The GRK2/AP-1 Signaling Axis Mediates Vascular Endothelial Dysfunction and Atherosclerosis Induced by Oscillatory Low Shear Stress

Abstract

Disturbed blood flow and the resulting oscillatory low shear stress (OSS) are key contributors to vascular endothelial dysfunction and the initiation of atherosclerosis. However, the molecular mediators that translate abnormal hemodynamic signals into pathological vascular endothelial responses remain unclear. G protein-coupled receptors (GPCRs) are classical mechanosensors in the vascular endothelium. Here, using vascular endothelial-specific knockout mice, in vitro parallel plate flow chamber systems, and phosphoproteomic analysis, G protein-coupled receptor kinase 2 (GRK2) is identified as a central mediator of OSS-induced vascular endothelial dysfunction. Mechanistically, OSS promotes GRK2 phosphorylation at serine 29, which subsequently activates the transcription factor activator protein-1 (AP-1), increasing the expression of the proinflammatory adhesion molecules intercellular cell adhesion molecule-1 (ICAM1) and vascular cell adhesion molecule 1 (VCAM1). In parallel, AP-1 promotes nuclear receptor subfamily 4 group A 1 (NR4A1) transcription, which anchors liver kinase B1 (LKB1) to the nucleus and suppresses downstream AMP-activated protein kinase (AMPK) signaling, leading to metabolic dysregulation and impaired vascular endothelial homeostasis. These findings underscore the GRK2/AP-1 signaling axis as a crucial mechanotransduction cascade linking disturbed flow to vascular endothelial dysfunction. Given the important role of GPCRs in mechanotransduction, targeting GRK2 may offer a novel therapeutic approach for atherosclerosis.

Keywords: G protein‐coupled receptor kinase 2; activator protein 1; atherosclerosis; shear stress; vascular endothelial dysfunction.

Figure 1

GRK2 activity is elevated in vascular endothelial cells exposed to disturbed flow. a) Schematic illustrating blood flow in the thoracic aorta and the inner curvature of the aortic arch. b) Enface immunofluorescence staining of GRK2^S29p (red), von Willebrand factor (vWF, endothelial marker, green), and DAPI (blue) in the endothelial cells of the thoracic aorta and the inner curvature of the aortic arch isolated from C57BL/6J mice. c) Quantification of the relative GRK2^S29p fluorescent intensity (n = 5, Student's t test). d) Schematic figure of the construction of a partial LCA ligation animal model to simulate disturbed flow and OSS in vivo. High‐resolution Doppler ultrasound imaging confirmed the successful model establishment, with significantly reduced flow velocity and evidence of diastolic flow reversal in the ligated LCA. Red arrows highlight diastolic flow reversal. The right carotid artery (RCA) served as an internal control (n = 5, Student's t test). e) Immunofluorescence staining for GRK2^S29p (red), vWF (yellow), and DAPI (blue) in the RCA endothelium under laminar flow and the LCA endothelium under disturbed flow. f) Quantification of the relative GRK2^S29p fluorescent intensity (n = 5, Student's t test). g) Immunofluorescence staining for F4/80 (red) and DAPI (blue) reflecting the infiltration of macrophages in the RCA and the LCA. h) Quantification of the relative F4/80 fluorescent intensity (n = 5, Student's t test). i) Schematic figure of the in vitro parallel plate flow chamber system, in which endothelial cells were under LSS (15 dyne cm^{− 2}, 0 Hz) and OSS (±2 dyne cm^{− 2}, 1 Hz) conditions. j) HUVECs were subjected to OSS for 0, 5, 15, 30, 60, or 120 min. The proteins in the cell lysate were analyzed via Western blotting with the indicated antibodies. k) Quantification of GRK2^S29p expression levels (n = 5, one‐way ANOVA). l–o) GRK2^S29p expression was elevated in HUVECs (l, n) and HAECs (m, o), as was the expression of ICAM1 and VCAM1 under OSS conditions compared with LSS conditions.

