Dendritic cells (DCs) primed with tumor antigens can effectively mediate the regression of a variety of established solid malignancies in both murine and human models. Despite such clinical efficacy, the optimal means of DC priming is unknown. The goal of this study was to compare three methods of tumor preparation: irradiation, boiling, or freeze thaw lysis for DC priming. Mouse bone marrow-derived DCs were loaded with defined ratios of E.G7 tumor cells expressing a model tumor antigen, OVA. Sensitized DCs were used for stimulation of OVA-specific CTLs derived from OT-1 T-cell receptor transgenic mice. IFN-gamma release, determined by ELISA at 24 and 48 h, was used to assess the expression of antigens by DCs. DCs loaded with irradiated tumors were effective stimulators for OT-1 CTLs, whereas DCs stimulated with freeze-thawed or boiled tumors did not stimulate IFN-gamma production. Freeze-thaw lysis appeared to inhibit CTL activity in vitro and in two of three cases, this effect was not overcome by the addition of OVA. The ability to load irradiated tumor cells was reproduced in two analogous human melanoma models using melanoma cell lines expressing gp100 and CTL clones specific for a gp100 melanoma antigen. Consistent with the in vitro data, only DC/irradiated tumor vaccines were effective in preventing or delaying outgrowth of E.G7 and a poorly immunogenic murine squamous cell carcinoma (SCCVII), on local tumor challenge. These data demonstrate that the method of tumor cell preparation clearly influences the ability of DCs to present antigen to T cells. Correlation of in vitro data with the generation of protective immunity in vivo suggests the utility of irradiated tumor-primed DCs as a means to generate protective immunity in patients with solid malignancies.