AUTOIMMUNITY AND GENETICS: THE ROLE OF ERAP1 IN THE PATHOGENESIS OF ANKYLOSING SPONDYLITIS AND MODULATION OF IMMUNE RESPONSES AS NOVEL THERAPIES. By Yuliya Pepelyayeva A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Microbiology and Molecular Genetics - Doctor of Philosophy 2018 ABSTRACT AUTOIMMUNITY AND GENETICS: THE ROLE OF ERAP1 IN THE PATHOGENESIS OF ANKYLOSING SPONDYLITIS AND MODULATION OF IMMUNE RESPONSES AS NOVEL THERAPIES. By Yuliya Pepelyayeva Autoimmune diseases are the second most common cause of chronic illness in the US. ERAP1 gene polymorphisms have been linked to multiple autoimmune and inflammatory diseases, including ankylosing spondylitis (AS), multiple sclerosis, inflammatory bowel disease, insulin-dependent diabetes mellitus, Bechet’s disease, and others. Understanding how ERAP1 functions contribute to the pathogenesis of autoimmunity will allow us to better understand mechanisms responsible for autoimmunity and will provide new targets for drug development. AS is a prototypic chronic arthritic disease predominantly affecting young individuals, with spinal inflammation and fusions as its key features. In this work we evaluated the intestinal, skeletal and immune systems of ERAP1-/- mice and discovered that they resemble key features of AS. We determined that these mice developed axial spinal ankylosis, sacroiliac erosions, inflammation at the intervertebral discs and systemic osteoporosis. Detailed analysis of the bone remodeling cells showed increased osteoclastogenesis and osteoclast activity, as well as reduced bone formation in vivo. Additionally, ERAP1-/- mice developed spontaneous dysbiosis and had increased susceptibility to chemically induced colitis, closely paralleling intestinal manifestations of AS. Correction of colonic dysbiosis in the ERAP1-/- mice from birth failed to reduce spinal fusions or the osteoporosis of the spine, suggesting that altered microbiome is the result of the disease, rather than its cause. Finally, immunologic analysis of the ERAP1-/- mice revealed a deficiency of type 1 regulatory T (Tr1) cells and tolerogenic dendritic cells, which are important for the development of Tr1 cells. Tr1 cells are thought to be important for tolerance and prevention of autoimmunity. It is possible that Tr1 cell deficiency is responsible for the multiple phenotypes present in the ERAP1-/- mice and ERAP1-associated diseases, including AS. Anti-TNF-α and anti-IL-17A blocking antibodies have been successful in providing symptomatic relief to AS patients. In our work, TNF-α and anti-IL-17A blockade failed to improve osteoporosis and ankylosis in ERAP1-/- mice, downplaying their involvement in the pathogenesis of skeletal abnormalities in these animals. While in autoimmune diseases, the therapeutic strategies involve reduction of the immune responses, there are situations where immune responses need to be strengthened via targeted therapies. For example, strategies targeting immune cell inhibitory and/or co-stimulatory molecules have been successfully used to treat multiple cancer types in humans. There is a need to develop novel cancer immunotherapies. The immune-cell receptor, CD2-like receptor activating cytotoxic cell (CRACC, also known as SLAMF7), is a member of the signaling lymphocytic activation molecules (SLAM) family of receptors which plays a critical role in immunoregulation. In this work, we targeted SLAMF7 signaling by using CRACC-Fc fusion protein expressed on an adenovirus platform (rAd5-mCRACC-Fc) for use as an immunomodulating agent against established CT26 colon adenocarcinoma tumors. We observed enhanced activation of innate and adaptive immunity, increased infiltration of tumors by lymphocytes, and improved tumor killing and survival of CT26 tumor-bearing mice. These data suggest that rAd5-mCRACC-Fc is a promising novel cancer immunotherapy that warrants further testing for clinical application. In summary, the work in this thesis touches upon multiple aspects of genetics and immunology and how the two fields intermingle on a variety of important clinical fronts. This thesis is dedicated to my best friend and partner Luul. Thank you for always believing in me and supporting me. iv ACKNOWLEDGMENTS I would like to sincerely thank my mentor Dr. Andrea Amalfitano for his guidance, understanding, patience, and mentorship during my graduate studies at Michigan State University. I have been fortunate to have an advisor who gave me the freedom to be independent and encouraged me to grow as a researcher. Dr. Amalfitano supported and encouraged me to develop scientific thinking, leadership and writing skills. Dr. Amalfitano has served as a role model physician, scientist and leader throughout my studies at Michigan State University, for which I am very grateful. I would like to thank Dr. Amalfitano for supporting and supervising the ERAP1 Ankylosing Spondylitis project, which contributed the most to my personal and scientific growth. I greatly appreciate the help provided by my guidance committee: Dr. John Fyfe, Dr. Laura McCabe, and Dr. Margaret Petroff. Their insightful comments and constructive discussions have significantly improved my research progress and helped me improve my scientific and critical thinking. I really appreciate their flexibility and patience. I would like to especially thank Dr. Laura McCabe for her mentorship and guidance in my learning of the osteobiology during her personal time and during journal clubs during which she challenged me to grow as a scientist. I would especially like to thank Dr. McCabe for her guidance and contribution in the development and writing of the manuscripts which describe the role of ERAP1 in osteobiology. I am so grateful for your time and guidance. I would also like to sincerely thank Dr. Yasser Aldhamen for his guidance and mentorship throughout the entirety of my studies at Michigan State University. He challenged v me to grow as a scientist. Thank you for your continuing support and guidance, and for giving me the opportunities to be involved in your projects and for challenging me to grow. I would like to extend special thanks to the D.O.-Ph.D. program. I sincerely thank Dr. Justin McCormick for believing in me and supporting me throughout the length of the program. Thank you for being my advocate and for finding financial resources for me to be able to afford the tuition. I would also like to sincerely thank Bethany Heinlein for being caring and helpful in every crisis I have encountered during the program no matter how big or small they were. You are truly an advocate for all the students in the D.O.-Ph.D. program and I am so thankful for all of your help and thoughtfulness. I would like to sincerely thank Dr. Brian Schutte for his mentorship, advice, and support during his appointment as director of the D.O.-Ph.D. program. Thank you to the Michigan Osteopathic College Foundation, Phyllis K. and Walter P. Dell Endowed, and Sills scholarships for providing me with financial support throughout my enrollment in the D.O.-Ph.D. program. I could not have been able to afford the cost of tuition without them. I would also like to thank my classmates and lab-mates, Maja Blake and Patrick O’Connell. I am deeply grateful for their help with big experiments and for their input and criticisms during the writing of my manuscripts. My sincere thanks to all current and former members of Dr. Amalfitano lab: Dr. Fadel Alaquoub, Dr. Charles Aylsworth, Sarah Godbehere Roosa, Cristiane Pereira-Hicks, Dr. Dionisia Quiroga, and Dr. David Rastall. All of them are good team players, and I am lucky that I had the opportunity to work with them. I would like to thank our current and former undergrads: Kristen Kenny, Doug Peters, Heather Newman, Ashley Raedy, Sean Hyslop and Ariana Angarita, who vi were a pleasure to teach. It was a pleasure to watch them grow and improve in their knowledge, maturity and critical thinking. My special thanks to Sarah Godbehere Roosa, who is a wonderful technician and an extremely pleasant person to work with. Thank you for always having my back, maintaining our huge mouse colony, and taking care of all the ordering, even if it was last minute. Thank you for your willingness to proof-read everything I’ve written during my graduate studies. Thank you for your support and friendship. Special thanks to all MSU core facilities, including animal care program and, in particular, all employees of Biochemistry animal housing facility; MSU Histopathology labs (Amy and Cathy); MSU Research Technology Support Facility: Flow Cytometry (Dr. Louis King) and Genomic Services for RT-PCR. Without their assistance, our research would not have been possible. I’m thankful to all faculty members and employees of the Department of Microbiology and Molecular Genetics, especially the Chair of the department, Dr. Victor DiRita and the Director of MMG Graduate program, Dr. Donna Kozlowsky for their kindness and full support provided during these years. Most importantly, none of this would have been possible without the love and patience of my family. My family has been a constant source of love, support, and strength all these years. I would like to express my heartfelt gratitude to my American family (especially April, Catlin, and Robert), who took me in back when I was only 15-years-old and treated me like their own daughter/sister from the moment we’ve met. I could not have achieved all the things that make me who I am today if it wasn’t for your love and generosity. My Ukrainian family (mom, Mark, and Lyuda) has encouraged me throughout this endeavor despite the fact that we are separated by vii thousands of miles. I always feel your love and support and I know that you are my biggest supporters. Your support and belief in me give me confidence and motivation every day to be the best version of myself and to keep going toward my dreams. And of course, I would like to sincerely thank my beloved partner and best friend Luul, who was there for me every step of my long journey. You encouraged me to apply to the D.O.-Ph.D. program and continued to support me throughout its duration. You challenge and push me the most, which helps me grow. You are always there for me, believing in me, encouraging me and reminding me that I can do anything. I am so grateful for the love, support, and understanding of my family. viii TABLE OF CONTENTS LIST OF TABLES ........................................................................................................................ xii LIST OF FIGURES ..................................................................................................................... xiii KEY TO ABBREVIATIONS ........................................................................................................xv KEY TO SYMBOLS ................................................................................................................. xxiv Chapter 1: Introduction ....................................................................................................................1 Autoimmunity ......................................................................................................................2 Innate immunity ..................................................................................................................3 Adaptive immunity .............................................................................................................4 T lymphocytes ..........................................................................................................5 CD4+ T cells ............................................................................................................6 CD8+ T cells ............................................................................................................8 B lymphocytes .........................................................................................................8 Tolerance............................................................................................................................10 Foxp3+ regulatory T cells ......................................................................................11 Type 1 regulatory T cells ..................................................................................................13 Ankylosing spondylitis .....................................................................................................16 Genes associated with AS and their role in its pathogenesis ............................................17 HLA-B27 ...............................................................................................................17 ERAP1 ...................................................................................................................19 ERAP1 in adaptive immunity ................................................................................20 ERAP1 innate immunity ........................................................................................21 Other genes associated with AS .............................................................................24 Skeletal features of AS.......................................................................................................25 Bone Remodeling ..............................................................................................................25 Gut inflammation and microbiota in AS ............................................................................28 AS therapies .......................................................................................................................29 Animal models of AS.........................................................................................................31 Harnessing auto-immunity: tumor immunotherapy ...........................................................34 SLAM receptors .................................................................................................................36 Use of adenoviruses as vectors ..........................................................................................38 Chapter 2: ERAP 1 deficient mice have reduced type 1 regulatory T cells and develop skeletal and intestinal features of ankylosing spondylitis ..........................................................................41 Introduction ........................................................................................................................42 Results ...............................................................................................................................44 ERAP1 deficiency causes spinal ankylosis and osteoporosis ...............................44 Deficiency of ERAP1 results in recruitment of immune cells to spinal joints .....48 ERAP1 deficient mice have increased susceptibility to chemically-induced colitis ......................................................................................................................51 ERAP1 deficient mice develop spontaneous gut dysbiosis ...................................54 ix ERAP1 deficient mice have reduced numbers of suppressor “Tr1-like” cells ......60 ERAP1-/- mice have reduced numbers of tolerogenic dendritic cells ....................63 Discussion ..........................................................................................................................65 Chapter 3: ERAP 1 deficient mice develop TNF-α independent osteoporosis, enhanced osteoclastogenesis and osteoclast activity ....................................................................................70 Introduction ........................................................................................................................71 Results ...............................................................................................................................73 ERAP1 deficiency causes systemic osteoporosis ..................................................73 ERAP1 deficient mice exhibit increased osteoclastic potential and osteoclast activity ...................................................................................................................75 ERAP1 deficient mice have reduced osteoblast activity in vivo ...........................77 Ankylosis and osteoporosis developed in ERAP1-/- mice were not improved after anti-TNF-α treatments ...........................................................................................79 Anti-TNF-α antibody therapy resulted in enhanced expression of pro-osteoblastic genes in ERAP1-/- mice ..........................................................................................83 Anti-TNF-α antibody therapy resulted in weight reduction in ERAP1-/- mice and negatively affected bone health ............................................................................86 Expression of human ERAP1 allele improved mean ankylosis score and BMC caused by global ERAP1 deficiency ......................................................................88 Discussion ..........................................................................................................................90 Chapter 4: Development of a novel adenovirus-based cancer immunotherapy that targets the CRACC pathway and suppresses established tumors ....................................................................95 Introduction ........................................................................................................................96 Results ...............................................................................................................................99 rAd5-mCRACC-Fc vector successfully infects cells and expresses mCRACC-Fc transgene .........................................................................................99 Overexpression of mCRACC-Fc results in enhanced activation of NK cells and increased expression of type I interferons ...........................................................102 Overexpression of mCRACC-Fc enhances dendritic cell maturation and activation ......................................................................................................103 Overexpression of mCRACC-Fc induces Th-1-skewed pro-inflammatory cytokine responses and STAT1 phosphorylation ..............................................................105 rAd5-mCRACC-Fc vaccination enhances adaptive immune responses to the co- administered antigens ..........................................................................................108 Intratumoral administration of rAd5-mCRACC-Fc vector reduces tumor growth and increases the survival of CT26 tumor-bearing mice ....................................112 Adenovirus and tumor lysate pre-vaccinations allow for reduced tumor growth and increased survival upon CT26 tumor challenge ...........................................116 rAd5-mCRACC-Fc injected tumors have increased CD8+ T and NK cell infiltration ............................................................................................................120 Discussion ........................................................................................................................122 Chapter 5: Materials and methods ..............................................................................................127 Animal procedures ..........................................................................................................128 x µCT imaging and ankylosis scoring ...............................................................................128 AS index ..........................................................................................................................129 Erosion scoring ...............................................................................................................129 DSS-induced colitis experiments ....................................................................................130 Microbial community analysis ........................................................................................130 Histology .........................................................................................................................131 Routine hematoxylin & eosin Stain ................................................................................132 Immunohistochemistry of the spines – IL-23 F4/80, IgM, TNF-α primary antibodies ........................................................................................................................132 Isolation of lymphocytes from spleen and lymph nodes ................................................134 Flow cytometry ...............................................................................................................135 Tr1 differentiation assay .................................................................................................136 Cross-fostering experiments ...........................................................................................137 Quantitative RT-PCR analysis ........................................................................................137 Osteoclast outgrowths .....................................................................................................139 Osteoclast pitting assay ...................................................................................................140 Osteoblast outgrowths .....................................................................................................140 Serum measurements ......................................................................................................141 Dynamic histomorphometry ...........................................................................................141 Adenovirus vector construction ......................................................................................142 Tumor challenge ..............................................................................................................143 Cell culture .......................................................................................................................143 Vaccination studies ..........................................................................................................144 Innate immune study .......................................................................................................145 Cytokine and chemokine analysis ...................................................................................145 ELISA analysis ...............................................................................................................145 ELISPOT analysis ...........................................................................................................146 Western blot ....................................................................................................................146 Immunohistochemistry analysis of tumors .....................................................................147 Statistical analysis ...........................................................................................................149 Chapter 6: Concluding remarks and future directions ................................................................150 BIBLIOGRAPHY ........................................................................................................................159 xi LIST OF TABLES Table 1: Summary of the bacterial genera that were significantly different between WT and ERAP1-/- mice.…………………………...…………………………………………………..…..56 Table 2: Details of antibodies used for immunohistochemistry staining of the spines…..….…134 Table 3: Specifications of antibodies used for immunohistochemistry of tumors……..............148 xii LIST OF FIGURES Figure 1: µCT analysis of the axial skeleton……………………………………….……….…..45 Figure 2: µCT analysis of the sacral region.………………………..…………………………...47 Figure 3: Histopathologic evaluation of ankylosed intervertebral joints.…………………..…...49 Figure 4: Induction of chemically-induced colitis.………………………………………..…….53 Figure 5: Evaluation of gut microbial composition and cross-foster experiments………..…….55 Figure 6: Relative abundance of phyla in WT, CF WT, ERAP1-/- and CF ERAP1-/- fecal samples.…………….……...…………...……...…………...……...…………..............................58 Figure 7: Histopathologic analysis of the ankylosed intervertebral joints from CF ERAP1-/- mice.……………..……...…………...……...…………...……...……………………………..…58 Figure 8: Evaluation of splenic Tr1 cells.…………….................................................................62 Figure 9: Evaluation of tolerogenic dendritic cells (tDCs)……………………………..……….64 Figure 10: ERAP1 deficiency results in systemic osteoporosis…………...……..…….…….....74 Figure 11: Deficiency of ERAP1 mediates increased osteoclastogenesis and osteoclast activity in vitro.……………..…...…………...…………...………...…………...…………………..…....76 Figure 12: Deficiency of ERAP1 results in reduced osteoblast activity in vivo.…………..…....78 Figure 13: Ankylosis and osteoporosis due to ERAP1 deficiency is TNF-α independent….......81 Figure 14: Ankylosis and osteoporosis due to ERAP1 deficiency is IL-17A independent..........82 Figure 15: Anti-TNFα treatment induced pro-osteoblastic gene expression in ERAP1-/- mice…………………………………………………………………………………………..…..85 Figure 16: Effect of the anti-TNF-α blockade on trabecular bone in ERAP1-/- mice……..…….87 Figure 17: Protective human ERAP1 allele improves bone fusions and BMC caused by ERAP1 deficiency…………...…………...…………...…………...…………...…………………..….….89 Figure 18: Effect of mCRACC-Fc overexpression on NK cell activity and function……...….100 xiii Figure 19: Flow cytometry analysis of dendritic cells and macrophage activation in response to mCRACC-Fc overexpression…...…………...…………...…………...…………...……….…..104 Figure 20: Cytokine and chemokine responses in rAd5-mCRACC-Fc treated mice………….107 Figure 21: Adaptive immune cell activation in rAd5-mCRACC-Fc treated mice...…………..110 Figure 22: Expression of mCRACC-Fc in CT26 cells…….……...………...............................114 Figure 23: Effect of rAd5-mCRACC-Fc intratumoral injections on anti-tumor responses...…115 Figure 24: Efficacy of tumor lysate in combination with rAd5-mCRACC-Fc as an anti-tumor vaccine…...…………...…………...…………...…………...………………...………………...118 Figure 25: CD3, CD8 and DX5 immunohistochemical analysis of tumors…………...………121 xiv KEY TO ABBREVIATIONS Ad – Adenovirus ADCC – Antibody-dependent cellular cytotoxicity AF – Annulus fibrosus AhR – Aryl hydrocarbon receptor AIM2 – Absent in melanoma 2 ANOVA – Analysis of variance APC – Antigen presenting ell APRIL – A proliferation inducing ligand AS – Ankylosing spondylitis ASAS – Assessment of SpondyloArthritis International Society BCR – B cell receptor BFR – Bone formation rate BM – Bone marrow BMC – Bone mineral content BMD – Bone mineral density BMP – Bone morphogenic protein BV/TV – Bone volume fraction C4BP – C4b binding protein CatK – Cathepsin K CCR7 – C-C chemokine receptor 7 CD – Cluster of differentiation CF – Cross-fostered xv CFSE – Carboxyfluorescein succinimidyl ester cGAS – Cyclic GMP–AMP (cGAMP) synthase CLR – C-type lectin receptors CMV – Cytomegalovirus Col1α1 – Collagen alpha-1(XI) chain CRACC – CD2-like receptor activating cytotoxic cell CTLA4 – Cytotoxic T-lymphocyte associated protein 4 CTLs – Cytotoxic lymphocyte CTR – Calcitonin receptor CXCL – C-X-C chemokine ligand CXCL – C-X-C motive chemokine CXCR – C-X-C motif chemokine receptor DAI – Disease activity index DAMP – Damage-associated molecular patterns DC – Dendritic cells DC-STAMP – Dendrocyte expressed seven transmembrane protein DKK-1 – Dickkopf-related protein 1 DMP1 – Dentin matrix acidic phosphoprotein 1 DNA – Deoxyribonucleic acid Dsg3 – Desmoglein 3 DSS – Dextran sodium sulfate EAT-2 – Ewing’s sarcoma-associated transcript 2 EBI2 – Epstein-Barr virus-induced receptor 2 xvi ELISA – Enzyme-linked immunosorbent assay ELISPOT – Enzyme-linked immunospot EOR – ER overload response ER – Endoplasmic reticulum ERAP1 – Endoplasmic reticulum aminopeptidase 1 ERT – EAT-2-related transducer FBS – Fetal bovine serum For – Forward Foxp3 – Forkhead box P3 GAPDH – Glyceraldehyde 3-phosphate dehydrogenase GC – Germinal center GFP – Green fluorescent protein GM-CSF – Granulocyte-macrophage colony-stimulating factor Gp120 – Glycoprotein 120 GSK3β – Glycoprotein synthase kinase-3 beta GVHD – Graft-versus-host disease GWAS – Genome-wide association studies H&E – Hematoxylin and eosin HDAC – Histone deacetylases HEK293 – Human embryonic kidney cells 293 HIV – Human immunodeficiency virus HLA – Human leukocyte antigen HPRT – Hypoxanthine-guanine phosphoribosyltransferase xvii HSCT – Hematopoietic stem cell transplantation I.P. – Intraperitoneal I.T. – Intratumoral I.V. – Intravenous IBD – Inflammatory bowel disease ICOS – Inducible T-cell costimulator IDDM – Insulin dependent diabetes mellitus IFN – Interferon Ig – Immunoglobulin IGF-1 – Insulin-like growth factor-1 IL – Interleukin IL-1RII – Type II IL-1 decoy receptor IL-23R – IL-23 receptor IL-6Rα – Interleukin receptor 6 alpha ILCs – Innate lymphoid cells ILT2 – Ig-like transcript 2 IPEX – Immune dysregulation, polyendocrinopathy, enteropathy, X-linked syndrome Irf7 – Interferon regulatory factor 7 ISG – Interferon-stimulated genes ITAM – Immunoreceptor tyrosine-based activation motif ITSMs – Immunoreceptor tyrosine-based switch motifs IVD – Intervertebral disc JAK2 – Janus kinase 2 xviii KC – Keratinocyte chemoattractant KIR – Killer-cell immunoglobulin-like receptors L6 – Lumbar vertebra 6 Lag3 – Lymphocyte-activation gene 3 Lef1 – Lymphoid enhancer-binding factor 1 LIR – Leukocyte Ig-like receptor LPR – lipoprotein receptor LPS – lipopolysaccharide MAR – Mineral apposition rate M-CSF – Macrophage-colony stimulating factor MDSCs – Myeloid-derived suppressor cells MFI – Mean fluorescent intensity mg – Milligram MHC – Major histocompability complex MIP-1α – Macrophage inflammatory protein-1 alpha mM – Milli-molar MRI – Magnetic resonance imaging mRNA – Messenger RNA MS – Multiple sclerosis MyD88 – Myeloid differentiation primary response 88 NFAT – Nuclear factor of activated T-cells NFATc1 – Nuclear factor of activated T-cells, cytoplasmic 1 NF-κB – Nuclear factor kappa B xix NK – Natural killer NKT – Natural killer T NLR – (Nod)-, leucine-rich repeat–containing receptors NOD – Nucleotide-binding oligomerization domain NP – Nucleus pulposus NS – Not significant NSAIDs – Nonsteroidal anti-inflammatory drugs OAS – Oligoadenylate synthase OPG – Osteoprotegerin OSCAR – Osteoclast associated receptor Osx – Osterix PAMPs – Pathogen associated molecular patterns PBMCs – Peripheral blood mononuclear cell PBS – Phosphate-buffered saline PD-1 – Programmed cell death protein 1 PD-L1 – Programmed death-ligand 1 PRRs – Pattern recognition receptors PSF – Penicillin streptomycin fungizone Qa-2 – Qa lymphocyte antigen 2 RA – Rheumatoid arthritis rAd5 – Recombinant adenovirus 5 RANK – Receptor activator of NF-κB RANKL – Receptor activator of NF-κB ligand xx RANTES –Regulated on activation, normal T cell expressed and secreted Rev – Reverse RLR – RIG-I like receptors RNA – ribonucleic acid RorγT – retinoic acid receptor-related orphan receptor gamma RT – Room temperature RT-PCR – Reverse transcription polymerase chain reaction Runx – Runt-related transcription factor S.Q. – Subcutaneous S1 – Sacral vertebra 1 SAP – SLAM-associated proteins SEM – Standard error of the mean SHP-1 – Src homology region 2 domain-containing phosphatase-1 SI – Sacroiliac SLAM – Signaling lymphocytic activation molecule SLAMF – Signaling lymphocytic activation molecule family of receptors SLE – Systemic lupus erythematosus SNP – Single nucleotide polymorphism Socs3 – Suppressor of cytokine signaling 3 SOST – Sclerostin SPF – Specific pathogen free STAT3 – Signal transducer and activator of transcription 3 TAA – Tumor-associated antigens xxi TAP – Transporter associated with antigen processing Tb.N – Trabecular number Tb.Sp – Trabecular spacing Tb.Th – Trabecular thickness TCR – T cell receptor tDCs – Tolerogenic dendritic cells Tfh – T follicular helper TGF – Transformation growth factor Th – T helper Tim-3 – T-cell immunoglobulin and mucin-domain containing-3 TLR – Toll-like receptors TME – Tumor microenvironment TNF-R1 – TNF receptor 1 TNF-α – Tumor necrosis factor alpha Tr1 – Type 1 regulatory cells TRAF6 – TNF receptor associated factor 6 TRAP – Tartate-resistant acid phosphatase Treg – Regulatory T cell TREM2 – Triggering receptor expressed in myeloid cells-2 TYK2 – Tyrosine kinase 2 UPR – Unfolded protein response v.p. – Viral particles WT – Wild type xxii Zap-70 – Zeta-chain-associated protein kinase 70 β2m – β2microglobulin µg – Microgram μCT – Micro computed tomography µL – Microliter µM – Micromolar xxiii α Alpha β Beta γ Gamma µ Micro º Degree ± Plus / Minus κ Kappa KEY TO SYMBOLS xxiv Chapter 1: Introduction 1 Autoimmunity Autoimmune diseases pose a significant burden on health care costs due to their chronic nature, a fact exacerbated by their associated cost and prevalence among younger individuals [1]. An estimated 50 million Americans are affected by autoimmune diseases, with an approximately 100 billion dollar associated healthcare cost. While autoimmune diseases greatly vary depending on the organs which they affect and how they present clinically, all autoimmune diseases progress through three sequential phases: initiation, propagation, and resolution [2]. The initiation phase is asymptomatic. In this phase, the autoimmunity is initiated due to a combination of genetic predispositions and environmental triggers. During the propagation phase, the unopposed inflammation causes tissue damage and the patients develop clinical symptoms. In the resolution phase, inflammation is resolved, however, patients often relapse. While there are some therapies available for many autoimmune diseases, these therapies have a major flaw of blocking the steps of the terminal (resolution) phase of the autoimmune disease, rather than targeting the key problem or the cause of the disease. Due to this flaw, such therapies require life-long adherence to the drugs, which is not only costly but also predisposes patients to the side-effects of these non-specific therapies being given over an extended period of time. Specifically, increased risk of malignancy, viral infections, and reactivation of latent diseases are a concern for patients who use biologic therapeutics [3]. It is important to understand the pathogenesis of the autoimmune disease in order to specifically target the disease at its source and eliminate it with a permanent cure. In general, the autoimmune diseases are thought to occur due to a failure of regulatory mechanisms, where the balance between effector and regulatory responses is lost [1, 2]. It is thought that therapies which can re-establish the equilibrium between the regulatory and effector functions, could provide a path to a cure for autoimmunity. 2 Innate immunity The immune system is divided into two main arms: the innate and the adaptive. The innate immune system is the first line of defense against pathogens and is comprised of multiple components such as anatomical barriers, the complement system, immune cells and other components such as antimicrobial peptides and others [4]. Anatomical barriers such as skin, epithelial cells lining gastrointestinal, respiratory and genitourinary tract surfaces and mucus provide a physical barrier protective against pathogen invasion, mechanical and chemical trauma. Defensins are a family of antimicrobial peptides that are widely expressed in epithelial cell and leukocytes also play an important role in the first line of defense by killing pathogens via membrane disruption [5]. Complement activation involves a reaction cascade that ultimately results in the pathogen cellular lysis, recruitment of inflammatory cells and an induction of phagocytosis. Innate immune cells include neutrophils, basophils, mast cells, macrophages, dendritic cells (DCs), natural killer (NK) cells, natural killer T (NKT) cells [6], as well as more recently defined players including myeloid-derived suppressor cells (MDSCs), and innate lymphoid cells (ILCs) [7]. The innate system mainly relies on recognition of “pathogen associated molecular patterns” (PAMPs) [6]. PAMPS are a variety of conserved molecular structures, which are only present on microbes and are relatively conserved between pathogens, with examples such as lipopolysaccharide (LPS), double stranded RNA, flagellin and many others. Upon recognition of PAMPs by the host’s pattern recognition receptors (PRRs), a signaling cascade is triggered, which results in expression of a variety of cytokines, chemokines, adhesion molecules, and immune receptors, which allow for a sufficient response against infection, and activation of the adaptive responses [6, 8-10]. PRRs can also recognize “damage-associated molecular patterns” 3 (DAMPs), which are endogenous signals released specifically during host cell stress or death [11]. PRRs include Toll-like receptors (TLRs), nucleotide-binding oligomerization domain (Nod) receptors, leucine-rich repeat-containing receptors (NLRs), RIG-I like receptors (RLRs), C-type lectin receptors (CLRs) and absent in melanoma 2 (AIM2) like receptors, as well as intracellular sensors of nucleic acids, including oligoadenylate synthase (OAS) proteins and cyclic GMP–AMP synthase (cGAS) [8-10]. Each PRR recognizes a particular PAMP or DAMP, triggering a particular signaling cascade. For example, LPS which is present only on the surface of gram-negative bacteria is recognized by TLR4 receptor, which upon activation, recruits MyD88 adaptor protein, and triggers downstream responses ultimately resulting in NF-kB transcription factor activation and proinflammatory cytokine gene expression [8, 12]. Although innate immune responses are non-specific, they are very important, because they develop very rapidly, in a matter of hours, allowing time for the adaptive responses to develop, which may take up to several weeks [13]. Moreover, activation of the innate immune system is a vital step for the activation of the adaptive immune system, which allows for more effective, specific responses with long-lasting memory [14]. Adaptive immunity The adaptive immune system as opposed to the innate, is capable of recognizing specific self- and non-self-antigens and then responding with heightened efficiency upon repeated exposure. Adaptive immune cells recognize antigens on a more specific level and develop long- lasting memory against them. The two main cell types belonging to the adaptive immune system 4 are B and T lymphocytes, which carry out antibody responses and cell-mediated immune responses, respectively. T lymphocytes T lymphocytes are produced in the bone marrow or fetal liver from a common lymphoid progenitor cell and migrate to the thymus for maturation and eventual expression of the T cell receptor (TCR) [15]. The role of TCR is to recognize antigens via its interaction with a major histocompatibility complex (MHC) bound to a peptide (self or foreign). During the maturation process, T cells undergo processes of positive and negative selection in the thymus [16]. Positive selection takes place in the cortex of the thymus, during which the T cells with TCRs incapable of binding to the thymic epithelial cells are destroyed in order to ensure that only T cells capable of interacting with MHC will survive. Negative selection process takes place in the medulla of the thymus, where the T cells with TCRs which bind with strong affinity to self-derived peptides are destroyed in order to prevent the formation of self-reactive T cells. It has been estimated that due to the process of somatic recombination as many as 1015 different antigen TCRs can be generated [17]. This allows for a tremendous diversity of pathogen recognition by the adaptive immune system. Due to somatic recombination of the TCR, T cells are capable of recognizing any peptide that could exist and are therefore well equipped to deal with the diversity of rapidly mutating pathogens and cancer cells. Mature antigen-naïve lymphocytes migrate from their primary lymphoid organs to secondary lymphoid organs, such as lymph nodes, spleen, and tissue sites [18], where they become activated via interaction of the TCRs with appropriate MHC molecules on the surfaces of specialized antigen presenting cells (APCs), namely macrophages and dendritic cells [19]. 