The Role and Mechanism of Action of CD200:CD200R1 Interaction in Breast Cancer

Transcription

1 The Role and Mechanism of Action of CD200:CD200R1 Interaction in Breast Cancer by Anna Podnos A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Institute of Medical Science University of Toronto Copyright by Anna Podnos 2015

2 The Role and Mechanism of Action of CD200:CD200R Interaction in Breast Cancer Abstract Anna Podnos PhD Institute of Medical Science University of Toronto 2015 Cancer cells can use immune inhibitory receptors to evade the host s anti-tumour responses and establish immunosuppressive networks in the tumour microenvironment. In this thesis, we investigated the interaction between the immunosuppressive molecule CD200 with its receptor, CD200R1, in breast cancer. We found that CD200 is expressed in the local tumour microenvironment in human breast cancer patients and developed a mouse model to study the effect of CD200 on tumour growth and metastasis. Using mouse-derived EMT6 breast cancer cells and BALB/c female hosts, we explored the effects of overexpressing and silencing the expression of CD200 and CD200R1 in hosts and tumour cells. CD200 expression by host and tumour cells enhanced tumour growth and metastasis to draining lymph nodes (DLN). Silencing CD200 expression in EMT6 tumour cells led to a reduction in primary tumour size and metastasis, as well as an increase in anti-tumour cytotoxic responses in the host. Lack of CD200R1 expression in the host resulted in a marked decrease in breast cancer development and CD200R1 -/- mice were able to mount specific anti-emt6 immune response that could be adoptively transferred to wild type naïve hosts. In addition, we extended our findings to a model in which anti-tumour immunity was explored in EMT6 tumour-bearing ii

3 hosts lacking CD200 expression and treated with a combination of immunotherapy with the nonconventional chemotherapeutic agent, metformin. The findings suggest that CD200 may be an important prognostic marker and a target for breast cancer treatment that could synergize with other therapies and improve outcomes in patients. iii

4 Acknowledgments First and foremost, I would like to express my gratitude to my supervisor and mentor, Reg Gorczynski, for his unconditional support, patience, generosity, and guidance. His passion for science has fueled my interest in biomedical research, and his vivacious energy and sense of humour have motivated me and made my time in his laboratory an enjoyable experience that I will never forget. I would also like to thank my colleagues and lab mates, who have supported me through the ups and downs of research over the last five years. I would especially like to thank our lab manager and a great teacher, Ismat Khatri, who taught me everything I know in the lab and has become a close personal friend and mentor. I am grateful for the encouragement and help from lab members Olha Kos and Fang Zhu, who have been reliable and supportive throughout this journey. Also, I thank Hassan Sadozai and Ramzi Khattar for their help with planning experiments and editing my thesis. I thank my program advisory committee members, Shannon Dunn and David Spaner, for their helpful comments and suggestions that have made this work better and for their guidance in resolving lab related problems. I also thank the Institute of Medical Science, particularly Mingyao Li, Cindy Morshead, and the administrative staff, for helping me navigate complicated situations and complete my thesis. Finally, I am thankful to Gillian Einstein, who has been my mentor for the Collaborative Graduate Program in Women s Health and has expanded my research interests during my PhD. I am eternally grateful to my fiancé, Tom, for believing in me, supporting me through the hard times, celebrating with me in good times, and editing countless copies of my thesis. I also want to thank my parents, who have taught me to work hard and persevere, and my little brother David, who can always put a smile on my face. Last but not least, I would like to give a shoutout to my dogs, Chuba and Xena, for their unconditional love and slobbery kisses. iv

5 Contributions Dr. Reg Gorczynski 1,2 assisted with limiting dilution assays, chromium release assays, and mouse injections. Dr. Ismat Khatri 2 assisted with human and mouse scd200 ELISAs and data analysis. Dr. Kai Yu 2 helped with genotyping and maintaining animal colonies. Dr. Nuray Erin 3 assisted with the mixed lymphocyte culture assays. The Ontario Tumour Bank provided the human breast cancer plasma and serum samples as well as histology sections. 1 Institute of Medical Science, University of Toronto, Toronto, ON, Canada 2 University Health Network, Toronto General Hospital, Toronto, ON, Canada 3 Department of Medical Pharmacology, School of Medicine, Akdeniz University, Antalya City, Antalya, Turkey v

6 Table of Contents Acknowledgments... iv Contributions... v Table of Contents... vi List of Tables... x List of Figures... xi List of Abbreviations... xiv List of Appendices... xix Chapter 1 Introduction and literature overview Introduction and literature overview Cancer and the immune system Immunity and cancer Tumour microenvironment Non-immune cells in the tumour microenvironment Innate immunity in cancer Adaptive immunity in cancer Immune inhibitory receptors in the tumour microenvironment Metastasis and the immune system Metastasis is a hallmark of cancer Immunity and metastasis Breast cancer Clinical features of breast cancer Tumour microenvironment in breast cancer Animal models of breast cancer vi

7 1.4 CD200 and CD200R CD200:CD200R overview CD200:CD200R in cancer EMT6 breast cancer model Objectives and hypotheses Chapter 2 Evidence for a role of tumour CD200 expression in breast cancer metastasis: decreased metastasis in CD200R1 -/- mice or using EMT6 sicd200 breast cancer cells Studies in WT and CD200R1 -/- mice with EMT6, EMT6 CD200tg, and EMT6 shcd200 tumour cells Abstract Introduction Materials and methods Mice Monoclonal antibodies EMT6 breast tumour cells, induction of tumour growth in BALB/c mice, and limiting dilution assays Production and use of lentiviral particles encoding shrna specific for mouse CD Preparation of cells and cytotoxicity, proliferation, and cytokine assays Statistics Results Comparison of metastasis of EMT6 or EMT6 CD200tg to DLN of wild-type (WT), CD200 tg or CD200R1 -/- mice EMT6 CD200tg tumour promotes metastasis of CD200 - tumour cells to DLN Metastasis of CD200 - EMT6 enhanced by EMT6 CD200 depends on host CD200R1 expression Decreased primary growth and absence of metastasis to DLN following injection of EMT6 sicd200 in WT BALB/c is rescued by co-injection with EMT6 CD200tg vii

8 2.4.5 Improved host immunity to EMT6 tumour cells in mice following injection of both EMT6 sicd200 in WT mice and control EMT6 tumour cells in CD200R1 -/- mice Discussion Tables Chapter 3 The role of CD200 expression by tumour and host cells in regulation of immunotherapy in the mouse EMT6 breast cancer model Studies in CD200 -/- and WT hosts with EMT6 and EMT6 sicd200 tumour cells Introduction Materials and methods Mice Tumour cells Growth of EMT6 cells in mice Tumour resection and immunization of mice Immunostaining and flow cytometry Antibodies Proteomics assay ELISArray Cytokine ELISAs Results Characterization of BALB/c CD200 -/- mice Analysis of cell populations harvested from tumours in WT and CD200 -/- BALB/c female mice Altering CD4 and CD8 T cell signaling in WT female mice bearing EMT6 or EMT6 sicd200 tumours Augmentation of adaptive immunity to EMT6 in mice receiving metformin Discussion Chapter 4 Soluble CD200 in plasma and serum of human breast cancer patients viii

9 4 Soluble CD200 levels in breast cancer patients Abstract Introduction Materials and methods Patient Samples Antibodies ELISA of soluble CD200 in plasma Immunological staining mrna extraction from formalin fixed paraffin embedded (FFPE) tissue sections and qrt-pcr Statistics Results Levels of scd200 in plasma of 30 breast cancer patients compared with 10 healthy age-matched women Levels of scd200 in plasma and serum samples from 100 breast cancer patients correlated with clinical characteristics of cancer Comparison of mrna expression and cell surface staining for CD200 and CD200R in breast tumour samples and adjacent normal tissue Discussion Tables Chapter 5 General Discussion and Future Directions References Appendices Bidirectional effect of CD200 on breast cancer development and metastasis, with ultimate outcome determined by tumour aggressiveness and a cancer-induced inflammatory response Cure of metastatic growth of EMT6 tumour cells in mice following manipulation of CD200:CD200R signaling ix

10 List of Tables Table 2-1: Comparative frequency of CD200 + EMT6 tumour cell clones in DLN of mice from Fig. 2-1 (pg. 76) Table 2-2: Comparative frequency of CD200 + EMT6 tumour clones in DLN of mice from Fig. 2-2 (pg. 77) Table 2-3: Comparative frequency of CD200 + EMT6 tumour clones in DLN of mice from Fig. 2-3 (pg. 78) Table 2-4: Comparative frequency of CD200 + EMT6 tumour clones in DLN of mice from Fig. 2-4 (pg. 79) Table 2-5: Comparative frequency of CD200 + EMT6 tumour clones in DLN of mice from Fig. 2-5 (pg. 80) Table 4-1: Demographic data from OTB samples (pg. 124) Table 4-2: Drug use in control volunteers (pg. 125) x

11 List of Figures Figure 1-1: Cells in the tumour microenvironment (pg. 4) Figure 1-2: Anti-tumour immunity vs. tumour-induced immunosuppression (pg. 18) Figure 1-3: The anatomy of a breast (pg. 26) Figure 1-4: Cytokines secreted by cells in the tumour microenvironment regulate breast cancer stem cell renewal (pg. 37) Figure 1-5: CD200 in the tumour microenvironment (pg. 46) Figure 1-6: CD200 expression on the surface of EMT6 cell (pg. 48) Figure 1-7: Predicted outcomes of EMT6 breast tumour growth and metastasis, according to the CD200:CD200R1 model (pg. 51) Figure 2-1: Primary tumour growth and limiting dilution analysis of frequency of cloneable EMT6 cells in DLN of WT, CD200 tg, or CD200R1 -/- mice receiving either control EMT6 or EMT6 CD200tg tumour cells (pg. 60) Figure 2-2: Comparison of primary tumour mass and frequency of tumour cells cloned by limiting dilution from DLN of WT (panel b) or CD200 tg BALB/c mice at 18 days post transplantation of EMT6 tumour cells (pg. 63) Figure 2-3: Differences in primary tumour growth and DLN metastasis of mixtures of control or EMT6 CD200tg tumour cells injected into WT or CD200R1 -/- mice (pg. 65) Figure 2-4: Decreased local growth and frequency of metastasis of EMT6 sicd200 tumour cells in WT mice is attenuated when cells are injected instead into CD200 tg recipients (pg. 67) Figure 2-5: Attenuated local growth and metastasis of EMT6 sicd200 cells to DLN in recipient WT mice is rescued by co-injection of EMT6 sicd200 cells with tumour cells themselves overexpressing CD200 (EMT6 CD200tg ) (pg. 69) xi

