Introduction

Top Summary Introduction Materials & Methods Results Discussion References

Cystic fibrosis is an autosomal-recessive disease that results from mutations in the gene encoding the CFTR (Anderson et al., 1992; Riordan, 1994). In epithelial cells from a variety of tissues, CFTR functions as a PKA-activated Cl channel (Kartner et al., 1991; Anderson et al., 1991b). In recent years, a cAMP-dependent Cl channel also has been identified in guinea pig (Bahinski et al., 1989; Harvey and Hume, 1989a; Ehara and Ishihara, 1990; Matsuoka et al., 1990), rabbit (Harvey and Hume, 1989b), cat (Zhang et al., 1994), simian (Warth et al., 1996), and human (Warth et al., 1996) cardiac ventricular myocytes. The whole-cell and single-channel current properties of the cardiac channel display strong similarity to those of the epithelial CFTR channel (Anderson et al., 1992; Riordan, 1994). Furthermore, with the exception of a deletion in 30 amino acids coded by exon 5 of the CFTR gene, the rabbit cardiac gene shares >90% homology with CFTR (Horowitz et al., 1993; Hart et al., 1996). Expression of the cloned cardiac CFTR channel in Xenopus laevis oocytes results in the appearance of PKA-activated Cl currents (Hart et al., 1996). Thus, cardiac tissue expresses an alternatively spliced variant of the CFTR channel.

Ion channel modulators, such as dihydropyridine Ca²⁺ channel antagonists and quaternary ammonium K⁺ channel blockers, have proved useful in elucidating channel gating mechanisms and mapping ion-permeation pathways. Although the CFTR Cl channel has been widely studied in both human epithelial tissues and heterologous cells expressing the CFTR gene (Anderson et al., 1992; Riordan, 1994), little quantitative data are available concerning the pharmacology of this channel. Previous studies of CFTR have focused on the compound DPC, which blocks the CFTR channel when applied at relatively high concentrations (200 µM to 3 mM) (Anderson et al., 1991b, 1992; McCarty et al., 1993). Greger and colleagues modified the structure of DPC to produce a group of arylaminobenzoate compounds that varied in the phenyl-to-benzoate group, carbon chain length (Wangemann et al., 1986). One of these arylaminobenzoates, NPPB, was identified as a potent blocker of Cl channels in the TAL of the kidney (Wangemann et al., 1986; Tilmann et al., 1991). However, the effects of NPPB and other arylaminobenzoates have not been studied in detail on the CFTR Cl channel. Thus, the goal of the current study was to quantify the blocking action of arylaminobenzoates on the cardiac cAMP-dependent Cl current.

	Materials and Methods

Top Summary Introduction Materials & Methods Results Discussion References

Isolation of cardiac myocytes. An enzymatic dissociation procedure based on that of Mitra and Morad (1985) was used to isolate the myocytes. Briefly, hearts were removed from adult guinea pigs (200-300 g), mounted on a Langendorf-type column, and perfused for 10 min with a Ca²⁺-free Tyrode's solution containing collagenase (0.25-0.32 units/ml) (type B; Boehringer-Mannheim Biochemicals, Mannheim, Germany) and protease (0.2 mg/ml) (type 14 or 25; Sigma Chemical, St. Louis, MO). After 10 min of perfusion with 0.2 mM Ca²⁺-containing Tyrode's, the heart was dissected into small pieces and single cells obtained by gentle agitation. Cells were stored at room temperature (22-25°) in normal Tyrode's solution (see below) and used 1-10 hr after isolation.

