACACGTCGCCTTCAACT 98485409 88 GTAATAGTGAAGCCACAGATGTATTACAGTTGAAG GCGACGTGTTTGCCTACTGCCTCGGA

Cd81 RHS4430- V3LHS_304 TGCTGTTGACAGTGAGCGACAAGGATGTGAAGCAG 101132069 176 TTCTATAGTGAAGCCACAGATGTATAGAACTGCTTC ACATCCTTGGTGCCTACTGCCTCGGA

2.2.19. Cell Lysis

Cells were lysed for 20 min at 4 °C in TNTE (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM

EDTA, 0.5 % Triton X-100) buffer supplemented with 25 mM NaF, 10 mM NaPPi, 1 mM

PMSF, 1 mM Na3VO4, and a protease inhibitor cocktail (consisting of 4 mg soybean trypsin inhibitor, 4 mg benzamidine HCl, 2 mg aprotinin, 400 mg antipain, 400 mg leupeptin, and 400 mg pepstatin). The lysates were centrifuged at maximum speed for 20 min at 4 °C to pellet the nuclei and cell debris. The supernatant was aliquoted to new tubes and stored at -20 °C until further use. For immunoblotting, proteins were reduced by adding Laemmli sample buffer (50 mM Tris-HCl, 5 % glycerol, 1 % SDS, 50 mM DTT and 0.05 % bromophenol blue).

2.2.20. Analysis of Human Stroma Gene Expression Data

The gene expression profiles from normal stroma from breast reduction specimens (n = 6) or invasive ductal carcinoma (n = 26) were downloaded from the NCBI GEO DataSets database

(GSE4823). The expression values for CD81, IGSF8, PTGFRN, CD82 and CD63 genes were obtained and analysed.

2.3. Results

2.3.1. L cell-secreted Factors Stimulate Breast Cancer Cell Protrusive Activity and Motility

In this project, I decided to study the ability of cancer cells to move as single cells as this may be a critical attribute of metastatic cells (Friedl and Alexander, 2011). While developing a motility assay, I found that media conditioned by mouse fibroblast L cells, henceforth referred to as active conditioned medium (ACM) (Figure 2.1A), had a strong stimulatory effect on the protrusive activity and motility of human breast adenocarcinoma MDA-MB-231 cells (Figure

2.1B-E). Under basal conditions, MDA-MB-231 cells exhibited overall a mesenchymal phenotype, characterized by an elongated and spindle-shape morphology with a leading front and a trailing edge (Figure 2.1B). Shortly after ACM treatment, MDA-MB-231 cells formed long cell protrusions that extended in different directions in a highly dynamic fashion (Figure

2.1B). Tracking of the cells in time (Figure 2.1C), revealed that ACM-treated cells moved at a significantly faster rate (Figure 2.1D) than control cells. Next, I grew MDA-MB-231 cells in the presence of ACM or control media (DMEM) for three weeks (Figure 2.1E). I found that when the cells were treated with ACM for three weeks followed by a wash out, their motility returned to the basal rate that could be re-stimulated with ACM. This indicates that the effect of 81

ACM in MDA-MB-231 cells is reversible and these cells are continuously dependent on L cell signals for induced motility. In addition, I tested other mammary cell lines of epithelial origin, including human carcinoma SUM-159PT, MDA-MB-468, and T-47TD cells, and mouse carcinoma EMT-6 cells, tumourigenic EpRas cells and immortalized EpH4 cells and in all cases

ACM induced cell protrusive structures and motility within a few hours of treatment (Figure 2.2

A-B). Altogether, these results demonstrate that L cell-secreted factor(s) have a strong and prevailing effect in inducing mammary cell protrusive activity and motility.

A B C L cell

3 days DMEM DMEM

ACM

18 hours

BCC ACM ACM

D E ns 10 *** 15 ns 8 6 10 4 5 2 Cell Speed (nm/s) 0 Cell Speed (nm/s) DMEM ACM 0 DMEM ACM DMEM ACM DMEM (3 wks) ACM (3 wks)

Figure 2.1 L cell-secreted Factors Stimulate BCC Protrusive Activity and Motility

(A) L cells, incubated with DMEM for 3 days, produce active conditioned media (ACM), which was used to treat BCCs. The protrusive activity and motility of individual BCCs was imaged by phase-contrast time-lapse microscopy for 18 hours. (B) Phase contrast images of MDA-MB- 231 cells treated with control media (DMEM) or ACM are shown. Short cell protrusions (white hollow arrow) and long cell protrusions (white arrow) are indicated. (C) Representative MDA- MB-231 cell migration tracks (light blue). The cell outline is shown at the beginning of each track and the arrow tip indicates the direction of movement. (D) The average speed of each cell was determined and results are plotted as a boxplot, where the box spans the 25th and 75th percentile, the line marks the median speed and the whiskers the minimum and maximum data

points. n = 30 cells per group. *** p < 0.0001. (E) After 3 weeks (wks) of treatment with DMEM or ACM, the response of MDA-MB-231 cells to fresh DMEM or ACM was determined using the motility assay as in (D). n = 33 cells per group. ns, not significant.

A SUM-159PT MDA-MB-468 T-47D EMT-6 EpRas EpH4 DMEM ACM

B 10 *** *** *** *** *** ***

4 Cell Speed (nm/s)

ACM ACM ACM ACM ACM ACM DMEM DMEM DMEM DMEM DMEM DMEM SUM-159PT MDA-MB-468 T-47D EMT-6 EpRas EpH4

Figure 2.2 ACM Promotes Protrusive Activity and Motility in Several BCCs

(A) Phase contrast images of SUM-159PT, MDA-MB-468, T-47D, EMT-6, EpRas and EpH4 cells, treated with control media (DMEM) or ACM, as described in Figure 2.1A. Note that cells incubated with DMEM formed short membrane protrusions (white hollow arrow), while ACM- treated cells extended highly dynamic and long cell protrusions (white arrow). (B) Cells treated with DMEM or ACM as described in (A) were traced and the average speed was determined. Results are plotted as box and whisker plots. n = 28 cells per group. *** p < 0.0001.

2.3.2. L cell-secreted Factors Stimulate Breast Cancer Cell Invasiveness and Metastatic Potential

Increased cell protrusive activity and motility generally correlate with increased invasiveness and metastatic potential (Lee et al., 2007). As a result, I examined MDA-MB-231 cells grown in a three dimensional Matrigel matrix and found that ACM stimulated the formation of MDA-

MB-231 cell protrusive structures (Figure 2.3A). Next, I aimed to test the effect of L cells- secreted factors in breast cancer cell metastasis. MDA-MB-231 cells are weakly metastatic

(Minn et al., 2005) and thus represented a good model system for investigating L cell-stimulated

BCC metastasis. Hence, MDA-MB-231 cells were injected either alone or together with L cells in the fourth mammary fat pad of severe combined immune deficient (SCID) mice, which are homozygous for the SCID mutation that affects both B and T lymphocyte differentiation, and thus have no reactivity against xenografts or allografts, but have normal natural killer cell, macrophage and granulocyte function (Dorshkind et al., 1984) (Figure 2.2B). Four weeks post implantation, the mice were sacrificed and their primary tumours and lungs were obtained for analysis. Tumours in mice coinjected with MDA-MB-231 cells and L cells were larger (Figure

2.3C) than tumours in mice injected with the same number of either MDA-MB-231 cells alone, or L cells alone, which are tumourigenic (Sanford et al., 1956). Next, the lungs of the tumour- bearing mice were stained for MDA-MB-231 metastatic lesions using an antibody specific for human vimentin (Figure 2.3D). Examination of the lungs revealed that MDA-MB-231 cells, when injected alone, formed few metastatic colonies consisting of just a few cells (Figure 2.2D).

In comparison, when MDA-MB-231 cells were coinjected with L cells, both the incidence and the size of MDA-MB-231 metastatic colonies were significantly increased (Figure 2.2E-F). To further investigate the ability of L cells to induce BCC metastasis, I decided to use the human carcinoma SUM-159PT cells as a second model system. SUM159-PT cells have a similar

migratory propensity as MDA-MB-231 cells and respond well to ACM (Figure 2.2B).

Moreover, like MDA-MB-231 cells, SUM-159PT cells are ER-, PR-, HER2-negative and have a similar proportion of stem-cell like cells in their population (Fillmore and Kuperwasser, 2008).

However, unlike MDA-MB-231 cells, which were derived from a pleural effusion, SUM-159PT cells were derived from an invasive ductal carcinoma and are not metastatic in orthotopic mouse models (Ma et al., 2007). Accordingly, orthotopic implantation of SUM-159PT cells alone produced almost no metastases at four weeks post injection. In contrast, orthotopic coimplantation of SUM-159PT cells with L cells promoted the appearance of SUM-159PT nests of cells in the lung parenchyma (Figure 3.4D-E). Importantly, results from subsequent experiments revealed that the increased frequency and size of metastatic lesions is not due to enhanced growth of the primary tumour in the coinjected groups (Figure 2.8A-F and Figure

3.4B-D). Altogether, these results reveal that L cell-secreted factors stimulate BCC metastasis in mouse models.

A DMEM ACM B MDA-MB-231

MDA-MB-231 + L cell 4 wks mfp

L cell

C 6 p = 0.0003 ) 3

2 Tumour Volume ( cm Volume Tumour 0 MDA-MB-231 MDA-MB-231 L cell + L cell

D MDA-MB-231 MDA-MB-231 + L cell E 6 p = 0.03

2 # Colonies / Lung

0 MDA-MB-231 MDA-MB-231 + L cell

F 5 p = 0.0004

4 Colony Area Colony

10 3 Log 2 MDA-MB-231 MDA-MB-231 + L cell

Figure 2.3 L cell-secreted Factors Promote MDA-MB-231 Invasiveness and Metastasis

(A) Phase-contrast images of MDA-MB-231 cells grown for three days in Matrigel in either DMEM or ACM. Higher power views of boxed areas are shown in lower panels. Scale bar = 90 µm. (B) Experimental schematic. MDA-MB-231 cells were injected alone or together with L cells in the fourth mammary fat pad (mpf) of ten-week-old female SCID mice. Four weeks (wks) post injection, the mice were sacrificed and their tumours and lungs were collected and analysed. (C) The volume of the primary tumours of mice injected with MDA-MB-231 only (n

= 4), MDA-MB-231 plus L cells (n = 7) or L cells only (n = 3) was determined and plotted as mean volume ± SD. p = 0.0003 using two tailed unpaired t-test with Welch’s correction. (D) The lungs of mice, injected with the indicated cells, were stained for human vimentin and representative micrometastases corresponding to the median size of the population are shown. Higher power views of boxed areas are shown in lower panels. Scale bar = 35 µm. (E) The number of metastatic colonies in the lungs of mice injected with MDA-MB-231 only (n = 4) or MDA-MB-231 plus L cells (n = 7) were quantified and plotted as mean colony number ± SEM. p = 0.03 using two tailed unpaired t-test with Welch’s correction. (F) The surface area of the metastatic colonies in (E) was determined and plotted as the mean area ± SEM. p = 0.0004 using two tailed unpaired t-test with Welch’s correction.

2.3.3. Identification and Characterization of L cell-secreted Exosomes

The above results suggested that L cells secrete a factor(s) that strongly stimulates BCC protrusive activity, motility and metastasis. To purify the factor(s), I fractionated ACM in twelve different fractions by size exclusion chromatography (SEC) (Figure 2.4A). Analysis of each fraction, revealed that the active factor(s) was concentrated in a single fraction (S6) (Figure

2.4B), which was considerably purified from total ACM and was rich in high molecular weight proteins (Figure 2.4C). Next, I subjected fraction S6 to ion exchange chromatography (IEC)

(Figure 2.4D), which allowed further purification of the active factor(s) in another single fraction (Q3) (Figure 2.4E) that was rich in high molecular weight proteins (Figure 2.4F). Then,

I resolved the proteins in fraction Q3 by gel electrophoresis, collected gel slices and then reduced, alkylated and digested the peptides in situ. Afterward, Dr. Liang Zhang subjected the prepared samples to mass spectrometry (MS) analysis. To our surprise, we found that the samples were enriched in membrane proteins instead of extracellular components (Table 2.1).

Using David bioinformatics tool (Huang da et al., 2009) and Exocarta (Mathivanan et al., 2012),

I found that several membrane proteins enriched in fraction Q3, including tetraspanin Cd81 and its partners Igsf8 and Ptgfrn, were previously reported to associate with exosomes (Table 2.1).

Altogether, these results indicate that L cells secrete exosomes that may mediate the activity of

ACM.

A B C 1xACM -- 15 MW(kDa) S3 S4 S5 S6 S7 S8 S9 S10 S11 S12 Input 170 130 95 50xACM 10 72 55 43 5 34

Superdex 200 Cell Speed (nm/s) Column 26 0 S5 S6 S7 S8 S9 S10 S11 S12 Input 17 Fractions S1-12

DE F Fractions 15 Input FT(S6)Q1 Q2 Q3 Q4 S6 MW(kDa) 170 130

+ +++ +++ 10 95 + +++ +++ 72 [NaCl] + +++ Mono Q +++ + +++ 55 ++ + Column + +++ +++ + +++ 5 43 + +++ + + + + ++ + ++ Cell Speed (nm/s) + + 34 +

0 26 Fractions FT Q1 Q2 Q3 Q4 Q1-4 Input

Figure 2.4 Fractionation of ACM by Protein Chromatography

(A) Schematic of the size exclusion chromatography experiment. 200ml of ACM was concentrated 50x with a Vivaspin sample concentrator. 4 ml of concentrated sample was loaded on a Superdex 200 gel filtration column and eluted in twelve fractions (S1-12) using phosphate buffer saline (PBS). (B) Fractions S5-12 and total ACM (input), described in (A), were diluted with DMEM and concentrated to the original volume using Vivaspin columns. Their activity was tested using the single cell motility assay. Cell speed was quantified and the results are plotted as boxplots. (C) Fractions S3-S12 and input, described in (A-B), were resolved by SDS- PAGE and stained with GelCode blue. Note that fraction S6 was eluted soon after the void volume (collected in fractions S1-S4) and contained high molecular weight (MW) components that were considerably purified from total media. (D) Schematic of the ion exchange chromatography experiment. Fraction S6 was applied to an anion exchange column and eluted in four fractions (Q1-Q4) with a sodium chloride gradient. (E) After rinsing with DMEM as in

(B), the activity of fractions Q1-Q4, flowthrough (FT) and input, described in (D), was tested using the single cell motility assay. Cell speed was quantified and the results are plotted as boxplots. (F) Fractions Q1-Q4, FT and input, described in (D-E), were resolved by SDS-PAGE and stained with silver nitrate.

Table 2-1 Proteins in Fraction Q3 Identified by Mass Spectrometry

The MS data table was obtained from ProHits. Genes are arranged in a descending order according to their Mascot Score. Highlighted in blue are genes clustered by David annotation tool as vesicle proteins and in red are some of the genes that I manually annotated according to Exocarta.

Gene Gene Mascot Unique Total % Hit Hit MW ID Symbol Score Peptide # Peptide # Coverage (kDa) 60595 Actn4 3001 46 57 57.5 105 15530 Hspg2 2847 46 47 12.7 469 16776 Lama5 2695 42 49 15 404 226519 Lamc1 2561 38 44 29.9 177 121021 Cspg4 2494 40 48 21.2 252 11603 Agrn 2168 32 34 21.7 198 15519 Hsp90aa1 2145 33 50 37.5 85 109711 Actn1 1982 16 39 44.3 103 18220 Nucb1 1860 32 46 60.8 53 22027 Hsp90b1 1858 32 34 38 92 14104 Fasn 1831 32 32 15.4 272 13602 Sparcl1 1771 27 44 41.7 72 14268 Fn1 1762 29 34 17.4 272 16777 Lamb1 1762 31 32 17.9 202 15481 Hspa8 1562 23 28 38.4 71 78908 Igsf3 1550 26 29 26.1 135 11568 Aebp1 1548 22 29 25.6 128 14828 Hspa5 1378 19 23 38.5 72 19221 Ptgfrn 1317 22 27 26.4 99 19039 Lgals3bp 1262 17 27 40.4 64 18822 Plod1 1211 17 19 28.7 84 14733 Gpc1 1160 18 29 36.4 61 235505 Cd109 1131 18 18 15 162 14115 Fbln2 929 15 19 15.7 126 11804 Aplp2 867 14 16 19.3 85 21826 Thbs2 783 12 12 12.9 130 18571 Pdcd6ip 770 12 12 15.8 97 17304 Mfge8 762 12 16 30.5 47 22154 Tubb5 712 11 14 24.3 50 56213 Htra1 688 12 13 28.8 51 12833 Col6a1 678 11 12 12.2 108 19173 Psmb5 664 10 10 38.3 29 104112 Acly 652 12 12 12.6 120 11820 App 648 10 12 17.4 78 18746 Pkm2 628 12 12 30.1 58 67300 Cltc 617 10 10 7.1 191 19170 Psmb1 607 8 10 40.4 26 15525 Hspa4 601 10 10 14.6 94 26444 Psma7 593 8 9 41.1 28 26442 Psma5 581 8 10 40.7 26 227613 Tubb4b 572 2 12 21.6 50

Table 2-1 continued

Gene Gene Mascot Unique Total % Hit Hit MW ID Symbol Score Peptide # Peptide # Coverage (kDa) 26443 Psma6 565 9 11 45.1 27 19166 Psma2 560 8 14 39.7 26 20174 Ruvbl2 557 8 8 23.1 51 12469 Cct8 547 8 8 16.8 60 26440 Psma1 547 9 12 33.5 30 26445 Psmb2 545 10 13 39.8 23 56505 Ruvbl1 539 8 8 23.2 50 433182 Gm5506 519 7 8 24 47 12466 Cct6a 517 9 10 21.3 58 14694 Gnb2l1 506 7 8 26.2 35 11461 Actb 505 8 9 31.7 42 19156 Psap 503 10 10 23.2 61 21810 Tgfbi 496 8 9 12.6 75 19167 Psma3 494 8 10 31.4 28 20256 Clec11a 484 8 9 28.7 36 100046120 LOC100046120 456 8 9 15.4 52 12977 Csf1 447 7 11 12 61 13437 Dnpep 445 8 8 21.8 52 20740 Spna2 442 8 8 3.6 285 11747 Anxa5 439 7 7 25.1 36 16728 L1cam 416 8 8 7.5 141 634386 Gm10116 414 6 6 47 21 22143 Tuba1b 414 6 7 19.1 50 50909 C1ra 410 6 7 10.7 80 236539 Phgdh 407 5 5 11.3 57 18830 Pltp 404 6 7 15.2 54 18007 Neo1 396 8 9 6.2 163 18148 Npm1 394 7 7 20.2 33 94242 Tinagl1 384 7 7 16.7 53 16211 Kpnb1 380 6 7 8.1 97 21991 Tpi1 376 5 5 21.1 32 19089 Prkcsh 375 6 6 10.6 59 21454 Tcp1 368 6 6 12.1 60 319178 Hist1h2bb 361 6 10 46.8 14 432551 Gm12141 355 6 6 11.3 61 18596 Pdgfrb 355 5 6 5.6 123 26446 Psmb3 353 4 5 28.3 23 16971 Lrp1 339 6 6 1.6 504 19172 Psmb4 337 5 6 22 29 12462 Cct3 317 5 5 10.5 61 57875 Angptl4 311 5 5 19.5 46

Table 2-1 continued

Gene Gene Mascot Unique Total % Hit Hit MW ID Symbol Score Peptide # Peptide # Coverage (kDa) 26441 Psma4 309 4 5 18.4 29 12461 Cct2 308 5 5 14.6 57 13629 Eef2 307 5 5 5.7 95 12306 Anxa2 306 5 5 18.3 39 223691 Eif3l 293 5 5 10.1 67 11803 Aplp1 291 5 6 9.9 73 18542 Pcolce 289 5 5 15.6 50 27050 Rps3 287 5 5 25.9 27 26433 Plod3 283 4 5 7.7 85 22321 Vars 282 4 4 3.9 140 12520 Cd81 275 3 4 27.5 26 11848 Rhoa 273 4 4 27.5 22 21894 Tln1 273 5 5 2.4 270 268373 Ppia 271 5 6 32.9 18 140559 Igsf8 270 4 4 9.7 65 319163 Hist1h2aa 268 3 6 30.2 14 50908 C1s 265 3 3 6.1 77 56752 Aldh9a1 264 5 5 10.8 56 70445 Cd248 264 5 5 7.1 82 217333 Trim47 263 4 4 11 61 27374 Prmt5 260 4 4 6.4 73 53378 Sdcbp 250 4 4 14.1 32 13030 Ctsb 248 4 4 12.7 37 17254 Slc3a2 247 5 5 10.1 62 18477 Prdx1 242 4 5 21.6 22 68794 Flnc 239 4 4 1.8 291 15267 Hist2h2aa1 236 1 6 30 14 17975 Ncl 233 5 5 7.5 77 19177 Psmb7 233 4 5 19.9 30 105348 Golm1 231 4 4 13.7 44 66656 Eef1d 230 3 3 13.2 31 14433 Gapdh 227 3 3 12.9 36 13627 Eef1a1 224 4 4 12.6 50 217119 Xylt2 221 4 4 5 97 56041 Uso1 219 4 4 5.5 107 22042 Tfrc 215 4 4 5.6 86 22631 Ywhaz 214 3 3 15.5 28 14319 Fth1 213 4 5 20.3 21 100046995 LOC100046995 212 4 4 10.1 64 11837 Rplp0 212 4 4 14.2 34 12822 Col18a1 209 3 3 3.9 134

Table 2-1 continued

Gene Gene Mascot Unique Total % Hit Hit MW ID Symbol Score Peptide # Peptide # Coverage (kDa) 20312 Cx3cl1 207 4 4 10.1 42 11746 Anxa4 202 3 3 11.3 36 216558 Ugp2 202 4 4 8.5 57 14678 Gnai2 199 4 4 14.6 40 22186 Uba52 198 4 4 36.7 15 107895 Mgat5 197 4 4 5.8 85 215449 Rap1b 197 4 4 28.8 21 19353 Rac1 196 4 4 17.2 21 17174 Masp1 195 3 3 5.3 80 21814 Tgfbr3 190 4 4 5.8 94 433283 Gm5525 189 3 3 9 48 16956 Lpl 184 2 2 6.1 53 16480 Jup 179 3 3 4.7 82 108075 Ltbp4 178 3 3 2.5 179 18538 Pcna 177 3 3 13.8 29 114228 Prss1 174 2 9 12.2 26 19175 Psmb6 172 3 3 15.5 25 27979 Eif3b 171 3 3 4.1 91 239673 4732456N10Rik 170 1 3 5.9 58 100044179 LOC100044179 170 1 3 17.2 23 14376 Ganab 167 4 4 4.7 109 97122 Hist2h4 166 3 3 31.1 11 16828 Ldha 160 3 3 9.9 36 14688 Gnb1 159 3 3 9.4 37 26364 Cd97 157 2 2 3.5 90 14679 Gnai3 155 1 3 10.7 41 27053 Asns 153 3 3 6.1 64 22352 Vim 153 2 4 6.7 54 66085 Eif3f 152 2 2 7.8 38 22235 Ugdh 152 3 3 6.7 55 22123 Psmd3 150 2 2 5.3 61 19384 Ran 149 3 3 15.7 24 12465 Cct5 148 2 2 4.8 60 56347 Eif3c 146 2 2 2.6 105 108679 Cops8 143 2 2 12.9 23 675857 LOC675857 142 3 3 4.4 89 12934 Dpysl2 141 2 2 4.5 62 20971 Sdc4 141 2 2 15.2 21 13681 Eif4a1 140 3 3 8.1 46 12282 Hyou1 137 3 3 2.9 111 19349 Rab7 134 2 2 12.1 23

