Antibody-targeted superantigen therapy induces tumor-infiltrating lymphocytes, excessive cytokine production, and apoptosis in human colon carcinoma.

Bacterial superantigens are the most potent known activators of human T lymphocytes. To engineer superantigens for immunotherapy of human colon carcinoma, the superantigen, staphylococcal enterotoxin A (SEA) was genetically fused to the Fab region of the colon carcinoma-reactive monoclonal antibody C242. In the present study the effector mechanisms involved in the anti-tumor response to C242 Fab-SEA were characterized. Immunohistochemistry and computer-aided image analysis were used in studies of cryopreserved tumor tissue to evaluate the phenotype of infiltrating cells and their cytokine profiles in response to therapy. Human T cells and monocytes were recruited to the tumor area and penetrated the entire tumor mass within hours after injection of C242 Fab-SEA. The production of cytokines at the single-cell level was found to be dominated by tumor necrosis factor (TNF)-alpha, interleukin (IL)-2, IL-4, IL-5, IL-10, IL-12, interferon (IFN)-gamma, granulocyte-macrophage colony-stimulating factor, and transforming growth factor-beta, whereas IL-1-alpha, IL-1ra, IL-1 beta, TNF-beta, IL-3, IL-6, and IL-8 were undetectable. Most of the TNF-alpha, IL-2, IL-12, and IFN-gamma were made by the infiltrating human leukocytes, while the colon carcinoma cells were induced to produce IL-4, IL-10, and TNF-alpha. Up-regulation of IFN-gamma receptors and TNF R p60 receptors was found, while the TNF R p80 receptor was absent. The cytokine production, T cell infiltration, and CD95 Fas receptor expression concomitantly occurred to induce programmed cell death in the tumor cells. This was followed by a strong reduction of the tumor mass that was seen within 24 h after C242 Fab-SEA infusion. These findings demonstrate that antibody-superantigen proteins efficiently recruit tumor-infiltrating lymphocytes actively producing a variety of cytokines likely to be essential for the therapeutic effects observed in the model. Although the humanized SCID model has obvious limitations in its predictive value for treatment of human cancer, we believe that these results encourage clinical evaluation of antibody-targeted superantigens.