The central role of the P2T receptor in amplification of human platelet activation, aggregation, secretion and procoagulant activity


Robert F. Storey, Cardiovascular Medicine, University Hospital, Queen's Medical Centre, Nottingham NG7 2UH, UK .


Adenosine diphosphate (ADP) is an important platelet agonist and ADP released from platelet dense granules amplifies responses to other agonists. There are three known subtypes of ADP receptor on platelets: P2X1, P2Y1 and P2T receptors. Sustained ADP-induced aggregation requires co-activation of P2Y1 and P2T receptors. AR-C69931MX, a selective P2T receptor antagonist and novel antithrombotic agent, was studied to characterize further the function of the P2T receptor. The roles of the P2Y1 receptor and thromboxane A2 were assessed using the selective P2Y1 antagonist A2P5P and aspirin respectively. Aggregation was measured by whole blood single-platelet counting and platelet-rich plasma turbidimetry, using hirudin anticoagulation. Dense granule release was estimated using [14C]-5-hydroxytryptamine (HT)-labelled platelets. Ca2+ mobilization, P-selectin expression, Annexin V binding and microparticle formation were determined by flow cytometry. P2T receptor activation amplified ADP-induced aggregation initiated by the P2Y1 receptor, as well as amplifying aggregation, secretion and procoagulant responses induced by other agonists, including U46619, thrombin receptor-activating peptide (TRAP) and collagen, independent of thromboxane A2 synthesis, which played a more peripheral role. P2T receptor activation sustained elevated cytosolic Ca2+ induced by other pathways. These studies indicate that the P2T receptor plays a central role in amplifying platelet responses and demonstrate the clinical potential of P2T receptor antagonists.

Adenosine diphosphate (ADP) activates platelets by binding to purinoceptors on the platelet surface and, in contrast to purinoceptors on other cell types, adenosine triphosphate (ATP) is a competitive antagonist for this process ( Gachet et al, 1996 ). Current evidence suggests that there are three types of ADP receptor on platelet surfaces, classified as P2X1, P2Y1 and P2T (P2TAC or P2YADP) receptors ( MacKenzie et al, 1996 ; Daniel et al, 1998 ; Fagura et al, 1998 ; Geiger et al, 1998 ; Jin et al, 1998 ; Jantzen et al, 1999 Leon et al, 1999 ). Both the P2X1 and the P2Y1 receptors, but not the P2T receptor, have been cloned ( Jin et al, 1998 ; Sun et al, 1998 ). The P2T receptor has been characterized pharmacologically, using selective antagonists, as the receptor linked, via Gi, to inhibition of adenylate cyclase, mediating a fall in the cyclic AMP level in response to ADP ( Mills & Smith, 1972; Daniel et al, 1998 ; Jin et al, 1998 ; Savi et al, 1998 ; Jantzen et al, 1999 ). Studies of a patient with ADP receptor deficiency support the concept that the P2T receptor is a single, distinct receptor subtype ( Leon et al, 1999 ). Co-activation of both the P2Y1 and the P2T receptors (G-protein coupled receptors) is required for platelet aggregation to occur, as detected by turbidimetry ( Jin & Kunapuli, 1998). The P2X1 receptor (a ligand-gated ion channel), which is selectively activated by α,β-methylene ATP, mediates rapid transient Ca2+ influx, but has not been found to contribute to platelet aggregation ( MacKenzie et al, 1996 ; Jin & Kunapuli, 1998; Kunapuli, 1998). The P2Y1 receptor activates phospholipase C (PLC), via Gq, and this accounts for most of the elevation in cytosolic Ca2+ induced by ADP, via formation of IP3 and release of Ca2+ from intracellular stores ( Daniel et al, 1998 ; Leon et al, 1999 ).

ADP may be released from platelet dense granules, where it is stored in high concentration, or erythrocytes and endothelial cells ( Gachet et al, 1996 ). ADP released by platelets stimulated by other agonists, such as thrombin or collagen, amplifies aggregation and secretion responses induced by these agonists ( Cattaneo et al, 1991 , 1997; Colman et al, 1994 ).

Analogues of ATP that bind specifically to the P2T receptor, including AR-C66096 and AR-C67085 (formerly ARL or FPL 66096 and ARL or FPL 67085 respectively) allow in vitro study of the function of the P2T receptor ( Humphries et al, 1994, 1995a, 1995b; Daniel et al, 1998 ; Fagura et al, 1998 ; Jin & Kunapuli, 1998; Jin et al, 1998 ). Blockade of the P2T receptor with these agents abolishes the turbidimetric response to ADP ( Jin & Kunapuli, 1998) and also renders aggregation induced by the thrombin receptor (PAR1)-activating peptide, TRAP, reversible ( Trumel et al, 1999 ).

