- Email: [email protected]
No evidence for the in vivo activity of aromatase-inhibiting flavonoids
Journal of Steroid Biochemistry & Molecular Biology 78 (2001) 231– 239 www.elsevier.com/locate/jsbmb
No evidence for the in vivo activity of aromatase-inhibiting flavonoids Niina Saarinen a,*, Suresh C. Joshi a, Markku Ahotupa a, Xiangdong Li a, Jenni A8 mma¨la¨ a, Sari Ma¨kela¨ a,b, Risto Santti a b
a Institute of Biomedicine, Uni6ersity of Turku, Kiinamyllynkatu 10, FIN-20520 Turku, Finland Unit for Pre6enti6e Nutrition, Department of Medical Nutrition, Karolinska Institute, S-14157 Huddinge, Sweden
Received 4 December 2000; accepted 11 April 2001
Abstract Measurements of the aromatase-inhibiting and antioxidative capacities of flavonoids in vitro showed that slight changes in flavonoid structure may result in marked changes in biological activity. Several flavonoids such as 7-hydroxyflavone and chrysin (5,7-dihydroxyflavone) were shown to inhibit the formation of 3H-17b-estradiol from 3H-androstenedione (IC50 B 1.0 mM) in human choriocarcinoma JEG-3 cells and in human embryonic kidney cells HEK 293 transfected with human aromatase gene (Arom+ HEK 293). Flavone and quercetin (3,3%,4%,5,7-pentahydroxyflavone) showed no inhibition (IC50 \ 100 mM). None of the requirements for optimal antioxidative capacity (2,3-double bond with 4%-hydroxy group, 3-hydroxyl group, 5,7-dihydroxy structure and the orthodihydroxy structure in the B-ring) is relevant for the maximum inhibition of aromatase by flavonoids. After oral administration to immature rats at a dose of 50 mg/kg body weight, which considerably exceeds amounts found in daily human diets, neither aromatase-inhibiting nonestrogenic flavonoids, such as chrysin, nor estrogenic flavonoids, such as naringenin and apigenin, induced uterine growth or reduced estrogen- or androgen-induced uterine growth. The inability of flavonoids to inhibit aromatase and, consequently, uterine growth in short-term tests may be due to their relatively poor absorption and/or bioavailability. © 2001 Elsevier Science Ltd. All rights reserved. Keywords: Aromatase; Flavonoid; Estrogen; Diet
1. Introduction The risk of human breast cancer (BC) is related to cumulative exposure of breast tissue to endogenous estrogens [1–4]. Experimental evidence also strongly favors the role of estrogens in the development and growth of BC. Strategies aiming at reduced estrogen production may be useful for the prevention of estrogen-related BC [5]. The incidence rates of BC are higher in populations consuming high-fat, low-fiber diets [2,6] than in populations with diets rich in fruit and vegetables [6,7]. In addition to the fiber effect [8], studies suggest that the possible beneficial effects of low-fat and high-fiber diets
* Corresponding author. Tel.: + 358-2-333-7223; fax: +358-2-3337352. E-mail address: [email protected] (N. Saarinen).
are related to weakly estrogenic compounds known as phytoestrogens [9–11]. Isoflavonoids are the most extensively studied phytoestrogens. The intake and serum concentrations of isoflavonoids, such as daidzein and genistein, mainly derived from soy, are higher in Asian countries, where the incidence of BC is low [9]. Several edible plants present in Asian as well as Western diets contain a wide variety of flavonoids with biologic actions similar to those of isoflavonoids [12]. They may also contribute to the chemopreventive actions of vegetable- and fruit-rich diets. The mechanism of the possible chemopreventive action of flavonoids is not fully known. Besides binding to the two known estrogen receptor subtypes (ERa and ERb) [13], flavonoids may compete with endogenous substrates for active sites of estrogen biosynthesizing and metabolizing enzymes, such as aromatase [14 –20] and 17b-hydroxysteroid oxidoreductase (HSOR), type 1
0960-0760/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved. PII: S 0 9 6 0 - 0 7 6 0 ( 0 1 ) 0 0 0 9 8 - X
N. Saarinen et al. / Journal of Steroid Biochemistry & Molecular Biology 78 (2001) 231–239
232
[21,22]. The inhibition of these enzymes lowers 17bestradiol concentrations in target cells, which then decrease the risk of BC. In addition, some flavonoids act as antioxidants, which may also play a role in the prevention of carcinogenesis in general [23]. To understand the significance and the complex role of individual flavonoids or their mixtures from various dietary sources in BC chemoprevention, it is important to recognize the critical structural properties determining the mechanisms of flavonoid actions. One of our aims was to identify the function of flavonoids as well as to understand the mode of their possible chemopreventive action. Earlier, we studied the inhibitory actions of several flavonoids on the 17b-HSORs type 1 [21] and 2 [22]. Here, we report on the effects of the same flavonoids on the formation of 3H-17b-estradiol from 3 H-andostenedione in JEG-3 cells, a human chorioncarcinoma cell line [24], and in Arom+HEK 293 cells transfected with the construct for the synthesis of aromatase enzyme. In addition to the ability of flavonoids to inhibit estrogen biosynthesis, we estimated their antioxidant properties and compared the structural requirements for selective inhibition of estrogen biosynthesis (aromatization and 17b-reduction) with those for antioxidative capacity and estrogenicity. The most potent aromatase inhibiting flavonoids were chosen for in vivo studies of estrogenicity and antiestrogenicity.
2. Materials and methods
2.1. Chemicals The test compounds were purchased from the following sources: apigenin (4%,5,7-trihydroxyflavone), chrysin (5,7-dihydroxyflavone), kaempferol (3,4%,5,7-tetrahydroxyflavone), quercetin (3,3%,4%,5,7-pentahydroxyflav-
Fig. 1. General structures of flavonoids.
one), 4-androstene-3,17-dione and diethylstilbestrol (DES) were obtained from Sigma Chemical Co. (St. Louis, MO, USA), 7-hydroxyflavone, fisetin (3,3%,4%,7tetrahydroxyflavone), kaempferide (3,5,7-trihydroxy-4%methoxyflavone), luteolin (3%,4%,5,7-tetrahydroxyflavone), galangin (3,5,7-trihydroxyflavone), naringenin (4%,5,7-trihydroxyflavanone) and pinostrobin (5-hydroxy-7-methoxyflavone) from Carl Roth GmbH (Karlsruhe, Germany). Fig. 1 shows the main chemical structures of the flavonoids. The cell culture materials were obtained from the following sources: Dulbecco’s modified Eagle medium (DMEM), fetal calf serum (FCS) and Trypsin–EDTA from Gibco BRL (Paisley, Scotland). The five-well plates and culture dishes were from NUNC (Roskilde, Denmark). Versene–EDTA in phosphate buffered saline contained NaCl (136 mM), KCl (2.68 mM), Na2HPO4 (8.1 mM), KH2PO4 (1.47 mM) and EDTA disodium salt (0.054 mM). 3H-androst-4-ene,3,17-dione and 3H-estrone were from NEN Life Science Products (Zaventem, Belgium). Unlabeled androstenedione, testosterone, 17b-estradiol, and estrone were obtained from Sigma Chemical Co. (St. Louis, MO, USA). Dichloromethane was purchased from Rathburn (Peebleshire, UK). MPV-2213ad aromatase inhibitor [25] was obtained from Hormos Medical Ltd (Turku, Finland).
2.2. In 6itro aromatase inhibition assays The aromatase activity in JEG-3, human choriocarcinoma cells, and in HEK 293 cells stably expressing the aromatase (Arom+HEK 293) was assayed by determining the ability of the cells to convert added 3 H-androstenedione to 3H-17b-estradiol and 3H-estrone. The procedure for the construction of the aromatase expression plasmid (pUbC-AROM) has been previously described [26]. For constructing the stable cell lines, pUbC-AROM-plasmid was transfected to HEK-293 cells grown in DMEM medium supplemented with 10% FCS using the DOTAP liposomal Transfection Kit (Boehringer Mannheim, Mannheim, Germany) according to the manufacturer’s instructions. After transfection, the cells were cultured for 48 h and transferred to the selection medium (DMEM, 10% FCS, 900 mgG418/ml medium). Cloned cells were applied in 96well plates at a density of 0.5 cells/well and grown for 2 months in the presence of the selection medium. The cloned cells lines were then screened for aromatase expression by RT-PCR and enzyme activity as previously described [26]. Cells were maintained in DMEM containing 10% of charcoal stripped heat inactivated FCS. The experiments with JEG-3 cells were done using cells from confluent monolayers (approximately 5 days after subculture) in 30×10 mm culture dishes. Arom+HEK 293 cells were cultured for 2–3 days prior to the
N. Saarinen et al. / Journal of Steroid Biochemistry & Molecular Biology 78 (2001) 231–239
experiments. The cells were washed with versene and detached with Trypsin– EDTA mixture. The cell suspension was centrifuged at 800 rpm, and the cells were resuspended in serum-free DMEM, counted, and diluted to 1.2×106 cells/ml. The incubation mixture contained 50 ml of 3H-androst-4-ene3,17-dione (in DMEM, final concentration 0.5 nM), 50 ml of unlabeled androstenedione (in DMEM, final concentration 0.5 nM), 100 ml of DMEM solution containing the test compound in ethanol, and 800 ml of cell suspension (i.e. 106 cells). Steroid and test compound stock solutions were prepared in ethanol. The amount of ethanol stock solution in DMEM was 1 ml in 1 ml of DMEM. Aromatase inhibition was tested at five different concentrations of the flavonoids (range 0.1– 10.0 mM) for the determination of IC50 values. After incubation for 4 h in JEG-3 cells and 3 h in Arom + HEK 293 cells, unlabeled carriers (androstenedione, testosterone, 17b-estradiol and estrone) were added. Steroids were extracted twice with 3.0 ml dichloromethane. The combined dichloromethane extracts were evaporated to dryness under nitrogen at +45 °C water bath. The extracted steroids were dissolved in 35% acetonitrile in water before HPLC run. HPLC was used for separation of steroid metabolites and quantification of 3H-17b-estradiol in JEG-3 cells and 3H-androstenedione and 3H-estrone in Arom+ HEK 293 cells. The methodologic details of the quantifications of 3H-labeled products have been described earlier [21]. The column system consisted of a guard column followed by a C18 150×3.9 mm ID analytical column (Technopak T-15, HPLC Technology, Wellington House, Cheshire, UK). The mobile phase was 35% acetonitrile in water, and the flow rate was 1.2 ml/min (analysis of incubation mixture of JEG-3 cells) or 1.0 ml/min (analysis of incubation mixture of Arom+ HEK 293 cells). For in-line detection of radioactive metabolites, the eluent of the HPLC column was continuously mixed with liquid scintillant and then monitored with an in-line radioactivity detector. Aromatase activity (formation of 3H-17b-estradiol from 3H-androstenedione in JEG-3 cells and formation of 3H-estrone in Arom+HEK 293 cells) was calculated as the percentage of 3H-androstenedione converted to metabolites described above. The number of JEG-3 cells was adjusted so that the conversion of the substrate to products was 5– 30% during incubation. In aromatase transfected Arom+ HEK 293 cells, the conversion of substrate to products was 50– 100%. In JEG-3 cells, the interconversion of estrone and estradiol was assayed by determining the ability of cells from confluent monolayers to convert added 3H-estrone to 3H-17b-estradiol. A cell suspension was prepared as described above for the measurement of aromatase activity. The incubation mixture contained 50 ml of 3 H-estrone (in DMEM, final concentration 0.5 nM), 50
233
ml of unlabeled estrone (in DMEM, final concentration 0.5 nM), 100 ml of DMEM solution containing the test compound, and 800 ml of cell suspension (i.e. 106 cells). No conversion of 3H-estrone to 3H-17b-estradiol occurred in transfected Arom+ HEK 293 cells. After incubation for 4 h, unlabeled carriers (estradiol and estrone) were added. The steroids were extracted twice with 3.0 ml of dichloromethane. The combined dichloromethane solutions were evaporated to dryness as described above. Two successive HPLC runs were used to isolate 3H-estrone and 3H-testosterone from 3 H-androstenedione. At the first run, the mobile phase was acetonitrile/water (35/65 vol). The retention times for estradiol, testosterone, estrone, and androstenedione were 9.8, 10.8, 12.2, and 13.0 min, respectively. Fractions containing 3H-estrone, 3H-androstenedione, and trace amounts of 3H-testosterone were collected and extracted with diethyl ether and rerun in water/ methanol/acetic acid (5:4:1 vol). The retention times of androstenedione, estrone, and testosterone were 9.2, 10.9, and 12.8 min, respectively. In addition to the conversion of estrone to 17b-estradiol, no other metabolites (17a-estradiol, 2-hydroxy-, or 4-hydroxy derivatives of estrone or estradiol) were detected.
