Toxic Injury from Mercuric Chloride in Rat Hepatocytes”

The relationship between cytosolic free Ca’+, mito- chondrial membrane potential, ATP depletion, pyri- dine nucleotide fluorescence, cell surface blebbing, and cell death was evaluated in rat hepatocytes exposed to HgC12. In cell suspensions, 50 pM HgClz oxidized pyri- dine nucleotides between l/z and 2 min, caused ATP depletion between 2 and 5 min, and produced an 89% loss of cell viability after 20 min. Rates of cell killing were identical in high (1.2 mM) and low (2.6 PM) Ca2+ buffers. video microscopy.

The relationship between cytosolic free Ca'+, mitochondrial membrane potential, ATP depletion, pyridine nucleotide fluorescence, cell surface blebbing, and cell death was evaluated in rat hepatocytes exposed to HgC12. In cell suspensions, 50 pM HgClz oxidized pyridine nucleotides between l/z and 2 min, caused ATP depletion between 2 and 5 min, and produced an 89% loss of cell viability after 20 min. Rates of cell killing were identical in high (1.2 mM) and low (2.6 PM) Ca2+ buffers.
Cytosolic free Ca2+ was determined in l-day cultured hepatocytes by ratio imaging of Fura-employing multiparameter digitized video microscopy. In high Ca2+ medium, HgC12 caused a 3-4-fold increase of free Ca2+ beginning after 6-7 min, but free Ca2+ did not change in low Ca2+ medium. Bleb formation occurred after about 4-5 min in both buffers prior to any increase of free Ca2+. Subsequently, in high Ca2+ medium, blebs became hot spots of free Ca2+ (>600 nM).
After about 2 min of exposure to HgC12, rhodamine 123 fluorescence redistributed from mitochondrial to cytosolic compartments signifying collapse of the mitochondrial membrane potential.
The results taken together demonstrate that bleb formation, ATP depletion, and the onset of cell death are not dependent on an increase of cytosolic free Ca2+. HgClz toxicity appears to be a consequence of inhibition of oxidative phosphorylation leading to ATP depletion and cell death.
The role of calcium in cell injury has been under intensive investigation in recent years. Work from several laboratories has led to the working hypothesis that anoxic and toxic injury causes cytosolic free calcium to increase which in turn initiates a sequence of events leading to irreversible injury and cell killing (1,2). As a consequence of increased free Ca*+, it is hypothesized that alterations of the cytoskeleton lead to cell surface bleb formation and subsequent rupture, that hydrolases (e.g. proteases, phospholipases, and endonucleases; Refs. [2][3][4] are activated further promoting lethal cell injury, and that Ca2+ enters mitochondria resulting in uncoupling and ATP depletion. Our recent work, however, has failed to support the calcium hypothesis of cell killing. Using inhibitors of glycolysis and oxidative phosphorylation to mimic ATP depletion of anoxia (chemical hypoxia), we observed that an increase of cytosolic free calcium did not precede cell surface blebbing or the onset of cell death (5,6). Nonetheless, increases of cell calcium may be important in other forms of cell injury. Our goal, therefore, has been to identify toxic chemicals which produce increases of cytosolic free Ca2+ so that the injurious effects, if any, of increased cytosolic free calcium may be characterized.
HgC12 is a potent nephrotoxin which is directly cytotoxic (7,8). Recently, Smith et al. (9) reported an increase of free calcium in suspensions of rabbit kidney proximal tubule cells following exposure to HgC12. This finding was interpreted to support the hypothesis that Ca2+ influxes induced by toxic chemicals initiate lethal cell injury. A second action of HgCl* is depolarization, uncoupling, and respiratory inhibition of mitochondria, effects which might also account for lethal cell injury (10,11). The objective of this investigation, therefore, was to compare the effects of HgClz on cytosolic free calcium and mitochondrial membrane potential in single living cells. The results show that cytosolic free Ca2+ does increase following exposure to HgC12 in Ca*+-containing medium but that this rise of free calcium does not accelerate cell killing. Rather, HgCl* appears to target mitochondria for its toxic effects.  (Table I).
Fluorescence Spectra of Fura-Released from Loaded Hepatocytes-Fluorescence was measured in supernatants of Fura-loaded cells after digitonin treatment. 340/365nm ratios in EGTA-and Ca2+-saturated buffers were nearly identical to those for Fura-free acid (Table II). However, 3401 380-nm ratios in Ca2+-saturated buffer averaged almost 50% less than for Fura-free acid, suggesting the presence of incomplete hydrolysis products of Fura-AM which Scanlon  The Kd of Fura-for Ca2+ was determined from measurements of fluorescence in EGTA buffers of varying Ca2+ concentrations as described under "Experimental Procedures." For 1 pM Fura-free acid, we observed a Kd of 168 nM (Table  II). In the presence of 50 pM HgC12, the Kd was 165 nM. Therefore, HgC12 at concentrations used in this study did not affect the affinity of Fura-for Ca2+. Similarly, HgC12 did not change the absolute fluorescence of Fura-(data not shown).
