TOXICOLOGICAL SCIENCES 76, 262–270 (2003) DOI: 10.1093/toxsci/kfh001

Epigallocatechin Gallate Modulates CYP450 Isoforms in the Female Swiss-Webster Mouse

Mette G. Goodin and Rhonda J. Rosengren1 Department of Pharmacology and Toxicology, University of Otago, Dunedin, New Zealand Downloaded from https://academic.oup.com/toxsci/article/76/2/262/1686012 by guest on 28 September 2021 Received August 4, 2003; accepted September 22, 2003 and epicatechin (EC). There has been scientific interest in this This study was designed to determine the effect of the in vivo chemical family due to their various pharmacological actions. administration of (EGCG) and epicat- For example, the inhibit the proliferation of human echin gallate (ECG) on involved in the synthesis and breast cells (Kavanagh et al., 2001; Morre et al., 2000; of . EGCG (12.5, 25, or 50 mg/kg/day, ip) or ECG (12.5 or 25 mg/kg/day, ip) was administered to female Swiss- Valcic et al., 1996) and epidemiological studies have suggested Webster mice for 7 days. The chemicals were well tolerated by the that green consumption is linked to a decrease in both the mice with the exception of EGCG given at 50 mg/kg, which rate of development and recurrence of (Nakachi resulted in severe hepatic necrosis and a 67% mortality rate. et al., 1998). Additionally, in the rat DMBA mammary carci- Following the administration of nontoxic doses of EGCG and nogenesis model, the administration of green-tea extract has ECG, (CYP19), CYP3A, CYP1A, and catechol decreased both tumor number (Tanaka et al., 1997) and weight O-methyltransferase (COMT) were measured. Additionally, the (Kavanagh et al., 2001). However, the effect of purified cat- activity of CYP2E1 was determined, since this CYP450 isoform is echins on tumor growth has only been examined in mice. The important in the bioactivation of numerous . The re- results demonstrated that EGCG caused regression of MCF-7 sults demonstrated that ovarian aromatase activity was inhibited 56% by EGCG (25 and 12.5 mg/kg), but not ECG, while hepatic cell implanted tumors and reduced the growth of implanted CYP3A catalytic activity and polypeptide levels were increased prostate tumors (Liao et al., 1995). The effect on prostate and 47 ؎ 2%, respectively, by 25 mg/kg of EGCG. tumors was -specific as EC, EGC, and ECG all failed 4 ؎ 31 However, ECG (but not EGCG) inhibited CYP1A catalytic activ- to decrease tumor growth. However, EGCG was the only ity and polypeptide levels (31 ؎ 5 and 47 ؎ 5%, respectively). catechin examined in the mammary tumor model. Overall, the Hepatic and renal COMT, as well as renal CYP3A remained results demonstrate that EGCG may have beneficial character- unchanged following catechin dosing. Hepatic CYP2E1 catalytic istics that could be exploited in the treatment of breast cancer. ؎ activity and polypeptide levels were significantly increased (37 Numerous studies with either green-tea extract or specific ؎ 3 and 22 3%) following administration of EGCG (25 mg/kg). members of the catechin family have postulated a wide variety These results indicate that EGCG modulates enzymes responsible of potential mechanisms of action, such as receptor for both the synthesis and metabolism of estradiol, which may provide a potential mechanism for the reported action of EGCG, (ER) antagonism (Komori et al., 1993), proapoptotic (Morre et reported action as an inhibitor of breast tumor growth. al., 2000), antiangiogenic (Kondo et al., 2002) and anti-oxida- Key Words: epigallocatechin gallate, CYP19, CYP3A, CYP1A, tive (Guo et al., 1999). None of these mechanisms have been catechol O-methyltransferase. conclusively proven in either breast cancer cells or breast tumor models. However, ER antagonism is not likely to be involved, since EGCG, EGC, and ECG all failed to antagonize ␣ The catechins are a group of polyphenolic chemicals that estradiol-induced responses in both uterotropic and ER re- belong to the flavanol family, which is a subclass of flavonoids. porter assays (Goodin et al., 2002). Alternatively, alter- are plant-derived chemicals that are contained in a ations in the synthesis and/or metabolism of estradiol may be variety of foods and beverages (Arts et al., 2000). In these involved in the antitumor actions of the catechins. Therefore, sources, the predominant catechins are epigallocatechin gallate aromatase (CYP19) would be a key target to examine, as it is (EGCG), (ECG), epigallocatechin (EGC), responsible for the synthesis of from and aromatase expression in breast tumor epithelial cells positively

1 correlates with cellular proliferation (Lu et al., 1996). CYP3A To whom correspondence should be addressed at the Department of Phar- macology and Toxicology, 18 Frederick Street, Adams Building, University and CYP1A are also important targets, as these CYP450 iso- of Otago, Dunedin, New Zealand. Fax: ϩ64 3 479 9140. E-mail: forms are responsible for the formation of the antiestrogenic [email protected]. metabolite, 2-hydroxyestradiol (Kerlan et al., 1992; Shou et Toxicological Sciences 76(2), © Society of Toxicology 2003; all rights reserved. al., 1997). This metabolite is then rapidly converted by cate-

262 EGCG MODULATES CYP450 ISOFORMS 263

chol O-methyltransferase (COMT) to 2-methoxyestradiol, methasone (75 mg/kg, ip, 4 consecutive days), or (4.8 g/kg, po, 16 h which displays antiangiogenic properties (Fotsis et al., 1994; prior to necropsy). Klauber et al., 1997). Therefore, the overall balance of the Plasma markers of hepatic and renal injury. Plasma ALT activity and activity of these enzymes is important in the regulation of BUN were used as indicators of hepatic and renal damage, respectively. Immediately following euthanasia, blood was collected from the inferior vena estradiol-mediated breast tumor growth. cava and stored on ice. Plasma ALT activity was determined kinetically, using The only previous reports of catechin-mediated alterations a Sigma diagnostic kit, and the results are expressed as IU/l. BUN was of CYP450 isoforms are restricted to either in vitro investiga- determined using a Sigma diagnostic kit, and the results are expressed as tions or to in vivo studies in which animals were treated with mg/dl. EC, (ϩ)-catechin, or herbal supplements containing a mixture histology. Liver slices were obtained from the distal portion of the of various plant extracts (Lhoste et al., 2003; Muto et al., 2001; left lateral lobe and the tissue was fixed for at least 48 h in 10% neutral buffered formalin. The samples were then embedded in paraffin, cut into 5 ␮m

