3-Methylcholanthrene and Other Aryl Hydrocarbon Receptor Agonists Directly Activate Estrogen Receptor α

Estrogen receptor α (ERα) and ERβ are members of the nuclear receptor superfamily of transcription factors and receptor ligands induce formation of ER homo- or heterodimers which then activate both nuclear and extranuclear signaling pathways (1, 2). The precise functions of ERα and ERβ have been defined in knockout mouse studies and it is apparent that both ER subtypes induce separable functions/genes as well as some overlapping activities (3, 4). ERα interacts with several transcription factors and nuclear coregulatory proteins, and this can result in diverse effects which depend on the interacting protein, gene promoter, and cell context (5, 6). For example, research in this laboratory has reported that 17β-estradiol (E2) activates multiple genes through interactions of ERα with Sp proteins which in turn bind their cognate GC-rich elements (7). E2 induces vascular endothelial growth factor (VEGF) expression in ZR-75 breast cancer cells through ERα/Sp1 and ERα/Sp3 interactions with proximal GC-rich motifs in the VEGF promoter (8). In contrast, E2 down-regulates VEGF in HEC1A endometrial cancer cells through ERα/Sp3 interactions with the same GC-rich promoter sequences (9). ER/AP1 interactions activate AP1-dependent genes; however, these responses also vary with ligand structure, ER subtype, and cell context (10).

Several studies have investigated interactions between ERα and the aryl hydrocarbon receptor (AhR), which is another ligand-activated nuclear transcription factor that forms a heterodimeric nuclear complex with the AhR nuclear translocator (Arnt) protein (11, 12). Both AhR and Arnt are members of the basic helix-loop-helix family of transcription factors (13). Inhibitory AhR-ERα crosstalk has been extensively investigated in multiple hormone-responsive tissues/cells including breast, endometrial, and ovarian cancer cells, rodent mammary tumors, and the rodent uterus (11, 12). Prototypical AhR agonists such as the high-affinity ligand 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) inhibit E2-induced proliferation and gene expression in these tissues/cells. The mechanisms of these interactions are complex and may be cell context– and gene-dependent (11, 12, 14, 15). Ohtake and coworkers investigated AhR-ERα crosstalk in MCF-7 and Ishikawa cells, and in vivo (mouse uterus) using 3-methylcholanthrene (3MC) as a ligand (16). Inhibitory effects were observed for some responses in which E2 and 3MC were coadministered; however, they also showed that 3MC activated E2-responsive genes in vitro and in vivo. It was concluded that the estrogenic activity of 3MC involved an AhR-ERα complex in which ERα bound cognate response elements.

The results for 3MC were reminiscent of our previous studies in breast cancer cells with the AhR ligand 6-methyl-1,3,8-trichlorodibenzofuran (6-MCDF) which inhibited E2-induced responses but also activated estrogenic activity (17, 18). However, it was concluded that 6-MCDF directly activated ERα independent of the AhR (18), and this complemented results showing that in the absence of the AhR, TCDD also activated ERα-dependent transactivation (19). In this study, we used 3MC and 3,3′,4,4′,5-pentachlorobiphenyl (PCB) as prototypical AhR agonists and investigated their estrogenic activities in MCF-7 cells and the mouse uterus. The results show that both compounds activate ERα-dependent signaling in the presence or absence of the AhR and interact with ERα in binding assays. Moreover, in living cells, 3MC, PCB, and E2 induce ERα homodimerization as determined by fluorescence resonance energy transfer (FRET), indicating that these compounds can directly activate ERα.

Cell lines, chemicals, biochemical, constructs, and oligonucleotides. MCF-7 cells were obtained from the American Type Culture Collection (Manassas, VA). DME/F12 with and without phenol red, 100× antibiotic/antimycotic solution was purchased from Sigma Chemical Co. (St. Louis, MO). Fetal bovine serum was purchased from Intergen (Purchase, NY). Antibodies for AhR, Arnt, ERα, Cyp1A1, TF_IIB, cathepsin D, IgG, and β-tubulin proteins were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). 17β-Estradiol (E2) and 3MC were obtained from Sigma. PCB was synthesized in this laboratory and >99% pure. Lysis buffer and luciferase reagents were purchased from Promega Corp. (Madison, WI). Small inhibitory RNA (siRNA) duplexes for AhR, ERα, luciferase (GL2), and the scrambled siRNA were prepared by Dharmacon Research (Lafayette, CO). The sequences of these siRNAs are shown below.

