NADPH oxidase 4 contributes to TRPV4-mediated endothelium-dependent vasodilation in human arterioles by regulating protein phosphorylation of TRPV4 channels

Abstract

Background

Impaired endothelium-dependent vasodilation has been suggested to be a key component of coronary microvascular dysfunction (CMD). A better understanding of endothelial pathways involved in vasodilation in human arterioles may provide new insight into the mechanisms of CMD.

Objective

To investigate the role of TRPV4, NOX4, and their interaction in human arterioles and examine the underlying mechanisms.

Methods and Results

Arterioles were freshly isolated from adipose and heart tissues obtained from 71 patients without coronary artery disease, and vascular reactivity was studied by videomicroscopy. In human adipose arterioles (HAA), ACh-induced dilation was significantly reduced by TRPV4 inhibitor HC067047 and by NOX 1/4 inhibitor GKT137831, but GKT137831 did not further affect the dilation in the presence of TRPV4 inhibitors. GKT137831 also inhibited TRPV4 agonist GSK1016790A-induced dilation in HAA and human coronary arterioles (HCA). NOX4 transcripts and proteins were detected in endothelial cells of HAA and HCA. Using fura-2 imaging, GKT137831 significantly reduced GSK1016790A-induced Ca²⁺ influx in the primary culture of endothelial cells and TRPV4-WT-overexpressing human coronary artery endothelial cells (HCAEC). However, GKT137831 did not affect TRPV4-mediated Ca²⁺ influx in non-phosphorylatable TRPV4-S823A/S824A-overexpressing HCAEC. In addition, treatment of HCAEC with GKT137831 decreased the phosphorylation level of Ser824 in TRPV4. Finally, proximity ligation assay (PLA) revealed co-localization of NOX4 and TRPV4 proteins.

Conclusion

Both TRPV4 and NOX4 contribute to ACh-induced dilation in human arterioles from patients without coronary artery disease. NOX4 increases TRPV4 phosphorylation in endothelial cells, which in turn enhances TRPV4-mediated Ca²⁺ entry and subsequent endothelium-dependent dilation in human arterioles.

Introduction

The coronary microcirculation includes small arteries and arterioles with diameters below 300–400 μm, and plays a critical role in the regulation of vascular resistance and tissue perfusion [8]. Accumulating evidence strongly suggests that coronary microvascular dysfunction (CMD) exists in many patients with coronary heart disease [30, 59], which can account for up to 50–65% of the patients with angina-like symptoms who failed to meet the diagnostic criteria of obstructive coronary artery disease by coronary angiography [42]. Additionally, CMD is associated with an increased risk of major adverse cardiovascular events [51]. Coronary endothelial impairment has been suggested to be a key component of CMD [49, 55]. A recent study has further shown that this impairment is mirrored in the systemic microvasculature, as indicated by reduced relaxation responses to the endothelium-dependent vasodilator acetylcholine (ACh) in adipose arterioles biopsied from patients with microvascular angina [19]. However, the pathogenic mechanisms underlying CMD remain largely unclear. A better understanding of endothelial signaling pathways, especially in human arterioles, may provide important insight into the mechanisms of microvascular dysfunction.

The endothelium regulates vascular tone by releasing various vasoactive factors, including nitric oxide (NO), prostacyclin (PGI₂), and endothelium-derived hyperpolarizing (EDH) factors [15] most prominently in resistance arteries and arterioles. A common pathway responsible for the stimulus-induced release of endothelial factors involves an increase in the intracellular Ca²⁺ concentration ([Ca²⁺]_i) in endothelial cells (EC) [44]. Transient receptor potential (TRP) vanilloid 4 (TRPV4), the fourth member of the TRP vanilloid subfamily, is a non-selective Ca²⁺-permeable cation channel expressed in vascular ECs of multiple species [18, 70]. It can be activated by diverse chemical and physical stimuli, such as the phorbol derivative 4α-phorbol-12,13-decanoate (4α-PDD) [63], arachidonic acid (AA) and its metabolites, hypotonic cell swelling [35, 57, 65], moderate heat [22], shear stress [20], and the endothelium-dependent vasodilator ACh [71]. Accordingly, TRPV4 has been implicated in a variety of physiological processes and diseases, including endothelium-dependent vasodilation [31, 46]. Previous studies have shown that TRPV4 contributes to ACh-induced dilation in various animal resistance arteries [18, 70], but whether TRPV4 mediates agonist-induced dilation in the human microcirculation remains understudied.

Nicotinamide adenine dinucleotide phosphate oxidase (NADPH oxidases, NOX) is a family of redox enzyme complexes producing the reactive oxygen species (ROS) such as the superoxide anion (O₂^·-) and hydrogen peroxide (H₂O₂) in a variety of cell types [13, 58, 75]. It is known that elevated ROS contributes to endothelial dysfunction in various diseases; however, ROS such as H₂O₂ can also act as physiological signaling molecules in the vascular wall. For example, NOX-derived H₂O₂ has been shown to regulate ACh-induced dilation in mouse mesenteric arteries [53] and rat intrarenal arteries [41]. Furthermore, Larsen et al. have reported that NOXs are expressed in human coronary endothelial cells and NOX-derived H₂O₂ is required for bradykinin (BK)-induced dilation in human arterioles [33]. However, the mechanisms by which NOX-derived ROS contribute to agonist-induced dilation are not fully understood. Our recent studies indicate that H₂O₂ induces robust protein phosphorylation of the TRPV4 channel to regulate its activity in EC [9]. The goal of this study was to determine the role of TRPV4, NOX and their potential interactions in the regulation of endothelium-dependent dilation in human arterioles. We examined the novel hypothesis that NOX-generated H₂O₂ stimulates phosphorylation and activation of endothelial TRPV4 to evoke Ca²⁺ entry and subsequent relaxation of the underlying smooth muscle.

Methods

Human tissue acquisition

Fresh human adipose arterioles (HAA) and human coronary arterioles (HCA) with internal diameters of 100–250 μm were isolated from adipose or heart tissues as discarded surgical specimens and unused donor hearts, as we described previously [45]. The study included patients without coronary artery disease (CAD) and with no or ≤ 2 risk factors for CAD, including hypertension, hyperlipidemia, and diabetes. All protocols were approved by the Institutional Review Board of the Medical College of Wisconsin and Froedtert Hospital on the use of human subjects in research. The tissue specimens were placed in ice-cold HEPES buffer (138 mM NaCl, 4 mM KCl, 2 mM MgSO₄, 1.6 mM CaCl₂, 1.2 mM KH₂PO₄, 6 mM glucose, and 10 mM HEPES, pH 7.4) and transferred to the lab for vessel isolation. Unless otherwise stated, vasoreactivity and molecular experiments were performed on endothelium-intact arterioles. For endothelium-denuded preparations, the endothelium was removed by slowly injecting 3–5 mL of air through the lumen.

A total of 71 patients who met the inclusion criteria were included in the study. Adipose samples were collected from 49 patients. Heart samples from 22 patients were used for this study. The demographic data of the patients were collected from the Generic Clinical Research Database at the Medical College of Wisconsin and are summarized in Table 1.

Table 1.

