Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Bisphenol A [BPA, 2,2-bis(4-hydroxyphenyl)propane; Fig. 1] is used as a monomer in the manufacture of polycarbonate plastics, epoxy resins, and composites, and consequently, has extensive applications in the food-packaging industry and in dentistry. BPA has weak estrogenic activity in vitro (Krishnan et al., 1993; Nagel et al., 1997). Although its structure is distinct from that of 17-estradiol (E₂), its ability to bind to the estrogen receptor (ER) might be rationalized if the two phenol rings mimicked the A- and D-rings of E₂ within the ligand binding domain of ER (Waller et al., 1996).

View larger version (18K): [in this window] [in a new window]   Fig. 1.   Structures of BPA and metabolites.

The estrogenic activity of BPA has been assessed by a variety of in vitro assays, including ER binding (Kuiper et al., 1997), yeast reporter-gene expression assays (Beresford et al., 2000), proliferation of MCF-7 human breast cancer cells (Perez et al., 1998), and induction of progesterone receptors in both human MCF-7 cells (Krishnan et al., 1993) and endometrial carcinoma cells (Bergeron et al., 1999). Studies in vivo have shown that BPA can mimic E₂ in stimulating prolactin secretion (Steinmetz et al., 1997), inducing growth, differentiation, and c-fos gene expression in the female rat reproductive tract (Steinmetz et al., 1998) and exhibiting uterotrophic activity in both rats (Ashby and Tinwell, 1998; Laws et al., 2000) and mice (Papaconstantinou et al., 2000).

There is concern that the estrogenicity of BPA may elicit toxicity to mammalian developmental and reproductive processes. BPA can effect early development of preimplantation mouse embryos (Takai et al., 2000), in addition to increasing prostate size as a consequence of low-dose fetal exposure (Nagel et al., 1997). Exposure to environmentally relevant doses of BPA has been shown to advance puberty and alter postnatal growth rate in mice (Howdeshell et al., 1999). In contrast, other studies using the same levels of fetal BPA exposure and the same mouse strain as Nagel et al. (1997) did not observe any effect on the prostate gland (Ashby et al., 1999; Cagen et al., 1999a). Also, male offspring from pregnant Wistar rats exposed to BPA in drinking water had normal reproductive organ development (Cagen et al., 1999b).

Metabolism can play an important role in modulating the estrogenic activity of xenoestrogens in vivo (Elsby et al., 2000). The metabolism of BPA has been well characterized in the rat with the major metabolite being the monoglucuronide (BPA glucuronide). BPA glucuronide constituted approximately 28% of the radioactivity found in urine (Knaak and Sullivan, 1966) and 68 to 100% of the plasma radioactivity (Pottenger et al., 2000) following oral administration of ¹⁴C-labeled BPA. Glucuronidation of BPA by rat liver microsomes is mainly catalyzed by the UDP-glucuronosyltransferase (UGT) isoform UGT2B1 (Yokota et al., 1999). Knaak and Sullivan (1966) also identified 5-hydroxybisphenol A [5- OHBPA, 2-(4,5-dihydroxyphenyl)-2-(4-hydroxyphenyl)propane], which has been postulated to be formed by rat liver microsomes (Atkinson and Roy, 1995). To date, the human metabolism of BPA, the estrogenic activity of 5-OHBPA (if any), and the modulating effect of metabolism on the estrogenicity of BPA have not been determined.

In a preliminary study BPA glucuronide was shown not to be estrogenic in both an ER binding assay and an MCF-7 cell gene expression assay (Matthews and Zacharewski, 1999). Therefore, when extrapolating the effects of BPA observed in experimental systems to effects likely to occur in humans an important point is the extent to which humans and rodents differ in their glucuronidation capacity. The present article explores this question with the aid of rat and human liver microsomes coupled with a yeast human ER reporter-gene expression assay (Elsby et al., 2001). In addition, we have assessed the estrogenic activity of the metabolite 5-OHBPA and the effects of 5-hydroxylation on the estrogenicity of BPA in the linked assay.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Chemicals. E₂, H-2 -glucuronidase-arylsulfohydrolase (from Helix pomatia), uridine diphosphate glucuronic acid (UDPGA), and NADPH were obtained from Sigma-Aldrich (Poole, Dorset, UK). Bisphenol A (4,4'-isopropylidenediphenol) and potassium nitrosodisulfonate (Fremy's salt) were supplied by Aldrich Chemical Co. (Gillingham, UK). Bis-hydroxy-methoxychlor [1,1,1-trichloro-2,2-bis-(4-hydroxyphenyl)ethane, bis-OH-MXC (99%)] was a gift from Dr. M. D. Shelby (National Institute of Environmental Health Sciences, Research Triangle Park, NC), originally supplied by CedraCorp (Austin, TX). HPLC-grade solvents were obtained from Sigma-Aldrich. All other chemicals were purchased from BDH (Poole, Dorset, UK).

