Genetic immunization through ex vivo transduction of dendritic cells has been suggested as an effective approach to enhance antitumor immunity by activating both CD4+ and CD8+ T cells. Immunizing mice with dendritic cells transduced with an adenovirus expressing the human melanoma antigen glycoprotein 100 (DCAdhgp100) as a cancer vaccine, we demonstrated complete protective immunity and a potent CTL response against melanomas expressing murine glycoprotein 100 in a CD4+ cell-dependent manner. Surprisingly, however, effective tumor rejection was not the result of cooperation between CD4+ and CD8+ T cells. Protective immunity was completely lost when CD4+ cells were depleted immediately before tumor challenge, whereas it was unaffected by removal of CD8+ cells, establishing a principal role for CD4+ cells in the effector phase of tumor rejection. Neither protective immunity nor CTL generation in this model required interleukin 12, in spite of high levels of IFN-gamma secretion by tumor-reactive T cells. Most notably, the DCAdhgp100 vaccine could elicit protective antitumor CD4+ cells in the absence of CD40 ligand, although it does not bypass the need for CD40-mediated signals to generate melanoma-reactive CTLs. Thus, in contrast to the current thinking that the optimal cancer vaccine should include determinants for both CD4+ and CD8+ cells, the potency of the DCAdhgp100 vaccine appears to be a result of its ability to directly prime autoreactive CD4+ cells through a process that does not require interleukin 12 and CD40 signals.