Procoagulant platelets: novel players in thromboinflammation

Abstract

Platelets play a key role in maintaining hemostasis. However, dysregulated platelet activation can lead to pathological thrombosis or bleeding. Once a platelet gets activated, it will either become an aggregatory platelet or eventually a procoagulant platelet with both types playing distinct roles in thrombosis and hemostasis. Although aggregatory platelets have been extensively studied, procoagulant platelets have only recently come into the spotlight. Procoagulant platelets are a subpopulation of highly activated platelets that express phosphatidylserine and P-selectin on their surface, allowing for coagulation factors to bind and thrombin to be generated. In recent years, novel roles for procoagulant platelets have been identified and they have increasingly been implicated in thromboinflammatory diseases. Here, we provide an up-to-date review on the mechanisms resulting in the formation of procoagulant platelets and how they contribute to hemostasis, thrombosis, and thromboinflammation.

PROCOAGULANT PLATELETS: WHAT IS IN A NAME?

Upon vessel wall trauma, platelets quickly adhere to the site of injury and become activated, resulting in the formation of aggregating and eventually procoagulant platelets. Together, these will cooperate in sealing of the injury site forming a fibrin-rich hemostatic plug. Although the platelet research community has uncovered the mechanisms of aggregating platelets in great detail, the mechanisms leading to and the function of procoagulant platelets remain a matter of debate.

Historically, procoagulant platelets were considered all platelets expressing high levels of phosphatidylserine (PS) on their surface. This allows for coagulation factors to bind and thrombin generation to initiate on the platelet surface (1, 2). There are two ways platelets can express PS on their surface; however, only one of them contributes to thrombosis and hemostasis. Similar to nucleated cells, dying platelets will express PS on their surface, marking them for clearance. This process is dependent on caspases and independent of Ca²⁺. Unique to platelets, strong platelet activation can also result in the expression of PS on the activated platelet surface independent of apoptosis and highly dependent on Ca²⁺ (3). In recent years, evidence emerged demonstrating apoptotic PS exposure has little or no role during hemostasis (4). Hence, the term procoagulant platelets is only used when discussing strongly activated PS-positive platelets. Moreover, procoagulant platelets and apoptotic platelets can easily be distinguished by the presence of high levels of P-selectin expression (or other α-granule proteins) exclusively on the procoagulant platelet surface (5).

MECHANISMS OF PROCOAGULANT PLATELET FORMATION

Quiescent platelets have a delicate plasma membrane lipid bilayer that has negatively charged PS on its inner layer. Homeostasis of the platelet phospholipid bilayer is maintained by ATP-dependent flippases, catalyzing the transport of phospholipids from the outer to the inner layer of the plasma membrane, and scramblases, responsible for the opposite (6, 7). For a platelet to become procoagulant, the balance between flippases and scramblases needs to be tipped in favor of scramblases. This will result in a PS-positive platelet. The key trigger for this is a high enough cytosolic calcium concentration as this will simultaneously activate scramblases (8) and inhibit flippases (9, 10). In general, it is assumed that the exposure of PS on the platelet surface is primarily driven by the scramblase transmembrane protein 16 F (TMEM16F), also known as Anocatamin 6 (Ano6) (11, 12).

Physiological Triggers for Procoagulant Platelet Formation

When procoagulant platelets were initially discovered, it was believed that for platelets to become procoagulant, both protease-activated receptors (PAR1 or PAR4) and immunoreceptor tyrosine-based activation motif (ITAM)-linked receptors [Glycoprotein VI (GPVI); Clec-2 or Fc Gamma RIIa (FcγRIIa)] needed to be activated simultaneously (Fig. 1) (12). However, this seems to mainly be the case in vitro. In vivo, the triggers for platelets to become procoagulant are less stringent and often times single agonist stimulation will suffice. From a signaling perspective, these various pathways will activate intracellular messengers that converge into the activation of distinct phospholipase C (PLC) isoforms (13). PARs signal through PLCβ, whereas ITAMs signal through PLCγ. Next, both PLC isoforms will cleave phosphatidylinositol-4,5-bisphosphate (PIP₂) into inositol trisphosphate (IP₃) and diacylglycerol (DAG). IP₃ interacts with the IP₃ receptor at the surface of the internal store of Ca²⁺, triggering the cytosolic release of Ca²⁺. Of note, these processes are not unique to procoagulant platelet formation and more research is needed to elucidate discrepancies in signaling events contributing to both subtypes of activated platelets.