Patient demographics

	HAA	HCA
Sex
Male (n)	16	14
Female (n)	33	8
Age (yr) (mean ± SEM)	48 ± 14	58 ± 14
Body Mass Index (mean ± SEM)	28 ±5	26 ± 7
History/Risk Factors
Hypertension (n)	8	10
Hypolipidemia (n)	2	7
Smoking (n)	5	13
Atrial Fibrillation (n)	0	1
Tumor (n)	5	0

Vascular cell dissociation

EC and smooth muscle cells (SMC) were freshly dissociated from arterioles as previously described [38, 69]. In brief, arteriole segments were thoroughly rinsed of luminal blood, cut into small rings, and incubated at room temperature for 10 min in a low-Ca²⁺ dissociation buffer consisting of (in mM) 145 NaCl, 4.0 KCl, 0.1 CaCl₂, 1.0 MgSO₄, 10 glucose, and 10 HEPES with 0.1% BSA (pH 7.4). The buffer was carefully removed, followed by sequential incubation with papain (1.0 mg/ml) and DTT (0.5mg/ml) in dissociation buffer for 15 min, and then collagenase (Sigma blend H; 1.0 mg/ml), trypsin inhibitor (1 mg/ml), and elastase (0.5 mg/ml) for 15–30 min at 37 °C. All enzymes and chemicals were purchased from Sigma. Arteriole segments were gently triturated to release EC and SMC. The cells were washed once with the dissociation solution by centrifugation at 250× g for 5 min. EC were separated from the cell suspension by Dynabeads coated with the anti-CD31 antibody (Invitrogen) at 4°C for 1 hour, followed by the beads pull-down by the magnet as previously described [32]. Isolated EC and SMC were stored on ice or at 4 °C and used the same day for the RNA extraction and PCR. For the primary culture of isolated EC, cells were plated on glass-bottom Peri dishes and incubated for 2–3 days in the complete EGM-2MV growth medium (Promo cell) before fura-2 calcium imaging and PLA assay.

Cell culture and transfection

Human coronary artery endothelial cells (HCAEC) purchased from PromoCell (Heidelberg, Germany) were maintained in the complete EGM-2MV growth medium (Promo cell) and incubated at 37°C with 5% CO₂. Cells were split at 1:3 to 1:4 ratios when the cells were at 90~95% confluence. HCAEC between passages 6 and 7 were used for the experiments. HEK293 cells between passages 6 to 10, were provided by Dr. David Wilcox (Medical College of Wisconsin) and grown in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% FBS, 1% penicillin G, and streptomycin, and incubated at 37 °C with 5% CO₂. HEK293 cells between passages 12 and 15 were used for TRPV4 cell surface and phosphorylation experiments.

The TRPV4-GFP constructs with wildtype and point mutations of TRPV4 protein (S824A, S824E, S823A/S824A) in a pWPTS lentiviral vector were prepared as described previously [9], and the recombinant lentiviruses were produced from HEK293 cells. All constructs were verified by DNA sequencing. For TRPV4 overexpression experiments, HCAEC at passage 6 or 7 were transduced with lentiviruses at an m.o.i (multiplicity of infection) of 5–10. To minimize the potential calcium overload in the cells from the overexpressed TRPV4 channels, the culture medium was replaced with the low calcium culture medium after 16 hours of transduction, with the calcium concentration reduced to ~0.6 mM by the addition of 1.2 mM EDTA. Cells were used 3 to 4 days after transduction for calcium imaging assay, 4–5 days after transduction for immunoblotting experiments.

Vascular reactivity

Freshly isolated arterioles (≈100–200 μm) were cannulated and pressurized under 60 mm Hg in Krebs – physiological salt solution (PSS) that contains (in mM): 123 NaCl, 4.7 KCl, 1.2 MgSO₄, 2.5 CaCl₂, 19 NaHCO₃, 1.2 KH₂PO₄, 0.026 EDTA, and 11 glucose, gassed with 21% O₂ – 5% CO₂ at 37 °C. Changes in internal diameter were measured by a videomicroscopy system as we described previously [45]. Vessels were preconstricted with endothelin-1 (0.1–0.3 nM) by ~30% of the baseline internal diameter. Two dose-response curves were performed on each arteriole. For endothelium-dependent vasodilation, dose-dependent relaxation responses to the endothelium-dependent vasodilator acetylcholine (ACh, log 10⁻⁹ – 10⁻⁵ M), or TRPV4 agonist GSK1016790A (1–100 nM) were determined in the presence or absence of the NOX1/4 inhibitor GKT137831 (1 μM), or TRPV4 inhibitor HC067046 (2 μM) in the vessel bath. To further identify the endothelial-dependent role of NOX4, responses to sodium nitroprusside (log 10⁻¹⁰- 10⁻⁴ M), an endothelium-independent vasodilator, were performed with or without GKT137831 (1 μM) in the bath. At the end of each experiment, papaverine (100 μM), an endothelium-independent vasodilator, was added to the bath for maximal vasodilation. The vessel dilation response was calculated using the equations: percent maximal dilation = [((diameter with test compound) – (Basal diameter)) / ((Maximum diameter) – (Basal diameter))] ×100. The percentage of maximal dilation was averaged and plotted from at least 6 independent experiments. The efficacy and specificity of the above pharmacological agonists and inhibitors have been reported in previous studies [4, 16, 26, 61].

RNA extraction and reverse transcription-polymerase chain reaction

Freshly dissected human adipose arterioles (200–400 μm) were snap-frozen in liquid nitrogen and stored at −80 °C until use. Enzymatically dissociated EC and SMC were used the same day. Total RNA was extracted with the RNeasy Plus Universal Mini Kit (Qiagen), and the cDNA was synthesized with SuperScript III reverse transcriptase kit (Invitrogen), according to the manufacturer’s instructions. For regular PCR, 2 ng of cDNA was amplified using Platinum PCR Supermix (Invitrogen) in the thermal cycler with a 40-cycle touch-down protocol. The negative control group was amplified with primers without reverse transcription (RT- group), or without template (H₂O group). For quantitative PCR (qPCR), 20 μL reaction volume containing 2–5 ng of cDNA,10 μL SYBR Green qPCR Mix (Qiagen), 1 μL of reverse and forward primer each and nuclease-free water was performed using CFX96 C1000 Thermal Cycler. Cycle threshold (Ct) values were determined by CFX Manager 3.1 software. H₂O was used instead of the template for negative control. The thermal profile employed 1 cycle of initial denaturation at 95°C for 5 min, 40 cycles of denaturation at 95°C for 10 sec, annealing at 60°C for 30 sec, and extension at 72°C for 15 sec. Melting curve analysis was performed immediately after the 40th cycle at 72 to 95°C with 0.5 °C increments for 5 sec. NOX1 to 5 primers were synthesized by Integrated DNA Technologies Inc. and sequence information is provided in Table 2. Since all NOX primers are designed to span an exon-intron boundary, this avoids the amplification of the possible small amount of contaminating genomic DNA. 18S rRNA primers as an internal control were obtained from Qiagen (Cat No. QT00199367), and the primer sequences are not publicly available.

Table 2.

