CYP2E1 overexpression in HepG2 cells induces glutathione synthesis by transcriptional activation of gamma-glutamylcysteine synthetase.

Induction of CYP2E1 (cytochrome P450 2E1) by ethanol appears to be one of the central pathways by which ethanol generates a state of oxidative stress. CYP2E1 is a loosely coupled enzyme; formation of reactive oxygen species occurs even in the absence of added substrate. GSH is critical for preserving the proper cellular redox balance and for its role as a cellular protectant. Since cells must maintain optimal GSH levels to cope with a variety of stresses, the goal of this study was to characterize the GSH homeostasis in human hepatocarcinoma cells (HepG2) that overexpress CYP2E1. This study was prompted by the finding that toxicity in CYP2E1-overexpressing cells was markedly enhanced after GSH depletion by buthionine sulfoximine treatment. CYP2E1-overexpressing cells showed a 40-50% increase in intracellular H(2)O(2); a 30% increase in total GSH levels; a 50% increase in the GSH synthesis rate; and a 2-fold increase in gamma-glutamylcysteine synthetase heavy subunit (GCS-HS) mRNA, the rate-limiting enzyme in GSH synthesis. This GCS-HS mRNA increase was due to increased synthesis since nuclear run-on assays showed increased transcription in CYP2E1-expressing cells, and the GCS-HS mRNA decay after actinomycin D treatment was similar in CYP2E1-expressing cells and empty vector-transfected cells. The facts that treatment with GSH ethyl ester almost completely prevented the increase in GCS-HS mRNA and decreased H(2)O(2) levels and that transient transfection with catalase (but not manganese-superoxide dismutase) produced a decrease in GCS-HS mRNA only in CYP2E1-expressing cells suggest a possible role for H(2)O(2) in the induction of GCS-HS gene transcription. In contrast to results with HepG2 cells expressing CYP2E1, no increase in GCS-HS mRNA was found with a HepG2 cell line engineered to express human cytochrome P450 3A4. In summary, CYP2E1 overexpression in HepG2 cells up-regulates the levels of reduced GSH by transcriptional activation of GCS-HS; this may reflect an adaptive mechanism to remove CYP2E1-derived oxidants such as H(2)O(2).

GSH, the most abundant antioxidant in cells, plays a major role in the defense against oxidative stress-induced cell injury and is essential in maintenance of intracellular redox balance and the essential thiol status of proteins (1). In the liver, the cellular GSH level is determined by a balance between the rate of its synthesis and the rates of its utilization (conjugation) and loss (export) (2). Endogenous hydrogen peroxide (H 2 O 2 ) is removed by GSH in the presence of glutathione peroxidase and can also be removed by catalase in the peroxisomes. Furthermore, conjugation with GSH, catalyzed by glutathione S-transferases (GST), 1 is an integral step in the detoxification and elimination of diverse classes of toxic chemical compounds (3). Either glutathione peroxidase or GST can reduce organic hydroperoxides in the presence of GSH. Exposure of cells to a number of xenobiotic agents (4 -8) and cytokines (9,10) has been shown to result in an increase in the total intracellular GSH content. The first step in GSH biosynthesis is rate-limiting and is catalyzed by ␥-glutamylcysteine synthetase (GCS). The GCS holoenzyme is a heterodimer composed of a heavy (GCS-HS; M r ϳ 70,000) and a light (GCS-LS; M r ϳ 30,000) subunit (11), which are encoded by different genes located on chromosomes 1 and 6, respectively. Changes in GCS activity have been reported to result from regulation at the transcriptional (9 -12) and/or post-transcriptional (13) level, and these changes may affect only the heavy or both the heavy and light subunits (13,14). Most of the studies reported in the literature have examined transcriptional regulation of GCS-HS.
Induction of CYP2E1 (cytochrome P450 2E1) and formation of reactive oxygen species (ROS) and lipid peroxidation derivatives appear to be one of the mechanisms by which ethanol is hepatotoxic (15,16). It has been demonstrated that CYP2E1, when reduced by NADPH-cytochrome P450 reductase, is a loosely coupled enzyme that displays high NADPH oxidase activity (17,18). Formation of reactive oxygen species can occur even in the absence of added substrates, e.g. formation of superoxide and H 2 O 2 by microsomes from CYP2E1-expressing cells was not altered by addition of substrates and ligands of CYP2E1 (19).
To directly demonstrate that ROS generated by CYP2E1 can promote hepatotoxicity even in the absence of substrate, a HepG2 cell line with enhanced expression of CYP2E1 was recently established (20). The CYP2E1-expressing cells (E47) have a slower growth rate than the empty vector-transfected control C34 cells when cellular GSH levels are maintained, but display a loss of cell viability when GSH synthesis is inhibited by treatment with BSO. However, we observed that despite increased production of ROS, GSH levels are not decreased in E47 cells compared with C34 cells. This suggested that a possible up-regulation of GSH synthesis and/or increased GSH turnover may occur in E47 cells in response to the increased oxidative stress. The goal of this study was to evaluate whether CYP2E1 overexpression could mediate an effect on GSH homeostasis.

