Role for Sulfatides in Platelet Aggregation
Background— Sulfatides are sulfated glycosphingolipids present on the surface of oligodendrocytes, renal tubular cells, and certain tumor cells. They appear to be involved in nerve conduction and cell adhesion, but their precise physiological function is not known.
Methods and Results— Here, we show a novel role for sulfatides as a major ligand for P-selectin in platelet adhesion and aggregation. Sulfatides are expressed on the platelet surface, and platelets expressing sulfatides adhere to P-selectin. Both sulfatide micelles and sulfatide-binding recombinant malaria circumsporozoite protein (MCSP) inhibit this adhesion. In parallel, platelets and CHO cells expressing P-selectin adhere to sulfatides, and anti-P-selectin antibodies inhibit this adhesion. Furthermore, both anti-P-selectin antibodies and sulfatide antagonist MCSP significantly reverse platelet aggregation induced by ADP, collagen, or thrombin receptor-activating peptide, suggesting that sulfatide-P-selectin interactions are necessary for the formation of stable platelet aggregates.
Conclusions— These results show that sulfatide interactions with P-selectin are important in platelet adhesion and platelet aggregation. The sulfatide interactions with P-selectin stabilize platelet aggregates, representing a new mechanism of platelet aggregation that may play a significant role in hemostasis and thrombosis.
Received February 14, 2001; revision received September 27, 2001; accepted October 4, 2001.
Sulfatides are 3-sulfated galactosyl ceramides expressed on the surface of oligodendrocytes, renal tubular cells, and certain tumor cells.1,2 Sulfatides appear to be involved in nerve conduction3 and cell adhesion,1 but their precise physiological function is not known. Sulfatides interact with several cell adhesion molecules, such as laminin, thrombospondin, von Willebrand factor (vWF), and P-selectin.1,2 Furthermore, cell surface sulfatides provide a binding site for human immunodeficiency virus,4 Helicobacter pylori,5 and malaria sporozoites.6
Platelets aggregate to form a hemostatic plug during hemostasis.7 After activation, platelet glycoprotein (GP) IIb/IIIa receptor becomes competent to bind soluble fibrinogen, which bridges GP IIb/IIIa between adjacent platelets.7 As activation progresses, platelets secrete their granular contents, and P-selectin, which is present in the α-granules, translocates to the outer surface.8 During this process, there is progressive stabilization of GP IIb/IIIa-fibrinogen interactions,9 necessary for the formation of stable platelet aggregates. We have recently shown that the expression of P-selectin on platelets determines the size and stability of the platelet aggregates.10
Here, we show that platelet sulfatides serve as a major ligand for P-selectin and that this interaction is necessary for the formation of stable platelet aggregates. This novel mechanism may play a significant role in the formation of stable platelet aggregates during hemostasis and thrombosis.
Antibodies and Reagents
Monoclonal anti-sulfatide antibody Sulph I was a generous gift from Dr J.-E. Mansson and Dr P. Fredman, University of Göteborg, Göteborg, Sweden. F(ab′)2 fragments were prepared from polyclonal anti-vWF and anti-thrombospondin antibodies (Calbiochem). The sources of all other antibodies have been described previously.10 Sulfatides obtained from Matreya and Avanti were chromatographically pure. All other reagents were purchased from Sigma.
