Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The epithelial sodium channel (ENaC) is a heteromeric channel made of three subunits, , , and , and localized in the apical membrane of epithelial cells (Garty and Palmer, 1997). Transport across the apical membrane through the sodium channel is the rate-limiting step in sodium reabsorption by the epithelial cells of the distal nephron, distal colon, and in airways. ENaC thereby plays a key role in the regulation of the sodium balance, extracellular fluid volume, and blood pressure by the kidney and in the controlled fluid reabsorption in the airways. The activity of ENaC is regulated by several hormones such as aldosterone and vasopressin; intracellular and nonhormonal extracellular factors also modulate this activity, but the mechanisms of these regulations are not yet completely understood (Garty and Palmer, 1997). It also has been proposed that ENaC, like other channels, may be regulated by a factor secreted by an associated ATP-binding cassette (ABC) protein, such as cystic fibrosis conductance regulator (CFTR) or the K channel-associated sulfonylureas receptor (SUR) (Al-Awqati, 1995). In this regard, the works of Stutts et al. (1995, 1997) and Mall et al. (1996) suggested a specific interaction between ENaC and CFTR, resulting in down-regulation of the sodium permeability by wild-type CFTR. In another study, a decrease of the open probability of a sodium channel was observed when it was coreconstituted with CFTR in planar bilayer (Ismailov et al., 1996).

We report here the effect of the glibenclamide, a high-affinity inhibitor of SUR (Aguilar-Bryan et al., 1995) and low-affinity inhibitor of CFTR (Schultz et al., 1996; Sheppard and Robinson, 1997), on the epithelial sodium channel. We have studied the epithelial sodium channel from three species: rat (Canessa et al., 1993, 1994), human (Lingueglia et al., 1994; Voilley et al., 1994; McDonald et al., 1995), and Xenopus (Puoti et al., 1995) ENaC, all expressed in Xenopus oocyte by coinjection of the cRNA of the three subunits, , , and . We also have studied the native sodium channel present in the Xenopus kidney A6 cell line.

We observed that glibenclamide induced a stimulation (40-50%) of the amiloride-sensitive current in oocytes expressing Xenopus or human ENaC and a similar stimulation of the transepithelial, short-circuit current in A6 cells, but not on the rat ENaC-expressing oocytes. Because glibenclamide is a known ligand of several ABC proteins, this observation suggests the role of a protein of this type associated with ENaC.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

A6 Cell Culture and Short-Circuit Current Measurements. The experiments were performed on the clone A6-2F3 obtained by limiting dilution of A6 cells (Verrey et al., 1987). The culture conditions were similar to those described earlier in detail (Broillet and Horisberger, 1991). Briefly, A6-2F3 cells, at passage 88 to 98, were grown on collagen-coated Transwell permeable filters of 4.7 cm² (Costar, Cambridge, MA). The cells were used for electrical measurements after 7 to 35 days of culture in an amphibian medium (Handler et al., 1979) supplemented with 5% fetal calf serum (PAA, Linz, Austria), 10⁷ M dexamethasone (Sigma, St. Louis, MO), 100 U/ml penicillin G, and 130 µg/ml streptomycin. To obtain a high level of transepithelial sodium transport the cells were treated with 300 nM aldosterone 24 to 48 h before measurements.

(1)

Expression of Rat and Xenopus ENaC in Xenopus Oocytes. In vitro transcribed cRNA for the , , and  subunits of Xenopus, rat, or human ENaC was injected into stage V/VI Xenopus oocytes (0.3-1 ng of cRNA of each subunit in a total volume of 50 nl) as described earlier (Canessa et al., 1994; Puoti et al., 1995). Human ENaC , , and  subunit clones were generously provided by M. Lazdunski (Institut de Pharmacologie Moléculaire et Cellulaire, Sophia Antipolis, Valbonne, France). Electrophysiological experiments were performed 1 or 2 days after cRNA injection.

Ion Current Measurement in Whole Oocytes. The amiloride-sensitive sodium current was measured as described previously (Canessa et al., 1993, 1998) by using the two-electrode voltage-clamp technique by means of a Dagan TEV voltage-clamp apparatus (Dagan Corp., Minneapolis, MN) at room temperature (22-25°C) and at a holding potential of 100 mV in a solution containing 100 mM sodium gluconate, 0.4 mM CaCl₂, 10 mM sodium HEPES (pH 7.4), 5 mM BaCl₂, and 10 mM tetraethylammonium chloride. The current signal was filtered at 20 Hz by using the internal filter of the Dagan apparatus and recorded continuously on a paper chart. Low chloride concentration and a 5 mM concentration of the K⁺ channel blocker Ba²⁺ were used to reduce the background membrane conductance. The dose-response relationship was obtained by following the same principle as for the I_s.c. in A6 cells (see above).

