CD226 mediates platelet and megakaryocytic cell adhesion to vascular endothelial cells.

Platelet adhesion to vascular endothelial cells is a pathophysiologically relevant cell-to-cell interaction. However, the mechanisms underlying this cellular interaction are incompletely understood. In search of the ligand for CD226 adhesion molecule expressed on platelets, we found that human umbilical vein endothelial cells (HUVEC) express significant amount of putative CD226 ligand. We demonstrated that thrombin-activated, but not resting, platelets bind to intact HUVEC. Anti-CD226 monoclonal antibody specifically inhibited the binding, indicating that CD226 mediates the intercellular binding between thrombin-activated platelets and HUVEC. We also demonstrated that platelet activation with thrombin induces tyrosine phosphorylation of CD226 as well as CD226-mediated platelet adhesion. Moreover, experiments using mutant transfectants suggested that the tyrosine at residue 322 of CD226 plays an important role for its adhesive function. CD226 was also expressed on primary megakaryocytes and megakaryocytic cell lines. Anti-CD226 monoclonal antibody inhibited binding of megakaryocytic cell lines to HUVEC. Taken together, these results reveal a novel mechanism for adhesion of platelets and megakaryocytic cells to vascular endothelial cells.

Cellular interaction between platelets and vascular endothelial cells plays a crucial role for maintaining the hemostatic system. Platelets roll on activated or inflated endothelial cells under high shear rate, which is mediated by P-selectin expressed on endothelial cells (1)(2)(3). Furthermore, once the endothelial cells are denuded, GPIb (CD42b) on platelet surface binds to immobilized von Willebrand factor (vWF) 1 at exposed subendothelium (4,5). Several other molecules, such as GPIIbIIIa (6,7), platelet-endothelial cell adhesion molecule-1 (8), fibrinogen (9), and ␤ 1 -integrin (10), have also been reported to be involved in this process, depending on both shear rate and activation status of platelets and endothelial cells. Recently, Harlan and colleagues (7) proposed a model for the interaction between activated platelets and intact endothelial cells, in which platelet-bound adhesive proteins such as fibrinogen, fibronectin, and vWF mediate bridging between human umbilical vein endothelial cells (HUVEC) and GPIIbIIIa (CD41/ CD61) expressed on platelets. However, in their experimental model, adhesion of activated platelets to intact HUVEC was only partially inhibited in the presence of Arg-Gly-Asp-Ser (RGDS) peptide, which blocks ligand binding to GPIIbIIIa (7), suggesting that there might be as yet undetermined adhesion molecules contributing to this interaction.
The hemostatic system is also maintained by the constant production of platelets from megakaryocytes. Platelets are assumed to be cell fragments torn from cytoplasm of mature megakaryocytes that reside in close proximity to bone marrow endothelial cells (11)(12)(13). Several lines of evidence suggest that cellular interaction between megakaryocytes and vascular endothelial cells plays a crucial role for megakaryopoiesis (14 -18). However, the molecular mechanism underlying this cellular interaction has not been well elucidated before now.
We previously identified a novel adhesion molecule DNAM-1 (CD226), which is a member of the immunoglobulin superfamily containing two Ig-like domains of the V-set and is encoded by a gene on human chromosome 18q22.3 (19). CD226 is a ϳ65-kDa glycoprotein expressed on the majority of NK cells and monocytes and a subset of T-lymphocytes, and involved in cytotoxicity and cytokine secretion mediated by NK cells and T-lymphocytes. It is intriguing that CD226 is also expressed on platelets (20). Cross-linking CD226 with an anti-CD226 monoclonal antibody induces platelet activation and aggregation (21), 2 suggesting that CD226 functions as a signal transducing adhesion molecule in platelets. We demonstrate here that CD226 is involved in adhesion of platelets and megakaryocytic cells to vascular endothelial cells.
Cell Lines and Cell Preparation-HEL, BW5147, COS-7, and 293T * This work was supported in part by Ministry of Education, Culture, Sports, Science, and Technology of Japan Grant 12670972 (to T. N.) and Grants 12051201 and 12470111 (to A. S.) and by special coordination funds from the Science and Technology Agency of the Japanese Government (to A. S.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 (22) and Dr. Takeyuki Sato (Chiba University, Chiba, Japan) (23), respectively. NH is an erythroleukemic cell line established in our laboratory. Plateletrich plasma (PRP) and gel-filtrated platelets were prepared as described elsewhere (24,25).
