- Email: [email protected]
Biotransformation of tamoxifen in a human endometrial explant culture model
Chemico-Biological Interactions 146 (2003) 237–249
Biotransformation of tamoxifen in a human endometrial explant culture model Minoti Sharma a , David E. Shubert b , Moheswar Sharma a , Jennifer Lewis a , Barbara P. McGarrigle b , Diane P. Bofinger c , James R. Olson b,∗ a
Department of Molecular and Cellular Biophysics, Roswell Park Cancer Institute, Elm and Carlton Streets, Buffalo, NY 14263, USA Department of Pharmacology and Toxicology, State University of New York at Buffalo, 102 Farber Hall, Buffalo, NY 14214, USA c Department of Biotechnical and Clinical Laboratory Science, State University of New York at Buffalo, Buffalo, NY 14214, USA
b
Received 28 March 2003; received in revised form 28 March 2003; accepted 27 June 2003
Abstract Although long-term tamoxifen therapy is associated with increased risk of endometrial cancer, little is known about the ability of endometrial tissue to biotransform tamoxifen to potentially reactive intermediates, capable of forming DNA adducts. The present study examined whether explant cultures of human endometrium provide a suitable in vitro model to investigate the tissue-specific biotransformation of tamoxifen. Fresh human endometrial tissue, microscopically uninvolved in disease, was cut into 1 × 2-mm uniform explants and incubated with media containing either 25 or 100 M tamoxifen in a 24-well plate. Metabolites were analyzed by reversed-phase HPLC using postcolumn, online, photochemical activation and fluorescence detection. Three metabolites, namely, ␣-hydroxytamoxifen, 4-hydroxytamoxifen, and N-desmethyltamoxifen were identified in culture medium and tissue lysates. N-desmethyltamoxifen was found to be the major metabolite in both tissue and media extracts of tamoxifen-exposed explants. Incubations of tamoxifen with recombinant human cytochrome P-450s (CYPs) found that CYP2C9 and CYP2D6 produced all three of the above tamoxifen metabolites, while CYP1A1 and CYP3A4 catalyzed the formation of ␣-hydroxytamoxifen and N-desmethyltamoxifen, and CYP1A2 and CYP1B1 only formed the ␣-hydroxy metabolite. CYP2D6 exhibited the greatest activity for the formation of all three tamoxifen metabolites. Western immunoblots of microsomes from human endometrium detected the presence of CYPs 2C9, 3A, 1A1 and 1B1 in fresh endometrium, while CYPs 2D6 and 1A2 were not detected. Immunohistochemical (IHC) analysis also confirmed the presence of CYPs 2C9, 3A and 1B1 in fresh human endometrium and in viable tissue cultured for 24 h with or without tamoxifen. Together, the results support the use of explant cultures of human endometrium as a suitable in vitro model to investigate the biotransformation of tamoxifen in this target tissue. In addition, the results support the role of CYPs 2C9, 3A, 1A1 and 1B1 in the biotransformation of tamoxifen, including the formation of the DNA reactive ␣-hydroxytamoxifen metabolite, in human endometrium. © 2003 Elsevier Ireland Ltd. All rights reserved. Keywords: Tamoxifen biotransformation; Human endometrium; Explant culture; Cytochrome P-450
1. Introduction ∗ Corresponding author. Tel.: +1-716-829-2319; fax: +1-716-829-2801. E-mail address: [email protected] (J.R. Olson).
The anti-estrogen, (Z)-1-[4-(2-dimethylaminoethoxy)phenyl]-1,2-diphenyl-1-butene (tamoxifen), a non-
0009-2797/$ – see front matter © 2003 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.cbi.2003.06.002
238
M. Sharma et al. / Chemico-Biological Interactions 146 (2003) 237–249
steroidal drug, is used widely in the treatment of breast cancer and more recently has been approved as a chemopreventive agent in women with an elevated risk of developing breast cancer. While beneficial for the treatment of breast cancer, tamoxifen treatment is associated with an increased incidence of proliferative and neoplastic endometrial changes [1,2]. Tamoxifen has also been shown to cause endometrial and liver tumors in rodents [3,4]. The mechanism of induction of endometrial cancer by tamoxifen is not known currently, but may involve the bioactivation of tamoxifen to a genotoxic metabolite. In rat liver, tamoxifen–DNA adducts were identified as an indication of tamoxifen genotoxicity [5,6]. Although the results are not consistent between studies, tamoxifen–DNA adducts have been reported in the endometrium of women exposed to tamoxifen [7–10]. The present study was designed to examine whether explant cultures of human endometrium can provide a suitable in vitro model to investigate the tissue-specific biotransformation of tamoxifen to potentially genotoxic metabolites. Two major pathways have been proposed for the metabolic activation of tamoxifen: sulfotransferase-catalyzed sulfation of ␣-hydroxylated tamoxifen and N-desmethyltamoxifen [11–14] and peroxidase-catalyzed oxidation of 4-hydroxytamoxifen [15–18]. The present study focused on the cytochrome P-450 (CYP)-dependent metabolism of tamoxifen by human endometrium. Tamoxifen is metabolized by CYP enzymes to N-desmethyl-, 4-hydroxy- and ␣-hydroxytamoxifen, the later of which is sulfonated, forming an electrophilic carbocation that is capable of reacting with DNA [11–13,19,20]. Metabolism of tamoxifen in the liver with subsequent accumulation of metabolites in this tissue has been reported in both rodents and humans [21]. While liver is the major site of drug metabolism, extrahepatic target tissues can also contribute to localized biotransformation/activation of tamoxifen. Currently, in vivo and in vitro data on the localized (i.e. tissue specific) metabolism of tamoxifen are very limited. The expression of CYP genes (mRNA) in human endometrial tissue suggests that this tissue may have the potential to biotransform tamoxifen and subsequently to generate genotoxic metabolites of tamoxifen [22–24]. However, the presence of mRNA for specific CYPs, does not necessarily reflect the presence of catalytically active
protein in human endometrium. The present report describes a human endometrial explant culture model that was found to be suitable for investigating the biotransformation of tamoxifen within this target tissue. In addition, this study reports on the expression of specific CYP proteins in human endometrium and on the ability of recombinant CYPs to biotransform tamoxifen to potentially genotoxic products.
2. Materials and methods 2.1. Surgical specimens Human endometrial tissue specimens, removed at hysterectomy, were procured under IRB approved protocols from the Tissue Procurement Facility at Roswell Park Cancer Institute, with donor’s consent, but without the patients’ identities. Endometrial tissue was obtained from a total of 15 individuals, approximately 35–45 years of age. 2.2. Tissue culture medium D-MEM/F-12 medium (phenol red-free with 15 mM HEPES buffer, l-glutamine, and pyridoxine HCl; Life Technologies, Grand Island, NY) was used in all stages of tissue preparation and explant culture. D-MEM/F-12 medium was supplemented with 3% charcoal stripped fetal bovine serum (FBS; Life Technologies), 1% antibiotic/antimycotic solution (penicillin/streptomycin/amphotericin; Life Technologies) and 17-estradiol (10 nM). Charcoal-stripped FBS was prepared by stirring FBS with washed and dried charcoal (30 g charcoal/1 l FBS) at 4 ◦ C overnight, then removing the charcoal with 1.22, 0.45 and 0.22 m filters and performing heat inactivation at 56 ◦ C for 30 min. Cultures were treated either with tamoxifen or an equivalent volume of ethanol (tamoxifen vehicle) such that the final concentration of ethanol in the medium was 0.1% and the final concentration of tamoxifen in the medium was 25 or 100 M. 2.3. Explant culture The surgical specimens were prepared and cultured under sterile conditions similar to the method described by Osteen et al. [25,26]. The time period
M. Sharma et al. / Chemico-Biological Interactions 146 (2003) 237–249
239
between the surgical removal and explant culture was within 2 h. Typically, each sample of fresh endometrial tissue, microscopically uninvolved in disease, was placed in 3% FBS–D-MEM/F-12 medium containing 17-estradiol (10 nM) and cut into uniform explants with a sterile scalpel blade. The pieces were immediately transferred at a concentration of eight to ten pieces per well (30–67 mg/well) to a 24-well plate (Costar Corp., Cambridge, MA) containing 1 ml medium/well with 25 or 100 M tamoxifen or vehicle (0.1% ethanol). The explants were incubated for 24 h at 37 ◦ C in a humidified 5% CO2–air environment. At the end of the incubation, the explants (10–53 mg/well) and culture media were harvested for HPCL metabolite analysis, or the explants were fixed and embedded in agar for morphology and immunohistochemistry.
tamoxifen with microsomes from lymphoblastoid cell lines only expressing vector (control+vector) served as a negative control. Background product formation from this control preparation was subtracted from all incubations containing recombinant CYPs to correct for any background oxidation not mediated by CYPs. Incubation mixtures contained the following components in a final volume of 1000 l:100 mM sodium phosphate buffer (pH 7.4), 1.4 mM NADPH, 5 mM MgCl2 , and 25 M tamoxifen and 10–30 pmol CYP. Immediately after incubation (60 min at 37 ◦ C) metabolites were extracted twice with 5 ml of 2% ethanol in hexane, dried under nitrogen, resuspended in 500 l of methanol, and stored at −20 ◦ C prior to analysis by HPLC.
2.4. Morphology and immunohistochemistry
Whole-tissue lysates were prepared from explants that were pooled from wells cultured under the same condition. The explants were collected by centrifugation (155 × g, 5 min), then suspended in PBS (a 1:10 dilution of ice-cold 10 × Dulbecco’s phosphate-buffered saline; Life Technologies) in a ratio of 10 mg tissue/0.1 ml PBS. The suspension was homogenized on ice by sonication (Heat Systems-Ultrasonics Model w-380 sonicator, Farmingdale, NY; 5–10% power, using intervals of five 0.5-s pulses with 1 min resting periods between intervals) until lysed. Metabolites were extracted from culture media using 2% ethanol in hexane (5 ml/ml medium × 2) [7]. The organic extract was evaporated to dryness by rotary evaporation, the residue was reconstituted with methanol (250 l), and an aliquot (25 l) was used for HPLC analysis. Metabolites were extracted from whole-tissue lysates (50 l) by treating the lysates with methanol (3 vol) and 10 l 6 N HCl to denature the protein [27]. The samples were vortexed for 1 min and centrifuged (200 × g, 5 min), and the supernatant (25–50 l) was used directly for HPLC analysis.
At the end of the culture incubation period, the medium from each well was removed and replaced with 10% buffered formalin for overnight fixation. The formalin was then removed and 1% agar (∼55 ◦ C) was placed in each well [26]. After the agar solidified, the agar plug was then paraffin-embedded and 5 m thick sections were cut and placed on Superfrost Plus slides (Fisher Scientific Co., Pittsburgh, PA). Slides were stained with hematoxylin and eosin for the characterization of morphology or with hematoxylin for immunohistochemical (IHC) localization of CYP proteins. CYP3A and CYP2C9 proteins were localized by a modification of the CYP1B1 localization procedure [26]. The antigen retrieval step was eliminated and the primary antibody was replaced with rabbit anti-human CYP3A or CYP2C9 (Gentest, Woburn, MA). UPC antibody (mouse myloma protein IgG 2A; Organon Teknik, Durham, NC), which also reacts with the secondary antibody, was used to stain the negative controls for the specimens.
