Antigen presentation by B cells and persistence of antigen-antibody complexes on follicular dendritic cells (FDC) have been implicated in sustaining T cell memory. In this study we have examined the role of B cells and antibody in the generation and maintenance of CD8+ cytotoxic T lymphocyte (CTL) memory. To address this issue we compared CTL responses to lymphocytic choriomeningitis virus (LCMV) in normal (+/+) versus B cell-deficient mice. The CTL response to acute LCMV infection can be broken down into three distinct phases: (a) the initial phase (days 3-8 after infection) of antigen-driven expansion of virus-specific CD8+ T cells and the development of effector CTL (i.e., direct ex vivo killers); (b) a phase of death (between days 10 and 30 after infection) during which >95% of the virus-specific CTL die and the direct effector activity subsides; and (c) the phase of long-term memory (after day 30) that is characterized by a stable pool of memory CTL that persist for the life span of the animal. The role of B cells in each of these three phases of the CTL response was analyzed. We found that B cells were not required for the expansion and activation of virus-specific CTL. The kinetics and magnitude of the effector CTL response, as measured by direct killing of infected targets by ex vivo isolated splenocytes, was identical in B cell-deficient and +/+ mice. Also, the expansion of CD8+ T cells was not affected by the absence of B cells and/or antibody; in both groups of mice there was an approximately 10,000-fold increase in the number of LCMV-specific CTL and a greater than 10-fold increase in the total number of activated (CD44hi) CD8+ T cells during the first week after virus infection. Although no differences were seen during the "expansion" phase, we found that the "death" phase was more pronounced in B cell-deficient mice. However, this increased cell death was not selective for LCMV-specific CTL, and during this period the total number of CD8+ T cells also dropped substantially more in B cell-deficient mice. As a result of this, the absolute numbers of LCMV-specific CTL were lower in B cell-deficient mice but the frequencies were comparable in both groups of mice. More significantly, the memory phase of the CTL response was not affected by the absence of B cells and a stable number of LCMV-specific CTL persisted in B cell-deficient mice for up to 6 mo. Upon reinfection, B cell-deficient mice that had resolved an acute LCMV infection were able to make accelerated CTL responses in vivo and eliminated virus more efficiently than naive B cell-deficient mice. Thus, CTL memory, as assessed by frequency of virus-specific CTL or protective immunity, does not decline in the absence of B cells. Taken together, these results show that neither B cells nor antigen-antibody complexes are essential for the maintenance of CD8+ CTL memory.