Ryu and Chung, 2003; Siegers et al., 1982; Wang et al., 1988). Downloaded from https://academic.oup.com/toxsci/article/76/2/262/1686012 by guest on 28 September 2021 sections, and stained with hematoxylin and eosin for examination by light Therefore, the aim of the present study was to determine if microscopy. The examiner of the slides was blind to the various treatment either EGCG or ECG would modulate the effect of enzymes groups. involved in the synthesis and metabolism of estradiol, namely, Tissue preparation. Hepatic and renal cytosolic extracts and microsomes aromatase, CYP1A, CYP3A, or COMT, at doses that were were prepared from individual mice by differential centrifugation (Guengerich, previously reported to cause tumor regression in athymic mice 1989). Ovarian tissue was pooled from 3 mice and microsomes were prepared (Liao et al., 1995). Since COMT inhibition has led to an as described (Kellis and Vickery, 1984). concentration of the resulting microsomes or cytosolic extract was immediately determined by the bicincho- increase in the renal carcinogenic potential of estradiol (Zhu ninic acid method (Smith et al., 1985). The microsomes and cytosolic extracts and Liehr, 1994), the effect of these two catechins on renal were stored at –80°C until experimentation. enzymes was also determined. Additionally, catechin-mediated Aromatase catalytic activity. The release of tritiated water from [3H]- changes in hepatic CYP2E1 were characterized, as this was used as an indicator of aromatase activity, as this assay is expressed in breast tumors (Iscan et al., 2001) and also has been validated as an appropriate indicator of murine ovarian aromatase activates numerous carcinogens (Constan et al., 1999; Sohn et activity (Toda et al., 2001). The assay was conducted as described (Vinggaard et al., 2000) with the following modifications. The incubation mixture con- al., 2001). EGCG and ECG were used in this investigation, as 3 tained 100 nM [ H]-androstenedione, 25 ␮g protein, 250 ␮M NADPH, and 50 catechins that contain a gallate group in the 3Ј position dem- mM KH2PO4 buffer in a total volume of 500 ␮l. Samples were preincubated onstrate potent inhibition of CYP450 in vitro (Muto et al., for 2 min at 37°C, and the reaction was initiated by the addition of NADPH. 2001; Wang et al., 1988). After 10 min, the reaction was terminated by the addition of and 0.9% NaCl. The samples were vortexed for 30 s and then centrifuged at 1700 ϫ g for 15 min. One hundred ␮l of the aqueous phase was transferred to MATERIALS AND METHODS scintillation vials and the radioactivity was counted on a Beckman LS3801 scintillation counter. Results are expressed as pmol/mg/h. Chemicals. EGCG, ECG, ␤-naphthoflavone, dexamethasone, acetone, CYP1A catalytic activity. Ethoxyresorufin O-deethylation was used to NADPH, alanine aminotransferase (ALT) kit, blood urea nitrogen (BUN) kit, determine changes in the catalytic activity of CYP1A (Ryan and Levin, 1990) S-adenosyl-L- (SAM), antirabbit IgG, erythromycin, , and and was performed as described previously (Bray et al., 2002). Results are ethoxyresorufin were purchased from Sigma Chemical Co. (St. Louis, MO). expressed as nmol/mg/min. 3 [ H]-androstenedione (25.9 Ci/mmol) was purchased from New England Nu- CYP3A catalytic activity. Erythomycin N-demethylation was used as a clear (Boston, MA). Nitroblue tetrazolium chloride (NBT), 5-bromo-4-chloro- selective probe for changes in the catalytic activity of CYP3A (Wrighton et al., 3-indoyl phosphate-p-toluidine salt (BCIP), avidin-alkaline phosphatase con- 1985) and was performed as described previously (Bray et al., 2002). Results jugate, biotinylated sodium dodecylsulfate-polyacrylamide gel electrophoresis are expressed as nmol/mg/min. (SDS-PAGE) broad-range standards, and nitrocellulose were purchased from Bio Rad (Hercules, CA). Primary antibodies antirat CYP1A, CYP3A, and CYP2E1 catalytic activity. Aniline was used as a selective CYP2E1 were purchased from Amersham (Auckland, N.Z.). All other chem- probe for changes in the catalytic activity of CYP2E1 and was performed as icals were of the highest purity commercially available. described previously (Inder et al., 1999). Results are expressed as nmol/mg/ min. Animals. Female Swiss-Webster mice (5–6 weeks old) were purchased from Hercus Taieri Resource Unit (Dunedin). All procedures were approved COMT catalytic activity. Methylation of 4-nitrocatechol was used as a by the University of Otago Animal Ethics Committee (#96/01). The animals selective probe for changes in COMT activity (Herblin, 1973) with the fol- were housed in micro-isolator cages on shredded paper bedding and had lowing modifications. Hepatic or renal cytosolic extract (2 mg) was incubated ␮ continuous access to Reliance rodent diet (Dunedin, NZ) and water. They were with 25 M 4-nitrocatechol, 0.2 mM SAM, 0.01 M MgCl2,and1mM maintained at 21–24°C with a 12-h light/dark cycle and allowed to acclimatize Tris-HCl buffer (pH 7.0) in a total volume of 2 ml. Samples were preincubated for one week before experimentation. Mice were randomly assigned to the for 2 min at 37°C and a 60-min incubation period was initiated by the addition various treatment groups, which consisted of 9 mice in each group. Mice were of SAM. The reaction was terminated by addition of 12 N NaOH and the ϫ dosed with EGCG (12.5, 25, or 50 mg/kg/day, ip), ECG (12.5 or 25 mg/kg, ip), samples were centrifuged at 1000 g for 10 min. Absorbance was determined or saline control (8 ml/kg) for 7 days. To control for potential catechin- at 520 nm and the results are expressed as nmol/mg/min. mediated weight loss, saline control mice were pair-fed. Mice were sacrificed Electrophoresis and Western blotting. SDS–PAGE was performed as