HUVEC were prepared by the standard procedure. Briefly, inner lumen of the umbilical vein was gently washed with sterile PBS and filled with trypsin-EDTA solution (Sigma). After a 30-min incubation at 37°C, detached cells were harvested and washed once with RPMI 1640 containing 10% fetal calf serum. Cells were resuspended in RPMI 1640 containing 15% horse serum, 6 units/ml heparin, 2 mM L-glutamine, 1 mM sodium pyruvate, 1 mM Hepes, and 1ϫ endothelial cell growth factor (Sigma), transferred to gelatin (Sigma)-coated flasks, and then cultured at 37°C in 5% CO 2 .
Establishment of BW5147 Transfectants Expressing Mutant CD226 -BW5147 transfectants expressing wild type and a site-specific CD226 mutant at serine residue 329 were described previously (26). To generate site-specific CD226 mutants at tyrosine residues 322, 325, or both, antisense PCR primers, which contained a codon for Phe 322 (TTT), Phe 325 (TTT), or both, were designed. The PCR products were subcloned into a retroviral vector pMX-neo (provided by Dr. Toshio Kitamura, DNAX Research Institute) with cloning sites of BamHI (5Ј) and NotI (3Ј). BOSC23 packaging cells were transfected with the CD226 cDNA in the retroviral vector using LipofectAMINE (Invitrogen) as described previously (27,28). BW5147 cells were infected with the CD226 retrovirus stock. Two days after infection, the cells expressing CD226 were cloned by flow cytometry. All mutant cDNAs were verified by sequencing.
Expression of the CD226-Ig Fusion Protein (FP) and Flow Cytometry-293T cells were transiently transfected with the plasmid of the CD226-Ig FP, which was constructed in pCDM8 expression vector containing a human CD8 leader segment followed by the Fc segment of human IgG at the COOH terminus (provided by Dr. Gerald Zurawski, DNAX Research Institute), and the CD226-Ig FP was purified by affinity chromatography using immobilized protein A (Bio-Rad). Cells (2 ϫ 10 5 ) were stained for 1 h on ice with 1 g/ml CD226-Ig FP or control FP. FITC-labeled anti-human IgG (Dako, Glostrup, Denmark) was used as a second step reagent. Flow cytometry was performed to detect the binding of CD226-Ig FP to HUVEC.
Adhesion Assay of Calcein-loaded Platelets-Platelet adhesion to HUVEC was determined by the method, as described (7), with modification. Briefly, PRP was treated with 1 mM aspirin for 20 min, and after adding 0.1 volume of ACD solution (2.2% (w/v) sodium citrate, 0.8% (w/v) citric acid, and 2.2% (w/v) dextrose), the PRP was centrifuged at 500 ϫ g for 20 min at room temperature. Platelets were resuspended in 200 l of Hepes-Tyrode buffer (137 mM NaCl, 2 mM KCl, 0.4 mM NaH 2 PO 4 , 1 mM MgCl 2 , 1 mM CaCl 2 , 5.6 mM glucose, 0.1% (w/v) BSA, 20 mM Hepes, pH 7.4) containing 10% ACD, and 0.5 l of 5 mM calcein AM (Molecular Probes, Eugene, OR) was added to the platelet suspension. After a 30-min incubation at room temperature in the dark with gentle shaking, platelets were washed twice with Hepes-Tyrode buffer containing 10% ACD and were resuspended in Hepes-Tyrode buffer (1 ϫ 10 9 /ml). Platelets were stimulated or not with thrombin (1 unit/ml, for 10 min at room temperature) under non-stirring condition, and then hirudin (2 units/ml, Sigma) was added to inactivate thrombin. After additional treatment with F(abЈ) 2 fragments of control Ig (Jackson Co., West Grove, PA) or anti-CD226 mAb alone or in combination with Arg-Gly-Asp-Ser (RGDS) peptide (500 M, Sigma) at room temperature for 10 min, 100 l of calcein-loaded platelets were transferred on HU-VEC prepared in each well of 12-well microtiter plates (Iwaki, Tokyo, Japan) containing 900 l of RPMI 1640 with 1 mM CaCl 2 . Platelets were co-cultured with HUVEC for 1 h at 37°C. After washing out the nonadherent platelets, HUVEC were detached by cell scraper, vigorously pipetted, and washed once. The platelet binding to HUVEC was detected by flow cytometry in triplicate experiments.
