Vitamin A and its derivatives (retinoids) exert modulatory effects on
epithelial differentiation and are used therapeutically against skin c
ancers, but the role of dietary vitamin A in ultraviolet (UV)-induced
carcinogenesis is far from clear. To study this process, 220 hairless
mice were given diets containing low (0.3-0.6 mg/kg; A-) or high (4-5
mg/kg; A+) amounts of retinol, which resulted after 2 months in an sim
ilar to 4-fold difference in liver and skin vitamin A levels as determ
ined by HPLC. Commencing after 1 month of diet, daily irradiations wit
h UVB (280-320 nm) or UVAB (280-380 nm) were given to 176 of the anima
ls for 18 weeks (cumulative doses of UVB and UVA: 26 J/cm(2) and 168 J
/cm(2), respectively). The first skin tumours, known to be squamous ce
ll carcinomas, appeared after 35 weeks in the UVAB-irradiated A+ anima
ls and 5-6 weeks later in the other groups. After one year the frequen
cy of tumour-bearing animals was 49-63% in the A+ groups and 28-39 % i
n the A-groups (P = 0.003). Two months later the corresponding figures
were 66-72% and 50-53%, respectively (P = 0.014). Disregarding the ef
fect of dietary vitamin A, there was no difference in the final tumour
incidence between UVB-and UVAB-irradiated animals. The epidermal vita
min A content at 72 h postirradiation was similar to 60% lower in A+ a
nimals and similar to 10% lower in A- animals compared with the non-ir
radiated controls. Rather than protecting against skin cancer, a diet
rich in vitamin A seems to facilitate UV carcinogenesis in hairless mi
ce. A possible explanation is that photodecomposition of excessive vit
amin A generates short-lived intermediates that may act as photosensit
izers during cutaneous carcinogenesis.