Transfection of MCF-7 breast cancer cells. Cells were cultured in six-well plates in 2 mL of DME/F12 medium supplemented with 5% fetal bovine serum. After 16 to 20 hours, when cells were 50% to 60% confluent, iRNA duplexes and/or reporter gene constructs were transfected using OligofectAMINE Reagent (Invitrogen, Carlsbad, CA). The effects of the E2, 3MC, PCB, and siRNAs on transactivation were investigated in MCF-7 cells cotransfected with (500 ng) pERE₃ or pDRE₃ constructs. Briefly, siRNA duplexes were transfected in each well to give a final concentration of 100 nmol/L. Twenty-four hours after transfection, cells were treated with Me₂SO, 10 nmol/L E2, 3 μmol/L 3MC, or 3 μmol/L PCB with 2.5% stripped serum media for 24 hours. Cells were then harvested, and luciferase activity (relative to β-galactosidase activity) was determined.

Western blot analysis. Cells were washed once with PBS and collected by scraping in 200 μL of lysis buffer [50 mmol/L HEPES, 0.5 mol/L sodium chloride, 1.5 mmol/L magnesium chloride, 1 mmol/L EGTA, 10% (vol/vol) glycerol, 1% Triton X-100, and 5 μL/mL of protease inhibitor cocktail (Sigma)]. Lysates from the cells were incubated on ice for 1 hour with intermittent vortexing followed by centrifugation at 40,000 × g for 10 minutes at 4°C. Equal amounts of protein (60 μg) from each treatment group were diluted with loading buffer, boiled, and loaded onto 10% and 12.5% SDS-polyacrylamide gel. Samples were electrophoresed and proteins were detected by incubation with polyclonal primary antibodies AhR (N-19), Arnt (C-19), CYP1A1 (G-18), ERα (G-20), cathepsin D (C-20), and β-tubulin (H-235) followed by blotting with the appropriate horseradish peroxidase–conjugated secondary antibody. After autoradiography, band intensities were determined by a scanning laser densitometer (Sharp Electronics Corporation, Mahwah, NJ) using Zero-D Scanalytics software (Scanalytics Corporation, Billerica, MA).

FRET microscopy and analysis. Cells were seeded in two-well Lab-Tek Chambered Coverglass slides (Nalge Nunc International, Rochester, NY) in DME/F12 medium supplemented with 5% charcoal-stripped serum and grown for 36 hours. Cells were then transfected for 16 hours with CFP-YFP chimera or CFP-ER alone or YFP-ER alone or in combination. Cells were washed with DME/F12 medium and then put on the stage of a Bio-Rad 2000 MP microscope system (Bio-Rad Laboratories, Hercules, CA) equipped with a Nikon T#300 inverted microscope with a 60× (NA1.2) water immersion objective lens and a titanium:sapphire laser tuned to 820 nm wavelength and an argon:krypton laser tuned to 488 nm excitation. Images were acquired between 8 and 20 minutes after the addition of each ligand. CFP and YFP FRET data in MCF-7 cells were collected using two photon 820 nmol/L excitation wavelength. Emission of CFP (CFP channel, donor signal) was collected using a 500DCLP dichroic and 450/80 nm filter, whereas emission of YFP (FRET channel, acceptor signal) was collected using a 528/50 nm filter. Donor bleed through signal to the FRET channel was calculated by measuring the FRET channel signal resulting from MCF-7 cells transfected only with the CFP fusion construct. Acceptor bleed through to the FRET channel was calculated by measuring the FRET channel signal resulting from MCF-7 cells transfected with YFP fusion construct alone. To correct for variations in fluorophore expression resulting from different transfection efficiencies, minimal levels of YFP expression and maximum levels of CFP were selected based on data collected from each experiment. Cells that did not match the selection criteria were eliminated from the FRET analysis. Negative (CFP empty and YFP empty) and positive (CFP-YFP chimera) controls were used to calculate the approximate FRET efficiency in cells treated with different ligands; it was assumed that the signal from cells transfected with the positive CFP-YFP chimera construct would exhibit 50% FRET efficiency when compared with signals from cells transfected with CFP/YFP empty constructs. For identification of region of interest and FRET analysis, MetaMorph software version 6.0 (Universal Imaging Corp., Downingtown, PA) was used. Acceptor signal acquired with the FRET channel was corrected by subtracting the background signal as well as the donor bleed through signal. At least 50 cells per treatment were analyzed by one-way ANOVA followed by Dunnett's test.