Synthesis of 5-Hydroxybisphenol A. The method of Atkinson and Roy (1995) for the synthesis of bisphenol quinone was modified for the synthesis of 5-OHBPA. BPA (100 mg) was dissolved in acetone (33 ml) and 10% (v/v) glacial acetic acid (33 ml) was added. Potassium nitrosodisulfonate (333 mg) was added and the reaction mixture was shaken vigorously for 15 min. An additional quantity of potassium nitrosodisulfonate (333 mg) was added and the mixture was again shaken for 15 min. This process was repeated a further three times, and 10% (v/v) glacial acetic acid (33 ml) was added in each case to dissolve the Fremy's salt. The final mixture was extracted twice with 50 ml of chloroform and the extracts were washed with 1 M HCl and with H₂O and evaporated to dryness on a rotary evaporator. The oily residue was reconstituted in hexane and purified by elution from a semipreparative silica column with ethyl acetate (40%) in petroleum ether. Fractions containing catechol were further purified by reversed phase HPLC; aliquots were eluted from an Ultracarb 5-µm C8 column with acetonitrile (30-50%, 0-10 min) in 0.1 M ammonium acetate (pH 6.9) at 1 ml/min. Peak fractions corresponding to 5-OHBPA were collected, pooled, and extracted with methyl tert-butyl ether (5 ml × 2). Organic phases were evaporated to dryness under a stream of N₂ at 40°C to give the catechol. The purity of the 5-OHBPA thus obtained was established by NMR, HPLC (Ultracarb 5 µm C-8 column) and LC-MS, which indicated the presence of one major peak of UV absorbance and the corresponding ion-current peak (m/z 243), and estimated to be 97%.

Animals. Adult and immature (21-25-day-old) female Wistar rats were obtained from a breeding colony maintained by the Biomedical Services Unit, University of Liverpool.

Isolation of Rat Hepatocytes. Hepatocytes were isolated from whole livers of adult female Wistar rats (200 g) by a two-step collagenase perfusion technique (Tettey et al., 1999). The viability of the cell suspension (typically 87%) was determined by trypan blue exclusion.

Incubations of BPA with Rat Hepatocytes in Suspension. Freshly isolated hepatocytes (12 × 10⁶ viable cells/ml) were incubated with either dimethyl sulfoxide (5 µl; drug-free control) or BPA (final concentration 100 or 500 µM in 5 µl of dimethyl sulfoxide) in freshly prepared Krebs-Henseleit buffer (pH 7.6), in rotating 50-ml round-bottomed flasks at 37°C under an atmosphere of O₂,CO₂ (95:5 v/v). The total volume was 5 ml. After a 2-h incubation the reactions were terminated by the addition of ice-cold acetonitrile (15 ml) and left on ice for 15 min to precipitate protein. Following centrifugation at 3000 rpm, the supernatant was removed and evaporated to dryness under N₂ at 40°C. The extraction was repeated twice and the extracts were reconstituted in methanol (200 µl) for analysis by LC-MS. Aliquots of the methanol extracts (20 µl) were eluted from an Ultracarb 5-µm C8 column with methanol (50-70-80%; 0-15-16 min) in 20 mM ammonium acetate (pH 6.5) at 0.9 ml/min. For confirmation of glucuronidation and sulfonation, extracts (100 µl) were hydrolyzed, after diluting with sodium acetate buffer (0.1 M, pH 5.0), with H-2 -glucuronidase-arylsulfohydrolase (15 µl) at 37°C for 3 h and subsequently analyzed by LC-MS.

Human Livers. Histologically normal livers were obtained from four male (age, 24-57 years) and four female (35-65 years) Caucasian transplant donors. The certified cause of death was in each case traumatic injury due to a 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 male or four female human liver microsomal preparations were homogenized together to obtain the preparation used for metabolism studies.

