ADP receptor P2Y12 is the capstone of the cross-talk between Ca²⁺ mobilization pathways dependent on Ca²⁺ ATPases sarcoplasmic/endoplasmic reticulum type 3 and type 2b in platelets

Abstract

Background

Blood platelet Ca²⁺ stores are regulated by 2 Ca²⁺-ATPases (SERCA2b and SERCA3). On thrombin stimulation, nicotinic acid adenosine dinucleotide phosphate mobilizes SERCA3-dependent stores, inducing early adenosine 5‘-diphosphate (ADP) secretion, potentiating later SERCA2b-dependent secretion.

Objectives

The aim of this study was to identify which ADP P2 purinergic receptor (P2Y1 and/or P2Y12) is(are) involved in the amplification of platelet secretion dependent on the SERCA3-dependent Ca²⁺ mobilization pathway (SERCA3 stores mobilization) as triggered by low concentration of thrombin.

Methods

The study used the pharmacologic antagonists MRS2719 and AR-C69931MX, of the P2Y1 and P2Y12, respectively, as well as Serca3^-/- mice and mice exhibiting platelet lineage-specific inactivation of the P2Y1 or P2Y12 genes.

Results

We found that in mouse platelets, pharmacological blockade or gene inactivation of P2Y12 but not of P2Y1 led to a marked inhibition of ADP secretion after platelet stimulation with low concentration of thrombin. Likewise, in human platelets, pharmacological inhibition of P2Y12 but not of P2Y1 alters amplification of thrombin-elicited secretion through SERCA2b stores mobilization. Finally, we show that early SERCA3 stores secretion of ADP is a dense granule secretion, based on parallel adenosine triphosphate and serotonin early secretion. Furthermore, early secretion involves a single granule, based on the amount of adenosine triphosphate released.

Conclusion

Altogether, these results show that at low concentrations of thrombin, SERCA3- and SERCA2b-dependent Ca²⁺ mobilization pathways cross-talk via ADP and activation of the P2Y12, and not the P2Y1 ADP receptor. The relevance in hemostasis of the coupling of the SERCA3 and the SERCA2b pathways is reviewed.

Essentials

• •Nicotinic-acid-adenine-dinucleotide-phosphate mobilizes SERCA3-ATPase Ca²⁺ stores initiating early ADP secretion in activated platelets.
• •The roles of the P2Y1 and P2Y12 ADP receptors were studied.
• •SERCA3-dependent ADP early release activates P2Y12 (and not P2Y1) and SERCA2b-dependent release.
• •Full platelet secretion relies on coupling of SERCA3 and SERCA2b stores mobilization via P2Y12.

