Regulation of Intracellular pH by Toadfish (Opsanus Beta) Hepatocytes

Various physiological states and environmental conditions produce extracellular and tissue acid-base disturbances in fish (Heisler, 1986). For example, following strenuous exercise in fish, blood pH can decrease by up to 0.5 units (Wood and Perry, 1985), and certain tissues (such as active muscles) can also incur substantial acid loads (Milligan and Wood, 1986, 1987). Several in vivo studies of acid-base status in fish during extracellular acidosis have shown that intracellular pH (pHi) of liver tissue is often regulated within relatively narrow limits and in some cases pHi may actually become more alkaline (Walsh and Moon, 1982; Milligan and Wood, 1986, 1987; Perry et al. 1988). The mechanisms of these responses in vivo are not well understood. As in other animal cells (Roos and Boron, 1981; Walsh and Milligan, 1989), fish hepatocytes have specific plasma membrane transporters that exchange acid-base equivalents with the extracellular space (i.e. Na⁺/H⁺ exchange, Cl⁻/HCO₃⁻ exchange; Walsh, 1986; Hugentobler et al. 1987) and whose rates of function are sensitive to changes in intracellular acid-base state (Walsh, 1986). However, these in vitro studies of intracellular pH regulation in fish hepatocytes have focused primarily on descriptions of mechanisms, employing largely non-physiological perturbations. In general, studies of pHi regulation have placed little emphasis on the physiological context in which these acid-base transporters operate or on other potential means of pHi regulation (i.e. buffering capacity and metabolism: Roos and Boron, 1981; Walsh and Milligan, 1989).

The purpose of the present study, therefore, was first to characterize the acid-base transporter(s) involved in pHi regulation in hepatocytes from a marine teleost, the gulf toadfish (Ospanus beta). I then attempted to simulate the in vivo acid-base perturbations associated with a specific physiological state (exercise) in suspensions of hepatocytes in vitro in the presence and absence of inhibitors of pHi regulation. By this approach I hoped to gain insight into the relative importance and in vivo physiological role of acid-base transporters in pHi regulation in fish liver tissue.

Specimens of mature gulf toadfish, Opsanus beta (body mass 40–490g), were collected by roller trawl in Biscayne Bay, Florida, by a local fisherman during August to December 1988. Upon return to the laboratory, toadfish were held in running saltwater aquaria (salinity 33 %o, temperature 23–28°C) and fed a diet of chopped squid twice a week. Fish were fasted for 3 days before use. Hepatocytes were isolated by collagenase perfusion of the liver, as previously described (Walsh, 1987). Media were identical to those in Walsh (1987), except that bovine serum albumin (BSA) was omitted (see below); [lactate]=0.4 mmol I⁻¹ and [glucose] = 1.3 mmol I⁻¹, unless noted.