Table 2-1 continued

Gene Gene Mascot Unique Total % Hit Hit MW ID Symbol Score Peptide # Peptide # Coverage (kDa) 226414 Dars 133 3 3 5.6 57 233726 Ipo7 132 2 2 2.2 119 78303 Hist3h2ba 131 2 2 19 14 100044728 LOC100044728 130 2 2 11.8 14 670464 Gm11596 126 1 4 14.6 21 17463 Psmd7 125 2 2 6.5 37 12842 Col1a1 123 3 3 2.1 138 19073 Srgn 123 2 3 19.7 17 17886 Myh9 121 2 2 1.3 226 100039355 Rps16-ps2 121 2 2 15.1 16 16913 Psmb8 120 1 2 9.1 30 11973 Atp6v1e1 117 2 2 11.9 26 320078 Olfml2b 117 2 2 3.2 83 11600 Angpt1 116 2 2 4.8 57 12512 Cd63 113 2 3 8.8 26 53605 Nap1l1 113 2 2 7.7 45 53381 Prdx4 111 2 2 7.7 31 14569 Gdi2 108 2 2 5.2 51 77590 Chst15 107 2 2 3.9 65 353172 Gars 107 2 2 3 82 16341 Eif3e 106 2 2 4.3 52 22144 Tuba3a 106 2 2 6.2 50 11657 Alb 105 1 1 2.1 69 12540 Cdc42 105 2 2 11 21 50918 Myadm 105 1 1 5.3 35 16541 Napsa 105 2 2 5.7 46 69748 Aldh16a1 104 1 1 1.7 85 20103 Rps5 104 1 1 7.4 23 13830 Stom 104 2 2 7 31 70247 Psmd1 103 2 2 2.4 106 100047521 LOC100047521 102 1 1 7.7 19 12831 Col5a1 101 2 3 1.5 184 19325 Rab10 101 2 2 11 23 384009 Glipr2 100 1 1 8.4 17 12215 Bsg 99 2 2 6.4 42 234734 Aars 98 2 2 2.3 107 12631 Cfl1 98 2 2 13.3 19 12464 Cct4 97 2 2 3.7 58 73830 Eif3k 97 2 2 6.9 25 21762 Psmd2 97 1 2 1.3 100 12468 Cct7 96 2 2 4 60

Table 2-1 continued

Gene Gene Mascot Unique Total % Hit Hit MW ID Symbol Score Peptide # Peptide # Coverage (kDa) 67025 Rpl11 94 2 2 12.9 20 20226 Sars 94 2 2 4.5 61 20102 Rps4x 90 1 1 4.9 30 20501 Slc16a1 87 1 1 2.6 53 21927 Tnfaip1 87 1 1 3.8 36 26894 Cops7a 86 2 2 7.6 30 670496 Gm11564 86 1 3 10 21 19646 Rbbp4 86 2 2 4 48 67373 2210010C04Rik 85 1 1 4.9 26 13628 Eef1a2 85 1 1 2.4 50 14086 Fscn1 85 2 2 3.9 54 18176 Nras 85 1 1 6.3 21 13669 Eif3a 83 1 1 1 162 11475 Acta2 82 2 2 5.6 42 18103 Nme2 82 2 2 13.8 17 18655 Pgk1 82 1 1 4.3 45 14466 Gba 80 1 1 2.5 58 54624 Paf1 80 1 1 2.6 60 27207 Rps11 80 1 1 7 18 14264 Fmod 79 2 2 5.1 43 246154 Vasn 79 1 1 1.5 72 12338 Capn6 78 2 2 2.7 75 70495 Atp6ap2 77 1 1 3.7 39 54130 Actr1a 76 1 1 4.3 43 51792 Ppp2r1a 76 1 1 1.7 65 20514 Slc1a5 74 1 1 2 58 63959 Slc29a1 74 1 1 2.6 50 26893 Cops6 73 1 1 4 36 12847 Copa 72 1 1 1 138 11966 Atp6v1b2 71 1 1 2.9 57 68981 Snrpa1 71 1 1 5.5 28 13660 Ehd1 70 1 1 2.2 61 18570 Pdcd6 69 1 1 6.8 22 66194 Pycrl 69 1 1 4 29 14863 Gstm2 68 1 1 4.1 26 16004 Igf2r 68 1 1 0.4 274 432548 Rpsa-ps4 68 1 1 4.4 33 12848 Cops2 67 1 1 2.9 52 215384 Fcgbp 67 1 1 0.4 275 22628 Ywhag 67 1 1 4 28 15270 H2afx 66 1 1 6.3 15

Table 2-1 continued

Gene Gene Mascot Unique Total % Hit Hit MW ID Symbol Score Peptide # Peptide # Coverage (kDa) 629242 Gm6958 65 1 1 4.4 33 192662 Arhgdia 63 1 1 7.8 23 68135 Eif3h 63 1 1 5.4 40 74488 Lrrc15 63 1 2 2.1 64 11964 Atp6v1a 62 1 1 1.9 68 56745 C1qtnf1 62 1 1 3.6 32 15078 H3f3a 62 1 2 8.1 15 56433 Vps29 62 1 1 5.5 20 16769 Dsg4 61 1 1 0.8 114 69077 Psmd11 61 1 1 3.1 47 75062 Sf3a3 61 1 1 2.4 59 12331 Cap1 59 1 1 2.3 52 20969 Sdc1 59 1 2 3.2 33 54161 Copg 58 1 1 1.4 97 66997 Psmd12 58 1 1 2.9 53 59029 Psmd14 58 1 1 4.2 35 20084 Rps18 58 1 1 7.2 18 17698 Msn 57 1 1 2.9 68 13138 Dag1 56 1 1 1.1 97 12521 Cd82 55 1 1 4.1 30 100044411 LOC100044411 55 1 1 2.3 45 55949 Eef1b2 54 1 1 5.8 25 100047061 LOC100047061 54 1 1 0.3 406 15469 Prmt1 54 1 1 3 42 22146 Tuba1c 54 1 1 3.3 50 75608 Chmp4b 53 1 1 4.9 25 101706 Numa1 53 1 1 0.5 235 20116 Rps8 53 1 1 6.3 24 100043806 Gm9816 52 1 1 2.3 55 20525 Slc2a1 52 1 1 2 54 107747 Aldh1l1 51 1 1 1 99 105722 Ano6 51 1 1 1.3 106 11826 Aqp1 51 1 1 3.3 29 15381 Hnrnpc 51 1 1 3.5 34 20750 Spp1 51 1 1 3.1 32 54709 Eif3i 50 1 1 3.7 36 50708 Hist1h1c 50 1 1 5.2 21 16783 Lamp1 49 1 1 3 44 101943 Sf3b3 49 1 1 1.2 135 94352 Loxl2 48 1 1 1.4 87 18187 Nrp2 48 1 1 1.1 105

Table 2-1 continued

Gene Gene Mascot Unique Total % Hit Hit MW ID Symbol Score Peptide # Peptide # Coverage (kDa) 20778 Scarb1 48 1 1 1.6 57 237858 Tusc5 48 1 1 6.9 19 71824 1700006A11Rik 47 1 1 1.8 60 12345 Capzb 47 1 1 3.7 31 404710 Iqgap3 47 1 1 0.7 185 23886 Gdf15 46 1 1 4 33 17448 Mdh2 46 1 1 3.3 36 18595 Pdgfra 46 1 1 0.8 123 19326 Rab11b 46 1 1 4.6 24 21366 Slc6a6 46 1 1 2.7 70 14635 Galk1 45 1 1 2.8 42 16403 Itga6 45 1 1 0.8 120 59287 Ncstn 44 1 1 1.1 78 230598 Nrd1 44 1 1 1 133 99543 Olfml3 44 1 1 2.7 46 11845 Arf6 43 1 1 5.7 20 68089 Arpc4 43 1 1 6.5 20 108664 Atp6v1h 43 1 1 1.9 56 26905 Eif2s3x 43 1 1 3 51 20005 Rpl9 43 1 1 5.2 22 71902 Cand1 42 1 1 1 136 27389 Dusp13 42 1 1 5.3 21 676974 LOC676974 42 1 1 5 38 67041 Oxct1 42 1 1 2.1 56 56012 Pgam2 42 1 1 4.3 29 65114 Vps35 42 1 1 1.8 92 230099 Car9 41 1 1 3 47 19045 Ppp1ca 41 1 1 3.3 38 623131 Prr19 41 1 1 2.2 40 321007 Serac1 41 1 1 1.9 71 41 1 1 0.3 339 69543 Capns2 40 1 1 4.5 27 56708 Clcf1 40 1 1 3.6 25 69654 Dctn2 40 1 1 2.2 44 57746 Piwil2 40 1 1 0.9 109 19038 Ppic 40 1 1 5.7 23 93765 Ube2n 40 1 1 7.2 17 70025 Acot7 39 1 1 2.9 42 211914 Asap2 39 1 1 0.8 106 56087 Dnahc10 39 1 1 0.2 529 26949 Vat1 39 1 1 2.7 43

100

Exosomes can be distinguished from other types of cell-derived microvesicles by several structural and molecular characteristics, including size, density and protein content (Simons and

Raposo, 2009; Thery et al., 2009). Hence, to confirm the presence of exosomes in ACM, I examined total ACM and the S6 and Q3 active fractions by electron microscopy (Figure 2.5A).

I observed cup-shaped vesicles with diameters ranging from 30nm-100nm that are characteristic of exosomes, as previously described (Simons and Raposo, 2009; Thery et al., 2009). Next, I investigated the identity of these microvesicles by immunostaining for the exosome markers,

Cd81, Igsf8 and Ptgfrn, and confirmed their association with the microvesicles in total ACM and the S6/Q3 fractions by EM (Figure 2.5B). Exosomes float at a density of 1.13-1.21 g/ml on a sucrose gradient during ultracentrifugation at 210,000 x g (Simons and Raposo, 2009; Thery et al., 2009). Thus, I subjected ACM microvesicles to ultracentrifugation in a sucrose gradient and analysed the proteomic content of the 1.13-1.21 g/ml density fractions by MS. Proteomic analysis of the samples and gene annotation using GeneCards (Stelzer et al., 2011) revealed that the exosome fraction was highly enriched in known exosome components, including tetraspanins (Cd63, Cd81 and Cd82) and endosomal proteins from the endoplasmic reticulum, the Golgi apparatus and the endocytic pathway (such as MVB proteins), which reflect the origin of exosomes from the endosomal membrane, thus further confirming their presence in ACM

(Figure 2.5C).

101

A Total ACM Fraction S6 Fraction Q3 B Cd81 Igsf8 Ptgfrn

100nm 100nm 100nm 100nm 100nm 100nm

C Secreted ECM remodeling Alb Fn1

Col23a1 Adam10 Clu Lyz1 Csn1s1Mif F10 Mfeg8

Bsg Channels/ Cyto- Ano6 Transporters Cell Igsf8 skeleton Adhesion Aqp1 Ptgfrn Ahnak Dmd Cell Actin Rac1 Signaling Atp1a1 Cd81 ER/Golgi Actrt2 RhoA Cacna1b Cd63 Arf5 Cfl1 Rras Calm1 Mvp Capn6 Tubb5 Gnai3 Rap1b Slc1a5 Cd82 Dennd5a Ergic3 Gnb1 Ywhae Gnb2 Ywhaz Slc2a1 Itgb1 Serinc5 Protein Fam129a Slc3a2 folding Gene Endocytosis Ubiquitin Slc7a6 Hsp90AB1 transcription / Exocytosis Ppia Kctd10 / translation Slc16a1 Trim47 Anxa2 Sdcbp Eef1a1 Maml3 Slc25a35 Anxa5 Sh3gl1 Tnfaip1 Uba52 H2afz Orc1l Ehd1 Ston2 Srrm2 Slc29a1 Ehd2 S100a6 Lysosome Flot1 Rab7 Unknown Ttyh3 Gapdh Rab10 Arl8b Function Hspa8 Lamp1 Lamp2 0610010O12Rik Litaf 1700027D21Rik Cd97 MVB Car9 Tsg101 Dip2b Kctd19 Ifitm3 Pdcd6IP EG432987Olfr570 pH Metabolism Fsip2 Rbm34 Scarb1 Chmp4b regulator Vps4a Akr1c12 Fth1 Gpr150 Veph1 Receptors Eno1 Ldha Fabp4 Ldhc Ftl1 Pkm2

Figure 2.5 The Active Fractions of ACM Contain Exosomes

(A) Images of exosomes in total ACM and the active fractions S6 and Q3 collected by electron microscopy. Cup-shaped microvesicles varying in size between 40 nm (arrowhead) to 100 nm (arrow) are indicated and enlarged in the bottom panel. (B) EM images of ACM exosomes immunolabelled with biotinylated anti-Cd81, anti-Igsf8 or anti-Ptgfrn antibodies and Streptavidin-Gold. (C) ACM exosomes purified on a sucrose gradient by ultracentrifugation were collected and subjected to proteomic analysis by MS. The identified proteins were 102

manually annotated using GeneCards and then were grouped by function. The exosome- associated proteins previously reported in the literature are highlighted in bold.

103

2.3.4. Functional Characterization of L cell-secreted Exosomes

Next, I aimed to determine whether L cell-secreted exosomes are the mediators of ACM activity. To establish whether L cell exosomes are sufficient to stimulate BCC protrusive activity and motility, I isolated exosomes from total ACM by differential ultracentrifugation and found that the 100,000 x g pellet, which contains the exosomes, possessed ACM stimulatory activity, while the supernatant was significantly depleted of activity (Figure 2.6A). Then, I used the cholesterol chelator methyl β-cyclodextrin (MβCD) to deplete ACM of cholesterol and thus disrupt exosome integrity (Figure 2.6B). I found that depletion of cholesterol resulted in disintegration of exosomes, detected by the loss of Cd81, and interfered with ACM-induced

MDA-MB-231 cell motility (Figure 2.6B). In addition, I used antibodies against the transmembrane proteins Cd81, Igsf8 or Ptgfrn to specifically immunocapture ACM exosomes. I subjected the resultant media to IB analysis and verified that Cd81, Igsf8 and Ptgfrn were effectively immunodepleted (Figure 2.6C). Also, I subjected the Cd81, Igsf8 or Ptgfrn immunoprecipitates to MS analysis and identified several exosome components, thus confirming that exosomes were successfully immunocaptured from ACM (Figure 2.6C). Then,

I tested the Cd81-, Igsf8- and Ptgfrn-immunodepleted media and found that the activity was significantly reduced when compared to control ACM (Figure 2.6C). Altogether, these results show that L cell-secreted exosomes are both sufficient and necessary to induce breast cancer cell motility.

104

A B 15 ** 15 * 10 10

5 5 Cell Speed (nm/s) Cell Speed (nm/s) 0 Control 10mM MCD 0 DMEM 2x103 10x103 100x103 100x103 ACM x g P x g P x g P x g S MCD (mM): - 10 IB:-Cd81 Pellet Sup.ACM Cd81 Speed (103 x g): 2 10 100 100 - ACM pellet Cd81 IB:-Cd81

Igsf8 IB:-Igsf8

Ptgfrn IB:-Ptgfrn

ACM C 15 LOC100044537 *** Gm6958 Angptl2 *** Rap1a Uba52 Aqp1 Tex21 Gm3579 Serinc3 ** Cd97 Kirrel Kng1 Nrp2 Rplp1 10 C2 Papln F11r Cd81 C3 Plekhb2 Igsf8 Gpnmb Cd44 Hpx Gapdh Trf F3 Myh9 5 Cp Slc3a2 Sdcbp Igsf8 Cell Speed (nm/s) Bsg Cd82 Tspan4 Pdcd6ip Cd81 Adam10 0 Slc1a5 Mfge8 Ptgfrn DMEM Ctrl IgG Cd81 Igsf8 Ptgfrn Tspan14

IgG: Ctrl IgGCd81Igsf8Ptgfrn Lamb1-1 Cd81 IB:-Cd81 Sdc4 Ptgfrn Pdgfrb Igsf8 IB:-Igsf8 Npm1 Ptgfrn IB:-Ptgfrn Mrgprf Slc39a14 Tspan9 Immunodepleted ACM Figure 2.6 L cell-secreted Exosomes Activate MDA-MB-231 Cell Motility

(A) ACM was subjected to differential ultracentrifugation at 2,000 x g, 10,000 x g and 100,000 x g and motility stimulation by the corresponding pellets (P) or the final supernatant (Sup. or S) was quantified and plotted (top panel; n = 29 cells per group. ** p < 0.001). Aliquots were also immunoblotted (IB) for the indicated exosome proteins (bottom panel). (B) The activity of ACM, previously incubated with methyl β -cyclodextrin (MβCD), was tested using the single cell motility assay and the results are plotted (top panel; * p <0.01). The exosome pellet was IB

105

for Cd81 (bottom panel). (C) Exosomes were captured from ACM with biotinylated antibodies against Cd81, Igsf8, Ptgfrn or control IgG. The resultant media was tested for activity (top panel; n = 31 cells per group. ** p < 0.001 and *** p < 0.0001). The exosome-depleted media was also subjected to IB (bottom panel). The Cd81, Igsf8 and Ptgfrn immunocomplexes were subjected to MS analysis and the associated proteins are indicated (right panel). Proteins previously reported in the literature as exosome components are highlighted in bold.

106

2.3.5. Cd81 Functions As a Component of L cell-secreted Exosomes to Promote BCC Motility

Exosomes are rich in diverse biologically active molecules, including proteins, RNAs and lipids that endow them with specific functions in different biological contexts (Simons and Raposo,

2009; Thery et al., 2009). Thus, to identify the exosome component(s) that are critical for promoting BCC protrusive activity and motility, I first subjected ACM to heat denaturation, which abolished its activity (Figure 2.7A). This suggested that proteins are likely the component(s) that mediate exosome activity. Next, I focused my attention on tetraspanins

Cd63, Cd81 and Cd82, as well as their interacting partners, Ptgfrn and Igsf8, all of which I found to be highly abundant in exosomes, as revealed by their proteomic analysis (Figure 2.5C).

To examine the role of Cd63, Cd81, Cd82, Igsf8 and Ptgfrn in promoting BCC motility, I knocked down their expression in L cells using siRNA and prepared conditioned media from the transfected cells. I found that downregulation of Cd81 expression in L cells led to loss of ACM- stimulated MDA-MB-231 cell motility (Figure 2.7B). In contrast, downregulation of Cd63,

Cd82 and Ptgfrn expression in L cells did not affect ACM-stimulated motility, whereas downregulation of Igsf8 expression promoted ACM activity (Figure 2.7B). To confirm the role of Cd81 in ACM exosomes, I generated L cells in which the expression of Cd81 was constitutively inhibited by two different shRNAs (Figure 2.7C). Similarly to the siRNA effect, downregulation of Cd81 expression in L cells by shRNAs inhibited ACM-induced MDA-MB-

231 cell motility. Importantly, loss of Cd81 expression did not affect exosome biogenesis and secretion, since the levels of the other exosome markers, namely Igsf8, Ptgfrn and Flotillin1, did not change (Figure 2.7D). Altogether these results indicate that Cd81 functions as a component of exosomes to stimulate BCC motility.

107

A B 20 * 15 *** 15 ***

10 10

5 5 Cell Speed (nm/s) Cell Speed (nm/s) 0 0 DMEM ACM ACM Mock siScr boiled siIgsf8 siCd81 siPtgfrnsiCd63siCd82 1.5

C 1.0 ** 12 *** 0.5

8 Relative mRNA Level Relative mRNA 0.0

4 siScr siIgsf8 siCd81 siPtgfrn siCd63 siCd82 Cell Speed (nm/s) 0 shScr shCd81#1 shCd81#2 D ACM ACM ACM CM: Orig. Sup. Pellet Scr + - + - + - shRNA Cd81#1 - + - + - + Cd81 Cd81 shRNA: Scr Cd81 IB:-Cd81 #1 #2 Igsf8 IB:-Igsf8 Cd81 IB:-Cd81 Ptgfrn IB:-Ptgfrn Flot.1 IB:-Flot.1 Actin IB:-Actin Tub. IB:-Tub. Actin Total Cell Lysate IB:-Actin

Figure 2.7 Cd81 Functions As a Component of Exosomes

(A) MDA-MB-231 cells were treated with control media (DMEM), ACM or boiled ACM. Cell speed was quantified and plotted. n = 36 cells per group. *** p <0.0001. (B) Conditioned media from L cells transfected with siRNA targeting different exosome components or a scramble sequence (Scr) was tested for activity in MDA-MB-231 cells. Cell speed was quantified and plotted (top panel). n = 32 cells per group. *** p <0.0001 and * p<0.01. Gene expression was determined by RT-qPCR and is shown as expression levels relative to siScr cells (bottom panel). (C) Conditioned media from L cells transfected with shRNAs targeting Cd81 (shCd81#1 or shCd81#2) or a scramble sequence (shScr) were tested for induction of MDA- MB-231 cell motility. Cell speed was quantified and plotted (top panel; n = 32 cells per group. *** p <0.0001). L cells total lysate was subjected to immunoblotting (IB) for Cd81 and actin (bottom panel). (D) Conditioned media (CM) prepared as in (C) was subjected to

108

ultracentrifugation. The original media (Orig.), the 100,000 x g supernatant (Sup.) and the 100,000 x g pellet were subjected to immunoblotting (IB) for the indicated exosome proteins.

109

2.3.6. Cd81 Is Necessary For L cell-stimulation of BCC Metastasis

Different tetraspanins have tumour-promoting (Tetraspanin 8 and Cd151) or tumour- suppressing functions (Cd9 and Cd82), although the molecular mechanisms involved are currently unclear (Zoller, 2009). To study the role of fibroblast-produced Cd81 in BCC metastasis, I coinjected L cells stably expressing shRNAs targeting either a control sequence or

Cd81 together with BCCs, using the orthotopic mouse model described above (Figure 2.8). In agreement with the previous in vivo experiments, the metastatic potential of MDA-MB-231 cells was significantly enhanced upon coinjection with control shRNA-expressing L cells

(Figure 2.8A-B). In contrast, in the presence of Cd81-deficient L cells, there was a dramatic suppression in the number of metastatic colonies (Figure 2.8A-B). Similarly, the metastatic potential of SUM-159PT cells was significantly suppressed by downregulation of Cd81 in L cells (Figure 2.8D-E). Importantly, downregulation of Cd81 expression in L cells did not affect significantly the growth rate or size of the primary tumour (Figure 2.8C and F), thus verifying that the enhanced metastatic potential of MDA-MB-231 cells was not due to accelerated tumour growth. Taken together, these results indicate that L cell-secreted Cd81-positive exosomes promote BCC metastasis by modulating cell motility.

