Glibenclamide, an ATP-Sensitive K+ Channel Blocker, Inhibits Cardiac cAMP-Activated Cl− Conductance
Abstract Stimulation of the β-adrenoceptor activates a time-independent Cl− conductance that is known to be regulated via phosphorylation by cAMP-dependent protein kinase in guinea pig ventricular myocytes. Since epithelial cystic fibrosis transmembrane conductance regulator Cl− channels are known to be sensitive to an antidiabetic sulfonylurea, glibenclamide, we tested whether the drug modulates cardiac cAMP-activated Cl− conductance. Bath application of isoproterenol (1 μmol/L, n=11) or forskolin (1 μmol/L, n=17) or the intracellular application of cAMP (1 mmol/L, n=9) activated whole-cell Cl− currents recorded from single myocytes at 36°C. External glibenclamide (≥10 μmol/L, n=26) inhibited the Cl− current induced by either of the stimulants in a concentration-dependent manner. The half-maximal inhibition concentration (IC50) of glibenclamide and the Hill coefficient were 24.5 to 37.9 μmol/L and 1.6 to 2.2, respectively. During current-clamp experiments, forskolin was found to shorten the action potential significantly (250±45 to 201±52 milliseconds, P<.05) in 7 of 11 cells tested. Glibenclamide antagonized the forskolin-induced shortening (to 243±54 milliseconds, n=7, P<.05). Intracellular administration of sodium orthovanadate (0.5 to ≈1 mmol/L, n=6) brought about persistent activation of Cl− current after brief bath application of forskolin. This Cl− current was not affected by H-89 (100 μmol/L, n=3), a specific inhibitor of cAMP-dependent protein kinase, and was suppressed by glibenclamide similarly, with an IC50 of 29.7 μmol/L. Thus, it is concluded that glibenclamide inhibits cardiac cAMP-activated Cl− channels at some step(s) downstream from the phosphorylation/dephosphorylation process.
Stimulation of the β-adrenoceptor leads to the activation of a time-independent Cl− channel current through a G protein/adenylate cyclase/cAMP-dependent protein kinase pathway in mammalian cardiac myocytes.1 2 3 4 5 6 7 8 The activation process requires ATP not only as a substrate for the kinase but also as an allosteric activator through binding to the Cl− channel.9 10 These properties of cardiac cAMP-activated Cl− channels are similar to those of CFTR Cl− channels.11 Indeed, Northern blot analysis using CFTR cDNA revealed that the CFTR Cl− channel is expressed in the heart.9 12 13
In the heart, Cl− channel blockers may exert actions of antiarrhythmic agents, such as Vaughan Williams class III,14 since the inhibition of the Cl− conductance should prevent the cellular depolarization and shortening of the action potential induced by β-adrenoceptor stimulation. The pharmacology of cardiac cAMP-activated Cl− channels is, however, less well established. The Cl− channel current was reported to be relatively insensitive to stilbene-derivative Cl− channel blockers, such as DNDS,3 15 16 17 DIDS (100 μmol/L),15 16 and SITS,16 at submillimolar concentrations, whereas DNDS and SITS were found to effectively diminish the Cl− current at 10 to 1000 μmol/L in an earlier report.2 DPC was reported to be effective in inhibiting the Cl− current at 1 mmol/L,17 although it was ineffective at 200 μmol/L.15 Other carboxylic acid analogues, such as 9-AC (100 to 500 μmol/L) and NPPB (50 μmol/L), were also found to strongly inhibit the Cl− current.5 17 18 However, Walsh17 has recently provided the data suggesting that these carboxylic acids inhibit, in a nonspecific manner, not only the Cl− channel but also the L-type Ca2+ channel.
Epithelial CFTR Cl− channels have recently been shown to be blocked by several compounds that modulate KATP channels, especially glibenclamide.19 CFTR Cl− channels share some properties with KATP channels. Both CFTR Cl− and KATP channels can be regulated by protein kinase A–dependent phosphorylation and intracellular ATP. Therefore, there is the possibility that cardiac cAMP-activated Cl− channels are also sensitive to glibenclamide.
