Understanding the role of prostaglandin E₂ in regulating human platelet activity in health and disease

Abstract

The platelet thrombus is the major pathologic entity in acute coronary syndromes, and antiplatelet agents are a mainstay of therapy. However, individual patient responsiveness to current antiplatelet drugs is variable, and all drugs carry a risk of bleeding. An understanding of the complex role of Prostaglandin E₂ (PGE₂) in regulating thrombosis offers opportunities for the development of novel individualized antiplatelet treatment. However, deciphering the platelet regulatory function of PGE₂ has long been confounded by non-standardized experimental conditions, extrapolation of murine data to humans, and phenotypic differences in PGE₂ response. This review synthesizes past and current knowledge about PGE₂ effects on platelet biology, presents a rationale for standardization of experimental protocols, and provides insight into a molecular mechanism by which PGE₂-activated pathways could be targeted for new personalized antiplatelet therapy to inhibit pathologic thrombosis without affecting hemostasis.

Introduction

Platelets perform a critical function by maintaining hemostasis and initiating thrombus formation. However, platelet hyperreactivity has been implicated in the pathophysiology of cardiovascular diseases, including acute coronary syndromes, myocardial infarction, and ischemic stroke. Antiplatelet therapy, such as with aspirin and P2Y12 antagonists, is a cornerstone in the management of cardiovascular disease. Despite use of these agents, platelet hyperreactivity persists in a substantial fraction of patients, contributing to on-treatment major adverse cardiovascular events including cardiovascular mortality. The benefit of antiplatelet drugs is counterbalanced by the risk of major bleeding, because they inhibit both arterial thrombosis and hemostasis. There is considerable interest in developing novel antiplatelet agents that decrease platelet reactivity, but do not affect hemostasis.

There are many naturally occurring molecules that inhibit platelet function, including prostaglandin I₂ (prostacyclin, PGI₂), prostaglandin D₂ (PGD₂), nitric oxide, and adenosine. Prostacyclin has been the most widely studied of these agents. It is produced by vascular endothelium and acts as a potent vasodilator that inhibits platelet aggregation. PGD₂ is produced by mast cells and activated platelets [1], is more stable than prostacyclin in blood plasma, is a vasodilator and inhibits platelet aggregation [2]. Prostaglandin E₂ (PGE₂) also modulates vascular tone and platelet reactivity. It is produced by the vascular endothelium, activated macrophages within atherosclerotic plaques, and platelets. Despite decades of research, the role of PGE₂ on platelet reactivity remains not well understood. Findings have long been confounded by differing effects based on experimental conditions, differences in murine and human biology, the use of synthetic PGE₂ receptor agonists, and unappreciated phenotypic differences in PGE₂ response. This review synthesizes past investigations and current research into the important role of PGE₂ in regulating thrombosis and provides insight into a potential new therapeutic approach that could inhibit pathologic thrombosis without affecting hemostasis.

The role of prostanoids in platelet function

The role of PGE₂ in regulating platelet function is best understood in the context of the larger family of related prostanoids that are produced in the vasculature from common polyunsaturated fatty acid precursors. Prostanoids are bioactive receptor ligands that trigger a variety of G-protein coupled receptor signaling pathways characteristic of individual cell types and activation states. Inflammation, endothelial function, and platelet reactivity interact in contributing to cardiovascular disease and form the mechanistic basis for targeting treatments and preventive strategies. In physiologic and pathophysiologic states, cellular activation releases arachidonic acid via phospholipase cleavage of membrane phospholipids (Figure 1). Arachidonic acid is the key polyunsaturated fatty acid metabolized to the precursor prostaglandin (PG) endoperoxides, PGG₂ and PGH₂, by cyclooxygenases COX-1 in platelets or COX-2 in the vasculature. PGH₂ is further converted to bisenoic prostanoids via specific synthases and isomerases. The most relevant prostanoids for platelet function are thromboxane A₂ (TxA₂), PGD₂, PGE₂, and prostacyclin (PGI₂). Originating from arachidonic acid and PG endoperoxides, prostanoids exhibit subtle differences in chemical structure that confer striking differences in biological responses attributable to characteristic specificity and activity profiles as ligands for the diverse family of prostanoid receptors. Prostanoids also exhibit a broad range of chemical stability and differences in susceptibility to and consequences of metabolism. The varieties and amounts of prostanoids produced vary depending on the specific cell type and activation status. Cell types that generate prostanoids in the vascular system include endothelial cells, vascular smooth muscle cells, platelets, leukocytes, mast cells, and macrophages. Prostanoids play key roles in vascular tone, inflammation, and thrombosis-hemostasis.

Fig. 1. Prostanoid Metabolism.

Arachidonic acid is released from cellular membrane phospholipids by phospholipase A2. Shown here is the metabolism by the cyclic pathway to prostaglandins and thromboxane. Leukotrienes are also derived from arachidonic acid by a separate linear pathway (not shown for clarity).