110

MDA-MB-231 + shScr L cell p = 0.046 A B 6

MDA-MB-231 + shCd81#1/2 L cell 4 wks mfp 4

MDA-MB-231

MDA-MB-231 + MDA-MB-231 + MDA-MB-231 + # Colonies / Lung shScr L cell shCd81#1 L cell shCd81#2 L cell MDA-MB-231 0 MDA + MDA + MDA + MDA shScr shCd81#1 shCd81#2 L cell L cell L cell

C MDA + shScr L cell MDA + shCd81#1 L cell MDA + shCd81#2 L cell 6 MDA ) 3

2 SUM159-PT + D shCd81#1 L cell Tumour Volume ( cm Volume Tumour 0 0 5 10 15 20 25 30 Days post injection

E p = 0.01 F SUM159-PT + shScr L cell SUM159-PT + shCd81#1 L cell 3 p = 0.01 SUM159-PT

) 6 3 shScr L cell 2 shCd81#1 L cell 4

1 2

0 # Metastatic Foci / Lung SUM159-PT SUM159-PTSUM159-PT ( cm Volume Tumour 0 + shScr + shCd81#1 0 8 16 24 32 L cell L cell Days post injection

Figure 2.8 Cd81 Is Necessary For L cell-stimulated BCC Metastasis

(A) Schematic of experiment (top panel). Lungs from mice harbouring the indicated tumours were stained for vimentin and representative images of MDA-MB-231 metastases are shown (bottom panel. Scale bar = 35 µm). (B) The number of metastases were quantified and plotted as mean colony number ± SEM. n = 4-7 mice per group. p = 0.046 using two tailed Mann Whitney test. (C) The volume of the primary tumours was measured at the indicated time points and plotted as mean volume ± SEM. (D) Lungs from mice injected with SUM-159PT and shScr L cells, were stained for vimentin and a representative SUM-159PT metastatic foci is 111

shown (top panel; a higher power view of the boxed area is shown in the lower panel. Scale bar = 70 µm). (E) The number of metastatic foci were quantified and plotted as mean number ± SEM. n = 9-10 mice per group. p = 0.01 using two unpaired t-test with Welch’s correction. (F) The volume of the primary tumours in mice injected with the indicated cell populations was measured at the indicated time points and plotted as mean volume ± SEM.

112

2.3.7. Human CAFs Secrete Cd81-positive Exosomes That Stimulate BCC Motility

My next goal was to gauge the relevance of our findings in human breast cancer. For this purpose, I prepared conditioned media from breast cancer-associated fibroblasts (CAFs), which were isolated from two different patients (Hosein et al., 2010). I found that both CAF populations, CAF1 and CAF2, secreted exosomes that contained Cd81, Igsf8 and Flotillin1

(Figure 2.9A) and stimulated MDA-MB-231 cell motility (Figure 2.9B). This result suggests that human CAFs, like L cells, secrete Cd81-positive exosomes in vivo to promote BCC motility. To investigate the significance of stromally expressed Cd81 in human breast cancer,

Dr. Liang Zhang analysed a publicly available gene expression data set (GSE4823) of stroma isolated from invasive ductal carcinoma and normal breast reduction tissue (Finak et al., 2006).

He found that CD81 expression was significantly upregulated in human breast tumour- associated stroma (Figure 2.9C). Altogether, these results suggest that human CAFs secrete

Cd81-positive exosomes and that Cd81 functions in vivo to promote tumour malignancy.

113

A B 10 *** *** 8 DMEM CAF1CM DMEM CAF2CM Cd81 IB:-Cd81 6 Igsf8 IB:-Igsf8 4 Flot.1 IB:-Flot.1 2 100,000xg Pellet Cell Speed (nm/s) 0 DMEM CAF1 DMEM CAF2 Pellet Pellet Pellet Pellet C 4 *** 3

CD81 Expression Levels 0 Normal Tumour Stroma Stroma

Figure 2.9 Human CAFs Secrete Cd81-positive Exosomes That Stimulate BCC Motility

(A) Conditioned media prepared from primary CAFs, which were isolated from two different patients (CAF1 and CAF2), were subjected to ultracentrifugation in parallel with control media (DMEM). Exosomes (100,000xg pellet) were immunoblotted (IB) for Cd81, Igsf8 and Flotillin 1 (Flot.1). (B) CAF exosomes, CAF1 Pellet and CAF2 Pellet, were tested for activity in comparison to the pellet from control media (DMEM Pellet). Cell speed was quantified and is plotted as boxplots. n = 33 cells per group. *** p <0.0001. (C) Gene expression profiling of CD81 in tumour-associated stroma from a publicly available multiple microarray data set (GSE 4823). CD81 expression was determined in stroma from breast reduction mammoplasty (normal) and stroma adjacent to invasive ductal carcinoma (tumour). The black and red bars indicate the mean ± SEM respectively. *** p<0.0001.

114

2.4. Discussion

Although CAFs play a key role in tumour progression, the paracrine signals that mediate direct cross talk between CAFs and cancer cells in metastasis are poorly understood. Here, I present evidence that fibroblasts, including primary human CAFs, secrete exosomes that stimulate BCC motility and metastasis in a Cd81-dependent manner. First, I show that L cell-secreted factors have a potent effect in the protrusive activity and motility of several mammary cell lines, including carcinoma MDA-MB-231, SUM-159PT, and EMT6 cells, and the relatively normal epithelial Eph4 cells. This effect is immediate and reversible, suggesting that fibroblast factors activate a signalling pathway that directly modulates cell protrusive structures and motility.

Next, I demonstrate that in orthotopic xenograft mouse models, the metastatic potential of two different BCC lines is strongly stimulated by L cell fibroblasts. There are two caveats associated with the role of L cells in BCC metastasis. First, due to rapid tumour growth, the mice had to be sacrificed at 4 weeks post injection, and thus it was not possible to assess whether L cell-stimulated-BCC micrometastases would give rise to macrometastases. Second,

L cells are tumourigenic, probably due to genomic instability acquired from their long term cell culture (Sanford et al., 1956). Although, it is debatable whether human CAFs generally acquire genomic abnormalities during tumour progression (Hosein et al., 2010), it has been shown that fibroblasts that incur DNA damage after chemotherapy exposure promote cancer cell resistance to treatment and cancer growth (Sun et al., 2012). In the future, it will be important to confirm results from our mouse model with other sources of fibroblasts, such as human primary CAFs.

Next, I identify exosomes as the active fibroblast-secreted factor that stimulates BCC protrusive activity and motility. This was an unexpected finding since published reports on the role of

115

signalling molecules, including growth factors and cytokines, dominate the literature on intracellular communication. Indeed, I carried out an siRNA-based screen for 310 different growth factors, ECM proteins, and signalling molecules expressed by L cells in order to pinpoint the active factor(s). To my surprise, interference with these genes, some of which were previously shown to regulate cell motility, did not affect ACM activity. Instead, using protein purification and subsequent proteomic analysis of the active fraction from ACM, I was able to identify fibroblast exosomes as strong modulators of BCC cell motility. This finding adds to the emerging body of evidence on the significant role of exosomes in intracellular communication.

In the future, it will be important to inject purified CAF exosomes in mice harbouring BCC tumours and then ascertain the effect of exosomes in BCC metastasis.

Although tetraspanins are constitutive components of exosomes, their function in this context is poorly understood. Here, I pinpoint Cd81 as a crucial component for exosome activity. My data suggests that Cd81 has a specific effect, as an exosome component, in stimulating BCC motility, since downregulation of other tetraspanins, Cd63 and Cd82, and a functional partner of

Cd81, namely Ptgfrn, did not affect ACM activity. Interestingly, I found that Igsf8, another functional partner of Cd81, hindered ACM activity. Within the context of a cell, both Igsf8 and

Ptgfrn mediate Cd81 function by linking the tetraspanin web to ERM proteins, which in turn regulate the actin cytoskeleton (Charrin et al., 2009b; Coffey et al., 2009; Montpellier et al.,

2011; Sala-Valdes et al., 2006). My data suggests that Cd81 controls the motility of recipient

BCCs within the context of exosomes through a different mechanism that has yet to be elucidated. Next, I found that Cd81 expression in L cells is necessary for L-cell-induced BCC metastasis because silencing of Cd81 lead to strong suppression of BCC lung metastases but did

116

not affect primary tumour growth. This indicates that fibroblast-secreted, Cd81-positive exosomes are involved in a metastasis-specific pathway in BCCs. Accordingly, we found that

Cd81 is significantly overexpressed in stroma associated with human invasive breast carcinoma.

Previously, it was reported that Tetraspanin 8 (Tspn8) functions as a component of tumour- derived exosomes to induce angiogenesis (Nazarenko et al., 2010). Now, my findings extend our understanding of tetraspanins as functional components of exosomes in stimulating BCC motility and metastasis.

117

Chapter 3 : Fibroblast Exosomes Mobilize the Wnt-PCP Pathway to Promote Breast Cancer Cell Motility and Metastasis

A version of this chapter was published by:

Valbona Luga, Liang Zhang, Alicia M. Viloria-Petit, Abiodun A. Ogunjimi, Mohammad R. Inanlou, Elaine Chiu, Abdel Nasser Hosein, Marguerite Buchanan, Mark Basik, and Jeffrey L. Wrana

in Cell 2012 Dec 21;151(7):1542-56.

Attributions:

The RT-qPCR reactions in Figures 3.1 and 3.7-8 and the immunofluorescence experiments in Figures 3.3-3.6, 3.8 and 3.10 were performed by Dr. Liang Zhang. The mouse experiments in Figure 3.2B-D were performed together with Dr. Liang Zhang and Dr. Alicia M. Viloria-Petit. All the remaining experiments were performed by Valbona Luga.

118

3.1. Introduction

PCP signalling regulates tissue polarity across a plane perpendicular to the apicobasal axis

(Gray et al., 2011; Seifert and Mlodzik, 2007; Zallen, 2007). PCP is manifested in both static and dynamic tissues. PCP signalling is transduced by an evolutionary conserved module of proteins, comprised of Frizzled (Fzd), Dishevelled (Dvl), Van Gogh-like (Vangl), Prickle (Pk),

Diego (Dgo) and EGF LAG seven-pass G-type receptor 1 (Celsr). PCP proteins assemble in discrete complexes, which are asymmetrically distributed and function to regulate each other’s subcellular localization in a manner that is propagated across the tissue (Gray et al., 2011;

Seifert and Mlodzik, 2007; Zallen, 2007). The signal(s) that trigger PCP have been intensively pursued, however they remain obscure. In vertebrates, Wnt ligands have been implicated in

PCP and CE movements (Gros et al., 2009; Heisenberg et al., 2000; Qian et al., 2007; Tada and

Smith, 2000). Due to posttranslational modifications, Wnts tether tightly to the plasma membrane and ECM (Bartscherer and Boutros, 2008; Mikels and Nusse, 2006; Port and Basler,

2010). How Wnts function as long range signalling molecules is unclear, however several mechanisms may be involved, including lateral diffusion of the ligand by heparan sulphate proteoglycans (Mikels and Nusse, 2006) and solubilisation by high-density lipoproteins

(Panakova et al., 2005), carrier proteins (Mulligan et al., 2012) and vesicles (Gross et al., 2012;

Morrell et al., 2008). The mechanism(s) that facilitate Wnt signalling in PCP are currently unknown.

Several core PCP pathway components are overexpressed in a number of cancers and are associated with tumour cell growth, migration and metastasis (Jessen, 2009). However, it is unknown whether these components function in the context of the core PCP pathway and what

119

molecular mechanisms underlie the potential role of Wnt-PCP signalling in cancer. Here I report that fibroblast exosomes, including CAF exosomes, stimulate breast cancer cell protrusive activity, motility and metastasis by activating the core PCP pathway. Upon exosome treatment, Fzd-Dvl and Vangl-Pk protein complexes distribute in a mutually exclusive fashion in the protrusive structures of single and motile BCCs, similar to their previously described subcellular distribution in normal planar-polarized epithelia. In addition, exosome-driven BCC metastasis is dependent on Pk1 expression. Moreover, I show that exosome activity is dependent on autocrine Wnt signalling in BCCs and demonstrate that fibroblast exosomes are internalized by BCCs where Wnt11 is loaded onto exosomes.

3.2. Materials and Methods

3.2.1. Cell Culture

MDA-MB-231 cells, SUM-159PT cells, L cells and human CAFs were described in Materials and Methods in Chapter 2.

3.2.2. Conditioned Media Preparation

Conditioned media was prepared as described in Materials and Methods in Chapter 2. MDA-

MB-231 cells were starved in DMEM supplemented with 50 U/ml penicillin/streptomycin

(Gibco), 0.25 mg/ml amphotericin B (Gibco) and 0.2 % FBS.

120

3.2.3. Single Cell Motility Assay

The assay was carried out as described in Materials and Methods in Chapter 2.

3.2.4. Cell Transfection

MDA-MB-231 cells were transfected with 40 nM siGENOME SMARTpool siRNAs

(Dharmacon) using DharmaFECT 4 (Dharmacon) transfection reagent according to manufacturer’s instructions. siCONTROL Non-Targeting siRNA Pool#2 (D-001206-14;

Dharmacon) was used as control siRNA. MDA-MB-231 cells were transfected with cDNA or shRNA expression vectors using LipofectamineTM LTX (Invitrogen) following according to manufacturer’s instructions. shRNAs for human Prickle1 (RHS4430-101132271, RHS4430-

101514564, RHS4430-101129727; Thermo Scientific) were purchased from Open Biosystems and the sequences are listed below. TRC Lentiviral eGFP shRNA (RHS4459; Thermo

Scientific) was used as control shRNA. MDA-MB-231 cells were selected with 3.6 mg/ml puromycin.

3.2.5. shRNA Sequences

Gene Item Cat. # Clone ID Hairpin Sequence

PK1 RHS4430- V3LHS_34 TGCTGTTGACAGTGAGCGATAAGACCGAGTTAAAG 101132271 7575 CAAAATAGTGAAGCCACAGATGTATTTTGCTTTAAC TCGGTCTTACTGCCTACTGCCTCGGA

PK1 RHS4430- V3LHS_34 TGCTGTTGACAGTGAGCGAAGCAGTGTGGTTTGAAG 101129727 7577 ATAATAGTGAAGCCACAGATGTATTATCTTCAAACC ACACTGCTCTGCCTACTGCCTCGGA

121

Gene Item Cat. # Clone ID Hairpin Sequence

PK1 RHS4430- V3LHS_34 TGCTGTTGACAGTGAGCGAAGGGTCAGATTTACTGC 101514564 7572 TCAATAGTGAAGCCACAGATGTATTGAGCAGTAAA TCTGACCCTGTGCCTACTGCCTCGGA

3.2.6. Cell Lysis And Immunoblotting

The cells were lysed as described in Materials and Methods in Chapter 2. Protein samples were subjected to SDS-PAGE followed by transfer onto nitrocellulose membrane (162-0115,

BioRad), probing with primary antibodies (as indicated below) and HRP-linked secondary antibodies (Santa Cruz), and detection with a SuperSignal chemiluminescence reagent (34095,

Thermo Scientific). The antibodies were used as follows: rabbit anti-Smurf1 (sc-25510, Santa

Cruz, 1:1000), rabbit anti-Smurf2 (sc-25511, Santa Cruz, 1:1000), rabbit anti-Dvl1 (AB5970,

Millipore, 1:1000), rabbit anti-Dvl2 (3224, Cell Signalling, 1:1000), rabbit anti-Dvl3 (3218, Cell

Signalling, 1:1000), mouse anti-Cd81 (sc-166029, Santa Cruz, 1:1000), rabbit anti-Igsf8

(ab74890, Abcam, 1:1000), rabbit anti-Ptgfrn (ab97567, Abcam, 1:1000), mouse anti-Flotillin1

(610820, BD, 1:1000), goat anti-GFP (600-101-215, Rockland, 1:1000), rat anti-HA

(11867423001, Roche, 1:5000), mouse anti-βTubulin (T5201, Sigma-Aldrich, 1:10000) and rabbit anti-Gapdh (G9545, Sigma-Aldrich, 1:10000).

3.2.7. RNA Extraction and Real-Time Quantitative PCR (RT-qPCR)

Total RNA was isolated and processed for RT-qPCR as described in Materials and Methods in

Chapter 2.

122

3.2.8. PCR Primers

Gene Species RefSeq ID Primer Sequence 5'-3' Amplicon Size (bp)

FZD3 Human NM_017412.2 CTCTGTATTTTGGGTTGGAAGC 150

GGAGTATTTGGATCCCTCAGG

FZD6 Human NM_003506.2 TTATGACCAGAGTATTGCCGC 150

AGTTTACGACAAGGTGGAACC

FZD7 Human NM_003507.1 CTACCACAGACTTAGCCACAG 120

AGCAGTTCTTTTCCCTACCG

PK1 Human NM_001144881 GTCGCCATTGGCACATGAAACACT 239

AGAGGCTTTACACTGGGCACAAGA

PORCN Human NM_022825 TCCCTGTTTGATGTCGATGTG 144

TCAGCCTATGAGACGGTAGAAG

VANGL1 Human NM_001172411 TGGATGCCCTCCTCTTCATCCATT 237

TATGCTTGGCTGCTCGGAATTTGG

WNT5A Human NM_003392 TCGCCCAGGTTGTAATTGAAG 124

TGAGAAAGTCCTGCCAGTTG

WNT11 Human NM_004626 CCAAGCCAATAAACTGATGCG 85

GCACTTACACTTCATTTCCAGAG

GAPDH Human NM_002046 AATCCCATCACCATCTTCCA 82

TGGACTCCACGACGTACTCA

HPRT Human NM_000194 ATGGACAGGACTGAACGTCTTGCT 80

TTGAGCACACAGAGGGCTACAATG

123

3.2.9. Immunofluorescence

The cells suspended at very low density in either DMEM or ACM assay media, prepared as described for the single cell motility assay, were seeded on coverslips (12-545-80, Thermo

Scientific). Following an overnight incubation with the assay media, the cells were fixed with 4

% paraformaldehyde in PBS, permeabilized with 0.25 % Triton X-100 in PBS and blocked with

2 % bovine serum albumin in PBS. The primary antibodies used for immunostaining were: goat anti-Fzd6 (AF3149, R&D, 1:250), rabbit anti-Dvl1 (D3320, Sigma-Aldrich, 1:500), rabbit anti-

Wnt11 (sc-50360, Santa Cruz, 1:2,000), an anti-Prickle1 Fab antibody (1ug/ml working concentration) generated in-house by Miss Elaine Fuchs using phage display technology, and

M2 monoclonal mouse anti-Flag (Sigma-Aldrich, 1:5,000). The secondary antibodies used for immunostaining were: DyLight 488-donkey-anti-goat (705-485-147, Jackson Immunoresearch,

1:500), DyLight 549-donkey-anti-mouse (715-507-003, Jackson Immunoresearch, 1:500) and

DyLight 549-donkey-anti-rabbit (611-742-127, Rockland, 1:500). The samples were counterstained with 4’, 6-Diamidino-2-phenylindole dihydrochloride (DAPI) (D9542, Sigma-

Aldrich) and Alexa-647-phalloidin (A22287, Invitrogen) and then were mounted onto slides

(12-550-143, Thermo Scientific). Images were acquired using an inverted microscope

(DMIRE2, Leica) equipped with a spinning disk confocal scanner (CSU10, Yokogawa), a

40x/NA 1.25 and a 63x/NA 1.32 oil immersion objective lenses (HCX PL APO, Leica) and an

EM-CCD camera (ImagEM, Hamamatsu).

3.2.10. L cell / MDA-MB-231 Orthotopic Mouse Model of Breast Cancer Metastasis

The assays were carried out as described in Materials and Methods in Chapter 2.

124

3.2.11. DNA Constructs

The human Wnt11 coding sequence (NM_004626) was PCR amplified in frame with a C- terminal HA tag and subcloned into the pCAGIP expression vector (Narimatsu et al., 2009) using MluI and BglII restriction enzymes. The mouse Cd81 (NM_133655) and EYFP (BD

Biosciences) coding sequences were PCR amplified and subcloned in the pLVX expression vector (Clontech Laboratories, Inc) using XhoI and XbaI restriction enzymes.

3.2.12. Exosome Isolation by Differential Ultracentrifugation

Exosomes were isolated as described in Materials and Methods in Chapter 2.

3.2.13. Exosome Internalization Assay: Immunoprecipitation

MDA-MB-231 cells were incubated with conditioned media collected from control or CD81-

EYFP-expressing L cells. Twenty hours later, MDA-MB-231 cells were rinsed three times with

PBS and lysed with TNTE buffer (as described in Materials and Methods in Chapter 2). Cd81-

EYFP internalized by MDA-MB-231 cells was immunoprecipitated following incubation with

GFP-Trap®_A beads (Chromotek) at 4 °C for 3 hrs. The precipitate was washed three times with lysis buffer and then was subjected to immunoblot analysis with a goat anti-GFP antibody

(600-101-215, Rockland, 1:1000).

125

3.2.14. Exosome Internalization Assay: Immunofluorescence and Dynasore Treatment

Conditioned media collected from control or CD81-EYFP-expressing L cells was concentrated

6-fold with Vivaspin concentrators (10 kDa MWCO; GE Healthcare) as described in Materials and Methods in Chapter 2 and was used to treat MDA-MB-231 cells for immunofluorescence experiments as described above. To inhibit endocytosis, the cells were also treated with 100uM of Dynasore (324410; EMD). The Cd81-EYFP protein was labelled with a goat anti-GFP antibody (600-101-215, Rockland, 1:600) and a DyLight 488-donkey-anti-goat (705-485-147,

Jackson Immunoresearch, 1:500).

3.2.15. Electron Microscopy

EM analysis of whole mounted exosomes was carried out as described in Materials and Methods in Chapter 2. Wnt11-HA containing exosomes were labelled with an anti-Wnt11 antibody

(ab31962, Abcam) and a gold-conjugated secondary antibody (25705, EMS). Images were captured using a transmission electron microscope (TecnaiTM 20, FEI) at 200 kV.

3.2.16. Exosome Immuno-Magnetic Extraction

Exosomes were immunocaptured using a biotinylated Cd81 antibody (NBP1-28138, Novus Biologicals) or an IgG isotype control (NBP120-18472, Novus Biologicals) as described in Materials and Methods in Chapter 2. Then, the immunocomplexes were subjected to immunoblotting with a rat anti-HA (11867423001, Roche, 1:10,000) and a mouse anti-Cd81

(sc-166029, Santa Cruz, 1:1000) antibodies.

126

3.3. Results

3.3.1. Core PCP Components Mediate Fibroblast-Induced BCC Protrusive Activity and Motility

In Chapter 2, I demonstrated that fibroblast exosomes strongly stimulate cancer cell protrusive activity, motility and metastasis. Smurf1 and Smurf2 ubiquitin ligases function to regulate cell polarity, protrusive activity and cancer cell motility (Ozdamar et al., 2005; Sahai et al., 2007;

Viloria-Petit et al., 2009; Wang et al., 2003). Hence, I investigated the role of Smurf1 and

Smurf2 in ACM-induced MDA-MB-231 cell motility and found that siRNA-mediated downregulation of either protein inhibited ACM activity (Figure 3.1A). Smurf1 and Smurf2 are also critical regulators of PCP and CE movements in mice (Narimatsu et al., 2009).

Consequently, I considered whether ACM-induced BCC protrusions and motility might be facilitated by the core PCP pathway. To test this hypothesis, I silenced the expression of the highly conserved, core PCP pathway components, including Fzd, Dvl, Pk and Vangl in BCCs and evaluated ACM-induced cell protrusive activity and motility (Figure 3.1). I found that

Fzd3, Fzd6 and Fzd7 receptors, which are most prominently associated with vertebrate PCP signalling (Ohkawara et al., 2011; Wang et al., 2006), mediated ACM activity (Figure 3.1B).

