Cleavage of platelet endothelial cell adhesion molecule-1 (PECAM-1) in platelets exposed to high shear stress

Authors

  • Y. Naganuma,

    1. Department of Laboratory Medicine, Faculty of Medicine, University of Yamanashi, Yamanashi, Japan and Department of Laboratory Medicine, Graduate School of Medicine, University of Tokyo, Japan
    Search for more papers by this author
  • K. Satoh,

    1. Department of Laboratory Medicine, Faculty of Medicine, University of Yamanashi, Yamanashi, Japan and Department of Laboratory Medicine, Graduate School of Medicine, University of Tokyo, Japan
    Search for more papers by this author
  • Q. Yi,

    1. Department of Laboratory Medicine, Faculty of Medicine, University of Yamanashi, Yamanashi, Japan and Department of Laboratory Medicine, Graduate School of Medicine, University of Tokyo, Japan
    Search for more papers by this author
  • N. Asazuma,

    1. Department of Laboratory Medicine, Faculty of Medicine, University of Yamanashi, Yamanashi, Japan and Department of Laboratory Medicine, Graduate School of Medicine, University of Tokyo, Japan
    Search for more papers by this author
  • Y. Yatomi,

    1. Department of Laboratory Medicine, Faculty of Medicine, University of Yamanashi, Yamanashi, Japan and Department of Laboratory Medicine, Graduate School of Medicine, University of Tokyo, Japan
    Search for more papers by this author
  • Y. Ozaki

    1. Department of Laboratory Medicine, Faculty of Medicine, University of Yamanashi, Yamanashi, Japan and Department of Laboratory Medicine, Graduate School of Medicine, University of Tokyo, Japan
    Search for more papers by this author

Yukio Ozaki, Department of Laboratory Medicine, Faculty of Medicine, University of Yamanashi, Shimokato 1110, Tamaho, Nakakoma, Yamanashi, 409–3898, Japan.
Tel.: +81 55 2739884; fax: +81 55 2736713; e-mail: yozaki@yamanashi.ac.jp

Abstract

Platelet endothelial cell adhesion molecule-1 (PECAM-1, CD31) is a 130 kDa transmembrane glycoprotein that belongs to the immunoglobulin superfamily and is expressed on the surface of endothelial cells, platelets, and other blood cells. Although the importance of this adhesion molecule in various cell–cell interactions is established, its functional role in platelets remains to be elucidated. In this study, we examined whether PECAM-1 underwent changes in platelets exposed to high shear stress. Platelet PECAM-1 was cleaved under high shear stress and was released into the extracellular fluid as a fragment with an approximate molecular weight of 118 kDa. The cleavage was inhibited by an anti-VWF MoAb, but not by recombinant VWF A1 domains. These findings suggest that the GPIb–VWF interaction is involved in PECAM-1 cleavage under high shear stress, and that the cleavage is independent of GPIb clustering by VWF multimers. Furthermore, EGTA or calpeptin inhibited PECAM-1 cleavage. This finding provides evidence for the involvement of calpain in PECAM-1 cleavage. Flow-cytometric analysis revealed that PECAM-1 expression on the platelet surface was decreased under high shear stress. This reduction occurred exclusively in a specific population of platelets, which corresponded to platelet-derived microparticles (PMP). In conclusion, PECAM-1 cleavage under high shear stress is closely related to the activation of calpain and the process of PMP formation mediated by the GPIb–VWF interaction.

Introduction

Platelet endothelial cell molecule-1 (PECAM-1, CD31), a 130 kDa glycoprotein of the immunoglobulin (Ig) superfamily, is expressed on the surface of platelets, endothelial cells, and some subsets of leukocytes [1]. PECAM-1 is composed of six Ig-like homology units, a transmembrane domain, and a cytoplasmic domain. In addition to its adhesive function, PECAM-1 has been shown to participate in signaling pathways [2]. The PECAM-1-mediated signaling is thought to occur, in part, via the immunoreceptor tyrosine-based inhibitory motif (ITIM) in its cytoplasmic domain. When the tyrosine residues in the PECAM-1 ITIM are phosphorylated, this motif mediates the binding of signaling molecules and adapter molecules having tandem SH2 domains.

PECAM-1 has multifunctional roles, including integrin regulation [3,4], migration of endothelial cells and neutrophils [5,6], and angiogenesis [7]. Recently, it has been reported that PECAM-1 is tyrosine-phosphorylated when endothelial cells are exposed to physiologic levels of shear stress [8]. The PECAM-1 tyrosine phosphorylation initiates a signaling cascade leading to extracellular signal-regulated kinase (ERK) activation in endothelial cells. These reports suggest that PECAM-1 has a possible role of mechanosensing in endothelial cells.

On the other hand, the role of PECAM-1 in platelets still remains elusive. It has been reported that activation of glycoprotein (GP) VI, one of the major collagen receptors on platelets, leads to tyrosine phosphorylation of PECAM-1 [9]. PECAM-1 –/– platelets in mice show increased adhesion to immobilized collagen and enhanced aggregation response to GPVI-specific agonists [10]. Thus, it is implied that PECAM-1, with its ITIM, negatively regulates the signaling pathway mediated by GPVI, which employs a number of signaling molecules related to tyrosine phosphorylation [11]. However, agonists such as thrombin or ADP, homophilic interaction of PECAM-1 or anti-PECAM-1 antibodies also induce tyrosine phosphorylation of PECAM-1, and its physiologic significance still awaits determination.

Recent evidence has indicated that shear-induced platelet aggregation (SIPA) is an important mechanism of thrombogenesis at sites of arterial bifurcations or stenosis [12–14]. The molecular mechanism of SIPA still remains incompletely understood. It is known that the interaction of fibrinogen with GPIIb/IIIa is required at low shear stress, while the interaction of von Willebrand factor (VWF) with GPIb initiates intracellular signaling under high shear stress [15]. Recent reports suggest that several proteins undergo tyrosine phosphorylation when platelets are exposed to high shear stress, and that the GPIb-mediated signaling pathway is closely related to tyrosine kinases [16]. As PECAM-1 with its ITIM may regulate the events related to tyrosine phosphorylation, and it has been implicated as a mechanoreceptor for shear stress in other cells, we were prompted to evaluate the involvement of PECAM-1 in platelets exposed to high shear stress.

