This procedure was carried out as previously described (Mishra et al., 2015, 2018). Rats were injected intraperitoneally (IP) with sodium pentobarbital (50 mg/kg) to induce surgical anesthesia and overdose. The cremaster muscles were then surgically removed and transferred to a cooled (4°C) dissection bath containing Krebs’ buffer (115 NaCl, 5 mM KCl, 25 mM NaHCO₃, 1.2 mM MgCl₂, 2.0 mM CaCl₂, 1.2 mM KH₂PO₄ and 10 mM D-glucose), with the pH adjusted to 7.4. Cremaster 1A and 2A arteries (150–220 μm maximal intraluminal diameter) were carefully isolated and cannulated on glass pipettes that had been fitted into a pressure myography chamber (Living Systems, Burlington, VT, United States). To isolate middle cerebral arteries (∼150–250 μm maximal intraluminal diameter), rats were decapitated following anesthesia, and the brain was removed and placed in the cooled dissection chamber. Glass cannulae with tapered ends were pulled from borosilicate capillary tubes (0.7 mm ID and 1.2 mm OD) using a Sutter P-97 pipette puller; tips were then manually broken to generate tip openings in the range of 40–60 microns. The vessel lumen was filled with Krebs’ solution containing 1% bovine serum albumin (BSA), as described by Duling (Duling et al., 1981). The chamber was transferred to the stage of an inverted microscope and the bath was perfused with Krebs’ buffer at 7 mL/min using a peristaltic pump and suction line. Bath solution was gassed with 95% air/5% CO₂ and temperature was maintained at ∼34°C for cremaster arteries and ∼37°C for cerebral arteries. Intraluminal pressure was increased in a stepwise manner under no flow conditions from 0 mmHg and sustained at 70 mmHg using a pressure column, allowing 5-min of equilibration at each pressure. Arteries that failed to develop > 20% myogenic tone within 20–30 min were discarded. Intraluminal vessel diameter (i.e., inner surface of the arterial wall) was measured throughout the protocol at a sampling frequency of 1 Hz using an automated diameter tracking system (IonOptix, Milton, MA, United States).

For experiments involving intraluminal perfusion of middle cerebral arteries, a combined pressure servo-controller and servo-pump system was employed (Living Systems Instrumentation, Burlington, VT, United States). Intraluminal flow (∼25 μl/min) was initiated as arteries were being pressurized to develop myogenic tone, and was maintained throughout the remainder of the experimental protocol. This level of intraluminal flow generated an average shear stress of 48.3 ± 15.9 dyn/cm² (mean ± SD, n = 3) as calculated from the equation:

Shear stress (τ) = 4ηQ/πr³, where η = fluid viscosity (0.008 dyn s/cm²), Q = intraluminal flow in ml/s and r = arterial radius in cm (Kuo et al., 1995). The average vessel radius in myogenically constricted arteries in the presence of intraluminal flow was 43.2 ± 5.3 microns (mean ± SD, n = 3).

The tip diameters of inflow and outflow cannulae used for intraluminal perfusion were matched as closely as possible for a given artery to achieve uniform intraluminal flow along the vessel length. Cannulae tips typically ranged in size from ∼40–60 μm, and were selected based upon the diameter of the isolated artery at the time of mounting. As the tips of mounting cannulae were typically smaller than the lumens of pressurized vessels, this difference will contribute to flow resistance through the arterial lumen. 2-APT was perfused through the vessel lumen during the experiment by switching the inflow cannula line from the control buffer to one containing 10 μM 2-APT, followed by return to the control buffer for 2-APT washout.

Maximal inhibition of developed myogenic tone by 2-APT and VAS2870 in cremaster and middle cerebral arteries and associated IC₅₀ values (see Figures 2, 3) were calculated from non-linear regression fits of the data to the Hill equation as follows (Mishra et al., 2015):

Where Rmax = the calculated maximal degree of drug-evoked inhibition, IC₅₀ is the calculated drug concentration producing half-maximal inhibition of the maximal response, [drug] represents the concentrations of drug used experimentally, and N = the calculated Hill coefficient.

