Purinergic dysregulation in pulmonary hypertension

This is the first study to report that deletion of CD39 results in high pulmonary arterial pressures and increases in the intravascular ATP/adenosine ratio and pulmonary vascular P2X1 receptors. It also shows that reconstitution of ectonucleotidase activity lessens, and P2X1 receptor blocking eliminates, elevated pulmonary arterial pressures.

Abstract

Despite the fact that nucleotides and adenosine help regulate vascular tone through purinergic signaling pathways, little is known regarding their contributions to the pathobiology of pulmonary arterial hypertension, a condition characterized by elevated pulmonary vascular resistance and remodeling. Even less is known about the potential role that alterations in CD39 (ENTPD1), the ectonucleotidase responsible for the conversion of the nucleotides ATP and ADP to AMP, may play in pulmonary arterial hypertension. In this study we identified decreased CD39 expression on the pulmonary endothelium of patients with idiopathic pulmonary arterial hypertension. We next determined the effects of CD39 gene deletion in mice exposed to normoxia or normobaric hypoxia (10% oxygen). Compared with controls, hypoxic CD39^−/− mice were found to have a markedly elevated ATP-to-adenosine ratio, higher pulmonary arterial pressures, more right ventricular hypertrophy, more arterial medial hypertrophy, and a pro-thrombotic phenotype. In addition, hypoxic CD39^−/− mice exhibited a marked increase in lung P2X1 receptors. Systemic reconstitution of ATPase and ADPase enzymatic activities through continuous administration of apyrase decreased pulmonary arterial pressures in hypoxic CD39^−/− mice to levels found in hypoxic CD39^+/+ controls. Treatment with NF279, a potent and selective P2X1 receptor antagonist, lowered pulmonary arterial pressures even further. Our study is the first to implicate decreased CD39 and resultant alterations in circulating purinergic signaling ligands and cognate receptors in the pathobiology of pulmonary arterial hypertension. Reconstitution and receptor blocking experiments suggest that phosphohydrolysis of purinergic nucleotide tri- and diphosphates, or blocking of the P2X1 receptor could serve as treatment for pulmonary arterial hypertension.

Listen to this article's corresponding podcast at http://ajpheart.podbean.com/e/purinergic-dysregulation-in-pulmonary-hypertension/.

NEW & NOTEWORTHY

This is the first study to report that deletion of CD39 results in high pulmonary arterial pressures and increases in the intravascular ATP/adenosine ratio and pulmonary vascular P2X1 receptors. It also shows that reconstitution of ectonucleotidase activity lessens, and P2X1 receptor blocking eliminates, elevated pulmonary arterial pressures.

pulmonary arterial hypertension (PAH) is a progressive disorder characterized by pulmonary arterial vasoconstriction, vascular remodeling, and smooth muscle cell proliferation (32). The resultant increase in pulmonary vascular resistance (PVR) leads to right ventricular afterload, hypertrophy and, ultimately, death due to right heart failure. Although the triggers of PAH are poorly understood, imbalances involving the prostacyclin, nitric oxide and endothelin-1 pathways have been implicated, and therapies targeting these pathways are currently used to treat the disease (32, 44). Common features of successful targeting of the implicated pathways include vasodilation and inhibition of vascular remodeling. However, the prognosis for PAH remains poor despite treatment targeting the known pathways, with a 15% one-year mortality despite modern treatment options (46). Thus the development of therapies targeting novel pathways that contribute to the pathobiology of PAH is of great importance.

The purinergic nucleotides adenosine triphosphate (ATP), adenosine diphosphate (ADP), adenosine monophosphate (AMP) and the nucleoside adenosine are extracellular signaling molecules(5) that can signal downstream effector targets to modulate endothelial and smooth muscle cell growth (31), apoptosis (12), coagulation (30), vascular tone (6, 7) and inflammation (15). These ligands interact with a variety of cognate P1 (adenosine) and P2 (ATP and ADP) receptors to produce effects that may be complimentary or antagonistic to one another, depending upon tissue-specific receptor subtypes and concentrations (6).

Intravascular nucleotide concentrations are regulated primarily by the ectonucleotidase CD39 [ectonucleoside triphosphate diphosphohydrolase 1 (ENTPD1)] and CD73 (5′-nucleotidase) (25, 41, 51). CD39 phosphohydrolyzes ATP and ADP to AMP, which is further dephosphorylated to adenosine by CD73. Thus these ectoenzymes play a critical role in maintaining extracellular nucleotide and adenosine homeostasis. The purinergic signaling contribution to the pathobiology of PAH has not been fully elucidated. For example, work by our laboratory showed increased CD39 expression on circulating endothelial microparticles from patients with idiopathic PAH (49), although the significance of this finding is unknown. This led us to hypothesize that an imbalance in extracellular nucleotide and adenosine ratios caused by altered CD39 expression may play a role in the pathobiology of PAH. A subsequent study supported this hypothesis by showing downregulation of CD39 on the pulmonary endothelium of patients with idiopathic PAH (IPAH) (22). To further investigate the role of CD39 and purinergic signaling in PAH we utilized the murine chronic hypoxia model. This model is commonly used to investigate pathways that may contribute to PAH (8, 9, 19, 45). We report here that deletion of the CD39 gene in the setting of chronic hypoxia results in significantly altered concentrations of plasma nucleotides and adenosine, significant upregulation of the lung P2X1 receptor, and a consistent and unexpectedly severe pulmonary hypertension phenotype. Furthermore, reconstitution of CD39 using a soluble apyrase mitigates the development of PH, while antagonism of the P2X1 receptor prevents the development of PH altogether.

MATERIALS AND METHODS

Procurement of Human Lung Tissue

Formalin-fixed, paraffin embedded human lung tissue from controls and patients with IPAH were obtained through the Cardiovascular Medical Research and Education Fund-Pulmonary Hypertension Breakthrough Initiative (PHBI) network. The PHBI study protocol was approved by the Institutional Review Boards of the participating sites in the network, and all sites were adherent to the requirements of the U.S. Federal Policy for the Protection of Human Subjects (45 CFR, Part 46), and supported the general ethical principles of the Declaration of Helsinki.

