Pflügers Arch – Eur J Physiol (1998) 436:270–279

© Springer-Verlag 1998

O R I G I NA L A RT I C L E

&roles:Matthias Rehn · Wolf-Michael Weber · Wolfgang Clauss

Role of the cytoskeleton in stimulation of Na+ channels in A6 cells by changes in osmolality

&misc:Received: 21 July 1997 / Received after revision: 4 December 1997 / Accepted: 16 December 1997

&p.1:Abstract Permeable supports with A6 cell monolayers were mounted in an Ussing chamber and bilaterally bathed with Ringer solution at room temperature. Shortcircuit current (Isc) was recorded continuously, and noise analysis revealed microscopic channel current characteristics. Our investigation focuses on the stimulation of apical Na+ entry caused by exposing the serosal surface of the A6 cell monolayers to hyposmotic Ringer solution. To evaluate the possible role of the cytoskeleton in the regulation of Na+ channels in response to a change in osmolality we used four different experimental approaches. In the control group, which were not exposed to any cytoskeleton-influencing drugs, there was a 1.5-fold increase in Isc and in the number of open Na+ channels after osmotic stimulation. For the second group cytochalasin D (0.1 µg/ml) was present on the serosal side during the experiments. Neither Isc nor the number of open Na+ channels increased after osmotic stimulation. In the third group colchicine (0.2 mM) or nocodazole (20 µM) was present on the serosal side, which resulted in 1.8-fold and 1.5-fold increases in Isc as well as 3-fold and 2-fold increases in the number of Na+ channels, respectively. In the fourth experimental group erythro-9-(2-hydroxy-3nonyl) adenine hydrochloride (EHNA, 0.5 mM), a dynein inhibitor, was present on the serosal side. In this group Isc decreased to about 0.4 µA/cm2, and subsequent application of amiloride abolished Isc completely. Under hyposmolar conditions EHNA abolished entirely the sensitivity of Isc to the osmotic challenge. Because of the EHNA-induced down-regulation of Isc, the density of apical Na+ channels in this experimental group could not be determined. These results show that the cytoskeleton is dominantly involved in osmotic channel regulation at the apical membrane, and that actin filaments, microtubules and molecular motors are involved in the recruitment of additional Na+ channels.

M. Rehn · W.-M. Weber · W. Clauss (✉) Institute of Animal Physiology, Justus-Liebig-University Giessen, Wartweg 95, D-35392 Giessen, Germany&/fn-block:

&kwd:Key words Epithelial monolayer · Kidney · Osmoregulation · Sodium transport · Noise analysis · Actin · Microtubules · Dynein&bdy:

Introduction Cultured renal distal tubule cells (A6) from the South African clawed toad Xenopus laevis grow to confluency on permeable supports and form a tight epithelium. Epithelia of this type and other tissues exhibit electrogenic Na+ transport which requires two steps: first, apical entry via the epithelial Na+ channel and, second, extrusion by the basolateral Na,K-ATPase. This transport can be stimulated by hormones and second messengers. A6 monolayers behave like other tissues when exposed to cAMP, arginine vasopressin (AVP) or aldosterone. Whereas the response of transepithelial Na+ transport to aldosterone or other corticoids occurs over a period of hours [1, 5, 9, 14, 15], the AVP-induced response [16, 17] is an example of a short-term reaction. In the case of aldosterone transepithelial short-circuit current (Isc) is increased and the Na+ transport is completely amiloride sensitive. In contrast, AVP causes additional Cl– secretion and amiloride only partially blocks the Isc. A short-term increase in amiloride-sensitive transepithelial Na+ transport can also be caused by osmotic stimulation [23, 32]. Hypotonic serosal solution causes an increase in Isc in confluent monolayers and cell swelling followed by a regulatory volume decrease in isolated cells [7], whereas hypertonic solution on the basolateral side of the epithelium causes cell shrinkage and a decrease in Isc [6, 32]. When the serosal side of the epithelium is exposed to hypotonic solution, the increased Isc depends only on cellular transport and not on the additional gradient across the epithelium, as indicated by the observation that the Isc is totally inhibited by amiloride. Hyposmotic solution on the apical side of the epithelium has no effect on Isc [32]. Furthermore, from studies of excised patches it is known that actin filaments regulate Na+ channel activity [3, 4]. The aim of this study was to investigate the short-term re-

271

sponse of the Na+ channel population in an intact epithelial monolayer to stimuation of its serosal side by hyposmotic solution, and the involvement of the cytoskeleton. Osmotic effects on actin filaments, microtubules and molecular motors were studied by the specific drugs cytochalasin D, colchicine, nocodazole and erythro-9-(2-hydroxy-3-nonyl)adenine hydrochloride (EHNA). Singlechannel current, channel density and open probability were obtained by noise analysis.

Materials and methods Glossary Isc

short-circuit current; µA/cm2

IAmilo INa INa0

amiloride-insensitive current; µA/cm2 macroscopic amiloride-sensitive Na+ current; µA/cm2 amiloride-sensitive Na+ current at zero blocker concentration; µA/cm2 single channel Na+ current in presence or absence of blocker; pA blocker concentration in the apical solution; µM ON-rate: open to blocked state rate coefficient; (µM*s) –1 OFF-rate: blocked to open state coefficient; s–1 equlibrium blocker coefficient of open channels≡kOFF/ kON; µM number of open Na+ channels in the absence of blocker open probability in the absence of blocker total channel number in the absence of the blocker low frequency plateau of power density spectrum corner frequency of power density spectrum transepithelial resistance; kΩ*cm2 normal Ringer solution; following called isosmotic Ringer