Primer sequences

Target gene	Orientation	Size (bp)	Primer Sequence
NOX1_HUMAN	Forward (22)	102	5’-GGT TTT ACC GCT CCC AGC AGA A-3’
	Reverse (22)		5’-CTT CCA TGC TGA AGC CAC GCT T-3’
NOX2_HUMAN	Forward (23)	132	5’-CTC TGA ACT TGG AGA CAG GCA AA-3’
	Reverse (22)		5’-CAC AGC GTG ATG ACA ACT CCA G-3’
NOX3_HUMAN	Forward (22)	160	5’-CCT GGA AAC ACG GAT GAG TGA G-3’
	Reverse (22)		5’-CCT CCC ATA GAA GGT CTT CTG C-3’
NOX4_HUMAN	Forward (23)	124	5’-CCA AGC AGG AGA ACC AGG AGA TT-3’
	Reverse (24)		5’-AGA AGT TGA GGG CAT TCA CCA GAT-3’
NOX5_HUMAN	Forward (22)	127	5’-CCA CCA TTG CTC GCT ATG AGT G-3’
	Reverse (22)		5’-GCC TTG AAG GAC TCA TAC AGC C-3’

RNA sequencing

Total RNA was extracted from arterioles with the RNeasy Plus Universal Mini Kit (Qiagen) and libraries prepared at the Genomic Science and Precision Medical Center (GSPMC at the Medical College of Wisconsin) according to the SMART-Seq Stranded Kit (Takara) utilizing 10ng of input. Briefly, RNA was reverse transcribed into cDNA using random priming and Illumina adapters and barcodes added by PCR. Degradation of ribosomal cDNA was completed with scZapR and scR-Probes as part of the kit protocol and final libraries amplified for sequencing. Final assessment, quantification, and pooling of the RNAseq libraries were completed with qPCR (Kapa Library Quantification Kit, Kapa Biosystems) and DNA high sensitivity fragment analysis (Agilent). The GSPMC completed next-generation sequencing on the Illumina NovaSeq600 with paired-end 100 base pair reads generated at >50 million reads per sample. Sequencing reads were aligned to the human Gencode v32 (based on Ensembl v98) and processed through the MAPR-Seq Workflow [29] with differential expression analysis completed using Bioconductor, edgeR v 3.8.6 software. Genes with a false discovery rate (FDR) less than 5% and an absolute fold change ≥ 4 will initially be filtered and considered significantly differentially expressed.

Immunoblotting

Fresh human arterioles were homogenized in ice-cold lysis buffer containing 25 mM Tris-HCL (pH 7.4), 150 mM NaCl, 0.1% SDS, and 1% NP-40, supplemented with protease and phosphatase inhibitors, and centrifuged at 12,000 g for 10 min at 4°C. Protein concentrations of the supernatants were measured by a BCA kit according to the manufacturer’s instructions. Samples were aliquoted and frozen in liquid N₂, stored at −80°C before use. Protein samples (20 μg) were separated by SDS-PAGE on 10% Bio-Rad TGX mini gels and transferred to polyvinylidene difluoride membranes. Membranes were blocked by 5% non-fat dry milk (NFDM) or bovine serum albumin (BSA) for 1 hour at room temperature and incubated overnight at 4 °C with a primary monoclonal antibody against NADPH oxidase 4 (kindly given by Dr. Doroshow at the NIH/NCI) (1:500 dilution in TBST) in TBST with 5% NFDM, or a monoclonal antibody against β-actin (1:40,000 dilution in TBST) in TBST with 2% NFDM. Blots were washed with TBST and then incubated with HRP-conjugated secondary antibody (1:20,000 dilution in TBST with 2% NFDM) for 1 hour at room temperature. Membranes were developed using ECL prime reagent and KONICA developer. The films were scanned using an EPSON scanner.

Biotinylation of cell surface proteins

Plasma membrane proteins from HCAEC were isolated by using a cell surface protein biotinylation kit (Pierce Co.). Cells were incubated with PBS containing 0.5 mg/ml sulfosuccinimidyl 2-(biotiamido) ethyl-1.3-dithiopropionate (Sulfo-NHS-SS-Biotin) for 30 min at 4 °C to label the surface protein. The biotinylation reaction was stopped by a quenching solution. After washing the cells with ice-cold Tris-buffered saline (TBS), a protein lysis buffer described above was added to extract total proteins. To isolate biotin-labeled surface proteins, total protein samples (150–200 μg) were gently mixed with 50 μl of NeutrAvidin agarose beads (supplied as 50% slurry) and rotated for 3 hours at 4 °C. The beads were collected after washing and centrifugation. Protein was then eluted from the beads use 100 μl of 1× Laemmli sample buffer with 50 mM DTT by heating the samples for 5 min at 95 °C. Western blot was followed to analyze the eluted protein.

Proximity ligation assay

The protein-protein interactions between TRPV4 and NOX4 in situ (distances < 40 nm) were detected by a proximity ligation assay (PLA) kit (Sigma-Aldrich #DUO92101). Briefly, primary EC isolated from HAA and HCAEC transduced with TRPV4-GFP were plated in 35-mm glass-bottom Petri dishes and grown to 60 – 70% confluence. Cells were fixed with 4% PFA in PBS for 20 min at room temperature (RT), followed by 0.1% Triton-X/0.3M glycine in PBS for 30 min at RT. Cells were blocked with block solution for 1 hour at RT, and two specific primary antibodies from different species that recognize and bind to TRPV4 (or TRPV4-GFP), and NOX4 proteins, including mouse anti-TRPV4 (Sigma MABS466) or mouse anti-GFP, and rabbit anti-NOX4 (47–6), were added and incubated overnight at 4°C. After 3 rinses in PBS, cells were incubated with anti-rabbit PLUS and anti-mouse MINUS PLA probes, which match the host of the primary antibodies, for 20 min at RT, followed by PCR amplification of fluorescent probes, according to the manufacturer’s instructions. Cells were then stained with DAPI and fluorescent images were captured using a fluorescence microscope (model TE200, Nikon) with a 20× objective. The cytoplasm size was estimated according to cellular fluorescence and DAPI-stained nuclei. PLA signals (red puncta) in the cell were counted and analyzed by Image J software.

Fura-2 calcium imaging

HECEC were plated and transduced with TRPV4-GFP wildtype or mutant in 35-mm glass-bottom Petri dishes and grown to 60 – 70% confluence. Cells were incubated with fura-2 AM (5 μM) (Molecular Probes) and 0.02% Pluronic F-127 at 37 °C for 30 min in the modified Hank’s balanced salt solution (HBSS) that contained (in mM): 123 NaCl, 5.4 KCl, 1.6 CaCl₂, 0.5 MgCl₂, 0.4 MgSO₄, 4.2 NaHCO₃, 0.3 Na₂HPO₄, 0.4 KH₂PO₄, 5.5 glucose, and 20 HEPES (pH 7.4 with NaOH). Fura-2 assay was used to monitor cytosolic Ca²⁺ signals as described previously [12, 36]. The emitted fura-2 fluorescence at 510 nm in cells that are alternately exposed to 340 and 380 nm excitation wavelength was recorded and analyzed by the MetaFluor software (Molecular Devices). Background fluorescence was subtracted before the study. The images of cell fluorescence were acquired every 3 seconds during the experiment. Changes in the intracellular Ca²⁺ concentration were presented as the ratio of the fluorescence intensity at 340 nm versus 380 nm excitation (F340/F380). The F340/380 ratio was averaged and plotted from at least 6 independent experiments. For the single experiment, 20 −40 cells were selected with the MetaFluor software based on the basal [Ca²⁺]_i. The cells with high basal [Ca²⁺]_i (F340/F380 ratio > 2.0) were excluded to minimize potential variations in response to TRPV4 agonists.

