αvβ3 Integrins Mediate Flow-Induced NF-κB Activation, Proinflammatory Gene Expression, and Early Atherogenic Inflammation

Abstract

Endothelial cell interactions with transitional matrix proteins, such as fibronectin, occur early during atherogenesis and regulate shear stress-induced endothelial cell activation. Multiple endothelial cell integrins bind transitional matrix proteins, including α5β1, αvβ3, and αvβ5. However, the role these integrins play in mediating shear stress-induced endothelial cell activation remains unclear. Therefore, we sought to elucidate which integrin heterodimers mediate shear stress-induced endothelial cell activation and early atherogenesis. We now show that inhibiting αvβ3 integrins (S247, siRNA), but not α5β1 or αvβ5, blunts shear stress-induced proinflammatory signaling (NF-κB, p21-activated kinase) and gene expression (ICAM1, VCAM1). Importantly, inhibiting αvβ3 did not affect cytokine-induced proinflammatory responses or inhibit all shear stress-induced signaling, because Akt, endothelial nitric oxide synthase, and extracellular regulated kinase activation remained intact. Furthermore, inhibiting αv integrins (S247), but not α5 (ATN-161), in atherosclerosis-prone apolipoprotein E knockout mice significantly reduced vascular remodeling after acute induction of disturbed flow. S247 treatment similarly reduced early diet-induced atherosclerotic plaque formation associated with both diminished inflammation (expression of vascular cell adhesion molecule 1, plaque macrophage content) and reduced smooth muscle incorporation. Inducible, endothelial cell-specific αv integrin deletion similarly blunted inflammation in models of disturbed flow and diet-induced atherogenesis but did not affect smooth muscle incorporation. Our studies identify αvβ3 as the primary integrin heterodimer mediating shear stress-induced proinflammatory responses and as a key contributor to early atherogenic inflammation.

Figure 1

Inhibition of αv but not α5 integrins blunts shear stress-induced NF-κB activation. A: Human aortic endothelial cells on fibronectin were treated with 50 μmol/L ATN-161 or 1 μmol/L S247 and exposed to acute onset of flow for 30 minutes. Activation of eNOS, FAK, PAK2, and NF-κB was assessed by Western blot analysis with phospho-specific antibodies and normalized to total protein levels. Representative blots are shown. Quantification of shear-induced activation of FAK (B), PAK2 (C), NF-κB (D), and ERK1/2 (E) is shown. n = 4 (B and E); n = 5 (C); n = 7 to 11 (D). ^∗∗P < 0.01, ^∗∗∗P < 0.001. eNOS, endothelial nitric oxide synthase; ERK, extracellular regulated kinase; FAK, focal adhesion kinase; IB, immunoblot; NT, no treatment; PAK2, p21-activated kinase 2.

Figure 2

Selective knockdown of αv integrins blunts shear stress-induced NF-κB activation. A: HAECs were transfected with siRNA for α5 or αv integrins, plated on fibronectin, and exposed to acute onset of shear stress for 30 minutes. Integrin expression was determined by Western blot analysis and shear-induced activation of eNOS, FAK, PAK2, and NF-κB was assessed by Western blot analysis with phospho-specific antibodies and normalized to total protein levels. Representative blots are shown. Quantification of shear-induced activation of FAK (B), PAK2 (C), NF-κB (D), and ERK1/2 (E) is shown. F and G: HAECs were treated as in A–E, and NF-κB activation was determined by immunocytochemistry for NF-κB nuclear translocation. Representative micrographs are shown. F: Quantification of nuclear translocation as described in E. HAECs were scored for nuclear NF-κB staining, at least 100 cells were counted per condition for each experiment. n = 5 (B); n = 8 (C–E); n = 4 (F). ^∗P < 0.05, ^∗∗P < 0.01, and ^∗∗∗P < 0.001. eNOS, endothelial nitric oxide synthase; ERK, extracellular regulated kinase; FAK, focal adhesion kinase; HAEC, human aortic endothelial cell; IB, immunoblot; NT, no treatment; PAK2, p21-activated kinase 2; Pos, positive.

