Biomacromolecules 2000, 1, 656-664
656
Bisphenol A and Its Biomaterial Monomer Derivatives Alteration of in Vitro Cytochrome P450 Metabolism in Rat, Minipig, and Human J. Michael Cannon,‡ Elisabet Kostoryz,† Karen A. Russo,† Robert E. Smith,† and David M. Yourtee*,† Toxicore Laboratory, Division of Pharmacology, School of Pharmacy, University of MissourisKansas City, Kansas City, Missouri 64108; and Office of Regulatory Affairs, Kansas City District, Food and Drug Administration, Office, Lenexa, Kansas 66214 Received May 18, 2000; Revised Manuscript Received August 17, 2000
Bisphenol A (BPA) is a common structural component in a wide variety of biomaterial monomers. The effects of BPA and the following derivatives: bisphenol A glycidyl methacrylate (BisGMA), bisphenol A glycidyl diacrylate (BAGDA), bisphenol A ethoxylate dimethacrylate (BAEDM), bisphenol A dimethacrylate (BADM), and bisphenol A diglycidyl ether (BADGE) on mixed function oxidases (MFOs) are reported in this study. The rate of formation of metabolites from isoform-specific substrates for the MFOs (or cytochromes) CYP 1A, 2A, 2C, 2E, 3A, and 4A in the absence (control) and presence of BPA and derivatives was used to assess inhibition or stimulation of human, rat (male and female) liver, and minipig liver microsomal MFO activity. For human preparations the strongest inhibition by BPA was observed for CYP 2C. The inhibition was most prominent when a lower dose of BPA was used on the complete post-mitochondrial fraction. BPA inhibited rat microsomal CYP 1A isoform-specific metabolite production to 29 ( 3% of control levels (100%). Biomaterial monomers exhibited mixed effects. For example, BPA stimulated CYP 4A in pooled human S9 to 129 ( 1% of control. Also, BADM and BAGDA stimulated CYP 4A to 141% and 142% of control values, respectively. Introduction A study by the U.S. Food and Drug Administration reported that more than 100 synthetic chemicals can be found in 234 food commodities.1 Chemicals used in food storage and in biomedical applications include those based on bisphenol A (4,4′-isopropylidene diphenol, BPA). There is evidence that these chemicals can leach from their polymeric matrix, allowing consumption by humans.2,3 BPA has been linked to estrogenic effects.4,5 It was estimated that 109 tons of BPA were released into air, surface water, or wastewater during 1998.6 Derivatives that contain the BPA core structure are shown in Table 1. They are used widely in the manufacture of epoxy, polycarbonate, and corrosion-resistant resins for packaging material.7,8 BADGE and BFDGE contain the BPA or bisphenol F core structure and diglycidyl ether substituents. They have been identified as leachates in other studies.8,9 The unreacted monomer has been detected in water from renovated pipes in the U.K. This and other concerns have elicited studies for the analytical determination of this material in food matrixes.9 BADGE, BFDGE, and others such as BisGMA (with significant dental use) may elicit toxicity in their own right. Also, additional concern exists as to the potential metabolism of these compounds to BPA, which has demonstrated toxicity.10 * Corresponding author, to whom all reprint requests should be directed. † University of MissourisKansas City. ‡ Food and Drug Administration.
