Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

K_ATPs are found in various types of cells, as for example the -cells of the endocrine pancreas, the smooth muscle cells of the vascular system, trachea and urinary tract, in various regions of the brain, in skeletal muscle, and in heart muscle cells (Ashcroft, 1988). In pancreatic -cells K_ATPs contribute to the cell's resting potential and are a critical link between blood glucose concentration and insulin secretion. In heart muscle cells, the physiological role of K_ATPs is still unclear. It is known that the channels are closed in normally functioning cardiomyocytes, but open when the intracellular ATP concentration falls below a critical value. Activation of K_ATPs in heart muscle cells was directly observed following anoxia (Vleugels et al., 1980), or metabolic inhibition (Noma, 1983). As the K_ATPs show a nearly linear behavior in the voltage range of an action potential, they are contributing to K⁺ efflux and subsequent interstitial K⁺ accumulation in ischemic regions (Hill and Gettes, 1980). In addition, opening of K_ATPs leads to shortening of the action potential duration (Wilde et al., 1990). Taken together, these effects may induce reentry arrhythmias and consequently cause sudden cardiac death (Janse and Wit, 1989). By means of sulfonylureas, which are potent inhibitors for K_ATPs, the contribution of these channels to ventricular fibrillation was directly demonstrated in a number of animal models, like isolated perfused rat and guinea pig hearts (Wolleben et al., 1989), (Kantor et al., 1990), (Gwilt et al., 1992; Tosaki and Hellegouarch, 1994), isolated perfused rabbit hearts (Chi et al., 1993; Bellemin-Baurreau et al., 1994) and postinfarcted conscious dogs (Billman et al., 1993). Moreover, there is evidence that the sulfonylurea glibenclamide reduces the incidence of ventricular tachycardia and ventricular fibrillation in diabetic patients with acute coronary artery diseases (Cacciapuoti et al., 1991; Lomuscio and Fiorentini, 1996; Davis et al., 1996). Thus, blockade of K_ATPs is regarded as a novel approach for prevention of ventricular fibrillation occurring in the course of ischemic events. However, sulfonylureas like glibenclamide cannot be used clinically as antifibrillatory agents because of their potent blood glucose lowering effects. This effect of glibenclamide on insulin release as well as its documented vasoconstrictive effect on coronary blood flow (Daut et al., 1990) exclude its use as an antiarrhythmic drug. As it was reported that pharmacological differences exist between K_ATPs in pancreatic -cells and heart muscle cells (Faivre and Findlay, 1989), our research focused on the development of novel sulfonylurea compounds with improved selectivity for cardiac muscle K_ATPs. In the present study we compare HMR 1883 and glibenclamide with respect to their effects on K_ATPs in cardiac and pancreatic preparations.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Experiments with papillary muscles. Guinea pigs of either sex (Marioth, animal breeding Hoechst, Frankfurt/Main, Germany), weighing 300 to 500 g, were killed by cervical dislocation and exsanguination. The hearts were rapidly removed and the right or left papillary muscles were excised and mounted in an organ bath which contained the following bathing solution (in mM): 136 NaCl, 3.3 KCl, 2.5 CaCl₂, 1.2 KH₂PO₄, 1.1 MgSO₄, 5 glucose, 10 HEPES, pH adjusted to 7.4 with NaOH, gassed with 100% O₂. In experiments where the pH of the solution was adjusted to 6.0 or 6.5, MES or PIPES, respectively, was used instead of HEPES buffer.

Experiments with the channel opener rilmakalim. After a 30 min equilibration period, the channel opener rilmakalim (Linz et al., 1992), was applied. The sulfonylurea was added 30 min after the channel opener and the APD₉₀ was recorded after an additional 60 min.

Experiments with hypoxia. Two types of experiments were performed: (1) after a 30 min equilibration period hypoxia was induced by gassing the bathing solution with 100% N₂ and omission of glucose. In addition, the pH-value of the hypoxic solution was adjusted to 6.5 (PIPES instead of HEPES buffer). After 60 min hypoxia, the sulfonylureas were added and the APD₉₀ was recorded after an additional 60 min. (2) After a 30 min equilibration period hypoxia was induced, while the sulfonylureas were added to the hypoxic solution. The APD₉₀ was recorded after an additional 60 min.