MATERIALS AND METHODS
In Vitro Model and Cell Culture Conditions-Two human hepatoma HepG2 sublines, which were established previously in our laboratory (20), were used as a model in this study. E47 and E43 cells contain the human CYP2E1 cDNA (kindly provided by Dr. F. J. Gonzalez, NCI, National Institutes of Health) inserted into the EcoRI restriction site of the pCI-neo expression vector (Promega, Madison, WI) in the sense orientation. C34 cells contain the pCI-neo vector alone. The HepG2 transduced clones C34 and E47 were cultured in minimal essential medium supplemented with 10% fetal calf serum, 100 units/ml penicillin, 100 g/ml streptomycin, and 2 mM glutamine in a humidified atmosphere in 5% CO 2 at 37°C. Cells were maintained in the presence of 0.5 mg/ml of Geneticin.
Most reagents were purchased from Sigma. The protein content of cell lysates was determined by the method of Lowry et al. (21) using the DC-20 protein assay kit (Bio-Rad), and 2Ј,7Ј-dichlorofluorescein diacetate (DCFDA) was purchased from Molecular Probes, Inc. (Eugene, OR). Specific reagents are described below. CYP2E1 levels were routinely monitored by assaying the oxidation of p-nitrophenol to p-nitrocatechol.
GSH Assay-Cells (1 ϫ 10 6 ) were seeded onto 100-mm plates and incubated overnight before any treatment. Cells were washed twice with 1ϫ PBS, detached by trypsinization, and treated with 10% (1:1, v/v) trichloroacetic acid to extract cellular GSH. The mixture was centrifuged at 13,000 ϫ g for 1 min to remove denatured proteins, and GSH was determined by the enzymatic method of Tietze (22). The GSH content was assayed by following the increase in absorbance at 412 nm for 2 min in a cuvette containing 0.1 M sodium phosphate buffer (pH 7.5), 5 M EDTA, 0.6 mM 5,5Ј-dithiobis(2-nitrobenzoic acid), 0.2 mM NADPH, 1 unit/ml glutathione reductase, and 10 l of sample (corresponding to ϳ100 g of protein) in a final volume of 1 ml. For determination of GSSG, the same 5,5Ј-dithiobis(2-nitrobenzoic acid) recycling assay was performed after using 2-vinylpyridine to remove reduced GSH (23). Briefly, 2 l of 2-vinylpyridine and 6 l of triethanolamine were simultaneously mixed with 100 l of sample, followed by incubation in the dark at room temperature for 1 h before initiating the recycling assay. The kinetics of the reaction were monitored for 10 min. The increment in absorbance at 412 nm was converted to GSH concentration using a standard curve with known amounts of GSH.
DCF Fluorescence as a Measure of Hydrogen Peroxide Production-Cells (1 ϫ 10 6 ) were seeded onto 100-mm plates and incubated overnight before any treatment. DCFDA was added at a final concentration of 2 g/ml, and plates were incubated for 30 min at 37°C in the dark. Cells were washed twice with 1ϫ PBS, trypsinized, and resuspended in 3 ml of 1ϫ PBS; and fluorescence was immediately read in a Perkin-Elmer 650-10S fluorescence spectrometer at 490 nm for excitation and 525 nm for emission with a slit width of 5 nm for both excitation and emission monochromators. Background readings from cells incubated without DCFDA were substracted. Results are expressed as arbitrary units of fluorescence.
GSH Synthesis Rate Using Monochlorobimane-Monochlorobimane can be used to measure activities of enzymes involved in GSH synthesis (24). Monochlorobimane forms a fluorescent adduct with GSH in a reaction catalyzed by GST. Cells were grown in 75-cm 2 flasks to 80% confluence, washed twice with 1ϫ PBS, scraped off in 0.01 M sodium phosphate plus 0.25 M sucrose (pH 7.4), sonicated for 3 s in a W220 sonicator (Ultrasonic, Farmingdale, NY), and centrifuged at 100,000 ϫ g for 60 min. Cell extracts were dialyzed overnight at 4°C to deplete the cytosolic GSH content in order to minimize feedback inhibition of GSH on GCS. The synthesis of GSH was followed at 37°C in a cuvette containing a mixture of 100 mM Tris-HCl, 150 mM KCl, 20 mM MgCl 2 , 2 mM EDTA, 2 mM ATP, 10 mM glutamate, 10 mM glycine, and 0.4 mM cysteine plus 0.4 mM dithiothreitol and 100 M monochlorobimane in a final volume of 2.5 ml at pH 7.3. The reaction was started by addition of cytosol (100 l, 1-10 mg of protein) that was or was not preincubated with BSO (20 mM for 20 min) to specifically inhibit GCS activity, and formation of the GSH-bimane fluorescent complex was monitored over time at 37°C using the Perkin-Elmer 650-10S fluorescence spectrometer with excitation at 385 nm and emission at 478 nm with a slit width of 5 nm for both excitation and emission monochromators. The difference in the rate of increase in fluorescence in the absence and presence of BSO is equivalent to the rate of GSH synthesis. The change in fluorescence was converted to GSH concentration using standard curves. Specifically, formation of fluorescent adduct was monitored by adding monochlorobimane (100 M) and GST (0.1 unit/ml) to a cuvette containing GSH standards.
GSH Efflux-To study GSH efflux, C34 and E47 cells were seeded onto 100-mm plastic dishes at 37°C in a 5% CO 2 and 95% air atmosphere and studied when confluency was almost reached. To measure GSH efflux, the cells were first incubated for 1 h with 0.5 mM acivicin to inhibit the enzyme ␥-glutamyl transpeptidase (25). The cells were then washed twice with Krebs-Henseleit buffer supplemented with 12.5 mM HEPES (pH 7.4) and incubated in 4 ml of Krebs buffer at 37°C for 90 min. Aliquots of the supernatant (500 l) were removed at 5, 30, 60, and 90 min, and GSH was determined as described above.
