[Frontiers in Bioscience 16, 2144-2160, June 1, 2011] |
|
|
STIM and Orai in hemostasis and thrombosis Attila Braun1, Timo Vogtle1, David Varga-Szabo2, Bernhard Nieswandt1
1 TABLE OF CONTENTS
1. ABSTRACT At sites of vascular injury, platelets rapidly adhere to the exposed subendothelial extracellular matrix, become activated and, together with the coagulation system, form a plug that seals the lesion. This process is essential to prevent blood loss, however, under pathological conditions it may lead to vessel occlusion. Agonist-induced elevation of intracellular Ca2+ levels is essential for platelet activation. It occurs through two different mechanisms: Ca2+ release from internal stores, involving phospholipase C (PLC)-dependent generation of inositol-1,4,5-trisphosphate (IP3) and activation of IP3 sensitive channels in the store membrane, and Ca2+ influx across the plasma membrane. Store operated Ca2+ entry (SOCE), triggered by store depletion, is the main influx pathway for extracellular Ca2+ in platelets, but the molecular mechanism underlying this pathway has long remained elusive. In the last years, however, the Ca2+ sensor stromal interaction molecule 1 (STIM1) and the channel protein Orai1 emerged as the key players in platelet SOCE. This review summarizes the current knowledge about the role of these proteins in platelet physiology and thrombus formation and discusses their suitability as antithrombotic targets. 2. INTRODUCTION Platelets are small anucleated cell fragments that originate from the cytoplasm of bone marrow megakaryocytes and circulate in the bloodstream for about ten days. In human blood, 150 - 450 x 106 platelets per ml are present, in mice around 1 x 109 platelets per ml. In the intact vasculature, platelets never undergo firm adhesion and are finally cleared from the circulation by macrophages in liver and spleen. However, when contacting damaged vessel walls, circulating platelets rapidly decelerate and adhere to the newly exposed subendothelial extracellular matrix (ECM). Sustained interaction of platelet receptors with components of the ECM triggers platelet activation, characterized by rearrangements of their cytoskeleton resulting in a change from discoid to spheric shape, fusion of intracellular alpha and dense granules with the plasma membrane and release of secondary platelet agonists (so-called second wave mediators), most notably adenosine diphosphate (ADP), adenosine triphosphate (ATP), serotonin and thromboxane A2 (TxA2). These agonists along with locally produced thrombin induce intracellular signaling events through receptors coupled to heterotrimeric G proteins, thereby enhancing platelet activation and recruitment of additional platelets from the bloodstream to the site of injury. Finally, this sophisticated interplay of extra- and intracellular signaling events leads to functional upregulation of adhesion receptors of the integrin class, most notably integrin alphaIIbbeta3 (also called glycoprotein (GP)IIb/IIIa) resulting in the establishment of stable platelet-ECM contacts and, through the binding of plasma fibrinogen, in platelet aggregation. In conjunction with an increase of the procoagulant activity of the cells, these events lead to the formation of a stable thrombus that seals the wound (1-2) (Figure 1). This process, called primary hemostasis, is essential to limit excessive blood loss after tissue trauma. On the other hand, in diseased vessels, uncontrolled thrombus formation can lead to acute vessel occlusion and ischemia or infarction of vital organs. Myocardial infarction, triggered by erosion or rupture of an atherosclerotic plaque in coronary arteries, and stroke, elicited by carotid thrombi that embolize to the brain, are two of the leading causes of disability and mortality worldwide (3). Therefore, a better understanding of platelet function and thrombus formation is required in order to develop new therapeutic strategies to prevent and treat those diseases. Calcium (Ca2+) serves as an essential second messenger in virtually all eukaryotic cell types, regulating fundamental cellular processes such as exocytosis, motility, apoptosis and gene transcription (4-6). Each cell type expresses a distinct set of Ca2+ sensitive effector proteins and regulators of intracellular free Ca2+ levels ((Ca2+)i) to assemble a Ca2+ signaling system that fulfills the cell's individual needs for spatial and temporal Ca2+ dynamics. In platelets, changes in (Ca2+)i contribute to a variety of responses including granule release (7-8), cytoskeletal rearrangements required for shape change (8-9), coagulant activity (10-11) and inside-out activation of integrins, indispensable for platelet aggregation (12-13). Various agonists, such as collagens, ADP, TxA2 or thrombin evoke a rise in (Ca2+)i that consists of two parts: release of compartmentalized Ca2+ from intracellular stores and influx of extracellular Ca2+ across the plasma membrane. This Ca2+ influx from the extracellular space potentiates and extends the Ca2+ signals derived from store release and further stimulates platelet activation (14). Already in 1991, it was found that the main Ca2+ entry mechanism in human platelets is store-operated Ca2+ entry (SOCE) (15-16), where Ca2+ influx is triggered by the depletion of Ca2+ from the intracellular stores (17). However, it took additional 15 years until the key components of the underlying molecular machinery were identified: the endoplasmic reticulum (ER)-resident Ca2+ sensor stromal interaction molecule 1 (STIM1) and the store operated Ca2+ channel Orai1 (also called CRACM1). The identification of the molecular basis of SOCE enabled genetic analyses and genetic manipulations of this process in mouse models that markedly increased our knowledge about Ca2+ signaling in platelet activation and thrombus formation. This review summarizes the major advances in understanding Ca2+ signaling in platelets in health and disease with a particular focus on the STIM and Orai proteins. 