Aryl Hydrocarbon Receptor Modulation of Estrogen Receptor α-Mediated Gene Regulation by a Multimeric Chromatin Complex Involving the Two Receptors and the Coregulator RIP140

Abstract

Although crosstalk between aryl hydrocarbon receptor (AhR) and estrogen receptor α (ERα) is well established, the mechanistic basis and involvement of other proteins in this process are not known. Because we observed an enrichment of AhR-binding motifs in ERα-binding sites of many estradiol (E2)-regulated genes, we investigated how AhR might modulate ERα-mediated gene transcription in breast cancer cells. Gene regulations were categorized based on their pattern of stimulation by E2 and/or dioxin and were denoted E2-responsive, dioxin-responsive, or responsive to either ligand. ERα, AhR, aryl hydrocarbon receptor translocator, and receptor interacting protein 140 (RIP140) were recruited to gene regulatory regions in a gene-specific and E2/dioxin ligand-specific manner. Knockdown of AhR markedly increased the expression of ERα-mediated genes upon E2 treatment. This was not attributable to a change in ERα level, or recruitment of ERα, phosphoSer5-RNA Pol II, or several coregulators but rather was associated with greatly diminished recruitment of the coregulator RIP140 to gene regulatory sites. Changing the cellular level of RIP140 revealed coactivator or corepressor roles for this coregulator in E2- and dioxin-mediated gene regulation, the choice of which was determined by the presence or absence of ERα at gene regulatory sites. Coimmunoprecipitation and chromatin immunoprecipitation (ChIP)-reChIP studies documented that E2- or dioxin-promoted formation of a multimeric complex of ERα, AhR, and RIP140 at ERα-binding sites of genes regulated by either E2 or dioxin. Our findings highlight the importance of cross-regulation between AhR and ERα and a novel mechanism by which AhR controls, through modulating the recruitment of RIP140 to ERα-binding sites, the kinetics and magnitude of ERα-mediated gene stimulation.

Estrogenic hormones are crucial for the regulation of many physiological processes in both reproductive and nonreproductive tissues, and they have a significant impact on the phenotypic properties of cancers, such as breast cancer, that develop in these tissues. These effects are exerted by the binding of estrogens to their receptors estrogen receptor α (ERα) and ERβ, which are members of the nuclear receptor superfamily of ligand-activated transcription factors (Hall et al., 2001; Katzenellenbogen and Katzenellenbogen, 2000; McDonnell and Norris, 2002).

Aryl hydrocarbon receptor (AhR) and its nuclear partner aryl hydrocarbon receptor translocator (ARNT) are also ligand-activated transcription factors, which belong to the helix-loop-helix, per/ARNT/Sim family of transcription factors. AhR and ARNT are involved in regulating physiological responses to polycyclic aromatic hydrocarbons and halogenated hydrocarbons that are ubiquitous in the environment and can cause toxicity in humans and other mammals due to AhR-dependent induction of phase I and phase II enzymes, which metabolize these chemicals into genotoxic and cytotoxic intermediates (Abel and Haarmann-Stemmann, 2010; Denison and Nagy, 2003; Schlezinger et al., 2006).

In the absence of ligand stimulation, AhR typically resides in a cytoplasmic complex comprising multiple heat shock proteins and the immunophilin-like XAP2 protein. In the presence of environmental chemicals, such as 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD, dioxin), ligand-bound AhR translocates to the nucleus where it partners with ARNT. The heterodimer then binds to dioxin response elements (DREs) or xenobiotic response elements (XREs) on DNA, acting as a nucleation site for the binding of coregulatory proteins that modify chromatin structure and accessibility and create a permissive environment for expression of target genes. Basal transcription machinery is then recruited, facilitating the transcription of genes such as CYP1A1 or CYP1B1 (Beischlag et al., 2008).

The findings of increased incidence of breast cancer in women who had higher concentrations of polycyclic aromatic hydrocarbons in adipose tissue (Muscat et al., 2003) and the observation of higher presence of nuclear AhR and CYP1B1 in human breast cancers versus normal breast imply that nuclear-localized AhR might play a significant role in the pathobiology and phenotypic properties of breast cancer cells (Schlezinger et al., 2006; Yang et al., 2008). This is further supported by additional studies that have demonstrated the involvement of AhR in cell growth (Chang et al., 2007; Marlowe et al., 2004; Marlowe and Puga, 2005; Puga et al., 2002; Yang et al., 2008), apoptosis (Kajta et al., 2009; Marlowe et al., 2008; Marlowe and Puga, 2005), and motility of cancer cells (Kung et al., 2009).

