Cytotoxicity and apoptosis produced by arachidonic acid in Hep G2 cells overexpressing human cytochrome P4502E1.

The goal of the current study was to evaluate the effects of arachidonic acid, as a representative polyunsaturated fatty acid, on the viability of a Hep G2 cell line, which has been transduced to express human cytochrome P4502E1 (CYP2E1). Arachidonic acid produced a concentration- and time-dependent toxicity to Hep G2-MV2E1-9 cells, which express CYP2E1, but little or no toxicity was found with control Hep G2-MV-5 cells, which were infected with retrovirus lacking human CYP2E1 cDNA. In contrast to arachidonic acid, oleic acid was not toxic to the Hep G2-MV2E1-9 cells. The cytotoxicity of arachidonic acid appeared to involve a lipid peroxidation type of mechanism since toxicity was enhanced after depletion of cellular glutathione; formation of malondialdehyde and 4-hydroxy-2-nonenal was markedly elevated in the cells expressing CYP2E1, and toxicity was prevented by antioxidants such as alpha-tocopherol phosphate, 6-hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid (trolox), propylgallate, ascorbate, and diphenylphenylenediamine, and the iron chelator desferrioxamine. Transfection of the Hep G2-MV2E1-9 cells with plasmid containing CYP2E1 in the sense orientation enhanced the arachidonic acid toxicity, whereas transfection with plasmid containing CYP2E1 in the antisense orientation decreased toxicity. The CYP2E1-dependent arachidonic acid toxicity appeared to involve apoptosis, as demonstrated by terminal deoxynucleotidyltransferase-mediated dUTP nick end labeling and DNA laddering experiments. Trolox, which prevented toxicity of arachidonic acid, also prevented the apoptosis. Transfection with a plasmid containing bcl-2 resulted in complete protection against the CYP2E1-dependent arachidonic acid toxicity. It is proposed that elevated production of reactive oxygen intermediates by cells expressing CYP2E1 can cause lipid peroxidation, which subsequently promotes apoptosis and cell toxicity when the cells are enriched with polyunsaturated fatty acids such as arachidonic acid. The Hep G2-MV2E1-9 cells appear to be a valuable model to study interaction between CYP2E1, polyunsaturated fatty acids, reactive radicals, and the consequence of these interactions on cell viability and to reproduce several of the key features associated with ethanol hepatotoxicity in the intragastric infusion model of ethanol treatment.

The goal of the current study was to evaluate the effects of arachidonic acid, as a representative polyunsaturated fatty acid, on the viability of a Hep G2 cell line, which has been transduced to express human cytochrome P4502E1 (CYP2E1). Arachidonic acid produced a concentration-and time-dependent toxicity to Hep G2-MV2E1-9 cells, which express CYP2E1, but little or no toxicity was found with control Hep G2-MV-5 cells, which were infected with retrovirus lacking human CYP2E1 cDNA. In contrast to arachidonic acid, oleic acid was not toxic to the Hep G2-MV2E1-9 cells. The cytotoxicity of arachidonic acid appeared to involve a lipid peroxidation type of mechanism since toxicity was enhanced after depletion of cellular glutathione; formation of malondialdehyde and 4-hydroxy-2-nonenal was markedly elevated in the cells expressing CYP2E1, and toxicity was prevented by antioxidants such as ␣-tocopherol phosphate, 6-hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid (trolox), propylgallate, ascorbate, and diphenylphenylenediamine, and the iron chelator desferrioxamine. Transfection of the Hep G2-MV2E1-9 cells with plasmid containing CYP2E1 in the sense orientation enhanced the arachidonic acid toxicity, whereas transfection with plasmid containing CYP2E1 in the antisense orientation decreased toxicity. The CYP2E1-dependent arachidonic acid toxicity appeared to involve apoptosis, as demonstrated by terminal deoxynucleotidyltransferase-mediated dUTP nick end labeling and DNA laddering experiments. Trolox, which prevented toxicity of arachidonic acid, also prevented the apoptosis. Transfection with a plasmid containing bcl-2 resulted in complete protection against the CYP2E1dependent arachidonic acid toxicity. It is proposed that elevated production of reactive oxygen intermediates by cells expressing CYP2E1 can cause lipid peroxidation, which subsequently promotes apoptosis and cell toxicity when the cells are enriched with polyunsaturated fatty acids such as arachidonic acid. The Hep G2-MV2E1-9 cells appear to be a valuable model to study interaction between CYP2E1, polyunsaturated fatty acids, reactive radicals, and the consequence of these interactions on cell viability and to reproduce several of the key features associated with ethanol hepatotoxicity in the intragastric infusion model of ethanol treatment.
There is current interest in the role of oxidative stress and generation of reactive radical species in the mechanism(s) by which ethanol is toxic to the liver and other tissues (1). Induction of CYP2E1 1 by ethanol appears to be one of the central pathways by which ethanol is believed to generate a state of oxidative stress. Microsomes isolated from rats treated chronically with ethanol display increased production of superoxide radical, H 2 O 2 , hydroxyl radical, and enhanced lipid peroxidation (2)(3)(4)(5)(6)(7)(8). Ethanol oxidation to the 1-hydroxyethyl radical is also elevated after ethanol consumption (9,10). Increased formation of reactive radical species and lipid peroxides after chronic ethanol treatment are prevented by anti-CYP2E1 IgG and by chemical inhibitors of CYP2E1, thus linking these increases to induction of CYP2E1 (4,11,12).
