The efficacy of antigen-specific CD8(+) CTLs depends not only on the quantity of CTLs generated but also perhaps, more importantly, on the avidity of the CTLs. To date, however, no strategy has been shown to preferentially induce higher-avidity human CTLs. In the present study, antigen-presenting cells (APC) generated from human peripheral blood mononuclear cells were infected with a recombinant avipox vector (rF-) containing the transgenes for a triad of costimulatory molecules (human B7.1, intercellular adhesion molecule-1, and LFA-3, designated as rF-TRICOM) and then used to elicit peptide-specific CTLs from autologous T cells. Compared with peptide-pulsed noninfected APCs or peptide-pulsed APCs infected with wild-type vector, peptide-pulsed APCs infected with rF-TRICOM induced not only more CTLs but also higher-avidity CTLs; this was shown by tetramer staining, tetramer dissociation, IFN-gamma production, and cytolytic assays. Peptide-pulsed rF-TRICOM-infected dendritic cells were also shown to induce CTLs with a >10-fold higher avidity than CTLs induced using CD40L-matured dendritic cells; the use of peptide-pulsed CD40L-matured dendritic cells infected with rF-TRICOM as APCs induced CTLs of even greater avidity. To our knowledge, these studies are the first to show a methodology to induce higher-avidity human CTLs and have implications for the development of more efficient vaccines for a range of human cancers.