5 There are two distinct T cells, CD4+ and CD8+ T cells (also termed as cytotoxic T lymphocytes - CTLs). CD4+ T cells bind to MHC-II (HLA-DR, HLA-DQ, and HLA-DP) complexes, which present peptides derived from extracellular antigens and are 18-20 amino acids in length. CD8+ T cells interact with shorter peptides of 9-11 amino acids in length, derived from cytosolic proteins, and presented on MHC-I (HLA-A, HLA-B, and HLA-C) molecules [20]. TCR receptors are composed of α and β subunits, forming a heterodimer. TCR is also linked to the CD3 signal transducing complex, which contains cytoplasmic immunoreceptor tyrosine-based activation motif (ITAM), which is phosphorylated via Src family kinases, allowing for recruitment and binding of Zap-70, and downstream signaling [21]. Upon interaction of the TCR complex with an MHC complex, a second co-stimulatory signal is required for activation of a T- cell. Co-stimulatory molecules such as CD28 and others, engage with B7-1 (CD80) or B7-2 (CD86) receptor on the APCs, thereby activating a signaling cascade and promoting T cell proliferation and expansion. Cytokines also play an important role in T cell responses. For instance, IL-2 is not only required for T cell expansion but is also required for differentiation and function of effector memory cells [22]. CD4+ T cells CD4+ T cells are the most abundant group of T cells in the body, with the majority of the CD4+ T cells functioning as T helper (Th) cells. Cytokines produced by APCs play an important role in driving the differentiation of helper T cells into a particular subtype. CD4+ T cells can differentiate into Th1, Th2, Th9, Th17, Th22, T follicular helper T cells (Tfh), and regulatory T cells (Tregs) subtypes, depending on the cytokines they are exposed to [23]. For example, IL-12 and IFN-γ drive differentiation of Th1, IL-4 drives differentiation of Th2 [24], and IL-6 and TGF-β drive differentiation of Th17 cell subtypes [25]. All CD4+ T cell subsets play important 6 roles, with each subtype being characterized by specific cytokines and cytokine-mediated T cell effector responses. For example, Th1 cell subset secretes IFN-γ and TNF, Th2 subset secretes IL-4, IL-5, and IL-13, and Th17 subset secretes IL-17. The Th1 cell subset is particularly important in eliminating viral infections and tumor responses due to its activation of NK cells and cytotoxic T cells [26]. The Th2 cell subset is particularly important for elimination of extracellular pathogens and parasites, as wells as induction of humoral responses [23, 27]. The Th17 cell subset plays a very important role in protection against extracellular pathogens, mucosal immunity, and activation of the adaptive responses via recruitment of neutrophils. Dysbalanced function in Th1 and Th2 cells subtypes has been proposed to play a role in autoimmune diseases such as multiple sclerosis (MS), insulin dependent diabetes mellitus (IDDM) and rheumatoid arthritis (RA) [23]. Interestingly, Th17 cells, in particular, have been linked to many autoimmune diseases including psoriasis, RA, MS, inflammatory bowel diseases (IBD) and even spondyloarthritides [28]. However, it seems that a complex combination of multiple cell types, cell subtypes, and cytokines likely each play a role in the development of autoimmunity, as they all play important roles in various aspects of normal immune responses. Once mature antigen-naïve T cells become activated via interactions with APCs, they undergo clonal expansion and differentiate into effector cells which migrate to the site of infection to assist in pathogen eradication [13]. Activated T cells produce a wide variety of cytokines, important for mediating innate and adaptive immune responses. CD4+ T cells play an important role in enhancing B cell activation and antibody production, enhancement of CD8+ T cell responses, macrophage and DC function, as well as regulation of the length and magnitude of the immune response. The effector cells then undergo apoptosis shortly after infection is cleared. 7 CD8+ T cells Upon interaction with primed APCs, similar to CD4+ T cells, CD8+ T undergo activation and expansion. They produce a variety of cytokines such as IL-2, TNF-α, and IFN-γ and migrate to the site of inflammation. The unique feature of CD8+ T cells is their ability to kill target host cells (virally infected or tumor cells) in a contact-dependent mechanism, with type I interferons (IFNs) playing an important role in CD8+ T cells expansion and cytotoxicity [29]. CTLs can induce apoptosis of the target cells via granzyme and perforin granules, where perforin granules disrupt the membrane, while granzyme proteases activate apoptosis. FAS-FASL activation by upregulation of FASL on the cell surface of CTLs can also induce apoptosis via classical caspase activation [30]. It is worth noting that the cytotoxic effector function of NK cells is similar to CD8+ T cells, where NK cells can kill target host cells via perforin/granzyme and FAS-FASL activation [31]. Upon activation, a subset of CD8+ T cells differentiates into memory cells, where IL-2 and IL-12 play an important role in this process [32, 33]. Memory cells are generally long-lived antigen-specific cells, capable of rapid expansion and enhanced elimination of the pathogen upon repeated exposure [34]. Interestingly, CD8+ T cells with regulatory functions have also been described, for example, a small subset of CD8+ CD25+ Foxp3+ T cells capable of suppressing effector T cell proliferation has been identified in the periphery of mice and humans [35]. B lymphocytes B cell development starts out in the bone marrow, where B cells progress through Pro-B, Pre-B and immature B cell stages [36]. Mature B cells acquire IgD and IgM antibodies on their surface, completing the antigen-independent maturation. The second, antigen-dependent phase of 8 B cell development takes place in the secondary lymphoid organs. Mature B cells become activated upon capture of an antigen via B cell receptor (BCR) and migrate to the T-B border via upregulation of C-C chemokine receptor 7 (CCR7) and Epstein-Barr virus-induced receptor 2 (EBI2) [37]. Activated B cell internalizes the antigen, processes it and loads peptides onto MHC- II complexes, thereby increasing its cell surface expression. At this point, B cells act much like APCs, where the MHC-II binds to the TCR, and costimulatory receptors such as B-7 and CD40 interact with CD28 and CD40L on T cells, respectively. When activated B cells interact with primed T cells, B cells expand and can become either memory cells or short-lived plasma cells [38]. Plasma cells are specialized cells, which secrete large quantities of antibodies. These plasma cells are short-lived and produce low-affinity antibodies [39]. Activated B cells can also enter the follicle of the lymph nodes and establish a germinal center (GC) where they interact with Thf cells and follicular dendritic cells, which ultimately results in the production of higher affinity B cells via class-switching and affinity maturation processes. Some GC B cells can differentiate into memory B cells and long-lived plasma cells [40]. The plasma cells migrate to the bone marrow (BM) via C-X-C chemokine receptor 4 (CXCR4), and are maintained by CXCL12, IL-6 and a proliferation inducing ligand (APRIL) secretion by the stromal cells. These plasma cells continue producing antibodies, providing serologic memory against subsequent exposures to the same antigen. Additionally, memory B cells establish a population in the lymphoid tissues and rapidly respond to re-infection. Interestingly, during affinity maturation, the process of somatic hypermutation takes place, during which cells may acquire self-reactivity and persist in the periphery, giving rise to autoimmunity. The best studied example of this is systemic lupus erythematosus (SLE) [41]. 9 Antibodies play a variety of important functions in the immune responses, including pathogen neutralization, phagocytosis of pathogens, activation of complement and activation of antibody-dependent cellular cytotoxicity (ADCC) [42]. Neutralization prevents attachment of pathogens, such as Salmonella enterica to epithelial cells [43]. Another example is anti-HIV-1 gp120 antibody, which prevents binding of the gp120 to CD4 [44]. Activation of the complement is an important mechanism which results in the lysis of the pathogen or of infected cell [45]. The Fc region of antibodies can engage with a variety of Fc receptors, having a multitude of effects, depending on the type of cell involved. For example, the bacteria-antibody complex can be phagocytosed upon interaction with FcγR on a phagocytic macrophage cell [42]. Antibody- dependent cell-mediated cytotoxicity responses are also mediated via Fc-FcR interactions which result in target cell death via apoptosis or cell lysis and are thought to be especially important for anti-tumor immunity mediated via anti-tumor antibodies [46]. Tolerance Both B and T lymphocytes undergo positive and negative selection processes during their maturation, allowing them to develop receptors capable of binding antigens, yet eliminating those cells, which contain receptors with strong reactivity to self-antigens [16, 47]. This is referred to as central tolerance. Interestingly, although central tolerance is an effective process, it does not destroy all self-reactive lymphocytes [48]. Peripheral self-tolerance is an important mechanism for the prevention of autoimmunity, by inhibiting and ensuring anergy of self- reactive lymphocytes which escaped negative selection [49]. An important family of T cell sub- types responsible for prevention of autoimmunity via peripheral self-tolerance is a group of 10 immunosuppressive T cells. The discovery was made in 1970 when two researchers found out that some T cells are capable of downregulating immune responses and named them T suppressor cells [50]. In addition to activation of innate immune responses, and activation and recruitment of the effector B and T lymphocytes to the site of inflammation, immunosuppressive classes of T cell are also recruited in order to balance the magnitude of the immune response and in order to maintain self-tolerance. Failure of the various types of immunosuppressive T cells to perform their tolerogenic activity due to the pathologic number and /or function may cause autoimmunity and chronic inflammation [51]. There are several subtypes of immunosuppressive T cells, some of which are outlined below. Foxp3+ regulatory T cells Conventional CD25+ regulatory T cells (Tregs) are produced in the thymus [52]. They are capable of suppressing responses toward self- and non-self-antigens. Functionally mature T regs persist in the periphery and suppress self-reactive T cells. Depletion of Tregs has been shown to cause multiple autoimmune diseases such as inflammatory bowel disease, due to loss of self-tolerance to commensal bacteria [53]. Conventional T regs express transcription factor forkhead box P3 (Foxp3), which is the master regulator of the Treg development and function. Foxp3 transcription factor binds to several other transcription factors forming an oligomeric complex, which includes NFAT, Runx1, HDAC, and possibly others. This transcription factor complex promotes the expression of multiple genes important for Treg function, such as IL2, IL4, CD25, CTLA4, and others [51, 54]. Mutations in the Foxp3 gene cause immune dysregulation, polyendocrinopathy, enteropathy, X-linked syndrome (IPEX) in humans, which is a severe multi-systemic autoimmune disease which develops in infancy [55]. 11 It is well established, that IL-2 signaling is essential for many cells including CD4+ and CD8+ T cells, B cells and NK cells [56]. It facilitates T cell differentiation and expansion, and at the same time promotes T cell apoptosis. Importantly, IL-2 signaling promotes Treg differentiation from naïve T cells facilitated by TGF-β [57]. The CD25 molecule is a component of the IL-2 receptor α chain IL-2R. CD25 and IL-2 are mandatory for Treg development, where IL-2 signaling activates JAK3 and STAT5 phosphorylation and translocation to the nucleus, where it binds to the promoter and induces Foxp3 transcription [58, 59]. Therefore, there is a negative feedback loop, where IL-2 promotes T cell responses and at the same time ensues halting of these responses by the promotion of Treg differentiation. Disruption of this feedback mechanisms results in autoimmunity [51]. Tregs suppress effector cells through several mechanisms, which include secretion of suppressive cytokines IL-10, TGF-β, and IL-35, all of which have been shown to be important mediators of Treg function in vivo. Other mechanisms include suppression by the cytolysis of effector cells, suppression by depletion of IL-2 and suppression through blockade of DC maturation via expression of inhibitory receptors CTLA-4 and Lag-3 [60]. Upon antigen-driven activation, Tregs are recruited to the site of infection via chemokines, where they compete with effector CD4+ T cells for interactions with DCs. Tregs suppress DC functions and maturation, thereby preventing activation of the effector T cells. Furthermore, they secrete immunosuppressive cytokines such as IL-10 and bind large amounts of IL-2, thereby depleting available resources important for effector cell expansion. Once Treg responses are activated, their suppressive functions are not limited to that particular antigen [61]. 12 Type 1 regulatory T cells Compared to Foxp3+ conventional Tregs, type 1 regulatory T cells (Tr1 cells) are the less studied immunosuppressive regulatory T cell due to the difficulty of identifying these cells. Tr1 cells are Foxp3- CD4+ T cells that are characterized by high levels of IL-10 and TGF-β cytokine secretion [62, 63]. Tr1 presence and normal function are also essential for attenuation of tissue inflammation and prevention of autoimmunity [64, 65]. These cells were first characterized by the Roncarolo group in 1997 who determined that a subset of antigen-specific T cells prevented colitis by inhibiting T cells responses [62]. Prior to 2013, these cells were identified via their unique cytokine expression profile of IL-10high, TGF-β+, IFN-γ+, IL-5+, IL-4-, and IL-2low [62, 63]. A discovery of surface markers Lag-3 and CD49b on Tr1 cells in 2013 has made their identification much easier [66]. Mechanisms of actions of Tr1 cells include direct inhibition of T cell proliferation via secretion of IL-10 and TGF-β cytokines [67-69], as well as indirect mechanisms via inhibition of IL-2 and IFN-γ production by effector T cells [70]. Additionally, Tr1 cells can exhibit their suppressive function via their effects on APCs via downregulation of co-stimulatory receptors and attenuation of proinflammatory cytokine release [71, 72]. Tr1 cells can also inhibit B cell functions and regulate antibody isotype switching, where Tr1 cells have been shown to promote IgG4 and prevent IgE antibody production [73, 74]. Finally, Tr1 cells have been reported to inhibit T cell responses via cytolysis of myeloid APCs mediated by granzyme B secretion [75, 76]. Tr1 cells have been implicated in multiple autoimmune diseases, including IDDM, MS, and pemphigus vulgaris. Specifically, IDDM patients have been shown to have IFN-γ-skewed 13 responses to pancreatic islet peptides, versus IL-10 responses elicited from T cells from non- diabetic individuals [77]. Similarly, Tr1-like cells from MS patients have been shown to produce reduced levels of IL-10 compared to healthy patients [78]. Moreover, CD4+ T cells from MS patients have a reduced propensity to differentiate into Tr1 cells compared to CD4+ T cells from healthy donors [31]. The IL-10 signaling pathway was found to be inhibited in the MS patients due to reduced STAT3 phosphorylation via SOCS3 inhibitory effects [31]. Finally, pemphigus vulgaris patients tend to lack desmoglein 3 (Dsg3) specific Tr1 cells compared to healthy patients [79]. Moreover, it was shown that Dsg3-specific Tr1 cells elicit their inhibitory function via IL-10 and TGF-β cytokine secretion. Based on the aforementioned evidence, it is thought that Tr1 cell reduction or functional impairment may be the underlying cause of loss of tolerance to self and therefore autoimmunity. Activation of Tr1 cells requires antigen-specific TCR interactions, but once they are activated they can induce suppression against other antigens via bystander suppression [72]. Unfortunately, there is no known master transcriptional regulator known to be responsible for Tr1 cells differentiation, in contrast to Foxp3 expression in Tregs. However, we know some players, which are important for Tr1 cell differentiation. IL-10 is a particularly important cytokine for Tr1 differentiation [80]. In the presence of IL-27, which is a key cytokine for IL-10 production [81], CD4+ T cells differentiate into Tr1s [82, 83]. This is thought to be STAT3 mediated, as its overexpression in CD4+ T cells, promotes Tr1 differentiation [84]. Differentiation of Tr1 cells is also regulated by DCs [49]. IL-27 produced by the DCs promotes the differentiation of CD4+ T cells into Tr1 cells and induces IL-10 production [81] via activation of the Blimp-1 and eomesodermin transcription factors [85]. Additionally, IL-27 induces aryl hydrocarbon receptor (AhR) and c-maf transcription factor upregulation [83], as 14 well as the production of ICOS and IL-21, all of which promote Tr1 differentiation [82]. A specific subtype of DCs is required for Tr1 cell differentiation, namely tolerogenic dendritic cells (tDCs), identified as CD11clowCD45RBhigh cells [86]. These cells are unique in their immature phenotype and ability to secrete high amounts of IL-10. The non-classical MHC-I molecule, HLA-G, plays an important role in Tr1 differentiation via complex interactions between tDCs and naïve T cells [87]. HLA-G is known for its high expression in the trophoblasts and its induction of fetal tolerance during pregnancy via interaction with Killer Cell Immunoglobulin- Like Receptor, Two Ig Domains And Long Cytoplasmic Tail 4 (KIR2DL4) inhibitory receptor on NK cells [88]. HLA-G on tDCs interacts with naïve T cells via Ig-like transcript 2 (ILT2), thereby promoting their differentiation into Tr1 cells [89]. In turn, HLA-G on the surface of T cells binds to ILT4 on DCs, thereby preventing their maturation and maintaining their phenotype of IL-10 producing cells [90]. IL-10 further strengthens this positive feedback loop by promoting HLA-G expression in DCs and T cells. Interestingly, despite years of research, it is still unclear, where Tr1 cells are generated, as they have been found to circulate in the periphery [66], at tissue sites, such as gut mucosa [62] and in spleens of mice [91]. It has been reported that Tr1 cells from intestinal mucosa can migrate to the periphery in the animal models of type 1 diabetes (also known as IDDM) [92]. Also, in-vitro stimulated human Tr1 cells from peripheral blood have been shown to express gut- homing receptors [93], and joint specific Tr1 reactive to collagen II have been found in the circulation of RA patients [94], suggesting that these cells differentiate at tissue sites and can migrate to the periphery. Despite limitations in the knowledge about Tr1 cell physiology and differentiation, there has been an increased interest in the use of Tr1 cells as a therapeutic modality for many 15 pathologies. Tr1 cell co-infusions have shown durable (long-term) prevention of graft-versus- host disease (GVHD) upon hematopoietic stem cell transplantation (HSTC) therapy in patients with hematological malignancies [95]. In addition to preventing GVHD, a recent study showed that Tr1 cell infusion along with HSCT therapy also promotes graft-versus-leukemia killing by directly killing myeloid leukemic cells in a granzyme B dependent manner, as well as indirectly via its effects on transplanted allogenic PBMCs [96]. Tr1 cell therapy has also shown some success in clinical trials involving patients with refractory Crohn’s disease, where 75% of patients had a reduction of the disease activity index 8 weeks after infusion. However, the remission rate was reduced from 38% to 25% between weeks 5 and 8, suggesting that higher dose or additional rounds of Tr1 infusion may be needed [97]. Until recently infusion of Tr1 cells was problematic due to the difficulty of expanding these cells in vitro. However, recently, a group of researchers developed a system utilizing a lentiviral vector that expressed IL-10 that promotes differentiation of conventional CD4+ T cells into Tr1-like cells [98]. This newly developed method will allow for scalability of Tr1 cell expansion. While there is much more optimization required before Tr1 cell therapy can be more widely utilized, it is an appealing modality for cell therapy that can be used for autoimmune and chronic inflammatory diseases and holds a lot of potential for future use in organ transplantations as well. Ankylosing spondylitis Ankylosing spondylitis (AS) is a prototypic chronic inflammatory disease affecting the axial skeleton. It belongs to the family of related immune-mediated arthropathies - 16 spondyloarthritides which also includes reactive arthritis, psoriatic arthritis, enteropathic arthritis, Crohn’s disease, undifferentiated spondyloarthritides and juvenile-onset spondyloarthritis [99, 100]. AS affects an estimated 0.9-1.4% of the adult population in the U.S. [101, 102]. AS has a wide range of manifestations, including spinal inflammation, chronic back pain, sacroiliitis and loss in the range of motion in the lumbar and thoracic spines [103]. Another common feature of AS is osteoporosis of the trabecular bone of the vertebral bodies, making them susceptible to vertebral fractures and spinal cord injuries [104, 105]. Osteoporosis of the femurs has also been reported [106]. Peripheral arthritis, acute anterior uveitis, and inflammatory bowel disease are also prevalent in AS patients, significantly impacting their quality of life [107]. Other less common manifestations include psoriatic lesions, aortitis, as well as renal and lung involvement [108]. The age of onset of AS symptoms is in the early twenties [109], although it is estimated that the diagnosis of AS is delayed by 8 to 10 years due to its gradual progression and ambiguous presentation [110]. Genes associated with AS and their role in its pathogenesis HLA-B27 AS is a highly heritable disease, as confirmed by multiple twin studies [111-113]. Interestingly, the risk of developing AS in first degree relatives with AS has been reported to be 6 to 16-fold higher than in general population [114, 115]. In 1973 human leukocyte antigen (HLA)*B27 was the first gene found to be associated with AS, where 88% of AS patients were found to be HLA*B27 positive [116, 117]. Since then multiple other HLA*B alleles have also been found to be associated with AS, including HLA*B13:02, HLA*B40:01, HLA*B40:02, 17 HLA*B47:01, and HLA*B51:01. Meanwhile, HLA*B07:02 and HLA*B57:01 are protective against AS [118]. HLA*B2705, HLA*B2702, HLA*B2704, and HLA*B2707 are the most common subtypes that have been shown to be associated with increased risk of AS, while two subtypes HLA*B2706 and HLA*B2709 are not associated [100]. HLA*B27 accounts for 20- 50% of total genetic susceptibility of AS [107]. Interestingly, while 90% of AS patients are HLA*B27 positive, the risk of developing AS in HLA*B27 positive individuals is about 5-10%, suggesting the involvement of other genes in the susceptibility of this disease. HLA-B27 is an MHC class I complex which is ubiquitously expressed in all cell types and is particularly highly expressed on APCs [100]. MHC-I is a heterotrimer, formed by an α- heavy chain, β2-microglobulin and a peptide derived from cytosolic or viral proteins intracellularly. Upon folding, the entire complex is shipped to the cell surface for antigen presentation. On the cell surface, MHC-I can be recognized by CD8+ T and NK cells. Foreign peptides presented on MHC-I are recognized by the TCR on CD8+ T cells, which result in activation of the CD8+ T cells and killing of the infected cell [20]. Interestingly, many pathogens and tumors disrupt the antigen presentation pathway in order to escape recognition by CD8+ T cells, thereby reducing MHC-I surface levels [119]. In turn, interactions of MHC-I with killer- cell immunoglobulin-like receptors (KIRs) on NK cells are required for inhibition of NK cell activity, where reduction or absence of surface MHC-I results in activation of NK cells and ultimate killing of the virus-infected or malignant cells. Despite knowing HLA*B27’s association with AS for over 40 years, we still do not fully understand the molecular mechanism through which it contributes to the pathogenesis of AS. Several hypotheses explaining the role of HLA*B27 in AS have been proposed. The arthritogenic peptide hypothesis proposes that arthritis-causing peptides derived from a pathogen 18 are preferentially presented by HLA*B27 and are recognized by autoreactive CD8+ T cells resulting in chronic inflammation due to microbial mimicry [120]. This hypothesis is, however, weakened by the fact that HLA*B27 transgenic rats still developed colitis and arthritis in the absence of CD8+ T cells [121]. The homodimer hypothesis is based on the propensity of HLA*B27 molecules to misfold and form heavy chain homodimers [122]. HLA*B27 has been shown to form free heavy chains and homodimers on PBMCs. These free heavy chains and homodimers have been shown to be recognized by KIR and Leukocyte Ig-like receptor (LIR) receptors with higher affinity than HLA*B27 [123]. Activation of these receptors can enhance T and NK cell responses. According to the misfolding hypothesis, the propensity of HLA*B27 to misfold causes its aggregation in the ER, which in turn activates the unfolded protein response (UPR) and ER overload response (EOR), which induce proinflammatory responses [124-126]. Interestingly, UPR activation enhances IL-23 secretion that has been shown to induce inflammation at the enthesis (junction between tendon and bone) [127]. This is a very intriguing finding because enthesis is thought to be the primary site of inflammation in AS, which later progresses to ankylosis [128]. ERAP1 In 2007, the Welcome Trust Case Control Consortium and Australo-Anglo-American Spondylitis Consortium identified five single nucleotide polymorphisms (SNPs) (rs27044, rs17482078, rs10050860, rs30187, rs2287987) in the ERAP1 gene to be associated with AS [129]. This study estimated that 26% of the risk of developing AS is attributed to ERAP1. These associations were later replicated in other studies in a variety of populations throughout the world [130-137]. 19 ERAP1 belongs to the oxytocinase subfamily of M1 zinc-metallopeptidases, and functions as a molecular ruler of the peptides destined for antigen display by MHC-I molecules [138]. Specifically, upon degradation of cytosolic proteins by the proteasome, peptides are transported into the endoplasmic reticulum (ER) via transporter associated with antigen processing (TAP). ERAP1 trims peptides that are 9-16 amino acids long from the N-terminus to the perfect size of 8-10 amino acids long, which are then loaded onto MHC-I complexes [139]. Epistatic gene-gene interactions between specific ERAP1 variants and HLA-B*27 have been demonstrated, where they together increase the risk of developing AS [118, 140, 141]. Additionally, ERAP1 and HLA-B27 interact functionally due to their involvement in the same molecular pathway of antigen presentation, where suppression of ERAP1 has been shown to increase free heavy chain and surface HLA*B27 levels [142]. Because of the epistatic relationship between ERAP1 and HLA*B27, initial mechanistic research following the GWAS studies focused on the role of ERAP1 in the antigen presentation pathway. ERAP1 in adaptive immunity It is well-accepted that ERAP1 affects antigen presentation, via its peptide trimming activity [143-146]. Our lab has demonstrated that depending on the ERAP1 allele, the trimming efficiency of ERAP1 changes, thereby affecting not only the availability of specific peptides for MHC-I, but also, globally affecting the surface levels of MHC-I. More specifically, when ERAP1 alleles harboring SNPs associated with high risk of AS were co-expressed with HLA*B27, there was a reduction in the HLA*B27 surface levels, compared to the ERAP1 alleles containing protective SNPs [147]. In our survey of ERAP1 allele activity, we showed that the two disease-associated variants (rs30187 and rs27044) had differing effects on enzymatic activity and substrate specificity [148], where one reduced and the other increased ERAP1 20 catalytic activity, accordingly, but ultimately they both resulted in reduced HLA-B27 surface level expression [149]. Additionally, we have shown that the absence of ERAP1 completely shifts peptide immunodominance by dictating which peptides are ultimately presented by MHC-I to CTLs [150]. Mice expressing human ERAP1 variant containing SNPs found to be associated with increased risk of AS generated unique antigen-specific T cell clones upon vaccination with foreign antigens, compared to mice expressing an AS protective ERAP1 allele [151]. ERAP1 alleles ultimately determine the overall surface MHC-I levels in these mice. Reeves et. al. showed that ERAP1 allele combinations seen in AS patients have a reduced ability to generate peptides ultimately resulting in reduced overall cell surface levels of HLA*B27 and MHC-I in general [152]. All AS combinations resulted in reduced peptide availability, regardless of whether the ERAP1 SNPs resulted in reduced or increased activity of ERAP1, as both under- trimming and over-trimming activities ultimately lowered the availability of correct sized peptides. This suggests that in diseases associated with ERAP1, its aminopeptidase activity plays at least a partial role in their pathogenesis, with the presence of specific ERAP1 SNPs changing ERAP1 activity and peptide specificity. ERAP1 in innate immunity Interestingly, ERAP1 has also been shown to play a direct role in innate immunity and may contribute to AS pathogenesis via these functions [153, 154]. For example, ERAP1 is involved in the proteolytic cleaving of several cytokine receptors, including TNF-R1, IL-6Rα and type II IL-1 decoy receptor (IL-1RII) [155-157]. This shedding function of ERAP1 allows for regulation of receptor availability on cell surfaces, thereby mediating immune responses. However, it is important to note that the shedding function of ERAP1 has not been confirmed in 21 ERAP1 knockout mice and soluble levels of receptors have been shown to be independent of ERAP1 polymorphisms in AS patients [158]. Variants in the human ERAP1 gene have also been shown to be associated with essential hypertension [159]. Overexpression of the ERAP1 gene in COS-7 fibroblast-like cells has been shown to induce its secretion outside of cells [160], suggesting that in certain conditions, ERAP1 may be secreted outside of cells and function as a soluble protein in the extracellular milieu or circulation. Dr. Tsujimoto’s group indeed confirmed that ERAP1 itself can be secreted outside the cell in response to LPS and IFN-γ activation [153]. Interestingly, while ERAP1 does not have an ER-retention sequence, it has been shown that exon 10 contains a sequence which is important for its retention in the ER [161]. The authors hypothesized that certain proteins might bind to the exon 10 sequence allowing for its retention in the ER, suggesting that saturation of the binding sites may cause secretion of ERAP1 outside the cells. It is also possible that there are other ER proteins, which compete for the same binding sites as ERAP1, allowing for ERAP1 secretion to take place; however, no such proteins have been identified to date. ERAP1 secretion from macrophages is mediated via activation of various TLRs, including TLR1, TLR2, TLR4, TLR6 and TLR9 in MyD88- and NF-kB- dependent manner, via the calmodulin pathway [162]. Furthermore, TLR-induced secretion of ERAP1 is mediated via IFN-γ, IFN-β, and TNF-α, where all three cytokines are needed for maximal ERAP1 secretion. Once ERAP1 is secreted by the macrophages, it stimulates their phagocytic activity [153]. Based on these results our lab wished to further investigate the function of secreted ERAP1 on immune cells. We have shown that extracellular functions of ERAP1 are dependent on its aminopeptidase activity and do not require re-uptake by the cells to elicit ERAP1’s extracellular effects. Exposure of total human PBMCs with extracellular ERAP1 protein variants, resulted in 22 activation of NK cells, DC and T cell activation, as well as enhanced production of cytokines and chemokines, through mechanisms involving the NLRP3 inflammasome and cathepsin B pathways [154]. Interestingly, ERAP1 variants had differing effects upon PBMCs, where exposure to AS predisposing variants (rs 30187 and rs 27044) enhanced production of pro- inflammatory cytokines such as IL-1β, TNF-α, and IL-6 and enhanced T cell and NK cell activation of treated human PBMCs. This suggests that polymorphisms in ERAP1 may predispose individuals to autoimmune diseases at least in part due to their extracellular function. Interestingly, ERAP1 functions are also important in NK cell biology. Our lab has extensively studied ERAP1-/- mice, and we have shown that NK cells from the ERAP1-/- mice exhibit increased activation in response to innate immune stimuli [163]. Evaluation of the splenic NK cells from ERAP1-/- mice revealed that ERAP1-/- mice have an increased percent of licensed and more terminally mature NK cells. Additionally, compared to identical innate immune stimulations of splenocytes derived from WT mice, stimulation of cells from ERAP1-/- mice resulted in increased levels of activation markers and proinflammatory cytokine secretion including TNF-α, IL-6, and IL-1β, all of which have been implicated in AS [163]. It is evident that beyond its function in antigen processing, ERAP1 also functions as a modulator of the innate immune responses, a function which needs to be further investigated in order understand its role in the pathogenesis of AS and in order to develop targeted therapies for this disease. Many ERAP1 polymorphisms in the ERAP1 gene have also been found to be associated with other autoimmune diseases through genome-wide association studies (GWAS). Interestingly, the rs30187 SNP was also found to be associated with IDDM [164]. Two GWAS studies found associations between ERAP1 SNPs and psoriasis. The Genetic Analysis of Psoriasis Consortium and the Welcome Trust Case Control Consortium 2 identified rs27525 SNP 23 [165] and another GWAS in Han Chinese population reported rs151823 to be associated with psoriasis [166]. Finally, rs30187 SNP has been shown to be associated with Crohn’s disease and multiple sclerosis [167]. Association of ERAP1 gene with multiple autoimmune diseases suggest its possible involvement in their pathogenesis. It is therefore important to investigate ERAP1 functions in order to gain molecular insight of the mechanisms linking ERAP1 to autoimmunity. Other genes associated with AS GWAS studies identified that rs11209026 SNP in the IL23R (IL-23 receptor) gene is protective against AS [129], where a non-synonymous substitution results in loss-of-function mutation [168] and reduced Th17 effector cells [169]. IL23/Th17 axis is thought to play a major role in AS pathogenesis. AS patients with active disease have been shown to have elevated IL-23 and IL-17 in their sera [170]. Moreover, macrophages from AS patients have been reported to secrete increased levels of IL-23 in response to stimulation [171]. Interestingly, associations in other genes involved in the IL23/IL17 pathways have also been reported, namely, STAT3, JAK2, TYK2, and several others [99]. IL-23 is a pro-inflammatory cytokine which plays an important role in regulating mucosal immunity and induces IL-17 production from Th17 cells [172]. Overexpression of IL-23 in mice resulted in entheseal inflammation promoted by IL-17 secreting RORγt+CD3+IL-23R+CD4-CD8- T cells [127]. It is thought that IL-17 promotes joint destruction by promoting osteoclast activity and suppressing proper bone regeneration [173]. While the exact cause for elevated IL-23 in AS patients is unclear, it is widely hypothesized that the elevated IL-23 stems from its enhanced production in the gut [174], possibly triggered by mucosal immune responses to gut dysbiosis. 24 Skeletal features of AS AS primarily affects the spine and the sacroiliac (SI) joints, with sacroiliitis being its hallmark feature, present early on in the disease [175]. The progression of AS is characterized by excessive osteoproliferation and syndesmophyte (bony bridges) formation between adjacent vertebrae, which ultimately results in joint fusion [176]. Based on imaging studies, we know that initially, subchondral erosions of the ilia are appreciated, which are followed by subchondral sclerosis and bony proliferation. AS is characterized by enthesitis [177], and it is hypothesized that an initial inflammatory insult at the enthesis results in local bone destruction and that in attempts to repair the damage, the new aberrant bone is laid down resulting in syndesmophyte formations. However, there is still much debate as to whether the syndesmophyte formation is secondary to initial joint destruction or is an intrinsic feature of the disease. Moreover, paradoxically, AS patients also experience trabecular bone loss, leading to osteoporosis [106]. It is thought that osteoporosis is secondary to inflammation, similarly to other inflammatory diseases such as RA, IBD and SLE [178]. In order to develop appropriate treatments for AS, the exact pathophysiological processes responsible for ankylosis and osteoporosis need to be understood. Bone remodeling Osteoclasts, osteoblasts, and the osteocytes are the main cells involved in bone remodeling and homeostasis. Bone remodeling is a process of bone maintenance, where microtears (microfractures) from stress are removed via osteoclasts and repaired by osteoblasts [179]. Osteoclasts perform bone resorption via secretion of hydrochloric acid and proteases such as cathepsin K, which degrade the bone mineral matrix [180]. Upon bone matrix degradation by 25 the osteoclasts, osteoblasts counteract this action via secretion of new bone matrix and mineralization. The loss of homeostasis between these two processes may result in the development of bone disease. For example, in postmenopausal osteoporosis, the bone resorption by osteoclasts outweighs the bone mineralization by the osteoblasts resulting in weakened bones [181]. Osteoclasts are the only bone resorbing, aka bone degrading cells. Increased osteoclast activity is the hallmark of many pathologies involving bone loss, such as osteoporosis, RA and others. Osteoclasts are derived from the monocyte/macrophage precursor lineage. Presence of macrophage-colony stimulating factor (M-CSF) and receptor activator of NF-κB ligand (RANKL) are required for osteoclast precursors to differentiate into osteoclasts [182]. M-CSF activates early osteoclast differentiation by binding to its receptor c-fms on osteoclast precursor cells and activating MAPK and ERK signaling. RANKL is a member of the TNF superfamily, which exists in a membrane-bound and soluble forms. RANKL binds to its receptor RANK on the surface of osteoclast precursors and induces their differentiation into osteoclast via NF-kB, phospholipase Cγ, c-fos and NFATc 1 activation [183]. Osteoclast associated receptor (OSCAR) and triggering receptor expressed in myeloid cells-2 (TREM2) are co-stimulatory receptors which are also important for osteoclast differentiation [184]. Osteoprotegerin (OPG) binds to RANKL, thereby preventing its interaction with RANK. Interestingly, the RANKL/OPG ratio is a major determining factor in the physiological state of bone remodeling [185]. Osteoblasts are derived from the mesenchymal stem cells, with bone morphogenetic proteins (BMPs) playing an important role in their differentiation [186]. BMPs belong to the TGF-β superfamily which exerts its functions via activation of SMAD. Runx2 is a master regulator of the osteoblast differentiation, an absence of which completely arrests osteoblast 26 maturation and mineralization function [187, 188]. Another important player in osteoblast differentiation is Osterix (Osx), a transcription factor which has been shown to be important for the commitment of mesenchymal stem cells to osteoblasts. It is also important for osteocalcin and Collagen alpha-1 (XI) chain (Col1α1) expression [189]. Wnt signaling pathway is the most well studied and understood pathway involved in osteoblast differentiation [190]. Wnt is a glycoprotein which binds to the Wnt receptor complex, composed of lipoprotein receptor (LPR), frizzled and GSK3β and transmits the signal intracellularly. Upon Wnt binding to the receptor complex, the β-catenin is rescued, which would otherwise be degraded. β-catenin translocates to the nucleus, where it binds to Tcf and Lef1 cofactors and induces expression of Runx2 and other pro-osteoblastic genes, thereby promoting osteoblast differentiation [191]. Interestingly, osteoclasts and osteoblasts are largely influenced by cytokines. For example, TNF-α is known to inhibit osteoblast differentiation and stimulate bone resorption [192]. TNF-α is thought to play an important role in the pathophysiology of chronic inflammatory diseases. For example, in RA, TNF-α is thought to contribute to joint inflammation and damage [193]. TNF-α has been reported to inhibit osteoblast differentiation via inhibition of insulin-like growth factor-1 (IGF-1) [194], Osx [195] and Runx2 [196-198]. Interestingly, depending on the concentration, exposure time and the cell type, TNF-α has been reported to have a paradoxical effect on osteoblasts [199]. However, there is an agreement that it inhibits osteoblastogenesis when acting on pre-osteoblasts, through induction of dickkopf-related protein 1 (DKK-1) which in turn inhibits Wnt pathway [200]. This mechanism has been proposed to be involved in AS [200]. TNF-α, as well as IL‐1α, have been shown to modulate osteoprotegerin (OPG) and RANKL expression in osteoblasts [201, 202]. TNF-α can also directly stimulate 27 osteoclast differentiation in the presence of M‐CSF and RANKL [203]. In fact, TNF-α is thought to be implicated in the pathogenesis of osteoporosis observed in chronic inflammatory disease [204, 205]. Additionally, IL-1 has been shown to stimulate osteoclast differentiation and is required for their resorptive activity [206, 207]. Gut inflammation and microbiota in AS The human gut is populated with an overwhelming estimate of 100 trillion microbial cells [208]. Over millennia, a symbiotic relationship between microbiota and the human host has evolved with human hosts benefitting from protection against pathogenic microorganisms and obtaining vital nutrients, in exchange for providing the commensals with a favorable living environment. It is not surprising that commensal bacteria also play an important role in the shaping of the immune system, since they exist in such close proximity to the mucosal immune system. Germ-free animals, which lack microbiota in their gut, as a result, have underdeveloped lymphoid structures in the gut, including Peyer’s patches and gut-associated lymphoid structures [209]. They also have altered mucus layers, and have reduced CD8+, CD4+, Th17, Treg cells, and IgG levels. Their intestinal epithelial cells also have altered phenotype with a reduction in TLR and MHC-II levels. There is increasing evidence of microbiota playing a role in a variety of inflammatory and autoimmune human diseases that extend beyond the intestinal system, such as MS and IDDM [210]. 