Regulation of platelet function by class B scavenger receptors in hyperlipidemia

Abstract

Platelets constitutively express class B scavenger receptors CD36 and SR-BI, two closely related pattern recognition receptors best known for their roles in lipoprotein and lipid metabolism. The biological role of scavenger receptors in platelets is poorly understood. However, in vitro and in vivo data suggest that class B scavenger receptors modulate platelet function and contribute significantly to thrombosis by sensing pathological or physiological ligands, inducing prothrombotic signaling, and increasing platelet reactivity. Platelet CD36 recognizes a novel family of endogenous oxidized choline phospholipids that accumulate in plasma of hyperlipidemic mice and in plasma of subjects with low HDL levels. This interaction leads to the activation of specific signaling pathways and promotes platelet activation and thrombosis. Platelet SR-BI, on the other hand, plays a critical role in the induction of platelet hyper-reactivity and accelerated thrombosis in conditions associated with increased platelet cholesterol content. Intriguingly, oxidized HDL, aSR-BI ligand, can suppress platelet function. These recent findings demonstrate that platelet class B scavenger receptors play roles in thrombosis in dyslipidemia and may contribute to acute cardiovascular events in vivo in hypercholesterolemia.

Introduction

Scavenger receptors (SR) are a group of structurally heterologous cell surface receptors that share an ability to recognize chemically modified or oxidized forms of LDL. SR belong to a wider family of pattern recognition receptors that mediate the innate immune host response which includes Toll-like receptors. Platelets express several SR including class B scavenger receptors CD36 and SR-BI, two closely related multiligand receptors, best known for their roles in lipoprotein and lipid metabolism¹. Other platelet SR include LOX-1 and CD68 ¹. Expression of class B SR in platelets is mainly constitutive while other receptors are rapidly exposed upon platelet activation. Ligands for these receptors can be roughly divided into three groups, physiological ligands, pathological endogenous (changed self) ligands, and pathological exogenous ligands. Pathological endogenous ligands for platelet SR may be present in circulation in a number of pathophysiological states related to dyslipidemia and oxidative stress. Pathological exogenous ligands may be present in cases of infections. The biological role of SR in platelets is not understood yet. However, evidence is accumulating that SR contribute significantly to thrombosis by sensing pathological or physiological ligands, inducing prothrombotic signaling, and increasing platelet reactivity. This in turn may lead to thrombosis in the presence of threshold concentrations of agonists. Platelet hyper-reactivity or increased platelet response to agonists is associated with augmented platelet adhesion, integrin activation, and aggregation ^2–4. Subjects with increased measures of platelet reactivity are at increased prospective risk for coronary events and death ^3–7. The pathophysiological significance of prothrombotic effects of platelet-hyper-reactivity may be very significant ⁸. The mechanisms responsible for enhancing platelet reactivity in vivo during dyslipidemia are gradually emerging, particularly due to the availability of novel murine knockout models, but these mechanisms are still poorly understood. The data demonstrating that class B scavenger receptors modulate platelet reactivity in conditions of hyperlipidemia and oxidative stress and contribute to cardiovascular events are reviewed below.

Platelet CD36, oxidized phospholipids, and thrombosis

CD36 is a multifunctional cellular receptor with broad ligand specificity that is expressed on macrophages, platelets, microvascular endothelial cells, and other cells⁹. It is structurally comprised of two transmembrane and two cytoplasmic domains, as well as a large heavily glycosylated extracellular domain. CD36 regulates cellular adhesion and angiogenesis, serving as a receptor for thrombospondin. It serves as a scavenger receptor in macrophages, mediating uptake of apoptotic cells and modified lipoproteins, and participates in carbohydrate and lipid metabolism, modulating insulin resistance and long chain fatty acid transport^10–15. CD36 has been implicated in a variety of pathologic conditions, including atherosclerosis, diabetes and cardiomyopathy^{10, 13, 16–17}.