3. ROLE OF INTRACELLULAR Ca2+ IN THROMBUS FORMATION Exposure of highly thrombogenic subendothelial ECM components upon vessel wall injury leads to rapid recruitment of platelets that become tethered to and "roll" along the site of injury in a stop-and-go manner (Figure 1). Under high shear flow conditions, such as found in arterioles or stenosed arteries, this process is mediated by the interaction of two molecules: the multimeric plasma glycoprotein von Willebrand factor (vWF), that becomes immobilized on the exposed subendothelial collagen, and the platelet receptor GPIb, itself part of the GPIb-V-IX complex (18). This interaction is, however, characterized by a fast dissociation rate and therefore does not allow stable platelet adhesion by itself (18). Thus, the principal function of this interaction is to reduce the velocity of platelets and to enable other platelet receptors, such as collagen receptors, to establish contacts with the thrombogenic surface. Besides this mandatory role for platelet adhesion, GPIb also acts as a signal transducer that can directly induce weak activation of the principal adhesion receptor in platelets, the integrin alphaIIbbeta3 (19). GPIb-dependent binding of platelets to vWF-coated surfaces under flow has been shown to induce transient Ca2+ spikes, which initiated transient platelet arrest and integrin alphaIIbbeta3 activation (20-21). These signals seem to derive from intracellular stores, as addition of the extracellular Ca2+ chelator EGTA had no inhibitory effect on the magnitude or duration of the Ca2+ spikes (20, 22). GPIb-dependent deceleration of platelets allows the establishment of contacts between collagen and the immunoglobulin-like receptor GPVI, the major activating platelet collagen receptor (2, 23) (Fig.1). Binding of GPVI to collagen does not mediate stable adhesion, but induces intracellular signaling events, which, through a strong and sustained rise in (Ca2+)i, finally leads to cellular activation, release of second wave mediators, most notably ADP and TxA2, and exposure of negatively charged phosphatidylserine (PS) on the platelet surface (24-25). The latter provides a platform for local thrombin generation by the coagulation system which, together with the released second wave mediators, further activates the platelet (26). These agonists signal through different G-protein coupled receptor (GPCR) pathways also resulting in Ca2+-release and SOCE thereby amplifying and sustaining the initial GPVI induced Ca2+ signals (see below) (27). In addition, released ATP enhances the activation processes triggered by GPCR and GPVI via the activation of the ATP gated Ca2+ channel P2X1 (28-30). All these signaling events converge in the final common step of platelet activation that is the shift of integrin alphaIIbbeta3 from a low to a high affinity state, which is a prerequisite for fibrinogen binding and stable thrombus formation (2) (Figure 1). A newly identified platelet receptor involved in thrombus formation is the C-type lectin-like receptor 2 (CLEC-2) (31-32). Studies using a CLEC-2 depleting antibody (31) or genetically modified mice (33) have shown that this receptor is of particular importance for the stabilization of growing thrombi. Albeit the physiological ligand of CLEC-2 in this process is not known, stimulation of CLEC-2 with the snake venom toxin rhodocytin revealed a downstream signaling pathway similar to GPVI (32, 34). Therefore, it is likely that, like in GPVI signaling (see below), Ca2+ plays also an important role in CLEC-2 signaling. However, this process has not been characterized in detail. 4. Ca2+ SENSITIVE PROTEINS INVOLVED IN PLATELET ACTIVATION In immune cells, Ca2+ signals are especially important for the activation of transcription factors, such as nuclear factor of activated T-cells (NFAT), which influence gene expression patterns and thereby shape the outcome of cell activation (6). In sharp contrast, platelets lack a nucleus and, upon vascular injury, have to be activated within seconds. Thus, Ca2+ signals have to be rapidly translated into a cellular response. Despite the paramount importance of Ca2+ in platelet signaling, not much is known about the molecular identity of the proteins involved in this process. An essential key regulator in Ca2+-dependent platelet activation is the highly Ca2+-sensitive Ca2+ and diacylglycerol regulated guanine nucleotide exchange factor I (CalDAG-GEFI) that has been characterized recently (35-37). In response to Ca2+, CalDAG-GEFI activates Rap1b (35-36), an abundant small GTPase in platelets, required for integrin activation and thus for normal platelet function and hemostasis (38). In consequence, CalDAG-GEFI deficient platelets exhibited marked defects in alphaIIbbeta3-mediated aggregation response to all tested agonists in vitro, causing a remarkable bleeding phenotype in tail bleeding assays and protection in different thrombosis models (35, 39). A more detailed analysis of CalDAG-GEFI-/- platelets revealed a critical role of the molecule in Erk-dependent TxA2 generation, providing a crucial feedback for protein kinase C (PKC) activation and ADP release, thus placing CalDAG-GEFI at the center of Ca2+ dependent platelet activation (37). Two other major Ca2+ dependent proteins in platelets are the cysteine proteases calpain-1 and 2 which are involved in the modification of the cytoskeleton, thus affecting cell spreading, aggregation and clot retraction by indirectly reducing the tyrosine phosphorylation status of several proteins involved in integrin signaling (40-41). Ca2+- and integrin-binding protein 1 (CIB1) seems also to be involved in signaling processes downstream of integrins, as its absence has been shown to interfere with the spreading of platelets on collagen (42). Besides this role in integrin signaling, calpains participate in the intracellular cleavage of different platelets receptors, thus downregulating platelet reactivity (43). The same effect may be achieved by extracellular cleavage of receptors by sheddases. This process is facilitated by the dissociation of the Ca2+ sensor calmodulin from the juxtamembrane cytoplasmic sequences of sheddase substrates, including GPVI and GPIb-IX-V and other receptors (44). This occurs upon platelet stimulation (45-46) or treatment with calmodulin inhibitors (47-48). As a consequence, Ca2+ positively regulates receptor proteolysis via calpains and calmodulin. Ca2+ is also a potent stimulus for platelet granule release but the molecular basis of this process has not been established. However, calmodulin, calpains, synaptotagmins and other proteins seem to be involved in this process (49). 