Figure 1.

Overview of mechanisms of procoagulant platelet formation. Platelet stimulation through agonists that induce ITAM or GPCR signaling in combination with shear or ROS can induce procoagulant platelet formation. These triggers induce a rapid rise in intracellular calcium, which enters the mitochondria. Once excessive calcium enters the mitochondria, CypD acts to form the mPTP, resulting in decrease mitochondrial membrane stability and the release of calcium back into the cytoplasm. The release of calcium through the mPTP activates TMEM16F, a phosphatidylserine (PS) scramblase, resulting in PS exposure on the platelet surface. Pathological formation of procoagulant platelets can induce thrombotic complications such as stroke and pulmonary and venous thrombosis. CypD, cyclophilin D; FcγRIIa, Fc Gamma RIIa; GPCR, G protein-coupled receptor; GPVI, glycoprotein VI; HIT, heparin-induced thrombocytopenia; ITAM, immunoreceptor tyrosine-based activation motif; mPTP, mitochondrial permeability transition pore; PAR, protease-activated receptor; ROS, reactive oxygen species; TMEM16F, transmembrane protein 16F; VITT, vaccine-induced thrombotic thrombocytopenia.

Under physiological conditions (i.e., in the presence of blood flow), platelets binding to collagen is sufficient to induce procoagulant platelets (14, 15). However, in the absence of flow, costimulation with PAR agonists is necessary to achieve similar procoagulant platelet responses. Nevertheless, platelets will almost always be subjected to some degree of shear stress in vivo. In fact, it has been reported that shear stress alone is sufficient to trigger procoagulant platelet formation and PS externalization increases with increasing shear stress (16). This could have implications for prosthetic cardiovascular devices that often induce very high shear stress. However, these results need validation, and the consensus is that rather than inducing procoagulant platelets, shear stress potentiates procoagulant platelet formation. This was elegantly demonstrated both in the absence and presence of platelet adhesion. Using a cone-plate rheometer, it was shown that arterial shear enabled low doses of a single agonist (thrombin or collagen) to induce full-scale procoagulant platelet activity (17, 18). Critically, in the setting of adherent immobilized platelets, low-shear flow conditions are sufficient to propagate procoagulant platelet formation. Mechanistically, this process requires platelet activation by von Willebrand factor (VWF) binding to platelet glycoprotein Ib (GPIb) in the presence of trace amounts of thrombin. Through a local enforcement loop, platelet PS will subsequently activate coagulation factors, which then generate sufficiently high amounts of thrombin to ensure focal fibrin formation. This was compellingly demonstrated by Cosemans et al. (19) using highly specific antibodies against the VWF-GPIb interaction, and plasma and blood from patients with von Willebrand’s disease or VWF-deficient mice. Of note, Nbeal2-deficient mice, lacking α-granules, also have strongly reduced procoagulant platelet responses (20). This indicates α-granules or components thereof (e.g., platelet-VWF) are critically required for platelet coagulant activity under shear stress.

Ca²⁺: Gatekeeper of Procoagulant Platelet Formation

A common feature of platelet activation is a rise in intracellular Ca²⁺ to initiate signaling events leading to either an aggregating or procoagulant platelet (21). Platelets take up Ca²⁺ from their surroundings mainly through Ca²⁺ channels ORAI1 and TRPC6. In conjunction with ORAI1, STIM1 will subsequently fill intracellular Ca²⁺ stores through a process known as store-operated calcium entry (SOCE). In parallel, sarco/endoplasmic reticulum Ca²⁺ ATPases (SERCA) will also fill the platelet Ca²⁺ stores. In platelets, three Ca²⁺ stores have been identified: the dense tubular system (equivalent to sarco/endoplasmic reticulum), lysosome-like acidic organelles, and mitochondria. From these stores, the dense tubular system is considered to be the major intracellular Ca²⁺ pool in platelets. In contrast, mitochondrial Ca²⁺ is considered to predominantly be involved in procoagulant platelet formation, since alterations in mitochondrial Ca²⁺ were reported to have little effect on classical platelet function (22–24). However, more studies are needed to investigate the contribution of mitochondrial Ca²⁺ to aggregatory platelet function.