NADPH oxidase activity was measured in tissue homogenates by a luminescence assay using NADPH as the primary substrate and lucigenin as an acceptor for oxidase generated superoxide molecules, as modified from Griendling et al. (1994). Experimentally, freshly isolated tissue was homogenized (5–10% w/v) using a small probe polytron in ice-cold buffer containing (final): 50 mM Tris–HCl, pH 7.5, 150 mM NaCl, 1 mM EGTA, 1 mM EDTA, 1 mM benzamidine, 1 mM PMSF, and 5 μg/ml each of leupeptin, pepstatin A and aprotinin. The crude homogenate was filtered through two layers of cotton gauze and the protein concentration was determined using a modified Lowry assay kit (Bio-Rad Laboratories). Typically, 100 μg of homogenate protein was aliquoted per tube and each experimental condition was assayed in triplicate. Tissue homogenate + NADPH (1 mM final) was pre-incubated at 37°C for 2 min, and the reaction was started by addition of lucigenin (50 μM final); total assay volume was 0.1 ml with a final homogenate protein concentration of 1 mg/ml. For the background control condition, tissue homogenate was replaced by homogenization buffer alone. When present, individual NOX inhibitors were incubated with tissue homogenates on ice for ∼30 min prior to the start of the assay. Lucigenin oxidation leads to photon emission, which was measured for 0.4 s at 0.1 Hz over a 5 min period using a Monolight 3010 luminometer. Recently, Rezende et al. (2016, 2017) have raised concerns that this assay may also detect cytochrome P450 activity in cellular homogenates.

Following isolation and removal of surrounding fat and connective tissue, total RNA was extracted from cremaster arteries and the aorta using the RNeasy micro kit (Qiagen) according to the manufacturer’s instructions. Total RNA from liver tissue was extracted to act as a positive control, with PCR wells minus the addition of cDNA acting as the negative, non-template control. The concentration and quality of extracted RNA were determined using a Nanodrop instrument. Complementary DNA was prepared from total RNA using a Quantinova synthesis kit (Qiagen) and quantitative RT-PCR was performed using the Kapa SYBR Fast Universal quantitative PCR (qPCR) kit (Kapa Biosystems) and validated primer sets obtained from Integrated DNA Technologies (see Table 1). The molecular targets within the tissues were NAPDH oxidase isozymes 1–4, along with superoxide dismutase 1 (SOD1); rat glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as the internal reference gene. Control reactions and those containing cDNA from cremaster arteries and the aorta were performed with 1 ng of template per reaction organized in a PCR well plate. The heating protocol per cycle consisted of 95°C for 5 s, 55°C for 10 s and 72°C for 10 s using an Eppendorf Realplex 4 Master cycler and was extended to 40 cycles. PCR specificity was examined by dissociation curve analysis. Assay validation was confirmed by testing serial dilutions of pooled template cDNAs, indicating a linear dynamic range of 2.8–0.0028 ng of template, which yielded percent efficiencies ranging from 100–118%. Control samples lacking template yielded no detectable fluorescence. Note that NOX3 mRNA could not be detected reproducibly in vascular tissue following 40 cycles of amplification, which prevented reliable quantification. This finding is consistent with the reported minimal expression of NOX3 isoform in cardiovascular tissues (Bedard and Krause, 2007; Lassegue et al., 2012). Expression of target genes in tissue-derived RNA samples was determined using the relative expression software tool (REST) version 2.0.13 (Pfaffl et al., 2002). Our reporting of qPCR data follows the published MIQE guidelines (Bustin et al., 2009).

Results are presented as mean ± S.D; N = the number of animals used experimentally in a given protocol. Statistical analyses were carried out using either SigmaPlot 14 or GraphPad Prism 8 software. A Student’s t-test was utilized for analysis of one or two sample groups; a one-way ANOVA with a Newman–Keuls post hoc test was used for comparisons among 3 or more groups.

Isolated and cannulated cremaster skeletal muscle 1A and 2A resistance arteries from adult Sprague–Dawley rats developed robust myogenic tone when pressurized to 70 mmHg (i.e., % decrease in maximal intraluminal diameter = 51.6 ± 7.4%, mean ± SD, n = 18). As illustrated by the representative tracing in Figure 1A, bath addition of H₂O₂ to myogenically constricted cremaster arteries produced a robust and reversible vasodilation that developed at concentrations above 10 μM, as quantified in Figure 1B. These myogenically active arteries also exhibited reversible vasodilations to acetylcholine (ACh), the smooth muscle K-ATP channel opener pinacidil and sodium nitroprusside (SNP), demonstrating intact endothelial and smooth muscle vasorelaxant signaling pathways in these vessels. Treatment of vessels with a bath solution containing nominally free CaCl₂ + 2 mM EGTA at the end of the stimulation protocol was used to determine the maximal passive diameter at 70 mmHg intraluminal pressure.

