The Hemostatic Mechanism

Top Abstract The Hemostatic Mechanism Primary Hemostasis Hemostasis and Thrombosis As... Ecto-ADPase/CD39 (NTPDase-1),... Identification of CD39 As... SolCD39: The Recombinant... Site-Directed Mutagenesis... Thromboregulation by CD39 in... Ecto-Nucleotidase in Cardiac... References

The hemostatic process consists of a series of physiologic and biochemical reactions that culminate in the arrest of bleeding from blood vessels that have been severed or traumatized physically or chemically. Hemostasis is accomplished via the interaction of three biological systems: 1) components of the vasculature per se, including endothelial cells; 2) blood platelets; and 3) plasma proteins comprising the intrinsic and extrinsic coagulation pathways (Marcus, 1999). Qualitative or quantitative deficiencies in any one of these systems result in a defect of hemostasis, coagulation, or both. These abnormalities can lead to a hemorrhagic diathesis that is clinically mild, moderate, or severe.

There is a paradoxical aspect to the high degree of efficiency of the hemostatic process. At pathologically damaged sites in blood vessels, such as necrotic or fissured atherosclerotic plaques, these structures become agonists for undesirable activation of the hemostatic system and the promotion of blood coagulation, culminating in thrombotic fibrin deposition. This may lead to the occurrence of arterial or venous thrombosis at critical sites in the circulation such as coronary, cerebral, or peripheral blood vessels. Thus, we define thrombosis as a pathologically misdirected form of hemostasis. Thrombosis is the major complication of atherosclerotic disease and accounts for 50% of mortality in the United States, Europe, and Japan (Marcus, 1999). Recently, in the United States, 2,500,000 thrombotic episodes, leading to 959,000 deaths were recorded. Thus, it can be calculated that there is one death every 33 s from an occlusive vascular event. In addition, 450,000 patients in the United States are affected by stroke annually.

	Primary Hemostasis

Top Abstract The Hemostatic Mechanism Primary Hemostasis Hemostasis and Thrombosis As... Ecto-ADPase/CD39 (NTPDase-1),... Identification of CD39 As... SolCD39: The Recombinant... Site-Directed Mutagenesis... Thromboregulation by CD39 in... Ecto-Nucleotidase in Cardiac... References

When the continuity of a blood vessel is interrupted, a series of biological and biochemical reactions are evoked that are defined as the primary hemostatic response. The initial events are modulated by exposed blood vessel components, such as microfibrils, basement membrane, and collagen. Concomitantly, platelet adhesion to the subendothelial matrix and consequent platelet activation occurs. At this point in time, proteins of the coagulation system are not directly involved, although tissue factor (a lipoprotein present in cell membranes which, when bound to factor VII, activates the extrinsic coagulation cascade) may play an earlier role than previously appreciated (Konigsberg et al., 2001).

Microscopic studies ex vivo have demonstrated that the vessel wall quickly retracts, and platelets are seen to immediately adhere to the subendothelium. This is especially true of platelet-collagen adhesion. von Willebrand Factor (vWF) from the plasma and from the subendothelial matrix rapidly adsorbs to the site of vascular damage and mediates further platelet adhesion through an interaction with the platelet glycoprotein Ib-IX-V receptor complex (Ruggeri, 1997, 2000; Kunicki and Ruggeri, 2001).

The above events are accompanied by activation of the platelet membrane glycoprotein IIb/IIIa (integrin _IIb3, GPIIb/IIIa). Activated platelets change shape from a disc to a spiny sphere, continue to spread, and further adhere to the damaged vessel surface. These platelets metabolically generate a releasate consisting of a large variety of biologically active compounds, which were originally stored in the granule compartments of the resting platelet. Components of the releasate serve as recruiting agents for platelets arriving at the site of injury from the general circulation. Platelet recruitment is the critical step in the generation of the evolving hemostatic plug. The recruitment process will ultimately promote total occlusion of the severed blood vessel by the platelet mass, which gradually becomes reinforced and consolidated by the development of fibrin deposition (Marcus et al., 2001).

	Hemostasis and Thrombosis As Models of Cell-Cell Interactions

Top Abstract The Hemostatic Mechanism Primary Hemostasis Hemostasis and Thrombosis As... Ecto-ADPase/CD39 (NTPDase-1),... Identification of CD39 As... SolCD39: The Recombinant... Site-Directed Mutagenesis... Thromboregulation by CD39 in... Ecto-Nucleotidase in Cardiac... References

It is now abundantly clear that thrombosis is a multicellular process (Marcus et al., 2001). Light and electron microscopy studies of the morphology of evolving thrombi indicated that erythrocytes, neutrophils, some monocytes, and platelets are all present in close proximity (Fig. 1) (Mustard et al., 1974). Subsequent in vitro, in vivo, and ex vivo studies verified that biochemical cell-cell interactions occur in the evolving and formed thrombus. Metabolic interactions between platelets, neutrophils, erythrocytes, and endothelial cellsactive components in the microenvironment of the thrombusare defined as "transcellular metabolism". Expanding on this concept, we define thromboregulation as a process or group of processes by which circulating hematologic cells and cells of the vessel wall interact to regulate or inhibit thrombus formation. Essentially, all thromboregulatory reactions are biochemical in nature and result in formation of biologically active metabolites that could have only arisen through interactions between heterogeneous cell types in the vasculature (Marcus et al., 1980, 1982, 1988, 1991; Valles et al., 1993, 2002; Serhan and Oliw, 2001). Thromboregulation is accomplished by biological compounds that are either cell-associated or released from the cell into the fluid-phase microenvironment. These compounds are either constitutive or elicited by agonists. For clinical purposes, vascular thromboregulators can also be classified as to whether they are aspirin-sensitive or not (see Table 2 in Marcus et al., 2001).

View larger version (36K): [in this window] [in a new window]   Fig. 1.   Events leading to formation of a hemostatic platelet plug or a thrombus. 1. Platelets adhere to exposed subendothelium as a consequence of vascular injury; they immediately adhere to collagen, basement membrane, and microfibrils. These events occur in the presence of von Willebrand factor and the platelet glycoprotein Ib/IX complex (GPI/IX). 2. The platelet release reaction: collagen, a major agonist induces ADP and serotonin (5-HT) release from platelet dense granules. In addition, the eicosanoid thromboxane A2 (TXA2) is enzymatically synthesized from released platelet arachidonic acid and secreted within seconds. Unmetabolized arachidonic acid is also released from platelets and metabolized by cells in the microenvironment (transcellular metabolism). This is important when cyclooxygenase(s) have been inhibited, e.g., following aspirin ingestion (Marcus et al., 1982). In parallel with adhesion and activation, endothelial cell defense initiates steps to limit thrombus size (thromboregulation). 3. Components of the platelet releasate initiate the recruitment phase of thrombus formation: ADP and TXA2 are important recruiting agents in the platelet releasate. Both 5-HT and TXA2 induce vasoconstriction, which consolidates the evolving thrombus and limits the velocity of blood flow proximal to the thrombus. Released ADP and TXA2 activate additional platelets arriving in the fluid phase. These activated platelets undergo shape change and aggregate upon the initial layer of adherent platelets. Phospholipoproteins on the activated platelet surface become available for catalytic interaction with proteins of the intrinsic coagulation system to promote thrombin formation. 4. Thrombin formation is the last stage of primary hemostasis. Thrombin is a powerful agonist for further platelet activation, release, and recruitment. By catalyzing fibrin formation, thrombin also initiates consolidation of the platelet plug via interlacing strands of fibrin ("viscous metamorphosis"). 5. Fibrin strands intercalate between activated platelets and become covalently cross-linked by factor XIIIa. The consolidated thrombus is virtually impermeable. The multicellular composition of the thrombus now is evident. 6. Polymorphonuclear leukocytes are seen in close contact with platelets. These cells actually serve an antithrombotic function because upon exposure to a platelet releasate, the leukocytes inhibit further platelet activation and recruitment (Valles et al., 1993). 7. Intact, metabolically viable erythrocytes are actually prothrombotic because they respond to an activated platelet releasate with production of biologically active compounds, which recruit additional platelets into an evolving thrombus. Releasates from collagen-stimulated platelets induce GPIIb/IIIa activation and P-selectin exposure. These responses are markedly amplified by releasates from combined platelet-erythrocyte suspensions. Thus, erythrocytes actively contribute to platelet-driven thrombogenesis and microvascular occlusion (Valles et al., 2002). This may explain why anemic patients have a prolonged bleeding time and patients with polycythemia are frequently "prothrombotic". 8. Formation of the platelet thrombus signals initiation of the fibrinolytic process. Thus, tissue plasminogen activator is then released from endothelial cells. Modified from Mustard et al., 1974 with permission.

Thromboregulators also have a chronological profile that describes their mode of action in relation to thrombin formation. For example, the protein C/protein S natural anticoagulant system is operative after thrombin has formed at a site of vascular injury. If the injury is severe, unregulated arterial thrombosis will occur because of the agonistic effect of injured tissue per se in addition to release and subsequent action of tissue factor. In those situations, the injury site escapes thromboregulation and the thrombotic diathesis is uncontrolled. Vascular thromboregulators can be classified into playing a role either early or late in the process of thrombus formation (see Table 3 in Marcus et al., 2001).

	Ecto-ADPase/CD39 (NTPDase-1), the Major Inhibitor of Platelet Activation and Recruitment

Top Abstract The Hemostatic Mechanism Primary Hemostasis Hemostasis and Thrombosis As... Ecto-ADPase/CD39 (NTPDase-1),... Identification of CD39 As... SolCD39: The Recombinant... Site-Directed Mutagenesis... Thromboregulation by CD39 in... Ecto-Nucleotidase in Cardiac... References

Appreciation of the importance of cell-cell interactions in the vasculature as well as transcellular metabolism as critical facets of thrombosis and inflammation is now widely recognized (Marcus et al., 1982; Karim et al., 1996; Marcus, 1999). These phenomena are pertinent with regard to platelets, leukocytes, erythrocytes, and endothelial cells. We currently believe that endothelial cells control platelet reactivity by at least four mechanisms, and it is probable that more will be discovered (see Table 2 in Marcus et al., 2001). First, there is the cell-associated ecto-ADPase/CD39 as well as three reactants in the fluid-phase: the eicosanoids, thromboxane A₂ (TXA₂; platelet-derived) and prostacyclin (PGI₂); and the autacoid, nitric oxide. All three soluble reactants are synthesized and released by activated endothelium and platelets (Ignarro et al., 1988; Broekman et al., 1991; Marcus, 1999).

Originally the role of eicosanoids such as PGI₂ was studied in platelet-endothelial cell interactions. At that time, inhibition of platelet aggregation was demonstrated to occur via generation of PGI₂ synthesized by aspirin-treated endothelial cells utilizing endoperoxides released from activated platelets in proximity (Marcus et al., 1980; Schafer et al., 1984). Subsequently, it was shown that human platelet reactivity could be inhibited by nitric oxide released from human umbilical vein endothelial cells in suspension (Moncada et al., 1988; Broekman et al., 1991). In extending these experiments further, we studied endothelial cells wherein nitric oxide production was neutralized by hemoglobin and both platelets and endothelial cells were treated with aspirin, thereby inhibiting all PGI₂ production. The results demonstrated that aspirin-treated, nitric oxide-deficient endothelial cells could still inhibit platelet function by metabolizing ADP released from activated platelets. This had been suggested previously (Heyns et al., 1974; Gordon, 1986). Additional studies led to the proposal that the major molecule responsible for platelet inhibition in the vasculature was the membrane-associated ecto-nucleotidase known as CD39, which is an ATP-diphosphohydrolase now classified as NTPDase-1 (Kaczmarek et al., 1996; Marcus et al., 1997). CD39 metabolizes ATP and ADP to AMP. A diagram depicting the domain structure of ecto-ADPase/CD39 is shown in Fig. 2D.

View larger version (29K): [in this window] [in a new window]   Fig. 2.   Inhibition and reversal of platelet aggregation by solCD39. Platelet-rich plasma from a donor who had ingested aspirin was stimulated and the aggregation response recorded for 4 min. Stimuli were 5 µM ADP (A), 2.5 µg/ml collagen (B), or the thrombin receptor agonist peptide TRAP6 (C). The in vitro platelet responses to these agonists were strongly inhibited by both abciximab (which blocks fibrinogen recognition by its receptor, GPIIb/IIIa), and a soluble form of human CD39 (solCD39), the domain structure of which is depicted compared with full-length (wild-type) CD39 (D). Two transmembrane regions are located near the amino and carboxyl termini; a hydrophobic domain is centrally located. The putative apyrase-conserved region is depicted on the left side as the apyrase domain, adjacent to the N-terminal portion. Cysteine residues are marked as C. The engineered form of soluble CD39, containing a Flag tag and an interleukin-2 (IL-2) secretion leader, lacking the two transmembrane domains is shown below for comparison (Gayle et al., 1998). Portions reprinted from Marcus et al., 2001 with permission.

	Identification of CD39 As the Endothelial Cell Ecto-ADPase

Top Abstract The Hemostatic Mechanism Primary Hemostasis Hemostasis and Thrombosis As... Ecto-ADPase/CD39 (NTPDase-1),... Identification of CD39 As... SolCD39: The Recombinant... Site-Directed Mutagenesis... Thromboregulation by CD39 in... Ecto-Nucleotidase in Cardiac... References

Prior to 1990, the prevailing hypothesis was that inhibition of platelet reactivity by endothelial cells was due to production of eicosanoids and/or nitric oxide by activated endothelial cells. Prevention of TXA₂ formation by activated platelets has led to the extensive use of low-dose aspirin as an antithrombotic agent in prevention and treatment of cardiovascular disorders. This prevents formation of TXA₂, but allows for the generation of PGI₂ by endothelial cells. Additional mechanisms responsible for platelet reactivity still exist, as evidenced by the broad variety of antiplatelet agents in current use. This includes such antithrombotic modalities as thrombolytics, anti-GPIIb/IIIa agents, anticoagulants, and ADP receptor blockers. For this review on inhibition of platelet reactivity, we will discuss purinergic signaling. This approach is highlighted by the recent identification of the P2Y₁₂ receptor (previously identified functionally as the P2Y_ADP, P2Y_AC, or P2T_AC receptor). The platelet P2Y₁₂ receptor and the P2Y₁ receptor are mainly responsible for ADP-mediated platelet responses (Gachet, 2001; Hollopeter et al., 2001). The uniqueness of the P2Y₁₂ receptor is underscored by its restricted expression in brain and platelets. In addition, the receptor is specifically inhibited by the clinically utilized thienopyridines, ticlopidine and clopidogrel.

Ten years before identification of the P2Y₁₂ receptor, the prevailing hypothesis was tested that endothelial cell-derived eicosanoids and nitric oxide were responsible for inhibition of platelet reactivity by these cells. This was approached by incubating aspirin-treated human umbilical vein endothelial cells (HUVEC) with radiolabeled ADP. Formation of PGI₂ (and PGD₂) was thereby inhibited, and in parallel, any nitric oxide generated was blocked by addition of purified oxyhemoglobin to the incubation system. The metabolic fate of the added ADP was determined, and any metabolites generated were measured by means of thin-layer radiochromatography. These experiments demonstrated an accumulation of AMP, which was further metabolized to adenosine by the 5'-nucleotidase present on endothelial cells. This was followed by uptake of the adenosine, and intracellular deamination to inosine and hypoxanthine. Importantly, supernatants from HUVEC incubated with [¹⁴C-ADP] were no longer capable of inducing aggregation in platelet-rich plasma. This meant that the added ADP had been metabolized by an endothelial cell ecto-nucleotidase (Marcus et al., 1991). Subsequently, the molecule responsible for the platelet inhibition described was determined to be CD39 (Kaczmarek et al., 1996; Marcus et al., 1997).

Historically, Handa and Guidotti (1996) purified a soluble apyrase from potato tubers (this had long been used in in vitro studies of platelet reactivity) and cloned its cDNA. Sequence analysis revealed 25% amino acid identity and 48% amino acid homology with human CD39. CD39 had originally been cloned as a cell-surface glycoprotein expressed on activated B cells by Maliszewski et al. (1994). Kansas and associates (1991) had shown that CD39 was present on natural killer cells and subsets of T cells as well as preparations of HUVEC. It is now known that nucleotidases with homology to CD39 and potato apyrase are expressed extensively throughout the animal and vegetable kingdoms, in species such as the garden pea, Caenorhabditis elegans, and Toxoplasma. There are at least four regions within the CD39 molecule that demonstrate extraordinary homology, designated apyrase-conserved regions (ACR) (Handa and Guidotti, 1996).

The identity of the HUVEC ADPase as CD39 was confirmed by several additional experimental observations (Kaczmarek et al., 1996; Marcus et al., 1997). With the use of antibodies to human CD39, all the ADPase activity from preparations of purified HUVEC membranes was immunoprecipitated (Marcus et al., 1997). Confocal microscopy and indirect immunofluorescence studies indicated that CD39 was localized to the cell surface of COS cells, which were transfected with human or murine CD39 cDNA (but not vector alone) (Marcus et al., 1997). CD39 strongly inhibited ADP-induced platelet aggregation (Fig. 2, A-C). Importantly, CD39-transfected COS cells metabolized ADP to AMP within 3 minthe time frame directly correlated with events leading to formation of a hemostatic plug or thrombus in vivo. Furthermore, the time point at which platelet inhibition by CD39-expressing cells became evident was also within 3 min following addition of ADP (Figs. 1 and 2). Thus, the time course for platelet inhibition by cells expressing CD39 correlates with their respective ADPase activities as measured biochemically. This suggests that CD39 represents the culmination of an evolutionary process directed toward metabolizing prothrombotic platelet-released nucleotides by an endothelial cell surface enzyme. CD39 thus controls excessive platelet recruitment and accumulation, and thereby maintains blood fluidity.

	SolCD39: The Recombinant Soluble Form of CD39/Ecto-ADPase

Top Abstract The Hemostatic Mechanism Primary Hemostasis Hemostasis and Thrombosis As... Ecto-ADPase/CD39 (NTPDase-1),... Identification of CD39 As... SolCD39: The Recombinant... Site-Directed Mutagenesis... Thromboregulation by CD39 in... Ecto-Nucleotidase in Cardiac... References

Our studies of the biochemical and biological properties of CD39 led us to realize that it could represent a novel therapeutic strategy for blockade of platelet reactivity in platelet-driven occlusive vascular diseases such as stroke (Pinsky et al., 2002). This premise was reinforced by further research on the mechanism of action of CD39 as an aspirin-independent control system that blocks platelet reactivity even in the setting of inhibited eicosanoid and nitric oxide production.

It is critically important to comprehend that the action of CD39 is not on the platelet per se; rather, CD39 metabolizes the ADP component of the releasate generated by activated platelets. This serves to abolish further platelet recruitment, in the absence of any direct effect on the platelet itself. Thus, we hypothesized that a soluble form of human CD39 (solCD39) would constitute an entirely new systemic antithrombotic modality for treatment of thrombosis-prone patients whose platelets have a low threshold for activation. Subsequent experiments in porcine and murine models, including a murine model of stroke (Pinsky et al., 2002) indicated that solCD39 did indeed efficiently inhibit platelet reactivity in the setting of experimental acute stroke.

The design of solCD39 was based on the structure of the full-length molecule (Fig. 2D), containing two transmembrane regions near the amino and carboxyl termini, respectively. These domains anchor the native protein in the cell membrane. Modeling studies, antibody epitope analyses, and sequence homology had demonstrated that the portion of the molecule between the transmembrane regions is external to the cell (Maliszewski et al., 1994; Marcus et al., 1997). This extracellular region contains the ACR characteristic of members of the E-NTPDase family. The external, luminal portion of CD39 is critical for its ecto-ADPase activity, both functionally and biochemically. The ability of CD39-expressing cells to metabolize extracellular nucleotides further supports the concept of extracellular localization of the enzymatic portion of the molecule. The fact that intracellular nucleotide concentrations are in the millimolar range suggests as well that the active site of CD39 does not interface with the cytoplasmic portion.

For generation of solCD39, the extracellular domain, encoding 439 amino acids, was isolated using oligonucleotide cassettes and polymerase chain reaction, and inserted into a mammalian expression vector. To insure secretion of the recombinant molecule an N-terminal interleukin-2 leader sequence was added. COS cells transfected with this solCD39-encoding plasmid generated ATPase and ADPase activity in the conditioned medium, increasing linearly for at least 5 days. solCD39 was isolated from this conditioned medium via immunoaffinity chromatography using a CD39 monoclonal antibody. These procedures resulted in a single ~66-kDa protein with both ATPase and ADPase activities. This indicated that the molecule was properly glycosylated by the COS cells. Upon removal of N-linked oligosaccharides by treatment with N-glycanase, SDS-polyacrylamide gel electrophoresis analysis yielded a protein band with the predicted molecular mass of 52 kDa (Gayle et al., 1998). The purified solCD39 was examined for its effects on platelet aggregation in vitro. solCD39 inhibited ADP, collagen, as well as thrombin receptor agonist peptide (TRAP₆)-induced platelet reactivity (Fig. 2). The inhibition noted with collagen and TRAP₆ indicated that these two agonists exert their effects via released ADP for aggregation and recruitment to a larger extent than previously appreciated.

	Site-Directed Mutagenesis Studies of Amino Acids in the ACR of SolCD39

Top Abstract The Hemostatic Mechanism Primary Hemostasis Hemostasis and Thrombosis As... Ecto-ADPase/CD39 (NTPDase-1),... Identification of CD39 As... SolCD39: The Recombinant... Site-Directed Mutagenesis... Thromboregulation by CD39 in... Ecto-Nucleotidase in Cardiac... References

To develop specific information concerning amino acid residues essential for enzyme catalysis, alanine scanning mutagenesis studies were performed, focusing on apyrase-conserved regions 1 through 4 of recombinant human solCD39. These experiments identified several key amino acids that are significant for CD39 function. Mutation of tyrosine 127 to alanine (Y127A) reduced both ADPase and ATPase activity by ~60%, whereas substitution of serine 57 with alanine (S57A) resulted in a 2-fold increase in ADPase enzymatic activity with no change in ATPase activity. Moreover, mutation of glutamate 174 to alanine (E174A) and serine 218 to alanine (S218A) resulted in total and ~90% loss of solCD39 enzymatic activity, respectively (Drosopoulos et al., 2000). In addition, kinetic analyses of aspartic acid mutants D54A and D213A demonstrated an increase in their catalytic rate, reflecting their increased enzymatic activity compared with wild-type. Kinetic analyses also revealed decreased affinity of D54A and D213A for the cofactor calcium, as well as for the substrates ADP and ATP (Drosopoulos, 2002).

Importantly, enzymatic activity of solCD39 mutants correlated with their biological activity in tests of in vitro platelet aggregation. In citrate-anticoagulated platelet-rich plasma (PRP), each mutant reversed platelet aggregation at a level that paralleled its relative ADPase activity, with the exception of D54A and D213A (Drosopoulos et al., 2000; Drosopoulos, 2002). For example, E174A, devoid of enzyme activity, failed to inhibit platelet aggregation, and S218A, with 91% loss of ADPase activity, was much less effective than wild-type in reversing platelet aggregation (Drosopoulos et al., 2000). However, D54A and D213A exhibited a decreased ability to inhibit platelet aggregation in citrated PRP, although they displayed increased enzymatic activity in citrate-free buffer systems. Upon further examination, this decrease induced by D54A and D213A was attributable to a reduction in available free calcium due to its chelation by the citrate anticoagulant used to prepare the PRP. In heparinized PRP, D54A and D213A completely reversed platelet aggregation (Drosopoulos, 2002). Thus, glutamate 174 and serine 218 are essential for both the enzymatic and biological activity of solCD39, and aspartates 54 and 213 are involved in calcium utilization (Drosopoulos et al., 2000; Drosopoulos, 2002). A distinct property of solCD39, displayed in all experiments, is the absolute correlation between its enzymatic and biological activity.

	Thromboregulation by CD39 in the Ischemic Brain

Top Abstract The Hemostatic Mechanism Primary Hemostasis Hemostasis and Thrombosis As... Ecto-ADPase/CD39 (NTPDase-1),... Identification of CD39 As... SolCD39: The Recombinant... Site-Directed Mutagenesis... Thromboregulation by CD39 in... Ecto-Nucleotidase in Cardiac... References

Given that stroke is the third leading cause of death and the principal cause of permanent morbidity in the United States, this disorder represents a major public health problem. Thus, 450,000 patients are affected by stroke each year. A murine model of ischemic stroke has been used to demonstrate a critical role for platelets in the progressive microvascular thrombosis that occurs distal to an obstruction of a major tributary in the cerebral vasculature. This microvascular thrombotic occlusion is characterized by distal accumulation of platelets and fibrin (Choudhri et al., 1998). This results in a lack of reflow (postischemic hypoperfusion and consequent neuronal injury). We observed that CD39 has the capacity to inhibit platelet function in an acute stroke and to reduce intravascular thrombosis in the absence of an increased risk of intracerebral hemorrhage that characterizes the therapeutic agents thus far used in the treatment of stroke (Pinsky et al., 2002).

Production of endogenous CD39 was augmented with solCD39 to demonstrate that it could inhibit ADP-mediated auto-amplification of platelet recruitment in distal microvessels and thereby reduce the thrombotic diathesis that follows stroke. solCD39 conferred cerebral protection in stroke without inducing intracerebral hemorrhage. In addition, CD39-null mice were generated by a gene-targeting vector in which exons 4 to 6, encoding apyrase-conserved regions 2 to 4 were replaced with a PGKneo cassette. Compared with wild-type mice, CD39-null mice exhibited diminished blood flow following reperfusion after being subjected to focal cerebral ischemia. When solCD39 was administered to CD39-null mice, they were "reconstituted" as demonstrated by increased postischemic blood flow. The large infarcts that were induced in CD39-null mice were actually reduced by reconstitution with solCD39 and became similar to the infarcts in the solCD39-treated animals with respect to infarct volume and intracerebral hemorrhage. These results suggest a possible new approach to antithrombotic therapy for stroke, based on metabolism of the major platelet agonist for vascular occlusion, platelet-released ADP (Pinsky et al., 2002).

Interestingly, the findings with our CD39^/ mice (Pinsky et al., 2002) differed from those reported with a different CD39-null mouse that, paradoxically, exhibited both thrombosis and hemorrhage (Enjyoji et al., 1999). To create their CD39 knockout mouse, Enjyoji and colleagues eliminated the CD39 translation start site and a portion of the 5'-untranslated region, while our knockout strategy targeted only the enzymatically active extracellular domain, exons 4 to 6 of CD39 containing ACR 2 to 4. These CD39-null mice had reduced platelet aggregation and reduced platelet interaction with damaged vasculature, suggesting that in vivo CD39 might maintain platelet responsiveness to ADP, as in null mice the platelet P2Y₁ receptor was apparently desensitized, although the concentration of circulating nucleotides was not significantly different from control. This might be a reflection of the activity of plasma phosphodiesterases, as discussed recently (Birk et al., 2002).

	Ecto-Nucleotidase in Cardiac Sympathetic Nerves

Top Abstract The Hemostatic Mechanism Primary Hemostasis Hemostasis and Thrombosis As... Ecto-ADPase/CD39 (NTPDase-1),... Identification of CD39 As... SolCD39: The Recombinant... Site-Directed Mutagenesis... Thromboregulation by CD39 in... Ecto-Nucleotidase in Cardiac... References

The role of NTPDases in cell physiology is expanding. Since they modulate the ultimate biologic effects of released nucleotides, NTPDases are important for hemostasis and thromboregulation and are also a key element in other aspects of purinergic signal transduction.

In adrenergic nerve cells, ATP and norepinephrine (NE) are stored together in vesicles, and are coreleased during sympathetic neurotransmission (von Kugelgen et al., 1994; Sneddon et al., 1999). Following its release, ATP is metabolized extracellularly to AMP by NTPDases, either directly or via ADP. Subsequently, it is further converted to adenosine by 5'-nucleotidase originating from myocytes and endothelial cells (Zimmermann and Braun, 1999). Following uptake by cells, adenosine is further catabolized to inosine and hypoxanthine.

In addition to its postsynaptic effects, ATP affects adrenergic transmission by acting on purinoceptors at sympathetic nerve endings (Burnstock, 1999). In primary cultures of dissociated rat superior cervical ganglion neurons, the ATP-gated ionotropic purinoceptor P2X (P2XR) enhances exocytosis of NE, whereas the metabotropic G-protein-coupled P2Y receptor (P2YR) may attenuate it (Boehm, 1999).

Recently, we reported (Sesti et al., 2002) that endogenous ATP also acts by autocrine feedback mechanisms in cardiac sympathetic terminals. Thus, ATP facilitates NE release from cardiac sympathetic nerves via a positive feedback mechanism mediated by P2XR and inhibits NE release via a negative feedback mechanism mediated by P2YR. Indeed, inasmuch as ATP is released upon nerve terminal depolarization with K⁺, the P2XR antagonist PPADS markedly inhibited NE release, whereas the P2Y₁R antagonist MRS-2179 potentiated it, as indicated by the shifts in K⁺ concentration-response curves (Fig. 3, A).

View larger version (13K): [in this window] [in a new window]   Fig. 3.   Release of endogenous NE from guinea pig heart synaptosomes by depolarization with K+. A, potentiation by the P2Y1R antagonist MRS-2179 and blockade by the P2XR antagonist PPADS. Synaptosomes were incubated for 5 min with increasing concentrations of K+ in the absence (control) or presence of 30 nM MRS-2179 or 10 µM PPADS (n = 4 and 4). Points represent mean increases in NE release above basal level (±S.E.M.). Basal NE level: 1.12 ± 0.06 pmol/mg of protein (n = 8). B, potentiation by the NTPDase inhibitor ARL-67156 and inhibition by the recombinant, soluble form of human NTPDase-1/CD39 (solCD39). Synaptosomes were incubated for 5 min with increasing concentrations of K+ in the absence (control) or presence of 30 µM ARL-67156 or 3 nM solCD39 (n = 16 and 8). Points represent mean increases in NE release above basal level (±S.E.M.). Basal NE level: 1.17 ± 0.11 pmol/mg of protein (n = 24). When not visible, error bars are included in the symbol. Reprinted from Sesti et al., 2002.

We found that lower concentrations of ATP will only activate P2YR, whereas higher concentrations of ATP will activate P2XR in sympathetic nerve endings isolated from the guinea pig heart (Sesti et al., 2002). This implies that the net sum resulting from combining the facilitatory and inhibitory components of the receptor-mediated action of ATP on NE release depends critically on the concentrations of ATP at the synaptic sites, as modulated by NTPDase activities.

Cardiac sympathetic nerve terminals express a nucleotidase activity that bears general similarity to that of NTPDase-1. It shows strict dependence on divalent cations (calcium) and a general insensitivity to specific inhibitors of Na⁺/K⁺- and P-type ATPases, alkaline phosphatase, and adenosine uptake. Activity is inhibited by sodium azide, the histidine and tyrosine modifier DEPC, and the selective NTPDase inhibitor ARL-67156 (Sesti et al., 2002). Thus, by metabolizing released ATP, NTPDase not only reduces activation of facilitatory P2XR but also favors activation of inhibitory P2YR, thus reducing release of NE from cardiac sympathetic nerve terminals. Indeed, K⁺-induced depolarization of the cardiac sympathetic nerve terminals elicited much more NE release in the presence of the NTPDase inhibitor ARL-67156 than under control conditions (Sesti et al., 2002). Conversely, in the presence of solCD39, the recombinant soluble form of human NTPDase-1/CD39 (Gayle et al., 1998), NE exocytosis was markedly attenuated (Sesti et al., 2002) (Fig. 3, B). It is unlikely that adenosine, formed as the final product of the sequential metabolic action of NTPDase and 5'-nucleotidase, played a role in the marked attenuation in NE exocytosis by an action on presynaptic inhibitory A₁-receptors. Using our thin layer radiochromatographic assay, we found no evidence of adenosine formation from ATP by cardiac synaptosomes in the presence or absence of added solCD39, unless the quantity of synaptosomes in the assay was increased to assure accumulation of sufficient AMP to (apparently) activate 5'-nucleotidase. Furthermore, synaptosomal NE release elicited by either ATP or K⁺-induced depolarization was unaffected by pharmacological blockade of adenosine A₁ receptors (unpublished).

Therefore, the data reveal a novel pathway regulating NE release from cardiac sympathetic nerve terminals. ATP coreleased with NE activates presynaptic P2XR and promotes NE exocytosis. NTPDase, by metabolizing released ATP, effectively decreases the release of NE, thereby playing a key role in the control of adrenergic function. It is also conceivable that 5'-nucleotidase in adjacent endothelial (Marcus et al., 1991) or smooth muscle cells would generate adenosine in the whole heart, which would then act on A₁ purinoceptors contributing to a further decrease in NE release (Imamura et al., 1994, 1996; Seyedi et al., 1997).

Enhanced adrenergic activity and NE release are known causes of clinical cardiac dysfunction, arrhythmias, and sudden cardiac death during myocardial ischemia (Braunwald and Sobel, 1988; Kubler and Strasser, 1994; Benedict et al., 1996). Thus, NTPDase could exert a cardioprotective action by reducing ATP-mediated NE release, thereby offering a novel therapeutic approach to myocardial ischemia and its consequences.

In conclusion, modulation of nucleotide-mediated signaling is important for maintenance of normal physiological processes. For platelets, it is essential that released ADP levels are regulated to prevent excessive platelet recruitment and the prothrombotic phenomena that may result. Results of our recent research have indicated that this paradigm is important for the cerebral and cardiac systems and the circulation. CD39/NTPDase-1 and its recombinant derivative, solCD39, play important roles in the control of purinergic signaling via prevention of excessive accumulation of extracellular adenine nucleotides.

Note Added in Proof. Since submission of this article, Belayev and colleagues demonstrated a neuroprotective effect of solCD39 following transient middle cerebral artery occlusion in Sprague-Dawley rat. solCD39 reduced neurological deficit, infarct size, and extent of edema [Belayev L, Khoutorova L, Deisher TA, Belayev A, Busto R, Zhang Y, Zhao W, and Ginsberg MD (2003) Neuroprotective effect of so1CD39, a novel platelet aggregation inhibitor, on transient middle cerebral artery occlusion in rats. Stroke 34:758-763]. These results are very similar to those reported by us (Pinsky et al., 2002). In addition, we also demonstrated potent inhibition of platelet aggregation ex vivo to ADP and decreased postischemic platelet and fibrin deposition following transient middle cerebral artery occlusion in a murine model (Pinsky et al., 2002).