Increased Mitochondrial Superoxide Production in Rat Liver ...

51, 224 –235 (1999) Copyright © 1999 by the Society of Toxicology

TOXICOLOGICAL SCIENCES

Increased Mitochondrial Superoxide Production in Rat Liver Mitochondria, Rat Hepatocytes, and HepG2 Cells following Ethinyl Estradiol Treatment Jinqiang Chen, Yunbo Li, Jackie A. Lavigne, Michael A. Trush, and James D. Yager 1 Department of Environmental Health Sciences, Division of Toxicological Sciences, Johns Hopkins School of Hygiene and Public Health, Baltimore, Maryland Received February 17, 1999; accepted May 18, 1999

Ethinyl estradiol (EE) is a strong promoter of hepatocarcinogenesis. Treatment of rats with EE and other hepatic promoters induces a mitosuppressed state characterized by decreased hepatocyte turnover and reduced growth responsiveness. Previously, we identified several nuclear and mitochondrial genome-encoded mitochondrial genes whose transcripts were increased during EEinduced hepatic mitosuppression in rats and in EE-treated HepG2 cells (Chen et al. Carcinogenesis, 17, 2783–2786, 1996 and Carcinogenesis, 19, 101–107, 1998). In both cultured rat hepatocytes and HepG2 cells, EE increased respiratory chain activity (reflected by increased mitochondrial superoxide production detected as increased lucigenin-derived chemiluminescence (LDCL). In this paper, we provide additional characterizations of these effects. Increased LDCL was detected in mitochondria isolated from EEtreated rats, documenting that these estrogen effects on mitochondrial function are not confined to cells in culture. EE and estradiol (E2) increased LDCL in cultured rat hepatocytes and HepG2 cells in a dose- (beginning at 0.25 mM levels) and time-dependent response. Inhibition of P450-mediated estrogen metabolism inhibited, while direct exposure to E2 catechol metabolites enhanced LDCL. Co-treatment with glutathione ester or with the specific antiestrogen, ICI 182708 inhibited LDCL. In contrast, estrogeninduced LDCL was enhanced by glutathione depletion, and by inhibition of catechol-o-methyltransferase. These results support a working hypothesis that in liver cells, increased respiratory chain activity induced by estrogen treatment requires both metabolism to catechols and an estrogen receptor-mediated signal transduction pathway. Key Words: Ethinyl estradiol; hepatic promoters; mitosuppression; rat; liver; superoxide.

Cumulative exposure to various estrogens has been associated with an increased risk for developing several common cancers in women, including a modest increase in liver 1

To whom correspondence should be addressed at Department of Environmental Health Sciences, Division of Toxicological Sciences, Johns Hopkins School of Hygiene and Public Health, 615 N. Wolfe St., Baltimore, MD 21205–2179. Fax: 410 –955– 0116. E-mail: [email protected].

cancer in association with prolonged use of oral contraceptives (Palmer et al., 1989). The synthetic estrogen ethinyl estradiol (EE), which is widely used in oral contraceptives, has been identified as a strong promoter of hepatocarcinogenesis using the rat initiation/promotion model (Yager and Liehr, 1996). In rat liver, at the non-hepatotoxic doses used in tumor promotion studies (#5 mg/day), EE induced an initial, transient stimulation of hyperplastic growth (Mayol et al., 1991; Yager et al., 1986) followed by a period of growth suppression (mitosuppression) (Yager et al., 1994). Several other liver tumor promoters, including clofibrate and phenobarbital (PB), cause similar effects (Barbason et al., 1983; Tanaka et al., 1992). For EE, both the initial mitogenic stimulation and promotion of diethylnitrosamine-initiated hepatocarcinogenesis were blocked by tamoxifen (Yager et al., 1986), indicating involvement of estrogen receptor signaling mechanisms. Our overall hypothesis is that mitosuppression represents a growth-negative, selective environment, which is fundamental to tumor promotion by EE. Under these conditions, the altered hepatic foci that develop during promotion by EE would represent clonal outgrowths of hepatocytes that became resistant to mitosuppression during initiation. In other words, initiated hepatocytes would be differentially resistant to EE-induced mitosuppression (Farber, 1990; Yager et al., 1994). In fact, hepatic tumor promotion by phenobarbital has been shown to involve differential mitoinhibitory effects in initiated versus surrounding non-initiated hepatocytes (Jirtle et al. 1994; Jirtle and Meyer, 1991; Mamsbach et al., 1996). The mechanisms underlying EE-induced mitosuppression in liver are not known. Studies by Jirtle and coworkers suggest that altered expression of transforming growth factor-beta (TGF-b) and the manose-6 phosphate/insulin-like growth factor II (M6P/IGF II) receptor are involved in PB-induced mitosuppression. They reported that in PB-induced, mitosuppressed rat livers, the levels of TGF-b1 protein and M6P/ IGF-II receptor mRNA and protein were increased (Jirtle et al.,

224

MITOCHONDRIAL O 2 •2 PRODUCTION AFTER ETHINYL ESTRADIOL TREATMENT

1994). Similar to their findings, we also detected increased mRNA levels of TGF-b1 and M6P/IGF-II receptor in EEinduced, mitosuppressed, female, rat liver (Chen et al., 1996). These results suggest that a TGF-b1/M6P/IGF-II receptormediated signaling pathway is involved in hepatic mitosuppression. But how PB and EE trigger this signaling pathway and how it leads to mitosuppression remain to be determined. Since several other hepatic promoters cause mitosuppression, elucidating these processes at the cellular and molecular levels may lead us to a better understanding of the mechanisms of carcinogenesis and tumor promotion. To investigate the mechanisms of EE-induced hepatic mitosuppression, we performed a differential display to identify genes whose expression was altered during mitosuppression (Chen et al., 1996). Nuclear genome-encoded mitochondrial ATP synthase subunit E was among the several genes whose transcript levels were increased. In a continuation of that study, we identified two additional cDNAs from the differential display as mitochondrial genome-encoded cytochrome c oxidase subunit III and ATPase synthase 6 (Chen et al., 1998). We found that EE, estradiol (E2) and the estradiol catechol metabolites 1,3,5(10)estretrien-3,4,17b-triol (4-OH-E2) and 1,3,5(10)estratrien-2,3,17b-triol (2-OH-E2) increased the levels of these and other mitochondrial genome-encoded transcripts in human hepatoma HepG2 cells (Chen et al., 1998). This increase, when caused by E2 and EE, was blocked by an inhibitor of cytochrome P450. Using lucigenin-derived chemiluminescence (LDCL) to detect intra-mitochondrial production of superoxide (Li et al., 1998), we observed that the increase in mitochondrial transcript levels was accompanied by increased mitochondrial production of superoxide, both in estrogen-treated cultured rat hepatocytes and in HepG2 human hepatoma cells (Chen et al., 1998). This demonstrated that the mechanistic characterization of this estrogen effect could be studied in vitro. In this paper, we extend our findings by showing that superoxide production was increased in mitochondria isolated from EE-treated rats. We also report the results of detailed characterizations of the estrogen-induced mitochondrial production of superoxide in HepG2 cells and cultured rat hepatocytes. MATERIALS AND METHODS Fischer 344 VF rats were obtained from Charles River (Willimgton, DE). Experiments involving rats were done according to protocols approved by the Animal Care and Use Committee of the Johns Hopkins School of Hygiene and Public Health. Reagents and media for isolation of rat hepatocytes and for culture of hepatocytes and HepG2 cells were obtained from Gibco BRL Life Technologies (Gaithersburg, MD) or from Collaborative Biomedical (Bedford, MA). Ethinyl estradiol, E2, 2-OH-E2, 4-OH-E2, 1,3,5,(10)estratrien-2,3,17btriol 2-methyl ether (2-methoxyestradiol), and 1,3,5,(10) estratrien-3,4,17btriol 4-methyl ether (4-methoxyestradiol) were from Steraloids, Inc. (Newport, RI). Lucigenin, L-buthionine-(S,R)-sulfoximine (BSO), glutathione ester, SKF-525, rotenone, and myxothiazol were from Sigma Chemical Co. (St. Louis, MO). The catechol-o-methyl transferase (COMT) inhibitor, RO 410960, was from Research Biochemicals, Inc. (Nadick, MA). The antiestrogen ICI 182708 was obtained from Dr. Nancy Davidson, Johns Hopkins

