Cyclic nucleotide‐dependent inhibitory signaling interweaves with activating pathways to determine platelet responses

Abstract Platelets are regulated by extracellular cues that impact on intracellular signaling. The endothelium releases prostacyclin and nitric oxide which stimulate the synthesis of cyclic nucleotides cAMP and cGMP leading to platelet inhibition. Other inhibitory mechanisms involve immunoreceptor tyrosine‐based inhibition motif‐containing receptors, intracellular receptors and receptor desensitization. Inhibitory cyclic nucleotide pathways are traditionally thought to represent a passive background system keeping platelets in a quiescent state. In contrast, cyclic nucleotides are increasingly seen to be dynamically involved in most aspects of platelet regulation. This review focuses on crosstalk between activating and cyclic nucleotide‐mediated inhibitory pathways highlighting emerging new hub structures and signaling mechanisms. In particular, interactions of plasma membrane receptors like P2Y12 and GPIb/IX/V with the cyclic nucleotide system are described. Furthermore, differential regulation of the RGS18 complex, second messengers, protein kinases, and phosphatases are presented, and control over small G‐proteins by guanine‐nucleotide exchange factors and GTPase‐activating proteins are outlined. Possible clinical implications of signaling crosstalk are discussed.

activates guanylate cyclase (sGC, NO-GC) directly to produce cGMP. cGMP activates cGMP-dependent protein kinase (isoform PKG Iβ) which phosphorylates substrate proteins that are often identical to the ones phosphoryated by PKA ( Figure 2). 1,2 Cytosolic levels of the cyclic nucleotides are further controlled by phosphodiesterases (PDE) that degrade cAMP and cGMP. The signaling pathways induced by PGI 2 and NO are thought to keep platelets in a quiescent state in the absence of damage.
In contrast, exposure of the extracellular matrix protein collagen and binding of von Willebrand factor (VWF) after vascular injury initiates platelet activation. Collagen stimulates the immunoreceptor tyrosine-based activation motif (ITAM)-containing glycoprotein VI (GPVI)-Fc receptor γ chain complex which signals via Src family and Syk tyrosine kinases, whereas VWF binds to the GPIb/ IX/V complex. In parallel, vascular damage also triggers the clotting cascade leading to thrombin generation which directly activates platelets via thrombin receptors and converts soluble fibrinogen into a fibrin mesh. Activated platelets release ADP and thromboxane A2 (TxA 2 ) which bind to specific GPCRs. The signaling events downstream of these receptors include the activation of diverse kinases and GPCR-linked heterotrimeric G-proteins as well as small G-proteins of the Ras and Rho families. Furthermore, calcium ions (Ca 2+ ) are released from intracellular stores leading to further signal amplification ultimately inducing membrane and cytoskeletal reorganization, granule release and integrin activation. [3][4][5] In addition to endothelium-dependent PGI 2 and NO pathways, platelet activation is limited by negative feedback through immunoreceptor tyrosine-based inhibition motif (ITIM)-containing receptors G6b-B and PECAM-1, 6 serine/threonine phosphatases, intracellular receptors and mechanisms of receptor desensitization and cleavage. 5 The ectonucleotidase CD39 expressed by endothelial cells hydrolyses ATP and ADP, thus limiting the availability of platelet agonists. The inhibitory functions of these molecules are often deduced from the phenotypes of the corresponding knockout mouse models (ie, hyper-reactive platelets).
Cyclic nucleotide-mediated inhibition and platelet activation are commonly described as separate systems. However, increasing evidence suggests an alternative model whereby simultaneous regulation of both systems determines platelet function. For example, to prevent platelet activation, inhibitory signaling targets key nodes in activating pathways. 7 On the other hand, platelet activation requires active blockage of endothelium-dependent inhibitory pathways, [8][9][10][11] including ADP-mediated inhibition of cAMP production via P2Y12 and thrombin-mediated cAMP degradation via PDE type 3A.
Interestingly, both P2Y12 and PDE3A are targets of multiple clinically used drugs which reduce thrombosis. 12 Once initiated, cyclic nucleotides are also able to reverse platelet aggregation leading to complete disaggregation of already formed platelet aggregates. [13][14][15][16] Given the powerful activity of endothelium-derived inhibitors combined with the relative abundance of endothelial cells compared to platelets, 17 cyclic nucleotide pathways are likely to contribute more significantly to the control of platelet activation than is usually appreciated. Recent comprehensive phosphoproteome studies indicate an increasing complexity in signaling events following binding F I G U R E 1 Prostacyclin signaling. Prostacyclin (PGI 2 ) is released by healthy endothelial cells. The prostacyclin receptor on platelets couples to the stimulatory heterotrimeric Gs protein complex. Prostacylin binding leads to a conformational change in the receptor activating its guanine-nucleotide exchange factor (GEF) activity towards Gαs resulting in the exchange of GDP by GTP. Gαs-GTP binds to and activates the transmembrane protein adenylate cyclase (AC) to synthesize cAMP from ATP. The second messenger cAMP has only one major target in platelets which is the cAMP-dependent protein kinase (PKA) family. cAMP binding to the regulatory subunits of PKA leads to activation of the catalytic subunits and to the phosphorylation of numerous substrate proteins resulting in a profound inhibition of most platelet functions  18,19 In this paper, we focus on the interdependence of cyclic nucleotide-dependent inhibitory and platelet activation pathways.
To modulate each other's actions, these pathways constantly interact at multiple levels. We will describe these interactions highlighting recently identified protein-protein assemblies involved in platelet control like the RGS18 complex. The potential clinical relevance of crosstalk between endothelium-dependent inhibition and platelet activation will also be considered.

