Stable gene transfer into hepatocytes might be used to compensate for a genetic deficiency affecting liver function or to deliver diffusible factors into the blood stream. In rats, we have combined retroviral-mediated gene transfer with a surgical procedure in which the liver is temporarily excluded from the circulation and infected in vivo. Partial hepatectomy was performed 24-48 hr before perfusion with virus to induce hepatocyte division and facilitate viral integration. A helper-free recombinant retrovirus coding for P-galactosidase with nuclear localization was used to score cells that expressed the transgene. For at least 3 months after gene transfer, up to 5% of hepatocytes expressed nuclear 3-galactosidase. Whereas in vitro reimplantation of genetically modified hepatocytes has proved to be inefficient in stably transferring genes into the liver, our approach provides a feasible alternative.A meethod for safe, efficient, and stable introduction of foreign DNA into hepatocytes would allow the development of protocols for the genetic treatment of many inborn errors of the metabolism.Replication-defective retroviral vectors ensure genomic integration of a few transgene copies in an unrearranged and permanent configuration transmitted to the cell progeny. The stability and long-term expression of transgenes introduced by retroviral vectors in vivo has been documented in a variety of somatic tissues (1-4). Use of helper-free packaging cell lines (5) has shown that the replication-defective recombinant retrovirus is not transmitted to nontargeted organs (4, 6) and a fortiori to other organisms.Freshly explanted hepatocytes are susceptible to retrovirus infection (7), and strategies for gene transfer into the liver involving in vitro infection of cultured liver cells followed by reimplantation have been proposed. However, whereas fresh hepatocytes can be efficiently engrafted into the spleen (8, 9) or the liver (10), most in vitro cultured and infected cells lose their capacity to be transplanted back (11). Therefore, the feasibility of large-scale gene transfer to the liver by this approach is unlikely. Our purpose was to develop an alternative method whereby the hepatocytes would directly receive the transgene in situ. Because cell division is required for genomic integration of retroviruses and because most hepatocytes are quiescent cells in the adult liver, retroviral infection of the steady-state organ is expected to be inefficient. On the other hand, partial resection of the liver is followed by a regenerative phase that rapidly reconstitutes a full-size organ. Many hepatocytes enter the cell cycle during this process (12, 13) and should be susceptible to retrovirus infection at this time. After complete restoration of the organ weight, hepatocytes return to a quiescent state (12, 13), and prolonged persistence of infected cells may be expected. Therefore, the in situ infection of hepatocytes with recombinant retroviral vectors may be a suitable method for long-term complementation of a genetic de...