Inhibition of cytokine signaling in human retinal endothelial cells through modification of caveolae/lipid rafts by docosahexaenoic acid

Abstract

Purpose: Docosahexaenoic acid (DHA(22:6,n3)) is the principal n3 polyunsaturated fatty acid (PUFA) in the retina. The authors previously demonstrated that DHA(22:6,n3) inhibited cytokine-induced adhesion molecule expression in primary human retinal vascular endothelial (hRVE) cells, the target tissue affected by diabetic retinopathy. Despite the importance of vascular inflammation in diabetic retinopathy, the mechanisms underlying anti-inflammatory effects of DHA(22:6,n3) in vascular endothelial cells are not understood. In this study the authors address the hypothesis that DHA(22:6,n3) acts through modifying lipid composition of caveolae/lipid rafts, thereby changing the outcome of important signaling events in these specialized plasma membrane microdomains.

Methods: hRVE cells were cultured in the presence or absence of DHA(22:6,n3). Isolated caveolae/lipid raft-enriched detergent-resistant membrane domains were prepared using sucrose gradient ultracentrifugation. Fatty acid composition and cholesterol content of caveolae/lipid rafts before and after treatment were measured by HPLC. The expression of Src family kinases was assayed by Western blotting and immunohistochemistry.

Results: Disruption of the caveolae/lipid raft structure with a cholesterol-depleting agent, methyl-cyclodextrin (MCD), diminished cytokine-induced signaling in hRVE cells. Growth of hRVE cells in media enriched in DHA(22:6,n3) resulted in significant incorporation of DHA(22:6,n3) into the major phospholipids of caveolae/lipid rafts, causing an increase in the unsaturation index in the membrane microdomain. DHA(22:6,n3) enrichment in the caveolae/raft was accompanied by a 70% depletion of cholesterol from caveolae/lipid rafts and displacement of the SFK, Fyn, and c-Yes from caveolae/lipid rafts. Adding water-soluble cholesterol to DHA(22:6,n3)-treated cells replenished cholesterol in caveolae/lipid rafts and reversed the effect of DHA(22:6,n3) on cytokine-induced signaling.

Conclusions: Incorporation of DHA(22:6,n3) into fatty acyl chains of phospholipids in caveolae/lipid rafts, followed by cholesterol depletion and displacement of important signaling molecules, provides a potential mechanism for anti-inflammatory effect of DHA(22:6,n3) in hRVE cells.

FIGURE 1

Incorporation of ¹⁴C- DHA_22:6,n3 into hRVE cellular lipids. hRVE cells were treated with ¹⁴C-DHA_22:6,n3 for the time periods indicated. Total lipids were extracted and subjected to thin-layer chromatography (TLC) analysis. Radioactivity was detected and quantified by phosphorimaging. A phosphorimage of total lipid TLC analysis is presented (A), as is phospholipid TLC analysis (B). Location of the authentic standards for polar lipid (PL), diacylglycerol (DAG), nonesterified fatty acid (NEFA), triacylglycerol (TG), and cholesterol ester (CE) is shown on the total lipid TLC (A). Location of the authentic standards for phosphatidylcholine (PC), phosphatidylinositol (PI), phosphatidylserine (PS), and phosphatidylethanolamine (PE) are shown on the phospholipid TLC (B).

FIGURE 2

Characterization of caveolae/lipid raft protein components in hRVE cells. Membrane lipids were fractionated to enrich for caveolae/lipid rafts and nonraft components. Each membrane fraction was separated on SDS-PAGE. Western blot analysis for the caveolae marker caveolin-1, the lipid raft marker flotillin-1, the plasma membrane marker Na⁺/K⁺ATPase, PKC-α, and Erk1/2 was performed to analyze protein distribution in each fraction. Fractions designated caveolae/raft and nonraft membrane are indicated.

