Materials and Methods

Chemicals.

Unless stated all chemicals were purchased from the Sigma-Aldrich (St. Louis, MO). Tangeretin was obtained from Indofine Chemical Co. (Somerville, NJ). Nobiletin from orange oil was a gift from Dr. William Widmer, Florida Department of Citrus, Lakeland, FL. PhIP was purchased from ICN Biomedicals Inc. (Costa Mesa, CA). Morphine was obtained from Sigma/RBI (Natick, MA).

Liver Specimens.

Human liver samples were obtained from the liver transplant unit at the Medical College of Wisconsin under a protocol approved by the Committee for the Conduct of Human Research. Microsomes from a mixture of four human liver samples (B, H, M, and P) were prepared by differential centrifugation as described previously (van der Hoeven and Coon, 1974) and mixed for use in these assays. The livers were chosen because they had average levels of estradiol-3-glucuronidation (UGT1A1) activities (Fisher et al., 2000a).

Assay for Modulation of Estradiol Glucuronidation Activity.

The primary focus of the study was to investigate modulation of UGT1A1 activity using estradiol-3-glucuronidation as a probe. Preliminary experiments (data not shown) were performed to ensure estradiol glucuronidation experiments were carried out under initial rate conditions. Preliminary range finding experiments (data not shown) were also performed at an estradiol concentration of 25 μM (near its S₅₀ value, substrate concentration at half-maximal activity) to generate information on the potency of each compound as a modulator (at 5, 10, 30, 50, and 100 μM modulator) of estradiol-3-glucuronidation. Under these conditions, the 16 compounds either resulted in inhibition of estradiol 3-glucuronidation or had no effect. The results obtained in the preliminary studies were used to generate the appropriate ranges of modulator concentrations for subsequent experiments. Structures of each of the compounds are shown in Table 1.

The reaction conditions for assay of modulation of estradiol glucuronidation were as follows. Before each incubation, human liver microsomes (0.5 mg/ml), 0.1 M potassium phosphate buffer (pH 7.1), and alamethicin (25 μg/ml) were mixed and placed on ice for 15 min. This was followed by the addition of MgCl₂ (1 mM, final concentration), saccharolactone (5 mM, final concentration), and estradiol (25 μM for preliminary range-finding experiments, and 5, 10, 25, 50, and 100 μM in subsequent experiments) added in methanol (0.5% final concentration). Compounds tested for modulation of estradiol glucuronidation activity were added in methanol (0.5% v/v in the final incubation) except for bilirubin, morphine, retinoic acid, and anthraflavic acid, which were added in dimethyl sulfoxide (0.5% v/v in the final incubation). As retinoic acid and bilirubin are light sensitive, these substrates were protected from light by aluminum foil throughout the experimental procedure. After a 3-min preincubation period in a shaking water bath (37°C), the reaction was started by the addition of UDPGA (5 mM, final concentration) to make a final incubation volume of 200 μl. After 30 min, the reaction was stopped by addition of ice-cold formic acid (25% v/v in buffer, 50 μl). α-Napthylglucuronide (4 nmol) was added as internal standard, and the tubes were kept on ice for 30 min. The mixtures were centrifuged and supernatant transferred to vials for mass spectrometry analysis. The effects of 17α-EE (10 μM) on estradiol glucuronidation at 1.25, 2.5, 5, 10, 15, 20, 30, 50, 75, and 100 μM estradiol were also examined.

Analyses of glucuronide formation were performed on a Micromass platform LCZ system (Micromass Ltd., Manchester, UK) equipped with a Waters Alliance 2690 separations module and column oven (30°C), a 3-μm, 100 × 2 mm Prodigy ODS (3) high performance liquid chromatography column (Phenomenex, Torrance, CA) and a SecurityGuard column (Phenomenex). The mobile phase solution A was 10 mM ammonium acetate, and solution B was 90% acetonitrile/10% water/10% ammonium acetate. Initial conditions were 85% A/15% B pumped at 0.25 ml/min. A linear gradient from 15 to 31% B between 0 and 8 min was used, followed by 1 min at 100% B, and a re-equilibration at 15% B. Analytes were detected as their [M − H]⁻ions using negative ion electrospray ionization. The source block temperature was 125°C, desolvation temperature was 400°C, capillary voltage was −2.70 KV, and the cone voltage was −30V. The internal standard α-napthylglucuronide, estradiol-3-glucuronide, and estradiol-17-glucuronide were detected by single ion monitoring atm/z 319, 447, and 447 and eluted at 5.6, 8.5 and 9 min, respectively. Standard curve correlation coefficients (r² ) were ≥0.99.