5-10% of AS patients have inflammatory bowel disease [107] and 46% of patients with spondyloarthritis have microscopic gut inflammation [211], suggesting involvement of the gut in the pathogenesis of AS. There has been an increased awareness given to the gut microbial health 28 in arthritic diseases, including AS in the recent years [212]. In 2014, the first study looking at gut microbiome in AS patients showed an increased presence of Prevotellaceae, Bacteroidaceae, Lachnospiraceae, Veillonellaceae and Porphyromonadaceae; and reduced abundance of Rikenellaceae and Ruminococcaceae species [213]. Since then several other studies have confirmed that AS patients have dysbiosis of the gut [214, 215]. However, it is unclear whether the dysbiosis is a symptom of the disease or the cause of the pathogenesis. AS therapies The mainstay therapies of AS are nonsteroidal anti-inflammatory drugs (NSAIDs) and physical therapy [107]. Patients who do not get relief of symptoms with NSAIDs are next placed on anti-TNF-α therapy. While both therapies are effective in reducing pain in 60% of patients, there is no convincing evidence in their ability to prevent radiographic progression of AS [107, 216]. Axial skeletal ankylosis and osteoporosis of vertebral bodies are common manifestations of AS [217]. Historically, the assessment of therapeutic efficacy in the patient uses clinical score assessments. The most recently adapted scoring system is Assessment of SpondyloArthritis International Society 20 (ASAS 20) improvement criteria, developed in 2001 [218]. This assessment method includes four main domains: the patient's global assessment of their health, pain, functionality, and inflammation, determined via standardized questionnaires. All clinical drug trials utilize this assessment to determine the efficacy of the tested drugs, where a successful therapy must achieve at least a 20% improvement in at least 3 domains. ASAS 40 requires an improvement of 40% or above in at least three domains to be considered a successful 29 therapy [219]. While alleviation of pain and other symptoms that contribute to the morbidity of AS is critical, the downside of using the clinical score assessment systems is that they do not provide us with information regarding the improvement of the skeletal abnormalities in AS patients. Therefore, it is important to include imaging modalities and blood work for evaluating the inflammatory status and the skeletal health in these patients in order to determine the therapeutic efficacy of the tested drugs relative to these major pathologies. While anti-TNF-α treatment is effective in reducing the clinical score of AS by reducing pain and inflammation 60 percent of patients [220], the efficacy of this therapy in regard to ankylosis and osteoporosis is under question. There is a discrepancy in reports regarding the efficacy of TNF-α blockade in improving the bone mineral density of the lumbar vertebra and prevention of vertebral fractures [221-224]. There is a similar discrepancy in regard to the efficacy of anti-TNF-α treatment in the prevention of syndesmophyte formation and ankylosis progression in AS patients [170, 225, 226]. A small open-label proof-of-concept trial showed human anti-IL23p40 antibody (ustekinumab) resulted in ASAS 40 improvement in 65 percent of AS patients and reduction in inflammation assessed by MRI [227]. Another clinical trial phase III testing anti-IL23p19 human monoclonal antibody is underway with results pending [228]. Data from phase III clinical trial has shown that anti-IL17A antibody (secukinumab) was successful in improving clinical score ASAS 40 in 60% of patients [229]. However, it is unclear whether this therapy improved inflammation or the ankylosis as no imaging modalities were used for evaluation of the spine. 30 There are other therapies that were tested in AS patient but did not show sufficient therapeutic efficacy. Blockade of IL-1 signaling by using anakinra IL1 receptor antagonist showed improvement in the clinical score of ASAS 40 only in 20 percent patients, and failed to improve inflammation as assessed by the MRI imaging and blood work [230]. Similarly, blockade of IL-6 signaling using humanized antibodies against IL-6R failed to show significant clinical improvement in AS patients with only 37 percent of patients achieving ASAS 20 [231]. Animal models of AS Multiple animal models of AS have been described to date [232-235], however, none of them successfully portray all of the manifestations seen in AS patients. HLA*B27 transgenic Lewis rats, which contain 150 copies of HLA*B27 and 90 copies of human β2m in order to observe a spontaneous phenotype [236, 237], primarily develop abdominal symptoms manifested as diarrhea and intestinal inflammation. Peripheral arthritis of the paws, ankylosis of the tail, epididymo-orchitis, and gut dysbiosis were also described in these animals [238]. More recently HLA*B27 rats expressing 55 copies of HLA-B27 and 29 copies of human β2m have been described to develop a similar disease, predominated by intestinal inflammation and accompanied by lumbar osteopenia, but no axial ankylosis was detected [239]. It is worth noting that only gastrointestinal manifestations develop in all of the animals, whereas arthritis develops in a portion of animals. Interestingly, animals with a lower copy number of HLA-B27 and β2m do not develop the disease, for example, the strain containing 6 copies of each gene are disease free. Homozygous animals with 20 copies of HLA*B27 and 15 copies of b2m developed intestinal and urogenital manifestation of the disease, while heterozygotes remain disease free [236]. The major downside of this animal model is that it requires overexpression of non- 31 physiological levels of HLA*B27 and β2m to see any phenotype, and even so, the main phenotype is gross intestinal inflammation rather than skeletal ankylosis, suggesting that these animals display a different disease process from that of ankylosing spondylitis. HLA*B27 transgenic mice (HLA*B27/hβ2m/mβ2m-/-) develop nail changes and peripheral arthritis limited to the hind paws and manifested as gross swelling and inflammation, which are induced in 14-66% of animals after they are transferred from specific pathogen free (SPF) environment to conventional area [240, 241]. This animal model fails to induce axial skeletal ankylosis and produces peripheral inflammation in an unpredictable manner. Our lab has not been able to reproduce consistent induction of phenotype in these mice. Homozygous mice expressing truncated ankylosis gene (ank) protein develop a progressive mineralization of all joints in the body, which starts out in the joints of the feet at 3 weeks of age and progress to the spine by 7 weeks of age resulting in severe stiffness and immobility [242]. Ank/ank homozygous mice develop the disease in 19-24 % of mice with 26% neonatal mortality and adult mortality by 6 months of age [243]. These mice have been proposed to resemble the skeletal disease observed in osteoarthritis and ankylosing spondylitis. In addition to the joint mineralization, these mice develop skin scaling on the surface of their paws, resembling psoriatic arthritis. The synovium from the affected joints is positive for hydroxyapatite crystals, which are thought to trigger the destruction of osteoarticular structures, joint narrowing and compensatory mineralization of the affected joints. Although this model does have some resemblance to human AS, the ankylosis in these mice is systemic and non- specific, whereas in AS, axial spine is preferentially affected in a significantly slower progression. AS patients have not been reported to have hydroxyapatite crystals in their 32 synovium, suggesting a major difference in the pathogenesis of the skeletal disease in ank/ank mice compared to AS. SKG mice are BALB/c mice that possess a missense mutation in ZAP-70, and spontaneously develop progressive peripheral arthritis closely resembling RA, caused by arthritogenic T cells [244]. When injected with curdlan (β-1,3-glucan) around 2 months of age, these arthritis-prone mice develop arthritic changes in the sacroiliac and peripheral joints, accompanied by erosions and syndesmophytes. They also develop enthesitis of the Achilles tendon, vertebral inflammation, unilateral uveitis and inflammation of the ileum closely resembling features of spondyloarthritides [245]. While this animal model does develop features seen in AS, it is questionable whether the pathogenesis induced by injection of bacteria-derived adjuvant is representative of the processes in the human disease. In a more recently described animal model, IL-23 overexpression in mice has been shown to induce aortic arch inflammation and more importantly, inflammation at the enthesis of the SI and intervertebral joints that are thought to be the primary sites of inflammation which progress to ankylosis in AS [127]. This animal model confirmed the hypothesis that IL-23 plays a role in the entheseal inflammation by inducing IL-17 production. This study also identified a new cell type possibly responsible for the arthritic and osteoproliferative changes leading to ankylosis. However, this model fails to mimic physiological conditions of AS, as it required systemic overexpression of IL-23 to induce the phenotype. Finally, mice overexpressing transmembrane TNF-α (tmTNFtg) has been shown to develop synovial inflammation and joint destruction at the peripheral joints, and spinal abnormalities [246]. While tmTNFtg mice have been proclaimed to develop skeletal ankylosis, 33 this was concluded based on the observation of crinkling of the tails; however, it was never confirmed via imaging such as μCT. This model provides an interesting insight into the role of TNF-α in the skeletal pathogenesis of AS. Although these animal models all have some similarities to the manifestations of AS in humans, none of them develop all key features of AS spontaneously. In chapter 2 we describe in detail the phenotype of ERAP1-/- mice, which are the first animal model described that spontaneously develops all key intestinal, immunological and most importantly skeletal features of AS in human patients. In chapter 2 we attempted to decipher the role of intestinal microbiota in the immunological and skeletal disease processes observed in ERAP1-/- mice with the goal of uncovering the role of microbiota in AS. Additionally, we surveyed the immune system of the ERAP1-/- mice with the goal of identifying a potential cellular culprit responsible for the phenotypes observed in ERAP1-/- mice. Based on the fact that osteoclasts and osteoblasts are involved in bone remodeling and that their functions can be influenced by proinflammatory cytokines elevated in AS patients and ERAP1-/- mice, we attempted to unravel the mechanism of the skeletal pathogenesis in these animals by surveying the bone remodeling cellular functions, as described in chapter 3. We also performed a therapeutic study aimed at determining the effect of anti-TNF-α and anti-IL-17 antibodies on osteoporosis and ankylosis observed in ERAP1-/- mice. Harnessing auto-immunity: tumor immunotherapy In-depth knowledge of the immune system is important for understanding the underlying processes in autoimmune and inflammatory diseases, where the goal is to develop therapies 34 halting immune responses, either by inhibiting effector cell functions or by harnessing inhibitory pathways and regulatory cell functions. On the other hand, a detailed understanding of the immune pathways can also be used to enhance immune responses in diseases fueled by immune suppression, such as cancer. It is well-accepted that the tumor microenvironment is immunosuppressive. The tumor microenvironment inhibits maturation of APCs and promotes activation of inhibitory pathways and recruitment of immunosuppressive T cells thereby allowing tumors to escape the immune surveillance [247]. In the 1990’s it became apparent that blockade of inhibitory T cell receptors, such as Cytotoxic T-Lymphocyte Associated Protein 4 (CTLA-4) allowed for enhanced anti-tumor responses in mice, tumor rejection and long-term immunity [248]. CTLA-4 acts as a checkpoint receptor. It is expressed on the surface of T cells in response to activation and competes with CD28 for B7 receptors, thereby halting T cell activation [249]. Anti-CTLA4 antibody prevents T cell inhibition and has been shown to be effective in a variety of tumor types [250, 251]. It even showed success in anti-tumor vaccines. The preclinical success of anti-CTLA4 therapy has been translated to human studies, where this therapy has shown regression of melanoma, renal cell carcinoma, prostate, urothelial and ovarian cancers [252]. Humanized anti-CTLA antibody (ipilimumab) is an FDA approved treatment for melanoma, which substantially improves 10-year survival. The success of anti-CTLA4 therapy was the dawn of what has now become known as “immune checkpoint therapy”, and paved the way for other therapies targeting intrinsic inhibitory pathways of T cells [253]. Another promising target for immune checkpoint therapy is programmed cell death protein 1 (PD-1) [254, 255]. PD-1, which is expressed on activated T cells, interacts with PD-1 ligands (PD-1L) expressed on APCs, resulting in inhibition and 35 apoptosis of T cells. Blockade of PD-1 and PD-L1 have shown success in melanoma, renal cell carcinoma and other tumors in human trials [256, 257]. Moreover, combination therapies targeting both, CTLA-4 and PD-1 have been shown to substantially improve tumor regression in several tumor types compared to individual therapies [258]. Currently, many other “immune checkpoints” are being investigated as targets in cancer immunotherapy, including other inhibitory T cell receptors such as Tim-3, Lag-3, the co- stimulatory molecules such as OX40 and 4-1BB are each targeted by blocking antibodies or agonistic antibodies, respectively [252]. Moreover, it is apparent that combination therapies allow for improved efficacy of these agents in cancer immunotherapy, due to synergistic effects of targeting multiple pathways simultaneously [259, 260]. Immune checkpoint inhibitors hold a great potential in the field of cancer therapies, and it is important to continue to identify new targets with the goal to further our understanding of how anti-tumor immune responses can be harnessed and strengthened in our search for curative therapies. SLAM receptors Signaling lymphocytic activation molecule (SLAM) family of receptors are immunomodulatory receptors that have been of particular interest due to their roles in autoimmunity, lymphoproliferative disease, humoral responses, lymphocyte development and survival, cytotoxicity and T cell signaling [261]. SLAM family includes six members: SLAMF1 (SLAM, CD150), SLAMF3 (Ly-9, CD229), SLAMF4 (2B4, CD244), SLAMF5 (CD84), SLAM6 (NTB-A, Ly108) and SLAMF7 (CRACC, CD319, CS1) [262]. These receptors are expressed on the surface of hematopoietic cells. With the only exception being SLAMF4, these 36 receptors are homotypic, meaning that these receptors can serve as self-ligands. Upon activation through the interaction of the extracellular domains of these receptors, intracellular adaptors are recruited to their cytoplasmic domains that contain immunoreceptor tyrosine-based switch motifs (ITSMs), thereby activating or inhibiting the intracellular signaling cascade, depending on the cell type and the adaptor protein [263]. The three adaptor SLAM-associated proteins (SAP) are SAP, Ewing’s sarcoma-associated transcript 2 (EAT-2) and EAT-2-related transducer (ERT). While EAT-2 and ERT signal via phosphorylation of the tyrosine residues on the C-terminal cytoplasmic domain, SAP adaptor recruits FynT tyrosine kinase [264]. SLAM receptors can also signal via SH2 domain-containing phosphatases SHP-1, SHP-2, and SHIP-1 in absence of adaptor molecules [262]. The SLAMF7 receptor, also known as CD2-like receptor activating cytotoxic cell (CRACC) is expressed on NK cells, macrophages, dendritic cells, and activated T and B cells [265-268]. In the presence of the EAT-2 adaptor protein, CRACC-CRACC interactions result in activating immune responses, while in the absence of EAT-2, CRACC interaction has inhibitory effects as has been shown in the EAT-2 negative subset of NK cells [269], and in T cells [267]. High levels of expression of SLAMF7 on malignant plasma cells in multiple myeloma made it an attractive target for therapeutic approaches. The humanized anti-SLAMF7 monoclonal antibody, elotuzumab, administered in combination with dexamethasone and lenalidomide has been shown to be effective in reducing disease progression and mortality in patients with refractory multiple myeloma in phase III clinical trials [270, 271]. Interestingly, although SLAMF7 is highly expressed on normal and myeloma plasma cells, elotuzumab is thought to work via its binding to the CD16 receptor on NK cells, activating ADCC mediated tumor cell killing [272]. Additionally, elotuzumab was shown to promote NK cell cytotoxicity against 37 SLAMF7+ multiple myeloma cells by directly binding to the SLAMF7 receptor and activating the NK cells. In this work, we utilized a mCRACC-Fc fusion protein to target SLAMF7 on a variety of immune cells. We hypothesized that engagement of SLAMF7 on NK cells will enhance their cytotoxic function and increase tumor killing. We also hypothesized that mCRACC-Fc fusion protein will bind to and occupy the SLAMF7 receptors on the APCs, preventing them from interacting with SLAMF7 on T cells, thereby preventing inhibitory effects of SLAMF7 in T cells. In chapter 4, we describe in detail the therapeutic benefit of mCRACC-Fc in established CT26 colon adenocarcinoma tumors. Use of adenoviruses as vectors One method for allowing high-level expression of a foreign protein, such as mCRACC- Fc fusion protein, is via the use of gene transfer vectors, including the adenovirus (Ad) based vectors. Adenovirus is the most commonly used vector for gene transfer therapy and vaccinations [273]. It is a non-enveloped, double stranded DNA virus with an icosahedral protein capsid, which generally causes a self-limiting course of infection with mild respiratory manifestations [274]. Although there are over 57 serotypes of human Ads, replication-deficient Ad5 is one of the most commonly used vectors [275]. The Ad genome is organized into four early (E1- E4) and five late (L1-L5) transcription units. In the replication-defective Ad, the E1A and E1B genes are deleted and replaced with an expression cassette, where the desired transgene can be inserted under an active promoter, such as cytomegalovirus (CMV) promoter. E1A encoded genes that are essential for replication and are responsible for the induction of delayed 38 genes encoded by E1B, E2, E3 and E4 regions. E1B proteins are important for host cell apoptosis of Ad-infected cells. E1A and E1B deleted vectors (E1- Ads) can be grown in HEK293 cells lines, which express the E1A and E1B encoded genes. The first generation of Ad vectors also has the E3 region deleted in addition to E1A and E1B (E1-, E3- Ads). E3 genes are not required for replication but encode proteins important for protecting the infected cells from being eliminated by the immune system. The first-generation Ads can exhibit leaky expression of delayed early and late genes, making them susceptible to recognition and elimination by T cell- mediated responses, which can ultimately reduce their ability to allow long-term expression of the transgene they express. Newer generation Ad vectors have more gene regions deleted that allows for an increased size of the transgene to be inserted and can also reduce leaky expression of Ad genes [276, 277]. Ads are very immunogenic. During intravenous injections in mice, Ads primarily infect the liver, where they are mainly taken up by the Kupffer cells [278]. Ad5 binds to the coagulation factor X and the complex is taken up by the scavenger receptor-A on the Kupffer cells. Factor X-Ad5 complex also triggers an immune response by binding to TLR4 [279]. It is known that Ad5 can also activate TLR2 and TLR9 on lymphoid cells [280]. Ad-mediated activation of the TLRs signals through the MyD88 adaptor and triggers NF-kB activation, which induces proinflammatory cytokine and chemokine secretion. Within first several hours of infection, various proinflammatory cytokines such as IL-1α, IL-1β, TNF-α, IL-6, IL-12, and IFN-γ as well as various chemokines, such as RANTES, MCP-1, KC, MIP-1α, MIP-1β, and CXCL10 are secreted [281]. Ad5-based vectors can also activate complement by binding to C4 and C4BP proteins in the classical and alternative complement pathways [282]. Ad DNA is also capable of activating inflammasome pathways [283]. Additionally, Ads are capable of improving 39 antigen-specific adaptive immune responses, which makes them a great tool for vaccine development [284]. Some of the reasons why Ads are the first-choice vector for clinical application include the existence of robust protocols for growing Ad vectors in high quantities, their efficient transduction of a variety of cell types, their safety profile associated with the fact that they do not integrate into host genome, and their capability to carry large DNA segments [275]. To date, Ad vectors have been tested in over 500 clinical trials with a majority of them tested in anticancer treatments. In chapter 4 of this dissertation, we utilized an Ad5 [E1-E3-] vector to transfer the mCRACC-Fc gene intratumorally into colon adenocarcinoma challenged mice with an attempt to deliver high levels of the gene directly into tumors in order to improve tumor immunogenicity. Additionally, we tested the efficacy of vaccinating animals with rAd5-mCRACC-Fc in combination with tumor antigens in order to build tumor-specific memory. 40 Chapter 2: ERAP 1 deficient mice have reduced type 1 regulatory T cells and develop skeletal and intestinal features of ankylosing spondylitis. This chapter is the edited version of the research article that was published in the Scientific Reports, volume 8, Article number: 12464, Aug 20, 2018. Authors: Yuliya Pepelyayeva*, David P.W. Rastall*, Yasser A. Aldhamen, Patrick O'Connell, Sandra Raehtz, Fadel S. Alyaqoub, Maja K. Blake, Ashley M. Raedy, Ariana M. Angarita, Abdulraouf M. Abbas, Cristiane N. Pereira-Hicks, Sarah G. Roosa, Laura McCabe, and Andrea Amalfitano *Both authors contributed to work equally. Author Contributions YP, DPR, and AA designed the experiments. Experiments in figure 1 were carried out by YP, SR, and DPWR; figure 2: YP, DPWR, and SR; figure 3: YP and DPWR; figure 4: YAA, FSA, AMR, AMA2, DPWR, PO, and YP; Figure 5: YP, AMA1, SGR, CNP, PO, and MKB; figure 6: YP, AMA1, SGR, CNP, PO, and MKB; figure 7: YP and AMA1; Figure 8: YP, PO, and MKB; figure 9: YP, PO and MKB. AA and LM supervised the project. YP and AA drafted the manuscript and all authors assisted in editing the manuscript. 41 Introduction Ankylosing spondylitis (AS) is a highly heritable autoimmune disease with an estimated prevalence of 1% worldwide [102]. AS is characterized by early-onset fusion of bones in the spine and pelvis, referred to as ankylosis [103]. Many AS patients also develop osteopenia or osteoporosis in both the axial and the peripheral skeletons, predisposing these individuals to pathologic fractures [105, 108]. Over 40 years ago, the genetic association of the Human Leukocyte Antigen - HLA-B*27 allele with AS was found, directly implicating immune mechanisms in AS susceptibility [117]. Despite this knowledge, the exact mechanism underlying pathogenesis of inflammation and spinal fusions in AS is still not understood. For example, while over 90% of AS patients have an HLA-B*27 variant, only 1-5% of individuals carrying HLA-B*27 develop AS, suggesting that additional risk factors must be present [103]. While HLA-B*27 is thought to contribute 23% of the genetic risk for AS, there are strong epistatic gene-gene interactions between the presence of specific ERAP1 variants and HLA-B*27 [140]. As one of ERAP1’s primary functions is to trim peptides prior to their loading onto MHC-I molecules, epistasis between ERAP1 and HLA alleles supports the hypothesis that deviations in antigen presentation pathways may underlie AS pathogenesis [118, 141]. Additionally, our work and that of others have also implicated ERAP1 in the suppression of innate and adaptive immune responses [153, 154, 163]. Adding to this genetic and immunologic complexity, gastrointestinal abnormalities are also thought to be involved in AS pathogenesis [285]. Patients with colitis are three times more likely to develop clinical AS, [286] and up to 60% of AS patients manifest microscopic inflammatory damage to their gut mucosa [108, 285]. It has also been found that AS patients have altered microbiota in the terminal ileum, [213] 42 indirectly suggesting that microbial interactions with the intestinal immune system may contribute to extra-intestinal immune abnormalities and pathogenesis of AS [212, 287, 288]. We wished to investigate the potential role of ERAP1 in the skeletal and intestinal pathogenesis of AS by critically evaluating the skeletal, intestinal and immunological phenotypes of ERAP1-/- mice. Here we show that deficiency of ERAP1 was sufficient to cause a spontaneous and rapid onset of spinal ankylosis, calcification of the anterior longitudinal ligament (ALL), sacroiliac (SI) erosions, and systemic osteoporosis in mice, making this a valuable model to use for studying the pathogenesis of these skeletal abnormalities that are also found in human AS. Moreover, ERAP1-/- mice developed intestinal features including spontaneous dysbiosis and increased susceptibility to dextran sodium sulfate (DSS)-induced colitis, paralleling intestinal features of AS. In our survey of the immune system of ERAP1-/- mice, we evaluated two major populations of suppressive T cells including Foxp3+Tregs and Type 1 regulatory T cells (Tr1s) [64]. While splenic Foxp3+Treg numbers and function did not differ between WT and ERAP1-/- mice, we detected a significant reduction in the numbers of CD3+CD4+ Foxp3-IL-10+ “Tr1-like” cells present peripherally in ERAP1-/- mice. In addition, we observed reduced numbers of tolerogenic dendritic cells (tDCs), which are thought to be important for Tr1 differentiation and function [86]. 43 ERAP1 deficiency causes spinal ankylosis and osteoporosis. Results ERAP1 SNPs have been genetically linked to increased susceptibility of AS [141]. We first surveyed the spinal morphology of ERAP1-/- mice for skeletal manifestations observed in human patients, such as sacroiliitis, syndesmophyte bridging, joint erosions and/or osteoporosis of the spine [108]. We performed μCT analysis of the axial skeletons of 14-weeks-old WT and ERAP1-/- mice and observed that ERAP1-/- mice exclusively developed spinal ankylosis between the transverse processes of L6 (lumbar vertebrae 6) and S1 (sacral vertebrae 1) [Fig. 1a, b]. We developed a scoring system ranging from 0 to 4, with 0 representing a normal L6-S1 joint, and 4 representing a joint architecture in which an abnormal syndesmophyte was extended along the full border of the iliac bone and was completely fused with S1 [Fig. 1a]. The left and right sides of this joint were scored independently, and the scores were averaged to get the mean ankylosis score for each individual mouse. Our detailed skeletal survey of ERAP1-/- mice showed that ERAP1-/- mice not only exclusively developed spinal ankylosis between L6 and S1 [Fig. 1b-c], but also developed other hallmarks of human AS, such as iliac erosions [Fig. 2a, b] and calcification of the anterior longitudinal ligament [Fig. 2c]. While there was no significant difference in the severity of fusions between male and female mice, we observed a clear trend in reduced severity of fusions and reduced prevalence of fusions in female mice, a result which parallels known gender differences for the human disease [102] [Fig 1g]. 44 Figure 1: µCT analysis of the axial skeleton. The spines from ERAP1-/- and WT mice were harvested and analyzed with µCT. (a) Scoring system with representative images from ERAP1-/- mice. (b) Representative isosurface images, demonstrating spinal ankylosis of the L6 to S1 vertebra in ERAP1-/- mice. Red arrows indicate the L6 vertebra. White arrows indicate remnants of the transverse process of L6. 14-weeks-old WT (n=10) and ERAP1-/- (n=16) male and female mice were assessed for ankylosis bilaterally. Scores were averaged, and Mean Ankylosis Scores were plotted on bar graph (c). (d) Representative isosurface images of the trabecular bone of the S1 vertebra. (e) Graph depicting bone mineral density (BMD) of the trabecular bone of S1 vertebra of 14-week-old female mice (n=5). (f) AS index was calculated by combining the fusion and an assigned BMD score, as described in methods, and graphed as a function of time. (g) Fusion scores of 14-weeks-old ERAP1-/- mice graphed as a function of sex. Values on graphs are the mean ± SEM. p values calculated by the Student's t-test, where *** - p<0.001, **** - p<0.0001 and NS- not significant. 45 To assess ERAP1-/- mice for osteoporosis of the spine, we analyzed the trabecular bone morphology of 14-week-old ERAP1-/- mice utilizing quantitative micromorphometry and MicroView software. The bone mineral density (BMD) of the S1 vertebral body in ERAP1-/- mice was significantly reduced compared to WT [Fig. 1d, e]. Other measures of the trabecular bone structure including bone volume fraction (BV/TV %), bone mineral content (BMC), trabecular number (Tb.N), trabecular thickness (Tb.Th), and trabecular spacing (Tb.Sp) confirmed the presence of an osteoporotic phenotype in ERAP1-/- mice [Fig. 2e-i]. We observed similar findings in the trabecular bone of L6 vertebral body and femurs of ERAP1-/- mice [data not shown]. Assessment of the trabecular bone of S1 vertebrae at different ages revealed that ERAP1-/- mice failed to gain maximal bone volume fraction at 14 weeks of age, which continued to be significantly lower compared to age-matched WT mice throughout their lifetime [Fig. 2d]. We observed the same trends in other trabecular measures [data not shown]. We developed a quantitative AS severity scoring system (AS Index), which combined ankylosis score and an assigned bone mineral density score, where 325 – 300 mg/cc received score 1, 299-275 – 2, 274- 250 – 3 and <250 – 4. The AS index was also used to evaluate disease progression between ERAP1-/- and WT mice over time [Fig. 1g]. The AS index was significantly increased in ERAP1- /- mice as compared to WT at 14 and 80 weeks of age, with some evidence of progression in severity of this score over time. This is the first report, to our knowledge, identifying a robust animal model in which young animals simultaneously present with spontaneous axial ankylosis and reduced bone density. 46 Figure 2: µCT analysis of the sacral region. The spines from 84-week-old female ERAP1-/- and WT female mice (n=8) were harvested and analyzed with µCT. (a) Representative 2D radiograph demonstrating erosions (black arrows) at the SI joint. (b) SI joints were assessed for erosions bilaterally, averaged per side and graphed. (c) Representative images demonstrating calcification of the anterior longitudinal ligament between L6 and S1 in ERAP1-/- animals, depicted with blue arrows. The spines from 14-week-old ERAP1-/- and WT female mice (n=5) were harvested and analyzed with µCT. (d) Graph comparing BV/TV% of the trabecular bone of S1 vertebra between WT and ERAP1-/- mice at various ages. Bar graphs demonstrating BV/TV (e), BMC (f), Tb.Th. (g), Tb.Sp. (h), and Tb.N. (i) of the trabecular bone of the S1 vertebra. Values on graphs are the mean ± SEM. p values calculated by the Student's t-test, where * - p<0.05 and ** - p<0.01 compared to age-matched WT mice. 47 Deficiency of ERAP1 results in recruitment of immune cells to spinal joints. We wished to evaluate the structure and integrity of the ankylosed joints in ERAP1-/- mice first using H&E staining. L6/S1 joints of WT mice demonstrated a normal architecture of concentrically lamellated annulus fibrosus (AF) and nucleus pulposus (NP), with characteristic cells and matrix [Fig 3a-c]. In contrast, the ankylosed L6/S1 joints from age-matched ERAP1-/- mice demonstrated severe pathology [Fig. 3d-f]. The NP structure was replaced by a dense cellular infiltrate dominated by mononuclear cells with large cytoplasm and heterochromatic nuclei [Fig. 2f]. The structure of AF was disrupted, and NP/AF interface was surrounded by ectopic bone [white arrows in Fig. 3e] and resident osteocytes [black arrows Fig. 3e]. To determine if the infiltrative mononuclear cells were of immune origin, we performed immunohistochemistry for IgM, F4/80, CD3, IL23, and TNF-α markers. IgM, F4/80, IL23, and TNF-α positive staining was detected in the NP of ERAP1-/- mice by 14 weeks, compared to age- matched WT mice [Fig. 3g-n]. IgM+ and F4/80+ staining suggests the presence of IgM deposits [Fig. 3h] and macrophages [Fig. 3j] in the NP, respectively. CD3 staining was negative, suggesting an absence of T cells (data not shown). The cytoplasm of large mononuclear cells in the NP of ERAP1-/- mice stained positive for IL-23 [Fig. 3l] and TNF-α [Fig. 3n] signifying the presence of abnormal local inflammatory responses in the L6/S1 intervertebral disc (IVD) of ERAP1-/- mice. 48 Figure 3: Histopathologic evaluation of ankylosed intervertebral joints. Spines from 84- week-old WT and ERAP1-/- female mice (n=3) were harvested, fixed, and stained. (a) Representative 10x H&E image of a WT L6/S1 joint demonstrating healthy annulus fibrosus (b), and nucleus pulposus (c). (d) Representative H&E of an ERAP1-/- L6/S1 joint demonstrating disruption of the disc by a large infiltrative mass of cells (black arrow and f) and ectopic bone (white arrows and e) within the joint. (e) 60x magnification of intervertebral bone formation with black arrows demonstrating osteocytes. Spines from 14-week-old male and female mice (n=3) were harvested, fixed, and stained with anti-IgM, IL-23, TNFα, and F4/80 antibodies. Immunohistochemistry representative images at 20X magnification demonstrating IgM+ deposits (white arrows) within the nucleus pulposus of the L6/S1 joint in ERAP1-/- (h), but not in WT mice (g). Some cellular infiltrates in the nucleus pulposus within the L6/S1 joints of ERAP1-/- mice stained positive for F4/80+ (white arrows), indicating macrophages (j), but not in WT mice (i). IL-23 positive staining of the L6/S1 nucleus pulposus infiltrate in ERAP1-/- (l), but not 49 Figure 3 (cont’d) in WT (k). TNFα positive staining of the nucleus pulposus in ERAP1-/- (n) but not in WT L6/S1 joint (m). 50 ERAP1 deficient mice have increased susceptibility to chemically-induced colitis. In addition to skeletal involvement, AS patients are known to have an increased susceptibility to inflammatory bowel disease (IBD), with 5-10% of AS patients developing IBD, [108] and up to 60% of AS patients having evidence of subclinical gut inflammation in the terminal ileum [108, 285]. To determine if ERAP1 functions impact the intestinal system to a degree that may influence the development and severity of colitis, we assessed the mortality rate of age- and sex-matched WT and ERAP1-/- mice after 3% DSS challenge, which is known to induce intestinal inflammation and colitis by altering the integrity of the intestinal epithelial barrier [289, 290]. We observed a significantly increased mortality rate of DSS-treated ERAP1-/- mice, beginning as early as 6 days after DSS challenge as compared to identically treated WT mice [Fig. 4a]. Only 27% of ERAP1-/- mice survived the challenge as compared to 53% survival in the WT controls (p<0.01) [Fig. 4a]. Mice were also monitored for weight loss, stool consistency, and rectal bleeding scores, all of which were used to calculate an overall colitis disease activity index (DAI). The DAI was significantly increased in DSS-treated ERAP1-/- mice, as compared to WT mice starting on day 4 following DSS administration (p<0.05) [Fig. 4b]. Histologic colon sections taken from WT and ERAP1-/- naïve mice on day 7 showed limited evidence of lymphocytic infiltration of the mucosa and submucosa, and little to no evidence of focal erosions or ulcerations [Fig. 4c]. In contrast, colonic sections of DSS-treated ERAP1-/- mice displayed extensive ulceration and inflammation with foci of proliferating lymphocytes [Fig. 4c]. Utilizing a previously described semi-quantitative scoring system [291], we confirmed the presence of significantly (p<0.01) increased inflammatory signs of colitis in DSS-treated ERAP1-/- mice, as compared to WT controls [Fig. 4d]. These results suggest that the changes observed in DSS-challenged ERAP1-/- mice were not due to anatomic differences prior to the 51 challenge, but rather due to abnormal immune responses to colonic irritation, findings consistent with other models of IBD [289, 290]. 52 Figure 4: Induction of chemically-induced colitis. 8-week-old male WT and ERAP1-/- mice (n=11) were treated with a 3% DSS solution in drinking water for 7 days, followed by regular drinking water for 11 days. The DSS solutions were made fresh at day 0 and 3 and left until day 7. Mice were monitored until day 18 after the start of DSS. (a) Survival plot of WT and ERAP1-/- DSS-treated mice. Differences in group survival were analyzed with the Kaplan-Meier test. Disease Activity Index of WT and ERAP1-/- mice was determined as described in materials and methods and graphed in (b) Two-way ANOVA with Bonferroni test were used to determine significant differences. (c) Representative images of H&E staining for histopathologic changes in colon tissue of naïve and DSS-treated WT and ERAP1-/- mice (n=4). Semiquantitative scoring of histopathology of colons was performed, as described in Materials and Methods section and plotted as a bar graph in (d). Data are expressed as the means ± SEM. p values calculated by the Student's t-test, where * - p<0.05, ** - p<0.01 and **** - p<0.0001. Data are representative of three independent experiments, with similar results. 