UNIVERSITY OF CALGARY The Role of Myeloid Derived Suppressor Cells in the Interleukin-10 Model of Colitis-Associated Cancer by Manmish Singh Bawa A THESIS SUBMITTED TO THE FACULTY OF GRADUATE STUDIES IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY GRADUATE PROGRAM IN GASTROINTESTINAL SCIENCES CALGARY, ALBERTA JULY 2015 © Manmish Singh Bawa 2015 Abstract Chronic inflammation in the colon is a risk factor for developing colorectal cancers, one of the leading causes of cancer related deaths in Canada. A hallmark feature of inflammation is the recruitment of leukocytes. Of interest in this thesis are the heterogenous population of Gr1+/CD11b+ Myeloid Derived Suppressor Cells (MDSC). These cells are recruited to tumour sites, but have recently also been shown to be increased during inflammation. A key regulator of inflammation and leukocyte recruitment to the gut is the innate immune receptor TLR4. In this thesis we use the IL-10-/- mouse model to examine the role of MDSC in colitis-associated cancer. Our preliminary data in an IL-10 TLR4 double knock out mouse revealed an increased cancer incidence in the absence of TLR4. We hypothesized that MDSC contributed to cancer development in the IL-10-/- and TLR4 signaling modulates cancer development through an effect on MDSC function or recruitment. Our studies demonstrated enhanced numbers of Gr1+/CD11b+ cells in the colons of IL-10-/- mice. Characterization of these cells revealed an active immunosuppressive phenotype demonstrated by activity of NOS II and Arginase I enzymes and suppression of T-cell proliferation. Pharmacological targeting of MDSC decreased dysplasia development. Enriching the MDSC pool using ii adoptive transfer was associated with enhanced neoplasia development in the IL-10-/- mice. These data suggest that MDSC contribute to inflammation associated tumour development. In the absence of TLR4, IL-10-/- mice recruited increased levels of MDSC to the colon. MDSC numbers correlated with the levels of cancer development in this model. TLR4 competence did not affect the immunosuppressive function of MDSC. We identified MCP-1 and SDF-1 as MDSC chemoattractants. Examination of colonic tissue demonstrated increased MDSC chemoattractant expression levels in the absence of TLR4 before tumour development in the IL10 deficient model. Our data identify a link between TLR4 signaling and MDSC recruitment to the gut during inflammation and demonstrate that MDSC recruitment drives cancer development. iii Acknowledgements I would like to acknowledge the love and support of my family that has made it possible for me to achieve everything that I have. My wonderful parents, Daljit and Baljinder, have always nurtured and supported me to the best of their ability and beyond. My brother and my best friend, Gharandip, has been a constant companion, source of guidance and inspires me to life in service. The love of my life, Navkiran, who went from being my girlfriend to fiancé and finally, wife, during the course of my graduate study years, is my reason for being. I would also like to acknowledge the support of my supervisor, Dr. Donna-Marie McCafferty. Donna-Marie has provided me with invaluable mentorship and guidance over the last 6 years. I would like to thank my supervisory committee: Dr. Joe Davison, Dr. Paul Beck and Dr. Oliver Bathe. Under the combined supervision of Donna-Marie and my committee, I have expanded my knowledge of the scientific method and how to put it into practice. It has been a privilege to be mentored and guided along in my career by these and many other individuals that I greatly admire. I would like to thank all members of the McCafferty lab over the years, especially, Ronald Chan, Dr. Rui Zhang, Dr. Ying Gao, Kelvin Ng and Dr. Maitham Khajah. I thank Dr. Stephan Urbanski for his great help with my project. iv I would like to thank the Queen Elizabeth II graduate scholarships for their support during my graduate career. Finally, I would like to thank the Gastrointestinal Research Group, the Department of Physiology and Pharmacology and the University of Calgary for providing an excellent environment to learn, do research, to interact and collaborate other scientists. v Dedication To my wife and family vi Table of Contents Abstract………………………………………………………….……………………….ii Acknowledgements………………………………………….………………………….iv Dedication………….………………………………………….………………………...vi Table of Contents...………………………………………….…………………………vii List of Tables….....………………………………………….…………………………...x List of Figures and Illustrations…………………………….………………………….xi List of Symbols, Abbreviations, Nomenclatures………………………………..….xiv Chapter 1: Introduction………………………………….…………………………...1 1.1 Inflammation and Cancer in the Gut……….…………………………......2 1.1.1 Colorectal Cancer………………………….………………………….....2 1.1.2 Inflammatory Bowel Disease…..……….…………………………........4 1.1.3 IBD and Cancer in the Colon…..……….…………………………........5 1.2 Myeloid derived suppressor cells…..…….…………………………......11 1.2.1 Introduction to myeloid derived suppressor cells….………………...11 1.2.2 MDSC Recruitment………………………………….………………….12 1.2.3 MDSC Function……………………………………….…………………14 1.3 Interleukin-10 and Toll like receptor 4 deficient model……..…………18 1.3.1 Models of Colitis-associated cancer………………..…………………18 1.3.2 The Interleukin 10 Deficient Mouse………………..……………….…20 1.3.3 Toll like receptor 4………………..……………….…………………….22 1.3.4 IL-10 TLR4 Double Deficient Mouse………………………………….25 1.4 Hypothesis and Objectives……………………………………………….26 Chapter 2: Materials and Methods…………………….………………………….28 vii 2.1 Animals……………..………………..……………….…………………...29 2.2 Evaluation of Inflammation and Cancer….……….…………………….30 2.3 Colon Immunohistochemistry for Gr1+ CD11b+ Cells…...…………….32 2.4 Flow Cytometry Experiments…………………………………………….33 2.5 MDSC Isolation……..………………..……………….…………………...37 2.6 MDSC Function in vitro……………..……………….…………………...38 2.7 Bone Marrow Transplant…………..……………….………………….....41 2.8 Molecular Techniques………………..……………….…………………..41 2.9 Chemotaxis Assays..………………..……………….…………………...45 2.10 Statistical Analysis..………………..……………….…………………...47 Chapter 3: The role of MDSC in the Interleukin-10 deficient mouse….…....48 3.1 Cancer and Gr1+CD11b+ counts in the IL-10 deficient mouse……….49 3.1.1 Colon cancer development is increased in the IL-10 deficient mouse…………………………………………………………………………...50 3.1.2 Gr1+CD11b+ cell recruitment to the colon is increased in the IL-10 deficient mouse………………………………………………………………...52 3.2 Gr1+CD11b+ cell Function in vitro……………………………...………..58 3.3 MDSC function in vivo…………………………………………………….65 3.3.1 Depleting MDSC in vivo decreases cancer development in the IL-10 deficient mouse………………………………………………………………...65 3.3.2 Adoptive transfer of MDSC in vivo increases cancer development in the IL-10 deficient mouse……………………………………………………..72 3.4 MDSC in the AOM/DSS Model of Colitis Associated Cancer………...79 Chapter 4: The role of MDSC in the Interleukin 10 Toll Like Receptor 4 double deficient mouse……………………………………………………………..83 4.1 Cancer incidence and MDSC levels in the IL-10 TLR4 double deficient mouse…………..……………………………………………………………….84 viii 4.1.1 Inflammation and Cancer levels in the IL-10 TLR4 Double Deficient mouse…………..……………………………………………………………….85 4.1.2 MDSC recruitment in the IL-10 TLR4 double deficient mouse……..87 4.1.3 MDSC and Cancer correlation…………………………………………91 4.2 Other leukocyte population levels in the absence of TLR4 in the IL-10 deficient mouse..……………………………………………………………….95 4.3 MDSC function in the IL-10 TLR4 double deficient mouse………...…99 4.4 TLR4 competence and cancer development…………………………104 Chapter 5: MDSC chemotaxis in vitro and in vivo……………………………111 5.1 Underagarose and Ibidi Chamber Chemotaxis Assays……………..114 5.2 Transwell Chamber Chemotaxis Assay…….………………..............120 5.3 MCP-1 increases MDSC chemotaxis in vivo………………...............127 5.4 Chemoattractant protein and mRNA levels in the IL-10 deficient mouse………………………………………………………………………….133 Chapter 6: Discussion and Future Directions…………………………………139 References…………………………………………………………………………...156 ix List of Tables Table 3-1 Combinations of primary and secondary antibodies used in Gr1+CD11b+ cell colon immunohistochemistry to check for non-specific binding…………………………………………………………………………………..54 Table 5-1 The number of migrating MDSC towards chemoattractants in the underagarose chemotaxis assay…………………………………………………...115 x List of Figures and Illustrations Figure 1-1 Comparison of IBD associated and Spontaneous Colon Cancers……8 Figure 1-2 Methods of MDSC-mediated immune suppression……………..…….16 Figure 3-1 Incidence of cancer increases with age in the IL-10 deficient mouse…………………………………………………………………………………...51 Figure 3-2 Lamina Propria Gr1+CD11b+ cell flow cytometry…………...…………55 Figure 3-3 Representative immunohistochemistry for Gr1+CD11b+ cells in colon tissue sections………………………………………………………………………....56 Figure 3-4 Gr1+CD11b+ cell colon immunohistochemistry…………………...…...57 Figure 3-5 IL-10 deficient Gr1+CD11b+ cells express arginase I activity………...62 Figure 3-6 IL-10 deficient Gr1+CD11b+ cells have NOS2 Activity………………..63 Figure 3-7 IL-10 deficient Gr1+CD11b+ cells suppress T-cell proliferation………64 Figure 3-8 Low dose chemotherapeutics deplete MDSC in the IL-10 deficient mouse sections………………………………………………………………………...67 Figure 3-9 Lower MDSC counts are associated with reduced polyp scores in the IL-10 deficient mouse………………………………………………………………....68 Figure 3-10 Lower MDSC counts are associated with reduced dysplasia scores in the IL-10 deficient mouse……………………………………………………….....69 Figure 3-11 Lower MDSC counts do not alter macroscopic inflammation in the IL10 deficient mouse….………………………………………………………………....70 Figure 3-12 Lower MDSC counts do not affect histological inflammation in the IL10 deficient mouse….………………………………………………………………....71 Figure 3-13 Adoptive transfer of MDSC does not affect MDSC levels in the colons of IL-10 deficient mice………………………………………………………...75 Figure 3-14 Adoptive transfer of MDSC increases neoplastic changes in the colons of IL-10 deficient mice………………………………………………………...76 Figure 3-15 Adoptive transfer of MDSC does not affect inflammation in IL-10 deficient mouse colons………………………………………………………………..77 xi Figure 3-16 Dysplasia and Gr1+CD11b+ cell counts are increased in the AOM/DSS model of colitis-associated cancer………………………………...……81 Figure 4-1 TLR4 deficiency increases cancer development and inflammation in the IL-10 deficient mouse……………………………………………………………..86 Figure 4-2 TLR4 deficiency does not affect MDSC levels in the bone marrow and spleen in the IL-10 deficient mouse………………………………………………….89 Figure 4-3 TLR4 deficiency increases MDSC recruitment to the colon in the IL-10 deficient mouse……………………..………………………………………………….90 Figure 4-4 MDSC recruitment correlates with neoplastic changes in the IL-10 deficient mouse model……………..………………………………………………….93 Figure 4-5 TLR4 deficiency does not affect peripheral blood leukocyte levels in the IL-10 deficient mouse…………..…………………………………………………97 Figure 4-6 MDSC are the only upregulated leukocyte population in the colon in the absence of TLR4 in the IL-10 deficient mouse….……………………………..98 Figure 4-7 TLR4 does not affect MDSC function in the IL-10 deficient mouse..101 Figure 4-8 TLR4 expression in MDSC in the IL-10 deficient mouse……………103 Figure 4-9 Bone marrow transplants between IL-10 TLR4 double deficient and IL10 deficient mice increase…………..………………………………………………107 Figure 4-10 TLR4 mRNA expression is decreased in the polyps of IL-10 deficient mice…………………………..………..………………………………………………109 Figure 5-1 Neutrophil and MDSC Underagarose Chemotaxis Assay…………..118 Figure 5-2 MDSC Ibidi Chemotaxis. ..……………………………………………..119 Transwell Chemotaxis Chamber Illustration……………………………...……….122 Figure 5-3 MDSC migrate towards MCP-1 in the Transwell Chemotaxis Assay…………..………..………………………………………………..…………...123 Figure 5-4 MDSC migration in the Transwell Chemotaxis Assay……..………..124 Figure 5-5 MDSC migrate towards SDF-1 in the Transwell Chemotaxis Assay.................................................................................................................125 xii Figure 5-6 TLR4 does not affect MDSC chemotaxis in the IL-10 deficient mouse…………..................................................................................................126 Figure 5-7 MCP-1 increases MDSC recruitment kinetics in vivo.......................131 Figure 5-8 MDSC recruitment kinetics in the colon in vivo................................132 Figure 5-9 MCP-1 mRNA levels are comparable in the IL-10 deficient mouse in the absence of TLR4………..………..……………………………………………...135 Figure 5-10 MCP-1 protein levels are significantly increased in in the IL-10 deficient mouse in the absence of TLR4…………………………………………..136 xiii List of Symbols, Abbreviations, Nomenclatures AOM Azoxymethane APC Adenomatous polyposis coli CARD15 Caspase activation and recruitment domain 15 CCR2 Monocyte chemoattractive chemokine receptor type CD Crohn’s Disease CD11b Integrin α 1 CD18 Integrin β 2 cDNA Complimentary DNA CO2 Carbon dioxide CpG Phosphodiester bond between Cytosine and Guanine CRC Colorectal Cancer DNA Deoxyribonucleic acid DSS Dextran sulfate sodium EDTA Ethylenediaminetetraacetic acid FBS Fetal bovine serum FITC Fluorescein isothiocyanate fMLP N-formyl-methionyl-leucyl-phenylalanine GAPDH Glyceraldehyde 3-phosphate dehydrogenase GMCSF Granulocyte macrophage colony stimulating factor H&E Hematoxylin and eosin HBSS Hank’s Balanced Salt Solution xiv protein receptor, C-C HEPES 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid HNPCC Hereditary nonpolyposis colorectal cancer IBD Inflammatory bowel disease IL Interleukin IL-1β Interleukin 1β IL-6 Interleukin 6 IL-10 Interleukin 10 IL-10-/- Interleukin 10 deficient IL-10-/-TLR4-/- Interleukin 10 Toll like Receptor 4 double deficient IL-12 Interleukin 12 IL-23R Interleukin 23 receptor L-arg L-arginine LPS Lipopolysaccharide KC Keratinocyte chemotactic peptide Mac-1 Macrophage-1 antigen MCP-1 Monocyte chemotactic protein 1, CCL2 MCSF Macrophage colony stimulating factor MDSC Myeloid derived suppressor cells MIP-2 Macrophage inflammatory protein 2, CXCL2 MyD88 Myeloid differentiation primary response gene 88 NF-κB Nuclear factor κ B NOD2 Nucleotide-binding oligomerization domain containing 2 xv NOS2 Nitric oxide synthase 2 OCT Optimal cutting temperature compound p53 Protein 53 PBS Phosphate buffered saline PCR Polymerase chain reaction PE Phycoerytherin PGE2 Prostaglandin E2 PMSF Phenylmethylsulfonyl fluoride PRR Pattern recognition receptor RAG2 Recombination activating gene 2 RANTES regulated upon activation, normal T-cell expressed and secreted RIPA Radio-immunoprecipitation assay buffer RNA Ribonucleic acid RPMI Roswell Park Memorial Institute RT-PCR Real time polymerase chain reaction SDF-1 Stromal derived factor 1, CXCL12 SDS Sodium dodecyl sulfate TH T-helper cell TLR Toll like receptor TLR4 Toll like receptor 4 TNF-α Tumor necrosis factor α TRIF TIR-domain-containing adapter-inducing interferon-β xvi UC Ulcerative Colitis VCl3 Vanadium III chloride VEGF Vascular endothelial growth factor WKYMVm W-peptide, Trp-Lys-Tyr-Met-Val-Met-NH2 WNT Wingless-related integration site xvii Chapter 1: Introduction 1 1.1 Inflammation and Cancer in the Gut 1.1.1 Colorectal Cancer Colorectal cancer (CRC) is responsible for approximately 608,000 deaths per year worldwide (Parkin et al., 2010). Colorectal cancer (CRC) is the third most common cancer diagnosis and the second leading cause of cancer related death in Canada (Singh et al., 2012a). In 2012 alone, there were an estimated 23,200 new cases diagnosed and 9,200 deaths from CRC in Canada (Dube, 2012). According to the Canadian Digestive Health Foundation, the likelihood of a Canadian developing CRC in their lifetime is 1 in 15 and the estimated lifetime treatment cost of each patient is $750,000 (http://www.cdhf.ca/en/statistics). This translates to a cost burden of half a billion dollars annually on the health care system. CRC is the formation of malignant neoplasia in the colon, rectum or appendix. Colon adenocarcinoma is the most common cancer in the gastrointestinal tract (Kumar, 2010). A combination of molecular events lead normal mucosa to the development of aberrant crypt foci, then adenomas and finally adenorcarcinomas. Adenomas are benign epithelial derived tumors within the mucosa that can become carcinomas when glandular tissue invades into the submucosal space. The development of cancer is the result of the accumulation of genetic mutations and epigenetic abnormalities, which drive the transformation of normal epithelia into adenomatous tissue (Kumar, 2010). Three different 2 pathways of genomic instability have been described in CRC: chromosomal instability, microsatellite instability and CpG island methylator phenotype (Kumar, 2010). Of these, the most common and best-studied pathway is the chromosomal instability pathway that involves the classic adenoma-carcinoma APC/β-catenin sequence (Fearon and Vogelstein, 1990; Pino and Chung, 2010; Vogelstein et al., 1988). The microsatellite instability pathway is found in 15% of colon cancers and involves the loss of DNA mismatch repair mechanisms (Boland and Goel, 2010). The CpG island methylator phenotype involves the hypermethylation of various promotor CpG island loci that leads to the silencing of key genes (Bae et al., 2013). In advanced stage CRC, these pathways can often be observed acting in congruence (Remo et al., 2012). CRCs is a multifactorial disease that can be initiated by many different types mutations and environmental factors (The Cancer Genome Atlas Network, 2012). Genetic causes of CRC such as hereditary syndromes, including Lynch syndrome (HNPCC) and familial adenomatous polyposis, account for approximately 20 % of CRC (Patel and Ahnen, 2012). The majority of CRC are related to lifestyle, rather than a specific underlying condition (Watson and Collins, 2011). The key predisposing factors in the development of CRC include advanced age, presence of polyps, familial history, genetics and a history of inflammatory conditions such as inflammatory bowel disease. A presence of a combination of these factors can greatly increase the probability of CRC 3 development (Ahmadi et al., 2009). Inflammation is present in colon and many other cancers regardless of the initiating sequence of events (Mantovani et al., 2008). 1.1.2 Inflammatory Bowel Disease Inflammatory bowel disease (IBD) refers to a group of idiopathic chronic relapsing inflammatory conditions that affect the gastrointestinal tract. The two most common forms of IBD are ulcerative colitis (UC) and Crohn’s disease (CD). The incidence of IBD ranges from about 10 to 200 cases per 100,000 individuals in North America and Europe. IBD is most prevalent in developed and urbanized nations and the incidence of the disease has been rising in the last half century (Bouma and Strober, 2003; Molodecky et al., 2012; Rocchi et al., 2012). The most commonly held hypothesis regarding the development of IBD is that in a genetically susceptible individual an environmental triggers leads to an aberrant immune response to a subset of enteric bacteria (Packey and Sartor, 2008). The immune response is overly aggressive and involves the adaptive immune system, especially T-cells. It is this uncontrolled immune response that is directly responsible for the mucosal damage that is observed in IBD (Bouma and Strober, 2003; Sartor, 2006). Genetic susceptibility has been identified as an important component in the pathogenesis of IBD via single nucleotide 4 polymorphism and candidate gene approaches in humans, and use of transgenic and knockout mice. Genes with nucleotide polymorphisms related to IBD include NOD2/CARD15, IL-23R and IL-10 (Hugot et al., 2001; Neuman and Nanau, 2012). Though a critically important component, twin studies have found genetic susceptibility does not entirely account for the development of IBD (Bouma and Strober, 2003; Loftus, 2004; Sartor, 2006). The interaction between the gut immune environment and enteric bacteria and their antigens also plays a key role in IBD. Inflammation does not develop in many induced and spontaneous models of IBD in germ free environments (Schultz et al., 1999). In humans, the use of antibiotics and some probiotic therapies has shown the importance of commensal bacteria in disease development and recovery (Sartor, 2005; Sartor, 2006). The combination of genetic susceptibility and bacterial antigen stimulation requires an environmental trigger for onset or relapse of the disease. The importance of environmental factors was first indicated by the differences in incidence of IBD across geographic regions, age and time. Potential environmental triggers include cigarette smoke, stress and diet (Loftus, 2004; Sartor, 2006). 1.1.3 IBD and Cancer in the Colon The link between inflammation and cancer development is well established and has been reported in a number of human diseases and animal models (Coussens and Werb, 2002; Dalgleish, 2006). Examples of inflammatory 5 conditions leading to cancer development include gastric, liver, prostate and colon cancers (De Marzo et al., 2007; Fattovich et al., 2004; Parsonnet et al., 1991). While this relationship is not yet fully understood, it is clear that the inflamed tissue microenvironment generates the right conditions for the activation of protumorigenic mechanisms (Dalgleish, 2006). In fact, inflammatory bowel disease patients are at an increased risk of developing colon cancer (Bernstein et al., 2001; Bernstein and Nabalamba, 2007; Rhodes and Campbell, 2002; Rutter et al., 2004). The risk of developing cancer in IBD patients increases with severity and duration of inflammation (Rutter et al., 2004). In UC patients, the risk of developing colorectal cancers is 1-3% at 10 years and 18% and 30 years of disease (Delaunoit et al., 2006; Eaden et al., 2001). One Swedish study reported an approximate six-fold increase in CRC development in UC patients compared to that of the general population (Ekbom et al., 1990). The risk of developing CRC is further increased in patients that develop both inflammatory bowel disease and primary sclerosing cholangitis (Thackeray et al., 2011). A recent Danish study suggested that a decreased risk of patients developing IBD related CRC was due to the availability of better treatment options and management (Jess et al., 2012). Another recent Danish population based cohort study showed an increased of 5-year mortality rate in colon cancer patients with a history of IBD (Ording et al., 2013). These studies suggest a strong link between IBD and cancer development. 6 CRC tend to follow heterogenous, but largely similar molecular and phenotypic pathways. While certain differences exist between inflammation associated and sporadic CRCs, they follow similar phenotypic stages in the transformation of normal epithelia into adenocarcinomas (Terzic et al., 2010). Inflammation associated and sporadic cancers also involve the accumulation of similar mutations and activation of common pathways (Rhodes and Campbell, 2002). However, the endoscopic features and the timing of specific genetic changes in IBD-associated colorectal cancer do differ from cancers that arise sporadically (Figure 1-1). In sporadic cancers, the loss of adenomatous polyposis coli occurs early and is frequent, whereas p53 mutations occur late and are less frequent. In contrast, in IBD-associated cancer the loss of adenomatous polyposis coli function generally occurs late and is infrequent, whereas p53 mutations occur early and are more frequent (Terzic et al., 2010). IBD associated cancers are often poorly differentiated and may be raised minimally above the surrounding mucosa, whereas sporadic cancers tend to be well differentiated and polypoid. (Delaunoit et al., 2006). 7 Figure 1-1 Comparison of IBD associated and Spontaneous Colon Cancers This figure adapted from a review in Gastroenterology by Terzic et al, shows the differences in progression mechanisms between colitis associated and spontaneously occurring colon cancers (Terzic et al., 2010). Mutations in the adenomatous polyposis coli, 61" nisms between colitis associated and spontaneously occurring colon cancers ncers lei followed by adenoma and carcinoma formation. In colitis associated cancer chronic inflammation promotes tumor development by activating proliferation and anti-apoptotic pathways in premalignant pathways. 8 While it was earlier proposed that the increased cancer development in IBD patients was directly caused by the genetic defects that predisposed them to IBD, there is now convincing evidence that the cancer risk is acquired and caused by the inflammation (Rhodes, 1996; Rhodes and Campbell, 2002). For example, a large Swedish epidemiological study found no significant increase in the risk of colorectal cancer in first-degree relatives of IBD patients (Ekbom et al., 1990). In UC patients, the development of cancer has generally found to be localized to inflamed areas, except in rectal inflammation (Ekbom et al., 1990). A recent study of neoplastic colon polyps found that adenoma size correlates well with the degree of inflammation in the bowel (Bilinski et al., 2012). Inflammatory status also correlates with the DNA methylation in colon mucosa, which is thought to contribute to cancer development by gene silencing (Saito et al., 2011). A better correlation still, is observed between cancer and inflamed sites in CD, even in the small bowel where cancers are otherwise rare (Church et al., 1985; Collier et al., 1985). A similar incidental and spatial correlation between inflammation and cancer has been reported in various animal models. Of special interest is the colitis-cancer association in the cotton-top tamarin. This new world primate universally develops colitis in captivity that is indistinguishable from UC in humans. Many of the monkeys also go on to develop colon cancer, but cancer has never been reported before the development of inflammation (Chalifoux and Bronson, 1981). A recent Illinois case-controlled study identified inflammation as 9 an independent risk factor for colon cancer development in UC patients (Rubin et al., 2013). Inflammation has been linked to cancer development through numerous and varied mechanisms (Kumar, 2010; Kundu and Surh, 2008). Cancer development and growth can be driven by inflammation at every stage. Inflammation can induce malignant cell transformation through genomic instability and epigenetic changes caused by oxidative damage by reactive oxygen and nitrogen species (Bertram, 2000; Costa et al., 2014; Rios-Arrabal et al., 2013). Cancer cell survival can be enhanced through increased proliferation and decreased apoptosis of transformed cells through expression of cytokines and activation of related pathways such as tumor necrosis factor α and tumor growth factor to(Atsumi et al., 2014; Seamons et al., 2013; Setia et al., 2013). Inflammation can help tumors grow by escaping immune surveillance and increased vascularization through production of pro-angiogenic factors (Gabrilovich and Nagaraj, 2009; Kong et al., 2013). Metastasis can be enhanced through increased contact between cancer and stromal cells as well increased mesenchymal to epithelial transition and angiogenesis (Fuxe and Karlsson, 2012). A key characteristic shared by both inflammatory and cancerous conditions is the presence of leukocytes. Leukocytes produce factors such as radical species and 10 cytokines that largely drive cancer development in the inflammatory microenvironment. Leukocytes involved in the innate and adaptive immune systems participate in the progression of IBD and colon cancers (Goldszmid and Trinchieri, 2012). Leukocytes can both limit and promote the progression of inflammation and cancer. A large body of work has identified inflammationassociated pathways in leukocytes as playing a key role in cancer development as well (Ben-Neriah and Karin, 2011; Demaria et al., 2010). For example, seminal work by Greten et al shows that the IκK B molecule expressed in the NFmo pathway links inflammation with cancer development in the colon (Greten et al., 2004). One leukocyte population that has been shown to play important roles in both inflammatory and cancerous conditions is the myeloid derived suppressor cell (Ostrand-Rosenberg and Sinha, 2009). 1.2 Myeloid derived suppressor cells 1.2.1 Introduction to myeloid derived suppressor cells Myeloid derived suppressor cells (MDSC) are a group of heterogenous cells of myeloid lineage with numerous pro-tumorigenic effects. The defining murine phenotype of MDSC is the expression of Gr-1, a glycophosphatidylinositol linked membrane protein and CD11b (Ly-40), a transmembrane protein that pairs with CD18 to form the Mac-1 integrin and an ability to suppress the immune system (Dolcetti et al., 2008; Kusmartsev and Gabrilovich, 2006). MDSC consist of “immature” macrophages, monocytes and dendritic cells, which have been 11 shown to play key role in tumor immune evasion, growth and progression. The normal hematopoiesis process of myeloid cells is disturbed and the immature cells are recruited out of the bone marrow leading to their peripheral accumulation. In the normal mouse, about 20-30 % of bone marrow cells and 24 % of splenocytes are in an immature state and are Gr-1+ CD11b+. A marked increase in systemic levels of MDSC, up to 50% of all splenocytes, is observed in various mouse models of cancer (Kusmartsev and Gabrilovich, 2003; Kusmartsev and Gabrilovich, 2006; Kusmartsev et al., 2000). Peripheral blood levels of MDSC have been found to be increased in human cancer patients and to correlate with metastatic tumor burden and cancer stage in a number of human cancers including colon and breast cancer (Almand et al., 2001; DiazMontero et al., 2008). Two recent studies have focused on the clinical significance of MDSC specifically in colorectal carcinomas. These studies report that MDSC levels in the periphery and at the tumor site of CRC patients correlate closely with disease state as well as nodal and distant metastasis (Sun et al., 2012; Zhang et al., 2013). 