Human CYP1B1 Is Regulated by Estradiol via Estrogen Receptor

Cytochrome P450s (CYP) are a multigene family of constitutive and inducible enzymes involved in the oxidative metabolic activation and detoxification of many endogenous and exogenous compounds (1, 2). CYP1B1 is a comparatively new isoform in the CYP1 family (3). Whereas CYP1B1 is capable of activating a variety of procarcinogens and promutagens including polycyclic aromatic hydrocarbons and aryl amines (4), it is induced by various polycyclic aromatic hydrocarbons as well as 2,3,7,8-tetrachlorodibenzo-p-dioxin (5). The induction is controlled by the aryl hydrocarbon receptor (AhR) and AhR nuclear translocator (6, 7). In our previous study (8), two xenobiotic responsive elements (XREs) that are critical for the induction were identified on the human CYP1B1 gene.

Recent findings suggest that human CYP1B1 is a key enzyme in the metabolism of 17β-estradiol (E2; Refs. 9, 10, 11). Of particular interest is the fact that CYP1B1 is highly expressed in estrogen-related tissues such as mammary, uterus, and ovary (4, 12), suggesting that CYP1B1 is important in the localized metabolic control of estrogen homeostasis. 4-Hydroxyestradiol, a catechol metabolite formed by CYP1B1 from E2, leads to decrease of the estrogenic activity. However, it is toxicologically active and appears to play a role in tumorigenesis, because it generates free radicals from the reductive-oxidative cycling with the corresponding semiquinone and quinone forms, which cause cellular damage (13, 14). It has also been reported that the 4-hydroxyestradiol level is elevated in human endometrial and breast cancers in comparison to normal tissue (15). Thus, the metabolism of the estrogens by CYP1B1 in the uterus may be an important determinant of the toxicological as well as physiological outcomes.

In the present study, we focused on estrogens as a potential regulator of human CYP1B1, because the enzyme is highly expressed in estrogen-regulated tissues. Many CYP isoforms are likely to be induced by the substrates themselves. In addition, computer-assisted homology search identified a potential estrogen receptor (ER) binding site on the CYP1B1 promoter. These backgrounds prompted us to investigate whether the human CYP1B1 gene is a target of E2.

E2 and 4-hydroxytamoxifen (4OHT) were obtained from Sigma-Aldrich (St. Louis, MO). 7α-[9(4,4,5,5,5-Pentafluoropentylsulfonyl)nonyl] estra-1,3,5(10)-triene-3,17β-diol (ICI 182,780) was purchased from Tocris Cookson Inc. (Bullwin, MO). pGL3-basic plasmid, phRL-TK plasmid, Tfx-20 reagent, and a dual-luciferase reporter assay system were from Promega (Madison, WI). pSG5 vector was obtained from Stratagene (La Jolla, CA). ERα expression vector (pSG5-HE0) and mutated ERα expression vectors (HE11, HE19, and HE38) constructed previously (16, 17) were gifts from Dr. Pierre Chambon (Institut de Genetique et de Biologie Moleculaire et Cellulaire, Strasbourg, France). [γ-³²P]ATP was from Amersham (Buckinghamshire, United Kingdom). All of the primers and oligonucleotides were commercially synthesized at Hokkaido System Sciences (Sapporo, Japan). Recombinant human ERα was purchased from PanVera (Madison, WI). Mouse antihuman ERα monoclonal antibodies and normal mouse IgG were from Santa Cruz Biotechnology (Santa Cruz, CA). Rabbit antihuman CYP1B1 polyclonal antibodies were from Alpha Diagnostic International (San Antonio, TX). All of the other chemicals and solvents were of the highest grade commercially available.

The human breast adenocarcinoma cell lines MCF-7 and MDA-MB-435 were obtained from American Type Culture Collection (Rockville, MD). The human endometrial adenocarcinoma cell line Ishikawa was a generous gift of Dr. Masato Nishida (Tsukuba University, Ibaraki, Japan). MCF-7 cells were cultured in DMEM (Nissui Pharmaceutical, Tokyo, Japan) supplemented with 0.1 mm nonessential amino acid (Invitrogen, Melbourne, Australia) and 10% fetal bovine serum (Invitrogen). MDA-MB-435 and Ishikawa cells were cultured in DMEM supplemented with 10% (v/v) fetal bovine serum. In the estrogen-induction assay, these cells were grown in phenol red-free DMEM medium (Invitrogen) containing 10% dextran-coated charcoal-treated fetal bovine serum for 48 h before treatment with estrogen. These cells were maintained at 37°C under an atmosphere of 5% CO₂-95% air.

