Recent studies have shown that nondividing primary cells, such as hepatocytes, can be efficiently transduced in vitro by human immunodeficiency virus-based lentivirus vectors. Other studies have reported that, under certain conditions, the liver can be repopulated with transplanted hepatocytes. In the present study, we combined these procedures to develop a model system for ex vivo gene therapy by repopulating rat livers with hepatocytes and hepatoblasts transduced with a lentivirus vector expressing a reporter gene, green fluorescent protein (GFP). Long-term GFP expression in vivo (up to 4 months) was achieved when the transgene was driven by the liver-specific albumin enhancer/promoter but was silenced when the cytomegalovirus (CMV) enhancer/promoter was used. Transplanted cells were massively amplified ( approximately 10 cell doublings) under the influence of retrorsine/partial hepatectomy, and both repopulation and continued transgene expression in individual cells were documented by dual expression of a cell transplantation marker, dipeptidyl peptidase IV (DPPIV), and GFP. In this system, maintenance or expansion of the transplanted cells did not depend on expression of the transgene, establishing that positive selection is not required to maintain transgene expression following multiple divisions of transplanted, lentivirus-transduced hepatic cells. In conclusion, fetal hepatoblasts (liver stem/progenitor cells) can serve as efficient vehicles for ex vivo gene therapy and suggest that liver-based genetic disorders that do not shorten hepatocyte longevity or cause liver damage, such as phenylketonuria, hyperbilirubinemias, familial hypercholesterolemia, primary oxalosis, and factor IX deficiency, among others, might be amenable to treatment by this approach.