Although both aggregating and procoagulant platelets require an increase in cytosolic Ca²⁺, procoagulant platelets need sustained levels of supramaximal Ca²⁺, whereas aggregating platelets only require transient increases in Ca²⁺ (24). The high and sustained increase in cytosolic calcium required for PS exposure is only achieved with strong platelet stimulation (combined ITAM and PAR signaling) and requires Ca²⁺ influx from the extracellular space. Early on during platelet activation, Ca²⁺ levels are considered oscillatory because platelets manage to handle excess Ca²⁺ through Ca²⁺ exporters [SERCA, the plasma membrane Ca²⁺ ATPases (PMCA), and the sodium calcium exchanger (NCX)] and mitochondrial Ca²⁺ uptake (21). However, in strongly stimulated platelets, over time, the NCX will reverse its direction and participate in sustaining high cytosolic Ca²⁺ levels (25). As a consequence, mitochondria will increasingly store more Ca²⁺ until a critical threshold is reached (24). Hereafter, the mitochondrial membranes will depolarize, initiating the process of platelet necrosis, resulting in the opening of the mitochondrial permeability transition pore (mPTP) (26, 27). Opening of the mPTP releases a significant amount of mitochondrial Ca²⁺ in the extracellular space. Critically, only platelets that have undergone mitochondrial depolarization will achieve the high and sustained Ca²⁺ levels required for activation of TMEM16F and subsequent PS exposure. In parallel, loss of the inner mitochondrial membrane potential will result in rapid ATP depletion, inhibiting flippases.

mPTP opening is dependent on cyclophilin D (CypD), as either pharmacological inhibition of CypD with cyclosporine A or genetic knockout of CypD strongly reduces procoagulant platelet formation (26, 27). Interestingly, in settings of oxidative stress, the Ca²⁺ threshold to activate CypD and mediate mPTP formation is lower, and platelets are more likely to become procoagulant (22, 28–30). However, oxidative stress alone has no effect on PS externalization, implying oxidative stress may play a role in priming platelets for procoagulant responses. As discussed later, this is of particular interest to several thrombotic diseases in which procoagulant platelets are known to be involved and where comorbidities greatly impact outcomes. Diabetes, for example, induces oxidative stress in platelets, increasing platelet reactive oxygen species and potentiating the formation of procoagulant platelets (28, 30).

Fluid Entry Influences Procoagulant Membrane Dynamics

Although Ca²⁺ is being mobilized to express PS on the platelet surface, fluids enter the procoagulant platelet inducing physical disruption of the membrane-cytoskeleton interaction, increasing internal hydrostatic pressure and resulting in irreversible membrane swelling (31). This membrane swelling is called ballooning and is required to support and localize thrombin generation on the platelet surface. In fact, the procoagulant quality of the platelet is dependent on the amount of coagulation factors they can bind on their enhanced surface area (32). Recently, it was shown that the transition to a balloon shape is dependent on a coordinated Na⁺, Cl^–, and water entry and that inhibition of either of these processes is sufficient to block platelet ballooning and thrombin generation (31). Interestingly, ballooning could be inhibited by acetazolamide, an FDA-approved inhibitor of aquaporin water channels (AQP). In other cells, it has been shown that AQP1 is a critical regulator of cell volume and membrane dynamics (33). However, experiments with mice deficient in AQP1 highlight the complexity of ballooning and procoagulant platelet formation. Although procoagulant platelet responses were suppressed in AQP1 KO mice, procoagulant membrane ballooning was unaffected (34). This highlights the multifaceted role of aquaporins in procoagulant platelet responses and demonstrates the need for more research on this topic.

