P2Y12 receptor blockade synergizes strongly with nitric oxide and prostacyclin to inhibit platelet activation

Aims In vivo platelet function is a product of intrinsic platelet reactivity, modifiable by dual antiplatelet therapy (DAPT), and the extrinsic inhibitory endothelial mediators, nitric oxide (NO) and prostacyclin (PGI2), that are powerfully potentiated by P2Y12 receptor blockade. This implies that for individual patients endothelial mediator production is an important determinant of DAPT effectiveness. Here, we have investigated this idea using platelets taken from healthy volunteers treated with anti‐platelet drugs. Methods Three groups of male volunteers (n = 8) received either prasugrel (10 mg), aspirin (75 mg) or DAPT (prasugrel + aspirin) once daily for 7 days. Platelet reactivity in the presence of diethylammonium (Z)‐1‐(N,N‐diethylamino)diazen‐1‐ium‐1,2‐diolate (DEA/NONOate) and PGI2 was studied before and following treatment. Results Ex vivo, PGI2 and/or DEA/NONOate had little inhibitory effect on TRAP‐6‐induced platelet reactivity in control conditions. However, in the presence of DAPT, combination of DEA/NONOate + PGI2 reduced platelet aggregation (74 ± 3% to 19 ± 6%, P < 0.05). In vitro studies showed even partial (25%) P2Y12 receptor blockade produced a significant (67 ± 2% to 39 ± 10%, P < 0.05) inhibition when DEA/NONOate + PGI2 was present. Conclusions We have demonstrated that PGI2 and NO synergize with P2Y12 receptor antagonists to produce powerful platelet inhibition. Furthermore, even with submaximal P2Y12 blockade the presence of PGI2 and NO greatly enhances platelet inhibition. Our findings highlight the importance of endothelial mediator in vivo modulation of P2Y12 inhibition and introduces the concept of refining ex vivo platelet function testing by incorporating an assessment of endothelial function to predict thrombotic outcomes better and adjust therapy to prevent adverse outcomes in individual patients.


CONCLUSIONS
We have demonstrated that PGI 2 and NO synergize with P2Y 12 receptor antagonists to produce powerful platelet inhibition. Furthermore, even with submaximal P2Y 12 blockade the presence of PGI 2 and NO greatly enhances platelet inhibition. Our findings highlight the importance of endothelial mediator in vivo modulation of P2Y 12 inhibition and introduces the concept of refining ex vivo platelet function testing by incorporating an assessment of endothelial function to predict thrombotic outcomes better and adjust therapy to prevent adverse outcomes in individual patients.

WHAT IS ALREADY KNOWN ABOUT THIS SUBJECT
• Platelet function is a product of intrinsic platelet reactivity. • This can be modified by dual anti-platelet therapy (DAPT), but also by the influence of the endothelial mediators, nitric oxide (NO) and prostacyclin (PGI 2 ). • NO and PGI 2 also independently amplify each other's effects.

WHAT THIS STUDY ADDS
• Three way synergy between PGI 2 , NO and P2Y 12 receptor antagonism produces powerful platelet inhibition. • Even with submaximal (25%) P2Y 12 blockade, the presence of PGI 2 and NO greatly enhances platelet inhibition. • Assessing endothelial mediator production and associations to platelet cyclic nucleotides in vivo could improve thrombotic outcomes in individual patients.

