Dendritic cells (DCs) are potent antigen-presenting cells that are capable of priming systemic antitumor immune responses in animal tumor models. However, many of the model tumor systems tested need definition of the specific tumor antigens involved. To use DCs in situations that are more relevant to the majority of human cancers, where the antigens are unknown, we have tested the adoptive transfer of immature DCs in mouse colorectal and melanoma models of varying immunogenicity but with undefined antigens. When DCs admixed with a syngeneic primary tumor inoculum were seeded s.c., the growth of the primary tumor was unchanged; however, if the primary tumor was then surgically excised and the animal was rechallenged with the same tumor, significant protection (75%) was generated when DCs were present in the original primary inoculum of a moderately immunogenic colorectal model (CMT93tk). This effect was not observed when a nonimmunogenic melanoma (B16) was tested in an identical protocol. Next, DCs were injected directly into 6-9-mm established tumors; again, protection (55%) was achieved against a secondary tumor challenge following excision of the primary, but only in the CMT93tk model of moderate immunogenicity. To increase the clinical relevance of this approach still further, we tested irradiated allogeneic K1735 melanoma cells mixed with syngeneic DCs as a vaccine against subsequent challenge with the poorly immunogenic syngeneic melanoma B16. The allogeneic vaccine alone was ineffective, but when admixed with DCs, a significant number of animals rejected a subsequent B16 challenge, suggesting that DCs are able to prime an immune response against melanoma antigens shared between K1735 and B16. The generation of systemic antitumor immunity by adoptive transfer of DCs has significant clinical potential because it is technically straightforward and does not require the definition of specific tumor antigens.