Concern that the reproductive health of humans is being affected by exposure to xenoestrogens has led to the development of various in vitro and in vivo screening assays for the identification of suspected xenoestrogens. However, the estrogenic activity of a chemical determined in vitro may not necessarily predict its activity in vivo if the chemical is metabolized during the assay and/or in vivo. Therefore, to investigate the role of metabolism in modulating the estrogenic activity of suspected xenoestrogens, we have devised a two-stage approach coupling incubations with either human or rat hepatic microsomes with a yeast estrogenicity (transcription) assay. We have assessed the activity of the proestrogenic pesticide 99.5% methoxychlor [1,1,1-trichloro-2,2-bis-(4-methoxyphenyl)ethane, MXC] (EC50 = 4.45 ± 1.9 μM, n = 6) and a structural analog, methoxybisphenol A [2,2-bis-(4-methoxyphenyl) propane, MBPA], in the yeast estrogenicity assay and also established that yeast (Saccharomyces cerevisiae), unlike human liver microsomes, are not able to demethylate MXC or MBPA to estrogenic metabolites. This indicates that the proestrogen MXC has weak intrinsic estrogenic activity. Using 99.5% MXC and 17β-estradiol as paradigms, we have demonstrated how metabolism can enhance or suppress, respectively, estrogenic activity. The effect of metabolism on the activities of the weak xenoestrogens 3,17β-bisdesoxyestradiol [1,3,5(10)-estratriene] and 6-hydroxytetralin (5,6,7,8-tetrahydro-2-naphthol) was also assessed. This two-stage approach can distinguish the estrogenic activity of a suspect chemical from the activity due to its more, or less, active metabolites and will aid in the evaluation of novel xenoestrogens and, more importantly, proestrogens.
Human exposure to xenoestrogens has been linked to disorders of the male reproductive tract, infertility, and to the increased incidences of certain cancers (Sharpe and Skakkebaek, 1993; Toppari et al., 1996;Ashby, 1997; Sonnenschein and Soto, 1998). Of particular concern are the proestrogens, because the majority of the current in vitro estrogenicity assays (receptor binding, cell proliferation, and yeast-based transcription assays) used to screen for suspect chemicals are likely to produce false-negative results for the prediction of estrogenic activity of such compounds in vivo due to a lack of metabolic capability.
The proestrogen methoxychlor [1,1,1-trichloro-2,2-bis-(4-methoxyphenyl)ethane; MXC] (Fig.1) is a broad-spectrum pesticide developed as a substitute for DDT. There is concern that the estrogenic metabolites of MXC may elicit toxicity to mammalian reproductive processes. In vivo studies have demonstrated that MXC can produce adverse effects in adult male mice such as blocking sexual arousal and lowering plasma testosterone levels following peri-implantation exposure (Amstislavsky et al., 1999), in addition to increasing prostate size as a consequence of low-dose fetal exposure (Welshons et al., 1999). MXC has also been shown to have adverse effects on fertility in female rats and in utero development (Cummings, 1997).
The metabolism of MXC has been well characterized in both human and rodent systems; the principal pathway being demethylation to mono-hydroxy-MXC [1,1,1-tri chloro-2-(4-hydroxyphenyl)-2-(4-methoxyphenyl)ethane, mono-OH-MXC] and bis-hydroxy-MXC [1,1,1-trichloro-2,2- bis-(4-hydroxyphenyl)ethane, bis-OH-MXC or HPTE] (Bulger et al., 1978, 1984; Dehal and Kupfer, 1994). In addition, rat liver microsomes catalyze ring hydroxylation resulting in the formation of a catechol, tris-hydroxy-MXC [1,1,1-trichloro-2-(4-hydroxyphenyl)-2-(3,4-dihydroxyphenyl)ethane] (Kupfer et al., 1990). Mono-OH-MXC and bis-OH-MXC are probably responsible for the estrogenic activity of MXC in vivo, confirmed by their ability and MXC's inability to bind to rat uterine estrogen receptor (ER) in vitro (Bulger et al., 1978; Ousterhout et al., 1981;Shelby et al., 1996). Bis-OH-MXC is the more potent metabolite, presumably due to the two hydroxyl groups enabling a tighter interaction with the ER ligand binding domain through hydrogen bond formation. Evaluation of the estrogenic activity of MXC in a yeast strain containing the human estrogen receptor (hER) revealed that 99% MXC displayed activity (Odum et al., 1997). However, it is possible that the response observed in this assay was a consequence of the estrogenic impurities in 99% MXC (Bulger et al., 1984). We have therefore determined the activity of purified MXC in the yeast estrogenicity assay.
