Functional NMDA receptors are expressed by human pulmonary artery smooth muscle cells

N-methyl-d-aspartate (NMDA) receptors are widely expressed in the central nervous system. However, their presence and function at extraneuronal sites is less well characterized. In the present study, we examined the expression of NMDA receptor subunit mRNA and protein in human pulmonary artery (HPA) by quantitative polymerase chain reaction (PCR), immunohistochemistry and immunoblotting. We demonstrate that both GluN1 and GluN2 subunit mRNAs are expressed in HPA. In addition, GluN1 and GluN2 (A–D) subunit proteins are expressed by human pulmonary artery smooth muscle cells (HPASMCs) in vitro and in vivo. These subunits localize on the surface of HPASMCs and form functional ion channels as evidenced by whole-cell patch-clamp electrophysiology and reduced phenylephrine-induced contractile responsiveness of human pulmonary artery by the NMDA receptor antagonist MK801 under hypoxic condition. HPASMCs also express high levels of serine racemase and vesicular glutamate transporter 1, suggesting a potential source of endogenous agonists for NMDA receptor activation. Our findings show HPASMCs express functional NMDA receptors in line with their effect on pulmonary vasoconstriction, and thereby suggest a novel therapeutic target for pharmacological modulations in settings associated with pulmonary vascular dysfunction.

www.nature.com/scientificreports/ one cell type expressing NMDA receptors in lung 10 . Activation of NMDA receptors on these cells leads to calcium release and airway contraction 10 . Analogous to the effect on neurons, excessive activation of NMDA receptors in lung triggers acute nitric oxide-dependent injury 11,12 , suggesting NMDA receptor-mediated excitotoxicity exists outside the central nervous system. We previously demonstrated that NMDA receptor activation is linked to tissue plasminogen activator (tPA)-mediated inhibition of pulmonary arterial contractility and induction of vascular permeability 13,14 . However, whether and where functional NMDA receptors are expressed in pulmonary artery has not been established. In the present study, we examined the expression of NMDA receptor mRNA and protein in human pulmonary artery (HPA) using real-time reverse transcription-polymerase chain reaction (RT-PCR) and immunohistochemistry followed by analysis of receptor function using electrophysiology and contractile response.

NMDA receptor expression by human pulmonary artery smooth muscle cells in vivo and in vitro.
To search for direct evidence of NMDA receptor expression within human pulmonary artery (HPA), we first performed RT-PCR using subunit-specific primers. Human brain was used as a positive control. mRNAs for both GluN1 and GluN2 subunits (GluN2A, GluN2B, GluN2C and GluN2D) were detected in HPA (Fig. 1). Of interest, the mobility of GluN2B cDNA differed between HPA (about 300 bp) (n = 3) and human brain (about 400 bp). Such difference was also detected in human pulmonary artery smooth muscle cells (HPASMCs) from a different donor ( Supplementary Fig. S1). In addition, a primer targeting a different region 15 of GluN2B amplified GluN2B mRNA in human brain but failed to amplify GluN2B mRNA in HPA (Data not shown). These results suggest that a splice variant of GluN2B is expressed in human pulmonary artery. We also performed immunohistochemistry on human lung tissue using validated NMDA receptor subunitspecific antibodies (Supplemental Fig. S2-S5). Antibody to PECAM-1 (Antibody validation: Supplemental  Fig. S6), a marker of vascular endothelium, revealed intense staining of the interior surface of pulmonary artery ( Fig. 2A). In contrast, immunostaining of GluN1 was associated with HPASMCs localized in smooth muscle cell layer (Fig. 2B). Strong immunoreactivity was also detected for all four subunits of GluN2 in HPASMCs ( Fig. 2C-F), but not with secondary antibody alone (Fig. 2G). These results indicate that both GluN1 and GluN2 subunits are expressed in vivo by HPASMCs.