Even though multiple lines of evidence suggest that CD36 may play a role in platelet activation, earlier studies of platelet function isolated from CD36 deficient patients did not demonstrate a significant role for platelet CD36 in physiological conditions^18–22. Since then, a number of various physiological and pathological ligands for CD36 have been identified raising a possibility that CD36 may play a role in platelet activation by pathological ligands. The incomplete list of CD36 ligands includes ghrelin, diacylated bacterial lipopeptide, lipoteichoic acid, phosphatydylserine, beta-amyloid, serum amyloid A, and specific oxidized phospholipids (oxPC_CD36) ^23–29. oxPC_CD36 are generated when LDL or cellular phospholipids containing polyunsaturated fatty acids at the sn-2 position undergo oxidative attack. In general, oxidized phospholipids that are generated via mild physiologically relevant oxidative processes have been shown to possess multiple biological activities and are associated with an increased risk of cardiovascular events ^30–33. oxPC_CD36 represent a small fraction of total oxidized phospholipids that possess high affinity for CD36 ^{29, 34}. Increased concentrations of oxPC_CD36 were detected at sites of oxidative stress such as human and animal atherosclerotic lesions ^34–35. oxPC_CD36 also accumulate in significant concentrations in plasma of hyperlipidemic apoE deficient and LDL receptor deficient mice ³⁶. In humans a subset of patients with low HDL levels has plasma oxPC_CD36 levels comparable to that of hyperlipidemic animals³⁶. In macrophages, oxPC_CD36 mediate uptake of oxidized LDL (oxLDL) via CD36 and, subsequently, foam cell formation. The roles of oxPC_CD36 and CD36 in atherosclerosis were extensively reviewed recently^{9, 31}. Since oxPC_CD36 are present in plasma, a hypothesis has been proposed that the interaction of oxPC_CD36 with platelet CD36 could modulate platelet function, potentially inducing pro-thrombotic signals associated with hyperlipidemia³⁶.

A defining feature of oxPC_CD36 is an sn-2 acyl group that incorporates a terminal gamma-hydroxy (or oxo)-alpha, beta-unsaturated carbonyl ²⁹ (Figure 1a). Recent studies of the conformation of oxPC_CD36 in the membranehave revealed that due to its polarity the oxidized sn-2 fatty acid moiety of oxPC_CD36 instead of being buried within the membrane, protrudes into the aqueous phase, rendering it accessible for the receptors (Figure 1b) ³⁷. Studies using a series of model oxidized phospholipids demonstrated that a terminal negatively charged carboxylate at the sn-2 position of oxPC suffices to generate high binding affinity to both class B scavenger receptors, CD36 and SR-BI ³⁸. The minimal binding domain of CD36 for oxPC_CD36, has been identified recently ³⁹. It contains two conserved lysine residues that are indispensable for the binding activity. Subsequently, interaction via an electrostatic mechanism of negatively charged sn-2 moiety of oxidized phosphatidylcholine, exposed on the phospholipid surface, with positively charged residues in CD36 has been shown to play a key role in high affinity binding ³⁹. It should be mentioned that in order to be a ligand for CD36, oxidized sn-2 residues should be part of the otherwise intact PC, containing the sn-1 hydrophobic chain and the sn-3 hydrophilic phosphocholine or phosphatidic acid group³⁸. Additional factors such as polarity, rigidity, optimal chain length of sn-2 and sn-3 positions, and a negative charge at the sn-3 position of phospholipids, further modulate the binding affinity ³⁸. Studies with model oxPC suggested that, in addition to oxPC_CD36, other oxidized phospholipids observed in vivo, possessing some but not all features of oxPC_CD36, represent weak ligands for class B scavenger receptors 38.

Figure 1. Structure and conformation of oxidized phospholipids specific ligands for CD36 (oxPC_CD36).

Truncated oxidized fatty acid esterified to the sn-2 position of oxidized phosphatidylcholine (PC) is shown. It contains a terminal aldehyde or carboxylic acid, a double bond in the βposition, and a ketone or alcohol in the γposition. The core structural motif conserved amongst the various oxidized PC species that are high affinity ligands for CD36 is shown in red. b) Unusual conformation of oxPC_CD36 in the phospholipid membrane. The truncated polar sn-2 fatty acid moiety of oxPC_CD36 protrudes into the aqueous phase, rendering it accessible for the receptors such as class B scavenger receptors.