Enzyme Assays-Glutathione peroxidase was assayed according to ; and the decrease in absorbance was monitored at 340 nm for 2 min. One unit of glutathione reductase is defined as the amount of enzyme that catalyzes the oxidation of 1 mol of NADPH/min/mg of protein. Total superoxide dismutase activity was measured according to Paoletti and Mocali (29). One unit of superoxide dismutase activity is defined as the amount of cell extract required to inhibit the rate of NADPH oxidation of the control by 50%. The control consisted of a sample in which the cell extract was omitted from the reaction mixture. The following solutions were sequentially added to a cuvette: 0.8 ml of 100 mM triethanolamine/diethanolamine (pH 7.4), 40 l of 7.5 mM NADPH, 25 l of 100 mM EDTA plus 50 mM MnCl 2 , and 100 l of dialyzed cell extract (ϳ0.5 mg of protein) or buffer. The contents were mixed, and the absorbance was read at 340 nm against air until a stable base line was observed as recorded over a 5-min period. Then, 100 l of 10 mM ␤-mercaptoethanol was added, followed by mixing; and the decrease in absorbance at 340 nm was monitored for over a 20-min reaction period. Measurements of the relative rates for both the control and cell extract were made from the linear portion of the curves.
Northern Blotting-Total RNA was isolated using the Trizol reagent (Life Technologies, Inc.). Ten micrograms of RNA was electrophoresed under denaturing conditions on formaldehyde-containing 0.9% agarose gels and transferred onto GeneScreen nylon membranes (NEN Life Science Products). RNA was UV-cross-linked at 1200 J ϫ 100 for 30 s and prehybridized in a hybridization oven for 2 h at 42°C in 5ϫ SSC, 50% formamide, 50 mM potassium phosphate (pH 6.5), and 1ϫ Denhardt's solution containing 0.1 mg/ml denatured salmon sperm DNA. cDNA probes were radiolabeled with [␣-32 P]dCTP (specific activity Ͼ 3000 Ci/mmol; Amersham Pharmacia Biotech) using a random primer DNA labeling kit (Amersham Pharmacia Biotech). Immobilized RNA was hybridized in 0.1% SDS, 5ϫ SSC, 50% formamide, 50 mM potassium phosphate (pH 6.5), 1ϫ Denhardt's solution, and 12.5ϫ dextran sulfate with 10 6 cpm/ml labeled cDNA and 0.1 mg/ml denatured salmon sperm DNA for 16 h at 42°C. Blots were washed four times for 15 min with 2ϫ SSC and 0.5% SDS at 65°C, exposed to a phosphor storage screen for a fixed period of time according to the hybridized cDNA, and scanned in a PhosphorImager (Molecular Dynamics, Inc., Sunnyvale, CA). Values were normalized to the housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA using ImageQuant software. The GCS-HS cDNA was obtained from American Type Culture Collection (ATCC 1064067, Manassas, VA). The GCS-LS probe was obtained by reverse transcription-polymerase chain reaction amplification of a fragment of 513 base pairs with the primers 5Ј-GGA GTT TCC AGA TGT CTT GG-3Ј and 5Ј-GGA ATG CTT TCC TGA AGA GC-3Ј; the complete GCS-LS cDNA sequence was obtained from a published sequence (GenBank TM /EBI accession L35546) (30).
Nuclear in Vitro Run-on Transcription Assays-Cultured cells were washed twice with sterile 1ϫ PBS, scraped off, and transferred into 15-ml Falcon tubes for centrifugation at 500 ϫ g at 4°C for 5 min. Pellets were loosened and resuspended in 4 ml of Buffer A (10 mM Tris (pH 7.4), 10 mM NaCl, and 3 mM MgCl 2 ). After addition of 0.5% (v/v) Nonidet P-40, samples were vortexed for 10 s and kept on ice for 10 min. After a second centrifugation, supernatants containing cytoplasmic RNA were discarded, and 4 ml of Buffer A was added. Samples were kept on ice for 10 min and spun down. Nuclei were collected in 100 l of storage buffer (50 mM Tris (pH 8.3), 40% glycerol, 5 mM MgCl 2 , and 0.1 mM EDTA) and immediately stored at Ϫ70°C until analyzed.
Transcription was assayed in vitro with 100 l of nuclei (isolated from ϳ25 ϫ 10 6 cells), 100 Ci of [ 32 P]UTP, and 100 l of assay buffer (2.5 mM MnCl 2 , 125 mM Tris (pH 7.4), 12.5 mM MgCl 2 , 2.5 mM dithiothreitol, 2.5 mM ATP, 1.25 mM GTP, 1.25 mM CTP, and 375 mM KCl). The reaction was carried out at 25°C for 45 min and stopped by adding DNase I (RNase-free) at a final concentration of 20 g/ml. After incubation at 30°C for 5 min, RNA was released from the nuclei by treatment with 1% SDS, 5 mM EDTA, 10 mM HEPES (pH 7.5), and 200 g/ml proteinase K at 42°C for 30 min. Newly transcribed RNA was isolated using the Trizol reagent.