Figure 2

Elevated GRK2^S29p expression in the endothelium of atherosclerotic lesions in human carotid arteries and aortic arteries of ApoE^‐/‐ hyperlipidemic mice. a) Immunofluorescence staining of GRK2^S29p (red), vWF (green), and DAPI (blue) in human carotid arteries with or without atherosclerotic plaques. b) Quantification of the relative GRK2^S29p fluorescent intensity (n = 3, Student's t test). c) Immunofluorescence staining for GRK2^S29p (red), vWF (green), and DAPI (blue) in aortas isolated from ApoE^−/− mice fed a normal diet and those fed a high‐fat Western diet. d) Quantification of the relative GRK2^S29p fluorescent intensity (n = 5, Student's t test). e,f) HUVECs (e) and HAECs (f) were treated with oxidized low‐density lipoprotein (x‐LDL, 50 µg mL⁻¹ for 48 h), and the proteins in the cell lysate were evaluated via Western blotting with the indicated antibodies. g) Quantification of GRK2^S29p, ICAM1, and VCAM1 protein expression levels (n = 5, Student's t test).

Figure 3

GRK2 knockout alleviates OSS‐induced vascular endothelial dysfunction and atherosclerosis. a,b) Immunofluorescence staining for ICAM1 (red), VCAM1 (green), vWF (yellow) and DAPI (blue) in partially ligated left common carotid arteries isolated from GRK2^flox/flox mice and GRK2^ECKO mice. (b) Quantification of the relative ICAM1 and VCAM1 fluorescent intensity (n = 5, Student's t test). c,d) Immunofluorescence staining for F4/80 (red) and DAPI (blue) indicating the infiltration of macrophages in partially ligated left common carotid arteries isolated from GRK2^flox/flox mice and GRK2^ECKO mice. (d) Quantification of the relative F4/80 fluorescent intensity (n = 5, Student's t test). e) Arterial tissues were isolated from GRK2^flox/flox; ApoE^‐/‐ mice and GRK2^ECKO; ApoE^‐/‐ mice fed a high‐fat Western diet, and their atherosclerotic lesions were examined via Oil Red O staining. f) Quantification of atherosclerotic lesions (n = 12, Student's t test). g) Aortic roots isolated from GRK2^flox/flox; ApoE^‐/‐ mice and GRK2^ECKO; ApoE^‐/‐ mice fed a high‐fat Western diet were sectioned for H&E staining and Oil Red O staining to examine atherosclerotic lesions. h,i) Quantification of atherosclerotic lesions (n = 12, Student's t test). j) ICAM1 and VCAM1 expression in GRK2 siRNA treated‐HUVECs and scramble siRNA treated‐HUVECs exposed to different shear stresses was detected via Western blotting. k) Quantification of ICAM1 and VCAM1 protein expression levels (n = 5, one‐way ANOVA). l) Adhesion of THP‐1 monocytes from GRK2 siRNA treated‐HUVECs and scramble siRNA treated‐HUVECs exposed to different shear stresses. m) GRK2^S29p, ICAM1 and VCAM1 expression levels were detected by Western blotting in constitutively activated GRK2^S29D‐overexpressing HUVECs, and dysfunctional GRK2^K220R‐overexpressing HUVECs exposed to LSS. n) Quantification of GRK2^S29p, ICAM1 and VCAM1 protein expression levels (n = 5; one‐way ANOVA). o) Compared with HUVECs overexpressing the wild‐type plasmid, HUVECs overexpressing the continuously activated GRK2^S29D plasmid presented increased monocyte adhesion under LSS conditions. p) RAW macrophages engulfed ox‐LDL after treatment with conditioned medium from constitutively activated GRK2^S29D‐overexpressing HUVECs and GRK2^K220R‐overexpressing HUVECs (n = 5, one‐way ANOVA).

