Response of the ke test to NCI/NTP-screened chemicals. I. Non-genotoxic carcinogens and genotoxic non-carcinogens.

The response of a physico-chemical carcinogen-screening test, the k(e) test, to 46 rodent carcinogens and 20 putative non-carcinogens that had been screened in long-term two-species bioassays by the National Cancer Institute/National Toxicology Program are reported. All of the chemicals screened are those that yield mutagenicity responses in the Ames Salmonella/microsome test that are either equivocal or contrary to the rodent carcinogenicity responses. The electron attachment rate constants, k(e)S, of the test chemicals in cyclohexane at 21 degrees C were measured using a pulse-conductivity technique. The k(e)S of 27 of the 46 rodent carcinogens (59%) are equal or greater than the diffusion-controlled k(e) of carbon tetrachloride, which is regarded as the boundary between a positive and negative response; the k(e)S of 8 of the 20 mutagenic non-carcinogens (40%) are less than diffusion-controlled. If the boundary between positive and negative k(e) responses is decreased to half the diffusion-controlled k(e), six additional carcinogens yield a positive ke response which increases the k(e) test sensitivity to 72% while the specificity to non-carcinogens remains at 40%. Comparison of these k(e)S with measures of the chemicals' electrophilicity that had been inferred from chemical structure indicates that k(e) provides a markedly different measure of electrophilicity and one that complements the Ames Salmonella assay. The use of the k(e) test as an analytical tool to indicate the presence of electron-attaching impurities in solvents such as benzene is discussed, as is the sensitivity of the k(e) test to rodent-liver carcinogens.