Thrombosis and thrombocytopenia are features of the antiphospholipid syndrome (APS), suggesting that antiphospholipid antibodies (aPL) may bind platelets, causing activation and aggregation of platelets and thrombosis. The intracellular events involved in aPL-mediated platelet activation are not fully understood and are therefore the subject of this study.
IgG fractions and their F(ab′)2 fragments were purified from the sera of 7 patients with APS and from the pooled sera of 10 healthy subjects (as controls). Phosphorylation of p38 MAPK, ERK-1/2, and [Ca2+]-dependent cytosolic phospholipase A2 (cPLA2) was determined in lysates of washed platelets pretreated with low doses of thrombin and aPL or control IgG or their F(ab′)2 fragments, by immunoblot. The effects of aPL on platelet aggregation in the presence or absence of a p38 MAPK inhibitor, SB203580, were examined. Thromboxane B2 (TXB2) production was detected by enzyme-linked immunosorbent assay on gel-filtered platelets treated with aPL and thrombin, with or without SB203580. Calcium mobilization studies were done utilizing a fluorometric assay.
Treatment of platelets with IgG aPL, or their F(ab′)2 fragments, in conjunction with subactivating doses of thrombin resulted in a significant increase in phosphorylation of p38 MAPK. Neither the IgG aPL nor their F(ab′)2 fragments increased significantly the phosphorylation of ERK-1/2 MAPKs. Furthermore, pretreatment of platelets with SB203580 completely abrogated the aPL-mediated enhanced aggregation of the platelets. Platelets treated with F(ab′)2 aPL and thrombin produced significantly larger amounts of TXB2 when compared with controls, and this effect was completely abrogated by treatment with SB203580. In addition, cPLA2 was also significantly phosphorylated in platelets treated with thrombin and F(ab′)2 aPL. There were no significant changes in intracellular [Ca2+] when platelets were treated with aPL and low doses of thrombin.
The data strongly indicate that aPL in the presence of subactivating doses of thrombin induce the production of TXB2 mainly through the activation of p38 MAPK and subsequent phosphorylation of cPLA2. The ERK-1/2 pathway does not seem to be involved in this process, at least in the early stages of aPL-mediated platelet activation.
Antiphospholipid antibodies (aPL) have been associated with recurrent thrombosis (arterial and/or venous) and recurrent pregnancy losses in patients with systemic lupus erythematosus and in those with the antiphospholipid syndrome (APS) (1). Thrombocytopenia is a frequent feature of APS, giving rise to the speculation that aPL may play a pathogenic role in thrombosis by binding platelets and causing platelet activation and aggregation (2).
Studies have demonstrated binding of affinity-purified aPL to platelets (2–5). Lellouche and colleagues (6) reported that urinary secretion of the major thromboxane metabolite, 11-dehydrothromboxane B2 (TXB2), was significantly increased in patients with lupus anticoagulant (LAC) as compared with normal controls. Studies from our group have also shown that affinity-purified anticardiolipin antibodies (aCL) from patients with APS, but not from patients with syphilis, enhanced activation of platelets treated with suboptimal doses of ADP, thrombin, or collagen (7). In a recent study by our group, platelets pretreated with suboptimal doses of thrombin receptor agonist peptide (TRAP) and aPL expressed enhanced levels of activated glycoprotein IIb/IIIa, indicating platelet activation (8). In another study, rabbit aCL were shown to enhance collagen-induced platelet activation (9). Robbins et al showed that aPL–β2-glycoprotein I (β2GPI) complexes significantly increased production of thromboxane A2 (TXA2), a proaggregatory prostanoid in platelets (10).
Platelets contain family members of the MAPKs, including ERK-1 (p44 MAPK), ERK-2 (p42 MAPK), and p38 MAPK (Figure 1). The MAPK p38 is a member of a family of proline-directed serine/threonine kinases that is dual-phosphorylated on a threonine and tyrosine residue, separated by 1 single amino acid (11, 12). In platelets, p38 MAPK is activated by stress, such as heat and osmotic shock, arsenite, H2O2, α-thrombin, collagen, and thromboxane analog (11, 12), and is involved in the phosphorylation of [Ca2+]-dependent cytosolic phospholipase A2 (cPLA2), with subsequent production of TXB2 (Figure 1). Thrombin has also been shown to induce phosphorylation of ERK-1/2, involving protein kinase C (PKC), phospholipase Cβ (PLCβ), and the intracellular mobilization of [Ca2+] (Figure 1) (13–15).