225

University, with the permission of Zeneca Pharmaceuticals. C 2-ceramide was from BioMol Research Laboratories, Inc. (Plymouth Meeting, PA). Cell culture. HepG2 cells were cultured in RPMI 1640 medium containing 10% fetal bovine serum, insulin (5 mg/ml)-selenium (5 ng/ml)/transferrin (5 mg/ml), ITS (Collaborative Biomedical), and gentamicin. Hepatocytes were isolated from male and female F-344 rats (approximately 200 g) by collagenase perfusion (Seglen, 1976) and cultured in a modified Chee’s medium as described in detail previously (Zurlo and Arterburn, 1996). Isolation of liver mitochondria. Mitochondria were isolated from male rat livers as described by Pedersen et al. (1978) and used immediately. The amount of mitochondrial protein was determined using the Bio-Rad assay (Bio-Rad Laboratories, Richmond, CA) with bovine serum albumin (BSA) as a standard. The coupling status, ratio of O 2 consumption in the presence of ADP 1 succinate/succinate, of the isolated mitochondria was determined to be 3.5 6 0.09 (mean 6 SD, n 5 3), indicating that the mitochondria are intact and undamaged. Lucigenin-derived chemiluminescence (LDCL) measurement. LDCL was used to assess intramitochondrial superoxide generation as a reflection of mitochondrial respiratory chain activity. The LDCL assay was performed using a Berthold LB9505 luminometer at 37°C for 30 min under conditions validated for use to detect superoxide anion production in enzymatic systems and intact cells, as described by Li et al. (1998). Isolated mitochondria or freshly prepared hepatocytes in suspension were incubated, in the luminometer, plus/minus estrogen. The mitochondria (500 mg protein equivalent) or hepatocytes (3 3 10 6) were suspended in 2.5 ml air-saturated respiration buffer (70 mM sucrose, 220 mM mannitol, 2 mM HEPES, 2.5 mM MgCl 2, 0.5 mM EDTA and 0.1% BSA, pH 7.4) plus the indicated agents at the indicated concentrations. The measurement of LDCL with isolated mitochondria was initiated by adding succinate and lucigenin to final concentrations of 6 mM and 10 mM, respectively. Luminescence was then monitored continuously for 30 min. The cultured rat hepatocytes and HepG2 cells were treated with the various estrogens at the concentrations and for the times indicated. Since the cells rapidly metabolize the estrogens, the media were renewed every 12 h during the culture period in order to remove conjugated metabolites and provide fresh estrogen. To perform the LDCL determinations, the cells were removed from the culture dishes, resuspended in 2.5 ml air-saturated PBS buffer (pH 7.4) (3 3 10 6 hepatocytes or 4 3 10 6 HepG2 cells), and assayed intact. The LDCL determination was initiated by adding lucigenin to the final concentration of 10 mM and followed continuously for 30 min in the luminometer. The chemiluminescence detected is expressed as counts/min (cpm). The LDCL data are shown both as the actual chemiluminescent curves generated over 30-minincubation periods, and in a numerical form based on the integration of the areas under the curves. In the latter cases, the data are presented as percent increase over control and the actual control values are given in the figure legends. While differences in actual chemiluminescence cpm do occur among separate experiments, differences among treatments are highly reproducible. Cell type-specific differences are also observed in the actual shapes of the chemiluninescence curves. Thus, HepG2 cells reproducibly yield cpm curves that increase gradually and extend through 15 to 20 min, whereas hepatocytes yield cpm curves that increase rapidly and are of shorter duration. The precise reasons for these differences are not known although the rates of substrate utilization and cell settling likely contribute to the differences. However, the specific treatments do not alter the overall shapes of the curves. Measurement of O 2 consumption. O 2 consumption was measured polarographically with a Clark-type oxygen electrode (YSI-53, Yellow Spring, OH) at 37°C in 2.5-ml reaction mixtures as described by Li et al. (1998). The buffer used, the concentrations of substrate, and the numbers of cells were identical to those used for LDCL determinations. Determination of NADH-dehydrogenase (Complex I) activity. NADHdehydrogenase activity was assayed spectrophotometrically by the rate of NADH-dependent ferricyanide reduction at 420 nm, as described by Singer (1994).

226

CHEN ET AL.

TABLE 1 LDCL and O 2 Uptake in Mitochondria Isolated from EE-Treated Rats

Treatment

LDCL (cpm 3 10 7/mg protein)

O 2 uptake (nMol/mg protein/min)

Control EE-treated

2.97 6 0.43 4.43 6 0.37*

26.6 6 1.6 33.7 6 2.0*

Note. Female F-344 rats were treated for 42 days with time-release tablets delivering EE at 5.0 mg/day as described previously (Yager et al., 1994). The rats were sacrificed, and mitochondria were prepared and assayed for LDCL and O 2 uptake, as described in Materials and Methods. Two separate experiments were conducted with 6 rats/group in the first and 4 rats/group in the second. * Significantly greater than control, p , 0.05, n 5 10 from 2 separate experiments.

Statistical analysis and replication of results. The results presented are representative of at least 3 to 5 independent experiments. Where indicated, statistical analysis was done using a one-way ANOVA and the differences considered significant when p # 0.05 using the Student-Newman-Keuls method for pairwise multiple comparisons.