53 ERAP1 deficient mice develop spontaneous gut dysbiosis. While there is strong evidence suggesting that the intestinal microbiota plays a role in modulating immune diseases that affect the intestine, such as IBD [292], there is less information available in regard to the direct role of the gut microbiota in causing arthritic autoimmune diseases, such as AS [212, 238]. Recently, dysbiosis of the terminal ileum has been noted in AS patients, suggesting that the microbiota may play a role in AS [213]. We undertook genetic phenotyping of the fecal microbiomes of similarly-housed ERAP1-/- and WT mice to determine if spontaneous dysbiosis was present in ERAP1-/- mice. Microbial communities present in the fecal samples of mice were determined by analysis of the V4 hypervariable region of the 16S rRNA gene. A non-metric multidimensional scaling analysis demonstrated that ERAP1-/- mice have a significantly altered fecal microbiome composition (p<0.05), as compared to similarly- housed WT mice [Fig. 5a, b]. Plotting of aggregated OTUs into phylum-level taxa revealed that fecal samples derived from 14-week-old ERAP1-/- mice were significantly enriched for Cyanobacteria (p<0.05) and Actinobacteria (p<0.05) [Fig. 6]. Genus-level analysis of ERAP1-/- samples revealed enrichment of Prevotella, Odoribacter, Bacteroides, and to lesser extent YS2, Parabacteroides and Clostridiales as compared to fecal samples from similarly-housed and age- matched WT mice [Fig. 5e and Table 1]. In addition, ERAP1-/- mice were deficient in the bacterial genus Lachnospiraceae and to a lesser extent Christensenellaceae and S24.7 [Fig. 5e and Table 1]. These results confirm that ERAP1-/- mice exhibit another phenotype found in AS patients, namely spontaneous intestinal dysbiosis [213]. 54 Figure 5: Evaluation of gut microbial composition and cross-foster experiments. 16s rRNA phenotyping was performed with Illumina on fecal samples of 14-week-old cross-fostered male and female mice (n=5). (a) A principal component analysis of the composition of the gut microbiome was represented by summarizing OTU abundances into Bray-Curtis dissimilarities and performing a non-metric multidimensional scaling (NMDS) ordination. To quantify the differences between the groups, PERMANOVA was performed. (b) A table summarizing Bonferroni p-values. Spines of 14-week-old cross-fostered male and female mice (n=6) were surveyed for spinal fusions and osteoporosis using μCT. Mean Ankylosis Score (c) and BMC (d) of the trabecular bone of S1 vertebra graphed as the mean ± SEM. Statistical significance was determined using One Way ANOVA, and Tukey post hoc test, where * - p<0.05, ** - p <0.01, *** - p<0.001, and NS – not significant. Data are expressed as the means ± SEM. Data are representative of three independent experiments, with similar results. (e) Relative abundance of genera in fecal samples of WT, CF WT, ERAP1-/-, and CF ERAP1-/- mice. Genus-level OTUs 55 Figure 5 (cont’d) significantly different between WT and ERAP1-/- fecal samples are summarized in Table 1. The microbiome data is a representative of two independent experiments with similar results. Table 1: Summary of the bacterial genera that were significantly different between WT and ERAP1-/- mice. Increased in ERAP1-/- Genus Bacteroides Clostridiales_unclassified Odoribacter Parabacteroides Prevotella YS2_unclassified Decreased in ERAP1-/- WT vs ERAP1-/- ERAP1-/- vs CF ERAP1-/- WT vs CF ERAP1-/- **** * *** * **** * NS ** *** NS * * ** NS NS * NS NS Genus WT vs ERAP1-/- ERAP1-/- vs CF ERAP1-/- WT vs CF ERAP1-/- Christensenellaceae_unclassified * Lachnospiraceae_unclassified ** S24.7_unclassified * NS NS NS * NS NS Table summarizing differences of the genus-level OTUs that were found to be significantly increased (top half of the table) or decreased (bottom half of the table) in fecal samples of ERAP1- /- mice compared to WT. Differences between ERAP1-/- and CF ERAP1-/- and WT and CF ERAP1-/- mice are summarized in columns 2 and 3, respectively, with arrows representing the direction of change in CF ERAP1-/- fecal samples. p values calculated by the Student's t-test, where * - p<0.05, ** - p<0.01, *** - p<0.001, **** - p<0.0001 and NS - not significant. 56 We next wished to determine if the intestinal microbiome plays a role in the development of skeletal pathology in ERAP1-/- mice. We undertook cross-fostering experiments, which have been shown to be an effective method of inducing an early and maintained shift in the commensal microbiota to that of a nursing mother [293]. Fecal 16S rRNA phenotyping of ERAP1-/- mice cross-fostered to WT dams (CF ERAP1-/-) showed significant changes in their microbiome compared to ERAP1-/- samples (p<0.05) [Fig. 5b], with normalization of Cyanobacteria and Actinobacteria phyla [Fig. 6]. We also observed correction of Prevotella, Odoribacter, Clostridiales, Lachnospiraceae, YS2, and S24.7 genus, but not Bacteroides, Parabacteroides, and Christensenellacea [Fig. 5e and Table 1]. Despite these significant alterations in the intestinal microbiota, μCT analysis of CF ERAP1-/- did not show any changes in their ultimate development of ankylosis [Fig. 5c] or osteoporosis [Fig. 5d] during the time frame of these experiments. Similarly, CF ERAP1-/- mice still had TNF-α, IL-23, and F4/80 positive staining, but not IgM in the NP of their L6/S1 disc spaces [Fig. 7]. Confirming the utility of cross-fostering, WT pups cross-fostered to ERAP1-/- dams (CF WT), indeed clustered similarly to ERAP1-/- samples [Fig. 5a]. Four cross-fostered mice created a cluster of their own, with the biggest contribution from Akkermansia genus, increasing the variability within cross- foster groups. Despite the transfer of the aberrant microbiota derived from ERAP1-/- mice to CF WT mice, no ankylosis or significant changes in the bone density measures were observed in the latter [Fig. 5c, d]. 57 Figure 6: Relative abundance of phyla in WT, CF WT, ERAP1-/- and CF ERAP1-/- fecal samples. 16s rRNA phenotyping was performed with Illumina on fecal samples of 14-week-old male and female mice (n=5). Stacked bar graph showing relative abundance of phyla in fecal samples of WT, CF WT, ERAP1-/-, and CF ERAP1-/- mice. Figure 7: Histopathological analysis of the ankylosed intervertebral joints from CF ERAP1-/- mice. 14-week-old cross-fostered ERAP1-/- female and male mice were sacrificed. Their spines were harvested, fixed, decalcified and stained. Representative 58 Figure 7 (cont’d) immunohistochemistry images of the L6/S1 intervertebral disc at 20X magnification for TNFα (a), IgM (b), F4/80 (c), with arrows pointing at F4/80+ macrophages, and IL-23 (d). 59 ERAP1 deficient mice have reduced numbers of suppressor “Tr1-like” cells. We have previously demonstrated that ERAP1-/- mice exhibit exaggerated innate and adaptive immune responses to antigenic stimuli, suggesting that they might lack a global immune suppressive function [163]. Two major populations of suppressive T cells include Foxp3+Tregs and Tr1 cells [64]. To investigate if ERAP1 function impacts Treg number or function, we first evaluated the number of conventional Foxp3+CD4+CD25+Tregs in the spleens of ERAP1-/- and WT mice but identified no significant differences between them [Fig. 8a]. We also evaluated the production of IL-10 and TGF-β, as well as analyzed the immune suppressive functional activity of conventional Tregs in cell culture systems, and again observed no differences in these activities between WT and ERAP1-/- derived Tregs [data not shown]. Additionally, we surveyed WT and ERAP1-/- spleens for numbers of Th17 cells, which have been implicated in playing a role in AS [294], but observed no differences [data not shown]. Although less-studied than conventional Tregs, Tr1 cell functions have been implicated in multiple autoimmune diseases, including insulin-dependent diabetes mellitus (IDDM) [92], Multiple Sclerosis (MS) [295], and IBD [62], all of which have associations with ERAP1 polymorphisms[296]. In contrasts to our analysis of Tregs, spleens derived from ERAP1-/- mice contained significantly reduced numbers of CD3+CD4+ Foxp3-IL-10+ “Tr1-like” cells, as compared to splenocyte populations derived from WT mice (p<0.05) [Fig. 8b, c]. These results were also confirmed when using two cell surface markers, LAG3 and CD49b, considered to be important in identifying human and murine Tr1 cells by flow cytometry analysis [64] [Fig. 8d, e]. We wished to assess whether reduced levels of peripheral “Tr1-like” cells in ERAP1-/- mice were due to a reduced ability of naïve T cells to differentiate into Tr1 cells. Naïve CD4+T 60 cells derived from WT and ERAP1-/- mice were isolated and cultured in the presence of anti- CD3, anti-CD28 antibodies, and recombinant IL-27 to induce Tr1 differentiation [297]. After 4 days of anti-CD3, anti-CD28 antibodies and recombinant IL-27 stimulation, there were significantly (p<0.001) reduced numbers of “Tr1-like” cells (CD4+ IL-10+/IFN-γ+ cells) derived from naïve CD4 T cells of ERAP1-/- mice, as compared to identical cultures derived from WT mice [Fig. 8f]. Finally, CF ERAP1-/- mice also had significantly (p<0.05) reduced number of splenic “Tr1-like” cells as compared to WT mice [Fig. 8d], indicating that transfer of microbiota from WT mice also did not correct the reduced Tr1 numbers noted in ERAP1-/- mice. 61 Figure 8. Evaluation of splenic Tr1 cells. Spleens from 14-week-old male and female WT(n=10), ERAP1-/- (n=12) and CF ERAP1-/- (n=8) mice were harvested, processed, stained and analyzed by flow cytometry. (a) The frequency of CD3+CD4+CD25+Foxp3+Treg cells in WT, ERAP1-/-, and CF ERAP1-/- mice graphed. (b) Gating strategy for Tr1 cells is highlighted in red. (c) The frequency of IL-10 producing CD3+CD4+Foxp3-“Tr1-like” cells in WT, ERAP1-/-, and CF ERAP1-/- mice graphed. (d) Graph depicting frequencies of CD3+CD4+Foxp3- Lag-3+ CD49b+ cells in WT, ERAP1-/-, and CF ERAP1-/- mice, with representative images in (e). Statistical significance was determined using One Way ANOVA, and Tukey post hoc test. Naïve CD4 T cells were isolated and cultured for 4 days in the presence or absence of CD3/28 and IL- 27 to induce Tr1 cell differentiation as described in methods. Flow cytometry analysis measuring the relative abundance of IL-10 and IFNγ producing CD4+T cells is graphed on scatter plot (f). Two-way ANOVA with Tukey post hoc test was used for determination of statistical significance. Data are representative of two independent experiments, with similar results. NS - not significant. Data are expressed as the mean ± SEM. 62 ERAP1-/- mice have reduced numbers of tolerogenic dendritic cells. CD45RBhigh CD11clow tDCs are thought to be required for optimal generation of Tr1 cells from naïve CD4 T cells [86], so we wished to assess the number and phenotype of these cells in ERAP1-/- mice. Splenocytes derived from ERAP1-/- had significantly reduced (p<0.05) numbers of CD45RBhigh CD11clow tDCs as compared to splenocytes derived from WT mice [Fig. 6a, b]. The non-classical HLA-G molecule is thought to be important for the differentiation of T cells into Tr1 cells in humans via its interactions between tDCs and naïve T cells [87, 89]. It is well known that ERAP1 functions directly influence classical MHC-I surface levels [147, 150, 151, 298]. In addition, it has been reported that siRNA mediated loss of ERAP1 can prevent HLA-G upregulation in trophoblast cell cultures [299]. We therefore measured surface levels of Qa-2 (the murine homolog of HLA-G) on antigen-presenting cells (APCs) of ERAP1-/- mice. Intriguingly, we observed reduced numbers of Qa-2+ macrophages (p<0.05) [Fig. 9c] and tDCs (p<0.01) in the spleens of ERAP1-/- mice [Fig. 9d, e]. Mean fluorescence intensity (MFI) of Qa-2 was also significantly reduced in the splenic macrophages [data not shown] and tDCs (p<0.01) of ERAP1-/- mice [Fig. 9f]. Percent of Qa-2 positive tDCs and macrophages, as well as the MFI surface levels of Qa-2 on tDCs also remained significantly reduced in CF ERAP1-/- mice [Fig. 9c, d, f]. An aberrant number of tDCs and Qa-2 surface levels suggest a possible mechanism responsible for the reduced numbers of Tr1 cells observed in ERAP1-/- mice, as further elucidated in the Discussion. 63 Figure 9. Evaluation of tolerogenic dendritic cells (tDCs). Spleens from 14-week-old male and female WT (n=10), ERAP1-/- (n=12), and CF ERAP1-/- (n=8) mice were harvested, processed, stained and analyzed by flow cytometry. (a) Relative frequencies of tolerogenic CD45RBhigh CD11clow dendritic cells (tDCs) in WT, ERAP1-/-, and CF ERAP1-/- spleens plotted on a scatter plot bar graph, with representative images in (b). (c) Graph of relative frequencies of Qa-2 expressing F4/80+ macrophages. (d) Graph, depicting percent of Qa-2 expressing CD45RBhigh CD11clow tDCs in WT, ERAP1-/-, and CF ERAP1-/- mice. (e) Representative histogram images demonstrating the relative abundance of Qa-2 positive tolerogenic DCs. (f) Graph depicting Mean Fluorescence Intensity (MFI) of surface Qa-2 on tDCs from WT, ERAP1- /-, and CF ERAP1-/- mice. Statistical significance was determined using One Way ANOVA, and Tukey post hoc test. Data are expressed as the means ± SEM. Data are representative of two independent experiments, with similar results. 64 Discussion Here we show that global deficiency of the ERAP1 gene caused the development of spontaneous axial ankylosis, spinal inflammation, and progressive systemic osteoporosis in mice. Furthermore, similar to AS patients, who have increased susceptibility to IBD [108] and dysbiosis in the terminal ileum [213], ERAP1-/- mice also developed intestinal phenotypes manifested as increased susceptibility to chemically induced colitis and spontaneous gut dysbiosis [213]. Together, these findings suggest that ERAP1-/- mice are an important and useful new animal model for studying the pathogenesis of the most important skeletal [103, 105, 108, 300] and intestinal manifestations found in AS patients [213, 285]. ERAP1-/- mice can also be a useful tool for testing therapeutic agents targeting skeletal, immune, and intestinal manifestations of AS. In addition, ERAP1-/- mice are a unique animal model which can be used as a tool for studying interactions of the gut-bone-immune axis. We and others have identified that ERAP1-/- mice manifest exaggerated innate and adaptive immune responses to a number of stimuli as compared to WT mice, suggesting a loss of global immune suppressive functions [147, 150, 153, 154, 163, 298]. While previous reports indirectly noted that Foxp3+Tregs are perturbed in AS [301, 302], we did not identify differences in the number nor the function of Foxp3+Tregs in the periphery of ERAP1-/- mice. While we cannot rule out entirely that a function of Foxp3+Tregs may be influenced by the loss of ERAP1, our analysis readily confirmed that another major class of regulatory T cells, the “Tr1-like” cells, were significantly reduced in ERAP1-/- mice. Tr1 cells are characterized by their ability to secrete high levels of IL-10 and TGF-β in the absence of Foxp3 expression, and their ability to suppress T cell and APC responses in the periphery [64]. We further identified that the number of tDCs was also significantly reduced in the spleens of ERAP1-/- mice. It is known that tDCs 65 induce naïve CD4 T cells to differentiate into Tr1 cells [86]. Namely, HLA-G expression on tDCs and its interaction with Ig-like transcript 2 (ILT-2) and ILT-4 receptors on naïve CD4 T cells are thought to be important for DC maturation and differentiation of Tr1 cells [87, 89]. While ERAP1’s role in classical MHC-I surface expression is well known, shRNA mediated loss of ERAP1 in human trophoblasts was also shown to reduce non-classical MHC-I HLA-G surface levels [299]. Our analysis showed that Qa-2, the murine homolog of HLA-G, was also significantly reduced on tDCs from ERAP1-/- mice. Together, our results suggest that the possible mechanism responsible for altered Tr1 differentiation in ERAP1-/- mice may be due to reduced tDC numbers and surface levels of Qa-2. We believe that Tr1 cell deficiency in ERAP1- /- mice may play a primary role in the pathogenesis of the phenotypes observed in these mice. Further studies are needed to better understand the mechanism via which ERAP1 mediates Tr1 cell differentiation. In addition to previous findings that ERAP1-/- mice have exaggerated innate and adaptive immune responses to a variety of stimuli [147, 150, 154, 163, 298], H&E staining revealed evidence of severe lumbosacral IVD degeneration in ERAP1-/- mice, with disruption of local structures and the presence of mononuclear cellular infiltrations in the NP. Immunohistochemical analysis of the ERAP1-/- spines revealed that mononuclear infiltrates stained positively for TNF- α and IL-23 in the cytoplasm. Both of these cytokines are known to play an important role in AS [100]. TNF-α is thought to play a major role in disc degeneration via recruitment of inflammatory cells to the IVD, destruction of the extracellular matrix, and calcification and hyperalgesia associated with IVD disease [303]. The IVD infiltrating cells also stained positive for IgM deposits and F4/80+ macrophages, both of which have been previously identified in vertebral biopsies derived from AS patients [300]. Proinflammatory cytokines and macrophage 66 infiltration at the NP are likely responsible for the disc degeneration observed in ERAP1-/- mice. It is possible that inflammation observed at the L6/S1 joint is in response to normal mechanical stress in the axial skeleton, [303] with exaggerated immune responses caused by ERAP1- dependent Tr1 reduction [163]. Commensal bacteria have been shown to play an important role in shaping the immune system, maintaining skeletal and gut homeostasis, and are thought to contribute to inflammatory diseases such as Rheumatoid Arthritis (RA), IBD, and MS [212, 287, 288, 304]. Using 16S rRNA fecal phenotyping we determined that ERAP1-/- mice spontaneously developed a significant fecal overabundance of specific bacterial species, including the Cyanobacteria and Actinobacteria phyla; as wells as Prevotella, Odoribacter. Bacteroides, YS2, Clostridiales and Parabacteroides genera. In addition, ERAP1-/- mice had a significant deficiency of genera Lachnospiraceae, Christensenellaceae, and S24.7. On the level of phylum, fecal samples from ERAP1-/- mice were significantly enriched for Cyanobacteria and Actinobacteria. An overabundance of Prevotella has been previously reported in HLA-B27 transgenic rats, [238] which, similarly to our ERAP1-/- mice develop osteoporosis [239]. Prevotella has also been reported to be enriched in the terminal ileum of AS patients [213]. Similarly to our results, increased Bacteroides has been reported to be increased in the fecal samples of AS patients [305]. While the survey of the terminal ileum of AS patients revealed increases in Lachnospiraceae, Ruminococcaceae, and reduction in Streptococcus and Actinomyces genera [213], and Firmicutes phylum [214], we did not observe such trends in the ERAP1-/- mice. Since the data on human intestinal microbiome in AS patients is limited, future, larger studies are warranted to expand the knowledge of microbiome influences on human AS. ERAP1-/- mice are 67 a suitable model for investigating interactions between gut dysbiosis, skeletal and immune systems. We and others have shown that ERAP1 alters the selection of immunodominant T cell epitopes and antigen-specific T cell functions [147, 150, 151, 298]. ERAP1’s role in immunodominance may also be responsible for the gut dysbiosis observed in ERAP1-/- mice due to altered immune tolerance allowing for aberrant microbial communities to colonize the intestines of ERAP1-/- mice. Cross-fostered ERAP1-/- mice did not resolve spinal inflammation or reduce the severity of ankylosis and osteoporosis, downplaying the role of the intestinal microbiota in these phenotypes. We must note, however, that because Bacteroides, Parabacteroides and Christensenellaceae genera levels were not normalized in cross-fostered ERAP1-/- mice, it is possible that these commensal microbes play a role in the skeletal phenotypes observed in ERAP1-/- mice. While the relationship between microbial communities (such as Bacteroides fragilis [306] and Clostridium species [288]) and Treg development and function has been shown, our work also suggests a possible link between dysbiosis and Tr1 cells. Future studies dissecting the relationship between dysbiosis and Tr1 cells are justified. ........... ERAP1 is a highly polymorphic gene, with many of its SNPs affecting substrate specify, thereby influencing ERAP1 functions [149, 307]. ERAP1 SNPs have been genetically linked to many autoimmune diseases including IDDM, MS, psoriatic arthritis, Bechet’s disease, celiac disease, and others [296]. Our results suggest that ERAP1 may be important for Tr1 cell production and therefore ERAP1-mediated Tr1 reduction could play a role in the pathogenesis of AS and other ERAP1-associated diseases. Indeed, defects in Tr1 cell numbers and functions have been implicated in IDDM [92], MS [295], and IBD [62]. Given these associations, we propose that therapeutic strategies targeting Tr1 development generally, and potentially ERAP1 68 functions specifically, may prove to be useful in the treatment of not only AS patients, but also many more individuals affected by ERAP1-associated diseases worldwide [296]. Further studies investigating the role of ERAP1 in Tr1 cell differentiation and function are now justified. 69 Chapter 3: ERAP 1 deficient mice develop TNF-α independent osteoporosis, enhanced osteoclastogenesis and osteoclast activity. This chapter is an edited version of the article that is being submitted for publication. Authors: Yuliya Pepelyayeva, Maja K. Blake, Sean Hyslop, Sandra Raehtz, Abdulraouf M. Abbas, David P.W. Rastall, Cristiane N. Pereira-Hicks, Sarah G. Roosa, Laura McCabe and Andrea Amalfitano. Author contributions: YP, LM, and AA designed the experiments. Experiments in figure 10 were carried out by YP and SR; figure 11: YP, DPWR, CPH, SGR, and AMA; figure 12: YP and MKB; figure 13: YP and MKB; figure 14: YP; figure 15: YP; figure 16: YP and MB; figure 17: YP. YP analyzed the data. YP and AA drafted the manuscript and all authors assisted in editing the manuscript. 70 Introduction Ankylosing Spondylitis (AS) is a chronic, highly heritable, inflammatory disease which affects an estimated 1 million Americans and many more millions worldwide [102]. AS is characterized by ankylosis (aka fusion) of the bones in the axial skeleton, and it is especially known for inducing inflammation and fusion of the sacroiliac joints [103]. Systemic osteopenia and osteoporosis predispose these affected individuals to pathologic fractures, making them susceptible to spinal cord injuries [105, 108]. Endoplasmic Reticulum Aminopeptidase 1 (ERAP1) polymorphisms have been identified and are highly associated with AS [134, 308]. Recently, we discovered that mice deficient in ERAP1 develop spinal inflammation, axial ankylosis, and osteoporosis, as well as extra-skeletal manifestations such as spontaneous dysbiosis and increased susceptibility to colitis [309], all of which are features of AS [103, 105, 108, 213, 285, 300]. Development of spontaneous skeletal abnormalities in ERAP1-/- mice suggests that ERAP1 plays an important role in the pathogenesis of ankylosis and bone remodeling. Previously, we published that ERAP1-/- mice exhibit hyperinflammatory responses to innate immune stimulation and produce elevated levels of multiple proinflammatory cytokines such as IL-1β, TNF-α and IL-6 [163]. All of these cytokines have been implicated in AS [310], and are known to affect bone remodeling and homeostasis via their effects on osteoblast and osteoclast functions [192]. While anti-TNF-α treatment is effective in reducing the clinical score of AS in affected individuals by reducing pain and inflammation, the efficacy of this therapy in regard to ankylosis and osteoporosis is unknown. Several studies suggest that TNF-α blockade can improve bone mineral density of the lumbar vertebra [223, 224], while others show its failure to reduce osteoporosis and frequency of vertebral fractures [221, 222]. Similarly, there is a discrepancy 71 between studies attempting to determine the efficacy of anti-TNF-α treatment in preventing syndesmophyte formation [170, 225, 226]. Since ERAP1-/- mice develop key skeletal manifestations of AS, namely axial skeletal ankylosis and osteoporosis of vertebral bodies [217], we wished to investigate the role of ERAP1 in the skeletal pathogenesis of AS by critically evaluating the skeletal physiology of ERAP1-/- mice. Additionally, we utilized these mice for evaluation of the efficacy of anti-TNF-α therapy in reducing the severity of the ankylosis and osteoporosis phenotypes. Our analysis showed that deficiency of ERAP1 resulted in reduced osteoblast function in vivo as determined by dynamic bone histomorphometry, which was accompanied by reduced expression of the Runx2 gene. Moreover, cell cultures of bone marrow (BM) cells, showed that ERAP1-/- cells have increased osteoclastogenesis and osteoclast function. TNF-α blockade in vivo did not resolve existent bone fusions and did not improve systemic osteoporosis in ERAP1-/- mice, downplaying the role of this cytokine in the skeletal pathogenesis of these mice. 72 ERAP1 deficiency causes systemic osteoporosis. Results ERAP1 SNPs have been genetically linked to increased susceptibility of developing AS [141]. Previously, we have shown that ERAP1-/- mice develop key skeletal features of AS [309], similar to those, observed in human patients, such as axial ankylosis, syndesmophyte bridging, joint erosions, spinal inflammation and osteoporosis of the spine [108]. Here we performed additional skeletal analysis of ERAP1-/- mice using μCT to further assess their bone health. Analysis of the trabecular bone of the femurs of 14-weeks-old ERAP1-/- mice revealed, reduced bone volume fraction (BV/TV %) (p<0.01), bone mineral density (BMD) (p<0.0001), trabecular number (Tb.N) (p<0.01) and increased trabecular spacing (Tb.Sp) (p<0.01) [Fig. 10B, D, F, G]. Trabecular thickness (Tb.Th) and bone mineral content (BMC) were also reduced in ERAP1-/- femurs, but the differences were not significant [Fig. 10C, E]. The representative isosurface images of the femur trabecular bone [Fig. 10A] and the bone density measures suggest that ERAP1-/- mice develop osteoporosis not only in the axial skeleton, but also in the long bones, suggesting that osteoporosis in these mice is systemic. 73 Figure 10. ERAP1 deficiency results in systemic osteoporosis. Femurs from 14-week-old ERAP1-/- and WT female mice (n=5) were harvested and analyzed with µCT. (A) Representative images of WT and ERAP1-/- trabecular bone of femurs. Graphs representing BV/TV % (B), BMC (C), BMD (D), Tb.Th. (E), Tb.Sp. (F), Tb.N. (G) in WT and ERAP1-/- mice. Graphs represent mean ± SEM, where p values were calculated by student's t-test. NS – not significant. 74 ERAP1 deficient mice exhibit increased osteoclastic potential and osteoclast activity. To understand the underlying bone cell mechanisms responsible for the development of osteoporosis in ERAP1-/- mice, we explored the functions of osteoclasts, the bone resorbing cells of the skeletal system [192]. Culture of bone marrow (BM) cells derived from ERAP1-/- mice resulted in a significantly (p<0.05) increased number of mature osteoclasts, in comparison to WT BM cells [Fig. 11A, B], demonstrating the increased osteoclastogenic potential of the ERAP1-/- BM cells. RANKL is a member of the TNF superfamily, which exists in a membrane-bound and soluble forms [311]. Consistent with its function of promoting osteoclast activation and bone resorption [312-314], we observed a significantly (p<0.05) increased percent of RANKL expressing CD4+ T cells in the BM of ERAP1-/- mice compared to WT [Fig. 11C]. Thus, the enhanced osteoclastogenesis in ERAP1-/- mice is likely due to heightened RANKL expression [315]. Next, we assessed whether the heightened propensity for osteoclast differentiation observed in ERAP1-/- mice was accompanied by increased osteoclast activity. BM cells from WT and ERAP1-/- mice were cultured on fluorescently labeled dentin discs in presence of osteoclast differentiation medium as described in methods. To determine the osteoclast activity, the size of the resorption pits per osteoclast was measured using confocal microscopy. In agreement with the finding of systemic osteoporosis, bone marrow-derived osteoclasts from ERAP1-/- mice developed significantly (p<0.05) larger resorption pits, as compared to osteoclasts derived from WT mouse BM cells [Fig. 11D, E]. Thus, ERAP1 deficiency leads to increased osteoclastogenesis and increased osteoclast activity. 75 Figure 11. Deficiency of ERAP1 mediates increased osteoclastogenesis and osteoclast activity in vitro. Osteoclastogenesis assays were performed on bone marrow from 14-week-old male mice (n=3), as described in methods. Representative images of TRAP+ osteoclasts differentiated from BM cultures is shown in (A). The number of osteoclasts was quantified and graphed in (B). BM cells from 14-week-old mice were collected, stained and analyzed by flow cytometry. Percent of RANKL positive CD4+ T cells is graphed in (C). Osteoclasts (depicted in red) were differentiated from bone marrow on FITC-stained dentin discs (depicted in green), then stained with Rhodamine Phalloidin and analyzed with confocal microscopy. The area invaded on the dentin surface was measured and represents osteoclast activity. Representative images shown in (D), quantified mean pit sizes are graphed in (E). Graphs represent mean ± SEM. Student’s t-tests were performed to determine statistical significance. 76 ERAP1 deficient mice have reduced osteoblast activity in vivo. Bone health is dependent on maintaining an intricate balance of osteoclast and osteoblast activity to ensure homeostasis. Disruption of homeostasis between these two cell types leads to bone diseases [192]. To assess the osteoblast activity in vitro, we cultured BM cells in presence of ascorbic acid and inorganic phosphate for 25 days and measured the level of calcium deposits they produced, using alizarin red release assay, as described in methods. ERAP1-/- derived osteoblasts showed no differences in their production of the calcified extracellular matrix in comparison to WT [Fig. 12A, B]. However, we did observe significantly (p<0.05) reduced mRNA levels of Runx2 gene [Fig. 12C], which is an important transcription factor required for osteoblast differentiation [316]. Interestingly, serum markers of osteoblast (osteocalcin) and osteoclast (TRAP5b) activity were not significantly different in ERAP1-/- mice compared to age- and sex-matched WT littermate mice [Fig. 12D, E]. Analysis of bone formation in the tibia of adult WT and ERAP1-/- indicated that while mineral apposition rate (MAR) was not significantly different [Fig. 12F], the bone formation rate (BFR) was significantly reduced in ERAP1-/- mice compared to WT, indicating a reduced number of active osteoblasts in vivo, [Fig. 12G]. Together these data suggest that ERAP1 deficiency may also affect osteoblast functions. 77 Figure 12. Deficiency of ERAP1 results in reduced osteoblast activity in vivo. Osteoblasts were derived from bone marrow of 14-week-old female mice (n=3) and stained with alizarin red to measure mineralization, as described in methods. Representative images of Alizarin Red staining in (A) and graphed in (B). Tibia from 14-week-old male mice were snap frozen and used for mRNA analysis. Runx2 mRNA levels are graphed in (C). Sera from 14 weeks old female mice (n=5) were analyzed for Osteocalcin and TRAP5b levels and graphed in (D), (E), respectively. 70-week-old female mice (n=8) were injected twice with calcein 11 days and 1 day prior to sacrifice and analyzed for MAR and BFR, graphed in (F) and (G) respectively. All graphs represent mean ± SEM. Student’s t-tests were performed to determine statistical significance. NS – not significant. 78 Ankylosis and osteoporosis developed in ERAP1-/- mice were not improved after anti-TNF- α treatments. TNF-α is elevated in the plasma of AS patients and anti-TNF-α therapies are an option for symptom management in AS patients. While anti-TNF-α treatment is effective in reducing the clinical score of AS by reducing pain and inflammation, the efficacy of this therapy in regard to ankylosis and osteoporosis is under question [170, 221-226]. Interestingly, TNF-α gene is transcribed at high levels in ERAP1-/- mice at baseline [163], paralleling elevated plasma TNF-α levels in AS patients [317]. Additionally, we have shown that the intervertebral discs between S1 and L6 vertebra have significant TNF-α and IL-23 expression in the nucleus pulposus [309]. These cytokines are thought to play a role in the pathogenesis of AS [318, 319], and are known to modulate bone remodeling cells [192]. We wished to test, whether prolonged anti-TNF-α antibody therapy would allow for reduction of axial ankylosis or osteoporosis in ERAP1-/- mice. ERAP1-/- and WT mice were subjected to weekly I.P. injections of an anti-TNF-α antibody or an isotype control for ten weeks, starting at 4 weeks of age. At the end of the study, mice were humanely sacrificed, and their spines were collected and analyzed using μCT. Anti-TNF-α treatments did not reduce the severity of the axial ankylosis between L6 and S1 vertebra in ERAP1-/- mice [Fig. 13A, B], suggesting that anti-TNF-α therapy is not effective for reduction of formed syndesmophytes. We next wished to test whether anti-TNF-α therapy can improve systemic osteoporosis in ERAP1-/- mice. Our analysis of the trabecular bone of the S1 vertebra showed that there were no significant changes in the BV/TV % [Fig. 13C], BMC, BMD, Tb.Th., Tb.Sp., or TB.N. [data not shown] in TNF-α treated ERAP1-/- mice compared to IgG treated controls. Interestingly, despite the fact that we observed no significant changes in the ankylosis or the bone density measures following anti-TNF-α treatment in ERAP1-/- mice, we saw 79 significantly (p<0.05) increased alkaline phosphatase positive osteoblast formation in BM cell cultures [Fig. 13D], suggesting that antibody delivery was successful. Recently, Anti-IL-17A antibodies have been shown to be very effective in reducing clinical symptoms in AS patients, who failed anti-TNF-α therapy [229]. To test the efficacy of anti-IL-17A antibody on the reduction of existing fusions and osteoporosis in ERAP1-/- mice we, performed a small pilot study (n=10), but did not observe significant improvement in the severity of fusions or BV/TV % [Fig. 14A, B], and other bone density measures [data not shown], suggesting that IL-17A is not involved in the pathogenesis of ankylosis and osteoporosis in ERAP1-/- mice. 80 Figure 13. Ankylosis and osteoporosis due to ERAP1 deficiency is TNF-α independent. 4- week-old male and female mice (n=10) were injected weekly with 250 μg/mouse of an anti- TNFα antibody or isotype control for ten weeks. Upon completion of the study, the mice were sacrificed at 14 weeks of age, and their spines were collected and analyzed via μCT. Mice were analyzed for axial ankylosis between S1 and L6 vertebra, with representative images in (A), and mean AS scores graphed in (B). S1 vertebrae from males were analyzed for bone density and architecture and BV/TV % was graphed in (C). Bone marrows from 14-week old male mice (n=4) were cultured and stained for alkaline phosphatase as described in methods. Percent alkaline phosphatase positive cells is graphed in (D). All graphs represent mean ± SEM. Student’s t-tests were performed to determine statistical significance. NS - not significant. 81 Figure 14. Ankylosis and osteoporosis due to ERAP1 deficiency is IL-17A independent. 4- week-old male and female mice (n=10) were injected weekly with 200 μg/mouse of anti-IL17A antibody or isotype control for ten weeks. Upon completion of the study the mice were sacrificed, and their spines were collected and analyzed via μCT. Mice were analyzed for axial ankylosis between S1 and L6 vertebra and S1 bone density measures. Mean AS scores are graphed in (A). BV/TV % of trabecular bone of S1 vertebrae from male mice is graphed in (B). All graphs represent mean ± SEM. Statistical analysis was performed using One-way ANOVA with Tukey’s post hoc test, where NS – not significant. 82 Anti-TNF-α antibody therapy resulted in enhanced expression of pro-osteoblastic genes in ERAP1-/- mice. Despite the fact that we did not observe significant changes in the gross skeletal morphology of ERAP1-/- mice, following anti-TNF-α treatment, we observed an upregulation of several genes important for bone remodeling, confirming that indeed the delivery of the antibody was successful. Specifically, we observed a significant (p<0.01) increase in Dmp1 in anti-TNF-α treated ERAP1-/- mice [Fig. 15A], a gene critical for mineralization of bone and dentin [320]. Similarly, we observed significantly elevated (p<0.01) levels of Trap gene expression (which is important for mineralization due to its role in dephosphorylation of bone matrix proteins [321]), and Osx, which is important for differentiation of mesenchymal cells into osteoblasts and affects bone mineral density [322], in ERAP1-/- mice treated with anti-TNF-α antibody [Fig. 15B, C]. Gene expression levels of Dmp1, Trap, or Osx were not changed in WT mice following anti- TNF-α antibody treatment [Fig. 15A-C]. These data together suggest that TNF-α blockade promotes osteoblast differentiation and mineralization activity in ERAP1-/- mice specifically, despite the failure to improve the osteoporosis in these mice. We also observed an increase in Opg gene in WT and ERAP1-/- mice in response to anti-TNF-α treatment [Fig. 15D], which encodes for the osteoprotegerin decoy receptor of RANKL, and prevents RANK/RANKL interaction, thereby inhibiting osteoclastogenesis [185]. Interestingly, we also observed elevated expression levels of several osteoclastic genes in ERAP1-/- mice following anti-TNF-α treatment. Dc-Stamp gene, which is important for osteoclast fusion and maturation [323], and CatK gene, which is important for actin ring formation and osteoclast activity [324], were significantly increased in ERAP1-/- mice upon TNF-α treatment [Fig. 15E, F]. Calcitonin is an important calcium-lowering hormone produced by thyroidal C cells [325]. Ctr gene, which encodes for the 83 calcitonin receptor was significantly (p<0.05) reduced in ERAP1-/- mice compared to WT and was significantly (p<0.05) upregulated following anti-TNF-α treatment [Fig. 15G]. Calcitonin exerts its anti-resorptive activity by binding to calcitonin receptors, which are especially highly expressed on osteoclasts, thereby inhibiting their activity [326]. Additionally, we evaluated gene expression levels of Rank, Rankl, Wnt10b, Sost, Dkk-1, Oc, Lpr6, Oscar, and Nfatc1, but did not observe significant differences between WT and ERAP1-/- tibia at baseline, or in response to anti-TNF-α treatment [data not shown]. 84 Figure 15. Anti-TNFα treatment induced pro-osteoblastic gene expression in ERAP1-/- mice. Tibia from 14-week-old male mice (WT n=4, ERAP1-/- n=12) were snap frozen and used for mRNA analysis. Relative mRNA levels of Dmp1 are graphed in (A), Trap in (B), Osx in (C), Opg in (D), Dc-Stamp in (E), CatK in (F) and Ctr in (G). All graphs represent mean ± SEM. Student’s t-tests were performed to determine statistical significance. 85 Anti-TNF-α antibody therapy resulted in weight reduction in ERAP1-/- mice and negatively affected bone health. To understand why we did not see significant changes in the bone density of anti-TNF-α treated ERAP1-/- mice, despite clear improvement in osteoblastogenesis and upregulation of pro- osteoblastic genes, we wished to assess bone density of femurs, due to the possibility that the antibody was not reaching the spine in sufficient amounts to elicit changes on the bone remodeling cells. We performed a μCT analysis of the trabecular bone of femurs and observed slight, but not significant reductions in BV/TV%, BMD, and BMC in ERAP1-/- mice following anti-TNF-α treatment [Fig. 16A-C]. We observed no change in the trabecular thickness, but significant (p<0.01) reduction in the trabecular number and significantly (p<0.05) increased trabecular spacing in the femurs of ERAP1-/- mice treated with anti-TNF-α antibody compared to IgG treated controls [Fig. 16D-F]. This result was surprising given known effects of TNF-α on osteoclast activity and osteoblast differentiation [192]. To further assess the effects of anti-TNF- α therapy, we assessed the weights of the ERAP1-/- mice. Indeed, the analysis showed that anti- TNF-α antibody therapy caused ERAP1-/- mice to weigh significantly (p<0.0001) less than the mice treated with isotype controls [Fig. 16G]. While it is unclear why the anti-TNF-α therapy resulted in reduced body weight in ERAP1-/- mice, the significantly lower body weights explain why the trabecular bone density measures were reduced in anti-TNF-α treated mice, as it is well established that bone density values are directly proportional to weight and mechanical loading [327]. 