The dynamics of T cell responses have been extensively studied during single virus infection of naïve mice. During a viral infection, viral antigen is presented in the context of MHC class I molecules on the surface of infected cells. Activated CD8 T cells that recognized viral antigens mediate clearance of virus through lysis of these infected cells. We hypothesize that the balance between the replicating speed of the virus and the efficiency at which the T cell response clears the virus is key in determining the disease outcome of the host. Lower T cell efficiency and delayed viral clearance can lead to extensive T cellmediated immunopathology and death in some circumstances. To examine how the efficiency of the immune response would impact immunopathology we studied several viral infection models where T cell responses were predicted to be less than optimal: 1. a model of co-infection with two viruses that contain a crossreactive epitope, 2. a viral infection model where a high dose infection is known to induce clonal exhaustion of the CD8 T cell response, 3. a neonatal virus infection model where the immune system is immature and 4. A model of beneficial heterologous immunity and T cell crossreactivity where mice are immunized as neonates when the T cell pool is still developing. Model 1. Simultaneous co-infections are common and can occur from mosquito bites, contaminated needle sticks, combination vaccines and the simultaneous administration of multiple vaccines. Using two distantly related arenaviruses, lymphocytic choriomeningitis virus (LCMV) and Pichinde virus (PICV), we questioned if immunological T cell memory and subsequent protection would be altered following a simultaneous co-infection, where two immune responses are generated within the same host at the same time. Coinfection with these two viruses, which require CD8 T cell responses to clear, resulted in decreased immune protection and enhanced immunopathology after challenge with either virus. After primary co-infection, each virus-specific immune response impacted the other as they competed within the same host and resulted in several significant differences in the CD8 T cell responses compared to mice infected with a single virus. Co-infected mice had a dramatic decrease in the overall size of the LCMV-specific CD8 T cell response and variability in which virus-specific response dominated, along with skewing in the immunodominance hierarchies from the normal responses found in single virus infected mice. The reduction in the number of LCMV-specific CD8 memory T cells, specifically cells with an effector memory-like phenotype, was associated with higher viral loads and increased liver pathology in co-infected mice upon LCMV challenge. The variability in the immunodominance hierarchies of co-infected mice resulted in an enhanced cross-reactive response in some mice that mediated enhanced immune-mediated fat pad pathology during PICV challenge. In both viral challenge models, an ineffective memory T cell response in co-infected mice facilitated increased viral replication, possibly leading to enhanced and prolonged accumulation of secondary effector T cells in the tissues, thereby leading to increased immune pathology. Thus, the magnitude and character of memory CD8 T cell responses in simultaneous co-infections differed substantially from those induced by single immunization. This has implications for the design of combination vaccines and scheduling of simultaneous immunizations. Model 2. The balance between protective immunity and immunopathology often determines the fate of the virus-infected host. Several human viruses have been shown to induce a wide range of severity of disease. Patients with hepatitis B virus (HBV), for example, show disease progression ranging from acute resolving infection to a persistent infection and fulminant hepatitis. Certain rapidly replicating viruses have the ability to clonally exhaust the T cell response, such as HBV and hepatitis C virus (HCV) in humans and the clone 13 strain of LCMV in mice. How rapidly virus is cleared is a function of initial viral load, viral replication rate, and efficiency of antigen-specific T cells. By infecting mice with three different inocula of LCMV clone 13, we questioned how the race between virus replication and T cell responses could result in different disease outcomes. A low dose of LCMV generated efficient CD8 T effector cells, which cleared the virus with minimal lung and liver pathology. A high dose of LCMV resulted in clonal exhaustion of T cell responses, viral persistence and little immunopathology. An intermediate dose only partially exhausted the CD8 T cell responses and was associated with significant mortality, and the surviving mice developed viral persistence and massive immunopathology, including necrosis of the lungs and liver. This was a T cell-mediated disease as T cell-deficient mice had no pathology and became persistently infected like mice infected with a high dose of LCMV clone 13. This suggests that for non-cytopathic viruses like LCMV, HCV and HBV, clonal exhaustion may be a protective mechanism preventing severe immunopathology and death. Model 3. Newborns are more susceptible to infections due to their lack of immunological memory and under-developed immune systems. Passive maternal immunity helps protect neonates until their immune systems have matured. We questioned