Treatment of AhR^−/− mice: semiquantitative reverse transcription-PCR analysis. AhR^−/− and AhR^+/+ mice (on C57BL/6J background) were purchased from The Jackson Laboratory (Bar Harbor, ME) and a breeding colony was maintained in the Institute of Biosciences and Technology laboratory animal facility (Houston, TX), and 21-day-old immature female mice were used for this study. A stock solution of 3MC was prepared in corn oil with a final concentration of 1.2 mg/mL. Mice were treated with 100 μL/10 g animal weight to give a final dose of 12 mg/kg. Control mice were treated with corn oil alone. Mice were treated with corn oil or 3MC (12 mg/kg) every 24 hours for 3 days. Animals were sacrificed, uterine wet weights relative to body weights were measured, and total uterine RNA was obtained with RNAzol B (Tel-Test, Friendswood, TX) according to the manufacturer's protocol. RNA concentrations were measured by UV 260:280 nm absorption ratio, and 200 ng/μL RNA was used in each reaction for reverse transcription-PCR. RNA was reverse transcribed at 42°C for 25 minutes using oligo d(T) primer (Promega) and subsequently PCR-amplified of reverse transcription product using 2 mmol/L MgCl₂, 1 μmol/L of each gene-specific primer, 1 mmol/L deoxynucleotide triphosphates, and 2.5 units AmpliTaq DNA polymerase (Promega). The gene products were amplified using 22 to 25 cycles (95°C, 30 seconds; 56°C, 30 seconds; and 72°C, 30 seconds). The sequence of the oligonucleotide primers used in this study was as follows:

Following amplification in a PCR express thermal cycler (Hybaid US, Franklin, MA), 20 μL of each sample were loaded on a 2% agarose gel containing ethidium bromide. Electrophoresis was done at 80 V in 1× TAE buffer for 1 hour, and the gel was photographed by UV transillumination using Polaroid film (Waltham, MA). Cyclin D1 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) band intensity values were obtained by scanning the Polaroid on a Sharp JX-330 scanner (Sharp Electronics); background signal was subtracted, and densitometric analysis was done on the inverted image using Zero-D software (Scanalytics). Results were expressed as cyclin D1 band intensity values normalized to GAPDH values.

Chromatin immunoprecipitation assay. MCF-7 cells (1 × 10⁷) were treated with Me₂SO (time 0), 10 nmol/L E2, 3 μmol/L 3MC, and 3 μmol/L PCB for 30 or 60 minutes. Cells were then fixed with 1.5% formaldehyde, and the cross-linking reaction was stopped by the addition of 0.125 mol/L glycine. After washing twice with phosphate-buffered saline (2×), cells were scraped and pelleted. Collected cells were hypotonically lysed, and nuclei were collected. Nuclei were then sonicated to the desired chromatin length (500 bp-1 kb). Chromatin was precleared by the addition of protein A-conjugated beads (Pierce, Rockford, IL), and then incubated at 4°C for 1 hour with gentle agitation. The beads were pelleted, and the precleared chromatin supernatants were immunoprecipitated with antibodies specific to IgG, TF_IIB, ERα, AhR, and Arnt (Santa Cruz Biotechnology) at 4°C overnight. The protein-antibody complexes were collected by the addition of protein A-conjugated beads at room temperature for 1 hour. The beads were extensively washed; the protein-DNA cross-links were eluted and reversed. DNA was purified by Qiaquick Spin Columns (Qiagen, Valencia, CA) followed by PCR amplification. The pS2 primers are: 5′-CTA GAC GGA ATG GGC TTC AT-3′ (forward), and 5′-ATG GGA GTC TCC TCC AAC CT-3′ (reverse), which amplify a 209-bp region of the human pS2 promoter containing an estrogen response element (ERE). The CYP1A1 primers are: 5′-CAC CCT TCG ACA GTT CCT CTC-3′ (forward), and 5′-GCT AGT GCT TTG ATT GGC AGA G-3′ (reverse), which amplify a 381-bp region of human CYP1A1 enhancer containing DREs. The positive control primers are: 5′-TAC TAG CGG TTT TAC GGG CG-3′ (forward), and 5′-TCG AAC AGG AGG AGC AGA GAG CGA-3′ (reverse), which amplify a 167-bp region of the human GAPDH gene. The negative control primers are: 5′-atg gtt gcc act ggg gat ct-3′ (forward), and 5′-TGC CAA AGC CTA GGG GAA GA-3′ (reverse), which amplify a 174-bp region of human CNAP1 exon. PCR products were resolved on a 2% agarose gel in the presence of 1:10,000 SYBR gold (Molecular Probes, Eugene, OR).

