Keywords
GPVI - antiplatelet therapy - drug discovery - protein–protein interaction
Introduction
Despite significant advances in the treatment of cardiovascular diseases, in 2016, over 15 millions of deaths were caused by ischemic heart disease and stroke, induced by artery thrombosis.[1] Platelets, anucleate blood cells derived from megakaryocytes, play a major role in the development of cardiovascular thrombosis. As shown in [Fig. 1], following endothelial damage caused by the growth of an atherosclerotic plaque and under high shear stress conditions, platelets will adhere to the subendothelial matrix, activate, and release prothrombotic mediators, recruiting additional platelets from the bloodstream to form an aggregate. If not treated, vascular occlusion may occur.[2] Platelet-mediated thrombosis is not restricted to events where extracellular matrix (ECM) is exposed to the blood vessel lumen, mainly of the artery. For instance, when vessel damage reaches the adventitial layer, tissue factors expressed in smooth muscle and adventitial cells can initiate thrombin production, thereby activating platelets. Another mechanism of platelet-mediated thrombosis, independent of ECM exposure, is through endothelial cell activation, which releases ultralarge von Willebrand factor (VWF) multimers that induce platelet activation via the glycoprotein (GP) Ib receptor complex. In each of these mechanisms, platelet-mediated thrombosis can result in uncontrolled clot formation and subsequently arterial or venous thrombosis.[4] Antiplatelet therapy has been an effective and efficient treatment that has led to reduced mortality and morbidity by targeting key pathways of platelet activation.[5]
[6] Current antiplatelet therapy includes antagonists for adenosine diphosphate (ADP) receptors (clopidogrel, ticlopidine, prasugrel), integrin αIIbβ3 inhibitors (abciximab, eptifibatide, tirofiban), phosphodiesterase inhibition (dipyridamole and cilostazol), antagonists for the thrombin receptor protease-activated receptor 1 (vorapaxar), and inhibition of thromboxane A2 (TXA2) synthesis (aspirin). However, such drugs present several limitations, including weak inhibition of platelet function, only a modest effect in primary prevention, and, most importantly, increase in bleeding events.[2]
[5]
[7]
Fig. 1 Platelet activation on atherosclerosis formation. Once an atherosclerotic plaque is established, the platelet activation process is inherently self-perpetuating. Briefly, following the damage to endothelial cells due to plaque growth, platelets will be exposed to the extracellular matrix (ECM) under a high shear stress flow condition. After adhering to ECM, platelets will become activated releasing prothrombotic factors that will recruit platelets from the bloodstream, forming an aggregate. Platelet aggregation at the injury site is crucial for thrombus formation and, if not controlled with antiplatelet drugs, may progress to vascular occlusion.[2] During atherosclerotic plaque formation, other key events take place, such as rolling and adhesion of monocytes to the endothelium, supported by platelets, then transform into macrophages, and subsequently in foam cells, which produce proinflammatory cytokines. The inflammatory process accelerates as T-cells, B-cells, and dendritic cells infiltrate the plaque.[3]
Hemorrhage is a major concern in all currently available antithrombotic treatments. Due to the weak inhibitory effect of particular drugs on platelets, such as the one of aspirin, a common medical management strategy is to combine different pharmacological therapies. In a meta-analysis of 18 randomized trials comprising 129,314 patients, it was reported that dual antiplatelet therapy presents a 50% increase in risk of bleeding compared with the single agent therapy.[8] In addition, the risk of hemorrhage can be more than threefold higher when triple therapy with warfarin is used to prevent and treat venous and arterial thrombosis, leading to an increase in the risk of fatal bleeding.[9] Furthermore, the risk of bleeding due to antiplatelet therapy increases with age. Recently, Li et al demonstrated, in a large population study, that the long-term risks and severity of bleeding in patients receiving aspirin increases dramatically with age (older than 75 years) and with substantial risk of fatal bleeding.[10] For these reasons, the potential benefit of preventing thromboembolic disease needs to be balanced with the potential harm of inducing hemorrhage. Therefore, antiplatelet therapy can be improved by developing new drugs capable of treating thrombotic events without causing bleeding as an adverse effect.
The Search for Inhibitors of GPVI, a Potential Target for “Bleeding-Free” Antithrombotic Therapy
The Search for Inhibitors of GPVI, a Potential Target for “Bleeding-Free” Antithrombotic Therapy
Several studies have recently reported GPVI as a potential target for antithrombotic therapy, which would not affect the physiological homeostasis.[11]
[12] Remarkably, reduction of thrombus formation by collagen ex vivo was observed in a patient with GPVI deficiency.[13] Additionally, the GPVI-receptor pathway is relevant in human cardiovascular disease, as the expression level and activity of GPVI correlates with the development of transient ischemic attack,[14] stroke,[14] acute coronary syndrome,[15] and myocardial infarct.[16] These data are in accordance with observations that GPVI inhibition leads to reduction of thrombus formation.[17]
[18]
[19]
[20]
Furthermore, GPVI deficiency was identified in patients who have, at the most, a slight tendency for mild bleeding,[13]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32] suggesting that isolated GPVI deficiency does not result in severe bleeding.[11] Similarly to humans, GPVI-deficient mice demonstrated no impact on bleeding time, but collagen-induced platelet activation and thrombus formation is dramatically reduced.[17]
[33] These observations strongly suggest that physiological hemostasis does not require GPVI. This supports the concept that pharmacological modulation of GPVI can open novel strategies for “bleeding-free” antithrombotic therapies. Finally, GPVI not only mediates platelet activation at the site of vascular injury where collagen is exposed, but also is implicated in the pathogenesis of other disorders, such as rheumatoid arthritis[34] and tumor metastasis.[35] These results suggest that pharmacological targeting of GPVI may lead to beneficial effects in a wide range of platelet-related diseases.
In summary, GPVI is a potential target for antithrombotic therapy because:
-
GPVI is only expressed in platelets and megakaryocyte (platelet precursor cells),[36] potentially leading to high cell specificity and reduced side effects.
-
GPVI inhibition decreases collagen-induced platelet function with minimal impact on bleeding.
-
The beneficial effect of GPVI inhibition on cardiovascular diseases is well-documented (for more information see review Andrews et al).[12]
Role of GPVI on Platelet Function and Hemostasis
Role of GPVI on Platelet Function and Hemostasis
Several studies have reported the effect of GPVI deficiency on human platelet function and their main findings are summarized in the [Table 1]. In 1987, the first case of human GPVI deficiency was reported, indicating the important role of GPVI on platelet function.[21] The subsequent studies using the patient's blood were crucial to expand the knowledge about GPVI by demonstrating: (1) the involvement of spleen tyrosine kinase (Syk) in GPVI intracellular signaling[37]
[38] (2) the validation of collagen-related peptide (CRP) as a GPVI-specific agonist[39] and (3) the crucial role of GPVI on platelet degranulation induced by collagen.[39] Next, additional studies, using GPVI-deficient human platelets, have demonstrated not only the importance of GPVI as a collagen receptor, but also that its absence leads, in the most severe cases, to mild bleeding (average bleeding time measured by the Duke's method – GPVI-deficient patients: 9 minutes vs. normal subjects: 3–5 minutes).[21]
[24]
[26]
[27]
[31] Importantly, several reports have shown thrombus formation is dramatically reduced when human platelets lack GPVI, indicating the GPVI receptor as a potential target for inhibition of thrombus formation, with a minimum impact on patients' bleeding time.[13]
[22]
[25]
[28] Here, it is important to highlight that the beneficial effect of preventing bleeding by GPVI inhibition might be either underestimated by the presence of other platelet-related disorders, such as thrombocytopenic purpura, systemic lupus erythematosus, and gray platelets syndrome, which affects the count and the structure of platelets and contributes to a longer bleeding time or overestimated due to the fact that FcRγIIA, a major FcR expressed only in higher primates, is most likely, activated in patients with GPVI deficiency induced by shedding.
