Background—The ability of apolipoprotein (apo)A-I to induce regression of preexisting atherosclerotic lesions has not been determined, and a mouse model of atherosclerosis regression has not yet been reported.

Methods and Results—LDL receptor–deficient mice were fed a western-type diet for 5 weeks to induce atherosclerotic lesions. A second-generation recombinant adenovirus encoding human apoA-I or a control adenovirus were injected intravenously in order to express apoA-I in the liver. Three days after injection, total apoA-I levels in mice injected with the apoA-I–expressing adenovirus were 216±16.0 mg/dL, compared with 68.0±3.0 mg/dL in control virus–injected mice (P<0.001). HDL cholesterol levels in mice injected with the AdhapoA-I vector 7 days after injection were 189±21.0 mg/dL, compared with 123±8.0 mg/dL in control virus-injected mice (P<0.02). Total and non-HDL cholesterol levels did not differ between the 2 groups. Atherosclerotic lesion area was quantified by en face analysis of the aorta and cross-sectional analysis of the aortic root. Compared with baseline mice, atherosclerosis progressed in mice injected with the control adenovirus. In contrast, in mice expressing apoA-I compared with baseline mice, total en face aortic lesion area was reduced by 70% and aortic root lesion was reduced by 46%. Expression of apoA-I was associated with a significant reduction in the fraction of lesions occupied by macrophages and macrophage-derived foam cells.

Conclusions—Liver-directed gene transfer of human apoA-I resulted in significant regression of preexisting atherosclerotic lesions in LDL receptor–deficient mice as assessed by 2 independent methods.