86 Figure 16. Effect of the anti-TNF-α blockade on trabecular bone in ERAP1-/- mice. Femurs from 14-week-old ERAP1-/- male mice (n=8) were harvested and analyzed with µCT. Graphs representing BV/TV % (A), BMD (B), BMC (C), Tb.Th. (D), Tb.N. (E), Tb.Sp. (F) in ERAP1-/- mice treated with IgG isotype control or anti-TNF-α antibody. Mouse weights are graphed in (G). Graphs represent mean ± SEM, where p values were calculated by student's t-test. 87 Expression of human ERAP1 allele improved mean ankylosis score and BMC caused by global ERAP1 deficiency. Finally, we wished to test the effect of transgenic expression of a human ERAP1 allele on skeletal phenotypes of ERAP1-/- mice. Using recombinase-mediated cassette exchange technology, we generated a strain of transgenic mice expressing human ERAP1 gene variant (hERAP1Low+/-), that contains five SNPs which are protective against AS (V349M, R528K, N575D, Q72R5, E730Q) [151] from the ubiquitously expressed ROSA26 locus. These mice were backcrossed with ERAP1-/- mice to discern the roles of human ERAP1 variants in the skeletal system in the absence of murine ERAP1 (mERAP1-/-/hERAP1Low+/-), as described previously [151]. These mice were scanned at 4, 14 and 33 weeks of age and were terminally sacrificed at 10-12 months of age. Their spines were harvested, fixed and scanned on μCT to evaluate mean ankylosis score and bone quality measures. Our analysis showed that mERAP1-/- /hERAP1Low+/- mice had reduced severity of fusion (p=0.05) compared to age-matched ERAP1- /- mice [Fig. 17A]. The severity of fusions did not change over time in these mice [data not shown]. We observed an overall trend of improved bone health in mERAP1-/-/hERAP1Low+/- mice compared to ERAP1-/- [Fig. 17C-G], however, only BMC was significantly (p<0.05) improved [Fig. 17B]. 88 Figure 17. Protective human ERAP1 allele improves bone fusions and BMC caused by ERAP1 deficiency. Sacra from 14-week-old mERAP1-/- and mERAP1-/- /hERAPlow+/- female and male mice (n=9) were harvested and analyzed with μCT. Graphs representing Mean Ankylosis Score are graphed in (A) BMC in (B), BMD in (C), BV/TV % in (D), Tb.Th. in (E), Tb.Sp. in (F), and Tb.N. in (G). Graphs represent mean ± SEM, where p values were calculated by the Student's t-test. 89 Discussion AS patients suffer from bone fusions in the axial spine and paradoxical systemic osteoporosis [103, 105, 106, 178, 217]. The exact mechanism for either one of these phenomena is not completely understood, so it is important to uncover their underlying pathogenesis in order to identify appropriate therapies. We observed a severe systemic osteoporosis in ERAP1-/- mice, which closely parallels the clinical presentation of AS patients. Based on this study, the systemic osteoporosis can be potentially explained by increased osteoclastogenesis and osteoclast activity observed in ERAP1-/- mice. Immune cells can indirectly stimulate osteoclast differentiation and promote their activity via secretion of multiple pro-inflammatory cytokines, including TNF-α, IL-1β, IL-6, and IL-17A [328]. Based on this, osteoporosis in chronic inflammatory diseases is thought to be modulated by the pro-inflammatory cytokines. In our previous work we discovered that in response to innate immune system activators, ERAP1-/- mice produced increased amounts of several proinflammatory cytokines, including IL-1β and TNF-α [163], both of which are known to increase osteoclast activity, and are elevated in AS patients [304]. We have additionally observed inflammatory cytokines in the spines of ERAP1-/- mice, namely IL-23 and TNF-α [309], suggesting that the localized increases in pro-inflammatory cytokines may be the driving mechanism of osteoporosis in these mice. RANKL is present on osteoblasts and is expressed by a multitude of immune cells [179]. Interaction of RANKL with RANK on osteoclast precursors stimulates osteoclast differentiation and activity. This interaction can be prevented by an OPG decoy receptor, which binds to RANKL and therefore inhibits osteoclastogenesis. TNF-α has been shown to stimulate RANKL expression on T cells [313]. Consistent with this CD4+ T cells from ERAP1-/- mice had significantly higher RANKL expression compared to WT, suggesting TNF-α’s involvement in 90 the pathogenesis of osteoporosis in ERAP1-/- mice. Higher RANKL expression in ERAP1-/- mice suggests that their RANKL/OPG ratio favors bone resorption [185], similar to observations made in human patients [329]. Upon TNF-α blockade in vivo, the OPG expression levels significantly increased, suggesting the involvement of this cytokine in osteoporosis of ERAP1-/- mice. However, some of the osteoclastic genes, namely cathepsin K and DC-Stamp encoding genes, which promote osteoclast maturation and activity [321, 322], were upregulated in response to TNF-α blockade in vivo. Similar to our findings, human studies have also shown increased osteoclast activity in vitro following TNF-α inhibition [319]. It is possible that a different cytokine is important for the increased osteoclastogenesis and osteoclast activity observed in ERAP1-/- mice. For example, IL-1β, which is known to be important for osteoclast activity [206], is produced at elevated levels from activated ERAP1-/- cells [163]. Additionally, IL-23, which is upstream of TNF-α is thought to contribute to bone resorption in AS [330]. It is possible that other cytokines need to be blocked, possibly in combination with TNF-α in order to inhibit bone resorption and prevent osteoporosis in these mice. We observed reduced osteoblast activity in vivo which was accompanied by reduced Runx2 gene expression in ERAP1-/- mice. This is consistent with the osteoporotic phenotype observed in ERAP1-/- mice. Previous studies have shown that TNF-α inhibits osteoblast differentiation by inhibiting Runx2 and Osx gene expression [195, 198]. Interestingly, Runx2 gene expression was not improved upon anti-TNF-α treatment and is possibly a contributing factor for lack of bone density improvement in ERAP1-/- mice following anti-TNF-α antibody treatment. While Osx was not significantly reduced in ERAP1-/- mice compared to WT, anti- TNF-α therapy resulted in upregulation of Osx transcript levels. Similarly, Dmp1 transcript levels also significantly increased in ERAP1-/- mice specifically, following TNF-α blockade. 91 Despite upregulation of transcript levels of Dmp1 and Osx genes, important for osteoblastogenesis and mineralization [320, 322], the bone density measures of ERAP1-/- mice treated with anti-TNF-α therapy did not improve, suggesting a potential involvement of other cytokines in the pathogenesis of osteoblast functions and osteoporosis of these mice. In this study ERAP1-/- mice failed to resolve spinal fusions and the systemic osteoporosis following anti-TNF-α treatment. This was surprising due to known effects of TNF-α cytokine on osteoclasts and osteoblasts and based on the proposed role of this cytokine in the pathogenesis of AS [313, 319]. We confirmed that the lack of benefit on anti-TFN-α treatment was not due to failure of the antibody delivery, as we observed changes in multiple genes and enhanced osteoblastogenesis in vitro in response to anti-TNF-α therapy. It is possible that there is no reduction in the severity of fusions because the existing fusions were no longer reversible by the time anti-TNF-α therapy was administered in these mice. In the future, an inducible ERAP1-/- mouse model would be useful to test whether TNF-α blockade can prevent the fusions from forming. It was also surprising that ERAP1-/- mice failed to improve osteoporosis after anti-TNF- α therapy. Interestingly, ERAP1-/- mice subjected to anti-TNF-α therapy had significantly reduced body weight compared to isotype treated controls. Although the cause for reduced body weight in response to anti-TNF-α therapy is unclear, the reduction in weight may have masked the beneficial effects of anti-TNF-α therapy on osteoporosis in ERAP1-/- mice. Reduced osteoblast function in ERAP1-/- mice was a somewhat surprising finding, considering that these mice develop skeletal fusions between the vertebra. However, this phenotype was consistent with the systemic osteoporosis observed in these animals. Nonetheless, it remains possible that osteoblasts derived specifically from the area of spinal fusions have different phenotype from those derived from long bones in these mice. Similarly, it is possible 92 that the expression pattern of bone resorption and osteogenic genes differs in the sacrum of ERAP1-/- mice compared to long bones. We have previously shown that ERAP1-/- mice have spinal inflammation, with TNF-α and IL-23 expression in the nucleus pulposus [309], so it is likely that due to enhanced localized proinflammatory presence in the intervertebral disc, the osteogenic cells behave differently from those located in the long bones. It has been shown that IL-23 [331] and TNF-α [193] have a destructive effect on joints causing erosions and joint damage, and it is generally thought that the improper repair of erosions by osteoblasts causes the formation of fusions. Therefore, it is possible that the enhanced production of TNF-α and IL-23 at the intervertebral joints may be contributing to the local erosions and fusions [309]. In the future, studies testing the efficacy of anti-IL-23 blockade alone and in combination with TNF-α in the prevention of skeletal pathologies in these mice may be useful. Finally, we evaluated the effect of transgenically expressing an AS protective human ERAP1 allele in ERAP1 deficient mice and observed some improvement of the ankylosis and osteoporosis in these animals. The trends of improved bone health are promising, especially when considering the extremely low levels of expression of the hERAP1Low gene in these mice [151]. In the future, it could be useful to generate homozygous ERAP1Low mice to be able to fully evaluate the protective effect of the ERAP1 allele in the skeletal pathogenesis of AS. How the loss of ERAP1 impacts bone remodeling remains an important question, and further studies are needed to elucidate the mechanisms responsible for osteoporosis and ankylosis in ERAP1-/- mice. It may be particularly helpful to evaluate the function of osteoclast and osteoblast cells specifically isolated from the sites of spinal fusions and inflammation. Testing of the effects of additional anti-inflammatory therapies alone and in combinations, as 93 well as further evaluation of mice expressing protective ERAP1 allele on bone remodeling cells, will allow for deeper understanding of ERAP1 modulated pathogenesis in AS. 94 Chapter 4: Development of a novel adenovirus-based cancer immunotherapy that targets the CRACC pathway and suppresses established tumors. This chapter is the edited version of a research article that is being submitted for publication. Authors: Yuliya Pepelyayeva, Patrick O’Connell, Sean Thomas Hyslop, Maja Kuranz Blake, Cristiane Pereira-Hicks, Sarah Godbehere Roosa, Andrea Amalfitano, and Yasser Ali Aldhamen. Author contributions: YP, AA, and YAA designed the experiments. Experiments in figure 18 were performed by YP, PO and MKB; figure 19: YP; Figure 20: YP, YAA, MKB, and CPH; figure 21: YP, MKB, and STH; figure 22: MKB and STH; figure 23: YP and YAA; figure 24: YP, PO, and STH; figure 25: YP. YP and YAA analyzed the data. CPH managed the cell lines and carried out western blot experiments. SGR and CPH prepared adenoviruses. YAA supervised the project. YP and YAA drafted the manuscript and all authors assisted in editing the manuscript. 95 Introduction In autoimmune diseases, the pathology arises from overactive immune responses, due to the failure of suppression mechanisms [2]. Treatment of autoimmunity involves targeted inhibition of inflammatory responses and/or enhancement of the regulatory T cell function [1]. Meanwhile, diseases where immune responses are suppressed, with examples of malignancy [332] and immunodeficiency [333], can benefit from targeted immune activation. Despite the expression of tumor-associated antigens (TAA), which should trigger potent immune responses, the immunosuppressive tumor microenvironment (TME) prevents the development of strong anti-tumor immune responses [334]. Tumor immunosuppressive mechanisms include inhibition of T cells via soluble factors or reduction of co-stimulation signals from antigen presenting cells (APCs) [334-336]. Blockade of T cell inhibitory receptors, such as CTLA-4 [337], PD-1 [338], Tim-3 [339], and others, have been proven to be effective anti-tumor therapies. Furthermore, agonistic antibodies targeting co-stimulatory molecules, such as CD134 (OX40) [340], CD137 (4-1BB) [341], and CD27 [342], are promising tumor immunotherapies in development. Additionally, combining multiple immune checkpoint inhibitors is superior to mono-therapies [259, 260], allowing for strengthened immune responses against tumors and potentially preventing tumor relapse in the future. Therefore, developing potent cancer immunotherapies that augment anti-tumor responses, either as a stand-alone therapy or in combination with other agents, is of significant importance. The CD2-like receptor activating cytotoxic cell (CRACC) receptor (also known as CD317, CS-1, SLAMF7) is a member of the signaling lymphocytic activation molecules (SLAM) family of receptors which is expressed on NK cells, macrophages, dendritic cells, and activated T and B cells [265-268]. CRACC is a homotypic receptor that interacts with the 96 Ewing’s sarcoma-associated transcript 2 (EAT-2) adaptor protein via the phosphorylated cytoplasmic immunoreceptor tyrosine-based switch motifs (ITSMs) of CRACC protein via its Src homology 2 (SH2) domain [267]. In the presence of EAT-2 adaptor protein, engagement of CRACC receptor generally results in immune cell activation, while in its absence, CRACC activation has inhibitory effects as has been shown in EAT-2 negative NK cells [269], and T cells [267]. We have previously demonstrated that specific targeting of CRACC, using murine mCRACC-Fc fusion protein during vaccination, allowed for blockade of the CRACC receptor and significantly enhanced NK cell activation, DC maturation, and antigen-specific CD8+ T cell responses [343]. We hypothesized that mCRACC-Fc fusion protein would make an effective and potent anti-tumor therapeutic by simultaneously activating the innate and adaptive arms of the immune system, thus allowing for enhanced tumor cell killing via simultaneous activation of APCs and enhancement of NK cell and cytotoxic T lymphocyte (CTL) immune responses. To test the ability of mCRACC-Fc to augment innate and adaptive immune responses within the TME, we constructed a mCRACC-Fc-expressing adenovirus and tested its anti-tumor activity in a well-established murine CT26 colon adenocarcinoma tumor model. We showed that transduction of the mCRACC-Fc gene using an adenovirus vector was successful and that expression of mCRACC-Fc resulted in augmented IL-12 production, enhanced activation of NK cells, and increased maturation and activation of APCs. We also noted a dramatic elevation of IFN-β and interferon-stimulated gene (ISG) responses in the spleens of rAd5-mCRACC-Fc treated mice. During the CT26 tumor challenge, intratumoral administration of rAd5-mCRACC- Fc allowed for reduced tumor growth and increased survival of CT26 tumor-bearing mice. Additionally, serial vaccinations against CT26 colon adenocarcinoma using a combination of 97 whole CT26 tumor lysate and rAd5-mCRACC-Fc allowed for the development of enhanced CT26 tumor-specific humoral and T cell responses. Tumors derived from rAd5-mCRACC-Fc vaccinated mice developed enhanced lymphocyte infiltration. Finally, antibodies derived from rAd5-mCRACC-Fc-treated mice showed an enhanced tumor killing via antibody-mediated cellular cytotoxicity (ADCC). Together, our data suggest that overexpression of mCRACC-Fc in TME is a novel cancer immunotherapy strategy that augments both innate and adaptive anti- tumor immune responses. 98 Results rAd5-mCRACC-Fc vector successfully infects cells and expresses mCRACC-Fc transgene. Previously, we constructed a murine mCRACC-Fc fusion protein that was composed of the CRACC extracellular domain fused to the murine IgG1 Fc domain [343]. We successfully validated the ability of mCRACC-Fc to bind to CRACC receptor in a dose-dependent manner and discovered its ability to enhance the activity of NK cells, the maturity of DCs, and antigen- specific T cell immune responses [343]. To further develop this technology for a variety of cancer-targeted immunotherapy applications, we devised an adenovirus platform to accomplish high-level TME delivery of the mCRACC-Fc fusion due to its cost-effectiveness, scalability, and long-term safety profile of using adenovirus vectors in humans [344]. We constructed recombinant Adenovirus 5 (rAd5) vector expressing murine mCRACC- Fc fusion protein and confirmed mCRACC-Fc expression and production by infecting modified HEK293 - C7 adenovirus producer cell line. Twelve hours following rAd5-mCRACC-Fc infection (MOI of 1,000), we noted over 600,000-fold induction of Slamf7 transcript expression over mock- and rAd5-Null-treated controls [Fig. 18A], proving effective transduction and robustness of utilizing Ad5 as a vector for expressing the mCRACC-Fc. 99 Figure 18. Effect of mCRACC-Fc overexpression on NK cell activity and function. Modified HEK-293 - C7 cells (n=3) were mock treated (PBS) or infected with MOI 1,000 of rAd5-Null or rAd5-mCRACC-Fc virus overnight. Trizol was used for mRNA isolation from cells as described in Materials and Methods. Slamf7 gene expression levels were normalized compared to GAPDH gene. (A) mRNA fold induction compared to mock is graphed as a mean ± SEM. 6-week-old Balb/c mice were injected I.V. with 1010 v.p. of rAd5-null or rAd5-mCRACC-Fc. 10 hours post- injection spleens were collected from injected (n=6) and naïve mice (n=3), processed, stained and analyzed by flow cytometry. (B) Representative histograms showing the percent of intracellular IFN-γ+ in CD3-Dx5+ NK cells. (C) Graph representing the percent of intracellular IFN-γ+ in CD3-Dx5+ NK cells. (D) Surface levels of CD69 on CD3-Dx5+ NK cells. 6-week-old Balb/c mice were injected I.V. with 1010 v.p. of Ad-Null or rAd5-mCRACC-Fc. 6 hours post- injection spleens were collected from injected (n=6) and naïve mice (n=3), flash frozen in liquid nitrogen and stored at -80º C for RNA analysis. Gene expression levels were normalized compared to Gapdh gene. mRNA fold induction over naïve mice is graphed for the following 100 Figure 18 (cont’d) genes: Ifnb (E), Ifna (F), Isg15 (G), and Oas2 (H). All graphs represent mean ± SEM. Statistical analysis was performed using one-way ANOVA with Tukey’s post hoc test, where * - p<0.05, # - p<0.01, ^ - p<0.001, & - p<0.0001, and NS - not significant. 101 Overexpression of mCRACC-Fc results in enhanced activation of NK cells and increased expression of type I interferons. Previously we have shown that co-administration of the mCRACC-Fc fusion protein and a rAd5-Null vector (mCRACC-Fc + rAd5-Null) increased IFN-γ secretion by NK cells 12 hours post-injection [343]. To investigate if mCRACC-Fc overexpression by Ad5 vector specifically enhances NK cell activation, we intravenously administered rAd5-mCRACC-Fc or rAd5-Null into Balb/c mice and performed flow cytometry analysis to assess CD69 and IFN-γ expression in NK cells. Ten hours post rAd5-mCRACC-Fc administration, we observed significant increases in the percentage of IFN-γ (p<0.0001) [Fig. 18B, C] and CD69 (p<0.001) [Fig. 18D] expressing NK cells in the spleen of rAd5-mCRACC, as compared to cells derived from rAd5-Null controls. Type I interferons (IFNs) have been shown to be important for NK cell maturation, function, and tumor cell killing [345]. Therefore, we wished to evaluate whether rAd5-mCRACC-Fc enhances NK cell activity by inducing type I IFN responses. Consistent with previous reports [346], rAd5- Null administration into Balb/c mice induced significant upregulation of Ifnb (p<0.0001) expression compared to mock-treated control; meanwhile, rAd5-mCRACC-Fc injection induced significant upregulation of Ifnb (p<0.0001) expression as compared to rAd5-Null-injected mice [Fig. 18E]. Ifna gene expression was also significantly (p<0.01) increased in the spleen of rAd5- mCRACC-Fc-treated mice, compared to mock-treated controls, while rAd5-Null did not induce significant Ifna gene expression at this time point [Fig. 18F]. We also evaluated the induction of ISGs after rAd5-mCRACC-Fc administration. Consistent with increased Ifnb expression, we observed enhanced Isg15 (p<0.0001) and Oas2 (p<0.01) expression in the spleens of rAd5- mCRACC-Fc treated mice, compared to rAd5-Null-injected mice [Fig. 18G, H]. Significant 102 upregulation of these genes upon rAd5-mCRACC-Fc infections suggested that mCRACC-Fc expression induced type I interferon gene activation and downstream ISG activation. Overexpression of mCRACC-Fc enhances dendritic cell maturation and activation. CRACC receptor is abundantly expressed on DCs and CD16+ monocytes [266]. In our previous work, we showed that co-administration of the mCRACC-Fc fusion protein and rAd5- Null virus-induced maturation and activation of DCs 12 hours following injection [343]. To determine if mCRACC-Fc overexpression by rAd5 vector enhances APC maturation, we intravenously injected Balb/c mice with rAd5-Null or rAd5-mCRACC-Fc vectors and evaluated APC phenotypes in splenocytes by flow cytometry. Ten hours-post-injection, we observed significantly (p<0.01) increased expression of CD86 [Fig. 19A, B] and CCR7 [Fig. 19C] co- stimulatory molecules on CD11c+CD11b- DCs of rAd5-mCRACC-Fc injected mice, as compared to rAd5-Null treated controls. We also observed increased CD40 expression on DCs of rAd5-mCRACC-Fc-treated mice compared to the naïve group; however, it was not significantly higher than in rAd5-Null controls [data not shown]. Evaluation of macrophage activation revealed significant upregulation of CD40 (p<0.05) [Fig. 19D] and CD86 (p<0.01) [Fig. 19E, F] on F4/80+CD11b+ macrophages in the spleens of rAd5-mCRACC-Fc as compared to rAd5- Null-treated mice. We did not observe significant differences in the CCR7 expression on F4/80+CD11b+ cells following rAd5-mCRACC-Fc administration [data not shown]. This data suggests that the rAd5-mCRACC-Fc vector is a potent innate immune cell stimulator, capable of inducing APC maturation and activation. 103 Figure 19. Flow cytometry analysis of dendritic cells and macrophage activation in response to mCRACC-Fc overexpression. 6-week-old Balb/c mice were injected I.V. with 1010 v.p. of rAd5-Null or rAd5-mCRACC-Fc. 10 hours post-injection spleens were collected from injected (n=6) and naïve mice (n=3), processed, stained and analyzed by flow cytometry. (A) Percent of CD86+ CD11c+CD11b- dendritic cells. (B) Representative histograms of CD86+ CD11c+CD11b- dendritic cells. Percent of CCR7+ CD11c+CD11b- dendritic cells graphed in (C). (D) Frequency of CD40+ F4/80+CD11b+ macrophages graphed. (E) Representative histogram images of CD86+ F4/80+CD11b+ macrophages, with graphed values in (F). All graphs represent mean ± SEM. Statistical analysis was performed using One-way ANOVA with Tukey’s post hoc test, where * - p<0.05, # - p<0.01, ^ - p<0.001, & - p<0.0001, and NS - not significant. 104 Overexpression of mCRACC-Fc induces Th-1-skewed pro-inflammatory cytokine responses and STAT1 phosphorylation. In addition to DC and macrophage activation and maturation, we observed increases in the gene expression levels of several proinflammatory cytokines and chemokines 6 hours following rAd5-mCRACC-Fc administration. We observed significant upregulation of Il12 (p<0.0001) transcripts in the spleens of rAd5-mCRACC-Fc injected mice, as compared to rAd5- Null controls [Fig. 20A]. Gene expression levels of granulocyte-macrophage colony-stimulating factor (Gm-csf), which is known to be an important growth factor for granulocyte and monocyte differentiation, was also significantly induced in rAd5-mCRACC-Fc injected mice compared to naïve (p<0.001) and trended to be higher than in rAd5-Null injected mice [Fig. 20B]. Additionally, we noted increased protein levels of IL-12p40 (p<0.05) and MIP-1β (CCL4) (p<0.001) [Fig. 20C, D] in the plasma of rAd5-mCRACC-Fc-treated mice, as compared to rAd5- Null. The levels of RANTES (CCL5) and KC (CXCL1) chemokines were also elevated in the plasma of rAd5-mCRACC-Fc injected mice but were not significantly increased compared to rAd5-Null controls [Fig. 20E, F]. To discern the signaling pathway responsible for increased proinflammatory cytokine production and innate cell activation we evaluated mRNA levels of Myd88, Tyk2, Jak1, Traf6, and Nf-kb, and did not observe significant induction 6 hours post rAd5-Null or rAd5-mCRACC- Fc injections [data not shown]. While we observed increased upregulation of Irf7, Irf9, Socs1, and Stat1 upon rAd5-mCRACC-Fc injection, mRNA transcript levels of these genes were not significantly higher than in rAd5-Null controls 6 hours post injection [data not shown]. To investigate if the enhanced type I IFN signaling and ISG upregulation in response to mCRACC- Fc overexpression are associated with increased STAT1 activation, [347] we performed WB 105 analysis for STAT1 phosphorylation. Compared to spleen lysates of rAd5-Null-treated mice, lysates of rAd5-mCRACC-Fc-treated mice showed an increased STAT1 phosphorylation [Fig. 20G, H]. 106 Figure 20. Cytokine and chemokine responses in rAd5-mCRACC-Fc treated mice. 6-week- old Balb/c mice were injected 1010 v.p. / mouse I.V. (n=6). 6 hours post-injection spleens were collected, flash frozen in liquid nitrogen and stored at -80ºC. RNA from spleens was extracted using Trizol reagent. mRNA expression was normalized against Gapdh and expressed as fold change over naïve (n=3) samples. Graphs depicting mRNA fold induction of IL12 (A) and GM- CSF (B). 6-week-old Balb/c mice (n=6) were injected I.V. with 1010 v.p. of rAd5-null or rAd5- mCRACC-Fc. 10 hours post injection, plasma was collected and was analyzed using 27-plex assay on Luminex 100. Graphed concentrations of IL-12p40 (C), MIP1β (D), RANTES (E) and KC (F). Flash Frozen spleens were homogenized for protein analysis 6 hours post injections as described in Materials and Methods. Images of Western Blots representing phosphorylated STAT-1 and β-actin are depicted in (G) and graphed in (H). All graphs represent mean ± SEM. Statistical analysis was performed using One-way ANOVA with Tukey’s post hoc test, where * - p<0.05, # - p<0.01, ^ - p<0.001, & - p<0.0001, and NS - not significant. 107 rAd5-mCRACC-Fc vaccination enhances adaptive immune responses to the co- administered antigens. In addition to activating innate immunity, type I IFNs are known to stimulate the adaptive arm of the immune system, both directly by impacting effector T-, regulatory T-, and B-cell responses, as well as indirectly by enhancing antigen cross-presentation to T and B cells [348, 349]. Since mCRACC-Fc overexpression induced dramatic upregulation of IFN-β and enhanced activation of DCs and macrophages, we wished to investigate whether rAd5-mCRACC-Fc enhanced adaptive immune responses. Balb/c mice were intravenously injected with rAd5-Null, rAd5-mCRACC-Fc, or not injected (naïve). Ten hours following rAd5 administration, splenocytes were isolated, stained, and analyzed by flow cytometry to assess the activation phenotype of the adaptive immune cells. We observed significantly (p<0.001) increased percentage of the CD69-expressing CD19+CD3- B [Fig. 21A, B], CD3+CD8- T [Fig. 21C], and CD3+CD8+ T cells [Fig. 21D], as compared to splenocytes derived from rAd5-Null injected mice. We then evaluated the ability of rAd5-mCRACC-Fc to induce memory B- and T-cell responses by utilizing CT26 tumor cell lysates. Balb/c mice were vaccinated twice over a period of one month with either CT26 tumor lysate alone, or combination of tumor lysates + rAd5-Null or tumor lysates + rAd5-mCRACC-Fc vectors. Splenocytes were harvested and cultured for 48 hours in the presence of CT26 tumor lysates to stimulate memory responses against tumor- associated antigen (TAAs). In this assay, we observed significantly (p<0.01) increased activation of CD3+CD8+ T cells from rAd5-mCRACC-Fc/CT26 lysate co-vaccinated mice, as compared to rAd5-Null/CT26 lysate, and lysate only groups [Fig. 21E]. We also observed similar trends in CD3+CD8-T cells [data not shown]. Additionally, we evaluated the number of, CT26-derived, 108 TAA-specific IFN-γ-producing memory T cells via ELISPOT assay. We observed significantly increased (p<0.05) number of CT26-specific IFN-γ+ T cells in rAd5-mCRACC-Fc/CT26 lysate co-vaccinated mice, as compared to naive and lysate only vaccinated groups [Fig. 21F]. Similarly, we observed significantly increased number of Ad5-specific IFN-γ+ T cells in rAd5- mCRACC-Fc/CT26 lysate-vaccinated mice, as compared to naive (p<0.01) and lysate only (p<0.05) groups [Fig. 21G]. Our data suggest that overexpression of mCRACC-Fc enhanced the early activation of the adaptive immune cells and induced memory T cell responses to the co- administered antigens. 109 Figure 21. Adaptive immune cell activation in rAd5-mCRACC-Fc treated mice. 6-week-old Balb/c mice were injected I.V. with 1010 v.p. of rAd5-Null or rAd5-mCRACC-Fc viruses. 10 hours post-injection spleens were collected from injected (n=6) and naïve mice (n=3). Percent CD69+ CD19+CD3- B cells analyzed using flow cytometry and graphed in (A), with representative histogram images in (B). Percent of CD69+ CD3+CD8- T cells is graphed in (C), and percent of CD69+ CD3+CD8+ T cells in (D). 6-week-old Balb/c mice were injected twice I.M. (n=6) over a period of 1 month (Days 0 and 12) with 200 μg of CT26 tumor lysate, tumor lysate + 1010 v.p. of rAd5-Null, tumor lysate + 1010 v.p. of rAd5-mCRACC-Fc, or not injected (non-vaccinated, n=4). On Day 27, mice were sacrificed, and their spleens were collected. Splenocytes were cultured in-vitro in the presence of 10 μg/mL of CT26 tumor lysate for 48 hours, after which cells were stained and analyzed by flow cytometry. Percent of CD69+ CD3+CD8+ T-cells is graphed in (E). Splenocytes from vaccinated mice were stimulated with 10 μg/mL of CT26 tumor lysate (F) or with 1010 v.p. of heat-inactivated rAd5-Null (G) overnight and subjected to ELISPOT analysis for IFN-γ+ cells. Statistical analysis was performed using 110 Figure 21 (cont’d) One-way ANOVA with Tukey’s post hoc test. Graphs represent mean ± SEM, where * - p<0.05, # - p<0.01, ^ - p<0.001 and & - p<0.0001, and NS - not significant. 111 Intratumoral administration of rAd5-mCRACC-Fc vector reduces tumor growth and increases the survival of CT26 tumor-bearing mice. Augmented NK cell responses and type I interferon production are known to elicit antitumor effects due to their ability to stimulate innate and adaptive immune responses [348, 349]. Based on the fact that mCRACC-Fc overexpression induces NK cell activity and type I IFN responses, as well as augments antigen-specific memory immune responses, we wished to test the efficacy of rAd5-mCRACC-Fc vector as an immune modulating vector against CT26 colon adenocarcinoma cancer cells in vivo. First, we confirmed successful transfer and expression of Slamf7 gene in CT26 infected cells in vitro [Fig. 22A] and in vivo [Fig. 22B] upon rAd5-mCACC-Fc infection. We then challenged Balb/c mice with CT26 colon adenocarcinoma tumors and when the tumors became palpable (about 150-200 mm3), mice were either intratumorally injected with 1010 viral particles (v.p.) of rAd5-Null, rAd5-mCRACC-Fc, or were not injected (untreated). Log-rank test showed a significantly (p<0.05) increased survival of rAd5-mCRACC-Fc injected mice (85% survival), as compared to rAd5-Null (46%) and untreated (43%) mice [Fig. 23A]. rAd5-Null treatment resulted in significantly (p<0.05) reduced tumor volume on days 10 through 19 compared to naive mice [Fig. 23A]. Additionally, rAd5- mCRACC-Fc vaccination resulted in significantly (p<0.01) reduced tumor volumes on days 10 through 21 and continuing through day 23 (p<0.05), as compared to untreated mice [Fig. 23B]. Although we noted reduced tumor volumes in rAd5-mCRACC-Fc treated mice compared rAd5- Null, those differences were not significant [Fig. 23B]. Three rAd5-mCRACC-Fc injected mice completely resolved their tumors. At the completion of the study, these mice were re-challenged with CT26 tumor cells via subcutaneous injection of 600,000 cells into the flank and were followed for 3 months. Importantly, none of the re-challenged mice developed tumors [data not 112 shown], suggesting that mCRACC-Fc overexpression facilitated the development of long-lasting and potent memory T cell responses against CT26 TAAs. 113 Figure 22. Expression of mCRACC-Fc in CT26 cells. CT26 cells (n=3) were mock treated (PBS) or infected with MOI 1,000 of rAd5-Null or rAd5-mCRACC-Fc virus overnight. Trizol was used for mRNA isolation from cells as described in methods. Slamf7 gene expression levels were normalized compared to Gapdh gene and graphed in (A). Statistical analysis was performed using one-way ANOVA with Tukey’s post hoc test. 8-week-old male Balb/c mice were injected subcutaneously (S.Q.) in the flank with 150,000 CT26 cells in 100 µL of PBS. 14 days later mice were either injected intratumorally (I.T.) with 1010 v.p. of rAd5-mCRACC-Fc (n=2), or not injected (n=2). Tumors were collected after 14 hours and snap frozen. RNA was extracted using TRIZOL. Slamf7 gene expression levels were normalized compared to Gapdh gene and graphed in (B). Statistical analysis was performed using Student’s unpaired t-test. The graphs represent mean ± SEM. 114 Figure 23. Effect of rAd5-mCRACC-Fc intratumoral injections on anti-tumor responses. 6- week-old Balb/c mice were injected S.Q. with 150,000 CT26 cells. 8 days later, mice were split into 3 groups: injected I.T. with 1010 v.p. of rAd5-Null (n=14), rAd5-mCRACC-Fc (n=14), or not injected (untreated, n=15). Mice were monitored every 2-3 days starting 5 days post tumor challenge. Tumor volume of 2,000 mm3 or the presence of ulcerations, were used as the humane endpoint. (A) Kaplan-Meier survival curve. Statistical analysis was performed using Log-ranked (Mantel-Cox) test. (B) Graph representing tumor volumes over time. Statistical analysis was performed using One-way ANOVA with Tukey’s post hoc test. Graphs represent mean ± SEM, where * - p<0.05 and # - p<0.01. 115 Adenovirus and tumor lysate pre-vaccinations allow for reduced tumor growth and increased survival upon CT26 tumor challenge. Strategies using agents that enhance innate immune responses [350] or block inhibitory immune pathways have proven to be useful as vaccine adjuvants [351]. Since we showed robust increases in the innate and adaptive immune activation following rAd5-mCRACC-Fc administration, we next wished to test the ability of rAd5-mCRACC-Fc/CT26 lysate co- vaccination to prevent CT26 tumor growth in-vivo. We therefore intraperitoneally injected Balb/c mice with a combination of 200 µg of CT26 tumor lysate and 1010 v.p of rAd5-Null, or rAd5-mCRACC-Fc three times over a 5-week period, after which, we subcutaneously injected the mice with 250,000 CT26 tumor cells. Percent survival at the completion of the study was significantly increased (p<0.05) in rAd5-Null/CT26 lysate (67%) and rAd5-mCRACC-Fc/CT26 lysate (63%) pre-vaccinated mice compared to non-vaccinated mice (11%) [Fig. 24A]. On days 22, 26, and 28 the tumor volumes were significantly reduced (p<0.01) in rAd5-mCRACC- Fc/CT26 lysate pre-vaccinated mice compared to non-vaccinated [Fig. 24B]. rAd5-Null/CT26 lysate pre-vaccinated mice also developed significantly reduced tumor sizes compared to non- vaccinated mice on days 22 (p<0.01), 26 (p<0.05), and 28 (p<0.01) [Fig. 24B]. There were no significant differences in the tumor volumes and survival between rAd5-Null and rAd5- mCRACC-Fc pre-vaccinated animals, indicating that three injections of CT26 lysate in combination with a rAd5 vector may have masked the benefit of mCRACC-Fc overexpression. To evaluate the tumor antigen-specific B cell memory responses developed by the vaccinated mice, we quantified anti-CT26 IgG antibody levels in the sera of vaccinated mice using CT26-lysate-specific IgG ELISA. Multiple dilutions of plasma were performed to dilute down the signal and be able to discern differences between the vaccinated and non-vaccinated 116 mice. While both viruses significantly increased tumor-specific antibody production, at dilutions 1:100 and 1:500 it was evident that rAd5-mCRACC-Fc/CT26 lysate co-vaccinated mice produced significantly (p<0.01 and p<0.05, respectively) increased CT26-specific antibodies, as compared to plasma derived from rAd5-Null-treated mice [Fig. 24C]. Based on the fact that the Fc region of IgG antibodies can bind to the CD16 receptor on NK cells and activate ADCC [352], we evaluated killing of CT26 tumor cells in the presence of plasma from co-vaccinated and non-vaccinated mice. Splenocytes or isolated NK cells from naïve animals were incubated with CFSE-labeled CT26 tumor cells in the presence of plasma from non-vaccinated, rAd5-Null/CT26 lysate or rAd5-mCRACC-Fc/CT26 lysate vaccinated mice for 18 hours, after which killing of tumors cells was assessed by flow cytometry as described in Materials and Methods. Interestingly, plasma from rAd5-mCRACC-Fc/CT26 lysate co-vaccinated mice significantly enhanced CT26 tumor killing by both total splenocytes (p<0.01) [Fig. 24D] and isolated NK cells (p<0.05) [Fig. 24E], as compared to plasma-derived from non- vaccinated mice. rAd5-Null/CT26 lysate co-vaccination did not induce significantly higher killing by splenocytes or NK cells compared to non-vaccinated mice [Fig. 24D, E]. The above data suggest that combining the use of the rAd5-mCRACC-Fc and CT26 lysate is an approach that improves tumor killing via enhanced ADCC activity. 117 Figure 24. Efficacy of tumor lysate in combination with rAd5-mCRACC-Fc as an anti- tumor vaccine. 6-week-old Balb/c mice were injected with 3 doses of 200 µg CT26 tumor lysate and 1010 v.p. of adenovirus intraperitoneally (I.P.) over a period of 5 weeks as described in Materials and Methods. Week 5 post 1st injection mice were given the 3rd dose of vaccination and were injected S.Q. with 300,000 CT26 tumor cells into the flank. Mice were monitored every 2-3 days starting 5 days post tumor challenge. Tumor volume of 2,000 mm3 or the presence of ulcerations were used as the humane endpoint. (A) Graphed percent survival. Statistical analysis was performed using Log-ranked test. (B) Graph representing tumor volumes over time. Statistical analysis was performed using One-way ANOVA with Tukey’s post hoc test. (C) Relative abundance of tumor-specific IgG antibodies in the plasma of vaccinated mice compared to naïve on day 41 post tumor cell injection. Each dilution was analyzed using One- way ANOVA with Tukey’s post hoc test to determine statistical significance. Splenocytes from naïve mice were incubated with CFSE-CT26 cells (20:1 E:T) and cultured for 18 hours in 118 Figure 24 (cont’d) presence of plasma 1:200 dilution from rAd5-Null/CT26 lysate, rAd5-mCRACC-Fc/CT26 lysate vaccinated, or non-vaccinated mice. Cells were stained with CellTrace Violet dye and analyzed by flow cytometry. Percent killing is graphed in (D). Percent killing of CFSE-CT26 cells by Dx5+ NK cells (1:1 E:T) overnight in presence of plasma (1:50 dilution) from vaccinated mice is graphed in (E). Graphs represent mean ± SEM, where * - p<0.05, # - p<0.01, ^ - p<0.001, & - p<0.0001, and NS – not significant. 119 rAd5-mCRACC-Fc injected tumors have increased CD8+ T and NK cell infiltration. Solid tumors infiltrated by TILs are associated with increased survival [353] and improved response to checkpoint inhibitors [354]. To test whether mCRACC-Fc overexpression promotes recruitment of TILs, upon completion of the tumor lysate and rAd5 combined pre- vaccination study, the tumors from each of the three study groups were harvested and subjected to immunohistochemistry analysis. Anti-CD3 antibody staining was remarkably higher in the tumors of rAd5-mCRACC-Fc vaccinated mice and was especially pronounced in the peripheral aspects of the tumors [Fig. 25A]. Tumors from non-vaccinated and rAd5-Null-vaccinated mice had visibly less pronounced and sparsely distributed CD3+ staining throughout the tumors [Fig. 25A]. In addition, we performed anti-CD8 and anti-Dx5 staining to evaluate the presence of CTLs and NK cells in the tumors. While we observed evidence of CD8+ staining in the tumors of rAd5-Null and rAd5-mCRACC-Fc pre-vaccinated animals, we observed none in the non- vaccinated group [Fig. 25B]. Interestingly, only tumors from rAd5-mCRACC-Fc injected mice showed signs of Dx5+ NK cell infiltration [Fig. 25C]. Our data suggest that combination of tumor lysate and rAd5-mCRACC-Fc enhanced lymphocyte migration to tumor sites and resulted in increased T cell and NK cell infiltration of the tumors. 