Table 1
Case reports of GPVI deficiency and their symptoms
|
Comorbidity
|
AB[a]
|
Platelet count[b]
|
Bleeding time[c]
|
Patient's platelet changes
|
Clinical cases
|
Ref
|
|
Thrombocytopenic purpura and hyperthyroidism
|
Yes
|
13–140
|
10 min (D)
|
Normal aggregation and ATP release in response to standard agonists but showed an absent aggregation response and adhesion to collagen. Low expression of GPVI receptor. Syk phosphorylation induced by collagen is reduced in GPVI-deficient platelets. Validation of CRP as a GPVI-specific agonist
|
55 y/o woman with recurrent cutaneous ecchymoses, frequent episodes of epistaxis and gum bleeding, thirst, excessive sweating, palpitation, and weight loss. A bone marrow aspirate showed hypercellularity with immature megakaryocytes. Serum triiodothyronine and total thyroxine levels were elevated, and thyroid-stimulating hormone level was decreased. No past or family history of bleeding
|
[21]
[37]
[38]
[39]
[40]
[41]
|
|
None
|
N/A
|
285
|
8.5 min (I)
|
Normal aggregation in response to standard agonists but collagen-induced response and adhesion were markedly reduced. Low expression of GPVI but normal level of other receptors. Platelets release ATP
|
26 y/o woman had a tendency to develop purpura since she was a young child, occasionally had hypermenorrhea. However, she never had any severe bleeding problems. Parents showed no history of bleeding
|
[23]
[42]
|
|
None
|
N/A
|
100–130
|
13 min (D)
|
Normal aggregation, ATP and Ca2+ release in response to standard agonists but not to collagen. Adhesion to collagen-I was decreased. Thrombin, but not collagen, promoted thromboxane B2 formation. Low expression of GPVI but normal level of other receptors
|
26 y/o man presented spontaneous superficial bleeding and petechiae and multiple ecchymoses over the four extremities and the trunk and hemorrhagic diathesis. No past or family history of a bleeding
|
[26]
|
|
None
|
N/A
|
110–140
|
5 min (D)
|
Normal aggregation and ATP release in response to standard agonists but showed an absent response to collagen. Platelet adhesion to collagen was markedly reduced. Low expression of GPVI but normal level of other receptors
|
56 y/o woman with purpura and epistaxis since childhood. After hysterectomy for myoma uteri developed excessive bleeding despite pretreatment with platelet transfusion. Parents showed no history of bleeding
|
[24]
|
|
Systemic lupus erythematosus and chronic thyroiditis
|
Yes
|
21–303
|
9 min (D)
|
Normal aggregation and ATP release in response to standard agonists but not to collagen. Platelet adhesion to collagen was markedly decreased. Low expression of GPVI but normal level of other receptors
|
47 y/o woman presented mild subcutaneous bleedings. She had been suffering from systemic lupus erythematosus since age 35. Platelet-associated IgG was increased. No history of bleeding
|
[27]
|
|
Thrombocytopenic purpura
|
Yes
|
105–135
|
> 15 min (I)
|
Normal aggregation and granule release in response to standard agonists but not to collagen and CRP. Adhesion to and poor thrombus formation on immobilized collagen under shear conditions. Binding to and spreading on CRP were reduced. Low expression of GPVI but normal level of other receptors
|
25 y/o woman had an onset of easy bruising, petechiae, weight loss, and fatigue over a 6-mo period. No family history of bleeding
|
[28]
|
|
Gray platelet syndrome
|
No
|
18–35
|
N/A
|
Platelets lacked α-granules. Normal aggregation in response to standard agonists but epinephrine-, CVX- and collagen-induced response was reduced. Low expression of GPVI and FcRγ but normal level of other receptors
|
55 y/o woman (parents are first cousins). She presented a lifelong bleeding syndrome with epistaxis, bleeding at the time of menarche, hemorrhage during surgery and gastric bleeding without detectable local lesions, and hypersplenism. The patient developed deep vein thrombosis. Bleeding problems during child births (2 sons), which are both healthy, no bleeding syndrome
|
[29]
|
|
Comorbidity
|
AB[a]
|
Platelet count[b]
|
Bleeding time[c]
|
Patient's platelet changes
|
Clinical Changes
|
Ref
|
|
Immune thrombocytopenic purpura
|
No
|
48
|
> 15 min (I)
|
Normal aggregation in response to standard agonists but, CVX-, CRP-, and collagen-induced response was reduced. The aggregation response to collagen was completely abolished by α2β1 inhibitor. Adhesion to collagen under static condition was reduced and abolished by α2β1 inhibitor. Adhesion to immobilized collagen under high shear rate but no formation of large aggregates. Low expression of GPVI and FcRγ but normal level of other receptors
|
20 y/o woman, with no previous history of bleeding, presented spontaneously occurring petechia and gingival bleeding
|
[22]
|
|
Immune thrombocytopenic purpura
|
Yes
|
2
|
N/A
|
Normal aggregation in response to standard agonists but, CRP- and collagen-induced response was reduced. PPP induced aggregation of PRP isolated from a healthy donor, which was impaired by inclusion of human recombinant GPVI, or partially blocked by FcγRIIa-blocking antibody. Prolonged collagen-epinephrine PFA-100 and collagen-ADP closure time. Low expression of GPVI but normal level of other receptors
|
55 y/o female presented with epistaxis and lower limb purpura. No past or family history of bleeding
|
[30]
|
|
None
|
N/A
|
208
|
N/A
|
Normal aggregation and ATP release in response to standard agonists but no response to collagen, CVX, and CRP. Perfusion of whole blood on collagen at intermediate shear stress showed a reduced aggregate formation, but also an increased adhesion of single platelets. PFA100 closure time for collagen/ADP was shortened. No P-selectin expression after stimulation by CRP. Reduction in GPVI expression. Normal expression of FcRγ receptor and other platelet receptors
|
31 y/o woman presented ecchymoses, epistaxis, several posttraumatic and postsurgery bleeding complications since her childhood, and menorrhagia. Parents showed no history of bleeding, but the patient carries a paternally inherited out-of-frame 16-bp deletion and a maternally inherited missense mutation p.S175N in a highly conserved GPVI residue
|
[25]
|
|
None
|
N/A
|
280
|
> 15 min (I)
|
Normal aggregation and ATP release in response to standard agonists but, no response to collagen and CVX. Platelet aggregate formation on collagen in flow conditions was impaired. Defect in collagen-activated platelet-catalyzed thrombin generation with decreased endogenous thrombin potential. Reduced GPVI expression. Normal expression of other platelet receptors
|
10 y/o girl presented bruising since infancy. Her parents were nonconsanguineous, and there was no family history of bleeding. Heterozygote for 2 GPVI mutations: incomplete protein deficiency, and loss of function Arg38 (R38) to Cys substitution in the first GPVI Ig-like loop
|
[13]
|
|
None
|
No
|
80
|
> 10 min (D)
|
Collagen- and epinephrine-induced platelet aggregation was impaired. Biotinylated CVX and anti-GPVI antibody binding was reduced. Low expression of GPVI and FcRγ, but normal level of other receptors
|
12 y/o girl presented cutaneous ecchymosis and recurrent epistaxis since 9 y/o. Several petechiae in the upper and lower limbs and chest. No family history of a bleeding tendency or consanguinity
|
[31]
|
|
None
|
N/A
|
Normal
|
13 and
9 min (I)
|
Normal aggregation response to arachidonate and ADP. Aggregation and serotonin secretion were reduced in response to collagen, CVX, and CRP. The GPVI receptor expression level was reduced
|
Two women (22 and 11 y/o) and two men (23 and 12 y/o) with easy bruising, epistaxis, and prolonged bleeding after minor injuries and gum bleeding. Symptoms started in early childhood. None of the patients had a family history of bleeding, there was no consanguinity
|
[32]
|
Abbreviations: ADP, adenosine diphosphate; ATP, adenosine triphosphate; CRP, collagen-related peptide; CVX, convulxin; GPVI, glycoprotein VI; IgG, immunoglobulin G; PPP, platelet-poor plasma; PRP, platelet-rich plasma.
a Is plasma antibody (AB) against GPVI detected?
b In (×109/L). Normal platelet count: 15–45 × 109/L.
c Normal bleeding time: (D)uke's test: normal 3–5 minute; (I)vy method: normal 3–8 minute.
It is well documented that FcγRIIa interacts with several platelet receptors and it has been reported that the cross-linking of GPVI with FcγRIIa seems to be important for platelet activation (for more information see review by Qiao et al[43]). In addition, the activation of FcγRIIa induces metalloproteinases-mediated ectodomain shedding of GPVI.[30] Based on these findings, one can speculate that the bleeding-free outcome induced by GPVI inhibition might be due to FcγRIIa activation, which could explain the pathological mechanism of patients with GPVI deficiency induced by shedding. Studies which aimed to clarify the importance of FcγRIIa on GPVI-induced platelet activation are contradictory, to date. For example, the antibody OM1, induces GPVI-dependent platelet aggregation without the requirement of FcγRIIa, whereas other GPVI antibodies, OM2 and 9O12.2, activated platelets in an FcγRIIa-dependent manner.[44]
[45] In addition, a recent report demonstrated that spreading and adhesion on fibrin is dramatically decreased in GPVI-deficient mouse platelets but are not altered in human platelets treated with monoclonal FcγRIIa-inhibiting antibody IV.3.[46] Therefore, further studies are needed to understand whether the beneficial antibleeding effect of GPVI inhibition is independent of FcγRIIa activation. This could be assessed by using, for example, platelets from GPVI-deficient patients.
