Endothelial pannexin 1–TRPV4 channel signaling lowers pulmonary arterial pressure in mice

Pannexin 1 (Panx1), an ATP-efflux pathway, has been linked with inflammation in pulmonary capillaries. However, the physiological roles of endothelial Panx1 in the pulmonary vasculature are unknown. Endothelial transient receptor potential vanilloid 4 (TRPV4) channels lower pulmonary artery (PA) contractility and exogenous ATP activates endothelial TRPV4 channels. We hypothesized that endothelial Panx1–ATP–TRPV4 channel signaling promotes vasodilation and lowers pulmonary arterial pressure (PAP). Endothelial, but not smooth muscle, knockout of Panx1 increased PA contractility and raised PAP in mice. Flow/shear stress increased ATP efflux through endothelial Panx1 in PAs. Panx1-effluxed extracellular ATP signaled through purinergic P2Y2 receptor (P2Y2R) to activate protein kinase Cα (PKCα), which in turn activated endothelial TRPV4 channels. Finally, caveolin-1 provided a signaling scaffold for endothelial Panx1, P2Y2R, PKCα, and TRPV4 channels in PAs, promoting their spatial proximity and enabling signaling interactions. These results indicate that endothelial Panx1–P2Y2R–TRPV4 channel signaling, facilitated by caveolin-1, reduces PA contractility and lowers PAP in mice.