ERα binding and coactivator recruitment. Fluorescence polarization (FP)-based assays were used to measure competitive ligand binding to ERα (ERα Competitor Assay, Green; Invitrogen) and ligand-induced coactivator peptide recruitment to ERα (ERα Coactivator Assay; Invitrogen). In the FP-based ERα-competitive ligand binding assay, serially diluted test compounds were competed against a fluorescent estrogen ligand (Fluormone ES2; K_d equals 4 ± 2 nmol/L) for binding human recombinant baculovirus-expressed ERα. If the test compound does not bind ERα, no competition occurs and a relatively large ERα/ES2 complex forms, exhibiting a high polarization value. If the test compound does bind ERα, the Fluormone ES2 is displaced from ERα, resulting in a smaller ERα/test compound complex and a low polarization value. E2 and ICI 182780 were dissolved in ethanol and mixed at 1% (vol/vol) with ES2 Screening Buffer; 3MC and PCB were dissolved in Me₂SO and mixed at 2% (vol/vol) with ES2 Screening Buffer. The test compounds were subsequently serially diluted in 2-fold increments with ES2 Screening Buffer and mixed 1:1 (vol/vol) with a 2× ERα/ES2 complex, such that the final concentration of ERα was 15 nmol/L and the final concentration of Fluormone ES2 was 1 nmol/L in a reaction volume of 100 μL. Competition for binding ERα between Fluormone ES2 and the test compounds was allowed to come to equilibrium for 2 hours and not more than 5 hours at room temperature in the dark.

In the FP-based ERα coactivator peptide recruitment assay, serially diluted test compounds were allowed to promote formation or disruption of a complex of human recombinant baculovirus-produced ERα and fluorescent coactivator peptide (D22). This peptide was rhodamine-labeled and contained an LXXLL interaction motif in a flanking sequence context found in known coactivators as identified by a phage display screen (20). Agonist-bound ERα results in increased affinity for the coactivator peptide, leading to a higher polarization value relative to a no-ligand control. In contrast, antagonist bound-ERα leads to decreased affinity for the coactivator peptide, and hence, a lower polarization value than a no-ligand control. E2 and ICI 182780 were dissolved in ethanol and mixed at 1% (vol/vol) with Coactivator Assay Buffer; 3MC and PCB were dissolved in Me₂SO and mixed at 2.5% (vol/vol) with Coactivator Assay Buffer. The test compounds were subsequently serially diluted in 2-fold increments in Coactivator Assay Buffer and mixed with ERα at a final concentration of 75 nmol/L and coactivator peptide D22 at 1× final concentration (relative to a 10× stock) in a reaction volume of 100 μL. Recruitment of coactivator peptide D22 to ERα in the presence of serially diluted test compounds was allowed to come to equilibrium for 2 hours and not more than 5 hours at room temperature in the dark. Polarization values (millipolarization units) were determined using a Beacon 2000 FP System equipped with a pair of 488 excitation and 535 nm emission filters for the competitive ligand binding assays and with a pair of 535 excitation and 590 nm emission filters for the coactivator recruitment assay.

