Oxidized LDL activates blood platelets through CD36/NOX2-mediated inhibition of the cGMP/protein kinase G signaling cascade

Abstract

Oxidized low-density lipoprotein (oxLDL) promotes unregulated platelet activation in dyslipidemic disorders. Although oxLDL stimulates activatory signaling, it is unclear how these events drive accelerated thrombosis. Here, we describe a mechanism for oxLDL-mediated platelet hyperactivity that requires generation of reactive oxygen species (ROS). Under arterial flow, oxLDL triggered sustained generation of platelet intracellular ROS, which was blocked by CD36 inhibitors, mimicked by CD36-specific oxidized phospholipids, and ablated in CD36(-/-) murine platelets. oxLDL-induced ROS generation was blocked by the reduced NAD phosphate oxidase 2 (NOX2) inhibitor, gp91ds-tat, and absent in NOX2(-/-) mice. The synthesis of ROS by oxLDL/CD36 required Src-family kinases and protein kinase C (PKC)-dependent phosphorylation and activation of NOX2. In functional assays, oxLDL abolished guanosine 3',5'-cyclic monophosphate (cGMP)-mediated signaling and inhibited platelet aggregation and arrest under flow. This was prevented by either pharmacologic inhibition of NOX2 in human platelets or genetic ablation of NOX2 in murine platelets. Platelets from hyperlipidemic mice were also found to have a diminished sensitivity to cGMP when tested ex vivo, a phenotype that was corrected by infusion of gp91ds-tat into the mice. This study demonstrates that oxLDL and hyperlipidemia stimulate the generation of NOX2-derived ROS through a CD36-PKC pathway and may promote platelet hyperactivity through modulation of cGMP signaling.

Figure 1

oxLDL stimulates ROS production in platelets. (A) Washed human platelets (1 × 10⁸ platelets/mL) were incubated with the superoxide-detection probe for 30 minutes at 37°C and then treated with either nLDL or oxLDL (1-50 µg/mL) for 15 minutes, and fluorescence was measured at 650 nm (n = 4). *P < .05 compared with basal. (B) As in panel A, except platelets were incubated with DHE (5 µmol/L) for 30 minutes followed by LDL (50 µg/mL) or thrombin (0.01 U/mL) for 15 minutes, and DHE oxidation product for superoxide anion was measured by LC-MS. The data are presented as % increase in superoxide anion above basal (n = 4). *P < .05 compared with basal. (C) Washed human platelets (5 × 10⁷ platelets/mL) were preincubated with a superoxide-detection probe for 30 minutes at 37°C and adhered to slides coated with oxLDL or nLDL (50 µg/mL) or human fibrinogen (100 µg/mL). Platelet fluorescence emission was then captured over time. Representative images of 5 separate experiments are shown. Scale bar, 20 µm. (D) Human platelets were stained with superoxide-detection probe, treated with nLDL or oxLDL (50 µg/mL), reconstituted with autologous RBCs, and perfused through fibrinogen-coated capillary tubes at arterial shear (1000 s⁻¹). (i) Images were then captured under bright-field or fluorescence microscopy (n = 4 separate experiments with different blood donors). (ii) The number of platelets in each field is presented as mean ± SEM (n = 4).

Figure 2

oxLDL stimulates platelet ROS production in a CD36- and NOX2-dependent manner. (A) Suspended human platelets (1 × 10⁸ platelets/mL) were incubated with the superoxide-detection probe for 30 minutes at 37°C, then treated with oxLDL (50 µg/mL) for 15 minutes, in the presence or absence of SSO (100 µmol/L) or fucoidan (50 µg/mL). Fluorescence was measured at 650 nm (n = 4). *P < .05 compared with oxLDL alone. (B) As in panel A, except platelets were treated with oxPC_CD36 or PAPC (5 µmol/L) (*P < .05, PAPC compared with oxPC_CD36). (C) Suspended human platelets (1 × 10⁸ platelets/mL) were incubated with DHE (5 µmol/L) for 30 minutes followed by oxPC_CD36 (5 µmol/L) or thrombin (0.1 U/mL) for 15 minutes, and the DHE oxidation product for superoxide anion was measured by LC-MS. The data are presented as % increase in superoxide above basal (n = 4). *P < .05 compared with basal. (D) WT and CD36^−/− murine platelets were stained with superoxide-detection probe and then treated with oxPC_CD36 (5 µmol/L). Reconstituted blood was perfused through fibrinogen-coated capillary tubes, and images of adherent platelets were then taken under bright-field or fluorescence microscopy (n = 4 separate experiments with different mice in each group). (i) Representative bright-field and fluorescence images are shown. (ii) The number of platelets in each field is presented as mean ± SEM (n = 4). (E) As in panel B, except WT and CD36^−/− murine platelets were used (n = 4). *P < .05, basal compared with oxPC_CD36. (F) Human platelets were treated with nLDL or oxLDL platelets in the presence of TEMPOL (1 mmol/L), MnTMPyP (100 µmol/L), gp91ds-tat, or sc-gp91ds-tat (2 µmol/L), or left untreated (control). Representative images of adherent platelets taken under bright-field or fluorescence microscopy are shown (n = 4). (G) Suspended murine platelets (1 × 10⁸ platelets/mL), NOX2^−/− or WT, were incubated with the superoxide-detection probe for 30 minutes at 37°C and then treated with oxPC_CD36 (5 µmol/L) for 15 minutes at 37°C, and fluorescence was measured at 650 nm (n = 4 individual mice). *P < .05 compared with basal. NS, not significant.

