Platelet Hyperreactivity in Diabetes: Focus on GPVI Signaling—Are Useful Drugs Already Available?

Platelet hyperreactivity is one of the most important factors in the initiation of thrombotic disease leading to heart attack and ischemic stroke and is a life-threatening complication of diabetes. Two-thirds of all cardiovascular deaths occur in patients with diabetes (1), who often have underlying atherosclerotic vascular disease associated with increased arterial shear stress. The reduction of thrombotic events in patients with diabetes, however, has been disappointingly minor in several trials of antiplatelet therapy (2). There is therefore an urgent need for alternative or improved antiplatelet therapy for patients with diabetes. Foremost, delineation of signaling pathways is critical to understanding the differential regulation of platelet activity so that pathological thrombosis is prevented without interfering with the physiological role of platelets in preventing bleeding. This review will focus on the underlying mechanisms responsible for platelet hyperreactivity in diabetes and why current antiplatelet agents have reduced efficacy in patients with diabetes. We will discuss three critical elements that are intimately involved in heightened platelet activity and the common link between them that can potentially be targeted by existing drugs. Interestingly, these three factors—high blood glucose, elevated shear stress, and aberrant redox leading to increased oxidative stress—are not affected by current antiplatelet drugs, but deleterious corollaries of each can be inhibited by existing drugs that impact the signaling pathway downstream of the platelet-specific receptor glycoprotein VI (GPVI). Significantly, converting existing drug treatments to new uses can accelerate translational research and provide optimism for patients with diabetes with hyperreactive platelets that are unsatisfactorily managed by standard antiplatelet drugs.

The capacity of platelets to rapidly adhere to exposed vascular matrix at arterial and pathological shear rates is dependent on two adhesion receptor complexes unique to platelets, the GPVI/FcRγ-chain complex and the GPIb-IX-V complex (3). Binding of ligand (collagen or von Willebrand factor [VWF], respectively) to either receptor initiates a cascade of signaling events leading to platelet activation and spreading, release of soluble agonists including thromboxane A₂ (TxA₂) and ADP, and activation of the integrin α_IIbβ₃ which, in binding VWF and fibrinogen, forms a bridge between platelets and results in platelet aggregation (Fig. 1). It is the signaling pathways involving soluble agonists that are targeted by antiplatelet drugs.

There are currently four different classes of antiplatelet drugs used for preventing and treating ischemic events associated with cardiovascular disease: cyclooxygenase 1 (COX-1) inhibitor (aspirin); inhibitors of the ADP receptor P2Y₁₂ (clopidogrel, prasugrel, ticagrelor); phosphodiesterase (PDE) inhibitor (dipyridamole); and integrin α_IIbβ₃ inhibitors (abciximab, tirofiban, eptifibatide) (see Fig. 1) (4). The U.S. Food and Drug Administration has recently (May 2014) approved for use another antiplatelet drug, an inhibitor of the thrombin receptor PAR1 (vorapaxar), although it is not approved for patients with a history of ischemic stroke or transient ischemic attack because of serious bleeding risks (5).

Aspirin irreversibly prevents the generation of TxA₂ from arachidonic acid by inhibiting the enzyme responsible, COX-1. P2Y₁₂ inhibitors prevent the binding of ADP, released from platelet granules, to its receptor, thereby inhibiting the subsequent signaling that would lead to integrin α_IIbβ₃ activation. Dipyridamole inhibits the uptake of adenosine into platelets, endothelial cells, and red blood cells and elevates levels of cAMP and cGMP through inhibition of the enzyme PDE, preventing platelet aggregation. It is usually a second-line treatment in combination with aspirin following ischemic stroke or transient ischemic attack only if P2Y₁₂ inhibition is not an option. Because integrin α_IIbβ₃ inhibitors can only be administered intravenously, their use is limited to acute, in-hospital procedures such as percutaneous coronary intervention. The standard antiplatelet agents in current use in the outpatient setting are aspirin and/or P2Y₁₂ inhibitors.

Many studies and trials, including those specifically conducted in patients with diabetes, have determined that aspirin for the primary prevention of cardiovascular events (i.e., delaying or preventing the occurrence of heart attack or strokes) in patients with diabetes cannot be recommended because of the lack of benefit and risk of major bleeding. For secondary prevention of cardiovascular events (i.e., prevention of the recurrence of events or the complications thereof), aspirin has shown to be less beneficial in patients with diabetes than in those without (6).

