Effects of Hyposmotic Shock on Ion Fluxes in Isolated Trout Hepatocytes

When subjected to rapid osmotic challenges, cells of a variety of types display a regulatory response whereby the resting volume of the cells tends to be restored after an initial phase in which they behave as osmometers following the Boyle–Van’t Hoff relationship.

The magnitude of cell swelling after exposure to hyposmotic bathing solutions changes in accordance with the change in osmolarity. This response is followed by a volume recovery phase, RVD (regulatory volume decrease) (Kregenow, 1971, 1981), which occurs despite the continued presence of the hyposmotic medium. The time taken by cells to restore their initial volume varies from some minutes to several hours depending on the cells studied. Volume regulatory behaviour has been observed in a variety of cell types, including mammalian red blood cells (Ellory, Hall & Stewart, 1985), trout erythrocytes (La Marre, 1983; Bourne & Cossins, 1984), human lymphocytes (Grinstein, Rothstein, Sarkadi & Gelfand, 1984), Ehrlich ascites tumour cells (Hoffmann, Simonsen & Lambert, 1984), mammalian renal cells (Law, 1985) and Necturus gallbladder epithelial cells (Larson & Spring, 1984).

The purpose of the present work was to study the behaviour of isolated trout hepatocytes subjected to hypotonic shock. Rainbow trout are freshwater, euryhaline (salmonid) fish, which can be acclimated in sea water. Transfer from one medium to another results in transient haemodilution or haemoconcentration that induces significant internal osmotic shock affecting liver cells, among others.

As it appears that volume regulation by vertebrate cells is a consequence of osmotic activation of previously quiescent ion transport pathways (K⁺, Na⁺, Cl⁻), the present work examines the role of these ions in the volume-regulatory responses of trout hepatocytes in a hypotonic medium. Ion-exchange mechanisms in these cells, in isotonic medium, are the subject of a separate paper (B. Fossat, L. Bianchini, J. Porthé-Nibelle & B. Lahlou, in preparation).

Trout (Salmo gairdneri, average mass 200 g) were obtained from a local dealer, kept in running tap water at 12°C and fed with commercial fish pellets.

NaCl, 66 mmol l⁻¹; KCl, 66 mmol l⁻¹; Hepes, l0 mmol l⁻¹; glucose, 5·6 mmol l⁻¹. L mol l⁻¹NaOH was added to obtain pH 7·4.

Collagenase/Dispase (Boehringer) was dissolved in buffer I to 0·8 mg ml⁻¹ and the buffer was filtered on Millipore 0·45 μm filters before use.

H-MEM contained (in mmol l⁻¹): NaCl, 136·8; KCl, 5·4; CaCl₂, 1·3; KH₂PO₄, 0·44; MgSO₄, 0·81; Na₂HPO₄, 0·34; Hepes, 10; and (in mg l⁻¹): vitamins, 8·1; glucose, 1000; phenol red, 10; amino acids, 566; glutamine, 292. The pH was adjusted to 7·4 with lmol l⁻¹ NaOH.

In Cl⁻-free H-MEM, Ca(NO₃)₂ and K₂SO₄ were substituted, respectively, for CaCl₂ and KCl; NaCl was replaced by sodium gluconate, sodium methane sulphonate or NaNO₃.

Osmotic shock was induced by a one-third dilution of H-MEM with a 6 mmol l⁻¹KCl solution buffered with 10 mmol l⁻¹ Hepes, pH 7·4 (dilution solution).

Ouabain (strophantin G) and the calcium ionophore A 23187 were purchased from Boehringer. Quinine and TEA⁺ (tetraethylammonium) were purchased from Sigma. Apamine was a gift from the Centre de Biochimie (Nice, France). Bumetanide (Lixil-Leo) and furosemide (Lasilix) are pharmaceutical products from, respectively, Leo and Hoechst Laboratories. Piretanide was a gift from Hoechst Pharmaceuticals, Hounslow, England.