In vitro studies have shown that oxPC_CD36 bind to platelets via CD36, activate platelets, and sensitize platelets to other agonists at pathophysiological concentrations ³⁶. In vivo relevance of these observations came from experiments with hyperlipidemic mice. A number of groups using several thrombosis models have demonstrated that hyperlipidemic apoE deficient mice or LDLR deficient mice do have a prothrombotic phenotype ^{36, 40–41}. The platelet contribution to this phenotype has been established only recently. In vitro studies demonstrated that platelets of hyperlipidemic apoE deficient mice are hyper-reactive to physiological agonists ³⁶, similar to platelets from hypercholesterolemic patients and other species of hypercholesterolemic animals ^{2, 42–44}. It has been demonstrated that a factorconferring hyper-reactivity resides in hyperlipidemic plasma and identified as oxPC_CD36 ³⁶. Genetic deletion of CD36 protected mice from hyperlipidemia-associated enhanced platelet reactivity and the pro-thrombotic phenotype. Moreover, studies in vitro have demonstrated that endogenous oxPC_CD36 in human plasma modulate the platelet response to agonists via CD36 ³⁶. oxPC_CD36 binding peptides significantly inhibited thrombosis in vivo in hyperlipidemic apoE−/− mice, and had only a weak effect in normolipidemic mice (S. Panigrahi and E. Podrez, unpublished data). Similar results were obtained using apoA-I mimetic peptide L-4F, capable of removing oxidized phospholipids from the plasma through direct binding ⁴⁵. Interestingly, recent findings demonstrated that amyloids appearing as a result of protein misfolding induces platelet aggregation via an activation pathway mediated by CD36, p38 (MAPK), and thromboxane A2 ⁴⁶. It has been established that oxidized LDL contains amyloid-like structures suggesting that this may be another CD36 dependent pathway activated by oxidized LDL in platelets ⁴⁷.

In resting platelets, membrane CD36 is palmitoylated and localized in caveolin negative cholesterol enriched lipid rafts which also contain several Src family tyrosine kinases ^48–49. There are also data indicating that in these lipid rafts CD36 exists as a multiprotein complex containing, in addition to Src-family kinases, tetraspanin proteinCD9 and fibrinogen receptor integrin αIIbβ3 or laminin receptor integrin α6β1 ^{48, 50–51}. The CD36 enriched fraction of platelet membranes is also enriched in GPIb, another major platelet glycoprotein essential for normal platelet adhesion and clot formation at sites of vascular injury⁴⁹. The functional significance of the association of CD36 with platelet glycoproteins is not clear even though it has been shown that signaling events induced via CD36 may lead to activation of αIIbβ3 ³⁶. Src-family kinases associated with CD36 in platelets include Fyn, Lyn, and c-Yes ⁴⁸. Lyn and CD36 do not interact directly, but do so through a lipid mediator⁵², further pointing to the importance of raft-associated lipids in CD36-mediated signal transduction. It has been proposed that in resting platelets CD36-anchored Lyn is negatively regulated by Csk homologous kinase (CHK)⁵³. Upon platelet activation CD36-anchored Lyn is released from suppression through the rapid translocation of CHK and subsequent dephosphorylation of its C-terminal negative regulatory tyrosine residue⁵³.

It has been long established that oxidized LDL induces platelet activation and adhesion and enhances platelet aggregation ^54–55. Signaling pathways induced by oxLDL in platelets include activation of Rho/Rho kinase ⁵⁶ and Src-family and Syk tyrosine kinases ^57–58. A number of recent studies provided evidence of the signaling events initiated by oxLDL in platelets and mediated by CD36. One of the earlier events in this activation is the recruitment and activation of Lyn and Fyn to CD36 ⁵⁹. This is followed by the downstream activation of the protein kinases MKK4 and JNK2 ⁶⁰ (Figure 2). The significance of this pathway is supported by the observation that pharmacologic inhibition of JNK prolonged thrombosis times in wild type mice, but not in CD36 deficient mice ⁵⁹. The downstream targets of Lyn and Fyn include the Vav family of proteins. Vavs are rapidly phosphorylated in platelets in a CD36-dependent manner ⁵⁹. Silverstein et al reported that genetic deletion of Vav family members rescued the high-fat diet induced prothrombotic phenotype in mice in a way similar to the rescue by CD36 knockout ⁵⁹. Downstream of CD36 is also p38 MAPK which, according to a recent report, requires the scavenger receptor, class A for maximal activation by oxLDL ⁶¹. Mechanisms described in other cell types may be relevant to platelet activation by oxLDL but require further studies. In macrophages, the generation of reactive oxygen species by oxLDL takes place through the formation of a complex between CD36 and a heterodimer of toll-like receptors (TLR4 and TLR6)⁶². Also in macrophages, oxLDL induces the phosphorylation of FAK and the inactivation of the phosphatase SHP2 that are responsible for cytoskeletal changes⁶³. Even though oxPC_CD36 is only one biologically active component in oxidized LDL we have observed that oxLDL and oxPC_CD36 share activation pathways. Platelet activation induced by both oxLDL60 and oxPC_CD36 can be prevented by Src-kinase inhibitors (A. Zimman and E. Podrez, unpublished data). Also, both oxLDL and oxPC_CD36 activate PKC and induce the phosphorylation of Syk ⁵⁷ (A. Zimman and E. Podrez unpublished data). However, it should be mentioned that a complex ligand such as oxidized LDL contains multiple biologically active components that can affect platelet function in a CD36 independent manner. Examples include oxidized lipids that increase intracellular Ca²⁺ through the PAF receptor⁶⁴, lysophosphatidylcholine⁶⁵, oxysterols and TBARS⁵⁸, lysophosphatydic acid⁶⁶ and amyloid-like structures ⁴⁶.