CONTRIBUTION OF PROCOAGULANT PLATELETS TO HEMOSTASIS AND THROMBOSIS

A reduced ability to form procoagulant platelets has been associated with mild bleeding disorders (35, 36). However, the most compelling evidence for a role of procoagulant platelets in hemostasis comes from patients with Scott syndrome. Patients with Scott syndrome patients lack the scramblase TMEM16F resulting in little PS expression on any surface and have a mild bleeding phenotype (37). Similarly, mice lacking TMEMF16 have a bleeding tendency (38). Critically, this bleeding phenotype was also observed in platelet-specific TMEM16F-deficient mice, implying a significant role for platelet PS in hemostasis (39). However, TMEM16F mediates both agonist-induced and apoptosis-induced PS scrambling, indicating the bleeding phenotype in these mice is not necessarily solely due to a lack of procoagulant platelet responses (8, 39). In contrast, inhibiting procoagulant platelet formation through deletion of platelet CypD had little or no effect on tail bleed times in mice (27). The discrepancy between bleeding associated with TMEMF16 deletion compared with CypD deficiency is currently unclear. However, differences in integrin αIIbβ3 inactivation between TMEMF16 and CypD deficient platelets may explain, in part, why increased bleeding is associated with TMEMF16 deletion. Furthermore, it is important to note that in either case, procoagulant platelet formation is not completely abolished (40). However, in TMEM16F-deficient animals and patients with Scott syndrome, platelet ballooning is averted (11). This phenotype has not been reported for CypD-deficient platelets. This raises the possibility that the combined reduction in PS and the inability for platelets to balloon could lead to a more pronounced hemostatic phenotype in the absence of TMEM16F. Taken together, these data suggest that in certain circumstances procoagulant platelet formation mediated though TMEM16F is critical for platelet hemostatic function, whereas CypD does not affect this. However, future exploration is needed to fully appreciate the role of procoagulant platelets on hemostasis.

General blocking of PS using either AnnexinA5 (41) or lactadherin (42) has been shown to block both arterial and venous thrombosis. However, the specific role of procoagulant platelets in thrombosis is less clear. Although deficiency of TMEM16F was associated with a reproducible reduction in arterial thrombosis(38, 39, 43), the role of CypD seems more model-dependent. Global and platelet-specific deletion of CypD results in accelerated times to occlusion in a photochemical arterial thrombosis model (27). The observed difference is independent of initial platelet adhesion and based on in vitro data, the acceleration in thrombosis is due largely to the inability of CypD-deficient platelets to inactivate αIIbβ3, resulting in a more proaggregatory platelet phenotype at the site of injury. However, in two different arterial thrombosis models, using FeCl₃ or laser-induced injury, Hua et al. (44) in their study observed decreased platelet adhesion and fibrin formation as well as reduced procoagulant platelet formation in platelet-specific CypD knockout mice, indicating procoagulant platelet PS is important in thrombosis when thrombin and fibrin are generated quickly after injury. To add to the nuanced role of procoagulant platelets in thrombosis, platelet CypD-deficient mice have increased thrombus formation in a venous thrombosis model (45) and increased microvascular occlusion in a collagen-epinephrine pulmonary thrombosis model (46). Data from the venous and pulmonary thrombosis models support the concept that procoagulant platelets potentially play antithrombotic roles by limiting integrin activation. However, in instances where platelet-dependent thrombin generation mediates thrombotic complications, inhibition of procoagulant platelets may be beneficial. A limitation to these observations is that most studies have focused on CypD and mPTP formation to examine the in vivo role of procoagulant platelets. The generation of different models examining other molecular players in procoagulant platelet formation may shed additional light on their role in hemostasis and thrombosis.

THROMBOINFLAMMATORY FUNCTIONS OF PROCOAGULANT PLATELETS

It has become increasingly clear that besides their roles in thrombosis and hemostasis, platelets play a critical role in inflammation. Platelets help recruit and activate immune cells at sites of infection and subsequently aid in the resolution of inflammation. However, also in the absence of infection, platelets can interact with leukocytes mediating thromboinflammation. Thromboinflammation occurs when thrombotic and inflammatory pathways converge causing collateral tissue damage through neutrophil extracellular traps (NETs) and intravascular thrombosis (47). Interestingly, the receptors mediating thromboinflammation differ from those in hemostasis, making them attractive therapeutic targets (48). In particular, procoagulant platelets have emerged as mediators of thromboinflammation. In vitro, several studies have reported on a preferential interaction of procoagulant platelets with neutrophils (49–51).

