T cell regulation of acute and chronic viral infection

Transcription

1 University of Iowa Iowa Research Online Theses and Dissertations Spring 2016 T cell regulation of acute and chronic viral infection Allison Fae Christiaansen University of Iowa Copyright 2016 Allison Fae Christiaansen This dissertation is available at Iowa Research Online: Recommended Citation Christiaansen, Allison Fae. "T cell regulation of acute and chronic viral infection." PhD (Doctor of Philosophy) thesis, University of Iowa, Follow this and additional works at: Part of the Microbiology Commons

2 T CELL REGULATION OF ACUTE AND CHRONIC VIRAL INFECTION by Allison Fae Christiaansen A thesis submitted in partial fulfillment of the requirements for the Doctor of Philosophy degree in Microbiology in the Graduate College of The University of Iowa May 2016 Thesis Supervisor: Professor Steven M. Varga

3 Graduate College The University of Iowa Iowa City, Iowa CERTIFICATE OF APPROVAL This is to certify that the Ph.D. thesis of PH.D. THESIS Allison F. Christiaansen has been approved by the Examining Committee for the thesis requirement for the Doctor of Philosophy degree in Microbiology at the May 2016 graduation. Thesis Committee: Steven M. Varga, Thesis Supervisor Kevin L. Legge Wendy J. Maury Stanley Perlman Thomas J. Waldschmidt

4 ACKNOWLEDGEMENTS I would like to acknowledge all the people who supported me as I have worked towards my doctorate. I would like to acknowledge my mentor, Dr. Steven Varga who has played a pivotal role in training me as a scientist. I would also I would also like to thank my past and present fellow lab members Dr. Kayla Weiss, Dr. Paola Boggiatto, Dr. Daniel McDermott, Dr. Cory Knudson, Stacey Hartwig and Megan Stoley. They provided both excellent scientific guidance as well as making the lab an enjoyable place to come to each day. Finally, I would like to thank my friends and family for their continuous support. I would especially like to thank my eclectic hometown friends for always providing me with something to laugh about and my parents for their endless encouragement. ii

5 ABSTRACT A balanced immune response is required to mediate clearance of a virus infection without immune-mediated disease. CD4 and CD8 T cells are capable of both exerting antiviral effector functions and regulating the immune response. The regulatory T cell (Treg) subset of CD4 T cells helps to modulate immune activation and inflammation. During respiratory syncytial virus (RSV) infection in mice, conventional CD4 T-cellmediated cytokine production has been shown to contribute to immune-mediated pathology. I demonstrate that Tregs are critical to control immunopathology during RSV infection. This was demonstrated through diphtheria toxin (DT)-mediated Treg elimination in a mouse strain expressing the DT receptor (DTR) under the control of the Foxp3 promoter. However, these mice were unable to maintain extended Treg depletion limiting the effectiveness of this model. In addition, DT-treated wild-type (WT) mice were found to be a necessary control for adverse DT-induced disease. In humans, I have shown that activated Tregs are reduced in the peripheral blood of RSV-infected infants compared to controls. RSV-infected infants also exhibited an increased proinflammatory cytokine response in nasal aspirates. However, the alarmin cytokine IL-33, which has been shown to mediate Treg homeostasis, was the only cytokine that exhibited reduced protein levels in RSV-infected infants compared to controls. Thus, severe RSV infection in infants may be due to lack of proper Treg-mediated immune regulation. Similar to RSV, regulation of the T cell response during chronic viral infection with lymphocytic choriomeningitis virus (LCMV) is vital to prevent immune-mediated pathology. During LCMV and human chronic viral infections, CD4 and CD8 T cells iii

6 exhibit T cell exhaustion where they lose the ability to exert effector functions. However, a functional CD4 and CD8 T cell response is required for viral clearance. During human chronic viral infection, an association between increased CD4 and CD8 T cell function and enhanced viral control has been identified that can be influenced by genetic factors. I aimed to identify the contribution of the host genetic factors that contribute to enhanced CD8 T cell function and viral control using the LCMV model. I found that increasing the major histocompatibility complex (MHC) diversity resulted in enhanced viral control in both a C57BL and BALB genetic background. Thus, induction of a broader T cell response was associated with enhanced viral control. However, mice expressing a heterozygous MHC on the C57BL background also exhibited mortality following chronic viral infection. Both CD4 and CD8 T cells were shown to contribute to this mortality and exhibited reduced T cell exhaustion during LCMV infection in these mice. Heterozygous MHC expression on the C57BL mouse background was also associated with an increased T helper (Th)-1 skewed CD4 T cell response compared to mice on the BALB background. Furthermore, CD4 T-cell-mediated IFN-γ production contributed to both CD8 T cell effector activity and mortality during chronic LCMV infection. Thus, both T cell epitope diversity and host genetics contribute to LCMV-induced mortality. Collectively, my data highlight both the need for effective immune-meditated viral control and regulation of T-cell-mediated pathology during both acute and chronic viral infections. iv

7 PUBLIC ABSTRACT The impact of infectious disease has been significantly reduced since the development of preventative vaccines and effective healthcare. Despite these advances, infectious diseases still account for approximately one-quarter of human deaths each year. Further development of new vaccines and treatments requires an in depth understanding of how the immune response eliminates the pathogen. In some infections an individual s own immune system causes more damage to the individual than the pathogen itself. My thesis focuses on understanding the regulatory mechanisms of T cells, a subset of immune cells, during infection. During respiratory syncytial virus (RSV), the leading cause of respiratory infection in infants, severe disease is mediated by an excessive immune response. I found that RSV infection induces a robust immune response that is not associated with a elevated regulatory T cell response in hospitalized infants. Thus, lack of immune regulation by regulatory T cells may contribute to disease severity in RSV-infected infants. Furthermore, chronic viral infections, such as human immunodeficiency virus (HIV), result from the inability of the immune system to eliminate the virus from the infected individual. The immune system s inability to eradicate the invading pathogen is partially due to excessive regulation of the T cell response. My project investigates the role of host genetic diversity on the T cell response during chronic infection in a murine model. I found that a specific subset of T cells are associated with reduced suppression of the T cell response. Thus, viral infections require a balance between immune-mediated viral clearance and immune-mediated disease. v

8 TABLE OF CONTENTS LIST OF TABLES...x LIST OF FIGURES... xi LIST OF ABBREVIATIONS... xiii CHAPTER I INTRODUCTION...1 T Cell Responses...1 Major histocompatibility complex...2 Respiratory Syncytial Virus...3 CD4 T-cell-mediated pathology in RSV infection...4 Lymphocytic Choriomeningitis Virus...7 LCMV replication and model strains...8 Cause of disease in LCMV...10 Regulatory T Cells...10 Mechanisms of regulation by Tregs...11 Treg activation and specificity...14 Tregs and viral infection...16 Tregs and RSV infection...18 T Cell Exhaustion...20 CD8 T cell exhaustion...20 CD4 T cell exhaustion...21 Antigen persistence...23 Inhibitory receptors...24 Soluble mediators of exhaustion...26 Cellular mediators of exhaustion...28 Thesis Objectives...29 CHAPTER II LIMITATIONS OF FOXP3 + TREG DEPLETION FOLLOWING VIRAL INFECTION IN DEREG MICE...31 Abstract...31 Introduction...32 vi

9 Material and Methods...33 Mice and infection...33 DT treatment...34 Blood collection and staining...34 Statistical analysis...34 Results...35 Foxp3 + Treg depletion in DEREG mice is transient...35 DT administration induces disease in both DEREG and WT naïve mice...36 Foxp3 + Treg depletion is not maintained following acute RSV infection...36 DT treatment enhances the severity of RSV-induced disease...37 DT treatment induces morbidity and mortality following LCMV Cl-13 infection...37 Discussion...38 CHAPTER III ALTERED TREG AND CYTOKINE RESPONSES IN RSV- INFECTED INFANTS...48 Abstract...48 Introduction...49 Materials and Methods...50 Study population...50 Nasal fluid cytokine determination...51 Peripheral blood mononuclear cell isolation...51 Flow cytometry...51 Statistical analysis...52 Results...52 Characteristics of the study population...52 Activated Tregs are reduced in RSV-infected infants...53 Cytokine concentrations increase in RSV-infected infants...54 Resting Treg frequencies decrease with age...55 Resting Treg frequencies decrease as RSV diseases progresses...55 Th17-associated cytokines are associated with reduced respiratory signs and symptoms...56 Discussion...57 vii

10 CHAPTER IV HOST GENETICS PLAY A CRITICAL ROLE IN CONTROLLING CD8 T CELL FUNCTION AND LETHAL IMMUNOPATHOLOGY DURING CHRONIC VIRAL INFECTION...73 Abstract...73 Introduction...74 Materials and Methods...76 Mice and infection...76 Plaque assays...77 Antibody depletion...77 Flow cytometry analysis...77 Histology and Evans Blue staining...79 Statistical analysis...79 Results...79 Host genetics influences viral persistence and Cl-13-induced mortality...79 H-2 bxd haplotype mice prioritize H-2 d -restricted CD8 T cell responses over H-2 b -restricted responses following acute LCMV infection...80 D2B6F1 mice maintain the highest frequency of functional CD8 T cells during chronic viral infection...81 Impact of host genetic background on the effector functions of CD8 T cells during chronic infection...83 CD8 T cell exhaustion is influenced by the genetic background of the host...84 D2B6F1 mice exhibit an increased CD4 T cell response during chronic viral infection...85 H-2 bxd haplotype mice on the C57BL genetic background maintain Th1- polarized CD4 T cells during chronic infection...85 CD8 and CD4 T cells contribute to Cl-13-induced mortality in D2B6F1 mice...86 IFN-γ contributes to mortality and viral control in D2B6F1 mice...87 D2B6F1 mice exhibit liver lesions following Cl-13 infection...87 Discussion...88 CHAPTER V DISCUSSION AND FUTURE DIRECTIONS Diptheria Toxin Receptor-Expressing Mouse Models IL-33 and Treg Responses During Human RSV Infection Host Genetics and Chronic Viral Infection Th1-biased immune response during chronic viral infection viii

11 MHC diversity and T-cell-mediated viral control T-Cell-Based Vaccines and Therapy Summary of Results Chapter II Chapter III Chapter IV REFERENCES ix

12 LIST OF TABLES Table 1. Demographic of study population Table 2. Medical intervention of RSV-infected infants Table 3. Medical intervention did not correlate with an altered immune response Table 4. Cytokines and chemokines in nasopharyngeal aspirates (pg/ml) Table 5. Correlations between CD4 T cell populations and cytokines (p-values) x

13 LIST OF FIGURES Figure 1. Foxp3 + Treg depletion is not maintained in naïve DEREG mice Figure 2. Additional DT treatment does not extend the duration of Foxp3 + Treg depletion Figure 3. DT treatment induces disease in naïve WT and DEREG mice Figure 4. Foxp3 + Treg depletion is not maintained following RSV infection Figure 5. DT treatment enhances disease in both WT and DEREG mice following RSV infection...46 Figure 6. DT treatment enhances disease and mortality in WT and DEREG mice following Cl-13 infection Figure 7. atreg frequencies are reduced in the peripheral blood of RSV-infected infants Figure 8. Conventional CD4 T cell frequencies are maintained in the peripheral blood of RSV-infected infants Figure 9. Proliferation of CD4 T cell subsets is not altered by RSV infection Figure 10. rtregs decrease with age Figure 11. rtreg frequencies decrease as RSV disease progresses Figure 12. Th17-associated cytokines correlate with reduced difficulty breathing and retraction Figure 13. Reduced immune regulation results in enhanced inflammatory cytokines and increased disease severity in RSV-infected infants Figure 14. H-2 bxd mice with the C57BL background succumb to Cl-13 infection Figure 15. H-2 bxd mice with the C57BL background clear Cl-13 infection Figure 16. H-2 bxd mice control chronic viral infection more rapidly than inbred mice Figure 17. D2B6F1 mice induce the largest frequency of activated CD8 T cells Figure 18. H-2 bxd haplotype mice maintain H-2 d -restricted responses xi

14 Figure 19. H-2 bxd haplotype mice increase the frequency of H-2 b -restricted responses following Cl-13 infection Figure 20. C57BL background H-2 bxd mice exhibit increased CD8 T cell effector functions compared to BALB background mice Figure 21. C57BL background mice exhibit reduced CD8 T cell exhaustion compared to BALB background mice Figure 22. The C57BL background induces a robust CD4 T cell response during chronic viral infection Figure 23. Cl-13 infection of mice on a C57BL background results in an enhanced Th1 response Figure 24. Both CD4 and CD8 T cells contribute to Cl-13-induced mortality in D2B6F1 mice Figure 25. CD4 T cell depletion results in reduced CD8 T cell activation in D2B6F1 mice Figure 26. IFN-γ production is mediated by CD4 T cells following Cl-13 infection Figure 27. IFN-γ contributes to mortality and viral control following Cl-13 infection in D2B6F1 mice Figure 28. IFN-γ contributes to CD4 and CD8 T cell activation following Cl-13 infection in D2B6F1 mice Figure 29. D2B6F1 mice express focal necrotic liver lesions during Cl-13 infection Figure 30. Robust Th1 responses enhance CD8 T-cell-mediated viral control and mortality in D2B6F1 mice following Cl-13 infection xii

15 LIST OF ABBREVIATIONS AIDS acquired immunodeficiency syndrome ANOVA analysis of variance APCs antigen presenting cells Arm Armstrong strain of LCMV atreg activated regulatory T cell B6 C57BL/6 mice BAC bacterial artificial chromosome BCF1 first generation cross BALB.B x BALB/c BHK baby hamster kidney Cl-13 clone 13 strain CTLA-4 cytotoxic T lymphocyte antigen 4 D2B6F1 first generation cross B10.D2 x C57BL/6 DC dendritic cell DEREG DEpletion of REGulatory T cell DT diptheria toxin DTR diptheria toxin receptor egfp enhancer green fluorescent protein F1 first generation FAS TNF receptor superfamily member 6 FCS fetal calf serum Foxp3 forkhead head box P3 GP glycoprotein HAV hepatitis A virus HB-EGF heparin-binding epidermal growth factor-like growth factor HBV hepatitis B virus HCV hepatitis C virus HIV human immunodeficiency virus HLA human leukocyte antigen HSV herpes simplex virus ICER inducible camp early repressor IDO indoleamine 2,3-dioxygenase IFN-I type I interferon IFN-γ interferon gamma IL interleukin i.p. intraperitoneal injection IPEX immunodysregulation polyendocrinopathy X-linked syndrome iregdc regulatory antigen presenting cell subset xiii

16 i.v. intravenous injection L polymerase LAG-3 lymphocyte activation gene-3 LCMV lymphocytic choriomeningitis virus MHC major histocompatibility complex Nfatc1 nuclear factor of activated T cell c1 NK natural killer NP nucleoprotein PBL peripheral blood lymphocytes PBMC peripheral blood mononuclear cell PD-1 programmed cell death protein 1 PD-L1 programmed death ligand 1 PFU plaque forming unit ptreg peripheral regulatory T cell R receptor RSV respiratory syncytial virus rtreg resting regulatory T cell SEM standard error of the mean SIV simian immunodeficiency virus STAT1 signal transducer and activator of transcription factor Tbet T-box transcription factor TCR T cell receptor Tfh T follicular helper cell TGF-β transforming growth factor beta Th T helper TIM-3 T cell immunoglobulin and mucin domain 3 TNF tumor necrosis factor Treg regulatory T cell ttreg thymic regulatory T cell VACV vaccinia virus WNV West Nile virus WT wild-type Z small polypeptide of unknown function of LCMV xiv

17 CHAPTER I INTRODUCTION A balanced immune response is crucial for pathogen clearance and to prevent immune-mediated pathology. Induction of the immune response creates an inflammatory environment to recruit cells necessary for either killing the pathogen or eliminating pathogen-infected cells. However, the process of eliminating a pathogen can cause damage to the host through excessive killing of host cells resulting in disruption of normal tissue functions. Thus, the immune system contains many regulatory mechanisms to prevent immune-mediated pathology. The adaptive immune response is tightly regulated by its requirement for antigen specificity for activation, however, this is not always sufficient to prevent T-cell-mediated disease. The T cell response contains both T cells specifically dedicated to immune suppression as well as T-cell-intrinsic inhibitory mechanisms to help regulate the immune response. This chapter will focus on these two major T cell regulatory pathways and their role in respiratory syncytial virus (RSV) and lymphocytic choriomeningitis (LCMV) infections. T Cell Responses CD4 and CD8 T cells are induced following recognition of antigens presented on the major histocompatibility complex (MHC). Productive T cell stimulation requires additional signals provided by co-stimulatory molecules such as CD28 (1-3), and cytokines such as type-i interferon (IFN-I) and interleukin (IL)-12 (4-6). These signals combine to induce optimal T cell activation, differentiation and proliferation. Following activation, CD4 T cells help to direct the immune response through the production of 1

18 cytokines. CD4 T cells play a vital role in promoting both CD8 T cell and B cell responses. CD8 T cells are effector cells that induce apoptosis of infected cells that express pathogen-derived antigens on MHC class I (MHC-I) molecules on their surface. Thus, the induction of the CD4 and CD8 T cell response is regulated by their T cell receptor (TCR)-specificity for peptide-mhc resulting in a directed T cell response specific for the pathogen. Major histocompatibility complex MHC-I molecules are expressed on all nucleated cells while MHC class II (MHC- II) molecules are expressed on antigen presenting cells (APCs) such as dendritic cells (DCs), macrophages, and B cells. In humans, MHC-II molecules are also expressed on endothelial cells, epithelial cells, and fibroblasts (7, 8). MHC-I molecules present peptides classically derived from intracellular proteins, however, exogenous peptides can also be presented through a process termed cross-presentation (8, 9). Peptides presented by MHC-II molecules are typically derived from exogenous proteins following APCmediated phagocytosis or endocytosis (7, 10, 11). Pathogen-derived proteins are presented following proteasome cleavage and processing in the ER into short peptide sequences (12). Peptides are then loaded onto the MHC complex and transported to the cell surface for presentation. MHC-I presents peptides to CD8 T cells while MHC-II presents to CD4 T cells. MHC molecules can only present peptides that specifically bind to the anchor residues in their peptide-binding groove. Thus, the MHC allelic variants available within the host determine what peptides can be presented. MHC genes are located on 2