Introduction
The pulmonary endothelium exerts a dilatory influence on small, resistance-sized pulmonary arteries (PAs) and thereby lowers pulmonary arterial pressure (PAP). However, endothelial signaling mechanisms that control PA contractility remain poorly understood. In this regard, pannexin 1 (Panx1), which is expressed in the pulmonary endothelium and epithelium (Navis et al., 2020), has emerged as a crucial controller of endothelial function (Begandt et al., 2017;Good et al., 2015). Panx1, the most studied member of the pannexin family, forms a hexameric transmembrane channel at the cell membrane that allows efflux of ATP from the cytosol (Bao et al., 2004;Lohman et al., 2012). Previous studies indicated that flow/shear stress increases ATP efflux through Panx1 in EC monolayers (Wang et al., 2016). Endothelial Panx1 (Panx1 EC ) has also been linked to inflammation in pulmonary capillaries (Sharma et al., 2018). Beyond this, however, the physiological roles of Panx1 EC in the pulmonary vasculature are largely unknown. eATP levels, suggesting that TRPV4 EC channels do not regulate Panx1 EC activity under basal conditions. Although eATP levels were also reduced in PAs from inducible, smooth muscle cell-specific Panx1 cKO (Panx1 cKO-SMC)  mice, the eATP levels in these mice were higher than Panx1 cKO-EC mice ( Figure 1B, Figure 1-figure supplement 1). Endothelial denudation also reduced eATP levels in PAs from control mice, which were reduced further in endothelium-denuded PAs from Panx1 cKO-SMC mice.
We recently demonstrated that right ventricular systolic pressure (RVSP), a commonly used in vivo indicator of PAP, was elevated in inducible EC-specific Trpv4 KO (Trpv4 cKO-EC) mice (Daneva et al., 2021). Similarly, Panx1 cKO-EC mice also showed elevated RVSP ( Figure 1C). The Fulton index, a ratio of right ventricular (RV) weight to left ventricle plus septal (LV + S) weight, was not altered in Panx1 cKO-EC mice compared to control mice, suggesting a lack of right ventricular hypertrophy in these mice (Table 1). Baseline RVSP was not altered in Panx1 cKO-SMC mice ( Figure 1C), indicating a lack of regulation of resting PAP by SMC Panx1. Functional cardiac MRI studies indicated no alterations in cardiac function in Panx1 cKO-EC mice compared to the control mice ( Table 1), confirming that the changes in RVSP were not due to altered cardiac function. Baseline TRPV4 EC sparklet activity and that induced by a low concentration (1 nmol/L) of the specific TRPV4 channel agonist, GSK1016790A (hereafter, GSK101), were significantly reduced in PAs from Panx1 cKO-EC mice compared to PAs from Panx1 fl/fl mice ( Figure 1D and E). Additionally, the number of TRPV4 EC sparklet sites per cell was decreased in PAs from Panx1 cKO-EC mice ( Figure 1E). At the agonist concentration that maximally activates TRPV4 EC sparklets in PAs (30 nmol/L GSK101; Daneva et al., 2021), sparklet activity per site and sparklet sites per cell were not different between Panx1 cKO-EC Panx1 and control mice (    2). Outward currents through TRPV4 EC channels, elicited by 10 nmol/L GSK101, were also lower in Panx1 cKO-EC than Panx1 fl/fl mice ( Figure 1F, left and center). However, when maximally activated, TRPV4 EC channel currents were not different between Panx1 cKO-EC and Panx1 fl/fl mice ( Figure 1F, right), suggesting that the maximum number of functional TRPV4 EC channels is not altered in Panx1 cKO-EC mice.
Endothelial Panx1-TRPV4 signaling lowers pressure-and agonistinduced PA constriction Isolated, pressurized PAs (50-100 μm, Figure 2A) from Trpv4 cKO-EC mice exhibited a greater intraluminal pressure-induced (myogenic) constriction than PAs from control mice ( Figure 2B, Figure 2-figure supplement 1), providing the first evidence that TRPV4 EC channels oppose myogenic constriction in PAs. This finding was further supported by a greater contractile response to the thromboxane A 2 receptor agonist U46619 in PAs from Trpv4 cKO-EC mice (1-300 nmol/L; Figure 2C). PAs from Panx1 cKO-EC mice also showed a higher myogenic constriction than PAs from control mice ( Figure 2D), offering the first evidence that endothelial Panx1 regulates myogenic constriction of PAs. U46619-induced constriction was also increased in PAs from Panx1 cKO-EC mice compared to PAs from control mice. Pretreatment of PAs from Panx1 cKO-EC mice with a low concentration of TRPV4 agonist (GSK101, 3 nmol/L) reduced the U46619-induced constriction to control levels, indicating that endothelial Panx1 dilates PAs through TRPV4 EC channels. The presence of apyrase also increased U46619-induced constriction of PAs from control mice, confirming the dilatory effect of eATP on PAs ( Together, these data provide the first evidence that Panx1 EC -eATP-TRPV4 EC channel signaling lowers PA contractility and resting PAP. To verify the possibility that flow/shear stress activates ATP efflux through endothelial Panx1, we measured luminal eATP levels in PAs following exposure to different intraluminal shear stress levels (4, 7, and 14 dynes/cm 2 ; Figure 2F; Ahn et al., 2017). Increase in shear stress elevated luminal eATP levels in PAs from control mice, but not in PAs from Panx1 cKO-EC mice ( Figure 2G), confirming a critical role for Panx1 EC in shear stress-induced increase in luminal eATP. Also, shear stress-induced increase in luminal eATP was not altered in PAs from Trpv4 cKO-EC mice compared to control mice ( Figure 2H), suggesting that TRPV4 EC channels do not influence the efflux of ATP through Panx1 EC in response to increase in shear stress. eATP acts through purinergic P2Y2R EC stimulation to activate TRPV4 EC channels.
Similar to Panx1 cKO-EC mice, P2ry2 cKO-EC mice showed elevated RVSP and unaltered Fulton index ( Figure 3E). Exogenous ATP (1 μmol/L)-induced dilation was abolished in PAs from P2ry2 cKO-EC mice ( Figure 3F), confirming an essential role of P2Y2R EC in ATP-induced dilation of PAs. Further, PAs from P2ry2 cKO-EC mice showed higher myogenic and U46619-induced constriction compared to PAs from control mice ( Figure 3G). As observed with PAs from Panx1 cKO-EC mice, pretreatment with a low concentration of TRPV4 channel agonist (GSK101, 3 nmol/L) reduced U46619-induced constriction to control levels in PAs from P2ry2 cKO-EC mice ( Figure 3H). Taken together, these findings The online version of this article includes the following source data and figure supplement(s) for figure 2: Source data 1. Endothelial TRPV4 knockout increases U46619-induced constriction of PAs.
Source data 3. Shear stress increases ATP efflux through endothelial Panx1 in PAs.
Source data 4. Endothelial TRPV4 channel does not contribute to shear stress-induced increase in luminal ATP.  Cav-1 EC provides a scaffold for Panx1 EC -P2Y2R EC -TRPV4 EC signaling We hypothesized that Cav-1 EC provides a signaling scaffold that supports and maintains the spatial proximity among the individual elements in the Panx1 EC -P2Y2R EC -TRPV4 EC pathway. Previous studies demonstrated that endothelium-specific knockout of Cav1 results in reduced TRPV4 EC channel current density and elevated PAP (Daneva et al., 2021). Here, we provide evidence that eATP-induced activation of TRPV4 EC sparklets is absent in PAs from Cav1 cKO-EC mice ( Figure 4A; knockout validation in Daneva et al., 2021). As observed with PAs from Trpv4 cKO-EC and P2ry2 cKO-EC mice, eATPinduced dilation was also abolished in PAs from Cav1 cKO-EC mice ( Figure 4B). These results provided the first functional evidence that Cav-1 EC is required for eATP-P2Y2R EC -TRPV4 EC signaling in PAs.
To provide additional evidence to support Cav-1 EC -dependent co-localization of Panx1 EC -P2Y2R EC -TRPV4 EC signaling elements in PAs, we performed in situ proximity ligation assays (PLAs), which allow the detection of two proteins in close proximity (<40 nm). PLA data confirmed that Cav-1 EC exists within nanometer proximity of Panx1 EC , P2Y2R EC , and TRPV4 EC channels in PAs ( Figure 4C). Nanometer proximity was also observed between TRPV4 EC channels and P2Y2R EC and between Panx1 EC and P2Y2R EC ( Figure 4D, Figure 4-figure supplement 1). TRPV4 EC :P2Y2R and P2Y2R:Panx1 co-localization was lost in PAs from Cav1 cKO-EC mice, further supporting the crucial scaffolding role of Cav-1 EC in Panx1 EC -P2Y2R EC -TRPV4 EC pathway. PA endothelium has also been shown to express another P2Y family receptor, P2Y1 (P2Y1R) (Konduri et al., 2004). The PLA data confirmed that P2Y1R does not occur in nanometer proximity with Cav-1 EC in PAs (Figure 4-figure supplement 2). Together, these data confirmed a crucial role for Cav-1 EC in facilitating the spatial proximity amongst the individual elements of the Panx1 EC -P2Y2R EC -TRPV4 EC pathway.
Figure supplement 2. Percent constriction of pulmonary arteries (PAs) from Panx1 fl/fl and Panx1 fl/fl plus apyrase (10 U/mL) mice in response to U46619 (U466; 1-100 nmol/L; n = 5; **p<0.01 vs. Panx1 fl/fl ; two-way ANOVA).    mediates P2Y2R EC -TRPV4 EC channel interaction in PAs. PLA experiments confirmed that PKC also exists in nanometer proximity with Cav-1 EC in PAs ( Figure 6A). The PKC dependence of Cav-1 EC activation of TRPV4 EC channels was confirmed by studies in HEK293 cells transfected with TRPV4 alone or TRPV4 channels plus Cav-1 ( Figure 6B), which showed that TRPV4 currents were increased in the presence of Cav-1. Further, the PKCα/β inhibitor Gö-6916 (1 μmol/L) reduced TRPV4 channel currents in Cav-1/ TRPV4-co-transfected cells to the level of that in cells transfected with TRPV4 alone ( Figure 6B and C). These results imply that Cav-1 enhances TRPV4 channel activity via PKCα/β anchoring. Experiments in which TRPV4 channels were co-expressed with PKCα or PKCβ showed that only PKCα increased currents through TRPV4 channels ( Figure 6D). Collectively, these results support the conclusion that Panx1 EC -P2Y2R EC -PKCα-TRPV4 EC signaling on a Cav-1 EC scaffold reduces PA contractility and lowers resting PAP ( Figure 6E).