Human CYP1B1 and glyceraldehyde-3-phosphate dehydrogenase mRNAs were quantified by real-time reverse transcription-PCR. Total RNA was isolated from the cells using ISOGEN (Nippon Gene, Tokyo, Japan), and cDNAs were synthesized as described previously (5). The forward and reverse primers for human CYP1B1 were described previously (5), and those for human glyceraldehyde-3-phosphate dehydrogenase were 5′-CCA GGG CTG CTT TTA ACT C-3′ and 5′-GCT CCC CCC TGC AAA TGA-3′. PCR was performed using the Smart Cycler (Cepheid, Sunnyvale, CA) with Smart Cycler software (Version 1.2b). The PCR condition for glyceraldehyde-3-phosphate dehydrogenase was as follows: after an initial denaturation at 95°C for 30 s, the amplification was performed by denaturation at 94°C for 4 s, annealing and extension at 64°C for 20 s for 45 cycles. The PCR condition for CYP1B1 was as follows: after an initial denaturation at 95°C for 30 s, the amplification was performed by denaturation at 94°C for 10 s, annealing and extension at 68°C for 20 s for 45 cycles. Amplified products were monitored directly by measuring the increase of the dye intensity of the SYBR Green I (Molecular Probes, Eugene, OR) that binds to double-strand DNA amplified by PCR. Copy number of mRNA in the cDNA samples was calculated using standard amplification curves.

pGL3-Basic plasmid containing the 5′-flanking region from −2299 to +25 of the human CYP1B1 gene was kindly provided by Dr. William A. LaMarr of the University of Massachusetts Medical School (Worcester, MA). A series of plasmids containing progressive deletion fragments (−1652/+25, −910/+25, and −152/+25) were constructed previously (8). A plasmid containing the 5′-flanking region from −39 to +25 (−39/+25) was constructed using restriction endonucleases. The orientation of these constructs was verified by restriction enzyme digestion or DNA sequencing. The mutated plasmid of pGL3 (−152/+25) mt was constructed by site-directed mutagenesis with a QuikChange site-directed mutagenesis kit (Stratagene). The forward and reverse mutagenic primers were 5′-CCT GCC AGTTCG CGC TGC ACT CCT TCT ACC-3′ and 5′-GGT AGA AGG AGTGCA GCG CGA ACT GGC AGG-3′, respectively (mutated sites are shown as bold letters). Nucleotide sequences were confirmed by DNA sequencing analyses.

CYP1B1/luc reporter gene plasmid and control (phRL-TK) reporter plasmid were transiently transfected into MCF-7 or Ishikawa cells using Tfx-20 reagent. Into the Ishikawa cells, ERα expression vector or pSG5 vector (control) was also cotransfected. After the incubation with transfection mixture at 37°C under an atmosphere of 5% CO₂-95% air for 1 h, growth medium was added to the cells. After 24 h, the medium was replaced with medium containing 0.1% (v/v) DMSO or 10 nm E2/0.1% DMSO. For inhibition studies, the transfected MCF-7 cells were incubated with 1 μm of 4OHT or ICI 182,780 in the absence or presence of E2. After 24 h-treatment, the cells were resuspended in passive lysis buffer, and then the luciferase activity was measured with a luminometer (Dainippon Pharmaceutical, Osaka, Japan) using the dual-luciferase reporter assay system.

MCF-7 cells were treated with 0.1% DMSO or 10 nm E2 for 12 h, and ChIP assay was carried out essentially using the ChIP assay kit (Upstate, Lake Placid, NY). Anti-ERα antibodies or normal mouse IgG (control) were used for immunoprecipitation of protein-DNA complexes. The forward and reverse primers for region 1 were 5′-TAC CGC ACA ATG GAA ACG T-3′ and 5′-TGA GTG GCG TCA ATT CCC A-3′, respectively. The forward and reverse primers for region 2 were 5′-AGC GGC TAC TTT TAG GGA TTC-3′ and 5′-GGT AAC CGC GCT TCA TCA CA-3′, respectively. PCR analyses were performed as follows: after an initial denaturation at 94°C for 3 min, the amplification was performed by denaturation at 94°C for 20 s, annealing at 55°C (region 1) or 58°C (region 2) for 20 s, and extension at 72°C for 30 s for 34 cycles. The PCR product was electrophoresed on a 2% agarose gel and visualized by ethidium bromide staining.