Physical and chemical characteristics of BPA and related derivatives are listed in Table 2. BPA posseses the required characteristics for rapid and complete diffusion across cellular membranes. Most importantly, BPA is nonionic despite the presence of two phenolic moieties, which have a pKa of approximately 10, and are un-ionized at physiological pH. The octanol/water partition coefficient of BPA is 3.52.5 BPA has been shown to produce damage to preputial gland, testes, kidneys, liver, spleen, pancreas, and lungs.10-13 The mechanisms of these toxic effects are not clearly defined but may be related to the formation of cellular DNA adducts.7,13 Sonnenschein and Soto15 described four possible effects of xenoestrogens such as BPA on normal basal endocrine functions: mimicry, antagonism, modification of hormone and/or hormone receptor levels, and alteration of metabolism. The metabolism of xenobiotics may be inhibited or increased by other chemicals acting on cytochrome oxidases.16 Although BPA is an extensively used chemical that has great potential for appearing in human tissues,6 there is limited information about its effect on key cytochromemetabolizing enzymes. In light of this, it was the purpose in this study to evaluate the hypothesis that BPA and some of its derivatives would affect mixed function oxidases. S9 preparations from uninduced animals were used for purposes of evaluating species differences and to permit direct comparison of control activities with those from previously published results. To
10.1021/bm005564+ CCC: $19.00 © 2000 American Chemical Society Published on Web 10/14/2000
Biomacromolecules, Vol. 1, No. 4, 2000 657
Alteration of in Vitro Cytochrome P450 Metabolism Table 1. Structure of BPA and Derivatives
Table 2. Molecular Properties of BPA and Derivatives surface areaa (Å2) volumeb (Å3) log Pc (O/W)
compound
MW
BPA BisGMA BADM BAEDM BAGDA BADGE BFDGE
228.2 512.6 364.5 452.5 484.6 340.4 312.4
263 596 431 518 540 404 370
219 485 351 432 449 326 291
3.52 4.64 5.29 6.29 4.02 3.98 3.18
a Property calculated using Alchemy 2000 software (Version 1, Tripos Associates, St. Louis, MO) after energy minimization using MM3. c Octanol/water (O/W) partition coefficient (log P) calculated using CLOGP software (Daylight Chemical Information Systems, URL http://www.daylight.com, February 15, 1999).
establish effects, cytochromes from rat, minipig, and human were evaluated. The cytochromes selected for this study were based on their significance to xenobiotic transformation as well as their role in the metabolism of endogenous substrates. Although over 26 human isoforms of cytochrome P450 have been characterized,17,18 the major proportion of human liver cytochromes are as follows: CYP 1A (13%), CYP 2A (4%), CYP 2C (18%), CYP 2E (7%), and CYP 3A (30%). Furthermore, CYP 1A, 2C, 2E, and 3A are involved in the metabolism of numerous drugs and environmental contaminants17 as well as endogenous compounds. Cumulatively, these families of cytochromes comprise about 70% of the biotransformation activities in humans and laboratory animals. 2. Materials and Methods 2.1. Chemicals. The following chemicals (with CAS numbers) were obtained from Aldrich Chemical Co., Mil-
waukee, WI: BPA (80-05-7), bisphenol A dimethacrylate (BADM, 3253-39-2), bisphenol A glycerolate diacrylate (BAGDA, 4687-94-9), carbon monoxide, and bisphenol A ethoxylate dimethacrylate (BAEDM, 41637-38-1). BisGMA (1565-94-2) and BADGE (1675-54-3) were obtained from 3M Corp., Minneapolis, MN. BFDGE (2095-03-6) was from Ciba Chemicals, Brewster, NY. Bicinchoninic acid solution (BCA), caffeine, copper(II) sulfate pentahydrate, lauric acid, 8-methoxypsoralen (8-MP), (()-miconazole, R-naphthoflavone (R-NF), NAD+ Type III (yeast), NADP+, potassium phosphate dibasic trihydrate, bovine serum albumin, terfenadine, theophylline (1,3-dimethylxanthine), sucrose, sulfaphenazole, theobromine (3,7-dimethylxanthine), trifluoroacetic acid (TFA), and 1,3,7-trimethyluric acid (137TMU) were obtained from Sigma Chemical, St. Louis MO. HPLC grade methanol and acetonitrile, glacial acetic acid, magnesium chloride, potassium chloride, potassium phosphate monobasic, Scintiverse LC, sodium phosphate dibasic anhydrous, sodium hydrosulfite (dithionite), sodium citrate tribasic, sodium phosphate monobasic dihydrate, and trichloroacetic acid (TCA) were from Fisher Scientific, Pittsburgh, PA. Coumarin, 1,7-dimethylxanthine (17DMX), D-glucose-6-phosphate (monosodium salt), glucose-6-phosphate dehydrogenase (EC 1.1.1.49), and 7-hydroxycoumarin (7OHC), were from Acros Chemicals, Pittsburgh, PA. Diclofenac sodium was from RBI, Natick, MA. 4′-Hydroxydiclofenac (4OHD) and terfenadine alcohol metabolite (TAM) were obtained from GenTest Corp., Woburn, MA. The 1-[14C]-lauric acid was obtained from American Radiolabeled Chemicals (ARC), St. Louis, MO.