The pattern is similar for the P2Y₁₂ inhibitor clopidogrel, and diabetes is associated with a higher prevalence of reduced responsiveness to dual antiplatelet therapy of aspirin and P2Y₁₂ inhibition (7), which has even been suggested to be harmful (8). Overall, patients with diabetes exhibit higher pretreatment platelet activity but a poorer response to standard antiplatelet therapy compared with patients without diabetes. Why might this be so? And what can be done about it?

High blood glucose, oxidative stress, and elevated vascular shear forces are prevalent in patients with diabetes. All predispose platelets to hyperreactivity that targeting of soluble agonists cannot prevent. Further, the coexistence of all three factors in diabetes can exacerbate the potential pathological forces exerted by each factor individually. For example, even short-term hyperglycemia is known to enhance platelet activation induced by high shear in type 2 diabetes (9). Hyperglycemia is also known to induce oxidative stress and reduce antioxidant defenses in normal individuals as well as patients with diabetes (10).

Patients with type 2 diabetes have an enhanced platelet surface expression of the collagen receptor GPVI compared with individuals without diabetes, and high levels of GPVI are associated with cardiovascular events such as acute coronary syndrome, myocardial infarction, ischemic stroke, and transient ischemic attacks (11). Increased GPVI expression is also associated with poor clinical outcome for cardiovascular events (12). Plasma levels of the soluble form of GPVI (which is cleaved from the platelet surface primarily by the metalloproteinase ADAM10) (13) are elevated in patients following acute ischemic stroke (14), indicating increased activation of GPVI in this circumstance. In an experimental model of stroke and reperfusion injury, depletion of GPVI by anti-GPVI antibody treatment of mice resulted in smaller infarct volumes compared with vehicle-treated controls, without an accompanying increase in bleeding complications (15). Inhibition of GPVI signaling could serve as a basis for antiplatelet agents that can inhibit pathological thrombosis rather than primary hemostasis, particularly in the setting of diabetes. Below, we describe how GPVI signaling is influenced by oxidative stress, hyperglycemia, and shear stress and how targeting GPVI signaling pathways can selectively disrupt platelet function under these pathological conditions.

Oxidative stress, resulting from an imbalance between reactive oxygen species (ROS) production and antioxidant defense, is increased in patients with diabetes. There are multiple sources of ROS: mitochondrial and nonmitochondrial, intracellular and extracellular. Platelets are able to generate intracellular ROS that are involved in platelet signaling pathways (16). Platelets from patients with a very rare (one case in a million) genetic deficiency in NADPH oxidase (Nox) subunits (chronic granulomatous disease) produce almost no ROS, indicating that Nox is the major source of platelet ROS. The adaptor protein, tumor necrosis factor receptor-associated factor 4 (TRAF4), is a binding partner of GPVI and GPIb-IX-V (17). TRAF4 selectively binds cytoplasmic sequences of GPIbβ (of the GPIb-IX-V complex) and GPVI, as well as p47^phox (of the Nox enzyme complex). TRAF4 also binds other redox-relevant signaling proteins such as Hic-5, which is constitutively associated with the Src family tyrosine kinase Lyn, an important factor for the propagation of signals downstream of GPVI and GPIb via spleen tyrosine kinase, Syk (Fig. 1). In this way, TRAF4 is the link between engagement of platelet adhesion receptors responsible for adhesion of platelets to the injured vessel wall at high shear rates and the generation of intracellular ROS generation. Of the different Nox isoforms, Nox2 and its regulatory subunits, including p47^phox, have been identified in platelets. Additionally, Nox1 was recently discovered to be expressed in platelets (18) and, at least in transfected cells, Nox1 is able to utilize p47^phox and p67^phox, indicating a potential role for Nox1 and/or Nox2 in platelet activation. In vivo platelet activation has recently been shown to be significantly and directly correlated with Nox2 activity (19). Reduced efficacy of aspirin in patients with type 2 diabetes could be related to oxidative stress–mediated overproduction of eicosanoids due to increased Nox2 activation in these patients relative to control subjects without diabetes (20). In an experimental model of type 1 diabetes, there was a selective exacerbation of GPVI-dependent ROS generation in platelets from diabetic monkeys with elevated glucose compared with healthy controls or diabetic monkeys with normalized blood glucose levels (21). This indicates that specific hyperreactivity to GPVI engagement is an early event in the progression of type 1 diabetes and is linked to high blood glucose levels. In this regard, while patients with diabetes are known to have elevated levels of platelet GPVI and ROS generation in combination with elevated HbA_1c, the precise causative and correlative relationships between these factors is not yet fully understood. Interestingly, oscillations in blood glucose levels, rather than chronically high levels, could be associated with aberrant production of ROS (22).