2.6. Tissue preparation and metabolite extraction
2.5. Incubations with recombinant CYPs 2.7. Metabolite analysis The relative ability of various CYPs to biotransform tamoxifen was assessed using lymphoblast expressed recombinant human CYP1A1, CYP1A2, CYP3A4, control+vector microsomes and CYP2C9, CYP2D6, CYP1B1 Supersomes (BD Gentest). Incubation of
The metabolites were analyzed by reversed-phase HPLC using postcolumn, online, photochemical activation and fluorescence detection [27–29]. The HPLC system consisted of a binary pump system and an
240
M. Sharma et al. / Chemico-Biological Interactions 146 (2003) 237–249
injection valve with variable loops (20–200 l) from Rainin Instruments Co. Inc., a Radial-Pak 8MBC18 LC cartridge (10 m, 8 mm i.d., 10 cm) with a compatible guard column from Waters, a postcolumn photochemical reactor from Aura Industries Inc. containing a 0.25 mm i.d., 5-m PTEF knitted reactor coil, and a 254-nm UV lamp which converted tamoxifen and the metabolites to fluorophores. The effluent from the photochemical converter was connected to a Shimadzu 530 RF fluorescence detector, operating at excitation and emission wavelengths of 260 and 375 nm, respectively. The detector signal was integrated by a Shimadzu Integrator CR501. HPLC grade solvents and analytical grade reagents were used to prepare the solvent system. All solvents were filtered through a Nylon-66 filter (0.2 m). A high-pressure inline filter (SSI, 0.5 m) was used as a further safe guard between each pump and the injector. The solvent system consisted of 90% methanol in 100 mM ammonium acetate, pH 5. The metabolites were analyzed under isocratic conditions at a flow rate of 0.5 ml/min. 2.8. Authentic standards for the identification of tamoxifen metabolites Tamoxifen and 4-hydroxytamoxifen were purchased from Sigma Chemical Co. (St. Louis, MO). ␣-Hydroxytamoxifen and N-desmethyltamoxifen were synthesized following reported procedures [30,31]. Spectroscopic characterization of the isolated products by mass spectrometry and nuclear magnetic resonance (nmr) were in agreement with the literature reports. 2.9. Electrophoresis and immunoblotting Microsomal protein fractions were prepared at 4 ◦ C from unincubated specimens of human endometrium as described by Drahushuk et al. [32], essentially following the protocol of Cinti et al. [33]. Protein concentrations were determined by the method of Bradford [34] using bovine serum albumin as a standard. The microsomal preparations were stored at −70 ◦ C until further analysis. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and western immunoblotting procedures were performed generally as described
earlier [32,35]. CYP1B1, CYP2C9, CYP2D6 and CYP3A proteins were quantitated by western immunoblotting as described by Drahushuk et al. [32] using chemiluminescent detection (CDP-Star; Tropix, Bedford, MA). Polyclonal antibody to CYP1B1 was generously provided by T.R. Sutter, University of Memphis, Memphis, TN. Monoclonal antibody to human CYP1A1 was generously provided by J. Stegeman, Woods Hole Institute, Woods Hole, MA. Specific and sensitive monoclonal antibody to human CYP2D6 and polyclonal antibody to human CYP2C9 and CYP3A4/3A7 were obtained from BD Gentest.
3. Results The biotransformation of tamoxifen was investigated using an explant culture model of human endometrial tissue. Routine histopathological analysis was conducted on fresh, unincubated tissue, as well as on explants cultured for 24 h under various conditions. The routine morphological analysis was necessary to assess not only tissue viability, but also to assess the homogeneity of tissue samples, including contamination by myometrium. Most cultured specimens were greater than 90% endometrium and were suitable for investigation. Endometrium remained viable in specimens that were cultured for up to 24 h in medium containing up to 100 M tamoxifen (see IHC analysis of CYPs, Fig. 5). Authentic standards of tamoxifen metabolites were analyzed by HPLC using online, postcolumn photochemical activation. Fig. 1 shows the reversed-phase HPLC separation of a mixture of ␣-hydroxytamoxifen (1), 4-hydroxytamoxifen (2), N-desmethyltamoxifen (3) and tamoxifen (4) (structures shown in Fig. 2). The elution order was based on polarity of the compounds, as expected. The intra- and interday variation in retention time and fluorescence signal (integrated peak area) of the peaks were within 3%. The signal for each compound was linear over a range of 1–100 ng/ml (correlation coefficient 0.999, n = 3). The flow velocity of the eluent through the capillary, and as a result, the residence time of the analytes in the irradiated zones appear to have some influence on the fluorescence signals (results not shown). The elution conditions were optimized to give the best resolution of the tamoxifen derivatives, not only from one
M. Sharma et al. / Chemico-Biological Interactions 146 (2003) 237–249
241
Table 1 Tamoxifen metabolites produced following 24 h of incubation of 100 M tamoxifen with explant cultures of human endometrium Metabolite
Media (pmol/ml)
Tissue lysate (pmol/mg)
␣-OH tamoxifen 4-OH tamoxifen N-desmethyltamoxifen
69.7 ± 61.4 128.8 ± 66.3 195.8 ± 73.3
1.2 ± 0.9 1.3 ± 0.8 24.9 ± 20.4
Values represent the mean ± S.D. of 15 individual specimens.
Fig. 1. HPLC profile of tamoxifen and authentic standards of three tamoxifen metabolites are shown using postcolumn, online, photochemical activation and fluorescence detection. Peaks 1–4 correspond to ␣-hydroxytamoxifen, 4-hydroxytamoxifen, N-desmethyltamoxifen and tamoxifen, respectively. HPLC conditions are described under Section 2.
another, but also from a large background of tamoxifen usually found in the exposed biological samples. Under these conditions (see Section 2), the lower limit of detection (S/N = 3) was 0.25, 0.13 and 0.14 pmol for 4-hydroxytamoxifen, ␣-hydroxytamoxifen and N-desmethyltamoxifen, respectively. Fig. 3 illustrates HPLC profiles of the metabolites extracted from the explants of human endometrial tissue exposed for 24 h to 0, 25, and 100 M tamoxifen in culture. In the explants exposed to 25 M tamoxifen (profile a), three metabolites (␣-hydroxytamoxifen, 4-hydroxytamoxifen and N-desmethyltamoxifen) were identified by cochromatography with the authentic standards (profile b). These peaks were not present in the control explants cultured under the same conditions using the vehicle (0.05% ethanol) (profile d). The metabolites were detected more readily using 100 M tamoxifen exposure (profile c). HPLC profiles of tamoxifen metabolites extracted from the media of the above cultures are identical to those illustrated in Fig. 3 (results not shown).
The effect of the incubation period on the stability of the metabolites detected in Fig. 3 was further determined by incubating a mixture of the authentic metabolites of known concentration (5 ng/ml) in the culture medium at 37 ◦ C. Aliquots were withdrawn at 0, 24 and 48 h of incubation, extracted and analyzed as described in Section 2. At each of the time points, the recoveries of each of the three metabolite standards ranged from 92 to 96%, demonstrating that the metabolites were stable for up to 48 h under the incubation conditions used in the experiments. Table 1 summarizes data on the relative quantity of each of the major metabolites of tamoxifen formed following 24 h of incubation of the explant cultures of human endometrial tissue with 100 M tamoxifen. The extraction efficiency of this procedure, determined by adding authentic standards of known concentration to the vehicle-exposed tissue lysates, was satisfactory (> 90%). Although all three metabolites were readily detectable, N-desmethyltamoxifen was found to be the major metabolite in both media and whole tissue lysates. Table 2 summarizes data on the relative ability of several recombinant human CYPs to metabolize tamoxifen (25 M). CYP2C9 and CYP2D6 produce all three tamoxifen metabolites, while CYP1A1 and CYP3A4 catalyzed the formation of the ␣-hydroxy and N-desmethyl metabolites but not the 4-hydroxy metabolite. CYP1A2 and CYP1B1 only catalyzed the formation of ␣-hydroxytamoxifen. CYP2D6 clearly exhibits the greatest activity for the formation of all three tamoxifen metabolites. Microsomal protein from uncultured human endometrial specimens was analyzed for the presence of specific CYPs using western immunoblotting (Fig. 4). CYP1A1, 1B1, 2C9, CYP3A proteins were detected in microsomal protein specimens from individual human endometrial samples from three of three donors.
242
M. Sharma et al. / Chemico-Biological Interactions 146 (2003) 237–249
Fig. 2. Structures of tamoxifen and metabolites.
M. Sharma et al. / Chemico-Biological Interactions 146 (2003) 237–249
243
Fig. 3. HPLC chromatographs of extracts from whole-tissue lysates of human endometrial explants exposed in culture for 24 h to media containing (a) 25 M tamoxifen; (b) cochromatography of (a) with authentic standards from Fig. 1; (c) 100 M tamoxifen; (d) vehicle only. Incubation and extraction conditions are described under Section 2. Chromatographic conditions are the same as in Fig. 2.
Table 2 Biotransformation of tamoxifen to ␣-hydroxytamoxifen, 4-hydroxytamoxifen, and N-desmethyltamoxifen by recombinant human CYPs
CYP1A1 CYP1A2 CYP1B1 CYP2C9 CYP2D6 CYP3A4
␣-HydroxyTAM
4-HydroxyTAM
N-desmethylTAM
80 ± 40.44 12.38 ± 1.60 8.73 ± 6.66 7.18 ± 4.18 173.05 ± 20.79 63.50 ± 10.91
ND ND ND 21.46 ± 1.78 614.23 ± 42.93 ND
15.39 ± 31.67 ND ND 25.94 ± 11.81 639.84 ± 67.16 123.55 ± 18.04
The rates of metabolite production are expressed in pmol/min per nmol CYP (mean ± S.D. of triplicate experiments). Incubation of tamoxifen with microsomes from lymphoblastoid cell lines only expressing vector (control+vector) served as a negative control. Background product formation from this control preparation was subtracted from all incubations containing recombinant CYPs to correct for any background oxidation not mediated by CYPs.
244
M. Sharma et al. / Chemico-Biological Interactions 146 (2003) 237–249
and B). The negative controls for the samples were stained with negative-control antibody in place of the antibody against specific human CYPs (Fig. 5A, C and E). All specimens were counterstained with Mayer’s hematoxylin (purple–blue stain). Figs. 5 and 6 also show that the morphology of the explants and the expression of these CYPs were generally well retained during the 24 h incubation period in the presence of 0, 25 or 100 M tamoxifen. While in culture, the stroma becomes more edemic and less dense, while the epithelial glands retain the CYPs (red staining) in the cytoplasm. Together, these results indicate that the human endometrial explant culture model retains CYPs that may contribute to the tissue-specific metabolism of tamoxifen.
Fig. 4. Representative western immunoblots are shown for human endometrium and human liver microsomal protein using specific, sensitive antibodies to human CYP1A1, CYP1B1, CYP2C9, CYP2D6, and CYP3A4/3A7. Lanes 1–4 contain 0.01, 0.025, 0.05 and 0.1 pmol of CYP1B1, 2D6, 3A4, and 3A7 supersome standards (Gentest), respectively. For the CYP1A1, and 2C9 standard, lanes 1–4 contain 0.1, 0.25, 0.5, and 1.0 pmol, respectively. Lanes 5–7 contain 50 g of human endometrium microsomal protein from three different donors and lane 8 contains 50 g of human liver microsomal protein. Note that the antibody to human CYP3A4/3A7 is sensitive and specific for purified supersome standards containing CYP3A4 and CYP3A7.