Subcellular Localization of Fura-in Single Cultured Hepatocytes-To localize Fura-within living cells, intracellular compartments were sequentially opened by increasing concentrations of digitonin. Digitonin at low concentrations causes selective permeabilization of the plasma membrane with sparing of intracellular organelles. At higher concentrations, cholesterol-containing organelles are permeabilized with sparing of low-cholesterol mitochondria (28,29). In order to document these selective effects of digitonin, we co-loaded additional fluorescent probes with specificity for mitochondria (rhodamine 123; Fig. 1C) or lysosomes (rhodamine-dextran; Fig. 2B Table I).
free, succinate-containing KRH buffer was employed to prevent mitochondrial swelling and depolarization after plasma membrane permeabilization. Preliminary experiments established that 20 and 100 pM digitonin were specific for permeabilizing the plasma membrane and nonmitochondrial organelles, respectively. After addition of the detergents, total fluorescence of Fura-inside single cells was measured (Table  I), and Figs. 1 and 2 show representative experiments demonstrating qualitative changes in the actual fluorescence images. After 20 pM digitonin, 83% of Fura-fluorescence at the Ca"-insensitive wavelength of 365 nm was released (Table I, Fig. 2C), but rhodamine 123 fluorescence and rhodaminedextran fluorescence (Fig. 20) Fig. 2E) and rhodamine-dextran fluorescence disappeared. Punctate mitochondrial labeling with rhodamine 123 remained after 100 pM digitonin (Fig.  lC), but after 0.1% Triton X-100, rhodamine 123 fluorescence was no longer detectable (data not shown). In good agreement with Fura-release from cell suspensions, these results indicated that 83% of Fura-was localized to cytosol, 11% to lysosomes and other endomembrane compartments (endosomes, endoplasmic reticulum, etc.) and 6% to mitochondria. On the basis of the predominantly cytosolic localization of Fura-2, we employed ratio-imaging of Fura-fluorescence to measure cytosolic free Ca *+ in single cultured hepatocytes with no attempt to correct measured fluorescence ratios for noncytosolic loading of Fura-2. Effect of H&Y, on Cytosolic Free Ca2'-Cytosolic free Ca2+ was measured in individual cultured hepatocytes after expo-Chloride in Rat Hepatocytes sure to 50 pM HgC12 (Fig. 3). In high Ca2' medium, a 3-4-fold increase of free Ca'+ occurred beginning after 6-7 min of exposure. Under these conditions, cells lost viability after 15-20 min as indicated by nuclear staining with propidium iodide (data not shown). In low Ca2' medium, cytosolic free calcium did not increase after 50 pM HgC12, although individual cells again lost viability after 15 to 20 min. Bleb formation observed by phase contrast optics occurred within 4-5 min of HgC12 addition in both high and low Ca2+ medium. Notably, bleb formation occurred prior to any change of free Ca*+. In Fig.  3, cytosolic free Ca2+ was measured at 2.5-min intervals. Since previous work by Smith and co-workers (9) suggested an early transient rise of Ca2+ after HgC12, we also measured Ca2+ at 20-s intervals but observed no transient changes (Fig. 4A).
To observe possible interference by HgC12 on Fura-fluorescence in situ, Fura-2-loaded cultured hepatocytes were treated with Br-A23187, a nonfluorescent Ca2+ ionophore. In control cells, 50 pM Br-A23187 produced an immediate increase of cytosolic free Ca2+ to a new steady state level of about 325 nM (Fig. 4B). A similar increase occurred when Br-A23187 was added 1 min after the addition of 50 pM HgC12 (Fig. 4A). Since in the absence of Br-A23187, Ca" did not increase until after 6-7 min of exposure to HgCL (Figs. 3 and 4A), these findings demonstrate directly that fluorescence of Fura-P-loaded cells remains responsive to changes of free Ca2+ after exposure to HgCl,.

