Defining mechanisms by which fever-range thermal stress regulates natural killer cell mediated recognition and killing of tumor cells

All circulating cells, including Natural Killer (NK) cells, normally experience temperature gradients as they move through compartments in the body; however inflammation-driven febrile conditions can expose them to elevated temperatures for several hours. Previous studies in mice suggest a potential for mild (fever-range) whole body hyperthermia to enhance tumor growth control by NK cells. Moreover, recent clinical studies in cancer patients indicate a positive overall survival benefit by combining hyperthermia with radiotherapy. Whether heat-induced enhancement of NK cell activity is involved in the benefits experienced by patients in these clinical trials is not yet clear. In general, few studies to date have studied whether a specific, thermally sensitive molecule or mechanism exists in the anti-tumor immune response which could mediate tumor cell killing. In this thesis research, we hypothesize that mild fever-range thermal stress results in enhancement of NK cell activity through at least two mechanisms: (1) specific upregulation of a surface molecule (MICA) on tumor cells that renders them more sensitive to NKG2D-mediated NK cell recognition and killing, and (2) transient clustering of NKG2D receptor on NK cells which could possibly decrease the NK cell activation threshold. We found that exposure of human peripheral blood NK cells and target tumor cells to fever-range temperatures (39.5°C) significantly enhances their ability to lyse Colo205 target cells. This effect is dependent upon the function of the NKG2D receptor of NK cells and its ligand MHC class I-related chain A (MICA), and is maximal when both NK and tumor cell targets are heated. A similar effect was not observed when long-term cultured NK cell lines or IL-2-activated peripheral blood NK cells were used as effectors, indicating that thermal sensitivity of effectors is activation state-dependent. On NK cells, plasma membrane reorganization may occur after mild heat stress, and disruption of membrane structures, termed lipid rafts, results in loss of thermal stress enhanced NK cell activity. It was observed that thermal exposure does not affect the total level of NKG2D surface expression, but does result in its distinct clustering, identical to that which occurs following IL-2-induced activation and this effect is transient; it is lost on NK cells within 24 hours. On tumor target cells, previously reported stress sensitive cellular targets include heat shock proteins, MICA and MHC Class I. We found that mild temperature elevation results in transcriptional up-regulation of MICA in a manner that correlates with increased sensitivity to cytolysis; however MHC Class I levels did not change. We also examined the levels of soluble MICA molecules after mild thermal stress, since these have been linked to the suppression of NK cell mediated activation through NKG2D. Although we observed a thermal enhancement of MICA on tumor cell surface, we did not observe an increase of the soluble form of MICA with thermal stress. To identify potential underlying mechanisms for the thermally increased MICA levels on tumor cells, we investigated the role of the transcription factor hsf1 on regulating MICA under mild thermal stress. We observed that blocking hsf1 in Colo205 cells using siRNA prevented thermal enhancement of MICA expression and significantly inhibited NK cell cytotoxic activity. Luciferase reporter assays utilizing an hsf1 binding site mutation in the MICA promoter region suggested that this site is required for thermally enhanced MICA promoter activity. We also demonstrate that MICA expression by thermal stress is transient as we investigated the half-life of thermally enhanced MICA on the tumor cell surface. In addition to investigations using a colon tumor cell line as a target, we also found that several patients' colon tumor xenografts also exhibit an enhanced expression of MICA after whole body heating of SCID mice. However, our studies reveal that not all tumor cell lines exhibit thermal sensitivity since HT29 cells did not show a similar enhancement of MICA activity, and further that normal human colon epithelial cells do not change their MICA or HLA expression with mild thermal stress. When we investigated different tumor models in vivo to validate physiological significance of the observed effects of hyperthermia in vitro, we observed an increase in message level of MICA orthologue Rae1 in mouse colon tumor CT26 and a consistent increase in NK cell infiltration into tumors accompanied by an elevated level of tumor cell apoptosis after whole body hyperthermia. Overall, these data suggest that hsf1 mediated MICA upregulation after mild thermal stress and elevated NKG2D clustering could result in enhanced natural killer cell cytotoxic activity. These data may help to identify possible mechanisms by which clinical hyperthermia protocols being tested in cancer patients improve long term control of tumor growth and increase overall survival.