The importance of dietary fat in alcoholic liver disease in humans is supported by epidemiological correlations which suggest that susceptibility to alcohol is related to different types of dietary fat (13,14). A major advance in ethanol hepatotoxicity studies has been the development of the intragastric infusion model of ethanol feeding, which leads to more significant liver injury than the classical liquid diets (15)(16)(17)(18)(19)(20). Liver injury occurs in this model when the rats consume diets containing polyunsaturated fatty acid (PUFA) but not saturated fatty acid. In these models, large increases in lipid peroxidation have been shown to correlate with CYP2E1 levels (15, 16, 18, 20 -22). The general hypothesis to account for the liver injury with this model is that elevated production of reactive radical species occurs due to induction of CYP2E1, and this results in lipid peroxidation when the diet is supplemented with PUFA (15, 16, 18, 20 -22).
In attempts to directly demonstrate that overexpression of CYP2E1 can result in hepatotoxicity of various agents, a Hep G2 cell line that constitutively expresses the human CYP2E1 was recently established (23). Electron spin resonance spectroscopy showed that microsomes from Hep G2-MV2E1-9 cells that express CYP2E1 produced superoxide radicals at rates about 10-fold greater than those from Hep G2-MV-5 cells that do not express CYP2E1; rates of H 2 O 2 production were about 3-fold greater with the Hep G2-MV2E1-9 microsomes. Rates of microsomal lipid peroxidation were also greater with the Hep G2-MV2E1-9 cells (23). Ethanol and acetaminophen were shown to be toxic to Hep G2-MV2E1-9 cells but not Hep G2-MV-5 cells (24,25). This model appears to be useful in efforts to establish a CYP2E1-dependent hepatotoxicity system and to evaluate the role of oxidative stress in the toxicity of compounds metabolized by CYP2E1.
Increased lipid peroxidation has been implicated as being associated with apoptosis, or programmed cell death. Direct exposure of various cell types to oxidants such as hydrogen peroxide or lipid hydroperoxides can directly induce apoptosis; in many experimental models pretreatment of the cells with antioxidants has been shown to protect against this form of cell death (26 -29). The prototypic regulator of mammalian apoptosis is the proto-oncogene bcl-2 (30). The functions of bcl-2 have been suggested to include acting as an antioxidant (31), modulating some aspects of nuclear transport (32), intervention in calcium signaling (33), and associating with several other proteins (34). Overexpression of bcl-2 leads to protection for many cell types against apoptosis induced by exposure to a wide variety of adverse conditions and stimuli, including lipid peroxidation, suggesting that bcl-2 controls a distal step in a signal transduction pathway leading to apoptosis (35)(36)(37)(38)(39)(40)(41)(42)(43).
The goal of the current study was to evaluate the cytotoxicity effects of arachidonic acid, a representative PUFA, to Hep G2 cells expressing CYP2E1 and to compare these effects to control cells that do not express CYP2E1. The effect of antioxidants and of bcl-2 on arachidonic acid toxicity and whether the toxicity was apoptotic in nature was also determined. It was hoped that this Hep G2 cell model might be a direct system that can establish linkage between CYP2E1, PUFA, oxidative stress, and cytotoxicity and thus mimic in a simple culture system the conditions believed to be representative of the gastric infusion model of ethanol toxicity.

MATERIALS AND METHODS
Cells and Chemicals-Hep G2-MV2E1-9 and Hep G2-MV-5 cells (23), human hepatocellular carcinoma Hep G2 sublines, were cultured in minimum essential medium (MEM), supplemented with 10% fetal calf serum, 100 units/ml penicillin, 100 mg/ml streptomycin, and 2 mM glutamine in a humidified atmosphere in 5% CO 2 at 37°C. Hep G2-MV2E1-9 cells contain a copy of the human CYP2E1 cDNA and constitutively express CYP2E1. Hep G2-MV-5 cells are the appropriate controls for Hep G2-MV2E1-9 cells as they contain only viral vector lacking the CYP2E1 cDNA. Most reagents were from Sigma. Specific reagents are described below.
Cytotoxicity Measurement: MTT Assay-Cytotoxicity of PUFA was primarily measured by the MTT assay (44). Tetrazolium salts such as MTT are metabolized by mitochondrial dehydrogenases to form a blue formazan dye and are therefore useful for the measurement of cytotoxicity. Approximately 2.0 -2.5 ϫ 10 4 cells, suspended in MEM containing 2 mM 4-methylpyrazole (added to stabilize CYP2E1 against degradation), were plated onto each well of a 24-well plate (Corning Co.) and incubated in 5% CO 2 at 37°C for 24 h. Test reagents, such as arachidonic acid and oleic acid, were then added to the culture medium for a designated preincubation time, typically 24 h. The culture medium was then replaced with normal MEM (without 4-methylpyrazole). After an additional incubation, typically 8 or 24 h, the medium was removed and cell viability was evaluated by the MTT assay, which was performed using the Promega Cell Titer 96 Non-radioactive Cell Proliferation Assay Kit. Briefly, 15% volume of dye solution was added to each well for a 1-h incubation at 37°C. An equal volume of solubilization/stop solution was then added to each well for an additional 1-h incubation. The absorbance of the reaction solution at 570 nm was recorded. The absorbance at 630 nm was used as reference. The net A 570 -A 630 was taken as the index of cell viability. The net absorbance from the wells of cells cultured with control medium was taken as the 100% viability value. The percent viability of the treated cells was calculated by the formula (A 570 Ϫ A 630 ) sample /(A 570 Ϫ A 630 ) control ϫ 100.
