Using the pH-sensitive dye 2′,7′-bis(2-carboxyethyl)-5(6)-carboxyfluorescein (BCECF), we examined the effect of hyperosmolar solutions, which presumably caused cell shrinkage, on intracellular pH (pHi) regulation in mesangial cells (single cells or populations) cultured from the rat kidney. The calibration of BCECF is identical in shrunken and unshrunken mesangial cells if the extracellular K+concentration ([K+]) is adjusted to match the predicted intracellular [K+]. For pHi values between ∼6.7 and ∼7.4, the intrinsic buffering power in shrunken cells (600 mosmol/kgH2O) is threefold larger than in unshrunken cells (∼300 mosmol/kgH2O). In the nominal absence of CO2/ , exposing cell populations to a HEPES-buffered solution supplemented with ∼300 mM mannitol (600 mosmol/kgH2O) causes steady-state pHi to increase by ∼0.4. The pHi increase is due to activation of Na+/H+exchange because, in single cells, it is blocked in the absence of external Na+ or in the presence of 50 μM ethylisopropylamiloride (EIPA). Preincubating cells in a Cl−-free solution for at least 14 min inhibits the shrinkage-induced pHi increase by 80%. We calculated the pHi dependence of the Na+/H+exchange rate in cell populations under normosmolar and hyperosmolar conditions by summing 1) the pHi dependence of the total acid-extrusion rate and 2) the pHi dependence of the EIPA-insensitive acid-loading rate. Shrinkage alkali shifts the pHi dependence of Na+/H+exchange by ∼0.7 pH units.
- acid-base transport
- intracellular pH
- volume regulation
normal physiological processes challenge the maintenance of constant cell volume (see Ref. 28). For example, the breakdown of proteins, glycogen, and triglycerides into their more osmotically active monomers can lead to cell swelling. Also, the activation of ion channels in a cell can elicit volume changes as H2O osmotically follows the movement of ions. For example, the activation of Na+ channels typically leads to the influx of Na+ and cell swelling, whereas the activation of K+ or Cl− channels generally leads to the efflux of K+ or Cl− and cell shrinkage. Finally, as discussed by Hoffman and Simonsen (22), excessive oral intake of H2O can create a hypotonic environment for epithelial cells of the gastrointestinal tract, and changes in the levels of antidiuretic hormone lead to changes in the osmolality of the medullary interstitium of the kidney.
Cells have acquired strategies to combat perturbations in their volume. These strategies can be categorized into chronic responses (hours to days) and acute responses (seconds to minutes). One important component of the chronic response is typically the synthesis or degradation of membrane-impermeant solutes (see Refs. 8 and 28). Also, long-term volume stresses on a cell may stimulate or inhibit the expression and/or activation of ion/osmolyte transport processes (see Ref.20).
Acute responses of a cell to volume disturbances involve the immediate stimulation of ion transport across plasma membranes. For example, cell swelling can stimulate the efflux of ions and thus H2O out of cells, a process termed volume regulatory decrease (2, 9, 15, 25, 27, 39). Conversely, cell shrinkage stimulates the influx of ions and thus H2O in cells. The subsequent increase in cell volume, a volume regulatory increase, can be caused by stimulation of 1) Na-K-Cl cotransport and 2) stimulation of Na+/H+exchange, which together with Cl-HCO3 exchange has the net effect of accumulating NaCl (see Refs. 10, 18, 20, 22, 28).
Several groups have studied the shrinkage-induced activation of Na+/H+exchange. Of course, decreased intracellular pH (pHi) can also activate Na+/H+exchange independently of shrinkage. Both shrinkage and low pHi activate Na+/H+exchange in human lymphocytes (17), C6 rat glioma cells (24,37), dog erythrocytes (32), barnacle muscle fibers (14), rabbit alveolar macrophages (21), and rat mandibular salivary glands (36). In barnacle muscle fibers, a G protein appears to be an intermediate in the transduction of the shrinkage signal to the activation of the Na+/H+exchanger (13, 23). The presence of intracellular Cl− also appears necessary for the shrinkage-induced activation of the transporter in both dog erythrocytes and barnacle muscle fibers. In the dog erythrocytes, the activation can be inhibited by removing external Cl−, which presumably depletes intracellular Cl−(31, 33). Indeed, in the barnacle muscle fiber, shrinkage-induced activation of the Na+/H+exchanger requires intracellular Cl− (14, 23) at a step at, or before, activation of a heterotrimeric G protein (23).
In experiments on the pHidependence of the Na+/H+exchanger in barnacle muscle fibers (14), the shrinkage-induced activation of the transporter is approximately threefold greater at relatively acidic values (pHi ≈ 6.8) than at relatively alkaline values (pHi ≈ 7.6). In this barnacle study, the authors defined the Na+/H+exchange rate as the amiloride-sensitive acid-extrusion rate in experiments in which they measured pHi with glass microelectrodes. Also, several groups have presented data consistent with the idea that shrinkage activates the Na+/H+exchanger in mammalian cells by shifting the flux vs. pHi profile of the exchanger in the alkaline direction (19, 21, 24, 36, 37).