Figure 3

αv Integrin knockdown inhibits shear stress-induced ICAM-1/VCAM-1 expression. HAECs were transfected with siRNA for α5 or αv integrins, plated on fibronectin, and exposed to laminar shear stress for 3 hours. VCAM-1 (A) and ICAM-1 (B) mRNA was analyzed by quantitative real-time PCR and normalized to β2-microglobulin expression. C–F: HAECs transfected with α5 or αv siRNA were plated on fibronectin and exposed to either laminar shear stress or 10 ng/mL TNFα for 5 hours. Cells were lyzed and analyzed for VCAM-1 (E) or ICAM-1 (F) expression by Western blot analysis. Representative blots are shown (C and D). G and H: HAECs transfected with α5 or αv siRNA were plated on fibronectin and exposed to OSS for 18 hours. Cells were lyzed and analyzed for VCAM-1 (G) or ICAM-1 (H) expression by Western blot analysis. n = 4 (B, G, and H); n = 4 to 7 (C and D). ^∗P < 0.05, ^∗∗P < 0.01, and ^∗∗∗P < 0.001. ERK, extracellular regulated kinase; Exp, expression; HAEC, human aortic endothelial cell; ICAM-1, intercellular adhesion molecule-1; M, mock; NT, no treatment; OSS, oscillatory shear stress; TNFα, tumor necrosis factor-α; VCAM-1, vascular cell adhesion molecule-1.

Figure 4

β3 siRNA limits shear-induced proinflammatory signaling and gene expression. A and B: HAECs were transfected with siRNA for β3 or β5 integrins, plated on a fibronectin and vitronectin matrix, and exposed to acute onset of flow for 30 minutes. NF-κB phosphorylation was determined by Western blot analysis. Representative images are shown. B: Quantification of NF-κB phosphorylation from A normalized to total NF-κB levels. HAECs treated as in A were exposed to laminar shear stress for 5 hours, and expression of VCAM-1 (C) and ICAM-1 (D) was determined by Western blot analysis. HAECs treated as in A were exposed to OSS for 18 hours and expression of VCAM-1 (E) and ICAM-1 (F) was determined by Western blot analysis. n = 4 (B–F). ^∗P < 0.05, ^∗∗P < 0.01, and ^∗∗∗P < 0.001. Exp, expression; HAEC, human aortic endothelial cell; ICAM-1, intercellular adhesion molecule-1; M, mock; OSS, oscillatory shear stress; VCAM-1, vascular cell adhesion molecule-1.

Figure 5

S247 limits inflammation after partial carotid ligation in ApoE knockout mice. A and B: ApoE knockout mice underwent partial ligation of their left carotid artery and treatment with saline, ATN-161 (5 mg/kg every third day), or S247 (40 mg/kg per day). After 7 days, plaque area in the left carotid artery exposed to disturbed flow was determined by immunohistochemistry for Mac2 (green; A) and SMA (red; A) or VCAM-1 (B) by using nuclear staining (DAPI, blue) as a counterstain. Representative micrographs from serial sections are shown. C: Quantification of total plaque area. D: Quantification of Mac2-positive and SMA-positive area. E: VCAM-1 intensity in the left carotid artery was quantified and normalized to VCAM-1 in the right carotid artery exposed to laminar shear stress. n = 5 to 9 mice per group (A, B, and E). ^∗P < 0.05. ApoE, apolipoprotein E; Mac2, macrophage; SMA, smooth muscle area; VCAM-1, vascular cell adhesion molecule-1.