BPA Glucuronidation. Incubations were carried out in capped 1.5-ml Eppendorf tubes using either human liver microsomes (0-1000 µM BPA, 30 min, 500 µg of protein) or immature rat liver microsomes (0-1000 µM BPA, 10 min, 50 µg of protein) in 50 mM Tris-HCl buffer (pH 7.5; final volume 200 µl) containing 10 mM MgCl₂ and Brij 58 surfactant (0.1 mg/ml). Glucuronidation activity was slightly lower in the absence of Brij 58. Substrate or UDPGA was omitted from control incubations. Glucuronide conjugation was confirmed by hydrolysis with H-2 -glucuronidase-arylsulfohydrolase at 37°C for 3 h. Following preincubation at 37°C for 2 min in a shaking water bath, the reaction was initiated by the addition of UDPGA in Tris-HCl buffer (final concentration, 3 mM). Reactions were terminated by the addition of ice-cold methanol (600 µl) and the internal standard (bis-OH-MXC, 4 µl of 1 mg/ml methanol solution) was added. After leaving the mixture on ice for 15 min and centrifuging at 10,000 rpm, the combined organic phases of two extractions of the protein pellet with methanol were evaporated to dryness under N₂ at 40°C and the residue reconstituted in methanol (100 µl) for immediate analysis by HPLC. Aliquots (50 µl) of the methanol solutions were eluted from an Ultracarb 5-µm C8 column with methanol (50-70-80%, 0-15-16 min) in 20 mM ammonium acetate (pH 6.5) at 1 ml/min. Peak area measurements of absorbance (280 nm) were used to quantify BPA glucuronide and were compared with standard solutions of BPA (0.5-100.0 µM). The initial rate was linear for time and protein concentration. The reaction did not conform to Michaelis-Menten kinetics; thus, the apparent V_max and K_m values, determined from glucuronidation activities (nmol of BPA glucuronide formed/min/mg of protein), were characterized as the maximal activity from the activity versus substrate concentration curve and the concentration required to produce 50% of this activity, respectively. Incubations were performed on four separate occasions in duplicate.

Microsomal BPA Oxidation. Incubations contained 1 mg of human or rat microsomal protein, 10 mM MgCl₂, and 200 µM BPA 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. Extracts of 15 individual incubations were pooled and evaporated to dryness under a stream of N₂ at 40°C for each incubation condition. The residue was reconstituted in methanol (100 µl) for immediate analysis by LC-MS. Aliquots (50 µl) of the methanol solutions were eluted from an Ultracarb 5-µm C8 column with methanol (50-70%; 0-15 min) in 20 mM ammonium acetate (pH 6.5) at 0.9 ml/min.

Yeast Estrogenicity Assay. The estrogenic activity of E₂, BPA, and 5-OHBPA was determined by the yeast assay of Routledge and Sumpter (1996). Briefly, in this system, the DNA sequence of hER 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 -galactosidase reporter gene (lac-Z). Upon binding an active ligand, the ER activates transcription of the reporter gene. Thus, -galactosidase is secreted into the medium where it hydrolyzes the chromogenic substrate chlorophenol red--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.

Coupled Microsomal Metabolism-Yeast Estrogenicity Assay. The coupled microsomal metabolism-yeast estrogenicity assay was carried out as described previously (Elsby et al., 2001). Microsomal incubations contained 0.5 mg of human or rat liver microsomal protein, 10 mM MgCl₂, 1 mM ascorbic acid, and 0 to 4 mM BPA in 67 mM phosphate buffer (pH 7.5) (for NADPH-mediated metabolism) or 10 mM MgCl₂ and 0 to 4 mM BPA in 50 mM Tris-HCl buffer (pH 7.5) (for UDPGA-mediated metabolism) to give a final volume of 200 µl. High substrate concentrations were used to take into account the dilution factor in the yeast assay. Substrate or cofactor was omitted from the controls. Following preincubation at 37°C for 2 min, the reaction was initiated by the addition of either NADPH (1 mM) or UDPGA (3 mM). After 45 min, the oxidation incubations were terminated with methyl tert-butyl ether (2 ml). The combined organic phases of two extractions were evaporated to dryness and reconstituted in methanol (200 µl). The glucuronidation incubations were terminated after 30 min with ice-cold methanol (600 µl) and the extracts reconstituted in methanol. Aliquots (10 µl) of the methanol solutions were incorporated into the yeast estrogenicity assay. For HPLC analysis, aliquots (20-50 µl) were eluted from an Ultracarb 5-µm C8 column with methanol (50-70-80%, 0-15-16 min) in 20 mM ammonium acetate (pH 6.5) at 1 ml/min.