In the present study, we performed experiments specifically designed to isolate the pHi dependence of the Na+/H+exchange rate in cultured rat mesangial cells under both normosmolar and hyperosmolar conditions. We found that exposing cells to a hyperosmolar solution, which presumably causes cell shrinkage, alkali shifts the pHi dependence of the transporter by ∼0.7 pH units. This shift explains why cell shrinkage causes steady-state pHi to increase.
Portions of this work have been published in abstract form (35).
All experiments were performed in the nominal absence of CO2/ . The standard HEPES-buffered solution (∼305 mosmol/kgH2O) contained (in mM) 125 NaCl, 5 KCl, 1 CaCl2, 1.2 MgSO4, 2 NaH2PO4, 32 HEPES, and 10.5 glucose and was titrated to 7.4 at 37°C with NaOH. Hyperosmolar solutions (600 mosmol/kgH2O) were made by adding ∼300 mM mannitol to a HEPES-buffered solution, thereby increasing the volume by ∼3%. In the Na+-free solutions, the Na+ substitute wasN-methyl-d-glucammonium (NMDG+). In the Cl−-free solutions, the Cl− substitute was gluconate. In some of the Cl−-free solutions, the Ca2+ concentration was increased to 3 mM to compensate for Ca2+chelation by the gluconate. The inhibitory effect of removing external Cl− on the shrinkage-induced alkalization (see results) was the same whether or not we compensated for Ca2+ chelation by gluconate. In NH3/ solutions, NaCl (or NMDG-Cl) was replaced by an equal concentration of NH4Cl. Both the normosmolar and hyperosmolar dye-calibration solutions, which were based on the corresponding HEPES-buffered solutions, contained 10 μM nigericin. In addition, in the normosmolar calibration solution, we replaced Na+ with 100 mM K+ [final K+ concentration ([K+]) = 105 mM] and NMDG+. In the hyperosmolar calibration solution, we replaced the Na+ with 205 mM KCl (final [K+] = 210 mM) and NMDG+ and adjusted the osmolality to 600 mosmol/kgH2O by adding ∼120 mM mannitol. We measured the osmolality of all solutions with a vapor-pressure osmometer (model 5100C; Wescor, Logan, UT). Ethylisopropylamiloride (EIPA) was diluted 1:1,000 from a 50 mM stock dissolved in dimethyl sulfoxide. The acetoxymethyl ester of 2′,7′-bis(2-carboxyethyl)-5(6)-carboxyfluorescein (BCECF-AM) was obtained from Molecular Probes (Eugene, OR). Ethanol was obtained from Pharmco Products (Brookfield, CT). EIPA was obtained from either E. J. Cragoe, Jr. (Nacogdoches, TX) or Research Biochemicals (Natick, MA). All other reagents were obtained from Sigma Chemical (St. Louis, MO).
Rat mesangial cells were prepared as previously described (6, 29). Briefly, cells from passages 3–7 were plated on glass coverslips and were grown at 37°C in Dulbecco’s modified Eagle’s medium (GIBCO-BRL, Life Technologies, Gaithersburg, MD) exposed to an atmosphere of 5% CO2-balance air. The medium was supplemented with 100 U/ml penicillin plus streptomycin (GIBCO-BRL), 5 μg/ml of insulin plus transferrin plus selenious acid (Becton-Dickinson Labware, Bedford, MA), and 10% fetal calf serum (Gemini Bioproducts, Calabasas, CA or GIBCO-BRL). To render cells quiescent at the time of experimentation, the cells were incubated for 16–18 h in a medium containing only 0.5% fetal calf serum. Experiments were performed on cells at least 80% confluent. A coverslip with attached cells was exposed to 10 μM BCECF-AM in the standard HEPES-buffered solution in an air incubator at 37°C for 20 min. The coverslip was then transferred to a temperature-regulated, flow-through cuvette and was perfused with the standard HEPES-buffered solution, prewarmed to 37°C, for ∼5 min before the start of the experiment to wash nonhydrolyzed BCECF-AM from the cells.
Measurement of pHi
Our technique for measuring pHi in a population of cells has been previously described (26), and it will only be briefly summarized here. The flow-through cuvette containing a coverslip was placed in the excitation light path of a SPEX Fluorolog-2 spectrofluorometer (model CM1T10E; SPEX Industries, Edison, NJ). Cells loaded with BCECF were alternately excited with light at the pH-sensitive wavelength of 502 nm (I502) and the relatively pH-insensitive wavelength of 440 nm (I440). The emission wavelength was 526 nm. The sample integration time was 1 s, and the sample frequency was once every 3 s. The fluorescence-excitation ratio (I502/I440) was normalized and converted to a pHi value using the high-K+/nigericin technique (38), as modified for a single-point calibration (6). Background I502 and I440 signals from cells without dye were subtracted from total I502 and I440 signals. An evaluation of shrinkage on the calibration of BCECF in mesangial cells is presented in results.
