Cytotoxicity of S-(1,2-dichlorovinyl)glutathione and S-(1,2-dichlorovinyl)-L-cysteine in isolated rat kidney cells.

S-(1,2-Dichlorovinyl)glutathione (DCVG) and S-(1,2-dichlorovinyl)-L-cysteine (DCVC) produced time- and concentration-dependent cell death in isolated rat kidney proximal tubular cells. AT-125 blocked and glycylglycine potentiated DCVG toxicity, indicating that metabolism by gamma-glutamyltransferase is required. S-(1,2-Dichlorovinyl)-L-cysteinylglycine, a putative metabolite of DCVG, also produced cell death, which was prevented by 1,10-phenanthroline, phenylalanylglycine, and aminooxyacetic acid, inhibitors of aminopeptidase M, cysteinylglycine dipeptidase, and cysteine conjugate beta-lyase, respectively. Aminooxyacetic acid and probenecid protected against DCVC toxicity, indicating a role for metabolism by cysteine conjugate beta-lyase and organic anion transport, respectively. DCVC produced a small decrease in cellular glutathione concentrations and did not change cellular glutathione disulfide concentrations or initiate lipid peroxidation. DCVC caused a large decrease in cellular glutamate and ATP concentrations with a parallel decrease in the total adenine nucleotide pool; these changes were partially prevented by aminooxyacetic acid. Both DCVG and DCVC inhibited succinate-dependent oxygen consumption, but DCVC had no effect when glutamate + malate or ascorbate + N,N,N',N'-tetramethyl-p-phenylenediamine were the electron donors. DCVC inhibited mitochondrial, but not microsomal, Ca2+ sequestration. These alterations in mitochondrial function were partially prevented by inhibition of DCVG and DCVC metabolism and were strongly correlated with decreases in cell viability, indicating that mitochondria may be the primary targets of nephrotoxic cysteine S-conjugates.

dichloroviny1)-L-cysteine (DCVC), respectively, have been employed as model compounds to study halogenated hydrocarbon-induced nephrotoxicity (1). Both conjugates are potent and specific nephrotoxins and produce proximal tubular necrosis and increases in b!ood urea nitrogen and urine glucose concentrations (1-4).
The hypothesis that the nephrotoxicity of glutathione Sconjugates is dependent on their metabolism to the corresponding cysteine S-conjugates, which are metabolized to the ultimate toxic species by cysteine conjugate @-lyase (EC 4.4.1. 13), has been validated by the elucidation of the metabolic pathway of DCVG and DCVC bioactivation and by use of specific enzyme inhibitors and nonmetabolizable analogues (1, 4, 5 ) . DCVG, like other glutathione S-conjugates, is converted to the corresponding cysteine S-conjugate, DCVC, by y-glutamyltransferase (EC 2.3.2.2) and cysteinylglycine dipeptidase (EC 3.4.13.6) or aminopeptidase M (EC 3.4.11.2), which are also responsible for degradation of GSH to cysteine (1,6). Subsequent metabolism of DCVC, however, may follow two routes: N-acetylation, catalyzed by N-acetyltransferase, to form the mercapturic acid or a P-lyase-catalyzed a,@-elimination to produce pyruvate, ammonia, and a reactive thiol (1, 7). A role for the 8-lyase reaction in DCVG-and DCVCinduced nephrotoxicity has been established (1,4).
Little is known about the mechanism by which DCVG and DCVC produce nephrotoxicity. Both compounds inactivate the organic ion transport system in rabbit renal proximal tubules (5, 8). DCVC causes a 50% decrease in nonprotein sulfhydryl groups in rabbit renal proximal tubules, and DCVC toxicity is potentiated by GSH depletion and is diminished by exogenous GSH (9). In rat kidney slices and in rat kidney and liver mitochondria, DCVC is a potent inhibitor of respiration (10, l l ) , indicating that mitochondria are targets within the cell. Specific interactions with individual enzymes also occur (12,13) and may partially explain DCVC toxicity. What is lacking, however, is an analysis of the biochemical processes in the cells affected by DCVG and DCVC and the relationship of these effects to cytotoxicity.