Synthetic oligonucleotides were labeled with [γ-³²P]ATP using T4 polynucleotide kinase (Toyobo, Osaka, Japan). The reaction mixture contained 3 pmol of the recombinant ERα protein, 2 μg of poly(dI·dC), 1 μg of salmon sperm DNA, and 30 fmol of the radiolabeled probe (∼20,000 cpm) in a final volume of 15 μl of binding buffer [25 mm HEPES-KOH (pH 7.9), 0.5 mm EDTA, 50 mm KCl, 10% glycerol, 0.5 mm DTT, and 0.5 mm (p-amidinophenyl) methanesulfonyl fluoride]. Binding reactions were performed on ice for 30 min. To determine the specificity of the binding to the oligonucleotides, competition experiments were conducted by coincubation with 100- and 500-fold excesses of unlabeled competitors. In some experiments, 2 μg of anti-ERα polyclonal antibodies or normal mouse IgG were preincubated with the recombinant ERα protein on ice for 15 min. The oligonucleotide sequences are shown in Table 1. The consensus oligonucleotide for estrogen response element (ERE) was from the Xenopus vitellogenin A2 gene (18). DNA-protein complexes were separated under nondenaturing conditions on 4% polyacrylamide gels with 0.5 × Tris-borate EDTA as the running buffer. The gels were dried, and then the DNA-protein complexes were detected and quantified with a Fuji Bio-Imaging Analyzer BAS 1000 (Fuji Film, Tokyo, Japan).

Human endometrial tissues were obtained from 20 patients undergoing hysterectomy as a treatment for benign neoplasms not associated with endometrial disease at Kanazawa University Hospital. Written informed consent was obtained from all of the patients. The collected endometrial tissues were sampled for histopathological diagnosis, and the remaining normal portions of the samples were used for immunohistochemistry (19 samples) or ChIP assays (1 sample). The phase of the menstrual cycle was determined by histological examination of the tissues and also confirmed by determination of the day of the menstrual cycle. These samples were confirmed to be derived from nonsmoking patients. For ChIP assay, endometrial glands were obtained from a patient (39 years old) during the proliferative phase as follows: Minced endometrial tissue was placed on plastic dishes in DMEM containing 350 units/ml of DNase I (Takara, Ohtsu, Japan) and 180 units/ml of collagenase type 3 (Washington Biochemical Corporation, Lakewood, NJ), and gently shaken for 40 min at 37°C as described previously (19). Individual glands on the bottom of the dishes were selected under a microscope and were collected into an Eppendorf tube.

Immunohistochemical analyses of CYP1B1 were performed using formalin-fixed, paraffin-embedded specimens of normal human endometrial tissues. The sections were autoclaved in 1 × Antigen Retrieval Solution (Biogenex, San Ramon, CA) for 10 min and then incubated with antihuman CYP1B1 antibodies at 4°C for 16 h. For the antibodies, no significant cross-reactivity to either human CYP1A1 or CYP1A2 protein has been reported (20). Staining reactions were performed using the ABC-elite kit (Vector Laboratories, Burlingame, CA). Staining of glands in functional layers was evaluated as positive when >10% of the constituting epithelial cells in a gland were stained. The results were judged by three independent pathologists as − (<10% of the glands at the region of interest), + (>10%), or ++ (>10% with high intensity).

Data are expressed as mean ± SD. Comparison of two groups was made with an unpaired, two-tailed Student’s t test. In immunohistochemistry, the statistical significance of differences in the extent of staining between the different phases of the menstrual cycle was tested by Fisher’s exact method. A value of P < 0.05 was considered statistically significant.

