The roles of α_IIbβ₃-mediated outside-in signal transduction, thromboxane A2, and adenosine diphosphate in collagen-induced platelet aggregation

Normal hemostasis depends in part on platelet signal transduction initiated by the adherence of platelets to collagen fibers exposed at sites of blood vessel injury. Because of the central role of collagen/platelet interactions in hemostasis, collagen-induced platelet aggregation has been studied extensively. Studies using platelets from patients with collagen-related bleeding disorders, along with characterization of platelet signaling and aggregation in response to a variety of snake venom proteins and the analysis of platelets from knockout mice, have provided much of the foundation for our understanding of collagen-induced platelet signaling.1,2Together, these studies indicate that the integrin α2β1 is the major receptor for platelet adhesion to collagen and that glycoprotein VI (GPVI) is the primary, but not exclusive, signaling receptor for collagen-induced platelet signal transduction.2,3 However, the role of α2β1 in collagen-induced signal transduction is controversial and unresolved.4-6 Nonetheless, it is clear that α2β1 contributes to this signaling but is unable to generate all the signaling required to induce aggregation.5 While collagen-related peptide (CRP) specifically induces GPVI signaling, collagen has the advantage of being a physiologic agonist for studying platelet activation. Collagen activation of platelets induces tyrosine phosphorylation of the signaling molecules Fc receptor γ-chain (FcRγ-chain), Syk, phospholipase C γ2 (PLC γ2), and the adapter molecule SLP-76. These signaling molecules have been reported to be required for collagen-induced platelet aggregation and dense-body secretion.7-10 In addition, the adapter molecule LAT (linker for activation of T cells) is tyrosine phosphorylated in response to collagen treatment of platelets11 and is required for irreversible platelet aggregation in response to a low, but not a high, level of collagen.12,13 Furthermore, LAT is required for thromboxane A2 (TxA2) production in response to all levels of collagen.12 

Despite the fact that a great deal is known about the participation of the above-mentioned signal transduction and adapter molecules in collagen-induced platelet signal transduction, the roles of TxA2 production, adenosine diphosphate (ADP) secretion, and α_IIbβ₃-mediated outside-in signaling in collagen-induced platelet activation have not been clarified. For example, a variety of studies support the view that TxA2 or ADP is absolutely required for collagen-induced human and murine platelet aggregation,14-16 but platelets from mice lacking TxA2 receptors17 or the ADP receptors P2Y118,19 or P2Y1220 aggregate in response to collagen. Additionally, the results of other studies suggest that platelet aggregation and therefore possibly the activation of α_IIbβ₃require collagen-induced secretion that is dependent on the FcRγ-chain,7 PLC γ2,10 and the function of Gαq.16 Although it is correct that activation of the FcRγ-chain is required for collagen-induced platelet aggregation and secretion, the results presented here demonstrate that platelet aggregation induced by a high level of collagen is not dependent on TxA2 and dense-granule secretion. The data presented here resolve the contradictions presented by the results cited above14-20and in so doing demonstrate that collagen activation of α_IIbβ₃ is not dependent on PLCγ2, Gαq, the thromboxane A2 receptor, or dense-body secretion. These results provide new insights into the signaling mechanisms underlying collagen-induced platelet aggregation and demonstrate that there are at least 2 distinct signaling pathways activated by collagen stimulation of platelets.

Collagen reagent Horm (native collagen fibrils from equine tendons) was purchased from Nycomed Arzneimittel (Munchen, Germany). Tp^−/− mice were derived as described.17Hamster control immunoglobulin G (IgG) was from Jackson ImmunoResearch Laboratories (West Grove, PA). 1B5 (hamster antimouse α_IIbβ₃ antibody) was a generous gift from Dr Barry Coller (Rockefeller University, NY). Apyrase, A3P5P, Ro31-8220, and mepacrine were purchased from Sigma-Aldrich (St Louis, MO), and AR-C69931MX was a generous gift from Astra-Zeneca (Loughborough, England).

Blood was collected from the abdominal aorta of isofluorane-anesthetized mice into syringes containing 2% (vol/vol) heparin. Platelet-rich plasma (PRP) was prepared by differential centrifugation. Aggregation tests were performed in a Chronolog aggregometer (CHRONO-LOG, Havertown, PA) using PRP (300 μL) adjusted to approximately 10⁶ platelets/μL with plasma. In cases where antibody was used, washed platelets were prepared by differential centrifugation and resuspension into Tyrode buffer (12 mM NaHCO3, 138 mM NaCl, 5.5 mM glucose, 2.9 mM KCl, 10 mM HEPES [N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid], 0.1% bovine serum albumen [BSA]), pH 7.4). Washed platelets were incubated with an antibody for 10 minutes at room temperature prior to stimulation by collagen.