Because ERα (Frasor et al., 2003; Hall et al., 2001) and AhR (Marlowe et al., 2008; Matthews and Gustafsson, 2006) can both greatly impact the gene expression and proliferative programs of breast cancer cells, the copresence of both receptors in breast cancer cells suggests that ERα and AhR crosstalk could be of physiologic and tumorigenic relevance and provide a potential regulatory route relevant to toxicology and breast cancer. Research to address the crosstalk between ERα and AhR has shown that agonist-bound AhR had antiestrogenic properties and that dioxin could inhibit E2 target genes (Krishnan et al., 1995; Matthews and Gustafsson, 2006; Porter et al., 2001; Safe et al., 2000; Wang et al., 2001), through inhibitory DREs or possibly by squelching of coactivators (Reen et al., 2002). More recently, liganded AhR was shown to act as an E3 ubiquitin ligase, targeting ERα protein for proteasomal degradation (Ohtake et al., 2007, 2009).

Although these studies have shed light on important aspects of this ERα and AhR nuclear receptor crosstalk, they did not explore the effect of unliganded AhR on kinetics of ERα-mediated gene regulation and the possible involvement of coregulator proteins that work along with nuclear receptors in regulating patterns of gene expression. The goal of this study was therefore to delineate interrelationships between ERα and AhR and the mechanisms governing ERα gene regulation by AhR and associated coregulators. In addition, in our recent studies where we characterized genome-wide binding sites of ERα (Madak-Erdogan et al., 2011), we observed an enrichment of AhR-binding motifs at ERα-binding sites, implying potential interrelationships of importance between these two receptors at the chromatin level.

Therefore, we have undertaken studies to explore the effect of AhR on the kinetics and magnitude of ERα-mediated gene regulation and the possible involvement of coregulator proteins in the interrelationships of these two nuclear receptors in regulating patterns of gene expression. Our findings reveal extensive crosstalk between these receptors and a novel mechanism in breast cancer cells by which AhR exerts control of ERα action through formation of a ternary complex involving the two receptors and the coregulator receptor interacting protein 140 (RIP140). In addition, the presence or absence of ERα at gene regulatory sites appears to play a crucial role in determining whether RIP140 functions as a coactivator or a corepressor in E2- and dioxin-mediated gene regulation.

MATERIALS AND METHODS

Ligands.

17β-Estradiol (E2) and dioxin were from Sigma-Aldrich (St Louis, MO) and were diluted in ethanol prior to addition to cell culture media.

Cell culture, RNA extraction, and real-time PCR analysis of gene expression regulation.

MCF-7 human breast cancer cells were maintained in culture as previously described (Frasor et al., 2003). At 6 days before vehicle (0.1% ethanol) or ligand treatment, cells were switched to phenol red-free media containing charcoal dextran–treated calf serum. Medium was changed on days 2 and 4 of culture, and cells were then treated with compounds as indicated. After cell treatments, total RNA was isolated, reverse transcribed, and analyzed by real-time PCR exactly as described previously (Frasor et al., 2003; Madak-Erdogan et al., 2008). Primers for the genes studied were as follows: LRRC54f-GGGCTACACGACGTTGGCT, LRRC54r-GAGGTCAAGCGACTCCAGGTA, HSPB8f-TGGATACGTGGAGGTGTCTGG, HSPB8r-GATCCACCTCTGCAGGAAGC, CYP1A1f-CTGACCCTGGGAAAGAACCC, CYP1A1r-CTGCTGGCTCATCCTTGACA, CYP1B1f-TGCAGGCAGAATTGGATCAG, CYP1B1r-CCATACAAGGCAGACGGTCC, AhRf-AGAGAGTCTTACTCTGCCGCCC, and AhRr-GCCAAGATTGTGCCACTGC. Primers for progesterone receptor and pS2 were as published (Frasor et al., 2003).

Chromatin immunoprecipitation and ChIP-reChIP assays.

