Gene transfer and expression of apolipoprotein A-I (apoA-I), the major protein component of high density lipoproteins (HDL), is a potentially attractive method for investigating the effects of apoA-I on atherosclerosis. We constructed a second generation recombinant adenovirus encoding the human apoA-I cDNA. This adenoviral vector or a control vector was injected intravenously into apoE-deficient mice fed a chow diet and low density lipoprotein (LDL) receptor (LDLR)-deficient mice fed Western diet, as well as control wild-type C57BL/6 mice. The mean peak plasma human apoA-I concentrations were 235, 324, and 276 mg/dL in apoE-deficient, LDLR-deficient, and wild-type mice, respectively. Human apoA-I concentrations decreased rapidly in apoE-deficient mice and were barely detectable 6 weeks after injection. In contrast, substantially higher levels of human apoA-I were sustained in LDLR-deficient mice. In wild-type mice, human apoA-I levels decreased more rapidly than in LDLR-deficient mice, but could still be detected in plasma for up to 8 months after virus injection. In apoE-deficient mice a substantial fraction of human apoA-I was found associated with triglyceride (TG)-rich lipoproteins; in contrast, in LDLR-deficient and wild-type mice the majority of human apoA-I was found in the HDL fraction. Finally, expression of human apoA-I caused a transient but significant increase in triglyceride levels in all three mouse models. In summary: 1) a second generation recombinant adenovirus resulted in high-level expression of human apoA-I in mice; 2) significantly higher levels of human apoA-I persisted for a longer time in LDLR-deficient mice compared with apoE-deficient mice; and 3) substantial human apoA-I was found associated with TG-rich lipoproteins in apoE-deficient but not LDLR-deficient mice.