Platelet activation can also be accompanied by the shedding of membrane microparticles, the presence of which is observed in some thrombotic and inflammatory disorders [17]. Shedding of platelet-derived microparticles (PMP) requires the activation of a calcium-dependent cysteine protease, calpain [18]. Its activity is regulated mainly by cytosolic Ca2+ and an endogenous inhibitor, calpastatin [19]. A number of endogenous calpain substrates have been reported so far: cytoskeletal proteins such as actin binding protein (ABP), talin, spectrin and enzyme proteins such as PKC, myosin light chain kinase (MLCK), and others. Cleavage of these endogenous substrates suggests calpain involvement during platelet activation. In this study, we report on the degradation and shedding of PECAM-1 under high shear stress and the involvement of calpain in this process.

Materials and methods

Antibodies and reagents

Botrocetin and convulxin were kindly provided byDr M. Berndt (Monash University, Clayton, Australia) and by Dr T. Morita (Meiji Pharmaceutical University, Tokyo, Japan), respectively. A rabbit polyclonal antibody (PoAb) against PECAM-1 cytoplasmic domain (BooBoo) [20] was a kind gift of Dr J. Madri (Yale University School of Medicine, CT, USA). Monoclonal antibodies (MoAbs) against PECAM-1 Ig-like domain 6th (P1.2) [21] and against the 15 amino acids of cytoplasmic tail (235.1) [22] were kindly donated by Dr P. Newman (The Blood Center of South-eastern Wisconsin, WI, USA). Anti-calpain PoAb [23] and anti-GPIb MoAb (WGA3) [24] were kindly donated by Dr H. Ishii (Showa Pharmaceutical University, Tokyo, Japan) and Dr M. Handa (Keio University, Tokyo, Japan), respectively. Recombinant VWF A1 domain (rA1 domain) [25] and anti-VWF MoAb Fab fragment, AJvW-2 [26], were supplied by Ajinomoto Co. (Kanagawa, Japan). VWF was purified from Confact F (factor VIII/VWF concentrate), kindly supplied by Kaketsuken (Kumamoto, Japan), with sepharose 4B chromatography as described previously [27].

The following materials were obtained from the indicated suppliers: anti-PECAM-1 MoAb, clone JC/70 A (DAKO A/S, Glostrup, Denmark); FITC- or PE-labeled anti-PECAM-1 MoAbs, FITC-labeled anti-GPIIa MoAb (Immunotech, Marseille, France); antiphosphotyrosine MoAb, PY20 (Transduction Laboratories, Lexington, KY, USA); antiphosphotyrosine MoAb, 4G10, anti-Src MoAb (Upstate Biotechnology, Lake Placid, NY, USA); collagen (Hormon-Chemie, Munich, Germany); thrombin (Green cross, Osaka, Japan); wheat germ agglutinin (WGA), bovine serum albumin (BSA), wortmannin, apyrase, phenylmethylsulfonyl fluoride (PMSF), Na3VO4, Triton X-100, E64 and E64d (Sigma Chemical Co., St. Louis, MO, USA); PP1 (Biomol, Plymouth Meeting, PA, USA); Gly-Arg-Gly-Asp-Ser (GRGDS) peptide (Peptide Institute, Osaka, Japan); protein A-Sepharose 4B and enhanced chemiluminescence (ECL) reaction reagent (Amersham Pharmacia Biotech, Buckinghamshire, UK); peroxidase-conjugated goat anti-mouse IgG PoAb (Cappel, Durham, NC, USA); anti-Cbl PoAb, peroxidase-conjugated goat anti-rabbit IgG PoAb (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA); calpain I and calpeptin (Calbiochem, San Diego, CA, USA); GM6001 (Chemicon international Inc., Temecula, CA, USA); ZVAD-FMK (R & D systems Inc., Minneapolis, MN, USA).

Cell preparation

Human venous blood was obtained from healthy volunteers who denied the use of any drugs for a minimum of 2 weeks before experiments. Platelet-rich plasma (PRP) was prepared as previously described [28], and incubated with 1 mm acetylsalicylic acid for 30 min to exclude any secondary effects of thromboxane A2. Washed platelets were prepared as previously described [28], and were adjusted to 5 × 108 cells mL−1 in a modified HEPES-Tyrode buffer (134 mmol L−1 NaCl, 0.34 mmol L−1 Na2HPO4, 2.9 mmol L−1 KCl, 12 mmol L−1 NaHCO3, 20 mmol L−1 HEPES, 5 mmol L−1 glucose, 1 mmol L−1 MgCl2, pH 7.4) and supplemented with 1 mm CaCl2 and 20 μg mL−1 VWF, unless otherwise stated.

Measurement of shear stress-induced platelet aggregation

Platelets were subjected to shear stress with a cone-plate viscometer as previously described [16]. Briefly, the system consists of a cone-plate streaming chamber, a light source, a photon counting unit and a microcomputer unit. Helium–neon laser light at 633 nm was passed through the streaming samples, and the transmitted light was recorded continuously. The instrument was calibrated with a platelet suspension for 100% OD and with buffer for 0% OD. Four hundred microliters of washed platelets was incubated with 20 μg mL−1 of VWF or 20 μg mL−1 of rA1 domain for 5 min, and was placed on the surface of a polymethylmethacrylate plate and exposed to shear stress at 30°C for the indicated time. Unless otherwise stated, platelets were stimulated with high shear stress, in which the rotation rate of the cone was 10 r.p.m. (6 dyne/cm2) for the first 15 s and was increased to 1800 r.p.m. (108 dyne/cm2) after a few seconds. For inhibition studies, platelets were incubated for 5 min with 10 μmol L−1 PP1, 0.1 μmol L−1 wortmannin, 2 mm EGTA, 1 U mL−1 of apyrase, 1 mmol L−1 of GRGDS, 20 μg mL−1 of AJvW-2 Fab fragment or 300 μmol L−1 calpeptin.

