Although live attenuated vaccines can provide potent protection against simian immunodeficiency virus (SIV) and simian-human immunodeficiency virus challenges, the specific immune responses that confer this protection have not been determined. To test whether cellular immune responses mediated by CD8؉ lymphocytes contribute to this vaccine-induced protection, we depleted rhesus macaques vaccinated with the live attenuated virus SIVmac239⌬3 of CD8 ؉ lymphocytes and then challenged them with SIVmac251 by the intravenous route. While vaccination did not prevent infection with the pathogenic challenge virus, the postchallenge levels of virus in the plasmas of vaccinated control animals were significantly lower than those for unvaccinated animals. The depletion of CD8 ؉ lymphocytes at the time of challenge resulted in virus levels in the plasma that were intermediate between those of the vaccinated and unvaccinated controls, suggesting that CD8؉ cell-mediated immune responses contributed to protection. Interestingly, at the time of challenge, animals expressing the Mamu-A*01 major histocompatibility complex class I allele showed significantly higher frequencies of SIV-specific CD8 ؉ T-cell responses and lower neutralizing antibody titers than those in Mamu-A*01 ؊ animals. Consistent with these findings, the depletion of CD8 ؉ lymphocytes abrogated vaccineinduced protection, as judged by the peak postchallenge viremia, to a greater extent in Mamu-A*01 ؉ than in Mamu-A*01 ؊ animals. The partial control of postchallenge viremia after CD8 ؉ lymphocyte depletion suggests that both humoral and cellular immune responses induced by live attenuated SIV vaccines can contribute to protection against a pathogenic challenge and that the relative contribution of each of these responses to protection may be genetically determined.Several unique biological features render human immunodeficiency virus (HIV) infection an especially challenging infection to prevent by vaccination (7,12,34). The development of an AIDS vaccine is further hindered by a lack of information on the types of immune responses that a vaccine should elicit in order to block or limit HIV infection (25). Both clinical observations of HIV-infected humans and experimental data from nonhuman primate models have indicated that antibodies induced by currently available immunogens are unlikely to elicit broadly protective immunity (6, 7). The temporal concordance of emerging virus-specific responses by CD8 ϩ T cells with declining viremia during primary HIV-1 infection (26), together with observations in nonhuman primates demonstrating that the experimental depletion of CD8 ϩ lymphocytes results in striking increases in virus levels (21,45), suggests that the induction of virus-specific cytotoxic T lymphocyte (CTL) responses should be a key objective of vaccine development strategies. However, vaccine approaches that elicit high-frequency virus-specific CTL responses have proven incapable of