by CO2 inhalation 24 h following the final catechin dose. The doses and route previously described (Bray et al., 2002). Briefly, 10 ␮g of microsomal protein of administration used were based on previous work, which demonstrated that was loaded onto a 10% gel with a 4% stacking gel. Polypeptide levels of EGCG (50 mg/kg/day, 14 days, ip) caused tumor regression in a murine CYP1A, CYP3A, and CYP2E1 were quantified by Western immunoblotting as MCF-7 cell implant model (Liao et al., 1995). Mice used as positive controls previously described (Bray et al., 2002). Upon transfer to a nitrocellulose were treated with ␤-naphthoflavone (80 mg/kg, ip, 3 consecutive days), dexa- membrane, were incubated with CYP1A, CYP3A, or CYP2E1 antirat 264 GOODIN AND ROSENGREN primary antibody. Each antibody was proven by the manufacturer to cross- react with the specific mouse CYP450 isoform examined. After washing, membranes were incubated with antirabbit IgG (an alkaline phosphate conju- gate) secondary antibody. The bands were visualized using the NBT/BCIP system and quantified by scanning densitometry. Statistical analysis. Individual groups were analyzed using a two-way ANOVA coupled with the Student-Newman-Keuls post hoc test with p Ͻ 0.05 as the minimum requirement for a statistically significant difference.

RESULTS

EGCG (50 mg/kg/d) administration resulted in a mortality Downloaded from https://academic.oup.com/toxsci/article/76/2/262/1686012 by guest on 28 September 2021 rate of 67%, as 5 mice died after either 4 or 5 days of EGCG administration and one mouse appeared sick and was eutha- nized on day 5. EGCG-induced significant weight loss and extensive as indicated by a mean body weight (bw) loss of 2.4 g, elevated plasma ALT activity (2123 Ϯ 1040 IU/l) (Table 1) and extensive hepatic necrosis throughout the entire liver (Fig. 1), while pair-fed vehicle control mice re- mained normal. Spleen, kidney, and liver weights were also increased (82, 23, and 19%, respectively) following EGCG (50 mg/kg) treatment, while the gross weight of the uterus and ovaries remained unchanged (Table 2). However, there was no major renal impairment as the levels of BUN (20–28 mg/dl) remained in the normal range (Table 2). The toxicity elicited by EGCG was dose-dependent, as lower doses of EGCG (25 and 12.5 mg/kg) produced normal ALT activities, normal liver histology, and no change in bw or organ weight. Since 50 mg/kg of EGCG produced extensive mortality, enzymatic parameters were only examined following lower doses of both EGCG and ECG (25 and 12.5 mg/kg). The FIG. 1. Hepatotoxicity following EGCG administration in female mice: results showed that both doses of EGCG decreased ovarian (A) Representative liver histopathology (H&E stain, ϫ110 magnification), following 7 days of EGCG (50 mg/kg). Severe hepatic necrosis is evident aromatase activity approximately 56% from control, while throughout the entire liver section; only a few viable hepatocytes remain there was no change in aromatase following ECG treatment (arrows). (B) Liver section from a pair-fed vehicle control mouse; the hepa- (Fig. 2). EGCG (25 mg/kg) also increased hepatic CYP3A tocytes all appear normal. catalytic activity 31 Ϯ 4% above control (Fig. 3). This was supported by a 47 Ϯ 2% increase in CYP3A polypeptide levels (Fig. 4). This response was dose-dependent, as the catalytic TABLE 1 activity and polypeptide levels for CYP3A remained un- Mortality, ALT Activity, BUN, and Body-Weight Change changed following 12.5 mg/kg of EGCG. The ability to mod- following Catechin Treatment ulate CYP3A was both catechin- and organ-specific, as ECG did not alter hepatic CYP3A (Figs 3 and 4), and renal CYP3A Mortality ALT BUN Body-weight Treatment (dose) (%) (IU/l) (mg/dl) change (g) remained unchanged following the administration of both cat- echins (data not shown). Interestingly, ECG (25 mg/kg) de- Saline (8 ml/kg) 0 25 Ϯ 324Ϯ 2 1.8 Ϯ 0.3 creased hepatic CYP1A catalytic activity by 31 Ϯ 5% of EGCG (12.5 mg/kg) 0 23 Ϯ 327Ϯ 3 1.3 Ϯ 0.4 control (Fig. 5) and polypeptide levels by 47 Ϯ 5% (Fig. 6), Ϯ Ϯ Ϯ EGCG (25 mg/kg) 0 32 5253 2.1 0.5 while EGCG failed to alter the levels of CYP1A. Examination EGCG (50 mg/kg) 67 2123 Ϯ 1040* 22 Ϯ 4 –2.4 Ϯ 1.1** ECG (12.5 mg/kg) 0 24 Ϯ 428Ϯ 3 1.5 Ϯ 0.4 of renal CYP1A failed to determine a catechin-specific effect, ECG (25 mg/kg) 0 25 Ϯ 524Ϯ 3 1.5 Ϯ 0.4 as renal CYP1A remained below levels that could be detected by either catalytic or Western immunoblotting techniques. Note. Mice were treated as indicated for 7 days. Each value represents the However, renal CYP1A catalytic activity was increased to Ϯ mean SE of 9 animals. Significance was determined with a 2-way ANOVA 0.53 Ϯ 0.10 nmol/mg/min and visualized by Western immu- coupled with the Students-Newman-Keuls post hoc test in which p Ͻ 0.05 was required for a statistically significant difference. noblotting following administration of the potent inducer *Significantly increased from vehicle control, p Ͻ 0.05. ␤-naphthoflavone. Another enzyme examined with relevance **Significantly decreased from vehicle control, p Ͻ 0.01. to estradiol metabolism was COMT. The results demonstrated EGCG MODULATES CYP450 ISOFORMS 265