Measurement of the platelet aggregation

Platelet aggregation induced by ristocetin or botrocetin in the presence of VWF was measured with the use of an AG-10 platelet aggregometer (Kowa Optimed Co. Ltd, Tokyo, Japan) as previously described [29]. Briefly, 300 μL of washed platelets was applied to the glass cuvette and was incubated with 20 μg mL−1 of VWF or 20 μg mL−1 of rA1 domain for 5 min. Platelets were activated by indicated agonists at 37°C under continuous stirring at 1000 r.p.m. The instrument was calibrated with a platelet suspension for 100% OD and with buffer for 0% OD.

Immunoprecipitation and immunoblotting assays

PECAM-1 was immunoprecipitated from platelet lysates, as previously described [30]. All immunoprecipitation steps were carried out at 4°C. Briefly, washed platelets were solubilized with an equal volume of 2× ice-cold lysis buffer (100 mmol L−1 Tris/HCl, pH 7.4, 2% Triton X-100, 5 mmol L−1 EGTA, 2 mmol L−1 Na3VO4, 1 mmol L−1 PMSF, 50 μg mL−1 leupeptin). The lysates were sonicated and were centrifuged at 15 000 × g for 5 min. The supernatants were precleared with protein A-sepharose beads for 30 min. For immunoprecipitation, the supernatants were incubated with appropriate antibodies, followed by the addition of protein A-sepharose beads. The precipitates obtained after centrifugation were washed three times in 1× lysis buffer before the addition of Laemmli sample buffer.

The precipitated proteins were separated by sodium dodecyl sulfate/polyacrylamide gel electrophoresis (SDS/PAGE) and transferred onto PVDF membranes. The membranes were blocked with 10% BSA in phosphate-buffered saline (PBS). After extensive washing of PVDF membrane with TBS-T (10 mm Tris/HCl, pH 7.4, 100 mm NaCl, 1% Tween 20), the immunoblots were performed with appropriate antibodies for 2 h at room temperature. Antibody binding was detected by using peroxidase-conjugated secondary antibodies and visualized with the ECL reaction reagent.

Flow cytometric analysis

Washed platelets, incubated with rA1 domains, were exposed to shear stress. The rA1 domains instead of VWF multimers were used to prevent platelet aggregation. Platelets were then incubated with fluorescence-labeled MoAb(s) for 20 min at room temperature in the dark. Platelets were mixed with 1% paraformaldehyde in PBS for 20 min at room temperature, and the samples were analyzed immediately by a FACScan (Becton Dickinson, San Jose, CA, USA).

Results

Cleavage of platelet PECAM-1 is induced by high shear stress

It has been reported that tyrosine phosphorylation of PECAM-1 is induced by thrombin, collagen, or convulxin [9]. We have also found that wheat germ agglutinin (WGA) specifically induces PECAM-1 tyrosine phosphorylation at a level that far exceeds those of other agonists [31]. Recently, it has been reported that tyrosine phosphorylation of PECAM-1 mediated by GPIb was induced by VWF/botrocetin interaction [32]. We hence evaluated the tyrosine phosphorylation of PECAM-1 in human platelets exposed to shear stress. While PECAM-1 underwent tyrosine phosphorylation in platelets activated by the combination of VWF and VWF modulators such as ristocetin or botrocetin, there was no tyrosine phosphorylation of PECAM-1 in platelets exposed to high shear stress in the presence of VWF, irrespective of time and magnitude of shear stress (data not shown). In the experiments confirming the recovery of PECAM-1 in immunoprecipitates of platelets exposed to shear stress, we noticed a faint band with an approximate molecular weight of 118 kDa (Fig. 1A). This band appeared 1 min after exposure to shear stress, and increased in intensity in a time-dependent manner until 11 min 30 s after shear stress was applied to platelets (Fig. 1B). Exposure to shear stress longer than 11 min 30 s was not evaluated in this study. The 118 kDa protein was not obtained with ristocetin or botrocetin plus VWF (Fig. 1A), collagen (100 μg mL−1), thrombin (1 U mL−1), WGA (0.1 mg mL−1) or convulxin (0.125 μg mL−1) stimulation (data not shown). Thus, high shear stress specifically induces the appearance of the 118 kDa protein. The 118 kDa protein could be a protein that associated with PECAM-1 under shear stress, which was non-specifically blotted by the anti-PECAM-1 MoAb (JC/70 A). However, the 118 kDa protein was not detected by antibodies such as antiphosphotyrosine MoAbs (4G10 and PY20), anti-Cbl polyclonal antibody (PoAb) or anti-Src MoAb (data not shown). Furthermore, the 118 kDa protein was also detected with another anti-PECAM-1 MoAb (9G11, data not shown). Based on these findings, we hypothesized that the 118 kDa protein was the degraded form of PECAM-1, and that the possibility of some other proteins that coprecipitated with PECAM-1 was remote. If this hypothesis holds correct, the 118 kDa protein probably exists in the supernatant fraction, but not in the pellets, as the PECAM-1 molecule consists of a long extracellular domain and a short cytoplasmic domain.

Figure 1.

A 118 kDa protein recognized by MoAb against PECAM-1 appears in response to high shear stress. (A) Washed human platelets in the presence of 20 μg mL−1 of VWF and 1 mmol L−1 Ca2+ were activated with high shear stress of 108 dyne/cm2 (a), 1.2 mg mL−1 of ristocetin (b) or 12 μg mL−1 of botrocetin (c) for the indicated durations. Shear stress-induced platelet aggregation was measured with a cone-plate viscometer and ristocetin- or botrocetin-induced platelet aggregation was measured with a platelet aggregometer, AG-10. After stimulation, the samples were lyzed and immunoprecipitated anti-PECAM-1 (JC/70 A). (B) Platelets were subjected to high shear stress of 108 dyne/cm2 up to 11 min 30 s, and then the samples were immunoprecipitated and blotted with MoAb against PECAM-1. (C) Platelets were activated in the presence of 20 μg mL−1 of VWF and 1 mm Ca2+ with high shear stress of 108 dyne/cm2. The samples were then centrifuged at 6000 × g for 30 s. The supernatants (s) and pellets (p) were subjected to immunoprecipitation and immunoblotting with anti-PECAM-1 MoAb. These results are representative of three to six experiments.