1. Introduction

Calcium (Ca²⁺) is central in cellular signaling and requires fine-tuned regulation to control its cytosolic concentration and to shape signals [[1], [2], [3]]. On stimulation, Ca²⁺ is imported from the extracellular medium (influx) or mobilized from internal stores (mobilization), leading to a cytosolic Ca²⁺ level increase. This increase may be global or localized and is counterbalanced by either export through the plasma membrane or recapture into internal stores or mitochondria. The major actors in downregulation of cytosolic Ca²⁺ are energy-dependent membrane ATPases, which are able to drive Ca²⁺ export against a concentration gradient from 100 nmol/L in the cytosol to the mmol/L range in the extracellular milieu or in internal stores (mainly, the endoplasmic reticulum [ER]). Plasma membrane calcium ATPases expel Ca²⁺ in the extracellular milieu, whereas sarco-endoplasmic reticulum calcium ATPases (SERCAs) recapture cytosolic Ca²⁺ into internal stores. In platelets, mobilization from internal stores has been shown to regulate such primordial functions as adhesion and aggregation [[4], [5], [6]]. The SERCA family is encoded by 3 genes (ATP2A1, ATP2A2, and ATP2A3), which produce each several alternate transcripts and protein isoforms: SERCA1a/b, SERCA2a-c, and SERCA3a-f. They are found in multiple tissues, but platelets exhibit only sarco-endoplasmic reticulum calcium ATPase type 2b (SERCA2b) and sarco-endoplasmic reticulum calcium ATPase type 3 (SERCA3) isoforms [[7], [8], [9], [10]]. SERCAs maintain a Ca²⁺ concentration gradient between the cytosol (100 nmol/L) and the ER (1 mmol/L) [11]. However, SERCAs redundancy may not be explained only by Ca²⁺ gradient maintenance but also by functional differences. Although SERCA enzymes exhibit comparable structures, they possess also different activities including: Ca²⁺ affinity higher for SERCA2b (K_1/2 approximately 0.27 μmol/Lol/L) than for SERCA3 (1 μmol/Lol/L), but Ca²⁺ uptake lower for SERCA2b (7 nmol/min/mg) than SERCA3 (21 nmol/min/mg) [12,13]. In addition, in platelets, subcellular localization studies by immuno-electron microscopy have suggested a differential topology for SERCAs, central for SERCA2b and peripheral for SERCA3 [14,15]. Functional studies have pointed to involvement of SERCA3 in store operated Ca²⁺ entry (SOCE) via association with STIM1, a Ca²⁺ sensor participating in Ca²⁺ stores replenishing [16], as well as with SERCA3 functional association with acidic Ca²⁺ stores [17]. Recently, we have provided evidence that the SERCA3-dependent Ca²⁺ mobilization pathway impacts the physiology of hemostasis as well as its pathophysiology: (1) SERCA3^-/- mice exhibit prolonged tail-clip bleeding time, mainly associated with rebleeding, showing that the SERCA3-dependent Ca²⁺ pathway is involved in hemostasis; (2) thrombosis of mesenteric vessel assay (FeCl3) induces unstable thrombi with emboli, indicating the SERCA3 pathway is involved in thrombus stabilization [2]; these effects have been correlated to the reduced activation of SERCA3^-/- platelets stimulated with low concentration agonists such as thrombin, as shown by (3) diminished aggregation, consistent with a role for SERCA3-dependent Ca²⁺ mobilization in αIIbβ3 integrin-mediated platelet-platelet interaction; and (4) lower secretion [2,18]. We further established that an initial mobilization of Ca²⁺ from internal stores specifically dependent on the Ca²⁺ pump SERCA3 and not on Ca²⁺ stores dependent on SERCA2b was necessary to obtain full platelet activation in these conditions [2]. Moreover, we found that SERCA3 stores and SERCA2b stores were mobilized differentially because SERCA3 stores were specifically mobilized by the second messenger nicotinic-acid-adenine-dinucleotide-phosphate (NAADP), as opposed to SERCA2b stores, mobilized only by 1,4,5, inositol trisphosphate (IP3) [18]. In addition, SERCA3 store mobilization drove an initial ADP secretion responsible for costimulation with thrombin of SERCA2b store mobilization and secretion amplification, dependent on thromboxane A₂. We also demonstrated a similar mechanism in human platelets [19].

Here, we keep characterizing the molecular mechanism underlying platelet Ca²⁺ mobilization of SERCA2b and SERCA3 stores by providing evidence that in both mouse and human platelets, the P2-class purinergic receptor type Y subtype 12 (P2Y12), but not the P2-class purinergic receptor type Y subtype 1 (P2Y1) ADP receptor, is engaged by ADP secreted via mobilization of SERCA3 stores, driving secretion through the mobilization of SERCA2b stores.

2. Methods

2.1. Materials

Adenosine 5‘-diphosphate (ADP) was obtained from Chrono-log Corp. ARC69931MX tetrasodium salt, Apyrase from potatoes (grade VII), bovine thrombin, MRS2179 (MRS) ammonium salt hydrate, SERCA inhibitor, thapsigargin (Tg) are from Sigma. The d-Phe-Pro-Arg chloromethylketone dihydrochloride (PPACK) and prostaglandin E1 were obtained from Santa Cruz Biotechnology. BAPTA1-AM Oregon Green 488 was purchased from Molecular Probes. The NAADP antagonist, Ned-19, was obtained from Tocris Bioscience. The ATP determination kit was from Fisher Scientific; the ADP and the serotonin detection kits were from Abcam. PAR1-AP (SFLLRN) and PAR4-AP (AYPGKF) were from BACHEM.

2.2. Mice

Serca3^-/- mice were originally generated by Dr G.E. Shull (University of Cincinatti, OH) [20], and Swiss black Serca3^-/- mice were crossed with C57BL/6 mice and provided by Dr P. Gilon (University of Louvain, Belgium) with the authorization of Dr G.E. Shull. Transferred Serca3^+/- heterozygous mice were obtained with the help of Charles River Laboratories and were intercrossed. Wild-type (WT) and Serca3^-/- homozygote mice were detected by polymerase chain reaction, using published oligonucleotide primers [21]. Then, homozygote WT and Serca3^-/- mice were kept apart, and genotypes of breeding couples were verified every 6 months. C57BL/6-inbred platelet factor 4 (Pf4)-Cre⁺ mice [C57BL/6-Tg(Pf4-icre)Q3Rsko/J; The Jackson Laboratory] originally created by the Skoda's laboratory [21] were a generous gift from Dr M.-P. Gratacap (U1048 INSERM, Toulouse, France). These mice are transgenics harboring a Cre recombinase gene driven by the platelet-lineage-specific promoter PF4, expressing the recombinase only in the platelet-megakaryocyte lineage. They were used in this study as controls for PF4-P2Y1^-/- and PF4-P2Y12^-/- knockout mice. PF4-P2Y1^-/- and PF4-P2Y12^-/- knockout mice exhibiting deletion of the P2Y1 or the P2Y12 receptor exclusively in the megakaryocytic lineage were generated by crossing of recombinant mice with floxed alleles (P2Y1^flox/flox and P2Y12^flox/flox mice) with C57BL/6-inbred platelet factor 4 (Pf4)-Cre⁺ mice [C57BL/6-Tg(Pf4-icre)Q3Rsko/J; The Jackson Laboratory]. Genotyping was performed on mouse tail DNA by using a polymerase chain reaction amplification method (Hechler et al., in preparation).

