In trout hepatocytes, hypotonic swelling is followed by a compensatory shrinkage called regulatory volume decrease (RVD). It has been postulated that extracellular ATP and other nucleotides may interact with type 2 receptors (P2) to modulate this response. In addition, specific ectoenzymes hydrolyze ATP sequentially down to adenosine, which may bind to type 1 receptors (P1) and also influence RVD. Accordingly, in this study, we assessed the role of extracellular nucleoside 5′-tri- and diphosphates and of adenosine on RVD of trout hepatocytes. The extent of RVD after 40 min of maximum swelling was denoted as RVD40, whereas the initial rate of RVD was called vRVD. In the presence of hypotonic medium (60% of isotonic), hepatocytes swelled 1.6 times followed by vRVD of 1.7 min−1 and RVD40 of 60.2%. ATP, UTP, UDP, or ATPγS (P2 agonists; 5 μM) increased vRVD 1.5–2 times, whereas no changes were observed in the values of RVD40. Addition of 100 μM suramin or cibacron blue (P2 antagonists) to the hypotonic medium produced no effect on vRVD but a 53–58% inhibition of RVD40. Incubation of hepatocytes in the presence of either 5 μM [γ-32P]ATP or [α-32P]ATP induced the extracellular release of [γ-32P]Pi (0.21 nmol·10−6 cells−1·min−1) and [α-32P]Pi (∼8 × 10−3 nmol·10−6 cells−1·min−1), suggesting the presence of ectoenzymes capable of fully dephosphorylating ATP. Concerning the effect of P1 activation on RVD, 5 μM adenosine, both in the presence and absence of 100 μM S-(4-nitrobenzil)-6-tioinosine (a blocker of adenosine uptake), decreased RVD40 by 37–44%, whereas 8-phenyl theophylline, a P1 antagonist, increased RVD40 by 15%. Overall, results indicate that ATP, UTP, and UDP, acting via P2, are important factors promoting RVD of trout hepatocytes, whereas adenosine binding to P1 inhibits this process.
- ectonucleoside triphosphate diphosphohydrolases
- extracellular adenosine 5′-triphosphate
over the past few years, several studies reported that many types of animal cells secrete ATP, as well as other nucleotides, to the extracellular space, where these molecules can accumulate at nano- to micromolar concentrations and influence various biological processes (10, 34).
Conditions known to elicit such an efflux of ATP from the cells include hypoxia, acidosis, mechanical deformation, hyposmotic shock, receptor stimulation, fluid shear stress, and membrane depolarization (see Ref. 4). In isolated hepatocytes, e.g., mechanical stimulation of single cells was found to evoke a release of a soluble factor (presumably ATP) that caused an unexpected Ca2+ rise in hepatocytes and bile duct epithelia having no direct contact to the stimulated cell (29). The involvement of extracellular ATP (ATPe) in mediating this effect was inferred from the observation that it could be blocked by the addition of apyrase, an ATP diphosphohydrolase (11).
Because of its size, charge, and concentration gradient, ATPe is unlikely to be taken up by the cells, and such cellular responses are generally believed to be mediated by interaction of ATPe with various membrane proteins at the outer leaflet of the plasma membrane, in particular, by P receptors (including purinoceptors and pyrimidinoceptors) and E-NTPDases (ectonucleoside triphosphate diphosphohydrolases; see Ref. 38. Regarding the former, it was observed that, in the liver, ATPe can activate type 2 P receptors (P2) in hepatocytes, colangiocytes, macrophages, and endothelial cells (35). The ubiquitous presence of P2 thus strongly indicates that ATP is present in the extracellular space in at least some physiological conditions (12).
At the same time, however, little is known regarding the cellular functions of E-NTPDases. Members of the E-NTPDase family hydrolyze either nucleoside 5′-triphosphates or both nucleoside 5′-tri- and diphosphates. One postulated role for these enzymes concerns the decrease of the effective concentration of nucleotides at the extracellular surface of the cell, which may serve to terminate the action of these nucleotides on P2 (38). In the liver, such a regulation of P2 by E-NTPDases could be important in the context of volume regulation. Hepatocytes are constantly subjected to osmotic stress as a consequence of solute uptake and metabolism, and this, in turn, may certainly result in changes of cell size (13). Moreover, various pathologies associated with ischemia, hypothermia, hyponatremia (22), and severe metabolic inhibition (3) can lead to an increase of intracellular osmolarity and cell swelling. In previous studies using rat hepatoma cells, a model has been proposed where ATPe, acting on P2, was suggested to modulate volume regulation of hyposmotically exposed cells. According to this model, cell swelling would initiate a flux of ATP to the extracellular medium (28), where it would then stimulate P2 and thereby activate the efflux of Cl− or KCl. This efflux is accompanied by an osmotic loss of water and would ultimately lead to a recovery of cell volume, a phenomenon generally referred to as regulatory volume decrease or RVD (16). This model is supported by the observation that swelling induced a loss of ATP from various epithelial cells (see Ref. 15) and perfused rat liver (see Ref. 28) and that removal of ATPe or blockage of P2 in hepatic cells inhibited both Cl− efflux and RVD. Despite the experimental evidence supporting this model, the effect of ATPe on RVD of vertebrate hepatocytes has not yet been studied in great detail, and particularly the role of the products of ATP hydrolysis in RVD has received only little attention (26). This, however, appears important inasmuch as one or more E-NTPDases, acting in coordination with E-5′-nucleotidases, are capable of fully dephosphorylating ATPe in physiological conditions (31). The resulting adenosine may interact with P1 receptors and thereby also influence RVD by yet unknown mechanisms.
