Platelet GPVI (Glycoprotein VI) and Thrombotic Complications in the Venous System

Abstract

The immunoglobulin receptor GPVI (glycoprotein VI) is selectively expressed on megakaryocytes and platelets and is currently recognized as a receptor for not only collagen but also a variety of plasma and vascular proteins, including fibrin, fibrinogen, laminin, fibronectin, and galectin-3. Deficiency of GPVI is protective in mouse models of experimental thrombosis, pulmonary thromboembolism as well as in thromboinflammation, suggesting a role of GPVI in arterial and venous thrombus formation. In humans, platelet GPVI deficiency is associated with a mild bleeding phenotype, whereas a common variant rs1613662 in the GP6 gene is considered a risk factor for venous thromboembolism. However, preclinical studies on the inhibition of GPVI-ligand interactions are focused on arterial thrombotic complications. In this review we discuss the emerging evidence for GPVI in venous thrombus formation and leukocyte-dependent thromboinflammation, extending to venous thromboembolism, pulmonary thromboembolism, and cancer metastasis. We also recapitulate indications for circulating soluble GPVI as a biomarker of thrombosis-related complications. Collectively, we conclude that the current evidence suggests that platelet GPVI is also a suitable cotarget in the prevention of venous thrombosis due to its role in thrombus consolidation and platelet-leukocyte complex formation.

Highlights.

• Platelets actively contribute to venous thrombosis with a suggested role of GPVI (glycoprotein VI).

• Platelet GPVI contributes to thrombosis in inflammation and cancers.

• Genetic variation in GP6 is a modest risk factor for venous thrombosis.

The immunoglobulin receptor GPVI (glycoprotein VI) has been widely studied as a platelet-activating receptor for collagen and is currently considered as a therapeutic target for arterial thrombotic complications. Here, we overview and discuss the indications for a role of GPVI also in venous thrombotic complications, including thromboembolism, thromboinflammation, and venous cancer thrombosis.

Interfaces of Platelet and Coagulation Activation

It is established that the processes of platelet activation and coagulation are highly interconnected in hemostasis and (arterial) thrombosis.^1,2 The reported connection mechanisms are multiple, and involve in particular: (1) collagen- and thrombin-induced exposure on platelets of the procoagulant phospholipid phosphatidylserine, (2) subsequent enhanced generation of factor Xa and thrombin on these procoagulant platelets, (3) enforcement of platelet activation by the coagulation products thrombin and fibrin, and (4) thrombin-mediated fibrinogen proteolysis to form fibrin and stabilize the platelet thrombus.^1,3,4 This concept of interactive thrombus-and-clot-formation is supported by static and flow studies, and it depicts platelets as thrombin- and fibrin-responsive cells, capable to aggregate and to generate massive amounts of thrombin and fibrin. Both the extrinsic (triggered by tissue factor) and intrinsic (activated by factor XIIa) coagulation pathways are considered to contribute to the thrombin generation process.⁵ Modeling studies furthermore support the concept of tissue factor and factor XII playing critical roles in the thrombus formation at both arterial and venous flow conditions.^6,7

In the human body, the process is considered to start with the GPVI-dependent platelet adhesion to collagen present in the subendothelial matrix or the lesioned-atherosclerotic plaque.⁸ In the context of thromboinflammation, that is when inflammation leads to thrombosis, also other cells come into the play. On the inflamed endothelium, circulating platelets interact with leukocytes (neutrophils, monocytes) again resulting in thrombin and fibrin generation.⁹ In this article, we critically review the evidence for a role of GPVI in venous thrombosis alongside its contribution to arterial thrombosis.

GPVI and Thrombus Formation

GPVI is expressed exclusively on platelets and megakaryocytes. In the platelet membrane, GPVI is associated with the FcR γ (Fc receptor γ)-chain, which is responsible for the signaling via its immunoreceptor-tyrosine-based-activation-motif.¹⁰ Upon GPVI-ligand interaction and dimerization, the 2 tyrosine residues in the immunoreceptor-tyrosine-based-activation-motif become phosphorylated by Src-family kinases, which results in the binding and phosphorylation of the tyrosine kinase Syk (spleen tyrosine kinase) through its tandem SH2 (Src homology 2) domains. The ensuing phosphorylation cascade in the LAT signalosome leads to activation of phospholipase Cγ2 and phosphoinositide 3-kinases, culminating in a prolonged intracellular Ca²⁺ increase and other platelet responses.¹¹

