Flightless-I Mediates the Repression of Estrogen Receptor Α Target Gene Expression by the Glucocorticoid Receptor in MCF-7 Cells

2019, 66 (1), 65-74

Original

Flightless-I mediates the repression of estrogen receptor α target gene expression by the glucocorticoid receptor in MCF-7 cells

Liu Yang and Kwang Won Jeong

Gachon Institute of Pharmaceutical Sciences, College of Pharmacy, Gachon University, Incheon 21936, Republic of Korea

Abstract. The human homologue of flightless-I (FLII) belong to the gelsolin protein family and contain a gelsolin-like domain at the C-terminus and a leucine-rich repeat (LRR) domain at the N-terminus. FLII regulates estrogen receptor alpha (ERα) and glucocorticoid receptor (GR)-mediated transcription by direct interaction through different domains, suggestive of its potential role in the crosstalk between the ERα and GR signaling pathway. Here, we demonstrate that FLII plays a critical role in GR-mediated repression of ERα target gene expression. In FLII-depleted cells, the reduction in 17-β-estradiol (E2)- induced ERα occupancy following treatment with dexamethasone (Dex) at the estrogen responsive element (ERE) site of the ERα target gene was significantly inhibited. The ERE binding of GR by the cotreatment with E2 and Dex was significantly inhibited by FLII depletion, indicating that FLII is required for the recruitment of GR at the ERE sites of ERα target genes. In addition, the recruitment of ERα-induced FLII to ERE sites was significantly reduced by Dex treatment. In protein binding assays, GR inhibited the E2-induced interaction between ERα and FLII, suggesting that GR interferes with the binding of ERα and FLII at the ERα target genes, resulting in the release of ERα and FLII from EREs. Taken together, our data reveal an unknown mechanism by which the transcription coactivator FLII regulates the GR-mediated repression of ERα target gene expression in MCF-7 cells.

Key words: Flightless-I, Estrogen receptor α, Glucocorticoid receptor, Dexamethasone

BREAST CANCER remains the leading cause of breast cancer is documented. Glucocorticoids have been cancer-related deaths among females worldwide [1-3]. shown to control apoptosis and proliferation in breast Numerous evidences have suggested that 17β-estradiol cancer models and may enhance the chemo-sensitivity of or estrogen (E2), a steroid hormone that regulates vari‐ breast cancer cells when used in combination [11, 12]. ous human physiological functions and influences However, the molecular mechanism underlying the diverse pathological processes, acts as one of the major effects of glucocorticoids in the therapy of breast cancer risk factors in breast cancer development and progres‐ is yet unclear. Similar to estrogen, glucocorticoid signal‐ sion [4, 5]. The effect of estrogen is mainly mediated via ing is regulated by binding to its cognate intracellular estrogen receptor alpha (ERα), a member of nuclear receptor known as glucocorticoid receptor (GR), a receptor superfamily of proteins that play a critical role member of the nuclear receptor superfamily of ligand- in cellular processes such as cell proliferation, differen‐ dependent transcription factors. Extensive studies have tiation, apoptosis, and migration, all of which influence focused on the crosstalk between GR and ERα to explore the development and progression of cancer [6, 7]. As a the underlying mechanism in breast cancer cells. Some result, antiestrogens that block ERα-mediated transcrip‐ findings have demonstrated that GR inhibits E2- tion and estrogen synthesis inhibitors have been devel‐ stimulated ERα target gene expression. GR occupies oped for the treatment of breast cancer [7-10]. ERα-binding regions (EBRs) through binding to the acti‐ The contribution of glucocorticoids in the treatment of vator protein 1 (AP-1) within EBRs [13, 14]. The recruit‐ ment of GR to EBRs destabilizes ERα and steroid Submitted Aug. 20, 2018; Accepted Oct. 4, 2018 as EJ18-0343 receptor coactivator-3 complex, leading to the repression Released online in J-STAGE as advance publication Oct. 26, 2018 of ERα target gene expression [15]. Furthermore, coacti‐ Correspondence to: Kwang Won Jeong, Gachon Institute of Phar‐ maceutical Sciences, College of Pharmacy, Gachon University, 191 vators such as steroid receptor coactivators (SRC-2 and Hambakmoero, Yeonsu-gu, Incheon 21936, Republic of Korea. SRC-3) and the mediator component MED14 have been E-mail: [email protected] reported to contribute to the interplay between GR and

©The Japan Endocrine Society 66 Yang et al.