Platelet suspensions in the resting state or after exposure to high sheer stress were centrifuged to obtain the supernatant fraction and pellets. The samples were lyzed, and then immunoprecipitated with anti-PECAM-1 MoAb (Fig. 1C). PECAM-1 with the original weight of 130 kDa existed in the pellets but not in the supernatant. On the other hand, the 118 kDa protein predominantly existed in the supernatant after exposure to high shear stress, although a small amount of it also existed in the pellets. These findings support the notion that PECAM-1 is cleaved upon exposure to high shear stress. A small amount of the 118 kDa protein in the pellet may represent the tethered portion of cleaved PECAM-1 on the platelet surface.

We then sought to determine whether the degradation of PECAM-1 and its shedding into the supernatant was a consequence of platelet aggregation induced by shear stress or whether they could be induced by the interaction of the GPIb-VWF without aggregation. Shear-induced platelet activation (SIPA) is induced by the simultaneous interaction of VWF with GPIb and GPIIb/IIIa, and the clustering of GPIb by VWF multimers is considered to be a prerequisite for the GPIb-mediated platelet activation [16]. The recombinant VWF A1 domain fragment fused with the maltose-binding protein (rA1 domain) lacks the binding site for GPIIb/IIIa and is a monomer in terms of the binding site for GPIb, thus lacking the ability to cluster GPIb molecules on the platelet membrane [25]. The rA1 domain did not induce SIPA (Fig. 2A), and in fact inhibited SIPA induced by high shear stress in the presence of VWF (data not shown). On the other hand, rA1 domain did not inhibit the degradation of PECAM-1 (Fig. 2B, lane b) and, moreover, it supported the shedding of PECAM-1 in the absence of VWF multimers. These findings suggest that the GPIb–VWF interaction without GPIb clustering or platelet aggregation is sufficient for the degradation of PECAM-1. That the GPIb/VWF interaction, but not the simple mechanical force, is required for PECAM-1 degradation was further supported by the use of a monoclonal antibody (MoAb) against VWF, AJvW-2, which blocks the interaction of GPIb and VWF [14]. AJvW-2 Fab fragments effectively inhibited SIPA (Fig. 2A), and the cleavage of PECAM-1 was also completely inhibited by AJvW-2 Fab (Fig. 2B, lane c). These findings taken together suggest that the cleavage of PECAM-1 is mediated specifically by the GPIb–VWF interaction, and that this process does not require the clustering of GPIb.

Figure 2.

The118 kDa protein appearance is mediated by GPIb signaling. (A) Platelets were preincubated with 20 μg mL−1 VWF (a), 20 μg mL−1 recombinant A1 domain (b) or 20 μg mL−1 recombinant A1 domain plus 20 μg mL−1 anti-VWF MoAb Fab fragment (AJvW-2) (c) for 5 min. The samples were activated with high shear stress of 108 dyne/cm2 for 5 min 45 s. The data indicated the trace of optical density. (B) Platelets were preincubated with 20 μg mL−1 VWF (a), 20 μg mL−1 recombinant VWF A1 domain (b) or 20 μg mL−1 anti-VWF MoAb Fab fragment (AJvW-2) (c) for 5 min. The samples were activated with high shear stress of 108 dyne/cm2 for the indicated durations, and then subjected to immunoprecipitation and immunoblotting with anti-PECAM-1 MoAb. These results are representative of three experiments.

In order to determine the site of PECAM-1 cleavage, immunoblotting was carried out with P1.2 that recognizes the Ig-like domain 6th of PECAM-1 [21]. P1.2 detected PECAM-1 and the 118 kDa protein, similar to JC/70 A, in the whole cell lysate after exposure to high shear stress, while it only detected PECAM-1 of the original molecular weight in the lysate of resting platelets (Fig. 3A). Taking into consideration the molecular weight of the shed form of the protein (118 kDa), this finding suggests that PECAM-1 is cleaved at a site close to the Ig-like domain 6th, but more proximal to the C-terminal. To detect the other end of PECAM-1, i.e. the cytoplasmic portion of PECAM-1, we then performed immunoprecipitation and blotting with a PoAb against the cytoplasmic portion of PECAM-1 (BooBoo) [20], or a MoAb raised against the C-terminal 15 amino acids of PECAM-1, 235.1 [22]. They both reacted with intact PECAM-1 but not with the 118 kDa fragment (Fig. 3B). With BooBoo, multiple bands with molecular weights ranging from 15 to 30 kDa were obtained with platelets exposed to high shear stress in some experiments (data not shown). However, this was not always reproducible. With 235.1, no proteins could be recovered with immunoprecipitation, and no bands other than 130 kDa PECAM-1 could be detected in platelets exposed to high shear stress by immunoprecipitation and immunoblotting. However, intact PECAM-1 as detected by BooBoo or 235.1 decreased after stimulation of high shear stress (Fig. 3B). These findings are consistent with the hypothesis that PECAM-1 is cleaved under high shear stress. Based on these findings, we assume that the C-terminal PECAM-1 is preferentially degraded after platelets are exposed to high shear stress, and that there may be multiple sites of cleavage. However, since JC/70 A and P1.2 only detected 130 kDa PECAM-1 and the 118 kDa fragment, but no other fragments of different molecular weights, it is suggested that there is no cleavage site in the 118 kDa fragment.

Figure 3.

PECAM-1 is cleaved at near the Ig domain 6th under high shear stress. (A) Platelets were activated with high shear stress of 108 dyne/cm2 for the indicated durations and the reaction was terminated by the addition of SDS sample buffer (R and NR stand for reduced- and non-reduced-samples, respectively). Proteins were resolved on 8% SDS-PAGE, and then immunoblotted with anti-PECAM-1 MoAbs, JC/70 A or P1.2. (B) Platelets were challenged with 108 dyne/cm2 high shear stress for 5 min 45 s. Immunoprecipitation and immunoblotting were performed with indicated antibodies. The results are representative of two to five experiments.