In the present study, glibenclamide was, for the first time, found to inhibit the cardiac cAMP-activated Cl− channels in a concentration-dependent manner, thereby counteracting shortening of the APD induced by β-adrenoceptor stimulation.
Materials and Methods
Preparation of Single Myocytes
Ventricular myocytes were isolated from adult male guinea pigs under pentobarbital anesthesia (50 mg/kg IP), as previously described.6 While the animals were under artificial ventilation, the aorta was promptly cannulated, and the heart was quickly excised. By retrograde perfusion at 36°C, oxygenated normal Tyrode’s solution was applied for 5 minutes; subsequently, nominally Ca2+-free Tyrode’s solution was applied until contraction stopped. The latter solution supplemented with 0.1 mg/mL collagenase (Yakult) and 0.7 mg/mL bovine serum albumin (Nacalai Tesque) was then perfused for 10 minutes. Finally, the heart was perfused with KB solution20 at room temperature to rinse away the collagenase. The partially digested heart was gently minced with scissors in ice-cooled KB solution. After filtering through a stainless steel mesh (65 μm in diameter), the myocyte suspension was stored at 4°C in KB solution.
Whole-cell recordings were carried out by using a patch-clamp amplifier (model EPC-7, List) at 36°C. The pipettes were fabricated from borosilicate glass capillaries (Hilgenberg) by using a two-step puller (model PP-83, Narishige). Patch pipettes had resistances of 2 to 3 MΩ when filled with the control pipette solution. Data were acquired on-line by computer (model PC9801VX, NEC) through a Bessel-type filter at 3 kHz and recorded on videotape by means of an AD converter (model PCM-501ES, Sony) for backup.
Membrane current recordings were performed by using the conventional whole-cell configuration in the control bath solution. On the other hand, for current-clamp experiments, the modified nystatin–perforated patch technique21 was used. After perforating the patch membrane with the pipette solution containing 250 μg/mL nystatin (Sigma Chemical Co) and 1 mg/mL fluorescein sodium (Nacalai Tesque), action potentials were elicited by intracellular current injection at a frequency of 0.1 Hz.
Solutions and Drugs
The composition of normal Tyrode’s solution was as follows (mmol/L): NaCl 143, NaH2PO4 0.3, KCl 5.4, CaCl2 1.8, MgCl2 0.5, glucose 5.5, and HEPES 5 (pH adjusted to 7.4 with NaOH). The control bath solution contained (mmol/L) NaCl 150, MgCl2 0.5, CdCl2 1, glucose 5.5, and HEPES-NaOH 5 (pH 7.4). When necessary, the Cl− concentration was reduced to 21 mmol/L by replacing NaCl with sodium gluconate. KB solution contained (mmol/L) l-glutamic acid 70, KCl 25, taurine 20, KH2PO4 10, MgCl2 3, EGTA 0.5 (Nacalai Tesque), glucose 10, and HEPES-KOH 10 (pH 7.4). The control pipette solution contained (mmol/L) aspartic acid 85, EGTA 10, tetraethylammonium chloride 20, Na2–creatine phosphate 5, MgATP 10, MgCl2 0.5, glucose 5.5, and HEPES-CsOH 10 (pH 7.4). During β-adrenoceptor stimulation, Na2-GTP (200 μmol/L) was added to the control pipette solution to minimize fade of the Cl− conductance.7 In some experiments, 0.5 or 1 mmol/L sodium orthovanadate (Sigma), an inorganic phosphate analogue, was added to the control pipette solution.
For the voltage-clamp experiments under the whole-cell conditions, K+ currents were eliminated by internal tetraethylammonium (20 mmol/L) and by omission of K+ from both pipette and bath solutions; Na+ and Ca2+ currents, by inactivating at −10 mV; any residual Ca2+ currents, by extracellular Cd2+ (1 mmol/L); Na+-K+ pump currents, by removal of external K+; and Na+-Ca2+ exchange currents, by the nominal absence of internal and external Ca2+.