A more recent and related P2T receptor antagonist, AR-C69931MX ( Ingall et al, 1999 ), has been found to be a highly selective competitive antagonist at the P2T receptor (unpublished observations) and is currently being developed as an intravenous antithrombotic agent. We have used this agent to study the role of the P2T receptor in platelet aggregation and secretion, as well as its role in TRAP-induced procoagulant activity, as determined by platelet microparticle formation and Annexin V binding ( Dachary-Prigent et al, 1993 ). We have related our findings to the role of thromboxane A2 synthesis using aspirin. We have also assessed the effects of AR-C69931MX on agonist-induced rises in cytosolic Ca2+ to explore further the mechanism whereby P2T receptor activation achieves its effects.


Materials ADP, epinephrine, U46619 (a thromboxane A2 mimetic), TRAP (SFLLRNPNDKYEPF), aspirin, platelet-activating factor (PAF; dissolved in 0·25% w/v human serum albumin), EGTA, probenecid and the P2Y2 receptor antagonist adenosine-2,5-biphosphate (A2P5P) were from Sigma. Collagen and dextrose buffer were from Nycomed. Saline was 0·9% NaCl from Baxter. AR-C69931MX [N-(2-methylthioethyl)-2-(3,3,3-trifluoropropylthio)-5′-adenylic acid, monoanhydride with dichloromethylenebisphosphonic acid] was provided by AstraZeneca R&D Charnwood, Loughborough, UK. Hirudin was recombinant desulphatohirudin (Revasc), a gift from Novartis. 5-Hydroxytryptamine (5-HT) and [14C]-5-HT (1·85 MBq/ml) were from Amersham International. Chlorimipramine hydrochloride was from Novartis. Scintillation fluid was from BDH Laboratories Supplies. Fixative solution consisted of saline with 4·6 mmol/l sodium EDTA, 4·5 mmol/l Na2HPO4, 1·6 mmol/l KH2PO4 and 0·16% w/v formaldehyde, pH 7·4 .

Flow cytometry was performed using a FACScan flow cytometer (Becton Dickinson), with light scatter and fluorescence channels set at logarithmic gain. Data were analysed using cellquest software. Fluo-3 AM from Molecular Probes (Oregon, USA) was prepared as a 500 µmol/l solution in anhydrous dimethylsulphoxide (DMSO). CD62P–fluoroscein isothiocyanate (FITC) antibody, mouse immunoglobulin (Ig)G–FITC and anti-CD42a-RPE fluorescent antibodies were from Serotec. Annexin V–FITC was from Alexis Biochemicals. HEPES/Tyrodes (HT) buffer was 129 mmol/l NaCl, 8·9 mmol/l NaHCO3, 2·8 mmol/l KCl, 0·8 mmol/l KH2PO4, 5·6 mmol/l dextrose and 10 mmol/l HEPES, pH 7·4. CaCl2·6H2O was from Fisons.

All concentrations given in subsequent sections are final concentrations in the reaction mixture.

Preparation of blood and platelet-rich plasma (PRP) Blood was obtained from healthy volunteers using a 19G needle and plastic syringe. All volunteers denied taking any non-steroidal anti-inflammatory drug in the previous 2 weeks. Aliquots of blood (9 ml) were transferred to polystyrene tubes containing anticoagulant. Hirudin (50 µg/ml; 900 anti-IIa units/ml) was mainly used as an anticoagulant to maintain physiological Ca2+ levels and avoid the artefactual enhancement of thromboxane A2 synthesis that occurs in medium containing low Ca2+ levels, such as citrate ( Mustard et al, 1975 ; Heptinstall & Mulley, 1977; Lages & Weiss, 1981; Packham et al, 1989 ). In some experiments, after an initial blood sample, healthy volunteers ingested aspirin 600 mg (UniChem) and further blood was venesected 2·5 h later for repeat analysis. Alternatively, in vitro aspirin 100 µmol/l was included with the anticoagulant. Trisodium citrate dihydrate (3·13% w/v; TCD) (1:9 TCD/blood) was used as an anticoagulant in some experiments for comparison with hirudin. For the release reaction experiments, [14C]-5-HT (0·521 µmol/l; 1·11 kBq/ml) was added to the anticoagulant (for uptake by platelets). For whole blood studies, the tubes were incubated at 37°C in a waterbath for a standard 30 min period before experimentation. PRP was prepared by centrifugation of blood at 180 g for 10 min and removal of the supernatant PRP. Platelet-poor plasma (PPP) was prepared by centrifugation of the remaining blood at 1500 g for 10 min and removal of the supernatant PPP. Platelet counts were performed on the PRP and the latter was diluted with PPP to obtain a final platelet count of 300 × 109/l. PRP turbidimetry was performed using a BioData PAP-4 aggregometer.

