Mechanisms underlying selective coupling of endothelial Ca²⁺ signals with eNOS versus IK/SK channels in systemic and pulmonary arteries

Abstract

Spatially localized Ca²⁺ signals activate Ca²⁺-sensitive intermediate- and small-conductance K⁺ (IK and SK) channels in some vascular beds and endothelial nitric oxide synthase (eNOS) in others. The goal of this study was to uncover the signaling organization that determines selective Ca²⁺ signal to vasodilatory target coupling in the endothelium. Resistance-sized mesenteric arteries (MAs) and pulmonary arteries (PAs) were used as prototypes for arteries with predominantly IK/SK channel- and eNOS-dependent vasodilation, respectively. Ca²⁺ influx signals through endothelial transient receptor potential vanilloid 4 (TRPV4_EC) channels played an important role in controlling the baseline diameter of both MAs and PAs. TRPV4_EC channel activity was similar in MAs and PAs. However, the TRPV4 channel agonist GSK1016790A (10 nM) selectively activated IK/SK channels in MAs and eNOS in PAs, revealing preferential TRPV4_EC-IK/SK channel coupling in MAs and TRPV4_EC-eNOS coupling in PAs. IK/SK channels co-localized with TRPV4_EC channels at myoendothelial projections (MEPs) in MAs, but lacked the spatial proximity necessary for their activation by TRPV4_EC channels in PAs. Additionally, the presence of the NO scavenging protein hemoglobin α (Hbα) within nanometer proximity to eNOS limits TRPV4_EC-eNOS signaling in MAs. In contrast, co-localization of TRPV4_EC channels and eNOS at MEPs, and the absence of Hbα, favor TRPV4_EC-eNOS coupling in PAs. Thus, our results reveal that differential spatial organization of signaling elements determines TRPV4_EC-IK/SK versus TRPV4_EC-eNOS coupling in resistance arteries.

INTRODUCTION.

Increases in endothelial cell (EC) Ca²⁺ are known to promote vasodilation and thereby reduce vascular resistance and blood pressure. A number of different mechanical stimuli and neurohumoral mediators cause vasodilation by increasing EC Ca²⁺. Intracellular Ca²⁺ can activate multiple vasodilator targets in ECs, including intermediate- and small-conductance K⁺ (IK or SK) channels (Murphy & Brayden, 1995; Burnham et al., 2002; Bychkov et al., 2002; Taylor et al., 2003; Ledoux et al., 2008b), endothelial nitric oxide synthase (eNOS) (Ignarro et al., 1987; Busse & Mulsch, 1990), and factors that hyperpolarize smooth muscle cells (SMCs) (Ellinsworth et al., 2016). Pressure myography studies in resistance-sized arteries have revealed that increases in endothelial Ca²⁺ cause vasodilation mainly via IK/SK channel activation in some vascular beds (Taylor et al., 2003; Ledoux et al., 2008b; Earley et al., 2009; Bagher et al., 2012; Sonkusare et al., 2012; Pires et al., 2015) and eNOS activation in other vascular beds (Hein & Kuo, 1999; Beyer & Gutterman, 2012; Marziano et al., 2017). Although the activation of IK/SK channels and eNOS by Ca²⁺ has been studied in detail (Taylor et al., 2003; Ledoux et al., 2008a; Ledoux et al., 2008b; Earley et al., 2009; Bagher et al., 2012; Nausch et al., 2012; Sonkusare et al., 2012; Marziano et al., 2017; Hong et al., 2018), the preferential activation of one target versus the other remains unexplained. The majority of vascular disorders are associated with a loss of IK/SK channel- or eNOS-dependent vasodilation (Feletou, 2009; Klinger et al., 2013; Ma et al., 2013; Seki et al., 2017). Understanding the molecular basis for preferential coupling of Ca²⁺ signals with IK/SK channel or eNOS is therefore essential for deciphering the selective impairment of vasodilatory mechanisms in vascular disorders.

Multiple ion channels increase Ca²⁺ levels in native ECs, including ion channels of the transient receptor potential (TRP) family (Earley et al., 2009; Sonkusare et al., 2012; Pires et al., 2015; Sullivan et al., 2015; Hong et al., 2018) and inositol triphosphate (IP3) receptors (Ledoux et al., 2008b; McCarron et al., 2010; Bagher et al., 2012; Tran et al., 2012; Heathcote et al., 2019). The TRP vanilloid 4 (TRPV4) channel has emerged as an important Ca²⁺-influx pathway in ECs from systemic and pulmonary resistance arteries (Bagher et al., 2012; Sonkusare et al., 2014; Marziano et al., 2017; Hong et al., 2018). Pharmacological activation of endothelial TRPV4 (TRPV4_EC) channels is known to dilate systemic and pulmonary arteries (Sonkusare et al., 2012; Marziano et al., 2017), and knockout of these channels specifically in the endothelium increases resting blood pressure (Ottolini et al., 2020). However, whether TRPV4_EC channels control resting vascular diameter has not been addressed. Interestingly, activation of TRPV4_EC channels dilates systemic mesenteric (Sonkusare et al., 2012), cremaster (Bagher et al., 2012), and cerebral arteries (Zhang et al., 2013) via IK/SK channel activation, and pulmonary arteries via eNOS activation (Marziano et al., 2017). TRPV4_EC channel-induced activation of IK/SK channels has been attributed to the localization of TRPV4_EC and IK/SK channels at myoendothelial projections (MEPs), which provide points of contact between ECs and overlying smooth muscle cells (SMCs) (Bagher et al., 2012; Sonkusare et al., 2014). However, it is not clear why TRPV4_EC channels do not activate eNOS in resistance systemic arteries or IK/SK channels in resistance pulmonary arteries. Therefore, we hypothesized that spatial proximity of TRPV4_EC channels with eNOS or IK/SK channels determines TRPV4-eNOS or TRPV4-IK/SK coupling in native endothelium.

On the one hand, activation of IK/SK channels causes EC membrane hyperpolarization, which is then transmitted to SMCs via myoendothelial gap junctions (MEGJs) in MEPs (Dora et al., 2003). SMC hyperpolarization results in deactivation of L-type Ca²⁺ channels and vasodilation (Knot & Nelson, 1998). On the other hand, activation of eNOS releases NO, which diffuses to SMCs and causes vasodilation via guanylyl cyclase-dependent and -independent mechanisms (Cohen et al., 1999). Interestingly, no patch-clamp studies have been performed on arteries that exhibit Ca²⁺-dependent activation of eNOS that would lead to the conclusion that IK/SK channels are present and can be activated in these arteries. It is similarly unknown whether eNOS is expressed and can be activated by Ca²⁺ in arteries that show selective Ca²⁺ signal-IK/SK channel coupling. Vasodilator effector molecules are a source of endothelial heterogeneity; therefore, it is important to understand the fundamental mechanisms that determine differential Ca²⁺ signal-target coupling in ECs.

In the current study, we used third-order mesenteric arteries (MAs) as a model for TRPV4-IK/SK channel coupling (Sonkusare et al., 2012; Sonkusare et al., 2014; Hong et al., 2018), and fourth-order pulmonary arteries (PAs) as a prototype for TRPV4-eNOS coupling (Marziano et al., 2017). We provide the first evidence for an essential role of TRPV4_EC channels in controlling baseline diameter in both MAs and PAs. Functional IK/SK channels and eNOS were present in ECs of both MAs and PAs, and TRPV4_EC and IK/SK channel current densities were similar between MAs and PAs. Stimulation of TRPV4_EC channels with the specific agonist GSK1016790A (hereafter, GSK101) selectively activated IK/SK channels in MAs, where TRPV4_EC and IK/SK channels were found to localize at MEPs; in contrast, stimulation of TRPV4_EC channels selectively activated eNOS in PAs. Moreover, a lack of TRPV4_EC-eNOS coupling in MAs correlated with co-localization of the NO-scavenging protein hemoglobin α (Hbα) with eNOS in MAs (Straub et al., 2012) but not in PAs. These results identify novel mechanisms of Ca²⁺ signal-to-vasodilator-target coupling in ECs, and may explain heterogeneity in endothelium-dependent vasodilator mechanisms.