Spatial Distribution
of Cytosolic Free Calcium-The spatial distribution of free Ca2+ within single cells was represented as pseudocolor maps. In a representative experiment in high Ca2+ medium, basal free Ca*+ averaged 175 nM and was evenly distributed (Fig. 5A). After 6 min of exposure to HgC12, cell surface blebs had formed as assessed by bright field phase contrast microscopy and also seen in the pseudo-color image, but average free Ca2+ was 160 nM and still homogeneous throughout the cell (Fig. 5B). After 16 min, free Ca*+ distribution was inhomogenous, and highest concentrations exceeding 600 nM were observed inside the blebs (Fig. 5C)  Fura-2-loaded cultured hepatocytes were incubated in KRH buffer. H&l2 (50 uM) in the absence (A. 0) and presence of Br-A23187 (50 pi) (A, n )'and Br-A23187 alone'(Bj were added where indicated (arrows). Cytosolic free Ca*+ was measured by ratio imaging of Fura-fluorescence using MDVM as described under "Experimental Procedures." One experiment representative of two.

Subsequently,
Fura-leaked to the extent that fluorescent images could no longer be obtained.
In a representative experiment in low Ca*+ medium, free Ca*+ averaged 115 nM, slightly lower than in high Ca2+ medium (Fig. 50). After 7 min of exposure to HgCl*, blebbing was apparent, but free Ca2+ concentration had not increased (Fig. 5E). After 16 min, free Ca2+ still had not increased (Fig.  5F), despite the presence of large, well developed blebs. After longer than 16 min, Fura-again leaked entirely from the cell.
Fura-Leakage after Exposure to HgC&-Absolute fluorescence at 340 and 365 nm was also monitored in order to follow Fura-leakage from the cells (Fig. 6). Before addition of HgC12, fluorescence at both wavelengths remained constant for more than 30 min without evidence of photobleaching or dye leakage from the cells (data not shown). Shortly after HgClz, however, a 40-50% increase of both 340 and 365 nm fluorescence occurred. The increases at the two wavelengths were proportionate such that 340/365-nm ratios were unchanged. Thereafter, fluorescence at 365 nm decreased at a relatively constant rate. After about 16 min, fluorescence dropped abruptly. The disappearance of Fura-f fluorescence coincided with the onset of propidium iodide staining (data not shown). These events signified hyperpermeability of the plasma membrane and onset of cell death.

Oxidation of Pyridine Nucleotides after Exposure to HgC&
Fluorescence of reduced pyridine nucleotides was measured in freshly isolated hepatocytes (Fig. 7). Before addition of HgC12, pyridine nucleotides were about 75% reduced. After 50 pM HgC&, fluorescence started to decrease within less than 30 s. After 2 min, pyridine nucleotides were maximally oxidized. The decrease of pyridine nucleotide fluorescence paralleled the increases of 340-and 365-nm fluorescence in Fura-2-loaded cells (Fig. 6).
Cell Viability in Freshly Isolated Hepatocyte Suspensions after Exposure to HgC&-After HgCl*, free Ca*+ increased in high Ca2+ medium but not in low Ca2+ medium. However, loss of cell viability appeared to occur identically.
To confirm the lack of effect of extracellular Ca2+ on cell killing, cell viability was measured in suspensions of hepatocytes using a propidium iodide cytotoxicity assay. Like trypan blue, propidium iodide is impermeable to viable cells, but binds to nuclei of nonviable cells. Since fluorescence of propidium iodide is enhanced following binding, this signal can be employed to monitor cell killing continuously in cell suspensions (14). Using the propidium iodide assay, HgClz was observed to cause dose-dependent cell killing in suspensions of freshly isolated hepatocytes (see "Experimental Procedures"). Notably, the rate of cell killing was virtually identical in high and low Ca2+ media (Fig. 8).

Effect of HgCl, on Mitochondrial Membrane Potential-
Changes of mitochondrial membrane potential in single cultured hepatocytes were monitored from rhodamine 123 fluorescence. After 50 pM HgC12, rhodamine 123 fluorescence began to increase to a maximal extent of about 50% after 2-3 min. Thereafter, total fluorescence decreased and a halfmaximal decrease occurred after about 8 min (Fig. 9). These changes were virtually identical in high and low Ca2+ media (Fig. 9).
As illustrated for a representative experiment in Fig. 10, before addition of HgC12, rhodamine 123 was localized to mitochondria (Fig. 1OA). By contrast, after 2.5 min of exposure to 50 pM HgC12, rhodamine 123 fluorescence was diffuse which indicated that dye had entered the cytosol from mitochondria. This abrupt redistribution of rhodamine 123 from mitochondria to cytosol was evidence of an early and rapid mitochondrial depolarization. Upon uptake into mitochondria, rhodamine 123 fluorescence is quenched by up to 75% (21). Thus, unquenching of rhodamine 123 as the fluorophore moves from the mitochondrial to the cytosolic compartment accounts for the rapid 50% increase of total cellular fluorescence observed after 2-3 min of HgC12 exposure (Fig. 9).