Platelet aggregation in whole blood Platelet aggregation studies were performed with whole blood using the single-platelet counting technique ( Fox et al, 1982 ; Storey et al, 1998 ). Aliquots (460 µl) of blood were placed in polystyrene tubes with a magnetic stirrer bar and 20 µl of saline or AR-C69931MX or A2P5P was added. Samples were then stirred at 37°C (stirring speed 1000 r.p.m.) for 2 min. Blood (15 µl) was removed and fixed in 30 µl fixative to measure initial platelet count before addition of 20 µl agonist or saline. Aliquots (15 µl) of blood were removed at different time points after addition of agonists, as referred to in the Results section, and fixed in 30 µl fixative. Fixed samples were counted using the Ultra-Flo 100 platelet counter (Becton Dickinson) and percentage aggregation calculated as percentage loss of single platelets compared with baseline.

The following agonists were assessed: ADP (0·3–100 µmol/l), collagen (0·5–8 µg/ml), U46619 (0·3–3 µmol/l), epinephrine (10 µmol/l), 5-HT (3–30 µmol/l), PAF (0·1–1 µmol/l), TRAP (3–30 µmol/l) and α,β-methylene ATP (100 µmol/l). The effects of aspirin, alone or in combination with AR-C69931MX, were assessed for selected concentrations of agonists, as described in the Results section, using AR-C69931MX 100–1000 nmol/l to cover the estimated therapeutic levels to be achieved clinically.

P-selectin expression Twenty microlitres of either saline or AR-C69931MX 100, 300 or 1000 nmol/l was added to 460 µl whole blood incubated at 37°C. After 2 min, 20 µl agonist was added (see Results section) without continuous stirring and an aliquot was removed at 4 min for fixing, as above. Unstimulated samples were used for baseline measurement. Samples were incubated with a saturating concentration of anti-CD62P for 20 min, washed and resuspended in FACSflow. Non-specific binding was determined using mouse IgG–FITC antibody (an antibody of irrelevant specificity) in place of anti-CD62P. Samples were diluted in FACSflow and analysed by flow cytometry. A gate was applied to the platelet region and 2000 platelet events were collected. P-selectin expression was determined as the median fluorescence of the entire platelet population and results were expressed as an increase over baseline values.

[14C]-5-HT release 5-HT release was determined by a previously validated method ( Heptinstall et al, 1980 ). Twenty microlitres of saline or AR-C69931MX and 20 µl chlorimipramine (1 µmol/l; to prevent 5-HT reuptake) were added to 440 µl blood containing [14C]-5-HT-labelled platelets and the sample was incubated at 37°C for 2 min. Agonist (20 µl) was then added and samples were stirred for 4 min. The reaction was terminated by placing on ice, as well as adding 50 µl of aspirin (1·26 mmol/l). Samples were centrifuged at 1500 g for 10 min and duplicate 50-µl samples of the supernatant plasma were placed in separate scintillation tubes that were then filled with scintillation fluid, capped and mixed. Sample radioactivity was counted in a scintillation counter. One hundred per cent release was estimated using duplicate 50-µl samples of a saline solution of [14C]-5-HT at the same concentration as that added to the whole blood sample, with adjustment of the results according to the packed cell volume of the blood sample to derive the initial plasma [14C]-5-HT concentration.

Microparticle formation and Annexin V binding Microparticle formation and Annexin V binding induced by TRAP 20 µmol/l were measured in PRP with simultaneous turbidimetry. Aspirin (100 µmol/l) was included with the anticoagulant in some tubes. Saline or AR-C69931MX (20 µl) (10–1000 nmol/l) was added to 460 µl PRP and the samples were warmed for 1 min, then stirred for 1 min with assessment of baseline light transmittance. TRAP (20 µl) (20 µmol/l) was then added and change in light transmittance was assessed for 4 min. At 4 min, 15 µl PRP was removed and added to 30 µl HT buffer + 5 µl Annexin V–FITC + 5 µl CaCl2 (100 mmol/l, final concentration 10 mmol/l) + 5 µl anti-CD42a-RPE. Anti-CD42a was used as a platelet membrane marker to gate out any artefact. Unstimulated PRP was used for a baseline measurement. Samples were incubated in the dark for 10 min at room temperature, then 1 ml HT buffer was added and samples were analysed by flow cytometry. Events were acquired for 10 s. Microparticle formation was assessed by setting the marker for forward scatter to include 1% of events with the lowest forward scatter for the baseline sample. Annexin V binding was assessed by setting the marker for Annexin V-positive events to include 1% events for the baseline sample. These markers were then applied to other samples. The baseline values were subtracted from the results for stimulated samples.