All experimental procedures were carried out in accordance with the European legislation concerning the use of laboratory animals and approved by the Animal Care and Ethical Committee of Université Paris-Saclay (APAFIS#14745-20180418165532119 v3).

2.2.1. Blood collection and mouse platelet preparation

Blood was collected by cardiac puncture of anesthetized (intraperitoneal injection of ketamine [100 mg/kg] and xylazine [10 mg/kg]) mice and 160 μmol/L D-phenylalanyl-L-prolyl-L-arginine chloromethyl ketone (PPACK) and 10% (vol/vol) ACD-C buffer (124 mmol/L sodium citrate, 130 mmol/L citric acid, and 110 mmol/L dextrose, pH6.5) were added. Washed platelets were isolated by centrifugation as previously described [2].

2.2.2. Blood collection and human platelet preparation

In accordance with the declaration of Helsinki, all healthy voluntary donors (men and women) gave written informed consent before blood collection. Human blood was collected in 10% (vol/vol) acid citric-citrate-dextrose A (ACD-A) buffer. Platelet-rich plasma (PRP) was obtained by centrifugation (120 g for 15 minutes at 20°C). Washed platelets were prepared as previously described [2]. Briefly, platelets were isolated from PRP (10% [vol/vol] ACD-A, 100 mU/mL apyrase, and 1 μmol/Lol/L prostaglandin E1) by centrifugation (1200 g for 12 minutes at 20°C) and washed 2 times with wash buffer in the presence of apyrase (100 mU/mL) and prostaglandin E1 (1 μmol/Lol/L) to minimize platelet activation. Platelets were adjusted to 3 x 10⁸ platelets/mL in calcium-free Tyrode buffer.

2.2.3. Platelet aggregation

Platelet aggregation was carried out as previously described [2], with washed platelets stimulated with thrombin. Light transmission was measured through the stirred suspension of platelets (3 x 10⁸ platelets/mL) during 3 minutes by using a Chronolog aggregometer (Chrono-Log Corporation).

2.2.4. Platelet dense granule secretion

Dense granule secretion was quantified by the assessment of ATP release by using an ATP determination kit (Molecular Probes), as previously described [2]. All experiments were run with triplicate samples. Serotonin secretion was assessed in platelet supernatants with the serotonin Enzyme-linked Immunosorbent Assay kit (Abcam).

2.2.5. Assay of cytosolic calcium by flow cytometry

Mouse or human platelets (3 × 10⁸ platelets/mL) were washed as indicated, resuspended in Ca²⁺-free (EGTA 0.1 mmol/L) Tyrode buffer and incubated at room temperature for 30 to 45 minutes with 1 μmol/L BAPTA-Oregon green 488; platelets were then diluted to 6 × 10⁵ platelets/mL in Ca²⁺-free Tyrode buffer, and Ca²⁺ mobilization was analyzed by using an Accuri C6 flow cytometer (BD Biosciences). Flow cytometer was setup to low speed (10μL/min) to analyze approximately 100 platelets/s. Platelets were activated with 5 μL agonists by gentle stirring with the pipet tip. Basal fluorescence before activation was arbitrarily set as 1 and variation of fluorescence intensity ratio over basal fluorescence represented evolution of Ca²⁺ concentration. Ca²⁺ mobilization was quantitated by area under the curve for 1 minute after stimulation and presented as arbitrary units.

2.2.6. Statistical analysis and data presentation

Analyses of statistical significance were performed with GraphPad Prism software (version 7.04; GraphPad Inc). Normal data distribution for platelet secretions and Ca²⁺ mobilization in response to thrombin was observed in previous studies [2,16,22] and in the present study and was assumed so all along the study. Depending on the conditions, values were analyzed with 1-way, 2-way, or 3-way analysis of variance (anova) or mixed-effects analysis (for groups with different sizes) followed by Tukey pair wise test or Bonferroni multiple comparisons test, as indicated in figure legends, and presented P values have been adjusted for multiple tests correction within experiments and have not been adjusted across tests. The results are presented either as graphs (mainly dot plots with means and SEM) or as representative images or tracings that were selected to illustrate data presented in the corresponding graph.

