Volume regulation in cortical collecting duct cells: role of AQP2

Biol. Cell (2005) 97, 687–697 (Printed in Great Britain) doi:10.1042/BC20040116 Research article

Volume regulation in corticalcollecting duct cells: role of AQP2Paula Ford*, Valeria Rivarola*, Osvaldo Chara*, Marcel Blot-Chabaud†, Francoise Cluzeaud†, Nicolette Farman†,Mario Parisi* and Claudia Capurro*1

*Laboratorio de Biomembranas, Departamento de Fisiologıa y Biofısica, Facultad de Medicina, Universidad de Buenos Aires,

Paraguay 2155, piso 7, 1121 Buenos Aires, Argentina, and †INSERM U478, Faculte de Medecine Xavier Bichat, Paris, France

Background information. The renal CCD (cortical collecting duct) plays a role in final volume and concentration ofurine by a process that is regulated by the antidiuretic hormone, [arginine]vasopressin. This hormone induces anincrease in water permeability due to the translocation of AQP2 (aquaporin 2) from the intracellular vesicles to theapical membrane of principal cells. During the transition from antidiuresis to diuresis, CCD cells are exposed tochanges in environmental osmolality, and cell-volume regulation may be especially important for the maintenanceof intracellular homoeostasis. Despite its importance, cell-volume regulation in CCD cells has not been widelyinvestigated. Moreover, no studies have been carried out till date to evaluate the putative role of AQPs during thisprocess in renal cells.

Results. In the present study, we have studied the regulatory cell-volume responses to hypo-osmotic or hyper-osmotic challenges in two CCD cell lines: one not expressing AQPs and the other stably transfected with AQP2.We have used a fluorescent probe technique in which the acquisition of single-cell kinetic data can be simul-taneously recorded with the intracellular pH. Experiments with hyperosmotic mannitol media demonstrated that,independent of AQP2 expression, CCD cells shrink but fail to show regulatory volume increase, at least under thestudied conditions. In contrast, under hypo-osmotic shocks, regulatory volume decrease occurs and the activationof these mechanisms is more rapid in AQP2 transfected cells. This regulatory response takes place in parallelwith intracellular acidification, which is faster in cells expressing AQP2. The acidification and the initial regulatoryvolume decrease response were inhibited by glibenclamide and BaCl2 only in AQP2 cells.

Conclusions. These results suggest that increases in the osmotic water permeability due to the expression ofAQP2 are critical for a rapid activation of regulatory volume decrease mechanisms, which would be linked to cysticfibrosis transmembrane conductance regulator and to barium-sensitive potassium channels.

IntroductionThe mammalian collecting duct plays a central rolein the final volume and concentration of urine. Two-thirds of the hypo-osmotic fluid entering the collect-ing duct is reabsorbed into the CCD (cortical col-lecting duct) due to osmotic equilibration with the

1To whom correspondence should be addressed ([email protected]).Key words: aquaporin, intracellular pH, regulatory volume decrease,osmolality, renal cell.Abbreviations used: AM, acetoxymethylester; AQP2-RCCD1, RCCD1 cellsstably transfected with AQP2; AQP, aquaporin; AVP, [arginine]vasopressin;BCECF, 2′,7′-bis-(2-carboxyethyl)-5(6)-carboxyfluorescein; CCD, corticalcollecting duct; CFTR, cystic fibrosis transmembrane conductance regulator;RVD, regulatory volume decrease; RVD30, cell-volume recovery at 30 min;RVI, regulatory volume increase; WT, wild-type.

nearby iso-osmotic interstitium. In this segment ofthe nephron, peritubular osmolality varies little, un-der normal physiological conditions, but significantchanges in luminal osmolality take place continu-ously. In the antidiuretic state, the presence of AVP([arginine]vasopressin) causes CCD cells to becomewater-permeable as a result of the apical insertionof AQP2 (aquaporin 2). Water is then passively re-absorbed into the interstitium, leading to increasesin luminal osmolality (Deen et al., 1994; Nielsenet al., 2002). During the diuretic state caused byinhibition of AVP release, AQP2 resides in the intra-cellular vesicles of principal cells of the CCD. Thislocation prevents water reabsortion and an excess-ive amount of fluid enters the medulla (Agre and

www.biolcell.org | Volume 97 (9) | Pages 687–697 687

P. Ford and others

Kozono, 2003). Therefore the reabsorption of waterin the CCD is important for the overall operation ofthe urinary concentration mechanism.

Alterations of luminal osmolality would inducechanges, at least transients, in cell volume. In re-cent years, it has become evident that the volume ofa given cell is an important factor in defining cellularfunctions, such as transepithelial transport, cell mi-gration, cell growth, cell death and the regulationof intracellular metabolism (O’Neill, 1999; Okadaet al., 2001). To avoid damage that may occur afteran acute change in volume, cells have developedimportant volume-regulatory mechanisms. In manysystems, hypo-osmotic shocks induce cell swellingand activation of RVD (regulatory volume decrease)mechanisms, involved in solutes and water efflux ofthe cell (Okada et al., 2001; Wehner et al., 2003).On the other hand, hyperosmotic shocks induce cellshrinkage, which activates RVI (regulatory volumeincrease) mechanisms, causing influx of solutes andwater into the cell (Waldegger et al., 1998).

Despite its importance, very little is known aboutcell-volume regulation in CCD cells (Strange, 1988;Schwiebert et al., 1994; Natke, 1996). Moreover, nostudies have been carried out so far to evaluate theputative role of AQPs during cell-volume regulationin renal cells.