Although recent studies have speculated that fibrin(ogen) is a potential GPVI agonist, platelets from GPVI deficiency patients were not tested upon stimulation with fibrin(ogen). Apart from one study in which platelets were reported to lack α-granules,[29] platelet aggregation and degranulation induced by non-GPVI agonists and the expression of other platelet receptors are preserved in patients with GPVI deficiency. In contrast, the aggregation response induced by the GPVI agonists, collagen-I, CRP, and convulxin (CVX), is dramatically impaired. In most cases, platelet aggregation is abolished in response to CRP and CVX, and it is reduced at high concentrations of collagen-I (> 10 μg/mL). These findings corroborate several studies in GPVI-knockout animals where platelet aggregation was reduced with a minor effect on platelet adhesion and a reduction of thrombus formation without any increase of bleeding time.[17]
[33] Collagen-induced platelet aggregation at high concentrations is only abolished when platelets are preincubated with α2β1-integrin inhibitor.[22] Collagen-I also binds to α2β1 integrin, and activation of this receptor triggers platelet aggregation, although it requires much longer time to reach the maximal aggregation, when compared with collagen or CVX.[47] In addition, human deficiency of α2β1 integrin on platelets results in severe bleeding, indicating a crucial role of this platelet collagen receptor for maintenance of hemostasis, and less relevance in platelet aggregation.[48]
[49]
[50]
[51]
Taken together, these results demonstrated that GPVI is a key mediator of collagen-induced platelet aggregation and, differently from α2β1 integrin, GPVI deficiency does not cause severe bleeding. In addition, the lack of GPVI does not interfere in platelet function (aggregation, adhesion, and degranulation) if induced by the activation of non-GPVI-specific platelet agonists.
Biochemical Structure of GPVI
Biochemical Structure of GPVI
According to its complementary deoxyribonucleic acid cloning, in humans, GPVI consists of 319 amino acids residues, with a 20-amino-acid signal sequence. The GPVI gene presents eight exons and is located in the chromosome 19q13.4 of the human genome.[52]
[53]
GPVI receptor is included in the immunoglobulin (Ig)-like receptor family, since its extracellular region possesses two Ig-like domains, designated D1 and D2, and a mucin-like Ser/Thr-rich region.[52]
[54] The domains D1 and D2 are linked by a single peptide strand that allows limited flexibility in the head of the receptor and a glycosylated stem connects D2 to the transmembrane domain.[54] One potential glycosylation site for the N-linked carbohydrate chain was identified at R272 and many O-linked carbohydrate chains are conjugated to the Ser/Thr-rich region.[55] Therefore, it is possible that the Ig-like domains extends and tower above the glycocalyx polysaccharide layer of the platelet surface, structurally similar to GPIb/V/IX.[52]
The cytoplasmic domain of GPVI, consisting of 51 amino acids in humans, has two unique sequences: a basic amino-acid-rich region near the transmembrane region and a proline-rich motif in the middle of the cytoplasmic region.[52]
[54] While the basic amino-acid-rich region binds to calmodulin,[56] the proline-rich sequence selectively binds to the proto-oncogene tyrosine-protein kinase (Src) homology 3 (SH3) domain of the Src family tyrosine kinases Fyn and Lyn.[57] The C-terminal region seems to be irrelevant for GPVI function, since the cytoplasmic region of mouse GPVI contains only 27 residues and lacks the additional 24 residues which make up the C-terminal sequence of human GPVI.[52]
As summarized in [Table 2], several site-directed mutagenesis studies were performed, revealing crucial amino acids residues for GPVI-agonist binding. Mutations at residues V34, K59, R60, and R166 markedly decreased the binding of recombinant GPVI to collagen or CRP, whereas N92, S94, and R272 decreased the cell response to CVX and CRP. In contrast, mutation of lysine 41 to alanine increases the binding of GPVI to both CRP and collagen. The importance of these amino acids residues for binding to collagen needs to be confirmed by crystallography. In 2006, Horii et al revealed its crystal structure and thus advanced our understanding of the GPVI receptor and its putative molecular mechanism of collagen binding. Moreover, the molecular structure of the receptor confirmed previously published data and, in addition, disclosed that two GPVI molecules associate into a back-to-back dimer. The drawback of this study is the fact that the GPVI-ligand complex was predicted by computational docking, only the noncrystal structure of the complex was presented.[63] Recently, an additional three crystal structures were added in the Protein Data Bank database (ID: 5OU7, 5OU8, and 5OU9) demonstrating GPVI-collagen-peptide complexes, which provided essential information to identify the collagen binding site within the GPVI receptor.[64] Unfortunately, all data and methodology related with these new complexes have yet to be published.
Table 2
Site-directed mutagenesis studies using GPVI receptor
|
Point of mutations
|
Effect
|
Ref
|
|
K59, R60, F91, R117, Y118, F120, R139, S164, R166
|
Mutation of K59 markedly decreases the binding of GPVI to CRP. Triple mutation (K59E, R117P, R166S) induced the most profound reduction of binding to CRP
|
[58]
|
|
K41, R60, R166, K59, F120, S164
|
Mutations of R60 and R166, as well as K59, reduce binding of GPVI to collagen, whereas K41 increases the binding of GPVI to both CRP and collagen
|
[59]
|
|
G30, V34, L36
|
V34 strongly impairs GPVI interaction with collagen and CRP, whereas L36A mutation slightly reduces the binding capacity. Triple mutation (G30, V34 and L36) fully inhibits the binding to collagen and CRP (but not to CVX). None of the recombinant GPVI mutations affects the maximal binding to the GPVI antibody 9O12
|
[60]
|
|
N92, S94, L95
|
Mutation of N92 and S94 dramatically reduces the adhesion of transfected-cells to CVX and CRP. Whereas mutation of L95 does not change the adhesion to GPVI-specific agonists. N-linked glycosylation at the consensus site Asparagine92-Glycine-Serine94 of human GPVI and the presence of complex oligosaccharides at this position is important for maximum binding of GPVI to both CVX and CRP
|
[61]
|
|
R272, W291, H292, S293
|
Mutation of R272 decreases the Ca2+ release in response to CVX in transfected cells. In addition, cells expressing the R272-mutated GPVI are unable to co-precipitate with FcR γ-chain
|
[55]
|
|
C274, C338
|
C338 is important for GPVI dimer formation in CVX-stimulated cells
|
[62]
|
Abbreviations: CRP, collagen-related peptide; CVX, convulxin; GPVI, glycoprotein VI.
GPVI Dimerization and Binding Specificity
GPVI Dimerization and Binding Specificity
Initially, GPVI was described to form a complex with the Fc receptor (FcR)γ-chain, which is expressed on other hematopoietic cells.[65] Complex formation with the FcRγ-chain is essential for GPVI signaling.[66] However, whether GPVI is exposed on the cell surface independently of complex formation is not clear. Some studies, where GPVI and FcRγ-chain were transfected into human cells, showed that GPVI is expressed on the cell surface, independently of complex formation, but is unable to signal on its own.[66] The interaction between GPVI and the FcRγ-chain depends on a salt bridge between the charged amino acid, R272, located in the transmembrane domain and a negatively charged aspartic acid residue, residing in the transmembrane domain of the FcRγ-chain.[55] The GPVI receptor appears on the platelet surface, also in its oligomeric form, then each GPVI molecule is in complex with the FcRγ-chain.[67] In addition, a monomeric GPVI exhibits low affinity for collagen when compared with a dimeric GPVI,[68] which may explain why the oligomeric form of this complex occurs on the surface of platelets.