Extracellular pH (pHe) was measured in suspension media with a Radiometer PHM 84 pH meter and capillary glass pH electrode. Total CO₂ content was measured with a Corning 965 total CO₂ analyser. Intracellular pH was measurement by an adaptation of the BCECF-fluorescence method for measuring cytoplasmic pHi (Rink et al. 1982). Hepatocytes were incubated for 1h with 10μmoll⁻¹ BCECF-AM [2’, 7’-bis(carboxyethyl)-5(6)-carboxyfluorescein acetoxymethylester; 1 μ1 of a stock solution of 10 mmol I⁻¹ in dimethylsulphoxide (DMSO) was added per ml of cell suspension] (Calbiochem, La Jolla, CA), a non-fluorescent, permeant form of the dye (converted to the fluorescent impermeant BCECF intracellularly by endogenous hepatic esterases). Insufficient dye accumulated in cells during shorter periods, whereas longer periods resulted in no further significant accumulation. Cells were washed (centrifuged at 200g for 3min and resuspended in fresh suspension medium) several times to remove any extracellular BCECF-AM or BCECF, and then approximately 15–30 mg of cells was suspended in 3.5 ml of suspension medium with a small magnetic flea in a cuvette mounted on a small magnetic stirrer (Hellma, Cuv-O-Stir, model 333, Jamaica, NY) in a Perkin-Elmer LS3B fluorescence spectrophotometer. Fluorescence was continuously measured at 518 nm and cells were excited primarily at 490 nm (a pH-sensitive wavelength) with periodic switching to excitation at 440 nm (a pH-insensitive wavelength). After correction for autofluorescence by the cells (measured prior to BCECF loading, see Results) and leaked dye (by monitoring the medium after cells had been centrifuged at 13000 g for lmin), the ratio of fluorescence when excited at 490 nm to that when excited at 440 nm was used to calculate pHi. pHi is calculated by comparing ratios obtained for the experimental cells with ratios obtained for cells exposed to high-potassium medium (as for normal toadfish saline, Walsh, 1987, except NaCl and KC1 concentrations are 5 mmol I⁻¹ and 130 mmol I⁻¹, respectively) in the presence of 10μgml⁻¹ nigericin, a K⁺/H⁺ ionophore (1 μl of a stock solution of 10 mg ml⁻¹ 75% dimethyl formamide/25 % ethanol was added per ml of cell suspension). This treatment equilibrates pHe and pHi, and allows placement of the cell ratio values on a standard curve vs pHe. These curves were also compared with those obtained for pure BCECF (Calbiochem) in K⁺-rich suspension medium. BSA was omitted from all media since it binds nigericin and renders it ineffective in the calibration protocol.

In selected experiments, pHi was measured by the dimethyloxazolidine (DMO) method with [¹⁴C]DMO (Amersham, Arlington Hts, IL), as previously applied to fish hepatocytes (Walsh, 1986). Non-bicarbonate buffering capacity (β_a) was measured by a modification of the methods of Boron (1977) using the difference between pHi at the end of a 30 min NH₄Cl prepulse and immediately after NH4CI washout with normal medium plus 0.5 mmol I⁻¹ amiloride. The addition of amiloride ensured that Na⁺/H⁺ exchange did not raise pHi and lead to overestimates of β_a.

When kinetic experiments were performed on the effects of various agents on pHi, cells were exposed either by adding small volumes of concentrated stock solutions to the cuvette, or by low-speed centrifugation followed by rapid resuspension in the new medium and transfer to the cuvette. In the latter case, the of time before fluorescence measurements began was typically 10–15 s. This is considered the ‘zero time’ point.

Membrane potential was measured from the degree of accumulation of S¹⁴CN⁻, as previously applied to rat hepatocytes (Hoek et al. 1980). Toadfish hepatocytes were incubated in 0.36μCiml⁻¹ S¹⁴CN⁻ (ICN Radiochemicals, Irvine, CA) and samples were centrifuged through bromododecane, as previously applied to measurements of intracellular compounds in fish hepatocytes (Walsh, 1986,1987). Values were corrected for a small extracellular fluid content (i.e. less than 15 % of total water) by incubation of hepatocytes with [¹⁴C]polyethylene glycol (New England Nuclear, Boston, MA). Corrections for intracellular binding of S¹⁴CN⁻ by hepatocytes was made by abolishing the membrane potential by incubation of cells in K⁺-rich medium (as in BCECF calibrations) with 10μgml⁻¹ of the potassium ionophore valinomycin (1 μl of a stock solution of 10 mg ml⁻¹ 75% dimethyl formamide/25 % ethanol was added per ml of cell suspension).

The BCECF-fluorescence method (Rink et al. 1982) for measurement of pH was successfully applied to toadfish hepatocytes using only a few modifications.

Optimal incubation time for toadfish hepatocytes with BCECF-AM was slightly longer than with mammalian cell types. Also, since toadfish hepatocytes showed some fluorescence without the dye, the excitation and emission wavelengths were chosen to minimize this interference. However, since this fluorescence by the hepatocytes could not be completely eliminated in the useful wavelength range of the dye, this endogenous fluorescence (usually less than 15 % of the signal caused by the dye at 490 nm) was routinely subtracted from the signal. By employing these modifications, the calibration lines for the dye in situ and free in solution agreed well (Fig. 1). Although the slopes of the two lines agreed precisely, the lines were slightly offset (Fig. 1); this has been observed previously in other cell types and is attributed to the slight modification of the fluorescence characteristics of BCECF by binding to intracellular components (Rink et al. 1982).