As an independent approach to study NMDA receptors expression, we performed immunofluorescence staining of cultured HPASMCs. Staining of permeabilized PASMCs demonstrated a network-like immunoreactivity for both GluN1 and GluN2 (A-D) subunits throughout the cell body ( Fig. 3) (Antibody validation: Supplemental Figs. S4, S5, S7, S8). Immunoreactivity was not observed with secondary antibody alone (Fig. 3).
To examine whether full length NMDA receptors are expressed, immunoprecipitation was performed on cultured HPASMCs with GluN1/GluN2 (A-D) subunit-specific antibodies. Full length GluN1 protein (120 kDa) was detected by immunoprecipitation using anti-GluN1 antibody to the intracellular domain (amino acids 660-811): the protein migrated at the same molecular weight as GluN1 protein from rat brain (Fig. 4A). No GluN1 immunoreactivity was detected with the IgG control (Fig. 4A). Similarly, addition of subunit-specific anti-GluN2A, anti-GluN2B, anti-GluN2C and anti-GluN2D antibodies immunoprecipitated GluN2A (MW = 180 kDa) (Fig. 4B), GluN2B (MW = 180 kDa) (Fig. 4C), GluN2C (MW = 140 kDa) (Fig. 4D) and GluN2D (MW = 150 kDa) (Fig. 4E), respectively, whereas no immunoreactivity was detected in the IgG control lanes. These results indicate that full length NMDA receptors are expressed by HPASMCs in culture. GluN2 subunits on cell surface is required to form fully functional channels. Therefore, we next examined surface localization of NMDA receptors on the plasma membrane of unpermeabilized HPASMCs using antibodies against the amino-terminal extracellular epitope of GluN1 and GluN2 (B, D) subunits (Antibody validation: Supplemental Figs. S3, S5, S10). Punctate immunostaining for GluN1 and GluN2 (2B, 2D) was observed on the cell surfaces of HPASMCs (Fig. 5A,B). As only heteromeric complexes including both GluN1 and GluN2 subunit are targeted to plasma membrane, this finding suggests that GluN1 and GluN2 subunits co-localize on plasma membrane of HPASMCs. Consistent with this inference, the punctate appearance of surface-labeled GluN1 overlapped with that of surface-labeled GluN2B (Fig. 5A) and GluN2D (Fig. 5B), respectively. Quantitative analysis of the distribution of GluN1 and GluN2 (2B, 2D) demonstrated that GluN1 is highly associated with GluN2 subunits (Fig. 5C,D) (Pearson's correlation coefficient: r = 0.779 and r = 0.859 for GluN2B and GluN2D, respectively), confirming that GluN1 and GluN2 subunits co-localize on the surface of HPASMCs. We next assessed if NMDA receptors on the surface of cultured HPASMCs are functional, we used whole-cell patch-clamp electrophysiology to record currents evoked by the application of NMDA and glycine in a Mg-free recording solution. Application of 100 μM NMDA and 10 μM glycine to cells clamped at − 50 mV evoked an inward whole-cell current of 10.9 ± 1.7 pA (mean ± SE, p < 0.01 by t-test) (Fig. 5E,F) (n = 16). The area under curve (AUC) is increased by 39% over baseline with NMDA/glycine application (p < 0.05 by t-test) (Fig. 5G) (n = 16). To confirm this was mediated by NMDA receptor activation, we added the NMDA receptor antagonist AP5 (50 μM) to the cells. In the presence of AP5, application of NMDA and glycine failed to evoke an inward whole-cell current (n = 2) (data not shown), indicating this NMDA-evoked current in HPASMCs was mediated by NMDA receptor activation and that pulmonary arterial NMDA receptors are indeed functional. NMDA receptor currents in HPASMCs were small compared with NMDA receptor-evoked currents in rat hippocampal neurons (about 200 pA) 16,17 . Serine racemase and VGLUT1 expression in HPASMCs. NMDA receptors are activated by glutamate and glycine, which bind to the GluN2 and GluN1 subunits, respectively. d-serine, another co-agonist acting at the glycine site, is generated from l-serine by serine racemase. To search for the potential endogenous co-agonist for NMDA receptor activation in HPASMCs, we stained cultured HPASMCs with an antibody against serine www.nature.com/scientificreports/ racemase (Antibody validation: Supplemental Fig. S11). Immunoreactivity for serine