²²Na (33 GBq mg⁻¹) and ³⁶C1 (0·43 MBq mg⁻¹) were obtained from the Radiochemical Centre, Amersham, England; ⁸⁶Rb (0·04–0·4 GBq mg⁻¹) and ³H₂O (5·5 MBq ml⁻¹) from the Commissariat à l’Energie Atomique, Saclay, France; and [¹⁴C]dextran (0·1 MBq mg⁻¹) from the New England Nuclear (Boston, USA).

The procedure followed was the collagenase method previously described by Porthé-Nibelle & Lahlou (1981) and slightly modified. Briefly, the trout was injected intraperitoneally with 1 ml of heparin (5000 i.u. ml⁻¹) 10 min before it was killed by a blow on the head. The hepatic portal vein was cannulated after ventral incision and section of the suprahepatic vein and the liver was perfused in situ in open circuit at room temperature (approx. 21 °C) with perfusion buffer I delivered by a peristaltic pump at a flow rate of 12 ml min⁻¹ for 6–7 min. The liver was then removed and perfused for 60 min in closed circuit at the same temperature and at a flow rate of 9 ml min⁻¹ with perfusion buffer II containing collagenase. The disaggregated liver cell suspension was then filtered through three layers of gauze, washed three times (20 s centrifugation at low speed) and resuspended in H-MEM at a final concentration of approx. 17·5 × 10⁶ cells ml⁻¹, corresponding to 3·3 mg protein ml⁻¹, i.e. 5·3×l0⁶cellsmg⁻¹protein. The cellular suspension was incubated with gentle shaking for 40min, for equilibration following separation. All experiments were performed at 21 °C. Cell viability (approx. 85%) was assessed by Trypan Blue exclusion.

Hepatocytes were preincubated for 10 min in H-MEM with ³H₂O (0·55 MBq ml⁻¹) and [¹⁴C]dextran (0·05 MBq ml⁻¹), the latter to label the extracellular space. Hypotonic shock was induced by the addition of the cellular suspension to a suitable volume of a dilute solution of the same specific radioactivity, to obtain a final osmolality of 220 mosmol kg⁻¹ (0·7×isotonicity).

At different times, 200-μl samples were taken in duplicate and sedimented by a 5 s centrifugation at 10000 g. The radioactive supernatant was removed and 50 μl was counted to determine the ³H:¹⁴C ratio for each sample. The pellet was resuspended in 1 ml of distilled water and sonicated for 10 s with a Sonicator W10 cell disruptor (Heat System Ultrasonics). 800 μl was used for the radioactivity measurement and 50 μl for the protein assay. For each pellet, intracellular water was determined from tritium counting corrected for extracellular contamination calculated from the ^J4C count.

Ionic uptake experiments were initiated by adding 2 ml of the cellular suspension to 0-8 ml of H-MEM or diluted solution containing radioactive tracer, in the presence or absence of ouabain (final concentration 10⁻³mol l⁻¹). At various times, 200-μl samples were removed in duplicate and each added to 1 ml of ice-cold sucrose (330 mmol I⁻¹) in conical microfuge Eppendorf tubes to stop the uptake. The cells were sedimented by a 5-s centrifugation at 10000 g, resuspended in chilled sucrose medium and centrifuged again, once for ⁸⁶Rb or twice for ²²Na and ³⁶C1 measurements. Pellets were resuspended in 1 ml of distilled water for counting of radioactivity and/or determinations of cellular contents (Na⁺, K⁺, Cl⁻ and protein). A 50-μl sample was used to determine the external specific radioactivity (RAS). Fluxes were expressed as nequiv min^{− 1}mg^{− 1} protein.