3. Results

3.1. Pharmacological inhibition of the ADP receptor P2Y12, but not P2Y1, alters thrombin-stimulated SERCA2b-dependent Ca²⁺ mobilization and secretion

To better understand how ADP released by mobilization of SERCA3-dependent Ca²⁺ stores (“SERCA3 stores”) stimulates secretion driven by mobilization of SERCA2b stores, we asked which of the P2Y1 and/or the P2Y12 ADP receptor is(are) activated. The effect of pharmacological inhibition of the ADP receptors on the SERCA3-dependent activation of platelets was assessed using the P2Y1 and P2Y12 antagonists, MRS and ARC69931MX (ARC), respectively; the efficacy of which was checked by Ca²⁺ mobilization by ADP (Supplementary Figure S1). First, thrombin-elicited Ca²⁺ mobilization in WT platelets as assessed by flow cytometry was not significantly diminished in the presence of the P2Y1 antagonist MRS. By contrast, the P2Y12 antagonist ARC induced a marked and significant reduction in Ca²⁺ mobilization, almost paralleling Ca²⁺ mobilization in Serca3^-/- platelets or in WT platelets treated with the ADP scavenger apyrase (Figure 1). Addition of MRS to ARC did not reinforce the antagonistic effect of ARC.

Figure 1.

Effect of P2-class purinergic receptor type Y subtype 1 (P2Y1) and P2-class purinergic receptor type Y subtype 12 (P2Y12) antagonists on Ca²⁺ mobilization in WT and Serca3^-/- mouse platelets. (A) Washed WT (blue tracings) or Serca3^-/- (orange tracings) platelets resuspended in the presence of EGTA (100 μmol/L) to prevent Ca²⁺ influx were preincubated with the Ca²⁺ fluorophore BAPTA for 30 minutes, then stimulated, in absence of external Ca²⁺ (100 μmol/L EGTA), with Thr (40 mU/mL) for 5 minutes, and fluorescence induced by Ca²⁺ mobilization assessed by cell sorting. The ratio of fluorescence of activated vs resting platelets (set as 1) is expressed as a function of time in seconds. Then, platelets were preincubated for 10 minutes with (B) the inhibitor of P2Y1 (MRS2179 [MRS], 100 μmol/L) or (C) the inhibitor of P2Y12 (ARC69931MX [ARC], 20 μmol/L) or (D) both, or (E) the ADP scavenger apyrase (5 U/mL) before stimulation with Thr (40 mU/mL). (F) Area under curves for 1 minute was assessed in arbitrary units for WT (blue) and Serca3^-/- (orange) platelets (n = 3). Data are expressed as mean ± SEM, and probability P calculated by the 2 way anova using the Tukey's multiple comparison test between all means; P ≤ 0.05, ∗; ≤ 0.01, ∗∗

Most importantly, the secretion kinetics of WT platelets stimulated with thrombin was not affected by either ADP receptor antagonist at 5 seconds, confirming no role for ADP or its receptors in initial secretion driven by SERCA3 stores mobilization. Accordingly, Serca3^-/- platelets did not secrete within this time frame (Figure 2A). At 10 seconds and thereafter, the P2Y12 antagonist ARC did strongly inhibit secretion, whereas MRS, the P2Y1 antagonist did show very little inhibition if any (Figure 2B–D). The inhibitory effect of ARC was not increased by addition of MRS. These results suggest that P2Y12, but not P2Y1, is involved in delayed but not in early secretion induced by the thrombin-stimulated SERCA3 store mobilization.

Figure 2.

Inhibition of the P2-class purinergic receptor type Y subtype 1 (P2Y1) and P2-class purinergic receptor type Y subtype 12 (P2Y12) ADP receptors and effect on secretion of WT and Serca3^-/- platelets. Mouse WT (blue) or Serca3^-/- (orange) platelets were washed, resuspended at 3 x 10⁸ platelets/mL in Tyrode buffer without Ca²⁺, and preincubated with buffer (-) or with the antagonists (+) MRS2179 (MRS) (P2Y1 specific, 100 μmol/L), ARC69931MX (ARC) (P2Y12 specific, 20 μmol/L) or both for 10 minutes. Platelets were then stimulated with Thr (40 mU/mL) in absence of stirring, for 5 seconds (A), 10 seconds (B), 1 minute (C), or 3 minutes (D) before stopping secretion by addition of phenylalanyl-L-prolyl-L-arginine chloromethyl ketone (PPACK) (100 μmol/L) and immediately chilling at 0°C for 5 minutes, recovering supernatants by centrifugation and assessing ATP concentrations (see Section 2). Data are expressed after subtraction of ATP values obtained in absence of thrombin stimulation. In other experiments, Thr (40 mU/mL) (E), Thr (40 mU/mL), and ADP (10 μmol/L) (F) were added (+) or not (-) and incubated for 3 minutes before stopping activation. Data are means ± SEM, and probability P calculated with 2 way anova Bonferroni multiple comparison test (A–D, n = 6) or the Tukey's multiple comparison test between all means (E, F, n = 3); not significant, ns; P ≤ 0.05, ∗; ≤ 0.01, ∗∗; ≤ 0.001, ∗∗∗; ≤ 0.0001, ∗∗∗∗

