Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Since Noma (1983) first identified a potassium channel that was inhibited by intracellular ATP in cardiac myocytes (referred to as K_ATP), it has been thought that K_ATP may play an important role in regulation not only of the resting membrane potential but also of membrane excitability. Many cytosolic metabolic regulators seem to have effects on K_ATP in a wide variety of tissues (reviewed by Ashcroft and Ashcroft, 1990; Edwards and Weston, 1993; Kitamura and Kuriyama, 1994). Therefore, much effort has been expended in the synthesis and development of more selective KCOs for use in several therapeutic areas (such as ischemic heart disease, hypertension, intermittent claudication, asthma and idiopathic detrusor instability).

KCOs have been introduced mainly as vasodilators in clinical treatment. Nicorandil (Sigmart), a nitro-compound, is one of the most effective KCOs in use. It is thought to produce its effect on smooth muscles either by abolishing action potentials (Furukawa et al., 1981; Itoh et al., 1981) or by increasing the K⁺ permeability (Yamanaka et al., 1985). However, using single-channel recordings, the target K⁺ channels for nicorandil in vascular smooth muscle cells have remained elusive. Nicorandil (20-200 µM) activates an extracellular and intracellular Ca⁺⁺-activated K⁺ channel that is suppressed by 4-AP, endothelin and intracellular ATP (porcine coronary artery: Inoue et al., 1989; Miyoshi et al., 1992; Wakatsuki et al., 1992; K_R channel). On the other hand, nicorandil (500 µM) opens an intracellular Ca⁺⁺-dependent K⁺ channel that is blocked by intracellular ATP, Mg⁺⁺ and extracellular application of both TEACl and 4-AP (rat portal vein: Kajioka et al., 1990: 10-pS K channel). Recently, Kamouchi and Kitamura (1994) have reported that in rabbit portal vein, nicorandil (300 µM-1 mM) activates a K_ATP that is sensitive to intracellular ATP, Mg⁺⁺ and NDPs. Thus, in spite of the common vasodilator action of nicorandil, it remains unresolved whether these quite different target K⁺ channels arise from different experimental conditions (such as experimental temperature, cultured cells, freshly dispersed cells and so on) or are species-dependent. Moreover, it still remains uncertain whether nicorandil activates K⁺ channels directly or does so through intracellular messengers such as cyclic GMP as a result of the activation of guanylate cyclase by nicorandil. Kubo et al. (1994) have clearly concluded that an increase in intracellular cyclic GMP directly modulates the gating of K_ATP in rat aorta using single-channel recordings.

Recently, glibenclamide has been shown to be a selective K_ATP blocker at submicromolar concentration (Bray and Quast, 1992; Edwards and Weston, 1993; Kitamura and Kuriyama, 1994; Teramoto and Brading, 1996) and, because of its potent selectivity, it has become an essential tool in defining whether a synthesized compound may be a KCO (Quast, 1993). Although nicorandil activates K_ATP in all the vascular smooth muscles reported above, no data have been presented about the inhibitory effect of glibenclamide either on the nicorandil-induced membrane currents or on the nicorandil-activated K⁺ channels. The present study was designed to investigate the effects exerted by nicorandil not only on the membrane currents but also on the K_GS in urethral smooth muscle and on the resting urethral tone. This is the first report in which patch-clamp techniques have been used to study the nicorandil-activated K⁺ channels in a nonvascular smooth muscle cell. We discuss both the target K⁺ channels for nicorandil and the involvement of cyclic GMP in its action on pig proximal urethra.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Female fresh pig urethras still attached to the bladder were obtained from a local abattoir and transported in a cold (6°C-8°C) modified Krebs solution (mM): 137 Na⁺, 5.9 K⁺, 1.2 Mg⁺⁺, 2.0 Ca⁺⁺, 132.7 Cl, 15.4 HCO₃, 1.2 H₂PO₄ and 11.5 glucose bubbled with 97% O₂ and 3% CO₂ (pH 7.25-7.3). The proximal urethra, 1 to 2 cm from the bladder neck (about 4-5 cm below the ureteric orifices), was excised, and both the connective tissue and the mucosa were removed in the above solution at room temperature (21°C-23°C) under a dissection microscope.

Tension measurement. To make possible the recording of isometric tension, fine strips (1.0-1.2 mm in length, 0.4-0.5 mm wide and 0.3-0.4 mm thick) were prepared as described by Teramoto and Brading (1996). Tension equivalent to 1 g weight was applied to each strip, which was then allowed to equilibrate for 1 h before the start of experiments (Bridgewater et al., 1993; Teramoto and Brading, 1996). To prevent both noradrenaline outflow from sympathetic nerve terminals and beta adrenoceptor stimulation, 3 µM guanethidine and 0.3 µM propranolol were added to the modified Krebs solution throughout the experiments. The data were recorded on a DAT recorder (48 kHz, SONY, DTC-1000ES, Tokyo, Japan) with the drug application time as a trigger pulse and were analyzed on a computer (Macintosh Quadra 610, Apple Computer UK Limited, Uxbridge, Middlesex, U.K.) using the commercial software MacLab 3.4.2 (ADInstruments Pty Ltd, Castle Hill, Australia) through 50-Hz digitalization. The tension experiments were performed at 37°C.