Statistical analysis. Statistical significance was determined by ANOVA and Scheffe's test, and the levels of probability are noted. The results are expressed as means ± SD for at least three separate (replicate) experiments for each treatment. For FRET experiments, one-way ANOVA followed by Dunnett's test was used to analyze the statistical differences between control and ligand-treated cells at P < 0.05 using Prism software version 4.0 (GraphPad Software, Inc., San Diego, CA).

Ohtake and coworkers (16) reported that 3MC activated estrogenic responses in vitro and in vivo and proposed a mechanism by which the liganded AhR complex acts as a coactivator of ERα-dependent transactivation. However, polynuclear aromatic hydrocarbons (PAH) and other AhR agonists including TCDD can directly activate ERα signaling (18, 19, 21–24). This study uses both 3MC and PCB as prototypical AhR ligands and investigates their activation of ERα-mediated responses and the mechanisms associated with this activity. Preliminary dose-response studies gave maximal induction using concentrations of 10 nmol/L for E2 and TCDD and 3 μmol/L concentrations of 3MC and PCB. Higher concentration of the latter compounds were not used due to solubility problems, and 1 μmol/L concentrations of PCB and 3MC also significantly induced activity (data not shown). The results in Fig. 1A summarize the effects of 10 nmol/L E2, 10 nmol/L TCDD, and 3 μmol/L 3MC and PCB on induction of luciferase in Ah-responsive MCF-7 cells transfected with pERE₃. The results show that E2, PCB, and 3MC but not TCDD, induced transactivation, and these responses were inhibited after cotreatment with the antiestrogen ICI 182780 (1 μmol/L). The estrogenic activity of these ligands was further investigated in MCF-7 cells transfected with pERE₃ and siRNAs for luciferase (iGL2), ERα (iER), and the AhR (iAhR) or a nonspecific RNA (iScr; Fig. 1B). E2, 3MC, and PCB activated luciferase activity in cells transfected with iScr, whereas all responses were decreased by >90% in cells transfected with iGL2. This is consistent with knockdown of the luciferase reporter gene in every treatment group. There was also a significant decrease in E2-, 3MC-, and PCB-induced luciferase activity in cells transfected with iERα, demonstrating the requirements of ERα for mediating these responses. After transfection with iAhR, activation of luciferase activity by E2, 3MC, and PCB was comparable to that observed in cells transfected with iScr. These results were similar to previous studies showing that 6-MCDF also induced transactivation in cells transfected with pERE₃ alone or in combination with iAhR (18). Thus, like E2, 3MC- and PCB-induced transactivation in MCF-7 cells transfected with pERE₃ was dependent on ERα but not AhR expression. Moreover, as previously reported, TCDD exhibited estrogenic activity in MCF-7 cells transfected with iAhR and this is consistent with the high binding affinity of TCDD for the AhR which essentially represses activation of ERα by TCDD (19). In contrast, 3MC and PCB bind with lower affinity to the AhR (25, 26) and can induce ERα-mediated transactivation in the presence or absence of AhR expression. In a parallel experiment using the same siRNAs, the ligand-dependent activation of the Ah-responsive pDRE₃ construct was investigated in MCF-7 cells (Fig. 1C). TCDD, 3MC and PCB, but not E2 or ICI 182780, induced luciferase activity in cells transfected with iScr or iERα, whereas in cells transfected with iGL2 or iAhR, the induced activities were significantly decreased. Thus, 3MC and PCB exhibit Ah- and ER-responsiveness which is dependent on AhR or ERα expression, respectively, and the estrogen responsiveness of these compounds is dependent on ERα but not AhR expression.

The efficiency of transfected iAhR and iERα on protein knockdown and loss of responsiveness was also determined at the protein level in MCF-7 cells. Results in Fig. 2A show that in whole cell lysates from MCF-7 cells transfected with iAhR, there was a >60% decrease in AhR protein levels compared with cells transfected with iScr. These results were obtained with whole cell lysates indicating highly efficient knockdown of AhR in the transfected cells (transfection efficiencies were 60-80%). The results are complemented by the decreased induction of CYP1A1 protein by 3MC and PCB in MCF-7 cells transfected with iAhR compared with CYP1A1 protein levels in cells transfected with iScr (Fig. 2B). The results in Fig. 2C also show that transfected iERα decreased levels of ERα protein in whole cell lysates from MCF-7 cells. In MCF-7 cells transfected with iScr, 10 nmol/L E2, 3 μmol/L 3MC, and 3 μmol/L PCB induced cathepsin D protein (relative to β-tubulin; Fig. 2D). In cells transfected with iERα, the induction of cathepsin D by these same compounds was clearly decreased, whereas in cells transfected with iAhR, cathepsin D was induced by E2, 3MC, and PCB. Similar results were also observed for the activation of c-fos by the same compounds (data not shown).