Thromboxane A₂

TxA₂ is the predominant prostanoid generated in activated platelets and has been studied extensively. Though chemically among the most unstable prostanoids, having a half-life of less than one minute, it acts locally via the TP_α receptor as an autocoid to potently activate platelets, stimulate platelet shape change, release platelet granule contents, and produce irreversible platelet aggregation. Besides its effects on platelets, TxA₂ mediates a number of physiologic and pathophysiologic responses, including vasoconstriction, inflammation, and angiogenesis [3]. TxA₂ is among the most potent known vasoconstrictors of both venous and arterial smooth muscle. Evidence suggests a significant, but not fully characterized function in mediating inflammatory responses. In septic shock models, TxA2 may be protective by reducing systemic vasodilation [4], but at the same time has been shown to contribute to renal dysfunction [5]. TxA₂ also has a role in angiogenesis and endothelial cell differentiation and proliferation. A TxA₂ mimetic that acts on the TP_β receptor was shown to inhibit VEGF-mediated endothelial cell migration and angiogenesis [6]. However, the TxA₂ agonist U46,619 stimulated IL-1beta-induced angiogenesis. Whereas the preferential agonist of the TP receptor is TxA2, the metabolic precursors of TxA₂, PGG₂ and PGH₂, are also TP agonists and thus may produce similar responses.

Excess thromboxane, measured as inactive metabolites including TxB₂, has been shown to contribute to thrombosis and myocardial ischemic events [7]. Deficiencies in thromboxane signaling have been associated with bleeding disorders [8, 9]. Irreversible inhibition of COX-mediated PG endoperoxide and TxA₂ production, primarily in platelets, is the key action of aspirin, leading to its benefit in reducing the risk of acute coronary syndromes and stroke. TxA₂ synthase inhibitors were developed as potential alternatives to aspirin that may have more favorable physiological properties. It was found that such drugs not only decrease TxA₂ formation, but they also favor PGI₂, PGE₂, PGF_2α, and PGD₂ production [10-13]. A limitation of the clinical use of thromboxane synthase inhibitors is that while TxA₂ production is reduced and platelet inhibitory PGI₂ and PGD₂ are increased, there is an opposing increase in proaggregatory cyclic endoperoxides [11, 12]. In particular, the parent PGH₂, which binds the TP receptor, accumulates in this setting. Therefore, it is understandable why the net anti-platelet effect of the TxA₂ synthase antagonist dazoxiben was minimal, unless combined with a TP antagonist [12]. Several TxA₂ synthase inhibitors and TxA₂ / PG endoperoxide (TP) receptor antagonists have been advanced to clinical development, but not one has been adopted into clinical use in the United States [14-16]. The TP receptor antagonist, terutroban, was compared to, and found to be no different from, aspirin in a large clinical trial (PERFORM) for prevention of stroke in patients with prior stroke or transient ischemic attacks [17].

Prostacyclin

Prostacyclin is the primary prostaglandin product of endothelial cells. Though chemically more stable than TxA₂, prostacyclin non-enzymatically converts to an inactive product, 6-keto-PGF_1α. Prostacyclin enhances microvascular blood flow by potent vasodilation. Through IP receptor mediated activation of platelet adenylate cyclase, prostacyclin impairs platelet adhesion and physiologically antagonizes platelet aggregation in response to all agonists as well as to vascular injury [18]. The counterbalancing actions of stimulatory thromboxane and inhibitory prostacyclin are thought, in part, to regulate systemic platelet activation, limit bleeding, and prevent thrombus overgrowth. In general, the potent hemodynamic effects of IP receptor agonists limit their clinical utility as antiplatelet agents. The stable prostacyclin mimetics epoprostenol, treprostinil, and iloprost, are used as vasodilators in patients with pulmonary arterial hypertension [19], and iloprost has been explored as a potential therapy in scleroderma and Raynaud's phenomenon [20]. Iloprost has also been studied as a potential benefit in critical limb ischemia [21-25], likely due to both antiplatelet and vasodilatory effects.

Prostaglandin D₂

PGD₂ is the primary prostaglandin produced by mast cells. PGD₂ has predominantly been studied in association with inflammation, particularly its role in asthma and other allergic disorders [26, 27]. More recently, PGD₂ has been linked to androgenic male pattern baldness. Prostaglandin D₂ inhibits hair growth and is elevated in the bald scalp of men with androgenetic alopecia [28]. PGD₂ signals in leukocytes and hair follicles through the DP2 receptor (CRTH2) to induce Th2, eosinophil, and basophil chemotaxis and inhibition of hair growth via G-alpha(i) [28, 29]. An important mechanism for niacin-induced flushing is vasodilation following release of PGD₂. This prompted development of laropiprant, a DP1 receptor antagonist, for use in combination with niacin to optimize niacin-induced cholesterol reduction and prevent niacin-induced facial flushing. However, this combination was abandoned because it caused more harm than good in a large clinical trial (HPS2-THRIVE) [30]. PGD₂ may also play roles in the central nervous system, including effects on sleep, body temperature regulation, and nociception.

PGD₂ interacts with DP1 receptors on both platelets and vascular smooth muscle cells that couple to G-alpha(s) to stimulate adenylate cyclase, which inhibits platelet aggregation and causes vasodilation [31]. However, the clinical importance of PGD₂ as an inhibitor of platelet function is not currently known, though laropiprant did not produce an overt prothrombotic effect.