Chromatin immunoprecipitation (ChIP) assays were performed as described (Madak-Erdogan et al., 2008). MCF-7 cells were treated with 0.1% control ethanol vehicle or 10nM E2 or 10nM dioxin for the times indicated before harvest. Cells were then fixed with formaldehyde and cross-linked chromatin complexes were collected, sonicated, and immunoprecipitated with the specific antibodies indicated. The antibodies ERα (HC-20), AhR (H-211), ARNT (H-172), and RIP140 (H-300) were purchased from Santa Cruz Biotechnology and pSer5-RNA Pol II—CTD4H8 antibody was purchased from Covance. Controls using rabbit IgG were routinely done in all ChIP assays. We also used a control non-ER binding region of the pS2 gene as an additional check for specificity of ChIP immunoprecipitations (Madak-Erdogan et al., 2008). ChIP-reChIP experiments were done following the same ChIP protocol. After the first pull down, immunoprecipitated material was recovered with 10mM dithiothreitol in IP buffer at 37°C for 30 min, diluted, and submitted to a second round of immunoprecipitation. Quantitative real-time PCR (qPCR) was used to calculate recruitment to the regions studied, as described before (Madak-Erdogan et al., 2008).

siRNA transfection.

Short interfering RNA (siRNA) knockdown of ERα, AhR, and RIP140 was done as previously described (Chang et al., 2008). Briefly, MCF-7 cells were transfected with siGENOME SMARTpool for ERα, AhR, or RIP140 or with control GL3 luciferase (no. D-001400-01) obtained from Dharmacon. All of the siRNAs were transfected into the cells at a final concentration of 20nM using the DharmaFECT transfection reagent (Dharmacon, Lafayette, CO) as per the manufacturer’s recommendations at 60 h prior to control vehicle or ligand treatment. The efficiency of ERα, AhR, and ARNT knockdowns were verified at the RNA level by qPCR and at the protein level by Western blot.

RESULTS

Bioinformatic Analysis in the Selection of Genes to Be Used in Examining ERα and AhR Crosstalk

Based on data from our previous genome-wide estrogen-regulated gene expression microarrays and ChIP-chip ERα-binding site analyses in breast cancer cells (Madak-Erdogan et al., 2011), we employed a bioinformatic approach using Cis-regulatory Element Annotation System and SeqPos to identify transcription factors whose binding motifs were enriched in the proximal promoter region of estrogen-regulated genes because these transcription factors might mediate or modulate the genomic actions of ERα. All of the algorithms identified enrichment of binding sites for AhR.

To investigate regulations between estrogen- and dioxin-regulated pathways, we selected for examination four estrogen-regulated genes, LRRC54, HSPB8, pS2, and PgR, and two AhR target genes, CYP1A1 and CYP1B1. When we treated MCF-7 breast cancer cells with 10nM E2 or 10nM dioxin alone, we observed that LRRC54, HSPB8, and PgR mRNAs were strongly upregulated by either E2 or dioxin (Fig. 1A). Interestingly, pS2 was markedly upregulated by E2 but only minimally affected by dioxin (Fig. 1B). On the other hand, the AhR targets CYP1A1 and CYP1B1 were markedly stimulated by dioxin, as expected, and were very minimally affected by E2 (Fig. 1C). Based on these results, we denoted LRRC54, HSPB8, and PgR as “Genes Stimulated by Either E2 or Dioxin,” pS2 as “E2 only Stimulated,” and CYP1A1 and CYP1B1 as “Dioxin only Stimulated” genes.

FIG. 1.

Time course of regulation of gene expression by estradiol or dioxin. MCF-7 cells were treated with 0.1% EtOH vehicle, 10nM 17β-estradiol (E) (solid line) or 10nM dioxin (D) (dashed line) for 0, 1, 2, 4, 6, and 8 h, and mRNA levels were monitored by qPCR. mRNA levels were normalized relative to 36B4 and fold change calculated relative to control. Results are the average ± SD of at least two independent experiments.

Patterns of Recruitment of ERα, AhR, and ARNT to ERα-Binding Sites of Regulated Genes by Estradiol and Dioxin

Next, we assessed the recruitment of ERα, AhR, and its partner ARNT to ERα-binding sites of regulated genes. ChIP analyses revealed that the genes could be divided into three groups, depending on their pattern of recruitment of the receptors ERα and AhR to the ERα-binding sites of these genes in the presence of their respective ligands (Fig. 2). For genes regulated by either E2 or dioxin (LRRC54, HSPB8), treatment with E2 or dioxin alone greatly increased the recruitment of both ERα and AhR. Of special note was our observation that the ligand for each receptor (e.g., dioxin) caused the recruitment of the unliganded form of the other receptor (e.g., ERα) to the binding sites studied (LRRC54_ERE and HSPB8_ERE). Thus, dioxin treatment stimulated the recruitment of ERα to these ERα-binding sites, as well as stimulating the recruitment of AhR and ARNT to these sites. Additionally, E2 treatment stimulated the recruitment of AhR and ARNT to ERα-binding sites of these genes (Fig. 2).

