Aim of the study was to investigate the usefulness of two human derived hepatoma cell lines (HepG2 and Hep3B) for the detection of dietary and lifestyle related DNA-reactive carcinogens. Comparisons of the sensitivity of HepG2 cells of different origin towards benzo[a]pyrene (B(a)P) showed that strong differences exist in the induction of micronuclei (MN). The most sensitive was used for all further experiments, in which we investigated the effects of aflatoxin B(1) (AFB(1)), B(a)P, As(2)O(3), CdCl(2), 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP), N-nitrosodimethylamine (NDMA), N-nitrosopyrrolidine (NPYR), 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK), ethanol, acetaldehyde and caffeic acid in micronucleus (MN) tests. Dose dependent effects were detected in HepG2 with AFB(1) (0.2microM), CdCl(2) (2.2microM), As(2)O(3) (8.1microM), B(a)P (22.7microM), PhIP (35.7microM), NDMA (22.7mM), acetaldehyde (11.2mM) and ethanol (442.2mM). Numbers in parentheses indicate the C(D) values (concentration that induced a two-fold increase over the background). NNK and caffeic acid gave negative results under all conditions. In Hep3B cells, the effects were generally weaker. With PhIP, As(2)O(3) and NDMA negative results were obtained; with caffeic acid and NPYR marginal but significant induction of MN was observed. Enzyme measurements showed that both cell lines possess CYP1A1, glutathione-S-transferase (three-fold higher in HepG2) as well as N-acetyltransferase (NAT) 1 and sulfotransferases (SULT1A1 and SULT1A3; two- and seven-fold higher in HepG2); other cytochrome P450 enzymes (CYP1A2, 2B1, 2E1) and NAT2 were not detectable. The differences in the activities of the various enzymes may explain the contrasting results obtained in the MN experiments. Overall, our results indicate that the HepG2 line is more sensitive towards dietary genotoxins than Hep3B, and support the assumption that the HepG2/MN assay enables the detection of genotoxic carcinogens which give negative results in other currently used in vitro assays.