12 Figure 2-6: Adaptive immune response in CD200R1 -/- tumour bearing hosts can be transferred to naïve WT mice (pg. 71) Figure 2-7: Importance of CD8 + T cells in specific lysis of EMT6 tumour cells (30:1 effector:target) using DLN cells harvested at 15 days post tumour inoculation into WT or CD200R1 -/- mice (pg. 73) Figure 3-1: Characterization of CD200 -/- mice in the context of EMT6 breast cancer model (pg. 87) Figure 3-2: EMT6 and EMT6 sid200 tumour microenvironment in WT and CD200 -/- female mice (pg. 90) Figure 3-3: Depleting CD8 + T cells in EMT6 and EMT6 sicd200 tumour bearing WT hosts (pg. 92) Figure 3-4: Depleting CD4 + T cells in EMT6 and EMT6 sicd200 tumour bearing WT hosts (pg. 94) Figure 3-5: Metformin augments anti-tumour immunity in CD200R1 -/- EMT6 tumour bearing hosts (pg. 96) Figure 3-6: Verteporfin negates the effect of metformin on EMT6 tumour growth and metastasis in CD200 -/- mice (pg 98) Figure 3-7: Cytokines in the EMT6 and EMT6 sicd200 tumour microenvironment (pg. 100) Figure 3-8: In vivo CD4 cell depletion in WT EMT6 tumour bearing mice (pg. 102) Figure 3-9: Metformin does not improve anti-tumour immunity in WT EMT6 tumour bearing hosts (pg. 103) Figure 4-1: Soluble CD200 levels in plasma of 30 breast cancer patients compared with 10 agematched controls (pg. 115) Figure 4-2: Soluble CD200 levels in plasma and serum of 100 breast cancer patients compared with normal plasma (pg. 117) xii

13 Figure 4-3: Correlation of scd200 levels with clinical characteristics of breast cancer in 100 patient samples (pg. 118) Figure 4-4: Correlation of circulating scd200 levels with ER, PR, and HER2/neu status in 100 breast cancer patients (pg. 120) Figure 4-5: Relative CD200 expression in paired tumour and healthy FFPE tissues from breast cancer patients (pg. 122) xiii

14 List of Abbreviations ADCC AKT ALDH1 ALL AML B7-H1 BCR BCSC BRCA1/2 BTLA CAF CD C/EBP CIA CLL CNS CSC CSF CTL CTLA-4 CXCL CXCR DC DCIS Antibody-dependent cellular cytotoxicity Protein kinase B Aldehyde dehydrogenase Acute lymphoid leukemia Acute myeloid leukemia B7-homologue 1 (PD-L1 or CD274) B cell receptor Breast cancer stem cell Breast cancer susceptibility gene ½ B and T lymphocyte attenuator (CD272) Cancer associated fibroblasts Cluster of differentiation CCAAT-enhancer binding proteins Collagen-induced arthritis Chronic lymphocytic leukemia Central nervous system Cancer stem cells Colony-stimulating factor Cytotoxic T lymphocyte Cytotoxic T lymphocyte antigen 4 (CD152) C-X-C motif chemokine ligand C-X-C motif chemokine receptor Dendritic cell Ductal carcinoma in situ xiv

15 EAE ECM EGF ELISA EMT ER ERK FACS FasL FOXP3 G418 G-CSF GAPDH GH GITR GM-CSF HLA HSC Experimental autoimmune encephalomyelitis Extracellular matrix Epidermal growth factor Enzyme-linked immunosorbent assay Epithelial to mesenchymal transition Estrogen receptor Extracellular ligand-regulated kinase Fluorescence-activated cell sorting Fas ligand Forkhead/winged helix box protein P3 Geneticin Granulocyte colony stimulating factor Glyceraldehyde 3-phosphate dehydrogenase Growth hormone Glucocorticoid-induced TNFR family related gene Granulocyte macrophage colony stimulating factor Human leukocyte antigen Hematopoietic stem cell HER2/neu Human epidermal growth factor receptor 2 HSPG HVEM IDC IDO IFN IGF Heparin sulfate proteoglycan Herpesvirus entry mediator Invasive ductal carcinoma Indoleamine-pyrrole 2,3-dioxygenase Interferon Insulin-like growth factor xv

16 IL inos IRF ITAM ITIM iv Jak LCIS LPS M-CSF MAPK MCP MDSC MHC MIP MLC MMP MMTV MS MSC NFAT NFκB NK NKG2D NOD Interleukin Inducible nitric oxide synthase Interferon regulatory factor Immunoreceptor tyrosine-based activation motif Immunoreceptor tyrosine-based inhibitory motif Intravenous Janus kinase Lobular carcinoma in situ Lipopolysaccharide Macrophage colony stimulating factor Mitogen-activated protein kinase Monocyte chemoattractant protein Myeloid derived suppressor cells Major histocompatibility complex Macrophage inflammatory protein Mixed lymphocyte culture Matrix metalloproteinase Mouse mammary tumour virus Multiple sclerosis Mesenchymal stem cells Nuclear factor of activated T cells Nuclear factor κb Natural killer Natural killer group 2 member D Non-obese diabetic xvi

17 PBL pcr PCR Peripheral blood leukocytes Pathologic complete response rate Polymerase chain reaction PD-1 Programmed cell death 1 pdc PDGF PI3K PKC PR RA RANTES RNS ROS SCID Plasmacytoid dendritic cell Platelet-derived growth factor Phosphatidylinositol 3-kinase Protein kinase C Progesterone receptor Rheumatoid arthritis Regulated on activation, normal T cell expressed and secreted Reactive nitrogen species Reactive oxygen species Severe combined immunodeficiency SDF-1 Stromal cell-derived factor 1 sirna shrna SMA STAT TAA TAM TCR TGF TIL TLR Small interfering RNA Short hairpin RNA Smooth muscle actin Signal transducer and activator of transcription Tumour associated antigen Tumour associated macrophages T cell receptor Transforming growth factor Tumour infiltrating lymphocytes Toll-like receptor xvii

18 TIMP TNBC TNFα VCAM VEGF Tissue inhibitor of metalloproteinases Triple-negative breast cancer Tumour necrosis factor α Vascular cell adhesion molecule Vascular endothelial growth factor xviii

19 List of Appendices Bidirectional effect of CD200 on breast cancer development and metastasis, with ultimate outcome determined by tumour aggressiveness and a cancer-induced inflammatory response (pg. 166) Cure of metastatic growth of EMT6 tumour cells in mice following manipulation of CD200:CD200R signaling (pg. 168) xix

20 1 Chapter 1 Introduction and literature overview This thesis describes studies investigating the role of CD200, an immunoregulatory molecule, and its receptor, CD200R, in breast cancer. Topics in cancer immunity relevant to this thesis are discussed in section 1.1. A summary of important concepts in metastasis is presented in section 1.2. An overview of breast cancer, including clinical features, key components in the breast tumour microenvironment, and animal models of breast cancer, are considered in section 1.3. Literature on CD200, CD200R, and the role of their interaction in cancer immunology are reviewed in section 1.4. Section 1.5 provides an overview of the EMT6 mouse breast cancer model, which is used in the studies presented in the following chapters. Lastly, the objectives and hypotheses of this study are discussed in section Introduction and literature overview 1.1 Cancer and the immune system Immunity and cancer A healthy immune system is important for controlling cancer, and immune suppression associated with malignancies contributes to their progression 1. Tumour infiltrating immune cells are found in most if not all neoplastic lesions, and these inflammatory cells can have both cancer-antagonistic and cancer-promoting effects 2. It is thought that only a minority of all cancers are caused by germline mutations, whereas 90% are linked to somatic mutations and environmental factors 3. Many environmental causes and risk factors of cancer are associated with chronic inflammation. Up to 20% of all cancers have been linked to chronic infections,

21 2 30% are attributed to tobacco smoking and inhaled pollutants (such as silica and asbestos), and 35% are thought to be related to dietary factors (20% of cancer burden is linked to obesity) 4. The existence of a functional relationship between inflammation and cancer was first proposed by Virchow in the 19 th century, when he observed that cancer often arose at sites of chronic inflammation 5, 6. The connection was further supported by findings that tumours resemble wounds that never heal 7. During normal wound healing, immune inflammatory cells appear transiently and then disappear. Conversely, inflammatory cells persist at sites of chronic inflammation, and their presence has been associated in some instances with various tissue pathologies, including fibrosis, aberrant angiogenesis, and neoplasia 2, 3, 8. At least two pathways are thought to link inflammation and cancer. In the intrinsic pathway, activation of different classes of oncogenes drives the expression of inflammationrelated programs, which guide the creation of an inflammatory microenvironment. This, in turn, drives oncogenesis, for example by releasing chemicals that are actively mutagenic (e.g., ROS) 2. In the extrinsic pathway, inflammatory conditions are hypothesized to promote cancer development (e.g., colitis-associated cancer of the intestine, cervical carcinoma with HPV, liver cancer with hepatitis B and C viruses) 1, 9. Inflammation can contribute to many functional hallmarks of cancer through growth factors that sustain proliferative signaling, survival factors that limit cell death, and extracellular matrix-modifying enzymes that facilitate angiogenesis, invasion, and metastasis 2, Tumour microenvironment Solid tumours can be viewed as organ-like structures composed of various cell types, whose interaction is required to drive and promote growth and metastasis (Fig. 1-1). Neoplasiaassociated angiogenesis and lymphangiogenesis help produce a chaotic vascular organization of

22 3 blood vessels and lymphatics, forming the tumour microenvironment 5, which contains immune cells, cancer cells, and their surrounding stroma. The inflammatory component of a developing neoplasm may include a diverse leukocyte population innate immune cells, like neutrophils, NK cells, macrophages, eosinophils, mast cells, as well as adaptive immune cells, including B cells, T cells, and dendritic cells (DC) 5. All of these cells are capable of producing an array of cytokines, chemokines, cytotoxic mediators, such as reactive oxygen species (ROS), serine and cysteine proteases, matrix metalloproteinases (MMPs) and membrane-perforating agents, as well as soluble mediators of cell killing, such as TNFα, interleukins and interferons. The expression of various immune modulators and the abundance and activation state of different cell types in the tumour microenvironment affects the balance between inflammation-promoted tumour growth and anti-tumour immunity 3. On one hand, the immune system specifically detects and targets infectious agents with the adaptive immune response (B and T cells), which is supported by cells of the innate system. On the other hand, the innate system is involved in wound healing and removal of dead cells and cellular debris. Cells of the innate immune system are one of the major sources of angiogenic, epithelial, and stromal growth factors and matrixremodeling enzymes, and these cells may be recruited and subverted to support neoplastic progression. Similarly, subclasses of B and T lymphocytes may facilitate the recruitment, activation, and persistence of wound-healing and tumour-promoting innate immune cells (neutrophils, macrophages) 2. In addition, and perhaps just as important, there are physical barriers that limit the access of the immune cells to the tumour cells 10. These include a dense extracellular matrix (ECM), high interstitial fluid pressure that is caused by a lack of lymphatic drainage, and molecular barriers that are expressed by endothelial cells, including regulator of G protein signaling 5 (RGS5) and endothelin B receptor (ETBR)

23 4 Figure 1-1 (Top) Cells in the tumour microenvironment. (Bottom) Cancer metastasis is a complex process that involves separation of tumour cells from the primary tumour, invasion through surrounding tissues and basement membranes, entry and survival in circulation, lymphatics, or peritoneal space, and arrest in a distant target organ. Adapted from Hanahan & Weinberg,

24 Non-immune cells in the tumour microenvironment Cancer associated fibroblasts Fibroblasts are found across the spectrum of carcinomas, and often are the preponderant cell population of the tumour stroma 2. Some cancer associated fibroblasts (CAFs) are similar to the fibroblasts that create the structural foundation supporting most normal epithelial tissues. Other CAFs are myofibroblasts, which express α-smooth muscle actin (SMA), and whose biological roles and properties differ markedly from those of tissue derived fibroblasts. Myofibroblasts transiently increase in abundance in wounds and are also found at sites of chronic inflammation, contributing to the pathological fibrosis observed in tissues such as lung, kidney, and liver 2. Recruited myofibroblasts and reprogrammed variants of normal tissue-derived fibroblastic cells can enhance cancer cell proliferation, angiogenesis, and metastasis 14, 15. CAFs secrete a variety of extracellular matrix components, which function in the formation of desmoplastic stroma that characterizes many advanced carcinomas 2, 16. For example, CAFs secrete elevated levels of the cytokine SDF-1 (also known as CXCL12) that stimulates carcinoma cell proliferation in vivo, acting through the CXCR4 receptor expressed on the surface of carcinoma cells 17, 18. CAFs also secrete high levels of TGF-β1, PDGF, and VEGF1 in many cancers, including head, neck, cervical, and prostate carcinomas 19 (Fig. 1-2). These soluble mediators modulate an extensive array of cellular functions that can promote tumourigenesis.