RESULTS

O 2 •2 Production by Liver Mitochondria from EE-Treated Female Rats Our investigation of the estrogen effects on mitochondria originated with finding increased expression of mitochondrial gene transcripts in rat liver during a state of EE-induced mitosuppression (Chen et al., 1996, 1998). While in cultured hepatocytes and HepG2 cells, the increase in mitochondrial transcript levels was associated with increased respiratory chain activity; an important question was whether estrogen treatment in vivo had similar effects on mitochondrial respiration. In two separate experiments, female F344 rats were treated for 42 days with time-release tablets delivering EE at 5 mg/day. This is the same dose of EE used previously to promote liver tumors (Yager et al., 1986) and to induce mitosuppression in the rat livers that served as a source for the mRNA used in the differential display (Chen et al., 1996; Yager et al., 1994). Controls received placebo tablets. Following euthanasia of the rats, mitochondria were isolated and used immediately for determination of LDCL and O 2 uptake. The results shown in Table 1 represent a composite from both experiments (total of 10 control and 10 EE-treated rats). The data show a statistically significant (p , 0.05) increase in both LDCL and O 2 consumption in liver mitochondria of EE-treated rats. These data demonstrate that the effects of EE on mitochondrial respiratory chain activity are not confined to cultured cells.

Immediate Effects of EE, E2, 2-OH- and 4-OH-E2 on Superoxide (O 2 •2) Production by Isolated Mitochondria and Hepatocytes in Suspension Figure 1A shows the effects of 2-OH-E2 on LDCL in freshly isolated mitochondria from untreated control rats. The integral areas under each curve represent LDCL units expressed as counts per min (cpm). These data show that, as the concentration of 2-OH-E2 increased from 0.5 to 30 mM, mitochondriagenerated LDCL levels decreased, indicating suppression of mitochondrial superoxide generation. Similar results were observed using 4-OH-E2, but not E2 or EE (data not shown), indicating that the catechol metabolites but not the parent estrogens are responsible for this effect. Figure 1B shows the effects of acute treatment of freshly isolated female rat hepatocytes incubated in suspension (30 min in the chemiluminometer) with EE, E2, 2-OH-E2, or 4-OH-E2. The areas under the LDCL curves were integrated and plotted as percent of an untreated control for each estrogen concentration. EE appeared to exert a small, transitory inhibition, whereas E2 had virtually no effect on LDCL generation. In contrast, the E2 catechol metabolites, 2-OH-E2 and 4-OH-E2, inhibited LDCL generation. This effect was nearly maximal at the lowest concentrations (1 mM) tested. The virtual lack of inhibitory effect of the parent estrogens in the acutely treated intact hepatocytes is similar to what was observed in isolated mitochondria. Furthermore, the methoxy-metabolites of the catechols, 2-methoxyestradiol and 4-methoxyestradiol, caused less inhibition than their respective catechols (data not shown). Together, these results provide support for a rapid inhibition of mitochondrial superoxide production by estrogen catechol metabolites. Consistent with this, we observed that 4-OH-E2 and 2-OH-E2, but not E2 or EE, inhibited mitochondrial complex I (NADH dehydrogenase) activity (Fig. 1C). Since complex I is a site of superoxide generation during electron transport from NADH, inhibition of complex I activity by the catechol metabolites could account for the reduction in LDCL under these conditions. Effects of EE, E2, 2-OH- and 4-OH-E2 on Mitochondrial Production of O 2 •2 by HepG2 cells and Hepatocytes vs. time in culture Next, we investigated the chronic effects of treatment with these estrogens on mitochondrial LDCL in cultured HepG2 cells and rat hepatocytes. In contrast to what was seen with acute exposure of isolated mitochondria and hepatocytes in suspension, we observed that treatment of cultured HepG2 cells with 2-OH-E2 (5 mM) for 48 h caused a 3-fold increase in superoxide levels over control (data not shown). Superoxide can be generated in the endoplasmic reticulum from microsomal cytochrome P450-catalyzed reactions such as P450-mediated redox cycling of catechol estrogens (Liehr and Roy, 1990; Liehr et al., 1986; Nelson et al., 1976) and other cellular processes, as well as within mitochondria by the respiratory


227

FIG. 1. The direct effects of EE, E2, 4OH E2, and 2OH E2 on LDCL production and NADH-dehydrogenase (Complex I) activity. Rat liver mitochondria (500 mg protein equivalent) or freshly isolated intact male rat hepatocytes (3 3 10 6) in suspension were incubated directly with the estrogens at the concentrations indicated. The effects of these agents on LDCL generation and on NADH-dehydrogenase activity were examined as described in Materials and Methods. (A) Effects of various 2-OH-E2 concentrations on LDCL production in freshly isolated rat liver mitochondria, shown as the actual LDCL curves; (B) Effects of EE, E2, 4-OH-E2, and 2-OH-E2 on LDCL production, shown in a numerical form that represents the integration of the areas under the curves plotted as % of untreated control. (A and B each contain data from a single experiment, but are representative of results obtained in at least 3 independent experiments.) (C). Effects of EE, E2, 4-OH-E2 and 2-OH-E2 on complex I activity in isolated rat liver mitochondria; * indicates p , 0.05, n 5 4.

chain activity. It has also been reported that oxidation/reduction of stilbene estrogens can occur in mitochondria (Thomas and Roy, 1995) raising a similar possibility for estrogen catechols. Such a process might also cause increased mitochondrial superoxide levels. To determine whether and to what extent these other potential sources of superoxide contribute to the 2-OH-E2-stimulation of LDCL, we used specific inhibitors of the mitochondrial electron transport. HepG2 cells were treated with 2-OH-E2 (5.0 mM) for 48 h. The cells were then removed from the culture dishes and LDCL was measured in the absence or presence of rotenone (an inhibitor of NADH dehydrogenase, complex I [Slater, 1967]) and myxothiazol (an inhibitor of ubiquinone cytochrome c oxidoreductase, complex

III [Von Jagow and Link, 1986]). As shown in Figure 2, LDCL produced by control HepG2 cells (line 1) was completely inhibited by a combination of rotenone plus myxothiazol (line 2). Treatment with 2-OH-E2 (5.0 mM) for 48 h caused a 3-fold increase in LDCL generation (line 3). This effect was also inhibited by the combination of rotenone plus myxothiazol (line 4) and by rotenone alone (line 5). These results demonstrate that the base line and 2-OH-E2-induced LDCL is derived mainly from superoxide produced by the mitochondrial electron transport chain. We examined the time- and concentration-dependent effects of EE, E2, and the E2-catechol metabolites on mitochondrial superoxide generation in both cultured HepG2 cells and rat

228

CHEN ET AL.