After each experiment in which a dye calibration was performed, the perfusion lines were washed with either ethanol followed by distilled H2O or albumin followed by ethanol and then H2O. These washing protocols substantially reduce nigericin contamination from one experiment to the next (Bevensee, Bashi, and Boron, unpublished observations). Moreover, because we report the Na+/H+exchange rate as the difference between the total and EIPA-insensitive acid-extrusion rates, our Na+/H+exchanger data should be immune to any residual nigericin contamination (Bevensee, Bashi, and Boron, unpublished observations).
For some experiments, we measured pHi in single mesangial cells using an inverted Zeiss IM-35 microscope equipped for epi-illumination. Our technique for measuring pHi in single cells is described in detail in Ref. 6.
Calculation of Acid-Extrusion and Acid-Loading Rates
Acid-extrusion and acid-loading rates are defined as the product of the rate of change in pHi(dpHi/dt) and the total intracellular buffering power (βTotal). Because it is difficult to measure the ratio of volume to surface area in many cells, it has become customary to report acid-extrusion and acid-loading rates in units of moles per liter of cell volume per second (i.e., μM/s), instead of the classical flux units of moles per unit area per unit time (i.e., μmol ⋅ cm−2 ⋅ s−1). We use the symbol ϕ to refer to these acid-extrusion and acid-loading rates, or “pseudofluxes.”
Calculation of Intrinsic Buffering Power
In the nominal absence of CO2/ , βTotal is equivalent to the intrinsic buffering power of the cell (βI). Cell shrinkage will concentrate intracellular H+buffers and likely increase βI. Therefore, we used the approach introduced by Boyarsky et al. (6) to calculate the pHi dependence of βI in mesangial cells exposed to both norm- and hyperosmolar conditions. βI experiments were performed in the absence of external Na+ to eliminate pHi changes caused by Na+-dependent processes (e.g., Na+/H+exchange). After initially exposing the cells to a Na+-free solution containing 40 mM NH3/ , we recorded step changes in pHielicited by progressively lowering extracellular [NH3/ ]. βI is defined as the change in intracellular [ ] per unit change in pHi.
In the experiment shown in Fig.1 A, mesangial cells in the nominally CO2/ -free, standard HEPES-buffered solution had a pHi of ∼7.1 beforepoint a. Exposing the cells to the HEPES-buffered solution supplemented with ∼300 mM mannitol (total osmolality = 600 mosmol/kgH2O) elicited an initial decrease in pHi (segment ab), followed by a larger increase to ∼7.5 (segment bc). These osmolality-induced changes in pHiare described in more detail in results. For the remainder of the experiment shown in Fig. 1 A(segment ck), the cells were exposed to hyperosmolar solutions. Removing external Na+ caused pHi to decrease to ∼6.7 (segment cd). In the continued absence of external Na+, switching the cells to a solution containing 40 mM NH3/ elicited a rapid increase in pHi(segment de). Subsequently, exposing the cells to solutions containing progressively lower concentrations of NH3/ caused stepwise decreases in pHi(segments ef,fg,gh,hi,ij, andjk). In some experiments (not shown), the pHi continued to decrease slowly after the initial step changes. For these experiments, we back-extrapolated the pHi vs. time record to determine more accurately the initial pHi changes elicited by the NH3/ solutions (1, 3, 12).
Using data from experiments similar to that shown in Fig.1 A, we plotted the pHi dependence of βI for mesangial cells under hyperosmolar conditions (Fig. 1 B). In other experiments, we used the same NH3/ step technique to determine the pHi dependence of βI for mesangial cells under normosmolar conditions (Fig. 1 B). As shown by the best-fit lines to the data, βI in shrunken cells is threefold greater than in unshrunken cells for pHi values between ∼6.7 and ∼7.4.
Data are reported as means ± SE. Levels of significance were assessed using the unpaired Student’st-test. AP value < 0.05 was considered significant. The pHi dependencies of βI were fitted by straight lines using a least-squares method. The dpHi/dtvalues were fitted by a third-order polynomial using a least-squares method.
Calibration of BCECF in Mesangial Cells Under Normosmolar and Hyperosmolar Conditions
For maintaining a stable pHi in nigericin calibration solutions, doubling the osmolality requires doubling the extracellular [K+].
As described in methods, we calibrated the pH-sensitive dye BCECF using the high-K+/nigericin technique. In our first series of experiments, we evaluated the effect of shrinkage on the calibration of the dye. If the cells were perfect osmometers, then increasing the extracellular osmolality from ∼300 to 600 mosmol/kgH2O should decrease cell volume by one-half and thereby double the concentration of intracellular constituents (e.g., K+), assuming no change in membrane permeability. Accordingly, the dye calibration solutions used for such shrunken cells should have two times the nominal [K+] (210 vs. 105 mM) to satisfy the calibration requirement that extracellular K+ concentration ([K+]o) equals the intracellular K+concentration ([K+]i).