Although several studies have shown that aPL enhance platelet activation in vitro in the presence of low doses of agonists (ADP, thrombin, collagen, or TRAP) (7–10, 16), the intracellular events involved in this process are not understood. To address this question, we examined the effects of aPL on phosphorylation of p38 MAPK, ERK-1/2 MAPKs, and cPLA2 on intracellular [Ca2+] mobilization and on TXB2 production in the presence of subactivating doses of thrombin. The effects of the specific inhibitor for p38 MAPK, SB203580 (4-[4-fluorophenyl]-2-[4-methylsulfinylphenyl]-5-[4-pyridyl] 1-imidazole), on aPL-mediated enhancement of platelet aggregation and on TXB2 production in the presence of thrombin were also determined.
PATIENTS AND METHODS
Purification of IgG and preparation of F(ab′)2 fragments.
IgG from 7 patients with APS (in accordance with the Sapporo criteria ) and from the pooled sera of 10 healthy subjects (as controls) were affinity purified by protein G (Gamma BindG Type 3; Amersham-Pharmacia, Piscataway, NJ) by elution with 0.5M acetic acid (pH 3) and neutralization of the fractions in 1M Tris buffer (pH 9). IgG were subsequently dialyzed in a 0.1M acetate buffer, pH 4.5, for 2 hours at 4°C, and digestion with pepsin to obtain F(ab′)2 fragments was performed by incubation with pepsin-agarose beads (Sigma-Aldrich, St. Louis, MO) for 18 hours at 37°C. The beads were separated by centrifugation for 10 minutes at 4,000 revolutions per minute, and the supernatants were dialyzed against phosphate buffer, pH 7.4, at 4°C. Intact Fc fragments and nondigested antibody were removed by protein G–sepharose chromatography. F(ab′)2 fragments showed a single band at ∼110 kd in nonreduced sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) (utilizing silver staining). IgG fractions and F(ab′)2 were passed through a 0.22-μm filter and were checked for the absence or presence of endotoxin (lipopolysaccharide [LPS]) by the Limulus amebocyte lysate assay (Amebolysate; ICN Biomedical, Costa Mesa, CA).
Determination of antibodies.
The aCL and antiphosphatidylserine (aPS) antibodies in the IgG fractions were determined by enzyme-linked immunosorbent assay (ELISA) as previously described (18), using an anti-human IgG antiserum labeled with alkaline phosphatase (γ-chain specific). Similarly, binding of the F(ab′)2 fragments to cardiolipin was determined by ELISA, using an anti-human κ and λ secondary antibody cocktail (1:1,000 dilution) labeled with alkaline phosphatase (Sigma-Aldrich). Titers of aCL were reported in IgG phospholipid (GPL) units, and levels of aPS antibodies were reported in net optical density (OD) units.
Anti-β2GPI antibodies were detected by ELISA as previously described, with some modifications, using irradiated Costar microtiter plates (Corning, Corning, NY) (19). Plates were coated with 20 μg/ml purified β2GPI, and blocked with 5% ovalbumin–phosphate buffered saline (PBS) solution. Anti-human IgG (γ-chain specific) and anti-κ and anti-λ secondary antibodies labeled with alkaline phosphatase were used. Anti-β2GPI activities were determined at 405 nm, with results expressed in net OD units. LAC activities of the IgG and of the F(ab′)2 fragments were determined by a modified activated partial thromboplastin time (APTT) as previously described (20).
Isolation of platelets.
Platelets were obtained from the blood of adult healthy volunteers who had not received medication for at least 10 days. Blood was collected in acid–citrate–dextrose anticoagulant (9/1 volume/volume). Plasma-rich platelets (PRPs) were obtained by centrifugation for 20 minutes at 120g at room temperature (21).
Preparation of washed platelets.
PRPs were centrifuged at 800g for 10 minutes at room temperature and resuspended in HEPES buffer (10 mM HEPES, 140 mM NaCl, 3 mM KCl, 0.5 mM MgCl2, 5 mM NaHCO3, 10 mM dextrose, pH 7.4) supplemented with 0.4 units/ml apyrase and 1 mM aspirin. The platelet count was adjusted to 1 × 109 platelets/ml and these preparations were used in the immunoblot assays and in the calcium mobilization experiments.
Preparation of gel-filtered platelets (GFPs).