Characterization of Recombinant Malaria Circumsporozoite Protein
The cDNA for recombinant malaria circumsporozoite protein (MCSP) was a generous gift from Dr Photini Sinnis, New York University Medical Center, New York, NY. Recombinant MCSP was purified from Escherichia coli HB2151 by metal chelate affinity chromatography and labeled with biotin. Wells of 96-well microtiter plates (MaxiSorp F96, Nunc) were coated with 1 μg glycolipids/well (by evaporation of 50 μL of a 20-μg/mL methanol solution) or soluble recombinant human P-selectin (by overnight incubation with 50 μL of a 20-μg/mL solution in Tris-buffered saline [TBS: 0.15 mol/L NaCl, 10 mmol/L, Tris, pH 7.5] at 4°C). Subsequently, the wells were blocked with 5% BSA, and biotin-labeled recombinant MCSP (2.5 μg/mL) was added to the wells in TBS containing 1% BSA. After a 60-minute incubation, the bound protein was quantified by peroxidase-labeled monoclonal anti-biotin antibody, with O-phenylenediamine as a substrate, and optical density was measured in an ELISA reader (MR 5000, Dynatech), as previously described.11 Nonspecific binding of biotin-labeled MCSP, determined in the presence of a 80-fold molar excess of unlabeled MCSP, was subtracted from total binding to obtain specific binding. Nonspecific binding was 11% of total binding.
Platelet Adhesion Assays
Washed platelets were prepared as previously described.11 Wells of 96-well microtiter plates were coated with glycolipids or soluble recombinant human P-selectin and blocked as described above. Platelets (1.25×106 to 5×106/well) were added in HEPES-buffered saline containing 1% BSA and 1 mmol/L CaCl2. After incubation for 60 minutes at 37°C, nonadherent platelets were removed by 3 vigorous washes, and the bound platelets were quantified by use of rabbit polyclonal anti-GP IIb/IIIa antibody, followed by peroxidase-conjugated protein A and O-phenylenediamine, as described before.11 The effects of various antibodies (each monoclonal antibody at 35 μg/mL, each polyclonal antibody at 100 μg/mL), recombinant MCSP (35 μg/mL), recombinant annexin V (35 μg/mL), sulfatide and phosphatidylcholine micelles (each 35 μg/mL), or dextran sulfate (MW 500 000) (10 μg/mL) were determined by incubating platelets at 37°C for 5 minutes before activation with 20 μmol/L thrombin receptor-activating peptide (TRAP), SFLLRNA. Platelet aggregation was performed as described previously.10
CHO Cell Adhesion Assays
Chinese hamster ovary (CHO) cells expressing a phosphatidylinositol glycan-linked form of P-selectin (CHO-P) were a generous gift from Dr Jose Lopez, Baylor College of Medicine, Houston, Tex. CHO cells and CHO-P cells were detached from culture dishes with 0.5 mmol/L EDTA, washed twice, and then resuspended at 2.5×106 cells/mL in DMEM/F12 containing 1% BSA and 1 mmol/L CaCl2, as previously described.12 Aliquots of 0.1 mL of the cell suspensions were added to microtiter wells that were coated with various glycolipids, as described above. After a 45-minute incubation at 37°C, nonadherent cells were removed by 3 rinses with DMEM/F12. Adherent cells were fixed with 4% paraformaldehyde. Subsequently, the adherent cells were stained with amido black solution (0.1% amido black in 10% methanol, 2% acetic acid) for 30 minutes. After elution of the dye by 150 μL 50-mmol/L NaOH, the adherent cells were quantified by measurement of the optical density at 630 nm.
Washed platelets (≈2.5×105/μL), either unactivated or activated with 20 μmol/L SFLLRNA, were incubated with 25 μg/mL Sulph I antibody or PL-1 for 45 minutes, fixed with paraformaldehyde (final concentration 1%), followed by 1 μg/mL FITC-labeled polyclonal goat anti-mouse IgG antibody for 30 minutes. In addition, platelets were incubated with 1 μg/mL biotin-labeled MCSP followed by 1 μg/mL FITC-labeled avidin (Sigma). The fluorescence intensity of events in a preset platelet gate was measured with a FACScan flow cytometer (Becton Dickinson).
All experimental values are represented as mean±SD. Statistical significance was evaluated by 1-way ANOVA followed by the Tukey-Kramer multiple comparisons test. Values of P<0.05 were considered statistically significant.