Immunohistochemistry and Immunofluorescence Microscopy

Formalin-fixed, paraffin-embedded lungs were sectioned at a thickness of 5 μm. Sections underwent deparaffinization, dehydration, antigen retrieval, and quenching of endogenous peroxidase activity. Blocking, primary antibody labeling, and immunoperoxidase staining were performed as recommended by the ImmPRESS Polymer Detection Kit and ImmPACT DAB peroxidase HRP substrate (Vector Laboratories, Burlingame, CA). Rabbit anti-human CD39 antibody (sc-33558, Santa Cruz Biotechnology, Dallas, TX) and guinea pig anti-mouse NTPDase1 antibody (http://ectonucleotidases-ab.com, Quebec, Canada) were used to detect CD39 in human and mouse lung tissue, respectively. Rabbit P2X1 receptor antibody (H-100, Santa Cruz Biotechnology) was for both mouse and human lung sections. Quantification of vessel CD39 staining was calculated as the percent of CD31-DAB-positive endothelial area that also stained positive for CD39. The area of positive antibody staining was determined using the color deconvolution followed by threshold functions within Fiji (42) (ImageJ, NIH, Bethesda, MD). Percent CD39 staining for 5 vessels from each group was calculated as: (area of CD39-positive staining/area of CD31-positive staining) × 100.

Assessment of mouse pulmonary arterial medial thickness was performed following αSMA primary antibody (ab5694, Abcam, Cambridge, MA) staining and immunofluorescence labeling application with the Tyramide Signal Amplification Plus System (PerkinElmer, Waltham, MA). Sections were then visualized using at 40× or 60× magnification using a Nikon TE2000E2 microscope (Melville, NY) with MetaMorph software (Molecular Devices, Sunnyvale, CA). Images were captured and ImageJ (NIH, Bethesda, MD) was used to measure overall vessel and medial diameter. Percent of media thickness for each vessel was computed as [(external diameter − internal diameter)/external diameter] × 100. Presented values are the mean of 10 fields from 6 mice in each group. Vessels were categorized based on diameter (<50 μm, 51–100 μm, >100 μm). All vessels were analyzed by an investigator blinded to study conditions.

Generation of CD39-Deficient Mice

All animal experiments were approved by, and carried out in accordance with the University of Michigan University Committee on Use and Care of Animals (UCUCA) guidelines. CD39-deficient mice (CD39^−/−) were generated from C57BL/6 mice (Charles River, Wilmington, MA, strain code 027) through the University of Michigan Transgenic Animal Model Core by deleting the first exon of CD39 using Cre-Lox recombination. A ploxPFLPneo cloning vector containing two loxP sites flanking the first exon of CD39 was used. BACPAC clone RP23-117D11 (derived from a C57BL/6 mouse) was used as source DNA for the insertion of BamHI sites at −490 and +950 of exon 1. This vector was introduced into C57BL/6-derived Bruce 4 embryonic stem (ES) cells and selected in G418. ES clones with successful insertion of LoxP were identified by qPCR, confirmed by Southern blot analyses of Bgl1/Sal1-digested DNA, and injected into C57BL/6 blastocysts. Probes were designed to detect the 5′ end and 3′ ends of the DNA targeted for homologous recombination in transfected stem cell lines. The 5′ and 3′ probes were radiolabeled and hybridized with genomic DNA purified from untransfected stem cells to confirm a successfully transfected stem cell line, and an F1 generation mouse. High contribution from the ES cell clone was assessed using coat color contribution. Successive breeding to a FLP recombinase-expressing mouse followed by a ubiquitously-expressed EIIa-CRE-recombinase-expressing mouse was used to ablate CD39. Homozygous animals were produced from the mating of hemizygous chimera offspring. Genotyping by PCR analysis of genomic DNA from tail tips was performed on all mice using primer sets to confirm wild-type (WT) CD39 (5′-TGGGAAGGG GTCAGCTCTATGTGGTA-3′ and 5′-CCTTCCCCTTCCTTCCTC TTTTCCTCCGTTAT-3′) or knockout (KO) CD39 genotype (5′-GTCATTTCACAGCTGGCA AGAGGTA-3′ and 5′-CAGGAAGTGGAGGTGATAGGGACAACA-3′). Quantitative RT-PCR analyses were used to assess CD39 mRNA in various adult mouse organs. Wild-type mice (CD39^+/+) of the same genetic background (C57BL/6) were purchased from Jackson Laboratory (Bar Harbor, ME). All mice were housed in a designated Animal Resource Facility at the University of Michigan under specific pathogen-free conditions.

Normobaric Hypoxia and Normoxia Exposure

Seven-week-old CD39^+/+ and CD39^−/− mice were maintained at 10% oxygen (normobaric) for 4 wk using a purpose-built, normobaric hypoxia chamber with automated oxygen-regulating capability. Circulating air was scavenged using charcoal and soda lime (Sigma-Aldrich, St. Louis, MO). Temperature and humidity were monitored, and the latter maintained within normal limits using Dririte (Sigma-Aldrich). Normoxic control mice were maintained in the same room at 21% oxygen, which was verified by a continuous oxygen sensor (BioSpherix, Lacona, NY). All animals were given free access to food and water, and were maintained on a 12:12-h dark/light cycle.

Hemodynamic Measurements

Mice were anesthetized using 2% isoflurane. A tracheostomy was performed, and animals were ventilated using 21% or 10% oxygen administered thorough the ventilator circuit. Optimal ventilator settings were confirmed using arterial blood gas measurements. After a median sternotomy was performed, hemodynamic measurements were made using a 1.2F solid-state pressure catheter (Scisense Transonic, London, ON) inserted through the right ventricular (RV) free wall into the RV cavity and advanced into the pulmonary artery (PA). Placement was verified using the pressure waveform. Data including the right ventricular systolic pressure (RVSP) and mean PA (mPA) pressures were collected and analyzed using LabScribe2 (iWorx, Dover, NH). Continuous EKG monitoring was performed using the MouseMonitor (Indus Instruments, Webster, TX).