Hepatocytes were loaded with radioactive isotope during the 40-min equilibration period, washed three times with H-MEM and resuspended in the same medium. At various times, 200-μl samples were collected in duplicate and treated as described above for uptake measurements. The amount of radioactivity remaining in cells was determined and expressed per mg protein. Efflux was calculated from the equation C_t = C_oe^{− kt}, where C_t and C_o are the intracellular radioactivities, respectively, at times t and 0, and k is the rate constant of the efflux. For first-order kinetics, ln(C_t/C₀) = −kt is a straight line, k is given by its slope and the biological half-time (T_1/2) by 0·693/k.

This procedure was used to determine at various times the instantaneous rate of ion uptake by using short-time incubations in three experimental conditions: H-MEM, ouabain H-MEM (10⁻³mol l⁻¹) and diluted ouabain H-MEM (0·7 × isotonicity). The experiment started with the addition of cells to the incubation medium. At timed intervals, 450μl of cellular suspension was added to 20μl of radioactive solution and gently agitated for 60s. 200-μl duplicate samples were then removed and immediately washed with ice-cold sucrose as described for uptake measurements.

To determine the initial rate of ion influx during the first minute of the experiment, 320μl of suspended cells was added to 150μl of each medium containing radioactive tracer.

The specific radioactivity (RAS) was determined for each time sample.

The experiments were repeated at least three times, the most representative being reported (Figs 1, 3–9). In Fig. 2 the mean and standard error for four determinations are given.

All isotopes were counted in an SL32 Intertechnique scintillation spectrometer, ⁸⁶Rb by the Cerenkov radiation in aqueous samples and ²²Na, ³⁶C1 and ³H after addition of scintillant (Aqua Luma from Kontron).

Ionic cellular contents were determined on individual samples after cell sonication of the counted sample for ⁸⁶Rb flux measurements or of a duplicate for ²²Na and ³⁶C1 flux measurements. Na⁺ and K⁺ were measured with an Eppendorf flame photometer and Cl⁻ by the Technicon colourimetric reaction (after protein precipitation by perchloric acid, 2% final concentration) with a DW2 doublebeam Aminco spectrocolorimeter.

Protein was assayed by the method of Lowry, Rosebrough, Farr & Randall (1951) adapted to a Technicon analyser, on a part of the counted sample, after sonication.

Osmolality of the solutions was evaluated by the freezing point method against NaCl standards on 50-μl samples, with a Knauer semimicro-osmometer.

In isotonic media, control cells maintained a constant intracellular water content of 3·10 ± 0·03 μl mg⁻¹ protein (N =15). When hepatocytes were exposed to the hypotonic medium there was an initial rapid increase in cell water content (Fig. 1). The maximal value (17–18% above the isotonic volume) was reached after 3 min. This was followed by a shrinking phase (RVD) and after 10-20 min the cell volume was not significantly different from control cells in isotonic media.

Diluting the medium enhanced by 5–7 times the ouabain-insensitive uptake of ⁸⁶Rb, a measure of K⁺ uptake (Fig. 2). However, the ouabain-sensitive influx was not modified by hypotonic shock (not shown in Fig. 2). Addition of isosmotic sucrose solution instead of 6 mmol l⁻¹ KCl medium to give an equivalent dilution of NaCl did not enhance the ouabain-insensitive uptake of ⁸⁶Rb. The observed stimulation of this flux was therefore a consequence of the osmotic pressure decrease.

Time course analysis of changes affecting the cellular ionic content indicates that after 20 min incubation in the presence of 1 mmol l⁻¹ ouabain, potassium fell to 80 ± 2% (N = 13) of its initial concentration in cells submitted to hypotonic shock compared with 93 ± 4 % (N = 6) in control cells maintained in isotonic medium. In

Fig. 3, which shows one experiment of this type, cell potassium level decreased further at 30min (66% in hypotonic vs 86% in control conditions).

In 10-min incubations, intracellular K⁺ concentration decreased as a function of the osmotic gradient (Fig. 4). Under the same conditions intracellular Na⁺ concentration was unchanged.

