|
|
||||||||
U553 INSERM: Hémostase, Endothélium et Angiogénèse, Institut dHématologie, Université Paris VII-Denis Diderot, IFR Saint-Louis, Hôpital Saint-Louis, Paris Cedex, France
1Correspondence: U 553 INSERM, Hôpital Saint-Louis, 1 Avenue Claude Vellefaux, 75475 Paris-cedex 10, France. E-mail: f.lafeve{at}chu-stlouis.fr
| ABSTRACT |
|---|
|
|
|---|
Key Words: platelet aggregation glycoprotein platelet-collagen interaction
| INTRODUCTION |
|---|
|
|
|---|
2ß1 integrin) (2)
Among the platelet reacting sites on collagen, the GPO motifs, making up
10% of type I and type III collagen (4)
, constitute binding sites for the GP VI receptor. GP Ia/IIa possesses multiple binding domains within type I collagen (5
, 6)
and type III collagen (7)
. The participation of GP Ia/IIa in type I collagen-induced platelet adhesion is well recognized, but its contribution to platelet adhesion to type III collagen is controversial and depends on the conformation of type III collagen (8
, 9)
. Within type III collagen, the
1(III)-CB4 fragment was shown to be a highly reactive site for platelets (5)
containing GP Ia/IIa (7)
and vWF binding sites (10)
. We previously demonstrated the existence of another platelet reactive site within this CB4 fragment from type III collagen corresponding to the octapeptide sequence KOGEOGPK (11)
. When immobilized, this isolated octapeptide induced platelet adhesion in both static and flow conditions (12)
but inhibited type III collagen-induced platelet aggregation (13)
. The platelet receptor for this octapeptide has been identified as a new platelet receptor for collagen and called type III collagen binding protein (TIIICBP) (12)
. This receptor, a 6872 kDa protein doublet, was shown to be a primary platelet receptor for type III collagen since blocking TIIICBP by the octapeptide or monoclonal antibodies prevented platelet contact and spreading on type III collagen under static and flow conditions. Moreover, the blockade of TIIICBP inhibited platelet adhesion and aggregation induced by type III collagen but not by type I collagen (14)
.
Interaction of platelets with collagen triggers platelet activation and signaling. Collagen binding to GP VI, a member of the immunoglobulin superfamily (15
16
17)
noncovalently coupled with the Fc receptor (FcR)
-chain (18
, 19)
, triggers phosphorylation of the FcR
-chain by Src family tyrosine kinases (20
, 21)
and leads to the recruitment and activation of Syk (22)
, resulting in PLC
2 activation. Several adaptor proteins have been shown to be critical in response to activation of GP VI, most notably the linker for activation of T cells (LAT) (23)
. Signaling by GP Ia/IIa was controversial until the recent studies by Inoue et al. (24
, 25)
demonstrating GP Ia/IIa-dependent tyrosine phosphorylation of many signaling proteins including Syk, PLC
2, and FAK but not LAT.
In the present study, we have focused on platelet adhesion and signaling induced by the immobilized KOGEOGPK octapeptide. We demonstrate that platelet adhesion to the immobilized octapeptide is independent of Ca2+ and Mg2+ cations and independent of GP VI. Moreover, we show that octapeptide-mediated platelet adhesion is potentiated by released ADP (Gi-coupled pathways), with evidence for very minor participation of TXA2. We demonstrate for the first time that platelet adhesion to a primary binding site from type III collagen triggers a signaling pathway similar to the one induced by GP VI through tyrosine phosphorylation of the FcR
-chain, LAT, Syk, PLC
2, and FAK. TIIICBP is phosphorylated after octapeptide-induced platelet adhesion and constitutively associated with the FcR
-chain and LAT but not with GPVI. However, our results suggest a potential cooperation between TIIICBP and GPVI during GPVI-mediated platelet activation.