TABLE 2 The Effect of Catechin Treatment on the Weight of Various Organs

Treatment (dose) Liver Kidney Spleen Uterus Ovary

Saline (8 ml/kg) 5.2 Ϯ 0.1 1.3 Ϯ 0.03 0.41 Ϯ 0.02 0.36 Ϯ 0.03 0.041 Ϯ 0.004 EGCG (12.5 mg/kg) 5.2 Ϯ 0.1 1.4 Ϯ 0.03 0.45 Ϯ 0.02 0.45 Ϯ 0.03 0.047 Ϯ 0.004 EGCG (25 mg/kg) 5.5 Ϯ 0.2 1.4 Ϯ 0.04 0.50 Ϯ 0.03 0.40 Ϯ 0.03 0.046 Ϯ 0.003 EGCG (50 mg/kg) 6.2 Ϯ 0.3* 1.6 Ϯ 0.05* 0.75 Ϯ 0.10* 0.38 Ϯ 0.04 0.048 Ϯ 0.005 ECG (12.5 mg/kg) 4.8 Ϯ 0.1 1.3 Ϯ 0.04 0.36 Ϯ 0.01 0.40 Ϯ 0.04 0.045 Ϯ 0.003 ECG (25 mg/kg) 5.3 Ϯ 0.1 1.3 Ϯ 0.03 0.38 Ϯ 0.02 0.39 Ϯ 0.09 0.035 Ϯ 0.003

Note. Mice were treated as indicated for 7 days. Each value is the indicated organ weight as a percentage of body weight and represents the mean Ϯ SE from Downloaded from https://academic.oup.com/toxsci/article/76/2/262/1686012 by guest on 28 September 2021 9 animals. Significance was determined with a 2-way ANOVA coupled with the Students-Newman-Keuls post hoc test in which p Ͻ 0.05 was required for a statistically significant difference. *Significantly increased from vehicle control, p Ͻ 0.01. that the catalytic activity of both hepatic and renal COMT tumor inhibitory properties of EGCG. However, administration remained unchanged following 7 days of catechin treatment, as of EGCG (50 mg/kg, ip) resulted in extensive hepatic necrosis, the values ranged between 16 Ϯ 1 and 19 Ϯ 3 nmol/mg/min. significant weight loss, and mortality in 67% of the mice. Other Lastly, CYP2E1 was also profiled following catechin treat- groups have reported a decrease in bw following the adminis- ment. While ECG had no effect on CYP2E1, EGCG (25 tration of (-)-epicatechin (EC) (30–35 mg/rat) (Lhoste et al., mg/kg/d, 7d) increased CYP2E1 catalytic activity by 37 Ϯ 3% 2003) or 5% extract in male rats (Satoh et al., 2002). (Fig. 7) and polypeptide levels by 22 Ϯ 3% (Fig. 8). Following 2 weeks of EC, the rats lost 14% of their body weight (Lhoste et al., 2003), while administration of green-tea DISCUSSION extract for 2–8 weeks caused a 10–15% loss in bw (Satoh et al., 2002). Only one other study has examined the more potent The ability of EGCG to modulate the enzymes responsible for the synthesis and metabolism of estradiol was initially intended to be conducted at doses relevant to the reported

FIG. 3. Hepatic CYP3A catalytic activity following treatment with EGCG FIG. 2. Ovarian aromatase activity following treatment with EGCG and and ECG: Mice were dosed with EGCG; ECG, 25 or 12.5 mg/kg, ip; or saline ECG: Mice were dosed with EGCG; ECG, 25 or 12.5 mg/kg, ip; or saline for for 7 days. Dexamethasone (Dex, 75 mg/kg, ip) was administered for 4 days. 7 days. Necropsies were performed 24 h after the final dose. The bars represent Necropsies were performed 24 h after the final dose. The bars represent the the mean Ϯ standard error (SE) of three independent determinations performed mean Ϯ SE from 9 animals. Significance was analyzed with a 2-way ANOVA in triplicate. Significance was analyzed with a 2-way ANOVA coupled with coupled with the Student-Newman-Keuls post hoc test in which p Ͻ 0.05 was the Student-Newman-Keuls post hoc test in which p Ͻ 0.05 was required for required for a statistical difference; *significantly increased from vehicle a statistical difference; *significantly decreased from vehicle treatment at a treatment at a level of p Ͻ 0.05; **significantly increased from all other level of p Ͻ 0.05. treatments at a level of p Ͻ 0.001. 266 GOODIN AND ROSENGREN

EGCG induced hepatotoxicity and mortality in adult female Swiss-Webster mice. Therefore, a species- and strain-specific response occurs following high doses of EGCG since hepato- toxicity was observed in female Swiss-Webster and C57Bl/6 mice, but not in rats or athymic mice. However, in the athymic mouse study, markers of hepatotoxicity were not included as part of the findings. Therefore, it is recommended that further investigations with this compound, especially in mice, include indicators of hepatotoxicity. This will help elucidate the exact species- and strain-specific response of EGCG.

Due to the hepatotoxicity and mortality that occurred, enzy- Downloaded from https://academic.oup.com/toxsci/article/76/2/262/1686012 by guest on 28 September 2021 matic activities were only performed following the adminis- tration of lower doses of the catechins. Aromatase was initially examined, as it is responsible for the synthesis of estradiol from , and a catechin-specific inhibition of aro- matase in vitro has previously been reported. Specifically,

