Vaccination remains one of the most effective means for preventing infectious diseases. During viral infection, activated CD8 T cells differentiate into cytotoxic effector cells that directly kill infected cells and produce anti-viral cytokines. Further T cell differentiation results in a population of memory CD8 T cells that have the ability to self-renew and rapidly proliferate into effector cells during secondary infections. However during persistent viral infection, T cell differentiation is disrupted due to sustained antigen stimulation resulting in a loss of T cell effector function. Despite the development of vaccines for a wide range of viral diseases, efficacious vaccines for persistent viral infections have been challenging to design. Immunization against virus T cell epitopes has been proposed as an alternative vaccination strategy for persistent viral infections, such as HIV. However, vaccines that selectively engage T cell responses can result in inappropriate immune responses that increase, rather than prevent, disease. Quantitative models of virus infection and immune response were used to investigate how virus and immune system variables influence pathogenic versus protective T cell responses generated during persistent viral infection. It was determined that an intermediate precursor frequency of virus-specific memory CD8 T cells prior to LCMV infection resulted in maximum T cell mediated pathology. Increased pathology was independent of antigen sensitivity or the diversity of TCR in the CD8 T cell response, but was dependent on CD8 T cell production of TNF and the magnitude of initial virus exposure. The threshold for exhaustion of responding CD8 T cells ultimately influences the precursor frequency that causes enhanced disease.In addition, viral infection can occur in the context of co-infection by heterologous pathogens that modulate immune responses and/or disease. Co-infection of two unrelated viruses in their natural host, Ectromelia virus (ECTV) and Lymphocytic Choriomeningitis virus (LCMV) infection in mice, were studied. ECTV infection can be a lethal infection in mice due in part to the blockade of antiviral cytokines, including Type I Interferons (IFN-I). It was determined that ECTV/LCMV co-infection results in decreased ECTV viral load and amelioration of ECTV-induced disease, presumably due to IFN-I induction by LCMV. However, immune responses to LCMV in ECTV co-infected mice were also lower compared to mice infected with LCMV alone and biased toward effector-memory cell generation. Thus, providing evidence for bi-directional effects of viral co-infection that modulate disease and immunity. Together the results suggest heterogeneity in T cell responses during vaccination with viral vectors may be in part due to heterologous virus infection or vaccine usage and that TNF-blockade may be useful for minimizing pathology while maintaining protection during virus infection. Lastly, quantitative mathematical models of virus and T cell immunity can be useful to generate predictions regarding which molecular and cellular pathways mediate T cell protection versus pathology.