The aim of the present work was to study cell-volume regulation after hypo- or hyperosmotic chal-lenges, in two CCD cell lines: one not expressingAQPs (WT-RCCD1, where WT stands for wild-typecells) (Capurro et al., 2001) and the other AQP2-RCCD1 (RCCD1 cells stably transfected withAQP2). For this purpose, we have used a fluorescentprobe technique, adapted to confluent RCCD1 cellsgrown on coverslips, in which the acquisition ofsingle-cell kinetic data can be simultaneously re-corded with the pHi (intracellular pH) (Hamannet al., 2002). Our results showed that hyperosmoticshocks, generated with mannitol, induced cell shrink-age but not RVI activation in both cell lines. In con-trast, under hypo-osmotic shocks, we demonstratedthat RVD occurs in both cell lines, but the activationof these mechanisms is more rapid in AQP2-RCCD1

cells. This regulatory response takes place in parallelwith intracellular acidification. Interestingly, cellu-lar acidification is more rapid in AQP2-RCCD1 cellsand sensitive to Ba2+ and glibenclamide. Altogether,these results suggest that increases in the osmotic

water permeability due to the expression of AQP2,could be critical for the rapid activation of cell-volume decrease mechanisms.

ResultsCell-volume measurements in RCCD1 cellmonolayers: fluorescence versus externalosmolalityTo study cell-volume changes in WT-RCCD1 andAQP2-RCCD1 cells, we have used the technique de-scribed by Hamann et al. (2002) for epithelial mono-layers. Figure 1(A) shows a representative experimentperformed in WT-RCCD1 cells. When cells wereloaded with BCECF [2′,7′-bis-2-carboxyethyl)-5(6)-carboxyfluorescein], the fluorescence intensity (F) in-side the cell was not homogeneous, being higher inthe central area compared with the peripheral area.Therefore we have studied the time evolution of F intwo small areas of the cell (pinhole), one located atthe centre (C) and the other at the periphery (P).Figure 1(B) shows that after hyperosmotic shocks(+100 mOsM mannitol) F, measured in P, alwaysdecreased. In contrast, no reproducible signals wereobserved when the pinhole was located at C.

Comparative experiments with the volume markercalcein yielded results similar to those obtained withBCECF, in line with reports from others (Wehneret al., 1995; Hamann et al., 2002; Pafundo et al.,2004). Moreover, decreases in the concentration of thedye or changes in the incubation time did not modifythe response (results not shown).

To investigate further the dependence of F on ex-ternal osmolality, different hyper- and hypo-osmoticgradients were applied. Figure 1(C) shows that theevolution of relative fluorescence [Ft/F0, where Ft isthe fluorescence from a region of the cell at time tand F0 is the fluorescence intensity obtained froma small region of the cell (pinhole) in iso-osmoticmedium], measured in P, was inversely proportionalto the external osmolality. Moreover, Ft/F0 showed alinear correlation with the relative external osmolality(OsM0/OsMt, where OsM0 stands for external osmo-lality at zero time and OsMt for external osmolality attime t) (Figure 1D). Similar results were obtained inAQP2-RCCD1 cells (results not shown). Fluorescencesignals were then transformed into relative cell-watervolume (V/V0, where V stands for cell-water volumeand V0 stands for cell-water volume at zero time) asdescribed in the Materials and methods section.

688 C© Portland Press 2005 | www.biolcell.org

AQP2 and cell-volume regulation Research article

Figure 1 Dependence of fluorescence intensity on the cell area and external osmolality in WT-RCCD1 cells(A) Upper panel: representative image of cells loaded with 2 µM BCECF; scale bar, 10 µm. Lower panel: three-dimensional

reconstruction of fluorescence intensity (F), showing the two small areas (pinhole) studied: peripheral (P) and central (C). (B) Time

evolution of F, in P and C, when a hyperosmotic shock (+100 mOsM mannitol) was applied. (C) Representative recording

of changes in relative F (Ft/F0, in P) in a cell exposed to a series of solutions having osmolalities as indicated in each box.

(D) Relationship between Ft/F0 and the reciprocal of the relative osmolality of the bathing solution (OsM0/OsMt); r represents

the regression coefficient. At least 20 cells were studied in each experiment.

Osmotic water permeability in WT-RCCD1 versusAQP2-RCCD1 cellsThe osmotic water permeability (Pf ) was estimatedin WT-RCCD1 and AQP2-RCCD1 cells. For thispurpose the time course of the relative cell volume(V/V0) was measured in response to a hyperosmoticgradient (+100 mOsM mannitol). Figure 2(A) showsthat the rate of cell volume shrinkage was signi-ficantly higher in AQP2-RCCD1 cells. However,within 2 min of exposure to the hyperosmotic solu-tion, both WT-RCCD1 and AQP2-RCCD1 cellsreach similar cell-volume values. Confocal immuno-fluorescence microscopy using a rat anti-AQP2 anti-body confirmed a constitutive expression of AQP2in the apical membrane (Figure 2A, inset). Pf wascalculated as described in the Materials and methodssection and values are summarized in Figure 2(B).Results show that Pf values were significantly higherin AQP2-RCCD1 cells with respect to WT-RCCD1

cells [Pf = (123 +− 7) × 10−4 and (10 +− 0.6) ×10−4 cm/s respectively, n = 6]. Moreover, Pf inAQP2-RCCD1 cells, but not in WT-RCCD1 cells,was significantly inhibited by HgCl2 [(22 +− 4) ×10−4 cm/s, n = 4].

Effects of mannitol and urea hyperosmolalityon cell volumeTo investigate further the characteristics of cell-volume response in both the cell lines, osmoticshocks were performed by adding an impermeablesolute (100 mOsM mannitol) or a permeable one(100 mOsM urea). Figure 3(A) shows the time courseevolution of V/V0 in WT-RCCD1 and AQP2-RCCD1

cells. It can be observed that, in both cell lines, hyper-osmotic mannitol induced cell shrinkage by approx.25% of their control volume. During the experi-mental period (10 min), both cell lines remained ina shrunken state, indicating no activation of RVI


P. Ford and others

Figure 2 Increased plasma-membrane waterpermeability in AQP2-RCCD1 cells(A) Relative volume (V/V0) time course of WT and

AQP2-transfected cells in response to a hyperosmotic gradi-

ent (+100 mOsM mannitol). Inset: X-Z image of AQP2-RCCD1

cells localization using a polyclonal rat anti-AQP2 antibody

(red) and nuclear staining using Sytox (green) as described

in the Materials and methods section. Scale bar, 20 µm.