Adhesion Assays of Various Cell Lines-Various cell lines were labeled with Na 2 51 CrO 4 (37 MBq/ml) (New England Nucleus, Boston, MA) and resuspended in RPMI 1640 (2 ϫ 10 5 /ml). 51 Cr-Labeled cells were treated with various concentrations of F(abЈ) 2 fragments of control Ig or anti-CD226 mAb at room temperature for 10 min, if required. One hundred microliters of 51 Cr-labeled cells were then transferred into each well of 48-well microtiter plates, in which HUVEC or COS-7 cells had been plated at least 48 h prior to assay. After a 1-h incubation at 37°C, non-adherent cells were removed by washing with RPMI 1640. Residual adherent cells were treated with lysis buffer (1% Nonidet P-40, 150 mM NaCl, 50 mM Tris-HCl, pH 8.0), and the radioactivities in the lysates were determined by a ␥-counter. The experiments were performed in triplicate.
For platelet adhesion to fibrinogen, 100-mm non-treated polystyrene dishes (Iwaki) were coated with 200 g/ml fibrinogen (fraction I; Sigma) for 16 h at room temperature, washed three times with PBS, and blocked with 1 mg/ml BSA for 1 h. Gel-filtrated platelets (3 ml, 4 ϫ 10 8 /ml) pretreated with apyrase (10 units/ml, Sigma) were placed on fibrinogen-coated dishes for 1 h at 37°C. After three washings with PBS, adherent platelets were lysed by 1 ml of RIPA buffer. Nonadherent platelets harvested from BSA-coated dishes were used as a negative control.
Lysates were precleared with Protein G-Sepharose CL-4B (Amersham Biosciences, Uppsala, Sweden) in RIPA buffer at 4°C for 1 h and reacted with anti-CD226 mAb at 4°C overnight. Immunocomplexes were precipitated with Protein G-Sepharose CL-4B, washed three times with RIPA buffer. Proteins were separated by SDS-PAGE and transferred onto a nitrocellulose membrane (Hybound TM ECL TM ; Amersham Biosciences). After overnight incubation in Tris-buffered saline (25 mM Tris-HCl, 140 mM NaCl, pH 7.4) containing 0.5% Tween 20 and 5% BSA, membranes were reacted with a specific antibody. Proteins were detected by using horseradish peroxidase-conjugated goat anti-mouse Ig (Amersham Biosciences) and developed by ECL-immunoblotting detection kit (Amersham Biosciences). Chemiluminescence was detected by autoradiography.
Immunofluorescent Microscopy-Bone marrow smears prepared from healthy volunteers were incubated with 5% horse serum in PBS for 30 min at room temperature and then stained with anti-CD226 mAb for 2 h at room temperature, washed with PBS, followed by staining with 1:2000 diluted Alexa 488-conjugated horse anti-mouse IgG (Molecular Probes) for 1 h at room temperature. After washings with PBS, slides were fixed with Vectashield mounting medium containing 4Ј,6diamidino-2-phenylindole (Vector Laboratories, Burlingame, CA), and observed under fluorescent microscopy.

HUVEC Express the Putative Ligand for CD226 (CD226L)-
To determine the role of CD226 in platelet adhesion to vascular endothelial cells, we first examined whether vascular endothelial cells express CD226L. We generated a soluble protein consisting of the extracellular domain of CD226 fused to the Fc portion of human IgG. The fusion protein of CD226 specifically bound to HUVEC, indicating that HUVEC express the putative CD226L (Fig. 1A). To examine whether the CD226L expressed on HUVEC mediate intercellular binding, we co-cultured 51 Crlabeled BW5147 transfectants expressing CD226 with HUVEC monolayers and determined specific intercellular binding. As shown in Fig. 1B, the BW5147 transfectants expressing CD226 bound HUVEC at significantly higher levels than the parental BW5147 cells. Moreover, the binding of the transfectants to HUVEC was specifically inhibited in the presence of F(abЈ) 2 fragments of anti-CD226 mAb. These results indicate that HU-VEC express the functional CD226L that mediates intercellular binding.