658
Biomacromolecules, Vol. 1, No. 4, 2000
2.2. Liver Preparations. Liver samples were from an uninduced male animal provided by the University of MissourisColumbia School of Veterinary Medicine. The animal received preanesthetics: atropine (10 mg/kg), rompun (2.25 mg/kg), ketamine (25 mg/kg), and heparin (1000 µg/ kg). Tissues were harvested after thiopental euthanasia (1020 mg/kg) and stored in ice cold 0.15 M KCl. S9 was prepared, as described in the next sections, within 6 h of harvest. Liver samples from three male and three female uninduced Sprague-Dawley rats (12 weeks of age) were obtained from Midwest Research Institute (Kansas City, MO). The animals were sacrificed using metofane, and the tissues were harvested and placed in cold 0.15 M potassium chloride and stored on ice. S9 was prepared from these tissues within 4 h of harvest. Sprague-Dawley rat livers (male and female) and human microsomal and S9 preparations were from XenoTech LLC (Kansas City, KS). Individual livers were weighed and combined with three volumes of ice cold 0.15 M KCl in a Waring blender. The contents were homogenized for 30-45 s and then transferred to centrifuge tubes. Centrifugation was performed at 9000g for 10 min using a Sorvall R2B refrigerated centrifuge. The supernatant was transferred to cryovials and stored at -70 to -80 °C. 2.3. Analytical Methods. The activity of the selected cytochromes in the absence (control) and presence of test materials was measured to evaluate the extent of the formation of metabolites from isoform-specific substrates. Methods specific for the assay of these metabolites were utilized as described in the literature. Two of these methods provided for the simultaneous evaluation of multiple isoforms. The formation of 17DMX and 137TMU from caffeine is reported to proceed via high affinity CYP 1A2 and CYP 3A4 enzymes, respectively,19,20 and consequently, the rate of formation of each metabolite was indicative of specific isoform activity and can be determined using a single HPLC system.21 The formation of 7OHC from coumarin was performed according to the method of Pearce and others.21 Microsomes or S9 (1-2 mg) were incubated at 37 °C with phosphate buffer (25 mM, pH 7.4), MgCl2 (3 mM), EDTA (1 mM), NADP+ (1 mM), glucose-6-phosphate (5 mM), glucose-6phosphate dehydrogenase (1 unit/mL), and coumarin (50 µM) in a final volume of 0.5 mL at the indicated final concentrations. Reactions were initiated by the addition of the NADPH generating system. After 30 min of incubation, trichloroacetic acid (30 µL of a 30% aqueous solution) was added followed by an aliquot of methanol (300 µL) to precipitate proteins. 8-Methoxypsoralen (2 µM), a mechanism-based inhibitor of CYP 2A, was used as positive control.24 The mixture was vortexed (10 s) and then centrifuged (20 min, 2000g). The supernatant was analyzed by reverse phase HPLC. The concentration of protein in the S9 preparations from male and female SD rat liver and male Yucatan minipig liver was determined using a commercially available kit (BCA Protein Assay, Sigma Chemical Co., St. Louis, MO). Bovine serum albumin (BSA), supplied with the kit, served as the standard. Solutions of protein (albumin standard or S9) in water were mixed with BCA and analyzed by visible light spectroscopy using an analytical wavelength of 562 nm with
Cannon et al.