Experiments with DOG. After an equilibration period of 30 min, glucose was replaced by DOG, and gassing was stopped. The substance to be tested was present in the DOG-solution and the APD₉₀ was recorded 60 min after induction of metabolic inhibition.

Experiments with Langendorff-Perfused Hearts

Investigation of ventricular fibrillation. Hearts of guinea pigs were quickly removed, cannulated via the aorta, and perfused at a constant flow of 1.2 ml/min by means of a roller pump with the following solution (in mM): 154 NaCl, 6.4 KCl, 1.6 CaCl₂, 2.4 NaHCO₃, 5 glucose, 30 mg/liter pinacidil, gassed with 95% O₂ and 5% CO₂. The temperature of the perfusion medium and the temperature of the isolated heart were maintained at 37°C. The electrogram was recorded by placing two Ag/AgCl electrodes on the surface of the heart. The signal was amplified (Hellige, Freiburg, Germany) and displayed on a chart recorder.

Experimental protocol. After cannulation, the heart was immediately perfused at low flow (1.2 ml/min) for 20 min, then the flow was increased to normal flow (7 ml/min) for 10 min. Pinacidil (30 mg/liter was always present in the perfusion solution. Ventricular fibrillation occurring within this period (20 min low flow plus 10 min normal flow) were counted. Each group of experiments consisted of 10 hearts. The drug to be tested was present from the beginning to the end of the experiment. For each drug concentration of glibenclamide and HMR 1883 a separate group of control experiments was performed.

Investigation of the coronary flow. In this set of experiments a latex balloon was inserted into the left ventricle and the coronary flow was recorded by a blood flow transducer (Hellige type E, Freiburg). The heart was immersed in the perfusion solution and was perfused with a constant pressure of 55 mmHg. The perfusion solution consisted of (in mM): 118 NaCl, 4.7 KCl, 2.5 CaCl₂, 25 NaHCO₃, 1.2 MgSO₄, 1.2 NaH₂PO₄, 5 glucose, 2 pyruvate, gassed with 95% O₂, 5% CO₂. Hypoxia was induced by gassing the perfusion solution with 20% O₂, 75% N₂, 5% CO₂.

Patch-Clamp Experiments

Isolation of ventricular myocytes. Isolated cells from ventricular muscle were prepared as described previously (Krause et al., 1995). Briefly, guinea pigs (Marioth, weight about 400 g) of either sex were sacrificed by cervical dislocation. Hearts were dissected and retrogradely perfused via the aorta at 37°C: first for 3 min with Tyrode solution (in mM): 143 NaCl, 5.4 KCl, 1.8 CaCl₂, 0.5 MgCl₂, 0.25 NaH₂PO₄, 5 HEPES, pH adjusted to 7.4 with NaOH; second with nominally Ca⁺⁺ free Tyrode solution (no Ca⁺⁺ added) for 5-7 min, and finally with nominally Ca⁺⁺ free Tyrode solution containing 0.2 mg/ml collagenase type 1 (Sigma, Deisenhofen, Germany). After 15-20 min of collagenase treatment, the heart was washed with storage solution (in mM): 70 KOH, 50 L-glutamic acid, 40 KCl, 20 taurine, 20 KH₂PO₄, 3 MgCl₂, 10 glucose, 10 HEPES, 0.5 EGTA, pH adjusted to 7.4 with KOH. The ventricles were cut into small pieces and myocytes were dispersed by gentle shaking and finally by filtration through a nylon mesh (365 µm). Thereafter, the cells were washed twice by centrifugation at 90 × g for 5 min and kept in storage solution at room temperature.