Adenosine trisphosphate (ATP) secretion was measured using CHRONO-LUME reagent (CHRONO-LOG) according to the manufacturer's protocol, with minor variations. After 4 to 6 minutes of platelet activation with stirring (1200 rpm), 10 μL luciferase-luciferin was added directly into the cuvettes. The luminescence intensity was measured at a luminescence setting of × 0.001.

TxA2 was assayed as TxB2 in the plasma after the platelets were removed at the end of the 4- to 6-minute aggregation period. The TxB2 EIA kit (Assay Designs, Ann Arbor, MI) was used according to the manufacturer's protocol to indirectly measure free TxA2. The plasma samples were diluted 1:50 with the supplied assay buffer for analysis. For the purpose of discussion, it is assumed that TxB2 levels reflect TxA2 levels.

Mepacrine is taken up into dense granules in platelets. The procedure for mepacrine uptake was followed as previously described.21 Washed platelets were incubated in modified Tyrode solution containing 10 μM mepacrine (quinacrine; Sigma Chemical, St Louis, MO) for 30 minutes at 37°C. After incubation, the platelets were removed from excess mepacrine by centrifugation and suspension of the platelets in modified Tyrode solution containing 5 mM EDTA (ethylenediaminetetraacetic acid). Resting platelets and platelets activated with 50 μg/mL collagen were stirred in the aggregometer for 4 minutes. After stirring, the platelets were washed twice and suspended in modified Tyrode solution containing 5 mM EDTA. Secretion of dense-body constituents was evaluated by measuring the fluorescence remaining in the platelets after stirring. For each measurement, 200 μL platelets were diluted 15-fold with modified Tyrode solution. Fluorescence intensity was measured in a Varian Cary Eclipse Fluorescence spectrophotometer (Walnut Creek, CA). The excitation wavelength for mepacrine is 488 nm; the emission wavelength selected was 530 nm. The fluorescence remaining in the control platelets was set as 100%. The fluorescence remaining in the platelets was assumed to represent unsecreted mepacrine. The loss of fluorescence by the collagen-treated platelets represents collagen-induced mepacrine secretion in the absence of platelet aggregation.

The studies that demonstrated the requirement for ADP signaling through P2Y1 and P2Y12 and TxA2 receptor signaling in collagen-induced platelet aggregation and secretion did not investigate these effects in response to high concentrations of collagen. Here, the relationship between secreted ADP, TxA2, collagen-induced α_IIbβ₃ activation, and platelet aggregation was investigated by treatment of thromboxane A2 receptor (Tp)–deficient platelets with a combination of apyrase, A3P5P, and AR-C69931MX to completely inhibit all signaling through TxA2 and ADP receptors in response to collagen. This study was done in 2 parts. Tp-deficient mouse platelets were stimulated with 2.5 or 50 μg/mL of collagen in the presence or absence of the combination of the inhibitors apyrase, A3P5P, and AR-C69931MX (the latter are antagonists for the ADP receptors P2Y1 and P2Y12, respectively). In the first part, platelet activation was characterized in response to a low level of collagen. In contrast to the wild-type platelets (Figure1A), Tp-deficient platelets aggregated less extensively and reversibly in response to a low level of collagen in the absence of inhibitors of ADP receptor signaling, indicating that TxA2 receptor signaling is required to cause irreversible aggregation in response to a low level of collagen. Tp-deficient platelets secreted about 66% less ADP and produced about 50% less TxA2 (measured as TxB2) than the wild-type platelets (Figure 1B) at a low collagen concentration without inhibitors. The diminished level of secretion and TxA2 production may be due to the diminished aggregation of Tp-deficient platelets, the lack of TxA2 receptor signaling, or both. Treatment with apyrase and the ADP receptor antagonists totally inhibited aggregation in response to a low level of collagen in both Tp-deficient and wild-type platelets (Figure 1C), indicating that ADP-induced signaling is absolutely required to cause aggregation in response to a low level of collagen. No ADP secretion was detected because of the treatment with apyrase, an ADP scavenger, but TxA2 production was measured. Tp-deficient and wild-type platelets in the presence of the aggregation inhibitors produced similar levels of TxA2 (Figure 1D), but the level of TxA2 production was less than the levels produced in the absence of inhibitors (Figure 1B,D). The results obtained (see Figure 1) in the presence of the ADP receptor antagonists and apyrase were no different from those obtained in the presence of the antagonists without apyrase (data not shown), indicating that any adenosine monophosphate (AMP) generated by apyrase had no significant effect on the results.