1.2.2 MDSC Recruitment MDSC numbers have been shown to be systemically upregulated in various models of infectious and inflammatory models such as ovalbumin challenge, polymicrobial sepsis and toxoplasmosis (Bronte et al., 1998; Kusmartsev and Gabrilovich, 2006; Mencacci et al., 2002). MDSC expansion during inflammation 12 has generally been found to be lower than in neoplastic conditions and transient (Gabrilovich and Nagaraj, 2009). Systemic levels of MDSC are increased in cancer models, the key location of action for MDSC is believed to be the tumor site where they are actively recruited (Dolcetti et al., 2008; Sawanobori et al., 2008). Peripheral accumulation of MDSC via disruption of myelopoiesis and recruitment is believed to be triggered by tumor derived soluble factors (Bronte et al., 2006). Tumor resection has been shown to completely normalize myelopoiesis and peripheral MDSC levels in numerous models (Salvadori et al., 2000; Sinha et al., 2005). Among the many tumor derived factors, vascular endothelial growth factor (VEGF) has received particular attention in its ability to cause the accumulation of MDSC (Dolcetti et al., 2008). VEGF has been shown to be responsible for recruitment of MDSC into the spleen and periphery in a transgenic mouse model with spontaneously occurring mammary carcinoma (Melani et al., 2003). Another study showed the ability of VEGF specifically derived from the tumor to inhibit proper dendritic cell maturation and cause MDSC accumulation (Gabrilovich et al., 1998). Interestingly, MDSC also produce VEGF and further increase its bioavailability by expressing matrix metalloproteinase 9 (Yang et al., 2004). The ability of tumors to interfere with the colony stimulating factor network, especially the functions of granulocyte macrophage colony stimulating factor (GMCSF) and macrophage colony stimulating factor (MCSF), may also be important in MDSC accumulation (Dolcetti et al., 2008; Hamilton, 2002; Menetrier-Caux et al., 1998; Rossner et al., 13 2005). Both monocyte chemotactic protein-1 (MCP-1, CCL2) and its corresponding receptor, CCR2, were found to be critically important in the migration of MDSC to the spleen and tumor site in a number of cancer models (Boelte et al., 2011; Huang et al., 2007). As aforementioned, non-tumor inflammatory responses, like those to antigen challenge or parasitic infection, are also capable of upregulating MDSC. Sinha et al. have reported the importance of the pro-inflammatory cytokines interleukin-1β, IL-6 and prostaglandin E2 in MDSC accumulation (Cheng et al., 2011; Obermajer et al., 2011a; Sinha et al., 2007b; Tu et al., 2008). These molecules are produced in inflammation well before any signs of related cancer and therefore Sinha et al. have hypothesized that inflammation may drive cancer development through the induction of MDSC, which diminish immune surveillance against pre-malignant and malignant cells (Obermajer et al., 2012; Sinha et al., 2007b). While work in recent years has helped recognize a number of factors that may be important in the accumulation and recruitment of MDSC, no clear pathway has been elucidated that explains this process (Dolcetti et al., 2008; Kusmartsev and Gabrilovich, 2006). MDSC recruitment has not been studied specifically in CRC models. 1.2.3 MDSC Function MDSC are involved in tumor immune evasion and progression processes both systemically and locally at the site of the tumor. They are capable of suppressing the function of various anti-tumor immune cells (Dolcetti et al., 2008; Kusmartsev 14 and Gabrilovich, 2006). MDSC have been shown to inhibit tumor specific and non-specific T-cells by inducing anergy or deletion. T-cells require L-arginine for their proper function and proliferation (Raber et al., 2012; Rodriguez et al., 2010). As shown in figure 1-2B, this process is dependent on the generation of radical species and L-arginine metabolism by the enzymes arginase 1 and nitric oxide synthase 2 (Dolcetti et al., 2008). MDSC also systemically inhibit natural killer cells via downregulation of perforin, and alpha-galactosylceramide-activated (anti-metastatic) natural killer T cells (Liu et al., 2007; Yanagisawa et al., 2006). The mechanism by which MDSC can induce antigen specific tolerance in CD8+ T cells has been described (Nagaraj et al., 2007). MDSC induce nitration of the tyrosine residues in the CD8 and T cell receptor complex in these cells, thereby inhibiting their ability to bind to specific peptide-major histocompatibility complex and respond to specific antigens (Figure 1-2A). The ability to respond to nonspecific stimulation in these CD8+ T cells remains functional. The nitration process is likely induced by release of reactive oxygen species and peroxinitrite during direct cell-to-cell contact (Nagaraj et al., 2007). MDSC also create a favourable microenvironment for tumor progression (Dolcetti et al., 2008; Kusmartsev and Gabrilovich, 2006). This is largely accomplished by the expression of a number of proangiogenic and prometastatic molecules such as VEGF, matrix metaloproteinases and basic fibroblast growth factor (Dolcetti et al., 2008). VEGF is a potent angiogenic and vasculogenic agent that has been linked to poor prognosis in a number of cancers (Patan, 2004). 15 Figure 1-2 Methods of MDSC-mediated immune suppression This figure, adapted from Nature Reviews Immunology by Gabrilovich and Nagaraj, shows the key mechanisms used by MDSC to suppress the immune system (Gabrilovich and Nagaraj, 2009). (A) In the periphery MDSC disrupt the function of tumor antigen specific CD8+ T-cells by binding with CD8+ cells in an antigen-specific manner and then disrupt the T-cell’s function by releasing radical oxygen and nitrogen species into the local microenvironment. (B) At the tumor site, MDSC are able to suppress the immune system in a non-specific manner by disrupting T-cell function by L-arginine depletion via arginase I and nitric oxide synthase II enzymes, as well as the generation of radical nitrogen species. 16 MDSC have been implicated in tumor refractoriness to anti-VEGF treatment, most likely due to their ability to both produce VEGF and to enhance its bioavailability by preventing breakdown (Shojaei et al., 2007). Interestingly, Yang et al. showed the ability of MDSC to enhance vascular density and maturation as well as decreased necrosis when directly injected into tumors and also that some of the MDSC themselves were directly incorporated into the endothelium (Yang et al., 2004). MDSC are thought to have a regulatory role in inflammation by limiting the inflammatory response via the same immunosuppressive mechanisms that make them pro-tumorigenic (Dolcetti et al., 2008; Gabrilovich and Nagaraj, 2009). A large body of work in recent years has shown the increased levels and a positive immunosuppressive role for MDSC in inflammation related conditions ranging from hepatitis and airway infections to trauma and sepsis (Arora et al., 2011; Cheng et al., 2011; Hegde et al., 2011; Makarenkova et al., 2006; Rodriguez et al., 2010; Sander et al., 2010). An immunosuppressive role has been proposed for MDSC in two different models of colitis (Haile et al., 2008; Singh et al., 2012b). Haile et al. also reported increased MDSC levels the peripheral blood of IBD patients and suggested an immunoregulatory role for MDSC in IBD (Haile et al., 2008). Many mechanisms of action of MDSC in inflammation and cancer models have been reported recently, but much work remains to be done to study and 17 confirm their role in specific cancers. The goal of this project is to study the role of these cells in an animal model of IBD and related cancer. 1.3 Interleukin-10 and Toll like receptor 4 deficient model 1.3.1 Models of Colitis associated cancer Much of the recent progress in our understanding of IBD and colon cancer can be attributed to the development of various animal models. A number of inducible and spontaneously occurring models have been developed to study both conditions, each with its own strengths and weaknesses (Kanneganti et al., 2011; Strober and Fuss, 2006; Wirtz and Neurath, 2007). Mouse models of IBD can be divided into three groups based on the defect in mucosal immunity that would be considered the key to the onset of inflammation: (1) defects in epithelial integrity, (2) defects in innate immunity and (3) defects in adaptive immunity. Chemically induced models of IBD tend to cause severe inflammation via defects in barrier function of the epithelial layer. Spontaneously occurring models of IBD are based on genetic defects that can affect any of the aforementioned criteria for disease onset (Strober and Fuss, 2006; Wirtz et al., 2007; Wirtz and Neurath, 2007). Like the human disease, not all models of colitis reliably develop cancer. A number of spontaneous and inducible models of colitis-associated cancer have been developed (Kanneganti et al., 2011). There are essentially three types of colitisassociated cancer mouse models: (1) a tumor suppressor gene knockout combined with an inflammation inducing chemical or infectious agent, (2) 18 chemical or infection induced inflammation which is often combined with a carcinogen to accelerate cancer development and (3) genetic knockouts that develop inflammation and cancer, but may also be combined with a carcinogen. Common tumor suppressor gene and inflammation inducing combinations include DSS treatment of with genetic knockouts such as p53 deficient and APCmin mice (Fujii et al., 2004; Tanaka et al., 2006). Cancer development mutations such as RAG2 can also be combined with infectious models such as Helicobacter hepaticus to induce inflammation-associated cancer (Erdman et al., 2003). Inducible models include chronic doses of inflammation causing agents such as carrageenan and dextran sulfate sodium (DSS) (Okayasu et al., 2002; Tobacman, 2001). These models can take a long time to develop neoplasia, but combining with doses of carcinogens such as axozymethane (AOM) and 1,2dimethylhydrazine can speed up cancer developmen (Kohno et al., 2005; Tanaka et al., 2003). The spontaneously occurring models of colitis-associated cancer involve mutations in key inflammatory and/or cancer pathways. The antiinflammatory cytokine interleukin-10 deficient mouse develops a CD like inflammation and adenocarcinomas (Zhang et al., 2007b). Mice deficient in Gα M, a g-protein signal transduction molecule, develop a UC like colitis and related carcinoma (Rudolph et al., 1995). Other spontaneously occurring models of colitis-associated cancer include, T-cell receptor βcell receptor eously occurring models of colitis-associated cancer include, Tted carcinoma playText>(Zhang et al., (Erdman et al., 2003; Kado et al., 2001). Recently, IL-10 and tumor necrosis 19 factor double deficient mouse has been described as developing a UC like inflammation and associated cancer (Hale and Greer, 2012). While a number of animal models of IBD related cancer have been developed, there is no prototypical animal model that can simulate every aspect of the human disease and each model has its own advantages and disadvantages. We are interested in the IL-10 deficient mouse model as the cancer arises from a chronic inflammatory driven by bacterial antigen, much like clinical IBD. 1.3.2 The Interleukin 10 Deficient Mouse The interleukin-10 knockout (IL-10-/-) mouse is a spontaneously developing model of IBD due to an aberrant immune response to normal enteric antigens (Kuhn et al., 1993; Rennick et al., 1995; Strober and Fuss, 2006). The loss of interleukin-10, an anti-inflammatory cytokine, leads to the development of severe enterocolitis characterized by massive mucosal infiltration by lymphocytes, and activated macrophages and neutrophils, mucosal lesions, abnormal mucosal architecture, thickening of the mucosal wall and even the thickening of the muscle layer in severely inflamed animals (Rennick et al., 1995; Strober and Fuss, 2006). The inflammation in these animals generally becomes apparent at the age of 8-12 weeks due to weight loss and predominantly affects the colon (Zhang et al., 2007b). The lack of IL-10 does not affect the normal development of T and B lymphocytes (Rennick et al., 1995). This model is dependent on 20 bacterial antigen stimulation because the inflammation can be altered in specific pathogen free environments and avoided entirely in a germ free environment (Rennick et al., 1995; Strober and Fuss, 2006). IL-10 is an anti-inflammatory cytokines with potent suppressor effects on TH1 cells and macrophage effector functions. A number of in vitro studies have established the ability of IL-10 to inhibit the production of inflammatory cytokine such as IL-12 and Tumor necrosis factor α (TNF-α), suppress the expression of various co-stimulatory molecules, inhibit T-cell proliferation and even polarize T-cell differentiation towards regulatory T-cells (Kuhn et al., 1993; Moore et al., 2001; Rennick et al., 1995; Strober and Fuss, 2006). The development of IBD in IL-10 deficient mice is likely due to the lack of IL-10 production by T-cells because a very comparable phenotype is observed in T-cell specific IL-10 knockout animals. Anti-IL-10 antibody treatment and IL-10 receptor β chain knockouts also have similar disease development as IL-10 knockout animals (Roers et al., 2004; Wirtz and Neurath, 2007). Neutralizing antibody treatment against any one of the cytokines that are modulated by IL-10, such as TNF-α, does not have any significant ablating effect on the disease in IL-10-/- animals. IL-10 cytokine treatment on the other hand has an ameliorating effect in IL-10 deficient mice (Rennick et al., 1995). Polymorphisms in the IL-10 gene have been related to both UC and CD development in a number of studies (Andersen et al., 2010; Fernandez et al., 2005; Zhu et al., 2013). Recently, polymorphisms in the IL-10 gene and IL-10 receptor gene have been linked to a severe form early onset IBD (Glocker et al., 21 2009; Kotlarz et al., 2012; Moran et al., 2013). In these patients, the lack of IL-10 or IL-10 receptor is likely important in myeloid derived cells, as a hematopoietic stem cell transplant is often curative (Engelhardt et al., 2013). The IL-10 deficient mouse model is a good model for studying IBD-associated cancer because it highly correlates with the human disease. The spontaneous occurrence of inflammation due to an uncontrolled immune response that is driven by bacterial antigen stimulation simulates the manner in which clinical IBD develops. Histologically and immunologically, this model resembles Crohn’s Disease, where discrete areas of inflammation are spread among normal mucosa and are believed to occur due to a TH1 cytokine and TH17 cytokine profiles (Rennick et al., 1995; Sartor, 2006; Strober and Fuss, 2006; Wirtz and Neurath, 2007; Zhang et al., 2007b). Cancer incidence in the IL-10 deficient mouse increases over time, as is the case in IBD patients (Rutter et al., 2004). MDSC have also recently been linked to inflammation in the IL-10 deficient mouse (Singh et al., 2012b). 1.3.3 Toll like receptor 4 Luminal bacterial antigens drive the pathogenesis of IBD and related colon cancer in both humans and animals (Hajishengallis et al., 2012; Tjalsma et al., 2012). In order to detect foreign antigens, the body uses sensory receptors 22 known as pattern recognition receptors. Pattern recognition receptors, such as the toll like receptor (TLR) family, have evolved to identify evolutionarily conserved molecular patterns in microbes, known as microbial-associated molecular patterns (Medzhitov and Janeway, 1999). Thirteen TLRs, which are members of the interleukin 1 family of receptors, have been discovered and they detect a variety of bacterial derived antigens ranging from flagellin to bacterial DNA. All TLRs are transmembrane proteins, with some residing on the surface of intracellular compartments. Stimulation of TLRs via their appropriate ligands results in the activation of a potent immune response, generally mediated by the adaptor molecule that is associated with all but one of the TLRs, Myeloid differentiation primary response gene 88 (MyD88). MyD88 activation leads to the activation of the nuclear factor κ B (NF-κB) pathway that causes production of pro-inflammatory cytokines and changes in surface protein expression (Fukata and Abreu, 2007; Fukata and Abreu, 2008; Takeda and Akira, 2004). Toll like receptor 4 (TLR4), which senses lipolysaccharide from the outer membrane of gram negative bacteria, was the first TLR to be discovered and is the best studied pattern recognition receptor, but its role in both intestinal inflammation and in related cancers remains controversial (Abreu et al., 2005; Fukata and Abreu, 2007; Zhang et al., 2007a). The intracellular events following the stimulation of TLR4 involves multiple mechanisms. In addition to MyD88 mediated NF-κB pathway, a second pathway mediated by TIR-domain 23 containing adapter-inducing interferon-β (TRIF) is also activated. The TRIF pathway leads to the nuclear localization of interferon regulatory factor 3 (IRF-3), which mainly causes the increased transcription of α and β interferons that are important in the induction of the immune response. This pathway can also further activate the NF-κB pathway, independent of MyD88 in a relatively delayed fashion (Takeda and Akira, 2004). Normally TLR4 signaling is downregulated in the epithelia following birth, which may reflect a tolerance mechanism to luminal colonization, however expression is modulated in epithelia and lamina propria cells in chronic inflammation and in malignancy (Abreu et al., 2005). Since TLRs play such an important role in mediating the immune response, it was initially postulated that suppressing either TLR4 or MyD88 function would decrease inflammation and alleviate IBD symptoms, but generally the reverse has been reported (Fukata et al., 2005). In the dextran sodium sulfate (DSS) induced model of colitis, TLR4 deficient mice undergo increased weight loss and rectal bleeding as well as increased susceptibility to bacterial translocation. The bacterial translocation is explained by the lack of a potent immune response in colon, which leads to decreased bacterial clearance. The increased bleeding may be caused by increased apoptosis accompanied by the loss of epithelial proliferation for which TLR4 is partially responsible (Abreu et al., 2005; Fukata and Abreu, 2007). These data taken together, suggest a critically important homeostatic role for TLR-4 that may be more important than its pro-inflammatory 24 role in the pathogenesis of IBD (Fukata and Abreu, 2007; Gonzalez-Navajas et al., 2010; Rakoff-Nahoum and Medzhitov, 2009; Zhang et al., 2007a). 1.3.4 IL-10 TLR4 Double Deficient Mouse Our laboratory has developed a novel IL-10 TLR4 double deficient mouse. Previous work from our lab shows that IL-10 TLR-4 double deficient mouse has greatly increased incidence of cancer. This is associated with a small increase in inflammation. In our hands, IL-10 knockouts develop adenocarcinoma in about 20% of mice at 6 months of age. We have found that 50% of double deficient mice develop cancer by just 3 months of age (Zhang et al., 2007a). The role of TLR-4 in the pathogenesis of inflammation-associated cancer remains unclear. While we report an increased incidence of cancer in absence of TLR-4 in the IL10 deficient model, others have reported a decrease in the induced chronic DSS and AOM model of colitis-associated cancer (Fukata et al., 2007). This is especially interesting, considering Greten et al. found that the main NF-κB pathway, which is activated by a number of mechanisms including TLR-4, in myeloid derived leukocytes is important for the growth and progression of the cancer in the same induced chronic DSS and azoxymethane model. They concluded this after finding that knocking out the main NF-κB pathway specifically in myeloid cells caused the average size of the tumor to be much decreased compared to controls (Greten et al., 2004). Alternations in the NF-κB 25 in signaling pathway, which would alter chemokine levels and affect leukocyte recruitment, appears to be very important for inflammation associated cancer development (Greten et al., 2004). We have found that IL-10 TLR-4 double deficient mice have significant changes in the message levels of various chemokines, including RANTES (regulated upon activation, normal T-cell expressed and secreted) and Macrophage inflammatory protein-1r, as compared to IL-10 deficient mice (Zhang et al., 2007a). Interestingly, our preliminary data showed that recruitment of MDSC into the colon is increased in IL-10 deficient mice as compared to wild type counterparts. MDSC recruitment was further increased IL-10 TLR4 double deficient as compared to IL-10 deficient mice. 1.4 Hypothesis and Objectives Inflammatory bowel disease is a key driver of colorectal cancers. Inflammatory pathways in both epithelial cells and leukocytes are associated with colon cancer initiation and development. MDSC are a protumorigenic cell type, whose peripheral recruitment correlates with colon cancer progression. MDSC have also been linked to inflammation in the gut, but their role remains unclear. We have found increased recruitment of MDSC of IL-10 deficient mouse model of colitisassociated cancer. MDSC recruitment is further increased in the IL-10 TLR4 double deficient mouse, which has a much-increased incidence of cancer. The study of MDSC in colon cancer in warranted due to their presence in high numbers and immunosuppressive properties, 26 despite their potentially immunosuppressive role in inflammation. The regulation and recruitment mechanisms involved are important and potential targets for therapeutic manipulation. The hypothesis for the work presented in this thesis is that myeloid derived suppressor cells exacerbate cancer development in colitis-associated cancer and TLR4, an important regulator of intestinal homeostasis and leukocyte recruitment, regulates recruitment and/or function of myeloid derived suppressor cells in the gut. Objectives: Ø To determine the role of MDSC in cancer development in the IL-10 deficient model Ø To determine whether TLR4 regulates cancer development and MDSC recruitment in the IL-10 deficient mouse model Ø To determine if TLR4 modulates MDSC function Ø To identify MDSC chemoattractants in vitro and in vivo 27 Chapter 2: Materials and Methods 28 2.1 Animals All experiments were approved by the Animal Care Committee of the University of Calgary and conform to the guidelines established by the Canadian Council for Animal Care. Mice were bred and housed in a specific pathogen-free environment on standard chow diet at the University of Calgary. IL-10-deficient (IL-10-/-) mice were generated by gene targeting in embryonic stem cells as previously described by Kuhn et al. on the 129SvEv background (Kuhn et al., 1993). IL-10 deficient mice were originally obtained from Dr. R Fedorek (University of Alberta, Canada). Mice deficient in IL-10 and TLR4 were generated from IL-10-/- mice heterozygous for TLR4 (IL-10-/- TLR4+/-). TLR4 deficient mice (Jackson Labs, Bay Harbour, USA) on the C57BL/10ScN background were backcrossed for a minimum of 8 generations onto the IL-10-/- 129SvEv background before breeding heterozygotes (IL-10-/- TLR4+/-). All mice were genotyped for IL-10 and TLR4 by PCR analysis for genomic DNA purified from tail biopsies with the DNeasy tissue kit (Qiagen, Germany). Following PCR, the amplified DNA was separated by gel electrophoresis run in a 1.5% agarose gel with ethidium bromide and visualized under ultraviolet light. The IL-10-/- sense primer was GCCTTCAGTATAAAAGGGGGACC and the anti-sense primer was GTGGGRGCAGTTATTGTCTTCCCG. The TLR4-/- sense primer sense primer was GCAAGTTTCTATATG the anti-sense primer was CCTCCATTTCCAATA. IL10-/- TLR4-/- breeding pairs and IL-10-/- TLR4+/+ breeding pairs were set up. Age 29 and sex matched mice housed in the same facilities were used in all experiments. 2.2 Evaluation of Inflammation and Cancer Mice were sacrificed by cervical dislocation. Hyperplasia and inflammation were evaluated macroscopically in excised colons. Rolled excised mouse colons were fixed in 10% formalin solution or frozen in optimal cutting temperature compound (OCT, VWR, West Chester, USA). Formalin fixed sections were dehydrated using graded alcohol immersions and embedded in paraffin wax. Colon sections (5-8μm) were hydrated and labeled with hematoxilyn and eosin (H&E). Dysplasia and inflammation were evaluated histologically in H&E labeled sections. Paraffin sections were used to evaluate morphology and evaluate dysplasia. OCT frozen sections were used for experiments where we had to both evaluate dysplasia and do immunohistochemistry for MDSC levels. Ø Macroscopic analysis of hyperplasia Mucosal hyperplasia in the colon was assessed macroscopically in excised colons via a previously published scoring system (Zhang et al., 2007b). Polyp scores were assigned according to the number of polyps observed with a score 30 of 0 for none, 1 for one to three individual polyps, 2 for three to six individual polyps and 3 for six or more polyps, merged polyps or raised plaques. Ø Macroscopic evaluation of Colonic Inflammation Inflammation was assessed macroscopically in excised colons via a previously published scoring system (Zhang et al., 2007b). The presence of each of the following variables was given a score of 1: erythema, edema, diarrhea, stricture, hemorrhage, adhesions and ulceration. Bowel wall thickness was measured in millimeters and added to the score. Ø Histological analysis of Dysplasia Dysplasia was evaluated histologically in colon sections by a pathologist, Dr. Stephan Urbanski (Alberta Health Services), and myself in a blinded fashion using a previously published scoring system based on changes in crypt architecture, epithelial dysplasia, presence of submucosal invasion and serosal adhesions (Zhang et al., 2007b). Crypts were given a score of 0 for normal, 1 for goblet cell depletion, 2 for branching, irregular, dilate lumen or back-to-back glands and 3 for complex budding and crypt reforming. Epithelia were given a score of 0 for normal, 1 for hyperplasia or aberrant crypt foci, 2 for low-grade dysplasia defined as nuclear enlargement, mild hyperchromatism, nuclear 31 crowding with stratification and 3 for high-grade dysplasia defined as nuclear stratification, prominent hyperchromatism, pleomorphism, loss of nuclear polarity. Submucosal invasion and serosal adhesions were scored as 1 each when present. Adenocarcinoma was defined as submucosal invasion by highly dysplastic crypts. Ø Histological evaluation of Colonic Inflammation Colitis was scored histologically in a blinded manner based on a previously published scoring system based on the level of damage to the normal mucosa (Zhang et al., 2007b). Scores were based on damage to normal mucosal architecture (0-3, based on severity), degree of cellular infiltration (0-3, based on severity), extent of muscle thickening (0-3, based on severity), loss of goblet cells (1) and presence of crypt abscesses (1). 2.3 Colon Immunohistochemistry for Gr1+ CD11b+ Cells To identify Gr1+CD11b+ cells in the colon we used immunohistochemistry in OCT frozen sections. This protocol was developed with the help of Dr. Pina Colarruso of the Live Cell Imaging facility at the University of Calgary. Entire colons were rolled and frozen in OCT. Cryosections (5-8 μ5) were fixed in cold acetone for 1 hour and incubated overnight at 4°C with 2μg/ml FITC conjugated rat anti- 32 mouse Gr1 antibody (clone RB6-8C5, eBioscience, Sand Diego, USA). Next, sections were incubated for 1 hour with 1% anti-rat Alexa488 anti-body to enhance the FITC signal. Sections were blocked with 10% normal rat serum for 1 hour and incubated overnight at 4°C with 2ug/ml PE conjugated rat anti-mouse CD11b antibody (clone M1/70, eBioscience, Sand Diego, USA). There was a 1X PBS wash step repeated 3 times between each step. Appropriate combinations of rat anti-mouse and rat isotype antibodies were used to control against nonspecific binding (Please see chapter 3.2 for more detail). Double labeled cells were counted following imaging of whole tissue sections on an Olympus IX70 inverted epi-fluorescence microscope (Olympus, Centre Valley, USA). Images were recorded with QImaging Retiga Exi 12-bit CCD camera (model 32-00558A150, QImaging, Surrey, Canada) and Volocity software (versions 4-6, Perkin Elmer, Waltham, USA). The total area of mucosal and submucosal tissue was determined using serial H&E stained tissue sections. Double positive cells were quantified and standardized per unit area (st2) of the mucosa and sub mucosa. 