METHODS

Ethical approval.

All animal studies were approved by the University of Virginia Animal Care and Use Committee (Protocols 4120 and 4051) and followed the principles and regulations of The Journal of Physiology (Grundy, 2015).

Animal care and use.

Male C57BL6/J, eNOS^−/− (10–14 weeks old, ~25 g; The Jackson Laboratory, Bar Harbor, ME, USA), and TRPV4_EC^−/− mice (Ottolini et al., 2020) were used in this study (n = 85 mice in total). Mice were housed in an enriched environment and maintained on a 12:12 h light-dark cycle at ~23°C with fresh tap water and standard chow diet available ad libitum. Mice were euthanized with pentobarbital (90 mg/kg, intraperitoneal; Diamondback Drugs, Scottsdale, AZ, USA) followed by decapitation for harvesting intestinal and lung tissues. TRPV4_EC^−/− mice were developed and validated as described earlier (Ottolini et al., 2020). Briefly, TRPV4^fl/fl (obtained from Dr. Wolfgang Liedtke, Duke University School of Medicine, Durham, NC) (Moore et al., 2013) were crossed with tamoxifen-inducible VE-Cadherin (Cdh5) Cre mice (Sorensen et al., 2009). TRPV4^fl/fl Cre⁺ mice were injected with tamoxifen (Sigma-Aldrich, St. Louis, MO, USA) at 6 weeks of age (40 mg/kg i.p. per day for 10 days), followed by a 2-week washout period. TRPV4^fl/fl Cre⁻ mice injected with tamoxifen were used as Wild-type (WT) control mice. Third-order MAs (~100 μm) and fourth order PAs (~100 μm) were dissected in cold Hepes-buffered physiological salt solution (Hepes-PSS; 10 mM Hepes, 134 mM NaCl, 6 mM KCl, 1 mM MgCl₂ hexahydrate, 2 mM CaCl₂ dihydrate, and 7 mM dextrose; pH adjusted to 7.4 using 1 M NaOH).

Pressure myography.

Isolated PAs and MAs were cannulated on glass pipettes, mounted in an arteriography chamber (Instrumentation and Model Facility, University of Vermont, Burlington, VT, USA) at areas lacking branching points, and pressurized to physiological pressure (15 mm Hg for PAs and 80 mm Hg for MAs). Arteries were superfused with PSS (119 mM NaCl, 4.7 mM KCl, 1.2 mM KH₂PO₄, 1.2 mM MgCl₂ hexahydrate, 2.5 mM CaCl₂ dihydrate, 7 mM dextrose, and 24 mM NaHCO₃) at 37°C and bubbled with 20% O₂/5% CO₂ to maintain the pH at 7.4. All drugs were added to the superfusing PSS. Arteries were preconstricted with 100 nM U46619, a thromboxane A2 agonist. Endothelial health was assessed by monitoring the response to NS309 (1 μM), a direct opener of endothelial IK/SK channels. Arteries that failed to dilate to NS309 were discarded. In some experiments, the endothelium was denuded by passing an air bubble through the artery for 60 seconds. Complete removal of the endothelial cell layer was verified by the absence of dilation to NS309. 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, USA). For all drug treatments, the incubation time was 5–10 minutes. At the end of each experiment, Ca²⁺‐free PSS (119 mM NaCl, 4.7 mM KCl, 1.2 mM KH₂PO₄, 1.2 mM MgCl₂ hexahydrate, 7 mM dextrose, 24 mM NaHCO₃, and 5 mM EGTA) was applied to assess the maximum passive diameter. Percent constriction was calculated by

$$\left[\frac{{\mathit{\text{Diameter}}}_{\mathit{\text{before}}}-{\mathit{\text{Diameter}}}_{\mathit{\text{after}}}}{{\mathit{\text{Diameter}}}_{\mathit{\text{before}}}}\right]*100,$$

where Diameter_before is the diameter before drug treatment and Diameter_after is the diameter after drug treatment.

Percent vasodilation was calculated by

$$\left[\frac{{\mathit{\text{Diameter}}}_{\mathit{\text{Dilated}}}-{\mathit{\text{Diameter}}}_{\mathit{\text{basal}}}}{{\mathit{\text{Diameter}}}_{\mathit{\text{cafree}}}-{\mathit{\text{Diameter}}}_{\mathit{\text{basal}}}}\right]*100,$$

where Diameter_basal is the diameter before drug treatment, ${\mathit{\text{Diameter}}}_{\mathit{\text{dilated}}}$ is the diameter after drug treatment, and Diameter Ca_free is the maximum passive diameter.

For a subset of MAs, pressure myography studies were performed in the absence of U46619 at a pressure of 80 mm Hg to induce myogenic tone. Endothelial health was tested with NS309, followed by treatment with the TRPV4 inhibitor GSK2193874.

Ca²⁺ imaging.

Ca²⁺-imaging studies were performed as described previously (Sonkusare et al., 2012; Marziano et al., 2017; Hong et al., 2018). Briefly, third-order MAs and fourth-order PAs were surgically opened and pinned down on a SYLGARD block with the endothelium facing up (en face preparation). MAs and PAs were incubated with Fluo-4 AM (10 μM) and pluronic acid (0.04%) at 30°C for 45 and 30 minutes, respectively, in the dark. Ca²⁺ images were acquired at 30 frames per second using an Andor Revolution WD (with Borealis) spinning-disk confocal imaging system (Andor Technology, Belfast, UK) comprising an upright Nikon microscope with a 60X water-dipping objective (numerical aperture, 1.0) and an electron multiplying charge-coupled device camera. Arteries were superfused with PSS (119 mM NaCl, 4.7 mM KCl, 1.2 mM KH₂PO₄, 1.2 mM MgCl₂ 160 hexahydrate, 2.5 mM CaCl₂ dihydrate, 7 mM dextrose, and 24 mM NaHCO₃), bubbled with 20% O₂ and 5% CO₂ to maintain the pH at 7.4. All experiments were performed at 37°C. Fluo-4 was excited using a 488-nm solid-state laser, and emitted fluorescence was captured using a 525/36 nm band-pass filter. Arteries were treated with cyclopiazonic acid (CPA; 20 μM), a sarco-endoplasmic reticulum (SR/ER) Ca²⁺-ATPase inhibitor, for 15 minutes at 37°C before imaging to eliminate intracellular Ca²⁺-release signals. CPA does not per se alter the activity of endothelial TRPV4 (TRPV4_EC) sparklets (Hong et al., 2018). TRPV4_EC sparklet activity was determined before and 5 minutes after the addition of a given pharmacological agent. Ca²⁺ images were analyzed using custom-designed SparkAn software (developed by Dr. Adrian Bonev, University of Vermont). Fractional fluorescence traces (F/F₀) were obtained by placing a 1.7 μm² (5×5 pixels) region of interest (ROI) at the peak event amplitude. Representative F/F₀ traces were filtered using a Gaussian filter and a cutoff corner frequency of 4 Hz.

Analysis of TRPV4_EC sparklet activity.

TRPV4_EC sparklet activity was assessed by measuring increases in fluorescence over an averaged image obtained from 10 images using previously established methods (Sonkusare et al., 2012; Marziano et al., 2017; Hong et al., 2018). The average TRPV4_EC sparklet activity is defined as NP_O, where N is the number of TRPV4_EC channels per site and P_O is the open state probability of the channel. NP_O was calculated with the Single Channel Search module of Clampfit using quantal amplitudes derived from all-points histograms (0.29 ΔF/F₀ for Fluo-4-loaded MAs) by applying the following equation:

$$NPo=\left[\frac{\left(T_{\mathit{\text{level}}1}+2T_{\mathit{\text{level}}2}+3T_{\mathit{\text{level}}3}+4T_{\mathit{\text{level}}4}\right)}{T_{\mathit{\text{total}}}}\right],$$

where T represents the dwell time at each quantal level, and T_total is the total recording duration. Average NP_O per site was obtained by averaging the NP_O for all sites in a field. The total number of sites per field corresponds to all sparklet sites per field averaged over different arteries.