ATP Hydrolysis-In suspensions of hepatocytes, ATP declined sigmoidally after exposure to HgC12 (Fig. 11). After 1.5 min, ATP changed little, but between 2 and 5 min, ATP declined rapidly to less than 15% of initial values. The rate and extent of ATP hydrolysis was the same in high and low Ca2+ media. Comparing ATP with other parameters measured (Figs. 8, 9, and ll), ATP hydrolysis followed closely the depolarization of mitochondria but preceded bleb formation and cell death.

DISCUSSION
In this study we investigated the relationships of cytosolic free Ca*+, mitochondrial membrane potential, pyridine nucleotide fluorescence, cell surface blebbing, and ATP content with respect to loss of cell viability after exposure of rat hepatocytes to HgC12. A series of cellular events occurred in the following order after 50 pM HgC12: oxidation of pyridine nucleotides (0.5-2 min), mitochondrial depolarization (2 min), ATP depletion (2-5 min), and bleb formation (4-5 min), increased cytosolic free Ca2+ in Ca'+-containing buffer (6-7 min), and cell death (15-20 min of these events provided insight into the cellular and subcellular basis for toxic injury by HgC&. H$+ binds with high affinity to protein and nonprotein thiols (10,30) with consequent glutathione depletion and oxidative stress (31). In our model of HgClz toxicity in rat hepatocytes, we observed a rapid oxidation of pyridine nucleotides beginning within 30 s and essentially complete after 2 min (Fig. 7) membrane potential using rhodamine 123, a cationic fluorescent probe which accumulates electrophoretically into negatively charged compartments (21,22). In normal cells, rhodamine 123 produced bright, punctate staining of mitochondria. After 50 FM HgCl*, an abrupt redistribution of rhodamine 123 fluorescence from mitochondria to cytosol occurred after about 2 min (Fig. 10). This redistribution signified a collapse of the mitochondrial membrane potential. Thus, we observed in situ the phenomenon described in isolated mitochondria whereby pyridine nucleotide oxidation leads to mitochondrial depolarization and collapse of ion gradients (32).

Simultaneously
with this abrupt redistribution of rhodamine 123 fluorescence, a 50% increase of total fluorescence of rhodamine 123 was observed (Fig. 9). When rhodamine 123 is taken up by isolated liver mitochondria, fluorescence is quenched by as much as 75% (21). Thus, the increase of total cellular fluorescence after HgClz represented unquenching of fluorescence as rhodamine 123 moved from mitochondria to cytosol. Because total cellular fluorescence of rhodamine 123 decreased relatively slowly as dye leaked from the cells, imaging of rhodamine 123 redistribution from mitochondria to cytosol within individual cells was essential to capture dynamic changes of the mitochondrial membrane potential (Fig.  10).
Our observations of mitochondrial depolarization in situ confirm and extend previous reports of uncoupling and depolarization of mitochondria exposed to HgCl*. Scott and Gamble (33) and Southard and co-workers (34) showed marked increases of permeability to K' and Mg2+ in isolated liver and heart mitochondria after 3-10 PM HgC12. Weinberg and co-workers (11) described uncoupling, respiratory inhibition, and increased atractyloside-insensitive adenine nucleotide translocation in isolated kidney mitochondria with as little as 1 pM HgCl*. Similar changes were observed in kidney mitochondria isolated from HgClz-treated animals (8). Chavez and Holguin (10) documented pyridine nucleotide oxidation, Ca2+ release, and mitochondrial depolarization in isolated kidney mitochondria after 5 pM HgC12. The latter authors implicated Ca*+ recycling as a cause of depolarization. The present work documents pyridine nucleotide oxidation and mitochondrial depolarization in situ after HgC12. However, increases of cytosolic free Ca2+ coincident with mitochondrial depolarization were not observed even transiently during measurements made at 20-s intervals (Fig. 4). The explanation for the absence of a change of cytosolic free Ca2+ may simply be that the concentration of mitochondrial free Ca2+ was insufficient to perturb cytosolic Ca2+ upon equilibration of the mitochondrial matrix and the cytosol. Mitochondrial free Ca*+ has been estimated to be 500 nM or less in intact cells (35), and one recent report suggests that mitochondrial free Ca2+ is actually less than cytosolic (36). In any event, if mitochondrial Ca2+ recycling was occurring after HgC&, then it was occurring without a net change of cytosolic free Ca'+. ATP hydrolysis after HgCL followed closely mitochondrial depolarization but preceded bleb formation, increased cytosolic Ca2+ (in high Ca2+ medium), and cell death (Fig. 11). Thus, ATP depletion appeared to be the consequence of uncoupling of oxidative phosphorylation. Interestingly, ATP depletion occurred at about the same rate and extent as when oxidative phosphorylation and glycolysis were inhibited with KCN and iodoacetate (chemical hypoxia) (14). Thus, the rapid decline of ATP after exposure to HgC& suggests that glycolytic ATP synthesis was also inhibited, possibly by Hg2 binding to reactive SH groups of glyceraldehyde-phosphate dehydrogenase.