Figure 4

AP‐1^S63p is a downstream of GRK2 and plays a role in OSS‐induced vascular endothelial dysfunction. a) Volcano plot revealing differentially expressed phosphorylated proteins in vascular endothelial cells exposed to OSS. b) GO functional enrichment analysis of differentially expressed phosphorylated proteins. c) KEGG pathway enrichment analysis of differentially expressed phosphorylated proteins. d) PPI network based on differentially expressed phosphorylated proteins. e) Core PPI network. f) Venn diagram showing the shared proteins between the key phosphorylated proteins in vascular endothelial cells exposed to OSS and the LC‒MS/MS‐detected proteins that potentially bind to GRK2. g) Enface immunofluorescence staining of AP‐1^S63p (red), vWF (green), and DAPI (blue) in the endothelium of the inner curvature of the aorta arch and the thoracic aorta isolated from C57BL/B6J mice. h) Quantification of the AP‐1^S63p fluorescent intensity level (n = 5, one‐way ANOVA). i) HUVECs subjected to OSS for 0, 5, 15, 30, 60, and 120 min. The proteins in the cell lysate were evaluated via Western blotting with the indicated antibodies. j) Quantification of AP‐1^S63p protein expression level (n = 5, one‐way ANOVA). (k) HUVECs and HAECs were subjected to LSS and OSS for 120 min. The proteins in the cell lysate were analyzed via Western blotting with the indicated antibodies. l) Quantification of AP‐1^S63p protein expression level (n = 5, student's t test). m) Immunofluorescence staining for AP‐1^S63p (red) and DAPI (blue) in GRK2 siRNA treated‐HUVECs and scramble siRNA treated‐HUVECs exposure to different shear stresses. n) Quantification of AP‐1^S63p fluorescent intensity level (n = 5, one‐way ANOVA). o) ICAM1 and VCAM1 expression in AP‐1 siRNA treated‐HUVECs and scramble siRNA treated‐HUVECs exposed to different shear stress stresses was detected via Western blot. p) Quantification of ICAM1 and VCAM1 protein expression levels (n = 5, one‐way ANOVA). q) AP‐1^S63p, ICAM1 and VCAM1 expression in constitutively activated AP‐1^S63D‐overexpressing HUVECs and inactivated AP‐1^S63A‐ overexpressing HUVECs was detected via Western blotting after exposure to LSS. r) Quantification of AP‐1^S63p, ICAM1 and VCAM1 protein expression levels (n = 5, one‐way ANOVA).

Figure 5

AAV9‐mediated overexpression of AP‐1^S63D under the control of the ICAM‐2 promoter in vascular endothelial cells promotes endothelial dysfunction and accelerates atherosclerosis. a) Schematic figure of the administration of ICAM2 promoter‐driven AAV9 carrying AP‐1^S63D or vector via tail vein injection in GRK2^ECKO mice. b) Co‐immunofluorescence staining of viral GFP (green) and vWF (red). c) Immunofluorescence staining for ICAM1 (red), VCAM1 (green), vWF (yellow) and DAPI (blue) in partially ligated left common carotid arteries isolated from GRK2^ECKO mice following the administration of ICAM‐2 promoter‐driven AAV9 carrying AP‐1^S63D or vector. d) Quantification of ICAM1 and VCAM1 protein expression levels (n = 5, Student's t test). e) Immunofluorescence staining for F4/80 (red) and DAPI (blue) indicating the infiltration of macrophages in partially ligated left common carotid arteries isolated from GRK2^ECKO mice following the administration of ICAM‐2 promoter‐driven AAV9 carrying AP‐1^S63D or vector. f) Quantification of the F4/80 fluorescent intensity (n = 5, Student's t test). g) Oil Red O staining indicating increased atherosclerotic lesions in endothelial cell specific AP‐1^S63D‐overexpressing GRK2^ECKO; ApoE^‐/‐ mice fed a high‐fat Western diet. h) Quantification of atherosclerotic lesions (n = 12, Student's t test). i) H&E staining and Oil Red O staining of the aortic roots of GRK2^ECKO; ApoE^‐/‐ hyperlipidemic mice following the administration of ICAM‐2 promoter‐driven AAV9 carrying AP‐1^S63D or vector. j,k) Quantification of atherosclerotic lesions (n = 12, Student's t test). l,m) Co‐IP experiment revealing the direct binding of AP‐1 to GRK2 in HEK293 cells (n = 3). n) Co‐IP experiments revealing the direct binding of AP‐1 to GRK2 in HUVECs (n = 3). o) Rescue Western blotting experiments demonstrating that AP‐1^S63D overexpression abrogated the protective effects of GRK2 knockdown on vascular endothelial cells exposed to OSS in HUVECs and HAECs. p) Quantification of ICAM1 and VCAM1 protein expression levels in HUVECs (n = 5, one‐way ANOVA). q) Rescue THP‐1 monocyte adhesion assay results indicating that overexpressing AP‐1^S63D abrogated the protective effects of GRK2 knockdown on vascular endothelial cells exposed to OSS. r) RAW macrophages engulf ox‐LDL after treatment with conditioned medium from GRK2^S29D‐overexpressing HUVECs following AP‐1 siRNA or scramble siRNA treatment.