Data and statistical analysis

All data are presented as means ± SEM. Molecular experiments were repeated at least three times unless otherwise stated. For vascular reactivity, data from the independent experiments were averaged, and significant differences in vasodilatory responses were evaluated by two-way (factor) repeated measures ANOVA (factor 1, doses; factor 2, different treatments such as a vehicle or an inhibitor), followed by the Holm-Sidak test for pairwise comparisons or comparisons versus control. For other studies, statistical comparisons were made by Student’s t-test or one-way ANOVA. All statistical analysis was performed using SigmaPlot (version 12.0). Significance in figures was depicted as *P < 0.05; ** P < 0.01.

Reagents

GSK1016790A, GSK2795039, SNP, and Endothelin-1 (ET-1) were obtained from Sigma-Aldrich. GKT1372831 was obtained from MCE MedChemExpress. HC067047 was obtained from Tocris Bioscience. All other chemicals were purchased from Sigma. Stock solutions were prepared in DMSO (1000× or higher). ET-1 stock was prepared in saline solution supplemented with 1% BSA. Working stocks were freshly prepared in distilled water.

Results

NOX4 mRNA and protein expression in human adipose and coronary arterioles

To characterize the NOX isoforms expressed in human arterioles and isolated vascular cells, we first examined their mRNA expression. RT-PCR analysis revealed that mRNA transcripts of both NOX2 (amplicon size: 132 bp) and NOX4 (amplicon size: 124 bp) are most abundantly expressed in HAA and HCA, as well as in the freshly isolated endothelial cells (EC) and smooth muscle cells (SMC; Fig. 1a). The purity of isolated cells has been confirmed with the endothelial cell marker, platelet endothelial cell adhesion molecule-1 (PECAM-1), and smooth muscle cell marker, transgelin (SM22), as we reported recently [32]. Using quantitative PCR (qPCR), NOX4 mRNA expression tends to be lower in endothelium-denuded HAA compared with endothelium-intact HAA (Fig. 1b). Although both NOX2 and NOX4 have been detected in human arterioles by RT-PCR, NOX4 is thought to be constitutively active and produce H₂O₂ rather than O₂^·- [17, 58]; thus, we focused on the function of NOX4 in the subsequent studies. Consistent with RT-PCR results, NOX4 protein (~55 kD) was detected in both HAA (Fig. 1c) and HCA (Fig 1d). The NOX4 protein band is slightly smaller than the expected size of 67 kD for the canonical isoform 1 of NOX4 (Uniprot ID Q9NPH5–1), which is not further explored in the present study but may represent some splicing variants of NOX4, such as isoform 2 (58 kD; Uniprot ID Q9NPH5–2). As controls, NOX4 was detected in NOX4 overexpressing but not in non-transfected HEK293 cells, confirming the specificity of the anti-NOX4 antibody [37] (Fig. 1c). These results demonstrate that NOX4 is expressed at both mRNA and protein levels in the endothelial cells of human arterioles.

Fig 1. NOX4 mRNA and protein expression in human adipose and coronary arterioles and isolated endothelial and smooth muscle cells.

a) RT-PCR analysis shows that NOX4, in addition to NOX2, is expressed in HAA and HCA (n=3 patients for each), as well as in endothelial cells (EC) and smooth muscle cells (SMC) freshly isolated from HAA (n=3 patients). b) NOX4 mRNA was further analyzed with qPCR in both endothelium-intact and denuded HAA (n=3), 18s was used as housekeeping gene (n=3). c) Representative gel images of NOX4 protein expression in HAA from 2 patients, HEK293 non-transfected cells (293 NT) as a negative control, and NOX4-overexpressed HEK293 cells (293 NOX4 O/E) as a positive control. β-actin was included as an internal reference and loading control. d) Representative gel images of NOX4 protein expression in endothelium-intact and -denuded HCA from 2 patients. Compared with endothelium-intact HCA, NOX4 protein expression is reduced in endothelium-denuded HAA; β-actin was included as an internal reference and loading control.

TRPV4 and NOX4 in acetylcholine-induced dilation in human adipose arterioles

To understand the vasomotor function of TRPV4 and NOX4 in the human microvasculature, we examined their roles in acetylcholine (ACh)-induced vasodilation in freshly isolated human adipose arterioles (HAA). In endothelin-1-preconstricted adipose arterioles, acetylcholine (log 10⁻⁹ to 10⁻⁵ M), an endothelium-dependent vasodilator, induced potent dilation in a concentration-dependent manner (Fig. 2). This dilation was significantly attenuated by preincubation of arterioles with the TRPV4-selective inhibitor HC067047 (2 μM; Fig. 2a). ACh-induced dilation was similarly inhibited by the NOX1/4 inhibitor GKT137831 (1 μM; Fig. 2b). In the presence of HC067047, however, GKT137831 did not significantly reduce ACh-induced vasodilation (Fig. 2c). Together, these results indicate that both TRPV4 and NOX4 contribute to ACh-induced dilation in human arterioles, likely acting in series through a similar signaling pathway. To further confirm the role of endogenous H₂O₂ in ACh-induced dilation, adipose arterioles were pretreated with peg-catalase (500 U/ml), a cell-permeable H₂O₂ scavenger. Similar to NOX1/4 inhibitor GKT137831, peg-catalase markedly inhibited ACh-induced dilation at log −7 M of ACh (Fig S1a). In the presence of peg-catalase, NOX1/4 inhibitor GKT137831, as well as TRPV4 inhibitor HC067047, did not significantly further reduce the dilation (Fig S1b–c) as compared to peg-catalase pretreated arterioles. These results provide additional support that NOX4, H₂O₂, and TRPV4 likely act through a common signaling pathway in the regulation of human arteriolar relaxation.

Fig 2. Role of TRPV4 and NOX4 in acetylcholine (ACh)-induced dilation in human adipose arterioles (HAA).

In arterioles preconstricted with endothelin-1 (ET-1), the TRPV4-selective blocker HC067047 (2 μM) significantly reduced ACh-induced vasodilation at the concentrations of log 10⁻⁷ – 10⁻⁵ M. b) The NOX1/4 inhibitor GKT137831 (1 μM) reduced ACh-induced dilation at the concentrations of log 10⁻⁸ – 10⁻⁷ M. c) In the presence of HC067047, GKT137831 did not further reduce ACh -induced dilation. Data a to c, * P < 0.05 vs control or HC067047 (n=5 vessels for each group).

Consistent with previous studies [6], ACh-induced dilation in adipose arterioles was reduced but not abolished by NG-nitro-L-arginine methyl ester (L-NAME, 100 μM), an inhibitor of NO synthase, combined with indomethacin (Indo, 10 μM), an inhibitor of cyclooxygenase, confirming that NO and prostacyclin (PGI₂) partially mediate ACh-induced dilation (Fig. S1d). Interestingly, in the presence of L-NAME and indomethacin, GKT137831 further inhibited ACh-induced dilation (Fig. S1e). These results suggest that NOX4 is mainly involved in non-NO and non-PGI₂-mediated dilation.

