To develop a primate model for liver-directed gene therapy, we studied several gene transfer vehicles and routes in eight rhesus monkeys (Macaca mulatta). For this purpose, we used first-generation, replication-deficient adenoviral vectors carrying the Escherichia coli lacZ gene (Ad.CMVlacZ) or a lacZ-containing plasmid (pCMV beta) with lipofectamine for transfection. The reporter gene construct was infused into either the portal vasculature, common bile duct, or saphenous vein. Adenovirus-mediated gene transfer via the portal vein resulted in expression of lacZ in over 70% of hepatocytes by days 3-7, but was accompanied by acute hepatitis. Adenovirus-mediated gene transfer via the common bile duct resulted in lacZ expression in less than 10% of hepatocytes and was accompanied by portal inflammation. The animals mounted a significant immune response, as demonstrated by adenoviral antigen-induced T-cell proliferation and production of neutralizing anti-adenovirus antibodies and antibodies to E. coli beta-galactosidase (beta-Gal). Activation of the immune response was associated with rapid decrease of the reporter gene by days 13-21. Lipofectamine-mediated gene transfer was inefficient, and no lacZ expression in the liver was detected. To limit the host immune response, 4 animals were immunosuppressed by cyclophosphamide/prednisone and then infused with the Ad.CMVlacZ via the portal vein or the saphenous vein. The monkeys showed sustained expression of lacZ for up to 35 days with no evidence of inflammation. The primates transduced via the saphenous vein showed a level of beta-Gal expression in the liver similar to that of the portal vein-infused animals. In conclusion, adenovirus-mediated gene transfer to non-human primate livers via the portal vein or saphenous vein is efficient, but it results in transient expression and is accompanied by an immune response to both vector and transgene products and acute hepatitis, whereas lipofectamine-mediated transfer is inefficient. Manipulation of the host immune response may expand potential applications of adenoviral vectors for liver-directed gene transfer.