Vaccine therapy of established tumors in the absence of autoimmunity.
PURPOSE: Many current clinical trials involve vaccination of patients with vaccines directed against tumor-associated antigens, which are, in actuality, "self-antigens" overexpressed in tumors as compared with normal tissues. As tumor vaccines become more potent through the addition of costimulatory molecules and cytokines and the use of diversified prime and boost regimes, the level of concern rises regarding the balance between antitumor immunity and pathological autoimmunity. Studies were conducted using mice bearing a transgenic self-antigen [human carcinoembryonic antigen (CEA)], which is expressed in some normal adult tissues, and tumor expressing the same self-antigen. These mice were vaccinated with recombinant poxviral vectors [recombinant vaccinia, recombinant fowlpox (rF)] encoding the CEA transgene as well as a triad of costimulatory molecules [B7-1, ICAM-1, and LFA-3 (TRICOM)]. Here we investigate the mechanism of tumor therapy and evaluate the safety of such a regimen in a self-antigen system. To our knowledge, the study reported here is the first description of a vaccine to a defined antigen where the regimen is potent enough to induce tumor therapy in the absence of autoimmunity. EXPERIMENTAL DESIGN: CEA transgenic mice were transplanted with CEA-expressing tumors. Fourteen days later, mice were vaccinated with recombinant vaccinia-CEA/TRICOM admixed with recombinant murine granulocyte macrophage colony-stimulating factor and then given low-dose interleukin 2. Mice were boosted on days 21, 28, and 35 with rF-CEA/TRICOM admixed with rF-granulocyte macrophage colony-stimulating factor and then given low-dose interleukin 2. Mice were monitored for survival and compared with groups of mice vaccinated in a similar manner with poxviral vectors containing CEA/B7-1 or CEA transgenes. To determine the mechanism of antitumor therapy, mice were depleted of T-cell subpopulations before vaccination with the CEA/TRICOM regimen. Mice successfully cured of tumor and age-matched control mice were monitored for 1 year. At 1 year, several clinical assays were carried out involving analysis of 9 serological parameters, 11 urinalysis parameters, and 14 immunological parameters. In addition, histopathology was performed on 42 tissues/mouse. RESULTS: The CEA/TRICOM vaccination regimen induced a therapeutic antitumor response as measured by increased survival, which was due largely to induced T-cell responses (both CD4(+) and CD8(+)) as determined by selective T-cell subset depletion. The CEA/TRICOM vaccination regimen induced a significant increase in proliferation of CD4(+) T cells to CEA protein and a significant increase in secretion of IFN-gamma from CD8(+) T cells in response to a defined CEA epitope. Despite CEA expression in normal adult gastrointestinal tissues, no toxicity was observed in the CEA/TRICOM-vaccinated group when an array of clinical serum and urine chemistry assays was conducted 1 year after vaccination. Moreover, a comprehensive histopathological evaluation of all tissues from these groups also showed no evidence of toxicity. CONCLUSIONS: Activation of T cells directed against a tumor-associated self-antigen, sufficient to mediate therapeutic antitumor immunity, was observed in vivo without the development of autoimmunity as analyzed by a comprehensive evaluation of biochemical, immunological, and histopathological criteria. These studies demonstrate that the use of vectors containing as many as three costimulatory molecules does not induce autoimmunity or other pathology. These studies thus demonstrate that a balance can indeed be achieved between the induction of an immune response to a self-antigen, which is capable of antitumor therapy, and the absence of autoimmunity.