Next, I knocked down all Dvl proteins in MDA-MB-231 cells using siRNA and found that interference with Dvl1 inhibited cell protrusions and motility, whereas interference with Dvl2 had a minor effect, possibly due to functional redundancy (Figure 3.1C). Examination of Pk and

Vangl mRNA levels in MDA-MB-231 cells by RT-qPCR revealed that Pk1 and Vangl1 are the only Prickle and Vangl isoforms respectively expressed in these cells (experiment carried out by

Dr. Liang Zhang). Importantly, silencing of Pk1 and Vangl1 resulted in loss of ACM-induced cell protrusive activity and motility, in particular, Pk1 silencing led to nearly complete cessation of movement (Figure 3.1D). In vertebrates, other proteins, such as Ryk and Ror2, are also 127

implicated in core PCP signalling (Andre et al., 2012; Gao et al., 2011). Consequently, I knocked down the expression of Ror2 and Ryk in MDA-MB-231 cells using siRNA and assessed ACM-induced cell motility. Interestingly, downregulation of Ryk expression resulted in enhancement of ACM activity, suggesting that there is a negative feedback loop between Ryk and core PCP components, whereas Ror2 expression was not necessary for ACM activity

(Figure 3.1E). Next, I tested the role of core PCP components in ACM-stimulated SUM-159PT cell motility and found that interference with Fzd7, Dvl3 and Pk1 significantly inhibited ACM activity (Figure 3.1F). In chapter 2, I demonstrated that human CAFs secrete exosomes that stimulate BCC motility. To test whether the PCP pathway is necessary for this, I knocked down

Pk1 in MDA-MB-231 cells and stimulated them with CAF exosomes (Figure 3.1G). I found that interference with Pk1 abolished exosome-stimulation of MDA-MB-231 cell motility

(Figure 3.1G). Collectively, these results indicate that the core PCP pathway plays a critical role in fibroblast-induced BCC protrusive activity and motility.

128

A siScr siSf1 siSf2 B siScr siFzd3 siFzd6 siFzd7

*** * 15 *** 20 *** siRNA:ScrSf1Sf2 * 1.5 15 10 Smurf1 IB:-Sf1 1.0 Smurf2 IB:-Sf2 10 5 Gapdh IB:-Gapdh 5 0.5 Total Cell Lysate Cell Speed (nm/s) 0 Cell Speed (nm/s) siScr siSf1 siSf2 0 0.0

siScr siFz3 siFz6 siFz7 Level Relative mRNA siScr siFz3 siFz6 siFz7

C siScr siDvl1 siDvl2 siDvl3 D siScr siPk1 siVangl1

15 siRNA: *** ScrDvl1Dvl2Dvl3 *** Dvl1 IB:-Dvl1 *** 1.5 10 1.0 Dvl2 IB:-Dvl2 5 0.5 Dvl3 IB:-Dvl3 0.0 Cell Speed (nm/s) Cell Speed (nm/s) -Tub IB: - Tub Level Relative mRNA 0 siScr siPk1siVangl1 siScr siPk1siVangl1 siScrsiDvl1siDvl2siDvl3 Total Cell Lysate

E F 10 G *** *** 15 8 *** ns*** 8 *** 6 10 6 4 4 5 2 Cell Speed (nm/s) Cell Speed (nm/s)

Cell Speed (nm/s) 2 0 0 siScr siScr siPk1 siScr siRor2 siRyk DMEMp CAF2p CAF2p 0 siScr siDvl3 siPk1siScr siFzd7

Figure 3.1 Core PCP Components Mediate ACM Activity in BCCs

MDA-MB-231 cells transfected with siRNAs targeting Smurf 1 (Sf1), Smurf 2 (Sf2) (A), Fzd3, Fzd6, Fzd7 (B), Dvl1, Dvl2, Dvl3 (C), Pk1, Vangl1 (D), Ror2, Ryk (E), or a scramble sequence (Scr) (A-E) were treated with ACM as in Figure 2.1A. Representative images are shown (A-D, top panels) with the appearance of short blunted protrusions (white hollow arrows) and long cell protrusions (white arrows) highlighted. Cell speed was quantified (A-D, bottom left panels) and plotted. n = 26-33 cells per group. *** p <0.0001. Total cell lysates were subjected to immunoblotting (IB) with the indicated antibodies to assess knockdown efficiency (A and C, 129

bottom right panels). Gapdh (A) and β-tubulin (β-tub, (C)) were used as loading control. Gene expression was determined by RT-qPCR and is shown as expression levels relative to siScr cells (B and D, bottom right panels). (F) SUM-159PT cells transfected with siRNAs targeting Dvl3, Pk1, Fzd7 or Scr were treated with ACM as in (A). Cell speed was quantified and plotted. n = 28 cells per group. *** p <0.0001. (G) MDA-MB-231 cells, transfected with the indicated siRNAs, were treated with CAF exosomes (CAF2p) or control pellet (DMEMp) as described in Figure 2.9A-B. Cell speed was quantified and plotted. n = 40 cells per group. *** p <0.0001.

130

3.3.2. Pk1 Expression Is Necessary For L cell-Stimulated BCC Metastasis

Next, I aimed to determine whether the enhanced metastatic potential of breast cancer cells promoted by L cells in vivo is dependent on the core PCP pathway. To test this hypothesis, I focused on Pk1, which as the only Prickle isoform expressed in MDA-MB-231 cells allowed efficient interference with the PCP pathway (Figure 3.1D and 3.1G). Similarly as in transient knock down experiments, stable silencing of Pk1 in MDA-MB-231 cells by three different shRNAs resulted in significant inhibition of motility in response to ACM (Figure 3.2A).

Subsequently, the cells stably expressing Pk1 or control shRNAs were coinjected with L cells in the mammary fat pad of mice (Figure 3.2B). In accordance with our previous in vivo experiments, the metastatic potential of MDA-MB-231 cells harbouring control shRNA was significantly enhanced upon coinjection with L cells (Figure 3.2B). In contrast, Pk1 downregulation led to a dramatic suppression in the number of L cell induced MDA-MB-231 metastatic foci in lungs (Figure 3.2B). Indeed, the number of micrometastases in mice coinjected with shPk1-expressing MDA-MB-231 cells and L cells was similar to animals injected with MDA-MB-231 cells alone (Figure 3.2C-D). Importantly, inhibition of Pk1 expression in MDA-MB-231 cells did not affect the growth rate or size of the primary tumour

(Figure 3.2C-D), indicating that PCP signalling specifically regulates breast cancer metastasis.

Altogether, these results indicate that the core PCP pathway in tumour cells is required for fibroblast stimulation of BCC metastasis in vivo.

131

A 15 *** *** B shScr/shPk1 MDA + L cell ** 4 wks mfp 10 shScr/shPk1 MDA

5 shScr MDA shPk1#1 MDA + L cell + L cell

Cell Speed (nm/s) 0 shScr shPk1shScrshPk1 shPk1 #1 #2 #3 1.5

1.0

0.5

Relative mRNA Level Relative mRNA 0.0 shScr shPk1shScrshPk1shPk1 #1 #2 #3 shScr MDA shPk1#1 MDA C D 5 p = 0.048 ns 2.0 p = 0.005 4 1.5 3 1.0 2

1 0.5 # Colonies / Lung # Colonies / Lung 0 0.0 shScr shPk1#1 shScr shPk1#1 shScr shPk1#2shPk1#3 shScr MDA + MDA + MDA MDA MDA + MDA + MDA + MDA L cell L cell L cell L cell L cell shScr MDA + L cell shScr MDA + L cell )

10 ) 3 shPk1#1 MDA + L cell 8 shPk1#2 MDA + L cell 3 shScr MDA shPk1#3 MDA + L cell 8 shPk1#1 MDA 6 shScr MDA 6 4 4 2 2 Tumour Volume ( cm Volume Tumour 0 ( cm Volume Tumour 0 0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 Days post injection Days post injection

Figure 3.2 Prickle1 Is Required For L cell-stimulated MDA-MB-231 Cell Metastasis

(A) MDA-MB-231 cells stably expressing three different shRNAs targeting Pk1 (shPk1#1, shPk1#2 or shPk1#3), or a scramble sequence (Scr) were treated with ACM as in Figure 2.1A. Cell speed was quantified and plotted (top panel). n = 28 cells per group. ** p<0.001 and *** p <0.0001. Gene expression was determined by RT-qPCR and is shown as expression levels relative to shScr cells (bottom panel). (B) Experimental schematic (top panel). MDA-MB-231 (MDA) cells were injected alone or together with L cells in the mammary fat pad (mpf) of mice

132

and lung metastases were detected as in Figure 2.3D (bottom panel; scale bar = 35 µm). (C-D) Metastatic colonies in mice injected with MDA-MB-231 cells in the presence (n = 8-10 mice per group) or absence (n = 6-8 mice per group) of L cells were quantified ((C-D) top panels; mean colony number ± SEM. p = 0.048 (C) and p = 0.005 (D) using two tailed unpaired t-test with Welch’s correction). Note that no metastases were observed in mice coinjected with shPk1#3 MDA-MB-231 cells and L cells (D, top panel). Primary tumour growth was measured at the indicated time points and is plotted as the mean volume ± SEM (C-D, bottom panels).

133

3.3.3. Core PCP Components Distribute Asymmetrically in ACM-treated BCCs

How PCP signalling is conducted intracellularly is currently unclear. However, it is known that core PCP proteins are asymmetrically distributed along the PCP axis (Seifert and Mlodzik,

2007; Zallen, 2007). In the fly wing cells, the proximal subset of proteins, consisting of Pk and

Vang and the distal subset, consisting of Fz, Dsh and Dgo, localize at the plasma membrane on opposite sides of cell-cell contacts (Seifert and Mlodzik, 2007; Zallen, 2007). Similarly, in vertebrates the asymmetric localization of core PCP components has been observed in different tissues, although a more thorough understanding of the localization of different PCP proteins in different biological contexts has yet to be realized (Ciruna et al., 2006; Gray et al., 2011;

Narimatsu et al., 2009; Yin et al., 2008). Based on this, I aimed to determine whether ACM stimulation affects the localization of core PCP components in individual breast cancer cells.

Hence, MDA-MB-231 cells were stimulated with ACM, similarly as in the motility assay, and immunostained for core PCP components using antibodies whose specificity was previously established by Dr. Liang Zhang. In ACM-stimulated MDA-MB-231 cells, endogenous Fzd6 accumulated and colocalized with endogenous Dvl1 at the leading front of cell protrusions, marked by filamentous, F-actin networks (Figure 3.3A). In contrast, in control cells, Fzd6 was evenly distributed across the cell membrane and displayed little colocalization with Dvl1

(Figure 3.3A). Quantification of Fzd6-Dvl1 colocalization revealed that ACM treatment stimulated the association of these proteins at the tip of cell protrusions (Figure 3.3B-C). Next,

Pk1 subcellular distribution was examined using a Pk1-specific Fab antibody fragment generated in house my Elaine Chiu. Endogenous Pk1 colocalized extensively with cortical F- actin, as previously reported (Veeman et al., 2003), and was aligned along the non-protrusive cell membrane, concentrating in regions flanking ACM-induced cell protrusions (Figure 3.4A).

Importantly, Pk1 staining at the base of ACM induced cell protrusions was mutually exclusive 134

to Fzd6 staining (Figure 3.4A). Likewise, ectopically expressed Flag-Vangl1 was enriched along the non-protrusive cell cortex in both control and ACM treated cells (Figure 3.4B).

Moreover, Vangl1 localization in ACM stimulated cell protrusions was mutually exclusive to

Fzd6 (Figure 3.4B). Similar analysis of SUM-159PT cells revealed that Fzd7 was enriched and colocalized with Dvl3 at the leading edge of ACM induced cell protrusions, marked by F-actin networks (Figure 3.5A-B). Importantly, CAF exosomes also stimulated the asymmetric accumulation of Fzd6 and Dvl1 at the tip of MDA-MB-231 cell protrusions (Figure 3.6A-B).

Collectively, these results demonstrate that Fzd-Dvl and Vangl-Pk complexes are asymmetrically distributed in exosome-stimulated BCC protrusions in a manner analogous to planar-polarized epithelial cells. This is the first time, to the best of our knowledge, that such asymmetric distribution of PCP components has been reported in single, motile and cancerous cells.

135

A Fzd6 Fzd6 Dvl1 Actin Dvl1 Actin DMEM

Fzd6 Fzd6 Dvl1 Actin Dvl1 Actin ACM

B C 1.5 Fzd6 Dvl1 Protrusion Actin 1.0

Cortex 2 Cortex 1 0.5 Cell Body Colocalization Coefficient 0.0 Protrusion Body Cortex1 Cortex2

Figure 3.3 Localization of Fzd6-Dvl1 Complexes in ACM-treated MDA-MB-231 Cells

(A) Representative confocal images of MDA-MB-231 cells cultured in DMEM or ACM are shown. Endogenous Fzd6 (green (first column) and white (second column)) was immunostained in parallel with endogenous Dvl1 (red (first column) and white (third column)). F-actin-rich cell protrusions were detected by phalloidin staining (blue (first column) and white (fourth column)). The localization of Fzd6 at the leading edge of ACM-stimulated cell protrusions is indicated (A-C; green arrows (second column)). The colocalization of Dvl1 (A, red arrows) with Fzd6 (green arrows) is indicated (yellow arrows). The inset shows a higher magnification of the indicated white, boxed area enclosing cell protrusions. Scale bar = 10 µm. (B) The regions of interest (ROI), including cell protrusions, the cell body and the cell cortex flanking the protrusions, were defined based on actin staining. (C) The colocalization coefficient between Dvl1 and Fzd6 voxels within ROIs were calculated and the results are plotted as box and whisker plots. n = 36 cells per group. 136

A Fzd6 Fzd6 Pk1 Actin Pk1 Actin DMEM

Pk1 Fzd6 Fzd6 Pk1 Actin Pk1 Actin ACM

B Fzd6 Fzd6 Vangl1 Actin Vangl1 Actin DMEM

Fzd6 Fzd6 Vangl1 Actin Vangl1 Actin ACM

Figure 3.4 Pk1 and Vangl1 Localize Asymmetrically in ACM-treated MDA-MB-231 Cells

Representative confocal images of cells cultured in DMEM or ACM are shown. Endogenous Fzd6 (green (first column) and white (second column)) was immunostained in parallel with endogenous Pk1 (A) or ectopically-expressed Flag-Vangl1 (B), (red (first column) and white (third column)). F-actin-rich cell protrusions were detected by phalloidin staining (blue (first column) and white (fourth column)). The localization of Fzd6 at the leading edge of cell protrusive structures is indicated (A-B; green arrows (second column)). Pk1 and Vangl1 staining along the non-protrusive cell membrane are indicated (red line). Note, Pk1 and Vangl1 are depleted from the leading edge of ACM-stimulated cell protrusions that is marked by rich F-

137

actin and Fzd6 staining (green arrows). The inset shows a higher magnification of the indicated white, boxed area enclosing ACM-stimulated cell protrusions. Scale bar = 10 µm.

138

A Fzd7 Fzd7 Dvl3 Actin Dvl3 Actin DMEM

Fzd7 Fzd7 Dvl3 Actin Dvl3 Actin ACM

B 1.0

0.8

0.6

0.4

0.2

Colocalization Coefficient 0.0 Protrusion Body Cortex1 Cortex2

Figure 3.5 Localization of Fzd7-Dvl3 Complexes in ACM-stimulated SUM-159PT Cells

(A) Representative confocal images of SUM-159PT cells cultured in DMEM or ACM are shown. Endogenous Fzd7 (green (first column) and white (second column)) was immunostained in parallel with endogenous Dvl3 (red (first column) and white (third column)). F-actin-rich cell protrusions were detected by phalloidin staining (blue (first column) and white (fourth column)). The localization of Fzd7 at the leading edge of ACM-stimulated cell protrusive structures is indicated (green arrows (second column)). The colocalization of Dvl3 (red arrows) with Fzd7 (green arrows) in cell protrusions is indicated (yellow arrows). The inset shows a higher magnification of the indicated white, boxed area enclosing cell protrusive structures. Scale bar = 10 µm. (B) The colocalization coefficient between Dvl3 and Fzd7 voxels within ROIs defined in ACM-treated cells as described in Figure 3.3B were calculated and the results are plotted as box and whisker plots. n = 25 cells per group.

139

A Fzd6 Dvl1

B 1.0

0.8

0.6

Fzd6 Actin 0.4 Dvl1 Actin 0.2

0.0 Colocalization Coefficient ProtrusionBody Cortex1Cortex2

Figure 3.6 Fzd6-Dvl1 Complexes Localize Asymmetrically in Exosome-stimulated BCCs

(A) Confocal images of MDA-MB-231 cells cultured with CAF exosomes are shown. Endogenous Fzd6 (white (top left panel) and green (bottom left panel)) was immunostained in parallel with endogenous Dvl1 (white (top right panel) and red (bottom left panel)). F-actin-rich cell protrusive structures were detected by phalloidin staining (blue (bottom left panel) and white (bottom right panel)). The localization of Fzd6 at the leading edge of cell protrusions is indicated (green arrows). The colocalization of Dvl1 (red arrows) with Fzd6 at cell protrusions is indicated (yellow arrows). The inset shows a higher magnification of the indicated white boxed area. Scale bar = 10 µm. (B) The colocalization coefficients between Dvl1 voxels and Fzd6 voxels within ROIs defined in exosome-treated MDA-MB-231 cells (as described in Figure 3.3B-C) were calculated and plotted as box and whisker plots. n = 26 cells per group.

140

3.3.4. Autocrine Wnts Mediate Fibroblast-Induced BCC Protrusive Activity and Motility

The function and subcellular localization of the core PCP components in ACM-stimulated BCCs suggested that fibroblasts secrete a factor(s) that promotes protrusive activity, motility and metastasis in BCCs via the PCP pathway. The signals that trigger the initial asymmetry in planar polarized tissue remain unknown, however genetic studies in different vertebrate model systems have revealed that Wnt ligands, primarily Wnt5a and Wnt11, are key regulators of PCP

(Gros et al., 2009; Heisenberg et al., 2000; Qian et al., 2007; Tada and Smith, 2000). Hence, I examined whether L cells express Wnt ligands that might trigger PCP signalling. Surprisingly, the expression of Wnt5a and Wnt11 in L cells was not detectable by RT-qPCR. Next, I knocked down the expression of porcupine, an acyltransferase essential for Wnt lipidation and secretion

(Herr and Basler, 2012; Port and Basler, 2010; Takada et al., 2006), in L cells using siRNA. I found that conditioned media prepared from porcupine-deficient L cells was as active as control media (Figure 3.7A). These results suggest that ACM activity is not mediated by L cell- secreted Wnt ligands. Therefore, I considered whether ACM might regulate autocrine Wnt signalling in BBCs. First, I silenced porcupine expression in MDA-MB-231 cells and observed potent inhibition of ACM-induced protrusive activity and motility (Figure 3.7B). Similarly, silencing of porcupine in SUM-159PT cells resulted in significant inhibition of ACM activity

(Figure 3.7C). Next, I knocked down the expression of Wnt5a and Wnt11 in MDA-MB-231 cells and found that interference with Wnt11, but not Wnt5a, inhibited ACM-stimulated BCC protrusive activity and motility (Figure 3.8A). Further, I aimed to test the association of Wnt11 with its putative receptor, Fzd6, by immunofluorescence. Endogenous Wnt11 staining in unstimulated cells revealed that the ligand is distributed in puncta concentrated inside cells

(Figure 3.8B), consistent with Wnt vesicular trafficking (Coudreuse and Korswagen, 2007;

Pfeiffer et al., 2002). Moreover, in unstimulated cells, Wnt11 displayed little colocalization 141

with Fzd6 (Figure 3.8B). In comparison, in ACM-stimulated cells, Wnt11 colocalized significantly with Fzd6 at the tip of cell protrusions (Figure 3.8B-C). Collectively, these results demonstrate that fibroblast exosomes promote autocrine Wnt ligand association with Fzd receptors and activation of PCP signalling during BCC protrusive activity and motility.

142

A ns 20

15 10

5 Cell Speed (nm/s) Relative mRNA Level Relative mRNA 0 DMEM Mock siScr siPorcn siScr siPorcn CM CM CM B Mock DMEM Mock ACM

15 1.5 ***

10 1.0

5 0.5 Cell Speed (nm/s) Relative mRNA Level Relative mRNA 0 0.0 Mock Mock siScr siPorcn siScr siPorcn DMEM ACM ACM ACM

siScr ACM siPorcn ACM C ***

Cell Speed (nm/s)

Mock Mock siScr siPorcn DMEM ACM ACM ACM

Figure 3.7 Porcupine Expression Is Necessary in BCCs

(A) The activity of conditioned media (CM) from L cells transfected with siRNA targeting Porcupine (Porcn) or a scrambled sequence (Scr), or from mock-transfected cells, was tested in MDA-MB-231 cells in parallel with control media (DMEM). Cell speed was quantified and plotted (left panel). n = 33 cells. ns stands for nonsignificant. Gene expression was determined as in Figure 3.1 (right panel). (B) Images of MDA-MB-231 cells transfected with the indicated siRNAs presented as in Figure 2.1 (left panel). Cell speed was quantified and plotted (middle

143

panel). n = 27 cells. *** p <0.0001. Gene expression was determined as in (A) (right panel). (C) The speed of SUM-159PT cells transfected with the indicated siRNAs and treated with ACM or DMEM was quantified and plotted. n = 31 cells per group. *** p <0.0001.

144

A Mock siScr

ns *** 15 1.5

10 1.0

5 0.5 Cell Speed (nm/s) Relative mRNA Level Relative mRNA 0 0.0 Mock siScr siWnt5a siWnt11 siScr siWnt5a siWnt11

siWnt5a siWnt11

B Fzd6 Fzd6 Wnt11 Actin Wnt11 Actin DMEM

Fzd6 Fzd6 Wnt11 Actin Wnt11 Actin ACM

C 1.5

1.0

0.5 Colocalization Coefficient 0.0 Protrusion Body Cortex1 Cortex2

Figure 3.8 BCC-produced Wnt11 Is Necessary For ACM Activity

(A) MDA-MB-231 cells transfected with siRNA targeting Wnt5a, Wnt11 or a scramble sequence (Scr), or under mock conditions were treated, imaged and presented as in Figure 2.1 (left panel). Cell speed was quantified (middle panel) and plotted. n = 30 cells per group. *** 145

p <0.0001. Wnt5a and Wnt11 gene expression was determined by RT-qPCR and is shown as expression levels relative to siScr cells (right panel). (B) MDA-MB-231 cells treated with DMEM or ACM were immunostained for endogenous Fzd6 (green (first column) and white (second column)) and endogenous Wnt11 (red (first column) and white (third column)). F- actin-rich cell protrusions were detected by phalloidin staining (blue (first column) and white (fourth column)). Fzd6 (green arrows) and Wnt11 (red arrows) colocalization at the leading edge of cell protrusions is indicated (yellow arrows). The inset shows a higher magnification of the indicated white, boxed area enclosing cell protrusions. Scale bar = 10 µm. (C) The colocalization coefficient between Wnt11 and Fzd6 voxels within ROIs defined in ACM stimulated cells as described in Figure 3.3B were calculated and the results plotted. n = 36 cells per group.

146

3.3.5. Fibroblast Exosomes Mobilize Autocrine Wnt-PCP Signalling in BCCs

Collectively, the data presented so far indicates that fibroblast Cd81-positive exosomes promote cell protrusive activity, motility and metastasis by activating Wnt-PCP signalling in recipient

BCCs. Next, I aimed to elucidate the molecular mechanisms whereby Cd81-positive exosomes activate autocrine Wnt-PCP signalling. First, I investigated whether cancer cells can produce

Cd81-positive exosomes. I isolated exosomes from media conditioned by similar number of

MDA-MB-231 cells and L cells and subjected them to SDS-PAGE and immunoblotting for

Cd81, Igsf8, Ptgfrn and Flotillin 1 (Figure 3.9A). I found that MDA-MB-231 cells secrete significantly less exosomes, which were particularly deficient for Cd81, when compared to L cells (Figure 3.9A). Notably, examination of Cd81 levels in total cell lysates revealed significant expression of this protein in MDA-MB-231 cells, suggesting that these cells are deficient in producing and/or secreting exosomes (Figure 3.9B). These results are in agreement with my aforementioned findings that both exosome secretion and Cd81 expression are important for activation of PCP signalling in BCCs.