The observed sensitivity of myogenically constricted cremaster arteries to H₂O₂ raised the possibility that H₂O₂ may serve as an endogenous signaling molecule for vasodilatory stimuli acting on the vascular endothelium and/or smooth muscle. As NADPH oxidase activity represents an important source of superoxide anions used for H₂O₂ production, we asked whether inhibition of vascular NOX enzymes by the reported pharmacological blockers 2-acetylphenothiazine (2-APT, aka ML171), VAS2870 and apocynin (Selemidis et al., 2008; Altenhöfer et al., 2015) would modify endothelium-dependent vasodilation in myogenically constricted cremaster arteries. We first investigated whether these inhibitors exerted potential direct effects on myogenic contraction in cremaster arteries, and utilized vasoactive agents acting via vascular endothelium (i.e., ACh) or smooth muscle (i.e., SNP, phenylephrine) to confirm the responsiveness of actively constricted vessels. Unexpectedly, bath addition of the selective NOX1 inhibitor 2-APT to cremaster arteries produced a robust, concentration-dependent vasodilation (i.e., inhibition of myogenic tone) that decayed in magnitude over the ∼8 min application period and readily reversed upon drug washout (Figure 2A). Exposure to 2-APT did not impair subsequent vasoactive responses to ACh, SNP or phenylephrine (PE). We further observed that addition of the selective NOX2 inhibitor VAS2870 to intact cremaster arteries also produced a potent, concentration-dependent inhibition of myogenic contraction (Figure 2B). However, the direct vasoactive effect of VAS2870 did not readily reverse upon washout, and we observed impaired contractile responses to both PE and 30 mM KCl in VAS2870 treated arteries (Figure 2B). The magnitude and potency of 2-APT and VAS2870-evoked inhibition of myogenic tone in intact cremaster arteries are quantified in Figures 2D,E, respectively. To examine the potential contribution of the endothelium to these actions of 2-APT and VAS2870, the endothelial layer was denuded in a subset of cremaster arteries by passing an air bubble through the lumen of cannulated vessels prior to pressurization. Denuded arteries still developed robust myogenic tone, exhibited normal vasoactive responses to the smooth muscle agents SNP and PE, but no longer dilated in response to ACh (Figure 2C). Endothelial denudation significantly blunted the magnitude and potency of 2-APT-evoked vasodilation, without affecting the recovery of myogenic contraction following 2-APT washout. These data demonstrate that the vascular endothelium plays a major role in the vasodilatory action of 2-APT in cremaster arteries, and that this agent can also directly affect vascular smooth muscle. Endothelial denudation did not diminish either the magnitude or potency of VAS2870-mediated inhibition of steady-state myogenic tone (Figure 2E), nor did it alter the irreversibility of this effect. This observation demonstrates that VAS2870 acts directly on smooth muscle in cremaster arteries.

To determine if the observed vasodilatory effects of 2-APT and VAS2870 are selective for skeletal muscle resistance arteries, we examined these agents in myogenically active, middle cerebral arteries from adult male Sprague–Dawley rats. Miller et al. (2005) have reported that middle cerebral arteries from male Sprague–Dawley rats generate 10-fold higher levels of superoxide anion compared with peripheral arteries, suggesting that inhibition of NOX activity in these vessels may have a more noticeable effect on vasoactive responsiveness. The representative tracing in Figure 3A shows that the active diameter of intact, myogenically active, middle cerebral arteries was 27.3 ± 4.8% (n = 4) less than their maximal passive diameter at 70 mmHg intraluminal pressure and exhibited normal responses to endothelium-dependent and –independent vasoactive agents, as we have previously reported (Mishra et al., 2015). Bath exposure to 2-APT produced a reversible, concentration-dependent inhibition of myogenic tone in intact middle cerebral arteries (Figure 3A) that was reduced in magnitude and potency, and was comparable to data observed in endothelium-denuded cremaster arteries (compare average data in Figures 2D, 3C). Interestingly, intraluminal application of 2-APT in intact cerebral arteries evoked a twofold greater vasodilatory response compared with bath application of 2-APT in the same vessels (Supplementary Figure 1).