Data Analysis.

Duplicate values for rate of estradiol glucuronide formation for each substrate concentration were fit to equations describing hyperbolic (Michaelis-Menten, eq. 1; Segel, 1975) or sigmoidal (Hill, eq.2; Segel, 1975) relationships using WinNonlin software (Pharsight Corporation, Mountain View, CA). The best fit of the data to a relationship was determined according to established criteria (Ring et al., 1994).ν=Vmax[S]/(Km+[S]) Equation 1ν=Vmax[S]n/(S50+[S]n) Equation 2Where enzyme kinetic data for estradiol glucuronidation best fit the Michaelis-Menten equation, Enzyme Kinetics!Pro (SynexChem, Fairfield, CA) and WinNonlin software were used to characterize the mode of inhibition (competitive, noncompetitive, mixed, or uncompetitive). Effects of each of the compounds at the lowest concentration of estradiol tested (5 μM) were determined, and if appropriate, IC₅₀ values were estimated. At this concentration of estradiol, compounds that increased or decreased estradiol glucuronidation by greater than 20% were respectively classed as stimulators or inhibitors (Table 1).

Results

Modulation of Estradiol-3-glucuronidation.

In the absence of added modulator, enzyme kinetics consistent with homotropic activation were observed for estradiol-3-glucuronidation over a concentration range of 5 to 100 μM in each of 16 experiments. The mean V_max (± standard deviation) and S₅₀ (± standard deviation) values for the 16 determinations were 433 ± 79 pmol/mg/min and 22 ± 5 μM. The mean Hill coefficient value (n), which gives an indication of the degree of sigmoidicity of the curve (n of 1 = Michaelis-Menten kinetics, no sigmoidicity) was 1.9 ± 0.6.

The structures of the 16 compounds examined are shown in Table 1. Figures 1 and2 show the modulatory effect of each of the flavonoids and nonflavonoid compounds, respectively, over the indicated range of concentrations on estradiol-3-glucuronidation. Results are expressed as a percentage of control activity at estradiol concentrations of 5, 10, 25, 50, and 100 μM. Interestingly, several patterns of interaction between the modulators and UGT1A1-catalyzed estradiol-3-glucuronidation were observed. The rates of estradiol 3-glucuronidation at all estradiol concentrations examined were basically unaffected by the addition of increasing concentrations of the UGT2B7 substrates ibuprofen and morphine (Fig. 2). The pattern expected for a competitive inhibitor of estradiol 3-glucuronidation is a decrease in activity as the modulator concentration increases. In addition for competitive inhibition, as substrate concentration increases the observed loss of glucuronide formation in the presence of inhibitor should decrease. Such patterns were observed for the effects of tangeretin (Fig. 1) and bilirubin (Fig. 2). However, as the kinetics of estradiol 3-glucuronide formation fitted the sigmoidal kinetic model best, and not hyperbolic kinetics described by the Michaelis-Menten equation, it was not appropriate to fit the inhibitory effects data to classic models of inhibition to generateK_i values.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Estradiol-3-glucuronidation activity in the presence of flavonoids in a pooled mix of human liver microsomes expressed as a percentage of control activity.

Symbols indicate the following estradiol concentrations: ●, 5 μM; ○, 10 μM; ▾, 25 μM; ▿, 50 μM; ▪, 100 μM. Mean control activities ± standard deviation (in the absence of modulator) from 16 experiments in pmol/mg/min were 43 ± 12 at 5 μM estradiol, 94 ± 15 at 10 μM, 210 ± 54 at 25 μM, 349 ± 94 at 50 μM, and 369 ± 106 at 100 μM. The mean S₅₀ (± standard deviation) was 22 ± 5 μM. The mean Hill coefficient, which gives an indication of the sigmoidicity of the curve, was 1.9 ± 0.6.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Estradiol-3-glucuronidation activity in the presence of estrogens, bilirubin, anthraflavic acid, PhIP, and the UGT2B7 substrates morphine, retinoic acid, and ibuprofen.