For the aggregation studies, GFPs were used. Platelets were filtered through a Sepharose 2B column equilibrated with HEPES buffer as previously described (21) without inhibitors, and the platelet count was adjusted to 250 ± 50 × 106/ml for the aggregation studies and 300 × 106 platelets/ml for the TXB2 assay. Platelets were maintained at 37°C in a closed (capped) tube and analyzed within 2 hours after drawing.
Immunoblot analysis for phosphorylation of p38 and p44/42 MAPKs.
One hundred–microliter aliquots of washed platelets were incubated in microtubes for 5 minutes at 37°C, under static conditions, with 50, 100, or 200 μg/ml IgG aPL from APS patients or control IgG, or with 200 μg/ml of F(ab′)2 aPL from APS patients or control F(ab′)2 IgG. Platelets were then treated with thrombin (0.005 units/ml). Platelets to be used as controls were treated with PBS and 1 unit/ml thrombin.
Platelets were subsequently lysed with Laemmli buffer (Bio-Rad, Richmond, CA) and 2-mercaptoethanol, and lysates were heated at 95°C for 10 minutes and centrifuged for 10 minutes at 11,000g. Samples were subjected to SDS-PAGE in a 12% gel. After the electrophoresis, proteins were transferred to nitrocellulose membranes at 10V for 30 minutes in a semidry transfer cell. The membrane was then blocked and subsequently incubated with one of the following antibodies: p38 MAPK (Thr180/Tyr182) 28b10 mouse anti-human monoclonal antibody (Cell Signals Technology, Beverly, MA) or p38 MAPK rabbit polyclonal antibody (Santa Cruz Biotechnology, Santa Cruz, CA), or with phosphorylated ERK-1/2 (Thr202/Tyr204) E10 mouse anti-human monoclonal antibody (Cell Signals Technology) or ERK-1/2 rabbit polyclonal antibody (Santa Cruz Biotechnology), followed by incubation with secondary antibody/goat anti-mouse or goat anti-rabbit IgG labeled with peroxidase (21) and detection by chemiluminescence (Amplight/BioRad, Hercules, CA). The intensity of the bands was quantitated by densitometric analysis using Gel Pro Analyzer software, version 4.5 (Fotodyne, Hartland, WI).
Phosphorylation of cPLA2.
One hundred microliters of purified platelets in a concentration of 1 × 109 platelets/ml (prepared as described above) was incubated with 50 μl (200 μg/ml final concentration) of F(ab′)2 aPL from APS patients (n = 5) or 50 μl control F(ab′)2 IgG or PBS for 5 minutes, and then stimulated with 0.005 units/ml thrombin, or treated with PBS and 1 unit/ml thrombin as a positive control. The reaction was stopped by the addition of an equal volume of 2× immunoprecipitation buffer (100 mM Tris HCl, pH 7.4, 300 mM NaCl, 2 mM EGTA, 5 μg/ml leupeptin, 5 μg/ml aprotinin, 2 mM phenylmethylsulfonyl fluoride, 2 mM Na3VO4, 2 mM NaF, 2% Nonidet P40, 0.5% sodium deoxycholate). The lysates were cleared by treating them with protein G–Sepharose for 45 minutes.
The cleared supernatants were incubated with 1 μg of cPLA2 antibody (Santa Cruz Biotechnology) for 2 hours at 4°C. The immune complexes were precipitated by addition of 90 μl of protein G–Sepharose for 45 minutes. After brief centrifugation, immunoprecipitates were washed 3 times with 1 ml of 1× immunoprecipitation buffer and then resuspended with 25 μl of SDS–sample buffer, and subjected to SDS-PAGE in a 10% gel for 210 minutes. The proteins in the gel were transferred to a nitrocellulose paper and were immunoblotted as described previously, using the same cPLA2 antibody (rabbit polyclonal cPLA2; Santa Cruz Biotechnology). The phosphorylated form of cPLA2 is detected as a slower-moving second band. Thrombin-induced anti-cPLA2 was detectable after 2 minutes of stimulation (approximately half of the protein was in the phosphorylated form) (21).
Aggregation of platelets was performed as follows: a 420-μl aliquot of GFPs and 1 μl of CaCl2 (1 mM final concentration) were placed in the aggregometer cuvettes at 37°C with 30 μl of IgG aPL antibodies from APS patients (final concentration 200 μg/ml) or 30 μl of control IgG (final concentration 200 μg/ml) for 5 minutes. In some experiments, platelets were pretreated for 20 minutes with SB203580 (0.1 and 1 μM) or with 1% dimethylsulfoxide (22). Aggregation was then started in a dual channel aggregometer (Minigator II; Payton Scientific, Buffalo, NY) and registered in a linear recorder. Platelet aggregation was evaluated as the amplitude of the aggregation curves (expressed in millimeters) after 5 minutes.
