Direct gene transfer into localized arterial segments can be performed in vivo by transfection with DNA–liposome complexes. This technique holds promise for the treatment of human diseases, including malignancy and cardiovascular disorders. We have previously characterized the potential toxicity of this form of treatment in mice in vivo (Stewart et al., 1992). In this report, we examine two issues relevant to long-term expression of foreign recombinant genes: (i) the potential for autoimmune damage to major organs and (ii) DNA localization in gonadal tissue. Autoimmunity and toxicity of allogeneic major histocompatibility (MHC) gene transfer was assessed in mice after induction of an immune response to a recombinant murine class I MHC gene by direct gene transfer in vivo. Histological examination of brain, heart, lung, liver, kidney, spleen, and skeletal muscle revealed no clinically significant immunopathology or organ damage. The toxicity of gene delivery by DNA liposomes was also analyzed in pigs and rabbits in vivo. No histopathology was observed following the introduction of plasmids encoding several different gene products, and analysis of serum following DNA liposome delivery revealed no abnormalities of serum biochemical parameters. The potential for transfer of recombinant DNA into testes and ovary in animals was evaluated by the polymerase chain reaction. Although evidence of recombinant plasmid was consistently observed in transfected, but not untransfected, arterial sites and occasionally in lung, kidney, spleen, and liver, no plasmid DNA was detected in testes or ovary. These studies suggest that uptake of recombinant DNA following direct gene transfer by liposomal transfection in major organs is not associated with autoimmunity, toxicity, or gonadal localization. This method of direct gene transfer is therefore appropriate for human gene therapy. The use of DNA–liposome complexes to transfer genes in vivo is now underway in an active clinical protocol. Mouse safety and acute toxicity data have been published previously (Hum. Gene Ther. 3, 267–275). In the present paper, Nabel et al. begin to provide long-term safety data, specifically with regard to the potential for autoimmune damage to major organs or for the transfer of the foreign genes into reproductive tissue.