Inhibition of glycolysis would also explain why fructose, and effective glycolytic substrate which can prevent ATP depletion and cell death of anoxic hepatocytes (14, 37), did not delay or prevent cell killing caused by HgC& (data not shown).
Jones and co-workers (38)(39)(40)  A cultured hepatocyte was loaded with rhodamine 123 and incubated with HgC12 as described in Fig. 8. Before addition of HgCl,, rhodamine 123 was localized to mitochondria (A). After 1 min of exposure to HgCb, fluorescence was still punctate and confined to mitochondria (B), but after 2.5 min, rhodamine 123 had redistributed from mitochondria into cytosol to make the cell diffusely fluorescent (C). After 11 min, most dye was lost from the cell (I)). Freshly isolated hepatocytes (100,000 cells/ml) were incubated in KRH containing low (0) or high (0) Ca*+ as described in Fig. 8. At various times after 50 /*M HgC12, the cell suspensions were quenched, and ATP was determined as described under "Experimental Procedures." nearly the same conditions. Thus, depolarization of the mitochondrial membrane potential may be an important factor, independent of ATP depletion, leading to the onset of cell death.
Cell surface blebbing is an early indication of hypoxic and toxic injury to hepatocytes (41)(42)(43). After HgClr, blebbing was evident after 4-5 min. Bleb formation followed ATP depletion, but preceded any increase of cytosolic free Ca'+. Our earlier work had demonstrated similar blebbing following ATP depletion during anoxia and chemical hypoxia (5,12). Thus, ATP depletion can account for blebbing during both reductive stress (anoxia, chemical hypoxia) and oxidative stress (HgCl* toxicity).
High energy phosphate compounds are required for normal assembly and turnover of cytoskeletal elements. ATP depletion may disrupt these cytoskeletal structures to cause blebbing.
We employed Fura-to measure cytosolic free Ca2+ in single hepatocytes. The probe was introduced into cultured hepatocytes as its membrane-permeant acetoxymethyl ester. With this type of loading, intracellular esterases liberate Furafree acid which is trapped inside the cells. For accurate measurement of cytosolic free Ca", Fura-AM must be lmin hydrolyzed completely and the free acid confined to cytosol. However, several reports indicate that this is not always the case. For example, Fura-AM may be incompletely hydrolyzed yielding fluorescent compounds which are insensitive to Ca'+ (27, 44). Furthermore, Fura-may localize to intracellular compartments other than cytosol (e.g. mitochondria, lysosomes, secretory granules, and endoplasmic reticulum, Refs. 27 and [44][45][46][47][48][49]. In view of these concerns, we evaluated the degree of hydrolysis and subcellular localization of the dye. In one set of experiments, we released Fura-from freshly isolated hepatocytes using digitonin and Triton X-100. Digitonin (100 pM) released 93% of the fluorescence released by Triton X-100 (0.1%) ( Table I). Since digitonin permeabilizes plasma membranes but not mitochondrial membranes, whereas Triton X-100 solubilizes all cellular membranes, we conclude that more than 90% of Fura-was in a nonmitochondrial compartment.
Compared with pure Fura-free acid, subtle differences were observed for Fura-fluorescence released from digitonintreated cells. The 340/380-nm ratio in high Ca*+ medium was lower than that for pure free acid, possibly because of products of incomplete hydrolysis or other metabolism (27) ( Table II). By contrast, the 340/365nm ratio was unchanged. Hence, we employed 340/365-nm fluorescence ratios to measure free Ca'+. This approach had the additional advantage of allowing us to monitor cellular dye content at 365 nm.
We further localized Fura-in loaded hepatocytes using MDVM.
HgCl, (50 PM) caused a 3-4-fold increase of free cytosolic Ca'+ beginning after 6-7 min in high Ca*+ medium (Fig. 3). In low Ca*+ medium, free Ca*+ did not change. Thus, the source of increased free Ca'+ was extracellular.
In high Ca*+ medium, spatial gradients of Ca*+ also developed within cells, and blebs became hot spots of free Ca*' (>600 nM) (Fig. 5).