FIG. 2.

ERα, AhR, and ARNT recruitment in the presence of estradiol or dioxin. MCF-7 cells were treated with 10nM 17β-estradiol (E) or 10nM dioxin (D) for 1, 2, and 4 h. ERα/DNA, AhR/DNA, or ARNT/DNA complexes were immunoprecipitated using specific ERα, AhR, or ARNT antibody, respectively, or rabbit IgG as a negative control. Immunoprecipitated DNA levels were measured by qPCR and % input was calculated. Values are the mean ± SD of three independent experiments.

A second pattern of chromatin binding was represented by pS2 estrogen response element (pS2_ERE) and CYP1B1 (CYP1B1_ERE) proximal promoter sites in which E2 was more effective in recruiting ERα and was equally as effective as dioxin in recruiting AhR. These sites showed substantial ERα recruitment with E2 treatment but only slight ERα recruitment with dioxin treatment (Fig. 2). There was also a mild increase in AhR and ARNT recruitment to both of the sites with either of the ligands.

A third pattern of chromatin binding was represented by the CYP1B1_XRE, which contains both XRE and ERE sequences located approximately 1 kb upstream of the transcription start site of the gene. At the CYP1B1_XRE site, there was a surprisingly high basal level of AhR and ARNT occupancy, which was further increased by dioxin but not E2 treatment. Both E2 and dioxin treatments, however, increased ERα recruitment to the CYP1B1_XRE. Of all the sites examined, the CYP1A1_XRE was unique in that no ERα recruitment occurred with either ligand. This site was strictly dioxin regulated, with only AhR and ARNT being recruited after dioxin treatment. Collectively, these data demonstrate that ERα, AhR, and ARNT occupy ERE- and XRE-binding sites in a gene-, binding site-, and ligand-specific manner.

AhR Knockdown Increases the Magnitude of Stimulation of Estrogen Target Genes

To investigate the effect of AhR on E2 regulation of gene expression, we used an RNAi approach that resulted in very efficient knockdown of AhR RNA and protein, with AhR protein no longer detectable by Western blot (Fig. 3A). With this AhR knockdown, there was no effect on ERα protein level (Fig. 3A). In AhR-depleted cells, the magnitude of stimulation of the ERα-target genes (LRRC54, HSPB8, PgR, and pS2) by E2 was significantly higher than that of control siRNA-treated cells (Fig. 3B). Moreover, in cells depleted of AhR, basal expression of the dioxin-only stimulated genes, CYP1A1 and CYP1B1, was lost in vehicle-treated cells, and no regulation of gene expression by E2 was observed (Fig. 3C). These findings, observable when AhR is depleted from cells, reveal that AhR itself, in the absence of its ligand, is normally suppressing E2-ERα action.

FIG. 3.

AhR knockdown increases E2 stimulation of genes responsive to either E2 or dioxin and genes responsive to E2 only. MCF-7 cells were transfected with control GL3 siRNA or AhR siRNA for 60 h and then treated with 0.1% EtOH vehicle (V) or 10nM E2 (E) for 4 h. Western blot in panel A and protein quantitation (AhR/β-actin) shows nearly complete knockdown of AhR with siAhR. mRNA levels were monitored by qPCR. mRNA levels for the genes studied were normalized relative to 36B4 and fold change calculated relative to control. Results are the average ± SD of at least two independent experiments. ***p < 0.001 or **p < 0.01, significantly different from Ctrl siRNA values.

AhR Knockdown Dampens E2-Mediated RIP140 Recruitment to ERα-Binding Sites but Has No Effect on ERα or RNA Polymerase II Recruitment

The increased gene stimulation by E2 that was observed after AhR depletion could be due to alterations at various steps of the transcriptional activation process, such as recruitment of ERα, coregulators, or the basal transcriptional machinery. To test these possibilities, we first examined the impact of AhR knockdown on recruitment of ERα or pSer5-RNA Pol II (active RNA polymerase II) to the ERα-binding sites of the regulated genes. AhR knockdown did not affect the magnitude of either ERα or pSer5-RNA Pol II recruitment to the regulatory sites with E2 treatment (Figs. 4A and B).

FIG. 4.