Arterial Blood Gas Analysis

Values including the partial pressure of oxygen (Po₂) and hemoglobin concentration were measured in heparinized arterial blood drawn from the left ventricle using an ABL800 flex analyzer (Radiometer America, Carlsbad CA).

Tissue Collection and Preparation

Lungs were gently perfused through the right ventricle using 20 mM EDTA in PBS at a constant pressure of 20 cmH₂O until the tissue blanched (3–5 min). Lungs used for histology were then perfused with 10% buffered formalin administered via the tracheostomy site, and transferred to 70% ethyl alcohol 24 h later. Lung samples for transcriptomics and Western blot analysis were placed in Allprotect Tissue Reagent (Qiagen, Vallencia, CA) and stored at −80°C. The heart was excised for weight measurements.

Right Ventricular Weight and Fulton's Index

The heart was excised, the right ventricular free wall was dissected and weighed, and the remaining septum and left ventricle were weighed. Right ventricular hypertrophy was assessed using Fulton's Index: (right ventricular weight)/(septum + left ventricular weight) × 100.

Blood Collection

Plasma for ATP, ADP, AMP and adenosine analysis was obtained by puncturing the right ventricle with a 26-gauge needle and drawing 500 μl of whole blood into a syringe prefilled with 500 μl of chilled “stop solution.” The “stop solution” used was based upon prior studies(17, 38) and contained 4.15 mM EDTA (arrests ATP catabolism), 5 nM NBTI (inhibits ATP release from erythrocytes), 10 μM forskolin (stabilizes platelets to prevent ATP release), 100 μM IBMX (inhibits cAMP phosphodiesterase), 40 μM dipyridamole (inhibits adenosine reuptake and adenosine deaminase), 10 μM EHNA (inhibits adenosine deaminase) and 10 μM 5-iodotubericidin (inhibits adenosine kinase). All reagents were obtained from Sigma-Aldrich (St. Louis, MO). The blood/stop solution mixture was centrifuged (13,000 g) for 2 min at 4°C, and the supernatant was recentrifuged for 2 min. Aliquots of the final supernatant were stored at −80°C. Blood for arterial blood gas analysis was drawn from the left ventricle using heparinized syringes and analyzed immediately.

Metabolomics Analysis

Plasma extraction.

Plasma ATP, ADP, AMP and adenosine extraction was performed using a mixture of methanol, acetone, and acetonitrile (1:1:1). The extraction solvent was added to plasma samples in Eppendorf tubes using a 4:1 solvent-to-sample ratio. The mixture was vortexed briefly, placed on ice for 5 min, vortexed again, and then centrifuged at 15,000 g for 5 min. Supernatant containing the metabolites was removed, dried and reconstituted in 100 μl of a 9:1 mixture of methanol and water. A series of calibration standards were prepared along with samples to facilitate quantification of metabolites.

Sample analysis (liquid chromatography-mass spectrometry).

An Agilent 1200 chromatography platform (Agilent Technologies, Santa Clara, CA) with a Luna NH2 HILIC (hydrophilic interaction chromatography) column (Phenomenex, Torrance, CA) was used for chromatographic separation. An Agilent 6410 series triple quadrupole mass spectrometer (Agilent Technologies, Santa Clara, CA) with electrospray ionization source (ESI) was operated in negative mode. The following transitions were used to identify and quantify metabolites: Adenosine: mass/charge (m/z) 266.1→m/z 134.1; AMP: m/z 346.1→m/z 79; ADP: m/z426.0→m/z 79; ATP: m/z 506.0→m/z 79. Data were processed by MassHunter workstation software, version B.04 (Agilent Technologies, Santa Clara, CA). Results were normalized to plasma volume and doubled (to adjust for the 1:1 dilution of whole blood into “stop solution”) to calculate final concentrations (in μM).

Quantitative RT-PCR Analysis of Purinergic Receptors

The 7500 Fast Real-Time PCR System, Taqman Gene Expression Master Mix, Human beta-2-microglobulin endogenous control, and Taqman Gene Expression Assays (Applied Biosystems, Grand Island, NY) were used to quantitate expression of the following purinergic receptor genes in whole lung homogenates: P2X1, P2X2, P2X3, P2X4, P2X5, P2Y1, P2Y2, P2Y4, P2Y6, P2Y12, ADORA1, ADORA2A, ADORA2B, ADORA3.

Ectonucleotidase Reconstitution and P2X1 Receptor Blocking

Alzet osmotic pumps (DURECT, Cupertino, CA) were used to deliver either 15 units of apyrase (A6410, Sigma-Aldrich), 25 mM of NF279 (Tocris Bioscience/Bio-Techne, Minneapolis, MN) or sterile 0.9% sodium chloride over a 4-wk period at a rate of 0.11 μl/h. NF279 has an IC₅₀ of 19 nM, and was selected for use in these experiments due to its selectivity over P2X2 (IC50 = 0.76 μM), P2X3 (IC50 = 1.62 μM), and P2X4 (IC50 > 300 μM)(39). In addition, NF279 is not degraded by ectonucleotidases (10).

Statistical Analysis

Statistical analyses and graph creation were performed using SPSS for Mac (IBM, Amnok, NY). Data are presented as means ± standard error of the mean (SE). Comparisons between groups were performed using a one-way between-groups analysis of variance with post hoc comparisons using the Tukey HSD test. A value of P < 0.05 was considered statistically significant.

