Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

L-type Ca⁺⁺ channels are crucial for the initiation and the maintenance of contraction in cardiac and smooth muscle (for review see McDonald et al., 1994). Drugs that block L-type Ca⁺⁺ channels (so-called Ca⁺⁺ antagonists) are widely used clinically in the treatment of hypertension, angina pectoris and cardiac arrhythmias (see Roberts and Zanchetti, 1996). Three major classes of structurally unrelated drugs, the phenylalkylamines (typified by verapamil), the dihydropyridines (typified by nifedipine) and the benzothiazepines (typified by diltiazem) belong to this important group of drugs. The common target of all known L-type Ca⁺⁺ -antagonists is the ₁-subunit of the channel molecule (Catterall and Striessnig, 1992), although different binding sites within the subunit have been identified, which are allosterically linked (reviewed by Hockerman et al., 1997). An interesting feature of the action of L-type Ca⁺⁺ -antagonists is their use-dependence, described by an increase in block at higher frequencies of stimulation and at more positive membrane potentials (McDonald et al. 1980; Lee and Tsien, 1983; Sanguinetti and Kass, 1984; Uehara and Hume, 1985). Two alternative hypotheses have been put forward which may account for the voltage- and time-dependent effects of L-type Ca⁺⁺-antagonists: the modulated receptor (Hondeghem and Katzung, 1984) and the guarded receptor (Starmer et al., 1989) theory. Both theories imply that the effects of the drugs are related to the existence of individual Ca⁺⁺ channel states during an excitation cycle (resting, open, and inactivated). The three channel states have a major impact for the effects of L-type Ca⁺⁺ -antagonists, and both binding studies and patch clamp experiments have shown that the binding affinities of these drugs are significantly determined by the channel states (McDonald et al., 1984; Uehara and Hume, 1985; Rakotoarisoa et al., 1990). Whereas some studies indicate that all three channel states are involved, most of them favor the sequence inactivated > open > resting, the inactivated state of the channels being the major determinant for drug binding (reviewed by McDonald et al., 1994; Hockerman et al., 1997). We have shown earlier that, in rat ventricular cardiomyocytes, verapamil exclusively binds to the inactivated state of L-type Ca⁺⁺ channels (Nawrath and Wegener, 1997). Under these conditions, the block of I_Ca(L) was purely use-dependent, without tonic component.

Drug binding to open channels would increase the decay of the Ca⁺⁺ current due to progressive block of open channels during depolarization. Such a phenomenon has been described in earlier studies for dihydropyridines and phenylalkylamines (Lee and Tsien, 1983; Timin and Hering, 1992). In the case of phenylalkylamines, however, the evidence for this is contradictory (Cohen and Lederer, 1987; Uehara and Hume, 1985). We also looked carefully for changes in the decay of I_Ca(L) in rat ventricular cardiomyocytes treated with verapamil, but found no convincing evidence for any change in the kinetics of the current (Nawrath and Wegener, 1997).

We report that another clinically used L-type Ca⁺⁺-antagonist, fendiline (belonging to the class of diphenylalkylamines), blocks Ca⁺⁺ channels preferentially in the open state. This is documented by 1) the acceleration of the current decay in the presence of the drug, 2) the enhancement of effects by favoring the open state of the channel and 3) reduced open times of single Ca⁺⁺ channels. A preliminary account of this work has been published (Nawrath and Rupp, 1997).

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Preparations. Sprague-Dawley rats (200-300 g) of either sex were anesthetized with ether and bled from the carotid arteries. The heart and the thoracic aorta were quickly removed and immersed in warmed and oxygenated solution A (containing in mM: NaCl 137, KCl 5.4, CaCl₂ 1.8, MgCl₂ 1, NaHCO₃ 12, NaH₂PO₄ 0.42, glucose 5.6; aerated with 95% O₂ + 5% CO₂; pH 7.4). After removal of the endothelium, the aorta was cut into rings of 3 to 5 mm and tied with silk ligatures after the connective tissue had been removed. Single ventricular cardiomyocytes were isolated as described previously (Wegener and Nawrath, 1995). Briefly, the hearts were enzymatically digested by perfusion with a collagenase-containing buffer solution via the aorta using the Langendorff-setup. Single myocytes were obtained from ventricular tissue pieces by mechanical dispersion.