ERα [16]. GR has been recently shown to bind to ERα- GGUGUUUGACAACGACdTdT-3' (sense) and 5'-GUC occupied enhancers depending on its SUMOylation and GUUGUCAAACACCUGCdTdT-3' (anti-sense); siNS, association with the N-CoR/SMRT-HDAC3 complex. 5'-UUCUCCGAACGUGUCACGUdTdT-3' (sense) and This complex may block the recruitment and the Mega‐ 5'-ACGUGACACGUUCGGAGAAdTdT-3' (anti-sense). Trans complex to repress ERα directed transcriptional Total RNA was isolated from MCF-7 cells with Trizol program [17]. (Invitrogen, Carlsbad, USA) after treatment with 10 nM Previous studies have demonstrated that flightless-I E2 or 100 nM dexamethasone (Dex; Sigma-Aldrich, (FLII) functions as a coactivator for the ERα-mediated Louis, USA) for 24 h alone or in combination. RNA was transcription [18, 19]. FLII comprises an N-terminal subjected to reverse transcription by iScript cDNA syn‐ leucine-rich repeat (LRR) and a C-terminal gelsolin-like thesis kit (Bio-Rad Laboratories, Hercules, CA, USA) in domain containing two large repeats (GelA and GelB) a total volume of 20 μL. A total of 2 μL of product was [20, 21]. The coactivator function of FLII in ERα signal‐ used for qPCR performed on a LightCycler 480II ing pathway mainly depends on the recruitment of the (Roche, Indianapolis, IN, USA) with the following SWI/SNF chromatin remodeling complex to promoters. primers: TFF1, 5'-GAACAAGGTGATCTGCG-3' To facilitate this process, FLII binds to ERα and BAF53, (forward) and 5'-TGGTATTAGGATAGAAGCACCA-3' an actin-related protein of the SWI/SNF complex, via C- (reverse); GREB1, 5'-CAAAGAATAACCTGTTGGCC terminal gelsolin-like domain in MCF-7 cells [18]. Fur‐ CTGC-3' (forward) and 5'-GACATGCCTGCGCTCTC thermore, FLII regulates GR-mediated transcription by ATACTTA-3' (reverse); CyclinD1, 5'-AAGCTCAAGTG direct binding to GR via the N-terminal LRR domain in GAACCT-3' (forward) and 5'-AGGAAGTTGTT A549 cells [22]. These findings suggest that FLII may GGGGC-3' (reverse); PgR, 5'-GTGCCTATCCTGCC play an important role in the crosstalk between ERα and TCTCAATC-3' (forward) and 5'-CCCGCCGTCGTAA GR. CTTTCG-3' (reverse); 18S, 5'-GAGGATGAGGTGG In the current study, we determined the function of AACGTGT-3' (forward) and 5'-TCTTCAGTCGCTCC FLII in the repression of ERα target gene expression by AGGTCT-3' (reverse). Results shown are mean and GR signaling in MCF-7 cells. In an attempt to elucidate range of variation of duplicate PCR reactions performed the underlying mechanism, we evaluated the role of FLII with the same cDNA sample. Relative expression levels in the regulation of ERα and GR recruitment to ERα tar‐ were normalized to the expression levels of 18S rRNA. get genes. Furthermore, we also observed the binding activity among ERα, GR, and FLII. Together, our data Chromatin immunoprecipitation (ChIP) assay describe the important role of FLII in the crosstalk We performed the ChIP assays according to previously between ERα and GR in the process of GR-mediated described protocols [18]. Briefly, MCF-7 cells were repression of ERα target gene expression. transfected with siRNAs and cultured for 3 days in hormone-free media. At approximately 90% confluency, Materials and Methods cells were treated with 100 nM E2 or 100 nM Dex alone or in combination for 60 min. After crosslinking with Plasmids and cell culture formaldehyde, the cell extracts were prepared from The following plasmids used were previously descri‐ control and hormone-treated MCF-7 cells. Sheared chro‐ bed [18, 23]: pGEX-ERα (LBD), pCDNA3.1-ERα, matin fragments were prepared by sonication. Immuno‐ pCDNA-hGR, and pTriEX-FLII. MCF-7 cells were cul‐ precipitation of sonicated chromatin solutions was tured in Dulbecco’s modified Eagle’s medium (HyClone, conducted by overnight incubation at 4°C with anti-ERα, South Logan, Utah) supplemented with 4 mM L- anti-GR, or anti-FLII antibodies (Santa Cruz Biotech‐ glutamine, 4,500 mg/L glucose, sodium pyruvate, and nology, Carlsbad, USA). Crosslinking was reversed by 10% fetal bovine serum (Gibco, Grand Island, USA) at heating overnight at 65°C, and protein-associated DNA