Analysis of TRPV4_EC sparklet localization at MEPs.

TRPV4_EC sparklet localization at MEPs was measured using Alexa Fluor 633 hydrazide staining of the internal elastic lamina (IEL). After performing Ca²⁺-imaging experiments, arteries were incubated with Alexa Fluor 633 hydrazide (10 μM) for 5 minutes (Marziano et al., 2017; Hong et al., 2018). Images were obtained from the same fields of view (matching X- and Y-coordinates) as used for recording TRPV4_EC sparklets, using an excitation wavelength of 640 nm and a band-pass emission filter (685/40 nm). Sparklet localization was assessed by overlaying the Fluo-4 image with Alexa Fluor 633 IEL staining. ROIs (1.7 μm²) corresponding to the peak sparklet fluorescence were placed on the overlaid IEL staining. Sparklet sites were considered to be localized to MEPs if the ROI was within 2 μm of the perimeter of holes in the IEL (Marziano et al., 2017; Hong et al., 2018). The remaining sparklet sites were considered to be non-MEPs sites.

Immunostaining.

Immunostaining was performed on en face preparations of MAs and PAs. Briefly, arteries were pinned down en face on Sylgard blocks and fixed with 4% paraformaldehyde at room temperature for 15 minutes. Fixed arteries were washed three times for 5 minutes with phosphate-buffered saline (PBS). The arteries were then treated with 0.2% Triton-X/PBS for 30 minutes at room temperature on a rocker. Following this permeabilization step, arteries were treated with 5% normal donkey serum or normal goat serum (Abcam plc, Cambridge, MA, USA) for 1 hour at room temperature and subsequently incubated overnight with antibodies against TRPV4, IK, SK, eNOS or Hbα overnight at 4°C (Table 1). Arteries were then washed three times with PBS and incubated with Alexa Fluor 568-conjugated donkey anti-rabbit or goat anti-rabbit secondary antibody (1:500; Life Technologies, Carlsbad, CA, USA), as appropriate, at room temperature for 1 hour in the dark. Thereafter, arteries were washed three times with PBS and incubated with 0.3 μM 4’,6-diamidino-2-phenylindole (DAPI; Invitrogen, Carlsbad, CA, USA) for 10 minutes at room temperature in the dark to stain nuclei. Images were obtained using the Andor imaging system as described previously (Marziano et al., 2017; Hong et al., 2018). Consecutive images were taken along the z-axis at a slice thickness of 0.2-μm from the top surface of ECs to the bottom surface where they contacted SMCs. DAPI immunostaining was imaged by exciting at 409 nm and collecting the emitted fluorescence with a 447/60-nm band-pass filter. The specificity of the antibody was determined using arteries from endothelial TRPV4^−/− knockout mice (Ottolini et al., 2020) and global eNOS^−/− mice for TRPV4 and eNOS, respectively. Blocking peptides were used in antibody control experiments for IK channel (REQVNSMVDISKMHMILYDL, Genscript USA Inc., Piscataway, NJ, USA), SK channel (ETQMENYDKHVTYNAERS, Genscript USA Inc.), and Hbα (Abcam plc., ab93083) antibodies. No immunostaining was observed under these conditions.

Table 1.

List of primary antibodies used for immunostaining.

Protein	Company	ID	Clonality	Concentration	References
TRPV4	LifeSpan BioScience, Inc.	LS_C94498	Polyclonal	1:200	(Ryskamp et al., 2011)
eNOS	BD Biosciences	610297	Monoclonal	1:100	(Laufs et al., 2000)
IK	Santa Cruz Biotechnology, Inc.	SC 365265	Monoclonal	1:200	(Lu et al., 2017)
SK	Alomone Labs	APC-028	Monoclonal	1:200	(Kramar et al., 2004)
Hbα	Abcam plc.	ab92492	Monoclonal	1:100	(Saha et al., 2017)

Automated co-localization analysis.

Imaris 9.3 image analysis software (Bitplane AG, Zurich, Switzerland) was used for automated analyses. Co-localization analyses were performed on three-dimensional (3D) images reconstructed from z-stack images from the top surface of the ECs to the point where MEPs contacted SMCs. The percentage of MEPs (black holes) that coincided with a given protein of interest was determined from automated counts of the total number of holes in each field of view. The green channel, showing the autofluorescence of the IEL, and black holes, indicating MEPs, were inverted so as to render MEPs in green. Background noise was then subtracted, and the image was filtered using a Gaussian filter (filter width, 0.2 μm). MEPs were automatically detected as green dots ≥ 2 μm in diameter. As a final visual confirmation of the accuracy of the automated detection of MEPs, the original green channel was re-inserted; the detection accuracy for MEPs was found to be ~95%. MEP detection was followed by automatic detection of the protein of interest, indicated by red immunostaining ≥ 2 μm in diameter. Finally, using a built-in Matlab R2019b co-localization feature in Imaris 9.3, MEPs and the protein of interest were considered to co-localize if the detected protein sites were within a distance of 4 μm from the detected MEPs. A similar procedure was followed for determining the percentage of immunostaining that co-localized with MEPs; immunostaining dots were detected first, followed by MEP detection, and co-localization analysis. To determine the probability that proteins were randomly distributed to MEPs, we randomly placed ROIs on immunostaining images and performed automated co-localization analyses. This analysis yielded 12% and 22% random co-localization of proteins with MEPs in MAs and PAs, respectively.

Patch-clamp analysis of freshly isolated ECs.

ECs were freshly isolated from third-order MAs and fourth-order PAs. Briefly, MAs and PAs were digested at 37°C for 60 and 30 minutes, respectively, in dissociation solution (55 mM NaCl, 80 mM Na-glutamate, 6 mM KCl, 2 mM MgCl₂, 0.1 mM CaCl₂, 10 mM glucose, 10 mM Hepes, pH 7.3) containing Worthington neutral protease (0.5 mg/mL). For MAs, collagenase (Worthington type 1, 0.5 mg/mL) was added to the enzyme solution after 60 minutes, and digestion was continued for two more minutes. Whole-cell currents were measured at room temperature using the perforated-patch configuration of the whole-cell patch-clamp technique. The bathing solution consisted of 10 mM Hepes, 134 mM NaCl, 6 mM KCl, 2 mM CaCl₂, 10 mM glucose, and 1 mM MgCl₂ (adjusted to pH 7.4 with NaOH). Patch electrodes were pulled from borosilicate glass (O.D., 1.5 mm; I.D., 1.17 mm; Sutter Instruments, Novato, CA, USA) using a Narishige PC-100 puller (Narishige International USA, Inc., Amityville, NY, USA) and polished using a MicroForge MF-830 polisher (Narishige International USA). The composition of the pipette solution for perforated-patch experiments was 10 mM Hepes, 30 mM KCl, 10 mM NaCl, 110 mM K-aspartate, and 1 mM MgCl₂ (adjusted to pH 7.2 with NaOH). Amphotericin B was dissolved in the intracellular pipette solution to reach a final concentration of 0.3 μM. The pipette resistance was 3–5 MΩ. IK and SK channel currents were elicited by adding 1 μM NS309 (IK/SK channel activator) or 10 and 100 nM GSK101 (TRPV4 channel agonist) to the superfusate. IK/SK channel currents were inhibited by adding TRAM-34 (IK channel inhibitor, 1 μM) and apamin (SK channel inhibitor, 300 nM) to the bath. Current traces obtained from freshly isolated ECs in the presence of TRAM-34 and apamin were subtracted from traces obtained in the presence of NS309 or GSK101 alone to yield TRAM-34 + apamin-sensitive IK/SK channel currents. IK/SK channel currents were recorded by applying 200-ms voltage ramps from -140 mV to +50 mV from a holding potential of −50 mV. The pipette solution for conventional patch clamp consisted of 10 mM HEPES, 123.2 mM KCl, 10 mM NaCl, 5.5 mM MgCl₂, 0.2 mM CaCl₂, and 5 mM HEDTA (adjusted to pH 7.2 with 16.8 mM KOH) and contained 3 μM free-Ca²⁺ and 1 mM free-Mg²⁺, as calculated using the Max-Chelator program (Chris Patton, Stanford University, CA, USA). TRPV4 channel current was measured in the presence of ruthenium red (1 μM) using the perforated-patch configuration as described previously (Sonkusare et al., 2012). Currents induced by the TRPV4 channel agonist GSK101 were assessed following application of a 200-ms voltage step from -50 to +100 mV. Data were acquired using a HEKA EPC 10 amplifier and PatchMaster v2X90 software (Harvard Bioscience, Holliston, MA, USA). Patch-clamp data were analyzed using FitMaster v2X73.2 (Harvard Bioscience) and MATLAB R2018a (MathWorks, Natick, MA, USA). TRPV4 channel currents were inhibited by applying 100 nM GSK2193874 (hereafter, GSK219), a selective TRPV4 antagonist, to the bath solution. The effect of each drug was studied 5 minutes after addition.