2.4 Flow Cytometry Experiments All flow cytometry experiments were conducted with an Attune Acoustic Focusing Cytometer (Life Technologies, Burlington, Canada) in the Flow Cytometry Core Facility, University of Calgary. Results were analyzed with Attune Cytometric Software (Life Technologies, Burlington, Canada). 33 Ø Flow Cytometry of Lamina Propria Gr-1+ CD11b+ MDSCs Flow cytometry was used to quantify Gr1+CD11b+ cells in the lamina propria of colon tissue. Excised entire colons were cut into 5 mm pieces. These pieces were washed five times with 20 mM HEPES HBSS solution by gentle stirring. Colon pieces were subjected to EDTA digestion to remove epithelium by gently stirring at 37°C in 20 ml of 5 mM EDTA 10% FBS HBSS solution for 15 minutes (4X). Next, tissues were gently stirred at 37°C in 20ml 100 units of activity/ml Collagenase Type II (Invitrogen, Grand Island, USA) in Complete RPMI for 1 hour. The collagenase digestion process that breaks down the mucosa into individual cells was repeated for a total of 3 times with lamina propria cells (supernatant) being collected and washed in Complete RPMI 1640 between digestions. After the last diegestion, cells were resuspended in 1X PBS and 1 x 106 cells were incubated with either 2x10-7 μg of FITC conjugated isotype (IgG2a, eBioscience, Sand Diego, USA) antibody, PE conjugated isotype (IgG2a, eBioscience, Sand Diego, USA) antibody, FITC conjugated rat anti-mouse Gr-1 antibody (clone RB6-8C5, eBioscience, Sand Diego, USA), PE conjugated rat anti-mouse CD11b antibody (clone M1/70, eBioscience, Sand Diego, USA) or both FITC conjugated Gr1 and PE conjugated CD11b antibodies. Flow cytometry analysis was performed to determine the percentage of Gr1+CD11b+ cells. The percentage of double positive cells was determined in the population gated for granulocytes and monocytes in the side and forward scatter plot. 34 Ø Flow Cytometry of Bone Marrow Gr-1+ CD11b+ MDSC Bone marrow cells were isolated by flushing of excised femur and tibia bones with 1X PBS and washed. Cells were labeled as previously described and flow cytometry analysis for the percentage of MDSC was performed. Ø Flow Cytometry of Gr-1+ CD11b+ MDSC Splenocytes Excised spleens were washed in 1X HBSS and mechanically disrupted in complete RPMI. A single cell suspension was obtained with a 100 μm cell strainer. Cells were labeled as previously described and flow cytometry analysis for the percentage of MDSC was performed. Ø Total and Differential White Blood Cell Counts Blood was collected from anesthetized mice via cardiac injection. 50 μl of blood was added to 440 μo of 3% acetic acid and 10 μo of 5% crystal violet solution. This solution was used to count the total number of white blood cells using a hemocytometer. Two droplets of blood were placed on a slide and a second slide was used to smear the blood across the first slide. Blood smears were stained using Hemacolor Harleco Kit (EMD Science, Missisauga, Canada). A total of 100 35 white blood cells were counted in a random manner on each slide and categorized into granulycyte, monocyte or lymphocyte like cells. Ø Lamina Propria Cell Leukocyte Differential Colonic lamina propria cells were isolated as previously described. 1 x 106 cells were each incubated with either 2x10-7 7g of CD4 (T-cells, Clone GK1.5), CD19 (B-cells, Clone eBio1D3), CD11c (dendritic cells. Clone eBio1D3), F4/80 (monocytes, Clone BM8) or isotype (IgG2aI and IgG2bn) antibodies (eBioscience, St Louis, USA). Gr1+CD11b+ cell flow cytometry was performed as previously described. The percentage of each cell type was determined by flow cytometry analysis with appropriate forward/side scatter plot gating for each population. Ø TLR4+ MDSC Flow Cytometry Gr1+CD11b+ cells were isolated from the bone marrow of IL-10-/- and IL-10-/TLR4-/- mice and Gr1+CD11b+ cells flow cytometry was performed as previously described. Cells were also labeled with allophycocyanin attached anti-mouse TLR4 antibody (Clone MTS-510, eBioscience, San Diego, USA) or triple labeled with anti-mouse Gr1, CD11b and TLR4 antibodies. Triple positive cells were determined by looking for TLR4+ cells in the population gated through Gr1+CD11b+ cells. 36 2.5 MDSC Isolation Ø Bone Marrow derived Gr1+CD11b+ cell Isolation Gr1+CD11b+ cells were isolated from the bone marrow of 6-month-old IL-10-/mice via percoll density gradient centrifugation as previously described (Kusmartsev et al., 2000). Briefly, femurs and tibiae of mice were removed and the marrow was washed out with 1X PBS. Gradient layers (2 ml each) of 100 % percoll with cells, 70 % percoll, 60 % percoll, 50 % percoll and 1X HBSS were centrifuged at 1800 x g for 30 minutes. Cells were collected from the interface between the 50 % and 60 % and washed with 1X PBS. Using flow cytometry analysis, we determined that Gr1+CD11b+ cell purity in isolates varied between 82-95 %. Ø Spleen derived Gr1+/CD11b+ Cell Isolation Gr1+CD11b+ cell isolation using the percoll gradient density centrifugation as described yielded a low purity of cells. We used a Myeloid-Derived suppressor cell isolation kit (Miltenyi Biotech, Cologne, Germany) to obtain a purer population as previously described (Brudecki et al., 2012). MDSC are a heterogenous group of cells with a number of sub-populations within them, which differ in function and phenotype (Gabrilovich and Nagaraj, 2009). The Gr1 epitope is found in both Ly-6G and Ly-6C glycoproteins. This kit allows for the isolation of either polymorphonuclear Gr-1 37 high + Ly-6G and mononuclear Gr- 1 1 dim dim – Ly-6G myeloid cells. We chose to isolate and study mononuclear Gr– Ly-6G myeloid cells from the spleen because mononuclear cells were the majority in MDSC isolated from the bone marrow. Briefly, splenocytes were isolated by mechanical digestion of spleens and incubated with Ly-6G Biotin and anti-biotin microbeads. Labelled cells were passed through magnetic MACS columns and negative cells were collected. Collected cells were incubated with anti-Gr1 biotin and streptavidin microbeads. Anti-Gr1 labeled cells were passed through MACS columns and positive cells were collected. Flow cytometry analysis of isolated cells showed greater than 92 % purity of Gr1+CD11b+ cells. MDSC Function in vitro Functional activity of Gr1+CD11b+ cells isolated from the bone marrow and spleen was measured in cells isolated from 3-month-old mice. Ø Arginase I Activity Assay Gr1+CD11b+ cells isolated from the bone marrow or spleen were tested for arginase I activity by measuring the production of L-ornithine and urea from Larginine (Rodriguez et al., 2004). Gr1+CD11b+ cells (5 x 105) were lysed in RIPA buffer (Radio-immunoprecipitation assay buffer, Sigma Aldrich, St. Louis, USA) for two hours. Cell lysates were added to 25 μl of Tris-HCl (50 mM; pH 7.5) 38 containing 10 mM MnCl2. This mixture was heated at 55–60°C for 10 min to activate the arginase I enzyme. Next, a solution containing 150 μl carbonate buffer (100mM; Sigma Aldrich, St. Louis, USA) and 50μl L-Arginine (100mM, Sigma Aldrich, St. Louis, USA) was added and incubated at 37°C for 20 min. The hydrolysis reaction from L-Arginine to urea was detected with diacetyl monoxime (Sigma Aldrich, St. Louis, USA) by spectrophotometry (540nm) following incubation at 95°C for 10 min. A standard curve for urea (Sigma Aldrich, St. Louis, USA) was used to convert absorbance values into μa urea. RIPA buffer was used as a negative control. Hepatocyte cell lysates, which are known to express high levels of arginase I enzyme, were used as a positive control. Ø Nitrite measurement Nitric oxide is produced from L-arginine by the nitric oxide synthase 2 (NOS2) enzyme. This reaction forms nitrates and nitrites as stable end products. The Griess reaction was used to measure total nitrite levels in Gr1+CD11b+ cell supernatants as an indication of nitric oxide synthase activity, after reduction of nitrates to nitrites as previously described (Miranda et al., 2001). The Griess reaction assay was performed using a kit according to manufacturer’s instructions (Invitrogen, Grand Island, USA). Briefly, Gr1+CD11b+ cells, isolated from the bone marrow or spleen, were cultured in 10% FBS (Invitrogen, Grand Island, USA) 2% PenStrep (Sigma Aldrich, St. Louis, USA) DMEM medium 39 (Invitrogen, Grand Island, USA) for 18 hours at 37°C. In 96 well plates, triplicate samples of sodium nitrate standards (100 μ.) or supernatant (100 μr) from 5 x 105 cultured Gr1+CD11b+ cells were studied. Vanadium III chloride (VCl3, Sigma Aldrich, St. Louis, USA) solution (80μo of 400 mg VCl3 in 50 mL of 1M HCl) was added to each well to reduce nitrates to nitrites followed by the Griess reagent (20 μr). Griess reagents are sulphanilic acid, which reacts with nitrites to form a diazonium salt, and N-alpha-naphthyl-etheylenediamine, which reacts with the diazonium salt and changes colour to pink. The plate was incubated at 37°C for 1 h and before the colourimetric changes were read at 550 nm using a spectrometer. Ø T-cell Proliferation Suppression Gr1+CD11b+ cells (1 x 105) isolated from the bone marrow or spleen were cultured in RPMI 1640 with CD4+ T-cells (5 x 105) isolated from spleens by T-cell enrichment kit (BD Biosciences, San Jose, USA) and stained with CFSE dye (Invitrogen, Grand Island, USA). T-cells proliferation was stimulated with 1 μg/ml anti-CD3 (Clone SPV-T3b, Invitrogen, Grand Island, USA) and 500 ng/ml antiCD28 (Clone 37.51.1, Invitrogen, Grand Island, USA) antibody for 72 hours. CFSE is equally divided between the daughter cells of proliferating cells, thereby lowering the total amount of dye in daughter cell. The ability of MDSC to 40 suppress T-cell proliferation was determined by flow cytometry analysis for the percentage of CFSE low cells. Unstimulated T-cells were used as controls. 2.7 Bone Marrow Transplant In order to study the role of TLR4 within leukocytes on cancer development we generated chimeric mice. Bone marrow chimeric mice were generated as previously described (Carvalho-Tavares et al., 2000). IL-10-/- and IL-10-/- TLR4-/mice at 3 months of age were irradiated with two doses of 5 Gy y th twGammacell, 137 Cs γ-irradiation source, Nordion International, Kanata, Canada) given three hours apart. Bone marrow was collected from donor IL-10-/- and IL10-/- TLR4-/- mice as previously described. Two hours after irradiation, the recipient mice were injected with 8 x 106 donor bone marrow cells from the opposite genotype via tail vein. Control mice were irradiated and then injected with bone marrow cells from donor mice of the same genotype. Donor mice were all males and recipients were all females to allow us to confirm that the transplant was successful. Irradiated mice were given 2% neomycin (Sigma Aldrich, St Louis, USA) in drinking water for 2 weeks. Mice were sacrificed at 6 months of age and inflammation and cancer development were evaluated as previously described. 2.8 Molecular Techniques 41 Ø Real time PCR TLR4, MCP-1 and SDF-1 message levels were determined in mouse colon tissue from ileo-cecal junction and distal colon by real time polymerase chain reaction as previously described (Zhang et al., 2007b). Briefly, total RNA was isolated from colon tissues using QIAzol (Qiagen Science, Missisauga, Canada). The RNA was further purified using Deoxyribonuclease Message Clean kit (Genhunter Corporation, Brookline, USA) according to manufacturer instructions. Total RNA was determined by measuring absorbance at 260 nm. Complimentary DNA (cDNA) was generated amplified using standard polymerase chain reaction (Ayala-Torres et al., 2000). Primers and probes for mouse TLR4, MCP-1 (CCL2) and SDF-1 (CXCL12) were purchased from Applied Bioscience (Life Technologies, Burlington, Canada). Three replicates of cDNA (1:5 diluted) were amplified using the ABI Prism 7900HT Sequence Detection System (Life Technologies, Burlington, Canada). The amplification reaction mixture contained 5 μl of cDNA, 5 μ of probe and primer mix and 2X Universal Master Mix (Life Technologies, Burlington, Canada). GAPDH was included as an internal control to normalize for varying quantity of cDNA, using 20X GAPDH Mix (Life Technologies, Burlington, Canada). Thermocycler parameters were 50°C for 2 min, 95°C for 10 min, 40 cycles of 95°C for 15 s and 60°C for 1 min. The results were analyzed using Applied 42 Biosystems RQ manager (Life Technologies, Burlington, Canada). Ø Western Blot In order to determine the levels of TLR4, MCP-1 and SDF-1 protein in mouse colons, western blots were performed as previously described (Khajah et al., 2013). 50 mg of excised colon tissue was lysed in RIPA buffer with 1:100 mamalian protease inhibitor cocktail (Sigma Aldrich, St Louis, USA). The supernatant was added to equal volume 2X SDS lysis buffer plus 100 μb PMSF, 100 μP Na3VO4, 10 μ 10F Aprotinin/Leupeptin, and 10 μ0 pepstatin. The samples were boiled for 10 minutes at 90°C and lysates were loaded into 10% sodium dodecyl sulfate-epolyacrylamide (SDS-PAGE) gel. After electrophoresis at constant voltage of 100V for 1 hour, proteins were transferred to a nitrocellulose membrane and were blocked with 5% milk for 1 hour. TLR4 (Clone 267518), MCP-1 (Clone 123616) and SDF-1 (Clone 79018) anti-mouse antibodies were purchased from R&D Systems, Minneapolis, USA. The membrane was incubated overnight at 4°C with 1/1000 anti-mouse antibody in 5% mlik. The next day, the membrane was washed and incubated for 1 hour with 1/10,000 dilution of anti-rabbit HRP-conjugated secondary antibody. The membrane was developed with super signal enhanced chemiluminescence and visualized with Kodak x-ray film. Bands were semi-quantitatively analyzed by 43 measuring density over β-actin using ImageJ (National Institutes of Health, Bethesda, USA). Ø Immunoprecipitation We were unable to visualize any protein bands for SDF-1 via western blot. To increase the SDF-1 protein for western blot we used immnoprecipitation-western blot technique. SDF-1 protein was pulled down using an immunoprecipitation kit according to manufacturer’s instructions (Cell Signaling Technologies, Danvers, USA). Whole excised colons were lysed in RIPA buffer with 1:100 mamalian protease inhibitor cocktail. 200 μc of lysed tissue was added to 200 il. 200 cktail. A agarose beads slurry. After a one hour incubation at 4°C the mixture was centrifuged and the supernatant was collected. 200μl of the supernantant was incubated with 1:1000 anti-mouse SDF-1 antibody at 4°C overnight. The next day, 20μl of 50% Protein A agarose beads slurry (Cell Signaling Technologies) was added to the supernatant and incubated at with gentle rocking for 3 hours at 4°C. After 30 seconds of microcentrifugation, the pellet was washed 5 times with 500μl of the lysis buffer. The pellet was resuspended in 20μl of 3X SDS sample buffer, heated at 95°C for 5 minutes and microcentrifuged. 30μl of the sample was run through the western blot protocol as previously described. 2.9 Chemotaxis Assays 44 We studied MDSC chemotaxis in vitro in three different assays. MDSC were isolated from the bone marrow of IL-10-/- and IL-10-/- TLR4-/- mice using percoll density gradient centrifugation as previously described. Chemoattractants included WKYMVm (an fMLP like peptide, Phoenix Pharmaceuticals, Burlingame, USA), KC (keratinocyte chemotactic peptide, R&D Systems, Minneapolis, USA), MIP-2 (macrophage inflammatory protein 2, CXCL2, R&D Systems, Minneapolis, USA), MCP-1 (monocyte chemotactic protein, CCL2, R&D Systems, Minneapolis, USA), SDF-1 (stromal derived factor 1, CXCL12, R&D Systems, Minneapolis, USA), Prostaglandin E2 (R&D Systems, Minneapolis, USA), IL-6 (R&D Systems, Minneapolis, USA), VEGF (R&D Systems, Minneapolis, USA) and IL-1&D S&D Systems, Minneapolis, USA). Ø Underagarose Chemotaxis Assay Underagarose assay for MDSC cell migration was performed as previously described (Heit and Kubes, 2003). Briefly, 1.2% agarose gels were poured in small petri dishes. Three equidistant wells were formed in-line 2.5 mm apart in the gel. Cells were placed in the outer two holes with chemoattractant in the middle and incubated in 5% CO2 at 37°C for 4 hours. Migrant cells were reported by counting the number of cells moving towards the chemoattract well minus the cells moving randomly out of the chamber in the opposite direction using a Zeiss Axiovert 135 microscope. 45 Ø Ibidi Chemotaxis Assay The ibidi chemotaxis assay was performed according to the manufacturer’s instructions as previously described (Zantl and Horn, 2011). Briefly, 3 x 102 cells are placed in a small slit between two reservoirs in the proprietary ibidi μ bchemotaxis chamber. The chemoattractant was placed in one of the large reservoirs. The movement of cells in the slit was observed using a Zeiss Axiotron 100 inverted microscope by time-lapse photography at one-minute intervals for 16 hours. The number of chemotactic cells was determined by subtracting the number of cells with net movement vector away from the reservoir with chemoattractant from the cells with net movement vector towards it. Ø Transwell Chemotaxis Assay Transwell chemotaxis assay was performed using Corning Transwell plates (Corning Incorporated, Lowell, USA) that contain two chambers as previously described by us (Schicho et al., 2010). Briefly, 5 x 105 cells were placed in 100 μ 1 of complete RPMI with 10% FBS in the upper chamber, which has 8 μi pores at its bottom surface. The lower chamber had the chemoattractant in 250 μi of complete RPMI with 10% FBS. Loaded transwell plates were placed in 5% CO2 at 37°C for 3 hours. Lower chambers containing only 10% FBS complete RPMI 46 were used as controls. Same chemoattracant concentration in both upper and lower chambers was used as a control to test chemokinesis. Cells that migrated to the lower chamber were considered chemotactic. In order to confirm that migrating cells were MDSC, migrating cells were transferred onto slides using a cytospin. Slides were stained with anti-mouse Gr1 and CD11b antibodies and MDSC were counted as previously described via immunohistochemistry. 2.10 Statistical Analysis Figures were generated and statistical analysis was performed using Graphpad Prism software (Version 4-5, Graphpad Software Inc, La Jolla, USA). Data are expressed as Mean ± Standard Error of Mean. Statistical significance was determined between two groups by Student’s T-test (two tailed) or by one-way ANOVA with Bonferroni’s correction for three for more groups. A P value less than 0.05 was considered statistically significant. 47 Chapter 3: The role of MDSC in the Interleukin-10 deficient mouse 48 Summary The following experiments show that MDSC play a key role in the progression of colitis associated cancer in the IL-10 deficient mouse model. First, the presence of Gr1+CD11b+ cells in IL-10 deficient colons is established via immunohistochemistry and flow cytometry analysis. Second, the activity of key Larginine metabolism pathways and the ability to suppress T-cell proliferation is shown in Gr1+CD11b+ cells isolated from IL-10 deficient mice. The Gr1 and CD11b markers as well as the immunosuppressive phenotype confirm that these cells are MDSC. Finally, we show that the pharmacological depletion of MDSC in vivo decreases neoplastic changes and the adoptive transfer of MDSC in vivo increases neoplastic changes in the IL-10 deficient mouse. Together these experiments establish an important protumorigenic role of MDSC in the IL-10 deficient mouse model of colitis-associated cancer. 3.1 Cancer and Gr1+CD11b+ counts in the IL-10 deficient mouse This section aims to establish cancer levels and MDSC counts in the colons of IL-10 deficient mice. Adenocarcinoma incidence was studied in colon sections with the help of a blinded pathologist. MDSC levels in the colon were studied using immunohistochemistry in colon sections and flow cytometry analysis of lamina propria cells. 49 3.1.1 Colon cancer development is increased in the IL-10 deficient mouse Cancer development was studied in hematoxylin and eosin labeled colon sections by a blinded pathologist, Dr. Stephan Urbanski of Alberta Health Services and Dr. Rui Zhang. Adenocarcinoma was described as submucosal invasion by highly dysplastic crypts. Histological analysis by a trained observer using H&E labeled slides is the definitive method of evaluating dysplasia and cancer development. Other methodologies such as polyp scoring and various blood tests can only indicate the possibility of neoplasia. Dr. Rui Zhang demonstrated the increasing incidence of adenocarcinoma formation in IL-10 deficient mice as they age (Figure 3-1). In our hands, IL-10-/- mice on the 129SvEv background have a 12% incidence of colitis-associated cancer at 3 months of age. The incidence of cancer increases to 20% at 6 months of age. No cancer is observed at 1.5 months (6 weeks) of age. 50 Incidence of Adenocarcinoma (%) 25 20 15 10 5 0 1.5 Months 3 Months 6 Months Figure 3-1 Incidence of cancer increases with age in the IL-10 deficient mouse This figure illustrates the incidence of colon adenocarcinoma as a percentage in IL-10 deficient mice raised in a specific pathogen free environment at 1.5, 3 and 6 months of age. These data were generated by Dr. Rui Zhang. n>10 51 3.1.2 Gr1+CD11b+ cell recruitment to the colon is increased in the IL-10 deficient mouse Our next step was to determine whether Gr1+CD11b+ were recruited to the colons of IL-10 deficient mice. Colon tissue was harvested from IL-10-/- mice at 3 and 6 months of age. The levels of Gr1+CD11b+ cells in mouse colons were determined via immunohistochemistry in colon sections and flow cytometry analysis of cells isolated from the colonic lamina propria. The percentage of Gr1+CD11b+ cells in the lamina propria determined by flow cytometry analysis following EDTA and collagenase digestion of colon tissues. The EDTA digestion helps remove the epithelial layer and the collagenase digestion is used to break down the lamina propria tissue into individual cells. Isolated cells were labeled with Gr1 and CD11b antibodies. The double positive analysis was performed on cells gated for granulocyte and monocyte size and density in the forward and side scatter plot. Non-specific binding was controlled using isotype antibodies. The percentage of Gr1+CD11b+ cells increases from 0.51 ± 0.40% in wild type mice to 3.0 ± 0.50% in IL-10-/- mice at 3 months of age (Figure 3-4). This difference is further increased at 6 months from 0.15 ± 0.01% to 4.4 ± 1.3%. Immunohistochemistry for Gr1+CD11b+ cells was performed on frozen sections of rolled wild type and IL-10-/- colons. Sections were labeled first with Gr1 antibody, 52 followed by a secondary plus blocking step and then with CD11b antibody. This protocol was designed with the help of Dr. Rui Zhang and Dr. Pina Colarusso, the head of the Live cell imaging facility at the University of Calgary. Non-specific binding was checked using combinations of isotype control antibody and the two primary antibodies. As shown in table 3-1, all combinations of isotype, Gr1 and CD11b antibodies were tried to check against non-specific binding. The isotypes only showed limited background staining (Figure 3-3A). Double positive cells (golden colour) were counted in each section histologically (Figure 3-3D). The area of the mucosa was determined histologically using hematoxylin and eosin labeled sections and data were expressed as double positive cells per unit area of mucosa to control for size variance between sections. The number of Gr1+CD11b+ cells per unit area of mucosa ( μ (2 ) determined by immunohistochemistry were significantly increased in the IL-10-/- mouse (21.3 ± 5.3, n=6) as compared to wild type mice (2.4 ± 2.0, n=5) at 3 months of age (Figure 3-3). The number of Gr1+CD11b+ cells/ce2 in IL-10 deficient mice were further increased to 95.6 ± 18.8 at 6 months of age (n=6). 53 Table 3-1 Combinations of primary and secondary antibodies used in Gr1+CD11b+ cell colon immunohistochemistry to check for non-specific binding 1st Antibody 2nd Antibody Gr-1 (FITC) CD11b (PE) ISO-FITC CD11b (PE) Gr-1 (FITC) ISO-PE ISO-FITC ISO-PE Both flow cytometry and immunohistochemistry analysis each have advantages and disadvantages. Lamina propria cell isolation and flow cytometry analysis can be performed within a day and it is easier to control for non-specific binding. Immunohistochemistry takes longer, especially in the case of double labeling, and checking against non-specific binding requires weeks if not months of work. Immunohistochemistry offers the advantage of allowing visualization of both the exact location of labeled cells in the mucosa as well as that of hematoxylin and eosin labeled sections for evaluation of inflammation and cancer. These data established the presence of Gr1+CD11b+ cells in the colons of IL-10 deficient mice. The levels of Gr1+CD11b+ cells increase with age and correlate with the development of cancer. 54 Gr1+CD11b+ Cells in Colon LP (%) 8 * 6 4 * 2 0 WT IL-10-/- 3 Months WT IL-10-/- 6 Months Figure 3-2 Lamina Propria Gr1+CD11b+ cell flow cytometry This figure illustrates the percentage of Gr1+CD11b+ cells in the colonic lamina propria of wild type and IL-10 deficient mice at 3 and 6 months of age. The percentage of Gr1+CD11b+ cells was determined by flow cytometry analysis of lamina propria cells isolated by EDTA and collagenase digestions. WT n = 2, IL10-/- n = 5, * indicates significant increase (P<0.05) from WT, One-way ANOVA followed by Bonferroni’s Post test 55 A B ______ 5μm ______ 5μm C D ______ 5μm ______ 5μm Figure 3-3 Representative immunohistochemistry for Gr1+CD11b+ cells in colon tissue sections This figure shows representative immuhistochemically stained serial colon sections from a 3-month-old IL-10 deficient mouse (400X). (A) Double staining with FITC and PE IgG2b isotypes shows only limited background staining. (B) Staining with anti-Gr1 (FITC) antibody shows positive cells in green. (C) Staining with anti-CD11b (PE) antibody shows positive cells in amber. (D) Double staining with anti-Gr1 (FITC) and anti-CD11b (PE) antibodies shows double positive cells in gold. 56 Gr1+CD11b+ Cell Count /Unit Area of Mucosa * 120 90 60 * 30 0 WT 3 months 6 months IL-10-/- Figure 3-4 Gr1+CD11b+ cell colon immunohistochemistry This figure illustrates the number of Gr1+CD11b+ cells per unit area of mucosa (μm2) in colonic sections from non-inflamed wild type (129SvEv) and 3 and 6 month old IL-10 deficient mice. The number of Gr1+CD11b+ cells was determined via immunohistochemistry of colon sections. significant increase (p<0.05) from WT, One-way ANOVA followed by Bonferroni’s Post test n ≥ 6, * indicates 57 3.2 Gr1+CD11b+ cell Function in vitro Having established the presence of Gr1+CD11b+ cells in the colons of IL-10 deficient mice, we next wanted to establish the presence of immunosuppressive pathways identified in literature to confirm these cells were suppressive. Arginase I and NOS2 enzyme activities have been identified as key MDSC immunosuppressive pathways and are used as identifiers of immunosuppressive activity (Dolcetti et al., 2008). We used three in vitro assays to test whether the Gr1+CD11b+ cells in the IL-10-/- model were immunosuppressive. Gr1+CD11b+ cells were isolated from the bone marrow of IL-10 deficient mice at 6 months of age via percoll density gradient centrifugation with purities greater than 80% as determined by flow cytometry analysis. Gr1+CD11b+ cells were isolated from the spleen following mechanical digestion via a proprietary magnetic bead based isolation kit with purities greater than 90% as determined by flow cytometry analysis. Arginase I activity was studied in cell lysates from 5 x 105 MDSC by measuring the catalysis of L-arginine by arginase I into L-ornithine and urea. Arginase I activity in bone marrow derived IL-10 deficient Gr1+CD11b+ cells (6.9 ± 2.1 x 10-3 μ3) was significantly increased from vehicle controls (8.7 ± 3.1 x 10-4 μ4) (Figure 3-5). Spleen derived IL-10-/- Gr1+CD11b+ cells exhibited further increased arginase I activity (0.20 ± 0.30 μ.). Liver tissue (0.26 ± 0.018 μ.) was used as a positive control. 58 Nitric oxide, an indicator for NOS2 acitivity, has a short half-life and is quickly broken down into nitrates and nitrites. Nitrate and nitrite production was measured by Griess reaction (Beda and Nedospasov, 2005). Nitrites were measured in the supernatant of 5 x 105 MDSC. Nitrates were first reduced to nitrites using Vanadium (III) Chloride. Nitrite production was greater in the bone marrow derived IL-10 deficient Gr1+CD11b+ cell supernatant (9.0 ± 0.34 μ.) than in vehicle controls (8.4 ± 1.0 x 10-2 μ2) (Figure 3-6). Spleen derived Gr1+CD11b+ cells exhibited further increased NOS2 activity (12.4 ± 0.57 μ.). Sodium nitrate (14.7 ± 0.67 μ.) was used as a positive control. So far, we had established the presence of immunosuppressive pathways in Gr1+CD11b+ cells. In order to directly assess their immunosuppressive capability we performed a T-cell proliferation suppression assay. Gr1+CD11b+ cell were coincubated with T-cells to determine whether they could prohibit proliferation. 