Ca2+ measurements Aliquots (2 ml) of PRP were incubated with fluo-3 AM (5 µmol/l) in the presence of probenecid (2·5 mmol/l) for 30 min at 37°C. Samples were then maintained at room temperature. Aliquots (10 µl) of labelled PRP were added to 1 ml HT buffer at 37°C containing CaCl2·6H2O (1 mmol/l) or EGTA (100 µmol/l), giving a 1:100 dilution of PRP. Aliquots (250 µl) of diluted PRP were then applied to the flow cytometer to measure the baseline fluorescence. A further 480-µl aliquot was added to a tube containing ADP (0·3 µmol/l), TRAP (20 µmol/l) or a combination of both agonists (to mimic the physiological effects of released ADP on the action of TRAP in undiluted PRP). The fluorescence of the sample was then measured after 5, 15, 30, 60 and 120 s. The median fluorescence intensity was determined for each time point.

Statistical analysis Data were analysed using anova for repeated measures on SPSS software and significance assigned to P-values less than 0·05. Data are expressed as means ± SEM. Correlation was measured using Excel 97 polynomial trend analysis.


Roles of the purinoceptor subtypes in ADP-induced platelet aggregation

The roles of the P2X1, P2Y1 and P2T receptors in ADP-induced aggregation were studied using whole blood single-platelet counting, which is sensitive to microaggregation, assessing the aggregation response to the selective P2X1 agonist α,β-methylene ATP and inhibition of ADP-induced aggregation by the selective P2Y1 and P2T antagonists A2P5P and AR-C69931MX respectively. α,β-Methylene ATP, at concentrations up to 100 µmol/l, did not induce even a transient aggregation response (including at 10 s after addition of this agonist), indicating that P2X1 receptor activation alone is not sufficient to induce microaggregation. This is consistent with previous reports that P2X1 receptor activation does not cause aggregation ( Savi et al, 1997 ; Kunapuli, 1998). Typical effects of the P2Y1 and P2T antagonists are represented by a comprehensive study of the time-course of aggregation induced by ADP 0·3 µmol/l, assessing the effects of a wide range of concentrations of AR-C69931MX and A2P5P ( Fig 1). AR-C69931MX, at increasing concentrations, reduced the extent of aggregation and reduced the time to peak aggregation, indicating that the P2T receptor sustains and amplifies ADP-induced aggregation. There was a residual early aggregation peak at 10 s that was resistant to increasing concentrations of AR-C69931MX, despite the low concentration of ADP used. Further studies revealed that each donor had a threshold ADP concentration at which increasing concentrations of AR-C69931MX (up to an excessive concentration of 100 µmol/l) failed to abolish this early aggregation response, indicating that P2Y1 receptor activation alone is capable of producing a transient aggregation response. In contrast, A2P5P tended to flatten the aggregation response consistent with a progressive reduction in the proportion of aggregating platelets with increasing concentrations of A2P5P. These observations are consistent with the P2Y1 receptor initiating ADP-induced aggregation and the P2T receptor amplifying and sustaining this response, as suggested previously ( Hechler et al, 1998 ; Jarvis et al, 2000 ). Simultaneous PRP turbidimetry and single-platelet counting, as described previously ( Storey et al, 1998 ), showed that AR-C69931MX could abolish the turbidimetric response to ADP concentrations up to 10 µmol/l, but there remained a transient microaggregation response (mediated by the P2Y1 receptor) that was detected by single-platelet counting but not by turbidimetry (data not shown).

Figure 1.

Inhibition of platelet aggregation induced by 0·3 µmol/l ADP by (A) the P2T antagonist AR-C69931MX and (B) the P2Y1 antagonist A2P5P, as assessed by whole blood single-platelet counting, illustrating different effects of the antagonists on the initial aggregation response (n = 6). Note that in (A) the effects of 1000 and 10 000 nmol/l AR-C69931MX are superimposed, indicating saturation of the effects of P2T receptor blockade. Results are means ± SEM.

