Late tumor bearers (LTB), that is, mice that had tumors 3 weeks or longer, can have a selective immune dysfunction and fail to respond to target antigens expressed by the cancer cells. Such mice, however, do respond to nonmalignant cells engineered to express the rejection antigen, and they can be vaccinated with such cells to reject growing tumors. In this study, we compared the efficacy of passive immunization with that of active immunization using the engineered vaccine. An allogeneic MHC class I molecule was used as model tumor antigen. We found that active immunotherapy was only effective for small tumors in early stages of growth. In a later stage of tumor growth, active immunotherapy did not cure any mice, whereas passive immunotherapy was successful in all animals. Reasons for the failure of these LTB to respond to active vaccination with the engineered vaccine may be related to the decreased primary or secondary response we observed in these mice after active immunization. It is suggested that normal antigen-presenting cells expressing the tumor rejection antigen can elicit, in the presence of IL-2, antigen-specific T-cell responses by LTB, and that such T cells may be curative when used in adoptive therapy. We also suggest that the stage of tumor growth and the immune status of LTB more closely simulate the conditions observed in cancer patients.