Cytotoxicity Measurement: LDH Assay-Leakage of lactate dehydrogenase (LDH) was measured as another index of cytotoxicity. Approximately 1-2 ϫ 10 6 cells were plated onto each well of a 6-well plate (Corning Co.) and incubated for 24 h. Cells were then treated with arachidonic acid or oleic acid for a 24-h period, followed by a second 24-h incubation period in the absence of fatty acid as described above. At the end of treatment, the combined media were collected to measure LDH activity (referred to as LDH out ). Cells were harvested by scraping, washed with PBS, suspended in 1 ml of PBS, and sonicated by using a Heat Systems-Ultrasonics Model W-375 Sonicator (5 s, duty cycle 25%, output control 40%). The LDH activity of the total cell lysate was measured (referred to as LDH in ). Lactate Dehydrogenase Assay Kit LD-L20 (Sigma) was used for the quantitative kinetic determination of LDH activity. The reagent contains 50 mM lactate plus 7 mM NAD ϩ in a pH 8.9 buffer system. To determine the LDH activity, 50 -200-l aliquots of cell tissue culture medium or of cell lysates were added to the LDH assay system, and the increase in absorbance at 340 nm due to NADH formation was recorded. The cytotoxicity index was expressed as the ratio of LDH out /LDH in .
TUNEL Analysis-Apoptosis in individual cells was assessed by terminal deoxynucleotidyltransferase-mediated dUTP nick end labeling (TUNEL) techniques as described by Gavrieli et al. (45) and Portera-Cailliau et al. (46) with modifications. Briefly, 5 ϫ 10 5 Hep G2-MV2E1-9 or Hep G2-MV-5 cells were plated onto each well of 6-well culture plates. After incubation with or without arachidonic acid, cells were washed twice with PBS ϩ 1% bovine serum albumin at 4°C, adjusted to a concentration of 0.2 ϫ 10 7 per 0.2 ml of PBS buffer, and fixed with 0.1 ml of freshly prepared 4% paraformaldehyde solution (in PBS, pH 7.4) for 30 min at room temperature. Cells were washed twice with PBS ϩ 1% bovine serum albumin and resuspended in 0.1 ml of permeabilization solution (0.1% Triton® X-100 in 0.1% sodium citrate) for 2 min on ice, followed by washing twice with PBS ϩ 1% bovine serum albumin. Cells were then resuspended in 50 l of TUNEL reaction mixture or label solution (without terminal transferase) as negative control, incubated for 60 min at 37°C in a humidified atmosphere in the dark, followed by washing twice in PBS ϩ 1% bovine serum albumin. Cells were analyzed by flow cytometry (EPICS® Profile Analyzer, Coulter Corp.).
DNA Agarose Gel Electrophoresis-DNA fragmentation was determined to evaluate apoptosis (47)(48)(49)(50). After incubation with or without arachidonic acid, cells were scraped off the 6-well culture plates with culture medium and were centrifuged at 1,200 rpm ϫ 10 min. The cell pellets were resuspended in 1 ml of lysis buffer (10 mM Tris-HCl, pH 7.5, 10 mM NaCl, 10 mM EDTA, 100 g/ml proteinase K, and 0.5% SDS) and incubated for 1 h at 50°C. After lysis, samples were extracted with 2 ml of phenol (neutralized with TE buffer, pH 7.5), followed by extraction with 1 ml of chloroform/isoamyl alcohol (24:1). The aqueous supernatants were precipitated with 2.5 volumes of ice-cold ethanol plus 10% volume of 3 M sodium acetate, pH 5.2, at Ϫ20°C overnight. After centrifugation at 13,000 ϫ g for 10 min, the pellets were air-dried, resuspended with 50 l of TE buffer, pH 7.5, supplemented with 0.1 g/ml RNase A, and electrophoretically separated on a 1.5% agarose gel in 0.5 ϫ TBE buffer containing 1 g/ml ethidium bromide at 100 V for 2 h. Pictures of the gels were taken by UV transillumination.
Transduction of Hep G2-MV2E1-9 Cells-The full-length human bcl-2 cDNA, excised from pSFFV-bcl-2 expression vector (kindly provided by Dr. George Acs and Dr. Beatriz Pogo, Mount Sinai School of Medicine, NY), and human CYP2E1 cDNA, excised from p91023(B)-2E1 (kindly provided by Dr. F. J. Gonzalez, National Cancer Institute, Bethesda), were inserted into the EcoRI restriction site of pCI-neo expression vector (Promega), in the sense and antisense (as) orientations to form pCI-bcl-2 or pCI-as-bcl-2, and pCI-2E1 or pCI-as-2E1, respectively. Transfection of Hep G2-MV2E1-9 cells was carried out by utilizing the LipofectAMINE reagent (Life Technologies, Inc.) as described by Hawley-Nelson et al. (51). Hep G2-MV2E1-9 cells were grown to 80 -90% confluence and harvested by trypsinization, and 1.5 ϫ 10 6 cells were seeded into a 100-mm culture dish and grown until 50 -70% confluence. Cells were rinsed with serum-free MEM before transfection. Solution A (15 g of the appropriate plasmid DNA in 800 l of serumfree MEM) and solution B (100 l of LipofectAMINE reagent in 800 l of serum-free MEM) were gently mixed and incubated at room temperature for 30 min to form a DNA-liposome complex. The complex was diluted with 6.4 ml of MEM, added to the Petri dish containing the Hep G2-MV2E1-9 cells, followed by incubation for 5 h at 37°C in a CO 2 incubator. 8 ml of MEM with 20% fetal calf serum was then added to each culture dish. After 18 h of incubation, fresh MEM was added, and the cells were incubated for an additional 2 days. The cells were collected by trypsinization and used for Western blot analysis and for studies with arachidonic acid.
Western Blot Analysis-Cell lysis was achieved by sonication (5 s, duty cycle 25%, output control 40%), followed by centrifugation at 5,000 ϫ g for 5 min. The supernatant was collected and protein determined with the DC-20 Protein Assay Kit (Bio-Rad). Protein (50 g for each sample) was resolved on a 10% SDS-polyacrylamide gel and transblotted onto nitrocellulose sheets (Bio-Rad) for Western blot analysis (52,53). Rabbit anti-human CYP2E1 polyclonal antibody (provided by Dr. J. M. Lasker, Mt. Sinai School of Medicine) and mouse anti-humanbcl-2 monoclonal antibody (Boehringer Mannheim) were used as the primary antibodies followed by treatment with alkaline phosphatase either conjugated to goat anti-rabbit IgG (Bio-Rad) or to rabbit antimouse IgG (Boehringer Mannheim) as the second antibody. Staining intensity was developed with the NBT-BCIP mixture (Promega).