RESULTS

Decreased CD39 on Pulmonary Endothelium and Plexiform Lesions from IPAH Patients

We hypothesized that patients with IPAH have decreased CD39 on both the pulmonary vascular endothelium of all vessel sizes, as well as within the angiomatoid proliferative lesions (plexiform lesions) commonly found in patients with severe IPAH (47). Immunohistochemistry using anti-CD39 antibody confirmed the presence of CD39 on CD31-positive pulmonary endothelium in vessels of all sizes (<50 μm, 50–100 μm, and >100 μm) from donors with healthy lungs (“control” in Fig. 1A). However, lung tissue from patients with IPAH showed markedly decreased pulmonary endothelial CD39 (Fig. 1A). Typical plexiform lesions containing vascular channels lined by CD31-positive endothelial cells were identified in lung samples from IPAH patients (Fig. 1B). Notably, CD39 was not present within these endothelial-rich lesions. (Fig. 1B). Quantification of the mean area of CD39-positive staining for 10 small (<50 μm) vessels selected at random from the lungs of 2 IPAH and 2 control patients (Fig. 1C) shows a statistically significant decrease in CD39-positive endothelial staining in IPAH patients. To determine if a decrease in CD39 occurs in nonvascular lung structures we examined expression in the pulmonary airway epithelium (Fig. 1D). Notably, there was no decrease in pulmonary airway epithelial CD39 expression in IPAH patients compared with controls without airway disease.

Fig. 1.

Decreased CD39 expression on endothelium of pulmonary arteries and plexiform lesions from patients with idiopathic pulmonary arterial hypertension (IPAH). A: representative immunohistochemical images of lung tissue stained for CD31 and CD39 shows a marked decrease in endothelial CD39 in the IPAH patient (right) compared with control tissue from donors with healthy lungs (left). Images are representative of findings in lung tissue from 3 patients in each group. B: a representative plexiform lesion from a patient with advanced IPAH shows a characteristic network of CD31-positive, endothelial-lined vascular channels that lack CD39 expression. Lesion is representative of those found throughout the lungs of 3 IPAH patients studied. A quantitative comparison of CD39 endothelial staining in 5 IPAH and 5 donor controls is shown in C. Expression is unchanged in pulmonary respiratory epithelium. D: representative lung tissue sections from two IPAH patients and two control patients without lung disease show similar bronchial epithelial CD39 staining (arrows). Arrowheads identify small, peri-bronchial vessels without (IPAH patient 2) and with (Control patient 2) endothelial CD39 staining. Scale bar, 50 μm.

Global Deletion of Murine CD39

To further investigate the contribution of decreased CD39 to the pathobiology of PAH we generated CD39-deficient mice utilizing a ploxPFLpneo targeting vector. PCR of tail tips from wild-type (CD39^+/+) and CD39 knockout (CD39^−/−) mice confirmed the genotypes of each animal. qRT-PCR and Western blotting confirmed a lack of CD39 expression in multiple organs from CD39^−/− mice, including the lungs (data not shown). Taken together, these experiments confirm deletion of CD39.

CD39 Deletion Results in Severe Pulmonary Arterial Hypertension in Mice Exposed to Hypoxia

Representative actual waveform tracings shown in Fig. 2, A and B, illustrate the extreme differences in right ventricular systolic pressure (RVSP) and mean pulmonary artery (PA) pressure, respectively, in hypoxic CD39^−/− mice compared with CD39^+/+ mice. Under normoxic conditions, there was no difference in mean RVSP (Fig. 2C) and mean PA pressure (Fig. 2D) between CD39^+/+ and CD39^−/− mice. Following 4 wk of hypoxia (10% oxygen), CD39^−/− mice developed a statistically significant increase in both pressure parameters compared with hypoxic CD39^+/+, normoxic CD39^−/−, and normoxic CD39^+/+ mice. To assess the possible effects of differences in heart rate on intra-cardiac pressures, heart rates were obtained after administering isoflurane, but prior to the surgical procedure. There were no differences in heart rate between the groups (Table 1). Systemic blood pressures increased in response to hypoxia (Table 1), although there were no significant differences in pressure increases between the CD39^+/+ and CD39^−/− mice. These findings confirm that the extreme increase in pulmonary arterial pressure in CD39^−/− mice compared with CD39^+/+ mice was secondary to an increase in pulmonary vascular resistance, rather than an increase in heart rate or induction of disproportionate systemic hypertension. Arterial blood gas analysis (Table 1) showed a similar increase in hemoglobin and decrease in Po₂ in both CD39^+/+ and CD39^−/− mice, confirming that the extreme pulmonary arterial hypertension in CD39^−/− mice was not secondary to disproportionate erythrocytosis or hypoxemia.

Fig. 2.

CD39 gene deletion exacerbates hypoxia-induced pulmonary arterial hypertension. Representative, superimposed waveforms of right ventricular systolic pressures [RVSP (A)] and pulmonary arterial pressures (B) are shown for CD39^−/− (black) and CD39^+/+ (green) mice exposed to hypoxia. Graphs compare the mean RVSP (C) and mean pulmonary arterial pressures (D) for CD39^+/+ and CD39^−/− mice exposed to normoxia or hypoxia for 4 wk. N = 6 for each group in C and D, and values are presented as means ± SE. **P < 0.005.

Table 1.

Physiologic parameters for normoxic and hypoxic mice based upon genotype

	Normoxia			Hypoxia
	CD39+/+	CD39−/−	P	CD39+/+	CD39−/−	P
Heart rate, beats/min	445 ± 26	431 ± 31	ns	429 ± 19	472 ± 13	ns
Systolic blood pressure, mmHg	95 ± 4	96 ± 2	ns	112 ± 2	105 ± 4	ns
Diastolic blood pressure, mmHg	66 ± 1	71 ± 5	ns	83 ± 1	82 ± 3	ns
Mean arterial pressure, mmHg	86 ± 2	86 ± 2	ns	102 ± 2	97 ± 4	ns
Hemoglobin, g/dl	13 ± 1	15 ± 1	ns	25 ± 1	26 ± 1	ns
Po2, mmHg	113 ± 5	104 ± 4	ns	52 ± 8	49 ± 8	ns

Values are means ± SE. ns, not significant.