25 Endothelial cells Endothelial cells may play an important role in the tumour microenvironment through direct interaction with the tumour cells as well as through their role in the formation of tumour associated vasculature. The angiogenic switch is a process that activates quiescent endothelial cells, causing them to form new blood vessels 20, 21. Tumour associated angiogenesis has been found to depend on several signaling pathways involving ligands of signal-transducing receptors expressed by endothelial cells (e.g., VEGF, angiopoietin, FGF, Notch) 2, 20, 21 (Fig. 1-2). The gene expression profiles of tumour associated endothelial cells are distinct from normal endothelial cells 22, 23. The initial angiogenic switch during tumour development is followed by neovascularization of variable intensity, which is controlled by both the cancer cells and the associated stromal microenvironment 24, Mesenchymal cells Mesenchymal stem and progenitor cells have been found to migrate into tumours from the marrow, where they may differentiate into various characterized stromal cell types, including osteoblasts, chondrocytes, adipocytes, fibroblasts, and myocytes 26. Some of these cells may persist in an undifferentiated (or partially differentiated) state, retaining functions that their more differentiated progeny lack 2. For example, MSCs present in the tumour stroma can secrete CCL5/RANTES in response to signals released by cancer cells. CCL5 then acts reciprocally on the cancer cells to stimulate invasive behaviour 27. Pericytes are a specialized mesenchymal cell type (related to smooth muscle cells), with finger-like projections that wrap around the endothelial tubing of blood vessels. In normal tissues they provide paracrine support signals to the normally quiescent endothelium and

26 collaborate with endothelial cells to synthesize the vascular membrane that helps vessel walls to withstand the hydrostatic pressure of blood flow 2, 28. Pericytes play a role in supporting the tumour endothelium For example, tumour vessels with less pericyte coverage appear more vulnerable to radiation and chemotherapy, suggesting that pericytes are critical to protect ECs and may promote therapeutic resistance 31, 32. Inhibition of platelet-derived growth factor (PDGF) receptor signaling by tumour associated pericytes results in reduced coating of tumour vessels and destabilization of vascular integrity and function 29, 30, 33. This is thought to enhance epithelial-to-mesenchymal transition (EMT, see below), cancer cell intravasation into the circulatory system, and enable cancer cell dissemination 2, 34. In addition, it was recently reported that in glioblastoma, glial stem cells can give rise to vascular pericytes that directly support neoangiogenesis, vessel function, and tumour growth Pericyte and CAF progenitors were found to originate from the bone marrow in various mouse models of cancer 26, 36, 37. Tumour-associated stromal cells in growing tumours may be derived from proliferation of pre-existing stromal cells; by differentiation in situ of local stem/progenitor cells originating in the neighboring normal tissue; or through recruitment of bone marrow-derived stem/progenitor cells Innate immunity in cancer According to the immune surveillance theory, innate immunity is responsible for early detection and elimination of foreign antigens and cancer cells. Innate immunity may be inefficient in patients who develop malignancy 1.

27 Tumour associated macrophages Circulating monocytes give rise to a variety of tissue-resident macrophages throughout the body, as well as to specialized cells such as dendritic cells (DCs) and osteoclasts. Monocytes originate in the bone marrow from a common myeloid progenitor (CMP) that is shared with neutrophils, and they are then released into the peripheral blood, where they circulate for several days before entering tissues and replenishing the tissue macrophage populations 38, 39. Macrophages are phagocytic cells that function in pathogen destruction, inflammation, and tissue repair. Macrophage differentiation results in mature cells polarized toward the M1 or M2 phenotype, which differ in terms of receptor expression, effector function, and cytokine and chemokine production 40. For instance, arginine metabolism is characterized by high levels of inducible nitric oxide synthase (inos) in M1 macrophages, whereas the arginase pathway predominates in M2 cells with generation of ornithine and polyamines 40. Furthermore, cells of the M1 phenotype produce cytokines like IL-12, IL-23, and tumour necrosis factor (TNF), whereas M2 macrophages typically produce IL-10, IL-1 receptor antagonist (IL-1ra) and the type II IL-1 decoy receptor (see Fig. 1-2). Chemokine receptors and ligands are differentially modulated in polarized macrophages. Differential production of chemokines that attract Th1 (e.g. CXCL9, CXCL10) and Th2 or regulatory T (Treg) cells (e.g. CCL22) integrates M1 and M2 macrophages into circuits of amplification and regulation of polarized T-cell responses 40, 41. Classically activated M1 macrophages are potent effector cells that kill microorganisms and tumour cells and produce copious amounts of proinflammatory cytokines. In contrast, M2 cells (induced by IL-10, IL-4 or IL-13, glucocorticoid hormones, vitamin D3) are hypothesized to modulate inflammatory responses and adaptive Th1 immunity, scavenge cellular debris, and promote angiogenesis, tissue remodeling and repair 40.

28 9 Tumour associated macrophages (TAMs) are a significant component of the inflammatory infiltrate in neoplastic tissues and are likely derived from a blood-borne macrophage precursor 42. Macrophages have the potential to express pro- and anti-tumour activity, the former prevailing in established neoplasia 40. Clinical studies have found that high numbers of intra-tumour macrophages correlate with high vessel density and tumour progression 43. TAMs and related cell types in mouse and human tumours generally have an M2 phenotype, which is thought to be oriented towards suppressing adaptive immunity, promoting tumour growth, remodeling tissues, and promoting angiogenesis 40, 44, 45 (Fig. 1-2). TAMs are required to regulate both the angiogenic switch and tumour metastasis in human cancers and in mouse tumour models 46, 47. For example, VEGF-C production by TAMs in human cervical tumours regulated peri-tumoural lymphangiogenesis and dissemination of cancer cells to lymph nodes 40. Macrophages and monocytes are recruited to tumours via chemotactic signaling. Expression of the chemokine CCL2/MCP-1 is frequently observed in tumours (sarcomas, gliomas, melanomas, lung, breast, cervix, and ovary tumours) in keeping with its description as a tumour-derived chemotactic factor 48, 49. CCL2 contributes to regulation of macrophage recruitment to tumours in mouse models of cancer and human malignancies 40, 49. Other CC chemokines related to CCL2, such as CCL7 and CCL8, are also produced by tumours and can recruit monocytes 43, 50. Vascular endothelial growth factor (VEGF) and macrophage-colony stimulating factor (M-CSF) also contribute to macrophage recruitment in tumours and promote their migration, proliferation, and survival 43, (Fig. 1-2).

29 Tumour associated neutrophils Neutrophils defend the host from invading microorganisms and assist in wound healing 54. Invading pathogens elicit an inflammatory response that recruits neutrophils to sites of infection, where they engulf and eliminate microorganisms using an arsenal of cytotoxic substances, proteinases, cytokines, and chemokines 54,55,56. The process of neutrophil recruitment and activation, observed in infection, is recapitulated within the tumour microenvironment. However, accumulating evidence suggests that, within the tumour microenvironment, neutrophils act to the detriment of the host. Neutrophils are the most abundant circulating leukocyte in humans and make up a significant portion of the inflammatory cell infiltrate found in a wide variety of murine cancer models and human cancers 57. Clinical studies have indicated that the presence of tumour associated neutrophils (TANs) confers a poor prognosis 57. For example, in patients with renal cell carcinoma, an increased number of neutrophils correlated with increased mortality 58. In addition, increased levels of neutrophils in the bronchioalveolar space of patients with bronchioalveolar carcinoma were associated with poor outcomes 59. Levels of IL-8, an important chemotactic molecule for neutrophils and macrophages, have been associated with neutrophil accumulation and reduced patient survival 60. Many cell types within the tumour microenvironment are capable of secreting neutrophil chemotactic substances. However, the tumour cells themselves may mediate neutrophil recruitment to sites of tumourigenesis by secreting CXC chemokines (e.g., interleukin-8 (IL-8)), suggesting that TANs are not a means of host defense. Studies report concrete examples of tumour-mediated signals eliciting pro-tumour responses from neutrophils (e.g., in mouse breast cancer and hepatocellular carcinoma) 57. Soluble factors produced by TANs include ROS, which

30 11 can initiate DNA damage and tumour establishment, nitric oxide, proteinases (important for ECM remodeling), elastase, and matrix metalloproneinases Adaptive immunity in cancer When mice genetically engineered to be deficient for various components of the innate and adaptive immune system were assessed for the development of carcinogen-induced tumours, it was observed that tumours arose more frequently and grew more rapidly in immunodeficient mice relative to immunocompetent controls 2. In particular, deficiencies in the development or function of CD8 + cytotoxic T lymphocytes (CTLs), CD4 + Th1 helper T cells, or NK cells each led to demonstrable increases in tumour incidence. Moreover, mice with combined immunodeficiencies in both T cells and NK cells were even more susceptible to cancer development. These results suggested that, at least in certain experimental models, both the innate and adaptive cellular arms of the immune system were able to contribute significantly to immune surveillance and thus tumour eradication 2, 61, 62. Anti-tumour immune responses are often reliant on the immunogenicity of a tumour, which varies greatly between cancer types and individuals 10. Clinical epidemiology also increasingly supports the hypothesis that an anti-tumour immune response exists in various types of human cancer For example, patients with colon and ovarian tumours that are heavily infiltrated with CTLs and NK cells often have a better prognosis than those that lack such abundant killer lymphocytes 65, 66. Additionally, immunosuppressed organ transplant recipients have been observed to develop donor-derived cancers, consistent with the idea that in the ostensibly tumour-free donors, the cancer cells were held in check, in a dormant state, by a fully functional immune system 2, 67 (Fig. 1-2).

31 Natural killer cells Natural killer (NK) cells mediate innate immunity against pathogens and tumours. They were originally discovered because of their ability to kill certain tumour cells in vitro 68. Studies in leukemia, lymphoma, and gastrointestinal stromal tumours support the idea that NK activation and cytotoxicity influence patient outcome Low NK cell activity has been reported in familial breast cancer patients as well as their clinically asymptomatic first degree relatives, suggesting that NK activity has an important hereditary and genetic component 72. Depletion of NK cells in vivo leads to enhanced tumour formation in mouse tumour models 73. Although several NK cell receptors have been implicated in the killing of tumours 74, the activating NKG2D receptor is unique because it recognizes defined antigens (e.g. stress-induced proteins) that are frequently overexpressed on many different tumours, indicating a role for this receptor in immune surveillance against cancer. Tumours have developed a number of strategies that might allow them to escape effective NK suveillance. Tumours can shed soluble ligands for activation receptors, secrete immunosuppressive cytokines, such as TGFβ and IL-10, which could impair NK cell or T cell effector functions, and produce the apoptotic Fas ligand, which can also lead to elimination of NK cells. Furthermore, not all cells within a tumour mass might express NKG2D ligands, and expression of an NKG2D ligand alone on tumours might be insufficient to activate NK cells. Finally, NK cells might not traffic to the primary tumour site, since histological examination of tumours rarely indicates the presence of NK cells 68 (Fig. 1-2).

