Reproducibility of microbial mutagenicity assays: II. Testing of carcinogens and noncarcinogens in Salmonella typhimurium and Escherichia coli.

A total of 63 chemicals were tested for mutagenicity in Salmonella typhimurium strains TA98, TA100, TA1535, TA1537, and TA1538, and Escherichia coli WP2 uvrA in a four-laboratory study. Sixty of the chemicals had been tested for carcinogenicity by the National Cancer Institute or the National Toxicology Program. All chemicals were tested for mutagenicity without metabolic activation and with liver S-9 preparations from uninduced and Aroclor 1254-induced F344 rats, B6C3F1 mice, and Syrian hamsters. The intra- and interlaboratory reproducibility of the Salmonella assay with regard to the overall judgment of mutagenic or nonmutagenic was good. The results in the E coli strain, however, exhibited a high degree of variability between laboratories. With one or two exceptions, the mutagens were detected with S-9 preparations from all three species. The uninduced liver S-9 preparations did not activate any chemicals to mutagens that were not also activated by induced S-9, but some chemicals were detected as mutagens only when induced S-9 was used. A positive mutagenic response in Salmonella was predictive of carcinogenicity 69% of the time; when equivocal carcinogens and borderline mutagens were included, the predictivity increased to 83%. Conversely, 76% of the carcinogens were mutagens. When the equivocal carcinogens were included, the proportion dropped to 75%. Relatively few chemicals (18%) were mutagenic in E coli. Not all the carcinogens induced tumors in both rats and mice, and the species-specific carcinogenicity could not be predicted from the S-9-specific mutagenicity.