Prostaglandin E₂

PGE₂ is produced in several cell types and contributes to homeostasis in the GI tract, kidney, immune system, and vasculature. PGE₂ is produced from PGH₂ by PGE synthase or spontaneous rearrangement of PGH₂ in aqueous solution [32]. PGE₂ is the major prostaglandin secreted by cultured vascular smooth muscle cells and macrophages. PGE₂ has been found to be the predominant prostanoid from endothelial cells in human microvessels [33] and rabbit coronary microvessels [34], and it likely plays a significant role in angiogenesis [35] and vascular tone [36, 37]. PGE₂ is also present in atherosclerotic plaque [38, 39]. Resident macrophages in atherosclerotic plaque constitutively express COX-1 and make small amounts of PGE₂ at baseline. Induction of COX-2 with lipopolysaccharide was shown to preferentially increase production of PGE₂. High nanomolar concentrations of PGE₂ produced by activated macrophages are implicated in cardiovascular pathologies such as plaque rupture and abdominal aortic aneurysm formation [40]. COX-2 and PGE synthase co-localize in symptomatic carotid plaques, and their expression correlates with the extent of inflammatory infiltration and matrix metalloproteinase activity [41]. PGE₂ derived from atherosclerotic plaque can exit the plaque and act directly on platelets [39]. Activated platelets themselves produce a large amount of PGE₂ [42]. The significant presence of PGE₂ in platelets, microvasculature, atherosclerotic plaque, and activated macrophages makes it natural to extrapolate that PGE₂ contributes importantly to vascular physiology and the pathophysiology of thrombosis. The remainder of the review will focus on the evolution to the current understanding of the role of PGE₂ in regulating platelet function.

Cyclic AMP role in platelet activation

Platelet homeostasis has long been ascribed to a balance of mechanisms that either raise or decrease cyclic adenosine 3'5'-monophosphate (cAMP) [43]. Prior to the discovery of the various prostaglandin receptors, Ashby et al. hypothesized that prostaglandins exert their effects via either stimulation or inhibition of cAMP production [44, 45]. While it is now recognized that cAMP is but one of many factors in the platelet regulatory pathway, a general appreciation of the role of cAMP is essential to understanding how prostaglandins, including PGE₂, modulate platelet activity. cAMP was shown to mediate the stronger antiplatelet effect of TxA₂ synthase inhibitors observed in pathologies associated with in vivo platelet activation [46]. An increase in cAMP is associated with platelet inhibition, and a decrease in cAMP promotes platelet aggregation induced by calcium mobilization[43] [47]. cAMP levels can be regulated by agents that either enhance production via adenylate cyclase or decrease its metabolism via cAMP phosphodiesterases. In general, G_s-coupled receptors stimulate adenylate cyclase, raising cAMP and inhibiting platelet function. For example, PGI₂ inhibits platelet aggregation by increasing intracellular cAMP via activation of its G_s-coupled IP receptor. In contrast, G_i-coupled receptors, such as the ADP receptor, P2Y12, inhibit adenylate cyclase, decrease cAMP, and facilitate platelet aggregation. Importantly, Gi receptor stimulation alone does not directly induce platelet aggregation, but acts synergistically with other receptors that induce mobilization of intracellular calcium, such as the G_q-coupled thromboxane receptor, TP [48, 49]

PGE₂ acts via multiple receptors

Initially it was thought that PGE₂ may act via the PGI₂ / IP receptor; however, it was later determined that PGE₂ preferably activates its own specific receptors [50]. PGE₂ has four receptor subtypes that were identified and subsequently cloned [51-53], termed EP1, EP2, EP3 and EP4. The EP receptors are present in numerous tissues, and the distribution of EP receptor subtypes varies among different tissues [54]. Eggerman et al. first showed that PGE₂ has a receptor on human platelets that is distinct from prostacyclin's receptor [50]. Using RT-PCR and Southern blot, Paul et al. showed that the EP3 and EP4 receptors are much more prominent in human platelets than the EP2 receptor [55]. The EP1 receptor is sparse in most tissues compared to the other EP receptor subtypes [54], and it has not been detected in human platelets thus far. These G protein coupled receptors differ in structure and signal transduction coupling (Figure 2). In order to understand the role of PGE₂ in regulating platelet activity, it is necessary to delve into the specific function of the different EP receptor subtypes.

Fig. 2. EP receptors and signaling pathways.

PGE₂ binds to four receptor subtypes: EP1, EP2, EP3, and EP4. Each receptor has distinct signaling pathways depending on the cell type in which it is expressed. All major mediators and second messengers are shown for each receptor subtype. The details of these pathways continue to be refined.

Elucidating EP receptor function with specific agonists/antagonists

Much research on the role of PGE₂ in platelet function has focused on the use of EP receptor subtype-selective agonists and antagonists. Frequently used selective EP receptor agonists include butaprost (EP2) [56], sulprostone (EP3) [57], and PGE1-OH (EP4) [57]. Iloprost is an EP1 agonist, but is poorly selective [57]. Newer synthetic EP receptor agonists include ONO-DI-004 (EP1) [58], ONO-AE1-259 (EP2) [58, 59], ONO-AE-248 (EP3) [60], and ONO-AE1-329 (EP4) [60]. Novel EP receptor antagonists include ONO-8713 (EP1) [58], ONO-AE3-240 (EP3) [61], DG-041 (EP3) [62], ONO-AE208 (EP4) [63], and MF-191 (EP4) [56].

EP1 Receptor Biochemistry and Function

The EP1 receptor acts predominantly via G_q, activating phospholipase C, protein kinase C, and releasing intracellular calcium [54]. However, it does not seem that EP1 is expressed in human platelets, as neither the selective EP1 agonist ONO-DI-004 nor the EP1 antagonist ONO-8713 have any effect on platelet aggregation induced by platelet activating factor (PAF) [58].