32 T cells T cells recognize peptides that are presented by human leukocyte antigen (HLA, also known as major histocompatibility complex (MHC)) molecules 10. Tumour specific antigens recognized by antibodies and T cells have been identified in cancer patients, yet their existence often does not control malignant growth 45. T lymphocyte precursors capable of responding to self-antigens, which are often overexpressed in tumours, are detectable at low levels in the circulation of most individuals 1 and are enriched in tumour tissues 75. Despite the TIL activation phenotype, they are functionally compromised and are enriched in TAA-specific memory T cells 1. There are many cellular and molecular mechanisms that mediate tumour escape from natural immune surveillance Studies of melanoma antigen (MLANA)-specific T cells confirmed that T cells specific for tumour antigen are rendered functionally tolerant once they are present in the tumour microenvironment 79, 80. Furthermore, these studies suggested that this locally induced tolerance may be reversible 81. Reduced expression of the major histocompatibility complex (MHC) on the cancer cell membrane and impaired antigenic peptide expression on the tumour cell surface are examples of mechanisms that can attenuate direct recognition of tumour antigens by T cells and direct priming of an immune response by a tumour 82. As a result, tumours can evade the host s immune response by being poor stimulators of T cells and by being poor targets for tumour specific CTLs 1. Tumours also directly interfere with the host immune system through release of factors that modulate the functions of immune cells or induce of apoptosis of these cells 1, 82. In addition to the wide variety of soluble immunosuppressive factors (TGFβ, IL-10, ROS, enzymes, and inhibitory ligands, such as FasL or TRAIL) that are released by tumour cells or other cells in the

33 tumour microenvironment, immune suppressor cell populations have been shown to play a key role in down-regulation of anti-tumour host immunity 83, 84. Conversion of naïve T cells into adaptive Treg cells is facilitated by direct release from tumours of TGF-β, IL-10, and indoleamine 2,3-dioxygenase (IDO), an immunosuppressive enzyme that depletes tryptophan, or through a tumour derived activation of release of such molecules from myeloid-derived suppressor cells (MDSC), TAMs and/or dendritic cells (DC) Tregs are a population of cells that play a critical role in the induction and maintenance of immunologic tolerance. Treg populations are heterogeneous and differ in phenotype, cytokine secretion profile, and suppressive mechanism. Subsets include Tregs that depend on cytokine production (IL-10 or TGFβ) for their function; CD4 + CD25 + FOXP3 + Tregs (naturally occurring or antigen-induced); and various other regulatory cells, including NKT cells, CD8 + Treg, γδtcr+ cells, and DN T cells. These cells can suppress tumour-specific T cell immunity, hamper NK cell activation and cytotoxicity, and inhibit maturation of DC, thus contributing to the progression of human and mouse tumours 86, 87. Depletion of Tregs in mice with anti-il-2r antibody or with low dose cyclophosphamide improves T cell-mediated tumour clearance, and depletion of CD4 + CD25 + T cells has been shown to promote tumour rejection in mice 88, 89, 90. In humans, Tregs can be defined as FOXP3+, CTLA4+, GITR+ T-cell subsets, and they are enriched among TIL in human tumours and are more abundant in the peripheral circulation of patients with cancer than of healthy controls 91. Tregs accumulate in both human and mouse tumours, as well as in secondary lymphoid organs 92, and are recruited and expanded by either the proliferation of preexisting Tregs 93 or the conversion of CD25-negative T cells 85, 94. In some tumour types, including breast cancer and hepatocellular carcinoma, increased numbers of Treg cells correlate with reduced overall survival 95, 96, whereas in other types, such as colorectal cancer and some

34 15 hematological malignancies, Treg cells are associated with improved survival 97, 98. These conflicting results may be due to the fact that in many of these studies, the function of FOXP3 + cells was not tested, and not all FOXP3 + T cells are functionally suppressive 98. It is possible that the effects of Tregs on the adaptive immunity or the differentiation of FOXP3 + cells into functional Tregs differs in various cancers. Alternatively, it is possible that the appropriate T cell compartment was not studied in the context of lymphoid malignancies. The utility of FOXP3 content and Treg function as prognostic factors in different cancers is an area of intense investigation, but much remains to be understood 98. Moreover, the controversial relationship between Treg populations and patient prognosis may be further impacted by other inflammatory cell subsets in the tumour microenvironment 99. TGF-β, a cytokine produced by Tregs, has been implicated in tumour progression. The TGF-β pathway has been linked to metastatic processes and has been shown to dramatically impact the ability of tumour cells to spread throughout the body 15, 100, 101. Tregs may express cell surface molecules that deliver negative signals to DC, including cytotoxic T lymphocyte antigen 4 (CTLA4) and lymphocyte activation gene 3 (LAG3). These signals inhibit the maturation of dendritic cells, block their expression of MHC and co-stimulatory molecules (CD80 and CD86) 102, activate their ability to produce IDO, and indirectly suppress genes encoding IL-6 and TNF 103. In addition, the release of adenosine and the secretion of TGF-β, IL-10, and IL-35 by Tregs may interfere with the activation and effector functions of T cells 102, 104. Finally, secretion of granzymes and perforin might have cytolytic effects on target T cells DC 82 (Fig. 1-2) B cells B cells can mediate tumour immunity by secreting pro-tumourigenic cytokines and altering Th1-to-Th2 ratios 105. Their importance in supporting tumour growth is evident in B cell

35 16 deficient mice, which exhibit resistance to engraftment of certain syngeneic tumours 106. Antibodies to tumour associated antigens (TAA) are often detectable in patients with cancer 107 and have been used as a biomarker of prognosis, as is the case with, for example, antibodies to p53 in a subgroup of patients with head and neck cancers 108. Regulatory B cells (Bregs) have been shown to play a role in tumour progression to metastasis 109. Similar to Tregs, the regulatory function of B cells is exerted via the production of regulatory cytokines, such as IL-10 and TGF-β, and the ability to express inhibitory molecules that suppress pathogenic T cells and autoreactive B cells in a cell-to-cell contact-dependent manner 110. Tumour-evoked Bregs induce the TGF-β-dependent conversion of resting CD4 + T cells to immune-suppressive FOXP3 + Tregs 111. Tumour-promoting B cells have been shown to facilitate the conversion of M1 macrophages to a tumour-promoting M2 phenotype through IL- 10 secretion 112. Additionally, B cells promote lymphangiogenesis and may actively promote metastasis, as has been shown for B cell-mediated lymphangiogenic metastasis in lymphoma and melanoma 113, 114 (Fig. 1-2) Dendritic cells Dendritic cells (DCs) are monocytic antigen-presenting cells that are derived from the bone marrow. DCs have been detected in several tumour types, including breast, lung, prostate, kidney, and ovarian carcinomas and melanoma Several subtypes of DC have been identified to infiltrate tumours, including Langerhan-like DC (positive for the marker Langerin), plasmacytoid DC (pdc), and myeloid-derived DC 43. DC can take up antigens and cross-present them to T cells, modulating their activity. However, local DC maturation is often blocked in the tumours by combinations of tumour-derived cytokines, such as M-CSF and IL-6, resulting in immunosuppresion.

36 17 An aberrant balance between immature and mature myeloid cells is a hallmark of cancer 118. Immature myeloid cells with specific inhibitory activities, called myeloid-derived suppressor cells (MDSCs), negatively regulate the immune response by suppressing CTL and NK cell activity 119. Although initially described in cancer patients, MDSCs are also present in other inflammatory settings, including solid organ transplantation, where tolerance is thought to be dependent, in part, on MDSCs that accumulate in the allografts 120, 121. In preclinical mouse models, the phenotype of MDSCs consists of co-expression of the myeloid lineage differentiation antigens Gr-1 (Ly6G) and CD11b (CR3, Mac-1) 122, 123. In contrast to murine models, the phenotype of MDSCs in humans is not as well defined 119. Identification of human MDSCs has been complicated by the lack of a specific marker and by the absence of a human homologue of mouse Gr Typically, human MDSCs are positive for the markers CD33 and CD11b but express low levels of HLA-DR 125. Accumulation of MDSCs in both preclinical models and in human samples was associated with defective dendritic cell function and inhibition of antigen specific T cell responses Several mechanisms have been described by which MDSCs suppress T cell responses. One is the depletion of specific amino acids such as l-arginine (l-arg), l-cysteine, or l- phenylalanine 124, which can inhibit expression of the CD3ζ chain and, thereby, T cell proliferation, causing arrest of T cells in the G0 G1 phase of the cell cycle 129. The production of ROS, such as H2O2, is another mechanism that affects immune regulation by inhibiting T cell proliferation 124. Furthermore, reactive nitrogen species (RNS) are produced when superoxide anion interacts with NO 130 and nitrate aromatic amino acids (such as tyrosine residues) in the T cell receptor (TCR) and CD8, resulting in a decreased recognition of peptide MHC (major histocompatibility complex) by the TCR 131. MDSCs can also secrete immune inhibitory cytokines that inhibit the activation of CD and CD8 +123, 126 T cells, attenuate the

37 cytotoxicity of NK cells 133, and polarize immunity toward a tumour-promoting type 2 phenotype through downregulation of IFN-γ and upregulation of IL , 134 (Fig. 1-2). 18 Figure 1-2 Anti-tumour immunity vs. tumour-induced immunosuppression. The balance between anti-tumour immunity and tumour-mediated immune suppression is affected by the abundance and activation state of many different cell types in the tumour microenvironment. Tumour cell specific antigens can be recognized by cytotoxic immune cells, leading to their destruction. Fibroblasts and macrophages within the tumour microenvironment contribute to a growth-suppressive state; however, these cells may later become educated by the tumour to acquire pro-tumourigenic functions. TAMs support tumour growth, angiogenesis, and invasion by secreting a pro-tumourigenic proteases, cytokines and growth factors (for example, EGF, which participates in a paracrine signaling loop through tumour-secreted CSF-1). As tumours grow, immune-suppressor cells, including MDSCs, Bregs, and Tregs are mobilized into the circulation in response to activated cytokine axes that are induced by tumourigenesis. MDSCs and Treg cells infiltrate the growing tumour to disrupt immune surveillance through multiple mechanisms, including disruption of antigen presentation by DCs, inhibition of T and B cell proliferation, and modulation of NK cell cytotoxicity. CAFs, which become activated by tumour-derived factors (for example, SDF-1, TGF-β, FGF or PDGF), secrete ECM proteins and

38 19 basement membrane components, regulate differentiation, modulate immune responses, and contribute to deregulated homeostasis. CAFs are also a key source of VEGF, which supports angiogenesis during tumour growth. Adapted from Quail & Joyce Immune inhibitory receptors in the tumour microenvironment There is growing evidence for a role for immunosuppressive molecules expressed by tumour cells themselves, or by host cells under the control of these tumour cells, in activating endogenous immune inhibitory pathways leading to tumour growth and suppression of the inflammatory responses of immune cells in the tumour microenvironment 136. Immune inhibitory receptors prevent overactivation and dysregulation of immune cells, which may cause extreme lymphoproliferation, chronic inflammation and autoimmunity. These pathways have recently become therapeutic targets to strengthen anti-tumour responses 137, since a number of immunoregulatory molecules are overexpressed by cancer cells of different tissue origins. The co-inhibitory molecule of the B7 family, B7-H1 (also known as PD-L1), is overexpressed in multiple myeloma, leukemia, ovarian cancer, and breast cancer 138. The B7 family consists of structurally related, cell-surface protein ligands, which bind to receptors on lymphocytes that regulate immune responses. B7 ligands deliver 'costimulatory' or 'coinhibitory' signals through the CD28 family of receptors on lymphocytes 139. In ovarian cancer, the inhibitory receptor for B7-H1, PD-1, was found on tumour infiltrating CD8 + T cells carrying TCRs with specificity for tumour-associated antigens but with defective effector functions 140. In breast cancer, a concurrent and abundant infiltration of T cells expressing the B7-H1 and FOXP3 molecules in the tumour microenvironment has been observed in high-risk patients 141.