Effects of several inhibitors on cleavage of PECAM-1 induced by high shear stress

SIPA is supported by extracellular calcium ions ([Ca2+]o) [33–35], ADP [36], fibrinogen [37] and VWF [38]. We then evaluated the effects of the antagonists of these substances on the cleavage of PECAM-1. We have already shown that anti-VWF MoAb inhibits high shear-induced degradation of PECAM-1 (Fig. 2B). On the other hand, GRGDS (an inhibitor of fibrinogen-GPIIb/IIIa or VWF–GPIIb/IIIa interaction) which moderately blocked SIPA (data not shown) had no effect on PECAM-1 cleavage (Fig. 4, lane g). It has been demonstrated that the GPIIb/IIIa–VWF interaction is also involved in SIPA [38]. Our finding that GRGDS had no effect on PECAM-1 cleavage suggests that PECAM-1 cleavage under high shear stress is not the consequence of platelet aggregation mediated by the GPIIb/IIIa–VWF interaction, and further substantiates the notion that PECAM-1 cleavage is induced by the GPIb–VWF interaction. Two millimolar EGTA completely inhibited PECAM-1 cleavage, while apyrase (ADP scavenger) had no effects on high shear-induced cleavage of PECAM-1 (Fig. 4, lane e and f).

Figure 4.

Effects of several inhibitors on PECAM-1 cleavage induced by high shear stress. Platelets were challenged with 108 dyne/cm2 high shear stress for 5min 45 s in the presence of vehicle (b), 10 μmol L−1 PP1 (c), 0.1 μmol L−1 wortmannin (d), 2 mm EGTA (e), 1 U mL−1 of apyrase (f), 1 mmol L−1 of GRGDS (g) or 20 μg mL−1 of AJvW-2 Fab fragment (h). Lane (a) represents the control without stimulation. Immunoprecipitation and immunoblotting were performed as described in the legend for Fig. 1. The results are representative of two experiments.

It has been recently reported that GPIb associates with PI-3 kinase and Src upon the GPIb/VWF interaction, and this association is involved in the signal transduction pathway mediated by GPIb [39]. PP1 (10 μmol L−1) and Wortmannin (0.1 μmol L−1), specific inhibitors of Src and PI-3 kinase, respectively, did not affect the high shear-induced cleavage of PECAM-1 (Fig. 4, lane c and d).

Involvement of calpain activation and calcium ions on cleavage of PECAM-1

We hitherto found that EGTA inhibited the cleavage of PECAM-1. This finding led us to hypothesize that some protease(s), which requires calcium ions for its activation, is released from platelets and activated under high shear stress with resultant cleavage of PECAM-1. As recent reports suggest the presence of matrix metalloproteinases and caspases in platelets [40,41], we used GM6001 (a broad spectrum inhibitor of metalloproteinases) and ZVAD-FMK (a general caspase inhibitor). However, these two proteinase inhibitors had no effect on cleavage of PECAM-1 (data not shown). We then investigated the role of calpain for PECAM-1 cleavage. Calpain is one of the proteinases which requires calcium ions for its activation, and the inactive form of μ-calpain has a molecular weight of 80 kDa. Upon activation it undergoes auto-degradation with resultant 76 kDa μ-calpain [42]. In the presence of VWF, the active form of μ-calpain was detected in platelets activated by ristocetin, botrocetin or high shear stress, which is in agreement with a previous report [43](Fig. 5A). However, μ-calpain was activated under high shear stress but not by ristocetin or botrocetin in the presence of rA1 domain (Fig. 5A). These findings suggest that GPIb clustering is required for calpain activation in the case of ristocetin or botrocetin. We have no clear explanation for the absence of PECAM-1 cleavage with ristocetin/botrocetin activation under no or low shear stress, in which calpain activation is present. For some factors, such signaling from mechanoreceptors may be required in addition to calpain activation to bring about PECAM-1 cleavage. The activation of μ-calpain and PECAM-1 cleavage under high shear stress were detected in the same samples with similar time courses, and they were both inhibited by the preincubation with 300 μmol L−1 calpeptin, a specific inhibitor of calpain [44] (Fig. 5B, upper panel, lane b).

Figure 5.

Involvement of calpain in PECAM-1 cleavage. (A) Platelets in the presence of 20 μg mL−1 VWF (a-d) or in the presence of 20 μg mL−1 of rA1 domain (a′-d′) were activated without (a, a′) or with 1.2 mg mL−1 ristocetin (b, b′), 3 μg mL−1 botrocetin (c, c′) or high shear stress of 108 dyne/cm2 (d, d′) for the indicated durations. Shear stress-induced platelet aggregation was measured with a cone-plate viscometer and ristocetin- or botrocetin-induced platelet aggregation was measured with a platelet aggregometer, AG-10. Immunoblotting was performed with anti-calpain PoAb. (B) Platelets preincubated with vehicle (a) or 300 μmol L−1 calpeptin (b) for 5 min in the presence of 20 μg mL−1 VWF were activated with high shear stress of 108 dyne/cm2 for the indicated durations. The samples were lyzed and immunoblotted with anti-calpain PoAb (upper panel) or anti-PECAM-1 MoAb (lower panel). These results are representative of three experiments.

Calpain is known to work on its substrates in the cytoplasm [45]. Alternatively, it is also possible that calpain after being released extracellularly acts on PECAM-1. In order to address this issue, we evaluated the effects of exogenous μ-calpain on PECAM-1 degradation. Platelet suspensions with exogenous μ-calpain for various frames of time at 30°C in the presence of 1 mm Ca2+. PECAM-1 cleavage was not detected even after 30-min incubation with μ-calpain (Fig. 6A, lane b). We then challenged platelets pretreated with μ-calpain for 30 min with high shear stress (Fig. 6A, lane c). There was little difference in the magnitude of PECAM-1 cleavage between μ-calpain-pretreated platelets (Fig. 6A, lane c) and the control platelets (data not shown), and PECAM-1 cleavage was inhibited by AJvW-2 (Fig. 6A, lane d). To confirm that exogenous calpain we used in this experiment was effective, intact PECAM-1 bound to protein A beads was prepared from platelet lysates by immunoprecipitation with anti-PECAM-1 MoAb. When μ-calpain was incubated with the beads for various frames of time at 30°C, PECAM-1 was degraded with the appearance of the 118 kDa fragment in a time-dependent manner (Fig. 6B).