CD226 Mediates Thrombin-activated Platelet Binding to HU-VEC-To explore the involvement of CD226 in platelet binding to vascular endothelial cells, HUVEC monolayers were incubated with calcein-loaded platelets. HUVEC were then harvested, and platelet binding was determined by flow cytometry. Whereas the binding of resting platelets to HUVEC was hardly detected (data not shown), thrombin-activated platelets effectively bound to HUVEC (Fig. 2). Pretreatment of thrombinactivated platelets with RGDS peptide, which blocks ligand binding to GPIIbIIIa, inhibited the platelet binding to HUVEC, consistent with a previous report (7). However, we observed that blockade of CD226 by F(abЈ) 2 fragments of anti-CD226 mAb alone also significantly inhibited the thrombin-activated platelet binding to HUVEC induced by thrombin. Combination of F(abЈ) 2 fragments of anti-CD226 mAb and RGDS peptide further inhibited platelet binding to HUVEC. These results indicate that CD226 is involved in thrombin-activated platelet adhesion to HUVEC.
Thrombin Induces Tyrosine Phosphorylation of CD226 in Platelets-Because platelet activation with thrombin induces CD226-mediated adhesion of platelets to HUVEC, we considered that thrombin may either up-regulate CD226 expression or modulate CD226 function on platelets. As the amount of platelet-surface CD226 did not change, as determined by flow cytometry, after stimulation with thrombin (data not shown), we supposed that platelet activation with thrombin modulates the avidity or affinity of CD226 molecule. We therefore investigated whether platelet activation by natural agonist such as thrombin or collagen induces signaling events in CD226. Because CD226 cytoplasmic portion contains the tyrosine at residue 322, which can be phosphorylated by the Fyn tyrosine kinase (29), we examined whether thrombin or collagen induces tyrosine phosphorylation of CD226. As demonstrated in Fig. 3 (A and B), stimulation of platelets with collagen or thrombin under stirring condition induced tyrosine phosphorylation of CD226. In a sharp contrast to the rapid phosphorylation by collagen that peaked at 1 min, tyrosine phosphorylation of CD226 by thrombin was achieved at the maximum level after 10 min. Judging from the slower time course of the phosphorylation in thrombin-stimulated platelets, we assumed that the phosphorylation is mediated mainly by fibrinogen binding to GPIIbIIIa and/or the resultant platelet aggregation. To explore this possibility, we examined whether thrombin-induced tyrosine phosphorylation of CD226 is observed under non-stirring condition, in which fibrinogen binding to GPIIbIIIa is reduced. As demonstrated in Fig. 3C, although platelet stimulation with thrombin under non-stirring condition induced significant level of tyrosine phosphorylation of CD226, the phosphorylation level was considerably lower than that induced under stirring condition. Moreover, addition of RGDS peptide, which blocks fibrinogen binding to GPIIbIIIa, significantly inhibited thrombin-induced tyrosine phosphorylation of CD226. These results suggest that fibrinogen binding to GPIIbIIIa is critical for achieving the maximal level of tyrosine phosphorylation of CD226 in thrombin-activated platelets. In fact, platelet adhesion to immobilized fibrinogen via GPIIbIIIa resulted in significant levels of tyrosine phosphorylation of CD226 (Fig. 3D).