| Gi signaling
The plasma membrane contains multiple receptor and regulatory proteins required for the initiation of signaling events leading to platelet inhibition and/or activation. Major receptors for platelet activators like ADP, epinephrine, and prostaglandin E2 (PGE2) work, at least in part, by inhibiting cAMP production, and thus by interfering with inhibitory pathways. Binding of ADP to the P2Y12 receptor causes a conformational change in the receptor allowing it to act as a guanine-nucleotide exchange factor (GEF) and activate the membrane-associated heterotrimeric G-protein of the Gi-family. The active GTP-bound Gαi-subunit dissociates from the β and γ subunits and binds to AC leading to diminished cAMP synthesis ( Figure 3 and Table 1). Simultaneously Gi triggers the phosphatidylinositol 3-kinase (PI3K) pathway, which converts phosphatidylinositol 4,5-bisphosphate to phosphatidylinositol 3,4,5-trisphosphate at the plasma membrane leading to the activation of the kinase Akt and inhibition of the Rap1-GAP RASA3, key activating pathways in platelets. [20][21][22] The synergism between P2Y12 function and blockage of cyclic nucleotide pathways has been shown for both, cAMP and cGMP components, although cGMP regulation by P2Y12 is not well understood. 9,23 These studies indicate that P2Y12-mediated inhibition of cyclic nucleotide signaling is mandatory to achieve full platelet activation. Accordingly, knockout of the P2Y12 receptor in mice, congenital deficiency of the receptor in humans, or expression of constitutively active Gαi2 in a mouse model confirm the importance of P2Y12 signaling for platelet aggregation and thrombus formation. 10,24,25 In thrombosis studies, the establishment of a tight vascular seal has been suggested to depend on P2Y12 function, 26 and thus on cross-inhibition of cAMP/cGMP pathways. P2Y12 and Gi signaling also contribute to the formation of the outer shell structure of hemostatic thrombi. 27 This type of crosstalk is not restricted to the P2Y12 receptor alone. The α2A-adrenergic and EP3 receptors for epinephrine and PGE2, respectively, are also coupled to Gi-family proteins. PGE2 can be released from activated platelets providing positive feedback on platelet activation at least in part through inhibition of cAMP synthesis. 28 Mathematical modelling suggests that the EP3 receptor might be even more potent than the P2Y12 receptor in inhibiting cAMP production. 29 Another example of a Gi-coupled receptor in platelets is the chemokine receptor CXCR4. 30,31 The importance of costimulatory Gi signaling was recently highlighted in a study of matrix metalloproteinase 2 which stimulates protease-activated receptor PAR1 without being able to induce platelet aggregation in itself because of a lack of concomitant Gi activation. 32 and increases binding of 14-3-3γ to RGS18. 7

