C-reactive protein promotes atherosclerosis by increasing LDL transcytosis across endothelial cells

Abstract

Background and purpose: The retention of plasma low-density lipoprotein (LDL) particles in subendothelial space following transcytosis across the endothelium is the initial step of atherosclerosis. Whether or not C-reactive protein (CRP) can directly affect the transcytosis of LDL is not clear. Here we have examined the effect of CRP on transcytosis of LDL across endothelial cells and have explored the underlying mechanisms.

Experimental approach: Effects of CRP on transcytosis of FITC-labelled LDL were examined with human umbilical vein endothelial cells and venous rings in vitro and, in vivo, ApoE(-/-) mice. Laser scanning confocal microscopy, immunohistochemistry and Oil Red O staining were used to assay LDL.

Key results: CRP increased transcytosis of LDL. An NADPH oxidase inhibitor, diphenylene iodonium, and the reducing agent, dithiothreitol partly or completely blocked CRP-stimulated increase of LDL transcytosis. The PKC inhibitor, bisindolylmaleimide I and the Src kinase inhibitor, PP2, blocked the trafficking of the molecules responsible for transcytosis. Confocal imaging analysis revealed that CRP stimulated LDL uptake by endothelial cells and vessel walls. In ApoE(-/-) mice, CRP significantly promoted early changes of atherosclerosis, which were blocked by inhibitors of transcytosis.

Conclusions and implications: CRP promoted atherosclerosis by directly increasing the transcytosis of LDL across endothelial cells and increasing LDL retention in vascular walls. These actions of CRP were associated with generation of reactive oxygen species, activation of PKC and Src, and translocation of caveolar or soluble forms of the N-ethylmaleimide-sensitive factor attachment protein.

Keywords: C-reactive protein; LDL; PKC; ROS; Src; atherosclerosis; caveolae; endothelial cells; transcytosis.

Figure 1

Schematic diagram of the model of LDL transcytosis. Passage of FITC-LDL through inserts with HUVECs monolayers was carried out at 37°C. LDL was added to the upper side of the inserts (50 μg·mL⁻¹ or 100 μg·mL⁻¹ FITC-LDL to all inserts and an additional sixfold excess of unlabelled LDL to naive inserts) for 3 h. Samples were collected from the outer chambers and further dialysed against PBS at 4°C. The FITC fluorescent intensity was assessed by fluorescence spectrophotometry.

Figure 2

Analysis of LDL transcytosis in an in vitro model. Receptor-mediated transport (LDL transcytosis) was calculated by subtracting the FITC intensity obtained in the presence of native LDL (paracellular transport, Figure 2A, B) from that obtained in the absence of native LDL (total transport, Figure 2A, B) and are summarized in Figure 2C, D respectively. A,B: After incubation with 50 μg·mL⁻¹ (A) or 100 μg·mL⁻¹ (B) FITC-LDL for the indicated amount of time, the total transport and paracellular transport were measured. **P < 0.01 versus 3 h group; *P < 0.05 versus 3 h group, n = 4; (C) Time-dependent transcytosis of 50 μg·mL⁻¹ and 100 μg·mL⁻¹ FITC-LDL. *P < 0.05 versus 3 h 50 μg·mL⁻¹ group, n = 4. (D) Concentration-dependent transcytosis of FITC-LDL at 3 h and 24 h. *P < 0.05 versus 3 h 50 μg·mL⁻¹ group, n = 4.

Figure 3

Effect of CRP on LDL transcytosis and ROS production in HUVECs. HUVECs were pretreated with 10 μmol·L⁻¹ DPI, or 30 μmol·L⁻¹ DTT for 30 min, and then exposed to 20 μg·mL⁻¹ CRP or 0.1 μmol·L⁻¹ H₂O₂ for 30 min. The intracellular ROS level was detected with DCF-DA. (A) LDL transcytosis in the absence or presence of CRP, DPI + CRP or DTT + CRP. **P < 0.01 versus Ctr; ^‡P < 0.01 versus CRP; n = 4. (B) Production of ROS in the absence or presence of CRP, DPI + CRP or DTT + CRP. *P < 0.05 versus Ctr; †P < 0.05 versus CRP; n = 4. (C) LDL transcytosis with or without H₂O₂ (0.1 μmol·L⁻¹). *P < 0.05 versus Ctr, n = 4.