With the exception of CYP1A1 and 1B1, these CYPs were also readily detected in human liver microsomes. In contrast, CYP2D6 could not be detected in the human endometrial specimens, although it was found at low levels in human liver microsomes. CYP1A2 was also not detected in microsomal protein from human endometrium (results not shown). The expression and retention of CYP2C9, 1B1, and 3A proteins were assessed by IHC analysis after 0 and 24 h exposures of human endometrial explant cultures to media containing 0, 25 or 100 M tamoxifen. The morphological analysis indicates that the endometrium contains a number of epithelial glands and that the tissue remains viable for 24 h in culture at tamoxifen concentrations up to 100 M (Fig. 5). IHC analysis indicates that explant cultures of human endometrium express CYP2C9, 1B1, and 3A constitutively, as illustrated by the presence of red stain in the stroma and epithelial glands (Fig. 5B, Fig. 6A
4. Discussion Although long-term tamoxifen therapy is associated with increased risk of endometrial cancer, little is known about the ability of endometrial tissue to biotransform tamoxifen to potentially reactive intermediates, capable of forming DNA adducts. The present study was designed to examine whether explant cultures of human endometrium can provide a suitable in vitro model to investigate the tissue-specific biotransformation of tamoxifen. This study used an explant culture model, developed by Osteen et al. [25], that uses fresh human endometrial biopsy tissue, containing both epithelial and stromal cells, similar in morphology to that observed in vivo. This model was recently used to investigate the effect of in vitro exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) on the expression of CYP1A1 and CYP1B1 in explant cultures of human endometrium [26]. The advantages of this model are the normal mixture of epithelial and stromal cells contained in the explants, and the ability to conduct morphological studies on the explants. The greatest disadvantage of the model is the small sample size, which limits extensive time and concentration-dependent metabolism studies, using drugs such as tamoxifen. In the present study, routine morphological analyses indicated that endometrial explants remain viable in specimens cultured for up to 24 h in medium containing 25 or 100 M tamoxifen (Figs. 5 and 6). For explant metabolism studies,
M. Sharma et al. / Chemico-Biological Interactions 146 (2003) 237–249
245
Fig. 5. IHC localization of CYP2C9 protein (red stain) in a fresh human endometrium specimen shown before (B) and after 24 h of explant culture in the absence (D) and presence of 100 M tamoxifen (F). Negative controls for the specimens were stained with a UPC negative-control antibody and are shown before (A) and after 24 h of explant culture in the absence (C) and presence of 100 M tamoxifen (E). The slides were counterstained with Mayer’s hematoxylin (purple blue). The images were photographed at an initial magnification of 200 ×. The scalar bar represents 50 M.
a 24 h incubation period with tamoxifen was chosen to optimize the CYP-mediated metabolism of tamoxifen, since CYP activities decrease with incubation time in primary culture models. IHC analysis indicated that explant cultures of human endometrium express CYP2C9, 3A4/3A7 and 1B1 constitutively (Figs. 5 and 6) and that the expression of these CYPs was generally well retained during 24 h in explant culture. Western immunoblot analysis also confirmed the presence of these immunoreactive CYP proteins, including low levels of CYP1A1, in microsomes from
human endometrium (Fig. 4). Thus, explant culture of human endometrium is a suitable in vitro model to investigate the CYP-dependent metabolism of tamoxifen in this target tissue. The expression of immunoreactive CYP1B1 in human endometrium is supported by an earlier study that reported the constitutive expression of CYP1B1 protein in explant cultures of human endometrial tissue, similar to that in uncultured biopsies of this tissue [26]. Hukkanen et al. [22] detected mRNA for CYPs 2C, 3A4, and 3A5 in human endometrium, but
246
M. Sharma et al. / Chemico-Biological Interactions 146 (2003) 237–249
Fig. 6. The IHC localization of CYP1B1 protein (red stain) in a fresh human endometrium specimen is shown before (A) and after 24 h of explant culture in the absence (C) and presence of 25 M tamoxifen (E). Panels on the right side illustrate the IHC localization of CYP3A protein (red stain) in a fresh human endometrium specimen shown before (B) and after 24 h of explant culture in the absence (D) and presence of 25 M tamoxifen (F). The slides were counterstained with Mayer’s hematoxylin (purple blue). The images were photographed at an initial magnification of 200 ×. The scalar bar represents 50 M.
did not detect message for CYPs 1A1, 1A2, 2A6, 2D6, 2F1, 3A7, and 19. In a related study, Vadlamuri et al. [23] reported that mRNA for CYP1A1 and CYP1B1 were constitutively expressed in human endometrial tissue. Although CYP3A7 is considered a human fetal liver CYP, Sarkar et al. [24] readily detected mRNA for CYP3A4 and CYP3A7 in human endometrium. In agreement with our IHC results in Fig. 6, in situ hybridization studies of Sarkar et al. [24] found that CYP3A7 expression was lo-
calized within the glandular epithelium as well as to some extent in the stroma surrounding the glandular structures. These investigators also found that the expression of CYP3A7 message in proliferative phase endometrium was significantly greater than in secretory phase tissue. Together, the above studies support our observations that immunoreactive CYP1A1, CYP1B1, CYP2C9 and CYP3A4/3A7 proteins are constitutively expressed in human endometrium.
M. Sharma et al. / Chemico-Biological Interactions 146 (2003) 237–249
Small sample size is a factor that often limits the determination of the role of tissue-specific biotransformation of chemotherapeutic agents in extra-hepatic tissues. In the present study, a limited amount of human endometrial tissue from explant culture (10–50 mg/well) was used to demonstrate the tissue-specific biotransformation of tamoxifen in this target tissue. Tamoxifen metabolites were measured both in the lysates of the cultured explants and in the culture media using HPLC with postcolumn, online, photochemical activation as a sensitive analytical tool. This sensitive tool enables the collection of data despite the limitations of small sample size. The quantity of metabolites was lower in the whole tissue lysates than in the culture media. This may be due to the irreversible binding of tamoxifen metabolites with proteins [36–38]. Although all three metabolites were readily detectable, N-desmethyltamoxifen was found to be the major metabolite in both media and whole tissue lysates. Similarly, N-desmethyltamoxifen was also found to be the major metabolite in endometrial samples collected during diagnostic hysteroscopy of breast cancer patients exposed to chronic tamoxifen therapy [39]. The considerable variation in the metabolite yields in both the culture media and the whole tissue lysates of the explants (Table 1) reflects the biological variation in the specimens from different individuals and possible variability in metabolism in proliferative and secretory phase tissue specimens. Incubation of tamoxifen with recombinant human CYP2C9 and CYP2D6 produced all three tamoxifen metabolites, while CYP1A1 and CYP3A4 only catalyzed the formation of the ␣-hydroxy and N-desmethyl metabolites (Table 2). CYP1A2 and CYP1B1 only mediated the formation of ␣-hydroxytamoxifen. CYP2D6 clearly exhibited the greatest activity for the formation of all three tamoxifen metabolites, however, the CYP2D6 protein was not detected in microsomes from three out of three human endometrial specimens. CYP2C9, CYP3A and CYP1B1 were detected by IHC analysis of explant cultures of human endometrium (Figs. 5 and 6) and these CYPs, in addition to CYP1A1, were detected by western immunoblotting of human endometrial microsomal protein (Fig. 4). Together, these results support the role of CYPs 2C9, 3A, 1A1, and 1B1 in the biotransformation of tamoxifen in human endometrium.
247
This conclusion is generally supported by earlier work [40,41], which reported that CYP2D6, 2C9 and 3A4, but not 1A1, 1A2, 2C19 and 2E1, contributed to the 4-hydroxylation of tamoxifen by human liver microsomes. Mani et al. [38] reported results utilizing rat and human liver microsomal protein that suggests that CYP3A activates tamoxifen to a reactive intermediate. The CYP3A family has also been reported to catalyze the N-demethylation of tamoxifen in human liver microsomes [42]. Recently, Boocock et al. [43] investigated the metabolism of tamoxifen by recombinant human CYPs and reported that CYP3A4 was the only form that catalyzed the formation of ␣-hydroxytamoxifen and the subsequent formation of tamoxifen–DNA adducts. While we also observed that CYP3A4 catalyzed the formation of ␣-hydroxytamoxifen, the results in Table 2 also show that ␣-hydroxytamoxifen was formed by CYPs 1A1, 1A2, 1B1, 2C9, and 2D6. Differences in source of enzyme, incubation conditions and/or analytical methods may in part explain the lack of complete agreement between these results. In agreement with our findings, Boocock et al. [43] also reported that CYP2D6 was the most active form in catalyzing the overall biotransformation of tamoxifen. However, while CYP2D6 may play a role in the hepatic metabolism of tamoxifen, particularly in individuals with higher hepatic levels of this enzyme, the lack of detectable CYP2D6 in the endometrium suggests that this enzyme may not play a significant role in tamoxifen metabolism in this target tissue. The genotoxicity of N-desmethyltamoxifen in rat liver, through a metabolic pathway involving ␣-hydroxylation, has been reported both in vitro and in vivo [44]. Although no DNA adducts of N-desmethyltamoxifen have been detected in extrahepatic tissue [5], the possibility of forming other types of DNA damage, such as DNA cleavage and 8-hydroxyguanosine formation cannot be ruled out [45]. Alpha-hydroxytamoxifen has been reported to be the major DNA-reactive metabolite in rat liver, thus contributing to the carcinogenic activity of tamoxifen in the rat and possibly the human [11–13]. DNA from the endometrium of women treated with tamoxifen was analyzed by a 32 P-postlabeling/HPLC on-line assay capable of detecting 2.5 adducts/1010 nucleotides [10]. With this sensitive assay, specimens from 8 of 16 subjects had detectable tamoxifen–DNA
248
M. Sharma et al. / Chemico-Biological Interactions 146 (2003) 237–249
adducts, providing further support for the genotoxicity of tamoxifen. This is in contrast to a similar study that found no convincing evidence for tamoxifen derived DNA adducts in human endometrium [8]. In summary, our results support the use of explant cultures of human endometrium as a suitable in vitro model to investigate the biotransformation of tamoxifen in this target tissue. In addition, the results support the role of CYPs 2C9, 3A, 1A1, and 1B1 in the biotransformation of tamoxifen, including the formation of the DNA reactive ␣-hydroxytamoxifen metabolite, in human endometrium. We are currently investigating the biotransformation and the potential formation of tamoxifen–DNA adducts in explant cultures of human endometrium and the effect of anti-oxidants on modulating these processes. This model will also provide a means to investigate mechanisms for the interindividual differences in the bioactivation of tamoxifen, which will assist future efforts to identify individuals that may be at greater risk of endometrial cancer will tamoxifen therapy.
Acknowledgements This study has been supported in part by grant CA 86875 from the NCI. We thank the Tissue Procurement Facility of Roswell Park Cancer Institute for providing the human endometrium specimens.
References [1] N. Colacurci, L. De Seta, P. De Franciscis, D. Mele, N. Fortunato, S. Cassese, Tamoxifen effects on endometrium, Panminerva Med. 42 (2000) 45–47. [2] B. Czernobilsky, B. Lifschitz-Mercer, Endometrial pathology, Curr. Opin. Obstet. Gynecol. 9 (1997) 52–56. [3] E. Mantyla, S. Karlsson, L. Nieminen, Induction of endometrial cancer by tamoxifen in the rat, in: J.J. Li, S.A. Li, J.A. Gustafsson, N. Nandi, L.I. Sekely (Eds.), Hormonal Carcinogenesis, vol. II, Springer, New York, 1996, pp. 442–445. [4] P. Hirsimaki, Y. Hirsimaki, L. Nieminen, S. Karlsson, E. Mantyla, On the safety studies of the antiestrogens toremifene and tamoxifen, Scand. J. Lab. Anim. Sci. 23 (1996) 147–153. [5] G. Gamboa da Costa, L.P. McDaniel-Hamilton, R.H. Heflich, M.M. Marques, F.A. Beland, DNA adduct formation and mutant induction in Sprague–Dawely rats treated with tamoxifen and its derivatives, Carcinogenesis 22 (2001) 1307– 1327.