As a main signaling receptor for collagen, GPVI has widely been studied in the context of arterial thrombosis, whereas patient examinations suggest a limited but non-negligible role in hemostasis.^12–14 An explanation is that in the absence of GPVI, hemostasis is preserved by parallel platelet pathways, for example, by platelet adhesion to collagen by the integrin α₂β₁ receptors and indirectly by GPIb-V-IX (glycoprotein Ib-V-IX) interacting with collagen-bound VWF (von Willebrand factor). Both GPVI and the FcR γ-chain have been found to be essential in murine arterial thrombosis models triggered by vascular injury and collagen exposure, regardless of the vascular bed, with a limited contribution to tail bleeding.¹⁵

The human GP6 gene contains 8 exons, of which the last encodes for a short intracellular and transmembrane domain (Figure [A] and [B]). Nine GP6 variants have been identified so far, the majority of which associates with a loss of receptor expression or function (Figure [B] through [D]). Individuals with platelet GPVI deficiency, carrying a common homozygous insertion in GP6 that prevents protein expression, were identified in 11 unrelated families in Chile.²² They all have normal platelet counts and no more than mild bleeding diathesis.¹² The platelets were shown to be dysfunctional in flow-dependent thrombus formation, but not in adhesion to collagen surfaces.²² Gene mapping of a cohort of 1212 blood donors set the mutation incidence at 2.9% in the Chilean population, suggesting the existence of a large number of mutation carriers without clear symptoms.²² Whether these individuals are protected from thrombosis is unknown but of great interest. Furthermore, few patients have been identified with an acquired immune deficiency in platelet GPVI and associated thrombocytopenia.¹³

Figure.

GPVI (glycoprotein VI) structure and natural variants. A, Schematic presentation of the human GP6 gene structure. The first seven exons encode for the protein extracellular domain, with exons 3-4 encoding for domains D1 and D2, respectively. Exon 8 encodes for the transmembrane and short intracellular region. B, Cartoon showing the human GPVI protein domains with amino acid positions of the nine GPVI natural variants and mutations indicated. C, Representation of crystal structure of the GPVI binding site interacting with CRP (collagen-related peptide). D, Table of reported effects of GPVI variants on protein expression and platelet functions.^16–21 FcR γ indicates Fc receptor γ; and VTE, venous thromboembolism.

In recent years, it has become clear that additional adhesive proteins in vessel wall or blood can function as ligands and agonists for GPVI, usually in conjunction with an integrin. These include vascular laminins (with integrin α₆β₁ as coreceptor),^23,24 fibrillar fibronectin,²⁵ and the basement membrane protein nidogen-1,²⁶ leading to the suggestion that it functions as a pattern recognition receptor.²⁷ In addition, it appeared that fibrinogen and fibrillar fibrin can bind and activate platelet GPVI.^28–30 However, in comparison to collagens or synthetic collagen-like peptides, the fibrin- or fibrinogen-mediated GPVI activation was found to have limited Syk-mediated signaling capacity and to rely on integrin α_IIbβ₃.³¹

Additive Roles of GPVI and Thrombin

As a general scheme, in vitro flow perfusion studies point to a synergy of GPVI- and collagen-dependent platelet activation, aggregation, and procoagulant activity, together with a tissue factor-mediated generation of thrombin and fibrin.^2,5,6 Similarly, mouse models of in vivo thrombus and fibrin formation have elucidated an additive contribution of GPVI-induced platelet activation and tissue factor–induced coagulation, in the mesenteric arterioles and venules.^32,33 Complementarity of GPVI and thrombin activities also appeared from the structure of in vivo thrombi raised by collagen exposure, with an inner core of highly activated platelets, a transition zone, and an outer shell of loosely packed platelets, with fibrin fibers stabilizing the inner core.³⁴ Similarly, GPVI was found to regulate the stability and hence thrombus structure.³⁵

Mouse studies furthermore revealed a role of GPVI in the thrombin-sensitive ischemic stroke models, for example, via transient middle cerebral artery occlusion, in which GPVI depletion suppressed arterial platelet adhesion and drastically reduced infarct size in the brain, independently of platelet aggregation.^36–38 Together, these findings draw attention to roles of platelet GPVI beyond the classical collagen-induced aggregation.

