The primary testing method for the identification of potential human carcinogens remains the two-year rodent bioassay where in-bred mice and rats are used as potential surrogates for human risk. We have proposed to supplement the two-year cancer bioassays with short-term bioassays conducted in selected transgenic mouse models. The rapid progress in the molecular understanding of cancer and the identification of genes that are directly involved in neoplasia have provided specific insights into the relationship between genes which control cell proliferation, suppression of tumor phenotype or modulators of these genes. The transgenic models utilized today are focused in these areas. To this end, the haploinsufficient p53 knockout (p53+/-) and zetaglobin v-Ha- ras (Tg.AC) transgenic mouse models have been compared with the results of conventional, two rodent species carcinogen bioassays. The results of these studies have shown some definitive capacity to identify a selected group of known human carcinogens, and that the results from the transgenic models provide direct mechanistic insights into the action of chemicals. For example, the p53+/- model has demonstrated the preferential identification of genotoxic/mutagenic carcinogens in which molecular changes in the wildtype allele can be demonstrated and associated directly with chemical exposure. In addition, in collaborative experiments, it has been possible to demonstrate the induction of subcutaneous sarcomas by implantation of transponders used for animal identification. Subsequent molecular analysis of these tumors has demonstrated specific p53 allelic loss. The Tg.AC model has demonstrated preferential responsive to nongenotoxic carcinogens. In studies in parallel with the conventional two-stage initiation/promotion model, data have been obtained to suggest that both the two-stage model and the Tg.AC transgenic model have the capacity to identify potential nongenotoxic carcinogens and tumor promoters in a much shorter period of time than conventional bioassays. In addition, the analysis of data obtained over a long period of time with tumor promotion models and analysis of long-term bioassays suggests that the property of tumor promotion is one component of complete nongenotoxic carcinogens. Thus, the mechanism of action of chemicals in the Tg.AC model is the preferential induction of transgene expression which leads to the development of a reporter phenotype of skin papillomas. Direct prospective comparison of studies conducted in conventional rodent bioassays with these transgenic models has demonstrated that the transgenic models are not supersensitive in that they do not detect agents which cause preferential species-specific effects in either mice or rats. Many critical questions remain to be answered in order to promote a scientific consensus on the use of these transgenic lines for the identification of potential human carcinogens, but the data developed to date suggests that these models can play an important role in chemical and drug safety assessments.