Furthermore, considering that ATPe may be rapidly transformed into UTP by ecto-nucleoside diphosphokinases (23), the action of this latter nucleotide on RVD appears to also deserve consideration.
In the present study, we therefore aimed at addressing this topic in a more comprehensive manner. For this purpose, we used trout hepatocytes as a cell model, the anisotonic responses of which have been amply documented by us (5–8, 20, 25). Employing this rather well-defined cell model, we studied the role of ATPe on volume regulatory processes after swelling, and we put special emphasis on the interaction between ATPe and its hydrolysis products with P1 and P2 receptors and E-NTPDases. In addition, we evaluated the role of ATPe in goldfish hepatocytes, since in these cell, in contrast to trout hepatocytes, we could previously not detect regulatory volume changes under anisotonic conditions (7).
MATERIALS AND METHODS
Collagenase (type IV), cibacron blue 3GA, adenosine, S-(4-nitrobenzil)-6-tioinosine (NBTI), suramin, poly-l-lysine, ouabain, apyrase (grade III), ATP, adenosine 5′-O-(1-thiotriphosphate) (ATPγS), UTP, ADP, UDP, α,β-methyleneadenosine 5′-triphosphate, and 8-phenyl theophylline (8-PT) were purchased from Sigma (St. Louis, MO). 2′,7′-Bis-(2-carboxyethyl)-5-(and-6)-carboxifluorescein (BCECF-AM) was obtained from Molecular Probes (Eugene, OR). [γ-32P]ATP (5.4 Ci/mg, ∼10 mCi/ml) and [α-32P]ATP (1.4 Ci/mg, ∼10 mCi/ml) were from NEN Life Science Products (Boston, MA). All other reagents were of analytical grade.
Na-K-ATPase was kindly provided by Dr. J. G. Nørby (Institute of Biophysics, University of Århus, Denmark). It was purified from pig kidney external medulla, as described by Jensen et al. (17). The specific activity of this enzyme measured under optimal conditions was 8 U/mg. In some experiments, the Na-K -ATPase was inactivated by heating the enzyme for 30 min at 65°C.
Rainbow trout Oncorhynchus mykiss (150–250 g; Walbaum) were obtained from the Center of Aquaculture from the Universidad del Comahue (Bariloche, Argentina) and were maintained in 200-liter tanks at 15°C. Goldfish Carassius auratus L. (10–30 g) were obtained commercially from local dealers and kept in 200-liter tanks at 20°C. Fish were acclimated to the above specified temperatures for at least 2 wk before being used.
Isolation of Hepatocytes
Fish were killed by a blow on the head and transection of the spinal cord. Hepatocytes were isolated by collagenase digestion methods, as described previously (19, 21, 33). After isolation, the cells were incubated in isotonic medium (see below) for 45 min in a shaking water bath at 20°C before use. The viability of isolated hepatocytes (>90%) was routinely assessed by Trypan blue exclusion (before the onset of each experiment) and retention of BCECF fluorescence (at the end of each experiment).
Except where otherwise stated, cells from trout and goldfish were incubated in media at 20°C (pH 7.6) having the following composition. Media for trout hepatocytes (in mM) were as follows: C (isotonic medium), 10 HEPES, 136.9 NaCl, 5.4 KCl, 1.5 CaCl2, 0.33 Na2HPO4, 0.44 KH2PO4, 1 MgSO4, 5 NaHCO3, and 5 glucose, osmolarity 292 mosmol/l; D, medium C without NaCl, osmolarity 53 mosmol/l; E, medium C plus 400 mM sucrose, osmolarity 672 mosmol/l; A, (hypotonic calibration medium) a 4:1 mixture of media C and D, osmolarity 238 mosmol/l; B, (hypertonic calibration medium) a 4:1 mixture of media C and E, osmolarity 334 mosmol/l. Media for goldfish hepatocytes (in mM) was as follows: CG (isotonic medium), 10 HEPES, 135.2 NaCl, 3.8 KCl, 1.3 CaCl2, 1.2 KH2PO4, 1.2 MgSO4, 10 NaHCO3, osmolarity 286 mosmol/l; DG, medium CG without NaCl, osmolarity 52.5 mosmol/l; EG, medium CG plus 400 mM sucrose, osmolarity 660 mosmol/l; AG, (hypotonic calibration medium) an 4:1 mixture of media CG and DG, osmolarity 250 mosmol/l; BG, (hypertonic calibration medium) an 4:1 mixture of media CG and EG, osmolarity 320 mosmol/l. Hypotonic media (denoted as HYPO for trout, and HYPOG for goldfish hepatocytes) were prepared by mixing one volume of isotonic medium with an equal volume of D (trout) or DG (goldfish), yielding an osmolarity of 160–170 mosmol/l, corresponding to 54–58% isotonic media.