Platelets and Venous Thrombosis

The effective treatment of venous thrombotic complications by a wide spectrum of anticoagulants implicates that especially thrombin (fibrin clotting) has an overall controlling role in venous thrombus formation, thus leaving a subordinate role for platelets. The spectrum of VTE includes deep venous thrombosis (DVT) and pulmonary thromboembolism. Worldwide VTE is the third common cause of cardiovascular mortality after coronary artery disease and stroke. Importantly, VTE is considered a long-term and phased disease since an initial (unprovoked) thromboembolism in the venous system can be followed by recurrent VTE, life-threatening emboli in the lungs, and a post-thrombotic syndrome.³⁹ Phenotypically, the accepted model is that impaired blood flow together with hypercoagulability and endothelial dysfunction (Virchow triad) gradually lead to the formation of large-size venous red thrombi, composed of fibrin, platelets, and red cells; a process that is driven by thrombin and can start in the valve pockets of large veins.^39,40

Activated platelets contribute to the venous thromboembolic events. Although less predominant than in arterial thrombi, platelet aggregates were found to comprise a relevant part of analyzed venous clots.^41,42 This agrees with the outcome of in vitro microfluidic studies, which assign a role of GPVI and other platelet receptors in thrombus formation at both venous and arterial flow conditions.⁴³ In adapted microfluidic chambers simulating venous flow disturbances around valves, it was found that platelets contribute here to the thrombus and clot growth by interacting with fibrin.⁷

Evidence for a consistent role of platelets in venous thrombosis furthermore comes from in vivo mouse studies. We note here that the present mouse venous thrombosis models have limitations in the way of triggering thrombus formation (usually flow restriction) and a short time of thrombus development (days to weeks). Nevertheless, it was shown that following veins flow restriction in such DVT models, platelets promote the recruitment of leukocytes (monocytes and neutrophils) to the activated endothelium. This resulted in a leukocyte-dependent coagulation process and the buildup of a venous thrombus or clot.⁴⁴ It was proposed that leukocyte-expressed tissue factor was instrumental in the induction of thrombin generation.

In a similar mouse model of venous flow restriction, the amyloid precursor protein was identified as negative regulator of platelet-neutrophil interactions, fibrin-thrombus formation, and embolization; this led to the suggestion that the platelet-derived amyloid precursor protein may limit VTE.⁴⁵ However, other mouse studies indicated that also platelet-expressed P-selectin and chemokines regulate leukocyte interactions.⁴⁶ Furthermore, mouse platelet deficiency in the secretion-controlling protein SNAP23 (a condition leading to thrombocytopenia) resulted in an impaired thrombosis tendency in both arteries and veins.⁴⁷ Collectively, these rodent studies suggest a partly leukocyte-dependent and partly fibrin-dependent role of platelets in the development of experimental VTE.

In humans, recent clinical studies to compare treatments of coronary artery disease advocate the combined use of antiplatelet and anticoagulant medication (dual pathway therapy). In the COMPASS trial (Cardiovascular Outcomes for People Using Anticoagulation Strategies), it was concluded that the combination of aspirin (weak platelet inhibitor) and rivaroxaban (factor Xa antagonist) resulted in a better cardiovascular outcome, when compared with single aspirin or rivaroxaban alone.⁴⁸ Although the rivaroxaban treatment led to more bleeding events than aspirin intake, the dual pathway therapy resulted in less VTE events when compared with the monotherapies. This suggested a favorable effect of aspirin also for patients with only venous thrombosis. The INSPIRE initiative with 1200 patients concluded that aspirin plus anticoagulant treatment reduced the overall risk of recurrent VTE by more than one-third.⁴⁹ However, later meta-analyses and post hoc studies did not confirm risk reduction of VTE by aspirin on top of anticoagulant.^50–52 It is remarked here that aspirin in vitro is a weak platelet inhibitor, suppressing collagen-induced, but not thrombin-induced platelet responses. Altogether, this led us to speculate that stronger antiplatelet drugs are more effective in preventing venous thrombotic events, for instance in VTE or in atrial fibrillation.