5: Ca2+ HOMEOSTASIS IN PLATELETS Platelet agonists increase (Ca2+)i by two mechanisms: Ca2+ release from intracellular stores and Ca2+ entry from the extracellular space into the cytosol. The dense tubular system (DTS) - an internal smooth ER membrane system - is the major Ca2+ store in platelets. It has a dual function in Ca2+ homeostasis: on the one hand it keeps platelets in a resting state by sequestering Ca2+, on the other hand it modulates platelet activation by releasing Ca2+ into the cytosol. Physiologically, Ca2+ is released from the DTS into the cytosol when activated platelet receptors couple to different isoforms of the phospholipase C (PLC) enzyme and trigger production of the second messenger inositol-1,4,5-trisphosphate (IP3) by the hydrolysis of phosphatidyl-1,4-bisphosphate (PIP2). In platelets, IP3 production is regulated by two different PLC subfamilies, beta and gamma. IP3 binds and activates IP3-receptors (IP3R) in the DTS membrane and activates Ca2+ store release (50). When the Ca2+ reservoir of the DTS becomes exhausted, extracellular Ca2+ enters into the cytoplasm. During this process, the thin elongated shape of the DTS changes to a rounded vesicular form (51) inducing redistribution of the cortical actin cytoskeleton (52-53). The other product of PIP2 hydrolysis, 1,2-diacyl-glycerol (DAG), induces PKC activation and receptor-operated Ca2+ entry (ROCE) (50). Platelet ROCE can also occur through an additional pathway, namely the ATP-gated cation channel, P2X1 (54). This induces transient Ca2+ influx through the plasma membrane and contributes to platelet activation by stimulating the PKC/Erk2 pathway leading to shape change and granule release (29, 54-56) (Figure 2). Platelets possess a Ca2+ toolkit (Figure 2) that upon Ca2+ influx rapidly removes Ca2+ from the cytosol. Ca2+ will be either pumped back into the DTS or into other specialized storage compartments like mitochondria and lysosome-like "acidic stores" (57) or it will be pumped out of the cell. Many well known Ca2+ pumps are linked to these control mechanisms, including sarcoplasmic endoplasmic reticulum Ca2+-ATPases (SERCAs) and plasma membrane Ca2+-ATPases (PMCAs). However, in platelets these Ca2+ store coupled regulatory processes are still poorly understood. Ca2+ store dependent platelet activation is triggered by subendothelial collagens, by TxA2 and ADP released from activated platelets, and by thrombin generated through the coagulation cascade. These agonists trigger different signaling pathways, all commonly leading to the activation of PLCs (Figure 2). ADP released from dense granules upon platelet activation activates shape change and reversible platelet activation by activating the G-protein coupled P2Y1 and P2Y12 receptors (58). TxA2 is produced by activated platelets from arachidonic acid and mediates platelet activation through its GPCRs thromboxane-prostanoid (TP)alpha and TPbeta (59). Thrombin also activates GPCRs, called protease activated receptors (PARs). Platelet activation through the Gq coupled PARs, TPs and P2Y1 is very fast as Gq directly binds to the PLCbeta enzyme (60). In contrast, GPVI-immunoreceptor tyrosine-based activation motif (ITAM) mediated IP3 production via the the PLCgamma2 enzyme is a slower process which is dependent on the subcellular localization, phosphorylation and activation of PLCgamma2 (61-63). 6. STIM PROTEINS REGULATE STORE CONTENT AND SOCE In the past few years, several models have been proposed for the activation of SOCE. The first model, proposing a diffusible messenger that could link store depletion to the activation of the Ca2+ channels in the plasma membrane, was published by Putney and colleagues (64). A second model was based on the conformational coupling of store and plasma membrane resident Ca2+ channels. This hypothesis favors the idea that upon store depletion conformational changes of the IP3R happen and this novel structure of the IP3R results in its physical attachment to the SOC channel to the plasma membrane (65-66). A third model suggested that upon store depletion, SOC channels are inserted into the plasma membrane through exocytosis of SOC containing vesicles (67). In this case, it has been proposed that in resting cells no or only very few SOC channels are expressed in the plasma membrane. This model was supported by experiments where blocking of vesicle secretion inhibited SOCE (68). To prove these models experimentally, SOCE was in the focus of platelet research for many years but the molecular identity of the SOC channel and its regulators as well as the physiological function of SOCE in vivo have long remained elusive. Stromal interaction molecules (STIMs) received much attention during the last few years because a number of results from different mammalian cell types strongly suggest that STIMs are the missing key modulators that link Ca2+ store release to Ca2+ entry through the plasma membrane and regulate the refilling process of the Ca2+ stores (69-71). Two different STIM isoforms, STIM1 and 2, are expressed in mammalians. The STIM1 isoform was originally identified as a potential tumor growth suppressor in rhabdomyosarcoma cell line (72). At that time it was not suspected to be involved in the regulation of SOCE and the function of the STIM2 protein was entirely unknown. Both STIM isoforms contain a signal peptide, an EF-hand motif, a sterile alpha motif, a transmembrane domain and a long cytoplasmic tail which contains protein interaction sites like a coiled-coil domain and proline rich sequences. The N-terminus of STIM1 including the EF-hand motif is located in the ER lumen where it binds Ca2+. Upon store depletion, Ca2+ dissociation from the EF-hand motif induces conformational changes of STIM1 allowing it to form plasma membrane near "puncta" structures. These "puncta" seem to represent a bridge linking the ER membrane to the plasma membrane, thereby enabling STIM1 to regulate the SOC channel function and the refill of intracellular stores (73). Interestingly, when key residues in the EF-hand - essential for Ca2+ binding - were mutated, STIM1 was