Discussion
Regulation of PA contractility and PAP is a complex process involving multiple cell types and signaling elements. In particular, the endothelial signaling mechanisms that control resting PAP remain poorly understood. Our studies identify a Panx1 EC , P2Y2R EC , and TRPV4 EC channel-containing signaling nanodomain that reduces PA contractility and lowers PAP. Although Panx1 EC and P2Y2R EC have been implicated in the regulation of endothelial function, their impact on PAP remains unknown. We demonstrate critical roles for several key, linked mechanistic, pathways showing that (1) Panx1 EC increases eATP levels in small PAs; (2) Panx1 EC -generated eATP, in turn, enhances Ca 2+ influx through TRPV4 EC channels, thereby dilating PAs and lowering PAP; (3) eATP acts through purinergic P2Y2R EC -PKCα signaling to activate TRPV4 EC channels; and (4) Cav-1 EC provides a signaling scaffold that ensures spatial proximity among the elements of the Panx1 EC -P2Y2R EC -PKCα-TRPV4 EC pathway. Our findings reveal a novel signaling axis that can be engaged by physiological stimuli to lower PAP and could also be therapeutically targeted in pulmonary vascular disorders. Moreover, the conclusions in this study may assist in future investigations of the mechanisms underlying pulmonary endothelial dysfunction.
Both ECs and SMCs control vascular contractility and arterial pressure. The expression of Panx1 and TRPV4 channels in both ECs and SMCs (Sharma et al., 2018;DeLalio et al., 2018;Martin et al., 2012;Ottolini et al., 2020a;Yang et al., 2006) makes it challenging to decipher the cell typespecific roles of Panx1 and TRPV4 channels using global knockouts or pharmacological strategies. Indeed, global Trpv4 knockout mice showed no systemic blood pressure or PAP phenotype (Xia et al., 2013;Zhang et al., 2009;Hong et al., 2018). However, inducible, Trpv4 cKO-EC mice had elevated systemic blood pressure and PAP (Daneva et al., 2021;Ottolini et al., 2020b). Lack of a phenotype in global knockout mice could be due to the deletion of TRPV4 channels from multiple cell types or compensatory mechanisms that have developed over time (reviewed by El-Brolosy and Stainier, 2017). Therefore, studies utilizing cell-specific knockout mice are necessary for a definitive assessment of the control of PAP by EC and SMC Panx1 and TRPV4 channels. Although SMC TRPV4 channels have been shown to contribute to hypoxia-induced pulmonary vasoconstriction, resting PAP is not altered in global Trpv4 knockout mice (Xia et al., 2013;Yang et al., 2012). Further, our studies indicate that SMC Panx1 and TRPV4 channels do not influence resting PAP. Taken together with findings from EC-knockout mice, these results provide strong evidence that endothelial, but not SMC, Panx1 and TRPV4 channels maintain low PA contractility and PAP under resting conditions. Despite the elevated PAP in EC-specific Panx1, P2ry2, and Trpv4 cKO mice (Daneva et al., 2021), right ventricular hypertrophy was not observed. These findings could be attributed to a short duration of inducible genetic deletion in our studies. Although the duration of the knockout is sufficient to result in elevated PAP, a longer duration or larger changes in PAP may be required for observing right ventricular hypertrophy in these mouse models.
The online version of this article includes the following source code for figure 3: Source data 1. Endothelial P2Y2R knockout increases U46619-induced constriction of PAs.   Recent studies in pulmonary fibroblasts and other cell types suggest that TRPV4 channel-mediated increases in cytosolic Ca 2+ can induce eATP release through Panx1 (Baxter et al., 2014;Rahman et al., 2018). However, the reverse interaction, in which Panx1-mediated eATP release activates TRPV4 channels, has not been explored in any cell type. Since Panx1 is activated by cytosolic Ca 2+ (Locovei et al., 2006) and eATP has been previously shown to activate TRPV4 EC channels (Marziano et al., 2017), bidirectional signaling between Panx1 and TRPV4 channels is conceivable. Our demonstration that baseline eATP levels are unchanged in PAs from Trpv4 cKO-EC mice rules out a role for TRPV4 EC channels in controlling eATP release under baseline conditions. Moreover, TRPV4 EC channels did not contribute to flow-induced efflux of ATP through Panx1 EC . Nevertheless, these data from pulmonary ECs do not rule out potential TRPV4-Ca 2+ -Panx1 signaling in other cell types.