a double beam spectrophotometer. The concentration of cytochrome P450 was determined according to the method of Matsubara and co-workers.23 The HPLC system was a Waters model 510 pump and a Waters model 712 WISP (Milford, MA). Separations were obtained with a Metachem Inertsil C18 column (250 × 4.6 mm ID, 5 µm) column. The mobile phase was water, methanol, and acetic acid (60/40/1, v/v/v), flowing at 1.2 mL/min. A Model RF-535 fluorometer (Shimadzu, Kyoto, Japan) was set for an excitation wavelength of 350 nm and an emission wavelength of 440 nm for detection of 7OHC. Data were acquired using a model C-R3A integrator (Shimadzu, Kyoto, Japan). Chromatographic suitability was verified from analysis of solution standards of 7OHC prepared in water. The retention time of 7OHC was approximately 7 min under the conditions of analysis. Solutions of coumarin and 4-hydroxycoumarin were injected to verify separation from the 7-hydroxy species. The retention times of coumarin and 4-hydroxy coumarin were approximately 4 and 11 min, respectively, as determined using detection at 280 nm. The formation of 17DMX and 137TMU from caffeine using S9 and microsomal preparations utilized the methodology of Pearce et al.21 Control samples were similarly prepared except that no test substance was added. Positive controls included activators (naphthoflavone, 100 µM) and inhibitors (miconazole, 20 µM; and diethyldithiocarbamate, 200 µΜ) of CYP 1A. The final composition and concentration of components in the 0.250 mL incubation mixture were as follows: potassium phosphate buffer (50 mM, pH 7.4), MgCl2 (3 mM), EDTA (1 mM), NADP+ (1 mM), glucose6-phosphate (5 mM), glucose-6-phosphate dehydrogenase (1 U/mL), protein (0.1-10 mg/mL), and caffeine (1 mM). Reactions were initiated by the addition of the NAPDH generating system and were terminated by the addition of 50 µL of 30% trichloroacetic acid. After briefly vortexing (∼2-4 s), the precipitated protein was removed by centrifugation (2000g, 20 min) and aliquots (100 µL) of the clear supernatant were analyzed by reverse phase HPLC for 17DMX and 137TMU content. The formation and analysis of 4′-hydroxydiclofenac from diclofenac in S9 and microsomal preparations utilized the methodology provided by GenTest Corp. (Woburn, MA). Aliquots of test substances (5 µL in ethanol) were added to individual glass tubes and the ethanol solvent was evaporated to dryness. Microsomal or S9 protein, diclofenac sodium (125 µL of a 0.4 mM solution in Tris buffer, pH 7.5, 0.1 M), and NADPH generating system (30 µL containing 13.3 mg/mL of NADP+, 13.3 mg/mL glucose-6-phosphate, 10 mg/mL MgCl2, and 13.3 units/mL glucose-6-phosphate dehydrogenase in 5 mM trisodium citrate) were added. Aliquots of Tris buffer were added to bring the final incubation volume to 500 µL. The reaction was initiated by addition of microsomal or S9 protein. Control samples were similarly prepared except that no test substance was added. Sulfaphenazole (200 µM), an inhibitor of CYP 2C,17 was incubated in the presence of substrate as a positive inhibition control. The reverse phase HPLC system used for 4′-hydroxy diclofenac also utilized the methodology provided by GenTest.The conversion of
Biomacromolecules, Vol. 1, No. 4, 2000 659
Alteration of in Vitro Cytochrome P450 Metabolism
Figure 1. Comparison of biomaterial effects on caffeine N3-demethylation (CYP 1A) in pooled rat and human liver microsomes. Functional group derivatives of BPA including dimethacrylates (BisGMA, BAEDM, BADM), acrylate (BAGDA), and epoxide (BADGE) from Table 1 are illustrated. Error bars represent the standard deviation of replicate determinations (n ) 3). Microsomes (0.33 mg of protein) were incubated at 37 °C for 1 h in 0.25 mL (final volume) incubation mixtures containing phosphate buffer (50 mM), MgCl2 (3 mM), NADP+ (1 mM), G-6-P (5 mM), G-6-P dehydrogenase (0.5 units), and caffeine (1 mM). The formation of 1,7-dimethylxanthine was determined by reverse phase HPLC with ultraviolet detection (280 nm).21 The N3-demethylation of caffeine in the presence of biomaterial monomers was compared to the corresponding control activity (/, p < 0.05; ///, p < 0.001). Control activities in rat and human microsomes were 66 and 93 pmol mg-1 min-1, respectively. Diethyldithiocarbamate (DEDTC), trioleadomycin (TAO), R-naphthoquinone (R-NF), and miconazole were used as positive inhibition controls.