We have previously shown that exogenous ADP added simultaneously to other agonists restored defective Ca²⁺ mobilization and secretion in Serca3^-/- platelets, through an IP3-SERCA2b pathway [2,18]. To verify that P2Y12 was the actual ADP receptor involved, ADP was added to thrombin, and secretion was assessed in the presence of the ADP receptor antagonists. In the presence of the P2Y1 antagonist MRS, addition of ADP (10 μmol/L) to thrombin did restore secretion in Serca3^-/- platelets almost to the same level than in WT platelets (Figure 2E,F), consistent with P2Y1 not being the principal ADP receptor involved. By contrast, no restoration of secretion by the addition of ADP to thrombin was observed in the presence of the P2Y12 antagonist ARC, either in WT or Serca3^-/- platelets (Figure 2E,F). This result strongly argues in favor of P2Y12 being the main ADP receptor triggered by early ADP released by the SERCA3 stores mobilization.

3.2. Genetic inactivation of the ADP receptor P2Y12 alters thrombin-stimulated cross-talk between SERCA3 and SERCA2b-dependent Ca²⁺ mobilization pathway

To confirm the role of P2Y12, we examined thrombin or PAR4-AP stimulation of platelets from P2Y1^-/-;PF4Cre (heretofore noted P2Y1^-/-) and P2Y12^-/-;PF4Cre mice (P2Y12^-/-) in which the gene of the corresponding ADP receptor was inactivated in the platelet-megakaryocyte lineage (see Section 2). Ca²⁺ mobilization in thrombin-stimulated P2Y1^-/- platelets was only minimally altered, contrary to P2Y12^-/- platelets where Ca²⁺ mobilization was severely diminished (Figure 3A,B). Moreover, pretreatment with the NAADP antagonist Ned19 inhibited thrombin-induced Ca²⁺ mobilization in P2Y1^-/- platelets to the same extent than WT platelets (approximately 40%) and also exhibited a strong inhibition on P2Y12^-/- platelets (Figure 3C). This latter result suggests that diminished mobilization in P2Y12^-/- platelets in response to thrombin is not because of the alteration of SERCA3 store mobilization and is thus likely because of the inhibition of the SERCA2b pathway. Secretion kinetics confirmed that although early secretion at 5 seconds, only dependent on the SERCA3 store mobilization, was comparable with WT, both in P2Y1^-/- as well as in P2Y12^-/- platelets, no secretion inhibition was observed at later time points with P2Y1^-/- (Figure 3D). By contrast, beyond 10 seconds, P2Y12^-/- platelets exhibited very low secretion. Similarly, early serotonin secretion at 5 seconds was not impaired in P2Y12^-/- platelets, but delayed serotonin secretion was completely blocked (Figure 3E). These results suggest that although P2Y1 is not involved in either SERCA3- or SERCA2b-dependent phases of thrombin-stimulated secretion, P2Y12, although not involved in the initial SERCA3-dependent phase of secretion, is essential to the SERCA2b-dependent phase of secretion.

Figure 3.

