Survivin is a good candidate for cancer immunotherapy since it is overexpressed in most common human cancers, poorly expressed in most normal adult tissues and is essential for cancer cell survival. Previously, we and others have demonstrated that survivin-specific immune responses can be generated in mice and cancer patients. These responses resulted in a substantial antitumor effect. However, the fact that survivin is expressed in normal hematopoietic progenitor cells and endothelial cells may potentially limit the use of vaccination against survivin in the clinic due to possible toxicity. In this study, we have evaluated this risk by using dendritic cells (DC) transduced with an adenovirus encoding mutant human survivin (Ad-surv DCs). Immunization of mice with Ad-surv DCs resulted in generation of CD8+ T cells recognizing multiple epitopes from mouse survivin. These responses provided significant antitumor effect against 3 different tumors EL-4 lymphoma, MC-38 carcinoma, and MethA sarcoma. Survivin-specific T-cells did not affect bone marrow hematopoietic progenitor cells and no autoimmune abnormalities were observed. However, as was the case with other tumor vaccines it provided only partial antitumor effect against established tumors. The existing paradigm suggests that generation of immune response against multiple tumor-associated antigens may provide a better antitumor effect. Here, we directly tested this hypothesis by combining vaccines targeting different tumor-associated proteins: survivin and p53. Despite the fact that combination of 2 vaccines generated potent antigen specific T-cell responses against both molecules they did not result in the improvement of antitumor effect in any of the tested experimental tumor models.