| MATERIALS AND METHODS |
|---|
|
|
|---|
-chain, anti-LAT antibodies, and mouse monoclonal antibody against phosphotyrosine (4G10) were purchased from Upstate Biotechnology (Lake Placid, NY, USA). Mouse monoclonal anti-Syk antibody was obtained from Chemicon International (Temecula, CA, USA). Rabbit polyclonal anti-PLC
2 and anti-FAK antibodies were from Santa Cruz Biotechnology (Santa Cruz, CA, USA). The convulxin (cvx) and the 3J24.2 and 9O12.2 monoclonal antibodies against GP VI were kindly donated by Dr. Martine Jandrot-Perrus (Inserm, Hôpital Bichat, Paris, France) and the 7F4 monoclonal antibody against TIIICBP was produced by our laboratory as described previously (14)
Platelet isolation
Blood was obtained from consenting healthy human donors and anticoagulated with acid-citrate-dextrose. Washed platelets were prepared as described previously (27)
. After the last wash, platelets were resuspended into a pH 7.5 Tyrodes buffer (137 mM NaCl, 2.7 mM KCl, 1.2 mM NaHCO3, 0.36 mM NaH2PO4, 2 mM CaCl2, 1 mM MgCl2, 5 mM HEPES, 5.5 mM glucose) and adjusted to 2 x 108 platelets/mL.
Platelet adhesion assay
F96-MaxiSorp plates (Nunc, Naperville, IL, USA) were coated with 100 µL per well of either 0.1 mg/mL type I collagen, type III collagen, or octapeptide in 0.1 M phosphate buffer (pH 8.0) for 2 h at room temperature; 0.1% BSA was used as a negative control in separate wells. The wells were then washed three times with 0.1 M phosphate buffer (pH 8.0). Washed platelet (107/well) preparations were usually preincubated with 1 mM RGDS for 10 min at 37°C to block the GP IIb/IIIa functions (except for kinetic analysis, where the global washed platelet deposition was studied) and added to the wells for 1 h at room temperature. After three washes in Tyrodes buffer, the number of adherent platelets was evaluated colorimetrically as described by Bellavite et al. (28)
. The number of adherent platelets was
15 to 21% of added platelets on both types of collagen or octapeptide and normalized at 100% adhesion (control) in each experiment.
Influence of divalent cations on platelet adhesion
For quantification of Ca2+ and Mg2+ effects on platelet adhesion to octapeptide, the number of adherent platelets was determined using washed platelets in a free MgCl2 and CaCl2 Tyrodes buffer, in the same buffer containing 1 mM MgCl2 or 2 mM CaCl2, or both 1 mM MgCl2 and 2 mM CaCl2.
Involvement of GP VI in platelet adhesion
Platelet suspensions were pretreated with the 9O12.2 antibody for 5 min at 37°C before their addition to the microtiter wells.
Influence of generated TXA2 and ADP receptors on platelet adhesion
To determine the influence of generated TXA2 on platelet adhesion, washed platelets were preincubated for 1 min at 37°C with 5 µM of indomethacin (Sigma), a cycloxygenase inhibitor, before their addition to the microtiter wells. The importance of P2Y1 and P2Y12 on platelet adhesion to the octapeptide was evaluated using 111In-labeled platelets (27)
pretreated for 10 min at 37°C with 100 µM of A3P5P (Sigma) or 10 µM of AR-C69931MX, respectively.
Platelet aggregation
Aggregation was monitored by measuring light transmission through a suspension of stirred washed platelets (2x108/mL) using a Beckman Chrono-Log aggregometer.
Immunoprecipitations
After 1 h adhesion, adherent platelets were lysed with a ice-cold lysis buffer containing 2% Triton X-100, a protease inhibitor cocktail (Roche Diagnostics, Mannheim, Germany), 2 mM sodium orthovanadate, and 10 mM sodium fluoride in free CaCl2 and MgCl2 Tyrodes buffer. Platelet lysates were incubated for 30 min at 4°C under gentle end-over-end mixing. After incubation, platelet lysates were centrifuged at 12,000 g for 10 min to remove insoluble materials and protein concentration in supernatant was determined. Immunoprecitation of the FcR
-chain, LAT, Syk, PLC
2, and FAK was performed using protein G SepharoseTM and their corresponding antibodies. Immunoprecipitation of TIIICBP receptor was performed using the 7F4 monoclonal antibody against TIIICBP.
