DNA adducts produced by oils, oil fractions and polycyclic aromatic hydrocarbons in relation to repair processes and skin carcinogenesis

Abstract

Ten polycyclic aromatic hydrocarbons (PAHs) mainly with three or four aromatic rings were tested for their ability to induce DNA adduct formation in mouse skin. Four of these were selected to investigate adduct formation and loss over a period of 8 days. Three mineral oils were also examined for their adduct forming ability and one was selected for adduct formation and loss over a period of 8 days. In addition, fractions derived from the same oil containing 2–3- and 4–6-ring aromatic compounds were applied to mouse skin in a non-carcinogenic oil vehicle and adduct levels were observed over an 8-day period. It was found that PAHs that had no mutagenic, initiating or carcinogenic activity and those that had mutagenic activity in bacteria but no initiating activity in mouse skin failed to produce DNA adducts in mouse skin. Two of the three PAHs with initiating activity and both complete carcinogens produced clear evidence of adduct formation, the adduct levels produced by complete carcinogens being 100–1000 times greater than those produced by initiators. Examination of adduct formation and loss with the carcinogenic PAHs benzo[a]pyrene and 5-methylchrysene over an 8-day period showed a peak at 24 h and an apparent two-phase process of adduct loss. It is suggested that the first steep loss was due to DNA repair and that the more gradual subsequent loss was probably due to epidermal hyperplasia and desquamation. With the initiator 1, 4-dimethylphenanthrene (three rings) a peak of adduct formation was seen at 2 days and adduct levels were not reduced much by 8 days. This suggested that, with initiators, adduct formation and repair may be spread over a longer period than with complete carcinogens. With the whole oils, clear evidence of adduct formation was seen with both a carcinogenic non-solvent-refined oil and with a non-carcinogenic residual oil. The level of adduct formation with the residual oil, however, was much lower than with the carcinogenic oil. When adduct formation by the carcinogenic oil was examined over 8 days, the pattern of adduct formation and loss was similar to that of a tumour initiator rather than a complete carcinogen. Peak adduct levels on the diagonal of the thin-layer chromatography (TLC) plates seemed to occur at 1 and 4 days after treatment, with no clear reduction after 8 days. From examination of adducts formed by the 2–3-ring and 4–6-ring aromatic fractions, it appeared that the main adduct spots produced by the carcinogenic oil were due to the 2–3-ring aromatic components of the oil. Adduct spots near the vertical axis of the TLC plates were also seen with the 2–3-ring and 4–6-ring fractions. The relevance of these spots is uncertain, but if they truly represent adducts, the findings suggest that they are due mainly to 4-ring PAHs. The studies suggest that the activity of carcinogenic oils is largely due to substituted 3- and 4-ring polycyclic aromatic compounds and that more attention should be paid to substituted 3-ring compounds in predicting the carcinogenic potential of oils from analytical data. Copyright © 2000 John Wiley & Sons, Ltd.