To test this hypothesis, we first calibrated BCECF in mesangial cells under normosmolar conditions using our standard 105 mM K+/nigericin calibration solution and then monitored pHi as we shrunk the cells in hyperosmolar (600 mosmol/kgH2O) calibration solutions containing either 105 or 210 mM K+. Shrinking cells in the presence of only 105 mM K+ would be expected to lead to a decrease in pHi because [K+]o(i.e., 105 mM) will be less than the predicted [K+]i(i.e., 210 mM), thereby promoting the nigericin-mediated exchange of internal K+ for external H+. On the other hand, shrinking the cells in a calibration solution containing 210 mM K+ should have no effect on nigericin-mediated pHi changes because [K+]owill match [K+]i.
During the initial portions of the recordings shown in Fig.2 (prior to point a), both groups of mesangial cells were exposed to our standard 105 mM K+/nigericin solution (pH 7.0). Thus, by definition, the pHi before point a is 7.0. Shrinking the cells in the standard calibration solution supplemented with ∼300 mM mannitol elicited a pHi decrease of ∼0.25 (segment ab). This acidification was likely caused by a nigericin-mediated exchange of internal K+ for external H+, due to the shrinkage-induced increase of [K+]i. As indicated by the broken line, the decrease in pHi was relatively constant for >40 min. Returning the cells to the normosmolar calibration solution caused pHi to increase (segment bc) to near the initial value at the onset of the experiment (prior to point a). In 10 experiments similar to that shown in Fig.2, shrinking mesangial cells with a 105 mM K+/∼300 mM mannitol calibration solution (pH 7.0) caused pHi to decrease from 7.0 to 6.76 ± 0.01 (P < 0.001). Clearly then, BCECF in shrunken cells cannot be adequately calibrated using the standard calibration solution containing 105 mM K+.
In contrast to what we observed with the 105 mM K+/∼300 mM mannitol calibration solution (600 mosmol/kgH2O), shrinking cells in a calibration solution containing 210 mM K+/∼120 mM mannitol (600 mosmol/kgH2O) elicited only a slight decrease in pHi, ∼0.04 (segment ab′). This small acidification was reversed by returning the cells to the normosmolar calibration solution (segment b′c′). In 16 experiments similar to that shown in Fig. 2, shrinking mesangial cells with a 210 mM K+/∼120 mM mannitol calibration solution caused pHi to decrease from 7.0 to 6.98 ± 0.01 (P = 0.03). Thus, at a pHi of 7.0, the calibration of BCECF in mesangial cells shrunken in a 600 mosmol/kgH2O solution containing 210 mM K+/nigericin solution is equivalent to the calibration of the dye in unshrunken cells incubated in an ∼300 mosmol/kgH2O solution containing 105 mM K+/nigericin. In similar experiments on BCECF-loaded rat thymic lymphocytes, Grinstein et al. (19) demonstrated that the fluorescence ratio of the dye remained constant when the cells were exposed to nigericin-containing solutions (pH 7.0) of different osmolarities but a constant [K+]-to-[Na+] ratio.
BCECF calibration is identical in shrunken and unshrunken mesangial cells, provided [K+]ois adjusted to match the predicted [K+]i.
Using an approach similar to that shown in Fig. 2, we extended our analysis of the effect of shrinkage and [K+]oon the intracellular calibration of BCECF for pHi values between ∼6 and ∼8.2. In Fig. 3, we plot the normalized fluorescence excitation ratio for the BCECF vs. pHi relationship for cells calibrated either under normosmolar conditions (∼300 mosmol/kgH2O) at a [K+]oof 105 mM or under hyperosmolar conditions (600 mosmol/kgH2O) with an external [K+] of 210 mM. The best-fit curves are the result of a nonlinear least-squares curve fit (see legend for Fig. 3). The two curves are virtually identical to one another. In other words, a single BCECF calibration can be used in experiments in which the mesangial cells undergo volume changes.
Effect of Shrinkage on Steady-State pHi
Shrinkage elicits an increase in steady-state pHi.
We evaluated the effect of shrinkage on the steady-state pHi of mesangial cells by exposing them to our standard HEPES-buffered solution supplemented with ∼300 mM mannitol (600 mosmol/kgH2O). Before point a in Fig.4, the cells had an average pHi of ∼7.05. Exposing the cells to a hyperosmolar solution containing ∼300 mM mannitol consistently elicited a small, transient decrease in pHi (segment ab), presumably due to the effect of concentrating and then reequilibrating H+ and intracellular buffers. This transient pHi decrease was followed by a robust increase in pHi(segment bc) to a value ∼0.7 pH units higher than the initial pHi(prior to point a). The hyperosmolar-induced pHi increase was at least partially reversed by returning to the normosmolar solution (segment cd). In a total of 54 experiments similar to that shown in Fig. 4, ∼300 mM mannitol caused an increase in pHi from 7.17 ± 0.02 to 7.56 ± 0.03 (P< 0.0001). The average pHiincrease of ∼0.4 was not dependent on the initial steady-state pHi before exposing the cells to the hyperosmolar solution (data not shown). Mesangial cells with an initial steady-state pHi of ∼7.0 were just as likely to alkalinize to the same extent as cells with an initial steady-state pHi of ∼7.4.