(B) Effect of 0.3 mM HgCl2 (inhibitor of AQP2) on the

osmotic water permeability coefficient (Pf) of WT and

AQP2-transfected cells. Values represent the means +− S.E.M.

for four to six experiments. *P < 0.001, Student’s t test for

unpaired data.

mechanisms. Increasing the length of the experi-mental period to 30 min did not change this observ-ation (results not shown).

When the bathing media were made hyperosmoticby the addition of urea, both RCCD1 cell lines res-ponded with a transient osmotic shrinkage (Fig-ure 3A). The maximal decrease in relative volumewas less than that observed with mannitol for WT-RCCD1. Cell-volume recovery was completely abol-

Figure 3 Volume changes in WT-RCCD1 and AQP2-RCCD1 cells under hyperosmotic shocks(A) At zero time, bathing osmolality was increased by the ad-

dition of 100 mOsM mannitol. Both WT-RCCD1 and AQP2-

RCCD1 cells decrease their volumes even though shrinkage is

more rapid in AQP2 cells. Cells were then returned to the con-

trol medium and, after equilibration, the bathing solution was

again increased but now by the addition of 100 mOsM urea.

(B, C) Effects of 300 µM phloretin on volume changes in WT-

RCCD1 and AQP2-RCCD1 cells respectively. Cells were first

incubated for 10 min in iso-osmotic solution + phloretin and

then treated with hyperosmotic urea solution + phloretin.

ished in the presence of phloretin, a specific inhibitorof urea transporters, in both cell lines (Figures 3Band 3C). These results indicate that urea enters thecell through a specific transporter and water followsit, reestablishing the initial cell volume.

RVD in WT-RCCD1 and AQP2-RCCD1 cellsWe have also investigated the regulatory volume re-sponse to hypo-osmotic solutions (−100 mOsM man-nitol) in RCCD1 cells. Figures 4(A) and 4(B) showthe time course of V/V0 in WT-RCCD1 and AQP2-RCCD1 cells respectively. Similar to the case of hyper-osmotic shocks, Pf values were significantly higher forAQP2-RCCD1 cells (Table 1). It can also be noticedthat both cell lines swelled to a similar maximumvolume. However, the time to reach this value was



Figure 4 RVD in WT-RCCD1 and AQP2-RCCD1 cells(A, B) Time course of relative volume changes (V/V0)

in WT-RCCD1 and AQP2-RCCD1 cells respectively. Hypo-

osmotic medium was achieved by subtracting 100 mOsM

mannitol from the control solution. Parameters used to eval-

uate the RVD response are: peak swelling (a), time to peak

(b) and RVD latency (c) values (summarized in Table 1). (C) In-

tracellular pH was simultaneously measured with cell volume.

Intracellular acidification occurs at 10 min in WT-RCCD1 cells

and at 1 min in AQP2-RCCD1. *P < 0.05, Student’s t test for

unpaired data.

faster for AQP2-RCCD1 cells (Table 1). After max-imal swelling, the delay before RVD activation (RVDlatency) was significantly higher for WT-RCCD1

cells. Then, 7 min after the hypo-osmotic shock,while WT-RCCD1 cells are still reaching their maxi-mal swelling, AQP2 cells have already regulated ap-prox. 15% of their volume. The initial rate of RVDactivation and the percentage cell-volume recoveryat 30 min (RVD30) were not significantly differentbetween both cell lines. All these values are summar-ized in Table 1.

Table 1 Osmotic swelling and RVD in WT-RCCD1 and AQP2-RCCD1 cellsPf is the osmotic water permeability coefficient calculated when ahypo-osmotic shock (−100 mOsM mannitol) was imposed. (a) Peakswelling represents the percentage of maximal swelling; (b) timeto peak represents the time (min) to reach the maximal swelling;(c) RVD latency is the time (min) between the peak and the start ofcell-volume regulation; (d) RVD rate is the initial (first 3 min) rateof RVD; and (e) percentage RVD30 denotes the percentage ofvolume regulation at 30 min. NS, not significant.

Variable WT-RCCD1 AQP2-RCCD1 P

Pf (×10−4) (cm/s) 8 +− 0.3 121 +− 6 P < 0.001

(a) Peak swelling (%) 54.33 +− 0.88 53.27 +− 0.39 NS

(b) Time to peak (min) 7.02 +− 0.30 1.55 +− 0.16 P < 0.001

(c) RVD latency (min) 4.21 +− 0.16 1.02 +− 0.07 P < 0.001

(d) RVD rate (%/min) −2.10 +− 0.26 −2.20 +− 0.16 NS

(e) Percentage RVD30 54.80 +− 1.7 55.44 +− 1.9 NS

pHi were simultaneously recorded with the volumechanges (Figure 4C). In WT-RCCD1 cells, pHi sig-nificantly decreased 10 min after the hypo-osmoticshock. In AQP2-RCCD1 cells, on the other hand, in-tracellular acidification was significantly faster (1 minafter cell swelling).

Effects of glibenclamide and BaCl2 on the RVDand pHi in RCCD1 cellsIn a later series of experiments, we have investi-gated the effects of 10−4 M glibenclamide [inhibitorof CFTR (cystic fibrosis transmembrane conductanceregulator)] and 2 × 10−4 M BaCl2 (inhibitor of K+channels) on RVD and pHi. Figure 5 shows that,in WT-RCCD1 cells, the time to reach the maximalswelling (time to peak) and the RVD latency were notaffected either by glibenclamide or by BaCl2 (Fig-ures 5A and 5B respectively). In contrast, in AQP2-RCCD1 cells, both parameters were significantly in-creased, reaching values similar to those observed inWT-RCCD1 cells. In both cell lines, the rate of RVDactivation was not modified by the blockers (Fig-ure 5C). Finally, the percentage of volume recoverywas significantly decreased by both blockers only inAQP2-RCCD1 cells (Figure 5D).