iNa, iNa0 [B]a kON kOFF kB N o0 P o0 N T0 So fc Rt NRS

Cell culture The A6 cell line (clone established by C. Johnson, Pittsburgh) was a generous gift from Willy Van Driessche (Leuven, Belgium). Passages 119–139 were grown in plastic flasks with a growth area of 25 cm2 (Nunc, Wiesbaden, Germany) at 25°C in a humidified incubator with 5% CO2 in air. The growth medium contained three parts of Coons F-12 (3.48 g/l) and seven parts of Leibovitz L-15 (10.36 g/l) culture media (Sigma, Deisenhofen, Germany), and 2.1 g/l NaHCO3 (Fluka, Neu-Ulm, Germany). This medium was supplemented with 10% fetal calf serum (FCS) (GIBCO, Eggenstein, Germany), 100 IU/ml penicillin and 100 µg/ml streptomycin (GIBCO). The cells were fed twice a week by replacing the growth medium with fresh medium. After 10–12 days the flask bottoms were coated with a confluent monolayer. After exhausting the growth medium, the cells were washed with 10 ml phosphatebuffered solution (PBS) containing (in mM): 137 NaCl, 2.7 KCl, 8.1 Na2HPO4 and 0.2 KH2PO4 (pH 7.4, adjusted with 1 M TRIZMA). Then the cells were detached with trypsin solution (1–2 ml per flask) containing 0.0625% trypsin in PBS and 0.005% ethylenediaminetetraacetate (EDTA). Then, 5–10 min later, the action of the enzyme was stopped with growth medium. After centrifugation (5 min) at 1900 g and resuspension in 3 ml growth medium, the cells were divided into six permeable inserts (Snapwell, polycarbonate membrane, pore size 0.4 µm, 1 cm2 growth area, Costar, Bodenheim, Germany). The cells reached confluence 10–14 days after seeding and the inserts were transferred to modified Ussing chambers.

Solutions and chemicals First both sides of the epithelium were bathed with identical isosmotic Ringer solution (NRS). This solution had an osmolality of about 240 mosmol/kg and contained (in mM): 120 NaCl, 3.4 KCl, 0.8 MgCl2, 0.8 CaCl2, 1 4-(2-hydroxyethyl)-1-piperazine-ethanesulphonic acid (HEPES) and 10 glucose (pH=7.4; adjusted with 1 M TRIZMA). These chemicals were purchased from Fluka. The solutions were gassed with O2. 6-Chloro-3,5-diamino-pyrazine-2-carboxamide (CDPC) and amiloride (50 µM; both Sigma) were added to the apical side of the epithelium to block epithelial Na+ channels. The CDPC concentration varied between 10 and 200 µM. In Na+free solution Na+ was replaced by equimolar N-methyl-D-glucamine (NMDG, Sigma). To change the serosal osmolality a hyposmolar solution with an osmolality of about 180 mosmol/kg was used, made of three parts NRS and one part distilled water (pH=7.4; adjusted with 1 M TRIZMA). The first experimental series was done without any drugs in order to obtain data under iso- and hyposmolar conditions. To estimate the involvement of the cytoskeleton in the regulation of Na+ transport during changes in osmolality, specific drugs were used on the serosal side during the experiments. Cytochalasin D (Sigma), a disrupter of actin filaments, was added at 0.1 µg/ml (equivalent to 0.2 µM). The role of the microtubules was investigated using colchicine (0.2 mM) and nocodazole (methyl[5(2-thienylcarbonyl)-1H-benzimidazol-2-yl]carbamate) (20 µM; both Sigma). To analyse the role of dynein, a compound of microtubules with ATPase activity, we used the inhibitor EHNA (Sigma) [20, 26]. In order to obtain maximal and half-maximal blocker concentrations, EHNA concentrations of between 0.01 and 0.5 mM were added to the serosal side. Subsequently 0.5 mM EHNA was used in all experiments. Finally noise analysis was carried out on data obtained in the presence of a low EHNA concentration (0.1 mM), to investigate if the Na+ channel density is affected by the drug. Electrical measurements All electrical measurements were performed at room temperature with the preparation placed in a modified Ussing chamber, allowing continuous perfusion of each side of the chamber. The solutions were exchanged without interruption of the electrical measurements. The apical flow rate was 14 ml/min and the basolateral was 12 ml/min. The voltage and current electrode pairs (both Ag/AgCl) were connected to the apical and basolateral bathing solutions via 1 M KCl-1% agarose bridges. The electrodes were connected to a low-noise voltage-clamp amplifier [30]. The transepithelial potential was clamped to zero and the Isc was continuously recorded. The transepithelial resistance was determined by Ohm’s law by applying continuous voltage pulses (+6 mV) and measuring the resulting current deflections. The data were digitized via an A/D transducer (MacLab, ADInstruments, Australia) and stored on a Macintosh LC II (Apple Computer, Calif., USA). The analogue data were transferred to noise analysis equipment [29] and to a dual-channel chart recorder (Kipp and Zonen BD 112, Delft, The Netherlands). Noise analysis The macroscopic current, which is equivalent to Isc, varies randomly. These fluctuations are due to small currents passing through a large number of channels which open and close independently of each other. The use of a specific blocker causes an interaction of the blocker with the Na+ channels. The blocker interrupts the single-channel current, which leads to a third blocked state. The random interactions of the blocker with the channels cause a Lorentzian component in the power density spectrum. This can be described with the following equation: S(f ) = So/1+(f /fc)2 (1) where So designates the plateau value and fc the corner frequency. The power density spectra were calculated from an average of 20 sweeps of data with a fundamental frequency of 0.5 Hz. Only the

272 frequency data between 1 and 366 Hz are shown because the relevant Na+ channels gate in this range. Data of higher frequencies represent the amplifier-induced noise and were cut-off. In the present study we recorded blocker-induced noise. The blocker concentration was increased during the experiment (see above). In the selected concentration range fc correlates with blocker concentration according to a linear regression model: 2π fc = kON [B]a+kOFF

(2)

The interaction of the blocker (B) with the open channel is characterized by the ON-rate (kON). The OFF-rate (kOFF) describes the dissociation of the blocker from the channel, and the subscript a refers to the apical side. The amiloride-insensitive current (IAmilo) was determined with 50 µM amiloride in the apical bathing solution. The macroscopic amiloride-sensitive current (INa) is the difference between Isc and IAmilo. The single-channel current (iNa) can be calculated from Eq. 3, where fc and So are the parameters from the Lorentzian fits, kON is from Eq. 2 and INa is the macroscopic amiloride-sensitive Na+ current: iNa = So π 2 fc2/INa kON[B]a

The probability of finding the channel in the open state in the absence of the blocker is calculated from: Po0 = [No0/No–1] · kB/CDPCa

(5)

by extrapolation to the zero blocker concentration. The inhibition constant kB is calculated by kOFF/kON. The total number of channels at zero blocker concentration (open+blocked+closed) was obtained from NT0 = No0/Po0

(6)

Statistics Data are expressed as mean ± standard error of mean (SEM). The Student’s test was used for estimating the significance of differences between paired mean data (n=number of culture inserts).

Results

(3)

The channel density was calculated according to the pseudo threestate model [12, 13]. The number of open channels in the absence of blocker, No0, can be calculated from the amiloride-sensitive Na+ current in the absence of the blocker (INa0) and the single-channel current at the extrapolated zero blocker concentration (iNa0), using:

Influence of an osmotic challenge on parameters (Isc, Rt) of macroscopic current

(4)

In a standard protocol prior to all experiments we determined Na+ channel density and single-channel current by noise analysis using CDPC. Subsequently the amiloride-

Fig. 1A–D Effect of a stepwise increase in 6-chloro-3,5-diaminopyrazine-2-carboxamide (CDPC) concentration (µM) at the apical side under iso- and hyposmotic conditions. Amiloride (50 µM) was added after the highest dose of CDPC. N-methyl-D-glucamine (NMDG) was used to detect non-amiloride-sensitive Na+ current. A Recording from a control experiment (no drugs). Osmotic stim-

ulation caused a shift in short-circuit current (Isc, n=6). B Influence of cytochalasin D (0.1 µg/ml) to the serosal side of the epithelium. Cytochalasin D prevented a shift in Isc (n=5). C Effect of colchicine (0.2 mM, serosal side) under isosmotic and hyposmotic conditions (n=5). D Effect of nocodazole (20 µM, serosal side) on Isc during isosmotic and hyposmotic conditions (n=6)&ig.c:/f

No0 = INa0/iNa0

t-Test between iso- and hyposmotic part of each treatment group: *significantly different (P<0.05); drugs applied on serosal side. Please note! Different base lines in Isc under isosmotic conditions were caused by varied properties of A6 passages!&/tbl.:

2.35 ± 0.27* 2.30 ± 0.25* 10.00 ± 2.80 1.68 ± 0.24 1.50 ± 0.26 8.50 ± 3.28 3.87 ± 0.65* 3.61 ± 0.71* 6.94 ± 0.46* 2.56 ± 0.54 2.43 ± 0.58 5.29 ± 0.63 10.64 ± 1.25* 10.15 ± 1.23* 3.91 ± 0.28* 6.47 ± 0.91 6.03 ± 0.95 5.26 ± 0.85 5.01 ± 0.53 4.35 ± 0.62 4.58 ± 0.43 5.70 ± 0.78 4.92 ± 0.73 4.81 ± 0.73 6.48 ± 0.98* 6.30 ± 0.98* 7.04 ± 0.52* 4.26 ± 0.82 4.19 ± 0.81 8.88 ± 0.59 Isc (µA/cm2) INa (µA/cm2) R (kΩ• cm2)

Hyposm. Isosm. Hyposm. Isosm. Hyposm. Hyposm.

Hyposm. Isosm. Isosm.

Isosm.

EHNA (0.1 mM) n=4 Nocodazole (20 µM) n=6 Colchicine (0.2 mM) n=5 Cytochalasin D (0.1 µg/ml) n=5 No drugs n=6 Parameter

Table 1 Short-circuit current (Isc), amiloride-sensitive Na+ current (INa) and transepithelial resistance (Rt) in control and drug-treated A6-monolayers in normal and hyposmotic Ringer. Data are means ± SE; n = number of experiments. Values are given during stabilazation in Isc in iso- and hyposmotic Ringer.&/tbl.c:&

273

sensitive portion (INa) and total Na+-mediated currents of the whole Isc were resolved (left part of current traces). Figure 1A shows a typical recording of Isc under control conditions (no drugs). The Isc raised when the serosal osmolality was changed from 240 to 180 mosmol/kg. First a fast up- and then down-regulation of Isc appeared, followed by a slow increase, so that the new base line of Isc became stable after 40 min. The stimulated Isc was completely blocked when amiloride (50 µM) was added to the apical solution. This indicates that, under hyposmotic conditions, Isc is equivalent to amiloride-sensitive Na+ current (INa). Isc and INa increased about 50% (P<0.05, Table 1) in comparison with the corresponding values in the presence of isosmotic Ringer solution. The transepithelial resistance (Rt) decreased significantly by about 20% (Table 1). In order to study the role of actin filaments during osmotic stimulation, cytochalasin D, an agent that disrupts actin, was used. Several control experiments with different concentrations of cytochalasin D revealed that more than 1 µg cytochalasin D per millililtre Ringer solution damages the epithelium irreversibly (loss of transepithelial resistance; not shown). Therefore, we chose a concentration which caused no apparent damage to cells (0.1 µg/ml). Figure 1B shows a typical trace of Isc when cytochalasin D was present during the experiment. When the drug was added at a stable Isc there was a slow increase in Isc in three (+1.69±0.87 µA/cm2) and a slow decrease (–0.68±0.15 µA/cm2) in two cases, a total variation in Isc of about 0.62±0.78 µA/cm2. Hyposmolar stimulation caused no shift in Isc, only a transient overshoot and decline in Isc appeared. Amiloride (50 µM) inhibited Isc completely. Rt was not affected by solutions of different osmolality (Table 1). Subsequently we investigated the influence of the tubulin system on the recruitment of additional Na+ channels. For this investigation we used colchicine and nocodazole, both disrupters of microtubules. An original trace of a colchicine experiment is shown in Fig. 1C. The stable level of Isc during the hyposmotic period was raised compared that when isosmotic Ringer was used. An immediate and transient increase and then decrease in Isc occurred when the osmolality was changed on the serosal side, followed by a slower increase. A new stable value of Isc was reached after 50 min. The increases of Isc and INa were to levels above 60% those measured under isosmotic conditions. Addition of amiloride (50 µM) to the apical side blocked Isc under both isosmotic and hyposmotic conditions. Rt before each CDPC-induced blockade decreased significantly between exposure to isosmotic and then hyposmotic solutions (Table 1). Isc and INa increased significantly when the osmotic conditions were varied (Table 1). Control experiments demonstrated that exposure to the cold alone or in combination with bilaterally applied nocodazole damages the epithelium (loss of transepithelial resistance or down-regulation of Isc; not shown). Another difficulty was the precipitation of nocodazole dissolved in dimethylsulphoxide (DMSO) when added to cold (4°C) Ringer. Therefore, nocodazole was added at room temperature to the serosal side. Figure 1D shows a recording

274

Fig. 2 Relative changes in Isc from the different experiments. The Isc before changing the osmolality is the value against which changes are normalized. The time course is shown from 10 min before to up to 40 min after the osmotic challenge. A quotient of 1 means no change. Whereas the hyposmotic challenge caused an increase in Isc, cytochalasin D prevented this response in the presence of a lower serosal osmolality.&ig.c:/f

from one of these experiments. The stable level of Isc before the first use of CDPC was about 2 µA/cm2. After the first wash-out Isc seemed to increase (+0.5 µA/cm2). By replacing the isosmotic solution with the hyposmotic solution, a transient overshoot and decline in Isc appeared, as it did in the other experimental series, followed by a slow increase of about 1 µA/cm2. Rt before each CDPC-induced blockade increased under osmotic stimulation. Under both osmotic conditions Isc was blocked by 50 µM amiloride. Figure 2 summarizes the experiments done with cytochalasin D, colchicine and nocodazole and compares the effects of these drugs on Isc. To obtain the kinetics of EHNA we used concentrations over a range of 0.01 mM to 0.5 mM on the serosal side. Figure 3A shows the dose-dependent behaviour of the amiloride-sensitive Isc upon increasing the EHNA concentration. We determined a half-maximal blocker concentration (k1/2) of 31±2.4 µM (Fig. 3B). In order to achieve maximal inhibition, in all further experiments we used 0.5 mM EHNA on the serosal side of the epithelium. Figure 3C shows a recording of the test series. Isc decreased to about 0.39±0.14 µA/cm2, and subsequent application of amiloride (50 µM) inhibited the Isc completely. Under hyposmolar conditions, EHNA abolished entirely the sensitivity of Isc to osmotic challenge (0.41±0.25 µA/cm2). In four of five experiments the removal of EHNA caused an increase in Isc to the former base line. Noise analysis was carried out on data obtained in the presence of 0.1 mM EHNA (n=4; traces not shown) on the serosal side of the epithelium during both osmotic conditions. In this case there was a significant drop in Isc from 5.20±0.43 to 1.68±0.24 µA/cm2 (Table 1) after the application of the dynein inhibitor to the isosmotic Ringer bathing the serosal side of the epithelium. Then the serosal solution was changed to hyposmotic Ringer and the Isc increased significantly to 2.35±

Fig. 3A–C Two different kinds of experiments showing the effects of erythro-9-(2-hydroxy-3-nonyl)adenine hydrochloride (EHNA) and typical recordings of Isc. A Effect of a stepwise increase in EHNA concentration (mM) at the serosal side. B Shown is the dose dependence of inhibition (n=5). C Influence of EHNA (0.5 mM) at the serosal side of the epithelium under isosmotic and hyposmotic conditions. Shown is Isc and its recovery after the removal of EHNA (n=5)&ig.c:/f

0.27 µA/cm2 (+40%; Table 1). This means a reduction of 20% in the capacity of Isc to contribute to osmotic regulation compared to control conditions (+50%). Rt was unaffected when the preparation was bathed in solutions of different osmolality (Table 1).

different (P<0.05) from corresponding value (isosmolar); all drugs applied on serosal side. Please note that different passages of A6 cells were used for each approach!

Fig. 4 Relative changes in open channel number (rel. No0), total channel number (rel. NT0), single-channel current (rel. iNa0) and the open probability (rel. Po0) between hyposmotic and isosmotic conditions in the absence of blocker. Data are obtained from the three-state model. No difference is indicated with “1”. Values over 1 indicate an increase and those below 1 a decrease in the parameters of microscopic current&ig.c:/f

Influence of an osmotic challenge on parameters of microscopic current

&/tbl.:

* Significantly

6.52±0.56 239.33±25.69 38.28±7.27 .258±32* 9.3±1.6* 0.14±0.02* 66.3±5.5* 6.49±0.65 264.47±7.49 42.16±4.94 .386±30 3.9±0.6 0.35±0.07 12.4±2.9 8.75±0.45* 239.25±29.16 28.41±4.54 .326±12 11.5±2.1* 0.15±0.03* 103.0± 31.9* 7.90±0.30 262.61±17.78 33.61±2.98 .387±48 6.4±1.4 0.43±0.09 20.0±6.0 7.92±0.54 8.36±0.42 208.73±28.13 182.49±8.38 27.81±5.40 22.17±1.90 .279±37 .217±15 22.6±3.2 48.2±7.7* 0.46±0.07 0.43±0.09 56.7±13.9 132.4±32.3* 6.35±0.29 6.26±0.56 238.38±14.45 250.45±11.34 37.77±2.55 41.09±3.61 .358±38 .356±18 14.3±2.2 12.2±1.6 0.43±0.06 0.30±0.06 37.4±8.5 57.4±22.3 6.53±0.38 6.68±0.25 204.22±13.95 209.97±11.60 32.34±3.72 31.66±2.46 .225±16 .222±6 19.2±3.6 28.4± 4.3* 0.217±0.067 0.181±0.046 157.1±65.7 223.9±61.9 kON (s•µM)–1 kOFF (s)–1 kB (µM) iNa0 (fA) No0 (106/cm2) P o0 NT0(106/cm2)

Hyposm. Isosm. Hyposm. Isosm. Hyposm. Hyposm.

Hyposm. Isosm. Isosm.

Isosm.

EHNA (0.1 mM) n=4 Nocodazole (20 µM) n=6 Colchicine (0.2 mM) n=5 Cytochalasin D (0.1µg/ml) n=5 No drugs n=6 Parameter

Table 2 Kinetic parameters of 6-chloro-3,5-diamino-pyrazine-2-carboxamide (CDPC) block and microscopic Na+ channel characteristics according to the three-state model. Values are means ± SE; n = number of experiments. On- and off-rate of CDPC channel reaction (kON, kOFF) are obtained from analysis of power density spectra; kB = kOFF/kON;

all following parameters are determined in the absence of blocker: iNa0 (single Na+ channel current), No0 (number of open channels), Po0 (probability in finding channels in open state), NT0 (total channel number, open, closed and blocked)&/tbl.c:&

275

Under control conditions blocker kinetic parameters (kON, kOFF, kB) were not affected by osmotic stimulation. Therefore, we could not discriminate between the electrophysiological properties of already operating and newly activated Na+ channels. In addition, the variations in single-channel parameters in the absence of blocker (iNa0, Po0) were not significant (Table 2). The number of open channels at zero blocker concentration (No0), however, increased by about 50% (Table 2), whereas the total amount of channels in the absence of blocker (NT0) was not increased. Relative changes in microscopic current parameters between isosmotic and hyposmotic conditions under control conditions (no drugs) are shown in the left columns of Fig. 4. Figure 5 shows the power density spectrum from a control experiment at two different blocker concentrations in a doublelogarithmic plot. Higher blocker concentrations caused a decrease in plateau values and higher values of fc. The fits to the data measured under hyposmotic conditions are shifted upwards compared to those obtained under isosmotic conditions. The fc was not altered by the differences in osmolality within the same blocker concentration. In the presence of cytochalasin D the kinetic parameters (kON, kOFF, kB) were not affected by changes in osmolality.

276

chicine did not prevent a shift in the curve fitted to the data obtained under hyposmotic conditions (spectra not shown). An osmotic challenge in the presence of nocodazole did not result in a change in the kinetic parameters of blockade, or in single-channel current in the absence of blocker (Table 2). The increase in the number of open Na+ channels in the absence of blocker was nearly twofold and that in the total number of Na+ channels was more than fivefold. Because of this immense increase in channel number the open probability decreased to two-thirds of its control value (Table 2). Power density spectra from such experiments do not differ from those measured under control conditions (no drugs). An increase in blocker concentration caused a decrease in plateau values and a shift in fc. Nocodazole did not prevent a shift in the curve fitted to the data obtained under hyposmotic conditions (spectra not shown). EHNA (0.1 mM) did not vary the blocker channel interaction (kON, kOFF, kB) when the preparation was bathed in isosmotic or hyposmotic solution (Table 2). The single-channel current and open probability in the absence of blocker decreased significantly when the isosmotic Ringer was replaced by hyposmotic solution (Table 2). Nevertheless, the number of open Na+ channels increased more than twofold and the total number of Na+ channels increased more than fivefold (Table 2). A shift in the spectra fitted to the data obtained under hyposmotic conditions was observed (not shown). Fig. 5 Power density spectra in the presence of 50 [iso.(●); hyposm.(l )] and 140 [iso.(■); hyposm.(n )] µM CDPC. Control experiment without drug treatment. The power density spectra are shifted in the presence of hyposmotic solution&ig.c:/f

The single-channel current and open probability at zero blocker concentration under hyposmotic conditions did not vary from those measured under isosmotic conditions (Table 2); also, the number of open channels and the total number of channels in the absence of the blocker were unchanged. Relative changes in microscopic current parameters between isosmotic and hyposmotic conditions in the presence of cytochalasin D are shown in Fig. 4. The use of cytochalasin D prevented the shift in the power density spectra caused by bathing the preparation in hyposmotic solution. Also, fc was not affected by changes in osmolality at constant concentrations of blocker (spectra not shown). In the colchicine-treated group the change in the osmolality of the bathing solution did not cause variation in kinetic parameters of blockade (kON, kOFF, kB). In addition, the single-channel current and open probability in the absence of blocker were not affected by a change in osmolality (Table 2). However, the increase in the number of open Na+ channels and the total number of Na+ channels when changing from isosmotic to hyposmotic conditions was more than twofold in the absence of the blocker. Relative changes in microscopic current parameters between isosmotic and hyposmotic conditions are shown in Fig. 4. Power density spectra are similar to those from control experiments. An increase in blocker concentration caused a decrease in plateau values and a shift in fc. The use of col-

Discussion For investigating the role of the cytoskeleton in the regulation of epithelial Na+ channels, detailed studies have so far concentrated on the role of actin filaments, and have been done mainly on excised patches under isosmotic conditions [3, 4]. One study addresses microtubules and also conducts noise analysis on data obtained from a frog skin preparation [8]. The aim of our study was to use an intact epithelial monolayer of a single cell type to investigate hyposmotic stimulation and to extend the investigation of the role of the cytoskeleton to the microtubule system and molecular motors. With the aid of noise analysis we were able to analyse macroscopic current parameters (Isc and Rt) and also the microscopic parameters, i.e. number, single-channel current and open probability of the Na+ channels. Influence of osmotic challenge on macroscopic current parameters (Isc, Rt) Our experiments show under control conditions (no drugs) that Isc increases by 50% when the serosal side of the preparation was exposed to hyposmolar solution. This is in good agreement with other reports: Granitzer et al. [11] report an increase of 25%; Niisato et al. [23] one of 50%; and Wills et al. [32], an increase of 75%. This variation may be the result of different cell culture conditions. The immediate transient increase in Isc after changing the serosal solution is likely to be due to Cl– secretion, and the transient decrease may be based on inverted Cl– flux. This

277

is reported by Ehrenfeld et al. [7], Crowe et al. [6], Brochiero et al. [2] and Marunaka and Eaton [21]. We focused our attention on the increased base line level of Isc after the osmotic challenge. The subsequent rise and stabilization of Isc after 40±10 min was also seen by other groups [7] and is caused by a slow but steady increase in Na+ absorption until a new increased stable base line is reached. In all our experiments Isc was completely inhibited by amiloride (50 µM), whatever the osmolality of the bathing solution. Therefore, it is very likely that the rise in Isc induced by the hyposmotic solution is carried entirely by Na+ ions via a transcellular pathway through amiloride-sensitive Na+ channels in the apical membrane. Another reason for the increase in Isc may be a change in the shunt resistance of the epithelium caused by lowering the osmolality on the serosal side along with the Na+ gradient across the epithelium evoked by dilution. This paracellular resistance does not seem to be affected by hyposmotic stimulation, because amiloride causes complete inhibition of Isc. If the paracellular resistance had been lowered, the Na+ gradient would have generated a paracellular flux, and the resulting current would not have been completely amiloride sensitive. Therefore, it is concluded that the increased Na+ transport rate occurs through the transcellular route. Influence of osmotic challenge on microscopic current parameters Considering that one channel can only pass at a certain quantity of ions per unit time, an increase in Na+ transport rate requires that either an increase in the number of Na+ channels or an increase in mean open time, equivalent to a raised open probability, occurs. To investigate these parameters we used noise analysis. In order to examine the channel type prevailing before and after osmotic stimulation, the interaction of a specific blocker (CDPC) with the channel was investigated. This is shown by the kinetic parameters of the blocker channel interaction. As there is no difference in the kB constant between the groups (Table 2), the additional channels are probably not of a novel type. A variation of kB could indicate a modification of the channel surface, such that it has a different affinity towards the channel blocker. An increase in pore size enables elevation of the single-channel current (iNa) and this is associated with a rise in Isc. To test this possibility, iNa in the absence of blocker was calculated. We could not find any significant change before and after osmotic stimulation, which indicates that the pore size of the additional channels is the same as that of the channels present under isosmotic conditions. Another important feature is the gating characteristic of the channel: a channel with an increased open probability can allow a greater flux of ions through it than a channel with a lower one. In our experiments hyposmotic stimulation induced no significant changes in open probability. In summary, none of the parameters that may cause an increase in Na+ flux discussed so far changed significantly in response to a hyposmotic challenge. This leaves only the number of channels. Indeed we observed a 50% increase in

the number of open channels, but no significant increase in total channel number (Table 2). In A6 cells adapted to isosmotic conditions, exposure of the serosal membrane to a hyposmolar solution caused Isc to rise, brought about through an increase in the number of Na+ channels in the apical membrane. This conclusion was also reached by Wills et al. [32]. In principle there are two possibilities for the additional recruitment of epithelial Na+ channels: exocytosis of subapically located Na+ channels in the apical membrane [10] or activation of quiescent channels [18] by the osmotic stimulation. The first possibility implies an increase in the total number of channels, and the second means an increase in the number of open channels but no increase in total channel number. Our results seem to support the second possibility for the control group. The role of actin filaments in the response to hyposmotic solution We investigated the role of the cytoskeleton in the regulation of open channel density following a change in osmolality. It has already been state that additional epithelial Na+ channels may be recruited. It is known that the cytoskeleton plays an important role in exocytotic insertion and endocytotic removal of integral membrane proteins. Microfilaments are responsible for the regulation of water and solute transport in tight epithelia. It has been shown in studies of excised patches of A6 cells that actin filaments regulate Na+ channel activity [3, 4]. We used cytochalasin D to disrupt the actin filaments. Ma et al. [19] reported that cytochalasin D itself causes an increase in the paracellular permeability and Prat et al. [28] reported that cytochalasin D (5 µg/ml) stimulates Na+ channels. However, this effect was only measurable at concentrations that damage the epithelium when applied over a longer period. Short-term application is fully reversible. In our study we investigated epithelial function over a longer time frame of several hours; therefore, we applied a lower concentration of 0.2 µM. This concentration did not cause a marked increase in permeability, as evidenced by an unchanged Rt (Table 2). Similar results were reported by Verrey et al. [31]. Our results show that actin filaments seem to be involved in the regulation of channel density. Cytochalasin D prevented the hyposmolality-induced increase in the amiloride-sensitve Isc, but had no significant effect on macroscopic or microscopic current parameters (Tables 1, 2). Therefore, the maintenance of actin organization appears to be important in the autoregulation of Na+ channel density. These results are in good agreement with the findings reported by Els and Chou [8], in their study of frog skin. If actin filaments are responsible for the recruitment of channels from subapical pools, disruption of these filaments should affect the insertion of new channels, such that total channel number would remain unchanged. Unfortunately under control conditions (intact actin filaments) we had a large error (SE) on the measurement of total channel number, so that any increase in total channel number may have been masked. Our results indicate that actin interacts with

278

quiescent Na+ channels, because actin disruption evoked no increase in open or total channel number. Role of the microtubular system in the response to hyposmotic solution Microtubules are responsible for the membrane-bound movement of intracellular components in polarized epithelia cells. Disruption of microtubules blocks the transfer of vesicles containing biological molecules to the apical membrane [24]. In our experiments colchicine (0.2 mM) did not prevent the effects of stimulation of Na+ channels by hyposmolar solution. Even at higher doses (more than 1.2 mM; data not shown), colchicine had no inhibitory effect on the increase in Isc. Another known microtubule disrupter, nocodazole (20 µM), had similar effects, which counters the results of Morris et al. [22], who reported that a decrease in Isc occurred after the application of colchicine and nocodazole. However, Rt was decreased by colchicine and increased by nocodazole, an adverse effect for which we have no explanation as yet. It shows, however, that the agents were indeed acting on the cells. We agree with Ma et al. [19] that the disruption of microtubules only affects epithelial Na+ conductance, since the application of microtubule inhibitors prevented the actions of amiloride. As kB and iNa were unaltered by changes in osmolality in the presence of colchicine and nocodazole, the same facts as previously discussed are valid. The number of open channels was doubled in both test series (Table 2, Fig. 4). The open probability was not altered by the application of colchicine; therefore, the rise in Isc must be brought about through an increase in the number of conducting channels. In the nocodazole series of experiments, the open probability even decreased significantly during hyposmotic stimulation (Table 2). In order to account for the increased Na+ flux the total number of Na+ channels must have increased dramatically (Table 2). In summary the results show that the integrity of the microtubule system is not essential for the immediate delivery of additional Na+ channels. In contrast, and very surprising to us, the loss of microtubule integrity caused the largest increase in total channel number. Similar results reported Els and Chou [8] in their study of the frog skin. Role of molecular motors in the response to hyposmotic solutions Vesicle transport via the microtubule system is based on trafficking motor proteins, such as dynein and kinesin. In epithelial cells dynein transports vesicles to the apical membrane [25]. Dynein possesses ATPase activity and EHNA is an ATPase inhibitor which specifically affects dynein activity [26]. Although EHNA concentrations above 1 mM also affect the basolateral Na,K-ATPase [27], we assume in our experiments, because of the lower dose, that the decrease in Isc is caused by a direct inhibi-

tory effect of EHNA on dynein. This compound completely prevented the induction of additional Na+ channels by exposure to hyposmolar solution. Noise analysis indicates that the channel density prevalent under isosmotic conditions is reduced by a low EHNA concentration alone, and this reduces the possibility of osmotic stimulation recruiting additional Na+ channels. Under the influence of a low EHNA concentration, the properties of the epithelia were not the same as those of untreated epithelia. This result may point to a permanent reconstruction and reorganization of biological membranes. Conclusions The combination of these findings may indicate that the hyposmolality-induced, short-term regulation of Na+ channel density might involve only those channels that are already in the apical membrane in the case of an intact cytoskeleton. Vesicle-bound channels may be transported with the aid of microtubules to form subapical pools. These channel vesicles might be fixed by the microtubule system and, upon stimulation, may be inserted into the apical membrane via actin filaments. The fixing component of the tubule system prevents overload of the apical membrane with channels, because when tubules were disrupted there was an immense rise in the number of open channels. Although a detailed understanding of all these events requires further research, the present experiments clearly demonstrate the importance of the cytoskeleton in the osmotic-induced activation of epithelial Na+ channels. &p.2:Acknowledgements We thank Dr. H. Bjerregaard (Roskilde, Denmark) for helpfull suggestions on A6 cell culture, and C. Schwarz, M. Hündt and H. Heidt for excellent technical assistance. This study was supported by the Deutsche Forschungsgemeinschaft (SFB 249).

References 1. Beron J, Mastroberardino L, Spillmann A, Verrey F (1995) Aldosterone modulates sodium kinetics of Na,K-ATPase containing an α1 subunit in A6 kidney cell epithelia. Mol Biol Cell 6: 261–271 2. Brochiero E, Banderali U, Lindenthal S, Raschi C, Ehrenfeld J (1995) Basolateral membrane chloride permeability of A6 cells: implication in cell volume regulation. Pflügers Arch 431: 32–45 3. Cantiello HF (1995) Role of the actin cytoskeleton on epithelial Na+ channel regulation. Kid Intern 48: 970–984 4. Cantiello HF, Stow JL, Prat AG, Ausiello DA (1991) Actin filaments regulate epithelial Na+ channel activity. Am J Physiol 261: C882–C888 5. Chalfant ML, Coupaye-Gerard B, Kleyman TR (1993) Distinct regulation of Na+ reabsorption and Cl– secretion by arginine vasopressin in the amphibian cell line A6. Am J Physiol 264: C1480–C1488 6. Crowe WE, Ehrenfeld J, Brochiero E, Wills NK (1995) Apical membrane sodium and chloride entry during osmotic swelling of renal (A6) epithelial cells. J Membr Biol 144: 81–91 7. Ehrenfeld J, Raschi C, Brochiero E (1994) Basolateral potassium membrane permeability of A6 cells and cell volume regulation. J Membr Biol 138: 181–195

279 8. Els WJ, Chou KY (1993) Sodium-dependent regulation of epithelial sodium channel densities in frog skin; a role for the cytoskeleton. J Physiol (Lond) 462: 447–464 9. Fischer H, Clauss W (1990) Regulation of Na+ channels in frog lung epithelium: a target tissue for aldosterone action. Pflügers Arch 416: 62–67 10. Garty H, Edelman IS (1983) Amiloride-sensitive trypsinization of apical sodium channels. J Gen Physiol 81: 785–803 11. Granitzer M, Bakos P, Nagel W, Crabbé J (1992) Osmotic swelling and membrane conductances in A6 cells. Biochim Biophys Acta 1110: 239–242 12. Granitzer M, Mountian I, Van Driessche W (1995) Effect of dexamethasone on sodium channel block and densities in A6 cells. Pflügers Arch 430: 493–500 13. Helman SI, Baxendale LM (1990) Analysis of a three-state model for apical Na+ channels of frog skin. J Gen Physiol 95: 647–678 14. Hoffmann B, Clauss W (1989) Time-dependent effects of aldosterone on sodium transport and cell membrane resistances in rabbit distal colon. Pflügers Arch 415: 156–164 15. Illek B, Fischer H, Clauss W (1990) Aldosterone regulation of basolateral potassium channels in alveolar epithelium. Am J Physiol 259: L230–L237 16. Kleyman TR, Ernst SA, Coupaye-Gerard B (1994) Arginine vasopressin and forskolin regulate apical cell surface expression of epithelial Na+ channels in A6 cells. Am J Physiol 266: F506–F511 17. Lang MA, Muller J, Preston AS, Handler JS (1986) Complete response to vasopressin requires epithelial organization in A6 cells in culture. Am J Physiol 250: C138–C145 18. Li JHY, Palmer LG, Edelman IS, Lindemann B (1982) The role of sodium channel density in the natriferic response of the toad urinary bladder to an antidiuretic hormone. J Membr Biol 64: 77–89 19. Ma TY, Hollander D, Thuy Tran L, Nguyen D, Hoa N, Bhalla D (1995) Cytoskeletal regulation of Caco-2 intestinal monolayer paracellular permeability. J Cell Physiol 164: 533–545 20. Marples D, Barber B, Taylor A (1996) Effect of a dynein inhibitor on vasopressin action in toad urinary bladder. J Physiol (Lond) 490: 767–774

21. Marunaka Y, Eaton DC (1990) Chloride channels in the apical membrane of a distal nephron A6 cell line. Am J Physiol 258: C352–C368 22. Morris R, Tousson A, Benos DJ, Schafer JA (1995) Microtubule disruption does not prevent vasotocin (AVT) activation of Na+ channels (abstract). J Am Soc Nephrol 6: 347 23. Niisato N, Marunaka Y (1997) Hyposmolality-induced enhancement of ADH action on amiloride-sensitive Isc in renal apithelial A6 cells. Jpn J Physiol 47: 131–137 24. Parczyk K, Haase W, Kondor-Koch C (1989) Microtubules are involved in the secretion of proteins at the apical cell surface of the polarized epithelial cell. Madin-Darby canine kidney. J Biol Chem 264: 16837–16846 25. Parisi M, Pisam M, Merot J, Chevalier J, Bourguet J (1985) The role of microtubules and microfilaments in the hydrosmotic response to antidiuretic hormone. Biochim Biophys Acta 817: 333–342 26. Penningroth SM, Cheung A, Bouchard P, Gagnon C, Bardin CW (1982) Dynein ATPase is inhibited selectively in vitro by erythro-9-[3-2-(hydroxynonyl)] adenine. Biochem Biophys Research Commun 104: 234–240 27. Penningroth SM, Rose P, Cheung A, Peterson DD, Rothacker DQ, Bershak P (1985) An EHNA-sensitive ATPase in unfertilized sea urchin eggs. Cell Motil 5: 61–75 28. Prat AG, Bertorello AM, Ausiello DA, Cantiello HF (1993) Activation of epithelial Na+ channels by protein kinase A requires actin filaments. Am J Physiol 265: C224–C233 29. Van Driessche W, Erlij D (1983) Noise analysis of inward and outward Na+ currents across the border of ouabain-treated frog skin. Pflügers Arch 398: 179–188 30. Van Driessche W, Lindemann B (1978) Low-noise amplification of voltage and current fluctuations arising in epithelia. Rev Sci Instrum 49: 52–57 31. Verrey F, Groscurth P, Bolliger U (1995) Cytoskeletal disruption in A6 kidney cells: impact on endo/exocytosis and NaCl transport regulation by antidiuretic hormone. J Membr Biol 145: 193–204 32. Wills NK, Millinoff LP, Crowe WE (1991) Na+ channel activity in cultured renal (A6) epithelium: regulation by solution osmolality. J Membr Biol 121: 79–90