The results indicate that the AhR agonists 3MC and PCB also directly activate ERα-mediated transcription independent of AhR and interaction of 3MC and PCB with ERα was further investigated in binding assays. The results in Fig. 3A illustrate the competitive hormone binding of E2, ICI 182780, PCB, and 3MC to ERα using a FP assay. The IC₅₀ values for displacement were 34.0 nmol/L, 79.4 nmol/L, 8.82 μmol/L and 17.7 μmol/L for E2, ICI 182780, PCB, and 3MC, respectively. The IC₅₀ ER binding values for 3MC and PCB (8.82 and 17.7 μmol/L, respectively) are higher than expected based on their transactivation responses which are observed using 3 μmol/L concentrations. These differences have also been observed for other estrogenic compounds that induce transactivation in the low micromolar range (27). For example, several alkylphenols, bisphenol, endosulfan, and kepone also exhibit high relative binding affinities compared with their more potent induction of transactivation (27). This may be related to the high lipophilicity of these compounds which may affect interactions with ERα over the relatively short incubation times for the binding assay. We also used a FP assay to measure ligand-induced coactivator peptide recruitment to ERα (20). The results (Fig. 3B) show that EC₅₀ values for E2, PCB, and 3MC were 27.7 nmol/L, 20.5 μmol/L, and 25.0 μmol/L, respectively. These data further confirm that both 3MC and PCB directly activate ERα. The results show that, like E2, both PCB and 3MC induce interactions with a coactivator peptide in this assay, although both compounds were much less active than E2. It is possible that PCB and 3MC resemble other xenoestrogenic compounds and induce ERα interactions with different sets of coactivators, and this is currently being investigated.

FRET can also be used to measure protein-protein interactions and previous studies have shown E2 and other ligand-dependent interactions of YFP-ERα and CFP-ERα by determining increased FRET efficiency (28–31). This assay has now been used to investigate 3MC- and PCB-induced interactions of YFP-ERα and CFP-ERα in MCF-7 cells transfected with these constructs. In Fig. 4A, the results show that PCB, 3MC, and E2 induce translocation of transfected CFP-ERα and YFP-ERα into the nucleus of COS-1 cells and representative images in all three treatment groups show a punctuate pattern of ERα staining. Compared with treatment with Me₂SO, E2 induced a significant increase in FRET efficiency and significantly enhanced FRET efficiency was also observed after treatment with 0.1 and 1.0 μmol/L PCB (Fig. 4B; ref. 31). In a parallel experiment (Fig. 4B), E2, 1.0 and 10 μmol/L 3MC also significantly increased FRET efficiency, thus confirming that both PCB and 3MC induce ERα dimerization in living cells. Moreover, it was also shown that PCB/MC-induced dimerization of ERα in the FRET assay was observed at concentrations lower than that required for transactivation (Fig. 1).

Several reports have shown that induction of E2-responsive genes such as pS2 by E2 results in recruitment of ERα to the pS2 gene promoter as determined in a chromatin immunoprecipitation (ChIP) assay (32–36). The results in Fig. 5A show that after treatment of MCF-7 cells with E2, 3MC, or PCB for 30 or 60 minutes, there is an enhanced band for ERα, suggesting that all three ligands induce ERα interactions with the region of the pS2 promoter containing a well-characterized nonconsensus ERE. In contrast, interactions of the AhR and Arnt proteins with this region of the pS2 promoter are relatively unchanged in the treatment groups compared with the solvent (Me₂SO)-treated cells (observed in duplicate experiments). Because 3MC and PCB induce both estrogenic and AhR agonist activity (Figs. 1 and 2), we also examined interactions of AhR/Arnt and ERα with the region of the CYP1A1 promoter containing DREs (Fig. 5A). The results show that 3MC and PCB induce recruitment of AhR/Arnt to the CYP1A1 promoter as previously described (36, 37). The results in Fig. 5A indicate that after treatment with PCB and 3MC (but not E2), there was enhanced binding of ERα to the CYP1A1 promoter. This is consistent with the AhR agonist activity of these compounds, and similar results have been observed in cells treated with TCDD (36, 38). Figure 5B is a control ChIP assay showing binding of TF_IIB to the GAPDH promoter but not CNAP1 exon. Thus, PCB and 3MC exhibit both AhR and ERα agonist activities and the latter response is due to direct interactions with ERα.