Figure 6

NR4A1 is transcriptionally regulated by AP‐1. a) Volcano plot showing the differentially expressed genes in the GSE211402 dataset. b) Volcano plot showing the differentially expressed genes in the GSE66360 dataset. c) Venn diagram showing the downstream factors that may be transcriptionally regulated by AP‐1 and play important roles in vascular endothelial dysfunction and atherosclerosis induced by OSS. d) qPCR experiment validating the effect of OSS on NR4A1 mRNA expression in HUVECs (n = 5, Student's t test). e) Validation of the effect of OSS on NR4A1 protein expression in HUVECs and HAECs via Western blotting (n = 5, Student's t test). f) Enface immunofluorescence staining of NR4A1 (red), vWF (green), and DAPI (blue) in the inner curvature of the aortic arch and thoracic aorta isolated from C57BL/6J mice. g) Quantification of the NR4A1 fluorescent intensity (n = 5, one‐way ANOVA). h) Potential binding sequences of the NR4A1 promoter region and AP‐1 predicted by the JASPAR database. i) Schematic diagram of the putative AP‐1 binding site in the NR4A1 promoter and the mutant. j) ChIP‒qPCR experiment validating the transcriptional regulation of NR4A1 by AP‐1 (n = 3, one‐way ANOVA). HUVECs were exposed to OSS and LSS, and then crosslinked chromatin was extracted and immunoprecipitated with anti‐AP‐1^S63p or IgG antibodies. The immunoprecipitated DNA was amplified by PCR. k) HUVECs were co‐transfected with the AP‐1 expression vector (AP‐1 WT) and different NR4A1 truncation promoter‐reporter plasmids to identify the promoter activity regions where AP‐1 regulates NR4A1 transcription (n = 5, one‐way ANOVA). l) HUVECs were cotransfected with the AP‐1 expression vector (AP‐1 WT), the NR4A1 promoter‐reporter plasmid (NR4A1 WT) or the NR4A1 promoter‐reporter binding site mutant (NR4A1 mutant). 48 h after transfection, cell lysates were prepared for the luciferase assay. The data represent the relative NR4A1 promoter activity normalized to Renilla luciferase activity (n = 3, Student's t test). m) NR4A1 expression in AP‐1 siRNA treated‐HUVECs and scramble siRNA treated‐HUVECs exposed to different shear stresses was evaluated via Western blotting and PCR. n) Quantification of NR4A1 protein and mRNA expression levels (n = 5, one‐way ANOVA). o) Immunofluorescence staining for NR4A1 (red) and DAPI (blue) in AP‐1 siRNA treated‐HUVECs and scramble siRNA treated‐HUVECs exposed to different shear stresses. p) Quantification of the relative NR4A1 fluorescent intensity (n = 5, one‐way ANOVA).

