Carcinogenicity of mutagens: predictive capability of the Salmonella mutagenesis assay for rodent carcinogenicity.

A total of 224 chemicals that have been tested in long-term studies for carcinogenicity in rats and mice by the National Cancer Institute and the National Toxicology Program were tested for mutagenicity in Salmonella typhimurium. Correlations between mutagenicity and carcinogenicity were examined. The influences of chemical structure, rodent species and organ responses, and bacterial strain responses on the carcinogenesis/mutagenesis correlations were also examined. Not all carcinogens induced tumors in both rats and mice. A clear mutagenic or equivocal mutagenic response in Salmonella was predictive for 77% of the carcinogens or equivocal carcinogens, although only 54% of the 149 carcinogens or equivocal carcinogens were mutagens, and 58% of the nonmutagens were carcinogens or equivocal carcinogens. The proportion of mutagens and equivocal mutagens that were not carcinogenic or equivocal was 23%. There was no apparent way to distinguish the mutagenic carcinogens from the mutagenic noncarcinogens by the responses of the specific Salmonella strains. The proportions of different chemical classes in the data base strongly affected the correlations; 40% of the chlorinated carcinogens were mutagens, whereas 75% of the amines and 100% of the nitro-containing carcinogens were mutagens. Because 29% of the chemicals (30% of the carcinogens) were chlorinated, the poor correlation of this class was reflected in the overall correlation. It is concluded that the use of the Salmonella mutagenicity assay is warranted for the identification of carcinogens, but not for noncarcinogens. The proportion of carcinogens detected as mutagens is dependent on the specific classes of chemicals tested and on the rodent species used to define the carcinogens.