EP2 Receptor Biochemistry and Function

The EP2 receptor couples to G_s, leading to increased production of cAMP. Thus, EP2 stimulation leads to inhibition of platelet aggregation. The selective EP2 agonist ONO-AE1-259 inhibits platelet aggregation [58, 59]. This inhibitory effect is also seen with the EP2 agonist butaprost [56].

EP3 Receptor Biochemistry and Function

The EP3 receptor is more complex than the other EP receptors, such that six isoforms have been identified [53, 64-66]. These isoforms result from alternative mRNA splicing, which alters the cytoplasmic domain and signal transduction pathways. Although EP3 was originally described to couple to G_i leading to reduction of intracellular cAMP [67], subsequent evidence suggests that different isoforms can variably increase cAMP via G_s or IP3 via G_q [68, 69]. Recent evidence shows that EP3 also can couple to G_αz [70, 71].

Stimulation of EP3 via the EP3-specific agonist sulprostone does not directly lead to platelet aggregation, but potentiates aggregation in response to agonists such as ADP and U46,619 in human [38, 56, 62, 72] and murine [73] platelet rich plasma. Additionally, PGE₂-mediated stimulation of EP3 decreased cAMP, consistent with its primary signaling in platelets being coupled to G_i [73]. The selective EP3 antagonist DG-041 abolishes the pro-aggregatory effects of both sulprostone [58] and low concentrations of PGE₂ [56, 63, 74].

EP4 Receptor Biochemistry and Function

Similar to the EP2 receptor, the EP4 receptor was initially characterized as a G_s-coupled receptor, which stimulates adenylate cyclase and cAMP production [75]. These cAMP-dependent pathways can involve Protein Kinase A (PKA), Epac (i.e. exchange protein activated by cAMP), and AMP Kinase (AMPK) [75-77]. Subsequent studies by Fujino and Regan have shown that stimulation of EP4 evokes coupling to a pertussis toxin-sensitive G_αi/0 leading to PI3-kinase-dependent activities [78-80]. Takayama and Libby described a novel EP4 receptor-associated protein (EPRAP) that participates in anti-inflammatory signaling in humans [81].

The EP4 agonist AE1-329 inhibited platelet aggregation induced by platelet activating factor (PAF), ADP, and collagen [58, 59, 82]. AE1-329 also prevented platelets from binding fibrinogen through the GPIIb/IIIa receptor [82]. Similarly, we have shown that inhibition of EP4 with a selective antagonist MF-191 completely abrogated the effect of PGE₂ on inhibiting U46,619-induced platelet aggregation [56]. While this inhibitory function is in keeping with its known coupling to G_s, other EP4-mediated signaling pathways could contribute to PGE₂ signaling in platelets. For example, when activated by PGE₂, EP4 can also recruit arrestins [83], leading to c-Src activation, a signaling pathway independent of G protein activation [84]. Interestingly, Arrestin 2 was also shown to regulate the PAR4-dependent signaling pathway in human platelets [85]. Thus, the complexity of the EP4 signaling pathways suggests that EP4 might function through both cAMP-dependent and -independent pathways.

Deciphering the physiologic role of PGE₂ in platelets

Early Studies – Conflicting results and biphasic concentration-dependent response

Early studies into the role of PGE₂ in regulating platelet function in humans have produced conflicting results. A consistent theme among studies is the finding that PGE₂ alone, without an additional agonist, does not lead to platelet aggregation. Thus, determining the effects of PGE₂ on platelet aggregation must be done with PGE₂ in combination with other agonists. An early study by Kloeze and colleagues showed that PGE₂ at concentrations of 5-10 μmol/L inhibited platelet aggregation induced by ADP [86]. Shio et al. further characterized two phases of aggregation, an initial phase, followed by a plateau, and then a secondary phase. They found that PGE₂ (50-200 μg/L = 0.14-0.57 μmol/L) decreased the primary phase of ADP-induced aggregation in all individuals, while it enhanced the secondary phase of aggregation in some individuals [87]. Though the same primary phase inhibition with second phase potentiation was also reported by McDonald et al. [88], enhancement of the primary phase of aggregation was described by Andersen et al. [89].

Whereas early investigations into the effects of low concentrations of PGE₂ on platelet aggregation were conflicting, studies using higher doses produced more consistent results. Heptinstall and Gray described a bimodal effect of human platelets to PGE₂, whereby it potentiated platelet aggregation at lower concentrations (0.1-10 μmol/L) and inhibited aggregation at a much higher concentration (100 μmol/L) [90, 91]. Such a biphasic effect of PGE₂ was demonstrated in multiple other studies [87, 92-94]. This biphasic response was hypothesized to be due to inhibition of cAMP production at low concentrations and stimulation of cAMP production at higher concentrations. Indeed, Robison et al. showed that cAMP levels were raised in human platelet rich plasma when PGE₂ was delivered at a high concentration known to inhibit ADP-induced platelet aggregation [95]. The studies of Ashby demonstrated that low dose PGE₂-mediated platelet aggregation was associated with decreased levels of cAMP [96]. Thus, well before the EP receptors were discovered, it was postulated that PGE₂ may act through multiple receptors that have opposing effects. It was thought that low concentrations of PGE₂ acted via a G_i-coupled receptor, and higher concentrations bound to a G_s-coupled receptor, such as the PGI₂ receptor [90].