Both AR-C69931MX and A2P5P induced disaggregation of ADP-induced aggregates ( Fig 2), indicating that continuous P2T and P2Y1 receptor occupancy by ADP are required for sustained ADP-induced aggregation.

Figure 2.

Disaggregation of platelet aggregates induced by ADP 10 µmol/l by addition of AR-C69931MX 1000 nmol/l (closed squares) and A2P5P 300 µmol/l (open squares) as indicated by arrow, compared with saline control (closed circles) after 4 min aggregation in whole blood, as assessed by single-platelet counting (n = 6). Results are means ± SEM .

Effects of AR-C69931MX and aspirin on platelet aggregation induced by different agonists

The effects of AR-C69931MX and aspirin were compared to determine the relative contributions of P2T receptor activation and thromboxane A2 synthesis to platelet responses to different agonists, as well as to develop further a rationale for the therapeutic use of AR-C69931MX. The effects of in vitro aspirin on aggregation and secretion were identical to the ex vivo effects of aspirin, therefore only the latter are illustrated.

Figure 3A illustrates the inhibition by aspirin and AR-C69931MX of aggregation induced by different agonists in hirudinized whole blood, determined at 4 min after addition of the agonist. AR-C69931MX inhibited ADP-induced aggregation in a concentration-dependent manner, whereas aspirin had no effect. Inhibition by AR-C69931MX of aggregation induced by agonists other than ADP was generally more pronounced at lower concentrations of agonist, as demonstrated by the lower concentrations of TRAP and collagen. Although there was not significant inhibition by AR-C69931MX of aggregation induced by U46619 1 µmol/l, in other experiments aggregation induced by U46619 0·3 µmol/l was inhibited by AR-C69931MX 100 nmol/l (26·1% vs. 7·0%; P < 0·01). It was therefore clearly demonstrated that AR-C69931MX had a broad-spectrum inhibitory effect on aggregation induced by all agonists, reflecting inhibition of P2T receptor activation by released ADP. Only when the concentration of some agonists was raised to a level sufficient to cause maximal aggregation in the absence of the contributory effects of P2T receptor activation by released ADP did AR-C69931MX fail to show a significant inhibitory effect on aggregation. The effect of AR-C69931MX on PAF-induced aggregation was primarily to render aggregation substantially more reversible: maximal aggregation induced by 1 µmol/l PAF occurred at 30 s and 100 nmol/l AR-C69931MX only reduced this from 83·9 ± 1·8% to 77·3 ± 2·5%, whereas at 4 min, after disaggregation had occurred, aggregation was reduced from 54·4 ± 29·1% to 8·5 ± 5·7% (P < 0·01). This suggests that P2T receptor occupancy by ADP is necessary for the effects of PAF to be sustained, as well as those of ADP. Aggregation induced by low concentrations of collagen and TRAP was also rendered more reversible by AR-C69931MX.

Figure 3.

Mean (A) percentage platelet aggregation (whole blood single-platelet counting), (B) percentage [14C]-5-HT release and (C) P-selectin expression (median fluorescence units) before and after ingestion of 600 mg aspirin with/without in vitro addition of 100 nmol/l AR-C69931MX 2 min before agonist (n = 6). Concentrations of agonists are as indicated either in µmol/l or, for collagen, in µg/ml. Results are means ± SEM. *P < 0·05 for effect of aspirin, effect of AR–C69931MX and interaction between AR-C69931MX and aspirin.

Aspirin treatment yielded significant inhibition of aggregation induced by the lower concentration of collagen (2 µg/ml), but no significant inhibition of aggregation induced by ADP, 5-HT, PAF, TRAP, U46619 (at any concentration), 8 µg/ml collagen or epinephrine. The combination of aspirin and AR-C69931MX yielded synergistic inhibition of aggregation induced by collagen.

[14C]-5-HT release

The effects of aspirin and AR-C69931MX in whole blood on [14C]-5HT release (reflecting dense granule release) are shown in Fig 3B. Again, AR-C69931MX inhibited 5-HT release induced by all the agonists, markedly so in the case of U46619, demonstrating a central role for the P2T receptor in the amplification of platelet dense granule secretion. Inhibition of 5-HT release by AR-C69931MX was evident even when there was no inhibition of aggregation induced by stronger agonist stimulation, reflecting the fact that greater platelet activation is required for maximal secretion than for maximal aggregation. Aspirin only inhibited 5-HT release induced by collagen and had no effect on 5-HT release induced by the other agonists. When citrate, rather than hirudin, was used as the anticoagulant, ADP-induced 5-HT release was enhanced, as previously described ( Mustard et al, 1975 ; Heptinstall & Mulley, 1977; Lages & Weiss, 1981; Packham et al, 1989 ), and this was then significantly inhibited by both aspirin and AR-C69931MX (data not shown).