[6] M.R. Osborne, A. Hewer, I.R. Hardcastle, P.L. Carmichael, D.H. Phillips, Identification of the major tamoxifen– deoxyguanosine adduct formed in the liver DNA of rats treated with tamoxifen, Cancer Res. 56 (1996) 66–71. [7] P.L. Carmichael, A.H. Ugwumadu, P. Neven, A.J. Hewer, G.K. Poon, D.H. Phillips, Lack of genotoxicity of tamoxifen in human endometrium, Cancer Res. 56 (1996) 1475–1479. [8] P.L. Carmichael, S. Sardar, N. Crooks, P. Neven, I. Van Hoof, A. Ugwumadu, T. Bourne, E. Tomas, P. Hellberg, A.J. Hewer, D.H. Phillips, Lack of evidence from HPLC 32 P-postlabelling for tamoxifen DNA adducts in the human endometrium, Carcinogenesis 20 (1999) 339–342. [9] S. Shibutani, N. Suzuki, I. Terashima, S.M. Sugarman, A.P. Grollman, M.L. Pearl, Tamoxifen–DNA adducts detected in the endometrium of women treated with tamoxifen, Chem. Res. Toxicol. 12 (1999) 646–653. [10] S. Shibutani, A. Ravindernath, N. Suzuki, I. Terashima, S.M. Sugarman, A.P. Grollman, M.L. Pearle, Identification of tamoxifen–DNA adducts in the endometrium of women treated with tamoxifen, Carcinogenesis 21 (2000) 1461–1467. [11] S. Shibutani, L. Dasaradhi, I. Terashima, E. Banoglu, M.W. Duffel, ␣-Hydroxytamoxifen is a substrate of hydroxysteroid (alcohol) sulfotransferase resulting in tamoxifen DNA adducts, Cancer Res. 58 (1998) 647–653. [12] H. Glatt, W. Davis, W. Meinl, H. Hermersdorfer, S. Venitt, D.H. Phillips, Rat, but not human, sulfotransferase activates a tamoxifen metabolite to produce DNA adducts and gene mutations in bacteria and mammalian cells in cultures, Carcinogenesis 19 (1998) 1709–1713. [13] S. Shibutani, P.M. Shaw, N. Suzuki, L. Dasaradhi, M.W. Duffel, I. Terashima, Sulfation of ␣-hydroxytamoxifen catalyzed by human hydroxysteroid sulfotransferase results in tamoxifen–DNA adducts, Carcinogenesis 19 (1998) 2007– 2011. [14] D.H. Phillips, A. Hewer, M.N. Horton, K.J. Cole, P.L. Carmichael, W. Davis, M.R. Osborne, N-demethylation accompanies ␣-hydroxylation in the metabolic activation of tamoxifen in rat liver cells, Carcinogenesis 20 (1999) 2003– 2009. [15] G.A. Potter, R. McCague, M. Jarman, A mechanistic hypothesis for DNA adduct formation by tamoxifen following hepatic oxidative metabolism, Carcinogenesis 15 (1994) 439– 442. [16] B. Moorthy, P. Sriram, D.N. Pathak, W.J. Bodell, K. Randerath, Tamoxifen metabolic activation: comparison of DNA adducts formed by microsomal and chemical activation of tamoxifen and 4-hydroxytamoxifen with DNA adducts formed in vivo, Cancer Res. 56 (1996) 53–57. [17] D.N. Pathak, K. Pongracz, W.J. Bodell, Activation of 4-hydroxytamoxifen and the tamoxifen derivative metabolite E by uterine peroxidase to form DNA adducts: comparison with DNA adducts formed in the uterus of Sprague–Dawley rats treated with tamoxifen, Carcinogenesis 17 (1996) 1785– 1790. [18] M.M. Marques, F.A. Beland, Identification of tamoxifen– DNA adducts formed by 4-hydroxytamoxifen quinone methide, Carcinogenesis 18 (1997) 1949–1954.
M. Sharma et al. / Chemico-Biological Interactions 146 (2003) 237–249 [19] L. Dasaradhi, S. Shibutani, Identification of tamoxifen–DNA adducts formed by ␣-sulfate tamoxifen and ␣-acetoxytamoxifen, Chem. Res. Toxicol. 10 (1997) 189–196. [20] C. Sanchez, S. Shibutani, L. Dasaradhi, J.L. Bolton, P.W. Fan, R.A. McClelland, Lifetime and reactivity of an ultimate tamoxifen carcinogen: the tamoxifen carbocation, J. Am. Chem. Soc. 120 (1998) 13513–13514. [21] E.A. Lien, E. Solheim, P.M. Ueland, Distribution of tamoxifen and its metabolites in rat and human tissue during steady state treatment, Cancer Res. 51 (1991) 4837–4844. [22] J. Hukkanen, M. Mantyla, L. Kangas, P. Wirta, J. Hakkola, P. Paakki, S. Evisalmi, O. Pelkonen, H. Raunio, Expression of cytochrome P450 genes encoding enzymes active in the metabolism of tamoxifen in human uterine endometrium, Pharmacol. Toxicol. 82 (1998) 93–97. [23] S.V. Vadlamuri, D.D. Glover, T. Turner, M.A. Sarkar, Regiospecific expression of cytochrome P4501A1 and 1B1 in human uterine tissue, Cancer Lett. 122 (1998) 143–150. [24] M.A. Sarkar, V. Vadlamuri, S. Ghosh, D.D. Glover, Expression and cyclic variability of CYP3A4 and CYP3A7 isoforms in human endometrium and cervix during the menstrual cycle, Drug Metab. Dispos. 31 (2001) 1–6. [25] K.G. Osteen, W.H. Rodgers, M. Gaire, J.T. Hargrove, F. Gorstein, L.M. Matrisian, Stromal–epithelial interaction mediated steroidal regulation of metalloproteinase expression in human endometrium, Proc. Natl. Acad. Sci. USA 91 (1994) 10129–10133. [26] D.P. Bofinger, L. Feng, L.H. Chi, J. Love, F.D. Stephen, T.R. Sutter, K.G. Osteen, T.G. Costich, R.E. Batt, S.T. Koury, J.R. Olson, Effect of TCDD exposure on CYP1A1 and CYP1B1 in explant cultures of human endometrium, Toxicol. Sci. 62 (2001) 299–314. [27] O. Merle, J. Guitton, M.D. Burke, M. Ollangnier, Rapid and simple method to determine tamoxifen and its major metabolites in human liver microsomes by high performance liquid chromatography with fluorescence detection, Anal. Lett. 31 (1998) 2067–2076. [28] K.M. Fried, I.W. Wainer, Direct determination of tamoxifen and its four major metabolites in plasma using coupled column high-performance metabolites in plasma using coupled column high-performance liquid chromatography, J. Chromatogr. B 655 (1994) 261–268. [29] M. Sharma, Analysis of tamoxifen–DNA adducts by high-performance liquid chromatography using postcolumn online photochemical activation, Biochem. Biophys. Res. Commun. 273 (2000) 40–44. [30] A.B. Foster, M. Jarman, O.-T. Leung, R. McCague, G. Leclercq, N. Devleeschouwer, Hydroxy derivatives of tamoxifen, J. Med. Chem. 28 (1985) 1491–1497. [31] K. Brown, R.T. Heydon, R. Jukes, I.N. White, E.A. Martin, Further characterization of the DNA adducts formed in rat liver after the administration of tamoxifen, N-desmethyltamoxifen or N,N-didesmethyltamoxifen, Carcinogenesis 20 (1999) 2011–2066. [32] A.T. Drahushuk, B.P. McGarrigle, H.-L. Tai, S. Kitareewan, J.A. Goldstein, J.R. Olson, Validation of precision-cut liver slices in dynamic organ culture as an in vitro model for
[33]
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
249
studying CYP1A1 and CYP1A2 induction, Toxicol. Appl. Pharmacol. 140 (1996) 393–403. D.L. Cinti, P. Moldeus, J.B. Schenkman, Kinetic parameters of drug-metabolizing enzymes in Ca2+ -sedimented microsomes from rat liver, Biochem. Pharmacol. 21 (1972) 3249– 3256. M.M. Bradford, A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein–dye binding, Anal. Biochem. 72 (1976) 248–254. V.K. Laemmli, Cleavage of structural protein during the assembly of the head of bacteriophage T4, Nature 227 (1970) 680–685. S.S. Dehal, D. Kupfer, Evidence that the catechol 3,4-dihydroxytamoxifen is a proximate intermediate to the reactive species binding covalently to proteins, Cancer Res. 56 (1996) 1283–1290. C. Mani, D. Kupfer, Cytochrome P-450-mediated activation and irreversible binding of the antiestrogen tamoxifen to proteins in rat and human liver: possible involvement of flavin-containing monooxygenases in tamoxifen activation, Cancer Res. 51 (1991) 6052–6058. C. Mani, R. Pierce, A. Parkinson, D. Kupfer, Involvement of cytochrome P-450 3A in catalysis of tamoxifen activation and covalent binding to rat and human liver microsomes, Carcinogenesis 15 (1994) 2715–2720. G. Giorda, L. Franceschi, D. Crivellari, M.D. Magri, A. Veronesi, C. Scarabelli, M. Furlanut, Determination of tamoxifen and its metabolites in the endometrial tissue of long-term treated women, Eur. J. Cancer 36 (Suppl. 4) (2000) S88–S99. H.K. Crew, S.W. Ellis, M.S. Lennard, G.T. Tucker, Variable contribution of cytochrome P450 2D6, 2C9 and 3A4 to the 4-hydroxylation of tamoxifen by human liver microsomes, Biochem. Pharmacol. 53 (1997) 171–178. S.S. Dehal, D. Kupfer, CYP2D6 catalyzes tamoxifen 4-hydroxylation in human liver, Cancer Res. 57 (1997) 3402– 3406. F. Jacolot, I. Simon, Y. Dreano, P. Beaune, C. Riche, F. Berthou, Identification of the cytochrome P450 IIIA family as the enzymes involved in the N-demethylation of tamoxifen in human liver microsomes, Biochem. Pharmacol. 41 (1991) 1911–1919. D.J. Boocock, K. Brown, A.H. Gibbs, E. Sanchez, K.W. Turteltaub, I.N. White, Identification of human CYP forms involved in the activation of tamoxifen and irreversible binding to DNA, Carcinogenesis 23 (2002) 1897–1901. G. Gamboa da Costa, L.P. Hamilton, F.A. Beland, M.M. Marques, Characterization of the major DNA adduct formed by ␣-hydroxy-N-desmethyltamoxifen in vitro and in vivo, Chem. Res. Toxicol. 13 (2000) 200–207. T. Okubo, F. Nagai, K. Ushiyama, Y. Yokoyama, S. Ozawa, K. Kano, S. Tomita, H. Kubo, I. Kano, DNA cleavage and 8-hydroxydeoxyguanosine formation caused by tamoxifen derivative in vitro, Cancer Lett. 122 (1998) 9–15.