Ohtake and coworkers (16) reported that treatment of 21-day-old AhR^−/− mice with 3MC (4 mg/kg) every day for 3 days did not affect uterine wet weight, whereas this dose induced a uterine wet weight increase and activation of E2-responsive genes in wild-type animals (16). We repeated this experiment in both AhR^−/− and AhR^+/+ mice but used a 3-fold higher dose of 3MC (12 mg/kg), and the effects on cyclin D1 and uterine wet weight increase were determined in at least three animals per treatment group (Fig. 6). The results show that 3MC induced cyclin D1 mRNA levels in both AhR^−/− (Fig. 6A) and AhR^+/+ (Fig. 6B) mice; uterine wet weight was also increased in AhR^−/− (Fig. 6C) and AhR^+/+ (Fig. 6D) mice treated with the same dose of 3MC. Thus, 3MC exhibited estrogenic activity in vivo and in vitro, and these responses were independent of AhR and due to direct interactions with ERα.

The AhR and ERα are ligand-induced nuclear transcription factors that are members of the basic helix-loop-helix and NR families, respectively. The AhR acts as a nuclear heterodimer with Arnt and activates genes through interactions with promoter DREs, whereas ERα primarily forms a homodimer which activates E2-responsive genes through interaction with consensus/nonconsensus promoter EREs. In addition, ER also activates genes through protein-protein interactions in which ER interacts with other DNA-bound transcription factors such as AP1 and Sp proteins (7, 10). Despite differences in the overall mechanisms of AhR and ER action and their resulting gene expression profiles, there is strong evidence for interactions between the two signaling pathways. For example, the AhR and ER physically interact (15, 16, 39), and several studies have reported ERα-dependent modulation of Ah-responsiveness in which ERα enhances (36, 40, 41), decreases (38, 42, 43), or does not change (44) a ligand-activated AhR response. Differences in these reported ERα-AhR interactions cannot be entirely resolved but may be due, in part, to cell context and the specific response. Research in our laboratories has investigated inhibitory AhR-ERα crosstalk in several hormone-responsive cancer cell lines/tumors and the rodent uterus, and a number of possible gene-specific mechanisms have been reported (11, 12).

Ohtake and coworkers (16) reported a novel mechanism of AhR-ERα interactions in which AhR ligands activate estrogenic responses in breast/endometrial cancer cells and the rodent uterus through AhR/Arnt-ERα (protein-protein) interactions where ERα binds cognate response elements. Thus, the AhR complex is acting as a ligand-induced coactivator of ERα. These studies primarily used 3MC as a ligand and although there was strong evidence for 3MC-induced estrogenic activity, 3MC also inhibited some E2-induced responses in cells and the mouse uterus that were consistent with inhibitory AhR-ERα crosstalk (16). Previous studies in this laboratory showed that 1 μmol/L 3MC inhibited several E2-induced responses and also down-regulated ER protein expression and estrogenic activity was not observed using this concentration (45, 46). However, it is possible that in one experiment which showed that 10 μmol/L 3MC induced a 70% decrease in ERα protein in MCF-7 cells (46), this response could be due to simultaneous proteasome-dependent degradation of ERα through interaction of 3MC with both ERα and AhR. It is also possible that the estrogenic activity of 3MC may be due to direct activation of ERα because several studies report that different structural classes of AhR agonists activate and bind both the AhR and ERα, and these include flavonoids, indolo[3,2]-carbazole, PAHs, and 6-MCDF (18, 19, 21–24). Moreover, it was also shown that in the absence of AhR expression (by RNA interference), the potent AhR agonist TCDD induced transactivation in MCF-7 cells transfected with pERE₃, and this response was inhibited by the antiestrogen ICI 182780 (19). The estrogenic activity of TCDD was somewhat unique and dependent on the lack of AhR expression due to the extremely high AhR-TCDD binding affinity. We hypothesized that 3MC may also directly activate ERα and, therefore, in this study, we used both 3MC and PCB which bind the AhR with moderate affinity (25, 26). The PCB congener was selected because this compound is more resistant to metabolism, whereas 3MC may be converted to hydroxylated metabolites and it is known that some PAH metabolites are more estrogenic than their parent hydrocarbons (21–23).