Effects of Estrogen Metabolism Inhibition, Increased/Decreased GSH Levels, and Inhibition of COMT Activity on Mitochondrial O 2 •2 Production

FIG. 2. Inhibition of 2-OH-E2-stimulated superoxide generation by specific inhibitors of mitochondrial electron transport. HepG2 cells were treated with 2-OH-E2 (5.0 mM) for 48 h. The cells were then removed from the culture dishes and LDCL was measured in the absence and presence of rotenone (10 mM) and myxothiazol (10 mM). The data represent the actual LDCL curves. These data are from a single experiment representative of 2 separate experiments.

hepatocytes (Fig. 3). The cells were treated with these compounds at the indicated concentrations for up to 60 h. Figure 3 shows that treatment of the HepG2 cells with 4-OH-E2 (Fig. 3C) or 2-OH-E2 (Fig. 3D) initially suppressed superoxide generation, similar to that seen with isolated mitochondria and hepatocytes in suspension (Figs. 1A and 1B). However, upon continued exposure of the cells to these catechols or to EE (Fig. 3A) or E2 (Fig. 3B), LDCL was enhanced in a concentration-, time-dependent manner, to levels 2- to 3-fold over control. In this figure, each point represents a single determination and reproducibility was determined through replicate experiments. However, to clearly demonstrate that the increases were statistically significant, on several occasions, triplicate determinations were made. In one such experiment, the cpm 3 10 6 for control and EE (2 mM)-treated (30 h) HepG2 cells were 1.03 6 0.11 and 2.34 6 0.16 (mean 6 SD), respectively, p , 0.003). Virtually identical results were obtained with cultured rat hepatocytes (data not shown). The stimulation of superoxide production by the different estrogens/metabolites showed somewhat different time- and dose-dependent patterns. Treatment with the higher doses (e.g., 2.5 mM and 5.0 mM) generally stimulated LDCL at earlier time points and to higher levels than the lower doses (0.25 mM and 0.5 mM), which caused increases but at later time points. We also examined the effects of these estrogens on O 2 consumption. As shown in Fig. 4 for cultured rat hepatocytes, treatment with EE, E2, 2-OH-E2 or 4-OH-E2 caused increased O 2 consumption, following a pattern similar to the increase they caused in mitochondrial superoxide production. These results provide additional support for the origin of the mitochondrial superoxide causing the increased LDCL, as being derived from an estrogen-induced increase in the activity of the mitochondrial electron transport chain.

Previously, we reported that the stimulatory effects of EE and E2 on mitochondrial gene mRNA levels was inhibited by the cytochrome P450 inhibitor, SKF-525 (Chen et al., 1998). We used this same approach to determine whether enhanced LDCL observed in EE- and E2-treated cells also requires metabolism of these estrogens. At 30 mM, SKF-525 inhibited estrogen metabolism .70% in HepG2 cells (Chen et al., 1998). As shown in Figure 5, as expected, E2 (10 mM, 48 h) caused an increase in LDCL, in this case, 3.5-fold. SKF-525 (30 mM) alone had little effect on HepG2 base line LDCL, while it caused a 45% reduction in the level of E2-enhanced LDCL. These results demonstrate that metabolism of E2 by cytochrome P450, presumably to its catechol metabolites, is required for the stimulation of mitochondrial respiratory chain activity. We investigated the effects of causing increased and decreased glutathione (GSH) levels on 4-OH-E2-induced LDCL. The results, shown in Figure 6A, demonstrate that exposure of HepG2 cells for 48 h to 7.5 mM 4-OH-E2 along with 100 mM GSH ester inhibited the induction of LDCL by 65%, as determined by comparison of the areas under the curves. In contrast, co-treatment of cultured hepatocytes with 2.5 mM 4-OH-E2 and 100 mM BSO, to reduce GSH levels, resulted in a 30% increase in 4-OH-E2-induced LDCL (Fig. 6B). The mechanism of inhibition by GSH ester is not unclear. Addition of GSH ester did not affect induction of mitochondrial transcript mRNA levels (data not shown). Given the generally high, mM, levels of GSH within hepatocytes, it is not expected that exposure to 100 mM GSH ester would significantly alter intercellular GSH levels, although changes in GSH compartmentalization cannot be discounted. Additional studies are required to reveal the mechanisms involved. Catechol-o-methyltransferase (COMT) catalyzes the o-methylation of both 2-OH-E2 and 4-OH-E2 which effectively blocks their estrogenic activity and ability to undergo further oxidative metabolism to more reactive semiquinone and quinone metabolites (Creveling, 1994; Zhu and Conney, 1998a). Since our results pointed to a role for catechol estrogens in mediating the estrogeninduced increase in mitochondrial respiratory chain activity, we next investigated the effects of inhibition of COMT on LDCL. Cultured rat hepatocytes and HepG2 cells were treated with 4-OH-E2 (2.5 mM) in the presence or absence of the specific COMT inhibitor, RO 410960 (10 mM) (Mannisto et al., 1992), for 48 h prior to measuring LDCL. At this concentration, RO 410960 inhibits COMT enzyme activity in HepG2 cells by approximately 94% with no detectable cytotoxic effects (Lavigne, unpublished observations). Compared to untreated controls, the COMT inhibitor alone had a small stimulatory effect on LDCL in cultured female rat hepatocytes (Fig. 7A), but no effect in the HepG2 cells (Fig. 7B). As expected, 4-OH-E2 induced mitochondrial super-


229

FIG. 3. The time- and concentration-dependent effects of chronic treatment with EE, E2, 4-OH-E2 and 2-OH-E2 on mitochondrial LDCL in cultured HepG2 cells. Cultured HepG2 cells were treated with EE (A), E2 (B), 4-OH-E2 (C), or 2-OH-E2 (D) at the indicated concentrations for up to 60 h. The medium containing these agents was renewed every 12 h during the culture period. The cells were removed from culture dishes at the indicated time points and assayed for LDCL. The data are shown in a numerical form that represents the integration of the areas under the LDCL curves, plotted as % of control values. These data are from single, representative experiments. Each experiment was independently repeated at least 3 times. In the experiments presented, the actual 0-time cpm values for the LDCL 3 10 6 were: EE, 1.9; E2, 2.8; 4-OH-E2, 2.4; and 2-OH-E2, 3.5.