Figure 7

NR4A1 functions in vascular endothelial cell energy metabolism disorders induced by OSS by binding to LKB1 in the nucleus. a,b) Co‐IP experiment revealing the direct binding of NR4A1 to LKB1 in HEK293 cells (n = 3). c) Co‐IP experiment revealing the direct binding of NR4A1 to LKB1 in HUVECs (n = 3). d) Immunofluorescence staining for LKB1 (green) and DAPI (blue) in HUVECs exposed to different shear stresses. e) NR4A1 siRNA treated‐HUVECs and scramble siRNA treated‐HUVECs were exposed to OSS and LSS, and the LKB1 protein expression levels in the nucleus and cytoplasm of the HUVECs were detected by Western blotting. f) Quantification of LKB1 protein expression levels in the nucleus and cytoplasm (n = 5, one‐way ANOVA). g) Immunofluorescence staining for LKB1^S428p (green) and DAPI (blue) in NR4A1 siRNA treated‐HUVECs and scramble siRNA treated‐HUVECs exposed to different shear stresses. h) Quantification of LKB1 immunofluorescent intensity (n = 5, one‐way ANOVA). i) NR4A1 siRNA treated‐HUVECs and scramble siRNA treated‐HUVECs were subjected to different shear stresses, and the protein expression levels of LKB1^S428p, LKB1, AMPK^T172p, and AMPK were detected by Western blotting. j) Quantification of LKB1^S428p and AMPK^T172p protein expression levels (n = 5, one‐way ANOVA). k) eNOS^S1177p expression in constitutively activated LKB1^S428D‐overexpressing HUVECs and vector control‐transfected HUVECs exposed to different shear stresses was detected via Western blotting. l) Quantification of eNOS^S1177p protein expression levels (n = 5, one‐way ANOVA). m) ROS levels were detected by DCFH‐DA (green) in constitutively activated LKB1^S428D‐overexpressing HUVECs and vector control HUVECs exposed to different shear stresses. n) Quantification of ROS levels (n = 5, one‐way ANOVA). o) Extracellular acidification rate (ECAR) profiles showing glycolytic function in vector‐, GRK2^S29D‐, AP‐1^S63D‐, and NR4A1‐overexpressing HUVECs. The vertical lines indicate the time of addition of glucose (10 mmol L⁻¹), oligomycin (3 µmol L⁻¹), and 2‐deoxy‐D‐glucose (2‐DG) (100 mmol L⁻¹). p) Quantification of glycolytic function parameters from (o); values are normalized to those of 10⁴ cells (n = 5, one‐way ANOVA). q) Oxygen consumption rate (OCR) profiles showing mitochondrial respiration function in vector‐, GRK2^S29D‐, AP‐1^S63D‐, and NR4A1‐overexpressing HUVECs. The vertical lines indicate the time of addition of oligomycin (3 µmol L⁻¹), trifluoromethoxy phenylhydrazone (FCCP) (1 µmol L⁻¹), antimycin A (1.5 µmol L⁻¹), or rotenone (3 µmol L⁻¹). r) Quantification of mitochondrial respiration function parameters from (q); values normalized to those of 10⁴ cells (n = 5, one‐way ANOVA).

Figure 8

A GRK2 inhibitor (paroxetine) alleviates OSS‐induced vascular endothelial dysfunction and atherosclerosis in vivo and in vitro. a) Immunofluorescence staining for ICAM1 (red), VCAM1 (green), vWF (yellow) and DAPI (blue) in partially ligated left common carotid arteries isolated from vehicle group mice and paroxetine‐treated mice (10 mg kg⁻¹ daily; I.g). b) Quantification of the relative ICAM1 and VCAM1 fluorescent intensities (n = 5, Student's t test). c) ICAM1 and VCAM1 were detected by Western blotting in paroxetine treated‐HUVECs and vehicle treated‐HUVECs exposed to different shear stresses. d) ICAM1 and VCAM1 were detected by Western blotting in fluoxetine treated‐HUVECs and vehicle treated‐HUVECs exposed to different shear stresses. e,f) Quantification of ICAM1 and VCAM1 protein expression levels (n = 5, one‐way ANOVA). g) The GRK2/AP‐1 signaling axis and its downstream targets were detected by Western blotting in paroxetine treated‐HUVECs and vehicle treated‐HUVECs exposed to different shear stresses. h) Quantification of the GRK2/AP‐1 signaling axis and its downstream targets expression levels (n = 5, one‐way ANOVA). i) Adhesion of THP‐1 monocytes in paroxetine treated‐HUVECs and vehicle treated‐HUVECs exposed to OSS. j) Oil Red O staining indicating decreased atherosclerotic lesions in paroxetine‐treated ApoE^‐/‐ mice fed a high‐fat Western diet. k) Quantification of atherosclerotic lesions (n = 12, Student's t test). l) H&E and Oil Red O staining of the aortic roots of hyperlipidemic ApoE^‐/‐ mice in the vehicle group and paroxetine‐treated (10 mg kg⁻¹ daily; I.g) group. m,n) Quantification of atherosclerotic lesions (n = 12, Student's t test).

Figure 9

Schematic illustration. OSS activates the GPCR in vascular endothelial cells and leads to the phosphorylation of GRK2, which in turn phosphorylates the downstream transcription factor AP‐1. AP‐1 promotes the transcription of the monocyte recruitment factors ICAM1 and VCAM1, causing inflammation in the vascular endothelium. In addition, AP‐1 also leads to increased expression levels of NR4A1, which binds with LKB1 in the nucleus, resulting in decreased LKB1 phosphorylation and activity. This ultimately leads to AMPK inactivation, oxidative stress, and reductions in eNOS activity.