We are focused on understanding the cancer-preventive actions of dietary chemopreventive agents such as selenium, curcumin and green tea to determine why these agents prevent cancer in some cases, fail to do so in others, and occasionally cause toxicity. Cancer prevention by a single agent may be less effective than multiple agents administered simultaneously. Just as resistance develops to cancer chemotherapy, cancer prevention may also be counteracted by resistance mechanisms. Similarly, a combination of agents may prevent the development of resistance in cancer prevention just as it does in chemotherapy. Instead of the random mixing of agents in an empiric manner, a mechanistically informed choice of agents is a more rational approach to identify effective combinations of agents for cancer prevention. We hypothesize that selenium and other cancer-preventive agents which generate reactive oxygen species exert their cancer-preventive actions, at least in part, through oxidation of unique cysteine-rich regions present in protein kinase C (PKC) isoenzymes. Considering that PKC is a known target for a variety of tumor promoters, this action is likely intimately involved in the chemopreventive actions of selenium and other chemopreventive agents. More importantly, this induced PKC oxidation can be reversed by selenoprotein thioredoxin reductase, an enzyme involved in maintaining protein thiols in the reduced state. Our initial studies revealed that the cancer-preventive actions of selenium may be negated both by an overexpression of PKC and thioredoxin reductase, conditions which naturally occur in tumor progression. Thus, advanced tumor cells may escape the cancer-preventive actions of selenium and other chemopreventive agents. Recently we have observed that curcumin potentiates the cancer-preventive actions of selenium by inactivating thioredoxin reductase. We postulate that curcumin may also potentiate the actions of other chemopreventive agents which induce sulfhydryl oxidation of PKC. We use a prostate carcinogenesis model employing carcinogen-initiated human epithelial cells. Prostate cancer cells representing early and advanced stages of malignancy are employed to determine the mechanistic basis for the differences in cellular sensitivity to selenium and other agents at various stages of prostate cancer. The results obtained will then be further validated by extending this model to nude mice fed with a diet supplemented with cancer-preventive agents alone and in combination. Understanding the mechanisms by which dietary agents prevent cancer or cause toxicity will allow us to optimize their use as safe and successful cancer-preventive agents.