Figure 6.

Involvement of purified μ-calpain on cleavage of PECAM-1. (A) Platelets were incubated without (a) or with 0.25 U mL−1μ-calpain (b–d) in the presence of 1 mm Ca2+ at 30 °C for 30 min, and then μ-calpain-treated platelets were incubated with vehicle (b), 20 μg mL−1 VWF (c) or 20 μg/mL AJvW-2 MoAb Fab fragment (d) for 5 min. Platelets were activated without (a, b) or with high shear stress of 108 dyne/cm2 for 5 min and 45 s (c,d). Immunoprecipitation and immunoblotting were performed as descried in the legend of Fig. 1. (B) Intact platelets were lyzed and immunoprecipitated with anti-PECAM-1 MoAb. The immunoprecipitates were incubated with vehicle or 0.25 U mL−1μ-calpain at 30 °C for the indicated durations in the presence of 1 mmol L−1 Ca2+ with gentle shaking. The samples were resolved on 8% SDS/PAGE, and then immunoblotted with anti-PECAM-1 MoAb. (C) Washed platelets in the presence of 1 mmol L−1 Ca2+ and 20 μg mL−1 VWF were preincubated with vehicle (a) or 146 μmol L−1 E64 (c) for 5 min at 37°C. E64d-loaded platelets (b) were prepared by incubating platelet-rich plasma with 146 μmol L−1 E64d for 15 min at 30°C. Platelets were washed, and resuspended in a modified HEPES-Tyrode buffer containing 1 mm Ca2+ and 20 μg mL−1 VWF. Platelets (a, b, and c) were activated with high shear stress of 108 dyne/cm2 for 5 min 45 s, and immunoprecipitation and immunoblotting were performed as descried in the legend of Fig. 1. These results are representative of three experiments.

It is known that GPIb is degraded with resultant production of glycocalicin by exogenously added calpain [46]. Flow cytometric analysis showed that approximately 40% of intact GPIb was present on the platelet surface after incubation with μ-calpain for 30 min (data not shown). However, the magnitude of SIPA with μ-calpain -treated platelets did not differ significantly from that of the control platelets. These findings suggest that exogenously added μ-calpain cleaved a considerable portion of GPIb but not PECAM-1, and that the remaining GPIb molecules on the platelet surface upon exposure to high shear stress sufficed to mediate activation signals that lead to PECAM-1 cleavage.

These findings taken together imply that calpain-catalyzed PECAM-1 cleavage does not take place in the extra cellular fluid. To confirm the intracellular activation of calpain, we then used E64d or E64, a membrane-permeable or membrane-impermeable calpain inhibitors, respectively [43]. After platelet-rich plasma was incubated with E64d for 15 min, platelets were washed and resuspended in a buffer containing 1 mmol L−1 Ca2+. Under this condition, cleavage of PECAM-1 was completely inhibited by E64d (Fig. 6C, lane b). The cleavage of PECAM-1 was not inhibited by pretreating platelets with E64 (Fig. 6C, lane c). These results strongly suggest that calpain activated by high shear stress cleaves PECAM-1 at the intracellular domain.

Relationship between the platelet-derived microparticles generation and a decrease in PECAM-1 expression on platelets

As the Western blotting studies hitherto revealed PECAM-1 degradation in platelets after exposure to high shear stress, the expression of PECAM-1 on the platelet surface was quantitatively assessed with FACS analysis. In Fig. 7(A), FACScan analysis revealed a decrease in the PECAM-1 expression after exposure to high shear stress, and simultaneous immunoprecipitation analysis also confirmed the cleavage of PECAM-1 (Fig. 7B). Low shear stress had no effects on PECAM-1 expression and cleavage of PECAM-1 (Fig. 7A,B). AJvW-2, which blocks the interaction between GPIb and VWF completely inhibited the decrease in PECAM-1 expression induced by high shear stress (Fig. 7C). In addition, calpeptin or EGTA almost totally inhibited the decrease in PECAM-1 expression (data not shown). These findings are also consistent with the effects of these inhibitors in PECAM-1 immunoprecipitation experiments.

Figure 7.

PECAM-1 expression on the platelet surface under high shear stress. Platelets were exposed to low shear stress of 6 dyne/cm2 or high shear stress of 108 dyne/cm2 in the presence of 20 μg mL−1 rA1 domain for 5 min 45 s. Aliquots of the samples (10 μL) were incubated with PE-labeled anti-PECAM-1 MoAb for flow cytometric analysis (A). The gray lines and dark lines represent low shear stress and high shear stress, respectively. The control sample gave results essentially similar to low shear stress. The residual samples were applied for immunoprecipitation and immunoblot analysis (B).  (C) In the presence of 1 mmol L−1 Ca2+ and 20 μg mL−1 of rA1 domain, platelets were incubated with 20 μg mL−1 of AJvW-2 for 5 min at 37°C, and then platelet were activated with high shear stress of 108 dyne/cm2 for 5 min 45 s. The gray lines and dark lines represent the control without stimulation and high shear stress, respectively. The control sample gave results essentially similar to high shear stress. These results are representative of three experiments.