GFPs were adjusted to a concentration of 3 × 108 platelets/ml in HEPES buffer and incubated for 5 minutes with (Fab′)2 aPL fragments from APS patients or control F(ab′)2 IgG fragments or PBS in the presence or the absence of 1 μM/liter SB203580, and then stimulated with 0.005 units/ml thrombin for 3 minutes. The production of TXB2 was stopped by the addition of indomethacin (0.1 mM) in HEPES buffer, and the preparations were centrifuged at 10,000 rpm for 30 seconds at 4°C. The supernatants were rapidly separated and stored at −20°C until analyzed. TBX2 was analyzed in the supernatants utilizing a commercial enzyme immunoassay system (Amersham-Pharmacia) as indicated by the manufacturer.
Intracellular [Ca2+] mobilization experiments.
Washed platelets were incubated for 20 minutes at 37°C with 5 μM visible light–excitable [Ca2+] indicator Fura Red–acetoxymethyl ester (AM) and 5 μM calcium green–AM (Molecular Probes, Eugene, OR). Subsequently, platelets were centrifuged at 800g for 10 minutes at room temperature and resuspended in HEPES buffer together with 0.4 units/ml apyrase VII, 0.1 μM prostaglandin E1, and 1 mM acetylsalicylic acid, and standardized at a platelet count of 1 × 109/ml. One hundred–microliter aliquots of platelets were then incubated with 200 μg/ml F(ab′)2 aPL from APS patients or 200 μg/ml control F(ab′)2 IgG or PBS for 5 minutes, and then stimulated with 0.01 units/ml thrombin. Each experiment included a calibration curve utilizing 0.01, 0.1, 0.5, and 1 units/ml thrombin.
The emission peaks of calcium green and Fura Red signal intensities were determined at 530/30 nm and 640/40 nm, respectively (21). Elevations in [Ca2+] result in increased fluorescence intensity of calcium green and decreased fluorescence intensity of Fura Red, at 485/20 nm excitation wavelength. Changes in [Ca2+] were evaluated on the basis of a ratio of signal intensity at the G-channel (calcium green, sensitive peak at 530 nm) relative to the R-channel (red, sensitivity peak at 640 nm) (21).
Data are presented as the mean ± SD. Student's t-test was used to compare mean values between treated platelets and control platelets. P values less than or equal to 0.05 were considered significant.
Characterization of IgG aPL and F(ab′)2 aPL.
All of the IgG aPL samples from the 7 APS patients were positive for aCL (>100 GPL units) and all of the IgG samples from the healthy controls were negative for aCL (<10 GPL units), when preparations were tested at a 200 μg/ml protein concentration. The anti-β2GPI antibody activities of the IgG aPL and control IgG were a mean ± SD 0.870 ± 0.218 net OD units and 0.140 ± 0.019 net OD units, respectively, when samples were tested at a 100 μg/ml protein concentration. Six of the 7 IgG aPL samples were positive for LAC (APTT ratio of patient plasma to normal plasma 1.21:1.50). All F(ab′)2 aPL samples were positive for binding to cardiolipin (25–186 GPL units/100 μg protein), and the anti-β2GPI activities of the F(ab′)2 aPL and control F(ab′)2 IgG were a mean ± SD 0.505 ± 0.111 net OD units and 0.068 ± 0.024 net OD units, respectively (significantly different; P = 0.0023). Binding to phosphatidylserine was tested by ELISA and was a mean ± SD 0.506 ± 0.112 net OD units for the F(ab′)2 aPL preparations and 0.018 ± 0.005 net OD units for control F(ab′)2 IgG (significantly different; P = 0.0002). All IgG preparations and F(ab′)2 fragments tested negative for LPS in the Limulus amebocyte lysate assay.
Effects of aPL on phosphorylation of p38 MAPK in platelets.