Proximity Ligation Assay (PLA).

Third-order MAs and fourth-order PAs were isolated and pinned down en face on a Sylgard block. Arteries were fixed in 4% paraformaldehyde for 15 minutes, washed three times with PBS, and then incubated in a solution of 0.2% Triton X for 30 minutes at room temperature. Following this latter permeabilization step, arteries were blocked by incubating with either 5% normal donkey serum (Abcam plc) or 300 mM glycine at room temperature for 1 hour. Arteries were then washed three times with PBS and incubated overnight at 4°C with primary antibodies. The following day, the PLA protocol was performed as described by the manufacturer of the Duolink PLA Technology kit (Sigma-Aldrich, St. Louis, MO, USA). After incubating arteries with 0.3 μM DAPI nuclear stain (Invitrogen, Carlsbad, CA, USA) for 10 minutes at room temperature in the dark, PLA imaging and analysis were performed using an Andor Revolution spinning-disk confocal imaging system and Imaris 9.3 software (Bitplane AG, Zurich, Switzerland), respectively. Images were obtained along the z-axis at a slice thickness of 0.02 μm from the top surface of ECs to the bottom surface where they contact SMCs.

NO measurements.

MAs and PAs were pinned down en face on a Sylgard block and incubated with 5 μM DAF-FM (4-amino-5 methylamino-2’,7’-difluorofluorescein diacetate), prepared in Hepes-PSS containing 0.02% pluronic acid (Marziano et al., 2017) for 20 minutes at 30°C in the dark. DAF-FM forms a fluorescent triazole compound after binding to NO. DAF-FM fluorescence was captured using an Andor Revolution WD (with Borealis) spinning-disk confocal imaging system. DAF-FM fluorescence was recorded using an excitation wavelength of 488 nm, and emitted fluorescence was captured with a 525/36-nm band-pass filter. Images were obtained along the z-axis at a slice thickness of 0.1 μm from the top surface of ECs to the bottom surface where they contact SMCs. For studying GSK101-induced NO release, arteries were treated with GSK101 (or vehicle for a control artery run in parallel) for 5 minutes, followed by GSK101 + DAF-FM (or vehicle + DAF-FM) for 20 minutes. Arteries were then placed in Hepes-PSS for 5 minutes before image acquisition, which was completed within 2 minutes. For studying the effect of HbαX peptide on GSK101-induced NO release, arteries were treated with HbαX (or vehicle) for 10 minutes, followed by GSK101 + HbαX (or vehicle) for 5 minutes and then GSK101 + HbαX (or vehicle) + DAF-FM for 20 minutes. DAF-FM vials were stored at −20°C and used within two months. Once dissolved in DMSO, DAF-FM solution was stored at −20°C and used within one week. The same DAF-FM solution was used for all comparison groups on each day of experiments.

DAF-FM fluorescence was analyzed using custom-designed SparkAn software. An outline was drawn around each endothelial cell to obtain the arbitrary fluorescence intensity of that cell. The plane with the peak fluorescence intensity was used for quantification. The background (intensity without laser) was then subtracted from the recorded fluorescence. The fluorescence values from all cells in a field of view were averaged to obtain a single fluorescence number for that field.

Western blotting.

MAs and PAs were lysed in radioimmunoprecipitation (RIPA) lysis buffer containing protease inhibitors (Life Technologies, Grand Island, NY, USA). Protein concentration was measured using a DC Protein Assay kit (Bio-Rad, CA, USA). Lysates (30 μg total protein) were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis on denaturing 4–12% gradient polyacrylamide ready-made gels (NuPAGE Bis-Tris gels; Life Technologies) and transferred onto PVDF (polyvinylidene difluoride) membranes. The membranes were blocked with 10% non-fat dried milk for 1 hour and incubated overnight with rabbit anti-eNOS antibody (NB-300–500, 1:500; Novus Biologicals, Littleton, CO, USA) at 4°C on a rotating shaker. The membranes were then washed three times with PBS containing 0.1% Tween-20 (PBST) and incubated with an anti-rabbit IgG-HRP (1:2500) secondary antibody for 90 minutes at room temperature. Immunoreactive proteins were visualized using enhanced chemiluminescence (ECL) detection reagents (SuperSignal West Femto Maximum Sensitivity Substrate; Life Technologies) as described previously (Kasetti et al., 2018). Blots were re-probed for β-actin to ensure equal protein loading. The relative density of bands was analyzed using Image J software.

Chemicals and reagents.

Cyclopiazonic acid, GSK1016790A, GSK2193874, NS309, apamin, and TRAM-34 were purchased from Tocris Bioscience (Minneapolis, MN, USA). L-NNA, Alexa Fluor 633 hydrazide, and DAF-FM were obtained from Thermo Fisher Scientific Inc. (Waltham, MA, USA). All other chemicals were purchased from Sigma-Aldrich. HbαX peptide and scrambled control peptide (Scr X) was obtained from Dr. Brant Isakson.

Statistical analysis.

Results are presented as means ± standard deviation. For imaging experiments (Ca²⁺ imaging, immunofluorescence, NO measurements, PLA) and pressure myography experiments, n=1 was defined as one artery, and for patch-clamp experiments, n=1 was defined as one cell. Data were obtained from at least three mice in experiments performed on at least two independent batches. All data were presented graphically using CorelDraw Graphics Suite X9 (Ottawa, ON, Canada) and were analyzed statistically using OriginPro 7.5 (Northampton, MA, USA), GraphPad Prism 8 (San Diego, CA, USA), and Matlab R2019b (Natick, MA, USA). The normality of data was determined by performing a Shapiro-Wilk test. Data were analyzed using two-tailed, paired, or independent t-tests for comparison of data collected from two different treatments, or one-way or two-way analysis of variance (ANOVA) for analysis of statistical differences among more than two different treatments. A post hoc Tukey’s correction was performed in cases where ANOVA results were significant. P-values < 0.05 were considered statistically significant.

RESULTS

1. TRPV4_EC channels control baseline diameter in PAs via eNOS activation and in MAs via IK/SK channel activation.

Pharmacological activation of TRPV4_EC channels dilates resistance arteries from multiple vascular beds (Bagher et al., 2012; Sonkusare et al., 2012; Marziano et al., 2017). However, the role of TRPV4_EC channels in regulating baseline arterial diameter remains unknown. Unlike MAs, which show robust pressure-induced constriction (myogenic tone) at physiological intravascular pressures of 60–80 mm Hg (Artamonov et al., 2018), PAs have little to no myogenic tone at their physiological pressure of 15 mm Hg. Therefore, to investigate the contribution of TRPV4 activity to baseline arterial diameter, we tested the effect of the specific TRPV4 inhibitor GSK2193874 (hereafter, GSK219) on baseline diameter in MAs and PAs from wild-type (WT) and TRPV4_EC^−/− mice pre-constricted with thromboxane receptor agonist U46619 (50 nM). GSK219 (100 nM) constricted both PAs and MAs from WT mice, an effect that was absent in arteries from TRPV4_EC^−/− mice, providing the first evidence for a dilator effect of TRPV4_EC channels on basal diameter in systemic and pulmonary resistance arteries (Fig. 1A and B). Interestingly, GSK219-induced constriction of PAs was abolished by NOS inhibition, which had no effect on GSK219-induced constriction of MAs, whereas GSK219-induced constriction of MAs was abolished by IK/SK channel inhibition, which had no effect on PAs (Fig. 1C). In MAs with myogenic tone, GSK219 caused a constriction similar to that in U46619–pre-constricted arteries. This effect was abolished by IK/SK channel inhibition, was unaffected by NOS inhibition and was absent in MAs from TRPV4_EC^−/− mice (Fig. 1D and E), suggesting that pre-constriction with U46619 did not alter TRPV4_EC channel-dependent downstream signaling. These results provide the first evidence for the control of basal diameter by TRPV4_EC channel–eNOS signaling in PAs and by TRPV4_EC channel–IK/SK channel signaling in MAs.