FIGURE 3

Modification of phospholipid fatty acyl composition of caveolae/lipid raft and total plasma membrane of hRVE cells by incubation with palmitic acid_16:0 or DHA_22:6,n3. hRVE cells were treated with 100 μM palmitic acid_16:0 (gray bars) or DHA_{22: 6,n3} (black bars) for 24 hours. Vehicle (20 μM BSA)-treated cells (white bars) were used as a control. Caveolae/lipid raft–enriched and nonenriched fractions were purified as described in Experimental Procedures. Lipids were extracted, and phospholipids and sphingolipids were enriched by aminopropyl column fractionation. Neutral phospholipids (PC, PE, SM) and acidic phospholipids (PI, PS, PA) were saponified and analyzed by RP-HPLC. The concentration for each fatty acid was obtained by normalizing to standards and was presented as the mole percent of total fatty acids in each fraction. The unsaturation index was calculated as the average number of double bonds per fatty acyl residue. Given that most phospholipids were in the neutral phospholipid fraction, only the data for neutral phospholipid fraction are presented. Data are the mean ± SD of results in three experiments. *P < 0.05 compared with control. ^#P < 0.05 compared with palmitic acid_16:0-treated cells.

FIGURE 4

DHA_22:6,n3 treatment depletes cholesterol in caveolae/lipid rafts. hRVE cells were treated with 100 μM palmitic acid_16:0 (gray bars) or DHA_22:6,n3 (black bars) for 24 hours. Vehicle (20 μM BSA)–treated cells were used as a control (white bars). Water-soluble cholesterol (25 μM cholesterol complexed with 250 μM MCD) was added to DHA_22:6,n3-treated cells for 30 minutes (striped bars). Caveolae/lipid raft–enriched domains were purified as described in Experimental Procedures and submitted to total lipid extraction and amino-propyl column fractionation. Neutral lipids were fractionated by normal-phase HPLC analyses. The amount of cholesterol was presented as nmol/μg protein. Data are the mean ± SD of results in three experiments. *P < 0.05 compared with control. #P < 0.05 compared with DHA_22:6,n3-treated cells without cholesterol.

FIGURE 5

Inhibition of TNF-α–induced NF-κB signaling by (MCD) pretreatment of hRVE cells. hRVE cells were pretreated with 8 mM MCD for 30 minutes before stimulation with TNF-α (20 ng/mL) for the indicated time. IκBα and ERK phosphorylation levels (A) and ICAM-1 expression levels (B) were assessed by immunoblot analyses. Equal amounts of protein were added to each lane, as confirmed by actin levels. Data are the mean ± SD of results in three experiments. *P < 0.05.

FIGURE 6

Cholesterol replenishment in caveolae/lipid rafts reverses DHA_22:6,n3 inhibition of TNF-α induced ICAM-1 expression. hRVE cells treated with 100 μM DHA_22:6,n3 (black bars) for 12 hours followed by TNF-α stimulation for 12 hours. Cells treated with palmitic acid_16:0 were used as a control (gray bars). Water-soluble cholesterol (25 μM cholesterol complexed with 250 μM MCD) was added to DHA_22:6,n3-treated cells for 30 minutes before TNF-α stimulation (striped bar). ICAM-1 expression levels were assessed by immunoblot analyses. Equal amounts of protein were added to each lane, as confirmed by actin levels. Data are the mean ± SD of results in three experiments, *P < 0.05 compared with palmitic acid_16:0 treated cells. #P < 0.05 compared with DHA_22:6,n3-treated cells without cholesterol.

FIGURE 7

Src family kinase enrichment in hRVE caveolae/lipid rafts. Caveolae/lipid raft fractions were purified, as described in Materials and Methods. Fractions 2 to 4 were combined and designated the caveolae/lipid raft (cav/raft) fractions, and fractions 5 to 9 were combined and designated the nonraft membrane fraction. Fractions were separated on SDS-PAGE and analyzed by Western blot. Results are representative of seven experiments.

FIGURE 8

Effect of SFK inhibition on TNF-α–induced VCAM-1 expression. hRVE cells were pretreated with PP2 (10 μM) for 30 minutes and stimulated with TNF-α (5 ng/mL) for 6 hours. VCAM-1 and ICAM-1 expression levels were assessed by immunoblot analyses. Equal amounts of protein were added to each lane, as confirmed by actin levels. Results are representative of three experiments.

FIGURE 9

DHA_22:6,n3 displaces Fyn and c-Yes from hRVE caveolae/lipid rafts. (A) hRVE cells were treated with 100 μM palmitic acid_16:0 or DHA_22:6,n3 for 24 hours. Vehicle (20 μM BSA)-treated cells were used as a control. Caveolae/lipid raft fractions were isolated and analyzed by Western blot. Amounts of Fyn (B) and c-Yes (C), localized in caveolae/lipid rafts isolated from fatty acid–treated cells, were quantified and presented as a fold change. Data are the mean ± SD of results in three experiments. *P < 0.05.