Genetic toxicity information is critical for the safety assessment of all xenobiotics. In the absence of carcinogenicity data, genetic toxicity studies may be used to draw conclusions about the carcinogenicity potential of chemicals. However, current in vitro assays have many limitations as they produce a high rate of irrelevant positive data and possible false negative data due to the weakness of the in vitro models used. Based on the knowledge that the majority of human genotoxic carcinogens require metabolic activation to become genotoxic, it is necessary to develop in vitro cell models that mimic human liver metabolism to replace the use of liver S9 fraction, which, though helpful for predicting the potential carcinogenicity of chemicals in rodents, is questionable in humans. We therefore investigate whether the recently described human hepatoma HepaRG cells, which express the major characteristics of liver functions similarly to primary human hepatocytes, could be a suitable model for human genotoxicity assessment. We determine the performance of comet and micronucleus assays in HepaRG cells to predict in vivo genotoxins based on the list of compounds published by European Centre for the Validation of Alternative Methods (ECVAM). Twenty compounds were tested in HepaRG cells with comet and micronucleus assays over a 24-h period. The specificity, the sensitivity, and the accuracy of the two tests were determined. We found that the comet assay had higher specificity (100%) than the micronucleus (MN) test (80%), whereas the latter was far more sensitive (73%) than the former (44%), resulting nonetheless in an accuracy of 72% for the comet assay and 75% for the MN test. Taken together, our data suggest that the HepaRG cell line can be of use in genetic toxicology and that efforts to develop competent human liver cell models should be increased.
Keywords: bioactivation; genotoxicity; hepatocytes; predictivity; sensitivity; specificity.