Potentiation of sarco-endoplasmic reticulum calcium ATPase type 2b (SERCA2b)-dependent secretion by sarco-endoplasmic reticulum calcium ATPase type 3 (SERCA3)-dependent ADP is altered in P2-class purinergic receptor type Y subtype 12 (P2Y12)- but not in P2-class purinergic receptor type Y subtype 1 (P2Y1)-defective mouse platelets. (A) Washed WT (blue tracings), P2Y1^-/- (gray tracings), or P2Y12^-/- (green tracings) platelets resuspended in the presence of EGTA (100 μmol/L) to prevent Ca²⁺ influx were preincubated with the Ca²⁺ fluorophore Oregon green 488 BAPTA for 30 minutes, then stimulated with Thr (40 mU/mL) for 5 minutes, and fluorescence induced by Ca²⁺ mobilization assessed by flow cytometry. The ratio of fluorescence of activated vs resting platelets (set as 1) is expressed as a function of time in seconds. (B) Area under curves for 1 minute was assessed in arbitrary units (A.U.) for WT (blue), P2Y1^-/- (gray), or P2Y12^-/- (green) platelets (n = 7). (C) Quantification of Ca²⁺ mobilization in the absence (-) or the presence of the nicotinic acid adenosine dinucleotide phosphate (NAADP) antagonist Ned-19 (10 μmol/L) is shown as area under curves for 2 minutes, assessed in arbitrary units (A.U.) for WT (blue), P2Y1^-/- (gray), or P2Y12^-/- (green) platelets after stimulation with Thr (40 mU/mL) (n = 3). (D) Kinetics of secretion of WT (blue dots), P2Y1^-/- (gray), or P2Y12^-/- (green) platelets were assessed after stimulation with thrombin (40 mU/mL) in absence of stirring, for 5, 10, 30, 60, or 180 seconds before stopping secretion by addition of phenylalanyl-L-prolyl-L-arginine chloromethyl ketone (PPACK) (100 μmol/L) and immediately chilling at 0°C for 5 minutes, recovering supernatants by centrifugation and assessing ATP concentrations (see Section 2) (n = 4). (E) WT (blue), Serca3^-/- (orange), or P2Y12^-/- (green) mouse platelets were washed, resuspended at 3 x 10⁸ platelets/mL in Tyrode buffer without Ca²⁺ and stimulated with thrombin (40 mU/mL) in the absence of stirring, for indicated times (5, 10, and 180 s) before stopping secretion by addition of PPACK (100 μmol/L) and immediately chilling at 0°C for 5 minutes, recovering supernatants by centrifugation and assessing serotonin concentration (see Section 2) (n = 3). (F) WT (blue) or P2Y12^-/- (green) platelets were preincubated with diluent (-) or (+) either Ned-19 (10 μmol/L) for 10 minutes, thapsigargin (Tg) (200 nmol/L) for 15 minutes or both, and stimulated (+) or not (-) with Thr (40 mU/mL) for 3 minutes. ATP secretion was measured as described (Methods) (n = 5). Data are expressed as mean ± SEM, and probability P calculated by the 2 way anova using the Tukey's multiple comparison test between all means except for (F), where was used the Bonferroni multiple comparison test between WT and P2Y12^-/-; not significant, ns; P ≤ 0.05, ∗; ≤ 0.01, ∗∗; ≤ 0.001, ∗∗∗; ≤ 0.0001, ∗∗∗∗

Examination of the respective effects of Ned-19 which blocks SERCA3 stores mobilization and Tg, an inhibitor of SERCA2b which leads to the emptying of SERCA2b-dependent Ca²⁺ stores, showed that on 3 minutes of thrombin stimulation, Ned-19-pretreated P2Y12^-/- platelets exhibited virtually no secretion compared with Ned-19-pretreated WT platelets (Figure 3F). By contrast, Tg inhibited approximately 50% of ATP secretion in WT platelets but did not further affect P2Y12^-/- platelet secretion, reaching the same level than Tg-inhibited WT platelets. Our results thus confirm that on low thrombin stimulation, P2Y12 is not involved in the initial SERCA3-dependent secretion but drives the secondary SERCA2b-dependent secretion. P2Y1 does not exhibit any role in the SERCA3/SERCA2b secretory mechanism stimulated by low concentration of thrombin.

3.3. P2Y12 also controls the SERCA2b-dependent secretion amplification step in human platelets

We next examined the role of the ADP receptor P2Y12 in Ca²⁺ mobilization induced by thrombin in human platelets. In Figure 4, on thrombin stimulation of platelets, inhibition of P2Y12 by ARC, scavenging ADP with apyrase, or both, led to a marked diminution of Ca²⁺ mobilization. Ned-19 and the blocking of SERCA3 store mobilization lowered Ca²⁺ mobilization within the same range (60% inhibition) of ARC or apyrase, and addition of either agent to Ned-19 did not show any additive effect (Figure 4H). Thus, global Ca²⁺ mobilization by thrombin involves SERCA3 stores mobilization, secreted ADP, and P2Y12, but not P2Y1. Similar results were obtained for serotonin (Supplementary Figure S2A–C).

Figure 4.

