Overexpression of the platelet P2X₁ ion channel in transgenic mice generates a novel prothrombotic phenotype

Adenosine triphosphate (ATP) is released as a cotransmitter from the sympathetic nerve endings, endothelium, and activated platelets. It is now established that ATP and other nucleotides act as extracellular signaling molecules.1 The receptors that mediate the action of the adenine nucleotides belong to 2 classes, the G-protein–coupled P2Y receptors and the P2X receptors, a family of ligand-gated ion channels.2 Seven distinct P2X purinergic receptors have been cloned from mammalian species (P2X_1-7) and have been found to be widely expressed in excitable and nonexcitable cells.3 Subunits of these receptors can assemble to form homomeric and heteromeric functional channels. All P2X receptors are cation-selective channels with almost equal permeability to Na⁺ and K⁺ and significant permeability to Ca²⁺.3 The Ca²⁺ permeation through P2X receptors is considered to be an important component of the physiologic and pathophysiologic responses mediated by these receptors in vivo (reviewed by Burnstock4 and North5).

In platelets, ATP and adenosine diphosphate (ADP) are present at high concentrations in the dense granules and are coreleased during platelet activation.6 ADP has long been recognized as an important platelet activator, playing an essential role in enhancing secretion and in amplifying platelet aggregation induced by other agonists. Biologic effects of ADP are mediated by 2 distinct metabotropic receptors, the G_q-protein–coupled P2Y₁receptor and the G_i-protein–coupled P2Y₁₂receptor. The latter is the target for specific antithrombotic drugs (reviewed by Gachet7). Platelets express the P2X₁ member of the P2X family of ligand-gated ion channels,8 which mediates a rapid ATP-induced Ca²⁺ influx.9,10 Because of the fast desensitizing property of the P2X₁ ion channel11 and the lack of specific platelet P2X₁ antagonists and because platelet studies have mainly been performed ex vivo at low extracellular Ca²⁺concentrations in citrated plasma, the function of P2X₁ in platelet activation only recently started to be unraveled, and a physiologic role of ATP in this process is now being considered.

In human platelets, the selective P2X₁ agonists αβ-methylene adenosine 5′-triphosphate (αβ-meATP) and βγ-methylene adenosine 5′-triphosphate (βγ-meATP) were shown to evoke a transient Ca²⁺ increase accompanied by reversible platelet shape change, provided that measures had been taken to avoid the activation and desensitization of P2X₁ by ATP spontaneously released during platelet preparation.12,13Despite the fact that the sole P2X₁ activation cannot cause platelet aggregation, the use of P2X₁ desensitization strategies indicated that this ion channel contributes significantly to human platelet aggregation induced by collagen, a major platelet agonist able to trigger dense granule release.13 Recently, we have reported that this effect depends on P2X₁-mediated activation of the extracellular signal-regulated kinase 2 (ERK2) mitogen-activated protein kinase (MAPK), which enhances platelet secretion initiated by collagen.14 According to our model, platelet stimulation with low doses of collagen rapidly causes minor dense granule release; ATP secreted during this early event activates a Ca²⁺- and a protein kinase C (PKC)–dependent P2X₁-ERK2 signaling cascade needed to complete platelet aggregation by enhancing release from collagen-primed dense granules.14 

To further investigate the physiologic role of the platelet P2X₁ ion channel, we have generated transgenic mice overexpressing human P2X₁ in the megakaryocytic cell lineage. Platelets from these mice displayed a gain of P2X₁functionality accompanied by a mild prothrombotic phenotype. Combining ex vivo and in vivo analyses of platelet function, this mouse model enabled us to demonstrate the involvement of P2X₁-mediated Ca²⁺ influx and the coupled ERK2 activation in platelet responses to collagen. We found that P2X₁ overexpression promotes platelet secretion induced by the thromboxane A₂mimetic U46619 and thereby enhances platelet aggregation. Moreover, P2X₁ overexpression also increases platelet activation and aggregate formation under shear stress. Together, our findings suggest a regulatory role for P2X₁ during in vivo hemostasis and thrombosis.

Polymerase chain reaction (PCR) cloning of the murine GPIIb promoter fragment extending from +23 to −508 relative to the initiation start site (TOPO TA cloning kit; Invitrogen, Carlsbad, CA) was performed using the following primers: sense, 5′-AGGAAGTGGGTAAATGTCCTACTC-3′; antisense, 5′-TCCCAAACGTCCTAAACAGGAATGG-3′. TheXhoI–HindIII promoter fragment was excised from the mGPIIb-PCR2.1–TOPO plasmid and inserted into the pGL3-basic luciferase reporter vector (Promega, Leiden, the Netherlands) digested with XhoI and HindIII. Megakaryocytic human erythroblastic leukemia (HEL) and nonmegakaryocytic HeLa cell lines were transiently transfected with the resultant reporter construct, and megakaryocytic-specific promoter activity was verified as described15; promoter-driven increases of luciferase expression were only measured in HEL cells, as expected. The promoter fragment was excised from the mGPIIb-pGL3 plasmid by digestion withKpnI and BamHI and inserted into theKpnI-BamHI–digested P2X₁-pcDNA3 vector16 in front of the human P2X₁(hP2X₁) cDNA. DNA constructs were verified by sequencing on the automated A.L.F. sequencer (Pharmacia Biotech, Uppsala, Sweden). The 2.2-kilobase (kb)KpnI–DraIII fragment (GPIIb-hP2X₁) was excised and purified for zygote injection. The GPIIb promoter has been successfully used to restrict transgene expression to the megakaryocytic cell lineage of mice.17 

Transgenic mice expressing human P2X₁ were generated by zygote injection into the Friend leukemia virus, strain B (FVB) background according to previously published procedures.18 Transgenic offspring were identified by PCR screening using genomic DNA extracted from tail samples. The following primer pair was used: sense, mGPIIb promoter, 5′-AGGAAGTGGGTAAATGTCCTACTC-3′; antisense, hP2X₁, 5′-TCAGGATGTCCTCATGTTCTCCTGCAGG-3′.

Total RNA was extracted from mouse washed platelets and leukocytes isolated from freshly drawn citrated blood using the High Pure RNA isolation kit (Roche Diagnostics, Brussels, Belgium). During reverse transcription–polymerase chain reaction (RT-PCR), specific amplification of the human P2X₁ cDNA in transgenic mouse samples was accomplished with the following primers: sense, hP2X₁, 5′-GTTCCAGGAGGAGCTGGCCGCCTTCC-3′; antisense, hP2X₁, 5′-GGTCTTCATGTGGGCAGCATTCAC-3′. For the specific amplification of mP2X₁ cDNA, the following primers were used: sense, mP2X₁, 5′-CTGCAGGATGAGCTGTCAGCCTTCTTC-3′; antisense, mP2X₁, 5′-GTAGAGGCATTTCTTCATGTAGGT-3′.

Adenosine 5′-diphosphate (ADP), αβ-meATP, βγ-meATP, apyrase (EC 3.6.1.5, grade 1: mixture of both high and low ATPase/ADPase ratio isoenzymes), and the thromboxane A₂mimetic U46619 were from Sigma (St Louis, MO). ADP, αβ-meATP, and βγ-meATP were purified by high-performance liquid chromatography (HPLC) on an Adsorbosphere HS C18 7-μm, 250 × 4.6-mm column (Alltech, Bad Segeberg, Germany) as described.12Fibrillar collagen (Horm-type 1 collagen) was from Nycomed (Munich, Germany) and thrombin (Dade Thrombin Reagent) was from Dade Behring (Marburg, Germany). The MEK1/2 inhibitor U0126 was purchased from BioMol Research Laboratories (Plymouth Meeting, MA), and D-Phe-Pro-Arg chloromethyl ketone (PPACK) was from Calbiochem (San Diego, CA). OG 488–annexin V was from Nexins Research (Hoeven, the Netherlands). Fura-2 acetoxymethyl ester and Pluronic F-127 came from Molecular Probes (Leiden, the Netherlands). Recombinant saratin was produced in the yeast Hansenula polymorpha as described.19 

Eight- to 12-week-old mice were bled under sodium pentobarbital anesthesia (6 mg/kg) from the retro-orbital plexus. Mouse blood was collected in a saline solution containing either 4 U/mL heparin, 20 μM PPACK, and 0.1 U/mL apyrase or 20 μg/mL hirudin. Platelet-rich plasma (PRP) was obtained by centrifugation at 800g for 30 seconds followed by 5 minutes at 150g. PRP from 3 animals were pooled, and the platelet counts were adjusted to 2.5 × 10⁵ platelets/μL with autologous platelet-poor plasma (PPP). Mouse washed platelets were prepared as previously described,14 using apyrase (1 U/mL) throughout the procedure. Platelets were resuspended in Ca²⁺-free Tyrode buffer containing 0.35% (wt/vol) human or bovine serum albumin and 1 U/mL apyrase, at a density of 2.5 × 10⁵ platelets/μL.

Platelet-rich fractions were immediately fixed overnight at 4°C in 2.5% (wt/vol) glutaraldehyde and 0.1 M phosphate buffer, pH 7.2. After centrifugation at 800g for 10 minutes, a condensed pellet of platelets was formed. After fixation in 1% OsO₄ (wt/vol), 0.1 M phosphate buffer, pH 7.2, and dehydration in a graded series of ethanol, the pellets were embedded in epoxy resin. Ultrathin sections were cut and stained with uranyl acetate and lead citrate before examination with a Zeiss EM 10 electron microscope (Oberkochen, Germany).

Light transmission during mouse platelet aggregation was recorded using apyrase-treated washed platelets in the presence of 2 mM CaCl₂ on an ELVI 840 aggregometer (Elvi Logos, Milan, Italy). Shear-induced platelet aggregations were performed in an annular ring-shaped viscometer generating laminar shear (Ravenfield viscometer; Heywood, Lancashire, United Kingdom) using mouse heparinized PRP. After 3 minutes, platelet samples were collected and fixed in 1% paraformaldehyde; the percentage of platelet aggregation was calculated by comparing single platelet counts before and after shearing. ATP secretion was monitored in hirudinized PRP in parallel with platelet aggregation by adding firefly luciferase and luciferin and comparing the luminescence generated by platelet ATP release or by an ATP standard (Chrono-Lume, Kordia, The Netherlands) as previously described.14 

Western blot detection of the human P2X₁ protein in transgenic mouse platelets (8 × 10⁸ platelets) was performed by using a polyclonal rabbit anti-hP2X₁antibody.16 Detection of ERK1/2 phosphorylation in human or mouse washed platelets (1 × 10⁷ platelets) was accomplished with the PhosphoPlus p44/42 MAP Kinase Antibody kit (New England Biolabs, Hitchin, United Kingdom) according to the instructions of the manufacturer.

Apyrase (2 U/mL)–treated mouse washed platelets (2 × 10⁵ platelets/μL) were loaded with 3.5 μM fura-2 acetoxymethyl ester in the presence of Pluronic F-127 for 15 minutes as described.20 The measurements were performed between 30 and 90 minutes after final platelet resuspension (0.7 × 10⁵ platelets/μL). CaCl₂ (2 mM) was added before the agonist. Fura-2 fluorescence was recorded from 0.2 mL aliquots of platelet suspension stirred at 37°C in an SLM-Aminco spectrofluorimeter (SLM Instruments, Rochester, NY) with excitation wavelengths of 340 and 380 nm and emission of 500 nm. Changes in intracellular Ca²⁺ concentration were monitored using the fura-2 340/380 fluorescence ratio and were calibrated according to the method of Grynkiewicz et al.21 The Ca²⁺signals evoked by αβ-meATP were assessed as the peak amplitude of the intracellular Ca²⁺ rise occurring within a few milliseconds after agonist application and returning to basal levels after 10 seconds.

Adhesion experiments under flow conditions were performed with anticoagulated mouse blood (4 U/mL heparin, 20 μM PPACK), basically as described.22 Whole blood was perfused for 4 minutes over a collagen-coated coverslip through a parallel-plate transparent flow chamber using a pulse-free pump, at a wall-shear rate of 1000 seconds⁻¹. During the perfusion, high-resolution microscopic transmission or fluorescent images were recorded in real-time with a Visitech digital imaging system (Sunderland, United Kingdom). Exposure of phosphatidylserine (PS) was detected by postperfusion with the heparinized rinsing buffer containing OG488-labeled annexin V (1 μg/mL). Phase-contrast and fluorescent images were obtained from at least 10 different collagen-containing microscopic fields that were arbitrarily chosen. When indicated, apyrase (0.1 U/mL) was added during blood sampling; in some experiments, blood was incubated with saratin (10 μg/mL) blocking von Willebrand factor (VWF) binding to collagen19 1 minute before perfusion. Area coverage from phase-contrast images was analyzed off-line using ImagePro software (Media Cybernetics, Silver Spring, MD). Area coverage by platelets stained with OG488-annexin V was determined with Quanticell software (Visitech).

Thromboembolism was induced by injection of a mixture of collagen (0.125 or 0.06 mg/kg) and epinephrine (60 μg/kg) into the jugular veins of anesthetized mice. When indicated, mice received 200 μg/kg U0126 1 minute before the induction of thromboembolism. For bleeding time measurements, mice were anesthetized, and 3 mm of the tail tip was amputated with a scalpel. The tail was then blotted with filter paper every 15 seconds until the paper was no longer blood stained.

Statistical analyses of the data were made using the nonpaired Student t test and the 2-tailed Tukey-Kramer multiple comparisons test. Survival data were analyzed using 2 × 2 contingency tables.

On zygote injection of a constructed GPIIb-hP2X₁transgene, among 19 offspring mice obtained, 4 animals were found to be transgenic by PCR screening. All founders transmitted the transgene in a Mendelian fashion. RT-PCR analyses using primer pairs that selectively amplify hP2X₁ versus mP2X₁mRNAs revealed the presence of hP2X₁transcripts in the platelets of the transgenic (TG) mice (Figure1A). The expression of the endogenous platelet mP2X₁ remained comparable to that of wild-type (WT) platelets (Figure 1A). No transgene expression was found in the leukocytes (Figure 1A). Overexpression of the hP2X₁ protein in platelets was demonstrated by immunoblotting of heterozygous (TG^+/−) and homozygous (TG^+/+) mouse platelet extracts (Figure 1B); similarly, immunohistochemistry of bone marrow sections revealed increased P2X₁ staining in the TG megakaryocyte membranes (not shown). The homozygous mice of 2 founder lines (denominated TG₁₆ and TG₁₈) overexpressing similar amounts of platelet hP2X₁ were characterized and showed identical phenotypes. These mice had no apparent physiologic abnormalities and displayed normal development, survival, and reproduction. Platelet count (Table1) and morphology (not shown), as well as other hematologic parameters, were identical to those of WT mice, with the exception of mild leukocytosis (Table 1). Values are represented as mean ± SD (P = .003).

Apyrase, by degrading ATP spontaneously released during blood sample handling, is needed ex vivo to protect platelet P2X₁channels from artificial desensitization.12 13 In the presence of this ectonucleotidase and physiologic Ca²⁺concentrations (2 mM CaCl₂), the nonhydrolyzable P2X₁ selective agonists αβ-meATP (1 μM, defined to be the optimal concentration) (Figure 2) and βγ-meATP (not shown) evoked a rapid Ca²⁺ influx in mouse platelets. The peak value of this intracellular Ca²⁺rise was increased by approximately 50% in TG platelets compared with WT platelets (TG, 146.5 ± 23.6 nM, n = 8; WT, 93.1 ± 21.4 nM, n = 8; P = .0003) (Figure 2). In the same experimental conditions, the WT and TG platelets displayed identical ADP-induced Ca²⁺ increases, reflecting P2Y₁-mediated Ca²⁺ responses (WT, 399 ± 66 nM, n = 5; TG, 478 ± 74 nM, n = 5; P = .112).

In human platelets, P2X₁ stimulation causes Ca²⁺ influx and subsequent platelet shape change.12 13 Similarly, αβ-meATP and βγ-meATP induced a quickly reversible shape change of WT mouse platelets proportional to Ca²⁺ influx. Platelets from TG mice exhibited a more prominent αβ-meATP–induced platelet shape change compared with WT platelet response (Figure 2B). Taken together, these data are compatible with a significant gain of P2X₁functionality in the TG platelets.

In an aggregometer, the P2X₁ selective agonists αβ-meATP and βγ-meATP did not cause platelet aggregation or secretion either in WT or in TG mice. Functional studies of apyrase-treated platelets from TG mice, where P2X₁desensitization is prevented, yet demonstrated strongly enhanced platelet aggregation evoked by low and intermediate doses of collagen (1 to 2 μg/mL) compared with WT platelets (Figure3A). A similar increase of TG platelet aggregation was observed after platelet stimulation with low concentrations of convulxin, a glycoprotein VI (GPVI)–selective agonist (0.025 to 0.04 μg/mL) (Figure 3A, right panel), indicating that P2X₁ overexpression enhances platelet aggregation mediated by the collagen receptor GPVI. Platelet aggregations induced by higher collagen (4 μg/mL or greater) or convulxin (0.05 μg/mL or greater) concentrations were identical for WT and TG platelets (not shown).

Threshold concentrations of collagen, which only caused shape change of WT platelets, produced full aggregation of TG platelets (Figure 3A). Consistent with an event requiring functional P2X₁channels, the enhanced reactivity to collagen was abrogated when the apyrase treatment was omitted (Figure 3B) or when these platelets were pretreated with αβ-meATP (1 μM) to selectively desensitize P2X₁ (Figure 3B). Blockade of ERK2 activation with the selective MEK1/2 inhibitor, U0126 (1 μM) (Figure 3C, last 2 lanes) abolished collagen-induced platelet aggregation for TG (1 μg/mL) and WT (2 μg/mL) platelets (Figure 3B). Interestingly, full aggregation of TG platelets already evoked by 1 μg/mL collagen coincided with maximal ERK2 phosphorylation (Figure 3C). These data demonstrated that P2X₁ and ERK2 are proximal and distal components of a common signaling cascade contributing to collagen-induced platelet aggregation and that this pathway has been up-regulated in TG platelets as a consequence of P2X₁ overexpression.

Further analyses of TG platelet aggregation revealed an increased response to low concentrations of the thromboxane A₂mimetic U46619 (less than 2 μM) (Figure4A). In contrast, TG platelet aggregation provoked by any concentration of HPLC-purified ADP or of thrombin occurred normally (Figure 4B).

To investigate the mechanism responsible for the observed increased aggregation of TG platelets, we compared platelet secretion induced by low concentrations of collagen or of U46619 in WT and TG PRP. As in human platelets, the desensitization of P2X₁with αβ-meATP strongly inhibited ATP secretion from WT platelets, indicating a P2X₁ contribution to this platelet response (not shown). Figure 5 shows increased ATP secretion of TG platelets (1.9 ± 0.6- and 3.3 ± 0.2-fold increases in response to 0.5 μg/mL collagen and 4 μM U46619, respectively [P < .001; n = 3]) paralleled with enhanced platelet aggregation. Thus, P2X₁ overexpression in TG platelets potentiates platelet-dense granule release initiated by low doses of collagen or of the thromboxane A₂ mimetic U46619, leading to the amplification of the initial platelet response and the completion of aggregation. ATP secretion triggered by ADP or thrombin was found to be normal (not shown).

To further investigate the TG platelet hyperreactivity to collagen, aggregate formation and exposure of coagulation-active negatively charged phosphatidylserine (PS) on a collagen surface were analyzed in whole mouse blood under conditions of flow. Using videomicroscopy, aggregate formation was monitored from phase-contrast images, and PS exposure (procoagulant activity) was imaged after annexin V staining. As previously described,22 when WT mouse blood was perfused over collagen at a shear rate of 1000 seconds⁻¹, platelets tethered, adhered, and assembled on the surface, with the adherent cells responding by a rapid increase in cytosolic Ca²⁺,22 and subsequent surface-exposure of PS (Figure 6). A similar picture was observed following perfusion of TG blood (Figure6). The use of ATP/ADP-degrading apyrase, needed to protect P2X₁ from ATP-provoked desensitization, considerably reduced aggregate formation and platelet PS exposure after the perfusion of WT blood (Figure 6A-B). This is compatible with the reported role of the P2Y₁ and P2Y₁₂ receptors interacting with released ADP during aggregate formation on collagen.23 24 Perfusion of apyrase-treated TG blood yet resulted in prominent platelet aggregation coinciding with high PS exposure (Figure 6A-B), suggesting that P2X₁ overexpression in TG platelets has compensated the inhibitory effects of apyrase by promoting ATP-dependent platelet adhesion, activation, and aggregation under these flow conditions.

Figure 6 also shows how aggregate formation and PS exposure by apyrase-treated TG platelets were abrogated by saratin (Figure 6A-B), which, by blocking VWF binding to collagen, prevents GPIbα-VWF–dependent platelet adhesion,19 suggestive of a role for P2X₁ in the initial control of platelet activation and aggregation on surface-bound VWF at the shear rate investigated.

Interestingly, area coverage by platelet aggregates and PS exposure was reduced by 30% to 35% during the perfusion of apyrase-treated TG blood in the presence of U0126 (Figure 6C-D). These data indicate the involvement of the ERK2 pathway in platelet aggregate formation on collagen under flow. U0126 did not affect aggregate formation or PS exposure when WT blood was perfused in P2X₁ nonprotective conditions (no apyrase) (not shown). These observations thus support the existence of a P2X₁-ERK2 pathway contributing to platelet activation by collagen and surface-bound VWF.

Platelet aggregation induced by shear stress can be measured in a viscometer producing a laminar flow. At high shear stress, VWF binding to GPIbα is essential to induce α_IIbβ₃-dependent platelet aggregation.25 We have investigated whether P2X₁ would contribute to shear-dependent platelet aggregation. For this purpose, platelet aggregations were performed in apyrase-treated heparinized PRP from WT and TG mice, at a shear rate of 9000 seconds⁻¹ corresponding to a shear stress of 124 dyne/cm². At this shear rate, the transgenic platelets underwent potent aggregation (52.4% ± 15.5% of aggregation) compared with negligible aggregation of WT platelets (9.9% ± 6.3%) (Figure 7), showing that the TG platelets exhibited increased ability to respond to shear stress.

Bleeding times, considered to reflect primary hemostasis in vivo, were identical for WT and TG mice (WT, 5.07 ± 2.05 minutes, n = 22; TG, 4.98 ± 2.41 minutes, n = 12). Because in vitro collagen-induced transgenic platelet activation and aggregation were greatly enhanced, we used an in vivo model of pulmonary thromboembolism by intravenous injection of a mixture of a low dose of collagen (0.06 mg/kg body weight) and epinephrine (60 μg/kg). As shown in Figure 8A, 80% of the TG mice died after 4 minutes compared with only a 30% mortality rate for WT mice. These results indicate that P2X₁ overexpression generates a prothrombotic phenotype.

To further examine the importance of the P2X₁-coupled ERK2 signaling pathway for platelet function in vivo, we analyzed the effect of a 1-minute pretreatment with U0126 (200 μg/kg) on the mortality of TG mice. All mice pretreated with U0126 survived the injection of collagen (0.06 mg/kg) plus epinephrine (60 μg/kg) (Figure 8A). Inhibition of ERK2 also conferred antithrombotic protection to WT mice injected with a mixture of a higher dose of collagen (0.125 mg/kg) plus epinephrine (60 μg/kg); the number of survivors reached 90% compared with 40% in controls (Figure8B).

We have generated transgenic mice overexpressing the human ATP-gated P2X₁ ion channel in the megakaryocytic cell lineage. Platelets from these mice displayed a gain of P2X₁ionotropic activity, as shown by more prominent Ca²⁺ influx and platelet shape change triggered by the 2 P2X₁-selective agonists, αβ-meATP and βγ-meATP. Transgenic platelets underwent enhanced aggregation in response to low doses of collagen or of the thromboxane A₂ mimetic U46619. We showed that the increased collagen- and U46619-induced platelet aggregations resulted from enhanced platelet-dense granule release. P2X₁overexpression also increased platelet reactivity to shear stress. Transgenic mice manifested increased in vivo thrombosis after injections of a mixture of low-dose collagen and epinephrine. Thus, the overexpression of P2X₁ in mouse platelets has generated a novel, mild, prothrombotic phenotype.

Enhanced secretion and aggregation of the transgenic platelets in response to collagen corroborate our previous findings in healthy human platelets.13,14 We have reported that low concentrations of collagen cause early minor ATP release that elicits a rapid P2X₁-mediated Ca²⁺ influx, contributing to enhancement of the platelet release reaction, thus completing platelet aggregation. It is therefore likely that the enhanced TG platelet response to collagen results from an increased number of functional ATP-responsive P2X₁ channels expressed at the surfaces of TG platelets. The higher level of P2X₁ activation thereby leads to enhanced platelet secretion and aggregation. Similarly, the increased TG platelet aggregation induced by low concentrations of the thromboxane A₂ mimetic U46619 was related to enhanced platelet secretion induced by this agonist. This is in agreement with earlier studies on the importance of secreted products, mainly ADP, during U46619-induced platelet aggregation.26 Collagen activates platelets by transducing signals through glycoprotein VI (GPVI). We investigated whether the observed enhanced platelet aggregation induced by collagen depended on this glycoprotein by using the GPVI-selective agonist convulxin. We found that TG platelet aggregation induced by low concentrations of convulxin was similarly enhanced, as with collagen, suggesting cooperation between secreted ATP- and GPVI-mediated signaling under mild stimulation of this receptor. In agreement with this finding and using ADP receptor antagonists, Quinton et al27 have reported a role for secreted ADP during platelet aggregation provoked by low concentrations of convulxin, whereas platelet aggregation at higher concentrations of convulxin was unaffected by these agents.

In human platelets, we have identified the ERK2 signaling pathway as an intracellular mechanism subserving the function of the ATP-gated P2X₁ ion channel during platelet aggregation induced by low concentrations of collagen. We have shown that the P2X₁-mediated ERK2 activation is needed to amplify dense-granule release initiated by this agonist. As in human platelets, the aggregation of wild-type mouse platelets evoked by low concentrations of collagen was abolished after blockade of the ERK2 pathway with the selective MEK1/2 inhibitor U0126. Accordingly, the increased collagen-induced aggregation of transgenic platelets coincided with the up-regulation of ERK2 and could equally be abolished by U0126. It appears that P2X₁ and ERK2 are 2 components of a common signaling cascade in human and mouse platelets.

Thus, even though the P2X₁-mediated Ca²⁺influx seems to be small, this necessarily local signal close to the plasma membrane, where ERK2 is also translocated, is likely to generate a significant local trigger for the promotion of Ca²⁺- and ERK2-dependent secretion.

To investigate the physiologic relevance of the P2X₁-coupled ERK2 pathway in thrombosis, we used an in vivo model of pulmonary thromboembolism by injection of a mixture of collagen and epinephrine. We showed that inhibition of the ERK2 pathway by U0126 fully protected transgenic mice against lethal thrombosis induced by a low dose of collagen plus epinephrine. Wild-type mice were also partly protected after the administration of a higher dose of collagen. Together with the observation that P2X₁-overexpressing transgenic platelets were hyperreactive to collagen in an ERK2-dependent fashion, our in vivo data support the existence of a P2X₁-ERK2 signaling axis involved—among several other components—in the control of platelet function during hemostasis.

Shear-induced platelet aggregation requires initial VWF binding to platelet GPIbα and subsequent α_IIbβ₃activation.25 The fact that shear stress triggered potent aggregation of transgenic platelets whereas wild-type platelets hardly aggregated supports an additional role for the P2X₁ ion channel in this process. Because the sustained elevation of cytosolic Ca²⁺ occurs in association with shear-induced platelet aggregation, possibly as the consequence of a rapid transmembrane ion flux,28 the contribution of P2X₁-mediated Ca²⁺ influx to platelet activation can be hypothesized. This would be in analogy to the shear stress–activated Ca²⁺ influx into human endothelial cells through another member of the P2X family, P2X₄.29 The authors proposed that P2X₄ has a “shear-transducer” property through which shear stress is perceived directly or indirectly (shear-induced ATP release) and is transmitted to the cell through Ca²⁺ signaling. Further analyses of transgenic platelet responses to shear stress will enable us to determine whether P2X₁ represents a platelet shear-transducer acting in conjunction with GPIb.

Where rapid blood flow creates high wall shear rates, such as in arterioles in the healthy circulation or in atherosclerotic arteries with restricted lumen, platelet thrombus formation depends on VWF immobilized on extracellular matrix components, in particular collagens.30 Compatible with the enhanced reactivity of transgenic platelets to collagen and shear stress, perfusion of whole blood from transgenic mice over a collagen surface at a shear rate of 1000 seconds⁻¹ resulted in greatly enhanced aggregate formation and surface exposure of negatively charged phosphatidylserine, which is instrumental in coagulation activation.31 These platelet responses were inhibited by saratin, which blocks the binding of VWF to collagen,19 indicating enhancement of an event requiring collagen-bound VWF. Furthermore, consistent with the existence of the P2X₁-ERK2 pathway, transgenic platelet aggregation on such a surface was significantly reduced after inhibition of the ERK2 pathway with U0126, possibly because of reduced secretion.

In vitro, because of the rapid desensitization of P2X₁ by spontaneously released ATP, a high concentration of apyrase is often required to demonstrate this ion channel function.12-14 In addition, in the present experiments with isolated mouse platelets, apyrase was needed to detect αβ-meATP–induced Ca²⁺influx and platelet shape change. Using this ectonucleotidase leads to the simultaneous degradation of actively secreted ATP, the major agonist of P2X₁, but also of ADP, acting at P2Y₁ and P2Y₁₂. Therefore, the use of apyrase for in vitro studies always constitutes a compromise between receptor protection and ligand breakdown. In vivo experiments in the transgenic mouse model, where endogenous ecto-ATPases prevent P2X₁desensitization, essentially confirm the conclusions of our in vitro analyses of this ion channel function and thus demonstrate a physiologic role for P2X₁ in platelets. This implies that, in vivo, mechanisms relying on fine-tuning ATPases32 must operate to control P2X₁ activity (desensitization and resensitization). The importance of these mechanisms for platelet function has been highlighted in mice lacking the endothelial ecto-nucleoside triphosphate diphosphohydrolase CD39 (NTPDase 1). Indeed, these mice showed disordered hemostasis and thromboregulation.33 

P2X₁ knock-out mice display male infertility resulting from reduced neurogenic vas deferens contraction.34 In agreement with our study, preliminary data recently presented indicate impaired in vitro platelet aggregation induced by low doses of collagen.35 Furthermore, the perfusion of whole blood from these mice over a collagen surface revealed reduced aggregate formation under flow conditions.35 

Finally, because P2X₁ overexpression in platelets generates a prothrombotic tendency, we can speculate that pathologic deregulation of P2X₁ expression may have a significant impact on platelet activation and may contribute to abnormal thrombosis. It is noteworthy that, in other tissues, the up-regulation of P2X₁ mRNA levels has been described in pathophysiologic conditions (reviewed by Burnstock4). Thus, P2X₁, as the predominant P2 receptor subtype in bladder smooth muscle, showed a considerably increased expression in the symptomatically obstructed bladder. The up-regulation of P2X₁ mRNA in the hearts of rats with congestive heart failure has been reported.

Overall, the present study provides evidence that the P2X₁ion channel plays a role in mediating the biologic effects of ATP during platelet activation. Platelet overexpression of P2X₁resulted in increased platelet secretion and aggregation triggered by collagen and thromboxane A₂ but also enhanced platelet responses under shear stress. A novel physiologic role of the P2X₁-ERK2 signaling pathway in hemostasis and thrombosis is also proposed.