5 x 105 T-cells were loaded with CFSE dye, proliferation was induced by CD3 and CD28 antibodies and detected by flow cytomery analysis for low CFSE level cells. Co-incubation was performed with 1 x 105 Gr1+CD11b+ cells. Stimulating CD4+ splenocytes with CD3 and CD28 antibodies increased proliferation detected by flow cytometry from 1.2 ± 0.20 % to 36.1 ± 2.1 %. Co-incubation of CD4+ cells with bone marrow derived Gr1+CD11b+ cells decreased proliferation to 24.5 ± 1.9% and was further reduced by spleen derived Gr1+CD11b+ cells (Figure 3-7). 59 The in vitro data reported here is generated from MDSC derived from the spleens and the bone marrow of mice. Ideally, we would have liked to the study the function of MDSC function in cells isolated from the colon. We were unable to accomplish this because of the difficulty involved in isolating a pure population from the colon. Percoll gradient centrifugation and the Miltenye magnetic beads isolation kit both produced cell purities below 60% when applied to cells isolated from colonic lamina propria. We were able to obtain usable numbers of MDSC from colonic lamina propria cell suspensions via fluorescence activated cell sorting (FACS) technique. FACS works in a similar manner to the flow cytometry analysis described earlier, but has the added advantage of selectively sorting out cells that match size, density and fluorescence criteria. The disadvantages of this technique are high cost and low yield. To isolate Gr1+CD11b+ cells via FACS, we first had to isolate lamina propria cells from the colon by EDTA and collagenase digestion and then label all of the isolated tens of millions of cells with Gr1 and CD11b antibodies. This process takes approximately 8 hour. After 5 hours of FACS sorting of these labeled cells, we only obtained approximately 1 x 105 Gr1+CD11b+ cells. This process did not generate enough cells to allow for extensive studies. We were unable to evaluate MDSC function in vitro in cells derived from the colon, but some alternative methodologies are available to explore this in future 60 experiments. A combination of flow cytometry, to identify Gr1+CD11b+ cells, and fluorescent in situ hybridization (FISH), to identify the expression of specific nucleic acid sequences, could be used to identify the messenger RNA expression of immunosuppressive proteins in lamina propria cell suspensions. FISH does not allow for the evaluation of protein expression and function, but it could be used in colon sections to identify whether Gr1+CD11b+ cells are expressing Arginase I and NOS2 mRNA sequences (Perske et al., 2010). FISH could also be potentially be used in combination with protein immunohistochemistry for Gr1+CD11b+ cells. This technique may be more challenging because of the potential for different antibodies and probes interfering with each other, but would have the added advantage of allowing for the comparison of MDSC function by location, for example in neoplastic versus inflammatory lesions. Intracellular NOS and Arginase production can also be evaluated by flow cytometry analysis (Obermajer et al., 2013). Combining these with Gr1+CD11b+ flow cytometry in colonic lamina propria cell suspensions would allow the evaluation of immunosuppressive protein function in colonic MDSC. This work confirmed that Gr1+CD11b+ cells in IL-10-/- mice were MDSC by establishing the presence of immunosuppressive immunosuppressive activity. 61 pathways and Urea Concentation (µM) 0.3 *$ 0.2 * 0.1 0.0 Positive Negative BM Spleen Figure 3-5 IL-10 deficient Gr1+CD11b+ cells express arginase I activity This figure illustrates arginase I enzymatic activity expressed as μ x urea produced in 5 x 105 Gr1+CD11b+ cell lysates derived from bone marrow (BM) and spleen of IL-10 deficient mice at 3 months of age. Lysate buffer and liver cell lysates were used as controls. n ≥ 4, * indicates significant increase (p<0.05) from negative controls. $ indicates significant increase (p<0.05) from BM, Oneway ANOVA followed by Bonferroni’s Post test 62 Nitrite Concentration (µM) 20 *$ 15 * 10 5 0 Positive Negative BM Spleen Figure 3-6 IL-10 deficient Gr1+CD11b+ cells have NOS2 Activity This figure illustrates NOS2 enzymatic activity represented as nitrite concentration, a stable product of nitric oxide production, in 5 x 105 Gr1+CD11b+ cells derived from IL-10 deficient bone marrow (BM) and spleen at 3 months of age. Carrier agent and sodium nitrate were used as controls. n ≥ 4, * indicates significant increase (P<0.05) from negative controls. $ indicates significant increase (P<0.05) from BM, One-way ANOVA followed by Bonferroni’s Post test 63 %Proliferating CD4+ Cells 50 40 * 30 *$ 20 10 0 T-Cells BM Spleen CD3 CD28 Stimulated T-Cells Figure 3-7 IL-10 deficient Gr1+CD11b+ cells suppress T-cell proliferation This figure illustrates the percentage of proliferating T-cells following stimulation with CD3 CD28 antibodies and co-incubation with Gr1+CD11b+ cells isolated from the bone marrow (BM) and spleen of IL-10 deficient mice at 3 months of age. T-cells were loaded with CFSE dye and proliferation was determined by flow cytometry analysis for low CFSE cells. Unstimulated T-cells were used as a control. n ≥ 4, * indicates significant decrease (P<0.05) from stimulated T-cells. $ indicates significant decrease (P<0.05) from BM groups, One-way ANOVA followed by Bonferroni’s Post test 64 3.3 MDSC function in vivo After confirming that the Gr1+CD11b+ cells in the IL-10 deficient mouse have immunosuppressive activity and therefore can be described as MDSC we next wanted to test whether these cells play a role in cancer development in this model in vivo. We tested this by depleting and transferring MDSC in vivo. 3.3.1 Depleting MDSC in vivo decreases cancer development in the IL-10 deficient mouse Low dose chemotherapeutic drugs such as gemcitabine and 5-fluorouracil (5-FU) have been reported to potently target MDSC both in vitro and in vivo (Suzuki et al., 2005; Vincent et al., 2010). These drugs are cytotoxic and were designed to target fast dividing cancer cells. At low doses though, they potently target MDSC and have been shown not to significantly affect cancer cells or immune cell types (Suzuki et al., 2005; Vincent et al., 2010). We used 5-fluorouracil and gemcitabine to deplete MDSC in IL-10 deficient mice and study cancer development in vivo. IL-10-/- mice at 6 weeks of age were given weekly injections of 60mg/Kg 5-FU (Sigma Aldrich, St. Louis, USA) or gemcitabine (Sigma Aldrich, St. Louis, USA) for 5 weeks. Mice were sacrificed at 12 weeks of age and MDSC counts, inflammation and cancer development were studied. Dysplasia scoring was done based on a previously published system that scores the levels of crypt reforming, epithelial dysplasia and submucosal 65 invasion (Zhang et al., 2007b). Both, 5-FU (9.48 ± 1.58 cells/5-2) and gemcitabine (14.26 ± 1.29 cells/em2), treatments reduced MDSC counts in colon as compared to saline control treated IL-10-/- mice (30.61 ± 4.24 cells/tm2), determined via immunohistochemistry (Figure 3-8). The decrease in MDSC counts was accompanied by a decrease in polyp and dysplasia scores in 5-FU and gemcitabine treated mice, indicating reduced cancer development. 5-FU treatment reduced polyp scores from 3.00 ± 0.00 to 0.33 ± 0.21 (Figure 3-9) and dysplasia scores from 4.40 ± 0.81 to 0.17 ± 0.17 (Figure 3-10). Gemcitabine treated IL-10-/- mice had polyp and dysplasia scores of 0.33 ± 0.21 and 2.00 ± 0.81, respectively. These data indicate that targeting MDSCs with low dose chemotherapeutics decreased neoplastic changes in IL-10 deficient mouse colons. Interestingly, depleting MDSCs did not affect inflammation development in the IL10 deficient mice. Inflammation in the colon was scored based on a previously published scoring system (Zhang et al., 2007b). Histologically, inflammation was scored based on the level of destruction of mucosal architecture, inflammatory cell infiltrate and muscle thickness. Treatment with 5-FU or gemcitabine did not affect the macroscopic (Figure 3-11) or histological (Figure 3-12) scores of inflammation. Mice treated with the drugs or with saline had similar levels of inflammatory mucosal damage and cell infiltration. 66 Gr1+CD11b+ Cell Count /Unit Area of Mucosa A 40 30 20 * 10 0 Control 5-FU Gr1+CD11b+ Cell Count /Unit Area of Mucosa B 40 30 * 20 10 0 Control Gem Figure 3-8 Low dose chemotherapeutics deplete MDSC in the IL-10 deficient mouse This figure illustrates the number of Gr1+CD11b+ cells in colon sections of IL-10 deficient mice following treatment with (A) 5-FU and (B) gemcitabine as determined via immunohistochemistry. Mice at 6 weeks of age were treated with a 60 mg/Kg dose weekly for 5 weeks and sacrificed at 12 weeks of age. Saline treated mice were used as controls. n ≥ 5, * indicates significant decrease (p<0.05) from control, Two-tailed Student’s t test 67 A Polyp Score 4 3 2 * 1 0 Control 5-FU B Polyp Score 4 3 2 * 1 0 Control Gem Figure 3-9 Lower MDSC counts are associated with reduced polyp scores in the IL-10 deficient mouse This figure illustrates the macroscopic polyp scores for the colons of IL-10 deficient mice following treatment with (A) 5-FU and (B) gemcitabine or saline controls. Mice at 6 weeks of age were treated with a 60 mg/Kg dose weekly for 5 weeks and sacrificed at 12 weeks of age. n ≥ 5, * indicates significant decrease (p<0.05) from control, Two-tailed Student’s t test 68 Histological Dysplasia Score 6 Histological Dysplasia Score A 6 4 2 0 * Control 5-FU B 4 * * 2 0 Control Gem Figure 3-10 Lower MDSC counts are associated with reduced dysplasia scores in the IL-10 deficient mouse This figure illustrates the histological dysplasia scores for colons of IL-10 deficient mice following treatment with (A) 5-FU and (B) gemcitabine or saline controls. Mice at 6 weeks of age were treated with a 60 mg/Kg dose weekly for 5 weeks and sacrificed at 12 weeks of age. n ≥ 5, * indicates significant decrease (p<0.05) from control, Two-tailed Student’s t test 69 Macroscopic Score of Inflammation A 4 3 2 1 0 Control 5-FU Macroscopic Score of Inflammation B 5 4 3 2 1 0 Control Gem Figure 3-11 Lower MDSC counts do not alter macroscopic inflammation in the IL-10 deficient mouse This figure illustrates the macroscopic inflammation scores for the colons of IL-10 deficient mice following treatments with (A) 5-FU and (B) gemcitabine or saline controls. Mice at 6 weeks of age were treated with a 60 mg/Kg dose weekly for 5 weeks and sacrificed at 12 weeks of age. n ≥ 5, Two-tailed Student’s t test 70 Histological Score of Inflammation A 6 4 2 0 Control 5-FU Histological Score of Inflammation B 6 4 2 0 Control Gem Figure 3-12 Lower MDSC counts do not affect histological inflammation in the IL-10 deficient mouse This figure illustrates the histological inflammation scores for the colons of IL-10 deficient mice following treatment with (A) 5-FU and (B) gemcitabine or saline controls. Mice at 6 weeks of age were treated with a 60 mg/Kg dose weekly for 5 weeks and sacrificed at 12 weeks of age. n ≥ 5, Two-tailed Student’s t test 71 3.3.2 Adoptive transfer of MDSC in vivo increases cancer development in the IL-10 deficient mouse To test whether increasing MDSC numbers in vivo could alter cancer development in the IL-10 deficient mouse mice were given tail vein injections of bone marrow derived MDSC. IL-10-/- mice at 6 weeks of age were given weekly injections of 3 x 106 MDSC weekly for 5 weeks and sacrificed at 12 weeks of age. To determine whether MDSC injected via tail vein were migrating to the colon, we injected CFSE labeled MDSC into one mouse. The colon from this mouse was excised and frozen sections were visualized using a microscope. Figure 313A shows a CFSE labeled MDSC in the colon of an IL-10 deficient mouse proving that adoptively transferred MDSC can migrate to the colon. The adoptive transfer of MDSC did not significantly affect macroscopic polyp scores (Figure 3-14A). This was because our polyp scoring scale goes ranges from 0 to 3 and saline treated IL-10 deficient mice at three months of age had a score close to the peak of the range at 2.75 ± 0.17. However, dysplasia scores were significantly higher in mice injected with MDSC (6.38 ± 0.50) as compared to saline controls (4.38 ± 0.50, Figure 3-14B). Importantly, 8 of 9 of the MDSC treated mice developed invasive cancers as compared to 1 of 8 in the saline treated controls. The adoptive transfer of MDSC did not affect macroscopic or histological scores of inflammation in the colons of IL-10 deficient mice (Figure 315). Despite an increase in cancer development, the adoptive transfer of MDSC 72 did not affect MDSC counts in the colon determined via immunohistochemistry (Figure 3-13B). We have already established that adoptively transferred MDSC are able to migrate to the colon, but this was an interesting finding that at first suggests that MDSC may not be reaching the colon in large numbers. However, it is important to look at this data in the context of life cycle of the cells being injected. The life cycle of MDSC in vivo is unknown. The next best approximation for MDSC life cycles can be made from that of the cells they mature into: neutrophils and monocytes. The life cycle of monocytes is also unknown. The lifecycle of inactivated neutrophils in circulation is 5 days and the life cycle of activated neutrophil is less than two days (Kumar, 2010). Since we only evaluated MDSC counts 7 days after the final injection of cells, it is likely that the injected MDSC would have lived through their life cycles by then. We chose to start treating animals at 6 weeks of age for our in vivo experiments because inflammation and dysplasia are not present in the colons of IL-10 deficient mice at this time point. MDSC adoptive transfer experiments previously have been performed by a number of other groups as an intervention in neoplastic and inflammatory conditions (Hegde et al., 2011; Ramachandran et al., 2013; Yin et al., 2010). All of these groups transferred 5 x 106 MDSC in their experiments. This number represents only a fraction of Gr1+CD11b+ cells present in spleen in animals suffering from inflammatory or cancerous conditions (Dolcetti et al., 2008; 73 Sawanobori et al., 2008). We chose a 3 x 106 cell dose because we are able to reliably able to isolate approximately 6 x 106 cells from the bone marrow of IL-10 deficient mice. This allowed us to inject two animals from MDSC derived from the bone marrow of one MDSC. Higher or lower doses of MDSC may have varying effects on disease development and should be followed up in future studies. We also conducted the MDSC adoptive transfer experiment with a single dose of cells given at 6 weeks of age. Animals were sacrificed at 12 weeks and the dysplasia scores showed statistically insignificant increase in treated animals. The fact that neutrophils have a relatively short life indicated that multiple injections may be more efficacious. We sacrificed the animals in our in vivo experiments at 12 weeks of age because by this time point the majority of IL-10 deficient mice have well established inflammation in the colon and approximately 12% of mice develop adenocarcinomas. This timing gave our treatments the opportunity to influence the development of inflammation and cancer. Trying in vivo interventions at different time frames could help further elucidate the role of MDSC in this model. A longer trial could show whether MDSC remain as critical to disease development in the longer term as they are in these first 6 weeks. Using low dose chemotherapeutics at a later time frame could help answer whether these drugs could be useful in patients at risk of developing colon cancer, for example patients with polyps. These data support a critical role for MDSC in the progression of dysplasia development in the IL-10 deficient model. 74 A Gr1+CD11b+ Cell Count /Unit Area of Mucosa B 40 30 20 10 0 Control MDSC Transfer Figure 3-13 Adoptive transfer of MDSC does not affect MDSC levels in the colons of IL-10 deficient mice (A) A CFSE stained bone marrow derived MDSC in a colon section of an IL-10 deficient mouse (400X). The colon was collected four hours after injection with labeled MDSC. (B) Gr1+CD11b+ cell counts in colon sections of IL-10 deficient mice following adoptive transfers of MDSC cells or saline controls as determined via immunohistochemistry. Bone marrow derived MDSC (3 x 106) were injected via tail-vein 5 times and animals were sacrificed at 6 weeks. n ≥ 5, (B) Two-tailed Student’s t test 75 A Polyp Score 4 3 2 1 0 Control MDSC Transfer Histological Dysplasia Score B 8 * 6 4 2 0 Control MDSC Transfer Figure 3-14 Adoptive transfer of MDSC increases neoplastic changes in the colons of IL-10 deficient mice This figure illustrates (A) polyp scores and (B) dysplasia scores for colons of IL10 deficient mice following adoptive transfers of MDSC cells or saline controls. Mice at 6 weeks of age were injected via tail vein with bone marrow derived MDSC (3 x 106) weekly 5 times and sacrificed at age of 12 weeks. n ≥ 5, * indicates significant increase (p<0.05) from Control, (B) Two-tailed Student’s t test 76 Macroscopic Score of Inflammation A 5 4 3 2 1 0 Control MDSC Transfer Control MDSC Transfer Histological Score of Inflammation B 5 4 3 2 1 0 Figure 3-15 Adoptive transfer of MDSC does not affect inflammation in IL10 deficient mouse colons (A) Macroscopic inflammation scores and (B) Histological inflammation scores assigned to colons sections of IL-10 deficient mice following adoptive transfers of MDSC cells or saline controls. Mice at 6 weeks of age were injected via tail vein with bone marrow derived MDSC (3 x 106) weekly 5 times and sacrificed at the age of 12 weeks. n ≥ 5, Two-tailed Student’s t test 77 The data from experiments with low dose chemotherapeutic drug targeting of MDSC support the role of MDSC in the development of cancer in this model. It is important to recognize that some or all of the effect observed in our model could be because of effect of the chemotherapeutic drugs on cancer cells rather than MDSC directly. Vincent et al. and others have done extensive work to show both in vitro and in vivo that MDSC are the only affected cell population at these low doses of 5-FU and gemcitabine (Le et al., 2009; Suzuki et al., 2005; Vincent et al., 2010). Vincent et al. also showed that the MDSC depletion effect occur within hours of the treatment. Other effects of treatment with low dose chemotherapeutic drugs, such as decreased cancer development, are observed at later time points suggesting that MDSC depletion could be affecting cancer development. This data in context together with the in vitro MDSC data and adoptive transfer data our results would suggest that it is likely that the decreased neoplastic changes are in part due to the depletion of MDSC by these drugs. Other potential mechanisms for therapeutically targeting MDSC in vivo include affecting their function using phosphodiesterase-5 inhibitors such as sildenafil, inducing their maturation with all-trans retinoic acid and targeting them with antiGr1 and CD11b antibodies (Kusmartsev et al., 2003; Serafini et al., 2006; Srivastava et al., 2012). We did not attempt the use of all-trans retinoic acid and Gr1 or Cd11b antibodies because these treatments would have unwanted effects 78 on other cell populations of myeloid origin. PDE-5 inhibitors such as sildenafil and vardenafil, which increase intracellular cyclic guanosine monophosphate availability, can decrease MDSC function by decreasing arginase I and NOS2 expresssion (Serafini et al., 2006). We attempted to target MDSC function in vivo with daily injections of 20 mg/Kg sildenafil (Sigma Aldrich, St. Louis, USA) in 6 week old IL-10-/- mice. Mice were sacrificed at 12 weeks of age and cancer development was studied. No significant effect on cancer development was observed with sildenafil treatment in the IL-10 deficient mice (data not shown). This could be because MDSC in this model are not affected by PDE-5 inhibition or perhaps the dose used was ineffective in the IL-10 deficient mouse. Further work is needed to make any conclusions from this experiment. For example, this could be further tested in vitro by treating MDSC with sildenafil before studying their function in the Arginase I or T-cell suppression assay. If MDSC function is decreased by sildenafil treatment in vitro then that would suggest that the PDE-5 pathway is important for MDSC function in this model. Further in vivo work could help tease out the right dose to target MDSC function in the IL-10 deficient mouse. 3.4 MDSC in the AOM/DSS Model of Colitis Associated Cancer Having already established the presence and protumorigenic role of MDSC in the IL-10 deficient mouse model of colitis-associated cancer, we were interested in determining whether such a link can be found in other models. A well established 79 model of inflammation associated colon cancer is the azoxymethane (Sigma Aldrich, St Louis, USA) and dextran sulfate sodium model (MP Biomedical, Santa Ana, USA) (Fukata et al., 2007; Kanneganti et al., 2011). Mice are treated with AOM, a carcinogen, followed by single or repeated rounds of DSS, which induces inflammation. In our model, we injected 2-month-old C57BL/6 mice with 10 mg/kg AOM. One week later, mice were given 2% DSS in drinking water for 7 days. Water only and AOM only treatments were used as controls. Mice were sacrificed 6 weeks after the end of DSS treatment and cancer development and Gr1+CD11b+ cell counts were determined. Dysplasia scores for mice treated with AOM/DSS (7.3 ± 0.31) were significantly higher than in C57BL/6 mice no treatment (0.75 ± 0.75, Figure 3-16A). Polyp scores were also significantly higher in AOM/DSS treated mice (data not shown). All eight AOM/DSS treated mice developed invasive adenocarcinomas. Interestingly, Gr1+CD11b+ counts were significantly increased in the colons of AOM/DSS treated mice (17.0 ± 2.8 cells/unit area of mucosa) as compared to controls (3.0 ± 1.0 cells/unit area of mucosa) as determined by immunohistochemistry in frozen colon sections (Figure 3-16B). Interestingly, MDSC counts are lower than IL-10 deficient mice of the same age (3 months) and much lower than 6-month-old IL-10 deficient mice. Only 12 % of IL-10 deficient mice develop invasive cancers by 3 months of age, which increases to 20 % at 6 months. Mice treated with AOM alone did not show significantly different results from water treatment. These data indicate that Gr1+CD11b+ cells are significantly increased in the AOM/DSS model. 80 A * Dysplasia Score 8 7 6 5 4 3 2 1 0 WT AOM only AOM/DSS B * Gr1+CD11b+ Cells /Unit area of Mucosa 20 15 10 5 0 WT AOM only AOM/DSS Figure 3-16 Dysplasia and Gr1+CD11b+ cell counts are increased in the AOM/DSS model of colitis-associated cancer This figure illustrates (A) dysplasia scores and (B) Gr1+CD11b+ cell counts in colons sections of C57BL/6 mice following treatment with AOM and DSS, AOM only and water only controls. 2 month old mice were injected with 10 mg/kg AOM. One week later mice were given 2% DSS in drinking water for 7 days. Mice were sacrificed 6 weeks after DSS treatment. n = 4-8, * indicates significant increase (p<0.05) from WT, One-way ANOVA followed by Bonferroni’s Post test 81 We have established the increased presence of Gr1+CD11+ cells in the colon. While we have been unable to isolate Gr1+CD11+ cells from the colon and confirm their immunosuppressive nature, the in vitro experiments on bone marrow and spleen derived MDSC demonstrate their immunosuppressive abilities in this model. The in vivo experiments together establish a protumorigenic role for MDSC in the IL-10 deficient model, where depleting them decreases neoplasia and their adoptive transfer increases neoplasia. Additionally, we also report the increased MDSC in an alternative model of inflammation associated colon cancer. These experiments support a role for Gr1+CD11b+, immunosuppressive, MDSC in the progression of inflammation associated cancer in the IL-10 deficient mouse. 82 Chapter 4: The role of MDSC in the Interleukin 10 Toll Like Receptor 4 double deficient mouse 83 Summary The following experiments study the role of MDSC in the IL-10 TLR4 double deficient mouse. First, the increased development of cancer in the IL-10 TLR4 double deficient mouse relative to the IL-10 deficient mouse is established. Second, the increased recruitment and correlation of MDSC with cancer development in this model is shown. Third, peripheral blood and colonic experiments show that MDSC are the only upregulated cell type in the IL-10 deficient mouse in the absence of TLR4. Finally, our data show that TLR4 expression does not affect MDSC function but that TLR4 in colon tissue modulates cancer development. Together these experiments further establish a protumorigenic role of MDSC in the IL-10 deficient mouse model and indicate that TLR4 expression in the colon modulates MDSC recruitment and cancer development. 4.1 Cancer incidence and MDSC levels in the IL-10 TLR4 double deficient mouse Chapter 3 shows the protumorigenic role for MDSC in vivo and in vitro in the IL10 deficient mouse. TLR4 plays an important role in the pathogenesis of both inflammation and cancer. Our laboratory and others have reported that TLR4 is a modulator of inflammation-associated cancer in the colon (Fukata et al., 2007; Zhang et al., 2007a). Others have shown the importance of the NFhe pathway, which is activated by TLR4, in the progression of colon cancer (Greten et al., 84 2004). Our laboratory has developed a novel IL-10 and TLR4 double deficient mouse. The first step in understanding the potential role of MDSC in this model was to establish the inflammation and cancer levels in this model. 4.1.1 Inflammation and Cancer levels in the IL-10 TLR4 Double Deficient mouse The IL-10-/- TLR4-/- mouse develops cancer at a markedly increased rate and incidence as compared to a TLR competent IL-10 deficient mouse. Colon cancer was evaluated with the help of a blinded pathologist. Adenocarcinomas were defined as invasive cancers where high dysplastic crypts have invaded into the submucosal layer. This work was performed my Dr. Rui Zhang, a post-doctoral fellow under Dr. McCafferty’s supervision. At three months of age 50% of IL-10-/- TLR4-/- mice develop adenocarcinomas compared to 12% of IL-10-/- mice (Figure 4-1A). Histological inflammation scores were evaluated based on factors such as leukocyte infiltration, muscle thickness and crypt architecture. Histological inflammation scores showed a small but significant increase in IL-10-/- TLR4-/- mice (6.8 ± 0.2) compared to IL-10-/- mice (6.0 ± 0.2, Figure 4-1B). Wild type 129SvEv mice had a baseline inflammation score of 1.4 ± 0.3. 85 Incidence of Adenocarcinoma (%) A 100 75 50 25 0 WT IL-10-/- IL-10-/-/TLR4-/- Histological Score of Inflammation B Figure 4-1 TLR4 * 7 6 *# 5 4 3 2 1 0 WT deficiency IL-10-/- increases IL-10-/-/TLR4-/- cancer development and inflammation in the IL-10 deficient mouse This figure illustrates (A) the incidence of adenocarcinomas and (B) histological inflammation scores in the colons of three month old mice on the 129SvEv background. n ≥ 20, * indicates significant increase (p<0.05) from WT, # indicates significant increase (p<0.05) from IL-10-/-, (B) One-way ANOVA followed by Bonferroni’s Post test 86 These data demonstrate that there is comparable, if not slightly higher, level of inflammation present in the absence of TLR4 in the IL-10 deficient mouse. This is not surprising as other have shown a homeostatic role for TLR4 in models of gut (Fukata et al., 2005; Gonzalez-Navajas et al., 2010). These data also establish a striking increase in the development of adenocarcinoma in the absence of TLR4 in this model. The reason for this increase is unclear. However, since our earlier data in the IL-10 deficient mouse have shown a protumorigenic role for MDSC, we investigated whether MDSC numbers and/or function were altered in the absence of TLR4. Having established a key role for MDSC in cancer development in the IL-10 model, we next wanted to see how MDSC recruitment was altered in this model in the absence of TLR4. 4.1.2 MDSC recruitment in the IL-10 TLR4 double deficient mouse A marked increase in MDSC levels is observed in the bone marrow, spleen and site of lesion in various animal and human inflammatory and cancerous conditions (Ostrand-Rosenberg and Sinha, 2009). We took a systematic approach in quantifying levels of MDSC in the spleen, bone marrow and the colon, the site of inflammation and cancer, in the presence and absence of TLR4. MDSC levels in the bone marrow were determined by flow cytometry analysis for Gr1+CD11b+ cells in cell suspensions obtained by flushing femur and tibia bones of 3-month-old mice. MDSC levels were significantly increased in the IL-10-/mouse bone marrow (55.4 ± 5.4 %) compared to wild type animals (34.8 ± 3.5 %). 