The importance of endothelium-derived mediators to the efficacy of dual anti-platelet therapy

Anti-platelet drugs supply prophylactic therapy for the many millions of patients at risk of secondary atherothrombosis. Importantly, these drugs interact with cyclic nucleotide systems in platelets, which are the same systems that mediate the key endothelium-dependent endogenous pathways of platelet regulation. Endothelial cells reduce the excitability of platelets through the production of nitric oxide (NO) and prostaglandin I₂ (PGI₂). Within platelets NO directly stimulates guanylyl cyclase to produce cGMP while PGI₂ acts on IP receptors to stimulate the production of cAMP by adenylyl cyclase. Elevation of either cGMP or cAMP alone reduces platelet reactivity, while cGMP and cAMP together are strongly synergistic and render platelets unresponsive. These are parts of a dynamic and balanced system and platelets also contain phosphodiesterase enzymes that act to rapidly degrade cAMP and cGMP. Therefore, endothelial cell mediators constantly stimulate the formation of cAMP and cGMP and intraplatelet systems constantly remove them [1–3].

An important consideration that is often overlooked in evaluating the roles of endothelial mediators in atherothrombosis is the localization of the interaction between platelets and endothelium. The ‘average’ person, that is, a 70 kg physiological man, has ≈5 l of blood and will have ≈1.25 trillion blood platelets. The same person’s body would contain ≈60 trillion endothelial cells [4]; that is, around 50-fold more endothelial cells than platelets. In a capillary of 8 μm diameter and 1 cm in length, the volume to internal surface area ratio is 0.5; in the proximal left anterior descending artery of 4 mm diameter and 1 cm length, the volume to internal surface area ratio is 1000. So there is a 2000-fold greater ratio of endothelial cells to platelets in capillaries than in large arterial vessels. Similarly, the cross-sectional area of the aorta, which receives the entire blood flow, is ≈3–5 cm² while that of the body’s total capillary bed is ≈4500–6000 cm². Finally, blood flows around 500 times faster in arteries than in capillaries; the blood flow in a healthy coronary artery may be from 10 to over 100 cm/s, in a capillary ≈0.1 cm/s. Therefore, the influence of the 50-fold excess of endothelial cells over platelets will predominantly occur in the capillaries where there is the time and space for this interaction to be exerted, and not in the large conduit vessels.

At rest human cardiac output matches the blood volume, at around 5 l/min, meaning that individual platelets are generally exposed to both the pulmonary and systemic capillary beds every minute. The endothelial cells in these beds make intimate contact with the platelets 2–3 times a minute exposing them to NO and PGI₂ and so elevating platelet levels of cGMP and cAMP. These platelets carry their cyclic nucleotide tone with them from their passage through capillary beds and will be exposed to residual circulating levels of PGI₂ (as well as other agents such as adenosine) that will continue to elevate the levels of cAMP. Because of this cyclic nucleotide tone platelets are continually and strongly inhibited at multiple levels. This includes inhibition of platelet binding to the blood vessel wall, platelet calcium extrusion, platelet degranulation, platelet actin polymerization, and platelet pro-coagulant activity. In accordance with this idea, platelets lacking IP receptors have exaggerated responses to arterial vessel damage, consistent with loss of a generally inhibitory effect of PGI₂ within the arterial circulation, and mice or humans with dysfunctional guanylyl cyclase, and so less responsive to NO, are at elevated risk of thrombosis following vascular insult [5,6].

Bearing in mind the strong inhibitory tone conferred upon platelets by endothelial cells, it is interesting to consider how fast flowing platelets can aggregate at sites of arterial breakage and so cause atherothrombosis. The answer appears to be that the first arriving platelets which stick to the damaged blood vessel wall release secondary mediators; most notably ADP that activates P2Y₁₂ receptors on platelets. This activation of P2Y₁₂ receptors leads to rapid blockade of the action of adenylyl cyclase, turning off the production of cAMP, and countering the inhibitory signaling actions of cGMP [3,7,8]. This combined inhibition of cAMP and cGMP pathways greatly potentiates platelet activation responses [7,8], a process that is enhanced by the rapid destruction of the cyclic nucleotides by the phosphodiesterases [1–3].

Thus we can draw together many lines of research to propose a single idea of atherothrombosis with the relationship between the endothelium and the platelets at its center. Endothelial cells produce NO and PGI₂ that (together with adenosine and others) elevate cAMP and cGMP in platelets, and these two systems synergize to produce the greatest inhibition of platelet reactivity. This process largely takes place within capillary beds and the rapidly moving platelets bring the effects with them to the large arterial vessels. Exposure of platelets to a damaged blood vessel wall at the site of atherosclerotic plaque rupture causes them to release ADP that turns off the cAMP and cGMP generating and signaling systems and in concert with very active phosphodiesterases rapidly switches the platelets to being strongly reactive. These systems are in balance with the time it takes platelets to pass from capillary beds to large vessels.

