A Novel Anti-Inflammatory Effect for High Density Lipoprotein

Abstract

High density lipoprotein has anti-inflammatory effects in addition to mediating reverse cholesterol transport. While many of the chronic anti-inflammatory effects of high density lipoprotein (HDL) are attributed to changes in cell adhesion molecules, little is known about acute signal transduction events elicited by HDL in endothelial cells. We now show that high density lipoprotein decreases endothelial cell exocytosis, the first step in leukocyte trafficking. ApoA-I, a major apolipoprotein of HDL, mediates inhibition of endothelial cell exocytosis by interacting with endothelial scavenger receptor-BI which triggers an intracellular protective signaling cascade involving protein kinase C (PKC). Other apolipoproteins within the HDL particle have only modest effects upon endothelial exocytosis. Using a human primary culture of endothelial cells and murine apo-AI knockout mice, we show that apo-AI prevents endothelial cell exocytosis which limits leukocyte recruitment. These data suggest that high density lipoprotein may inhibit diseases associated with vascular inflammation in part by blocking endothelial exocytosis.

Fig 1. Lipoprotein electrophoresis.

HDL-3 was isolated from the blood of a healthy human volunteer, fractionated by SDS-PAGE, and stained with Coomassie blue. Purified apolipoproteins were also loaded and fractionated as controls. The HDL-3 fraction contains a major apolipoprotein with the same relative mobility as ApoA-I.

Fig 2. HDL decreases endothelial exocytosis.

(A) Dose-response. Endothelial cells were pre-treated with various concentrations of purified HDL-3 within the human serum reference range for 2 h. (B). Time course. Endothelial cells were pre-treated with 0.5 mg/dL HDL-3 for 2 h. Cells were washed and stimulated with thrombin 1 U/ml, and the amount of VWF released over 1 h was measured by an ELISA (n = 3 ± S.D. *P < 0.05 vs. thrombin and 0 HDL-3).

Fig 3. Purified apolipoproteins decrease endothelial exocytosis.

(A-D) Dose-response for ApoA-I (human serum reference range is 0.75–1.75 mg/mL), and equivalent concentrations of Apo-AII, ApoC-I, and ApoE. Endothelial cells were pre-treated with various concentrations of purified apolipoproteins for 2 h. Cells were washed and stimulated with thrombin 1 U/ml, and the amount of VWF released over 1 h was measured by an ELISA (n = 3 ± S.D. *P < 0.05 vs. thrombin and 0 aplipoprotein). (E) Purified apolipoproteins are not cytotoxic to endothelial cells. Endothelial cells were pre-treated with IC_50% concentrations of purified apolipoproteins for 2 h. Hydrogen peroxide (H₂O₂) is a positive control for death. Cell viability was measured via formazan release with the MTS assay (n = 3 ± S.D. *P < 0.05 vs. control. **P < 0.01 vs. control.) The dashed line represents control levels.

Fig 4. ApoA-I inhibits endothelial exocytosis via SR-BI.

(A) Reaction of albumin (lane 1) with maleic acid (lane 2) leads to maleylated albumin with enhanced electrophoretic mobility, shown by Coomassie staining. (B) Maleylated albumin relieves the inhibition of endothelial exocytosis by ApoA-I. (C) SR-BI Expression: HAEC lysate, lane 1; Human Embryonic Kidney (HEK) 293 cell lysate, lane 2; wild-type mouse liver lysate, lane 3 were analyzed by immunoblotting with an SR-BI antibody. The signal at 82 kDa is consistent with the molecular weight of SR-BI; the band at 60 kDa is consistent with the molecular weight of unglycosylated SR-BI (asterisk). (D) SR-BI blocking antibody. Endothelial cells were pre-treated with an SR-BI blocking antibody, followed by the addition of 10⁻⁴ mg/mL apoA-I for 2 h. Cells were washed, stimulated with thrombin 1 U/ml, and the amount of VWF released over 1 h was measured by an ELISA (n = 3 ± S.D. *P < 0.05, vs. 0 apoA-I and 0 SR-BI receptor antibody).

Fig 5. HDL and ApoA-I inhibit leukocyte adhesion to endothelial cells in vitro.

(A-B) HDL inhibits p-selectin-mediated leukocyte adhesion to endothelial cells in vitro. Endothelial cells were pre-treated with 0.5 mg/dL HDL-3 for 2 h, washed, and stimulated with or without 1 U/mL thrombin for 1 h. Antibody to P-selectin or control IgG was added to some cells. (C-D). ApoA-I inhibition of leukocyte adherence. Endothelial cells were pre-treated with apoA-I 10⁻⁴ mg/ml for 2 h, washed, and stimulated with 1 U/mL thrombin for 1 h. The adherence of HL-60 cells labeled with BCEF-AM was measured as above, and imaged using a digital camera (n = 5 ± S.D. *P < 0.05, vs. thrombin only condition). Calibration bar is 50 μm.

Fig 6. PKC mediates apoA-I inhibition of endothelial exocytosis.