Evidence from Animal Models

Several reports have linked procoagulant platelets to thrombotic complications during ischemia-reperfusion injury through their interactions with other cells. In the setting of inflammation, thrombotic complications can occur at distant locations compared with the site of injury. For example, gut ischemia followed by reperfusion induces thrombosis in the pulmonary circulation, which is dependent on procoagulant platelets mediating neutrophil recruitment. Interestingly, similar neutrophil-rich thrombi were found in the lungs of patients with human acute respiratory distress syndrome (50). Mechanistically, the formation of thrombi in the lungs after gut ischemia-reperfusion was both P-selectin dependent and independent. In particular, the rolling of neutrophils over PS-positive procoagulant platelet membranes was found as a key driver of pathological neutrophil macroaggregate formation. Importantly, deletion of platelet CypD reduces pulmonary thrombosis after gut ischemia-reperfusion. In contrast, mice deficient in Bak and Bax, both regulators of apoptotic PS expression, are not protected, providing a role for procoagulant platelets in reperfusion injury (50).

Similar observations were reported in a murine model of cerebral ischemia-reperfusion injury. After transient middle cerebral artery occlusion (tMCAO), circulating PS-positive platelets are increased systemically and were shown to preferentially interact with neutrophils (51). This results in increased neutrophil recruitment to the brain and platelet-mediated NET formation in the cerebral vasculature (52). Interestingly, platelet-specific CypD knockout mice are protected from ischemic stroke due to reduced PS-platelet-dependent neutrophil brain recruitment. The reduction of neutrophils to the brain after ischemia improves cerebral blood flow, resulting in reduced infarct injury and improved neurological and motor function (51). These studies indicate that procoagulant platelets are primed to interact with neutrophils to promote thrombosis during ischemia-reperfusion. Intriguingly, similar to the discrepancies in hemostasis, in cerebral ischemia-reperfusion injury, TMEM16F deficiency does not recapitulate the CypD knockout mice, highlighting the complexity of procoagulant platelets also in reperfusion injury (39).

In contrast to their detrimental role in sterile inflammation, procoagulant platelets were recently shown to maintain vascular integrity in the inflamed lung. Through intravital microscopy, Kaiser et al. (53) in their study demonstrated that procoagulant platelets binding to exposed collagen at sites of lung injury are crucial for the initiation of coagulation and the prevention of inflammatory bleeding in a mechanism that is CypD and TMEM16F dependent. Their study also identified that combined integrin-mediated outside-in-signaling and GPVI signaling were required to trigger procoagulant activity in the inflamed lung and that inhibition of both pathways was needed to breach vascular integrity. These results shed new light on the role of procoagulant platelets in the setting of inflammation. Depending on the setting, procoagulant platelets can either promote thrombosis or prevent bleeding. It is tempting to speculate that similar mechanisms that are preventing inflammatory bleeding in the lung are what is driving detrimental procoagulant platelet formation in the brain. Crucially, neither CypD (51) nor TMEM16F (39)-deficient mice showed signs of inflammatory brain bleeding when subjected to stroke, highlighting the tissue-specific role of platelets in maintaining vascular integrity in the setting of inflammation.

Clinical Implications for Procoagulant Platelets

In humans, procoagulant platelet formation is associated with clinical outcomes in certain diseases such as stroke, COVID-19, heparin-induced thrombocytopenia (HIT), and vaccine-induced thrombotic thrombocytopenia (VITT).

Similar to the murine stroke data, procoagulant platelet formation is associated with adverse outcomes in patients with ischemic stroke. In one study using patients with a previous lacunar stroke, procoagulant platelet formation predicted recurrent stroke within a year of the initial stroke (54). In a separate study, high levels of procoagulant platelet formation in patients with asymptomatic carotid artery stenosis at the time of enrollment were predictive of ipsilateral stroke or transient ischemic attack during a 40-mo follow-up period (55). In contrast, increased procoagulant platelet formation is associated with a reduced risk of hemorrhage after nonlacunar ischemic stroke (56). Likewise, lower levels of procoagulant platelets were linked to increased mortality after intracerebral hemorrhage (57).