Shrinkage-induced increase in pHi is blocked by removing external Na+ or by applying 50 μM EIPA.
Using an inverted microscope equipped for epi-illumination, we also performed experiments on single mesangial cells. Exposing single cells (not shown) to a HEPES-buffered solution supplemented with ∼300 mM mannitol (600 mosmol/kgH2O) elicited a series of pHi changes very similar to those in Fig. 4: a transient decrease in pHi, followed by a rapid and reversible increase to a value ∼0.6 pH units greater, on average, than the pHi at the onset of the experiments. In the experiment shown in Fig.5 A on a single mesangial cell, removing external Na+ caused pHi to decrease from ∼7.2 to ∼6.7 (segment ab), probably due to reversal of Na+/H+exchange and/or unmasking of background acid loading. In the continued absence of external Na+, exposing the cell to the HEPES-buffered solution supplemented with ∼300 mM mannitol did not elicit a pHi increase (segment cd). In other experiments similar to that shown in Fig. 5 B, exposing single cells to 50 μM EIPA caused pHi to decrease from ∼7.15 to ∼6.75 (segment ab), due to unmasking of background acid loading that is usually balanced by Na+/H+exchange at steady-state pHi. In the continued presence of EIPA, exposing the cell to a solution supplemented with ∼300 mM mannitol did not elicit a pHi increase (segment bc). In summary, the pHi increase elicited by hyperosmolality can be inhibited by either removing external Na+ or by applying 50 μM EIPA. Therefore, shrinkage appears to increase the steady-state pHi of mesangial cells by activating the Na+/H+exchanger.
Shrinkage-induced increase in pHi is reduced by incubating the cells for 14 or 90 min in a Cl−-free medium.
As noted in the introduction, Cl− is required for the shrinkage-induced activation of Na+/H+exchange in dog erythrocytes and barnacle muscle fibers. We therefore tested if the shrinkage-induced pHi increase in mesangial cells could be reduced by exposing the cells to a Cl−-free solution. In Fig.6, we show the results from two experiments in which cells were exposed to the HEPES-buffered solution supplemented with ∼300 mM mannitol either in the absence or presence of external Cl−. In both experiments, the mannitol solution elicited a small decrease in pHi (segment ab), similar to segment ab in Fig. 4. However, the subsequent pHi increase was 0.36 pH units greater in the control cells (segment bc) than in those exposed to the Cl−-free medium for ∼90 min before the experiment (segment bc′). In five experiments, we found that the mean pHi increase elicited by hyperosmolality was 0.09 ± 0.05 in the absence of Cl−. In seven, week-matched control experiments conducted in the presence of Cl−, the pHi increase averaged 0.45 ± 0.07. Thus Cl− removal reduced the hyperosmolar-induced pHi increase by 80%.
Effect of Shrinkage on pHi Dependence of Na+/H+Exchange
Shrinkage increases the rate of pHirecovery from an acid load.
To determine the effect of shrinkage on the pHi dependence of Na+/H+exchange in mesangial cells, we used the approach introduced by Boyarsky et al. (5). For cells subjected to either normosmolar or hyperosmolar conditions, we determined1) the pHi dependence of “total acid extrusion” (ϕE) during the pHi recovery from an acid load and2) the pHi dependence of “EIPA-insensitive acid loading” (ϕEIPA). We obtained the pHi dependence of the Na+/H+exchange rate by adding ϕEIPA to ϕE for identical pHi values.
The results from two experiments, performed on unshrunken and shrunken mesangial cells, are shown in Fig. 7. In both cases, we used the -prepulse technique to acid load the cells (4) and then monitored the pHi recovery from the acid load. For the hyperosmolar experiment, cells in the standard HEPES-buffered solution had an average pHi of ∼7.3. Shrinking the cells in the solution supplemented with ∼300 mM mannitol elicited the usual transient decrease and rapid increase in pHi (segment abc). Using the -prepulse technique (4), we then acid loaded the cells (segment cdef). After the acid load, pHi recovered rapidly (segment fg). After the pHi recovery was complete, applying 50 μM EIPA elicited a decrease in pHi (segment gh) that represents an unmasking of background acid loading that is normally balanced by acid extrusion mediated by Na+/H+exchange.
For the normosmolar experiment, we only show the record for the recovery of pHi from the acid load (segment f′g′) and the decrease in pHi from the subsequent application of EIPA (segment g′h′).
It is important to compare the data from these two experiments at comparable pHi values, because acid-base transport rates can vary markedly with changes in pHi. In comparing thesegment fg andsegment f ′g′ pHi recoveries in shrunken and unshrunken cells, respectively, we see that there is only a narrow pHi range (near ∼7.2) in which the two pHi recoveries overlap. At this single pHi of 7.2, the pHi recovery rate is substantially greater in shrunken cells (point f) than in unshrunken cells (point g′).