37°C and in an atmosphere containing 5% CO2. sequence was purified by phenol-chloroform extraction and ethanol precipitation. The purified DNA sequence RNA interference and quantitative reverse was dissolved in 100 μL of nuclease free water and ana‐ transcription PCR (RT-qPCR) lyzed by real-time quantitative PCR using LightCycler Small-interfering RNA experiments followed by RT- 480II system with SYBR Green I Master (Roche, Indian‐ qPCR were performed according to a previously pub‐ apolis, IN, USA). Results shown are mean and range of lished method [18, 24]. Transfection of MCF-7 cells was variation of duplicate PCR reactions from a single performed with Oligofectamine (Invitrogen, Carlsbad, experiment which is representative of at least three inde‐ USA) according to the manufacturer’s protocol. The pendent experiments. Results were expressed as percent‐ sequences of siRNAs were as follows: siFLII, 5'-GCA age of input chromatin (before immunoprecipitation). FLII in GR-mediated gene repression 67

The primers used were as follows: TFF1(ERE1), 5'-CCG sodium deoxycholate, 0.1% sodium dodecyl sulfate, and GCCATCTCTCACTATGAA-3' (forward) and 5'-CCTT 2 mM EDTA). Immunoblotting was performed as previ‐ CCCGCCAGGGTAAATAC-3' (reverse); TFF1(ERE2), ously described [18] using anti-ERα (Santa Cruz Bio‐ 5'-CCTCCCCAGCTCACGTTGT-3' (forward) and 5'- technology, Carlsbad, USA). GGGTTGCATTTAAGGGACCTT-3' (reverse); TFF1 (ERE3), 5'-GTCGTTGCCAGCGTTTCC-3' (forward) Statistical analysis and 5'-CTTCTCCACGCCCTGTAAATTT-3' (reverse); Statistical analyses for qPCR was carried out using GREB1(ERE1), 5'-GTGGCAACTGGGTCATTCTGA-3' GraphPad Prism software v5.01 (La Jolla, CA, USA). (forward) and 5'-CGACCCACAGAAATGAAAAGG-3' Data were analyzed using one-way analysis of variance (reverse); GREB1(Enh3), 5'-GAAGGGCAGAGCTGAT (ANOVA) followed by Tukey’s multiple comparison AACG-3' (forward) and 5'-GACCCAGTTGCCACACT test. P < 0.05 was regarded as significant. TTT-3' (reverse). Results Protein binding assay and immunoblotting For in vitro protein-protein binding assay, the proteins FLII is required for the repression of ERα-mediated ERα, GR, and FLII were synthesized by transcription transcription by GR and translation in vitro using the TNT-Quick-coupled It has been previously reported that Dex inhibited E2- transcription/translation system (Promega, Madison, stimulated ERα target gene expression [25-27]. The USA) according to the manufacturer’s protocol. In vitro mRNA levels of several endogenous ERα target genes binding assay was performed after pre-cleaning. Briefly, were measured in MCF-7 cells treated with E2, Dex, or synthesized ERα, GR, and FLII proteins were added to both. As expected, E2 treatment significantly induced NETN buffer. After incubation with protein A/G PLUS- the expression of ERα target genes (e.g., TFF1, GREB1, Agarose beads (Santa Cruz Biotechnology, Carlsbad, CCND1, and PgR), whereas Dex treatment had no effect USA) for overnight at 4°C with slow rotation, beads on these gene expressions. In comparison with E2 treat‐ were precipitated by centrifugation at a gentle speed. The ment, treatment of cells with E2 and Dex resulted in the supernatant was collected and immunoprecipitation was inhibition of the expression of these genes (Fig. 1A). conducted by incubating the supernatant with 1 μg of These results support the previous finding that Dex anti-FLII or anti-ERα antibodies and protein A/G PLUS- represses E2-induced ERα target gene expression [15]. Agarose beads at 4°C overnight in the presence of vehi‐ To further explore the contribution of FLII in Dex- cle, E2 (1 μM), Dex (1 μM), or E2 + Dex. Beads were mediated repression of ERα target gene expression, FLII washed thrice with NETN buffer and immunoblotting was depleted in MCF-7 cells by siRNA and the effect of was performed using anti-GR or anti-FLII antibodies. E2 or Dex treatment was analyzed. In comparison with Glutathione-S-transferase (GST) pull-down assay was cells transfected with non-specific siRNA (siNS), those performed based on a previously described procedure transfected with FLII-specific siRNA (siFLII) showed [18]. GST-ERα(LBD) was expressed in Escherichia coli depletion of FLII expression without any effect on ERα (BL21), purified by incubation with glutathione agarose and GR protein levels (Fig. 1B). In comparison with beads (Macherey-Nagel, Duren, Germany), and washed siNS-transfected cells, FLII-depletion resulted in a sig‐ with NETN buffer (300 mM sodium chloride [NaCl], 1 nificant inhibition of E2-induced ERα target gene (e.g., mM ethylenediaminetetraacetic acid [EDTA], 20 mM TFF1, CCND1, and PgR) expression. However, the Tris-HCl [pH 8.0], and 0.01% NP-40). GST fusion pro‐ repressive effect of Dex on E2-stimulated endogenous teins attached to beads were overnight incubated with the ERα target gene expression was compromised in FLII- in vitro synthesized FLII or GR using the TNT-Quick- depleted MCF-7 cells (Fig. 1C), suggesting that FLII is coupled transcription and translation system (Promega, essential for the repression of E2-induced endogenous Madison, USA) at 4°C, followed by analysis with west‐ ERα target gene expression by GR. ern immunoblotting using anti-FLII antibody. For co- immunoprecipitation assay, MCF-7 cells were plated in a FLII regulates the ERα and GR occupancy at ERα 10-cm dish. The next day, cells were washed twice with target genes upon Dex stimulation phosphate-buffered saline (PBS) and incubated in Having identified the potential function of FLII in the hormone-free media. After 72 h, cells were treated with process of Dex-mediated repression of ERα target gene E2 (100 nM) and Dex (100 nM) alone or in combination expression, we evaluated the underlying mechanism. A for indicated time points. Cell extracts were prepared in previous study indicated that the increase in GR recruit‐ 1.0 mL of radioimmunoprecipitation assay (RIPA) buffer ment and the decrease in ERα recruitment on ERα target (50 mM Tris-Cl [pH 8.0], 150 mM NaCl, 1% NP-40, 1% genes in response to E2 + Dex treatment may be respon‐ 68 Yang et al.

