We previously reported that ATP coreleased with norepinephrine from cardiac sympathetic nerves activates presynaptic P2X purinoceptors (P2XR), thereby enhancing norepinephrine exocytosis. Blockade of ectonucleoside triphosphate diphosphohydrolase 1 (E-NTPDase1/CD39) potentiates norepinephrine exocytosis, whereas recombinant soluble CD39 (solCD39) in-hibits it. This suggested that CD39 gene (Entpd1) deletion would enhance purinergic and adrenergic signaling by preserving ATP and its norepinephrine-releasing activity. However, we found that the neurogenic contractile response of vasa deferentia from Entpd1-null (CD39–/–) mice was attenuated and accompanied by reduced activity of pre- and postsynaptic P2XR, whereas contractile responses to K+ or norepinephrine remained intact. In addition, the magnitude of ATP and norepinephrine exocytosis from cardiac synaptosomes was decreased in CD39–/– mice. Inhibition of E-NTPDase1/CD39, or solCD39 administration, did not affect the attenuated contractile response of vasa deferentia from CD39–/– mice. Notably, Entpd1 deletion and pharmacological P2XR desensitization in control mice similarly attenuated vasa deferentia responses. Thus, excessive and prolonged ATP exposure resulting from CD39 deletion desensitizes pre- and postjunctional P2XR at the sympathetic neuromuscular junction. This diminishes purinergic activity directly and adrenergic activity indirectly. It remains to be determined whether this desensitization results from receptor internalization, changes in receptor conformation or phosphorylation. Shutdown of ATP signaling in CD39–/– mice may represent a defense mechanism for the prevention of purinergic overstimulation. Our findings emphasize the cardioprotective role of neuronal CD39: by reducing presynaptic facilitatory effects of neurotransmitter ATP, CD39 attenuates norepinephrine release and its dysfunctional consequences. Moreover, by virtue of its antithrombotic action CD39 can potentially prevent the transition from myocardial ischemia to infarction.
Efficient regulation of neuronal function and synaptic transmission in the autonomic nervous system is achieved with precise control of transmitter release and subsequent clearance. The actions of amino acids and monoamines are terminated by neuronal reuptake via specific high-affinity transporters; in contrast, the effects of acetylcholine and ATP are terminated by biotransformation in the extracellular space (Amara and Kuhar, 1993; Zimmermann, 2000).
During the past decade, many ectoenzymes involved in hydrolysis and interconversion of extracellular nucleotides have been characterized, among them the CD39 family, the ectonucleotide pyrophosphatase/phosphodiesterases, alkaline phosphatases, and ectoprotein kinases (Zimmermann, 2001; Robson et al., 2006). Many of these ectonucleotidases catalytically remove nucleotides in tandem with other ectoenzymes to generate nucleosides at very high turnover rates. CD39, an apyrase, ectonucleotidase or ectonucleoside triphosphate diphosphohydrolase 1 (E-NTPDase1) was identified as the first member of the family of E-NTPDases (Zimmermann et al., 2000). CD39 is the dominant ectonucleotidase, and its main function is to hydrolyze nucleoside tri- and diphosphates, yielding monophosphates of both purine and pyrimidine nucleosides (Zimmermann, 2000; Zimmermann et al., 2000; Robson et al., 2006). CD39 is widely expressed, in endothelium (Kaczmarek et al., 1996; Marcus et al., 1997), in lymphocytes (Maliszewski et al., 1994), and in the sympathetic nervous system (Sesti et al., 2002; Machida et al., 2005). Transgenic mice with targeted disruption of CD39 have been developed. These CD39-null mice exhibit major alterations in hemostasis and in inflammatory and thrombotic reactions to reperfusion injury. They have major defects in healing responses and angiogenesis (Enjyoji et al., 1999; Pinsky et al., 2002).
When present in extracellular compartments, ATP and other nucleotides act as agonists for P2X purinoceptors (P2XR) at the cell surface. To date, seven subtypes of the ionotropic P2XR family and eight members of the P2YR family have been identified (Burnstock, 2007). We recently reported that ATP coreleased with norepinephrine (NE) from cardiac sympathetic nerve endings activates presynaptic P2XR, thereby enhancing NE exocytosis, whereas P2YR activation reduces NE exocytosis, thus partially counteracting the effects of P2XR stimulation (Sesti et al., 2002). We additionally identified E-NTPDase1/CD39 in cardiac sympathetic nerve endings (Machida et al., 2005). By metabolizing ATP, this enzyme terminates purinergic signaling, effectively decreasing NE release (Sesti et al., 2002; Machida et al., 2005). Accordingly, NE and ATP are coreleased upon depolarization of cardiac sympathetic nerve endings, and the ATP enhances NE exocytosis by a process modulated by E-NTPDase1/CD39 activity.
Evidence for purinergic cotransmission is particularly strong for the sympathetic innervation of smooth muscle in arterioles and vas deferens (Sneddon and Westfall, 1984). Due to its very dense sympathetic plexus and easy accessibility, the vas deferens is frequently used to evaluate effects of various drugs at the sympathetic junction. Thus, when sympathetic nerves of the vas deferens are stimulated, ATP and NE are released and then act predominantly at postjunctional purinergic P2X1- and adrenergic α1-receptors, respectively (Westfall et al., 1996a; Mulryan et al., 2000). Most of the ATP is released at the onset of stimulation, whereas NE is released throughout the entire stimulation period. Neuronally released ATP elicits excitatory junction potentials, which in turn initiate action potentials, resulting in the phasic component of the neurogenic contractile response. The second contractile phase is noradrenergic and more prolonged (Todorov et al., 1996).
We have demonstrated that pharmacological inhibition of E-NTPDase1/CD39 potentiates NE exocytosis, whereas its soluble recombinant form, solCD39 (Gayle et al., 1998), inhibits NE release (Sesti et al., 2002; Machida et al., 2005). We therefore postulated that inactivation of CD39 would result in enhancement of purinergic and adrenergic signaling. To test this hypothesis, we studied vasa deferentia and cardiac sympathetic nerve endings (cardiac synaptosomes) isolated from CD39 null mice. Our evidence indicates that the excessive and prolonged purinergic and adrenergic stimulation associated with CD39 gene deletion leads to subsequent desensitization of pre- and postjunctional P2XR, thereby decreasing sympathetic activity.
Materials and Methods
Tissue Harvest. All animal studies were approved by the Weill Cornell Medical College Institutional Animal Care and Use Committee. CD39-null mice, generated as described previously (Pinsky et al., 2002), were obtained from Dr. David J. Pinsky (University of Michigan, Ann Arbor, MI), with permission from Immunex/Amgen Inc. (Seattle, WA). They were backcrossed onto the C57BL/6 background at least nine generations. All studies were performed on male young adult mice at 4 to 5 months of age. Following pretreatment with heparin (100 IU i.p.), mice were anesthetized with CO2, and then they were killed by cervical dislocation. Hearts and vasa deferentia were quickly excised and cooled in ice-cold modified Krebs-Henseleit (KH) solution (118 mM NaCl, 4.7 mM KCl, 1.2 mM MgSO4·7H2O, 24 mM NaHCO3, 1.1 mM KH2PO4, 10 mM glucose, 0.5 mM pyruvic acid, and 2.5 mM CaCl2·2H2O). KH solution was continuously equilibrated with 95% O2 + 5% CO2, as described previously (Schaefer et al., 2006).
Mouse Vas Deferens Experiments. The mid-portion of the vas deferens was rapidly mounted into vessel chambers containing Krebs-Ringer (KR) solution (118 mM NaCl, 4.7 mM KCl, 1.2 mM MgSO4·7H2O, 25 mM NaHCO3, 1.2 mM KH2PO4, 8.3 mM glucose, and 2.5 mM CaCl2·2H2O) equilibrated with 95% O2 + 5% CO2, as described previously (Schaefer et al., 2006). One end was connected to a force transducer. Dissected vessels were allowed to equilibrate for at least 1 h at a resting tension of 250 mg at 37°C. Aerated Krebs-Ringer solution was replaced every 30 min. At the beginning of the experiment, a brief contraction with 80 mM K+ was induced, and the developed tension was continuously recorded with Power-Lab/8SP (ADInstruments, Colorado Springs, CO). After washout and an additional equilibration period of 60 min, sympathetic nerves were stimulated by electrical-field stimulation (EFS) at 2, 4, 8, 12, and 16 Hz for 15 s at 5-min intervals with a pulse width of 1 ms and supramaximal voltage. EFS was applied using an S48 stimulator and SIU5 stimulus isolation unit (Grass Technologies, West Warwick, RI). Contractile responses to EFS were individually analyzed, and they are expressed as percentage of the response to 80 mM K+. When used, drugs were incubated for at least 15 min, or as otherwise indicated, and the response to EFS was analyzed and compared with the response in the absence of drug. To differentiate between pre- and postsynaptic sites of action, concentration-response curves for ATP (0.1–100 μM) and NE (0.01–100 μM) were constructed, and values are expressed as percentage of the response to K+.
Preparation of Cardiac Synaptosomes. After excision of the heart, the aorta was cannulated with an 18-gauge custom-made steel cannula, and the heart was perfused at constant pressure (100 cm of H2O) with KH buffer at 37°C. Two hearts per each mouse type were perfused for 20 min to ensure that blood components were removed from the coronary vasculature. Both hearts were minced together in ice-cold 0.32 M sucrose containing 1 mM EGTA, pH 7.4, and preparation of cardiac synaptosomes was performed as described previously (Koyama et al., 2003). In brief, minced tissue was digested with 40 to 75 mg of collagenase (type II; Worthington Biochemicals, Freehold, NJ) per 10 ml of HEPES-buffered saline solution (HBS; containing 50 mM HEPES, pH 7.4; 144 mM NaCl; 5 mM KCl; 1.2 mM CaCl2; 1.2 mM MgCl2; 1 mM pargyline HCl to prevent enzymatic destruction of synaptosomal NE) per gram wet heart weight for 1 h at 37°C. After low-speed centrifugation (10 min at 120g, 4°C), the resulting pellet was suspended in 10 vol of 0.32 M sucrose and homogenized with a Teflon/glass homogenizer and respun. The pellet, which contained cellular debris, was discarded and the supernatant equally subdivided into 4 to 6 tubes before centrifugation (20,000g for 20 min, 4°C). The resulting pellets containing cardiac synaptosomes were resuspended in HBS to a final volume of 500 μl in the presence or absence of drugs. HBS contained 1 mM pargyline and 1 mM tropolone, 1 μM atropine, 1 μM desipramine, and 1 μM yohimbine. Each suspension was incubated in a water bath at 37°C with either K+ or α,β-methylene ATP (α,β-MeATP). K+ was incubated at 37°C for 5 min, and α,β-MeATP was incubated for 5 s (Sesti et al., 2002). In each experiment, one sample was untreated and incubated for the same length of time. Following incubation, each sample was centrifuged for 20 min at 20,000g and 4°C, and the supernatant was assayed for NE and ATP content. The pellet was assayed for protein content by a modified Lowry procedure (Koyama et al., 2003).
Mouse Platelet Aggregation Assay. Mice were anesthetized with a ketamine/xylazine mixture (80/5 mg/kg i.p., respectively). Using a short length (∼1 cm) of heparinized hematocrit tubing, a small blood sample (750 μl) was collected from the retrobulbar venous plexus directly into a mixture of saline (165 mM NaCl; 675 μl) and sodium citrate (3.2%; 75 μl). The mixture was centrifuged for 5 min at 120g and 25°C to obtain PRP (Pinsky et al., 2002; Kopp et al., 2006). Aggregometry was performed using an aggregometer system (model 490; Chrono-Log Corp., Havertown, PA) (Pinsky et al., 2002; Kopp et al., 2006). ADP (10 μM) was used as agonist. Aggregation was allowed to proceed for a minimum of 5 min. Extent of aggregation was calculated as both maximal extent (amplitude) and as “area under the curve” (as if the curve were a chromatographic peak; arbitrary units; duration 5 min), as determined by the Aggro/Link software, version 5.2.2 (Chrono-Log Corp.).
Norepinephrine Assay. Synaptosomal supernatants were assayed for NE by high-performance liquid chromatography with electrochemical detection as described previously (Koyama et al., 2003). The detection limit was ∼0.05 pmol.
ATP Assay. ATP levels were measured by a luminescence assay (ATP bioluminescence assay kit HS II; Roche Diagnostics, Indianapolis, IN). Samples (50 μl) of each supernatant were pipetted into appropriate test tubes, placed in a TD-20/20 luminometer (Turner Designs, Sunnyvale, CA), and processed by autoinjection of 50 μlof luciferin/luciferase reagent. ATP concentrations were calculated from a calibration curve constructed the same day using ATP standards included in the kit. The amount of ATP was expressed as femtomoles per milligram of protein (Sesti et al., 2003).
Drugs and Chemicals. ATP disodium salt, NE HCl, prazosin HCl, pyridoxal phosphate-6-azophenyl-2′,4′-disulfonic acid (PPADS), α,β-MeATP, and ARL67156 were obtained from Sigma-Aldrich (St. Louis, MO).
Statistics. Values refer to mean ± S.E.M. One-way analysis of variance followed by Dunnett's post test, paired and unpaired Student's t test were used where appropriate as indicated. P < 0.05 was considered significant.
CD39 Gene Deletion Attenuates the Neurogenic Contractile Response of Mouse Vas Deferens. The smooth muscle of the vas deferens responds to EFS with an initial rapid contraction, attributed to the activation of postjunctional P2X1R by ATP released from sympathetic nerves, followed by a sustained contraction, due to activation of postjunctional α1-adrenoceptors by NE coreleased with ATP (Todorov et al., 1996). Some overlap is likely, since neuronally released ATP also activates prejunctional P2XR, thereby enhancing NE release (Todorov et al., 1996; Sesti et al., 2002). Frequency-response curves for the EFS-induced increase in contractility in vasa deferentia isolated from wild type (WT) and CD39–/– mice are shown in Fig. 1. In the 2- to 16-Hz frequency range, the magnitude of both contractile phases progressively increased as a function of stimulation frequency (Fig. 1, A and B). The slopes of the frequencyresponse curves obtained from CD39–/– vasa were markedly lower than those from WT vasa; indeed, the mean maximal effect of EFS (Emax) was ∼60 and 70% lower in CD39–/– vasa than in WT vasa for the purinergic and adrenergic phases, respectively (Fig. 1, A and B).
Preincubation of vasa deferentia isolated from WT mice with the P2XR antagonist PPADS at 1 and 10 μM caused a concentration-dependent reduction of the slopes of the frequency-response curves for both purinergic and adrenergic phases. Thus, the Emax of each phase was reduced by ∼25% with 1 μM PPADS, whereas with 10 μM PPADS, the Emax of the purinergic and adrenergic phases was reduced by ∼60 and 35%, respectively (Fig. 2, A and B). Notably, the frequency-response curves for the purinergic and adrenergic phases in WT vasa incubated with 10 μM PPADS were superimposable with the curves obtained in untreated CD39–/– vasa. Only at 16 Hz was the adrenergic response in the presence of 10 μM PPADS higher than the corresponding response in untreated CD39–/– vasa. Indeed, at the higher stimulation frequencies NE release is nearly maximal, and facilitatory presynaptic P2XR contribute only minimally to the adrenergic response (Todorov et al., 1996). Thus, a large adrenergic response is achieved at high stimulation frequencies despite the desensitization of presynaptic P2XR that facilitate NE release. Frequency-response curves in CD39–/– vasa were not affected by PPADS at 1 and 10 μM(n = 4; data not shown).
Recombinant CD39 Administration Fails to Correct the Attenuated Vas Deferens Response in CD39-Null Mice. Administration of our soluble recombinant form of CD39 (i.e., solCD39) (Gayle et al., 1998) to vasa deferentia isolated from WT mice and subjected to EFS induced a concentration-dependent downward shift of the frequency-response curves for both purinergic and adrenergic contractile phases (Fig. 3, A and B). This was consistent with diminished activation of both pre- and postjunctional P2XR, due to decreased availability of transmitters ATP and NE at the neuroeffector junction. In contrast, administration of solCD39 to vasa from CD39–/– mice did not modify the frequency-response curve of the purinergic phase, whereas an upward, albeit not significant trend was observed in the frequency-response curve of the adrenergic phase (Fig. 3, C and D).
Nucleotidase Inhibition Fails to Potentiate the Vas Deferens Response in CD39 Null Mice. When vasa deferentia isolated from WT mice were preincubated with the nucleotidase inhibitor ARL67156 at 100 μM, the purinergic and adrenergic phases of the contractile response to EFS were both potentiated, as indicated by an upward shift of the frequency-response curves (Emax were increased by ∼50–60%; Fig. 4, A and B). In contrast, preincubation with 100 μM ARL67156 of vasa from CD39–/– mice did not modify the frequency-response curve of the purinergic phase (Fig. 4C). An upward, albeit not significant trend was observed in the frequency-response curve of the adrenergic phase (Fig. 4D).
CD39 Gene Deletion and P2X-Receptor Desensitization Similarly Attenuate the Mouse Vas Deferens Response. Desensitization of P2XR induced in vasa deferentia from WT mice by a 30-min preincubation with a high concentration of α,β-MeATP (30 μM) caused a marked decrease in the slope of the purinergic frequency-response relationship (Fig. 5A) and a moderate downward shift of the adrenergic curve in the 2 to 12-Hz EFS frequency range, but not at 16 Hz (Fig. 5B). In vasa deferentia from CD39–/– mice, preincubation with α,β-MeATP (30 μM) induced a slight, but not significant, downward shift of the purinergic frequency-response curve (Fig. 5A). Preincubation with 30 μM α,β-MeATP failed to affect the adrenergic curve obtained from CD39–/– vasa (Fig. 5B). Analogous to α,β-MeATP, preincubation with 100 μM ATP for 30 min markedly reduced the purinergic response in WT vasa but hardly at all in CD39–/– vasa (data not shown).
CD39 Gene Deletion Attenuates the Response of Murine Platelets to ADP. Platelet aggregation responses to 10 μM ADP were lower in PRP of CD39–/– than in that of WT mice (Fig. 6). Lack of enzymatically active CD39 in both endothelial cells and leukocytes of CD39–/– mice, resulting in reduced metabolism of extracellular nucleotides, most likely plays a role in the decreased responsiveness to ADP in CD39–/– mice (Enjyoji et al., 1999; Pinsky et al., 2002). Thus, locally higher levels of circulating nucleotides result in partial desensitization of platelet nucleotide receptors. This may occur in the absence of a measurable change in plasma adenine nucleotides (Enjyoji et al., 1999), due to alternative plasma ATP/ADP metabolism via phosphodiesterase(s) (Birk et al., 2002). It is well known that especially platelet P2X1R are easily desensitized, remaining so for a prolonged period, reducing platelet responsiveness. This contributes to the known increase in susceptibility to thrombotic disorders of CD39 null mice (Enjyoji et al., 1999; Pinsky et al., 2002; Dwyer et al., 2004).
CD39 Gene Deletion Attenuates the Mouse Vas Deferens Response to Postsynaptic P2XR Activation. Because the neurogenic response to EFS of the isolated vas deferens includes two components, pre- and postjunctional (Todorov et al., 1996), we questioned whether changes observed in vasa deferentia from CD39–/– mice also involved postsynaptic sites. Therefore, we assessed postjunctional responses to K+, NE, α,β-MeATP, and ATP of vasa from WT and CD39–/– mice. The magnitude of contractile responses to 80 mM K+ in vasa from CD39–/– mice was equal to that of vasa from WT mice (Fig. 7A). This excludes a decreased responsiveness in vasa deferentia from CD39–/– mice. Similarly, the concentration-response curve for the contractile response to exogenous NE in CD39–/– vasa was superimposable on the curve for WT vasa (Fig. 7B). In contrast, contractile responses to α,β-MeATP and ATP in vasa of CD39–/– mice were markedly attenuated compared with those of WT vasa, as indicated by the downward shift of the concentration-response curves (Fig. 7, C and D).
CD39 Gene Deletion Attenuates NE and ATP Exocytosis from Cardiac Sympathetic Nerve Terminals. Since our findings suggested that prejunctional mechanisms play a role in the reduced neurogenic response of vasa deferentia from CD39–/– mice, we next determined whether this involved a deficient release of ATP and NE. For this, we chose the cardiac synaptosomal preparation, an optimal model for the investigation of NE and ATP exocytosis from sympathetic nerve endings (Sesti et al., 2003; Schaefer et al., 2006). Upon K+-induced depolarization (30–100 mM), synaptosomes released NE and ATP in a K+ concentration-dependent manner (Fig. 8). Synaptosomal NE release in basal conditions (i.e., pre-K+ depolarization) was the same in WT and CD39–/– preparations (Fig. 8A), whereas basal release of ATP was 2- to 3-fold greater in CD39–/– than in WT synaptosomes (Fig. 8C), likely resulting from accumulated ATP due to lack of its hydrolysis. The magnitude of depolarization-induced NE release was greater in WT than in CD39–/– synaptosomes (Fig. 8B). Notably, when depolarized with 30 to 100 mM K+, CD39–/– synaptosomes released relatively less ATP than WT synaptosomes (Fig. 8D), a phenomenon probably related to the much higher basal release of ATP in CD39–/– than in WT synaptosomes (Fig. 8C).
Since our findings collectively implied a malfunction of prejunctional P2XR, we compared NE exocytosis elicited by the P2XR agonist α,β-MeATP in synaptosomes from WT and CD39–/– mouse hearts. Concentration-response curves for the release of NE induced by α,β-MeATP are shown in Fig. 9A. Although synaptosomes from WT and CD39–/– hearts responded to increasing concentrations of α,β-MeATP with graded increases in NE release, there was a wide separation between the two concentration-response curves in the 0.3 and 3 μM range. Indeed, α,β-MeATP was ∼10-fold less effective in CD39–/– than in WT synaptosomes in releasing synaptosomal NE (EC50 value for α,β-MeATP was 0.32 ± 0.01 and 3.28 ± 0.1 μM for WT and CD39–/– synaptosomes, respectively) (Fig. 9A). Although the N-type Ca2+ channel blocker ω-conotoxin GVIA at 100 nM was equally effective in its antiexocytotic action in both CD39–/– and WT synaptosomes (Fig. 9B), the P2XR antagonist PPADS at 10 μM was significantly less effective in inhibiting 100 mM K+-induced NE exocytosis in CD39–/– than in WT synaptosomes (Fig. 9B). This points to a role for CD39 in the regulation of nucleotide-induced responses, in the absence of a direct effect on NE release via other signaling pathways.
Extracellular ATP and ADP as well as UTP and UDP partake in biological signaling by activating nucleotide P2 receptors (Burnstock, 2007). The concentration of extracellular nucleotides is regulated by a variety of surface-located enzymes known as ectonucleotidases (Zimmermann et al., 2000; Robson et al., 2006). The most prominent of these ectoenzymes are members of the ectonucleoside 5′-triphosphate diphosphohydrolase (NTPDase) family (Robson et al., 2006). Although seven NTPDases (1–6 and 8) have been extensively studied at biochemical and molecular levels, their cellular actions remain to be defined (Zimmermann et al., 2000; Zimmermann, 2001; Bigonnesse et al., 2004; Robson et al., 2006). The most widely expressed NTPDase, CD39/NTPDase1, as well as NTPDase2, exhibits tissue distributions (e.g., neural, vascular, and secretory) that coincide with the distribution of several P2XR and P2YR. CD39/NTPDase1 is, however, the dominant ectonucleotidase of the vasculature and immune system and rate-limiting in the formation of adenosine. At synapses where transmitter lifetime is controlled by enzymatic hydrolysis, inhibition of inactivating enzymes can profoundly affect changes in membrane conductance elicited by neurotransmitters. Indeed, inhibition of ecto-ATPase by the selective inhibitor ARL67156 potentiates excitatory junction potentials, ATP overflow, and development of spontaneous excitatory junction potentials in guinea pig vas deferens (Sneddon et al., 2000; Ghildyal et al., 2006). Therefore, at the sympathetic neuroeffector junction, ecto-ATPase modulates the prejunctional release of ATP as well as its postjunctional actions.
In immunohistochemical studies using the mouse monoclonal antibody BU61 directed against human CD39, we have demonstrated that CD39 is expressed in sympathetic nerves (Machida et al., 2005). We also reported that ATPase activity, with a pharmacological profile similar to CD39, is released upon depolarization of cardiac sympathetic nerve terminals (Sesti et al., 2001). This release of ATPase activity has also been reported by Westfall and associates, who demonstrated a marked effect on nucleotide metabolism in vasa deferentia of various species (Westfall et al., 2000, 2002; MihaylovaTodorova et al., 2002). Therefore, in addition to CD39 anchored to plasma membranes, a releasable form of CD39 may exist, at least in the microenvironment of sympathetic nerves.
Studies using selective pharmacological reagents, or knockout mice lacking P2 purinoceptors or nucleotide-metabolizing enzymes, have enhanced our comprehension of purinergic signaling by ATP and its metabolites. This occurs in hemostasis, neuronal function, fertility, inflammation, and ion transport. Thus, deletion of E-NTPDase1/CD39 should favor accumulation of nucleotides and thereby enhance purinergic signaling at pre- and postsynaptic sites. In contrast, we found the opposite to be true. Indeed, the neurogenic contractile response to EFS was markedly depressed in vasa deferentia isolated from CD39–/– mice. Furthermore, we found that CD39 deletion resembled blockade of P2XR with the selective P2XR antagonist PPADS and that PPADS was almost without effect in CD39–/– mouse vas deferens. Thus, a dysfunction of postsynaptic P2XR seemed likely. Indeed, we found that the contractile response of the vas deferens to exogenously applied ATP or α,β-MeATP was markedly depressed, whereas the response to exogenously applied NE or K+ was maintained. Similarly, the responses to ADP of CD39-null platelets was reduced (Fig. 6), confirming previous reports (Enjyoji et al., 1999; Pinsky et al., 2002; Dwyer et al., 2004). That CD39 deletion leads to P2R desensitization had been observed previously in angiogenesis and immune cell models (Goepfert et al., 2001; Mizumoto et al., 2002). Thus, purinoceptor desensitization as a result of CD39 deletion seems to be a general phenomenon.
Not only was the purinergic phase of the vas deferens contractile response reduced in CD39–/– mice but also the adrenergic phase was blunted, despite an intact postsynaptic response to NE. Thus, we predicted that NE release would be decreased in CD39-null animals to account for this observation. Previously, we and others had demonstrated that presynaptic P2 receptors have a major influence on NE release: P2XR enhance NE release, whereas P2YR attenuate it (Sesti et al., 2002, 2003; Queiroz et al., 2003). Thus, diminished NE exocytosis in cardiac synaptosomes from CD39–/– mice might have resulted from a dysfunction of presynaptic P2XR. This was supported by our finding that activation of presynaptic P2XR in cardiac synaptosomes from CD39–/– mice yielded much less NE than in WT synaptosomes. Moreover P2XR blockade with PPADS was significantly less effective in preventing NE exocytosis in CD39–/– synaptosomes than in WT controls. Thus, decreased activity of presynaptic P2XR plays an important role in the attenuation of the neurogenic response of the vas deferens and the reduced NE exocytosis from sympathetic nerve endings. It is also plausible that presynaptic P2YR may play an additional inhibitory role. We next questioned whether the decrease in ATP metabolism would lead to measurable increases in ATP levels in CD39–/– synaptosomes. Thus, we found that basal ATP release from cardiac nerve terminals was 3.5-fold higher in CD39–/– mice than in WT controls. This increase was probably sufficient to desensitize P2XR, in a manner similar to the degradation-stable agonist α,β-MeATP (Burnstock, 2007).
The relevance of E-NTPDase1/CD39 in sympathetic transmission is clearly evident upon administration of recombinant CD39 (i.e., solCD39). In the presence of solCD39, the contractile response of vasa deferentia from WT mice (Fig. 3), and NE exocytosis from guinea pig heart synaptosomes (Sesti et al., 2002), were markedly attenuated. Furthermore, the E-NTPDase1/CD39 inhibitor ARL67156 significantly increased the contractile response of vasa deferentia from WT mice (Fig. 4) as well as enhanced NE exocytosis from guinea pig heart synaptosomes (Sesti et al., 2002). In contrast, ARL67156 was without effect in vasa deferentia from CD39–/– mice. Enhanced neurogenic responses of vas deferens in the presence of ARL67156 have been reported previously (Westfall et al., 1996b; Queiroz et al., 2003). This indicates that endogenous ATP, released by depolarization of sympathetic nerve endings, exerts an autocrine P2XR-mediated facilitatory effect on NE exocytosis (Sesti et al., 2002). We propose that in CD39–/– mice, these P2XR become dysfunctional (i.e., desensitized) due to lack of ATP hydrolysis and consequent buildup of ATP concentrations at the sympathetic neuromuscular junction.
As mentioned previously, P2 receptors that mediate facilitation of NE release belong to the P2XR subtype (Sesti et al., 2003; Kubista and Boehm, 2006). Immunohistochemical studies have shown that P2X1, P2X2, and P2X3 subunits are expressed in rat vas deferens. The P2X2 and P2X3 subunits are preferentially expressed in nerve fibers and terminals, and P2X1 in smooth muscle (Vulchanova et al., 1996). Targeted deletion of the P2X1R attenuates the neurogenic contractile response of the vas deferens, and it abolishes the response to the P2XR agonist α,β-MeATP (Mulryan et al., 2000). P2X1R undergo rapid desensitization (Boué-Grabot et al., 2000). Nanomolar ATP concentrations drive significant fractions of the rapidly desensitizing P2X1R pool into a long-lasting refractory state. Recovery is slow, with a time constant of 12 min (Rettinger and Schmalzing, 2003). Therefore, attenuation of postsynaptic responses to EFS and α,β-MeATP, as we observed in vasa deferentia of CD39–/– mice, is most likely due to P2X1R desensitization.
Because reconstitution with solCD39 failed to restore the response of vasa deferentia from CD39–/– mice to normality, a long-lasting malfunction of P2 receptors (i.e., desensitization due to ATP accumulation) occurred. Ligand-gated receptors have been shown to undergo receptor internalization: indeed, more than 50% P2X1R are quickly internalized during continued exposure to α,β-MeATP (Ennion and Evans, 2001). In addition, conversion between inactive and active states has been proposed, since 10 min after agonist stimulation, 80% P2X1R activity reappeared in the vas deferens. However, only 40% of the contractile response was regained (Ennion and Evans, 2001). In addition, another laboratory reported that 80 to 90% of the green fluorescent protein-tagged P2X1R signal was internalized after P2X1R stimulation (Dutton et al., 2000). Whether internalization of P2X1R is necessary for recovery from desensitization remains to be determined. Transitions in and out of the inactivated state would be the rate-limiting step in desensitization and recovery of P2X1R. In general, decreased recycling of P2X1Rtothe cell surface may be the mechanism of long-term down-regulation (Ennion and Evans, 2001).
It has been proposed that without P2X1R desensitization, persistently elevated levels of ATP would lead to cell toxicity due to permanently open cation channels. Extracellular ATP concentrations can reach ∼1 to 20 μM even in the absence of cell lysis, and this could increase ATP levels to millimolar amounts (Mahaut-Smith et al., 2004). Therefore, P2 receptor desensitization is actually protective for normal homeostasis. As opposed to high ATP concentrations within the synaptic junction, low ATP concentrations may be present at the border of a synaptic cleft. Due to the high ATP sensitivity of the P2X1R, a silencing of P2X1R responsiveness is likely when the nerve terminal is repeatedly activated. In contrast, a low-sensitivity ATP receptor may repeatedly respond to rapid rises in synaptic ATP, when transmitter concentrations rise and fall rapidly in both physiological and pathological conditions. Due to their low ATP sensitivity, these remaining P2XR shut down immediately when ATP concentrations fall to the submicromolar level, and they are therefore readily prepared to respond to the next rise in ATP concentration. Both of these two receptor types may complement each other in subserving particular demands in neuronal excitability. Gene deletion of CD39 disturbs this fundamental balance of purinergic signaling due to increased and persistent P2 receptor stimulation. This therefore leads to long-lasting attenuation of autonomic transmission in vas deferens and the heart. In general, our data support the concept that pre- and postsynaptic P2XR differ in terms of their responsiveness to excess extracellular ATP. This may result from a difference in P2X subunits (e.g., P2X2 and P2X3 presynaptic and P2X1 postsynaptic), from a difference in the mechanisms responsible for desensitization, or both.
In conclusion, we have demonstrated that targeted deletion of CD39 leads to attenuation of sympathetic neurotransmission, which we attribute to P2XR desensitization at both pre- and postjunctional sites of the sympathetic neuromuscular junction. Whether this desensitization results from receptor internalization, from changes in receptor conformation or phosphorylation (e.g., interactions with protein kinases A and C, β-arrestins, clathrin, and mitogen-activated protein kinases) remains to be determined. We propose that our findings underscore the key role of neuronal CD39 in cardiac sympathetic function; not only does CD39 attenuate NE release and its dysfunctional consequences by reducing the presynaptic facilitatory effects of transmitter ATP but its thromboregulatory action can also potentially prevent the transition from myocardial ischemia to infarction. This occurs via its capacity to block platelet activation and recruitment (Marcus et al., 1997; Pinsky et al., 2002).
We thank Christopher Morrey for excellent graphical assistance.
This work was supported by National Institutes of Health Grants R01 HL34215, R01 HL73400, and P01 HL46403 (to R.L.) and R37 HL47073, P01 HL46403, and R01 NS41462 (to A.J.M. and M.J.B.), as well as by Merit Review grants from the Department of Veterans Affairs (to A.J.M. and M.J.B.).
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
ABBREVIATIONS: E-NTPDase1, ectonucleoside triphosphate diphosphohydrolase 1; P2XR, P2X purinoceptor(s); P2YR, P2Y purinoceptor(s); NE, norepinephrine; solCD39, soluble recombinant form of CD39; KH, Krebs-Henseleit; EFS, electrical-field stimulation; HBS, HEPES-buffered saline; α,β-MeATP, α,β-methylene ATP; PRP, platelet-rich plasma; PPADS, pyridoxal phosphate-6-azophenyl-2′,4′-disulfonic acid; ARL67156, trisodium 6-N,N-diethyl-β-γ-dibromomethylene-d-adenosine-5′-triphosphate; WT, wild type; NTPDase, ectonucleoside 5′-triphosphate diphosphohydrolase.
- Received May 4, 2007.
- Accepted June 11, 2007.
- The American Society for Pharmacology and Experimental Therapeutics