In VitroandIn VivoHepatoma Cell-Specific Expression of a Gene Transferred with an Adenoviral Vector

Abstract

Recombinant adenoviruses are widely used for the transfer of foreign genes into various mammalian cells. However, the utilization of these vectors for cancer gene therapy requires the specific and efficient expression of the transferred gene in tumor cells. To obtain targeted expression in hepatoma cells, we constructed recombinant adenoviral vectors containing transcriptional elements from either the rat α-fetoprotein (AFP) or the human insulin-like growth factor II (IGFII) genes driving expression of the nuclear β-galactosidase gene (nls lacZ). In vitro infection revealed that the AFP but not the IGFII transcriptional regulatory sequence controlled nls lacZ expression specifically in hepatoma cells. The same specificity was obtained in vivo in subcutaneous human hepatic tumors generated by engraftment of Huh7 hepatoma cells in nude mice as well as in primary liver tumors developed in rats and mice. No marker gene expression was detectable after AFP-nls lacZ gene transfer to normal rat liver parenchyma despite evidence for the presence of DNA encoding the nls lacZ gene. However, in vivo experiments with primary liver tumors in rats and mice also revealed that primary hepatoma cells were poorly infected by adenoviral vectors. Peritumoral and normal tissues were infected efficiently by adenoviral vectors. We conclude that hepatoma cell-specific expression of a transgene can be achieved with AFP regulatory sequences but that adenoviral vectors may not be the preferable vector for transferring genes in vivo in primary liver tumors. The aim of the present study was to design adenoviral vectors able to express a transgene specifically and with high efficiency in primary hepatocellular carcinoma. We demonstrate that the α-fetoprotein but not the insulin-like growth factor II regulatory region was able to achieve correct specific expression in liver cancer cells in vitro as well as in vivo. However, we also present evidence that adenoviral vectors, although able to infect normal hepatocytes or liver tumor cells transplanted in nude mice, were poorly infectious for primary liver tumors in vivo. These results emphasize the need for relevant in vivo models of primary tumors to test the efficiency of the vectors and better characterize basic aspects of targeted gene transfer to malignant tumors.