39 Another co-inhibitory receptor, B and T lymphocyte attenuator (BTLA), was found to be expressed at high levels on human melanoma tumour antigen-specific effector CD8 + T cells that were susceptible to inhibition by the BTLA ligand HVEM 142. BTLA + CD8 + T cells from the tumour microenvironment were shown to produce less IFNγ than their BTLA - counterparts 143. These studies support the hypothesis that immunoregulatory molecules expressed by tumour cells modulate anti-tumour immune response. 20 Clinical trials with blocking antibodies against inhibitory receptors, like CTLA4 (the inhibitory receptor for CD80 and CD86) and PD-1, have yielded promising results for the treatment of various cancers (e.g. melanoma, gastric and esophageal adenocarcinomas), and CTLA4 blockade is already in use for stage IV melanoma patients 144, 145. Immunoregulatory molecule blockade is a promising new approach to cancer therapy, and many other inhibitory immune interactions remain to be characterized. This thesis focuses on characterization of another inhibitory receptor, CD200R (and its ligand CD200), in the regulation of breast cancer, in both mouse models and humans. 1.2 Metastasis and the immune system Metastasis is a hallmark of cancer Metastatic disease that is resistant to therapy is the major cause of death from cancer 146. The multistep process of invasion and metastasis has been schematized as a sequence of discrete steps, referred to as the invasion-metastasis cascade 147. This succession of events begins with local invasion, then intravasation by cancer cells into nearby blood and lymphatic vessels, and transit of cancer cells through the lymphatic and hematogenous systems. This is followed by

40 extravasation of cancer cells, the formation of micrometastases, and finally the colonization of distant organs by macroscopic tumours 2 (Fig. 1-1). 21 Epithelial-mesenchymal transition (EMT) is a process in which transformed epithelial cells lose their cell polarity and cell-cell adhesion and acquire the abilities to invade, resist apoptosis, and disseminate By co-opting a process involved in various steps of embryonic morphogenesis and wound healing, cancerous cells are hypothesized to acquire concomitantly multiple attributes that enable invasion and metastasis 2. These phenotypic changes are driven by alterations in signal transduction pathways that activate EMT-associated transcription factors, such as Snail, Zeb, and Twist family members, which directly alter epithelial gene expression programs, particularly suppression of E-cadherin adhesion protein encoding gene Cdh1 151, 152. Interestingly, EMT has been reported to promote the generation of CAFs, through dedifferentiation of tumour cells of epithelial origin (for example, in breast and prostate cancers) to generate a mesenchymal-like cell population that expresses CAF markers 135, 153. In addition to the alterations that occur in the primary tumour microenvironment, roles for host-derived cells and mediators at sites distal to the tumour have also been reported 154. An important concept in tumour metastasis is the formation of premetastatic niches, in which malignant tumours prepare the environment of remote organs to receive metastatic cells by altering host homeostasis in these organs before tumour cell arrival 155. MDSCs have been identified as an important component of these premetastatic niches To metastasize, the tumour must possess certain unique characteristics, including the ability to penetrate the endothelium and acquire mobility within tissues as well as lymphatics or blood vessels 159. Not surprisingly, solid tumour cells are able to adopt the phenotypic characteristics of lymphoid cells that enable them to migrate using similar mechanisms 160.

41 22 Circulating tumour cells are particularly sensitive to lysis by NK cells or monocytes. In the presence of anti-tumour Abs, these effector cells of innate immunity are also able to mediate antibody-dependent cellular cytotoxicity (ADCC), thus efficiently eliminating tumour targets 161. A tumour cell that manages to avoid such immune intervention in the peripheral blood or lymphatic circulation and arrives at a new tissue site is dependent on the local microenvironment for growth factors and structural support by the extracellular matrix (ECM), which can be produced by immune cells responding to TAA, thus promoting metastasis formation Immunity and metastasis Communication between cancer cells and the stromal cells in the tumour microenvironment is involved in the acquisition of invasive growth and metastatic capabilities 2, 105, 162, 163. Tumour-infiltrating immune cells secrete cytokines and chemokines that can promote EMT, including CAFs, macrophages, MDSCs, granulocytes, and lymphocytes 109, Macrophages in the tumour microenvironment of pancreatic islet cancers, mammary tumours, and lung metastases were activated by IL-4 secreted by cancer cells to produce high levels of cathepsin protease, which is thought to promote invasion 167. TNF-α-stimulated MSCs can secrete CCL2, which enhanced recruitment of monocytes, macrophages, and neutrophils to the tumour in a mouse lymphoma model, and this promoted malignancy through macrophagedependent mechanisms 168. Moreover, TNF-α secreted by TAMs activated NF-κB-mediated transcription of Snail1 and Zeb, leading to diminished E-cadherin expression on pancreatic carcinoma cells 169. Leukocyte-secreted cytokines, such as IL-6 and IL-23, can also initiate EMT via the activation of STAT3 signaling resulting in inhibition of E-cadherin expression and loss of cell cell adhesion in breast cancer cells 109, 170. TAMs, MDSCs, and Tregs can also produce

42 23 large quantities of TGF-β and are important inducers of EMT 171. Furthermore, cancer invasiveness can be regulated by interactions with immune inflammatory cells that assemble at the boundaries of the tumours, producing the extracellular matrix-degrading enzymes and other factors that enable invasive growth 105, 163, Breast cancer Clinical features of breast cancer With an estimated 1.15 million new cases each year, female breast cancer is the second most common cancer in the world and the most common cancer among women, accounting for 411,093 cancer deaths per year 173, 174. It is estimated that in 2014, 24,400 women will be diagnosed with breast cancer in Canada (1 in 9 women over their lifetime), which represents 26% of all new cancer cases in Canadian women 175. This disease is the 2 nd leading cause of death from cancer in women (following lung cancer), and it kills 1 in 29 Canadian women 176. Breast cancer can also occur in men, but it is rare only about 1 in 150 breast cancers is diagnosed in males 177. Breast cancer incidence has increased steadily in developed countries over the past few decades, but the mortality caused by breast cancer has decreased in recent years. This could be caused by earlier detection of breast cancer, improved screening techniques, surgical and radiotherapy interventions, understanding of the pathogenesis of the disease, and the use of chemotherapies in a more efficacious manner 178. Despite local and systemic therapies, women with breast cancer are at risk of developing metastases (often in bones, lungs, liver, and brain) throughout their life, and approximately 30% of breast cancer

43 24 patients relapse 179. Macroscopic metastases may erupt from dormant micrometastases decades after a primary tumour has been surgically removed or pharmacologically destroyed 2, 180, 181. Several factors have been associated with an increased risk of breast cancer, including family history, nulliparity, early menarche, advanced age, and a personal history of breast cancer 182, 183. Five to 10% of all women with breast cancer may have a germline mutation of BRCA1 and BRCA2 genes; women with BRCA1/BRCA2 mutations have a 40% to 85% lifetime risk of developing breast cancer 184 and an increased risk of ovarian and other cancers 185, 186. The risk of a primary breast cancer in the contralateral breast ranges from 3% to 10% at 10 years after diagnosis 187, but for BRCA1/2 mutation carriers diagnosed at a young age, the risk of contralateral breast cancer can reach 50% in the following 25 years 188, 189. Breast cancer is commonly treated by various combinations of surgery, radiation therapy, chemotherapy, hormone therapy, and immunotherapy. Several clinical and pathological features can influence the combination of therapies the patient receives, such as the patient s age and menopausal status; disease stage; tumour size; lymph node status; histologic and nuclear grade of the primary tumour; estrogen receptor (ER) and progesterone receptor (PR) status of the tumour; human epidermal growth factor type 2 receptor (HER2/neu) overexpression; and the proliferative capacity of the tumour 190, 191. About 60 to 70% of human breast cancers are ER-positive and estrogen-dependent 192, and ER and PR expression are an important indicator of potential responses to hormonal therapy 193. Based on the molecular classification of breast cancers, ER + tumours fall under luminal A (~40%; ER high, HER2 low ) and luminal B subtypes (~20%; ER low, HER2 low ), with luminal A subtype having a better outcome 194, 195. The ER + group is the most numerous and diverse, with several genomic tests to assist in predicting outcomes for ER + patients receiving

44 endocrine therapy 196, 197. ER is a member of the superfamily of nuclear receptors that function as ligand-inducible transcription factors 198 and mediates the biological (proliferative) effects of estrogens (steroid sex hormones) 199. Clinical epidemiological studies have shown a strong correlation between the actions of ovarian steroid hormones, particularly 17β-estradiol, and carcinogenesis in the mammary gland and uterus 200, 201. In addition to endogenous estrogens in the body, aromatase enzyme can aromatize androgens into estrogens. Hormone therapies include ER modulators, like tamoxifen or raloxifene, which block ERs on breast cancer cells, and aromatase inhibitors that lower the amount of estrogen in the body by stopping the conversion of other hormones into estrogens Human epidermal growth factor receptor HER2 overexpression is present in approximately 20 30% of breast cancer tumours. HER2 overexpression is associated with a more aggressive disease, higher recurrence rate, and shortened survival 203.The discovery of HER2 (also called ERBB2) group 204 is of interest given the effective therapeutic targeting of HER Trastuzumab is a humanized monoclonal antibody targeting the HER2 receptor, which was approved for use in The mechanisms of action of trastuzumab include extracellular mechanisms involving antibody-dependent cellular cytotoxicity (ADCC), and intracellular mechanisms involving apoptosis and cell cycle arrest. Inhibition of angiogenesis, and prevention of DNA repair following chemotherapy-induced damage may also play a role in the mechanism of action 205, 206. Trastuzumab has been shown to be effective in combination with chemotherapy, for the treatment of early stage and metastatic HER2 positive breast cancer 203. The American Joint Committee on Cancer (AJCC) has designated disease staging by TNM classification to define breast cancer 207 ; T describes the size of the primary tumour, N describes the regional lymph node involvement, and M describes distant metastasis. The stage of breast cancer is defined by the combination of these three factors: stage 0 (Tis, N0, M0),

45 stages IA-B (T0-1, N0-1, M0), stages IIA-B (T0-3, N0-1, M0), stages IIIA-C (T0-4, N0-3, M0), stage IV (any T, any N, M1). 26 Figure 1-3 The anatomy of a breast. Adapted from Canadian Cancer Society. Out of several histologic classification for breast cancer (ductal, lobular, nipple see Fig. 1-3 for breast anatomy), the infiltrating or invasive ductal carcinoma (IDC) is the most common type, comprising 70% to 80% of all cases 208. Ductal carcinoma in situ (DCIS) (Stage 0 breast cancer) is a noninvasive condition, but it can sometimes progress to become an invasive cancer, like IDC. Surgery (breast-conserving lumpectomy or total mastectomy) is a common treatment option for patients with DCIS; there is a 25% to 50% incidence of recurrence following limited surgery for palpable tumour, with 50% of those recurrences being invasive carcinoma 209, 210. Usually, surgery is accompanied by radiation therapy, although not every patient benefits from it, and researchers are looking for a way to identify these unresponsive subsets of patients to avoid unnecessary postoperative radiation therapy In addition, patients with DCIS may receive a regimen of tamoxifen 214, an ER antagonist. Clinical trials