P-selectin expression

The effects of AR-C69931MX and aspirin in whole blood on P-selectin expression (reflecting alpha-granule release) are shown in Fig 3C. 5-HT and epinephrine did not induce significant P-selectin expression (data not shown). The lower concentrations of TRAP and collagen were not studied. Again, AR-C69931MX inhibited P-selectin expression induced by all the agonists that produced a significant response and, again, there was dramatic inhibition by AR-C69931MX of the intensity of secretion induced by U46619.

Aspirin inhibited P-selectin expression induced by collagen, but had no effect on P-selectin expression induced by ADP, PAF, TRAP or U46619. When citrate was used as the anticoagulant, aspirin also weakly inhibited TRAP-induced P-selectin expression (P < 0·05).

Effect of concentration of AR-C69931MX

AR-C69931MX inhibited ADP-induced aggregation and secretion in a concentration-dependent manner. Whole blood concentrations of AR-C69931MX up to 100 nmol/l inhibited aggregation and secretion induced by other agonists in a concentration-dependent manner, but there was little or no significant difference between the effects of 100 nmol/l AR-C69931MX and 1000 nmol/l AR-C69931MX on aggregation, P-selectin expression or 5-HT release induced by agonists other than ADP, indicating a potent yet saturable effect of AR-C69931MX on responses to these agonists (data not shown). The most significant difference was seen for collagen-induced responses, e.g. [14C]-5-HT release induced by 2 µg/ml collagen was reduced from 34·0 ± 4·5% to 17·1 ± 2·3% by 100 nmol/l AR-C69931MX and to 14·3 ± 2·8% by 1000 nmol/l AR-C69931MX (effect of AR-C69931MX concentration: P = 0·001).

Microparticle formation and Annexin V binding

The effects of aspirin and AR-C69931MX on microparticle formation and Annexin V binding induced by TRAP 20 µmol/l in PRP were assessed ( Fig 4), with simultaneous turbidimetry (not shown). AR-C69931MX rendered TRAP-induced aggregation reversible, as shown for AR-C66096 ( Trumel et al, 1999 ), and inhibited microparticle formation and Annexin V binding in a concentration-dependent manner (P < 0·001). Concentrations of > 100 nmol/l AR-C69931MX completely reversed aggregation and abolished microparticle formation and Annexin V binding. There was a strong correlation between reversibility of TRAP-induced aggregation in the presence of AR-C69931MX and inhibition of these procoagulant responses (R2 = 0·894, inhibition of Annexin V binding vs. inhibition of aggregation at 4 min). Aspirin, on the other hand, showed no significant effect on aggregation, microparticle formation or Annexin V binding.

Figure 4.

Inhibition of procoagulant activity by AR-C69931MX assessed by (A) microparticle formation and (B) Annexin V binding induced by 20 µmol/l TRAP after 4 min aggregation in platelet-rich plasma (n = 6). Closed circles represent control samples and open circles represent samples that have been incubated with 100 µmol/l aspirin. Results are means ± SEM.

Ca2+ measurements

Data are presented as mean percentages of the maximum fluorescence obtained with the combination of ADP and TRAP as agonists ( Fig 5). ADP evoked similar increases in cytosolic Ca2+ in the presence and absence of extracellular Ca2+ and these were abolished by the P2Y1 antagonist A2P5P, as previously described ( Jin et al, 1998 ). AR-C69931MX had no significant effect on the initial rise in intracellular Ca2+ but accelerated the decay of Ca2+, both in the presence and absence of extracellular Ca2+, indicating that the P2T receptor sustains elevated Ca2+ induced by P2Y1 receptor activation. TRAP induced only slightly greater increases in peak cytosolic Ca2+ than ADP with a more sustained response, and these increases were not significantly affected by either A2P5P or AR-C69931MX. However, analyses were performed on PRP samples that had been diluted 100-fold, and this dilution would be expected to remove most of the effects of released ADP. Therefore, to mimic the physiological effects of TRAP, the combination of TRAP and ADP was studied. ADP added substantially to the initial increase in cytosolic Ca2+ induced by TRAP and led to greater Ca2+ levels over the 2-min period compared with TRAP alone. A2P5P attenuated this additional increase in peak Ca2+ at 5 s, but failed to prevent a subsequent rise in Ca2+ to levels achieved in the absence of A2P5P and had no effect on the slope of decay. In contrast, AR-C69931MX both attenuated the 5-s peak Ca2+ level and accelerated the decay of cytosolic Ca2+ in an identical fashion to its effects in the studies of ADP alone. The combination of A2P5P and AR-C69931MX restored the response to that seen with TRAP alone. Thus, P2T receptor activation sustains elevated Ca2+ levels induced by TRAP as well as by P2Y1 receptor activation. Removal of extracellular Ca2+ by EGTA slightly enhanced the rate of decay of cytosolic Ca2+, but had no effect on the inhibitory profiles of A2P5P or AR-C69931MX.