19 chromosome 17 in mice termed the H2 genes and chromosome 6 in humans termed the human leukocyte antigen (HLA) genes. The MHC loci in humans contain more than 200 genes and are the most polymorphic of all human genes (13). In addition, MHC allele expression is codominant allowing for heterozygous individuals to express a greater variety of MHC molecules. Each individual can induce a unique T cell response to a single pathogen resulting in varying control of infections (14-16). This allows for MHC controlled genetic diversity to contribute to effective T cell responses with pathogen clearance as well as excessive T cell responses resulting in immune-mediated disease. In addition, each pathogen induces a specific T cell response dependent on the peptides that can be presented by MHC molecules that is distinct from other pathogen responses within the same individual. Respiratory Syncytial Virus RSV is the leading cause of lower respiratory tract disease in children worldwide (17), with the majority of children infected with RSV by the age of two (18). An estimated 132, ,000 RSV-associated hospitalizations occur annually in the United States in children under the age of five (19). In addition, 3-4 million children are hospitalized and an estimated 66, ,000 children die annually worldwide from RSV-associated acute lower respiratory infection (17). With the high global burden of RSV-induced disease, development of an efficacious vaccine has been a high priority. However, despite persistent efforts over the last 50 years, there is currently no licensed RSV vaccine. Although the host inflammatory response is necessary for viral clearance, it also induces damage to the lung. The CD4 T 3

20 cell response has been implicated in contributing to immune-mediated pathology during RSV infection in children (20). The mouse model of RSV infection has been extensively used to examine the role of CD4 T cells in the induction and regulation of the immune response and their impact on disease severity. However, a better understanding of the balance between the protective and pathogenic properties of RSV-specific CD4 T cells is needed in the ongoing struggle to create a successful RSV vaccine. CD4 T-cell-mediated pathology in RSV infection CD4 T cell depletion in mice prior to RSV infection results in amelioration of RSV-induced disease (21). Furthermore, RSV infection can induce mucus production and airway hyperreactivity, signs that are associated with CD4 T helper (Th)2 immune responses. Infants are more prone towards Th2 responses than adults as they express higher levels of IL-4 receptor (R) on CD4 T cells in cord blood following in vitro RSV stimulation compared to CD4 T cells from the peripheral blood of adults (22). In addition, compared to adults, neonatal dendritic cells (DCs) produce less IL-12 resulting in reduced interferon gamma (IFN-γ) production by neonatal CD4 T cells (23). Together these unique features of the immature immune system in neonates predispose infants towards Th2 responses. Conflicting reports exist regarding the presence of Th2-associated cytokines in RSV-infected infants. While higher levels of IFN-γ than IL-4 appeared to be present in nasal washes of RSV-infected infants (24), the ratio of IL-4 to IFN-γ was increased with more severe disease (25). Furthermore, Murai et al. observed higher levels of IFN-γ than IL-4 in nasal washes but only IL-4 exhibited a significant increase in cytokine levels 4

21 compared to uninfected controls due to basal levels of IFN-γ expression in the controls (26). Similar results are found following stimulation of peripheral blood mononuclear cells (PBMCs) from RSV-infected infants (27-29). Furthermore, increased severity of RSV-induced disease was associated with higher cortisol levels, which is known to inhibit IFN-γ responses (28). These data suggest that both Th1 and Th2 CD4 T cells are present in both the airways and blood during RSV infection. A recent study examining over 450 RSV-infected infants under one year of age collected over 5 years with a range of disease from mild to severe did not find detectable levels of either Th1- (i.e. IFN-γ) or Th2- (i.e. IL-4, IL-5, IL-13) associated cytokines in nasal washes (30). However, the lower limit of detection in their assay appears to have been above those levels detected by other groups for some cytokines (24, 25, 30). This large study highlights the low levels of Th2 cytokines detected in RSV-infected infants indicating that these cytokines may not play a critical role in disease. Furthermore, multiple Th2 cytokines are only detectable in the upper respiratory tract as measured in nasal wash samples, but in the same infants only IL-4 is detected in the lower respiratory tract from tracheal aspirates. This indicates that Th2 cytokines may not be strongly associated with severe lower respiratory tract disease following a primary RSV infection. Thus, the contribution of Th2 cells in adding to the severity of RSV-induced disease in infants remains unclear. The magnitude of Th2 responses following RSV infection in mice appears to be dependent on the viral strain. The line 19 and 2-20 strains of RSV induced significantly greater amounts of the Th2-associated cytokine IL-13 in the lung than the A2 and Long strains, resulting in increased mucus production and airway hyperreactivity (31, 32). 5

22 Blockade of IL-13 reduced mucus production and disease in all RSV strains (31, 33). During infection with RSV strains where minimal Th2 responses are observed, disruption of the immune response can result in an increased Th2 response. For example, regulatory T cell (Treg) depletion and natural killer (NK) cell depletion have both been shown to increase the pro-inflammatory CD4 T cell response, including an increased Th1 as well as Th2 response, resulting in enhanced disease (34, 35). Thus, data from murine studies indicate that an increased pro-inflammatory CD4 T cell response, including IL-13, results in an increase in RSV-induced disease severity. The Th17 subset of CD4 T cells has been shown to induce airway hyperreactivity during asthma (36). IL-17 was elevated in the tracheal aspirates but not nasal washes of RSV-infected infants compared to controls (37). The presence of IL-17 specifically in the lower airways indicates that IL-17, unlike the Th2 cytokines, may play a more significant role in the lower respiratory tract rather than in the upper airways. In addition, IL-17, but not the Th2-associated cytokines IL-4 and IL-13, was shown to stimulate human primary tracheobronchial epithelial cells cultured in an air liquid interface to produce the mucusrelated mrna transcripts, muc5b and muc5ac (38, 39). Following RSV infection in mice, IL-17 neutralizing antibody treatment resulted in reduced mucus production as measured by both mucus-associated mrna transcripts and histology (37). Thus, Th17 cells may play a critical role in the induction of mucus production and airway hyperreactivity in the lower respiratory tract during RSV infection. It is clear that RSV infection results in a pro-inflammatory response resulting in mucus production and airway hyperreactivity. However, the cellular source of these 6

23 cytokines is less certain. In addition to CD4 T cells, multiple cell types in the lung are capable of producing IL-4, IL-5, IL-13 and IL-17, including eosinophils, mast cells, type- 2 innate lymphoid cells and macrophages (40-43). However, IL-4, IL-5, IL-13 and IL-17 cytokine production by these cells have not been examined in RSV infection. Furthermore, the disease signs and cytokine profile observed during RSV infection is similar to what is seen during some asthmatic responses of which allergen-specific CD4 T cells are known to play a major role. Owing to the similarities with asthma observed in humans and the pro-inflammatory role that CD4 T cells play during murine RSV infection, CD4 T cells may be a major contributor to immune-mediated disease during RSV infections in infants. No matter the cell type responsible, the presence of the proinflammatory environment created during RSV infection indicates that regulation of the immune response is vital to prevent excessive damage to host tissue. Lymphocytic Choriomeningitis Virus Chronic viral infections pose a serious healthcare concern resulting in substantial mortality worldwide. Currently, the World Health Organization estimates that 35 million people are chronically infected with human immunodeficiency virus (HIV). In 2013, an estimated 1.5 million people worldwide died from acquired immunodeficiency syndrome (AIDS)-related illnesses. In addition, more than 130 million people suffer from a chronic hepatitis C virus (HCV) infection and 500 thousand individuals succumb to the disease each year (44, 45). Both of these chronic viral infections pose a substantial health risk to humans and lack an effective vaccine. Increasing our understanding of how the immune 7

24 response combats chronic viral infection is vital to prevent further infections and to eliminate this crippling epidemic. The inability of human chronic infections such as HIV and HCV to induce a productive infection in non-human models has required the use of a model pathogen to characterize the immune response during chronic viral infection. The well-characterized natural mouse pathogen LCMV has been used to delineate many basic features of immunology including MHC restriction, immune suppression and T cell memory. LCMV strains that result in persistent infections have been used to define T cell exhaustion which is now known to occur in humans during both chronic infections and tumor development. However, many questions still exist surrounding the mechanisms of T cell exhaustion and the development of chronic viral infection. LCMV replication and model strains LCMV is an enveloped negative-sense RNA virus of the arenaviridae family. The LCMV genome consists of two segments, short and long, with each segment encoding two proteins in opposite orientations (46, 47). The short segment contains the nucleoprotein (NP) and the two subunits of the glycoprotein (GP). The long segment contains the RNA polymerase (L) and a small polypeptide of unknown function (Z). The organization of the viral genome allows for increased early expression of the NP and L proteins as they are translated and transcribed directly from the viral genome (48). However, the GP and Z genes are coded in the opposite orientation of NP and L and require the replication of a positive-sense genome intermediate prior to transcription and translation (48). This has been shown to impact the level of viral protein expression and 8

25 resulting antigen presentation which influences the antigen-dependent CD8 T cell immunodominace hierarchy (49). Two strains of LCMV are regularly utilized to examine the differences between the development of acute and chronic infection in mice. The Armstrong (Arm) strain induces an acute infection whereas the clone-13 (Cl-13) strain causes a chronic infection. The Cl-13 strain was derived from mice neonatally infected with LCMV Arm (50). Three amino acid mutations were identified in Cl-13 that differ from Arm, two of which have been implicated in promoting viral persistence (51). A mutation in L (K1079Q) allows for increased viral replication (51). In addition, a mutation in GP (F260L) mediates enhanced binding to the virus receptor alpha-dystroglycan, resulting in increased infection of CD11c + DCs and altered tissue tropism (52-54). These two mutations are sufficient to induce persistent viral infection for 70 days within the blood of mice (51). Importantly, all known CD8 T cell epitopes are identical between the two strains (55). Thus, LCMV Arm and Cl-13 strains allow study of the development of acute versus chronic infections. Acute LCMV Arm infection is cleared in mice by the CD8 T cell response (56, 57). In contrast, Cl-13 infection induces rapid viral replication that cannot be cleared by the CD8 T cell response (51, 55). While both CD4 and CD8 T cells help to control viral replication, Cl-13 is eventually cleared following the development of a B cell response and antibody production (58, 59). In addition, the antibody response is dependent on CD4 T cell help (60, 61). Thus, LCMV provides an ideal model to study the induction of both T and B cell responses during chronic viral infection. 9

26 Cause of disease in LCMV The idea of immune-mediated disease was first described using the LCMV mouse model (62, 63). Studies in LCMV determined that transfer of T cells following an intracranial infection with the Arm strain resulted in illness independent of virus (64, 65). Further studies determined that this LCMV-induced meningitis and death was dependent on CD8 T cells (62). In the absence of CD8 T cells mice survive but undergo weight loss mediated by IFN-γ production from antigen-specific CD4 T cells (64, 66). No disease is observed in either BALB/c or C57BL/6 (B6) mice following either intraperitoneal (i.p.) or intravenous (i.v.) infection with the Arm strain. In contrast, high dose i.v. infection of either BALB/c or B6 mice with Cl-13 results in substantial weight loss without mortality. However, Cl-13 infection of NZB mice results in lethal vascular leakage mediated by IFN-I (67). These various routes and strains of LCMV infection demonstrate multiple ways in which the host immune response can mediate immunopathology. Thus, LCMV provides an ideal tool to understand the regulatory mechanisms of the immune response in order to prevent immune-mediated disease. Regulatory T Cells Tregs maintain immune homeostasis and inhibit excessive inflammatory responses to prevent damage to the host (68). Tregs are a subset of CD4 T cells that develop either in the thymus termed thymic Tregs (ttreg) or in the periphery from conventional CD4 T cells termed peripheral Tregs (ptreg). ttregs are generally specific for self-peptide whereas ptregs are generally specific for pathogen-derived peptides, however, pathogen-specific ttregs and self-specific ptregs have also been identified. The 10

27 transcription factor that serves as the master regulator of Treg differentiation was first identified in immunodysregulation polyendocrinopathy X-linked syndrome (IPEX) patients that lack expression of forkhead box P3 (Foxp3) (69). These individuals develop severe autoimmunity from lack of immune regulation (69). In addition, Tregs have also been shown to play a critical role in limiting excessive inflammatory responses and immune-mediated disease during infection. Thus, Tregs play a vital role in both immune maintenance and infection. Mechanisms of regulation by Tregs Multiple mechanisms of Treg function have been identified. Tregs can mediate immune suppression through both cytokine production and direct cell-to-cell interactions. Tregs have been shown to produce the inhibitory cytokines IL-10 (70), IL-35 (71), and transforming growth factor (TGF)-β (72). Many early in vitro studies showed that either blocking with neutralizing antibodies or using IL-10- or TGF-β-deficient Tregs did not alter their suppressive ability (73-75). However, additional studies with altered stimulation conditions and in vivo models have revealed an essential role for both IL-10 and TGF-β in Treg suppressive function. IL-10 deficiency has been shown to mediate spontaneous severe colitis, which can be abrogated by the addition of Tregs in an IL-10- dependent manner (70). Specific ablation of IL-10 production within Foxp3-expressing cells resulted in spontaneous colitis, augmented airway hyperreactivity, and increased skin hypersensitivity but not systemic autoimmunity (76). Tregs can also produce both soluble and membrane bound TGF-β (77). Whereas wild-type (WT) Tregs are sufficient to prevent the development of colitis, TGF-β-deficient Tregs are not (77, 78). 11

28 Furthermore, TGF-β production by Tregs was shown to be necessary for Treg suppression of allergen-induced airway responsiveness (79). Theses studies revealed distinct suppressive roles for both IL-10 and TGF-β. The recently discovered cytokine IL-35 has also been shown to contribute to Treg function (71). IL-35 is preferentially expressed by Tregs and Treg deficiency in either the Ebi3 or the il12a (p35) cytokine subunit reduces Treg suppressive activity and the ability of Tregs to prevent colitis (71, 80). Thus, IL-10, TGF-β and IL-35 all contribute to Treg-mediated suppression and the ability of Tregs to prevent the development of autoimmune disease. Suppression of IL-2 production is a mechanism by which Tregs modulate T cell proliferation (81). Tregs do not produce IL-2 due to Foxp3-mediated transcriptional suppression of the il2 gene (82, 83). The majority of IL-2 production comes from conventional CD4 T cells and to a lesser extent by CD8 T cells, natural killer (NK) cells and DCs (68, 81, 84). Both conventional T cells and Tregs require IL-2 for survival and proliferation. However, Tregs exhibit increased cell surface expression of the high affinity IL-2R subunit CD25 as compared to conventional T cells (85). However, conflicting reports exist over the role of IL-2 consumption as a mechanism of Tregmediated suppression. Several studies have shown that Treg suppression is independent of CD25 expression and addition of anti-cd25 to block IL-2 binding does not alter Treg suppressive activity (86, 87). However, many studies have shown that Tregs can suppress IL-2 mrna in conventional T cells through promoting nuclear translocation of the inducible camp early repressor (ICER) to suppress transcription of nuclear factor of activated T cell c1 (Nfatc1) (75, 88, 89). In addition, Tregs have been shown to regulate 12

29 the survival of IL-2-producing cells (85, 87). This results in a reciprocal relationship between the number of Tregs and the number of IL-2-producing cells. Thus, Tregs regulate the host immune response through control of IL-2 production and IL-2- producing cells. Tregs have also been shown to suppress the immune response by cytolysis of activated immune cells. Tregs have been shown to express both granzyme A and B (90-92). Granzyme B-deficient Tregs resulted in reduced Treg-mediated T effector cell death (90). In addition, human Tregs were shown to mediate activated T cell and monocyte cell death in a perforin-dependent manner (91). A murine model of graft versus host disease revealed a vital role for granzyme B expression by Tregs in Treg-mediated tolerization (93). Thus in both humans and mice, cytolysis of effector T cells and infiltrating monocytes represents an important suppressive mechanism by Tregs. Tregs express multiple inhibitory ligands that block maturation and function of antigen-presenting cells. Tregs in both mice and humans have been shown to express cytotoxic T lymphocyte antigen 4 (CTLA-4) and lymphocyte activation gene-3 (LAG-3) (94-97). CTLA-4 exerts its inhibitory function on DCs by trans-endocytosing its ligands, CD80 and CD86, removing them from the surface of DCs and preventing CD28 costimulation of effector T cells (98, 99). This results in a CTLA-4-dependent suppression by Tregs of the activation and proliferation of conventional T cells (100). In addition, studies have linked Treg expression of CTLA-4 with DC production of the T cell suppressor molecule indoleamine 2,3-dioxygenase (IDO), which catalyzes the degradation of tryptophan to kynurenine resulting in cell cycle arrest of effector T cells 13