Platelet Adhesion to Sulfatides
We examined the adhesion of platelets to various glycolipids. Platelets did not bind significantly to galactosyl ceramides (cerebrosides) I and II, gangliosides GM3 and GD3, sphingomyelin, or phosphatidylcholine (Figure 1a). Platelets did adhere significantly to sulfatides, however, which are 3-sulfated galactosyl ceramides, and this adhesion increased by up to 100% after activation (Figure 1a). When viewed with the phase-contrast microscope, the adherent platelets were spread on the surface in a single cell layer without aggregates. Platelet adhesion to sulfatides was maximal at a sulfatide concentration of 1 μg/well (data not shown).
The monoclonal anti-P-selectin antibodies CLB-thromb/6 and G1, both directed against the lectin domain of P-selectin,15 inhibited platelet adhesion to sulfatides by ≈60%, whereas an isotype-matched control antibody had no effect (Figure 1c). Furthermore, F(ab′)2 fragments of polyclonal antibodies against vWF or thrombospondin as well as the chimeric anti-GP IIb/IIIa Fab fragment abciximab had no effect (Figure 1c). These results indicate that platelet adhesion to sulfatides was mediated via platelet P-selectin but not via vWF, thrombospondin, or GP IIb/IIIa. Thus, the increased binding of platelets to sulfatides after activation is explained by the fact that platelets express more P-selectin on their surfaces after activation.
MCSP bound strongly to sulfatides, with a 6-fold lower binding to phosphatidylserine and no significant binding to other glycolipids, including phosphatidylinositol, even at saturating levels of lipid coating (1 μg/well) (Figure 1b). MCSP inhibited the adhesion of platelets to sulfatides by 75% (Figure 1c). Because MCSP bound weakly to phosphatidylserine, we examined the effect of annexin V, which strongly binds to phosphatidylserine on platelet adhesion to sulfatides (Figure 1c). Annexin V had no effect on platelet adhesion to sulfatides, suggesting that phosphatidylserine did not contribute significantly to the interaction of platelets with sulfatides. Furthermore, the sulfatide antagonist high-molecular-weight dextran sulfate inhibited the adhesion of platelets to sulfatides by ≈60% (Figure 1c).
CHO-P Cell Adhesion to Sulfatides
CHO-P cells exhibited significant adherence to sulfatides, whereas control CHO cells adhered only minimally (n=3, P<0.001) (Figure 2a). The monoclonal anti-P-selectin antibody CLB-thromb/6 inhibited the adhesion of CHO-P cells to sulfatides almost completely (n=3, P<0.001) (Figure 2a). In agreement with the observed platelet adhesion pattern, CHO-P cells adhered mainly to sulfatides but not to other glycolipids, such as cerebrosides I and II, gangliosides GM3 and GD3, sphingomyelin, or phosphatidylcholine (Figure 2b).
Sulfatide Expression on Platelets
The presence of sulfatides in human platelets has been described previously.1 Here, we investigated the expression of sulfatides on the platelet surface using the binding of a monoclonal anti-sulfatide antibody, Sulph I, and biotin-labeled MCSP to platelets. The binding specificity of Sulph I has been well characterized.16 Using flow cytometric analysis, we observed significant binding of Sulph I and MCSP to unactivated platelets that further increased, by 70% to 80%, on activation of platelets with TRAP (Figure 3). Under similar conditions, we did not detect significant binding of the anti-P-selectin glycoprotein ligand-1 (PSGL-1) antibody PL-1 to platelets, suggesting that little, if any, PSGL-1 is present on platelets.
Platelet Adhesion to P-Selectin
Platelets adhered significantly to surfaces coated with human P-selectin (Figure 4a). This adhesion increased by >100% after platelet activation (Figure 4a). Sulfatide micelles inhibited the platelet adhesion to P-selectin in a concentration-dependent saturable manner with an IC50 of 6.9 μmg/mL (Figure 4b). Sulfatide-binding MCSP also inhibited this interaction by 85%, whereas annexin V had no significant effect (Figure 4c). Furthermore, high-molecular-weight dextran sulfate inhibited platelet adhesion to P-selectin by ≈70% (Figure 4c). These results indicate that platelet sulfatides mediate platelet adhesion to P-selectin.