Introduction
Compromise in the integrity of the vascular endothelium precipitates rapid platelet activation as platelets become exposed to sub-endothelial collagen and tissue factor. This activation is driven by a cascade of complex intracellular signalling pathways leading to the production of secondary platelet agonists, notably thromboxane (TX) A 2 and ADP [1,2]. Dual anti-platelet therapy (DAPT) is recommended for the secondary prevention of atherothrombotic events in patients with acute coronary syndromes or following percutaneous coronary intervention [3,4] and targets these two pathways with a P2Y 12 receptor antagonist, such as clopidogrel, prasugrel or ticagrelor, and aspirin. The P2Y 12 receptor blockers inhibit platelet aggregation by blocking the amplifying effects of ADP acting on platelet P2Y 12 receptors [5], while aspirin targets the TXA 2 -dependent pathway by inhibiting the cyclooxygenase (COX) enzyme within platelets [6]. Whilst DAPT is associated with an improvement in patient outcomes, thrombotic events do still occur. An often explored hypothesis is that the risk of experiencing a thrombotic event is associated with the level of platelet blockade, i.e. those individuals with less effective blockade provided by aspirin and, particularly P2Y 12 receptor blockers, are more at risk of thrombotic events. However, studies have failed to show any benefits from ex vivo monitoring of platelet function and subsequent tailoring of treatment in patients receiving dual anti-platelet therapy [7][8][9][10]. This failure is possibly because the ex vivo platelet tests used in these trials do not consider the environment in which platelets reside in vivo. Namely that within the circulation endothelium-derived autacoids, nitric oxide (NO) and prostacyclin (PGI 2 ), reduce platelet reactivity and prevent inappropriate platelet activation [11,12]. Indeed, within the circulation each platelet is balanced by approximately 50 endothelial cells (e.g. 1.25 trillion platelets vs. 60 trillion endothelial cells in a 70 kg man) [13]. NO diffuses freely into platelets activating guanylyl cyclase (GC) to increase intracellular cGMP levels [14], while PGI 2 binds to IP receptors activating adenylyl cyclase (AC) to increase intracellular cAMP levels [15]. Elevations in the intracellular levels of individual cyclic nucleotides promotes a generalized inhibition of platelet function [16] and the two pathways synergize to produce particularly strong inhibition [12]. NO and PGI 2 also individually synergize with P2Y 12 inhibition producing robust anti-aggregatory platelet effects [17,18].
Taking account of the above observations we hypothesized that within the circulation the levels of endothelium-derived mediators are an important determinant of the efficacy of DAPT. Therefore, for individual patients in vitro measures of platelet reactivity do not accurately predict the in vivo effectiveness of DAPT due to the confounding of differences in endothelial mediator production.
To test this hypothesis we added NO and PGI 2 to standard ex vivo tests of platelet function in blood taken from healthy volunteers receiving anti-platelet therapies.

Study participants
Twenty-four healthy, non-smoking male volunteers (aged 18-40 years) were recruited and participated in the study. Health status was determined though medical history and physical examination, including blood pressure, pulse rate, blood chemistry and urinalysis. Volunteers with normal clinical profiles were included in the study. The study was approved by St Thomas's Hospital Research Ethics Committee (Ref. 07/Q0702/24) and all volunteers gave written consent before entering the study.

Study protocol
Healthy volunteers abstained from aspirin, non-steroidal anti-inflammatory drugs (NSAIDs) and any other antiplatelet therapy for 14 days before commencing the study. The volunteers were divided into groups of eight. The first group received aspirin (75 mg; Nu-Seals Cardio 75, Alliance Pharmaceuticals Ltd, Chippenham, UK), the second prasugrel (10 mg; Effient®, Eli Lilly, RA Houten, The Netherlands) and the third both aspirin (75 mg) and prasugrel (10 mg) to represent DAPT for 7 days. Adherence was assessed by interview. Blood samples were collected before and after drug treatment.

Blood collection
Blood for platelet aggregation was collected by venepuncture into tri-sodium citrate (0.32% final; Sigma, Poole, Dorset, UK). Platelet-rich plasma (PRP) was obtained by centrifugation at 175 × g for 15 min at 25°C. Platelet-poor plasma (PPP) was obtained by centrifugation of PRP at 15 000 × g for 2 min. All experiments were completed within 2 h of blood collection.

Incubation with platelet function inhibitors
For in vitro incubation experiments, PRP was treated with either vehicle (0.5% DMSO) or the P2Y 12 receptor blocker prasugrel active metabolite (PAM; a kind gift of AstraZeneca) at 1.5 μM, 3 μM or 6 μM, to represent 25%, 50% or 100% of the concentration needed for complete P2Y 12 receptor blockade, respectively, in the absence or presence of aspirin (acetylsalicylic acid, ASA, 30 μM) for 30 min at 37°C.