Shrinkage alkali shifts the pHidependence of the Na+/H+exchange rate by ∼0.7 pH units.
From segment fg and segment f′g′ pHi trajectories similar to those shown in Fig. 7, we computed the pHi dependence of total acid extrusion (ϕE) for both shrunken and unshrunken cells (Fig.8 A). We also used the EIPA-unmasked acidifications similar to those insegments gh andg′h′ in Fig. 7 to obtain the pHidependence of EIPA-insensitive acid loading (ϕEIPA) for both shrunken and unshrunken cells (Fig. 8 B). Some of the normosmolar ϕEIPA data points at the lower pHi values in Fig. 8 B were obtained by monitoring the recovery of pHi from an acid load (analogous to segments fg andf′g′ in Fig. 7) in the presence of EIPA (data not shown). In principle, we would have obtained normosmolar ϕEIPA data for pHi values between 6.9 and 6.6 if either EIPA had caused pHi to fall below pH 6.9 or if pHi had recovered from an acid load in the presence of EIPA above pH 6.6. In Fig. 8 C, we plot the pHi dependence of the Na+/H+exchange rate (ϕNa-H) for both shrunken cells and unshrunken cells. We obtained the ϕNa-H plots in Fig.8 C by adding the respective ϕE and ϕEIPA data sets for norm- and hyperosmolar conditions. For both shrunken and unshrunken cells, ϕNa-H decreases linearly with increasing values of pHi. However, the plot of ϕNa-H vs. pHi for the shrunken cells is alkaline shifted by ∼0.7 pH units compared with the same plot for the unshrunken cells.
Na+/H+Exchange in Shrunken Cells
Shrinkage activates the Na+/H+exchanger in mesangial cells.
In the present study, we demonstrate that shrinking mesangial cells in a hyperosmolar solution containing ∼300 mM mannitol (600 mosmol/kgH2O) elicits a sustained increase in steady-state pHi of ∼0.4 pH units. The increase is caused by stimulation of Na+/H+exchange because it can be inhibited by removing external Na+ or by applying 50 μM EIPA. The ∼0.4 pH unit increase is substantially larger than the maximal ∼0.15 pH unit increase observed when mesangial cells are stimulated by the growth factor arginine vasopressin in the nominal absence of CO2/ (16). Therefore, shrinkage is a potent stimulator of Na+/H+exchange in mesangial cells. In other experiments in which we monitored the pHi recovery from acid loads in the presence or absence of EIPA, the pHi dependence of ϕNa-H is alkali shifted by ∼0.7 pH units in shrunken vs. unshrunken mesangial cells. As discussed in the , there is no reason to expect that this ∼0.7 pH unit alkali shift should be of the same magnitude as the shrinkage-induced increase in steady-state pHi.
Results from previous studies are consistent with shrinkage alkali shifting the pHi dependence of ϕNa-H.
As mentioned in the Introduction, several groups have provided evidence consistent with the idea that cell shrinkage alkali shifts the pHi dependence of ϕNa-H in mammalian cells. For example, in acid-loaded human lymphocytes, shrinkage alkali shifted the Na+-dependent acid extrusion vs. pHi relationship by 0.2–0.3 pH units (19). However, the Na+-dependent acid-extrusion rate includes the contributions of all Na+-dependent acid-base transporters, including the nigericin used to acid load the cells (Bevensee, Bashi, and Boron, unpublished observations and Ref. 34). In Na+-depleted C6 rat glioma cells, shrinkage alkali shifted the22Na+flux vs. pHi relationship by 0.3–0.4 pH units (24). However, the total, unidirectional22Na+influx includes the fluxes of all Na+ transport processes and cannot distinguish Na+/Na+exchange from Na+/H+exchange. In another study on acid-loaded rabbit alveolar macrophages, shrinkage alkali shifted the bafilomycin-resistant acid extrusion vs. pHi relationship by ∼0.2 pH units (21). However, the bafilomycin-resistant acid-extrusion rate represents the net effect of all acid-loading and acid-extruding processes other than the vacuolar H+ pump. Finally, in acid-loaded rat mandibular salivary glands, shrinkage alkali shifted the total acid-extrusion rate vs. pHi relationship by 0.15 pH units (36). However, the total acid-extrusion rate represents the algebraically summed contributions of all acid-loading and acid-extruding mechanisms. Thus, although the above four studies show that shrinkage causes an alkali shift in the pHi dependence of some combination of acid-base parameters, the studies were not specifically designed to isolate the pHi dependence of ϕNa-H from the pHi dependence of other acid-extrusion and acid-loading mechanisms.
Evidence has been presented that cell shrinkage alkali shifts the pHi dependence of ϕNa-H in rat C6 glioma cells (37). However, the activity of the Na+/H+exchanger at alkaline pHi values was underestimated because there was no compensation for background acid loading unmasked by amiloride at the high pHi values reached during cell shrinkage.