147

A B MDA-MB-231 MCM 1-3d Cells:MDA-MB-231 L cell exosomes Day: 1 2 3 1 2 3 Cd81 IB:Cd81 ACM Lcell 1-3d Ponceau exosomes

Total Cell Lysate Pellet Supernatant CM: MCM ACM MCM ACM Day: 1 2 3 1 2 3 * 1 2 3 1 2 3 Cd81 IB:-Cd81 Igsf8 IB:-Igsf8 Ptgfrn IB:-Ptgfrn Flot.1 IB:-Flot.1 Conditioned Media

Figure 3.9 BCCs Are Deficient For Cd81-positive Exosomes

(A) Conditioned media prepared from MDA-MB-231 cells (MCM) or L cells (ACM) incubated with DMEM for one, two or three days (1-3d) as indicated (top panel) were subjected to ultracentrifugation to isolate exosomes prior to immunoblotting (IB) the pellet and the supernatant for the indicated exosome markers (bottom panel; * lane contained molecular weight markers). (B) Total cell lysates from MDA-MB-231 cells or L cells in (A) were IB for Cd81. Ponceau staining was used to detect total protein loading.

148

My next goal was to determine how fibroblast exosomes interact with BCCs to activate autocrine Wnt11-PCP signalling. Currently, it is not clear how exosomes deliver their cargo in target cells. However, evidence suggests that exosomes can be retained on the cell surface

(Segura et al., 2007), can fuse with the plasma membrane (Parolini et al., 2009), or are internalized by recipient cells (Fitzner et al., 2011; Morelli et al., 2004). To examine these possibilities, I generated L cells stably expressing a Cd81-EYFP fusion protein, which was secreted as a component of exosomes (Figure 3.10A). Then, I treated MDA-MB-231 cells with conditioned media (CM) from L cells expressing either a control vector or Cd81-EYFP (Figure

3.10B). Immunoprecipitation of Cd81-EYFP from total lysates of recipient MDA-MB-231 cells revealed that Cd81-EYFP associated with these cells (Figure 3.10B). Also, immunofluorescence studies of Cd81-EYFP in recipient MDA-MB-231 cells revealed that

Cd81-EYFP accumulated in vesicular structures, where it colocalized with endogeneous Wnt11

(Figure 3.10C). Furthermore, inhibition of endocytosis by Dynasore, a potent and specific inhibitor of dynamin (Macia et al., 2006), impeded the accumulation of Cd81-EYFP in recipient

MDA-MB-231 cells (Figure 3.10D). Altogether, there results show that fibroblast exosomes are endocytosed by BCCs and that exogenous Cd81 colocalizes with endogenous Wnt11 in vesicular structures in BCCs.

149

A B Cd81-EYFP Cd81-EYFP MDA-MB-231 _CM _CM Cd81-EYFP_L cell 20h 3d 100,000 x g pellet + IB + IF (exosomes)

105 x g 105 x g Pellet Sup. Total pLVX_CM: + - + - - CM: ACM pLVX Cd81-EYFP Cd81-EYFP_CM: - + - + + CM: ACM pLVX Cd81-EYFP Cd81- IP: GFP Cd81- Cd81- IB:GFP IB: GFP EYFP IB: GFP EYFP EYFP MDA-MB-231 Total CM Conditioned Media Total Lysate C D pLVX_CM Cd81-EYFP_CM Cd81-EYFP_CM + DMSO + DMSO + Dynasore Cd81-EYFP Cd81-EYFP Wnt11 Cd81-EYFP Cd81-EYFP Cd81-EYFP Wnt11 Wnt11 Wnt11 Wnt11 DAPI DAPI DAPI DAPI pLVX_CM

Cd81-EYFP Cd81-EYFP Wnt11 Cd81-EYFP Cd81-EYFP Cd81-EYFP Wnt11 DAPI Cd81-EYFP_CM

3 p = 0.0001

1 (Artificial unit) of EYFP signal of EYFP Relative intensity 0 DMSO Dynasore

Figure 3.10 MDA-MB-231 Cells Internalize L cell-secreted Exosomes

(A) Exosomes isolated from conditioned media (CM) from L cells stably expressing Cd81- EYFP (Cd81-EYFP_CM) or pLVX vector (pLVX_CM) were subjected to immunoblotting (IB) with an anti-GFP antibody. (B) BCCs were treated for 20h with CM prepared in (A) (top panel). Total CM from L cells (bottom left panel) or anti-GFP immunoprecipitates from BCCs (bottom right panel) were IB for GFP. (C) Confocal images of BCCs treated as in (B) and stained for GFP (green (left panels) and white (middle panels)) and Wnt11 (red (left panel) and white (right panels)) and counterstained for DAPI (blue (left panels)). Scale bar = 10 µm. (D) Confocal images of BCCs treated for 16 h with the indicated CM in the presence of 100 mM Dynasore or DMSO (control) and immunostained as in (C). The GFP signal was measured 150

(bottom panel; mean intensity signal ± SEM. n= 30 cells per group. p = 0.0001 using two tailed unpaired t-test).

151

In Wnt producing cells, Wnts are trafficked via endosomes from the plasma membrane in a process that is thought to be necessary for ligand maturation and long range signalling

(Coudreuse and Korswagen, 2007; Pfeiffer et al., 2002). Importantly, exosomes originate from the endocytic system (Babst, 2011; Stoorvogel et al., 2002). Moreover, Wnt ligands are secreted in an exosome-tethered form by Wnt producing cells (Beckett et al., 2012; Gross et al.,

2012; Koles et al., 2012). My results indicated that exogenous Cd81-EYFP colocalizes with endogenous Wnt11 in vesicular structures in the perinuclear region of the cell (Figure 3.10C), consistent with previously reported subcellular localization of the ligand (Coudreuse and

Korswagen, 2007). Hence, I sought to determine whether exogenous exosomes access Wnts within the endocytic system of BCCs and are secreted back by recipient cells in association with the ligand. For this, I generated MDA-MB-231 cells stably expressing Wnt11 tagged with HA at the C-terminus, which enabled detection of the ligand. Tagging of the ligand at the C- terminus does not impede its secretion or function (Herr and Basler, 2012). Then, I treated

Wnt11-HA-expressing MDA-MB-231 cells with either control (DMEM) or ACM media for twenty hours to allow exosome trafficking in recipient cells (Figure 3.11A). I found that ACM treatment did not affect the steady state levels of Wnt11-HA in total lysates of MDA-MB-231 cells, suggesting that ACM does not affect ligand secretion or turnover (Figure 3.10A). Next, I subjected the media conditioned by MDA-MB-231 cells to differential ultracentrifugation and immunoblotting for Wnt11-HA and Cd81 (Figure 3.11B). This revealed that Wnt11-HA associated with the exosome pellet isolated from ACM that was conditioned by Wnt11-HA- expressing MDA-MB-231 cells (Figure 3.11B). In contrast, Wnt11-HA was not observed in the

100,000 x g pellet from DMEM that was conditioned by Wnt11-HA-expressing MDA-MB-231 cells. Of note, Wnt11-HA was not detected in total conditioned media, probably due to a weak signal (Figure 3.11B). Furthermore, immunogold labelling and EM confirmed that Wnt11-HA

152

was tethered to exosomes isolated from MDA-MB-231-conditioned ACM (Figure 3.11C), whereas no ligand was detected in exosomes isolated from ACM or MDA-MB-231-conditioned

DMEM. Since the ligand was not detected in total media (Figure 3.11B), I wondered whether

Wnt11-HA was secreted in DMEM conditioned by MDA-MB-231 cells. To test this, I treated parental, non-Wnt11-HA-expressing, MDA-MB-231 cells with DMEM or ACM that was conditioned by Wnt11-HA-expressing cells (Figure 3.11C). Then, I analysed parental cell lysates for internalized Wnt11-HA and found that similar amounts of Wnt11-HA were internalized in cells treated with either conditioned DMEM or ACM (Figure 3.11D). This indicates that Wnt11-HA is present in conditioned control media and ligand secretion and internalization by BCCs is not dependent on exosomes. Since Wnt11-HA was secreted in conditioned DMEM, I next asked whether Wnt11 might associate with exosomes in the culture media in the absence of cellular trafficking. For this, I incubated DMEM that was conditioned by Wnt11-HA overexpressing cells with naïve ACM in the absence of cells (Figure 3.11E).

Then, I isolated the exosomes either by ultracentrifugation or anti-Cd81 immunoprecipitation and immunoblotted for Wnt11-HA (Figure 3.11E). I found that when ACM was simply mixed with conditioned DMEM, little or no Wnt11-HA associated with exosomes, despite their efficient purification under these conditions (Figure 3.11E). In contrast, the ligand associated with exosomes when ACM was directly conditioned by Wnt11-HA-expressing MDA-MB-231 cells (Figure 3.11E). Altogether, my data support a model in which fibroblast exosomes are internalized by BCCs via the endocytic system where they are loaded with Wnt11 and then recycled back to the extracellular milieu to activate PCP signalling, which drives BCC motility and metastasis (Figure 3.12).

153

A Wnt11-HA _DMEM DMEM exosomes IB / cDNA: pCAGIP Wnt11-HA Wnt11-HA_ 5 (10 x g pellet) EM Media: D A D A MDA-MB-231 20h + Wnt11-HA IB:-HA lyse cells IB + 20h Ponceau exosomes IB / 5 Total Cell Lysate ACM (10 x g pellet) EM Wnt11-HA _ACM

B Orig. 104 x g 105 x g 105 x g C ACM pellet pellet sup. Media: A D A A D A A D A A D A Wnt11-HA: - + + - + + - + + - + + Wnt11-HA IB:-HA

Cd81 IB:-Cd81 50nm 50nm 50nm Conditioned Media D E Wnt11-HA Wnt11-HA Wnt11-HA_ MDA-MB-231 _DMEM/ACM _DMEM1 ACM2 DMEM1 + ACM2 Total exosomes 20h lyse cells + + IP:HA IP Cd81 +ve exosomes Media: p_D p_A W_D W_A Wnt11-HA Wnt11-HA_ _ACM1 DMEM2 ACM1 + DMEM2 Total Wnt11-HA IB:-HA exosomes + Cd81 +ve IP: -HA IP exosomes

Media1: p_D p_A W_D W_A Media1: p_D p_A W_D W_A Media2: A D A D Media2: A D A D

Wnt11-HA IB:-HA Wnt11-HA IB:-HA

Cd81 IB:-Cd81 Cd81 IB:-Cd81

Total Exosomes IP:-Cd81

Figure 3.11 Wnt11 Associates With L cell Exosomes Upon Trafficking In BCCs

(A) MDA-MB-231 BCCs stably expressing an empty vector (pCAGIP) or Wnt11-HA were incubated with DMEM (D) or ACM (A) for 20 h to produce BCC-conditioned media, pCAGIP_DMEM/ACM and Wnt11-HA_DMEM/ACM (indicated in yellow, left panel). Then, BCC total lysates were subjected to immunobloting (IB) with an anti-HA antibody (right panel). Exosomes isolated from the media produced as in (A) were subjected to IB with the indicated antibodies (B) or EM with an anti-Wnt11 antibody (C). (D) Media conditioned by pCAGIP (p)- or Wnt11-HA (W)-expressing BCCs as in (A) was used to treat wild type BCCs (top panel). 154

Total lysates of the recipient cells were subjected to anti-HA immunoprecipitation (IP) and IB (bottom panel). (E) Media conditioned as described in (A) (Wnt11-HA_DMEM1 or Wnt11- HA_ACM1, yellow in schematic, top panel) was mixed with an equal volume of original ACM (ACM2) or DMEM (DMEM2), to produce Wnt11-HA_DMEM1+ACM2 and Wnt11- HA_ACM1+DMEM2 respectively (orange in schematic, top panel). Exosomes were purified by ultracentrifugation (bottom left panel) or immunoprecipitated (IP) using an anti-Cd81 antibody (bottom right panel) and IB as in (B).

155

Fibroblast

Motile/Metastatic BCC BCC

Cell Protrusions PCP Cell Motility Metastatic Potential

Fibroblast Cd81 Fzd6 Wnt11 Vangl1 Pk1 Fibroblast/BCC exosome exosome (Cd81 positive) (Cd81/Wnt11 positive)

Figure 3.12 A Model For Fibroblast Exosomes Stimulating Wnt-PCP Signalling In BCCs

Fibroblasts secrete Cd81-positive exosomes, which are endocytosed by BCCs. BCC-produced Wnt11 tethers to fibroblast exosomes within the endocytic system of BCCs. Exosome-tethered Wnt11 is secreted by BCCs and interacts with Fzd6 at the leading edge of cell protrusions to activate core PCP signalling, which drives BCC cell protrusive activity, motility and metastasis.

156

3.4. Discussion

The Wnt-PCP pathway is required for convergent extension movements during vertebrate gastrulation and neurulation. However, it is unclear whether the core PCP pathway is involved in cancer cell motility and metastasis. Here, I demonstrate that the expression of Fzd, Dvl,

Vangl and Pk, is necessary for MDA-MB-231 and SUM-159PT cell protrusive activity and motility. Hence, my data suggests that the function of the core PCP pathway components in cell motility is conserved in cancer cells. Interestingly, silencing of two other Wnt receptors, Ryk and Ror2, which regulate vertebrate PCP (Andre et al., 2012; Gao et al., 2011), resulted in enhancement of ACM activity or no effect respectively. In agreement with this, I found that silencing of Wnt5a, which serves as a ligand for both Ryk and Ror2 (Andre et al., 2012; Gao et al., 2011), did not affect BCC motility in response to ACM stimulation. Altogether, these results suggest that the core PCP signalling pathway is distinct from the Wnt5a-Ryk/Ror2 pathways. Next, I reveal that Pk1 expression is required for fibroblast-induced BCC metastasis but is not necessary for primary tumour growth. This is consistent with the role of Pk1 in BCC migration, which is a critical for cancer metastasis. Importantly, overexpression of Vangl1 in colon cancer cells or squamous cancer cells results in stimulation of cell motility and metastasis

(Lee et al., 2004; Lee et al., 2009). In the future, it will be of interest to examine the role of the other core PCP components in cancer cell metastasis.

The mechanisms involved in effectuating core PCP signalling in vertebrates are poorly defined.

The polarized subcellular distribution of core PCP components seems to be conserved from the fruit fly to vertebrates (Ciruna et al., 2006; Devenport and Fuchs, 2008; Narimatsu et al., 2009;

Yin et al., 2008). However, this has been challenging to demonstrate in vertebrates due to lack

157

of antibodies and the dynamic nature of PCP-regulated processes. Here, I present data that suggests that Fzd-Dvl and Vangl-Pk complexes are asymmetrically distributed in exosome- stimulated BCC protrusions in a manner analogous to planar-polarized cells in the fly wing. To the best of my knowledge, this is the first time that asymmetrical distribution of core PCP components has been reported in single cancerous cells. The mechanism through which Fzd accumulates in the tip of ACM-induced BCC cells is unknown. However, it is possible that transport of Fzd along MTs contributes to Fzd enrichment at membrane protrusions similarly to what has been observed in the fly wing epithelium (Shimada et al., 2006). Next, it will important to determine whether Vangl-Pk complexes are excluded from the Fzd-Dvl complexes at the tip of exosome-stimulated cell protrusion in a similar manner to that observed in the fly wing. Indeed, we have already shown that upon Wnt stimulation, phosphorylated Dvl in association with Par6 recruit Smurf ubiquitin ligases, which target Prickle for proteasomal degradation (Narimatsu et al., 2009).

Genetic studies have revealed that Wnt ligands, mainly Wnt5a and Wnt11, play a critical role in vertebrate PCP (Gros et al., 2009; Heisenberg et al., 2000; Qian et al., 2007; Tada and Smith,

2000). However, the mechanisms involved are poorly characterized. Here, I first demonstrate that the expression of Wnt11, but not Wnt5a, in BCCs is necessary for ACM-stimulated BCC protrusive activity and motility. Next, I show that ACM induces the association of Wnt11 with

Fzd6 at the tip of BCC protrusions. Furthermore, I reveal that fibroblast exosomes are internalized by BCCs, where Cd81 colocalizes with Wnt11 in vesicular structures. Our preliminary data suggests that these vesicles are late endosomes, since exogenous Cd81 and

Wnt11 colocalize with Rab7, a late endosome marker, in BCCs. Moreover, I show by

158

biochemistry and EM that Wnt11 tethers to Cd81-positive exosomes upon trafficking of the latter in BCCs. Importantly, my data suggests that fibroblast exosomes do not affect Wnt ligand secretion by Wnt-producing BCCs or Wnt internalization by recipient BCCs. Instead, my data indicates that fibroblast exosomes mobilize autocrine Wnts in BCCs and promote Wnt association with Fzd receptors, thus activating Wnt-PCP signalling during BCC protrusive activity and motility. This is the first time that such a mechanism of Wnt mobilization and function has been proposed. My findings have several implications in the fields of PCP signalling, Wnt trafficking and exosome function and I will consider them further in Chapter 4.

159

Chapter 4 : Thesis Summary and Future Directions

160

4.1. Overview

The cross-talk between cancer cells and associated stromal cells is critical for tumour progression (Egeblad et al., 2010; Kalluri and Zeisberg, 2006). However, stromal signals that specifically promote cancer cell metastasis are poorly defined. Here, I presented my findings, which indicate that fibroblasts secrete Cd81-positive exosomes that promote BCC motility and metastasis by mobilizing autocrine Wnt-PCP signalling in tumour cells. First, I found that L cells secrete exosomes, which have a potent stimulatory effect on the protrusive activity and motility of several BCCs. Next, I identified Cd81 as a component of L cell-secreted exosomes that is necessary for stimulation of BCC motility. Further, I established that Cd81 specifically mediates L cell-stimulation of BCC metastasis because silencing of Cd81 led to suppression of metastases but had no effect on the growth of the primary tumour. Importantly, I demonstrate that primary human CAFs also secrete stimulatory Cd81-positive exosomes and that Cd81 expression is upregulated in stroma associated with human breast invasive carcinoma.

Altogether, these results indicate that fibroblasts, including CAFs, secrete Cd81-positive exosomes that stimulate cancer cell migration during metastasis.

Several PCP components, including Fzd, Dvl and Vangl, have been reported to regulate cancer cell motility and invasion (Jessen, 2009). However, the role of PCP signalling in cancer has not been elucidated. Here, I provide evidence that the Wnt-PCP pathway promotes BCC metastasis.

Firstly, all the core PCP proteins, which I tested, were necessary for ACM-induced BCC protrusive activity and motility. Secondly, silencing of Pk1 expression in BCCs led to a strong inhibition of L cell-stimulated BCC metastasis but had no effect on primary tumour growth.

Thirdly, I reveal that Fzd-Dvl and Vangl-Pk complexes display asymmetrical subcellular

161

distribution in BCC protrusions stimulated with ACM. Fourthly, I show that CAF exosomes promote BCC motility in a Pk1-dependent manner and Fzd-Dvl complexes accumulate at the tip of BCC protrusions upon stimulation with CAF exosomes. Lastly, I demonstrate that the expression of Wnt11 is essential for the migratory potential of ACM-treated BCCs and that

Wnt11 associates with Fzd receptors at the tip of cell protrusions upon ACM-treatment. Based on these results, I propose that Wnt-PCP signalling regulates BCC protrusive activity and motility in a similar manner to its role in CE movements during vertebrate development.

Wnt ligands are tightly associated with the plasma membrane due to posttranslational modifications, which render them highly hydrophobic and insoluble in the extracellular milieu

(Coudreuse and Korswagen, 2007; Willert and Nusse, 2012). Several factors have been shown to mobilize Wnts and assist their long range signalling (Willert and Nusse, 2012). For instance, exosomes produced by Wnt-secreting cells carry Wnts in the extracellular milieu (Beckett et al.,

2012; Gross et al., 2012; Ratajczak et al., 2006). Here, I present data that suggests that exosomes from fibroblasts mobilize Wnt ligands produced by BCCs upon their trafficking in

BCCs. Wnts are palmitoylated, in a manner that is dependent on Porcupine, and this modification is critical for their secretion and function (Willert and Nusse, 2012). Accordingly,

I found that silencing of Porcupine expression in BCCs inhibited ACM-induced BCC protrusive activity and motility. Next, I found that fibroblast-secreted Cd81 is internalized by BCCs in a dynamin-dependent manner and colocalizes with BCC-produced Wnt11 in vesicular structures.

Further, I demonstrated that autocrine Wnt11 associates with Cd81-positive exosomes upon trafficking in BCCs. Altogether, my findings indicate that autocrine Wnt ligands that are

162

tethered to fibroblast-secreted exosomes within endocytic vesicles of recipient BCCs activate

Wnt-PCP signalling in cancer cell motility and metastasis.

4.2. Unanswered Questions and Future Directions

My findings indicate that fibroblast exosomes stimulate BCC metastasis. Although it is technically challenging, the direct effect of exosomes in metastasis has yet to be shown by injecting purified CAF exosomes in mice harbouring tumours. Additionally, it will be important to establish whether exosome secretion from fibroblasts is relevant in human disease. To do so, first, it will be necessary to identify specific markers for fibroblast exosomes in order to distinguish them from tumour exosomes. Proteomic analysis of L cell exosomes revealed a number of previously reported exosome components as well as several novel ones (Figure 2.5C and 2.6C), which may serve as markers for the identification of CAF exosomes. Then, exosomes from cancer patients, which could be isolated from bodily fluids, including blood, saliva and urine, can be used to determine whether the concentration of exosomes correlates with disease progression. Importantly, the concentration of tumour exosomes in sera from human patients correlates with disease progression (Peinado et al., 2012; Taylor and Gercel-

Taylor, 2008). Next, it will be interesting to determine whether Cd81 protein levels in CAF exosomes correlate with disease stage. Our analysis of published gene expression data of carcinoma-associated stroma (Finak et al., 2006), indicates that Cd81 expression is upregulated in human CAFs (Figure 2.9C). Thus, it is likely that Cd81 levels are elevated in CAF exosomes, which may serve as diagnostic biomarkers for cancer detection. Finally, another line of investigation would be the analysis of exosomes from CAFs associated with other types of cancers.

163

My data indicates that Cd81 functions as a component of exosomes to regulate BCC cell motility and metastasis. However, what is the mechanism through which Cd81 functions?