if a noncytolytic virus that produces strong T cell responses in adult mice would also induce an equally effective response in neonatal mice. Neonates were infected with very low doses of LCMV Armstrong and surprisingly the majority succumbed to infection between days 7-11, which is the peak of the T cell response in adult mice infected with LCMV. Death was caused by T cell-dependent pathology and not viral load as 100% of T cell deficient neonates survived with minimal lung and liver pathology. This is similar to the adult model of medium dose LCMV clone 13, but T cell responses in neonates were not partially clonal exhausted. Furthermore, surviving neonates were not persistently infected, clearing virus by day 14 post infection. In adult mice direct intracranial infection leads to LCMV replication and CD8 T cell infiltration in the central nervous system (CNS), causing CD8 T cell-mediated death. However, this does not occur in adults during LCMV intraperitoneal (ip) infections. We questioned if unlike adults LCMV could be gaining access to the CNS in neonates following ip infection. Replicating LCMV was found in the brain of neonates after day 5 post infection along with virus-specific CD8 T cells producing IFNγ at day 9 post infection. Neonates lacking perforin had complete survival when followed until day 14 post infection, suggesting perforin-mediated T cell-dependent immunopathology within the CNS of neonates was causing death after LCMV infection. Passive immunity from LCMV-immune mothers also protected 100% of pups from death by helping control viral load early in infection. We believe that the maternal antibody compensates for the immature innate immune response of neonates and controls viral replication early so the neonatal T cell response induced less immunopathology. Neonates are commonly thought to have less functional immune systems, but these results show that neonates are capable of producing strong T cell responses that contribute to increased mortality. Model 4. Due to their enhanced susceptibility to infection neonatal and infant humans receive multiple vaccines. Several non-specific effects from immunizations have been observed, for example, measles or Bacillus Calmette- Guerin (BCG) vaccines have been linked to decreased death of children from infections other than measles virus or tuberculosis. These studies mirror the concepts of beneficial heterologous immunity, where previous immunization with an unrelated pathogen can result in faster viral clearance. LCMV-immune mice challenged with vaccinia virus (VV) have lower viral loads then naïve mice and survive lethal infections, but some mice do develop fat pad immunopathology in the form of panniculitis or acute fatty necrosis (AFN). We questioned how immunological T cell memory formed during the immature neonatal period would compare to memory generated in fully mature adults during a heterologous viral challenge. Mice immunized as neonates had comparable reduction in VV load and induction of AFN, indicating that heterologous immunity is established during viral infections early in life. Interestingly, the LCMV-specific memory populations that expanded in mice immunized as neonates differed from that of mice immunized as adults. In adult mice 50% of the mice have an expansion of LCMVNP205- specific CD8 T cells while the majority of neonates expanded the LCMVGP34- specific CD8 T cell pool. This alteration in dominant crossreactivities may be due to the limited T cell receptor repertoire of neonatal mice. In naïve neonatal mice we found altered Vβ repertoires within the whole CD8 T cell pool. Furthermore, there was altered Vβ usage within virus-specific responses compared to adult mice and a wide degree of variability between individual neonates, suggesting enhanced private specificity of the TCR repertoire. Beneficial heterologous immunity is maintained in neonates, but there was altered usage of crossreactive responses. As neonatal mice were found to be so sensitive to LCMV infection we questioned if neonates could control another arena virus that did not replicate as efficiently in mice, PICV. Unlike LCMV infection, neonatal mice survived infection with PICV even with adult-like doses. However, viral clearance was protracted in neonates compared to adults, but was cleared from fat pad and kidney by day 11 post infection. The peak of the CD8 T cell response was similarly delayed. PICV infected neonates showed dose-dependent PICV-specific CD8 T cell responses, which were similar to adult responses by frequency, but not total number. As with LCMV infection there were changes in immunodominance hierarchies in neonates. Examination of the immunodominance hierarchies of PICV-infected neonates showed that there were adult-like responses to the dominant NP38- specific response, but a loss of the NP122-specific response. Six weeks post neonatal infection mice were challenged with LCMV Armstrong and there was a strong skewing of the PICV immunodominance hierarchy to the crossreactive NP205-specific response. These data further support the hypothesis that heterologous immunity and crossreactivity develop following neonatal immunization, much as occurs in adults, although TCR repertoire and crossreactive patterns may differ. Changing the balance between T cell efficiency and viral load was found to altered the severity of the developing immunopathology after viral infection.