constitutively active, staying permanently in the "puncta" structure and keeping the SOC channels open without change in the Ca2+ store content (70-71). The in vivo biological function of an EF-hand mutant STIM1 was first studied in platelets. Mice homozygously expressing a constitutively active mutant form of STIM1 (Stim1Sax), due to one single amino acid substitution in the EF-hand motif (D84G), died in utero and showed severe diffuse bleedings in different regions of the body, indicating a hemostatic defect. In contrast, mice expressing only one mutant allele (Stim1Sax/+) were viable but developed severe macrothrombocytopenia associated with megakaryocyte hyperplasia, splenomegaly, bone marrow fibrosis and a bleeding disorder (74). Detailed analysis of megakaryocyte function showed, however, that thrombopoiesis was not impaired in these mice. In contrast, platelets were rapidly cleared from the circulation due to a preactivated state as evidenced by elevated levels of activated integrins on the surface of these cells (74). Ca2+ signaling was investigated in Stim1Sax/+ platelets to better understand this complicated in vivo phenotype. Increased basal Ca2+ levels were found in the cytosol of unstimulated Stim1Sax/+ platelets that were based on permanently opened SOC channels in the plasma membrane. SOCE through the two major signaling pathways - namely the Gq-PLCbeta and GPVI-ITAM-PLCgamma2 pathways (Figure 2) - was also investigated. Using different agonists, interestingly only the GPVI-ITAM-PLCgamma2 dependent signaling was defective in Stim1Sax/+ platelets both in vitro and in vivo (74). These very first experiments already indicated that STIM1 might play a pivotal role in the regulation of platelet SOCE, and suggested that SOCE is not equally important for all signaling pathways, at least in platelets. Shortly after, mice lacking STIM1 were generated and platelet function was studied (75). Since Stim1-/- mice show marked growth retardation and die around 4 weeks after birth, these studies were performed in irradiated wild-type mice transplanted with Stim1−/− bone marrow. In these mice, injected Stim1−/− bone marrow stem cells repopulated all lineages of blood cells, including megakaryocytes, and mutant platelets were analyzed. In Stim1−/− platelets a severely defective Ca2+ response to all major agonists was found, clearly establishing SOCE as the major way of Ca2+ entry in platelets. Like for Stim1Sax/+ platelets, flow-cytometric measurements of integrin activation and degranulation revealed a selective signaling defect in the GPVI-ITAM-PLCgamma2 pathway in STIM1 deficient platelets. This was also observed in aggregation studies, where the response of STIM1 deficient platelets towards GPVI-agonists was completely abolished at low agonist concentrations and strongly reduced at higher concentrations, whereas it was normal in response to GPCR-agonists at all concentrations. Despite the severe reduction in Ca2+ entry, a residual Ca2+ influx was still detected in Stim1-/- platelets (Figure 3). This indicates that other molecules can also regulate Ca2+ influx in platelets, but only to a minor extent. One possible candidate molecule is STIM2 that has been shown to activate SOC channels in eukaryotic cells and is also expressed in platelets. Alternatively, the residual Ca2+ entry could be mediated by ROCE, independent of SOCE. DAG could be one of the metabolites which induces ROCE by direct activation of transient receptor potential (TRP) channels. TRPC6 is a non-selective cation channel that is activated directly by DAG in a membrane-delimited fashion, independently of PKC (76). In line with this, TRPC6 has indeed been proposed as a candidate mediating store-independent Ca2+ entry in human platelets (77). In addition, the ATP-gated Ca2+ channel P2X1 must be also considered to contribute to the residual Ca2+ influx in Stim1-/- platelets. Besides the severely impaired SOCE, a reduced Ca2+ release from intracellular stores upon agonist-induced platelet activation was seen in Stim1−/− platelets (75). Similar observations have also been made in Stim1-/- mast cells (78) indicating that STIM1 could be involved in the refilling process of the intracellular Ca2+ store. Although STIM2 is expressed in platelets, no obvious abnormalities in Ca2+ homeostasis of Stim2-/- platelets could be shown so far (79). The normal Ca2+ store release and SOCE of these cells indicate that STIM2 has a minor role in platelet Ca2+ homeostasis or its function can be fully compensated by STIM1. 7. IDENTIFICATION OF THE MAJOR SOC CHANNEL IN PLATELETS Studies on SOCE in platelets have long been hampered by the lack of identified SOC channels. The first potential candidate SOC channel, a TRP Ca2+ channel, was discovered in Drosophila (80). This finding accelerated the efforts to identify more members of the TRP gene family. Searching genome-project databases, three Drosophila TRP genes and at least 33 TRP genes in mammalian systems were identified (81). Based on sequence identities, the TRP protein family has been divided into seven sub-families; especially the TRPC subfamily received growing attention in platelet research (82). The subunit structure of all TRPCs has been preserved during evolution with six putative transmembrane helices and unique protein binding motifs in the amino- (ankyrin repeat) and carboxyl terminal regions (Dystrophin-homologous sequence). Two members of the TRPC family are expressed in platelets: TRPC1 and TRPC6. The already mentioned TRPC6 is highly abundant in the platelet plasma membrane and mediates nonselective-cation influx in response to activation with DAG (77). For over a decade, TRPC1 was suggested to be the major SOC channel in platelets. The de novo conformational coupling model proposed TRPC1 to mediate SOCE by a coupling-dependent activation through interaction of the protein with the type II IP3R upon intracellular store depletion (83-85). Other investigators, however, suggested TRPC1 to be located rather in intracellular membranes of human platelets, indicating a role for TRPC1 other than SOCE (77). Mice lacking TRPC1 display fully intact SOCE and generally unaltered Ca2+ homeostasis in platelets (86). In addition, platelet activation in vitro and in vivo