Fig. 1 FLII is required for the repression of ERα-mediated transcription by GR. (A) Effect of hormones on endogenous ERα target gene expression in MCF-7 cells. MCF-7 cells were grown in hormone-free media for 72 h, followed by treatment with vehicle (ethanol), E2 (10 nM), Dex (100 nM), or E2 + Dex (10 nM + 100 nM). After 24 h treatment, cells were harvested for RNA isolation. Total RNA was analyzed by RT-qPCR. The mRNA level was normalized to corresponding 18S RNA values. (B) Depletion of endogenous FLII by transfection of specific siRNA for FLII (siFLII). MCF-7 cells were transfected with siRNA specific for FLII or non-specific siRNA (siNS). After being cultured in hormone-free media for 72 h, cells were collected and lysed. The lysates were measured for the protein levels of FLII, GR, ERα, and β-actin by western blotting. (C) MCF-7 cells were transfected with siNS or siFLII and grown in hormone-free media for 72 h. The cells were treated with vehicle (ethanol), E2 (10 nM), Dex (100 nM), or E2 + Dex (10 nM + 100 nM) for 24 h. Total RNA was collected and analyzed by RT-qPCR. The mRNA levels of TFF1, PgR, and CCND1 were normalized to 18S RNA level. The asterisk indicates a significant difference compared with the control group. *p < 0.05 .