AhR knockdown does not affect recruitment of ERα or pSer5 Pol II to ERα binding sites. MCF-7 cells were transfected with control GL3 siRNA or AhR siRNA for 60 h and then treated with 0.1% EtOH vehicle (V) or 10nM E2 (E) for 1 h. (A) ERα/DNA or (B) pSer5-RNA Pol II/DNA complexes were immunoprecipitated using specific ERα antibody or pSer5-RNA Pol II antibody or rabbit IgG as a negative control. Immunoprecipitated DNA levels were measured by qPCR, and % input was calculated. Values are the mean ± SD of two independent experiments.

Therefore, we considered that modulation of the magnitude of the E2-ERα response by AhR might be due to either an increased recruitment of a coactivator or a decreased recruitment of a corepressor. Because several studies have reported interaction of AhR with the coregulators SMRT (Nguyen et al., 1999; Rushing and Denison, 2002), RIP140 (Kumar et al., 1999), and SRC1, SRC2, and SRC3 (Beischlag et al., 2002; Kumar and Perdew, 1999), we assessed the recruitment of these coregulators by ChIP assays with and without AhR knockdown. Of the coregulators tested (NCOR, SMRT, SRC1, SRC2, SRC3, and RIP140), only RIP140 showed decreased recruitment in cells depleted of AhR. The greatly reduced recruitment of RIP140 to regulatory sites by E2 after siAhR treatment of cells is shown in Figure 5. Moreover, after ERα knockdown, we completely lost recruitment of RIP140 to the regulatory regions (Fig. 5). These data suggest that a complex containing ERα, AhR, and RIP140 is recruited to these regulatory regions with E2 treatment and that ERα and AhR are both required for full RIP140 recruitment to these regions.

FIG. 5.

AhR and ERα knockdown greatly reduces RIP140 recruitment to regulatory sites. MCF-7 cells were transfected with control GL3 siRNA, ERα siRNA, or AhR siRNA for 60 h and then treated with 0.1% EtOH vehicle (V) or 10nM E2 (E) for 1 h. RIP140/DNA complexes were immunoprecipitated using specific RIP140 antibody or rabbit IgG as a negative control. Immunoprecipitated DNA levels were measured by qPCR, and % input was calculated. Values are the mean ± SD of at least three independent experiments. ***p < 0.001, significantly different from Ctrl siRNA values.

To explore this further, we examined the dynamics of RIP140 recruitment to these chromatin-binding sites over time (Fig. 6). By ChIP assays, we observed that RIP140 showed greatly increased recruitment to ERα-binding sites by 1 h, after which the occupancy by RIP140 was either maintained at this elevated level or decreased slightly over the 3-h period examined (Fig. 6A). Moreover, by ChIP-reChIP experiments (Fig. 6B), we found that AhR, ERα, and RIP140 were present in the same complex at the ERα-binding sites upon E2 treatment.

FIG. 6.

E2 treatment promotes formation of a tripartite ERα, AhR, and RIP140 complex at gene regulatory sites. ChIP, ChIP/reChIP, and immunoprecipitation studies. (A) MCF-7 cells were treated with 10nM E2 for 0, 1, 2, and 3 h. RIP140/DNA complexes were immunoprecipitated using specific RIP140 antibody or rabbit IgG as a negative control. (B) MCF-7 cells were treated with 10nM E2 for 0–3 h. RIP140/DNA complexes were immunoprecipitated using specific RIP140 antibody or rabbit IgG as a negative control. The complexes were eluted and subjected to a second immunoprecipitation with either ERα or AhR antibody. Immunoprecipitated DNA levels were measured by qPCR, and % input was calculated. (C) MCF-7 cells were treated with 10nM E2 for 0, 1, and 2 h. AhR or RIP140 complexes were immunoprecipitated from cell lysates using AhR or RIP140 antibodies or rabbit IgG as a negative control. The complexes were eluted and resolved in 10% SDS gels. Proteins were transferred to nitrocellulose membrane, and ERα was detected using Odyssey/LICOR detection system.

To further demonstrate the mutual presence of these three factors in the same complex, we conducted coimmunoprecipitation experiments and detected E2-dependent formation of a complex containing AhR, RIP140, and ERα that was observable by 1 h (Fig. 6C). Thus, our data suggest that RIP140 is recruited to regulatory sites together with ERα and AhR. ERα appears to be essential for the formation of this tripartite complex, as observed through knockdown experiments, whereas AhR presence is partially required, as shown in AhR knockdown studies followed by ChIP analysis of RIP140 recruitment (Fig. 5).

