352. Intentional Killing of Normal Melanocytes by Gene Transfer Generates Curative Anti Melanoma Immunity without Obligate Autoimmunity

Abstract

Most tumor vaccination strategies depend upon the molecular identification of specific tumor rejection antigens and/or the use of patient derived tumor materials. In addition, there is a strong correlation between the use of cancer vaccines and the induction of autoimmune reactivity. We hypothesised that in situ, inflammatory killing of a normal tissue from which a tumor derives may generate autoimmune reactivity to self antigens expressed in that tissue as well as in the tumor cells. Here we report that simple intradermal injection of 2 plasmids promoted tissue specific, inflammatory killing of melanocytes and induced an immune response that eradicated systemically established B16 tumors. The therapeutic response was CD8+T cell dependent but, significantly, was rapidly suppressed in vivo and did not induce autoimmune disease in animals in which tumors had been cured. The T cell response was characterized by selection of B16 tumor variants which grow extremely aggressively in vivo, are amelanotic and which have lost expression of the tyrosinase and TRP-2 antigens. The mechanism of antigen loss was, at least in part, through epigenetic mechanisms. These could be reversed by long term growth in culture or, more rapidly, by treatment with the de-methylating agent 5-azacytidine. Depletion of putative regulatory T cells within the CD25+ T cell population improved therapeutic efficacy. Finally, vaccination could be further improved using adoptive transfer of activated T cells with specificity for an antigen that was not a target of the initial T cell response raised by melanocyte killing. These data demonstrate that inflammatory killing of a normal tissue can lead to the immunoselection of tumor cell growth variants. They also raise the prospect that such vaccinations could be effectively used in combination with other conventional anti tumor treatment modalities.