Collagen is an important GPVI ligand. In the vasculature, different collagens have been detected, including: collagen types I, III–VIII, XIII–XVI, XVIII, and XIX.[69] However, only some collagen types bind to GPVI, such as the large fibril-forming collagen I, II, and III. To evaluate collagen IV as a ligand may prove to be difficult as it is a component of the complex basement membrane (BM), which is composed of a variety of proteins, and it is present as a thin sheet-like structure.[70] The reduced affinity of collagen V to GPVI can be explained by either the small size of the collagen V fibrils not covering the binding site of the GPVI dimer properly[71] or due to its high content of glycosylated hydroxylysine residues within the triple helical molecule portion which may impair their interaction.[72] The adhesion of GPVI to collagens IV and V has so far provided confounding results. Some authors have described platelets from patients lacking the GPVI not adhering to bovine collagens IV and V.[23]
[73] Others have observed adhesion to these collagens, but they attributed it to possible contamination of the preparation with other collagens, such as collagen-I.[71] Yet, another study demonstrated that type IV and type V collagens did not have any affinity for GPVI.[52] Nonetheless, a role in GPVI-dependent platelet activation, at superficial vascular injury sites, has been attributed to collagen IV.[74]
GPVI-collagen interaction and platelet activation occur specifically through the amino acid motif glycine–proline–hydroxyproline (GPO), which comprises approximately 10% of the triple helix in collagen molecules.[75] The observation that 5 or more GPO repeats spontaneously align into triple helical structures that are similar to tissue collagen molecules prompted the development of synthetic peptides.[76] Moreover, such triple-helical collagenous molecules, with three chains of repetitive GPO motifs, were laterally bundled by covalently cross-linkage of their N- and C-terminal cysteine or lysine residues, forming the so-called CRP, a powerful synthetic platelet agonist, specifically binding to GPVI.[77]
During collagen-dependent platelet activation, multiple GPVI binding sites are clustered. Furthermore, under flow conditions, platelet adhesion to collagen-I depends on interaction with other collagens, such as collagen-III, indicating that ligand interactions may support platelet receptor clustering.[71] Another study has shown that not only integrin α2β1,[47] but also GPVI form clusters along collagen fibers, with the homotrimeric collagen-III being the most effective substratum for cluster formation.[78]
Both fibrinogen and fibrin have recently been described as GPVI ligands. Several studies demonstrated that fibrin and fibrinogen induce platelet signaling, as well as support thrombus formation and stabilization.[79] A marked increase of tyrosine phosphorylation, including in the FcRγ chain and Syk in human platelets, increased calcium (Ca2+) signaling, as well as attachment and spreading of platelets, have been observed in response to fibrin(ogen), supporting the hypothesis that fibrin(ogen) is a potential GPVI ligand, in addition to binding to the platelet fibrin receptor, integrin αIIbβ3.[80]
[81]
Among other endogenous ligands, the adipocyte-derived cytokine, globular adiponectin (gAd), was also proposed to bind to GPVI. This interaction leads to rapid aggregation of platelets, in a tyrosine kinase-dependent mechanism.[82] Studies using FcRγ null mouse platelet showed that aggregation by gAd was abolished in platelets, and the authors concluded that gAd may promote adverse platelet activation at sites of vascular injury.[82] However, the functional relevance of this putative GPVI-gAd interaction has remained elusive, since circulating levels of gAd are not sufficient to mediate platelet activation.[70]
Laminin-411 and -511 are major components of the blood vessel BM. These laminin isoforms might be additional endogenous ligands of GPVI. Isolated laminin from human placenta, with a high content of laminin-511, supported the spreading of platelets through a GPVI-dependent mechanism.[83] This interaction depends on initial integrin α6β1 adhesion to laminin.[83] The authors suggested that GPVI interaction with laminin might contribute to platelet spreading in vivo. Inoue et al also observed that platelets adhere to laminin under shear conditions, although forming smaller aggregates than the platelets that adhered to collagen, which the authors explained by the fact that GPVI has lower affinity for laminin as compared with collagen.[84] The mechanism of shear-resistant attachment of platelets to laminin depends on the interaction of VWF and GPIb-IX-V.[84] Another study by Schaff et al attributed to GPVI yet another important function, not only in the adhesive process, but in the mediation of platelet activation induced by α5-containing laminins.[85] Consequently, the study concluded that laminins contribute to thrombus formation in vivo. This latter interaction could be critical for vessel repair following minor damage, without the risk of forming occlusive thrombi. Indeed, after mild injury, where the sporadic detachment of an endothelial cell can occur, these subendothelial BM proteins are the first ECM components that platelets come into contact with.[70] However, in the prevention of major bleeds occurring after severe damages reaching far beyond the vessel wall, laminins might be irrelevant for platelets' thrombotic functions, as it induced only a weak activation of GPVI, as compared with the fibrillary collagens of the interstitial stroma.[70]
Intracellular Signaling Downstream of GPVI
Intracellular Signaling Downstream of GPVI
Firm platelet adhesion to collagen and fibrin via their respective integrin receptors requires prior integrin activation through “inside-out” signals, generated by GPVI signaling and reinforced by the release of ADP and TXA2.[76]
Platelet signaling is initiated through the immunoreceptor tyrosine-based activation motif (ITAM) that consists of two YXXL motifs, separated by 6–12 residues,[86] in the cytosolic domain of several signaling receptors. The FcRγ chain harbors such an ITAM motif. Other platelet receptors, such as CLEC-2, contain only one singular YXXL motif, termed (hem)ITAM, but complement the two hemITAM motifs, resulting in a fully functional ITAM structure, by dimerization of two molecules. The GPVI-FcRγ complex is critical for the signaling activity of platelets and thrombus formation, similar to other immune and C-type lectin receptors.[87]
[88] Moreover, FcγRIIA mediates platelet integrin αIIbβ3 “outside-in” signaling, thus contributing to general thrombus stabilization.[89]
The cross-linking of GPVI triggers the phosphorylation of the two cytosolic motifs in the FcRγ-chain ITAM by Src family tyrosine kinases, leading to binding and activation of the two tandem Src Homology 2 (SH2) domains by Syk. Consequently, this triggers the downstream signaling cascade that culminates in the activation of phospholipase Cγ2 (PLCγ2).[90] However, prior to activation of PLCγ2, various adapter and effector proteins are recruited. The ITAM phosphorylation occurs in the cholesterol-rich membrane lipid rafts, that are enriched in essential GPVI signaling molecules.[91] Interestingly, studies using rat GPVI-expressing RBL-2H3 cells and human platelets, have shown that the GPVI-FcRγ complex does not constitutively associate with rafts, but upon ligand stimulation, it is recruited to lipid rafts.[91] Within lipid rafts, an essential protein for the formation of the signalosome can be found, which is the linker for activation of T cells (LAT).[92] However, some studies have shown that phosphorylation of Syk, LAT, and PLCγ2 does not depend on the integrity of the lipid rafts in human platelets.[93] The Src kinases, Fyn and Lyn, are also present in rafts and bind to a conserved proline-rich domain in the juxtamembrane region of GPVI, resulting in the phosphorylation of the FcRγ immune receptor ITAM sequences, which are essential for triggering downstream signals.[57] While Lyn binds preferentially to the proline-rich domain and is required for rapid cell adhesion to collagen at high shear, Fyn does not bind to this domain, as it seems to be more important to sustain than to initiate the signaling events.[94] Studies by Quek et al which knocked out Lyn, Fyn, or both in mouse models further revealed that, although Fyn deficiency apparently does not affect the tyrosine phosphorylation of PLCγ2, its activity is significantly reduced, possibly due to the fact that phosphorylation of adapter proteins, such as LAT, was diminished. This resulted in reduced recruitment of PLCγ2 to the membrane.[95]
The activation of tyrosine kinase Syk leads to the phosphorylation of tyrosine on the adapter protein LAT.[70] After phosphorylation, LAT recruits the SH2-containing adapter proteins Grb2 and Gads, which constitutively associate with SLP-76.[96] PLCγ2 binds to the complex, through adapter proteins LAT and SLP-76, and generates the second messengers, Ca2+ and diacylglycerol (DAG).[87] Syk also plays an essential role in activating PLCγ2, since it phosphorylates SLP-76.[97] While SLP-76 is essential for tyrosine phosphorylation and activation of the phospholipase, the absence of LAT resulted in a limited degree of tyrosine phosphorylation activation of PLCγ, but it is unclear whether LAT is required for full platelets activation.[98]
Additional adapter and effector proteins were identified to assemble into the ITAM signalosome, such as the signal transducer and activator of transcription 3 (STAT3),[99] the Tec family tyrosine kinases Btk[100] and Tec,[101] the small G protein Rac1,[102] the GTP exchange factors Vav1 and Vav3,[103] the ubiquitinating protein, c-Cbl,[104] and the α and β isoforms of phosphatidylinositol 3-kinase (PI3K).[105] All these proteins are involved in the signaling process within the platelet. Upon collagen stimulation, platelets activate STAT3, which serves as a protein scaffold to facilitate the catalytic interaction between the Syk kinase and its substrate, PLCγ2, which releases further secondary messengers.[99] Studies using human platelets and STAT3 null mouse platelets have indicated that STAT3 facilitates the crosstalk between proinflammatory cytokines and hemostasis/thrombosis signals in platelets.[99] Btk works together with Tec, since Tec appears to have a functional role in the regulation of PLCγ2 in the absence of Btk. Nevertheless, these proteins are essential prerequisites for platelet activation and aggregation by GPVI, regulation of tyrosine phosphorylation of PLCγ2, as well as increase in Ca2+.[101] Additionally, Vav1 and Vav3, guanine nucleotide exchange factors (GEFs) for the Rho/Rac family, have a role in the activation of PLCγ2.[103] This process happens by tyrosine phosphorylation upon GPVI activation and platelet deficiency in both Vav1 and Vav3 decrease aggregation and spreading, as a consequence of reduced phosphorylation of PLCγ2.[103] The c-Cbl protein has been described to be tyrosine-phosphorylated in response to stimulation of GPVI; however, the role of this protein is not clear. Some studies have described it as an adaptor protein that promotes PI3Kβ activation and consequently, activation of PLCγ2 in a knockout mouse model.[104] An earlier study showed that phosphorylation of several proteins, such as Syk and PLCγ2, was increased in the absence of c-Cbl on the surface of murine platelets, leading to the conclusion that it negatively regulates platelet responses to GPVI agonists.[106] PI3K-α and -β are involved in GPVI-mediated inositol 1,4,5-triphosphate (IP3) production and Ca2+ mobilization, via PLCγ2, and contributes to Rap1b activation.[105] Activated PLCγ2 hydrolyzes phosphatidylinositol 4,5-bisphosphate, generating the second messengers DAG and IP3, which in turn, mediates the release of Ca2+, whereas DAG activates protein kinase C (PKC).[107] The GTPase Rap1 is activated downstream of PLCγ2 in a Ca2+-dependent process, where Ca2+ is sensed by GEF, CalDAG-GEFI, which will then facilitate this activation.[87]
The GPVI signaling pathway can be inhibited by Src kinases if a relevant tyrosine residue in an inhibitory site is phosphorylated. Hence, dephosphorylation of this site is critical for activation[70] and seems to depend on the membrane tyrosine phosphatase, CD148. In the absence of this phosphatase, Src family kinase activity is reduced and GPVI-mediated activation is impaired, resulting in defective thrombosis.[108] Lyn plays an inhibitory role in mouse platelet activation through tyrosine phosphorylation of platelet endothelial cell adhesion molecule-1 and subsequent binding of the SH2 domain-containing phosphatase.[109] Additionally, this complex suppresses GPVI signaling.