To study the mechanism of S-conjugate nephrotoxicity in more detail, we have employed isolated rat kidney proximal tubular cells as an in vitro system. We report herein that DCVG and DCVC produced both time-and concentrationdependent cell death and that inhibition of mitochondrial function correlated with the observed cytotoxicity, indicating that mitochondria are primary targets. A preliminary report of some of this work has been presented (1).
Cell Isolation-Isolated rat kidney proximal tubular cells were prepared by the collagenase perfusion method of Jones et al. (18) from male Fischer 344 rats (200-300 g; Charles River Laboratories, Wilmington, MA). Cell concentration was determined in the presence of 0.2% (w/v) trypan blue in a hemacytometer, and cell viability was measured by trypan blue exclusion or by lactate dehydrogenase leakage into the medium. Typically, 85-95% of the cells excluded trypan blue, and 10-18% of the total lactate dehydrogenase activity leaked from the cells. Cell yield was approximately 20 X lo6 cells/kidney. Cells were incubated at 37 "C in Krebs-Henseleit buffer, pH 7.4, supplemented with 25 mM Hepes, 2.5 mM CaCl2, 25 mM NaHC03, and 2% (w/v) bovine serum albumin. All buffers were equilibrated Assays-@-Lyase activity was measured by the formation of 2mercaptobenzothiazole from BTC, as previously described (17) except that substrate was dissolved in supplemented Krebs-Henseleit buffer, pH 7.4. Cells (1 X 106/ml) were incubated at 37 "C with 0.1-2 mM BTC. After 2.5 min, 0.3 ml of 30% (w/v) trichloroacetic acid was added to 1.5 ml of cell suspension, and the absorbance of the acid extract was measured at 321 nm.
Cellular GSH, GSSG, and glutamate concentrations were measured by ion-exchange high-pressure liquid chromatography on a 10-p U1trasil-NH2 column (4.6 X 250 mm; Beckman Instruments) with a methanol-acetate solvent system according to the method of Reed et al. (19). Isolated cells (2.5 X 106/ml) were incubated at 37 "C in supplemented Krebs-Henseleit buffer, pH 7.4, with the indicated additions. At various times, 1-ml samples of the cell suspension were centrifuged in an Eppendorf microcentrifuge for 1 min. The cells were resuspended in 0.5 ml of 0.9% (w/v) NaCl, and 0.1 ml of 30% (w/v) trichloroacetic acid was added. A sample (0.5 ml) of the acid extract was neutralized with approximately 0.1 ml of 1 M KzC03, and 0.1 ml of 40 mM iodoacetic acid was added. After 1 h of incubation at room temperature, 0.5 ml of 1.5% (v/v) l-fluoro-2,4-dinitrobenzene in absolute ethanol was added, and the samples were incubated for 4 h in the dark. Derivatives were separated by gradient elution, were detected at 365 nm, and were quantified by comparison to standards. GSH (10 nmol/106 cells) was added during acid extraction to assess recovery; 9.4 nmol of GSH (94.0%) was recovered, and 8.6 nmol (91.5%) was in the form of GSH and 0.8 nmol(8.5%) was in the form of GSSG (in GSH equivalents), indicating little autooxidation of GSH during the derivatization procedure.
For measurement of total cellular adenine nucleotide concentrations, 0.2 ml of 30% (w/v) trichloroacetic acid was added at various times to 1-ml samples of cell suspensions (4-5 X lo6 cells/ml) incubated at 37 "C with supplemented Krebs-Henseleit buffer, pH 7.4, with the indicated additions. Acid extracts, neutralized with 5 M KzC03, were analyzed for ATP (20) and ADP and AMP (21) by pyridine nucleotide-linked enzyme assays.