is not altered in the absence of TRPC1. Furthermore, studies on human platelets revealed that the presumably inhibitory anti-TRPC1 antibodies used in previous studies to demonstrate the function of the channel, had no specific effect on SOCE and failed to bind to the protein (86). Together, these results provided evidence that SOCE in platelets is mediated by Ca2+ channels other than TRPC1 or that the lack of TRPC1 function could be fully compensated in platelets. The discovery of a novel putative SOC channel, Orai1, led to the long awaited breakthrough in this field. Orai1 was originally identified in a Drosophila cell line by three independent groups (87-89). It is a 33-kDa plasma membrane protein with four transmembrane domains. Glycosylation of Orai1 in human platelets increases its apparent molecular weight to 50 kDa (90). It appears that tetramer of Orai1 proteins assemble to form a functional SOC unit (91). In vitro studies have shown that upon store depletion Orai1 co-localizes and directly binds to STIM1 and allows extracellular Ca2+ to enter into the cytosol (92). The physiological role of Orai1, as a potential mediator of SOCE, was first studied in two human patients suffering from severe combined immunodeficiency (SCID) syndrome, caused by a point mutation in the Orai1 gene (93). Shortly after this discovery, a role for Orai1 in platelet SOCE was proposed (94), based mainly on the finding, that quantitative RT-PCR analysis of candidate SOC channels in human platelets showed a >30-fold higher expression of Orai1 when compared to TRPC1 or other TRPC channels. The function of this novel Ca2+ channel in platelets was first studied in mice lacking Orai1 (90). Orai1 deficiency in platelets resulted in severely reduced SOCE (Figure 3) in response to all major physiological agonists, but in contrast to STIM1 deficiency it had no effect on the store Ca2+ content. These data suggest that functional SOCE is not a prerequisite for proper store refilling and indicates that STIM1 presumably plays a direct, yet unidentified, role in this process. Further, lack of Orai1 resulted in strongly reduced, but not completely abolished SOCE, indicating that other channels might mediate this residual Ca2+ influx. Alternatively, the residual Ca2+ entry could be mediated by DAG induced ROCE, similar to Stim1-/- platelets. In vitro, the functional consequences of reduced SOCE were rather mild. Orai1 deficient platelets aggregated normally in response to GPCR agonists, like ADP, thrombin or the TxA2 analog U46619. Aggregation towards stimulation with GPVI agonists was diminished, but, in contrast to Stim1-/- platelets, these defects were completely overcome at higher agonist concentrations. This selective impairment of GPVI-ITAM signaling was also observed in flow-cytometric measurements of integrin activation and degranulation. At the same time, Bergmeier et al. generated Orai1 knock-in mice expressing the Orai1R93W mutant protein (95), which is the mouse equivalent to the Orai1R91W mutation that abrogates SOCE in T cells of human SCID patients (87). Mutant platelets were characterized by markedly reduced SOCE and impaired agonist-induced increase in (Ca2+)i. Further, Orai1R93W platelets showed reduced integrin activation and impaired degranulation when stimulated with low agonist concentrations under static conditions. This defect did, however, not significantly affect the ability of Orai1R93W platelets to aggregate or to adhere to collagen under arterial flow conditions ex vivo (see below). 8. STIM1 AND ORAI1 IN THROMBUS FORMATION Because of the pivotal importance of Ca2+ signaling for platelet activation, molecules involved in this process may serve as target structures for antithrombotic therapy. In different genetic mouse models, it has been demonstrated that attenuation of intracellular Ca2+ signals or the deletion of a Ca2+ responsive protein leads to impaired thrombus formation. For instance, the genetic ablation of the ATP sensitive Ca2+ channel P2X1 or the Ca2+ sensor CalDAG-GEFI both resulted in a reduced aggregation response towards collagen in vitro and resistance against lethal collagen/epinephrine induced pulmonary thromboembolism in vivo (29, 35). Conversely, the overexpression of human P2X1 in the megakaryocytic cell lineage produced a prothrombotic phenotype (30). Furthermore, enhanced SOCE has been suggested to contribute to the hypercoagulable state of platelets from patients suffering from Diabetes mellitus type 2 (96). In line with this, the strongly decreased Ca2+ influx in response to all major agonists found in Stim1-/- and Orai1-/- platelets resulted in dramatically impaired thrombus formation when blood of these mice was perfused over a collagen-coated surface ex vivo (97). At high shear rates, resembling arterial flow conditions, wild-type platelets adhere to collagen fibers and build aggregates that grow into large thrombi during the perfusion period. In the case of Stim1-/- and Orai1-/- platelets, the surface area covered by thrombi was reduced by 42% and 60%, respectively, and the total thrombus volume was reduced by even more than 80% (75, 90). Given that under static in vitro conditions integrin activation, granule release and aggregation of Stim1-/- and Orai1-/- platelets was not affected in response to GPCR agonists and only partially defective towards GPVI agonists, these results indicate that STIM1 and Orai1 mediated SOCE is especially important under flow conditions where agonist potency becomes limiting because of rapid dilution. In contrast to Orai1-/- platelets, Orai1R93W platelets exhibited normal aggregation to fibrillar collagen in a flow adhesion assay and the surface area covered by platelets was not significantly reduced compared to wild-type (95). These differences may be explained by different experimental setups; e.g. the lower flow rate used (1,700s-1 vs 1,000s-1 in (90) and (95), respectively), or, as the authors assume, the Orai1R93W mutation may not completely abolish Orai1 function in murine platelets. In line with the flow adhesion results, Stim1-/- and Orai1-/- bone marrow chimeric (BMc) mice were protected in different in vivo models of thrombus formation (75, 90). This effect was most profound in a mechanical injury