For measurements of relative cell volume (Vr) in Cl−-free medium, cells were isolated using media similar to C where the salts NaCl, CaCl2, and KCl were replaced by equimolar concentrations of Na(NO3), Ca(NO3)2, and K(NO3), respectively. The same replacement was performed to get Cl−-free media similar to A, B, and HYPO.
Hepatocytes attached to poly-l-lysine-coated glass coverslips were mounted on a perfusion chamber filled with isotonic medium C and placed on the stage of a Nikon TE-200 epifluorescence inverted microscope. Cells were then loaded with 4 μM BCECF-AM. Dye loading was monitored fluorometrically by sampling the signal of single cells every 60 s until fluorescence of the cells reached ∼5–10 times the autofluorescence level. Subsequently, the solution was washed out with medium C for at least 1 h before starting the experiment, and chamber perfusion was stopped. During experimental manipulations, all media were removed from or introduced in the chamber manually. Changes in cell water volume were inferred from readings of the fluorescence intensity recorded by exciting BCECF at 440 nm, where the fluorochrome is pH insensitive (the isosbestic point; see Ref. 2). Under these conditions, changes in fluorescence intensity recorded from a small region of dye-loaded cells reflect changes in intracellular fluorophore concentration and, therefore, alterations of cell water volume. Fluorescence images were acquired by use of a charge-coupled device camera (Hamamatsu C4742–95) connected to a computer and the Metafluor acquisition program (Universal Imaging).
Values of Vr change were then computed from monitored changes in relative fluorescence (Ft/Fo), with Fo representing the signal obtained from a pinhole region of the cell equilibrated with isotonic medium, and Ft denoting the fluorescence of the same region of the cell at time = t. Thus Vr represents a fractional volume, where volume before treatments is one, and volume changes are related to this value. Preliminary comparative experiments with the volume marker calcein yielded similar results to those obtained with BCECF, in line with reports from others (37a). A detailed description of the technique can be found elsewhere (1, 2).
The RVD associated with the volumetric response of cells exposed to hypotonic medium was calculated by use of the following equation (adapted from Ref. 1): (1) where Vr max is the maximal value of Vr attained during hypotonic swelling, and Vrt represents the value of Vr observed at different times after reaching Vr max. RVD thus denotes the magnitude of volume regulation, with 100% RVD indicating complete volume regulation and 0% RVD indicating no volume regulation. To compare the effects of the different experimental conditions employed, the following two parameters were calculated to characterize RVD. On the one hand, we calculated the extent of RVD observed 40 min after Vr max to cover the whole time frame of the regulatory volume response studied. This parameter was denoted as RVD40. On the other hand, the initial rate of RVD or vRVD was calculated, and this was done by fitting data of RVD vs. time by the second-order polynomial A0 + vRVD t + Ct2, where t is time, A0 is value of RVD at t = 0, and C is coefficient. The first derivative of this function at t = 0 corresponds to vRVD. The RVD40 covers the time frame for the whole RVD response, whereas the vRVD provides information on regulatory volume changes that occur acutely.
For all treatments examined, hepatocytes were exposed to hypotonic medium to induce an increase in cell volume, and the subsequent volume changes were observed. Before the hypotonic challenge, the cells were briefly exposed to isotonic medium (C; 292 mosmol/l) and to moderately anisotonic media (A = 253 mosmol/l; B = 328 mosmol/l). These treatments served to calibrate experiments and allowed transformation of the fluorometric signal to values of Vr (see above). This calibration procedure was identical in all experiments investigating Vr changes vs. time.
Effects of Different Treatments on RVD
Under the assumption that swollen cells may release ATP to the extracellular medium, and that this ATP could be converted to other nucleotides, Na-K-ATPase (3 U/ml, suspended in medium C or HYPO) and apyrase (3 U/ml, dissolved in medium C or HYPO) were used as scavengers of extracellular nucleotides by accelerating the reactions from ATP to ADP + Pi (Na-K-ATPase) and from NTP to NMP + 2 Pi (apyrase), where NTP, NDP, and NMP stand for nucleoside tri-, di-, and monophosphate, respectively.