High doses of PGE₂ bind the PGD₂ Receptor

Both the PGI₂ receptor (IP) and PGD₂ receptor (DP1) mediate platelet inhibition, and it was hypothesized that high concentrations of PGE₂ could interact with one of these receptors. PGE₂ has a relatively low affinity for IP in humans, and selective inhibition of IP does not prevent the inhibitory effects of PGE₂ [58]. PGE₂ does act, however, on the DP1 receptor [2]. The affinity of PGE₂ for DP1 has been shown to be at a Ki of 307 +/− 106 nmol/L [97]. This Ki predicts that PGE₂ concentrations above 500 nmol/L will activate DP1.

Studies in mice – important data, but limited translation to humans

In part because of the high variability in the response of human platelets to PGE₂, particularly at low doses, investigators extensively studied the effect of PGE₂ on platelets using mice. In contrast to humans, low concentrations of PGE₂ consistently have a pro-aggregatory effect in mice. Concentrations of PGE₂ up to 100 μmol/L potentiate platelet aggregation induced by sub-efficacious concentrations of a thromboxane receptor agonist, U46,619, and by collagen (18-21).

Within many tissue types, including platelets, the presence of the EP receptors differs in mice compared to humans. In murine platelets, EP1 is essentially absent, and the expression levels of EP3 are over 200-fold higher than EP4 and over 1000-fold higher than EP2 [98]. Therefore, the relative contribution of EP2 and EP4 pathways to PGE₂ effects on platelets is negligible in mice. For example, the dose of the EP2 specific agonist butaprost required for platelet inhibition is much higher in mice than humans [IC₅₀ 87 ± 3 μmol/L vs. 2.57 ± 0.82 μmol/L, respectively) [58] [59]. Likewise, the EP4 agonist AE1-329 had a much less robust inhibitory effect in mouse platelets compared to human platelets [58, 59, 82]. EP3 activation almost entirely drives the effect of PGE₂ in mice, explaining why PGE₂ has no detectable effect on platelets in EP3 ⁻/⁻ knockout mice [73].

Similar to humans, very high doses of PGE₂ (greater than 1 μmol/L) in mice inhibit platelets [73]. However, in mice, this inhibitory effect is not mediated by PGE₂ binding of DP1, as this receptor is not present in murine platelets [99]. Rather, high doses of PGE₂ bind the murine prostacyclin receptor IP [73]. Thus, for numerous reasons the findings in mice cannot be extrapolated directly to humans.

As PGE₂ mediates in vitro platelet aggregation in mice via EP3, investigators explored how modulating EP3 may contribute to in vivo thrombosis. Mice deficient in EP3 had decreased susceptibility to atherothrombosis triggered by intravenous arachidonic acid infusion [98] or mechanical plaque rupture [39]. Tilly et al showed that blocking EP3 in mice with the EP3 antagonist DG-041 inhibited atherothrombosis in vivo caused by ferric chloride or arachidonic acid exposure followed by mechanical plaque disruption [100]. Whereas mice that completely lack functional EP3 demonstrate increased in vivo bleeding times [98], mice treated with DG-041 have preserved hemostasis [100, 101].

Biased agonism – PGE₂ and EP specific agonists may have different effects

Early studies in humans were done without the understanding that human platelets express three different PGE₂ receptors. The discovery of the EP receptor subtypes launched a number of investigations into how the diverse effects of each receptor mediate the overall response to PGE₂. Most of these studies were performed using EP receptor specific agonists, and the results were extrapolated to determine the effects of PGE₂ itself. However, our data suggest that activation of EP2 in human platelets by its selective agonist butaprost or by PGE₂ itself may have different effects: although butaprost inhibits platelet aggregation induced by U46,619, PGE₂ does not affect platelet aggregation when both EP3 and EP4 are blocked with their respective antagonists (DG-041 and MF-191) [56]. These results are in keeping with the growing evidence suggesting that synthetic ligands of a target GPCR may activate distinct pathways, leading to coupling of the receptor to different intracellular effectors [102, 103] in a process termed “biased agonism” [104, 105]. Similar to EP2, differences based on “biased agonism” have recently been illustrated with subtype selective ligands of EP4 [106].

Current understanding - PGE₂ response is determined by a relative effect of EP receptor activation

In 2010, we and others reported that the net effect of PGE₂ in human platelets reflects a balance between the potentiating effects mediated by EP3 and the inhibitory effects mediated by EP4 [56, 58, 59, 107]. These data suggest that PGE₂ may contribute to platelet homeostasis by modulating the overall response to other activators and inhibitors. The following section will discuss how inter-individual phenotypic differences in the effects of low dose PGE₂ initially hindered the field, but eventually led to an improved understanding of the role of PGE₂.