Satoh et al. (2002), reported IC50 values of 60 ␮M and 100 ␮M for EGCG and EGC, respectively, while 100 ␮M of ECG only produced 20% inhibition. However, other groups have reported that EGC (Kao et al., 1998) and (-)-catechin fail to inhibit aromatase (Campbell and Kurze, 1993). Our in vivo results support a catechin-specific inhibition of aromatase, since EGCG (12.5 or 25 mg/kg) inhibited ovarian aromatase activity FIG. 4. Hepatic CYP3A immunoreactive polypeptide levels following by 56%, while ECG did not alter aromatase activity. A de- treatment with EGCG and ECG: Mice were dosed with EGCG, ECG (25 or 12.5 mg/kg, ip), or vehicle for 7 days. Dexamethasone (Dex, 75 mg/kg, ip) was crease in aromatase in vivo would reduce the production of administered for 4 days. Necropsies were performed 24 h after the final dose. estradiol and thus decrease estradiol-mediated events. There- (A) Representative Western immunoblot for CYP3A. The positions of the molecular weight markers are indicated on the right (kDa). The lanes are, from left to right, (1) dexamethasone; (2) saline; (3) EGCG, 12.5 mg/kg; (4 and 5) EGCG, 25 mg/kg; (6) ECG, 12.5 mg/kg; (7 and 8) ECG, 25 mg/kg. (B) Results from scanning densitometry of Western immunoblots: the bars represent the mean Ϯ SE of the optical density from 9 animals. Significance was analyzed with a 2-way ANOVA coupled with the Student-Newman-Keuls post hoc test in which p Ͻ 0.05 was required for a statistical difference; *significantly increased from vehicle treatment at a level of p Ͻ 0.05; **significantly increased from all other treatments at a level of p Ͻ 0.01. catechins. Specifically, 85 mg/kg/day of EGCG (but not EC, EGC or ECG) for 7 days caused a 15–21% loss of initial body weight in male Sprague-Dawley rats, while 92 mg/kg/d for 7 days caused a 10% loss of body weight in female rats (Kao et al., 2000). This study supports our results, which demonstrate that only EGCG causes a loss of body weight in female C57Bl/6 (Goodin et al., 2002) and Swiss-Webster mice. While the studies in rats provide conflicting results regarding cate- chin-mediated weight loss, none of them have reported hepa- totoxicity. Therefore, we are the only group to have previously reported mild hepatotoxicity following EGCG treatment FIG. 5. Hepatic CYP1A catalytic activity following treatment with EGCG (Goodin et al., 2002). Since Liao et al. (1995) administered and ECG: Mice were dosed with EGCG, ECG (25 or 12.5 mg/kg, ip), or EGCG (50 mg/kg, ip) to athymic nude female mice for 14 to 28 vehicle for 7 days. ␤-Naphthoflavone (␤-NF, 80 mg/kg, ip) was administered days and did not report mortality or hepatotoxicity, we had for 3 days. Necropsies were performed 24 h after the final dose. The bars Ϯ previously postulated that the mild hepatotoxicity elicited by represent the mean SE from 9 animals. Significance was analyzed with a 2-way ANOVA coupled with the Student-Newman-Keuls post hoc test in EGCG (50 mg/kg/d, 3d) in immature C57Bl/6 female mice was which p Ͻ 0.05 was required for a statistical difference; *significantly de- an age-specific response (Goodin et al., 2002). However, this creased from vehicle treatment at a level of p Ͻ 0.05; *significantly increased clearly is not the case, as four or more days of 50 mg/kg of from all other treatments at a level of p Ͻ 0.001. EGCG MODULATES CYP450 ISOFORMS 267

but detectable levels following administration of the classic inducer, dexamethasone (Zerilli et al., 1998). While there are no other in vivo studies with EGCG or ECG, two studies have measured CYP450 activity in male rats following treatment with either (ϩ)-catechin or EC. Specifically, Siegers et al. (1982) demonstrated that (ϩ)-catechin (200 mg/kg/day, 7 or 28 days, po) did not alter total hepatic CYP450 or CYP2E1 activity in Sprague-Dawley rats. However, Lhoste et al. (2003) reported that both (ϩ)-catechin and EC (30–35 mg/rat/day, 14 days, po) decreased CYP2E1 (37 and 25%, respectively) in

male F344 rats. The discrepancy in the response following Downloaded from https://academic.oup.com/toxsci/article/76/2/262/1686012 by guest on 28 September 2021 (ϩ)-catechin in rats may be due to the fact that Lhoste et al. (2003) administered the catechins in 10% in order to simulate catechin exposure from drinking wine. However, it appears as though the effect of catechins on CYP2E1 activity is both catechin- and species-specific, as we demonstrated a modest increase in hepatic CYP2E1 following EGCG. It is worth noting that the majority of the in vivo data is not supported by in vitro inhibition studies. Specifically, studies, using bacterial cells transfected with plasmids for specific human CYP450 isoforms, demonstrated that both ECG and EGCG acted as either mixed or noncompetitive inhibitors of CYP1A1, CYP1A2, CYP2A6, CYP2C9, CYP2E1, and FIG. 6. Hepatic CYP1A immunoreactive polypeptide levels following CYP3A4 (Muto et al., 2001). ECG exhibited the greatest treatment with EGCG and ECG: Mice were dosed with EGCG, ECG (25 or potency against CYP1A2 as 10 ␮M caused 50% inhibition. 12.5 mg/kg, ip), or vehicle for 7 days. ␤-Naphthoflavone (␤-NF, 80 mg/kg, ip) was administered for 3 days. Necropsies were performed 24 h after the final However, the effect of ECG on CYP1A in vivo is exhibited via dose. (A) Representative Western immunoblot for CYP1A: the positions of molecular weight markers are indicated on the right (kDa). The lanes from left to right: (1) ␤-naphthoflavone; (2) saline; (3) EGCG, 12.5 mg/kg; (4 and 5) EGCG, 25 mg/kg; (6) ECG, 12.5 mg/kg; (7 and 8) ECG, 25 mg/kg. (B) Results from scanning densitometry of Western immunoblots: the bars represent the mean Ϯ SE of the optical density from 9 animals. Significance was analyzed with a 2-way ANOVA coupled with the Student-Newman-Keuls post hoc test in which p Ͻ 0.05 was required for a statistical difference; *significantly decreased from vehicle treatment at a level of p Ͻ 0.05; **significantly increased from all other treatments at a level of p Ͻ 0.001. fore, aromatase inhibition may be partly responsible for the antitumor actions of EGCG in rodent models of mammary cancer. However, other mechanisms are likely to be involved in the antitumor effect produced by EGCG, as other aromatase inhibitors have been reported to cause a decrease in tumor growth but not tumor regression (Yue and Brodie, 1993). Further studies examining the effect of EGCG on breast tumor aromatase activity in a murine model will be conducted to confirm this hypothesis. The effect of EGCG on breast tumor growth could also be due to changes in estradiol metabolism via the modulation of FIG. 7. Hepatic CYP2E1 catalytic activity following treatment with either CYP1A, CYP3A, or COMT. Our results demonstrated EGCG and ECG: Mice were dosed with EGCG, ECG (25 or 12.5 mg/kg, ip) that EGCG increased the catalytic activity of hepatic, but not or vehicle for 7 days. Necropsies were performed 24 h after the final dose with renal CYP3A, and ECG decreased the catalytic activity of the exception of acetone (Ace, 4.8 mg/kg, po), which was administered 16 h Ϯ hepatic CYP1A, while renal CYP1A remained below detect- prior to necropsy. The bars represent the mean SE from 9 animals. Signif- icance was analyzed with a 2-way ANOVA coupled with the Student-New- able levels. The renal enzymatic results are not unexpected: we man-Keuls post hoc test in which p Ͻ 0.05 was required for a statistical have previously reported that renal CYP1A is undetectable difference; *significantly increased from vehicle treatment at a level of p Ͻ (Bray et al., 2001), while rodent renal CYP3A remains at low 0.05; **significantly increased from all other treatments at a level of p Ͻ 0.01. 268 GOODIN AND ROSENGREN