2.3. In 6itro antioxidati6ity The antioxidative capacity of flavonoids was estimated on the basis of their potency to inhibit tertbutylhydroperoxide-induced lipid perioxidation (t-BuOOH-LP) in rat liver microsomes in vitro [27]. The test for t-BuOOH-LP was done as follows: the buffer (50 mM sodium carbonate, pH 10.2, with 0.1 mM EDTA) was pipetted in a volume of 0.8 ml in the luminometer cuvette. Twenty microliters of diluted liver microsomes, final concentration 1.5 mg protein/ml, was added, followed by luminol (6 ml, 0.5 mg/ml) and test chemicals. The test compounds were added to incubation mixtures in a small volume diluted in ethanol or dimethylsulfoxide (2% of incubation volume), and their potency in inhibiting lipid peroxidation was compared to that of the vehicle (ethanol or dimethyl sulfoxide). The reaction was initiated using 0.05 ml of 0.9 mM t-BuOOH at 33 °C. Chemiluminescence was measured for 45 min at 1 min cycles, and the integral was calculated. Chemiluminescence measurements were made using a model 1251 Luminometer (Bio-Orbit, Turku, Finland) connected to a personal computer, with dedicated software used for the assays.
2.4. In 6i6o uterotropic estrogenicity and aromatase inhibition assays in immature rats The in vivo estrogenicity and ability to inhibit aromatase of selected flavonoids was evaluated by uterotropic tests in immature rats. Sprague-Dawley female
234
N. Saarinen et al. / Journal of Steroid Biochemistry & Molecular Biology 78 (2001) 231–239
rats (originally obtained from Harlan, Horst, Netherlands) were used in the study. Female pups were weaned from dams at the age of 18 days and evenly divided between the treatment groups. Immature animals (3–5 per cage) were housed under a 12:12-h light – dark cycle at 21 °C with 50% humidity and free access to RM3 soy-free diet (SDS Special Diet Services, Witham Essex, UK). In estrogenicity assays, positive control rats (n= 9) were daily gavaged with DES (2 mg per kg body weight). Control rats (n = 10) received daily p.o. 100 ml per 10 g of body weight rape seed oil. The aromatase inhibitor, MPV-2213ad, was administered p.o. to rats (n= 6) at a daily dose of 10 mg/kg body weight. Naringenin (n=4) or chrysin (n = 6) suspended in oil was administered to the rats (50 mg per kg body weight) for 7 days. On day 8 after weaning, i.e. after 7 days of treatment, the animals were suffocated with carbon dioxide and weighed. The uteri were immediately removed and weighed after careful removal of any fluid from the uterine cavity. To sensitize the in vivo aromatase inhibition assay, the uterotropic effects of different doses of androstenedione (5, 10, or 30 mg per kg body weight) given s.c. in rape seed oil (10 ml per 10 g of body weight) were analyzed. Immature female Sprague-Dawley rats were divided into four treatment groups (n = 4 in all dose groups) at the age of 18 days. The control group was treated s.c. daily with rape seed oil (10 ml per 10 g of body weight). All animals were gavaged daily with rape seed oil (100 ml per 10 g of body weight). After 7 days of treatment, the animals (aged 25 days) were killed and weighed. The uteri were removed and weighed as described above. On the basis of the analysis of the different doses of androstenedione, the 30 mg/kg dose was selected for in vivo aromatase inhibition assays with flavonoids. The aromatase inhibition of flavonoids with a daily dose of 50 mg/kg body weight was tested in immature female rats treated s.c. with androstenedione (30 mg/kg of body weight). In the control group (n = 7), 18-dayold rats were injected s.c. with rape seed oil. In the positive control group (n =9), the aromatase inhibitor MPV-2213ad (10 mg/kg. n =6), was administered p.o. in rape seed oil. Apigenin (n =4), naringenin (n =5), or chrysin (n=5) were gavaged to the animals in oil as described above. The weight gain of the animals was followed daily. Twenty-five-day-old animals were suffocated with carbon dioxide, weighed, decapitated, and uterine wet weights were measured.
2.5. Statistical analysis Analyses were done using STATISTICA version 5.1 software for Windows. Normally distributed uterine growth test data was analyzed using the one-way analy-
sis of variance followed by Tukey’s least-significance test. The acceptable level of significance for these analyses was PB0.05.
3. Results
3.1. Aromatase inhibitory and antioxidant properties of fla6onoids in 6itro Table 1 shows the IC50 values of the flavonoids for the inhibition of 3H-17b-estradiol formation from 3Handrostenedione in JEG-3 cells and 3H-estrone formation from 3H-androstenedione in aromatase gene transfected Arom+ HEK 293 cells. Two flavonoids, 7-hydroxyflavone and chrysin, had an IC50 value lower than 1 mM in JEG-3 cells as well as in Arom+HEK 293 cells, whereas flavone and quercetin showed no inhibition (IC50 \ 100 mM). A good correlation was seen between the IC50 values of chrysin and naringenin obtained from JEG-3 cells and Arom+HEK 293 cells. A marked difference was found between the IC50 values of apigenin and naringenin (0.18 vs. 1.4 and 1.4 vs. 3.2, respectively). In JEG-3 cells, the rate of conversion of 3 H-estrone to 3H-estradiol (17b-reduction) was 2.1- to 2.5-fold compared to the formation of 3H-17b-estradiol from 3H-androstenedione (aromatization plus 17b-reduction). Further, apigenin caused an 18% inhibition of 3 H-estrone reduction at a concentration of 1.0 mM while inhibiting 3H-estradiol-17b formation from 3Handrostenedione by 68%. Under the same incubation conditions, 7-hydroxyflavone had no effect on estrone reduction but inhibited aromatization by 70%. Even though no accumulation of 3H-estrone in the culture medium in the presence of 1.0 mM apigenin or 7-hydroxyflavone was detected, it is possible that the lower IC50 value in JEG-3 cells is due to the simultaneous inhibition of estrone formation and reduction to 17bestradiol. Table 1 shows the antioxidative capacities of the flavonoids (IC50 values for inhibition of t-butylperoxide-induced lipid peroxidation). Kaempferide, quercetin, and fisetin appeared to be potent antioxidants but had no or only weak inhibitory effects on aromatization. Luteolin, galangin, and kaempferol showed both inhibitory and antioxidant capacities.
3.2. Estrogenicity and inhibition of androstenedione-induced uterine growth in 6i6o The administration of naringenin or chrysin (p.o.) to immature female rats at a dose of 50 mg per kg body weight had no significant effect on uterine weight gain (Fig. 2). Aromatase inhibitor (MPV-2213ad) reduced uterine weight gain (PB 0.05).
N. Saarinen et al. / Journal of Steroid Biochemistry & Molecular Biology 78 (2001) 231–239
235
Table 1 Aromatase inhibition and antioxidant properties of flavonoids in vitro Compounds
IC50 (mM) Inhibition of 17b-estradiol formation in JEG-3 cellsa and in Arom+HEK 293 cellsb
Fla6ones Flavone \100 7-Hydroxyflavone 0.35 6-Hydroxyflavone 5.5 Chrysin (5,7-dihydroxyflavone) 0.5 (0.6) Apigenin (4%,5,7-trihydroxyflavone) 0.18 (1.4) Luteolin 1.7 (3%,4%,5,7-tetrahydroxyflavone) Galangin (3,5,7-trihydroxyflavone) 12 Kaempferol 11 (3,4%,5,7-tetradyroxyflavone) Fisetin 55 (3,3%,4%,7-tetrahyroxyflavone) Querceting \100 (3,3%,4%,5,7-pentahydroxyflavone) Fla6anones Naringenin 1.4 (3.2) (4%,5,7-trihydroxyflavanone) Methoxyfla6ones Pinostrobin 4 (5-hydroxy-7-methoxyflavone) Acacetin 4 (5,7-dihydroxy-4%,methoxyflavone) Kaempferide (3,5,7-trihydroxy-4%,ethoxyflavone)
a b
Antioxidative capacity (inhibition of t-BuOOH-LP)
\1000 170 120 18 7 0.25 0.44 0.9 0.3 0.4
9
29 7
80
0.5
Values determined using the means of 2–6 measurements at each concentration. Values in parentheses determined by using Arom+HEK 293 cells.
Fig. 2. Estrogenicity of naringenin and chrysin in immature animals. Number of animals in control group 10, DES group 9, MPV-2213ad group 6, naringenin group 4, and chrysin group 6. Each compound was given orally in oil (100 ml per 10 g of body weight) daily for 7 days. *Significantly different from control.
Fig. 3 shows the dose-dependency of uterine growth in androstenedione-treated animals. As expected, aromatase inhibitor MPV-2213ad reduced the increase in uterine weight in androstenedione-treated animals.
However, flavonoids (apigenin, naringenin, chrysin) did not significantly reduce androstenedione-induced uterine growth, indicating a lack of aromatase-inhibiting effect in vivo (Fig. 4). No significant differences in
236
N. Saarinen et al. / Journal of Steroid Biochemistry & Molecular Biology 78 (2001) 231–239
animal weight gain were seen between the different treatment groups (data not shown).