Because Na-K-ATPase can only utilize ATP as a substrate, an effect of this enzyme will point to the presence of ATPe affecting RVD. Alternatively, the enzyme apyrase was used in parallel experiments. Although this enzyme is not as specific as Na-K-ATPase for identifying ATP, it has nevertheless proven very useful for removing tri- and diphosphonucleotides (nucleosides di- and triphosphate of adenosine, guanosine, and uridine) from the extracellular medium (14).
Before their use in the experiments, the hydrolytic capacity of Na-K-ATPase and apyrase was checked by measuring their associated ATPase activity. These measurements were conducted at 20°C by following the release of [γ-32P]Pi from [γ-32P]ATP, as described previously (31, 32).
Regarding the actual concentration of ATP in the intercellular space of trout hepatocytes, there is no information available in the literature. However, because it has been documented that submicromolar concentrations of the nucleotide may accumulate in the extracellular medium of several other cell types (34), we tested the activity of Na-K-ATPase and apyrase at 0.5 and 5 μM ATP.
Agonists and antagonists of P1 and P2.
The involvement of P1 and P2 receptors as modulators of RVD was evaluated by using agonists and antagonists of these receptors. Agonists of P2 included 5 μM of either ATP, ADP, ATPγS, UTP, or UDP and 100 μM α,β-methylene-ATP (a P2X agonist in mammalian systems), whereas adenosine was used as a P1 agonist. As an inhibitor of adenosine uptake, we employed 100 μM NBTI.
As antagonists of P2 either 100 μM suramin or 100 μM cibacron blue 3GA were employed, whereas 100 μM 8-PT was used as a P1 antagonist. The following stock solutions were prepared: 1 mM ATP, ADP, ATPγS, UTP, or UDP, all neutralized with imidazole, 1 mM adenosine in DMSO, 20 mM suramin in bidistilled water, and 20 mM cibacron blue 3GA in DMSO. The 8-PT was directly dissolved in the assay medium. NBTI was taken from a 20 mM stock solution in DMSO.
In trout hepatocytes exposed to hypotonic media, part of the RVD is mediated by the loss of Cl− and K+ from the cells, and it has been reported that Cl− cannot be replaced by NO3− as a mediator of RVD (5). Therefore, to establish whether ATPe exerts its effect on RVD via a hypothetical interaction with the KCl leakage pathway, one series of experiments was conducted with hepatocytes isolated and subsequently exposed to Cl−-free media, both during isotonic and hypotonic conditions.
The DMSO used to dissolve some of the reagents applied yielded a final concentration in the cell suspension of 0.5% (vol/vol). Although this caused an increase in the osmolarity of the media used, preliminary experiments showed that this did not affect Vr of hepatocytes over 45 min. The lack of osmotic effect of DMSO on Vr can be explained by assuming that this reagent equilibrates fast within the intra- and extracellular compartments.
E-ATPase Activity and Adenosine Production
Because ATP added to a hepatocyte suspension is impermeable to hepatocytes, any hydrolysis of ATP into ADP + γ-Pi in the cell suspension can be defined as E-ATPase activity (31) and can be assigned to the one or more membrane proteins of the E-NTPDase superfamily (38). Thus we used the ATPase assay described above as a measure of the rate at which isolated hepatocytes hydrolyze ATPe (9, 31).
Production of adenosine from ATP was estimated at 20°C by following the release of [α-32P]Pi from [α-32P]ATP using a method similar to that described for measuring E-ATPase activity. Under these conditions, one adenosine is formed for every [α-32P]Pi produced.
The rates of [α-32P]Pi and [γ-32P]Pi release were measured both in isotonic and in hypotonic conditions. For this purpose, cells (106 cells/ml) were diluted 1:1 in medium C (isotonic condition) or medium D (hypotonic condition) under continuous stirring, and, at 0, 2, 20, and 40 min, aliquots were withdrawn to assess the rates of release of [γ-32P]Pi (from [γ-32P]ATP) and [α-32P]Pi (from [α-32P]ATP), as described previously.
ATPase activity was estimated by following the time course of [γ-32P]Pi release from [γ-32P]ATP. Next, Eq. 2 was fitted to experimental data. (2) where Y and Y0 are the values of Pi at any time (t) and that at t = 0, respectively; A represents the maximal value for the increase of Y with time, and k is a rate coefficient. The parameters of best fit resulting from the regression were used to calculate Na-K-ATPase activity and the half-time of ATP hydrolysis rate (t1/2).
Activity of Na-K-ATPase was calculated as k·A, i.e., that of the first derivative of Eq. 2 for t tending to zero. To calculate the rate of [α-32P]Pi release from [α-32P]ATP, linear functions were fitted to the experimental data.