14-3-3 bound RGS18
is less active leading to enhanced Gq signaling and Ca 2+ -release ( Figure 5, RGS18 complex inactive). Loss of SHP-1 during platelet activation leads to binding of the serine/threonine phosphatase PP1 to spinophilin which is associated with reduced PP1 activity towards myosin light chain (MLC) potentially contributing to enhanced MLC phosphorylation and platelet activation. 36 Cyclic nucleotide inhibitory pathways also disrupt the RGS18/spinophilin/SHP-1 complex, 37 however, they inhibit binding of 14-3-3γ to RGS18 ( Figure 5, RGS18 complex active). 7 This free RGS18 binds more effectively to Gαi2 and attenuates Gq signaling leading to reduced Ca 2+ -release and thus contributing to platelet inhibition. 7,37 The reorganization of RGS18 complexes is mediated by multiple de-/phosphorylation events. Thrombin and TxA 2 induce SHP-1 Y536 phosphorylation leading to phosphatase activation and de-phosphorylation of tyrosines on spinophilin thus contributing to dissolution of the spinophilin/SHP-1/RGS18 complex. 37 The role of SHP-1 phosphatase activity in inducing complex dissociation is supported by pharmacological studies using a nonselective SHP-1/SHP-2 inhibitor (NSC-87877). 35 Thrombin, TxA 2 , and ADP induce phosphorylation of RGS18 on S49 resulting in enhanced 14-3-3 binding to RGS18. 38 In contrast, PKA phosphorylates S216 of RGS18 leading to de-phosphorylation of S49 and S218, presumably by PP1 ( Figure 5, RGS18 complex in transition a), and to detachment of 14-3-3 from RGS18. 38 The PKA-induced release of RGS18 from spinophilin is supported by phosphorylation of spinophilin on S94. 37 Interestingly, spinophilin knockout platelets exhibit reduced Ca 2+ -release and aggregation which could, at least in part, be due to dysregulated RGS18 function. 35 Absence of spinophilin also results in reduced cAMP synthesis in response to PGI 2 , highlighting multiple functions of this scaffold protein in platelets. Studies on ulating GPVI and integrin αIIbß3 signaling but did not reveal any defect in GPCR signaling. 39 Because SHP-1 is ablated during megakaryocyte differentiation in these mice, it is possible that the role of SHP-1 in regulating GPCR signaling is not apparent because of compensatory mechanisms and/or redundancy with the closely related SHP-2 phosphatase. Indeed, platelets deficient in both SHP-1 and SHP-2, display marked defects in PAR4 receptor signaling, suggesting a level of redundancy. 39 Whether a spinophilin/ RGS18 complex exists in SHP-1 deficient platelets remains to be determined.

| GPIb/IX/V complex
The GPIb/IX/V complex is a highly and specifically expressed VWF PKA might impact indirectly on the GPIbα/filamin interaction by phosphorylating filamin A leading to its protection against proteolysis. 1 A key site in GPIbα that is phosphorylated in resting platelets and dephosphorylated during activation by an unknown kinase and phosphatase, respectively, is S609. 43 GPIbα S609 phosphorylation is required for 14-3-3 binding and receptor signaling. 44 The GPIbβ subunit can be phosphorylated by PKA on S166 possibly contributing to binding of 14-3-3 to the GPIb/IX/V complex. 1 S166 phosphorylation has been suggested to depend on localization of PKA to lipid rafts possibly via the A-kinase anchoring protein moesin. 45 The functional role of this phosphorylation is uncertain, as mutations of S166 have resulted in enhanced, as well as reduced receptor function, 1 and the role of phosphorylated S166 for 14-3-3 binding has been questioned. 40 PKA has been described to inhibit shedding of extracellular regions of GPIbα. 46 Similarly, PKA inhibits shedding of the semaphorin family member Sema4D, a receptor that contributes to thrombus growth and stability. 47 These shedding processes are mediated by the metalloproteinase ADAM17, however, the mechanisms involved in PKAinduced inhibition of ADAM17 are not known. These data indicate a possible new role of PKA in the long-term stabilization of certain membrane receptors which could contribute to thrombus growth.

