A. Recombinant anti-tumor vaccines: Using murine tumor models, we have shown that recombinant vaccinia viruses (rVv) are capable of eliciting powerful immune responses resulting in lymphocytes capable of destroying tumor cells in vitro. We have synthesized fusion peptides with sequences representing hydrophobic endoplasmic reticulum insertion signal sequences linked to minimal antigenic determinants that are capable of eliciting anti-tumor immune responses against model antigens in vitro. Finally, in one in vivo tumor model, we have successfully immunized mice against subsequent development of a tumor challenge using a rVV. Importantly, we have also successfully used this rVV in an active immunization schema to treat mice with established cancer. B. Immune recognition of tumor cells: We have shown that some tumor cells present endogenous antigen poorly. One mouse tumor, MCA 101, has a nearly complete block in its ability to present endogenous antigens to CD8+ T cells. Human cancer cells can escape from immune recognition by the loss of functional beta2-microglobulin, deletion of specific major histocompatibility alpha-chain alleles, and downregulation of the antigen processing machinery. Most importantly, we have shown that each of these mechanisms can be corrected in vitro. C. Gene-modified tumor cells: The antigen processing capabilities of some tumor cells can be enhanced by gene modification with interferon-gamma. We have shown that a nonimmunogenic tumor called MCA 101, which was not previously capable of producing effective TIL, could elicit CD8+ T cells therapeutically active against the wild-type tumor after retroviral transduction with the cDNA for interferon-gamma. The same genetic maneuver with another nonimmunogenic murine fibrosarcoma, a clone of MCA 102, could result in the generation of T cells in draining lymph nodes capable of treating established parental metastatic tumor.