The immune response to an infection is determined by a number of factors, which also affect the generation of memory T cells afterwards. The immune response can also affect the stability of the pre-existing memory populations. The memory developed after an infection can influence the response to subsequent infections with unrelated pathogens. This heterologous immunity may deviate the course of disease and alter the disease outcome. The generation and stability of memory CD8 T cells and the influence of the history of infections on subsequent heterologous infections are studied in this thesis using different viral infection sequences. Previous studies using mice lacking individual immunoproteasome catalytic subunits showed only modest alterations in the CD8 T cell response to lymphocytic choriomeningitis virus (LCMV). In this study, I found that the CD8 T cell response to LCMV was severely impaired in mice lacking all three catalytic subunits of the immunoproteasome, altering the immunodominance hierarchy of the CD8 T cell response and CD8 T cell memory. Adoptive transfer experiments suggested that both inefficient antigen presentation and altered T cell repertoire contribute to the reduction of the CD8 T cell response in the immunoproteasome knockout mice. Immune responses generated during infections can reduce pre-existing memory T cell populations. Memory CD8 T cells have been shown to be reduced by subsequent heterologous infections. In this study, I re-examined the phenomenon using immune mice infected with LCMV, murine cytomegalovirus (MCMV) and vaccinia virus (VACV) in different infection sequences. I confirmed that memory CD8 T cells were reduced by heterologous infections, and showed that LCMV-specific memory CD4 T cells were also reduced by heterologous infections. Reduction of the memory CD8 T cells is thought to be the result of apoptosis of memory CD8 T cells associated with the peak of type I interferon early during infection. I showed that memory CD4 T cells were similarly driven to apoptosis early during infection; however, Foxp3+ CD4+ regulatory T cells were relatively resistant to virus infection-induced apoptosis, and were stably maintained during LCMV infection. The stability of Treg cells during viral infections may explain the relatively low incidence of autoimmunity associated with infections. The history of infections can deviate the course of disease and affect the disease outcome, but this heterologous immunity is not necessarily reciprocal. Previous studies have shown the effects of heterologous immunity during acute infections. In this thesis, I showed that the history of LCMV infection led to higher viral titers during persistent MCMV infection, caused more severe immunopathology at the beginning of infection, and reduced the number of MCMV-specific inflationary memory CD8 T cells after the period of memory inflation. In a different context of infection, the history of LCMV infection can be beneficial. LCMV-immune mice have been shown to have lower viral titers after VACV infection, but VACV-immune mice are not protected during LCMV infection. I found that memory CD8 T cells generated from LCMV and VACV infections were phenotypically different, but the differences could not explain the nonreciprocity of heterologous immunoprotection. By increasing the number of crossreactive VACV A11R198-205-specific memory CD8 T cells, however, I showed that some VACV-immune mice displayed reduced viral titers upon LCMV challenge, suggesting that the low number of potentially cross-reactive CD8 T cells in VACV-immune mice may be part of the reasons for the non-reciprocity of immunoprotection between LCMV and VACV. Further analysis deduced that both number of potentially cross-reactive memory CD8 T cells and the private specificity of memory CD8 T cell repertoire played a part in determining the outcome of heterologous infections.