lauric acid to the 11-hydroxy (ω-1 hydroxylation) and 12hydroxy (ω hydroxylation) metabolites utilized the general methods described by Pearce et al.21 Diethyl dithiocarbamate (DEDTC, 200 µM), an inhibitor of CYP 2E,17 was used as positive control. Reactions were initiated by the addition of the NAPDH generating system and were terminated by the addition of 30 µL of 30% trichloroacetic acid and 300 µL methanol. After briefly vortexing (∼2-4 s), the precipitated protein was removed by centrifugation (2000g, 20 min). Additional (200 µL) of the clear supernatant were analyzed by reverse phase HPLC using on-line radiometric detection. Additional aliquots of the supernatant (50 µL) combined with scintillation cocktail (8 mL) and counted using LSC to determine the counting efficiency of the on-line radiochemical detector. The reverse phase HPLC system used for the separation of lauric acid and metabolites was also based on the method reported by Pearce and co-workers.21 The formation of terfenadine alcohol metabolite (TAM) from terfenadine utilized the methods of Ling and co-workers.23 2.4. Data Interpretation. Statistical significance testing for means comparison was performed using Excel functions for probability using two-tailed tests assuming equal variance between samples. The protocol of study accepted differences as statistically significant with p e 0.05. Testing of Pearson’s correlation also utilized commercial spreadsheets and one-
Table 3. Effects of Biomaterial Monomers on TAM Formation with Liver S9 (CYP 3A)a biomaterial BPA BisGMA BAEDM BADM BAGDA BADGE BFDGE c
TAM formation with liver S9b (%) 100 ( 4.2 101.2 ( 17 167.2 ( 39.9 137.6 ( 32.6 85.3 ( 14.1 169.8 ( 44.2 87.7 ( 1.9c
100 ( 11 86.7 ( 28.1 164.6 ( 27.7c 195.8 ( 28.7c 114.5 ( 28.3 136.4 ( 16.7c 160.4 ( 30.8c
100 ( 6.4 89.3 ( 3 133.9 ( 4c 101.1 ( 7.8 107.9 ( 5.2 118.8 ( 8.8c 130.4 ( 10c
a Biomaterial doses were 30-32 µM. b Human liver microsomes. Different from control, p < 0.05.
tailed significance tables as utilized by Pearce and coworkers.22 Enzyme kinetics (Km, Vmax), and inhibition constants (Ki) were calculated using “Enzyme Kinetics”, Version 1.5 (Trinity Software, Camden, MA). 3. Results 3.1 Controls. In Table 3 are shown control activities for the cytochromes. Pooled rat and human liver preparations were the most active isoform preparations. The activity of coumarin 7-hydroxylase found for human liver microsomal preparations was 915 pmol mg-1 min-1. This value was consistent with the range 280-4750 pmol mg-1 min-1
660
Biomacromolecules, Vol. 1, No. 4, 2000
reported by Pearce et al.21 and a Vmax of 179-2470 pmol mg-1 min-1 reported by Draper et al.24 A Km of 1 µM was determined for coumarin in the presence of pooled human liver microsomes which was consistent with literature values of 0.2-0.7 µM.21,24 Only minor amounts (