Tetraspanins have been proposed to mediate exosome adhesion to target cells ((Nazarenko et al.,

2010). Thus, it is possible that Cd81 is involved in exosome binding to BCCs and/or internalization by BCCs. In addition, tetraspanins are thought to regulate sorting of cargo molecules in exosomes (Stoorvogel et al., 2002; Wubbolts et al., 2003). Thus, it is possible that

Cd81 functions inside L cells where it recruits a second exosome component, which may have a more direct effect in BCC motility. Interestingly, Cd81 interacts with and regulates the activity of Adam10 metalloprotease within the context of tetraspanin-enriched microdomains (TEMs) in plasma membranes (Arduise et al., 2008; Xu et al., 2009). My preliminary work shows that

Adam10 is a component of L cell exosomes and is required for ACM activity. Thus, it will be of interest to study the relationship between Cd81 and Adam10 as active components of L cell exosomes. Alternatively, Cd81 could function in BCCs upon exosome internalization and assimilation of Cd81 in the membrane of recipient cells. Within the context of TEMs in the plasma membrane, tetraspanins associate with multiple proteins, several of which regulate the actin cytoskeleton (Zoller, 2009) (Figure 1.9). Furthermore, Cd82 interacts with both Vangl1

(Lee et al., 2004) and Cd81 within the context of a cell. Thus, it may be that Cd81 affects Vangl activity by regulating its distribution in TEMs. Finally, Cd81 could also regulate sorting of palmitoylated and/or GPI-linked Wnts in exosomes within MVBs of BCCs. Indeed, my preliminary data suggests that fibroblast Cd81 is critical for association of BCC-produced

Wnt11 with exosomes. Importantly, I found that Cd81 has a specific role in stimulating BCC motility, since Cd63 and Cd82, which are highly enriched in exosomes, were not necessary for

ACM activity (Figure 2.7B). Next, it will be of interest to determine whether Cd9, which is 164

most closely related to Cd81, is also important for ACM activity. Then, comparison of Cd81 with Cd9, Cd63 and Cd82 might provide important clues as to what determinant regulates Cd81 function in exosomes.

Here, I demonstrated that the core PCP pathway regulates BCC protrusive activity and motility.

However, the mechanism(s) that mediate core PCP signalling are poorly understood. In vertebrates, this problem is further complicated by the existence of several new components that regulate different PCP processes. We previously demonstrated that Smurf ubiquitin ligases regulate PCP and CE movements in mice by associating with Dvl that is in complex with Par6 and subsequently targeting Pk for degradation (Narimatsu et al., 2009) (Figure 1.7). In addition,

Smurfs recruited at the tips of cell protrusions by the Par6 polarity complex, regulate cancer cell migratory and invasive phenotype by targeting RhoA for degradation (Sahai et al., 2007;

Viloria-Petit et al., 2009; Wang et al., 2003) (Figure 1.4). I similarly demonstrated that Smurfs are required for ACM-stimulated BCC protrusive activity and motility (Figure 3.1A).

Moreover, my preliminary data suggests that overexpressed Par6 enhances MDA-MB-231 cell motility. Importantly, the Par6/Par3/aPKC polarity complex regulates apicobasal polarity of epithelial cells (Ozdamar et al., 2005; Thiery et al., 2009) and front-rear polarity of migratory cells (Etienne-Manneville, 2008). Therefore, I propose that the Smurf-Par6 complex also cooperates with the core PCP pathway to regulate cancer cell motility and metastasis. In the future, it will be of interest to determine whether exosomes also promote the association of

Smurf with Par6 and degradation of Prickle. It will also be of interest to determine whether the core PCP pathway regulates the motility of other cancer cell types.

165

I also established that Wnt11 is tethered onto fibroblast exosomes upon their trafficking in

BCCs, but how and where does Wnt11 associate with exosomes within cells? Currently, it is unclear how exosomes are internalized and what their fate is upon internalization. My work suggests that endocytosis of exosomes in BCCs is important (Figure 3.10D) and that Wnt colocalizes with Cd81 (Figure 3.10C) and Rab7, a marker for MVBs, in vesicular structures

(preliminary data). It is known that Wnts localize to MVBs and recycling endosomes within

Wnt-producing cells, and this has been suggested to be necessary for Wnt maturation and signalling (Coudreuse and Korswagen, 2007). Thus, I propose that Wnts associate with fibroblast exosomes in MVBs of BCCs where Wnt11 integrates with fibroblast exosomes, which are then secreted to activate Wnt-PCP signalling (Figure 3.12). It will be of interest to test this hypothesis by interfering with the function of the endocytic system components, including ESCRT and Rab proteins, which regulate exosome biogenesis and secretion respectively (Babst, 2011; Bobrie et al., 2011). Related to this, determining what Wnt- associated factors regulate Wnt targeting to exosomes will be of interest. A common theme is emerging that is Wnts are trafficked within cells in association with transmembrane proteins. In particular, Wnts are retained in the ER by the transmembrane proteins Porcupine and Oto, which are involved in Wnt palmitoylation (Takada et al., 2006) and GPI-linkage (Zoltewicz et al.,

2009) respectively, and are trafficked from the ER to Golgi by two p24 transmembrane family proteins, Éclair and Emp24 (Port et al., 2011), which also act as cargo receptors for GPI-linked proteins, and from Golgi to the plasma membrane by the seven-pass transmembrane protein

Wntless (Banziger et al., 2006; Bartscherer et al., 2006; Goodman et al., 2006). I found that

Porcupine is essential for ACM activity (Figure 3.7B-C) and it has been reported that

Porcupine-mediated acylation of Wnts is necessary for ligand association with lipid rafts (Zhai et al., 2004), which, like TEMs, are also enriched in cholesterol and sphingomyelin. Thus, it is

166

possible that Porcupine and other Wnt-associated factors regulate Wnt targeting to exosomes.

Moreover, it is possible that Wnts require other transmembrane proteins, such as Cd81, to be sorted in exosome membranes, as suggested above.

In conclusion, my studies reveal that fibroblast exosomes play a key role in stimulating autocrine Wnt-PCP signalling in breast cancer cells to drive cell motility and metastasis. My findings have several important implications in the fields of tumour-stroma cross-talk, PCP signalling in cancer, Wnt mobilization and exosome-mediated intercellular communication.

Future testing of the hypotheses derived from my studies may lead to a more thorough understanding of tumour-stroma signalling in metastasis and may bring us closer to developing targeted therapeutics for the treatment of cancer.

167

References

Abounit, S., and Zurzolo, C. (2012). Wiring through tunneling nanotubes--from electrical signals to organelle transfer. J Cell Sci 125, 1089-1098.

Al-Hajj, M., Wicha, M.S., Benito-Hernandez, A., Morrison, S.J., and Clarke, M.F. (2003). Prospective identification of tumorigenic breast cancer cells. Proc Natl Acad Sci U S A 100, 3983-3988.

Al-Nedawi, K., Meehan, B., Kerbel, R.S., Allison, A.C., and Rak, J. (2009a). Endothelial expression of autocrine VEGF upon the uptake of tumor-derived microvesicles containing oncogenic EGFR. Proc Natl Acad Sci U S A 106, 3794-3799.

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

Alvarez-Erviti, L., Seow, Y., Yin, H., Betts, C., Lakhal, S., and Wood, M.J. (2011). Delivery of siRNA to the mouse brain by systemic injection of targeted exosomes. Nat Biotechnol 29, 341- 345.

Anastas, J.N., and Moon, R.T. (2013). WNT signalling pathways as therapeutic targets in cancer. Nat Rev Cancer 13, 11-26.

Andre, F., Schartz, N.E., Movassagh, M., Flament, C., Pautier, P., Morice, P., Pomel, C., Lhomme, C., Escudier, B., Le Chevalier, T., et al. (2002). Malignant effusions and immunogenic tumour-derived exosomes. Lancet 360, 295-305.

Andre, P., Wang, Q., Wang, N., Gao, B., Schilit, A., Halford, M.M., Stacker, S.A., Zhang, X., and Yang, Y. (2012). The wnt coreceptor ryk regulates wnt/planar cell polarity by modulating the degradation of the core planar cell polarity component vangl2. J Biol Chem 287, 44518- 44525.

Angers, S., and Moon, R.T. (2009). Proximal events in Wnt signal transduction. Nat Rev Mol Cell Biol 10, 468-477.

Antic, D., Stubbs, J.L., Suyama, K., Kintner, C., Scott, M.P., and Axelrod, J.D. (2010). Planar cell polarity enables posterior localization of nodal cilia and left-right axis determination during mouse and Xenopus embryogenesis. PLoS One 5, e8999.

Arduise, C., Abache, T., Li, L., Billard, M., Chabanon, A., Ludwig, A., Mauduit, P., Boucheix, C., Rubinstein, E., and Le Naour, F. (2008). Tetraspanins regulate ADAM10-mediated cleavage of TNF-alpha and epidermal growth factor. J Immunol 181, 7002-7013. 168

Attisano, L., and Wrana, J.L. (2002). Signal transduction by the TGF-beta superfamily. Science 296, 1646-1647.

Babst, M. (2011). MVB vesicle formation: ESCRT-dependent, ESCRT-independent and everything in between. Current opinion in cell biology 23, 452-457.

Baeg, G.H., Lin, X., Khare, N., Baumgartner, S., and Perrimon, N. (2001). Heparan sulfate proteoglycans are critical for the organization of the extracellular distribution of Wingless. Development 128, 87-94.

Baietti, M.F., Zhang, Z., Mortier, E., Melchior, A., Degeest, G., Geeraerts, A., Ivarsson, Y., Depoortere, F., Coomans, C., Vermeiren, E., et al. (2012). Syndecan-syntenin-ALIX regulates the biogenesis of exosomes. Nature cell biology 14, 677-685.

Balaj, L., Lessard, R., Dai, L., Cho, Y.J., Pomeroy, S.L., Breakefield, X.O., and Skog, J. (2011). Tumour microvesicles contain retrotransposon elements and amplified oncogene sequences. Nat Commun 2, 180.

Banziger, C., Soldini, D., Schutt, C., Zipperlen, P., Hausmann, G., and Basler, K. (2006). Wntless, a conserved membrane protein dedicated to the secretion of Wnt proteins from signaling cells. Cell 125, 509-522.

Barrios-Rodiles, M., Brown, K.R., Ozdamar, B., Bose, R., Liu, Z., Donovan, R.S., Shinjo, F., Liu, Y., Dembowy, J., Taylor, I.W., et al. (2005). High-throughput mapping of a dynamic signaling network in mammalian cells. Science 307, 1621-1625.

Bartscherer, K., and Boutros, M. (2008). Regulation of Wnt protein secretion and its role in gradient formation. EMBO Rep 9, 977-982.

Bartscherer, K., Pelte, N., Ingelfinger, D., and Boutros, M. (2006). Secretion of Wnt ligands requires Evi, a conserved transmembrane protein. Cell 125, 523-533.

Bayly, R., and Axelrod, J.D. (2011). Pointing in the right direction: new developments in the field of planar cell polarity. Nature reviews Genetics 12, 385-391.

Beckett, K., Monier, S., Palmer, L., Alexandre, C., Green, H., Bonneil, E., Raposo, G., Thibault, P., Borgne, R.L., and Vincent, J.P. (2012). Drosophila S2 cells secrete Wingless on exosome- like vesicles but the Wingless gradient forms independently of exosomes. Traffic.

Bhowmick, N.A., Chytil, A., Plieth, D., Gorska, A.E., Dumont, N., Shappell, S., Washington, M.K., Neilson, E.G., and Moses, H.L. (2004). TGF-beta signaling in fibroblasts modulates the oncogenic potential of adjacent epithelia. Science 303, 848-851.

169

Bianco, N.R., Kim, S.H., Ruffner, M.A., and Robbins, P.D. (2009). Therapeutic effect of exosomes from indoleamine 2,3-dioxygenase-positive dendritic cells in collagen-induced arthritis and delayed-type hypersensitivity disease models. Arthritis Rheum 60, 380-389.

Bissell, M.J., and Hines, W.C. (2011). Why don't we get more cancer? A proposed role of the microenvironment in restraining cancer progression. Nature medicine 17, 320-329.

Blair, S.S. (2012). Cell polarity: overdosing on PCPs. Curr Biol 22, R567-569.

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

Borovina, A., Superina, S., Voskas, D., and Ciruna, B. (2010). Vangl2 directs the posterior tilting and asymmetric localization of motile primary cilia. Nat Cell Biol 12, 407-412.

Bosveld, F., Bonnet, I., Guirao, B., Tlili, S., Wang, Z., Petitalot, A., Marchand, R., Bardet, P.L., Marcq, P., Graner, F., et al. (2012). Mechanical control of morphogenesis by Fat/Dachsous/Four-jointed planar cell polarity pathway. Science 336, 724-727.

Brabletz, T. (2012). To differentiate or not--routes towards metastasis. Nat Rev Cancer 12, 425- 436.

Brazzoli, M., Bianchi, A., Filippini, S., Weiner, A., Zhu, Q., Pizza, M., and Crotta, S. (2008). CD81 is a central regulator of cellular events required for hepatitis C virus infection of human hepatocytes. Journal of virology 82, 8316-8329.

Buijs, J.T., van der Horst, G., van den Hoogen, C., Cheung, H., de Rooij, B., Kroon, J., Petersen, M., van Overveld, P.G., Pelger, R.C., and van der Pluijm, G. (2012). The BMP2/7 heterodimer inhibits the human breast cancer stem cell subpopulation and bone metastases formation. Oncogene 31, 2164-2174.

Campbell, P.J. (2012). Telomeres and cancer: from crisis to stability to crisis to stability. Cell 148, 633-635.

Campellone, K.G., and Welch, M.D. (2010). A nucleator arms race: cellular control of actin assembly. Nat Rev Mol Cell Biol 11, 237-251.

Carreira-Barbosa, F., Concha, M.L., Takeuchi, M., Ueno, N., Wilson, S.W., and Tada, M. (2003). Prickle 1 regulates cell movements during gastrulation and neuronal migration in zebrafish. Development 130, 4037-4046.

Cha, S.W., Tadjuidje, E., White, J., Wells, J., Mayhew, C., Wylie, C., and Heasman, J. (2009). Wnt11/5a complex formation caused by tyrosine sulfation increases canonical signaling activity. Curr Biol 19, 1573-1580. 170

Charrin, S., le Naour, F., Silvie, O., Milhiet, P.E., Boucheix, C., and Rubinstein, E. (2009a). Lateral organization of membrane proteins: tetraspanins spin their web. Biochem J 420, 133- 154.

Charrin, S., Yalaoui, S., Bartosch, B., Cocquerel, L., Franetich, J.F., Boucheix, C., Mazier, D., Rubinstein, E., and Silvie, O. (2009b). The Ig domain protein CD9P-1 down-regulates CD81 ability to support Plasmodium yoelii infection. J Biol Chem 284, 31572-31578.

Cherukuri, A., Carter, R.H., Brooks, S., Bornmann, W., Finn, R., Dowd, C.S., and Pierce, S.K. (2004). B cell signaling is regulated by induced palmitoylation of CD81. J Biol Chem 279, 31973-31982.

Chevallet, M., Luche, S., and Rabilloud, T. (2006). Silver staining of proteins in polyacrylamide gels. Nat Protoc 1, 1852-1858.

Ciruna, B., Jenny, A., Lee, D., Mlodzik, M., and Schier, A.F. (2006). Planar cell polarity signalling couples cell division and morphogenesis during neurulation. Nature 439, 220-224.

Clark, K.L., Oelke, A., Johnson, M.E., Eilert, K.D., Simpson, P.C., and Todd, S.C. (2004). CD81 associates with 14-3-3 in a redox-regulated palmitoylation-dependent manner. J Biol Chem 279, 19401-19406.

Clayton, A., Harris, C.L., Court, J., Mason, M.D., and Morgan, B.P. (2003). Antigen-presenting cell exosomes are protected from complement-mediated lysis by expression of CD55 and CD59. Eur J Immunol 33, 522-531.

Clevers, H. (2011). The cancer stem cell: premises, promises and challenges. Nat Med 17, 313- 319.

Coffey, G.P., Rajapaksa, R., Liu, R., Sharpe, O., Kuo, C.C., Krauss, S.W., Sagi, Y., Davis, R.E., Staudt, L.M., Sharman, J.P., et al. (2009). Engagement of CD81 induces ezrin tyrosine phosphorylation and its cellular redistribution with filamentous actin. J Cell Sci 122, 3137-3144.

Collado, M., and Serrano, M. (2010). Senescence in tumours: evidence from mice and humans. Nat Rev Cancer 10, 51-57.

Coudreuse, D., and Korswagen, H.C. (2007). The making of Wnt: new insights into Wnt maturation, sorting and secretion. Development 134, 3-12.

Dang, C.V. (2012). MYC on the path to cancer. Cell 149, 22-35.

Delandre, C., Penabaz, T.R., Passarelli, A.L., Chapes, S.K., and Clem, R.J. (2009). Mutation of juxtamembrane cysteines in the tetraspanin CD81 affects palmitoylation and alters interaction with other proteins at the cell surface. Experimental cell research 315, 1953-1963. 171

Devenport, D., and Fuchs, E. (2008). Planar polarization in embryonic epidermis orchestrates global asymmetric morphogenesis of hair follicles. Nat Cell Biol 10, 1257-1268.

Dirat, B., Bochet, L., Dabek, M., Daviaud, D., Dauvillier, S., Majed, B., Wang, Y.Y., Meulle, A., Salles, B., Le Gonidec, S., et al. (2011). Cancer-associated adipocytes exhibit an activated phenotype and contribute to breast cancer invasion. Cancer Res 71, 2455-2465.

Dorshkind, K., Keller, G.M., Phillips, R.A., Miller, R.G., Bosma, G.C., O'Toole, M., and Bosma, M.J. (1984). Functional status of cells from lymphoid and myeloid tissues in mice with severe combined immunodeficiency disease. J Immunol 132, 1804-1808.

Dvorak, H.F. (1986). Tumors: wounds that do not heal. Similarities between tumor stroma generation and wound healing. The New England journal of medicine 315, 1650-1659.

Egeblad, M., Nakasone, E.S., and Werb, Z. (2010). Tumors as organs: complex tissues that interface with the entire organism. Developmental cell 18, 884-901.

Erez, N., Truitt, M., Olson, P., Arron, S.T., and Hanahan, D. (2010). Cancer-Associated Fibroblasts Are Activated in Incipient Neoplasia to Orchestrate Tumor-Promoting Inflammation in an NF-kappaB-Dependent Manner. Cancer cell 17, 135-147.

Etienne-Manneville, S. (2008). Polarity proteins in migration and invasion. Oncogene 27, 6970- 6980.

Etienne-Manneville, S., and Hall, A. (2001). Integrin-mediated activation of Cdc42 controls cell polarity in migrating astrocytes through PKCzeta. Cell 106, 489-498.

Etienne-Manneville, S., and Hall, A. (2003). Cdc42 regulates GSK-3beta and adenomatous polyposis coli to control cell polarity. Nature 421, 753-756.

Evans, M.J., von Hahn, T., Tscherne, D.M., Syder, A.J., Panis, M., Wolk, B., Hatziioannou, T., McKeating, J.A., Bieniasz, P.D., and Rice, C.M. (2007). Claudin-1 is a hepatitis C virus coreceptor required for a late step in entry. Nature 446, 801-805.

Fang, Y., Wu, N., Gan, X., Yan, W., Morrell, J.C., and Gould, S.J. (2007). Higher-order oligomerization targets plasma membrane proteins and HIV gag to exosomes. PLoS biology 5, e158.

Fillmore, C.M., and Kuperwasser, C. (2008). Human breast cancer cell lines contain stem-like cells that self-renew, give rise to phenotypically diverse progeny and survive chemotherapy. Breast Cancer Res 10, R25.

172

Finak, G., Sadekova, S., Pepin, F., Hallett, M., Meterissian, S., Halwani, F., Khetani, K., Souleimanova, M., Zabolotny, B., Omeroglu, A., et al. (2006). Gene expression signatures of morphologically normal breast tissue identify basal-like tumors. Breast Cancer Res 8, R58.

Fitzner, D., Schnaars, M., van Rossum, D., Krishnamoorthy, G., Dibaj, P., Bakhti, M., Regen, T., Hanisch, U.K., and Simons, M. (2011). Selective transfer of exosomes from oligodendrocytes to microglia by macropinocytosis. Journal of cell science 124, 447-458.

Flint, M., von Hahn, T., Zhang, J., Farquhar, M., Jones, C.T., Balfe, P., Rice, C.M., and McKeating, J.A. (2006). Diverse CD81 proteins support hepatitis C virus infection. Journal of virology 80, 11331-11342.

Folkman, J. (2007). Angiogenesis: an organizing principle for drug discovery? Nat Rev Drug Discov 6, 273-286.

Friedl, P., and Alexander, S. (2011). Cancer invasion and the microenvironment: plasticity and reciprocity. Cell 147, 992-1009.

Friedl, P., and Wolf, K. (2009). Proteolytic interstitial cell migration: a five-step process. Cancer metastasis reviews 28, 129-135.

Gaggioli, C., Hooper, S., Hidalgo-Carcedo, C., Grosse, R., Marshall, J.F., Harrington, K., and Sahai, E. (2007). Fibroblast-led collective invasion of carcinoma cells with differing roles for RhoGTPases in leading and following cells. Nat Cell Biol 9, 1392-1400.

Gao, B., Song, H., Bishop, K., Elliot, G., Garrett, L., English, M.A., Andre, P., Robinson, J., Sood, R., Minami, Y., et al. (2011). Wnt signaling gradients establish planar cell polarity by inducing Vangl2 phosphorylation through Ror2. Developmental cell 20, 163-176.

Gao, C., and Chen, Y.G. (2010). Dishevelled: The hub of Wnt signaling. Cellular signalling 22, 717-727.

Gao, D., Nolan, D.J., Mellick, A.S., Bambino, K., McDonnell, K., and Mittal, V. (2008). Endothelial progenitor cells control the angiogenic switch in mouse lung metastasis. Science 319, 195-198.

Gay, L.J., and Felding-Habermann, B. (2011). Contribution of platelets to tumour metastasis. Nat Rev Cancer 11, 123-134.

Gesierich, S., Berezovskiy, I., Ryschich, E., and Zoller, M. (2006). Systemic induction of the angiogenesis switch by the tetraspanin D6.1A/CO-029. Cancer research 66, 7083-7094.

173

Giampieri, S., Manning, C., Hooper, S., Jones, L., Hill, C.S., and Sahai, E. (2009). Localized and reversible TGFbeta signalling switches breast cancer cells from cohesive to single cell motility. Nat Cell Biol 11, 1287-1296.

Giansanti, V., Torriglia, A., and Scovassi, A.I. (2011). Conversation between apoptosis and autophagy: "Is it your turn or mine?". Apoptosis : an international journal on programmed cell death 16, 321-333.

Goetz, J.G., Minguet, S., Navarro-Lerida, I., Lazcano, J.J., Samaniego, R., Calvo, E., Tello, M., Osteso-Ibanez, T., Pellinen, T., Echarri, A., et al. (2011). Biomechanical remodeling of the microenvironment by stromal caveolin-1 favors tumor invasion and metastasis. Cell 146, 148- 163.

Goldstein, B., Takeshita, H., Mizumoto, K., and Sawa, H. (2006). Wnt signals can function as positional cues in establishing cell polarity. Developmental cell 10, 391-396.

Gong, Y., Mo, C., and Fraser, S.E. (2004). Planar cell polarity signalling controls cell division orientation during zebrafish gastrulation. Nature 430, 689-693.

Goodman, R.M., Thombre, S., Firtina, Z., Gray, D., Betts, D., Roebuck, J., Spana, E.P., and Selva, E.M. (2006). Sprinter: a novel transmembrane protein required for Wg secretion and signaling. Development 133, 4901-4911.

Goodrich, L.V., and Strutt, D. (2011). Principles of planar polarity in animal development. Development 138, 1877-1892.

Gordon-Alonso, M., Sala-Valdes, M., Rocha-Perugini, V., Perez-Hernandez, D., Lopez-Martin, S., Ursa, A., Alvarez, S., Kolesnikova, T.V., Vazquez, J., Sanchez-Madrid, F., et al. (2012). EWI-2 association with alpha-actinin regulates T cell immune synapses and HIV viral infection. J Immunol 189, 689-700.