CLEC-2 a GPVI Homologous Receptor
CLEC-2 a GPVI Homologous Receptor
The C-type lectin receptor 2 (CLEC-2) is a recently discovered platelet receptor that shares structural and functional similarities with GPVI.[88] It is a type-II membrane protein with a C-terminal extracellular domain comprising a stalk region and a carbohydrate-like recognition domain that lacks the conserved amino acids necessary for binding carbohydrates.[70] For this reason, CLEC-2 is considered a C-type lectin-like domain-containing protein.[110] The cytoplasmic N-terminal tail of 31 amino acids, contains a single and conserved YxxL sequence, an hemITAM, starting at position 7, downstream of a tri-acidic amino acid region (DEDG).[70] Similarly to GPVI, CLEC-2 is highly expressed on megakaryocytes and platelets but also at low levels on primary human liver sinusoidal endothelial cells and on neutrophils, and possibly on other hematopoietic cells.[70]
The human CLEC-2 gene, also known as CLEC1B, is located on chromosome 12 in a cluster with six other C-type lectin receptors. An example of this is Dectin-1, which, like CLEC-2, is a member of group V of the type II C type lectin-like receptors.[70]
CLEC-2 has an exogenous ligand, rhodocytin, a snake venom component, which upon binding to CLEC-2 triggers tyrosine phosphorylation of the CLEC-2 hemITAM sequence and, consequently, other signaling proteins.[111] The first endogenous ligand for CLEC-2 was identified by the group of Yukio Ozaki, and described as a transmembrane sialoglycoprotein, podoplanin.[112] Moreover, podoplanin expressed by tumor cells mediates aggregation of platelets, in a Src kinases/ PLCγ2-dependent manner, similarly to rhodocytin.[112] Other authors linked podoplanin from kidney HEK-293T cells to platelet activation.[113] Podoplanin can be found in different tissues like brain, heart, kidney, lungs, osteoblasts, and lymphoid organs,[114] but not on blood cells or platelets.[87] This raises a controversy on the importance of CLEC-2 and its main ligand in hemostasis, leading to different studies trying to address this issue. Studies from May et al, using CLEC-2-deficient mouse platelets, demonstrated that this C-type lectin receptor is an essential mediator for platelet activation in vitro and in vivo, and that it is involved in aggregate and thrombus formation, and hence in hemostasis.[115] However, another study using CLEC-2-deficient mice showed normal hemostasis in these animals.[116] CLEC-2 has also been implicated in thrombus stabilization, both in vitro and in vivo, despite the fact that bleeding tendency was not increased, contrary to previous data.[117] Nevertheless, if there is a role for CLEC-2 in hemostasis, this cannot be attributed to podoplanin binding, since it is absent from the vascular injury site.[87] The only circumstance in which podoplanin could be relevant in the thrombotic process is in advanced stages of atherosclerosis, since its expression was significantly enhanced in advanced atherosclerotic lesions, compared with early lesions.[118] CLEC-2 actions in hemostasis itself, even if minor, would have to be triggered by ligands other than podoplanin, which have been already suggested before but not yet identified.[117]
CLEC-2 exists as a homodimer on the cell surface, which enables Syk's tandem SH2 domains to bind to the two phosphorylated hemITAM motifs of CLEC-2 homodimer, leading to tyrosine phosphorylation of downstream adapter proteins and effector enzymes, including PLCγ2, and consequently, platelets activation.[119]
Although CLEC-2 and the GPVI with its associated FcRγ-chains belong to the same family of ITAM receptors, and they trigger similar signaling events downstream of Syk, there is increasing evidence that the events in their signaling cascades are mechanistically distinct.[70]
[87] The intracellular domain of CLEC-2 contains a single YxxL motif, similar to GPVI.[87] CLEC-2 signals through Src and Syk tyrosine kinases, leading to tyrosine phosphorylation and recruitment of adapter proteins, Tec family tyrosine kinases, and various effector proteins including PI 3-kinase, Vav, Rac1, and PLCγ2.[120]
Platelet activation by rhodocytin, but not by collagen or by CVX, results in phosphorylation of the tyrosine residue in the YXXL motif. This phosphorylation depends on Syk and PLCγ2, and partially on LAT, SLP-76, and Vav1/Vav3.[111] The dependence on these signaling proteins differ between CLEC-2 and GPVI, since SLP-76 is an absolute requirement in GPVI signaling.[121] Although CLEC-2 also translocates to lipid rafts upon ligand engagement, phosphorylation of CLEC-2 critically correlates with actin polymerization, Rac1 activation, and release of ADP and TXA2, while the FcRγ-chain phosphorylation is independent of these requirements.[122]
A study by Bender et al targeted both GPVI and CLEC-2 simultaneously in mice with an antibody treatment, which led to defective hemostasis and impaired thrombus formation. The same was observed when GPVI and CLEC-2 were genetically inactivated in mice.[123] These studies further revealed an unexpected functional redundancy of the two receptors in hemostasis and thrombosis. Moreover, they highlighted both receptors as promising targets for antithrombotic protection, which upon inactivation cause only moderately increased bleeding times in mice.[123] These findings support the potential use of anti-GPVI and anti-CLEC-2-based agents in the prevention of thrombotic diseases with patients at lower risk for hemorrhage.
Current GPVI Inhibitors: Antibodies, GPVI-Fc Fusion Protein, and Bioactive Compounds from Plant Extracts and Snake Venom
Current GPVI Inhibitors: Antibodies, GPVI-Fc Fusion Protein, and Bioactive Compounds from Plant Extracts and Snake Venom
With the aim of decreasing thrombosis, clinical antiplatelet therapies have been developed, especially for the cohort of patients with higher tendency for cardiovascular events. A drawback of such antithrombotic therapies is the increased risk of bleeding. In several experimental models, blockage or depletion of GPVI not only impaired platelet activation and thrombus formation, but also did not result in major bleeding complications, thereby revealing striking benefits in treatment of thrombosis and thromboinflammation.[12] This identified the role of GPVI as nonessential in hemostasis but nevertheless important in thrombosis, making its therapeutic targeting unlikely to lead to severe bleeding. Another advantage of targeting GPVI is the high specificity, since GPVI expression is restricted to platelets and megakaryocytes, which is important for minimizing potential side effects of a therapeutic agent.[12] The available antiplatelet therapeutic agents and their targets in the different stages of platelet activation were listed by Andrews et al, in a review that can be consulted for more detailed information.[12]
In this section, we gathered the spectrum of various inhibitors that have been specifically developed to inhibit GPVI. The signaling pathways and the effect on platelet activation induced by GPVI-inhibition are summarized in [Table 3].