Figure 1. TRPV4_EC channels control resting diameter in resistance PAs and MAs.

Data are presented as means ± standard deviation. A, representative diameter traces showing effects of the TRPV4 channel inhibitor GSK219 (100 nM) on PAs (left) and MAs (right) isolated from WT and TRPV4_EC^−/− mice, and on WT PAs and MAs in the presence of IK/SK channel inhibitors (1 μM TRAM-34, 300 nM apamin) or NOS inhibitor L-NNA (100 μM). Fourth‐order PAs (pressurized to 15 mm Hg) and third‐order MAs (pressurized to 80 mm Hg) were pre-constricted with the thromboxane analog U46619 (50 nM). B, percentage GSK219 (100 nM)-induced constriction of PAs and MAs isolated from WT and TRPV4_EC^−/− mice (n=5; one-way ANOVA). C, percentage GSK219 (100 nM)-induced constriction of PAs and MAs in the presence of IK/SK channel inhibitors (1 μM TRAM-34 + 300 nM apamin) or NOS inhibitor (100 μM L-NNA) (n=5; one-way ANOVA). D, representative diameter traces of GSK219-induced constriction of MAs with myogenic tone in the absence (left) and presence (right) of IK/SK channel inhibitors. E, percentage GSK219 (100 nM)-induced constriction of MAs with myogenic tone in the presence of IK/SK channel inhibitors or NOS inhibitor in WT mice, or in MAs from TRPV4_EC^−/− mice (n=5–7; one-way ANOVA). The vehicle itself (DMSO) did not alter the diameter of PAs and MAs (n=5; t-test, paired; P=0.5349 for PAs and 0.6079 for MAs).

2. Low-level activation of TRPV4_EC channels induces IK/SK channel currents in ECs from MAs, but not PAs.

Activation of TRPV4_EC channels with relatively low concentrations (10 nM) of the specific agonist GSK1016790A (hereafter, GSK101) has been shown to elicit localized Ca²⁺ influx signals through TRPV4_EC channels, termed TRPV4_EC sparklets, without increasing global Ca²⁺ levels (Sonkusare et al., 2012; Marziano et al., 2017). To determine whether TRPV4_EC channel activation induces IK/SK currents, we performed whole-cell patch-clamp studies in freshly isolated ECs from MAs and PAs treated with 10 nM GSK101. GSK101 induced robust IK/SK channel currents in ECs from MAs that were inhibited by TRAM-34 (IK channel inhibitor, 1 μM) + apamin (SK channel inhibitor, 300 nM) (Fig. 2A and B). In sharp contrast, GSK101 failed to evoke IK/SK channel currents in ECs from PAs (Fig. 2A and B). This absence of an effect of GSK101 was not attributable to decreased IK/SK expression in PAs, since direct activation of IK/SK channels with NS309 (1 μM) induced similar IK/SK channel current densities in ECs from MAs and PAs (Fig. 2C and D). These results provide the first direct evidence of functional IK/SK channels in native ECs from resistance PAs and suggest that TRPV4_EC sparklet–IK/SK channel coupling occurs in resistance MAs, but not in resistance PAs.

Figure 2. Whole-cell, but not localized, increases in intracellular Ca²⁺ activate IK/SK currents in PAs.

Data are presented as means ± standard deviation. A, representative traces for ionic currents in freshly isolated ECs from MAs and PAs under baseline conditions, in the presence of GSK101 (10 nM), and GSK101 + TRAM-34 (1 μM)/apamin (Apa, 300 nM), recorded in the perforated-patch configuration. B, current density plot of outward currents at 0 mV under baseline conditions, in the presence of GSK101 (10 nM), and GSK101 + TRAM-34/apamin (n=5; t-test, unpaired). C, representative traces for ionic currents under baseline conditions, in the presence of NS309 (1 μM), and NS309 + TRAM-34/apamin in ECs from MAs and PAs in the perforated-patch configuration. D, current density plot of outward currents at 0 mV under baseline conditions, in the presence of NS309, and NS309 + TRAM-34/apamin (n=7; t-test, unpaired). E, grayscale images of en face preparations of PAs and MAs loaded with Fluo-4AM before (left) and after the addition of GSK101 (100 nM) (right). F, averaged whole-cell Fluo-4AM fluorescence in ECs from PAs and MAs before (Baseline) and after the addition of GSK101 (10 and 100 nM, n=5, one-way ANOVA). G, representative traces for baseline currents and GSK101 (100 nM)-activated, TRAM-34 (1 μM)/apamin (300 nM)-sensitive IK/SK channel currents in freshly isolated ECs from PAs in the perforated patch-clamp configuration. H, representative traces of TRAM-34/apamin-sensitive IK/SK currents evoked by 3 μM free Ca²⁺ in freshly isolated ECs from PAs and MAs in the whole-cell configuration. I, density plots for outward currents at 0 mV under baseline conditions, and TRAM-34/apamin-sensitive IK/SK channel currents activated by GSK101 (100 nM) or 3 μM free Ca²⁺ (n=5; one-way ANOVA).

3. Whole-cell increases in intracellular Ca²⁺ activate IK/SK channel currents in pulmonary ECs.

IK/SK channels are primarily activated by increases in intracellular Ca²⁺ (Ledoux et al., 2006). Both IK and SK channels possess a calmodulin (CaM)-binding site, and Ca²⁺/CaM binding to this site induces a conformational change that opens the channel pore (Fanger et al., 1999; Ledoux et al., 2006). To verify that IK/SK channels in PAs are not sensitive to Ca²⁺/CaM-induced activation, we assessed IK/SK channel activity in ECs in response to increased whole-cell Ca²⁺ levels, induced using a high concentration of GSK101 and by increasing the concentration of free Ca²⁺ in the patch pipette. At a low concentration (10 nM), GSK101 induced spatially restricted TRPV4_EC sparklet activity but did not alter whole-cell Ca²⁺ fluorescence (Fig. 2E and F). In contrast, excessive TRPV4 channel activation with 100 nM GSK101 increased whole-cell Ca²⁺ fluorescence intensity in both PAs and MAs (Fig. 2E and F). These whole-cell increases in endothelial Ca²⁺ correlated with an increase in IK/SK channel currents in PAs (Fig. 2G), suggesting that whole-cell increases in Ca²⁺ can activate IK/SK channels in PAs. The Ca²⁺-sensitivity of IK/SK channels in ECs from PAs was further tested using the conventional whole-cell patch-clamp configuration under conditions in which the concentration of free Ca²⁺ in the patch pipette was increased to 3 μM to cause maximal Ca²⁺-dependent activation of IK/SK channels. Under these conditions, the density of IK/SK channel currents was similar between ECs from PAs and MAs (Fig. 2H and I). These results confirm that increases in intracellular Ca²⁺ can activate IK/SK channels in ECs from PAs and that the maximum density of IK/SK channel currents is similar in ECs from resistance PAs and MAs.

