Overexpression of ABCG1 protein attenuates arteriosclerosis and endothelial dysfunction in atherosclerotic rabbits

Abstract

The ABCG1 protein is centrally involved in reverse cholesterol transport from the vessel wall. Investigation of the effects of ABCG1 overexpression or knockdown in vivo has produced controversial results and strongly depended on the gene intervention model in which it was studied. Therefore, we investigated the effect of local overexpression of human ABCG1 in a novel model of vessel wall-directed adenoviral gene transfer in atherosclerotic rabbits. We conducted local, vascular-specific gene transfer by adenoviral delivery of human ABCG1 (Ad-ABCG1-GFP) in cholesterol-fed atherosclerotic rabbits in vivo. Endothelial overexpression of ABCG1 markedly reduced atheroprogression (plaque size) and almost blunted vascular inflammation, as shown by markedly reduced macrophage and smooth muscle cell invasion into the vascular wall. Also endothelial function, as determined by vascular ultrasound in vivo, was improved in rabbits after gene transfer with Ad-ABCG1-GFP. Therefore, both earlier and later stages of atherosclerosis were improved in this model of somatic gene transfer into the vessel wall. In contrast to results in transgenic mice, over-expression of ABCG1 by somatic gene transfer to the atherosclerotic vessel wall results in a significant improvement of plaque morphology and composition, and of vascular function in vivo.

Keywords: ABCG1; adenoviral vector; atherosclerosis.; gene transfer.

Figure 1

Representative histological sections of the left carotid artery: significant plaque formation in GFP-infected cholesterol-fed rabbits (Ad-GFP) as determined after staining with hematoxylin eosin and elastica van Gieson (VG) staining. Local Ad-ABCG1-GFP infection of the left carotid artery resulted in reduced plaque formation. The scale bar corresponds to a length of 50 µm. The different thicknesses of the intima surrounded by elastic tuna (arrow) in VG elastic staining are shown.

Figure 2

Quantitative assessment demonstrated a significant reduction of macrophages and smooth muscle cells, with a trend towards reduced T-lymphocyte invasion after local Ad-ABCG1-GFP gene transfer (^**indicates P<0.01 compared to Ad-GFP). The means ±SEM of 6 animals are shown.

Figure 3

Transgene expression of the recombinant ABCG1-GFP fusion protein was verified by Western blotting with an anti-ABCG1 antibody, showing a clear signal at the expected size. Lane 1, protein size marker; lane 2, ABCG1-GFP; lane 3, control membrane preparation.

Figure 4

Expression of the reporter gene GFP was detected in the left carotid arteries of all rabbits after gene transfer in vivo with Ad-ABCG1-GFP (co-expressing GFP with ABCG1) and Ad-GFP. Representative histological sections show specific fluorescence in the left carotid artery after infection with Ad-GFP (middle panel) as well as after Ad-ABCG1-GFP (right panel). For comparison, only a slight background fluorescence occurred in the right carotid artery of identical rabbits without local vascular gene transfer (left panel).

Figure 5

Representative en face macroscopic preparations of left carotid arteries from atherosclerotic rabbits. Four weeks after local gene transfer with Ad-ABCG1-GFP plaque formation was vastly reduced compared to those infected with Ad-GFP only. All preparations lack a small 5 mm block in the middle which was taken out for histological analysis.

Figure 6

Quantitative assessment of plaque formation by digital image analysis showed significant reduction after local gene transfer with Ad-ABCG1-GFP. Other vascular sites without gene transfer served as internal controls demonstrating homogenous plaque progression in rabbits after high cholesterol diet. The means ±SEM of 6 animals are shown (^*indicates P<0.05 significance compared to Ad-GFP).

Figure 7

Quantitative assessment of intima and media areas by digital image analysis showed a marked reduction of intima area after local gene transfer with Ad-ABCG1-GFP. The means ±SEM of 6 animals are shown (^*indicates P<0.05 significance compared to Ad-GFP).

Figure 8

Representative images of cell invasion into atherosclerotic plaques, as visualized by brown staining using immunohistochemistry: macrophages. The scale bars correspond to a length of 50 µm. Positive cells are indicated by arrows.

Figure 9

Representative images of cell invasion into atherosclerotic plaques, as visualized by brown staining using immunohistochemistry: smooth muscle cells. The scale bars correspond to a length of 50 µm. Positive cells are indicated by arrows.

Figure 10

Representative images of cell invasion into atherosclerotic plaques, as visualized by brown staining using immunohistochemistry: T lymphocytes. The scale bars correspond to a length of 50 µm. Positive cells are indicated by arrows.

Figure 11

Endothelial dysfunction was investigated in rabbits by vascular ultrasound in vivo by studying the flow-induced vascular reactivity by infusion of a high volume of saline. The effect of Ad-ABCG1-GFP and of control adenovirus (Ad-GFP) in atherosclerotic rabbits was compared to the effect of Ad-GFP in healthy control rabbits. The relative change in diameter of the left carotid artery compared to baseline (%) is demonstrated after 1–2 min (directly after infusion) and 10 min after infusion. ^*indicates P<0.05 compared to Ad-GFP-ABCG1. The means ±SEM of at least 6 animals are shown.