Reduction of misleading ("false") positive results in mammalian cell genotoxicity assays. II. Importance of accurate toxicity measurement.

In a previous publication, Fowler et al. [4] demonstrated that the seemingly high rate of false or misleading positive results obtained in in vitro cytogenesis assays for genotoxicity - when compared with in vivo genotoxicity or rodent carcinogenicity data - was greater when rodent cell lines were used that were also reported to have mutant or non-functional p53. As part of a larger project for improvement of in vitro mammalian cell assays, we have investigated the impact of different toxicity measures, commonly used in in vitro cytogenetic assays, on the occurrence of misleading positive results. From a list of 19 chemicals that produce "false" positive results in in vitro mammalian cell assays [10], six substances that had given positive responses in CHO, CHL and TK6 cells [4], were evaluated for micronucleus induction in vitro, with different measures of toxicity for selection of the top concentration. The data show that estimating toxicity by relative cell count (RCC) or replication index (RI) consistently underestimates the toxicity observed by other measures (Relative Population Doubling, RPD, or Relative Increase in Cell Count, RICC). RCC and RI are more likely to lead to selection of concentrations for micronucleus scoring that are highly cytotoxic and thus could potentially lead to artefacts of toxicity being scored (elevated levels of apoptosis and necrosis), generating misleading positive results. These results suggest that a further reduction in the frequency of misleading positive results in in vitro cytogenetic assays can be achieved with this set of chemicals, by avoiding the use of toxicity measures that underestimate the level of toxicity induced.