46 have shown that with tamoxifen, ipsilateral invasive breast cancer decreased from 4.2% to 2.1% at 5 years (P=.03), and the incidence of contralateral breast neoplasms (invasive and noninvasive) decreased from 0.8% per year to 0.4% per year (P =.01) Lobular neoplasia, or lobular carcinoma in situ (LCIS), is often a multifocal and bilateral disease that identifies women at an increased risk for subsequent development of invasive breast cancer. Most women with LCIS have disease that can be managed without additional local therapy after biopsy. The use of tamoxifen has been shown to decrease the risk of developing subsequent breast cancers in women with LCIS 216. Stage I, II, IIIA, and operable IIIC breast cancer often requires a multimodal approach to treatment, depending on ER, PR and HER2/neu status of the primary tumour 217. Treatment options include different combinations of lumpectomy/mastectomy, radiation therapy, axillary lymph node surgery, adjuvant or neoadjuvant chemotherapy, hormone therapy (tamoxifen, aromatase inhibitors), and targeted therapy like monoclonal antibodies (trastuzumab for HER2 + patients). Stage IIIB, inoperable IIIC, IV, recurrent, and metastatic breast cancer may often be responsive to different combinations of therapies, but the treatment for metastatic breast cancer is rarely curative. Triple-negative breast cancer (TNBC) is defined by the absence of staining for ER, PR, ad HER2/neu. TNBC is unresponsive to some of the most effective therapies available for breast cancer treatment, including HER2-directed therapy, such as trastuzumab, and endocrine therapies, such as tamoxifen or the aromatase inhibitors. It is characterized by diffuse erythema and edema, no palpable mass, early age at diagnosis, poor nuclear grade, and poor survival outcome 218. Combination cytotoxic chemotherapy remains the standard therapy for early-stage TNBC 219.

47 28 In Chapter 4 of this thesis I will describe data from our studies suggesting a novel classification of human breast cancer, which takes into account expression of the CD200 molecule. Preliminary data suggest this may highlight a previously unrecognized heterogeneity in this human cancer population, and this may open up new avenues of treatments Tumour microenvironment in breast cancer Components of the breast tumour microenvironment Cancer stem cells Cancer stem cells (CSCs) are defined by their ability to initiate tumours in immunocompromised mice upon serial passage (a demonstration of self-renewal) as well as by their ability to differentiate into the non self-renewing cells forming the tumour bulk Epigenetic changes that occur during carcinogenesis often affect stem cell regulatory pathways, such as Notch, Hedgehog, Wnt, PI3K, NF-κB, and Jak/STAT pathways In breast cancer, these pathways are frequently dysregulated by signals from the tumour microenvironment 228. MSCs, CAFs, endothelial cells, and various immune cells interact with CSCs via networks of growth factors and cytokines that form positive feedback loops that promote tumour cell invasion and metastasis 27, 229, 230 (Fig. 1-4). Breast cancer stem cells (BCSC) have been found to express specific cell surface markers CD44 + CD25 -/lo (human) and CD24 + CD29 hi CD49f hi (mouse) 222, 231, Mesenchymal stem cells In human breast cancers, MSCs may be recruited from the bone marrow 229 or from the normal breast stroma, where they integrate into the tumour-associated stroma and increase the

48 potential for metastasis 27. Mobilization of MSCs into the circulation of patients with advanced highly vascularized breast cancer is associated with chemoresistance 233. Aldehyde dehydrogenase 1 (ALDH1) is one of several cancer stem cell markers found in breast cancer, and its expression also identifies MSCs that are selectively recruited to sites of growing tumour. MSCs interact with breast cancer stem cells (BCSCs) through cytokine loops involving IL-6 and CXCL7 229, which stimulate the self-renewal of BCSCs 234. Furthermore, CAFs, TAMs, and MSCs have been shown to secrete IL-6, IL-8, and CXCL7, which in turn activate STAT3/NFκB signaling, leading to self-renewal of BCSCs 229 (Fig. 1-4). This generates a positive feedback loop between the tumour microenvironment and tumour cells. Immunohistochemical analysis has supported the idea of the existence of such MSC:BCSC interactions in biopsies obtained from breast cancer patients 229. Expression of stem cell markers such as ALDH1 in breast cancer cells was shown to be an independent predictor of poor outcome in women with breast cancer 221. In addition, MSCs have the ability to differentiate into adipocytes and CAFs, which can interact with and influence tumour cells Cancer associated fibroblasts The gene expression profile of CAFs (activated fibroblasts within desmoplastic lesions that are associated with malignant tumours 15 ) resembles that of wound-activated fibroblasts, and this profile is associated with poor prognosis in breast cancer 236, 237. Breast CAFs confer a mesenchymal-like phenotype and can enhance the metastasis of both premalignant and malignant mammary epithelial cells, whereas normal fibroblasts promote an epithelial-like phenotype and suppress metastasis 135, 238. Cytokines, such as SDF-1 produced by breast CAFs (but not normal fibroblasts), may promote proliferation of tumour cells, which express the SDF-

49 1 receptor CXCR4 17. High levels of expression of SDF-1 and CXCR4 have been associated with poor survival in breast cancer patients 239, 240. Additional factors produced by cells in the tumour microenvironment that regulate tumour proliferation, invasion, and metastasis include insulinlike growth factor (IGF), platelet-derived growth factor (PDGF), Wnt, Notch ligand, Hedgehog ligands, and matrix metalloproteinases (MMPs) CAFs in the breast tumour microenvironment can select for bone-specific metastatic characteristics in primary tumour cells due to selective interaction between breast cancer cells with high Src activity with CAFs that secrete CXCL12 and IGF CAFs in breast cancer can secrete exosomes, which are vesicles nm in size that carry small signaling molecules, including mrnas and micrornas 248. CAF-derived exosomes can promote breast cancer cell migration through WNT-PCP signaling Endothelial cells Endothelial cells have an important function in the tumour microenvironment by direct interaction with the tumour cells as well as by their role in blood vessel formation 250, 251. Newly formed blood vessels carry oxygen and nutrients to the growing tumours, facilitating growth and metastasis. Endothelial cells also produce a variety of cytokines that influence in breast tumour progression and can directly regulate BCSCs 252, 253 (Fig. 1-4). Although many pro-angiogenic factors have been identified, vascular endothelial growth factor (VEGF) is the primary mediator of angiogenesis 254, and, as a result, it has been the principal target of anti-angiogenic therapies. A humanized monoclonal antibody targeting VEGF (bevacizumab) as well as multi-kinase VEGF inhibitors (sorafenib and sunitinib) are currently approved for clinical use in a number of tumour types, including renal and colon cancers 255, 256. Bevacizumab was initially approved for use in metastatic breast cancer on the basis of reports demonstrating that it prolonged the time of

50 tumour progression when used in combination with paclitaxel 257. More recently, however, it was found that the effect on tumour progression was limited, and that the addition of bevacizumab to cytotoxic chemotherapy failed to increase patient survival Tumour associated macrophages T cells and monocytes that differentiate into TAMs at the tumour site are the most common breast tumour infiltrating leukocytes Genes associated with leukocyte or macrophage infiltration, like CD68, have been identified as part of molecular signatures that correlate with poor prognosis in breast carcinomas 264. High TAM density is a strong negative prognostic indicator in breast cancer. There is a positive relationship between high levels of TAMs and lymph node metastases in breast carcinoma, and the density of TAMs is associated with clinical aggressiveness 259, 261, 263. In addition, in human breast and oesophagus cancers, CCL2 levels correlated with the extent of macrophage infiltration, lymph node metastasis, and clinical aggressiveness 265, 266. The tumour-promoting properties of TAM may be related to their ability to express growth factors for breast cancer cells, angiogenic mediators, extracellular 6, 48, 262, matrix-degrading enzymes, inflammatory cytokines, and reactive oxygen intermediates 267, 268. In an experimental model of metastatic breast cancer, TAMs supplied epidermal growth factor (EGF) to breast cancer cells, while the cancer cells reciprocally stimulated the macrophages with CSF-1 (Fig. 1-2). The interactions of TAMs and cancer cells were hypothesized to facilitate extravasation into the circulatory system and metastatic dissemination 163, 269. Blocking TAMs with clodronate liposomes or through ablation of the CSF1 gene significantly interfered with extravasation and metastatic outgrowth of breast tumour cells in the lung 270. In addition, VCAM-1 positive breast cancer cells bound to VLA-4 (also

51 32 called integrin α4β1)-expressing macrophages during metastasis of to the lung 271. This interaction activated phosphatidylinositol 3-kinase (PI3K)-AKT signaling in breast tumour cells, protecting them from caspase-induced apoptosis, and interruption of this interaction rendered metastatic cells susceptible to apoptotic insult 271. Interestingly, VCAM-1 also interacts with a different integrin partner, α4β1, in osteoclasts, which can contribute to bone metastasis T cells In contrast to TAMs, the correlation between the extent and type of T cell infiltration in the primary tumour and breast cancer progression is less clear 263. A tumour-directed immune response involving cytolytic CD8 + T cells, Th1 cells, and NK cells appears to protect against tumour progression, while an immune response involving the activation of humoral immunity and a Th2 polarized response likely promotes tumour development 178. T cell anti-tumour responses are reported to be impaired in advanced stages of breast cancer and tumour-specific T cells can undergo inhibitory regulation and become anergic in tumour-bearing hosts Increased Treg levels have been observed in the peripheral blood, primary tumour, and draining lymph nodes (DLN) of breast cancer patients 277. Studies have shown that FOXP3 + cells were associated with a more advanced disease in breast cancer, consistent with findings in many other cancers 278. It was also found that intratumoural FOXP3 expression had a linear association with invasion, size, and vascularity in breast carcinoma 279. Metastatic breast cancer was found to be associated with an expansion of peripheral blood CD4 + CD25 high FoxP3 + GITR + CD152 + Treg, whose immunosuppressive properties do not differ from those of healthy subjects 178, 280. Furthermore, FOXP3-positive Treg numbers represent an important marker for the identification of breast cancer patients at risk of late relapses 95.