Figure 4

Effect of CRP on the activity of PKC and Src kinase in HUVECs. HUVECs were pretreated with 10 μmol·L⁻¹ DPI, or 30 μmol·L⁻¹ DTT for 30 min, and then exposed to 20 μg·mL⁻¹ CRP for 5 min. (A) PKC activity in the absence or presence of CRP, DPI + CRP, DTT + CRP, BIM I + CRP or PP2 + CRP. **P < 0.01 versus Ctr; ^‡P < 0.01 versus CRP; §P > 0.05 versus CRP; n = 4. (B) Src kinase activity in the absence or presence of CRP, DPI + CRP, DTT + CRP, BIM I + CRP or PP2 + CRP. **P < 0.01 versus Ctr; ^‡P < 0.01 versus CRP; §P > 0.05 versus CRP; n = 4.

Figure 5

Effect of CRP on the expression of proteins involved in LDL endocytosis and exocytosis in lipid rafts in HUVECs. HUVECs were first incubated with 10 μmol·L⁻¹ DPI, 30 μmol·L⁻¹ DTT, 3 mmol·L⁻¹ MβCD, 10 μmol·L⁻¹ NEM for 30 min or 5 μmol·L⁻¹ BIM I, 5 μmol·L⁻¹ PP2 for 1 h, followed by 20 μg·mL⁻¹ CRP exposure for 3 h. (A) Representative Western blot showing the subcellular localization of caveolin-1 without CRP stimulation. (B) Representative Western blot showing the expression of proteins involved in LDL endocytosis and exocytosis in lipid rafts in the absence or presence of CRP, DTT + CRP, BIM I + CRP, PP2 + CRP, MβCD + CRP or NEM +CRP. (C–F) Summary bar graph showing the expression of proteins involved in LDL endocytosis (C–D) and exocytosis (E–F) in lipid rafts in the absence or presence of CRP, DTT + CRP, BIM I + CRP, PP2 + CRP, MβCD + CRP or NEM + CRP respectively. *P < 0.05 versus Ctr; †P < 0.05 versus CRP; n = 4.

Figure 6

Confocal analysis of FITC-LDL uptake in HUVECs. HUVECs were first incubated with 50 μg·mL⁻¹ FITC-LDL for 24 h, and then incubated with 10 μmol·L⁻¹ DPI, 30 μmol·L⁻¹ DTT, 3 mmol·L⁻¹ MβCD, 10 μmol·L⁻¹ NEM for 30 min or 5 μmol·L⁻¹ BIM I, 5 μmol·L⁻¹ PP2 for 1 h, followed by 20 μg·mL⁻¹ CRP exposure for 3 h. (A) Confocal microscopy images of FITC-LDL uptake stimulated by CRP alone or after pretreatment with DPI, DTT, BIM I, PP2, MβCD or NEM in HUVECs. Scale bars = 50 μm. (B) Quantification summary of FITC-LDL uptake in HUVECs. *P < 0.05 versus Ctr; †P < 0.05 versus CRP; n = 4.

Figure 7

Confocal analysis of FITC-LDL retention in human umbilical venous wall. The human umbilical venous rings were incubated with 50 μg·mL⁻¹ FITC-LDL, and/or 20 μg·mL⁻¹ CRP, 10 μmol·L⁻¹ DPI, 30 μmol·L⁻¹ DTT, 5 μmol·L⁻¹ BIM I, 5 μmol·L⁻¹ PP2, 3 mmol·L⁻¹ MβCD, 10 μmol·L⁻¹ NEM for 3 h. (A) Confocal microscopic images of FITC-LDL retention stimulated by CRP alone or with DPI, DTT, BIM I, PP2, MβCD or NEM in human umbilical venous rings. Scale bars = 100 μm. (B) Quantitative summary of FITC-LDL retention in vessels. *P < 0.05 versus Ctr; †P < 0.05 versus CRP; n = 4.

Figure 8

CRP increases atherosclerotic lesion formation, CD154 staining within plaques in ApoE-/- mice. (A) The treatment protocols used for the analysis of the effects of CRP. (B) Oil Red O–stained aortic root sections. Scale bars = 600 μm. (C) Quantitative summary of the percentage of the area of atherosclerotic lesion in the aortic root. *P < 0.05 versus Ctr; †P < 0.05 versus CRP; n = 7. ORO: Oil Red O. (D) Immunostaining for CD154 in aortic root sections. Scale bars = 200 μm. (E) Quantitative summary of the expression of CD154. *P < 0.05 versus Ctr; †P < 0.05 versus CRP; n = 7.

Figure 9

Proposed model of CRP-induced LDL transcytosis in endothelial cells.