racemase was observed in HPASMCs (Fig. 6A). Serine racemase protein (37 kDa) was also detected on western blots of whole cell lysates from HPASMCs cultures (Fig. 6B) at the same molecular weight as serine racemase from rat brain neuronal lysates. This indicates that d-serine might be synthesized in HPASMCs and function as a potential local endogenous source for NMDA receptor activation. We also stained cultured HPASMCs with antibody against vesicular glutamate transporter 1 (VGLUT1) (Antibody validation: Supplemental Fig. S12), which is responsible for vesicular storage of glutamate and subsequent release through exocytosis in both neuronal cells and endocrine cells 18,19 . Immunoreactive VGLUT1 was observed in HPASMCs (Fig. 6C). VGLUT1 protein (62 kDa) was also detected in whole cell lysates of HPASMCs (Fig. 6D), indicating that glutamate might be released from PASMCs and act with D-serine to activate NMDA receptors in an autocrine manner.
NMDA receptors regulate contractility of pulmonary arteries in response to vasoconstrictor phenylephrine. Lastly, we investigated the functional significance of activating pulmonary arterial SMC NMDA receptors. We previously reported that the vasoreactivity of isolated pulmonary artery depends on activation of NMDA receptors 13,14 . In order to study the involvement of NMDA receptors in regulation of vasoreactivity of pulmonary vasculature in the context of the surrounding lung parenchyma, we measured the contraction of pulmonary arteries in human precision cut lung slices (hPCLS), as previously described 20,21 . Contraction was measured in response to the vasoconstrictor phenylephrine in the absence or presence of NMDA receptor antagonist MK801 (10 μM). Phenylephrine induced pulmonary arterial contraction in a dose-dependent manner under both normoxic and intermediate duration hypoxic conditions ( To confirm the effect of NMDA receptors in pulmonary vasoconstriction, mouse precision cut lung slices were treated with endothelin-1 (ET-1) (10 -9 -10 -7 M), a potent endogenous vasoconstrictor. While ET-1 initiated

Discussion
In the present study, we demonstrate that both GluN1 and each of the four GluN2 (2A-D) subunit proteins are expressed by HPASMCs. GluN1 and GluN2 subunits localize on the surface of HPASMCs and form functional receptors as evidenced by the whole-cell patch-clamp electrophysiology analysis performed in cultured HPASMCs ( Fig. 5E-G) and by the difference in the maximal extent of contraction of pulmonary arteries measured ex vivo in hPCLSs in response to phenylephrine in the absence or presence of NMDA receptor antagonist MK801 under hypoxic conditions (Fig. 7). As NMDA receptor activation leads to the influx of calcium, which regulates vascular smooth muscle cell contractility, our results also suggest a potential role of NMDA receptor activation in pulmonary vasoconstriction under pathological conditions such as hypoxia.
The presence and activity of NMDA receptors in the lung has been demonstrated previously [9][10][11][12]22 . Excessive activation of NMDA receptors in the lung causes excitotoxicity leading to acute injury accompanied by disruption of pulmonary arterial contractility leading to increased permeability 11 , similar to findings in the CNS in which NMDA receptors-mediated excitotoxicity in neurons caused by pathologically high levels of glutamate leads to ischemia and neurodegeneration. Moreover, we previously reported that urokinase plasminogen activator, at pathological concentrations found in acute lung injury, inhibits pulmonary arterial contractility and promotes pulmonary vascular permeability through processes involving docking to NMDA receptors on pulmonary vascular smooth muscle cells 14 . Recent studies also point to the contribution of NMDA receptors in the pathophysiology of pulmonary arterial hypertension through mechanism involving HPASMC proliferation and migration 22,23 . The findings in this paper support these observations by providing physical evidence that functional NMDA receptors are expressed on pulmonary artery smooth muscle cells.
