Further evidence against a direct genotoxic mode of action for arsenic-induced cancer.

Arsenic in drinking water, a mixture of arsenite and arsenate, is associated with increased skin and other cancers in Asia and Latin America, but not the United States. Arsenite alone in drinking water does not cause skin cancers in experimental animals; therefore, it is not a complete carcinogen in skin. We recently showed that low concentrations of arsenite enhanced the tumorigenicity of solar UV irradiation in hairless mice, suggesting arsenic cocarcinogenesis with sunlight in skin cancer and perhaps with different carcinogenic partners for lung and bladder tumors. Cocarcinogenic mechanisms could include blocking DNA repair, stimulating angiogenesis, altering DNA methylation patterns, dysregulating cell cycle control, induction of aneuploidy and blocking apoptosis. Arsenicals are documented clastogens but not strong mutagens, with weak mutagenic activity reported at highly toxic concentrations of inorganic arsenic. Previously, we showed that arsenite, but not monomethylarsonous acid (MMA[III]), induced delayed mutagenesis in HOS cells. Here, we report new data on the mutagenicity of the trivalent methylated arsenic metabolites MMA(III) and dimethylarsinous acid [DMA(III)] at the gpt locus in Chinese hamster G12 cells. Both methylated arsenicals seemed mutagenic with apparent sublinear dose responses. However, significant mutagenesis occurred only at highly toxic concentrations of MMA(III). Most mutants induced by MMA(III) and DMA(III) exhibited transgene deletions. Some non-deletion mutants exhibited altered DNA methylation. A critical discussion of cell survival leads us to conclude that clastogenesis occurs primarily at highly cytotoxic arsenic concentrations, casting further doubt as to whether a genotoxic mode of action (MOA) for arsenicals is supportable.