30 (99, 101). While CTLA-4 is not the only mechanism of Treg suppression, Treg-specific ablation of CTLA-4 resulted in fatal autoimmune disease underlying the importance of CTLA-4 in Treg-mediated suppression (100). However, CTLA-4-deficient Tregs are still suppressive and able to prevent the development of autoimmunity in models of colitis (102) and experimental autoimmune encephalomyelitis (103). LAG-3, an additional inhibitory ligand expressed by Tregs, is a CD4 homolog that binds MHC class II and has been shown to negatively regulate effector T cell proliferation and activation ( ). In addition, Tregs have been shown to inhibit DC maturation in a LAG-3-dependent manner (108). However, LAG-3-deficient mice, unlike CTLA-4-deficient mice, did not develop autoimmune disease (94), indicating a less vital role for this inhibitory ligand in maintaining tolerance. It is clear that Treg expression of inhibitory ligands modulates the capacity of DCs to stimulate an effector T cell response resulting in suppression of T effector cells. Multiple suppressive mechanisms have been identified in Tregs that play vital roles in their suppressive function. It has been shown that these mechanisms are not redundant, but instead, they each play a distinct suppressive role specific for particular tissues and inflammatory settings. Thus, understanding the suppressive mechanism employed by Tregs in a particular disease setting is necessary for targeted manipulation of the Treg response. Treg activation and specificity Tregs require continuous TCR-activation by their cognate antigen to induce their suppressive effector functions. TCR ablation following Treg development was shown to 14

31 result in a naïve-like Foxp3 + Treg population with reduced suppressive activity (109, 110). Thymic-derived Tregs specific to self-peptides can respond when their self-peptide is expressed under inflammatory conditions during autoimmune diseases. It is also possible that pathogen-specific thymic Tregs may be cross-reactive for self-peptide (111). It is clear that pathogen-specific Tregs exist and have also been identified for several infectious agents including HCV (112), human papillomavirus (113) and RSV (114). The lack of phenotypic markers to distinguish thymic versus peripheral Tregs makes it difficult to determine the origin of Tregs responding during particular disease settings. While TCR stimulation appears to be required for Treg effector activity, inflammatory signals also contribute to Treg activation. Along with TCR stimulation, IL- 2 is required to prevent Treg anergy and to induce proliferation and activation (115, 116). TGF-β and retinoic acid are both important in the conversion of conventional CD4 T cells into Tregs following TCR activation ( ). Furthermore, signaling through Toll-like receptors expressed on Tregs can induce Treg suppressive activity (120). In addition, once activated, Tregs can enact their suppressive function locally in a nonspecific manner (121). Thus, if a Treg-specific antigen is present, Tregs can mediate their suppressive functions against T cells responding to another antigen in a bystander manner (111, 122). Thus, while stimulation through the TCR is required for Treg activation, Tregs specific for the invading pathogen may not be vital to Treg-mediated immune suppression during infections as long as a Treg-specific antigen is present at the site of infection. 15

32 Tregs and viral infection Tregs play a major role in limiting immunopathology during infection. At the same time, Treg-mediated suppression can prevent effective immune-mediated pathogen clearance contributing to the development of a chronic infection. Thus, examination of the role of Tregs during specific infections has determined if Tregs are either beneficial or detrimental to their host. The role of Tregs during chronic viral infections has been extensively studied. However, conflicting data exist as to whether Tregs are either detrimental in the establishment of a chronic infection or beneficial in preventing immune-mediated disease. During HIV infection, Tregs have been shown to be detrimental as an increased Treg frequency within the CD4 T cell compartment in lymphoid tissues and gastrointestinal mucosa has been associated with HIV progressors versus non-progressors ( ). In addition, the gp-120 protein expressed during HIV has been shown to promote Treg survival implying enhanced Treg survival promotes HIV replication (123). Furthermore, Tregs have been shown to inhibit HIV-specific CD8 T cell responses (127, 128). However, other groups have shown that overall Treg numbers are reduced in the peripheral blood during HIV infection due to CD4 T-cell-mediated depletion, correlating with disease progression ( ). These studies suggest that Tregs limit T cell hyperactivation and immune dysfunction thereby limiting immune-mediated disease (130). Thus, for therapeutic purposes it is unclear how manipulation of the Treg response during HIV infection could be utilized to enhance viral clearance and limit disease progression. 16

33 Similar to HIV, the role of Tregs during HCV and herpes simplex virus (HSV) infections are beneficial in some cases and detrimental in others. Tregs have been shown to be increased in the peripheral blood of HCV-infected patients compared to controls and Tregs were shown to exhibit suppressive activity (131, 132). In addition, elevated levels of the suppressive cytokines IL-10 and TGF-β were found in the peripheral blood of chronically infected individuals versus controls (112, 133). However, at sites of hepatic lesions, Treg numbers were associated with reduced histology scores (131). During HSV-1 infection, Tregs have been shown to limit the effector T cell response and viral clearance during the acute and memory phases of infection (134, 135). However, Tregs have also been shown to protect mice from chronic ocular immunoinflammatory lesions, which in humans can result in blindness, by limiting inflammatory cell infiltration and CD4 T cell activation (136, 137). These studies indicate that systemic Treg responses are detrimental to effective immune responses and pathogen clearance, however, Treg-mediated immune suppression can be beneficial at localized inflammatory sites in order to prevent immune-mediated tissue damage during tissue-specific chronic infection. During acute viral infection a strong immune response is induced that mediates clearance of the pathogen. However, in some acute infections disease occurs independently of pathogen clearance due to excessive immune responses and localized tissue damage. Tregs have been shown to play a vital role in the prevention of immunopathology of the lung following RSV infection and will be discussed in detail later in this chapter. However, while Tregs were shown to accumulate in the lung during 17

34 influenza infection, they did not significantly impact either disease severity or viral clearance (138, 139). In the skin, Tregs have been shown to limit cutaneous inflammatory disease following vaccinia virus (VACV) scarification in mice (140). During HSV-2 infection of the vaginal mucosa of mice, Tregs contributed to survival by mediating proper recruitment of effector cells to prevent viral spread to the nervous system (141). In addition to these acute localized infections, Tregs have also been associated with limiting disease during acute systemic infection. During West Nile virus (WNV) infection, which is predominantly asymptomatic, reduced Treg frequencies were associated with symptomatic infections in both humans and mice indicating a protective role for Tregs (142). Thus Tregs both limit inflammation and limit disease severity during acute viral infections. Tregs and RSV infection Tregs have been shown to play a vital role in preventing immune-mediated pathology during RSV infection in mice (143). Tregs expand in the lung and bronchoalvelar lavage following RSV infection in mice (143, 144). This expansion coincides with conventional CD4 and CD8 T cell expansion. RSV-specific Tregs have been identified in C57BL/6 mice by MHC class II tetramer staining following RSV infection, indicating that at least some Tregs are virus-specific (145). However, following RSV infection it is unclear if all Tregs are virus-specific or if other signals can trigger Treg activation and expansion. Treg depletion by either anti-cd25 antibody treatment or diptheria toxin (DT) treatment in Foxp3-diptheria toxin receptor (DTR) mice results in a significant increase 18

35 in the inflammatory response to RSV infection (35, 143, 144, 146). Treg depletion during RSV infection results in an increased influx of innate immune cells including NK cells, eosinophils and neutrophils into the lung after day 6 post-infection coinciding with when Tregs induce their suppressive effect (35, 143, 144, 146). In addition, there is an increase in the number of conventional CD4 and CD8 T cells following Treg depletion during RSV infection. This increased inflammatory environment in the absence of Tregs results in enhanced weight loss and increased mucus production following RSV infection (35, 143, 144, 146). These data indicate that Tregs play a vital role in limiting inflammation during RSV infection in mice. Tregs regulate the immune response to RSV through both direct interactions as well as by cytokine production. Following RSV infection in mice, Tregs upregulate the inhibitory molecule CTLA-4 (143). In addition, Tregs have been shown to express and release granzyme B following RSV infection in mice (146). This indicates that Tregs may be able to regulate the immune response by directly killing activated CD4 and CD8 T cells. Tregs, in addition to conventional CD4 T cells, can also broadly downregulate the immune system through the production of the immunosuppressive cytokine IL-10 during RSV infection ( ). These inhibitory actions by Tregs make this CD4 T cell subset vital in limiting the severity of RSV-induced disease. In humans, the role of Tregs during RSV infection is currently unclear. Activated conventional human CD4 T cells can transiently express Foxp3, requiring the use of additional markers to identify Tregs (150). In addition, obtaining lung tissue from RSVinfected infants to identify Tregs is not feasible. Thus, further study of the Treg response 19

36 following RSV infection in humans is necessary to increase our understanding of the role of Tregs in modulating RSV-mediated disease. T Cell Exhaustion During acute infection, naïve T cells are activated by their cognate antigen and inflammatory signals resulting in T cell proliferation and expansion to create a large effector population to mediate pathogen clearance. Following antigen clearance and the resolution of inflammation, the majority effector T cells die and those remaining develop a memory phenotype. These memory T cells require reactivation by their cognate-antigen to induce their effector functions. However, in the presence of persistent antigen presentation this program of T cell contraction and memory is altered. If effector T cell responses were maintained in the presence of persistent expression of antigen and inflammation this would result in continuous immune activation. This would ultimately induce immune-mediated pathology and mortality of the host. However, T cells contain an intrinsic regulatory pathway to shut down their effector functions in the presence of continuous antigen presentation and inflammation termed T cell exhaustion. This process, while allowing for the pathogen to persist, prevents T-cell-mediated damage to host tissues. Extensive studies in both mice and humans have examined this T cell inhibitory pathway. CD8 T cell exhaustion CD8 T cell exhaustion was first described in chronic LCMV infection (57). However, CD8 T cell exhaustion has also been observed in HIV, chronic hepatitis B virus 20

37 (HBV) and HCV infections ( ). Exhausted CD8 T cells were identified as antigenspecific T cells that lose their cytotoxic capacity as viral antigen persists (57). During exhaustion, loss of effector functions occurs in a hierarchical manner where polyfunctional T cells capable of co-producing cytokines are lost first. Initial stages of T cell exhaustion are characterized by a reduction of IL-2 production and proliferation (154). Further exhaustion results in a reduced frequency of antigen-specific CD8 T cells capable of producing tumor necrosis factor (TNF) (154). At late stages of chronic infection, exhausted CD8 T cells lose their capacity to exert cytotoxicity and the ability to produce IFN-γ (154). Finally, the exhausted CD8 T cell is completely deleted from the host (57, 154, 155). This decrease in the functional capacity of CD8 T cells promotes pathogen persistence but prevents immune-mediated pathology. However, CD8 T cells still maintain some antiviral function, as in their absence viral loads increase in LCMV Cl-13 (156, 157) and simian immunodeficiency virus (SIV) infections (158, 159). Thus, following chronic viral infection CD8 T cells loose their effector functions resulting in reduced viral control. CD4 T cell exhaustion Similar to CD8 T cells, CD4 T cells undergo a similar loss of IFN-γ, TNF, and IL-2 production during chronic infection ( ). However, CD4 T cells exhibit more diverse functional properties and increased plasticity than CD8 T cells. CD4 T cell exhaustion is often measured by loss of their Th1 phenotype. Th1 cells have been shown to contribute to clearance of viral infections (163). Th1-mediated IFN-γ production has been shown to enhance CD8 T cell effector functions during acute LCMV-Arm infection 21

38 (164). In addition, restoration of CD4 T-cell-mediated IFN-γ production is associated with enhanced viral control in humans (165, 166). However, during chronic viral infection in both mice and humans, CD4 T cells gradually convert from a Th1 to a T follicular helper (Tfh) cell phenotype (60, 167). Tfh cells produce high levels of IL-21 and migrate to germinal centers to promote B cell-mediated antibody production, maintenance of CD8 T cell function and viral control (60, ). However, it is currently unclear whether Th1 or Tfh cells play a more critical role in the control of chronic viral infection. Importantly, CD4 T cells greatly impact the extent of CD8 T cell exhaustion and viral control. CD4 T cell depletion in mice results in more rapid CD8 T cell exhaustion during chronic LCMV infection (157). Furthermore, addition of antigen-specific CD4 T cells can restore CD8 T cell effector function during chronic LCMV infection (171). IL- 21 production by Tfh cells limits the extent of CD8 T cell exhaustion observed during late stages of chronic LCMV infection (60, 168) and HIV infection (172). Furthermore, IL-21R-deficient CD8 T cells showed no defect during acute infection and memory development ( ). However, IL-21R deficient CD8 T cells exhibited a more rapid loss of function during chronic infection ( ), indicating a specific role for IL-21 signaling in exhausted CD8 T cells. Furthermore, Tfh cells are vital for antibody generation that eventually mediates viral clearance observed in chronic LCMV infection (61, 163). Thus, while chronic viral infection alters CD4 T cell function, CD4 T cells still play a vital role in maintaining CD8 T cell function and controlling viral titers. 22

39 Antigen persistence T cell exhaustion is driven by chronic antigen stimulation. Evidence for this comes from the differential rates of exhaustion observed in antigen-specific CD8 T cells that correlate with differential levels of antigen presentation in chronic LCMV infection (154, 173). Correlations with the level of antigen load and T cell exhaustion were also observed in human HIV infection (174). Further human studies in HIV and HBV demonstrate that suppression of viral replication and the resulting reduction of antigenpresentation allows for restoration of IFN-γ, TNF and IL-2 production by both CD4 (166, 175) and CD8 T cells (165, 176). CD4 T cell exhaustion and conversion to a Tfh phenotype is also dependent on antigen presentation as well as IL-6 (177, 178). In addition, proliferation capacity and IFN-γ production by exhausted CD8 T cells was restored when cells were removed from HIV-infected individuals and cultured in vitro (179, 180). The requirement of antigen presentation as a driver for T cell exhaustion was further shown through transfer of CD8 T cells primed during a chronic LCMV infection into uninfected mice (181, 182). Following transfer the CD8 T cells exhibited restoration of function and full development into a memory CD8 T cell phenotype (181, 182). Furthermore, continuous antigen-presentation alone in the absence of inflammation was capable of driving CD8 T cell exhaustion (155). Thus, CD8 T cell exhaustion requires the presence of persistent antigen. The longer CD8 T cells are in the presence of their cognate antigen, the more functionally exhausted they become, and the more difficult it becomes to restore their function (182, 183). As viral persistence progresses more CD4 T cells convert from a Th1 23

40 to a Tfh phenotype (60). In addition, the longer the CD8 T cells are exposed to chronic antigen presentation, the less their function can be restored when they are removed from the chronic antigen environment (181, 182). In fact, fully exhausted CD8 T cells require antigen persistence to survive (184). This is in contrast to memory CD8 T cells which undergo antigen-independent maintenance (185). These studies have shown that continuous antigen presentation drives the progressive loss of CD8 T cell effector function and Tfh conversion. Inhibitory receptors Inhibitory receptors also play a crucial role in mediating CD4 and CD8 T cell exhaustion. Inhibitory receptors are upregulated on both CD4 and CD8 T cells following T cell activation in both acute and chronic infections ( ). However, during chronic infection CD4 and CD8 T cells maintain increased cell surface expression of inhibitory receptors (185, 186). During chronic infection both CD4 and CD8 T cells maintain high expression levels of programmed cell death protein 1 (PD-1), LAG-3, T cell immunoglobulin and mucin domain (TIM)-3, CTLA-4 and CD39 in both mice (186, ) and humans ( ). Blockade of individual inhibitory receptors allows for restoration of some effector functions (189). Co-expression of inhibitory markers substantially reduces T cell function and develops with progressive length of exhaustion (183). Thus, inhibitory receptor expression has become a hallmark of T cell exhaustion. Importantly, inhibitory receptors utilize multiple mechanisms to modulate T cell function. Inhibitory receptors can block stimulation of T cells by competing for binding with activating receptors. For example, CTLA-4 can prevent co-stimulation via CD28 as 24

41 described in its inhibitory role on Tregs above. LAG3 competes with CD4 for binding to MHC-II and contains an intracellular signaling domain of unknown function (107, 195). In addition, inhibitory receptors can also reduce intracellular signaling through the TCR. PD-1 and TIM-3 have been shown to cluster with the TCR and dephosphorylate proximal TCR signaling molecules through their intracellular inhibitory motif (196, 197). By this mechanism, PD-1 and TIM-3 both reduce TCR microcluster formation and prevent TCR signaling (196, 197). CD39 hydrolyzes extracellular ATP which is processed into the inhibitory molecule adenosine to prevent phosphorylation of TCR proximal signaling (198, 199). In addition, some inhibitory receptors can utilize multiple inhibitory mechanisms (196, 197). However, the signaling mechanisms by which inhibitory ligands enact their intracellular functions and downstream gene regulation have not been fully elucidated. The biological importance of many of these inhibitory receptors has been determined through blockade of either the inhibitory receptor or its ligands. Most thoroughly studied, PD-1 and programmed death ligand 1 (PD-L1) blockade increases CD8 T cell effector functions and decreases viral load in both chronic LCMV infection in mice (189) and SIV infection in primates (200). Furthermore, in vitro PD-1 blockade in CD8 T cells isolated from HIV-infected patients allowed for increased proliferation and cytokine production (201). However, PD-1 blockade did not fully restore CD8 T cell function, indicating a role for other inhibitory pathways. LAG-3 blockade or LAG-3 knock-out mice alone did not alter CD8 T cell exhaustion (202), however, combined PD- 1 and LAG-3 blockade allowed for a greater increase in CD8 T cell function and viral 25