The treatment of washed platelets with varying concentrations of thrombin (0.005–5 units/ml) produced a dose-dependent phosphorylation of p38 MAPK (Figure 2A). Treatment of platelets with control IgG and 0.005 units/ml thrombin produced phosphorylation that was not different from that of platelets treated with 0.005 units/ml thrombin alone. In contrast, platelets treated with IgG aPL (n = 7) at 200 μg/ml protein concentration and 0.005 units/ml thrombin produced a significant increase in phosphorylation of p38 MAPK (among the 7 IgG aPL samples, fold increases of 4.4, 7.8, 7.4, 5.8, 7.5, 6.3, and 5.0, when compared with platelets treated with control IgG and thrombin) (Figure 3A). A 7.6-fold increase was observed in platelets treated with 0.2 units/ml of thrombin alone (positive control).
The effect was dependent on the amount of aPL used. For example, for sample IgGaPL1, a 4.4-, 3.0-, and 1.9-fold increase in p38 MAPK phosphorylation was observed with 200, 100, and 50 μg/ml protein concentrations, respectively. Treatment of platelets with IgG aPL alone (n = 7) (in the absence of thrombin) did not produce any effect on p38 MAPK phosphorylation (Figure 3B).
To determine whether the effects produced by IgG aPL were due to interactions of the antibodies with platelets either through the Fab fragment or through the Fc portion of the immunoglobulins, the effects of F(ab′)2 fragments on phosphorylation of p38 MAPK on platelets were determined by immunoblot. As shown in Figures 4A and B, 4 of the 5 F(ab′)2 aPL preparations produced a significant increase of phosphorylation (12.6-, 4.6-, 3.7-, and 12.6-fold increases) when compared with platelets treated with control F(ab′)2 IgG and 0.005 units/ml thrombin. No increase in phosphorylation of p38 MAPK was observed in platelets treated with control F(ab′)2 IgG and thrombin when compared with platelets treated with thrombin alone.
Effects of aPL on phosphorylation of ERK-1/2.
Treatment of platelets with different concentrations of thrombin (0.005–5 units/ml thrombin) produced a dose-dependent phosphorylation of ERK-1/2 (Figure 2B). Phosphorylation of ERK-1/2 was observed starting at 0.2 units/ml thrombin. Neither the platelets treated with F(ab′)2 aPL (n = 5) at 200 μg/ml protein concentration and 0.005 units/ml thrombin nor the platelets treated with control F(ab′)2 IgG and 0.005 units/ml thrombin induced phosphorylation of ERK-1/2 after 2 minutes of treatment (Figure 4C).
Effects of a specific inhibitor of p38 MAPK on aPL-mediated platelet aggregation.
To confirm whether the phosphorylation of p38 MAPK is involved in aPL-mediated platelet activation, aggregation studies were performed in the presence and in the absence of the specific inhibitor of p38 MAPK, SB203580. As shown in Figure 5, platelets treated with IgG aPL and 0.005 units/ml thrombin produced an aggregation of 55%, whereas there was no aggregation in platelets treated with control IgG and 0.005 units/ml thrombin. Pretreatment of the platelets with 0.1 μM or 1 μM SB203580 abrogated the aggregatory effects of aPL by 89% and 100%, respectively.
Effects of aPL on phosphorylation of cPLA2.
Treatment of washed platelets with 1 unit/ml thrombin induced strong phosphorylation of cPLA2 after 2 minutes (Figure 6). Four of the 5 F(ab′)2 aPL fragments induced significant phosphorylation of cPLA2 in platelets pretreated with 0.005 units/ml thrombin. Neither treatment with 0.005 units/ml thrombin nor treatment with control F(ab′)2 IgG and 0.005 units/ml thrombin induced phosphorylation of cPLA2 (Figure 6).
Effects of aPL on TXB2 production.
The production of TXB2 was significantly increased in platelets treated with F(ab′)2 aPL and thrombin, when compared with platelets treated with control F(ab′)2 IgG and thrombin (mean ± SD levels of TXB2 535.0 ± 50.9 pg/ml versus 251.0 ± 12.7 pg/ml; P = 0.0027) or with thrombin alone (230.0 ± 12.5 pg/ml). The mean values of TXB2 produced by platelets treated with control F(ab′)2 IgG and those treated with 0.005 units/ml thrombin alone were not statistically significantly different (Table 1). The effect of F(ab′)2 aPL on TXB2 production was completely abrogated by pretreatment of the platelets with SB203580 (mean ± SD levels of TXB2 280.3 ± 57.0 pg/ml), indicating that the activation of p38 MAPK is involved in aPL-mediated production of TXB2 by platelets (Table 1).