oxide production in both cell types. In the presence of COMT inhibitor, 4-OH-E2-induced LDCL was enhanced further, 2.2-fold in hepatocytes and 1.6-fold in the HepG2 cells compared to their respective inhibitor-only-treated controls. Similar results were observed in cells treated with E2 and EE in the presence and absence of the COMT inhibitor (data not shown). These results provide additional support for a role of the estrogen catechol metabolites in causing the increased LDCL. Effects of the Antiestrogen ICI 182708 on Mitochondrial O 2 •2 Production The catechol metabolites of E2 are also estrogenic, with the 4-OH-E2 having somewhat greater estrogenicity than 2-OH-E2 (Schutze et al., 1993). In order to determine whether the enhanced LDCL caused by EE, E2, and its catechol metabolites involved the estrogen receptor, we used the specific antiestrogen ICI 182708 (Wakeling, 1993). HepG2 cells were treated with EE (5 mM) in the presence or absence of ICI 182708 (15 mM) for 48 h, at which time the cells were harvested for LDCL measurement. ICI 182708 alone caused a slight inhibition of LDCL (15%),

compared to untreated controls (Fig. 8A). As expected, EE (5 mM) caused a 3.0 fold increase in LDCL. This effect was inhibited by 96% in the presence of 15 mM ICI 182708. The effect of ICI 182708 on EE-induced superoxide generation was concentration-dependent. At a ratio of ICI 182708/EE of 1:1 (5 mM/5 mM), the inhibition was approximately 30% (data not shown). Identical effects of ICI 182708 were observed when 4-OH-E2 was used as the estrogen (data not shown). Additional experiments were conducted to check the specificity of the inhibitory effect of the antiestrogen. First, the addition of ICI 182708 just at the time of the LDCL determination, caused only a slight (;15%) reduction in LDCL (data not shown). This small percentage demonstrates that ICI 182708 has only a small antioxidant effect and that its large inhibition of estrogen-induced superoxide production requires its simultaneous presence in the culture medium. Second, we have observed that ceramide also causes an increase in LDCL in HepG2 cells (Fig. 8B), presumably through a signaling pathway not involving the estrogen receptor. In this experiment, ICI 182708 caused only ;13–15% reduction in LDCL in

230

CHEN ET AL.

FIG. 4. The time effects of chronic treatment with EE, E2, 4-OH-E2, or 2-OH-E2 on mitochondrial O 2 consumption in cultured rat hepatocytes. Cultured rat hepatocytes were treated with EE, E2, 4-OH-E2, or 2-OH-E2. The hepatocytes were removed from culture dishes at the indicated times for determination of O 2 consumption as described in Materials and Methods. The O 2 consumption rates of treated cells are expressed as % of untreated controls. These data are from a single representative experiment where the 15-h control value was 3 nM O 2/min/10 6 cells.

both control and ceramide-treated (5 mM, 48 h) cells, consistent with the small antioxidant properties of ICI 182708.

drial gene transcripts reported previously (Chen et al., 1996, 1998). This important finding demonstrates that the cultured hepatocytes and HepG2 cells represent valid models for investigation of the mechanisms involved. Second, in liver mitochondria isolated from untreated rats and in suspensions of freshly isolated hepatocytes, an acute, 30-min, exposure to the catechol metabolites 2-OH-E2 and 4-OH-E2, but not to E2 itself or EE, caused an inhibition of superoxide production. LDCL was blocked by the addition of specific inhibitors of the electron transport chain, immediately prior to addition of lucigenin. These results indicated that the LDCL was due to superoxide produced by the electron transport chain and suggested that the inhibition of LDCL by the E2 catechols was due to inhibition of electron transport. This was confirmed by the observation that the catechols, but not E2 or EE, inhibited complex I activity. In culture, treatment of HepG2 cells and hepatocytes with the catechol metabolites, and to a lesser extent EE and E2, also caused an inhibition in mitochondrial superoxide production. However, in culture, this inhibition was transient and with continuous treatment was followed by a concentration- and time-dependent increase in mitochondrial LDCL. A similar pattern was observed in oxygen consumption, lending additional support to the conclusion that the estrogen effects on LDCL reflect increased mitochondrial respiratory chain activity. The increase in LDCL was usually first observed after 36 to 48 h of continuous treatment, and generally continued to increase through 60 h. In contrast, estrogen-induced increases in mitochondrial gene transcript levels were first detected after 12 h, continued to increase through 24 h, and then either increased further, held constant, or declined somewhat, depending on the gene, through 48 h (Chen et al., 1998). This overall pattern of an initial inhibition

DISCUSSION

Our overall goal is to understand the toxicological effects of EE that contribute to its strong tumor-promoting potency in liver. Chronic treatment of female rats with EE and other hepatic promoters leads to the onset of a mitosuppressed state. Mitosuppression is characterized by decreased hepatocyte turnover, reduced growth responsiveness, and increased transcript levels of several nuclear and mitochondrial, genome-encoded mitochondrial genes (Barbason et al., 1983; Chen et al., 1996, 1998; Jirtle and Meyer, 1991; Tanaka et al., 1992; Yager et al., 1994). This estrogen-induced increase in mitochondrial transcript levels was also observed in cultured hepatocytes and HepG2 cells, and was accompanied by increased mitochondrial superoxide production, representing increased respiratory chain activity (Chen et al., 1998). Here, we have provided additional characterizations of this effect of estrogens on mitochondria and presented several novel observations. First, we have demonstrated that mitochondria isolated from the livers of EE-treated female rats show increased respiratory chain activity along with the increased levels of the mitochon-

FIG. 5. Inhibition of E2-stimulated mitochondrial superoxide generation by the cytochrome P450 inhibitor, SKF-525. HepG2 cells were treated with E2 (10 mM) for 48 h in the presence and absence of SKF-525 (30 mM). The cells were then removed from culture dishes for LDCL determination. The data are shown as the actual LDCL curves, and are from a single, representative experiment that was repeated 3 times.


231

FIG. 6. Effects of increasing or decreasing glutathione levels on4-OH-E2-induced LDCL. (A) HepG2 cells were treated with 4-OH-E2 (7.5 mM) in the presence and absence of glutathione ester (100 mM) for 48 h. (B) Cultured hepatocytes were treated with 4-OH-E2 (2.5 mM) in the presence or absence of BSO (100 mM), an inhibitor of GSH biosynthesis, for 48 h. The cells were then removed from the culture dishes for LDCL determination. The data represent the actual LDCL curves. Preliminary studies in MCF-7 cells indicated that 50 mM BSO caused a .80% reduction in GSH levels, suggesting the likelihood that 100 mM in HepG2 cells would cause at least this level of reduction in GSH. These data are from single independent experiments, each of which was repeated 2 times.

of superoxide production followed by increased mRNA transcript levels and then by increased mitochondrial superoxide production suggests a cause/effect relationship. As discussed previously (Chen et al., 1998), others have detected increased levels of several mitochondrial gene transcripts following estrogen treatment in various tissues and cell

types (Bettini and Maggi, 1992; Law et al., 1994; and Van Itallie and Dannies, 1988). However, we have extended these observations by demonstrating that this effect is associated with increased mitochondrial respiratory chain activity. Third, our results indicate that cytochrome P450-mediated biotransformation of EE and E2 is required for both the in-

FIG. 7. The effect of COMT inhibition on 4-OH-E2-induced mitochondrial superoxide production. Cultured rat hepatocytes (A) and HepG2 cells (B) were treated with 4-OH-E2 (2.5 mM) in the presence or absence of the specific COMT inhibitor, RO410960 (10 mM), for 48 h. The cells were then removed from dishes for LDCL determination. The data represent the actual LDCL curves. These data are from separate independent experiments, each of which were repeated 2 times.