120 Figure 25. CD3, CD8 and DX5 immunohistochemical analysis of tumors. Tumors from non- vaccinated (n=1), rAd5-Null/CT26 lysate (n=3) and rAd5-mCRACC-Fc/CT26 lysate (n=3) vaccinated mice were collected and fixed at the end of the study (day 41 post tumor challenge). Immunohistochemistry for CD3, CD8, and Dx5 was performed. A) Representative images of CD3 stained tumor slides, with images taken using 10X magnification. (B) Representative images of CD8 stained tumor sections, with arrows pointing at CD8+ cells. Images were taken using 20X magnification. (C) Representative images of DX5 stained tumor slides, with arrows pointing at DX5+ cells. Images were taken using 40X magnification. 121 Discussion The self-ligand CRACC receptor mediates both positive and negative signals in immune cells, a function that is governed by the presence or absence of the EAT-2 adaptor [268, 355]. In this study, we targeted CRACC receptor by overexpressing mCRACC-Fc fusion protein using an adenoviral vector as a therapy in a solid-tumor model. We have demonstrated previously that co- administration of a recombinant mCRACC-Fc fusion protein enhances vaccine-induced innate and antigen-specific T cell immune responses [343], suggesting that mCRACC-Fc could function as a novel immune-modulating agent. Building upon those proof-of-principle studies, we have now developed a gene therapy approach that targets the CRACC pathway using recombinant Ad5 as a vector. Overexpression of mCRACC-Fc by an adenoviral vector allowed for an enhanced production of critical, Th-1-skewed cytokines, as well as early activation of innate and adaptive immune cells. Additionally, serial injections of tumor antigen in combination with rAd5-mCRACC-Fc enhanced TAAs-specific memory T- and B-cell responses, as well as infiltrations of tumors by leukocytes. Most importantly, intratumoral administration of the rAd5- mCRACC-Fc vector reduced tumor growth and improved survival rate in mice with established CT26 colon adenocarcinoma tumors. While the molecular mechanism as to how rAd5-mCRACC-Fc achieves these beneficial results is not fully defined, in this study we observed dramatic upregulation of type I IFNs, especially IFN-β, in response to mCRACC-Fc overexpression. Type I IFNs are known to play an important role in mediating anti-tumor immune responses [347]. In addition to preventing malignant cellular transformation, via maintaining p53 tumor suppressor gene expression [356] and triggering apoptosis of certain cancers [357], type I IFN-mediated anti-tumor activity is thought to be mainly indirect, via the activation of other immune cells [358]. For example, type I 122 IFNs increase APC survival and antigen cross-presentation, thereby promoting CD8+ T cell survival, and inhibiting regulatory T cell responses [345, 347]. Additionally, type I IFNs are known to play a particularly important role in driving NK cell anti-tumor activity by directly promoting NK cells maturation and activation, and indirectly via IL-15 secretion by type I IFN- activated conventional DCs [345]. It is also known that delivery of a strong type I IFN signal specifically into tumors and combination of type I IFN targeted strategies with immune checkpoint inhibitors allows for the development of robust anti-cancer responses [358, 359]. Therefore, the enhanced NK cell, APC, and T cell activation following rAd5-mCRACC-Fc administration may be the result of upregulation of type I interferons and their downstream effects. Interestingly, this suggests that targeting the CRACC signaling pathway, may prove to be a useful strategy for promoting interferon-induced responses generally. Expression of the mCRACC-Fc via an adenovirus vector also resulted in enhanced NK cell activation and increased IFN-γ production early after rAd5-mCRACC-Fc administration, an NK cell phenotype that is important for the development of Th-1-skewing innate and adaptive immune responses [360, 361]. This finding is very important because NK cells are vital for tumor cell killing and prevention of metastasis [362]. Additionally, Th-1 skewed immune responses promote anti-tumor immunity [26]. While the exact mechanism responsible for the enhanced NK activity by rAd5-mCRACC-Fc is not completely defined, we noted that administration of rAd5-mCRACC-Fc induced higher levels of IL-12 production, early after administration, which is important for NK cell activity and induction of Th-1 responses [363]. Therefore, it is possible that the enhanced NK cells activation and IFN-γ production is mediated by rAd5-mCRACC-Fc-regulated IL-12 release [363]. Moreover, mice serially vaccinated with CT26 tumor lysate and rAd5-mCRACC-Fc developed increased levels of tumor-specific IgG 123 antibodies, and as a result, showed an increased killing of tumor cells by inducing ADCC response ex vivo. ADCC is an important mechanism in tumor-killing [352], which is exploited in anti-tumor immunotherapy approaches [364, 365]. The ability of rAd5-mCRACC-Fc to enhance tumor-specific antibody production with an enhanced ADCC-inducing capability is another feature that makes CRACC receptor a favorable target for immune-modulation. In addition to NK cell activation, we also noted that CRACC modulation enhanced the maturation and activation of CD11c+ conventional DCs, a phenotype that might also be mediated by the increased IFN-γ production from NK cells [366]. Our data suggest that CRACC could function as an inhibitory receptor in DCs and monocytes, and that blockade of CRACC- CRACC self-ligation by the use of CRACC-Fc-producing rAd5 vectors enhances DC- and macrophage-mediated innate immune responses. Consistent with this, activation of human monocytes with LPS has been shown to reduce EAT-2 levels, which resulted in reduced TNF-α and IL12p70 production upon CRACC receptor activation [367]. These findings suggest that in LPS-treated monocytes CRACC receptor has an inhibitory function due to loss of EAT-2. In agreement with this, we have shown previously that treatment of murine RAW264.7 macrophages with mCRACC-Fc fusion protein increased EAT-2 levels and prevented LPS- mediated EAT-2 downregulation [343]. Traditional anti-cancer treatments are generally highly immunosuppressive [368]. Preclinical data suggest that enhancement of innate immune responses can ensure the development of long-lasting adaptive anti-cancer immune responses [364]. It is also thought that strategies that enhance both the innate and adaptive immune responses might allow for the development of more effective and long-lasting anti-tumor immunity. In this study, we observed that overexpression of mCRACC-Fc resulted in increased activation of DCs and macrophages as 124 evidenced by upregulation of CD86 activation receptor. CD86 is a co-stimulatory molecule that binds to the B7 receptor, which upon stimulation, induces T cell activation [369]. Increased surface levels of maturation markers on DCs and macrophages suggests that mCRACC-Fc overexpression enhances the activation of the adaptive arm of the immune system via antigen cross-presentation [369]. Consistent with this, serial vaccinations with tumor lysate and rAd5- mCRACC-Fc enhanced tumor-specific adaptive immune responses, as was evidenced by the enhanced production of tumor-specific antibodies and IFN-γ-expressing T cells. Recently, a small phase Ib clinical trial in patients with metastatic melanoma reported an improved response to combined intralesional injection of modified human herpes simplex virus and systemic anti-PD-1 [354]. This was accompanied by conversion of cold (non-inflamed) tumors to hot (inflamed) tumors, as was evidenced by enhanced T cells infiltrations into the tumors in response to therapy. Additionally, it is evident that increased infiltration of tumors by CD3+ and CD8+ T-cells is associated with better prognosis and increased survival in patients affected by various solid tumors [353]. Increased infiltration of CD3+, CD8+, and Dx5+ TILs seen in the CT26 challenged mice treated with rAd5-mCRACC-Fc intratumorally, suggests that increased lymphocyte recruitment to the tumors may be the mechanism responsible for the increased survival rates and reduced tumor growth observed in these mice. Interestingly, we observed an increased production of MIP-1β chemokine in response to rAd5-mCRACC-Fc administration, a chemokine that has been shown to reduce tumor size and enhance survival in CT26 challenged mice via the recruitment of T- and NK-cells to tumors [370]. In an effort to test the efficacy of combined tumor antigens and rAd5-mCRACC-Fc as a preventative vaccine against CT26 colon adenocarcinoma, we subjected mice to repeated doses of combination vaccines containing tumor lysate prior to CT26 challenge as a proof-of-principle. 125 Although the differences between tumor volumes were not significant, tumor volumes from rAd5-mCRACC-Fc treated mice showed a trend of smaller volumes compared to rAd5-Null and exhibited less variability in tumor sizes. The efficacy of rAd5-Null in reducing tumor volumes is not surprising as it is a double stranded DNA virus that can activate innate immune responses via TLR2 and TLR9 pathogen recognition receptors and induce downstream signaling despite it being a non-replicating virus [280]. Ad5 can also enhance adaptive immune responses [371]. It is likely that the robust immune activation induced by the three-dose vaccination of rAd5 vectors potentially masked some of the beneficial effects of mCRACC-Fc overexpression. Additionally, tumor lysates alone have low immunogenicity [372], therefore the use of more immunogenic epitopes along with rAd5-mCRACC-Fc vaccine would likely result in more impressive responses. Despite the use of poorly immunogenic tumor lysate, rAd5-mCRACC-Fc and tumor lysates co-vaccination triggered enhanced CT26 tumor-specific antibody production and ADCC responses. In summary, our work showed that blockade of CRACC signaling by using rAd5- mCRACC-Fc vector enhanced IL-12 production, increased type I IFN responses, improved APC activation and maturation, augmented NK cell responses, and enhanced tumor-specific memory immune responses. Most importantly, rAd5-mCRACC-Fc intratumoral injections resulted in reduced tumor growth and enhanced survival rate of tumor-bearing mice, a phenotype that was accompanied with enhanced recruitment and infiltration of T and NK cells into tumors. Together, our data suggest that CRACC receptor immunomodulation strategy is a novel approach for enhancing anti-tumor immune responses, which should be explored further as a stand-alone immunotherapy and/or in combination with other immune-modulating agents. 126 Chapter 5: Materials and Methods 127 Animal procedures All experiments were performed on C57BL/6 (WT) or C57BL/6 global ERAP1 knockout mice (ERAP1-/-), which were a kind gift from Dr. Kenneth Rock (Professor, University of Massachusetts Medical School). All mice used in experiments were bred in-house. All animal procedures were reviewed and approved by the Michigan State University EHS, IBC, and IACUC (http://iacuc.msu.edu.proxy1.cl.msu.edu/) and conformed to NIH guidelines (AUF number: 02/13-045-00 for work in chapters 2 and 3; AUF: 09-14-166-00 for work in chapter 4). Care for mice was provided in accordance with PHS and AAALAC standards. Numbers, age and sex of animals used for each experiment are specified on the corresponding figure legends. mERAP1-/-hRAP1low+/- were generated as previously described [151]. µCT imaging and ankylosis scoring Vertebrae were harvested, fixed, and scanned using a GE Explore Locus microcomputed tomography (μCT) system as described previously [373]. Scans were performed using 20 μm voxel resolution obtained from 720 views. Beam angle increment of 0.5, beam strength of 80 peak kV and 450 µA were used. Each run included age-matched WT, and ERAP-/- spines, and a calibration phantom for standardization and consistency. Isosurface images were created using an averaged autothreshold value of 1009, to separate bone from soft tissue using MicroView software. Right and left sides of the lumbar vertebrae 6 (L6) were scored separately and the scores were averaged to calculate the Mean Ankylosis Score. A “0” represented WT morphology of L6; a 1 represented the beginning of a syndesmophyte growing from L6 0.5-1 mm in length; a 2 represented a syndesmophyte extending 1-2 mm; a 3 represented a bridging syndesmophyte 128 extending over 2 mm and contacting the iliac bone, but not fused with S1; and a 4 represented a fully fused joint. Representative scores are shown in Figure 1A. Trabecular bone analysis of the S1 vertebra was performed using advanced ROI generated using splines in the transverse plane outlining the entire length of the vertebral body excluding the outer cortical bone. Average threshold values of 1232 for sacra in Chapter 2 were used in order to separate bone from bone marrow. In chapter 3, bone measurements of femurs were done using fixed threshold value of 847.4 for trabecular bone. Trabecular bone analysis of femurs was performed in the region starting from the growth plate and extending 1% of the total femur length proximally toward diaphysis excluding the cortical bone. Bone measurements of S1 in chapter 3 were done using a fixed threshold value of 877 for trabecular bone. Bone mineral density, content, bone volume fraction, trabecular thickness, spacing and number values, as well as the representative isosurface images were generated using GE Healthcare MicroView software. Bone measurements were blinded until the final analysis. AS index AS index was calculated by taking the sum of the Mean Ankylosis Score and an assigned BMD score, where 325 – 300 mg/cc received score 1, 299-275 – 2, 274-250 – 3 and <250 – 4. Erosion scoring SI joints were assessed and scored for erosions using MicroView software in the sagittal plane, where “0” represented a smooth joint; “1” represented a joint with mild erosions; and “2” 129 represented erosions >1mm in size or >3 erosions in one joint. Both SI joints were scored separately and averaged. DSS-induced colitis experiments 8-week-old male mice were treated with 3% DSS (MP Biologicals) dissolved in sterile distilled water. Mice were given DSS in sterile drinking water ad libitum for the experimental days 1-7 (controls received sterile water only) followed by regular water until the end of the study. The DSS solutions were made fresh on day 0 and 3. Body weight, stool consistency, and the presence of blood in the stool were determined daily, as previously described [374]. The baseline clinical score was determined on day 0. Scoring procedures were performed, as follows. Weight loss relative to baseline: 0, no weight loss; 1, 1%- 5% weight loss; 2, 5%- 10% weight loss; 3, 10%- 20% weight loss; and 4, >20% weight loss. Stool scores were determined as follows: 0, well-formed pellets; 1, semiformed stools that did not adhere to the anus; 2, pasty and semiformed stools that adhered to the anus; 3, liquid stools that adhered to the anus. Bleeding scores were determined as follows: 0, no blood; 1, minor blood traces in stool visible; 2, visible blood traces in stool; 3, visible blood traces and dark stool with no rectal bleeding; 4, gross rectal bleeding. DAI was calculated by the addition of body weight, stool consistency, and rectal bleeding scores and dividing the total number by 3. Microbial community analysis Fecal samples were shipped to Microbiome Insights company on dry ice, who performed the sequencing and analysis. 16S (V4 region) genes were sequenced on an Illumina MiSeq. Raw 130 Fastq files were quality-filtered and clustered into 97% similar operational taxonomic units (OTUs) using the mothur software pipeline [http://www.mothur.org]. 3.23428 x 105 high-quality bacterial reads were obtained. The final dataset had 3469 OTUs (including those occurring once with a count of 1) and a read range of 1.1499 x 104 and 1.9824 x 104. High-quality reads were classified using Greengenes (v. 18_9) as the reference database. OTU abundances were aggregated into genera and plotted as the relative abundances. OTU abundances were summarized with the Bray-Curtis index and a non-metric multidimensional scaling (NMDS) analysis was performed to visualize microbiome similarities in ordination plots. Permutational multivariate analysis of variance (PERMANOVA) with Bonferroni correction was used to calculate significant differences between groups. Histology Colon sections were harvested from DSS-treated WT or ERAP1-/- mice at day 7, fixed, stained with hematoxylin and eosin (H&E) and scored, as previously described [291]. Mice received a score of 0-3 for inflammation with 1, 2, and 3 corresponding to lymphocytic infiltration of the epithelium, mucosa, and transmural inflammation, respectively. Colons were then scored 0-3 for pathologic changes to intestinal architecture with 1, 2, and 3 corresponding to focal erosions, extensive erosions, and extensive ulcerations with granulation tissue, respectively. Spines were harvested, fixed for 48-72, hours and further processed for H&E and immunohistochemistry as described below. 131 Routine hematoxylin & eosin stain Tissue samples previously fixed in 10% Neutral Buffered Formalin were processed and vacuum infiltrated with paraffin on the Sakura VIP 2000 tissue processor; followed by embedding with the ThermoFisher HistoCentre III embedding station. Once blocks were cooled, excess paraffin was removed from the edges, placed on a Reichert Jung 2030 rotary microtome and faced to expose tissue sample. Once the block is faced it is cooled and finely sectioned at 4- 5 microns. Sections were dried in a 56°C slide incubator to ensure adherence to the slides for 2 – 24 hours not exceeding this temperature which would potentially destroy antigenic components. Slides were removed from the incubator and stained with a routine Hematoxylin and Eosin method as follows: Two changes of Xylene – 5 minutes each, two changes of absolute ethanol – 2 minutes each, two changes of 95% ethanol – 2 minutes each, running tap water rinse for 2 minutes, endure Hematoxylin (Cancer Diagnostics – Durham, NC) for 1 ½ minutes followed directly by a 10 – 15 second differentiation in 1% aqueous glacial acetic acid and running tap water for 2 minutes to enhance nuclear detail. Upon completion of running tap water slides were placed in one change of 95% ethanol – 2 minutes, 1% Alcoholic Eosin-Phloxine B – 2 minutes to stain cytoplasm, one change of 95% ethanol for 2 minutes, four changes of 100% ethanol – 2 minutes each, four changes of Xylene – 2 minutes each followed by coverslipping with synthetic mounting media for permanent retention and visualization with light microscopy. Immunohistochemistry of the spines – IL-23 F4/80, IgM, TNF-α primary antibodies Specimens were decalcified in 14% EDTA, processed, embedded in paraffin and sectioned on a rotary microtome at 4s. Sections were placed on slides coated with 2% 3- 132 Aminopropyltriethoxysilane and dried at 56C overnight. The slides were subsequently deparaffinized in Xylene and hydrated through descending grades of ethyl alcohol to distilled water. Slides were placed in Tris Buffered Saline pH 7.4 (Scytek Labs – Logan, UT) for 5 minutes for pH adjustment. Following TBS, slides underwent heat-induced epitope retrieval utilizing Scytek Citrate Plus Retrieval pH 6.0 or Enzyme Retrieval followed by rinses in several changes of distilled water. Endogenous Peroxidase was blocked utilizing 3% Hydrogen Peroxide / Methanol bath for 30 minutes followed by running tap and distilled water rinses. Following pre- treatments standard micro-polymer complex staining steps were performed at room temperature on the IntelliPath™ Flex Autostainer. All staining steps were followed by rinses in TBS Autowash buffer (Biocare Medical – Concord, CA). After blocking for the non-specific proteins with Rodent Block M (Biocare) for 5 or 10 minutes; sections were incubated with specific primaries in normal antibody diluent (NAD-Scytek) and incubated 60 minutes. Micro-Polymer (Biocare) reagents were subsequently applied for specified incubations followed by reaction development with Romulin AEC™ (Biocare) and counterstained with Cat Hematoxylin (see Table 2 for details). 133 Table 2: Details of antibodies used for immunohistochemistry staining of the spines. Primary Antibody Ab Vendor: Pretreatment: Primary: Staining System: Rabbit anti – IL23 Abcam #115759 Polyclonal Cambridge, MA 0.03% Pronase E in TBS for 10 minutes at 37°C 1:100 in NAD ProMark Rabbit on Rodent HRP Polymer™ - 30 minutes – 2 Hours AEC – 5 minutes CATHE Hematoxylin 1:10 – 1 minute Rat anti – F4/80 Bio-Rad # MCA497G Monoclonal Hercules, CA Overnight Retrieval overnight at 56°C 1:100 in NAD ProMark Rat on Mouse HRP Polymer™ Overnight 15 minutes – Probe Rabbit anti – IgM Polyclonal ThermoFisher #61-6800 Rockford, IL Overnight Retrieval overnight at 56°C 1:300 in NAD – 60 minutes 15 minutes – Polymer AEC – 5 minutes CATHE Hematoxylin 1:10 – 1 minute ProMark Rabbit on Rodent HRP Polymer™ - 30 minutes AEC – 5 minutes CATHE Hematoxylin 1:5 – 5 minutes Rabbit anti- TNF alpha Polyclonal Abcam #6671 Cambridge, MA Overnight Retrieval overnight at 56°C 1:100 in NAD ProMark Rabbit on Rodent HRP Polymer™ - 30 minutes – 2 Hours AEC – 5 minutes CATHE Hematoxylin 1:10 – 1 minute Isolation of lymphocytes from spleen and lymph nodes Spleen or Lymph node tissues were physically disrupted (by passage through a 40 µm sieve), followed by RBCs lysis by using 2 ml of ACK lysis buffer (Invitrogen) per sample. Cells 134 were subsequently washed twice with complete RPMI medium (RPMI 1640 (Invitrogen) supplemented with 10% FBS, 1% PSF (penicillin, streptomycin, fungizone)), resuspended, and counted. Flow cytometry For surface staining, 2 million cells were incubated with Fcγ block (BD Biosciences) and with the appropriate antibodies on ice for 45 minutes, and washed twice with FACS buffer. For intranuclear staining, following surface staining and washing, 4 million cells were fixed and stained with intranuclear antibodies using BD Pharmingen Transcription Factor Buffer Set (BD Bioscience) per manufacturer’s protocol. For intracellular staining, 4 million splenocytes were plated in a 96-well plate in presence of 50 ɳg/mL PMA (Calbiochem) and 1µg/mL Ionomycin (Calbiochem) for 30 min, after which 1 µL/well of GolgiPlug (BD Biosciences) was added, and cells were incubated for additional 5.5 hours. Cells were washed twice with FACS buffer and surface stained as above, after which BD Cytofix/Cytoperm Fixation/Permeabilization kit (BD Biosciences) was used per manufacturer’s protocol. List of antibodies used in chapter 2: PECy7-anti-CD4 (BD Biosciences), APC-Efluro780- anti-CD3 (eBioscience), Alexa488-anti-CD25 (Invitrogen), Pacific Blue-anti-Foxp3 (Invitrogen), APC-anti-IL-10 (BioLegend), PE Cy7-anti-CD49b (Invitrogen), PE-anti-Lag3 (BioLegend), PE- anti-F4/80 (eBioscience), Pacific Blue-anti-CD8 (BD Biosciences), APC-Cy7-anti-CD11b (BD Biosciences), PECy7-anti-CD11c (BD Biosciences), BV650-anti-Qa2 (BD Biosciences), FITC- 135 anti-CD45RB (BD Biosciences), all at 4 μg/ml. Live/dead Fixable Aqua Dead Cell Stain Kit (Invitrogen) was used per manufacturer’s protocol to exclude dead cells from analysis. List of antibodies in chapter 3: PE-anti-RANK (Invitrogen), FITC-anti-CD3 (Invitrogen), Pacific Blue-anti-CD4 (BD Pharmigen). Live/Dead Fixable Aqua Dead Cell Stain Kit (Invitrogen) was used per manufacturer’s protocol to eliminate dead cells from analysis. List of antibodies used in chapter 4: mix 1: PE-Cy7 – CD11c (BD Biosciences), APC- Cy7 – CD11b (BD Biosciences), PE – F4/80 (eBioscience), V450 – CD86 (BD Horizon), FITC – CD40 (eBioscience), PerCP-Cy5.5 – CD107 (CCR7) (BD Pharmigen); mix 2: APC-Cy7 – CD3e (BD Pharmigen), Alexa Fluor 700 – CD8 (BD Pharmigen), PerCP-Cy5.5 – CD19 (BD Pharmigen), PE-Cy7 – CD49b clone Dx5 (eBioscience), FITC – CD69 (BD Pharmigen, eFluor450 – CD107a (eBioscience); mix 3: APC – CD3 (BD Pharmigen), Alexa Fluor 700 – CD8a (BD Pharmigen), PE-Cy7 – CD49b clone Dx5 (eBioscience), Alexa Fluor 488 – IFN-γ (BD Pharmigen), eFluor450 – CD107a (eBioscience). Data were collected on BD LSR II instrument and analyzed using FlowJo software (Tree Star). Tr1 differentiation assay Splenocytes and lymphocytes were isolated from the spleens and lymph nodes of 4 WT and 4 ERAP1-/- mice. Cells from both tissues were processed into a homogenous single-cell suspension and combined. Naïve CD4+ T cells were isolated from this cell mixture using the MACS Mouse Naïve CD4+ T cell Isolation Kit (Miltenyi Biotec) per manufacturer’s instructions. Cells were plated in complete RPMI medium at 1x105 cells/well in a 96-well high binding plate 136 pre-coated with 2 µg/mL anti-CD3 mAb (BioLegend) and 2 µg/mL anti-CD28 mAb (BioLegend). IL-27 (Peprotech) was added to indicated wells at 25 ng/mL. The supernatant was collected from cultures at 96 hours at which point the media was replaced with media containing 0.1 µg/mL PMA (Calbiochem), 1µg/mL Ionomycin (Calbiochem), 0.2 µL/well of GolgiPlug (BD Biosciences) and 1.95µM Monesin (Sigma Aldrich). After 5 hours the cells were collected, washed and stained. Cross-fostering experiments ERAP1-/- and WT breeding cages were set up simultaneously to synchronize pregnancies. Upon showing physical signs of pregnancy, dams were removed from sires and housed separately. Within 48 hours after birth, ERAP1-/- pups were transferred to lactating WT dams (CF ERAP1-/-), and vice versa, WT pups were transferred to ERAP1-/- dams (CF WT). Control cages where WT pups were transferred to different WT dams and ERAP1-/- pups were transferred to different ERAP1-/- dams were also set up. Pups were weaned by 28 days of age. At 14 weeks of age, mice were sacrificed and their spines were harvested and fixed for µCT and histology analysis. An average threshold value of WT control mice of 948 was used for trabecular bone analysis of S1 of cross-fostered and control animals using µCT. Fecal samples were collected at 4 and 14 weeks of age and frozen in liquid nitrogen. Quantitative RT-PCR analysis cDNA was generated using SuperStrand First-Strand Synthesis Kit III (Invitrogen) from Trizol isolated RNA per the manufacturer's protocol. Quantitative RT-PCR was performed using 137 SYBR green PCR Mastermix (Life Technologies) and analyzed on a QuantStudio7 system (Thermofisher). The following primers were used in chapter 3: Hprt: For: 5’GTAATGATCAGTCAACGGGGGAC3’, Rev: 5’ CCAGCAAGCTTGCAACCTTAACCA’3; Dmp-1: For: 5’TGAGGAAGACAGTGACTCTCAGGA3’, Rev: 5’TTCTCTCACCATGTGTGCTCTGAC3’; Trap: For: 5’ACTTCCCCAGCCCTTACTACCG3’, Rev: 5’TCAGCACATAGCCCACACCG3’; Dc- Stamp: For: 5’ TGGAAGTTCACTTGAAACTACGTG3’, Rev: 5’ CTCGGTTTCCCGTCAGCCTCTCTC3’; Catk: For: 5’ ACGGAGGCATCGACTCTGAA 3’, Rev: 5’ GATGCCAAGCTTGCGTCGAT3’;Ctr: For: 5’ TGCAGACAACTCTTGGTTGG3’, Rev: 5’ TCGGTTTCTTCTCCTCTGGA3’; Opg: For: 5’ACAGTTTGCCTGGGACCAAA3’, Rev: 5’TCACAGAGGTCAATGTCTTGGA3’. To measure relative gene expression, Hprt gene measurement and the comparative threshold cycle method were used for all samples. Induction of gene expression was calculated as the relative change from the level of WT sample transcripts. 6-week-old Balb/c male mice were injected intravenously (I.V.) with 1010 v.p. of rAd5- Null (n=6), rAd5-mCRACC-Fc (n=6) or not injected - naïve (n=3). After 6 hours, mice were sacrificed, and their spleens were snap frozen in liquid nitrogen for Western Blot and RT-PCR analysis and stored at -80 ºC. RNA was extracted using the Trizol reagent (Life Technologies), per manufacturer's protocol. The following primers were used in chapter 4: Isg15: For: 5’GGTGTCCGTGACTAACTCCAT3’, Rev: 5’TGGAAAGGGTAAGACCGTCCT’3; Oas2: For: 5’TTGAAGAGGAATACATGCGGAAG3’, Rev: 5’GGGTCTGCATTACTGGCACTT3’; Ifna: For: 5’GCCTTGACACTCCTGGTACAAATGAG3’, Rev: 5’CAGCACATTGGCAGAGGAAGACAG3’; Ifnb: For: 138 5’TGGGTGGAATGAGACTATTGTTG3’, Rev: 5’CTCCCACGTCAATCTTTCCTC3’; Il6: For: 5’TAGTCCTTCCTACCCCAATTTCC3’, Rev: 5’TTGGTCCTTAGCCACTCCTTC3’; Il12: For: 5’TGGTTTGCCATCGTTTTGCTG3’, Rev: 5’ACAGAGGTTCACTGTTTCT3’; Gm- csf: For: 5’GGCCTTGGAAGCATGTAGAGG3’, Rev: 5’GGAGAACTCGTTAGAGACGACTT3’; mGapdh: For: 3’AGAACATCATCCCTGCATCC3’, Rev: 5’CACATTGGGGGTAGGAACAC3’. HEK-293 derived C7 cells were plated at 2*106/ mL in 12-well plates in 500 µL of complete RPMI (GIBCO) media. rAd5-Null and rAd5-mCRACC viruses were added at the multiplicity of infectivity (MOI) of 1,000 v.p. per cell per well in 10 µL. Mock-treated cells were treated with 10 µL of PBS. Cells were incubated overnight. After 12 hours of incubation, RNA was extracted using Trizol. Primers used: Slamf7: For: 5’GGCACATGCGTGATCAATCT3’, Rev: 5’ATCGCCAAGCGATACTCAGA3’; hGAPDH: For: 5’GGGTGTGAACCATGAGAAGTATGAC3’, Rev: 5’GCCATCCACAGTCTTCTGGGT3’. To measure relative gene expression, Gapdh gene measurement and the comparative threshold cycle method were used for all samples. Induction of gene expression was calculated as the relative change from the level of mock-treated cell transcripts to the level of recombinant Ad5-treated cell transcripts. Osteoclast outgrowths Bone marrow cells were differentiated into osteoclasts in cell culture as described previously [375], with several modifications. 250,000 cells per well were plated in 48-well plates. After overnight incubation in αMEM (Gibco) containing 10% FBS, and 50 µg/mL of 139 penicillin and streptomycin, the cells were treated with 5 ɳg/ml RANKL (R&D Systems) and 30 ɳg/ml M-CSF (Gibco) to induce osteoclastogenesis. After 5 days of culture, cells were stained with TRAP (Sigma) and imaged using light microscopy. Osteoclast size and number was determined using ImageJ software. Osteoclast pitting assay Confocal osteoclast pitting assay was essentially performed as previously described [376]. Briefly, dentine discs (Fisher, NC0443942) were biotinylated with 0.5 mM Biotinylation Reagent (Fisher, 50-492-586) for 1 hour at room temperature. Prior to seeding with osteoclasts, dentine discs were incubated in 96-well plates in αMEM (Gibco) containing 10% FBS, and 50 µg/mL of penicillin and streptomycin. Freshly isolated bone marrow cells were seeded over dentine discs and incubated overnight with 250,000 cells per well. The next day cells were supplemented with 5 ɳg/ml RANKL (R&D Systems) and 30 ɳg/ml M-CSF (Gibco). After 2 days of culture, media was replaced with acidic Resorption media. After 2 days of culture, cells were washed, fixed, and permeabilized with Triton-100. Cells were stained with FITC-streptavidin (ThermoFisher, SA10002) and Rhodamin Phalloidin (Thermo Fisher, R415) and analyzed as previously described [376]. Osteoblast outgrowths Bone marrow cells were obtained from mouse femurs and tibia and plated overnight in 48 well plates at 500,000 cells per well in αMEM (Gibco) containing 10% FBS and 50 µg/mL of penicillin and streptomycin. The next day the cells were treated with 50 µg/mL of ascorbic acid 140 (Sigma, 50-81-7) and 10 mM β-glycerophosphate (Sigma, 154804-51-0) to promote osteoblastogenesis. Media was changed every 2-3 days. For osteoblastogenesis analysis, after 5 days of culture cells were washed with PBS and fixed with 10% formalin for 60 seconds. Cells were incubated with BCIP/NBT substrate (Sigma, B5655) solution for 10 minutes and evaluated for alkaline phosphatase positive cells using ImageJ software. For calcium deposition assay, after 28 days of incubation, cells were washed, fixed, and stained with 2% Alizarin Red S (EMD Millipore, C.I. 58005) solution as described previously. For quantification of staining, alizarin release assay was performed as described before [377] and the absorbance at 405nm was measured. Serum measurements At the time of harvest, blood was collected via cardiac puncture. It was left to clot for 10 minutes at room temperature and was centrifuged at 4,000 rpm for 10 minutes. Serum was collected, frozen in liquid nitrogen, and stored at -80°C. TRAP5b and Osteocalcin were measured using a Mouse TRAP (Immunodiagnostic Systems, SB-TR103) and OC assay kits (Biomedical Technologies Inc., BT-470) per the manufacturer’s protocol. Dynamic histomorphometry As previously described, mice were injected intraperitoneally with 200μl of 10mg/ml sterile calcein (Sigma) at 14 and 4 days prior to harvest [378]. Tibias were fixed with 10% formalin for 24 hours and transferred to 70% ethanol 48 hours later. Samples were embedded, sectioned and examined using fluorescent microscopy. Ten images were taken per tibia. The 141 distance between the calcein lines and their length along the bone surface was measured and used to calculate the bone formation rate (BFR) and mineral apposition rate (MAR). Adenovirus vector construction rAd5-Null was constructed and purified, as previously described [379]. For rAd5- mCRACC-Fc vector, the CRACC extracellular domain (ECD) (NCBI Reference Sequence: NM_144539.5 (http://www.ncbi.nlm.nih.gov/nuccore/NM_144539.5)) was fused to mIgG1-Fc portion and cloned into pShuttle CMV. The rAd5-mCRACC-Fc vector was constructed and purified, as described [379]. CRACC-ECD/mIgG1-Fc was excised using primers flanked by EcoRI and HindIII restriction endonucleases (NEB, Ipswich, MA) from a plasmid (Biomatik, Delaware, USA) and sub-cloned into the pShuttle vector, which contains a CMV expression cassette. The resulting pShuttle-mCRACC-Fc plasmid was linearized with PmeI restriction enzyme and homologously recombined with the pAdEasyI Ad5 vector genome yielding pAd- mCRACC-Fc. HEK293 cells were transfected with PacI-linearized plasmid and a viable virus was obtained and amplified after several rounds of expanding infection. rAd5-mCRACC-Fc virus was purified using a CsCl2 gradient. To confirm that rAd5-mCRACC-Fc vector expresses mCRACC-Fc transgene, we performed qRT-PCR analysis to validate the expression of mCRACC-Fc following rAd5-mCRACC-Fc or rAd5-Null infection. All viruses were found to be replication competent adenovirus (RCA)-free by both RCA PCR (E1 region amplification) and direct sequencing methods. 142 Tumor challenge Intratumoral study: 8-week-old male Balb/c mice were injected subcutaneously (S.Q.) in the flank with 150,000 CT26.CL25 (ATCC, CRL-2639) cells in 100 µL of PBS. 8 days later, once visible tumors formed, mice were split randomly into 3 groups and were either injected intratumorally (I.T.) with 1010 viral particles (v.p.) of rAd5-Null (n=15), rAd5-mCRACC-Fc (n=15), or not injected (n=14). Mice were monitored every 2-3 days and their tumor width and length were measured. Using formula ½*(Length × Width2), tumor volumes were calculated. Tumor volume of 2,000 mm3 or presence of ulcerations on the tumors were used as humane end- points. Pre-vaccination using tumor lysate and adenoviruses: 6-week-old Balb/c male mice were given three doses of 200 µg of CT26 tumor lysate and 1010 v.p. of rAd5-Null (n=8), or rAd5- mCRACC-Fc (n=9) intraperitoneally (I.P.), or not injected (non-vaccinated, n=9) over a period of 5 weeks, with the 2nd dose - 3 weeks after the first, and 3rd dose - 12 days after 2nd dose. On the same day as the 3rd dose, mice were injected with 250,000 CT26 tumor cells. Mice were monitored every 2-3 days starting on day 5 post tumor challenge. Measurements and volume calculation were performed as described above. Tumor volume of 2,000 mm3 or presence of ulcerations on the tumors were used as humane end-points. Cell culture CT26.CL25 colon adenocarcinoma cells were purchased from ATCC (CRL-2639™) and utilized within 4 weeks of thawing from a master stock generated on 06/15/2016. The cells were passaged up to 6 times for in vivo experiments and up to 12 times for in vitro experiments. Cells 143 were not authenticated in the last year. Tumor lines were also verified to be mycoplasma negative by Taconic Biosciences via RAPIDMAP-27 PCR analysis. Tumor cells were cultured in modified RPMI-1640 media (ATCC, 30-2001) supplemented with 10% heat-inactivated FBS, and 1X penicillin, streptomycin, and fungizone (PSF) (Invitrogen). HEK-293C7 cells were cultured in DMEM media, supplemented with 10% FBS, 1X PSF, and hygromycin. For ADCC assays, CT26 cells were stained with 5 µM of 5(6)-Carboxyfluorescein diacetate N-succinimidyl ester (CFSE) (Sigma) in PBS at room temperature (RT) for 12 minutes, washed twice with 5% FBS and plated at 100,000 cells/well in U-bottom 96-well plates. Total splenocytes from naïve mice were incubated with CFSE-CT26 cells (20:1 Effector:Target) and cultured for 18 hours in presence of plasma (1:200 dilution) from rAd5-Null/CT26 lysate, rAd5- mCRACC-Fc/CT26 lysate vaccinated, or non-vaccinated mice. Or, Dx5+ NK cells were isolated using Miltenyi kit (130-052-501) and incubated with CFSE-CT26 cells (1:1 E:T) overnight in presence of plasma (1:50 dilution) from vaccinated mice. Cultures contained 2 ng/mL of murine IL-2 (R&D). Cells were trypsinized with 0.25% trypsin for 5 minutes at 37°C, washed, and stained with CellTrace Violet viability dye and analyzed by flow cytometry on a BD LSR II instrument. Vaccination studies 6-week-old Balb/c male mice were injected twice I.M. (n=6) over a period of 1 month (Days 0 and 12) with 200 μg of CT26 tumor lysate, tumor lysate + 1010 v.p./mouse of rAd5-Null, tumor lysate + 1010 v.p./mouse of rAd5-mCRACC-Fc, or not injected (non-vaccinated, n=4). On Day 27, mice were sacrificed, and their spleens were collected. Splenocytes were cultured in- 144 vitro in presence of 10 μg/mL of CT26 tumor lysate overnight for ELISPOT analysis or for 48 hours for Flow cytometry analysis. Innate immune study 6-week-old Balb/c male mice were injected intravenously (I.V.) with 1010 v.p. of rAd5- Null (n=6), rAd5-mCRACC-Fc (n=6), or not injected - naïve (n=3). After 10 hours, plasma and spleens were collected for Bioplex and Flow cytometry analysis, as previously described [343]. Cytokine and chemokine analysis Mouse 27-plex multiplex-based assay was used to determine cytokine/chemokine plasma concentrations via Luminex 100 per manufacturer’s protocol, and as previously described [343]. ELISA analysis ELISA assessing levels of CT26-specific antibodies in the plasma of tumor challenged mice was assessed as described previously [379]. 100 µg of CT26 lysate was plated per well in a high-binding 96-well flat-bottom plate (Corning) and incubated at 4 °C overnight. Following incubation, plates were washed with PBS, containing 0.05% Tween (Sigma-Aldrich) and incubated with blocking buffer (PBS containing 3% bovine serum albumin) for 1 hour at room temperature. Plasma was plated at 1:10, 1:50, 1:100, 1:200, 1:500 and 1:1000 dilutions. After incubation, wells were washed with wash buffer 5 times. Wells were coated with 100 µL of horseradish peroxidase (HRP)-conjugated goat anti-mouse anti IgG antibody (Bio-Rad) diluted 145 1:7,000 and incubated for 1 hour at RT. After washing wells with wash buffer 5 times, 100 µL of Tetramethylbenzidine (TMB) substrate (Sigma-Aldrich) was added to each well to initiate the spectrophotometric reaction, which was stopped with 50 µL of 2 N sulfuric acid after 30 minutes of incubation. Plates were analyzed using an automatic microplate reader at 450 nm absorbance. ELISPOT analysis Splenocytes were incubated with media alone (unstimulated), media containing 20 μg/mL of whole CT26 tumor lysate, or with 1010 v.p. of heat inactivate rAd5-Null. Plates were then incubated for 18 h in a 37ºC, 5% CO2 incubator. Plates were stained and developed using Ready-set Go IFN-γ kit (eBioscience) per the manufacturer’s protocol. Spots were counted and photographed by an automated ELISPOT reader system (Cellular Technology). Western blot Spleens from Mock, rAd5-Null, and rAd5-mCRACC-Fc intravenously injected mice 6 hours post injections were harvested, snap frozen and stored at -80ºC. For cell lysate preparation, spleens were homogenized in ice-cold lysis buffer containing 1% NP-40 Lysis Buffer (Life Technologies), protease inhibitor (Sigma-Aldrich) and phosphatase inhibitors (ThermoFisher Scientific). The concentrations of cell lysates were determined by a BCA assay (ThermoFisher Scientific). 