model of the abdominal aorta, where thrombus formation is triggered predominantly by collagens and thus occurs in a highly GPVI/PLCgamma2-dependent manner (98). Likewise Orai1-/- BMc mice were protected in a model of otherwise lethal pulmonary thromboembolism, triggered by intravenous injection of collagen and epinephrine (90). These results are in accordance with the dramatically reduced SOCE and the selectively impaired aggregation response of STIM1 and Orai1 deficient platelets in response to the GPVI-specific agonist collagen related peptide or collagen in vitro. On the contrary, in a chemical injury model, where vessel injury is induced by topical application of FeCl3 on mesenteric arterioles, thrombus formation is more dependent on soluble mediators such as thrombin (99), and thus on GPCR signaling. In this model, only Stim1-/- BMc mice displayed impaired thrombus formation. The appearance of first aggregates was delayed in 50% of these chimeras and formation of stable thrombi was almost completely abrogated (75). In contrast, normal thrombus formation was observed in Orai1-/- BMc mice in this model (90) which is most likely attributed to the slightly higher (Ca2+)i achieved after stimulation compared to Stim1-/- platelets (Fig 3). The mechanistic explanation for this observation is the decreased store content found in Stim1-/- but not in Orai1-/- platelets. This indicates that already minor alterations in (Ca2+)i can make the difference between stable and unstable thrombi. In conclusion, these studies showed that SOCE is of particular importance for thrombus growth under conditions of high shear flow, which is mainly driven by the GPIb-GPVI-ITAM axis. The protective effect of STIM1, but not of Orai1 deficiency in the more thrombin-driven chemical injury model is likely to arise from the role of STIM1 in refilling intracellular stores and appears therefore not to be attributable to the defect in SOCE. The hemostatic and thrombotic function of platelets is not only dependent on their ability to adhere and aggregate at sites of vascular injury, but also on their ability to support blood coagulation. Procoagulant conversion of platelets is associated with specific biochemical and morphological changes, including calpain (and Ca2+)-dependent proteolytic cleavage of actin cytoskeleton proteins, activation of caspases, blebbing, membrane contraction and microvesiculation. In the process of procoagulant conversion, the redistribution of PS on the platelet surface is one of the early characteristics. This allows binding and assembly of enzyme complexes of the coagulation pathways and accelerates the clotting process. In non-stimulated platelets, PS is mostly sequestered in the inner layer of the plasma membrane (100). Upon platelet activation, platelets lose plasma membrane phospholipid asymmetry and PS is rapidly exposed on the surface. Although an increase in (Ca2+)i is essential to trigger the redistribution of PS, the Ca2+ dependent signaling mechanism underlying this externalization process is not fully understood. It has been demonstrated that PS exposure is dependent on store release and extracellular Ca2+ influx since a continuously elevated (Ca2+)i is obligate to maintain the externalization process (101). Ca2+ mobilizing agents - ionophores, thapsigargin (an inhibitor of SERCA) or thrombin - induce PS exposure and the formation of microvesicles in platelets (102). Platelet stimulation with single G-protein coupled agonists, like thrombin or ADP, results in limited PS exposure (103-104) whereas GPVI activation causes appreciable procoagulant activity (105-106). The basis of this difference might be that thrombin or ADP induced activation of Gq-PLCbeta is a rather rapid but transient process, whereas GPVI-ITAM stimulation causes more sustained activation of PLCgamma2 dependent signaling (107). Combined stimulation of the GPVI-ITAM-PLCgamma2 and the Gq-PLCbeta pathway results in high PS exposure (11). Accordingly, the moderate procoagulant response caused by thrombin is strongly amplified in the presence of thapsigargin in platelets, emphasizing the important role of store release and sustained high level of (Ca2+)i in this process. However, Ca2+ release from the stores alone is insufficient to explain such an effect (108) and SOCE has also been proposed to regulate, at least in part, the transmembrane movements of PS. Inhibition of SOCE with extracellular Ca2+ chelators reduces PS externalization, thereby supporting this notion (109). Analysis of STIM1-, STIM2- and Orai1- deficient platelets has shown that the Stim1-Orai1 complex also contributes to PS exposure (79). Platelets from Stim1-/- and Orai1-/- BMc mice formed only small aggregates on collagen under flow in vitro, and PS exposure was almost completely absent. The procoagulant index, calculated as the ratio of the surface coverage of PS-exposing platelets to the surface coverage of all platelets, was also greatly reduced in both knockout mice. Quantitative analysis of Ca2+ rises in single platelets demonstrated a nearly complete ablation of the average Ca2+ signal in Stim1-/- and a greatly reduced signal in Orai1-/- platelets. These data pointed out the importance of Stim1-Orai1 dependent Ca2+ entry in PS exposure under conditions of anticoagulation (thrombin inhibition). In line with these results, Bergmeier et al. also found that Orai1R93W platelets are defective in surface PS exposure, suggesting that Orai1 mediated SOCE is crucial for the procoagulant response of the platelet rather than for other Ca2+-dependent cellular responses (95). However, deficiency in STIM1 and Orai1 did not abolish PS exposure and thrombus formation under non-anticoagulation conditions in vitro. In the presence of thrombin, GPVI dependent PS exposure and procoagulant activity were only partly affected (79). In situ thrombin generation seemed to be sufficient for high PS exposure even in the absence of SOCE. This result suggested the presence of a compensating Ca2+ entry mechanism, which could be triggered either by STIM2 or by ROCE. In contrast, normal Ca2+ store release and influx in response to all agonists were found in Stim2-/- platelets (79). Furthermore, normal SOCE, thrombus formation and PS exposure were observed both under coagulantion and non-coagulation conditions (79).The role of ROCE in PS exposure in the absence of SOCE was also investigated by using SKF96365, a selective Ca2+ entry blocker. Strikingly, in both, Stim1-/- and Orai1-/- platelets SKF96365 further suppressed the GPVI-ITAM-PLCgamma2 and the Gq-PLCbeta induced Ca2+ influx, suggesting ROCE as an alternative Ca2+ entry pathway in platelets which independently regulates PS exposure (79). 