Recording procedure. The patch-clamp method (Hamill et al., 1981) was used to record whole-cell ventricular currents using PC-501 (Warner Instrument, Hamden, CT), L/M EPC-7 (Adams and List, Westbury, NY), and Axopatch 200 (Axon Instruments, Foster City, CA) amplifiers. Pipettes were made from Gold Seal Accu-fill 90 Micropets (Clay Adams, Parsippany, NJ) and had resistances of 1-3 M when filled with internal solution. Series resistance was determined by measuring the time constant of the capacity current and the membrane capacitance. For a typical set of experiments, series resistance ranged from 5.2 to 7.9 M in the guinea pig ventricular myocytes with a mean ± standard error of 6.2 ± 0.2 M (18 cells). Typically, > 50% of the series resistance could be compensated electronically. Membrane currents were measured with 12-bit analog/digital converters (Scientific Associates and Axon Instruments). Data were sampled at 10 KHz, filtered at 2-3 KHz with a low-pass Bessel filter (Frequency Devices, Haverhill, MA), and stored using personal computers [Northgate (Edn Prairie, MN) and Dell (Austin, TX)].

Measurement of the cardiac I_Cl. Isolated cells were initially placed in a normal Tyrode's solution consisting of 132 mM NaCl, 5 mM KCl, 2 mM MgCl₂, 1 mM CaCl₂, 5 mM dextrose, and 5 mM HEPES, pH 7.4. After establishment of the whole-cell voltage-clamp, the solution was changed to a K⁺-free solution containing 140 mM NaCl, 1 mM MgCl₂, 1 mM CaCl₂, 5 mM dextrose, 5 mM HEPES, and 1 mM BaCl₂, pH 7.4 (total [Cl] = 146 mM) (osmolarity, 280 mOsM). I_Ca was eliminated by the addition of 200-500 nM nisoldipine (Miles Laboratories, West Haven, CT) to the external solution. I_Na was eliminated by holding the membrane potential at 40 mV and adding 10 µM tetrodotoxin to the bath. Patch electrodes were filled with a pipette solution consisting of 70 mM CsCl, 40 mM Cs-aspartate, 2 mM MgCl₂, 1 mM CaCl₂, 11 mM EGTA, 5 mM ATP (K⁺ salt), and 10 mM HEPES, pH 7.3 (total [Cl] = 76 mM) (osmolarity, 280 mOsM). The ratio of EGTA/CaCl₂ in these solutions sets the free intracellular Ca²⁺ concentration to ~10 nM (Fabiato, 1988).

Expression of CFTR in X. laevis oocytes. The cDNA for the human epithelial CFTR channel was generously supplied by Dr. Alan Smith (Genzyme, Cambridge, MA), and CFTR transcripts were prepared using the mMessage mMachine kit (Ambion, Austin, TX). Stage V and VI oocytes were injected with 50 nl of cRNA (0.1-0.2 mg/ml) using a microinjector (Drummond Scientific, Broomall, PA). CFTR Cl currents were measured 1-3 days after injection using a TEV 200 two microelectrode voltage clamp (Dagan, Minneapolis, MN). The oocyte bathing solution consisted of 96 mM NaCl, 2 mM KCl, 1 mM MgCl₂, and 5 mM HEPES, pH 7.4. CFTR currents were activated by application of a cAMP cocktail containing forskolin (10 µM), 8-chlorphenylthio cAMP (500 µM) and 3-isobutyl 1-methylxanthine (1 mM).

Preparation and use of the arylaminobenzoate compounds. The structure of the drugs BNBA, NPEB, NPPB, and NPBA are shown in Fig. 1. These arylaminobenzoates were generously supplied by Dr. Rainer Greger (Albert-Ludwigs University, Freiburg, Germany). The drugs differ in the length of the carbon chain between the benzoate (with carboxyl group) and phenyl rings. The pK_a of NPPB was determined titrimetrically to be ~4.5. This is consistent with a previous reported calculation (Wangemann et al., 1986). The percentage of charged drug (C) was calculated using the equation:

(1)

where K_a is the acid dissociation constant for the drug molecule. Stock solutions of the drugs were prepared in 100% DMSO and diluted into the external solution so the final volume of DMSO was 0.1%.

View larger version (18K): [in this window] [in a new window]   Fig. 1.   Structure of arylaminobenzoates used in this study.