Figure 5.

Cytosolic Ca2+ measurements using fluo-3 AM-labelled platelets and 0·3 µmol/l ADP (A and D), 20 µmol/l TRAP (B and E) or both agonists (C and F) in the presence (A–C) or absence (D–F) of extracellular Ca2+. The effects of 1000 nmol/l AR-C69931MX (closed squares) and 300 µmol/l A2P5P (open squares) were assessed in comparison with saline control (closed circles), as well as the combination of both antagonists (open circles, C and F only). Results are means ± SEM of three experiments and are expressed as a percentage of the maximum fluorescence obtained with ADP + TRAP. Maximum fluorescence was 87·9 ± 13·0 and 95·3 ± 15·1 fluorescence units in the presence and absence of extracellular calcium respectively.


Central role of the P2T receptor

These studies demonstrate that the P2T receptor plays a central role in the amplification of platelet aggregation, secretion and procoagulant activity induced by the whole range of natural agonists or their mimetics. We have also demonstrated an association between this role and the effects of the P2T receptor on cytosolic Ca2+ levels that have been elevated by other receptor pathways. Considerable evidence has accumulated recently to support these findings: congenital deficiency of the P2T receptor is associated with impaired dense granule release in response to ADP, U46619 and collagen ( Nurden et al, 1995 ; Cattaneo et al, 1997 ; Leon et al, 1999 ); the P2T receptor antagonist AR-C66096 inhibits processes associated with platelet secretion, namely P-selectin expression induced by some radiographic contrast media ( Heptinstall et al, 1998 ) and aggregation and secretion induced by sera from patients with heparin-induced thrombocytopenia ( Polgar et al, 1998 ); P2T receptor activation has been previously shown to be necessary for sustained aggregation induced by TRAP, via stimulation of PI 3-kinase ( Trumel et al, 1999 ); and several studies have shown that thienopyridines, which act on the P2T receptor, inhibit dense granule release induced by some agonists, including low concentrations of thrombin and platelet-activating factor, but not others, including collagen ( Hardisty et al, 1990 ; Cattaneo et al, 1991 ; Heptinstall et al, 1995 ). The dramatic effects of AR-C69931MX on some responses, such as the secretion response to the thromboxane A2 mimetic U46619 or the procoagulant response to TRAP, show how crucial co-activation of the P2T receptor by released ADP is for a full response to these agonists. Thus, ADP, although a relatively weak agonist by itself, is a powerful co-factor in the responses to stronger agonists via activation of the P2T receptor.

Our studies of cytosolic Ca2+ show how the P2T receptor does not contribute to the initial rise in Ca2+ induced by ADP, which is known to be mediated by the P2Y1 and, to a lesser extent, P2X1 receptors. Studies in the presence of EGTA, which chelates extracellular Ca2+ and therefore abolishes the effect of P2X1 receptor activation, showed little difference from those studies performed in the presence of Ca2+, indicating no significant contribution of the P2X1 receptor activation to the results obtained. We have shown how the P2T receptor sustains the elevated cytosolic Ca2+ induced not only by P2Y1 receptor activation but also by PAR1 activation with TRAP. Studies of the effect of clopidogrel on Ca2+ measurements using fura-2-loaded platelets showed no effect either on cytosolic Ca2+ rises induced by ADP or on the subsequent decay of Ca2+ ( Hechler et al, 1998 ), reflecting either differences in methodology or incomplete P2T receptor blockade by clopidogrel. Similar to our findings, activation of the α2A-adrenergic receptor, which like the P2T receptor is linked to Gi and inhibition of adenylate cyclase, has been shown to potentiate Ca2+ release in response to stimuli such as thrombin ( Keularts et al, 2000 ). The close association between cytosolic Ca2+ and platelet aggregation and secretion is well established ( Hawiger et al, 1994 ). There is also close association between cytosolic Ca2+ and procoagulant changes in transbilayer lipid distributions as aminophospholipid translocase, which maintains phospholipid asymmetry, is inhibited by micromolar concentrations of Ca2+ and lipid scramblase activity is Ca2+ dependent ( Zwaal & Schroit, 1997). Thus, maintenance of cytosolic Ca2+ levels by P2T receptor activation may at least partly explain the effects of AR-C69931MX on platelet aggregation, secretion and procoagulant activity. The signalling pathways whereby P2T receptor activation achieves these amplification effects are currently unclear and the relevance of cyclic AMP has not been established.