High Performance Liquid Chromatography. HPLC was performed with an Ultracarb 5-µm C8 column (25 × 0.46 cm; Phenomenex, Macclesfield, Cheshire, UK) connected to a Dionex ASI-100 automated sample injector [Dionex (UK) Ltd., Macclesfield, Cheshire, UK], a Dionex P580 pump, and a Dionex UVD170S UV detector. Data was processed by Chromeleon software [Dionex (UK) Ltd.]. Metabolites were identified as chromatographic peaks of UV absorbance that were absent from control incubations (minus substrate or cofactor).

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; Phenomenex) with a gradient of methanol (50-70-80%, 0-15-16 min) in 20 mM ammonium acetate, pH 6.5. 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 high voltage and radio frequency lens voltage 0.5 and 0.1 kV, respectively; and the photomultiplier voltage 650 V. The mass spectrometer acquired spectra between m/z 100 to 1050 over a scan duration of 5 s or via selected-ion monitoring (SIM) (dwell time of 200 ms; interchannel delay of 20 ms). Data were processed with MassLynx 2.0 software (Micromass Ltd.).

Statistical Analysis. Data are expressed as the mean ± S.D. of four separate experiments performed in duplicate. For kinetic constants the data are expressed as the mean ± S.E.M. Comparison of positive and control wells in the yeast estrogenicity assay was made using the Kruskal-Wallis multiple comparison test. Results from the coupled microsomal metabolism-yeast estrogenicity assay were compared by the Mann-Whitney test and, for the EC₅₀ values determined, the unpaired t test. The maximal rate of glucuronidation by human and rat liver microsomes was compared by the Mann-Whitney test. Statistical significance was taken at p values <0.05.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Metabolism of BPA by Rat Hepatocytes. Incubation of BPA (500 µM) with female rat hepatocytes yielded one major and two minor metabolites with retention times of 6.8 and 7.5 and 8.7 min, respectively; unmetabolized BPA had a retention time of 17.5 min (Fig. 2). When analyzed by LC-MS, the most polar metabolite yielded m/z 403, corresponding to the monoglucuronide of BPA ([M  1]). The minor metabolites gave m/z 323 (R_t = 7.5 min) and m/z 307 (R_t = 8.7 min) corresponding to 5-OHBPA sulfate and BPA sulfate, respectively. BPA glucuronide was the only metabolite formed during incubation of 100 µM BPA with hepatocytes. Enzymatic hydrolysis of extracts from hepatocyte incubations resulted in the disappearance of chromatographic peaks corresponding to 5-OHBPA sulfate and BPA sulfate and a large reduction in the peak corresponding to BPA glucuronide (complete hydrolysis was observed for incubations containing 100 µM substrate). This was accompanied by an increase in the absorbance of BPA and the appearance of a peak with a similar retention time to authentic 5-OHBPA (R_t = 13.1 min) (Fig. 2). Conjugation was also confirmed by LC-MS fragmentation. BPA glucuronide (Fig. 3) fragmented by the loss of the glucuronic acid moiety to give the aglycone [M  1] (m/z 227). BPA sulfate fragmentation resulted in two peaks of m/z 227 and m/z 212 corresponding to BPA and demethylated aglycone [227  CH₃], respectively (Fig. 3). Fragmentation of 5-OHBPA sulfate (Fig. 3) resulted in loss of SO₃ to give the aglycone [M  1] (m/z 243).

View larger version (26K): [in this window] [in a new window]   Fig. 2.   HPLC separation (methanol-ammonium acetate) with electrospray MS detection of the metabolites from an incubation of BPA (500 µM) with female rat hepatocytes. The SIM chromatograms represent [M  1] ions with masses corresponding to BPA glucuronide (m/z 403), 5-OHBPA sulfate (m/z 323), and BPA sulfate (m/z 307). Metabolites were characterized as ion-chromatographic peaks that were absent from control incubations. The absorbance chromatogram represents the liberation of BPA and 5-OHBPA from these conjugates by -glucuronidase/sulfatase.

