Retinoic acid and photocarcinogenesis--a controversy.

There is considerable evidence that retinoids, including retinoic acid, prevent or normalize many malignant transformations. Contrary reports are few in number and are often problematic in design or interpretation. The mechanism of action of retinoids in differentiation and in neoplasia is not understood. The effects, however, are multifarious, and may be exerted directly on normal tumor cells, or indirectly via cytotoxic mechanisms and the immune system. After extensive testing and after almost 25 years of use on human skin, retinoic acid has not been found to be a carcinogen. When applied topically to non-irradiated mouse skin for up to 18 months it does not produce tumors (130). It was also negative in the Ames test. Moreover, the role of retinoic acid in normal differentiation of skin and mucous membranes appears to make it a physiologic necessity. Since UV radiation apparently destroys epidermal vitamin A (127, 128), its metabolite, retinoic acid, may need to be replenished continuously. The extreme vulnerability of the hairless mouse to UV radiation makes it a valuable animal for many photobiologic studies, including studies of carcinogenesis. However, this same extraordinary vulnerability means that we must be cautious in making casual extrapolations to humans. This problem is compounded when active topical agents are added, especially when application is made to the entire dorsum of the mouse, in contrast to limited areas of human skin. In most cases, animal studies have to be interpreted very carefully. The final judgment must rest on the human experience.