Rodent studies of furan are associated with liver cell necrosis, release of liver-associated enzymes, increased hepatocyte proliferation, and hepatocarcinogenesis. For carcinogens whose proposed mode of action is cytolethality, it is hypothesized that the dose-response curve for tumor development would parallel the dose-response curve for cell death with compensatory proliferation in the target organ. To prospectively test this hypothesis, female B6C3F(1) mice were exposed to furan at carcinogenic doses and lower for 3 weeks or 2 years. At 3 weeks and in the 2-year study, there were dose-dependent and significant increases in hepatic cytotoxicity at 1.0, 2.0, 4.0, and 8.0mg furan/kg. For cell proliferation as measured by 5-bromo-2'-deoxyuridine (BrdU) labeling index (LI), there was a statistically significant trend with increasing dose levels of furan and increased LI at 8.0mg/kg. There was an increased incidence of foci of altered hepatocytes, hepatocellular adenomas, and adenomas or carcinomas at 4.0 and 8.0mg/kg and carcinomas at 8.0mg/kg. The multiplicity of microscopic tumors was increased and latency was decreased in mice exposed to 8.0mg/kg. Prevalence of hepatic nodules at necropsy was increased in mice exposed to 4.0 and 8.0mg/kg. Data demonstrate an association among furan-induced hepatic cytotoxicity, compensatory cell replication, and liver tumor formation in mice; at high doses >or=4.0mg/kg, furan induced hepatotoxicity, compensatory cell replication and tumorigenesis in a dose-related manner, while furan did not produce tumors at cytotoxic doses of 1.0 and 2.0mg/kg.