Abstract
Mibefradil is a tetralol derivative chemically distinct from other calcium channel antagonists. It is a very effective antihypertensive agent that is thought to achieve its action via a higher affinity block for low-voltage-activated (T) than for high-voltage-activated (L) calcium channels. Estimates of affinity using Ba2+ as the charge carrier have predicted a 10- to 15-fold preference of mibefradil for T channels over L channels. However, T channel IC50values are reported to be ∼1 μM, which is much higher than expected for clinical efficacy because relevant blood levels of this drug are ∼50 nM. We compared the affinity for mibefradil of the newly cloned T channel isoforms, α1G, α1H, and α1I with an L channel, α1C. In 10 mM Ba2+, mibefradil blocked in the micromolar range and with 12- to 13-fold greater affinity for T channels than for L channels (∼1 μM versus 13 μM). When 2 mM Ca2+ was used as the charge carrier, the drug was more efficacious; the IC50 for α1G shifted to 270 nM and for α1H shifted to 140 nM, 4.5- and 9-fold higher affinity than in 10 mM Ba. The data are consistent with the idea that mibefradil competes for its binding site on the channel with the permeant species and that Ba2+ is a more effective competitor than Ca2+. Raising temperature to 35°C reduced affinity (IC50 792 nM). Reducing channel availability to half increased affinity (∼70 nM). This profile of mibefradil affinity makes these channels good candidates for the physiological target of this antihypertensive agent.
Mibefradil, a tetralol derivative chemically distinct from other calcium channel antagonists, i.e., the dihydropyridines (e.g., nifedepine), the phenylalkylamines (e.g., verapamil), and the benzothiazapenes (e.g., diltiazem), has been reported to preferentially block T-type calcium channel currents in many tissues, including heart, brain, and vascular smooth muscle (Table 1). Early reports of mibefradil action were reported under its development name Ro 40-5967. It was briefly used clinically as Posicor but was withdrawn by Hoffmann-La Roche because of interactions of the drug with liver enzymes, i.e., cytochrome P450.
Until recently, the action of mibefradil on T-type calcium currents could only be studied in native cells, where usually it is small and difficult to separate from other inward calcium currents. Despite such difficulties, it has generally been agreed that mibefradil blocks T-type calcium currents at lower concentrations than it blocks other calcium currents (Table 1), although a considerable range of IC50 values has been reported, from a low of 130 nM in vascular smooth muscle (Clozel et al., 1997) to 4.7 μM in mouse spermatogenic cells (Arnoult et al., 1998), with estimates in cardiac myocytes at 1 to 2 μM (Benardeau and Ertel, 1998). The available data were obtained in a wide range of preparations, raising the possibility of channel isoform differences between cell types, and they also were gathered under disparate recording conditions, i.e., both Ca2+ and Ba2+ have been used as the charge carrier and concentrations have ranged from 5 to 30 mM.
The recent cloning of T-type calcium channels (Perez-Reyes et al., 1998, Cribbs et al., 1999) allows resolution of some of the source of these diverse affinities. In these experiments, we electrophysiologically determined IC50 values of mibefradil for α1G and α1H using both calcium and barium as the charge carrier. To easily make comparisons between such solutions, we held cells at sufficiently negative potentials and pulsed at sufficiently slow frequencies to guarantee full channel availability, conditions in the literature that are generally accepted to evaluate rested state blocking ability of the agent under study. In addition, we examined the change in efficacy of mibefradil at a temperature close to physiological (35°C) and evaluated the change in affinity of the drug when channel availability was reduced to half. Some of these data have been reported in abstract and review form (Martin et al., 1998b, Perez-Reyes et al., 1999).
Materials and Methods
Expression of Low-Voltage Activated and High-Voltage Activated Channels in Mammalian Cells.
T-type calcium channels (α1G, α1H, and α1I) were stably expressed in HEK293 cells using a calcium phosphate transfection protocol and G418 selection (600 or 1000 mg/ml for selection and 600 or 200 mg/ml for continued maintenance). L-type calcium channels, α1C with β2a, and α2δsubunits were stably expressed in HEK293 cells using a calcium phosphate transfection protocol and G418 selection (1000 mg/ml for selection and maintenance).