Fig. 5 shows that, under normal experimental conditions, the hypotonic shock induced an increase of ⁸⁶Rb efflux from loaded hepatocytes. Two phases may be observed: for 8–10 min the efflux rate constant was 0·0204 ± 0·0022 min⁻¹ (N = 5) with a T_1/2 of 34 min, then these parameters returned to values [k = 0·0060 ± 0·0008 min⁻¹ (N = 5), T_1/2 = 115min] similar to those measured in H-MEM with or without ouabain (B. Fossat, L. Bianchini, J. Porthé-Nibelle & B. Lahlou, in preparation).

Analysis of intracellular K⁺ content (Fig. 6) indicated similar kinetics: a rapid loss during the initial 10min (70 nequiv mg⁻¹protein) followed by a steady-state K⁺ concentration. When the uptake was determined simultaneously with ⁸⁶Rb flux, this enabled the corresponding efflux to be calculated. The net flux was sixfold higher than the uptake (Fig. 6 inset). This demonstrates that the release of K⁺ was seven times greater than the uptake.

Under the same conditions, intracellular Na⁺ content was remarkably stable for 20 min.

Transitory events occurring after osmotic shock cannot be accurately analysed by cumulative uptake measurements. Therefore pulse, or short-time (1 min), incubation experiments were carried to circumvent this difficulty. In isotonic H-MEM and ouabain H-MEM, ⁸⁶Rb uptake was constant with time, although the influx was four times lower in ouabain H-MEM than in H-MEM (Fig. 7). In contrast, in hypotonic media, ⁸⁶Rb uptake increased, reached a maximal value at 6min and then returned to the initial level after 15 min. Immediate effects of ouabain and hypotonic shock explain the discrepancy in the initial rates for the measurements in the first minute.

In hypotonic media, owing to dilution of the external Na⁺, accumulation of ²²Na into hepatocytes was lower than in isotonic ouabain H-MEM. However, the ratio of ²²Na uptake to sodium external concentration was constant in both cases (results not shown).

Likewise, hypotonicity did not modify the ²²Na efflux parameters (k = 0·030 ± 0·004 min⁻¹, N = 3, T_1/2 = 23min) which were analogous to those obtained in ouabain H-MEM (k = 0·036 ± 0·005 min⁻¹, N = 3, T_1/2 = 19·5 min), as has been confirmed by measurements of Na⁺ intracellular content.

As demonstrated elsewhere (B. Fossat, L. Bianchini, J. Porthé-Nibelle & B. Lahlou, in preparation), chloride exchange was very much faster than Na⁺ and K⁺ movements. Because of the higher magnitude of chloride fluxes, it was not possible to determine if Cl⁻ was involved in RVD by using cumulative uptake measurements. It was therefore necessary to apply the ‘pulse’ technique to ³⁶C1 uptake measurement, since a decrease of intracellular Cl⁻ (54nequiv mg⁻¹ protein 10 min⁻¹) was observed in hypotonic media.

As for Rb, in hypotonic media, ³⁶C1 influx analysed for 60-s periods every 2 min increased and reached a maximum between 4 and 6 min, returning to the control level over the same time period (Fig. 8).

Substituting impermeant anions (gluconate or methane sulphonate) for Cl⁻ during the hypotonic treatment allowed for analysis of the interdependence between K+ and Cl⁻. As indicated in Fig. 9, in H-MEM gluconate (as in H-MEM methane sulphonate) hypotonic shock did not stimulate ⁸⁶Rb uptake. Such an inhibition was observed when Cl⁻ was replaced by NO3⁻, an anion of the Hofmeister series.

To explore further the Cl⁻-dependence of K⁺ influx, ⁸⁶Rb pulse experiments were performed in the presence of loop diuretics. At 10⁻⁴moll⁻¹, piretanide, bumetanide and furosemide did not change ⁸⁶Rb influx consistently (less than 15%) in hypotonic media. Nevertheless, in the presence of 10⁻³moll⁻¹ furosemide, the stimulated ⁸⁶Rb uptake measured 6 min after transfer of hepatocytes to a diluted medium was inhibited by 78·8 ± 6·4 % (N = 4).