Odum et al. (1997) postulated that the activity of MXC they observed resulted from the compound's metabolism by yeast to the diphenol metabolite. A recent study, using the same yeast assay, reported that 95% MXC produced a submaximal response (compared with bis-OH-MXC) after a 3-day incubation and a maximal response after a 4-day incubation, which was taken by Beresford et al. (2000) to be consistent with the slow conversion of MXC to an estrogen. We have investigated whether yeast are able to metabolize MXC under the conditions of the assay. Methoxybisphenol A [2,2-bis(4-methoxyphenyl)propane, MBPA] was also used to assess the metabolic competence of the yeast as it is structurally similar to MXC.
Assessing the metabolic competence of the yeast estrogenicity assay is important because the activity determined with this major assay may not reflect the activity in vivo (such as in the uterotrophic assay) if a test chemical is metabolically activated or inactivated. Using MXC,Bulger et al. (1984) developed an in vitro assay combining rat liver microsomes with uterine ER binding for the detection of proestrogens. However, this approach will only indicate whether metabolites are able to bind to ER, not whether they are agonists or antagonists at the receptor. We have therefore developed a two-stage approach coupling liver microsomal incubations with a yeast transcription assay to assess the role of metabolism in modulating the estrogenic activities of various test chemicals. E2 and 99.5% MXC were chosen as paradigm compounds to demonstrate how metabolic inactivation and activation, respectively, can affect estrogenic activity.
Materials and Methods
17β-Estradiol (E2), 2-hydroxyestradiol (2-OH-E2), mestranol, MXC (95%), and NADPH were obtained from Sigma-Aldrich (Poole, Dorset, UK). Bisphenol A (4,4′-isopropylidenediphenol, BPA) and 6-hydroxytetralin (5,6,7,8-tetrahydro-2-naphthol, 6-OH-tetralin) were supplied by Aldrich Chemical Co. (Gillingham, UK). Bis-OH-MXC (99%) was a gift from M. D. Shelby (National Institute of Environmental Health Sciences, Research Triangle Park, NC), originally supplied by CedraCorp (Austin, TX). MBPA was prepared by methylation of BPA using dimethyl sulfate in the standard manner. It was purified by chromatography using a Florisil column and eluting with chloroform. NMR, elemental analysis, and mass spectrometry confirmed its structure. 3,17β-Bisdesoxyestradiol [1,3,5(10)-estratriene, bisdesoxy-E2] was synthesized as described previously (Elsby et al., 2000). HPLC grade solvents were obtained from Sigma-Aldrich. All other chemicals were purchased from BDH (Poole, Dorset, UK).
Purification of Methoxychlor.
MXC (95%) was purified according to the method of Bulger et al. (1978). The base-washed MXC was eluted from an Ultracarb 5-μm C8 column with acetonitrile (50–70%, 15 min) in 0.1 M ammonium acetate (pH 6.9) at 1 ml/min. Peak fractions corresponding to MXC were collected, pooled, and extracted with ethyl acetate (4 ml × 2). Organic phases were evaporated to dryness under a stream of N2 at 40°C to give MXC. The purity of the MXC thus obtained was established by normal phase HPLC (Spherisorb CN column) and LC-MS, which indicated the presence of one peak of UV absorbance and the absence of an ion-current peak (m/z 316) corresponding to bis-OH-MXC, and estimated by NMR to be 99.5%. MXC that had been subjected to base washing and HPLC purification is referred to as 99.5% MXC. Bis-OH-MXC was not detectable in either 95 or 99.5% MXC (10 mM); bis-OH-MXC is approximately 50-fold more potent, indicating that an impurity of no less than 2% would have to be present to be responsible for the estrogenic activity observed with MXC. This would correspond to a concentration of 200 μM in a 10 mM MXC stock solution, which is far greater than the limit of detection for the HPLC (0.5 μM bis-OH-MXC) analytical method.