The monoclonal anti-P-selectin antibody CLB-thromb/6 inhibited the platelet adhesion to P-selectin by ≈75%, whereas an isotype-matched control antibody had no effect (Figure 4c). The monoclonal anti-human PSGL-1 antibody PL-1, which blocks PSGL-1 binding to P-selectin,17 or anti-human GP Ib antibody WM23, which inhibits the binding of P-selectin expressing CHO cells to GP Ib,12 had no effect on platelet adhesion to P-selectin (Figure 4c). These results indicate that platelet adhesion to P-selectin is not mediated via known P-selectin ligands, such as PSGL-1 or GP Ib. Abciximab also had no effect on the interaction of platelets with P-selectin (Figure 4c), suggesting that platelet-P-selectin interactions are not GP IIb/IIIa-mediated.
Reversal of Platelet Aggregation by Sulfatide Antagonist and Anti-P-Selectin Antibodies
The effects of anti-P-selectin antibodies and sulfatide antagonist on platelet aggregation induced by ADP or TRAP were examined. None of these agents had a significant effect on the initial slope of platelet aggregation in human platelet-rich plasma (PRP) after activation with ADP (Figure 5a) or TRAP (Figure 5b). All of these agents, however, reversed platelet aggregation significantly, even though the agonists were used at concentrations sufficient to induce irreversible platelet aggregation. After activation with ADP, both the anti-P-selectin antibodies G1 and CLB-thromb/6 and the sulfatide antagonist MCSP inhibited platelet aggregation by 75% to 85%, whereas recombinant annexin V had no significant effect (Figure 5a). The inhibitory effect was even more pronounced for platelets activated with TRAP, for which a 90% to 100% inhibition of platelet aggregation was observed with G1, CLB-thromb/6, or MCSP (Figure 5b). Similar results were obtained with platelets activated with collagen (data not shown).
Neither anti-P-selectin antibodies (CLB-thromb/6 and G1) nor the recombinant MCSP had any effect on binding of monoclonal antibody PAC-1 to platelets (data not shown). PAC-1 recognizes an activation-dependent epitope at the fibrinogen-binding site of GP IIb/IIIa.18 These results suggest that neither anti-P-selectin antibodies nor sulfatide antagonists interfere with fibrinogen binding to platelets. The functional monoclonal antibodies PL-1, directed against PSGL-1, and WM23, directed against GP Ib, had no significant effect on platelet aggregation induced by either ADP or TRAP (Figure 5a and 5b). These results leave sulfatides as a major candidate ligand for P-selectin in platelet aggregation. The possibility that platelet-leukocyte interactions were responsible for the inhibitory effect of anti-P-selectin antibodies and sulfatide antagonist is unlikely, because the leukocyte number was <0.1% of the platelet number in PRP, as determined by flow cytometric analysis with an anti-CD45 antibody. Furthermore, the agonists used (ADP and TRAP) would not have stimulated leukocytes, thereby inducing the release of leukocyte metabolites that influence platelet-platelet interactions.
In this study, we show that sulfatides are expressed on the platelet surface and serve as a major ligand for P-selectin. Platelets bind to sulfatides, and anti-P-selectin antibodies inhibit this interaction. In parallel, platelets bind to P-selectin, and both sulfatide micelles and sulfatide-binding MCSP inhibit this interaction. Furthermore, both anti-P-selectin antibodies and the sulfatide antagonist MCSP significantly reverse platelet aggregation induced by ADP or TRAP. On the basis of these results, we propose that sulfatides exposed on the platelet surface bind to P-selectin on adjacent platelets, reinforcing interactions between already fibrinogen-bridged platelets, thereby allowing the formation of large, stable platelet aggregates. P-selectin, the largest protein of the C-type lectin family, protrudes out of the membrane because of its increased length.19 As a result, the amino-terminal lectin domain may be more likely to interact with sulfatides on adjacent platelets in an aggregate. Even so, the initial bridging of platelets by GP IIb/IIIa-fibrinogen interactions might be necessary to approximate platelet sulfatides with P-selectin, because Glanzmann’s thrombasthenic platelets do not aggregate despite the presence of P-selectin.7 In addition to their role in platelet aggregation, we show a role for sulfatides in platelet adhesion. Thus, unstimulated platelets expressing sulfatides could adhere to activated endothelial cells expressing P-selectin, thereby providing a nidus for further platelet adhesion and activation.