In the second part of this study, Tp-deficient platelets were treated with a high level of collagen in the presence and absence of the inhibitors. When Tp-deficient platelets were activated with a high level of collagen, they aggregated (Figure2A), secreted ADP, and produced TxA2 to similar extents as wild-type platelets (Figure 2B), indicating that TxA2 receptor signaling is not necessary for aggregation in response to a high level of collagen. Furthermore, wild-type platelets that were treated with apyrase, A3P5P, and AR-C69931MX aggregated to a similar level as platelets that were untreated (Figure 2A,C). In contrast, Tp-deficient platelets treated with apyrase, A3P5P, and AR-C69931MX aggregated slowly and only about 38% as extensively as the wild-type platelets in response to a high level of collagen (Figure 2C), even though the level of TxA2 production was similar to that of the wild-type platelets treated with apyrase and the ADP receptor antagonists (Figure 2D). So, the lack of both ADP and TxA2 receptor signaling results in diminished aggregation, and, in contrast to the situation in response to a low level of collagen, either ADP or TxA2 suffices to cause maximum aggregation in response to a high level of collagen. Moreover, α_IIbβ₃ activation in response to a high level of collagen does not require ADP or TxA2, because platelets that lack the TxA2 receptor and P2Y1- and P2Y12-propagated signaling can undergo a low level of aggregation in response to a high concentration of collagen. That apparent aggregation was not simply collagen-mediated clumping or α2β1-mediated adhesion because it was inhibited by the antimurine α_IIbβ₃ monoclonal antibody (mAb) 1B5 (Figure 2E), which inhibits aggregation of mouse platelets by preventing the binding of fibrinogen to its receptors.22 

Although Tp-deficient platelets treated with apyrase, A3P5P, and AR-C69931MX underwent a low level of aggregation in response to a high level of collagen, those results do not necessarily mean that collagen can induce platelet aggregation in the absence of dense-granule secretion. This issue was addressed by treating the platelets with Ro31-8220, an inhibitor of protein kinase C (PKC) and secretion. Tp-deficient platelets in plasma were treated with Ro31-8220 and tested for aggregation, ATP secretion, and TxA2 production in response to 50 μg/mL collagen. Tp-deficient platelets aggregated slowly (Figure3A) and only to an extent similar to that of the Tp-deficient platelets treated with apyrase and ADP receptor antagonists (Figure 2C). As expected, Tp-deficient platelets treated with 3 μg/mL collagen in the presence of Ro31-8220 did not aggregate (data not shown). Furthermore, even though treatment with Ro31-8220 prevented ATP secretion (Figure 3C), the Tp-deficient platelets produced an undiminished level of TxA2 (Figure 3D). These results indicate that collagen-induced signaling can cause aggregation, α_IIbβ₃ activation, and TxA2 production in the absence of demonstrable dense-granule secretion and sufficient PKC activation to cause ADP secretion. The aggregation caused by a high level of collagen in the presence of Ro31-8220 reflects activation of α_IIbβ₃ because that aggregation was inhibited by the mAb 1B5, but not nonimmune IgG (Figure 3B).

As shown in Figure 2, TxA2 and dense-body secretion are not required to induce a low level of platelet aggregation in response to a high level of collagen. Contrary to the published data, these results support the view that platelets that lack PLCγ210 or Gαq16 would be expected to aggregate to a high, but not a low, level of collagen, assuming that the reported absence of aggregation is due to the lack of secretion and only if PLCγ2 or Gαq are not required to directly activate α_IIbβ₃. These predictions were tested by characterizing the aggregation response of platelets from both PLCγ2- and Gαq-deficient mice treated with a high concentration of collagen. As predicted, PLCγ2-deficient platelets aggregated in response to a high level of collagen (Figure 4A), although the aggregation was slow and the level was diminished in comparison with the wild-type platelets. This low level of aggregation was α_IIbβ₃-dependent aggregation because it was inhibited by mAb 1B5, but not IgG (Figure 4B). As expected, the extent of aggregation of the PLCγ2-deficient platelets was similar to that of the Tp-deficient platelets treated with Ro31-8220 (Figure 3A). Furthermore, the presence of apyrase, A3P5P, AR-C69931MX, and indomethacin did not decrease the extent or rate of the aggregation of PLCγ2-deficient platelets in response to a high level of collagen (Figure 4C), confirming that the aggregation was ADP and TxA2 independent. Likewise, treatment of wild-type platelets with the inhibitors rendered their aggregation in response to a high level of collagen similar to that of the PLCγ2-deficient platelets with or without inhibitors (Figure 4D). Therefore, the diminished aggregation of the PLCγ2-deficient platelets demonstrates that secretion and TxA2 are not absolutely required for a high level of collagen to cause α_IIbβ₃ activation and aggregation, although maximum aggregation requires secretion and/or TxA2.