Biotransformation of tamoxifen in a human endometrial explant culture model Minoti Sharma a , David E. Shubert b , Moheswar Sharma a , Jennifer Lewis a , Barbara P. McGarrigle b , Diane P. Bofinger c , James R. Olson b,∗ a
Department of Molecular and Cellular Biophysics, Roswell Park Cancer Institute, Elm and Carlton Streets, Buffalo, NY 14263, USA Department of Pharmacology and Toxicology, State University of New York at Buffalo, 102 Farber Hall, Buffalo, NY 14214, USA c Department of Biotechnical and Clinical Laboratory Science, State University of New York at Buffalo, Buffalo, NY 14214, USA
b
Received 28 March 2003; received in revised form 28 March 2003; accepted 27 June 2003
Abstract Although long-term tamoxifen therapy is associated with increased risk of endometrial cancer, little is known about the ability of endometrial tissue to biotransform tamoxifen to potentially reactive intermediates, capable of forming DNA adducts. The present study examined whether explant cultures of human endometrium provide a suitable in vitro model to investigate the tissue-specific biotransformation of tamoxifen. Fresh human endometrial tissue, microscopically uninvolved in disease, was cut into 1 × 2-mm uniform explants and incubated with media containing either 25 or 100 M tamoxifen in a 24-well plate. Metabolites were analyzed by reversed-phase HPLC using postcolumn, online, photochemical activation and fluorescence detection. Three metabolites, namely, ␣-hydroxytamoxifen, 4-hydroxytamoxifen, and N-desmethyltamoxifen were identified in culture medium and tissue lysates. N-desmethyltamoxifen was found to be the major metabolite in both tissue and media extracts of tamoxifen-exposed explants. Incubations of tamoxifen with recombinant human cytochrome P-450s (CYPs) found that CYP2C9 and CYP2D6 produced all three of the above tamoxifen metabolites, while CYP1A1 and CYP3A4 catalyzed the formation of ␣-hydroxytamoxifen and N-desmethyltamoxifen, and CYP1A2 and CYP1B1 only formed the ␣-hydroxy metabolite. CYP2D6 exhibited the greatest activity for the formation of all three tamoxifen metabolites. Western immunoblots of microsomes from human endometrium detected the presence of CYPs 2C9, 3A, 1A1 and 1B1 in fresh endometrium, while CYPs 2D6 and 1A2 were not detected. Immunohistochemical (IHC) analysis also confirmed the presence of CYPs 2C9, 3A and 1B1 in fresh human endometrium and in viable tissue cultured for 24 h with or without tamoxifen. Together, the results support the use of explant cultures of human endometrium as a suitable in vitro model to investigate the biotransformation of tamoxifen in this target tissue. In addition, the results support the role of CYPs 2C9, 3A, 1A1 and 1B1 in the biotransformation of tamoxifen, including the formation of the DNA reactive ␣-hydroxytamoxifen metabolite, in human endometrium. © 2003 Elsevier Ireland Ltd. All rights reserved. Keywords: Tamoxifen biotransformation; Human endometrium; Explant culture; Cytochrome P-450
1. Introduction ∗ Corresponding author. Tel.: +1-716-829-2319; fax: +1-716-829-2801. E-mail address: [email protected] (J.R. Olson).
The anti-estrogen, (Z)-1-[4-(2-dimethylaminoethoxy)phenyl]-1,2-diphenyl-1-butene (tamoxifen), a non-
0009-2797/$ – see front matter © 2003 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.cbi.2003.06.002
238
M. Sharma et al. / Chemico-Biological Interactions 146 (2003) 237–249
steroidal drug, is used widely in the treatment of breast cancer and more recently has been approved as a chemopreventive agent in women with an elevated risk of developing breast cancer. While beneficial for the treatment of breast cancer, tamoxifen treatment is associated with an increased incidence of proliferative and neoplastic endometrial changes [1,2]. Tamoxifen has also been shown to cause endometrial and liver tumors in rodents [3,4]. The mechanism of induction of endometrial cancer by tamoxifen is not known currently, but may involve the bioactivation of tamoxifen to a genotoxic metabolite. In rat liver, tamoxifen–DNA adducts were identified as an indication of tamoxifen genotoxicity [5,6]. Although the results are not consistent between studies, tamoxifen–DNA adducts have been reported in the endometrium of women exposed to tamoxifen [7–10]. The present study was designed to examine whether explant cultures of human endometrium can provide a suitable in vitro model to investigate the tissue-specific biotransformation of tamoxifen to potentially genotoxic metabolites. Two major pathways have been proposed for the metabolic activation of tamoxifen: sulfotransferase-catalyzed sulfation of ␣-hydroxylated tamoxifen and N-desmethyltamoxifen [11–14] and peroxidase-catalyzed oxidation of 4-hydroxytamoxifen [15–18]. The present study focused on the cytochrome P-450 (CYP)-dependent metabolism of tamoxifen by human endometrium. Tamoxifen is metabolized by CYP enzymes to N-desmethyl-, 4-hydroxy- and ␣-hydroxytamoxifen, the later of which is sulfonated, forming an electrophilic carbocation that is capable of reacting with DNA [11–13,19,20]. Metabolism of tamoxifen in the liver with subsequent accumulation of metabolites in this tissue has been reported in both rodents and humans [21]. While liver is the major site of drug metabolism, extrahepatic target tissues can also contribute to localized biotransformation/activation of tamoxifen. Currently, in vivo and in vitro data on the localized (i.e. tissue specific) metabolism of tamoxifen are very limited. The expression of CYP genes (mRNA) in human endometrial tissue suggests that this tissue may have the potential to biotransform tamoxifen and subsequently to generate genotoxic metabolites of tamoxifen [22–24]. However, the presence of mRNA for specific CYPs, does not necessarily reflect the presence of catalytically active
protein in human endometrium. The present report describes a human endometrial explant culture model that was found to be suitable for investigating the biotransformation of tamoxifen within this target tissue. In addition, this study reports on the expression of specific CYP proteins in human endometrium and on the ability of recombinant CYPs to biotransform tamoxifen to potentially genotoxic products.
2. Materials and methods 2.1. Surgical specimens Human endometrial tissue specimens, removed at hysterectomy, were procured under IRB approved protocols from the Tissue Procurement Facility at Roswell Park Cancer Institute, with donor’s consent, but without the patients’ identities. Endometrial tissue was obtained from a total of 15 individuals, approximately 35–45 years of age. 2.2. Tissue culture medium D-MEM/F-12 medium (phenol red-free with 15 mM HEPES buffer, l-glutamine, and pyridoxine HCl; Life Technologies, Grand Island, NY) was used in all stages of tissue preparation and explant culture. D-MEM/F-12 medium was supplemented with 3% charcoal stripped fetal bovine serum (FBS; Life Technologies), 1% antibiotic/antimycotic solution (penicillin/streptomycin/amphotericin; Life Technologies) and 17-estradiol (10 nM). Charcoal-stripped FBS was prepared by stirring FBS with washed and dried charcoal (30 g charcoal/1 l FBS) at 4 ◦ C overnight, then removing the charcoal with 1.22, 0.45 and 0.22 m filters and performing heat inactivation at 56 ◦ C for 30 min. Cultures were treated either with tamoxifen or an equivalent volume of ethanol (tamoxifen vehicle) such that the final concentration of ethanol in the medium was 0.1% and the final concentration of tamoxifen in the medium was 25 or 100 M. 2.3. Explant culture The surgical specimens were prepared and cultured under sterile conditions similar to the method described by Osteen et al. [25,26]. The time period
M. Sharma et al. / Chemico-Biological Interactions 146 (2003) 237–249
239
between the surgical removal and explant culture was within 2 h. Typically, each sample of fresh endometrial tissue, microscopically uninvolved in disease, was placed in 3% FBS–D-MEM/F-12 medium containing 17-estradiol (10 nM) and cut into uniform explants with a sterile scalpel blade. The pieces were immediately transferred at a concentration of eight to ten pieces per well (30–67 mg/well) to a 24-well plate (Costar Corp., Cambridge, MA) containing 1 ml medium/well with 25 or 100 M tamoxifen or vehicle (0.1% ethanol). The explants were incubated for 24 h at 37 ◦ C in a humidified 5% CO2–air environment. At the end of the incubation, the explants (10–53 mg/well) and culture media were harvested for HPCL metabolite analysis, or the explants were fixed and embedded in agar for morphology and immunohistochemistry.
tamoxifen with microsomes from lymphoblastoid cell lines only expressing vector (control+vector) served as a negative control. Background product formation from this control preparation was subtracted from all incubations containing recombinant CYPs to correct for any background oxidation not mediated by CYPs. Incubation mixtures contained the following components in a final volume of 1000 l:100 mM sodium phosphate buffer (pH 7.4), 1.4 mM NADPH, 5 mM MgCl2 , and 25 M tamoxifen and 10–30 pmol CYP. Immediately after incubation (60 min at 37 ◦ C) metabolites were extracted twice with 5 ml of 2% ethanol in hexane, dried under nitrogen, resuspended in 500 l of methanol, and stored at −20 ◦ C prior to analysis by HPLC.
2.4. Morphology and immunohistochemistry
Whole-tissue lysates were prepared from explants that were pooled from wells cultured under the same condition. The explants were collected by centrifugation (155 × g, 5 min), then suspended in PBS (a 1:10 dilution of ice-cold 10 × Dulbecco’s phosphate-buffered saline; Life Technologies) in a ratio of 10 mg tissue/0.1 ml PBS. The suspension was homogenized on ice by sonication (Heat Systems-Ultrasonics Model w-380 sonicator, Farmingdale, NY; 5–10% power, using intervals of five 0.5-s pulses with 1 min resting periods between intervals) until lysed. Metabolites were extracted from culture media using 2% ethanol in hexane (5 ml/ml medium × 2) [7]. The organic extract was evaporated to dryness by rotary evaporation, the residue was reconstituted with methanol (250 l), and an aliquot (25 l) was used for HPLC analysis. Metabolites were extracted from whole-tissue lysates (50 l) by treating the lysates with methanol (3 vol) and 10 l 6 N HCl to denature the protein [27]. The samples were vortexed for 1 min and centrifuged (200 × g, 5 min), and the supernatant (25–50 l) was used directly for HPLC analysis.
At the end of the culture incubation period, the medium from each well was removed and replaced with 10% buffered formalin for overnight fixation. The formalin was then removed and 1% agar (∼55 ◦ C) was placed in each well [26]. After the agar solidified, the agar plug was then paraffin-embedded and 5 m thick sections were cut and placed on Superfrost Plus slides (Fisher Scientific Co., Pittsburgh, PA). Slides were stained with hematoxylin and eosin for the characterization of morphology or with hematoxylin for immunohistochemical (IHC) localization of CYP proteins. CYP3A and CYP2C9 proteins were localized by a modification of the CYP1B1 localization procedure [26]. The antigen retrieval step was eliminated and the primary antibody was replaced with rabbit anti-human CYP3A or CYP2C9 (Gentest, Woburn, MA). UPC antibody (mouse myloma protein IgG 2A; Organon Teknik, Durham, NC), which also reacts with the secondary antibody, was used to stain the negative controls for the specimens.