The value of t1/2 was calculated as: (3) In experiments of Fig. 10B, intact hepatocytes were used to assess the effect of exogenous ATPe (ranging from 0.5 to 1,000 μM) on E-ATPase activity. Next, Eq. 4 was fitted to experimental data (4) where vmax represents the apparent maximal E-ATPase activity and K1/2 the substrate concentration at which a half-maximal hydrolysis rate is obtained under the specific conditions of the experiment.
Results of ATPase activity were expressed as nanomoles Pi·10−6 cells·min−1 or as a percentage (%) of total hydrolyzed ATP.
The absence or presence of RVD was evaluated by determining whether the slope of Vr vs. time was significantly different from zero (P < 0.05), using a Pearson Correlation test.
The effect of the different treatments on RVD was evaluated by means of one-way ANOVA followed by a Tukey-Kramer test of multiple comparisons. P ≤ 0.05 was considered significant. For all experiments of Vr vs. time, 55–60 cells from four to five independent preparations were used.
Volumetric Response Under Standard Conditions
In a first series of experiments, trout hepatocytes were exposed to hypotonic medium (HYPO; medium 160–170 mosmol/l) under standard conditions, i.e., in the absence of agents potentially promoting or inhibiting the action of ectoenzymes and/or nucleoside receptors (Fig. 1). The RVD was assessed both by the initial rate of RVD change (denoted as vRVD) and by the extent of RVD at 40 min, i.e., RVD40 (see materials and methods). Hypotonic exposure induced cell swelling to a Vr max value of 1.52 ± 0.04, followed by an RVD with vRVD = 1.71 ± 0.36 min−1 and RVD40 = 60.18 ± 3.23%. These values served as the control response to which the treatments to be described in the following were compared.
Effect of Nucleotide Scavengers
As outlined in the Introduction, cell swelling evoked by hypotonic exposure may cause the release of ATP from hepatocytes. However, in the presence of sufficient amounts of either Na-K-ATPase or apyrase, the accumulation of the nucleotide in the extracellular space may be prevented. Thus, in the next series of experiments, the effect of these nucleotide scavengers on hypotonically induced volume changes was evaluated. Before the use of these enzymes in volumetric experiments, we determined their hydrolytic capacity in cell-free media.
As depicted in Fig. 2, under these conditions, Na-K-ATPase (3 U/ml) activity assayed with 0.5 and 5 μM ATP amounted to 5.7 ± 2.1 and 54.5 ± 4.4 nmol·mg−1·min−1, respectively. The t1/2 values for ATP hydrolysis determined in these assays amounted to 0.26 ± 0.17 min (0.5 μM ATP) and 0.92 ± 0.31 min (5 μM ATP). ATPase activity of Na-K-ATPase determined with 5 μM ATP was not significantly affected by the presence of 1 mM UTP. In contrast, the inactivation of Na-K-ATPase by heat caused the enzyme to lose nearly all ATPase activity.
In the case of apyrase (3 U/ml) with 0.5 μM ATP present, ATPase activity and t1/2 amounted to 1.7 ± 1.7 μmol·mg−1·min−1 and 0.0043 ± 0.0019 min, respectively, whereas, for 5 μM ATP, ATPase activity was 22.0 ± 10.4 nmol·mg−1·min−1 and t1/2 was 0.035 ± 0.028 min.
The effect of these nucleotide scavengers on hypotonically induced volume changes are summarized in Fig. 3. As can be seen in Fig. 3A, none of the added enzymes altered the initial change of Vr in the presence of hypotonic medium, which reached a maximum of 1.51–1.55. In contrast, both nucleotide scavengers significantly affected volume regulatory responses of the cells (Fig. 3B). In the presence of Na-K-ATPase, vRVD was diminished by 95% and RVD40 decreased by 78%. RVD measured in the presence of heat-inactivated Na-K-ATPase was not different from control values. In the presence of apyrase, vRVD and RVD40 decreased 52 and 60%, respectively.
RVD in the Presence of P2 Agonists and Antagonists
A series of experiments conducted under isotonic conditions indicated that 5 μM ATP had no effect on steady-state volume of hepatocytes, which was maintained at 1.004 ± 0.002 (48 cells from 4 independent preparations) during 45 min of incubation with the nucleotide. In further experiments, trout hepatocytes were incubated in hypotonic medium containing either 5 μM ATP, ATPγS (a nonhydrolyzable analog of ATP), UTP, ADP, or UDP. In addition, cells were also exposed to hypotonic medium, including 100 μM α,β-methylene-ATP. It was found that none of the P2 agonists affected RVD40 (Figs. 4 and 5) and that neither ADP nor α,β-methylene-ATP had any effect on the values of vRVD compared with hypotonic controls (Fig. 4).
