Short- and long-wave ultraviolet light (UVB and UVA) induce similar mutations in human skin cells

The mutagenic and carcinogenic properties of sunlight have been attributed to the short-wavelength range (UVB, 290-320 nm) of the solar UV spectrum. Despite being well established that UVA (320-400 nm) can also damage DNA and that it has mutagenic and carcinogenic properties (Stary and Sarasin, 2000(, the relevance of these effects for solar mutagenesis and skin carcinogenesis remains unproven. There are several indications that UVA in particular might play an important role in the pathogenesis of cutaneous malignant melanoma, the deadliest type of skin cancer (Moan et al., 1999; Rünger, 1999; Wang et al., 2001). However, this has recently been questioned, as only UVB, but not UVA, induced melanoma in a transgenic mouse model (De Fabo et al., 2004). The exact mechanisms of UVA mutagenesis are still a matter of debate (Douki et al., 1999). It is generally accepted that pyrimidine dimers, namely cyclobutane pyrimidine dimers (CPDs) and pyrimidine (6,4) pyrimidone photoproducts, are the major mutagenic lesions for UVB. Several lines of evidence indicate that pyrimidine dimers are not the only or major mutagenic lesions with UVA: Enninga et al. (1986) reported that mutagenicity per dimer increased with increasing wavelength, in particular within the UVA range. Further evidences include differences in the spectrum of UVA- and UVB-induced mutations (Drobetsky et al., 1995; Robert et al., 1996; Besaratinia et al., 2004), a lower incidence of p53 mutations in UVA- as compared to UVB-induced skin tumors in mice (van Kranen et al., 1997), and a second peak of tumor formation in the UVA range of the squamous cell carcinoma action spectrum in mice, without a second peak of dimer formation in that wavelength range (de Gruijl et al., 1993). However, to our knowledge, epigenetic differences in the cellular responses to UVA and UVB have as yet not been discussed to account for the different mutagenic properties of UVA and UVB. UVA mutagenesis has been suggested to result from oxidative DNA base modifications (Darr and Fridovich, 1994; Stary and Sarasin, 2000), such as 7,8-dihydro-8-oxoguanine (8-oxoG). This mutagenic lesion is a likely candidate for UVA mutagenesis, as one peak of its formation has been shown for the UVA range (Kielbassa et al., 1997). If this is true, one would expect a high proportion of mutations typically induced by 8-oxoG in the spectrum of UVA-induced mutations (G-->T transversions, generated by mispairing of the template 8-oxoG with adenine, or A-->C transversions, generated by misincorporation of 8-oxoG as the substrate opposite adenine (Epe, 1991; Cheng et al., (1992). Indeed, Besaratinia et al. (2004) found 25 % of UVA-induced mutations in rodent cells to be G-->T transversions, and Persson et al. (2002) found three G-->T transversions in the p53 gene of single keratinocytes derived from UVA-irradiated human skin. Drobetsky et al. (1995) did find a large proportion of A-->C transversions (but not G-->T transversions) in the spectrum of UVA-induces mutations in hamster cells, but not with UVB, and suggested that these mutations are fingerprints for exposure to UVA. Description of this mutation in human squamous skin tumors has been interpreted to indicate a role for UVA in human skin carcinogenesis (Agar et al., 2004). Other oxidative DNA lesions or oxygen-radical-induced DNA strand breaks might also mediate mutation formation following UVA-induced oxidative stress (Peak et al., 1987; Kuluncsics et al., 1999). If, for example, thymine glycol, an oxidized pyrimidine lesion, contributed to UVA mutagenesis, T-->C transition mutations would be expected in the spectrum of UVA-induced mutations, as thymine glycol has been reported to induce such lesions (Essigmann et al., 1989). Based on the assumption that oxidative base damage contributes to solar mutagenesis, the use of antioxidants in sunscreens has been widely advertised for the prevention of photocarcinogenesis. Despite its relatively weak ability to induce pyrimidine dimers through direct excitation of the DNA molecule, such lesions have nevertheless been suggested to contribute at least in part also to UVA mutagenesis. If this is true, C-->T transitions and CC-->TT tandem mutations would be expected in the spectrum of UVA-induced mutations. Recently, it was suggested that UVA generates CPDs via a photosensitized triplet energy transfer (Douki et al., 2003; Rochette et al., 2003) in contrast to formation via direct excitation of DNA by UVB. Currently available investigations of UVA-induced mutations produced conflicting results and used model systems that might not reflect well the situation in humans (Drobetsky et al., 1995; Robert et al., 1996). So far, a detailed spectrum of UVA-induced mutation in primary human skin cells has not been available and an understanding of the exact mechanism of UVA mutagenesis in humans remains elusive today. In order to study in detail the mechanisms of mutation formation following UVA exposure, we generated, sequenced, and compared mutations in the hypoxanthine-phosphoribosyl-transferase (hprt) gene generated by exposure to UVA or UVB in primary human skin fibroblasts.