Phenotypic differences in PGE₂ effect on human platelets

Despite the conclusion by many investigators that low dose PGE₂ promotes aggregation, a number of studies have noted variable responses among human subjects [72, 87, 91, 108, 109]. In the study by Shio et al., certain subjects seemed to lack the second phase of potentiation, such that the overall response to PGE₂ was only inhibition [87]. Gresele et al. showed that PGE₂ promoted platelet aggregation to arachidonic acid in patients deemed “nonresponders” to thromboxane synthase inhibitors, but inhibited aggregation in those termed “responders” [109]. At that time, there was significant interest in the development of thromboxane synthase inhibitors, and several other investigators described the phenomenon of responders and non-responders to these drugs [110-113]. Matthews et al. showed that PGE₂ (0.01-5 μmol/L) inhibited aggregation in response to ADP in 2 out of 8 blood donors and potentiated aggregation in the other 6 donors [72]. This inter-individual variability in humans remained unexplained, but was suggested to be due to the differential “response of their adenylate cyclase to PGE₂” [109] or variation in the ratio of PGD₂ to TxA₂ + PGE₂ produced [112].

Our group first described in 2010 that the effects of PGE₂ at low concentration (100 nmol/L) on human platelets segregates into different phenotypic groups [56]. Since then, we have refined our characterization of these groups and clarified the terminology. The phenotypic groups are: inhibitory (PGE₂ inhibits platelet aggregation), and potentiating (PGE₂ permits or potentiates aggregation) (Figure 3). Testing 104 healthy volunteers revealed that 45% have the inhibitory phenotype and 55% have the potentiating phenotype. This variable PGE₂ effect was lost when platelets were activated with higher concentrations of agonist that elicit full activation [56]. Importantly, the volunteers have been tested on multiple occasions over several years without variation in the phenotypic group assignment, demonstrating reproducibility and suggesting independence from environmental factors. We have also analyzed patients with coronary artery disease who are receiving dual antiplatelet therapy (aspirin and P2Y12 inhibitors) and have shown that this treatment does not affect phenotypic determination.

Fig. 3. Representative traces of the two phenotypic groups analyzed by light transmission aggregometry.

Healthy volunteers or patients undergoing PCI were consented following a protocol approved by the Vanderbilt Institutional Review Board. Platelet rich plasma (PRP) was obtained from citrated blood, and platelet count was adjusted to 250,000 cells per μl with autologous platelet poor plasma (PPP). PRP was preincubated with PGE₂ 100 nmol/L (PGE₂) or vehicle (control) for 30 s, followed by a sub-maximal concentration of U46,619. Aggregation was recorded for 6 min as described by Smith et al [56]. The frequency of each phenotypic group is indicated (n=104).

PGE₂ has the same inhibitory or potentiating effect on platelet aggregation whether in the presence of agonist for the thromboxane receptor [56] or the thrombin receptor protease activated receptor type 4 (PAR4) (Figure 4). In contrast, PGE₂ has no effect on platelets activated with a protease activated receptor type 1 (PAR1) specific agonist peptide (Figure 4). Of note, MacIntyre and Gordon showed that PGE₂ had no effect on platelet aggregation induced by collagen in citrated platelet rich plasma [114], results that we reproduced (data not shown). These data outline the complexity of the regulation of platelet aggregation by PGE₂ and reinforce the need to better understand the mechanism responsible for the two phenoptypes.

Fig. 4. Effect of PGE2 on platelet aggregation induced by agonist peptides specific for the two thrombin receptors expressed on human platelets.

Healthy volunteers were consented following a protocol approved by the Vanderbilt Institutional Review Board. PRP was obtained from citrated blood, and platelet count was adjusted to 250,000 cells per μl with autologous platelet poor plasma (PPP). 250 μl PRP was preincubated with vehicle (−) or with 100 nM PGE₂ (+) for 30 s, followed by sub-maximal concentrations of agonist peptides specific for PAR1 (PAR1-AP) or PAR4 (PAR4-AP). Aggregation was recorded for 6 min as described previously [56]. Values represent mean ± S.E.M (n = 3). Statistical significance was evaluated by paired t-test.

We observed that all individuals express EP2, EP3, and EP4, and that these receptors respond similarly to their specific agonists butaprost, sulprostone, and PGE1-OH, respectively [56]. However, the net result of activation of these three receptors by their endogenous agonist, PGE₂, leads to the different phenotypes, suggesting that there are differences in specific signaling pathways. Importantly, our data showed that an EP3 antagonist, DG-041, converts subjects from the potentiating to the inhibitory phenotype [56] (Figure 5). Further work is necessary to determine the signaling mechanisms by which PGE₂ affects platelet aggregation in each phenotypic group.

Fig. 5. Effect of PGE₂ on platelet activation in humans.

The net effect of PGE₂ reflects the differential activation of the two receptors EP3 and EP4. EP3 antagonists shift phenotype from potentiating to inhibitory.

Several reasons may explain why the two different PGE₂ response phenotypes were not recognized and characterized earlier: 1) The presence of the different EP receptor subtypes in human platelets was not initially appreciated; 2) Much of the data was obtained using murine models, in which the PGE₂ response results entirely from EP3 activation; 3) Many experiments have been conducted with higher concentrations of agonists that cause full platelet activation, a condition in which PGE₂ has no effect; and 4) Most recent experiments with human platelets have been performed using EP-specific agonists, rather than PGE₂, possibly leading to “biased agonist” effects.