NOX4 in TRPV4 agonist-induced dilation in human adipose and coronary arterioles

NOX4 has been recently shown to contribute to endothelial-dependent dilation in intrarenal arterioles [41]. To further investigate whether NOX4 participates in TRPV4-mediated vasodilation in human adipose and coronary arterioles, TRPV4 agonist GSK1016790A (GSK) was used to dilate the arterioles. As a synthetic agonist of TRPV4, GSK activates TRPV4 by directly binding to its transmembrane domains in a ligand-like manner [62], which in turn increases the intracellular Ca²⁺ concentration of endothelial cells, leading to endothelial-dependent vasodilation. As shown in Fig 3a–b, GSK (1 – 100 nM) dilated the arterioles in a dose-dependent manner. In HAA, after preincubation with NOX1/4 inhibitor GKT137831 (1 μM), the percentage of maximum dilation of GSK was significantly reduced (Fig 3a). In human coronary arterioles (HCA), GSK-induced vasodilation was also reduced after preincubation with GKT137831 (Fig. 3b). GKT137831 did not affect the response to the direct smooth muscle vasorelaxing agent, sodium nitroprusside (SNP), indicating an endothelium-dependent effect of GKT137831 (Fig 3c). In contrast to ACh, preincubation with L-NAME (100 μM) and indomethacin (10 μM) eliminated GSK-induced dilation, indicating that NO and PGI₂ are the main mediators for GSK-induced vasodilation (Fig S2a). The specificity of GSK1016790A was confirmed by TRPV4 selective antagonist HC0670147 (2 μM; Fig S2b). Together, these results indicate that NOX4 is involved in TRPV4-mediated dilation in human adipose and coronary arterioles.

Fig 3. Role of NOX4 in the TRPV4 agonist GSK1016790A-induced dilation in human adipose and coronary arterioles.

a) In HAA, preincubation with GKT137831 (1 μM) significantly reduced GSK1016790A-induced dilation at 10 to 100 nM. b) In human coronary arterioles (HCA), preincubation with GKT137831 also reduced GSK1016790A-induced dilation at 10 to 100 nM. c) For the endothelium-independent vasodilator sodium nitroprusside (SNP)-induced dilation, there is no significant difference between control and GKT137831-treated vessels. * P < 0.05 vs control (Data in a and b, n=6 vessels for each group; data in c, n=4 vessels).

NOX4 regulates GSK-induced Ca²⁺ influx in primary endothelial cells of human adipose arterioles

With the finding that GKT137831 largely inhibited TRPV4 agonist-induced dilation, it is reasonable to consider that NOX4 may act downstream of TRPV4 activation; however, NOX4 regulation of TRPV4 activation cannot be excluded and may also play an essential role. Since the vasomotor function of TRPV4 is mainly due to Ca²⁺ influx through the channel, we then examined whether NOX4 regulates TRPV4-mediated vasodilation by affecting TRPV4-mediated Ca²⁺ influx. Endothelial cells were freshly isolated from HAA and briefly cultured for 2–3 days before Ca²⁺ assay. The intracellular Ca²⁺ concentrations were monitored by the ratiometric calcium indicator fura-2 and the results are presented as F340/F380 ratios, with the corresponding fluorescence pseudo-colored from blue to red on a scale of 0.1 to 1. As shown in Fig 4a, the red fluorescence increased after additions of TRPV4 agonist 3 to 10 nM GSK, and with the addition of TRPV4-selective antagonist HC067047 (2 μM), the red fluorescence decreased rapidly. Representative traces in Fig 4b show that GSK induced Ca²⁺ increase in a concentration-dependent manner and this increase was largely reversed by HC067047. After preincubation with NOX1/4 inhibitor GKT137831 (1 μM), GSK-induced Ca²⁺ influx was reduced, and the remaining response was further inhibited by HC067047 (Fig 4c). In summarized data, Ca²⁺ influx induced by 3 and 10 nM GSK was significantly reduced by GKT137831 (Fig 4d), indicating that NOX4 regulates Ca²⁺ influx through the TRPV4 channel.

Fig 4. GSK1016790A-induced Ca²⁺ responses in the primary culture of endothelial cells isolated from human adipose arterioles.

a) Representative fura-2 fluorescence images in control cells (baseline), and after sequential additions of TRPV4 agonist GSK1016790A (GSK) 1 nM, GSK 3 nM, GSK 10 nM, and HC067047 (HC) 2 μM, respectively. The F340/F380 ratio is set on a scale of 0.1 to 1, with the corresponding color changes from blue to red. b-c) Representative traces. GSK1016790A induced Ca²⁺ increase in a concentration-dependent manner, and this increase was largely reversed by TRPV4-selective antagonist HC067047 (2 μM). After preincubation with NOX1/4 inhibitor GKT137831 (GKT137, 1 μM), GSK-induced Ca²⁺ response was reduced at GSK 3 and 10 nM, and the remaining response was further inhibited by HC067047. d) Summary data of concentration-dependent Ca²⁺ response induced by GSK1016790A in the absence and presence of GKT137831. GSK-induced response is reversible by HC under both conditions. * P < 0.05 vs control (n=6 independent experiments in each group).

NOX4 regulates GSK-induced Ca²⁺ influx and Ser824 phosphorylation in TRPV4-overexpressing human coronary artery endothelial cells

To further investigate the role and underlying mechanisms of NOX4 regulation, wild-type (WT) TRPV4- overexpressing human coronary artery endothelial cell (HCAEC) was used as in our previous studies [27, 73]. In control TRPV4-WT overexpressing HCAEC, GSK- induced Ca²⁺ influx in a concentration-dependent manner (Fig 5a). Consistent with the results obtained from the primary EC (Fig 4), preincubation with NOX1/4 inhibitor GKT137831 (1 μM) significantly inhibited GSK-induced Ca²⁺ response (Fig 5b–c). Given that TRPV4 as a channel protein requires expression on the plasma membrane, we confirmed the cell surface expression of transfected WT and the mutants by biotinylation cell surface assay (Supplemental Fig S3).

Fig 5. Effect of NOX1/4 inhibitor GKT137831 on GSK1016790A-induced Ca²⁺ influx in TRPV4 wild-type (WT)-overexpressing human coronary artery endothelial cells (HCAEC).

a) In control HCAEC overexpressed with TRPV4-WT, GSK1016790A induced Ca²⁺ influx in a concentration-dependent manner, the arrows indicate the administration of GSK1016790A (GSK). The intracellular Ca²⁺ concentrations were monitored by the ratiometric calcium indicator fura-2 and the results are presented as F340/F380 ratios. b) In WT TRPV4-overexpressing HCAEC, GSK1016790A-induced Ca²⁺ response at 1 nM and 3 nM was inhibited by the preincubation with GKT137831 (1 μM). c) Summary data of concentration-dependent Ca²⁺ responses induced by GSK1016790A in the absence and presence of GKT137831. * P < 0.05 vs control (n=6 independent experiments in each group).

Our previous work had shown that TRPV4 is regulated by Ser824 phosphorylation through the H₂O₂-PKA pathway [9]. NOX4 as an endogenous source of H₂O₂ may regulate the phosphorylation of the TRPV4 channel at Ser824. Therefore, the non-phosphorylatable TRPV4 mutant (S823A/S824A) was used to further investigate whether Ser824 phosphorylation is involved. As shown in Fig 6, GSK activated TRPV4 S823A/S824A mutant in control, but this activation was no longer affected by preincubation with NOX1/4 inhibitor GKT137831 (1 μM). To directly examine S824 phosphorylation, we performed immunoblotting using a phosphoserine motif antibody against the motif RXRXXS*/T* as previously reported [9]. As shown in Figs. 6d–e, GKT137831 significantly reduced TRPV4 phosphorylation at Ser824. The specific detection of Ser824 phosphorylation was verified by the non-phosphorylatable mutant-expressing cells in the absence and presence of PKC activator PMA (1 μM) stimulation (Supplemental Fig S4). The above results suggest that S824 phosphorylation is required for TRPV4 regulation by NOX4.