Figure 3 continued
Elevated capillary TRPV4 EC channel activity has been linked to increased endothelial permeability (Thorneloe et al., 2012;Yin et al., 2008), lung injury (Alvarez et al., 2006), and pulmonary edema (Thorneloe et al., 2012;Yin et al., 2008). Moreover, Panx1 EC -mediated eATP release is associated with vascular inflammation at the level of capillaries (Sharma et al., 2018). The physiological roles of Panx1 EC and TRPV4 EC channels in PAs, however, remain unknown. ECs from pulmonary capillaries and arteries are structurally and functionally different. Whereas PAs control pulmonary vascular resistance and PAP, capillaries control vascular permeability. TRPV4 EC channels couple with distinct targets in arterial and capillary ECs (Sonkusare et al., 2012;Longden et al., 2017). Our data identify physiological roles of Panx1 EC -TRPV4 EC channel signaling in PAs, but whether such signaling operates in the capillary endothelium and is essential for its physiological function is unclear.
Purinergic signaling and the endogenous purinergic receptor agonist eATP are essential controllers of pulmonary vascular function (Konduri and Mital, 2000;Konduri et al., 2004;Hennigs et al., 2019;Kylhammar et al., 2014). Our discovery of the Panx1 EC -P2Y2R EC -TRPV4 EC channel pathway establishes a signaling axis in ECs that regulates pulmonary vascular function. The pulmonary vasculature is a high-flow circulation, yet the flow-induced signaling mechanisms are poorly understood in PAs. Our results confirm that flow/shear stress increases ATP efflux through Panx1 EC in PAs, which could be a potential mechanism for flow-induced dilation of PAs. Further investigations are needed to verify flow/ shear stress-induced, eATP-dependent activation of P2Y2R EC -PKCα-TRPV4 EC signaling in PAs. Several purinergic receptor subtypes are expressed in the pulmonary vasculature, including P2YRs and P2XRs (Konduri et al., 2004;Hennigs et al., 2019;Syed et al., 2010). Although only P2Y2R EC appears to mediate eATP activation of TRPV4 EC channels, our studies do not rule out potentially important roles for other P2Y or P2X receptors in the pulmonary endothelium.
The online version of this article includes the following figure supplement(s) for figure 5:
Cav-1 EC /PKCα-dependent signaling is a novel endogenous mechanism for activating arterial TRPV4 EC channels and lowering PAP. Proximity to PKCα appears to be crucial for the normal function of TRPV4 channels. Evidence from the systemic circulation suggests that co-localization of TRPV4 channels with scaffolding proteins enhances their activity (Mercado et al., 2014;Sonkusare et al., 2014), and we specifically demonstrated that PKC anchoring by AKAP150 enhances the activity of TRPV4 EC channels in mesenteric arteries (Ottolini et al., 2020b). Here, we show that PKC anchoring by Cav-1 EC enables PKC activation of TRPV4 EC channels in PAs. This discovery raises the possibility that disruption of PKC anchoring by Cav-1 EC could impair the Panx1 EC -P2Y2R EC -TRPV4 EC signaling axis under disease conditions. A lack of PKC anchoring by scaffolding proteins in systemic arteries has been demonstrated in obesity and hypertension (Ottolini et al., 2020b;Sonkusare et al., 2014). Further studies of pulmonary vascular disorders are required to establish whether the Panx1 EC -P2Y2R EC -PKCα-TRPV4 EC signaling axis is impaired in pulmonary vascular disorders.
In conclusion, Panx1 EC -P2Y2R EC -TRPV4 EC channel signaling reduces PA contractility and maintains a low resting PAP. This mechanism is facilitated by eATP released through Panx1 EC and subsequent activation of P2Y2R EC -PKCα signaling. Cav-1 EC ensures the spatial proximity among Panx1 EC , P2Y2R EC , and TRPV4 EC channels and also anchors PKCαclose to TRPV4 EC channels. These findings identify a novel endothelial Ca 2+ signaling mechanism that reduces PA contractility. Further investigations are needed to determine whether impairment of this pathway contributes to elevated PAP in pulmonary vascular disorders and whether this pathway can be targeted for therapeutic benefit.

