Paul T. Golumbek

To compare the ability of different cytokines and other molecules to enhance the immunogenicity of tumor cells, we generated 10 retroviruses encoding potential immunomodulators and studied the vaccination properties of murine tumor cells transduced by the viruses. Using a B16 melanoma model, in which irradiated tumor cells alone do not stimulate significant anti-tumor immunity, we found that irradiated tumor cells expressing murine granulocyte-macrophage colony-stimulating factor (GM-CSF) stimulated potent, long-lasting, and specific anti-tumor immunity, requiring both CD4+ and CD8+ cells. Irradiated cells expressing interleukins 4 and 6 also stimulated detectable, but weaker, activity. In contrast to the B16 system, we found that in a number of other tumor models, the levels of anti-tumor immunity reported previously in cytokine gene transfer studies involving live, transduced cells could be achieved through the use of irradiated cells alone. Nevertheless, manipulation of the vaccine or challenge doses made it possible to demonstrate the activity of murine GM-CSF in those systems as well. Overall, our results have important implications for the clinical use of genetically modified tumor cells as therapeutic cancer vaccines.

Many tumors express tumor-specific antigens capable of being presented to CD8+ T cells by major histocompatibility complex (MHC) class I molecules. Antigen presentation models predict that the tumor cell itself should present these antigens to T cells. However, when conditions for the priming of tumor-specific responses were examined in mice, no detectable presentation of MHC class I-restricted tumor antigens by the tumor itself was found. Rather, tumor antigens were exclusively presented by host bone marrow-derived cells. Thus, MHC class I-restricted antigens are efficiently transferred in vivo to bone marrow-derived antigen-presenting cells, which suggests that human leukocyte antigen matching may be less critical in the application of tumor vaccines than previously thought.

The generation of antigen-specific antitumor immunity is the ultimate goal in cancer immunotherapy. When cells from a spontaneously arising murine renal cell tumor were engineered to secrete large doses of interleukin-4 (IL-4) locally, they were rejected in a predominantly T cell-independent manner. However, animals that rejected the IL-4-transfected tumors developed T cell-dependent systemic immunity to the parental tumor. This systemic immunity was tumor-specific and primarily mediated by CD8+ T cells. Established parental tumors could be cured by the systemic immune response generated by injection of the genetically engineered tumors. These results provide a rationale for the use of lymphokine gene-transfected tumor cells as a modality for cancer therapy.