Single-Channel Recordings. Before patch-clamp experiments, the oocytes were placed for 3 to 5 min at room temperature in a hypertonic medium (475 mOsM) with the following composition: 200 mM potassium aspartate, 20 mM KCl, 1 mM MgCl₂, 10 mM EGTA, 10 mM sodium HEPES, pH 7.4. The vitelline membrane then could be removed manually from the cell by using fine forceps (Methfessel et al., 1986). The oocytes then were transferred immediately to the recording chamber.

Chemicals and Drugs. Amiloride, glibenclamide, and tolbutamine were obtained from Sigma. A stock solution of 0.2 M glibenclamide in dimethyl sulfoxide (DMSO) was used. In all experiments performed with the A6 cells and in part of the experiments performed with the oocytes, appropriate amounts of DMSO were added to the solutions to obtain the same DMSO concentration in control and glibenclamide-containing solutions.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Effect of Glibenclamide on the Macroscopic, Amiloride-Sensitive Current in Xenopus Oocytes Expressing ENaC. As shown in the example of Fig. 1a, 100 µM glibenclamide induced a stimulation about 50% of the amiloride-sensitive sodium current, I_Na(am), in an oocyte expressing Xenopus ENaC. The increase in current reached a maximum after a few seconds and tended to decrease slightly thereafter. It was rapidly reversible. When the sodium channel first was inhibited by 10 µM amiloride, no effect of glibenclamide could be detected. Under our experimental conditions (i.e., with K⁺ channels blocked by 5 mM barium), no effect of 100 µM glibenclamide could be detected in noninjected oocytes (n = 9, data not shown).

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Effect of glibenclamide on the amiloride-sensitive current in oocyte expressing ENaC. a, original current recording in an oocyte expressing Xenopus ENaC. The current sensitive to 10 µM amiloride was measured at 100 mV holding potential before and during exposure to 100 µM glibenclamide. When amiloride was applied first, the same concentration of glibenclamide had no detectable effect. b, dose-response curve for the effect of glibenclamide on INa(am) in oocytes expressing Xenopus ENaC () or rat ENaC (). The curve was obtained from eq. 1, with the best-fitting parameters of K1/2 = 39 µM and a maximal increase of .50. Hardly any effect was detectable on the oocytes expressing rat ENaC except for a small inhibition at high concentrations. The absolute values of the current in the control (without glibenclamide) solution was 4.5 ± 0.6 µA (n = 12) for the Xenopus ENaC and 2.6 ± 1.1 (n = 4) for the rat ENaC.

Effect of Apical Glibenclamide on the sodium Transport by A6 Cells. Glibenclamide also induced an increase of the amiloride-sensitive transepithelial sodium transport in A6 cells grown on permeable support. Addition of 100 µM glibenclamide to the continuously perfused apical medium on A6 cells produced an increase of the amiloride-sensitive short-circuit current, I_s.c.(am), from 103 ± 7 to 146 ± 6 µA/dish (n = 14), P < .001 (paired t test). This effect had roughly the same time course as that observed in oocytes, although the rate of change was limited by the slower solution-exchange rate in the Ussing chamber. The effect also was reversible within a few minutes (Fig. 2a).

View larger version (13K): [in this window] [in a new window]   Fig. 2.   Effect of apical glibenclamide on sodium transport by A6 cells. a, the effect of 100 µM glibenclamide on the amiloride-sensitive short-circuit current, Is.c.(am). Although there was a ~20% spontaneous decrease in Is.c. over the ~15 to 20 min between the first and second control measurement, glibenclamide clearly induced a significant and reversible increase in INa(am) (n = 14). b, dose response for the effect of glibenclamide on the amiloride-sensitive short-circuit current, Is.c.(am). Four concentrations (10, 30, 100, and 300 µM) of glibenclamide were applied successively. The curve was obtained from eq. 1, with the best-fitting parameters of K1/2 = 46 µM and a maximal effect of .33. In this series of experiments, the absolute value of Is.c.(am) in the control solution was 138 ± 9 µA/dish (n = 12).

ENaC Subunits Involved in the Effect of Glibenclamide. With the purpose to find the precise mechanism of action of glibenclamide and to localize its effect to a specific ENaC subunit, we took advantage of the difference of response of rat and Xenopus ENaC and we tested effects of 100 µM glibenclamide on different heteromeric channels composed of rat and Xenopus ENaC subunits. As shown in Fig. 3, only the Xenopus  + rat  combination could be activated by glibenclamide with an amplitude similar to that observed with the Xenopus ENaC: the mean increase of I_Na(am) in this group was 49 ± 8% (n = 23). In the other subunit combination, the glibenclamide-induced change in I_Na(is) amounted to, at most, a few percent and was, in all cases, significantly smaller than that observed with the Xenopus ENaC. These results suggest that both the  and  subunits are essential for the action of glibenclamide, whereas the  subunit would not have any influence.

View larger version (49K): [in this window] [in a new window]   Fig. 3.   Effect of glibenclamide on different heteromeric channels produced by all combinations of rat and Xenopus ENaC subunits. Only the Xenopus and the Xenopus/ rat combination displayed a sizable stimulation by glibenclamide. The asterisks indicate a statistically significant difference (P < .001) from the Xenopus group; the solid circles indicate a statistically significant difference (P < .001) from the rat group. The numbers in parentheses indicate the number of measurements.

Effect of Glibenclamide at the Single-Channel Level. The effects of glibenclamide on the single-channel current also were examined in Xenopus oocytes expressing Xenopus ENaC by using the cell-attached mode and excised patches in the outside-out configuration. Currents in the cell-attached configuration were recorded with 100-mM LiCl solutions in the pipette, and the value of the single-channel conductance measured under these conditions was 7.3 ± 0.1 pS, a value similar to that reported earlier (Canessa et al., 1994). In the outside-out configuration, the identity of the channel could be verified by the reversible blocking effect of amiloride as described earlier (Chraïbi et al., 1998).

View larger version (41K): [in this window] [in a new window]   Fig. 4.   Effect of oocyte superfusion with glibenclamide on sodium channel activity. a, original patch-clamp channel recording showing a single, active sodium channel in the cell-attached configuration. The pipette was filled with a solution containing 100 mM LiCl, and the pipette potential was +100 mV. A downward deflection indicates current flowing from the pipette into the cell. The closed channel level is indicated by the mark to the left of the recording. At the time indicated by the bar, a solution containing 100 µM glibenclamide was perfused around the oocyte (but could not reach the extracellular side of the patch membrane, contained in the pipette). Glibenclamide did not induce any evident change in the channel activity. b, the mean effect of glibenclamide on the N · Po is shown in the bar graph (n = 4). Because of a large variation in the number of channels in each patch, the N · Po values are normalized to the value in the control solution. Glibenclamide did not induce any statistically significant change in N · Po (paired t test).

View larger version (57K): [in this window] [in a new window]   Fig. 5.   Effect of extracellular glibenclamide on sodium channel activity in outside-out excised patches. The pipette was filled with a 100-mM K (80 mM K gluconate, 20 mM KCl) solution, and the bath solution contained 100 mM LiCl. The pipette potential was held at 100 mV. A downward deflection indicates current flowing into the pipette, and the marks to the left of the recordings point to the current level when all sodium channels are closed. a, in this original recording, amiloride (200 nM) added to the bath (extracellular side) solution reduced the activity of the sodium channels by more than half. This effect was quickly reversible. b, this recording was obtained with the same patch as the one shown in a. Addition of glibenclamide (100 µM) to the external solution induced a reversible increase of the channel activity either by increasing the open probability or activating previously silent channels. c, average values of N · Po. The effects of 200 nM amiloride (left) and 100 µM glibenclamide (right) are reported (n = 8).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Glibenclamide increases the current mediated by Xenopus or human (but not rat) ENaC by a rather small but reproducible amount. This effect can be observed on the channel expressed naturally in epithelial cells originating from Xenopus kidney (the A6 cell line) as well as on the same channel expressed in Xenopus oocytes. The amplitude of I_Na(am) increase was somewhat smaller in A6 cells than in oocytes, and the apparent affinity was somewhat lower; however, considering the different experimental conditions, in terms of membrane potential or intracellular ionic concentration, the agreement is good enough to suggest that both effects are due to the same mechanism. This effect, therefore, is not an artifact of the expression system resulting from, for instance, the association of ENaC with a oocyte-specific protein.

The same effect clearly is detected at the single-channel level as an approximate doubling of the N · Po, with no changes of the single-channel conductance. Because of the rapid on- and off-rate of this effect, it is highly improbable that the number of channels present at the cell surface is changed, and, thus, we attribute this effect to an increase of the open probability of sodium channels.

What are the molecular mechanisms responsible for the effect of glibenclamide? The results of cell-attached- and excised-patch experiments indicated that the action must be local, i.e., on the channel protein itself or on a protein associated closely enough to the channel to be contained in the same, ~1-µm² patch of membrane. We recently have shown that trypsin and other serine proteases are able to increase the activity of ENaC expressed in oocytes or in A6 cells (Vallet et al., 1997; Chraïbi et al., 1998). The effect of proteases is of a much larger amplitude (2- to 10-fold with Xenopus ENaC), occurs with a slower time course, and is very slowly reversible.

To evaluate the possibility that the effect of glibenclamide was due to a release of the sodium "self-inhibition" (Garty and Benos, 1988), we measured the effect of 100 µM glibenclamide on the amiloride-sensitive inward current (at 100 mV) in a 5-mM Na⁺ extracellular solution. Although the amplitude of the stimulation was slightly smaller in 5 mM Na⁺ than in 100 mM Na⁺ solutions, this difference was quantitatively too small to allow one to conclude that the effect of glibenclamide is due to inhibition of the sodium self-inhibition.

Glibenclamide is a high-affinity ligand of the SUR expressed in pancreatic islet  cells, with measured kDa in the nanomolar range (Zünkler et al., 1988; Aguilar-Bryan et al., 1995), and inhibits K⁺ current carried by the channel formed by the association of SUR with an inward-rectifying K channel (Gribble et al., 1997; Tucker and Ashcroft, 1998). It inhibits inward-rectifying K channels in other organs with similar or lower affinity (Ashcroft and Ashcroft, 1992). It also inhibits the CFTR, but with a much lower apparent affinity (Schultz et al., 1996; Sheppard and Robinson, 1997). In addition, McNicholas et al. (1996) showed that coexpression of CFTR with the inward-rectifier K channel ROMK2 enhanced the sensitivity of ROMK2 to glibenclamide. These effects of glibenclamide, mediated by ABC proteins associated with ion channels, suggest the hypothesis that glibenclamide also acts on the sodium channel by binding to an ABC protein closely associated with ENaC. If an ENaC-associated glibenclamide receptor indeed does exist, it is probably not CFTR itself because we do not find any evidence of a cAMP-activated Cl current in native oocytes, whereas large, cAMP-activated Cl current can be induced by expression of exogenous mammalian CFTR (Chraïbi et al., 1998). However, our data do not allow one to exclude that glibenclamide binds directly to some part of the ENaC protein itself. Our attempt to localize the action of glibenclamide to a single subunit of the channel failed to give a simple answer. The simplest interpretation of our results is that both the  and  subunits of the rat ENaC are responsible for the resistance to this effect; however, other interpretations are possible.

Stutts et al. (1995) have provided evidence supporting a functional interaction between ENaC and CFTR in airways: the activation of CFTR would decrease the open probability of an associated sodium channel. This phenomenon would explain the high rate of sodium transport in the airways of cystic fibrosis patients (Boucher et al., 1986; Willumsen and Boucher, 1991) or of CFTR knockout mice (Grubb et al., 1994; Grubb and Boucher, 1997). Although we have no indication of the presence of CFTR expressed in oocytes, other ABC protein(s) could play the role of channel regulator, whether ENaC is artificially expressed in oocyte or naturally present in urinary epithelia.

Glibenclamide and several congener drugs are widely used for their stimulatory effect on insulin release by pancreatic  cells in the treatment of type II diabetes. Although these drugs allow a rather good control of glycemia in many cases, treatment with oral sulfonylurea does not decrease the high incidence of cardiovascular diseases commonly associated with type II diabetes (Meinert et al., 1970). Concerning arterial hypertension, patients treated by sulfonylurea drugs tend to have a higher blood pressure than those treated by insulin despite similar control of glycemia (Schmitt and Moore, 1993; Schmitt and Johns, 1995). In Liddle's disease, a genetically altered, hyperfunctioning ENaC leads to inappropriate sodium reabsorption and severe salt-sensitive hypertension (Hansson et al., 1995). Our results raise the question of a possible sodium-retention effect of glibenclamide when these drugs are used in the treatment of diabetic type II patients. This hypothesis obviously would need to be supported by measurements of the concentrations of glibenclamide in distal tubular fluid.

In summary, we have shown that glibenclamide induces a rapid activation of amiloride-sensitive sodium current mediated by the sodium channel naturally present in amphibian epithelial cells or by Xenopus or human ENaC expressed in Xenopus oocytes. Although the mechanism of this effect is not known, the available evidence suggests that glibenclamide acts on a closely associated protein or on the sodium channel itself.