K_V7 channels contribute to paracrine, but not metabolic or ischemic, regulation of coronary vascular reactivity in swine

K_V7 channels are expressed in some coronary artery cell types and may participate in coronary dilation mediated by the endothelium (i.e., paracrine responses). However, K_V7 channels do not appear to function in coronary vasodilation in response to more relevant physiological and pathophysiological stimuli such as increased metabolic activity or ischemia.

Abstract

Hydrogen peroxide (H₂O₂) and voltage-dependent K⁺ (K_V) channels play key roles in regulating coronary blood flow in response to metabolic, ischemic, and paracrine stimuli. The K_V channels responsible have not been identified, but K_V7 channels are possible candidates. Existing data regarding K_V7 channel function in the coronary circulation (limited to ex vivo assessments) are mixed. Thus we examined the hypothesis that K_V7 channels are present in cells of the coronary vascular wall and regulate vasodilation in swine. We performed a variety of molecular, biochemical, and functional (in vivo and ex vivo) studies. Coronary arteries expressed KCNQ genes (quantitative PCR) and K_V7.4 protein (Western blot). Immunostaining demonstrated K_V7.4 expression in conduit and resistance vessels, perhaps most prominently in the endothelial and adventitial layers. Flupirtine, a K_V7 opener, relaxed coronary artery rings, and this was attenuated by linopirdine, a K_V7 blocker. Endothelial denudation inhibited the flupirtine-induced and linopirdine-sensitive relaxation of coronary artery rings. Moreover, linopirdine diminished bradykinin-induced endothelial-dependent relaxation of coronary artery rings. There was no effect of intracoronary flupirtine or linopirdine on coronary blood flow at the resting heart rate in vivo. Linopirdine had no effect on coronary vasodilation in vivo elicited by ischemia, H₂O₂, or tachycardia. However, bradykinin increased coronary blood flow in vivo, and this was attenuated by linopirdine. These data indicate that K_V7 channels are expressed in some coronary cell type(s) and influence endothelial function. Other physiological functions of coronary vascular K_V7 channels remain unclear, but they do appear to contribute to endothelium-dependent responses to paracrine stimuli.

NEW & NOTEWORTHY

K_V7 channels are expressed in some coronary artery cell types and may participate in coronary dilation mediated by the endothelium (i.e., paracrine responses). However, K_V7 channels do not appear to function in coronary vasodilation in response to more relevant physiological and pathophysiological stimuli such as increased metabolic activity or ischemia.

major questions remain incompletely answered in coronary vascular physiology. For example, what is(are) the principle mechanism(s) that couple coronary blood flow to metabolism and ischemia in vivo? It is clear that multiple pathways of regulation exist, including both open-loop and negative feedback controls (3, 14). Recent evidence suggests that myocardial production of hydrogen peroxide (H₂O₂) may be a signal responsible for the near lockstep increases of coronary blood flow that accompany intensified metabolic demand (21, 40, 47). H₂O₂ is generated in a variety of cellular processes and may be the primary transmitter of physiological redox signals (9, 45). For example, in coronary microvessels, H₂O₂ is an endogenous hyperpolarizing factor released by mechanical and paracrine stimulation of the endothelium (31, 46). The specific redox-sensitive targets mediating H₂O₂-induced vasodilation of coronary vascular smooth have not been firmly established. Our previous findings suggest that voltage-dependent K⁺ (K_V) channels may play an important role, but the identities of individual K_V channels involved remain elusive (4, 5, 12, 38, 39).

K_V7 channels are expressed in a variety of vascular smooth muscle cell types, including those from the coronary circulation (20, 23). Moreover, K_V7 channels are redox sensitive, because they are activated by H₂O₂ (13, 25). These qualities make K_V7 channels appropriate candidates to regulate coronary vascular resistance. Accordingly, it has been demonstrated ex vivo that K_V7 channels play a role in coronary vasodilation. Specifically, using Langendorff-perfused rat hearts, Khanamiri et al. (20) demonstrated that linopirdine reduces coronary flow and attenuates reactive hyperemia. Importantly, however, these data have not been supported by more physiologically relevant in vivo studies of coronary blood flow. Additionally, crystalloid-perfused hearts suffer from inadequate oxygen delivery (2, 41), which may influence results. Moreover, a more recent study in mice by Lee et al. (23) clearly indicated that the function and expression of K_V7 channels is much reduced in coronary arteries compared with cerebral arteries. Details regarding the presence and roles K_V7 channels in the coronary circulation are not clear and require further investigation.

Because pharmacological agents selective for K_V7 channels are available, it is possible to perform in vivo experiments to ascertain whether K_V7 channels are involved in coupling coronary blood flow to metabolism, participate in ischemic vasodilation, or mediate responses to paracrine stimulation of the endothelium. In the present study, we tested the hypothesis the hypothesis that K_V7 channels are present in cells of the coronary vascular wall and regulate vasodilation in swine. We used swine as experimental models and performed coronary blood flow measurements as well as isometric tension, patch-clamp electrophysiology, quantitative polymerase chain reaction (qPCR), Western blot, and immunohistochemistry (IHC) studies. We examined the role of K_V7 channels in 1) regulating coronary vascular resistance at the resting heart rate, 2) ischemic vasodilation (reactive hyperemia induced by a 15-s coronary occlusion), 3) coronary vasodilation to exogenous H₂O₂, 4) coronary metabolic vasodilation elicited by pacing-induced tachycardia, and 5) paracrine-mediated (bradykinin induced) endothelium-dependent coronary vasodilation. We also investigated the expression of KCNQ genes and K_V7 channels using molecular (qPCR), biochemical (IHC, Western blot), and functional (isometric tension, patch clamp) assays.

METHODS

All experiments involving animals were approved by an Institutional Animal Care and Use Committee and performed in accordance with the Guide for the Care and Use of Laboratory Animals (NIH Publication No. 85-23, Revised 2011). Lean adult male Ossabaw swine (n = 10) or domestic swine (n = 6) were sedated with telazol, xylazine, and ketamine (5, 2.5, and 2.5 mg/kg sc) before being anesthetized with morphine (3 mg/kg im) and α-chloralose (100 mg/kg iv). Following the completion of in vivo experimental protocols, hearts were fibrillated and excised for ex vivo studies.

qPCR analysis of KCNQ expression.

Epicardial coronary arteries (right coronary artery, RCA; left anterior descending coronary artery, LAD; and left circumflex coronary artery; LCx) were dissected from the heart. Total RNA was isolated using the RNeasy fibrous tissue kit (Qiagen) and stored at −80°C. cDNA synthesis was performed using 1 μg of total RNA and the iScript cDNA synthesis kit (Bio-Rad). Quantitative RT-PCR was performed using SYBR Green Supermix (Bio-Rad). The specificity of PCR amplification products was validated by melting curve analysis. Primers for KCNQ1, KCNQ2, KCNQ4, and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) for swine were designed using Premier Primer 5 software (PREMIER Biosoft). Primers for KCNQ3 and KCNQ5 were those described previously by Svalø et al. (44). Primer sequences and amplicon sizes are given in Table 1. KCNQ gene expression was normalized to a reference gene (GAPDH) and analyzed with Bio-Rad CFX Manager software.

Table 1.