The effects of E2 treatment on the CYP1B1 mRNA expression were investigated in the human breast adenocarcinoma MCF-7 and MDA-MB-435 cells, and the human endometrial adenocarcinoma Ishikawa cells. These cells were treated with 10 nm E2 for 12 h, and the mRNA expression was examined by real-time reverse transcription-PCR. As shown in Fig. 1, CYP1B1 mRNA was significantly (P < 0.05) induced by E2 treatment in ER-positive MCF-7 cells by ∼3-fold. In contrast, no significant induction was observed in ER-negative MDA-MB-435 cells. These results suggest that E2 induced CYP1B1 mRNA expression in an ER-dependent manner. Although Ishikawa cells are ER positive, the induction of CYP1B1 by E2 was not observed.

To investigate the transcriptional activation of the human CYP1B1 gene by E2, luciferase assays were performed in which a series of reported plasmids containing the 5′-flanking region of the human CYP1B1 gene were transiently transfected into MCF-7 cells (Fig. 2,A). The reporter activity of the pGL3 (−2299/+25) plasmid was significantly (P < 0.05) induced by 2-fold by E2 treatment (Fig. 2 A). Deletion from the 5′-end to −152 led to 7-fold activation of the transcription (P < 0.01). In contrast, the transcriptional activation was not observed with the pGL3 (−39/+25) plasmid. These results suggest the existence of an E2-responsive region between −152 and −39.

The ER dependency of the transcriptional activation by E2 was examined. MCF-7 cells transfected with the pGL3 (−152/+25) plasmid were treated with ER antagonists, 4OHT or ICI 182,780, in the absence or presence of E2 for 24 h (Fig. 2 B). In the absence of E2, the reported activity was significantly (P < 0.05) increased by 4OHT or ICI 182,780 treatments. In contrast, the transcriptional activity enhanced by E2 was significantly (P < 0.05) inhibited by 4OHT or ICI 182,780. These results suggest that ER could be involved in the transcriptional activation of the human CYP1B1 by E2.

A computer-assisted homology search revealed imperfect palindromic ERE to which ERα may bind in the human CYP1B1 gene between −63 and −49 (Fig. 3,A). The sequence (AGGTCGCGCTGCCCT) is different from the consensus ERE (AGGTCANNNTGACCT) in two bases (Table 1). To determine whether ERα is associated with the putative ERE on the human CYP1B1 gene, ChIP assays were performed. DNA was extracted from MCF-7 cells treated with DMSO or E2 for 12 h after cross-linking treatment and incubated with specific anti-ERα antibodies or normal mouse IgG. As shown in Fig. 3,B, PCR was performed with a primer set for region 1 (−209 to +81) including the putative ERE or a primer set for region 2 (−534 to −206). In the presence of E2, immunoprecipitants obtained with anti-ERα antibodies generated a distinct PCR product with the primer set for region 1, whereas those obtained with mouse IgG generated no product (Fig. 3 C). In the absence of E2, no distinct band was observed. These results indicated that the incubation with E2 facilitated the binding of ERα to the putative ERE. Using the primer set for the region 2, immunoprecipitants obtained with anti-ERα antibodies or mouse IgG did not generate the PCR product in the absence or presence of E2. These results suggest that ERα is associated with the putative ERE on the human CYP1B1 gene.

To examine whether ERα binds to the putative ERE on the human CYP1B1 gene, gel shift assays were performed (Fig. 4). Oligonucleotides used for the gel shift analyses are shown in Table 1. We first confirmed the specific binding complex formed with recombinant ERα and consensus oligonucleotide for ERE used as a probe (Fig. 4,A). This band was competed out by the putative ERE on human CYP1B1 gene (1B1/ERE) used as a cold competitor. The specific band was also observed with the recombinant ERα and 1B1/ERE used as a probe that was super-shifted by anti-ERα antibodies and competed out by consensus oligonucleotide for ERE and 1B1/ERE but not by mutant 1B1/ERE (Fig. 4,B). No significant band was observed with mutant 1B1/ERE as a probe (Fig. 4 C). These results indicated that ERα specifically binds to the putative ERE on the human CYP1B1 gene.