Clinical Implications

Risk for cardiovascular events

It is understood that vascular injury stimulates platelet adhesion in part via activation of platelet glycoproteins, followed by release of chemical mediators, such as thrombin, ADP, TxA₂, and norepinephrine that recruit more platelets to the site of the wound. However, a gradient of agonist forms around the thrombus, and platelets circulating in the low concentration part of the gradient may not be recruited to the wound, but would be partially activated, making them more “sticky.” These “sticky” platelets express adhesion molecules at their surface and may contribute to pathologic thrombus formation [115]. PGE₂ may be an underappreciated mediator, generated from the macrophages in atherosclerotic plaques and activated platelets, that plays a role in regulation of platelet reactivity. We have shown that low concentrations of PGE₂, such as those generated locally by activated platelets, can exert either inhibitory or potentiating effects. In individuals with the inhibitory phenotype, the local PGE₂ could slow the growth of the associated thrombus by increasing the threshold for activation of circulating platelets in the proximity of the atheroma (Figure 5). This inhibition would be mediated by the effects of PGE₂ on EP4. In contrast, in individuals with the potentiating phenotype, platelet aggregation may be enhanced via the potentiating EP3 receptors without effective inhibition through EP4 receptors. Thus, we would postulate that in the setting of atherosclerotic plaque rupture, individuals with the potentiating phenotype might be at higher risk of thrombotic cardiovascular events including myocardial infarction or stroke. We further would postulate that phenotyping individuals for their platelet responses to PGE₂ could serve as a risk prediction tool that might also guide treatment selectively for those at higher risk.

Potential for novel therapeutic options

Increasing platelets threshold to agonist stimulation could constitute an appealing new therapeutic approach in the management of cardiovascular thrombotic events [115], as it might inhibit thrombosis without affecting hemostasis. Proposed targets for intervention include the EP3 and EP4 receptors via EP3 antagonists and EP4 agonists.

Based on encouraging preclinical studies in mice [100, 101], the in vivo effects of EP3 antagonism have been investigated in human subjects. Tilly et al. treated 10 healthy volunteers with DG-041 for 7 days and found that the drug reduced platelet aggregation to collagen, while not affecting cutaneous bleeding time [100]. In a separate study by Fox et al., healthy volunteers were given DG-041 alone or in various combinations with clopidogrel and aspirin [116]. The study showed that DG-041 neither increased bleeding time alone nor further increased bleeding time to clopidogrel. At the same time, the effect of DG-041 further decreased platelet aggregation. In 2006, deCODE Genetics (Reykjavik, Iceland) began a Phase IIa randomized, placebo-controlled clinical trial of DG-041 in 144 patients with peripheral arterial disease, examining the safety and tolerability of the drug and its effect on various markers of disease severity. While the results have not been published, the company internally reported no serious adverse events and favorable effects on inflammatory biomarkers, such as c-reactive protein and monocyte chemotactic protein 1, as well as ankle-brachial index measurements [117].

While these studies are very promising, Schober et al. cautioned that the concentration of PGE₂ in human atherosclerotic plaques is low, which may render the effects of PGE₂ in vessels minimal. In their study PGE₂ from plaque homogenates had no effect on platelet aggregation [38]. Furthermore, the EP3 antagonists AE5-599 and AE3-240 did not prevent platelet aggregation, dense granule release, or GP IIb/IIIa exposure induced by plaque homogenates. These results are important as they indicate that more studies are needed to understand fully the effect of PGE₂ on atherothrombosis in humans. However, our data predict that plaque homogenates should not be used as platelet agonist: 1) plaque homogenate contains collagen and thrombin and other agonists that are not regulated by PGE₂. Moreover, the concentration of these agonists in the assay must have elicited full platelet activation, a condition in which PGE₂ has no effect. 2) The platelet donor must be phenotyped for PGE₂ response prior to doing the experiment to make sure that the data can be interpreted correctly: a donor with the potentiating phenotype would fail to show platelet inhibition. 3) PGE₂ half-life in biological milieu such as blood or plasma is in the minute range, and the cellular machinery generating it in plaques is disturbed during homogenization, suggesting that there might have been no native PGE₂ left in the homogenate. In studying the complex downstream effects of PGE_2, it is essential to have a precise and sophisticated experimental design, which takes into account the concentration-sensitive effects of the molecule and the interplay with the variable characteristics of the donor platelets.

In addition to the clinical use of an EP3 receptor antagonist, a selective EP4 receptor agonist has been proposed to be of clinical utility. Apart from its effects on platelet aggregation, PGE₂ via EP4 may have beneficial effects on reducing inflammation and myocardial injury in the setting of ischemia [118, 119]. In vitro antithrombotic effects of the EP4 agonist AE1-329 have been seen using platelet aggregation assays [58, 59, 82] and flow chamber experiments that studied thrombus formation in human blood passed over a collagen column [82]. Indeed, platelet inhibition to AE1-329 was similar to the effects of aspirin and was synergistic in combination [82]. Whether the EP4 agonist increases bleeding time is not known, and in vivo studies in mice cannot be relied upon to predict effects in humans for the reasons described above. Additionally, as the EP4 receptor plays a regulatory role in numerous tissues, the non-platelet effects of systemic administration would need to be closely monitored. For example, in an animal model of type 1 diabetes, an EP4 agonist exerted negative pro-inflammatory effects on the kidney leading to fibrosis and albuminuria [120]. EP4 agonists also may promote tumor progression and metastasis in various malignancies [121-123]. Nevertheless, an EP4 agonist was explored as a potential therapy for ulcerative colitis in a phase II clinical trial, and the safety profile was favorable [124].