Table 3
Inhibitors of GPVI activity
|
Inhibitor class
|
Compound
|
Effect
|
Ref
|
|
Snake venom derived compounds
|
Trowaglerix venom polypeptides
|
Specific inhibition of collagen-induced human platelet aggregation
Decrease of thrombus formation without prolonging in vivo bleeding time in mice
Protein docking studies suggest that Trowaglerix bound at the lower surface of D1 domain and outer surface of D2 domain of the GPVI
|
[124]
|
|
Kistomin
|
Inhibition of collagen-, convulxin- and GPVI-specific antibody-induced human platelet aggregation in a concentration- and time-dependent manner
Decrease in platelet GPVI
Formation of GPVI cleavage fragments by cleavage near the mucin-like region (at Glu205/Ala206 and at Val218/Phe219)
Reduction of tyrosine kinase phosphorylation induced by a GPVI-specific antibody
Decrease of human platelet adhesion to collagen
|
[125]
|
|
Natural bioactive compounds
|
Buddleja globose
(polyphenols/flavonoids)
|
Inhibition of human platelet aggregation in response to collagen, convulxin and ADP
Decrease of P-selectin in platelets and PAC-1 binding
Inhibition of PLC-γ2 and PKC phosphorylation
|
[126]
|
|
Rabdosia japonica
(Glaucocalyxin A)
|
Inhibition of collagen- and CRP- triggered human platelet aggregation
Inhibition of collagen-stimulated tyrosine phosphorylation of Syk, LAT, and PLCγ2
Inhibition of P-selectin secretion and integrin activation by convulxin
Inhibition of thrombin-induced platelet activation
Attenuation of platelet adhesion on collagen surfaces in high shear condition.
Increase of the time for complete occlusion upon vascular injury in mice without extended tail-bleeding time
|
[127]
|
|
Solanum lycopersicum (lycopene)
|
Inhibition of ADP- and collagen-induced platelet aggregation in vitro, ex vivo, and in vivo
Inhibition of the phosphoinositide breakdown, intracellular Ca+2 mobilization, thromboxane B2, and PKC phosphorylation
|
[128]
[129]
|
|
Anopheline antiplatelet protein (AAPP)
|
AAPP blocks the binding site of GPVI on collagen I and III inhibiting platelet adhesion. There is no direct interaction between AAPP and GPVI, since AAPP failed to inhibit GPVI-induced platelet function on platelets
|
[130]
|
|
Cy-3-g (anthocyanin)
|
Decrease of platelet adhesion and aggregation to collagen at both venous and arterial shear stresses
Inhibition of PAC-1, P-selectin, CD63 and CD40L, and of fibrinogen binding
Decrease in thrombus growth without prolonging bleeding time in mice
Downregulation of collagen-induced GPVI signaling by decreasing phosphorylation of Syk, LAT, and SLP-76, as well as the expression of Lyn, Fyn, and PLC-γ2
|
[131]
|
|
GPVI mimetic
|
GPVI-Fc or Revacept
|
High affinity binding to immobilized collagen
Competitive inhibition of binding to collagen
Inhibition of collagen-induced human platelet aggregation
Decrease of thrombus formation in vitro and in vivo, without significantly affecting the bleeding time
|
[132]
[133]
[134]
|
|
Inhibitory monoclonal antibodies
|
JAQ1
|
Depletion of GPVI in mice platelets
Abolition of the responses to collagen and CRP
Protection from thromboembolism in mice with moderate tail bleeding times
Transient decrease of thrombin-induced activation of integrin αIIbβ3, P-selectin expression
Temporary protection from lethal tissue factor-induced pulmonary thromboembolism in mice
|
[19]
[33]
|
|
mF1232
|
Inhibition of collagen-induced human platelet aggregation in vitro
Immunodepletion without thrombocytopenia nor GPVI shedding in monkeys
Mutation in the following residues decreased the binding of mF1232 to GPVI: Gly30, Val34, and Leu36
|
[135]
|
|
Inhibitory antibody Fab fragments
|
9O12.2 Fab fragments/ scFv
|
Inhibition of collagen- and CRP-induced platelet aggregation
Inhibition of the procoagulant activity of collagen-stimulated platelets and adhesion to collagen under static condition
Prevention of platelet adhesion and thrombi formation under flow conditions
Simultaneous mutation of Gly-30, Val-34, and Leu-36 resulted in full inhibition of the binding to CRP
|
[45]
[60]
[136]
|
|
OM2
|
Inhibition of ex vivo collagen-induced platelet aggregation
Minor increase of skin bleeding time in cynomolgus monkeys
No induction of thrombocytopenia
No GPVI depletion
|
[137]
|
|
OM4
|
Inhibition of collagen-induced aggregation of rat platelets in vitro/ex vivo
Acute thrombocytopenia
No GPVI depletion
Reduction of the number of complete occlusions in an animal model of cyclic flow reduction, without prolongation of bleeding time
|
[138]
|
|
Fab SAR264565
|
Binding to GPVI with high affinity
Blockage of GPVI function in human platelets
Reduction in platelet and fibrin deposition in the perfusion chamber
|
[139]
|
|
b-, m-Fab
|
b- and m-Fab bind to GPVI dimer but not to GPVI monomer
Platelet aggregation is induced by b-Fab while m-Fab blocks GPVI-induced platelet aggregation
|
[140]
|
|
ACT017
|
Binding to GPVI with high specificity and affinity
Inhibition of collagen-induced platelet aggregation ex vivo
No thrombocytopenia in macaques
No GPVI depletion
No bleeding side effects
|
[141]
|
|
Inhibitory antibody scFvs fragments
|
10B12
|
Binding to the collagen-binding site between the two Ig-like domains
Inhibition of CRP- and collagen-dependent activation of platelets in aggregometry and thrombus formation in whole blood perfusion
Complete blockage of human platelet thrombus formation onto lipid-rich atheromatous plaques in flowing blood
Impairment of the initial platelet response, shape change, induced by plaque
Residue K59 is critical to 10B12 binding
|
[18]
[58]
|
|
1C3
|
Potentiation of the effect of 10B12 on platelet aggregation induced by collagen and cross-linked collagen-related peptide
Inhibition of a proportion of GPVI binding to its ligands, most likely indirectly
Inhibition of clustering of the Ig-like domains of GPVI on collagen/CRPs
The mutation of D1D2 I148A abolished the ability to bind 1C3
|
[142]
|
|
Inhibitory antibody dAbs fragments
|
BLO8–1
|
Binding to the collagen binding domain of GPVI
Specific inhibition of the binding of recombinant human GPVI to cross-linked CRP in vitro
Prevention of cross-linked CRP induced platelet aggregation
Inhibition of thrombus formation in whole blood
Residue K59 is critical for BLO8–1 binding
|
[20]
|
|
GPVI signaling pathway inhibitors
|
LFM-A13
|
Specific inhibition of BTK
Inhibition of collagen-induced activation ultrastructural changes Inhibition of collagen-induced platelet aggregation
Prolongation the tail bleeding times of mice
Improvement of survival in mouse model of agonist-induced invariably fatal pulmonary thromboembolism
|
[143]
|
|
R406
|
Abrogation of shape change and aggregation induced by activation of GPVI and CLEC-2
Reduction in platelet spreading on fibrinogen
Inhibition of tyrosine phosphorylation of signaling proteins, including PLCγ2
Inhibition tyrosine phosphorylation of CLEC-2 and Syk downstream of CLEC-2 activation, whereas phosphorylation of Syk downstream of GPVI and integrin alphaIIbbeta3 was unaffected
|
[144]
|
|
PRT-060318
|
Inhibition of heparin-induced mouse thrombocytopenia
Reduction in thrombosis in mice
|
[145]
|
|
Other inhibitors
|
GW3965
|
Reduction of the size and the stability of thrombi in vivo
Reduction in CRP-induced platelet aggregation, degranulation and activation of integrin αIIbβ3
Association of LXR with signaling components proximal to GPVI
|
[146]
|
|
Panobinostat and romidepsin
|
Impairment of mouse platelets ability to respond to CRP
Increase in acetylation of megakaryocytic GPVI resulting in degradation to a ∼10 kDa GPVI
Reduction in the surface and total expression of GPVI
Inhibition of Syk phosphorylation and activation
Reduction in CRP-induced activation of integrin αIIbβ3
|
[147]
|
|
2-APT
|
Abolition of collagen-dependent whole blood thrombus formation
Impairment of washed platelet aggregation in response to collagen
Reduction in CRP-induced activation of integrin αIIbβ3 and PKC
Inhibition of Syk downstream collagen-activation
Inhibition of the superoxide ion generation stimulated by platelet adhesion on collagen and fibrinogen
|
[148]
|
|
EXP3179 (Losartan metabolite)/ Losartan
|
Inhibition of collagen-I and GPVI-dependent platelet aggregation and activation
Inhibition of collagen-I and GPVI-dependent PAC-1 expression
Reduction of human atherosclerotic plaque material-induced platelet aggregation in vitro and of murine platelet adhesion after acute vessel injury in vivo (just EXP3179)
Decrease of adhesion of GPVI-receptor expressing Chinese hamster ovarian cells on collagen-I under shear conditions
|
[149]
|
|
γ-aminobutyric acid
|
Binding to GPVI
Inhibition of convulxin- and collagen-induced human platelet activation
Decrease of Ca2+ mobilization, PLC-γ2, PKC, Akt phosphorylation, and hydroxyl radical formation
The occlusion time of platelet plug formation was significantly prolonged by GABA in mice
|
[150]
|
Abbreviations: 2-APT, 2-acetylphenothiazine; ADP, adenosine diphosphate; CRP, collagen-related peptide; GABA, γ-aminobutyric acid; GPIV, ; Ig, immunoglobulin; LAT, linker for activation of T cell; PKC, protein kinase C.