Immuno- and ligand blotting
Analysis of protein tyrosine phophorylation was performed on immunoprecipitates of whole cell lysates of adherent or resting platelets. The same protein concentration for each aliquot was resuspended in Laemmli buffer, heated at 100°C for 5 min and subjected to SDS-PAGE on 7.5% acrylamide or 4%-20% tricine gels. Proteins were transferred on a nitrocellulose membrane and probed with 4G10. After stripping, the membrane was first incubated with antibodies against the FcR
-chain, LAT, Syk, PLC
2 or FAK followed by HRP-labeled anti-mouse or -rabbit antibodies. In the case of TIIICBP, the membrane was first probed with biotinylated-octapeptide, followed by HRP-coupled streptavidin. Proteins were visualized by ECL.
Statistical evaluation
Results are reported as mean of three experiments run in triplicate ± SEM and statistical analysis were performed using the unpaired Students t test.
| RESULTS |
|---|
|
|
|---|
10 min, then increased gradually. At 60 min, the number of deposited platelets on the octapeptide (18.8±2.6x105) or on type III collagen (14.8±5.8x105) was not significantly different. Platelet suspensions were next preincubated with 1 mM RGDS at a concentration that completely blocked collagen-induced platelet aggregation in stirred suspension (data not shown) to inhibit the GP IIb/IIIa participation in ligand binding, platelet aggregation, and avoid an overestimation of the adhesion rate. Under these conditions, the number of adherent platelets at 60 min was decreased by 33% on the octapeptide and by 37% on type III collagen.
|
Platelet adhesion to immobilized octapeptide
Platelet adhesion to octapeptide is Ca2+ and Mg2+ independent
GP Ia/IIa-mediated platelet adhesion to type I collagen has been shown to be dependent on Mg2+ cations (2)
. To evaluate the importance of divalent cations in octapeptide-induced platelet adhesion, adhesion experiments were performed 1) in the presence of both 2 mM CaCl2 and 1 mM MgCl2 (considered as 100% adhesion), 2) in the presence of either CaCl2 or MgCl2, 3) in the absence of CaCl2 and MgCl2, and compared with type I and type III collagen-induced platelet adhesion. Contrary to type I collagen, platelet adhesion to immobilized octapeptide or type III collagen was independent of either Ca2+ or Mg2+ cations (Fig. 2
), suggesting that GP Ia/IIa did not participate in octapeptide-induced platelet adhesion.
|
Platelet adhesion to octapeptide is GP VI independent
To look for an eventual participation of GP VI in platelet adhesion to the octapeptide, we used the recently described 9O12.2 antibody against GP VI (31)
. As shown in Fig. 3
, preincubation of platelets with 20 µg/mL 9O12.2 had no effect on platelet adhesion to octapeptide whereas 9O12.2 significantly decreased platelet adhesion to type III collagen by 30 ± 3%.
|
Effects of ADP and TXA2 on platelet adhesion to octapeptide
The involvement of ADP receptors P2Y1 and P2Y12 in platelet adhesion to the octapeptide was next determined by using the P2Y1 antagonist A3P5P (100 µM) and the P2Y12 inhibitor AR-C69931MX (10 µM), at concentrations currently used in the literature (32
, 33)
. A3P5P had no effect on platelet adhesion to the octapeptide whereas a partial inhibition (26±6%) was observed in the presence of AR-C69931MX. In contrast, platelet adhesion to type III collagen was significantly decreased by 41 ± 7% and 76 ± 3% using A3P5P and AR-C69931MX, respectively (Fig. 4
A). The effect of generated TXA2 on platelet adhesion was determined using the cycloxygenase inhibitor indomethacin (Fig. 4B
). At 5 µM, indomethacin slightly but significantly decreased platelet adhesion to octapeptide (9±2%) and type III collagen (20±1%), indicating that TXA2 contributes minimally to platelet adhesion to the octapeptide.
|
Platelet signaling induced by the octapeptide
Kinetics of protein phosphorylation
The ability of the octapeptide to induce platelet protein tyrosine phosphorylation was investigated. After 1 h adhesion, the tyrosine phosphorylation profiles induced by platelet deposition to the octapeptide or to type III collagen were similar and blocking GP IIb/IIIa by 1 mM RGDS slightly increased the whole phosphorylation (Fig. 5
A). Kinetic analysis revealed that tyrosine phosphorylation induced by platelet adhesion to the octapeptide is a rapid phenomenon (Fig. 5B
), as many phosphorylated bands were observed after only 1 min adhesion. Because platelets did not significantly adhere to type III collagen before 10 min, it was not possible to compare their phosphorylation state at early times.
