Clinical trials have shown a significant increase in incidence of lung cancer among smokers and asbestos workers supplemented with beta-carotene, suggesting a tumor-promoting activity for this agent. We set out to test possible tumor-promoting and chemopreventive activities of dietary and topical beta-carotene in the two-stage 7,12-dimethylbenz[a]anthracene-12-O-tetradecanoylphorbol 13-acetate (TPA) model of mouse skin carcinogenesis. In the first study, the effects of three levels of dietary beta-carotene (6, 60, and 600 micrograms/g purified diet containing no other retinoid or carotenoid) were assessed over a period of 42 weeks. Carcinoma yield was reduced by approximately 50% in the 600 micrograms/g diet group (mean 0.22 carcinomas/mouse) compared with the 6 micrograms/g diet group (mean 0.44 carcinomas/mouse, p = 0.003). The 60 micrograms/g diet group showed a pattern of inhibition similar to the 600 micrograms/g diet group. A protective effect (25% reduction) of beta-carotene (in the 600 micrograms/g diet group) on papilloma formation was also found. However, the intermediate 60 micrograms/g diet group showed the same incidence as the low 6 micrograms/g diet group. This points to a lack of correlation between papilloma and carcinoma incidence, as we also found in previous work on dietary retinoids and carotenoids. The purpose of the second study was to assess whether topical beta-carotene (2 micrograms) has tumor-promoting or chemopreventive activity in the two-stage protocol. In the absence of TPA, beta-carotene had no significant tumor-promoting activity. Instead, papilloma yield induced by TPA was decreased by topical beta-carotene from an average of 20 to approximately 10 papillomas/mouse (p = 2.5 x 10(-7)). The effect of topical beta-carotene persisted beyond the treatment period (Week 24) until the termination of the study at Week 42. Western blot analysis of mouse skin extracts showed that topical beta-carotene upregulated retinoic acid receptor-alpha and -gamma expression in the dorsal skin. This finding suggests that beta-carotene may work as a chemopreventive agent by upregulating the expression of retinoid receptors in mouse skin. In conclusion, our data show that beta-carotene prevents skin carcinoma formation, induces retinoic acid receptor expression, and fails to act as a tumor promoter in the two-stage model of skin tumorigenesis.