Fig 6. Role of Ser824 phosphorylation in the regulation of TRPV4-mediated Ca²⁺ influx by NOX4.

a) In control HCAEC overexpressing TRPV4-S823A/824A (3A/4A) mutant, GSK10167890A induced Ca²⁺ influx in a dose-dependent manner. b) After the preincubation with GKT137831, a similar Ca²⁺ response to GSK1016790A was observed in TRPV4-S823A/824A mutant-overexpressing HCAEC. c) Summary data of concentration-dependent Ca²⁺ responses induced by GSK1016790A. There is no significant difference in GSK1016790A responses between control and GKT137831-treated cells. d) Representative immunoblots of phospho-Ser824 (p-S824) of TRPV4 in the vehicle, NOX2 inhibitor GSK2795039 (GSK279, 1 μM)- and NOX1/4 inhibitor GKT137831 (GKT137, 1 μM)-pretreated TRPV4-WT overexpressing HCAEC. TRPV4 Ser824 phosphorylation was analyzed with a phosphoserine motif antibody against the motif RXRXXS*/T* (p-S824 antibodies), and the same blot was re-probed with GFP antibodies to detect total cellular TRPV4 proteins. E) Quantitative analysis of immunoblots. * P < 0.05 vs control (n=3 independent experiments in each group).

The protein-protein interactions of TRPV4 and NOX4

Based on the above findings that NOX4 and TRPV4 contribute to a similar pathway of the endothelial agonist ACh-induced dilation and NOX4 regulates TRPV4 channel phosphorylation and Ca²⁺ influx, we next examined whether these two proteins are colocalized in endothelial cells. Duolink proximity ligation assay (PLA assay) was performed using both primary EC isolated from HAA and TRPV4-GFP-overexpressing HCAEC.

Mouse- anti-TRPV4 (or mouse-anti-GFP) and rabbit-anti-NOX4 antibodies were used as primary antibodies to recognize and bind TRPV4 (or TRPV4-GFP) and NOX4 protein, respectively. After ligation and amplification, the PLA signal was detected in primary cultured HAA-EC (Fig 7a) and TRPV4-GFP- overexpressing HCAEC (Fig 7b) as red puncta using fluorescence microscopy, with the nuclei signal of DAPI shown in blue. No PLA signal was detected in control cells treated with no primary, but only secondary antibodies. The mean red puncta per cell were counted and analyzed. Compared with control, the PLA signal was significantly higher in cells incubated with TRPV4 and NOX4 antibodies, for both primary HAA-EC and TRPV4-GFP-overexpressing HCAEC, indicating that NOX4 and TRPV4 are closely localized in endothelial cells.

Fig 7. Detection of protein-protein interactions between TRPV4 and NOX4 in primary cultured EC isolated from HAA and TRPV4-GFP overexpressed HCAEC by proximity ligation assay (PLA).

PLA puncta are present in both a) primary cultured EC isolated from HAA and b) TRPV4-GFP overexpressed HCAEC incubated with anti-TRPV4 and anti-NOX4 antibodies, and with anti-GFP and anti-NOX4 antibodies, respectively. No PLA signal was detected in the control group without the addition of two primary antibodies. Anti-CD31 beads used to isolate primary HAA-EC show low-level red fluorescence. In panels a-b, images from left to right indicate PLA probe fluorescence in cells (red), cell nuclei stained by DAPI (blue), and overlays of two images. The mean PLA signal per cell was counted and quantified by Image J software, and summarized as bar graphs on the right of each panel. * P < 0.05 vs no antibody control (Neg Ctrl) (n=3 independent experiments in each group)

Discussion

To our knowledge, this represents the first study to characterize the role of TRPV4 and its interaction with NOX4 in endothelium-dependent vasodilation in human arterioles. The major findings of this study are as follows: first, both TRPV4 and NOX4 contribute to ACh-induced dilation in adipose arterioles from subjects without CAD or significant risk factors, and both act through a similar vasodilation pathway. Second, as one of the major NOX isoforms expressed in arteriolar EC, NOX4 colocalizes with TRPV4 in arteriolar EC and regulates TRPV4 agonist-induced dilation in adipose and coronary arterioles. Third, NOX4 increases Ser824 phosphorylation of TRPV4, which in turn facilitates Ca²⁺ influx through TRPV4 to regulate TRPV4-mediated vasodilation. Together, these results demonstrate a novel interaction between TRPV4 and NOX4 in regulating endothelium-dependent vasodilation in human arterioles. A schematic diagram of the proposed mechanisms is shown in Figure 8.

Fig 8. Schematic diagram illustrating the proposed mechanism by which NOX4 regulates TRPV4-mediated vasodilation in human arterioles.

In endothelial cells of human arterioles, NOX4 and TRPV4 proteins are functionally expressed and closely localized on or near the cell membrane. Through this close localization, NOX4 enhances TRPV4 agonist GSK1016790A (GSK)-induced channel activation by increasing protein phosphorylation of TRPV4 at Ser824, which in turn facilitates Ca²⁺ influx through TRPV4 channels to induce NO-mediated vasodilation. The protein-protein interaction of NOX and TRPV4 also contributes to the physiological agonist acetylcholine (ACh)-induced vasodilation in human arterioles, which involves NO and prostaglandin I₂ (PGI₂), as well as a non-NO, non-prostanoid component of dilation.

Role of TRPV4 and NOX4 in human arteriolar vasodilation

As one of the most extensively studied TRP channels, TRPV4 has been implicated in receptor agonist- and flow-induced endothelium-dependent vasodilation in various animal vascular beds [2, 31, 38, 46, 48, 52, 71]. Our previous studies further indicate that TRPV4 mediates flow-induced dilation in coronary arterioles from subjects with CAD [5]. However, TRPV4 does not seem to contribute to flow-mediated dilation in subjects without CAD (unpublished observations), raising an intriguing question about the role of TRPV4 under normal conditions. In this study, we used adipose arterioles freshly isolated from non-CAD subjects with no or ≤2 risk factors, and vasodilator responses to receptor agonist ACh were examined in cannulated preparations to mimic in vivo physiological conditions. After treatment with TRPV4-selective inhibitor HC067047, the maximum dilation to ACh was reduced, indicating that TRPV4 contributes to ACh-induced dilation in non-CAD arterioles. This proposed role is further supported by the findings that the direct TRPV4 agonist GSK1016790A induced significant Ca²⁺ entry in arteriolar EC and concentration-dependent dilation in adipose and coronary arterioles. These results are in line with those of earlier studies from our and other groups that TRPV4 contributes to ACh-induced dilation in normal mouse arteries [48, 71]. In contrast, Heathcote et al. have recently reported that, although TRPV4 channels modulate vascular tone by increasing endothelial Ca²⁺, ACh-induced vasodilation in rat mesenteric arteries was preserved after TRPV4 inhibition by HC067047 [23]. The reasons for this discrepancy remain unclear but underscore potential differential regulation of TRPV4 channels or compensation for loss of TRPV4 channels in different species and vascular beds. Similarly, whether these mechanisms contribute to the contrasting roles of TRPV4 in ACh- vs. flow-induced vasodilation of non-CAD arterioles remains to be determined.