To investigate whether the transcriptional activation of the human CYP1B1 gene is ERα-dependent or not, luciferase assays were performed using Ishikawa cells that expressed low levels of ERα. Notably, E2 treatment of the Ishikawa cells failed to induce the reporter activity of the pGL3 (−152/+25) plasmid containing the putative ERE. However, when ERα was overexpressed, the reporter activity was significantly (P < 0.01) induced up to 10-fold by the E2 treatment (Fig. 5 A). In contrast, no activation was observed with the pGL3 (−152/+25) mt and pGL3 (−39/+25) plasmids lacking the putative ERE by E2 treatment. Therefore, the putative ERE located between −63 and −49 appeared to be indispensable for the E2-dependent transactivation.

To confirm the requirement of ERα for the transactivation of CYP1B1, various mutant ERα expression vectors were cotransfected with pGL3 (−152/+25) plasmid into Ishikawa cells (Fig. 5 B). The transactivation by E2 was 9-fold with wild-type ERα (P < 0.05). However, the transactivation by E2 was not observed with the coexpressions of HE11 and HE38, which expressed mutant protein lacking the DNA-binding domain and the ligand-induced transcriptional activation domain AF-2, respectively. With the HE19 plasmid expressing mutant protein lacking the constitutive transcriptional activation domain AF-1, a significant (P < 0.01) transactivation by E2 was observed (6.7-fold). Therefore, the DNA-binding domain and the ligand-dependent transcriptional activation domain AF-2 appeared to be required for the E2-dependent transcriptional activation of CYP1B1.

Because CYP1B1 is expressed in uterus and the endometrium is highly regulated by estrogens, human endometrial glandular epithelial cells were used for ChIP assays to investigate the association of ERα with the CYP1B1 promoter in intact cells. As shown in Fig. 6, the immunoprecipitant obtained with anti-ERα antibodies generated a distinct PCR product when the primer set for region 1 was used, but not with mouse IgG. Using the primer set for region 2, the immunoprecipitants obtained with anti-ERα antibodies or mouse IgG generated no PCR product. The association of ERα with the putative ERE on the human CYP1B1 gene was demonstrated in human endometrial cells.

The endometrium would be a suitable tissue to investigate the regulation of CYP1B1 by E2. Immunohistochemistry of CYP1B1 was performed in normal endometrial tissues obtained from 19 patients at different phases of the menstrual cycle. Immunostaining of CYP1B1 was detected in both the glandular epithelial or stromal cells (Fig. 7). We evaluated the staining of CYP1B1 in the glandular epithelial cells in functional layers, because they are mostly susceptible to E2. CYP1B1 protein was strongly stained in glandular epithelial cells in the proliferative phase (Fig. 6,A). It was still apparent in the early secretory phase, but with subsequent diminishment with the progression to the secretory phase. Of 11 endometria in the proliferative phase, 6 and 5 samples were judged as ++ and +, respectively (Table 2). Of 5 endometria in the early secretory phase, 2 and 3 samples were judged as ++ and +, respectively. Three endometria in the late secretory phase were judged as −. The expression level of CYP1B1 in the glandular epithelial cells was significantly (P < 0.001) different according to the phase of the menstrual cycle (Table 2).

In the present study, we demonstrated that E2 induced the expression of human CYP1B1 mRNA in ER-positive cells, which is mediated by the direct interaction of ERα with an ERE on the CYP1B1 gene. This is the first study in which the specific activation of estrogen-metabolizing enzyme by estrogen itself was demonstrated. The phenomenon was highly dependent of the function of the ER. Whereas ER-positive MCF-7 cells sensitively responded to E2, ER-negative MDA-MB-435 cells did not. Of particular interest is the poor responsiveness of ER-positive Ishikawa cells to E2. This cell line is well known to express ER but at low levels (21, 22). The overexpression of ERα in this cell line was able to induce CYP1B1 transcriptional activity in the presence of E2. Even MCF-7 cells that constitutively express high levels of ER exhibited dramatic transcriptional activation of CYP1B1 up to 80-fold when ER was overexpressed (data not shown). ER-antagonist 4OHT or ICI 182,780 clearly inhibited the E2-mediated CYP1B1 transcriptional activation. Taken together, these findings support that ER expression plays crucial roles in the E2-mediated activation of human CYP1B1.