Primers used for quantitative PCR

Gene	Channel	Primers (5′→3′)	Amplicon, bp	Accession No.
KCNQ1	KV7.1	F TCGCCACCTCAGCCATCA	219	AB033207.1
		R CCACTTGGCCGGACTCATT
KCNQ2	KV7.2	F GTGGTGTTTGGCGTGGAGT	140	AF110020.1
		R GCAATGGAGGCGATGAGC
KCNQ3	KV7.3	F AAGACGGGACCCTACTGCTG	127	XM_005662845.1
		R CGGTACTTGGCGTTGTTCCT
KCNQ4	KV7.4	F ACTCACGGTGGACGATGTTAT	100	XM_003128110.4
		R CAGCGTCTCCTTGAATTTCCTT
KCNQ5	KV7.5	F CCCTGTGGACAGCAAAGACC	126	XM_003121265.4
		R GTCTGGGCGCTGAACTCATT
GAPDH		F TGGGAAACTGTGGCGTGAT	123	AF017079.1
		R AAGGCCATGCCAGTGAGC

F, forward; R, reverse.

Western blot for K_V7.4.

Segments of the LCx artery were homogenized in lysis and extraction buffer containing (in mM) 50 β-glycerophosphate, 0.1 Na₃VO₄, 2 MgCl₂, 1 EGTA, 1 DTT, 0.02 pepstatin, 0.02 leupeptin, and 1 PMSF, as well as 0.5% Triton X-100 and 0.1 U/ml aprotinin. Lysate samples (300 μg of protein, measured by Bradford assay; Bio-Rad) were loaded on 4–12% Bis-Tris gels (Life Technologies) and subjected to electrophoresis. Proteins were transferred overnight at 4°C to nitrocellulose membranes, which were then blocked for 1 h at room temperature with 5% bovine serum albumin. Per the manufacturers' datasheets, immunoreactivity of K_V7.4 (Alomone; product no. APC-164) and GAPDH (a loading control; Abcam) appear at ∼110 and 37 kDa, respectively. Thus membranes were bisected between those molecular weights and each half processed separately. Membranes were incubated overnight at 4°C in buffer containing primary antibody (1:400 for anti-K_V7.4 and 1:1,000 for GAPDH). After washing was completed, membranes were placed for 1 h at room temperature in buffer containing horseradish peroxidase-linked secondary antibody (1:2,000 of goat anti-rabbit for K_V7.4 and 1:10,000 goat anti-mouse for GAPDH; both from Santa Cruz Biotechnology). K_V7.4 immunoreactivity was visualized with ECL Prime (Amersham), whereas GAPDH immunoreactivity was visualized with 3,3′-diaminobenzidine (DAB; Bio-Rad).

K_V7.4 IHC.

Methods to detect the tissue distribution of K_V7.4 included IHC with chromogenic and fluorescent indicators. Isolated coronary arteries and sections of myocardium were prepared for IHC studies using horseradish peroxidase detection. The myocardial sections were examined for K_V7.4 immunoreactivity in a small-caliber resistance vessel, whereas cardiomyocytes served as a K_V7.4-positive control. Formalin-fixed, paraffin-embedded sections of swine tissues were stained using an antibody against K_V7.4 (sc-50417; Santa Cruz Biotechnology) in conjunction with the Indiana University Immunohistochemistry Laboratory Core (Indiana University Health, Department of Pathology). Slides were imaged on an Aperio Scan Scope CS (Leica Biosystems) at ×40 magnification, and images were processed with associated ImageScope software.

To more closely examine the cellular localization of K_V7.4 in isolated epicardial coronary arteries, we used a fluorescently labeled secondary antibody. Segments of LCx arteries were fixed in paraformaldehyde and embedded in Richard-Allan Neg 50 frozen section medium (Thermo Scientific). Tissue was sectioned (8 μm) and placed on Superfrost Plus slides (Thermo Scientific). Samples were blocked with 10% donkey serum in PBS containing 0.1% Triton X-100 for 1 h at room temperature. Primary antibody (1:25; sc-50417; Santa Cruz Biotechnology) was added to sections in 2.5% donkey serum and allowed to incubate overnight at 4°C. Washed slides were incubated with secondary antibody (1:100 donkey anti-rabbit Alexa Fluor 594; Invitrogen) and Hoechst 33,342 stain (1:5,000; Invitrogen) for 1 h at room temperature. Slides were washed, Shandon Immu-Mount (Thermo Scientific) was applied, and coverslips were mounted. Red (K_V7.4 immunoreactivity) and blue (nuclear staining) fluorescence was visualized and photographed.

Isometric tension studies.

The aorta was cannulated to perfuse the coronary tree with cold (4°C) Ca²⁺-free Krebs solution (containing in mM: 131.5 NaCl, 5 KCl, 1.2 NaH₂PO₄, 1.2 MgCl₂, 25 NaHCO₃, and 10 glucose) to remove blood. LCx arteries were cut into ∼3-mm rings and mounted in water-jacketed organ baths (37°C) filled with Krebs solution containing 4 mM CaCl₂. Repeated contractions with 60 mM KCl at successively greater internal circumferences determined the optimal length of coronary artery rings as passive tension was increased in 1-g increments until K⁺-induced tension changed less than 10%. Arteries were constricted with 1 μM U46619 and relaxed with flupirtine or bradykinin in the presence or absence of linopirdine.

K_V7 channel modulators.

Flupirtine and linopirdine stocks were made with DMSO. The final concentration of vehicle did not exceed 0.1% in any experiment.

Patch-clamp electrophysiology.

Segments of LCx arteries were enzymatically digested to isolate coronary smooth muscle cells as previously described (12). Currents were measured at room temperature in conventional dialyzed patches. The bath solution contained (in mM) 125 NaCl, 5 KCl, 2 MnCl₂, 1 MgCl₂, 10 glucose, 10 HEPES, and 5 Tris; pH 7.4. The pipette solution contained (in mM) 140 KCl, 3 Mg-ATP, 1 Na-GTP, 1 EGTA; 10 HEPES, and 5 Tris; pH 7.1. After whole cell access was established, series resistance and membrane capacitance were compensated as fully as possible (Axon 200B amplifier and pClamp 9 software).

Coronary blood flow measurements.

Acute in vivo experiments were conducted in open-chest anesthetized pigs with the use of an extracorporeal pressure-clamped perfusion system. Anesthetized pigs were intubated and ventilated with O₂-supplemented room air. Catheters were placed into the thoracic aorta via the right femoral artery (for blood pressure and heart rate monitoring), into the left femoral artery (to draw a supply of blood for the extracorporeal perfusion system), and into the right femoral vein (for supplemental anesthetic, heparin, and sodium bicarbonate). Coronary perfusion pressure was maintained at 100 mmHg by a servo-controlled peristaltic pump throughout the study. Blood gas parameters were kept within normal physiological limits through periodic blood gas analyses and appropriate adjustments to breathing rate, tidal volume, and/or bicarbonate supplementation. The LCx artery was ligated proximally and cannulated with the pressure-controlled line of blood from the left femoral artery. An ∼30-min equilibration period elapsed before experiments commenced. Baseline parameters were obtained and experimental protocols performed as detailed below.