In middle cerebral arteries denuded of their endothelium (confirmed by loss of dilatory response to SKA-31), the concentration-dependent vasodilatory response to 2-APT was qualitatively similar to that observed in intact arteries, although the magnitude of effect was slightly blunted (Figure 3B). In contrast to the observed effects of 2-APT, bath application of VAS2870 to intact middle cerebral arteries produced a robust and largely irreversible inhibition of myogenic tone that was similar in magnitude and potency to that observed in intact cremaster arteries (compare Figures 3A,D with Figures 2B,E). In endothelium-denuded cerebral arteries, the magnitude and potency of VAS2870 evoked vasodilation were similar to those observed in intact vessels (Figures 3B,D). Following washout of VAS2870 from the bath, we observed that the PE-induced vasoconstrictor response was severely blunted, similar to observations in cremaster arteries (Figures 2B,C). In myogenically contracted, cerebral arteries denuded of endothelium, active intraluminal diameter was 23.6 ± 3.9% (n = 5) less than their maximal passive diameter at 70 mmHg intraluminal pressure.

To confirm that 2-APT and VAS2870 act as bona fide inhibitors of NADPH oxidase activity in vascular tissue, we examined the ability of these agents to inhibit enzymatic activity in aortic tissue homogenates from adult male Sprague–Dawley rats. Using an optimized in vitro biochemical assay to measure de novo superoxide anion generation, we independently verified that these agents directly inhibit NADPH oxidase activity at concentrations comparable to those reported for inhibition of NOX isozymes in other cell types (see Table 2) (Selemidis et al., 2008; Altenhöfer et al., 2015). To complement these biochemical data, we carried out q-PCR analysis and identified mRNA encoding NOX 1, 2 and 4 in rat cremaster arteries and aorta (Figure 8). Robust expression of mRNA for SOD 1, an enzyme responsible for the conversion of superoxide to H₂O₂, was also detected in both cremaster arteries and aorta.

To interrogate the potential impact of NADPH oxidase inhibition on vasoactive responses to endothelium-dependent and –independent agents, we examined the vasomotor effects of ACh, SNP, PE and the endothelial KCa channel activator SKA-31 (Sankaranarayanan et al., 2009) in intact, myogenically contracted cremaster arteries. As shown in Figure 4A, treatment with the NOX1 selective inhibitor 2-APT (3 μM) did not alter vasoactive responsiveness to these exogenous stimuli, which is further confirmed by quantification of individual responses observed in the absence and presence of 2-APT (Figure 4B). Our biochemical data indicate that the selected concentration of 2-APT (i.e., 3 μM) inhibited total NADPH oxidase activity in aortic tissue homogenates by 86% (Table 2). Functionally, exposure to 3 μM 2-APT caused a transient inhibition of basal myogenic tone that reversed within 10 min (denoted in Figure 4A by the asterisk above the dilatory response at the start of 2-APT addition). This transient vasodilatory effect was commonly observed at 2-APT concentrations < 30 μM, as illustrated in Figure 2A. We were unable to examine the effects of VAS2870 treatment on endothelium-dependent and -independent vasoactive stimuli, given its robust inhibition of basal myogenic tone at concentrations that produce meaningful inhibition of vascular NADPH oxidase activity (see Figure 2B).