CD39 Deletion Exacerbates Hypoxic Pulmonary Arterial and Right Ventricular Remodeling

The effect of hypoxia on CD39 is shown in Fig. 3A where representative CD39 immunostaining shows a similar pattern in CD39^+/+ mice exposed to normoxia and hypoxia, and no staining in CD39^−/− mice. The effect of hypoxia on pulmonary arterial remodeling was determined by measuring the medial thickness of small, medium and large arteries. Representative photomicrographs of small pulmonary arteries from CD39^+/+ and CD39^−/− mice exposed to normoxia and hypoxia are shown in Fig. 3B. Immunolabeling with anti-αSMA shows a significantly thicker medial layer in hypoxic CD39^−/− mice, compared with the other groups. Analysis of pulmonary arterial medial thickness in pulmonary arteries of various sizes confirmed that hypoxia induced significantly greater medial thickness in the small (<50 μm) and medium (51–100 μm) pulmonary arteries of CD39^−/− mice compared with hypoxic CD39^+/+, normoxic CD39^−/−, and normoxic CD39^+/+ mice (Fig. 3C). It is important to note that normoxic CD39^−/− mice also had significantly increased medial thickness compared with normoxic CD39^+/+ mice.

Fig. 3.

CD39 deletion exacerbates pulmonary arterial remodeling and right ventricular hypertrophy in mice exposed to hypoxia. A: representative immunohistochemical staining of CD39 on small pulmonary arteries (arrows) and airway epithelium (arrowheads) in CD39^+/+, but not CD39^−/−, mice exposed to both normoxia and hypoxia. 60× magnification. Scale bars are equal to 30 μm. B: representative photomicrographs using α-SMA antibody and immunofluorescence labeling shows increased pulmonary arterial remodeling in hypoxic CD39^−/− mice (60× magnification). Scale bars are equal to 30 μm. C: differences in pulmonary arterial medial thickness were determined in small (<50 μm), medium (51–100 μm), and large (>100 μm) pulmonary vessels labeled with α-SMA antibody. For each vessel size category, the values presented are the mean ± SE using 10 fields from 6 mice from each experimental condition. D: comparison of right ventricular hypertrophy using Fulton's Index [ratio of RV:(LV + septum) weights in wild-type (CD39^+/+) and CD39 knockout (CD39^−/−) mice exposed to normoxia and hypoxia]. N = 6 for each group; values presented as means ± SE. *P < 0.05, **P < 0.005.

Hypoxia also induced greater right ventricular hypertrophy in CD39^−/− compared with CD39^+/+ mice (Fig. 3D), as shown by a statistically significant increase in Fulton's Index (the ratio of RV weight to the sum of the left ventricular plus septal weights) of hypoxic CD39^−/− mice, compared with the other groups.

CD39 Deletion Increases the Plasma ATP-to-Adenosine Ratio in Mice Exposed to Hypoxia

To investigate the mechanism by which CD39 deletion results in severe hypoxia-induced pulmonary hypertension, we measured intravascular concentrations of the purinergic signaling molecules ATP, ADP, AMP and adenosine. We hypothesized that global deletion of CD39 would alter nucleotides and adenosine both in the circulation and in the local milieu, creating an ATP-rich and adenosine-poor intravascular environment. Using a liquid chromatography-mass spectrometry (LC-MS) platform, we determined that hypoxic CD39^+/+ and CD39^−/− mice both exhibited significantly increased plasma ATP concentrations, compared with normoxic controls (Fig. 4A). Among the hypoxic groups, the ATP concentration was significantly higher in CD39^−/− mice, compared with wild-type mice (Fig. 4A). Hypoxic CD39^−/− mice were also found to have higher ADP concentrations compared with hypoxic wild-type controls (Fig. 4B). Both CD39^+/+ and CD39^−/− mice had a hypoxia-induced decrease in AMP concentrations, with a trend toward lower concentrations in both normoxic and hypoxic CD39^−/− mice, compared with wild-type controls (Fig. 4C). Normoxic wild-type and CD39^−/− mice had similar adenosine concentrations (Fig. 4D). In wild-type mice, hypoxia induced a significant increase in plasma adenosine. However, a significant decrease in adenosine concentration was noted in hypoxic CD39^−/− mice (Fig. 4D).

Fig. 4.

Circulating plasma nucleotide and adenosine concentrations are significantly altered in hypoxic CD39^−/− mice. Shown are plasma ATP (A), ADP (B), AMP (C), and adenosine (D) concentrations, and the ATP/AMP (E) and ATP/adenosine ratios (F), as determined by liquid chromatography-mass spectrometry. N = 6 per group. Values presented as means ± SE. *P < 0.05, **P < 0.005.

To assess the overall effect of CD39 gene deletion on a complex system that involves the sequential generation of products that become substrates for subsequent enzymatic reactions, two important ratios were compared. The ATP-to-AMP ratio compares the starting substrate (ATP) and final product (AMP) of CD39 ectonucleotidase activity. As shown in Fig. 4E, hypoxic CD39^−/− mice had a significantly elevated ratio compared with hypoxic wild-type, and both normoxic, controls. The ATP-to-adenosine ratio reflects the combined actions of CD39 and CD73, as the latter converts AMP to adenosine. Hypoxic CD39^−/− mice had strikingly higher plasma ATP-to-adenosine ratios compared with all other groups (Fig. 4F). Of note, hypoxic CD39^+/+ mice exhibited an increase in both plasma ATP and adenosine concentrations compared with their normoxic controls, which is consistent with prior studies (4, 13). Interestingly, hypoxic but not normoxic CD39^−/− mice exhibited a statistically significant increase in ATP and ADP compared with CD39^+/+ controls.

CD39 Deletion Results in Altered Lung Purinergic Receptor Expression

Given that intravascular nucleotides are ligands for purinergic receptors that regulate vascular tone, we profiled the transcription levels of major receptors in whole lung homogenates from each of the mouse groups (Tables 2, 3, and 4). Among the ligand-gated P2X purinoreceptors, hypoxia significantly upregulated the P2X1 receptor in CD39^+/+ (almost a 9-fold increase) and, to an even greater extent, CD39^−/− mice (a 17-fold increase), compared with normoxic controls (Table 2 and Fig. 5A). Immunostaining of the P2X1 receptor in murine pulmonary vessels (Fig. 5B) is consistent with an increase of this receptor in hypoxic CD39^−/− lungs. P2Y2 was the only G protein-coupled purinoreceptor significantly affected by hypoxia, and CD39^−/− mice exhibited a significant upregulation of this receptor compared with both hypoxic CD39^+/+ mice and normoxic controls (Table 3). The adenosine A2A was upregulated by hypoxia in both CD39^+/+ (a 2-fold increase) and CD39^−/− (a 6-fold increase) mice, compared with normoxic controls (Table 4).