Measurement of tension. Aortic rings were suspended in organ baths (5 ml) containing oxygenated solution A at 36 ± 1°C. One end was fixed to a hook of a muscle holder and the other end was connected to an inductive force-displacement transducer whose output was fed to a carrier frequency preamplifier (Carrier amplifier/TA2000, Gould, Cleveland, OH). Resting tension was set to 10 mN. Aortic rings were precontracted by high extracellular [K⁺] solution (in mM: NaCl 94, KCl 48, CaCl₂ 1.8, MgCl₂ 1, NaHCO₃ 12, NaH₂PO₄ 0.42, glucose 5.6; aerated with 95% O₂ + 5% CO₂; pH 7.4). Removal of endothelium was verified by the lack of any relaxation in the presence of carbachol (3 µM). Drugs were added from stock solutions to the organ bath to achieve the final concentrations as indicated.

Whole-cell recordings. Electrophysiological experiments were performed on rod-shaped myocytes with clear cross striations using the whole-cell configuration of the patch-clamp technique (Hamill et al., 1981). During the experiments, the myocytes were continuously superfused with solution B (in mM: NaCl 137, KCl 5.4, MgCl₂ 0.5, CaCl₂ 1.8, glucose 5, HEPES 10). When Ba⁺⁺ ions were used as charge carrier, the superfusing solution B was changed to solution C (in mM: NaCl 137, KCl 5.4, MgCl₂ 0.5, BaCl₂ 1.8, glucose 5, HEPES 10). The recording patch pipettes were built from borosilicate glass (Science Products, Frankfurt, FRG). The resistances of the pipettes ranged from 1 to 1.3 M when filled with pipette solution (composition in mM: CsCl 125, MgCl₂ 6, CaCl₂ 0.15, K₂ATP 5, Na₂GTP 0.1, EGTA 5, HEPES 10; pH was adjusted with CsOH to 7.3). Membrane currents were recorded by an EPC-7 amplifier (List, Darmstadt, Germany) which received rectangular voltage pulses from a PC equipped with a Labmaster interface (Scientific Solutions, Solon, OH). Current signals were filtered at 3 kHz (8-pole Bessel Filter, Rockland System Corp., Rockland, ME), digitized at a sampling rate of 1 kHz, stored on an AT compatible computer running pClamp software (Axon Instr. Inc., Foster City, CA) which was additionally used for the generation of voltage pulses and data analysis. During the experiments, the myocytes were voltage-clamped at a holding potential of -80 mV. To inactivate the fast sodium current, a 20-msec prepulse to -40 mV was set before activating the Ca⁺⁺-current. L-type Ca⁺⁺-currents [I_Ca(L)] were elicited by 180 msec depolarizing voltage pulses to 0 mV at 0.2 Hz. Whole-cell currents through Ca⁺⁺ channels carried by Ba⁺⁺ ions [I_Ba(L)] were elicited by 180 or 300 msec depolarizing voltage pulses from -80 mV or -40 mV to +10 mV at 0.2 Hz. The experiments were performed at 36 ± 1°C.