50 μg of total protein was loaded onto 12 % gel Mini-Protean TGX Precast Gels (Bio-Rad). The proteins were transferred to nitrocellulose membrane (Amersham Protran) for 1 hour at room temperature. The membrane was blocked for 1 hour in Odyssey Blocking Buffer (Licor Biosciences), then incubated for overnight at room temperature with primary monoclonal 146 rabbit anti-human anti-STAT-1-P (1:200, R&D, cat# 1086B) or anti-mouse anti–β-actin (1:3000; Abcam, cat #8224). The blot was washed with TBS-T three times and then incubated with IRDye anti-mouse (926-32210; Licor) or anti-rabbit (925-68070) secondary Ab diluted in blocking buffer (1:10,000) for 1 h at room temperature. The blotted membrane was washed and developed on the Licor Odyssey (Licor). Once, the nitrocellulose paper was scanned and analyzed for STAT1-P, it was stripped using NewBolt IR Stripping Buffer (Licor) and re-blotted and analyzed for β-actin. Densitometric analysis was done using ImageJ software. Immunohistochemistry analysis of tumors At the end of the tumor challenge, tumors were collected from non-vaccinated (n=1), rAd5-Null/CT26 lysate (n=3) and rAd5-mCRACC-Fc/CT26 lysate (n=3) vaccinated mice. The samples were fixed in 10% Neutral Buffered Formalin. Sampled were processed, embedded in paraffin and sectioned on a rotary microtome at 4µm. Sections were placed on positively charged slides and dried at 56ºC overnight. Following deparaffinization in Xylene, slides underwent heat- induced epitope retrieval in a steamer utilizing Scytek Citrate Plus Retrieval pH 6.0 followed by rinses in several changes of distilled water. Endogenous Peroxidase was blocked via 3% Hydrogen Peroxide/Methanol bath for 30 minutes. After blocking for the non-specific proteins with Rodent Block M (Biocare) for 10 or 20 minutes, sections were incubated with appropriate primary antibodies as described in Table 3. Micro-Polymer (Biocare) reagents were subsequently applied for specified incubations followed by reaction development with Romulin AEC™ (Biocare) and counterstained with Cat Hematoxylin. 147 Table 3: Specifications of antibodies used for immunohistochemistry of tumors. Ab Vendor: Pretreatment: Primary: Staining System (BioCare Medical): Heat Retrieval – Citrate Buffer pH 6.0 – Pascal Pressure Cooker – 125°C for 15 sec, 80°C for 1 min, room temperature with lid off for 30 min Heat Retrieval – Citrate Buffer pH 6.0 – Steamer for 30 min, room temperature with lid off for 10 min 1:450 in NAD – 1 Hour Rodent Block M – 20 minutes ProMark Rabbit on Rodent HRP Polymer™ - 35 minutes AEC Chromogen – 5 minutes CATHE Hematoxylin 1:10 – 1 minute 1:100 in NAD - 1 Hour Rodent Block M – 10 minutes ProMark Rat on Mouse HRP Probe™ - 10 minutes ProMark Rat on Mouse HRP Polymer™ - 10 minutes AEC Chromogen – 5 minutes CATHE Hematoxylin 1:10 – 1 minute No Pretreatment 1:100 in Rodent Block M – 20 minutes NAD – 1 Hour ProMark Rabbit on Rodent HRP Polymer™ - 20 minute AEC Chromogen – 5 minutes CATHE Hematoxylin 1:10 – 1 minute 148 Primary Antibody Rabbit anti – CD3 Polyclonal Abcam #GR3194253- 3 Cambridge, MA Rat anti – CD8 Dianova #DIA-808 Monoclonal Hamburg, Germany Rabbit anti – Integrin Alpha 2 (DX5) Monoclonal Abcam #GR196223- 25 Cambridge, MA Statistical analysis Statistical tests performed for each data set and significance levels are indicated in the figure legends. Data in all graphs are presented as a mean ± standard error. All statistical analysis was performed using GraphPad Prism 7 (GraphPad Software). 149 Chapter 6: Concluding remarks and future directions 150 The immune system is a complex network of cellular and molecular components which provides protection from pathogens to the host and is essential for its survival. Immunodeficiency is characterized by failure of the immune system to appropriately respond to the invading pathogens, making affected individuals susceptible to infection and cancer. However, overactive immune responses can also be harmful by causing damage to host tissues. It is, therefore, a fine balance, where a properly functioning immune system is needed to provide protection to the host from pathogens and malignancy, while avoiding overactive responses that can result in damage to the host. Autoimmune diseases affect 50 million Americans and cost 100 billion of healthcare dollars [1, 2]. Chronic inflammation in autoimmune diseases is thought to occur due to a failure of regulatory mechanisms to offset the effector cell functions. It is important to study and to understand the molecular basis of the pathogenesis of autoimmune and autoinflammatory diseases in order to identify molecular and/or cellular players that can be targeted in search for cures. AS is a prototypic chronic arthritic disease characterized by spinal inflammation and ankylosis, which primarily develops in young adults [102]. Despite years of research, the exact mechanism underlying the causation of AS has not been elucidated. GWAS studies have identified multiple ERAP1 polymorphisms associated with increased risk of developing AS [129]. In chapter 2 we evaluated the skeletal system of the ERAP1-/- mice and determined that these mice developed spinal fusions between L6 and S1 vertebral transverse processes. These mice also developed erosions at the sacroiliac joints, inflammatory infiltrates at the intervertebral disc spaces and systemic osteoporosis. In chapter 3 we performed detailed analysis of the bone remodeling cells in ERAP1-/- mice, which showed increased osteoclastogenesis and osteoclast 151 activity, along with increased RANKL levels on the CD4+ T cells which may be responsible for the enhanced osteoclastogenesis [315]. Osteoblast analysis revealed a reduced function in vivo, which was accompanied by reduced Runx2 gene expression levels, important for osteoblast differentiation and function [316]. Altered bone remodeling functions in these mice were consistent with the systemic osteoporosis in ERAP1-/- mice. Both IL-17 and TNF-α blockade failed to improve ankylosis and osteoporosis in the spines of ERAP1-/- mice downplaying the role of these cytokines in the skeletal pathogenesis. Interestingly, while TNF-α blockade improved osteoblastogenesis in vitro, this therapy still failed to improve osteoporosis. In lieu of recent studies suggesting therapeutic potential of IL-23 blockade in AS [227], in addition to our observation of IL-23 positive staining in the spines of ERAP1-/- mice, it will be interesting and important to test efficacy of IL-23 blockade alone and in combination with anti-TNF-α in preventing spinal fusions in the ERAP1-/- mice. In chapter 2 in addition to skeletal abnormalities, we also showed that ERAP1-/- mice developed spontaneous gut dysbiosis and had increased susceptibility to chemically induced colitis. This closely parallels human disease, since AS patients have altered gut microbiota [213], develop microscopic inflammation in the terminal ileum [108, 285], and have increased susceptibility to inflammatory bowel disease [108]. While we discovered an interesting phenomenon of dysbiosis being present in the ERAP1-/- mice, our experiments did not support a causative link between the dysbiosis and bone pathogenesis in ERAP1-/- mice. Since transfer of unhealthy microbiota from ERAP1-/- mice did not induce osteoporosis or spinal fusions in WT mice, and vice versa correction of unhealthy microbiota in ERAP1-/- mice did not improve skeletal phenotypes, it is apparent that the imbalance of the commensal bacteria in the gut is not the cause of pathogenesis in ERAP1-/- mice. While it is still possible that several microbial 152 species that were not corrected in ERAP1-/- mice, namely Bacteroides, Parabacteroides, and Christensenellacea genus are responsible for the pathogenesis in these mice, it is also possible that the dysbiosis is simply a consequence of the variation in the antigen presentation in ERAP1- /- mice. Since ERAP1 is absent in these mice, their peptide repertoire available for loading into MHC-I molecules is different from those of WT mice, resulting in tolerance toward different bacteria from that of WT mice with functional ERAP1. To completely rule out the role of microbiota in the pathogenesis observed in ERAP1-/- mice, individual transfer of the genera which did not successfully colonize ERAP1-/- upon cross-fostering would be required. Finally, it would be of great interest to determine whether correction of dysbiosis in ERAP1-/- mice upon cross-fostering improves its susceptibility to DSS-induced colitis, as there is a lot of supporting evidence for the role of microbiota in intestinal health [380]. Immunologic analysis of the ERAP1-/- mice revealed reduced levels of tolerogenic dendritic cells and type 1 regulatory T cell, which may be responsible for overactive immune responses in the ERAP1-/- mice. It is known that HLA-G is important for the development of tDCs, which are in turn required for Tr1 cell development [381]. In fact, patients with reduced HLA-G levels have been shown to have deficits in tDCs and Tr1 cells. Interestingly, silencing of ERAP1 in trophoblastic cells has been shown to reduce HLA-G expression levels [299]. This suggests that ERAP1 may play a role in Tr1 cell development via its function as a molecular ruler influencing HLA-G expression levels. In agreement with this, we also observed reduced levels of Qa-2, a functional homolog of HLA-G in mice [382, 383], which is consistent with previously published results [384]. Taken together, our hypothesis is that ERAP1 polymorphisms affecting ERAP1’s trimming activity result in reduced peptide loading and surface expression of HLA-G, which ultimately leads to reduced interactions between tDCs and naïve CD4+ T cells, 153 resulting in reduced Tr1 cell differentiation and loss of tolerance. It is important to show the direct relationship between ERAP1-mediated deficiency of Tr1 cells in humans. Once this relationship is proven, it will be important to investigate the relationship between ERAP1 SNPs and HLA-G levels in patients with autoimmune diseases and in healthy controls. Several in-vitro experiments can be performed to confirm ERAP1’s role in the pathogenesis of Tr1 cell development. This can be done by first measuring HLA-G levels on human PBMCs in the presence and absence of ERAP1 shRNA. If HLA-G levels go down with silencing of ERAP1, then the direct relationship will be confirmed between ERAP1 and HLA-G in human cells. To confirm ERAP1’s role in tDC (also called DC-10s in humans) development, monocytes isolated from human PBMCs can be cultured in presence of GM-CSF, IL-4 and IL-10 for 5 days and LPS for two more days in order to stimulate their differentiation into DC-10s [87]. If in the presence of ERAP1 shRNA, the differentiation of monocytes into DC-10s is reduced, this will support our hypothesis that ERAP1 modulates DC-10 levels via HLA-G surface expression. Similar experiments can be performed on naïve T cells isolated from human PBMCs, where the ability of naïve T cells to differentiate into Tr1 cells is measured in absence or presence of ERAP1, by utilizing ERAP1 shRNA lentivirus in presence of IL-10 and IL-27 cytokines. IL-27 promotes the differentiation of naive CD4+ T cells into Tr1 cells and induces them to produce IL-10 [81]. In chapter 2 we showed that culturing naive CD4+ T cells in the presence of IL-27 in cell culture did not correct the Tr1 cell deficit. It is possible that ERAP1-/- T cells are not able to express sufficient levels of IL-10 in response to IL-27 cytokine signaling. While it is possible, this would be surprising since we have not previously noted defects in IL-10 production or STAT signaling in ERAP1-/- mice [163]. We have not been successful thus far in obtaining a 154 sufficient number of cells to perform an IL-10 secretion assay or RT-PCR analysis of Tr1 cells. However, in the future experiments RNA and protein level analyses performed on isolated Tr1 cells should be used to investigate transcription and translation levels of the molecular players involved in IL-27 mediated Tr1 differentiation and IL-10 production. Use of IL-10 expressing lentivirus has been shown to promote differentiation of conventional CD4+ T cell into IL-10 producing “Tr1-like” cells [98]. It will be important to test whether ERAP1 deficient naïve T cells can differentiate into Tr1 cells utilizing IL-10 lentivirus and whether overexpression of IL- 10 can overcome the ERAP1-mediated Tr1 deficit. Co-culture experiments will also provide an important insight into the importance of DC10 – naive CD4+ T cell interactions for ERAP1- dependent Tr1 differentiation. Different combinations of "ERAP+" or "ERAP1-" DC10s and naïve T cell co-cultures can be evaluated. If the underlying problem of generating Tr1 cells in ERAP1 deficient cells is due to reduced DC-10s and insufficient DC10-naive CD4+ T cell interactions, then the co-culture of "ERAP1+" DC10 and "ERAP1-" CD4+T cells, will allow for a normal production of Tr1 cells. Vice versa, culturing "ERAP1-" DC10 and "ERAP1+" CD4+ T cells will result in reduced production of Tr1 cells if the underlying cause stems from reduced HLA-G mediated interactions between DC-10s and naive CD4+T cells. Finally, it is possible that the deficiency of Tr1 cells in the ERAP1-/- mice are not due to aberrant production but are due to reduced survival of these cells. This can be reconciled by performing experiments where the lifespan of ERAP1-/- derived Tr1 cells is measured compared to WT over time. This can be done in vitro where CFSE-labeled Tr1 cells are expanded in culture, and in vivo, where CFSE-labeled Tr1 cells in infused into the mice and tracked over time compared to WT [385]. To prove the causative relationship between Tr1 cells and the phenotypes observed in ERAP1-/- mice, it is important to perform in vivo studies correcting the deficiency of Tr1 cells in 155 ERAP1-/- mice. This can be done in several ways. One way is to correct the Tr1 levels by anti- CD3 antibody intranasal administration. This method allows for the proliferation of Tr1 cells in vivo via stimulation of the DCs in the upper respiratory airways to produce IL-27 [386]. Another way involves an adoptive transfer of WT Tr1 cells into ERAP1-/- mice to test their effect on various phenotypes. First and foremost, testing the effect of correcting Tr1 cell deficiency on the skeletal phenotypes in ERAP1-/- mice will be important. Based on our work, while the severity of osteoporosis, spinal inflammation, and erosions progress over time, the severity of fusions does not substantially change, so it will be important to assess the effects of Tr1 cell therapy on all skeletal phenotypes including the mean AS fusions, trabecular bone measurements, severity of erosions and inflammation of the spine. Since there is no certainty that formed fusions can be reversed, it will be important to assess spinal inflammation in response to Tr1 cell infusion, as it may be the underlying mechanism for local tissue damage and subsequent osteoproliferation [177]. Spinal inflammation may be monitored via magnetic resonance imaging or histologically. It will also be important and interesting to test whether adoptive transfer of Tr1 cells can reduce DSS susceptibility in ERAP1-/- mice. DSS-challenged ERAP1-/- mice developed severe inflammation in the gut along with increased disease activity index. Given that Tr1 cells have been implicated in IBD [62] and their infusion has been shown to provide symptomatic relief to Crohn's patients [97], adoptively transferred Tr1 may aid in halting immune responses and reducing tissue damage in the DSS-challenged ERAP1-/- mice, proving their involvement in the pathogenesis of increased susceptibility to colitis in these mice. By three weeks of age ERAP1-/- pups develop axial fusions, and it is not entirely clear whether these fusions can be reversed. It is possible that the mice are born with the fusions and they remain stable over time. To address this question, fetal pups need to be scanned in-utero 156 close to delivery. If the pups develop the fusions in-utero, the ERAP1-/- mice are not a good model for testing curative therapies and prevention strategies for ankylosis. If this is the case, the best solution for future work involving therapy development targeted at prevention of ankylosis would be to generate inducible ERAP1-/- mice. This, for example, can be done by utilizing the Cre-loxP system, where ERAP1 exons will be flanked with loxP sites in mice that will be bred with the Cre-ERt2 mice where Cre recombinase is fused with a modified ligand-binding portion of estrogen receptor [387]. Upon administration of inducer agent (in this case, tamoxifen), which will specifically bind to Cre-ERt2 allowing for its translocation to the nucleus and excision of the floxed ERAP1 gene. This would allow for controlled induction of spinal fusions in skeletally mature mice, where interventions can be given simultaneously with the inducer agent. In summary, in this dissertation, we demonstrated that ERAP1-/- mice are the first animal model to spontaneously develop all key skeletal and intestinal features of AS, making it a valuable model for studying the pathogenesis of AS and for testing efficacy of therapeutic agents. We also discovered that ERAP1 is important for Tr1 cell development suggesting the possible role of Tr1 cells in the pathogenesis of AS. ERAP1 gene polymorphisms have also been linked to multiple autoimmune and inflammatory diseases, such as MS [167], IBD, IDDM [164], psoriasis [165] and others. It is possible that the mechanism underlying the pathogenesis of ERAP1 associated diseases may be mediated via ERAP1’s functions in Tr1 cell differentiation. This new insight justifies future studies aimed at dissecting the link between ERAP1 and Tr1 cell functions. In autoinflammatory diseases such as AS, the pathology arises from overactive immune responses, which can be reduced via inhibition of inflammatory responses or via enhancement of the regulatory T cell function. Meanwhile, on the opposite side of the spectrum, the TME is 157 largely immunosuppressive in order to evade the immune surveillance. Immune stimulating strategies can be used to augment immune responses and enhance tumor killing. Interestingly, since the absence of ERAP1 gene strengthens innate immune responses, and NK cell activity in particular [163], ERAP1 may be a great target for tumor immunomodulation. Interestingly, it has been shown that inhibition of ERAP1 enhanced NK cytotoxic activity and resulted in an enhanced killing of lymphoblastoid cells [388]. Utilization of ERAP1 inhibition as a cancer immunotherapy needs further investigation. In this work, we utilized mCRACC-Fc fusion protein on an Ad5 platform to target SLAMF7 receptor as an immunomodulation strategy. This strategy was successful in stimulating innate and adaptive immune responses, enhancing tumor-specific memory responses, promoting leukocyte infiltration of the tumors and augmenting tumor killing by NK cells. Most importantly, rAd5-mCRACC-Fc intratumoral injections increased survival of CT26 colon adenocarcinoma challenged mice. This is a great example of utilizing immunomodulation in order to enhance anti-tumor immune responses. In the future, it will be important to test this strategy in other tumor models, such as the more aggressive B16 melanoma model. Additionally, it will be interesting to test the mCRACC-Fc fusion protein in combination with other immune modulators such as anti- PD-1, CTLA-4, and Tim-3 targeted therapies. SLAMF7 is a promising target for tumor immune-modulation that will hopefully be developed further and eventually tested in human trials. 158 BIBLIOGRAPHY 159 BIBLIOGRAPHY 1. 2. 3. 4. 5. 6. 7. 8. 9. Rosenblum, M.D., et al., Treating human autoimmunity: current practice and future prospects. Sci Transl Med, 2012. 4(125): p. 125sr1. Rosenblum, M.D., K.A. Remedios, and A.K. Abbas, Mechanisms of human autoimmunity. J Clin Invest, 2015. 125(6): p. 2228-33. McCluggage, L.K., Safety of TNF inhibitors in adolescents and children. Adolesc Health Med Ther, 2011. 2: p. 1-8. Turvey, S.E. and D.H. Broide, Innate immunity. J Allergy Clin Immunol, 2010. 125(2 Suppl 2): p. S24-32. Hazlett, L. and M. Wu, Defensins in innate immunity. Cell Tissue Res, 2011. 343(1): p. 175-88. Medzhitov, R., Recognition of microorganisms and activation of the immune response. Nature, 2007. 449(7164): p. 819-26. Gasteiger, G., et al., Cellular Innate Immunity: An Old Game with New Players. J Innate Immun, 2017. 9(2): p. 111-125. Kawai, T. and S. Akira, The role of pattern-recognition receptors in innate immunity: update on Toll-like receptors. Nat Immunol, 2010. 11(5): p. 373-84. Rathinam, V.A., S.K. Vanaja, and K.A. Fitzgerald, Regulation of inflammasome signaling. Nat Immunol, 2012. 13(4): p. 333-42. 10. Wu, J. and Z.J. Chen, Innate immune sensing and signaling of cytosolic nucleic acids. Annu Rev Immunol, 2014. 32: p. 461-88. 11. 12. 13. 14. Tang, D., et al., PAMPs and DAMPs: signal 0s that spur autophagy and immunity. Immunol Rev, 2012. 249(1): p. 158-75. O'Neill, L.A. and A.G. Bowie, The family of five: TIR-domain-containing adaptors in Toll-like receptor signalling. Nat Rev Immunol, 2007. 7(5): p. 353-64. Bonilla, F.A. and H.C. Oettgen, Adaptive immunity. J Allergy Clin Immunol, 2010. 125(2 Suppl 2): p. S33-40. Kabelitz, D. and R. Medzhitov, Innate immunity--cross-talk with adaptive immunity through pattern recognition receptors and cytokines. Curr Opin Immunol, 2007. 19(1): p. 1-3. 15. Hedrick, S.M., Thymus lineage commitment: a single switch. Immunity, 2008. 28(3): p. 297-9. 160 16. Ma, D., Y. Wei, and F. Liu, Regulatory mechanisms of thymus and T cell development. Dev Comp Immunol, 2013. 39(1-2): p. 91-102. 17. Davis, M.M. and P.J. Bjorkman, T-cell antigen receptor genes and T-cell recognition. Nature, 1988. 334(6181): p. 395-402. 18. Dustin, M.L., The cellular context of T cell signaling. Immunity, 2009. 30(4): p. 482-92. 19. 20. 21. 22. Sprent, J., Antigen-presenting cells. Professionals and amateurs. Curr Biol, 1995. 5(10): p. 1095-7. Fleury, S.G., G. Croteau, and R.P. Sekaly, CD4 and CD8 recognition of class II and class I molecules of the major histocompatibility complex. Semin Immunol, 1991. 3(3): p. 177- 85. Smith-Garvin, J.E., G.A. Koretzky, and M.S. Jordan, T cell activation. Annu Rev Immunol, 2009. 27: p. 591-619. Pipkin, M.E., et al., Interleukin-2 and inflammation induce distinct transcriptional programs that promote the differentiation of effector cytolytic T cells. Immunity, 2010. 32(1): p. 79-90. 23. Raphael, I., et al., T cell subsets and their signature cytokines in autoimmune and inflammatory diseases. Cytokine, 2015. 74(1): p. 5-17. 24. Berger, A., Th1 and Th2 responses: what are they? BMJ, 2000. 321(7258): p. 424. 25. Korn, T., et al., IL-17 and Th17 Cells. Annu Rev Immunol, 2009. 27: p. 485-517. 26. 27. Xu, H.M., Th1 cytokine-based immunotherapy for cancer. Hepatobiliary Pancreat Dis Int, 2014. 13(5): p. 482-94. Sallusto, F. and A. Lanzavecchia, Heterogeneity of CD4+ memory T cells: functional modules for tailored immunity. Eur J Immunol, 2009. 39(8): p. 2076-82. 28. Tesmer, L.A., et al., Th17 cells in human disease. Immunol Rev, 2008. 223: p. 87-113. 29. 30. Le Bon, A., et al., Direct stimulation of T cells by type I IFN enhances the CD8+ T cell response during cross-priming. J Immunol, 2006. 176(8): p. 4682-9. Trapani, J.A. and M.J. Smyth, Functional significance of the perforin/granzyme cell death pathway. Nat Rev Immunol, 2002. 2(10): p. 735-47. 31. Martinez-Forero, I., et al., IL-10 suppressor activity and ex vivo Tr1 cell function are impaired in multiple sclerosis. Eur J Immunol, 2008. 38(2): p. 576-86. 32. Pearce, E.L. and H. Shen, Generation of CD8 T cell memory is regulated by IL-12. J Immunol, 2007. 179(4): p. 2074-81. 161 33. Williams, M.A., A.J. Tyznik, and M.J. Bevan, Interleukin-2 signals during priming are required for secondary expansion of CD8+ memory T cells. Nature, 2006. 441(7095): p. 890-3. 34. 35. 36. 37. 38. 39. 40. Samji, T. and K.M. Khanna, Understanding memory CD8(+) T cells. Immunol Lett, 2017. 185: p. 32-39. Churlaud, G., et al., Human and Mouse CD8(+)CD25(+)FOXP3(+) Regulatory T Cells at Steady State and during Interleukin-2 Therapy. Front Immunol, 2015. 6: p. 171. LeBien, T.W. and T.F. Tedder, B lymphocytes: how they develop and function. Blood, 2008. 112(5): p. 1570-80. Hoffman, W., F.G. Lakkis, and G. Chalasani, B Cells, Antibodies, and More. Clin J Am Soc Nephrol, 2016. 11(1): p. 137-54. Shinnakasu, R. and T. Kurosaki, Regulation of memory B and plasma cell differentiation. Curr Opin Immunol, 2017. 45: p. 126-131. Allen, C.D., T. Okada, and J.G. Cyster, Germinal-center organization and cellular dynamics. Immunity, 2007. 27(2): p. 190-202. King, C., S.G. Tangye, and C.R. Mackay, T follicular helper (TFH) cells in normal and dysregulated immune responses. Annu Rev Immunol, 2008. 26: p. 741-66. 41. Meffre, E. and H. Wardemann, B-cell tolerance checkpoints in health and autoimmunity. Curr Opin Immunol, 2008. 20(6): p. 632-8. 42. Forthal, D.N., Functions of Antibodies. Microbiol Spectr, 2014. 2(4): p. 1-17. 43. Forbes, S.J., et al., Association of a protective monoclonal IgA with the O antigen of Salmonella enterica serovar Typhimurium impacts type 3 secretion and outer membrane integrity. Infect Immun, 2012. 80(7): p. 2454-63. 44. Klasse, P.J., et al., How can HIV-type-1-Env immunogenicity be improved to facilitate antibody-based vaccine development? AIDS Res Hum Retroviruses, 2012. 28(1): p. 1-15. 45. Walport, M.J., Complement. First of two parts. N Engl J Med, 2001. 344(14): p. 1058-66. 46. 47. Clynes, R.A., et al., Inhibitory Fc receptors modulate in vivo cytotoxicity against tumor targets. Nat Med, 2000. 6(4): p. 443-6. Cyster, J.G., S.B. Hartley, and C.C. Goodnow, Competition for follicular niches excludes self-reactive cells from the recirculating B-cell repertoire. Nature, 1994. 371(6496): p. 389-95. 48. Gallegos, A.M. and M.J. Bevan, Central tolerance: good but imperfect. Immunol Rev, 2006. 209: p. 290-6. 162 49. 50. 51. 52. Kushwah, R. and J. Hu, Role of dendritic cells in the induction of regulatory T cells. Cell Biosci, 2011. 1(1): p. 20. Gershon, R.K. and K. Kondo, Cell interactions in the induction of tolerance: the role of thymic lymphocytes. Immunology, 1970. 18(5): p. 723-37. Sakaguchi, S., et al., Regulatory T cells and immune tolerance. Cell, 2008. 133(5): p. 775-87. Sakaguchi, S., T. Takahashi, and Y. Nishizuka, Study on cellular events in post- thymectomy autoimmune oophoritis in mice. II. Requirement of Lyt-1 cells in normal female mice for the prevention of oophoritis. J Exp Med, 1982. 156(6): p. 1577-86. 53. Singh, B., et al., Control of intestinal inflammation by regulatory T cells. Immunol Rev, 2001. 182: p. 190-200. 54. Wu, Y., et al., FOXP3 controls regulatory T cell function through cooperation with NFAT. Cell, 2006. 126(2): p. 375-87. 55. Ochs, H.D., S.F. Ziegler, and T.R. Torgerson, FOXP3 acts as a rheostat of the immune response. Immunol Rev, 2005. 203: p. 156-64. 56. Malek, T.R. and I. Castro, Interleukin-2 receptor signaling: at the interface between tolerance and immunity. Immunity, 2010. 33(2): p. 153-65. Laurence, A., et al., Interleukin-2 signaling via STAT5 constrains T helper 17 cell generation. Immunity, 2007. 26(3): p. 371-81. Burchill, M.A., et al., Interleukin-2 receptor signaling in regulatory T cell development and homeostasis. Immunol Lett, 2007. 114(1): p. 1-8. Yao, Z., et al., Nonredundant roles for Stat5a/b in directly regulating Foxp3. Blood, 2007. 109(10): p. 4368-75. Vignali, D.A., L.W. Collison, and C.J. Workman, How regulatory T cells work. Nat Rev Immunol, 2008. 8(7): p. 523-32. Takahashi, T., et al., Immunologic self-tolerance maintained by CD25+CD4+ naturally anergic and suppressive T cells: induction of autoimmune disease by breaking their anergic/suppressive state. Int Immunol, 1998. 10(12): p. 1969-80. Groux, H., et al., A CD4+ T-cell subset inhibits antigen-specific T-cell responses and prevents colitis. Nature, 1997. 389(6652): p. 737-42. Bacchetta, R., et al., Growth and expansion of human T regulatory type 1 cells are independent from TCR activation but require exogenous cytokines. Eur J Immunol, 2002. 32(8): p. 2237-45. 163 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. Roncarolo, M.G., et al., Tr1 cells and the counter-regulation of immunity: natural mechanisms and therapeutic applications. Curr Top Microbiol Immunol, 2014. 380: p. 39-68. Zeng, H., et al., Type 1 regulatory T cells: a new mechanism of peripheral immune tolerance. Cell Mol Immunol, 2015. 12(5): p. 566-71. Gagliani, N., et al., Coexpression of CD49b and LAG-3 identifies human and mouse T regulatory type 1 cells. Nat Med, 2013. 19(6): p. 739-46. Taga, K. and G. Tosato, IL-10 inhibits human T cell proliferation and IL-2 production. J Immunol, 1992. 148(4): p. 1143-8. Gorelik, L., S. Constant, and R.A. Flavell, Mechanism of transforming growth factor beta-induced inhibition of T helper type 1 differentiation. J Exp Med, 2002. 195(11): p. 1499-505. Gorelik, L. and R.A. Flavell, Transforming growth factor-beta in T-cell biology. Nat Rev Immunol, 2002. 2(1): p. 46-53. Vieira, P.L., et al., IL-10-secreting regulatory T cells do not express Foxp3 but have comparable regulatory function to naturally occurring CD4+CD25+ regulatory T cells. J Immunol, 2004. 172(10): p. 5986-93. de Waal Malefyt, R., et al., Interleukin 10 (IL-10) and viral IL-10 strongly reduce antigen-specific human T cell proliferation by diminishing the antigen-presenting capacity of monocytes via downregulation of class II major histocompatibility complex expression. J Exp Med, 1991. 174(4): p. 915-24. Gregori, S., K.S. Goudy, and M.G. Roncarolo, The cellular and molecular mechanisms of immuno-suppression by human type 1 regulatory T cells. Front Immunol, 2012. 3: p. 30. Satoguina, J.S., et al., T regulatory-1 cells induce IgG4 production by B cells: role of IL- 10. J Immunol, 2005. 174(8): p. 4718-26. 74. Meiler, F., et al., Distinct regulation of IgE, IgG4 and IgA by T regulatory cells and toll- like receptors. Allergy, 2008. 63(11): p. 1455-63. 75. Grossman, W.J., et al., Differential expression of granzymes A and B in human cytotoxic lymphocyte subsets and T regulatory cells. Blood, 2004. 104(9): p. 2840-8. 76. Magnani, C.F., et al., Killing of myeloid APCs via HLA class I, CD2 and CD226 defines a novel mechanism of suppression by human Tr1 cells. Eur J Immunol, 2011. 41(6): p. 1652-62. 77. Arif, S., et al., Autoreactive T cell responses show proinflammatory polarization in diabetes but a regulatory phenotype in health. J Clin Invest, 2004. 113(3): p. 451-63. 164 78. 79. Astier, A.L., et al., Alterations in CD46-mediated Tr1 regulatory T cells in patients with multiple sclerosis. J Clin Invest, 2006. 116(12): p. 3252-7. Veldman, C., et al., Type I regulatory T cells specific for desmoglein 3 are more frequently detected in healthy individuals than in patients with pemphigus vulgaris. J Immunol, 2004. 172(10): p. 6468-75. 80. Brockmann, L., et al., IL-10 Receptor Signaling Is Essential for TR1 Cell Function In Vivo. J Immunol, 2017. 198(3): p. 1130-1141. 81. Murugaiyan, G., et al., IL-27 is a key regulator of IL-10 and IL-17 production by human CD4+ T cells. J Immunol, 2009. 183(4): p. 2435-43. 82. 83. 84. 85. Apetoh, L., et al., The aryl hydrocarbon receptor interacts with c-Maf to promote the differentiation of type 1 regulatory T cells induced by IL-27. Nat Immunol, 2010. 11(9): p. 854-61. Pot, C., et al., Cutting edge: IL-27 induces the transcription factor c-Maf, cytokine IL-21, and the costimulatory receptor ICOS that coordinately act together to promote differentiation of IL-10-producing Tr1 cells. J Immunol, 2009. 183(2): p. 797-801. Schmetterer, K.G., et al., STAT3 governs hyporesponsiveness and granzyme B-dependent suppressive capacity in human CD4+ T cells. FASEB J, 2015. 29(3): p. 759-71. Zhang, P., et al., Eomesodermin promotes the development of type 1 regulatory T (TR1) cells. Sci Immunol, 2017. 2(10). 86. Wakkach, A., et al., Characterization of dendritic cells that induce tolerance and T regulatory 1 cell differentiation in vivo. Immunity, 2003. 18(5): p. 605-17. Gregori, S., et al., Differentiation of type 1 T regulatory cells (Tr1) by tolerogenic DC-10 requires the IL-10-dependent ILT4/HLA-G pathway. Blood, 2010. 116(6): p. 935-44. Ferreira, L.M.R., et al., HLA-G: At the Interface of Maternal-Fetal Tolerance. Trends Immunol, 2017. 38(4): p. 272-286. Gregori, S., C.F. Magnani, and M.G. Roncarolo, Role of human leukocyte antigen-G in the induction of adaptive type 1 regulatory T cells. Hum Immunol, 2009. 70(12): p. 966- 9. Ristich, V., et al., Tolerization of dendritic cells by HLA-G. Eur J Immunol, 2005. 35(4): p. 1133-42. Gagliani, N., et al., Transplant tolerance to pancreatic islets is initiated in the graft and sustained in the spleen. Am J Transplant, 2013. 13(8): p. 1963-75. 165 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. Yu, H., et al., Intestinal type 1 regulatory T cells migrate to periphery to suppress diabetogenic T cells and prevent diabetes development. Proc Natl Acad Sci U S A, 2017. 114(39): p. 10443-10448. Pellerin, L., et al., Peanut-specific type 1 regulatory T cells induced in vitro from allergic subjects are functionally impaired. J Allergy Clin Immunol, 2018. 141(1): p. 202-213 e8. Brun, V., et al., Isolation of functional autologous collagen-II specific IL-10 producing Tr1 cell clones from rheumatoid arthritis blood. Int Immunopharmacol, 2011. 11(8): p. 1074-8. Bacchetta, R., et al., Immunological Outcome in Haploidentical-HSC Transplanted Patients Treated with IL-10-Anergized Donor T Cells. Front Immunol, 2014. 5: p. 16. Locafaro, G., et al., IL-10-Engineered Human CD4(+) Tr1 Cells Eliminate Myeloid Leukemia in an HLA Class I-Dependent Mechanism. Mol Ther, 2017. 25(10): p. 2254- 2269. Desreumaux, P., et al., Safety and efficacy of antigen-specific regulatory T-cell therapy for patients with refractory Crohn's disease. Gastroenterology, 2012. 143(5): p. 1207- 1217 e2. Andolfi, G., et al., Enforced IL-10 expression confers type 1 regulatory T cell (Tr1) phenotype and function to human CD4(+) T cells. Mol Ther, 2012. 20(9): p. 1778-90. Brown, M.A., T. Kenna, and B.P. Wordsworth, Genetics of ankylosing spondylitis-- insights into pathogenesis. Nat Rev Rheumatol, 2016. 12(2): p. 81-91. 100. Zambrano-Zaragoza, J.F., et al., Ankylosing spondylitis: from cells to genes. Int J Inflam, 2013. 2013: p. 501653. 101. Reveille, J.D., J.P. Witter, and M.H. Weisman, Prevalence of axial spondylarthritis in the United States: estimates from a cross-sectional survey. Arthritis Care Res (Hoboken), 2012. 64(6): p. 905-10. 102. Dean, L.E., et al., Global prevalence of ankylosing spondylitis. Rheumatology (Oxford), 2014. 53(4): p. 650-7. 103. Braun, J. and J. Sieper, Ankylosing spondylitis. Lancet, 2007. 369(9570): p. 1379-90. 104. Mitra, D., et al., The prevalence of vertebral fractures in mild ankylosing spondylitis and their relationship to bone mineral density. Rheumatology (Oxford), 2000. 39(1): p. 85-9. 105. Klingberg, E., et al., Osteoporosis in ankylosing spondylitis - prevalence, risk factors and methods of assessment. Arthritis Res Ther, 2012. 14(3): p. R108. 106. Magrey, M. and M.A. Khan, Osteoporosis in ankylosing spondylitis. Curr Rheumatol Rep, 2010. 12(5): p. 332-6. 166 107. Taurog, J.D., A. Chhabra, and R.A. Colbert, Ankylosing Spondylitis and Axial Spondyloarthritis. N Engl J Med, 2016. 374(26): p. 2563-74. 108. El Maghraoui, A., Extra-articular manifestations of ankylosing spondylitis: prevalence, characteristics and therapeutic implications. Eur J Intern Med, 2011. 22(6): p. 554-60. 109. Feldtkeller, E., et al., Age at disease onset and diagnosis delay in HLA-B27 negative vs. positive patients with ankylosing spondylitis. Rheumatol Int, 2003. 23(2): p. 61-6. 110. van Tubergen, A., The changing clinical picture and epidemiology of spondyloarthritis. Nat Rev Rheumatol, 2015. 11(2): p. 110-8. 111. Brown, M.A., et al., Susceptibility to ankylosing spondylitis in twins: the role of genes, HLA, and the environment. Arthritis Rheum, 1997. 40(10): p. 1823-8. 112. Breban, M., Genetics of spondyloarthritis. Best Pract Res Clin Rheumatol, 2006. 20(3): p. 593-9. 113. Breban, M., et al., The genetics of spondyloarthropathies. Joint Bone Spine, 2006. 73(4): p. 355-62. 114. Calin, A., et al., Genetic differences between B27 positive patients with ankylosing spondylitis and B27 positive healthy controls. Arthritis Rheum, 1983. 26(12): p. 1460-4. 115. van der Linden, S., H. Valkenburg, and A. Cats, The risk of developing ankylosing spondylitis in HLA-B27 positive individuals: a family and population study. Br J Rheumatol, 1983. 22(4 Suppl 2): p. 18-9. 116. Schlosstein, L., et al., High association of an HL-A antigen, W27, with ankylosing spondylitis. N Engl J Med, 1973. 288(14): p. 704-6. 117. Brewerton, D.A., et al., Ankylosing spondylitis and HL-A 27. Lancet, 1973. 1(7809): p. 904-7. 118. Cortes, A., et al., Major histocompatibility complex associations of ankylosing spondylitis are complex and involve further epistasis with ERAP1. Nat Commun, 2015. 6: p. 7146. 119. Bessoles, S., et al., Adaptations of Natural Killer Cells to Self-MHC Class I. Front Immunol, 2014. 5: p. 349. 120. Smith, J.A., E. Marker-Hermann, and R.A. Colbert, Pathogenesis of ankylosing spondylitis: current concepts. Best Pract Res Clin Rheumatol, 2006. 20(3): p. 571-91. 121. Taurog, J.D., et al., Spondylarthritis in HLA-B27/human beta2-microglobulin-transgenic rats is not prevented by lack of CD8. Arthritis Rheum, 2009. 60(7): p. 1977-84. 167 122. Bird, L.A., et al., Lymphoblastoid cells express HLA-B27 homodimers both intracellularly and at the cell surface following endosomal recycling. Eur J Immunol, 2003. 33(3): p. 748-59. 123. Chan, A.T., et al., Expansion and enhanced survival of natural killer cells expressing the killer immunoglobulin-like receptor KIR3DL2 in spondylarthritis. Arthritis Rheum, 2005. 52(11): p. 3586-95. 124. Mear, J.P., et al., Misfolding of HLA-B27 as a result of its B pocket suggests a novel mechanism for its role in susceptibility to spondyloarthropathies. J Immunol, 1999. 163(12): p. 6665-70. 125. Colbert, R.A., et al., HLA-B27 misfolding and spondyloarthropathies. Adv Exp Med Biol, 2009. 649: p. 217-34. 126. DeLay, M.L., et al., HLA-B27 misfolding and the unfolded protein response augment interleukin-23 production and are associated with Th17 activation in transgenic rats. Arthritis Rheum, 2009. 60(9): p. 2633-43. 127. Sherlock, J.P., et al., IL-23 induces spondyloarthropathy by acting on ROR-gammat+ CD3+CD4-CD8- entheseal resident T cells. Nat Med, 2012. 18(7): p. 1069-76. 128. Tseng, H.W., et al., Inflammation-driven bone formation in a mouse model of ankylosing spondylitis: sequential not parallel processes. Arthritis Res Ther, 2016. 18: p. 35. 129. Wellcome Trust Case Control, C., et al., Association scan of 14,500 nonsynonymous SNPs in four diseases identifies autoimmunity variants. Nat Genet, 2007. 39(11): p. 1329- 37. 130. Choi, C.B., et al., ARTS1 polymorphisms are associated with ankylosing spondylitis in Koreans. Ann Rheum Dis, 2010. 69(3): p. 582-4. 131. Davidson, S.I., et al., Association of ERAP1, but not IL23R, with ankylosing spondylitis in a Han Chinese population. Arthritis Rheum, 2009. 60(11): p. 3263-8. 132. Harvey, D., et al., Investigating the genetic association between ERAP1 and ankylosing spondylitis. Hum Mol Genet, 2009. 18(21): p. 4204-12. 133. Li, C., et al., ERAP1 is associated with ankylosing spondylitis in Han Chinese. J Rheumatol, 2011. 38(2): p. 317-21. 134. Maksymowych, W.P., et al., Association of a specific ERAP1/ARTS1 haplotype with disease susceptibility in ankylosing spondylitis. Arthritis Rheum, 2009. 60(5): p. 1317-23. 135. Pazar, B., et al., Association of ARTS1 gene polymorphisms with ankylosing spondylitis in the Hungarian population: the rs27044 variant is associated with HLA-B*2705 subtype in Hungarian patients with ankylosing spondylitis. J Rheumatol, 2010. 37(2): p. 379-84. 168 136. Pimentel-Santos, F.M., et al., Association of IL23R and ERAP1 genes with ankylosing spondylitis in a Portuguese population. Clin Exp Rheumatol, 2009. 27(5): p. 800-6. 137. Szczypiorska, M., et al., ERAP1 polymorphisms and haplotypes are associated with ankylosing spondylitis susceptibility and functional severity in a Spanish population. Rheumatology (Oxford), 2011. 50(11): p. 1969-75. 138. Kochan, G., et al., Crystal structures of the endoplasmic reticulum aminopeptidase-1 (ERAP1) reveal the molecular basis for N-terminal peptide trimming. Proc Natl Acad Sci U S A, 2011. 108(19): p. 7745-50. 139. Alvarez-Navarro, C. and J.A. Lopez de Castro, ERAP1 structure, function and pathogenetic role in ankylosing spondylitis and other MHC-associated diseases. Mol Immunol, 2014. 57(1): p. 12-21. 140. Reveille, J.D., Genetics of spondyloarthritis--beyond the MHC. Nat Rev Rheumatol, 2012. 8(5): p. 296-304. 141. Evans, D.M., et al., Interaction between ERAP1 and HLA-B27 in ankylosing spondylitis implicates peptide handling in the mechanism for HLA-B27 in disease susceptibility. Nat Genet, 2011. 43(8): p. 761-7. 142. Haroon, N., et al., Endoplasmic reticulum aminopeptidase 1 (ERAP1) exhibits functionally significant interaction with HLA-B27 and relates to subtype specificity in ankylosing spondylitis. Ann Rheum Dis, 2012. 71(4): p. 589-95. 143. Akram, A., et al., HLA-B27, but not HLA-B7, immunodominance to influenza is ERAP dependent. J Immunol, 2014. 192(12): p. 5520-8. 144. Hammer, G.E., et al., The aminopeptidase ERAAP shapes the peptide repertoire displayed by major histocompatibility complex class I molecules. Nat Immunol, 2006. 