Symbols indicate the following estradiol concentrations: ●, 5 μM; ○, 10 μM; ▾, 25 μM; ▿, 50 μM; ▪, 100 μM. Mean control activities ± standard deviation (in the absence of modulator) from 16 experiments in pmol/mg/min were 43 ± 12 at 5 μM estradiol, 94 ± 15 at 10 μM, 210 ± 54 at 25 μM, 349 ± 94 at 50 μM, and 369 ± 106 at 100 μM. The mean S₅₀ (± standard deviation) was 22 ± 5 μM. The mean Hill coefficient, which gives an indication of the sigmoidicity of the curve, was 1.9 ± 0.6.

Naringenin inhibited estradiol 3-glucuronidation to a similar extent at all naringenin concentrations tested (Fig. 1) and therefore did not act as a competitive-type inhibitor. Flavone, quercetin (both Fig. 1), and retinoic acid (Fig. 2) were weak inhibitors of estradiol 3-glucuronidation at the modulator concentrations tested, however, it was difficult to determine from these results whether they were competitive-type inhibitors or not. For chrysin (5, 7-dihydroxyflavone), flavanone, nobiletin, silymarin (all Fig. 1), estriol, and anthraflavic acid (both Fig. 2), the greatest inhibitory effect on estradiol 3-glucuronidation was at substrate concentrations above 25 μM. In addition, for chrysin, flavanone, nobiletin, silymarin (all Fig. 1), estriol, and anthraflavic acid (Fig. 2), there seemed to be little effect, or possibly slight stimulation, at the lowest concentrations of estradiol examined (5 and 10 μM), principally at the lowest effector concentration. PhIP had a slight stimulatory effect at the lowest estradiol concentrations but had no inhibitory effects at any of the PhIP concentrations tested (Fig. 2).

Clear demonstration of heterotropic activation of UGT1A1 activity by 17α-EE was observed in two separate experiments designed to test for this phenomenon (Figs. 2 and 3). Figure 2shows that at a substrate concentration of 5 μM, estradiol-3-glucuronidation was stimulated to 180% of the untreated value by 10 μM 17α-EE. However, at this substrate concentration, as the 17α-EE concentration increased, stimulation of estradiol-3-glucuronidation was attenuated so that by 50 μM 17α-EE, estradiol-3-glucuronide formation was 125% of the control activity and by 75 μM 17α-EE, approximately 20% inhibition was observed. An Eadie-Hofstee plot of this data (Fig. 3) clearly shows that increasing concentrations of 17α-EE (10 and 50 μM in this example) shift the shape of the line from a “curve” (observed in the absence of 17α-EE), which is typical of homotropic activation to a more linear relationship observed for Michaelis-Menten behavior. To more fully evaluate the effect of 17α-EE on estradiol glucuronidation, 10 μM 17α-EE was added to a wide range of estradiol concentrations, and its effects on estradiol 3-glucuronidation and estradiol 17-glucuronidation were examined (Fig. 4). The results demonstrate that 10 μM 17α-EE was shown to stimulate estradiol-3-glucuronidation by as much as 180% of control values at the low estradiol concentrations of 1.25 to 10 μM (below the S₅₀ value). This confirmed the observation of heterotropic activation at low 17α-EE and of substrate concentrations in the previously described experiment (Fig. 2) and that stimulation decreased with increasing substrate concentration. However, at higher effector concentrations, estradiol 3-glucuronidation formation was inhibited by 17α-EE (Fig. 2). Interestingly, as would be expected for competitive inhibition by 17α-EE, estradiol 17-glucuronidation was decreased by 10 μM 17α-EE at lower substrate concentrations (Fig.4). In addition, as shown in Fig. 6, estradiol 17-glucuronide formation was inhibited by 17α-EE in a concentration-dependent manner with the greatest inhibition occurring at the lowest substrate concentration.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Eadie-Hofstee plot of estradiol 3-glucuronidation in the absence and presence of 17α-EE.

In the absence of 17α-EE there is a “curved” relationship typical of homotropic activation kinetics. However, this moves toward a linear relationship, which is more typical of Michaelis-Menten kinetics in the presence of increasing concentrations (10 and 50 μM) of 17α-EE. This data are replotted data from the graph shown in Fig. 2.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Heterotropic activation of estradiol-3-glucuronidation and inhibition by 17α-EE-dependence on substrate concentration.

With a wide range of estradiol concentrations, 17α-EE (10 μM) is seen to stimulate estradiol-3-glucuronidation at 1.25 to 10 μM estradiol and inhibit estradiol-3-glucuronidation at 15 to 100 μM, whereas estradiol-17-glucuronidation is inhibited at all estradiol concentrations tested.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6

Estradiol-17-glucuronidation activity in the presence of estrogens, bilirubin, anthraflavic acid, PhIP, and the UGT2B7 substrates morphine, retinoic acid, and ibuprofen.

Results are expressed as pmol estradiol-17-glucuronide/mg protein/min. The mean V_max (± standard deviation) value for 16 experiments was 111 ± 26 pmol/mg/min. The meanK_m value (± standard deviation) was 7 ± 5 μM. Symbols indicate the following estradiol concentrations: ●, 5 μM; ○, 10 μM; ▾, 25 μM; ▿, 50 μM; ▪, 100 μM.

Since estradiol-3-glucuronidation activity exhibited homotropic activation at low substrate concentration, and the compounds investigated appeared to differentially modulate the UGT1A1 enzyme at low versus high estradiol concentration, a classification of the effect of each compound was made at 5 μM estradiol (Table 1). Flavone, chrysin, quercetin, nobiletin, tangeretin, bilirubin, and retinoic acid were inhibitors of estradiol 3-glucuronidation at 5 μM. Heterotropic activation was observed for 17α-EE, PhIP, and anthraflavic acid. Flavanone, naringenin, ibuprofen, and estriol had no effect on estradiol 3-glucuronidation at a substrate concentration of 5 μM. Information on the effects of morphine and silymarin at 5 μM were not available due to technical difficulties with the assays. For silymarin, estradiol 3-glucuronidation was 150% of control activity at 25 μM estradiol and 10 μM silymarin (Fig. 1). There is no other data point above 120% of control activity in this graph (Fig. 1). It is therefore not appropriate to classify this compound as a stimulator (heterotropic activator), since the data at 5 μM estradiol was incomplete for this compound, and the apparent “stimulation” arising from a sporadic single data point does not seem to follow a trend with regard to estradiol concentration, as observed for 17α-EE, for example (described above and shown in Fig. 2).

Subtle differences in flavonoid structure had significant effects on the ability of the flavonoid to modulate estradiol 3-glucuronidation (Table 1). The ring nomenclature for flavonoids is shown on flavone. A hydroxyl group at position 3 is indicative of a flavonol (e.g., quercetin, Table 1), whereas saturation of the carbon atoms between position 2 and 3 is indicative of a flavonone (Table 1). The most potent inhibitors of estradiol-3-glucuronidation at a concentration of 5 μM estradiol were the flavones, tangeretin (4′,5,6,7,8-pentamethoxyflavone, IC₅₀ = 1 μM) and chrysin (5,7-dihydroxyflavone, IC₅₀ < 10 μM). The presence of a methoxy group at the 3′ position appears to significantly reduce the inhibitory potency of flavones [e.g., nobiletin (3′, 4′,5,6,7,8-hexamethoxyflavone, IC₅₀ > 10 μM) compared with tangeretin]. Flavone, flavanone, silymarin, naringenin, and quercetin were considerably less potent than tangeretin or nobiletin in inhibiting estradiol 3-glucuronidation at a concentration of 5 μM estradiol (see Fig. 1 and Table 1).

Modulation of Estradiol-17-glucuronidation.

Figures 5 and6 show the effect of each of the flavonoids and nonflavonoid compounds, respectively over the indicated range of concentrations on estradiol 17-glucuronidation expressed as a percentage of control activity at each of the 5 estradiol concentrations examined. Hyperbolic Michaelis-Menten kinetics for the formation of estradiol 17-glucuronide were observed in each experiment yielding average V_max andK_m values (± standard deviations) for the 16 determinations of, respectively, 111 ± 26 pmols/mg/min and 7 ± 4 μM. 17α-EE was a competitive inhibitor of estradiol-17-glucuronidation (Ki = 20 μM) at the modulator concentrations tested. Estriol and naringenin were found to be noncompetitive inhibitors of estradiol 17-glucuronide and less potent inhibitors (Ki >50 μM) than 17α-EE. Interestingly, unlike the behavior observed for estradiol 3-glucuronidation, none of the examined compounds demonstrated a pattern of modulation consistent with stimulation of estradiol 17-glucuronidation (Figs. 5 and 6), although sporadic data points above 120% of control activity were observed in some cases (e.g., at 25 μM estradiol and 10 μM silymarin, and at 100 μM estradiol and 50 μM flavone).

• Download figure
• Open in new tab
• Download powerpoint

Figure 5

Estradiol-17-glucuronidation activity in the presence of flavonoids in a pooled mix of human liver microsomes expressed as a percentage of control activity.

Symbols indicate the following estradiol concentrations: ●, 5 μM; ○, 10 μM; ▾, 25 μM; ▿, 50 μM; ▪, 100 μM. Mean control activities ± standard deviation (in the absence of modulator) from 16 experiments in pmol/mg/min were 46 ± 19 at 5 μM estradiol, 63 ± 25 at 10 μM, 83 ± 24 at 25 μM, 98 ± 23 at 50 μM, and 100 ± 25 at 100 μM. The meanV_max (± standard deviation) value for 16 experiments was 111 ± 26 pmol/mg/min. The meanK_m value (± standard deviation) was 7 ± 5 μM.

Discussion

The results presented in this study demonstrate that, depending on effector and substrate concentrations, UGT1A1-catalyzed estradiol-3-glucuronidation activity is differentially sensitive to the effects of additional UGT1A1 substrates and other compounds. For those compounds (modulators) that stimulated estradiol 3-glucuronidation at estradiol concentrations below its S₅₀ value, the greatest stimulation was observed at the lowest modulator concentrations (Figs. 1 and 2). This was confirmed by a separate, second experimental design using 17α-EE as modulator (Fig. 4). At higher concentrations of both modulator and substrate, no effect or inhibition was observed (Figs. 1 and 2). Although this is the first observation of heterotropic activation for a UGT enzyme, the in vitro observation of crossover from stimulation to inhibition as modulator concentration increases has also been observed for other enzymes involved in endo- and xenobiotic metabolism [e.g., cytochrome P450 (Ekins et al., 1998; Maenpaa et al., 1998; Stresser et al., 2000)]. The demonstration of heterotropic activation of UGT1A1-catalyzed estradiol-3-glucuronidation in the current study is consistent with the previous studies (Fisher et al., 2000a,b) indicating homotropic activation. Interestingly, homotropic activation kinetics for estradiol 3-glucuronidation activity have also been demonstrated using microsomes expressing recombinant UGT1A1 in our laboratory, indicating that this is a true behavior of the enzyme (unpublished observations).

Several compounds in the study were observed to be inhibitors of estradiol 3-glucuronidation by UGT1A1. Flavonoids were the most potent inhibitors with tangeretin being the most potent inhibitor of estradiol 3-glucuronidation. Unlike naringenin and the other hydroxylated flavonoids, which provide hydroxy groups for glucuronidation and are therefore substrates for UGT1A1, tangeretin has methoxy groups at various positions. It therefore appears that the most potent inhibitor determined in this study of UGT1A1-catalyzed estradiol-3-glucuronidation, tangeretin, is not a substrate of the enzyme.

The Hill equation (see “Data Analysis” under Materials and Methods section) is a useful mathematical tool for describing the degree of sigmoidicity of the substrate concentration/enzyme activity relationship. It was originally proposed to provide an indication of the number of subunits in a multimeric enzyme that bound successive ligands in a cooperative manner. The Hill coefficient value for UGT1A1-catalyzed estradiol 3-glucuronidation in the current set of experiments was on average 1.9. There is evidence for UGT1A1 existing as a tetramer (Peters and Jansen, 1986) and another UGT, UGT2B1 existing as a dimer (Meech and Mackenzie, 1997). Therefore, it is tempting to speculate that UGT1A1 behaves as a cooperative ligand-binding multisubunit enzyme in microsomes with regard to UGT1A1-catalyzed estradiol 3-glucuronidation, thus explaining the activation (homotropic and heterotropic) of estradiol 3-glucuronidation.

There are alternative explanations for the cooperative behavior of UGT1A1 in human liver microsomes. One potential explanation is that allosteric effector sites exist on the UGT1A1 enzyme or that two substrate molecules occupy the UGT1A1 active site. For example, a kinetic model (Korzekwa et al., 1998) describing this binding and originally proposed to explain activation at low substrate concentrations and inhibition by effector at high substrate concentrations for CYP3A-catalyzed reactions (Ueng et al., 1997) adequately explains our data (results not shown). Therefore, should this proposal also hold for UGTs, these two sites may be within the UGT1A1 active site or on separate subunits, as described above. There is evidence of overlapping but distinct binding locations for 17α-EE and bilirubin in the UGT1A1 active site (Ciotti and Owens, 1996). However, there is no information to date on whether two estrogen-type molecules can simultaneously occupy the active site.

It may be argued that the observed heterotropic and homotropic activation of estradiol 3-glucuronidation is an artifact of the in vitro systems examined. The lumenal location of UGT enzymes in the endoplasmic reticulum presents a particular problem for investigating UGT-catalyzed enzyme reactions. Sonication and detergents have been used to overcome this problem but ultimately result in disruption of the natural microenvironment of the enzymes. The alamethicin treatment (used in the current study) of microsomes is not thought to be as disruptive but does result in the free transit of substrate, cofactors, and products to and from the UGT enzyme (Fisher et al., 2000a). Thus the assay conditions used in the current studies appear to provide an optimal system for examining the behavior of the UGT enzymes. Furthermore, the lack of homotropic or heterotropic activation of estradiol 17-glucuronidation over the same effector and substrate concentration ranges demonstrates that activation is not a generalized artifact of the incubation conditions. It is therefore appropriate to conclude that the observations of homotropic and heterotropic activation for estradiol-3-glucuronidation are true behaviors of UGT1A1 at least in vitro. However, further studies will need to be performed to evaluate whether this in vitro phenomenon also occurs in vivo.

Observed K_m/S₅₀values for UGT enzymes are often much higher than the plasma concentrations of substrates (Senafi et al., 1994). For example, plasma concentrations of estradiol are in the nanomolar range (Berrino et al., 2001), and the S₅₀ value for estradiol binding to UGT1A1 is 22 μM. If UGT1A1 behaves in vivo as it does in vitro, then at this low substrate concentration many compounds will be stimulators of estradiol 3-glucuronidation activity (Table 1). The findings of this study also raise an interesting point as to the ability of alamethicin-treated human liver microsomes to predict drug-drug interactions that result from the alteration of the catalytic activity of UGT1A1. It seems appropriate that for the most accurate prediction of drug-drug interactions, near-therapeutic concentrations of substrate (and inhibitor) should be used rather than the higher concentrations reflecting the substrate K_m or S₅₀ value. This study design may provide a more relevant observation of enzyme behavior, as a compound predicted to be an inhibitor from an experiment using the substrate atK_m concentrations may actually be a stimulator or have no effect at therapeutic concentrations in vivo.

Although it has not been possible to identify all of the enzymes that contribute to estradiol 17-glucuronidation in human liver microsomes, UGT2B7 is known to be a contributor (Gall et al., 1999). In the current study, three compounds were found to inhibit estradiol 17-glucuronidation. The competitive pattern of inhibition observed for estradiol 17-glucuronidation in human liver microsomes suggests that 17α-EE, estriol, and naringenin displace estrogen from the active site(s) of enzymes responsible for estradiol 17-glucuronidation. As already indicated, activation was not observed for this biotransformation. Thus it appears that Michaelis-Menten relationships between inhibitors and the substrate can be applied to the estradiol 17-glucuronidation results obtained in the current study.

In summary, evidence is provided for differential modulation, both activation and inhibition of estradiol-3-glucuronidation by other UGT1A1 substrates. For example, UGT1A1-catalyzed estradiol-3-glucuronidation is stimulated by 17α-EE at low substrate concentrations. In contrast, bilirubin is a weak competitive-type inhibitor. The most potent inhibitor of this activity at the lowest estradiol concentration tested is the polymethoxyflavone, tangeretin (IC₅₀ = 1 μM at 5 μM estradiol concentration). For compounds with similar structures to estrogens and anthraflavic acid, the stimulatory effects on estradiol 3-glucuronidation appear to be dominant over the inhibitory effects at low modulator concentrations, whereas at higher modulator concentrations, the inhibitory effects are dominant.