Immature female Wistar rats (21–25 days old) were obtained from a breeding colony maintained by the Biomedical Services Unit, University of Liverpool.
Histologically normal livers were obtained from four female Caucasian transplant donors (age, 35–65 years). The certified cause of death was traumatic injury due to road traffic accident. The livers were removed and transferred to the laboratory within 30 min of death. They were portioned, frozen in liquid nitrogen, and stored at −80°C. Approval was granted by the relevant ethical committees and prior consent was obtained from the donors' relatives.
Preparation of Microsomes.
Livers were removed from eight immature female Wistar rats immediately after they were killed by cervical dislocation, pooled, and homogenized in two volumes of ice-cold 67 mM potassium phosphate buffer (pH 7.5) containing 0.15 M potassium chloride. Samples (10–20 g) of the stored human livers were homogenized individually. Microsomal fractions were prepared by differential centrifugation according to the method of Gill et al. (1995). Microsomal protein concentrations were determined by the method of Lowry et al. (1951). Equal amounts of the four human liver microsomal preparations were homogenized together to obtain the preparation used for metabolism studies.
Incubations contained 1 mg of human microsomal protein, 10 mM MgCl2, and 1 mM substrate in 67 mM phosphate buffer (pH 7.5) to give a final volume of 1 ml. Substrate or NADPH was omitted from control incubations. Following preincubation at 37°C for 2 min in a shaking water bath, the reaction was initiated by the addition of NADPH (final concentration, 1 mM). After a 60-min incubation with a further addition of NADPH at 30 min, the reaction was terminated by the addition of methyl tert-butyl ether (4 ml). The organic phases of two 15-min extractions were pooled for each incubation and evaporated to dryness under a stream of N2 at 40°C. The residue was reconstituted in methanol (100 μl) for immediate analysis by HPLC or LC-MS. Aliquots (10 μl) of the methanol solutions were eluted from an Ultracarb 5-μm C8 column with methanol (50-70-80%; 0-15-16 min) in 0.1 M ammonium acetate (pH 6.9) at 1 ml/min for HPLC and 0.9 ml/min for LC-MS.
Kinetics of MXC Bis-Demethylation.
The initial rate was linear for time (5–15 and 5–30 min for rat and human, respectively) and protein concentration (0.1–2.0 and 0.25–2.0 mg for rat and human, respectively). Incubations were carried out as described above using human liver microsomes (0–20 μM MXC, 30 min, 1 mg of protein) or immature rat liver microsomes (0–500 μM MXC, 15 min, 1 mg of protein) in a final volume of 1 ml. The internal standard (mestranol, 4 μl of 1 mg/ml) was added following termination with methyltert-butyl ether (4 ml). After extraction and evaporation under N2, the residues were reconstituted in methanol (200 μl) for HPLC analysis. Aliquots (50 μl) of the methanol solutions were eluted from an Ultracarb 5-μm C8 column with acetonitrile (50–70%, 15 min) in 0.1 M ammonium acetate (pH 6.9) at 1 ml/min. Peak area measurements of absorbance (280 nm) were used to quantify bis-OH-MXC and were compared with standard solutions. ApparentK m andV max values were determined from demethylation activities (picomoles of bis-OH-MXC formed/min/mg of protein) using the Grafit package (Sigma, St. Louis, MO). Incubations were performed on four separate occasions in duplicate.
Yeast Estrogenicity Assay.