Figure 3

oxLDL-induced ROS generation and activation of gp91^phox/NOX2 requires PKC. (A) Washed human platelets incubated with apyrase, indomethacin, and EGTA were treated with PP2 and PP3 and then with LDL before being lysed with ice-cold lysis buffer. Lysates were immunoblotted for phosphorylated PKC substrates. (i) Blots are representative of at least 3 experiments using different blood donors. (ii) Densitometric analysis of the blots. *P < .05 compared with oxLDL alone. (B) Washed human platelets (5 × 10⁷ platelets/mL) were incubated with BIMI or BIMV (both 10 µmol/L) for 20 minutes at 37°C or left untreated (control) for 20 minutes at 37°C, incubated with a fluorescent superoxide detection probe for 30 minutes at 37°C, and adhered to slides coated with nLDL or oxLDL (50 µg/mL). Images are representative of 3 independent experiments with separate blood donors. Scale bar, 20 µm. (C) Washed human platelets were incubated with apyrase, indomethacin, and EGTA to prevent secondary signaling. In addition, they were also incubated with PP2/PP3 (20 µmol/L); U73122 (5 µmol/L); BIMI, BIMV, and Ro31-8220 (10 µmol/L); FA6-152; or a control IgG (both 5 µg/mL), then lysed with ice-cold lysis buffer. p47^phox was then immunoprecipitated (IP) and immunoblotted (IB) using an antibody against phosphorylated PKC substrates. (i) Blots are representative of at least 3 experiments using different blood donors. (ii) Densitometric analysis of the blots. *P < .05, **P < .01 compared with oxLDL alone. (D) Washed human platelets were incubated with nLDL or oxLDL (50 µg/mL) for 15 minutes at 37°C, then diluted 1:1 with cell fractionation buffer (320 mmol/L sucrose, 4 mmol/L N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid, and 0.5 mmol/L Na₃VO₄, protease inhibitor cocktail [pH 7.4]). The suspensions were immediately snap-frozen in liquid nitrogen and lysates separated into soluble (S) and particulate (P) fractions by ultracentrifugation. Proteins from each fraction were separated by SDS-PAGE and immunoblotted for p47^phox, p22^phox, CD36, and Syk. (i) Blots re representative of at least 3 experiments using different blood donors. (ii) Densitometric analysis of the blots. *P < .05 compared with oxLDL alone.

Figure 4

oxLDL modulates cGMP-mediated inhibition of platelet aggregation and arrest under conditions of flow. (A) Washed human platelets (3 × 10⁸ platelets/mL) were incubated with 8-pCPT-cGMP (50 µmol/L) for 2 minutes alone or with nLDL or oxLDL (50 μg/ml) for 15 minutes followed by 8-pCPT-cGMP (50 µmol/L) for 2 minutes. Thrombin-stimulated aggregation was then measured under constant stirring (1000 rpm) at 37°C for 3 minutes. Aggregation was recorded for 3 minutes. (i) Representative aggregation traces. (ii) Percent inhibition of thrombin induced aggregation by cGMP is presented as mean ± SD (n = 4, *P < .05 compared with absence of oxLDL). (B) As in panel A, except platelets were incubated with gp91ds-tat or its scrambled control (both 2 µmol/L) prior to the addition of oxLDL. (i) Representative aggregation traces. (ii) Percent inhibition of thrombin induced aggregation by cGMP is presented as mean ± SD (n = 4). *P < .05. (C) As in panel A, except WT and NOX2^−/− murine platelets were treated with oxPC_CD36 or PAPC (5 µmol/L) for 15 minutes followed by 8-pCPT-cGMP (50 µmol/L) for 2 minutes. (i) Representative aggregation traces are shown. (ii) Percent inhibition of thrombin induced aggregation by cGMP is presented as mean ± SD (n = 3). *P < .05 compared with the absence of oxPC_CD36.