The NOX family is an endogenous source of ROS such as H₂O₂ in the vasculature [34]. NOXs contain 7 isoforms, including NOX1–5, DUOX1, and DUOX2. Munoz et al. reported that NOX2 and NOX4, as endothelial sources of H₂O₂, contribute to endothelium-dependent vasodilation in intrarenal arteries [40, 41]. The studies from Zinkevich et al. and Larsen et al. have also shown that inhibiting NOX2 reduces flow-induced and bradykinin-induced vasodilation in HCA from subjects with CAD [33, 74]. In this study, we examined the novel role of NOX in non-CAD human arterioles. By characterizing the mRNA and protein expression of NOX isoforms in non-CAD arterioles and freshly isolated vascular cells, we found that NOX4, which constitutively produces H₂O₂ [47, 58], is abundantly expressed in adipose and coronary arterioles and arteriolar EC. Furthermore, inhibition of NOX4 reduced ACh-induced but not SNP-induced vasodilation in non-CAD arterioles, providing the first evidence that NOX4, like TRPV4, contributes to endothelium-dependent vasodilation under normal conditions. For ACh-induced dilation, while inhibition of TRPV4 and NOX4 each reduced the dilation, inhibition of both TRPV4 and NOX4 did not further reduce the response, indicating TRPV4 and NOX4 act through a similar vasodilatory pathway. Consistent with our previous results that H₂O₂ regulates TRPV4 function in cultured EC [9], we also found that NOX4 inhibition largely abolished TRPV4 agonist-induced dilation in HCA and HAA. Together, these results demonstrate a novel role of NOX4 and its interaction with TRPV4 in the regulation of endothelium-mediated dilation in non-CAD arterioles. Given the findings that NOX isoforms have similar expression in HAA and HCA, and NOX4 regulates TRPV4-mediated dilation in both vascular beds, HAA may serve as a surrogate of coronary function. This further supports the concept that the regulation mechanism in human arterioles is often conserved across the arteriolar beds [3].

Since NOX4 mainly produces H₂O₂, we hypothesize that the regulation of ACh- and TRPV4 agonist-induced dilation by NOX4 is at least partially mediated through NOX4-derived H₂O₂. Such NOX4-H₂O₂ pathway has been reported by others in that H₂O₂ production in NOX4 siRNA-treated cells was significantly reduced as compared with the control group [66, 67]. Besides, Munoz et al. had demonstrated that NOX4 is involved in endothelium-dependent relaxations and ROS production in renal arterioles [41]. In this study, similar to NOX4 inhibition, H₂O₂ scavenger peg-catalase inhibited the ACh-induced dilation in HAA, and in the presence of peg-catalase, ACh-induced dilation was not further reduced by NOX4 inhibitor GKT137831, as well as by TRPV4 inhibitor HC067047, as compared with peg-catalase alone (Supplemental Fig S1a–c), which further supported that a shared signaling pathway involved in the NOX4, H₂O₂, and TRPV4 in regulation of human arteriolar relaxation.

NOX1 has also been reported as an important source of O₂^·- in endothelial and smooth muscle cells of some vascular beds[48, 60]. Recent studies have shown that NOX1 deletion reduces the oxidant stress and restores microvascular health in a mouse model of diet-induced obesity [60]. In the present study, NOX1 mRNA is below the detection limit in non-CAD human arterioles, which is consistent with RNA sequencing results of the NOX family in human coronary arteries according to the GTEx database, as well as with our pilot sequencing in human adipose arterioles (Fig. S5). Thus, NOX1 may not play a significant role in the observed effects of NOX1/4 inhibitor GKT137831 on ACh- and TRPV4 agonist-induced dilation in human arterioles. However, it will be interesting to determine whether the expression of NOX1 is upregulated in human arterioles under pathological conditions such as CAD.

Since TRPV4 is a Ca²⁺-permeable cation channel, the mechanism of TRPV4 activation by receptor agonists such as ACh likely involves receptor-operated Ca²⁺ (ROC) entry following the binding of the agonist to the membrane receptor in EC [21, 28, 43, 44]. To investigate the endothelial factors released in responses to ACh and TRPV4 agonist GSK1016790A, arterioles were pretreated with NO synthase and cyclooxygenase inhibitors, L-NAME, and indomethacin, respectively. ACh-induced dilation was partially inhibited by L-NAME and indomethacin and the remaining dilation could be further inhibited by the NOX4 inhibitor GKT137831, indicating that NOX4 may contribute to vasodilation independent of NO and PGI₂. In contrast to ACh, TRPV4 agonist-induced dilation was largely abolished by L-NAME and indomethacin, as well as by GKT137831. This indicates that NOX4 can also contribute to NO- and PGI₂-mediated dilation depending on the endothelial stimuli employed. How this differential coupling to vasodilator factors occurs remains to be explored in future studies.

NOX4 regulation of TRPV4 function in EC

There is abundant evidence that NOX-derived H₂O₂ acts as a diffusible EDH factor to relax smooth muscle cells in human arterioles, especially in disease conditions such as CAD [33, 74]. However, in some vascular beds under normal conditions, exogenous application of H₂O₂ can also enhance EDH-type relaxation by elevating Ca²⁺ release from intracellular stores in EC [11, 14]. Given the finding that NOX4 inhibitors markedly reduced the TRPV4-mediated vasodilation in non-CAD arterioles, NOX4 may contribute to vasodilation by modulating TRPV4-mediated Ca²⁺ entry. To directly test this hypothesis, we first examined TRPV4-mediated Ca²⁺ responses in the primary culture of endothelial cells freshly isolated from human adipose arterioles, and then further verified the results in TRPV4-overexpressing HCAEC. In both endothelial systems, inhibition of NOX4 significantly reduced GSK1016790A (3 nM)-induced Ca²⁺ entry. Together with the findings of vasodilation studies, these results suggest that NOX4 can regulate endothelium-mediated dilation by enhancing TRPV4-mediated Ca²⁺ signaling.

Protein phosphorylation is one of the endogenous mechanisms that regulate TRPV4 channels. Fan et al. have reported that TRPV4 channel activation is enhanced by phosphorylation via different serine/threonine protein kinases [17]. Located at the C-terminal domain of TRPV4 channels, Ser824 residue is an important protein kinase A (PKA)-dependent phosphorylation site responsible for TRPV4 sensitization [17]. Our previous studies indicate that H₂O₂ at pathophysiological concentrations induces robust TRPV4 phosphorylation at Ser824 and enhances TRPV4-mediated Ca²⁺ entry, and both effects can be inhibited by PKA inhibitors. In addition, physiological lipid mediator arachidonic acid (AA)-induced TRPV4 activation is strongly regulated by Ser824 phosphorylation through the H₂O₂-PKA signaling pathway [9]. In this study, we examined Ca²⁺ response to the increasing dose of GSK1016790A in Ser823A/Ser824A double-mutant, non-phosphorylatable TRPV4-expressing HCAEC. Unlike the results seen in WT-TRPV4-expressing HCAEC discussed above, NOX4 inhibitors did not significantly affect Ca²⁺ responses to GSK1016790A in Ser823A/Ser824A-TRPV4-expressing HCAEC. Furthermore, in WT-TRPV4-overexpressing HCAEC, PKI, an inhibitor of PKA, shifted the Ca²⁺ dose-response curve induced by GSK1019790A (Supplemental Fig S6) to the right. Thus, PKA-mediated protein phosphorylation of TRPV4 at Ser823/Ser824 is likely required for the regulation of TRPV4-mediated Ca²⁺ entry by NOX4.