4. Discussion Various flavonoids were tested to determine the structural properties needed for the inhibition of estro-
gen biosynthesis, estrogenicity, and lipid oxidation. Naturally occurring flavonoids such as chrysin and nondietary 7-hydroxyflavone were found to inhibit the formation of 17b-estradiol in JEG-3 and estrone formation in Arom+ HEK 293 cells with an IC50 value lower than 1 mM. The determination of the inhibition was based on quantitative measurements of the endproduct, 3H-17b-estradiol in JEG-3 cells or 3H-estrone
Fig. 3. Uterotropic activity of s.c. androstenedione (AN) in different doses given daily for 7 days, and inhibition of AN-induced uterine growth by aromatase inhibitor (MPV-2213ad). Four animals in each treatment group. *Significantly different from control.
Fig. 4. Inhibition of androstenedione (AN)-induced uterine growth by flavonoids in immature animals. Number of animals in control group 7, AN 30 mg/kg group 9, AN 30 mg/kg+ MPV2213ad group 6, AN 30 mg/kg+naringenin group 5, AN 30 mg/kg+ chrysin group 5 and AN 30 mg/kg+ apigenin group 4. Bars indicated by different letters indicate statistically significant differences (PB 0.05) tested by Tukey’s least significant difference test.
N. Saarinen et al. / Journal of Steroid Biochemistry & Molecular Biology 78 (2001) 231–239
in Arom +HEK 293 cells. In JEG-3 cells, 3H-17bestradiol is formed from 3H-androstenedione through two sequential steps: (1) the aromatization of the A ring with three hydroxylations, loss of the angular methyl group at C19, cis elimination of 1b- and 2b-hydrogens; and (2) 17b-reduction of 3H-estrone. The de3 H-17b-estradiol from creased formation of 3 H-androstenedione in the presence of apigenin and 7-hydroxyflavone was shown to be at least partly due to the inhibition of the aromatization (3H-estrone formation). This became evident when the 17b-reduction activity was shown to be higher than aromatization in JEG-3 cells, and 17b-reduction inhibition by apigenin and 7-hydroxyflavone was lower than aromatization. Moreover, Arom+HEK 293 cells were shown to lack 17b-reductase activity, and the measurement of aromatization was based on estrone accumulation. Our findings largely confirm and extend earlier observations made on the inhibition of placental and preadipocyte aromatases by means of an assay quantitating the production of 3H2O released from 3H-andostenedione after aromatization [14,15,17– 20]. They are also in agreement with a detailed account on the structural requirements of flavonoids for the inhibition of aromatase provided by site-directed mutagenesis studies [28]. When aromatase inhibition is assayed using androstenedione as the substrate for enzyme, flavonoids inhibit estrone formation competitively with respect to a substrate. Kao et al. [28] suggested that flavonoids bind to the active site in an orientation such that their A and C rings mimic rings D and C of the steroid, respectively. The number and location of hydroxyl groups in flavones are important for the aromatase inhibition. As for hydroxylations in ring A, it seems that the 7-hydroxy group in flavone molecule is essential for maximal inhibition [20,28]. The farther the hydroxyl group is from the C-4 keto group, the higher is the inhibitory activity of the compound (6-hydroxyflavone vs. 7-hydroxyflavone in the present study). 4%-Hydroxylation decreases the inhibition capacity of flavone indicated by inhibitory profiles of chrysin and apigenin. A conversion of C-2, C-3 double bond to a single bond decreases the inhibition of aromatase (apigenin vs. naringenin). The hydroxyl group in position 3 of flavones (flavonol structure) (galangin vs. chrysin; kaempferol versus apigenin; quercetin vs. luteolin) and C-3%, C-4%-dihydroxy (luteolin vs. apigenin) reduce their inhibitory capacity. The presence of a hydroxy group in position 3 of the flavone significantly also reduces its ability to bind aromatase [28]. Methylation of the 7- and 4%-OH group reduces the inhibitory effect on 17b-estradiol formation (pinostrobin vs. chrysin and acacetin vs. apigenin) in JEG-3 cells. The structural requirements for aromatase inhibition do not correlate with structures known for participa-
237
tion in antioxidative capacity of flavonoids. Roughly, the best aromatase inhibitors are the poorest antioxidants. The specific structural criteria defining the antioxidant activities of flavonoids include: (i) 4%-hydroxy group; (ii) 3-hydroxyl group in conjugation with the 4-keto group in the C ring; (iii) 5,7-dihydroxy structure in the A ring and (iv) orthodihydroxy structure in the B ring [29–31,15,32]. Our data are in agreement with these conclusions and provide new information. A single hydroxyl group in the 4%-position of the B ring in the flavonoids does not contribute to antioxidant activity (apigenin vs. chrysin; kaempferol vs. galangin). The presence of hydroxyl substituents on the flavone structure enhances antioxidative activity, whereas substitution by the methoxy group diminishes antioxidant activity. O-methylation of 4%-hydroxyl inactivates antioxidant activities of flavones, confirming earlier findings by Cao et al. [30] and Arora et al. [29]. In conclusion, none of the requirements for optimal antioxidative capacity (3-hydroxyl group, 5,7-dihydroxy structure and the orthodihydroxy structure in the B ring) is relevant for the maximum inhibition of aromatase by flavonoids. Structural demands for the inhibition of aromatase also differ from those of the 17b-HSOR type 1 enzyme (E.C. 1.1.1.62) [21]. In cultured breast cancer cells (T-47D cells), this enzyme is primarily a reductase, converting the weak endogenous estrogen (estrone) to the hormonally more potent estrogen 17b-estradiol. The 5- and 4%-OH groups in the flavone structure are essential for 17b-HSOR inhibition but not for aromatase inhibition (chrysin vs. 7-hydroxyflavone and kaempferol vs. galangin). The most striking difference was that 7-hydroxyflavone was one of the best inhibitors of aromatase, whereas it was completely inactive on 17b-HSOR type 1 [21,22]. Further, chrysin was a potent aromatase inhibitor but inhibited 17b-HSOR type 1 weakly. Comparisons between the chemical requirements of aromatase and 17b-HSOR type 1 inhibition also suggest some similarities in the structural requirements. Apigenin, naringenin, luteolin, and acacetin seem to inhibit both aromatase and 17bHSOR type 1 enzyme at 1 mM concentrations [22]. In both cases, the hydroxyl group in position 3 of flavones (flavonol structure) (galangin vs. chrysin; kaempferol vs. apigenin; quercetin vs. luteolin) and C-3%, C-4%-dihydroxy (luteolin vs. apigenin) reduce their inhibitory activity. The structural requirements for aromatase inhibition do not correlate closely with structures known for their participation in the estrogenicity of flavonoids [13,33,34]. The main feature required to confer estrogenicity is the presence of hydroxyl substituents in the 4% and 7 positions of the flavan nuclei [33]. An additional hydroxyl group in the 5 position may increase
238
N. Saarinen et al. / Journal of Steroid Biochemistry & Molecular Biology 78 (2001) 231–239
estrogenic activity in some cases. Two hydroxyl groups in positions 4% and 5 are not essential structural features for the compound as aromatase inhibitor. Chrysin and 7-hydroxyflavone are nonestrogenic but potent aromatase inhibitors. Further, the flavanol structure may increase estrogenicity while decreasing the capacity of flavones for inhibiting aromatization. These differences in the structural requirements for aromatase inhibition and estrogenicity may be explained by a difference in the binding orientation between flavones and steroids in the binding sites of aromatase and ERs [28]. According to Miksicek [33], mono- and dihydroxyflavones compete with 17b-estradiol for binding to ERa only when present in a 1000- to 10 000-fold molar excess over the natural steroid, and their affinities to ERa show a range of 0.3– 10 mM. In the study of Kuiper et al. [13], the binding affinities of several flavonoids to both ER subtypes (ERa and ERb) were determined. As for flavone or 5,7-dihydroxyflavone (chrysin), no significant competition with 17b-estradiol for binding to either ER was detected at concentrations of up to 10 mM (i.e. affinities more than 10 000-fold lower than that of 17b-estradiol). In contrast, the IC50 values of flavonoids for steroid metabolizing enzymes are approximately 1 mM, which is in the same range as the Km values that these enzymes have for steroid substrates. Mono- and dihydroxyflavones can thus probably better compete for the active sites of the enzymes than those of ERs and be more effective in altering steroid metabolism than in replacing steroids from the receptor binding sites, as also recently suggested by Baker [35]. On the other hand, some tri- and tetrahydroxyflavonoids bind to ERs, particularly ERb, at relatively high affinity. For example, 4%,5,7-trihydroxyflavone (apigenin) has been reported to bind to ERa and ERb with affinities about 300- and 20-fold lower than that of 17b-estradiol, respectively [13]. Therefore, the possible simultaneous interaction of certain flavonoids with steroid metabolizing enzymes and ERs (ERb) cannot be excluded. Obviously, the effects of flavonoids in vivo cannot be predicted on the basis of these in vitro results alone. In vivo experiments are vital. The administration of nonestrogenic (chrysin) or estrogenic (naringenin and apigenin) flavonoids neither induced uterine growth nor reduced estrogen- or androgen-induced uterine growth. The dose, 50 mg/kg body weight, considerably exceeds amounts found in daily human diets [36]. The inability of flavonoids to inhibit aromatase and, consequently, uterine growth may be due to their poor absorption and/or bioavailability. Recently, in the study of Breinholt et al. [37], oral apigenin (100 mg/kg body weight) had no estrogenic or antiestrogenic effect on uterine growth in immature mice. Small portions (a few percentages maximally) of administered dose excreted in the urine are consistent with these explanations [38–41].
Further, little is known about the metabolism of flavonoids. According to the proposed general concept for changes in the flavonoid B-ring, 4%-hydroxyl and catechol (3%- and 4%-dihydroxyl) structures may be the primary products if flavonoids are biotransformed by the mono-oxygenases of the endoplasmic reticulum [42]. This would decrease, but not abolish, their ability to inhibit aromatase (apigenin vs. chrysin and luteolin vs. apigenin). Dealkylation of flavonoids such as acacetin is also possible, which would be beneficial for their inhibitory capacity. This makes it unlikely that metabolic inactivation accounts for the lack of biological activity of aromatase-inhibiting flavonoids in vivo. In conclusion, minor changes in flavonoid structure may result in major changes in biological activity. Theoretically, some flavonoids could be important in reducing breast cancer risk by decreasing the production of endogenous estrogens via inhibition of the key enzymes in estrogen biosynthesis. Other flavones, although very similar by structure, may be completely inactive, and/or exert other actions important in cancer risk modulation, such as ER-mediated effects or antioxidation. Therefore, when assessing the role of dietary flavonoids in the regulation of endogenous estrogen production, and subsequently in breast cancer chemoprevention, it is important to evaluate the intake of and exposure to individual compounds, shown to interact with aromatase and/or 17b-HSOR. Further, the compounds should be sufficiently absorbed and bioavailable. The consumption of ‘flavonoid-rich’ food items only indicates exposure to a highly variable mixture of active and inactive compounds, and does thus not allow any further conclusions. Furthermore, it should be considered that flavonoids may be metabolized, possibly resulting in crucial changes in their biological activity.