Two membrane glycoproteins, PSGL-1 and GP Ib, have been identified as P-selectin ligands. PSGL-1 is expressed on neutrophils and monocytes,17 and several posttranslational modifications of PSGL-1, including tyrosine sulfation near the amino terminus, are necessary for its interaction with P-selectin.20 The platelet GP Ib complex, which mediates platelet adhesion to subendothelium at sites of injury, provides a highly acidic region with 3 sulfated tyrosine residues at its amino terminus for the interaction with P-selectin.12 Thus, sulfate moieties are a consistent component of ligand recognition sites for P-selectin, and it may be that the recognition sites in PSGL-1 or GP Ib are structurally related to sulfatides. This concept is supported by the observation that PSGL-1-bearing myeloid cells and sulfatides have overlapping binding sites on P-selectin.21 Recently, the crystal structures of P-selectin containing the lectin and EGF domains cocomplexed with the N-terminal domain of human PSGL-1 have been solved.22 The amino acid residues Arg85 and His114 of P-selectin make direct critical contacts to the sulfates on Tyr7 and Tyr10 of PSGL-1,22 providing further evidence for sulfate moieties as a component of P-selectin binding sites.
Although PSGL-1 and GP Ib are P-selectin ligands, functional monoclonal antibodies to PSGL-1 and GP Ib had no effect on platelet adhesion to P-selectin or platelet aggregation (Figures 4 and 5⇑). In contrast, sulfatide antagonist inhibited both platelet adhesion to P-selectin and platelet aggregation (Figures 4 and 5⇑). These results suggest that sulfatides, and not PSGL-1 or GP Ib, are major ligands for P-selectin in platelet adhesion and aggregation.
Polyclonal antibody F(ab′)2 fragments, which completely inhibited vWF binding to sulfatides, only partially inhibited platelet adhesion to sulfatides.23 Moreover, preincubation of sulfatides with vWF did not increase platelet adhesion to sulfatides.23 This suggests that vWF is not a major factor mediating platelet adhesion to sulfatides. Thrombospondin, a granule protein secreted by platelets after activation, also binds to sulfatides1 and to the surface of activated platelets.24 A monoclonal antibody that strongly inhibits the interaction of thrombospondin with sulfatides did not inhibit platelet aggregation,1 however, suggesting that thrombospondin interaction with sulfatides was not involved in platelet aggregation. We found that F(ab′)2 fragments of polyclonal anti-vWF or anti-thrombospondin antibodies had no significant effect on platelet adhesion to sulfatides, whereas monoclonal anti-P-selectin antibodies inhibited this interaction and platelet aggregation (Figures 1 and 5⇑). Furthermore, the anti-GP IIb/IIIa antibody abciximab had no effect on platelet adhesion to sulfatides, suggesting that GP IIb/IIIa does not play a role in platelet interactions with sulfatides, in agreement with a previous report.23
Sulfatides and their analogues have been used in vivo to block selectin functions. In animal models, these compounds inhibit P-selectin-dependent organ injuries.25 Indeed, inhibition of P-selectin function also had antithrombotic effects in many studies. P-selectin inhibition accelerated thrombolysis in a primate model of arterial thrombosis,25 reduced the extent of venous thrombosis in another primate model,26 and reduced recurrent coronary arterial thrombosis in dogs.27 Although attributed to the inhibition of platelet-leukocyte interactions, a significant component of the observed beneficial effects of P-selectin blockade may instead be due to the reversal of platelet aggregation.