As treatment with 3 μM 2-APT may not effectively inhibit NOX isoforms 2–4, given the reported IC₅₀ values of 2-APT for these enzymes (Selemidis et al., 2008; Altenhöfer et al., 2015), we examined the effects of the non-selective NADPH oxidase inhibitor apocynin (10 and 100 μM) on vasoactive responses to ACh, SNP, PE and SKA-31 in myogenically active cremaster arteries. Parallel biochemical assays revealed that apocynin at these concentrations inhibited total NADPH oxidase activity in aortic tissue homogenates by 49% and 88%, respectively (Table 2). Apocynin is also reported to exhibit anti-oxidant behavior at similar concentrations (Heumuller et al., 2008). As shown by the representative tracings in Figures 5A,B, apocynin treatment alone evoked additional constriction in myogenically active vessels, leading to a greater reduction in intraluminal diameter that reversed upon washout. Interestingly, this observed increase in contractility was opposite to the effects of 2-APT and VAS2870 on basal myogenic tone (Figure 2A). Despite the increased contractile tone, treatment with apocynin at either 10 or 100 μM did not modify vasoactive responses to either endothelium-dependent or –independent agents, as quantified in Figure 5C. Collectively, the data presented in Figures 4, 5 indicate that treatment of myogenically active cremaster arteries with the NOX inhibitors 2-APT and apocynin did not modify acute vasoactive responses to endothelium-dependent and –independent agents, despite having pronounced inhibitory effects on vascular NADPH oxidase activity at the same experimental concentrations.

Several studies have reported that reactive oxygen species and/or H₂O₂ may contribute to vasodilatory signaling under conditions of endothelial dysfunction (i.e., reduced NO bioavailability) (Liu et al., 2003, 2011; Miura et al., 2003; Ellinsworth et al., 2015; Gutterman et al., 2016). Cremaster resistance arteries from spontaneous Type 2 Diabetic Goto-Kakizaki (T2D GK) rats exhibit endothelial dysfunction and impaired vasodilatory signaling that are consistent with decreased NO availability (Mishra et al., 2021). To determine the impact of NOX inhibition on vasoactive signaling under conditions of endothelial dysfunction, we examined vasodilatory and vasoconstrictor responses in cremaster resistance arteries from adult T2D GK rats. As shown in Figure 6A, treatment with 3 μM 2-APT caused a modest decrease of basal myogenic tone (control = 38.3 ± 4.2% vs. 32.1 ± 5.7% in the presence of 2-APT, n = 3), but did not alter the magnitude of vasoactive responses to stimuli acting via the endothelium or vascular smooth muscle (Figure 6C). In contrast, exposure of T2D GK cremaster arteries to 100 μM apocynin produced a substantial and reversible constriction of intraluminal diameter (control = 38.1 ± 3.9% vs. 54.6 ± 7.1% in the presence of apocynin, mean ± SD, n = 5, P = 0.002) (Figure 6B), but did not modify vasodilation in response to either endothelium-dependent or –independent agents (Figure 6D). However, the PE-induced contractile response was dampened in the presence of apocynin.

In view of the unexpected and robust vasodilatory action of 2-APT on myogenically constricted cremaster arteries (see Figure 2), we explored the potential underlying cellular mechanism by examining 2-APT effects in the absence and presence of established inhibitors of prominent endothelial-signaling pathways. Vasodilatory responses to ACh and SNP were utilized as internal controls to verify the effectiveness of individual treatment protocols on targeted cellular pathways (Figure 7D). As shown in Figure 7A, inhibition of NO synthase/cyclo-oxygenase signaling by L-NAME + indomethacin did not noticeably alter either the magnitude or sensitivity of 2-APT-evoked vasodilation in intact cremaster arteries from adult Sprague–Dawley rats. A similar result was observed for 2-APT action in the presence of UCL-1684 and TRAM-34, selective inhibitors of KCa2.3 and KCa3.1 channels that mediate endothelium-dependent hyperpolarization (Félétou, 2009; Köhler et al., 2016; Khaddaj Mallat et al., 2017; Figure 7B). However, exposure of cremaster arteries to the guanylyl cyclase inhibitor ODQ (10 μM) significantly reduced the magnitude of 2-APT evoked vasodilation (Figure 7C). Figure 7E quantifies the effects of all three treatment conditions on the extent and sensitivity of 2-APT induced vasodilation.

Although inhibition of endothelial and smooth muscle signaling pathways had either modest or no effects on the magnitude of 2-APT evoked vasodilation, it was evident from the original tracings that the vasodilatory responses to 2-APT decayed more rapidly under these treatment conditions. By calculating the fraction of peak 2-APT evoked dilation remaining following ∼8 min of 2-APT exposure, we found that all the treatments described above hastened the recovery of myogenic contraction to baseline levels in the continued presence of 2-APT, which was most prominent for 2-APT concentrations > 1 μM (Figure 7F). Exactly how these varied treatments promoted this decay or “desensitization” of the 2-APT mediated vasodilation remains unclear.