Mutations in the glycogen synthase kinase 3beta (GSK3beta) phosphorylation sites of the beta-catenin gene exon 3 are found in 20-30% of human primary hepatocellular carcinoma (HCC), whereas mutations in the APC or AXIN genes are found in other HCC populations. These data strongly suggest that the Wnt signaling pathway is involved in hepatocarcinogenesis. To determine the role of beta-catenin in intestinal tumorigenesis, we earlier constructed a mutant mouse strain Catnb(lox(ex3)), in which exon 3 of the beta-catenin gene was sandwiched by loxP sequences. By genetic crosses of these mice with the Fabpl-cre transgenic mice that express the cre gene controlled by the fatty acid binding protein gene promoter, we introduced the beta-catenin stabilizing mutation into the small intestine and liver. Although numerous polyps were formed in the small intestine, we did not find any neoplastic (i.e., dysplastic) foci in the liver, and the mice died in 5 weeks after birth because of acute liver damage accompanying mitochondrial swelling. When a recombinant adenovirus that expresses the cre gene from a human cytomegalovirus early gene promoter was constructed and inoculated at a high multiplicity (10(9) plaque-forming units/mouse), the Catnb(lox(ex3)) mice showed marked hepatomegaly, with similar mitochondrial swelling in the hepatocytes, and died within 3 weeks after infection. On the other hand, when inoculated at lower multiplicities of infection (10(7) and 10(8) plaque-forming units/mouse, respectively), the Catnb(lox(ex3)) mice survived >6 months without any neoplastic foci in the liver, although the nuclear localization of beta-catenin was found in some hepatocytes even after 6 months. These results suggest that, in contrast to intestinal polyposis, the Wnt pathway activation by stabilized beta-catenin is not sufficient for hepatocarcinogenesis, but additional mutations or epigenetic changes may be required.