|
Immunoprecipitation of signaling proteins
To check for phosphorylation of TIIICBP, the receptor for octapeptide, TIIICBP, was immunoprecipitated from lysates of platelets adherent to either immobilized octapeptide or type III collagen, using the 7F4 monoclonal antibody. As shown in Fig. 6
A, platelet adhesion to the octapeptide or to type III collagen induced tyrosine phosphorylation of the 68 kDa subunit of TIIICBP.
|
Participation of the FcR
-chain and LAT in octapeptide-mediated platelet signaling was evaluated. We observed phosphorylation of the FcR
-chain and LAT after platelet adhesion to the octapeptide. However, the intensity of their phosphorylation was weaker than for type III collagen-adherent platelets (Fig. 6B, C
). The FcR
-chain and LAT coimmunoprecipitated in TIIICBP immunoprecipitates, even in nonactivated platelets (Fig. 6D
). A coimmunoprecipitation of GP VI with TIIICBP was not observed in our conditions (not shown).
Syk and PLC
2 were also immunoprecipitated from platelets adherent to either immobilized octapeptide or type III collagen in order to study their participation in platelet signaling. Figure 6E, F
shows that platelet adhesion to immobilized octapeptide, as to type III collagen, strongly promoted tyrosine phosphorylation of Syk and PLC
2 even in the presence of RGDS (not shown). PLC
2 coimmunoprecipitated in Syk immunoprecipitates and vice versa, demonstrating association of these two signaling molecules in octapeptide-mediated platelet activation.
Immunoprecipitation of FAK
Focal adhesion complexes connect membrane integrins to actin cytoskeleton and constitute a starting point for signaling in platelets. The nonreceptor tyrosine kinase FAK is one of the more representative proteins present in focal adhesion complexes prone to be tyrosine phosphorylated during platelet activation (34
, 35)
. We therefore tested the ability of the octapeptide to induce formation of focal adhesion complexes during platelet adhesion. As observed in Fig. 6G
, FAK was tyrosine phosphorylated in platelets adherent to immobilized octapeptide or type III collagen. This phosphorylation was also observed in the presence of RGDS (not shown).
Cooperation between GP VI and TIIICBP during GP VI-induced platelet aggregation
Although an association of GP VI with TIIICBP was not observed during immunoprecipitation experiments, we have studied the effects of the 7F4 monoclonal antibody or the octapeptide on platelet aggregation induced by two agonists of GP VI: convulxin and CRP. Pretreatment of platelets with 7F4 (50 µg/mL) or the octapeptide (0.5 mM) decreased by respectively 36% and 45% platelet aggregation induced by 150 pM convulxin (Fig. 7
A) and by respectively 44% and 32% platelet aggregation induced by 75 ng CRP (Fig. 7B
). These effects were specific since type III collagen-mediated platelet aggregation was completely inhibited by 7F4 or the octapeptide as described previously (12
, 14)
, whereas thrombin-mediated platelet aggregation, independent of TIIICBP, was not affected (not shown).
|
| DISCUSSION |
|---|
|
|
|---|
We first investigated platelet deposition to immobilized octapeptide in static conditions. When performed in the absence of aggregation inhibitor, the model used is a mixture of platelet contact, adhesion, and aggregation. Whereas platelet deposition mediated by the immobilized octapeptide occurred rapidly, deposition to type III collagen presented a lag phase of
10 min, probably corresponding to the time needed for primary interactions mediated in part by octapeptide sequences present in lesser number within the collagen molecule. At 60 min, the tyrosine phosphorylation pattern induced by platelet deposition to the immobilized octapeptide was similar to the one obtained with type III collagen, suggesting activation of similar signaling pathways. When aggregation was blocked by RGDS, a slight increase in the whole phosphorylation was observed, possibly due to inhibition of the GP IIb/IIIa-dependent dephosphorylation step, as described recently by Francischetti et al. (36)
. This dephosphorylation step was shown to be dependent of platelet aggregation and inhibited by RGDS (37)
.