Lipid Peroxidation Assay-Malondialdehyde (MDA) and 4-hydroxyalkenals, such as 4-hydroxy-2-nonenal (4-HNE), end products derived from peroxidation of PUFA and related esters, provide a convenient index as a measure for lipid peroxidation (54). Lipid peroxidation in Hep G2-MV2E1-9 and Hep G2-MV-5 cells was monitored by measuring total MDA and 4-HNE production, utilizing the lipid peroxidation assay kit, LPO-586 (Calbiochem). Briefly, after incubating the cells with varying concentrations of arachidonic acid, the tissue culture medium from the first and second 24-h incubation period was collected and assayed. The cells were collected by scraping and centrifugation. The pellets were resuspended in 20 mM Tris-HCl, pH 7.4, buffer, lysed by sonication, and centrifuged at 5,000 ϫ g for 5 min. The protein content of the cell lysates was determined (Bio-Rad DC-20 Protein Assay Kit) followed by the LPO-586 assay.

RESULTS
Arachidonic Acid Cytotoxicity in Hep G2-MV2E1-9 Cells-It has been show that dietary fat composition and subsequent elevated lipid peroxidation are related to the severity of alcohol-induced liver injury in the intragastric feeding rat model. To evaluate a role of PUFA in alcohol-related toxicity, Hep G2-MV2E1-9 and Hep G2-MV-5 cells were loaded with arachidonic acid (20:4) for 24 h, the medium was removed, and the cells were rinsed and continuously incubated at 37°C for an additional 24 h in normal MEM. Cell viability was then assessed by the MTT assay. Pretreatment with 0.03 mM arachidonic acid caused 43-72% (mean of 62%) loss of viability to Hep G2-MV2E1-9 cells, whereas no significant loss of viability (0 -13%, mean of 4%) was found with Hep G2-MV-5 cells. Compared with arachidonic acid, oleic acid (18:1) showed no significant toxicity to the Hep G2-MV2E1-9 cells even at concentrations (0.05 mM) in which arachidonic acid was highly cytotoxic (Table I). Arachidonic acid toxicity was also evaluated by morphology and by the LDH leakage assay. As shown in Fig.  1, arachidonic acid caused a 3-fold increase of LDH leakage, in terms of LDH out /LDH in ratio, with Hep G2-MV2E1-9 cells, 24 h after removal of arachidonic acid. Only a small increase (30%) in LDH leakage was found with Hep G2-MV-5 cells. In contrast to arachidonic acid, preloading cells with oleic acid did not result in increased LDH leakage by the Hep G2-MV2E1-9 (and Hep G2-MV-5) cells (Fig. 1). Arachidonic acid caused substantial morphological changes when added to the Hep G2-MV2E1-9 cells as many cells were detached and floated to the top of the culture dish; cells were shrunken and dispersed and a monolayer was not formed (Fig. 2). No such changes in morphology were evident when arachidonic acid was added to the Hep G2-MV-5 cells (Fig. 2).
To characterize the cytotoxicity produced by arachidonic acid, time course and dose-dependent experiments were con-ducted. The cytotoxic effect of preloading with various concentrations of arachidonic acid is shown in Fig. 3A. At concentrations of 0.005 or 0.01 mM, there was no significant toxicity by arachidonic acid in either cell line. At 0.02-0.03 mM, arachidonic acid caused significant toxicity to Hep G2-MV2E1-9 cells but not to Hep G2-MV-5 cells. At a concentration of 0.05 mM, arachidonic acid caused more than 80% loss of viability of Hep G2-MV2E1-9 cells; some toxicity was also observed in Hep G2-MV-5 cells although it was significantly lower than that in the Hep G2-MV2E1-9 cells. As shown in Fig. 3B, some toxicity by arachidonic acid could be observed immediately after the initial 24-h preincubation period, and this toxicity became more pronounced during the second incubation period after removal of the arachidonic acid. No significant cytotoxicity was observed in Hep G2-MV-5 cells over the same incubation period. At 36 h after preloading, viability of Hep G2-MV2E1-9 cells was lowered by 73% by the arachidonic acid treatment,  1. Arachidonic acid-induced cytotoxicity as measured by the LDH release assay. Hep G2-MV2E1 and Hep G2-MV-5 cells were cultured in control medium, medium containing 0.03 mM oleic acid, or medium containing 0.03 mM arachidonic acid for 24 h. The medium was then replaced with normal MEM for an additional 24-h incubation. At the end of this treatment, the supernatants were collected to measure the LDH activity (designated as LDH out ), and cells were harvested by scraping for measurement of LDH activity (designated as LDH in ). The cytotoxicity index was expressed as the ratio of LDH out /LDH in . whereas viability of Hep G2-MV-5 cells was decreased 26%.
Treatment with 0.1 mM BSO caused an approximate 90% depletion of GSH in both cell lines (24). Since GSH is known to protect cells against the toxicity of numerous agents, the effect of removal of GSH on the arachidonic acid toxicity was evaluated. In the presence of 0.1 mM BSO, arachidonic acid was more toxic to both cell lines. BSO treatment cause about a 2-3-fold increase in toxicity by arachidonic acid in both Hep G2 cell lines (Fig. 3A). However, the BSO treatment did not potentiate the toxicity of oleic acid to the Hep G2-MV2E1-9 cells (Table I).
Since GSH depletion potentiated the cytotoxicity of arachidonic acid, GSH appears to be important in protecting the Hep G2 cells against arachidonic acid-induced toxicity.