Eight micrograms of each cDNA of interest was denatured in 0.25 N NaOH and 0.5 M NaCl for 10 min at room temperature and diluted in 0.1ϫ SSC and 0.125 N NaOH. Samples were blotted onto GeneScreen membranes, neutralized in 0.5 N NaCl and 0.5 M Tris (pH 7.5), and UV-cross-linked. Prehybridization was carried out for 6 h at 42°C, and purified denatured riboprobes (1 ϫ 10 6 cpm/ml) were used for hybridization for 72 h at 42°C. The composition of the hybridization buffer was the same as described for the Northern blot assay, using 0.1 mg/ml bakers' yeast tRNA instead of the salmon sperm DNA. After four washes of 15 min each with 2ϫ SSC and 0.5% SDS at 65°C, membranes were exposed to the phosphor storage screen, and the signals were quantified using the GAPDH mRNA signal as a control.
Transient Transfection Experiments-Transient transfection experiments were carried out with pCI vector containing CYP2E1 in the antisense (pCI-AS-2E1) or sense (pCI-2E1) orientation, pcDNA3-MnSOD (kindly provided by Dr. Larry Oberley, University of Iowa, Iowa City, IA), or pZeoSV2-catalase (kindly provided by Dr. A. Melendez, Albany Medical College, Albany, NY). Controls for these experiments included transfections with the empty vectors pCI, pZeoSV2, and pcDNA3. Transfection of C34 and E47 cells was carried out with 10 g of the appropriate plasmid, 3 g of pL7-Lac (kindly provided by Dr. Robert Krauss, Mount Sinai School of Medicine, New York, NY), and 30 l of Fugene 6 (Roche Molecular Biochemicals). After 48 h, the efficiency of the transfection was determined by X-gal staining; and at 48 and 96 h, total RNA was isolated to study GCS-HS and GAPDH mRNA expression by Northern blotting.
Statistics-Results refer to means Ϯ S.D. and are average values from three to six values per experiment; experiments were repeated at least twice. Statistical evaluation among groups was carried out using Student's t test, and p Ͻ 0.05 was considered significant.

GSH Levels and Intracellular H 2 O 2
Production-GSH is among the most important intracellular antioxidants. Since ROS generated from CYP2E1 or other sources can be removed either by direct reaction with GSH (e.g. the hydroxyl radical) or by the glutathione peroxidase reaction (e.g. H 2 O 2 ), we investigated the effect of CYP2E1 overexpression on the regulation of total cellular GSH equivalents (GSH ϩ GSSG) and the generation of hydrogen peroxide as an indicator of ROS. Hydrogen peroxide was monitored by quantification of fluorescence of DCF, a highly fluorescent probe sensitive to peroxides. BSO is an effective inhibitor of ␥-glutamylcysteine synthetase, the rate-limiting enzyme for GSH synthesis. The intracellular GSH levels in E47 and C34 cells with or without BSO treatment were evaluated (Fig. 1). In the absence of BSO treatment, E47 cells cultured in normal minimal essential medium had a significant 30% increase in total GSH compared with C34 cells (Fig. 1A); GSSG levels were very low and similar in both cell lines (data not shown). As expected, BSO treatment caused a decline in intracellular GSH in both cell lines; although the fall in intracellular GSH in E47 cells was linear over the 8-h reaction period, the decline in GSH in C34 cells showed a lag of ϳ4 h. Thereafter, the fall in GSH was similar for E47 and C34 cells (Fig. 1B). To determine if the decline in GSH levels was due, at least in part, to GSH efflux from the cells, the GSH content in the medium was assayed. GSH efflux could not be observed unless acivicin, an inhibitor of ␥-glutamyl transpeptidase was added, confirming reports of others (25). The initial levels of GSH in the medium of E47 cells were ϳ25% higher than the levels in the medium of C34 cells, most likely a reflection of the 30% higher intracellular GSH levels in E47 cells. The rate of GSH efflux, as assayed over a 90-min time period (Fig. 1C), was 36.0 Ϯ 4.0 pmol/min/mg of protein in C34 cells, whereas the rate in E47 cells was 50.3 Ϯ 3.5 pmol/min/mg of protein (ϳ40% higher). From the rate of GSH efflux, ϳ6 nmol of GSH/mg of protein would be released from E47 cells in 2 h, which is equivalent

FIG. 1. Intracellular GSH levels in C34 and E47 cells (A), time course for the effect of BSO on GSH levels (B), and GSH efflux (C).
A, steady-state GSH levels were determined as described under "Materials and Methods." B, C34 and E47 cells were treated with 0.1 mM BSO for the times indicated. Thereafter, the medium was removed, and cells were trypsinized and treated with 10% trichloroacetic acid. GSH was determined, and the results are expressed as nmol of GSH consumed after BSO treatment. C, GSH efflux was measured after inhibition of ␥-glutamyl transpeptidase by acivicin pretreatment (0.5 mM for 1 h). The results are expressed as means Ϯ S.D. (n ϭ 6). *, p Ͻ 0.05 versus C34 cells. prot, protein.
to the loss of intracellular GSH in 2 h (Fig. 1B). Thus, most of the decline in intracellular GSH in E47 cells, after BSO treatment to prevent new synthesis, is due to GSH efflux from the cells. Similar calculations with C34 cells indicated that ϳ4 nmol of GSH/mg of protein would be excreted from the cells in 2 h, which is greater than the 1-nmol decline in intracellular GSH found at 2 h. The reasons for this discrepancy and the lag in the fall in GSH in C34 cells are not known.
As shown in Fig. 2, E47 cells labeled with DCFDA displayed a significant 40 -50% increase in DCF fluorescence compared with C34 cells; this fluorescence was further elevated after treatment with BSO for 24 -48 h, indicating the critical role of GSH in detoxification of ROS generated by CYP2E1 in E47 cells. It is interesting that increased DCF fluorescence was observed in E47 cells even though GSH levels were elevated in these cells. The small increase in DCF fluorescence in C34 cells 24 -48 h after BSO treatment likely reflects the effectiveness of the residual GSH (ϳ20% of control values) in maintaining antioxidant function coupled with the lack of a significant oxidative stress in these cells as compared with E47 cells (20). Indeed, C34 cells remained fully viable after BSO treatment, in contrast to the loss of viability of E47 cells.
Antioxidant Enzyme Activities-To maintain the intracellular GSH levels, GSSG is reduced back to GSH by glutathione reductase at the expense of NADPH, or de novo synthesis of GSH occurs. Organic hydroperoxides can be reduced by either glutathione peroxidase or GST. Superoxide is converted to H 2 O 2 by superoxide dismutase. H 2 O 2 can be removed either by reduction by GSH in the cytosol as catalyzed by glutathione peroxidase or by catalase in the peroxisomes. As shown in Table I, there was no difference in the activities of glutathione reductase and superoxide dismutase in C34 and E47 cells. There was a 30% decrease in glutathione peroxidase activity in cells overexpressing CYP2E1. However, there was a 2-fold increase in GST activity in E47 cells compared with C34 cells. Thus, there appears to be a mixed response by enzymes involved in GSH homeostasis to the presence of CYP2E1 in E47 cells, with increases (GST), decreases (glutathione peroxidase), or no change (glutathione reductase) being observed.
Effect of CYP2E1 on GCS-HS mRNA Expression-In view of these results, we investigated possible mechanisms that could account for the up-regulation of GSH levels in E47 cells. The first step of GSH biosynthesis is rate-limiting and catalyzed by GCS. GCS is composed of a heavy (GCS-HS; M r ϳ 73,000) and a light (GCS-LS; M r ϳ 30,000) subunit, which are encoded by different genes. The heavy subunit exhibits all the catalytic activity as well as feedback inhibition by GSH. The light subunit alone is inactive, but allows the holoenzyme to be catalytically more efficient and less subject to inhibition by GSH than the heavy subunit alone. Northern blots of GCS-HS and GAPDH (used as a loading control) mRNA levels are shown in Fig. 3A. Control cells (C34) expressed two transcripts of GCS-HS mRNA with sizes of 4.1 and 3.2 kilobases. CYP2E1expressing cells (E47) showed a 2-fold increase in these two transcripts of GCS-HS mRNA. E43 cells, a HepG2 cell line that was developed at the same time as E47 cells and that expresses similar levels of CYP2E1, also showed a 2-fold induction of GCS-HS mRNA (Fig. 3B). The ratio of GCS-HS to GAPDH mRNA was assigned an arbitrary value of 1 in C34 cells. This ratio was increased ϳ2-fold in E47 and E43 cells. Northern blotting for GCS-LS was also carried out (Fig. 3C). A 1.4kilobase transcript was detected in C34 cells, and there was a 2-fold increase in this transcript in E47 cells. Thus, there appears to be coordinate regulation of the mRNAs for both subunits of GCS in E47 cells. An increase in GCS-HS mRNA may reflect either an increased transcription rate of the GCS-HS gene or a stabilization of the mRNA. The half-life of GCS-HS mRNA in C34 and E47 cells was assessed in the presence of actinomycin D (10 g/ml). As shown in Fig. 4, the half-life in C34 and E47 cells was similar (ϳ4 h), indicating that post-transcriptional modification of GCS-HS mRNA cannot account for the increase in GCS-HS mRNA in CYP2E1-overexpressing cells. To demonstrate increased transcription of the GCS-HS gene, nuclear run-on experiments were carried out. Indeed, the nuclear run-on revealed an increased capacity of E47 cells to transcribe the GCS-HS gene (Fig. 5). The ratio of newly synthesized GCS-HS to GAPDH mRNA was elevated from a value of 1 in C34 cells to a value of 1.7 in E47 cells. This increase in transcription under steady-state conditions is similar to the increase in mRNA levels as detected by Northern blotting.
GSH Synthesis Rate-To determine if the increase in GCS-HS mRNA was correlated with a higher enzyme activity, we determined the de novo capacity to synthesize GSH. Cytosolic fractions from CYP2E1-overexpressing cells showed a 50% increase in the rate of GSH synthesis as detected using monochlorobimane as a probe; the rate of GSH synthesis in C34 cells was 0.235 Ϯ 0.05 nmol of GSH/min/mg  of protein, whereas the rate in E47 cells was 0.360 Ϯ 0.04 (n ϭ 3; p Ͻ 0.05).
Effect of Glutathione Ethyl Ester-To further characterize the nature of CYP2E1-induced transcriptional activation of GCS-HS mRNA, GSH, which may act as an antioxidant or feedback regulator of GSH synthesis, was added to the culture medium, and the effect on the GCS-HS mRNA levels was determined. Glutathione ethyl ester (GSH-EE) has been reported to be effectively transported into human cells and converted intracellularly to GSH (31) and therefore was used as the source of added GSH. As shown in Fig. 6A, GSH-EE increased the content of GSH in C34 cells by 50%, but produced only a 14% increase in the already elevated GSH levels in E47 cells; in fact, GSH-EE equalized the GSH content in the two cell lines. Importantly, GSH-EE lowered DCF fluorescence in E47 cells to the level found in C34 cells (Fig. 6B). The quenching of ROS as reflected by the decreased DCF fluorescence by added GSH-EE may account for the lack of a significant enhancement of intracellular GSH levels in E47 cells by added GSH-EE. GSH-EE had no effect on GCS-HS mRNA levels in C34 cells, but largely prevented the increase in this mRNA found in E47 cells (Fig. 6C), consistent with the decrease in DCF fluorescence that GSH-EE produced in these cells. Because GSH-EE generates ethanol, controls were treated with an equimolar (2.5 mM) dose of ethanol. This dose of ethanol did not produce any effect on the parameters measured.
Transfection with Catalase Decreases GCS-HS mRNA Ex- pression in E47 Cells-Several antioxidants, primarily preventive against lipid peroxidation, such as trolox, vitamin E, vitamin C, uric acid, and the iron chelator desferrioxamine, did not lower the elevated GCS-HS mRNA levels in E47 cells (data not shown). The effectiveness of GSH-EE suggested the possibility that H 2 O 2 may be a critical ROS important in the increase in GCS-HS mRNA found in E47 cells. We therefore transiently transfected C34 and E47 cells with catalase and, as an additional control, besides empty vector, with manganese-superoxide dismutase. Plasmids pcDNA3-MnSOD and pZeoSV2-catalase and control plasmids pZeoSV2 and pcDNA3 were used for transient transfection (Fig. 7A). Transfection with catalase produced an ϳ30% decrease in the GCS-HS mRNA content in CYP2E1-expressing cells. Manganese-superoxide dismutase had no effect on the levels of GCS-HS mRNA (data not shown). There was no effect of catalase (Fig. 7A) or manganese-superoxide dismutase on the GCS-HS mRNA levels in C34 cells. The efficiency of the transfections, assessed by X-gal staining, was 25% in both cell lines.
Role of CYP2E1 in the Induction of GCS-HS mRNA-The only apparent difference between E47 and C34 cells is the presence of CYP2E1. It appears that the induction of GCS-HS mRNA in E47 cells is caused by the expression of CYP2E1 in these cells. To validate this idea, we transiently transfected plasmids pCI, pCI-AS-2E1, and pCI-2E1 into E47 and C34 cells (Fig. 7B). The presence of CYP2E1 in C34 cells, after transfection with pCI-2E1, caused a 20% increase in GCS-HS mRNA. pCI-AS-2E1 had no effect, consistent with the absence of CYP2E1 in C34 cells. In E47 cells, transfection with pCI-AS-2E1 decreased the levels of GCS-HS mRNA ϳ10%; however, there was a 30% increase in GCS-HS mRNA in cells transiently transfected with pCI-2E1. The efficiency of transfection was between 20 and 30% for all plasmids used with both cell lines.
Comparison between CYP2E1 and CYP3A4 in Mediating GCS-HS mRNA Induction-The control C34 cells, like the parental cell line, do not contain or express very low levels of P450 isoforms. To ascertain if induction of the GCS-HS gene in E47 cells could be due to the overexpression of any P450, we evaluated the ability of CYP3A4 (cytochrome P450 3A4) to increase GCS-HS mRNA levels. A HepG2 cell line that constitutively and stably expresses CYP3A4 has recently been developed by methods similar to those used for establishing E47 cells. 2 The CYP3A4-overexpressing HepG2 cells contain the human CYP3A4 cDNA (kindly provided by Dr. F. J. Gonzalez) inserted into the EcoRI site of the pCI-neo expression vector. Successful expression of CYP3A4 in these cells was demonstrated by Northern and Western blot analyses and by oxidation of a typical CYP3A4 substrate, fentanyl. 2 The content of CYP3A4 in these cells, ϳ40 pmol/mg of microsomal protein, was similar to the content of CYP2E1 in E47 cells. In two independent experiments of DCF fluorescence as detected by flow cytometry, ROS production with the CYP3A4-expressing cells was 22% greater than that found with the control C34 cells. By contrast, DCF fluorescence with E47 cells was 145% greater than that with control cells. A comparison of GCS-HS mRNA levels in C34, E47, and CYP3A4-expressing cells was made. As shown in Fig.  8, CYP3A4-expressing cells showed no increase in GCS-HS mRNA as compared with empty vector-transfected cells (C34). This result suggests that there may be some selectivity in the induction of the GCS-HS gene by the CYP2E1-expressing cells, at least as compared with one other P450 isoform. It is interesting to speculate that the lack of increase in GCS-HS mRNA in the CYP3A4-expressing cells, in contrast to the increase found in E47 cells, may reflect the similar rates of production of ROS, as detected by DCF fluorescence, in the CYP3A4-expressing cells and C34 cells. DISCUSSION As summarized in a recent publication (32), the effects of chronic ethanol treatment on hepatic levels of GSH are not clear, with reports of decreased GSH levels (33,34), unchanged levels (35), or even increased levels (36). A more consistent finding is that chronic ethanol treatment lowers mitochondrial GSH levels, due to defects in mitochondrial GSH transport (37). A recent study using the intragastric infusion model of ethanol administration reported that liver GSH levels declined by 40%, although GCS activity and GCS-HS mRNA levels increased 2-fold (32). The decline in GSH despite the induction of GCS was suggested to reflect other factors, including dietary, e.g. cysteine availability, important in determining the steadystate GSH levels. The above study (32) is the first report that chronic ethanol ingestion increased GCS-HS expression and activity. Among several biochemical alterations that occur in 2 D. E. Feierman, unpublished observations. FIG. 7. Effect of transfection with catalase or sense and antisense CYP2E1 in C34 and E47 cells. C34 and E47 cells were transfected with catalase (A) or sense and antisense CYP2E1 (B), and total RNA was isolated after 96 h. Northern blotting for GCS-HS and GAPDH was performed. Quantification of Northern blots, discussed under "Results," was performed as relative GCS-HS mRNA signal corrected for differences in loading using the GAPDH signal. The ratio of GCS-HS to GAPDH in C34 cells was assigned a value of 1.

FIG. 8. Comparison of induction of GCS-HS mRNA in HepG2 cells.
Northern blotting of GCS-HS and GAPDH was performed with RNA isolated from C34, E47, and CYP3A4-expressing cells. Quantification of Northern blots was performed as relative GCS-HS mRNA signal corrected for differences in loading using the GAPDH signal. The ratio of GCS-HS to GAPDH in C34 cells was assigned a value of 1. the intragastric infusion model are development of oxidative stress and striking increases in the content of CYP2E1. Whether CYP2E1 and oxidative stress play a role in the induction of GCS in this model remains to be determined. This led to the current investigation on the effect of CYP2E1 on glutathione homeostasis and possible induction of GCS activity and expression. An additional reason for the current study was that previous results showed that GSH appears to be essential in protecting HepG2 cells against CYP2E1-derived oxidative stress since BSO treatment, which depletes cellular GSH, resulted in a striking loss of viability of E47 cells, without any effect on the control C34 cells (20). E47 cells grow at a slower rate than the control C34 cells, yet despite this slow growth rate, E47 cells are fully viable as long as GSH levels are maintained. This suggested a possible up-regulation of GSH synthesis and/or increased turnover (20).
In the current study, an increase in GSH levels and in ROS production in CYP2E1-expressing HepG2 cells was observed. The results with DCFDA as a fluorogenic probe provide evidence for an increase in production of ROS in E47 cells, and the further increase after BSO treatment confirms the major role of GSH detoxification of ROS especially in CYP2E1-expressing cells. Although studies with isolated microsomes or reconstituted systems have demonstrated an increased ability of CYP2E1 to produce ROS (17)(18)(19), the current results extend this capability to intact cells. To our surprise, GSH levels were higher in E47 cells, even in the presence of increased production of ROS. This suggested the possibility that GSH synthesis might be elevated perhaps as an adaptive response by E47 cells to a state of oxidative stress. GSH efflux from E47 cells was also higher than that from C34 cells, consistent with an upregulation of GSH synthesis. These observations led to subsequent studies to determine why GSH levels are elevated in E47 cells.
Because all eukaryotic cells maintain high concentrations of GSH or its analogs (38), the maintenance of intracellular GSH levels is indispensable for normal cell function and survival. The ability to maintain cellular functions under conditions of oxidative stress depends on the rapid induction of protective antioxidant systems. GSH synthesis and GCS-HS mRNA transcription have been shown to be increased as an adaptive mechanism against different kinds of insults leading to oxidative stress (6) such as exposure to tumor necrosis factor (9), nitric oxide (10), 4-hydroxy-2-nonenal (13), ␤-naphthoflavone (7), and okadaic acid and menadione (39). The increase in GSH levels in E47 cells is due to an increase in de novo synthesis of GSH since this increase correlates with a 50% increase in GCS activity, the rate-limiting enzyme involved in GSH synthesis. The GCS-HS mRNA is increased 2-fold in cells overexpressing CYP2E1. Run-on transcription assays and mRNA stabilization assays were performed to show that this increase in GCS-HS mRNA was due to an increase in transcription. A 2-fold increase in GCS-LS mRNA is also observed in E47 cells, suggesting the possibility of coordinate induction of the two genes encoding the two subunits of GCS. These results suggest that CYP2E1 up-regulates the levels of reduced GSH by transcriptional activation of GCS genes, perhaps as an adaptive mechanism for removal of CYP2E1-derived ROS such as H 2 O 2 . To our knowledge, this appears to be the first report of a gene up-regulated by CYP2E1 (or CYP2E1-derived ROS; discussed below).
Increased levels of GSH can protect cells against oxidative stress by directly acting as an antioxidant or by serving as a substrate for removal of ROS or reactive metabolites by antioxidant enzymes. In this respect, it is interesting that the activity of GST was found to be increased in E47 cells. Prelim-inary experiments have shown that the elevated GST activity is associated with increased mRNA levels for the A1 and A2 Alpha GST isoforms; the Alpha GST isoforms are the major GST present in HepG2 cells (40). Thus, Alpha GST may be another gene up-regulated by CYP2E1. Recent experiments have confirmed the importance of human Alpha GST isoforms such as A1 and A2 in protecting cells against oxidative stress (41). There appears to be some selectivity in E47 cells for increasing the activity of enzymes important for GSH homeostasis, e.g. GST and GCS, as glutathione reductase activity is not altered, whereas glutathione peroxidase activity is decreased. The last is known to be sensitive to oxidative stress (42).
Since the cloning of the promoter of GCS-HS mRNA (43) To further evaluate the nature of CYP2E1 induction of the GCS-HS gene, the effect of antioxidants was studied. Lipid peroxidation reactions do not appear to play an important role in explaining the up-regulation of the GCS-HS gene since addition of classical inhibitors of lipid peroxidation does not prevent the increase in GSH content or GCS-HS mRNA. These agents were previously shown to prevent the toxicity by arachidonic acid or ferric nitrilotriacetate to E47 cells, demonstrating their effectiveness as antioxidants and inhibitors of lipid peroxidation (44,45). GSH-EE was effective in preventing the increase in GCS-HS mRNA in E47 cells. This action of GSH-EE appears to be due to detoxification of ROS as detected by decreased DCF fluorescence and not to feedback inhibition by elevated cellular levels of GSH because GSH-EE did not decrease the level of GCS-HS mRNA in C34 cells, although GSH levels were raised to the same value as in E47 cells. In E47 cells, GSH-EE did lower the DCF fluorescence to levels found in C34 cells; and under these conditions of lowered ROS levels, GCS-HS mRNA levels also decreased. These findings indicate that GSH is crucial in combating the oxidative stress derived from CYP2E1 overexpression.
As a second approach, transfection with plasmids containing the antioxidant enzyme catalase or manganese-superoxide dismutase was performed. Transfection with catalase resulted in a 30% decrease in GCS-HS mRNA levels in CYP2E1-expressing cells, whereas there was no effect in control cells. This suggests that H 2 O 2 plays an important role in CYP2E1-mediated induction of GCS, which is consistent with the enhanced DCF fluorescence and the effect of GSH-EE in E47 cells. Transfection with manganese-superoxide dismutase did not cause any change in the GCS-HS mRNA levels in E47 or C34 cells.
To provide further evidence for a direct role of CYP2E1 in the transcriptional activation of GCS-HS, transient transfection experiments were performed with vectors containing CYP2E1 in the sense (pCI-2E1) and antisense (pCI-AS-2E1) orientations and control pCI plasmid. In the pCI-2E1-transfected cells, there was a further 30% increase in the GCS-HS mRNA level in the CYP2E1-expressing cells (E47) and a 20% increase in the control cells (C34). Transfection with the plasmid pCI-AS-2E1 had no effect in C34 cells, but produced a slight decrease in GCS-HS mRNA in E47 cells. These findings show an association between levels of CYP2E1 expression and GCS-HS induction, as the levels of GCS-HS mRNA followed this order: E47 plus pCI-2E1 Ͼ E47 Ͼ E47 plus pCI-AS-2E1 Ͼ C34 plus pCI-2E1 Ͼ C34 ϭ C34 plus pCI-AS-2E1.
The finding that at least another cytochrome P450-overexpressing HepG2 cell line (CYP3A4-expressing cells) showed no induction of GCS-HS mRNA provides evidence for some specificity in the CYP2E1 induction of GCS-HS gene transcription in E47 cells and confirms the importance of the CYP2E1 isoform in mediating oxidative stress and toxicity. These results suggest the possibility that induction of CYP2E1 by ethanol and the resulting increased production of ROS may play a role in the induction of GCS when ethanol is administered in the intragastric infusion model (32). Experiments are in progress to evaluate the effect of in vivo induction of CYP2E1 on GCS expression and activity. Are the levels of CYP2E1 expressed in E47 cells of biological relevance? There is considerable variability in the content of CYP2E1 in human liver microsomes, e.g. oxidation of dimethylnitrosamine, a preferential substrate for CYP2E1, by several preparations of human liver ranged from 0.28 to 0.98 nmol/min/mg of microsomal protein (46). Parkinson (47) reported a CYP2E1 content of 22 pmol/mg of human microsomal protein; the content of CYP2E1 in E47 cells is ϳ50 pmol/mg of microsomal protein. The rate of oxidation of pnitrophenol by E47 cell microsomes is ϳ0.3-0.4 nmol/min/mg of protein, which is within the range of p-nitrophenol oxidation by human liver microsomes (0.4 -1 nmol/min/mg of protein). Thus, the level of CYP2E1 in E47 cells is within the physiological range for this P450 in human liver. Induction of GCS may have additional consequences beyond alcohol metabolism/effects and oxidative stress. Drugs such acetaminophen can be metabolized by several P450 isoforms, especially CYP2E1, to the reactive N-acetyl-p-benzoquinoneimine, which displays little toxicity when GSH is available for conjugation and detoxification (48 -54). Lowering of GSH levels promotes acetaminophen toxicity by preventing quenching of the N-acetyl-pbenzoquinoneimine. Acetaminophen is highly toxic after chronic alcohol treatment, probably a reflection of induction of CYP2E1 (and other P450 isoforms), and is especially toxic when GSH is depleted. Indeed, in a previous study with HepG2 E9 cells, which contain lower levels of CYP2E1 than E47 cells, acetaminophen produced a concentration-and time-dependent loss of cell viability only when GSH levels were lowered by treatment with BSO (55). Studies are currently in progress to assess acetaminophen toxicity in E47 cells, which display higher levels of CYP2E1, but also higher levels of GSH; the former should promote toxicity of acetaminophen, whereas the latter should, at least initially, be protective. Treatment with BSO would be expected to strikingly elevate acetaminophen toxicity in E47 cells.
In summary, CYP2E1 overexpression in HepG2 cells up-regulates the levels of reduced GSH, which may serve as an adaptive mechanism to attenuate the CYP2E1-derived oxidative stress. Transfection with catalase or addition of GSH-EE decreases the CYP2E1-mediated induction of GCS-HS, suggesting that this induction may be H 2 O 2 -mediated. There appears to be some selectivity in the response of the HepG2 cells to CYP2E1, as cells expressing CYP3A4 do not show altered levels of GCS-HS mRNA. These results further emphasize the critical role of GSH in protecting against CYP2E1-mediated oxidative stress and toxicity.