9. INHIBITION OF STIM1 AND ORAI1 FUNCTION: A THERAPEUTICAL OPTION FOR STROKE? Despite the marked protection of Stim1-/- and Orai1-/- BMc mice in collagen-dependent models of arterial thrombosis, these and also the Orai1R93W mice, exhibited only mildly prolonged tail bleeding times, indicating that SOCE is more important for arterial thrombus formation than for primary hemostasis. This makes STIM1 and Orai1 potential targets for antiplatelet therapy. However, problems may arise from the widespread expression of STIM1 and Orai1 as well as from differences in STIM1/Orai1 expression and function between mice and humans (93). The most prominent defect in Stim1-/- and Orai1-/- mice is a pronounced perinatal lethality. Surviving mice display strongly reduced body size and weight and have a dramatically shortened life time (75, 78, 90, 110-113). The exact cause of death is not known, but myopathy, due to defects in skeletal muscle (112-113) and (cardio) pulmonary defects (75, 78) have been observed. Within the hematopoietic system, defects in mast cells (78, 113), macrophages (114) and B cells (110) have been found, while Stim1-/- T cells, despite exhibiting a pronounced hyperproliferative phenotype (111, 115), were still able to fulfill some effector functions (115). These impairments within the hematopoietic system are, however, not life limiting, as Stim1-/- and Orai1-/- BMc mice were viable and did not show obvious signs of disease (75, 90, 110). It is important to note, however, that laboratory mice grow up and are kept in a pathogen free environment and the ability of Stim1-/- and Orai1-/- BMc mice or conditional knock-out mice to cope with complex infections has not been tested. In contrast to mice, in humans lacking either functional STIM1 or ORAI1, a defect within the immune system is the life limiting factor (93, 116-117). The few patients described all suffered from combined immunodeficiency due to impaired lymphocyte activation resulting in recurrent and often lethal infections. The most apparent non-immune phenotypes observed in STIM1 and ORAI1 deficient patients are congenital myopathy and defects in dental enamel calcification (116-117). Importantly, patients deficient in either STIM1 or ORAI1 function did not exhibit an overt bleeding diathesis. These results correlate well with the largely intact primary hemostasis in Stim1-/- and Orai1-/- BMc mice. The thrombocytopenia observed in all three STIM1 and in one ORAI1 deficient patient was caused by autoantibodies against platelet surface antigens and not by developmental defects (93). Of note, ORAI1 seems to have a non-redundant function in human platelets, as Ca2+ influx after treatment with thapsigargin in platelets of a patient with a loss of function mutation in ORAI1 was profoundly impaired (116, 118). In conclusion, inhibition of STIM1 or Orai1 function may be a suitable approach for antithrombotic therapy in human, however, not for long-term therapy, due to the expected side effect of immunosuppression. However, it may be useful in some clinical settings where a short term inhibition of platelet function is sufficient to achieve the therapeutic goal, e.g. in the treatment of acute ischemic stroke (119). Stroke is the second leading cause of mortality worldwide (3). In most cases, stroke is caused by occlusion of a major or multiple smaller intracerebral arteries leading to focal cerebral ischemia and subsequent neuronal cell death in the center of the ischemic territory. At present, early thrombolysis of vessel occluding thrombi is the only established therapeutic option (120-121). This recanalization of major intracerebral arteries is a prerequisite for salvage of tissue; however, sufficient reflow does not guarantee the prevention of infarct growth, as secondary thrombus formation in the microvasculature at distal sites of the infarct area may occur (122). Platelets have a multifaceted role in ischemic stroke, as they are critically involved in its pathogenesis but are also required for prevention of intracranial bleeding (121). To test the functional relevance of STIM1 and Orai1 mediated SOCE in stroke, Stim1-/- and Orai1-/- BMc mice have been subjected to transient middle cerebral arterial occlusion (tMCAO) (119, 123). In this animal model of ischemic stroke, a thread is advanced through the carotid artery into the middle cerebral artery, reducing blood flow by about 95% (119). After 1 hour, the thread is withdrawn, allowing reperfusion of the ischemic area, thus mimicking the clinical situation with vessel occlusion, followed by thrombolysis and reperfusion. Remarkably, in both Stim1-/- and Orai1-/- BMc mice infarct volumes 24 hours after induction of tMCAO were reduced to ~30% of those observed in control BMc mice (75, 90) (Figure 3). Magnetic resonance imaging, performed in living animals, further showed that this protection was sustained, since no secondary infarct growth was observed during the entire observation of 7 or 5 days after infarct induction, respectively. Moreover, no intracranial hemorrhage was detected in T2-weigthed gradient echo images, demonstrating that the protective effect did not occur at the expense of intracranial bleeding. Importantly, the reduced infarct sizes were of functional relevance, as both, Stim1-/- and Orai1-/- BMc mice reached significantly better scores than control mice in tests assessing the motoric or global neurological function (75, 90). Of note, STIM2 deficient mice have also been shown to be protected in the tMCAO model, but this phenotype could be linked to alterations within the neuronal system and was not due to altered platelet function (124). Taken together, these results revealed STIM1 and Orai1 as essential mediators of ischemic brain infarction and established these two proteins and SOCE as