For the current-clamp experiments, a K+-rich pipette solution containing (mmol/L) potassium aspartate 110, KCl 20, K2-ATP 5, Na2–creatine phosphate 5, MgCl2 5, EGTA 5, and HEPES-KOH 5 (pH 7.4) was used.
The following agents were added to bath solutions: 1 μmol/L isoproterenol (Nacalai Tesque), 1 μmol/L forskolin (Nippon Kayaku), 100 μmol/L H-89 (Seikagaku Corp), and 1 to 200 μmol/L glibenclamide (Hoechst). Stock solutions of isoproterenol (1 mmol/L in distilled water), forskolin (10 mmol/L in ethanol), H-89 (50 mmol/L in DMSO), and glibenclamide (100 mmol/L in DMSO) were diluted to the desired final concentrations immediately before use. The final concentration of DMSO was ≤0.2%. DMSO alone (≤1%) did not affect the cardiac cAMP-activated Cl− conductance. cAMP (Sigma) was dissolved in the standard pipette solution (10 mmol/L, stock solution) and diluted to 1 mmol/L immediately before use.
Statistical differences of the APD and RMP were evaluated by paired Student’s t test and considered significant at P<.05. Values are given as mean±SD in the text.
Fig 1A⇓ shows whole-cell currents recorded from a myocyte held at −10 mV. β-Adrenoceptor stimulation with isoproterenol (1 μmol/L) produced outward currents with considerable increases in current noise. Reduction of the external Cl− concentration from 153 to 21 mmol/L produced a prompt decrease of the outward current. The outward current level was restored after returning to the control bath solution (153 mmol/L Cl−). The I-V relation under the transmembrane Cl− gradient (intracellular, 21 mmol/L; extracellular, 153 mmol/L) showed slight outward rectification with a reversal potential of −41 mV (Fig 1B⇓, a), which is fairly close to the estimated Cl− equilibrium potential (ECl, −53 mV). The reversal potential was shifted in a positive direction (to −13 mV; Fig 1B⇓, b) by reducing the extracellular Cl− concentration (to 21 mmol/L). These observations indicate that isoproterenol activates the current mainly carried by Cl−, as previously reported.1 2 3 4 5 6 7 8
Glibenclamide suppressed isoproterenol-induced outward currents in a concentration-dependent manner (Fig 1A⇑). The compound produced virtually complete inhibition of the isoproterenol-induced current at 100 μmol/L. It is notable that the current noise was also largely reduced by glibenclamide. The effect of glibenclamide was slow in onset, and the outward current rarely recovered after washout of glibenclamide especially at higher concentrations.
The glibenclamide effect was evident over the entire range of membrane potential examined. Glibenclamide-sensitive currents (Fig 1B⇑, a-c) also showed slight outward rectification and reversed at a similar potential (−43 mV). Thus, it appears that the Cl− conductance activated by isoproterenol is sensitive to glibenclamide.
Direct stimulation of adenylate cyclase by forskolin (1 μmol/L), a diterpene alkaloid, similarly activated the Cl− conductance1 22 (Fig 2A⇓). The forskolin-induced outward current was also blocked by glibenclamide (100 μmol/L). Fig 2B⇓ shows whole-cell current responses to voltage steps. Both forskolin-induced (Fig 2B⇓, b-a) and glibenclamide-sensitive currents (Fig 2B⇓, b-c) were time independent. Both I-V curves exhibited slight outward rectification and the same reversal potential (−35 mV, Fig 2C⇓). These results indicate that the forskolin-activated Cl− conductance is also sensitive to glibenclamide and that the site of glibenclamide action is located downstream from cAMP production by adenylate cyclase.
Activation of outward current (Fig 3A⇓) with a similar I-V relation (Fig 3B⇓) was obtained when internal dialysis of myocytes with cAMP (1 mmol/L) was begun after rupture of a cell-attached patch. The I-V curve showed slight outward rectification with a reversal potential of −36 mV (Fig 3B⇓, a). Glibenclamide produced a concentration-dependent decrease in the outward current (Fig 3A⇓). The I-V relation for the glibenclamide-sensitive current again showed slight outward rectification (Fig 3B⇓, a-b) and had a reversal potential of −37 mV. Thus, it is evident that glibenclamide inhibits the cAMP-induced Cl− conductance.
As shown in Fig 4⇓, essentially similar concentration-inhibition relations were obtained from the pooled data of glibenclamide effects on Cl− currents induced by isoproterenol (1 μmol/L, 4A), forskolin (1 μmol/L, 4B), and intrapipette cAMP (1 mmol/L, 4C). The IC50 values for isoproterenol, forskolin, and cAMP were 32.0, 24.5, and 37.9 μmol/L, respectively, and the Hill coefficients were 2.2, 1.9, and 1.6, respectively.
The Cl− current activated by forskolin (1 μmol/L) was promptly reduced after washout of forskolin under control conditions (Fig 5A⇓). H-89 (100 μmol/L), a specific inhibitor of cAMP-dependent protein kinase,23 virtually abolished the forskolin-induced Cl− current (Fig 5B⇓). In contrast, in the presence of 0.5 mmol/L sodium orthovanadate in the pipette, brief application of forskolin (1 μmol/L) produced persistent activation of the Cl− current even after washout of forskolin (Fig 5C⇓). Neither isoproterenol nor acetylcholine affected this persistently activated Cl− current (data not shown). H-89 was also without effect on the Cl− current in the presence of 1 mmol/L sodium orthovanadate (Fig 5D⇓). The current noise during the channel activation in the presence of sodium orthovanadate (Fig 5C⇓ and 5D⇓) was considerably smaller than that in the absence of sodium orthovanadate and even comparable to the basal current (without forskolin) (Fig 5A⇓ and 5B⇓), suggesting that the channel activation is independent of phosphorylation/dephosphorylation. Subsequent exposure to glibenclamide (100 μmol/L) produced concentration-dependent inhibition of the Cl− current (Fig 5C⇓) with IC50 of 29.7 μmol/L (Fig 5E⇓). Glibenclamide again caused a rapid increase in noise during partial inhibition of the current by 30 μmol/L.
To investigate how the cardiac action potential is modulated by the cAMP-activated Cl− current, current-clamp experiments were conducted by using a nystatin–perforated patch method. Fig 6A⇓ shows the example of the effects of forskolin (1 μmol/L) and glibenclamide (100 μmol/L) on action potentials. APD was shortened by forskolin (Fig 6A⇓, b) and thereafter restored by glibenclamide (Fig 6A⇓, c). As shown in Fig 6B⇓ and 6C⇓, forskolin significantly shortened APD90 in 7 of 11 cells from 250.1±44.6 to 200.8±51.6 milliseconds (P<.05) and induced a slight depolarization of the RMP (−74.3±2.2 to −72.3±2.6 mV, P<.05). In all 7 cells, subsequent application of glibenclamide was found to prolong APD90 (to 243.0±53.8 milliseconds, P<.05) and recover, in part, the RMP (to −74.0±2.6 mV, P<.05). Forskolin increased APD90 in two cells (238.0 to 260.5 and 240.5 to 268.5 milliseconds) and had no effect on APD in two other cells. In these four cells, glibenclamide added subsequently tended to prolong their APDs (data not shown).
Glibenclamide-Induced Blockade of Cardiac cAMP-Activated Cl− Currents
The major findings of the present study are as follows: (1) Glibenclamide is a potent inhibitor of the cardiac cAMP-activated Cl− conductance. (2) Its inhibition site is located downstream from the phosphorylation/dephosphorylation steps for the channel activation.
Since their first characterization in the guinea pig heart,1 2 3 cardiac cAMP-activated Cl− channel currents have been extensively studied, particularly with respect to the molecular mechanism underlying their modulation.3 4 5 6 7 8 9 On the other hand, consistent information has not been obtained regarding the pharmacology.2 3 5 15 16 17 18 The present study showed that glibenclamide is the most potent inhibitor of the cardiac cAMP-activated Cl− conductance. The effects of glibenclamide developed slowly and were rarely washed out, as reported in the epithelial CFTR Cl− channel currents,19 presumably because of the lipophilic nature of the compound.
In the present study, glibenclamide was found to inhibit the isoproterenol-, forskolin-, and cAMP-activated Cl− currents with similar IC50 values. Glibenclamide, in virtually the same concentration range, also inhibited the Cl− current that had been persistently activated by forskolin and intracellular sodium orthovanadate and caused a considerable noise during partial inhibition of the current. As recently demonstrated by using excised giant patch membranes of ventricular cells,24 it is likely that the inorganic phosphate analogue, sodium orthovanadate, can lock Cl− channels in the open state, which is independent of both phosphorylation and dephosphorylation, by directly interacting with the nucleotide-binding domains. Therefore, glibenclamide inhibits the Cl− channel at some step(s) downstream from the phosphorylation/dephosphorylation process.
Cardiac cAMP-Activated Cl− Channels and CFTR Cl− Channels
CFTR has been shown to be an epithelial Cl− channel, the function of which is impaired in patients with cystic fibrosis.25 It is now known that the molecular structure of the cardiac cAMP-activated Cl− channel resembles, at least in part, that of the epithelial CFTR Cl− channel.12 13 Both channels exhibit, in common, relatively small single-channel conductances (≈13 picosiemens),4 9 11 activation by cAMP-dependent protein kinase–mediated phosphorylation,2 11 time-independent kinetics,1 2 25 and the anion selectivity sequence (Br−>Cl−>I−>F−).26 27 Recently, it was shown that the epithelial CFTR Cl− channel is blocked by glibenclamide, with an IC50 value of 21.8 μmol/L.19 The present study showed that cardiac cAMP-activated Cl− currents are blocked by glibenclamide, with a similar IC50 value, and therefore provides additional evidence supporting the idea that cardiac cAMP-activated Cl− channels are an isoform of epithelial CFTR Cl− channels.
However, there is a difference between the property of the cardiac Cl− channel and that of the epithelial CFTR Cl− channel. The Hill coefficient for CFTR Cl− currents was reported to be 0.8 for the glibenclamide action,19 whereas that for the cardiac Cl− conductance was 1.6 to 2.2 (Figs 4⇑ and 5E⇓), which is closer to the Hill coefficient (1.3) for the effect of glibenclamide on cardiac KATP channels.28
Physiological Implication of Activation and Blockade of Cardiac cAMP-Activated Cl− Currents
Activation of the Cl− channels modulates cardiac action potentials under autonomic nervous control1 5 18 29 30 : at membrane potentials negative to ECl it produces a depolarizing inward current, whereas at plateau-phase potentials it produces a hyperpolarizing outward current, thereby accelerating repolarization. In the present study, activation of the Cl− current by forskolin did indeed bring about APD shortening as well as a decrease of RMP (Fig 6⇑) in most cells tested. Glibenclamide reversed the forskolin-induced APD shortening (Fig 6⇑). In this experimental condition, ICa and IK were supposed to be enhanced by forskolin.18 However, in the forskolin-stimulated myocytes, glibenclamide at 100 μmol/L was found to be without effect on ICa (n=4) and have only a small inhibiting effect on IK (by 12.0±12.9%, n=6; M. Tominaga, unpublished data, 1995).
Since the possibility has been raised that the Cl− current activation has an arrhythmogenic nature5 30 in the guinea pig heart, the glibenclamide sensitivity of the cardiac cAMP-activated Cl− channel may be of value for the design and synthesis of new types of antiarrhythmic agents.
Selected Abbreviations and Acronyms
|APD||=||action potential duration|
|APD90||=||APD at 90% repolarization|
|CFTR||=||cystic fibrosis transmembrane conductance regulator|
|DPC||=||diphenyl carboxylic acid|
|ICa||=||L-type Ca2+ current|
|IK||=||delayed rectifier K+ current|
|KATP channels||=||ATP-sensitive K+ channels|
|RMP||=||resting membrane potential|
The authors are grateful to Dr A.F. James and Dr S. Oiki for critical reading of the manuscript.
Reprint requests to Makoto Tominaga, MD, Department of Cellular and Molecular Physiology, National Institute for Physiological Sciences, Okazaki 444, Japan.
Presented as preliminary data in abstract form (Circulation. 1992;86[suppl I]:I-695).
- Received February 21, 1995.
- Accepted April 20, 1995.
- © 1995 American Heart Association, Inc.
Harvey RD, Hume JR. Autonomic regulation of a chloride current in heart. Science. 1989;244:983-985.
Harvey RD, Clark CD, Hume JR. Chloride current in mammalian cardiac myocytes: novel mechanism for autonomic regulation of action potential duration and resting membrane potential. J Gen Physiol. 1990;95:1077-1102.
Hwang T-C, Horie M, Nairn AC, Gadsby DC. Role of GTP-binding proteins in the regulation of mammalian cardiac chloride conductance. J Gen Physiol. 1992;99:465-489.
Hwang T-C, Horie M, Gadsby DC. Functionally distinct phospho-forms underlie incremental activation of protein kinase-regulated Cl− conductance in mammalian heart. J Gen Physiol. 1993;101:629-650.
Hwang T-C, Nagel G, Nairn AC, Gadsby DC. Regulation of the gating of CFTR Cl− channels by phosphorylation and ATP hydrolysis. Proc Natl Acad Sci U S A. 1994;91:4698-4702.
Levesque PC, Hart PJ, Hume JR, Kenyon JL, Horowitz B. Expression of cystic fibrosis transmembrane regulator Cl− channels in heart. Circ Res. 1992;71:1002-1007.
Horowitz B, Tsung SS, Hart P, Levesque PC, Hume JR. Alternative splicing of CFTR Cl− channels in heart. Am J Physiol. 1993;264:H2214-H2220.
Hwang T-C, Horie M, Dousmanis AG, Gadsby DC. Regulation of PKA-activated Cl conductance in guinea pig ventricular myocytes: whole-cell studies. J Gen Physiol. 1992;100:69a. Abstract.
Walsh KB. Agents which block heart Cl− channels also inhibit Ca2+ current regulation. Biophys J. 1994;66:A239. Abstract.
Sheppard DN, Welsh MJ. Effect of ATP-sensitive K+ channel regulators on cystic fibrosis transmembrane conductance regulator chloride currents. J Gen Physiol. 1992;100:573-591.
Horie M, Hwang T-C, Gadsby DC. Forskolin activation of PKA-regulated chloride current in guinea pig ventricular myocytes. Biophys J. 1992;61:396a. Abstract.
Chijiwa T, Hagiwara M, Hidaka H. Inhibition of forskolin-induced neurite outgrowth and protein phosphorylation by a newly synthesized selective inhibitor of cyclic AMP-dependent protein kinase, N-[2-(p-bromocunnamylamino) ethyl]-5-isoquinoline sulfonamide (H-89), of PC12D pheochromocytoma cells. J Biol Chem. 1990;265:5267-5272.
Anderson MP, Rich DP, Gregory RJ, Smith AE, Welsh MJ. Generation of cAMP-activated chloride currents by expression of CFTR. Science. 1991;251:679-682.
Overholt JL, Hobert ME, Harvey RD. On the mechanism of rectification of the isoproterenol-activated chloride current in guinea-pig ventricular myocytes. J Gen Physiol. 1993;102:871-895.
Anderson MP, Gregory RJ, Thompson S, Souza DW, Paul S, Mulligan RC, Smith AE, Welsh MJ. Demonstration that CFTR is chloride channel by alteration of its anion selectivity. Science. 1991;253:202-205.
Findlay I. Inhibition of ATP-sensitive K+ channels in cardiac muscle by the sulphonylurea drug glibenclamide. J Pharmacol Exp Ther. 1992;261:540-545.