52 33 IDO, an immunosuppressive enzyme that depletes tryptophan, has been found to promote Treg differentiation and may become a suitable target to abrogate the development of T cell tolerance and to promote an effective immune response to breast cancer 278. In a 4T1 mouse model of breast cancer, attenuation of the immune suppressive Treg function using a soluble homodimeric form of mouse GITRL, with adenovirus-mediated intratumoural murine GM-CSF and IL-12 gene delivery, led to elevated IFNγ production, tumour-specific cytolytic T-cell activities, tumour rejection, and long-term survival in 65% of the animals Myeloid derived suppressor cells Circulating MDSC levels are positively correlated with clinical breast cancer stage and metastatic tumour burden 119. Immature DCs are present in more than 90% of breast cancers, whereas mature DCs are confined to peritumoural areas 115. Breast cancer patients with lower numbers of circulating MDSCs had a significantly higher probability of achieving a complete response to chemotherapy 282. In mouse models of breast cancer, limiting the number/function of MDSCs may help improve the efficacy of breast cancer therapies and enhance anti-tumour immunity. For example, 4T1 cells implanted into the flanks of BALB/c mice resulted in the production of nitric oxide by MDSC and attenuation of the IFN responsiveness of T cells 283. Furthermore, bone marrow derived myeloid progenitors can be recruited to the pre-metastatic lung in breast cancer 284, where they can induce the metastasis of tumour cells through downregulation of SMAD2 signaling (the canonical TGF-β pathway) and a switch to metastatic growth 285. In mouse models of breast and ovarian cancers, tumours with the ability to release significant amounts of CSFs or VEGF are associated with an expansion of a population of MDSC 241. This

53 34 may not only help the tumours suppress immune reactions, but also aid in the construction of new blood vessels for tumour growth 286, Signaling in the breast tumour microenvironment Different components of the breast tumour microenvironment described above communicate with each other by secreting cytokines and chemokines involved in inflammation, as well as by expressing immunoregulatory molecules on their surface. In most cases, inflammation associated with cancer is similar to that seen with chronic inflammation, which can be characterized by slow onset, long duration, large mononuclear cell infiltration, and tissue fibrosis and angiogenesis. Occasionally the balance seems to shift towards a more acute inflammatory process, favouring immune effector function activation 45. Chronic inflammation is correlated with risk of breast cancer recurrence in women after primary therapy 288 and may be mediated by cytokines including TNFα, IL-1β, IL-6, and IL-8 5 (Fig. 1-2, 4). TNFα is one of the key chemical mediators implicated in inflammation-associated cancers. High doses of TNFα can cause hemorrhagic necrosis via selective destruction of tumour blood vessels and generation of specific T cell anti-tumour immunity. However, when produced in the tumour microenvironment, TNFα can act as an endogenous tumour promoter 289, 290. TNFα has been linked to breast cancer development. Inhibition of TNFα and NF-κB transcription factor can protect from chemically-induced mammary gland carcinogenesis 291, and in vitro activation of the TNFα/NF-κB axis can induce an invasive and malignant behaviour in breast cancer cells 292. Human breast cancer cells have been shown to express TNFα 261, and its chronic expression in breast tumours may support breast tumour growth 293, together with other inflammatory cytokines, like IL1β, CCL2, and CCL5 294.

54 35 The release of chemokines by breast tumour cells can mediate the migration of leukocytes, primarily monocytes, from the circulation to breast tumours. CCL2, a regulator of macrophage recruitment and differentiation, was found to be highly expressed by the breast tumour cells, and expression levels were correlated with TAM accumulation. High expression of CCL2 was an indicator of early relapse and poor prognosis in breast cancer patients 266, 295. Inhibition of CCL2-CCR2 signaling in mouse breast cancer models prevented metastasisassociated macrophage accumulation and reduced metastasis to the lungs 296. Moreover, the ability of human breast carcinoma MDA-231 cells expressing CCL2 to form lung micrometastases in mice was inhibited by neutralizing antibodies to CCL In addition to CCL2, high levels of CCL5 (RANTES) were shown to correlate with advanced breast carcinoma 298, 299, and overexpression of CCL5 in MCF-7 breast tumour cells resulted in increased invasiveness 300. In cell lines of breast carcinoma, the expression of CCL5 was elevated by TNFα, alone or in synergism with IFNγ 265, 301. Consistent with the data above, samples of peripheral blood cells from stage II breast cancer patients showed higher expression of the chemokine receptor CXCR4 than PBL of controls and patients in earlier stages of disease 302. CXCL12 is a chemokine that binds to a CXCR4. CXCL12 is expressed in various tumours and may be important in tumour growth and invasion 303. Blockade of IL-6:IL-6R and IL-8:CXCR1 pathways reduced the proliferation of BCSCs in mouse xenograft models of breast cancer and resulted in a reduction of tumour growth and metastasis in these models 304 (Fig. 1-4). Serum levels of IL-6 and IL-8 have been associated with poor patient outcome in breast cancer 234, 305. The production of these inflammatory cytokines is regulated by the NF-κB signaling pathway 306. IL-6 has been shown to be a direct

55 regulator of BCSC self-renewal 224 and is a key component of a positive feedback loop involving MSCs and BCSCs 229. IL-6 is also an important mediator of the expansion and recruitment of MDSCs 307, 308. IL-8 receptor CXCR1 is highly expressed on BCSCs, and IL-8 has been shown to stimulate their self-renewal 304. Blocking CXCR1 in mouse xenografts significantly reduced the number of BCSCs, leading to decreased tumourigenicity and metastasis. Levels of another inflammatory cytokine, IL-1β, have been shown to be significantly higher in invasive carcinoma (like IDC) than in DCIS or in benign lesions, implying that elevated levels of IL-1β are also directly correlated with a more advanced disease 309. IL-6 and IL-1β increase the rate of accumulation and T cell suppressive activity of MDSCs 310, 311. IL-1β-driven inflammation also increases MDSC suppression of innate immunity by facilitating cross-talk between MDSCs and macrophages 312. Again, in a study using a 4T1 mouse model of breast cancer, the ectopic expression of IL-1β by 4T1 cells was shown to increase MDSC levels and stimulate growth and metastasis of tumours in vivo GM-CSF, a cytokine that drives myeloid and specifically dendritic cell differentiation, and VEGF, which drives angiogenesis, are additional inflammation-associated molecules that induce the accumulation of MDSC 84. Furthermore, GM CSF produced by breast cancer cells elicited the production of the IL-6 like cytokine oncostatin M by neutrophils in co-culture experiments. In turn, oncostatin M stimulated breast cancer cells exhibited increased VEGF production and increased invasiveness in Matrigel invasion assays 84.

56 37 Figure 1-4 Cytokines secreted by cells in the tumour microenvironment regulate breast cancer stem cell renewal. A model of cytokine networks mediating the interaction between MSCs, breast cancer cells, TAMs, CAFs, endothelial cells, and BCSC. IL-6 regulates CXCL7 production by MSCs and mediates MSC chemotaxis. The interaction between mesenchymal cells and tumour cells is regulated by a positive feedback loop that includes CXCL7 and IL-6. CAFs, TAMs, and MSCs secrete IL-6, IL-8, and CXCL7, which in turn activate Stat3/NF-κB signaling, leading to self-renewal of BCSCs, generating a positive feedback loop between the tumour microenvironment and tumour cells Adapted from Liu et. al., and Korkaya, Animal models of breast cancer Mouse models for breast cancer can be characterized into four main groups: syngeneic models, genetically engineered mice, xenograft models, and chemically/virally/ionizing radiation-induced models 313. Animal models have a limited role in cancer research, because the

57 biology of rodents and their tumours differs significantly from that of humans and human cancer (e.g., size, lifespan and time of tumour progression, numbers and differentiation of targets for oncogenic transformation, transformation efficiency of cancer cells). For example, about onehalf of human breast cancers are hormone-responsive at diagnosis, while the vast majority of mouse tumours are hormone-independent with much lower levels of ER and PR than human tumours 314. Although the basic mutation frequency is similar in both species, sporadic cancers are quite rare in WT rodents, and cells of rodent origin are much easier to transform in vitro by oncogene transfection or chemical carcinogens Chemical carcinogens, in particular polycyclic hydrocarbons (e.g. DMBA) and alkylating agents (e.g. MNU, ENU) have been widely used to study mammary tumourigenesis in mice. These models have been useful for identification of oncogenes and tumour-suppressor genes, mapping of tumour susceptibility traits, and the assessment of the carcinogenic or chemopreventative effects of various compounds 316. However, environmentally-induced cancer models develop a restricted subset of tumour types and grades with incomplete penetrance and variable latency 316. Mouse strains susceptible to mammary cancer were isolated many years ago, with vertical transmission subsequently shown to be due to a mouse mammary tumour virus (Bittner milk factor) 315. Mouse breast cancer cell lines that are commonly used include 4T1, 4T07, and EMT6 cells. Every cancer cell line has unique properties. For example, 4T1 mammary carcinoma (originally derived from a spontaneous mouse mammary tumour of a BALB/C mouse) is a transplantable cell line that is highly tumourigenic and invasive and can spontaneously metastasize from the primary tumour in the mammary gland to multiple distant sites including lymph nodes, blood, liver, lung, brain, heart, and bone 317. In comparison to 4T1, EMT6 tumours are less inflammatory and aggressive. EMT6 cells secrete much lower levels of

58 IL-6 than 4T1 cells, which corresponded with lower levels of infiltrating MDSC in the primary breast tumour 318 (see EMT6 Model for a full description). 39 The site of injection and the specific tropism of the breast cancer cell line used have largely defined primary and secondary metastatic growth. Orthotopic or ectopic implantation of cancer cells in the skin or mammary fat pad with subsequent formation of primary tumour metastases resembles the multiple stages involved in malignant breast cancer development in patients 319. Tail vein injection results mainly in lung seeding, whereas portal vein injection results in colonization of the liver. Intracardiac infusion gives rise to a broader target organ spectrum, including bone. The direct introduction of cancer cells into the blood circulation should be considered an assay of organ colonization and not a true metastatic process 320. Transgenic mice that express oncogenes under the mammary-specific promoters, such as the tumour virus long terminal repeat (MMTV-LTR) or the whey acidic protein gene (Wap), were the first generation of genetically modified mice for modeling breast cancer 321. Since then, hundreds of transgenic strains have been generated to test the biological relevance of several oncogenic pathways for the initiation of neoplastic transformation of mammary epithelial cells 322. Cell-type specific promoters to limit gene expression to specific target tissues (tetracycline-based systems) and promoter-specific recombinase-based (Cre-Lox) mechanisms for eliminating transgenes from specific tissues have proven to be useful methods for discovering genes and pathways that are involved in tumour initiation 322. Xenograft models are widely used in preclinical breast cancer studies. There are many human breast cancer cell lines available, most are derived from either established cancer cell lines or immortalized normal breast epithelial cells 315. Among the more commonly used are the metastatic human breast adenocarcinoma cells (MCF-7, MCF10AT) and the metastatic human

59 40 breast ductal carcinoma cells (T-47D), which result in the formation of solid tumours after a relatively short latency post-orthotopic injection in immunocompromised recipients. MDA-MB- 231 cells (derived from the pleural effusion of a cancer patient) are a widely used model of an estrogen-independent breast cancer cell line, which is able to colonize bone, liver, lung, adrenal glands, ovary, and brain after intravenous injection into immunocompromised mice 320. Cancer cell lines, like the ones mentioned above, that have been adapted to grow in culture are likely to have different environmental requirements than primary breast tumour cells. The establishment of a primary tumour in vitro is a rare event, found in no more than 1% of primary cancers 315. Cell selection in conversion to a continuous cell culture line, genetic drift, as well as viral or Mycoplasma infection are factors that can impact the validity of these xenograft models. Several other important limitations to xenograft models include the lack of immune responses in immunocompromised mice (which are key for tumour development); stromal components (like CAFs) that are not of tumour origin 17 ; and human-mouse cell-host incompatibility 320, 323. Humanized mouse models of breast cancer have recently become a major tool for studying this disease 324. Such models can be used to humanize metabolic enzymes, glycosylation enzymes, telomere structure and the immune system, which are useful in biomarker discovery as well as preclinical testing of therapeutics 316. Humanizing the murine immune system has consisted of ablating the endogenous immune system followed by engraftment with human immune cells 325. Advances in humanizing the mouse mammary fat pad, including reconstituting tumour-promoting immune and fibroblast compartments may improve breast xenograft models 326. Further modifications that rely on transgenic expression of human immunoreceptor genes have primarily been performed in mice expressing the homologous murine receptors that could heterodimerize with the human gene products 316, 327.