The estrogenic activity of E2, 2-OH-E2, 95% MXC, 99.5% MXC, bis-OH-MXC, MBPA, BPA, bisdesoxy-E2, and 6-OH-tetralin was determined by the recombinant estrogen receptor-reporter gene expression assay of Routledge and Sumpter (1996). Briefly, in this system, the DNA sequence of hERα is integrated into the genome of Saccharomyces cerevisiae, which also contains transfected expression plasmids comprising the yeast 3-phosphoglycerate kinase promoter, estrogen responsive sequences, and a β-galactosidase reporter gene (lac-Z). Upon binding an active ligand, the ER activates transcription of the reporter gene. Thus, β-galactosidase is secreted into the medium where it hydrolyzes the chromogenic substrate chlorophenol red-β-d-galactopyranoside, resulting in a color change from yellow to red that is measured spectrophotometrically (550 nm) after 3 days. The criterion for activity in the assay is a reproducible and statistically significant (Kruskal-Wallis multiple comparison test) dose-related increase in the absorbance of test wells compared with controls.
MXC (99.5%) was also assessed using a 4-day incubation period: the 96-well plates, containing test chemicals and yeast, were incubated at 32°C for the first 3 days, according to the method of Routledge and Sumpter (1996), and were then kept at room temperature for the final day, before being read spectrophotometrically. The reduction in incubation temperature allowed the suppression of the gradual increase in the background absorbance of the medium caused by additional constitutive expression of β-galactosidase by the yeast. This ensured a sufficient signal-to-noise ratio to distinguish between absorbance due to ligand activation and background absorbance.
The purpose of the fourth day of incubation was to establish whether a full dose-response curve (i.e., the response obtained with bis-OH-MXC) was achieved for 99.5% MXC on that day, as demonstrated previously for 95% MXC by Beresford et al. (2000), who interpreted the longer incubation time, required for MXC to produce a full dose-response curve (compared with bis-OH-MXC following a 3-day incubation), as an indication that the MXC was metabolized slowly to bis-OH-MXC.
Yeast Antiestrogenicity Assay.
The antiestrogen screen has been described previously (Routledge and Sumpter, 1997). Briefly, E2 was added to the yeast and growth medium at a concentration of 1 × 10−10 M. This concentration produced a submaximal response. Chemicals that were able to antagonize the activity of the natural ligand resulted in a concentration-dependent decrease in the absorbance of the medium.
Yeast Metabolism Studies.
Yeast [obtained by a 24-h culture of a 10-times concentrated stock (Routledge and Sumpter, 1996)] were incubated in growth medium (10 ml) containing either 1 mM 99.5% MXC or 1 mM MBPA. Substrate or yeast was omitted from controls. Following incubation at 32°C in a naturally ventilated incubator for 3 days, reaction mixtures were centrifuged at 2000 rpm for 10 min to separate growth medium and yeast. Yeast cells were lysed by the addition of distilled H2O (1 ml) followed by sonication. The lysed yeast and growth medium (2-ml aliquots) from each incubation were extracted separately with methyl tert-butyl ether (4 ml × 2). The combined organic phases were evaporated to dryness under a stream of N2 at 40°C, and the residue was reconstituted in methanol (100 μl) for analysis by HPLC or LC-MS. Aliquots (20 μl) were eluted from an Ultracarb 5-μm C8 column with methanol (50-70-80%, 0-15-16 min) in 0.1 M ammonium acetate (pH 6.9) at 1.0 ml/min for HPLC or 0.9 ml/min for LC-MS.
Coupled Microsomal Metabolism-Yeast Estrogenicity Assay.
Microsomal incubations contained 1 mg of human or rat liver microsomal protein, 10 mM MgCl2, and either 0 to 40 nM E2 or 0 to 3 mM 99.5% MXC in 67 mM phosphate buffer (pH 7.5) to give a final volume of 200 μl. Bisdesoxy-E2 (0–2 mM), 6-OH-tetralin (0–8 mM), or MBPA (0–3 mM) was incubated with human liver microsomes. Substrate or NADPH was omitted from control incubations. Following preincubation at 37°C for 2 min in a shaking water bath, the reaction was initiated by the addition of NADPH (1 mM). After 30 min the reaction was terminated by the addition of methyl tert-butyl ether (2 ml). This mixture was subject to rotation mixing for 15 min. The combined organic phases of two extractions were evaporated to dryness under N2 at 40°C and the residue was reconstituted in methanol (200 μl). Aliquots (10 μl) of the methanol solutions were incorporated into the yeast estrogenicity assay as described above. Estrogenic activities measured following incubations in the presence or absence of NADPH were compared by the test of Mann-Whitney. For HPLC analysis, aliquots (50 μl) were eluted from an Ultracarb 5-μm C8 column with acetonitrile (50–70%, 0–15 min, 99.5% MXC or MBPA; 50-70-80%, 0-12-13 min, bisdesoxy-E2; 30–50%, 0–12 min, 6-OH-tetralin) in 0.1 M ammonium acetate (pH 6.9) at 1 ml/min. High substrate concentrations were used to take into account the dilution factor in the yeast assay.