One of the first described snake venom proteins that interacts with GPVI is CVX, isolated from the venom of Crotalus durissus. CVX is an agonist and activates platelets.[151] Although numerous snake venom compounds have been described to activate platelets, recently, a peptide derived from a snake venom compound was shown to interact with GPVI, however, in an antagonistic way. The partial sequence of trowaglerix α subunit, a potent specific GPVI-targeting snake venom C-type lectin protein, isolated from the venom of Tropidolaemus wagleri, specifically inhibited collagen-induced human platelet aggregation.[124] It is worth noting that there is a binding epitope on GPVI for polypeptides, such as hexa/decapeptides, which shows great potential for the development of lead compounds, aimed as small-mass GPVI antagonists for thrombogenic diseases.[124] A mechanistically different snake venom metalloproteinase purified from the venom of Calloselasma rhodostoma, kistomin, inhibited collagen- and CVX-induced human platelet aggregation by cleavage of GPVI.[125]
A large number of natural bioactive compounds (NBCs), such as polyphenols, terpenoids, alkaloids, and fatty acids, were reported to have an apparent inhibitory activity on human platelets.[152] However, their mode of action has remained elusive. Some of these compounds have been associated with GPVI and its signaling pathway. Buddleja globose, also known as matico, is a medicinal plant rich in polyphenols and flavonoids which present high antioxidant and anti-inflammatory activity.[126] Matico inhibits human platelet aggregation response to collagen, CVX, and ADP in a concentration-dependent manner, as well as the expression of P-selectin, a maker of platelet activation. This antiplatelet activity most likely is caused by its high content in polyphenols.[126] A potential mechanism of action for these compounds is through the inhibition of phosphorylation of PLC-γ2 and PKC-β2, which are signaling proteins downstream of GPVI.[126] These findings have to be confirmed in clinical studies, to establish matico's suggested cardioprotective effects. Glaucocalyxin A, an ent-diterpenoid isolated and purified from the aerial parts of Rabdosia japonica var. glaucocalyx (Maxim.), disrupts CRP- and CVX-triggered human platelet activation in addition to thrombus formation in mice, also through the inhibition of collagen-stimulated tyrosine phosphorylation of Syk, LAT, and PLCγ2.[127] Fractions of aqueous and methanolic extracts from Solanum lycopersicum, the well-known tomato, also possesses antiaggregating activity as they inhibit collagen-induced rat platelet aggregation, potentially via GPVI receptor.[128] Furthermore, another publication describes lycopene, a natural carotenoid antioxidant presents in tomatoes, as an active compound with in vitro and in vivo antiplatelet activity, concentration-dependently inhibiting the breakdown of phosphoinositide, intracellular Ca+2 mobilization, TXB2 formation upon stimulation by collagen, and PKC phosphorylation.[129]
Interestingly, the inhibitory effect of NBCs on platelets goes beyond the direct interaction with GPVI. As reported by Yoshida et al, anopheline antiplatelet protein (AAPP), a saliva protein from malaria vector mosquito, inhibits human platelet adhesion by blocking the binding sites of GPVI on collagen. Interestingly, AAPP has no direct effect on GPVI-induced platelet activation, and therefore, it could be used as a tool to indirectly inhibit the GPVI-induced platelet functions with the benefits of potential minor side effects.[130]
Another potential NBCs mechanism of action involves peroxisome proliferator-activated receptor (PPAR) and the inhibition of protein kinase A and G, cyclooxygenase-1, TXA2, and cytosolic Ca2+. An interesting study by Akbiyik et al associated PPAR-γ of platelets with GPVI signaling.[153] By treating collagen-stimulated platelets with PPAR-γ ligands, such as 15d-PGJ2 and rosiglitazone, the authors demonstrated that these ligands modulate the activity of the GPVI signaling pathway, resulting in a decrease of human platelet activation, aggregation, and thrombus formation under arterial flow conditions.[154] Cyanidin 3-O-β-glucopyranoside, an anthocyanin from the flavonoid family, attenuated thrombus growth without prolonging bleeding time in mice, downregulated collagen-induced GPVI signaling by significantly decreasing phosphorylation of Syk, LAT, and SLP-76, as well as the expression of Lyn, Fyn, and PLC-γ2.[131]
To date, the main clinical strategies developed focus on interfering with the GPVI–collagen interaction by two main approaches: using collagen-binding molecules or through GPVI-function blocking reagents.
A GPVI mimetic has been developed, a recombinant fusion protein that consists of the extracellular collagen-binding domain of GPVI fused with the C-terminal of human Ig Fc domain, forming a soluble GPVI dimeric called GPVI-Fc or Revacept.[132] This Fc fusion protein specifically binds to immobilized collagen, with high affinity, and, as a decoy receptor, it competitively inhibits binding of collagen to platelet-anchored human GPVI.[132] Furthermore, in response to Revacept, thrombus formation is attenuated in vitro when platelets are under shear conditions. In an in vivo setting, the platelet-vessel wall interactions are compromised by abolishing stable arrest and platelet aggregation following vascular injury.[132]
[133] Most importantly, when delivered locally at the site of vascular injury, the amount of GPVI-Fc is sufficient to reduce thrombus formation without systemic antiplatelet effects, as shown in a rabbit model.[133] In a clinical phase I study, Revacept did not significantly affect bleeding time, while collagen-induced platelet aggregation was dose-dependently inhibited.[134] Some authors have assessed the combined administration of Revacept with first line treatments, as an alternative thrombolytic therapy. The main goal of the work from Reimann et al was to establish a thrombolytic therapy regimen using a mouse model of acute ischemic stroke, with concomitant administration of the first line treatment, recombinant tissue plasminogen activator (rtPA) and Revacept.[155] By combining lower doses of rtPA with Revacept, treatment efficacy improved without increased risk of intracerebral hemorrhage, while higher doses of rtPA sustained a significant risk of bleeding risk.[155] The overall results of Revacept characterize this compound as a safe and reliable alternative for antiplatelet therapy. A phase II study has taken place to assess the detailed clinical settings and pathologies where this therapy could carry advantageous outcomes (NCT01645306).