7(1): p. 103-12. 145. Hammer, G.E., et al., In the absence of aminopeptidase ERAAP, MHC class I molecules present many unstable and highly immunogenic peptides. Nat Immunol, 2007. 8(1): p. 101-8. 146. Kanaseki, T., et al., ERAAP synergizes with MHC class I molecules to make the final cut in the antigenic peptide precursors in the endoplasmic reticulum. Immunity, 2006. 25(5): p. 795-806. 147. Seregin, S.S., et al., Endoplasmic reticulum aminopeptidase-1 alleles associated with increased risk of ankylosing spondylitis reduce HLA-B27 mediated presentation of multiple antigens. Autoimmunity, 2013. 46(8): p. 497-508. 148. Stamogiannos, A., et al., Effects of polymorphic variation on the mechanism of Endoplasmic Reticulum Aminopeptidase 1. Mol Immunol, 2015. 67(2 Pt B): p. 426-35. 169 149. Evnouchidou, I., et al., Cutting Edge: Coding single nucleotide polymorphisms of endoplasmic reticulum aminopeptidase 1 can affect antigenic peptide generation in vitro by influencing basic enzymatic properties of the enzyme. J Immunol, 2011. 186(4): p. 1909-13. 150. Rastall, D.P., et al., ERAP1 functions override the intrinsic selection of specific antigens as immunodominant peptides, thereby altering the potency of antigen-specific cytolytic and effector memory T-cell responses. Int Immunol, 2014. 26(12): p. 685-95. 151. Rastall, D.P.W., et al., Mice expressing human ERAP1 variants associated with ankylosing spondylitis have altered T-cell repertoires and NK cell functions, as well as increased in utero and perinatal mortality. Int Immunol, 2017. 29(6): p. 277-289. 152. Reeves, E., et al., Functionally distinct ERAP1 allotype combinations distinguish individuals with Ankylosing Spondylitis. Proc Natl Acad Sci U S A, 2014. 111(49): p. 17594-9. 153. Goto, Y., et al., Secretion of endoplasmic reticulum aminopeptidase 1 is involved in the activation of macrophages induced by lipopolysaccharide and interferon-gamma. J Biol Chem, 2011. 286(24): p. 21906-14. 154. Aldhamen, Y.A., et al., Autoimmune disease-associated variants of extracellular endoplasmic reticulum aminopeptidase 1 induce altered innate immune responses by human immune cells. J Innate Immun, 2015. 7(3): p. 275-89. 155. Cui, X., et al., Identification of ARTS-1 as a novel TNFR1-binding protein that promotes TNFR1 ectodomain shedding. J Clin Invest, 2002. 110(4): p. 515-26. 156. Cui, X., et al., Shedding of the type II IL-1 decoy receptor requires a multifunctional aminopeptidase, aminopeptidase regulator of TNF receptor type 1 shedding. J Immunol, 2003. 171(12): p. 6814-9. 157. Cui, X., et al., An aminopeptidase, ARTS-1, is required for interleukin-6 receptor shedding. J Biol Chem, 2003. 278(31): p. 28677-85. 158. Haroon, N., et al., Serum cytokine receptors in ankylosing spondylitis: relationship to inflammatory markers and endoplasmic reticulum aminopeptidase polymorphisms. J Rheumatol, 2010. 37(9): p. 1907-10. 159. Yamamoto, N., et al., Identification of 33 polymorphisms in the adipocyte-derived leucine aminopeptidase (ALAP) gene and possible association with hypertension. Hum Mutat, 2002. 19(3): p. 251-7. 160. Hattori, A., et al., Molecular cloning of adipocyte-derived leucine aminopeptidase highly related to placental leucine aminopeptidase/oxytocinase. J Biochem, 1999. 125(5): p. 931-8. 170 161. Hattori, A., Y. Goto, and M. Tsujimoto, Exon 10 coding sequence is important for endoplasmic reticulum retention of endoplasmic reticulum aminopeptidase 1. Biol Pharm Bull, 2012. 35(4): p. 601-5. 162. Goto, Y., et al., TLR-mediated secretion of endoplasmic reticulum aminopeptidase 1 from macrophages. J Immunol, 2014. 192(9): p. 4443-52. 163. Aldhamen, Y.A., et al., Endoplasmic reticulum aminopeptidase-1 functions regulate key aspects of the innate immune response. PLoS One, 2013. 8(7): p. e69539. 164. Fung, E.Y., et al., Analysis of 17 autoimmune disease-associated variants in type 1 diabetes identifies 6q23/TNFAIP3 as a susceptibility locus. Genes Immun, 2009. 10(2): p. 188-91. 165. Genetic Analysis of Psoriasis, C., et al., A genome-wide association study identifies new psoriasis susceptibility loci and an interaction between HLA-C and ERAP1. Nat Genet, 2010. 42(11): p. 985-90. 166. Sun, L.D., et al., Association analyses identify six new psoriasis susceptibility loci in the Chinese population. Nat Genet, 2010. 42(11): p. 1005-9. 167. Guerini, F.R., et al., A functional variant in ERAP1 predisposes to multiple sclerosis. PLoS One, 2012. 7(1): p. e29931. 168. Di Meglio, P., et al., The IL23R R381Q gene variant protects against immune-mediated diseases by impairing IL-23-induced Th17 effector response in humans. PLoS One, 2011. 6(2): p. e17160. 169. Sarin, R., X. Wu, and C. Abraham, Inflammatory disease protective R381Q IL23 receptor polymorphism results in decreased primary CD4+ and CD8+ human T-cell functional responses. Proc Natl Acad Sci U S A, 2011. 108(23): p. 9560-5. 170. Milanez, F.M., et al., IL-23/Th17 axis is not influenced by TNF-blocking agents in ankylosing spondylitis patients. Arthritis Res Ther, 2016. 18: p. 52. 171. Zeng, L., M.J. Lindstrom, and J.A. Smith, Ankylosing spondylitis macrophage production of higher levels of interleukin-23 in response to lipopolysaccharide without induction of a significant unfolded protein response. Arthritis Rheum, 2011. 63(12): p. 3807-17. 172. Stritesky, G.L., N. Yeh, and M.H. Kaplan, IL-23 promotes maintenance but not commitment to the Th17 lineage. J Immunol, 2008. 181(9): p. 5948-55. 173. Sieper, J. and D. Poddubnyy, Axial spondyloarthritis. Lancet, 2017. 390(10089): p. 73- 84. 174. Ciccia, F., et al., Overexpression of interleukin-23, but not interleukin-17, as an immunologic signature of subclinical intestinal inflammation in ankylosing spondylitis. Arthritis Rheum, 2009. 60(4): p. 955-65. 171 175. Braun, J., J. Sieper, and M. Bollow, Imaging of sacroiliitis. Clin Rheumatol, 2000. 19(1): p. 51-7. 176. Ghasemi-Rad, M., et al., Ankylosing spondylitis: A state of the art factual backbone. World J Radiol, 2015. 7(9): p. 236-52. 177. Lories, R.J., F.P. Luyten, and K. de Vlam, Progress in spondylarthritis. Mechanisms of new bone formation in spondyloarthritis. Arthritis Res Ther, 2009. 11(2): p. 221. 178. Carter, S. and R.J. Lories, Osteoporosis: a paradox in ankylosing spondylitis. Curr Osteoporos Rep, 2011. 9(3): p. 112-5. 179. Schett, G., et al., Mechanisms of Disease: the link between RANKL and arthritic bone disease. Nat Clin Pract Rheumatol, 2005. 1(1): p. 47-54. 180. Teitelbaum, S.L., Bone resorption by osteoclasts. Science, 2000. 289(5484): p. 1504-8. 181. Boyce, B.F., Z. Yao, and L. Xing, Osteoclasts have multiple roles in bone in addition to bone resorption. Crit Rev Eukaryot Gene Expr, 2009. 19(3): p. 171-80. 182. Teitelbaum, S.L. and F.P. Ross, Genetic regulation of osteoclast development and function. Nat Rev Genet, 2003. 4(8): p. 638-49. 183. Park, J.H., N.K. Lee, and S.Y. Lee, Current Understanding of RANK Signaling in Osteoclast Differentiation and Maturation. Mol Cells, 2017. 40(10): p. 706-713. 184. Takayanagi, H., Osteoimmunology: shared mechanisms and crosstalk between the immune and bone systems. Nat Rev Immunol, 2007. 7(4): p. 292-304. 185. Boyce, B.F. and L. Xing, Functions of RANKL/RANK/OPG in bone modeling and remodeling. Arch Biochem Biophys, 2008. 473(2): p. 139-46. 186. Kamiya, N. and Y. Mishina, New insights on the roles of BMP signaling in bone-A review of recent mouse genetic studies. Biofactors, 2011. 37(2): p. 75-82. 187. Komori, T., et al., Targeted disruption of Cbfa1 results in a complete lack of bone formation owing to maturational arrest of osteoblasts. Cell, 1997. 89(5): p. 755-64. 188. Stein, G.S., et al., Runx2 control of organization, assembly and activity of the regulatory machinery for skeletal gene expression. Oncogene, 2004. 23(24): p. 4315-29. 189. Nakashima, K., et al., The novel zinc finger-containing transcription factor osterix is required for osteoblast differentiation and bone formation. Cell, 2002. 108(1): p. 17-29. 190. Canalis, E., Notch signaling in osteoblasts. Sci Signal, 2008. 1(17): p. pe17. 191. Takada, I., A.P. Kouzmenko, and S. Kato, Wnt and PPARgamma signaling in osteoblastogenesis and adipogenesis. Nat Rev Rheumatol, 2009. 5(8): p. 442-7. 172 192. Walsh, M.C., et al., Osteoimmunology: interplay between the immune system and bone metabolism. Annu Rev Immunol, 2006. 24: p. 33-63. 193. Gravallese, E.M., et al., Identification of cell types responsible for bone resorption in rheumatoid arthritis and juvenile rheumatoid arthritis. Am J Pathol, 1998. 152(4): p. 943-51. 194. Gilbert, L., et al., Inhibition of osteoblast differentiation by tumor necrosis factor-alpha. Endocrinology, 2000. 141(11): p. 3956-64. 195. Lu, X., et al., Transcriptional regulation of the osterix (Osx, Sp7) promoter by tumor necrosis factor identifies disparate effects of mitogen-activated protein kinase and NF kappa B pathways. J Biol Chem, 2006. 281(10): p. 6297-306. 196. Lee, H.L., et al., Tumor necrosis factor-alpha enhances the transcription of Smad ubiquitination regulatory factor 1 in an activating protein-1- and Runx2-dependent manner. J Cell Physiol, 2013. 228(5): p. 1076-86. 197. Zhao, L., et al., Tumor necrosis factor inhibits mesenchymal stem cell differentiation into osteoblasts via the ubiquitin E3 ligase Wwp1. Stem Cells, 2011. 29(10): p. 1601-10. 198. Gilbert, L., et al., Expression of the osteoblast differentiation factor RUNX2 (Cbfa1/AML3/Pebp2alpha A) is inhibited by tumor necrosis factor-alpha. J Biol Chem, 2002. 277(4): p. 2695-701. 199. Osta, B., G. Benedetti, and P. Miossec, Classical and Paradoxical Effects of TNF-alpha on Bone Homeostasis. Front Immunol, 2014. 5: p. 48. 200. Diarra, D., et al., Dickkopf-1 is a master regulator of joint remodeling. Nat Med, 2007. 13(2): p. 156-63. 201. Brandstrom, H., et al., Tumor necrosis factor-alpha and -beta upregulate the levels of osteoprotegerin mRNA in human osteosarcoma MG-63 cells. Biochem Biophys Res Commun, 1998. 248(3): p. 454-7. 202. Hofbauer, L.C., et al., Interleukin-1beta and tumor necrosis factor-alpha, but not interleukin-6, stimulate osteoprotegerin ligand gene expression in human osteoblastic cells. Bone, 1999. 25(3): p. 255-9. 203. Lam, J., et al., TNF-alpha induces osteoclastogenesis by direct stimulation of macrophages exposed to permissive levels of RANK ligand. J Clin Invest, 2000. 106(12): p. 1481-8. 204. Chen, G. and D.V. Goeddel, TNF-R1 signaling: a beautiful pathway. Science, 2002. 296(5573): p. 1634-5. 173 205. Marotte, H., et al., A 1-year case-control study in patients with rheumatoid arthritis indicates prevention of loss of bone mineral density in both responders and nonresponders to infliximab. Arthritis Res Ther, 2007. 9(3): p. R61. 206. Nguyen, L., et al., Interleukin-1 beta stimulates bone resorption and inhibits bone formation in vivo. Lymphokine Cytokine Res, 1991. 10(1-2): p. 15-21. 207. Lee, Y.M., et al., IL-1 plays an important role in the bone metabolism under physiological conditions. Int Immunol, 2010. 22(10): p. 805-16. 208. Ley, R.E., D.A. Peterson, and J.I. Gordon, Ecological and evolutionary forces shaping microbial diversity in the human intestine. Cell, 2006. 124(4): p. 837-48. 209. Spiljar, M., D. Merkler, and M. Trajkovski, The Immune System Bridges the Gut Microbiota with Systemic Energy Homeostasis: Focus on TLRs, Mucosal Barrier, and SCFAs. Front Immunol, 2017. 8: p. 1353. 210. Wu, H.J. and E. Wu, The role of gut microbiota in immune homeostasis and autoimmunity. Gut Microbes, 2012. 3(1): p. 4-14. 211. Van Praet, L., et al., Microscopic gut inflammation in axial spondyloarthritis: a multiparametric predictive model. Ann Rheum Dis, 2013. 72(3): p. 414-7. 212. Jethwa, H. and S. Abraham, The evidence for microbiome manipulation in inflammatory arthritis. Rheumatology (Oxford), 2017. 56(9): p. 1452-1460. 213. Costello, M.E., et al., Brief Report: Intestinal Dysbiosis in Ankylosing Spondylitis. Arthritis Rheumatol, 2015. 67(3): p. 686-691. 214. Gill, T., et al., The intestinal microbiome in spondyloarthritis. Curr Opin Rheumatol, 2015. 27(4): p. 319-25. 215. Wen, C., et al., Quantitative metagenomics reveals unique gut microbiome biomarkers in ankylosing spondylitis. Genome Biol, 2017. 18(1): p. 142. 216. Maksymowych, W.P., Update on the treatment of ankylosing spondylitis. Ther Clin Risk Manag, 2007. 3(6): p. 1125-33. 217. Davey-Ranasinghe, N. and A. Deodhar, Osteoporosis and vertebral fractures in ankylosing spondylitis. Curr Opin Rheumatol, 2013. 25(4): p. 509-16. 218. Anderson, J.J., et al., Ankylosing spondylitis assessment group preliminary definition of short-term improvement in ankylosing spondylitis. Arthritis Rheum, 2001. 44(8): p. 1876- 86. 219. Brandt, J., et al., Development and preselection of criteria for short term improvement after anti-TNF alpha treatment in ankylosing spondylitis. Ann Rheum Dis, 2004. 63(11): p. 1438-44. 174 220. Liu, W., et al., Efficacy and safety of TNF-alpha inhibitors for active ankylosing spondylitis patients: Multiple treatment comparisons in a network meta-analysis. Sci Rep, 2016. 6: p. 32768. 221. Maas, F., et al., Radiographic vertebral fractures develop in patients with ankylosing spondylitis during 4 years of TNF-alpha blocking therapy. Clin Exp Rheumatol, 2016. 34(2): p. 191-9. 222. Haroon, N.N., et al., Effect of TNF-alpha inhibitor treatment on bone mineral density in patients with ankylosing spondylitis: a systematic review and meta-analysis. Semin Arthritis Rheum, 2014. 44(2): p. 155-61. 223. Arends, S., et al., Clinical studies on bone-related outcome and the effect of TNF-alpha blocking therapy in ankylosing spondylitis. Curr Opin Rheumatol, 2014. 26(3): p. 259-68. 224. Siu, S., et al., Meta-analysis of tumor necrosis factor inhibitors and glucocorticoids on bone density in rheumatoid arthritis and ankylosing spondylitis trials. Arthritis Care Res (Hoboken), 2015. 67(6): p. 754-64. 225. Maas, F., et al., Reduction in spinal radiographic progression in ankylosing spondylitis patients receiving prolonged treatment with TNF-alpha inhibitors. Arthritis Care Res (Hoboken), 2016. 226. Zhang, J.R., et al., Effects of tumor necrosis factor-alpha inhibitors on new bone formation in ankylosing spondylitis. Joint Bone Spine, 2016. 83(3): p. 257-64. 227. Poddubnyy, D., et al., Ustekinumab for the treatment of patients with active ankylosing spondylitis: results of a 28-week, prospective, open-label, proof-of-concept study (TOPAS). Ann Rheum Dis, 2014. 73(5): p. 817-23. 228. Babaie, F., et al., The role of gut microbiota and IL-23/IL-17 pathway in ankylosing spondylitis immunopathogenesis: New insights and updates. Immunol Lett, 2018. 196: p. 52-62. 229. Baeten, D., et al., Secukinumab, an Interleukin-17A Inhibitor, in Ankylosing Spondylitis. N Engl J Med, 2015. 373(26): p. 2534-48. 230. Haibel, H., et al., Open label trial of anakinra in active ankylosing spondylitis over 24 weeks. Ann Rheum Dis, 2005. 64(2): p. 296-8. 231. Sieper, J., et al., Assessment of short-term symptomatic efficacy of tocilizumab in ankylosing spondylitis: results of randomised, placebo-controlled trials. Ann Rheum Dis, 2014. 73(1): p. 95-100. 232. Breban, M., Animal models and in vitro models for the study of aetiopathogenesis of spondyloarthropathies. Baillieres Clin Rheumatol, 1998. 12(4): p. 611-26. 175 233. Adarichev, V.A. and T.T. Glant, Experimental spondyloarthropathies: animal models of ankylosing spondylitis. Curr Rheumatol Rep, 2006. 8(4): p. 267-74. 234. Braem, K. and R.J. Lories, Insights into the pathophysiology of ankylosing spondylitis: contributions from animal models. Joint Bone Spine, 2012. 79(3): p. 243-8. 235. Zhang, Y., Animal models of inflammatory spinal and sacroiliac joint diseases. Rheum Dis Clin North Am, 2003. 29(3): p. 631-45. 236. Taurog, J.D., et al., Susceptibility to inflammatory disease in HLA-B27 transgenic rat lines correlates with the level of B27 expression. J Immunol, 1993. 150(9): p. 4168-78. 237. Hammer, R.E., et al., Spontaneous inflammatory disease in transgenic rats expressing HLA-B27 and human beta 2m: an animal model of HLA-B27-associated human disorders. Cell, 1990. 63(5): p. 1099-112. 238. Lin, P., et al., HLA-B27 and human beta2-microglobulin affect the gut microbiota of transgenic rats. PLoS One, 2014. 9(8): p. e105684. 239. Rauner, M., et al., Loss of bone strength in HLA-B27 transgenic rats is characterized by a high bone turnover and is mainly osteoclast-driven. Bone, 2015. 75: p. 183-91. 240. Khare, S.D., et al., HLA-B27 heavy chains contribute to spontaneous inflammatory disease in B27/human beta2-microglobulin (beta2m) double transgenic mice with disrupted mouse beta2m. J Clin Invest, 1996. 98(12): p. 2746-55. 241. Khare, S.D., et al., Spontaneous inflammatory disease in HLA-B27 transgenic mice does not require transporter of antigenic peptides. Clin Immunol, 2001. 98(3): p. 364-9. 242. Sweet, H.O. and M.C. Green, Progressive ankylosis, a new skeletal mutation in the mouse. J Hered, 1981. 72(2): p. 87-93. 243. Mahowald, M.L., H. Krug, and J. Taurog, Progressive ankylosis in mice. An animal model of spondylarthropathy. I. Clinical and radiographic findings. Arthritis Rheum, 1988. 31(11): p. 1390-9. 244. Sakaguchi, S., et al., Spontaneous development of autoimmune arthritis due to genetic anomaly of T cell signal transduction: Part 1. Semin Immunol, 2006. 18(4): p. 199-206. 245. Ruutu, M., et al., beta-glucan triggers spondylarthritis and Crohn's disease-like ileitis in SKG mice. Arthritis Rheum, 2012. 64(7): p. 2211-22. 246. Edwards, C.K., 3rd, et al., Soluble human p55 and p75 tumor necrosis factor receptors reverse spontaneous arthritis in transgenic mice expressing transmembrane tumor necrosis factor alpha. Arthritis Rheum, 2006. 54(9): p. 2872-85. 247. Rabinovich, G.A., D. Gabrilovich, and E.M. Sotomayor, Immunosuppressive strategies that are mediated by tumor cells. Annu Rev Immunol, 2007. 25: p. 267-96. 176 248. Leach, D.R., M.F. Krummel, and J.P. Allison, Enhancement of antitumor immunity by CTLA-4 blockade. Science, 1996. 271(5256): p. 1734-6. 249. Walunas, T.L., C.Y. Bakker, and J.A. Bluestone, CTLA-4 ligation blocks CD28- dependent T cell activation. J Exp Med, 1996. 183(6): p. 2541-50. 250. Hurwitz, A.A., et al., CTLA-4 blockade synergizes with tumor-derived granulocyte- macrophage colony-stimulating factor for treatment of an experimental mammary carcinoma. Proc Natl Acad Sci U S A, 1998. 95(17): p. 10067-71. 251. van Elsas, A., A.A. Hurwitz, and J.P. Allison, Combination immunotherapy of B16 melanoma using anti-cytotoxic T lymphocyte-associated antigen 4 (CTLA-4) and granulocyte/macrophage colony-stimulating factor (GM-CSF)-producing vaccines induces rejection of subcutaneous and metastatic tumors accompanied by autoimmune depigmentation. J Exp Med, 1999. 190(3): p. 355-66. 252. Sharma, P. and J.P. Allison, Immune checkpoint targeting in cancer therapy: toward combination strategies with curative potential. Cell, 2015. 161(2): p. 205-14. 253. Sharma, P., et al., Novel cancer immunotherapy agents with survival benefit: recent successes and next steps. Nat Rev Cancer, 2011. 11(11): p. 805-12. 254. Dong, H., et al., Tumor-associated B7-H1 promotes T-cell apoptosis: a potential mechanism of immune evasion. Nat Med, 2002. 8(8): p. 793-800. 255. Freeman, G.J., et al., Engagement of the PD-1 immunoinhibitory receptor by a novel B7 family member leads to negative regulation of lymphocyte activation. J Exp Med, 2000. 192(7): p. 1027-34. 256. Brahmer, J.R., et al., Safety and activity of anti-PD-L1 antibody in patients with advanced cancer. N Engl J Med, 2012. 366(26): p. 2455-65. 257. Powles, T., et al., MPDL3280A (anti-PD-L1) treatment leads to clinical activity in metastatic bladder cancer. Nature, 2014. 515(7528): p. 558-62. 258. Wolchok, J.D., et al., Nivolumab plus ipilimumab in advanced melanoma. N Engl J Med, 2013. 369(2): p. 122-33. 259. Larkin, J., et al., Combined Nivolumab and Ipilimumab or Monotherapy in Untreated Melanoma. N Engl J Med, 2015. 373(1): p. 23-34. 260. Moynihan, K.D., et al., Eradication of large established tumors in mice by combination immunotherapy that engages innate and adaptive immune responses. Nat Med, 2016. 22(12): p. 1402-1410. 261. Cannons, J.L., S.G. Tangye, and P.L. Schwartzberg, SLAM family receptors and SAP adaptors in immunity. Annu Rev Immunol, 2011. 29: p. 665-705. 177 262. Wu, N. and A. Veillette, SLAM family receptors in normal immunity and immune pathologies. Curr Opin Immunol, 2016. 38: p. 45-51. 263. Sidorenko, S.P. and E.A. Clark, The dual-function CD150 receptor subfamily: the viral attraction. Nat Immunol, 2003. 4(1): p. 19-24. 264. Veillette, A., et al., Importance and mechanism of 'switch' function of SAP family adapters. Immunol Rev, 2009. 232(1): p. 229-39. 265. Bouchon, A., et al., Activation of NK cell-mediated cytotoxicity by a SAP-independent receptor of the CD2 family. J Immunol, 2001. 167(10): p. 5517-21. 266. Calpe, S., et al., The SLAM and SAP gene families control innate and adaptive immune responses. Adv Immunol, 2008. 97: p. 177-250. 267. Cruz-Munoz, M.E., et al., Influence of CRACC, a SLAM family receptor coupled to the adaptor EAT-2, on natural killer cell function. Nat Immunol, 2009. 10(3): p. 297-305. 268. Tassi, I. and M. Colonna, The cytotoxicity receptor CRACC (CS-1) recruits EAT-2 and activates the PI3K and phospholipase Cgamma signaling pathways in human NK cells. J Immunol, 2005. 175(12): p. 7996-8002. 269. Guo, H., et al., Immune cell inhibition by SLAMF7 is mediated by a mechanism requiring src kinases, CD45, and SHIP-1 that is defective in multiple myeloma cells. Mol Cell Biol, 2015. 35(1): p. 41-51. 270. Lonial, S., et al., Elotuzumab Therapy for Relapsed or Refractory Multiple Myeloma. N Engl J Med, 2015. 373(7): p. 621-31. 271. Richardson, P.G., et al., A phase 2 trial of lenalidomide, bortezomib, and dexamethasone in patients with relapsed and relapsed/refractory myeloma. Blood, 2014. 123(10): p. 1461-9. 272. Collins, S.M., et al., Elotuzumab directly enhances NK cell cytotoxicity against myeloma via CS1 ligation: evidence for augmented NK cell function complementing ADCC. Cancer Immunol Immunother, 2013. 62(12): p. 1841-9. 273. Wold, W.S. and K. Toth, Adenovirus vectors for gene therapy, vaccination and cancer gene therapy. Curr Gene Ther, 2013. 13(6): p. 421-33. 274. Russell, W.C., Adenoviruses: update on structure and function. J Gen Virol, 2009. 90(Pt 1): p. 1-20. 275. Appaiahgari, M.B. and S. Vrati, Adenoviruses as gene/vaccine delivery vectors: promises and pitfalls. Expert Opin Biol Ther, 2015. 15(3): p. 337-51. 276. Amalfitano, A., et al., Production and characterization of improved adenovirus vectors with the E1, E2b, and E3 genes deleted. J Virol, 1998. 72(2): p. 926-33. 178 277. Raper, S.E., et al., Selective gene transfer into the liver of non-human primates with E1- deleted, E2A-defective, or E1-E4 deleted recombinant adenoviruses. Hum Gene Ther, 1998. 9(5): p. 671-9. 278. Khare, R., et al., Advances and future challenges in adenoviral vector pharmacology and targeting. Curr Gene Ther, 2011. 11(4): p. 241-58. 279. Doronin, K., et al., Coagulation factor X activates innate immunity to human species C adenovirus. Science, 2012. 338(6108): p. 795-8. 280. Appledorn, D.M., et al., Adenovirus vector-induced innate inflammatory mediators, MAPK signaling, as well as adaptive immune responses are dependent upon both TLR2 and TLR9 in vivo. J Immunol, 2008. 181(3): p. 2134-44. 281. Muruve, D.A., The innate immune response to adenovirus vectors. Hum Gene Ther, 2004. 15(12): p. 1157-66. 282. Appledorn, D.M., et al., Complex interactions with several arms of the complement system dictate innate and humoral immunity to adenoviral vectors. Gene Ther, 2008. 15(24): p. 1606-17. 283. Barlan, A.U., et al., Adenovirus membrane penetration activates the NLRP3 inflammasome. J Virol, 2011. 85(1): p. 146-55. 284. Barouch, D.H., Novel adenovirus vector-based vaccines for HIV-1. Curr Opin HIV AIDS, 2010. 5(5): p. 386-90. 285. Ciccia, F., A. Rizzo, and G. Triolo, Subclinical gut inflammation in ankylosing spondylitis. Curr Opin Rheumatol, 2016. 28(1): p. 89-96. 286. Thjodleifsson, B., et al., A common genetic background for inflammatory bowel disease and ankylosing spondylitis: a genealogic study in Iceland. Arthritis Rheum, 2007. 56(8): p. 2633-9. 287. Yadav, S.K., et al., Gut dysbiosis breaks immunological tolerance toward the central nervous system during young adulthood. Proc Natl Acad Sci U S A, 2017. 114(44): p. E9318-E9327. 288. Atarashi, K., et al., Induction of colonic regulatory T cells by indigenous Clostridium species. Science, 2011. 331(6015): p. 337-41. 289. Araki, A., et al., MyD88-deficient mice develop severe intestinal inflammation in dextran sodium sulfate colitis. J Gastroenterol, 2005. 40(1): p. 16-23. 290. Zaki, M.H., et al., IL-18 production downstream of the Nlrp3 inflammasome confers protection against colorectal tumor formation. J Immunol, 2010. 185(8): p. 4912-20. 179 291. Erben, U., et al., A guide to histomorphological evaluation of intestinal inflammation in mouse models. Int J Clin Exp Pathol, 2014. 7(8): p. 4557-76. 292. Ni, J., et al., Gut microbiota and IBD: causation or correlation? Nat Rev Gastroenterol Hepatol, 2017. 14(10): p. 573-584. 293. Daft, J.G., et al., Cross-fostering immediately after birth induces a permanent microbiota shift that is shaped by the nursing mother. Microbiome, 2015. 3: p. 17. 294. Glatigny, S., et al., Proinflammatory Th17 cells are expanded and induced by dendritic cells in spondylarthritis-prone HLA-B27-transgenic rats. Arthritis Rheum, 2012. 64(1): p. 110-20. 295. Astier, A.L. and D.A. Hafler, Abnormal Tr1 differentiation in multiple sclerosis. J Neuroimmunol, 2007. 191(1-2): p. 70-8. 296. Fierabracci, A., et al., The putative role of endoplasmic reticulum aminopeptidases in autoimmunity: insights from genomic-wide association studies. Autoimmun Rev, 2012. 12(2): p. 281-8. 297. Chihara, N., et al., Differentiation and Characterization of Tr1 Cells. Curr Protoc Immunol, 2016. 113: p. 3 27 1-3 27 10. 298. York, I.A., et al., Endoplasmic reticulum aminopeptidase 1 (ERAP1) trims MHC class I- presented peptides in vivo and plays an important role in immunodominance. Proc Natl Acad Sci U S A, 2006. 103(24): p. 9202-7. 299. Shido, F., et al., Endoplasmic reticulum aminopeptidase-1 mediates leukemia inhibitory factor-induced cell surface human leukocyte antigen-G expression in JEG-3 choriocarcinoma cells. Endocrinology, 2006. 147(4): p. 1780-8. 300. Bron, J.L., et al., Discovertebral (Andersson) lesions of the spine in ankylosing spondylitis revisited. Clin Rheumatol, 2009. 28(8): p. 883-92. 301. Ciccia, F., et al., Expansion of intestinal CD4+CD25(high) Treg cells in patients with ankylosing spondylitis: a putative role for interleukin-10 in preventing intestinal Th17 response. Arthritis Rheum, 2010. 62(12): p. 3625-34. 302. Guo, H., et al., Functional defects in CD4(+) CD25(high) FoxP3(+) regulatory cells in ankylosing spondylitis. Sci Rep, 2016. 6: p. 37559. 303. Molinos, M., et al., Inflammation in intervertebral disc degeneration and regeneration. J R Soc Interface, 2015. 12(104): p. 20141191. 304. Yan, J. and J.F. Charles, Gut Microbiome and Bone: to Build, Destroy, or Both? Curr Osteoporos Rep, 2017. 15(4): p. 376-384. 180 305. Stebbings, S., et al., Comparison of the faecal microflora of patients with ankylosing spondylitis and controls using molecular methods of analysis. Rheumatology (Oxford), 2002. 41(12): p. 1395-401. 306. Round, J.L. and S.K. Mazmanian, Inducible Foxp3+ regulatory T-cell development by a commensal bacterium of the intestinal microbiota. Proc Natl Acad Sci U S A, 2010. 107(27): p. 12204-9. 307. Goto, Y., et al., Glutamine-181 is crucial in the enzymatic activity and substrate specificity of human endoplasmic-reticulum aminopeptidase-1. Biochem J, 2008. 416(1): p. 109-16. 308. Huang, J., et al., Novel non-HLA-susceptible regions determined by meta-analysis of four genomewide scans for ankylosing spondylitis. J Genet, 2008. 87(1): p. 75-81. 309. Pepelyayeva, Y., et al., ERAP1 deficient mice have reduced Type 1 regulatory T cells and develop skeletal and intestinal features of Ankylosing Spondylitis. Sci Rep, 2018. 8(1): p. 12464. 310. Hreggvidsdottir, H.S., T. Noordenbos, and D.L. Baeten, Inflammatory pathways in spondyloarthritis. Mol Immunol, 2014. 57(1): p. 28-37. 311. Hofbauer, L.C., et al., The roles of osteoprotegerin and osteoprotegerin ligand in the paracrine regulation of bone resorption. J Bone Miner Res, 2000. 15(1): p. 2-12. 312. Burgess, T.L., et al., The ligand for osteoprotegerin (OPGL) directly activates mature osteoclasts. J Cell Biol, 1999. 145(3): p. 527-38. 313. Fuller, K., et al., TNFalpha potently activates osteoclasts, through a direct action independent of and strongly synergistic with RANKL. Endocrinology, 2002. 143(3): p. 1108-18. 314. Lacey, D.L., et al., Osteoprotegerin ligand is a cytokine that regulates osteoclast differentiation and activation. Cell, 1998. 93(2): p. 165-76. 315. Dai, S.M., et al., Interleukin-18 enhances monocyte tumor necrosis factor alpha and interleukin-1beta production induced by direct contact with T lymphocytes: implications in rheumatoid arthritis. Arthritis Rheum, 2004. 50(2): p. 432-43. 316. Komori, T., Regulation of osteoblast differentiation by Runx2. Adv Exp Med Biol, 2010. 658: p. 43-9. 317. Schulz, M., H. Dotzlaw, and G. Neeck, Ankylosing spondylitis and rheumatoid arthritis: serum levels of TNF-alpha and Its soluble receptors during the course of therapy with etanercept and infliximab. Biomed Res Int, 2014. 2014: p. 675108. 318. Adamopoulos, I.E., et al., IL-23 is critical for induction of arthritis, osteoclast formation, and maintenance of bone mass. J Immunol, 2011. 187(2): p. 951-9. 181 319. Perpetuo, I.P., et al., Effect of Tumor Necrosis Factor Inhibitor Therapy on Osteoclasts Precursors in Ankylosing Spondylitis. PLoS One, 2015. 10(12): p. e0144655. 320. Beniash, E., et al., Possible role of DMP1 in dentin mineralization. J Struct Biol, 2011. 174(1): p. 100-6. 321. Hayman, A.R., Tartrate-resistant acid phosphatase (TRAP) and the osteoclast/immune cell dichotomy. Autoimmunity, 2008. 41(3): p. 218-23. 322. Tang, W., et al., Osteoblast-specific transcription factor Osterix (Osx) is an upstream regulator of Satb2 during bone formation. J Biol Chem, 2011. 286(38): p. 32995-3002. 323. Yagi, M., et al., DC-STAMP is essential for cell-cell fusion in osteoclasts and foreign body giant cells. J Exp Med, 2005. 202(3): p. 345-51. 324. Wilson, S.R., et al., Cathepsin K activity-dependent regulation of osteoclast actin ring formation and bone resorption. J Biol Chem, 2009. 284(4): p. 2584-92. 325. Copp, D.H. and B. Cheney, Calcitonin-a hormone from the parathyroid which lowers the calcium-level of the blood. Nature, 1962. 193: p. 381-2. 326. Chambers, T.J. and A. Moore, The sensitivity of isolated osteoclasts to morphological transformation by calcitonin. J Clin Endocrinol Metab, 1983. 57(4): p. 819-24. 327. Iwaniec, U.T. and R.T. Turner, Influence of body weight on bone mass, architecture and turnover. J Endocrinol, 2016. 230(3): p. R115-30. 328. Appel, H., C. Loddenkemper, and P. Miossec, Rheumatoid arthritis and ankylosing spondylitis - pathology of acute inflammation. Clin Exp Rheumatol, 2009. 27(4 Suppl 55): p. S15-9. 329. Taylan, A., et al., Biomarkers and cytokines of bone turnover: extensive evaluation in a cohort of patients with ankylosing spondylitis. BMC Musculoskelet Disord, 2012. 13: p. 191. 330. Jo, S., et al., A novel role for bone-derived cells in ankylosing spondylitis: Focus on IL- 23. Biochem Biophys Res Commun, 2017. 491(3): p. 787-793. 331. Shukla, P., M.N. Mansoori, and D. Singh, Efficacy of anti-IL-23 monotherapy versus combination therapy with anti-IL-17 in estrogen deficiency induced bone loss conditions. Bone, 2018. 110: p. 84-95. 332. Munn, D.H. and V. Bronte, Immune suppressive mechanisms in the tumor microenvironment. Curr Opin Immunol, 2016. 39: p. 1-6. 333. Mogensen, T.H., et al., Innate immune recognition and activation during HIV infection. Retrovirology, 2010. 7: p. 54. 182 334. Adler, A.J., Mechanisms of T cell tolerance and suppression in cancer mediated by tumor-associated antigens and hormones. Curr Cancer Drug Targets, 2007. 7(1): p. 3-14. 335. Gabrilovich, D.I., S. Ostrand-Rosenberg, and V. Bronte, Coordinated regulation of myeloid cells by tumours. Nat Rev Immunol, 2012. 12(4): p. 253-68. 336. Juneja, V.R., et al., PD-L1 on tumor cells is sufficient for immune evasion in immunogenic tumors and inhibits CD8 T cell cytotoxicity. J Exp Med, 2017. 214(4): p. 895-904. 337. Hodi, F.S., et al., Improved survival with ipilimumab in patients with metastatic melanoma. N Engl J Med, 2010. 363(8): p. 711-23. 338. Topalian, S.L., et al., Safety, activity, and immune correlates of anti-PD-1 antibody in cancer. N Engl J Med, 2012. 366(26): p. 2443-54. 339. Anderson, A.C., N. Joller, and V.K. Kuchroo, Lag-3, Tim-3, and TIGIT: Co-inhibitory Receptors with Specialized Functions in Immune Regulation. Immunity, 2016. 44(5): p. 989-1004. 340. Weinberg, A.D., et al., Science gone translational: the OX40 agonist story. Immunol Rev, 2011. 244(1): p. 218-31. 341. Ascierto, P.A., et al., Clinical experiences with anti-CD137 and anti-PD1 therapeutic antibodies. Semin Oncol, 2010. 37(5): p. 508-16. 342. Roberts, D.J., et al., Control of established melanoma by CD27 stimulation is associated with enhanced effector function and persistence, and reduced PD-1 expression of tumor infiltrating CD8(+) T cells. J Immunother, 2010. 33(8): p. 769-79. 343. Aldhamen, Y.A., et al., CRACC-targeting Fc-fusion protein induces activation of NK cells and DCs and improves T cell immune responses to antigenic targets. Vaccine, 2016. 34(27): p. 3109-3118. 344. Sharon, D. and A. Kamen, Advancements in the design and scalable production of viral gene transfer vectors. Biotechnol Bioeng, 2018. 115(1): p. 25-40. 345. Muller, L., P. Aigner, and D. Stoiber, Type I Interferons and Natural Killer Cell Regulation in Cancer. Front Immunol, 2017. 8: p. 304. 346. Zhu, J., X. Huang, and Y. Yang, A critical role for type I IFN-dependent NK cell activation in innate immune elimination of adenoviral vectors in vivo. Mol Ther, 2008. 16(7): p. 1300-7. 347. Cheon, H., E.C. Borden, and G.R. Stark, Interferons and their stimulated genes in the tumor microenvironment. Semin Oncol, 2014. 41(2): p. 156-73. 183 348. Le Bon, A., et al., Cutting edge: enhancement of antibody responses through direct stimulation of B and T cells by type I IFN. J Immunol, 2006. 176(4): p. 2074-8. 349. Santini, S.M., et al., Type I interferon as a powerful adjuvant for monocyte-derived dendritic cell development and activity in vitro and in Hu-PBL-SCID mice. J Exp Med, 2000. 191(10): p. 1777-88. 350. Coffman, R.L., A. Sher, and R.A. Seder, Vaccine adjuvants: putting innate immunity to work. Immunity, 2010. 33(4): p. 492-503. 351. Pardoll, D.M., The blockade of immune checkpoints in cancer immunotherapy. Nat Rev Cancer, 2012. 12(4): p. 252-64. 352. Kobayashi, M., et al., Identification and purification of natural killer cell stimulatory factor (NKSF), a cytokine with multiple biologic effects on human lymphocytes. J Exp Med, 1989. 170(3): p. 827-45. 353. Gooden, M., et al., HLA-E expression by gynecological cancers restrains tumor- infiltrating CD8(+) T lymphocytes. Proc Natl Acad Sci U S A, 2011. 108(26): p. 10656- 61. 354. Haanen, J., Converting Cold into Hot Tumors by Combining Immunotherapies. Cell, 2017. 170(6): p. 1055-1056. 355. Veillette, A., SLAM-family receptors: immune regulators with or without SAP-family adaptors. Cold Spring Harb Perspect Biol, 2010. 2(3): p. a002469. 356. Takaoka, A., et al., Integration of interferon-alpha/beta signalling to p53 responses in tumour suppression and antiviral defence. Nature, 2003. 424(6948): p. 516-23. 357. Yang, Y., et al., Exploiting synthetic lethality for the therapy of ABC diffuse large B cell lymphoma. Cancer Cell, 2012. 21(6): p. 723-37. 358. Zitvogel, L., et al., Type I interferons in anticancer immunity. Nat Rev Immunol, 2015. 15(7): p. 405-14. 359. Sokolowska, O. and D. Nowis, STING Signaling in Cancer Cells: Important or Not? Arch Immunol Ther Exp (Warsz), 2018. 66(2): p. 125-132. 360. Martin-Fontecha, A., et al., Induced recruitment of NK cells to lymph nodes provides IFN-gamma for T(H)1 priming. Nat Immunol, 2004. 5(12): p. 1260-5. 361. Reid-Yu, S.A., C.L. Small, and B.K. Coombes, CD3(-)NK1.1(+) cells aid in the early induction of a Th1 response to an attaching and effacing enteric pathogen. Eur J Immunol, 2013. 43(10): p. 2638-49. 362. Kim, S., et al., In vivo natural killer cell activities revealed by natural killer cell-deficient mice. Proc Natl Acad Sci U S A, 2000. 97(6): p. 2731-6. 184 363. Wu, Y., Z. Tian, and H. Wei, Developmental and Functional Control of Natural Killer Cells by Cytokines. Front Immunol, 2017. 8: p. 930. 364. Goldberg, J.L. and P.M. Sondel, Enhancing Cancer Immunotherapy Via Activation of Innate Immunity. Semin Oncol, 2015. 42(4): p. 562-72. 365. Boyerinas, B., et al., Antibody-Dependent Cellular Cytotoxicity Activity of a Novel Anti- PD-L1 Antibody Avelumab (MSB0010718C) on Human Tumor Cells. Cancer Immunol Res, 2015. 3(10): p. 1148-1157. 366. Piccioli, D., et al., Contact-dependent stimulation and inhibition of dendritic cells by natural killer cells. J Exp Med, 2002. 195(3): p. 335-41. 367. Kim, J.R., et al., CS1 (SLAMF7) inhibits production of proinflammatory cytokines by activated monocytes. Inflamm Res, 2013. 62(8): p. 765-72. 368. Whiteside, T.L., Inhibiting the inhibitors: evaluating agents targeting cancer immunosuppression. Expert Opin Biol Ther, 2010. 10(7): p. 1019-35. 369. van Vliet, S.J., et al., Innate signaling and regulation of Dendritic cell immunity. Curr Opin Immunol, 2007. 19(4): p. 435-40. 370. Luo, X., et al., Intratumoral expression of MIP-1beta induces antitumor responses in a pre-established tumor model through chemoattracting T cells and NK cells. Cell Mol Immunol, 2004. 1(3): p. 199-204. 371. Aldhamen, Y.A., S.S. Seregin, and A. Amalfitano, Immune recognition of gene transfer vectors: focus on adenovirus as a paradigm. Front Immunol, 2011. 2: p. 40. 372. Win, S.J., et al., Cross-presentation of epitopes on virus-like particles via the MHC I receptor recycling pathway. Immunol Cell Biol, 2011. 89(6): p. 681-8. 373. Irwin, R., et al., Colitis-induced bone loss is gender dependent and associated with increased inflammation. Inflamm Bowel Dis, 2013. 19(8): p. 1586-97. 374. Lee, T., et al., beta-Arrestin-1 deficiency protects mice from experimental colitis. Am J Pathol, 2013. 182(4): p. 1114-23. 375. Britton, R.A., et al., Probiotic L. reuteri treatment prevents bone loss in a menopausal ovariectomized mouse model. J Cell Physiol, 2014. 229(11): p. 1822-30. 376. Coxon, F.P., Fluorescence imaging of osteoclasts using confocal microscopy. Methods Mol Biol, 2012. 816: p. 401-24. 377. Gregory, C.A., et al., An Alizarin red-based assay of mineralization by adherent cells in culture: comparison with cetylpyridinium chloride extraction. Anal Biochem, 2004. 329(1): p. 77-84. 185 378. Porter, A., et al., Quick and inexpensive paraffin-embedding method for dynamic bone formation analyses. Sci Rep, 2017. 7: p. 42505. 379. Aldhamen, Y.A., et al., Vaccines expressing the innate immune modulator EAT-2 elicit potent effector memory T lymphocyte responses despite pre-existing vaccine immunity. J Immunol, 2012. 189(3): p. 1349-59. 380. Becker, C., M.F. Neurath, and S. Wirtz, The Intestinal Microbiota in Inflammatory Bowel Disease. ILAR J, 2015. 56(2): p. 192-204. 381. Amodio, G., et al., HLA-G expression levels influence the tolerogenic activity of human DC-10. Haematologica, 2015. 100(4): p. 548-57. 382. Comiskey, M., et al., Evidence that HLA-G is the functional homolog of mouse Qa-2, the Ped gene product. Hum Immunol, 2003. 64(11): p. 999-1004. 383. Dias, F.C., et al., Human leucocyte antigen-G (HLA-G) and its murine functional homolog Qa2 in the Trypanosoma cruzi Infection. Mediators Inflamm, 2015. 2015: p. 595829. 384. Yan, J., et al., In vivo role of ER-associated peptidase activity in tailoring peptides for presentation by MHC class Ia and class Ib molecules. J Exp Med, 2006. 203(3): p. 647- 59. 385. Lyons, A.B., Analysing cell division in vivo and in vitro using flow cytometric measurement of CFSE dye dilution. J Immunol Methods, 2000. 243(1-2): p. 147-54. 386. Wu, H.Y., et al., In vivo induction of Tr1 cells via mucosal dendritic cells and AHR signaling. PLoS One, 2011. 6(8): p. e23618. 387. McLellan, M.A., N.A. Rosenthal, and A.R. Pinto, Cre-loxP-Mediated Recombination: General Principles and Experimental Considerations. Curr Protoc Mouse Biol, 2017. 7(1): p. 1-12. 388. Cifaldi, L., et al., ERAP1 regulates natural killer cell function by controlling the engagement of inhibitory receptors. Cancer Res, 2015. 75(5): p. 824-34. 186