View larger version (24K): [in this window] [in a new window]   Fig. 3.   Electrospray spectra of BPA glucuronide (A), BPA sulfate (B), and 5-OHBPA sulfate (C) metabolites from an incubation of BPA (500 µM) with female rat hepatocytes. The position of conjugation for 5-OHBPA sulfate is not known.

BPA Glucuronidation Kinetics. BPA glucuronidation was assayed by HPLC (Fig. 4A). With pooled male and female human liver microsomes the mean V_max for BPA glucuronidation was 5.9 ± 0.4 and 5.2 ± 0.3 nmol/min/mg of protein, respectively; the mean apparent K_m being 77.5 ± 8.3 and 66.3 ± 7.5 µM, respectively (Fig. 4B). This compared with a V_max and K_m of 31.6 ± 8.1 nmol/min/mg of protein and 27.0 ± 1.2 µM, respectively, in immature female rat microsomes (Fig. 4). There was a significant difference (p < 0.05) between the V_max of BPA glucuronidation by human and immature female rat liver microsomes.

View larger version (18K): [in this window] [in a new window]   Fig. 4.   A, HPLC separation (methanol-ammonium acetate) of the metabolite from an incubation of BPA with either pooled male or female human liver microsomes, or immature female rat liver microsomes in the presence of UDPGA. B, activity versus substrate-concentration curves for the formation of BPA glucuronide (nmol/min/mg of protein) in incubations of BPA with either pooled male or female human liver microsomes (500 µg of protein, 30 min) or immature female rat liver microsomes (50 µg of protein, 10 min).

Microsomal Oxidative Metabolism. Incubation of BPA with female human or rat liver microsomes, in the presence of NADPH, yielded one metabolite (Fig. 5A), which coeluted with authentic 5-OHBPA and yielded an identical mass spectrum (Fig. 5B).

View larger version (18K): [in this window] [in a new window]   Fig. 5.   A, HPLC separation (methanol-ammonium acetate) with electrospray MS detection of the analyte from an incubation of BPA (200 µM) with either pooled female human or immature female rat liver microsomes, in the presence of NADPH. The SIM chromatograms represent [M  1] for ions with masses corresponding to 5-OHBPA (m/z 243) or BPA (m/z 227). Metabolites were characterized as ion-chromatographic peaks that were absent from control incubations. B, fragmentation spectrum obtained for the catechol metabolite of BPA following incubations with either pooled female or immature female rat liver microsomes, in the presence of NADPH.

Yeast Estrogenicity Assay. E₂, BPA, and 5-OHBPA were active in the yeast assay (Fig. 6). The rank order of EC₅₀ was E₂ (4.3 × 10¹¹ ± 1.8 × 10¹¹ M) > BPA (7.8 × 10⁷ ± 1.4 × 10⁷ M) > 5-OHBPA (6.1 × 10⁶ ± 1.9 × 10⁶ M).

View larger version (16K): [in this window] [in a new window]   Fig. 6.   Response of the yeast estrogenicity assay to BPA and its metabolite 5-OHBPA. Data are expressed as means ± S.D. (n = 4) and are compared by the Kruskal-Wallis multiple comparison test.

Modulation of the Estrogenic Activity of BPA by Metabolism in Vitro. The estrogenic activity of BPA was significantly reduced following incubation with human liver microsomes in the presence of UDPGA (Fig. 7A). There was approximately a 3-fold decrease in activity (EC₅₀ = 2.7 ± 0.6 and 6.5 ± 0.9 µM, in the absence and presence of UDPGA, respectively; p < 0.001). In contrast, the estrogenic activity was decreased approximately 7-fold by incubation with immature rat liver microsomes and UDPGA (EC₅₀ = 2.8 ± 0.5 and 15.8 ± 1.0 µM, in the absence and presence of UDPGA, respectively; p < 0.001) (Fig. 7B). There was a significant difference (p < 0.001) between the reduction of estrogenicity by human female and immature female rat liver microsomes. HPLC analysis of the incubations with either human or rat liver microsomes confirmed the formation of BPA glucuronide.

View larger version (19K): [in this window] [in a new window]   Fig. 7.   Response of the yeast estrogenicity assay to analytes from incubations of BPA with pooled female human liver microsomes (A) and immature female rat liver microsomes (B), in the presence or absence of UDPGA. Data are expressed as means ± S.D. (n = 4) and activities in the presence or absence of UDPGA are compared by the test of Mann-Whitney.