sible for the GR-mediated repression of these genes [15]. those treated with E2 and Dex showed reduction in the To gain insight into the function of FLII in the recruit‐ recruitment of ERα to EREs of ERα target genes (Fig. ment of ERα and GR to EREs of the enhancer or pro‐ 2A), consistent with the results of a previous study [15]. moter regions of the ERα target genes, such as ERE1-3 In the absence of FLII, the reduction in ERα occupancy of TFF1 gene [28] and ERE1, and enh3 of GREB1 gene at EREs of ERα target genes by Dex treatment was com‐ [29], we depleted endogenous FLII using siRNA against promised in cells treated with E2 and Dex (Fig. 2A). It is protein FLII and performed the ChIP assay. Upon E2 noteworthy that the depletion of endogenous FLII had no treatment, the recruitment of ERα to EREs or enhancer effect on ERα recruitment in response to E2 treatment regions of ERα target genes was significantly induced alone. (Fig. 2A). Dex treatment alone had no effect on ERα By contrast, GR recruitment to EREs of ERα target recruitment as compared with the vehicle control. How‐ genes showed a different profile. We noted that in the ever, in comparison with cells treated with E2 alone, presence of FLII, GR was loaded onto several EREs of FLII in GR-mediated gene repression 69

Fig. 2 Effect of FLII on the recruitment of ERα and GR to EREs of ERα target genes. (A & B) MCF-7 cells transfected with siRNA and grown in hormone-free media for 72 h were treated with vehicle (ethanol), E2 (100 nM), Dex (100 nM), or E2 + Dex (100 nM + 100 nM) for 60 min and subjected to ChIP assay using anti-ERα (A) or anti-GR (B) antibody. Immunoprecipitated DNA was analyzed by real-time qPCR. The results are plotted as the percentage of total input before immunoprecipitation. Error bars denote the range of variation of duplicate qPCR reactions from a single experiment which is representative of at least three independent experiments. *p < 0.05 . the promoter or enhancer regions in cells treated with E2 ERα target gene expression, we conducted ChIP assays and Dex; no GR recruitment to these regions was of FLII in MCF-7 cells treated with E2, Dex, or both. As observed in cells treated E2 or Dex alone, which sup‐ expected, FLII was recruited to EREs of the enhancer ports the previous report [15]. However, in the absence and promoter regions of ERα target genes (e.g., TFF1 of FLII, GR failed to load onto these regions even after and GREB1) upon E2 treatment (Fig. 3). Dex treatment the treatment of cells with E2 and Dex (Fig. 2B). Taken alone had no effect on FLII recruitment to these regions. together, these results suggest that FLII regulates GR However, the combination treatment of E2 and Dex occupancy at the promoter or the enhancer regions of resulted in a significant reduction in FLII recruitment to ERα target genes, resulting in the loss of ERα from these these regions as compared with E2 treatment alone. This regions in response to E2 and Dex treatment, thus, con‐ result suggests that Dex inhibited ERα target gene tributing to GR-mediated repression of ERα transcrip‐ expression likely through the attenuation of the occu‐ tional activity. pancy of FLII at ERα target genes.

GR attenuates FLII occupancy at EREs Effect of GR on the binding between ERα and FLII FLII is recruited to ERα target genes in response to E2 After confirming the regulatory function of FLII, we treatment and, thus, is critical for ER target gene expres‐ aimed to identify the possible mechanisms associated sion [18, 19]. To determine whether Dex-stimulated GR with FLII regulation. We have previously shown that affects FLII recruitment and to seek additional evidence FLII binds to GR through the LRR fragment [22]. We to substantiate the importance of FLII in GR-mediated performed an in vitro protein-protein binding analysis 70 Yang et al.