Figure 2. Signal transduction pathways induced by oxLDL and mediated by CD36 in platelets.

oxPC_CD36 in oxLDL induces specific signaling cascades after binding to the scavenger receptor CD36. (?) indicates pathways identified in macrophages only.

SR-BI, dyslipidemia, oxidative stress, and platelet function

SR-BI shares 30% sequence homology with CD36, has similar structural organization, and shares with CD36 several common ligands including oxidized lipoproteins and oxidized phospholipids ^67–69. Recently the presence of this receptor has been demonstrated on the surface of human and murine platelets ^70–71. The major physiological function of SR-BI is to serve as a receptor for HDL and to promote selectiveuptake of HDL cholesteryl esters in steroidogenic tissues and liver ⁷². In addition, SR-BI stimulates the bidirectional flux of free cholesterol between cells and lipoproteins, modifies membrane cholesterol distribution, and induces signaling events⁷³. Since the antithrombotic function of HDL is well established, it seems logical that HDL interaction with platelet SR-BI may inhibit platelet function. However, previous reports on the effect of HDL on platelets have been contradictory ^74–78. One explanation for the inconsistency may be the inadvertent oxidation of HDL during the isolation procedure. When extreme care was taken to prevent HDL oxidation during isolation, HDL or HDL subfractions HDL2 and HDL3 had no significant effect on the activation and aggregation responses of isolated platelets ⁷¹. This result looks particularly surprising, since HDL induces cholesterol efflux from cells, and in turn, the removal of cholesterol from platelets using cholesterol chelator MβCD is associated with a strong inhibition of platelet activation ⁷⁹. Recently, Calkin et al showed that reconstituted HDL, but not HDL, does inhibit platelet activation in vivo ⁸⁰. These authors suggested the reason for the difference is that reconstituted HDL has a higher capacity for cholesterol efflux than native HDL. Intriguingly, HDL oxidized by various pathways potently inhibits platelet activation and aggregation induced by physiological agonists ⁷¹. Studies using murine CD36^−/− or SR-BI^−/− platelets demonstrated that the anti-thrombotic activity of oxHDL requires binding to platelet SR-BI but not to CD36 and that oxidized phospholipids in oxHDL are likely to play a role in this activity ⁷¹. Studies involving SR-BI in platelets are very limited for now. In endothelial cells, one of the most important downstream targets of SR-BI is the activation of eNOS upon HDL binding to this receptor ⁸¹. c-Src and PDZK1 form a complex with SR-BI in this cell type. Our preliminary studies have shown the association of SR-BI to c-Src in platelets (Podrez et al, unpublished data). However, the effect of oxHDL on platelets is mediated by a pathway different from the eNOS/Akt pathway operating in endothelial cells⁷¹. A global survey of changes in protein phosphorylation by mass spectrometry ⁸² revealed that oxHDL induces multiple changes in the phosphorylation of platelet proteins, including a group of proteins associated with focal adhesion and tight junctions (A.Zimman and E.Podrez, unpublished data). Taken together this surprising finding suggests that lipid oxidation and oxidative stress appear to trigger not only prothrombotic effects mediated by CD36 but also to produce a previously unrecognized defense mechanism against thrombosis mediated by SR-BI.