Patients with COVID-19 are at increased risk for thrombotic complications. This is, in part, due to increased platelet reactivity (58). The role of procoagulant platelets in COVID-19 seems to depend on the disease severity and state. Procoagulant platelet formation of isolate platelets induced by dual agonists stimulation is significantly reduced in patients with moderate and severe COVID-19 compared with healthy donors (46, 59). In addition, mitochondrial depolarization is also significantly impaired in patients with COVID-19, suggesting COVID-19 induces mitochondrial dysfunction, which impairs the ability of platelets to become procoagulant in response to ex vivo stimulation (46). What the implications of these findings are, is currently not understood. However, although levels of procoagulant platelet potential are decreased in patients with COVID-19 early in the disease time course, they increase over time (46, 59). Furthermore, patients with COVID-19 with higher levels of procoagulant platelet potential were at increased risk for death (59). In agreement with this, baseline circulating levels of procoagulant platelets are significantly increased in severe COVID-19 compared with healthy donors as well as to other critically ill patient populations (60). Importantly, the level of circulating PS-positive platelets positively correlates with D-dimer levels and illness severity as well as associates with thrombotic complications (60). Interestingly, procoagulant platelet formation is mediated by COVID-19-induced antibody formation (60–62). Circulating immunoglobulins (IgGs) during COVID-19 activate platelets through FcγRIIa, altering cytoplasmic calcium and cyclic adenosine-monophosphate (cAMP) levels and activating AKT (61, 62). Therapeutically, inhibition of cAMP with IIoprost (62) or blocking AKT phosphorylation (61), inhibits COVID-19 IgG-induced procoagulant platelet formation.

HIT is an immune-mediated thrombocytopenia mediated by IgGs against platelet factor 4 (PF4)-heparin complexes, which in turn activates platelets through FcγRIIa (63). IgG-dependent platelet activation can result in aggregation and the release of procoagulant microparticle particles, which can participate in thrombin generation to promote a thrombotic environment (64). Interestingly, HIT antibodies in complex with PF4 are capable of inducing procoagulant platelet formation (65, 66). This is dependent on platelet FcγRIIa and the generation of monocyte-derived tissue factor and monocyte-supported thrombin generation (65, 67, 68). In a separate study, HIT plasma only induces procoagulant platelets in the presence of heparin and PAR stimulation (PAR1 or ADP) (65, 66). Mechanistically, HIT IgG binding to FcγRIIa induces procoagulant platelet formation, through robust calcium flux, resulting in calpain activation (64, 66). More recently, a similar procoagulant platelet phenotype was observed in patients who developed VITT after adenoviral-based COVID-19 vaccine administration (69, 70). Sera from patients with VITT contain high levels of anti-PF4 antibody and these antibodies significantly increase procoagulant platelet formation compared with healthy donors (69, 71). Similar to HIT, this is dependent on the presence of PF4 and heparin and the binding of antibody complexes to platelet FcγRIIa (69, 71). A similar finding of VITT-induced procoagulant platelet formation was also observed by Lee et al. (71), but also required PAR1 signaling. Although these studies suggest a role for procoagulant platelet formation in the pathophysiology of HIT and VITT, additional studies are necessary to mechanistically delineate their role in the disease process.

CONCLUSIONS AND PERSPECTIVES

Procoagulant platelets are increasingly recognized as mediators of thromboinflammation and therefore therapeutic targets for thromboinflammatory diseases. The discrepancy between procoagulant platelets and aggregating platelets and how they are distinctly formed raises the possibility that targeting procoagulant platelets could block pathological thrombosis while keeping hemostasis intact. However, there are currently no pharmacological inhibitors available that only block procoagulant platelet formation. Given the bleeding phenotype observed in patients with Scott syndrome (37), general scramblase blocking using TMEM16F inhibitors may not be ideal. However, drugs such as cyclosporin A (26) and acetazolamide (31) are FDA-approved drugs that, as a side effect, efficiently block procoagulant platelet formation without known bleeding consequences. In contrast, desmopressin has been shown to increase procoagulant platelet formation and reduce bleeding (36, 72). Nevertheless, a thorough characterization of these molecules in models of thrombosis, hemostasis, and thromboinflammation is needed to determine their safety and efficacy profile.

GRANTS

This work was supported by National Institutes of Health (NIH) Grants K01AG059892 and R01HL163019 (to R.A.C.) and the American Heart Foundation Grant 2021Post830138 (to F.D).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

F.D. and R.A.C. prepared figures; F.D. and R.A.C. drafted manuscript; F.D. and R.A.C. edited and revised manuscript; F.D. and R.A.C. approved final version of manuscript.