We report here the use of the highly attenuated SIVmac142 clone, unable to establish permanent infection of rhesus macaques, in a vaccine trial. Four rhesus macaques were immunized over a long time period with HUT-78 cells infected with wild-type SIVmac142 or, in order to reinforce the safety use of the vaccine, a deleted mutant with similar in vitro infectivity. The first two injections were done with living cells and the remaining boosts with cells emulsified in muramyl dipeptide adjuvant. Three control macaques were injected with uninfected HUT-78 cells. Over 3 years, we have been unable except once to detect viral infections by three methods. However, antibodies directed against the viral Gag proteins and envelope glycoproteins were detected by immunoblots and/or in vitro neutralization assays. All macaques were challenged intravenously with a low dose (10 animal infectious doses) of a highly pathogenic biological clone of SIVmac251 grown on macaque PBMCs. All seven animals became persistently viremic following challenge. The cell-associated viral loads of the vaccinated monkeys were not reduced relative to those of unvaccinated controls during the first weeks postchallenge even if vaccinated monkeys did not present a transient CD4 decrease. Thus, our data reinforced the notion that the efficacy of live attenuated SIV requires the establishment of persistent infection.