Statistics and data analysis
Data were analyzed using Prism 6.0e. Summary data (IC 50 , EC 50 ) were obtained by fitting of data to a logistic equation and tested by Student's t-test (two groups) or one way ANOVA (>two groups). Flow data were analyzed using FlowJo v8.7 (Treestar, Ashland, USA) where the 'single platelet' population was gated based on forward scatter and CD61-APC immunoreactivity (FL-4 mean fluorescence intensity). Statistical significance was determined by two way ANOVA with Dunnett's post hoc test unless otherwise stated and data sets considered different if P < 0.05. P2Y 12 nomenclature conforms to the BJCP guidelines.

Light transmission aggregation responses following in vivo mono and dual anti-platelet therapy
In individuals taking aspirin, standard LTA responses to AA (1 mM) were strongly inhibited, as were responses to collagen. Responses to ADP (5 μM) were also significantly reduced, although to a lesser degree, while those to U46619 (10 μM) were unaffected (Supplementary Figure 1A). In individuals taking prasugrel, aggregatory responses to AA, collagen, ADP and U46619 were all significantly reduced (Supplementary Figure 1B). Aggregations induced by AA, collagen and ADP were abolished in individuals taking DAPT (Supplementary Figure 1C) while responses to U46619 were strongly reduced.
Involvement of cAMP and cGMP in the synergistic effects of P2Y 12 blockade, PGI 2 and DEA/NONOate We found no significant change in cGMP levels in the platelets in response to DEA/NONOate and/or PGI 2 after incubation with aspirin, PAM or aspirin + PAM after platelet aggregation stimulated by collagen ( Figure 3A) or TRAP-6 ( Figure 3B).
Phospho (Ser 239 )-VASP, a downstream marker of PKG activation, remained unchanged in most conditions studied but was increased following TRAP-6 stimulation in the presence of PAM in all cases and most so in the presence of PGI 2 (28 ± 5 to 47 ± 15 units, P < 0.05) or DEA/NONOate + PGI 2 (27 ± 4 to 46 ± 11 units, P < 0.05, Figure 3F).

Discussion
Here we show in healthy individuals receiving standard DAPT leading to consensus levels of platelet inhibition that ex vivo responses to the strong primary platelet activators collagen and TRAP-6 are powerfully influenced by the presence of NO and PGI 2 . The strong synergies between P2Y 12 inhibitors and the cAMP and cGMP signalling systems mean that the in vivo platelet reactivity in patients receiving DAPT will be a function of the level of P2Y 12 receptor blockade and the levels of endothelium-derived NO and PGI 2 . This provides an explanation for different thrombotic outcomes in the presence of similar levels of platelet blockade, i.e. individual patients with different levels of endothelial function, or indeed disease-driven endothelial dysfunction, would have different levels of in vivo platelet inhibition for the same level of DAPT activity, as determined by ex vivo testing.