Five different kinds of experiments were performed to test whether K_V7 channels were involved in regulating various functions of the coronary circulation. In all five experiments, coronary blood flow was measured in real time in the cannulated, pressure-controlled coronary circulation. Before experiments were initiated, blood gas parameters were obtained to establish that the animal was under balanced anesthesia as well as to determine the hematocrit. Based on hematocrit and coronary blood flow, the plasma volume flow of the coronary perfusion territory was determined. Harvard Apparatus syringe pumps were used to infuse drugs into the coronary perfusion line, and dose rates of drugs were calculated for specific coronary plasma concentrations listed below. 1) We examined the capacity of K_V7 channels to change coronary vascular resistance at the resting heart rate by infusing K_V7 channel modulators directly into the LCx artery (both channel activators and inhibitors were used; maximum concentrations achieved in the coronary plasma for flupirtine and linopirdine were 30 and 100 μM, respectively). 2) We investigated the role of K_V7 channels in ischemic vasodilation by occluding the arterial cannula for 15 s and measuring reactive hyperemia before and during intracoronary infusion of 30 μM linopirdine. 3) We quantified the role of K_V7 channels in the coronary vasodilation to exogenous H₂O₂ by infusing 0.3–3 μM H₂O₂ in to the coronary perfusion line in the absence or presence of linopirdine (10 μM; infused before and simultaneously with H₂O₂). 4) We investigated the role of K_V7 channels in coronary metabolic vasodilation (using a subset of animals; n = 2) by pacing the heart to higher rates before and during intracoronary infusion of 10 μM linopirdine. In this experiment, arterial and coronary venous blood were sampled simultaneously at each heart rate to calculate myocardial oxygen consumption by the Fick principle (5). We assessed the role of K_V7 channels in endothelium-dependent vasodilation by performing intracoronary bradykinin concentration-response protocols before (control) and during simultaneous infusion of 30 μM linopirdine.

Statistical analyses.

A variety of statistical tests were used. In all tests, a value of P < 0.05 was considered significant. Normalized qPCR data and isometric tension data were analyzed by two-way analysis of variance (ANOVA). Post hoc analyses followed ANOVA when applicable, and specific differences are indicated in figures, legends, or text. The data regarding K_V7 channel modulators on resting coronary blood flow were analyzed by one-way repeated-measures ANOVA (RM-ANOVA), linear regression, and nonlinear regression (concentration-response fits). The data regarding the effect of linopirdine on H₂O₂- and bradykinin-induced coronary vasodilation were analyzed by two-way RM-ANOVA. The data regarding the effect of linopirdine on coronary metabolic vasodilation were analyzed by linear regression and analysis of covariance. The data regarding the effect of linopirdine on reactive hyperemia parameters were analyzed by paired t-test. The data regarding effects of K_V7 modulators on currents in coronary vascular smooth muscle cells (current-voltage and conductance-voltage relationships) were analyzed by two-way RM-ANOVA. The data obtained in Ossabaw and domestic swine were analyzed for variance between strains (both groups were lean and metabolically healthy). No differences between strains were found; therefore, the data were pooled.

RESULTS

KCNQ genes and K_V7.4 protein are expressed in swine coronary arteries.

K_V7 channels are encoded by KCNQ genes, five of which are known (KCNQ1–KCNQ5). We performed qPCR experiments to determine the expression of KCNQ genes relative to a reference gene, GAPDH (Fig. 1A). The segments of coronary artery that were used are expected to contain a number of cell types (e.g., smooth muscle cells, endothelial cells, and adipocytes, as well as sympathetic nerve endings). A survey of the literature (Table 2) suggests that KCNQ1, KCNQ4, and KCNQ5 are commonly expressed in arteries and veins, whereas KCNQ2 and KCNQ3 are observed less frequently. We detected expression of all five KCNQ genes in the LAD, LCx, and RCA (Fig. 1). Group data (n = 3) indicate that there were no differences between arteries (P = 0.86), but there were significant differences in the level of individual KCNQ transcripts (P < 0.0001). KCNQ4 and KCNQ5 were the most abundant transcripts (the higher expression of these 2 transcripts was statistically indistinguishable from each other; Fig. 1B). There were significantly lower levels of KCNQ1, KCNQ2, and KCNQ3 (the more limited expression of these 3 transcripts was statistically indistinguishable from each other; Fig. 1B). KCNQ4 was one of the most abundant transcripts, and a recent study implicated K_V7.4 channels in regulating coronary vascular responses (20); therefore, we performed Western blots to establish whether K_V7.4 protein is expressed in the LCx artery. A band of immunoreactivity consistent with K_V7.4 was detected (Fig. 1C).

Fig. 1.

Expression of KNCQ genes and K_V7.4 protein in coronary arteries. A: representative quantitative PCR (qPCR) amplification plot. B: group data (n = 3) for KNCQ gene expression normalized to GAPDH. C: representative immunoblot for K_V7.4 and GAPDH in the swine left circumflex coronary artery (LCx). LAD, left anterior descending coronary artery; RCA, right coronary artery.

Table 2.

Expression of KCNQ mRNA and K_V7 channel protein in vascular tissues

Tissue	KCNQ1 KV7.1	KCNQ2 KV7.2	KCNQ3 KV7.3	KCNQ4 KV7.4	KCNQ5 KV7.5	Ref.
Mouse portal vein	+					35
Mouse aorta, carotid, femoral, mesenteric arteries	+			+	+	50
Mouse portal vein	+			+	+	48
Rat pulmonary artery	+			+	+	18
Rat cerebral arteries	+			+	+	52
Arteries from human adipose and mesentery	+		+	+	+	33
Rat basilar arteries	+	+	+	+	+	28
Rat coronary artery	+		+	+	+	20
Rat gracilis artery	+		+	+	+	51
Pig coronary artery	+			+	+	16
Mouse mesenteric artery				+	+	8
Rat pulmonary artery					+	24
Mouse mesenteric artery	+		+	+	+	42
Rat aorta and vena cava	+				+	36

Immunohistochemical detection of K_V7.4.

We performed IHC studies on sections of myocardium and isolated epicardial coronary arteries. The myocardium was strongly positive for K_V7.4 immunoreactivity (example marked by a star in Fig. 2A; note cytosolic and membranous brown staining of cardiomyocytes). Moreover, coronary microvessels in those myocardial sections were positive for K_V7.4 immunoreactivity (multiple examples marked by arrows in Fig. 2A). Conduit coronaries were also positive for K_V7.4 immunoreactivity (brown color; Fig. 2B). To more closely examine the cellular distribution of K_V7.4 immunoreactivity in the LCx artery, a fluorescent secondary antibody was used (Fig. 2C). The distribution of K_V7.4 reactivity was not uniform across the vascular wall. Staining for K_V7.4 was most prominent in the intimal layer (endothelial) and adventitial layer (containing loose connective tissue, adipocytes, sympathetic nerve endings, etc.). There was less K_V7.4 staining in the medial (smooth muscle) layer. Figure 2C, inset, shows the red (K_V7.4) channel only, and an arrow points to the thin line of immunostaining on the intimal surface.

Fig. 2.

Immunostaining for K_V7 channels. A: section of swine myocardium stained for K_V7.4. Brown color indicates positive immunoreactivity. Cardiomyocytes are strongly positive for K_V7.4 immunostaining (example highlighted by star). Coronary microvessels are also K_V7.4 positive (arrows point to several examples). B: section of swine LCx artery stained for K_V7.4. There is diffuse brown staining in the sample. C: overlay of light and fluorescent microscope images of an LCx artery section. K_V7.4 immunofluorescence is red (see inset for red-only channel; arrow points to endothelium) and nuclei are blue (4,6-diamidino-2-phenylindole).

Flupirtine-induced relaxations are largely endothelium dependent.

There is no single drug that opens all five known K_V7 channels, but flupirtine is an agonist of K_V7.2, K_V7.3, K_V7.4, and K_V7.5 (27). Thus flupirtine was chosen to determine the effects of a K_V7 channel opener on coronary vascular tone. The LCx artery was utilized for isometric tension experiments with flupirtine (3–300 μM). All artery rings were tested for functional endothelium before addition of any K_V7 channel modulators. Artery segments were preconstricted with the stable thromboxane A₂ mimetic U46619 (1 μM), and relaxation to 1 μM bradykinin was 94 ± 2, 95 ± 3, and 91 ± 2% in the rings serving as the control, linopirdine-treated, and DMSO-treated groups, respectively. Artery segments were again preconstricted with U46619, and increasing concentrations of flupirtine were added. Flupirtine relaxed coronary artery rings in the control group in a concentration-dependent manner (Fig. 3A; n = 7). To determine whether K_V7 channels mediated the effects of flupirtine, two separate groups of rings were pretreated with the pan-K_V7 antagonist linopirdine (10 or 30 μM; n = 7 per group) and subjected to identical flupirtine concentration-response protocols. Flupirtine-induced relaxation was significantly attenuated in the presence of linopirdine (Fig. 3A). Linopirdine pretreatment had no effect on baseline tension (i.e., the K_V7 channel blocker did not elicit a contraction by itself) or the tension developed in response to U46619. Time control responses to vehicle (0.1% DMSO) demonstrated no significant change in active tension (11 ± 6% relaxation, n = 3).

Fig. 3.

Flupirtine relaxes coronary artery rings ex vivo. A: group data (n = 7) demonstrating the ability of flupirtine to relax LCx artery segments contracted with U46619. Pretreating 2 additional sets of rings with either 10 or 30 μM linopirdine inhibited flupirtine-induced relaxation (n = 7 each). *P < 0.05 indicates a difference between the control and linopirdine-treated groups. #P < 0.05 indicates a difference between 10 and 30 μM linopirdine. B: relaxation elicited by 300 μM flupirtine in rings denuded of endothelium in the presence or absence of 30 μM linopirdine (n = 3 each).

To determine whether the endothelium played a role in flupirtine-induced coronary artery relaxation, a separate set of rings (n = 3) was mechanically denuded of endothelium, preconstricted with U46619, and challenged with 300 μM flupirtine (Fig. 3B). Endothelial denudation significantly reduced flupirtine-induced relaxation of coronary artery rings (49 ± 8% in Fig. 3B vs. 91 ± 5% in Fig. 3A; P = 0.03). Whereas linopirdine inhibited the flupirtine-induced relaxation of intact rings (Fig. 3A), 30 μM linopirdine had no effect on the magnitude of flupirtine-induced relaxation in rings denuded of endothelium (Fig. 3B; 58 ± 2%; P = 0.34).

Linopirdine inhibits bradykinin-induced relaxation of coronary artery rings.

Having established a role for a functional endothelium in mediating coronary artery relaxation to the K_V7 activator flupirtine, we assessed the role of K_V7 channels in endothelium-dependent relaxation to bradykinin. Isolated coronary artery rings were preconstricted with 1 μM U46619 and exposed to increasing doses of bradykinin (1 to 30 μM) under control conditions or in the presence of 30 μM linopirdine (n = 3). Bradykinin relaxed coronary artery rings in a concentration-dependent manner (Fig. 4). Linopirdine diminished relaxation across the range of bradykinin concentrations tested, resulting in a right shift in the concentration-response curve (P = 0.02; Fig. 4). To establish whether this linopirdine effect resulted from alterations in the endothelial response to bradykinin or the vascular smooth muscle response to relaxing factors, rings were exposed to increasing doses of sodium nitroprusside (1 nM to 30 μM). The NO donor relaxed rings in a concentration-dependent manner, and linopirdine had no statistically significant inhibitory effect (95 ± 3% vs. 85 ± 8%; P = 0.44).

Fig. 4.

Linopirdine attenuates bradykinin-induced relaxation of coronary artery rings. The graph shows group data illustrating the effect of 30 μM linopirdine on bradykinin-induced relaxation of coronary artery rings (n = 3). Two-way RM ANOVA indicates a significant inhibitory effect of linopirdine (P < 0.001, but post hoc analysis failed to indicate where specific differences exist). Linopirdine appears to shift the bradykinin concentration-response curve to the right.

Effects of linopirdine on K⁺ current in coronary artery smooth muscle.

Linopirdine inhibits K⁺ current in smooth muscle cells from the LCx artery (Fig. 5A). Specifically, linopirdine reduced current 28–31% at potentials between −10 and +40 mV (Fig. 5B; n = 5). Moreover, linopirdine decreased tail currents 29–50% following conditioning steps between −30 and +40 mV (Fig. 5C), but linopirdine did not shift the half-activation potential (V_1/2: −14 ± 1 and −15 ± 1 mV before and after linopirdine; Fig. 5D).

Fig. 5.

Linopirdine and K_V current. A: current traces from a representative cell (top); voltage template is shown at bottom. Current was inhibited by 10 μM linopirdine. B: plot of group data (n = 5 cells) for the current-voltage relationship before and after addition of linopirdine (data taken from period indicated by horizontal bar in A, left). C: group data conductance-voltage curve (i.e., analysis of tail currents in same cells described for B; data come from area indicated by arrow in A, left). Maximal conductance was reduced by linopirdine. D: normalized conductance (G/G_max)-voltage curve. Linopirdine did not shift the half-activation voltage (V_1/2) of the tail currents (i.e., the voltage sensitivity of the channels was unaffected). *P < 0.05, between control and linopirdine.

It is possible that the contribution of K_V7 channels to smooth muscle K⁺ current may best appreciated at the end of relatively long (4–5 s) voltage steps. This is because the presence of auxiliary/modulatory subunits (such as KCNE) could slow the activation kinetics of K_V7 channels dramatically. Thus we more thoroughly examined the effect of linopirdine on coronary smooth muscle K⁺ current with long voltage steps (Fig. 6). Additionally, in these experiments, the concentration of the K_V7 channel antagonist was doubled to increase efficacy. When the smooth muscle membrane was voltage-clamped to test potentials for nearly 5 s, a slow inactivation of current was seen at the more depolarized potentials (Fig. 6A). Linopirdine (20 μM) clearly reduced current at the earlier time points (Fig. 6B) as expected from the previous experiments (Fig. 5). Interestingly, however, the noninactivating current (or the current at the steady-state that might be attributed to K_V7 channels) was not blocked by 20 μM linopirdine (Fig. 6D). Current in the presence of 20 μM linopirdine was 88–109% of control at voltages between −10 and +40 mV (there was no statistically significant effect).

Fig. 6.

Linopirdine and K_V current: long voltage steps to assess steady-state current. A and B: current traces from the same cell before (A) and after (B) addition of 20 μM linopirdine. Early (peak) current was inhibited by linopirdine, similar to the effect of 10 μM linopirdine in Fig. 5. C: voltage template. D: group data (n = 5 cells) for the current-voltage relationship (data taken from period indicated by horizontal bar in A). There was no statistically significant effect of linopirdine on steady-state current.

K_V7 channel modulators and coronary blood flow at the resting heart rate.

Our first in vivo experiment was designed to determine whether K_V7 channels are active in the coronary circulation and regulate blood flow at the resting heart rate. There are two logical predictions to make with regard to K_V7 channel modulators. First, the addition of a K_V7 channel opener might increase coronary blood flow. This would be similar to what we have shown previously for another K⁺ channel opener, pinacidil (12), and would agree conceptually with the flupirtine-induced relaxation of coronary artery rings ex vivo (isometric tension data presented in Fig. 3A). Second, administration of a K_V7 channel blocker might decrease coronary blood flow at rest. This would be similar to what we have shown previously with a different K⁺ channel blocker, 4-aminopyridine (12). Pigs were instrumented to measure blood flow in myocardium perfused by the LCx artery while it was pressure clamped at 100 mmHg.

Flupirtine is a K_V7 channel agonist that can significantly relax coronary artery rings ex vivo (Fig. 3A). Flupirtine was infused directly into the coronary perfusion line at dose rates calculated to give specific coronary plasma concentrations between 1 and 30 μM to determine whether opening K_V7 channels could regulate baseline coronary vascular resistance. Flupirtine had no effect on coronary blood flow at the resting heart rate over the tested range of concentrations (Fig. 7A). The data were analyzed three ways, all of which returned the same negative results with regard to flupirtine. Specifically, 1) one-way RM-ANOVA did not indicate any differences in absolute or delta blood flow at the various doses; 2) subjecting the data to a linear regression gave a slope that was not significantly different from zero (linear regression analysis produces a line with a slope of 0.009 ± 0.007; not different from zero); and 3) no meaningful concentration-dependent curve (nonlinear regression) could be applied.

Fig. 7.

K_V7 modulators and coronary blood flow. A: plot of change (Δ) in coronary blood flow vs. intracoronary plasma concentration of flupirtine. Data are from 10 pigs. There is no significant relationship between Δcoronary blood flow and the concentration of K_V7 channel agonist. That is, flupirtine has no effect on coronary blood flow at the resting heart rate (there is no flupirtine-induced vasodilation). B: plot of Δcoronary blood flow vs. intracoronary plasma concentration of linopirdine. Data are from 13 pigs. There is no significant relationship between Δcoronary blood flow and the concentration of the K_V7 channel inhibitor. That is, linopirdine has no effect on coronary blood flow at the resting heart rate (there is no linopirdine-induced vasoconstriction).

Linopirdine is a K_V7 channel antagonist that antagonized, at least partly, flupirtine-induced relaxation of LCx artery rings ex vivo (Fig. 3A). Linopirdine was infused directly into the coronary perfusion line to give specific coronary plasma concentrations between 1 and 100 μM to determine whether inhibiting K_V7 channels could influence baseline coronary vascular resistance. Linopirdine had no effect on coronary blood flow at the resting heart rate (Fig. 7B). There was no discernable effect of linopirdine in vivo when evaluated by 1) RM-ANOVA on absolute or delta flow values, 2) linear regression (slope of 0.014 ± 0.006; not different from zero), or 3) nonlinear regression (unable to fit with a dose-response curve).

To address whether the apparent lack of responses to K_V7 channel modulators on baseline coronary vascular resistance were due to the choice of pharmacological agents, a subset of animals was subjected to similar intracoronary concentration-response experiments with additional, structurally unrelated K_V7 channel drugs. Two more K_V7 channel activators, L364-373 (0.03–30 μM) and retigabine (1–30 μM), were assessed in an additional two pigs. Neither of these K_V7 channel openers increased coronary blood flow. The average change in flow with 0.03, 0.1, 0.3, 1, 3, 10, and 30 μM L364-373 was −5, 3, −3, 4, −6, −1, and 6%, respectively. The average change in flow with 1, 3, 10, and 30 μM retigabine was 1, 0, 0, and 1%, respectively. Chromanol 293B is a K_V7 channel blocker, but intracoronary infusion of this drug (0.1–30 μM) had no effect on coronary blood flow at the resting heart rate (data from 1 pig). The change in flow with 0.1, 0.3, 1, 3, 10, and 30 μM chromanol 293B was −1, 0, −1, 1, 2, and −3%, respectively. Thus, regardless of the K_V7 channel modulator employed in vivo, there was no appreciable change in coronary blood flow at the resting heart rate.

Additional cardiovascular parameters were recorded during the intracoronary infusion of K_V7 channel modulators and are presented in Tables 3 and 4. We did not expect major changes in systemic hemodynamics because the drugs were infused directly into the coronary circulation. Neither flupirtine (Table 3) nor linopirdine (Table 4) significantly changed any parameter we measured. The same is true for L364-373, retigabine, and chromanol 293B (data not shown). Heart rate and mean arterial pressure remained stable.

Table 3.

Cardiac and hemodynamic parameters with intracoronary flupirtine infusion

		Flupirtine, μM
	Baseline	1	3	10	30
Systolic pressure, mmHg	111 ± 6	103 ± 5	103 ± 9	108 ± 6	106 ± 9
Diastolic pressure, mmHg	76 ± 4	71 ± 4	70 ± 6	73 ± 5	71 ± 6
Mean pressure, mmHg	92 ± 5	86 ± 4	86 ± 7	89 ± 5	88 ± 7
Heart rate, beats/min	67 ± 8	72 ± 8	79 ± 12	71 ± 7	79 ± 12

Table 4.

Cardiac and hemodynamic parameters with intracoronary infusion of linopirdine

		Linopirdine, μM
	Baseline	1	3	10	30	100
Systolic pressure, mmHg	104 ± 5	105 ± 5	103 ± 5	103 ± 5	103 ± 5	94 ± 8
Diastolic pressure, mmHg	72 ± 4	72 ± 4	71 ± 4	71 ± 4	71 ± 4	65 ± 6
Mean pressure, mmHg	88 ± 4	88 ± 4	86 ± 4	86 ± 4	86 ± 4	79 ± 7
Heart rate, beats/min	74 ± 7	73 ± 8	76 ± 9	76 ± 9	75 ± 8	84 ± 11

Linopirdine and coronary reactive hyperemia.

Our second in vivo experiment was designed to determine whether K_V7 channels play a role in coronary reactive hyperemia (i.e., vasodilation in response to an ischemic stimulus). If K_V7 channels are responsive to vasodilators produced by the ischemic myocardium, then linopirdine should reduce the peak hyperemic response and/or the repayment of flow debt. This would be similar to what we have shown previously for the K⁺ channel blocker 4-aminopyridine (12). To elicit coronary reactive hyperemia, the arterial perfusion cannula was occluded for 15 s and then released. Blood flow responses were measured under control conditions and during infusion of linopirdine to a plasma concentration of 30 μM (Fig. 8; n = 6 pigs). Data presented in Table 5 quantify the effects of linopirdine on reactive hyperemia. There were no changes in baseline flow, peak hyperemic flow (which reflects the minimum vascular resistance obtained), debt area, repayment area, or flow measured 30 s after the release of occlusion.

Fig. 8.

Linopirdine and ischemic vasodilation (reactive hyperemia). A plot of coronary blood flow vs. time is shown for a reactive hyperemia protocol. A clamp on the coronary perfusion line was used to reduce blood flow to zero for 15 s and then released. Coronary blood flow quickly increased and then returned the baseline level over the next minute. The protocol was repeated in the same pig during infusion of linopirdine to reach a coronary plasma concentration of 30 μM. Linopirdine had no significant effect on ischemic vasodilation (see also Table 5).

Table 5.

Quantification of coronary reactive hyperemia

	Control	Linopirdine (30 μM)
Baseline flow, ml·min−1·g−1	0.66 ± 0.07	0.80 ± 0.11
Flow debt, ml/g	0.17 ± 0.02	0.20 ± 0.03
Peak flow, ml·min−1·g−1	4.40 ± 0.56	4.57 ± 0.54
Repayment volume, ml/g	1.32 ± 0.31	1.09 ± 0.36
Flow 30 s after reperfusion, ml·min−1·g−1	1.19 ± 0.16	1.16 ± 0.33

Linopirdine and H₂O₂-induced coronary vasodilation.

Our third in vivo experiment was designed to determine whether K_V7 channels mediate the coronary blood flow increase elicited by H₂O₂. We reasoned that if K_V7 channels were the redox-sensitive targets of H₂O₂, linopirdine should attenuate H₂O₂-induced coronary vasodilation. This would be similar to what we have shown previously for another K⁺ channel blocker, 4-aminopyridine (12). Pigs (n = 5) were instrumented for measuring coronary blood flow in the pressure-clamped LCx artery. Coronary vasodilation induced by H₂O₂ can be seen in Fig. 9. H₂O₂ was infused to reach plasma concentrations of 0.3, 1, 3, and 10 μM, and this increased coronary blood flow 9 ± 6, 38 ± 15, 71 ± 22, and 126 ± 48%, respectively. After coronary blood flow returned to the control level, an intracoronary infusion of linopirdine was commenced to reach a plasma concentration of 10 μM and the H₂O₂ infusion protocol was repeated. Linopirdine did not inhibit H₂O₂-induced coronary vasodilation, because the same four concentrations of H₂O₂ increased coronary blood flow 12 ± 4, 47 ± 11, 94 ± 21, and 190 ± 40%, respectively (Fig. 9).

Fig. 9.

Linopirdine and the vasodilation in response to exogenous H₂O₂. The graph shows group data (n = 5) for coronary blood flow vs. intracoronary concentration of exogenous H₂O₂. H₂O₂-induced coronary vasodilation was quantified twice in each pig: before (control) and during simultaneous administration of a K_V7 channel antagonist (10 μM linopirdine). H₂O₂ significantly reduced coronary vascular resistance during both challenges (i.e., coronary pressure was held constant at 100 mmHg and flow increased). There was, however, no effect of linopirdine on the response, because the K_V7 channel blocker did not attenuate H₂O₂-induced coronary vasodilation.

To further test the role of K_V7 channels in coronary vasodilation, we designed an in vivo experiment to determine whether K_V7 channels play a role in the increase of coronary blood flow that accompanies an escalating rate of myocardial metabolism. Heart rate is a major determinant of myocardial oxygen consumption; therefore, accelerating the heart rate by electrical pacing is a way to increase myocardial metabolic demands and increase coronary blood flow. Our reasoning was that if K_V7 channels were involved in coronary metabolic vasodilation, then linopirdine should diminish pacing-induced coronary vasodilation. This would be similar to what we have shown previously for 4-aminopyridine, an unrelated K⁺ channel blocker (40). To elicit coronary metabolic vasodilation (Fig. 10), the heart was paced +25, +50, and +75 beats/min above the intrinsic rate. This was done before (control) and during intracoronary infusion of linopirdine to reach a concentration of 10 μM in the coronary plasma in two pigs. Under control conditions, pacing increased coronary blood flow an average of 6, 26, and 48% at +25, +50, and +75 beats/min, respectively. Linopirdine had no effect on coronary metabolic vasodilation, as coronary blood flow in the presence of this drug increased an average of 3, 28, and 51% at +25, +50, and +75 beats/min, respectively. Table 6 contains relevant arterial and coronary venous blood gas data obtained during these experiments; linopirdine treatment did not reduce coronary venous Po₂ or oxygen content.

Fig. 10.

Linopirdine and coronary metabolic vasodilation. Coronary blood flow is plotted against the rate of myocardial oxygen consumption. Data are from 2 pigs. Measurements were made at resting heart rate and when heart rate was paced +25, +50, and +75 beats/min. This was done before (control; open symbols and dashed line for linear regression) and during intracoronary infusion of a K_V7 channel antagonist (10 μM linopirdine; filled symbol and solid line for linear regression). Tachycardia significantly increased coronary blood flow during both pacing protocols. There was, however, no effect of linopirdine on the pacing-induced coronary blood flow response. In other words, linopirdine did not attenuate coronary metabolic vasodilation.

Table 6.

Blood gas parameters during cardiac pacing before and during linopirdine infusion

		Pacing, beats/min
	Baseline	+25	+50	+75
Arterial
pH, mmHg
Control	7.44	7.44	7.44	7.43
Linopirdine	7.45	7.45	7.47	7.42
Pco2, mmHg
Control	38	42	42	43
Linopirdine	43	41	38	38
Po2, mmHg
Control	93	93	91	87
Linopirdine	91	92	96	95
O2 content, ml/dl
Control	17.7	17.4	17.3	17.9
Linopirdine	17.9	17.9	17.4	17.4
Venous
pH, mmHg
Control	7.41	7.40	7.39	7.39
Linopirdine	7.40	7.41	7.41	7.40
Pco2, mmHg
Control	55	57	58	57
Linopirdine	57	55	56	55
Po2, mmHg
Control	18	18	18	19
Linopirdine
O2 content, ml/dl	18	18	17	22
Control	4.1	3.8	4.2	4.3
Linopirdine	3.5	3.4	3.5	4.5

Note that the venous analysis was performed on blood drawn from a coronary vein. Data are represented as the average from 2 pigs.

Linopirdine and bradykinin-induced coronary vasodilation.

Our final in vivo experiment examined the effect of linopirdine on bradykinin-induced (endothelium dependent) coronary vasodilation. The rationale for this experiment was based on 1) our demonstrated role for the endothelium in flupirtine-induced relaxation of coronary artery rings (Fig. 3) and 2) the inhibition by linopirdine of bradykinin-induced relaxation in coronary artery rings (Fig. 4). For this experiment, coronary blood flow was measured across a range of intracoronary bradykinin infusion rates (0.03–0.3 μg/min) before and during intracoronary infusion of linopirdine to a plasma concentration of 30 μM (n = 4). Under control conditions, bradykinin infusion significantly increased coronary blood flow at all doses (Fig. 11). Simultaneous infusion of linopirdine significantly attenuated bradykinin-induced coronary vasodilation (Fig. 11). Neither the infusion of bradykinin nor linopirdine affected heart rate or systolic, diastolic, and mean blood pressures (Table 7).

Fig. 11.

Linopirdine inhibits bradykinin-induced coronary vasodilation in vivo. The plot shows group data (n = 4) for Δcoronary blood flow vs. intracoronary plasma concentration of bradykinin. Coronary blood flow responses were measured twice in each pig: the first set of bradykinin responses was elicited under control conditions, whereas the second set of bradykinin responses was measured during simultaneous infusion of linopirdine to reach 30 μM in the coronary plasma. *P < 0.05 for Δcoronary blood flow values between groups.

Table 7.

Cardiac and hemodynamic parameters during in vivo bradykinin experiments

		Bradykinin, μg/min
	Baseline	0.03	0.1	0.3
Systolic pressure, mmHg
Control	109 ± 12	105 ± 10	103 ± 10	105 ± 10
Linopirdine	100 ± 9	106 ± 7	108 ± 6	107 ± 6
Diastolic pressure, mmHg
Control	76 ± 9	72 ± 7	70 ± 6	71 ± 7
Linopirdine	72 ± 5	74 ± 4	73 ± 4	71 ± 4
Mean pressure, mmHg
Control	91 ± 10	87 ± 8	85 ± 7	86 ± 8
Linopirdine	84 ± 6	88 ± 5	88 ± 4	87 ± 4
Heart rate, beats/min
Control	85 ± 15	88 ± 17	91 ± 18	91 ± 18
Linopirdine	96 ± 10	95 ± 15	90 ± 17	88 ± 18

DISCUSSION

By measuring coronary blood flow in anesthetized, open-chest swine, we tested the hypothesis that K_V7 channels are present in cells of the coronary vascular wall and regulate vasodilation in vivo. We also used qPCR, Western blot, isometric tension recording, patch-clamp electrophysiology, and IHC to study K_V7 channels (and their genes) in the porcine coronary artery. Our data indicate that KCNQ genes and K_V7 channel protein are expressed in the vascular wall of coronary arteries. These K_V7 channels are unlikely to be regulators of baseline coronary vascular resistance or redox-sensitive targets of H₂O₂. K_V7 channels do not appear responsive to vasodilators produced by the ischemic myocardium or normal working myocardium (i.e., metabolic vasodilators). However, despite the apparent absence of an influence on some major physiological aspects of coronary vascular regulation, K_V7 channels may play an important role in paracrine-mediated endothelium-dependent responses, because linopirdine significantly diminished coronary vasodilator responses to bradykinin ex vivo and in vivo.

KCNQ genes and K_V7.4 protein are expressed in the coronary arteries of swine (Figs. 1 and 2). Functional K_V7 channels are expressed in some cell type(s) of the vascular wall, because flupirtine relaxes coronary arteries in a linopirdine-sensitive manner ex vivo (Fig. 3). It is not clear what cell type(s) of the coronary vascular wall express K_V7 channels, because they might be in the smooth muscle (Fig. 5) and/or the other cell types such as the endothelium (Figs. 2, 4, and 11). We did not observe effects of K_V7 channel modulators by themselves (i.e., no flupirtine-induced vasodilation or any linopirdine-induced vasoconstriction; Fig. 7). Thus K_V7 channels apparently do not determine the basal level of coronary vascular resistance [although they are seemingly expressed in coronary resistance vessels <200 μm in diameter (11); Fig. 2]. Similarly, there was no linopirdine-sensitivity in the responses to some of the most physiologically relevant stimuli (tachycardia or ischemia; Figs. 8 and 10).

In isolated coronary rings, we observed flupirtine-induced and linopirdine-sensitive vascular relaxation (Fig. 3) that might be attributed to K_V7 channels, particularly K_V7 channels of the endothelium (Fig. 2). K_V7 channel modulators could theoretically influence vascular tone in at least two ways. The first way would be a direct effect whereby K_V7 channel agonists and antagonists change the membrane potential of vascular smooth muscle by binding to K_V7 channels in the sarcolemma. This direct action has been demonstrated using current-clamp techniques on single vascular smooth muscle cells from the mouse portal vein, where K_V7 channel inhibition depolarized the membrane potential (49). Our data do not strongly support this idea of a direct smooth muscle action of K_V7 modulators, at least with regard to the coronary circulation of swine, because linopirdine did not have inhibitory effects on the steady-state current in smooth muscle cells where one would expect to see K_V7 channels function (Fig. 6). The second mode of action would be an indirect effect in which K_V7 channel agonists and antagonists do not act on the vascular smooth muscle cell itself, but rather on neighboring cell types (e.g., neurons, endothelium, or adipocytes) to influence the production and/or release of vasoactive factors (e.g., norepinephrine, nitric oxide, or hydrogen sulfide, etc.). This type of indirect action has been demonstrated for sympathetic nerves innervating rat mesenteric arteries, where K_V7 channel inhibition enhances excitatory purinergic neurotransmission (19). Our data do support an indirect action of K_V7 channel modulators and indicate 1) a role for the endothelium in flupirtine-induced relaxation of coronary artery rings (Fig. 3), 2) the function of linopirdine-sensitive K_V7 channels in bradykinin-induced coronary relaxation ex vivo (Fig. 4) and vasodilation in vivo (Fig. 11), and 3) endothelial expression of K_V7.4 immunoreactivity (Fig. 2). The direct and indirect mechanisms of K_V7 channel modulators are likely not exclusive, because they could interact to regulate vascular tone. For example, K_V7 channel inhibition contracts murine arteries and also augments contractions to the adrenergic agonist phenylephrine (50).

Members of the K_V7 channel family are encoded by five known genes (KCNQ1–KCNQ5). Expression of these genes produces subunits that make up delayed rectifier (noninactivating) K⁺ channels such as the classic M-current, named for its suppression by muscarinic agonists (6). K_V7 channels are widely recognized to regulate the excitability of cardiac myocytes and neurons. Moreover, defects in these genes (at least KCNQ1 through KCNQ4) cause channel dysfunction and underlie a variety of pathologies such as long QT syndrome, epilepsy, and deafness. The first description of K_V7 channels in vascular smooth muscle was made over a decade ago, when KCNQ1 and a novel splice variant, KCNQ1b, were observed in the mouse portal vein (29, 35). Since that seminal observation, it has been found that KCNQ isoforms are also expressed in gastrointestinal, myometrium, airway, and bladder smooth muscle (1, 7, 17, 29). K_V7 channels can be homo- or heterotetramers; however, data regarding which subunits can form heteromers in vascular tissues is recent (10), and we do not yet know the exact subunit composition of K_V7 channels in swine coronary vascular smooth muscle. Subunit composition is likely to influence the responses to physiological regulators (8), including responses to endothelium-derived relaxing factors (16), paracrine molecules (10), and transmembrane signaling machinery (43).

We found mRNA encoding K_V7 channels to be present in the coronary artery, but we do not definitively know what cell type(s) is(are) expressing the genes. Specifically, our qPCR and Western blot results (Fig. 1) do not necessarily indicate that K_V7 channels are expressed in the vascular smooth muscle, and our IHC data (Fig. 2) suggest that the mRNA signal may possibly arise from cells in the adventitia or intima. The presence of K_V7 channels in various cells of the vascular wall could potentially explain the flupirtine-induced and linopirdine-sensitive relaxation of LCx artery rings (Fig. 3). That is, the K_V7 drugs may be acting on cells that release factors responsible for the vascular relaxation (endothelium- and/or adipocyte-derived relaxing factors). The presence of K_V7 channels in cells of the intimal layer could potentially explain the endothelial dependence of flupirtine-induced relaxation (Fig. 3) and the linopirdine-sensitivity of bradykinin-induced vascular relaxation ex vivo (Fig. 4) and bradykinin-induced coronary vasodilation in vivo (Fig. 11). Mechanisms might conceivably include K_V7 channel contributions to endothelial membrane potential, Ca²⁺ influx, and the production of relaxing factors.

It was recently reported that K_V7 channels are expressed in porcine coronary arteries and involved in hypoxic vasodilation (16). Similar to our experiments, Hedegaard et al. (16) used whole artery samples and detected KCNQ1–KCNQ5 transcripts by RT-PCR as well as K_V7.4 (and K_V7.5) protein by Western blot analysis. Our IHC data extend their work and demonstrate a nonuniform distribution of K_V7.4 immunoreactivity across the cell types of the vascular wall (Fig. 2). Our patch-clamp work differs from theirs in that we used different K_V7 channel blockers (linopirdine vs. XE991), but we both demonstrate some moderate inhibition of K⁺ current in coronary vascular smooth muscle at more depolarized test potentials, but no apparent membrane potential change (i.e., the voltage at which the whole cell current reverses does not visibly depolarize with application of a K_V7 channel antagonist).

The findings of the present study agree with the recent findings of Lee et al. (23), who demonstrated marked heterogeneity in K_V7 channel function in cerebral and coronary smooth muscle of mice. These investigators studied the expression and function of K_V7 channels in the cerebral (circle of Willis, basilar artery) and coronary (LAD artery) circulations. KCNQ4 was the predominant transcript observed, and although K_V7.4 protein was detected in the LAD artery, its expression was greatly reduced compared with that in the cerebral vessels. There was no significant K⁺ current in murine LAD artery myocytes that was sensitive to K_V7 channel modulators, unlike myocytes from the circle of Willis or basilar artery. The absence of major K_V7 channel function in coronary smooth muscle certainly contrasts with numerous studies in a wide variety of vascular beds (cf. most references in Table 2), but it should be recognized that KCNQ knockout mice have no known coronary phenotypic differences, and moreover, no vascular (or other smooth muscle) abnormalities have been reported for mice lacking KCNQ genes. These knockout animals do, however, demonstrate profound deficits in other body systems, including cardiac long QT syndrome, inner ear problems, deafness, intestinal epithelium dysfunction, pancreatic disorders, epilepsy, etc.

A caveat for our in vivo studies is the use of pharmacological agents and whether the coronary plasma concentrations of K_V7 activators and inhibitors might possibly be insufficient. Existing literature seems to validate the doses (intracoronary plasma concentrations) we used. For example, intravenous linopirdine at 0.10 mg/kg increases mean arterial pressure in Sprague-Dawley rats (26), and values for rat blood volume (22) and hematocrit (37) allow one to translate that dose to initial plasma concentrations between 6.5 and 9.8 μM. Linopirdine blocks all five K_V7 subtypes with IC₅₀ values between 5 and 16 μM (15). We used 10–30 μM linopirdine in most experiments (e.g., Figs. 8–11) and up to 100 μM in the experiment shown in Fig. 7, which should be an effective range. Furthermore, we infused linopirdine directly into the coronary circulation, alleviating any concerns about hepatic metabolism. A plasma concentration of 10 μM should be sufficient, because doubling the dose to 20 μM linopirdine has no further effect on the constriction of isolated rat isolated mesenteric arteries (26). Moreover, we assessed a number of K_V7 channel modulators in vivo in search of coronary vascular effects. For example, we tried three different K_V7 agonists (flupirtine, L364-373, and retigabine), none of which were able to increase coronary blood flow. Flupirtine should be an ideal K_V7 modulator for our in vivo studies, because it has been in therapeutic use for over 30 years, has excellent bioavailability, and shows high selectivity for K_V7 channel activation [with reports of significant vasodilation at concentrations ≥100 nM (32)]. However, we obtained no evidence suggesting that flupirtine (or any other K_V7 channel modulator) altered coronary vascular tone. Another caveat concerns the patch-clamp measurements we made at room temperature (Figs. 5 and 6) and their comparison to results of ex vivo and in vivo experiments at physiological temperatures (Figs. 3, 4, and 8–11). K_V7 channels are temperature sensitive, but not in a manner that we would expect to change our conclusion. This is because increasing temperature accelerates activation and inactivation kinetics and increases the maximal conductance, but the voltage dependence of activation and pharmacology are little affected (30). Thus we would predict that experiments at more physiological temperatures would increase the size of any K_V7 currents and change their shape (i.e., kinetics); however, elevated temperature would be unlikely to change the effects of K_V7 channel modulators.

The data lead us to conclude that KCNQ genes (and K_V7.4 protein) are expressed in some cells of the coronary vascular wall and that K_V7 channels play a role in endothelium-dependent vasodilation. The mechanism(s) by which K_V7 channels modulate endothelial responses have not been identified. Additionally, it may be important to further study the distribution of K_V7.5 in the coronary vascular wall, because KCNQ5 expression was also high. The search for specific K⁺ channels that regulate baseline coronary vascular resistance and function as end effectors in metabolic and ischemic coronary vasodilation would likely benefit from focus on other K_V channel families, such as K_V1 (12, 34).

GRANTS

This study was supported by National Heart, Lung, and Blood Institute Grants U01HL118738 (to D. Beard and G. Kassab) and R01HL11760 (to J. Tune and K. Mather). A. G. Goodwill was supported by American Heart Association Award 13POST1681001813. J. N. Noblet and D. Sassoon were supported by National Center for Advancing Translational Sciences, Clinical and Translational Sciences Award TL1 TR000162 (to A. Shekhar).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

A.G.G., L.F., J.N.N., E.D.C., D.J.S., Z.B., G.S.K., J.D.T., and G.M.D. conception and design of research; A.G.G., L.F., J.N.N., E.D.C., D.J.S., Z.B., G.S.K., J.D.T., and G.M.D. performed experiments; A.G.G., L.F., J.N.N., E.D.C., D.J.S., Z.B., G.S.K., J.D.T., and G.M.D. analyzed data; A.G.G., L.F., J.N.N., E.D.C., D.J.S., Z.B., G.S.K., J.D.T., and G.M.D. interpreted results of experiments; A.G.G., L.F., J.N.N., E.D.C., D.J.S., Z.B., G.S.K., J.D.T., and G.M.D. prepared figures; A.G.G., L.F., J.N.N., E.D.C., D.J.S., Z.B., G.S.K., J.D.T., and G.M.D. edited and revised manuscript; A.G.G., L.F., J.N.N., E.D.C., D.J.S., Z.B., G.S.K., J.D.T., and G.M.D. approved final version of manuscript; G.M.D. drafted manuscript.