Fig. 3 GR attenuates FLII occupancy at EREs. (A & B) Chromatin immunoprecipitation assays were conducted with MCF-7 cells in 150-mm dishes. Cells were cultured in hormone-free media for 72 h. Chromatin were prepared from cells treated with vehicle (ethanol), E2 (100 nM), Dex (100 nM), or E2 + Dex (100 nM + 100 nM) for 20 min. Cross-linked chromatin fragments were immunoprecipitated with anti-FLII antibody. Immunoprecipitated DNA was analyzed by real-time qPCR. The results are plotted as the percentage of total input before immunoprecipitation. Error bars denote the range of variation of duplicate qPCR reactions from a single experiment which is representative of at least three independent experiments. *p < 0.05. using expressed proteins with the TNT-Quick-coupled assay using ERα-LBD, which shows no binding to GR transcription/translation system. FLII can bind to GR but can bind to FLII. The GST pull-down assay revealed both in the absence and presence of Dex, suggesting that that GR compromised the binding between FLII and the binding activity is hormone-independent (Fig. 4A). ERα-LBD induced by E2, suggesting that GR is capable However, as GR is translocated into the nucleus upon of inhibiting FLII binding to ERα-LBD regardless of its ligand binding, the binding of GR to FLII at the target binding to the N-terminus of ERα (Fig. 4C). Next, we genes is dependent on Dex. performed co-immunoprecipitation analysis for intracel‐ It was shown that FLII enhanced the ERα tran‐ lular binding studies. FLII showed strong binding to ERα scriptional activity by direct binding with ERα [18]. To in E2-treated cells, whereas co-treatment of E2 and Dex elucidate the mechanism underlying the GR-mediated weakened the binding between ERα and FLII induced by repression of ERα target gene expression by FLII, we E2, supporting our in vitro binding studies (Fig. 4D). tested the effect of GR on the binding between ERα and Taken together, our protein binding studies suggest FLII. Our in vitro binding assay results demonstrate that that GR recruited to the ERE through the interaction with E2 induced the binding between ERα and FLII (Fig. 4B). the LRR domain of FLII negatively regulates the binding In the presence of GR, the binding between these pro‐ between ERα and FLII. The results of our previous study teins was inhibited, while this effect was unaffected in show that the binding activity between ERα and FLII the presence of Dex. Both FLII and GR are known to played an important role in ERα transcriptional activity bind to ERα [15, 17, 18, 30]. However, ERα-ligand bind‐ [18]. The present study results demonstrate that FLII ing domain (LBD) is the major binding site for FLII plays an important role in the mechanism underlying [18], whereas the N-terminus of ERα is responsible for GR-mediated inhibition of ERα target gene expression. binding to GR [30]. We performed the same binding FLII in GR-mediated gene repression 71

Fig. 4 Effect of GR on the binding between ERα and FLII. (A & B) In vitro binding assay using FLII, GR, and ERα proteins, which were transcribed and translated in vitro. Anti-FLII or anti-ERα antibodies were used for the immunoprecipitation in the presence or absence of hormone. After an overnight incubation at 4°C, beads were washed with NETN buffer and immunoblotting was performed using anti-GR or anti-FLII antibodies. (C) GST pull-down assay using GST-fused ERα-LBD expressed in E. coli and in vitro translated FLII or GR. Bound protein was analyzed by immunoblotting using anti-FLII antibody. (D) Co-immunoprecipitation assay to monitor ERα and FLII interaction in MCF-7 cells. After being cultured in hormone-free media for 72 h, cells were treated with ethanol or hormones for indicated time points. The cell lysate was prepared in RIPA buffer. Immunoprecipitation was performed using normal mouse IgG or anti-FLII antibodies. Immunoblotting was performed using anti-ERα antibody.

Discussion recognition of EREs or through indirect interaction with other factors such as AP-1 [35]. GR may occupy ERα- Studies have demonstrated that GR and ERα alter each binding regions via tethering with AP1 in the presence of other’s regulatory and phenotypic roles. For instance, FOXA1. GR loading onto EBRs results in the destabili‐ GR expression is associated with more favorable clinical zation of ERα and SRC-3 complex, leading to the inhibi‐ outcomes in breast cancer [30-33]. Previous studies have tion of ERα activity. Another report have shown that GR documented the repressive actions of GR on specific loading onto prolactin (PRL) array, comprising estrogen- ERα transcriptional responses [15]. The activation of responsive transcriptional reporter gene derived from the both GR and ERα in cells results in GR-mediated modu‐ promoter and enhancer regions of the rat prolactin gene, lation of chromatin accessibility, allowing the binding requires co-regulators MED14, SRC-2, and SRC-3 [16]. between ERα and new set of genomic regions through an Furthermore, GR has been recently implicated in trans- assisted loading mechanism [34, 35]. The loading of repression of ERα-regulated gene expression through its ERα to the newly accessible sites depends on the direct recruitment to the ERα-bound enhancers in breast cancer. 72 Yang et al.

Fig. 5 Proposed model for the function of FLII in GR-mediated inhibition of ERα signaling. (A) FLII and other transcriptional complexes are recruited to ERα response elements in response to E2, resulting in the activation of ERα-mediated gene expression. (B) In the presence of Dex, GR recruited to EREs through the interaction with FLII and AP-1 induces the dissociation of the coactivator complex and ERα from EREs.

