A 7-week treatment with the tobacco carcinogen NNK induced 8-10 lung adenomas per A/J mouse. NNK suppressed humoral and cellular immune responses and increased plasma PGE2 and LTB4 levels. This protocol is particularly suitable for testing NSAIDs and lipoxygenase inhibitors as cancer preventive agents. Sulindac and ASA inhibited lung tumorigenesis by 52 and 60%, respectively, attenuated the suppressive effect of NNK, and lowered the plasma PGE2 to basal levels. In contrast, naproxen neither inhibited lung tumorigenesis nor increased NNK-suppressed NK cell cytotoxicity. NSAIDs and lipoxygenase inhibitors had additive preventive efficacies against NNK-induced lung tumorigenesis. However, sulindac was not effective in preventing lung tumorigenesis induced by B[a]P, which lacks immunosuppressive activity. These results and those published by other investigators lead to the following hypothesis: Reactive intermediates derived from NNK interfere with the stimulation of the complex NF-kappa B/I kappa B. NF-kappa B is involved in the regulation of immune and inflammatory responses. The authors propose that NNK-derived intermediates induce the expression of COX-2 and lipoxygenase involved in NNK activation. This hypothesis provides a rationale for the lack of efficacy of naproxen to prevent tumorigenesis, to attenuate NNK-induced synthesis of PGE2, and to increase NK cell cytotoxicity. According to this hypothesis, PGE2 synthesis and induction of apoptosis contribute to varying degrees to the mechanism of cancer prevention.