P2-class purinergic receptor type Y subtype 12 (P2Y12), not P2-class purinergic receptor type Y subtype 1 (P2Y1), is involved in Ca²⁺ mobilization elicited by Thrombin and in late secretion. Washed control human platelets resuspended in the presence of EGTA (100 μmol/L) to prevent Ca²⁺ influx were preincubated with the Ca²⁺ fluorophore Oregon green 488 BAPTA for 30 minutes, then stimulated with thrombin (40 mU/mL) for 5 minutes, and fluorescence induced by Ca²⁺ mobilization assessed by flow cytometry. The ratio of fluorescence of activated vs resting platelets (set as 1) is expressed as a function of time in seconds. Before activation, platelets were preincubated for 10 minutes with buffer (A), the P2Y12 antagonist ARC69931MX (ARC) (20 μmol/L, B), the nucleoside phosphate scavenger apyrase (5 U/mL, C), apyrase and ARC (D), the nicotinic acid adenosine dinucleotide phosphate antagonist Ned-19 (10 μmol/L) (E), Ned-19 and ARC (F), and apyrase and Ned19 (G). Area under curves for 1 minute was assessed in arbitrary units (A.U.) after stimulation with Thrombin in the absence (-) or the presence (+) of ARC, apyrase or Ned-19 (n = 3). (H) Tracings illustrated in the figure are representative of 3 independent experiments. (I) Kinetics of secretion of washed human platelets was assessed after pre-incubation for 10 minutes with Tyrode buffer (blue) or the P2Y12 antagonist (ARC) (orange) and stimulation with Thr (40 mU/mL) in absence of stirring or of added Ca²⁺ for 2, 3, 4, 5, 10, 30, 60 seconds or 3 minutes, before stopping secretion by the addition of phenylalanyl-L-prolyl-L-arginine chloromethyl ketone (PPACK) (100 μmol/L) and immediate chilling at 0°C for 5 minutes, recovering supernatants by centrifugation and assessing ATP concentrations (see Section 2) (n = 3). Data are expressed as mean ± SEM, and probability P calculated by anova; P≥0.05, ns; ≤ 0.05, ∗; ≤ 0.01, ∗∗; ≤ 0.001,∗∗∗; ≤ 0.0001;∗∗∗∗

Next, we examined the role of P2Y12 in secretion elicited by thrombin. Secretion assessment on thrombin stimulation (40 mU/mL) of human platelets showed that inhibition of P2Y12 by ARC did not affect secretion until 3 seconds, after which, it almost completely blunted secretion until 3 minutes (Figure 4I). This suggests that similar to mouse platelets, human platelet P2Y12 is not involved in the first wave of secretion elicited by thrombin and is consistent with the initial secretion driven by the SERCA3 being independent of P2Y12. A decline in the secreted ATP level between 3 and 4 seconds when P2Y12 is blocked by ARC suggests that beyond this time point, in absence of active P2Y12, ATP (and probably ADP) undergoes either degradation or uptake by platelets. Altogether, these results confirm that thrombin-induced early secretion corresponding to the SERCA3-dependent secretion is independent of P2Y12, whereas delayed secretion is dependent on P2Y12 and is likely to correspond to the SERCA2-dependent secretion.

4. Discussion

We have previously observed that the absence of SERCA3 in mice or the reduced expression of SERCA3 protein in human platelets is associated with decreased platelet activation in response to low concentrations of agonists, whereas platelets normally respond to higher doses [2,19].

Considering the interest for a way to reduce unwanted activation of platelets without completely inhibiting their function, we studied the involved signaling pathways.

We have then shown that the 2 distinct platelet Ca²⁺ mobilization pathways involving stores controlled by the Ca²⁺-ATPases SERCA2b and SERCA3, respectively, are regulated differently and most importantly control 2 distinct phases of ADP secretion. Of note, we found that in both mouse and human platelets, the early phase of ADP secretion dependent on SERCA3 stores potentiated the delayed phase of ADP secretion controlled by SERCA2b stores. Therefore, the SERCA3 and SERCA2b store mobilization pathways cross-talk using ADP. Here we have addressed the next question: which of the P2Y1 and/or P2Y12 ADP receptors is (are) involved in this cross-talk? We show that ADP released by SERCA3 store mobilization potentiates thrombin-stimulated secretion dependent on the SERCA2b store mobilization through the P2Y12 and not the P2Y1 ADP receptor.

These results raise a number of questions, one of which is why does P2Y12 appear as the main ADP receptor activating the SERCA2b pathway, P2Y1 playing a very minor or no role? In fact, of the 2 platelet ADP receptors, P2Y12 is considered playing a more relevant role than P2Y1 in platelet physiology [23], and this is because of the much lower number of P2Y1 compared with P2Y12 receptors [24], and the ensuing weaker signaling, probably moreover "swamped" by P2Y12 as well as thrombin signaling. Alternatively, since P2Y1 (but not P2Y12) desensitization is PKC-dependent [25], it may be desensitized by thrombin activation. At any rate, our finding that P2Y12 is the predominant ADP receptor acting in the potentiation step following SERCA3 store Ca²⁺ mobilization and early secretion is consistent with this view.