Single channel recordings. Single Ca⁺⁺ channel currents were recorded using the cell-attached configuration of the patch-clamp technique (Hamill et al., 1981). Patch pipettes (resistances 1-3 M) were coated with Silicone resin (Sylgard, Dow Corning Company, München, Germany) and then heat-polished in a microforge. The patch pipettes were filled with solution (in mM: BaCl₂ 100, HEPES 5, tetrodotoxin 0.03; pH was adjusted with Tris to 7.4). During the experiment, the myocytes were continuously superfused with bath solution (in mM: K⁺ aspartate 100, KCl 40, Mg₂Cl 1, HEPES 5, EGTA 5, glucose 10; pH was adjusted with KOH to 7.4). The membrane potential was corrected for the junction potential between pipette and the bath solution (-19 mV). For the activation of Ca⁺⁺ channels, test pulses of 100-msec duration from a holding potential of -80 to 0 mV were delivered at 0.1 Hz. The number of depolarizing voltage steps amounted to 300 under control and test conditions. Currents were digitized at a sampling rate of 5 KHz and recorded as described above. After gigaseal formation, channel activity was monitored under control and test conditions. In this study, about 55 of 460 membrane patches (seal resistances: 10-100 G) showed channel activity. Membrane patches lacking channel activity were rejected. A sufficient amount of data for analysis of single channel activity was obtained from 26 patches; 7 patches were discarded from analysis due to disappearance of channel activity under control conditions.

Chemicals. All salts and solvents used were at least p. a. grade and purchased from Sigma Chemical Co. (St Louis, MO). BayK8644 was obtained from Bayer (Leverkusen, Germany), fendiline and tetrodotoxin from Sigma. Verapamil was a gift from Knoll (Ludwigshafen, Germany). The structural formulae of fendiline and verapamil are shown in figure 1. 

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Structural formulae of verapamil and fendiline.

Evaluation of results. Data are presented as original recordings or expressed as means ± S.E.M. (in % of control values). Changes in aortic tension were expressed in % of K⁺-induced tension. Cardiac I_Ca(L) and I_Ba(L) were measured as the difference of peak inward and steady-state current at the end of the voltage pulse. Single channel currents were measured as difference between closed and open current level. The analysis of single channel currents was performed using ASCD software (provided by Dr. Guy Droogmans, Department of Physiology, Leuven, Belgium) and pClamp software (Axon Instr. Inc). Capacitative and leak currents were digitally subtracted using the average current of the blank sweeps. For construction of mean current, all records including blanks were averaged. Mean open and mean shut times were calculated as arithmetic values from idealized openings; for the construction of idealized openings, transitions between the open and closed states were detected as crossings at the half amplitude levels. Events shorter than 0.2 msec were rejected from analysis. In addition, open times were compiled in frequency histograms of 0.2 msec bin width for further statistical analysis. Decay of I_Ca(L) was fitted to single or two exponential functions. Open time distributions and concentration-response curves were fitted to two exponential and sigmoidal functions, respectively, (correlation coefficient > 0.99) using GraphPad Prism 2.0 (GraphPad Software Inc., San Diego, CA). Statistical analysis was performed using either paired or unpaired Student's t test. Differences were considered as significant at P < .05.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

I_Ca(L) was elicited repetitively at a frequency of 0.2 Hz by depolarizing voltage clamp steps (shown in fig. 2A) both under control conditions and in the presence of either verapamil or fendiline. Both drugs diminished the magnitude of I_Ca(L) in a concentration-dependent manner (fig. 2B). The EC₅₀ values of verapamil and fendiline amounted to 0.9 and to 13 µM, respectively. The original records in figures 2C and D show the effects of about half maximally effective concentrations of either drug on I_Ca(L). Verapamil (1 µM) reduced peak I_Ca(L) by 41% 2 min after the addition of the drug (fig. 2C). The time course of I_Ca(L) decay remained virtually unchanged;  amounted to 10.1 and to 9.8 msec, under control conditions and 2 min after the addition of verapamil, respectively. In the same cell, fendiline (10 µM; after 10 min washout of verapamil) reduced peak I_Ca(L) by 38%. Different to verapamil, fendiline accelerated the decay of I_Ca(L);  amounted to 10.7 and to 7.5 msec, under control conditions and 2 min after the addition of fendiline, respectively.