42 control than PD-1 blockade alone (183). TIM-3 blockade, similar to LAG-3 blockade, exhibited little effect alone, but when combined with PD-L1 blocking antibodies, CD8 T cell function was greatly increased compared to PD-L1 blockade alone (203). The majority of LAG-3 + and TIM-3 + cells co-express PD-1 (183). Thus blockade of either LAG-3 or TIM-3 alone may be ineffective due to residual PD-1 expression, whereas PD- 1 single positive cells can be restored, as they do not express other inhibitory receptors. The inhibitory ligand CTLA-4 is expressed mainly on exhausted CD4 T cells (204, 205). Thus, CTLA-4 blockade was shown to reduce inhibitory cytokine expression and restore Th1 cytokine production in CD4 T cells during SIV infection (206). These studies indicate that each inhibitory receptor has a distinct role in T cell exhaustion. Importantly, PD-1 has been shown to play an important role in regulating the immune response early during chronic viral infections. Pretreatment with PD-1/PD-L1 antibody blockade and PD-1-deficient mice infected with chronic LCMV succumbed to immune-mediated mortality (189, 207). However, PD-1 antibody blockade or PD-1 genetic deficiency had no impact on T cell development in naïve mice or on the T cell response during acute LCMV infection indicating that PD-1 only plays a role during chronic infection (189). Thus, whereas inhibitory receptor expression serves to limit CD8 T cell effector functions during chronic stages of infection, they are beneficial during initial infection at preventing immunopathology. Soluble mediators of exhaustion Both inflammatory and inhibitory cytokines have been implicated in contributing to T cell exhaustion. The most well-studied include IFN-I and IL-10. IFN-I during acute 26

43 infection is vital for the induction of antiviral responses. However, during chronic viral infection IFN-I has been shown to contribute to immune suppression and viral persistence (208, 209). IFN-I expression was only detectable in the first two days of chronic LCMV infection (208), however, IFN-I-inducible genes are detected well into chronic stages of infection in mice (209), primates (210) and humans (211). In addition, IFN-I was significantly decreased in non-pathogenic SIV infection compared to pathogenic infection (212). Blockade of IFN-I early during chronic LCMV infection mediated an increased frequency of IFN-γ- and TNF-producing CD4 T cells resulting in enhanced viral control (208, 209). However, CD8 T cell exhaustion was not altered in the absence of IFN-I (208, 209). Thus, IFN-I appears to induce genes and an environment that contributes to viral persistence without altering CD8 T cell exhaustion. The immunomodulatory cytokine IL-10 is increased during both chronic LCMV (213) and HIV infection (214, 215). An increased level of IL-10 was associated with HIV progressors compared to controllers (215). In addition, IL-10 has been shown to mediate enhanced CD4 and CD8 T cell exhaustion both in chronically infected mice (213, 216) and in cell cultures from chronically infected HIV and HCV patients (217, 218). Interestingly, blockade of IFN-I significantly reduced IL-10 levels but did not alter CD8 T cell exhaustion during chronic LCMV infection (208, 209). Importantly, blockade of IL-10 either prior to or during chronic LCMV infection mediated enhanced viral control (213, 216). However, the efficacy of IL-10 blockade was dependent on the chronic viral strain of LCMV examined (219). Thus, IL-10 blockade during human chronic viral 27

44 infection may not represent an effective therapeutic intervention for all chronic infections and individuals. Cellular mediators of exhaustion The induction of T cell exhaustion requires signals from other cells through both cytokines and direct cell-to-cell interactions. Antigen presenting cells (APCs) provide many of the inhibitory ligands and soluble mediators mentioned above. A regulatory APC subset termed iregdcs were recently found to produce IL-10 and express PD-L1 ( ). These cells have been shown to be the major source of IL-10 during chronic viral infection (222). iregdcs were identified by expression of CD95 and CD39 and required both IFN-γ and IFN-I for their development (221). Thus, iregdcs provide a vital link between IFN-I, IL-10 and CD8 T cell exhaustion. In addition to APCs, other cells have been shown to contribute to T cell exhaustion. The chronic LCMV model has revealed an inhibitory role for Tregs by maintaining CD8 T cell exhaustion. Depletion of Tregs at day 0 or 45 after infection was shown to increase the expansion of exhausted T cells (223, 224). Tregs were shown to play a beneficial role in limiting disease early after infection by preventing LCMVinduced mortality (224). However, Tregs were shown to promote the development of chronic infection when Tregs were experimentally expanded by IL-2 immune complexes (using monoclonal antibody clone JES6-1) (224). Importantly, a combined Treg depletion and PD-L1 blockade given therapeutically resulted in reduced viral titers, and to a greater extent than PD-L1 blockade alone (223). This shows that Tregs contribute to suppression of the CD8 T cell response during chronic viral infection. 28

45 NK cells have also been shown to both activate the immune response and contribute to the induction of CD8 T cell exhaustion during chronic infection. Increased frequencies of activated NK cells have been associated with chronic viral infection in both mice (225) and humans (226, 227). However, their beneficial versus detrimental role in HIV and HCV infection is unclear. NK cells can provide innate defense against viral infections by producing effector cytokines and directly killing virus-infected cells. In addition, both HIV and HCV have been shown to directly interact with receptors on NK cells and alter their function (226, 228). The resulting dysfunctional NK cells have been associated with reduced clearance and control of both HIV (229, 230) and HCV infections (228, 231). Alternatively, NK cells can produce immunomodulatory IL-10 (232) and directly kill effector T cells (233, 234) during chronic viral infection in mice. NK cell depletion in mice resulted in enhanced CD8 T cell effector function and enhanced viral clearance (225, 233, 234). NK cells enhanced CD8 T cell exhaustion by killing activated CD4 T cells which are required to limit CD8 T cell effector function (234). Together these studies show that NK cells have both beneficial and detrimental roles during chronic viral infection. Thesis Objectives Immune regulation is necessary for the induction of a proper immune response. The host immune response is modulated by multiple overlapping regulatory mechanisms. Tregs exist specifically to inhibit highly activated cells and to produce regulatory cytokines. T cells also exhibit self-regulatory mechanisms such as the induction of T cell exhaustion in the presence of excessive antigen and inflammation. Manipulation of these 29

46 regulatory mechanisms may contribute to more rapid pathogen clearance and reduced disease. Thus, my thesis aims to address the following objectives: 1. Investigate the use of DEpletion of REGulatory T cells (DEREG) mice as a tool to study Tregs during viral infections. 2. Determine the role of Tregs during RSV infection in infants. 3. Determine the contribution of MHC heterogeneity and host genetic background on CD8 T cell exhaustion and the development of chronic viral infection. 30

47 CHAPTER II LIMITATIONS OF FOXP3 + TREG DEPLETION FOLLOWING VIRAL INFECTION IN DEREG MICE Abstract Regulatory T cells play a critical role in maintaining tissue homeostasis and preventing the development of immunopathology. DEREG mice express a DTRenhanced green fluorescent protein (egfp) transgene under the control of the Foxp3 promoter allowing for Treg depletion following DT administration. DEREG mice have been utilized to investigate the role of Tregs in a wide range of disease settings. Administration of DT to naïve DEREG mice resulted in the rapid depletion of Foxp3 + Tregs from the peripheral blood. However, by day 4 post-dt administration, a GFP - Foxp3 + Treg population emerged that lacked expression of the DTR transgene and was resistant to further depletion by additional DT treatment. I further evaluated the impact of Treg depletion during both acute and chronic viral infections. Similar to naïve mice, Treg numbers rapidly rebounded during an inflammatory setting following an acute viral infection. DT treatment of both WT and DEREG mice following both acute and chronic viral infections induced exacerbated disease as compared to PBS-treated controls. Furthermore, following a chronic systemic viral infection, DT treatment resulted in nearly 100 percent mortality in both WT and DEREG mice while the PBS-treated controls survived. My results demonstrate that Treg depletion in DEREG mice is transient and that DT administration can have adverse effects during virus-induced 31

48 inflammation and highlights the critical need to include DT-treated WT mice when using DTR models to control for DT-mediated toxicity. Introduction Tregs represent a subset of CD4 + T cells defined by expression of the transcription factor, Foxp3, which is encoded by the X-linked gene Foxp3 (235, 236). Deficiencies in Foxp3 function result in autoimmunity very early in life in both mice and humans due to a lack of proper immune regulation (237, 238). In addition, Tregs have been shown to limit pathogen-specific T cell responses during both acute (143, 146, 239, 240) and chronic infections (224, 241). The vital role of Tregs in immune homeostasis has led to the development of novel tools to further dissect the functions of Tregs. DEREG mice were created to allow for the specific depletion of Foxp3 expressing Tregs (242, 243). DEREG mice express a DTR-eGFP transgene controlled by a Foxp3 promoter that is inserted via a bacterial artificial chromosome (BAC). Following DT treatment, Foxp3 expressing cells are specifically depleted. These mice have been utilized to study a wide range of immune disorders including cancer ( ), allergy ( ), and various infections (136, 146, 224). My studies were aimed at gaining a better understanding of the impact of Treg depletion on the adaptive immune response following an acute respiratory infection as well as during a chronic systemic infection. Following DT treatment of naïve DEREG mice, Treg depletion was transient with the frequency of Foxp3 + Tregs returning to WT 32

49 levels within 6 days. In addition, the newly emerged Foxp3 + Tregs no longer expressed the DTR-eGFP transgene, thus preventing prolonged depletion. Similar results were observed in mice undergoing either an acute or chronic viral infection. Furthermore, DT treatment in both transgenic DEREG mice as well as WT mice resulted in enhanced morbidity and mortality. These findings limit the potential versatility of the DEREG mouse model and identify adverse effects of DT treatment on WT mice during inflammation associated with an active viral infection. Material and Methods Mice and infection BALB/c and C57BL/6 female mice were bred in the University of Iowa animal facility and were utilized between 6-8 weeks of age. BALB/c and C57BL/6 DEREG female mice expressing DTR-eGFP under the control of the Foxp3 promoter were obtained from Dr. Katharina Lahl and Dr. Tim Sparwasser (Twincore, Hanover, Germany). The A2 strain of RSV was a gift from Dr. Barney Graham (National Institutes of Health; Bethesda, MD) and was propagated on HEp-2 cells (American Type Culture Collections; ATCC, Manassas, VA). The clone-13 strain of LCMV was a gift from Dr. Raymond Welsh (University of Massachusetts Medical School, Worcester, MA) and was propagated in baby hamster kidney (BHK)-21 cells (ATCC). For RSV infections, mice were anesthetized with isoflurane and infected intranasally with 2-3 x 10 6 plaque forming units (PFU). For LCMV C-13 infection, mice were infected i.v. with 1.8 x 10 6 PFU. Disease was assessed by monitoring weight loss and pulmonary function using an unrestrained whole-body plethysmograph (Buxco Electronics, Wilmington, NC). Penh 33

50 values were recorded for 5 min per mouse to obtain an average Penh value. All experimental procedures involving mice were approved by the University of Iowa Animal Care and Use Committee. DT treatment DEREG and WT controls were administered 1 µg DT (Lot # D and D Calbiochem, San Diego, CA; Lot# SLBC6372V Sigma-Aldrich, St. Louis, MO) i.p. diluted in endotoxin-free PBS on the indicated days. Blood collection and staining Peripheral blood was collected at indicated time points and red blood cells were lysed using NH 4 Cl. Cells were fixed and permeabilized over night using the mouse regulatory T cell staining buffer kit (ebioscience, San Diego, CA). Cells were stained with antibodies against Foxp3 (clone FJK-16s; ebioscience), CD4 (clone RM4-5; Biolegend, San Diego, CA), CD90.2 (clone ; ebioscience). After staining cells were resuspended in FACS buffer prior to analysis on a BD LSRFortessa. Data were analyzed using FlowJo software (Tree Star, Ashland, OR). Statistical analysis Graphical and statistical analysis was performed using Prism software (Graphpad Software Inc., San Diego, CA), with error bars representing the standard error of the mean (SEM). Two-way analysis of variance (ANOVA)s with a Dunnett s post test were calculated to determine the statistical significance. P values were considered significant when p<

51 Results Foxp3 + Treg depletion in DEREG mice is transient To determine the longevity of Foxp3 + Treg depletion, mice were treated with 1 µg DT for two consecutive days to deplete Tregs as previously reported (242). The frequencies of Foxp3 + and GFP + CD4 T cells in the blood were monitored every other day for 6 days (Figure 1). Less than 1 percent of CD4 T cells expressed Foxp3 on day 2 after 2 days of DT treatment. However, by day 6 post-treatment, the frequency of Foxp3 + Tregs in the peripheral blood lymphocytes (PBL) had returned to levels similar to the controls, indicating that Foxp3 + Treg depletion was transient (Figure 1A and C). In addition, egfp expression, a marker for the continued expression of the DTR-eGFP transgene, was observed on a very low frequency of Tregs appearing on days 4 and 6 post-treatment (Figure 1B and C). The lack of egfp expression suggests that the newly emerging Foxp3 + Tregs no longer express the DTR-eGFP transgene. To verify that the newly emerging egfp negative Foxp3 + Tregs were resistant to further depletion, a third DT treatment was administered on day 3 post-initial treatment (Figure 2). This additional DT treatment did not prevent the rapid return of Foxp3 + Tregs in the PBL by day 4 (Figure 2A and C). Furthermore, the majority of Tregs appearing on days 2-8 post-treatment did not express egfp (Figure 2B and C). These data indicate that additional DT treatment on day 3 does not alter the longevity of Foxp3 + Treg depletion in DEREG mice. Thus, Treg depletion cannot be maintained in naïve DEREG mice due to the emergence of a Treg population that no longer expresses the DTR-eGFP transgene. 35

52 DT administration induces disease in both DEREG and WT naïve mice To determine the physiological impact of DT treatment and transient Foxp3 + Treg depletion, weight loss was monitored daily following DT administration (Figure 3). DTtreated DEREG mice expressed significant (p<0.01) weight loss compared to the PBS- DEREG and DT-WT controls (Figure 3A). In addition, a third treatment of DT on day 4 post-initial treatment induced significant weight loss in both DT-treated DEREG (p<0.001) and WT (p<0.01) mice (Figure 3B). Thus, DT administration induces toxicity in both DEREG and WT naïve mice. Foxp3 + Treg depletion is not maintained following acute RSV infection Respiratory pathogens such as RSV have been shown to induce acute inflammation in the lung resulting in respiratory disease (143, 147). I sought to determine if virus-induced inflammation would alter the kinetics of Treg depletion. Mice were treated with 1 µg DT on days 3 and 4 post-rsv infection (Figure 4). My lab previously reported that the total number of Tregs in the lung begins to increase by day 4 post-rsv infection (143), thus Tregs were depleted prior to this expansion. Similar to naïve DEREG mice, administration of 2 consecutive injections of DT resulted in effective Treg depletion in RSV-infected mice (Figure 4A and C). Similar results were observed using DT from 2 different sources (ie. DT1 and DT2). However, by day 9 post-infection (6 days post-treatment), Foxp3 + Treg frequencies in the PBL had returned to control levels. In addition, the returning Foxp3 + Tregs lacked egfp expression, similar to naïve DTtreated DEREG mice (Figure 4B and C). Thus, similar to naïve DEREG mice, Treg depletion is not maintained following acute RSV infection. 36

53 DT treatment enhances the severity of RSV-induced disease Given the enhanced weight loss observed in naïve DEREG and WT mice following DT administration, I sought to examine the impact of DT treatment and Treg depletion on RSV-induced disease severity. Weight loss and Penh, as a measure of pulmonary function, were monitored daily in DEREG and WT mice administered DT on days 3 and 4 post-infection (Figure 5). Treatment with DT1 (Calbiochem) induced significantly enhanced disease as measured by increased Penh compared to both PBStreated and WT controls (Figure 5A). However, administration of either DT1 or DT2 (Sigma-Aldrich) induced increased weight loss in both WT and DEREG mice following RSV infection (Figure 5A and B). Slight variations can occur in DT batches from various vendors due to preparation differences affecting DT activity and purity. However, with both vendors tested DT-treated WT mice exhibited enhanced disease compared to the PBS-DEREG controls. Thus, depletion of Tregs in DEREG mice further enhanced disease. However, DT-treated WT mice also exhibited increased morbidity following acute RSV infection. DT treatment induces morbidity and mortality following LCMV Cl-13 infection To determine the efficacy of the DEREG model during a chronic viral infection, the impact of DT treatment during chronic LCMV Cl-13 infection was examined. WT and DEREG mice were treated with DT on days 10 and 11 following Cl-13 infection. Recent work has indicated that the number of Tregs increases after day 10 post-infection (250), thus day 10 was chosen for depletion. Weight loss and survival were monitored daily following infection (Figure 6). Both DT-treated WT and DEREG mice exhibited 37

54 increased morbidity compared to PBS-treated DEREG controls (Figure 6). In addition, DT treatment induced 95% mortality in WT mice and 100% mortality in DEREG mice following LCMV Cl-13 infection while PBS-treated DEREG controls survived (Figure 6). Similar results were observed with two vendors of DT (DT1 Calbiochem and DT2 Sigma-Aldrich) inducing similar weight loss and mortality (Figure 6A and B). These data indicate a role for DT-induced toxicity during a chronic viral infection in WT mice as well as DEREG mice. Discussion This study examined the kinetics of Treg depletion in naïve and virus infected DEREG mice. Following two consecutive days of DT treatment, Tregs were 95% depleted in both naïve and RSV-infected mice. However, by day 4 post-depletion approximately 50% of the Foxp3 + Treg population had returned in the PBL (Figure 1 and Figure 4). Furthermore, by day 6 post-treatment the frequency of Foxp3 + Tregs in DTtreated DEREG mice had returned to levels similar to controls. In addition, the emerging Treg population following depletion no longer expressed the DTR-eGFP transgene and was resistant to further depletion with additional DT treatment (Figure 2). The emergence of an egfp - population in DEREG mice following DT administration has previously been noted in the literature (224, 242, 251). However, to my knowledge, this is the first report detailing the kinetics of the emerging egfp - Treg population in DEREG mice following depletion. It is unclear what causes the emergence of this egfp - Foxp3 + population, one hypothesis is that the BAC transgene has delayed expression in newly generated Tregs (242), however this has not been formally proven. Alternatively, the 38