To characterize further the nature of the volume-induced K⁺ flux, we investigated the effects of various drugs on the uptake of ⁸⁶Rb measured after 10 min.

Two types of drugs were used to try to identify the nature (voltage-or Ca²⁺-dependence) of the induced K⁺ permeability (for a review see Latorre & Miller, 1983). The Ca²⁺ ionophore A23187 (10⁻⁵mol l⁻¹) or Ca²⁺-dependent K⁺ transport blockers, such as apamine (10⁻⁶mol l⁻¹) or quinine (10⁻³mol l^−l), were without effect. Likewise blocking agents of voltage-dependent K⁺ channels, Ba²⁺ (5×l0⁻³mol l⁻¹) and TEA⁺ (10⁻³mol l⁻¹), were completely ineffective.

Many vertebrate cells regulate their volume in hypotonic media (e.g. lymphocytes, red blood cells, Ehrlich ascites tumour cells, renal tubular cells). Regulatory volume decrease (RVD) following cell osmotic swelling is accomplished by the loss of osmotically active effectors that creates an ‘osmotically obliged’ water efflux leading to cell shrinking towards the initial isotonic volume. In most invertebrates, osmotically active particles involved in RVD are organic molecules, such as free amino acids (Gerard, 1975; Moran & Pierce, 1985) but in vertebrate cells inorganic ions are major determinants of the volume restoration.

In fishes, cell volume regulation has been examined in the erythrocytes of flounder (Fugelli, 1967; Cala, 1977) and trout (La Marre, 1983; Bourne & Cossins, 1984). Hitherto, fish hepatocytes have been models for other investigations, such as the transport of organic molecules (steroids: Porthé-Nibelle & Lahlou, 1981; amino-acids: Shuttleworth & Goldstein, 1984), gluconeogenesis (Seibert, 1985) and intracellular pH regulation (Walsh, 1986).

In a separate paper (B. Fossat, L. Bianchini, J. Porthé-Nibelle & B. Lahlou, in preparation) we have examined the characteristics of the exchange mechanisms by which trout hepatocytes maintain their ionic equilibrium in an isotonic environment. When these cells were exposed to hypotonic H-MEM (0·70 × isotonicity) they swelled to a maximum of 1-17 times the control volume after 3 min and, by 15 min the cell volume had returned to normal. This time course is similar to that observed in rat hepatocytes (Berthon, Claret, Mazet & Poggioli, 1980; Bakker-Grunwald, 1983) and human lymphocytes (Bui & Wiley, 1981; Grinstein et al 1984). Faster volume regulation (2 min for swelling and RVD) was reported by Larson & Spring (1984) for Necturus gallbladder epithelial cells.

In trout hepatocytes, hypotonic shock enhanced ouabain-insensitive ⁸⁶Rb uptake and efflux, indicating an increase in K⁺ permeability. Release of K⁺ was seven times greater than uptake, as demonstrated by net flux analysis, giving a net loss of K⁺ and water from the cell during RVD. Pulse experiments confirmed the transitory nature of the K⁺ permeability increase that had been observed in efflux experiments.

A loss of cellular K⁺ seems to be a common feature of RVD in most cells. However, in RVD it is necessary for the ionic fluxes to be electroneutral and osmotically effective. Three types of mechanisms may be evoked: (1) an electrogenic K⁺ pathway (Grinstein, Dupre & Rothstein, 1982b, in human lymphocytes) accompanied by a Cl⁻ movement of the same magnitude (Grinstein, Clarke, Dupre & Rothstein, 1982b); (2) cotransport of K⁺ and Cl⁻ (Larson & Spring, 1984, in Necturus gallbladder; La Marre, 1983, in trout erythrocytes); and (3) antiport K⁺/H⁺ associated with Cl⁻/HCO₃⁻ exchange (Kregenow, 1981; Cala, 1983, in Amphiuma red blood cells).