potential targets for the prevention and treatment of ischemic stroke. This is of particular relevance because the utility of current antithrombotic or anticoagulant therapeutics is limited by their inherent effect on primary hemostasis and thus the risk of intracranial bleeding complications. Recent experiments have provided evidence that administration of platelet GPIb inhibiting antibodies after the induction of stroke can be as beneficial as prophylactic treatment, indicating that platelet inhibition can be effective even at later stages of disease progression (119). Thus, a single bolus of a STIM1 and Orai1 inhibitor, applied to a patient with acute ischemic stroke, might be sufficient to improve the clinical outcome, while avoiding side effects on other cell types which would likely arise from long term application (see above). It is important to note, however, that data obtained in mice cannot directly be extrapolated to the human situation as differences in the pathomechanisms of ischemic tissue injury may exist. However, in the case of integrin alphaIIbbeta3 blockade the results obtained in mice (119) were predictive for the effect observed in humans (125). Orai1 seems to be a good target protein for the development of SOCE inhibiting drugs, because it is located in the plasma membrane and therefore easily accessible. STIM1 has also been reported to be located to some extent in the plasma membrane in some cell types, including human platelets (71, 126). However, the suitability of plasma membrane STIM1 as a therapeutical target appears questionable, as a recent study raised the interesting possibility that plasma membrane STIM1 may participate in the inhibition of SOCE (126). Different pharmaceutical companies are currently working on Orai1 inhibitors, mainly with the purpose to develop novel immunosuppressive drugs (127). However, those drugs should also be effective in the inhibition of platelet SOCE given the dominant role of Orai1 in this process (94, 118). 10. CONCLUSION AND OPEN QUESTIONS The identification of STIM and Orai as the key molecules mediating SOCE was a milestone in cell biology, and has important implications for platelet biology. Analysis of genetically modified mice clearly established STIM1 and Orai1 as the main store Ca2+ sensor and SOC channel in platelets, respectively. Considering the well-known importance of Ca2+ signals in platelet physiology and the reported defective T-cell function in the absence of SOCE, it was quite surprising that platelets in the absence of SOCE still developed normally and were able to fulfill their role in primary hemostasis. The activation of SOCE involves receptor stimulation, signaling from the membrane to the ER, store depletion, clustering of STIM1 and finally activation of SOC channels. The time required for this process, may be too long for platelets as they have to react very rapidly to activating stimuli. Thus, store release of Ca2+ appears to be sufficient to trigger the main responses of initial platelet activation, like integrin activation and granule release. In contrast, SOCE seems to be especially important for subsequent events, like growth and stabilization of thrombi (75, 90) and coagulant activity (79, 95). Primary hemostasis and thrombus formation have long been considered to be intrinsically tied together. STIM1/Orai1-mediated SOCE, however, seems to be of greater relevance for pathological processes than for primary hemostasis, given the protective effect of STIM1- and Orai1 deficiency in different thrombosis models and the absence of spontaneous hemorrhages and excessive bleeding after injury. Thus, selective targeting of pathological thrombus formation without or only mildly affecting the cell�s ability to seal wounds at sites of injury, may be possible. Future developments will have to show if these observations can be transferred to the human system and if novel powerful, yet safe, antithrombotics can be developed. These recent advances, however, did not only answer many questions but also raised new ones. One first question - not only for platelet biology - is how STIM1 regulates store content and which other proteins or signaling pathways are involved in this process. Also, it is not known why defective SOCE impairs GPVI-ITAM coupled receptor signaling while GPCR signaling remains largely intact. Further open questions are: What is the function of STIM2 in platelets? Is Orai1 the only SOC channel in platelets? Is there a pathway that negatively regulates the STIM1-Orai1 interaction? New insights into the regulation of Ca2+ signaling and SOCE will be of pivotal importance to understand the complexity of platelet function. 11. REFERENCES 1. Z. M. Ruggeri: Platelets in atherothrombosis. Nat Med, 8(11), 1227-34 (2002) Abbreviations: ECM: extracellular matrix; ADP: adenosine diphosphate; ATP: adenosine triphosphate; TxA2: thromboxane A2; Ca2+: calcium; (Ca2+)i: intracellular free Ca2+ levels; SOCE: store-operated Ca2+ entry; ER: endoplasmic reticulum; STIM1: stromal interaction molecule 1, vWF: von Willebrand factor; GP: glycoprotein; PS phosphatidylserine; GPCR: G-protein coupled receptor; CLEC-2: C-type lectin- like receptor 2; CalDAG-GEFI: Ca2+ and diacylglycerol regulated guanine nucleotide exchange factor I; PKC: protein kinase C; CIB1: Ca2+- and integrin-binding protein 1; DTS: dense tubular system; PLC: phospholipase C; IP3: inositol-1,4,5-trisphosphate; IP3R: IP3-receptor; PIP2: phosphatidyl-1,4-bisphosphate; DAG: 1,2-diacyl-glycerol, ROCE: receptor operated Ca2+ entry; SERCA: sarcoplasmic endoplasmic reticulum Ca2+-ATPases; PMCA: plasma membrane Ca2+-ATPase; TP: thromboxane-prostanoid; PAR: protease activated receptor; ITAM: immunoreceptor tyrosine-based activation motif; TRP: transient receptor potential l; SCID: severe combined immunodeficiency, tMCAO: transient middle cerebral arterial occlusion Key Words: Platelets, Thrombus Formation, Hemostasis, Review Send correspondence to: Bernhard Nieswandt, Vascular Medicine, University Hospital Wurzburg, Rudolf Virchow Center, DFG Research Center for Experimental Biomedicine, Josef-Schneider-Strabe 2, 97080 Wurzburg, Germany, Tel: 49 931-31-80405, Fax: 49-931-201-61652, E-mail:bernhard.nieswandt@virchow-uni-wuerzburg.de |