Introduction

Top Abstract Introduction Methods Results Discussion References

The treatment of heart failure is very important in the field of cardiovascular medicine. Recent clinical trials with angiotensin-converting enzyme inhibitors demonstrated the improvement of quality of life and the reduction of mortality in patients with severe chronic heart failure (CONSENSUS Trial Study Group, 1987; SOLVD Investigators, 1991). However, results with positive inotropic agents, such as PDE inhibitors and beta 1 agonists, were disappointing in terms of excess morbidity and mortality without producing important clinical benefits (Packer, 1988; Packer et al., 1991).

A new class of cardiovascular drugs, the myofilament Ca⁺⁺ sensitizers, has been developed for the treatment of both acute and chronic heart failure (Packer, 1988; Nielsen-Kudsk and Aldershvile, 1995). Compared with previous positive inotropic agents that enhance intracellular Ca⁺⁺, the Ca⁺⁺-sensitizing drugs, including pimobendan, EMD 53998, MCI-154 and levosimendan, increase myocardial contractility by producing more force for a given amount of intracellular Ca⁺⁺ (Nielsen-Kudsk and Aldershvile, 1995). Among these agents, levosimendan has a unique property in that it binds to cardiac cTnC in a Ca⁺⁺-dependent manner (Pollesello et al., 1994; Haikala et al., 1995a, 1995b; Nielsen-Kudsk and Aldershvile, 1995). The positive inotropic effect, due to increased Ca⁺⁺ sensitivity of the contractile proteins, has been reported to be exerted at concentrations of 0.03 to 10 µM in skinned fibers from guinea pig papillary muscles (Edes et al., 1995; Haikala et al., 1995b). In addition, levosimendan causes vasodilation in both experimental animal models (Harkin et al., 1995; Pagel et al., 1995; Vegh et al., 1995) and clinical studies (Lilleberg et al., 1995; Sundberg et al., 1995; Vegh et al., 1995). Although levosimendan (0.1-0.3 µM) was suggested to act as a PDE inhibitor (Edes et al., 1995), this drug had favorable preventive effects against ventricular tachyarrhythmias induced by ischemia-reperfusion (Kaszala et al., 1994).

Although levosimendan (5-10 µM) increased the I_Ca(L) in guinea pig ventricular cells, presumably by the PDE inhibition (Virag et al., 1996; Boknik et al., 1997), little data are available concerning the electrophysiological effects of levosimendan. In the present study, we explored the effects of levosimendan on membrane currents and action potentials in rat ventricular cells using the patch-clamp technique of whole-cell and single-channel recordings. Levosimendan shortened APD, and our data suggest this occurred through opening of K_ATP channels.

	Methods

Top Abstract Introduction Methods Results Discussion References

Young (10-18 days old) Sprague-Dawley rats were used for this study. The rats were handled in accordance with the Declaration of Helsinki and the Guiding Principles in the Care and Use of Animals.

Cell preparation. Freshly isolated single cells were prepared from ventricles of young rats as previously described (Yokoshiki et al., 1996). In brief, the rats were decapitated while under CO₂ anesthesia, and the hearts were removed and rinsed in oxygenated Tyrode's solution and, then immersed in Ca⁺⁺-free Tyrode's solution for 20 min. After the spontaneous beatings had ceased, the ventricles were dissected, and small pieces were enzymatically digested for 50 min (37°C) in Ca⁺⁺-free Tyrode's solution containing collagenase (1 mg/ml; Wako Chemicals, Osaka, Japan). The cells were mechanically dispersed in the modified KB solution at room temperature using a Pasteur pipette. The cell suspensions were stored in a refrigerator (4°C) and were used for 2 to 6 hr after isolation. Usually, 40% to 60% of the isolated cells were rod shaped in Tyrode's solution (containing 1.8 mM Ca⁺⁺). Rod-shaped cells with smooth surfaces and clear cross-striations were selected for experiments. All experiments were performed at room temperature (20-22°C).

Whole-cell recordings. The standard patch-clamp technique was applied in the whole-cell configuration with a patch-clamp amplifier (Axopatch-1D; Axon Instruments, Foster City, CA). Action potentials and whole-cell currents were measured under current-clamp or voltage-clamp mode, respectively. Voltage-clamp experiments were performed by applying either voltage ramp or step pulses. The patch electrodes (2-5 M) were made from borosilicate glass capillary tubing (World Precision Instruments, Sarasota, FL). The cell suspension was placed into a small chamber (0.5 ml) on the stage of an inverted microscope (TMD-Diaphot; Nikon, Tokyo, Japan). The bath was superfused with the normal Tyrode's solution. The composition of the internal (pipette) solution for the whole-cell experiments was (in mM) KCl 140, MgCl₂ 1, HEPES 5, EGTA 10 and ATP-Na₂ 5, pH adjusted to 7.2 with KOH.