Among the approaches of directly targeting GPVI, inhibiting antibodies have been developed. The first monoclonal antibody (mAb) against GPVI was described by Nieswandt et al, as JAQ1. It inhibited mouse platelet aggregation induced by collagen, however, when cross-linked, this antibody had the opposite effect.[156] This is a disadvantage of the monoclonal antibodies, as the Fc domain can cross-link and cluster the GPVI.[12] Nevertheless, upon JAQ1 treatment, GPVI was depleted from circulating platelets in mice and the thrombotic responses to collagen and CRPs were abolished, resulting in protection from thromboembolism with moderate tail bleeding times.[19]
[33] Other mAbs, the human GPVI-specific mouse mF1201 and mF1232, were also assessed in in vitro studies with different effects. The antibody mF1201 induced human platelet activation and GPVI shedding, while mF1232 inhibited collagen-induced human platelet aggregation.[135] This was reflected in studies in vivo, where monkeys treated with mF1232 suffered from long-term antiplatelet effects due to GPVI shedding-dependent immunodepletion without significant thrombocytopenia.[135]
To overcome the potential undesired platelet activation triggered by mAbs, inhibitory Fab fragments were developed. GPVI-blocking antibody fragments, such as antigen-binding Fab or F(ab′)2 fragments, are generated from monoclonal anti-GPVI antibodies by limited proteolysis with soluble papain or immobilized pepsin, respectively.[45] Interestingly, the development of these fragments was based on the pathological case of a patient suffering from purpura and prolonged bleeding time despite an adequate number of platelets.[45] The F(ab')2 fragments of antibodies isolated from the patient's blood induced TXB2 synthesis, increased cytoplasmic Ca2+ levels in platelets and specifically inhibited aggregation induced by collagen.[21] In line with these findings, antagonistic Fab fragments generated from the mAb 9O12.2 have been developed. This mAb was selected because, although 9O12.2 IgGs induced platelet activation, it also disrupted the interaction of GPVI with collagen.[45] 9O12.2 Fab fragments inhibit collagen- and CRP-induced platelet aggregation and prevent thrombus formation under flow conditions.[45] In addition, a successful screening has identified Fab fragments that specifically bind to GPVI dimer, but do not bind to GPVI monomer.[140] The promising results spurred the design and humanization of the single-chain variable fragment (scFv) from the 9O12 mAb by recombinant protein engineering. By preserving the functional properties, these Fab fragments can potentially be exploited for therapeutic applications.[136] For example, ACT017 is a high affinity humanized Fab that inhibits collagen-induced human platelet aggregation, and showed no thrombocytopenia, GPVI depletion, or bleeding side effects in macaques[141] and is currently being tested in phase II studies (NCT03803007). The inhibition of platelet function by fragments (Fab, scFvs, and dAbs) generated from other antibodies was also later reported and their effects are summarized in the [Table 3].[18]
[20]
[58]
[137]
[138]
[139]
[142]
Not included in any of the named categories, additional inhibitors to GPVI have been described. Compounds targeting tyrosine kinases and adaptor proteins downstream of GPVI also block the signaling pathway and function of GPVI. One of the first studies describing an inhibitor of the GPVI pathway was a leflunomide metabolite analog, LFM-A13, and it was designed to specifically bind to one of the TEC family protein tyrosine kinase, Btk.[143] Later, two other Syk inhibitory molecules, R406[144] and PRT-060318,[145] emerged as tools to study the GPVI and CLEC-2 signaling pathways. All compounds inhibit platelet activation. R406 decreases the tyrosine phosphorylation of signaling proteins, such as PLC-γ2, in human platelets and PRT-060318 reduces heparin-induced thrombocytopenia and thrombosis in mice.[143]
[144]
[145] Other compounds also interfere with the GPVI signaling pathway but not by targeting signalosome molecules. An example is GW3965, a transcription factors ligand involved in the regulation of cholesterol homeostasis and liver X receptors, which is expressed in platelets.[146] Spyridon et al, attributed the atheroprotective properties of GW3965 to the antiplatelet and antithrombotic effects, since it nongenomically modulated human platelet aggregation and reduced the size and stability of thrombi in an in vivo mouse model.[146] These effects are suggested to be mechanistically related to the GPVI signaling pathway.[146] Inhibitors of the enzyme histone deacetylase, panobinostat, and romidepsin, although structurally different, both impaired mouse platelets ability to respond to CRP, increased acetylation of megakaryocytic GPVI resulting in degradation to approximately 10 kDa GPVI and inhibited Syk phosphorylation and activation.[147] Interestingly, NADPH oxidases (NOXs) are involved in platelet activation (mechanism remains largely unknown) and was elucidated by the NOX inhibitor 2-acetylphenothiazine (2-APT).[148] 2-APT also abolished the collagen-dependent formation of whole blood thrombus and aggregation of washed human platelets. Moreover, the CRP-triggered activation of integrin αIIbβ3, PKC, and collagen-activation of Syk was inhibited by 2-APT.[148] The signaling molecules of the GPVI pathway are an appealing target for therapy, but the lack of specificity poses a significant challenge, as Syk and other components of the GPVI signalosome are also important in immune cell signal transduction.[157] Another potential concern of this targeting is the effect of disrupting GPVI receptor signaling via the ITAM signaling receptors on platelets, which could also negatively affect the outside-in signaling of the αIIbβ3 and β1 integrins. Nevertheless, early studies using bone marrow chimeric Syk-deficient mice or pharmacological inhibition of Syk in vivo reported virtually normal hemostasis. Therefore, targeting GPVI-mediated production of reactive oxygen species or the transcription factors may present more subtle effects by diminishing but not blocking GPVI-signaling function.[12]
[157]
More blocking inhibitors of GPVI have also been found in an intuitively less expected manner. EXP3179 is an active metabolite that arises from the angiotensin II type 1 (AT1)-receptor antagonist losartan (LOS) after hepatic metabolization by the cytochrome-P450 pathway. Both EXP3179 and LOS, inhibit collagen I- and GPVI-dependent human platelet aggregation and activation.[149] EXP3179 reduces platelet aggregation induced by human atherosclerotic plaque material in vitro, as well as murine platelet adhesion after acute vessel injury in vivo.[149] The endogenously produced molecule, γ-aminobutyric acid (GABA), also showed a direct inhibitory effect on GPVI. GABA is the major inhibitory neurotransmitter in the central nervous system, and specifically inhibits collagen-induced platelet activation, by binding to GPVI, resulting in decreased Ca2+ mobilization, PLC-γ2, PKC, Akt phosphorylation, and hydroxyl radical formation.[150]
Recently, our group has initiated drug discovery attempts to identify new small molecules capable of inhibiting the activity of GPVI. For that, a 384-well plate assay to test platelet function in a high-throughput screening format was developed and, using compounds selected by structure-based virtual screening, we identified several hits that markedly decreased the human platelet activation induced by CRP.[158] Currently, additional experiments are being performed to test the specificity of these hit compounds and their effect on thrombus formation.
Conclusion
This review summarizes several reports demonstrating the importance of GPVI receptor on hemostasis, platelet function, and thrombus formation by means of human GPVI-deficiency studies. In addition, the recent discoveries related to the biochemical structure, intracellular signaling, and potential inhibitors of GPVI receptor were included in this review. Due to its characteristic features, GPVI is crucial for thrombus formation and is a potential drug target for thrombosis. The targets of the different inhibitors on platelet GPVI, as well as downstream signaling pathways are illustrated in [Fig. 2].
Fig. 2 Inhibitors of glycoprotein (GP) VI activity. Snake venom derived compounds: Trowaglerix venom polypeptides (Troα10); Kistomin (Kis); Natural bioactive compounds: Buddleja globose (Bjg); Rabdosia japonica - Glaucocalyxin A (GLA); Solanum lycopersicum (lycopene) – LPE; Anopheline antiplatelet protein (AAPP); Anthocyanin (cyanidin-3-glucoside) – Cy-3-g; GPVI mimetic: GPVI-Fc (Revacept); Inhibitory monoclonal antibodies: JAQ1; mF1232; Inhibitory antibody Fab fragments: 9O12.2; OM2; OM4; SAR264565; m-Fab; ACT017; Inhibitory antibody scFvs fragments: 10B12; 1C3; Inhibitory antibody dAbs fragments: BLO8–1; GPVI signaling pathway inhibitors: LFM-A13; R406; PRT-060318; Other inhibitors: GW3965; Panobinostat - Pan; romidepsin - Romi; 2-APT; EXP3179 (Losartan metabolite); γ-aminobutyric acid – GABA. * GPVI-induced degranulation inhibition; inhibition of GPVI-induced αIIbβ3 activation.
Differently from human deficiency of α2β1 integrin, mutations of GPVI receptor in patients lead, at most, to mild bleeding. In addition, the role of GPVI in platelet function seems to induce a rapid activation of αIIbβ3 integrin and platelet degranulation, crucial steps for aggregation formation. Therefore, GPVI is an appealing therapeutic target, and parallel efforts have led to the development of several compounds that are able to inhibit GPVI activity. The insights that these studies provided and the early stage of development of an ideal antiplatelet agent to target GPVI, have prompted further research aimed at developing a new GPVI-inhibiting drug. In addition, the “bleeding-free” feature of GPVI inhibitors ensures a clinical advantage compared with all antiplatelet drugs available on the market. Taken together, the purpose of this review was to provide guidance for future investigation in the platelet field, specifically for GPVI-related research, and potentially help with the development of a novel drug class for antithrombotic therapy.