The patch-clamp method. Whole-cell currents were recorded in the tight-seal whole-cell mode of the patch-clamp technique (Hamill et al., 1981). In addition, cell potentials were recorded in the whole-cell mode when the amplifier (EPC-7, List, Darmstadt, Germany or EPC-9, HEKA, Lambrecht, Germany) was switched to the current-clamp mode (clamp current = 0 pA). Patch pipettes were pulled from borosilicate glass capillaries with 0.3 mm wall thickness and 1.5 mm outer diameter, and their tips were heat polished. The series resistance was in the range of 2 to 10 M and was compensated by 50% by means of the EPC'sec compensation circuit. The cell capacitance was calculated from the current response after applying a voltage pulse from the holding potential (80 mV) to 70 mV. The capacitance of the cells we studied was about 170 pF. Current-voltage (I/V) relations were measured by applying voltage ramps from 140 mV to + 50 mV. The current recorded at 0 mV clamp potential was evaluated and displayed in the figures. At this voltage most of the time- and voltage-dependent currents are inactive. Action potentials were recorded in the current-clamp mode (whole-cell current clamped to 0 nA), and were evoked by applying brief (1 msec) pulses via the patch electrode to the cell. The inward-rectifying current I_K1 was measured at 100 mV clamp potential, when voltage-ramps from 150 mV to +50 mV were applied within 500 msec. The fast and slow component of the delayed outward current (I_Kr and I_Ks, respectively) were recorded by applying voltage pulses of 3 sec duration from the holding potential of 40 mV to either 10 mV (for I_Kr) or +50 mV (for I_Ks). In order to block the L-type Ca⁺⁺ channels, nifedipine (10 µM) was present in the experiments where I_Kr and I_Ks were recorded.

Solutions

The pipette solution contained (in mM): 140 KCl, 10 NaCl, 1.1 MgCl₂, 1 EGTA, 1 Mg-ATP, 10 HEPES, pH = 7.2, adjusted with KOH, and the bathing solution was (in mM): 140 NaCl, 4.7 KCl, 1.1 MgCl₂, 1.3 CaCl₂, 10 glucose, 10 HEPES, pH = 7.4, adjusted with NaOH.

Sign convention. Outward movement of positive ions (from the cell to the extracellular side) are depicted as positive currents. The sign of the electrical potential refers to the cytosolic side with respect to the grounded extracellular side.

Temperature. Experiments were performed at 34 ± 1°C.

Patch-clamp experiments with RINm5F cells. RINm5F cells were maintained in RPMI 1640 tissue culture media, containing 11 mM glucose, supplemented with 10% (vol/vol) fetal calf serum, 2 mM glutamine and 50 µg/ml gentamycine. Cells were seeded out every two to three days onto petri dishes and kept in a humidified atmosphere of 95% O₂ and 5% CO₂ at a temperature of 37°C. For patch-clamp experiments, cells were isolated by incubation in a Ca⁺⁺-free medium containing 0.25% trypsin for about 3 min. Single cells and clusters of 2-3 cells were obtained after centrifugation with 800 rpm and were stored on ice until use. The tight-seal whole-cell patch-clamp technique was applied to single cells. The pipette solution was (in mM): 140 KCl, 10 NaCl, 1.1 MgCl₂, 0.5 EGTA, 1 Mg-ATP, 10 HEPES, pH = 7.2, adjusted with KOH, and the bathing solution was (in mM): 140 NaCl, 4.7 KCl, 1.1 MgCl₂, 2 CaCl₂, 10 HEPES, pH = 7.4, adjusted with NaOH.

Statistics

All averaged data are presented as the mean ± SEM. The Student's t-test was used to determine the significance of paired or unpaired observations. Differences were considered significant at P < .05.

The values for half-maximal inhibition (IC₅₀) and the Hill coefficient were calculated by fitting the data points of the concentration/response curves to the logistic function: where a represents the plateau-value at low drug concentration and d the plateau-value at high drug concentration; c represents the IC₅₀ value and n the Hill-coefficient.

The curve-fitting, the Student's t-test, and the Fisher Exact test were performed with the computer program Sigma-Plot.

Materials

Glibenclamide and HMR 1883 (fig. 1) were produced at Hoechst Marion Roussel Frankfurt, Research Chemistry. HMR 1883 differs from glibenclamide in the following points: The sulfonylurea moiety is shifted from the para to the meta position; a methoxy group is attached to the para position; the urea moiety is modified to thiourea; and finally the cyclohexyl moiety attached to the terminal nitrogen is replaced by a methyl group.