Measurements of pHi by BCECF compared reasonably well with measurements by the DMO method, except that pHi measured by BCECF changed less at low and high extremes of pHe (Fig. 2).

In preliminary experiments, toadfish hepatocytes were subjected to pHi perturbations by a variety of non-physiological means (30 mmol I⁻¹ DMO, sodium acetate or NH4CI, and 50mmoll⁻¹ NaHCO₃/5% CO₂). DMO, sodium acetate and CO₂ all produced acidifications of pHi, as did an ammonium chloride prepulse followed by washout (see De Weer, 1978). However, the NH4CI prepulse method gave the most consistent acidifications and rates of recovery, so it was employed in further experiments. Toadfish hepatocytes responded to NH4CI with an initial alkalinization followed by a slow return to normal values by about 30 min (Fig. 3A). When cells were centrifuged and resuspended in normal medium (final measured NH₃ values were always less than 0.2 mmol I⁻¹), there was a distinct acidification of pHi followed by a slow recovery (Fig. 3B) and recovery was complete by 30–45 min. Rates of acid extrusion for the first 10 min of this recovery period (Table 1) compare reasonably with rates of acid extrusion for other tissues (Nuccitelli and Deamer, 1982). To determine which membrane transport system(s) was responsible for this recovery, hepatocytes were incubated with known inhibitors of acid-base transporters during recovery from acidification by NH4CI prepulse. SITS (4-acetamido-4’-isothiocyanostilbene-2,2’-disulphonic acid) had no effect on recovery (Table 1). However, amiloride substantially increased the extent of the initial acidification (Fig. 3B) and significantly slowed the rate of recovery from acidification (Fig. 3B; Table 1). Similar results were obtained when extracellular [Na⁺] was lowered by substituting N-methyl-D-glucamine (Table 1)

Non-bicarbonate buffering capacity in toadfish hepatocytes was calculated by using the zero time pHi values for amiloride recovery and 30 min pHi values for NH4CI prepulse and found to be –42.80 ±6.18 mmol pH unit⁻¹1⁻¹ (N=7). Membrane potential in toadfish hepatocytes was calculated from corrected distribution ratios of S¹⁴CN⁻ and found to be –33.67±3.07mV (N=3). This value did not vary substantially with changes in pHe or with pHi perturbations, although the binding correction factor did increase with decreasing pHe (results not shown). When toadfish hepatocytes were exposed to a physiological range of extracellular variables that change during strenuous exercise in vivo (P. J. Walsh, unpublished data), only changes in and pHe had significant effects on pHi (Table 2). Some of the variables had no effect on pHi (i.e. elevated [glucose], [epinephrine], [norepinephrine]), and when hepatocytes were exposed to these changes in the presence of amiloride to block Na⁺/H⁺ exchange, changes in pHi were similar to those that occurred under normal circumstances (i.e. a rapid acidification of about 0.1 – 0.15 units with little further change; Table 2). The changes in pHi caused by elevated were relatively small (approx. – 0.2 units), and these changes were corrected rapidly (within 10min). Exposure of cells to pHe values lower than normal led to a readjustment of pHi to a lower value (Fig. 2; Table 2), and these adjustments were almost complete within 30 s and complete by 5 min. (Kinetics are not shown for these very rapid changes. One limitation of the BCECF method, which applies to fish hepatocyte suspensions but not to immobilized cells, is that some pHi adjustments can occur in the brief time it takes to resuspend cells in a new medium and return them to the fluorometer.)

