Necturus gallbladder epithelial cells bathed in 10 mM HCO3/1% CO2 display sizable basolateral membrane conductances for Cl− (GClb) and K + (GKb). Lowering the osmolality of the apical bathing solution hyperpolarized both apical and basolateral membranes and increased the K +/Cl− selectivity of the basolateral membrane. Hyperosmotic solutions had the opposite effects. Intracellular free-calcium concentration ([Ca2+]i) increased transiently during hyposmotic swelling (peak at ∼30 s, return to baseline within ∼90 s), but chelation of cell Ca2+ did not prevent the membrane hyperpolarization elicited by the hyposmotic solution. Cable analysis experiments showed that the electrical resistance of the basolateral membrane decreased during hyposmotic swelling and increased during hyperosmotic shrinkage, whereas the apical membrane resistance was unchanged in hyposmotic solution and decreased in hyperosmotic solution. We assessed changes in cell volume in the epithelium by measuring changes in the intracellular concentration of an impermeant cation (tetramethylammonium), and in isolated polarized cells measuring changes in intracellular calcein fluorescence, and observed that these epithelial cells do not undergo measurable volume regulation over 10–12 min after osmotic swelling. Depolarization of the basolateral membrane voltage (Vcs) produced a significant increase in the change in Vcs elicited by lowering basolateral solution [Cl−], whereas hyperpolarization of Vcs had the opposite effect. These results suggest that: (a) Hyposmotic swelling increases GKb and decreases G Clb. These two effects appear to be linked, i.e., the increase in G Kb produces membrane hyperpolarization, which in turn reduces G Clb. ( b) Hyperosmotic shrinkage has the opposite effects on GKb and G Clb. ( c) Cell swelling causes a transient increase in [Ca2+]i, but this response may not be necessary for the increase in GKb during cell swelling.
Introduction
The Necturus gallbladder (NGB)1 epithelium is a useful model system for functional studies of a family of epithelia that perform near-isosmotic fluid absorption. The mechanisms accounting for ion and water transport in NGB epithelium under control conditions have been characterized in detail and considerable information is also available on regulation of transport (for reviews, see Reuss, 1988; Reuss and Altenberg, 1995). An issue requiring further study is the regulation of basolateral membrane Cl− and K+ channels, especially its relationship to the mechanism of cross talk between apical and basolateral membranes, in which cell volume changes could play a major role (Schultz and Hudson, 1991). For instance, we have proposed that elevation of cAMP causes a reduction in Cl− efflux across the basolateral membrane. This is based on the effects of cAMP on fluid transport (Petersen and Reuss, 1983), intracellular ion contents, and cell volume (reviewed in Reuss and Altenberg, 1995; Reuss et al., 1991). Further, exposure to cAMP reduces basolateral membrane conductance (Copello et al., 1993). We suspected that this may include a decrease in basolateral membrane GCl (GClb) brought about by the cell shrinkage produced by cAMP. In other words, the hypothesis is that the changes in apical membrane conductances elicited by cAMP result in a loss of cell KCl, which in turn causes cell shrinkage, and that the shrinkage results in a decrease in basolateral membrane G Cl. Accordingly, we assessed the effects of cell volume and membrane voltage on the Cl− and K+ conductances of the basolateral membrane. Specifically, we investigated the rapid effects of cell swelling and shrinkage on these conductances, in order to test the hypothesis that cell volume changes are involved in cross talk between apical and basolateral membranes. Inasmuch as GClb is small relative to the basolateral K + conductance (GKb), in addition to osmotic-shrinkage we carried out osmotic-swelling experiments. If cell swelling increases G Clb, then swelling experiments would yield relative changes in G Clb easier to measure than those elicited by cell shrinkage. In contrast with the apical membrane pathway (Copello et al., 1993; Heming et al., 1994), the G Clb appears to be cAMP independent (Copello et al., 1993) and is activated by HCO 3−/CO2 (Stoddard and Reuss, 1988a). The results of these studies demonstrate that GClb is increased by cell shrinkage and decreased by cell swelling, probably via changes in cell membrane voltage (depolarization during cell shrinkage, hyperpolarization during cell swelling). The basolateral K + conductance (GKb) is apparently voltage insensitive (Stoddard and Reuss, 1988 ,b), and possibly activated by elevations in internal Ca2+ (Bello-Reuss et al., 1981). Here, we demonstrate that the K+ conductance is stimulated by cell swelling and inhibited by cell shrinkage.
Methods
General
Mudpuppies (Necturus maculosus) were purchased from Kon's Scientific (Germantown, WI) or Nasco Biologicals (Ft. Atkinson, WI) and maintained at 5–10°C. Anesthesia was accomplished by immersion in a 1 g/liter tricaine methanesulfonate solution. The gallbladder was excised, opened, washed, and mounted in a modified Ussing chamber (Altenberg et al., 1990). When the gallbladders were mounted serosal side up, a patch of connective tissue was removed by dissection, to allow for microelectrode impalements through the basolateral membrane (Altenberg et al., 1990). Some experiments were carried out with polarized epithelial cells isolated from the intact tissue (“figure-eight cells”), using a method previously described (Torres et al., 1996b). The control bathing solution (bPSS) contained, in mM: 90 NaCl, 10 NaHCO3, 2.5 KCl, 1.8 CaCl2, 1.0 MgCl2, and 0.5 NaH2PO4, gassed with 1% CO2/99% air, pH 7.66, at room temperature, with an average osmolality of 207 mosmol/kg. A low-NaCl but isosmotic solution was prepared by replacing 35 mM NaCl with 70 mM sucrose. This solution was made hyposmotic (by ∼17 or 34%) by removing sucrose, or hyperosmotic (also by ∼17 or 34%) by adding a further 35 or 70 mM sucrose. Changes in bathing solution osmolality were made only in the apical bathing medium, inasmuch as reductions in basolateral solution [NaCl], with or without changes in osmolality, elicited long-lasting oscillations in membrane voltage and conductance. To assess cell membrane ionic selectivity, Cl− was replaced with gluconate, and Na+ was replaced with K+. These replacements were isomolar.
Electrophysiological Techniques
Transepithelial voltage (Vms, referenced to the basolateral bathing solution; Vsm, referenced to the apical bathing solution) and cell membrane voltages (apical = Vmc, basolateral = Vcs, referenced to the respective adjacent solutions) were measured as previously described (Altenberg et al., 1990). The ground electrode was an Ag-AgCl pellet separated from the apical bathing solution by a short Ringer-agar bridge. The basolateral bathing solution electrode was a flowing, saturated-KCl bridge in series with a calomel half-cell. Hence, liquid junction potentials upon changes in the basolateral solution were minimized. The transepithelial resistance, Rt (ΔVms/It, where ΔVms is the change in Vms elicited by a constant current pulse of density It), and the apparent ratio of cell membrane resistances (Ra/Rb, where the subscripts a and b denote apical and basolateral membranes, respectively) were determined from the steady-state voltage deflections elicited by DC pulses of 50 μA/cm2 and 1-s duration, applied across the tissue through Ag-AgCl electrodes. The voltage deflections were corrected for series resistances. The absolute values of Ra and Rb were estimated from two-point cable analysis, as previously described (Petersen and Reuss, 1985; Copello et al., 1993). Intracellular-microelectrode studies were also carried out on isolated polarized cells attached with Cell-Tak® (Collaborative Medical Products, Bedford, MA) to a dialysis membrane mounted in the microelectrode chamber (Torres et al., 1996b).
Intracellular Cl− activity (aCli) and tetramethylammonium (TMA+) activity (aTMAi) were measured with double-barrel ion-sensitive microelectrodes constructed and calibrated as previously described (Altenberg et al., 1990). Validation of impalements was as described before (Altenberg et al., 1990). To measure changes in cell water volume, epithelia were loaded with TMA+ using transient exposure to nystatin, and the intracellular activity of TMA+ was determined with double-barrel, TMA+-sensitive microelectrodes (Reuss, 1985; Altenberg et al., 1990).
Fluorescence Techniques
Changes in cell water volume were assessed in isolated polarized cells attached to a coverslip coated with Cell Tak®, loaded with calcein,AM (3 μM, for 20 min) and then superfused with isosmotic solution for 15 min before starting the experiment. Calcein is a good choice among fluorescent indicators to assess changes in cell water volume because it undergoes less photobleaching than other probes and is retained better in the cells (Altenberg et al., 1994; Alvarez-Leefmans et al., 1995; Crowe et al., 1995). The experiments were performed using a digital video confocal laser imaging system (Odyssey; Noran Instruments, Middleton, WI). Excitation light was 488 nm, and emitted light was measured at wavelengths longer than 515 nm. Pictures were obtained at 30-s intervals, and fluorescence of a 15–20 μm2 area in the center of a cell was measured. The records were corrected for fluorescence decay independent of cell volume changes (primarily due to dye photobleaching), which was fit by a single exponential. The data are presented as Fo/ Ft, where Fo = fluorescence in isosmotic solution, at t = 0, and Ft = corrected fluorescence at time = t. The ratio Fo/Ft is proportional to cell volume and was ∼75% of the ideal (“osmometric”) responses after exposure to hypo or hyperosmotic solutions.
Intracellular free [Ca2+] ([Ca2+]i) was estimated from the 340/380 nm Fura-2 fluorescence ratio (F340/380) in isolated polarized cells. Isolated polarized cells (Torres et al., 1996b) were loaded at room temperature with Fura-2,AM (5 μM for 1 h). The cells were then attached to a coverslip coated with Cell-Tak®, mounted in a chamber and superfused with isosmotic solution for ∼15 min before starting the experiment. Measurements were carried out essentially as described (Altenberg et al., 1994). Data were acquired at 1 Hz. At the end of each experiment, cells were exposed to 10–20 μM ionomycin to obtain saturating free [Ca2+]i, and then to 1–2 mM MnCl2 in the continuous presence of ionomycin, to quench the dye and thus correct for background fluorescence. Data are presented as F340/380 after background correction. In some experiments, cells were loaded with acetoxymethyl ester of 1,2-bis-(2-aminophenoxy)ethane-N,N,N ′,N ′-tetraacetic acid (BAPTA,AM), Half-BAPTA,AM, or N,N,N ′,N ′-tetrakis(2-pyridylmethyl) ethylenediamine (TPEN), a heavy-metal chelator (Kao, 1994). All of these chemicals were obtained from Molecular Probes (Eugene, OR).
Statistical Analysis
Results are given as means ± SEM. Statistical comparisons were done by t tests for paired or unpaired data, as appropriate. A value of P < 0.05 was considered significant.
Results
Effects of Changes in Apical Bathing Solution Osmolality on Voltages and Resistances
The effects of lowering the apical solution osmolality (from 200 to 135 mosmol/kg) on the basic electrophysiological properties of the epithelium are illustrated in Fig. 1. In this experiment, a cell was impaled from the apical side with a conventional microelectrode. Voltages and resistances were recorded before, during, and after exposure to hyposmotic solution on the apical side. Reducing bathing solution osmolality caused a hyperpolarization of Vcs and an increase in Ra/Rb. Both effects were reversible. Table I summarizes the results of experiments such as the one shown in Fig. 1. The hyperpolarization at 3 min was 8 ± 1 mV, and the repolarization was complete after 4 min of perfusion with isosmotic solution.
The experiment shown in Fig. 2 illustrates the effects of exposure to a hyperosmotic solution (the apical solution osmolality was increased from 200 to 270 mosmol/ kg), and Table II summarizes the results with this experimental protocol. The main effects of hyperosmotic solution were cell membrane depolarization and a decrease in Ra/Rb, i.e., opposite changes to those elicited by hyposmotic solution. As in the experiments with hyp-osmotic solution, the effects were reversible. The effect of anisosmotic solutions on Vms was small, but significant, apical side negative with hyposmotic solution and opposite in polarity with hyperosmotic solution. These changes can be explained by paracellular pseudo-streaming potentials caused by water flow from apical to basolateral solution with hyposmotic solution and from basolateral to apical solution with hyperosmotic solution. These voltage changes, because of their small magnitude, do not contribute significantly to the changes in cell membrane voltages (Reuss et al., 1992a, 31b). There were also significant changes in Rt: a decrease with hyp-osmotic solution and an increase with hyperosmotic solution. These changes are in the directions expected for widening and narrowing of lateral-intercellular spaces, respectively (Reuss et al., 1992a), but changes in junctional resistance cannot be ruled out.
Effects of Changes in Apical Bathing Solution Osmolality on Relative K+ and Cl− Conductances of the Basolateral Membrane
The changes in Vcs elicited by increasing basolateral [K+] from 2.5 to 25 mM (ΔVcsK), or by lowering basolateral [Cl −] from 98.1 to 8.1 mM (ΔVcsCl), were measured in the same cells, during exposure to isosmotic apical bathing solution and at 20 min of exposure to hyposmotic solution. Similar results were obtained between 5 and 20 min in hyposmotic bathing medium (not shown). As illustrated in Fig. 3 and summarized in Table I, exposure to hyposmotic solution increased ΔVcsK and decreased ΔVcsCl, relative to the respective values measured in isosmotic solution. Further, ΔVcsCl in hyposmotic solution was not significantly different from zero.
Similar experiments were carried out to assess the changes in basolateral membrane ionic selectivity after a 5–10 min exposure to hyperosmotic solution. The results are illustrated in Fig. 4 and summarized in Table II. In hyperosmotic medium, ΔVcsK was reduced significantly, while ΔVcsCl was increased. Both changes are opposite to those elicited by exposure to hyposmotic solution.
Because of the low paracellular electrical resistance of the NGB epithelium, the values of ΔVcsK and ΔVcsCl are less than the corresponding changes in zero-current membrane voltages (the ionic substitutions produce changes in intraepithelial current flow). There are also concomitant paracellular diffusion potentials, but these are small and in both instances (high-K+ and low-Cl−) tend to compensate in part for the shunting effect of the paracellular pathway (see Reuss and Finn, 1975).
These results demonstrate that the relative GKb increases and relative G Clb decreases during exposure to hyposmotic solution. Conversely, the relative G Kb decreases, and the relative G Clb increases during exposure to a hyperosmotic solution. To estimate the changes in absolute values of G Kb and G Clb in both sets of experiments, we performed the experiments described in the next section.
Effects of Exposure to Anisosmotic Solutions on Cell Membrane Resistances
Fig. 5, A and B, depicts representative two-point cable analysis experiments under control conditions and during transient exposure to hyposmotic and hyperosmotic solutions. Two cells were simultaneously impaled, and both transepithelial and intracellular current pulses were applied at intervals. Downward (hyperpolarizing) voltage spikes denote ΔVx, i.e., the voltage changes elicited by current injection into another (electrically coupled) cell. Exposure to hyposmotic solution caused a decrease in ΔVx, which reached a stable value in ∼4 min and returned to control levels within 4–5 min of restoring superfusion with isosmotic solution (Fig. 5,A). In contrast, exposure to hyperosmotic solution increased ΔVx. This effect was complete within 4 min and returned to control levels after a 4-min exposure to isosmotic solution (Fig. 5,B). Fig. 5 C depicts fits of the Bessel function K0 to the two-point cable analysis values obtained with hyposmotic and hyperosmotic solutions, respectively. From these data, we calculated Rz (resistance for current flow out of the cell, i.e., Ra and Rb in parallel); from Rz and the concomitant values of Ra/Rb and Rt, we calculated the resistances of the cell membranes and the paracellular pathway. The method used requires normalization of the data to the values observed under control conditions. The bases, assumptions, and limitations of this method have been discussed in detail elsewhere (Petersen and Reuss, 1985; Copello et al., 1993).
The results of these calculations are summarized in Table III. Exposure to anisosmotic solutions elicits mostly changes in basolateral membrane resistance: Rb decreases during superfusion with hyposmotic solution and increases during superfusion with hyperosmotic solution. Whereas hyposmotic solution causes exclusively a decrease in Rb, hyperosmotic solution both increases Rb and decreases Ra. The latter effect could be due to activation of apical membrane maxi-K+ channels by depolarization (Segal and Reuss, 1990).
The ionic-substitution experiments illustrated in Figs. 3 and 4 and summarized in Tables I and II demonstrated an increase in relative GKb with hyposmotic solution and a decrease in relative G Kb with hyperosmotic solution. These results, together with the resistance data, conclusively demonstrate that the changes in basolateral membrane conductance are principally due to changes in G Kb in the same direction: increase with hyposmotic solution and decrease with hyperosmotic solution. Rather unexpectedly, we found that the changes in GKb and G Clb were in opposite directions: when G Kb increased (hyposmotic solution), G Clb decreased; when G Kb decreased (hyperosmotic solution), G Clb increased. We next investigated the mechanism of the changes in G Kb and G Clb.
Cell Swelling Causes Elevation of [Ca 2+]i , But the Latter Is Not Necessary for the Increase in GKb
Entry of Ca 2+ has been shown to play a signaling role in cell volume regulation from hyposmotic swelling by activation of a Ca2+-dependent GK (Christensen, 1987). Therefore, the effects of hyposmotic solution on [Ca2+]i in isolated polarized cells were estimated from the F340/F380 Fura-2 fluorescence ratio. The results are shown in Fig. 6. The cells were exposed to a 34%-hyposmotic solution under control conditions or after loading with the Ca2+ chelator BAPTA,AM. After the onset of superfusion with hyposmotic solution, [Ca2+]i rose rapidly and then (during continuous exposure to this solution) decreased monotonically to control levels in <2 min. In cells preloaded with BAPTA, the elevation in [Ca2+]i by exposure to hyposmotic solution was abolished. Experiments were also carried out with 50 μM Half-BAPTA,AM and 50 μM TPEN, a heavy-metal chelator (Kao, 1994). Neither compound binds Ca2+ in the nanomolar range. In both instances, the change in [Ca2+]i after exposure to hyposmotic solution was similar to that observed under control conditions (not shown). Hence, the effect of BAPTA cannot be attributed to nonspecific toxic effects or to heavy-metal chelation. Because of the effect of pH on Fura-2 fluorescence (Reers et al., 1989), we also measured pHi with BCECF, using the same protocol as for the [Ca2+]i measurements. Hyposmotic solution did not cause measurable changes in pHi (data not shown). These control experiments validate the conclusion that [Ca2+]i does in fact rise transiently during exposure of NGB epithelial cells to hyposmotic solutions.
To determine the role of [Ca2+]i in the response of isolated polarized cells to hyposmotic swelling, we examined the effect of swelling (34% hyposmotic solution) on membrane voltage (Vm) with or without preloading the cells with 50 μM BAPTA (see methods). Without BAPTA, Vm hyperpolarized by 4 ± 1 mV; with BAPTA, the hyperpolarization was 3 ± 1 mV (not significantly different; n = 4 paired experiments).
GClb Is Modulated by Membrane Voltage
The changes in G Clb described above could result from: ( a) direct effect of changes in cell volume, (b) changes in membrane voltage produced by the activation or inhibition of GKb, or ( c) other mechanisms. In other experiments, we have observed decreases in GClb associated with basolateral membrane hyperpolarization (Stoddard and Reuss, 1989; Altenberg et al., 1992; Altenberg et al., 1993). Cell swelling activates G Kb; this in turn would hyperpolarize the membrane (because E K \> Vcs), and the hyperpolarization would cause a decrease in GClb. The opposite effects, i.e., decrease of G Kb, depolarization, and activation of G Clb, would be elicited by cell shrinkage. To test the hypothesis that the effects of cell volume changes on G Clb are mediated by changes in membrane voltage, we studied the effects of baseline V cs on ΔVcsCl. Depolarization of V cs was accomplished either by elevating apical solution [K+] from 2.5 to 25 mM (the change in Vcs results from a loop-current change, see Reuss and Finn, 1975); hyperpolarization of Vcs was accomplished by a transepithelial current clamp. The results are shown in Fig. 7. With depolarization of Vcs, ΔVcsCl increased, and with hyperpolarization of V cs, it decreased significantly. In the latter experiment, an extracellular electrode within 5 μm from the impaled cell was used as reference to avoid the effect of changes in voltage drops in the solutions (produced by the different conductivities of high-Cl−and low-Cl− solutions). We conclude that GClb is voltage dependent: decreased by hyperpolarization and increased by depolarization of V cs.
In conclusion, these experiments show that after swelling of NGB epithelial cells GKb increases, causing basolateral membrane hyperpolarization, which may explain the parallel decrease in G Clb. In previous studies, we found no increase of apical membrane Cl − conductance after cell swelling by exposure to hyposmotic solutions (Heming et al., 1994). This is consistent with the lack of change in Ra (see above). Hence, our studies do not support the notion that NGB epithelial cells undergo volume-regulatory decrease by loss of K+ and Cl−via channels. However, cell volume measurements using an optical-sectioning technique have been reported to show virtually complete cell volume regulation within a few minutes of the onset of the osmotic perturbation (Persson and Spring, 1982; Larson and Spring, 1984; Furlong and Spring, 1990). Hence, it was of interest to test whether cell volume regulation takes place regardless of the decrease in GClb. This was done in the next series of experiments.
Effects of Exposure to Hyposmotic Solution on Cell Water Volume and Intracellular aCl i
Cell water volume and aCli were measured before, during, and after exposure to hyposmotic solution (apical side alone). In Fig. 8,A we illustrate a cell water volume measurement in TMA+-loaded cells. Exposure to hyp-osmotic bathing solution caused an increase in cell volume, evidenced by the decrease in [TMA+]i. However, there was no measurable cell volume regulation, although during this time GKb is clearly activated (see Figs. 1 and 3 and Tables I and III). One could argue that the experimental method, namely the TMA + loading by transient exposure to nystatin, caused a fall in GKb or otherwise changed the properties of the cells, abolishing the volume regulatory response. To test this possibility, we measured aCl i before, during, and after exposure to hyposmotic solution, without TMA+ loading. The expectations are that hyposmotic solution will initially cause a dilution of intracellular Cl−; if there is cell volume regulation by KCl efflux, then aCli would fall further during the regulatory response; if there is cell volume regulation by a different mechanism, without Cl− loss, then aCli would rise during the regulatory response. Fig. 8,B illustrates the result of this experiment. Exposure to hyposmotic apical solution produced a monotonic change in aCli, i.e., there was no indication of more than one process (water influx) changing aCli. Further, Fig. 8,C depicts the intracellular TMA+ and Cl− activities during the first 5 min of exposure to the hyposmotic solution (records from Fig. 8, A and B). The linear relationship indicates that the rates of change in both ion activities are the same, i.e., that the time courses of aCli and aTMAi mirror each other. In addition, the maximal fractional changes in the two activities did not differ significantly (see Fig. 8). Inasmuch as TMA+ is a cell water volume marker, i.e., effectively impermeant, this result indicates that during the 5-min period Cl− was also effectively impermeant, i.e., its cell content did not change. If cell volume regulation were very fast, one could envision a monotonic “subosmometric” change in aTMAi after hyposmotic swelling (cell water volume tends to increase because of water influx and to decrease because of regulatory efflux of solute and water). However, the magnitude of the change in aTMAi is that expected for a cell behaving as an ideal osmometer, and the parallel measurements of aCli do not show the higher rate of decrease predicted if Cl− were involved in the regulatory response, or the lower rate of decrease expected if the regulatory volume decrease did not involve Cl−.
The possibility of cell volume regulation after exposure to hyposmotic solution was further explored in a preparation of isolated NGB epithelial cells that retain structural and functional polarity (Torres et al., 1996a, 40b). Changes in cell water volume were estimated from the changes in intracellular fluorescence in cells loaded with calcein, as illustrated in Fig. 9,A, B, and C. Exposure of isolated polarized cells to hyposmotic (Fig. 9,D) or hyperosmotic solutions (Fig. 9 E) elicited rapid changes in intracellular fluorescence, indicative of swelling and shrinkage, respectively, with no regulatory volume changes during the >10-min period of exposure to anisosmotic solution. These experiments confirm the results described in the preceding section, and argue against the possibility that the lack of volume regulatory responses results from artefacts related to microelectrode impalements.
To eliminate the possibility of an artifactual inability to measure a putative volume regulatory response, we exposed cAMP-stimulated cells to an isosmotic, high-K+ external solution ([K+] was raised from 2.5 to 67.5 mM, replacing Na+), at constant external [Cl−] (63.1 mM). The result, illustrated in Fig. 9 F, was rapid and reversible cell swelling, consistent with previous observations made with the TMA+ technique in the assembled epithelium (Cotton and Reuss, 1991). These results further strengthen the conclusion that, in the experimental conditions of these studies, osmotically swollen NGB epithelial cells do not undergo short-term regulatory volume decrease.
Discussion
Schultz (1981; see also Schultz and Hudson, 1991) developed the important notion that changes in transport rate at one of the membrane domains of an epithelium must result in “matching” changes in transport rate at the opposite membrane domain, so that in the steady state cell volume and composition remain near constant. This adaptive mechanism requires a signaling system between the cell membranes, i.e., intermembrane cross talk. The study of the mechanisms of cross talk between apical and basolateral membranes is of great importance to understand the regulation of salt and water transport in gallbladder and other epithelia. Parameters possibly involved in cross talk are cell volume, intracellular ionic activities, and membrane voltage.
One instance in which parallel changes in transport rates of apical and basolateral membranes occur in NGB epithelium is that resulting from the effect of cAMP. After an increase in cAMP levels, it has been demonstrated that an apical Cl− conductance, not present under control conditions, develops and dominates the cell membrane conductances (Copello et al., 1993; Heming et al., 1994; Reuss and Altenberg, 1995). This results in net Cl− loss across the apical membrane (Cotton and Reuss, 1991), because intracellular Cl− is above the value predicted from electrochemical equilibrium as a consequence of the operation of the apical membrane Cl−/HCO3− exchanger (Reuss, 1988). Gar-vin and Spring (1992) have also suggested that cAMP elevation results in a change in the dominant mechanism for Na + and Cl− transport across the apical membrane: from Na+/H+ and Cl−/HCO3− exchanges to Na+-Cl− cotransport. Further, Dausch and Spring (1994) have proposed that protein kinase C activation causes the opposite effect. Regardless of these possibilities, the increase in apical membrane GCl causes membrane depolarization and activation of voltage-sensitive apical membrane maxi-K+ channels (Cotton and Reuss, 1991; Reuss, 1991). The parallel increases in apical membrane GCl and GK cause net KCl efflux, intracellular Cl− and K+ contents fall, the cells shrink (Cotton and Reuss, 1991), and transepithelial salt and water transport decrease. The latter result indicates that net transport across the basolateral membrane must also be reduced. Our starting hypothesis was that cAMP causes a reduction in GClb. Hence, we assessed the effects of changes in cell water volume on basolateral membrane Cl − and K+ conductances.
Cell Swelling Stimulates and Cell Shrinkage Inhibits the Basolateral Membrane K+ Conductance
Our results show that cell swelling causes a selective increase in basolateral membrane conductance, attributable to an increase in K+ conductance. This effect is teleologically appropriate for cell volume regulation, inasmuch as it would facilitate K+ loss from the swollen cells.
The opposite experimental perturbation, i.e., cell shrinkage by exposure to a hyperosmotic solution, caused depolarization of both cell membranes, an increase in basolateral membrane resistance, and a decrease in the relative GKb. These effects are opposite to those elicited by exposure to hyposmotic solution and indicate that cell shrinkage decreases G Kb. Again, this effect is teleologically appropriate, i.e., net K + efflux across the basolateral membrane would decrease in a cell that has undergone shrinkage.
The elevation in [Ca2+]i after exposure to a hyposmotic solution could be solely or in part responsible for the increase of the GKb, inasmuch as this conductance is activated by maneuvers expected to elevate [Ca 2+]i (Bello-Reuss et al., 1981). Activation of nonselective cation channels, followed by Ca2+ entry and activation of Ca2+-activated K+ channels, has been demonstrated in other epithelia (Christensen, 1987). Our measurements demonstrated that [Ca2+]i does rise during cell swelling; chelation with BAPTA prevents the elevation in [Ca2+]i but not the membrane hyperpolarization. Although this result could be interpreted to indicate that [Ca2+]i plays no role in the activation of GKb, we cannot rule out that, in the presence of BAPTA, [Ca 2+]i rose in a small compartment, undetectable by our method. An alternative to Ca2+ activation is that the basolateral K+ channels are activated by stretch.
Cell Swelling Inhibits the Basolateral Membrane Cl− Conductance
Cell swelling reduced the relative GClb to a value not different from zero, indicating that the basolateral membrane Cl − conductance decreases. This effect is contrary to expectations, inasmuch as it is opposite to the effect of cell swelling on GKb and would, per se, tend to prevent KCl efflux via conductive pathways.
The effect of cell swelling on G Clb could be indirect, since it occurs during membrane hyperpolarization. We found that membrane hyperpolarization per se reduces G Clb, as indicated by basolateral solution Cl −-substitution experiments at three membrane voltages. Atypical osmosensitivity of Cl− channels has been observed in other cell types (Chesnoy-Marchais and Fritsch, 1994). It is possible that cell swelling and membrane hyperpolarization have separate effects on GClb, but the membrane voltage changes appeared to dominate under the present experimental conditions.
Implications of the Results for Cell Volume Regulation
Osmotic swelling experiments yielded the expected increase in the total basolateral membrane conductance. The unexpected feature was that changes in G Clb do not parallel the changes in G Kb, because the basolateral membrane Cl − conductance is voltage sensitive. Hence, it is possible that the effects of cell volume (e.g., via membrane stretch) are directly exerted on GKb and that the resulting change in membrane voltage is responsible for the change in G Clb.
Our cell volume measurements suggest that NGB epithelial cells do not regulate their volume after exposure to anisosmotic solution. Our measurements of intracellular [TMA +] and aCli are in apparent contradiction with the results of Spring and associates (Persson and Spring, 1982; Larson and Spring, 1984; Furlong and Spring, 1990), which suggested that the epithelial cells from NGB undergo regulatory volume decrease from hyposmotic swelling by basolateral exit of K+ and Cl−, probably via basolateral membrane channels (Furlong and Spring, 1990). In contrast, our studies suggest that there is no measurable short-term cell volume regulation in NGB epithelial cells undergoing osmotic swelling. The experiments on isolated, polarized cells strengthen our conclusions in that they rule out artifacts resulting from TMA+ loading and microelectrode impalement as possible explanations for the lack of cell volume regulation. Further, measurements of cell volume changes in isolated polarized cells in which apical membrane GCl was activated by cAMP showed that although these cells do not regulate rapidly their volume after osmotic swelling (data not shown), they undergo fast, sizable swelling after elevation of medium [K+], i.e., under isosmotic conditions. This again indicates that our results cannot be attributed to a methodological inability to detect cell swelling in isosmotic conditions. The explanation for the differences between our results and those of Spring and associates (e.g., Furlong and Spring, 1990) is not clear.
Implications of the Results for Cross Talk
Our results indicate that cell swelling causes rapid and reversible changes in basolateral membrane Cl−- and K+-channel activities. The demonstration that GKb is elevated by cell swelling supports the notion that cell volume is a signal responsible for cross talk between apical and basolateral membranes. However, under these circumstances G Clb not only does not increase, but in fact decreases. This would not be the expected response if the cell had been swollen by increased solute entry across the apical membrane. Under these conditions, the expected cross talk response would be to increase solute exit across the basolateral membrane. It is in principle possible that net solute exit is activated by cell swelling, but not via conductive pathways. A pathway for basolateral membrane Cl − efflux in NGB epithelium is KCl cotransport (Corcia and Armstrong, 1981; Reuss, 1981). Activation or inactivation of this mechanism during cell swelling and cell shrinkage, respectively, would fulfill the requirements of intermembrane cross talk, but would also result in cell volume regulation, and hence this possibility is not supported by our results.
The present observations suggest that GClb is regulated mainly or exclusively by V m changes, although a minor role of cell volume is possible. The dominant role of membrane voltage suggests that this regulatory mechanism is designed to preserve intracellular Cl− content instead of cell volume, i.e., there is an inverse relationship between electrical driving force and electrodiffusive permeability. Because of the hyperpolarization elicited by activation of GKb after cell swelling, G Clb falls, and this prevents Cl − loss and tends to maintain aCli closer to control levels. The need for this mechanism from the point of view of cell homeostasis is not obvious; cell Cl− content ends up being preserved at the expense of cell volume. In other cell types, intracellular [Cl−] has been shown or claimed to modulate membrane transporters (e.g., Robertson and Foskett, 1994) and regulators, such as protein kinases (Treharne et al., 1994). In NGB epithelium, in addition to effects derived from these “messenger” roles of intracellular Cl−, a likely consequence of a decrease in intracellular [Cl−] is a fall in intracellular pH. The apical membrane of this epithelium expresses a highly active anion (Cl−/HCO3−) exchanger (Reuss, 1988). A fall in aCl i at constant [Cl−] in the apical bathing solution enhances Cl− influx and HCO3− efflux across the apical membrane, which would cause cell acidification. In conclusion, osmotic swelling of NGB epithelial cells increases G Kb, but reduces G Clb, and hence there is no appreciable short-term regulatory volume decrease. If cell swelling is involved in cross talk between apical and basolateral membranes, then the effects of the increase in cell volume are not sufficient to provide an effective negative-feedback mechanism to restore cell volume to control levels, suggesting a more complex adaptive mechanism.
Acknowledgments
We thank Drs. S.A. Lewis and S.A. Weinman for comments on a preliminary version of the manuscript, F. Bavarian and K. Spilker for technical support, and L. Durant for secretarial help.
This work was supported by National Institutes of Health Grant DK-38734.
references
Dr. Torres's present address is Department of Experimental Medicine, University of Chile, Independencia 1027, Santiago, Chile.
Abbreviations used in this paper: ΔVcsCl, changes in V cs elicited by lowering external [Cl−]; ΔVcsK, changes in V cs elicited by raising external [K+]; NGB, Necturus gallbladder; Vcs, basolateral membrane voltage; Vmc, apical membrane voltage; Vms, transepithelial voltage.
Author notes
Address correspondence to Luis Reuss, M.D., Department of Physiology and Biophysics, University of Texas Medical Branch, Galveston, TX 77555-0641. Fax: 409-772-3381; E-mail: lreuss@mspo2.med.utmb.edu