The Tyrosine (Tyr 322 ) Is Involved in CD226-mediated Adhesion-We previously reported that phosphorylation of the serine at residue 329 by PKC plays an important role for CD226mediated adhesion (26). However, we also observed that the site-specific mutant CD226 at residue 329 (Ser 329 3 Phe) is still capable of binding to the putative CD226L (data not shown and Ref. 26), suggesting that CD226-mediated adhesion depends in part on undetermined molecular events other than the phosphorylation of Ser 329 . Here we have demonstrated that thrombin induces both CD226-mediated platelet adhesion to HUVEC and tyrosine phosphorylation of CD226. These results suggest that tyrosine phosphorylation of CD226 may also be involved in CD226-mediated adhesion. To examine this possibility, we established BW5147 transfectants expressing either wild type or site-directed mutant CD226 at residues 322, 325, or both (Tyr 322 3 Phe, Tyr 325 3 Phe, or Tyr 322 3 Phe/Tyr 325 3 Phe, respectively). As shown in Fig. 4B, the introduction of the site-directed mutations at residues 322 (Tyr 322 3 Phe) and both 322 and 325 (Tyr 322 3 Phe/Tyr 325 3 Phe), but not at 325 alone (Tyr 325 3 Phe), completely abolished tyrosine phospho-rylation of CD226 after stimulation with pervanadate, suggesting that Tyr 322 as well as Ser 329 may play an important role for CD226-mediated adhesion. For binding assay, we used COS-7 cells, instead of HUVEC, which also express the putative CD226L, as confirmed by flow cytometry using the CD226-Ig fusion protein (data not shown), because COS-7 cells are hardly detached from plates by vigorous washing to extremely reduce nonspecific binding of 51 Cr-labeled BW5147 transfectants. Whereas these transfectants express comparable amount of CD226 (Fig. 4A), the transfectants expressing mutated CD226 at residue(s) 322 (Tyr 322 3 Phe) or both 322 and 325 (Tyr 322 3 Phe/Tyr 325 3 Phe), but not 325 alone (Tyr 325 3 Phe), bound to COS-7 cells at a significantly lower level, as compared with that expressing wild type CD226 (Fig. 4C). Taken together, these results suggest that tyrosine phosphorylation of CD226   FIG. 3. Natural agonists induce tyrosine phosphorylation of CD226 in platelets. A and B, gel-filtrated platelets were stimulated either with 20 g/ml collagen (A) or with 1 unit/ml thrombin (B) under stirring condition for indicated duration. C, gel-filtrated platelets treated or not with RGDS peptide (500 M) at 37°C for 5 min were stimulated for 5 min with thrombin (1 unit/ml) or control solvent under stirring (stir.) or non-stirring (non-stir.) condition as denoted. D, gel-filtrated platelets pretreated with apyrase (10 units/ml, for 10 min) were placed on fibrinogencoated dishes for 1 h at 37°C. Adherent platelets were then lysed, and CD226 was immunoprecipitated (IP) from the lysates. Non-adherent platelets harvested from BSA-coated dishes were used as a negative control. The precipitates were analyzed by immunoblotting (IB) with anti-phosphotyrosine or anti-CD226 mAbs. The data are representative of several independent experiments.

FIG. 4. Tyr 322 is involved in CD226-mediated adhesion.
A, BW5147 cells transfected with mock control vector or wild type (WT) or site-directed mutant CD226 were stained with anti-CD226 mAb, followed by an FITC-conjugated anti-mouse Ig. The expression of CD226 molecule in each transfectants was analyzed by flow cytometry. B, BW5147 transfectants were stimulated with pervanadate for 10 min at 37°C and lysed. Immunoprecipitation (IP) and immunoblotting (IB) were performed, as indicated. C, 51 Cr-labeled BW5147 cells transfected with mock control vector or wild type or mutant CD226 at tyrosine residue 322 or 325, or both 322 and 325 (Y-F 322 , Y-F 325 , or Y-F 322ϩ325 ) were incubated on COS-7 cells prepared in 48-well microtiter plates at 37°C for 1 h. The adherent BW5147 cells were lysed, and the radioactivities in the lysates were assayed by a ␥-counter. Total radioactivities after 51 Cr labeling were confirmed to be statistically identical between each of the transfectants. Results are presented as the mean radioactivities in triplicate experiments, after subtraction of background counts/min (Ͻ3500 cpm) generated by binding of the transfectant with mock control vector from counts/min generated by transfectants with CD226. Error bars denote standard deviations (*, p Ͻ 0.05).
at residue 322 plays an important role for CD226-mediated adhesion of the transfectant to COS-7.