The precise combination of NMDA receptor GluN1 subunits with the various GluN2 subunits determines the biophysical and pharmacological properties of the NMDA receptors 24 . For example, GluN1/GluN2A heterodimers possess faster onset and shorter decay times than GluN1/GluN2B and GluN1/GluN2C channels, whereas GluN2D-containing NMDA receptors have the slowest decay times [24][25][26] . There is also evidence for the co-assembly of GluN2A/GluN2B, GluN2A/GluN2C, GluN2B/GluN2D, and other combinations in different regions of the brain and neuronal populations 24 , although the functional output and biophysical properties of these "mixed" channels are not understood in depth. GluN2 subunits are heterogeneously expressed in different regions of the brain. Here, we detected all four types of GluN2 subunits in pulmonary artery smooth muscle www.nature.com/scientificreports/ cells, implying the possible presence of multiple combinations of GluN1 and GluN2 subunits. Whether there is regional variation in subunit expression along the vasculature and in disease states will require further study. NMDA receptor function heterogeneity is determined by the C-terminal domain of GluN2 subunits interacting with various signaling proteins. The C-terminal domain is also subject to post-translation modification such as phosphorylation, which can potentiate NMDA receptor function 27,28 . The phosphorylation state also controls the internalization rate of NMDA receptors 29 , leading to altered NMDA receptor numbers on cell surfaces. Interestingly, hypoxia-ischemia in neonatal brain increases tyrosine phosphorylation of NMDA receptors, which correlates with enhanced association of Src protein-tyrosine kinase with NMDA receptors 30 . Whether such events contribute to the vasoconstriction under hypoxic conditions observed here remains to be investigated. Further investigations will also be needed to see if the presence of multiple GluN2 subunits in PASMCs diversifies the function of NMDA receptors through interactions with different signaling molecules, thus allowing pulmonary artery to be finely regulated in different physiological or pathological conditions. Consistent with the previous reports that the localizations of d-serine and serine racemase are in close vicinity to NMDA receptors 31 , we found high levels of serine racemase in PASMCs, suggesting that d-serine could function as a co-agonist for NMDA receptors in PASMCs. As previously described 22 , we also observed expression of VGLUT1 in PASMCs, which confers a glutamatergic phenotype in neurons 19 . Previous findings have also identified VGLUT1 expression outside the brain in osteoblasts from which glutamate is released in a regulated manner 32 . The presence of VGLUT1 in PASMCs provides strong indirect evidence for presynaptic glutamatergic signaling events in PASMCs and also suggests that glutamate might mediate the intercellular communications between non-neuronal cells in an autocrine and/or paracrine manner. In addition to release from PASMCs 22 , glutamate in the blood stream may be taken up by pulmonary arterial endothelial cells (PAECs) through sodium-dependent or sodium-independent transporters 33 , from where it can be released and exert its action on NMDA receptors on adjacent PASMCs. Interestingly, hypoxia increases the rate of glutamate uptake Deficiency of smooth muscle cell specific NMDA receptors attenuates hypoxic vascular remodeling and pulmonary hypertension through mechanisms involving decreased endothelial dysfunction, resistance to apoptosis and perivascular inflammation 22 . Our finding that NMDA receptor blockade reduces vasoconstriction under hypoxic conditions might suggest both a role in pulmonary hypertension and a beneficial effect of NMDA receptor blockade on ventilation/perfusion mismatch in acute hypoxia. The differential effects of NMDA receptor blockade on human and murine pulmonary vasoconstriction under normoxic conditions may signify species difference in NMDA receptor expression levels, post-translational modification and/or the extent of glutamate release. This inference may also be applicable to hypoxia [34][35][36][37] , as NMDA receptor blockade attenuates pulmonary vasoconstriction under hypoxia but has no effect under normoxia. In addition, hypoxia regulates endothelin-1, a potent vasoconstrictor released by endothelial cells 38,39 . Activation of endothelin-1 receptor increases calciumdependent glutamate release from PASMCs and NMDA receptor phosphorylation in PASMCs 22 , highlighting the crosstalk between endothelin-1 receptor and NMDA receptor upon pulmonary vasoconstriction. NMDA receptor activation may represent a common pathway downstream of pulmonary vasoconstrictors. Future study will be needed to delineate signaling mechanisms between these receptors in health and disease states.