Cellular oxygen consumption was measured polarographically with a Clark-type electrode at 37 "C. The electrode was calibrated with air-saturated buffer at 37 "C, and zero oxygen concentration was obtained by addition of sodium sulfite. Measurements were begun by addition of 3.3 mM succinate, 4 mM glutamate + 2 mM malate, or 1 mM ascorbate + 0.2 mM TMPD as respiratory substrates to 1 X lo6 cells. The endogenous rate of cellular oxygen consumption (ie. in the absence of added respiratory substrates) in control cells was 15 nmol of O2/min/1O6 cells. At all time points examined (up to 3 h), endogenous rates of cellular oxygen consumption in the absence and presence of S-conjugates were 30-40% of succinate-stimulated rates.
For the measurement of cellular Ca2+ sequestration, isolated kidney cells (4-5 X 106/ml) were separated from the incubation medium by rapid centrifugation through a suspension of Caz+-and Mp-free Hanks' solution (22) and Percoll (20%, v/v). The cells were then resuspended in the modified Hanks' solution. Intracellular Ca2+ compartmentation was determined by dual-wavelength spectroscopy with 95% 0 2 , 5 % COz.

RESULTS
Isolated kidney cells were incubated with DCVG and cell viability, determined by trypan blue exclusion and lactate dehydrogenase leakage, was assessed at various times and at various DCVG concentrations (Fig. 1). DCVG (1 mM) reduced cell viability from approximately 90% at the start of the incubation to less than 40% at 4 h. Addition of 0.25 mM AT-125, a potent inhibitor of y-glutamyltransferase (27), protected the cells from DCVG toxicity. In contrast, addition of 1 mM glycylglycine, a y-glutamyl acceptor, caused a slight, but statistically significant, increase in DCVG cytotoxicity at 0.5 and 1 h. These results show that metabolism of DCVG by y-glutamyltransferase is required for the expression of toxicity and indicate that the cysteinylglycine analogue DCVCG is the product of the above reaction. To test this point, the effects of 1,lO-phenanthroline, an inhibitor of aminopeptidase M (28), and phenylalanylglycine, an inhibitor of cysteinylglycine dipeptidase (6), on DCVG cytotoxicity were examined (data not shown). Both inhibitors protected the cells from DCVG toxicity, indicating that the ultimate toxin is generated by DCVCG metabolism. This conclusion was supported by studies with DCVCG itself; DCVCG produced cell death comparable to that produced by DCVG, and 1,lO-phenanthroline and phenylalanylglycine protected against DCVCG-induced cell death (data not shown).
After metabolism of DCVG to DCVCG, the next reaction in the bioactivation pathway is removal of the glycyl residue to form the cysteine conjugate DCVC (1). Isolated kidney cells were incubated with DCVC, and cell viability was assessed at various times and at various DCVC concentrations (Fig. 2). DCVC, at a concentration of 1 mM, reduced cell viability from approximately 90% to 35% during a 4-h incubation. Addition of the P-lyase inhibitor AOAA (1, 4) or the renal anion transport inhibitor probenecid protected the cells from DCVC toxicity, indicating that probenecid-sensitive  transport and @-lyase-catalyzed metabolism are important in DCVC cytotoxicity. Furthermore, AOAA protected against DCVG and DCVCG toxicity, but had no effect on y-glutamyltransferase activity (data not shown). AOAA (0.1 mM) or AT-125 (0.25 mM) alone produced less than a 10% loss in cell viability relative to control cells over the 4-h incubation. These results supported the hypothesis that DCVG bioactivation occurs by this sequence of reactions. Hassall et al. (9) reported that exposure of rabbit renal tubules to DCVC produced a 50% decrease in the content of nonprotein sulfhydryl groups. Hence the effect of DCVC on total cellular GSH, GSSG, and glutamate concentrations in rat kidney cells was determined (Table I). Glutamate concentrations were measured, because glutamate is normally present at relatively high concentrations in the kidney and serves as a good substrate for cellular energy production (29). Incubation of isolated kidney cells with 1 mM DCVC for 30 min caused a 34% reduction in the GSH concentration, a 48% reduction in the glutamate concentration, and no significant change in the GSSG Concentration. Incubations up to 2 h produced no further decreases in GSH concentrations, but glutamate concentrations declined further to 35% of control (data not shown). The addition of 0.1 mM AOAA prevented most of the loss in glutamate and nearly half of the loss in GSH concentration. The modest decline in GSH concentration and the lack of a change in GSSG concentration due to exposure to DCVC indicate that production of an oxidative stress is probably not a primary mechanism of cysteine Sconjugate-induced cytotoxicity. In agreement with this conclusion, lipid peroxidation, as assessed by formation of thiobarbituric acid-reactive material, was not detected (data not shown).
The decrease in the glutamate concentration indicates that DCVC affects cellular energy metabolism. Therefore, the effect of DCVC on three parameters of mitochondrial function in isolated cells was examined adenine nucleotide status, oxygen consumption, and Ca2+ compartmentation. Several nephrotoxic agents, such as aminoglycosides and heavy metals, and pathological conditions, such as renal ischemia, cause mitochondrial dysfunction (30). A prominent and early effect of these diverse agents is a dramatic fall in ATP concentrations with a parallel decline in total adenine nucleotide content (30). Similarly, isolated kidney cells incubated with 1 m M DCVC for 30 min exhibited a 67% decline in both ATP and total adenine nucleotide concentrations (Table 11). The large decrease in ATP concentration, combined with the moderate (30-40%) increases in ADP and AMP, caused the cellular ATP/ADP ratio and energy charge to fall, indicating that the ability of the cells to maintain ATP-dependent functions will be severely impaired. Addition of AOAA to inhibit DCVC bioactivation provided substantial protection against DCVCinduced alterations in adenine nucleotide status.
DCVG caused a time-dependent inhibition of succinatedependent oxygen consumption, and addition of AT-125 prevented this effect (Fig. 3). Similarly, DCVC inhibited succinate-dependent oxygen consumption, and inhibition of DCVC metabolism by inclusion of AOAA partially prevented this effect (Fig. 4A). The concentration-response curve for DCVC inhibition of oxygen consumption was relatively steep; DCVC concentrations as low as 0.1 mM caused substantial inhibition (Fig. 4B).
The effect of DCVC on cellular oxygen consumption with respiratory substrates other than succinate was also studied (Table 111). Incubation times of 2 h were chosen to compare the substrate specificity of DCVC-induced inhibition of cellular respiration because earlier incubation times, while giving qualitatively similar results, produced only modest inhibition of oxygen consumption. In contrast to metabolite levels (cf .  Tables I and 11), changes involving integrated cellular function, such as maintenance of cell viability and cellular respiration, require longer to occur. In contrast to the inhibitory effect of DCVC with succinate as respiratory substrate, little effect was observed when glutamate + malate or ascorbate + TMPD were the electron donors. This shows that DCVC specifically inhibits succinate oxidation.
Mitochondria and the endoplasmic reticulum play major roles in regulation of hepatic (22,31) and renal (25,32) intracellular Ca2+ homeostasis by sequestering Ca2+, thus buffering changes in cytosolic Ca2+ concentrations. Therefore, the effect of DCVC on cellular Ca2+ compartmentation was investigated (Fig. 5). Incubation of isolated kidney cells with 1 mM DCVC reduced mitochondrial Ca2+ sequestration by 62% in 2 h. In contrast, microsomal Ca2+ sequestration was not impaired by DCVC. The plasma membrane is also involved in intracellular Ca2+ homeostasis. Addition of a low Effect of DCVC on cellular adenine nucleotide concentrations Isolated kidney cells (4-5 X 106/ml) were incubated with the indicated additions and were processed as described under "Experimental Procedures" for analysis of cellular ATP, ADP, and AMP concentrations by pyridine nucleotide-linked enzyme assays. Energy charge = (ATP + %ADP)/(ATP + ADP + AMP). Results are the mean f S.E. of 3 cell preparations.   concentration of digitonin, which permeabilizes the plasma membrane without affecting the mitochondrial and endoplasmic reticular membranes, caused only a slight increase (40%) in FCCP-releasable Ca2+ (data not shown). Thus, as was found in isolated hepatocytes (23), plasma membrane Ca2+ transport mechanisms efficiently extruded Ca2+ mobilized from mitochondria by FCCP. In cells treated with 1 mM DCVC, the amount of Ca2+ released by FCCP, but not released from the cells in the absence of digitonin, was unchanged at

TABLE I11
Influence of respiratory substrate on DCVC-induced inhibition of cellular oxygen consumption Cellular oxygen consumption was measured polarographically with a Clark-type electrode at 37 "C. Isolated cells (1 X 106/ml) were incubated in the absence or presence of 1 mM DCVC for 2 h. Measurements were begun by addition of respiratory substrates to the cell suspension. Results are the mean f S.E. of 3 cell preparations.  5. Effects of DCVC on mitochondrial (A), microsomal   (B), and total cellular (0 Caz+ sequestration. Cells (4-5 X IO6/ ml) were incubated in the absence (0) or presence (0) of 1 mM DCVC and were analyzed as described under "Experimental Procedures" for spectrophotometric measurement of the intracellular Ca2+ pools released by the sequential addition of 10 p~ FCCP and 15 p~ A23187. Total cellular Ca2+ is that released by FCCP and A23187. Results are the mean of 4 cell preparations. Standard errors were less than 10% of the means. 0.5 and 1 h and was increased by 25% at 2 h, indicating that plasma membrane Ca2' transport is not a primary target in DCVC nephrotoxicity (data not shown).

DISCUSSION
The present studies describe the effects of the S-haloalkenyl conjugates DCVG and DCVC on a variety of cellular functions and correlate these changes with the cytotoxicity of these compounds. Isolated rat kidney proximal tubular cells are ideally suited for these studies, because the cells are uniformly exposed to exogenous materials and can be well oxygenated and thus maintained for relatively long periods of time (greater than 4-5 h) with sufficient viability (cf. Figs. 1 and  2). Moreover, the cell preparation used has been well characterized with regard to drug metabolism, mitochondrial respiration, hemoprotein content, GSH metabolism, and amino acid transport (6, 18, 33). The use of isolated cells allows effects of S-conjugates on integrated cellular processes, such as mitochondrial respiration, cellular Ca2+ homeostasis, and maintenance of cell ivability, to be studied. Finally, proximal tubular cells are the i n vivo target for nephrotoxic cysteine Sconjugates, and hence isolated proximal tubular cells are a particularly relevant model.
Protection from the cytotoxic effects of DCVG, DCVCG, and DCVC provided by inhibitors of the bioactivating enzymes y-glutamyltransferase, cysteinylglycine dipeptidase and aminopeptidase M, and the @-lyase verifies the suggested role of these enzymes in S-haloalkenyA conjugate nephrotox- icity (1, 4). Thus, toxicity produced by DCVG is due to its metabolism to DCVCG and subsequently to DCVC, followed by 8-lyase-catalyzed generation of the presumed ultimate toxin 1,2-dichlorovinylthiol. The incomplete protection of the cells from DCVC cytotoxicity by AOAA is probably due to the relatively rapid nonenzymatic rate of DCVC degradation, which is attributable to the effectiveness of 1,2-dichlorovinylthiol as a leaving group. Previous in vivo and in vitro studies in this laboratory demonstrated that the same bioactivation pathway is involved in the nephrotoxicity of the glutathione and cysteine S-conjugates of chlorotrifluoroethylene (14,34).
The data presented in this study indicate that inhibition of cytosolic or other extramitochondrial processes in the kidney is probably not an important part of the mechanism of DCVCinduced nephrotoxicity. The modest decline in cellular GSH concentrations without concomitant oxidation of GSH to GSSG and the absence of detectable lipid peroxidation are inconsistent with production of an oxidative stress as an important mechanism of S-conjugate-induced nephrotoxicity. The absence of measurable lipid peroxidation in the present study may be due to species differences or may indicate that under certain conditions, DCVC produces an oxidative stress. The loss of GSH that was observed may be due to the rapid turnover of this tripeptide (t,,, = 0.5 h) in these cells and to a depletion of the amino acid precursors and ATP required for its resynthesis (36).
The inhibition of cellular respiration with succinate as the substrate and the rapid and dramatic decline in both ATP concentrations and the total adenine nucleotide pool caused by DCVC are similar to effects produced by many toxic agents and pathological states affecting the kidneys (30). The renal proximal tubule is highly dependent upon oxidative metabolism for generation of ATP (29). Furthermore, the ATP pool turns over rapidly in these cells, and there is a close coupling between oxidative metabolism and ATP-dependent functions, particularly active transport (37,38). An acute decline in ATP supply is not, however, associated with irreversible cell damage (39), indicating that perturbations in adenine nucleotide metabolism are not key events in DCVC-induced nephrotoxicity. Of greater significance than the decrease in ATP concentration produced by DCVC is the decline in the ATP/ADP ratio and the cellular energy charge (Table 11). Concentration ratios, rather than concentrations of the individual components of the adenylate system, are important in regulation of energy metabolism (40). The energy charge is maintained within a narrow range in normal cells due to the kinetics of adenine nucleotide metabolism (30, 39, 40). As both ATP concentrations and the energy charge decrease, AMP deaminase and 5'-nucleotidase are activated, which leads to increases in IMP and adenosine concentrations and to an irreversible loss of AMP; the activation of these enzymes supports maintenance of the energy charge and thus normal cell function. The decrease in the size of the total adenine nucleotide pool observed in the present study indicates that similar processes occur in response to DCVC exposure.
The selective inhibitory effect of DCVC on oxygen consumption with different electron donors indicates that succinate oxidation is specifically inhibited in these cells. In agreement with this conclusion, DCVC inhibits succinate:cytochrome c oxidoreductase activity in rat kidney mitochondria.' DCVC also inhibited the ability of mitochondria to regulate intracellular Ca2+ distribution, providing further support for the conclusion that DCVC-induced toxicity is primarily associated with mitochondrial dysfunction. Comparison of the rapid changes in the adenine nucleotide pool (Table 11) with the slower time course of inhibition of mitochondrial Cazc sequestration (Fig. 5) indicates the importance of maintenance of Ca2+ homeostasis in the renal proximal tubule cell. Moreover, Ca" uptake takes priority over oxidative phosphorylation in the hierarchy of mitochondrial function, thereby effectively uncoupling oxidative phosphorylation (41).
Much recent work has focused on the role of Ca2+ in cellular injury and cell death. Maintenance of a low cytosolic Ca2+ concentration is essential for normal cell function (25,31,32). Farber and colleagues (42, 43) proposed that influx of extracellular Ca2+ leading to Ca2+ overload is the final common mediator of irreversible cell injury, although more recent studies support a critical role for redistribution of endogenous  cytotoxicity (44, 45). Therefore, inhibition of mitochondrial Ca2+ sequestration may be a key component of the mechanism of DCVC nephrotoxicity.
Correlation analysis revealed a firm relationship between cell viability and cellular oxygen consumption or mitochondrial Ca2+ sequestration (Fig. 6). DCVC inhibited mitochondrial, but not microsomal or plasma membrane, Ca2+ sequestration; hence DCVC-induced alterations in intracellular Ca2+ distribution can be attributed to a mitochondrial effect. Moreover, mitochondria account for most of the cellular oxygen consumption. These observations are consistent with the cytotoxicity of DCVC, and, therefore, of DCVG and DCVCG, being due to mitochondrial effects. This view is supported by the finding that DCVC lowered the cellular energy charge, which is highly dependent on mitochondrial respiration, and did not appear to produce an oxidative stress in renal proximal tubular cells.
In conclusion, this study demonstrates the usefulness of isolated rat kidney proximal tubular cells in the investigation of the mechanism of glutathione and cysteine S-conjugateinduced nephrotoxicity. More importantly, the studies reported herein validate and extend the hypothesis asserted above that the cytotoxicity of glutathione S-conjugates is dependent on metabolism to the corresponding cysteine Sconjugates, which undergo bioactivation by cysteine conjugate P-lyase to produce a reactive thiol, the presumed ultimate reactive species. Finally, the observation that the cytotoxicity of cysteine S-conjugates is associated with mitochondrial dysfunction, rather than with the production of oxidative stress, warrants detailed investigation.