| NO and sGC
The classical pathway for the activation of cGMP production involves endothelium-derived NO stimulating platelet sGC. sGC represents the main receptor for endothelial NO in platelets and NO-binding stimulates cGMP production up to 100-fold. 50 57 Similarly, platelets from patients lacking sGC expression do not aggregate well in response to low ADP concentrations, and platelet adhesion to collagen and VWF is attenuated. 59 However, reduced sGC expression has also been linked to an increased risk of myocardial infarction possibly due to enhanced platelet activation. 60 Interestingly, low NO concentrations have been seen to slightly potentiate thrombin-induced Ca 2+ -elevations thus potentially contributing to platelet activation. 61,62 Platelet activators like thrombospondin-1 (TSP1) have been shown to cross-inhibit sGC through some unknown mechanism. 63 Further studies are required to clarify interactions between different receptors, NO, sGC, cGMP, and Ca 2+ F I G U R E 5 Activation/inactivation cycle of RGS18. Regulator of G-protein signaling 18 (RGS18) terminates Gq signaling and is regulated by both platelet activating and inhibitory pathways. In freshly isolated resting platelets RGS18 is found in a complex with the adapter protein spinophilin, the tyrosine phosphatase SHP-1 and the phospho-serine/threonine binding protein 14-3-3γ attached to phosphorylated serine 218 of RGS18 (transition state on the way towards inactivation, transition i). Platelet activators like thrombin and TXA 2 induce the phosphorylation of serine 49 of RGS18 generating a second 14-3-3 binding site leading to enhanced 14-3-3 binding. Simultaneously, SHP-1 is activated and detaches from spinophilin (involving de-phosphorylation of tyrosine residues on spinophilin), the serine/threonine phosphatase PP1 binds to spinophilin instead, and spinophilin dissociates from RGS18. In this state the RGS18 complex is inactive. Platelet inhibitors like prostacylin and nitric oxide induce the phosphorylation of serine 216 on RGS18 and serine 94 on spinophilin which lead to the activation of PP 1 and detachment of spinophilin from RGS18 (transition state on the way towards activation, transition a). Active PP 1 removes phosphate groups from serines 49 and 218 of RGS18 leading to the loss of 14-3-3 binding. This state of the RGS18 complex is characterized by free catalytically active RGS18 which can hydrolyse Gαq-GTP to form inactive Gαq-GDP (Figure 4). Phosphorylation sites linked to platelet inhibition are highlighted in red, whereas sites linked to platelet activation are marked in green

| Calcium
Gq-coupled receptor and GPVI signaling trigger phospholipase C activation, leading to an increase in cytosolic inositol-3-phosphate (IP 3 ) levels. IP 3 binds to Ca 2+ channels on the dense tubular system, called IP 3 -receptors, resulting in Ca 2+ -release into the cytosol followed by store-operated calcium entry through the plasma membrane ultimately contributing to platelet activation. 66 Steady state, ADP-and thrombin-induced elevation of cytosolic Ca 2+ levels have been modelled resulting in the validation of many static and kinetic parameters of the enzymes and ion channels involved. 67,68 The well established antagonism between activators and endothelial inhibitors in the regulation of Ca 2+ levels has recently been confirmed in a systematic study comparing effects of ADP, thrombin, convulxin and U46619 in combination with activators of cAMP and cGMP pathways. 61 The Gq GAP RGS18 contributes to differential control of Ca 2+ levels by platelet activators and inhibitors. 7,35 PKA also inhibits IP 3 -induced Ca 2+ -release through phosphorylation of the IP 3 -receptor and the IP 3 -receptor-associated G-kinase substrate (IRAG, MRVI1). 1 In the absence of IRAG cGMP is less potent in inhibiting platelet aggregation and dense granule release. 69 The IP 3 -receptor/PKG complex includes PDE type 5 which might support spatial control of cGMP signaling. 1

