Evaluation of potential mammalian genotoxins using Drosophila: the need for a change in test strategy

Recent developments in the field of genetic toxicology testing, in particular the outcome of several international collaborative studies, necessitate a re-appraisal of the potential use of Drosophila assays for mutagen testing. For an evaluation of the test performance of the classical sex-linked recessive lethal (SLRL) assay to detect as mutagens mammalian carcinogens, the parameters sensitivity, specificity and accuracy were compared, using as a database the Gene-Tox Report and the results of two international collaborative studies (CSSTT and IPCS). A characteristic of the assay's performance is its low sensitivity (0.33-0.79) and low accuracy (0.50-0.73), when genotoxins other than direct-acting agents and simple promutagens (single-step activation) were included in the analysis. However, the high specificity (0.86-1.0) of the SLRL assay means that in general a positive result has considerable value for prediction of potential genotoxicity in mammals. Experience in the field of carcinogenesis dictates that organ specificity in one species cannot be predicted on the basis of observations in another. It is concluded here that any attempt to use Drosophila assays on the basis of just exposure dose as predictors of effects likely to occur in specific organs in mammals will fail. Examples are provided by the procarcinogens diethylnitrosamine (DEN), vinyl chloride (VC) and hexamethylphosphoramide (HMPA), as well as by the direct-acting mutagens diethylsulphate (DES), N-methyl-N'-nitro-N-nitrosoguanidine (MNNG) and hycanthone. This view is that it would be highly desirable to compare the Drosophila results on germ cells with in vivo mammalian germ-line tests on the basis of molecular dosimetry studies as proposed by Lee. This approach then should enable the comparison on an adequate basis of mechanism of mutagenesis in Drosophila and rodents. It is specifically concluded that the potential value of the SLRL and any other Drosophila germ-cell assay should be judged against their capability of predicting mammalian genotoxicity in a broad sense but certainly not in specific organs. In this function, germ-line assays will have to compete with the novel tests measuring somatic mutation/mitotic recombination (SMART) in Drosophila. The values obtained for sensitivity (0.75-0.78) and accuracy (0.83-0.86) for the latter tests were clearly better than those found for the SLRL test, suggesting that SMART assays present a promising new development. However, the final judgement of their probable role as complementary or confirmatory methods of genotoxic activity observed in Salmonella must await the outcome of the evaluation studies in progress in several laboratories.