87 MDSC levels were comparable in the bone marrow of IL-10-/- TLR4-/- mice (60.3 ± 3.4 %) and IL-10-/- counterparts (Figure 4-2A). MDSC levels in the spleen were determined by flow cytometry analysis for Gr1+CD11b+ cells in cell suspensions obtained by mechanical digestion of mouse spleens. MDSC levels were significantly increased in the IL-10-/- mouse spleens (9.4 ± 1.7 %) compared to wild type animals (1.6 ± 0.3 %). MDSC levels in the spleens of IL-10-/- TLR4-/mice (8.6 ± 2.4) were comparable to IL-10-/- mice (Figure 4-2B). This work was performed by Dr. Rui Zhang, a postdoctoral fellow under the supervision of Dr. McCafferty. We next studied MDSC recruitment to the colon via colon section immunohistochemistry and lamina propria flow cytometry analysis for Gr1+CD11b+ cells. We have already established that MDSC levels were significantly increased in the colons of IL-10-/- mice compared to wild type animals at both three and six months of age (Page 47). Lamina propria cell flow cytometry showed a similar increase in the percentage of Gr1+CD11b+ cells in the IL-10-/- TLR4-/- from IL-10-/- counterparts at both 3 (5.9 ± 0.5 %) and 6 (8.8 ± 1.2 %) months of age (Figure 4-3A). Gr1+CD11b+ cell immunohistochemistry showed significantly higher cell numbers in the IL-10 TLR4 double deficient colon sections as compared to IL-10 deficient sections at both 3 (64.6 number>84</rec2) and 6 (145.9 ± 17.8 cells/μm2) months of age (Figure 4-3B). 88 Gr1+CD11b+ Cells in Bone Marrow (%) A * 70 60 * 50 40 30 20 10 0 WT IL-10-/- IL-10-/-/TLR4-/- * * IL-10-/- IL-10-/-/TLR4-/- Gr1+CD11b+ Cells in Spleen (%) B 12 10 8 6 4 2 0 WT Figure 4-2 TLR4 deficiency does not affect MDSC levels in the bone marrow and spleen in the IL-10 deficient mouse This figure illustrates MDSC levels in (A) the bone marrow and (B) spleens of three month old mice. MDSC levels were determined via flow cytometry for Gr1+CD11b+ cells. This work was performed by Dr. Rui Zhang. n ≥ 10, * indicates significant increase (p<0.05) from WT, One-way ANOVA followed by Bonferroni’s Post test 89 A Gr1+CD11b+ Cells in Colon LP (%) 15 *# 10 *# 5 WT IL-10-/IL-10-/-/TLR4-/- * * 0 6 Months Old 3 Months Old Gr1+CD11b+ Cell Count /Unit Area of Mucosa B # 175 150 WT IL-10-/IL-10-/-/TLR4-/- 125 100 *# 75 50 25 * 0 3 Months Old 6 Months Old Figure 4-3 TLR4 deficiency increases MDSC recruitment to the colon in the IL-10 deficient mouse This figure illustrates MDSC levels in colons of three and six month old mice. MDSC levels were determined via (A) flow cytometry for Gr1+CD11b+ cells in colonic lamina propria cells and (B) Gr1+CD11b+ cell immunohistochemistry in colon sections. n = 4-6, * indicates significant increase (p<0.05) from WT, # indicates significant increase (p<0.05) from IL-10-/-, One-way ANOVA followed by Bonferroni’s Post test 90 These data show that MDSC levels are comparable in the spleens and bone marrow of IL-10-/- and IL-10-/- TLR4-/- mice. Interestingly, MDSC levels are significantly increased in the colon, the site of inflammation and cancer, in the absence of TLR4 in the IL-10 deficient mouse. This indicates a correlation of MDSC presence with cancer development. 4.1.3 MDSC and Cancer correlation Our data so far suggest MDSC are recruited to the colon in increased numbers in the absence of TLR4 and that these cells contribute to the development of dysplasia in the colon. Next we examined the location of MDSC within the colon tissue. Using immuohistochemically stained and H&E labeled serial sections we examined the general distribution of MDSC in the colon. H&E sections allowed us to identify inflamed areas and areas with adenocarcinoma. Using Gr1+CD11b+ cell immunohistochemistry we were able to count MDSC cells in inflamed and tumor areas within the same colon section. MDSC counts per unit area of mucosa (μm2) were greatly increased in areas of adenocarcinoma (1177 ± 230 cells/μm2) than in inflamed areas (122 ± 54 cells/μm2) of approximately the same size in four IL-10-/- and IL-10-/- TLR4-/- mouse colon sections (Figure 4-4A). We determined the correlation of cancer development with inflammation in this model by plotting MDSC counts against histological scores of inflammation in colons from WT, IL-10-/-, and IL-10-/- TLR4-/- mice. A positive correlation was observed between MDSC levels and inflammation in the colon (r2=0.67, n=15, Figure 4-4B). 91 Next, we studied the correlation between MDSC levels and cancer development in this model. When we plotted dysplasia scores against whole tissue MDSC counts from WT, IL-10-/-, and IL-10-/- TLR4-/- mice a stronger positive correlation was observed (r2=0.82, n=28, Figure 4-4C). Our data illustrate that MDSC congregate in areas of dysplasia over inflammation. These data suggest that both inflammation and MDSC recruitment are drivers of cancer development in the IL-10 deficient mouse model. These data establish a strong correlation between MDSC and cancer development in the IL-10-/- and IL10-/- TLR4-/- mouse. The increased cancer development in the IL-10-/- mouse in the absence of TLR4 could in part by explained by an increased recruitment of MDSC. 92 Histological Inflammation Score B Gr1+CD11b+ Cell Count /Unit Area of Mucosa A * 1500 1000 500 0 Inflamed Areas 8 6 4 2 0 0 50 100 150 MDSC Count r2 = 0.67 Cancerous Areas C Dysplasia Score 9 6 3 0 0 50 100 150 200 250 MDSC Count r2 = 0.82 Figure 4-4 MDSC recruitment correlates with neoplastic changes in the IL10 deficient mouse model (A) Compares the level of MDSC in inflamed and cancerous areas (adenocarcinoma) in the colons of IL-10 deficient and IL-10 TLR4 double deficient mice (n=8). This figure also illustrates the correlation between (B) histological inflammation scores or (C) dysplasia scores and MDSC levels in the colons of wild type, IL-10 deficient and IL-10 TLR4 double deficient mice at 3 and 6 months of age (n=30). The slopes of the correlation were significantly changed from 0 (p<0.05). MDSC levels were determined via Gr1+CD11b+ cell 93 immunohistochemistry in colon sections. * indicates significant increase (p<0.05) from Inflamed Areas, (A) Two-tailed Student’s t test, (B) and (C) Linear regression analysis 94 4.2 Other leukocyte population levels in the absence of TLR4 in the IL-10 deficient mouse Having established an increased recruitment of MDSC in the colons of IL-10 deficient mice in the absence of TLR4 we next wanted to determine whether other white blood cell populations were affected. We first performed a leukocyte differential count on peripheral blood from IL-10-/- and IL-10-/- TLR4-/- mice. Total white blood cell counts in peripheral blood were comparable in IL-10-/- (5.5 ± 0.42 x 106 cells/ml) and IL-10-/- TLR4-/- (5.7 ± 0.81 x 106 cells/ml) as determined by hemocytometer counting (n=4). The differential count illustrated that the percentage of monocytes, granulocytes and lymphocytes were comparable in peripheral blood of IL-10-/- and IL-10-/- TLR4-/- mice (Figure 4-5). Next we used flow cytometry to identify specific leukocyte populations in colonic lamina propria following EDTA and collagenase digestions of colon tissue from 3month-old IL-10-/- and IL-10-/- TLR4-/- mice. The percentage of T-cells, B-cells, macrophage and dendritic cells was determined by flow cytometry analysis for CD4+, CD19+, F4/80+ and CD11c+ cells, respectively. The percentage of B-cells, macrophage and dendritic cells was comparable in IL-10-/- and IL-10-/- TLR4-/colons (Figure 4-6). The percentage of MDSC was also measured by flow cytometry analysis for Gr1+CD11b+ cells to confirm previous results. As in the previous work, MDSC levels were significantly increased in the IL-10-/- TLR4-/mouse colons (10.0 ± 0.99 %) as compared to IL-10-/- (3.9 ± 0.44 %). 95 Interestingly, the percentage of T-cells was decreased in IL-10 TLR4 double deficient mice (17.8 ± 0.41 %) from IL-10 deficient mice (30.3 ± 1.7 %). This decrease in T-cells might be expected as we have shown MDSC suppress the proliferation of T-cells in this model. These data show that white blood cell levels in peripheral blood are comparable in IL-10-/- and IL-10-/- TLR4-/- mice. We also demonstrate that MDSC are the only upregulated cell type in the colon in the absence of TLR4 in the IL-10 deficient mouse model. 96 yt e M on oc yt Ly m ph oc cy lo nu ra G e IL-10-/IL-10-/-/TLR4-/- te %Leukocyte 55 50 45 40 35 30 25 20 15 10 5 0 Figure 4-5 TLR4 deficiency does not affect peripheral blood leukocyte levels in the IL-10 deficient mouse This figure illustrates differential leukocyte counts in the peripheral blood of threemonth-old mice on the 129SvEv background. Peripheral blood was collected via cardiac puncture. Leukocyte differential counts were determined histologically using H&E labeled blood smears. n = 3, Two-way ANOVA followed by Bonferroni’s Post test 97 35 * %Leukocyte 30 IL-10-/IL-10-/-/TLR4-/- 25 20 15 10 * 5 e ag C D ls M ac ro B ph -C el ls el C T- M D SC 0 Figure 4-6 T-cells are decreased in the colon in the absence of TLR4 in the IL-10 deficient mouse This figure illustrates percentage of various leukocyte populations in the colons of three-month-old mice on the 129SvEv background. Leukocyte levels were determined via flow cytometry of colonic lamina propria cells isolated following EDTA and collagenase digestions. n = 3, * indicates significant change (p<0.05) from IL-10-/-, Two-way ANOVA followed by Bonferroni’s Post test 98 Our data so far have illustrated an enhanced recruitment of MDSC to the colon in TLR4 deficient mice. We next asked the question whether their function is altered in the absence of TLR4 in the IL-10 deficient model. We investigated this using the in vitro tests of L-arginine metabolism and suppression of T-cell proliferation. 4.3 MDSC function in the IL-10 TLR4 double deficient mouse MDSC are levels are highest in the bone marrow, spleen and at the sites of inflammation or cancer (Gabrilovich and Nagaraj, 2009). We have been unable to isolate purified MDSC from the colon, the site of cancer in IL-10 deficient and IL10 TLR4 double deficient mice. We have been able to isolate pure MDSC populations from the bone marrow (>80 %) and spleen (>92 %). In vitro function in bone marrow and spleen derived MDSC was measured by arginase I assay, griess reaction for nitrate production and suppression of T-cell proliferation assay. MDSC derived from the bone marrow and spleens of IL-10-/- and IL-10-/-TLR4-/had comparable levels of arginase I activity, nitric oxide production and T-cell proliferation suppression ability (Figure 4-7, spleen data not shown). These data indicate that TLR4 does not affect the function of MDSC in this model, only their recruitment. Having demonstrated that TLR4 competence does not affect MDSC function, we were interested in determining what proportion of MDSC that express TLR4 in the IL-10-/- mouse. MDSC were treated with lipopolysaccharide (LPS) in order to 99 induce TLR4 expression. We were not able to detect TLR4 without stimulation with LPS, a gram-negative bacterial cell wall component. LPS has previously been used to induce TLR4 expression in MDSC and other cell types (Bunt et al., 2009). We performed flow cytometry analysis for TLR4+ cells in MDSC from 3month-old IL-10-/- mice. MDSC purified from the bone marrow were triple labeled with Gr1, CD11b and TLR4 antibodies. The percentage of TLR4+ cells in Gr1+CD11b+ cell poplulations from IL-10-/- TLR4-/- mice, which were used as controls, was 0.16 ± 0.017 % (Figure 4-8). The percentage of TLR4+ cells in Gr1+CD11b+ cell populations from IL-10-/- mice was 5.1 ± 0.30 %. The percentage of TLR4+ MDSC in IL-10 deficient mice, while much higher than the negative controls, suggests that only a small proportion of MDSC express detectable levels of TLR4 in this models. Since the majority of MDSC do not express TLR4 and as our findings suggest, it is unlikely to affect their function. 100 Po si tiv e N eg at iv e IL -1 IL 0 -/-1 0 -/B M TL R 4 -/IL B -1 IL M 0 -/-1 0 -/Sp TL le en R 4- / Sp le en Nitrite Concentration (µM) Po si tiv e N eg at iv e IL IL -1 -1 00//B M TL R 4/IL B -1 M IL 0 -/-1 0 -/Sp TL le en R 4 -/Sp le en Urea Concentation (µM) A B 0.3 0.2 * * 0.1 12 8 * * 0.0 16 * * * * 4 0 C 101 40 # 30 20 # # # 10 el R IL 0 -/TL IL -1 IL -1 0 -/B M C T- M D 4 -/SC B -1 M IL 0 -/-1 M D 0 -/Sp SC TL le en R 4 -/M D Sp SC le en M D SC 0 ls %Proliferating CD4+ Cells 50 CD3 CD28 Stimulated T-Cells Figure 4-7 TLR4 does not affect MDSC function in the IL-10 deficient mouse (A) Illustrates arginase I enzymatic activity expressed as μI urea produced in 5 x 105 MDSC lysates. Lysate buffer and liver cell lysates were used as controls. (B) Illustrates NOS2 enzymatic activity represented as nitrite concentration, a stable product of nitric oxide production, in 5 x 105 MDSC cells. Carrier agent and sodium nitrate were used as controls. (C) Illustrates the percentage of proliferating T-cells following stimulation with CD3 CD28 antibodies and coincubation with MDSC cells. MDSC were isolated from the bone marrow of 3month-old mice via percoll density gradient centrifugation and from the spleen via Miltenyi magnetic beads isolation. n≥4, * indicates significant increase (p<0.05) from negative controls, # indicates significant decrease (p<0.05) CD3 CD28 stimulated T-cells, One-way ANOVA followed by Bonferroni’s Post test 102 %TLR4+ MDSC 6 4 2 0 IL-10-/-/TLR-4-/- IL-10-/- Figure 4-8 TLR4 expression in MDSC in the IL-10 deficient mouse This figure illustrates percentage of TLR4+ MDSC (Gr1+CD11b+TLR4+) cells in the bone marrow of 3-month-old mice. TLR4+ MDSC levels were determined via flow cytometry of bone marrow cells. n=3-4 103 4.4 TLR4 competence and cancer development Our data on MDSC in the IL-10 TLR4 double deficient mouse indicate that functionally they are comparable to cells from the IL-10 deficient mouse, but recruited to the colon in greater numbers. This suggests that the TLR4 deficiency likely induces enhanced MDSC recruitment by affecting the tissue microenvironment. To test this directly, we transferred the bone marrow, the source of MDSC, between TLR4 competent and TLR4 deficient IL-10-/- mice. If TLR4 incompetent bone marrow were to increase cancer development in the IL10 deficient colon, it would indicate that TLR4 competency in the MDSC plays a role in cancer development. Bone marrow transplant experiments were performed in IL-10-/- and IL-10-/- TLR4/- mice. Mice at two months of age were irradiated to eliminate the recipient marrow and injected with bone marrow cell suspensions from the opposite genotype via tail vein (n=5). Mice were given broad-spectrum antibiotics in drinking water for 2 weeks following the transplant to avoid infections in an immunocompromised state and sacrificed at 6 months of age. Bone marrow transplants between the same genotype (eg: IL-10-/- to IL-10-/-) were used as controls. We assessed macroscopic and histological parameters of inflammation. Macroscopic scores of colon inflammation were comparable in all of the groups (Figure 4-9A). Histological scores of inflammation were also comparable in all control and treatment groups (Figure 4-9B). Interestingly, the histological scores 104 of inflammation were relatively high in all groups for their age. Next, we examined parameters of cancer development. Macroscopic polyp scores, an indicator of hyperplasia, were unchanged among all groups (Figure 4-9C). Surprisingly, no dysplastic changes were observed in any of the treatment groups. Therefore, no cancer development was observed even at this advanced age. This was most likely an affect of the transplant experiment, the antibiotic treatment or a combination of the two. Irradiation has a greater effect on fast dividing cell types such as immune cells that play a key role in inflammation-associated cancer. Radiation is also used as a treatment for various cancers because it causes DNA damage and cell death in fast dividing cancer cells. The radiation treatment in this experiment could have killed off the cancer cells in the colon, leading to decreased cancer development. In all groups, antibiotic treatment post irradiation and BMT could affect the microflora in the gut and affect cancer development. For example, Antibiotic treatment has been shown to decrease inflammation and cancer development in the gut by decreasing the microflora (Hale and Greer, 2012). The nature of the techniques used to conduct this experiment likely interfered with the outcome being tested. Repeat trials should be performed with a lower dose of antibiotics and the animal sacrifice should be delayed until 8 or 10 months of age. Alternative controls, such as transferring wild type bone marrow into IL-10 deficient and IL-10 TLR4 double deficient mice, could be added to future experiments to study the effects of normal bone marrow interaction with the inflamed colon tissue microenvironments. 105 Having already established increased cancer development in the IL-10-/- TLR4-/mouse colons, we were interested in studying how TLR4 expression is modulated in IL-10 deficient colons. The modulation of TLR4 expression within IL-10 colons could further indicate whether TLR4 plays a role in the tissue microenvironment in the pathogenesis of cancer in this model. This work was performed in collaboration with a project student under the supervision of Dr. McCafferty, Mattias Svensson. TLR4 mRNA expression was studied colons of IL10 deficient mice. Polyps and cancer are almost always found in the proximal colon, close to the ileo-ceacal junction in this model (Zhang et al., 2007b). Wild type colon TLR4 levels were set as standards and TLR4 mRNA was quantified in IL-10 deficient distal colons and polyps collected from the proximal colon (Figure 4-10). Interestingly, TLR4 expression was significantly lower in the IL-10-/- polyp region (0.38 ± 0.098) as compared to the distal colon (0.86 ± 0.22). We also attempted to analyze TLR4 protein in the IL-10 deficient colon via western blot analysis with two different TLR4 antibodies. These attempts were unsuccessful as we were unable to visualize any bands, including the positive controls. These data are interesting because they demonstrate a differential expression of TLR4 within the colon tissue around sites of dysplasia. They suggest TLR4 expression as a potential marker for the transformation from inflammation to cancer. It also indicates that the TLR4 deficiency is important in the colon tissue for modulating protumorigenic changes. 106 A B 6 Histological Score of Inflammation 4 2 8 6 4 2 0 4 4- R LR /-T 0-1 IL to in 0-1 0IL -1 /- LR /-T /-T 0-1 TL -/10 L- -I 4- LR /- 4- 0-1 IL IL IL /- in in to to IL IL -1 -1 0- 0- /- /- 4LR /-T 0-1 IL to in /0- -1 IL /- /- 4 R TL -/10 L- -I /4LR IL -1 IL 0- -1 0- /-T /-T IL LR -1 4- 0- /- /- in in to to IL IL -1 -1 0- 0- /- /- 0 /- Macroscopic Score of Inflammation 10 Macroscopic Polyp Score C 4 3 2 1 /- 4 4- R LR TL /-T -/- 0- 10 -1 L- IL -I to /- in 4IL -1 0- /- LR 0-1 IL IL -1 0- /-T /-T IL LR -1 4- 0- /- /- in in to to IL IL -1 -1 0- 0- /- /- 0 Figure 4-9 Bone marrow transplants between IL-10 TLR4 double deficient and IL-10 deficient mice do not affect inflammation and cancer development This figure illustrates (A) macroscopic inflammation scores (B) histological inflammation scores and (C) macroscopic polyp scores for mouse colons from bone marrow transplant recipient mice. Mice at 2 months of age were irradiated and injected with donor marrow. Mice were treated with neomycin for 2 weeks 107 following the irradiation and sacrificed at 6 months of age. IL-10 deficient mice were given IL-10 TLR4 double deficient bone marrow and vise versa. Bone marrow transplants between the same genotypes were used as controls. n = 5, One-way ANOVA followed by Bonferroni’s Post test 108 Realtive quantification of TLR4 1.2 0.8 * 0.4 0.0 WT Distal Colon Polyp Area IL10-/- Figure 4-10 TLR4 mRNA expression is decreased in the polyps of IL-10 deficient mice This figure illustrates the relative quantification of TLR4 message expression in the colon regions of three-month-old mice. TLR4 mRNA expression was determined by RT-PCR using wild types as standards. n = 7, * indicates significant change (p<0.05) from Distal Colon, Two-tailed Student’s t test 109 The data shown in this chapter show that the absence of TLR4 increases cancer development in the IL-10 deficient mouse model. MDSC are the only upregulated leukocytes in the absence of TLR4 and their levels highly correlate with neoplastic changes in this model. TLR4 deficiency does not affect MDSC function. The majority of MDSC do not express TLR4 and TLR4 expression is decreased in IL-10-/- polyps. These data suggest that increased MDSC recruitment and cancer development in the colon in the absence of TLR4 is likely a result of changes in the colon tissue. 110 Chapter 5: MDSC chemotaxis in vitro and in vivo 111 Summary The following experiments elicit a potential mechanism for the increased recruitment of MDSC in the absence of TLR4 in the IL-10 deficient mouse. In vitro and in vivo chemotaxis assays were used to identify potential MDSC chemoattractants in the IL-10 deficient mouse. We also determined that TLR4 does not modulate MDSC chemotaxis in this model. Finally, we show that MDSC chemoattractant protein levels are significantly increased in the absence of TLR4 in the IL-10 deficient mouse. Our findings suggest that TLR4 modulates MDSC recruitment in the IL-10 deficient mouse by increasing the bioavailability of MDSC chemoattractants. Our data so far show a key role for MDSC and TLR4 in cancer development in the IL-10 deficient mouse model of inflammation-associated cancer. TLR4 deficiency increases MDSC recruitment and cancer development in the colon. TLR4 message is decreased in hyperplastic areas in IL-10 deficient mouse colons. These data together suggest that the absence of TLR4 in the tissue microenvironment causes a change that leads to increased MDSC recruitment and cancer development. In this chapter, we investigated the potential mechanisms involved in the increased recruitment of MDSC observed in the IL10-/- mice in the absence of TLR4. Our first step was to identify factors that recruit MDSC in this model. Next, we would examine whether these factors are affected by TLR4 deficiency in IL-10-/- mouse colons. 112 A number of agents, including IL-6, prostaglandin E2 and MCP-1, have been proposed to be involved in MDSC recruitment (Cheng et al., 2011; Huang et al., 2007; Obermajer et al., 2011a; Sinha et al., 2007b; Tu et al., 2008). These studies were largely conducted in vivo, using pharamacological antagonism or genetic knockouts. Huang et al. used antibodies to decrease MCP-1 bioavailability and genetics knockouts for its receptor, CCR2, to show that MCP-1 is crucial for MDSC recruitment to the cancer site (Huang et al., 2007). Sinha et al. have shown the importance of prostaglandin E2 in MDSC recruitment by knocking out its receptor, EP2, and pharmacological blockade of the cyclooxygenase 2 pathway that generates prostaglandin E2 in mouse model of breast cancer (Sinha et al., 2007b). While these studies suggest that these factors play a role in MDSC recruitment, it is unlikely that all are direct chemoattractants. We focused on identifying potential direct chemoattractants of MDSC because their bioavailability in the colon would likely determine MDSC recruitment levels. We first investigated MDSC chemotaxis in vitro, which to our knowledge we are the first to do. In order to examine MDSC recruitment in vitro we first had to establish an appropriate system. We examined MDSC migration in three chemotaxis assays described in literature: underagarose assay, ibidi chamber assay and transwell chamber assay. 113 5.1 Underagarose and Ibidi Chamber Chemotaxis Assays Our laboratory has previously used the underagarose assay to study neutrophil chemotaxis in vitro (Khajah et al., 2013). All in vitro chemotaxis assays rely on cells migrating along a gradient of chemoattractant concentration. This assay is performed by tracking the movement of cells under a layer of agarose from the well containing cells towards the chemoattractant (Heit and Kubes, 2003). The net number of migrating cells is determined by counting the cells moving towards the chemoattractant well minus the cells moving away in the opposite direction. Vehicle was used as control. We used neutrophil chemotaxis in this assay with a previously identified chemoattractant, fMLP, for comparison with MDSC. Neutrophils were isolated from the bone marrow of wild type mice via percoll gradient density centrifugation. Neutrophil migration significantly increased towards fMLP as compared to PBS controls (Figure 5-1A). Migration levels were comparable to previously reported levels (Khajah et al., 2013). MDSC were isolated from IL-10-/- mouse bone marrow via percoll density gradient centrifugation. The purity of MDSC derived in this manner as confirmed by flow cytometry analysis for Gr1+CD11b+ cells was greater than 80%. MDSC migration towards neutrophil chemoattractants including MIP-2 (0.25 μ( - 2 μ ), KC (0.25 μ( - 2 μ-) and fMLP (0.5 - 2 μ2) was unchaged from control levels. Table 5-1, shows the number of migrating cells towards one representative concentration from each of these chemoattractants. Interestingly, while 114 chemotaxis levels did not significantly increase with these chemoattractants, the trend and variability was higher than control levels. This could indicate that these chemoattractants are either inducing the random movement of cells or attracting the small proportion of impure cells. Table 5-1 The number of migrating MDSC towards chemoattractants in the underagarose chemotaxis assay (n = 3-6) Chemoattractant Net Migrating Cells Control 0.94 ± 1.1 MIP-2 (1 μI) 2.1 ± 1.6 KC (0.5 μ.) 1.4 ± 1.1 fMLP (1 μ() 3.2 ± 2.7 After some preliminary data showed promising results, we generated a dose response curve for MDSC chemotaxis towards MCP-1 (0.1 - 8 μ t) in the underagarose assay (Figure 5-1B). MDSC chemotaxis was significantly increased towards 4 μw MCP-1 (16.9 ± 2.1 net migrating cells) as compared to controls (-0.33 ± 0.67 net migrating cells). An issue with this assay is that it does not allow us to specifically extract the migrating cells and confirm that they are Gr1 and CD11b positive. Therefore, while the there was a significant increase in 115 the number of migrating cells towards MCP-1, we could not confirm whether these cells were Gr1+CD11b+ or from the approximately 20% impure cell population. We next tried the Ibidi chamber chemotaxis assay (Heit et al., 2008; Zantl and Horn, 2011). In this assay, cells are placed in a small channel (0.4x17x4.8 mm) between two larger reservoirs (80 μe). Chemoattractant conditions are placed in one of the large reservoirs and the other is filled with vehicle, leading to a concentration gradient across the channel. The movement of cells in the channel was tracked closely over 16 hours using a camera attached to an inverted microscope. Migrating cells were considered chemotactic if their net movement was greater than 2 mm towards the chemoattractant reservoir. MDSC were isolated from IL-10-/- mice via percoll density gradient centrifugation. The advantage of ibidi chambers over underagarose is that cells in the reservoir can be easily fixed and labeled with antibodies. We used the same protocol as Gr1+CD11b+ immunohistochemistry to check which chemotactic cells were Gr1+CD11b+ (Figure 5-2A). Using the ibidi chemotaxis, we found that MDSC migration towards 1 μ MCP-1 (18.3 ± 2.1 migrating cells) was significantly increased as compared to PBS controls (4.0 ± 0.77 migrating cells, Figure 5-2B). These data suggest that MCP-1 is a chemoattractant for Gr1+CD11b+ cells from IL-10 deficient mice. 116 The Ibidi chemotaxis chamber has the advantage of allowing us to confirm that migrating cells were Gr1+CD11b+. The ibidi chamber also has the added advantages of using a small amount of chemoattractant per well and allows us to determine the speed and pathway of cell movement. However, only three ibidi chemotaxis chambers fit on one slide that has to be monitored by a camera attached to a microscope. Therefore only three replicates can be monitored simultaneously on one microscope. In addition, data collection from these three replicates was time consuming and required approximately 24 hours of preparation, experimentation and analysis time. This method was not ideal due to increased cost in time and in animal sacrifice. 117 A * Net Migrating Cells 50 40 30 20 10 0 -10 PBS fmlp (1µM) B * Net Migrating Cells 20 10 0 0. P 1µ M BS M 0. 5µ CP -1 M M C 1µ PM 1 M 2µ CP -1 M M 4µ CP -1 M M 6µ CP -1 M M 8µ CP -1 M M C P1 -10 Figure 5-1 Neutrophil and MDSC Underagarose Chemotaxis Assay This figure illustrates the number of net migrating (A) neutrophils and (B) MDSC in the underagarose chemotaxis assay. Cells were isolated from the bone marrow via percoll density gradient centrifugation. Neutrophils were isolated from wild type mice and MDSC were isolated from IL-10-/- mice. Migration was determined following by counting the number of migrating cells following a 4 hour incubation at 37°C in 5% CO2. n = 3-7, * indicates significant increase (p<0.05) from PBS, (A) Two-tailed Student’s t test, (B) One-way ANOVA followed by Bonferroni’s Post test 118 A B Migrating Cells 25 * 20 15 10 5 0 PBS MCP-1 (1 µM) Figure 5-2 MDSC Ibidi Chemotaxis This figure illustrates (A) a representative image of Gr1+CD11b+ cells and (B) the number of migrating MDSC in the Ibidi chamber chemotaxis assay. MDSC were isolated from the bone marrow of IL-10-/- mice via percoll density gradient centrifugation. Cell movement was monitored via an inverted microscope over 16 hours at 37°C. Cells with net movement greater than 2 mm towards the chemoattractant reservoir were considered chemotactic. Cells were fixed and labeled with Gr1 (FITC) and CD11b (PE) antibodies. Double-labeled cells are golden coloured. n = 6, * indicates significant increase (p<0.05) from PBS, Twotailed Student’s t test 119 Our data in MDSC chemotaxis assays show that MCP-1 is a chemoattractant. The ideal assay would allow the testing of a number of potential chemoattractants with a reasonable use of resources and allow for the confirmation of migrating cells as being Gr1+CD11b+. We were successful in achieving both of these goals with the transwell chamber chemotaxis assay. 