RIP140 Knockdown Reveals a Dual Role for RIP140 as a Corepressor or Coactivator of Gene Expression Determined by the Ligand and Gene Promoter Context

Our data thus far imply that AhR and ERα form a chromatin complex containing RIP140 and that when AhR or ERα is low, a reduced amount of RIP140 is present in the complex. To test the functional significance of RIP140 recruitment to gene regulatory sites, we depleted RIP140 from MCF-7 cells using siRNA and then treated the cells with E2 and examined differences in ER-mediated gene stimulation over time. We observed that both the magnitude and duration of E2 stimulation of these target genes was markedly greater in cells after RIP140 knockdown (Fig. 7). The findings support a role for RIP140 as a corepressor that acts as a brake to keep E2-ERα actions under tight control.

FIG. 7.

RIP140 acts as a corepressor of E2-ERα gene regulation. MCF-7 cells were transfected with control GL3 siRNA or RIP140 siRNA for 60 h and were then treated with 10nM E2 for the indicated times. mRNA levels were normalized relative to 36B4 and fold change calculated relative to control. Results are the average ± SD of at least two independent experiments.

RIP140 Acts as a Coactivator at Chromatin Binding Sites That Do Not Recruit ERα

To characterize the function of RIP140 in the presence of liganded AhR, we first performed ChIP experiments to verify recruitment of RIP140 to chromatin in the presence of dioxin. As seen in Figure 8A, there was robust recruitment of the coregulator to AhR-binding sites with dioxin treatment. To assess whether RIP140 could also act as a corepressor for liganded AhR, we treated control and RIP140 knockdown cells with dioxin (Fig. 8B) and monitored effects on gene expression. In control cells, we observed increased gene expression with dioxin (Fig. 8B). In cells with knockdown of RIP140, we observed enhanced expression of the LRRC54 and HSPB8 genes in response to dioxin (Fig. 8B), as had been observed for E2 stimulation of these genes in cells depleted of RIP140 (Fig. 7). On the other hand, for CYP1A1, at which we did not observe any ERα recruitment to the gene regulatory site as shown before in Figure 2, RIP140 acted as a coactivator because RIP140 knockdown reduced the dioxin stimulation of CYP1A1 gene expression (Fig. 8B).

FIG. 8.

RIP140 has dual actions on dioxin-regulated gene expression. RIP140 acts as a corepressor of dual ligand-regulated genes where ERα and liganded AhR are both present at chromatin binding sites, whereas RIP140 acts as a coactivator of dioxin-only regulated genes where only liganded AhR is recruited to chromatin binding sites. (A) MCF-7 cells were treated with 10nM dioxin for 1 h. RIP140/DNA complexes were immunoprecipitated using specific RIP140 antibody or rabbit IgG as a negative control. (B) MCF-7 cells were transfected with control GL3 siRNA or RIP140 siRNA for 60 h and then treated with 10nM dioxin for the indicated times. mRNA levels were normalized relative to 36B4 and fold change calculated relative to control. Results are the average ± SD of at least two independent experiments.

Hence, our data imply that the direction of RIP140 regulation, repression, or activation is determined by the ligand stimulus and by the composition of the protein complexes present at specific chromatin binding sites (Fig. 9). In the presence of E2, a tripartite complex containing ERα, AhR, and RIP140 forms at dual-regulated gene sites. In this context, RIP140 acts as a corepressor, as schematized in panel A. In the presence of dioxin, two distinct complexes form (panel B); one containing ERα, AhR, and RIP140, in which RIP140 acts as a corepressor, as is seen with E2 also at dual ligand–regulated genes, and a second containing only AhR and RIP140, in which RIP140 acts as a coactivator. AhR is important in recruitment of RIP140 to chromatin, but we have demonstrated that ERα is the main determinant of the functional switch of RIP140 from a coactivator to a corepressor that controls the magnitude and duration of estrogen and dioxin signaling in these breast cancer cells.

FIG. 9.