Hyperglycemia and heightened cardiovascular risk are linked. Many studies have therefore investigated the effect of glycemic control on cardiovascular risk factors. It was the purpose of the ACCORD (Action to Control Cardiovascular Risk in Diabetes) trial in 2007 to aggressively use drugs to lower three key risk factors, blood glucose, blood pressure, and triglycerides, in patients with type 2 diabetes with the aim of preventing cardiovascular disease (23). The glycemia arm of the trial was terminated early, however, because of higher mortality in the intensive compared with the standard blood glucose–lowering strategies. Antiplatelet strategies are also ineffective in the face of heightened blood glucose. Acute, short-term hyperglycemia during a 4-h hyperglycemic clamp enhanced the platelet activation induced by high shear stress in patients with type 2 diabetes (24), and this exaggerated platelet response was resistant to aspirin despite adequate inhibition of COX-1 in these individuals.

Treatment of healthy human platelets with elevated levels of glucose in vitro has been reported to specifically enhance GPVI-dependent signaling (25). In diabetic monkeys, inhibition of Syk activity (with BAY61–3606) reduced GPVI-dependent ROS generation regardless of the level of glycemic control in the animals, revealing that hyperreactivity associated with hyperglycemia in this diabetes model was due to increased signaling downstream of Syk rather than an upregulation of receptor proximal signaling (21). This pointed to Syk inhibition as a potentially effective antiplatelet strategy for patients with diabetes due to Syk inhibition having little impact on hemostasis in animal models of arterial thrombosis (mice, rabbits, and pigs) or patients (26). This finding highlights the importance of investigating the role of Syk in thrombus formation in humans.

Aldose reductase (AR) is the first enzyme of the polyol pathway, which converts excess glucose to sorbitol accompanied by an increase in the cytosolic NADH/NAD+ ratio (Fig. 1). Under normal glycemic conditions, AR is only a minor consumer of glucose; however, during hyperglycemia, AR activity is significantly enhanced and is thought to contribute to the vascular complications associated with diabetes by increasing oxidative and osmotic stress on cells. Two important recent studies have linked AR and GPVI signaling. First, proteomic analysis of differentially altered proteins revealed that AR activity and expression were upregulated following GPVI-dependent platelet activation (27). These changes were functionally relevant because inhibition of AR activity resulted in reduced GPVI-dependent platelet aggregation. Second, AR was shown to play a central role in GPVI-dependent signal transduction (increased PLCγ2/PKC/p38 MAPK activation leading to increased TxA₂ generation), and this signaling pathway was enhanced in hyperglycemic conditions (25).

Patients with diabetes often have underlying atherosclerotic vascular disease associated with altered blood flow and increased shear stress. Defining the effect of shear on platelet function in a developing thrombus is significant because the adhesion of platelets to damaged arterial walls results in major cardio/cerebrovascular incidents such as heart attack and stroke, primarily due to altered vascular rheology and shear in diseased vessels. It has recently been identified that shear gradients experienced by platelets trigger sustained aggregation of discoid platelets independent of soluble agonists (ADP, TxA₂) (28). This is critical because the major antiplatelet agents, aspirin and clopidogrel, prevent release of soluble agonists (TxA₂ and ADP, respectively) but sustained aggregation of platelets at high shear such as that experienced in diseased blood vessels is independent of these agonists, meaning that their administration has very limited efficacy under these conditions, particularly in individuals such as patients with diabetes with severe vascular disease. Currently very little is understood about how shear gradients control thrombosis. Ultimately, targeting platelets selectively within diseased regions of high shear variation (e.g., within a stenotic vessel) but not systemically, thus avoiding many of the dangerous side effects of current treatments, could improve safety.

Under conditions of high shear stress found in arterioles and atherosclerotic arteries, GPIb-IX-V and GPVI cooperate to arrest platelets at sites of vascular damage. Interaction of GPIb with VWF on exposed collagen initially tethers the platelets, and collagen interaction with GPVI results in platelet activation and growth of the thrombus. Experimental and clinical inhibition of GPVI signaling have indicated that targeting this locus is likely to be selective for pathological thrombus formation rather than normal hemostasis because at low shear rates the fibrinogen receptor, α_IIbβ₃, is the primary receptor supporting platelet adhesion and plug formation.

Alterations in mechanical shear stress are sufficient to activate Nox and generate ROS (29). In particular, areas of disturbed flow typically occurring at sites of atherosclerotic lesions result in upregulation of Nox complex subunits, increased superoxide production, and reduced nitric oxide generation. Taken together, the combination of high blood glucose, elevated oxidative stress, and severe vascular disease (resulting in pathologically high shear stress) prevalent in patients with diabetes means that this patient group has specific health care needs when it comes to effective antithrombotic therapies.