RVSP and Fulton index measurement
Mice were anesthetized with pentobarbital (50 mg/kg bodyweight; intraperitoneally) and bupivacaine HCl (100 μL of 0.25% solution; subcutaneously) was used to numb the dissection site on the mouse. RVSP was measured as an indirect indicator of PAP. A Mikro-Tip pressure catheter (SPR-671; Millar Instruments, Huston, TX), connected to a bridge amp (FE221), and a PowerLab 4/35 4-channel recorder (ADInstruments, Colorado Springs, CO), was inserted through the external jugular vein into the right ventricle. Right ventricular pressure and heart rate were acquired and analyzed using LabChart8 software (ADInstruments). A stable 3 min recording was acquired for all the animals, and 1 min continuous segment was used for data analysis. When necessary, traces were digitally filtered using a low-pass filter at a cutoff frequency of 50 Hz. At the end of the experiments, mice were euthanized, and the hearts were isolated for right ventricular hypertrophy analysis. Right ventricular hypertrophy was determined by calculating the Fulton index, a ratio of the right ventricular (RV) heart weight over the left ventricular (LV) plus septum (S) weight (RV/ LV + S).
Luciferase assay for total ATP release ATP assay protocol was adapted from Yang et al., 2020. Fourth-order PAs (~50 μm diameter) were isolated in cold HEPES-buffered physiological salt solution (HEPES-PSS, in mmol/L, 10 HEPES, 134 NaCl, 6 KCl, 1 MgCl 2 hexahydrate, 2 CaCl 2 dihydrate, and 7 dextrose, pH adjusted to 7.4 using 1 mol/L NaOH). Isolated PAs were pinned down en face on a Sylgard block and cut open. PAs were placed in black, opaque 96-well plates and incubated in HEPES-PSS for 10 min at 37 °C, followed by incubation with the ectonucleotidase inhibitor ARL 67156 (300 μmol/L, Tocris Bioscience, Minneapolis, MN) for 30 min at 37 °C. 50 μL volume of each sample was transferred to another black, opaque 96-well plate. ATP was measured using ATP bioluminescence assay reagent ATP Bioluminescence HSII kit (Roche Applied Science, Penzberg, Germany). Using a luminometer (FluoStar Omega), 50 μL of luciferin:luciferase reagent (ATP bioluminescence assay kit HSII; Roche Applied Science) was injected into each well and luminescence was recorded following a 5 s orbital mix and sample measurement at 7 s. ATP concentration in each sample was calculated from an ATP standard curve. For some experimental groups, PAs were first mounted on a pressure myography chamber and were denuded by pushing air through the lumen for 1 min.