Figure 5

oxLDL inhibits cGMP signaling through a mechanism that requires CD36, Src kinases, and PKC. (A) Human platelets (3 × 10⁸ platelets/mL) were incubated with nLDL (50 µg/mL), oxLDL (50 µg/ml), oxPC_CD36 (5 µmol/L), or PAPC (5 µmol/L) prior to the addition of 8-pCPT-cGMP (50 µmol/L) for 2 minutes. Platelets were washed and lysed, and intracellular cGMP concentrations were measured by enzyme immunoassay (n = 4). (B) Human platelets (5 × 10⁸ platelets/mL) were treated with nLDL or oxLDL (50 μg/mL) for 15 minutes, lysed, and PKG1 immunoprecipitated. Immunoprecipitates were incubated with exogenous cGMP and PKG activity measured spectrophotometrically at 450 nm. *P < .05 compared with the absence of oxLDL. (C) In an in vitro kinase assay, constitutively active rPKG (0.5 µg) was incubated with rRhoA (1 µg) for 30 minutes at 37°C. In some cases, rPKG was preincubated for 30 minutes with either xanthine (100 µmol/L)/xanthine oxidase (5 mU/mL) (X/XO) or KO₂ (1 mmol/L) before addition to rRhoA. Phosphorylation was terminated by Laemmli buffer, the mixture separated by SDS-PAGE and immunoblotted for phosphoRhoA-Ser¹⁸⁸, total RhoA, or PKG1. **P < .01 for rPKG alone compared with the presence of X/XO or KO₂. (D) Washed human platelets (5 × 10⁸ platelets/mL) preincubated with apyrase (2 U/mL), indomethacin (10 µmol/L), and EGTA (1 mmol/L) were treated with oxLDL (50 µg/mL) in the presence or absence of the CD36 blocking antibody FA6-152 or IgG (5 µg/mL) prior to the addition of 8-pCPT-cGMP (50 µmol/L) for 2 minutes. Platelets were lysed, separated by SDS-PAGE, and immunoblotted (IB) for phospho–VASP-ser²³⁹ and β-tubulin. *P < .05 compared with the absence of oxLDL and oxLDL in the presence and absence of FA6. (E) As in panel D, except platelets were treated with PP2 or PP3. *P < .05 compared with the absence of oxLDL and oxLDL in the presence and absence of PP2). (F) As in panel D, except platelets were treated with Ro31-8220, BIMI, BIMV, U73122, or BAPTA-AM prior to the addition of oxLDL (50 µg/mL) for 15 minutes and 8-pCPT-cGMP (50 µmol/L) for 2 minutes. In all figures, subpanel i shows blots representative of at least 3 experiments using different blood donors, and subpanel ii shows densitometric analysis of the blots. *P < .05 compared with oxLDL alone.

Figure 6

oxLDL inhibits cGMP-signaling through a mechanism that requires NOX2 and intracellular ROS. (A) Platelets (5 × 10⁸ platelets/mL) were incubated with apyrase (2 U/mL), indomethacin (10 µmol/L), and EGTA (1 mmol/L) and treated with MnTMPyP (100 µmol/L) or TEMPOL (1 mmol/L) prior to the addition of 8-pCPT-cGMP (50 µmol/L) for 2 minutes. Platelets were lysed, separated by SDS-PAGE, and immunoblotted (IB) for phospho–VASP-ser²³⁹. (i) Blots are representative of at least 3 experiments using different blood donors. (ii) Densitometric analysis of the blots. **P < .01 compared with the absence of inhibitors. (B) As in panel A, except that in some cases, platelets were pretreated with gp91ds-tat or its scrambled control (both 2 µmol/L). Blots are representative of at least 3 separate experiments. *P < .05 for CGMP alone, cGMP/oxLDL, and oxLDL in the presence and absence of gp91da-tat. (C) As in panel A, except NOX2-deficient or WT murine platelets were treated with 8pCPT-cGMP (50 µmol/L) and oxPC_CD36 (5 µmol/L). *P < .05 compared with the absence of oxPC_CD36. NS, not significant.

Figure 7

oxLDL and hyperlipidemia caused decreased platelet sensitivity to cGMP through NOX2. (A) Human platelets were stained with DiOC₆ (1 µmol/L) for 10 minutes at 37°C, reconstituted with RBCs, treated with either nLDL or oxLDL for 15 minutes, and then perfused through fibrinogen (1 mg/mL) or bovine serum albumin (BSA)-coated capillary tubes for 2 minutes at a shear rate of 1000 s⁻¹. (i) Shown are images representative of 3 independent experiments with separate blood donors. Scale bar, 20 µm. (ii) Data are presented as surface area coverage (%) (n = 3). **P < .01, fibrinogen compared with cGMP, and oxLDL in the presence and absence of inhibitors. (B) Blood from WT and ApoE^−/− mice were stimulated with ADP (10 µM) in the presence and absence of 8pCPT-cGMP (50 µmol/L) and platelet fibrinogen binding measured by flow cytometry. (i) Data are presented as % inhibition of fibrinogen (Fgn) binding and are expressed as mean ± SEM taken from 9 individual mice for each group. *P < .05. Representative fluorescence-activated cell sorter histograms of fibrinogen binding for WT (ii), sc-gp91ds-tat (iii), and gp91ds-tat (iv). Each histogram shows fibrinogen binding at basal (red line), ADP (10 µM) (green line), and 8pCPT-cGMP (50 µmol/L)/ADP (blue line). (C) Summary of the proposed signaling pathway downstream of CD36 through which oxLDL activates gp91^phox/NOX2 and suppresses PKG signaling. sGC, soluble guanylyl cyclase; NS, not significant.