Table 1. Effects of F(ab′)2 antiphospholipid antibodies (aPL) on thromboxane B2 (TXB2) production in platelets*
TXB2 production, mean ± SD pg/ml
Platelets were treated with 0.005 units/ml thrombin and with F(ab′)2 aPL (n = 5) or F(ab′)2 IgG from normal healthy subjects (IgG-NHS) (n = 5), and TXB2 was measured by enzyme-linked immunosorbent assay, as described in Patients and Methods. In some experiments, platelets were pretreated with 1 μM SB203580. Experiments were repeated 4 times.
P < 0.0027 versus platelets treated with F(ab′)2 IgG-NHS plus thrombin.
P not significant versus platelets treated with F(ab′)2 IgG-NHS plus thrombin.
Effects of aPL on intracellular [Ca2+] mobilization.
The treatment of platelets with different concentrations of thrombin (0.01, 0.1, 1, and 5 units thrombin) produced a significant and dose-dependent increase in intracellular calcium, with a peak between 2 and 3 minutes of treatment (G-channel to R-channel mean ratios 2.1, 2.8, 3.2, and 5.7, respectively) (Figure 7). There was no significant increase in intracellular calcium concentrations for a period of 5 minutes when platelets were treated with IgG aPL and 0.01 units/ml thrombin, as compared with platelets treated with 0.01 units/ml thrombin alone or platelets treated with control IgG and thrombin (mean G-channel to R-channel ratios at 3 minutes 2.22, 2.1, and 2.2, respectively) (Figure 7). The experiments were run 7 times and the results shown in Figure 7 correspond to 1 representative experiment.
Studies have shown conclusively that aPL are thrombogenic in in vivo animal models (23–25). The prothrombotic properties of aPL may be explained in part by their ability to enhance the activation of platelets. Moreover, aPL have been shown to increase production of TXB2 in the presence of low doses of ADP, collagen, or thrombin (7, 10, 26, 27). A recent study by our group showed a significant increase in the expression of activated glycoprotein IIb/IIIa on platelets treated with aPL and TRAP (8). Furthermore, aPL-enhanced thrombosis in vivo can be abrogated by infusions of a glycoprotein IIb/IIIa antagonist (1B5) monoclonal antibody and in β3-null (glycoprotein IIb/IIIa–deficient) mice (Vega-Ostertag M, et al: unpublished observations). In this study, we confirmed that aPL have a direct effect on platelet activation in the presence of thrombin, and we examined the intracellular pathways involved in this process.
The activation of platelets by thrombin has been shown to involve more than one pathway, comprising the p38 MAPK pathway, including the downstream calcium-dependent phosphorylation of cPLA2, or the PLCβ pathway, involving activation of PKC, phosphorylation of ERK1/ERK2 (Figure 1), and the intracellular mobilization of [Ca2+]. In both cases, TXB2 is produced and platelets are activated (11–15).
In this study, we found that phosphorylation of p38 MAPK mediates aPL-induced activation of platelets in vitro. The degree of phosphorylation of this enzyme varied among the 7 different preparations of aPL (4.4–7.9-fold increase) and the F(ab′)2 aPL fragments (3.7–12.6-fold increase over the control) used in this study. This variability is not surprising. Antiphospholipid antibodies are known to be heterogeneous in specificity and function, and have been shown to bind negatively charged phospholipids, β2GPI, prothrombin, annexin V, and other proteins of the coagulation cascade. A wide variety of functions have been attributed to aPL, from activation of endothelial cells, up-regulation of tissue factor in monocytes, and platelet activation.
The present data conclusively show that aPL induce phosphorylation of p38 MAPK after pretreatment with low (subactivating) doses of thrombin. Furthermore, these effects are dependent on the dose of antibody utilized and are abrogated by pretreatment of the platelets with the specific inhibitor of the enzyme, SB203580, as shown in the aggregation studies and in the TXB2 production experiments.
We also show that aPL up-regulate production of TXB2 in platelets. This is consistent with the study by Opara et al that recently demonstrated an increase in platelet TXB2 production and in aggregation by aCL–β2GPI complexes (9, 28). Furthermore, in our study, we show that pretreatment of platelets with SB203580 completely abrogates the production of TXB2 induced by aPL and low doses of thrombin, and that cPLA2 is significantly phosphorylated in platelets treated with low doses of thrombin and F(ab′)2 aPL, an event downstream of p38 MAPK activation (Figure 1). Therefore, the data conclusively show involvement of that enzyme in aPL-mediated platelet activation. These data are also consistent with the suggestions by Gonzalez-Buritica et al, who, in 1988, hypothesized that phospholipase A2 plays a role in platelet activation in patients with aPL (29).