Issues to Consider When Measuring Changes in Acid-Extrusion and Acid-Loading Rates Elicited by Changes in Cell Volume
Shrinking mesangial cells does not alter the BCECF calibration curve. Because shrinkage will increase [K+]i, we calibrated the intracellular dye by exposing the cells to an extracellular solution with a [K+]ochosen to match the predicted, shrinkage-increased [K+]i. We found that simultaneously doubling both osmolality (from ∼300 to 600 mosmol/kgH2O) and [K+]o(from 105 to 210 mM) had little effect on the pHi of cells already clamped with nigericin to 7.0 (segment ab′ in Fig. 2). Thus the doubling of osmolality must have doubled [K+]i, so that we can conclude that the cells behave as near-perfect osmometers.
At least under the conditions of our calibration experiments (i.e., neither Na+ nor CO2/ were present), shrunken mesangial cells did not exhibit volume recoveries for at least 40 min after cell shrinkage. The evidence is that pHi failed to increase1during segment ab in Fig. 2. On the other hand, cell volume may have recovered during our other experiments (e.g., conducted in the presence of Na+). If so, then our postexperiment pHi calibrations would have been erroneously high because [K+]iwould have been lower than we thought. However, the size of this error is expected to be very small because the likely degree of volume recovery is very low. For example, when we exposed cells to ∼300 mM mannitol in Fig. 4, Na+/H+exchange caused pHi to increase by ∼0.7. This alkalinization represents the extrusion of ∼14 mM H+ (and thus the uptake of 14 mM Na+), assuming a buffering power of 20 mM (Fig. 1). The complete recovery of cell volume would have required the uptake of ∼300 mM Na+. Thus the computed uptake of 14 mM represents less than a 5% volume recovery. A 5% volume recovery would lead to a 5% error in the estimated [H+]i, which translates to an overestimation of 0.02 pH units. Thus it is unlikely that volume recovery seriously affected either our calibration or its application to physiological experiments.
In extending our analysis to other pHi values, we deduced that the calibration of intracellular BCECF is the same in unshrunken and shrunken cells (600 mosmol/kgH2O) provided that the appropriate [K+]ois used. Therefore, at least in mesangial cells, the calibration of intracellular BCECF is insensitive to cell-volume changes per se. Apparently, the pH sensitivity of intracellular BCECF in mesangial cells is unaffected by shrinkage-induced increases in ionic strength and concentrations of intracellular constituents (excluding K+).
Shrinking mesangial cells increases βI.
We had anticipated that a doubling of osmolality, and a subsequent halving of cell volume, would have doubled the intracellular concentrations of proton buffers and thus have increased βI by twofold. However, over the pHi range of 6.7–7.4, the apparent βI was threefold greater in shrunken (600 mosmol/kgH2O) vs. unshrunken cells. One possible explanation for the larger-than-expected increase in βI is cytosolic crowding of proton buffers. Macromolecular crowding can cause changes in protein-protein interactions that are disproportionately larger than the increase in protein concentration (30). If macromolecular crowding were to cause the pK values of buffers to shift closer to the physiological pH, the effect would be an increase in buffering power. Regardless of the mechanism for the increase in βI, given identical pHi recovery rates, the calculated acid-extrusion rate will be threefold greater in shrunken vs. unshrunken mesangial cells. Our βI results are at least qualitatively consistent with those previously reported in barnacle muscle fibers (14) and rat C6glioma cells (37).
Shrinkage can lead to overestimates in calculating acid-extrusion and acid-loading rates. One can report acid-base transport rates either in units of moles per surface area per time (J) or in units of moles per volume per time (ϕ). Shrunken cells have less volume but presumably the same plasma membrane surface area as unshrunken cells. Therefore, shrinkage will have different effects on acid-base transport rates reported in terms of J vs. ϕ.
As discussed in methods, it has become customary to report acid-base fluxes in terms of ϕ for small mammalian cells in which it is difficult to compute the volume-to-surface ratio. In these pseudoflux units, the acid-base transport rate would be higher for the smaller of two cells, even though theJ values for the two cells would be identical. Ideally, one would correct for cell volume changes when reporting acid-base fluxes in terms of ϕ. However, it is extremely difficult to measure volume changes accurately in most small cells. In at least one example though, Seo et al. (36) used proton nuclear magnetic resonance spectroscopy to measure cell-volume changes and then corrected acid-extrusion rates in perfused rat mandibular salivary glands under hyperosmolar conditions.