2.6. Tissue preparation and metabolite extraction
2.5. Incubations with recombinant CYPs 2.7. Metabolite analysis The relative ability of various CYPs to biotransform tamoxifen was assessed using lymphoblast expressed recombinant human CYP1A1, CYP1A2, CYP3A4, control+vector microsomes and CYP2C9, CYP2D6, CYP1B1 Supersomes (BD Gentest). Incubation of
The metabolites were analyzed by reversed-phase HPLC using postcolumn, online, photochemical activation and fluorescence detection [27–29]. The HPLC system consisted of a binary pump system and an
240
M. Sharma et al. / Chemico-Biological Interactions 146 (2003) 237–249
injection valve with variable loops (20–200 l) from Rainin Instruments Co. Inc., a Radial-Pak 8MBC18 LC cartridge (10 m, 8 mm i.d., 10 cm) with a compatible guard column from Waters, a postcolumn photochemical reactor from Aura Industries Inc. containing a 0.25 mm i.d., 5-m PTEF knitted reactor coil, and a 254-nm UV lamp which converted tamoxifen and the metabolites to fluorophores. The effluent from the photochemical converter was connected to a Shimadzu 530 RF fluorescence detector, operating at excitation and emission wavelengths of 260 and 375 nm, respectively. The detector signal was integrated by a Shimadzu Integrator CR501. HPLC grade solvents and analytical grade reagents were used to prepare the solvent system. All solvents were filtered through a Nylon-66 filter (0.2 m). A high-pressure inline filter (SSI, 0.5 m) was used as a further safe guard between each pump and the injector. The solvent system consisted of 90% methanol in 100 mM ammonium acetate, pH 5. The metabolites were analyzed under isocratic conditions at a flow rate of 0.5 ml/min. 2.8. Authentic standards for the identification of tamoxifen metabolites Tamoxifen and 4-hydroxytamoxifen were purchased from Sigma Chemical Co. (St. Louis, MO). ␣-Hydroxytamoxifen and N-desmethyltamoxifen were synthesized following reported procedures [30,31]. Spectroscopic characterization of the isolated products by mass spectrometry and nuclear magnetic resonance (nmr) were in agreement with the literature reports. 2.9. Electrophoresis and immunoblotting Microsomal protein fractions were prepared at 4 ◦ C from unincubated specimens of human endometrium as described by Drahushuk et al. [32], essentially following the protocol of Cinti et al. [33]. Protein concentrations were determined by the method of Bradford [34] using bovine serum albumin as a standard. The microsomal preparations were stored at −70 ◦ C until further analysis. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and western immunoblotting procedures were performed generally as described
earlier [32,35]. CYP1B1, CYP2C9, CYP2D6 and CYP3A proteins were quantitated by western immunoblotting as described by Drahushuk et al. [32] using chemiluminescent detection (CDP-Star; Tropix, Bedford, MA). Polyclonal antibody to CYP1B1 was generously provided by T.R. Sutter, University of Memphis, Memphis, TN. Monoclonal antibody to human CYP1A1 was generously provided by J. Stegeman, Woods Hole Institute, Woods Hole, MA. Specific and sensitive monoclonal antibody to human CYP2D6 and polyclonal antibody to human CYP2C9 and CYP3A4/3A7 were obtained from BD Gentest.
3. Results The biotransformation of tamoxifen was investigated using an explant culture model of human endometrial tissue. Routine histopathological analysis was conducted on fresh, unincubated tissue, as well as on explants cultured for 24 h under various conditions. The routine morphological analysis was necessary to assess not only tissue viability, but also to assess the homogeneity of tissue samples, including contamination by myometrium. Most cultured specimens were greater than 90% endometrium and were suitable for investigation. Endometrium remained viable in specimens that were cultured for up to 24 h in medium containing up to 100 M tamoxifen (see IHC analysis of CYPs, Fig. 5). Authentic standards of tamoxifen metabolites were analyzed by HPLC using online, postcolumn photochemical activation. Fig. 1 shows the reversed-phase HPLC separation of a mixture of ␣-hydroxytamoxifen (1), 4-hydroxytamoxifen (2), N-desmethyltamoxifen (3) and tamoxifen (4) (structures shown in Fig. 2). The elution order was based on polarity of the compounds, as expected. The intra- and interday variation in retention time and fluorescence signal (integrated peak area) of the peaks were within 3%. The signal for each compound was linear over a range of 1–100 ng/ml (correlation coefficient 0.999, n = 3). The flow velocity of the eluent through the capillary, and as a result, the residence time of the analytes in the irradiated zones appear to have some influence on the fluorescence signals (results not shown). The elution conditions were optimized to give the best resolution of the tamoxifen derivatives, not only from one
M. Sharma et al. / Chemico-Biological Interactions 146 (2003) 237–249
241
Table 1 Tamoxifen metabolites produced following 24 h of incubation of 100 M tamoxifen with explant cultures of human endometrium Metabolite
Media (pmol/ml)
Tissue lysate (pmol/mg)
␣-OH tamoxifen 4-OH tamoxifen N-desmethyltamoxifen
69.7 ± 61.4 128.8 ± 66.3 195.8 ± 73.3
1.2 ± 0.9 1.3 ± 0.8 24.9 ± 20.4
Values represent the mean ± S.D. of 15 individual specimens.
Fig. 1. HPLC profile of tamoxifen and authentic standards of three tamoxifen metabolites are shown using postcolumn, online, photochemical activation and fluorescence detection. Peaks 1–4 correspond to ␣-hydroxytamoxifen, 4-hydroxytamoxifen, N-desmethyltamoxifen and tamoxifen, respectively. HPLC conditions are described under Section 2.
another, but also from a large background of tamoxifen usually found in the exposed biological samples. Under these conditions (see Section 2), the lower limit of detection (S/N = 3) was 0.25, 0.13 and 0.14 pmol for 4-hydroxytamoxifen, ␣-hydroxytamoxifen and N-desmethyltamoxifen, respectively. Fig. 3 illustrates HPLC profiles of the metabolites extracted from the explants of human endometrial tissue exposed for 24 h to 0, 25, and 100 M tamoxifen in culture. In the explants exposed to 25 M tamoxifen (profile a), three metabolites (␣-hydroxytamoxifen, 4-hydroxytamoxifen and N-desmethyltamoxifen) were identified by cochromatography with the authentic standards (profile b). These peaks were not present in the control explants cultured under the same conditions using the vehicle (0.05% ethanol) (profile d). The metabolites were detected more readily using 100 M tamoxifen exposure (profile c). HPLC profiles of tamoxifen metabolites extracted from the media of the above cultures are identical to those illustrated in Fig. 3 (results not shown).
The effect of the incubation period on the stability of the metabolites detected in Fig. 3 was further determined by incubating a mixture of the authentic metabolites of known concentration (5 ng/ml) in the culture medium at 37 ◦ C. Aliquots were withdrawn at 0, 24 and 48 h of incubation, extracted and analyzed as described in Section 2. At each of the time points, the recoveries of each of the three metabolite standards ranged from 92 to 96%, demonstrating that the metabolites were stable for up to 48 h under the incubation conditions used in the experiments. Table 1 summarizes data on the relative quantity of each of the major metabolites of tamoxifen formed following 24 h of incubation of the explant cultures of human endometrial tissue with 100 M tamoxifen. The extraction efficiency of this procedure, determined by adding authentic standards of known concentration to the vehicle-exposed tissue lysates, was satisfactory (> 90%). Although all three metabolites were readily detectable, N-desmethyltamoxifen was found to be the major metabolite in both media and whole tissue lysates. Table 2 summarizes data on the relative ability of several recombinant human CYPs to metabolize tamoxifen (25 M). CYP2C9 and CYP2D6 produce all three tamoxifen metabolites, while CYP1A1 and CYP3A4 catalyzed the formation of the ␣-hydroxy and N-desmethyl metabolites but not the 4-hydroxy metabolite. CYP1A2 and CYP1B1 only catalyzed the formation of ␣-hydroxytamoxifen. CYP2D6 clearly exhibits the greatest activity for the formation of all three tamoxifen metabolites. Microsomal protein from uncultured human endometrial specimens was analyzed for the presence of specific CYPs using western immunoblotting (Fig. 4). CYP1A1, 1B1, 2C9, CYP3A proteins were detected in microsomal protein specimens from individual human endometrial samples from three of three donors.
242
M. Sharma et al. / Chemico-Biological Interactions 146 (2003) 237–249
Fig. 2. Structures of tamoxifen and metabolites.
M. Sharma et al. / Chemico-Biological Interactions 146 (2003) 237–249
243
Fig. 3. HPLC chromatographs of extracts from whole-tissue lysates of human endometrial explants exposed in culture for 24 h to media containing (a) 25 M tamoxifen; (b) cochromatography of (a) with authentic standards from Fig. 1; (c) 100 M tamoxifen; (d) vehicle only. Incubation and extraction conditions are described under Section 2. Chromatographic conditions are the same as in Fig. 2.
Table 2 Biotransformation of tamoxifen to ␣-hydroxytamoxifen, 4-hydroxytamoxifen, and N-desmethyltamoxifen by recombinant human CYPs
CYP1A1 CYP1A2 CYP1B1 CYP2C9 CYP2D6 CYP3A4
␣-HydroxyTAM
4-HydroxyTAM
N-desmethylTAM
80 ± 40.44 12.38 ± 1.60 8.73 ± 6.66 7.18 ± 4.18 173.05 ± 20.79 63.50 ± 10.91
ND ND ND 21.46 ± 1.78 614.23 ± 42.93 ND
15.39 ± 31.67 ND ND 25.94 ± 11.81 639.84 ± 67.16 123.55 ± 18.04
The rates of metabolite production are expressed in pmol/min per nmol CYP (mean ± S.D. of triplicate experiments). Incubation of tamoxifen with microsomes from lymphoblastoid cell lines only expressing vector (control+vector) served as a negative control. Background product formation from this control preparation was subtracted from all incubations containing recombinant CYPs to correct for any background oxidation not mediated by CYPs.
244
M. Sharma et al. / Chemico-Biological Interactions 146 (2003) 237–249
and B). The negative controls for the samples were stained with negative-control antibody in place of the antibody against specific human CYPs (Fig. 5A, C and E). All specimens were counterstained with Mayer’s hematoxylin (purple–blue stain). Figs. 5 and 6 also show that the morphology of the explants and the expression of these CYPs were generally well retained during the 24 h incubation period in the presence of 0, 25 or 100 M tamoxifen. While in culture, the stroma becomes more edemic and less dense, while the epithelial glands retain the CYPs (red staining) in the cytoplasm. Together, these results indicate that the human endometrial explant culture model retains CYPs that may contribute to the tissue-specific metabolism of tamoxifen.
Fig. 4. Representative western immunoblots are shown for human endometrium and human liver microsomal protein using specific, sensitive antibodies to human CYP1A1, CYP1B1, CYP2C9, CYP2D6, and CYP3A4/3A7. Lanes 1–4 contain 0.01, 0.025, 0.05 and 0.1 pmol of CYP1B1, 2D6, 3A4, and 3A7 supersome standards (Gentest), respectively. For the CYP1A1, and 2C9 standard, lanes 1–4 contain 0.1, 0.25, 0.5, and 1.0 pmol, respectively. Lanes 5–7 contain 50 g of human endometrium microsomal protein from three different donors and lane 8 contains 50 g of human liver microsomal protein. Note that the antibody to human CYP3A4/3A7 is sensitive and specific for purified supersome standards containing CYP3A4 and CYP3A7.