It is surprising that the TRPV4 channel response to synthetic agonists such as GSK101 was affected by protein phosphorylation. However, Cenac et al. have reported that PKC activators increased the phosphorylation of TRPV4 and its responsiveness to TRPV4 agonists 4αPDD [10]. Mohapatra et al. have also demonstrated that TRPV1 phosphorylation by PKA regulates the channel sensitivity, and consistent with our GSK101 data, PKA-mediated protein phosphorylation increased the sensitivity of TRPV1 channels to capsaicin but did not affect the maximal response to this agonist [39]. It is of note that TRPV4-mutant (S823A/S824A)-expressing HCAEC showed more Ca²⁺ influx than that in TRPV4-WT-overexpressing HCAECs (Fig 6), which may be due to mutation-related differences in the phosphorylation of other sites that regulate TRPV4 channel response to the agonists [17].

To further corroborate the results of Ca²⁺ experiments, Ser824 phosphorylation was directly measured by Western blotting with p-Ser824 antibodies [9]. Consistent with previous studies by Fan et al. [17], we found that TRPV4 channels exhibit basal S824 phosphorylation. Interestingly, as compared to the no-treatment control, pretreatment of NOX4 inhibitor GKT137831 significantly reduced the basal phosphorylation of TRPV4 at Ser824 in WT-TRPV4 expressing HCAEC. Since EC were not stimulated by PKA or PKC activators, these results indicate that TRPV4 channels exhibit a basal level of phosphorylation, and this basal phosphorylation is likely maintained by NOX4-dependent processes so that NOX4 inhibitors can down-regulate this basal S824 phosphorylation to regulate TRPV4 activity. This notion is supported by the results of the PLA assay indicating that TRPV4 and NOX4 may colocalize with the opportunity for functional interactions on or near the plasma membrane of EC. In the present study, a siRNA approach was also used to understand the role of NOX4 in S824 phosphorylation (Supplemental Fig S7). Overall, results obtained with NOX siRNA further support the role of NOX4 (and possibly NOX2) in the regulation of TRPV4 function, although some discrepancies in the relative effectiveness of NOX4/2 siRNA and inhibitors were also noted. The underlying causes of this discrepancy remain to be determined but may be related to an upregulation of other NOX or ROS enzymes by NOX4/2 siRNA [50] or a disruption of protein-protein interactions by siRNA [64]. For example, given that TRPV4 and NOX4 proteins are functionally closely localized, using NOX4 siRNA to knock down NOX4 protein expression may disrupt the proposed interaction between NOX4 and TRPV4, which in turn may reduce the inhibitory effects of NOX4 siRNA on S824 phosphorylation. In contrast, GKT137831 as a small-molecule inhibitor blocks the enzyme activity of NOX4, while the scaffold for TRPV4 and NOX4 interaction is likely preserved.

Study limitations

We could not control all patient characteristics since tissues used in the study were obtained as surgical discards. The demographic information, like medication history, presence of non-cardiovascular diseases, results of clinical tests of the patients, or other components of the medical history is limited. Our non-CAD patient cohort includes subjects without clinically diagnosed CAD and has no or ≤2 risk factors for CAD, and thus not all may be considered as true healthy subjects. For fura-2 Ca²⁺ assay, we used the primary culture of EC isolated from HAA as a surrogate for native EC. We have tried to measure TRPV4-mediated Ca²⁺ responses in EC in situ of freshly isolated human arterioles, but unlike mouse vessels that we have reported previously [71], responses of EC in situ of HCA and HAA were small and it is thus difficult to compare control vs treatments. Further study is needed to answer whether this difference in TRPV4-mediated Ca²⁺ responses is intrinsic to human arterioles or due to other technical variables.

In the present study, NOX1/4 inhibitor GKT137831 partially inhibited TRPV4-mediated Ca²⁺ response but largely blocked vasodilation, indicating that other underlying mechanisms could be involved. For example, besides TRPV4 phosphorylation examined in this study, there is evidence that during angiogenesis, NOX4 activates eNOS by eNOS phosphorylation [12]. In addition, NOX-derived H₂O₂ may act as a diffusible endothelial factor to relax smooth muscle, especially during CAD or other disease conditions [7, 33]. Since the same concentration of GKT137 (1 μM) was used in both the relaxation (HAA) and Ca²⁺ imaging (HAA-EC, HCAEC) studies, we expect that GKT137831 would be as effective, if not more effective, in a simpler EC system as compared to intact HAA. Although the NOX1/4 inhibitor GKT137831 (Setanaxib) has been tested extensively in the literature and has also entered into clinical trials [54], current vasorelaxation studies obtained with this inhibitor could be strengthened by a complementary siRNA approach.

Using PLA assay on both primary cultured EC and cells overexpressing TRPV4-GFP, our results indicate that TRPV4 is localized in close proximity to NOX4 in endothelial cells. However, further studies are required to determine how TRPV4 and NOX4 (or NOX2) proteins functionally interact with each other in endothelial cells. We have employed a co-immunoprecipitation technique to further test the protein-protein interaction between TRPV4 and NOX4 in HCAEC, HAA, and HCA, but the results were inconclusive (data not shown). It is possible that the expression level of TRPV4 or NOX4 proteins in the arterioles is relatively low and the interaction of these proteins may be transient or disrupted during sample preparation. Alternatively, NOX4 could be adjacent to but not necessarily physically interact with TRPV4 to regulate TRPV4 channel function in EC.

Perspective on clinical applications

There is increasing consensus that CMD plays a pivotal role in myocardial ischemia and predicts outcomes in patients with coronary heart disease, even in the absence of obstructive CAD. Coronary microvascular impairment as a role of adjunct cardioprotection target has been gradually explored in recent years [24, 72], however, due to the limited knowledge of the underlying mechanisms of microvascular dysfunction, the clinical efficacy of such therapy remains suboptimal. Although animal models of CMD recapitulate certain characteristics of microvascular dysfunction seen in patients, a perfect translation of experimental findings obtained from animal models to the clinical setting remains an important challenge in the field [56]. Therefore, a better understanding of the mechanisms of microvascular regulation in human arterioles may provide needed insight into the pathogenic processes in patients with CMD.

Currently, both TRPV4 and NOXs have been implicated in many physiological and pathological processes in cardiovascular and other systems. For example, studies have shown that TRPV4 channels play an important role in a broad range of diseases such as cardiac remodeling [1], systemic hypertension [48], and pulmonary hypertension [36, 68]. It has been recently proposed that increased vascular expression of NOXs may be involved in heart failure with preserved ejection fraction (HFpEF) [25]. Our previous studies have shown that TRPV4 channels contribute to flow-mediated dilation in CAD arterioles [73]. The present data revealed a novel role for TRPV4 channels and its interplay with NOX4 in the regulation of the receptor agonist ACh-induced dilation in non-CAD arterioles. Given that ACh-induced dilation in adipose arterioles may serve as a more clinically accessible surrogate for coronary endothelial function in patients with CMD [19], our new findings will guide future research to elucidate the precise role and mechanism of TRPV4 and NOXs, as well as their therapeutic potential, in CMD and other cardiovascular diseases.

Availability of data and material

All data generated or analyzed during this study are included in this published article.