To characterize platelet adhesion to the immobilized octapeptide and the associated signaling events, subsequent experiments were performed in the presence of RGDS. Our results demonstrate that platelet adhesion to immobilized octapeptide is independent of extracellular Ca2+ and Mg2+, providing evidence that GP Ia/IIa, which requires divalent cations to function (2)
, does not contribute to platelet adhesion to the immobilized octapeptide and that TIIICBP does not belong to the integrin family. This confirms our previous results on the failure of anti-GP Ia/IIa antibody (6F1) to inhibit platelet adhesion to the immobilized octapeptide (12)
.
The immobilized octapeptide contains the three amino acids that could form GPO motifs able to react with GP VI. However, using the 9O12.2 antibody against GP VI (31)
, we demonstrated that platelet adhesion to the octapeptide is independent of GP VI; thus, it is unlikely that immobilized octapeptide forms GPO motifs in the right conformation recognized by GP VI.
Nakamura et al. (38)
have proposed a direct link between GP VI and generation of TXA2 during platelet-collagen adhesive interactions and the necessity of the collagen polymeric fibrillar structure to induce generation of TXA2. Using indomethacin, an inhibitor of TXA2 generation, we show that the TXA2-dependent pathway plays a minor role in GP IIb/IIIa-independent platelet adhesion to type III collagen and is almost negligible for adhesion to the octapeptide. This is consistent with the report from Nakamura et al. (39)
, who proposed that, during platelet adhesion to collagen, generated TXA2 was involved in the activation of GP IIb/IIIa and had a minimal effect on platelet adhesion and spreading.
Using inhibitors of the P2Y1 and P2Y12 ADP receptors (40)
, we observed that released ADP was a potent cofactor of platelet adhesion to type III collagen. In contrast, platelet adhesion to the immobilized octapeptide was less sensitive to released ADP, as it was dependent only on the P2Y12 ADP receptor. Because the contact phase is an initial event in platelet-collagen interaction involving nonactivated platelets (8)
, it is likely that platelet adhesion to the immobilized octapeptide is independent of secondary secreted agonists. Thereby, platelet interactions with octapeptide sequences could contribute to the ADP-independent residual adhesion to type III collagen. Moreover, the slight ADP-sensitive part of platelet adhesion to the immobilized octapeptide could be subsequent to the initial platelet contact mediated by TIIICBP and involve activation of a subset of additional TIIICBP or other receptors, leading to enhanced granule secretion and platelet recruitment. Until the recent revision of the two-step/two-site model by Nieswandt and Watson (4)
, GP Ia/IIa was considered the main adhesive receptor for collagen, with GP VI principally involved in platelet activation and signaling (41
, 42)
. This model now places GP VI in a central position in platelet-collagen interactions. GP Ia/IIa binding to collagen requires prior activation by intracellular signals generated by GP VI and is reinforced by released ADP and TXA2. Moreover, cooperation between GP VI and GP Ia/IIa is necessary to obtain the full platelet response to collagen in vitro (43)
and for collagen-induced thrombus formation (44)
. Nevertheless, our results raise the question of a GP VI requirement in initial events of platelet adhesion to type III collagen since a type III collagen-related sequence, the KOGEOGPK octapeptide, which is involved in the first steps (e.g., platelet contact and adhesion) of platelet interaction with collagen (12
, 14)
, is able to promote platelet adhesion independently of GP VI.