In contrast, both ATP and ATPγS enhanced vRVD 1.8–2 times with respect to control values, with the extent of stimulation of vRVD being similar in the presence of either of these nucleotides. Similarly, UDP and UTP increased vRVD 1.5–1.8 times (Fig. 5). The observed increase in vRVD with UTP could have been the result of the extracellular conversion of this nucleotide into ATP and the subsequent action of ATP on RVD. This is why, in separate experiments, we exposed the cells to HYPO + Na-K-ATPase (to remove all ATP present in the extracellular medium) and 0 or 5 μM UTP. Under these conditions, we found that, compared with hypotonically exposed cells with Na-K-ATPase only, RVD for HYPO + Na-K-ATPase was 12-fold (vRVD) and 2.8-fold (RVD40) higher in the presence than in the absence of UTP (Fig. 5).
In Fig. 6A, the time course of Vr changes of trout hepatocytes exposed to HYPO in the presence and absence of P2 antagonists is shown. Cells were exposed to 100 μM cibacron blue and to 100 μM suramin in the presence and absence of 5 μM ATP. Compared with the controls (HYPO), none of the treatments produced any effect on vRVD. At the same time, however, both cibacron blue and suramin caused a significant decrease of RVD40 by 53–58%, the latter effect persisting even in the presence of 5 μM ATP (Fig. 6B).
Effect of Extracellular Adenosine and 8-PT on RVD
The hypothetical involvement of P1 receptors in the volume regulatory response of trout hepatocytes was assessed by exposing cells to hypotonic saline containing 5 μM adenosine, both in the presence and absence of 100 μM NBTI (a blocker of the uptake of adenosine). Results showed that both adenosine and adenosine + NBTI decreased RVD40 by 37–44% but exerted no significant effects on vRVD (Fig. 7). Values of RVD with NBTI only were not significantly different from controls.
The effect of 100 μM 8-PT (a P1 antagonist) on RVD was evaluated in the presence and absence of 5 μM ATP (Fig. 8). Inclusion of the antagonist alone in hypotonic medium produced no significant effect on vRVD but caused a 22% increase of RVD40 compared with control conditions. When administered in concert with ATP, 8-PT evoked a 60% increase of vRVD and, noteworthy, a 22% increase of RVD40, an effect not seen with ATP alone.
Impact of Cl−-free Conditions on RVD
For these measurements, trout hepatocytes were isolated and preincubated in media where Cl− had been replaced by NO3− (Fig. 9). This procedure allowed removal of Cl− from the intra- and extracellular compartments, thereby creating a condition referred to as “Cl− free.” As can be seen in Fig. 9A, in isotonic Cl−-free medium, Vr remained constant for 30 min at 1.003 ± 0.003. During subsequent exposure to hypotonic Cl−-free conditions, vRVD was not significantly altered compared with controls (HYPO), irrespective of the absence or presence of 5 μM ATP (Fig. 9B). In contrast, Cl−-free conditions produced a significant 61–69% decrease of RVD40, this effect being independent from added ATP as well.
E-ATPase Activity and Adenosine Production
In light of the suggested role of extracellular nucleotides in the volume regulatory response of trout hepatocytes, this series of experiments investigated the capacity of the cells to hydrolyze ATPe. To this end, hepatocytes (0.25–3 × 106 cells/ml) were incubated in the presence of 5 μM of either [γ-32P]ATP or [α-32P]ATP (Fig. 10A). We observed that [γ-32P]Pi accumulation increased nonlinearly, with the initial rate of [γ-32P]Pi release (a measure of E-ATPase activity) amounting to 0.21 ± 0.02 nmol·10−6 cells·min−1 (n = 5 independent preparations). Using 0.5 × 106 cells, the t1/2 for ATP hydrolysis in ADP + γ-Pi was 10.1 ± 1.7 min (Fig. 10A).
On the other hand, the rate of [α-32P]Pi release (a measure of adenosine production) could be described by the sum of two linear functions, with the rate amounting to 7.1 ± 0.4 × 10−3 nmol·10−6 cells·min−1 during the first 20 min of incubation and to 9.1 ± 0.8 × 10−3 nmol·10−6 cells·min−1 for the residual 70 min investigated. In the absence of cells, the rates of [γ-32P]Pi and [α-32P]Pi release were negligible. In separate experiments (data not shown), cells (0.5 × 106 cells/ml) were incubated in hypotonic medium for 0, 2, 20, and 40 min with 5 μM of either [γ-32P]ATP or [α-32P]ATP. Under these conditions, hypotonic exposure did not significantly affect the release of [γ-32P]Pi or [α-32P]Pi.
In Fig. 10B, a substrate curve for E-ATPase activity is shown. Enzyme activity vs. ATP (0.5–1,000 μM) could be described by a single hyperbole with K1/2 = 246 ± 105 μM.