Cardiac magnetic resonance imaging (MRI)
MRI studies were conducted under protocols that comply with the Guide for the Care and Use of Laboratory Animals (NIH publication no. 85-23, revised 1996). Mice were positioned in the scanner under 1.25% isoflurane anesthesia and body temperature was maintained at 37 °C using thermostatic circulating water. A cylindrical birdcage RF coil (30 mm diameter, Bruker, Ettlingen, Germany) with an active length of 70 mm was used, and heart rate, respiration, and temperature were monitored during imaging using a fiber optic, MR-compatible system (Small Animal Imaging Inc, Stony Brook, NY). MRI was performed on a 7 Tesla (T) Clinscan system (Bruker) equipped with actively shielded gradients with a full strength of 650 mT/m and a slew rate of 6666 mT/m/ms (Vandsburger et al., 2007). Six short-axis slices were acquired from base to apex, with slice thickness of 1 mm, in-plane spatial resolution of 0.2 × 0.2 mm 2 , and temporal resolution of 8-12 ms. Baseline ejection fraction (EF), end-diastolic volume (EDV), end-systolic volume (ESV), myocardial mass, wall thickness, stroke volume (SV), and cardiac output (CO) were assessed from the cine images using the freely available software Segment version 2.0 R5292 (http:// segment. heiberg. se).

Pressure myography
Isolated mouse PAs (~50 μm) were cannulated on glass micropipettes in a pressure myography chamber (The Instrumentation and Model Facility, University of Vermont, Burlington, VT) at areas lacking branching points and were pressurized at a physiological pressure of 15 mm Hg (Ottolini et al., 2020a). Arteries were superfused with PSS (in mmol/L, 119 NaCl, 4.7 KCl, 1.2 KH 2 PO 4 , 1.2 MgCl 2 hexahydrate, 2.5 CaCl 2 dihydrate, 7 dextrose, and 24 NaHCO 3 ) at 37 °C and bubbled with 20 % O 2 /5 % CO 2 to maintain the pH at 7.4. All drug treatments were added to the superfusing PSS. PAs were pre-constricted with 50 nmol/L U46619 (a thromboxane A2 receptor agonist). All other pharmacological treatments were performed in the presence of U46619. Before measurement of vascular reactivity, arteries were treated with NS309 (1 μmol/L), a direct opener of endothelial IK/SK channels, to assess endothelial health. Arteries that failed to fully dilate to NS309 were discarded. Changes in arterial diameter were recorded at a 60-ms frame rate using a charge-coupled device camera and edge-detection software (IonOptix LLC, Westwood, MA;Sonkusare et al., 2012;Sonkusare et al., 2014). All drug treatments were incubated for 10 min. At the end of each experiment, Ca 2+ -free PSS (in mmol/L, 119 NaCl, 4.7 KCl, 1.2 KH 2 PO 4 , 1.2 MgCl 2 hexahydrate, 7 dextrose, 24 NaHCO 3 , and 5 EGTA) was applied to assess the maximum passive diameter. Percent constriction was calculated by where Diameter before is the diameter of the artery before a treatment and Diameter after is the diameter after the treatment. Percent dilation was calculated by (2) w here Diameter basal is the stable diameter before drug treatment, Diameter dilated is the diameter after drug treatment, and Diameter Ca-free is the maximum passive diameter.

Flow/shear stress-induced ATP release
Flow/shear stress was measured using a protocol modified from Ahn et al., 2017. Briefly, isolated PAs (~50 μm) were cannulated on glass micropipettes in a pressure myography chamber (The Instrumentation and Model Facility, University of Vermont) at areas lacking branching points and were pressurized at a physiological pressure of 15 mm Hg (Ottolini et al., 2020a). Arteries were superfused with PSS (in mmol/L, 119 NaCl, 4.7 KCl, 1.2 KH 2 PO 4 , 1.2 MgCl 2 hexahydrate, 2.5 CaCl 2 dihydrate, 7 dextrose, and 24 NaHCO 3 ) at 37 °C and bubbled with 20% O 2 /5% CO 2 to maintain the pH at 7.4. The arteries were treated luminally with 300 μmol/L ARL-67156 (ecto-ATPase inhibitor; Sigma-Aldrich) to avoid ATP degradation throughout the duration of the experiment. The tips of the cannulating pipettes were always arranged with smaller pipettes upstream and larger pipettes downstream. The average tip size was 20.1 ± 0.4 μm at the upstream end and 23.6 ± 0.4 μm at the downstream end. Both ends of the vessel were secured, and the vessel was maintained at an intraluminal pressure of 15 cmH 2 O by elevating the inflow reservoir. Flow/shear stress was increased by adjusting the height of the reservoir. Flow-induced luminal solution was collected at the outflow pipette end. After a 30 min equilibration period, a baseline sample was collected for luminal ATP measurement. Shear stress was calculated from the flow rate in the vessel lumen and the diameter of the vessels using the equation (Zemskov et al., 2011) : τ = 4(µQ/(πr 3 )) , where μ is viscosity, . Q is volumetric flow rate, and r is internal radius of the vessel. The volumetric flow rate was measured as the volume of the flowthrough at different pressures. Vessel diameter was measured at each flow rate. The shear stress range was 4-14 dynes/ cm 2 . Luminal outflow samples per shear stress range were obtained every 30 min. The samples were used for luciferase assays for total ATP release, as described above.