In summary, we show an important role for sulfatide-P-selectin interactions in platelet adhesion and aggregation. Sulfatide-P-selectin interactions stabilize platelet aggregates, representing a new mechanism of platelet aggregation. This mechanism may play a significant role in arterial thrombosis, where platelet thrombi are formed. Consequently, therapeutic interventions directed against sulfatides or P-selectin may be beneficial in the treatment of arterial thrombosis.
This work was supported by NIH grant HL-65096, a Grant-in-Aid from the American Heart Association (9750553N)- and a grant from the Charles Slaughter Foundation.
Guest editor for this article was Ralph L. Nachman, MD, Cornell University, NY.
Bosio A, Binczek E, Stoffel W. Functional breakdown of the lipid bilayer of the myelin membrane in central and peripheral nervous system by disrupted galactocerebroside synthesis. Proc Natl Acad Sci U S A. 1996; 93: 13280–13285.
Huesca M, Goodwin A, Bhagwansingh A, et al. Characterization of an acidic-pH-inducible stress protein (hsp70), a putative sulfatide binding adhesin, from Helicobacter pylori. Infect Immun. 1998; 66: 4061–4067.
Pancake SJ, Holt GD, Mellouk S, et al. Malaria sporozoites and circumsporozoite proteins bind specifically to sulfated glycoconjugates. J Cell Biol. 1992; 117: 1351–1357.
Hsu-Lin S, Berman CL, Furie BC, et al. A platelet membrane protein expressed during platelet activation and secretion: studies using a monoclonal antibody specific for thrombin-activated platelets. J Biol Chem. 1984; 259: 9121–9126.
Merten M, Thiagarajan P. P-selectin expression on platelets determines size and stability of platelet aggregates. Circulation. 2000; 102: 1931–1936.
Merten M, Pakala R, Thiagarajan P, et al. Platelet microparticles promote platelet interaction with subendothelial matrix in a glycoprotein IIb/IIIa-dependent mechanism. Circulation. 1999; 99: 2577–2582.
Romo GM, Dong JF, Schade AJ, et al. The glycoprotein Ib-IX-V complex is a platelet counterreceptor for P-selectin. J Exp Med. 1999; 190: 803–814.
Kean EL. Rapid, sensitive spectrophotometric method for quantitative determination of sulfatides. J Lipid Res. 1968; 9: 319–327.
Ruchaud-Sparagano MH, Malaud E, Gayet O, et al. Mapping the epitope of a functional P-selectin monoclonal antibody (LYP20) to a short complement-like repeat (SCR 4) domain: use of human-mouse chimaera and homologue-replacement mutagenesis. Biochem J. 1998; 332: 309–314.
Fredman P, Mattsson L, Andersson K, et al. Characterization of the binding epitope of a monoclonal antibody to sulphatide. Biochem J. 1988; 251: 17–22.
Moore KL, Patel KD, Bruehl RE, et al. P-selectin glycoprotein ligand-1 mediates rolling of human neutrophils on P-selectin. J Cell Biol. 1995; 128: 661–671.
Taub R, Gould RJ, Garsky VM, et al. A monoclonal antibody against the platelet fibrinogen receptor contains a sequence that mimics a receptor recognition domain in fibrinogen. J Biol Chem. 1989; 264: 259–265.
Ushiyama S, Laue TM, Moore KL, et al. Structural and functional characterization of monomeric soluble P-selectin and comparison with membrane P-selectin. J Biol Chem. 1993; 268: 15229–15237.
Toombs CF, DeGraaf GL, Martin JP, et al. Pretreatment with a blocking monoclonal antibody to P-selectin accelerates pharmacological thrombolysis in a primate model of arterial thrombosis. J Pharmacol Exp Ther. 1995; 275: 941–949.
Ikeda H, Ueyama T, Murohara T, et al. Adhesive interaction between P-selectin and sialyl Lewisxplays an important role in recurrent coronary arterial thrombosis in dogs. Arterioscler Thromb Vasc Biol. 1999; 19: 1083–1090.