View larger version (19K): [in this window] [in a new window]   Fig. 8.   Response of the yeast estrogenicity assay to analytes from incubations of BPA with pooled female human liver microsomes (A) and immature female rat liver microsomes (B), in the presence or absence of NADPH. Data are expressed as means ± S.D. (n = 4) and activities in the presence or absence of NADPH are compared by the test of Mann-Whitney.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The results of the present study have demonstrated that metabolism can significantly modulate the estrogenic activity of the xenoestrogen BPA. We have shown that the physiologically inactive (Matthews and Zacharewski, 1999) BPA glucuronide is the major metabolite of BPA in rat hepatocytes, which is in agreement with previous studies in vitro and in vivo (Knaak and Sullivan, 1966; Nakagawa and Tayama, 2000; Pottenger et al., 2000). However, the metabolism of BPA in humans or human cell fractions has not been described. Therefore, the kinetics of BPA glucuronidation was determined in both male and female human liver microsomes and immature female rat liver microsomes. Microsomes from immature female rats were used because these rats are used in the rodent uterotrophic assay for predicting estrogenicity in humans (Odum et al., 1997).

There was no sex difference in the glucuronidation of BPA in human liver microsomes, unlike in Fischer 344 rats where females glucuronidate BPA to a greater extent than males (Pottenger et al., 2000). The difference between the maximal rate of BPA glucuronidation by human and immature female rat liver microsomes has potentially important implications for predicting the estrogenicity of BPA in humans from the rodent uterotrophic assay. Uterotrophic activity of BPA in immature rats is dependent upon the route of administration, with the oral route requiring larger doses than the subcutaneous route to produce a uterotrophic response (Ashby and Tinwell, 1998; Laws et al., 2000). This can be attributed to the extent of glucuronidation, which is expected to be greatest following oral administration of BPA due to first-pass metabolism (Pottenger et al., 2000). Human exposure to BPA is likely to be greatest via the oral route and therefore uterotrophic assays using this route of administration would be appropriate. However, the predictive power of the uterotrophic assay might be compromised by differences in the rates of BPA glucuronidation, because estrogen target tissues in humans may be subject to greater exposure to BPA than the tissues of the immature female rat. This would suggest that assessment of the estrogenicity of BPA using the immature rat uterotrophic assay might well underestimate the potency of BPA in humans. This hypothesis was supported by coupling microsomal metabolism with the yeast estrogenicity assay, when there was a greater reduction in the estrogenic activity of BPA (determined by EC₅₀) following glucuronidation by immature female rat liver microsomes compared with human liver microsomes. The reduction in estrogenicity in vitro was only 2.5-fold, and therefore it is uncertain whether this effect will be significant in vivo. However, the greater uterotrophic activity of BPA in the rat following subcutaneous administration, compared with oral administration (Ashby and Tinwell, 1998), can be related to both lower plasma BPA glucuronide levels and consequently higher plasma BPA levels (Pottenger et al., 2000). The recent finding that BPA was inactive in the immature mouse uterotrophic assay (300 mg/kg oral administration) (Tinwell et al., 2000) suggests that the mouse may glucuronidate BPA to an even greater extent. Therefore, it would be of interest to determine whether an inverse relationship exists between the rate of BPA glucuronidation in vitro and the uterotrophic activity of BPA in various mammalian species. Thereby, it may be possible to determine whether there is a minimum disparity between rates of BPA glucuronidation in vitro that can be related to a statistically significant disparity in biological response.

UGT2B1 is the major UGT isozyme responsible for BPA glucuronidation in rat liver (Yokota et al., 1999) and shares similar sequence homology to human UGT2B7 and UGT2B17 (Bélanger et al., 1998; Carrier et al., 2000). Both human isozymes are expressed in liver, but unlike UGT2B1, are also expressed in several steroid target tissues, such as brain, uterus, mammary gland, and testis (Bélanger et al., 1998; King et al., 1999), indicating the possibility that humans may have a degree of local protection toward BPA in these tissues. However, this may be further complicated by possible interindividual variation in glucuronidation as a consequence of a polymorphism in the UGT2B7 gene (Lampe et al., 2000).

Species differences in glucuronidation can be highly substrate-specific. For example, it was found that for -hydroxytamoxifen the rate of glucuronidation in human liver microsomes was 50-fold greater than that in rat microsomes (Boocock et al., 2000). In contrast, human and rat liver microsomes effectively glucuronidate 4-hydroxytamoxifen at similar rates (personal communication: D. J. Boocock, 2000).