With the above concepts in mind we can consider the effects of dual anti-platelet therapy, that is, aspirin plus a P2Y₁₂ receptor blocker. Aspirin is established in clinical practice as the default anti-platelet therapy in cardiovascular disease. This is based upon robust, widely accepted data that for ‘at-risk’ patients low-dose aspirin reduces thrombotic events by around 30% [9]. Aspirin acts by irreversibly blocking the cyclooxygenase enzyme within platelets and consequently inhibiting the production of thromboxane A₂. Thromboxane A₂, when unchecked, drives further aggregation through stimulation of its cognate TP receptors on neighboring platelets. As knowledge regarding anti-thrombotic therapy has developed, it has become accepted that administration of blockers of the ADP P2Y₁₂ receptor together with aspirin further reduces the risk of acute thrombotic events [10–12].

Importantly, due to the manner in which this therapy has evolved, landmark trials have all been carried out in the presence of aspirin and so have not tested the efficacies of P2Y₁₂ receptor blockers alone. We have reported previously that addition of aspirin to more recently developed strong P2Y₁₂ receptor blockers, e.g. prasugrel and ticagrelor, produces relatively little additional inhibition of platelet function as assessed outside the body [13]. In turn, we have suggested that in the presence of strong P2Y₁₂ receptor blockade, the addition of aspirin may actually produce a net reduction in anti-thrombotic efficacy [14]. This suggestion follows from the now widely accepted idea that whole body inhibition of cyclooxygenase increases the risk of thrombosis [15], and from multiple studies that have shown aspirin can inhibit cyclooxygenase at sites other than platelets. Reduction in vascular PGI₂ production, for instance, caused by aspirin could reduce platelet cyclic AMP, increase platelet reactivity, and so increase the potential for thrombosis. The incremental increase in platelet inhibition provided by aspirin inhibiting platelet thromboxane A₂ production on top of strong P2Y₁₂ receptor blockade may well be insufficient to balance the negative effect of PGI₂ suppression on the cardiovascular system.

As a further layer of complexity, blockade of platelet P2Y₁₂ receptors also increases the sensitivity of platelets to the inhibitory effects of both PGI₂ [7] and NO [8]. As previously mentioned, PGI₂ and NO have synergistically inhibitory effects upon platelets. The interaction with P2Y₁₂ receptor blockers actually provides a powerful three-way synergistic effect: NO, PGI₂, and P2Y₁₂ receptor blockade are inhibitory individually, synergize with each other in individual pairs, and synergize markedly more as a trio [3]. Understanding this interaction provides even further cause to question the benefit of addition of aspirin to strong P2Y₁₂ receptor blockade. Inhibition of the production of PGI₂ will lessen this key three-way synergy and increase platelet reactivity beyond that predicted for loss of PGI₂ alone (which by itself has been suggested as responsible for the pro-thrombotic effects of non-steroid anti-inflammatory drugs).

Future of dual anti-platelet therapy without aspirin

In addition to our experimental modeling, recent data from clinical trials also indicate that we should re-evaluate the role of aspirin in dual anti-platelet therapy [11,16,17] and trials are underway specifically to test this idea (e.g. GLOBAL-LEADERS, https://clinicaltrials.gov/ct2/show/NCT01813435; and TWILIGHT, https://clinicaltrials.gov/ct2/show/NCT02270242). If these trials demonstrate that aspirin provides little additional benefit to strong P2Y₁₂ receptor blockade, then use of drugs such as prasugrel or ticagrelor alone, that is, without aspirin, could become standard of care. This approach may well reduce both thrombotic and bleeding risk in patients. Furthermore, the addition to P2Y₁₂ receptor blocker single therapy of novel drugs such as direct guanylyl cyclase activators, to compensate for failing endothelial cell function, or targeted phosphodiesterase inhibitors, to increase the levels of intraplatelet cyclonucleotides, could provide enhanced anti-thrombotic protection. More generally an improved understanding of the dynamic regulation of platelet responsiveness by the endothelium could help us better understand issues such as ‘resistance’ to anti-platelet therapies [18] and pave the way for new and more effective therapeutic strategies.

Financial & competing interests disclosure

TD Warner has received grants from AstraZeneca for research into the mechanisms of action of P2Y₁₂ receptor blocking drugs. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.