To our surprise, there was no significant left shift in PECAM-1 fluorescence, suggesting that in the majority of platelets, there was virtually no change in the level of PECAM-1 expression between the resting state and after exposure to high shear stress. It is also suggested that the decrease in PECAM-1 expression occurs in a specific population of platelets. There have been several reports on the production of platelet microparticles (PMP) induced by shear stress [47], the size of which is smaller than ordinary platelets. We then set separate gatings for PMP and platelets of normal size, and sought to evaluate the relationship between PMP formation and changes in PECAM-1 expression. As shown in the R2 region which represents PMP in Fig. 8(B, upper panel), there was a considerable production of PMP after exposure to high shear stress for 5 min 45 s, and PECAM-1 expression was significantly decreased in the R2 region (Fig. 8B, lower panel). On the other hand, the PECAM-1 expression in R1 which represented platelets of normal size was unaffected by exposure to high shear stress. It may be argued that the level of PECAM-1 expression in PMP is low simply because of their decreased surface area. To address this issue, we compared the ratio of the expression of PECAM-1 or GPIIa, respectively, in PMP (R2 in Fig. 8B, upper panel) to that of intact platelets of normal size (R1 in Fig. 8A, upper panel). It has been reported that there was no significant difference in the expression of GPIaIIa on platelets surface between resting and activated states [48]. The ratio of the mean fluorescence level (R2 vs. R1) was 20.4% for PECAM-1, and 46.1% for GPIIa (the mean of three experiments). These findings demonstrate that PECAM-1 is specifically degraded in PMP produced under high shear stress, and suggest that PECAM-1 cleavage is closely related to the process of PMP formation. It is also of interest that the mean fluorescence value of platelets of normal size had no significant change after exposure to high shear stress (Fig. 8B, lower panel). This finding implies that the process of PMP production and PECAM-1 degradation occurs in an all-or-none manner, in which a specific population of platelets undergoes PECAM-1 degradation and complete disruption to produce PMP, while the majority of platelets remain unaffected. Alternatively, it is possible that a portion of cleaved PECAM-1 remained tethered to platelets of normal size and therefore there is no apparent decrease in PECAM-1 expression as assessed by flow cytometry. This issue awaits elucidation.

Figure 8.

Relationship between PECAM-1 cleavage and platelet-derived microparticle generation under high shear stress. Platelets were preincubated with 20 μg mL−1 of rA1 domain and 1 mmol L−1 Ca2+, and activated with high shear stress of 108 dyne/cm2 for 5 min 45 s (B). The samples were incubated with PE-labeled anti-PECAM-1 MoAb and FITC-labeled anti-GPIIa MoAb, and the fluorescence was analyzes by flow cytometry. (A) represents the control without stimulation. Upper panels represent dot plot analysis (FSC vs. SSC) of the control sample and that of high shear stress, respectively. R1 and R2 represent the regions of platelets and platelet-derived microparticles, respectively. Lower panels show histograms of PECAM-1 fluorescence corresponding to the upper panels. In lower panels, the left ordinate represents the event counts for R1, and the right ordinate for R2. Solid lines and gray regions represent those of the control and high shear stress, respectively. These results are representative of five experiments.

Discussion

In this study, we demonstrate that PECAM-1 is cleaved under high shear stress. That the 118 kDa protein represents the cleaved fragment of PECAM-1 is suggested by several lines of evidence; this protein was recognized by three different monoclonal antibodies that react with the extracellular domain of PECAM-1. P1.2, a MoAb, which recognizes the Ig-like domain 6th of PECAM-1, reacted with the 118 kDa protein (Fig. 3A). Furthermore, a PoAb, BooBoo that recognizes the cytoplasmic domain of PECAM-1 did not react with the 118 kDa protein (Fig. 3B). These findings taken together suggest that upon exposure to high shear stress, on the platelet membrane PECAM-1 is cleaved at a site somewhere below the 6th Ig-like domain, producing the 118 kDa fragment which is shed extracellularly. While the cleavage of PECAM-1 should produce other fragments of lower molecular weight, it was difficult to detect them with two antibodies against the cytoplasmic domain of PECAM-1. It is likely that the shorter C-terminal fragment of PECAM-1 has multiple sites of cleavage.

In this study, PECAM-1 cleavage was detectable when platelets were exposed to high shear stress for a few minutes. It may be argued that in vivo platelets are exposed to high shear stress for a short period of time at sites of arterial bifurcation or stenosis. However, we assume that platelets are repetitively exposed to high shear stress in vivo, and that the experimental conditions in this study would represent the accumulated effects of repetitive exposures to high shear stress. It has been reported that the level of soluble PECAM-1 is elevated in stroke [49] or acute myocardial infarction [50]. As arterial stenosis with generation of high shear stress can be expected in these disorders, it is likely that platelets are repetitively exposed to high shear stress. Our notion that PECAM-1 is degraded under high shear stress may explain the elevated level of soluble PECAM-1 in these clinical settings.

As the combination of high concentrations of A23187 and thrombin induced PECAM-1 degradation (data not shown), the phenomenon is not confined to high shear stress. However, taking into consideration the fact that PECAM-1 cleavage was below the detectable level with supraoptimal concentrations of collagen, thrombin or WGA at our hand (data not shown), it appears to be specific for high shear stress under ordinary conditions. It is to be noted that high shear stress, as platelet stimulus, is rather weak in terms of degranulation and Ca2+ mobilization, in comparison with these three agonists [51]. What factors are required for PECAM-1 cleavage in addition to the signaling molecules involved in platelet activation induced by thrombin, collagen or WGA awaits to be elucidated.

The Fab fragment of AJvW-2, a MoAb against the A1 domain of VWF completely inhibited PECAM-1 cleavage as well as SIPA, suggesting that PECAM-1 cleavage is mediated specifically by the interaction between GPIb and VWF. It is of particular interest that the combination of botrocetin and VWF, which is widely accepted to mediate the GPIb-related signal transduction pathway, does not induce PECAM-1 cleavage. We have also found other differences between high shear stress and botrocetin stimulation in this study, in addition to PECAM-1 cleavage; PECAM-1 tyrosine phosphorylation is present with VWF plus botrocetin, but not with high shear stress. PMP release is induced by high shear stress, but not by VWF plus botrocetin (data not shown). Calpain is activated by high shear stress, botrocetin or ristocetin in the presence of VWF (Fig. 5A), which is in agreement with a previous report [43], while in the presence of the rA1 domain, high shear stress but not botrocetin or ristocetin activates calpain (Fig. 5A). We have no clear explanation for this apparent discrepancy, as both SIPA and that of VWF and botrocetin are known to be mediated by the GPIb complex. Miyazaki et al. also found that PMPs were abundantly generated by high shear stress but not by ristocetin or botrocetin in the presence of VWF [47]. While it is possible that the involvement of certain mechano-receptors in addition to the role of GPIb is required for PECAM-1 cleavage, we do not have any concrete evidence to support this notion, and this issue awaits to be addressed in the future.