Platelets and Tumor Growth

It is becoming increasingly recognized that there is a significant interaction between tumor cells and platelets. Platelet activation in cancer patients increases the risk of thrombosis, a major complication of cancer [125]. More recent data has shown that platelets may contribute to metastasis by inducing epithelial-mesenchymal transition, vascular remodeling, endothelial adhdesion, and even immune evasion [126-128]. Platelets recently were shown possibly to function as a chemo-attractant to cancer cells [129].

Platelets can also induce COX-2 overexpression in cancer cells themselves, demonstrated in a colon cancer cell line [130]. Subsequently, in both platelets and tumor cells, PGE₂ is thought to contribute to a variety of signaling pathways that may ultimately lead to tumor growth and metastasis. Inhibition of COX via aspirin or NSAIDs has already been explored as agents to inhibit this cascade [131]. Specific antagonists that act on the EP receptors have been proposed to be novel anti-tumor agents in colorectal and breast cancer [132-134].

Conclusions

The critical need to elucidate platelet biology has driven over 40 years of research into the function of PGE₂. PGE₂ can have either stimulatory or inhibitory roles in human platelets depending on experimental conditions, and individuals segregate into different phenotypic groups exhibiting either PGE₂-mediated potentiation or inhibition of platelet aggregation. The discovery of the EP receptors launched a series of investigations into their effects in platelets. Using specific agonists, it was shown that the EP3 receptor promotes platelet aggregation and that the EP4 and EP2 receptors inhibit platelet aggregation. Investigators including our group and Heptinstall et al. have shown that the effect of PGE₂ on human platelets is determined by the net activation of EP4 and EP3 [56, 58]. Since mice functionally only express EP3 in platelets, murine studies cannot be expected to predict outcomes in humans. In addition, biased agonism must be appreciated as a potential confounder when designing experiments with synthetic agonists to study the effect of PGE₂. Taken together, these considerations demand that we standardize the experimental conditions when studying the effect of PGE₂ on platelet function (Table 1).

Table 1.

Proposed Standardization of PGE₂ Platelet Assays

Key Points
• Studies must be performed on human platelets, because the relative expression of EP receptor subtypes in mice is different.
• PGE2 must be utilized rather than EP-specific agonists due to the potential for biased agonism; the contribution of each receptor sub-type can be studied by blocking the other sub-types with highly selective antagonists.
• The concentration of PGE2 must not exceed 100-200 nmol/L to ensure that it does not stimulate DP1 receptors.
• The concentration of agonist used to stimulate platelet aggregation must be sub-maximal and determined for each individual by constructing a dose-response relationship with and without PGE2.
• Recognition of the presence of two human phenotypic groups should be reflected in the data analysis and its interpretation.

We have made the seminal observation that humans segregate into two distinct platelet response phenotypes: 1) inhibitory phenotype, whereby a low concentration of PGE₂ increases the activation threshold for agonists; 2) potentiating phenotype, whereby a low concentration of PGE₂ decreases the activation threshold for agonists. The presence of these phenotypes may explain much of the inter-individual variability in PGE₂-mediated responses that has been described over the years, including the variable responses to TxA₂ synthase inhibitors. When performing and analyzing human platelet function studies, especially those involving prostanoids, platelets with different phenotypes may be best studied with slightly different protocols and analyzed separately to avoid confounding results.

Finally, persistent on-treatment platelet hyper-reactivity remains a challenge in the management of cardiovascular diseases, as it is correlated with adverse clinical outcomes. There is great interest in developing novel antiplatelet agents that allow true personalized or tailored antiplatelet therapy. Thus far, studies using point-of-care platelet function testing have failed to show benefit to this approach. DG-041, an EP3 antagonist, showed promise in its ability to inhibit platelet function while not increasing bleeding. The use of an EP3 receptor antagonist in conjunction with knowledge of the individual's PGE₂ response phenotype offers a simple and elegant approach to personalize anti-platelet therapy. An EP4 agonist also may inhibit platelet function, but the effects on bleeding and on normal function in other tissues are not known. Further research into the biological properties of the EP receptors and their signaling pathways that underlie the inhibitory and potentiating phenotypes may create additional options for safe and effective treatment of cardiovascular diseases.

Highlights.

Human platelet's response to PGE₂ segregates in two phenotypic groups

PGE₂ effects on mouse platelets cannot be translated to human platelets

Studies must use PGE₂ rather than receptor-specific agonist because of biased agonism

Published data in platelets are inconsistent from lack of standardized protocols

PGE₂ receptors could be new targets for antiplatelet drugs that preserve hemostasis

Fig. 6. Schematic representation of a blood vessel demonstrating the hypothetical effects of PGE₂ on thrombus growth in inhibitory and potentiating phenotypes.

At the site of atherosclerotic plaque rupture, a gradient of agonist is released, and platelets are recruited to form an initial plug. PGE₂ is produced by macrophages in the plaque as well as by activated platelets. The inhibitory phenotype is accompanied by a increased threshold for platelet aggregation, preventing platelet hyperreactivity. In contrast, the potentiating phenotype is associated with a decreased threshold for platelet aggregation, promoting formation of an occlusive thrombus.

Abbreviations

AMPK

AMP Kinase

cAMP

cyclic adenosine 3'5'-monophosphate

COX

Cyclooxygenase

PAF

Platelet activating factor

PG

Prostaglandin

DP

Prostaglandin D₂ receptor

EP

Prostaglandin E₂ receptor

PKA

Protein Kinase A

Tx

Thromboxane