In contrast to the whole collagen molecule, the octapeptide is unable to promote platelet aggregation when presented in solution, but instead must be immobilized on a plastic surface to induce platelet activation and signaling, probably due to its monovalent structure. Using type II collagen-related CNBr peptides, Guidetti et al. (45)
have proposed that the inability of CNBr peptides to induce platelet aggregation is due to their inability to form fibrils. Therefore, once immobilized, the octapeptide could promote multiple interactions and possibly oligomerization of TIIICBP, necessary to trigger platelet activation and signaling. Our present results demonstrate that, like the whole type III collagen molecule, the octapeptide sequence when immobilized triggers platelet activation and strong GP IIb/IIIa-independent tyrosine phosphorylation of Syk and PLC
2. Tyrosine phosphorylation of the FcR
-chain and the transmembrane adaptor LAT were observed but to a lesser extent than those induced by type III collagen, suggesting the likely contribution of GP VI for full phosphorylation of the FcR
-chain and LAT. In addition, this study demonstrates that TIIICBP associates with the FcR
-chain and LAT independent of platelet activation. Although GP VI-independent phosphorylation of the FcR
-chain and LAT has already been reported in platelets activated by vWF/ristocetin, vWF/botrocetin (46)
or by peptides related to the carboxyl-terminal domain of thrombospondin-1 (47
, 48)
, a cooperation between TIIICBP and GP VI could be considered in order to elicit full tyrosine phosphorylation of the FcR
-chain and LAT during platelet-collagen adhesive interactions. This potential cooperation between TIIICBP and GP VI is supported by the fact that TIIICBP participates in GP VI-induced platelet aggregation. This involvement of TIIICBP seems to be specific since it was observed with two different GP VI agonists, convulxin and CRP, and two different TIIICBP inhibitors, 7F4 and soluble octapeptide. A dual specificity of convulxin for another platelet receptor, native GP Ib, has been demonstrated in the recent study by Kanaji et al. (49)
. This raises the question of a potential indirect contribution of GP Ib in the inhibitory effect of TIIICBP inhibitors on GP VI-induced platelet aggregation. However, since convulxin- and CRP-mediated platelet aggregations are decreased to a similar extent when blocking TIIICBP, an indirect contribution of GP Ib seems unlikely.
Platelet adhesion to the octapeptide also induces phosphorylation of TIIICBP. It remains to be determined whether phosphorylation of the TIIICBP receptor immediately follows its binding to the octapeptide sequence or results from subsequent events such as association of TIIICBP with other platelet receptors like GP VI. Once tyrosine phosphorylated, TIIICBP, in association with the FcR
-chain and/or LAT, could represent a docking site for other signaling proteins and for GP VI leading to cooperation between these two collagen receptors for subsequent tyrosine phosphorylation of the FcR
-chain, LAT, and downstream signaling proteins leading to full platelet activation. How this cooperation between GP VI and TIIICBP is carried out and the respective contribution of each receptor in signaling induced by platelet adhesion to type III collagen remains to be determined. Since TIIICBP is constitutively associated with the FcR
-chain, we postulate that the FcR
-chain could be the physical link between TIIICBP and GP VI during type III collagen-mediated platelet activation, as already demonstrated for GP Ib
and GP VI during activation by alboaggregin A, which binds to GP Ib
and GP VI (50)
. After receptor stimulation, GP VI is recruited in specialized areas of the platelet membrane called "lipid rafts" in an FcR
-chain-dependent manner (51)
. Inhibition of TIIICBP by 7F4 or soluble octapeptide could partially block this specific GP VI recruitment, resulting in a decrease of platelet aggregation.
Immunoprecipitation experiments have revealed the GP IIb/IIIa-independent tyrosine phosphorylation of FAK during octapeptide-induced platelet adhesion, suggesting the formation of focal adhesion complexes and a link between TIIICBP and the platelet cytoskeleton. However, a direct association between TIIICBP and proteins of focal adhesion complexes remains to be determined.
In conclusion, the present study demonstrates for the first time a critical role for the type III collagen-related primary binding sequence, the KOGEOGPK octapeptide, in mediating platelet adhesion independent of GP Ia/IIa and GP VI. Platelet adhesion to the immobilized octapeptide is able to trigger a signaling pathway similar to the one induced by type III collagen, as attested by tyrosine phosphorylation of key signaling proteins such as Syk, the FcR
-chain, LAT, PLC
2, FAK, and TIIICBP itself. We show that TIIICBP participates in GP VI-mediated platelet aggregation and postulate for a potential cooperation between TIIICBP and GP VI with the FcR
-chain as the physical link in the platelet membrane. Although the FcR
-chain and LAT are constitutively associated with TIIICBP, a physical interplay between TIIICBP and GP VI remains to be clarified. Platelet primary interactions with the octapeptide should then be considered for future research dealing with platelet-type III collagen adhesive interactions and thrombus formation in vivo. Interaction of platelets with the octapeptide may also explain why GP Ia/IIa or GP VI deficiency in mice has a relatively minor effect on bleeding time (52
, 53)
.
| ACKNOWLEDGMENTS |
|---|
Received for publication October 29, 2003. Accepted for publication April 27, 2004.
| REFERENCES |
|---|
|
|
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |