Changes in cell volume measured with an electrophysiologic technique

Transcription

1 Proc. Natl. Acad. Sci. USA Vol. 82, pp , September 1985 Physiological Sciences Changes in cell volume measured with an electrophysiologic technique (water transport/epithelial transport/ion-sensitive microelectrodes) Luis REUSS Department of Physiology and Biophysics, Washington University School of Medicine, St. Louis, MO Communicated by Gerhard Giebisch, April 29, 1985 ABSTRACT Epithelial cells of the gallbladder of Necturus maculosus were loaded with tetramethylammonium (Me4N+) by transient exposure of the apical (lumen-facing) surface to a solution of high Me4N+ concentration containing also the polyene antibiotic nystatin. Upon removal of nystatin, in the continued presence of Me4N+, spontaneous restoration of the native ionic permeability of the apical cell membrane was observed. At this time, external Me4N' was removed; intracellular [Me4N+] measured with ion-sensitive microelectrodes was 2-15 mm and remained unchanged for several hours. Changes in cell volume were estimated from the changes in intracellular [Me4N+] produced by alterations in the osmolality of the mucosal bathing solution. Assuming that such changes are caused entirely by water fluxes across the apical membrane, the minimum value of its hydraulic permeability coefficient (Lp) was 1-3 x 10-3 cm-sec'-1(osmoles/kg)'1, suggesting that an osmolality difference across the apical membrane as small as 1-3 milliosmoles/kg could explain the average rate of transepithelial water transport. These results agree with optical measurements [Persson, B. 0. & Spring, K. R. (1982) J. Gen. Physiol. 79, ]. The effective thickness of the apical unstirred layer was estimated from the time courses of both the apical membrane voltage and the response of an extracellular K'-sensitive micr-oelectrode to an increase in [K'] in the mucosal bath. Since changes in concentration of the osmotically active solute at the membrane surface were thus shown to be significantly delayed by diffusion, the Lp value, calculated assuming a step-change in osmolality, is an underestimate. The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C solely to indicate this fact Accurate and rapid measurements of changes in cell volume can provide information about the mechanisms of water transport across cell membranes and of cell-volume maintenance and regulation. In combination with measurements of intracellular ionic activities, they allow estimation of net transmembrane ion fluxes with excellent time resolution. Cell volume measurements are particularly difficult in situ. An elegant computer-assisted video technique in which thin serial sections of the cell are measured has been applied to epithelial cells of gallbladder and isolated renal tubules (1-3). However, this method imposes restrictions on the optical properties of the preparation, requires expensive equipment, and in flat epithelia is rather slow, allowing a measurement only every 6 sec. I have developed an alternative technique, which can, in principle, be applied to any cell type suitable for intracellular microelectrode recordings, and used it to measure volume changes in epithelial cells of the gallbladder of Necturus maculosus. By means of transient exposure to the polyene antibiotic nystatin, cells were loaded with tetramethylammonium (Me4N+), an impermeant ion whose activity can be accurately measured with intracellular microelectrodes, allowing for calculation of the intracellular concentration. After the loading procedure the intracellular content of this probe is constant, so that changes in its concentration simply reflect changes in cell volume. MATERIALS AND METHODS Preparation and Solutions. Gallbladders were removed from anesthetized animals as described (4) and mounted apical (mucosal)-side-up in a chamber that allows for continuous separate superfusion of mucosal and serosal sides. The control physiologic salt solution (NaCl-Ringer's solution) had the following composition: 90 mm NaCl/10 mm NaHCO3/2.5 mm KCl/1.8 mm CaCl2/1.0 mm MgCl2/0.5 mm NaH2PO4; it was equilibrated with a 1% C02/99% air mixture and had a ph of about 7.65 and an osmolality ofabout 200 milliosmoles/kg. For Me'N' loading, NaCl was replaced with an equimolar concentration of Me4N cyclamate of Me4N sulfate (in the latter case, sucrose was added to maintain tonicity). Chloride removal during most of the period of exposure to nystatin was necessary to prevent cell swelling (5). Hyposmotic and hyperosmotic solutions were prepared by varying the sucrose or mannitol concentration of a solution containing 70 mm NaCl plus the same concentrations of all other salts listed above for NaCl-Ringer's solution or by altering the NaCl concentration of the Ringer's solution. The range of change in osmolality was ±20%. Osmolalities of all solutions were measured in triplicate. Electrophysiologic Techniques. Cell-membrane voltages and intracellular Me4N' concentration ([Me4N+]I) were determined from simultaneous impalements of two neighboring epithelial cells with a conventional microelectrode (filed with 3 M KCl) and a K+-sensitive microelectrode (see below), a valid method because, in this preparation, all cells are electrically coupled (4, 6). After Me4N' loading, the membrane voltages measured in several cells in the same preparation were essentially the same. Microelectrode construction and impalement validation were as described (7, 8). K+-sensitive, liquid-membrane microeletrodes, such as the ones based on the exchanger potassium tetrakis (pchlorophenylborate), are known to be far more responsive to quaternary ammonium ions than to K+ (9). The voltage response to Me4N' of such a microelectrode is shown in Fig. 1. Since in some cases the slope of the relationship between voltage and log[me4n+] varied with [Me4N+], all calibrations were carried out at [K+] = 120 mm (a value close to that measured in the epithelial cells under control conditions) and [Me4N+] ranging from 0.1 to 20 mm. Only electrodes with constant slopes of at least 53 mv/log[me4n+] in the range 1-10 mm were used. Within this concentration range, the selectivity coefficient (Me4N+/K+) was on the average 0.9 x 103. To calculate [Me4N+]i from the membrane voltage and the voltage output of the ion-sensitive microelectrode, I Abbreviation: [Me4N']j, intracellular tetramethylammonium concentration.

2 140 r Physiological Sciences: Reuss Proc. NatL. Acad Sci. USA 82 (1985) 6015 I Nystatin7 E 100lo 80k _ [Me4N'I, mm FIG. 1. Calibration of a Me4N+-sensitive microelectrode. V* is the voltage recorded by the microelectrode in a rapid-flow chamber with respect to a calomel half-cell connected to the solution by a saturated-kcl, flowingjunction. The rate of exchange of the solution was high, to avoid changes in [K+]. The calibration solutions consisted of 120 mm KCl plus the Me4NCl concentration given in the abscissa. In the range 1-20 mm, the slope was 54 mv/log[me4n+]. assumed equal ionic strengths in intra- and extracellular compartments. Me4N' Loading. The native cell membranes of the epithelium of Necturus gallbladder are impermeable to Me4N +, but the apical membrane develops a high Me4N' permeability when treated with the pore-forming antibiotic nystatin (see below). Accordingly, the following protocol was developed to incorporate Me4N' into the intracellular compartment and to prevent its later loss. First, Me4N' was allowed to enter the cells by exposing their apical surface to a "high-me4n`" solution (isomolar replacement of Na' with Me4N', see above) containing nystatin (Sigma) at units/ml. Then, nystatin was removed, whereas the high extracellular Me4N' concentration was maintained until the conductive properties of the cell membrane, among them its Me4N' impermeability, were completely restored. After cell membrane "resealing" was ascertained by the criteria described below, Me4N' was removed from the bathing medium without causing measurable loss from the intracellular compartment, and [Me4N+]i was continuously measured by simultaneous impalements with conventional and K+-sensitive microelectrodes. The changes in [Me4N+]i and hence in cell volume elicited by alterations of the bathing medium could be thus recorded with high accuracy and excellent time resolution. RESULTS AND DISCUSSION Validation of the Me4N' Loading Technique. The apical membrane of Necturus gallbladder epithelial cells is predominantly K+-conductive and has a low electrodiffusive Na+ permeability (10, 11). The changes in its electrical properties during and after Me4N' loading are illustrated in Fig. 2. Exposure to nystatin caused dramatic falls in the cell membrane voltages and in the apparent ratio of electrical resistances of the two cell membranes (apical/basolateral, Ra/Rb), indicating an increase in ionic permeability of the apical membrane. These observations agree with those reported for rabbit urinary bladder epithelium (5). During nystatin treatment, replacing mucosal Na+ with K+ caused further cell membrane depolarization (data not shown), whereas replacing the mucosal Na+ with Me4N' caused a o Time, min FIG. 2. Effects of nystatin ( units/ml) on electrical properties of Necturus gallbladder epithelial cells. (Upper) Apical membrane voltage (Vmc) (cell-mucosal bathing medium) with NaCl- Ringer's solution on both sides (open bars), during an increase in mucosal solution [K+] to 82.5 mm (hatched bars), and during mucosal exposure to Me4N sulfate-ringer's solution (solid bars). (Lower) Apparent ratio of cell membrane resistances (Ra/Rb), calculated from the voltage deflections elicited across each cell membrane by transepithelial current pulses of ca. 50,uA cm2, after correction for series resistances. o, Values during bilateral exposure to NaCl-Ringer's solution. e, Values during mucosal exposure of Me4N sulfate. Nystatin causes a large cell membrane depolarization, a mucosa-negative change in transepithelial voltage, and a fall in Ra/Rb to a value not different from zero. Replacing mucosal NaCl with Me4N sulfate repolarized partially the apical membrane and increased significantly Ra/Rb. Removal of nystatin, during continuing exposure to Me4N sulfate, resulted in slow recovery of Vmc and Ra/Rb toward control values. Forty minutes after removal of the antibiotic, the tissues were exposed again to NaCl-Ringer's solution on both sides. Vmc, Ra/Rb, and the change in apical membrane voltage produced a high mucosal [K+] (difference between open and hatched bar) were not different from the analogous values measured before nystatin treatment (pair analysis). Means ± SEM are shown; n = 6 experiments. significant but small increase in Ra/Rb and an increase in the voltages across both cell membranes. These results indicate that in the nystatin-treated membrane, the cationic electrodiffusive permeabilities follow the sequence K+ > Na+ > Me4N'. The Me4N' permeability of the nystatin-treated membrane is sizable, as shown by the low values of membrane voltage and of Ra/Rb and by the demonstration of a value of [Me4N+]j in the millimolar range upon completion of the loading procedure (see below). Upon removal of nystatin, while the apical surface of the cells continued to be exposed to Me4N cyclamate or sulfate, Ra/Rb rose slowly to a value close to that observed before exposure to the antibiotic. When, about 40 min after removal of nystatin, the tissues were exposed again to NaCI-Ringer's solution, the membrane voltages, the value of Ra/Rb, and the apical-membrane voltage changes elicited by an increase in mucosal K+ concentration (isomolar replacement of Na+) were not different from those observed before nystatin treatment (Fig. 2). These results show that after removal of nystatin, the cation electrodiffusive permeabilities of the apical membrane

3 6016 Physiological Sciences: Reuss are fully restored. At this time, in six preparations, [Me4N+]J ranged from about 2 to about 15 mm. Although thereafter the preparations were exposed to Me4N+-free bathing solutions, the [Me4N+]i did not change appreciably for as long as 3 hr. Native intracellular ions, such as K+, Na', or Cl-, could in principle be employed as cell volume markers, as done by others (12). However, preliminary experiments with K+ and Cl- revealed rapid changes in intracellular content during alterations in external osmolality, indicating rapid net fluxes across the cell membrane. Measurement of Changes in Cell Volume. Continuous measurements of [Me4N+]i while the osmolality of the mucosal bathing solution was rapidly changed as shown in Fig. 3. As expected, exposure to hyposmotic and hyperosmotic mucosal solutions caused reversible decreases and increases in [Me4N+]i, respectively. The time courses of these changes were related to the volume of the mucosal bathing solution and its rate of exchange. With a typical mucosal volume of about 1 ml and a mucosal solution flow rate of about 30 ml-min-1, the changes in [Me4N+]i approached a steady state in about 20 sec. At this time, the fractional cell-volume change was not different from that predicted from the fractional change in external osmolality if the cell behaves as a perfect osmometer. These results agree with those obtained by others with optical techniques (2). Inasmuch as the osmolality of the serosal bathing solution was not altered in these experiments, the "osmometric" behavior of the cells suggests either that the basolateral membrane has a relatively low water permeability or that a small compartment located under the basolateral cell membrane (predominantly lateral intercellular spaces) underwent rapid osmotic equilibration with the mucosal solution and the cell interior. Since optical measurements suggest that the osmotic water permeability coefficient of the basolateral membrane is higher than that of the apical membrane (2), the second hypothesis is more likely. In principle, the osmotic water flow into this compartment could be transcellular, intercellular (i.e., across junctional complexes), or both. The noise levels of the reference and the ion-sensitive Proc. NatL Acad ScL USA 82 (1985) microelectrodes allow for reliable measurements of changes in the differential voltage trace of 1 mv or less. Hence, the technique permits measurements of cell volume changes of 5% or less. Estimate of Apical Membrane Hydraulic Permeability Coefficient. A major objective of this kind of study is to estimate the hydraulic permeability coefficient (Lp) and the related osmotic water permeability coefficient (P., Lp RT/VW) of the apical membrane (where R, T, and V. are the gas constant, the absolute temperature, and the partial molar volume of water, respectively). In principle, provided that the reflection coefficient of the solute added to or removed from the apical surface is unity, Lp can be calculated from the ratio Jv/Ar, where Jv is the initial water flow [AVolceii/(areatime)] and AT is the change in osmotic pressure. Such analysis is not straightforward, however, for the following reasons: (i) The cell volume before the osmotic challenge must be known. (ii) The water flow causing the change in cell volume is assumed to occur only across the apical membrane; osmotic water flow might also occur across the junctions, changing the osmolality of the fluid in the lateral intercellular spaces; this would, in turn, alter cell volume by osmotic water flow across the lateral membranes. (iii) The change in osmolality at the cell surface is assumed to be instantaneous upon the solution substitution (an unlikely possibility) or its time course must be known; such a calculation is possible only when the effective thickness of the mucosal unstirred layer and the diffusion coefficient of the osmotic probe are known (see below). (iv) "Sweeping-away" effects [i.e., changes in solute concentrations on both sides of the membrane produced by the osmotic water flow (13, 14)] must either be assumed to be small or, if not, must be calculated. (v) The volume of the subepithelial space undergoing osmotic equilibration is taken to be small compared to the volume of the cells. Because of these difficulties in obtaining the correct value for Lp of the apical membrane, a definitive calculation of this parameter is not possible yet. However, assuming, as stated in (ii) above, that the changes in [Me4N+]i illustrated in Fig. 3 are caused entirely by water flows through the apical a -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Vi -Vcs, 1502 mv b -60J_ V( A L mv -8 V.9 LL I 1-17 mosmol/kg +32 mosmol/kg 10 sec FIG. 3. Cell volume changes produced by rapid alterations in the osmolality of the mucosal solution. The records in a and b were obtained during impalements of the same cells with a conventional microelectrode and a Me4N+-sensitive microelectrode. In each panel, the transepithelial voltage (Vms), the cell membrane voltage (Vcs = Vcell - Vserosa), and the difference between the voltages of the ion-sensitive and conventional microelectrode (Vi Vc5) - are depicted. Transepithelial current pulses were applied before the changes in osmolality for impalement validation (7). (a) Effect of reducing [NaCl] in the mucosal medium by 17 milliosmoles/kg, equivalent to an 8.5% reduction of the control osmolality. Note the fall in [Me4N'] i (negative deflection in the differential voltage trace), which indicates cell swelling. After about 20 sec, cell volume had increased to 109% of control. The initial rate of change of [Me4N+]i, calculated by a least-squares linear regression to the digitized voltages, was -0.7%/sec. (b) Effect of increasing [NaCl] in the mucosal medium by 32 milliosmoles/kg, equivalent to a 16% increase in osmolality. Twenty seconds after the onset of the increase in [Me4N+]i, cell volume had decreased to 85% of control. The initial rate of change of [Me4N']j was -1.8%/sec.

4 membrane, the minimal apical membrane Lp, calculated from the initial rate of change of [Me4N+]i and for a cell height of 35,m (1), neglecting apical membrane foldings, is 1-3 x 10-3 cm-sec-'(osmol/kg)-l (P,, = x 10-3 cm-sec-1). Since the average spontaneous fluid transport rate in tissues incubated in NaCl-Ringer's solution on both sides is about 3.8 x 10-6 cm sec-1 (15), an osmolality difference of 1-3 milliosmoles/kg across the apical membrane would suffice to account for salt-coupled water transport. This estimate agrees well with the one derived from optical measurements -1 Physiological Sciences: VK VM I / Time, sec Reuss I10 mv FIG. 4. Measurement of the effective thickness of the mucosal unstirred layer (8) in Necturus gallbladder. (Upper) The tissue was mounted horizontally and continuously superfused by gravity with NaCl-Ringer's solution, at a rate of =30 ml-min-. The solution formed an -1-ml segment of sphere on top of the tissue. Inflow and outflow were at opposite points. The voltage traces depict the outputs of an intracellular conventional microelectrode (Vmc) and an extracellular K+-sensitive microelectrode (VK), both referred to a calomel half-cell connected to the mucosal solution with a flowing, saturated-kcl bridge. The distance between the K+-sensitive microelectrode and the tissue surface (h) was estimated from h a-sin = a, where a is the distance traveled from tissue surface to final position (measured to the nearestgum with a manual micromanipulator) and a is the angle of the microelectrode axis with respect to the tissue surface (horizontal). In this experiment, h 19 = gm. The effective unstirred layer thickness was estimated from the time coursesof VK and Vmc upon a rapid replacement of NaCl with KCl in the perfusate. The voltage responses of the K+ microelectrode and the membrane were assumed to have the formav = S log{[k+]t/[k+]0}, where S is assumed to be a constant (in each case) throughout the transition, [K+], is the [K+] at the membrane surface at any time, and [K+]o is (i.e., the voltage value at which [K+] at the membrane has risen to half of the final change) was estimated. At AV1/2, t = ti12, and 8 can be calculated from the control [K+] (2.5 mm). From this equation, AV1,2 = t1/ /D (16). The results were 118,um and 136,um for the VK and Vmc traces, respectively. The difference (18 /.Lm) is in excellent agreement with the measured distance between the tip of the K+ microelectrode and the cell membrane (19 Am). (Lower) Calculated fractional changes in concentration at the cell membrane surface when either NaCl (0) or sucrose (0) is added instantaneously to the bulk solution. The values were computed from the solution to the diffusion equation from bulk solution into unstirred layer (17) for = 136,um and diffusion coefficients of cm2sec-1 and cm2 sec1 for NaCl and sucrose, respectively. Proc. NatL. Acad. Sci USA 82 (1985) 6017 (2), but both are based on the assumptions and simplifications listed above. Unstirred Layers. As eloquently stated by Diamond (13), a major source of uncertainty in water-transport studies in flat epithelia is the presence of unstirred layers, which not only make it impossible to achieve an instantaneous change in osmolality at the membrane surface but also permit the development of sweeping-away effects. The contribution of the latter effects is diminished if the water flow (or AVOlcell) can be estimated rapidly and with small osmotic gradients, as with the technique described. However, the osmotic pressure difference at the cell surface must be known during the transient; i.e., the effective thickness of the unstirred layer (8) must be determined. From 8, the concentration (at the membrane surface) of the osmotic probe "instantaneously" added to or removed from the bulk solution can be estimated if its diffusion coefficient is known. I have estimated 8 under experimental conditions similar to those shown in Fig. 3, by the method illustrated in Fig. 4. The changes in apical membrane voltage, measured with an intracellular microelectrode, and in the voltage output of an extracellular Ks-sensitive microelectrode, whose tip was positioned at a known distance from the cell surface, were recorded before, during, and after a rapid increase in mucosal [K+]. The K+-sensitive microelectrodes had rise times (10 to 90%) ofca. 200 msec. The value of 8 was calculated from the time course of the voltage changes and hence includes the contributions of delays caused by mixing in the bulk solution and by the true unstirred layer; from 8 and the respective diffusion coefficients, the time course of a change in NaCl or sucrose concentration at the cell surface was computed. It can be seen that at the time at which Jv is estimated-i.e., within the initial 10 sec-aosmolality is far short of the final value. Hence, if the osmotic water flow is only across the apical membrane, the values of Lp and Pos would be higher than those estimated above assuming a step change in osmolality. Simultaneous measurements of [Me4N']s, 8, and unstirred layer [NaCl] will be required for a full, quantitative analysis of this problem. In addition, new techniques will have to be developed to estimate transcellular and paracellular water fluxes. Conclusions. In summary, I have described and illustrated a specific application of a method to measure, with electrophysiologic techniques, changes in cell volume. The main advantages of this method are excellent time resolution, independence of optical properties of the preparation, relatively low cost, and relative technical ease. Its main shortcoming may prove to be the difficulty in making the measurements in small cells. Alone or in combination with other intracellular or extracellular measurements, this method may be a useful tool in a number of areas of cellular physiology. I thank Drs. E. Bello-Reuss, P. De Weer, and A. Roos for comments on the manuscript, Jozianne Bazile for technical assistance, and Sue Eads for secretarial help. This work was supported by National Institutes of Health Grants AM19580 and AM Spring, K. R. & Hope, A. (1979) J. Gen. Physiol. 73, Persson, B.-O. & Spring, K. R. (1982) J. Gen. Physiol. 79, Strange, K. & Spring, K. R. (1984) J. Gen. Physiol. 84, 22 (abstr.). 4. Reuss, L. & Finn, A. L. (1975) J. Membr. Biol. 25, Lewis, S. A., Eaton, D. C., Clausen, C. & Diamond, J. M. (1977) J. Gen. Physiol. 70, Fromter, E. (1972) J. Membr. Biol. 8, Weinman, S. A. & Reuss, L. (1982) J. Gen. Physiol. 80, Reuss, L., Reinach, P., Weinman, S. A. & Grady, T. P. (1983) Am. J. Physiol. 244, C336-C347.

5 6018 Physiological Sciences: Reuss 9. Neher, E. & Lux, H. D. (1973) J. Gen. Physiol. 61, Van Os, C. H. & Slegers, J. F. G. (1975) J. Membr. Biol. 24, Reuss, L. & Finn, A. L. (1975) J. Membr. Biol. 25, Zeuthen, T. (1982) J. Membr. Biol. 66, Diamond, J. M. (1979) J. Membr. Biol. 51, Proc. Natl. Acad.a SeL USA 82 (1985) 14. Barry, P. H. & Diamond, J. M. (1984) Physiol. Rev. 64, Reuss, L. (1984) J. Gen. Physiol. 84, Diamond, J. M. (1966) J. Physiol. (London) 138, Jacobs, M. H. (1967) Diffusion Processes (Springer, New York).