Electrophysiological Recordings and Data Analysis.
Currents were recorded using an Axopatch 200 (Axon Instruments, Inc., Foster City, CA) and pClamp data acquisition software (Axon Instruments, Inc.). Patch pipettes were constructed with aluminasilicate glass capillary tubes (0.8–1.5 MΩ). The pipette solution contained 130 mM KCl , 11 mM EGTA, 10 mM HEPES, 5 mM Mg2+-ATP, pH = 7.4. The bath solution contained 140 mM NaCl , 2 mM CaCl2, 10 mM HEPES, pH = 7.4. Mibefradil was diluted in the bath solution to the desired concentration (1 nM–10 mM) from a stock solution (1 mM in distilled water). Current measurements were made at 20–23 or 35–37°C. Temperature was controlled with a Sensortek thermocouple Peltier feedback system (TS-4; Physiotemp Instruments, Inc., Clifton, NJ). Currents were capacity corrected using 16 to 64 subthreshold responses (voltage steps of 10 or 20 mV) and leak subtracted, based on linear interpolation between the current at the holding potential and 0 mV. The effect of mibefradil was assessed using a voltage clamp protocol that stepped to −30 mV for 100 ms from a holding potential of −100 mV once every 5 s or once every 10 s. Because it is well known that channel gating is sensitive to the divalent ions and concentrations, a rather negative holding potential was selected to allow full channel availability under all experimental conditions. Summary data are presented as means ± standard error of the mean. Data were fit with appropriate equations using Matlab (Mathworks, Natick, MA), Prism (GraphPad, San Diego, CA), or SAS (Cary, NC), and fitted parameters are reported with standard errors of the estimate.
Results
Comparison of α1G, α1H, and α1I.
T channel clones were studied in stable cell lines established in HEK293 cells. Figure1 illustrates families of currents recorded in typical cells and mean current-voltage relationships for six cells of each type with 2 mM Ca2+ as the charge carrier. Time course of the currents (Fig. 1A) and time to peak of the currents (Fig. 1C) were similar to those reported for native T-type calcium currents and previous reports of these clones (Cribbs et al., 1998; Perez-Reyes et al., 1998; Lee et al., 1999). Current densities of these stable lines were similar (Fig. 1B), and also, as expected from the previous reports of these clones, peak current-voltage relationships were similar among the clones.
Block by Mibefradil.
To study block by mibefradil, we held cells at a sufficiently negative potential that under all ionic conditions studied, channels were fully available (−100 mV). Cells were lifted from the chamber bottom and transferred to a second chamber in which solutions with various concentration of drugs were flowing. Reversal of action was achieved by transferring the cell back to the drug-free chamber. In some experiments, drug was washed into the recording chamber. There were no statistical differences in the efficacy of mibefradil block determined in these two ways. Figure2 shows data from a typical cell, expressing α1H, studied with 2 mM Ca2+ as the charge carrier. Figure 2A depicts superimposed sample current traces obtained during mibefradil washin and washout with the time course of the change in peak current displayed below. The cell was depolarized to −30 mV, once every 10 s until current magnitude in drug stabilized. This low frequency of stimulation was chosen to avoid accumulation of channels in an inactivated conformation, which itself would reduce current magnitude but could also affect drug block if, as suggested by some investigators, drug affinity for inactivated channels was greater than for rested channels. Figure 2B illustrates the current-voltage relationships for this cell before, during, and after exposure to mibefradil. Block was not dependent on the test potential between −40 and +20 mV, indicating that, if state dependence of drug affinity is present, drug on- and off-rates were not rapid enough to be manifest during 100-ms depolarizations. Given this, we routinely evaluated mibefradil block at a test potential of −30 mV. Current magnitude returned almost to control after washout of the drug. The peak current-voltage data (Fig. 2B) also illustrate a small leftward shift in the current-voltage relationship that was often noted after drug washout. This shift in kinetics was not related to mibefradil; it occurred over time without an experimental intervention. Although such shifts in gating have not been commonly reported in the native T-type current literature, they have often been observed for voltage-gated sodium channels (Hanck and Sheets, 1992; Shcherbatko et al., 1999). The speed of the leftward shift in kinetic parameters could be retarded by inclusion of 5 or 10 mM Mg2+-ATP in the pipette (Zhang et al., 2000) and was usually less than 5 mV for the cells included in this study.
Comparison of α1G, α1H, and α1I to α1C.
Available data suggest that T channels have a higher affinity for mibefradil than L channels. We, therefore, compared drug efficacy under “standard” recording conditions, i.e., 10 mM Ba2+, for the three isoforms. L-type ICa was studied in HEK293 cells stably transfected with α1C, α2δ, and β2a. Figure 3 shows the fraction of current remaining after exposure to mibefradil concentrations between 100 nM and 100 μM. Averaged data were fit with a single site binding relation (eq. 1): blocked fraction = [mibefradil]/(IC50 + [mibefradil]). IC50 values for α1H, α1G, and α1I were indistinguishable at 1.1 ± 0.2, 1.2 ±0.2, and 1.5 ± 0.1 μM, respectively. In contrast, the IC50 for α1C (heart L channel) was 13-fold lower at 12.9 ± 1.3 μM. These values confirm that the cloned T channels indeed have a higher affinity for mibefradil than do L channels. IC50values for the three T channels were not statistically different from each other and, therefore, were similar to only a subset of those reported in the literature and were higher than others (Table1).
Block by Mibefradil in Physiological Calcium.
Experimenters often choose Ba2+ as the charge carrier for studying ICa, although T channels do not have the same preference for Ba2+ over Ca2+ as do high-voltage-activated calcium channels. We, therefore, measured IC50 values for the new clones with 2 mM Ca2+ as the charge carrier. Summary data are shown in Fig.4A. With physiological calcium, affinity of mibefradil was much greater. For α1H, the IC50 shifted 9-fold to 0.14 ± 0.2 μM, and for α1G, it shifted 4.5-fold to 0.27 ± 0.03 μM, causing α1H to display a 2-fold preference for mibefradil in physiological calcium, a feature that was not evident with 10 mM Ba2+ as the charge carrier. Note that at 10 nM mibefradil for both isoforms there was greater block than predicted by a single site dose-response relationship.
Relative Contribution of Charge Carrier and Concentration to the IC50 of Mibefradil.
Many drugs and toxins that block in the pore compete with permeant ions and, therefore, have IC50 values that are dependent on permeant concentration and/or charge carrier. The two dose-response relationships represent changes in permeant concentration, 2 and 10 mM, as well as changes in the charge carrier itself, Ba2+ or Ca2+. To evaluate the contribution of each of these to the IC50, we additionally evaluated block by 1 μM mibefradil in 2 mM Ba2+ and 10 mM Ca2+ (2–4 cells in each condition) and compared the IC50values the block predicted with those already determined. Figure5 illustrates graphically the differences with each by showing superimposed normalized data of the washin of 1 μM mibefradil. Changing from 2 mM Ca2+ to 2 mM Ba2+ increased the IC502.5-fold from 0.14 to 0.34 μM (reduced affinity), whereas increasing Ca2+ from 2 to 10 mM was only slightly more effective, increasing the IC50 2.9-fold to 0.41 μM. However, increasing Ba2+ concentration from 2 to 10 mM increased the IC50 almost 4-fold (0.34 to 1.3 μM), suggesting that Ba2+ competes more effectively with mibefradil than does Ca2+.
Effect of Holding Potential on Block.
Dependence of mibefradil block on holding potential is controversial in the literature (seeDiscussion). To examine this for the cloned T-type channels, we evaluated block at −80 mV, a potential at which half the channels were inactivated. To minimize cell-to-cell variability, the midpoint of steady-state inactivation was determined by comparing the currents measured from a range of holding potentials to that obtained at −110 mV. A typical result using α1H is shown in Fig.6A. In this cell, the peak current measured at a holding potential of −110 mV was −872 nA, and this current was reduced 52% by switching the holding potential to −83 mV. Addition of submicromolar doses of mibefradil caused a rapid inhibition with little effect on the kinetics (Fig. 6A; also see Fig. 2). The average dose-response analysis from eight cells is shown in Fig. 6B. Fits to the data with a single site dose-response relationship (eq. 1) yielded an IC50 of 69 ± 23 nM. In these experiments inactivation averaged 55 ± 4%, and the holding potential was −81 ± 1 mV. Recovery of the current on washout of the drug at this potential was very slow, making it difficult to achieve precontrol current levels. However, recovery was significantly faster and more complete (82 ± 5% of control, n= 6) if the holding potential was hyperpolarized to −100 or −110 mV.
Block by Mibefradil near Physiological Temperatures.
The issue of how to compare data under voltage clamp to efficacy seen in animals requires that not only should physiological ion concentrations be used but temperatures at least close to physiological should be examined. Only one previous study used a temperature near physiological (Table 1;McDonough and Bean, 1998). There is no a priori way to predict the effect of temperature on block, and, therefore, we examined the question experimentally. Figure 7summarizes data from these experiments. Panel A shows typical data for a cell expressing α1H, recorded at 23 and 35°C. Panel B shows grouped data summarizing the peak current-voltage relationships and time to peak of the currents as a function of potential. As one would expect from studies on native T channels (Nobile et al., 1990), current magnitude increased dramatically at the higher temperature and current kinetics, both turn-on and decay were accelerated. Panel C shows washin and washout of 1 μM mibefradil at the two temperatures and the dose-response data for the two temperatures. The IC50 at 35°C was 792 ± 127 nM, an increase of 5-fold over that measured at room temperature.
Discussion
Our results show that mibefradil has a greater affinity for the cloned T channel isoforms α1G, α1H, and α1I over a cloned L channel, α1C. When 10 mM Ba2+ was used as the charge carrier, the IC50 values for mibefradil were in the micromolar range for α1G, α1H, α1I and 12- to 13-fold higher for α1C. When 2 mM Ca2+ was used as the charge carrier, the affinity of mibefradil increased 4.5-fold for α1G and 9-fold for α1H. A lesser affinity for drug was observed by Mehrke et al. (1994) in native T currents in human medullary thyroid carcinoma cells (hMTC); where KD decreased from 2.7 to 0.7 μM (affinity increased) when the permeant species was changed from Ba2+ to Ca2+. One straightforward explanation then for the disparate data in the literature is that mibefradil competes with the permeant species for its binding site on the channel. The change in block that occurred when Ba2+ concentration was reduced from 10 to 2 mM, i.e., the IC50 for mibefradil decreased 4-fold, is consistent with this interpretation. Similarly, when the Ca2+ concentration was reduced from 10 to 2 mM, the IC50 for mibefradil decreased 2.5-fold. These data taken together with the higher IC50 obtained in Ba2+ compared with Ca2+suggest that Ba2+ is a more effective competitor. It is the case, however, that channel kinetics is dependent on the divalent ion present and the concentration, so that it is formally possible that changes in kinetics, i.e., differences in distribution of closed and inactivated channels under the different ionic conditions, were responsible. However, the experimental conditions, negative holding potential and slow pulse rate, were chosen to minimize contributions of this sort, so it is more reasonable that the results reflect competition between drug and the permeant cations.
Most early studies reported use dependence of mibefradil block both in native preparations expressing T currents as well as other calcium currents. Use dependence usually implies voltage and/or state dependence of drug binding/unbinding, but studies addressing such voltage/state dependence have yielded conflicting results. For example,Mehrke et al. (1994) found no voltage or use dependence in T channels from hMTC cells and concluded that mibefradil bound to the rested state of the channel. Data of Klugbauer et al. (1998), who studied T current in hMTC cells, also supported a primary role for rested state block. In addition, Mishra and Hermsmeyer (1994a) concluded, based on experiments in rat vascular muscle, that mibefradil blocked the rested state.
In contrast, use-dependent block was observed for the high-voltage activated channel α1Cb expressed in Chinese hamster ovary cells, and Welling et al. (1995) suggested that for that isoform, mibefradil preferentially binds to an active state of the channel. Similarly, in native T channels in guinea pig atrial myocytes (Benardeau and Ertel, 1998), bovine adrenal glomerulosa cells (Rossier et al., 1998), and rat dorsal root ganglion cells (Todorovic and Lingle, 1998) mibefradil showed use dependence.
Direct evidence for greater drug affinity in high-voltage activated channels for inactivated conformations comes from Bezprozvanny and Tsien (1995), who noted both inactivated and open channel mibefradil block of recombinant high-voltage activated channels (α1C, α1B, α1A, and α1E) expressed in Xenopus oocytes and byJiménez et al. (2000) in high-voltage activated channels expressed in mammalian cells. McDonough and Bean (1998) found a dramatic increase in mibefradil affinity when the holding potential was reduced in T channels from rat cerebellar Purkinje neurons. Their results indicated that the increase in affinity can be partly explained by a decreased off-rate of the drug from the inactivated state. Similarly, Benardeau and Ertel (1998) reported a dramatic increase in drug efficacy at depolarized potentials for T currents in guinea pig atrial myocytes.
In this study, we observed no sensitivity of mibefradil to the test potential. These data do not rule out state dependence of drug action, but they do say that, if state dependence is present, the on- and off-rate of drug is sufficiently slow that changes in drug binding do not appreciably occur over 100 ms (the duration of the depolarizations). We did observe an increased affinity of the drug for α1H channels when the holding potential was reduced from −100 to −83 mV; the IC50 decreased 2-fold when channel availability was reduced to half. This suggests that mibefradil has a higher affinity for the inactivated state of the channel. Such a modest effect of holding potential could easily have been missed in earlier experiments where inactivated channel block was not observed. However, the increased affinity of drug was much less than that observed in rat cerebellar Purkinje neurons. Although only three isoforms of T channels have been found, there is quite a bit of alternative splicing that occurs (e.g., Mittman et al., 1999), and it is possible that differences observed in various native preparations could reflect different expression patterns of splice variants. Another possible cause for the differences observed could relate to experimental conditions. T channel clones exhibit complex development and recovery from inactivation (Martin et al., 1998a, 1999). Our experimental conditions were chosen to minimize accumulation in slow inactivated conformations (slow pulsing rates). It remains to be investigated whether various inactivated conformations preferentially bind the drug. Such studies will undoubtedly be helpful in establishing the physiological effects of mibefradil because in many cells resting potentials are such that T channels are partially inactivated.
When the affinity of mibefradil was determined at physiological temperature, it was found to decrease affinity 5-fold relative to room temperature. This result might suggest that mibefradil acts as an open channel blocker because at the elevated temperature the channel is open for a shorter period of time, especially given that there is some evidence in the literature that mibefradil acts as an open channel blocker, i.e., of native T-type current expressed in mouse spermatogenic cells (Arnoult et al., 1998) and of α1A channels expressed in Xenopus oocytes (Aczel et al., 1998). However, this explanation seems somewhat unlikely because the time spent in the rested and/or inactivated state dominated the experimental design so that there was only a very small change in the proportion of the time the channel spent in the open state at the higher temperature. Nonetheless, the 5-fold reduction in IC50 with only a 12°C change in temperature supports the idea of state dependence in that this Q10 is much higher than would be predicted for a purely physical effect (i.e., diffusion or static protein-protein interactions).
Cloning of the T-type calcium channel isoforms has simplified their electrophysiological and pharmacological study. These data suggest that mibefradil is a potent inhibitor of these channels. Because of possible interaction between the permeant species and the drug being investigated, it would seem prudent to take such effects into consideration especially when it is not possible to study drug action with physiological ion concentrations and near physiological temperature. This would allow for a more direct correlation with clinical studies.
Footnotes
-
Send reprint requests to: D. A. Hanck, Ph.D., Cardiology (MC6094), University of Chicago, 5841 South Maryland Ave., Chicago, IL 60637. E-mail: d-hanck{at}uchicago.edu
-
↵1 This research was supported by the National Heart, Lung and Blood Institute, National Institutes of Health Grants HL-PO1-20592 (D.A.H.) and HL-58728 (E.P.-R.) and by a grant-in-aid from the American Heart Association to D.A.H.
- Abbreviations:
- hMTC
- human medullary thyroid carcinoma
- Received March 13, 2000.
- Accepted June 30, 2000.
- The American Society for Pharmacology and Experimental Therapeutics