232

CHEN ET AL.

FIG. 8. The effect of the anti-estrogen ICI 182708 on EE-induced mitochondrial superoxide production. Cultured HepG2 cells were treated with either EE (5 mM) (A) or C2-ceramide (5 mM) (B) in the presence and absence of ICI 182708 (15 mM) for 48 h. The cells were then removed from the dishes for LDCL determinations. The data represent the actual LDCL curves. These data are from separate, independent experiments. The experiment presented in A was repeated 3 times and that in B once.

creases in mitochondrial transcript levels (Chen et al., 1998) and respiratory chain activity. The HepG2 cells that are not induced to increase P450 metabolism convert 8 to 10% of ethyl acetate extractable 3H-E2 to non-ethyl acetate extractable, water-soluble metabolites within 180 min (data not shown). Treatment with 30 mM SKF-525 caused a .70% inhibition in E2 metabolism over this time period (Chen et al., 1998). Continuous exposure to E2 plus SKF-525 caused a 90 –95 % inhibition in estrogen-induced mitochondrial transcript levels (Chen et al., 1998) and a 45% inhibition in mitochondrial superoxide production (Fig. 5), providing solid support for estrogen metabolites causing these mitochondrial effects. Two additional findings suggest that the metabolites responsible for increased LDCL are the estrogen catechols. First, treatment of HepG2 cells and rat hepatocytes with the catechols themselves, either 2-OH-E2 or 4-OH-E2, caused increased mitochondrial gene transcript levels and respiratory chain activity. The second pertains to the results observed by inhibiting COMT activity. Catechol-o-methyltransferase catalyzes the o-methylation of both 2-OH- and 4-OH-E2, effectively eliminating their estrogenicity while at the same time producing potentially protective methoxy estrogen metabolites (Creveling, 1994; Zhu and Conney, 1998a,b). Inhibition of COMT would be expected to result in increased cellular levels of these pro-oxidant catechol metabolites. Although we have not yet definitively demonstrated this in these cells, inhibition of COMT does increase the levels of catechol metabolites in human breast cancer cells (Lavigne, unpublished observations). In our study, we found that inhibition of COMT activity in estrogen-treated hepatocytes and HePG2 cells increased mitochondrial superoxide production. This increase was not caused by further oxidative metabolism of the catechols but by

superoxide produced by the respiratory chain. This is supported by the observations, mentioned previously, that the increase in superoxide was inhibited by specific respiratory chain inhibitors and associated with increased oxygen utilization. As also mentioned above, we did observe an initial, transient inhibition of mitochondrial respiration caused by the catechols, but not the parent estrogens, most likely due to direct inhibition of complex I by the catechols. This suggested the hypothesis that the subsequent increase in mitochondrial gene transcript levels, followed by respiratory chain activity, may be a direct response to this initial inhibitory effect. As a result of this hypothesis, we did not expect that these mitochondrial effects would be mediated through the estrogen receptor. However, our fourth major finding pertains to the estrogen receptor in the mitochondrial response to estrogen treatment. HepG2 cells express high affinity nuclear estrogen receptors (ER) and are estrogen-responsive with regard to growth and apolipoprotein mRNA transcription and secretion (Coezy et al., 1987; Jin et al., 1998; Tam et al., 1985). In order to test our idea that the mitochondrial response to estrogen would occur independent of the ER, we exposed HepG2 cells to EE 6 the specific anti-estrogen ICI 182708. Unexpectedly, the EE-induced increase in LDCL was completely inhibited. In other studies, we observed that treatment of HepG2 cells with ceramide also caused an increase in mitochondrial superoxide production, presumably through a non-estrogen receptor-mediated process. As expected, we found that the ceramideinduced increase was not inhibited by ICI 182708, demonstrating its specificity for blocking the induction of mitochondrial respiratory chain activity by EE. These results altered our concept regarding the signaling pathways involved and led to development of the working hypothesis shown in Figure 9. At


FIG. 9. Working hypothesis for the mechanism of increased mitochondrial transcript levels and respiratory chain activity in liver cells by estrogens.

this point, our results indicate that the induction of mitochondrial gene expression and subsequent increase in respiratory chain activity require the cytochrome P450-mediated biotransformation of EE and E2 to catechol metabolites, which then bind to estrogen receptors to induce increased expression of mitochondrial genes. This is followed by increased respiratory chain activity. It is likely that this effect is mediated by the ER-mediated induction of a mitochondrial gene transcription factor (Larsson et al., 1996; Parisi and Clayton, 1991). It is also possible that nuclear genome-encoded mitochondrial genes have estrogen-receptor binding consensus sequences in their promoter regions and that this may account for the increased expression of these genes. It is possible too that the induction of these genes by the catechol metabolites is mediated through the ligand-estrogen receptor complex acting through another regulatory element, such as an AP1 site (Paech et al., 1997). There is also a possibility that estrogens may have direct effects on mitochondria. For glucocorticoids, sequences with at least partial similarity to glucocorticoid-receptor response elements are present in the mitochondrial genome (Demonacos et al., 1996; Ioannou et al.,1988), and there is evidence that the glucocorticoid receptor is rapidly imported into liver mitochondria after dexamethasone treatment (Demonacos et al., 1993). While there is no evidence for a similar process occurring with the estrogen receptor, there are mitochondrial DNA sequences that show partial similarity to the estrogen response element consensus sequence (Demonacos et al., 1996). Future studies will explore the mechanisms of estrogen effects on mitochondrial transcript levels. In summary, our results demonstrate that estrogen treatment of rats in vivo and liver cells in culture results in increased mitochondrial gene expression followed by increased mitochondrial superoxide production, reflecting increased respiratory chain activity. This estrogen effect requires metabolism to catechols and is mediated through the estrogen receptor. Some