Acknowledgements The authors would like to thank Dr Matti Poutanen, for helping in generating the aromatase expressing 293 cells, Tuula Tanner for skillful technical assistance and Leena Gro¨ nlund for assistance in the preparation of the manuscript. This work was funded by Tekes, the National Technology Agency of Finland, Project 51583/00 and EC grant ENV4-CT96-0204.
References [1] G.A. Colditz, Relationship between estrogen levels, use of hormone replacement therapy, and breast cancer, J. Natl. Cancer Inst. 90 (1998) 814 – 832. [2] C. Lopez-Otin, E.P. Diamandis, Breast and prostate cancer: an analysis of common epidemiological, genetic, and biochemical features, Endocr. Rev. 19 (1998) 365 – 396.
N. Saarinen et al. / Journal of Steroid Biochemistry & Molecular Biology 78 (2001) 231–239 [3] A. Ekbom, D. Trichopoulos, H.O. Adami, C.C. Hsieh, S.J. Lan, Evidence of prenatal influences on breast cancer risk, Lancet 340 (1992) 1015 – 1018. [4] B.E. Henderson, R.K. Ross, M.C. Pike, J.T. Casagrande, Endogenous hormones as a major factor in human cancer, Cancer Res. 42 (1982) 3232 –3239. [5] G.J. Kelloff, R.A. Lubet, R. Lieberman, K. Eisenhauer, V.E. Steele, J.A. Crowell, E.T. Hawk, C.W. Boone, C.C. Sigman, Aromatase inhibitors as potential cancer chemopreventives, Cancer Epidemiol. Biomarkers Prev. 7 (1998) 65 –78. [6] G.R. Howe, T. Hirohata, T.G. Hislop, J.M. Iscovich, J.M. Yuan, K. Katsouyanni, F. Lupin, E. Marubini, B. Modan, T. Rohan, P. Toniolo, Y. Shunzhang, Dietary factors and risk of breast cancer: combined analysis of 12 case-control studies, J. Natl. Cancer Inst. 82 (1990) 561 –569. [7] J.D. Potter, K. Steinmetz, Vegetables, fruit and phytoestrogens as preventive agents, in: B.W. Stewart, D. McGregory, P. Kleihues (Eds.), Principles of Chemoprevention. In: IARC Scientific Publications, vol. 129, International Agency for Research on Cancer, Lyon, 1996, pp. 61 –89. [8] M. Gerber, Fibre and breast cancer, Eur. J. Cancer Prev. 7 (Suppl. 2) (1998) S63 – S67. [9] H. Adlercreutz, Phytoestrogens: epidemiology and a possible role in cancer protection, Environ. Health Perspect. 103 (Suppl. 7) (1995) 103 – 112. [10] M.S. Kurzer, X. Xu, Dietary phytoestrogens, Annu. Rev. Nutr. 17 (1997) 353. [11] A.L. Murkies, G.L. Wilcox, S.R. Davis, Phytoestrogens, J. Clin. Endocrinol. Metab. 83 (1998) 297 –303. [12] S.M. Kuo, Dietary flavonoid and cancer prevention: evidence and potential mechanism, Crit. Rev. Oncog. 8 (1997) 47 – 69. [13] G.G. Kuiper, J.G. Lemmen, B. Carlsson, J.C. Corton, S.H. Safe, P.T. van der Saag, B. van der Burg, J.A. Gustafsson, Interaction of estrogenic chemicals and phytoestrogens with estrogen receptor beta, Endocrinology 139 (10) (1998) 4252 – 4263. [14] H-J. Jeong, Y.G. Shin, J-H. Kim, J.M. Pozzuto, Inhibition of aromatase activity by flavonoids, Arch. Pharm. Res. 22 (1999) 309 – 312. [15] S.A. Van Acker, D.J. van den Berg, M.N. Tromp, D.H. Griffionen, W.P. van Bennekom, W.J. van der Vijg, A. Bast, Structural aspects of antioxidant activity of flavonoids, Free Radic. Biol. Med. 20 (1996) 331 –342. [16] C. Wang, T. Ma¨ kela¨ , T. Hase, H. Adlercreutz, M.S. Kurzer, Lignans and flavonoids inhibit aromatase enzyme in human adipocytes, J. Steroid Biochem. Mol. Biol. 50 (1994) 205 – 212. [17] D.R. Campbell, M.S. Kurzer, Flavonoid inhibition of aromatase enzyme activity in human preadipocytes, J. Steroid Biochem. Mol. Biol. 46 (1993) 381 – 388. [18] F.H. De Jong, K. Oishi, R.B. Hayes, J.F.A.T. Bogdanowicz, A.R. Ibrahim, Y.J. Abul Hajj, Aromatase inhibition by flavonoids, J. Steroid Biochem. Mol. Biol. 37 (1990) 257 – 260. [19] A.R.. Ibrahim, Y.J. Abul-Hajj, Aromatase inhibition by flavonoids, J. Steroid Biochem. Mol. Biol. 37 (2) (1990) 257 – 260. [20] J.T. Kellis Jr., L.E. Vickery, Inhibition of human estrogen synthetase (aromatase) by flavones, Science 225 (1984) 1032 – 1034. [21] S. Ma¨ kela¨ , M. Poutanen, J. Lehtima¨ ki, M.-L. Kostian, R. Santti, R. Vihko, Estrogen-specific 17b-hydroxysteroid oxidoreductase type I (E.C. 1.1.1.62) as a possible target for the action of phytoestrogens, P.S.E.B.M. 208 (1995) 51 –59. [22] S. Ma¨ kela¨ , M. Poutanen, M.-L. Kostian, N. Lehtima¨ ki, L. Strauss, R. Santti, R. Vihko, Inhibition of 17b-hydroxysteroid oxidoreductase by flavonoids in breast and prostate cancer cells, P.S.E.B.M. 217 (1998) 310 – 316. [23] J.L. Slavin, Mechanisms for the impact of whole grain fodds on cancer risk, J. Am. Coll. Nutr. 19 (Suppl. 3) (2000) 300S – 307S.
239
[24] M.D. Krekels, W. Wouters, R. Decoster, R. Van Ginckel, A. Leonares, P.A. Janssen, Aromatase in the human choriocarcinoma JEG-3: inhibition by R76713 in cultured cells and in tumors grown in nude mice, J. Steroid Biochem. Mol. Biol. 38 (1991) 415 – 422. [25] O. Ahokoski, K. Irjala, R. Huupponen, K. Halonen, E. Salminen, H. Scheinin, Hormonal effects of MPV-2213ad, a new selective aromatase inhibitior in healthy male subjects. A phase I study, Br. J. Clin. Pharmacol. 45 (2) (1998) 141 – 146. [26] X. Li, E. Nokkala, W. Yan, T. Streng, N. Saarinen, A. Wa¨ rri, R. Santti, I. Huhtaniemi, S. Ma¨ kela¨ , M. Poutanen, Reproductive tract dysfunction in aromatase over-expressing transgenic male mice maintaining elevated serum estradiol and low testosterone concentrations, Endocrinology 142 (6) (2001) 2435 – 2441. [27] M. Ahotupa, E. Ma¨ ntyla¨ , L. Kangas, Antioxidant properties of the triphenylethylene antiestrogen drug toremifene, NaunynSchmiedebergs. Arch. Pharmacol. 356 (1997) 297 – 302. [28] Y.C. Kao, C. Zhou, M. Sherman, C.A. Laughton, S. Chen, Molecular basis of the inhibition of human aromatase (estrogen synthetase) by flavone and isoflavone phytoestrogens: a site-directed mutagenesis study, Environ. Health Perspect. 106 (1998) 85 – 92. [29] A. Arora, M.G. Nair, G.M. Strasburg, Structure – activity relationships for antioxidant activities of a series of flavonoids in liposomal system, Free Radic. Biol. Med. 24 (1998) 355 –363. [30] G. Cao, E. Sofice, R.L. Prior, Antioxidant and prooxidant behaviour of flavonoids: structure – activity relationships, Free Radic. Biol. Med. 22 (1997) 749 – 760. [31] C. Rice-Evans, N.J. Miller, G. Paganga, Structure – antioxidant activity relationships of flavonoids and phenolic acids, Free Radic. Biol. Med. 20 (1996) 933 – 956. [32] A. Saija, M. Scalese, M. Lanza, D. Marzullo, F. Bonina, F. Castelli, Flavonoids as antioxidant agents: importance of their interaction with biomembranes, Free Radic. Biol. Med. 19 (1995) 481 – 486. [33] R.J. Miksicek, Estrogenic flavonoids: structural requirements for biological activity, P.S.E.B.M. 208 (1995) 44 – 50. [34] J.C. Le Bail, F. Varnat, J.C. Nicolas, G. Habrioux, Estrogenic and antiproliferative activities on MCF-7 human breast cancer cells by flavonoids, Cancer Lett. 130 (1998) 209 – 216. [35] M.E. Baker, Hormone mimics and molecular fossils, Chemtech August (1997) 26 – 34. [36] A. Scalbert, G. Williamson, Dietary intake and bioavailability of polyphenols, J. Nutr. 130 (2000) 2073S – 2085S. [37] V. Breinholt, A. Hossaini, G.W. Svendsen, C. Brouwer, S.E. Nielsen, Estrogenic activity of flavonoids in mice. The importance of estrogen receptor distribution, metabolism and bioavailability, Food Chem. Toxicol. 38 (2000) 555 – 564. [38] B. Ameer, R.A. Weintraub, J.V. Johnson, R.A. Yost, R.L. Rouseff, Flavanone absorbtion after naringin, hesperidin, and citrus administraion, Clin. Pharmacol. Ther. 60 (1996) 34 –40. [39] P.C.H. Hollman, Bioavailability of flavonoids, Eur. J. Clin. Nutr. 51 (Suppl. 1) (1997) S66 – S69. [40] K. Jansen, R.P. Mensink, F.J.J. Cox, J.L. Harryvan, R. Hovenier, P.C.H. Hollman, M.B. Katan, Effeects of the flavonoids quercetin and apigenin on hemostasis in healthy volunteers: results from an in vitro and a dietary supplement study, Am. J. Clin. Nutr. 67 (1998) 255 – 262. [41] S.E. Nielsen, J.F. Young, B. Daneshvar, S.T. Lauridsen, P. Knuthsen, B. Sandstrom, L.O. Dragsted, Effect of parsley (Petroselinum crispum) intake on urinary apigenin excretion, blood antioxidant enzymes and biomarkers for oxidative stress in humans, Br. J. Nutr. 81 (1999) 447 – 455. [42] S.E. Nielsen, V. Breinholt, U. Justesen, C. Cornett, L.O. Dragsted, In vitro biotransformation of flavonoids by rat liver microsomes, Xenobiotica 28 (1998) 389 – 401.
No evidence for the in vivo activity of aromatase-inhibiting flavonoids Niina Saarinen a,*, Suresh C. Joshi a, Markku Ahotupa a, Xiangdong Li a, Jenni A8 mma¨la¨ a, Sari Ma¨kela¨ a,b, Risto Santti a b
a Institute of Biomedicine, Uni6ersity of Turku, Kiinamyllynkatu 10, FIN-20520 Turku, Finland Unit for Pre6enti6e Nutrition, Department of Medical Nutrition, Karolinska Institute, S-14157 Huddinge, Sweden
Received 4 December 2000; accepted 11 April 2001
Abstract Measurements of the aromatase-inhibiting and antioxidative capacities of flavonoids in vitro showed that slight changes in flavonoid structure may result in marked changes in biological activity. Several flavonoids such as 7-hydroxyflavone and chrysin (5,7-dihydroxyflavone) were shown to inhibit the formation of 3H-17b-estradiol from 3H-androstenedione (IC50 B 1.0 mM) in human choriocarcinoma JEG-3 cells and in human embryonic kidney cells HEK 293 transfected with human aromatase gene (Arom+ HEK 293). Flavone and quercetin (3,3%,4%,5,7-pentahydroxyflavone) showed no inhibition (IC50 \ 100 mM). None of the requirements for optimal antioxidative capacity (2,3-double bond with 4%-hydroxy group, 3-hydroxyl group, 5,7-dihydroxy structure and the orthodihydroxy structure in the B-ring) is relevant for the maximum inhibition of aromatase by flavonoids. After oral administration to immature rats at a dose of 50 mg/kg body weight, which considerably exceeds amounts found in daily human diets, neither aromatase-inhibiting nonestrogenic flavonoids, such as chrysin, nor estrogenic flavonoids, such as naringenin and apigenin, induced uterine growth or reduced estrogen- or androgen-induced uterine growth. The inability of flavonoids to inhibit aromatase and, consequently, uterine growth in short-term tests may be due to their relatively poor absorption and/or bioavailability. © 2001 Elsevier Science Ltd. All rights reserved. Keywords: Aromatase; Flavonoid; Estrogen; Diet
1. Introduction The risk of human breast cancer (BC) is related to cumulative exposure of breast tissue to endogenous estrogens [1–4]. Experimental evidence also strongly favors the role of estrogens in the development and growth of BC. Strategies aiming at reduced estrogen production may be useful for the prevention of estrogen-related BC [5]. The incidence rates of BC are higher in populations consuming high-fat, low-fiber diets [2,6] than in populations with diets rich in fruit and vegetables [6,7]. In addition to the fiber effect [8], studies suggest that the possible beneficial effects of low-fat and high-fiber diets
* Corresponding author. Tel.: + 358-2-333-7223; fax: +358-2-3337352. E-mail address: [email protected] (N. Saarinen).
are related to weakly estrogenic compounds known as phytoestrogens [9–11]. Isoflavonoids are the most extensively studied phytoestrogens. The intake and serum concentrations of isoflavonoids, such as daidzein and genistein, mainly derived from soy, are higher in Asian countries, where the incidence of BC is low [9]. Several edible plants present in Asian as well as Western diets contain a wide variety of flavonoids with biologic actions similar to those of isoflavonoids [12]. They may also contribute to the chemopreventive actions of vegetable- and fruit-rich diets. The mechanism of the possible chemopreventive action of flavonoids is not fully known. Besides binding to the two known estrogen receptor subtypes (ERa and ERb) [13], flavonoids may compete with endogenous substrates for active sites of estrogen biosynthesizing and metabolizing enzymes, such as aromatase [14 –20] and 17b-hydroxysteroid oxidoreductase (HSOR), type 1
0960-0760/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved. PII: S 0 9 6 0 - 0 7 6 0 ( 0 1 ) 0 0 0 9 8 - X
N. Saarinen et al. / Journal of Steroid Biochemistry & Molecular Biology 78 (2001) 231–239
232
[21,22]. The inhibition of these enzymes lowers 17bestradiol concentrations in target cells, which then decrease the risk of BC. In addition, some flavonoids act as antioxidants, which may also play a role in the prevention of carcinogenesis in general [23]. To understand the significance and the complex role of individual flavonoids or their mixtures from various dietary sources in BC chemoprevention, it is important to recognize the critical structural properties determining the mechanisms of flavonoid actions. One of our aims was to identify the function of flavonoids as well as to understand the mode of their possible chemopreventive action. Earlier, we studied the inhibitory actions of several flavonoids on the 17b-HSORs type 1 [21] and 2 [22]. Here, we report on the effects of the same flavonoids on the formation of 3H-17b-estradiol from 3 H-andostenedione in JEG-3 cells, a human chorioncarcinoma cell line [24], and in Arom+HEK 293 cells transfected with the construct for the synthesis of aromatase enzyme. In addition to the ability of flavonoids to inhibit estrogen biosynthesis, we estimated their antioxidant properties and compared the structural requirements for selective inhibition of estrogen biosynthesis (aromatization and 17b-reduction) with those for antioxidative capacity and estrogenicity. The most potent aromatase inhibiting flavonoids were chosen for in vivo studies of estrogenicity and antiestrogenicity.
2. Materials and methods
2.1. Chemicals The test compounds were purchased from the following sources: apigenin (4%,5,7-trihydroxyflavone), chrysin (5,7-dihydroxyflavone), kaempferol (3,4%,5,7-tetrahydroxyflavone), quercetin (3,3%,4%,5,7-pentahydroxyflav-
Fig. 1. General structures of flavonoids.
one), 4-androstene-3,17-dione and diethylstilbestrol (DES) were obtained from Sigma Chemical Co. (St. Louis, MO, USA), 7-hydroxyflavone, fisetin (3,3%,4%,7tetrahydroxyflavone), kaempferide (3,5,7-trihydroxy-4%methoxyflavone), luteolin (3%,4%,5,7-tetrahydroxyflavone), galangin (3,5,7-trihydroxyflavone), naringenin (4%,5,7-trihydroxyflavanone) and pinostrobin (5-hydroxy-7-methoxyflavone) from Carl Roth GmbH (Karlsruhe, Germany). Fig. 1 shows the main chemical structures of the flavonoids. The cell culture materials were obtained from the following sources: Dulbecco’s modified Eagle medium (DMEM), fetal calf serum (FCS) and Trypsin–EDTA from Gibco BRL (Paisley, Scotland). The five-well plates and culture dishes were from NUNC (Roskilde, Denmark). Versene–EDTA in phosphate buffered saline contained NaCl (136 mM), KCl (2.68 mM), Na2HPO4 (8.1 mM), KH2PO4 (1.47 mM) and EDTA disodium salt (0.054 mM). 3H-androst-4-ene,3,17-dione and 3H-estrone were from NEN Life Science Products (Zaventem, Belgium). Unlabeled androstenedione, testosterone, 17b-estradiol, and estrone were obtained from Sigma Chemical Co. (St. Louis, MO, USA). Dichloromethane was purchased from Rathburn (Peebleshire, UK). MPV-2213ad aromatase inhibitor [25] was obtained from Hormos Medical Ltd (Turku, Finland).
2.2. In 6itro aromatase inhibition assays The aromatase activity in JEG-3, human choriocarcinoma cells, and in HEK 293 cells stably expressing the aromatase (Arom+HEK 293) was assayed by determining the ability of the cells to convert added 3 H-androstenedione to 3H-17b-estradiol and 3H-estrone. The procedure for the construction of the aromatase expression plasmid (pUbC-AROM) has been previously described [26]. For constructing the stable cell lines, pUbC-AROM-plasmid was transfected to HEK-293 cells grown in DMEM medium supplemented with 10% FCS using the DOTAP liposomal Transfection Kit (Boehringer Mannheim, Mannheim, Germany) according to the manufacturer’s instructions. After transfection, the cells were cultured for 48 h and transferred to the selection medium (DMEM, 10% FCS, 900 mgG418/ml medium). Cloned cells were applied in 96well plates at a density of 0.5 cells/well and grown for 2 months in the presence of the selection medium. The cloned cells lines were then screened for aromatase expression by RT-PCR and enzyme activity as previously described [26]. Cells were maintained in DMEM containing 10% of charcoal stripped heat inactivated FCS. The experiments with JEG-3 cells were done using cells from confluent monolayers (approximately 5 days after subculture) in 30×10 mm culture dishes. Arom+HEK 293 cells were cultured for 2–3 days prior to the
N. Saarinen et al. / Journal of Steroid Biochemistry & Molecular Biology 78 (2001) 231–239
experiments. The cells were washed with versene and detached with Trypsin– EDTA mixture. The cell suspension was centrifuged at 800 rpm, and the cells were resuspended in serum-free DMEM, counted, and diluted to 1.2×106 cells/ml. The incubation mixture contained 50 ml of 3H-androst-4-ene3,17-dione (in DMEM, final concentration 0.5 nM), 50 ml of unlabeled androstenedione (in DMEM, final concentration 0.5 nM), 100 ml of DMEM solution containing the test compound in ethanol, and 800 ml of cell suspension (i.e. 106 cells). Steroid and test compound stock solutions were prepared in ethanol. The amount of ethanol stock solution in DMEM was 1 ml in 1 ml of DMEM. Aromatase inhibition was tested at five different concentrations of the flavonoids (range 0.1– 10.0 mM) for the determination of IC50 values. After incubation for 4 h in JEG-3 cells and 3 h in Arom + HEK 293 cells, unlabeled carriers (androstenedione, testosterone, 17b-estradiol and estrone) were added. Steroids were extracted twice with 3.0 ml dichloromethane. The combined dichloromethane extracts were evaporated to dryness under nitrogen at +45 °C water bath. The extracted steroids were dissolved in 35% acetonitrile in water before HPLC run. HPLC was used for separation of steroid metabolites and quantification of 3H-17b-estradiol in JEG-3 cells and 3H-androstenedione and 3H-estrone in Arom+ HEK 293 cells. The methodologic details of the quantifications of 3H-labeled products have been described earlier [21]. The column system consisted of a guard column followed by a C18 150×3.9 mm ID analytical column (Technopak T-15, HPLC Technology, Wellington House, Cheshire, UK). The mobile phase was 35% acetonitrile in water, and the flow rate was 1.2 ml/min (analysis of incubation mixture of JEG-3 cells) or 1.0 ml/min (analysis of incubation mixture of Arom+ HEK 293 cells). For in-line detection of radioactive metabolites, the eluent of the HPLC column was continuously mixed with liquid scintillant and then monitored with an in-line radioactivity detector. Aromatase activity (formation of 3H-17b-estradiol from 3H-androstenedione in JEG-3 cells and formation of 3H-estrone in Arom+HEK 293 cells) was calculated as the percentage of 3H-androstenedione converted to metabolites described above. The number of JEG-3 cells was adjusted so that the conversion of the substrate to products was 5– 30% during incubation. In aromatase transfected Arom+ HEK 293 cells, the conversion of substrate to products was 50– 100%. In JEG-3 cells, the interconversion of estrone and estradiol was assayed by determining the ability of cells from confluent monolayers to convert added 3H-estrone to 3H-17b-estradiol. A cell suspension was prepared as described above for the measurement of aromatase activity. The incubation mixture contained 50 ml of 3 H-estrone (in DMEM, final concentration 0.5 nM), 50
233
ml of unlabeled estrone (in DMEM, final concentration 0.5 nM), 100 ml of DMEM solution containing the test compound, and 800 ml of cell suspension (i.e. 106 cells). No conversion of 3H-estrone to 3H-17b-estradiol occurred in transfected Arom+ HEK 293 cells. After incubation for 4 h, unlabeled carriers (estradiol and estrone) were added. The steroids were extracted twice with 3.0 ml of dichloromethane. The combined dichloromethane solutions were evaporated to dryness as described above. Two successive HPLC runs were used to isolate 3H-estrone and 3H-testosterone from 3 H-androstenedione. At the first run, the mobile phase was acetonitrile/water (35/65 vol). The retention times for estradiol, testosterone, estrone, and androstenedione were 9.8, 10.8, 12.2, and 13.0 min, respectively. Fractions containing 3H-estrone, 3H-androstenedione, and trace amounts of 3H-testosterone were collected and extracted with diethyl ether and rerun in water/ methanol/acetic acid (5:4:1 vol). The retention times of androstenedione, estrone, and testosterone were 9.2, 10.9, and 12.8 min, respectively. In addition to the conversion of estrone to 17b-estradiol, no other metabolites (17a-estradiol, 2-hydroxy-, or 4-hydroxy derivatives of estrone or estradiol) were detected.
2.3. In 6itro antioxidati6ity The antioxidative capacity of flavonoids was estimated on the basis of their potency to inhibit tertbutylhydroperoxide-induced lipid perioxidation (t-BuOOH-LP) in rat liver microsomes in vitro [27]. The test for t-BuOOH-LP was done as follows: the buffer (50 mM sodium carbonate, pH 10.2, with 0.1 mM EDTA) was pipetted in a volume of 0.8 ml in the luminometer cuvette. Twenty microliters of diluted liver microsomes, final concentration 1.5 mg protein/ml, was added, followed by luminol (6 ml, 0.5 mg/ml) and test chemicals. The test compounds were added to incubation mixtures in a small volume diluted in ethanol or dimethylsulfoxide (2% of incubation volume), and their potency in inhibiting lipid peroxidation was compared to that of the vehicle (ethanol or dimethyl sulfoxide). The reaction was initiated using 0.05 ml of 0.9 mM t-BuOOH at 33 °C. Chemiluminescence was measured for 45 min at 1 min cycles, and the integral was calculated. Chemiluminescence measurements were made using a model 1251 Luminometer (Bio-Orbit, Turku, Finland) connected to a personal computer, with dedicated software used for the assays.
2.4. In 6i6o uterotropic estrogenicity and aromatase inhibition assays in immature rats The in vivo estrogenicity and ability to inhibit aromatase of selected flavonoids was evaluated by uterotropic tests in immature rats. Sprague-Dawley female
234
N. Saarinen et al. / Journal of Steroid Biochemistry & Molecular Biology 78 (2001) 231–239
rats (originally obtained from Harlan, Horst, Netherlands) were used in the study. Female pups were weaned from dams at the age of 18 days and evenly divided between the treatment groups. Immature animals (3–5 per cage) were housed under a 12:12-h light – dark cycle at 21 °C with 50% humidity and free access to RM3 soy-free diet (SDS Special Diet Services, Witham Essex, UK). In estrogenicity assays, positive control rats (n= 9) were daily gavaged with DES (2 mg per kg body weight). Control rats (n = 10) received daily p.o. 100 ml per 10 g of body weight rape seed oil. The aromatase inhibitor, MPV-2213ad, was administered p.o. to rats (n= 6) at a daily dose of 10 mg/kg body weight. Naringenin (n=4) or chrysin (n = 6) suspended in oil was administered to the rats (50 mg per kg body weight) for 7 days. On day 8 after weaning, i.e. after 7 days of treatment, the animals were suffocated with carbon dioxide and weighed. The uteri were immediately removed and weighed after careful removal of any fluid from the uterine cavity. To sensitize the in vivo aromatase inhibition assay, the uterotropic effects of different doses of androstenedione (5, 10, or 30 mg per kg body weight) given s.c. in rape seed oil (10 ml per 10 g of body weight) were analyzed. Immature female Sprague-Dawley rats were divided into four treatment groups (n = 4 in all dose groups) at the age of 18 days. The control group was treated s.c. daily with rape seed oil (10 ml per 10 g of body weight). All animals were gavaged daily with rape seed oil (100 ml per 10 g of body weight). After 7 days of treatment, the animals (aged 25 days) were killed and weighed. The uteri were removed and weighed as described above. On the basis of the analysis of the different doses of androstenedione, the 30 mg/kg dose was selected for in vivo aromatase inhibition assays with flavonoids. The aromatase inhibition of flavonoids with a daily dose of 50 mg/kg body weight was tested in immature female rats treated s.c. with androstenedione (30 mg/kg of body weight). In the control group (n = 7), 18-dayold rats were injected s.c. with rape seed oil. In the positive control group (n =9), the aromatase inhibitor MPV-2213ad (10 mg/kg. n =6), was administered p.o. in rape seed oil. Apigenin (n =4), naringenin (n =5), or chrysin (n=5) were gavaged to the animals in oil as described above. The weight gain of the animals was followed daily. Twenty-five-day-old animals were suffocated with carbon dioxide, weighed, decapitated, and uterine wet weights were measured.
2.5. Statistical analysis Analyses were done using STATISTICA version 5.1 software for Windows. Normally distributed uterine growth test data was analyzed using the one-way analy-
sis of variance followed by Tukey’s least-significance test. The acceptable level of significance for these analyses was PB0.05.
3. Results
3.1. Aromatase inhibitory and antioxidant properties of fla6onoids in 6itro Table 1 shows the IC50 values of the flavonoids for the inhibition of 3H-17b-estradiol formation from 3Handrostenedione in JEG-3 cells and 3H-estrone formation from 3H-androstenedione in aromatase gene transfected Arom+ HEK 293 cells. Two flavonoids, 7-hydroxyflavone and chrysin, had an IC50 value lower than 1 mM in JEG-3 cells as well as in Arom+HEK 293 cells, whereas flavone and quercetin showed no inhibition (IC50 \ 100 mM). A good correlation was seen between the IC50 values of chrysin and naringenin obtained from JEG-3 cells and Arom+HEK 293 cells. A marked difference was found between the IC50 values of apigenin and naringenin (0.18 vs. 1.4 and 1.4 vs. 3.2, respectively). In JEG-3 cells, the rate of conversion of 3 H-estrone to 3H-estradiol (17b-reduction) was 2.1- to 2.5-fold compared to the formation of 3H-17b-estradiol from 3H-androstenedione (aromatization plus 17b-reduction). Further, apigenin caused an 18% inhibition of 3 H-estrone reduction at a concentration of 1.0 mM while inhibiting 3H-estradiol-17b formation from 3Handrostenedione by 68%. Under the same incubation conditions, 7-hydroxyflavone had no effect on estrone reduction but inhibited aromatization by 70%. Even though no accumulation of 3H-estrone in the culture medium in the presence of 1.0 mM apigenin or 7-hydroxyflavone was detected, it is possible that the lower IC50 value in JEG-3 cells is due to the simultaneous inhibition of estrone formation and reduction to 17bestradiol. Table 1 shows the antioxidative capacities of the flavonoids (IC50 values for inhibition of t-butylperoxide-induced lipid peroxidation). Kaempferide, quercetin, and fisetin appeared to be potent antioxidants but had no or only weak inhibitory effects on aromatization. Luteolin, galangin, and kaempferol showed both inhibitory and antioxidant capacities.
3.2. Estrogenicity and inhibition of androstenedione-induced uterine growth in 6i6o The administration of naringenin or chrysin (p.o.) to immature female rats at a dose of 50 mg per kg body weight had no significant effect on uterine weight gain (Fig. 2). Aromatase inhibitor (MPV-2213ad) reduced uterine weight gain (PB 0.05).
N. Saarinen et al. / Journal of Steroid Biochemistry & Molecular Biology 78 (2001) 231–239
235
Table 1 Aromatase inhibition and antioxidant properties of flavonoids in vitro Compounds
IC50 (mM) Inhibition of 17b-estradiol formation in JEG-3 cellsa and in Arom+HEK 293 cellsb
Fla6ones Flavone \100 7-Hydroxyflavone 0.35 6-Hydroxyflavone 5.5 Chrysin (5,7-dihydroxyflavone) 0.5 (0.6) Apigenin (4%,5,7-trihydroxyflavone) 0.18 (1.4) Luteolin 1.7 (3%,4%,5,7-tetrahydroxyflavone) Galangin (3,5,7-trihydroxyflavone) 12 Kaempferol 11 (3,4%,5,7-tetradyroxyflavone) Fisetin 55 (3,3%,4%,7-tetrahyroxyflavone) Querceting \100 (3,3%,4%,5,7-pentahydroxyflavone) Fla6anones Naringenin 1.4 (3.2) (4%,5,7-trihydroxyflavanone) Methoxyfla6ones Pinostrobin 4 (5-hydroxy-7-methoxyflavone) Acacetin 4 (5,7-dihydroxy-4%,methoxyflavone) Kaempferide (3,5,7-trihydroxy-4%,ethoxyflavone)
a b
Antioxidative capacity (inhibition of t-BuOOH-LP)
\1000 170 120 18 7 0.25 0.44 0.9 0.3 0.4
9
29 7
80
0.5
Values determined using the means of 2–6 measurements at each concentration. Values in parentheses determined by using Arom+HEK 293 cells.
Fig. 2. Estrogenicity of naringenin and chrysin in immature animals. Number of animals in control group 10, DES group 9, MPV-2213ad group 6, naringenin group 4, and chrysin group 6. Each compound was given orally in oil (100 ml per 10 g of body weight) daily for 7 days. *Significantly different from control.
Fig. 3 shows the dose-dependency of uterine growth in androstenedione-treated animals. As expected, aromatase inhibitor MPV-2213ad reduced the increase in uterine weight in androstenedione-treated animals.
However, flavonoids (apigenin, naringenin, chrysin) did not significantly reduce androstenedione-induced uterine growth, indicating a lack of aromatase-inhibiting effect in vivo (Fig. 4). No significant differences in
236
N. Saarinen et al. / Journal of Steroid Biochemistry & Molecular Biology 78 (2001) 231–239
animal weight gain were seen between the different treatment groups (data not shown).
4. Discussion Various flavonoids were tested to determine the structural properties needed for the inhibition of estro-
gen biosynthesis, estrogenicity, and lipid oxidation. Naturally occurring flavonoids such as chrysin and nondietary 7-hydroxyflavone were found to inhibit the formation of 17b-estradiol in JEG-3 and estrone formation in Arom+ HEK 293 cells with an IC50 value lower than 1 mM. The determination of the inhibition was based on quantitative measurements of the endproduct, 3H-17b-estradiol in JEG-3 cells or 3H-estrone
Fig. 3. Uterotropic activity of s.c. androstenedione (AN) in different doses given daily for 7 days, and inhibition of AN-induced uterine growth by aromatase inhibitor (MPV-2213ad). Four animals in each treatment group. *Significantly different from control.
Fig. 4. Inhibition of androstenedione (AN)-induced uterine growth by flavonoids in immature animals. Number of animals in control group 7, AN 30 mg/kg group 9, AN 30 mg/kg+ MPV2213ad group 6, AN 30 mg/kg+naringenin group 5, AN 30 mg/kg+ chrysin group 5 and AN 30 mg/kg+ apigenin group 4. Bars indicated by different letters indicate statistically significant differences (PB 0.05) tested by Tukey’s least significant difference test.
N. Saarinen et al. / Journal of Steroid Biochemistry & Molecular Biology 78 (2001) 231–239
in Arom +HEK 293 cells. In JEG-3 cells, 3H-17bestradiol is formed from 3H-androstenedione through two sequential steps: (1) the aromatization of the A ring with three hydroxylations, loss of the angular methyl group at C19, cis elimination of 1b- and 2b-hydrogens; and (2) 17b-reduction of 3H-estrone. The de3 H-17b-estradiol from creased formation of 3 H-androstenedione in the presence of apigenin and 7-hydroxyflavone was shown to be at least partly due to the inhibition of the aromatization (3H-estrone formation). This became evident when the 17b-reduction activity was shown to be higher than aromatization in JEG-3 cells, and 17b-reduction inhibition by apigenin and 7-hydroxyflavone was lower than aromatization. Moreover, Arom+HEK 293 cells were shown to lack 17b-reductase activity, and the measurement of aromatization was based on estrone accumulation. Our findings largely confirm and extend earlier observations made on the inhibition of placental and preadipocyte aromatases by means of an assay quantitating the production of 3H2O released from 3H-andostenedione after aromatization [14,15,17– 20]. They are also in agreement with a detailed account on the structural requirements of flavonoids for the inhibition of aromatase provided by site-directed mutagenesis studies [28]. When aromatase inhibition is assayed using androstenedione as the substrate for enzyme, flavonoids inhibit estrone formation competitively with respect to a substrate. Kao et al. [28] suggested that flavonoids bind to the active site in an orientation such that their A and C rings mimic rings D and C of the steroid, respectively. The number and location of hydroxyl groups in flavones are important for the aromatase inhibition. As for hydroxylations in ring A, it seems that the 7-hydroxy group in flavone molecule is essential for maximal inhibition [20,28]. The farther the hydroxyl group is from the C-4 keto group, the higher is the inhibitory activity of the compound (6-hydroxyflavone vs. 7-hydroxyflavone in the present study). 4%-Hydroxylation decreases the inhibition capacity of flavone indicated by inhibitory profiles of chrysin and apigenin. A conversion of C-2, C-3 double bond to a single bond decreases the inhibition of aromatase (apigenin vs. naringenin). The hydroxyl group in position 3 of flavones (flavonol structure) (galangin vs. chrysin; kaempferol versus apigenin; quercetin vs. luteolin) and C-3%, C-4%-dihydroxy (luteolin vs. apigenin) reduce their inhibitory capacity. The presence of a hydroxy group in position 3 of the flavone significantly also reduces its ability to bind aromatase [28]. Methylation of the 7- and 4%-OH group reduces the inhibitory effect on 17b-estradiol formation (pinostrobin vs. chrysin and acacetin vs. apigenin) in JEG-3 cells. The structural requirements for aromatase inhibition do not correlate with structures known for participa-
237
tion in antioxidative capacity of flavonoids. Roughly, the best aromatase inhibitors are the poorest antioxidants. The specific structural criteria defining the antioxidant activities of flavonoids include: (i) 4%-hydroxy group; (ii) 3-hydroxyl group in conjugation with the 4-keto group in the C ring; (iii) 5,7-dihydroxy structure in the A ring and (iv) orthodihydroxy structure in the B ring [29–31,15,32]. Our data are in agreement with these conclusions and provide new information. A single hydroxyl group in the 4%-position of the B ring in the flavonoids does not contribute to antioxidant activity (apigenin vs. chrysin; kaempferol vs. galangin). The presence of hydroxyl substituents on the flavone structure enhances antioxidative activity, whereas substitution by the methoxy group diminishes antioxidant activity. O-methylation of 4%-hydroxyl inactivates antioxidant activities of flavones, confirming earlier findings by Cao et al. [30] and Arora et al. [29]. In conclusion, none of the requirements for optimal antioxidative capacity (3-hydroxyl group, 5,7-dihydroxy structure and the orthodihydroxy structure in the B ring) is relevant for the maximum inhibition of aromatase by flavonoids. Structural demands for the inhibition of aromatase also differ from those of the 17b-HSOR type 1 enzyme (E.C. 1.1.1.62) [21]. In cultured breast cancer cells (T-47D cells), this enzyme is primarily a reductase, converting the weak endogenous estrogen (estrone) to the hormonally more potent estrogen 17b-estradiol. The 5- and 4%-OH groups in the flavone structure are essential for 17b-HSOR inhibition but not for aromatase inhibition (chrysin vs. 7-hydroxyflavone and kaempferol vs. galangin). The most striking difference was that 7-hydroxyflavone was one of the best inhibitors of aromatase, whereas it was completely inactive on 17b-HSOR type 1 [21,22]. Further, chrysin was a potent aromatase inhibitor but inhibited 17b-HSOR type 1 weakly. Comparisons between the chemical requirements of aromatase and 17b-HSOR type 1 inhibition also suggest some similarities in the structural requirements. Apigenin, naringenin, luteolin, and acacetin seem to inhibit both aromatase and 17bHSOR type 1 enzyme at 1 mM concentrations [22]. In both cases, the hydroxyl group in position 3 of flavones (flavonol structure) (galangin vs. chrysin; kaempferol vs. apigenin; quercetin vs. luteolin) and C-3%, C-4%-dihydroxy (luteolin vs. apigenin) reduce their inhibitory activity. The structural requirements for aromatase inhibition do not correlate closely with structures known for their participation in the estrogenicity of flavonoids [13,33,34]. The main feature required to confer estrogenicity is the presence of hydroxyl substituents in the 4% and 7 positions of the flavan nuclei [33]. An additional hydroxyl group in the 5 position may increase
238
N. Saarinen et al. / Journal of Steroid Biochemistry & Molecular Biology 78 (2001) 231–239
estrogenic activity in some cases. Two hydroxyl groups in positions 4% and 5 are not essential structural features for the compound as aromatase inhibitor. Chrysin and 7-hydroxyflavone are nonestrogenic but potent aromatase inhibitors. Further, the flavanol structure may increase estrogenicity while decreasing the capacity of flavones for inhibiting aromatization. These differences in the structural requirements for aromatase inhibition and estrogenicity may be explained by a difference in the binding orientation between flavones and steroids in the binding sites of aromatase and ERs [28]. According to Miksicek [33], mono- and dihydroxyflavones compete with 17b-estradiol for binding to ERa only when present in a 1000- to 10 000-fold molar excess over the natural steroid, and their affinities to ERa show a range of 0.3– 10 mM. In the study of Kuiper et al. [13], the binding affinities of several flavonoids to both ER subtypes (ERa and ERb) were determined. As for flavone or 5,7-dihydroxyflavone (chrysin), no significant competition with 17b-estradiol for binding to either ER was detected at concentrations of up to 10 mM (i.e. affinities more than 10 000-fold lower than that of 17b-estradiol). In contrast, the IC50 values of flavonoids for steroid metabolizing enzymes are approximately 1 mM, which is in the same range as the Km values that these enzymes have for steroid substrates. Mono- and dihydroxyflavones can thus probably better compete for the active sites of the enzymes than those of ERs and be more effective in altering steroid metabolism than in replacing steroids from the receptor binding sites, as also recently suggested by Baker [35]. On the other hand, some tri- and tetrahydroxyflavonoids bind to ERs, particularly ERb, at relatively high affinity. For example, 4%,5,7-trihydroxyflavone (apigenin) has been reported to bind to ERa and ERb with affinities about 300- and 20-fold lower than that of 17b-estradiol, respectively [13]. Therefore, the possible simultaneous interaction of certain flavonoids with steroid metabolizing enzymes and ERs (ERb) cannot be excluded. Obviously, the effects of flavonoids in vivo cannot be predicted on the basis of these in vitro results alone. In vivo experiments are vital. The administration of nonestrogenic (chrysin) or estrogenic (naringenin and apigenin) flavonoids neither induced uterine growth nor reduced estrogen- or androgen-induced uterine growth. The dose, 50 mg/kg body weight, considerably exceeds amounts found in daily human diets [36]. The inability of flavonoids to inhibit aromatase and, consequently, uterine growth may be due to their poor absorption and/or bioavailability. Recently, in the study of Breinholt et al. [37], oral apigenin (100 mg/kg body weight) had no estrogenic or antiestrogenic effect on uterine growth in immature mice. Small portions (a few percentages maximally) of administered dose excreted in the urine are consistent with these explanations [38–41].
Further, little is known about the metabolism of flavonoids. According to the proposed general concept for changes in the flavonoid B-ring, 4%-hydroxyl and catechol (3%- and 4%-dihydroxyl) structures may be the primary products if flavonoids are biotransformed by the mono-oxygenases of the endoplasmic reticulum [42]. This would decrease, but not abolish, their ability to inhibit aromatase (apigenin vs. chrysin and luteolin vs. apigenin). Dealkylation of flavonoids such as acacetin is also possible, which would be beneficial for their inhibitory capacity. This makes it unlikely that metabolic inactivation accounts for the lack of biological activity of aromatase-inhibiting flavonoids in vivo. In conclusion, minor changes in flavonoid structure may result in major changes in biological activity. Theoretically, some flavonoids could be important in reducing breast cancer risk by decreasing the production of endogenous estrogens via inhibition of the key enzymes in estrogen biosynthesis. Other flavones, although very similar by structure, may be completely inactive, and/or exert other actions important in cancer risk modulation, such as ER-mediated effects or antioxidation. Therefore, when assessing the role of dietary flavonoids in the regulation of endogenous estrogen production, and subsequently in breast cancer chemoprevention, it is important to evaluate the intake of and exposure to individual compounds, shown to interact with aromatase and/or 17b-HSOR. Further, the compounds should be sufficiently absorbed and bioavailable. The consumption of ‘flavonoid-rich’ food items only indicates exposure to a highly variable mixture of active and inactive compounds, and does thus not allow any further conclusions. Furthermore, it should be considered that flavonoids may be metabolized, possibly resulting in crucial changes in their biological activity.
Acknowledgements The authors would like to thank Dr Matti Poutanen, for helping in generating the aromatase expressing 293 cells, Tuula Tanner for skillful technical assistance and Leena Gro¨ nlund for assistance in the preparation of the manuscript. This work was funded by Tekes, the National Technology Agency of Finland, Project 51583/00 and EC grant ENV4-CT96-0204.
References [1] G.A. Colditz, Relationship between estrogen levels, use of hormone replacement therapy, and breast cancer, J. Natl. Cancer Inst. 90 (1998) 814 – 832. [2] C. Lopez-Otin, E.P. Diamandis, Breast and prostate cancer: an analysis of common epidemiological, genetic, and biochemical features, Endocr. Rev. 19 (1998) 365 – 396.
N. Saarinen et al. / Journal of Steroid Biochemistry & Molecular Biology 78 (2001) 231–239 [3] A. Ekbom, D. Trichopoulos, H.O. Adami, C.C. Hsieh, S.J. Lan, Evidence of prenatal influences on breast cancer risk, Lancet 340 (1992) 1015 – 1018. [4] B.E. Henderson, R.K. Ross, M.C. Pike, J.T. Casagrande, Endogenous hormones as a major factor in human cancer, Cancer Res. 42 (1982) 3232 –3239. [5] G.J. Kelloff, R.A. Lubet, R. Lieberman, K. Eisenhauer, V.E. Steele, J.A. Crowell, E.T. Hawk, C.W. Boone, C.C. Sigman, Aromatase inhibitors as potential cancer chemopreventives, Cancer Epidemiol. Biomarkers Prev. 7 (1998) 65 –78. [6] G.R. Howe, T. Hirohata, T.G. Hislop, J.M. Iscovich, J.M. Yuan, K. Katsouyanni, F. Lupin, E. Marubini, B. Modan, T. Rohan, P. Toniolo, Y. Shunzhang, Dietary factors and risk of breast cancer: combined analysis of 12 case-control studies, J. Natl. Cancer Inst. 82 (1990) 561 –569. [7] J.D. Potter, K. Steinmetz, Vegetables, fruit and phytoestrogens as preventive agents, in: B.W. Stewart, D. McGregory, P. Kleihues (Eds.), Principles of Chemoprevention. In: IARC Scientific Publications, vol. 129, International Agency for Research on Cancer, Lyon, 1996, pp. 61 –89. [8] M. Gerber, Fibre and breast cancer, Eur. J. Cancer Prev. 7 (Suppl. 2) (1998) S63 – S67. [9] H. Adlercreutz, Phytoestrogens: epidemiology and a possible role in cancer protection, Environ. Health Perspect. 103 (Suppl. 7) (1995) 103 – 112. [10] M.S. Kurzer, X. Xu, Dietary phytoestrogens, Annu. Rev. Nutr. 17 (1997) 353. [11] A.L. Murkies, G.L. Wilcox, S.R. Davis, Phytoestrogens, J. Clin. Endocrinol. Metab. 83 (1998) 297 –303. [12] S.M. Kuo, Dietary flavonoid and cancer prevention: evidence and potential mechanism, Crit. Rev. Oncog. 8 (1997) 47 – 69. [13] G.G. Kuiper, J.G. Lemmen, B. Carlsson, J.C. Corton, S.H. Safe, P.T. van der Saag, B. van der Burg, J.A. Gustafsson, Interaction of estrogenic chemicals and phytoestrogens with estrogen receptor beta, Endocrinology 139 (10) (1998) 4252 – 4263. [14] H-J. Jeong, Y.G. Shin, J-H. Kim, J.M. Pozzuto, Inhibition of aromatase activity by flavonoids, Arch. Pharm. Res. 22 (1999) 309 – 312. [15] S.A. Van Acker, D.J. van den Berg, M.N. Tromp, D.H. Griffionen, W.P. van Bennekom, W.J. van der Vijg, A. Bast, Structural aspects of antioxidant activity of flavonoids, Free Radic. Biol. Med. 20 (1996) 331 –342. [16] C. Wang, T. Ma¨ kela¨ , T. Hase, H. Adlercreutz, M.S. Kurzer, Lignans and flavonoids inhibit aromatase enzyme in human adipocytes, J. Steroid Biochem. Mol. Biol. 50 (1994) 205 – 212. [17] D.R. Campbell, M.S. Kurzer, Flavonoid inhibition of aromatase enzyme activity in human preadipocytes, J. Steroid Biochem. Mol. Biol. 46 (1993) 381 – 388. [18] F.H. De Jong, K. Oishi, R.B. Hayes, J.F.A.T. Bogdanowicz, A.R. Ibrahim, Y.J. Abul Hajj, Aromatase inhibition by flavonoids, J. Steroid Biochem. Mol. Biol. 37 (1990) 257 – 260. [19] A.R.. Ibrahim, Y.J. Abul-Hajj, Aromatase inhibition by flavonoids, J. Steroid Biochem. Mol. Biol. 37 (2) (1990) 257 – 260. [20] J.T. Kellis Jr., L.E. Vickery, Inhibition of human estrogen synthetase (aromatase) by flavones, Science 225 (1984) 1032 – 1034. [21] S. Ma¨ kela¨ , M. Poutanen, J. Lehtima¨ ki, M.-L. Kostian, R. Santti, R. Vihko, Estrogen-specific 17b-hydroxysteroid oxidoreductase type I (E.C. 1.1.1.62) as a possible target for the action of phytoestrogens, P.S.E.B.M. 208 (1995) 51 –59. [22] S. Ma¨ kela¨ , M. Poutanen, M.-L. Kostian, N. Lehtima¨ ki, L. Strauss, R. Santti, R. Vihko, Inhibition of 17b-hydroxysteroid oxidoreductase by flavonoids in breast and prostate cancer cells, P.S.E.B.M. 217 (1998) 310 – 316. [23] J.L. Slavin, Mechanisms for the impact of whole grain fodds on cancer risk, J. Am. Coll. Nutr. 19 (Suppl. 3) (2000) 300S – 307S.
239
[24] M.D. Krekels, W. Wouters, R. Decoster, R. Van Ginckel, A. Leonares, P.A. Janssen, Aromatase in the human choriocarcinoma JEG-3: inhibition by R76713 in cultured cells and in tumors grown in nude mice, J. Steroid Biochem. Mol. Biol. 38 (1991) 415 – 422. [25] O. Ahokoski, K. Irjala, R. Huupponen, K. Halonen, E. Salminen, H. Scheinin, Hormonal effects of MPV-2213ad, a new selective aromatase inhibitior in healthy male subjects. A phase I study, Br. J. Clin. Pharmacol. 45 (2) (1998) 141 – 146. [26] X. Li, E. Nokkala, W. Yan, T. Streng, N. Saarinen, A. Wa¨ rri, R. Santti, I. Huhtaniemi, S. Ma¨ kela¨ , M. Poutanen, Reproductive tract dysfunction in aromatase over-expressing transgenic male mice maintaining elevated serum estradiol and low testosterone concentrations, Endocrinology 142 (6) (2001) 2435 – 2441. [27] M. Ahotupa, E. Ma¨ ntyla¨ , L. Kangas, Antioxidant properties of the triphenylethylene antiestrogen drug toremifene, NaunynSchmiedebergs. Arch. Pharmacol. 356 (1997) 297 – 302. [28] Y.C. Kao, C. Zhou, M. Sherman, C.A. Laughton, S. Chen, Molecular basis of the inhibition of human aromatase (estrogen synthetase) by flavone and isoflavone phytoestrogens: a site-directed mutagenesis study, Environ. Health Perspect. 106 (1998) 85 – 92. [29] A. Arora, M.G. Nair, G.M. Strasburg, Structure – activity relationships for antioxidant activities of a series of flavonoids in liposomal system, Free Radic. Biol. Med. 24 (1998) 355 –363. [30] G. Cao, E. Sofice, R.L. Prior, Antioxidant and prooxidant behaviour of flavonoids: structure – activity relationships, Free Radic. Biol. Med. 22 (1997) 749 – 760. [31] C. Rice-Evans, N.J. Miller, G. Paganga, Structure – antioxidant activity relationships of flavonoids and phenolic acids, Free Radic. Biol. Med. 20 (1996) 933 – 956. [32] A. Saija, M. Scalese, M. Lanza, D. Marzullo, F. Bonina, F. Castelli, Flavonoids as antioxidant agents: importance of their interaction with biomembranes, Free Radic. Biol. Med. 19 (1995) 481 – 486. [33] R.J. Miksicek, Estrogenic flavonoids: structural requirements for biological activity, P.S.E.B.M. 208 (1995) 44 – 50. [34] J.C. Le Bail, F. Varnat, J.C. Nicolas, G. Habrioux, Estrogenic and antiproliferative activities on MCF-7 human breast cancer cells by flavonoids, Cancer Lett. 130 (1998) 209 – 216. [35] M.E. Baker, Hormone mimics and molecular fossils, Chemtech August (1997) 26 – 34. [36] A. Scalbert, G. Williamson, Dietary intake and bioavailability of polyphenols, J. Nutr. 130 (2000) 2073S – 2085S. [37] V. Breinholt, A. Hossaini, G.W. Svendsen, C. Brouwer, S.E. Nielsen, Estrogenic activity of flavonoids in mice. The importance of estrogen receptor distribution, metabolism and bioavailability, Food Chem. Toxicol. 38 (2000) 555 – 564. [38] B. Ameer, R.A. Weintraub, J.V. Johnson, R.A. Yost, R.L. Rouseff, Flavanone absorbtion after naringin, hesperidin, and citrus administraion, Clin. Pharmacol. Ther. 60 (1996) 34 –40. [39] P.C.H. Hollman, Bioavailability of flavonoids, Eur. J. Clin. Nutr. 51 (Suppl. 1) (1997) S66 – S69. [40] K. Jansen, R.P. Mensink, F.J.J. Cox, J.L. Harryvan, R. Hovenier, P.C.H. Hollman, M.B. Katan, Effeects of the flavonoids quercetin and apigenin on hemostasis in healthy volunteers: results from an in vitro and a dietary supplement study, Am. J. Clin. Nutr. 67 (1998) 255 – 262. [41] S.E. Nielsen, J.F. Young, B. Daneshvar, S.T. Lauridsen, P. Knuthsen, B. Sandstrom, L.O. Dragsted, Effect of parsley (Petroselinum crispum) intake on urinary apigenin excretion, blood antioxidant enzymes and biomarkers for oxidative stress in humans, Br. J. Nutr. 81 (1999) 447 – 455. [42] S.E. Nielsen, V. Breinholt, U. Justesen, C. Cornett, L.O. Dragsted, In vitro biotransformation of flavonoids by rat liver microsomes, Xenobiotica 28 (1998) 389 – 401.