It is well known that the CYP1B1 gene is under the regulatory control of the AhR. Many polycyclic aromatic hydrocarbons are known to induce CYP1B1 and their own metabolism through binding to and activation of the AhR. A recent report (23) demonstrated that AhR might be responsible for E2-induced CYP1B1 expression in ER-positive cells. In that study, AhR expression was significantly increased in response to E2 stimulation. Because AhR is known to interact with XRE on the CYP1B1 gene and activate the transcription without exogenous AhR ligand, it was speculated that E2-induced CYP1B1 expression was mediated by the AhR function activated by E2. Our results do not support this hypothesis for the following reasons. First, the pGL3 (−910/+25) plasmid containing two critical XRE sites (8) for AhR-dependent transactivation showed low E2-responsiveness in our reporter assays. Second, the introduction of a mutation in XRE in the pGL3 (−910/+25) plasmid did not affect the E2 responsiveness (data not shown). Third, there is no XRE sequence up to −152 bp in the CYP1B1 gene in which prominent ER responsiveness was observed, and our ChIP assays and gel shift analyses failed to confirm binding of the AhR to the E2-response element on the CYP1B1 promoter. We thus concluded that the ER-mediated pathway, rather than the AhR-pathway, plays essential roles in E2-induced CYP1B1 expression.

It has been reported that some estrogen-regulated genes are indirectly regulated by the cooperation of Sp1 and ER within a GC-box and ERE half site (24, 25). For the human CYP1B1 gene, two Sp1 binding sequences have been reported to be located at −84 and −68 (26) that are near the putative ERE identified in the present study. It will be important to clarify whether Sp1 and ER cooperatively regulate the transcription of the human CYP1B1 gene.

Our immunohistochemical analyses revealed predominant staining of CYP1B1 in endometrial glandular cells in a proliferative phase. The staining was attenuated with the progression to the secretory phase. This pattern of expression is consistent with change in the local concentration of E2 in endometrium (27), increasing E2 levels peaking at the ovulatory phase, with a subsequent decrease in the secretory phase, which supports the presence of an E2-dependent activation pathway of CYP1B1. Furthermore, it was clearly demonstrated with the ChIP assays using endometrial glandular cells that ERα binds to the CYP1B1 promoter in vivo.

An interesting aspect of our result is the potential pathway of E2-induced carcinogenesis in estrogen-target tissues, especially in endometrium. CYP1B1 catalyzes the 4-hydroxylation of estrogens (9). The induced expression of CYP1B1 in the endometrium would result in the increased production of 4-hydroxyestradiol. 4-Hydroxyestradiol is a catechol estrogen that can lead to carcinogenesis by the generation of free radicals (13, 14). This notion is supported by several previous studies, showing elevated levels of 4-hydroxyestradiol in breast tumors and endometrial adenocarcinoma (15, 28). Therefore, the enhanced metabolism of the estrogens by CYP1B1 in the uterus may facilitate carcinogenesis. Of particular interest is the action of tamoxifen, a specific antagonist of ER. In our reporter assays, 4OHT effectively inhibited E2-induced CYP1B1 expression. However, in the absence of E2, this reagent exhibited an agonistic function, up-regulating the CYP1B1 promoter. Whereas tamoxifen effectively blocks the action of E2, it increases the risk of endometrial cancer (29), although the precise molecular mechanisms remain unclear. The up-regulation of CYP1B1 by 4OHT may therefore be one component of the carcinogenesis pathway in the endometrium.

In summary, we proposed that the human CYP1B1 gene is a novel target of estrogen. ERα plays critical roles in this regulation. Significant levels of CYP1B1 expression were observed in uterus and were tightly correlated with the local levels of E2. These findings may provide insights into the molecular mechanisms of the homeostasis of estrogens as well as estrogen-dependent carcinogenesis.

We thank Dr. William A. LaMarr of University of Massachusetts Medical School (Worcester, MA) for providing pGL3-basic plasmid containing 5′-flanking region of human CYP1B1 gene. We are grateful to Dr. Pierre Chambon of Institut de Genetique et de Biologie Moleculaire et Cellulaire (Strasbourg, France) for the gift of ERα expression vectors (pSG5-HE0, HE11, HE19, and HE38). We also thank Dr. Masato Nishida of Tsukuba University (Ibaraki, Japan) for providing Ishikawa cells. We acknowledge Brent Bell for reviewing the manuscript.