Table 1A
In vitro effects of aspirin and PAM on platelet aggregation. PRP from healthy volunteers (n = 4) was treated with PAM (1.5 μM, 3 μM and 6 μM) to represent 25%, 50% and 100% maximum concentration for total P2Y 12 receptor inhibition, respectively. Tables show % final aggregation in response to ADP (20 μM), collagen (4 μg ml À1 ) and TRAP-6 (25 μM) in the presence of vehicle (NaOH, 10 mM), NO (100 nM), PGI 2 (1 nM) or NO + PGI 2 in the (A) absence and (B) presence of aspirin (30 μM  DAPT, aspirin plus a P2Y 12 receptor blocker, is the preventative therapy provided to patients at particular risk of coronary thrombosis, notably for the first 12 months following coronary stent implantation or an acute coronary syndrome [20,21]. Despite this therapeutic approach coronary thrombosis still occurs, and there have been great efforts made to find ex vivo tests that could predict for clinical outcomes [22,23]. Deductive reasoning leads to the conclusion that less effective platelet blockade would leave individuals at increased risk of thrombosis and so multiple efforts have been made to link levels of platelet reactivity in ex vivo tests to clinical outcomes. Despite the attractive logic of this approach, tailoring anti-platelet therapy to ex vivo platelet responses has failed to provide any improvement in clinical outcomes as noted in large scale studies such as ADRIE [24] and several large scale prospective, randomized clinical trials, such as GRAVITAS [7], ARCTIC [8], TRIGGER-PCI [9] and TRILOGY [10]. In patients receiving clopidogrel, there are wellcharacterized metabolic differences that can produce suboptimal levels of its active metabolite and consequently result in suboptimal levels of P2Y 12 receptor blockade [25]. There are also some reports of variability in the effects of prasgurel and ticagrelor, although to a much lesser extent than for clopidogrel [26]. Biochemical resistance to the effects of aspirin are also particularly rare [27]. Allowing for differences dependent upon adherence to therapy, individuals on DAPT may in fact present a rather more homogenous level of platelet inhibition than can be associated to different clinical outcomes. Having recently reported that blockade of P2Y 12 receptors greatly increases the inhibitory effects of NO, and knowing that not only was there a similar interaction with the inhibitory effects of PGI 2 but that NO and PGI 2 powerfully synergize to inhibit platelets, we reasoned that differences in the levels of Table 2A In vitro effects of aspirin and PAM on ATP release. PRP from healthy volunteers (n = 4) was treated with PAM (1.5 μM, 3 μM and 6 μM) to represent 25%, 50% and 100% maximum concentration for total P2Y 12 receptor inhibition, respectively. Tables show ATP release (nM) in response to collagen (4 μg ml À1 ) and TRAP-6 (25 μM) in the presence of vehicle (NaOH, 10 mM), NO (100 nM), PGI 2 (1 nM) or NO + PGI 2 in the (A) absence and (B) presence of aspirin (30 μM  NO and PGI 2 in the presence of the same levels of P2Y 12 receptor blockade would produce different levels of platelet inhibition. By testing this hypothesis in individuals receiving standard DAPT we show here that strong and synergistic interactions between P2Y 12 receptor blockade and endothelium-derived mediators produce profound inhibitory effects upon platelets. We firstly established that the drug regime given in our studies elicited satisfactory reduction in baseline reactivity, therefore establishing effectiveness of P2Y 12 and/or COX inhibition in accordance with suggested analytical cutoffs [28]. These reductions were against high pre-treatment levels of platelet reactivity (>70% response to 5 μM ADP) [29]. In these studies, and others presented here, we took care to include the standard measures of platelet function as determined in consensus statements [28,30]. Then to make data readily accessible we have presented results in the form of heat maps that move from red to green, indicating movement from full platelet activation to no platelet activation. It is well known that NO and PGI 2 synergize to inhibit platelets [12] and it has been demonstrated by ourselves and others that P2Y 12 antagonists potentiate the inhibitory actions of both PGI 2 , dependent upon cAMP [17], and NO, dependent upon cGMP generation [18]. Here, we have shown that this synergy, in the presence of NO and PGI 2 is mostly cAMP dependent. Therefore phosphodiesterase 3 (PDE3) inhibitors, such as cilostazol, may have an enhanced effect compared with PDE5 inhibitors, such as sildenafil. Though not widely commented upon, the body contains many more endothelial cells than platelets, in the order of 50 times more, and the two populations constantly interact. In the circulation DAPT exerts its effects upon platelets in the presence of endothelium-derived mediators while these are absent in ex vivo testing. In the studies presented here we found that the interactions of NO, PGI 2 and P2Y 12 receptor blockade in inhibiting platelets were markedly synergistic as noted by isobolographic analysis and measures of aggregation, ATP release, activation of GP IIb/IIIa receptors and P-selectin expression. In volunteers taking DAPT we noted inhibition of responses to ADP and AA that were in keeping with consensus statements of effective DAPT, i.e. in our study the drugs were working to an effective level of clinical efficacy. Despite this level of effective inhibition, high concentrations of the strong primary platelet activators, TRAP-6 or collagen, still caused notable platelet activation. Addition of low concentrations of NO and PGI 2 , to model the environment within the blood vessel, had little effect on their own but led to almost complete inhibition in platelets from individuals treated with DAPT. Similarly, while NO, PGI 2 or DAPT alone had relatively little effect upon platelet granule release, determined as ATP release, when combined they caused more than 50%  inhibition. These results indicate that even in the presence of effective DAPT, i.e. within consensus guidelines, the presence of NO and PGI 2 lead to very much higher levels of platelet inhibition.