CD226 Mediates Binding of Megakaryocytic Cells to HU-VEC-We examined the expression of CD226 on megakaryocytes or megakaryocytic cell lines. Fig. 5A shows that human primary megakaryocytes express significant amount of CD226, as determined by immunohistochemistry. Analyses by flow cytometry and immunoblotting showed that CD226 is also expressed on various megakaryocytic cell lines, including UT7/ TPO, CMK, and HEL, but not on an erythroleukemic cell line, NH (Fig. 5, B and C). Therefore, experiments were undertaken to examine whether CD226 is involved in interaction between megakaryocytes and vascular endothelial cells. HUVEC monolayers were incubated with 51 Cr-labeled HEL cells in the presence or absence of anti-CD226 mAb, and specific binding of HEL cells to HUVEC was determined. As shown in Fig. 6, HEL cell binding to HUVEC was significantly inhibited in the presence of F(abЈ) 2 fragments of anti-CD226 mAb, indicating that CD226 participates in HEL cell binding to HUVEC. Similar results were observed when UT7/TPO or CMK cells were cul-tured with HUVEC (data not shown). These results indicate that CD226 is involved in interaction between megakaryocytic cells and vascular endothelial cells. DISCUSSION In the present study we have found that the putative ligand for CD226 is expressed on HUVEC (Fig. 1). This observation led us to examine whether CD226 mediates intercellular binding between platelets and HUVEC. We have demonstrated that thrombin-activated, but not resting, platelets bind to HUVEC, and CD226 is in part responsible for this binding (Fig. 2). We have also shown that platelet activation with thrombin induces tyrosine phosphorylation of CD226 (Fig. 3B) as well as CD226mediated adhesion. Significance of tyrosine phosphorylation of CD226 for its adhesive function was demonstrated at least in CD226-expressing BW5147 transfectants (Fig. 4). Taken together with these and our previous observations (26,29), it may be feasible to consider that thrombin modulates CD226 adhesive function possibly via CD226 tyrosine phosphorylation, which enables CD226-mediated platelet adhesion to HUVEC. Although we have demonstrated that the tyrosine phosphorylation at residue 322 of CD226 is in part responsible for CD266-mediated adhesion in BW5147 transfectant system (Fig. 4), it remains undetermined whether the tyrosine phosphorylation is involved also in thrombin-induced platelet adhesion to HUVEC. Indeed, thrombin gave rise to CD226-mediated platelet adhesion to HUVEC under non-stirring condition (Fig.  2), in which CD226 was only weakly tyrosine-phosphorylated (Fig. 3C). One possible explanation is that, besides tyrosine phosphorylation of CD226, other signaling events induced by thrombin co-operatively modulate CD226 adhesive function. In fact, in our previous experiment, we demonstrated that serine phosphorylation of CD226 is critical for its adhesive function in T-lymphocytes (26). Alternatively, a small amount of tyrosine phosphorylation of CD226 may be sufficient to exert its adhesive function.
Fibrinogen binding to GPIIbIIIa, which should be induced by platelet activation by thrombin, seems to play an important role for thrombin-induced tyrosine phosphorylation of CD226 (Fig. 3, C and D). We previously demonstrated that crosslinking CD226 on NK cells derived from a patient with leukocyte adhesion molecule deficiency, whose leukocytes lack LFA-1 expression, fails to induce NK cell activation (29). Furthermore, CD226 is associated with LFA-1 in NK cells and activated T-lymphocytes (29), indicating physical and functional relationship between CD226 and LFA-1. Presently, the signaling pathway leading to CD226 tyrosine phosphorylation in thrombin-activated platelets is uncertain. However, the requirement of ligand binding to GPIIbIIIa for the tyrosine phosphorylation in thrombin-activated platelets may indicate functional association between GPIIbIIIa and CD226, corresponding with the association of LFA-1 and CD226 in lymphocytes. On the other hand, the rapid time course of the tyrosine phosphorylation in collagen-stimulated platelets suggests a possibility that an Src family tyrosine kinase Fyn, which is rapidly activated by collagen in platelets (30), may phosphorylate the tyrosine at residue 322 of CD226, as was demonstrated in lymphocytes (29).
Cellular interaction between megakaryocytes and vascular endothelial cells is crucial not only for maturation and differentiation of megakaryocytes (14 -18) but also for migration into the inner lumen of vascular endothelial cells, whereby platelets are generated (31). We have described here that human primary megakaryocytes and several megakaryocytic cell lines express significant amount of CD226, which is involved in intercellular binding between megakaryocytic cell lines and HUVEC. Because cross-linking CD226 induces platelet activation and aggregation (20), 2 CD226 may mediate signals for cytoplasmic processing in megakaryocytes, possibly contributing to megakaryocyte maturation and/or platelet production.
In summary, we have revealed a novel mechanism for adhesion of platelets and megakaryocytic cells to vascular endothelial cells. The present studies suggest that CD226 may play an important role for thrombosis and hemostasis. Further studies are under way to identify the CD226L and determine the structural requirements involved in CD226 binding. Activation mechanisms of CD226 during platelet activation by natural agonists should also be clarified in future experiments.