Effect of Exogenous ATPe on RVD of Goldfish Hepatocytes
In the last set of experiments, the impact of added ATPe on RVD of hypotonically challenged goldfish hepatocytes was investigated. As depicted in Fig. 11A, in isotonic medium (296 mosmol/l), goldfish cells maintained a constant Vr (1.00 ± 0.02, n = 4). Exposure of the cells to hypotonic medium (HYPOG) caused an acute increase of Vr to a maximum at 1.69 ± 0.02, followed by an only slight decrease of cell volume, which yielded no significant RVD (P = 0.15). However, when goldfish hepatocytes were exposed to HYPOG in the presence of 5 μM ATP, a vRVD of 0.9 ± 0.1 min−1 and RVD40 of 28.5 ± 3.2% were detected, both parameters being significantly different from the values determined in the absence of ATP.
In accordance with previous observations, when trout hepatocytes were subjected to hypotonic conditions, they first swelled and then underwent a secondary decrease of cell volume called RVD (Fig. 2). Results of the present study provide unequivocal evidence that extracellular nucleotides, released from the cells under these conditions, are an important mediator of this regulatory response. Thus addition of Na-K-ATPase and apyrase to cell suspensions in sufficient amounts to acutely remove all ATP liberated from the cells inhibited 52–95% of vRVD and 60–78% of RVD40 (Fig. 3B).
An investigation into the mechanistic basis of this phenomenon revealed that, in line with various other reports (24, 37), in trout hepatocytes the effect of ATP on RVD required the interaction of the nucleotide with P2 receptors. Although antagonists of this type of receptor did not affect vRVD, they inhibited RVD40 by 53–58% (Fig. 7B), whereas exogenously added ATP and ATPγS, two agonists of P2, enhanced vRVD by 80–100% (Fig. 5). Moreover, in the presence of suramin, a P2 antagonist, the presence of ATPe did not alter RVD, supporting the hypothesis that volumetric regulation of hypotonically challenged cells requires the interaction of ATPe with P2. Surprisingly, we observed that the effects of agonists and antagonists on P2 on vRVD and RVD40 were not fully consistent, but we consider an interpretation of this finding too speculative at present.
In mammalian systems, P2 receptors have been grouped in two families: G protein-coupled receptors termed P2Y and intrinsic ion channels termed P2X receptors (27). In trout hepatocytes, the specific P2 subtype responsible for modulating RVD seems to be P2Y, since, as shown in Fig. 5, the P2X agonist α,β-methylene-ATP caused no changes in RVD, whereas ATP and its nonhydrolyzable analog ATPγS significantly enhanced this parameter. In addition, we found that added ADP did not enhance RVD of the trout cells, which agrees with the comparatively low potency of ADP to stimulate most members of this receptor subtype family (36).
Besides a direct interaction of ATPe with membrane receptors, the nucleotide could also affect RVD through the generation of hydrolysis byproducts by the activity of ectoenzymes. During incubation of cell suspensions with [α-32P]ATP, the liberation of [α-32P]Pi may be expected only if such ectoenzymes were present that sequentially liberate γ-, β -, and α-Pi of ATP at the cell surface. This is indeed what we observed in our experiments, where after addition of [α-32P]ATP a release of [α-32P]Pi was seen (for every [α-32P]Pi one adenosine is liberated), providing evidence that ATP is fully dephosphorylated in the extracellular medium (Fig. 10A).
The adenosine produced in this process may then act on P1 receptors and thereby inhibit RVD. This conclusion is supported by the observation that addition of exogenous adenosine caused a reduction of RVD40 that remained unaltered after blocking the uptake of nucleosides with NBTI (Fig. 8). Moreover, the P1 antagonist 8-PT, in the absence of exogenous adenosine, caused a 15–22% increase of RVD40. To our knowledge, this is the first report in cells of extracellular adenosine inhibiting RVD via P1. Taken together, the above-described results suggest that adenosine appears to have accumulated in sufficient amounts to activate P1 receptors. This may indicate that, despite the observed low rate of adenosine production from ATP (Fig. 1), sufficiently high local concentrations might have been generated in close proximity of the cell surface. This may be particularly true for the situation found in the intact liver, where, according to a rough estimate, a concentration of hepatocytes equivalent to ∼7 × 108 cells/ml is to be expected, that is, a cell density ∼2 × 105 times higher than the one used in the assay chamber.