The recruitment of GR depends on its SUMOylation and bution of FLII in GR-mediated repression of the expres‐ requires N-CoR/SMRT-HDAC3 complex, which disas‐ sion of ERα target genes (Fig. 5). We have demonstrated sembles the E2-induced MegaTrans complex at ERα- that FLII is required for Dex-mediated inhibition of E2- activated enhancers [17]. induced ERα transcriptional activity. This effect of FLII We have recently shown that FLII augments GR- appears to be dependent on the ability to control the ERE mediated transcription, wherein its functions are exerted recruitment of ERα and GR. The ChIP assay results in through the LRR domain [22]. These results led us to FLII-depleted cells provide evidence that GR binding to speculate that FLII may be a potential regulator of ERα EREs depends on FLII. Furthermore, the inhibition of and GR. The molecular mechanism underlying the role ERα transcriptional activity by GR was associated with of FLII in ERα signaling pathway has been well studied the inhibitory effect of GR on the binding between ERα in a previous study [18, 19]. FLII directly binds to ERα- and FLII. The interaction between ERα and FLII was LBD via G3 fragment within the GelA domain and to the compromised by GR in vitro in MCF-7 cells. Suppres‐ actin-related segment of the SWI/SNF complex, BAF53, sion of the binding between ERα and FLII by GR was through G1. These binding activities contribute to the probably owing to the destabilization of the FLII-bound recruitment of the SWI/SNF chromatin remodeling com‐ protein complex in the presence of GR and FLII. It plex to promoters, which further enhance ERα-mediated appears that GR binding to the ERE site weakens the transcription. binding of coactivators such as ERα and FLII, resulting Here, we propose a possible mechanism for the contri‐ in the release of FLII from the ERα target gene. FLII is FLII in GR-mediated gene repression 73 associated with CARM1, SENP3 and MLL1/2 complex, determining the susceptibility to dexamethasone as an which catalyze histone H3 methylation and lead to tran‐ anti-cancer therapy. scription activity [36, 37]. This complex also determines RNA polymerase II recruitment to ERα target genes [18, Acknowledgements 19]. Release of FLII from EREs by GR recruitment is likely to result in the loss of these FLII-interacting tran‐ This research was supported by Basic Science scription coactivators, including SWI/SNF chromatin Research Program through the National Research Foun‐ remodeling complex. These results suggest that FLII dation of Korea (NRF) funded by the Ministry of Educa‐ plays an important role in the crosstalk between ERα and tion (2017R1D1A1B03031165). GR. The role of FLII in breast cancer cells seems to be bilateral. Increased expression of the ER target gene by Conflict of Interest FLII may be beneficial for the proliferation of estrogen- induced breast cancer cells, while at the same time the None of the authors has any potential conflicts of expression level of FLII may be an important factor in interest associated with this research.

