Connexin (Cx) 26 and 32 are the major gap junction proteins in liver. We recently demonstrated that Cx32 is essential for phenobarbital (PB)-mediated tumor promotion in mouse liver. To investigate whether Cx26 plays a similar role, an initiation-promotion experiment was conducted using mice with a liver-specific knockout of Cx26. Control and Cx26-deficient mice were injected a single dose of N-nitrosodiethylamine (DEN, 90 microg/g b.wt.) at 6 weeks of age and groups of mice were subsequently kept on a PB (0.05%) containing or control diet for 35 weeks. At the end of the experiment, the carcinogenic response in the liver was monitored. Mice from PB treatment groups showed strongly increased liver weights compared with mice treated with DEN alone, which was mostly due to a much higher tumor burden. The tumor response in PB-treated mice of both strains was quite similar, but the number of smaller tumors and of enzyme-altered neoplastic lesions was somewhat larger in PB-treated Cx26 knockout (Cx26 KO) compared with wild-type mice, whereas the volume fraction of enzyme-altered lesions was slightly reduced in PB-treated Cx26-deficient mice. There was no significant difference in tumor prevalence between Cx26 KO and wild-type mice. Altogether our present data show that elimination of Cx26 has only minor effects on chemically induced mouse hepatocarcinogenesis, in striking contrast to the effects seen in Cx32 KO mice.