60 41 With the advent of recombinase-mediated genomic replacement, it might now be possible to knock in the human loci for the entire T-cell and B-cell receptors that span hundreds of kilobases. In an analogous fashion, specific major histocompatability complexes (MHCs) can be introduced into mice that lack endogenous MHC genes, thus making it theoretically possible to reconstitute the major components of the human immune system 328. Another mechanism of humanization is the insertion of human genes into the mouse genome. Knock-in of a human cdna into the corresponding murine genomic locus is a frequently used strategy 316, 329. A recent method called recombinase-mediated genomic replacement could be used to replace hundreds to thousands of kilobases of mouse genomic sequence with syntenic human sequence and to analyze the contribution of non-coding mutations that affect oncogene or tumour suppressor gene expression 330. Furthermore, there is a fundamental difference in telomere length in mice and humans, related to the limited or absent telomerase expression in most somatic human cells 331, as opposed to persistent telomerase expression in mouse tissues 332. Telomeres maintain chromosomal integrity and telomere dysfunction has an important role in oncogenesis, so the replacement of murine telomere components with human is another way to better recapitulate the neoplastic process 316. The model systems employed in the thesis below have focused on mouse transplantable tumour cell lines EMT6 and 4T1, which, as will become evident, have allowed characterization of host immune/ inflammatory processes in murine breast cancers, which differ in their growths, and metastatic potentials.

61 1.4 CD200 and CD200R CD200:CD200R overview CD200 is a transmembrane glycoprotein that belongs to the type I immunoglobulin family and exists in both soluble and membrane-bound forms 333. CD200 is widely expressed on leukocytes, including B cells, DCs and T cells. It is also found on endothelial cells, neurons, skin and tissue stem cells 334. CD200 has a short cytoplasmic tail with no known signaling motifs, and it functions through binding to CD200R, an immune inhibitory receptor that is predominantly expressed by cells of the myeloid lineage 335, resulting in an increased threshold of immune activation and attenuation of inflammatory responses In mouse, five CD200 receptors, CD200R1-R5, have been described. Of the five receptors, only two, CD200R1 and R2, are found in human. CD200R1 is the major receptor for CD200 in both mouse and human, and CD200R1 contains the NPXY motif in its cytoplasmic tail that is responsible for delivering inhibitory signals downstream 337 through Dok2 and RasGAP 339. Functional properties have been attributed to the alternative receptors CD200R2-R5 in mouse. However, in human, the function of CD200R2 remains unknown 340. The immunoregulatory function of CD200:CD200R1 interaction has been demonstrated in a number of models. CD200 knock-out (CD200 -/- ) C57B/6 mice are phenotypically normal, but have an increased susceptibility to autoimmune diseases, such as experimental autoimmune encephalomyelitis (EAE) and collagen-induced arthritis (CIA) 341. CD200 -/- mice have an increased activation of CD200R1 + myeloid-derived cells, including macrophages and DCs 341. CD200R + plasmacytoid DCs (pdcs) in mice are not tolerogenic under basal conditions, but engagement of CD200R on pdcs was induces IDO expressions, initiating the tolerogenic pathway of tryptophan catabolism with the help of type I interferons 342, 343. CD200:CD200R

62 43 interaction skews cytokine production by macrophages and T cells towards the TH2 arm and inhibits degranulation of mast cells in a variety of mouse models, including allergy and allotransplantation studies 344. Our lab has reported attenuation of CIA and reduced allo- and xeno- graft rejection by infusion of a soluble recombinant form of CD200 named CD200Fc (synthesized by linking the extracellular domain of CD200 to a murine IgG2aFc region) In this molecule, the Fc r and complement binding regions of the Fc region of IgG2a had been eliminated. CD200Fc suppresses T cell allostimulation and type-1 cytokine production (IL-2, IFNγ ) in vitro and in vivo 348. Tolerance and long-term survival of both skin and cardiac allografts occurred in mice with systemic overexpression of a CD200 transgene (CD200 tg mice) 349. Allografts in CD200 tg mice had more infiltrating FOXP3 + Treg cells and non-degranulating mast cells 350. These results support the characterization of CD200 as a co-regulatory molecule, which controls the outcome of the TCR-antigen encounter 336, 347. Several viral homologs of CD200 have been identified, indicating that the CD200:CD200R axis has been exploited by viruses as a mean to control host immune responses 351. It was reported that CD200R1 signaling deficiency in female mice affects the clearance of viral infection. Female CD200R1-deficient mice had significantly greater immunemediated pathology during influenza A infection than male mice 352. In mice, engagement of CD200R2 appears to skew differentiation of DCs toward a phenotype that is capable of inducing CD4 + CD25 + FOXP3 + Treg cells CD200:CD200R in cancer CD200 levels are elevated in patients with various solid cancers including: renal, colon, melanoma, and ovarian carcinomas as well as cancers of the hematopoietic system, including

63 44 myeloma, acute myeloid leukemia (AML), and chronic lymphocytic leukemia (CLL) Elevated CD200 has a negative effect on likelihood of survival, and expression of CD200 on cancer cells has been correlated with poor chances of recovery among AML patients 355. Clinical trials of an antibody to CD200 found reduced tumour size in leukemia and lymphoma patients and a decrease in the number of circulating Treg 359. The role of CD200 in immune tolerance suggests that CD200 may be a marker for cells that are better able to evade the immune system. Established tumours may evade the immune system by generating a tolerogenic response mediated by CD200 (Fig. 1-5). The role of CD200 in cancer has also been supported by work in preclinical models. CD200 -/- mice were shown to be resistant to chemical-induced skin carcinogenesis 360, and decreased tumour growth in these mice was accompanied by increased expression of proinflammatory cytokines by DCs in skin-draining lymph nodes, indicating that tumour growth in the presence of CD200 was likely a result of CD200-induced immunosuppression. Administering CD200Fc to mice with EL4 thymoma tumours increased tumour growth 347, 361. Human cancer stem cells from prostate, breast, brain, and colon cancers expressed CD200 on their surface 358, 362, and CD200 + breast cancer cells, but not CD200 - cells, grew in SCID mice to form a tumour 363. Expression of membrane CD200 on human CLL cells caused suppression of autologous effector T cells, and CD200 blockade using a rat monoclonal anti-cd200 antibody produced a reduction in the number of CD25 + CD4 + FOXP3 + Treg in vitro 364, 365. Furthermore, increased expression of CD200R, which is required for signalling following CD200 engagement, was detected on a subpopulation of CD4 + T cells in the spleen of CLL patients relative to controls.

64 Importantly, CD200 expression on tumour cells can be modulated upon immune challenge in vivo. Using a murine breast tumour cell line EMT6 with no detectable CD200 expression in vitro, we showed that CD200 expression on EMT6 cells was increased in vivo after transplantation into immunocompetent BALB/c mice 366 (see EMT6 breast cancer model below). These findings suggest that the presence of CD200 on tumour cells may identify cancer cells that are better able to evade the patient s immune response CD200:CD200R signaling may regulate anti-cancer responses by altering the proliferation and activation of regulatory cells in the tumour microenvironment. CD200 affects the cytokine production profile by immune cells 344, which may influence immune cell infiltration and activation in the tumour and the DLN, enhancing metastasis via distal action. For example, overexpression of CD200 is associated with increased production of TGFβ 367, an important molecule in the regulation of anti-tumour immunity 368. In mice, CD200 stimulates the development of IL-10/TGFβ-producing suppressor T cells (Ts) that prevent an inflammatory immune response 367, 369 and may protect breast tumour cells from the host s anti-tumour inflammation, thus resulting in enhanced tumour growth and metastasis. The thesis that follows demonstrates a role for CD200 expression by tumour/host on both local and metastatic growth of breast cancer cells in mice. A recent study has supported these findings in a different mouse model of squamous cell carcinoma, where CD200 expression was shown to be highly induced in cells that have metastasized to the lymph node and other solid organs. These CD200 + squamous carcinoma cells appeared to have a selective advantage to metastasize, possibly through modulating CD200R + myeloid cells in the lymph node microenvironment 370.

65 46 Figure 1-5 CD200 in the tumour microenvironment. (Upper) CD200 expressed by tumour or host cells interacts with CD200R either on an immune effector cell, directly suppressing antitumour activity, or on an APC, which delivers an indirect immunosuppressive signal. (Lower) Antibodies blocking the CD200:CD200R interaction may be a promising therapeutic approach to enhance anti-tumour immune responses in cancer. Adapted from Rygiel & Meyaard,

66 Soluble CD200 A soluble form of CD200 (scd200) was detected in normal human plasma, and its levels were increased in the plasma of CLL patients, where scd200 levels were correlated with tumour burden, late stage disease, and disease aggressiveness 371. High scd200 levels enhanced engraftment of splenic human CLL cells in vivo, when NOD.SCID IL-2Rγ-/- mice were supplemented with scd200 hi CLL plasma. Engraftment was attenuated when scd200 was preabsorbed from the plasma or when the mice were treated with an anti-cd200 mab 371. Recently we found that CD200Fc increased lung and liver metastasis of three human breast cancer lines tested in immunocompromised mice (MCF7; HTB19; MDA-MB-231). Furthermore, tumour metastasis was associated with changes in gene expression in tumour tissue, with significant alterations in genes encoding MMPs and transcription factors modulating inflammation EMT6 breast cancer model We have developed a mouse model to study the role of CD200 in breast cancer growth and metastasis. Murine EMT6 breast tumour cells are injected into the mammary fat pads of female BALB/c mice, where the tumours develop over days. The EMT6 cell line was established several decades ago from a transplantable murine mammary carcinoma that arose in a BALB/cCRGL mouse after implantation of a hyperplastic mammary alveolar nodule. The resulting tumour line (named KHJJ) was propagated in BALB/cKa mice and adapted to tissue culture after the 25th animal passage, and the cell line was named EMT. EMT6 is a clonal isolate of EMT isolated in 1971 at Stanford University 373.

67 48 EMT6 breast tumour cells grown in vitro do not express CD200 on their surface. However, when EMT6 cells are injected into immunocompetent, wild type (WT) BALB/c mice, approximately one third of the EMT6 cells isolated from the tumour were found to express CD200 on their surface 366. The in vivo induced expression of cell-surface CD200 is lost after 7-10 days in culture. In the presence of a compromised immune system (in NOD.SCID IL-2Rγ-/- and CD200 tg mice), this selection for CD200 + tumour cells is not observed (Fig. 1-6) 366. Thus, in female mice, we infer that a competent immune system selects for growth of breast cancer cells expressing CD200 on their surface. CD200, in turn, binds to receptor(s) on the host s immune cells within the tumour microenvironment and inhibits generation and expression of anti-tumour immune response. This hypothesis supports the idea that cancer cells may hijack the body s own immune system to protect themselves from being eradicated by the host s anti-tumour immunity. Figure 1-6 CD200 expression on the surface of EMT6 cells grown in culture, in immunocompromised NOD.SCID IL-2Rγ-/- mice, and in immunocompetent WT BALB/c mice. To study the effect of host- or tumour-derived CD200 on EMT6 breast cancer growth and metastasis, we have developed four animal strains and three cancer cell lines. In addition to the WT female host, our lab generated CD200R1 -/- mice, CD200 -/- mice, and CD200 tg mice