55 egfp - Foxp3 + population may be an outgrowth of the small population of non-transgenic cells. My data indicate that Treg depletion in DEREG mice is transient, preventing its use to evaluate the impact of prolonged Treg depletion. Moreover, the transient nature of Treg depletion must be taken into account when interpreting results using DEREG mice. I observed exacerbated morbidity in naïve and virally infected DEREG and WT mice following DT administration. A study by zu Hörste and colleagues reported similar morbidity in C57BL/6 mice following DT treatment from multiple vendors (252). However the mechanism by which DT induces disease in WT mice remains unclear. Mice have been shown to be less susceptible to DT-induced cell death but they are not completely resistant (253, 254). A recent study found that DT administration to naïve mice induced proteinuria and podocyte injury, indicating a need for caution when utilizing DTR transgenic models to study either hypertension or renal disease (255). These studies, in addition to my study, indicate a need for caution when utilizing DTR transgenic systems due to DT-mediated toxicity. In this study I observed that DT treatment enhanced morbidity following both acute and chronic viral infections. Following RSV infection, DT-treated DEREG mice exhibited the greatest signs of disease as measured by weight loss and Penh (Figure 5). These data are consistent with a role for Tregs in limiting RSV-induced disease as has been previously reported with anti-cd25 depletion and DT depletion of DEREG mice (143, 146). However, the intermediate disease exhibited by DT-treated WT mice indicates that not all of the disease measured can be attributed to the elimination of Tregs. A recent study by Loebbermann et al. utilized DEREG mice to examine the role of Tregs 39

56 following RSV infection (146). While they found significant differences following RSV infection between untreated WT and DT-treated DEREG mice, DT-treated WT controls were not included in the study. Further analysis of Tregs following RSV infection using Foxp3 DTR mice (256) has recently been reported, however DT-treated WT mice were not included as controls (35). Thus, it is currently unclear to what extent the enhanced disease severity observed following Treg depletion during acute RSV infection is caused by elimination of the Tregs versus the toxicity induced by DT administration. I observed enhanced mortality following chronic LCMV Cl-13 infection and DT treatment in both DEREG and WT mice while PBS-treated mice survive (Figure 6). However, two recent studies examining DT-mediated Treg depletion following chronic LCMV infection did not report DT-mediated morbidity in WT mice (223, 224). This difference in morbidity may be due to viral strain differences as Schmitz et al. utilized the chronic LCMV-docile strain which only induced mortality in DEREG mice (224). This difference in mortality may also be due to earlier DT administration and the absence of tracking past day 15 post-infection by the Schmitz et al. group (224). An additional study by Penaloza-MacMaster et al. examined DT-mediated depletion during Cl-13 infection but did not include DT-treated WT control mice. These data further highlight the importance of including DT-treated WT control mice when utilizing DTR transgenic systems. My results demonstrate that following prolonged inflammation induced by a chronic viral infection such as Cl-13, DT-induced disease is indistinguishable from disease resulting from a lack of Tregs. Thus, DT-treated WT mice are a critical control 40

57 necessary to distinguish the disease contribution caused by cellular depletion from DTinduced toxicity. DEREG mice have been utilized to extensively study Treg functions in a number of disease settings. However, my results highlight the limitations of this model to examine Treg responses during a viral infection. The limited number of days in which Tregs are depleted restricts the versatility of the DEREG model and complicates the interpretations of the data at later time points. In addition, the DT-induced morbidity and mortality in infected WT mice introduces the additional complication of DT-induced toxicity. While DT-induced mortality in WT mice may only be observed in a highly inflammatory environment such as during chronic LCMV Cl-13 infection, further studies will be necessary to determine the full extent of this toxic side effect. My results highlight a complication in all DTR models that must be carefully considered when utilizing these model systems. 41

58 Figure 1. Foxp3+ Treg depletion is not maintained in naïve DEREG mice. Naïve DEREG and WT mice were treated with either 1 µg DT or PBS i.p. on days 0 and 1. The frequency of Foxp3 + (A) CD4 T cells were monitored in the peripheral blood. The frequency of egfp + expression within the Foxp3 + Treg population was also monitored (B). Gated Thy1.2 + CD4 + cells depicting Foxp3 expression (x-axis) and egfp expression (y-axis) are shown (C). Plots are concatenated from one of three independent experiments (n = 4-5). Data in A and B depict cumulative data from two to three independent experiments (n = 8-12). Statistical significance was determined by a two-way ANOVA with a Dunnett s post-test comparing DT-DEREG to PBS-DEREG (B) or to both DT-WT and PBS-DEREG (A). ***, p <

59 Figure 2. Additional DT treatment does not extend the duration of Foxp3 + Treg depletion. Naïve DEREG and WT mice were treated with either 1 µg DT or PBS i.p. on days 0, 1 and 3. The frequency of Foxp3 + (A) and egfp + Foxp3 + (B) CD4 T cell frequencies were monitored in the peripheral blood. Gated Thy1.2 + CD4 + cells depicting Foxp3 expression (x-axis) and egfp expression (y-axis) are shown (C). Plots are concatenated from one of two independent experiments (n = 3-4). Data in A and B depict cumulative data from two independent experiments (n = 7-8). Statistical significance was determined by a two-way ANOVA with a Dunnett s post-test comparing DT-DEREG to PBS-DEREG (B) or to both DT-WT and PBS-DEREG (A). **, p < 0.01; ***, p <

60 Figure 3. DT treatment induces disease in naïve WT and DEREG mice. Naïve DEREG and WT mice were treated with either 1 µg DT or PBS i.p. on days 0 and 1 (A) or days 0,1, and 3 (B). Weight loss was monitored daily following treatment. Data in A and B depict cumulative data from 2-3 independent experiments (n = 7-12). Statistical significance was determined by a two-way ANOVA with a Dunnett s post-test comparing PBS-DEREG to DT-DEREG (*) or DT-WT (+). *, p < 0.05; **, p < 0.01; ***, p < 0.001; ++, p <

61 Figure 4. Foxp3 + Treg depletion is not maintained following RSV infection. Following i.n. RSV infection DEREG and WT mice were treated with either 1 µg DT (Calbiochem DT 1; Sigma Aldrich DT 2) or PBS i.p. on days 3 and 4 post-infection. The frequency of Foxp3 + (A) and egfp + Foxp3 + (B) CD4 T cells were monitored in the peripheral blood. Gated Thy1.2 + CD4 + cells depicting Foxp3 expression (x-axis) and GFP expression (y-axis) are shown (C). Plots are concatenated from one of two independent experiments (n = 3-4). Data in A and B depict cumulative data from two independent experiments (n = 7-8). Statistical significance was determined by a two-way ANOVA with a Dunnett s post-test comparing both DT1- and DT2-DEREG to PBS-DEREG (B) or to DT1/2-WT and PBS-DEREG (A). ***, p <

62 Figure 5. DT treatment enhances disease in both WT and DEREG mice following RSV infection. Following i.n. RSV infection DEREG and WT mice were treated with either 1 µg DT1 from Calbiochem (A), DT2 from Sigma-Aldrich (B) or PBS i.p. on days 3 and 4 postinfection. Weight loss was monitored daily following infection. Data in A and B depict cumulative data from 2-5 independent experiments (n = 8-22). Statistical significance was determined by a two-way ANOVA with a Dunnett s post-test comparing each group to PBS-DEREG. Data from (A) and (B) were collected simultaneously and PBS-DEREG values are repeated on each graph. *, p < 0.05; ***, p <

63 Figure 6. DT treatment enhances disease and mortality in WT and DEREG mice following Cl-13 infection. Following i.v. Cl-13 infection DEREG and WT mice were treated with either 1 µg DT1 from Calbiochem (A), DT2 from Sigma-Aldrich (B) or PBS i.p. on days 10 and 11 postinfection. Weight loss and survival were monitored daily following infection. Data in A and B depict cumulative data from 3 independent experiments (n = 3-11). Survival statistics were determined by a log rank Mantel-Cox test. Data from (A) and (B) were collected simultaneously and PBS-DEREG values are repeated on each graph. *, p <

64 CHAPTER III ALTERED TREG AND CYTOKINE RESPONSES IN RSV-INFECTED INFANTS Abstract Respiratory syncytial virus is the leading cause of bronchiolitis and pneumonia in children under one year of age in the United States. The host immune response is believed to contribute to RSV-induced disease. I hypothesize that severe RSV infection in infants is mediated in part through a lack of sufficient Tregs resulting in immunopathology. Blood and nasal aspirates from 23 RSV-infected and 17 control infants under 1 year of age were collected. Treg frequencies were determined by flow cytometry from peripheral blood mononuclear cells. Analysis of 24 cytokines was measured by multiplex assay on nasal aspirates. I demonstrate that the frequency of activated Tregs is significantly reduced in the peripheral blood of RSV-infected infants compared to age-matched controls. In addition, analysis of cytokine levels in nasal secretions revealed increased amounts of a wide range of both Th1- and Th2-associated cytokines compared to controls. Of particular note, the amount of IL-33 protein, a cytokine important in maintaining Treg homeostasis in mucosal tissues, was decreased in RSV-infected children. These results suggest that reduced production of IL-33 following RSV infection may lead to decreased Treg numbers and an inability to properly control the host inflammatory response. 48

65 Introduction Respiratory syncytial virus is the leading cause of lower respiratory tract infections in young children worldwide (17). While a number of risk factors have been identified that increase the chance of severe RSV infection, the majority of RSV hospitalizations occur in previously healthy children (257). Currently, the only approved RSV-specific therapy is prophylactic treatment with palivizumab (258, 259). However, preventative treatment with palivizumab is restricted to high-risk infants and therefore there is currently no method to prevent severe RSV-induced disease in otherwise healthy children (260). The development of an RSV vaccine is needed to reduce the global disease burden and mortality observed from RSV infection. Thus, an increased understanding of the host immune response to severe RSV infection in infants would help facilitate the development of an effective vaccine and/or therapeutic intervention. The host immune response is thought to contribute to the severity of RSV-induced disease in both humans (261) and animal models (21, 262). Specifically CD4 T cells have been implicated in contributing to RSV-induced disease in mice (21). CD4 T cell-induced disease is believed to be mediated by an improper skewing of CD4 T cell subsets (263, 264). Th1 polarized CD4 T cells that produce IFN-γ and TNF contribute to the clearance of viral infections, however, RSV infection often induces both Th2 and Th17 responses (33, 37, 264). Thus, regulation and induction of a proper immune response is vital to prevent RSV-induced disease (147). Tregs have been shown to play a critical role in modulating disease in RSV-infected mice (143, 146, 240, 265). Tregs have been shown to limit lung inflammation and immune-mediated pathology in RSV-infected mice (143, 49

66 146, 240). However, the role of Tregs during RSV infection in humans is currently unclear. In this study I observe a reduction in the frequency of activated Tregs (atregs) in the peripheral blood of RSV-infected infants compared to controls. I also find that resting Treg (rtreg) frequencies decrease as RSV disease progresses. To understand the relationship between Treg expression and respiratory cytokine levels, I also measured a panel of cytokines in the nasal secretions of each infant. I found that RSV-infected infants display an increased level of a wide range of pro-inflammatory cytokines. Importantly, I observed a reduction in IL-33, a cytokine required for Treg maintenance within mucosal tissues. Taken together my results demonstrate a reduction in Treg frequencies and an increase in inflammatory cytokines in the airways in RSV-infected infants indicating that a failure to properly regulate the host immune response may contribute to the severity of RSV-induced disease. Materials and Methods Study population The study population was comprised of infants under 1 year of age hospitalized at the University of Iowa Hospitals and Clinics during two RSV seasons between January 2013 and June 2014 that presented with suspected viral respiratory illness. Viral infection was confirmed by PCR of nasopharyngeal secretions. Infants less than 1 year of age admitted to the hospital without respiratory signs were included as controls. Demographic information and family history were obtained from the parents. Physical examination and 50

67 treatment information were obtained from medical records and the primary medical team. Children were excluded if they had a history of immunodeficiency, congenital cardiac disease, hemodynamic instability, deemed at risk of death, or lacked parental consent. Patient consent and samples were collected with the help of the UIHC research nurses, Dr. Sameer Kamath, and Dr. Muhammad Syed. The study was approved by the University of Iowa Human Subjects Office (IRB ) and written consent was obtained from all patients guardians. Nasal fluid cytokine determination Nasal washes and tracheal aspirates were treated with an equal volume of 0.1% Triton X-100 PBS, 1ul/ml protease inhibitor cocktail (P8340; Sigma-Aldrich, St. Louis, MO) and frozen at -80 C. Cytokine concentrations were measured using a multiplex assay per the manufacturer s instructions (EMD Millipore, Billerica, MA). Peripheral blood mononuclear cell isolation PBMCs were isolated from 2 ml of blood collected in 4ml Na Heparin green top tubes (BD, Franklin Lakes, NJ). PBMCs were isolated using Ficoll-Paque Plus (GE Healthcare Life Sciences, Pittsburgh, PA). Cells were resuspended at 2 x 10 6 cells/ml in 10% dimethyl sulfoxide fetal calf serum (FCS) and frozen at -80 C followed by longterm storage in liquid nitrogen. Flow cytometry PBMCs were thawed and stained with anti-cd25 (BC96) and anti-cd45ra (HI100) extracellularly. Cells were fixed using Foxp3 fixation buffer and stained in 51

68 permeabilization buffer with antibodies specific to CD3 (OKT3), CD4 (SK3), Foxp3 (236A/E7; all antibodies were obtained from ebioscience, San Diego, CA) and Ki-67 (B65, BD Bioscience, San Jose, CA). Samples were run on a BD LSR Fortessa (BD Biosciences) and the data were analyzed using FlowJo software (Tree Star Inc, Ashland, OR). Statistical analysis Statistical analyses were performed using GraphPad Prism software (La Jolla, CA) to compare T cell frequencies between control and RSV patients. Groups were compared using the Mann-Whitney test for unpaired samples. Additional statistical analyses were performed using SAS (SAS Institute Inc, Cary, NC) by Dr. Patrick Ten Eyck. Pearson correlations were calculated to determine associations between T cell populations, cytokine concentrations and patient data (disease level, days post-symptoms, and age). A linear modeling framework was used to determine associations between predictors (patient survey information an physical exam results) and outcomes (T cell populations and cytokine concentration). Results Characteristics of the study population In this study, 40 infants presenting with respiratory infection were enrolled. PCR and cultures of respiratory secretions determined that 25 (62.5%) of the infants were infected with RSV. Two of the RSV-infected infants exhibited co-infections, one with adenovirus and the other with influenza. The remaining infants were infected with human 52

69 metapneumovirus (10%), adenovirus (7.5%), parainfluenza virus (5%), influenza (2.5%) or bacteria (12.5%). As expected (266, 267), RSV was the most prevalent pathogen detected in infants hospitalized with respiratory infections. Control samples were collected from infants admitted to the hospital for elective surgery with no report of respiratory illness. Sex and gestational age were not found to be significantly different between the controls and RSV-infected infants (Table 1). However, infant age at sample collection was significantly lower in controls compared to RSVinfected infants (Table 1). Patient history showed no difference between the controls and RSV-infected infants with regard to a family history of asthma, second hand exposure to tobacco smoke, previous respiratory infections, infants with siblings or household pets (Table 1). In addition, a small portion of infants from both controls and the RSV-infected infants were administered palivizumab prior to hospitalization but no difference was observed between the control and RSV-infected infants (Table 1). Overall, no critical differences were noted in the characteristics of the study populations. RSV-infected infants were administered a variety of treatments to support their severity of illness (Table 2). However, no intervention was associated with a significant change in either the frequency of T cell populations or cytokine concentrations (Table 3) Activated Tregs are reduced in RSV-infected infants Human regulatory T cells can be divided into two distinct subsets. rtregs are defined by high expression of CD45RA and intermediate expression of either Foxp3 or CD25. atregs are defined by low expression of CD45RA and high expression of either 53

70 Foxp3 or CD25 as previously reported (268). Using a similar gating strategy (Figure 7A), the frequency of rtregs in the peripheral blood did not exhibit a significant difference between RSV-infected infants and controls (Figure 7B). However, I observed a reduced frequency of atregs identified by CD45RA lo CD25 hi in the peripheral blood of RSVinfected infants compared to age-matched controls (Figure 7C). Furthermore, the frequencies of conventional naïve and memory CD4 T cell populations were not altered in RSV-infected infants (Figure 8A and B). The frequency of proliferating CD4 T cells, as measured by Ki-67 expression, did not significantly differ between RSV-infected infants and controls (Figure 9). Thus, while atreg frequencies were reduced in RSVinfected infants, the proportion of proliferating atregs was similar. Overall, RSVinfected infants displayed a reduction in the frequency of atregs in the peripheral blood, with no perturbation observed in any other CD4 T cell subset examined. Cytokine concentrations increase in RSV-infected infants To understand the relationship between Treg expression and respiratory cytokine levels I also measured a panel of 24 cytokines and chemokines. Of those cytokines analyzed, eight were observed at significantly greater levels in nasal excretions of RSVinfected infants compared to controls (Table 4). The Th1-associated cytokines IFN-γ and TNF were significantly increased in RSV-infected infants (Table 4). The Th2-associated cytokine IL-4 was also significantly increased in RSV-infected infants, however, the Th2-associated cytokine IL-5 was undetectable and IL-13 showed no significant difference between groups (Table 4). Furthermore, the pro-inflammatory and pyrogenic cytokines IL-1α and IL-6 were also increased in RSV-infected infants. IL-1β exhibited 54