Single-channel recordings. Single-channel current recordings were made in the open cell-attached patch configuration (Kakei et al., 1985; Ohya and Sperelakis, 1989) with the same patch-clamp amplifier used in the whole-cell experiments. In brief, after making cell-attached patch with recording pipette, one end of the cell was mechanically disrupted using another glass pipette containing the bath solution. The tip of the patch electrode was coated with Sylgard (Dow Corning, MI), and its resistance ranged from 5 to 8 M when it was filled with the following pipette (extracellular) solution (in mM): KCl 150, MgCl₂ 0.5, CaCl₂ 1.8 and HEPES 5, pH adjusted to 7.4 with KOH. The bath (intracellular) solution contained (in mM) KCl 150, MgCl₂ 0.5, HEPES 5, and EGTA 1, pH adjusted to 7.3 with KOH. Current signals were filtered at 1 or 2 kHz, sampled at 1 or 10 kHz and stored in a personal computer. Data recorded from only one functioning channel with a sampling rate of 10 kHz and a cutoff frequency of 2 kHz were used for the channel kinetic analysis (see fig. 5C). The storing of the digitized signals was carried out using AxoTape (version 2, Axon Instruments), and analysis was performed with pCLAMP (version 6.02, Axon Instruments).

View larger version (43K): [in this window] [in a new window]   Fig. 5.   Conductance and kinetic properties of the KATP channel currents in open cell-attached patch. A, KATP channel currents recorded in open cell-attached patch (symmetrical 150 mM K+) from young rat ventricular cells at various membrane potentials are given. Bath solution contained no ATP. B, Current-voltage relationship of the KATP channel. The slope conductance of unitary inward current was 80 pS (n = 4). C, Open-time histograms of the unitary KATP channel currents at 80 mV. The histogram at 80 mV was fitted by a single exponential curve with a time constant of 1.3 msec. Data were filtered at 2 kHz and sampled at 10 kHz.

Drugs and chemicals. Levosimendan was a gift from Orion-Farmos (Espoo, Finland). Various NDPs, nucleotide triphosphates and glibenclamide were purchased from Sigma Chemical (St. Louis, MO).

Data analysis. All data are given as mean ± S.E.M. Statistical significance was evaluated by Student's paired t test, and P < .05 was considered to be significant.

	Results

Top Abstract Introduction Methods Results Discussion References

Shortening of APD. Action potentials in young (days 10-18) rat ventricular cells were recorded in current-clamp at a stimulation rate of 0.1 Hz (fig. 1). Bath application of 10 µM levosimendan produced shortening of APD with slight hyperpolarization of RP (table 1). This effect usually started 1 to 5 min after exposure to levosimendan and reached a maximum within 1 to 2 min from the onset of the response. As shown in traces d and e of figure 1A, the cells exposed to 10 µM levosimendan became inexcitable at its maximal effect. This effect could be reversed after washing out of the levosimendan (fig. 1A, right). Levosimendan (3 µM) also abbreviated the APD (similar to trace c) but did not abolish excitability [i.e., the overshoot of the action potential (above the zero potential) remained]. These effects on the action potentials are summarized in table 1.

View larger version (33K): [in this window] [in a new window]   Fig. 1.   Shortening of APD by levosimendan in young (days 10-18) rat ventricular cells in current-clamp mode. A, Action potentials recorded in current-clamp. Extracellular application of 10 µM levosimendan caused shortening of APD; this effect could be reversed after washing out (W.O.). Arrows, peak levels of the action potentials in d and e. B, Time courses of changes in APA, APD at 60 mV (APD60) and resting potential (RP). a-h, Time points at which action potentials in A were recorded.

Stimulation of K_ATP current. Superimposed current traces evoked by voltage-clamp ramps are shown in figure 2A, and the time courses of the currents (measured at 0 and 100 mV) are illustrated in figure 2B. The ramp pulses were applied every 15 sec from +40 to 100 mV (at 40 mV/sec). The outward current was rapidly increased by levosimendan. The average reversal potential (V_rev) was 76.3 ± 0.7 mV (n = 5), which was close to the equilibrium potential for K⁺ (E_K) value (83 mV) [E_K = 59 mV × log (140/5.4)], suggesting that most of the current stimulated by levosimendan was a K⁺ current. This effect usually started 1 to 5 min after exposure to levosimendan and reached a maximum within 1 to 2 min from the onset of the response. Glibenclamide (1 µM), a relatively specific inhibitor of K_ATP channels, abolished the levosimendan-stimulated current (n = 4), as shown in figure 2. Therefore, the levosimendan-stimulated current is similar to a K_ATP current.

View larger version (26K): [in this window] [in a new window]   Fig. 2.   Stimulation of K+ current by levosimendan in rat ventricular cells. A, Superimposed current traces evoked by voltage-clamp ramps (40 mV/sec) between +40 mV and 100 mV. The outward current was rapidly increased by levosimendan (b) and intersected with the abscissa at ~75 mV. The reversal potential (Vrev) was close to the EK (83 mV), suggesting that most of the current stimulated by levosimendan was a K+ current. Glibenclamide (1 µM) abolished the levosimendan-stimulated currents (c). B, Time courses of the currents measured at 0 and 100 mV. a-c, Time points at which the currents in A were recorded.

View larger version (19K): [in this window] [in a new window]   Fig. 3.   Dose-dependent activation of K+ currents by levosimendan. A, Current-voltage relations for the currents stimulated by different doses of levosimendan. The currents were evoked by voltage-clamp ramps (40 mV/sec) between +40 and 100 mV. B, Dose-response relation for stimulation of K+ currents by levosimendan. Currents were evoked by ramp pulses and measured at 0 mV when the currents were maximally activated. Difference in currents between the maximal activated and control (before application of levosimendan) were plotted against each concentration. Number of cells studied are shown in parentheses. Data points were fitted by the Hill equation. The EC50 value was 4.7 µM, and the Hill coefficient (nH) value was 2.3.

No effect on I_Ca(L). I_Ca(L) was recorded under the condition in which all K⁺ currents were blocked by extracellular TEA and 4-AP and intracellular Cs⁺. Fast Na⁺ current was also blocked by substitution of extracellular Na⁺ with TEA. I_Ca(L) was evoked every 15 sec by a test potential to +10 mV from a holding potential of 40 mV. As shown in figure 4, bath application of 10 µM levosimendan had no effect on I_Ca(L); peak current amplitude after a 5-min exposure to levosimendan was 96.5 ± 0.8% of control (n = 4). In contrast, isoproterenol (1 µM) markedly stimulated I_Ca(L) (204 ± 16% of control). Lower doses of levosimendan [0.2 µM (n = 5) and 1 µM (n = 4)] also had no effect on I_Ca(L) (0.2 µM, 94.4 ± 1.8% of control; 1 µM, 97.7 ± 1.0% of control).

View larger version (17K): [in this window] [in a new window]   Fig. 4.   Lack of effects of levosimendan on rat cardiac ICa(L). ICa(L) was evoked every 15 sec by a test potential to +10 mV from a holding potential of 40 mV. Levosimendan (10 µM) had no effects on ICa(L), which was markedly stimulated by the subsequent application of 1 µM isoproterenol (ISO).

Stimulation of K_ATP channels. Single-channel K⁺ currents were recorded in the open cell-attached patch mode (symmetrical 150 mM K⁺) from rat ventricular cells at various membrane potentials (fig. 5A). The K⁺ channel activities appeared within 30 sec after the open cell-attached patches were made by mechanical disruption of one end of the cell in bath solution containing no ATP. The opening of the channels appeared in bursts, and the flickerings within bursts decreased when the membrane was depolarized, as previously reported (Kakei et al., 1985; Tung and Kurachi, 1991). The current-voltage relation for this channel is shown in figure 5B. The conductance of the unitary inward current was 80 pS (n = 4), and a slight inward rectification was observed at positive membrane potentials. The open-time histograms (at 80 mV) was fitted by a single exponential curve with a time constant of 1.3 msec (fig. 5C). The conductance and kinetic properties of this K⁺ channel are similar to those of the K_ATP channel previously reported for cardiac cells (Kakei et al., 1985; Tung and Kurachi, 1991).

View larger version (31K): [in this window] [in a new window]   Fig. 6.   Stimulation of KATP channels by levosimendan in the presence of ATP. After formation of open-cell attached patches in ATP-free solution, spontaneous variation of KATP channel activities was observed. KATP channel activities were inhibited by 0.3 mM ATP in the bath. Levosimendan (10 µM) activated the KATP channels that had been inhibited by ATP, and the channels were abolished by the bath application of 10 µM glibenclamide (Glib). Membrane potential (Vm) was held at 80 mV. Top trace, sampling rate of 0.25 kHz; expanded traces, 10 kHz. The cutoff frequency of the filter was 2 kHz.

View larger version (27K): [in this window] [in a new window]   Fig. 7.   Lack of effect of levosimendan on KATP channels in the absence of ATP. KATP channels exhibiting high spontaneous activity (A) or run-down (B) in ATP-free solution were not stimulated by 10 µM levosimendan.

View larger version (26K): [in this window] [in a new window]   Fig. 8.   Levosimendan synergistically activates run-down KATP channels with nucleotide diphosphates (NDPs). After run-down of KATP channels, 100 µM ADP (A) and 3 mM UDP (B) could stimulate the channels. Subsequent application of levosimendan (10 µM) produced further stimulation of the channels.

View larger version (30K): [in this window] [in a new window]   Fig. 9.   Effect of levosimendan on KATP channels in the presence of AMP-PNP, a nonhydrolyzable ATP analog. A, Levosimendan (10 µM) had no effect on KATP channels inhibited by 0.5 mM AMP-PNP. However, subsequent application of 10 µM levosimendan in the presence of 0.3 mM ATP was capable of stimulating the channels. B, KATP channels inhibited by AMP-PNP were activated by levosimendan when ADP (30-50 µM) was present in the bath solution.

	Discussion

Top Abstract Introduction Methods Results Discussion References

In the present study, we demonstrated that the new Ca⁺⁺-sensitizing positive inotropic agent levosimendan activated K_ATP channel currents in rat ventricular myocytes under physiological conditions. In contrast, levosimendan (0.2-10 µM) had no effect on I_Ca(L). Single-channel recordings (in open cell-attached patches) showed that levosimendan stimulated the K_ATP channels in the presence of ATP, whereas stimulation was not observed in channels exhibiting a high degree of spontaneous activity in ATP-free condition. Levosimendan also could not stimulate rundown K_ATP channels. Although the channels inhibited by AMP-PNP, a nonhydrolyzable ATP analog, could not be activated by levosimendan, the presence of NDPs restored the ability of levosimendan to potentiate the channel activity. Furthermore, levosimendan synergistically stimulated the rundown K_ATP channels when NDPs were present.

In addition to the Ca⁺⁺-sensitizing positive inotropic effect, levosimendan (0.1-100 µM) increased the cAMP level in guinea pig cardiomyocytes (Edes et al., 1995; Boknik et al., 1997). On the basis of these findings, levosimendan was suggested to produce PDE inhibition, and levosimendan was found to stimulate I_Ca(L) (Virag et al., 1996; Boknik et al., 1997). On the contrary, in the present study, I_Ca(L) of rat cardiomyocytes was not affected by levosimendan (0.2-10 µM). This discrepancy might arise from the different species studied, which have different PDE isozymes (Polson, 1996). For example, in isolated rat ventricular cells, several selective PDE III inhibitors (to which milrinone belongs) had little or no effect on cAMP level, but a PDE IV inhibitor, Ro 20-1724, had a potent action in increasing cAMP (Kelso et al., 1993). Thus, cAMP could still be hydrolyzed by the PDE IV isozyme (shown to be abundant in rat ventricle; Bode et al., 1991), even when the PDE III was inhibited. In addition, the basal level of cAMP could be different in the two cases.

The activation of K_ATP channels by levosimendan is not due to any metabolic impairment of the cell because levosimendan (30 µM) did not change the activity of the myofibrillar ATPase that consumes ATP for cross-bridge cycling (Haikala et al., 1995b). Furthermore, I_Ca(L), which is metabolically regulated in several tissues (Sperelakis and Schneider, 1976; Irisawa and Kokubun, 1983; O'Rourke et al., 1992; Yokoshiki et al., 1997), was not affected by levosimendan (0.2-10 µM) in the present study.

The Ca⁺⁺-sensitizing effect of levosimendan in skinned fiber was reported to increase dose-dependently from 0.03 to 10 µM (Edes et al., 1995; Haikala et al., 1995b). However, in intact muscles (guinea pig papillary muscles), high doses (>1 µM or 10 µM) of levosimendan produced lesser stimulation or even negative inotropy (Haikala et al., 1995b; Boknik et al., 1997). Therefore, stimulation of K_ATP channels observed in the present study may account for both the lesser stimulation (or negative inotropy) at the higher doses and the vasodilatory action.

Although levosimendan produced PDE inhibition at the doses of 0.1 to 0.3 µM (Edes et al., 1995), this drug prevented ventricular fibrillation induced by ischemia-reperfusion in anesthetized dogs (Kaszala et al., 1994), thought to be due to Ca⁺⁺ overload and subsequent triggered activity (Janse and Wit, 1989). Levosimendan also reduced ischemic damage induced by ligation of the coronary artery in isolated rabbit hearts (Rump et al., 1994a, 1994b), an effect observed even at 0.1 µM (Rump et al., 1994b). At this concentration, the increase in the coronary flow was relatively small, and the authors speculated that factors beyond coronary dilation might contribute to the anti-ischemic effects of levosimendan. On the other hand, opening of K_ATP channels is generally considered to be cardioprotective against ischemia-related events (Hearse, 1995). Therefore, our results may in part account for the antiarrhythmic and anti-ischemic effects reported in the previous studies (Kaszala et al., 1994; Rump et al., 1994a, 1994b) because activation of K_ATP channels by levosimendan would likely occur in ischemic regions in which intracellular ADP concentration ([ADP]_i) is increased and intracellular ATP concentration ([ATP]_i) is decreased.

The therapeutic concentration of levosimendan as a positive inotropic agent is thought to be ~50 ng/ml (i.e., 0.18 µM) in patients with left ventricular dysfunction (Lilleberg et al., 1995; Sandell et al., 1995). This value is lower than that required for KCO action of levosimendan in the present study. Therefore, levosimendan at this therapeutic concentration may produce little or no activation of K_ATP channels in human hearts under normal conditions. However, levosimendan might more readily activate the K_ATP channels in ischemic myocardium because the ratio of [ADP]_i to [ATP]_i ([ADP]_i/[ATP]_i) may be increased, as mentioned above. Because the synergistic action of levosimendan with ADP could be observed at low concentrations (30-50 µM ADP), levosimendan might protect ischemic myocardium more effectively than nicorandil, which required higher ADP concentrations [0.5 mM (Shen et al., 1991) to 1 mM (Jahangir et al., 1994)] for activation of K_ATP channels. However, one must be careful in extrapolating these experimental results to the clinical setting.

The site of action of KCOs remains unclear (Henry and Escande, 1994). It might be simplest to assume that there is competition between KCOs and ATP at the ATP-binding site that closes the channel. The parallel shift produced by KCOs in the concentration-response curve for ATP inhibition of the channels (Thuringer and Escande, 1989; Nakayama et al., 1990; Ripoll et al., 1990) may support this hypothesis. However, the possibility exists that ATP and KCOs act at different sites with opposing effects on the channel. Because levosimendan was not able to stimulate the channel that had been inhibited by AMP-PNP, the competition hypothesis at the ATP-binding site (inhibitory) would not account for the K⁺ channel-opening action of levosimendan.

A synergistic activation of K_ATP channels by the KCO drug diazoxide and ADP was reported in pancreatic  cells (Larrson et al., 1993). The authors concluded that channel stimulation by diazoxide is dependent on the binding of Mg-ADP to a cytosolic regulatory constituent of the channel. Therefore, the synergistic action with NDPs might be shared by other KCOs as in the case of levosimendan. In addition, the functional classification of KCOs has recently been postulated based on their mechanisms of action (Terzic et al., 1995). According to their model, KCOs were classified into three types because they seem to have at least three sites of action: (1) the ATP-binding unit (inhibitory), (2) the NDP-binding unit (stimulatory) and (3) the transducer unit. Nicorandil, a NDP-acting drug ("type 3" in their classification), is considered to act in the presence of NDP by both enhancing the maximal channel activity and decreasing sensitivity of K_ATP channels to inhibition by ATP (Shen et al., 1991; Terzic et al., 1995). With respect to mechanism of action, levosimendan may act similarly to nicorandil. One difference is that activation of the channel by nicorandil was not observed in the presence of ATP (0.5 mM) unless ADP (0.5 mM) was also present (Shen et al., 1991). On the other hand, levosimendan could stimulate the channel in the presence of ATP alone, that is, in the absence of added ADP; however, some endogenous ADP was undoubtedly present. Because levosimendan could antagonize the channel inhibition by AMP-PNP if ADP (30-50 µM) were present, it might be possible that levosimendan synergistically activated the channels with a small amount of ADP generated by ATP hydrolysis in the vicinity of the channels. For example, hydrolysis of ATP occurs continuously within the sarcolemmal membrane in association with cellular activities such as Na⁺-K⁺ pump, Ca⁺⁺-ATPase (Sperelakis, 1995) and actin filament assembly (Stossel, 1993) to maintain homeostasis.

In summary, levosimendan shortened APD in rat ventricular cells, presumably through opening of K_ATP channels. This stimulation of K_ATP channels by levosimendan may result from the synergistic action with NDPs.