Figure 6

S247 limits spontaneous atherosclerosis in Western diet-fed Apoe knockout mice. Apoe knockout mice were fed a high-fat, Western diet for 8 weeks and treated with saline or 40 mg/kg per day S247 for the final 4 weeks. A: Oil Red O staining of the aortic arch was performed to visualize plaque formation by en face imaging. Quantification of Oil Red O-positive areas in the aortic arch or whole aorta from the aortic cusp to renal branchpoints is shown. B and C: Analysis of atherosclerotic plaque area in the aortic root by Russell-Movat Pentachrome staining. Representative images are shown with insets of the entire vessel. Quantification of plaque cross-sectional area determined at multiple regions along the aortic root (C). D: Representative micrographs of plaques from the aortic root were analyzed for Mac2 area (green) and SMA (red), using nuclear staining (DAPI, blue) as a counterstain. E: Quantification of Mac2-positive area. F: Quantification of SMA-positive. Percentage of the atherosclerotic plaque that stained positive for either Mac2 (G) or SMA (H). I: Plaques were scored as either positive or negative for SMA-containing fibrous caps and the percent positive are shown with a pie graph. J: Scatter plots for the relation between SMA and plaque area. The slopes showing the relation between plaque size and SMA are shown. n = 5 mice per group. n = 23 to 30 plaques from 4 to 5 mice per condition (I). ^∗P < 0.05. Original magnification: ×40 (B and C, main images); ×10 (insets). Mac2, macrophage; SMA, smooth muscle area.

Figure 7

Reduced inflammation after partial carotid ligation in iEC-αv KO mice. A: After tamoxifen treatment, lung endothelial cells were isolated to analyze αv protein levels by Western blot analysis. B: iEC-Control and iEC-αv KO mice underwent partial carotid ligation, and the intimal mRNA was isolated after 48 hours. Expression of Klf2, ICAM-1, and VCAM-1 was determined by quantitative real-time PCR. C and D: After 7 days, plaque area in the left carotid artery was determined by immunohistochemistry for Mac2 (green, C) and SMA (red, C) or VCAM-1 (D) by using nuclear staining (DAPI, blue) as a counterstain. Representative micrographs from serial sections are shown. E: Quantification of total plaque area. F: Quantification of Mac2-positive and SMA-positive area. G: VCAM-1 intensity in the left carotid artery was quantified and normalized to VCAM-1 in the right carotid artery exposed to laminar shear stress. n = 4 mice per group (A), n = 5 mice per group (B–D and G). ^∗P < 0.05, ^∗∗P < 0.01. Original magnification: ×40 (C and D); ×10 (C and D, insets). Ctrl, control; IB, immunoblot; ICAM-1, intercellular adhesion molecule-1; iEC, inducible epithelial cell; Klf2, Krüppel-like Factor 2; KO, knockout; LC, left carotid; Mac2, macrophage; RC, right carotid; SMA, smooth muscle area; Tub, tubulin; VCAM-1, vascular cell adhesion molecule-1.

Figure 8

Endothelial αv deletion limits spontaneous atherosclerosis. iEC-Control and iEC-αv KO mice were treated with tamoxifen, and atherosclerosis was induced by Western diet feeding for 8 weeks. A: Oil Red O staining of the aortic arch was performed to visualize plaque formation by en face imaging. Quantification of Oil Red O-positive areas in the aortic arch or whole aorta from the aortic cusp to renal branchpoints is shown. B: Representative images show analysis of atherosclerotic plaque area in the aortic root by Russell-Movat Pentachrome staining. C: Quantification of plaque cross-sectional area determined at multiple regions along the aortic root. D: Representative micrographs of the plaques from the aortic root were analyzed for Mac2 (green) and SMA (red) areas, using nuclear staining (DAPI, blue) as a counterstain. E: Quantification of Mac2-positive area. F: Quantification of SMA-positive. Percentage of the atherosclerotic plaque staining positive for either Mac2 (G) or SMA (H). I: Plaques were scored as either positive or negative for SMA-containing fibrous caps and the percent positive is shown with a pie graph. J: Scatter plots for the relation between SMA and plaque area. The slopes show the relation between plaque size and smooth muscle. n = 7 mice per group; n = 24 to 39 plaques from 6 to 7 mice per condition (I and J). ^∗P < 0.05, ^∗∗P < 0.01. Original magnification: ×40 (main images); ×10 of the entire vessel (inset). iEC, inducible epithelial cell; KO, knockout; Mac2, macrophage; Pos, positive; SMA, smooth muscle area.