Implications of arsenic genotoxicity for dose response of carcinogenic effects.

Epidemiological data relating arsenic ingestion and skin and internal cancers strongly suggest a sublinear or threshold relationship. Physiological saturation of methylation-based arsenic detoxification has been proposed as one explanation for a sublinear response. We have evaluated the molecular bases for sublinearity in light of new data and hypotheses regarding arsenic genotoxicity and chemical carcinogenesis. A review of the dose-response relationships observed in arsenic genotoxicity assays is presented. With the exception of sister chromatid exchanges, sublinear dose-response relationships for arsenic-induced chromosomal aberrations were observed repeatedly in different mammalian and human cell systems. Arsenic also enhanced the clastogenicity and mutagenicity of other DNA damaging agents with a sublinear dose response. Consistent with the dose response of arsenic-induced genetic alterations, arsenic also inhibited DNA ligases I and II, enzymes which play a role in DNA repair, with a sublinear dose response. In some cases, protective effects of relatively low exposures to arsenic have been observed, again consistent with sublinearity. We discuss recent theories on the mechanism of arsenic carcinogenicity and the potential implications for dose-response modeling and risk assessment. Overall, based on available arsenic genotoxicity data, we conclude that it is likely that arsenic indirectly induces genetic damage with a sublinear dose response in humans, thus providing a biological basis for a sublinear dose-response relationship for human cancer. Furthermore, these results suggest that linear dose-response modeling from populations experiencing high arsenic exposures is likely to overpredict cancer risks at relatively low arsenic levels.