Table 2.

Average mRNA fold-change of P2X receptors compared with CD39+/+ normoxic mice

	P2X Receptor, Average mRNA Fold Change vs. CD39+/+ Normoxic Mice
	P2X1	P2X2	P2X3	P2X4	P2X5
CD39+/+ normoxic	1.0	1.0	1.0	1.0	1.0
CD39−/− normoxic	1.4	1.2	1.1	1.0	1.0
CD39+/+ hypoxic	8.7*	1.0	1.1	0.9	1.3
CD39−/− hypoxic	17.1*	1.1	1.3	1.0	1.3

^*P < 0.05 for CD39 −/− hypoxic vs. CD39 +/+ hypoxic.

Table 3.

Average mRNA fold-change of P2Y receptors compared with CD39+/+ normoxic mice

	P2Y Receptor, Average mRNA Fold Change vs. CD39+/+ Normoxic Mice
	P2Y1	P2Y2	P2Y4	P2Y6	P2Y12
CD39+/+ normoxic	1.0	1.0	1.0	1.0	1.0
CD39−/− normoxic	0.9	1.0	1.1	1.0	1.0
CD39+/+ hypoxic	0.7	1.7*	1.3	0.8	0.8
CD39−/− hypoxic	1.6	3.3*	2.0	0.9	1.9

^*P < 0.05 for CD39 −/− hypoxic vs. CD39 +/+ hypoxic.

Table 4.

Average mRNA fold-change of adenosine receptors compared with CD39+/+ normoxic mice

	Adenosine Receptor, Average mRNA Fold Change vs. CD39+/+ Normoxic Mice
	A1	A2A	A2B	A3
CD39+/+ normoxic	1.0	1.0	1.0	1.0
CD39−/− normoxic	1.2	0.9	0.8	0.7
CD39+/+ hypoxic	1.1	2.2*	0.9	0.4
CD39−/− hypoxic	1.5	6.5*	0.9	0.6

^*P < 0.05 for CD39 −/− hypoxic vs. CD39 +/+ hypoxic.

Fig. 5.

P2X1 is upregulated in the lungs of CD39^−/− mice and treatment with apyrase or a P2X1 antagonist mitigates severe pulmonary hypertension. P2X1 receptor mRNA increases 8-fold in CD39^+/+ and 17-fold in CD39^−/− mice exposed to 4 wk of hypoxia (A). Representative P2X1 receptor (P2RX1) immunohistochemical staining in CD39^+/+ and CD39^−/− mice exposed to normoxia or hypoxia shows increased vessel P2RX1 expression in hypoxic CD39^+/+ and, to a greater extent, hypoxic CD39^−/− mice (B). The effects of 4 wk of continuously-administered saline control, apyrase (apyr) or the P2RX1 antagonist NF279 on right ventricular systolic pressure (RVSP) in the mouse groups indicated are shown in C. N = 5 for each group; values presented as means ± SE. *P < 0.05, **P < 0.005.

Reconstitution of Hypoxic CD39^−/− Mice with Soluble Apyrase Decreases, and Treatment with a P2X1 Receptor Antagonist Prevents, Severe Pulmonary Arterial Pressure Elevations

Based upon our findings of increased intravascular ATP and ADP, significant upregulation of the P2X1 receptor, and severe pulmonary hypertension in hypoxic CD39^−/− mice, we designed two sets of rescue experiments aimed at reversing the PH phenotype. The first rescue experiment involved the reconstitution of intravascular ectonucleotidase activity. Soluble potato tuber apyrase has known ATPase and ADPase activity, and has been used in prior studies involving CD39^−/− mice (23, 40). Sequence homology of this apyrase is similar to human and murine CD39 (18). As shown in Fig. 5C, continuous subcutaneous administration of soluble apyrase during 4 wk of hypoxia significantly decreased the RVSP in normoxic CD39^+/+ and hypoxic CD39^−/− mice, compared with control mice that received continuous saline infusions. Of note, apyrase did not decrease the RVSP in hypoxic CD39^−/− mice to levels found in normoxic CD39^+/+ mice, indicating that the conversion of ATP and ADP to downstream products such as AMP and adenosine alone was not sufficient to completely reverse the PH phenotype. The second rescue experiment involved the continuous infusion of NF279, a potent and selective P2X1 antagonist, during 4 wk of hypoxia. As shown in Fig. 5C, NF279 prevented the elevation of pulmonary arterial pressures in all groups, including hypoxic CD39^−/− mice.

Pulmonary Vessels from Patients with IPAH Have Increased P2X1 Receptors

Our mouse experiments implicated increased pulmonary vascular P2X1 receptors in the pathobiology of experimental PH. To determine the significance of this finding in humans, we performed immunohistochemical staining for the P2X1 receptor in human lung sections. As shown in Fig. 6, pulmonary vessels from 2 patients with IPAH exhibit increased P2X1 receptor staining in plexiform lesions, remodeled vessels with medial hypertrophy, and in larger vessels compared with individuals without lung disease.

Fig. 6.

Increased pulmonary vascular P2X1 receptor expression in IPAH patients. Representative lung tissue sections from two IPAH patients show increased medial P2X1 staining in plexiform lesions, muscularized vessels, and large pulmonary arteries compared with vessels from control patients without pulmonary disease. Size bars equal 50 microns.

DISCUSSION

In this report we show, for the first time, that deletion of CD39 in vivo results in high pulmonary arterial pressures, significant pulmonary artery remodeling, and right ventricular hypertrophy. Our investigations into the mechanism responsible for this impressive pulmonary hypertension phenotype implicate altered intravascular purinergic signaling molecules and cognate receptors. Most notably, we identify marked increases in both circulating intravascular ATP and cognate lung P2X1 receptors in our model, which is a previously unidentified mechanism contributing to pathobiology of pulmonary hypertension. Furthermore, reconstitution of CD39^−/− mice using soluble apyrase with ATPase and ADPase activities lowered pulmonary arterial pressures to levels seen in hypoxic wild-type mice, and treatment with a potent and specific P2X1 receptor prevented the development of pulmonary hypertension altogether. Taken together, these findings support a role for altered CD39 activity, a resultant shift towards a high ATP and low adenosine intravascular environment, and an increase in the P2X1 receptor in the development of PH. We also identify both a decrease in CD39 and an increase in the P2X1 receptor in pulmonary vessels from patients with IPAH, making our findings relevant to this human disease.

This is the first in vivo study to demonstrate impressive alterations in a full panel of circulating plasma nucleotide and adenosine concentrations as a direct result of targeted deleting of the CD39 gene. Our methods included the use of a LC-MS platform to analyze plasma drawn directly into a “stop solution” formulated to minimize the interconversion of ATP, ADP, AMP, and adenosine. This proved to be an accurate and reproducible method for studying plasma ATP and its metabolites, paving the way for future animal and human studies involving the role of nucleotides/adenosine regulation in pulmonary arterial hypertension. The role of intravascular nucleotides and adenosine as extracellular signaling molecules is well-established (6). ATP is released in both lytic and nonlytic manners from the major blood vessel wall cell types, including endothelial cells (3), vascular smooth muscle cells (26), perivascular sympathetic nerves (28), and erythrocytes (2). Our finding of a hypoxia-induced increase in circulating plasma ATP in CD39^+/+ wild-type mice is consistent with prior studies in which hypoxia has been shown to increase luminal ATP concentrations with contributions from circulating erythrocytes (2) and endothelium (3). Our finding of an extreme elevation of plasma ATP (and a lesser although significant increase in ADP) in hypoxic CD39^−/− mice was surprising. These findings indicate that the lack of the ectoenzyme CD39 results in a large buildup of upstream substrates (ATP and ADP) and dearth of downstream products (adenosine and a trend toward decreased AMP).

While CD39 is the major extracellular ectonucleotidase (37), it is important to note that other enzymes such as the ectonucleotide pyrophosphatase/phosphodiesterase (E-NPP) family and alkaline phosphatases are capable of hydrolyzing ATP, while adenylate kinase, nucleoside diphosphate kinase and ATP synthase can regenerate ATP (51). In addition, the concentration of intravascular nucleotides can be altered by cell lysis, release channels, transporters and exocytosis (51). These other mechanisms may explain the finding of increased circulating ATP and ADP in hypoxic, but not normoxic, CD39^−/− mice (Fig. 2, A and B). We hypothesize that this multifaceted system is capable of maintaining nucleotide homeostasis in normoxic conditions despite a lack of CD39, but hypoxia overwhelms these mechanisms resulting in the high ATP and ADP concentrations noted in our study. Future studies are needed to assess the specific effects of CD39 deletion on these other enzymes.

Based upon the finding of an ATP-rich and adenosine-poor intravascular environment in hypoxic CD39^−/− mice, we hypothesized that these changes would perturb purinergic signaling homeostasis, leading to increased pulmonary arterial pressures. The intravascular nucleotides and adenosine are important signaling molecules that regulate the cardiovascular system by acting upon P1 (adenosine) and P2 receptors (ATP and ADP). The overall contribution of purinergic signaling to vascular tone results from the integration of a variety of complex vasoconstrictive and vasodilatory signals. For example, ATP causes vasoconstriction by activation of the P2X₁ receptor in vascular smooth muscle cells (SMCs), but triggers nitric-oxide-mediated vasodilatory effects via activation of endothelial P2Y₁ and P2Y₂ receptors (6). In addition to its role as a potent mediator of platelet activation and thrombosis, ADP has been shown to increase pulmonary vascular resistance in the setting of hypoxia (27). Adenosine is capable of triggering both pulmonary vasodilation [through A_2A(50) and A_2B receptor stimulation (21, 36)] and vasoconstriction (through stimulation of the A₁ receptor leading to thromboxane A₂ release) (43). Thus purinergic regulation of vascular tone is complex, and the net effect upon the pulmonary vasculature results from the sum total input from multiple receptor subtype activation by the circulating nucleotide/adenosine pool. For this reason, our in vivo model gave us the optimal means of assessing the true physiological consequences of perturbing purinergic signaling through CD39 deletion.

The continuous infusion of soluble apyrase, which cleaves inorganic phosphate from ATP and ADP, provided partial protection against the development of pulmonary hypertension in CD39^−/− mice exposed to 4 wk of hypoxia. This finding of a protective role for a substance with both ATPase and ADPase activity (like CD39) against the development of PAH sheds light on our previous discovery of increased functional CD39 on circulating plasma microparticles in patients with PAH (49). We now hypothesize that microparticle-based CD39 in humans is increased as a compensatory response to mitigate PAH. Our findings are congruent with previous studies that have shown CD39 to be protective in hypoxic environments contributing to other disease states (14, 20). Interestingly, continuous administration of apyrase did not completely protect CD39^+/+ mice from developing pulmonary hypertension (Fig. 5C). This finding suggests that decreasing the buildup of ATP and ADP alone was not sufficient to completely prevent pulmonary hypertension, and encouraged us to also consider the effects of combined CD39 deletion and hypoxia on purinergic receptors.

Our purinergic receptor transcriptomic profiling of lungs from CD39^+/+ and CD39^−/− mice identified significant upregulation of three receptors in response to hypoxia. The most significant change was in the P2X1 receptor, which showed a nearly 9-fold increase in hypoxic wild-type mice and a 17-fold increase in hypoxic CD39^−/− mice. These marked elevations provide mechanistic insight into the role of purinergic signaling in the development of pulmonary hypertension. The P2X1 receptor is expressed at high levels in vascular smooth muscle and platelets, and to lesser degrees in the heart and inflammatory cells (15). Prior studies have identified a role for a P2X1-receptor-mediated vasoconstriction in the regulation of afferent renal arterioles (24), mesenteric arteries, vas deferens (34), and urinary bladder (48). P2X1 receptor mRNA has been identified in the smooth muscle of adult rat pulmonary arteries (35), and a prior study suggested that the P2X1 receptor likely mediated a small vasoconstrictive response in isolated adult porcine pulmonary arteries in response to the nonhydrolyzable ATP analog α,β-meATP (33). To our knowledge, ours is the first study to utilize an in vivo model to identify a major role for P2X1 receptor and the ATP ligand in the pathogenesis of pulmonary hypertension. The potent and selective P2X1 antagonist NF279 (10, 39) protected both CD39^+/+ and CD39^−/− mice from the development of hypoxic pulmonary hypertension. This finding emphasizes the importance of the P2X1 receptor in the development of hypoxic pulmonary hypertension and supports a disease paradigm in CD39 deletion that results in upregulation of both the P2X1 receptor and its cognate ligand (ATP).

In addition to a marked upregulation of the P2X1 receptor, our profiling also identified a significant increase in the pulmonary A2A receptor (Table 4) and a more modest increase in the pulmonary P2Y2 receptor (Table 3) in response to hypoxia. In both cases, the mRNA was highest in the lungs of CD39^−/− mice. A prior study found that deletion of the A2A receptor resulted in pulmonary hypertension in an experimental model (50), and another study speculated that activation of the A2A receptor was a theoretical mechanism for the reversal of PH in a monocrotaline rat model (1). Thus it is likely that the increase in the A2A receptor in our model represents a compensatory mechanism. P2Y2 is thought to play a role in experimental models of systemic hypertension (29), but its role in the pathobiology of pulmonary hypertension has not been elucidated. Additional studies are needed to more fully characterize the contributions of the P2Y2 and A2A receptors in both experimental and human pulmonary hypertension.

Based upon the findings in our mouse model we explored a possible role for dysregulated purinergic signaling in the pathobiology of human PAH using lung sections from patients with idiopathic pulmonary arterial hypertension (IPAH) and sex- and gender-matched controls. Immunohistochemical staining of vascular CD39 showed a decrease in IPAH lungs compared with control lungs. These findings are congruent with a recent study that showed attenuated pulmonary arterial endothelial CD39 in patients with IPAH (22). However, we have expanded upon these findings using our in vivo model, which identifies a likely mechanism by which decreased CD39 leads to the development of pulmonary hypertension as a result of increases in both the P2X1 receptor and its cognate ATP ligand. This is relevant to human IPAH, as immunohistochemical staining of human IPAH lungs showed increased vascular P2X1 staining, compared with control lungs. It is important to note that systemic pressures were not affected by CD39 deletion in the mouse, but pulmonary arterial pressures increased markedly in the setting of hypoxia. This specificity is also present in humans with pulmonary arterial hypertension, who do not exhibit associated systemic hypertension even in the setting of severe elevations of the pulmonary arterial pressures. Thus our model recapitulates this common presentation of the human disease.

Our study has limitations. No animal model of pulmonary hypertension has been shown to fully recapitulate human pulmonary arterial hypertension, although transgenic mouse models have been shown to be helpful in the study of pathways related to the pathobiology of PAH (11).

Like most other murine models of PH (16), lung remodeling in CD39-deleted mice was limited to medial hypertrophy. However, it is important to note that, unlike most other murine models, our gene deletion resulted in significantly increased medial thickness even in normoxia (Fig. 3).This suggests that CD39 plays a protective role against the initiation of a progressive pulmonary vasculopathy even in normoxic conditions, possibly due to maintaining homeostasis of circulating nucleotide concentrations or the P2X1 receptor. This, along with our other findings, indicate that future studies are needed to explore the role of altered CD39 activity in human IPAH.

In summary, this study supports a novel mechanism by which severe hypoxia-induced PH develops as a result of CD39 ectonucleotidase dysregulation, a resulting shift in intravascular nucleotide and nucleoside concentration toward an ATP/ADP-rich and AMP/adenosine-poor milieu, and a significant upregulation of pulmonary vascular P2X1 receptor. Reconstitution of ATPase and ADPase activity lessened the degree to which pulmonary arterial pressures increased in this model, while blockade of the P2X1 receptor prevented an increase in pulmonary arterial pressures altogether. The findings of decreased CD39 and increased P2X1 receptor expression in pulmonary vessels from IPAH patients support dysregulated purinergic signaling as a potential mechanism contributing to this disease. The potential therapeutic benefits of soluble CD39 delivery and/or P2X1 receptor antagonism in IPAH should be explored.

GRANTS

This work was supported by National Institutes of Health (NIH) Grants 1K-23HL-119623 (S. H. Visovatti), 1K-08HL-123621 (S. Goonewardena), R01-HL-127151 (D. J. Pinsky), NS-087147 (D. J. Pinsky), NIH T32 Training Grants (S. Visovatti, Y. Kanthi, S. Goonewardena), and the Taubman Medical Research Institute of the University of Michigan (D. J. Pinsky) and the Ruth Professorship (D. J. Pinsky). Funding for the PHBI was provided by the Cardiovascular Medical Research and Education Fund (CMREF). This work utilized metabolomics core services supported by NIH Grant DK-097153 to the University of Michigan.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

S.H.V., M.C.H., S.N.G., Y.K., C.F.B., and D.J.P. conception and design of research; S.H.V., M.C.H., A.C.A., Y.K., P.R., J.W., and R.R. performed experiments; S.H.V., S.N.G., A.C.A., Y.K., P.R., J.W., D.P.-D., R.R., C.F.B., and D.J.P. analyzed data; S.H.V., M.C.H., S.N.G., A.C.A., Y.K., D.P.-D., C.F.B., and D.J.P. interpreted results of experiments; S.H.V. prepared figures; S.H.V. and D.J.P. drafted manuscript; S.H.V., M.C.H., S.N.G., A.C.A., Y.K., P.R., J.W., D.P.-D., R.R., C.F.B., and D.J.P. edited and revised manuscript; S.H.V., M.C.H., S.N.G., A.C.A., Y.K., P.R., J.W., D.P.-D., R.R., C.F.B., and D.J.P. approved final version of manuscript.