Expression of GPVI is limited only to platelets and their precursor cells in the bone marrow, the megakaryocytes, making GPVI a promising target for antiplatelet agents because of the lack of off-target effects in other cell types. Experimentally, fragment antigen binding (Fab) fragments of anti-GPVI monoclonal antibodies have been proven to be potent inhibitors of platelet activation in vitro and ex vivo and in animal models of thrombosis without prolonging bleeding time (30). Inhibition of GPVI function in humans, mainly due to autoantibodies, has been associated with a mild prolongation of bleeding time (31).

Given that fibrous collagen is the ligand for GPVI and that collagen is exposed at sites of vascular injury or atherosclerotic plaque rupture, an alternative strategy to prevent thrombotic events in at-risk individuals such as patients with diabetes is to prevent platelet binding to vascular collagen, thereby avoiding unwanted effects on circulating platelets. Even in the absence of plaque rupture, GPVI plays a role in platelet adhesion to atherosclerotic endothelium (32). A soluble GPVI-Fc dimeric fusion protein (Revacept) works by competitive inhibition of the GPVI binding sites on exposed collagen (33). A phase I study determined the safety and tolerance of Revacept in healthy males and established that there was no interference with normal hemostasis (34). Testing in animal models revealed improved endothelial function and reduced vessel-wall thickening in cholesterol-fed, atherosclerotic rabbits (35) and inhibition of thrombus formation, reduced infarct area, and improvement in functional outcomes following ischemic stroke (36). Other researchers have reported, however, that stronger protection against thrombosis is achieved with direct GPVI inhibition rather than with the GPVI-Fc dimer (37). Despite this, Revacept is now in phase II trials in patients with coronary artery disease, symptomatic coronary stenosis, transient ischemic attacks, and ischemic stroke. Thus far, no relevant adverse effects have been observed.

The drawback of using anti-GPVI antibody– or fusion protein–based therapies is that the administration of these antithrombotic agents requires intravenous injection so they are not appropriate for use outside the clinic. Orally bioavailable inhibitors would have much greater utility. Other antiplatelet agents associated with inhibition of GPVI signaling have been proposed (38,39) and are the subject of ongoing research, but an alternative strategy involves the repurposing of existing drugs for a new use.

On average, discovery of a new drug takes 13 years, costs $3 billion, and may have a failure rate of approximately 95% (40). It is therefore of great benefit to be able to repurpose current drugs for the successful treatment of a different disease. Aspirin is one such example of a drug successfully repurposed as an antiplatelet agent. Originally used in pain relief, aspirin is now additionally used alone or in combination with other antiplatelet treatments as a standard therapy for the primary and secondary prevention of heart attack and stroke. Losartan (an angiotensin II receptor 1 inhibitor) used in the treatment of hypertension was discovered using structure-based repurposing to have anti-GPVI effects (41). In a recent study investigating the effects of a therapeutic dose of losartan on collagen-induced platelet activation, however, no statistically significant differences were observed between treated and nontreated patients with respect to a biological antiplatelet effect (42). While not representative of successful repurposing in this case, the study does highlight a novel mechanism for inhibiting GPVI function, as well as at least illustrating the potential of repurposing as a high-gain strategy for yielding new antiplatelet agents.

Focusing on drugs that could have particular benefit in the diabetes population (inhibitors of AR, Nox, and/or signaling molecules in the GPVI pathway) could potentially circumvent the platelet hyperreactivity in patients with diabetes that results from the combination of high glucose, excess ROS, and elevated shear stress.

Epalrestat is an AR inhibitor approved for use in Japan to prevent progression of diabetic neuropathy and diabetic retinopathy/nephropathy (43). AR inhibition by epalrestat not only directly impacts the GPVI signaling pathway to reduce platelet reactivity but also alleviates the oxidative stress associated with the development of diabetes (44). Further, inhibition of AR inhibits protein kinase C signaling pathways and upregulates nitric oxide production by endothelial cells (45), inhibiting platelet activation and providing an important antithrombotic effect.

GKT137831 is a first-in-class, selective Nox1/4 inhibitor in phase II clinical development for the treatment of diabetic nephropathy (46). Trials with this inhibitor have shown that it is very well tolerated and reduces inflammatory marker levels in patients with type 2 diabetes (47). The discovery of Nox1 expression in platelets warrants the investigation of this compound as a novel antithrombotic in the setting of diabetes, especially given that chronic hyperglycemia is known to upregulate Nox1 and Nox2 expression and amplifies Nox-derived ROS generation (20,48). There are currently no Nox2-specific inhibitors under clinical evaluation. A study by Gray et al. (46) showed that genetic deletion of Nox2 in the context of diabetes was associated with increased rate of infections and mortality, suggesting that complete inhibition of Nox2 in diabetes is not desirable.