| Cyclic nucleotides
In addition to the cyclic nucleotide-synthesizing enzymes AC and sGC, PDEs control the spatial and temporal distribution of cAMP and cGMP. Platelets contain three PDEs, two of which degrade cAMP (PDE2A and PDE3A) and one is cGMP-specific (PDE5A). 1 PDE3A inhibition by cilostazol or JF959602 leads to increased cAMP levels and inhibits the initial accumulation of platelets at sites of injury as well as attachment and detachment of platelets from growing thrombi in vivo. 70 Thrombin stimulates the enzymatic function of PDE3A resulting in reduced cAMP levels. 11  has been suggested to play a role as gate-keeper for binding of 14-3-3 to PDE3A (Figure 6). In contrast, the cAMP/PKA pathway triggers the phosphorylation of S312 alone which is not linked to 14-3-3 binding but which also stimulates the enzymatic function of PDE3A. These data indicate that activating pathways rapidly phosphorylate and stimulate PDE3A to reduce cAMP levels below an inhibitory threshold whereas inhibitory PKA pathways may use PDE3A in a negative feedback loop. 14-3-3 binding to PDE3A might help to discriminate between activating and inhibitory pathways. 71 Modelling of cyclic nucleotide levels confirms that PDEs are highly regulated exhibiting very low basal catalytic activity. 72 These studies indicate that cAMP does not influence total cGMP levels. In contrast, cGMP might stimulate PKA function through an inhibition of PDE3A. 73 In fact, a synergistic inhibitory role for NO/cGMP in the cAMP/PKA pathway has been proposed. 74

| Kinases
Opposite regulation of MLC kinase and MLC phosphorylation levels by activating and cyclic nucleotide-mediated inhibitory pathways has been known for a long time, 76 and this regulation is likely to contribute to the control of actin dynamics, platelet adhesion, and aggregation. 77 Extracellular signal-regulated kinase (ERK) and p38 mitogen-activated protein kinase (MAPK) activation induced by VWF, thrombin, collagen and TxA 2 are generally thought to be F I G U R E 6 PDE3A signaling. Upon binding of active Gαs-GTP adenylate cyclase (AC) is stimulated to synthesize cAMP from ATP. cAMP is required for the activation of cAMP-dependent protein kinase (PKA). cAMP levels can be reduced by phosphodiesterase PDE3A which specifically degrades cAMP to AMP. PDE3A is regulated by multiple phosphorylation events. For example, platelet activation leads to activation of protein kinase C isoforms (PKC, Figure 4) which phosphorylate PDE3A on serine 428 leading to attachment of the phospho-serine/threonine binding protein 14-3-3 and to stimulation of PDE3A function. In this way platelet activation interferes with inhibitory cAMP pathways signaling has been suggested to be involved in regulating thrombus size. 27 Thus, PKA might be activated during thrombus formation by an endothelium-independent mechanism to limit thrombus growth.
Platelet activation by oxidized low-density lipoproteins has been suggested to involve Src kinases, PKC, p47 phox phosphorylation, NADPH oxidase 2, and the production of reactive oxygen species ultimately leading to an inhibition of PKG. 81 This mechanism might be similar to TSP1-induced inhibition of PKA which has also been suggested to involve Src family kinases. 82 Src family kinases are crucial for initiating and propagating signals from major platelet recepors, 83 supporting the concept that platelet activation requires blockage of cyclic nucleotide signaling. Recent mass spectrometry based proteome studies indicate that many kinases are phosphorylated upon PGI 2 and/or ADP treatment of platelets probably leading to complex inhibition/activation patterns. 18,19,84