5.2 Transwell Chamber Chemotaxis Assay The transwell chamber, as shown in the attached image, has an upper chamber and a lower chamber connected by an 8 μc porous membrane. Cells (5 x 105 bone marrow derived MDSC) were placed in the upper chamber, which has pores at the bottom to allow for cells to migrate to the lower chamber, which contains the chemoattractant. Migrated cells suspensions were cytospun onto slides. These slides were labeled with Gr1+CD11b+ antibodies with the same protocol as the colon immunohistochemistry. Gr1+CD11b+ cells were counted as net migrating cells. The percentage of migrating cells that were Gr1+CD11b+ was always greater than 90%. Vehicle only in the lower well was used as a control. Chemoattractant in both the upper and lower chambers was used as a control for chemokinesis. The transwell assay allowed us to test a large number of conditions to be performed on the same day and confirm that migrating cells are Gr1+CD11b+. 120 Confirming results from previous assays, our preliminary data in the transwell assay showed that MCP-1, but not fMLP, was a chemoattractant for MDSC. Our data also showed that greater than 90% of cells migrating towards MCP-1 were Gr1+CD11b+. Having established the assay, we next performed a dose response curve for MDSC migration towards MCP-1. Migration towards MCP-1 was significantly increased from control levels at 50 and 100 ng/ml concentrations (Figure 5-3). Next, we tested various factors identified as important for MDSC recruitment in literature and common leukocyte chemoattractants in the transwell chamber chemotaxis assay. We found no significant increase in MDSC migration towards IL-1β (10 ng/ml - 2 mltio), IL-6 (100 ng/ml - 2 g/ml), prostaglandin E2 (1 ng/ml 1 μg/ml), KC (0.1 μM - 2 μM), fMLP (0.5 ng/ml - 2 μg/ml) and VEGF (100 ng/ml - 1 mll V) as compared to control levels. Figure 5-4 shows the number of migrating cells towards one representative concentration from these chemoattractants. Using the transwell chamber chemotaxis assay, we also identified SDF1/CXCL12 as a chemoattractant for MDSC in this model (Figure 5-5). Migration towards SDF-1 was significantly increased from controls at 100, 150 and 200 ng/ml concentrations. We tested SDF-1 in our model because SDF-1 and its receptor have previously been identified as being key to MDSC recruitment in 121 ovarian cancer and related to MDSC in a model of breast cancer (Liu et al., 2010; Obermajer et al., 2011b). Chemotaxis is the movement of cells along a gradient towards a chemoattractant. Some agents induce chemokinesis in cells, which is movement in random directions. We confirmed that MCP-1 and SDF-1 induce chemotaxis and not chemokinesis in MDSC by loading the concentration of the chemoattractant that induced highest migration of cells in both the upper and lower chamber (Figure 5-3 & 5-5). This did not increase the migration of cells into the lower chamber for MCP-1 (100 ng/ml) or SDF-1 (150ng/ml). We next compared MDSC migration in cells derived from TLR4 competent and deficient IL-10-/- mice. Interestingly, migration was comparable in MDSC derived from IL-10-/- and IL-10-/- TLR4-/- mice towards both MCP-1 and SDF-1 (Figure 56). This finding is in agreement with previous data that suggested that TLR4 competence does not affect MDSC function. These data show that MCP-1 and SDF-1 are MDSC chemoattractants in the IL-10 deficient mouse model. Both MCP-1 and SDF-1 have been related to MDSC recruitment in previous studies, but we are the first to show that they are direct chemoattractants. 122 8 6 * * 4 2 0 C on 12 t r o .5 ng l / 2 5 ml ng / 50 ml ng 10 / m 0n l g 20 /m 0n l g 40 /m 0n l g/ C 800 ml he n m g/m ok in l es is Migrating Gr1+CD11b+ Cells (x 10^4 cells/ml) 10 Figure 5-3 MDSC migrate towards MCP-1 in the Transwell Chemotaxis Assay This figure illustrates a dose response curve of the number of migrating MDSC towards MCP-1 in the transwell chamber chemotaxis assay. MDSC were isolated from the bone marrow of 3-month-old IL-10-/- mice via percoll density gradient centrifugation. Cells well allowed to migrate at 37°C for three hours. Migrating cells were confirmed as Gr+CD11b+ via immunohistochemistry. n ≥ 3, * indicates significant increase (p<0.05) from Control, One way ANOVA followed Bonferroni’s Post Test 123 by Migrating Gr1+CD11b+ Cells (x 10^4 cells/ml) 10 * 8 6 4 2 l) /m /m -6 IL β (1 µg ng -1 IL PG (1 µ l) E2 g/ m K (1 µ l) C g VE (10 / m l) 0 G n F (5 g/m fM 0 LP 0 n l) (1 g/m 00 l) ng /m l) M C P- 1 (1 00 C on tr ol 0 Figure 5-4 MDSC migration in the Transwell Chemotaxis Assay This figure illustrates the number of migrating MDSC towards representative doses of various factors in the transwell chamber chemotaxis assay. MDSC were isolated from the bone marrow of 3-month-old IL-10-/- mice via percoll density gradient centrifugation. Cells were allowed to migrate at 37°C for three hours. Migrating cells were confirmed as Gr+CD11b+ via immunohistochemistry. n ≥ 3, * indicates significant increase (p<0.05) from Control, One way ANOVA followed by Bonferroni’s Post Test 124 is l es C he m ok in m l g/ 40 0n g/ m l * 0n 20 15 0n g/ m l m l 0n g/ /m 10 ng tr 50 on C * * ol Migrating Gr1+CD11b+ Cells (x 10^4 cells/ml) 11 10 9 8 7 6 5 4 3 2 1 0 Figure 5-5 MDSC migrate towards SDF-1 in the Transwell Chemotaxis Assay This figure illustrates a dose response curve of the number of migrating MDSC towards SDF-1 protein in the transwell chamber chemotaxis assay. MDSC were isolated from the bone marrow of 3-month-old IL-10-/- mice via percoll density gradient centrifugation. Cells were allowed to migrate at 37°C for three hours. Migrating cells were confirmed as Gr+CD11b+ via immunohistochemistry. n ≥ 3, * indicates significant increase (p<0.05) from Control, One way ANOVA followed by Bonferroni’s Post Test 125 A 10 Migrating Cells (*10^4 Cells/ml) 8 IL-10-/IL-10-/-/TLR4-/- * * 6 4 2 C on 12 t r o .5 ng l / 2 5 ml ng / 50 ml ng 10 / m 0n l g 20 /m 0n l g 40 /m 0n l g/ C 800 ml he ng m ok /ml in es is 0 B Migrating Cells (*10^4 Cells/ml) 12 * * 8 IL-10-/IL-10-/-/TLR4-/- * 4 is l es ok in l m C he m 40 0n g/ m l g/ 0n 20 15 0n g/ m l l m g/ /m 0n 10 ng 50 C on tr ol 0 Figure 5-6 TLR4 does not affect MDSC chemotaxis in the IL-10 deficient mouse This figure illustrates the number of migrating MDSC towards (A) MCP-1 and (B) SDF-1 in the transwell chamber chemotaxis assay. MDSC were isolated from the bone marrow of 3-month-old mice via percoll density gradient centrifugation. Cells were allowed to migrate at 37°C for three hours. Migrating cells were confirmed as Gr+CD11b+ via immunohistochemistry. n ≥ 3, * indicates significant increase (p<0.05) from Control, Statistical analysis: One way ANOVA followed by Bonferroni’s Post Test 126 5.3 MCP-1 increases MDSC chemotaxis in vivo Our in vitro data have identified MCP-1 and SDF-1 as MDSC chemoattractants. Next, we wanted to determine whether this would translate into increased recruitment of MDSC in vivo. We employed the use of intravital microscopy to study in vivo recruitment kinetics in leukocyte-endothelial cell interaction in skeletal muscle vasculature as previously described (McCafferty et al., 2002). These studies are challenging and require significant time and practice to perfect. In order to complete these studies for my thesis the following experiment was performed in collaboration with Dr. Ying Gao, a post-doctoral fellow under Dr. McCafferty’s supervisions, and Dr. Katarzyna Stevens of the intravital core, University of Calgary, who routinely use this technique. I isolated MDSC from the bone marrow and labeled them while Dr. Gao or Stevens performed the intravital studies and analysis. Wild type mice at 3 months of age were given MCP-1 (300 ng in 0.1 ml) or saline (0.1 ml) subcutaneously. Mice were then anesthetized with a ketamine-xylazine cocktail and the skeletal muscle was isolated with blood vessels intact as previously described (McCafferty et al., 2002). MDSC were isolated from IL-10-/- at 6 months of age and labeled with CFSE dye. Cells (upto 10 x 106) were incubated with 1 ml of 1 μ CFSE dye for 15 minutes. Three hours after the subcutaneous injection of MCP-1 or saline, 4 x 106 CFSE labeled MDSC were injected by intravenous jugular route. The adherence of green fluorescent cells was recorded for 1 minute in 5 post-capillary venules (20 - 40 μ 2 in size) every fifteen minutes for 30 minutes. MDSC adhesion was defined as a 127 fluorescent cell being stationary for 30 seconds. Data are presented as the average number of adherent cells in 5 vessels at the 30-minute time point. Our initial data showed a low number of adhering cells in capillaries, so we also recorded a field of view (40X magnification) for 1 minute every fifteen minutes for one hour. Data are presented as the average adhering cells in fields of views at the 30-minute time point. Adhesion of MDSC was significantly increased upon MCP-1 treatment. The average number of adhering MDSC in 5 vessels increased from 0.65 ± 0.17 adherent cells in saline treated mice to 1.8 ± 0.54 adherent cells in MCP-1 treated mice (Figure 5-7A). The number of adherent cells were also increased in the field of view at 40X magnification from 24.0 ± 6.5 adherent cells in saline treated mice to 47.1 ± 7.4 adherent cells in MCP-1 treated mice (Figure 5-7B). Next, we also studied MDSC-endothelial cell interaction in colon vasculature of wild type, IL-10-/- and IL-10-/- TLR4-/- animals as previously described (Qi et al., 2005). The purpose of this experiment was to compare MDSC recruitment kinetics in untreated mice in the vasculature of the colon, the site of inflammation and cancer in our models. CFSE labeled MDSC (4 x 106) isolated from 6-monthold IL-10 deficient mice were injected into untreated 3-month-old wild type, IL-10/- and IL-10-/- TLR4-/- mice via femoral vein. The adherence of fluorescent cells was recorded in five fields of view (40X magnification) in the ascending colon 128 from the serosal side at 30 minutes post cell injection. Data are presented as the average adhering cells at the 30-minute time point. MDSC adherence was comparable in IL-10-/- (33.8 ± 5.6 adherence cells) and IL-10-/- TLR4-/- (35.3 ± 12.2 adherence cells) mouse colons. MDSC adherence trended higher in these mutant mice than wild type (15.6 ± 5.6 adherence cells), but the change was not significant. Our in vivo data suggest that MCP-1 increases MDSC recruitment in skeletal muscle vasculature. There is a possibility that cells activated during the isolation procedure are binding in a non-specific manner. However, increased cell adhesion is observed upon MCP-1 treatment indicating that the binding is not non-specific. Using fMLP as a negative control would further help confirm that this was not non-specific binding. We observed a relatively low number of cells adhering per vessel even after MCP-1 treatment. This experiment could be further optimized by modulating the dose of MCP-1, time between drug treatment and cell injection or the number of cells injected. We used skeletal muscle to study MDSC recruitment kinetics in vivo because it is easily transilluminated compared to the colon. It is however possible that MDSC-endothelial interactions in skeletal vasculature differ from that in the colon. Our initial data do not show MDSC adherence was increased in IL-10-/- and IL-10-/- TLR4-/- than in wild type mouse colons. There is however a trend towards increased recruitment that may be significant in a larger sample size. From our previous data, we would expect 129 increased MDSC adherence in the absence of TLR4 in the IL-10-/- colon. Interestingly, MDSC adherence levels were comparable in IL-10-/- and IL-10-/TLR4-/- mouse colons. This may be explained by a number of factors. We did not score the colons for inflammation and cancer. The IL-10 deficient mouse is a variable model and the level of disease development would affect cell recruitment kinetics. It is possible in a sample size of 3 that dysplasia levels were not reflective of the difference between the two mutant mice. These data suggest no differences in MDSC-endothelial adherence but cell recruitment is a multi-step process that involves rolling, adherence and transmigration across the endothelial barrier. TLR4 may affect transmigration rather than adherence. Future investigations could compare the expression of molecules involved in adherence and transmigration in IL-10-/- and IL-10-/- TLR4-/- colons. We studied intact colons from the serosal end using a spinning disk microscope, which allows us to observe vasculature in the muscle and submucosal layers of the colon. Dysplasia is initiated and largely found in the mucosal layer. MDSC recruitment kinetics maybe different in the mucosal layer than those observed from the serosa. Further intravital studies from the mucosal end could help delineate this difference. These data show that MCP-1 significantly enhances MDSC adhesion in vivo. Taken together with the in vitro data, our data make a strong case for MCP-1 as a chemoattractant for MDSC in the IL-10 deficient mouse. 130 Adhering Cells (Average in 5 Vessels) A * 2.5 2.0 1.5 1.0 0.5 0.0 Saline MCP-1 (300ng) B Adhering Cells (4X Field of View) 60 * 40 20 0 Saline MCP-1 (300ng) Figure 5-7 MCP-1 increases MDSC recruitment kinetics in vivo This figure illustrates the number of adherent MDSC in (A) 5 post-capillary venules (20 and 40 μ 0) and (B) field of view at 40X magnification in the cremaster muscles of 3-month-old wild type mice in vivo. Mice were given subcutaneous injections of saline or MCP-1. MDSC were isolated from the bone marrow of 6-month-old IL-10-/- and labeled with CFSE. 4 x 106 labelled MDSC were injected and adherence was observed using a spinning disk microscope. n = 3-4, * indicates significant increase (p<0.05) from Control, Two-tailed Student’s t test 131 Adhering Cells (4X Field of View) 50 40 30 20 10 0 Wild type IL-10-/- IL-10-/- TLR4-/- Figure 5-8 MDSC recruitment kinetics in the colon in vivo This figure illustrates the number of adherent MDSC in field of view at 40X magnification in the ascending colon of 3-month-old untreated mice. MDSC were isolated from the bone marrow of 6-month-old IL-10-/- mice and labeled with CFSE. 4 x 106 labelled MDSC were injected and adherence was observed using a spinning disk microscope. Wild type n = 2, Mutant n = 3, Two-tailed Student’s t test between mutant mouse groups 132 The data from in vitro and in vivo recruitment studies suggest that MDSC can be recruited to SDF-1 and MCP-1. However, TLR4 deficiency does not alter their ability migrate towards these factors in vitro. These data further suggest that the increased MDSC recruitment into the colon in the absence of TLR4 is mediated by a change in the tissue microenvironment. Our previous data indicate that levels of MDSC in the spleen were comparable in IL-10-/- and IL-10-/- TLR4-/- mice. We therefore examined how TLR4 deficiency effects the expression of these chemoattractants in colon tissue. 5.4 Chemoattractant protein and mRNA levels in the IL-10 deficient mouse We determined chemoattractant levels in the proximal and distal colons of IL-10-/and IL-10-/- TLR4-/- rather than whole colons because earlier experiments indicated that TLR4 levels are modulated differently in the proximal and distal part of the colon. Chemoattractant levels were determined in 6-week and 6month-old mice to determine how chemoattractant levels are changed after disease development. MCP-1 and SDF-1 mRNA levels were determined in colon tissue by the real time polymerase chain reaction. RNA was extracted from tissue with Trizol and purified. Purified RNA was reverse transcribed into cDNA and amplified with chemoattractant and GAPDH, the housekeeping gene, probe and primers. One of samples from the 6-week-old IL-10-/- group was given the relative value of 1. 133 MCP-1 relative quantification levels over GAPDH were unchanged between IL10-/- and IL-10-/- TLR4-/- colons in the proximal and distal colon in both age groups (Figure 5-9). The SDF-1 RT-PCR did not generate quantifiable results. Next, we determined protein levels in colons via western blot analysis. Laemmli buffer was added to lysed colon tissue and boiled. This solution was run through stracking and separating SDS-PAGE gels by electrophoresis. Protein was transferred to nitrocellulose and incubated with chemoattractant antibody overnight. The next day, the membrane was incubated with secondary antibody and visualized. A representative blot is provided in Figure 5-10A. MCP-1 protein levels were comparable in the proximal colon at 6 weeks of age. MCP-1 protein levels were significantly increased in IL-10-/- TLR4-/- colons as compared to IL-10-/- in the proximal colon at 6 months of age and in the distal colon at both 6 weeks and 6 months of age (Figure 5-10B). We have shown that MCP-1 protein is significantly increased in the IL-10 deficient mouse colon in the absence of TLR4. These data suggest that TLR4 deficiency increases MCP-1 levels in the colon leading to increased recruitment of protumorigenic MDSC. MCP-1 protein is subject to methylation and posttranslational modifications, especially in inflammatory environments. A western blot the protein can show up in stretched or multiple bands (Fujita et al., 2010; Grammas and Ovase, 2001). In our blots, we found MCP-1 protein in two distinct bands (Figure 5-10A). 134 Relative Quantification (MCP-1/GAPDH) 15 10 5 -1 IL0 -/- 10 TL /-P R -/ ro - x IL I Pr im -1 L- ox al 0 -/- 10 - im TL /-D al R -/ is - ta D l is ta l IL IL -1 IL0 -/- 10 TL /-P R -/ ro - x IL I Pr im -1 L- ox al 0 -/- 10 - im TL /-D al R -/ is - ta D l is ta l 0 6 Week Old 6 Months Old Figure 5-9 MCP-1 mRNA levels are comparable in the IL-10 deficient mouse in the absence of TLR4 This figure illustrates the levels of MCP-1 messenger RNA in mouse colon tissue from IL-10 deficient and IL-10 TLR4 double deficient mice. RNA was purified from colon tissue with Trizol and complimentary DNA was generated. Messenger RNA levels were quantified by real time polymerase chain reaction. n = 3, Oneway ANOVA followed by Bonferroni’s Post test 135 A I 0 -/- L-1 MCP-1/Actin Average Intensity TL 0 -/P R 4 -/- rox IL i -1 IL Pro ma 0 -/- -1 x l i TL 0 -/- ma R Di l 4 -/- st D al is ta l B 2.5 * 2.0 1.5 1.0 0.5 * * I 0 -/- L-1 TL 0 -/R Pr 4 -/- ox IL i -1 IL Pro ma 0 -/- -1 x l i TL 0 -/- ma R Di l 4 -/- st D al is ta l -1 IL IL -1 0.0 6 Week Old 6 Months Old Figure 5-10 MCP-1 protein levels are significantly increased in in the IL-10 deficient mouse in the absence of TLR4 This figure illustrates (A) representative western blots of MCP-1 and βoactin protein in colon tissue from 6-month-old mice and (B) MCP-1 protein levels quantified by densitometry over actin in mouse colon tissue. Protein levels were determined by western blot analysis. n = 3, * indicates significant change between groups (p<0.05), One-way ANOVA followed by Bonferroni’s post test 136 SDF-1 has previously been reported to be expressed in human colons and colonic cells lines as well as mouse colons (Ding et al., 2013; Mikami et al., 2008; Schimanski et al., 2008). We used the same western blot protocol as MCP-1 to determine SDF-1 protein levels in the colon. We were unable to visualize any SDF-1 protein from mouse colons via western blot analysis. Only the SDF-1 protein positive control band was visualized at the correct size. As SDF-1 protein may be present only in minute quantities in our model, we decided to increase the potential protein in the western blot by immunoprecipitation. Agarose bead slurry was incubated with lysed colon tissue to remove non-specific binding particles. After centrifugation, the supernatant was incubated with SDF-1 antibody overnight. The next day this solution was incubated with agarose beads. After centrifugation, the pellet was washed, resuspended in SDS buffer and boiled. This solution was run through the western blot as previously described. We were again unable to visualize any bands except the positive control. It remains unclear whether SDF-1 plays a role in MDSC recruitment to the colon in this model. It is possible that either SDF-1 does not play a role in this model or that protein levels are too low to be detectable by immunoprecipitation and western blotting techniques. A radiometric assay or enzyme linked immunosorbent assay may be able to detect SDF-1 in IL-10 deficient colons. Another possibility is that the two different antibodies used are not compatible with the SDF-1 protein in 129SvEv mice. 137 This chapter aimed to elucidate the relationship between the absence of TLR4 and increased MDSC recruitment in the IL-10 deficient mouse. To understand how TLR4 was affecting MDSC recruitment to the colon, we first identified MCP1 and SDF-1 as MDSC chemoattractants via both in vitro and in vivo assays. Then our data showed that in the absence of TLR4, MCP-1 protein levels are increased in the IL-10 deficient mouse. These data taken together show a potential sequence of events where as disease development progresses in the IL-10 deficient mouse, TLR4 deficiency leads to increased MDSC chemoattractants in the colon. These chemoattractants increase MDSC recruitment into the colon. This increased MDSC recruitment, as our earlier findings have shown, is associated with increased cancer development in colitisassociated cancer. 138 Chapter 6: Discussion and Future Directions 139 Colorectal cancers are neoplasms of the colon, rectum and appendix. CRCs have a heavy disease burden and are a leading cause of cancer related deaths in Canada and worldwide (Parkin et al., 2010; Singh et al., 2012a). Colon adenocarcinoma is the most common presentation of CRC and a heterogeneous combination of events can lead to its development (Kumar, 2010). Long-standing inflammation in the colon is related to an increased risk of developing adenocarcinoma (Bernstein and Nabalamba, 2007). While the link between inflammation and cancer development has been established in CRC and many other types of cancers, the underlying mechanisms remain poorly understood. A key characteristic of both inflammation and cancer is the presence of leukocytes (Goldszmid and Trinchieri, 2012). One leukocyte population that plays a role in both inflammatory bowel disease and colon cancer is the myeloid derived suppressor cell (Ostrand-Rosenberg and Sinha, 2009). Murine MDSC are identified by Gr1 and CD11b markers, and the ability to suppress the immune response (Melani et al., 2003). These are immature cells that are recruited away from the bone marrow to sites of cancer. There MDSC are thought to promote tumor formation by suppressing the immune response and promoting angiogenesis. The MDSC also likely play an immunosuppressive role in inflammation and this is still being explored (Kong et al., 2013). The role of MDSC and their mechanisms have not been colitis associated cancer have not been studied. 140 Three main types of experimental models of colitis-associated cancer are studied in literature. Ours is unique in that in the absence of IL-10, it is spontaneous chronic inflammation with key similarities to clinical IBD that drives adenocarcinoma development as mice age (Zhang et al., 2007b). We found that MDSC recruitment was elevated to the colon in this model. We are also in the unique position of being able to study the role of TLR4 in cancer development with an IL-10 and TLR4 double knockout mouse (Zhang et al., 2007a). TLR4 and related pathways are thought to be homeostatic regulators of intestinal inflammation (Gonzalez-Navajas et al., 2010; Greten et al., 2004). The loss of this innate immune receptor is involved in the progression from inflammation to the cancer phenotype at a much earlier time point (Zhang et al., 2007a). Furthermore, novel data presented in this thesis illustrate a profound recruitment of MDSC to the gut of IL-10 deficient mice as they age, which is exacerbated in the absence of TLR4 and therefore make a potential contribution to disease progression. Our studies examined the role of MDSC in the IL-10 deficient model of inflammation-associated cancer and the potential role of TLR4 in their regulating recruitment and function. Despite the presence of MDSC in cancer models, little is known about their recruitment factors. We hypothesized that MDSC contribute to cancer development in this model and TLR4 modulates their recruitment to affect cancer development in colitis-associated cancer. 141 First, we established the presence of MDSC in the IL-10 deficient colon by using immunohistochemistry and flow cytometry analysis for Gr1+CD11b+ cells. Using flow cytometry we demonstrated increased MDSC levels with age, which is also associated with cancer development. This is the first time MDSC have been visualized in colon sections via immunohistochemistry. Employing this technique gave us the added advantage of being able correlate dysplasia development and MDSC levels in the same colon. Colon tissues were examined for location of MDSC and we found cells accumulate in areas of cancer, supporting data in literature that suggest MDSC are recruited to the tumor microenvironment (Gabrilovich and Nagaraj, 2009). Interestingly, in IL-10-/- mice on the C57BL/6 background Singh et al. reported that low levels of MDSC associated with inflammation (Singh et al., 2012b). They showed that at 6 months of age IL-10-/mice on the C57BL/6 background approximately 0.4 % of colonic lamina propria cells are MDSC. In our 6-month-old IL-10-/- mice on the 129SvEv background MDSC counts are 8.8 ± 1.2 % determined by flow cytometry analysis for Gr1+CD11b+ cells. This difference is striking even accounting for slightly altered cell isolation protocols and different antibody suppliers. This is probably due to C57BL/6 mice being less susceptible to developing inflammation (Mahler et al., 1998). This study associated MDSC with inflammation in the IL-10 deficient mouse, but our interest is in their role in colitis-associated cancer. 142 A striking observation in these studies is that MDSC recruitment in the colon is increased in the IL-10 deficient mouse in the absence of TLR4. TLR4 deficiency did not however, alter MDSC levels in the bone marrow or spleen. These data suggest TLR4 in some manner regulates MDSC recruitment to the gut. We observed a strong correlation between MDSC levels with dysplasia scores in the colons of wild type, IL-10 deficient and IL-10 TLR4 double deficient mice. MDSC were the only cell type significantly increased in the colon the absence of TLR4. Interestingly, TLR4 deficiency decreased T-cell counts in the colon further supporting an immunosuppressive role of MDSC in this model. We also found elevated Gr1+CD11b+ cell counts in the AOM/DSS model of colitis-associated cancer. This demonstrates that MDSC presence is not limited to the IL-10 model and may play a role in other models of inflammation-associated cancer. Further in vivo and in vitro MDSC functional studies in the AOM/DSS and other models are warranted to determine if MDSC have a universal role in colitis-associated cancer. Next, we studied Gr1+CD11b+ cell function in vitro in the IL-10 deficient mouse. It was critical to establish immunosuppressive functionality in these cells to confirm they are MDSC. Gr1 and CD11b can be present in a number of myeloid cell populations, such as neutrophils, but it is the ability to both express these markers and to suppress the immune response that characterizes MDSC (Gabrilovich and Nagaraj, 2009). We achieved this by establishing the presence 143 of known immunosuppressive pathways in MDSC, arginase I and NOS 2 enzymatic activity, and by showing the ability of MDSC in the IL-10 deficient model in suppressing the proliferation of T-cells. We found that approximately 5% of MDSC express TLR4. TLR4 competence did not affect MDSC function in the arginase I activity, NOS2 activity and T-cell proliferation suppression assays. These are three of the most commonly used assays in assessing MDSC immunosuppressive functionality (Gabrilovich and Nagaraj, 2009; OstrandRosenberg and Sinha, 2009). MDSC suppress the immune response and contribute to the local microenvironment by other mechanisms, such as direct cell-to-cell contact CD8+ cell suppression, and production of immunosuppressive and proangiogenic factors, such as VEGF, that were not tested here (Dolcetti et al., 2008; Nagaraj et al., 2007). Interestingly, one of the immunosuppressive factors produced by MDSC is IL-10, which suppresses the immune response in macrophage and Tcells by decreasing the production of cytokines such as IL-12 (Koike et al., 2012; Obermajer and Kalinski, 2012; Sinha et al., 2007a; Srivastava et al., 2012). TLR4 has been proposed to regulate MDSC cross talk and function. Koike et al report that IL-10 production by MDSC is MyD88 dependent, which is activated by pattern recognition receptors, such as TLR4 (Koike et al., 2012). Interestingly, TLR4 deficiency decreased MDSC function, including IL-10 production, by cutting off cross talk between MDSC and macrophage (Bunt et al., 2009). The lack of 144 functional differences between TLR4 competent and deficient MDSC in our model could be due to the added deficiency in IL-10. We have already established the ability of IL-10 deficient MDSC in suppressing T-cell proliferation. Therefore IL-10 is not essential for MDSC immunosuppression in our model. However, it is possible that IL-10 competent MDSC could have a further enhanced ability to suppress T-cell or macrophage function. A comparison of IL10 competent and deficient MDSC could indicate the importance of this immunosuppressive mechanism in MDSC. The potential role of TLR4 in MDSC immunosuppressive function could be addressed with functional assays in TLR4 deficient but IL-10 competent MDSC. One of the issues with this experiment is the low levels of MDSC in healthy wild type mice. An alternate model that induces MDSC expansion would have to be used or MDSC would have to be generated from bone marrow cells ex vivo, a methodology for this has been described recently (Lechner et al., 2010). We observed enhanced immunosuppressive functionality in spleen derived MDSC over bone marrow derived cells from both IL-10 deficient and IL-10 TLR4 double deficient mice. MDSC could be going through a type of activation step through recruitment factors and/or by peripheral exposure to various factors. Exposure to a number of inflammation related factors including prostaglandin E2, IL-1β, interferon γ and tumor necrosis factor α has been shown to induce increased functionality in MDSC (Sevko and Umansky, 2013; Yang et al., 2013). 