Genetic Variation in GP6 and Venous Thrombosis

In 2007, the incidence of DVT was estimated at 1 per 1000 individuals per year, with a 10-year recurrent risk of about 30%.⁵³ Several risk factors for DVT have been recognized, including age, hospitalization, cancer, pregnancy, anticonception, and surgery.⁵³ Family studies have estimated that approximately half of the DVT cases are heritable.⁵⁴ Yet, the identified genetic factors, although of large effect sizes, account for only a minority of all DVTs. Genetic factors link to deficits in the anticoagulant proteins, antithrombin protein C and protein S, as well as to a gain-of-function in the procoagulant proteins factor V (factor V Leiden or Padua) and prothrombin (G20210A mutation).^54,55 Next to the genes of these (anti)coagulant factors, efforts have been made to search for additional (common) variants that associate with DVT, for a better risk prediction and understanding of the disease.

In a first population-based case-control study, the Leiden group examined 20 000 single nucleotide variants, which after a 4-stage refinement protocol resulted in 3 single nucleotide variants that significantly associated with DVT.⁵³ A minor allele of the antithrombin gene SERPINC1 (T, rs2227589) had a modest prothrombotic tendency (Table 1). Interestingly, the same held for a major allele of GP6 (G, rs1613662), which is linked to a higher GPVI expression on platelets. A meta-analysis of 5 studies encompassing 4000 VTE cases and 6100 controls indicated that the 2 single nucleotide variants of SERPINC1 and GP6 counted as 15% higher risk factors for venous thrombosis in White patients (Table 1). Although was not seen for Black patients, the sample size was underpowered here for a definitive conclusion.⁶⁴ In a prospective Danish case-cohort study, it was later confirmed that the heterozygous presence of this GP6 allele (G, rs1613662) moderately, but significantly, increased the hazard ratio for VTE.⁶⁵ Since 2015, the research has identified 17 VTE-associated genes, predominantly of (anti)coagulation factors, including GP6, ABO, and secretion-regulating genes appearing as network disconnected entities.⁵⁴ In a recent study of the association between this GP6 SNP and VTE risk in cancer, it was found that cancer-free AA-allele carriers had a 12-29% higher risk of VTE or DVT (nonsignificant) and a 53% to 61% higher risk of pulmonary thromboembolism.¹⁶ The different risk between genotypes, however, disappeared for the (lower numbers of) patients with active cancer, which suggested that the cancer state overruled mild effects of a variable GPVI expression.¹⁶

Table 1.

Epidemiological Support for a Role of Variation in Genes of GPVI, FcR γ-Chain, and Antithrombin in the Venous Thrombosis Risk

The risk GP6 variant rs1613662 consists of an A/G conversion, which introduces a serine to proline substitution in amino acid 219 and is considered to affect the expression of GPVI on platelets (Figure [B] and [D]).¹⁶ Platelets carrying the minor G-allele thus express less GPVI receptors and can be impaired in collagen-dependent adhesion and activation properties.^66,67 Shear-dependent thrombus formation was found to be reduced in these individuals to a similar extent as in subjects carrying a variant (rs3557) of FCER1G encoding for the FcRγ-chain, which also associates with lower GPVI expression (Table 1).^68,69 In agreement with the linkage between GPVI and FcRγ-chain, it appears that human platelet deficiency in GPVI is accompanied by a lower expression of the FcRγ-chain.²²

In addition to the common variant P219S of GPVI, several other rare mutations in the GP6 gene are described that associate with an altered collagen-receptor binding or a reduced GPVI expression level on platelets (Figure [D]). Markedly, heterozygous deficiency in human GP6 seems to be accompanied by a no more than subtle effect on platelet function.^12,22 Similarly, in heterozygous GP6^+/− mice, no obvious platelet phenotype was found.⁷¹

Mouse GPVI and Venous Thrombosis

In mice, it is well established that antibody-mediated or genetic deficiency of either GPVI or FcRγ-chain results in abrogation of thrombus formation in the arterial circulation, with limited effects on tail bleeding.^15,72 A few available reports also point to a role of mouse GPVI in venous thrombosis, although knowledge is still scarce here. In the murine microcirculation, it was observed that injury-induced thrombus formation in the venules required collagen-dependent platelet activation and tissue factor–induced thrombin generation, with a procoagulant role herein of phosphatidylserine-exposing platelets.³² Upon injury of the mouse mesenteric tissue, thrombus formation was reduced to a similar extent in the venules and arterioles, after the depletion of platelet GPVI with JAQ1 monoclonal antibody against GPVI or after genetic deletion of the FcRγ-chain.⁷³