View larger version (16K): [in this window] [in a new window]   Fig. 1.   Chemical structures of glibenclamide and HMR 1883.

In patch-clamp experiments, stock solutions of glibenclamide and HMR 1883 in DMSO (100 µM) were produced and further diluted into the respective bathing solution. In experiments with papillary muscles and isolated perfused hearts the compounds were dissolved in 1,2-propanediol and were further diluted into the respective bathing solution. The concentration of propanediol was maximally 0.5%.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Experiments with Guinea Pig Papillary Muscles

Effects under control conditions. In left papillary muscles possible effects of glibenclamide and HMR 1883 on the action potential recorded under control conditions were investigated. The action of the drugs was recorded 60 min after application. Neither with 2 and 20 µM glibenclamide, nor with 2 and 20 µM HMR 1883 a significant effect could be observed on the parameters APD₉₀, resting membrane potential, amplitude of the action potential, upstroke velocity, and refractory period. Moreover, when the Ca⁺⁺-dependent (slow) action potential was recorded, no significant effects of either 20 µM glibenclamide or 20 µM HMR 1883 were detected on the cell's resting potential, the amplitude of the action potential, the upstroke velocity, and the action potential duration, indicating that the drugs have no effect on Ca⁺⁺ channels (data not shown).

Effects on rilmakalim-induced shortening of the APD. In guinea pig right papillary muscles, the effects of the drugs on the rilmakalim-induced shortening of the action potential duration was investigated. As demonstrated in figure 2a, rilmakalim (1 µg/ml) induced a pronounced shortening of the APD, which was associated with a decrease in the amplitude of the action potential. 30 min after application of rilmakalim the sulfonylureas were added to the bath in the presence of rilmakalim. The pH value of the bathing solution was kept at 7.4 throughout the experiments. Both substances prolonged the APD in a concentration dependent manner. As a typical example, figure 2a demonstrates that 2 µM HMR 1883 caused a pronounced prolongation of the APD in the presence of rilmakalim. Figure 2b shows the fitted mean values, yielding half maximal inhibition with 0.33 µM glibenclamide and 1.8 µM HMR 1883. When the inhibitors were applied together with rilmakalim, the APD shortening was inhibited with similar potency as described above. The cell's resting potential did not change significantly after application of rilmakalim and after subsequent addition of either glibenclamide or HMR 1883 (data not shown).

View larger version (21K): [in this window] [in a new window]   Fig. 2.   Inhibition of the rilmakalim (2.4 µM)-induced shortening of the action potential duration in guinea pig right papillary muscles at extracellular pH 7.4 a, Original recordings of action potentials under control conditions, 30 min after application of 1 µg/ml rilmakalim, and after additional application of HMR 1883 (2 µM). The recordings were obtained from the same papillary muscle. The shortening of the APD90 recorded 30 min after addition of rilmakalim is set to 100%. b, Dose-response curves for inhibition of the rilmakalim-induced shortening of the APD90 by glibenclamide and HMR 1883. The substances were added 30 min after application of rilmakalim. The recordings were performed 60 min after application of the sulfonylureas. The number of experiments was 5 for each data point.

View larger version (13K): [in this window] [in a new window]   Fig. 3.   Dose-response curves for inhibition of the rilmakalim (7.3 µM)-induced shortening of the APD90 by glibenclamide and HMR 1883 at an extracellular pH of 6.0 in guinea pig right papillary muscles. The substances were added 30 min after application of rilmakalim. The recordings were performed 60 min after application of the sulfonylureas. The shortening of the APD90 recorded 30 min after addition of rilmakalim is set to 100%. The number of experiments was 5 for each data point.

Effects on hypoxia-induced shortening of the APD. In order to produce a more physiological shortening of the action potential duration, guinea pig right papillary muscles were exposed to hypoxic solution, which was free of glucose and oxygen and had a pH of 6.5. After 60 and 120 min of hypoxia, the APD₉₀ decreased from 177 ± 9 msec to 62 ± 15 msec and 38 ± 10 msec (n = 4), respectively. This time-course of hypoxia-induced APD₉₀ shortening corresponds with previously published observations (Hayashi et al., 1997). The cell's resting potential did not change significantly during hypoxia, and after subsequent addition of glibenclamide or HMR 1883.

View larger version (20K): [in this window] [in a new window]   Fig. 4.   Original recordings demonstrating the effect of HMR 1883 on the hypoxia-induced shortening of the action potential duration in a guinea pig right papillary muscle at an extracellular pH of 6.5. The substances were added 60 min after induction of hypoxia. a, Effect of 2 µM HMR 1883. b, Effect of 20 µM HMR 1883.

View larger version (35K): [in this window] [in a new window]   Fig. 5.   Statistical mean values of the hypoxia-induced shortening of the action potential duration in guinea pig right papillary muscle in the presence of glibenclamide (a) and HMR 1883 (b) The drugs were added 60 min after hypoxia was induced. The extracellular pH was 6.5. The number of experiments is given in brackets. Statistically significant difference to control values is denoted by asterisks.

View larger version (20K): [in this window] [in a new window]   Fig. 6.   Effects of 2 µM glibenclamide (a) or 20 µM HMR 1883 (b) on the contractile force under hypoxia in guinea pig right papillary muscles. The extracellular pH was 6.5. The asterisks denote statistically significant differences with respect to the control values.

Effects on deoxyglucose-induced shortening of the APD. As demonstrated in figure 7, metabolic inhibition, induced by replacing all glucose by deoxyglucose, caused a shortening of the APD₉₀ by 123 ± 5 msec (n = 27). This APD₉₀ shortening was markedly reduced by 2 µM glibenclamide, but no additional effects could be observed with 20 µM and even with 100 µM of the drug (fig. 7a). The effect of HMR 1883 was less pronounced: 2 µM had only a slight, but statistically significant effect, whereas 20 µM reduced the APD₉₀ shortening to 55 ± 7 msec (n = 12). Increasing the HMR 1883 concentration to 100 µM did not produce a further effect on the APD₉₀ shortening (fig. 7b).

View larger version (40K): [in this window] [in a new window]   Fig. 7.   Effects of glibenclamide (a) and HMR 1883 (b) on the deoxyglucose-induced shortening of the action potential duration in guinea pig left papillary muscles. The substances were applied simultaneously with the DOG solution. The extracellular pH was 6.8. The APD90 were recorded 60 min after bath exchange with DOG + drug solution. The number of experiments is given in brackets. The asterisks denote statistically significant differences with respect to the control values.

Experiments with Langendorff-perfused hearts. When subjected to low coronary flow in the presence of pinacidil, 8 to 10 of a total of 10 hearts fibrillated (denoted as "control" in table 1. In the presence of 10 nM of either glibenclamide or HMR 1883 the number of fibrillating hearts was not statistically significantly different from the respective control group. However, as shown in table 1 only 5 hearts fibrillated in the presence of 100 nM HMR 1883 (P = .016), whereas non of the glibenclamide (100 nM) treated hearts fibrillated. In the presence of 1 µM HMR 1883 none of 10 hearts fibrillated.

Patch-Clamp Experiments

Effects on I_K1, I_Kr and I_Ks. In 6 experiments, I_K1 was 85 ± 10% of control after application of 100 µM HMR 1883. This difference was statistically not significant. Both the I_Kr and I_Ks showed a time dependent run down: 15 min after establishing the whole-cell configuration, the I_Kr declined to 86 ± 2% (n = 9) and the I_Ks to 88 ± 4% (n = 9) of control. When 100 µM HMR 1883 was present, I_Kr was 87 ± 2% (n = 6) and I_Ks was 84 ± 6% (n = 4) of control, when recorded 15 min after obtaining the whole cell mode. Thus, 100 µM HMR 1883 had no significant effects on the I_Kr and I_Ks current.

Effects on rilmakalim-activated whole-cell current and APD₉₀. In isolated guinea pig cardiomyocytes the whole-cell currents were recorded by applying voltage ramps to isolated guinea pig cardiomyocytes. Figure 8a demonstrates a typical experiment, were K_ATP currents were evoked by 1 µM rilmakalim at pH_o = 7.4. As shown previously with guinea pig (Terzic et al., 1994) and rat cardiomyocytes (Krause et al., 1995), the current increased both in the outward and the inward direction. When the rilmakalim-induced current had reached a steady state, the sulfonylureas were applied in increasing concentrations. Both glibenclamide and HMR 1883 inhibited the current dose-dependently and completely, with half-maximal concentration of 20 nM for glibenclamide and 0.8 µM for HMR 1883 (fig. 8b). Additional experiments were performed in isolated rat ventricular myocytes, where the whole-cell current was activated by 10 µM rilmakalim (pH_o = 7.4). Half-maximal inhibition of the rilmakalim-activated K_ATP current was achieved with 8 nM glibenclamide and with 0.7 µM HMR 1883 (data not shown).

View larger version (21K): [in this window] [in a new window]   Fig. 8.   Whole-cell patch-clamp recordings in guinea pig isolated ventricular myocytes a, Current traces recorded by applying voltage-ramps from 140 mV to 20 mV in one cell demonstrating the activation of inward- and outward currents by rilmakalim (1 µM) and its dose-dependent inhibition by HMR 1883. b, Dose-response curves demonstrating the inhibition of rilmakalim-activated currents by glibenclamide and by HMR 1883 (pHo = 7.4). The numbers in brackets indicate the number of experiments. c, Recordings of the action potential under current-clamp conditions, demonstrating the inhibition of rilmakalim (0.25 µM)-induced shortening of the action potential duration (APD). The inhibition of the APD90 shortening in per cent is plotted on the ordinate (the rilmakalim-induced shortening is set to 100%). The extracellular pH was 6.5.

Investigation of the coronary flow. When added to the perfusate of Langendorff-perfused guinea pig hearts, 10 and 100 µM HMR 1883 had no significant effect on the coronary flow (data not shown). Gassing of the perfusate with low oxygen (20% O₂, 75% N₂, 5% CO₂) in the absence of drugs, caused an increase in the coronary flow from 9.3 ± 0.4 ml/min to 14.6 ± 0.4 ml/min. As demonstrated in figure 9a, addition of 10 µM HMR 1883 slightly attenuated the hypoxia-induced increase in CF (18%, 7 observations), and 100 µM attenuated by 45% (n = 5). In contrast, 1 µM HMR 1883 had no significant effect, whereas 1 µM glibenclamide inhibited the hypoxia-induced increase in CF completely (data not shown). The inhibitory effect of glibenclamide is in agreement with previous studies (Daut et al., 1990). Figure 9b demonstrates that addition of 10 µM HMR 1883 to the perfusate had no significant effect, and that the increase in CF was not impaired by the drug (CF increased from 7.8 ± 0.5 ml/min to 14.1 ± 1.5 ml/min, n = 7). Subsequent wash-out of the drug under maintained hypoxia showed no effect on the CF (fig. 9b). These data indicate that half-maximal inhibition of the hypoxia-induced CF occurs at HMR 1883 concentrations far above 10 µM.

View larger version (16K): [in this window] [in a new window]   Fig. 9.   Effects of HMR 1883 on the coronary flow induced by hypoxia. a, First hypoxia was induced, and then 10 µM HMR 1883 was added. b, HMR 1883 was applied before hypoxia was induced.

Experiments with pancreatic -cells (RINm5F). The cell potential of RINm5F cells was investigated with the whole-cell mode of the patch-clamp technique. Under resting conditions the potential varied between 20 mV and 60 mV (mean 45 ± 2 mV, n = 54), and often depolarizing spikes were observed. As demonstrated in figure 10a, after addition of 100 µM diazoxide to the bathing solution, the cells hyperpolarized to 79 ± 1 mV (n = 54), where the potential remained stable. Therefore, all experiments were performed in the presence of 100 µM diazoxide. In order to test the responsiveness to K_ATP blockers, in all experiments 1 mM tolbutamide was tested. The drug caused a fast and reversible depolarization of the cell potential to 29 ± 2 mV (n = 54) (fig. 10a). Addition of glibenclamide to cells which were incubated in diazoxide-containing solution caused a concentration-dependent depolarization of the cell potential, yielding 9 nM for half-maximal inhibition of the glibenclamide-sensitive component (fig. 10b). In contrast, addition of HMR 1883 showed no significant effects at concentrations below 1 µM, whereas 10 µM of the drug depolarized the cells to 61 ± 8 mV (n = 5), and 100 µM caused depolarization to 28 ± 5 mV (n = 5) (fig. 10a). Higher concentrations of HMR 1883 were not tested in order to avoid too high amounts of the solvent DMSO. As no saturation in the dose-response curve was reached, the data were not fitted by the logistic function. Estimation by eye, however, indicates that half-maximal inhibition occurs with a HMR 1883 concentration above 10 µM.

View larger version (22K): [in this window] [in a new window]   Fig. 10.   Recordings of cell potentials in pancreatic -cells (RINm5F) with the whole-cell patch-clamp method. a, Original recording, demonstrating the hyperpolarization induced by diazoxide, the depolarization due to tolbutamide (tolbut.) and the effects of HMR 1883 (in the presence of diazoxide). The inhibitory effect of HMR 1883 was reversible after 2 min. b, Dose-response curves, summarizing the depolarizing effects of glibenclamide and HMR 1883 in RINm5F cells in the presence of 100 µM diazoxide. The number of experiments is indicated in brackets.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

Experiments under control conditions. In the present study with papillary muscles, addition of either 2 µM or 20 µM of glibenclamide or HMR 1883 had no significant effects on the action potential recorded under normal conditions (NaCl, pH = 7.4 solution). These data indicate that the substances have no appreciable effects on other K⁺ channels in guinea pig papillary muscle. The data are in agreement with previous reports, where 10 µM glibenclamide had no significant effects on the action potentials recorded in rabbit or guinea pig papillary muscles under normal conditions (Deutsch et al., 1991), (Nakaya et al., 1991). Moreover, in the present experiments no effects on the upstroke velocity of the normal or the Ca⁺⁺-dependent (slow) action potential could be observed, indicating that 20 µM of glibenclamide and HMR 1883 have neither significant effects on the fast Na⁺ channels nor on the transient Ca⁺⁺ channels. Our patch-clamp experiments directly demonstrated that 100 µM HMR 1883 had no effects on the I_K1, I_Kr and I_Ks, and thus confirm the results obtained in the papillary muscle.

Experiments with papillary muscles. HMR 1883 inhibited APD shortening caused by the potent K_ATP opener rilmakalim as well as by hypoxia and metabolic inhibition, indicating that the substance blocks K_ATPs. In contrast to the complete inhibition of rilmakalim-induced APD shortening, hypoxia- and DOG-induced APD shortening could not be fully prevented by HMR 1883 or by glibenclamide, although the concentration dependence showed saturation behavior (figs. 5 and 7). The observations are in agreement with previous studies, where pretreatment with glibenclamide prevented anoxia-induced APD shortening only partially in guinea pig papillary muscle (Nakaya et al., 1991), (Takizawa et al., 1996). In experiments with papillary muscles, where hypoxia was induced in the absence of drugs, a shortening of the APD₉₀ by about 115 msec was observed after 60 min hypoxia and by about 139 msec after 120 min. These results are in good agreement with recent observations (Hayashi et al., 1997) where a shortening of the APD to about 20% of control was recorded after 60 min, but only a slightly further shortening could be recorded after 120 min of hypoxia. When added after 60 min of hypoxia, both glibenclamide and HMR 1883 were able to partially prolong the APD₉₀. It is interesting to note that with 2 µM glibenclamide the prolongation was only 49%, and there was no further prolongation with 20 µM. These data are in good agreement with previous studies, where incomplete APD prolongation was observed when 10 µM glibenclamide was added 30 min after induction of hypoxia (Wang et al., 1996). The present study shows that HMR 1883 was less potent than glibenclamide, but induced a clear prolongation of the APD₉₀ at 2 µM when applied 60 min after hypoxia. The incomplete block of K_ATPs by the sulfonylureas could be due to the fact that the substances partially lose their potency in blocking channels which were activated by metabolic inhibition (Wilde et al., 1990), (Findlay, 1993), (Krause et al., 1995). As reported previously, it is unlikely that reduction in Ca⁺⁺ currents plays a significant role in the shortening of the APD under hypoxic conditions (Noma and Shibasaki, 1985).

Patch-clamp experiments. In guinea pig cardiomyocytes the rilmakalim-activated whole-cell current was blocked by glibenclamide with an apparent IC₅₀ value of 20 nM in the presence of 1 µM rilmakalim at pH_o = 7.4 (fig. 8b). In previous experiments an even higher potency of glibenclamide (IC₅₀ of about 6 nM and 8 nM) was reported for inhibition of whole-cell currents activated by the channel opener SR 44866 (Findlay, 1992a) or rilmakalim (Krause et al., 1995). This discrepancy could be due to different experimental conditions. When results from papillary muscles and isolated cells are compared, it is obvious that glibenclamide inhibits with higher potency in the latter preparation. A possible explanation could be that in papillary muscles glibenclamide cannot reach underlying tissue, whereas underlying cells are electrically coupled to surface cells, from which recordings are obtained. Interestingly, the difference between potency of inhibition between papillary muscle and single cells does not exist for HMR 1883. Possibly, HMR 1883 can reach the underlying tissue in papillary muscle in contrast to glibenclamide.

Experiments with Langendorff-perfused hearts. The present study demonstrates that ventricular fibrillation could be induced with high incidence in Langendorff-perfused guinea pig hearts in a similar way as it was reported for rabbit hearts (Chi et al., 1993). These observations indicate that K⁺ channel openers have pro-fibrillatory effects in guinea pig hearts, and thus confirm previous observations made in isolated rabbit hearts (Chi et al., 1993) and in a conscious canine model of sudden cardiac death (Chi et al., 1990). As observed by Chi et al. (Chi et al., 1993) in their rabbit heart model, we showed that glibenclamide was able to prevent the occurrence of VF in our model. In addition we demonstrated that HMR 1883 prevented VF in a dose-dependent manner. Approximately 50% of the hearts were protected by a concentration of 100 nM. This low concentration of HMR 1883 showed slight effects on the rilmakalim-induced shortening of action potentials at pH_o = 6.0 (fig. 3) as well as on rilmakalim-induced currents in isolated guinea pig cardiomyocytes (fig. 8). Obviously, the blockade of a small number of open K_ATPs may be sufficient to exert a cardioprotective effect.

Experiments with pancreatic (RINm5F) cells. Like rat pancreatic -cells, RINm5F cells possess K_ATPs which are activated by diazoxide and inhibited by tolbutamide and glibenclamide (Dunne et al., 1987). The fact that glibenclamide caused only a partial inhibition of the cell potential indicates that K_ATPs do not contribute exclusively to the cell's resting potential. This observation is in agreement with data from mouse pancreatic cells, where four different types of K⁺ channels were observed (Rorsman and Trube, 1986).

View this table: [in this window] [in a new window]   TABLE 2 Relative potency estimates (IC50 values) for inhibition of KATP channels by glibenclamide and HMR 1883 in cardiac tissues (guinea pigs), pancreatic -cell (rat, RINm5F) and coronary vascular system (guinea pig) under hypoxic conditions

Limitations of this study. In the present study we described a number of in vitro experiments performed with different tissues from different species. Especially, most experiments performed with cardiac tissue were done with guinea pig, whereas the pancreatic effect was investigated with cultured cells derived from rat (RINm5F). Some experiments performed with isolated rat cardiomyocytes confirmed the results obtained with guinea pigs, yielding high potency for glibenclamide (IC₅₀ value 8 nM) to inhibit the rilmakalim-activated K_ATP current, whereas HMR 1883 blocked half-maximally with 0.7 µM. Preliminary investigations in conscious dogs indicated that no significant insulin release was observed at plasma concentrations of HMR 1883 up to 30 µM, whereas protection of conscious dogs from ventricular fibrillation occurred at plasma concentrations of approximately 3 µM (own, unpublished results). Moreover, no effects on the coronary flow could be observed at the antifibrillatory dose in dog (Billman and Englert, 1998).