The interactions between CD80 and CD86 on antigen-presenting cells and CD28 on T cells serve as an important costimulatory signal in the activation of T cells. Although the simplistic two-signal hypothesis has been challenged in recent years by the identification of different costimulators, this classical pathway has been shown to significantly impact antiviral humoral and cellular immune responses. How the CD80/CD86-CD28 pathway affects the control of chronic or latent infections has been less well characterized. In this study, we investigated its role in antiviral immune responses against murine gammaherpesvirus 68 (MHV-68) and immune surveillance using CD80/CD86 ؊/؊ mice. In the absence of CD80/CD86, primary antiviral CD8 ؉ T-cell responses and the induction of neutralizing antibodies were severely impaired. During long-term immune surveillance, the virus-specific CD8 ؉ T cells were impaired in IFN-␥ production and secondary expansion and exhibited an altered phenotype. Surprisingly, a low level of viral reactivation in the lung was observed, and this effect was independent of CD28 and CTLA-4. Thus, CD80 and CD86, signaling through CD28 and possibly another unidentified receptor, are required for optimal immune surveillance and antiviral immune responses to murine gammaherpesvirus.

In herpesvirus infections, the virus persists for life but is contained through T-cell-mediated immune surveillance. How this immune surveillance operates is poorly understood. Recent studies of other persistent infections have indicated that virus persistence is associated with functional deficits in the CD8 ؉ T-cell response. To test whether this is the case in a herpesvirus infection, we used a mutant murine gammaherpesvirus that is defective in its ability to persist in the host. By comparing the immune response to this virus with a revertant virus that can persist, we were able to dissect the changes in the antiviral CD8 ؉ T-cell response that are induced by virus persistence. Surprisingly, persistently infected mice controlled a secondary challenge infection more rapidly than nonpersistently infected mice, indicating enhanced rather than diminished effector functions. Consistent with this, virus-specific CD8 T cells from these mice exhibited faster upregulation of the cytotoxic mediator granzyme B. Another unexpected finding was that CD8 ؉ T cells from neither infection responded efficiently to homeostatic cytokines. The unresponsiveness of the memory cells from the nonpersistently infected mice appears to be linked to the prolonged replication of virus within the lungs. Other changes seen in different chronic infection models were also observed, such as changes in Bcl-2 levels, interleukin-2 production, and the immunodominance hierarchy. These data show persistence of gammaherpesvirus type 68 alters the properties of CD8 ؉ T cells and illustrates that immune surveillance does not require CD8 T cells with the same attributes as "classical" memory CD8 ؉ T cells.

IL-15 is known to be critical in the homeostasis of Ag-specific memory CD8+ T cells following acute viral infection. However, little is known about the homeostatic requirements of memory CD8+ T cells during a latent viral infection. We have used the murine gammaherpesvirus-68 (MHV-68) model system to investigate whether IL-15 is necessary for the maintenance of memory CD8+ T cells during a latent viral infection. IL-15 is not essential either for the initial control of MHV-68 infection or for the maintenance of MHV-68-specific memory CD8+ T cells. Even at 140 days postinfection, the proportion of CD8+ T cells recognizing the MHV-68 epitopes were the same as in control mice. The maintenance of these memory CD8+ T cells was attributable to their ability to turn over in vivo, probably in response to the presence of low levels of Ag. IL-15−/− mice had a significantly higher turnover rate within the virus-specific memory CD8+ T cell population, which was the result of increased levels of viral gene expression rather than an increase in viral load. These cells did not accumulate in the spleens of the IL-15−/− mice due to an increased sensitivity to apoptosis as a result of decreased Bcl-2 levels. Intriguingly, memory CD8+ T cells from latently infected mice failed to undergo homeostatic proliferation in a naive secondary host. These data highlight fundamental differences between memory CD8+ T cells engaged in active immune surveillance of latent viral infections vs memory CD8+ T cells found after acute viral infections.