Figure 6 shows the effects of glibenclamide andBaCl2 on pHi in both cells lines. In WT-RCCD1

cells, pHi was not affected by the inhibitors, either at1 min or at 10 min after the hypo-osmotic swelling(Figure 6A). In contrast, in cells expressing AQP2,the initial swelling-induced cytosolic acidification(1 min after the hypo-osmotic shock) was completelystopped (Figure 6B). An intracellular acidification


P. Ford and others

Figure 5 Effects of glibenclamide and BaCl2 on RVDresponse in WT-RCCD1 and AQP2-RCCD1 cellsRCCD1 cells were first treated for 10 min with control solu-

tion containing 10−4 M glibenclamide (Glib) or 2 × 10−4 M

BaCl2. Then cells were treated with a hypo-osmotic solution

(−100 mOsM mannitol) always in the presence of inhibitors.

Control experiments were performed in the absence of block-

ers. Parameters calculated to evaluate the RVD response, in

the absence or presence of blockers, are: (A) time to peak

(min), (B) RVD latency (min), (C) RVD rate (%/min) and (D) per-

centage RVD30. Results are the means +− S.E.M. for three to

five independent experiments (at least 20 cells were recorded

in each experiment). *P < 0.02, **P < 0.01, ***P < 0.001, com-

pared with controls (Student’s t test for unpaired data).

Figure 6 Effects of glibenclamide and BaCl2 on pHi inRCCD1 cellsWT-RCCD1 (A) and AQP2-RCCD1 (B) cells were treated with

10−4 M glibenclamide (Glib) or 2 × 10−4 M BaCl2. The effects

of inhibitors on pHi are represented at 1 and 10 min after the

hypo-osmotic shock. Glib and BaCl2 significantly prevented

intracellular acidification (at 1 min) and decreased the mag-

nitude of cytosolic acidification (at 10 min) only in AQP2 cells.

Results shown are the means +− S.E.M. for three to five inde-

pendent experiments (at least 20 cells were recorded in each

experiment). *P < 0.05, **P < 0.02, ***P < 0.01, compared with

controls (Student’s t test for unpaired data).

appeared 10 min after the hypo-osmotic shock, andit was significantly lower than that in controls(Figure 6B).

Altogether these results indicate that, in cells ex-pressing AQP2, the rapid RVD activation and theinitial swelling-induced acidification would be asso-ciated with the same transporters.

DiscussionGiven the constantly changing nature of transepi-thelial solute and water transport during the trans-ition from antidiuresis to diuresis, cells of the



collecting tubule are faced with continuous chal-lenges to intracellular homoeostasis. It is now becom-ing apparent that cells possess a variety of mechanismsto respond to volume changes. However, cell-volumeregulation and the involvement of AQP2 in theseevents in CCD cells have not yet been investigated.Therefore, in the present study, we have used a fluor-escence technique to study the cell-volume regu-lation after osmotic shocks in WT-RCCD1 andAQP2-RCCD1 cells.

Measurement of cell-volume changes in confluentRCCD1 monolayersMeasurement of water transport across cell mem-branes, with high-temporal resolution and sensitiv-ity, is essential to determine osmotic water permeab-ility and changes in cell volume. During the lasttwo decades, several methods have been developed,founded on distinct biophysical principles (Verkman,2000). In single isolated cells, the use of fluorescentdye dilution, based on an inversely proportional rela-tionship between relative fluorescence intensity andrelative cell-water volume, proved to be useful formeasuring volume changes (Muallem et al., 1992;Crowe et al., 1995). However, the application of thistechnique to confluent monolayers of epithelial cellsresulted in a directly proportional relationship, op-posite to what was expected (Wehner et al., 1995;Raat et al., 1996; Hamann et al., 2002). These resultswere interpreted by some of the authors as resultingfrom fluorescence self-quenching.

Our present results show that, in RCCD1 cellsgrown on coverslips and loaded with BCECF, fluor-escence intensity had a bell-shape distribution (Fig-ure 1A). After an osmotic shock, different responseswere obtained when the studied area (pinhole) waslocated either at the centre (C) or at the periphery(P) of the cell. Whereas, in P, changes in fluorescence(F) were always directly proportional to cell-watervolume, they varied in C. These responses were notmodified by loading RCCD1 cells with low concen-trations of the dye, by changing the incubation timeor using another fluorophore (calcein).

It has previously been discussed that one limita-tion of fluorescence video microscopy as a methodfor cell-volume measurements is that it collects theemitted light from a layer comparable with the cellheight (Zelenina and Brismar, 2000). To solve thislimitation, some geometrical considerations can be

Figure 7 Geometrical parameters involved in volumechanges when using videomicroscopy fluorescencetechniquesSchematic representation of an isolated and monolayer cell

when treated with an iso-osmotic (A) or hypo-osmotic (B)

solution. l, thickness of the layer in which the emitted light

is collected; d, path length; h, cell height; C, central area; P,

peripheral area; hP, cell height in P; dp, path length in P; hC,

cell height in C; dC, path length in C; FP, fluorescence intensity

measured in P; FC, fluorescence intensity measured in C.

taken into account (Figure 7). Three parameters canbe defined: the thickness of the layer in which theemitted light is collected (l), the path length (d) andthe cell height (h). As described by the Lambert–Beerlaw, F is directly proportional to the concentration ofthe fluorophore (c) and to d (F ∝ cd).

In isolated spherical cells, h probably exceeds l, andd is always equal to l. In this case, if an osmotic shockis applied, the path length value remains constant(Figures 7A and 7B). Therefore F is inversely pro-portional to the cell volume since only c is changingas it was reported previously (Muallem et al., 1992;Crowe et al., 1995).

When monolayers are used, the situation is quitedifferent. In our experiments, fluorescence has a bell-shaped distribution, which probably reflects that cellheight is bigger at the centre than in the periphery(hC > hP, Figure 7A). The path length in P (dP) is al-ways lower than l and it significantly changes when anosmotic shock is applied (Figure 7B). Consequently, ifchanges in dP are bigger than changes in fluorophoreconcentration, F would be directly proportional to


P. Ford and others

cell volume. In RCCD1 monolayers, the differencesobserved between both the studied regions (P and C)may be explained by the fact that cells can easilyincrease (under hypo-osmotic shocks) or decrease(under hyperosmotic shocks) their height in P, withlittle or no modification of this parameter in C.