233

evidence suggests that these mitochondrial effects could have a role in estrogen-induced mitosuppression and tumor promotion. As mentioned previously, other hepatic promoters such as phenobarbital and certain peroxisomal proliferators also induce mitosuppression in rat liver. TGF-b is a known inhibitor of growth and inducer of apoptosis in hepatocytes (Oberhammer and Qin, 1995). Interestingly, phenobarbital, nafenopin, methyl clofenapate and Wy-14,643 inhibit spontaneous and TGF-binduced apoptosis in rat hepatocytes (Baylay, et al., 1994; James and Roberts, 1996; Oberhammer and Qin, 1995). Although the exact mechanism is unknown, apoptosis inhibition caused by the peroxisome proliferators occurs through the peroxisome-proliferator receptor alpha (Roberts et al., 1998). Whether or not these agents affect mitochondrial function is unknown, although it is known that mitochondria have an important regulatory role in initiating the apoptotic process (Green and Reed, 1998). Estradiol has been shown to inhibit apoptosis in human breast cancer cells (Wang and Phang, 1995), and preliminary experiments in our laboratory indicate that EE also inhibits spontaneous and TGF-b-induced apoptosis (Gokhale et al., unpublished observations). It has also been reported that estradiol can stabilize mitochondrial function and protect PC12 neural cells from induction of apoptosis by mutant presenilin-1 (Mattson et al., 1997). As mentioned previously, mitosuppression is associated with increased expression of TGF-b (Chen et al., 1996; Jirtle et al., 1991), which would be expected to cause both growth inhibition and increased apoptosis. However, the ability of these tumor promoters to inhibit apoptosis would confine the TGF-b effects to growth inhibition, accounting for mitosuppression. Studies are underway to define the relationship between EE-induced mitosuppression, inhibition of apoptosis, and increased mitochondrial respiratory chain activity in liver cells in EE hepatocarcinogenesis. ACKNOWLEDGMENTS This research was supported by US PHS grant R01 CA 36701. Maintenance and use of shared equipment was supported by an NIEHS Center Grant P30 ES03819. JC was a postdoctoral trainee and JAL a doctoral trainee, both of whom were supported by NIEHS Training Grant T32 ES 07141.

REFERENCES Barbason, H., Rassenfosse, C., and Betz, E. H. (1983). Promotion mechanism of phenobarbital and partial hepatectomy in DENA hepatocarcinogenesis cell kinetic effect. Br. J. Cancer 47, 517–525. Bayly, A. C., Roberts, R. A., and Dive, C. (1994). Suppression of liver cell apoptosis in vitro by the non-genotoxic hepatocarcinogen and peroxisome proliferator nafenopin. J. Cell Biol. 125, 197–203. Bettini, E., and Maggi, A. (1992). Estrogen induction of cytochrome c oxidase subunit III in rat hippocampus. J. Neurochem. 58, 1923–1929. Cavalieri, E. L., Stack, D. E., Devanesan, P. D., Todorovic, R., Dwivedy, I., Higginbotham, S., Johansson, S. L., Patil, K. D., Gross, M. L., Gooden,

234

CHEN ET AL.

J. K., Ramanathan, R., Cerny, R. L., and Rogan, E. G. (1997). Molecular origin of cancer: Catechol estrogen 3,4-quinones as endogenous tumor initiators. Proc. Natl. Acad. Sci. U S A 94, 10937–10942. Chen, J. Q., Gokhale, M., Li, Y., Trush, M. A., and Yager, J. D. (1998). Enhanced levels of several mitochondrial mRNA transcripts and mitochondrial superoxide production during ethinyl estradiol-induced hepatocarcinogenesis and after estrogen treatment of HepG2 cells. Carcinogenesis 19, 2187–2193. Chen, J. Q., Schwartz, D. A., Young, T. A., Norris, J. S., and Yager, J. D. (1996). Identification of genes whose expression is altered during mitosuppression in livers of ethinyl estradiol-treated female rats. Carcinogenesis 17, 2783–2786. Coezy, E., Auzan, C., Lonigro, A., Philippe, M., Menard, J., and Corvol, P. (1987). Effect of mestranol on cell proliferation and angiotensinogen production in HepG2 cells: Relation with the cell cycle and action of tamoxifen. Endocrinology 120, 133–141. Creveling, C. R. (1994). Catechol-O-methyltransferase: Factors relating to the carcinogenic potential of catecholestrogens. Polycyclic Aromatic Compds. 6, 253–259. Demonacos, C. V., Karayanni, N., Hatzoglou, E., Tsiriyiotis, C., Spandidos, D. A., and Sekeris, C. E. (1996). Mitochondrial genes as sites of primary action of steroid hormones. Steroids 61, 226 –232. Demonacos, C., Tsawdaroglou, N., Djordjevic-Markovic, R., Papalopoulou, M., Galanopoulos, V., Papadogeorgaki, S., and Sekeris, C. E. (1993). Import of glucocorticoid receptor into rat liver mitochondria in vivo and in vitro. J. Steroid Biochem. Molec. Biol. 46, 401– 414. Farber, E. (1990). Clonal adaptation during carcinogenesis. Biochem. Pharmacol. 39, 1837–1846.

Liehr, J. G., and Roy D., (1990). Free radical generation by redox cycling of estrogens. Free Radic. Biol. Med. 8, 415– 423. Liehr, J. G., Ulubelen, A. A., and Strobel H. W. (1986). Cytochrome P-450mediated redox cycling of estrogens, J. Biol. Chem. 36, 16865–16870. Mannisto, P. T., Ulmanen, I., Lundstrom, K., Taskinen, J., Tenhunen, J., Tilgmann, C., and Kaakkola, S. (1992). Characteristics of catechol Omethyltransferase (COMT) and properties of selective COMT inhibitors. Prog. Drug Res. 39, 291–350. Mansbach, J. M., Mills, J. J., Boyer, I. J., DeSouza, A. T., Hankins, G. R., and Jirtle, R. L. (1996). Phenobarbital selectively promotes initiated cells with reduced TGF-b receptor levels. Carcinogenesis 17, 171–174. Mattson, M. P., Robinson, N., and Guo, Q. (1997). Estrogens stabilize mitochondrial function and protect neural cells against the pro-apoptotic action of mutant presenilin-1. Neuroreport 17, 3817–3821. Mayol, X., Neal, G. E., Davies, R., Romero, A., and Domingo, J. (1991). Ethinyl estradiol-induced cell proliferation in rat liver: Involvement of specific populations of hepatocytes. Carcinogenesis 13, 2381–2388. Nelson, S. D., Mitchell, J. R., Dybing, E., and Sasame, H. A. (1976). Cytochrome P-450-mediated oxidation of 2-hydroxyestrogens to reactive intermediates, Biochem. Biophys. Res. Commun. 70, 1157–1165. Oberhammer, F. A. and Qin, H.-M. (1995). Effect of three tumour promoters on the stability of hepatocyte cultures and apoptosis after transforming growth factor-b1. Carcinogenesis 16, 1363–1371. Paech, K., Webb, P., Kuiper, G. G. J. M., Nilsson, S., Gustafsson, J.-A., Kushner, P. J., and Scanlan, T. S. (1997). Differential ligand activation of estrogen receptors ERa and ERb at AP1 sites. Science 277, 1508 –1510.

Green, D. R., and Reed, J. C. (1998). Mitochondria and apoptosis. Science 281, 1309 –1312.

Palmer, J. R., Rosenberg, L., Kaufman, D. W., Warshauer, M. E., Stolley, P., and Shapiro, S. (1989). Oral contraceptive use and liver cancer. Am. J. Epidemiol. 130, 878 – 882.

Ioannou, I. M., Tsawdaroglou, N., and Sekeris, C. (1988). Presence of glucocorticoid responsive elements in the mitochondrial genome. Anticancer Res. 8, 1405–1410.