Next using an in vitro approach we modelled events in the presence of suboptimal levels of P2Y 12 receptor blockade by using concentrations of PAM that were 50% and 25% of the effective concentration. Under these  conditions we noted that relative to the consensus levels of DAPT we did not achieve significant reduction in platelet aggregation. Notably, however, in the presence of NO and PGI 2 effective levels of inhibition were achieved, even when platelets were exposed to only 25% of the effective concentration of PAM. As we express in heat maps, there is a clear interaction between DAPT and the endothelial mediators that move platelets from reactive ('red') to unreactive ('green'). Interestingly, these comparisons indicate that 25% of the effective concentration of PAM plus NO and PGI 2 produces a stronger inhibition in LTA, the 'gold standard test', than 100% of the effective concentration of PAM in the absence of NO and PGI 2 (i.e. the normal conditions for testing ex vivo platelet responsiveness). This suggests that in individuals in whom suboptimal P2Y 12 inhibition is achieved, such as poor clopidogrel metabolizers, anti-platelet efficacy may be particularly sensitive to any changes in endothelial function. Our in vitro data also demonstrate that the triple synergy between P2Y 12 blockade, NO and PGI 2 can be explained by changes in cAMP signalling, which is consistent with known interactions between NO and PGI 2 [31] and PGI 2 and P2Y 12 [17]. We show that following standard DAPT the level of platelet reactivity is a function of the level of P2Y 12 receptor blockade and the levels of NO and PGI 2 . While we added NO and PGI 2 exogenously they are surrogates for the effects of endogenous NO and PGI 2 and other elevators of platelet cyclic nucleotides such as adenosine. We propose that, since in vivo platelet function is a product of both internal platelet responsive signalling reactivity and the external influence of the endothelium, an assessment of endothelial mediator production could be combined with results from ex vivo platelet testing to predict thrombotic outcomes better in individual patients. Furthermore, with the emergence of this complex and very powerful synergy between PGI 2 , NO and P2Y 12 inhibitors (Figure 4), we should perhaps consider optimizing the availability and activity of endotheliumderived mediators (such as PDE inhibitors), or providing mimetic drugs, rather than adding in further anti-platelet therapies. These findings could, eventually, be applied in Figure 4 Summary of the interaction between the endothelium and P2Y 12 antagonism. In the healthy intact circulation platelets are kept in a non-activated state in part by the action of endothelium-derived mediators, NO and PGI 2 . At areas of endothelial damage platelets become activated leading to platelet adherence and activation. This effect is partly driven by stimulation of P2Y 12 receptors following from platelet release of ADP. This stimulation of P2Y 12 receptors inhibits the effects of cAMP and cGMP within platelets, making platelets more excitable, and amplifying platelet activation. When P2Y 12 receptors are blocked, cAMP and cGMP pathways are not inhibited by ADP and the inhibitory effects of NO and PGI 2 are sustained. The inhibitory effectiveness of P2Y 12 receptor blockade and DAPT in vivo is therefore strongly dependent upon the production of NO and PGI 2 within the circulation a personalized medicine framework where the endothelial mediator production of individuals is assessed and appropriate add-on therapy applied. In a more generalized approach these additional therapies could also be supplied to patient groups with known endothelial dysfunction, such as diabetics. This approach could provide increased anti-platelet efficacy while avoiding the increased risk of bleeding events associated with the approach of triple anti-platelet therapy.

Supporting Information
Additional Supporting Information may be found in the online version of this article at the publisher's web-site:

Figure S5
Representative control data for flow cytometry experiments. GPIIb/IIIa activation by PAC-1 binding in the (A) absence and (B) presence of DAPT, P-selectin expression in the (C) absence and (D) presence of DAPT and VASP phosphorylation (Ser 239 ) in the (E) absence and (F) presence of DAPT was measured by flow cytometry in PRP stimulated with TRAP-6 (25 μM) in the presence of vehicle (NaOH, 10 mM), DEA/NONOate (100 nM), PGI 2 (1 nM) or DEA/NONOate + PGI 2 . Histograms are representative of n = 3