We propose here that, in RCCD1 cells, the thick-ness of the sample, more than self-quenching alone,can be intimately related to the direct relationship ob-served between fluorescence and cell-water volume.

The linear relationship always observed in P, be-tween relative fluorescence (Ft/F0) and external rela-tive osmolality (OsM0/OsMt) validates this tech-nique. Moreover, our present results show that Pf

values in WT-RCCD1 cells are in good agreementwith those obtained, using other techniques, for dif-ferent renal cell lines (Farinas et al., 1995; Valentiet al., 1996; Deen et al, 1997; Maric et al., 2001), aswell as in lipid bilayers not containing water channels(Finkelstein, 1987). Pf values in AQP2-RCCD1 cells,which have a constitutive expression of AQP2 in theapical plasma membrane, were approx. 12-fold higherthan those in WT-RCCD1 cells and were signifi-cantly inhibited by HgCl2. Similarly, previous worksin AQP2-transfected cells reported a 2–10-fold in-crease in Pf after AVP or forskolin stimulation(Katsura et al., 1995; Valenti et al., 1996; Deen et al.,1997; Zelenina and Brismar, 2000; Maricet al., 2001).

Cell-volume regulation in renal cellsExperiments with hyperosmotic mannitol mediademonstrated that, independent of AQP2 expression,RCCD1 cells shrink but fail to show RVI, at leastunder the studied conditions (cells grown on cover-slips). However, we have previously demonstratedthat, in RCCD1 cells grown on permeable supports,the isoform NHE2 of the Na+/H+ exchanger (oneof the transporters involved in RVI) is expressed inthe basolateral membrane and is mainly activatedafter basolateral hyperosmotic shock (Ford et al.,2002). In view of the fact that, in the present study,cells are grown on coverslips, osmotic gradients aremore likely to be achieved from the luminal bath.The absence of hyperosmotic RVI, after luminal os-motic shocks, may be better explained not by thelack of specific volume regulatory transporters but bythe fact that they are less or not activated as they arelocalized in the basolateral membrane.

Volume changes in the presence of urea gradients(permeable solute) were the result of solute diffusion(blocked by phloretin) and not of volume regulation.However, this observation also allowed us to con-firm that RCCD1 cells are capable of recovering theirvolume when a solute is entering.

Regarding hypo-osmotic shock experiments, wedemonstrated that RVD occurs in RCCD1 cells andthe activation of these mechanisms is more rapid inAQP2-RCCD1 cells. Moreover, the time between themaximal swelling and the beginning of cell-volumeregulation (RVD latency) is significantly shorter inAQP2-RCCD1 cells. Our results also showed thatRVD takes place in parallel with intracellular acidi-fication. This decrease in pHi was observed previouslyin AQP2-RCCD1 cells in line with the faster RVDactivation.

In many mammalian cells, it was reported previ-ously that the predominant pathway for RVD is theopening of K+ and anion channels. The K+ channelsinvolved are usually large-conductance Ca2+-activ-ated channels (O’Neill, 1999). The molecular basisof the anion conductance is unknown and may com-prise more than one channel such as CIC-2, maxi Cl−channel, outwardly rectifying channel (VSOAC) andCFTR (Strange et al., 1996; O’Neill, 1999). Particu-larly, the anion channel CFTR was shown to accelerateand potentiate the RVD response (Schwiebert, 1999).We reported previously that RCCD1 cells expressCFTR, at the mRNA and protein levels (Djelidi et al.,1999), and also possess a functional Ba2+-sensitiveapical K+ channel (Blot-Chabaud et al., 1996). In thepresent study, we demonstrated that the rapid activ-ation of RVD mechanisms that occurs in AQP2 ex-pressing cells was inhibited by glibenclamide (in-hibitor of CFTR) and BaCl2. Moreover, the initial(1 min) swelling-induced acidification disappearedin the presence of these blockers (Figures 5 and 6).These results suggest a direct relationship betweenthe rapid RVD activation and the immediate intra-cellular swelling-induced acidification. Despite theeffects of blockers on the rapid activation of RVDand the initial intracellular acidification, AQP2 cellsare still able to recover their volume. However, thepercentage of volume recovery and the subsequentacidification were low.

Several works reported that cell swelling may leadto (i) increase in intracellular Ca2+ activity; (ii) modi-fication of the architecture of the cytoskeleton and



modification of the expression of some proteins (e.g.K+ channels, Na+ channels and CFTR); (iii) modifi-cation of the phosphorylation of a variety of proteins;and (iv) the stimulation of ATP release (Waldeggeret al., 1998; Schwiebert, 1999). Especially, ATP re-lease was found to be enhanced by the expressionof CFTR and would facilitate the RVD processes(Hazama et al., 2000). Previously, some authorshave also proposed that CFTR directly or indirectlyprovides a bicarbonate conductance (Clarke et al.,2001), while others suggested that it may serve asa ‘HCO3

− sensor’ to regulate the activity of othertransporters (Park et al., 2002).

Results of the present study thus allow us tohypothesize that, in AQP2 cells, a rapid swell-ing occurs and it induces the activation of CFTRand K+ channels. Consequently, this would cause theexit of HCO3

− and K+ from the cells, resulting ina rapid intracellular acidification. The disappearanceof this initial cytosolic acidification after blockingthe Ba2+-sensitive K+ channel could be an indirecteffect.

Whether these processes can be directly or indi-rectly linked to AQP regulation must be further in-vestigated. To date, only one work has been recentlyreported that proposed the participation of a waterchannel in the RVD processes. In that study, the au-thors demonstrated an impairment of RVD in thecorneal epithelial cells of mice knocked out for AQP1(Kuang et al., 2004). Hill et al. (2004) proposed amechanistic model to demonstrate how AQPs mightfunction as sensors. The model suggests that AQPsmay constitute an osmosensor that is required to ac-tivate RVD or RVI by signalling to the channelsor transporters that define the steady-state osmolytecontent (Hill et al., 2004). Our results are in linewith this proposal in view of the fact that they high-light the importance of AQP osmosensing for a rapidactivation of the RVD mechanisms. However, otherslowly activated RVD mechanisms may also be in-volved in the overall response in cells expressing ornot expressing AQP2.