In contrast to the PLCγ2-deficient platelets, Gαq-deficient platelets released normal levels of ADP (Figure5B) and TxA2 (Figure 5C) and aggregated to an extent similar to that of the control platelets from the littermate mice (Figure 5A) in response to a high level of collagen. Therefore, Gαq signaling is not required for aggregation induced by a high level of collagen. Thus, collagen-induced platelet aggregation is dependent on PLCγ2 and Gαq only in response to a low level of collagen. Specifically, collagen-induced platelet aggregation appears to be dependent on PLCγ2 and Gαq only in response to levels of collagen that cause secretion-dependent aggregation. Nonetheless, even though TxA2 and secretion are absolutely required only for irreversible aggregation in response to a low level of collagen, they enhance collagen-induced aggregation regardless of the level of collagen used to stimulate the platelets. Furthermore, PLCγ2- or Gαq-elicited signaling are not required for collagen-induced α_IIbβ₃ activation, because aggregation induced by a high level of collagen is TxA2 and secretion independent.

The observations that Tp-deficient platelets produced the same amount of TxA2 and secreted a similar amount of ADP as the control platelets in response to a high concentration of collagen but produced less TxA2 and secreted less ADP than the control platelets in response to a low level of collagen (Figure 3) support the hypothesis that TxA2 production and possibly ADP secretion are driven, at least in part, by aggregation in response to a low level of collagen. This hypothesis was tested by treating control platelets with the antimouse α_IIbβ₃ mAb 1B5 (Figure6A), which inhibits platelet aggregation by preventing fibrinogen binding.22 Platelets treated with a low level of collagen in the presence of the mAb 1B5 secreted and produced about 80% less ADP (Figure 6B) and TxA2 (Figure 6C), respectively, than platelets treated with control IgG. Thus under certain conditions, α_IIbβ₃-mediated outside-in signal transduction plays an important role in collagen-induced secretion and platelet aggregation.

Although α_IIbβ₃ outside-in signaling is required for maximum TxA2 production and ADP secretion in response to a low level of collagen, the following results demonstrate that this signaling is not absolutely required for those functions in response to high levels of collagen. Because 1B5 could not completely prevent platelet aggregation to a high level of collagen (50 μg/mL), EDTA was used to prevent aggregation under these conditions (Figure7A). Platelets treated with 50 μg/mL collagen in the presence of EDTA produced about 30% of the amount of TxA2 produced in its absence (Figure 7B). The cause of the diminished production of TxA2 is not known; but regardless of the cause, 50 μg/mL of collagen caused a substantial level of TxA2 production in the absence of platelet aggregation. The significance of these results is that TxA2 production occurred in the absence of aggregation (despite the ability of EDTA to interfere with the binding of collagen to α2β1). Dense-body secretion was measured using mepacrine21 because EDTA precludes using CHRONO-LUME to measure ATP secretion. As with TxA2 production, a high level of collagen appeared to have caused secretion of about 55% of the dense granule mepacrine (Figure 7C), which is inhibited by Ro31-8220, in the absence of aggregation (not shown). These results support the conclusion that treatment of platelets with high levels of collagen can cause a significant level of TxA2 production and ADP secretion in the absence of α_IIbβ₃-mediated outside-in signaling. Thus, collagen treatment of platelets can result in both aggregation-dependent and aggregation-independent TxA2 production and ADP secretion, demonstrating the occurrence of 2 distinct signaling mechanisms in response to collagen.

The investigation of the cross-talk between receptors and signaling molecules in response to collagen activation of platelets described here has led to the conclusion that ADP and TxA2 signaling are not required for activation of α_IIbβ₃and aggregation in response to a high level of collagen. Contrary to the published data using platelets from mice deficient in PLCγ210 or Gαq,16 our results demonstrate that PLCγ2- or Gαq-deficient platelets aggregate to a high level of collagen, indicating that neither PLCγ2 nor Gαq are required for α_IIbβ₃ activation and aggregation under those conditions. The diminished level of PLCγ2-deficient platelet aggregation may be explained by the fact that no secretion or TxA2 production occurred, supporting the idea that there is cross-talk between collagen receptors with receptors of secreted products. Thus, α_IIbβ₃ activation induced by a level of collagen that causes secretion-independent aggregation is not dependent on PLCγ2, Gαq, dense-granule secretion, and the TxA2 receptor.