Parisi, M. A., and Clayton, D. A. (1991). Similarity of human mitochondrial transcription factor 1 to high-mobility group proteins. Science 252, 969 – 969.

James, N. H., and Roberts, R. A. (1996). Species differences in response to peroxisome proliferators correlate in vitro with induction of DNA synthesis rather than suppression of apoptosis. Carcinogenesis 17, 1623–1632.

Pedersen, P. L., Greenawalt, J. W., Reynafarje, B., Hullihen, J., Decker, G. L., Soper, J. W., and Bustamente, E. B. (1978). Preparation and characterization of mitochondria and submitochondrial particles of rat livers and liverderived tissues. Methods Cell. Biol. 20, 411– 481.

Jin, F.-Y., Kamanna, V. S., and Kashyap, M. L. (1998). Estradiol stimulates apolipoprotein A-I- but not A-II-containing particle synthesis and secretion by stimulating mRNA transcription rate in Hep G2 cells. Aterioscler. Thromb. Vasc. Biol. 18, 999-1006. Jirtle, R. L. Hankins, G. R., Reisenbichler, H., and Boyer, I. J. (1994). Regulation of mannose 6-phosphate/insulin-like growth factor-II receptors and transforming growth factor beta during liver tumor promotion with phenobarbital. Carcinogenesis 15, 1473–1478. Jirtle, R. L., and Meyer, S. A. (1991). Liver tumor promotion: Effect of phenobarbital on EGF and protein kinase C signal transduction and transforming growth factor-beta expression. Dig.. Dis. Sci. 36, 659 – 668. Larsson, N.-G., Garman, J. D., Oldfors, A., Barch, G. S., and Clayton, D. A. (1996). A single mouse gene encodes the mitochondrial transcription factor A and a testis-specific nuclear HMG-box protein. Nat. Genet. 13, 296 –302. Law, S. W., Apostolakis, E.M., Samora, P. J., O’Malley, B. W., and Clark, J. H. (1994). Hormonal regulation of hypothlamic gene expression: Identification of multiple novel estrogen-induced genes. J. Steroid Biochem. Mol. Biol. 51, 131–136. Li, Y., Trush, M. A., and Yager, J. D. (1994). DNA damage caused by reactive oxygen species originating from a copper-dependent oxidation of the 2-hydroxyl catechol of estradiol. Carcinogenesis 15, 1421–1427. Li, Y., Zhu, H., Kuppusamy, P., Roubaud, V., Zweier, J. L., and Trush, M. A. (1998). Validation of lucigenin (bis-N-methylacridinium) as a chemilumigenic probe for detecting superoxide anion radical production by enzymatic and cellular systems. J. Biol. Chem. 273, 2015–2023.

Roberts, R. A., James, N. H., Woodyatt, N. J., Macdonald, N., and Tugwood, J. D. (1998). Evidence for the suppression of apoptosis by the peroxisome proliferator-activated receptor alpha (PPARa). Carcinogenesis 19, 43– 48. Schutze, N., Vollmer, G., Tiemann, I., Geiger, M., and Knuppen, R. (1993). Catechol estrogens are MCF-7 cell estrogen receptor agonists. J. Steroid Biochem. Mol. Biol. 46, 781– 89. Seglen, P. O. (1976). Preparation of isolated rat liver cells. Methods Cell. Bio. 13, 29 – 83. Singer, T. P. (1994). Determination of the activity of succinate, NADH, choline, and a-glycerophosphate dehydrogenase. Methods Biochem. Anal. 22, 123–175. Slater, E. C. (1967). Application of inhibitors and uncouplers for a study of oxidative phosphorylation. Methods Enzymol. 10, 48 –57. Tam, S.-P., Archer, T. K., and Deeley, R. G. (1985). Effects of estrogen on apolipoprotein secretion by the human hepatocarcinoma cell line HepG2. J. Biol. Chem. 260, 1670 –1675. Tanaka, K., Smith, P. F., Stromberg, P. C., Eydelloth, R. S. Herold, E. G., Grossman, S. J, Frank, J. D., Hertzog, P. R., Soper, K. A., and Keenan, K. P. (1992). Studies of early hepatocellular proliferation and peroxisomal proliferation in Sprague-Dawley rats treated with tumorigenic doses of clofibrate, Toxicol. Appl. Pharmacol. 116, 71–77. Thomas, R. D., and Roy, D. (1995). Mitochondrial enzyme-catalyzed oxida-

MITOCHONDRIAL O 2 •2 PRODUCTION AFTER ETHINYL ESTRADIOL TREATMENT tion and reduction reactions of stilbene estrogen. Carcinogenesis 16, 891– 895. Van Itallie, C. M., and Dannies, P. S. (1988). Estrogen induces accumulation of the mitochondrial ribonucleic acid for subunit II of cytochrome oxidase in pituitary cells. Mol. Endocrinol. 2, 332–337. von Jagow, G., and Link, T. A. (1986). Use of specific inhibitors on the mitochondrial bc 1 complex. Methods Enzymol. 126, 253–271. Wakeling, A. E. (1993). The future of new pure antiestrogens in clinical breast cancer. Breast Cancer Res. Treat. 25, 1–9. Wang, T. T. Y., and Phang, J. M. (1995). Effects of estrogen on apoptotic pathways in human breast cancer cell line MCF-7. Cancer Res. 55, 2487– 2489. Yager, J. D., and Liehr, J. G. (1996). Molecular mechanisms of estrogen carcinogenesis. Ann. Rev. Pharmacol. Toxicol. 36, 203–232. Yager, J. D., Roebuck, B. D., Puluszcyk, T. L., and Menoli, V. A. (1986). Effects of ethinyl estradiol and tamoxifen in liver DNA turnover and new

235

synthesis and appearance of gamma glutamyl transpeptidase-positive foci in female rats. Carcinogenesis 7, 2007–2014. Yager, J. D., Zurlo, J., Sewall, C. H., Lucier, G. W., and He, H. (1994). Growth stimulation followed by growth inhibition in livers of female rats treated with ethinyl estradiol. Carcinogenesis 15, 2117–2123. Zhu, B. T., and Conney, A. H. (1998a). Functional role of estrogen metabolism in target cells, review and perspectives. Carcinogenesis 19, 1–27. Zhu, B. T., and Conney A. H. (1998b). Is 2-methoxyestradiol an endogenous estrogen metabolite that inhibits mammary carcinogenesis? Cancer Res. 58, 2269 –77. Zhu, B. T., and Liehr, J. G. (1994) Quercetin increased the severity of estrodiol-induced tumorigenesis in hamster kidney. Toxicol. Appl. Pharmacol. 125, 149 –158. Zurlo, J., and Arterburn, L. M. (1996). Characterization of a primary hepatocyte culture system for toxicological studies. In vitro Cell. Dev. Biol. Anim. 32, 211–220.