In summary, functional NMDA receptors are expressed by PASMCs. Activation of NMDA receptors can regulate pulmonary vasoconstriction. Characterization of number, distribution and regulation of pulmonary vascular NMDA receptors may provide insight into disorders associated with increased and impaired pulmonary vascular contractility, e.g. sepsis. Targeting NMDA receptor function may also represent a critical locus for preventing ventilation/perfusion mismatch, a site where acute lung injury is initiated in diverse settings and a potential locus for intervention to remediate dysregulation of vascular tone and permeability [40][41][42] .

Study approval. Postmortem human lung tissue was commercially obtained from the National Disease
Research Interchange (NDRI, Philadelphia, PA) and the International Institute for the Advancement of Medicine (IIAM, Edison, NJ). The study protocol on human and murine lung tissue was approved by the Rutgers University Institutional Review Board and the Institutional Animal Care and Use Committee of the University of Pennsylvania, respectively. All experiments were performed in accordance with standard regulations.

Real-time reverse transcription-polymerase chain reaction (RT-PCR).
HPA total RNA was purified using Qiagen's RNeasy Mini kit (Qiagen, Germantown, MD) and then subject to RT-PCR using OneStep RT-PCR kit (Qiagen, Germantown, MD). 2 μg total RNA were used per reaction. Human brain RNA (Agilent, Cedar Creek, TX) was used as a positive control. The primer sequences for GluN2A, GluN2C and GluN2D were used as described 15   . In all experiments, HPASMCs were studied before passage 6 prior to any change in cell morphology.

Western blot.
Western blot was carried out as previously described 43  Human precision cut lung slices preparation and measurement of vasoconstriction. Human precision cut lung slices (hPCLS) were prepared and vasoconstriction was measured as previously described 21 , with modifications for vessels (instead of airways). Briefly, whole human lungs from 26 non-diseased donors (Supplemental Table S1 including all the donor information) were inflated using 2% (wt/vol) low melting point agarose and cooled to solidify the agarose. From the solidified lobes, tissue cores (8 mm in diameter) containing a small vessel were prepared and sliced at a thickness of 350 µm using a VF300 vibratome (Precisionary, Greenville, NC). The slices were maintained in supplemented Ham's F12 medium at 37 °C in a humidified air-CO 2 (95-5%) incubator, with 3 medium changes during the 48 h. Adjacent slices containing contiguous segments of the same vessel served as controls for the experimental treatments. Following the wash period, the slices were incubated in normoxic (21% O 2 ) or intermediate duration hypoxic (5% O 2 ) conditions. To assess the involvement of NMDA receptor in regulation of vasoreactivity of pulmonary artery, slices were preincubated with MK801 (10 µM) for one hour and then vascular constriction were measured in the presence of varying concentrations of phenylephrine (10 -8 -10 -4 M).
To measure vasoconstriction, the vessel in each slice was visualized under a microscope (Nikon Eclipse, × 40 magnification) and images were captured using an in-line camera (Evolution QEi, Silver Spring, MD). Luminal areas of the imaged vessels were quantified using Image-Pro Plus Software (version 6, Media Cybernetics, Silver Spring, MD) and represented in units of square micrometers. The luminal area at each concentration of phenylephrine was obtained and constriction of the vessels was calculated as a percentage of baseline luminal area. Unpaired Student's t-tests were performed to determine statistical significance at the p < 0.05 confidence level when comparing different treatment groups.

Data availability
All data generated or analysed during this study are included in this published article (and its Supplementary  Information files).