An increase in [lactate] to levels similar to post-exercise levels had no significant effect on pHi (Table 3). Since lactate is a moderately strong acid (pKa=4) this lack of effect was interesting and further experiments were designed to determine the mechanistic basis of this response. When Na⁺/H⁺ exchange was inhibited by amiloride, in addition to the initial acidification noted under normal conditions (Table 3), upon addition of lactate there was a slow but steady further acidification of pHi, indicating that lactate uptake and/or metabolism created an intracellular acid load. To determine if transport, metabolism or both processes caused the acid load, further experiments were conducted. When lactate metabolism was inhibited by 1 mmol I⁻¹ each of 3-mercaptopicolinic acid (3-MPA) and aminooxyacetic acid (AOA) (Walsh, 1987), pHi increased, and it continued to increase during the 30 min exposure to lactate (Table 3), indicating that lactate metabolism, not lactate transport, was responsible for generating the acid load.

Intracellular pH values obtained by the BCECF method for toadfish hepatocytes in vitro at normal (‘resting’) pHe (7.70) (Fig. 2) agree well with in vivo pHi values for fish liver obtained by the DMO method (Walsh and Moon, 1982; Milligan and Wood, 1986, 1987; Perry et al. 1988). At resting pHe, pHi determinations in toadfish hepatocytes in vitro by the BCECF and DMO methods agree closely (Fig. 2). However, as pHe becomes very alkaline or acidic, intracellular pH values measured by the BCECF method appear to be more ‘narrowly regulated’ than those measured by the DMO (Fig. 2). In previous studies using BCECF, it was demonstrated that this marker is confined primarily to the cytoplasmic compartment (Rink et al. 1982), whereas DMO yields an average pHi including most, if not all, of the intracellular compartments (Hinke and Menard, 1978). Therefore, it is likely that the differences in pHi values between the two methods at low and high pHe (Fig. 2) can be attributed to the differential regulation of the cytoplasmic vs ‘other’ compartments. Despite the close tracking of average (DMO) pHi with pHe in hepatocytes from toadfish (Fig. 2) and other fish (Anguilla rostrata, Walsh and Moon, 1983; Salmo gairdneri, Walsh, 1986), pHi is a regulated variable in fish hepatocytes. Most values of pHi fall below the line of equality for pHi and pHe (Fig. 2; Walsh and Moon, 1983; Walsh, 1986). Furthermore, in the present study I determined the membrane potential in toadfish hepatocytes to be – 33.76±3.O7mV (see Results), a value in close agreement with previous studies of skate hepatocytes (Wondergem et al. 1985) and rat hepatocytes (Hoek et al. 1980). This measurement allows the estimate (using the Nernst equation; Eckert and Randall, 1983) that pHi would be approximately 0.6 units below pHe if protons were in electrochemical equilibrium. Since this difference is never observed in toadfish hepatocytes incubated under physiological conditions (Tables 2, 3; Fig. 2), we can conclude that pHi is truly regulated in toadfish hepatocytes and probably in fish hepatocytes in general.

Further evidence that pHi is regulated in fish hepatocytes comes from studies of pHi perturbations by a variety of non-physiological means. pHi is rapidly corrected from acid loading by Na⁺/H⁺ exchange in rainbow trout hepatocytes (Walsh, 1986). Recovery from acid loading by toadfish hepatocytes is also inhibited by amiloride and low extracellular [Na⁺] (Table 1), suggesting that toadfish hepatocytes also possess a Na⁺/H⁺ exchanger. Interestingly, rainbow trout (Walsh, 1986) and skate (Hugentobler et al. 1987) hepatocytes possesses SITS-sensitive anion exchanger (i.e. a C1⁻/HCO₃⁻ exchanger that may also transport SO₄²⁻), and inhibition of this exchanger with SITS (or removal of Cl⁻) in rainbow trout hepatocytes elevates (DMO) pHi by 0.15 units, regardless of pHe (Walsh, 1986). However, SITS had no effect on pHi or on pHi recovery in toadfish hepatocytes (Table 1). Whether this difference is due to methodological biases, or represents a real species difference, deserves further attention.

Despite considerable potential to recover from acid loading by Na⁺/H⁺ exchange, it appears that this system has only a modest role in determining pHi under resting or basal conditions. When Na⁺/H⁺ exchange is inhibited by 0.5 mmol I⁻¹ amiloride, pHi is lowered by only about 0.1 unit and, at least during a 30 min incubation, pHi did not continue to drift downwards (Table 2). This observation suggests that under control conditions, proton leakage and metabolic acid loading are slow, and/or that intracellular buffers are capable of stabilizing pHi. In this respect, non-bicarbonate intracellular buffering capacity in toadfish hepatocytes is substantial, as measured by NH₄C1 prepulse and washout (–42.80±6.18mmolpHunit⁻¹l⁻¹, see Results), and agrees well with previous measurements on fiver tissue by HC1 and CO₂ titrations (Heisler, 1986; Milligan and Wood, 1986, 1987). Total intracellular buffering capacity would be the sum of non-bicarbonate buffering capacity and 2.3 × [HCO₃⁻] (Burton, 1978) or about –43+(–6) = –49mmolpH units⁻¹l⁻¹ under physiological conditions.

The importance of the Na⁺/H⁺ exchanger to pHi regulation increases under simulated exercise stress in toadfish hepatocytes. This exchanger is responsible for recovery from increased (respiratory acidosis), and for maintenance of pHi (albeit at lower values than resting pHi) at decreased pHe and constant (metabolic acidosis), in that inhibition with 0.5 mmol I⁻¹ amiloride prevents recovery or maintenance of pHi (Table 2). In addition, this sytem is important in protecting the cells from acidosis generated by lactate metabolism. When hepatocytes were exposed to elevated lactate levels typical of peak values in toadfish blood following exercise (5 mmol I⁻¹, P. J. Walsh, unpublished data), there was no effect on pHi (Table 3). However, when Na⁺/H⁺ exchange was inhibited, in addition to the immediate depression of pHi noted under resting conditions, pHi continued to drift downwards (Table 3). This acidification by lactate (which is normally compensated for by Na⁺/H⁺ exchange) could be the result of transport of lactate into the cell, metabolism of lactate, or a combination of the two processes. When metabolism of lactate is inhibited at elevated [lactate], there is an initial increase in pHi, followed by a further increase (Table 3), suggesting that lactate metabolism at elevated [lactate] is responsible for the acidification. Since lactate transport is by passive diffusion in toadfish hepatocytes, and continues to some degree even when metabolism is inhibited (Walsh, 1987), these data also suggest that transport of lactate per se does not acidify the cell.

In two in vivo studies of the effects of strenuous exercise on intracellular acid-base status in fish, pHi of liver was largely unaffected during the initial large extracellular acidosis (rainbow trout, Milligan and Wood, 1986; starry flounder, Milligan and Wood, 1987). The results of the present study with toadfish hepatocytes and a previous study with rainbow trout hepatocytes indicate that this in vivo pHi homeostasis is largely the result of intracellular buffering capacity and a Na⁺/H⁺ exchange system. Interestingly, in both the starry flounder and the rainbow trout, hepatic pHi steadily increases (by approx. 0.5 and 0.2 units, respectively) during the later stages of recovery (Milligan and Wood, 1986, 1987). Despite the examination of several relevant variables, the present in vitro study did not provide a mechanistic insight into the in vivo increase in pHi associated with long-term recovery from exercise. At least in short-term exposures, none of the post-exercise variables caused an increase in hepatocyte pHi (Tables 2, 3). Thus, the observed increase in hepatic pHi in vivo may be due to longer exposure to these factors, or to some other parameter not investigated. Experiments with hepatocytes in long-term culture may offer mechanistic insights into this interesting in vivo phenomenon. Alternatively, an increase in hepatic pHi during recovery from exercise may not be a general phenomenon in fish, and the presence or absence of this phenomenon should be examined in the toadfish.

I wish to thank Dr W. Ross Ellington for critically reading an earlier draft of this manuscript. I am grateful to Jimbo Luznar for supplying the toadfish and to Smith Kline & French for their generous gift of 3-mercaptopicolinic acid. This research was supported by an NSF grant (DCB-8608728) to PJW.