In contrast to the major shifts in K⁺, cellular Na⁺ content and Na⁺ fluxes remained unchanged after hypotonic shock. These results indicate that this ion is not implicated in RVD as observed in most cells studied except dog erythrocytes (Parker, 1983) and rabbit renal tubular cells (Grantham, Lowe, Dellasega & Cole, 1977; Paillard, Leviel & Gardin, 1979).

The failure of Ca²⁺-dependent (Gardos effect) and voltage-dependent (Cook & Haylett, 1985) K⁺ channel inhibitors to characterize the identity of the volume-induced K⁺ pathway makes the first mechanism unlikely in the present case.

In hypotonic H-MEM, the ³⁶C1 and ⁸⁶Rb influx pulse experiment indicated a similar time course for modifications in permeability. The Cl⁻-dependence of ⁸⁶Rb uptake is shown by the disappearance of increased K⁺ permeability in Cl⁻-free media (nitrate H-MEM, gluconate H-MEM and methane sulphonate H-MEM).

Similar results have been reported for trout erythrocytes (La Marre, 1983; Bourne & Cossins, 1984): after hypotonic shock, cellular volume increase enhanced a furosemide-sensitive K⁺ efflux (i.e. Cl⁻-dependent). Likewise, in Necturus gallbladder, Larson & Spring (1984) demonstrated that RVD results from KCl exit through the epithelial cell basolateral membrane, and in human red blood cells (Dunham, Stewart & Ellory, 1980) RVD involved activation of a furosemide-sensitive KCl cotransport.

In isotonic media, trout red blood cells respond to β-adrenergic stimulation by an increase in cell volume (Baroin, Garcia-Romeu, La Marre & Motais, 1984). The response of a decrease in volume involved a passive Cl⁻-dependent K⁺ loss (Borgese, Garcia-Romeu & Motais, 1987), which may be blocked by the Replacement of Cl⁻ with NO₃⁻. Such a substitution is effective in trout hepatocyte RVD.

Although our results suggest that RVD is mediated by a K⁺/C1⁻ cotransport, loop diuretics failed to inhibit the K⁺ permeability increase. Inhibition by 10^{− 3}mol l⁻¹ furosemide must be considered with caution since it is less specific than bumetanide (for reviews see Lauf, 1985; Geek & Heinz, 1986; Hoffmann, 1986).

The Cl⁻-dependence of the K⁺ fluxes suggests that K⁺/H⁺ and Cl⁻/HCO₃⁻ mechanism, as reported by Cala (1983) in Amphiuma red blood cells, may be ruled out in the present case. However, it is important to establish the role of the membrane potential, and changing intracellular pH, in the possible genesis of the volume-sensitive flux. The fact that NO₃⁻ cannot substitute effectively for Cl⁻argues for the involvement of a K⁺/Cl⁻ cotransport system of the red cell type.

The nature of the signal, induced by a change in cellular volume and how it is perceived by the transport system is not yet known. In Necturus gallbladder epithelial cells, Foskett & Spring (1985) reported a role for the intact microfilament network in activation of KCl transport during RVD. This conclusion is also reached for MDCK cells: dibutyryl cyclic AMP or agents known to elevate endogenous level of cyclic AMP alter the state of microfilaments in a way that changes cell volume. However, in trout hepatocytes, cytochalasin failed to reduce the hypotonic-dependent ⁸⁶Rb fluxes (our unpublished results).

The present study deals with volume changes in fish hepatocytes for the first time. Further investigations on the mechanism of KCl transport and the molecular nature of the triggering events for RVD are clearly important for elucidating this system.

Financial support from the Commissariat à l’Energie Atomique, Département de Biologie, is gratefully acknowledged.