References

1. Jemal A, Bray F, Center MM, Ferlay J, Ward E, et al. (2006) Genome-wide analysis of estrogen receptor bind‐ (2011) Global cancer statistics. CA Cancer J Clin 61: 69– ing sites. Nat Genet 38: 1289–1297. 90. 14. Kushner PJ, Agard DA, Greene GL, Scanlan TS, Shiau 2. Bray F, Ren JS, Masuyer E, Ferlay J (2013) Global esti‐ AK, et al. (2000) Estrogen receptor pathways to AP-1. J mates of cancer prevalence for 27 sites in the adult popu‐ Steroid Biochem Mol Biol 74: 311–317. lation in 2008. Int J Cancer 132: 1133–1145. 15. Karmakar S, Jin Y, Nagaich AK (2013) Interaction of glu‐ 3. Group EBCTC (1990) Treatment of early breast cancer. 1. cocorticoid receptor (GR) with estrogen receptor (ER) α Worldwide evidence 1985–1990, Oxford University Press, and activator protein 1 (AP1) in dexamethasone-mediated USA. interference of ERα activity. J Biol Chem 288: 24020– 4. Deroo BJ, Korach KS (2006) Estrogen receptors and 24034. human disease. J Clin Invest 116: 561–570. 16. Bolt MJ, Stossi F, Newberg JY, Orjalo A, Johansson HE, 5. Manavathi B, Dey O, Gajulapalli VNR, Bhatia RS, et al. (2013) Coactivators enable glucocorticoid receptor Bugide S, et al. (2013) Derailed estrogen signaling and recruitment to fine-tune estrogen receptor transcriptional breast cancer: an authentic couple. Endocr Rev 34: 1–32. responses. Nucleic Acids Res 41: 4036–4048. 6. Thomas C, Gustafsson JÅ (2011) The different roles of 17. Yang F, Ma Q, Liu Z, Li W, Tan Y, et al. (2017) Gluco‐ ER subtypes in cancer biology and therapy. Nat Rev Can‐ corticoid receptor: MegaTrans switching mediates the cer 11: 597–608. repression of an ERα-Regulated transcriptional program. 7. Early Breast Cancer Trialist’ Collaborative Group Mol Cell 66: 321–331. (EBCTCG) (2005) Effects of chemotherapy and hormonal 18. Jeong KW, Lee Y-H, Stallcup MR (2009) Recruitment of therapy for early breast cancer on recurrence and 15-year the SWI/SNF chromatin remodeling complex to steroid survival: an overview of the randomised trials. Lancet hormone-regulated promoters by nuclear receptor coacti‐ 365: 1687–1717. vator flightless-I. J Biol Chem 284: 29298–29309. 8. Kistner RW, Smith OW (1960) Observations on the use of 19. Jeong KW (2014) Flightless I (Drosophila) homolog facil‐ a non-steroidal estrogen antagonist: MER-25. Surg Forum itates chromatin accessibility of the estrogen receptor 10: 725–729. alpha target genes in MCF-7 breast cancer cells. Biochem 9. Cole MP, Jones CT, Todd ID (1971) A new anti- Biophys Res Commun 446: 608–613. oestrogenic agent in late breast cancer: an early clinical 20. Liu Y-T, Yin HL (1998) Identification of the binding part‐ appraisal of ICI46474. Br J Cancer 25: 270–275. ners for flightless I, a novel protein bridging the leucine- 10. Kiang DT, Kennedy BJ (1977) Tamoxifen (antiestrogen) rich repeat and the gelsolin superfamilies. J Biol Chem therapy in advanced breast cancer. Ann Intern Med 87: 273: 7920–7927. 687–690. 21. Kopecki Z, Cowin AJ (2008) Flightless I: an actin- 11. Moutsatsou P, Papavassiliou AG (2008) The glucocorti‐ remodelling protein and an important negative regulator of coid receptor signalling in breast cancer. J Cell Mol Med wound repair. Int J Biochem Cell Biol 40: 1415–1419. 12: 145–163. 22. Jin HL, Yang L, Jeong KW (2017) Flightless-I homolog 12. Vilasco M, Communal L, Mourra N, Courtin A, Forgez P, regulates glucocorticoid receptor-mediated transcription et al. (2011) Glucocorticoid receptor and breast cancer. via direct interaction of the leucine-rich repeat domain. Breast Cancer Res Treat 130: 1–10. Mol Biol Rep 44: 243–250. 13. Carroll JS, Meyer CA, Song J, Li W, Geistlinger TR, et al. 23. Lim MS, Jeong KW (2014) Role of Flightless-I (Droso‐ 74 Yang et al.

phila) homolog in the transcription activation of type I 14: 707–719. collagen gene mediated by transforming growth factor 31. Abduljabbar R, Negm OH, Lai CF, Jerjees DA, Al-Kaabi beta. Biochem Biophys Res Commun 454: 393–398. M, et al. (2015) Clinical and biological significance of 24. Lee JM, Park SJ, Im DS (2017) Calcium signaling of glucocorticoid receptor (GR) expression in breast cancer. lysophosphatidylethanolamine through LPA1 in human Breast Cancer Res Treat 150: 335–346. SH-SY5Y neuroblastoma cells. Biomol Ther (Seoul) 25: 32. Pan D, Kocherginsky M, Conzen SD (2011) Activation of 194–201. the glucocorticoid receptor is associated with poor prog‐ 25. Perou CM, Jeffrey SS, Van De Rijn M, Rees CA, Eisen nosis in estrogen receptor-negative breast cancer. Cancer MB, et al. (1999) Distinctive gene expression patterns in Res 71: 6360–6370. human mammary epithelial cells and breast cancers. Proc 33. West DC, Kocherginsky M, Tonsing-Carter EY, Dolcen Natl Acad Sci U S A 96: 9212–9217. DN, Hosfield DJ, et al. (2018) Discovery of a glucocorti‐ 26. Perou CM, Sørlie T, Eisen MB, van de Rijn M, Jeffrey SS, coid receptor (GR) activity signature using selective GR et al. (2000) