NTPDase1 (CD39) controls nucleotide-dependent vasoconstriction in mouse

Abstract

Aims

Extracellular nucleotides are vasoactive molecules. The concentrations of these molecules are regulated by ectonucleotidases. In this study, we investigated the role of the blood vessel ectonucleotidase NTPDase1, in the vasoconstrictor effect of nucleotides using Entpd1^−/− mice.

Methods and results

Immunofluorescence, enzyme histochemistry, and HPLC analysis were used to evaluate both NTPDase expression and activity in arteries and isolated vascular smooth muscle cells (VSMCs). Vascular reactivity was evaluated in vitro and mean arterial blood pressure was recorded in anesthetized mice after nucleotide i.v. infusion. Expression of nucleotide receptors in VSMCs was determined by RT–PCR. Entpd1^−/− mice displayed a dramatic deficit of nucleotidase activity in blood vessel wall in situ and in VSMCs in comparison to control mice. In aortic rings from Entpd1^−/− mice, UDP and UTP induced a potent and long-lasting constriction contrasting with the weak response obtained in wild-type rings. This constriction occurred through activation of P2Y₆ receptor and was independent of other uracil nucleotide-responding receptors (P2Y₂ and P2Y₄). UDP infusion in vivo increased blood pressure and this effect was potentiated in Entpd1^−/− mice. In addition, pressurized mesenteric arteries from Entpd1^−/− mice displayed an enhanced myogenic response, consistent with higher local concentrations of endogenously released nucleotides. This effect was inhibited by the P2 receptor antagonist RB-2.

Conclusion

NTPDase1 is the major enzyme regulating nucleotide metabolism at the surface of VSMCs and thus contributes to the local regulation of vascular tone by nucleotides.

1. Introduction

Extracellular nucleotides take part in a wide range of physiological and pathological processes including the regulation of vascular tone.¹ In the circulation, sources of extracellular nucleotides are numerous. Red blood cells, endothelial cells (ECs), activated platelets, and neutrophils release adenine nucleotides (ATP and ADP) in a non-lytic manner. This release occurs in response to mechanical constraints, such as shear stress, and also to hypoxia, hyperoxia, or agonist stimulation.² In the vessel wall, sympathetic nerve terminals release ATP that is co-stored with noradrenaline. The release of adenine nucleotides has been more extensively studied due to accessible methods to measure these nucleotides while uracil nucleotides are more difficult to quantify.³ Nevertheless, UTP and UDP were shown to be released from ECs and activated platelets, and UTP release has also been reported in kidney microcirculation.⁴

Once released, nucleotides and nucleosides (adenosine) exert their biological effect via the activation of membrane bound P2 and P1 receptors to regulate cell function in an autocrine or paracrine manner.² To date, seven ligand-gated P2X (P2X_{1 – 7})⁵ and eight G protein-coupled P2Y receptors (P2Y_{1,2,4,6,11 – 14})⁶ were cloned and characterized in humans. While P2X receptors are all activated by ATP, P2Y receptors respond differently to ATP, ADP, UTP, UDP, or UDP-glucose. Three other receptors, CysLT_1,2 and GPR17 are closely related to the P2Y family and respond to both leukotriens and uracil nucleotides.⁶

Nucleotides participate to local and systemic control of blood flow at different levels.² Endothelial P2 receptor activation induces a local vasorelaxation. This effect involves three major dilating factors, nitric oxide, prostacyclin and endothelium-derived hyperpolarizing factor, depending on the species and the vascular territories considered. Several receptors have been implicated in these processes. While P2X₁ and P2X₄ receptors were shown recently to mediate ATP-induced vasorelaxation in resistance arteries,^7,8 P2Y₁ and P2Y₂ are responsible for ADP and UTP/ATP-induced relaxation.¹ In contrast, the activation of P2 receptors from vascular smooth muscle cells (VSMCs) promotes vasoconstriction via P2X⁹ or pyrimidine-sensitive P2Y receptors that take part in the neurogenic response of resistance arteries.^10,11

The concentrations of extracellular nucleotides, and thus P2 receptor activation, are regulated by ectonucleotidases that are membrane-bound enzymes with an extracellularly facing catalytic site.¹² Among these enzymes, the nucleoside triphosphate diphosphohydrolase-1 (NTPDase1) and NTPDase2 are expressed in the vasculature.¹³ NTPDase1 is expressed on ECs and VSMCs and NTPDase2 in the adventitia surrounding blood vessels, probably at the surface of fibroblasts. By hydrolysing ADP in the blood stream, a key player in platelets activation, NTPDase1 was first proposed to contribute to the antithrombotic property of ECs¹⁴ before its genetic identification as the lymphocyte marker CD39.¹⁵ The study of NTPDase1 deficient mice confirmed the antithrombogenic role of the enzyme^16,17 and its implication in the preservation of vascular permeability.^18,19 To date, the reported functions of vascular NTPDase1 are associated to its endothelial and not muscular expression. We show here that NTPDase1 is the major enzyme hydrolysing nucleotides at the surface of VSMCs and that its absence results in an increase in nucleotide-dependent vasoconstriction.

2. Methods

2.1 Animals

Experiments were carried out in accordance with the guidelines of the Institutional Ethical Committee for Experimental Animals and conform to Guide for the Care and Use of Laboratory Animals (NIH Publication No. 85-23, revised 1996).

2.2 Immunostaining

Cryosections of OCT-embedded tissues were fixed with acetone containing 0.5% phosphate-buffered formalin and immunostained with either mN1-2_C (guinea pig anti-mouse NTPDase1 antisera), mN2-36_L (rabbit anti-mouse NTPDase2 antisera),²⁰ anti α-actin monoclonal antibody (clone 1A4, Sigma, Oakville, ON), or anti-mouse CD31 (clone MEC13.3, BD Pharmingen, San Diego, CA) for endothelium-labelling, in TBS 2% goat serum. Alexa Fluor^® 633, 594, and 488-conjugated goat anti-mouse, anti-guinea pig, and F(ab′)2 fragment of anti-rabbit were used, respectively, for immunofluorescence. Sections were mounted using antifading mowiol and analysed with an Olympus IX70 microscope. Staining was performed with Vectastain ABC elite kit (Vector Laboratories, Burlingame, CA) with 3,30-diaminobenzidine as chromogen.

2.3 Enzyme histochemistry

Nucleotide hydrolysis in situ was evaluated as previously described.²¹

2.4 NTPDase1 expression and activity in VSMCs

VSMCs primary cultures were obtained as previously described.²² NTPDase1 expression was evaluated by western blot using mN1-2_C. For ectonucleotidase activity, nucleotides were incubated at 37°C with confluent VSMCs and supernatants were analysed by HPLC at indicated time-points.²³

2.5 Isometric contraction of aortic rings

Vascular tone of thoracic aortas was measured in vertical organ chambers containing Krebs solution.

2.6 Determination of myogenic responses in mesenteric artery

The reactivity of mesenteric arteries (diameter 150 μm) was determined with a wire myograph, and the myogenic tone (MT) with a pressure arteriograph.

2.7 Blood pressure measurement

The carotid artery of anesthetized mice was cannulated with PE-10 polyethylene tubing (Becton Dickinson, Oakville, ON) containing 50 U/mL heparin in saline. A cannula, inserted in the contralateral jugular vein was used for intravenous infusions of agonists. Changes in mean arterial blood pressure (ΔMAP) were detected through the carotid with a pressure transducer connected to a Blood Pressure Analyser-200A (Micro-Med, Tustin, CA).

2.8 Statistical analyses

The raw data were compared using unpaired Student’s t-test. For Figures 5 and 7 measurements were compared by ANOVA followed by a Bonferroni post hoc test for multigroup comparisons.

Figure 5.

P2Y₆ receptor mediates uracil nucleotide-induced vasoconstriction in mouse aorta. Constriction experiments were performed in the presence of the NTPDase1 inhibitor ARL 67156 (100 μmol/L). UTP and UDP-induced an equivalent concentration-dependent contraction in P2ry2^−/− (filled triangle), P2ry4^−/− (filled square), and C57Bl/6 control (filled circle) aortic rings which was absent in P2ry6^−/− (filled diamond) rings. ATPγS-dependent contraction was partially diminished in P2ry2^−/− aortic rings. Results are expressed as the percentage of KCl (50 mmol/L)-induced contraction and represent the mean ±SEM of five independent experiments for each group of mice tested except for P2ry6^−/− that were carried out with the aortic sections from three individuals. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 7.

Increased myogenic tone in Entpd1^−/− arteries. Third-order mesenteric arteries reactivity was evaluated with a wire myograph. (A) Uracil nucleotide-induced contractions were enhanced in Entpd1^−/− arteries. These responses were inhibited by the P2 receptor antagonist RB-2 (30 μmol/L). (B) An arteriograph was used for MT determination (see Methods). The luminal pressure was increased in 25 mmHg steps from 25 to 150 mmHg. The myogenic response (% MT) was expressed by the active (AD, determined with Ca²⁺) and passive diameters (PD, determined without Ca²⁺) such that %MT= [(PD−AD)/PD] · 100%. (C) The passive dilatation of arteries (without calcium) was equivalent in both strains of mice. Data in each panel represent the mean ± EM of measurements performed on arteries from eight different mice. Data were analysed by ANOVA followed by a Bonferroni multiple comparison post hoc test. **P < 0.01, ***P < 0.001.

Sources of the different mice strains and a complete detailed method are available in the online supplement.

3. Results

3.1 NTPDase1 is the major ectonucleotidase in VSMCs

In mouse aorta, NTPDase1 is expressed in VSMCs of the media, as indicated by its colocalization with α-actin (Figure 1A). NTPDase2 is expressed in the surrounding adventitia. The aortic endothelium, whose integrity was controlled by uniform PECAM labelling (data not shown), did not display strong immunoreactivity for NTPDase1, contrasting with previous observations in smaller arterioles that showed a strong immunolabelling of NTPDase1 in ECs and weaker staining in VSMCs.¹³

Figure 1.

Deficit of nucleotidase activity in Entpd1^−/− mouse vasculature. (A) Entpd1^+/+ mouse aortic sections were analysed for expression of NTPDase1 and NTPDase2 by immunofluorescence (left panels: lum, lumen of the aorta; med, media; adv, adventitia). Purple colour on the merge panel materialize the colocalization of NTPDase1 (red) and α-actin (blue). NTPDase2 (green) is expressed by adventitial cells. (B) As shown by brown precipitates, Entpd1^+/+ mouse aortas’ sections display a broad nucleotidase activity absent from the media of Entpd1^−/− sections. (C) Similar deficit of nucleotide hydrolysis was observed in heart arteries. NTPDase2 activity is easily distinguished in the proximal adventitia of Entpd1^−/− arteries with triphosphate nucleotides as substrates (B and C).

The comparison of nucleotide hydrolysis in situ revealed a broad nucleotidase activity in the media of Entpd1^+/+ mice aortic sections that was not detected in Entpd1^−/− tissues (Figure 1B). Such deficit of activity was observed in the Entpd1^−/− arterioles of other tissues including heart (Figure 1C), lungs, and liver (data not shown), suggesting that NTPDase1 is the main nucleotidase expressed in VSMCs. The activity of NTPDase2, which is a nucleoside triphosphatase with low diphosphatase activity,²⁴ was more evident in Entpd1^−/− aorta with ATP or UTP as substrates (Figure 1B). In accordance to its expression pattern, NTPDase2 activity was restricted to the adventitia.

NTPDase1 expression in VSMCs primary cultures was confirmed by western blot (Figure 2A), and its biochemical activity by HPLC. While ATP, ADP, UTP, and UDP were rapidly hydrolysed by Entpd1^+/+ cells, this hydrolysis was negligible in the presence of Entpd1^−/− cells (Figure 2B), highlighting the major contribution of NTPDase1 in nucleotide hydrolysis at the surface of VSMCs.

Figure 2.

NTPDase1 is the major ectonucleotidase at the surface of VSMCs. (A) Western blot analysis showing NTPDase1 expression in NTPDase1 transfected cells (COS mN1), Entpd1^+/+ VSMCs, but not in Entpd1^−/− VSMCs. The loading control with the anti-α-actin antibody revealed a band in both cultures but was absent in COS cells. (B) The nucleotide hydrolysis of Entpd1^+/+ and Entpd1^−/− VSMCs supernatants was evaluated by HPLC at the indicated time-points after addition of the indicated nucleotides (100 μmol/L). Results were normalized to spontaneous degradation of nucleotides. Data represent the mean + SEM of experiments performed in duplicates on 3–5 different cell cultures. *P < 0.05, **P < 0.005, ***P < 0.0005.

3.2 Nucleotide-induced vasoconstriction is potentiated in Entpd1^−/− mouse aortic rings

The role of NTPDase1 in nucleotide-induced vasoconstriction was assessed by comparing the isometric tension developed by aortic rings from Entpd1^+/+ and Entpd1^−/− mice. Comparable responses in both genotypes to a depolarizing KCl (Fig 3A), or U46619 (Fig 3G), a thromboxane A₂ analogue, confirmed that Entpd1^−/− vessels could contract normally. In contrast, nucleotide-induced vasoconstriction was drastically enhanced in Entpd1^−/− aortic rings. As illustrated by the effect of UDP, the force and the steadiness of contraction were augmented in the absence of NTPDase1 (Figure 3B, C, H ). Similar results were obtained for UTP, ATP, and ADP. Figure 3C–F shows the dose response curve obtained by cumulative concentrations of nucleotides. UDP and UTP induced a strong constriction of Entpd1^−/− rings (E_max = 0.57 g and 0.54 g, respectively), in comparison to Entpd1^+/+ rings (<0.1 g tension at 100 μM, Figure 3C, D), being the second most potent vasoconstrictors after U46619 (E_max = 0.69 g), nearly two times more potent than the α-adrenergic agonist phenylephrine (E_max = 0.33 g). ADP and ATP were less potent vasoconstrictors with EC₅₀ in the high micromolar range (Fig 3E, F). Pharmacological inhibition of NTPDase1 in Entpd1^+/+ aortic rings with ARL 67156 concentration-dependently potentiated UDP-induced vasoconstriction (Figure 3H). ARL 67156 had no effect in Entpd1^−/− aortic rings confirming that its effect on Entpd1^+/+ aortic rings was due to NTPDase1 inhibition.

Figure 3.

Contractile effect of nucleotides is unmasked in Entpd1^−/− aortas. (A) While the constriction produced by depolarizing KCl (50 mmol/L) was similar in both strain of mice, (B) the vasoconstrictor effect of UDP (100 μmol/L) was greatly enhanced and more stable in Entpd1^−/− compared with wild-type aortic rings. (C–F) Nucleotides induced a sigmoid shaped dose-dependent contraction in Entpd1^−/− aortic rings (open circles) but only a weak and progressive tension in Entpd1^+/+ rings (filled circles). (G) The dose–response curve induced by U46619 was similar in both genotypes. (H ) Pharmacological inhibition of NTPDase1 with ARL 67156 potentiated the contraction induced by UDP in a dose-dependent manner in Entpd1^+/+ but not in Entpd1^−/− aortic rings. Each dose-response curve was built from the experiment performed on rings from four to seven aortas from different mice.

3.3 Expression of nucleotide receptors in mouse aorta

The potential receptor(s) involved in the contractile response to uracil nucleotides were identified by RT–PCR on both intact thoracic aorta and VSMC cDNAs. Primers for nucleotide-responding, or potentially responding receptors, except P2X receptors that are exclusively activated by ATP, were used to screen for the presence of P2Y, CysLT_1,2, GPR17, and two orphan receptors closely related to P2Y receptor family, GPR34 and GPR87.⁶ The method and primer pairs are available in the online supplement. P2Y₁, P2Y₂, and P2Y₆ receptors were highly expressed in both entire thoracic aorta and VSMCs (Figure 4). P2Y_12,13, CysLT₁, and GPR34 messengers were amplified in thoracic aorta but not in VSMCs, P2Y₄ and P2Y₁₄ were barely detectable, and CysLT₂ was absent. The differential expression between aorta and VSMCs may be due to contaminating platelets (P2Y₁₂), red blood cells (P2Y₁₃), and lymphocytes (GPR34) that are impossible to completely eliminate from aortic preparations. Endothelium denudation before thoracic aorta RNA extraction did not modify the expression pattern of these receptors (data not shown). Thus, VSMCs mainly express P2Y₁, P2Y₂, and P2Y₆ receptors.

Figure 4.

Expression of nucleotide responding receptors in aorta and VSMCs cultures. Expression of P2Y, orphan GPR17, GPR34, GPR87 as well as CysLT₁ and CysLT₂ receptors was examined in aortas and VSMCs cultures (passage 2–4) by RT–PCR. Positive controls (Ctrl) were achieved with appropriate cDNAs. Expected PCR fragment sizes are listed in Supplementary material online.

3.4 Nucleotide-induced vasoconstriction in mouse aorta: evidence for P2Y₆ receptor

The pharmacological profile of nucleotide-induced vasoconstriction was evaluated in Entpd1^−/− aortic rings. As shown by their EC₅₀ values (Table 1), the rank order of potency revealed a predominant effect of uracil nucleotides with the selective P2Y₆ receptor agonist, 3-phenacyl-UDP,²⁵ among the most potent agonists (EC₅₀ = 23 ±12 μM, Table 1). The pA₂ of three non-selective P2 receptor antagonists (RB-2, suramin and PPADS) on UDP and UTP-induced constrictions were similar, suggesting that a unique receptor population is responsible for the constrictor effect of UDP and UTP (Table 2). Adenine nucleotides displayed a weak effect (EC₅₀ > 200 μM) theoretically ruling out the contribution of P2Y₁ (ADP), P2Y₂/P2Y₄ (UTP and ATP), and P2X receptors (ATP). To definitively determine the receptor(s) involved, we used knockout mice for uracil nucleotide responding P2Y receptors, namely P2Y₂, P2Y₄, and P2Y₆. UDP and UTP induced similar contractions in P2ry2^−/−, P2ry4^−/− and wild-type aortic rings but was abolished in P2ry6^−/− aortas (Figure 5). As a control, ATPγS-induced contraction was exclusively diminished in P2ry2^−/− aortic rings, suggesting that, as reported for the endothelial receptors,²⁶ VSMCs P2Y₂ receptor mediates the effect of ATPγS, but not the one of UTP. These results demonstrate that P2Y₆ receptor is the mediator of the strong vasoconstrictor property of UDP and UTP in mouse aorta.

Table 1.

EC₅₀ values of nucleotide-induced vasoconstriction of Entpd1^−/− mice thoracic aorta

Nucleotide (n)	EC50 (μM)
UTP (5)	1.06 ±0.64
UDP (7)	1.76 ±0.94
UP4A (3)	3.43 ±0.11
3 phenacyl-UDP (2)	22.8 ±12.4
TDP (4)	24.4 ±4.9
TTP (3)	27.7 ±9.2
ATPγS (3)	130 ±86
ADPβS (1)	53.6
ITP (3)	109 ±20
IDP (3)	120 ±60
GTP (3)	190 ±59
GDP (3)	323 ±65
ATP (5)	>200
ADP (2)	>200

Table 2.

pA₂ values of P2 receptor antagonists on UDP and UTP-induced vasoconstriction of Entpd1^−/− mice thoracic aorta

	UDP	UTP
PPADS	4.25 ±0.19	4.13 ±0.11
Reactive blue 2	4.31 ±0.16	4.16 ±0.15
Suramin	4.08 ±0.14	4.05 ±0.15

3.5 Effect of UDP on blood pressure

Intravenous infusion of UDP (10 μmol/kg) displayed a biphasic effect on blood pressure in mice: a slight transient decrease followed by a marked and sustained increase that reached a maximum between 2 and 3 min after UDP infusion (Figure 6A). The initial slight drop in blood pressure is likely linked to endothelial P2Y₆ receptor since its activation has been shown to produce relaxation in vitro.^26,27 The pressor effect was significantly more important in Entpd1^−/− than in Entpd1^+/+ mice upon 1 and 10 μmol/kg UDP infusions (Figure 6B). Angiotensin II induced a similar increase in blood pressure in both strains of mice suggesting that the increased hypertensive effect in the absence of NTPDase1 was specific to UDP (Figure 6B). These data fit with the exacerbated constrictor effect of UDP obtained in Entpd1^−/− arteries in vitro and show that this phenomenon can influence blood pressure in vivo.

Figure 6.

In vivo effect of UDP. (A) The intravenous infusion of UDP (10 μmol/kg) induced an initial fast hypotensive response followed by a prolonged increase in blood pressure that reached a maximum between 2 and 3 min following infusion that was significantly higher in Entpd1^−/− mice. (B) The maximal increase in blood pressure was significantly higher in Entpd1^−/− for 1 and 10 μmol/kg UDP infusions. As a control, similar increase pressor effect was observed in response to angiotensin II (AgII, 100 pmol/kg). Data represent the mean ± SEM of four experiments performed on different mice. *P < 0.05, **P < 0.01.

3.6 Enhanced MT in Entpd1^−/− arteries

MT, the reduction in diameter in response to an increase in intraluminal pressure, is the hallmark of resistance arteries. This response is entirely dependent on smooth muscle contraction, although it is regulated by mechanisms including local changes in flow, endothelial function, nerve activation, and autacoids.²⁸ Stretch, among other cellular stresses, induces the release of nucleotides. Since we observed an enhanced contractile effect of exogenous nucleotides in Entpd1^−/− aorta (Figure 3), mesenteric resistance arteries of these mice were used to assess the potential contribution of endogenously released nucleotides in MT. Pharmacological reactivity of mesenteric arteries was investigated with a wire myograph. In agreement with the results obtained with aortas, the constrictor effect of uracil nucleotides was facilitated in Entpd1^−/− mesenteric arteries (Figure 7A) while U46619-induced contraction was equivalent in both genotypes (data not shown). EC₅₀ of UDP and UTP-induced contraction were 3.5 and 4.5 μM in Entpd1^+/+ arteries, and 0.28 and 0.27 μM in Entpd1^−/− arteries. The greater difference obtained in the contractile effect of uracil nucleotides between Entpd1^+/+ and Entpd1^−/− aortas is likely linked to its thicker SMC layer that slow nucleotides diffusion favouring their exposition to NTPDase1. The P2 receptor antagonist RB-2 potently inhibited those responses (Figure 7A). MT was measured with a pressure arteriograph. Mesenteric arteries developed MT between 25 and 150 mmHg intraluminal pressure which was significantly enhanced in Entpd1^−/− arteries, denuded (Figure 7B) or not (data not shown) of their endothelium. The passive dilatation of arteries (absence of extracellular calcium) was equivalent in both strains of mice (Figure 7C). In the same setting, the pharmacological response to phenylephrine was similar: 67±8 vs. 71±7% of contraction for wild-type and Entpd1^−/− arteries, respectively. RB-2 prevented the increase in MT observed in Entpd1^−/− arteries but did not have a significant effect in wild-type arteries (Figure 7B). These results suggest that nucleotides released by stretching the vessel wall contribute to enhance the MT that can be easily measured in Entpd1^−/− arteries.

4. Discussion

To date, the reported vascular functions of NTPDase1 (thromboregulation, vascular permeability) were linked to its endothelial expression. We show here that NTPDase1 is the major ectonucleotidase expressed in VSMCs. Its expression correlates with the high hydrolytic activity previously reported on these cells²⁹ and is absent in the media layer of Entpd1^−/− arteries in situ and at the surface of Entpd1^−/− VSMCs primary cultures (Figures 1 and 2). Importantly, the absence of NTPDase1 unmasks a strong constrictor effect of UDP and UTP in aortas and mesenteric arteries (Figures 3 and 7).

In agreement with an important nucleotidase activity at the VSMCs surface, hydrolysis-resistant analogues UDPβS and UTPγS are 1000-fold more potent as vasoconstrictor agents than UDP and UTP.¹⁰ Based on our pharmacological and molecular data, we found that P2Y₆ receptor is involved in the uracil nucleotide-dependent vasoconstriction in mouse vessels. Using knockout mice, we definitely excluded P2Y₂ and P2Y₄ receptors (Figure 5). It was not really surprising for P2Y₄ receptor since its transcript was barely detected in mouse aorta (Figure 4), and absent in mesenteric artery,³⁰ contrasting with rat vasculature where P2Y₄ receptor was detected in both VSMCs and intact aorta.³¹ The result was not expected for P2Y₂ since this receptor is largely expressed in aorta and VSMCs (Figure 4), and was proposed as the main receptor involved in UTP- and ATP-dependent inositol polyphosphate formation in rat VSMCs.³² Indeed, we found that ATPγS-dependent contraction was partly dependent on P2Y₂ receptor (Figure 5), but not this of UTP. As P2Y₂ is required for VSMCs migration and proliferation in response to UTP,³³ it appears that distinct P2Y receptors are involved in different functions of VSMCs.

The fact that UTP and UDP were found equipotent vasoconstrictors apparently did not fit with the P2Y₆ as the only receptor type involved. Indeed, UDP was reported to be 100-fold more potent than UTP at the human recombinant P2Y₆ receptor.³⁴ Such pharmacological profile (UDP and UTP equipotent) was previously reported in the vasculature of mouse³⁰ and rat,³⁵ and was attributed to a P2Y₆-like receptor. Interestingly, UDP was more potent than UTP in wild-type aorta, while the two nucleotides were equipotent in Entpd1^−/− aortas. As UTP is a much better substrate than UDP for NTPDase1,²³ the enzyme may influence the apparent pharmacology of P2Y₆, affecting the real potency of UTP. Further investigations are required to validate this hypothesis that may conciliate the pharmacological differences between P2Y₆ and P2Y₆-like receptor.

The fact that NTPDase1 deficiency was associated with a gain of vessel contractility is interesting as, so far, studies involving Entpd1^−/− mice showed that the absence of NTPDase1 was associated with P2 receptor desensitization.^{17,36 – 38} Thus, NTPDase1 plays different roles: considering easily desensitized receptors such as P2Y₁, NTPDase1, by scavenging extracellular nucleotides, prevents their desensitization,¹⁷ in contrast, for poorly desensitizing receptors such as P2Y₆, NTPDase1 limits their activation and its absence/inhibition facilitates it. It is likely that NTPDase1 will be essential in view to terminate P2Y₆ receptor activation, this being especially true considering the poverty of uracil metabolizing enzymes apart from the NTPDases.³⁹

The absence of NTPDase1 unmasks the contractile effect of UDP also in vivo since intravenous infusion of UDP led to an enhanced pressor effect in Entpd1^−/− compared with control mice (Figure 6). This was due to an increase in peripheral resistance, since no modification in the heart rate occurred (data not shown). UTP did not display the same effect being mostly hypotensive (data not shown) contrasting with in vitro experiments in which UTP and UDP were vasoconstrictors (Figure 3). This discrepancy may be due to the involvement of endothelial P2 receptors, such as P2Y₂ which may functionally antagonize constrictor effect on VSMCs. It is noteworthy that Entpd1^−/− mice do not display elevated blood pressure (data not shown). This may be linked to the fact that the absence of NTPDase1 also facilitates the activation of endothelial P2 receptors, resulting in a facilitated vasodilation that balances the enhanced VSMCs P2 receptor activation (Kauffenstein et al., submitted for publication).

MT controls local blood flow in resistance arteries and depends on the intrinsic property of VSMCs to contract in response to intraluminal pressure increase. Mechanisms underlying MT development are largely unknown but autacoids released by VSMCs such as 20-HETE⁴⁰ or sphingosine1-phosphate⁴¹ have been proposed to participate in this adaptive vascular response. We found that MT was significantly more pronounced in Entpd1^−/− when compared with Entpd1^+/+ arteries (Figure 7B). This may be attributable to the release of nucleotides upon cell stretching. In agreement with these data, uracil nucleotides exert a potent vasoconstriction in mouse mesenteric arteries (Figure 7A and Vial and Evans³⁰). The enhanced MT observed in Entpd1^−/− arteries was strongly inhibited by RB-2, in agreement with the hypothesis that released nucleotides reinforce MT. Several arguments play in favour of uracil nucleotides as autocrine stimulators of SMC. First, uracil nucleotides are released following cellular mechanical stress.^3,42,43 VSMCs themselves could well be a source of uracil nucleotides since vascular smooth muscles have been reported to differ from skeletal muscles and other tissues by a particularly low intracellular ATP/UTP ratio. It has also been reported that pyrimidines are released by rat perfused hind limb in response to constrictor agents (NA, vasopressin, AgII).⁴⁴ Thus, mechanical distension of SMC could well release extracellular uracil nucleotides. Recently Nishida et al.⁴⁵ reported that an autocrine loop involving uracil nucleotides release and P2Y₆ receptor activation occurs in cardiomyocytes and participates to the development of cardiac fibrosis. This autocrine loop is similar to the one that, as suggested in the present study, participates in MT amplification. Further investigations will be required to identify the nucleotides released and P2 receptors taking part to this process. Nonetheless, P2Y₆ receptors seem to be valid candidates since antagonists of this receptor type inhibit MT in mouse mesenteric arteries.⁴⁶

Altogether, our results suggest that modulation of the expression level or activity of NTPDase1 in VSMCs influence the constrictor effect of nucleotides. As a consequence, in inflammatory or oxidative environment, where NTPDase1 activity is diminished,⁴⁷ nucleotide receptor activation and the resulting vasoconstriction may be facilitated. On the opposite, an increase in NTPDase1 activity should reduce the constrictor effect of nucleotides. This situation may occur following hypoxic conditions where the expression of the enzyme is dramatically increased.¹⁹

Besides high affinity transporters, ectoenzymes constitute an efficient mechanism for extracellular mediators’ clearance. The local activity of fatty acid amide hydrolase/monoacyl-glycerol lipase, dipeptidyl peptidase IV, and Sphingosine-1-Phosphate Phosphohydrolase which hydrolyse endocannabinoids,⁴⁸ neuropeptides,⁴⁹ and sphingosine1-phosphate,⁵⁰ respectively, regulate their bioavailability and their associated vasoactive effects. Among these enzymes, NTPDase1 represents a potential target to modulate vascular tone.

In conclusion, NTPDase1 is the major enzyme hydrolysing nucleotides at the surface of VSMCs where it prevents nucleotide-induced vasoconstriction. This effect is unmasked in Entpd1^−/− vessels leading to greater P2Y₆ receptor-dependent vasoconstriction in vitro and pressor effect of exogenous infusion of UDP in vivo. The deficit in nucleotidase activity also potentiates MT strengthening a role of endogenously released nucleotides and NTPDase1 in this auto-regulatory process. Further work will be required to establish the relevance of these mechanisms in relation with physiopathological conditions where NTPDase1 expression/activity is affected. The complete identification of the actors of the purinergic signalling in VSMCs will help to identify potential targets to modulate VSMCs physiology and pathology-associated dysfunctions such as those encountered in hypertension and atherosclerosis.