High Performance Liquid Chromatography.
HPLC was performed with an Ultracarb 5-μm C8 column (25 × 0.46 cm; Phenomenex, Macclesfield, Cheshire, UK) or a Spherisorb CN column (25 × 0.46 cm; Phenomenex) connected to a Spectra-Physics SP8800 ternary solvent delivery system, a Spectra-Physics analytical UV1000 UV detector (Spectra-Physics, San Jose, CA), and a Radiomatic A250 data system (Canberra-Packard, Pangbourne, Berks, UK). Metabolites were identified as chromatographic peaks of UV absorbance that were absent from control incubations (minus substrate or NADPH).
Liquid Chromatography-Mass Spectrometry.
A Quattro II tandem quadrupole instrument (Micromass Ltd., Manchester, UK) fitted with the standard LC-MS interface and electrospray source was used in the negative ion monitoring mode. The LC system consisted of two Jasco PU980 pumps (Jasco UK, Great Dunmow, Essex, UK) and a Jasco HG-980-30 mixing module. Analytes were resolved on an Ultracarb 5-μm C8 column (25 × 0.46 cm) with a gradient of methanol (50-70-80%, 0-15-16 min) in 0.1 M ammonium acetate, pH 6.9. The flow rate was 0.9 ml/min. Eluate split-flow to the LC-MS interface was ca. 40 μl/min. Nitrogen was used as the nebulizing and drying gas. The interface temperature was 70°C; the capillary voltage, 3.9 kV; the HV and RF lens voltage, 0.5 kV and 0.1 kV, respectively; and the photomultiplier voltage, 650 V. The mass spectrometer acquired spectra betweenm/z 100 to 1050 over a scan duration of 5 s. Data were processed with Masslynx 2.0 software (Micromass Ltd.).
Incubation of 95% MXC with human liver microsomes, in the presence of NADPH, yielded two metabolites with retention times of 20.0 and 23.7 min; 95% MXC had a retention time of 31.1 min (Fig. 2A). When analyzed by LC-MS, the more polar metabolite yielded m/z 316 corresponding to bis-OH-MXC ([M − 1]−). This metabolite coeluted with authentic bis-OH-MXC. The second metabolite gave m/z 330 corresponding to mono-OH-MXC.
LC-MS analysis of an incubation of MBPA with human liver microsomes revealed metabolites that yielded anions at m/z227 (Rt = 17.5 min) and 241 (Rt = 22.7 min) (Fig. 2B). The former coeluted with authentic BPA. The latter had a mass corresponding to desmethyl-MBPA. MBPA had a retention time of 29.9 min.
Determination of Bis-Demethylation Kinetics.
With human liver microsomes the mean V max for MXC bis-demethylation was 20.2 ± 0.5 pmol/min/mg of protein; the mean apparent K m being 0.804 ± 0.081 μM (n = 4) (Fig. 3A). This compared with a V max andK m of 101.2 ± 3.3 pmol/min/mg of protein and 5.943 ± 1.009 μM (n = 4), respectively, in immature female rat microsomes (Fig. 3B). Eadie-Hofstee plots indicated the involvement of more than one enzyme for the bis-demethylation of MXC in both human and rat liver microsomes.
Yeast Estrogenicity Assay.