Targeted inhibition of Syk in the treatment of hematological malignancies has seen the instigation of phase I and II trials of the orally available inhibitor fostamatinib disodium (49), which is also under clinical investigation for the treatment of rheumatoid arthritis, immune thrombocytopenic purpura, and allergies. While phase III trials of fostamatinib in the treatment of rheumatoid arthritis have been disappointing, the drug has met its primary end point in the first of two double-blind phase III trials for the treatment of chronic immune thrombocytopenic purpura. Because of the role of Syk in GPVI-dependent signaling and its link with oxidative stress, as well as the observation that signaling pathways downstream of Syk are heightened under hyperglycemic conditions, Syk inhibitors represent an additional opportunity for repurposing as antiplatelet agents in diabetes. Recently, a novel, orally available Syk inhibitor, BI1002494, with greater selectivity and potency than fostamatinib, has shown great promise for the prevention of arterial thrombosis and limitation of brain infarct whether administered before or after acute stroke in a mouse model (50).

Other drugs that could be potentially investigated for their utility in diabetes are statins, which are the cornerstone therapy for lowering LDL cholesterol. In a prelude to results from ongoing, larger, randomized clinical trials, the RATIONAL (aspiRin stAtins or boTh for the reductIon of thrOmbin geNeration in diAbetic peopLe) trial in Argentina investigated whether aspirin, statins, or both could be of benefit in a smaller cohort (30 patients) of people with diabetes. The trial demonstrated that in patients with type 2 diabetes without previous cardiovascular events, administration of statins but not aspirin was beneficial in reducing thrombotic risk, as assessed by a reduction in thrombin generation (51). Statins are already proving to have potential benefits beyond their primary lipid-lowering ability, and, mechanistically, platelets are involved in many of the effects (52). Statins are known to alter membrane fluidity and enhance ADAM10-dependent shedding of amyloid precursor protein (53). ADAM10 is also the enzyme involved in shedding of GPVI from the platelet surface (13). Whether this is involved in the reduced thrombotic risk associated with statin use in patients with diabetes is yet to be determined.

Finally and importantly, all of the inhibitors mentioned have been assessed in at least phase I clinical trials, and no serious adverse effects have been reported—in particular, no bleeding complications. It remains to be seen whether any of these drugs—or rationally designed derivatives—can be successfully repurposed as antiplatelet therapies, either alone or in combination.

Patients with diabetes are a group at particular risk of cardiovascular events because, unlike individuals without diabetes, high platelet reactivity persists despite the use of recommended antiplatelet regimens. A total of 65% of all CVD deaths occur in people with diabetes or prediabetes. There is therefore a massive unmet need for alternative therapies in patients with diabetes. Three independent but often interrelated variables that contribute to platelet hyperreactivity—high blood glucose, oxidative stress, and elevated vascular shear forces—coexist in patients with diabetes, creating a critical confluence for cardiovascular risk, and the platelet-specific receptor GPVI is involved in the signaling pathways of all three. Clearly, converting existing drug treatments to new uses (repurposing) can accelerate the pace of translational research.

In this brief review, we have discussed how current pharmacological agents that target the redox-relevant proteins (Syk, Nox, AR) in the GPVI signaling pathway important for platelet function specifically under high shear conditions could reduce bleeding risk and prevent thrombosis. In the future, there are at least three critical areas for maximizing the advantages of successful drug-repurposing approaches. First, it is clear that key relationships between diabetic hyperglycemia and abnormal GPVI expression and function resulting in platelet hyperreactivity require greater understanding from advanced experimental approaches and pilot clinical studies. Second, given some of the limiting side effects or lack of efficacy associated with initial repurposing studies to date, the identification and refinement of selective and pharmacologically suitable inhibitors of GPVI signaling among the increasing number of inhibitors available in the cancer and other fields need to be further analyzed for antiplatelet activity experimentally and clinically. Finally, it is evident that far greater efficiencies of cost, time, and scale are achievable if improved analytical methods for assessing platelet GPVI expression/function can be applied in clinical trials of relevant inhibitors to provide vital information on potential antiplatelet effects. Together, such initiatives can provide renewed hope for patients with diabetes with hyperreactive platelets that are refractory to standard antiplatelet drugs.

Acknowledgments. The authors thank their colleagues for valuable discussions. The authors take full responsibility for the contents of this review.

Duality of Interest. K.J.-D. received a small research grant from Genkyotex, manufacturer of GKT137831. No other potential conflicts of interest relevant to this article were reported.