Cell dispersion. The cell preparation was a modification of the previous method (Teramoto and Brading, 1996). Briefly, thin strips of smooth muscle (8-10 mm × 2-3 mm) were dissected from the proximal urethral wall and placed in a salt solution (mM): 140 Na⁺, 5 K⁺, 0.5 Mg⁺⁺, 0.5 Ca⁺⁺, 147 Cl, 10 glucose, HEPES/Tris, titrated to pH 7.35 to 7.4, containing papain (0.2-0.3 mg/ml) bubbled with O₂ at 6°C to 8°C for 20 min. The digested strips were washed in a Ca⁺⁺-free salt solution complemented with 1 mg/ml BSA and preincubated in this solution at 35°C for 4 to 5 min. The strips were then incubated in Ca⁺⁺-free salt solution containing 0.3 to 0.4 mg/ml collagenase (Type IA) at 35°C for 10 to 12 min. The strips were washed in 100 µM Ca⁺⁺-containing salt solution and placed in a test tube. Then the test tube was tapped gently until sufficient cells were yielded (the tapping method, Teramoto and Brading, 1996). They were stored at 4°C and were normally used within 5 h for experiments.

Solutions. In tension measurement, modified Krebs solution was used (mM): 137 Na⁺, 5.9 K⁺, 1.2 Mg⁺⁺, 2.0 Ca⁺⁺, 132.7 Cl, 15.4 HCO₃, 1.2 H₂PO₄ and 11.5 glucose bubbled with 97% O₂ and 3% CO₂ (pH 7.35-7.40 at 37°C). The following solutions were used for whole-cell recording. The PSS in the bath had the following composition (mM): 140 Na⁺, 5 K⁺, 1.2 Mg⁺⁺, 2.0 Ca⁺⁺, 151.4 Cl, 10 glucose, 10 HEPES and was titrated to pH 7.35 to 7.40 with Tris base; 60 mM K⁺ PSS was obtained by replacing 55 mM Na⁺ with equimolar K⁺. The pipette solution contained (mM): 140 K⁺, 5 Mg⁺⁺, 150 Cl, 5 EGTA, 5 glucose, 10 HEPES/Tris (pH 7.35-7.40). For the single-channel recordings (cell-attached configuration and inside-out patches), the composition of both pipette and bath solution was the same (mM): 140 K⁺, 140 Cl, 5 EGTA, 10 HEPES/Tris (pH 7.35-7.40), i.e., symmetrical K⁺ conditions. The concentrations of Mg-ATP and free Ca⁺⁺ ([Ca⁺⁺]_i) were calculated using the commercial software EQCAL (Biosoft, Cambridge, U.K.). Cells were allowed to settle in the small experimental chamber (80 µl in volume). The bath solution was superfused by gravity throughout the experiments at a rate of 2 ml/min. The following chemicals were used: ADP, GDP, IDP, ATP, BSA, EGTA, 8-bromo cyclic GMP, HEPES, collagenase, glibenclamide, guanethidine, propranolol, papain and SNP (Sigma, Dorset, U.K.), methylene blue and Tris (BDH Chemicals Ltd., Dorset, U.K.). NDPs and ATP were added as Na salt, and 100 mM stock solutions (dissolved in 140 mM KCl solution) of these were titrated to pH 7.4 and frozen at 70°C. Dilution of the stock solution was made immediately before the application. Nicorandil (kindly provided by the Chugai Pharmaceutical Co. Ltd, Tokyo, Japan) was dissolved daily in 1 N HCl as 1 M and immediately diluted with the experimental bath solution used to 100 mM as a stock solution, adjusting pH (7.35-7.40). Levcromakalim (kindly provided by SmithKline Beecham Pharmaceuticals, Harlow, U.K.) was prepared daily as 300 mM stock solutions in DMSO. Glibenclamide was also dissolved to 100 mM (for patch-clamp experiments) or 250 mM (for tension recordings) in DMSO. Both drugs were diluted just before use. The final concentration of DMSO was less than 0.1%, and this concentration did not affect either the membrane currents or the potassium channels.

Recording procedure and data analysis. The experimental system used was essentially the same as described previously (Hamill et al., 1981; Teramoto and Brading, 1996). Briefly, patch-clamp experiments were performed through a L/M-EPC 7 patch-clamp amplifier (List-Medical-Electronic, Darmstadt, Germany) in conjunction with an AD/DA converter (DT2801A, Data Translation, Marlboro, MA). The sampled current data were filtered at 10 kHz and stored, together with potential records, on videotape using a pulse code modulation unit (16-bit resolution, SONY PCM-701, Tokyo, Japan) coupled to a video recorder (Panasonic AG-6200, Osaka, Japan) for subsequent off-line analysis. All experiments were carried out at room temperature (21°C-23°C). The whole-cell current data were low-pass-filtered at 500 Hz (3 dB) by an 8-pole Bessel filter, sampled at 25 ms (continuous traces) or 1 ms (ramp currents) and analyzed on a computer (Macintosh Quadra 610) using the commercial software Mac Lab 3.4.2 (ADInstruments Pty Ltd, Castle Hill, Australia); leakage current was not subtracted. For single-channel recording, the stored data were low-pass-filtered at 2 kHz (3 dB) and sampled into the computer with an interval of 80 µs using the PAT program (kindly provided by Dr. Dempster, the University of Strathclyde). Single-channel events were detected using a half-amplitude criterion (Colquhoun and Sigworth, 1995) and were inspected manually. Because we could not determine the total number of functioning channels in each patch, we limited examination to the opening kinetics only. Open lifetime distribution was log-binned using the method of McManus et al. (1987). When the square root of the number of events in a bin was plotted against the open lifetime, the component of the open lifetime distribution appeared as a clear peak with the respective time constant falling in the vicinity of the distribution peak (Sigworth and Sine, 1987). Conditional probability density function was fitted to the open lifetime distribution by the method of maximum likelihood. However, events briefer than 500 µs were not included in the evaluation, and no correction was made for missed events. The all-point amplitude histogram was obtained from a continuous recording of 30 s or 2 min and fitted with the Gaussian distribution function using a least-squares fitting. Continuous traces in the figures (>4 s) were obtained from records filtered at 500 Hz for presentation. In the present experiments, we did not define the total number of channels present in each patch membrane, so the channel activity was calculated by using the following equation from an all-point amplitude histogram and expressed as an NP_o value. where N is the number of channels, P_o is the open-state probability, t_j is the time spent at each current level corresponding to j = 0, 1, 2, ... , N, and T is the duration of the recording.

Statistical analysis. Statistical analyses were performed with Student's t test for paired values. Changes were considered significant at P < .01. Data are expressed as the mean ± S.D.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Effects of nicorandil on membrane currents in pig proximal urethra. Nicorandil (500 µM) evoked a sustained amplitude outward current under quasi-physiological conditions (pipette solution, 140 mM KCl containing 5 mM EGTA; bath solution, PSS) when the holding membrane potential was kept at 50 mV in pig proximal urethra. An additional application of 10 µM glibenclamide completely suppressed this current to slightly below the control level (fig. 1A). The minimum concentration of nicorandil that produced a detectable outward current was 30 µM, and the peak amplitude of the nicorandil-induced outward current increased in a concentration-dependent manner (fig. 1B). To make a rough estimation of the ion selectivity of this outward current and to obtain both the current-voltage (I-V) relationships and the reversal potential, voltage ramps were applied (see inset in fig. 2A) and the extracellular potassium concentration ([K⁺]_o) was changed by iso-osmotic substitution of sodium. Figure 2A shows the experimental protocol. In the absence of nicorandil (control), I-V relationships were obtained by the application of six ramp pulses in solutions containing 5 mM K⁺ followed by 60 mM K⁺ and then back to 5 mM K⁺. Nicorandil (500 µM) was then applied in the bath solution and caused a sustained outward current. In the presence of 500 µM nicorandil, the same voltage protocol was performed. When [K⁺]_o was raised from 5 mM to 60 mM, the basal sustained nicorandil-induced membrane current at 50 mV changed from outward to inward. When [K⁺]_o was returned to 5 mM, the nicorandil-induced current became outward current again. Figure 2B shows the average of the six ramp currents before and during the application of 500 µM nicorandil (5 mM K⁺, 60 mM K⁺) for the cell shown in figure 2A. In each [K⁺]_o condition, the net membrane current activated by 500 µM nicorandil was obtained by subtracting the averaged control current from the mean nicorandil-induced current (fig. 2C). The reversal potential of the nicorandil-induced membrane current in this cell was 75 mV in 5 mM K⁺ (average 77 ± 4 mV, n = 4) and 23 mV in 60 mM K⁺ (average 24 ± 2 mV, n = 4). These values were very close to the theoretical E_K in each [K⁺]_o condition (5 mM K⁺, E_K = 84.2 mV; 60 mM K⁺, E_K = 21.4 mV). These results suggest that in pig proximal urethra, nicorandil-induced membrane currents are carried mainly by potassium ions through channels that are sensitive to glibenclamide.

View larger version (12K): [in this window] [in a new window]   Fig. 1.   Effects of nicorandil on whole-cell membrane current recorded from dispersed smooth muscle cells of pig proximal urethra. The holding membrane potential was kept at 50 mV. A) Effects of 500 µM nicorandil and 10 µM glibenclamide on the membrane current. The bath solution was PSS (2.0 mM Ca++), and the pipette solution was 140 mM KCl containing 5 mM EGTA. The dashed line indicates the zero-current level. B) Relationship between peak amplitude of the membrane outward current (pA) and concentration of nicorandil (µM). The peak amplitude was measured from the base-current level just before nicorandil was applied. Open circles indicate the peak amplitude of nicorandil-induced current. Each symbol indicates the mean of 4 to 6 observations with S.D.

View larger version (21K): [in this window] [in a new window]   Fig. 2.   The nicorandil-induced membrane current is mainly a potassium current due to increased K+ permeability. Holding potential was kept at 50 mV. Bath solution was initially PSS, and [K+]o was transiently raised from 5 mM K+ to 60 mM K+. Pipette solution was 140 mM KCl containing 5 mM EGTA. Similar observations were obtained in three other cells. A) Ramp currents induced by the six ramp potential pulses from 120 mV to 0 mV for 600 ms after a 300-ms conditioning pulse (120 mV, see inset) applied every 10 s before and during application of 500 µM nicorandil. In the absence of nicorandil (control), the ramp membrane currents were obtained in either 5 mM K+ condition () or 60 mM K+ condition (). Nicorandil (500 µM) caused an outward current in PSS. In the presence of nicorandil, the same voltage protocol was performed. When [K+]o was raised from 5 mM () to 60 mM (), the basal sustained 500 µM nicorandil-induced current at 50 mV changed from outward to inward, indicating that the currents were through potassium-selective channels. The vertical deflections indicate ramp currents. The dashed line indicates the zero-current level. B) The mean ramp membrane currents on an expanded time scale in several conditions. Each symbol is the same as in fig. 2A. Nicorandil (500 µM) shifted the potential at which the averaged ramp current crossed the zero-current level from 35 mV () to 68 mV (). When [K+]o was raised to 60 mM, the mean ramp current changed sign at 19 mV (). In the presence of nicorandil, the averaged ramp membrane current intersected at 21 mV (). C) Net membrane currents evoked by nicorandil when [K+]o was either 5 mM or 60 mM. Net membrane current was obtained by subtraction of the two ramp membrane currents (shown in fig. 2B) recorded before and during application of 500 µM nicorandil in each [K+]o condition. The reversal potential of nicorandil-induced current in 5 mM K+ was 75 mV, close to the calculated EK value. The reversal potential in 60 mM K+ was 23 mV.

Nicorandil activates a K_GS in unitary current recordings. To investigate further the nicorandil-induced glibenclamide-sensitive outward currents, single-channel experiments were performed in symmetrical K⁺ conditions (pipette and bath solution: 140 mM KCl containing 5 mM EGTA). To minimize the activation of any voltage- or Ca⁺⁺-dependent K⁺ channels, the holding membrane was kept at 50 mV, and the free Ca⁺⁺ concentration was maintained below 2 nM. In cell-attached configuration, opening of the large-conductance K⁺ channels (Teramoto and Brading, 1994) was sometimes observed even at 50 mV. When 500 µM nicorandil was applied in the bath solution, another channel of small amplitude started to activate (fig. 3A). The amplitude of the small channel was 2.14 pA at 50 mV from an all-points amplitude histogram (fig. 3B). Additional application of 10 µM glibenclamide gradually inhibited the activity of this small channel and, about 1 min later, caused complete suppression. Upon removal of glibenclamide, the channel appeared again and the channel activity recovered (fig. 3C). In the presence of 500 µM nicorandil, I-V relationships were obtained by changing the holding membrane potential from 100 mV to 0 mV with 10-mV increments every 30 s. Figure 4B shows the channel records at the indicated membrane potentials. Figure 4A indicates the I-V relationships and reveals that the reversal potential is about 0 mV (i.e., symmetrical 140 mM K⁺ conditions, E_K = 0 mV) and that the conductance of the small amplitude channel is about 43 pS (43.2 ± 0.8 pS, n = 5).

View larger version (21K): [in this window] [in a new window]   Fig. 3.   Effects of 500 µM nicorandil on the unitary current recording using cell-attached patch in symmetrical 140 mM K+ conditions. Patch membranes contained Ca++-activated K+ channels (large downward deflections) at 50 mV. The dashed line indicates the current base line where the channel is not open. A) When 500 µM nicorandil was applied in the bath, a small amplitude channel was activated. The lower current trace is expended from the upper trace; a and b show the duration of the traces analyzed in fig. 3B. B) The all-points amplitude histograms in the absence a and presence b of 500 µM nicorandil. Continuous lines in the histograms are theoretical curves fitted with a Gaussian distribution function. The abscissa shows the amplitude of the current (pA), and the ordinate shows the percentage value of the probability density function (%). C) In the presence of 500 µM nicorandil, 10 µM glibenclamide reversibly inhibits the nicorandil-induced K+ channel.

View larger version (19K): [in this window] [in a new window]   Fig. 4.   Relationship between the holding membrane potential and the amplitude of the single-channel current activated by 500 µM nicorandil in symmetrical 140 mM K+ conditions. A) Current-voltage relationship obtained using cell-attached patch. The amplitude of the K+ channel currents was taken from the all-points amplitude histograms for 30 min. The line was fitted by the least-squares method. The channel conductance was 43.2 ± 0.8 pS (n = 5). B) Traces are channel activities recorded from the same patch membrane at the indicated membrane potentials. The dashed line indicates the current base line where the channel is not open.

Effects of intracellular Mg-ATP on K_GS. In the presence of 300 µM nicorandil (the bath solution), the opening of the 43-pS K⁺ channel was observed under cell-attached configuration. After the patch membrane was excised (ATP-free internal solution), "run-down" of the channels occurred rapidly even in the presence of nicorandil (fig. 5A). Although the time course of the run-down phenomenon varied from cell to cell, the channel openings disappeared within 60 s. After the run-down was complete, 1 mM Mg-ATP was applied in the bath solution (internal surface of the patch) and the channel was reactivated. This channel was inhibited by the application of 10 µM glibenclamide (6 out of 6 patches, data not shown).

View larger version (19K): [in this window] [in a new window]   Fig. 5.   Application of either 1 mM Mg-ATP or various NDPs to the intracellular side of the inside-out patch reactivated the 43-pS K+ channel after patch excision. The dashed line indicates the current level when the channel is not open. The arrow indicates the time when the inside-out patch was established by excising from the cell. A) Nicorandil (300 µM) activated a 2.14-pA K+ channel in symmetrical K+ conditions (containing 5 mM EGTA) using the cell-attached configuration at 50 mV. After patch excision (inside-out configuration) at the indicated time, "run-down" of the 43-pS K+ channel occurred even in the presence of 300 µM nicorandil (the bath). After the 43-pS K+ channel activity had disappeared, Mg++ and ATP were applied in the bath solution (the internal side of the patch membrane). Mg-ATP (1 mM) reactivated a channel of the same amplitude. We calculated the concentration of Mg-ATP using Mg++ and ATP binding constants (see "Materials and Methods"). B) NDP (1 mM) reactivated the channel-opening activity of the 43-pS K+ channel using an inside-out configuration after channel "run-down." Nicorandil (300 µM) was present in the bath solution.

NDPs reactivate K_GS after run-down. Figure 5B shows that after "run-down," when 1 mM ADP was applied to the intracellular surface of the patch in the presence of 300 µM nicorandil, a K⁺ channel of the same amplitude was reactivated at 50 mV, showing burst-like openings of longer duration. Similar results were obtained by the application of other NDPs (1 mM), such as GDP and IDP. The activity of the K⁺ channel reactivated by 1 mM ADP was completely suppressed by 10 µM glibenclamide (data not shown).

Nicorandil and levcromakalim activate the same K⁺ channel. To investigate the target K⁺ channels for KCOs in pig urethra, we used levcromakalim, a well-known KCO from a class with a different chemical structure from that of nicorandil. In cell-attached configuration, 300 µM nicorandil activated a 2.14-pA K⁺ channel at 50 mV. After washout of nicorandil, the channel completely disappeared. Approximately 5 min later, the same concentration of levcromakalim (300 µM) was applied in the bath and immediately activated a K⁺ channel of the same amplitude (fig. 6A). If the NP_o value in the presence of 300 µM nicorandil is normalized as 1.0, then the relative NP_o value in the presence of 300 µM levcromakalim was increased by approximately 5 times (4.6 ± 1.5, n = 4). To compare further the nicorandil-induced and levcromakalim-induced K⁺ channels, we analyzed the open lifetime during recordings from patches showing only single-channel openings because we could not estimate the total number of channels per patch. The open-lifetime histogram was obtained from 30-s recordings in the presence of each KCO (fig. 6, B and C). There was no significant difference in the mean open lifetime (1.7 ± 0.1 ms, n = 5, in nicorandil compared with 1.6 ± 0.2 ms, n = 5, in levcromakalim). These results suggest that nicorandil and levcromakalim activate the K⁺ channels with not only the same amplitude but also similar opening kinetics, although the potency of the compounds is different (levcromakalim  nicorandil). Thus the target K⁺ channel for KCOs in pig proximal urethra seems to be K_GS.

View larger version (25K): [in this window] [in a new window]   Fig. 6.   Nicorandil (300 µM) activated a K+ channel of the same amplitude as 300 µM levcromakalim at 50 mV, showing similar channel-opening kinetics in the same cell-attached membrane patches of pig urethra. A) Nicorandil (300 µM) activated the 2.14-pA K+ channel in symmetrical K+ conditions at 50 mV. After washout, the activity of the nicorandil-induced channel disappeared. About 5 min later, the application of 300 µM levcromakalim caused the activity of the same amplitude K+ channel. The dashed line indicates the level where the channel is not open. B and C) The open lifetime in the presence of either 300 µM nicorandil (B) or 300 µM levcromakalim (C) at 50 mV. The abscissa shows the log of the open lifetime (ms), and the ordinate shows the square root of the number of events (n1/2). The solid curve indicates a single exponential fitting using the least-squares method. The time constants for the mean open lifetime were 1.7 ms (1.7 ± 0.1 ms, n = 5, nicorandil) and 1.6 ms (1.6 ± 0.2 ms, n = 5, levcromakalim), respectively. We obtained the data from the same patch, filtering at 2 kHz for analysis.

Cyclic GMP did not modulate the activity of K_GS in pig urethra. Nicorandil is well known as a nitro-compound and is therefore thought to be an activator of guanylate cyclase in smooth muscles. To determine whether nicorandil directly activates the 43-pS K⁺ channel in pig urethra or whether it is cyclic GMP that modulates this channel activity, we tested the channel opening using a soluble guanylate cyclase inhibitor, a nitro-compound and a membrane-permeable cyclic GMP. In cell-attached patches, the 43-pS K⁺ channel was activated by 300 µM nicorandil. Figure 7A shows that applying methylene blue (10-100 µM), a selective soluble guanylate cyclase inhibitor, did not inhibit the nicorandil-induced channel activity. However, 10 µM glibenclamide produced reversible inhibition. SNP, a potent nitro-compound, did not activate any channels when applied to the bath at from 10 to 100 µM using cell-attached configuration (fig. 7B). Furthermore, the application of 1 mM 8-bromo cyclic GMP, a membrane-permeable cyclic GMP, did not activate any channels. About 2 min after washout of these drugs, 300 µM nicorandil was applied in this patch. Nicorandil activated the 43-pS K⁺ channel in the same patch membrane. Similar results were obtained in five other patches. These results show that activation by nicorandil of the 43-pS K⁺ channel in pig proximal urethra is unrelated to either the guanylate cyclase pathway or cyclic GMP modulation.

View larger version (15K): [in this window] [in a new window]   Fig. 7.   The nicorandil-induced 43-pS K+ channel was not affected by modulation of the guanylate cyclase pathway or cyclic GMP at 50 mV in pig proximal urethra. Both the pipette solution and the bath solution were 140 mM K+ containing 5 mM EGTA. The dashed line indicates the level where the channel is not open. A) The 300 µM nicorandil-induced K+ channel was not suppressed by methylene blue (10 and 100 µM) but was suppressed by 10 µM glibenclamide. B) Neither SNP (10 and 100 µM) nor 1 mM 8-bromo cyclic GMP activated any channels in a patch that contained the nicorandil-induceable K+ channels. Similar results were obtained in five other patches.

Effects of nicorandil as compared with levcromakalim on the urethral resting tone. Application of nicorandil (10 µM) caused a concentration-dependent relaxation in pig proximal urethra (fig. 8, A and B; n = 10). Although 500 nM nicorandil did not cause any significant urethral relaxation, the same concentration of levcromakalim applied to the same strip about 25 min after nicorandil had been washed off elicited relaxation (n = 6, fig. 8C). Because 5 µM nicorandil induced less relaxation than 5 µM levcromakalim, nicorandil is less potent than levcromakalim at inducing urethral relaxation (n = 5, fig. 8C). Figure 8D shows that after 10 µM nicorandil-induced relaxation had stabilized, application of 10 µM methylene blue partially inhibited the relaxation (n = 5). Additional application of 1 µM glibenclamide caused a further inhibition, although a small component of nicorandil-induced relaxation remained. Further application of both 100 µM methylene blue and 10 µM glibenclamide did not influence this remaining component, which recovered to the control level upon the removal of 10 µM nicorandil (n = 5).

View larger version (22K): [in this window] [in a new window]   Fig. 8.   Effects of nicorandil on the resting tone of pig proximal urethra. Guanethidine (3 µM) and propranolol (0.3 µM) were present throughout the experiments. A) Nicorandil was applied for 4 min in the modified Krebs solution. The results illustrated were obtained from a single smooth muscle strip. Just before applying each concentration of each drug, we measured the average resting tone for 4 min. The amplitude of relaxation was obtained from the last 30 s of application of each drug as the mean value after subtracting the average resting tone. B) Relative amplitude of nicorandil-induced relaxation. The abscissa shows the concentration of nicorandil (µM), and the ordinate shows the relative relaxation amplitude, 1 mM nicorandil-induced relaxation being normalized as 1.0. Each symbol represents the mean (n = 10) with ± S.D. shown by the vertical bar. C) The amplitude of nicorandil-induced urethral relaxation (6-min application) was compared with that of levcromakalim-induced relaxation at the same concentration (500 nM, 5 µM) by use of the same tissue strips (n = 5). D) Effects of both glibenclamide and methylene blue on 10 µM nicorandil-induced relaxation in pig urethra. Application of 10 µM methylene blue partially inhibited the 10 µM nicorandil-induced relaxation. Glibenclamide (1 µM) was subsequently added to the bath and caused a further inhibition of the nicorandil-induced relaxation. Note that both 10 µM glibenclamide and 100 µM methylene blue did not cause a further significant inhibition of the 10 µM nicorandil-induced relaxation (n = 5).

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

In the present experiments, we have been able to demonstrate that nicorandil induces glibenclamide-sensitive outward currents through the activation of a K_GS in smooth muscle cells dispersed from pig proximal urethra and causes a concentration-dependent relaxation in the resting tone of intact tissue.

The target K⁺ channels for nicorandil in smooth muscle cells. The nicorandil-induced K_GS in pig urethra differs in many respects from the nicorandil-induced K_ATP described in either porcine coronary artery (K_R, Inoue et al., 1989; Miyoshi et al., 1992; Wakatsuki et al., 1992) or rat portal vein (10-pS K channel, Kajioka et al., 1990). 1) Although the experimental conditions (especially the K⁺ gradient) are slightly different, the conductance of the nicorandil-activated K⁺ channel in pig urethra is quite different [K_R channel in porcine coronary artery, 30 pS (pipette 145 mM K⁺/bath 145 mM K⁺); K-channel in rat portal vein, 10 pS (pipette 6 mM K⁺/bath 138 mM K⁺)]. 2) We have not been able to record channel openings of K_GS in pig urethra in the absence of KCOs. However, both K_R and the 10-pS K channel obviously open even in the absence of nicorandil, with high open-state probabilities. 3) The channel opening of K_GS in pig urethra is quite different from that of K_R, which shows not only a semiconductance level but also long channel openings without channel flickering. 4) When the patch membrane is excised in ATP-free solution, "run-down" of the channel in pig urethra occurs rapidly, even in the presence of nicorandil. However, both K_R and the 10-pS K channel still clearly open, showing a high open-state probability even in inside-out configuration. 5) Intracellular application of 1 mM Mg-ATP reactivates K_GS in pig urethra. However, intracellular application of 5 mM Mg-ATP has little effect on the activation of the 10-pS K channel in rat portal vein, showing only the blocking effect due to intracellular Mg⁺⁺. 6) In inside-out patches, K_GS reactivates upon application of either Mg-ATP or NDPs to the intracellular surface of the patch membrane, showing burst-like channel opening even in the presence of 5 mM EGTA. These results suggest little intracellular Ca⁺⁺ sensitivity in this channel. However, both K_R and the 10-pS K channel have a potent intracellular Ca⁺⁺ sensitivity (the intracellular Ca⁺⁺ sensitivity of the 10-pS K channel is much higher than that of the large-conductance Ca⁺⁺-activated K⁺ channels in rat portal vein). These observations suggest that K_GS in pig urethra may be a quite different type of K⁺ channel from K_R and the 10 pS K-channel.

Does nicorandil directly activate K_GS in pig urethra? It is well known that in vascular smooth muscles, nicorandil causes relaxation with a dual mechanism. It not only causes a membrane hyperpolarization as a KCO but also acts as an NO releaser, increasing intracellular cyclic GMP levels via the stimulation of guanylate cyclase (reviewed by Taira, 1989). It is still unclear whether nicorandil directly activates K_ATP in smooth muscles or cyclic GMP modulates K_ATP after the activation of guanylate cyclase by nicorandil. Using the patch-clamp technique, Kubo et al. (1994) concluded that an increase in intracellular cyclic GMP directly modulates the gating of K_ATP in rat aorta. In the present single-channel experiments, methylene blue (10-100 µM) did not suppress the nicorandil-induced K_GS openings that were completely suppressed by 10 µM glibenclamide. On the other hand, neither SNP nor 8-bromo cyclic GMP induced channel openings in patches that contained the nicorandil-activated K⁺ channels. These results suggest that nicorandil directly activates K_GS in pig urethra without involving the guanylate cyclase pathway.

Nicorandil-induced relaxation in smooth muscles. In tension measurement, Yoneyama et al. (1990) proposed that nicorandil exerts its effects on the coronary circulation predominantly as a K_ATP opener. On the other hand, Kukovetz et al. (1991) reported that in bovine coronary artery, although lower concentrations of nicorandil (about 50 µM) cause relaxation predominantly through activation of K_ATP, higher concentrations lead to an increase in cyclic GMP through stimulation of guanylate cyclase. Miwa et al. (1993) concluded that the relative involvement of the two mechanisms of nicorandil-induced relaxation in porcine coronary artery depends on the location. Furthermore, Kreye et al. (1991) reported that in rabbit aorta, the relaxant action of nicorandil depends primarily on its nitrovasodilator-like properties. Therefore, even in vascular smooth muscles, it still remains uncertain which mechanism plays the major role in the nicorandil-induced relaxation. The potent relaxant effects of nicorandil on nonvascular smooth muscles in intact tissues have also been widely investigated (guinea pig trachealis, Allen et al. 1986; guinea pig taenia caeci, Weir and Weston, 1986; guinea pig ileum, Sun and Benishin, 1994; rat urinary bladder, Zhou et al., 1995) and indicate a possible clinical role for nicorandil in several diseases (Andersson, 1992; see Introduction). In the present tension experiments, both glibenclamide (1-10 µM) and methylene blue (10-100 µM) were used to investigate the relaxation mechanism in pig urethra. Application of either inhibitory compound caused a partial inhibition on the nicorandil-induced relaxation, which suggests that a dual nicorandil-induced relaxation mechanism may be present in pig proximal urethra as well as vascular smooth muscle. However, a component (about 30%) of the relaxation was insensitive to combined application of both blockers. Glibenclamide and methylene blue at these concentrations are thought to be specific inhibitors of K_GS and soluble guanylate cyclase, respectively (glibenclamide, Teramoto and Brading, 1996; methylene blue, Martin et al., 1985), so this may suggest that a third nicorandil-induced relaxation mechanism is present in pig urethra. Cornwell et al. (1994) suggested that higher concentration of NO-generating compounds (1 µM) produce cyclic GMP-independent actions in rat aortic vascular smooth muscle cells. Higher concentrations of NO (1 µM) itself have been reported to produce a number of cyclic GMP-independent effects in tissues (such as the inhibition of enzyme-containing ferrous-sulfhydryl groups, the production of peroxynitrite and the ADP-ribosylation of protein under certain conditions). Thus the cyclic GMP-specific effects may be associated with low concentration (<1 µM) of NO or NO-generating compounds (reviewed by Lincoln et al., 1996). Further studies are required to characterize the glibenclamide- and methylene blue-insensitive nicorandil-induced relaxing component in pig urethra.

The target K⁺ channel for KCOs in pig proximal urethra and clinical implication of KCOs in urology. To avoid the confusion about the designation of KCOs, we use the term exclusively to mean openers of K_ATP according to the definition of Quast (1993) in smooth muscles. KCOs were introduced into the field of urology in in vivo studies (Foster et al., 1989b; Malmgren et al., 1989; Hedlund et al., 1991). They can suppress unstable bladder contractions and are potentially useful drugs for the treatment of urge incontinence. Investigations of the in vitro effects of KCOs on guinea pig, pig and human urinary bladder (Foster et al., 1989a, b; Fujii et al., 1990) suggest the presence of K_ATP. Ideal KCOs for treatment of urge incontinence should relax the bladder but have little effect on either cardiovascular activity or urethral resting tone. KCOs that had additional relaxant effects on the resting urethral tone would be unsuitable for the treatment of patients with unstable bladders. Recently, Howe et al. (1995) reported that ZD6169, a newly synthesized K_ATP opener, has selectivity for bladder over the cardiovascular system in vivo in the anesthetized dog. However, the influence of the drug on the urethral resting tone and on urethral pressure was not mentioned. Because of similarities between the pig and human in micturition mechanisms (Melick et al., 1961; Crowe and Burnstock, 1989), and because of our group's use of an in vivo pig model, we selected the pig urethra as a source of tissue for investigation. That study revealed the presence of K_GS in the proximal urethra (Teramoto and Brading, 1996). We have now focused on the effects of nicorandil on the resting urethral tone and studied the characteristics of K_GS in this tissue. We have shown that both nicorandil and levcromakalim relax the urethra and activate K_GS, which suggests that the target K⁺ channel for KCOs in pig proximal urethra might be K_GS. These results could provide useful in the design of KCOs for use in urology. KCOs that relax urethral tone might be useful in the treatment of patients with bladder outflow obstruction, but urethral relaxation would be an undesirable side effect of KCOs designed to reduce the activity of the unstable bladder in the treatment of urge incontinence.