E2, 95% MXC, 99.5% MXC, bis-OH-MXC, and BPA were active in the yeast assay, whereas MBPA was inactive (Fig. 4). The rank order of potency was E2 > bis-OH-MXC > BPA > 95% MXC > 99.5% MXC (Table1). Figure 5 shows the activity of E2, 2-OH-E2, bisdesoxy-E2, and 6-OH-tetralin. The activity of 99.5% MXC did not produce a full response even after a 4-day incubation (data not shown). Coincubation of MBPA (300 μM) with 99.5% MXC, over the concentration range shown in Fig. 4, abolished the activity of the latter (data not shown). In contrast, coincubation of MBPA (300 μM) with E2(9.8 × 10−13–2 × 10−9 M) caused a parallel shift to the right in the concentration-response curve of E2.
Yeast Antiestrogenicity Assay.
MXC (99.5%) produced a concentration-dependent increase in the submaximal response of E2. MBPA resulted in a concentration-dependent decrease in the submaximal response of E2. MBPA was antiestrogenic over the concentration range 75 to 300 μM (Fig.6).
Yeast Metabolism Studies.
Incubation of either 99.5% MXC (1 mM) or MBPA (1 mM) with yeast did not result in metabolism to bis-OH-MXC or BPA, respectively (Fig. 7). The greatest recovery of parent compound was present in extracts of lysed yeast cells compared with extracts of growth medium based on LC-MS with UV detection. Concentration of the analytes (100-fold) from these incubations ensured that any formation of bis-OH-MXC or BPA would be higher than the limit of sensitivity for the LC-MS method (100 μM). Absence of metabolism was confirmed by HPLC analysis, for which the limit of detection was 0.5 μM bis-OH-MXC or BPA.
Modulation of the Estrogenic Activity of Test Chemicals by Metabolism in Vitro.
The estrogenic activity of E2 was significantly reduced (approximately a 7-fold decrease in potency) following incubation with human liver microsomes in the presence of NADPH (Fig.8A). In contrast, the estrogenic activity was almost abolished by incubation with rat liver microsomes and NADPH (Fig. 8B).
Following incubation in the presence of NADPH with either human or rat liver microsomes, the estrogenic activity of 99.5% MXC was significantly enhanced; as demonstrated by the 20- to 30-fold shift in potency and the production of a maximal response (Fig. 8 ).As demonstrated in Fig. 2, 99.5% MXC was found to be metabolized to bis-OH-MXC and mono-OH-MXC by both human and rat liver microsomes. In addition, rat liver microsomes formed two unidentified metabolites.
There was no difference in the activity of MBPA in the yeast estrogenicity assay following incubation with human liver microsomes in the presence or absence of NADPH (data not shown). MBPA was metabolized to BPA and desmethyl-MBPA, as shown in Fig. 2.
The steroidal analog, bisdesoxy-E2, was active in the yeast assay following incubation with human liver microsomes in the absence of NADPH (Fig. 9A). Incubation of bisdesoxy-E2 with human liver microsomes in the presence of NADPH significantly increased the estrogenic activity (approximately an 80-fold increase in potency; Fig. 9A). HPLC analysis of incubations containing bisdesoxy-E2 and human liver microsomes, in the presence of NADPH, revealed the formation of one metabolite that corresponded to 17β-desoxyestradiol.
There was no significant difference in the estrogenic activity of 6-OH-tetralin following incubation with human liver microsomes in either the presence or absence of NADPH (Fig. 9B). Analysis of incubations containing 6-OH-tetralin and human liver microsomes, in the presence of NADPH, did not indicate the formation of metabolites.
Xenoestrogens encompass a diverse range of chemicals thought to be responsible for the increase in adverse effects on reproductive health and other endocrine disruption in both humans and wildlife (Toppari et al., 1996; Sonnenschein and Soto, 1998; Tyler et al., 1998). A variety of in vitro and in vivo assays have been developed to aid in the risk assessment of suspected xenoestrogens (Routledge and Sumpter, 1996;Odum et al., 1997; Jobling, 1998). Competitive receptor binding, cell proliferation, and yeast-based transcription assays are examples of commonly used in vitro methods. In vivo assays, such as the rodent uterotrophic assay, produce a more accurate estimate of a chemicals' estrogenic activity since they allow the evaluation of numerous biological end points, in addition to enabling the chemical to act on an intact reproductive system while being subject to metabolism and excretion. However, they are often expensive and time-consuming. Discrepancies between in vitro and in vivo assays can occur when the former do not have the capability to metabolize test compounds (Elsby et al., 2000). This could result in false negatives or false positives for the prediction of estrogenic activity in vivo if an inert compound or estrogen was metabolically activated or inactivated, respectively, in mammals. In addition, assessment of weak xenoestrogens using these in vitro assays may underestimate activity in vivo if such compounds act as sources of more potent estrogenic metabolites. The combination of liver microsomal incubations with a yeast estrogenicity assay, using two paradigm compounds, introduces metabolism into a functional in vitro screening assay.
The proestrogen MXC was chosen to represent the enhancing effects of metabolism on estrogenic activity. However, it has been shown that MXC can contain estrogenic impurities (Bulger et al., 1984) and this has been confirmed in the present study by the maximal response produced by 95% MXC compared with the submaximal response of the purified 99.5% MXC in the yeast estrogenicity assay.
It has been hypothesized that the activity of MXC in the assay is consistent with metabolic conversion to bis-OH-MXC (Odum et al., 1997;Beresford et al., 2000). Yeast are able to metabolize a variety of foreign compounds (Kärenlampi et al., 1982; Yoshida and Aoyama, 1984; Miller et al., 1986) due to the presence of cytochrome P450 (Kärenlampi et al., 1980). Indeed, the major P450 isozyme present in yeast (S. cerevisiae) is responsible for the 14α-demethylation of the sterol lanosterol (Yoshida and Aoyama, 1984). Therefore, the metabolic capability of the relevant yeast strain was investigated. The inability of the transformed yeast to metabolize 99.5% MXC in the conditions of the assay, and the absence of a maximal response from the 4-day assay, indicates that 99.5% MXC has weak intrinsic estrogenicity. Presumably, the responses seen with MXC in previous studies (Odum et al., 1997; Beresford et al., 2000) were a consequence of this activity, in addition to the presence of impurities, and not a result of metabolism.
Since MBPA, like MXC, was sequentially demethylated by human liver microsomes to the mono-OH and bis-OH metabolites, the yeast might have metabolized MBPA to BPA and produced an estrogenic response. The lack of activity of MBPA in the estrogenicity assay might have been due to an inability to enter the yeast cell, because the yeast cell wall can be a barrier to some chemicals, resulting in false negatives for estrogenic activity in the yeast (Sonnenschein and Soto, 1998). However, in yeast metabolism studies, by far the greatest proportion of MBPA, like 99.5% MXC, was observed in extracts of lysed yeast cells compared with the minor amounts seen in growth medium, thus indicating that the test chemicals were able to cross the cell wall. In fact MBPA was antiestrogenic, as demonstrated by both the antiestrogenicity assay and competitive reversal of antagonism by E2, so it was possible that the compound's lack of activity in the yeast estrogenicity assay derived from low metabolic turnover to BPA, resulting in the antiestrogenicity masking the metabolite's estrogenicity. Absence of metabolism of MBPA by yeast indicated that the inactivity in the estrogenicity assay was due solely to the compound's antiestrogenic properties. This emphasizes the possibility of false negatives for predicting antiestrogenic activity unless a test chemical deemed nonestrogenic in the estrogenicity assay is assessed for potential antiestrogenicity.
In the coupled microsomal metabolism-yeast estrogenicity assay immature female rat liver microsomes were used because immature female rats are used in the rodent uterotrophic assay (Odum et al., 1997). This approach can provide an indication of a suspect chemical's activity in the uterotrophic assay.
The estrogenic activity observed with the lowest concentration of 99.5% MXC incubated with human liver microsomes is greater than that from an incubation with rat liver microsomes. Even though the apparentV max for MXC bis-demethylation in rat is approximately 5-fold greater than that in human liver microsomes, maximal activity isn't reached until higher substrate concentrations, whereas the apparent V max has already been achieved with human liver microsomes at this concentration. Therefore, at this substrate concentration, the amount of bis-OH-MXC formed by rat liver microsomes is lower than that formed by human liver microsomes. When the apparent V max for bis-OH-MXC formation has been achieved, the maximal estrogenic response produced with increasing 99.5% MXC concentrations will result partly from the intrinsic activity of the parent compound but is mainly due to the increased formation of the mono-OH-MXC metabolite. The formation of mono-OH-MXC and bis-OH-MXC by both human and rat liver microsomes is in agreement with Dehal and Kupfer (1994). The unidentified metabolites from incubations of 99.5% MXC with rat liver microsomes are most likely ring-hydroxylated derivatives of MXC and mono-OH-MXC based upon retention time and the proposed pathway of metabolism of MXC by Dehal and Kupfer (1994). However, it is not known whether these metabolites contribute to the estrogenic activity observed in the yeast assay.