71 the greatest increase in RSV-infected infants indicating inflammasome activation (Table 4). Only IL-33, a danger signal induced in response to tissue damage (269), was notably reduced in RSV-infected infants (Table 4). Six cytokines and chemokines, including IL- 5, IL-9, IL-17F, IL-17E, Eotaxin, and TSLP were undetectable in both RSV-infected infants and controls. Together, my data demonstrate an increase in Th1, Th2 and proinflammatory cytokines while IL-33 is decreased in RSV-infected infants compared to controls. Resting Treg frequencies decrease with age rtreg cells have been shown to be highly prevalent in cord blood and found to decrease with age (268). I wanted to confirm that my rtreg populations followed a similar pattern. Newborn infants from both controls and those with RSV infection exhibited the highest frequencies of CD45RA hi Foxp3 int rtregs in PBMCs. With increasing age, both control and RSV-infected infants displayed reduced frequencies of rtregs in their blood (Figure 10). An inverse correlation between rtreg frequencies and the age of the infant was found to be statistically significant (Figure 10; p = ). These data confirmed the previous findings of rtreg frequencies decreasing in the blood after birth (268). Resting Treg frequencies decrease as RSV diseases progresses Upon TCR activation, rtregs upregulate Foxp3 expression and convert to atregs (268). Thus, if Tregs are activated following RSV infection, I would expect to observe a decrease in rtregs and an increase in atregs as RSV infection progressed. Therefore, I examined the change in the rtreg and atreg population over the course of an acute 55

72 respiratory infection in infants. A significant decline in the rtreg frequency was observed in the blood of RSV-infected infants as the days post-disease onset increased (Figure 11). Thus, rtreg prevalence in the blood decreased as RSV disease progressed. However, I did not observe a subsequent increase in atregs. This suggests that while the decline in rtregs may be due to conversion to atregs, the atregs were either migrating from the blood or undergoing apoptosis during RSV infection. Th17-associated cytokines are associated with reduced respiratory signs and symptoms The impact of IL-17 on disease severity following RSV infection has been controversial. Two previous studies have found infants with moderate RSV infection expressed more IL-17 than those with severe infection (270, 271). However, IL-17 was found to be pathogenic in the murine model of RSV resulting in enhanced mucus production and neutrophilic infiltrate (37). Here I found that Th17-associated cytokines IL-1β, IL-17A and IL-23 were all elevated when either difficulty breathing or retraction was not reported by the infant s doctor following examination (Figure 12). Moreover, no cytokines were associated with any medical treatment that the children received (Table 3). Thus, Th17-associated cytokines may not be detrimental in the upper airways during infant RSV infection. 56

73 Discussion Respiratory syncytial virus infection has been shown to alter the inflammatory milieu of infected infants (25, 264), however, regulation of this inflammatory response has not been well studied. Here I examined both the cytokine milieu and the expression of regulatory and conventional CD4 T cells in infants hospitalized with RSV infection compared to uninfected controls. I found that RSV infection in infants induces a significant reduction in atregs but not rtregs compared to controls. These findings are similar to a recent study performed by Raiden et al. in Argentina (22). However, Raiden et al. also observed a reduction in all Foxp3 expressing CD4 T cells irrespective of subset (22). The differences in responsiveness of the rtreg and activated CD4 subsets observed between the studies may be due to differences in circulating RSV strains or in the timing of sample collection after infection. However, in agreement with my results, Raiden et al. also observed a reduction in the frequency of atregs in the blood that may be the result of atreg migration from the blood to the site of the infection in the lung. Alternatively, atreg numbers may be reduced in all peripheral tissues following RSV infection. I also observe that rtreg frequencies in the peripheral blood are reduced as RSV disease progresses. rtregs are believed to represent a long-lived Treg population that upon activation proliferate and convert to atregs (268). Thus, a reduction in rtregs should result in an increase in atregs, however, I did not observe this in my cohort of RSVinfected infants. Consequently, RSV infection appears to drive Treg activation whereupon Tregs either leave the blood to enter the lung or undergo apoptosis. Further understanding the fate of Tregs is vital to understanding the regulation of the RSV 57

74 immune response in infants. This reduction of Tregs in the blood is not observed in the murine model of RSV (143). In mice, Treg frequencies show a slight increase in the blood and are greatly expanded in the lung following severe RSV infection (143). Tregs have been shown to be vital in limiting the lung immunopathology observed in mice following RSV infection by regulating T cell recruitment, activation and cytokine production (35, 143, 146, 240). It appears that Tregs do not follow a similar distribution pattern in RSV-infected infants and thus may not be present in the lung in sufficient numbers to properly regulate the host immune response. To understand the correlation between Treg subset distribution and respiratory cytokine levels I also measured a panel of cytokines in the same RSV-infected infants and controls. I found that RSV-infected infants exhibited an increase in both Th1- and Th2-associated cytokines, though both are expressed at very low levels. A previous study with nasal samples from over 800 RSV-infected children under 5 years of age, also observed an increase in the Th1-associated cytokine TNF but did not observe any detectable levels of Th2 cytokines (30), however, their assay may have had a higher limit of detection. Nevertheless, it is difficult to determine if the low levels of the Th1- and Th2-associated cytokines I observed are sufficient to impact RSV disease severity. Furthermore, Tregs are known to control both Th1- and Th2-associated cytokine responses in mice, however, no correlation was observed between the level of these cytokines and the frequency of either atregs or rtregs within each individual (Table 5). While the majority of cytokines measured were increased following RSV infection, IL-33 was found to be reduced in RSV-infected infants compared to controls. 58

75 This is in contrast to recent studies in which IL-33 was observed to increase in mice and human infants following RSV infection ( ). However, human RSV strain differences and severity of disease may explain the observed differences in IL-33 levels in infants. The RSV infected infants in my study exhibited severe RSV infection requiring hospitalization and intensive care whereas the infants enrolled in the study by Saravia et al. only required clinical presentation of RSV symptoms (272), thus increased disease severity may result in reduced IL-33 levels. IL-33 is believed to play an important role in RSV infection through its role as an alarmin that serves as a warning of tissue damage to the host ( ), and thus may play an important role in activating the immune response. Furthermore, IL-33 has been shown to be necessary for Treg accumulation and maintenance in the mucosal site of the intestines (279). IL-33 may play a similar role for Treg recruitment and maintenance in other mucosal sites such as the lung. Importantly, my observed reduction in IL-33 levels may explain the reduction in the frequency of atregs observed in the RSV-infected infants. Furthermore, the reduction in IL-33 from basal levels may indicate apoptotic rather than necrotic death of endothelial cells which has been shown to reduce IL-33 levels (280). The observed reduction in IL- 33 may play an important role in explaining the severity of disease observed in these RSV-infected infants. However, my observations were made in two separate compartments, with IL-33 in the nasal washes and Tregs in the blood. Further, studies of Tregs in human airways will help to determine the connection between IL-33 and Tregs during RSV infection. 59

76 While Th17-associated cytokines were not found to change between RSVinfected infants and controls, specific symptoms observed within the RSV-infected infants group were associated with the presence of Th17-associated cytokines. I observed a reduction in Th17-associated cytokines but not in either Th1- or Th2-cytokines when difficulty breathing and retraction were observed. This indicates that Th17-associated cytokines may be beneficial in preventing specific disease symptoms within the RSVinfected infants. Two previous studies had similar findings where infants with moderate RSV infection expressed more IL-17 than those with severe infection (270, 271). However, IL-17 was found to be pathogenic in the murine model of RSV resulting in enhanced mucus production and neutrophilic infiltrate (37). While they found that IL-17- mediated enhanced disease, they also observed enhanced viral clearance in RSV-infected mice (37). IL-17 is present during RSV infection in both mice and humans, however, IL- 17 may play a protective role in human RSV infection. My results indicate that atregs decrease in the peripheral blood following severe RSV infection in infants. This reduced regulation is associated with increased Th1-, Th2-, and inflammasome-associated cytokines. These results together with a reduction in the alarmin cytokine IL-33 indicate a potential lack of regulation resulting in an unbalanced immune response following RSV infection in infants. Understanding how the virus modulates the host inflammatory response can help to determine what immune targets are beneficial in treatment and prevention of RSV-infection in infants. 60

77 Table 1. Demographic of study population. Control (n = 17) RSV (n = 23) p-value # Male/Female 9/8 14/ Age* (range) 1.7 ( ) 3.3 ( ) Gestational age* (range) 37.3 (30-39) 36.8 (28-39) Family history of asthma 13.3% 21.7% Household smokers 13.3% 21.7% Prev. respiratory infection 6.7% 13% Siblings 66.6% 85.7% Household pets 37.5% 13% Palivizumab 6.3% 8.7% *Average in months # p-value calculated by Fisher s exact test except for ( + ) labeled values + p-values calculated by Mann-Whitney test 61

78 Table 2. Medical intervention of RSVinfected infants. RSV (n = 23) Oxygen received 21 (91%) Use of CPAP 8 (35%) Use of heliox 0 (0%) Mechanical ventilation 5 (22%) Steroids 3 (13%) Albuterol 8 (35%) 62

79 Table 3. Medical intervention did not correlate with an altered immune response. Received Received Mechanical Received Received Received Oxygen CPAP Ventilator Palivizumab Steroids Albuterol atreg* # rtreg IL-1α IL-1β IL IL IL IL IL IL-12p IL-12p IL IL IL-17A IL IL IL TNF IFN-γ MDC Rantes *atreg and rtreg frequencies were determined by flow cytometry + Cytokine concentrations were determined by multiplex immune assay # p-values were calculated based on a linear modeling framework to determine associations between predictors (patient treatment) and outcomes (T cell populations and cytokine concentration) 63

80 Figure 7. atreg frequencies are reduced in the peripheral blood of RSV-infected infants. PBMCs from RSV-infected infants and controls were stained to identify subsets of conventional and regulatory CD4 T cells. (A) Representative plots depict CD45RA hi Foxp3 int /CD25 int resting Tregs (rtreg, I), CD45RA lo Foxp3 hi /CD25 hi activated Tregs (atreg, II), CD45RA hi Foxp3 - naïve conventional CD4 T cells (III), and CD45RA lo Foxp3 - memory conventional CD4 T cells (IV) from control and RSV-infected infants. The cumulative frequencies of resting (B) and activated (C) Tregs are shown for controls and RSV-infected infants by two staining methods. All samples are gated on CD4 + CD3 + T cells. Data represent the mean ± SEM for both controls (n = 13) and RSV (n = 17) groups. **, p <

81 Figure 8. Conventional CD4 T cell frequencies are maintained in the peripheral blood of RSV-infected infants. PBMCs from RSV-infected infants and controls were stained to identify subsets of conventional and regulatory CD4 T cells. (A) Naïve and (B) memory conventional CD4 T cell frequencies are shown from all groups. All samples are gated on CD4 + CD3 + T cells. Data represent the mean ± SEM for both controls (n = 13) and RSV (n = 17) groups. 65

82 % Ki-67 + of CD4 + T cells Control RSV Naive Memory rtreg atreg Figure 9. Proliferation of CD4 T cell subsets is not altered by RSV infection. PBMCs from RSV-infected infants and controls were stained to identify subsets of conventional and regulatory CD4 T cells. Ki-67 expression from each CD4 T cell subset defined by CD45RA and Foxp3 expression is shown for both control and RSVinfected infants. All samples are gated on CD4 + CD3 + T cells. Data represent the mean ± SEM for both controls (n = 13) and RSV (n = 17) groups. 66

83 Table 4. Cytokines and chemokines in nasopharyngeal aspirates (pg/ml). Control (n = 17) RSV (n = 23) LOD Avg Median 75 th Q 25 th Q Avg Median 75 th Q 25 th Q P Sig IL-1α IL-1β ** IL * IL IL IL IL-12p IL-12p IL IL IL-17A IL IL ** IL TNF IFN-γ MDC RANTES Avg = Average Q = Quartile LOD = Limit of detection P = p-value determined by two-tailed t-test Sig = Significance; 0.1, * 0.05, ** 0.01, and

84 Figure 10. rtregs decrease with age. The coefficients of determination CD45RA hi Foxp3 int rtregs and infant age at blood collection were calculated along with the p-value measuring significance of association. The association is shown for the combined control and RSV-infected infant data. 68

85 Figure 11. rtreg frequencies decrease as RSV disease progresses. The coefficients of determination between CD45RA hi Foxp3 int rtregs (A) or CD45RA lo Foxp3 hi atregs (B) and parent reported symptom start in RSV-infected infants were calculated along with the p-values measuring significance of association. 69

86 Figure 12. Th17-associated cytokines correlate with reduced difficulty breathing and retraction. Physical exam findings of difficulty breathing (left) and retraction (right) are associated with Th17-associated but not Th1- or Th2-associated cytokine concentrations in nasal excretions from RSV-infected infants. Data represent the mean ± SEM for both controls (n = 13) and RSV (n = 17) groups and statistics were measured by unpaired t test. *, p <

87 Table 5. Correlations between CD4 T cell populations and cytokines (p-values). Naïve* rtreg atreg Memory IL-1α IL-1β IL IL IL IL IL IL-12p IL-12p IL IL IL-17A IL IL IL TNF IFN-γ MDC RANTES p-values indicate significance of the Pearson correlation from RSV-infected individuals *CD4 T cell frequencies were determined by flow cytometry + Cytokine concentrations were determined by multiplex immune assay 71

88 IL-33 Tregs Inflammatory cytokines Increased disease severity Figure 13. Reduced immune regulation results in enhanced inflammatory cytokines and increased disease severity in RSV-infected infants. RSV-infected infants exhibit reduced IL-33 cytokine levels in nasal washes and reduced atreg frequencies in the peripheral blood compared to control infants. IL-33 helps to maintain Treg populations in mucosal sites. RSV-infected infants also exhibit increased proinflammatory cytokines in nasal washes. The lack of regulation due to a reduced Treg response may contribute to the enhanced inflammatory cytokines levels and hospitalization of RSV-infected infants. 72

89 CHAPTER IV HOST GENETICS PLAY A CRITICAL ROLE IN CONTROLLING CD8 T CELL FUNCTION AND LETHAL IMMUNOPATHOLOGY DURING CHRONIC VIRAL INFECTION Abstract Effective CD8 T cell responses are vital for the control of chronic viral infections. Many factors of the host immune response contribute to the maintenance of effector CD8 T cell responses versus CD8 T cell exhaustion during chronic infection. Specific MHC alleles and the degree of MHC heterogeneity are associated with enhanced CD8 T cell function and viral control during human chronic infection. However, it is currently unclear to what degree host genetics influences the establishment of chronic viral infection. In order to examine the impact of MHC heterogeneity versus non-mhc host genetics on the development of chronic viral infection, an F1 cross of B10.D2 x B6 (D2B6F1) and BALB.B x BALB/c (BCF1) mice were infected with the Cl-13 strain of LCMV. Following chronic Cl-13 infection both D2B6F1 and BCF1 mice (i.e. H-2 bxd ) demonstrated increased viral control compared to homozygous mice. Strikingly, H-2 bxd mice on a C57BL genetic background (i.e. D2B6F1) mice exhibited mortality following Cl-13 infection. CD8 T cell depletion prevented mortality in Cl-13-infected D2B6F1 mice indicating that mortality was CD8 T cell dependent. D2B6F1 mice maintained more CD8 T cell effector cytokine production and reduced expression of the T cell exhaustion marker PD-1. In addition, LCMV infection of D2B6F1 mice also induced a larger Th1 response than BCF1 mice and Cl-13-induced mortality in D2B6F1 mice was also 73

90 dependent on CD4 T-cell-mediated IFN-γ production. My results indicate that increasing the degree of MHC heterogeneity can have a profound impact on the breadth and effectiveness of the antiviral T cell response. However, the host genetic background can also greatly influence the maintenance of the CD8 T cell effector activity, resulting in immune-mediated mortality. Thus, following a chronic viral infection, increased functionality of the CD8 T cell response allowed for more rapid viral clearance at the cost of enhanced immunopathology dependent on both MHC diversity and the genetic background of the host. Introduction Chronic viral infection such as HIV and HCV affect an estimated 185 million people worldwide. In the course of these infections HIV and HCV are able to evade the host immune response and establish viral persistence. However, genetic host factors that influence the host T cell response have been shown to greatly impact viral load and host disease progression ( ). Expression of either heterozygous human leukocyte antigen (HLA) Class I alleles or the specific HLA alleles HLA B*5701 and HLA B*2701 are associated with enhanced CD8 T cell responses and reduced viral loads (283, ). Clearance of HCV infection has also been associated with the expression of heterozygous HLA alleles as well as with the expression of specific HLA alleles (282, 284). These host advantages driven by HLA designation demonstrate the critical importance of the T cell response during chronic viral infection. In both humans and animal models of chronic viral infection, the capacity to maintain a functional CD8 T cell response has been shown to play a critical role in limiting viral replication ( , 74

91 289). However, constant CD8 T cell stimulation during chronic viral infection can drive CD8 T cells into a state of exhaustion where antiviral CD8 T cells progressively lose their effector functions, preventing effective clearance of infection ( ). CD8 T cell exhaustion is believed to occur in order to limit immunopathology to the host. In addition to CD8 T cells, CD4 T cells also play a crucial role during the development of chronic viral infections. During HIV and murine infection with the persistent Cl-13 strain of LCMV, depletion of CD4 T cells inhibits CD8 T cell effector functions and results in reduced viral control (157, 166). Similar to CD8 T cells, CD4 T cell responses are altered by the presence of persistent antigen. Chronic viral infection drives CD4 T cell conversion from a Th1 phenotype critical to viral clearance to a Tfh phenotype (60, 172). Thus, maintaining CD4 and CD8 T cell effector functions is vital to the control of chronic viral infection. Murine models have often been used to delineate the contribution of MHC alleles versus host genetic background on mouse strain susceptibility to infection. BALB/c and B6 mice have been shown to differ in their susceptibility to infections in both MHCdependent ( ) and -independent manners ( ). However, the contribution of MHC alleles versus host genetic background has not been examined during chronic viral infection. In the present study, I sought to determine the genetic influences on CD8 T cell exhaustion and the development of persistent viral infection using MHC congenic inbred and first generation (F1) crosses of various inbred mouse strains. My results indicate that 75