It is known that SIPA is dependent on the interaction of GPIb and VWF, and on that of GPIIb/IIIa and VWF. GRGDS, which interferes with the GPIIb/IIIa–VWF interaction, did not block PECAM-1 cleavage. The rA1 domain is a monovalent A1 domain linked to the maltose-binding protein, and thus cannot support platelet agglutination/aggregation induced by VWF multimers. We found that rA1 domain inhibited SIPA but not PECAM-1 cleavage and it sufficed to induce PECAM-1 cleavage under high shear stress. These findings suggest that PECAM-1 cleavage is mediated via the GPIb–VWF interaction, and that it is not the consequence of SIPA. Furthermore, our findings also suggest that PECAM-1 cleavage does not require the clustering of GPIb complex, in contrast to a widely held notion that the GPIb-mediated platelet activation requires GPIb clustering by VWF multimers.

PECAM-1 cleavage implies the involvement of certain proteases. Ilan et al. reported that endothelial PECAM-1 is cleaved in the process of apoptosis, and this cleavage is mediated by caspases and matrix metalloproteinases (MMP) [52]. It has been reported that MMPs are released from platelets during aggregation induced by ADP, thromboxane A2 or collagen [53]. Li et al. reported that mRNAs of several caspases exist in platelets, and caspases 3 is activated during storage [54]. We thus used several inhibitors of MMPs and caspases on PECAM-1 cleavage induced by shear stress, only to find that these inhibitors did not modify PECAM-cleavage.

Calpain is an intracellular calcium-activated cysteine protease, found ubiquitously in mammalian and avian tissues and cells [19], and it was reported that calpain is activated in SIPA [55]. As EGTA, a potent calcium chelator, inhibited PECAM-1 cleavage induced by high shear stress, we explored the possibility that calpain is involved in this process. We confirmed that calpain auto-degradation, which represents its activation occurs in response to high shear stress, is blocked by calpeptin, a membrane-permeable inhibitor of calpain. As expected, calpeptin also potently inhibited the cleavage of PECAM-1.

Pretreatment of platelets with a membrane-permeable calpain inhibitor, E64d [43], completely inhibited PECAM-1 cleavage induced by high shear stress. On the other hand, a membrane-impermeable calpain inhibitor, E64 [43], could not inhibit cleavage of PECAM-1 under high shear stress. These findings imply that calpain activation and its action on PECAM-1 takes place in the cytoplasm, but not in the extracellular fluid after its release. There are additional lines of evidence to support this hypothesis. Exogenously added active μ-calpain effectively cleaved PECAM-1 purified from platelet lysates, while the same concentration of μ-calpain added to intact platelets failed to do so, while glycocalicin production was detected under the same condition, suggesting that GPIb was effectively cleaved (data not shown). Exogenous μ-calpain also did not augment PECAM-1 degradation in platelets exposed to high shear stress (Fig. 6A, lane c). As the theoretical molecular weight of the cytoplasmic domain of PECAM-1 (AA527-626) is 11 kDa, the molecular weight of the cleaved form of PECAM-1, 118 kDa, is also compatible with the notion that calpain acts intracellularly.

We have found that high shear stress induces PECAM-1 degradation and its shedding in human platelets. It is directly mediated by the GPIb–VWF interaction, but not the consequence of platelet aggregation. The functional role of this phenomenon remains largely unknown. However, in this study we present some lines of evidence to suggest that PECAM-1 degradation is related to PMP formation. In fact, PECAM-1 degradation is exclusively present in PMP. In platelets of normal size, there was virtually no change in PECAM-1 expression. Based on these findings, we assume that PECAM-1 cleavage does not occur uniformly in all platelets, but that it is confined to a certain population of platelets. Once PECAM-1 cleavage occurs, platelets may undergo almost complete disruption and PMP formation. This all-or-none manner of PECAM-1 degradation and PMP formation is still a bold hypothesis, which needs to be reinforced by further investigation. Miyazaki et al. reported that high shear stress leads to the generation of PMP, and that its mechanism involves the binding of VWF to GPIb, influx of extracellular calcium, and activation of platelet calpain [47]. Our findings in this study are in good agreement with theirs.

It is known that PMP exhibited prothronbinase activity [56], and that local generation of PMP in small atherosclerotic arteries or arterioles may promote acute arterial occlusion by providing and expanding a catalytic surface for coagulation cascade. Microparticle generation by high shear stress can occur in atherosclerotic arteries and arterioles in various clinical settings including occlusive diseases.

PMP production is a sign of apoptosis in a number of cells [57]. It has been recently reported that PECAM-1 degradation and the resulting short cytoplasmic domain of PECAM-1 accelerates apoptosis in endothelial cells [52]. The cytoplasmic domain of PECAM-1 contains an immunoreceptor tyrosine-based inhibitory motif (ITIM) [11]. Phosphorylated ITIM binds to SH2-containing signaling proteins, and PECAM-1 degradation may thus generate intracellular signals leading to apoptosis. By analogy, it is possible that in a certain population of platelets, presumably aged platelets, PECAM-1 is apt to be influenced by high shear stress with resultant PECAM-1 cleavage and PMP formation. It is feasible that PECAM-1 in a physiological environment serves to suppress destruction of platelets, and our study may provide clues for the measures that preserve platelet viability during storage of platelets.

In conclusion, we have found in this study that the cleavage of PECAM-1 is induced by high shear stress. The GPIb–VWF interaction is a prerequisite for this phenomenon, but GPIb clustering or aggregation is not required. Calpain activation appears to play an essential role for PECAM-1 cleavage. In addition, we suggest that PECAM-1 cleavage is related to PMP production under high shear stress.

Acknowledgements

We are grateful to Dr J. Madri at Yale University School of Medicine; Dr P. Newman at The Blood Center of South-eastern Wisconsin; Dr M. Berndt at Monash University; Dr T. Morita at Meiji Pharmaceutical University; Dr M. Handa at Keio University; Dr Ishii at Showa Pharmaceutical University for providing antibodies and snake venoms. This work was supported in part by Mitsubishi Pharma Research Foundation and by a Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology, Japan.

Ancillary