Model showing crosstalk between ERα and AhR and the key role of the coregulator RIP140 as a corepressor or coactivator in gene regulation in breast cancer cells. (A) In cells containing ERα and AhR, treatment with E2 results in the formation of a multimeric chromatin complex containing liganded ERα, unliganded AhR, ARNT, and RIP140. In this situation, AhR and RIP140 act as corepressors of E2-ERα gene expression. (B) In cells exposed to dioxin only, and not E2, at gene sites containing EREs and DREs, liganded AhR recruits RIP140 to the unliganded ERα, AhR-ARNT complex, and RIP140 acts as a corepressor of dioxin-AhR gene expression. However, at genes such as CYP1A1, which has only DREs, RIP140 acts as a coactivator of gene expression. Hence, in panel A, the presence of AhR fine-tunes and dampens the magnitude of gene stimulation by E2-ERα. In panel B (left), dioxin-AhR gene activity is modulated by ERα and RIP140 presence at genes with both ERE- and DRE-binding sites. At sites with DREs only (panel B, right), in the absence of ERα chromatin binding, RIP140 acts as a coactivator of gene regulation. Thus, the presence or absence of ERα at a gene regulatory site appears to be the determinant of whether RIP140 functions as a corepressor or coactivator.

DISCUSSION

Although reciprocal inhibitory crosstalk between AhR and ERα has been previously documented, most of the studies have focused on the effect of ERα on dioxin-mediated gene expression, that is, when AhR is ligand occupied, and the studies examined receptor binding to chromatin at only one fixed time (Ahmed et al., 2009). Ours is the first study to examine the impact of the AhR on E2-induced ERα-mediated gene expression and to define the mechanisms by which unliganded AhR modulates ERα signaling. In doing so, we have uncovered a previously unknown mechanism for AhR modulation of ERα signaling. We have shown that E2 and dioxin can regulate common as well as distinct groups of genes in MCF-7 cells. LRRC54 and PgR were previously reported to be stimulated by dioxin (Hsu et al., 2007; Tanaka et al., 2007) and, consistent with these findings, we observed that the expression of both genes, as well as another estrogen-regulated gene, HSPB8 (Madak-Erdogan et al., 2008), was increased very effectively by either E2 or dioxin. By contrast, pS2 was stimulated only by E2, and CYP1A1 and CYP1B1 were only stimulated in the presence of dioxin.

For all genes studied except CYP1A1, ChIP assays revealed a ligand-dependent recruitment of ERα, AhR, and ARNT to the gene regulatory regions. Both E2 and dioxin were very effective in recruiting ERα, AhR, and ARNT to the ERα-binding site of genes responsive to either E2 or dioxin. As expected, E2 highly induced ERα recruitment to the pS2 ER-binding site and CYP1B1 proximal promoter (CYP1B1_ERE), but we could also detect some minor AhR and ARNT recruitment with either E2 or dioxin treatment, which was more pronounced at the CYP1B1_XRE. In contrast to an early report, which suggested recruitment of ERα to the CYP1A1 XRE with dioxin treatment (Matthews et al., 2005), later reports and this study by us did not observe any ERα recruitment to this XRE (Beischlag and Perdew, 2005; Ohtake et al., 2003; Wihlen et al., 2009), whereas we did detect very efficient recruitment of AhR and ARNT with dioxin treatment. In keeping with our observations, Ahmed et al. (2009) in ChIP-chip studies with promoter-focused microarrays showed that dioxin treatment increased the interaction between AhR and ERα at human gene promoters and increased the overlap of genomic regions bound by both receptors.

To study the significance of AhR recruitment to ER-binding sites of E2-regulated genes, we knocked down AhR without affecting ERα levels and showed that with reduction of AhR, the magnitude of E2-mediated gene stimulation was increased and that this was independent of changes in ERα or active RNA Pol II recruitment. It was important to establish that ERα levels remained unaffected because previous studies have reported that AhR is an E3 ubiquitin ligase for ERα leading to its degradation, but this process was found to occur only with cell exposure to AhR ligands, which also resulted in degradation of AhR itself (Ohtake et al., 2003, 2007, 2009; Wormke et al., 2003).

Because the increased expression of ER target genes observed upon depletion of AhR by siRNA knockdown was not associated with increased recruitment of ERα or active RNA Pol II, we examined recruitment of coregulators known to interact with both AhR and ERα. These include the coactivators SRC1, 2, and 3 (Beischlag et al., 2002; Hestermann and Brown, 2003; Kumar and Perdew, 1999), the corepressor, silencing mediator of retinoic acid and thyroid hormone receptors (SMRT) (Nguyen et al., 1999; Rushing and Denison, 2002), and RIP140 (Kumar et al., 2001, 1999). Of the various coregulators we tested, that included SRC1, 2, and 3, SMRT, and RIP140, we found that recruitment of only RIP140 was changed, i.e., reduced, after depletion of AhR.