Acknowledging the limitations of current large-vein thrombosis models in mouse (eg, artificial flow restriction), these do point to a key role of platelets in general and of GPVI in particular. In transgenic mice with low levels of anticoagulant factors (antithrombin or protein C), it was established that a (fatal) thrombotic occlusion of large vessels was dependent on platelet activity next to tissue factor.⁷⁴ In this context, considering that galectin-3–binding proteins can also act as GPVI ligands,⁷⁵ an interesting finding is that the injection of galectin-3 enhanced the venous thrombosis induced by stasis.⁷⁶ It should be stated, however, that next to GPVI, also platelet GPIb-V-IX and its ligand VWF can act as functional players in mouse venous thrombosis models.^77,78 Moreover, in mouse embolization models, it was reported that GPVI depletion causes a transient protection against tissue factor–induced pulmonary thromboembolism.⁷⁹ Collectively, these mouse studies lead to the attractive, but still unproven, supposition that GPVI acts as a central platelet receptor in the context of venous thrombus formation and embolization and by implication that anti-GPVI cotreatment might provide antithrombotic protection by affecting thrombus formation in a both collagen- and thrombin-dependent way.

Soluble GPVI as a Biomarker in Thrombotic Diseases

Both in humans and mice, GPVI is stably expressed on resting platelets, but it can be extracellularly cleaved in the presence of antibodies or agonists, resulting in shedding of the GPVI extracellular ectodomain. The proteolysis results in a platelet population that is essentially devoid of functionally active GPVI.^80,81 Shedding of GPVI is primarily mediated by ADAM proteases (a disintegrin and metalloprotease), which become enzymatically active upon platelet activation. From experiments with mice deficient in ADAM10 or ADAM17, it was postulated that yet another enzyme can contribute to antibody-induced GPVI shedding.⁸² In humans, ADAM10 is primarily responsible for the GPVI cleavage secondary to conditions including receptor activation, high shear stress, or exposure to Ca²⁺-ionophore or factor Xa.^81,83–85 In contrast to the GPVI shedding, ADAM17-induced cleavage of GPIbα appears to be a constitutive process upon platelet aging, which occurs independently of GPVI ligands.⁸⁶

Because GPVI is selectively expressed on platelets and megakaryocytes, the presence in the circulation of sGPVI (soluble GPVI) has been postulated as a biomarker reflecting platelet activation in vivo. Elevated plasma levels of sGPVI are reported in circumstances with a prothrombotic propensity, such as acute ischemic stroke, thrombotic microangiopathy, rheumatoid arthritis, disseminated intravascular coagulation, atrial fibrillation, and DVT (Table 2).^56–58,61 Along the same line, elevated sGPVI accompanied the major bleeding events in patients with heparin-induced thrombocytopenia, that is, a condition relying on prior platelet activation through the low-affinity Fc receptor, FcγRIIA.⁶⁰ Higher levels of circulating sGPVI are also seen during sepsis progression, in worse-outcome trauma patients, or at gout flares.^59,62

Table 2.

Changes in Circulating sGPVI Observed in Thrombotic and Related Diseases

Although the majority of studies on cleaved GPVI relate to arterial thrombotic conditions, elevated sGPVI was also observed in patients developing DVT postoperatively, suggesting that the cleaved fragment can act as a biomarker in venous thrombosis.⁵⁸ In support of this, GPVI shedding can be induced by fibrin clots in sepsis or trauma.⁶² Taken together, these data suggest that the presence of soluble GPVI reflects in situ platelet activation in cases of multiple longer-term cardiovascular-related pathologies. Other diagnostic markers of platelet activation, such as β-thromboglobulin and platelet factor 4,⁸⁷ or of coagulant activity, such as thrombin-antithrombin complexes and D-dimers,⁸⁸ may rather detect more acute thrombotic events. A future well-structured study to compare these various platelet activation-dependent biomarkers may help to better understand the pathologies of distinct thrombotic diseases.

Platelet GPVI in the Wider Context of Thromboinflammatory Conditions

The term thromboinflammation is used to describe a condition where thrombotic and inflammatory events are pathological and lead to organ damage. It also implies a role of platelets in the vascular-related inflammation, often occurring in the microcirculation. Thromboinflammation was first described as a pathological process in sepsis and in ischemia-reperfusion injury.⁹ The thromboinflammatory process is supposed to be driven by coagulation activation, with anticoagulant therapies being effective although with bleeding side effects.⁸⁹ Mechanistically, is not well understood how platelets contribute to a vascular inflammatory potential, but they likely support this by the release of chemokines and cytokines.⁹⁰ An attractive model here is that the activated endothelial cells of the inflamed vessel wall express P-selectin and VWF, which triggers leukocyte and platelet adhesion, after which leukocyte-expressed tissue factor induces the formation of a venous clot.⁴⁴ Support for a platelet secretory role in the clotting process comes from the use of mice with platelet secretion defects.⁴⁷ In mouse models of nonsterile thromboinflammation, platelets were found to stimulate neutrophil granular release in a GPVI-dependent manner.⁹¹ In pneumonia-induced sepsis, GPVI appeared regulate the formation of platelet-leukocyte complexes.⁹² However, based on this limited knowledge, it appears that at the inflamed vascular beds GPVI can contribute to the inflammation process, besides its involvement in thrombus formation.

Platelets furthermore help to maintain vascular integrity and prevent blood loss from leaky vessels in inflammatory conditions. The few published mouse studies on inflammatory hemostasis indicate that platelet GPVI plays a role in the regulation of vascular integrity albeit in an organ- and stimulus-dependent manner, alongside the receptors CLEC2 (C-type lectin 2) and GPIb-IX-V.^93–96 However, there are no reports on inflammatory hemostasis or bleeding in GPVI-deficient individuals.

Inflammation-propagating effects of platelets have also been examined in the context of coronavirus disease 2019 (COVID-19). An infection by severe acute respiratory syndrome coronavirus-2 can lead to a wide range of clinical manifestations, varying from absence of symptoms to severe pneumonia which can progress into an acute respiratory distress syndrome and sepsis. In subjects at risk (elderly, patients with comorbidities), massive vascular inflammation is frequently observed, culminating in disseminated intravascular coagulopathy, arterial or venous thrombosis, and pulmonary thromboembolism.^97,98 The underlying condition is characterized as (pulmonary) endothelialitis and thromboinflammation.⁹ An assumption is that platelets, likely by interacting with leukocytes, promote the thromboinflammatory activity in SARS-CoV-2 infections.⁹⁹ Reports are, however, inconsistent regarding alterations in platelet responses in severely diseased patients with COVID-19, ranging from negative priming^100,101 to positive priming.^102,103 Because low platelet counts are uncommon,^97,104 it is unlikely that platelet activation is a disease trigger. A clinical study on the effect of GPVI antagonist glenzocimab in severe acute respiratory syndrome coronavirus-2 syndrome is in current progress (https://www.clinicaltrials.gov; Unique identifier: NCT04659109).

Role of GPVI in Cancer-Induced Thrombosis

The relation between cancers and (venous) thrombosis was already discovered in the nineteenth century.¹⁰⁵ Nowadays, it appears that (venous) thrombotic events are the second leading cause of death in cohorts of cancer patients, whereas conversely, a history of idiopathic venous thrombosis increases the susceptibility for developing cancer.¹⁰⁶ A high risk score for VTE is a predictive variable for earlier mortality in treated cancer patients.¹⁰⁷

In spite of the fact that cancers are different in origin, development, and fate, there is increasing evidence that platelet interaction with a tumor exposed to the circulation promotes cancer dissemination and metastasis.¹⁰⁸ Platelets can influence tumor cells by several mechanisms, including (1) the release of granular growth factors, matrix proteins, and inflammatory mediators, (2) the expression of P-selectin and other cell-adhesive receptors, and (3) the fibrin-mediated interaction of immune cells with a tumor.^108–110

The few available mouse studies point to a role of platelet GPVI especially in tumor metastasis. Platelet GPVI deficiency resulted in less metastatic foci after the implantation of tumor cells into mice.^75,111 A recently suggested mechanism is that GPVI mediates platelet interaction with cancer cell-derived galectin.⁷⁵ We speculate that GPVI can also promote metastasis via the binding to fibrin clots, which are formed around tumor cells expressing tissue factor.

Clinical Possibilities for GPVI Antagonism and Agonists

Selected inhibitors of platelet GPVI are close to enter the clinic for treatment of thrombosis. A clinical study is under way with a GPVI-blocking Fab (ACT017, glenzocimab), aiming to treat acute ischemic stroke.¹¹² In transgenic mice carrying the human GP6 gene, glenzocimab was found to be effective in thrombus suppression, without impacting GPVI-dependent inflammatory hemostasis.¹¹³ However, multiple inhibitors of protein tyrosine kinases downstream of GPVI, that is, Syk and Btk (Bruton tyrosine kinase), have extensively been evaluated in clinical trials and are currently prescribed for the treatment of B cell malignancies.^114–117 This type of chemotherapy is mostly well tolerated in diseased patients while causing only limited side effects such as bleeding, nausea, vomiting, or diarrhea.^118,119 GPVI-dependent antiplatelet effects can be expected from the use of these drugs as reported for the Btk inhibitor ibrutinib, prescribed to target B cells in chronic lymphocytic leukemia, along with effects on GPIb-mediated platelet responses.¹²⁰ With the introduction into the clinic of more tyrosine kinase inhibitors, it will be interesting to see if the antithrombotic effects eventually reported will link to a suppression of GPVI activation in platelets, such as already shown for ibrutinib in the setting of deep vein.^119,121

GPVI-Fibrin-Thrombin Loop as a Novel Target

Given the recent evidence for a role of GPVI in fibrin(ogen)-dependent thrombus formation and stability,^28,35 we speculate that in the venous thrombosis setting especially fibrin may act as a relevant GPVI agonist. Although the signaling strength of fibrin to GPVI is only low,³¹ we note that venous thrombus formation is usually a slow process, in comparison to arteries. This raises the possibility that a weak but prolonged GPVI signal might be pathophysiologically relevant. Because in the venous setting, GPVI will act on platelets in conjunction with thrombin, anti-GPVI treatment is likely to enforce the antithrombotic effect of anticoagulant drugs in venous thrombotic diseases. However, trials still need to be performed to support this idea.

Conclusive Model and Perspective

Collectively the studies in this review support the concept of VTE as a disease process, where platelet GPVI may play a role at different stages, although multiple questions remain (Table 3). A mechanistical question is whether fibrin is indeed the principal GPVI agonist under conditions of venous thrombosis and embolism. Assuming that the answer is yes, we propose that the formation of a venous thrombus is steered by a slow but continuous GPVI-thrombin-fibrin feedforward loop, in which fibrin-bound platelets expose low levels of phosphatidylserine in a GPVI-dependent manner, after which the activity of locally generated thrombin is dampened by its binding to fibrin. Under flow conditions in vitro, we have observed that GPVI consolidates rather than propagates the formation of a thrombus.³¹ However, once stasis is reached, this loop can still allow the gradual growing of a clot over time, balanced in vivo by endothelial activity. Future elucidation of the crystal structure resolving the binding sites of collagen and fibrin(ogen) to GPVI (Figure [C]), and development of selective inhibitors based on the fibrin binding site, will help to resolve the importance of this feedforward loop. So far, it is also known that binding to fibrinogen relies on the avidity of GPVI-GPVI interactions.¹²²

Table 3.

A Summary of What Is Currently Known and Unknown Regarding Platelet GPVI, Murine VT, and VTE

At the inflamed and then less antithrombotic endothelium, we propose that GPVI can influence thrombus formation in multiple ways. The endothelial activation leads to VWF release and P-selectin expression, which results in capturing and activation of both platelets and neutrophils in part via GPVI.⁴⁴ In this setting, GPVI can increase the thromboinflammatory status by promoting neutrophil granular release⁹¹ and support the thrombus formation through fibrin binding.^28,31 The GPVI-activated and procoagulant platelets will furthermore contribute to the generation of thrombin and fibrin.

At present, it is unclear how the interindividual variation in platelet GPVI levels fits in this scenario. However, in analogy to the other genetic risk factors of venous thrombosis (defects in the anticoagulants protein C, protein S, antithrombin), one can assume that GPVI expression level determines the activity of these interactions as in the proposed GPVI-thrombin-fibrin loop. All these intriguing aspects together make GPVI an attractive receptor for further studies on VTE and for a deeper understanding of the sets of mouse and human data (Table 3).

Article Information

Acknowledgments

We thank Sam Montague for textual suggestions and Alex Slater (both University of Birmingham) for contributing to the figure. All authors cowrote, revised, and reviewed the article.

Sources of Funding

G. Perrella is supported by a Birmingham-Maastricht PhD studentship. S.P. Watson acknowledges the support of the British Heart Foundation (BHF; CH03/003).

Disclosures

None.