According to the above discussion, acid-extrusion and acid-loading rates reported in terms of ϕ, as in the present study, should be greater in shrunken (600 mosmol/kgH2O) vs. unshrunken (300 mosmol/kgH2O) mesangial cells, even if shrinkage had no effect on acid-base transport per se (i.e., even if shrinkage had no effect onJ). Thus the ϕ-J effect would predict that mesangial cells shrunken to one-half their initial volume would display a doubling of their acid-extrusion and acid-loading rates when expressed in terms of ϕ but no change when expressed in terms ofJ. This doubling arises because of the effect of shrinkage on surface-to-volume ratio2; any effects on βI are automatically incorporated in the calculated ϕ. However, we found that shrinkage produced a 5- to 12-fold increase in ϕNa-H, 2.5- to 6-fold greater than the doubling expected because of the ϕ-J effect. Thus we conclude that shrinkage increases the Na+/H+exchange rate by a factor of at least 2.5–6.
Two other arguments support our conclusion that shrinkage does indeed stimulate Na+/H+exchange, independent of a ϕ-Jeffect. First, shrinkage leads to an increase in steady-state pHi that is blocked by removing Na+ or applying EIPA. Second, shrinkage produces a substantial alkali shift of the ϕNa-H vs. pHi relationship, without substantially changing its slope. A ϕ-J effect could affect the magnitude of the transport parameter at any given pHi but could not produce such a shift.
In conclusion, shrinkage of mesangial cells leads to a substantial elevation in steady-state pHi by stimulating the Na+/H+exchanger. The pHi increase requires external Na+ and can be inhibited by EIPA. The kinetic basis of the Na+/H+exchange stimulation is a large alkali shift in the ϕNa-H vs. pHi relationship.
We thank Dr. Bruce A. Davis, Emilia M. Hogan, and Dr. Bernhard M. Schmitt for evaluating the manuscript and providing useful comments.
Relationship Between a Shift in the pHiDependence of an Acid-Base Transporter and the Resultant Change in Steady-State pHi
pHi is controlled by the sum of fluxes or other processes that extrude acid from the cell (J E) and those that load acid into the cell (J L). As shown in Fig. 9,J E tends to decrease with increasing pHi, whereas J L tends to increase. Imagine that a mesangial cell has only one acid loader (e.g., H+ influx) and one acid extruder (e.g., Na+/H+exchange). J E1describes the kinetics of the Na+/H+exchanger under control conditions, andJ L1 describes the kinetics of acid loading. The steady-state pHi of the cell is 7.1, determined by the point whereJ E1 =J L1( point a). We now switch to a hyperosmolar solution, which alkali shifts the pHi dependence of the Na+/H+exchanger by ∼0.7 pH units (J E1 →J E2). Will this 0.7 pH unit alkali shift ofJ E elicit a 0.7 pH unit increase in steady-state pHi? In the unlikely event that the plot of J Lvs. pHi is horizontal (J L2), thenJ E1 →J E2 will indeed increase steady-state pHi by 0.7 pH units, from 7.1 to 7.8 (point a→ b). However, in the equally unlikely event that the plot ofJ L vs. pHi is vertical (J L3), thenJ E1 →J E2 will not increase steady-state pHi at all (point a →c). In all likelihood, the plot ofJ L vs. pHi will slope upward, as indicated by J L1in Fig. 9. Therefore, the shrinkage stimulus that elicits a 0.7 pH unit alkali shift ofJ E will result in a smaller shift of the steady-state pHi (point a → d). Thus the steeper the slope ofJ L, the smaller the change in steady-state pHi. Similarly, the steeper the slope ofJ E, the smaller the effect on steady-state pHi of shifting the pHi profile of an acid loader.
If the cell has multiple acid extruders, then, all things being equal, an alkali shift of the Na+/H+exchanger will elicit an even smaller alkali shift in steady-state pHi, inasmuch as the Na+/H+exchanger makes a smaller contribution to the overall acid-extrusion rate. One should generally expect the shift in steady-state pHi to be smaller than the instigating shift in the pHiprofile of either an acid loader or acid extruder.
Address for reprint requests: M. O. Bevensee, Cellular and Molecular Physiology, Yale Univ. School of Medicine, 333 Cedar St., Rm. B-127 SHM, New Haven, CT 06520 (E-mail:).
This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Program Project Grant PO1DK-17433.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.
↵1 If the volume had increased, [K+]ishould have fallen, and nigericin should have mediated an influx of K+ and an efflux of H+ (i.e., produced a pHi increase).
↵2 Imagine that we shrink a cell to one-half its initial volume and that this shrinkage causes both βI and surface-to-volume ratio (S/V) to double, but has no effect on Na+/H+exchange per se. If we now acid load the cell, the rate of pHi recovery from this acid load will be the same as that under control conditions because the doubling of βI will tend to cut dpHi/dtin half, whereas the doubling of S/V will tend to double dpHi/dt. If we report the Na+/H+ exchange rate in terms of ϕ = dpHi/dt× βI, we will see that ϕ will double because dpHi/dtis unchanged and βI is doubled. On the other hand, if we report the Na+/H+exchange rate in terms of J = dpHi/dt× βI/(S/V), we will seeJ will be unchanged because the doubling of βI is canceled by the doubling of S/V.
- Copyright © 1999 the American Physiological Society