The role of SR-BI deficiency and, specifically, platelet SR-BI deficiency in platelet function and thrombosis in vivo has been studied directly using SR-BI^−/− mice. It should be taken into account that complete SR-BI deficiency is characterized by severe dyslipoproteinemia associated with platelet abnormalities, includinghigh unesterified cholesterol content, morphological changes, and thrombocytopenia. Platelets of SR-BI^−/− mice display either reduced or normal responses to selected physiological agonists ⁸³. This phenotype is particularly surprising taking into account that platelet enrichment in unesterified cholesterol has previously been directly linked to increased platelet reactivity ^{42, 84–86}. To separate the effects of SR-BI dependent dyslipidemia and platelet specific SR-BI deficiency in platelet function, a bone marrow transplantation approach has been used. SR-BI^−/− platelets isolated from normolipidemic wild type mice chimeras have normal platelet cholesterol content, normal morphology and show mild reduction in response to physiological agonists that was mostly evident at high concentrations of agonists⁸⁷. In vivo studies demonstrated that wild type recipient mice with SR-BI deficient bone marrow have delayed thrombosis in ferric chloride induced carotid artery thrombosis assay, demonstrating that this seemingly mild defect still contributes significantly to thrombosis⁸⁷. Wild type platelets isolated from dyslipidemic SR-BI^−/− chimeric mice had cholesterol content increased by 60% and exhibited augmentation of activation responses to physiological agonists, in agreement with the published data on the impact of cholesterol on platelets ⁸⁵. Results from this study suggest that the unusual platelet phenotype observed in SR-BI deficient mice is a reflection of the opposite effects of platelet SR-BI deficiency and platelet cholesterol overload on platelet reactivity. Most interestingly, in vitro studies revealed that SR-BI deficient platelets are remarkably resistant to hyper-reactivity induced by cholesterol overload, suggesting that platelet SR-BI may play a particularly prominent role in thrombosis in conditions of hyper-cholesterolemia. Indeed, in vivo studies confirmed that platelet SR-BI expression is prothrombotic in vivo in dyslipidemia associated with either apoE or SR-BI deficiency. These data strongly suggest that platelet SR-BI plays an important role in the widely known induction of platelet hyper-reactivity by high platelet cholesterol content ^{42, 84–86}. It should be mentioned that the relevance of these studies to humans is not clear yet. No SR-BI deficient patients have been reported so far. However, a single nucleotide polymorphism of the SCARB1 gene was associated with changes in SR-BI protein levels, which in turn were inversely associated with HDL cholesterol levels and HDL particle size ⁸⁸, resembling HDL changes in SR-BI^−/− mice. SCARBI genetic polymorphisms in humans are associated with variations in plasma lipoproteins and coronary artery disease ^89–91. Whether platelet reactivity contributes or not to variations in coronary artery disease is not known. However, SCARB1 polymorphisms show associations with the risk for peripheral artery disease, a condition where platelet dependent thrombosis plays a prominent role ⁹².

The molecular mechanism of SR-BI involvement in the regulation of platelet function is currently unknown. The function of SR-BI demonstrated in platelets is related to platelet unesterified cholesterol content. Thus it is of particular interest that the C-terminal transmembrane domain of SR-BI directly binds cholesterol ⁹³. This cholesterol binding property is unique to SR-BI among three class B scavenger receptor family members. It has been proposed recently that SR-BI senses alterations in the cholesterol environment in the plasma ⁹³, leading to changes in association of SR-BI with partner protein(s) critical to downstream signaling. We observed that the sensitivity of platelets to physiological agonists is upregulated with increasing platelet cholesterol content if platelets express SR-BI. In the absence of SR-BI expression, responses to ADP and convulxin are oblivious to platelet cholesterol content, and the increase in response to thrombin is significantly blunted⁸⁷. One possibility is that enrichment in plasma membrane cholesterol leads to conformational changes in SR-BI, and upregulates SR-BI-dependent signaling pathways in platelets. On the other hand, membrane depletion of cholesterol inhibits the SR-BI dependent component of platelet signaling, leaving only a “basal” level of signaling. It should be noted, however, that SR BI expression in other cell types is associated with significant changes in the organization of cholesterol in cell membranes ^94–95. Cell membrane cholesterol and, particularly, cholesterol in specialized membrane microdomains is important in the assembly of signaling complexes and plays an important role in signaling events in platelets⁹⁶. SR-BI is highly enriched in caveolae in several cell types ⁹⁷. Thus, it may affect signaling events in platelets in a less specific way, by changing membrane structure. In our unpublished studies we observed that cholesterol flux to HDL is reduced by ~50% in SR-BI deficient platelets compared to wild type platelets, potentially reflecting changes in cholesterol organization.

In conclusion

In the context of dyslipidemia and enhanced oxidative stress, platelet CD36 interaction with a small subset of endogenous oxidized phospholipids leads to pro-thrombotic signaling and accelerated thrombosis. The effects of other CD36 ligands on platelet function and thrombosis need further investigation. The role of SR-BI in platelet function is more complex. Liver deficiency of SR-BI leads to platelet hyper-reactivity and accelerated thrombosis due to high cholesterol content in platelets. Platelet SR-BI plays a critical role in the induction of platelet reactivity by high platelet cholesterol content. On the other hand, specific ligands that are generated in oxidative stress and interact with platelet SR-BI may induce suppression of platelet function. It is tempting to speculate that platelet class B scavenger receptors might be a specific target for antiplatelet therapies specifically aimed at platelet hyper-reactivity in dyslipidemia.