4. ECs from MAs and PAs show similar TRPV4_EC channel activity.

Lower expression/activity of TRPV4_EC channels could be one possible explanation for the lack of GSK101-induced IK/SK channel currents in PAs. We addressed this possibility by recording currents through TRPV4_EC channels in ECs from MAs and PAs and by recording the activity of TRPV4_EC sparklets in en face preparations of MAs and PAs. Outward currents through TRPV4_EC channels were recorded in the presence of GSK101 (10 nM) and ruthenium red (RuR, 1 μM), the latter of which blocks Ca²⁺ influx at negative voltages, preventing Ca²⁺ overload and Ca²⁺-dependent activation of IK/SK currents (Sonkusare et al., 2012). The remaining outward currents were inhibited by the selective TRPV4 channel inhibitor GSK219 (Fig. 3A). GSK101 induced comparable GSK219-sensitive outward currents in ECs from MAs and PAs, suggesting similar TRPV4_EC channel activity in these arteries (Fig. 3A and B). Furthermore, recordings of elementary Ca²⁺ influx through TRPV4_EC channels (TRPV4_EC sparklets) in the intact endothelium of MA and PA en face preparations in the absence or presence of the TRPV4_EC channel activator GSK101 (0, 3, 10 nM) showed that TRPV4_EC sparklet activity was similar between PAs and MAs (Fig. 3C–E). Spatial localization analysis confirmed that TRPV4_EC sparklet activity was higher at MEPs than non-MEP sites in both PAs and MAs (Fig. 3F), and that overall TRPV4_EC sparklet activity was similar between PAs and MAs at both MEP and non-MEP locations. These data suggest that TRPV4_EC channel activity is similar between PAs and MAs and confirm that the absence of TRPV4_EC-IK/SK channel coupling in PAs is not attributable to diminished TRPV4_EC channel activity.

Figure 3. GSK101 induces similar TRPV4_EC channel activity in ECs from MAs and PAs.

Data are presented as means ± standard deviation. A, representative traces for GSK101 (10 nM)-induced outward TRPV4_EC currents in freshly isolated ECs from PAs (left) and MAs (right). Application of the TRPV4 channel inhibitor GSK219 (100 nM) attenuated TRPV4_EC currents. TRPV4_EC currents were evoked by a 200-ms pulse from −50 mV to +100 mV. Ruthenium red (1 μM) was present throughout the experiment to inhibit Ca²⁺-induced IK/SK channel currents and prevent Ca²⁺ overload. B, scatter plot showing GSK101-induced outward currents at +100 mV, and inhibition by GSK219 (n=5; one-way ANOVA). C, representative images for en face preparations of PAs (left) and MAs (right) loaded with Fluo-4AM and counterstained with Alexa Fluor 633 hydrazide (10 μM). TRPV4_EC sparklet sites localized at holes in the IEL were indicative of MEPs. D, representative F/F₀ traces for TRPV4_EC sparklets in PAs (left) and MAs (right) in the presence of 10 nM GSK101, to activate TRPV4_EC channels, and cyclopiazonic acid (CPA, 20 μM), to eliminate Ca²⁺ release from intracellular stores. E, baseline (CPA alone) and GSK101 (3 and 10 nM)-induced TRPV4_EC sparklet activity in PAs and MAs, expressed as NP_O per site, where N is the number of channels and P_O is open state probability (n=5; one-way ANOVA). F, GSK101 (3 nM)-induced TRPV4_EC sparklet activity at MEP and non-MEP sites in PAs and MAs (n=5–13, one-way ANOVA).

5. TRPV4_EC channels signal through eNOS activation to dilate PAs and through IK/SK channel activation to dilate MAs.

Previous studies have shown that TRPV4_EC channel activation causes endothelium-dependent dilation of resistance MAs and PAs (Sonkusare et al., 2012; Marziano et al., 2017). Studies using the fluorescent NO indicator DAF-FM showed that TRPV4 channel activation with GSK101 (10 nM) increased endothelial NO levels in PAs, but not MAs (Fig. 4A and B). Moreover, GSK101-induced NO release was absent in PAs from eNOS^−/− mice, confirming operation of the TRPV4_EC-eNOS-NO pathway in these arteries (Fig. 4A and B). In pressurized PAs, vasodilation induced by GSK101 (3–30 nM) was completely abolished by NOS inhibition with L-NNA and was not affected by inhibition of IK/SK channels (Fig. 4C and D). In contrast, GSK101-induced dilation of MAs was largely abolished by IK/SK channel inhibition with TRAM-34 + apamin, but was unaffected by NOS inhibition with L-NNA (Fig. 4C and D). The direct IK/SK channel activator NS309 caused dilation of both PAs and MAs, indicating that communication via myoendothelial junctions occurs in both vascular beds (Fig. 4E). These results, together with our previous demonstration that TRPV4_EC channel-induced vasodilation is absent in PAs from eNOS^−/− mice (Marziano et al., 2017), support vasodilatory TRPV4_EC-eNOS coupling in PAs (but not in MAs) and TRPV4_EC-IK/SK channel coupling in MAs (but not in PAs).

Figure 4. TRPV4_EC channels act through IK/SK channel activation to dilate MAs and through eNOS-NO signaling to dilate PAs.

Data are presented as means ± standard deviation. A, representative images for DAF‐FM fluorescence in ECs of PAs and MAs under basal conditions (left) and in the presence of the TRPV4 channel agonist GSK101 (10 nM) (right). B, averaged DAF‐FM fluorescence in ECs from PAs and MAs of WT and eNOS^−/− mice in the absence or presence of 10 nM GSK101 (n=5, one-way ANOVA). C, representative diameter traces for GSK101 (3–30 nM)-induced dilation of PAs (left) and MAs (right) pre-constricted with the thromboxane analog U46619 (50 nM), and effects of TRAM-34 (1 μM) + apamin (300 nM) and the NOS inhibitor L-NNA (100 μM). Third‐order MAs were pressurized to 80 mm Hg, and fourth‐order PAs were pressurized to 15 mm Hg. D, percent dilation of PAs (left) and MAs (right) in response to GSK101 (3–30 nM) under control conditions, in the presence of TRAM-34 + apamin, and in the presence of L-NNA (n=5, P value vs Control, two-way ANOVA). E, scatter plot for NS309 (1 μM)-induced dilation of PAs and MAs (n=6–7; t-test, unpaired).

6. TRPV4_EC and IK/SK channels localize at MEPs in MAs but not in PAs.

The activation of IK/SK channels in pulmonary ECs by whole-cell increases in Ca²⁺, but not by TRPV4_EC sparklets, suggests that the spatial proximity between TRPV4_EC channels and IK/SK channels necessary for local coupling of TRPV4_EC sparklets to IK/SK channels is altered in pulmonary ECs. To test whether the spatial proximity of TRPV4_EC channels with eNOS in PAs and with IK/SK channels in MAs determines the specific TRPV4_EC sparklet-target coupling, we immunohistochemically assessed co-localization of TRPV4_EC channels with eNOS and IK/SK channels in PAs and MAs. TRPV4_EC channels localized mainly at MEPs in PAs and MAs, and almost all MEPs in both vascular beds showed TRPV4_EC channel expression (Fig. 5A–D), consistent with the localization of TRPV4_EC sparklets. While the majority of IK and SK channels were localized to MEPs in MAs, only a small fraction of these channels were found at MEPs in PAs, suggesting that IK and SK channels do not co-localize with TRPV4_EC channels at MEPs in PAs. Interestingly, ~50% of MEPs showed robust eNOS staining in both PAs and MAs, suggesting close proximity of TRPV4_EC with eNOS in both arteries (Fig. 5A–D). We also detected eNOS expression at other locations, possibly indicating Golgi/ER localization, as previously reported (Fig. 5A) (Fulton et al., 2002; Govers et al., 2002; Sehgal et al., 2007). These results support the concept that the absence of TRPV4_EC-IK/SK channel coupling in PAs is attributable to the lack of spatial proximity between TRPV4_EC and IK/SK channels, but do not explain the preferential TRPV4_EC-eNOS signaling in PAs and its absence in MAs.

Figure 5. TRPV4_EC-IK/SK channels localize at MEPs in MAs but not in PAs.