Gray, R.S., Roszko, I., and Solnica-Krezel, L. (2011). Planar cell polarity: coordinating morphogenetic cell behaviors with embryonic polarity. Dev Cell 21, 120-133.

Gros, J., Serralbo, O., and Marcelle, C. (2009). WNT11 acts as a directional cue to organize the elongation of early muscle fibres. Nature 457, 589-593.

Gross, J.C., Chaudhary, V., Bartscherer, K., and Boutros, M. (2012). Active Wnt proteins are secreted on exosomes. Nat Cell Biol.

Guo, N., Hawkins, C., and Nathans, J. (2004). Frizzled6 controls hair patterning in mice. Proc Natl Acad Sci U S A 101, 9277-9281.

174

Han, C., Yan, D., Belenkaya, T.Y., and Lin, X. (2005). Drosophila glypicans Dally and Dally- like shape the extracellular Wingless morphogen gradient in the wing disc. Development 132, 667-679.

Hanahan, D., and Weinberg, R.A. (2011). Hallmarks of cancer: the next generation. Cell 144, 646-674.

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.

Harley, C.B. (2008). Telomerase and cancer therapeutics. Nat Rev Cancer 8, 167-179.

Hashimoto, M., Shinohara, K., Wang, J., Ikeuchi, S., Yoshiba, S., Meno, C., Nonaka, S., Takada, S., Hatta, K., Wynshaw-Boris, A., et al. (2010). Planar polarization of node cells determines the rotational axis of node cilia. Nat Cell Biol 12, 170-176.

Hedlund, M., Stenqvist, A.C., Nagaeva, O., Kjellberg, L., Wulff, M., Baranov, V., and Mincheva-Nilsson, L. (2009). Human placenta expresses and secretes NKG2D ligands via exosomes that down-modulate the cognate receptor expression: evidence for immunosuppressive function. J Immunol 183, 340-351.

Heisenberg, C.P., Tada, M., Rauch, G.J., Saude, L., Concha, M.L., Geisler, R., Stemple, D.L., Smith, J.C., and Wilson, S.W. (2000). Silberblick/Wnt11 mediates convergent extension movements during zebrafish gastrulation. Nature 405, 76-81.

Heldin, C.H., Vanlandewijck, M., and Moustakas, A. (2012). Regulation of EMT by TGFbeta in cancer. FEBS letters 586, 1959-1970.

Hemler, M.E. (2005). Tetraspanin functions and associated microdomains. Nat Rev Mol Cell Biol 6, 801-811.

Hemler, M.E. (2008). Targeting of tetraspanin proteins--potential benefits and strategies. Nat Rev Drug Discov 7, 747-758.

Henne, W.M., Buchkovich, N.J., and Emr, S.D. (2011). The ESCRT pathway. Developmental cell 21, 77-91.

Hergenreider, E., Heydt, S., Treguer, K., Boettger, T., Horrevoets, A.J., Zeiher, A.M., Scheffer, M.P., Frangakis, A.S., Yin, X., Mayr, M., et al. (2012). Atheroprotective communication between endothelial cells and smooth muscle cells through miRNAs. Nature cell biology 14, 249-256.

Herr, P., and Basler, K. (2012). Porcupine-mediated lipidation is required for Wnt recognition by Wls. Developmental biology 361, 392-402. 175

Hood, J.L., San, R.S., and Wickline, S.A. (2011). Exosomes released by melanoma cells prepare sentinel lymph nodes for tumor metastasis. Cancer research 71, 3792-3801.

Hosein, A.N., Wu, M., Arcand, S.L., Lavallee, S., Hebert, J., Tonin, P.N., and Basik, M. (2010). Breast carcinoma-associated fibroblasts rarely contain p53 mutations or chromosomal aberrations. Cancer research 70, 5770-5777.

Huang da, W., Sherman, B.T., and Lempicki, R.A. (2009). Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat Protoc 4, 44-57.

Hunter, M.P., Ismail, N., Zhang, X., Aguda, B.D., Lee, E.J., Yu, L., Xiao, T., Schafer, J., Lee, M.L., Schmittgen, T.D., et al. (2008). Detection of microRNA expression in human peripheral blood microvesicles. PLoS One 3, e3694.

Hyodo-Miura, J., Yamamoto, T.S., Hyodo, A.C., Iemura, S., Kusakabe, M., Nishida, E., Natsume, T., and Ueno, N. (2006). XGAP, an ArfGAP, is required for polarized localization of PAR proteins and cell polarity in Xenopus gastrulation. Dev Cell 11, 69-79.

Iden, S., and Collard, J.G. (2008). Crosstalk between small GTPases and polarity proteins in cell polarization. Nat Rev Mol Cell Biol 9, 846-859.

Imai, T., and Yoshie, O. (1993). C33 antigen and M38 antigen recognized by monoclonal antibodies inhibitory to syncytium formation by human T cell leukemia virus type 1 are both members of the transmembrane 4 superfamily and associate with each other and with CD4 or CD8 in T cells. J Immunol 151, 6470-6481.

Ishida-Takagishi, M., Enomoto, A., Asai, N., Ushida, K., Watanabe, T., Hashimoto, T., Kato, T., Weng, L., Matsumoto, S., Asai, M., et al. (2012). The Dishevelled-associating protein Daple controls the non-canonical Wnt/Rac pathway and cell motility. Nat Commun 3, 859.

Iyengar, P., Espina, V., Williams, T.W., Lin, Y., Berry, D., Jelicks, L.A., Lee, H., Temple, K., Graves, R., Pollard, J., et al. (2005). Adipocyte-derived collagen VI affects early mammary tumor progression in vivo, demonstrating a critical interaction in the tumor/stroma microenvironment. The Journal of clinical investigation 115, 1163-1176.

Janda, C.Y., Waghray, D., Levin, A.M., Thomas, C., and Garcia, K.C. (2012). Structural basis of Wnt recognition by Frizzled. Science 337, 59-64.

Janowska-Wieczorek, A., Wysoczynski, M., Kijowski, J., Marquez-Curtis, L., Machalinski, B., Ratajczak, J., and Ratajczak, M.Z. (2005). Microvesicles derived from activated platelets induce metastasis and angiogenesis in lung cancer. Int J Cancer 113, 752-760.

Jessen, J.R. (2009). Noncanonical Wnt signaling in tumor progression and metastasis. Zebrafish 6, 21-28.

176

Jiang, D., Munro, E.M., and Smith, W.C. (2005). Ascidian prickle regulates both mediolateral and anterior-posterior cell polarity of notochord cells. Current biology : CB 15, 79-85.

Jung, T., Castellana, D., Klingbeil, P., Cuesta Hernandez, I., Vitacolonna, M., Orlicky, D.J., Roffler, S.R., Brodt, P., and Zoller, M. (2009). CD44v6 dependence of premetastatic niche preparation by exosomes. Neoplasia 11, 1093-1105.

Kadowaki, T., Wilder, E., Klingensmith, J., Zachary, K., and Perrimon, N. (1996). The segment polarity gene porcupine encodes a putative multitransmembrane protein involved in Wingless processing. Genes & development 10, 3116-3128.

Kaji, K., Oda, S., Miyazaki, S., and Kudo, A. (2002). Infertility of CD9-deficient mouse eggs is reversed by mouse CD9, human CD9, or mouse CD81; polyadenylated mRNA injection developed for molecular analysis of sperm-egg fusion. Dev Biol 247, 327-334.

Kalluri, R., and Zeisberg, M. (2006). Fibroblasts in cancer. Nat Rev Cancer 6, 392-401.

Kaplan, R.N., Riba, R.D., Zacharoulis, S., Bramley, A.H., Vincent, L., Costa, C., MacDonald, D.D., Jin, D.K., Shido, K., Kerns, S.A., et al. (2005). VEGFR1-positive haematopoietic bone marrow progenitors initiate the pre-metastatic niche. Nature 438, 820-827.

Karner, C.M., Chirumamilla, R., Aoki, S., Igarashi, P., Wallingford, J.B., and Carroll, T.J. (2009). Wnt9b signaling regulates planar cell polarity and kidney tubule morphogenesis. Nature genetics 41, 793-799.

Karnoub, A.E., Dash, A.B., Vo, A.P., Sullivan, A., Brooks, M.W., Bell, G.W., Richardson, A.L., Polyak, K., Tubo, R., and Weinberg, R.A. (2007). Mesenchymal stem cells within tumour stroma promote breast cancer metastasis. Nature 449, 557-563.

Karnoub, A.E., and Weinberg, R.A. (2008). Ras oncogenes: split personalities. Nat Rev Mol Cell Biol 9, 517-531.

Kilian, B., Mansukoski, H., Barbosa, F.C., Ulrich, F., Tada, M., and Heisenberg, C.P. (2003). The role of Ppt/Wnt5 in regulating cell shape and movement during zebrafish gastrulation. Mech Dev 120, 467-476.

Kim, J.W., Wieckowski, E., Taylor, D.D., Reichert, T.E., Watkins, S., and Whiteside, T.L. (2005). Fas ligand-positive membranous vesicles isolated from sera of patients with oral cancer induce apoptosis of activated T lymphocytes. Clin Cancer Res 11, 1010-1020.

Koles, K., Nunnari, J., Korkut, C., Barria, R., Brewer, C., Li, Y., Leszyk, J., Zhang, B., and Budnik, V. (2012). Mechanism of evenness interrupted (Evi)-exosome release at synaptic boutons. The Journal of biological chemistry 287, 16820-16834.

177

Koppenol, W.H., Bounds, P.L., and Dang, C.V. (2011). Otto Warburg's contributions to current concepts of cancer metabolism. Nat Rev Cancer 11, 325-337.

Korkut, C., Ataman, B., Ramachandran, P., Ashley, J., Barria, R., Gherbesi, N., and Budnik, V. (2009). Trans-synaptic transmission of vesicular Wnt signals through Evi/Wntless. Cell 139, 393-404.

Korpal, M., Ell, B.J., Buffa, F.M., Ibrahim, T., Blanco, M.A., Celia-Terrassa, T., Mercatali, L., Khan, Z., Goodarzi, H., Hua, Y., et al. (2011). Direct targeting of Sec23a by miR-200s influences cancer cell secretome and promotes metastatic colonization. Nat Med 17, 1101-1108.

Kosaka, N., Iguchi, H., Yoshioka, Y., Takeshita, F., Matsuki, Y., and Ochiya, T. (2010). Secretory mechanisms and intercellular transfer of microRNAs in living cells. The Journal of biological chemistry 285, 17442-17452.

Kreuzaler, P., and Watson, C.J. (2012). Killing a cancer: what are the alternatives? Nat Rev Cancer 12, 411-424.

Landskroner-Eiger, S., Qian, B., Muise, E.S., Nawrocki, A.R., Berger, J.P., Fine, E.J., Koba, W., Deng, Y., Pollard, J.W., and Scherer, P.E. (2009). Proangiogenic contribution of adiponectin toward mammary tumor growth in vivo. Clin Cancer Res 15, 3265-3276.

Lapidot, T., Sirard, C., Vormoor, J., Murdoch, B., Hoang, T., Caceres-Cortes, J., Minden, M., Paterson, B., Caligiuri, M.A., and Dick, J.E. (1994). A cell initiating human acute myeloid leukaemia after transplantation into SCID mice. Nature 367, 645-648.

Laulagnier, K., Motta, C., Hamdi, S., Roy, S., Fauvelle, F., Pageaux, J.F., Kobayashi, T., Salles, J.P., Perret, B., Bonnerot, C., et al. (2004). Mast cell- and dendritic cell-derived exosomes display a specific lipid composition and an unusual membrane organization. Biochem J 380, 161-171.

Lawrence, P.A., Struhl, G., and Casal, J. (2007). Planar cell polarity: one or two pathways? Nature reviews Genetics 8, 555-563.

Lee, G.Y., Kenny, P.A., Lee, E.H., and Bissell, M.J. (2007). Three-dimensional culture models of normal and malignant breast epithelial cells. Nat Methods 4, 359-365.

Lee, J.H., Park, S.R., Chay, K.O., Seo, Y.W., Kook, H., Ahn, K.Y., Kim, Y.J., and Kim, K.K. (2004). KAI1 COOH-terminal interacting tetraspanin (KITENIN), a member of the tetraspanin family, interacts with KAI1, a tumor metastasis suppressor, and enhances metastasis of cancer. Cancer research 64, 4235-4243.

178

Lee, J.K., Bae, J.A., Sun, E.G., Kim, H.D., Yoon, T.M., Kim, K., Lee, J.H., Lim, S.C., and Kim, K.K. (2009). KITENIN increases invasion and migration of mouse squamous cancer cells and promotes pulmonary metastasis in a mouse squamous tumor model. FEBS letters 583, 711-717.

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

Levy, S., and Shoham, T. (2005). The tetraspanin web modulates immune-signalling complexes. Nature reviews Immunology 5, 136-148.

Levy, S., Todd, S.C., and Maecker, H.T. (1998). CD81 (TAPA-1): a molecule involved in signal transduction and cell adhesion in the immune system. Annual review of immunology 16, 89- 109.

Li, W., Cooper, J., Karajannis, M.A., and Giancotti, F.G. (2012a). Merlin: a tumour suppressor with functions at the cell cortex and in the nucleus. EMBO Rep 13, 204-215.

Li, X., Li, J.J., Yang, J.Y., Wang, D.S., Zhao, W., Song, W.J., Li, W.M., Wang, J.F., Han, W., Zhang, Z.C., et al. (2012b). Tolerance induction by exosomes from immature dendritic cells and rapamycin in a mouse cardiac allograft model. PLoS One 7, e44045.

Li, Z., Hannigan, M., Mo, Z., Liu, B., Lu, W., Wu, Y., Smrcka, A.V., Wu, G., Li, L., Liu, M., et al. (2003). Directional sensing requires G beta gamma-mediated PAK1 and PIX alpha- dependent activation of Cdc42. Cell 114, 215-227.

Little, K.D., Hemler, M.E., and Stipp, C.S. (2004). Dynamic regulation of a GPCR-tetraspanin- G protein complex on intact cells: central role of CD81 in facilitating GPR56-Galpha q/11 association. Mol Biol Cell 15, 2375-2387.

Liu, G., Zhang, J., Larsen, B., Stark, C., Breitkreutz, A., Lin, Z.Y., Breitkreutz, B.J., Ding, Y., Colwill, K., Pasculescu, A., et al. (2010). ProHits: integrated software for mass spectrometry- based interaction proteomics. Nat Biotechnol 28, 1015-1017.

Livak, K.J., and Schmittgen, T.D. (2001). Analysis of relative gene expression data using real- time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods 25, 402-408.

Lyden, D., Hattori, K., Dias, S., Costa, C., Blaikie, P., Butros, L., Chadburn, A., Heissig, B., Marks, W., Witte, L., et al. (2001). Impaired recruitment of bone-marrow-derived endothelial and hematopoietic precursor cells blocks tumor angiogenesis and growth. Nat Med 7, 1194- 1201.

Ma, L., Teruya-Feldstein, J., and Weinberg, R.A. (2007). Tumour invasion and metastasis initiated by microRNA-10b in breast cancer. Nature 449, 682-688.

179

Macia, E., Ehrlich, M., Massol, R., Boucrot, E., Brunner, C., and Kirchhausen, T. (2006). Dynasore, a cell-permeable inhibitor of dynamin. Dev Cell 10, 839-850.

Maecker, H.T., and Levy, S. (1997). Normal lymphocyte development but delayed humoral immune response in CD81-null mice. J Exp Med 185, 1505-1510.

Mani, S.A., Guo, W., Liao, M.J., Eaton, E.N., Ayyanan, A., Zhou, A.Y., Brooks, M., Reinhard, F., Zhang, C.C., Shipitsin, M., et al. (2008). The epithelial-mesenchymal transition generates cells with properties of stem cells. Cell 133, 704-715.

Mantovani, A., Allavena, P., Sica, A., and Balkwill, F. (2008). Cancer-related inflammation. Nature 454, 436-444.

Martin-Belmonte, F., and Perez-Moreno, M. (2012). Epithelial cell polarity, stem cells and cancer. Nat Rev Cancer 12, 23-38.

Massague, J. (2012). TGFbeta signalling in context. Nat Rev Mol Cell Biol 13, 616-630.

Mathivanan, S., Fahner, C.J., Reid, G.E., and Simpson, R.J. (2012). ExoCarta 2012: database of exosomal proteins, RNA and lipids. Nucleic Acids Res 40, D1241-1244.

Mathivanan, S., Ji, H., and Simpson, R.J. (2010). Exosomes: extracellular organelles important in intercellular communication. J Proteomics 73, 1907-1920.

Mattila, P.K., and Lappalainen, P. (2008). Filopodia: molecular architecture and cellular functions. Nat Rev Mol Cell Biol 9, 446-454.

Mikels, A.J., and Nusse, R. (2006). Wnts as ligands: processing, secretion and reception. Oncogene 25, 7461-7468.

Minn, A.J., Gupta, G.P., Siegel, P.M., Bos, P.D., Shu, W., Giri, D.D., Viale, A., Olshen, A.B., Gerald, W.L., and Massague, J. (2005). Genes that mediate breast cancer metastasis to lung. Nature 436, 518-524.

Mitra, S.K., Hanson, D.A., and Schlaepfer, D.D. (2005). Focal adhesion kinase: in command and control of cell motility. Nat Rev Mol Cell Biol 6, 56-68.

Mittelbrunn, M., Gutierrez-Vazquez, C., Villarroya-Beltri, C., Gonzalez, S., Sanchez-Cabo, F., Gonzalez, M.A., Bernad, A., and Sanchez-Madrid, F. (2011). Unidirectional transfer of microRNA-loaded exosomes from T cells to antigen-presenting cells. Nat Commun 2, 282.

Miyazaki, T., Muller, U., and Campbell, K.S. (1997). Normal development but differentially altered proliferative responses of lymphocytes in mice lacking CD81. EMBO J 16, 4217-4225.

180

Montcouquiol, M., Sans, N., Huss, D., Kach, J., Dickman, J.D., Forge, A., Rachel, R.A., Copeland, N.G., Jenkins, N.A., Bogani, D., et al. (2006). Asymmetric localization of Vangl2 and Fz3 indicate novel mechanisms for planar cell polarity in mammals. J Neurosci 26, 5265- 5275.

Montpellier, C., Tews, B.A., Poitrimole, J., Rocha-Perugini, V., D'Arienzo, V., Potel, J., Zhang, X.A., Rubinstein, E., Dubuisson, J., and Cocquerel, L. (2011). Interacting regions of CD81 and two of its partners, EWI-2 and EWI-2wint, and their effect on hepatitis C virus infection. J Biol Chem 286, 13954-13965.

Moon, R.T., Campbell, R.M., Christian, J.L., McGrew, L.L., Shih, J., and Fraser, S. (1993). Xwnt-5A: a maternal Wnt that affects morphogenetic movements after overexpression in embryos of Xenopus laevis. Development 119, 97-111.

Morel, A.P., Lievre, M., Thomas, C., Hinkal, G., Ansieau, S., and Puisieux, A. (2008). Generation of breast cancer stem cells through epithelial-mesenchymal transition. PLoS One 3, e2888.

Morelli, A.E., Larregina, A.T., Shufesky, W.J., Sullivan, M.L., Stolz, D.B., Papworth, G.D., Zahorchak, A.F., Logar, A.J., Wang, Z., Watkins, S.C., et al. (2004). Endocytosis, intracellular sorting, and processing of exosomes by dendritic cells. Blood 104, 3257-3266.

Morrell, N.T., Leucht, P., Zhao, L., Kim, J.B., ten Berge, D., Ponnusamy, K., Carre, A.L., Dudek, H., Zachlederova, M., McElhaney, M., et al. (2008). Liposomal packaging generates Wnt protein with in vivo biological activity. PLoS One 3, e2930.

Morse, M.A., Garst, J., Osada, T., Khan, S., Hobeika, A., Clay, T.M., Valente, N., Shreeniwas, R., Sutton, M.A., Delcayre, A., et al. (2005). A phase I study of dexosome immunotherapy in patients with advanced non-small cell lung cancer. J Transl Med 3, 9.

Mulligan, K.A., Fuerer, C., Ching, W., Fish, M., Willert, K., and Nusse, R. (2012). Secreted Wingless-interacting molecule (Swim) promotes long-range signaling by maintaining Wingless solubility. Proceedings of the National Academy of Sciences of the United States of America 109, 370-377.

Nardella, C., Clohessy, J.G., Alimonti, A., and Pandolfi, P.P. (2011). Pro-senescence therapy for cancer treatment. Nat Rev Cancer 11, 503-511.

Narimatsu, M., Bose, R., Pye, M., Zhang, L., Miller, B., Ching, P., Sakuma, R., Luga, V., Roncari, L., Attisano, L., et al. (2009). Regulation of planar cell polarity by Smurf ubiquitin ligases. Cell 137, 295-307.

Nazarenko, I., Rana, S., Baumann, A., McAlear, J., Hellwig, A., Trendelenburg, M., Lochnit, G., Preissner, K.T., and Zoller, M. (2010). Cell surface tetraspanin Tspan8 contributes to

181

molecular pathways of exosome-induced endothelial cell activation. Cancer research 70, 1668- 1678.

Nomachi, A., Nishita, M., Inaba, D., Enomoto, M., Hamasaki, M., and Minami, Y. (2008). Receptor tyrosine kinase Ror2 mediates Wnt5a-induced polarized cell migration by activating c- Jun N-terminal kinase via actin-binding protein filamin A. J Biol Chem 283, 27973-27981.

Nurnberg, A., Kitzing, T., and Grosse, R. (2011). Nucleating actin for invasion. Nat Rev Cancer 11, 177-187.

Ocana, O.H., Corcoles, R., Fabra, A., Moreno-Bueno, G., Acloque, H., Vega, S., Barrallo- Gimeno, A., Cano, A., and Nieto, M.A. (2012). Metastatic colonization requires the repression of the epithelial-mesenchymal transition inducer Prrx1. Cancer Cell 22, 709-724.

Odorizzi, G. (2006). The multiple personalities of Alix. Journal of cell science 119, 3025-3032.

Oft, M., Peli, J., Rudaz, C., Schwarz, H., Beug, H., and Reichmann, E. (1996). TGF-beta1 and Ha-Ras collaborate in modulating the phenotypic plasticity and invasiveness of epithelial tumor cells. Genes & development 10, 2462-2477.

Ohkawara, B., Glinka, A., and Niehrs, C. (2011). Rspo3 binds syndecan 4 and induces Wnt/PCP signaling via clathrin-mediated endocytosis to promote morphogenesis. Dev Cell 20, 303-314.

Olson, M.F., and Sahai, E. (2009). The actin cytoskeleton in cancer cell motility. Clin Exp Metastasis 26, 273-287.

Orimo, A., Gupta, P.B., Sgroi, D.C., Arenzana-Seisdedos, F., Delaunay, T., Naeem, R., Carey, V.J., Richardson, A.L., and Weinberg, R.A. (2005). Stromal fibroblasts present in invasive human breast carcinomas promote tumor growth and angiogenesis through elevated SDF- 1/CXCL12 secretion. Cell 121, 335-348.

Ozdamar, B., Bose, R., Barrios-Rodiles, M., Wang, H.R., Zhang, Y., and Wrana, J.L. (2005). Regulation of the polarity protein Par6 by TGFbeta receptors controls epithelial cell plasticity. Science 307, 1603-1609.