BPA sulfate and 5-OHBPA sulfate were minor metabolites identified following incubations of BPA with rat hepatocytes. BPA sulfate has been detected in rats following intraperitoneal administration (Pottenger et al., 2000). 5-OHBPA was the only metabolite observed in incubations of BPA with immature female or human liver microsomes in the presence of NADPH. A hydroxylated metabolite of BPA was identified in vivo, following oral administration of BPA (800 mg/kg), which constituted approximately 35% of fecal radioactivity (Knaak and Sullivan, 1966). In contrast, Pottenger et al. (2000) did not report the presence of 5-OHBPA in feces following oral, intraperitoneal, or subcutaneous administration of BPA (10 or 100 mg/kg) to Fischer 344 rats. Of the seven minor unidentified metabolites present in feces, none represented more than 7% of the administered dose (Pottenger et al., 2000). It was hypothesized that the presence of the hydroxylated metabolite identified in the study of Knaak and Sullivan (1966) was evidence that oxidative metabolism only occurs in vivo at high doses following saturation of other metabolic pathways (Pottenger et al., 2000). In rat hepatocytes, both BPA sulfate and 5-OHBPA sulfate were only formed at the higher substrate concentration (0.5 mM), which may be indicative of saturation of the glucuronidation pathway.

5-OHBPA exhibits weak estrogenic activity (approximately 10-fold less potent than BPA) in vitro in the yeast assay, however, there was no significant effect of oxidative metabolism on the estrogenicity of BPA, as determined from coupled microsomal incubations and the yeast assay. This lack of effect can be attributed to the slow formation of 5-OHBPA in the microsomal system. Glucuronidation of BPA, not 5-hydroxylation, is likely to be the crucial determinant of the extent of estrogenic activity in vivo.

Human exposure to environmental levels of BPA are likely to have the most significant estrogenic impact during critical "windows" of development, such as in the fetus or neonate (Howdeshell et al., 1999; Takai et al., 2000). Moreover, limited glucuronidation activity has been demonstrated in human fetal liver (Ring et al., 1999). The majority of UGT isozymes is not expressed until after birth, with the full complement being expressed by 3 months of age, although at reduced levels (approximately 25%) compared with adult (Coughtrie et al., 1988). Indeed, UGT2B7 expression in 20-week human fetal liver is significantly lower than that of adults (King et al., 1999). This could suggest that the human fetus might be unable to efficiently glucuronidate BPA following maternal exposure. In contrast to the adult, limited UGT activity in the fetus may mean that alternative phase II metabolic pathways predominate (Ring et al., 1999); fetal phenol-sulfotransferase activity toward paracetamol is known to be greater than UGT activity in humans (Ring et al., 1999). Indeed, this same human sulfotransferase isozyme has been shown to sulfonate environmental estrogen-like chemicals such as BPA in vitro (Suiko et al., 2000). Therefore, it is possible that fetal liver has the capability to remove BPA via sulfonation.

Unlike the situation pertaining in human, significant amounts of phenol UGT activity, responsible for the glucuronidation of planar phenols, have been detected in rat fetal liver (>30% of adult levels) (Coughtrie et al., 1988). A recent study demonstrated that the rate of clearance of BPA, following administration of a high dose (1 g/kg) to pregnant rats, was far greater in the adult compared with the fetus, which resulted in a higher fetal blood concentration of BPA compared with adult over time (Takahashi and Oishi, 2000). The differences observed for the clearance of BPA from adult or fetal rat likely reflect a reduced ability of the fetus to glucuronidate BPA. The lack of observed reproductive toxicity in adult Wistar rats following fetal exposure to low doses of BPA (Cagen et al., 1999b) may reflect the ability of the pregnant rat to efficiently glucuronidate BPA, thereby reducing the levels of BPA reaching the fetus, where it can be further conjugated by limited fetal UGT activity, thus reducing the estrogenic burden placed upon the fetus. However, the effects could be different if the mother and, in turn the fetus, exhibited lower activity toward the glucuronidation of BPA.

The findings of the present study confirm that glucuronidation decreases the estrogenic activity of BPA. Consequently, the difference between the rates of glucuronidation in human liver microsomes and immature female rat liver microsomes suggests that humans may be exposed to a higher estrogenic burden than immature rat for the same dose of BPA.