(2)

	Results

Top Summary Introduction Materials & Methods Results Discussion References

Arylaminobenzoate block of the cardiac cAMP-dependent Cl channel. Fig. 2 (top) displays whole-cell background currents recorded from a cardiac ventricular cell during 100-msec voltage steps applied to various potentials. The external application of 2 µM forskolin caused the activation of a time-independent, outward-rectifying Cl current when recorded with a Cl concentration of 76 mM in the pipette and 146 mM in the bath (E_Cl = 16 mV). The physiological properties of this cardiac I_Cl have been described previously (Bahinski et al., 1989; Harvey and Hume, 1989a; Matsuoka et al., 1990; Walsh and Long, 1992). As shown in Fig. 2, the addition of a 20 µM concentration of NPPB caused a strong reduction in both inward and outward I_Cl. Overall in six myocytes examined, 20 µM NPPB decreased the inward (at 90 mV) and outward (at +60 mV) I_Cl by 94 ± 4% and 91 ± 4%, respectively. Partial recovery from NPPB block could be observed during long periods of drug washout (5-10 min). In three washout experiments, I_Cl recovered to within 16 ± 10% of the initial amplitude measured in the absence of the drug. A more rapid and complete recovery from drug block was obtained in experiments with lower concentrations of NPPB (see Fig. 9).

View larger version (23K): [in this window] [in a new window]   Fig. 2.   Block of the cardiac ICl by NPPB. Top, whole-cell background currents recorded during 100-msec voltage steps applied from a holding potential of 40 mV to 80 mV through +60 mV in 20-mV increments. Currents were measured under basal conditions (CON) and after the addition of 2 µM forskolin (FORS). The forskolin-activated ICl was inhibited by 100% and 96%, respectively, at 80 and +60 mV by a 20 µM concentration of NPPB. Cell AB4-368. Bottom, current-voltage relationship for forskolin-sensitive ICl obtained in the presence and absence of 20 µM NPPB. Values are mean ± standard error for six experiments. Values for NPPB were obtained at 5 min after addition of the drug to the bath; this represents a time period when block of ICl by NPPB had saturated (see Fig. 4).

Voltage-dependence of arylaminobenzoate block. The expressed epithelial CFTR Cl channel is blocked by the arylaminobenzoate DPC (200 µM) in a strongly voltage-dependent manner (McCarty et al., 1993; McDonough et al., 1994). Because voltage-dependent block of the cardiac channel was not evident in the initial experiments with 20 µM NPPB (Fig. 2), it was determined whether voltage-dependent block might be observed with lower concentrations of the drug. The application of 2 µM NPPB produced a clear voltage-dependent block of the channels with > 30% reduction in the inward current and no decrease in the outward current (Fig. 3). The electrical distance sensed by the NPPB molecule, , was determined using Woodhull analysis (Woodhull, 1973) as described in Materials and Methods. An average  of 32 ± 3% (four myocytes) was calculated with the orientation from the inside membrane. A similar value ( = 28 ± 5%) was measured during internal dialysis of NPPB. Relatively low concentrations of other arylaminobenzoates also blocked I_Cl in a voltage-dependent manner. For example, the drug BNBA (25 µM) blocked the inward current (at 90 mV) by 34 ± 7% but inhibited the outward current (at +60 mV) by only 4 ± 4%. In four experiments, an average  of 31 ± 2% was determined with this drug concentration.

View larger version (15K): [in this window] [in a new window]   Fig. 3.   Voltage-dependent block of the cardiac ICl by NPPB. Left, forskolin (FORS)-activated ICl recorded during 100- msec voltage steps applied to either 90 or +60 mV in the presence and absence of 2 µM NPPB. ICl was inhibited by 32% and 0%, respectively, at 90 and +60 mV. Right, current-voltage relationship for ICl shown on the left. Cell AB4-C08.

View larger version (14K): [in this window] [in a new window]   Fig. 4.   Time course of NPPB inhibition of the cardiac ICl. Left, forskolin-activated ICl recorded during 100-msec voltage steps applied to either 90 or +60 mV both before (FORS) and 30 sec after the addition of 20 µM NPPB. Right, time course for inhibition of ICl by NPPB measured at either 90 or +60 mV. Percentage inhibition is relative to that obtained just before NPPB addition (time = 0). Cell AB4-452. In separate experiments, the solution exchange "dead space" for the chamber used in these experiments was determined by measuring the block of the inward rectifier K+ current by 1 mM BaCl2. The "dead space" value was 2-10 sec.

Effect of increasing carbon chain length on arylaminobenzoate potency. Previous studies have shown that increasing the carbon chain length between the benzoate and phenyl rings of the arylaminobenzoates increases the potency for Cl current inhibition in the thick ascending limb of the kidney (Wangemann et al., 1986). To determine whether a similar order of potency exists for block of the cardiac cAMP-dependent channel, we tested the effect on I_Cl of the drugs shown in Fig. 1. Fig. 5 displays concentration-versus-inhibition curves for the drugs BNBA, NPEB, NPPB, and NPBA. Because block of I_Cl by the arylaminobenzoates was voltage dependent, especially at low drug concentrations (see Fig. 3), all the data displayed in Fig. 5 were recorded at 90 mV. Increasing the carbon chain length produced a clear increase in the potency of arylaminobenzoate block of the cardiac cAMP-dependent channel, with IC₅₀ values of 47 µM, 17 µM, and 4 µM for BNBA (one carbon atom), NPEB (two carbon atoms), and NPPB (three carbon atoms), respectively. Increasing the carbon chain further with NPBA (four carbon atoms) caused no additional increase in drug potency over that of NPPB (IC₅₀ for NPBA = 4 µM).

View larger version (16K): [in this window] [in a new window]   Fig. 5.   Concentration-versus-response curves for inhibition of ICl by the arylaminobenzoates. Values are mean ± standard error of three to six experiments. The theoretical curves for the data are given by [drug]/(IC50 + [drug]) with a IC50 values of 47, 17, 4, and 4 µM for BNBA, NPEB, NPBA, and NPPB, respectively, providing the best least-squares fit to the points.

pH dependence of arylaminobenzoate block. All of the drugs shown in Fig. 1 contain a negatively charged carboxyl group that causes the pK_a of the compounds to be < 5. The pK_a of NPPB was determined to be close to 4.5, and thus according to eq. 1, the drug molecules are predominately charged (>99%) in external solution at pH 7.4. Increasing the pH of the solution to 10.0 should not significantly alter the amount of charged to neutral form of the drug (100% of NPPB molecules are ionized at pH 10.0). It was surprising to find, therefore, that block of the cardiac I_Cl by NPPB was almost completely abolished when the pH of the external solution was increased to 10.0 (Fig. 6). Even at an NPPB concentration of 50 µM, which produces >95% block at pH 7.4, there was only a small and nonsignificant decrease in the current at pH 10.0 (10 ± 6%, four cells, p > 0.2). However, a strong and rapid block of the current could be induced at this time by replacing the external solution with a drug-containing external solution buffered to pH 7.4 (results not shown), as expected from the previous experiments (see Fig. 2). Increasing the internal solution pH also decreased the block of the I_Cl by NPPB. At a concentration of 20 µM, NPPB caused a 94 ± 4% (n = 4) decrease in the current at pH 7.3 and 61 ± 6% (five cells) decrease at pH 8.4. This difference was statistically significant (p < 0.05).

View larger version (16K): [in this window] [in a new window]   Fig. 6.   Increasing extracellular pH prevents the block of ICl by NPPB. Left, current-voltage relationship for ICl measured in the presence and absence of 50 µM NPPB in external solution at pH 10.0. Cell AB4-606. FORS, forskolin. Right, concentration-versus-response curve for inhibition of ICl by NPPB. Data at pH 7.4 are taken from Fig. 5. For the data at pH 10.0, values are mean ± standard error of three or four experiments.

View larger version (16K): [in this window] [in a new window]   Fig. 7.   Decreasing extracellular pH augments the block of ICl by NPPB. Left, current-voltage relationship for ICl measured in the presence and absence of 1 µM NPPB in external solution at pH 6.4. Cell AB4-651. FORS, forskolin. Right, concentration-versus-response curve for inhibition of ICl by NPPB. Data at pH 7.4 are taken from Fig. 5. For the data at pH 6.4, values are mean ± standard error of three to five experiments. The theoretical curve for the data at pH 6.4 provided an IC50 value of 0.6 µM.

Internal block by NPPB. The results displayed in Figs. 2-7 indicate that arylaminobenzoate block of the cardiac cAMP-dependent Cl channel occurs preferentially at negative membrane potentials and is enhanced by lowering the pH of the external solution. One explanation for these results would be that the drugs act primarily through a membrane-accessible pathway. Decreasing the external pH might enhance the movement of the drugs into the membrane by altering charged groups on the drug. Once inside the cell, block by the charged drug molecules (at the internal pH 7.3) would be favored at negative membrane potentials. To test this model, NPPB was added to the patch pipettes and dialyzed into the myocytes.

View larger version (20K): [in this window] [in a new window]   Fig. 8.   Block of ICl by internally applied NPPB. Current-voltage relationships for ICl measured during dialysis with internal solution containing DMSO (FORS) or 10 µM NPPB. In each experiment, a 10-12-min period was allowed for dialysis after breakthrough into the whole-cell configuration. After this dialysis period, the cells were exposed to 2 µM forskolin, and the peak current was measured. Results were obtained in external solution at pH 7.4 (left) or 10.0 (right). Values are mean ± standard error of three or four experiments. There was a significant difference (p < 0.02) between the control and NPPB data. In contrast, there was no significant difference (p > 0.5) between the NPPB values shown on the left and right.

View larger version (13K): [in this window] [in a new window]   Fig. 9.   Influence of external pH on the recovery of ICl block by NPPB. ICl was blocked by 5 µM NPPB at pH 7.4, and the time course of recovery determined during washout with external solution at pH 7.4 (left) and 10.0 (right). In each experiment, the currents recorded at 90 mV were normalized to the control data measured before NPPB application. Time = 0 represents the last point taken before starting NPPB washout. The data were fit with a single exponential function with time constants of 2.9 (pH 7.4) and 3.3 (pH 10.0) min. Cells AB4-841 and AB4-808.

NPPB block of the epithelial CFTR channel. NPPB was found to block the cardiac I_Cl in a voltage- and pH-dependent manner (Figs. 3, 6, and 7). Because the cardiac I_Cl is believed to represent an alternatively spliced variant of the CFTR Cl channel (Horowitz et al., 1993; Hart et al., 1996), it was important to determine whether NPPB has similar actions on the epithelial form of the channel. Fig. 10 displays the results of experiments in which the effect of NPPB was determined on the human epithelial CFTR channel expressed in X. laevis oocytes. At a concentration of 50 µM, externally applied NPPB caused a voltage-dependent block of the CFTR current (Fig. 10, left) with a 25 ± 2% decrease measured at 90 mV, and a 3 ± 2% change at +60 mV (four oocytes). As was the case with the cardiac I_Cl, block of the CFTR current could be abolished completely when the pH of the external solution was increased to pH 10 (Fig. 10, right) (percentage change at 90 and +60 mV = 1 ± 3% and 6 ± 2%, respectively; five oocytes).

View larger version (15K): [in this window] [in a new window]   Fig. 10.   Block of the human epithelial CFTR Cl channel by NPPB. Left, current-voltage relationship for CFTR current expressed in a X. laevis oocyte and measured in the presence and absence of 50 µM NPPB in external solution at pH 7.4. In this experiment, NPPB caused a 30% decrease in the current measured at 90 mV. Oocyte 8202. CON, control. Right, current-voltage relationship for the CFTR current measured in the presence and absence of 50 µM NPPB in external solution at pH 10.0. No change was recorded in the current measured at 90 mV. Oocyte 8222.

	Discussion

Top Summary Introduction Materials & Methods Results Discussion References

Arylaminobenzoate block of the cardiac cAMP-dependent channel. The goal of the current study was to begin quantifying the structural requirements necessary for drug block of the cardiac cAMP-dependent Cl channel. This was accomplished by studying the effect of a group of arylaminobenzoate compounds that varied in the phenyl-to-benzoate ring, carbon chain length. Our approach was based on previous findings that the arylaminobenzoate compound DPC blocks both the epithelial (Anderson et al., 1992; McCarty et al., 1993) and cardiac (Walsh and Wang, 1996) isoforms of the CFTR channel. McCarty et al. (1993) found that DPC and the structurally related chemical flufenamic acid (at 200 µM concentrations) produce a ~40% block of the CFTR Cl current in X. laevis oocytes expressing the human epithelial channel. Single-channel analysis suggested that DPC is an open channel blocker of CFTR that permeates through the membrane to reach its binding site (McCarty et al., 1993). In guinea pig ventricular myocytes, DPC reduces the whole-cell I_Cl with an IC₅₀ value of 270 µM (Walsh and Wang, 1996).

Voltage and pH dependence of arylaminobenzoate block. Arylaminobenzoate block of the cardiac cAMP-dependent current occurred in both a voltage- and pH-dependent manner. Voltage-dependent block of the epithelial CFTR channel has been reported for DPC and flufenamic acid ( = 41% from the inside) (McCarty et al., 1993), as well as glutamate ( = 31%) and gluconate ( = 30%) (Linsdell and Hanrahan, 1996). In the current study, voltage-dependent block was most evident using relatively low concentrations of the drugs or immediately after the application of higher concentrations of the compounds. NPPB has length of ~16 Å (Wangemann et al., 1986), yet it blocked the I_Cl with the same  (~30%) as BNBA, which has a length of ~12 Å. This implies that although these two compounds are of different size, they act at a similar site ~30% within the electric field. This apparent discrepancy can be reconciled if the lipophilic phenyl ring of the compounds penetrates to varying amounts into the lipid membrane (see above). This might allow the negatively charged carboxyl group on each of the arylaminobenzoates to interact at a similar binding site either on or near the channel. For the epithelial CFTR channel, a serine at position 341 represents at least one hypothetical binding site for the carboxyl group of DPC (McDonough et al., 1994). Mutations at this site reduce the potency of DPC block of the channel (McDonough et al., 1994). If the phenyl ring is involved in docking the compounds within the lipid membrane, substitutions on this ring, which reduce the lipophilicity, may also reduce the drug potency.

Relevance of the study to cardiac pharmacology. A -adrenergic-activated Cl channel has been identified in guinea pig (Bahinski et al., 1989; Harvey and Hume, 1989a; Ehara and Ishihara, 1990; Matsuoka et al., 1990), rabbit (Harvey and Hume, 1989b), cat (Zhang et al., 1994), and simian (Warth et al., 1996) cardiac ventricular cells. Molecular analysis strongly suggests that these cells express an alternatively spliced variant of the CFTR channel (Horowitz et al., 1993; Hart et al., 1996). Although an initial report suggested that CFTR is expressed in human cardiac tissue (Levesque et al., 1992), other studies have failed to identify a cAMP-dependent Cl current in human cardiac myocytes (Oz and Sorota, 1995; Sakai et al., 1995). If present in the human heart, the CFTR channel could play an important role in controlling the action potential duration. Due to the properties of the cardiac CFTR current-voltage relationship, activation of this channel during -adrenergic stimulation will cause a decrease in the duration of the cardiac action potential and shorten the QT interval of the electrocardiogram. Block of the I_Cl by the Cl channel blocker anthracene-9-carboxylate increases the duration of the action potential in guinea pig ventricular myocytes during exposure to isoproterenol (Levesque et al., 1993). By lengthening the cardiac action potential duration, future arylaminobenzoate derivatives might represent a new and unique group of class III antiarrhythmic agents that would be effective during sympathetic stimulation. The action of other putative CFTR channel blockers, such as sulfonylureas (Sheppard and Welsh, 1992; Venglarik et al., 1996) and clofibric acid analogues (Walsh and Wang, 1996), also will require careful attention.