92 increasing MHC diversity allows for a broader CD8 T cell response that is able to more rapidly control persistent LCMV Cl-13 infection. However, F1 mice expressing a H-2 bxd haplotype on different genetic backgrounds exhibit differences in both CD8 T cell exhaustion and CD4 T cell polarization, resulting in differences in viral control and disease. Unexpectedly, H-2 bxd mice on a C57BL background succumb to LCMV Cl-13 infection in a T-cell- and IFN-γ-dependent manner. Thus, both MHC haplotype and the remaining genetic background together influence host resistance to CD8 T cell exhaustion and disease susceptibility following LCMV Cl-13 infection. Materials and Methods Mice and infection BALB/c and C57BL/6 mice were purchased from National Cancer Institute (Frederick, MD) and bred in the University of Iowa animal facility. BALB.B and B10.D2 mice were purchased from the Jackson Laboratory (Frederick, MD) and bred in the University of Iowa animal facility. First generation BALB.B x BALB/c (BCF1) mice and B10.D2 x C57BL/6 (D2B6F1) mice were created in the University of Iowa animal facility. All mice were utilized between 6-8 weeks of age. The Arm and Cl-13 strains of LCMV were gifts from Dr. Raymond Welsh (University of Massachusetts Medical School, Worcester, MA) and propagated in BHK-21 cells (ATCC). For LCMV Arm infection, mice were infected i.p. with 2 x 10 5 PFU. For LCMV Cl-13 infection, mice were infected i.v. with 1.8 x 10 6 PFU. Disease was assessed daily by monitoring weight loss and survival. The University of Iowa Animal Care and Use Committee approved all experimental procedures involving mice. 76

93 Plaque assays Serum, lung, liver, and kidneys were harvested from Cl-13-infected mice. Serum was isolated from whole blood collected in heparinized tubes (Fisher; Pittsburgh, PA), capped (Lecia Microsystems; Richmond, IL) and spun at 2500 rpm for 30 min. Lung, liver and kidneys were homogenized and spun at 2000 rpm for 10 min. Serum and supernatants were collected and snap-frozen in liquid nitrogen prior to storage at -80 C. Plaque assays were performed on Vero cells and infected with serial dilutions of murine supernatants in serum-free MEM. Plates were incubated for 90 minutes at 37 C and overlayed with 3 ml/6-well of a 1:1 mixture of complete EMEM and 1% agarose. After a 3 day incubation at 37 C plates were stained with 2 ml/well of a 1:1 mixture of complete EMEM and 1% agarose supplemented with 170 µl/ml 1% neutral red. Plaques were counted after an additional 24 hour incubation at 37 C. Antibody depletion For T cell depletion, mice were administered 200 µg of rat IgG (MP Biomedicals, Aurora, OH) or anti-cd4 (clone GK1.5) and/or anti-cd8 (clone 2.43) i.p. on days -2, 2, 6, 10 relative to infection. Depletion was verified in the blood on day 6 post-infection. For cytokine depletion, mice were administered 200 µg of rat IgG (MP Biomedicals), anti-ifn-γ (clone XMG1.2), or anti-tnf (clone XT22) i.p. on days -1, 1, 3, 5, 7, and 9 relative to infection. Flow cytometry analysis Cells were isolated from the blood and spleen. A single-cell suspension was collected from the spleen following homogenization between two micro slides (Lecia 77

94 Biosystems). Red blood cells from both the blood and spleen were lysed in NH 4 Cl. For detection of antigen-specific cytokine responses cells were stimulated with 1µM NP , GP 33-41, GP , NP , GP , or GP (Biosynthesis, Lewisville, TX) in the presence of 10 µg/ml BFA (Sigma-Aldrich) for 5 hr at 37 C in 10% FCS-supplemented RPMI. For tetramer staining, cells were incubated with a single tetramer (NP , GP , GP 66-77, and NP from NIH Tetramer Core Facility; NP , GP 33-41, and GP made in house) for 30 min at 4 C for CD8 tetramers or 90 min at 37 C for the CD4 tetramer. Following tetramer staining or stimulation, cells were fixed using 1-step Fix/Lyse Solution (ebioscience, San Diego, CA). For staining of transcription factors, cells were fixed with Foxp3 staining buffer set (ebioscience) and stained intracellularly for T-box transcription factor (Tbet; clone 4B10). Cells were stained with the extracellular mabs specific to CD90.2 (clone ), CD8α (clone ), CD4 (clone RM4-5), CD11a (clone M17/4), CD49d (clone R1-2), and PD-1 (clone RMP1-30, all previous antibodies from ebioscience and Biolegend) for 15 min at 4 C. For CXCR5 staining, cells were stained with the mab for CXCR5 (biotin, BD Pharmingen) at room temperature for 30 min followed by streptavidin-pe (ebioscience) in FACS buffer. Cells were stained for the intracellular mabs specific to IFN-γ (XMG1.2), and TNF (clone MP6-XT22) in FACS Buffer containing 0.5% saponin (Sigma-Aldrich) for 30 min at 4 C. Samples were run on a BD LSRFortessa flow cytometer and data were analyzed using FlowJo software (Tree Star Inc, Ashland, OR). 78

95 Histology and Evans Blue staining Brains, lungs, livers and kidneys from day 7 Cl-13 infected B6, B10.D2 and D2B6F1 mice were perfused with 10% neutral buffered formalin (Fischer Scientific, Fair Lawn, NJ) via the heart and stored in 10 ml of 10% neutral buffered formalin for 48 hours. Organs were stained for hematoxylin and eosin (H&E). Tissues were examined for gross pathology in a manner masked to experimental groups by Dr. David Meyerholz. For Evan Blue staining, mice were i.v. injected with 0.5 mg of Evans Blue (Sigma- Aldrich, St. Louis, MO) diluted in PBS 1 hour prior to euthanasia. Mice were perfused with 10 ml of PBS through the heart. Pictures were then taken to assess vascular permeability. Statistical analysis Graphical and statistical analyses were performed using Prism software (GraphPad Software Inc., San Diego, CA) with error bars representing the SEM. P values were considered significant when p < Results Host genetics influences viral persistence and Cl-13-induced mortality Here I sought to determine the impact of host genetics on the course of a chronic LCMV Cl-13 infection. Mice expressing an H-2 b, H-2 d or H-2 bxd haplotype on either a C57BL or BALB background were utilized. MHC H-2 bxd haplotype mice on the C57BL background were generated via an F1 cross of C57BL/6 and B10.D2 (D2B6F1) mice. Using a similar strategy H-2 bxd haplotype mice on a BALB background were created via 79

96 an F1 cross of BALB/c and BALB.B (BCF1) mice. To determine susceptibility to chronic infection, mice were infected with Cl-13 and disease and viral clearance were monitored. While all strains developed weight loss following LCMV infection, only D2B6F1 mice exhibited mortality (Figure 14). D2B6F1 mice also displayed early control of viral replication and mice that survived exhibited enhanced viral clearance in both serum and systemic organs (Figure 15, upper panel; Figure 16). In addition, BCF1 mice exhibited enhanced viral clearance at day 30 post-infection as compared to MHC congenic mice on the same background (Figure 15, lower panel; Figure 16). These results indicate that H-2 bxd haplotype mice mediate more rapid viral control than either of their H-2 d or H-2 b homozygous controls. However, on the C57BL genetic background, D2B6F1 mice control viral replication more rapidly but also succumb to infection, indicating a combined role for both MHC-linked genes and non-mhc background genes in the mortality resulting from LCMV Cl-13 infection. H-2 bxd haplotype mice prioritize H-2 d -restricted CD8 T cell responses over H-2 b - restricted responses following acute LCMV infection To determine the magnitude and epitope hierarchy of the CD8 T cell response induced in MHC H-2 bxd haplotype mice under ideal T cell priming conditions, the CD8 T cell response was measured 8 days after an acute LCMV-Arm infection (Figure 17A). The total number and frequency of virus-specific CD8 T cells, as measured by expression of the surrogate activation marker CD11a, did not differ between heterozygous (F1) and homozygous MHC haplotype inbred control mice following acute LCMV infection (Figure 17A). Further examination of epitope-specific responses revealed maintenance of 80

97 both immunodominant (NP 118 ) and subdominant (GP 283 ) H-2 d -restricted responses in the MHC H-2 bxd haplotype mice compared to homozygous MHC controls (Figure 18A, white bars). However, within H-2 b -restricted epitopes, both immunodominant (NP 396, GP 33, and GP 276 ) and subdominant (NP 205 ) responses were significantly reduced in F1 mice compared to inbred controls (Figure 19A, white bars). These data indicate that MHC H- 2 bxd haplotype mice induce a T cell response of similar magnitude as homozygous control mice by prioritizing H-2 d -specific CD8 T cell responses over H-2 b -specific responses following acute LCMV infection. Thus, following an acute LCMV infection NP 118 represents the immunodominant epitope and all of the H-2 b -specific epitopes become subdominant. D2B6F1 mice maintain the highest frequency of functional CD8 T cells during chronic viral infection While the relative epitope-specific distribution within the virus-specific CD8 T cell response was altered in H-2 bxd haplotype mice following acute infection, the magnitude of the overall virus-specific CD8 T cell response as determined by the number of CD11a + CD8 T cells was not altered (Figure 17A). However, during chronic viral infection, D2B6F1 mice exhibited an increased frequency of CD11a + CD8 lo T cells (Figure 17B). This did not result in a corresponding increase in the total number of CD11a + CD8 lo T cells (Figure 17B, lower panel). At day 8 post-infection, D2B6F1 mice exhibit severe weight loss and dehydration resulting in reduced cellularity. Thus, to understand the magnitude of the immune response relative to the size of the mouse frequencies of cells were utilized. Consequently, this indicates that a higher proportion of 81

98 CD8 T cells were responding to Cl-13 infection in D2B6F1 mice than the inbred control mice. Previous studies have shown that in B6 mice the immunodominace hierarchy is altered between LCMV Arm and Cl-13 infection (154, 173). NP 396, represents the largest immunodominant epitope following Arm infection. However, NP 396 -specific CD8 T cells are the most rapid to exhaust and be deleted during Cl-13 infection (154, 173). In contrast, the frequency of CD8 T cells specific to subdominant epitopes are maintained allowing for the emergence of subdominant epitope-specific CD8 T cells during chronic infection (154, 173). Examination of individual CD8 T cell epitopes following chronic viral infection revealed that D2B6F1 mice either maintained or increased the frequency of CD8 T cells responding to many of the epitopes regardless of the proportion of CD8 T cells that responded during acute viral infection (Figure 18; Figure 19). Moreover, both BCF1 and D2B6F1 mice maintained a higher proportion of antigen-specific CD8 T cells than inbred mice following chronic versus acute infection (Figure 18; Figure 19). In the F1 mice, a larger proportion of the H-2 b -epitopes were expressed during chronic Cl-13 infection compared to during acute Arm infection, potentially due to their altered immunodominance hierarchy within the total CD8 T cell response (Figure 19). Thus, by broadening the number of epitopes that respond in MHC heterozygous mice, I observed increased maintenance of CD8 T cell epitopes and an increase in the total frequency of CD8 T cells responding based on CD11a expression (Figure 17B). Taken together, my results indicate that increasing MHC heterogeneity results in an increase in the magnitude of the CD8 T cell response that is induced during a chronic viral infection. 82

99 Impact of host genetic background on the effector functions of CD8 T cells during chronic infection H-2 bxd, D2B6F1 mice on the C57BL genetic background were the only strain tested that succumbed to Cl-13 infection (Figure 14). D2B6F1 mice also mounted a larger magnitude CD8 T cell response as determined by the frequency of cells that have upregulated CD11a, than mice on the BALB/c background (Figure 17B). Previous studies have shown that CD8 T cell exhaustion early following Cl-13 infection prevents CD8 T-cell-induced mortality (189, 207). An increase in the CD8 T cell response in D2B6F1 mice suggests that CD8 T cells may play a role in mediating the Cl-13-induced mortality observed in this strain. Therefore, I sought to determine the functionality of the CD8 T cell response and the level of CD8 T cell exhaustion. Both H-2 b - (Figure 20A) and H-2 d - restricted (Figure 20B) CD8 T cell epitopes from D2B6F1 mice exhibited a greater frequency of IFN-γ-producing CD8 T cells than BCF1 mice (Figure 20C). In addition, the total number of NP 118 -specific CD8 T cells, representing the largest of the CD8 T cell epitopes, was significantly increased in D2B6F1 compared to BCF1 mice (p < 0.05; Figure 20D). This result indicates that in addition to an increase in the magnitude of the virus-specific CD8 T cell response, D2B6F1 mice retain more functional CD8 T cells than BCF1 mice. This further suggests a role for non-mhc host genetic background genes in the induction of the CD8 T cell response and the resulting Cl-13-induced mortality in D2B6F1 mice. 83

100 CD8 T cell exhaustion is influenced by the genetic background of the host CD8 T cell exhaustion has been shown to occur in a stepwise fashion, where CD8 T cells initially lose the ability to produce TNF, followed by IFN-γ and eventually resulting in CD8 T cell deletion (154, 162). In addition, exhausted CD8 T cells upregulate expression of inhibitory receptors such as PD-1 (183, 189). To determine the extent of CD8 T cell exhaustion in D2B6F1 and BCF1 mice, I examined both PD-1 expression and epitope-specific IFN-γ production during Cl-13 infection. As early as day 8 post-infection, D2B6F1 mice exhibited reduced PD-1 expression on their CD8 T cells (Figure 21A). By day 30 post-infection, virus was cleared from both LCMV Arminfected mice and surviving Cl-13-infected D2B6F1 mice and PD-1 was no longer expressed on the CD8 T cells (Figure 21B). However, Cl-13-infected BCF1 mice maintained high levels of virus and PD-1 expression on their CD8 T cells at this time point (Figure 21B). This difference in PD-1 expression highlights a major difference between the MHC H-2 bxd haplotype response on the C57BL and BALB background. At both day 8 (Figure 21C) and day 30 (Figure 21D) post-infection, D2B6F1 mice expressed a higher frequency of functional CD8 T cells than BCF1 mice. Thus, CD8 T cells from D2B6F1 mice exhibited reduced CD8 T cell exhaustion compared to BCF1 counterparts. This increased CD8 T cell functionality may contribute to both the rapid viral control but also the Cl-13-mediated mortality observed in D2B6F1 mice. 84

101 D2B6F1 mice exhibit an increased CD4 T cell response during chronic viral infection CD4 T cells have been shown to play a vital role in limiting CD8 T cell exhaustion during chronic viral infection in B6 mice (157, 171). To determine if the CD4 T cell response contributes to the reduced CD8 T cell exhaustion observed in D2B6F1 mice, I next examined the frequency and function of the virus-specific CD4 T cell response. During LCMV Cl-13 infection, D2B6F1 mice exhibited a significantly increased frequency of antigen-specific CD4 T cells as measured by the surrogate activation markers CD11a and CD49d (296), as compared to CD4 T cells from BCF1 mice (p<0.0001; Figure 22A). This increased frequency of virus-specific CD4 T cells coincided with a reduced expression of the exhaustion marker PD-1 on the responding CD4 T cells (Figure 22B). D2B6F1 mice also exhibited an increased frequency of GP 61 - specific CD4 T cells by both IFN-γ production and GP 66 tetramer staining (Figure 22C, D and E). My results indicate that expression of the MHC H-2 bxd haplotype on the C57BL background induces a larger CD4 T cell response than the H-2 bxd haplotype on the BALB background. H-2 bxd haplotype mice on the C57BL genetic background maintain Th1-polarized CD4 T cells during chronic infection Following a chronic Cl-13 infection, CD4 T cells have been shown to convert from Th1 to a Tfh phenotype (60). In an acute LCMV Arm infection, the Th1 cells are maintained and IFN-γ production has been shown to play a pivotal role in supporting CD8 T cell effector activity (60, 164). Furthermore, the Th1 cells are required for viral 85

102 clearance during an acute infection (297). In a chronic infection, IL-21 production by Tfh cells has been shown to be the primary mechanism by which CD4 T cells help maintain CD8 T cell effector functions at late stages of infection (60, 168, 169). Here I find that D2B6F1 mice, which exhibit reduced CD8 T cell exhaustion and more rapid viral clearance, maintain a higher proportion of Th1 (Tbet + CD4 + ) cells while BCF1 mice exhibit more Tfh (CXCR5 + CD4 + ) cells (Figure 23). Thus, my results indicate that maintaining Th1-biased CD4 T cell responses may prevent CD8 T cell exhaustion and viral persistence early during Cl-13 infection allowing for more rapid viral control. CD8 and CD4 T cells contribute to Cl-13-induced mortality in D2B6F1 mice CD8 T cell exhaustion occurs in order to prevent CD8 T-cell-mediated immunopathology (189, 207). In the absence of inhibitory receptors during the early stages of chronic viral infection, CD8 T cells have been shown to both clear virus and induce lethal immunopathology (189, 207). During LCMV Cl-13 infection, CD4 T cells provide vital help for CD8 T-cell-mediated viral control (157, 171). To determine if the increased effector functions observed for both CD8 and CD4 T cells in D2B6F1 mice contributes to their mortality, I depleted CD8 and CD4 T cells prior to Cl-13 infection. In the absence of either CD4 or CD8 T cells, D2B6F1 mice survived Cl-13 infection (Figure 24A and B). In addition, CD8 T cell depletion did not alter the frequency of CD4 T cell response, indicating that CD4 T cells alone cannot induce mortality in the absence of CD8 T cells (Figure 25A). In contrast, CD4 T cell depletion resulted in a significant reduction in the frequency of activated and epitope-specific CD8 T cells (Figure 25B). This indicates that CD4 depletion may enhance the survival of D2B6F1 mice by affecting 86