145 Increased levels of these cytokines have been reported in the IL-10 deficient mouse (Kawachi et al., 2000). IL-10-/- TLR4-/- mice have similar levels of inflammation to IL-10-/- mice and their MDSC have similar immunosuppressive abilities suggesting that they are exposed to a similar activating environment. IL1e has been shown to expand MDSC numbers and to enhance their function through a NF-κh dependent pathway (Tu et al., 2008). Our data demonstrate that no difference between TLR4 competent and deficient MDSC functionality indicating that NF-κu is not important or is activated by a different mechanism in MDSC function. Prostaglandin E2 has been shown to induce MDSC and enhance their expression of arginase I and NOS2 enzymes in both human and murine models (Obermajer et al., 2011a; Obermajer et al., 2011b; Obermajer et al., 2012; Sinha et al., 2007b). Prostaglandin E2 may be an interesting target for future studies in this model because it is secreted by colon cancer cells and our work has shown that COX-2 expression, which is part of the prostaglandin E2 production pathway, is enhanced in this model in the absence of TLR4 (data not shown). Interestingly, Watanabe et al. have reported that Gr1+CD11b+ cells found in the spleens of tumor-bearing mice are potent immunosuppresors but Gr1+CD11b+ cells from wild type spleens had much lower suppressive effects (Watanabe et al., 2008). Our results are consistent with the hypothesis that inflammatory cytokines stimulate MDSC in the periphery to enhance their function. 146 To demonstrate a role of MDSC in vivo in our model of colitis-associated cancer we took two approaches: depletion and supplementation. Low dose chemotherapeutics such as gemcitabine and 5-FU have previously been employed to target MDSC and reduce cancer development in other models (Suzuki et al., 2005; Vincent et al., 2010). We recognize that there is an inherent issue with using cytotoxic drugs in this study. However, in vivo and in vitro work by these other groups showed that low doses of these drugs target MDSC more potently and efficaciously than cancer cells and other leukocytes. We showed that repeated low doses of 5-FU and gemcitabine significantly decrease MDSC counts and dysplasia development in the colon. Vincent et al. reported that 5-FU targets MDSC in more potent manner than gemcitabine. This observation was repeated in our results where we found that at the same dose, 5-FU decreased MDSC counts and dysplasia further than gemcitabine. Vincent et al. used a single dose gemcitabine and 5-FU in a model of thymoma cancer cell line injection murine model of cancer. Le et al. used both single and multiple doses of gemcitabine in a 4T1 mammary carcinoma cell line injection model of murine cancer (Le et al., 2009; Vincent et al., 2010). Interestingly, while both these studies showed decreased tumor sizes, neither showed the dramatic decrease in dysplasia development observed by us in the IL-10 deficient mouse, where we observed no adenocarcinoma formation, only low-grade adenomas in some mice. This may be explained by differences between the models, but also potentially a more critical role for MDSC in dysplasia development in colitis associated cancer 147 and warrants further study. Recently, others have shown that the adoptive transfer of MDSC increases immunosuppression in models of inflammation or cancer (Parekh et al., 2013; Schmidt et al., 2013). Schmidt et al. showed that MDSC adoptive transfer increases immunosuppression acutely in a model of hepatocellular carcinoma. Another recent study showed increased mortality following an adoptive transfer of MDSC, which suggested an important role for MDSC in cancer progression in a model of multiple myeloma (Ramachandran et al., 2013). Our data indicate increased cancer development with adoptive transfers of MDSC in a model of inflammation-associated cancer. To our knowledge, this work is the first to show increased dysplasia with an adoptive transfer for MDSC. A number of groups have shown MDSC as having an anti-inflammatory effect in colitis and other models of inflammation (Haile et al., 2008; Parekh et al., 2013; Singh et al., 2012b). In light of these recent studies, a surprising finding from our work was the lack of effect on inflammation despite significant differences in dysplasia with in vivo MDSC depletion and adoptive transfer studies. This may be explained by a number of factors. We scored inflammation macroscopically and histologically through scoring systems, which may not incorporate minute but significant changes in inflammation that could be reflected in more sensitive assays such as myeloperoxidase levels or assays for cytokines levels. Haile et al. showed MDSC are immunosuppressive in vivo in a T-cell transfer model of colitis 148 by co-injecting MDSC with the inflammation inducing T-cells. This is similar to decreasing T-cell responsiveness in vitro by co-incubation with MDSC and does not necessarily suggest that MDSC reduce established colitis severity. Singh et al. showed resveratrol treatment decreases inflammation and increases MDSC recruitment in the IL-10 deficient mouse, but did not establish a cause and effect relationship between the two changes (Haile et al., 2008; Singh et al., 2012b). Zhang et al. have suggested that MDSC recruitment is elevated and their adoptive transfer decreases inflammation in DSS colitis (Zhang et al., 2011). Our data also show that MDSC are recruited to cancer lesions in nearly 5X higher numbers than to inflammatory lesions within the same colons. In our model, where dysplastic lesions are common, it is likely that MDSC are not recruited to inflammatory lesions in large enough numbers to affect a change. We looked at inflammation at a point where dysplasia is evident, but earlier time points may show different results. Future studies could examine inflammation at an earlier time point where cancer phenotype has not yet developed. We showed that TLR4 mRNA is downregulated in IL-10 polyps compared to nonhyperplastic mucosa. This evidence suggests that TLR4 may play a role in the modulation of dysplasia development through alterations in colonic tissue. We hypothesized that the absence of TLR4 signaling is enhancing cancer development in the colon by increasing MDSC recruitment through a change in the microenvironment, such as the modulation of chemoattractant levels. To 149 determine how TLR4 was enhancing MDSC recruitment in the colon, we first had to determine which factors affect MDSC recruitment. Numerous studies have identified factors that play a role in MDSC recruitment (Sevko and Umansky, 2013). These studies have generally been conducted in vivo with genetic or pharmacological knockouts. We identified MDSC chemoattractants using in vitro and in vivo chemotaxis assays. We found that MDSC from the IL-10 deficient mouse migrated towards MCP-1 and SDF-1. Both of these factors have been previously related to MDSC recruitment (Huang et al., 2007; Liu et al., 2010). Migration towards either MCP-1 or SDF-1 was not affected by TLR4 deficiency, again indicating that TLR4 does not affect MDSC directly. Interestingly, while message levels were unchanged, we found that MCP-1 protein was significantly increased in the colons of IL-10 deficient mice in the absence of TLR4 at 6 months of age. In 6-week-old IL-10 deficient mice when no inflammation or dysplasia are apparent, MCP-1 protein was significantly increased in the distal colon in the absence of TLR4. This suggests that the lack of TLR4 signaling in the colon increases chemoattractant levels such as MCP-1 , which can recruit MDSC. MCP-1, a CC family chemokine, is primarily secreted by monocyte/macrophage and dendritic cells upon activation and is a potent chemoattractant and activator of macrophage populations (Popivanova et al., 2009). Innate immune receptors such as TLRs play a key role in the initial activation of macrophage and dendritic 150 cells by allowing them to recognize the presence of foreign antigen and initiating the process of activating an immune response via the NF-κi pathway (Danese and Mantovani, 2010; Takeda and Akira, 2004). Later recruitment of macrophages that produce chemokines such as MCP-1 in large quantities happens due to the cytokines released by activated T-helper cells and macrophage at the site of inflammation (Danese and Mantovani, 2010; De Paepe et al., 2009). MCP-1 has been previously related to inflammation and cancer in numerous models. MCP-1 levels are increased in IBD patients and the AOM/DSS model of colitis-associated cancer (Popivanova et al., 2008; Reinecker et al., 1995). Pharmacological blockage or deficiency of the MCP-1 receptor results in decreased cancer development in a model of colitisassociated cancer (Popivanova et al., 2009). MCP-1 and its receptor have been previously related to MDSC recruitment in vivo (Boelte et al., 2011; Huang et al., 2007; Jackson et al., 2013). The mechanism of how TLR4 affects MCP-1 levels and MDSC recruitment to the colon in the IL-10 deficient mouse is an important question that remains to be addressed. Our work supports increased MCP-1 levels in the colon in the absence of TLR4 as the key to elevated MDSC recruitment. The increased MCP-1 levels are unlikely to be a direct effect due to lack of TLR4 signaling. Innate immune receptors such as TLR4 are thought to induce inflammation related cytokines and chemokines such as MCP-1 (Lapara and Kelly, 2010; Yoshimura and Takahashi, 151 2007). However, TLR4 and related pathways also play an important homeostatic role and the loss of this function leads to the development of inflammation in the colon (Fukata and Abreu, 2007; Gonzalez-Navajas et al., 2010; Zhang et al., 2007a). This inflammatory response would certainly involve the activation of the NF-κF pathway in immune cells including MCP-1 secreting macrophage and dendritic cells. The NF-κh pathway has been identified by Greten et al. as a key modulator of inflammation associated cancer in the colon (Greten et al., 2004). TLR4 signaling is also important for epithelial proliferation, a key step in the resolution of inflammation (Ruemmele et al., 2002). Studies in various disease models show that lack of TLR signaling increases inflammation and injury including the production of cytokines such as MCP-1 (Hayashi et al., 2012; Zampell et al., 2012). Therefore, a potential pathway linking TLR4 and MDSC recruitment is the activation of an altered and persistent immune response after the loss of TLR4’s homeostatic functionality leading to increased MCP-1 production and MDSC recruitment. This potential pathway could be verified in further studies with TLR4 antagonists or antibody in the IL-10 deficient mouse. Results from MCP-1 antagonist and antibody studies could determine whether MCP-1 plays a key role in MDSC recruitment in vivo. Successful bone marrow transplant experiments between IL-10 deficient and IL-10 TLR4 double deficient mice could explain whether it is TLR4 deficiency in myeloid cells, colon tissue or a combination of both is important in determining MCP-1 levels, MDSC recruitment and cancer development in this model. 152 We find that TLR4 protects against cancer development in the IL-10 deficient model of colitis-associated cancer. This is in striking contrast to results published by Fukata et al., who reported that TLR4 promotes cancer development in the AOM/DSS model of colitis-associated cancer (Fukata et al., 2007). A number of factors may explain these opposing results. The IL-10 deficient mouse and AOM/DSS are both models of colitis-associated cancer, but the mechanisms of cancer pathogenesis differ greatly. In the AOM/DSS model, the carcinogen, AOM, initiates cancer development and barrier disruption and inflammation induced by DSS are used to drive cancer development. In the IL-10 deficient mouse, cancer initiation and development are both driven by the chronic inflammatory microenvironment in the mucosa. Another key difference between the two models is the extent of MDSC recruitment. MDSC recruitment to the colon in the AOM/DSS model in our hands, where 100 % of mice developed invasive cancer, was four times lower than in 6-month-old IL-10 deficient mice, when only 20 % of mice develop invasive cancers. These data suggest a greater recruitment and better correlation with cancer development for MDSC in the IL-10 deficient than the AOM/DSS model of colitis associated cancer. We have established a key role for MDSC in carcinogenesis in the IL-10 deficient mouse and MDSC recruitment to the colon is even greater in the absence of TLR4. Therefore, our data suggest that MDSC recruitment is one of the key differences that may explain the 153 opposing roles for TLR4 in cancer development in the IL-10 deficient and AOM/DSS models. The link between inflammation and cancer is well established but not fully understood (Ben-Neriah and Karin, 2011; Rutter et al., 2004). Though we do not find MDSC to play an anti-inflammatory role in the IL-10 deficient model, inflammation does recruit MDSC to the colon in our model and numerous others. We have established the critical role for MDSC in cancer development in the colon. Therefore, our data suggest that MDSC may be a critical component in the link between inflammation and cancer. Once recruited by the inflammatory microenvironment MDSC could help transformed cells in immune escape and growth (Ostrand-Rosenberg and Sinha, 2009). In recent years, a number of MDSC subpopulations with functional, phenotypic and morphological heterogeneity have been described (Gabrilovich and Nagaraj, 2009; Haile et al., 2010; Peranzoni et al., 2010). Reports have suggested MDSC in various models may have different roles based on such things as monocytic vs granulocytic lineage, the level of Gr1 expression and the presence of markers such as CD49d (Duffy et al., 2013; Haile et al., 2010). This is one interesting aspect of MDSC biology that is not explored here. Further studies could tease out if MDSC subpopulations have specific roles in the IL-10 deficient mouse. It is 154 possible that certain subpopulations have a more potent immunosuppressive function, produce more proangiogenic factors or may migrate differently towards chemoattractants. The work presented here has shown that active immunosuppressive MDSC are present in the IL-10 deficient mouse model of colitis-associated cancer. We are the first to study their role in a model of inflammation-associated cancer. We have shown that MDSC contribute to cancer development in this model. Our in vivo data show a profound effect on cancer development by MDSC as compared to other models. In a novel IL-10 TLR4 double deficient mouse, cancer development is significantly increased. TLR4 is generally thought to be a proinflammatory mediator, but our data supports growing evidence of its homeostatic role. We find that in the absence of TLR4, MDSC recruitment is also elevated in the IL-10 mouse and correlates with cancer development. We identified MCP-1 and SDF-1 as MDSC chemoattractants. This work contributes to the field by being the first to identify direct MDSC chemoattractants in vitro. We propose TLR4 modulation of MCP-1 levels in the colon tissue as a mechanism for increased MDSC recruitment leading to increased cancer development in the IL-10 deficient model. Our data strongly suggest that the tissue microenvironment and MDSC may be a critical part of that link between inflammation and cancer. 155 References Abreu, M.T., M. Fukata, and M. Arditi. 2005. TLR signaling in the gut in health and disease. J Immunol 174:4453-4460. Ahmadi, A., S. Polyak, and P.V. Draganov. 2009. Colorectal cancer surveillance in inflammatory bowel disease: the search continues. World J Gastroenterol 15:61-66. Almand, B., J.I. Clark, E. Nikitina, J. van Beynen, N.R. English, S.C. Knight, D.P. Carbone, and D.I. Gabrilovich. 2001. Increased production of immature myeloid cells in cancer patients: a mechanism of immunosuppression in cancer. J Immunol 166:678-689. Andersen, V., A. Ernst, J. Christensen, M. Ostergaard, B.A. Jacobsen, A. Tjonneland, H.B. Krarup, and U. Vogel. 2010. The polymorphism rs3024505 proximal to IL-10 is associated with risk of ulcerative colitis and Crohns disease in a Danish case-control study. BMC Med Genet 11:82. Arora, M., S.L. Poe, A. Ray, and P. Ray. 2011. LPS-induced CD11b+Gr1(int)F4/80+ regulatory myeloid cells suppress allergen-induced airway inflammation. Int Immunopharmacol 11:827-832. Atsumi, T., R. Singh, L. Sabharwal, H. Bando, J. Meng, Y. Arima, M. Yamada, M. Harada, J.J. Jiang, D. Kamimura, H. Ogura, T. Hirano, and M. Murakami. 2014. Inflammation amplifier, a new paradigm in cancer biology. Cancer Res 74:8-14. 156 Ayala-Torres, S., Y. Chen, T. Svoboda, J. Rosenblatt, and B. Van Houten. 2000. Analysis of gene-specific DNA damage and repair using quantitative polymerase chain reaction. Methods 22:135-147. Bae, J.M., J.H. Kim, and G.H. Kang. 2013. Epigenetic alterations in colorectal cancer: the CpG island methylator phenotype. Histol Histopathol Beda, N., and A. Nedospasov. 2005. A spectrophotometric assay for nitrate in an excess of nitrite. Nitric Oxide 13:93-97. Ben-Neriah, Y., and M. Karin. 2011. Inflammation meets cancer, with NF-kappaB as the matchmaker. Nat Immunol 12:715-723. Bernstein, C.N., J.F. Blanchard, E. Kliewer, and A. Wajda. 2001. Cancer risk in patients with inflammatory bowel disease: a population-based study. Cancer 91:854-862. Bernstein, C.N., and A. Nabalamba. 2007. Hospitalization-based major comorbidity of inflammatory bowel disease in Canada. Can J Gastroenterol 21:507-511. Bertram, J.S. 2000. The molecular biology of cancer. Molecular aspects of medicine 21:167-223. Bilinski, C., J. Burleson, and F. Forouhar. 2012. Inflammation associated with neoplastic colonic polyps. Ann Clin Lab Sci 42:266-270. Boelte, K.C., L.E. Gordy, S. Joyce, M.A. Thompson, L. Yang, and P.C. Lin. 2011. Rgs2 mediates pro-angiogenic function of myeloid derived suppressor 157 cells in the tumor microenvironment via upregulation of MCP-1. PLoS One 6:e18534. Boland, C.R., and A. Goel. 2010. Microsatellite instability in colorectal cancer. Gastroenterology 138:2073-2087 e2073. Bouma, G., and W. Strober. 2003. The immunological and genetic basis of inflammatory bowel disease. Nat Rev Immunol 3:521-533. Bronte, V., S. Cingarlini, I. Marigo, C. De Santo, G. Gallina, L. Dolcetti, S. Ugel, E. Peranzoni, S. Mandruzzato, and P. Zanovello. 2006. Leukocyte infiltration in cancer creates an unfavorable environment for antitumor immune responses: a novel target for therapeutic intervention. Immunol Invest 35:327-357. Bronte, V., M. Wang, W.W. Overwijk, D.R. Surman, F. Pericle, S.A. Rosenberg, and N.P. Restifo. 1998. Apoptotic death of CD8+ T lymphocytes after immunization: induction of a suppressive population of Mac-1+/Gr-1+ cells. J Immunol 161:5313-5320. Brudecki, L., D.A. Ferguson, D. Yin, G.D. Lesage, C.E. McCall, and M. El Gazzar. 2012. Hematopoietic stem-progenitor cells restore immunoreactivity and improve survival in late sepsis. Infect Immun 80:602-611. Bunt, S.K., V.K. Clements, E.M. Hanson, P. Sinha, and S. Ostrand-Rosenberg. 2009. Inflammation enhances myeloid-derived suppressor cell cross-talk by signaling through Toll-like receptor 4. J Leukoc Biol 85:996-1004. 158 Carvalho-Tavares, J., M.J. Hickey, J. Hutchison, J. Michaud, I.T. Sutcliffe, and P. Kubes. 2000. A role for platelets and endothelial selectins in tumor necrosis factor-alpha-induced leukocyte recruitment in the brain microvasculature. Circ Res 87:1141-1148. Chalifoux, L.V., and R.T. Bronson. 1981. Colonic adenocarcinoma associated with chronic colitis in cotton top marmosets, Saguinus oedipus. Gastroenterology 80:942-946. Cheng, L., J. Wang, X. Li, Q. Xing, P. Du, L. Su, and S. Wang. 2011. Interleukin6 induces Gr-1+CD11b+ myeloid cells to suppress CD8+ T cell-mediated liver injury in mice. PLoS One 6:e17631. Church, J.M., F.L. Weakley, V.W. Fazio, B.A. Sebek, E. Achkar, and M. Carwell. 1985. The relationship between fistulas in Crohn's disease and associated carcinoma. Report of four cases and review of the literature. Dis Colon Rectum 28:361-366. Collier, P.E., P. Turowski, and D.L. Diamond. 1985. Small intestinal adenocarcinoma complicating regional enteritis. Cancer 55:516-521. Costa, A., A. Scholer-Dahirel, and F. Mechta-Grigoriou. 2014. The role of reactive oxygen species and metabolism on cancer cells and their microenvironment. Semin Cancer Biol Coussens, L.M., and Z. Werb. 2002. Inflammation and cancer. Nature 420:860867. 159 Dalgleish, A.G., Haefner, B., editor. 2006. The Link Between Inflammation and Cancer. Spring US, New York. 256 pp. Danese, S., and A. Mantovani. 2010. Inflammatory bowel disease and intestinal cancer: a paradigm of the Yin-Yang interplay between inflammation and cancer. Oncogene 29:3313-3323. De Marzo, A.M., E.A. Platz, S. Sutcliffe, J. Xu, H. Gronberg, C.G. Drake, Y. Nakai, W.B. Isaacs, and W.G. Nelson. 2007. Inflammation in prostate carcinogenesis. Nat Rev Cancer 7:256-269. De Paepe, B., K.K. Creus, and J.L. De Bleecker. 2009. Role of cytokines and chemokines in idiopathic inflammatory myopathies. Curr Opin Rheumatol 21:610-616. Delaunoit, T., P.J. Limburg, R.M. Goldberg, J.F. Lymp, and E.V. Loftus, Jr. 2006. Colorectal cancer prognosis among patients with inflammatory bowel disease. Clin Gastroenterol Hepatol 4:335-342. Demaria, S., E. Pikarsky, M. Karin, L.M. Coussens, Y.C. Chen, E.M. El-Omar, G. Trinchieri, S.M. Dubinett, J.T. Mao, E. Szabo, A. Krieg, G.J. Weiner, B.A. Fox, G. Coukos, E. Wang, R.T. Abraham, M. Carbone, and M.T. Lotze. 2010. Cancer and inflammation: promise for biologic therapy. J Immunother 33:335-351. Diaz-Montero, C.M., M.L. Salem, M.I. Nishimura, E. Garrett-Mayer, D.J. Cole, and A.J. Montero. 2008. Increased circulating myeloid-derived suppressor cells correlate with clinical cancer stage, metastatic tumor burden, and 160 doxorubicin-cyclophosphamide chemotherapy. Cancer Immunol Immunother Ding, X., Z. Zhao, W. Duan, S. Wang, X. Jin, and L. Xiang. 2013. Expression patterns of CXCR4 in different colon tissue segments of patients with Hirschsprung's disease. Exp Mol Pathol 95:111-116. Dolcetti, L., I. Marigo, B. Mantelli, E. Peranzoni, P. Zanovello, and V. Bronte. 2008. Myeloid-derived suppressor cell role in tumor-related inflammation. Cancer Lett Dube, C. 2012. Tackling colorectal cancer as a public health issue: what can the gastroenterologist do? Can J Gastroenterol 26:417-418. Duffy, A., F. Zhao, L. Haile, J. Gamrekelashvili, S. Fioravanti, C. Ma, T. Kapanadze, K. Compton, W.D. Figg, and T.F. Greten. 2013. Comparative analysis of monocytic and granulocytic myeloid-derived suppressor cell subsets in patients with gastrointestinal malignancies. Cancer Immunol Immunother 62:299-307. Eaden, J.A., K.R. Abrams, and J.F. Mayberry. 2001. The risk of colorectal cancer in ulcerative colitis: a meta-analysis. Gut 48:526-535. Ekbom, A., C. Helmick, M. Zack, and H.O. Adami. 1990. Ulcerative colitis and colorectal cancer. A population-based study. N Engl J Med 323:1228-1233. Engelhardt, K.R., N. Shah, I. Faizura-Yeop, D.F. Kocacik Uygun, N. Frede, A.M. Muise, E. Shteyer, S. Filiz, R. Chee, M. Elawad, B. Hartmann, P.D. Arkwright, C. Dvorak, C. Klein, J.M. Puck, B. Grimbacher, and E.O. 161 Glocker. 2013. Clinical outcome in IL-10- and IL-10 receptor-deficient patients with or without hematopoietic stem cell transplantation. J Allergy Clin Immunol 131:825-830. Erdman, S.E., T. Poutahidis, M. Tomczak, A.B. Rogers, K. Cormier, B. Plank, B.H. Horwitz, and J.G. Fox. 2003. CD4+ CD25+ regulatory T lymphocytes inhibit microbially induced colon cancer in Rag2-deficient mice. Am J Pathol 162:691-702. Fattovich, G., T. Stroffolini, I. Zagni, and F. Donato. 2004. Hepatocellular carcinoma in cirrhosis: incidence and risk factors. Gastroenterology 127:S35-50. Fearon, E.R., and B. Vogelstein. 1990. A genetic model for colorectal tumorigenesis. Cell 61:759-767. Fernandez, L., A. Martinez, J.L. Mendoza, E. Urcelay, M. Fernandez-Arquero, J. Garcia-Paredes, M. Diaz-Rubio, and E.G. de la Concha. 2005. Interleukin10 polymorphisms in Spanish patients with IBD. Inflamm Bowel Dis 11:739-743. Fujii, S., T. Fujimori, H. Kawamata, J. Takeda, K. Kitajima, F. Omotehara, T. Kaihara, T. Kusaka, K. Ichikawa, Y. Ohkura, Y. Ono, J. Imura, S. Yamaoka, C. Sakamoto, Y. Ueda, and T. Chiba. 2004. Development of colonic neoplasia in p53 deficient mice with experimental colitis induced by dextran sulphate sodium. Gut 53:710-716. 162 Fujita, K., C.M. Ewing, R.H. Getzenberg, J.K. Parsons, W.B. Isaacs, and C.P. Pavlovich. 2010. Monocyte chemotactic protein-1 (MCP-1/CCL2) is associated with prostatic growth dysregulation and benign prostatic hyperplasia. Prostate 70:473-481. Fukata, M., and M.T. Abreu. 2007. TLR4 signalling in the intestine in health and disease. Biochem Soc Trans 35:1473-1478. Fukata, M., and M.T. Abreu. 2008. Role of Toll-like receptors in gastrointestinal malignancies. Oncogene 27:234-243. Fukata, M., A. Chen, A.S. Vamadevan, J. Cohen, K. Breglio, S. Krishnareddy, D. Hsu, R. Xu, N. Harpaz, A.J. Dannenberg, K. Subbaramaiah, H.S. Cooper, S.H. Itzkowitz, and M.T. Abreu. 2007. Toll-like receptor-4 promotes the development of colitis-associated colorectal tumors. Gastroenterology 133:1869-1881. Fukata, M., K.S. Michelsen, R. Eri, L.S. Thomas, B. Hu, K. Lukasek, C.C. Nast, J. Lechago, R. Xu, Y. Naiki, A. Soliman, M. Arditi, and M.T. Abreu. 2005. Toll-like receptor-4 is required for intestinal response to epithelial injury and limiting bacterial translocation in a murine model of acute colitis. Am J Physiol Gastrointest Liver Physiol 288:G1055-1065. Fuxe, J., and M.C. Karlsson. 2012. TGF-beta-induced epithelial-mesenchymal transition: a link between cancer and inflammation. Semin Cancer Biol 22:455-461. 163 Gabrilovich, D., T. Ishida, T. Oyama, S. Ran, V. Kravtsov, S. Nadaf, and D.P. Carbone. 1998. Vascular endothelial growth factor inhibits the development of dendritic cells and dramatically affects the differentiation of multiple hematopoietic lineages in vivo. Blood 92:4150-4166. Gabrilovich, D.I., and S. Nagaraj. 2009. Myeloid-derived suppressor cells as regulators of the immune system. Nat Rev Immunol 9:162-174. Glocker, E.O., D. Kotlarz, K. Boztug, E.M. Gertz, A.A. Schaffer, F. Noyan, M. Perro, J. Diestelhorst, A. Allroth, D. Murugan, N. Hatscher, D. Pfeifer, K.W. Sykora, M. Sauer, H. Kreipe, M. Lacher, R. Nustede, C. Woellner, U. Baumann, U. Salzer, S. Koletzko, N. Shah, A.W. Segal, A. Sauerbrey, S. Buderus, S.B. Snapper, B. Grimbacher, and C. Klein. 2009. Inflammatory bowel disease and mutations affecting the interleukin-10 receptor. N Engl J Med 361:2033-2045. Goldszmid, R.S., and G. Trinchieri. 2012. The price of immunity. Nat Immunol 13:932-938. Gonzalez-Navajas, J.M., S. Fine, J. Law, S.K. Datta, K.P. Nguyen, M. Yu, M. Corr, K. Katakura, L. Eckman, J. Lee, and E. Raz. 2010. TLR4 signaling in effector CD4+ T cells regulates TCR activation and experimental colitis in mice. J Clin Invest 120:570-581. Grammas, P., and R. Ovase. 2001. Inflammatory factors are elevated in brain microvessels in Alzheimer's disease. Neurobiol Aging 22:837-842. 164 Greten, F.R., L. Eckmann, T.F. Greten, J.M. Park, Z.W. Li, L.J. Egan, M.F. Kagnoff, and M. Karin. 