Role of CYP2E1 in Arachidonic Acid Cytotoxicity-Inasmuch as the only apparent difference between Hep G2-MV2E1-9 and Hep G2-MV-5 cells is the expression of CYP2E1 in the former, but not the latter, it would appear that the greater toxicity caused by arachidonic acid in Hep G2-MV2E1-9 cells is due to the presence of CYP2E1 in these cells. To validate the role of CYP2E1 in the elevated arachidonic acid cytotoxicity in Hep G2-MV2E1-9 cells, a plasmid, pCI-as-2E1, containing cDNA encoding antisense CYP2E1 was transfected into Hep G2-MV2E1-9 cells to block CYP2E1 production. Alternatively, a plasmid (pCI-2E1) containing human CYP2E1 cDNA was used to enrich the CYP2E1 content of the Hep G2-MV2E1-9 cells. Western blot analyses of the CYP2E1 content after transfection with the CYP2E1 sense and antisense plasmid indicated that the expression of CYP2E1 was decreased by about 80 -90% with pCI-as-2E1 as compared with control transfection with pCI, whereas expression of CYP2E1 was elevated about 3-fold after transfection with pCI-2E1 (Fig. 4, lanes 1, 3, and 5). Arachidonic acid toxicity in the cells transfected with control plasmid was very similar to that found previously with the non-transfected Hep G2-MV2E1-9 (Fig. 5A, pCI curve, compared with Fig. 3A E9 without BSO curve). Transfection with pCI-as-2E1 partially prevented the arachidonic acid toxicity; in fact, the arachidonic acid toxicity curve in the presence of pCI-as-2E1 (Fig. 5A) was similar to the toxicity curve found for the Hep G2-MV-5 cells (Fig. 3A, without BSO curve). This suggests that transfection with pCI-as-2E1 largely protected against the CYP2E1-dependent arachidonic acid toxicity. Transfection with pCI-2E1 plasmid increased the toxicity by arachidonic acid compared with the control pCI transfection (Fig. 5A). Thus, arachidonic acid toxicity is dependent upon CYP2E1 expression under these reaction conditions and at these concentrations of arachidonic acid.
Fatty acids can be metabolized by cytochrome P450 (55-58); CYP2E1 catalyzes -1 hydroxylation of arachidonic acid to a variety of complex products (59 -61). Since CYP2E1 is a loosely coupled cytochrome P450, i.e. can generate reactive oxygen species such as superoxide and H 2 O 2 in the absence of a metabolic substrate (4,62), it was necessary to evaluate two possible roles for CYP2E1 in promoting the toxicity of arachidonic acid, i.e. CYP2E1 directly oxidized arachidonic acid to reactive metabolites that produced the toxicity (Equation 1) or CYP2E1 generated superoxide and H 2 O 2 which then reacted with arachidonic acid to produce toxicity (Equation 2). The latter possibility would not require direct oxidation of arachidonic acid by CYP2E1. 4-Methylpyrazole (4-MP) is a ligand of CYP2E1 and an effective inhibitor of CYP2E1-catalyzed oxidation of substrates (63). 4-MP was shown to completely prevent the toxicity of acetaminophen and ethanol to the Hep G2-MV2E1-9 cells, indicating that this toxicity required metabolism of these agents by CYP2E1 (24,25). 4-MP was added to the culture medium of Hep G2-MV2E1-9 cells during the 24-h loading period with arachidonic acid and during the 24-h post-loading period. 4-MP did not protect against the cytotoxicity produced by arachidonic acid in the absence of BSO; a small protection was observed against the enhanced toxicity found in the presence of BSO (Fig. 5B); however, this protection was much less than the complete protection afforded by 4-MP against toxicity of acetaminophen and ethanol (24,25). As will be discussed below, other CYP2E1 ligands and competitive substrates such as Me 2 SO or ethanol did not significantly protect the Hep G2-MV2E1-9 cells against arachidonic acid toxicity. These results suggest that direct metabolism of arachidonic acid by CYP2E1 may not play an important role in arachidonic acid toxicity in Hep G2-MV2E1-9 cells.
Enhanced Lipid Peroxidation Induced by Arachidonic Acid in Hep G2-MV2E1-9 Cells-The possible mechanism for arachidonic acid toxicity suggested in Equation 2 directly implicates lipid peroxidation as playing a central role in the toxicity. Lipid peroxidation of Hep G2-MV2E1-9 and Hep G2-MV-5 cells was assessed by measuring production of the lipid peroxidation end products MDA and 4-HNE. As shown in Fig. 6, arachidonic acid induced lipid peroxidation in Hep G2-MV2E1-9 cells in a concentration-dependent manner; enhanced formation of MDA and 4-HNE was observed in both cell lysate (Fig. 6A) and in the culture medium from the cells (Fig. 6B). Arachidonic acid (up to 0.03 mM) caused little or no lipid peroxidation in Hep G2-MV-5 cells. The significant difference in lipid peroxidation between the two cell sublines suggests that overexpression of CYP2E1 enhanced the PUFA-induced lipid peroxidation. Subsequent studies were carried out to evaluate whether the enhanced FIG. 5. Role of CYP2E1 in arachidonic acid-induced cytotoxicity. A, Hep G2-MV2E1-9 cells were transiently transfected with plasmids pCI, pCI-2E1, or pCI-as-2E1 and then subjected to arachidonic acid treatment for 24 h. After removal of medium containing PUFA followed by an additional 24 h of culture in normal medium, cell viability was determined by the MTT assay. B, Hep G2-MV2E1-9 cells were pretreated with 0.02, 0.03, or 0.04 mM arachidonic acid for 24 h, in the absence or presence of 2 mM 4-MP or absence or presence of 0.1 mM BSO. After removal of the arachidonic acid medium, fresh control medium (with the same amount of 4-MP or BSO respectively) was added, and cell viability was determined after an additional 24-h incubation period by the MTT assay.

FIG. 6. Lipid peroxidation in Hep G2-MV2E1-9 and Hep G2-MV-5 cells after arachidonic acid pretreatment.
After 24 h of arachidonic acid (0.01-0.03 mM) pretreatment followed by an additional 24 h of culture in normal medium, Hep G2-MV2E1-9 and Hep G2-MV-5 cells were harvested and assayed for lipid peroxidation (A). The culture medium was collected at the end of the arachidonic acid treatment and at the end of the additional 24-h culture, and the combined total was used to assay for lipid peroxidation (B). The lipid peroxidation level was monitored by the production of MDA and 4-HNE, utilizing the LPO-586 kit, as described under "Materials and Methods." lipid peroxidation was responsible for the cytotoxicity and cell damage produce by arachidonic acid.
Effect of Antioxidants on Arachidonic Acid Cytotoxicity-To characterize further the nature of arachidonic acid cytotoxicity, several antioxidants were added to the culture medium, and their effect on arachidonic acid toxicity was determined. As shown in Table II, ascorbic acid, the iron chelator desferrioxamine, and several typical inhibitors of lipid peroxidation, such as trolox, ␣-tocopherol phosphate, propylgallate, and DPPD, produced efficient protection against 0.03 mM arachidonic acid toxicity in the Hep G2-MV2E1-9 cells. Me 2 SO (5-50 mM) and ethanol (25-160 mM) as ligands for CYP2E1 and as hydroxyl radical scavengers failed to prevent arachidonic acid toxicity. These results suggest that the arachidonic acid toxicity in Hep G2-MV2E1-9 is due to, at least in part, the enhanced lipid peroxidation. Aspirin, an inhibitor of the cyclooxygenase pathway for arachidonic acid metabolism, did not protect against the toxicity of arachidonic acid (Table II).
In Situ DNA Nick End Labeling of Hep G2-MV2E1-9 Cells-Two distinct modes of cell death, apoptosis and necrosis, can be distinguished based on differences in morphological, biochemical, and molecular changes of dying cells. Experiments were carried out to determine whether apoptotic cell death occurs in arachidonic acid-induced cytotoxicity to Hep G2-MV2E1-9 cells. In general, cells undergoing apoptosis display a characteristic pattern of structural changes in nucleus and cytoplasm, including rapid blebbing of the plasma membrane and nuclear disintegration. The nuclear collapse is associated with extensive damage to chromatin and DNA cleavage into oligonucleosomal length DNA fragments (47)(48)(49)(50). After 24 h of arachidonic acid preloading, Hep G2-MV2E1-9 and Hep G2-MV-5 cells were placed in normal MEM for an additional 8 h of incubation. Cells were then harvested for in situ DNA nick end labeling as determined by the TUNEL method. In the absence of arachidonic acid, the intensity of FITC labeling was similar for the Hep G2-MV2E1-9 (mean, 0.41-1.02) and Hep G2-MV-5 (mean, 0.5-0.74) cells (Fig. 7). Arachidonic acid enhanced the FITC labeling with both cell lines; however, the intensity of FITC labeling in the Hep G2-MV2E1-9 cells preincubated with 0.03 mM arachidonic acid (mean, 12.11-13.34) was significantly higher than that of Hep G2-MV-5 (mean, 2.12-2.84). Results from several TUNEL experiments are summarized in Fig. 8

TABLE II
Effect of anti-oxidative agents on the cytotoxicity produced by arachidonic acid to Hep G2-M2E1-9 cells Hep G2-MV2E1-9 cells were incubated with or without 0.03 mM arachidonic acid for 24 h in the presence of the indicated additions. The respective media were removed and replaced by MEM supplied with the corresponding agents but without arachidonic acid, and a second incubation for 24 h was carried out in the absence of arachidonic acid. Cell viability was evaluated by the MTT assay. Net absorbance refers to absorbance at 570 nm minus absorbance at 630 nm, and percent viability refers to net absorbance of arachidonic acid treated, divided by net absorbance of control, times 100. ited by trolox (histogram in Fig. 7; quantitation in Fig. 8B). These results suggest that enhanced lipid peroxidation caused by arachidonic acid preincubation induced apoptosis and cytotoxicity in Hep G2-MV2E1-9 cells. Apoptosis in Hep G2-MV2E1-9 Cells-Apoptotic cells often produce a unique ladder composed of nucleotide fragments at an interval of 200 base pairs, which can be visualized by DNAagarose electrophoresis. The TUNEL in situ labeling suggested that arachidonic acid toxicity in Hep G2-MV2E1-9 cells is apoptotic in nature. To study this further, DNA fragmentation within Hep G2-MV2E1-9 cells was determined. Hep G2-MV2E1-9 cells were harvested at various times after arachidonic acid incubation (6, 12, and 24 h) and 8 h after removal of arachidonic acid. Total DNA was purified for the agarose gel electrophoresis assay. During the 24-h preloading period, 0.02-0.04 mM arachidonic acid did not induce significant DNA fragmentation (Fig. 9A, lanes 3-5, 7-9, and 11-13 compared with  lanes 2, 6, and 10). However, 8 h after preloading, 0.03 and 0.04 mM arachidonic acid caused DNA fragmentation in the Hep G2-MV2E1-9 cells (Fig. 9A, lanes 16 and 17 compared with lane  14, no arachidonic acid added). Eight hours after the initial 24-h preloading with 0.03 mM arachidonic acid, Hep G2-MV-5 cells did not show a significant DNA ladder (Fig. 9B, lane 6, compared with Hep G2-MV2E1-9 cells shown in lane 7). The DNA fragmentation in Hep G2-MV2E1-9 cells was completely blocked by 0.1 mM trolox (Fig. 9B, lanes 8 and 9, compared with  lane 7).

DISCUSSION
The primary goal of the present study was to investigate the toxicity of arachidonic acid in a human liver cell line in which the major or the only significant cytochrome P450 isoform is CYP2E1. Induction of CYP2E1 and the formation of reactive intermediates, including reactive metabolites, reactive oxygen species, lipid peroxidation derivatives appears to be one of the mechanisms that is receiving much current interest in studies evaluating how ethanol is hepatotoxic. It has been demonstrated that relative to several other cytochrome P450 isozymes, CYP2E1 displays high NADPH oxidase activity, is loosely coupled, and is more reactive in oxidizing ethanol to the 1-hydroxyethyl radical (4, 9 -11, 62). Microsomes from ethanoltreated rats are more reactive than the controls in producing a variety of reactive oxygen intermediates by reactions sensitive to anti-CYP2E1 antibodies and to chemical inhibitors of CYP2E1 (2)(3)(4)(5)(6)(7)(8). Correlation between induction of CYP2E1, lipid peroxidation, and ethanol-induced liver injury has been reported with the continuous intragastric infusion model of ethanol feeding (16,18,22). The studies using the intragastric model of rat feeding indicated that a high content of polyunsaturated fatty acids would lead to enhanced CYP2E1-dependent lipid peroxidation and pathogenesis of alcoholic liver disease (20). To establish direct linkage between CYP2E1, PUFA toxicity, and the role of lipid peroxidation and oxidative stress, we utilized a previously established human hepatoma Hep G2 subline, Hep G2-MV2E1-9 clone, which was transduced with human CYP2E1 cDNA by using a retrovirus shuttle vector (23). An advantage of this model is the stable, constitutive expression of CYP2E1, in contrast to the rapid decline of the isoform in primary cultured hepatocytes. Experiments were carried out to evaluate whether arachidonic acid, a representative PUFA, is more toxic to cells expressing CYP2E1 compared with control cells not expressing CYP2E1, whether the elevated toxicity is associated with enhanced lipid peroxidation, whether antioxidants can rescue the cells against PUFA cytotoxicity, whether the cytotoxicity is apoptotic in nature, and whether bcl-2 can protect the cells against the PUFA cytotoxicity.
Hep G2-MV2E1-9 cells expressing CYP2E1 and Hep G2-MV-5 cells that do not have detectable CYP2E1 expression were first incubated with arachidonic acid for 24 h, followed by removal of the PUFA, addition of fresh medium not containing added PUFA, and analysis for toxicity. Indices of toxicity included LDH leakage, morphology, and decreased vital dye reduction (MTT assay). Arachidonic acid (0.03 mM) induced cytotoxicity in Hep G2-MV2E1-9 cells, whereas significantly lower or no cytotoxicity was found in the control Hep G2-MV-5 cells. The cytotoxicity produced by arachidonic acid was concentration-and time-dependent. An important control is the observation that oleic acid was not toxic to the CYP2E1 expressing cells under conditions in which arachidonic acid was toxic, indicating that toxicity is not due to fatty acid metabolism per se but rather due to the presence of a PUFA. This suggests that lipid peroxidation plays a role in the arachidonic acid cytotoxicity to Hep G2-MV2E1-9 cells. Three lines of experiments are supportive for a role for lipid peroxidation in the PUFA toxicity to Hep G2-MV2E1-9 cells. Depletion of GSH by BSO treatment increased arachidonic acid toxicity to the Hep G2-MV2E1-9 (and the Hep G2-MV-5 cells). GSH is known to protect cells against oxidative stress and damage caused by lipid peroxidation (64,65). Formation of characteristic end products of lipid peroxidation, malondialdehyde and 4-hydroxy-2-nonenal, was strikingly elevated in the Hep G2-MV2E1-9 cell extracts and in the culture medium from the Hep G2-MV2E1-9 cells after addition of arachidonic acid, whereas only a small increase in these lipid aldehydes was found with Hep G2-MV-5 cells. A variety of antioxidants that prevent lipid peroxidation including ␣-tocopherol phosphate, trolox, ascorbate, propylgallate, DPPD, and the iron chelator, desferrioxamine, were effective in preventing the toxicity by arachidonic acid to the Hep G2-MV2E1-9 cells. Me 2 SO and ethanol, besides being substrates or ligands for CYP2E1, are also effective hydroxyl radical scavenging agents. These compounds afforded little protection against the arachidonic acid toxicity, suggesting that either hydroxyl radical-like species were not involved in the PUFA toxicity (e.g. hydroxyl radical scavengers do not prevent micro- Hep G2-MV2E1-9 cells were transiently transfected with plasmids pCI, pCI-bcl-2, or pCI-as-bcl-2 and then subjected to arachidonic acid treatment for 24 h. After an additional 24 h of culture in normal medium without arachidonic acid supplement, cell viability was evaluated by the MTT assay. somal lipid peroxidation (66,67)) or that secondary radicals produced from the interaction of Me 2 SO (methyl radical) or ethanol (1-hydroxyethyl radical) were themselves toxic.
It is not likely that the enhanced PUFA toxicity to Hep G2-MV2E1-9 cells is mediated via arachidonic acid metabolism to eicosaenoid products since there should be no difference in cyclooxygenase pathways between the Hep G2-MV2E1-9 and the Hep G2-MV-5 cells. In addition, aspirin did not significantly protect against the PUFA toxicity, whereas transfection with a plasmid containing antisense CYP2E1 cDNA lowered the PUFA toxicity. The significant difference between the Hep G2-MV2E1-9 and Hep G2-MV-5 cells is the expression of CYP2E1 in the former and not in the latter.