has been demonstrated with the in vivo administration of other gallate-containing compounds (Mannisto and Kaakkola, 1999). Therefore, the rapid conversion to 2-methoxyestradiol should not be altered by EGCG, thus allowing the induction of CYP3A to shift the metabolism toward the production of 2-methoxyestradiol. Since nM and ␮M concentrations of 2-methoxyestradiol disrupt microtubule formation (D’Amato et al., 1994), inhibit and induce (La- khani et al., 2003), small increases in 2-methoxyestradiol may have significant inhibitory effects on breast tumor growth.

Overall, the modest increase in CYP3A activity, paired with Downloaded from https://academic.oup.com/toxsci/article/76/2/262/1686012 by guest on 28 September 2021 the 56% inhibition of aromatase, may work together to both decrease estradiol production and shift the metabolism of es- tradiol toward the production of antiestrogenic and antiangio- genic metabolites. However, CYP3A is also responsible for the formation of 15–20% of 4-hydroxyestradiol, a metabolite with estrogenic properties (Weisz et al., 1992). Therefore, to deter- mine the exact effect on estradiol metabolism, future studies will measure the production of estradiol and its metabolites following EGCG treatment in tumor-bearing mice. Addition- ally, future studies will also determine the specific isoform of CYP3A (i.e., 3A11, 3A41, and/or 3A44) that is increased by EGCG. In addition to examining the enzymes responsible for the synthesis and metabolism of estradiol, the modulatory effect on FIG. 8. CYP2E1 immunoreactive polypeptide levels following treatment CYP2E1 by EGCG and ECG was also determined, as this with EGCG and ECG: Mice were dosed with EGCG, ECG (25 or 12.5 mg/kg, enzyme plays an important role in of numerous ip), or vehicle for 7 days. Necropsies were performed 24 h after the final dose carcinogens (Constan et al., 1999; Sohn et al., 2001) and is also with the exception of acetone (Ace, 4.8 mg/kg, po), which was administered expressed in breast tumors (Iscan et al., 2001). Therefore, this 16 h prior to necropsy. (A) Representative Western immunoblot for CYP2E1:. the positions of molecular weight markers are indicated on the right (kDa). The examination would demonstrate if catechin administration had lanes from left to right: (1) acetone; (2) saline; (3) EGCG, 12.5 mg/kg; (4 and the potential to produce drug/chemical interactions. While 5) EGCG, 25 mg/kg); (6) ECG, 12.5 mg/kg; (7 and 8) ECG, 25 mg/kg. (B) ECG failed to alter CYP2E1, EGCG (25 mg/kg/day, 7days) Results from scanning densitometry of Western immunoblots: the bars repre- increased CYP2E1 catalytic activity by 37% and polypeptide Ϯ sent the mean SE of the optical density from 9 animals. Significance was levels by 22%. Unfortunately this increase in CYP2E1 oc- analyzed with a 2-way ANOVA coupled with the Student-Newman-Keuls post hoc test in which p Ͻ 0.05 was required for a statistical difference; *signifi- curred following the same dose of EGCG that modulated cantly increased from vehicle treatment at a level of p Ͻ 0.05; **significantly aromatase and CYP3A. This small increase in CYP2E1 is increased from all other treatments at a level of p Ͻ 0.01. unlikely to significantly increase the activation of carcinogens. However, the combined increase in CYP2E1 and CYP3A may a decrease in protein and not through competitive enzyme cause drug-interactions with substances, such as acetamino- inhibition, as CYP1A polypeptide levels and catalytic activity phen, which utilize both of these enzymes in their metabolism were both decreased to the same extent. Since EGCG, but not (Kostrubsky et al., 1997; Raucy et al., 1989). This effect may ECG induces weight loss and hepatotoxicity in mice, it is not be limited to the mouse, as, in the male rat, CYP2E1 is either surprising that their effects on CYP450 isoforms in vivo are unchanged (Siegers et al., 1982) or decreased by (ϩ)-catechin distinct both from each other and in in vitro experiments. and EC (Lhoste et al., 2003). However, further work needs to This catechin-specific modulation of CYP450 isoforms may be performed to determine the extent of this potential species- in part explain why EGCG, but not ECG, was able to decrease specific response. tumor growth in mice (Liao et al., 1995). Specifically, EGCG In summary, this is the first study to examine both hepatic (25 mg/kg) was a weak inducer of hepatic CYP3A activity and and renal enzymes responsible for the synthesis and metabo- polypeptide levels. This is important, as CYP3A and CYP1A lism of estradiol following in vivo administration of specific are responsible for the formation of approximately 85% of catechins. The resulting catechin-specific alteration in ovarian 2-hydroxyestradiol, which is then rapidly converted to 2-me- aromatase and hepatic CYP3A by EGCG may prove to be an thoxyestradiol by COMT (Kerlan et al., 1992; Shou et al., important component of the mechanism of EGCG-mediated 1997). Importantly, EGCG did not inhibit COMT activity as breast tumor suppression. However, this result is tempered by EGCG MODULATES CYP450 ISOFORMS 269 the finding that high doses of EGCG produce significant hep- Kavanagh, K. T., Hafer, L. J., Kim, D. W., Man, K. K., Sherr, D. H., Rogers, atotoxicity and mortality in female mice. A. E., and Sonenshein, G. E. (2001). Green tea extracts decrease - induced mammary tumor burden in rats and rate of breast cancer cell proliferation in culture. J. Cell. Biochem. 82, 387–398. ACKNOWLEDGMENTS Kellis, J. J., and Vickery, L. (1984). Inhibition of human estrogen synthetase (aromatase) by flavones. Science 225, 1032–1034. The authors would like to thank M. J. Le Nedelec and B. J. Bray for their technical assistance. The work was funded by a Deans’ Bequest Grant from the Kerlan, V., Dreano, Y., Bercovici, J. P., Beaune, P. H., Floch, H. H., and University of Otago. Berthou, F. (1992). Nature of P450 involved in the -2/4- of estradiol in human liver microsomes. Biochem. Pharma- col. 44, 1745–1756. REFERENCES Klauber, N., Parangi, S., Flynn, E., Hamel, E., and D’Amato, R. J. (1997). Inhibition of angiogenesis and breast cancer in mice by the microtubule Downloaded from https://academic.oup.com/toxsci/article/76/2/262/1686012 by guest on 28 September 2021 Arts, I. C. W., van de Putte, B., and Hollman, P. C. H. (2000). Catechin inhibitors 2-methoxyestradiol and taxol. Cancer Res. 57, 81–86. contents of foods commonly consumed in The Netherlands: 1. Fruits, Komori, A., Yatsunami, J., Okabe, S., Abe, S., Hara, K., Suganuma, M., Kim, vegetables, staple foods, and processed foods. J. Agric. Food Chem. 48, S. J., and Fujiki, H. (1993). Anticarcinogenic activity of green tea polyphe- 1746–1751. nols. Japan. J. Clin. Oncol. 