Paget, S. (1989). The distribution of secondary growths in cancer of the breast. 1889. Cancer metastasis reviews 8, 98-101.

Pan, B.T., and Johnstone, R.M. (1983). Fate of the transferrin receptor during maturation of sheep reticulocytes in vitro: selective externalization of the receptor. Cell 33, 967-978.

Panakova, D., Sprong, H., Marois, E., Thiele, C., and Eaton, S. (2005). Lipoprotein particles are required for Hedgehog and Wingless signalling. Nature 435, 58-65.

182

Parolini, I., Federici, C., Raggi, C., Lugini, L., Palleschi, S., De Milito, A., Coscia, C., Iessi, E., Logozzi, M., Molinari, A., et al. (2009). Microenvironmental pH is a key factor for exosome traffic in tumor cells. The Journal of biological chemistry 284, 34211-34222.

Pegtel, D.M., Cosmopoulos, K., Thorley-Lawson, D.A., van Eijndhoven, M.A., Hopmans, E.S., Lindenberg, J.L., de Gruijl, T.D., Wurdinger, T., and Middeldorp, J.M. (2010). Functional delivery of viral miRNAs via exosomes. Proceedings of the National Academy of Sciences of the United States of America 107, 6328-6333.

Peinado, H., Aleckovic, M., Lavotshkin, S., Matei, I., Costa-Silva, B., Moreno-Bueno, G., Hergueta-Redondo, M., Williams, C., Garcia-Santos, G., Ghajar, C., et al. (2012). Melanoma exosomes educate bone marrow progenitor cells toward a pro-metastatic phenotype through MET. Nat Med 18, 883-891.

Pfeiffer, S., Ricardo, S., Manneville, J.B., Alexandre, C., and Vincent, J.P. (2002). Producing cells retain and recycle Wingless in Drosophila embryos. Curr Biol 12, 957-962.

Pileri, P., Uematsu, Y., Campagnoli, S., Galli, G., Falugi, F., Petracca, R., Weiner, A.J., Houghton, M., Rosa, D., Grandi, G., et al. (1998). Binding of hepatitis C virus to CD81. Science 282, 938-941.

Ploss, A., Evans, M.J., Gaysinskaya, V.A., Panis, M., You, H., de Jong, Y.P., and Rice, C.M. (2009). Human occludin is a hepatitis C virus entry factor required for infection of mouse cells. Nature 457, 882-886.

Polager, S., and Ginsberg, D. (2009). p53 and E2f: partners in life and death. Nat Rev Cancer 9, 738-748.

Pollard, T.D. (2007). Regulation of actin filament assembly by Arp2/3 complex and formins. Annual review of biophysics and biomolecular structure 36, 451-477.

Port, F., and Basler, K. (2010). Wnt trafficking: new insights into Wnt maturation, secretion and spreading. Traffic 11, 1265-1271.

Port, F., Hausmann, G., and Basler, K. (2011). A genome-wide RNA interference screen uncovers two p24 proteins as regulators of Wingless secretion. EMBO reports 12, 1144-1152.

Psaila, B., and Lyden, D. (2009). The metastatic niche: adapting the foreign soil. Nat Rev Cancer 9, 285-293.

Putz, U., Howitt, J., Doan, A., Goh, C.P., Low, L.H., Silke, J., and Tan, S.S. (2012). The Tumor Suppressor PTEN Is Exported in Exosomes and Has Phosphatase Activity in Recipient Cells. Science signaling 5, ra70.

183

Qian, D., Jones, C., Rzadzinska, A., Mark, S., Zhang, X., Steel, K.P., Dai, X., and Chen, P. (2007). Wnt5a functions in planar cell polarity regulation in mice. Dev Biol 306, 121-133.

Raiborg, C., and Stenmark, H. (2009). The ESCRT machinery in endosomal sorting of ubiquitylated membrane proteins. Nature 458, 445-452.

Raposo, G., Nijman, H.W., Stoorvogel, W., Liejendekker, R., Harding, C.V., Melief, C.J., and Geuze, H.J. (1996). B lymphocytes secrete antigen-presenting vesicles. J Exp Med 183, 1161- 1172.

Rasanen, K., and Vaheri, A. (2010). Activation of fibroblasts in cancer stroma. Experimental cell research 316, 2713-2722.

Ratajczak, J., Miekus, K., Kucia, M., Zhang, J., Reca, R., Dvorak, P., and Ratajczak, M.Z. (2006). Embryonic stem cell-derived microvesicles reprogram hematopoietic progenitors: evidence for horizontal transfer of mRNA and protein delivery. Leukemia 20, 847-856.

Ridley, A.J., Schwartz, M.A., Burridge, K., Firtel, R.A., Ginsberg, M.H., Borisy, G., Parsons, J.T., and Horwitz, A.R. (2003). Cell migration: integrating signals from front to back. Science 302, 1704-1709.

Rijsewijk, F., Schuermann, M., Wagenaar, E., Parren, P., Weigel, D., and Nusse, R. (1987). The Drosophila homolog of the mouse mammary oncogene int-1 is identical to the segment polarity gene wingless. Cell 50, 649-657.

Rubinstein, E., Ziyyat, A., Prenant, M., Wrobel, E., Wolf, J.P., Levy, S., Le Naour, F., and Boucheix, C. (2006). Reduced fertility of female mice lacking CD81. Dev Biol 290, 351-358.

Sahai, E., Garcia-Medina, R., Pouyssegur, J., and Vial, E. (2007). Smurf1 regulates tumor cell plasticity and motility through degradation of RhoA leading to localized inhibition of contractility. J Cell Biol 176, 35-42.

Sala-Valdes, M., Ursa, A., Charrin, S., Rubinstein, E., Hemler, M.E., Sanchez-Madrid, F., and Yanez-Mo, M. (2006). EWI-2 and EWI-F link the tetraspanin web to the actin cytoskeleton through their direct association with ezrin-radixin-moesin proteins. The Journal of biological chemistry 281, 19665-19675.

Samavarchi-Tehrani, P., Golipour, A., David, L., Sung, H.K., Beyer, T.A., Datti, A., Woltjen, K., Nagy, A., and Wrana, J.L. (2010). Functional genomics reveals a BMP-driven mesenchymal-to-epithelial transition in the initiation of somatic cell reprogramming. Cell stem cell 7, 64-77.

Sanders, T.A., Llagostera, E., and Barna, M. (2013). Specialized filopodia direct long-range transport of SHH during vertebrate tissue patterning. Nature.

184

Sanford, K.K., Hobbs, G.L., and Earle, W.R. (1956). The tumor-producing capacity of strain L mouse cells after 10 years in vitro. Cancer research 16, 162-166.

Schlessinger, K., McManus, E.J., and Hall, A. (2007). Cdc42 and noncanonical Wnt signal transduction pathways cooperate to promote cell polarity. J Cell Biol 178, 355-361.

Segura, E., Guerin, C., Hogg, N., Amigorena, S., and Thery, C. (2007). CD8+ dendritic cells use LFA-1 to capture MHC-peptide complexes from exosomes in vivo. J Immunol 179, 1489-1496.

Seifert, J.R., and Mlodzik, M. (2007). Frizzled/PCP signalling: a conserved mechanism regulating cell polarity and directed motility. Nat Rev Genet 8, 126-138.

Sharma, C., Yang, X.H., and Hemler, M.E. (2008). DHHC2 affects palmitoylation, stability, and functions of tetraspanins CD9 and CD151. Mol Biol Cell 19, 3415-3425.

Sheldon, H., Heikamp, E., Turley, H., Dragovic, R., Thomas, P., Oon, C.E., Leek, R., Edelmann, M., Kessler, B., Sainson, R.C., et al. (2010). New mechanism for Notch signaling to endothelium at a distance by Delta-like 4 incorporation into exosomes. Blood 116, 2385-2394.

Sherer, N.M., and Mothes, W. (2008). Cytonemes and tunneling nanotubules in cell-cell communication and viral pathogenesis. Trends Cell Biol 18, 414-420.

Shimada, Y., Yonemura, S., Ohkura, H., Strutt, D., and Uemura, T. (2006). Polarized transport of Frizzled along the planar microtubule arrays in Drosophila wing epithelium. Dev Cell 10, 209-222.

Shnitsar, I., and Borchers, A. (2008). PTK7 recruits dsh to regulate neural crest migration. Development 135, 4015-4024.

Shulman, J.M., Perrimon, N., and Axelrod, J.D. (1998). Frizzled signaling and the developmental control of cell polarity. Trends Genet 14, 452-458.

Silvie, O., Rubinstein, E., Franetich, J.F., Prenant, M., Belnoue, E., Renia, L., Hannoun, L., Eling, W., Levy, S., Boucheix, C., et al. (2003). Hepatocyte CD81 is required for Plasmodium falciparum and Plasmodium yoelii sporozoite infectivity. Nat Med 9, 93-96.

Simons, M., and Mlodzik, M. (2008). Planar cell polarity signaling: from fly development to human disease. Annual review of genetics 42, 517-540.

Simons, M., and Raposo, G. (2009). Exosomes--vesicular carriers for intercellular communication. Curr Opin Cell Biol 21, 575-581.

Skog, J., Wurdinger, T., van Rijn, S., Meijer, D.H., Gainche, L., Sena-Esteves, M., Curry, W.T., Jr., Carter, B.S., Krichevsky, A.M., and Breakefield, X.O. (2008). Glioblastoma microvesicles 185

transport RNA and proteins that promote tumour growth and provide diagnostic biomarkers. Nature cell biology 10, 1470-1476.

Song, M.S., Salmena, L., and Pandolfi, P.P. (2012). The functions and regulation of the PTEN tumour suppressor. Nat Rev Mol Cell Biol 13, 283-296.

Sopko, R., and McNeill, H. (2009). The skinny on Fat: an enormous cadherin that regulates cell adhesion, tissue growth, and planar cell polarity. Curr Opin Cell Biol 21, 717-723.

Stelzer, G., Dalah, I., Stein, T.I., Satanower, Y., Rosen, N., Nativ, N., Oz-Levi, D., Olender, T., Belinky, F., Bahir, I., et al. (2011). In-silico human genomics with GeneCards. Human genomics 5, 709-717.

Sternlicht, M.D., Lochter, A., Sympson, C.J., Huey, B., Rougier, J.P., Gray, J.W., Pinkel, D., Bissell, M.J., and Werb, Z. (1999). The stromal proteinase MMP3/stromelysin-1 promotes mammary carcinogenesis. Cell 98, 137-146.

Stoorvogel, W., Kleijmeer, M.J., Geuze, H.J., and Raposo, G. (2002). The biogenesis and functions of exosomes. Traffic 3, 321-330.

Sun, D., Zhuang, X., Xiang, X., Liu, Y., Zhang, S., Liu, C., Barnes, S., Grizzle, W., Miller, D., and Zhang, H.G. (2010). A novel nanoparticle drug delivery system: the anti-inflammatory activity of curcumin is enhanced when encapsulated in exosomes. Mol Ther 18, 1606-1614.

Sun, Y., Campisi, J., Higano, C., Beer, T.M., Porter, P., Coleman, I., True, L., and Nelson, P.S. (2012). Treatment-induced damage to the tumor microenvironment promotes prostate cancer therapy resistance through WNT16B. Nat Med 18, 1359-1368.

Tada, M., and Smith, J.C. (2000). Xwnt11 is a target of Xenopus Brachyury: regulation of gastrulation movements via Dishevelled, but not through the canonical Wnt pathway. Development 127, 2227-2238.

Takada, R., Satomi, Y., Kurata, T., Ueno, N., Norioka, S., Kondoh, H., Takao, T., and Takada, S. (2006). Monounsaturated fatty acid modification of Wnt protein: its role in Wnt secretion. Developmental cell 11, 791-801.

Takeuchi, M., Nakabayashi, J., Sakaguchi, T., Yamamoto, T.S., Takahashi, H., Takeda, H., and Ueno, N. (2003). The prickle-related gene in vertebrates is essential for gastrulation cell movements. Curr Biol 13, 674-679.

Talmadge, J.E., and Fidler, I.J. (2010). AACR centennial series: the biology of cancer metastasis: historical perspective. Cancer Res 70, 5649-5669.

186

Tammela, T., and Alitalo, K. (2010). Lymphangiogenesis: Molecular mechanisms and future promise. Cell 140, 460-476.

Taylor, D.D., and Gercel-Taylor, C. (2008). MicroRNA signatures of tumor-derived exosomes as diagnostic biomarkers of ovarian cancer. Gynecol Oncol 110, 13-21.

Taylor, D.D., and Gercel-Taylor, C. (2011). Exosomes/microvesicles: mediators of cancer- associated immunosuppressive microenvironments. Semin Immunopathol 33, 441-454.

Thali, M. (2011). Tetraspanin functions during HIV-1 and influenza virus replication. Biochem Soc Trans 39, 529-531.

Thery, C., Amigorena, S., Raposo, G., and Clayton, A. (2006). Isolation and characterization of exosomes from cell culture supernatants and biological fluids. Curr Protoc Cell Biol Chapter 3, Unit 3 22.

Thery, C., Ostrowski, M., and Segura, E. (2009). Membrane vesicles as conveyors of immune responses. Nat Rev Immunol 9, 581-593.

Thiery, J.P., Acloque, H., Huang, R.Y., and Nieto, M.A. (2009). Epithelial-mesenchymal transitions in development and disease. Cell 139, 871-890.

Thiery, J.P., and Sleeman, J.P. (2006). Complex networks orchestrate epithelial-mesenchymal transitions. Nat Rev Mol Cell Biol 7, 131-142.

Topczewski, J., Sepich, D.S., Myers, D.C., Walker, C., Amores, A., Lele, Z., Hammerschmidt, M., Postlethwait, J., and Solnica-Krezel, L. (2001). The zebrafish glypican knypek controls cell polarity during gastrulation movements of convergent extension. Developmental cell 1, 251- 264.

Trajkovic, K., Hsu, C., Chiantia, S., Rajendran, L., Wenzel, D., Wieland, F., Schwille, P., Brugger, B., and Simons, M. (2008). Ceramide triggers budding of exosome vesicles into multivesicular endosomes. Science 319, 1244-1247.

Tsai, J.H., Donaher, J.L., Murphy, D.A., Chau, S., and Yang, J. (2012). Spatiotemporal regulation of epithelial-mesenchymal transition is essential for squamous cell carcinoma metastasis. Cancer Cell 22, 725-736.

Tsitsikov, E.N., Gutierrez-Ramos, J.C., and Geha, R.S. (1997). Impaired CD19 expression and signaling, enhanced antibody response to type II T independent antigen and reduction of B-1 cells in CD81-deficient mice. Proc Natl Acad Sci U S A 94, 10844-10849.

187

Valadi, H., Ekstrom, K., Bossios, A., Sjostrand, M., Lee, J.J., and Lotvall, J.O. (2007). Exosome-mediated transfer of mRNAs and microRNAs is a novel mechanism of genetic exchange between cells. Nat Cell Biol 9, 654-659.

Valenti, R., Huber, V., Filipazzi, P., Pilla, L., Sovena, G., Villa, A., Corbelli, A., Fais, S., Parmiani, G., and Rivoltini, L. (2006). Human tumor-released microvesicles promote the differentiation of myeloid cells with transforming growth factor-beta-mediated suppressive activity on T lymphocytes. Cancer research 66, 9290-9298. van den Boorn, J.G., Schlee, M., Coch, C., and Hartmann, G. (2011). SiRNA delivery with exosome nanoparticles. Nat Biotechnol 29, 325-326. van Niel, G., Porto-Carreiro, I., Simoes, S., and Raposo, G. (2006). Exosomes: a common pathway for a specialized function. J Biochem 140, 13-21.

Varelas, X., Samavarchi-Tehrani, P., Narimatsu, M., Weiss, A., Cockburn, K., Larsen, B.G., Rossant, J., and Wrana, J.L. (2010). The Crumbs complex couples cell density sensing to Hippo-dependent control of the TGF-beta-SMAD pathway. Dev Cell 19, 831-844.

Veeman, M.T., Slusarski, D.C., Kaykas, A., Louie, S.H., and Moon, R.T. (2003). Zebrafish prickle, a modulator of noncanonical Wnt/Fz signaling, regulates gastrulation movements. Curr Biol 13, 680-685.

Viaud, S., Thery, C., Ploix, S., Tursz, T., Lapierre, V., Lantz, O., Zitvogel, L., and Chaput, N. (2010). Dendritic cell-derived exosomes for cancer immunotherapy: what's next? Cancer research 70, 1281-1285.

Vidal, M., Mangeat, P., and Hoekstra, D. (1997). Aggregation reroutes molecules from a recycling to a vesicle-mediated secretion pathway during reticulocyte maturation. J Cell Sci 110 ( Pt 16), 1867-1877.

Viloria-Petit, A.M., David, L., Jia, J.Y., Erdemir, T., Bane, A.L., Pinnaduwage, D., Roncari, L., Narimatsu, M., Bose, R., Moffat, J., et al. (2009). A role for the TGFbeta-Par6 polarity pathway in breast cancer progression. Proc Natl Acad Sci U S A 106, 14028-14033.

Visvader, J.E. (2011). Cells of origin in cancer. Nature 469, 314-322.

Wallingford, J.B. (2012). Planar cell polarity and the developmental control of cell behavior in vertebrate embryos. Annual review of cell and developmental biology 28, 627-653.

Wallingford, J.B., Ewald, A.J., Harland, R.M., and Fraser, S.E. (2001a). Calcium signaling during convergent extension in Xenopus. Curr Biol 11, 652-661.

188

Wallingford, J.B., Vogeli, K.M., and Harland, R.M. (2001b). Regulation of convergent extension in Xenopus by Wnt5a and Frizzled-8 is independent of the canonical Wnt pathway. Int J Dev Biol 45, 225-227.

Wang, H.R., Zhang, Y., Ozdamar, B., Ogunjimi, A.A., Alexandrova, E., Thomsen, G.H., and Wrana, J.L. (2003). Regulation of cell polarity and protrusion formation by targeting RhoA for degradation. Science 302, 1775-1779.

Wang, Y. (2009). Wnt/Planar cell polarity signaling: a new paradigm for cancer therapy. Mol Cancer Ther 8, 2103-2109.

Wang, Y., Guo, N., and Nathans, J. (2006). The role of Frizzled3 and Frizzled6 in neural tube closure and in the planar polarity of inner-ear sensory hair cells. J Neurosci 26, 2147-2156.

Weber, U., Gault, W.J., Olguin, P., Serysheva, E., and Mlodzik, M. (2012). Novel regulators of planar cell polarity: a genetic analysis in Drosophila. Genetics 191, 145-162.

Wehner, P., Shnitsar, I., Urlaub, H., and Borchers, A. (2011). RACK1 is a novel interaction partner of PTK7 that is required for neural tube closure. Development 138, 1321-1327.

Willert, K., and Nusse, R. (2012). Wnt proteins. Cold Spring Harb Perspect Biol 4.

Wirtz, D., Konstantopoulos, K., and Searson, P.C. (2011). The physics of cancer: the role of physical interactions and mechanical forces in metastasis. Nat Rev Cancer 11, 512-522.

Witsch, E., Sela, M., and Yarden, Y. (2010). Roles for growth factors in cancer progression. Physiology (Bethesda) 25, 85-101.

Wolfers, J., Lozier, A., Raposo, G., Regnault, A., Thery, C., Masurier, C., Flament, C., Pouzieux, S., Faure, F., Tursz, T., et al. (2001). Tumor-derived exosomes are a source of shared tumor rejection antigens for CTL cross-priming. Nature medicine 7, 297-303.

Wu, G., Huang, X., Hua, Y., and Mu, D. (2011). Roles of planar cell polarity pathways in the development of neural [correction of neutral] tube defects. Journal of biomedical science 18, 66.

Wu, J., and Mlodzik, M. (2009). A quest for the mechanism regulating global planar cell polarity of tissues. Trends Cell Biol 19, 295-305.

Wubbolts, R., Leckie, R.S., Veenhuizen, P.T., Schwarzmann, G., Mobius, W., Hoernschemeyer, J., 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. The Journal of biological chemistry 278, 10963-10972.

189

Xu, D., Sharma, C., and Hemler, M.E. (2009). Tetraspanin12 regulates ADAM10-dependent cleavage of amyloid precursor protein. FASEB journal : official publication of the Federation of American Societies for Experimental Biology 23, 3674-3681.

Xu, L., and Hynes, R.O. (2007). GPR56 and TG2: possible roles in suppression of tumor growth by the microenvironment. Cell Cycle 6, 160-165.

Yang, C., and Robbins, P.D. (2011). The roles of tumor-derived exosomes in cancer pathogenesis. Clin Dev Immunol 2011, 842849.

Yang, M., Chen, J., Su, F., Yu, B., Su, F., Lin, L., Liu, Y., Huang, J.D., and Song, E. (2011). Microvesicles secreted by macrophages shuttle invasion-potentiating microRNAs into breast cancer cells. Molecular cancer 10, 117.

Yang, Y. (2012). Wnt signaling in development and disease. Cell & bioscience 2, 14.

Yen, W.W., Williams, M., Periasamy, A., Conaway, M., Burdsal, C., Keller, R., Lu, X., and Sutherland, A. (2009). PTK7 is essential for polarized cell motility and convergent extension during mouse gastrulation. Development 136, 2039-2048.

Yin, C., Kiskowski, M., Pouille, P.A., Farge, E., and Solnica-Krezel, L. (2008). Cooperation of polarized cell intercalations drives convergence and extension of presomitic mesoderm during zebrafish gastrulation. J Cell Biol 180, 221-232.

Yu, S., Liu, C., Su, K., Wang, J., Liu, Y., Zhang, L., Li, C., Cong, Y., Kimberly, R., Grizzle, W.E., et al. (2007). Tumor exosomes inhibit differentiation of bone marrow dendritic cells. J Immunol 178, 6867-6875.

Zallen, J.A. (2007). Planar polarity and tissue morphogenesis. Cell 129, 1051-1063.

Zhai, L., Chaturvedi, D., and Cumberledge, S. (2004). Drosophila wnt-1 undergoes a hydrophobic modification and is targeted to lipid rafts, a process that requires porcupine. J Biol Chem 279, 33220-33227.

Zhang, X.A., Bontrager, A.L., and Hemler, M.E. (2001). Transmembrane-4 superfamily proteins associate with activated protein kinase C (PKC) and link PKC to specific beta(1) integrins. J Biol Chem 276, 25005-25013.

Zitvogel, L., Regnault, A., Lozier, A., Wolfers, J., Flament, C., Tenza, D., Ricciardi-Castagnoli, P., Raposo, G., and Amigorena, S. (1998). Eradication of established murine tumors using a novel cell-free vaccine: dendritic cell-derived exosomes. Nature medicine 4, 594-600.

Zoller, M. (2009). Tetraspanins: push and pull in suppressing and promoting metastasis. Nat Rev Cancer 9, 40-55. 190

Zoltewicz, J.S., Ashique, A.M., Choe, Y., Lee, G., Taylor, S., Phamluong, K., Solloway, M., and Peterson, A.S. (2009). Wnt signaling is regulated by endoplasmic reticulum retention. PLoS One 4, e6191.

191

Figures 2.1 (A, C-D), 2.2-2.9, and 3.1-3.11 were adapted with permission from Luga V, Zhang L, Viloria-Petit AV, Ogunjimi AA, Inanlou MR, Chiu E, Buchanan M, Hosein AN, Basik M, Wrana JL. Exosomes mediate stromal mobilization of autocrine Wnt-PCP signaling in breast cancer cell migration. Cell. 151, 1542 (2012).

192