103 the magnitude of the CD8 T cell response. Furthermore, CD8 and/or CD4 T cell depletion resulted in increased viral titers, confirming their critical role in the enhanced viral control observed in D2B6F1 mice (Figure 24C). These results indicate that the increased CD8 T cell response observed in D2B6F1 mice results in both enhanced viral control and mortality in D2B6F1 mice. IFN-γ contributes to mortality and viral control in D2B6F1 mice To determine the role of Th1-associated cytokines on Cl-13-induced mortality in D2B6F1 mice, IFN-γ was depleted prior to infection. IFN-γ is produced by both CD8 and CD4 T cells during chronic viral infection and has been shown to contribute to viral clearance and immunopathology (64, 66, 163). Depletion of CD4 T cells, but not CD8 T cells, resulted in a significant decrease in the amount of IFN-γ protein in the serum during Cl-13 infection (p = ; Figure 26). CD8 T cell depletion resulted in enhanced IFN-γ levels indicating that CD8 T cells facilitate reduced IFN-γ levels (Figure 26). Importantly, IFN-γ neutralization resulted in increased survival (Figure 27A). Furthermore, IFN-γ neutralization resulted in increased viral titers in the serum (Figure 27B). In addition, IFN-γ neutralization reduced the magnitude of both the CD4 and CD8 T cell response in D2B6F1 mice (Figure 28). These results indicate that CD4 T-cellmediated IFN-γ production plays a critical role in CD8 T cell induction, viral control and the mortality observed in D2B6F1 mice during LCMV Cl-13 infection. D2B6F1 mice exhibit liver lesions following Cl-13 infection To determine the cause of morbidity and mortality observed in D2B6F1 mice following Cl-13 infection, vascular permeability and pathology were examined. 87

104 Assessment of pathology on day 7 post-infection in the brain, lung, liver and kidney did not identify distinct differences between D2B6F1 mice, which succumb to Cl-13 infection beginning on day 8, and B6 or B10.D2, mice which survive Cl-13 infection (personal communications with Dr. Meyerholz). Assessment of vascular permeability by Evans Blue injection on day 8 post-infection did not identify differences in vascular permeability in either the lung or the brain between B6 and D2B6F1 mice following Cl- 13 infection. No signs of Evans Blue permeability were identified in the brain following Cl-13 infection (Figure 29C). Equally high levels of Evans Blue were noted in the lung following Cl-13 infection in both D2B6F1 and B6 mice, indicating similar levels of vascular permeability (Figure 29B). In contrast, D2B6F1 mice did exhibit increased vascular permeability in the liver (Figure 29A). This resulted in focal Evans Blue staining as a result of necrotic liver lesions only observed in D2B6F1 mice (Figure 29A, arrows). These focal lesions indicate increased liver tissue damage in D2B6F1 mice. Discussion Host susceptibility to the establishment of a persistent viral infection is linked to the ability of the host to mount an effective CD8 T cell response to control virus replication. However, CD8 T-cell-mediated viral control is limited by T cell exhaustion during chronic infection, which occurs to prevent T-cell-mediated pathology (207). Thus, understanding the mechanisms that induce an effective CD8 T cell response to prevent viral persistence without the induction of immunopathology is vital to the prevention of chronic infection. Here I show MHC heterogeneity and additional host genetic background genes significantly impact CD8 T cell exhaustion and viral clearance. 88

105 CD8 T cells are important for clearing acute LCMV Arm infection and help to maintain viral control during chronic LCMV Cl-13 infection (156, 157). Increasing the diversity of viral antigens presented during chronic viral infection often increases viral control (282, 285, 286). I show that both D2B6F1 mice and BCF1 mice induce an increased frequency of activated CD8 T cells compared to their inbred controls resulting in enhanced viral control. This is similar to the enhanced viral control observed in HLA heterozygote individuals during HCV and HIV infection (282, 286). In addition, Cl-13 infection in D2B6F1 mice induced both a larger CD8 T cell response and maintained increased effector functions compared to BCF1 mice. Similar to CD8 T cell depletion during SIV infection of macaques (180), CD8 T cell depletion in D2B6F1 mice ameliorates the enhanced viral control observed during Cl-13 infection. Thus, individuals that induce a CD8 T cell response that recognize a broader range of epitopes are more capable of controlling and clearing chronic viral infections. Increasing CD8 T cell function is desired for enhanced control and clearance of chronic viral infection. However, the purpose of CD8 T cell exhaustion is to prevent CD8 T-cell-mediated immunopathology. Previous studies have shown that PD-1 deficiency early during infection results in CD8 T-cell-mediated mortality (189, 207). This is due to an increase in CD8 T cell numbers and function in PD-1-deficient mice (22, 23). Similarly, I observe the mortality in D2B6F1 mice is associated with an increased CD8 T cell response and reduced CD8 T cell exhaustion. In the absence of PD-1, mice succumb to chronic infection due to perforin-mediated killing of endothelial cells by CD8 T cells resulting in increased vascular permeability (207). In D2B6F1 mice I also observed focal 89

106 necrotic liver lesions similar to PD-1-deficient mice (Figure 29). However, D2B6F1 mice do express low levels of PD-1 expression potentially resulting in a less severe and delayed mortality compared to PD-1-deficient mice. Increasing MHC diversity resulted in CD8 T-cell-mediated mortality that was dependent on the host genetic background. Increasing MHC diversity allows for induction of a broader CD8 T cell response. However, during chronic viral infection T cell exhaustion limits the size and function of the virus-specific T cell response (154, 162). My results indicate that the genetic background of the host influences the extent of CD8 T cell exhaustion. CD4 T cells play an important role in CD8 T cell exhaustion (157, 171) and mouse strain susceptibility to infection often differs based on their ability to mount an appropriate CD4 T cell response (294, ). During acute infection, a Th1-biased response is induced to aid in rapid viral clearance (177, 294, ). However, BALB/c mice often induce weaker Th1 responses than B6 mice (298, 299). For example, BALB/c mice are more susceptible to Leishmania and RSV infections due to their inability to mount robust Th1 responses (293, 294, 299). Thus, B6 mice are able to more easily clear Th1-dependent pathogens (298). In addition, perforin-deficient mice succumb to acute LCMV Arm infection in a IFN-γ-dependent manner (301). However, this effect is only observed in B6 mice, not BALB/c mice (302). This difference is believed to be due to a weaker IFN-γ response in BALB/c mice, and only when memory CD8 cells are generated prior to LCMV infection do perforin-deficient BALB/c mice succumb to IFN-γ-mediated death (302). While both BALB/c and B6 mice are susceptible to persistent viral infection, only H-2 bxd expressing mice on the C57BL 90

107 background succumb to Cl-13 infection. This may be due to differences in the CD4 T cell phenotype that is induced. At the peak of the T cell response, D2B6F1 mice were found to induce a larger Th1 response than BCF1 mice. This increased Th1 response in D2B6F1 mice may help to mediate the larger and more functional CD8 T cell response observed in D2B6F1 mice versus BCF1 mice (Figure 30). Thus, both the immune environment and the increased MHC diversity contribute to the CD8 T-cell-dependent viral clearance and mortality observed in D2B6F1 mice. The majority of studies examining CD8 T cell exhaustion following LCMV Cl-13 infection are conducted in inbred B6 mice. However, my findings demonstrate that the host genetic background can profoundly influence the outcome of a LCMV Cl-13 infection. Thus, utilization of hosts with more diverse genetic backgrounds are likely to yield additional factors involved in CD8 T cell exhaustion and the establishment of chronic infections opening new potential avenues to prevent chronic infection in humans. 91

108 Figure 14. H-2 bxd mice with the C57BL background succumb to Cl-13 infection. Mice were infected i.v. with LCMV Cl-13. (A) Weight loss and (B) survival were monitored following infection. Graphs depict cumulative data from 2-4 independent experiments (n = 7-18). Survival statistics were determined by Mantel-Cox Log-rank test. ***, p <

109 Figure 15. H-2 bxd mice with the C57BL background clear Cl-13 infection. Mice were infected i.v. with LCMV Cl-13. Viral titers were determined by plaque assay at the indicated time points in the serum following infection. Graphs depict cumulative data from 2-4 independent experiments (n = 7-18). Titer differences were determined by t test at each time point. *, p < 0.05; ***, p <

110 Figure 16. H-2 bxd mice control chronic viral infection more rapidly than inbred mice. Mice were infected i.v. with LCMV Cl-13 and viral titers were determined by plaque assay at day 30 following infection. Data depict cumulative results from 4 independent experiments (n = 3-15). Titer differences were determined by one-way ANOVA with Tukey s post-test. *, p < 0.05; **, p < 0.01; ***, p <

111 Figure 17. D2B6F1 mice induce the largest frequency of activated CD8 T cells. Mice were infected with either LCMV Arm i.p. or Cl-13 i.v. and spleens were harvested 8 days later. The frequency (top) and total number (bottom) of CD11a + CD8 T cells are shown following (A) Arm and (B) Cl-13 infections. Data depict cumulative results from 3-4 independent experiments (n = 8-17). Statistics were determined by one-way ANOVA with Tukey s multiple comparison test. *, p < 0.05; **, p < 0.01; ***, p <

112 Figure 18. H-2 bxd haplotype mice maintain H-2 d -restricted responses. Mice were infected with either LCMV Arm i.p. or Cl-13 i.v. and spleens were harvested 8 days later. The frequency of tetramer-specific CD8 T cells were identified for (A) H-2 d - specific epitopes following Arm (left) and Cl-13 (right) infection. The fold change in the frequency of tetramer + from Cl-13 compared to Arm infection was examined for (B) H-2 d - specific epitopes. Data depict cumulative results from 3-4 independent experiments (n = 8-17). Statistics were determined by one-way ANOVA with Tukey s multiple comparison test. *, p < 0.05; **, p < 0.01; ***, p <

113 Figure 19. H-2 bxd haplotype mice increase the frequency of H-2 b -restricted responses following Cl-13 infection. Mice were infected with either LCMV Arm i.p. or Cl-13 i.v. and spleens were harvested 8 days later. (A) The frequency of tetramer-specific CD8 T cells were identified for H-2 b - specific epitopes following Arm (left) and Cl-13 (right) infection. (B) The fold change in the frequency of tetramer + from Cl-13 compared to Arm infection was examined for H-2 b - specific epitopes. Data depict cumulative results from 3-4 independent experiments (n = 8-17). Statistics were determined by one-way ANOVA with Tukey s multiple comparison test. *, p < 0.05; **, p < 0.01; ***, p <

114 Figure 20. C57BL background H-2 bxd mice exhibit increased CD8 T cell effector functions compared to BALB background mice. Mice were infected with LCMV Cl-13 i.v and spleens were harvested 8 days later. The frequency of IFN-γ + (y-axis) and TNF + (x-axis) CD CD8 + T cells are shown following H-2 b (A) or H-2 d (B) peptide stimulation. Cumulative frequencies (C) and total numbers (D) of IFN-γ + CD8 T cells are shown in F1 mice. Data depict cumulative results from 3-4 independent experiments (n = 8-16). Statistics were determined by two-way ANOVA with Sidak s multiple comparison test. *, p < 0.05; ***, p <

115 Figure 21. C57BL background mice exhibit reduced CD8 T cell exhaustion compared to BALB background mice. PD-1 expression from day 8 (A) and day 30 (B) CD11a + CD8 T cells from LCMV Cl-13- infected BCF1 (dotted line, top frequency), D2B6F1 (solid line, bottom bold frequency), and Arm infected (shaded) mice. The frequency of tetramer + CD8 T cells expressing IFN-γ + from day 8 (C) and day 30 (D) Cl-13-infected BCF1 (white) and D2B6F1 mice (black). Data depict cumulative results from 2-4 independent experiments (n = 5-16). Statistics were determined by two-way ANOVA with Sidak s multiple comparison test. *, p < 0.05; **, p < 0.01; ***, p <

116 Figure 22. The C57BL background induces a robust CD4 T cell response during chronic viral infection. Mice were infected with LCMV Cl-13 i.v and spleens were harvested 8 days later. (A) The frequency of CD11a hi CD49d + CD4 T cells following infection. (B) PD-1 expression in Cl-13 infected BCF1 (dotted line, top frequency), D2B6F1 (solid line, bottom bold frequency) mice and Arm infected mice (filled line). (C) Flow plots and (D) cumulative data of IFN-γ + and TNF + cells following GP 61 peptide stimulation. (E) The frequency of GP 66 -tetramer + CD4 T cells. Data depict cumulative results from 3-5 independent experiments (n = 12-21). Statistics were determined by t test. ***, p <

117 Figure 23. Cl-13 infection of mice on a C57BL background results in an enhanced Th1 response. Mice were infected with LCMV Cl-13 i.v and spleens were harvested 8 days later. (A) Flow plots depicting Tbet and CXCR5 expression on CD4 T cells 8 days post-infection. (B) Cumulative frequencies of CXCR5 and Tbet expressing CD4 T cells. (C) Pie charts depict the proportions of CXCR5 and Tbet expressing CD4 T cells. Data depict cumulative results from 3-5 independent experiments (n = 5-21). Statistics were determined by two-way ANOVA with Sidak s multiple comparison test. ***, p <

118 Figure 24. Both CD4 and CD8 T cells contribute to Cl-13-induced mortality in D2B6F1 mice. D2B6F1 mice were treated with 200 µg IgG, anti-cd4, anti-cd8 or both on day -2, 2, 6, 10 following LCMV Cl-13 infection. (A) Survival (living/infected) and (B) weight loss were monitored following infection. (C) Serum viral titers were determined on indicated days post-infection. Data depict cumulative results from 2 independent experiments (n = 5-9). Titer statistics were determined by two-way ANOVA with Sidak s multiple comparison test. Survival statistics were determined by log rank Mantel-Cox test compared to IgG treated mice. *, p < 0.05; ***, p <

119 Figure 25. CD4 T cell depletion results in reduced CD8 T cell activation in D2B6F1 mice. D2B6F1 mice were treated with 200 µg IgG, anti-cd4, anti-cd8 or both on day -2, 2, 6, 10 following LCMV Cl-13 infection. (A) The frequency of CD11a hi CD49d + and GP 61 - peptide specific CD4 T cells following stimulation were assessed on day 8 post-infection in the PBL. (B) The frequency of CD11a + and peptide-specific CD8 T cells following stimulation were assessed on day 8 post-infection in the PBL. Data depict cumulative results from 2 independent experiments (n = 5-9). Statistics were determined by t test or one-way ANOVA with Tukey s post-test. **, p < 0.01; ***, p <

120 Figure 26. IFN-γ production is mediated by CD4 T cells following Cl-13 infection. D2B6F1 mice were treated with 200 µg IgG, anti-cd4, anti-cd8 or both on day -2, and 2 following LCMV Cl-13 infection and serum was collected on day 6 for IFN-γ ELISA. Data depict cumulative results from 2 independent experiments (n = 5-9). Statistics were determined by t test. **, p < 0.01; ***, p <

121 Figure 27. IFN-γ contributes to mortality and viral control following Cl-13 infection in D2B6F1 mice. D2B6F1 mice were treated with 200 µg IgG or anti-ifn-γ on day -1, 1, 3, 5, 7, and 9 days post-lcmv Cl-13 infection. (A) Survival (living/infected) and weight loss were monitored following infection. (B) Serum viral titers were measured on day 10 postinfection. Data depict cumulative results from 2-4 independent experiments (n = 6-27). Survival statistics were determined by log rank Mantel-Cox test compared to IgG treated mice. One-way ANOVA and Tukey s post-test were used to determined titer statistics. ***, p <

122 Figure 28. IFN-γ contributes to CD4 and CD8 T cell activation following Cl-13 infection in D2B6F1 mice. D2B6F1 mice were treated with 200 µg IgG or anti-ifn-γ on day -1, 1, 3, 5, 7, and 9 days post-lcmv Cl-13 infection. The frequency of activated CD4 (A) and CD8 (B) T cells were analyzed on day 8 post-infection. Data depict cumulative results from 2 independent experiments (n = 7-9). Statistics were determined by t test. ***, p <

123 Figure 29. D2B6F1 mice express focal necrotic liver lesions during Cl-13 infection. Mice were i.v. treated with PBS (Naive) or infected with LCMV Cl-13 as indicated 8 days prior to harvest. Mice were i.v. injected with 0.5 mg of PBS (No EB) or Evans Blue 30 min prior to harvest. Data depict (A) livers, (B) lungs and (C) brains. Cl-13 infected D2B6F1 and B6 mice were all given Evans Blue. 107

124 D2B6F1 BCF1 Th1 Tfh Th1 Tfh IFN-γ IFN-γ CD8 T cell CD8 T cell Enhanced CD8 T cell effector functions Enhanced viral control Reduced survival Reduced CD8 T cell effector functions Reduced viral control Survival Figure 30. Robust Th1 responses enhance CD8 T-cell-mediated viral control and mortality in D2B6F1 mice following Cl-13 infection. CD4 and CD8 T cells contribute to both viral clearance and mortality following Cl-13 infection of H-2 bxd on the C57BL background (D2B6F1 mice) but not on the BALB background (BCF1 mice). CD4 T cells mediate IFN-γ production and enhance CD8 T cell effector functions following Cl-13 infection. The frequency of Th1-skewed CD4 T cells but not the frequency of Tfh cells is enhanced in D2B6F1 mice compared to BCF1 mice. IFN-γ production by Th1 cells mediates a heightened CD8 T cell effector response resulting in enhanced viral control at the cost of mortality of the host during Cl-13 infection of D2B6F1 mice. 108