Injecting a virus that carries the gene for human alpha‑1 antitrypsin (AAT) into the liver of mice can make the animals produce therapeutic levels of the protein in their blood, using far less virus than when the same gene is injected into muscle. The liver cells take up the gene, keep it as a circular DNA piece, and release the protein without harming the liver.

Previous work from our group showed that recombinant adeno-associated virus (rAAV) vectors mediated long-term secretion of therapeutic serum levels of human alpha-1 antitrypsin (hAAT) after a single injection in murine muscle. We hypothesized that hepatocyte transduction could be even more efficient, since these cells represent the natural site of AAT production and secretion. To test this hypothesis, rAAV vectors containing the hAAT cDNA driven by either the human elongation factor 1 alpha promoter, the human cytomegalovirus immediate-early promoter (CMV), or the CMV-chicken beta actin hybrid (CB) promoter were injected into the portal or tail veins of adult C57Bl/6 mice. Potentially therapeutic serum levels of hAAT (600 microg/ml) were achieved after portal vein injection of doses of 4 x 10(9) infectious units (IU), a 10-fold lower dose than that required for similar levels of expression via the i.m. route. Serum levels greater than 1 mg/ml were achieved at doses of 3 x 10(10) IU. Southern blotting of liver DNA revealed the presence of circular episomal vector genomes. Immunostaining showed that transgene expression was scattered throughout the liver parenchyma. Similar results were obtained with a rAAV-CB-green fluorescent protein (GFP) vector. There was no evidence of hepatic toxicity. These data indicate that liver-directed rAAV-based gene therapy is effective in the murine model, and hence might be feasible for treatment of human AAT deficiency.