Data are presented as means ± standard deviation. A, representative merged images from en face preparations of fourth-order PAs (left) and third-order MAs (right) showing IEL autofluorescence (green) and TRPV4_EC, IK, SK, or eNOS immunofluorescence (red). Black holes represent MEPs. Traces under each image indicate representative fluorescence intensity plot profiles for 0.2-μm-thick lines across MEPs. Green line, IEL; red line, protein of interest; gray rectangles, MEPs. White arrows in the inset images indicate eNOS localization at MEPs; scale= 10 μm. B, representative antibody control images for IK and SK channel antibodies (using blocking peptides in MAs) and eNOS antibody (PAs from eNOS^−/− mice). TRPV4 antibody has previously been validated using TRPV4_EC^−/− mice (Ottolini et al., 2020). C, percent of MEPs localized with immunofluorescence (n=6; one-way ANOVA). D, percentage of immunofluorescence staining localized with MEPs (n=5–9; one-way ANOVA).

7. Hemoglobin α (Hbα) localizes with eNOS at MEPs in MAs but not in PAs.

To more precisely compare eNOS levels between PAs and MAs, we studied eNOS expression using western blotting. The expression of eNOS protein was not different between PAs and MAs (Fig. 6A), suggesting that regulation of eNOS activity, rather than its expression, underlies the differential contribution of eNOS in PAs and MAs. Because Hbα has been shown to interact with eNOS and scavenge NO, thereby limiting NO bioavailability (Straub et al., 2012; Keller et al., 2016; Shu et al., 2019), we hypothesized that Hbα limits the role of eNOS-NO signaling in MAs. Immunostaining studies showed minimal expression of Hbα in ECs from PAs but confirmed strong expression and localization of Hbα at MEPs in MAs (Fig. 6B–D). Additionally, proximity ligation assays (PLAs) confirmed nanometer proximity between eNOS and Hbα in MAs, but not in PAs (Fig. 6E and F). These results support the concept that the presence of Hbα close to eNOS may be responsible for limiting the role of eNOS in MAs. Inhibiting NOS with L-NNA caused a ~12% constriction in PAs but did not affect the diameter of MAs (Fig. 6G). HbαX, a peptide that inhibits the interaction of Hbα with eNOS, (Straub et al., 2014; Keller et al., 2016; Shu et al., 2019), was used to determine whether Hbα-eNOS interactions limit the role of eNOS in MAs; a scrambled control peptide (Scr X) was used as a control. Both HbαX and Scr X possess a tat tag, which has previously been shown to render these peptides cell permeable (Straub et al., 2014; Keller et al., 2016; Shu et al., 2019). Strikingly, MAs treated with HbαX were capable of constricting in response to L-NNA, suggesting that Hbα and eNOS normally interact in these arteries and that disruption of Hbα-eNOS interactions with HbαX disinhibited TRPV4_EC-eNOS signaling. Consistent with this, L-NNA was unable to constrict MAs in the presence of Scr X. Neither HbαX nor Scr X affected L-NNA–induced constriction of PAs (Fig. 6G), supporting the conclusion that Hbα and eNOS do not normally interact in these cells. Interestingly, GSK101-induced increases in NO levels were enhanced in MAs in the presence of HbαX, but were unaffected by HbαX in PAs (Fig. 6H). Additionally, the increase in NO in the presence HbαX was absent in MAs from eNOS^−/− mice, supporting the eNOS-specific nature of the HbαX effect in MAs. These results confirm that the spatial proximity of TRPV4_EC channels with IK/SK/eNOS and of eNOS with Hbα determines the vasodilatory signaling target of TRPV4_EC channels in resistance arteries (Fig. 7).

Figure 6. Hbα co-localizes with eNOS at MEPs in MAs but not in PAs.

Data are presented as means ± standard deviation. A, Western blot (left) and densitometric analysis (right) of eNOS levels showing similar eNOS expression in both MAs and PAs (n=4; t-test, unpaired). B, representative merged images from en face preparations of PAs (left) and MAs (right) showing IEL autofluorescence (green) and Hbα immunofluorescence (red). Black holes represent MEPs. Traces under each image are fluorescence intensity plot profiles for 0.2-μm-thick lines across MEPs. Green line, IEL; red line, protein of interest. C, percent of MEPs localized with red immunofluorescence (left) (n=5; t-test, unpaired) and percentage of total immunofluorescence (red) staining at MEPs (right) (n=5–8, t-test, unpaired). D, representative Hbα immunostaining image from an antibody control experiment using a blocking peptide in MA. E, representative PLA merged images showing EC nuclei (blue) and red puncta for eNOS:Hbα co-localization in en face preparations of PAs (left) and MAs (right). F, quantification of eNOS:Hbα PLA puncta in PAs and MAs (n=8–9; t-test, unpaired). G, percent constriction to L-NNA (100 μM) in the absence or presence of the control scrambled peptide Scr X (5 μM) or inhibitory peptide HbαX (5 μM) in PAs and MAs (n=5; one-way ANOVA). H, averaged fold increase in GSK101 (10 nM)-induced DAF-FM fluorescence in the presence or absence of HbαX (5 μM) in PAs and MAs from WT mice and in MAs from eNOS^−/− mice (n=5, one-way ANOVA).

Figure 7. Schematic depicting the mechanisms underlying preferential TRPV4_EC sparklet-eNOS versus TRPV4_EC sparklet-IK/SK channel coupling in resistance PAs and MAs.

Endothelial projections to SMCs in PAs and MAs are shown. EC: endothelial cell; eNOS: endothelial nitric oxide synthase; NO: nitric oxide; SMC: smooth muscle cell; TRPV4_EC: endothelial cell transient receptor potential vanilloid 4 channel; IK/SK: intermediate and small conductance Ca²⁺-sensitive K⁺ channels; Hbα: hemoglobin alpha; MEP: myoendothelial projection.

DISCUSSION

Heterogeneity in endothelial vasodilatory mechanisms is well known; however, the preferential activation of one vasodilatory pathway over others remains poorly understood. Pressure myography studies have shown that IK/SK channel-induced hyperpolarization is the predominant mechanism for dilation of systemic resistance arteries (Taylor et al., 2003; Ledoux et al., 2008b; Earley et al., 2009; Bagher et al., 2012; Sonkusare et al., 2012; Tran et al., 2012; Pires et al., 2015; Sullivan et al., 2015), whereas Ca²⁺ signal-eNOS coupling dominates in resistance PAs (Marziano et al., 2017). The current study advances our understanding of heterogeneity in endothelial vasodilatory pathways via three main findings: 1) TRPV4_EC channels control the basal diameter of resistance PAs and MAs; 2) although functional IK/SK channels and eNOS are present in ECs from both PAs and MAs, TRPV4_EC channels preferentially activate eNOS in PAs and IK/SK channels in MAs; and 3) spatial proximity of TRPV4_EC channels with IK/SK channels or eNOS and that of eNOS with Hbα determines the vasodilatory target activated by TRPV4_EC channels. Moreover, the dilatory effect of TRPV4_EC channels on baseline diameter may underlie the higher resting blood pressure in TRPV4_EC^−/− mice (Ottolini et al., 2020). Distinct factors regulate systemic and pulmonary arterial pressures; therefore, understanding the different mechanisms that control endothelial function in systemic and pulmonary microcirculations is a crucial first step in achieving selective regulation of one versus the other. Additionally, many cardiovascular disorders have been associated with selective impairment of either eNOS-mediated or IK/SK channel-mediated vasodilation (Forstermann & Munzel, 2006; Klinger et al., 2013; Ma et al., 2013; Sonkusare et al., 2014; Seki et al., 2017). Spatial uncoupling of TRPV4_EC channels from IK/SK channels or eNOS, or of eNOS and Hbα, can be explored as a potential mechanism for endothelial dysfunction in vascular disorders.