The endogenous estrogen, E2, was used to demonstrate the effects of metabolism on suppressing estrogenicity. One of the major hepatic pathways for E2 metabolism in both rat and human is 2-hydroxylation resulting in the formation of the catecholestrogen 2-OH-E2 (Kerlan et al., 1992; Suchar et al., 1995; Zhu and Conney, 1998). 2-OH-E2 is approximately 100-fold less potent than E2 in the yeast estrogenicity assay. Comparison of the yeast data for authentic 2-OH-E2 with that obtained following microsomal incubations of E2, in the presence of NADPH, suggests that the reduced estrogenic activity observed following metabolism is due to the loss of E2 through conversion to 2-OH-E2; the amount of 2-OH-E2 formed even at the highest substrate concentration would not be sufficient to produce an estrogenic response. Therefore, the difference in the estrogenic activity of E2 following metabolism by rat liver microsomes compared with human is likely a consequence of its greater conversion to the catecholestrogen.
MBPA is an antiestrogen that undergoes metabolic activation to estrogenic metabolites. However, metabolic turnover by human liver microsomes was not sufficient to enable the metabolites' estrogenicity to counteract the antiestrogenicity of MBPA in the yeast estrogenicity assay. Bisdesoxy-E2 is a weak xenoestrogen that has been shown to be hydroxylated by immature female rat liver microsomes to the more potent 17β-desoxyestradiol (Elsby et al., 2000). This pathway of metabolic activation has been confirmed in this study using human liver microsomes coupled with the yeast estrogenicity assay. In contrast, 6-OH-tetralin is a very weak xenoestrogen whose activity does not appear to be modulated by metabolism. It might undergo hydroxylation to a catechol but such a reaction is evidently too slow to have any impact on the estrogenic activity of the parent compound as determined in the yeast assay.
It is important to acknowledge that the extent of an estrogenic response through this two-stage approach will depend on the rate of metabolic turnover for a given compound. Additionally, there was no direct microsomal affect on estrogenic response: extracts from incubations in the absence of substrate and NADPH did not enhance or suppress basal absorbance of the medium. When considering this linked assay system, it should be borne in mind that modulation of xenoestrogens' activity by hepatic metabolism may differ from that effected by metabolism in reproductive “target” organs/tissues, such as the uterus. In addition, the effect of sex on the metabolism of xenoestrogens, and its relationship to estrogenic activity, needs to be assessed.
The findings of the present study demonstrate that a combination of human liver microsomes and a yeast estrogenicity assay can be a useful initial screen for assessing the effects of metabolism on the activity of suspected xenoestrogens and, in particular, proestrogens. Importantly, for human risk assessment, this system will help to overcome the problems associated with the generation of false positives or false negatives from animal studies by use of quantitative in vitro bridging studies.
We thank Dr. M. D. Shelby for the gift of bis-OH-MXC.
- Received July 5, 2000.
- Accepted October 5, 2000.
Send reprint requests to: Professor B. K. Park, Department of Pharmacology and Therapeutics, University of Liverpool, New Medical Bldg., Ashton St., Liverpool L69 3BX, UK. E-mail:
This work was supported by a collaborative studentship between the Medical Research Council and AstraZeneca Central Toxicology Laboratory (to R.E.). B.K.P. is a Wellcome Principal Fellow. The LC-MS system was purchased and maintained with grants from the Wellcome Trust.
- estrogen receptor
- methoxybisphenol A
- bisphenol A
- high performance liquid chromatography
- liquid chromatography-mass spectrometry
- retention time
- The American Society for Pharmacology and Experimental Therapeutics