A further important aspect of the present study relates to the finding that not only ATPe, but also UTP, affected RVD of trout hepatocytes (Fig. 5). The stimulation of vRVD by UTP could have been because of the direct action of UTP on P2Y (36) or by means of its hydrolysis product UDP. Alternatively, an extracellular nucleotide disphosphokinase known to be present in various cell types (23) could have transferred the γ-Pi of UTP to a molecule of ADP to generate ATPe, which, in turn, could have modulated RVD. However, incubation of trout hepatocytes in the presence of UTP and an excess of exogenous Na-K-ATPase so as to hydrolyze all the ATP potentially formed in this way confirmed an effect of UTP as such, since both vRVD and RVD40 of hypotonically challenged trout hepatocytes were still significantly enhanced by UTP (Fig. 5). The fact that UDP was also capable of increasing RVD might point to the coexpression, in trout hepatocytes, of several subtypes of P2Y receptors, since at least one of these is highly specific for both UTP and UDP (36a).
Effects of ATP on RVD of Goldfish Hepatocytes
As outlined in the Introduction, we have recently found that, although goldfish hepatocytes show substantial cell swelling in hypotonic medium, they nevertheless show no or only a slight RVD (7). On the other hand, we have previously also obtained indirect evidence that these cells possess P2Y receptors (31). Therefore, we reasoned, goldfish hepatocytes represent an interesting model to check whether ATPe, even in the absence of spontaneous RVD, can still elicit volume regulation via P2Y activation.
As shown in Fig. 11, exposure of cells to hypotonic medium induced a slight RVD that, however, was not statistically significant. In contrast, upon addition of micromolar concentrations of ATP, a significant RVD was triggered, clearly suggestive of a hypothetical role of ATPe also in goldfish hepatocytes.
There may be several reasons underlying the lack of RVD of goldfish hepatocytes under standard conditions (i.e., in the absence of exogenous ATP). These may include an inability of these cells to release ATP to the extracellular space upon hypotonic challenge or a very high ectoenzyme activity of goldfish hepatocytes, degrading any released nucleotides so rapidly that the effective concentration at the cell surface is not enough to active P2Y receptors. This, however, seems unlikely, as E-ATPase activity of goldfish hepatocytes was found to be lower, and not higher, than those of trout cells (31). So third, since it is known that ATPe can also act in a paracrine way, the RVD of these cells could, under conditions prevailing in the intact liver, be activated by nucleotides liberated from other cells types like endothelial cells or macrophages, cells that are not present in the isolated liver cell preparation used in our studies.
Together with the results from trout hepatocytes, our findings indicate that the basic model of volume regulation by ATPe is highly conserved among vertebrates.
In Fig. 12, we present a model compatible with the experimental results obtained in this study for trout hepatocytes. After exposure to hypotonic medium, trout hepatocytes swell rapidly, and this causes the release of ATP to the extracellular space via an as yet unidentified pathway(s). The resulting ATPe then interacts with P2 and thereby activates effector mechanisms mediating the loss of osmolytes along with osmotically obliged water and in consequence the recovery of cell volume. Our finding that the RVD response triggered by ATP is dependent on the presence of Cl− indicates that it involves the activation of well-described pathways of ion release, such as Cl− channels or KCl cotransporters, and our model does not postulate novel, as yet unidentified, effector mechanisms. UTP and UDP have an effect on RVD qualitatively similar to that of ATP. In addition to P2 activation, ATP may be hydrolyzed by E-NTPDases and 5′-nucleotidases, the activities of which are insensitive to swelling, and this hydrolysis both terminates the effect of ATP and, together with the slow diffusion of the nucleotide in the extracellular space, limits the action of ATP on distant cells. As can be seen in Fig. 10B, because of the relatively high K1/2 of E-ATPase activity, an increase in the concentration of ATPe, in the micromolar range, would lead to a concomitant increase in the activity of the ectoenzyme. Although the role of the intermediate ADP remains unclear at present, production of adenosine from ATP is apparently sufficient to act on P1, and this exerts an inhibitory effect on RVD. Regarding the generation of uridine from UTP, there are no known receptors for that nucleoside, so that in principle uridine, unlike adenosine, would be unable to affect volume via cell surface receptors.
Concerning physiological aspects, one has to consider that an euryhaline species like the trout has to adapt to a changing extracellular osmolarity, so that swelling constitutes a physiological disturbance requiring an effective volume regulatory mechanism (13, 20). Results of this study suggest that extracellular nucleotides are important factors enabling such volume regulation.
This work was supported by grants from Consejo Nacional de Investigaciones Cientificas y Técnicas (CONICET), F. Antorchas, Universidad de Buenos Aires, and Agencia Nacional de Promoción Cientifica y Tecnológica (no. 11017) of Argentina. G. Krumschnabel was supported by the Fonds zur Förderung der wissenschaftlichen Forschung in Österreich, project no. P16154-B06. R. M. González-Lebrero and P. J. Schwarzbaum are career researchers from CONICET.
Present address for P. Mut: Laboratorio de Biomembranas (Facultad de Medicina), Universidad de Buenos Aires, C1121ABG Buenos Aires, Argentina.
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. Section 1734 solely to indicate this fact.
- Copyright © 2004 the American Physiological Society