23, 186–190. Bray, B. J., Goodin, M. G., Inder, R. E., and Rosengren, R. J. (2001). The Kondo, T., Ohta, T., Igura, K., Hara, Y., and Kaji, K. (2002). Tea catechins effect of retinol on hepatic and renal drug-metabolising enzymes. Food inhibit angiogenesis in vitro, measured by human endothelial cell growth, Chem. Toxicol. 39, 1–9. migration, and tube formation through inhibition of VEGF receptor binding. Bray, B. J., Perry, N. B., Menkes, D. B., and Rosengren, R. J. (2002). St. Cancer Lett. 180, 139–144. John’s wort extract induces CYP3A and CYP2E1 in the Swiss-Webster Kostrubsky, V., Lewis, L., Wood, S., Sinclair, P., Wrighton, S., and Sinclair, mouse. Toxicol. Sci. 66, 27–33. J. (1997). Effect of taxol on P450 3A and acetaminophen Campbell, D., and Kurze, M. (1993). inhibition of aromatase en- toxicity in cultured rat hepatocytes: Comparison to dexamethasone. Toxicol. zyme activity in human preadipocytes. J. . Biochem. Mol. Biol. 46, Appl. Pharmacol. 142, 79–86. 381–388. Lakhani, N. J., Sarkar, M. A., Venitz, J., and Figg, W. D. (2003). 2-Me- Constan, A., Sprankle, C., Peters, J., Kedderis, G., Everitt, J., Wong, B., thoxyestradiol, a promising anticancer agent. Pharacotherapy 23, 165–172. Gonzalez, F., and Butterworth, B. (1999). Metabolism of chloroform by Lhoste, E. F., Ouriet, V., Bruel, S., Flinois, J.-P., Brezillon, C., Magdalou, J., 2E1 is required for induction of toxicity in the liver, Cheze, C., and Nugon-Baudon, L. (2003). The human colonic microflora kidney, and nose of male mice. Toxicol. Appl. Pharmacol. 160, 120–126. influences the alterations of xenobiotic-metabolizing enzymes by catechins D’Amato, R. J., Lin, C. M., Flynn, E., Folkman, J., and Hamel, E. (1994). in male F344 rats. Food Chem. Toxicol. 41, 695–702. 2-Methoxyestradiol, an endogenous mammalian metabolite, inhibits tubulin Liao, S., Umekita, Y., Guo, J., Kokontis, J. M., and Hiipakka, R. A. (1995). polymerization by interacting at the cochicine site. Proc. Natl. Acad. Sci. U Growth inhibition and regression of human prostate and breast tumors in SA91, 3964–3968. athymic mice by tea epigallocatechin gallate. Cancer Lett. 96, 239–243. Fotsis, T., Zhang, Y., Pepper, M. S., Adlercreutz, H., Montesano, R., Nawroth, Lu, Q., Nakmura, J., Savinov, A., Yue, W., Weisz, J., Dabbs, D., Wolz, G., and P. P., and Schweigerer, L. (1994). The endogenous oestrogen metabolite Brodie, A. (1996). Expression of aromatase protein and messenger ribonu- 2-methoxyoestradiol inhibits angiogenesis and suppresses tumour growth. cleic acid in tumor epithelial cells and evidence of functional significance of Nature 368, 237–239. locally produced estrogen in human breast . Endocrinology 137, Goodin, M. G., Fertuck, K. C., Zacharewski, T. R., and Rosengren, R. J. 3061–3068. (2002). -mediated actions of polyphenolic catechins in Mannisto, P., and Kaakkola, S. (1999). Catechol-O-methyltransferase vivo and in vitro. Toxicol. Sci. 69, 354–361. (COMT): Biochemistry, molecular biology, pharmacology, and clinical Guengerich, F. P. (1989). Characterization of human microsomal cytochrome efficacy of the new selective COMT inhibitors. Pharmacol. Rev. 51, 593– P-450 enzymes. Ann. Rev. Pharmacol. Toxicol. 29, 241–264. 628. Guo, Q., Zhao, B., Shen, S., Hou, J., Hu, J., and Xin, W. (1999). ESR study on Morre, D. J., Bridge, A., Wu, L.-W., and Morre, D. M. (2000). Preferential the structure- activity relationship of tea catechins and their inhibition by (-)-epigallocatechin-3-gallate of the cell surface NADH oxi- epimers. Biochim. Biophys. Acta 1427, 13–23. dase and growth-transformed cells in culture. Biochem. Pharmacol. 60, 937–946. Herblin, W. F. (1973). A simple colorimetric assay for catechol-O-methyl . Anal. Biochem. 51, 19–22. Muto, S., Fujita, K., Yamazaki, Y., and Kamataki, T. (2001). Inhibition by green tea catechins of metabolic activation of procarcinogens by human Inder, R. E., Bray, B. J., Sipes, I. G., and Rosengren, R. J. (1999). Role of cytochrome P450. Mutat. Res. 479, 197–206. cytochrome P4502E1 in retinol’s attenuation of carbon tetrachloride hepa- totoxicity in the mouse. Toxicol. Sci. 52, 130–139. Nakachi, K., Suemasu, K., Suga, K., Takeo, T., Imai, K., and Higashi, Y. (1998). Influence of drinking tea on breast cancer malignancy among Jap- Iscan, M., Klaavuniemi, T., Coban, T., Kapucuoglu, N., Pelkonen, O., anese patients. Jap. J. Cancer Res. 89, 254–261. and Raunio, H. (2001). The expression of cytochrome P450 enzymes in human breast tumours and normal breast tissue. Breast Can. Res. Treat. 70, Raucy, J., Lasker, J., Lieber, C., and Black, M. (1989). Acetaminophen 47–54. activation by human liver cytochromes P450IIE1 and P450IA2. Arch. Bio- chem. Biophys. 271, 270–283. Kao, Y.-C., Zhou, C., Sherman, M., Laughton, C., and Chen, S. (1998). Molecular basis of the inhibition of human aromatase (estrogen synthetase) Ryan, D., and Levin, W. (1990). Purification and characterization of hepatic by flavone and isoflavone : A site-directed mutagenesis microsomal cytochrome P-450. Pharmacol. Ther. 45, 153–239. study. Environ. Health Perspect. 106, 85–92. Ryu, S.-D., and Chung, W.-G. (2003). Induction of the procarcinogen-activat- Kao, Y.-H., Hiipakka, R. A., and Liao, S. (2000). Modulation of endocrine ing CYP1A2 by an herbal in rats and humans. Food systems and food intake by green tea epigallocatechin gallate. Endocrinol- Chem. Toxicol. 41, 861–866. ogy 141, 980–987. Satoh, K., Sakamoto, Y., Ogata, A., Nagai, F., Mikuriya, H., Numazawa, M., 270 GOODIN AND ROSENGREN

Yamada, K., and Aoki, N. (2002). Inhibition of aromatase activity by green Wymer, J., and Guillen, J. M. (1996). Inhibitory effect of six green-tea tea extract catechins and their endocrinological effects of oral administration catechins and on the growth of four selected human tumor cell lines. in rats. Food Chem. Toxicol. 40, 925–933. Anticancer Drugs 7, 461–468. Shou, M., Korzekwa, K., Brooks, E., Krausz, K., Gonzalez, F., and Gelboin, H. Vinggaard, A., Hnida, C., Breinholt, V., and Larsen, J. (2000). Screening of (1997). Role of human hepatic cytochrome P450 1A2 and 3A4 in the selected for inhibition of CYP19 aromatase activity in vitro. metabolic activation of . Carcinogenesis 18, 207–214. Toxicol. In vitro 14, 227–234. Siegers, C. P., Larseille, J., and Younes, M. (1982). Effects of dithiocarb and Wang, Z. Y., Das, M., Bickers, D. R., and Mukhtar, H. (1988). Interaction of (ϩ)-catechin on microsomal enzyme activities of rat liver. Res. Comm. epicatechins derived from green tea with rat hepatic cytochrome P-450. Chem. Pathol. Pharmacol. 36, 61–73. Drug Met. Disp. 16, 98–103. Weisz, J., Bui, Q., Roy, D., and Liehr, J. (1992). Elevated 4-hydroxylation of Smith, P. K., Krohn, R. I., Hermanson, G. T., Mallia, A. K., Gartner, F. H., estradiol by hamster kidney microsomes: A potential pathway of metabolic Provenzano, M. D., Fujimoto, E. K., Goeke, N. M., Olson, B. J., and Klenk, activation of estrogens. Endocrinology 131, 655–661. D. C. (1985). Measurement of protein using bicinchoninic acid. Anal. Downloaded from https://academic.oup.com/toxsci/article/76/2/262/1686012 by guest on 28 September 2021 Biochem. 150, 76–85. Wrighton, S., Schuetz, E., Watkins, P., Maurel, P., Barwick, J., Bailey, B., Hartle, H., Young, B., and Guzelian, P. (1985). Demonstration in multiple Sohn, O., Fiala, E., Requeijo, S., Weisburger, J., and Gonzalez, F. (2001). species of inducible hepatic cytochromes P450 and their mRNAs related to Differential effects of CYP2E1 status on the metabolic activation of the the glucocorticoid-inducible cytochrome P450 of the rat. Mol. Pharmacol. colon carcinogen azoxymethane and methylazoxymethanol. Cancer Res. 61, 28, 312–321. 8435–8440. Yue, W., and Brodie, A. (1993). MCF-7 human breast carcinomas in nude Tanaka, H., Hirose, M., Kawabe, M., Sano, M., Takesada, Y., Hagiwara, A., mice as a model for evaluating aromatase inhibitors. J. Steroid. Biochem. and Shirai, T. (1997). Postinitiation inhibitory effects of green-tea catechins Mol. Biol. 44, 671–673. on 7,12-dimethylbenz[a]anthracene-induced mammary gland carcinogene- Zerilli, A., Lucas, D., Dreano, Y., Picart, D., and Berthou, F. (1998). Effect of sis in female Sprague-Dawley rats. Cancer Lett. 116, 47–52. pyrazole and dexamethasone administration on cytochrome P450 2E1 and Toda, K., Takeda, K., Okada, T., Akira, S., Saibara, T., Kaname, T., 3A isoforms in rat liver and kidney: Lack of specificity of p-nitrophenol as Yamamura, K., Onishi, S., and Shizuta, Y. (2001). Targeted disruption of a of P450 2E1. . Clin. Exper. Res. 22, 652–657. the aromatase P450 gene (Cyp19) in mice and their ovarian and uterine Zhu, B. T., and Liehr, J. G. (1994). increases the severity of ␤ responses to 17 -oestradiol. J. Endocrinol. 170, 99–111. estradiol-induced tumorigenesis in hamster kidney. Toxicol. Appl. Pharma- Valcic, S., Timmerman, B. N., Aberts, D. S., Wachter, G. A., Krutzsch, M., col. 125, 149–158.