2004. IKKbeta links inflammation and tumorigenesis in a mouse model of colitis-associated cancer. Cell 118:285-296. Haile, L.A., J. Gamrekelashvili, M.P. Manns, F. Korangy, and T.F. Greten. 2010. CD49d is a new marker for distinct myeloid-derived suppressor cell subpopulations in mice. J Immunol 185:203-210. Haile, L.A., R. von Wasielewski, J. Gamrekelashvili, C. Kruger, O. Bachmann, A.M. Westendorf, J. Buer, R. Liblau, M.P. Manns, F. Korangy, and T.F. Greten. 2008. Myeloid-derived suppressor cells in inflammatory bowel disease: a new immunoregulatory pathway. Gastroenterology 135:871881, 881 e871-875. Hajishengallis, G., R.P. Darveau, and M.A. Curtis. 2012. The keystone-pathogen hypothesis. Nat Rev Microbiol 10:717-725. Hale, L.P., and P.K. Greer. 2012. A novel murine model of inflammatory bowel disease and inflammation-associated colon cancer with ulcerative colitislike features. PLoS One 7:e41797. Hamilton, J.A. 2002. GM-CSF in inflammation and autoimmunity. Trends Immunol 23:403-408. Hayashi, C., G. Papadopoulos, C.V. Gudino, E.O. Weinberg, K.R. Barth, A.G. Madrigal, Y. Chen, H. Ning, M. LaValley, F.C. Gibson, 3rd, J.A. Hamilton, and C.A. Genco. 2012. Protective 165 role for TLR4 signaling in atherosclerosis progression as revealed by infection with a common oral pathogen. J Immunol 189:3681-3688. Hegde, V.L., P.S. Nagarkatti, and M. Nagarkatti. 2011. Role of myeloid-derived suppressor cells in amelioration of experimental autoimmune hepatitis following activation of TRPV1 receptors by cannabidiol. PLoS One 6:e18281. Heit, B., and P. Kubes. 2003. Measuring chemotaxis and chemokinesis: the under-agarose cell migration assay. Sci STKE 2003:PL5. Heit, B., L. Liu, P. Colarusso, K.D. Puri, and P. Kubes. 2008. PI3K accelerates, but is not required for, neutrophil chemotaxis to fMLP. J Cell Sci 121:205214. Huang, B., Z. Lei, J. Zhao, W. Gong, J. Liu, Z. Chen, Y. Liu, D. Li, Y. Yuan, G.M. Zhang, and Z.H. Feng. 2007. CCL2/CCR2 pathway mediates recruitment of myeloid suppressor cells to cancers. Cancer Lett 252:86-92. Hugot, J.P., M. Chamaillard, H. Zouali, S. Lesage, J.P. Cezard, J. Belaiche, S. Almer, C. Tysk, C.A. O'Morain, M. Gassull, V. Binder, Y. Finkel, A. Cortot, R. Modigliani, P. Laurent-Puig, C. Gower-Rousseau, J. Macry, J.F. Colombel, M. Sahbatou, and G. Thomas. 2001. Association of NOD2 leucine-rich repeat variants with susceptibility to Crohn's disease. Nature 411:599-603. 166 Jackson, A.R., V.L. Hegde, P.S. Nagarkatti, and M. Nagarkatti. 2013. Characterization of endocannabinoid-mediated induction of myeloidderived suppressor cells involving mast cells and MCP-1. J Leukoc Biol Jess, T., J. Simonsen, K.T. Jorgensen, B.V. Pedersen, N.M. Nielsen, and M. Frisch. 2012. Decreasing risk of colorectal cancer in patients with inflammatory bowel disease over 30 years. Gastroenterology 143:375-381 e371; quiz e313-374. Kado, S., K. Uchida, H. Funabashi, S. Iwata, Y. Nagata, M. Ando, M. Onoue, Y. Matsuoka, M. Ohwaki, and M. Morotomi. 2001. Intestinal microflora are necessary for development of spontaneous adenocarcinoma of the large intestine in T-cell receptor beta chain and p53 double-knockout mice. Cancer Res 61:2395-2398. Kanneganti, M., M. Mino-Kenudson, and E. Mizoguchi. 2011. Animal models of colitis-associated carcinogenesis. J Biomed Biotechnol 2011:342637. Kawachi, S., S. Jennings, J. Panes, A. Cockrell, F.S. Laroux, L. Gray, M. Perry, H. van der Heyde, E. Balish, D.N. Granger, R.A. Specian, and M.B. Grisham. 2000. Cytokine and endothelial cell adhesion molecule expression in interleukin-10-deficient mice. Am J Physiol Gastrointest Liver Physiol 278:G734-743. Khajah, M., G. Andonegui, R. Chan, A.W. Craig, P.A. Greer, and D.M. McCafferty. 2013. Fer kinase limits neutrophil chemotaxis toward end target chemoattractants. J Immunol 190:2208-2216. 167 Kohno, H., R. Suzuki, S. Sugie, and T. Tanaka. 2005. Beta-Catenin mutations in a mouse model of inflammation-related colon carcinogenesis induced by 1,2-dimethylhydrazine and dextran sodium sulfate. Cancer Sci 96:69-76. Koike, Y., T. Kanai, K. Saeki, Y. Nakamura, M. Nakano, Y. Mikami, Y. Yamagishi, N. Nakamoto, H. Ebinuma, and T. Hibi. 2012. MyD88-dependent interleukin-10 production from regulatory CD11b(+)Gr-1(high) cells suppresses development of acute cerulein pancreatitis in mice. Immunol Lett 148:172-177. Kong, Y.Y., M. Fuchsberger, S.D. Xiang, V. Apostolopoulos, and M. Plebanski. 2013. Myeloid derived suppressor cells and their role in diseases. Curr Med Chem 20:1437-1444. Kotlarz, D., R. Beier, D. Murugan, J. Diestelhorst, O. Jensen, K. Boztug, D. Pfeifer, H. Kreipe, E.D. Pfister, U. Baumann, J. Puchalka, J. Bohne, O. Egritas, B. Dalgic, K.L. Kolho, A. Sauerbrey, S. Buderus, T. Gungor, A. Enninger, Y.K. Koda, G. Guariso, B. Weiss, S. Corbacioglu, P. Socha, N. Uslu, A. Metin, G.T. Wahbeh, K. Husain, D. Ramadan, W. Al-Herz, B. Grimbacher, M. Sauer, K.W. Sykora, S. Koletzko, and C. Klein. 2012. Loss of interleukin-10 signaling and infantile inflammatory bowel disease: implications for diagnosis and therapy. Gastroenterology 143:347-355. Kuhn, R., J. Lohler, D. Rennick, K. Rajewsky, and W. Muller. 1993. Interleukin10-deficient mice develop chronic enterocolitis. Cell 75:263-274. 168 Kumar, V., Abbas, A., Fausto, N., Aster, J., editor. 2010. Robbins and Cotran Pathologic Basis of Disease. Saunders Elsevier, Philadelphia, PA. Kundu, J.K., and Y.J. Surh. 2008. Inflammation: gearing the journey to cancer. Mutation research 659:15-30. Kusmartsev, S., F. Cheng, B. Yu, Y. Nefedova, E. Sotomayor, R. Lush, and D. Gabrilovich. 2003. All-trans-retinoic acid eliminates immature myeloid cells from tumor-bearing mice and improves the effect of vaccination. Cancer Res 63:4441-4449. Kusmartsev, S., and D.I. Gabrilovich. 2003. Inhibition of myeloid cell differentiation in cancer: the role of reactive oxygen species. J Leukoc Biol 74:186-196. Kusmartsev, S., and D.I. Gabrilovich. 2006. Role of immature myeloid cells in mechanisms of immune evasion in cancer. Cancer Immunol Immunother 55:237-245. Kusmartsev, S.A., Y. Li, and S.H. Chen. 2000. Gr-1+ myeloid cells derived from tumor-bearing mice inhibit primary T cell activation induced through CD3/CD28 costimulation. J Immunol 165:779-785. Lapara, N.J., 3rd, and B.L. Kelly. 2010. Suppression of LPS-induced inflammatory responses in macrophages infected with Leishmania. J Inflamm (Lond) 7:8. Le, H.K., L. Graham, E. Cha, J.K. Morales, M.H. Manjili, and H.D. Bear. 2009. Gemcitabine directly inhibits myeloid derived suppressor cells in BALB/c 169 mice bearing 4T1 mammary carcinoma and augments expansion of T cells from tumor-bearing mice. Int Immunopharmacol 9:900-909. Lechner, M.G., D.J. Liebertz, and A.L. Epstein. 2010. Characterization of cytokine-induced myeloid-derived suppressor cells from normal human peripheral blood mononuclear cells. J Immunol 185:2273-2284. Liu, B.Y., I. Soloviev, P. Chang, J. Lee, X. Huang, C. Zhong, N. Ferrara, P. Polakis, and C. Sakanaka. 2010. Stromal cell-derived factor-1/CXCL12 contributes to MMTV-Wnt1 tumor growth involving Gr1+CD11b+ cells. PLoS One 5:e8611. Liu, C., S. Yu, J. Kappes, J. Wang, W.E. Grizzle, K.R. Zinn, and H.G. Zhang. 2007. Expansion of spleen myeloid suppressor cells represses NK cell cytotoxicity in tumor-bearing host. Blood 109:4336-4342. Loftus, E.V., Jr. 2004. Clinical epidemiology of inflammatory bowel disease: Incidence, prevalence, and environmental influences. Gastroenterology 126:1504-1517. Mahler, M., I.J. Bristol, E.H. Leiter, A.E. Workman, E.H. Birkenmeier, C.O. Elson, and J.P. Sundberg. 1998. Differential susceptibility of inbred mouse strains to dextran sulfate sodium-induced colitis. Am J Physiol 274:G544551. Makarenkova, V.P., V. Bansal, B.M. Matta, L.A. Perez, and J.B. Ochoa. 2006. CD11b+/Gr-1+ myeloid suppressor cells cause T cell dysfunction after traumatic stress. J Immunol 176:2085-2094. 170 Mantovani, A., P. Allavena, A. Sica, and F. Balkwill. 2008. Cancer-related inflammation. Nature 454:436-444. McCafferty, D.M., A.W. Craig, Y.A. Senis, and P.A. Greer. 2002. Absence of Fer protein-tyrosine kinase exacerbates leukocyte recruitment in response to endotoxin. J Immunol 168:4930-4935. Medzhitov, R., and C.A. Janeway, Jr. 1999. Innate immune induction of the adaptive immune response. Cold Spring Harb Symp Quant Biol 64:429435. Melani, C., C. Chiodoni, G. Forni, and M.P. Colombo. 2003. Myeloid cell expansion elicited by the progression of spontaneous mammary carcinomas in c-erbB-2 transgenic BALB/c mice suppresses immune reactivity. Blood 102:2138-2145. Mencacci, A., C. Montagnoli, A. Bacci, E. Cenci, L. Pitzurra, A. Spreca, M. Kopf, A.H. Sharpe, and L. Romani. 2002. CD80+Gr-1+ myeloid cells inhibit development of antifungal Th1 immunity in mice with candidiasis. J Immunol 169:3180-3190. Menetrier-Caux, C., G. Montmain, M.C. Dieu, C. Bain, M.C. Favrot, C. Caux, and J.Y. Blay. 1998. Inhibition of the differentiation of dendritic cells from CD34(+) progenitors by tumor cells: role of interleukin-6 and macrophage colony-stimulating factor. Blood 92:4778-4791. Mikami, S., H. Nakase, S. Yamamoto, Y. Takeda, T. Yoshino, K. Kasahara, S. Ueno, N. Uza, S. Oishi, N. Fujii, T. Nagasawa, and T. Chiba. 2008. 171 Blockade of CXCL12/CXCR4 axis ameliorates murine experimental colitis. J Pharmacol Exp Ther 327:383-392. Miranda, K.M., M.G. Espey, and D.A. Wink. 2001. A rapid, simple spectrophotometric method for simultaneous detection of nitrate and nitrite. Nitric Oxide 5:62-71. Molodecky, N.A., I.S. Soon, D.M. Rabi, W.A. Ghali, M. Ferris, G. Chernoff, E.I. Benchimol, R. Panaccione, S. Ghosh, H.W. Barkema, and G.G. Kaplan. 2012. Increasing incidence and prevalence of the inflammatory bowel diseases with time, based on systematic review. Gastroenterology 142:4654 e42; quiz e30. Moore, K.W., R. de Waal Malefyt, R.L. Coffman, and A. O'Garra. 2001. Interleukin-10 and the interleukin-10 receptor. Annu Rev Immunol 19:683765. Moran, C.J., T.D. Walters, C.H. Guo, S. Kugathasan, C. Klein, D. Turner, V.M. Wolters, R.H. Bandsma, M. Mouzaki, M. Zachos, J.C. Langer, E. Cutz, S.M. Benseler, C.M. Roifman, M.S. Silverberg, A.M. Griffiths, S.B. Snapper, and A.M. Muise. 2013. IL-10R polymorphisms are associated with very-early-onset ulcerative colitis. Inflamm Bowel Dis 19:115-123. Nagaraj, S., K. Gupta, V. Pisarev, L. Kinarsky, S. Sherman, L. Kang, D.L. Herber, J. Schneck, and D.I. Gabrilovich. 2007. Altered recognition of antigen is a mechanism of CD8+ T cell tolerance in cancer. Nat Med 13:828-835. 172 Neuman, M.G., and R.M. Nanau. 2012. Single-nucleotide polymorphisms in inflammatory bowel disease. Transl Res 160:45-64. Obermajer, N., and P. Kalinski. 2012. Generation of myeloid-derived suppressor cells using prostaglandin E2. Transplant Res 1:15. Obermajer, N., R. Muthuswamy, J. Lesnock, R.P. Edwards, and P. Kalinski. 2011a. Positive feedback between PGE2 and COX2 redirects the differentiation of human dendritic cells toward stable myeloid-derived suppressor cells. Blood 118:5498-5505. Obermajer, N., R. Muthuswamy, K. Odunsi, R.P. Edwards, and P. Kalinski. 2011b. PGE(2)-induced CXCL12 production and CXCR4 expression controls the accumulation of human MDSCs in ovarian cancer environment. Cancer Res 71:7463-7470. Obermajer, N., J.L. Wong, R.P. Edwards, K. Chen, M. Scott, S. Khader, J.K. Kolls, K. Odunsi, T.R. Billiar, and P. Kalinski. 2013. Induction and stability of human Th17 cells require endogenous NOS2 and cGMP-dependent NO signaling. J Exp Med Obermajer, N., J.L. Wong, R.P. Edwards, K. Odunsi, K. Moysich, and P. Kalinski. 2012. PGE(2)-driven induction and maintenance of cancer-associated myeloid-derived suppressor cells. Immunol Invest 41:635-657. Okayasu, I., M. Yamada, T. Mikami, T. Yoshida, J. Kanno, and T. Ohkusa. 2002. Dysplasia and carcinoma development in a repeated dextran sulfate sodium-induced colitis model. J Gastroenterol Hepatol 17:1078-1083. 173 Ording, A.G., E. Horvath-Puho, R. Erichsen, M.D. Long, J.A. Baron, T.L. Lash, and H.T. Sorensen. 2013. Five-Year Mortality in Colorectal Cancer Patients with Ulcerative Colitis or Crohn's Disease: A Nationwide Population-based Cohort Study. Inflamm Bowel Dis 19:800-805. Ostrand-Rosenberg, S., and P. Sinha. 2009. Myeloid-derived suppressor cells: linking inflammation and cancer. J Immunol 182:4499-4506. Packey, C.D., and R.B. Sartor. 2008. Interplay of commensal and pathogenic bacteria, genetic mutations, and immunoregulatory defects in the pathogenesis of inflammatory bowel diseases. J Intern Med 263:597-606. Parekh, V.V., L. Wu, D. Olivares-Villagomez, K.T. Wilson, and L. Van Kaer. 2013. Activated invariant NKT cells control central nervous system autoimmunity in a mechanism that involves myeloid-derived suppressor cells. J Immunol 190:1948-1960. Parkin, D.M., J. Ferlay, M.P. Curado, F. Bray, B. Edwards, H.R. Shin, and D. Forman. 2010. Fifty years of cancer incidence: CI5 I-IX. Int J Cancer 127:2918-2927. Parsonnet, J., G.D. Friedman, D.P. Vandersteen, Y. Chang, J.H. Vogelman, N. Orentreich, and R.K. Sibley. 1991. Helicobacter pylori infection and the risk of gastric carcinoma. N Engl J Med 325:1127-1131. Patan, S. 2004. Vasculogenesis and angiogenesis. Cancer Treat Res 117:3-32. Patel, S.G., and D.J. Ahnen. 2012. Familial colon cancer syndromes: an update of a rapidly evolving field. Curr Gastroenterol Rep 14:428-438. 174 Peranzoni, E., S. Zilio, I. Marigo, L. Dolcetti, P. Zanovello, S. Mandruzzato, and V. Bronte. 2010. Myeloid-derived suppressor cell heterogeneity and subset definition. Curr Opin Immunol 22:238-244. Perske, C., N. Lahat, S. Sheffy Levin, H. Bitterman, B. Hemmerlein, and M.A. Rahat. 2010. Loss of inducible nitric oxide synthase expression in the mouse renal cell carcinoma cell line RENCA is mediated by microRNA miR-146a. Am J Pathol 177:2046-2054. Pino, M.S., and D.C. Chung. 2010. The chromosomal instability pathway in colon cancer. Gastroenterology 138:2059-2072. Popivanova, B.K., K. Kitamura, Y. Wu, T. Kondo, T. Kagaya, S. Kaneko, M. Oshima, C. Fujii, and N. Mukaida. 2008. Blocking TNF-alpha in mice reduces colorectal carcinogenesis associated with chronic colitis. J Clin Invest 118:560-570. Popivanova, B.K., F.I. Kostadinova, K. Furuichi, M.M. Shamekh, T. Kondo, T. Wada, K. Egashira, and N. Mukaida. 2009. Blockade of a chemokine, CCL2, reduces chronic colitis-associated carcinogenesis in mice. Cancer Res 69:7884-7892. Qi, W., K.V. Ebbert, A.W. Craig, P.A. Greer, and D.M. McCafferty. 2005. Absence of Fer protein tyrosine kinase exacerbates endotoxin induced intestinal epithelial barrier dysfunction in vivo. Gut 54:1091-1097. 175 Raber, P., A.C. Ochoa, and P.C. Rodriguez. 2012. Metabolism of L-arginine by myeloid-derived suppressor cells in cancer: mechanisms of T cell suppression and therapeutic perspectives. Immunol Invest 41:614-634. Rakoff-Nahoum, S., and R. Medzhitov. 2009. Toll-like receptors and cancer. Nat Rev Cancer 9:57-63. Ramachandran, I.R., A. Martner, A. Pisklakova, T. Condamine, T. Chase, T. Vogl, J. Roth, D. Gabrilovich, and Y. Nefedova. 2013. Myeloid-derived suppressor cells regulate growth of multiple myeloma by inhibiting T cells in bone marrow. J Immunol 190:3815-3823. Reinecker, H.C., E.Y. Loh, D.J. Ringler, A. Mehta, J.L. Rombeau, and R.P. MacDermott. 1995. Monocyte-chemoattractant protein 1 gene expression in intestinal epithelial cells and inflammatory bowel disease mucosa. Gastroenterology 108:40-50. Remo, A., M. Pancione, C. Zanella, and R. Vendraminelli. 2012. Molecular pathology of colorectal carcinoma. A systematic review centred on the new role of the pathologist. Pathologica 104:432-441. Rennick, D., N. Davidson, and D. Berg. 1995. Interleukin-10 gene knock-out mice: a model of chronic inflammation. Clin Immunol Immunopathol 76:S174-178. Rhodes, J.M. 1996. Unifying hypothesis for inflammatory bowel disease and associated colon cancer: sticking the pieces together with sugar. Lancet 347:40-44. 176 Rhodes, J.M., and B.J. Campbell. 2002. Inflammation and colorectal cancer: IBD-associated and sporadic cancer compared. Trends Mol Med 8:10-16. Rios-Arrabal, S., F. Artacho-Cordon, J. Leon, E. Roman-Marinetto, M. Del Mar Salinas-Asensio, I. Calvente, and M.I. Nunez. 2013. Involvement of free radicals in breast cancer. SpringerPlus 2:404. Rocchi, A., E.I. Benchimol, C.N. Bernstein, A. Bitton, B. Feagan, R. Panaccione, K.W. Glasgow, A. Fernandes, and S. Ghosh. 2012. Inflammatory bowel disease: a Canadian burden of illness review. Can J Gastroenterol 26:811-817. Rodriguez, P.C., C.P. Hernandez, K. Morrow, R. Sierra, J. Zabaleta, D.D. Wyczechowska, and A.C. Ochoa. 2010. L-arginine deprivation regulates cyclin D3 mRNA stability in human T cells by controlling HuR expression. J Immunol 185:5198-5204. Rodriguez, P.C., D.G. Quiceno, J. Zabaleta, B. Ortiz, A.H. Zea, M.B. Piazuelo, A. Delgado, P. Correa, J. Brayer, E.M. Sotomayor, S. Antonia, J.B. Ochoa, and A.C. Ochoa. 2004. Arginase I production in the tumor microenvironment by mature myeloid cells inhibits T-cell receptor expression and antigen-specific T-cell responses. Cancer Res 64:58395849. Roers, A., L. Siewe, E. Strittmatter, M. Deckert, D. Schluter, W. Stenzel, A.D. Gruber, T. Krieg, K. Rajewsky, and W. Muller. 2004. T cell-specific inactivation of the interleukin 10 gene in mice results in enhanced T cell 177 responses but normal innate responses to lipopolysaccharide or skin irritation. J Exp Med 200:1289-1297. Rossner, S., C. Voigtlander, C. Wiethe, J. Hanig, C. Seifarth, and M.B. Lutz. 2005. Myeloid dendritic cell precursors generated from bone marrow suppress T cell responses via cell contact and nitric oxide production in vitro. Eur J Immunol 35:3533-3544. Rubin, D.T., D. Huo, J.A. Kinnucan, M.S. Sedrak, N.E. McCullom, A.P. Bunnag, E.P. Raun-Royer, R.D. Cohen, S.B. Hanauer, J. Hart, and J.R. Turner. 2013. Inflammation is an independent risk factor for colonic neoplasia in patients with ulcerative colitis: a case-control study. Clin Gastroenterol Hepatol 11:1601-1608 e1601-1604. Rudolph, U., M.J. Finegold, S.S. Rich, G.R. Harriman, Y. Srinivasan, P. Brabet, G. Boulay, A. Bradley, and L. Birnbaumer. 1995. Ulcerative colitis and adenocarcinoma of the colon in G alpha i2-deficient mice. Nat Genet 10:143-150. Ruemmele, F.M., J.F. Beaulieu, S. Dionne, E. Levy, E.G. Seidman, N. CerfBensussan, and M.J. Lentze. 2002. Lipopolysaccharide modulation of normal enterocyte turnover by toll-like receptors is mediated by endogenously produced tumour necrosis factor alpha. Gut 51:842-848. Rutter, M., B. Saunders, K. Wilkinson, S. Rumbles, G. Schofield, M. Kamm, C. Williams, A. Price, I. Talbot, and A. Forbes. 2004. Severity of inflammation 178 is a risk factor for colorectal neoplasia in ulcerative colitis. Gastroenterology 126:451-459. Saito, S., J. Kato, S. Hiraoka, J. Horii, H. Suzuki, R. Higashi, E. Kaji, Y. Kondo, and K. Yamamoto. 2011. DNA methylation of colon mucosa in ulcerative colitis patients: correlation with inflammatory status. Inflamm Bowel Dis 17:1955-1965. Salvadori, S., G. Martinelli, and K. Zier. 2000. Resection of solid tumors reverses T cell defects and restores protective immunity. J Immunol 164:2214-2220. Sander, L.E., S.D. Sackett, U. Dierssen, N. Beraza, R.P. Linke, M. Muller, J.M. Blander, F. Tacke, and C. Trautwein. 2010. Hepatic acute-phase proteins control innate immune responses during infection by promoting myeloidderived suppressor cell function. J Exp Med 207:1453-1464. Sartor, R.B. 2005. Probiotic therapy of intestinal inflammation and infections. Curr Opin Gastroenterol 21:44-50. Sartor, R.B. 2006. Mechanisms of disease: pathogenesis of Crohn's disease and ulcerative colitis. Nat Clin Pract Gastroenterol Hepatol 3:390-407. Sawanobori, Y., S. Ueha, M. Kurachi, T. Shimaoka, J.E. Talmadge, J. Abe, Y. Shono, M. Kitabatake, K. Kakimi, N. Mukaida, and K. Matsushima. 2008. Chemokine-mediated rapid turnover of myeloid-derived suppressor cells in tumor-bearing mice. Blood 111:5457-5466. Schicho, R., M. Bashashati, M. Bawa, D. McHugh, D. Saur, H.M. Hu, A. Zimmer, B. Lutz, K. Mackie, H.B. Bradshaw, D.M. McCafferty, K.A. Sharkey, and M. 179 Storr. 2010. The atypical cannabinoid O-1602 protects against experimental colitis and inhibits neutrophil recruitment. Inflamm Bowel Dis Schimanski, C.C., P.R. Galle, and M. Moehler. 2008. Chemokine receptor CXCR4-prognostic factor for gastrointestinal tumors. World J Gastroenterol 14:4721-4724. Schmidt, K., S. Zilio, J.C. Schmollinger, V. Bronte, T. Blankenstein, and G. Willimsky. 2013. Differently immunogenic cancers in mice induce immature myeloid cells that suppress CTL in vitro but not in vivo following transfer. Blood 121:1740-1748. Schultz, M., S.L. Tonkonogy, R.K. Sellon, C. Veltkamp, V.L. Godfrey, J. Kwon, W.B. Grenther, E. Balish, I. Horak, and R.B. Sartor. 1999. IL-2-deficient mice raised under germfree conditions develop delayed mild focal intestinal inflammation. Am J Physiol 276:G1461-1472. Seamons, A., P.M. Treuting, T. Brabb, and L. Maggio-Price. 2013. Characterization of Dextran Sodium Sulfate-Induced Inflammation and Colonic Tumorigenesis in Smad3 (-/-) Mice with Dysregulated TGFbeta. PLoS One 8:e79182. Serafini, P., K. Meckel, M. Kelso, K. Noonan, J. Califano, W. Koch, L. Dolcetti, V. Bronte, and I. Borrello. 2006. Phosphodiesterase-5 inhibition augments endogenous antitumor immunity by reducing myeloid-derived suppressor cell function. J Exp Med 203:2691-2702. 180 Setia, S., B. Nehru, and S.N. Sanyal. 2013. Activation of NF-kappaB: Bridging the gap between inflammation and cancer in colitis-mediated colon carcinogenesis. Biomedicine & pharmacotherapy = Biomedecine & pharmacotherapie Sevko, A., and V. Umansky. 2013. Myeloid-derived suppressor cells interact with tumors in terms of myelopoiesis, tumorigenesis and immunosuppression: thick as thieves. J Cancer 4:3-11. Shojaei, F., X. Wu, A.K. Malik, C. Zhong, M.E. Baldwin, S. Schanz, G. Fuh, H.P. Gerber, and N. Ferrara. 2007. Tumor refractoriness to anti-VEGF treatment is mediated by CD11b+Gr1+ myeloid cells. Nat Biotechnol 25:911-920. Singh, H., E. Shu, A. Demers, C.N. Bernstein, J. Griffith, and K. Fradette. 2012a. Trends in time to diagnosis of colon cancer and impact on clinical outcomes. Can J Gastroenterol 26:877-880. Singh, U.P., N.P. Singh, B. Singh, L.J. Hofseth, D.D. Taub, R.L. Price, M. Nagarkatti, and P.S. Nagarkatti. 2012b. Role of resveratrol-induced CD11b(+) Gr-1(+) myeloid derived suppressor cells (MDSCs) in the reduction of CXCR3(+) T cells and amelioration of chronic colitis in IL-10(/-) mice. Brain Behav Immun 26:72-82. Sinha, P., V.K. Clements, S.K. Bunt, S.M. Albelda, and S. Ostrand-Rosenberg. 2007a. Cross-talk between myeloid-derived suppressor cells and 181 macrophages subverts tumor immunity toward a type 2 response. J Immunol 179:977-983. Sinha, P., V.K. Clements, A.M. Fulton, and S. Ostrand-Rosenberg. 2007b. Prostaglandin E2 promotes tumor progression by inducing myeloidderived suppressor cells. Cancer Res 67:4507-4513. Sinha, P., V.K. Clements, and S. Ostrand-Rosenberg. 2005. Reduction of myeloid-derived suppressor cells and induction of M1 macrophages facilitate the rejection of established metastatic disease. J Immunol 174:636-645. Srivastava, M.K., L. Zhu, M. Harris-White, U.K. Kar, M. Huang, M.F. Johnson, J.M. Lee, D. Elashoff, R. Strieter, S. Dubinett, and S. Sharma. 2012. Myeloid suppressor cell depletion augments antitumor activity in lung cancer. PLoS One 7:e40677. Strober, W., and I.J. Fuss. 2006. Experimental models of mucosal inflammation. Adv Exp Med Biol 579:55-97. Sun, H.L., X. Zhou, Y.F. Xue, K. Wang, Y.F. Shen, J.J. Mao, H.F. Guo, and Z.N. Miao. 2012. Increased frequency and clinical significance of myeloidderived suppressor cells in human colorectal carcinoma. World J Gastroenterol 18:3303-3309. Suzuki, E., V. Kapoor, A.S. Jassar, L.R. Kaiser, and S.M. Albelda. 2005. Gemcitabine selectively eliminates 182 splenic Gr-1+/CD11b+ myeloid suppressor cells in tumor-bearing animals and enhances antitumor immune activity. Clin Cancer Res 11:6713-6721. Takeda, K., and S. Akira. 2004. TLR signaling pathways. Semin Immunol 16:3-9. Tanaka, T., H. Kohno, R. Suzuki, K. Hata, S. Sugie, N. Niho, K. Sakano, M. Takahashi, and K. Wakabayashi. 2006. Dextran sodium sulfate strongly promotes colorectal carcinogenesis in Apc(Min/+) mice: inflammatory stimuli by dextran sodium sulfate results in development of multiple colonic neoplasms. Int J Cancer 118:25-34. Tanaka, T., H. Kohno, R. Suzuki, Y. Yamada, S. Sugie, and H. Mori. 2003. A novel inflammation-related mouse colon carcinogenesis model induced by azoxymethane and dextran sodium sulfate. Cancer Sci 94:965-973. Terzic, J., S. Grivennikov, E. Karin, and M. Karin. 2010. Inflammation and colon cancer. Gastroenterology 138:2101-2114 e2105. Thackeray, E.W., P. Charatcharoenwitthaya, D. Elfaki, E. Sinakos, and K.D. Lindor. 2011. Colon neoplasms develop early in the course of inflammatory bowel disease and primary sclerosing cholangitis. Clin Gastroenterol Hepatol 9:52-56. The Cancer Genome Atlas Network. 2012. Comprehensive molecular characterization of human colon and rectal cancer. Nature 487:330-337. Tjalsma, H., A. Boleij, J.R. Marchesi, and B.E. Dutilh. 2012. A bacterial driverpassenger model for colorectal cancer: beyond the usual suspects. Nat Rev Microbiol 10:575-582. 183 Tobacman, J.K. 2001. Review of harmful gastrointestinal effects of carrageenan in animal experiments. Environ Health Perspect 109:983-994. Tu, S., G. Bhagat, G. Cui, S. Takaishi, E.A. Kurt-Jones, B. Rickman, K.S. Betz, M. Penz-Oesterreicher, O. Bjorkdahl, J.G. Fox, and T.C. Wang. 2008. Overexpression of interleukin-1beta induces gastric inflammation and cancer and mobilizes myeloid-derived suppressor cells in mice. Cancer Cell 14:408-419. Vincent, J., G. Mignot, F. Chalmin, S. Ladoire, M. Bruchard, A. Chevriaux, F. Martin, L. Apetoh, C. Rebe, and F. Ghiringhelli. 2010. 5-Fluorouracil selectively kills tumor-associated myeloid-derived suppressor cells resulting in enhanced T cell-dependent antitumor immunity. Cancer Res 70:3052-3061. Vogelstein, B., E.R. Fearon, S.R. Hamilton, S.E. Kern, A.C. Preisinger, M. Leppert, Y. Nakamura, R. White, A.M. Smits, and J.L. Bos. 1988. Genetic alterations during colorectal-tumor development. N Engl J Med 319:525532. Watanabe, S., K. Deguchi, R. Zheng, H. Tamai, L.X. Wang, P.A. Cohen, and S. Shu. 2008. Tumor-induced CD11b+Gr-1+ myeloid cells suppress T cell sensitization in tumor-draining lymph nodes. J Immunol 181:3291-3300. Watson, A.J., and P.D. Collins. 2011. Colon cancer: a civilization disorder. Dig Dis 29:222-228. 184 Wirtz, S., C. Neufert, B. Weigmann, and M.F. Neurath. 2007. Chemically induced mouse models of intestinal inflammation. Nat Protoc 2:541-546. Wirtz, S., and M.F. Neurath. 2007. Mouse models of inflammatory bowel disease. Adv Drug Deliv Rev 59:1073-1083. Yanagisawa, K., M.A. Exley, X. Jiang, N. Ohkochi, M. Taniguchi, and K. Seino. 2006. Hyporesponsiveness to natural killer T-cell ligand alpha- galactosylceramide in cancer-bearing state mediated by CD11b+ Gr-1+ cells producing nitric oxide. Cancer Res 66:11441-11446. Yang, L., L.M. DeBusk, K. Fukuda, B. Fingleton, B. Green-Jarvis, Y. Shyr, L.M. Matrisian, D.P. Carbone, and P.C. Lin. 2004. Expansion of myeloid immune suppressor Gr+CD11b+ cells in tumor-bearing host directly promotes tumor angiogenesis. Cancer Cell 6:409-421. Yang, W.C., G. Ma, S.H. Chen, and P.Y. Pan. 2013. Polarization and reprogramming of myeloid-derived suppressor cells. J Mol Cell Biol Yin, B., G. Ma, C.Y. Yen, Z. Zhou, G.X. Wang, C.M. Divino, S. Casares, S.H. Chen, W.C. Yang, and P.Y. Pan. 2010. Myeloid-derived suppressor cells prevent type 1 diabetes in murine models. J Immunol 185:5828-5834. Yoshimura, T., and M. Takahashi. 2007. IFN-gamma-mediated survival enables human neutrophils to produce MCP-1/CCL2 in response to activation by TLR ligands. J Immunol 179:1942-1949. Zampell, J.C., S. Elhadad, T. Avraham, E. Weitman, S. Aschen, A. Yan, and B.J. Mehrara. 2012. Toll-like receptor deficiency worsens inflammation and 185 lymphedema after lymphatic injury. American journal of physiology. Cell physiology 302:C709-719. Zantl, R., and E. Horn. 2011. Chemotaxis of slow migrating mammalian cells analysed by video microscopy. Methods Mol Biol 769:191-203. Zhang, B., Z. Wang, L. Wu, M. Zhang, W. Li, J. Ding, J. Zhu, H. Wei, and K. Zhao. 2013. Circulating and tumor-infiltrating myeloid-derived suppressor cells in patients with colorectal carcinoma. PLoS One 8:e57114. Zhang, R., S. Ito, N. Nishio, Z. Cheng, H. Suzuki, and K.I. Isobe. 2011. Dextran sulphate sodium increases splenic Gr1(+)CD11b(+) cells which accelerate recovery from colitis following intravenous transplantation. Clin Exp Immunol 164:417-427. Zhang, R., Y. Li, P.L. Beck, and D.M. McCafferty. 2007a. Toll-like receptor 4 regulates colitis-associated adenocarcinoma development in interleukin10-deficient (IL-10(-/-)) mice. Biochem Soc Trans 35:1375-1376. Zhang, R., A. Ma, S.J. Urbanski, and D.M. McCafferty. 2007b. Induction of inducible nitric oxide synthase: a protective mechanism in colitis-induced adenocarcinoma. Carcinogenesis 28:1122-1130. Zhu, H., X. Lei, Q. Liu, and Y. Wang. 2013. Interleukin-10-1082A/G polymorphism and inflammatory bowel disease susceptibility: a metaanalysis based on 17,585 subjects. Cytokine 61:146-153. 186