These results confirm the view that there are at least 2 distinct signaling pathways that can mediate platelet activation in response to collagen. Our data do not identify the mechanism underlying the selection between these pathways, but do demonstrate that selection between these pathways appears to be a function of the concentration of collagen used to activate the platelets. Interestingly, the pathway utilized to cause secretion-dependent aggregation (low levels of collagen) uses 2 phospholipases to cause discernible aggregation. In contrast, the signaling pathway activated by exposure to a high level of collagen that causes secretion-independent aggregation is not dependent on phospholipase activity. Secretion/Gαq-dependent aggregation induced by a low concentration of collagen is thought to use PLCβ to activate platelets.16 However, as reported elsewhere,10 aggregation induced by a low level of collagen is also dependent on PLCγ2.

The requirement for 2 distinct phospholipases that presumably have the same enzymatic function for platelet activation in response to collagen requires explanation. The relationship between the 2 phospholipases is revealed by the fact that PLCγ2-deficient platelets do not produce TxA2 or release ATP in response to any level of collagen (Figure 4), whereas the Gαq-deficient platelets release TxA2 and ATP in response to a high level, but not a low level, of collagen (Figure 5). Thus, the function of Gαq and consequently PLCβ in response to collagen is subordinate to PLCγ2. The significance of this relationship appears to be that Gαq amplifies the signal initiated by PLCγ2. So, the data presented here and elsewhere10 support the view that PLCγ2 function is required for activation of Gαq, and Gαq function is required for signal amplification to drive extensive platelet aggregation in response to a low level of collagen.16 Presumably, activation of PLCγ2 by a low level of collagen results in TxA2 and ADP release, which in turn activates Gαq via their receptors. Then Gαq activates PLCβ, which causes the release of TxA2 and ADP by platelets, thereby causing signal amplification and the recruitment of more platelets into growing aggregates.

The signaling pathway used by platelets treated with a high level of collagen causes secretion-independent aggregation. This pathway requires the function of GPVI/FcR γ-chain, but not LAT,12,13 PLCγ2 (Figure 4), SLP-76,13 or Gαq (Figure 5). Since platelets deficient in Fyn/Lyn,23Gαq (Figure 5), or LAT12 secrete dense granules in response to high concentrations of collagen, it is not known whether those signaling molecules are required for secretion-independent aggregation, because platelets from those knockout mice have not been tested in response to a high level of collagen in the presence of the required combination of inhibitors. However, platelets can activate α_IIbβ₃ in the absence of Fyn/Lyn, Gαq, or LAT, indicating that collagen-stimulated platelets can cause Fyn/Lyn-, Gαq-, or LAT-independent α_IIbβ₃activation. Furthermore, the absolute requirement of PLCγ2 for collagen-induced secretion and TxA2 production, but not for aggregation (Figure 4), demonstrates that PLCγ2 is not required for collagen-induced activation of α_IIbβ₃ and subsequent secretion-independent aggregation. It is also known that SLP-76 is not required for aggregation induced by a high level of collagen.13 The observations of others rule out the requirement for α-granule or lysosomal secretion for activation of α_IIbβ₃ in response to collagen, since it is well known that GPVI/FcRγ-chain deficient platelets do not secrete P-selectin or CD63 (granulophysin) in response to collagen but do activate α_IIbβ₃.5,24 The signaling pathway that is utilized in response to a low level of collagen requires FcR γ-chain, P2Y1 and P2Y12 (Figure 1), Gαq,16 PLC γ2,10 SLP-76,9and PKC (Figure 3) for α_IIbβ₃ activation in response to low concentrations of collagen. The function of these proteins is that they facilitate TxA2 production and dense-granule secretion, since the lack of or inhibition of these proteins substantially decreases TxA2 production and secretion.

The results presented here show that aggregation is required to cause maximal TxA2 production and dense-granule secretion (Figure 6), and that TxA2 and secretion seem to be required for maximal aggregation (Figure 2). Previous reports also provide evidence that coordinated signaling events between α_IIbβ₃-mediated outside-in signaling and P2Y1 and P2Y12-mediated signaling are essential for TxA2 production.25 Furthermore, cross-talk between GPVI and Gi-coupled receptors during collagen-induced platelet aggregation has been established.26 Thus, it appears that aggregation drives TxA2 production and secretion and the released agonists can induce aggregation, making it seem likely that these events participate in positive feedback induced by collagen throughout the aggregation process.

We thank Dr Stefan Offermanns for providing us with the Gαq knockout mice and Dr Barry Coller for providing us with the anti–mouse α_IIbβ₃ antibody 1B5.