Because we demonstrated that AhR was controlling ERα signaling through RIP140 recruitment, we reasoned that RIP140 knockdown should mimic AhR knockdown. Confirming our hypothesis, RIP140 knockdown reproduced AhR knockdown, resulting in greatly increased magnitude and duration of estrogen stimulation of target genes.

Our findings suggest that, in the absence of AhR, ERα is capable of recruiting some RIP140, though probably less efficiently, and this recruitment is sufficient to partially control the magnitude and duration of E2 stimulation. Consonant with this, RIP140 has been previously demonstrated to act as a corepressor of ERα in MCF-7 cells (Carroll et al., 2006; White et al., 2005), and RIP140 mRNA levels are upregulated by both E2 and dioxin treatments (Augereau et al., 2006a,b), indicating that the mechanism we propose might serve as a negative feedback loop for controlling the magnitude and duration of both ER and AhR signaling. Thus, upregulation of RIP140 mRNA levels by dioxin and recruitment of RIP140 to ERα-containing chromatin complexes might be reinforcing and important ways in which AhR carefully regulates ERα signaling.

To explore this mechanism further, we examined gene stimulation after RIP140 knockdown and observed an increase in magnitude and duration of gene stimulation by E2 or dioxin at genes responsive to either ligand. By contrast, we observed a dampening of dioxin-mediated stimulation of the dioxin-only regulated gene CYP1A1 after RIP140 knockdown, which suggests a coactivator role for RIP140 at this dioxin-only regulated gene. The determinant of this difference seen at the CYP1A1 gene is likely that it lacks ERα recruitment to the XRE with either of the ligands. Thus, our data suggest that ERα presence in the regulatory complex may determine the coactivator or corepressor character of RIP140 (Fig. 9).

Since RIP140 can be posttranslationally modified by acetylation (Vo et al., 2001), phosphorylation (Ho et al., 2008), or sumoylation (Rytinki and Palvimo, 2008), it is possible that enzymes performing any of these modifications might be involved in controlling the functional switch of RIP140 activity because some of these enzymes have been shown to interact with ERα-containing complexes on chromatin (Madak-Erdogan et al., 2011), thus fine tuning the magnitude and duration of estrogen signaling.

Our findings highlight that although ERα is the major mediator of RIP140 recruitment to ERα-binding sites of estrogen-regulated genes in MCF-7 breast cancer cells, AhR plays a key role in modulating RIP140 recruitment and gene transcriptional response to estrogen. Furthermore, all three proteins, ERα, AhR, and RIP140, are found in the same complex at gene regulatory sites with E2 treatment. An important observation was that AhR modulation of estrogen signaling by ERα did not require addition of an AhR ligand. The majority of previous studies have focused on the impact of unliganded ERα on liganded AhR or of liganded AhR on liganded ERα (Ahmed et al., 2009), whereas in this study, we report important actions of unliganded AhR on E2-ERα gene regulatory activity. Thus, even in the absence of added AhR ligand, we find that AhR itself acts as a repressor of ERα activity, perhaps by facilitating the recruitment of RIP140 as a corepressor. This effect of unliganded AhR on E2-ERα activity that we have studied here is distinct from the well-known antiestrogenic properties of dioxin-bound AhR (Matthews and Gustafsson, 2006; Safe et al., 2000).

In summary, we describe AhR modulation of estrogen signaling in a gene- and ligand-specific manner that involves formation of a multimeric complex of ERα, AhR, and RIP140 at the chromatin level. AhR finely controls the time course and extent of ERα-mediated gene transcription by E2 by facilitating effective recruitment of RIP140, which acts as a corepressor at estrogen-stimulated genes. However, if the chromatin complex is devoid of ERα, RIP140 recruited by AhR/ARNT then acts as a coactivator and elicits full activation of gene expression (Fig. 9). The observations support the idea that gene-specific regulation of transcription in cells containing ERα and AhR is determined by an exquisite combination of ligand stimuli, DNA binding elements, and coregulators such as RIP140 to obtain a cellular response that is appropriately and carefully controlled in its magnitude and duration.

FUNDING

The Breast Cancer Research Foundation (to B.S.K.); National Institutes of Health (P50 AT006268 from the National Center for Complementary and Alternative Medicines, the Office of Dietary Supplements, and the National Cancer Institute [to B.S.K.]; T32ES007326 from NIEHS [to Z.M.E.]).