Our results show that 3MC and PCB induce CYP1A1 and luciferase activity (in pDRE₃-transfected cells) and thereby exhibit their expected AhR agonist activity (Figs. 1C and 2B). Moreover, in a ChIP assay, both 3MC and PCB recruited the AhR/Arnt complex to the CYP1A1 promoter (Fig. 5B). 3MC and PCB also induced transactivation in MCF-7 cells transfected with pERE₃ (Fig. 1A) and cathepsin D (Fig. 2D). Results of RNA interference studies show that the estrogenic activity was dependent on ERα but not AhR expression (Figs. 1 and 2). Further evidence for direct activation of ERα by 3MC and PCB was determined in ERα binding and FRET assays that showed direct competitive ER binding by both ligands and increased ligand-induced FRET efficiency through CFP-ERα/YFP-ERα interactions, respectively (Figs. 3 and 4). 3MC- and PCB-induced FRET efficiencies were observed in both MCF-7 and AhR-negative COS-1 cells, indicating that ERα homodimerization was AhR-independent.

Matthews and coworkers (36) recently used the ChIP assay to investigate ligand-induced ERα and AhR/Arnt interactions on prototypical Ah-responsive (CYP1A1) and E2-responsive (pS2) gene promoters. They used E2 and TCDD as ligands for their study in MCF-7 cells, and the results obtained for TCDD were observed in cells expressing AhR and therefore reflected only the AhR agonist activity of this compound. 3MC and PCB simultaneously activated both the AhR and ERα in MCF-7 cells, whereas TCDD exhibited estrogenic activity only in the absence of the AhR (Fig. 1B) due to the extremely high binding affinity of TCDD for the AhR. In Ah-responsive MCF-7 cells treated with TCDD, there was an increased recruitment of AhR/Arnt and ERα to the CYP1A1 promoter, whereas the ERα complex was not recruited to the pS2 promoter due to preferential TCDD-AhR binding (36). Results in Fig. 5 show that like TCDD, 3MC and PCB recruited AhR/Arnt and ERα to the CYP1A1 promoter but did not affect AhR/Arnt interactions with the pS2 promoter. However, like E2, both 3MC and PCB recruited ERα to the pS2 promoter, and these results are consistent with the estrogenic activity of these ligands through direct interactions with ERα. 3MC also induced an increase in uterine wet weight and cyclin D1 mRNA levels in AhR^−/− and AhR^+/+ mice, suggesting that the failure to observe induction in the previous study may be due to the low dose used. These in vivo results coupled with the in vitro data show that 3MC and PCB induce estrogenic responses independent of the AhR.

In summary, this study shows that 3MC and PCB represent a novel subclass of xenoestrogenic compounds that activate both ERα and AhR signaling pathways. Moreover, the results indicate that activation of ERα occurs through direct interactions with this receptor and are independent of the AhR complex. The results of this study extend the number of ligands that bind and activate both ERα and AhR. Selective ER modulators are extensively used for the treatment of breast cancer and other hormone-related diseases (47), and there is evidence that selective AhR modulators may also have some clinical applications (11, 12, 17) for cancer chemotherapy. Studies on selective modulators for these receptors should evaluate cross-receptor activation and its effect on potential clinical applications for these compounds.

Grant support: NIH grants ES04176, ES04917, and ES09106 (S. Safe), Specialized Program of Research Excellence in Breast Cancer P50-Ca89018 (V.C. Jordan), the Texas Agricultural Experiment Station (S. Safe), and a Fellowship from the Eli Lilly Company (E. Ariazi).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.