The involvement of P2Y12 in the potentiation of SERCA2b store mobilization may appear counter-intuitive since P2Y12 does not activate phospholipase C [26] and thus cannot directly mobilize SERCA2b stores via IP3 release. However, we have previously established that SERCA3 store mobilization potentiated SERCA2b store-dependent secretion via the thromboxane A2 pathway, at least in part (about 30%) [18]. Since the Gq-dependent platelet thromboxane A2 receptor TPα activates phospholipase C and releases IP3 [27], we conclude that the thromboxane A2 pathway is one of the pathways triggered by P2Y12 leading to further SERCA2b store mobilization (see Figure 5), although another IP3-producing pathway yet unidentified is likely to be involved.

Figure 5.

Illustration of the links between sarco-endoplasmic reticulum calcium ATPase type 3 (SERCA3)- and sarco-endoplasmic reticulum calcium ATPase type 2b (SERCA2b)-dependent pathway in platelet activation. Thrombin stimulation at low concentration activates 2 signaling pathways, the first one leading to the production of nicotinic acid adenosine dinucleotide phosphate (NAADP) promoting SERCA3 stores mobilization and the second one activating phospholipase C and production of 1,4,5, inositol trisphosphate (IP3), releasing Ca²⁺ from SERCA2b stores. The first pathway results in a rapid secretion of ADP (red arrow) that potentiates the second SERCA2b-dependent pathway via the ADP receptor P2-class purinergic receptor type Y subtype 12 (P2Y12) (not P2-class purinergic receptor type Y subtype 1 [P2Y1]) [18]. Note that P2Y12 cannot activate Phospholipase C directly but probably indirectly by TxA₂ and its receptor TPα, which activate phospholipase C and the release of IP3, as suggested by the fact that the potentiation step is indomethacin-sensitive [18]. Note that ADP released via this pathway (blue arrow) may also activate P2Y12.

Interestingly, the SERCA3-NAADP signaling pathway appears to be relevant only for full platelet activation under conditions of low stimulation. In the presence of a high dose of thrombin, full platelet secretion or platelet aggregation is still reached even in conditions of absence or inhibition of this SERCA3-NAADP pathway. This may thus be interpreted as the SERCA3-NAADP pathway being an accessory, dispensable pathway. However, the physiological and pathological impact of its absence or inhibition strongly suggest the opposite, not to mention the evolutionary conservation of the pathway from mice to humans. We thus infer from these considerations that the SERCA3-NAADP pathway provides fine tuning for platelet activation, allowing may be a more gradual response. In the scheme proposed in Figure 5, we suggest that a high level of stimulation could produce enough IP3 to completely mobilize Ca²⁺ from SERCA2b-dependent stores, whereas ADP-dependent IP3 production would only be required in the presence of a low level of agonist. This would imply an autocrine role for ADP secretion. It is also possible that the early secreted ADP has a paracrine action contributing to the activation of platelets that are not at the site of activation and thus helping to build a stronger thrombus. This may be in agreement with our initial in vivo observations in SERCA3-deficient animals, where we observed significantly more rebleeding or embolism [2]. At this point, we cannot tell which mechanism is really at play.

One of the questions relating to early secretion driven by the SERCA3 store mobilization is whether it corresponds to actual dense granule secretion or to another non-vesicular secretory mechanism, such as export through the ATP transporter pannexin1 [28]. We tested this question by the assessment of serotonin secretion, not exported by pannexin1, but cosecreted with ATP or ADP from dense granules. We found parallel secretion kinetics between ATP and serotonin in mouse platelets (Figure 3E) and in human platelets (Supplementary Figure S2A). Moreover, serotonin secretion within the first 5 seconds of secretion behaved like ATP and was inhibited by Ned-19, and not by the SERCA2b inhibitor Tg (Supplementary Figure S2B). These results show that serotonin and ATP are cosecreted by SERCA3 store mobilization and are thus consistent with dense granule secretion. Interestingly, SERCA3-dependent secretion represents approximately 5 pmoles of ATP (for 3 × 10⁶ platelets in our assay) within the first seconds of activation (eg, see Figure 4). In total, 5 pmoles of ATP represent approximately 10% of the total of releasable ATP (50 pmoles) assessed with thrombin at high concentration (2 U/mL) after 3 minutes [2,29]. Since mouse platelets exhibit 8 to 10 granules [30], 10% might correspond to a single granule. This is close to 2 granules released by 0.1 U/mL thrombin in 10 seconds observed by Eckly et al. [30], the difference being probably because of our milder conditions (0.04 U/mL in 5 seconds). Whether this granule is specifically associated (functionally or physically) with SERCA3 store mobilization or is randomly recruited by proximity with SERCA3-dependent stores and the plasma membrane remains to be established.

Finally, the pathway of SERCA3 store mobilization being a platelet activation booster, it may be a pertinent antithrombotic target with limited impact on normal hemostasis, deserving close experimental assessment.

Supporting Information

Supplementary Figure S1.

Supplementary Figure S2.