In summary, results of the present study suggestthat increases in the osmotic water permeability dueto the expression of AQP2 is critical for a rapid ac-tivation of cell-volume decrease mechanisms. The in-vestigation of whether AQPs may form part of thecellular device for volume sensing and regulation isof physiological relevance.

Materials and methodsCell cultureRCCD1 cells were grown in modified DM medium [Dulbecco’smodified Eagle’s medium/Ham’s F-12 (1:1, v/v), 14 mMNaHCO3, 3.2 mM glutamine, 5 × 10−8 M dexamethasone,3 × 10−8 M sodium selenite, 5 µg/ml insulin, 10 µg/ml epi-dermal growth factor, 5 × 10−8 M tri-iodothyronine, 10 units/ml penicillin–streptomycin, 20 mM Hepes, pH 7.4] and 2%(v/v) fetal bovine serum (Gibco BRL, Gaithersburg, MD, U.S.A.)(Blot-Chabaud et al., 1996). All experiments were performed onconfluent cells, between the 20th and 40th passages, grownon coverslips for 2 or 3 days.

Stable transfection of RCCD1 cells with AQP2and immunolocalizationThe cDNA coding for rat AQP2 was inserted into the KpnI–BamHI restriction sites of the pcB6 expression vector. Cells(5 × 105/well) were plated in six-well dishes 24 h before trans-fection. For each well, 100 µl of OPTIMEM (Invitrogen,Carlsbad, CA, U.S.A.) containing 1 µg of recombinant plas-mid was incubated at room temperature (24◦C) for 15 min. Themixture was then combined with 5 µg of Lipofectin®, dilutedto 100 µl of OPTIMEM, and incubated again at room tempera-ture for 15 min. After the addition of 800 µl of OPTIMEM, theblend was added to the cells. Cells were then incubated overnightat 37◦C before replacement of the medium by DM containing2% fetal bovine serum. The cells were grown for 2 days, thentrypsinized and seeded at very low density. A selection of cellscontaining transfected DNA was obtained with a medium con-taining Geneticin (400 µg/ml; Life Technologies, Gaithersburg,MD, U.S.A.). Resistant clones were isolated, amplified and char-acterized for AQP2 expression. One of these clones was selectedfor further studies.

AQP2 expression was examined using the rabbit polyclonalanti-rat AQP2, generated against a peptide corresponding toamino acids 250–271 of AQP2, conjugated with keyhole-limpethaemocyanin (Neosystem, Strassburg, France) (Shi et al., 2001).Cells, grown on permeable filters, were fixed with methanolfor 10 min at −20◦C, and incubated with the AQP2 antibody(1:100 dilution) for 1 h at room temperature. After three rinsesin PBS, the samples were incubated with a CY3-conjugated goatanti-rabbit IgG antibody (1:200 dilution; Jackson Immuno-research Laboratories, West Grove, PA, U.S.A.) for 30 min atroom temperature. Nuclei were counterstained with Sytox Green(1:500 dilution; Molecular Probes, Eugene, OR, U.S.A.). Allsamples were examined under a Zeiss photomicroscope equippedwith epifluorescence optics (Zeiss, Oberkochen, Germany) andby confocal laser scanning microscopy (Leica, Wetzlar,Germany).

Cell-volume changesConfluent WT-RCCD1 and AQP2-RCCD1 cells grown on cover-slips were mounted on a chamber and loaded with 2 µMBCECF/AM (where AM stands for acetoxymethylester; Molecu-lar Probes) for 30 min at 37◦C. The chamber was then placedon the stage of a Nikon TE-200 epifluorescence inverted micro-scope (Nikon Planfluor ×40 oil immersion objective lens). Cellswere subsequently bathed at 20◦C in a dye-free solution for atleast 15 min before use in experiments. During experiments,the bathing solution was exchanged by aspirating the media andadding a new one. This fluid exchange was completed within


P. Ford and others

5 s. Fluorescence intensity was recorded by exciting BCECF atthe isosbestic point (440 nm) where the fluorochrome is pH-insensitive. Comparative experiments were performed using thevolume marker calcein (Molecular Probes). Various concentra-tions of dyes and different incubation time periods were alsotested. One advantage of BCECF is that it permits the simul-taneous measurement of changes in cell-water volume and pHi.

Fluorescence data were acquired every 10 s by using a charge-coupled-device camera (Hamamatsu C4742-95) connected toa computer and the Metafluor acquisition program (UniversalImaging, West Chester, PA, U.S.A.). The procedure to estimatecell-water volume was similar to the one described for retinalpigment epithelial cells (Hamann et al., 2002). The change incell-water volume can be calculated as

V

V0=

(Ft

F0

)− fb

1 − fb

where V is the cell-water volume at time t and V0 is the cell-water volume at zero time; F0 represents the signal obtained froma small region of the cell (pinhole) equilibrated with iso-osmoticmedium having an osmolality OsM0; Ft is the fluorescence fromthe same region at time t in a solution having an osmolality OsMt;and f b is the relative background. This parameter correspondsto the y intercept of a plot of Ft/F0 versus OsM0/OsMt, which isthe relative fluorescence when no osmolality change occurred.

Osmotic water permeability and cell-volume regulationThe time course of V/V0 for each experiment was fitted witha single exponential function. The osmotic water permeability(Pf ) was calculated from the exponential time constant (τ) byusing the relation

Pf = V0

τ · A · �OsM · Vw

where A is the cell-surface area, V0 is the initial cell-watervolume, �OsM is the osmotic gradient and VW is the partialmolar volume of water (18 cm3/mol).