Some toxicity by higher concentrations of arachidonic acid was also observed with the Hep G2-MV-5 cells that do not express CYP2E1. This toxicity by higher concentrations of arachidonic acid most likely reflects a non-CYP2E1-mediated lipid peroxidation process since (a) toxicity was enhanced after BSO treatment to lower cellular GSH levels (Fig. 3A), (b) small increases in malondialdehyde and 4-hydroxynonenal were produced upon incubating the Hep G2-MV-5 cells with 0.03 mM arachidonic acid (Fig. 6); and (c) the small increase in FITC labeling found when arachidonic acid was incubated with the Hep G2-MV-5 cells, analogous to the large increase found with the CYP2E1-expressing cells, was prevented by the antioxidant trolox ( Figs. 7 and 8). Most likely, reactive oxygen species are being produced from other cellular sources than cytochrome P450 mixed function oxidase activity, e.g. mitochondria may be the predominant source of reactive oxygen species under many conditions. There are numerous studies in the literature showing that enrichment of hepatocytes or tumor cells with arachidonic acid or other PUFAs results in lipid peroxidation and cellular toxicity that can be prevented by antioxidants such as vitamin E or DPPD (68 -73), analogous to the toxicity produced by high concentrations of arachidonic acid to the Hep G2-MV-5 cells.
Human CYP2E1 has been shown to metabolize arachidonic acid to -1-hydroxy-arachidonic acid (59 -61). A CYP2E1 inhibitor, 4-methylpyrazole, did not prevent the toxicity of arachidonic acid at concentrations that prevented toxicity of ethanol, CCl 4 , and acetaminophen (24,25). CYP2E1 substrates, such as ethanol and Me 2 SO, also did not effectively inhibit the arachidonic acid toxicity (Table II). These results suggest that the direct metabolism of arachidonic acid to potentially toxic products by CYP2E1 does not contribute significantly to the PUFA toxicity. To validate the role of CYP2E1 in the enhanced PUFA toxicity to the E9 cells, transfection experiments with plasmids containing CYP2E1 cDNA in the sense and antisense orientations were carried out. Compared with transfection with the control plasmid, transfection with sense CYP2E1 cDNA increased arachidonic acid toxicity, whereas transfection with antisense CYP2E1 cDNA decreased PUFA toxicity to the level observed with Hep G2-MV-5 cells. Isolated microsomes from Hep G2-MV2E1-9 cells have been shown to produce superoxide radical and H 2 O 2 at elevated rates compared with Hep G2-MV-5 microsomes (23). Low concentrations of ferric-ATP effectively catalyzed lipid peroxidation with Hep G2-MV2E1-9 microsomes but not with Hep G2-MV-5 microsomes (23). Taken as a whole, these results suggest that elevated production of reactive oxygen intermediates due to the presence of CYP2E1 in Hep G2-MV2E1-9 cells can result in the formation of potent oxidants that can initiate lipid peroxidation if sufficient levels of PUFA are available.
Since CYP2E1 is a "loosely coupled" enzyme (4, 62), formation of reactive oxygen intermediates occurs even in the absence of added substrates. In fact, formation of superoxide and H 2 O 2 by microsomes from the Hep G2-MV2E1-9 cells was not altered by the addition of substrates and ligands of CYP2E1, including ethanol and 4-methylpyrazole (23), which probably explains why these agents did not protect against arachidonic acid toxicity.
DNA fragmentation assessed by DNA-agarose gel electrophoresis and the TUNEL method showed that the toxicity induced by arachidonic acid in the Hep G2-MV2E1-9 cells involved apoptosis. Trolox, a vitamin E analog and a lipid peroxidation inhibitor, prevented Hep G2-MV2E1-9 cells from apoptosis and cytotoxicity induced by arachidonic acid, suggesting that lipid peroxidation played a role in the developing apoptosis and in the cytotoxicity. Intracellular reactive oxygen species and elevated levels of lipid peroxidation have been implicated as being associated with apoptosis (31,74,75). Our results suggest that enrichment of the polyunsaturated fatty acid levels in biological membranes of Hep G2-MV2E1-9 cells is critical for development of apoptosis induced by CYP2E1dependent oxidative stress. bcl-2 inhibits many types of apoptotic cell death, although the mechanism is not completely clear (35)(36)(37)(38)(39)(40)(41)(42)(43). bcl-2 is localized to intracellular sites of reactive oxygen species generation including mitochondria, endoplasmic reticulum, and nuclear membranes (31, 76 -78). When Hep G2-MV2E1-9 cells were transfected with bcl-2, they became resistant to the arachidonic acid toxicity, which is consistent with the protection by various antioxidants. Interestingly, the transfectants from plasmid containing antisense bcl-2 cDNA showed an increased toxicity by arachidonic acid, probably due to the suppression of the low level of endogenous bcl-2 in the Hep G2 cells (Fig. 4B). Transfection with the bcl-2 sense or antisense plasmid did not affect expression of CYP2E1 as compared with transfection with control plasmid (Fig. 4A).
In summary, experiments have been carried out that demonstrate that arachidonic acid is toxic to cells that express CYP2E1 but not to cells that do not express CYP2E1. The PUFA toxicity is associated with increased lipid peroxidation and can be diminished by antioxidants that prevent lipid peroxidation. The toxicity appears to be apoptotic in nature and can be prevented by overexpression of bcl-2. Since production of reactive oxygen intermediates is elevated with microsomes isolated from cells expressing CYP2E1 compared with controls, it is proposed that this elevated generation of reactive intermediates can initiate lipid peroxidation, which subsequently causes apoptosis and cellular damage, when the cells are preloaded with PUFA. These results indicate that enrichment of cells that express CYP2E1 with PUFA results in cytotoxicity. The Hep G2-MV2E1-9 cells appear to be a useful model to study interactions between CYP2E1, PUFA, and free radicals and the consequences of these interactions on cell viability. They also appear to reproduce, in a simple cell culture model, several of the key features associated with ethanol hepatotoxicity in the intragastric infusion model of ethanol treatment.