| Phosphatases
Proteome studies document de-phosphorylation of many sites confirming an important role for phosphatases in platelet control. 18,85 Certain kinases are able to activate phosphatases, as shown for thrombin-induced phosphorylation and activation of SHP-1 leading to de-phosphorylation of spinophilin, 35 and for PKA-mediated activation of PP1 leading to detachment of 14-3-3 from RGS18. 38 Although tyrosine phosphatases are important during platelet activation, 86 their role in cyclic nucleotide-mediated platelet inhibition remains unclear. Interestingly, of the many phosphorylation sites induced by ADP treatment only half are reversed upon activation of cAMP pathways while the remaining sites affected by ADP are not altered. However, these cAMP-induced changes are sufficient to reverse ADP-induced platelet aggregation. 18

| Rap1
Small G-proteins of the Ras and Rho families like Rap1, Rac1, and RhoA are critical regulators of platelet function. 87 Similarly to heterotrimeric G-proteins, small G-proteins are dependent on specific GEFs and GAPs for activation and inhibition, respectively. 88,89 These GEFs and GAPs have been shown to be controlled by activation pathways as well as by cyclic nucleotide signaling leading to differential changes in small G-protein activity. Rap1 is the most highly expressed small G protein in platelets and is the main activator of the fibrinogen receptor integrin αIIbβ3. CalDAG-GEFI (CD-GEFI, RasGRP2), an important GEF of Rap1, is activated by Gq-coupled receptor signaling leading to increased Rap1-GTP levels and integrin αIIbβ3 activation (Figure 7). PKA inhibits CalDAG-GEFI by phosphorylation of S116, S117, and S587 resulting in lowered Rap1-GTP. 90,91 ADP treatment is also able to reduce S587 phosphorylation thus reversing PKA effects. 18

| Rac1
Rac1 integrates signals from several platelet activation pathways and is involved in platelet lamellipodia formation and thrombus stability.
Similar to Rap1, parallel regulation of the Rac1 regulating proteins ArhGEF6 and ArhGAP17 has been described whereby PKA and PKG simultaneously inhibit ArhGEF6 and stimulate ArhGAP17 leading to attenuated Rac1 function. 94 These effects are mediated by ArhGEF6 S684 phosphorylation and 14-3-3 binding, and by ArhGAP17 S702 phosphorylation resulting in detachment of the actin regulator Cdc42-interacting protein 4, CIP4.

| RhoA
RhoA is activated downstream of G13 and is an essential regulator of platelet shape change. RhoA is also known to be differentially regulated by cyclic nucleotides and platelet activators. 95 RhoA inhibition by PKA/PKG involves RhoA S188 phosphorylation and reduced association of RhoA with ROCK2 and the myosin-binding regulatory subunit of PP1, MYPT1. 77,96

| Arf6
ADP-ribosylation factor 6 (Arf6) belongs to the Arf family of small G-proteins and is involved in receptor trafficking in platelets. 97 In contrast to Rap1, Rac1, and RhoA, Arf6 is inhibited during platelet activation, whereas cyclic nucleotide pathways elevate Arf6-GTP levels. 98 Cytohesin-2 is a GEF for Arf6 which plays a role in the constitutive inhibition of dense granule secrection in resting platelets. 99 Cytohesin-2 is phosphorylated by PKC during platelet activation leading to the dissociation of Arf6 and its GEF 99