The RVD associated with the volumetric response of cellsexposed to hypo-osmotic medium was calculated by the use ofthe following equation:

RVDt =

(V

V0

)max

−(

V

V0

)t(

V

V0

)max

− 1

× 100

where (V/V0)max is the maximal value of V/V0 attained duringhypo-osmotic swelling (peak), and (V/V0)t represents the valueof V/V0 observed at time t. RVDt thus denotes the magni-tude of volume regulation at time t, with 100% RVD indicatingcomplete volume regulation and 0% RVD indicating no volumeregulation.

Intracellular pH experimentspHi was recorded simultaneously along with cell volume. Thecells were loaded with BCECF/AM as described above and il-luminated with alternate excitation wavelengths of 440 and490 nm. Changes in pHi were inferred from the 440/490 ratioof the fluorescence emitted from dye-loaded cells. At the end of

each experiment, the ratio was calibrated in terms of pH valueby incubating the cells in ‘high K+ solution’ (in mM: 140 KCl,4.6 NaCl, 1 MgCl2, 2 CaCl2, 10 Hepes and 10 glucose, pH 7.4)containing 5 µM nigericin. Then the bathing solution waschanged by solutions of known pH (pH range 6.4–8.2), allow-ing the fluorescence ratio to reach a steady state at each pH.

Solutions and chemicalsThe iso-osmotic solution (osmolality, 320 +− 4 mOsM) contained(in mM): 90 NaCl, 10 NaHCO3, 5 KCl, 1 CaCl2, 0.8 MgSO4,1 MgCl2, 100 mannitol, 20 Hepes and 5 glucose. Aniso-osmoticsolutions were prepared by mannitol addition or removal to ob-tain the desired osmolality, thus maintaining the ionic strength.In some experiments, hyperosmotic solutions were prepared bythe addition of 100 mOsM urea to the iso-osmotic solution. Theosmolalities were always measured by a pressure vapour osmo-meter (Wescor, Logan, UT, U.S.A.). All solutions were titratedto pH 7.40 using Tris (Sigma) and bubbled with atmosphericair.

In some experiments, 0.3 mM HgCl2 or 300 µM phlor-etin (3-[4-hydroxyphenyl]-1-[2,4,6-tri-hydroxyphenyl]-1-pro-panone; C15H14O5; Sigma) were used as water- and urea-transport inhibitors respectively.

In another experimental series, 10−4 M glibenclamide(Sigma) and 2 × 10−4 M BaCl2 (Sigma) were used. Gliben-clamide was prepared as a 10−2 M stock solution in DMSO anddiluted to its appropriate bath solution. In control experiments,no effect of vehicle was seen (results not shown).

StatisticsResults are reported as means +− S.E.M., and n is the number ofexperiments performed. In each experiment, at least 20 indi-vidual cells were evaluated. For all comparisons, Student’s t testfor unpaired data was applied and P < 0.05 was considered to bestatistically significant.

AcknowledgmentsThis work was supported by grants fromFondo Nacional para la Ciencia y la Tecnologıa(Argentina), Consejo Nacional de InvestigacionesCientıficas y Tecnicas (Argentina), Universidadde Buenos Aires (Argentina) and FundacionANTORCHAS (Argentina).

ReferencesAgre, P. and Kozono, D. (2003) Aquaporin water channels: molecular

mechanisms for human diseases. FEBS Lett. 555, 71–78Blot-Chabaud, M., Laplace, M., Cluzeaud, F., Capurro, C.,

Cassingena, R., Vandewalle, A., Farman, N. and Bonvalet, J.P.(1996) Characteristics of a rat cortical collecting duct cell line thatmaintains high transepithelial resistance. Kidney Int. 50, 367–376

Capurro, C., Rivarola, V., Kierbel, A., Escoubet, B., Farman, N.,Blot-Chabaud, M. and Parisi, M. (2001) Vasopressin regulateswater flow in a rat cortical collecting duct cell line not containingknown aquaporins. J. Membr. Biol. 179, 63–70

Clarke, L.L., Stien, X. and Walker, N.M. (2001) Intestinal bicarbonatesecretion in cystic fibrosis mice. JOP 2, 263–267



Crowe, W.E., Altamirano, J., Huerto, L. and Alvarez-Leefmans, F.J.(1995) Volume changes in single N1E-115 neuroblastoma cellsmeasured with a fluorescent probe. Neuroscience 69, 283–296

Deen, P.M., Verdijk, M.A., Knoers, N.V., Wieringa, B., Monnens, L.A.,van Os, C.H. and van Oost, B. (1994) Requirement of human renalwater channel aquaporin-2 for vasopressin-dependentconcentration of urine. Science 264, 92–95

Deen, P.M., Rijss, J.P., Mulders, S.M., Errington, R.J., van Baal, J.and van Os, C.H. (1997) Aquaporin-2 transfection of Madin-Darbycanine kidney cells reconstitutes vasopressin-regulatedtranscellular osmotic water transport. J. Am. Soc. Nephrol. 8,1493–1501

Djelidi, S., Fay, M., Cluzeaud, F., Thomas-Soumarmon, A., Bonvalet,J.P., Farman, N. and Blot-Chabaud, M. (1999) Vasopressinstimulates long-term net chloride secretion in cortical collectingduct cells. FEBS Lett. 460, 533–538

Farinas, J., Simanek, V. and Verkman, A.S. (1995) Cell volumemeasured by total internal reflection microfluorimetry: applicationto water and solute transport in cells transfected with waterchannel homologs. Biophys. J. 68, 1613–1620

Finkelstein, A. (1987) Water Movement through Lipid Bilayers, Poresand Plasma Membranes. John Wiley, New York

Ford, P., Rivarola, V., Kierbel, A., Chara, O., Blot-Chabaud, M.,Farman, N., Parisi, M. and Capurro, C. (2002) Differential role ofNa+/H+ exchange isoforms NHE-1 and NHE-2 in a rat corticalcollecting duct cell line. J. Membr. Biol. 190, 117–125

Hamann, S., Kiilgaard, J.F., Litman, T., Alvarez-Leefmans, F.J.,Winther, B.R. and Zeuthen, T. (2002) Measurement of cell volumechanges by fluorescence self-quenching. J. Fluorescence 12,139–145