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Effects of fendiline and verapamil on ICa(L). A, Voltage protocol used for activation of ICa(L). After establishment of the whole-cell configuration, the myocytes were depolarized from a holding potential of -80 mV to -40 mV for 20 msec to inactivate fast Na+ currents. After this period, a further depolarization to 0 mV was performed for 180 msec to activate ICa(L). B, Concentration-dependent effects of verapamil () and fendiline () on peak ICa(L). The EC50 values amounted to 0.9 µM for verapamil and to 13 µM for fendiline. Data points represent mean values ± S.E.M. (n > 6 for each condition). C and D, Original recordings of ICa(L). The effect of verapamil (1 µM) was studied first; then, after wash-out for 10 min, the effect of fendiline (10 µM) was investigated in the same cell. Current traces under control and test conditions are superimposed. Peak ICa(L) was reduced within 2 min to about 59% of control by 1 µM verapamil (C) and to 62% of control by 10 µM fendiline (D). The decay of ICa(L) was not influenced by verapamil (1 µM; C) but accelerated by fendiline (10 µM; D).

Using Ba⁺⁺ as the charge carrier, the current (I_Ba(L)) through Ca⁺⁺ channels is much more pronounced than I_Ca(L), due to a larger conductance of the channels for Ba⁺⁺ ions and inhibition of the Ca⁺⁺-induced inactivation of the current (McDonald et al., 1994). The effects of fendiline (10 µM) were more pronounced if Ba⁺⁺ instead of Ca⁺⁺ was used as the charge carrier; the EC₅₀ values were 8 µM for the inhibition of I_Ba(L) and 13 µM for the inhibition of I_Ca(L) (data not shown). In addition, fendiline (10 µM) had an even greater influence on the voltage-dependent decay of the current (fig. 3A). ₁ amounted to 21 msec and ₂ to 121 msec under control conditions and to 14 and 69 msec, respectively, 3 min after the addition of fendiline. One explanation for the faster decay of the current in the presence of fendiline is an increasing block of the channels during the depolarizing voltage clamp pulse due to access or binding of the drug at 10 mV. The analysis of fractional current changes has allowed to determine the time constant of block development during the pulse (Jahnel et al., 1994). The block by fendiline of I_Ba(L) developed monoexponentially during the pulse to 10 mV ( = 96 msec) and was virtually complete within 300 msec (fig. 3B).

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Effect of fendiline on the time course of IBa(L). A, Original recording of IBa(L). Current traces under control conditions and in the presence of fendiline (10 µM) are superimposed. IBa(L) was activated by a depolarizing voltage pulse from -40 to +10 mV for 300 msec at 0.2 Hz as indicated by the voltage clamp pulse. Fendiline (10 µM) did not significantly influence the peak current value within 3 min but accelerated the decay of IBa(L). B, Fraction (1-IE/IC) of the current traces shown in A. (1-IE/IC) is plotted against the duration of the depolarizing voltage pulse. The fraction increased during the voltage pulse indicating a time-dependent block of IBa(L). The experimentally obtained values (circles) were fitted by a monoexponential function ( = 96 msec; correlation coefficient > 0.99; solid line). Similar results were obtained in two other experiments.

From these results, it was tentatively assumed that fendiline blocks open Ca⁺⁺ channels and therefore accelerates the decay of the current, in contrast to verapamil that had no influence on the decay and that has been shown to block channels in the inactivated state (Nawrath and Wegener, 1997). If this interpretation holds true, a condition that favors the open state of the channels should facilitate the effects of fendiline, but not of verapamil. In the presence of the Ca⁺⁺ agonist BayK8644, peak I_Ca(L) was increased, an effect due to favoring the transition of mode 1 into mode 2 of open calcium channels (Hess et al. 1984). Under these conditions, the effects of fendiline on I_Ca(L) were indeed greatly enhanced, whereas those of verapamil were reduced (fig. 4). In the presence of BayK8644 (10 µM), fendiline (10 µM) and verapamil (10 µM) depressed peak I_Ca(L) (in % of control) to 29 ± 8% (n = 6) and 86 ± 8% (n = 6), respectively. Without BayK8644, peak I_Ca(L) was reduced (in % of control) to 63 ± 6% (n = 6) by fendiline (10 µM) and to 15 ± 4% (n = 6) by verapamil (10 µM; fig. 2B). In addition, fendiline (10 µM), but not verapamil, accelerated the decay of I_Ca(L) in the presence of BayK8644 (fig. 4A and B).

View larger version (18K): [in this window] [in a new window]   Fig. 4.   Effects of fendiline and verapamil on ICa(L) in the presence of BayK8644. A and B, Original recordings of ICa(L) in the presence of BayK8644 (10 µM). Peak ICa(L) was reduced within 5 min to 33% of control by 10 µM fendiline (A) and to 79% of control by 10 µM verapamil (B). The decay of ICa(L) was not influenced by verapamil (B) but accelerated by fendiline (A). C, In the presence of BayK8644 (10 µM), mean values ± S.E.M. of peak ICa(L) (in % of control without drug) were 128 ± 13 (n = 6); the values were decreased to 29 ± 8 (n = 6) by fendiline (10 µM) and to 86 ± 8% (n = 6) by verapamil (10 µM). The effects of fendiline and verapamil were significantly different (P < .01), evaluated by unpaired Student's t test.

In patch clamp experiments on single channels, using Ba⁺⁺ as the charge carrier and in the presence of BayK8644, fendiline (1 µM) significantly reduced the overall open time of the single channel by 34% leading to a reduction of the ensemble average current by 43% at 0.1 Hz (fig. 5). The main effect of fendiline was a major change of the open time histogram; ₁ amounted to 0.8 msec and ₂ to 3.3 msec under control conditions and to 0.5 and 1.3 msec, respectively, 20 min after the addition of fendiline (fig. 5). On average, fendiline reduced the mean current from 0.56 ± 0.15 to 0.32 ± 0.09 pA (n = 5) resulting mainly from changes in mean open time of single channels (table 1); in addition, channel availability was slightly reduced. Unitary current amplitude, latency to first event, and the number of mean openings/record were not significantly affected. A similar reduction by fendiline (1 µM) in ensemble average current was also seen in multichannel recordings (not shown) that were not subject to further analysis.

View larger version (29K): [in this window] [in a new window]   Fig. 5.   Effect of fendiline on Ca++ channel activity in the presence of BayK8644 (1 µM) and using Ba++ ions as charge carrier. A, Single Ca++ channel fluctuations in the course of 10 consecutive voltage pulses are shown under control conditions (left column) and in the presence of fendiline (1 µM; right column). The channels were activated by depolarizing voltage steps from -80 to 0 mV at 0.1 Hz, as indicated by the voltage protocol. Channel openings appear as downward deflections. B, The open time distributions of Ca++ channel activity are shown under control conditions and in the presence of fendiline (1 µM). The number of channel openings (N) is plotted against the duration of the open time. Bin width was set to 0.2 msec, channel openings shorter than 0.2 msec were discarded. The open time distributions were fitted by two exponential functions. The time constants amounted to 1 = 0.8 msec and 2 = 3.3 msec under control conditions and to 0.5 and 1.3 msec, respectively, in the presence of fendiline (1 µM). C, Mean currents of Ca++ channel activity, averaged out of five experiments, are shown under control conditions and in the presence of fendiline (1 µM).

For comparison, the effects of verapamil (1 µM) were investigated under the same experimental conditions. Verapamil did not reduce the magnitude of the ensemble average current and, more specifically, did not modify the open time histogram (fig. 6). ₁ amounted to 0.3 msec and ₂ to 3.4 msec under control conditions and to 0.3 and 3.3 msec, respectively, 10 min after the addition of verapamil. Correspondingly, the kinetic characteristics of elementary Ca⁺⁺ currents remained virtually unchanged in the presence of this drug (table 2).

View larger version (31K): [in this window] [in a new window]   Fig. 6.   Effect of verapamil on Ca++ channel activity in the presence of BayK8644 (1 µM) and using Ba++ ions as charge carrier. A, Single Ca++ channel fluctuations in the course of 10 consecutive voltage pulses are shown under control conditions (left column) and in the presence of verapamil (1 µM; right column). The channels were activated by depolarizing voltage steps from -80 to 0 mV at 0.1 Hz, as indicated by the voltage protocol. Channel openings appear as downward deflections. B, The open time distributions are shown under control conditions and in the presence of verapamil (1 µM). The number of channel openings (N) is plotted against the duration of the open time. Bin width was set to 0.2 msec, channel openings shorter than 0.2 msec were discarded. The open time distributions were fitted by two exponential functions. The time constants amounted to 1 = 0.3 msec and 2 = 3.4 msec under control conditions and to 0.3 msec and 3.3 msec, respectively, in the presence of verapamil (1 µM). C, Mean currents of Ca++ channel activity are shown under control conditions and in the presence of verapamil (1 µM).

We have shown so far that the Ca⁺⁺ agonist BayK8644 can reverse the potency ratio of fendiline/verapamil in cardiomyocytes with respect to the effects on Ca⁺⁺/Ba⁺⁺ currents through Ca⁺⁺ channels. A similar reversal of the potency of these drugs, with respect to changes in tension, was also seen in aortic smooth muscle, depolarized by an increase in [K⁺]_o. High [K⁺]_o -induced tension in smooth muscle is thought to be mediated by a given proportion of open Ca⁺⁺ channels. At [K⁺]_o 48 mM, assuming [K⁺]_i of 150 mM, a resting potential of about -30 mV can be calculated. At this potential, 5 to 10% of the channels may be in the open and about 60% in the inactivated state, according to the position of the steady-state activation and inactivation curves of Ca⁺⁺ channels in smooth muscle (McDonald et al., 1994). Verapamil (1 µM) completely relaxed rat aortic rings previously contracted by [K⁺]_o 48 mM, whereas fendiline (1 µM) diminished tension by about 50% (fig. 7). BayK8644 (10 µM) increased high [K⁺]_o-induced tension by about 10%. Assuming that the effects of BayK8644 in smooth muscle are also mediated by inducing mode 2 of Ca⁺⁺ channels (Yatani et al., 1987), it was not surprising that the effect of verapamil (1 µM) was drastically reduced by BayK8644 (10 µM), whereas the effect of fendiline was significantly enhanced (fig. 7).

View larger version (13K): [in this window] [in a new window]   Fig. 7.   Effects of verapamil and fendiline on high [K+]o-induced tension in aortic rings in the absence and presence of BayK8644. Without Bay K8644, high [K+]o-induced tension was reduced to -3 ± 4% of control (n = 3) by verapamil (1 µM) and to 47 ± 3% of control (n = 3) by fendiline (1 µM). In the presence of Bay K8644 (10 µM), K+-induced tension was reduced to 73 ± 5% of control (n = 3) by verapamil (1 µM) and to 18 ± 10% of control (n = 3) by fendiline (1 µM). Under both conditions, the effects of verapamil and fendiline were significantly different (P < .01), calculated by unpaired Student's t test.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

This study shows that a Ca⁺⁺-antagonistic drug, fendiline, inhibits preferentially open Ca⁺⁺ channels. The evidence for this is derived from several findings. 1) Fendiline accelerated the decay of I_Ca(L) using either Ca⁺⁺ or Ba⁺⁺ ions as charge carriers that indicates development of block during depolarization. 2) The effects of fendiline were enhanced in the presence of BayK8644 that favors a conducting channel state characterized by long openings (Hess et al., 1984). 3) The effects of fendiline were stronger if Ba⁺⁺ instead of Ca⁺⁺ was used as the charge carrier. 4) Fendiline shortened the mean open time of single channels. In contrast, verapamil did not accelerate the decay of I_Ca(L), was less effective in the presence of BayK8644, and did not have an influence on the mean open time of single channels.

The differential effects of both drugs can explain that fendiline is a weak and verapamil a more potent Ca⁺⁺ antagonist in physiological conditions. This is related to the fact that, during repetitive activity in the heart and also in smooth muscle, the probability of Ca⁺⁺ channels to be in the conducting (open) state is relatively small (about 10%), compared with the probability to be in the closed state (about 90%; either in the resting or in the inactivated state). Consequently, fendiline is allowed only for a short time to bind to its receptor sites, depending on Ca⁺⁺ channel openings. In contrast, verapamil is bound more effectively in the inactivated state. The ratio of the open/closed times of Ca⁺⁺ channels can be drastically changed by BayK8644 (Hess et al., 1984; Kokubun and Reuter, 1984) resulting into large increases in I_Ca(L), accompanied by increases in force of contraction and smooth muscle tone. The open channel blocker fendiline is then given more time to bind to its receptor sites, and higher concentrations of the drug will be built up during repetitive depolarizations. Conversely, the inactivated state blocker verapamil is given less time to reach its receptor sites under these conditions, because inactivated (closed) times are reduced.

The described effects of fendiline and verapamil on I_Ca(L) were strictly dependent on the pulse protocol. The interpulse duration in single channel experiments was intentionally set to 10 sec to allow sufficient unbinding of the drugs at rest (-80 mV). Especially at shorter pulse intervals, the occurrence of blanks and changes in fast gating can occur due to uncomplete unbinding of the drugs previously bound by the preceding voltage pulse (Pelzer et al., 1985). Therefore, a first pulse can significantly determine the events during a second pulse, if the interpulse duration is significantly shorter than the time required for the unbinding of drugs from its receptor sites at rest. In the case of verapamil, a first pulse is indeed the prerequisite to observe an effect during a second pulse, because this drug is exclusively bound to the inactivated state. As a consequence, verapamil is ineffective at extremely low driving frequencies, as shown in this and in previous papers (Ehara and Kaufmann, 1978; Nawrath and Wegener, 1997). Our findings may also explain the earlier observations that fendiline exerted stronger effects in smooth muscle (Spedding and Berg, 1984) and in the myocardium (Schreibmayer et al., 1992) in the presence of BayK8644.

In the latter study, the increase in potency of fendiline in the presence of BayK8644 was explained by an allosteric interaction between fendiline and BayK8644 binding resulting in an inhibitory effect of the dihydropyridine agonist. Recently, it has been suggested that single amino acid residues may be involved in the formation of the distinct binding sites of different classes of drugs that may allow an allosterical interaction of drug binding (Hockerman et al., 1997).

We propose the following hypothesis for the action of verapamil and fendiline on Ca⁺⁺ channels. Ca⁺⁺ channels can fluctuate among the resting (R), open (O) and inactivated (I) state, determined by Hodgkin-Huxley kinetics (fig. 8). Whereas Ca⁺⁺ antagonists can principally bind to any of these channel states leading to their corresponding drug-bound states R*, O* and I*, fendiline binds preferentially to the open state and verapamil to the inactivated state leading to corresponding changes in the Hodgkin-Huxley kinetics. It would be most interesting to know whether these differences in drug action may also be related to different drug binding sites in the channel.

View larger version (16K): [in this window] [in a new window]   Fig. 8.   Simplified model of Ca++ channel states under control conditions and in the presence of calcium channel blockers. Under control conditions, voltage-dependent Ca++ channels exist in three main states during an excitation cycle: the resting (R), open (O), and inactivated (I) state. Calcium channel blockers may bind to each channel state resulting in drug-bound channel states: the resting (R*), open (O*) and inactivated (I*) channel state. As pointed out in the "Discussion," fendiline is assumed to bind mainly to the open channel state leading to O*, whereas verapamil is thought to bind preferentially to the inactivated state leading to I*. () Hodgkin-Huxley kinetics, () kinetics of drug binding/unbinding, (- - -) Hodgkin-Huxley kinetics, as modified by drug binding/unbinding.

It is in line with our observations that fendiline is less effective than verapamil also in clinical conditions (for review see Bayer and Mannhold, 1987). However, this situation may change when Ca⁺⁺ channel activity is increased either by the activation of the adenylyl cyclase/cAMP system or, more directly, by the therapeutic use of Ca⁺⁺ agonists such as Bay y 5959 for the treatment of heart failure (Bechem et al., 1997).