Autoantibodies to heat shock protein 60 promote thrombus formation in a murine model of arterial thrombosis

Abstract

Background and objectives: Anti-heat shock protein (HSP)60 autoantibodies are associated with atherosclerosis and are known to affect endothelial cells in vitro. However, their role in thrombus formation remains unclear. We hypothesized that anti-HSP60 autoantibodies could potentiate thrombosis, and evaluated the effect of anti-murine HSP60 antibodies in a ferric chloride (FeCl3)-induced murine model of carotid artery injury.

Methods: Anti-HSP60, or control, IgG was administered to BALB/c mice 48 h prior to inducing carotid artery injury, and blood flow was monitored using an ultrasound probe.

Results: Thrombus formation was more rapid and stable in anti-HSP60 IGG-treated mice than in controls (blood flow=1.7%+/-0.6% vs. 34%+/-12.6%, P=0.0157). Occlusion was complete in all anti-HSP60 IgG-treated mice (13/13), with no reperfusion being observed. In contrast, 64% (9/14) of control mice had complete occlusion, with reperfusion occurring in 6/9 mice. Thrombi were significantly larger in anti-HSP60 IgG-treated mice (P=0.0001), and contained four-fold more inflammatory cells (P=0.0281) than in controls. Non-injured contralateral arteries of anti-HSP60 IgG-treated mice were also affected, exhibiting abnormal endothelial cell morphology and significantly greater von Willebrand factor (VWF) and P-selectin expression than control mice (P=0.0024 and P=0.001, respectively).

Conclusions: In summary, the presence of circulating anti-HSP60 autoantibodies resulted in increased P-selectin and VWF expression and altered cell morphology in endothelial cells lining uninjured carotid arteries, and promoted thrombosis and inflammatory cell recruitment in FeCl3-injured carotid arteries. These findings suggest that anti-HSP60 autoantibodies may constitute an important prothrombotic risk factor in cardiovascular disease in human vascular disease.

Fig. 1

Repeated immunization with mycobacterial heat shock protein (HSP)65 induces a break in tolerance to murine self-HSP60 in BALB/c mice. BALB/c mice were immunized with mycobacterial HSP65 in the presence of complete Freund’s adjuvant (CFA). Sera from immunized mice were evaluated for IgG antibody to (A) mycobacterial HSP65 or (B) murine HSP60. Each point represents the mean of duplicate samples from an individual mouse, and the bar represents the mean binding (OD₄₀₅) for the entire group of mice (n = 10). These data are representative of five independent experiments.

Fig. 2

Anti-heat shock protein (HSP)60 antibodies promote thrombus stabilization in a murine model of carotid injury. (A) Blood flow was monitored (23 min) following application of 6% FeCl₃ to the carotid arteries of mice injected with anti-HSP60 IgG or control IgG. Initial blood flow through the carotid artery was considered to be maximal (100%). Points represent the mean value for each group of mice, and bars represent standard error of the mean (SEM). The graph inset represents blood flow (mean and SEM) during the thrombus stabilization phase for mice treated with phosphate-buffered saline, control IgG, or anti-HSP60 IgG. (B) The time required to achieve 20%, 50% or 80% occlusion was determined from the curves in (A), and the mean time and SEM for each group are indicated (50% and 80%occlusion occurred significantly faster in anti-HSP60 IgG-treated mice than in control IgG-treated mice).

Fig. 3

Anti-heat shock protein (HSP)60 antibodies promote the generation of larger thrombi. Representative histologic transverse sections of carotid arteries injured with 6%FeCl₃, stained with Verhoeff, illustrate the size of the thrombus and the structure of the elastic laminae of the vessels. The size of the thrombi in the carotid arteries of anti-HSP60 IgG-treated mice (B, D) was visually greater than that of those in control IgG-treated mice (A, C). In addition, the elastic fibers were stretched and unfolded in the anti-HSP60 IgG-treated mice (B, D), whereas those isolated from the control IgG-treated group (A, C) appeared normal, as indicated by the wavy folded structure of the elastic laminae. (E) The cross-sectional area of the thrombus was measured by computer-assisted planimetry, and the thrombus area for each carotid artery was calculated inmm² by standard formulae. The mean thrombus size and standard error of the mean for each group are indicated. All data are representative of four mice from each group. (C) and (D) show the segment of the artery [indicated by the square in (A) and (B), respectively] at higher magnification [(A, B) original magnification, ×20; (C, D) original magnification, ×100]. Scale bars: 0.05 mm.

Fig. 4

Anti-heat shock protein (HSP)60 antibodies favor inflammatory cell recruitment to the site of the thrombus. Representative histologic transverse sections of carotid arteries injured with 6% FeCl₃, stained with hematoxylin and eosin, illustrate the structure of the thrombus and the presence of inflammatory cells. Few inflammatory cells were detected in the thrombi from control IgG-treated mice (A, C), as compared to a larger number of clustered cells in the thrombi of anti-HSP60 IgG-treated mice (B, D, E). (F) The number of inflammatory cells recruited into the thrombi was evaluated by counting the number of nuclei in a transverse section of the thrombus, using a computer-assisted enumeration method. The mean value and standard error of the mean for each group are indicated. All data are representative of four mice from each group. (D) and (E) show the segments of the artery [indicated by the squares on the right and left, respectively, in (B)] at higher magnification [(A, B) original magnification, ×20; (C, D, E) original magnification, ×40]. Scale bars: 0.05 mm.

Fig. 5

Anti-heat shock protein (HSP)60 antibodies induce endothelial von Willebrand factor (VWF) overexpression. Representative histologic transverse sections of intact carotid arteries, contralateral to the arteries injured with 6%FeCl₃, were stained for VWF. VWF detection is represented by dark brown staining on the surface of endothelial cells [examples of positive cells are indicated by arrows in (C) and (D)]. Note that VWF staining is more widespread and intense in intact carotid arteries from anti-HSP60 IgG-treated mice (B, D) than in intact carotid arteries from control IgG-treated mice (A, C). (E) The percentage of positive staining for VWF was evaluated by determining the ratio of the endothelial surface stained with anti-VWF (mm²) to the total endothelial surface (mm²), using computer-assisted technology. The mean value and standard error of the mean for each group are indicated. All data are representative of four mice from each group. (C) and (D) show the segments of the artery [indicated by the squares in (A) and (B), respectively] at higher magnification [(A, B) original magnification, ×20; (C, D) original magnification, ×40]. Scale bars: 0.05 mm.

Fig. 6

Anti-heat shock protein (HSP)60 antibodies induce endothelial P-selectin overexpression. Representative histologic transverse sections of intact carotid arteries, contralateral to the arteries injured with 6%FeCl₃, were stained for P-selectin. P-selectin detection is represented by dark brown staining on the surface of endothelial cells [examples of positive cells are indicated by arrows in (C) and (D)]. Note that P-selectin staining is more widespread and intense in intact carotid arteries from anti-HSP60 IgG-treated mice (B, D) than in intact carotid arteries from control IgG-treated mice (A, C). (E) The percentage of positive staining for P-selectin was evaluated by determining the ratio of the endothelial surface stained with anti-P-selectin (mm²) to the total endothelial surface (mm²), using computer-assisted technology. The mean value and standard error of the mean for each group are indicated. All data are representative of four mice from each group. (C) and (D) show the segments of the artery [indicated by the squares in (A) and (B), respectively] at higher magnification [(A, B) original magnification, ×20; (C, D) original magnification, ×40]. Scale bars: 0.05 mm.