Autonomous parvoviruses preferentially replicate in and kill in vitro-transformed cells and reduce the incidence of spontaneous and implanted tumors in animals. Because of these natural oncotropic and oncolytic properties, parvoviruses deserve to be considered as potential antitumor vectors. Here, we assessed whether parvovirus H1 is able to kill human hepatoma cells by induction of apoptosis but spares primary human liver cells, and whether the former cells can efficiently be transduced by H1 virus-based vectors. Cell death, infectivity, and transgene transduction were investigated in Hep3B, HepG2, and Huh7 cells and in primary human hepatocytes with natural and recombinant H1 virus. All hepatoma cells were susceptible to H1 virus-induced cytolyis. Cell death correlated with H1 virus DNA replication, nonstructural protein expression, and with morphological features of apoptosis. H1 virus-induced apoptosis was more pronounced in p53-deleted Hep3B and p53-mutated Huh7 cells than in HepG2 cells which express wild-type p53. In Hep3B cells, apoptosis was partially inhibited by DEVD-CHO, a caspase-3 inhibitor. In contrast, H1 virus-infected primary hepatocytes were neither positive for nonstructural protein expression nor susceptible to H1 virus-induced killing. Infection with a recombinant parvovirus vector carrying the luciferase gene under control of parvovirus promoter P38 led to higher transgene activities in hepatoma cells than in the hepatocytes. Taken together, H1 virus kills human hepatoma cells at low virus multiplicity but not primary hepatocytes. Thus, recombinant H1 viruses carrying antitumor transgenes may be considered as potential therapeutic options for the treatment of hepatocellular carcinomas.