With the exception of CYP1A1 and 1B1, these CYPs were also readily detected in human liver microsomes. In contrast, CYP2D6 could not be detected in the human endometrial specimens, although it was found at low levels in human liver microsomes. CYP1A2 was also not detected in microsomal protein from human endometrium (results not shown). The expression and retention of CYP2C9, 1B1, and 3A proteins were assessed by IHC analysis after 0 and 24 h exposures of human endometrial explant cultures to media containing 0, 25 or 100 M tamoxifen. The morphological analysis indicates that the endometrium contains a number of epithelial glands and that the tissue remains viable for 24 h in culture at tamoxifen concentrations up to 100 M (Fig. 5). IHC analysis indicates that explant cultures of human endometrium express CYP2C9, 1B1, and 3A constitutively, as illustrated by the presence of red stain in the stroma and epithelial glands (Fig. 5B, Fig. 6A
4. Discussion Although long-term tamoxifen therapy is associated with increased risk of endometrial cancer, little is known about the ability of endometrial tissue to biotransform tamoxifen to potentially reactive intermediates, capable of forming DNA adducts. The present study was designed to examine whether explant cultures of human endometrium can provide a suitable in vitro model to investigate the tissue-specific biotransformation of tamoxifen. This study used an explant culture model, developed by Osteen et al. [25], that uses fresh human endometrial biopsy tissue, containing both epithelial and stromal cells, similar in morphology to that observed in vivo. This model was recently used to investigate the effect of in vitro exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) on the expression of CYP1A1 and CYP1B1 in explant cultures of human endometrium [26]. The advantages of this model are the normal mixture of epithelial and stromal cells contained in the explants, and the ability to conduct morphological studies on the explants. The greatest disadvantage of the model is the small sample size, which limits extensive time and concentration-dependent metabolism studies, using drugs such as tamoxifen. In the present study, routine morphological analyses indicated that endometrial explants remain viable in specimens cultured for up to 24 h in medium containing 25 or 100 M tamoxifen (Figs. 5 and 6). For explant metabolism studies,
M. Sharma et al. / Chemico-Biological Interactions 146 (2003) 237–249
245
Fig. 5. IHC localization of CYP2C9 protein (red stain) in a fresh human endometrium specimen shown before (B) and after 24 h of explant culture in the absence (D) and presence of 100 M tamoxifen (F). Negative controls for the specimens were stained with a UPC negative-control antibody and are shown before (A) and after 24 h of explant culture in the absence (C) and presence of 100 M tamoxifen (E). The slides were counterstained with Mayer’s hematoxylin (purple blue). The images were photographed at an initial magnification of 200 ×. The scalar bar represents 50 M.
a 24 h incubation period with tamoxifen was chosen to optimize the CYP-mediated metabolism of tamoxifen, since CYP activities decrease with incubation time in primary culture models. IHC analysis indicated that explant cultures of human endometrium express CYP2C9, 3A4/3A7 and 1B1 constitutively (Figs. 5 and 6) and that the expression of these CYPs was generally well retained during 24 h in explant culture. Western immunoblot analysis also confirmed the presence of these immunoreactive CYP proteins, including low levels of CYP1A1, in microsomes from
human endometrium (Fig. 4). Thus, explant culture of human endometrium is a suitable in vitro model to investigate the CYP-dependent metabolism of tamoxifen in this target tissue. The expression of immunoreactive CYP1B1 in human endometrium is supported by an earlier study that reported the constitutive expression of CYP1B1 protein in explant cultures of human endometrial tissue, similar to that in uncultured biopsies of this tissue [26]. Hukkanen et al. [22] detected mRNA for CYPs 2C, 3A4, and 3A5 in human endometrium, but
246
M. Sharma et al. / Chemico-Biological Interactions 146 (2003) 237–249
Fig. 6. The IHC localization of CYP1B1 protein (red stain) in a fresh human endometrium specimen is shown before (A) and after 24 h of explant culture in the absence (C) and presence of 25 M tamoxifen (E). Panels on the right side illustrate the IHC localization of CYP3A protein (red stain) in a fresh human endometrium specimen shown before (B) and after 24 h of explant culture in the absence (D) and presence of 25 M tamoxifen (F). The slides were counterstained with Mayer’s hematoxylin (purple blue). The images were photographed at an initial magnification of 200 ×. The scalar bar represents 50 M.
did not detect message for CYPs 1A1, 1A2, 2A6, 2D6, 2F1, 3A7, and 19. In a related study, Vadlamuri et al. [23] reported that mRNA for CYP1A1 and CYP1B1 were constitutively expressed in human endometrial tissue. Although CYP3A7 is considered a human fetal liver CYP, Sarkar et al. [24] readily detected mRNA for CYP3A4 and CYP3A7 in human endometrium. In agreement with our IHC results in Fig. 6, in situ hybridization studies of Sarkar et al. [24] found that CYP3A7 expression was lo-
calized within the glandular epithelium as well as to some extent in the stroma surrounding the glandular structures. These investigators also found that the expression of CYP3A7 message in proliferative phase endometrium was significantly greater than in secretory phase tissue. Together, the above studies support our observations that immunoreactive CYP1A1, CYP1B1, CYP2C9 and CYP3A4/3A7 proteins are constitutively expressed in human endometrium.
M. Sharma et al. / Chemico-Biological Interactions 146 (2003) 237–249
Small sample size is a factor that often limits the determination of the role of tissue-specific biotransformation of chemotherapeutic agents in extra-hepatic tissues. In the present study, a limited amount of human endometrial tissue from explant culture (10–50 mg/well) was used to demonstrate the tissue-specific biotransformation of tamoxifen in this target tissue. Tamoxifen metabolites were measured both in the lysates of the cultured explants and in the culture media using HPLC with postcolumn, online, photochemical activation as a sensitive analytical tool. This sensitive tool enables the collection of data despite the limitations of small sample size. The quantity of metabolites was lower in the whole tissue lysates than in the culture media. This may be due to the irreversible binding of tamoxifen metabolites with proteins [36–38]. Although all three metabolites were readily detectable, N-desmethyltamoxifen was found to be the major metabolite in both media and whole tissue lysates. Similarly, N-desmethyltamoxifen was also found to be the major metabolite in endometrial samples collected during diagnostic hysteroscopy of breast cancer patients exposed to chronic tamoxifen therapy [39]. The considerable variation in the metabolite yields in both the culture media and the whole tissue lysates of the explants (Table 1) reflects the biological variation in the specimens from different individuals and possible variability in metabolism in proliferative and secretory phase tissue specimens. Incubation of tamoxifen with recombinant human CYP2C9 and CYP2D6 produced all three tamoxifen metabolites, while CYP1A1 and CYP3A4 only catalyzed the formation of the ␣-hydroxy and N-desmethyl metabolites (Table 2). CYP1A2 and CYP1B1 only mediated the formation of ␣-hydroxytamoxifen. CYP2D6 clearly exhibited the greatest activity for the formation of all three tamoxifen metabolites, however, the CYP2D6 protein was not detected in microsomes from three out of three human endometrial specimens. CYP2C9, CYP3A and CYP1B1 were detected by IHC analysis of explant cultures of human endometrium (Figs. 5 and 6) and these CYPs, in addition to CYP1A1, were detected by western immunoblotting of human endometrial microsomal protein (Fig. 4). Together, these results support the role of CYPs 2C9, 3A, 1A1, and 1B1 in the biotransformation of tamoxifen in human endometrium.
247
This conclusion is generally supported by earlier work [40,41], which reported that CYP2D6, 2C9 and 3A4, but not 1A1, 1A2, 2C19 and 2E1, contributed to the 4-hydroxylation of tamoxifen by human liver microsomes. Mani et al. [38] reported results utilizing rat and human liver microsomal protein that suggests that CYP3A activates tamoxifen to a reactive intermediate. The CYP3A family has also been reported to catalyze the N-demethylation of tamoxifen in human liver microsomes [42]. Recently, Boocock et al. [43] investigated the metabolism of tamoxifen by recombinant human CYPs and reported that CYP3A4 was the only form that catalyzed the formation of ␣-hydroxytamoxifen and the subsequent formation of tamoxifen–DNA adducts. While we also observed that CYP3A4 catalyzed the formation of ␣-hydroxytamoxifen, the results in Table 2 also show that ␣-hydroxytamoxifen was formed by CYPs 1A1, 1A2, 1B1, 2C9, and 2D6. Differences in source of enzyme, incubation conditions and/or analytical methods may in part explain the lack of complete agreement between these results. In agreement with our findings, Boocock et al. [43] also reported that CYP2D6 was the most active form in catalyzing the overall biotransformation of tamoxifen. However, while CYP2D6 may play a role in the hepatic metabolism of tamoxifen, particularly in individuals with higher hepatic levels of this enzyme, the lack of detectable CYP2D6 in the endometrium suggests that this enzyme may not play a significant role in tamoxifen metabolism in this target tissue. The genotoxicity of N-desmethyltamoxifen in rat liver, through a metabolic pathway involving ␣-hydroxylation, has been reported both in vitro and in vivo [44]. Although no DNA adducts of N-desmethyltamoxifen have been detected in extrahepatic tissue [5], the possibility of forming other types of DNA damage, such as DNA cleavage and 8-hydroxyguanosine formation cannot be ruled out [45]. Alpha-hydroxytamoxifen has been reported to be the major DNA-reactive metabolite in rat liver, thus contributing to the carcinogenic activity of tamoxifen in the rat and possibly the human [11–13]. DNA from the endometrium of women treated with tamoxifen was analyzed by a 32 P-postlabeling/HPLC on-line assay capable of detecting 2.5 adducts/1010 nucleotides [10]. With this sensitive assay, specimens from 8 of 16 subjects had detectable tamoxifen–DNA
248
M. Sharma et al. / Chemico-Biological Interactions 146 (2003) 237–249
adducts, providing further support for the genotoxicity of tamoxifen. This is in contrast to a similar study that found no convincing evidence for tamoxifen derived DNA adducts in human endometrium [8]. In summary, our results support the use of explant cultures of human endometrium as a suitable in vitro model to investigate the biotransformation of tamoxifen in this target tissue. In addition, the results support the role of CYPs 2C9, 3A, 1A1, and 1B1 in the biotransformation of tamoxifen, including the formation of the DNA reactive ␣-hydroxytamoxifen metabolite, in human endometrium. We are currently investigating the biotransformation and the potential formation of tamoxifen–DNA adducts in explant cultures of human endometrium and the effect of anti-oxidants on modulating these processes. This model will also provide a means to investigate mechanisms for the interindividual differences in the bioactivation of tamoxifen, which will assist future efforts to identify individuals that may be at greater risk of endometrial cancer will tamoxifen therapy.
Acknowledgements This study has been supported in part by grant CA 86875 from the NCI. We thank the Tissue Procurement Facility of Roswell Park Cancer Institute for providing the human endometrium specimens.
References [1] N. Colacurci, L. De Seta, P. De Franciscis, D. Mele, N. Fortunato, S. Cassese, Tamoxifen effects on endometrium, Panminerva Med. 42 (2000) 45–47. [2] B. Czernobilsky, B. Lifschitz-Mercer, Endometrial pathology, Curr. Opin. Obstet. Gynecol. 9 (1997) 52–56. [3] E. Mantyla, S. Karlsson, L. Nieminen, Induction of endometrial cancer by tamoxifen in the rat, in: J.J. Li, S.A. Li, J.A. Gustafsson, N. Nandi, L.I. Sekely (Eds.), Hormonal Carcinogenesis, vol. II, Springer, New York, 1996, pp. 442–445. [4] P. Hirsimaki, Y. Hirsimaki, L. Nieminen, S. Karlsson, E. Mantyla, On the safety studies of the antiestrogens toremifene and tamoxifen, Scand. J. Lab. Anim. Sci. 23 (1996) 147–153. [5] G. Gamboa da Costa, L.P. McDaniel-Hamilton, R.H. Heflich, M.M. Marques, F.A. Beland, DNA adduct formation and mutant induction in Sprague–Dawely rats treated with tamoxifen and its derivatives, Carcinogenesis 22 (2001) 1307– 1327.
[6] M.R. Osborne, A. Hewer, I.R. Hardcastle, P.L. Carmichael, D.H. Phillips, Identification of the major tamoxifen– deoxyguanosine adduct formed in the liver DNA of rats treated with tamoxifen, Cancer Res. 56 (1996) 66–71. [7] P.L. Carmichael, A.H. Ugwumadu, P. Neven, A.J. Hewer, G.K. Poon, D.H. Phillips, Lack of genotoxicity of tamoxifen in human endometrium, Cancer Res. 56 (1996) 1475–1479. [8] P.L. Carmichael, S. Sardar, N. Crooks, P. Neven, I. Van Hoof, A. Ugwumadu, T. Bourne, E. Tomas, P. Hellberg, A.J. Hewer, D.H. Phillips, Lack of evidence from HPLC 32 P-postlabelling for tamoxifen DNA adducts in the human endometrium, Carcinogenesis 20 (1999) 339–342. [9] S. Shibutani, N. Suzuki, I. Terashima, S.M. Sugarman, A.P. Grollman, M.L. Pearl, Tamoxifen–DNA adducts detected in the endometrium of women treated with tamoxifen, Chem. Res. Toxicol. 12 (1999) 646–653. [10] S. Shibutani, A. Ravindernath, N. Suzuki, I. Terashima, S.M. Sugarman, A.P. Grollman, M.L. Pearle, Identification of tamoxifen–DNA adducts in the endometrium of women treated with tamoxifen, Carcinogenesis 21 (2000) 1461–1467. [11] S. Shibutani, L. Dasaradhi, I. Terashima, E. Banoglu, M.W. Duffel, ␣-Hydroxytamoxifen is a substrate of hydroxysteroid (alcohol) sulfotransferase resulting in tamoxifen DNA adducts, Cancer Res. 58 (1998) 647–653. [12] H. Glatt, W. Davis, W. Meinl, H. Hermersdorfer, S. Venitt, D.H. Phillips, Rat, but not human, sulfotransferase activates a tamoxifen metabolite to produce DNA adducts and gene mutations in bacteria and mammalian cells in cultures, Carcinogenesis 19 (1998) 1709–1713. [13] S. Shibutani, P.M. Shaw, N. Suzuki, L. Dasaradhi, M.W. Duffel, I. Terashima, Sulfation of ␣-hydroxytamoxifen catalyzed by human hydroxysteroid sulfotransferase results in tamoxifen–DNA adducts, Carcinogenesis 19 (1998) 2007– 2011. [14] D.H. Phillips, A. Hewer, M.N. Horton, K.J. Cole, P.L. Carmichael, W. Davis, M.R. Osborne, N-demethylation accompanies ␣-hydroxylation in the metabolic activation of tamoxifen in rat liver cells, Carcinogenesis 20 (1999) 2003– 2009. [15] G.A. Potter, R. McCague, M. Jarman, A mechanistic hypothesis for DNA adduct formation by tamoxifen following hepatic oxidative metabolism, Carcinogenesis 15 (1994) 439– 442. [16] B. Moorthy, P. Sriram, D.N. Pathak, W.J. Bodell, K. Randerath, Tamoxifen metabolic activation: comparison of DNA adducts formed by microsomal and chemical activation of tamoxifen and 4-hydroxytamoxifen with DNA adducts formed in vivo, Cancer Res. 56 (1996) 53–57. [17] D.N. Pathak, K. Pongracz, W.J. Bodell, Activation of 4-hydroxytamoxifen and the tamoxifen derivative metabolite E by uterine peroxidase to form DNA adducts: comparison with DNA adducts formed in the uterus of Sprague–Dawley rats treated with tamoxifen, Carcinogenesis 17 (1996) 1785– 1790. [18] M.M. Marques, F.A. Beland, Identification of tamoxifen– DNA adducts formed by 4-hydroxytamoxifen quinone methide, Carcinogenesis 18 (1997) 1949–1954.
M. Sharma et al. / Chemico-Biological Interactions 146 (2003) 237–249 [19] L. Dasaradhi, S. Shibutani, Identification of tamoxifen–DNA adducts formed by ␣-sulfate tamoxifen and ␣-acetoxytamoxifen, Chem. Res. Toxicol. 10 (1997) 189–196. [20] C. Sanchez, S. Shibutani, L. Dasaradhi, J.L. Bolton, P.W. Fan, R.A. McClelland, Lifetime and reactivity of an ultimate tamoxifen carcinogen: the tamoxifen carbocation, J. Am. Chem. Soc. 120 (1998) 13513–13514. [21] E.A. Lien, E. Solheim, P.M. Ueland, Distribution of tamoxifen and its metabolites in rat and human tissue during steady state treatment, Cancer Res. 51 (1991) 4837–4844. [22] J. Hukkanen, M. Mantyla, L. Kangas, P. Wirta, J. Hakkola, P. Paakki, S. Evisalmi, O. Pelkonen, H. Raunio, Expression of cytochrome P450 genes encoding enzymes active in the metabolism of tamoxifen in human uterine endometrium, Pharmacol. Toxicol. 82 (1998) 93–97. [23] S.V. Vadlamuri, D.D. Glover, T. Turner, M.A. Sarkar, Regiospecific expression of cytochrome P4501A1 and 1B1 in human uterine tissue, Cancer Lett. 122 (1998) 143–150. [24] M.A. Sarkar, V. Vadlamuri, S. Ghosh, D.D. Glover, Expression and cyclic variability of CYP3A4 and CYP3A7 isoforms in human endometrium and cervix during the menstrual cycle, Drug Metab. Dispos. 31 (2001) 1–6. [25] K.G. Osteen, W.H. Rodgers, M. Gaire, J.T. Hargrove, F. Gorstein, L.M. Matrisian, Stromal–epithelial interaction mediated steroidal regulation of metalloproteinase expression in human endometrium, Proc. Natl. Acad. Sci. USA 91 (1994) 10129–10133. [26] D.P. Bofinger, L. Feng, L.H. Chi, J. Love, F.D. Stephen, T.R. Sutter, K.G. Osteen, T.G. Costich, R.E. Batt, S.T. Koury, J.R. Olson, Effect of TCDD exposure on CYP1A1 and CYP1B1 in explant cultures of human endometrium, Toxicol. Sci. 62 (2001) 299–314. [27] O. Merle, J. Guitton, M.D. Burke, M. Ollangnier, Rapid and simple method to determine tamoxifen and its major metabolites in human liver microsomes by high performance liquid chromatography with fluorescence detection, Anal. Lett. 31 (1998) 2067–2076. [28] K.M. Fried, I.W. Wainer, Direct determination of tamoxifen and its four major metabolites in plasma using coupled column high-performance metabolites in plasma using coupled column high-performance liquid chromatography, J. Chromatogr. B 655 (1994) 261–268. [29] M. Sharma, Analysis of tamoxifen–DNA adducts by high-performance liquid chromatography using postcolumn online photochemical activation, Biochem. Biophys. Res. Commun. 273 (2000) 40–44. [30] A.B. Foster, M. Jarman, O.-T. Leung, R. McCague, G. Leclercq, N. Devleeschouwer, Hydroxy derivatives of tamoxifen, J. Med. Chem. 28 (1985) 1491–1497. [31] K. Brown, R.T. Heydon, R. Jukes, I.N. White, E.A. Martin, Further characterization of the DNA adducts formed in rat liver after the administration of tamoxifen, N-desmethyltamoxifen or N,N-didesmethyltamoxifen, Carcinogenesis 20 (1999) 2011–2066. [32] A.T. Drahushuk, B.P. McGarrigle, H.-L. Tai, S. Kitareewan, J.A. Goldstein, J.R. Olson, Validation of precision-cut liver slices in dynamic organ culture as an in vitro model for
[33]
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
249
studying CYP1A1 and CYP1A2 induction, Toxicol. Appl. Pharmacol. 140 (1996) 393–403. D.L. Cinti, P. Moldeus, J.B. Schenkman, Kinetic parameters of drug-metabolizing enzymes in Ca2+ -sedimented microsomes from rat liver, Biochem. Pharmacol. 21 (1972) 3249– 3256. M.M. Bradford, A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein–dye binding, Anal. Biochem. 72 (1976) 248–254. V.K. Laemmli, Cleavage of structural protein during the assembly of the head of bacteriophage T4, Nature 227 (1970) 680–685. S.S. Dehal, D. Kupfer, Evidence that the catechol 3,4-dihydroxytamoxifen is a proximate intermediate to the reactive species binding covalently to proteins, Cancer Res. 56 (1996) 1283–1290. C. Mani, D. Kupfer, Cytochrome P-450-mediated activation and irreversible binding of the antiestrogen tamoxifen to proteins in rat and human liver: possible involvement of flavin-containing monooxygenases in tamoxifen activation, Cancer Res. 51 (1991) 6052–6058. C. Mani, R. Pierce, A. Parkinson, D. Kupfer, Involvement of cytochrome P-450 3A in catalysis of tamoxifen activation and covalent binding to rat and human liver microsomes, Carcinogenesis 15 (1994) 2715–2720. G. Giorda, L. Franceschi, D. Crivellari, M.D. Magri, A. Veronesi, C. Scarabelli, M. Furlanut, Determination of tamoxifen and its metabolites in the endometrial tissue of long-term treated women, Eur. J. Cancer 36 (Suppl. 4) (2000) S88–S99. H.K. Crew, S.W. Ellis, M.S. Lennard, G.T. Tucker, Variable contribution of cytochrome P450 2D6, 2C9 and 3A4 to the 4-hydroxylation of tamoxifen by human liver microsomes, Biochem. Pharmacol. 53 (1997) 171–178. S.S. Dehal, D. Kupfer, CYP2D6 catalyzes tamoxifen 4-hydroxylation in human liver, Cancer Res. 57 (1997) 3402– 3406. F. Jacolot, I. Simon, Y. Dreano, P. Beaune, C. Riche, F. Berthou, Identification of the cytochrome P450 IIIA family as the enzymes involved in the N-demethylation of tamoxifen in human liver microsomes, Biochem. Pharmacol. 41 (1991) 1911–1919. D.J. Boocock, K. Brown, A.H. Gibbs, E. Sanchez, K.W. Turteltaub, I.N. White, Identification of human CYP forms involved in the activation of tamoxifen and irreversible binding to DNA, Carcinogenesis 23 (2002) 1897–1901. G. Gamboa da Costa, L.P. Hamilton, F.A. Beland, M.M. Marques, Characterization of the major DNA adduct formed by ␣-hydroxy-N-desmethyltamoxifen in vitro and in vivo, Chem. Res. Toxicol. 13 (2000) 200–207. T. Okubo, F. Nagai, K. Ushiyama, Y. Yokoyama, S. Ozawa, K. Kano, S. Tomita, H. Kubo, I. Kano, DNA cleavage and 8-hydroxydeoxyguanosine formation caused by tamoxifen derivative in vitro, Cancer Lett. 122 (1998) 9–15.