Hazama, A., Fan, H.T., Abdullaev, I., Maeno, E., Tanaka, S.,Ando-Akatsuka, Y. and Okada, Y. (2000) Swelling-activated,cystic fibrosis transmembrane conductance regulator-augmentedATP release and Cl− conductances in murine C127 cells.J. Physiol. (Cambridge, U.K.) 523, 1–11

Hill, A.E., Shachar-Hill, B. and Shachar-Hill, Y. (2004) What areaquaporins for? J. Membr. Biol. 197, 1–32

Katsura, T., Verbavatz, J.M., Farinas, J., Ma, T., Ausiello, D.A.,Verkman, A.S. and Brown, D. (1995) Constitutive and regulatedmembrane expression of aquaporin 1 and aquaporin 2 waterchannels in stably transfected LLC-PK1 epithelial cells. Proc. Natl.Acad. Sci. U.S.A. 92, 7212–7216

Kuang, K., Yiming, M., Wen, Q., Li, Y., Ma, L., Iserovich, P., Verkman,A.S. and Fischbarg, J. (2004) Fluid transport across cultured layersof corneal endothelium from aquaporin-1 null mice.Exp. Eye Res. 78, 791–798

Maric, K., Wiesner, B., Lorenz, D., Klussmann, E., Betz, T. andRosenthal, W. (2001) Cell volume kinetics of adherent epithelialcells measured by laser scanning reflection microscopy:determination of water permeability changes of renal principalcells. Biophys. J. 80, 1783–1790

Muallem, S., Zhang, B.X., Loessberg, P.A. and Star, R.A. (1992)Simultaneous recording of cell volume changes and intracellularpH or Ca2+ concentration in single osteosarcoma cellsUMR-106-01. J. Biol. Chem. 267, 17658–17664

Natke, Jr, E. (1996) Cell volume regulation of rabbit cortical collectingtubule in anisotonic media. Am. J. Physiol. 258, F1657–F1665

Nielsen, S., Frokiaer, J., Marples, D., Kwon, T.H., Agre, P. andKnepper, M.A. (2002) Aquaporins in the kidney: from molecules tomedicine. Physiol. Rev. 28, 205–244

Okada, Y., Maeno, E., Shimizu, T., Dezaki, K., Wang, J. andMorishima, S. (2001) Receptor-mediated control of regulatoryvolume decrease (RVD) and apoptotic volume decrease (AVD).J. Physiol. (Cambridge, U.K.) 532, 3–16

O’Neill, W.C. (1999) Physiological significance of volume-regulatorytransporters. Am. J. Physiol. 276, C995–C1011

Pafundo, D.E., Mut, P., Perez Recalde, M., Gonzalez-Lebrero, R.M.,Fachino, V., Krumschnabel, G. and Schwarzbaum, P.J. (2004)Effects of extracellular nucleotides and their hydrolysis productson volume regulatory decrease of hepatocytes from trout(Oncorhynchus mykiss). Am. J. Physiol. 287, R833–R843

Park, M., Ko, S.B., Choi, J.Y., Muallem, G., Thomas, P.J.,Pushkin, A., Lee, M.S., Kim, J.Y., Lee, M.G., Muallem, S. et al.(2002) The cystic fibrosis transmembrane conductance regulatorinteracts with and regulates the activity of the HCO3

− salvagetransporter human Na+-HCO3

− cotransport isoform 3.J. Biol. Chem. 277, 50503–50509

Raat, N.J., De Smet, P., van Driessche, W., Bindels, R.J. andVan Os, C.H. (1996) Measuring volume perturbation of proximaltubular cells in primary culture with three different techniques.Am. J. Physiol. 271, C235–C241

Schwiebert, E.M. (1999) ABC transporter-facilitated ATP conductivetransport. Am. J. Physiol. 276, C1–C8

Schwiebert, E.M., Mills, J.W. and Stanton, B.A. (1994) Actin-basedcytoskeleton regulates a chloride channel and cell volume in a renalcortical collecting duct cell line. J. Biol. Chem. 269, 7081–7089

Shi, H., Levy-Holzman, R., Cluzeaud, F., Farman, N. and Garty, H.(2001) Membrane topology and immunolocalization of CHIF inkidney and intestine. Am. J. Physiol. 280, F505–F512

Strange, K. (1988) RVD in principal and intercalated cells of rabbitcortical collecting tubule. Am. J. Physiol. 255, C612–C621

Strange, K., Emma, F. and Jackson, P.S. (1996) Cellular andmolecular physiology of volume-sensitive anion channels.Am. J. Physiol. 270, C711–C730

Valenti, G., Frigeri, A., Ronco, P.M., D’Ettorre, C. and Svelto, M.(1996) Expression and functional analysis of water channels in astably AQP2-transfected human collecting duct cell line.J. Biol. Chem. 271, 24365–24370

Verkman, A.S. (2000) Water permeability measurement in living cellsand complex tissues. J. Membr. Biol. 173, 73–87

Waldegger, S., Steuer, S., Risler, T., Heidland, A., Capasso, G.,Massry, S. and Lang, F. (1998) Mechanisms and clinicalsignificance of cell volume regulation. Nephrol. Dial. Transplant. 13,867–874

Wehner, F., Sauer, H. and Kinne, R.K. (1995) Hypertonic stressincreases the Na+ conductance of rat hepatocytes in primaryculture. J. Gen. Physiol. 105, 507–535

Wehner, F., Olsen, H., Tinel, H., Kinne-Saffran, E. and Kinne, R.K.(2003) Cell volume regulation: osmolytes, osmolyte transport,and signal transduction. Rev. Physiol. Biochem. Pharmacol. 148,1–80

Zelenina, M. and Brismar, H. (2000) Osmotic water permeabilitymeasurements using confocal laser scanning microscopy.Eur. Biophys. J. 29, 165–171

Received 31 August 2004/17 November 2004; accepted 17 November 2004

Published as Immediate Publication 29 April 2005, doi:10.1042/BC20040116


Volume regulation in cortical collecting duct cells: role of AQP2

Documents

Transcript of Volume regulation in cortical collecting duct cells: role of AQP2