Roles of the P2T and P2Y1 receptors in ADP-induced aggregation

The effects of specific P2T and P2Y1 antagonists on platelet microaggregation in response to ADP have not been previously demonstrated and these studies add further weight to previous observations on the relative contribution of the P2T and P2Y1 receptors. Previously, it has been stated that both receptors are required for aggregation as assessed by turbidimetry, a measure of macroaggregation ( Jin & Kunapuli, 1998). Using the more sensitive measure of single-platelet counting, we have shown that microaggregation can occur when the P2T receptor is blocked; however, this aggregation is not sustained for sufficient time to allow the formation of macroaggregates, explaining why macroaggregation is not detected by turbidimetry. In contrast, P2Y1 receptor blockade is sufficient to abolish aggregation. Our findings concur with previous work suggesting that the P2Y1 receptor initiates platelet aggregation, with the P2T receptor playing a synergistic role ( Hechler et al, 1998 ; Jarvis et al, 2000 ). The rapid disaggregatory effects of AR-C69931MX and A2P5P demonstrate how ADP-induced aggregation requires continued occupancy of both the P2T and the P2Y1 receptors by ADP and suggests that displacement of ADP from either receptor leads to deactivation of GPIIb/IIIa complexes and release of fibrinogen. The mechanism by which this occurs warrants further study.

Limitations of aspirin therapy and the potential role of P2T receptor antagonists

These studies demonstrate the limitations of aspirin therapy with regard to inhibition of platelet responses when studied at physiological divalent cation levels. Although it is clear that thromboxane A2 is an important agonist in thrombotic disease and that inhibiting thromboxane A2 production will reduce secretion and aggregation induced by this agonist (as mimicked in these studies by U46619), it is also clear that the majority of platelet responses remain intact despite aspirin treatment. Previous studies of aspirin have shown how in vitro lowering of the divalent cation level with citrate anticoagulation facilitates thromboxane A2 production, particularly in response to ADP ( Mustard et al, 1975 ; Heptinstall & Mulley, 1977; Lages & Weiss, 1981; Packham et al, 1989 ). This explains why the effects of aspirin appear more limited when physiological divalent cation levels are maintained in vitro by using a direct thrombin inhibitor such as hirudin. Our studies show how P2T receptor antagonism yields much greater inhibition of platelet function than the relatively weak effects of aspirin. However, the additive effects of aspirin and AR-C69931MX on collagen-induced responses illustrate the rationale for using the combination of these two agents for anti-platelet therapy.

The effects of AR-C69931MX on platelet secretion and procoagulant activity are important with regard to its potential role in the management of arterial thrombosis. Secreted ADP and 5-HT play an important role in recruiting platelets to arterial thrombosis, and 5-HT contributes to vasoconstriction and reduced arterial flow ( Willerson et al, 1989 ). Procoagulant activity of activated platelets contributes to local thrombin generation and thrombin is another important soluble agonist that recruits platelets to arterial thrombi, as well as generating fibrin ( Badimon & Badimon, 1996). Platelet microparticles may also contribute to further activation of platelets, as well as to activation of monocytes and endothelial cells ( Barry & FitzGerald, 1999). Thus, the effects of AR-C69931MX in vitro appear to support strongly further investigation of its use as an antithrombotic agent for managing arterial thrombosis. The first phase II study of AR-C69931MX in patients with acute coronary syndromes demonstrated that an infusion dose of 4 µg/kg/min achieves a mean plasma level equivalent to a plasma AR-C69931MX concentration of 484 nmol/l ( Storey et al, 1999 ), showing that the in vitro studies presented here are representative of the likely in vivo effects of this agent.


The authors are grateful to AstraZeneca R&D Charnwood for supplying AR-C69931MX and supporting this work, and are grateful to Novartis for the gift of Revasc.