The ERK-1/2 phosphorylation pathway may also be initiated in platelets by thrombin (Figure 1) (14, 15). Interestingly, the data from our study show that treatment with aPL and low doses of thrombin does not induce phosphorylation of ERK-1/2. Initial in vitro studies performed in transfected cells and in HeLa cells suggested that p42 MAPK phosphorylates cPLA2 at Ser505, which lies within a consensus sequence for MAPK (Pro-Xaa-Ser/Thr-Pro). Subsequently, other investigators reported the concomitant activation of MAPK and phosphorylation of cPLA2 in stimulated cells (30–34).
Several recent reports, however, dissociated cPLA2 phosphorylation from MAPK activation in platelets. One study showed that phosphorylation of ERK-1/2 is not required for phosphorylation of cPLA2 in thrombin-stimulated platelets (35). Those authors concluded that cPLA2 is the physiologic target of p38 MAPK, and that ERK-1/2 phosphorylation of cPLA2 is not required for its receptor-mediated activation in platelets (35). Borsch-Haubold et al have shown no effect on phosphorylation of cPLA2 or release of thromboxane when the specific inhibitor of PKC, Ro31-8220, was used in platelets stimulated with thrombin or collagen (12, 35–38). Similarly, the same group of authors showed that inhibition of MAPK kinase using PD98059 did not affect platelet responses to the physiologic stimuli, thrombin and collagen, indicating a role for p38 MAPK in primary activation of human platelets, independent of ERK-1/2 (35–39).
Altogether, these results are in agreement with our findings and provide evidence against a role for ERK-1/2 in primary aPL-mediated platelet aggregation. However, we do not exclude the possibility that these MAPKs may play a role in postaggregation events in platelets mediated by aPL.
We did not find significant changes in intracellular [Ca2+] when platelets were treated with F(ab′)2 aPL and low doses of thrombin, supporting the hypothesis that the PLCβ pathway and the downstream phosphorylation of ERK-1/2 activation are not involved in aPL-mediated platelet activation (see Figure 7). The p38 MAPK–dependent activation of cPLA2 in platelets appears to be dependent on intracellular calcium concentrations. In our studies, although no significant changes in calcium concentrations were observed when aPL were added to the system, treatment of platelets with 0.01 units/ml thrombin induced a modest increase in calcium (as shown in Figure 7). We speculate that the observed change may have been sufficient to initiate cPLA2 activation, but these observations would need to be further evaluated to confirm this hypothesis.
The effects of aPL on phosphorylation of p38 MAPK and of cPLA2 and production of TXB2 by platelets reported in this study are due to the influence of the F(ab′)2 fragment and not to the Fc portion of the antibody. This finding is consistent with the findings of the study by Robbins et al, which showed that F(ab′)2 aCL fragments significantly stimulated TXB2 production in platelets (10).
The present study did not focus on establishing the nature of the receptors to which aPL bind on platelets. Studies have shown that aPL bind only to platelets previously exposed to low doses of agonists or to platelets that have been frozen and thawed repeatedly, exposing negatively charged phospholipids (i.e., phosphatidylserine) (2). However, the precise nature of the receptors for aPL in platelets is not known. Opara et al previously hypothesized that β2GPI might mediate aCL binding to the activated platelet cell surface by binding with phosphatidylserine, thereby promoting increased platelet activation by the aCL–β2GPI complexes (28). A recent study demonstrated that dimeric β2GPI binds to members of the low-density lipoprotein receptor family in platelets and induces increased platelet adhesion to collagen (40). This effect was increased by addition of anti-β2GPI monoclonal antibodies to the system and was abrogated by inhibition of thromboxane synthesis.
In summary, our study is the first to show that aPL-mediated platelet activation occurs selectively through the p38 MAPK pathway. Upon priming of the platelets with aPL and low doses of thrombin, cPLA2 is phosphorylated and TXB2 is produced. PKC and ERK-1/2 activation do not seem to be involved in this response. These findings may be important in designing new approaches to targeted treatment of thrombosis in APS patients.
We are grateful to Dr. Jacob Rand of Albert Einstein School of Medicine (New York, NY) for his helpful comments on this work.