The platelet “sugar high” in diabetes

In this issue of Blood, Zhu and colleagues clearly demonstrate that advanced glycation end products (AGE), generated under hyperglycemic conditions, can specifically interact with CD36 on platelets.¹  This interaction can trigger CD36-dependent JNK2 activation, enhance platelet aggregation, and accelerate thrombus formation. Thus, AGE-CD36–mediated platelet hyperreactivity may play an important role in the increased risk of arterial thrombosis in diabetic patients.

It has long been observed that diabetes mellitus is associated with arterial thrombosis. Emerging data suggest that accumulation of AGE in blood may relate to platelet hyperreactivity,²  a risk factor for thrombosis, but no evidence exists to directly link AGE to platelet function and no mechanism has been elucidated. It has been reported that AGE can bind to the receptor of AGE (RAGE), as well as several scavenger receptors including SR-A, SR-B1, and CD36 on various cells.³  However, the specific interaction between AGE and CD36 on platelets has not been established and the biologic consequences of this interaction are unclear. Zhu et al first used biotinylated AGE-BSA to detect AGE-platelet interactions in a flow cytometric assay. They found that AGE specifically binds to wild-type (WT) platelets but not to platelets of CD36 deficient (CD36^−/−) mice. They demonstrated that this interaction can be inhibited by NO₂⁺LDL, a specific ligand for CD36, but not by control NO₂⁻LDL. Furthermore, using a recombinant CD36 fusion protein, which contains a large section of the CD36 N-terminal extracellular domain, the authors clearly demonstrated an interaction between AGE and CD36 in a cell-free ELISA system. Interestingly, although RAGE, SR-A, and SR-B1 were also detected on WT and CD36^−/− platelets, they showed that AGE-platelet binding is primarily mediated by CD36.

Platelets are highly sensitive responders to environmental change. At sites of vascular injury, rapid platelet adhesion and aggregation are key events to maintain normal hemostasis. However, the same processes may also lead to thrombotic disorders; the formation of platelet plugs at sites of atherosclerotic lesion rupture is the most common mechanism leading to myocardial or cerebral infarction. It has been documented that the platelet GPIb complex and several integrin family proteins initiate platelet tethering and subsequent adhesion.⁴  Platelet aggregation is then mediated by platelet β3 integrin and its ligand fibrinogen, although fibrinogen-independent platelet aggregation can also occur.^5,6  It is clear that platelet activation plays important roles in both platelet adhesion and platelet aggregation. Suboptimal platelet activation or a hyperreactive state induced by different platelet signaling pathways (eg, triggering CD36 with oxidized LDL),⁷  may facilitate platelet adhesion and aggregation, and accelerate thrombosis.

To test whether AGE-CD36 interaction may affect platelet activation, Zhu et al examined platelet aggregation in response to a low dose of adenosine diphosphate (ADP). They found that pretreatment of murine platelet-rich plasma (PRP) with AGE-BSA, but not native BSA, significantly increased platelet aggregation. Similar results were also observed in healthy human PRP, but not in PRP of CD36^−/− mice. Importantly, after injections of AGE-BSA into mice, thrombosis was accelerated in a carotid artery thrombosis model in WT mice, but this prothrombotic effect was significantly attenuated in CD36^−/− mice. However, it appears that the absence of CD36 cannot completely protect against the AGE-BSA–mediated enhancement of thrombosis, suggesting that other CD36-independent mechanisms may also be involved in this model of in vivo thrombosis.

To further explore the relevance of these findings in diabetes, Zhu and colleagues established both type 1 and type 2 diabetes models (ie, drug-induced pancreatic islet destruction and diet-induced insulin resistance, respectively) in WT or CD36^−/− mice. After induction of hyperglycemia, they found that thrombosis in carotid arteries was significantly accelerated in WT mice; vessel occlusion time was 30% to 40% shorter compared with nondiabetic control mice. Consistently, CD36 deficiency almost completely rescued the prothrombotic phenotype in both type 1 and type 2 diabetes models. They found that plasma AGE levels increased approximately 12-fold and 2-fold in type 1 and type 2 diabetic mice, respectively. Although no significant difference in AGE levels was observed between WT and CD36^−/− mice, more AGE accumulated in the thrombi of diabetic WT mice. Interestingly, the absence of CD36 did not affect AGE binding to the vessel wall, which is consistent with their observation that CD36 may be the predominant AGE receptor on platelets but not on endothelial or smooth muscle cells.

Because the authors previously found that CD36-mediated platelet activation by oxidized LDL requires activation of the MAP kinase JNK, they tested whether the same pathway was also activated after AGE-CD36 interaction. They found that platelets treated with AGE-BSA, but not native BSA, markedly increased phosphorylated JNK2 (p-JNK2) in a CD36-dependent manner. The levels of p-JNK2 also increased 2- to 3-fold in platelets isolated from WT, but not CD36^−/−, diabetic mice. Furthermore, immunohistochemical staining of carotid artery thrombi showed that p-JNK2 was 20% to 50% higher in WT diabetic mice than CD36^−/− diabetic mice. These data elegantly demonstrated that platelet CD36 is a prothrombotic receptor for AGE, and that AGE-CD36–mediated JNK2 activation may play an important role in the arterial thrombosis observed in diabetic patients.

Platelets are versatile. In addition to their acute responses to vascular injury that lead to hemostatic plug or thrombus formation, platelets can also de novo synthesize proteins in response to inflammation, alterations in plasma fibrinogen levels, and other mild environmental changes.^8-10  This chronic response may have important biologic functions. However, it is currently unknown whether the platelet response to AGE via CD36 can also lead to alterations of platelet proteins and granule components, which may affect the long-term process of atherosclerosis as well as other diabetic vascular complications such as the renal failure and retinal damage observed in the later stage of diabetes. It is also unclear whether polymorphisms of CD36 affect the AGE-CD36 interaction and the progression of diabetes. In addition, because platelets and platelet activation are also involved in deep vein thrombosis, it is possible that the AGE-CD36-JNK2 pathway may also play a role in venous thromboembolism. These questions, including potential targeting of the CD36-JNK2 pathway for antithrombotic therapy, are worthwhile for future studies.