Although IK/SK channel currents have been reported in cultured pulmonary ECs (Lin et al., 2015), whether functional IK/SK channels are present in native ECs from the pulmonary circulation had remained unknown. Our results represent the first recordings of IK/SK channel currents in freshly isolated ECs from resistance PAs. Moreover, vasodilation in response to direct IK/SK channel activation (Fig. 4E) provides functional evidence for endothelial-to-smooth muscle communication via MEGJs in PAs. It is important to note that the number of functional IK/SK channels per EC is comparable between MAs and PAs. While TRPV4_EC channel-dependent activation of IK/SK channels in PAs requires whole-cell increases in endothelial Ca²⁺, localized TRPV4_EC sparklets are sufficient to activate IK/SK channels in MAs. These findings, together with immunostaining results showing that IK/SK channels localize at MEPs in MAs but not in PAs (Fig. 5A–D), suggest that the absence of TRPV4_EC-IK/SK channel coupling in PAs is attributable to the lack of proximity between TRPV4_EC and IK/SK channels.

TRPV4_EC sparklets do not couple with IK/SK channels in PAs, which begs the question: what is the role of endothelial IK/SK channels in PAs? IK/SK channels may be activated by another localized Ca²⁺ signal in PAs, possibly pointing to a physiological stimulus–Ca²⁺ signal-IK/SK channel linkage that is unique to PAs. The exact physiological stimulus or Ca²⁺ signal that activates IK/SK channels in PAs remains unknown. Pulmonary arterial pressure (PAP) is regulated by a plethora of factors, and there is likely to be some redundancy in the mechanisms that can maintain PAP homeostasis. The presence of IK/SK channels in the PA endothelium could be a redundant mechanism that comes into play in response to pathological stimuli. eNOS regulates basal PA diameter (Fig. 4C and D) and resting PAP (Fagan et al., 1999; Marziano et al., 2017), and NO signaling is known to be impaired in pulmonary vascular disorders, including pulmonary arterial hypertension (PAH) (Klinger et al., 2013). Interestingly, it has been shown that expression of endothelial Hbα is increased in pulmonary hypertension, which may explain the loss of NO signaling (Alvarez et al., 2017). A reasonable speculation is that a change in the localization of IK/SK channels at MEPs can, to a certain extent, offset the loss of NO signaling under pathological conditions.

Pressure myography studies are thought to more closely resemble physiological conditions compared with wire myography or tension recording studies (Buus et al., 1994; Falloon et al., 1995; Schubert et al., 1996; Jadeja et al., 2015). Indeed, profound differences in endothelial vasodilatory pathways have been observed between pressure myography and wire-myography/tension recording experiments (Boettcher & de Wit, 2011). Resistance-sized systemic arteries are known to develop pressure-induced constriction (myogenic tone) (Bayliss, 1902; Davis & Hill, 1999)—a physiological autoregulatory mechanism. Because all MAs developed myogenic tone at 80 mmHg and only a small fraction of PAs showed myogenic tone at their physiological pressure of 15 mmHg, vascular reactivity was studied under conditions in which both MAs and PAs were pre-constricted with U46619. GSK219-induced, IK/SK channel-dependent constriction was similar in MAs with myogenic tone and those pre-constricted with U46619. Moreover, previous studies in pressurized MAs have shown that the IK/SK channel-dependent nature of TRPV4_EC channel-induced vasodilation is preserved in the presence of U46619 (Marziano et al., 2017). Thus, it can be assumed that the presence of U46619 alone did not affect the downstream targets of TRPV4_EC channels.

Anchoring proteins that influence the activity of TRPV4 channels, IK/SK channels, and eNOS may also contribute to the heterogeneity of vasodilatory endothelial signaling. For example, the activity of TRPV4_EC channels at MEPs is enhanced by A kinase anchoring protein 150 (AKAP150) in MAs (Sonkusare et al., 2014); however, AKAP150 is not detected in ECs from PAs (Marziano et al., 2017). The activity of TRPV4_EC channels in PAs is likely controlled by other scaffolding proteins. In this regard, caveolin-1 (Cav-1), a crucial protein in the pulmonary circulation, has been shown to interact with both TRPV4 channels and eNOS in ECs (Saliez et al., 2008; Goedicke-Fritz et al., 2015); (Ju et al., 1997; Bernatchez et al., 2005; Chen et al., 2018). The possibility that Cav-1–enriched membrane invaginations provide a signaling scaffold for TRPV4_EC-eNOS signaling in PAs remains to be tested.

The mechanisms underlying the differential spatial coupling between TRPV4_EC channels and IK/SK channels or eNOS, and between eNOS and Hbα are not known. Compared to systemic resistance arteries, pulmonary resistance arteries are exposed to a lower pressure, higher flow, and a lower dissolved oxygen content. It is plausible that chronic exposure to vastly different environments ultimately results in the differential signaling organization observed in this study. Moreover, the functional consequences of differential coupling of TRPV4_EC channels with eNOS or IK/SK channels are not known. In PAs, TRPV4_EC channel-induced NO was shown to limit TRPV4_EC sparklet activity (NP_O per site) (Marziano et al., 2017). This NO-dependent negative feedback loop was absent in MAs, and inhibition of IK/SK channels did not alter the activity of TRPV4_EC sparklets in MAs (0.048 ± 0.003 and 0.054 ± 0.04 in the presence of 3 nM GSK101 before and after the addition of TRAM-34 and apamin; n=3, P=0.098), possibly reflecting tighter regulation of TRPV4_EC channel activity and its functional consequences in the pulmonary microcirculation.

One of the limitations of our study is that it used autofluorescence of the IEL as a marker of endothelial projections. While this commonly used approach allows localization analyses of Ca²⁺ signals and proteins, it should be noted that the presence of a hole in the IEL is not a direct indicator of the presence of MEGJs. Other matrix proteins, including collagen, fibronectin and laminin, may be present in the areas of the IEL that lack elastin, potentially complicating localization analyses. Our automatic detection of co-localization suggested that ~90% of holes in the IEL are sites of TRPV4_EC channel co-localization, supporting the presence of endothelial projections in most of the holes. However, the lengths of endothelial projections inside these holes, and the percentage of projections that go on to form MEGJs, is not known.

In conclusion, our results demonstrate that selective TRPV4_EC channel-IK/SK channel coupling in resistance MAs is made possible by the spatial proximity of TRPV4_EC channels with IK/SK channels and by the presence of Hbα, which serves to limit eNOS-NO signaling (Fig. 7). In contrast, preferential TRPV4_EC channel-eNOS signaling in PAs can be explained by the spatial proximity between TRPV4_EC channels and eNOS, the absence of Hbα, and the lack of proximity between TRPV4_EC channels and IK/SK channels in this vascular bed. Thus, our studies reveal the unique spatial organizations that underlie endothelial heterogeneity in vasodilatory signaling pathways. Notable in this context, systemic hypertension has been associated with a loss of IK/SK-channel dependent vasodilation in MAs (Kohler et al., 2010; Sonkusare et al., 2014), whereas pulmonary hypertension has been associated with a loss of eNOS-mediated vasodilation in PAs (Klinger et al., 2013). Impaired Ca²⁺ signal-target coupling may provide a potential mechanism for endothelial dysfunction in vascular disorders.

KEY POINTS.

• Endothelial cell TRPV4 (TRPV4_EC) channels exert a dilatory effect on the resting diameter of resistance mesenteric and pulmonary arteries.

• Functional IK/SK channels and eNOS are present in the endothelium of mesenteric and pulmonary arteries.

• TRPV4_EC sparklets preferentially couple with IK/SK channels in mesenteric arteries and with eNOS in pulmonary arteries.

• TRPV4_EC channels co-localize with IK/SK channels in mesenteric arteries, but not in pulmonary arteries, which may explain TRPV4_EC-IK/SK channel coupling in mesenteric arteries and its absence in pulmonary arteries.

• The presence of the nitric oxide-scavenging protein, hemoglobin α, limits TRPV4_EC-eNOS signaling in mesenteric arteries.

• Spatial proximity of TRPV4_EC channels with eNOS and the absence of hemoglobin α favor TRPV4_EC-eNOS signaling in pulmonary arteries.