In order to test the controversially discussed hypothesis that oval cells are part of a liver stem cell compartment and can give rise to cholangiocellular as well as hepatocellular carcinomas in the course of liver carcinogenesis, we transfected an oval cell line established in our laboratory with an oncogenically activated genomic Ha-ras clone (pUC EJ 6.6), carrying a valine at position 12 instead of the wild-type glycine, or a rat p53 cDNA mutated by site-directed mutagenesis at codon 247, which corresponds to codon 249 in the human p53. This codon is of particular interest since it represents a mutation hotspot observed in hepatocellular carcinoma especially in regions with high aflatoxin B1 exposure. Independent Ha-rasVal12 and p53Ser247 recombinant clones were subcutaneously injected into syngeneic newborn rats and the resulting tumours were analysed histopathologically. Each of two p53Ser247 clones gave negligible tumour yields (one tumour out of 13 injected animals), whereas each of two Ha-rasVal12 clones gave marked tumour yields (four tumours out of 13 and seven out of 12 treated animals, respectively). In addition, the p53Ser247-induced tumours appeared only after 11 months and were small, whereas the Ha-rasVal12-induced tumours appeared already after 6-8 weeks and grew rapidly. Histopathological analysis of the tumours revealed only undifferentiated carcinomas. Interestingly, one tumour that arose upon injection of Ha-rasVal12-transfected cells stained positive for albumin, showing at least a partial hepatocytic differentiation.