Ca 2+ imaging
Measurements of TRPV4 EC Ca 2+ sparklets in the native endothelium of mouse PAs were performed as previously described (Sonkusare et al., 2012). Briefly, fourth-order (~50 μm) PAs were pinned down en face on a Sylgard block and loaded with Fluo-4-AM (10 μmol/L) in the presence of pluronic acid (0.04%) at 30 °C for 30 min. TRPV4 EC Ca 2+ sparklets were recorded at 30 frames per second with Andor Revolution WD (with Borealis) spinning-disk confocal imaging system (Oxford Instruments, Abingdon, UK) comprised an upright Nikon microscope with a 60× water dipping objective (numerical aperture 1.0) and an electron multiplying charge coupled device camera (iXon 888, Oxford Instruments). All experiments were carried out in the presence of cyclopiazonic acid (20 μmol/L, a sarco-endoplasmic reticulum [ER] Ca 2+ -ATPase inhibitor) in order to eliminate the interference from Ca 2+ release from intracellular stores. Fluo-4 was excited at 488 nm with a solid-state laser and emitted fluorescence was captured using a 525/36 nm band-pass filter. TRPV4 EC Ca 2+ sparklets were recorded before and 5 min after the addition of specific compounds. To generate fractional fluorescence (F/F 0 ) traces, a region of interest defined by a 1.7-μm 2 (5 × 5 pixels) box was placed at a point corresponding to peak sparklet amplitude. Each field of view was ~110 × 110 μm and covered ~15 ECs. Representative F/ F 0 traces were filtered using a Gaussian filter and a cutoff corner frequency of 4 Hz. Sparklet activity was assessed as described previously using the custom-designed SparkAn software (Sonkusare et al., 2012;Sonkusare et al., 2014).

Calculation of TRPV4 sparklet activity per site
Activity of TRPV4 Ca 2+ sparklets was analyzed as described previously (Sonkusare et al., 2012;Ottolini et al., 2020b;Sonkusare et al., 2014). Area under the curve for all the events at a site was determined using trapezoidal numerical integration ([F−F 0 ]/F 0 over time, in seconds). The average number of active TRPV4 channels, as defined by NP O (where N is the number of channels at a site and P O is the open state probability of the channel), was calculated by NPO = (Tlevel1 + 2Tlevel2 + 3Tlevel3 + 4Tlevel4)/Ttotal ( 3) where T is the dwell time at each quantal level detected at TRPV4 sparklet sites and T total is the duration of the recording. NP O was determined using Single Channel Search module of Clampfit and quantal amplitudes derived from all-points histograms (Marziano et al., 2017) (ΔF/F 0 of 0.29 for Fluo-4-loaded PAs).
Total number of sparklet sites in a field was divided by the number of cells in that field to obtain sparklet sites per cell.

All-points histograms
All-points amplitude histograms were constructed as described previously (Sonkusare et al., 2012;Ottolini et al., 2020b). Briefly, images were filtered with a Kalman filter (adopted from an ImageJ plug-in written by Christopher Philip Mauer, Northwestern University, Chicago, IL; acquisition noise variance estimate = 0.05; filter gain = 0.8). The inclusion criteria were a stable baseline containing at least five steady points and a steady peak containing at least five peak points. Sparklet traces were exported to ClampFit10.3 for constructing an all-points histogram, which was fit with the multiple Gaussian function below: where F/F 0 represents fractional fluorescence, a represents the area, μ represents the mean value, and σ2 represents the variance of the Gaussian distribution. While the detected sparklets can have multiple amplitudes corresponding to quantal level 1, 2, 3, or 4, the baseline (level 0) was the same for all the detected sparklets regardless of the amplitude of the sparklets. Therefore, the baseline corresponds to a higher count compared to all other events.

Immunostaining
Immunostaining was performed on fourth-order PAs (~50 μm) pinned en face on SYLGARD blocks. PAs were fixed with 4% paraformaldehyde (PFA) at room temperature for 15 min and then washed three times with phosphate-buffered saline (PBS). The tissue was permeabilized with 0.2% Triton-X for 30 min, blocked with 5% normal donkey serum (ab7475, Abcam, Cambridge, MA) or normal goat serum (ab7475, Abcam), depending on the host of the secondary antibody used, for 1 hr at room temperature. PAs were incubated with the primary antibodies (Key resources table) overnight at 4 °C. Following the overnight incubation, PAs were incubated with secondary antibody 1:500 Alexa Fluor 568-conjugated donkey anti-rabbit (Life Technologies, Carlsbad, CA) for 1 hr at room temperature in the dark room. For nuclear staining, PAs were washed with PBS and then incubated with 0.3 mmol/L DAPI (Invitrogen, Carlsbad, CA) for 10 min at room temperature. Images were acquired along the z-axis from the surface of the endothelium to the bottom where the EC layer encounters the smooth muscle cell layer with a slice size of 0.1 μm using the Andor microscope described above. The internal elastic lamina (IEL) autofluorescence was evaluated using an excitation of 488 nm with a solid-state laser and collecting the emitted fluorescence with a 525/36 nm band-pass filter. Immunostaining for the protein of interest was evaluated using an excitation of 561 nm and collecting the emitted fluorescence with a 607/36 nm band-pass filter. DAPI immunostaining was evaluated using an excitation of 409 nm and collecting the emitted fluorescence with a 447/69 nm band-pass filter. The specificity of Panx1 and P2Y2R antibodies was confirmed by a lack of signal in PAs from endothelial knockout mice. The specificity of TRPV4, Cav-1, and PKC antibodies was confirmed previously (Daneva et al., 2021;Ottolini et al., 2020b).

In situ PLA
Fourth-order (~50 μm) PAs were pinned en face on SYLGARD blocks. PAs were fixed with 4% PFA for 15 min followed by three washes with PBS. PAs were then permeabilized with 0.2% Triton X for 30 min at room temperature followed by blocking with 5% normal donkey serum (Abcam plc, Cambridge, MA) and 300 mmol/L glycine for 1 hr at room temperature. After three washes with PBS, PAs were incubated with the primary antibodies (Key resources table) overnight at 4 °C. The PLA protocol from Duolink PLA Technology kit (Sigma-Aldrich) was followed for the detection of co-localized proteins. Lastly, PAs were incubated with 0.3 μmol/L DAPI nuclear staining (Invitrogen) for 10 min at room temperature in the dark room. PLA images were acquired using the Andor Revolution spinning-disk confocal imaging system along the z-axis at a slice size of 0.1 μm. Images were analyzed by normalizing the number of positive puncta by the number of nuclei in a field of view. The specificity of the PLA antibodies was determined using PAs from endothelial knockout mice for one of the protein pairs.

Plasmid generation and transfection into HEK293 cells
HEK293 cells authenticated with STR profiling were obtained from ATCC USA. Mycoplasma contamination was not detected as per ATCC website. The TRPV4 coding sequence without stop codons was amplified from mouse heart cDNA. The amplified fragment was inserted into a plasmid backbone containing a CMV promoter region for expression and, in addition, is suitable for lentiviral production by Gibson assembly. The in-frame FLAG tag was inserted into the 3′-primer used for amplification. Constructs were verified by sequencing the regions that had been inserted into the plasmid backbone. HEK293 cells were seeded (7 × 10 5 cells per 100 mm dish) in Dulbecco's Modified Eagle Medium with 10% fetal bovine serum (Thermo Fisher Scientific Inc, Waltham, MA) 1 day prior to transfection. Cells were transfected using the LipofectamineLTX protocol (Thermo Fisher Scientific Inc). TRPV4 was co-expressed with PKCα and PKCβ, obtained from Origene Technologies (Montgomery County, MD).

Statistical analysis
Results are presented as mean ± SEM. The n = 1 was defined as one artery in the imaging experiments (Ca 2+ imaging, PLA), one cell for patch-clamp experiments, one mouse for RVSP measurements, one artery for pressure myography experiments, one mouse for functional MRI, one mouse for ATP measurements, and one mouse for qPCR experiments. The data were obtained from at least three mice in experiments performed in at least two independent batches. The individual data points are shown for each dataset. For in vivo experiments, an independent team member performed random assignment of animals to groups and did not have knowledge of treatment assignment groups. All the in vivo experiments were blinded; information about the groups or treatments was withheld from the experimenter or from the team member who analyzed the data. All data are shown in graphical form using CorelDraw Graphics Suite X7 (Ottawa, ON, Canada) and statistically analyzed using GraphPad Prism 8.3.0 (Sand Diego, CA). A power analysis to determine group sizes and study power (>0.8) was performed using GLIMMPSE software (α = 0.05; >20% change). Using this method, we estimated at least cells per group for patch-clamp experiments, five arteries per group for imaging and pressure myography experiments, and mice per group for RVSP measurements and MRI. A Shapiro-Wilk test was performed to determine normality. The data in this article were normally distributed; therefore, parametric statistics were performed. Data were analyzed using two-tailed, paired or independent t-test (for comparison of data collected from two different treatments), one-way ANOVA or twoway ANOVA (to investigate statistical differences among more than two different treatments). Tukey correction was performed for multiple comparisons with one-way ANOVA, and Bonferroni correction was performed for multiple comparisons with two-way ANOVA. Statistical significance was determined as a p-value <0.05.