| CLINICAL IMPLICATIONS
Inherited disorders affecting genes involved in the positive or negative regulation of platelet cyclic nucleotide levels underscore the importance of this regulatory system to hemostasis and thrombosis in humans. For example, Gs hyperfunction leads to an enhanced tendency for bleeding, 100 enhanced NO signaling correlates with F I G U R E 7 Regulation of Rap1 by GEFs and GAPs. Integrin αIIbß3 is a major platelet integrin required for platelet aggregation. The small G-protein Rap1 is a positive regulator of integrin αIIbß3 activation and cycles between an inactive, GDP-bound and an active, GTP-bound state. Formation of Rap1-GTP and integrin activation are enabled by the calcium ion (Ca 2+ ) dependent guanine-nucleotide exchange factor CalDAG-GEFI, whereas hydrolysis of Rap1-GTP to inactive Rap1-GDP requires the GTPase-activating proteins (GAPs) Rap1GAP2 and RASA3. Rap1GAP2 interacts with the phospho-serine/threonine binding protein 14-3-3 through phosphorylated serine 9 on Rap1GAP2. During platelet activation levels of free intracellular Ca 2+ rise ( Figure 4) leading to the activation of CalDAG-GEFI, increased Rap1-GTP formation and integrin activation. In contrast, cyclic nucleotide-dependent inhibitory pathways on one hand suppress intracellular Ca 2+ levels, thus indirectly inhibit CalDAG-GEFI activation, and on the other hand result in phosphorylation of CalDAG-GEFI on serines 116, 117 and 587 inhibiting its activity. Simultaneously, cyclic nucleotide-dependent inhibitory pathways induce Rap1GAP2 phosphorylation on serine 7 leading to an inhibition of 14-3-3 binding, activation of Rap1GAP2, reduced Rap1-GTP levels and lowered Rap1 activity. In contrast, platelet activation leads to phosphorylation of Rap1GAP2 on serine 9 resulting in enhanced 14-3-3 binding and inhibition of Rap1GAP2. In parallel, αIIbβ3-mediated outside-in signaling leads to phosphatidylinositol 3-kinase (PI3K) activation, which in turn reduces the GAP activity of RASA3. Platelet activators also suppress the phosphorylation of the inhibitory serine 587 site on CalDAG-GEFI.  101 whereas Gs or sGC dysfunction culminates in enhanced susceptibility to thrombosis. 60,102 In particular, loss of sGC expression, through a combination of mutations in the α1 subunit of sGC and in the sGC-interacting chaperone CCTη, was shown to reduce platelet cGMP levels and to increase the risk of myocardial infarction. 60 Recently, a meta-analysis identified PDE3A as a novel stroke risk locus. 103 The clinical relevance of crosstalk between activating and cyclic nucleotide-dependent inhibitory pathways is further highlighted by the success of drugs targeting either the Gi-coupled P2Y12 receptor or PDEs 3A and 5A, 12,104,105 which prevent the blockade of cyclic nucleotide production or degradation, respectively. 106  The state of inhibitory pathways can be monitored by measuring intracellular phosphorylation of the PKA/PKG substrate protein VASP in platelets, which has been used to study the efficiency of P2Y12 inhibition in patients. 109,110 Pharmacological studies of P2Y12 receptor based anti-platelet therapy show that addition of low concentrations of PGI 2 and NO to platelets from volunteers receiving dual anti-platelet therapy with aspirin and the P2Y12 blocker prasugrel strongly enhances platelet inhibition in the presence of TRAP-6 or collagen. 8 This synergy between P2Y12 inhibitors and cAMP/cGMP signaling suggests that in the presence of similar levels of platelet blockade thrombosis risk might critically depend on the availability of PGI 2 and NO and thus on endothelial function. Since endothelial dysfunction is commonly observed in patients with cardiovascular disease it could be beneficial to include endothelial function testing in the assessment of anti-platelet therapy. 8 Combining stimulators of endothelium-dependent cAMP pathways such as the PDE3A inhibitor cilostazol with P2Y12 inhibitors might improve anti-thrombotic protection. 17 Interestingly, cilostazol treatment has been associated with fewer hemorrhagic events compared to aspirin. 104

| CONCLUSIONS
Cyclic nucleotide-mediated inhibitory signaling is emerging to be closely intertwined with activating pathways. On the other hand, effective platelet activation mandates interference with inhibitory pathways at multiple levels, including blocking cyclic nucleotide production and promoting their degradation, and dephosphorylating PKA substrates. Inhibitors and activators tend to regulate the same signaling nodes in opposite directions (Table 1)