(A) HDL-3 and apoA-I activate PKC. Cells were treated with apoA-I (10⁻⁴ mg/mL) or HDL (0.5 mg/dL) and with or without the PKC inhibitor RO318220 for 2 h. PKC activation was assessed using a pan-phospho-PKC antibody (upper); asterisk indicates a non-specific band. eNOS activation was assessed using a phospho-eNOS (S1177)-specific antibody (akt phosphorylation site). (B) HDL and ApoA-I activate ERK1/2. Cells were pre-treated with or without the MEK inhibitor UO126 for 2 h, and then treated with apoA-I (10⁻⁴ mg/mL) or HDL (0.5 mg/dL) for 1 h. ERK activation was assessed using a phospho-ERK1/2 antibody (upper). This blots are representative of two others with similar results. (C) Cells treated with apoA-I (10⁻⁴ mg/mL), HDL-3 (0.5 mg/dL) for 1hr, or thrombin (1 U/mL) for 1 hr or 3hrs, and NFκB activation (p-NF κB) was assessed. Thrombin appears to activate NFkB (phospo-NFkB). (D) Treatment of endothelial cells with apoA-I (10⁻⁴ mg/mL) or HDL-3 (0.5 mg/dL) neither inhibits NFκB activation nor changes VCAM-1 expression following stimulation of cells with thrombin (1 U/mL) for 0.5–3 hrs. This suggests apoA-I and HDL-3 at the doses and in the time frame used do not exert an anti-inflammatory effect through these signaling pathways. Western blotting was conducted for phospo-NFkB, IkB-α (p65 subunit), vascular cell adhesion protein 1 (VCAM-1), or GAPDH. These blots are representative of two others with similar results.

Fig 7. PKC mediates apoA-I inhibition of endothelial exocytosis.

(A) PKC mediates apoA-I inhibition of endothelial exocytosis: PKC activation. Cells were pre-treated with the PKC activator PMA for 2 h. The cells were washed, then stimulated with thrombin for 1 h. The amount of VWF released over 1 h was measured by an ELISA (n = 3 ± S.D. *P < 0.05 for condition vs. thrombin alone). (B) Cells were pre-treated with the MAPKK inhibitor U0126, the PKC inhibitor RO318220 for 2 h. The cells were washed, treated with 0.5 mg/dL HDL-3 for 2 h, and then stimulated with thrombin for 1 h. The amount of VWF released over 1 h was measured by an ELISA (n = 3 ± S.D. *P < 0.05 for condition vs. thrombin + 0 HDL-3). (C) PKC mediates apoA-I inhibition of exocytosis. Cells were pre-treated with the MAPKK inhibitor U0126, the PKC inhibitor RO318220, or the NOS inhibitor L-NAME for 2 h. The cells were washed, treated with 0.5 mg/dL HDL-3 for 2 h, and then stimulated with thrombin for 1 h. The amount of VWF released over 1 h was measured by an ELISA (n = 3 ± S.D. *P < 0.05 for condition vs. thrombin + 0 apoA-I).

Fig 8. ApoA-I and HDL-3 inhibit leukocyte adhesion to endothelial cells in vitro via a PKC-mediated signaling pathway.

Cells were pre-treated with the MAPKK inhibitor U0126 or the PKC inhibitor RO318220 for 2 h. The cells were then treated with apoA-I or HDL-3 for 1 h, and next stimulated with thrombin for 1 h. (A). The endothelial cells were incubated with HL-60 cells, washed and then imaged with a digital camera. (n = 5 ± S.D. *P < 0.05).

Fig 9. ApoA-I inhibits leukocyte adhesion to endothelial cells in vivo.

(A) Venules from the small intestine were isolated in mice. Histamine (1μM) was superfused onto the venule and, after 5 mins, rhodamine 6G-labeled leukocytes were imaged using a digital fluorescence camera. Representative images are shown above. Calibration bar = 100 μm. P-selectin knockout mice (p-selectin^-/-) do not display leukocyte-to-endothelial cell adhesion after histamine stimulation, validating this as an in vivo acute inflammatory adhesion assay. (B) ApoA-I knockout mice (apoA-I ^-/-) have low plasma HDL concentration compared to WT mice and show increased adhesion of leukocytes to the vascular endothelium in vivo (decreased rolling velocity) compared to WT mice or apoA-I ^-/- mice injected with apo-AI (0.5mg, intraperitoneal) 24 hours before. Plasma HDL concentration in the apoA-I ^-/- mouse is noted below the blot. Changes in leukocyte rolling velocity (Δv, μm/ms) are noted above the graph for each treatment group. Western blotting was conducted for plasma apoAI (1:20 dilution) using an anti-apoA-I antibody. Leukocyte rolling velocity was assessed using 3 leukocytes per frame for n = 3–7 mice (mean rolling velocity ± SEM, *P < 0.05.

Fig 10. Nitration of apoA-I attenuates apoA-I inhibition of exocytosis.

(A) Purified apoA-I (10⁻⁴ mg/mL) was incubated with increasing concentrations of peroxynitrite (left) or degraded peroxynitrite as a control (right). Mixtures were then fractionated by SDS-PAGE and immunoblotted with antibody to nitrotyrosine in order to assess tyrosine nitration of apoA-I. The position of apoA-I on the gel was confirmed with Coomassie staining (not shown). This experiment was repeated three times with similar results. (B). ApoA-I was treated with increasing concentrations of peroxynitrite or degraded peroxynitrite as a control. Endothelial cells were then incubated with treated or non-treated apoA-I 10⁻⁴ mg/mL for 2 h, washed, and stimulated with thrombin (1 U/ml). The amount of VWF released over 1 h was measured by an ELISA (n = 3 ± S.D. *P < 0.05, peroxynitrite vs. degraded peroxynitrite.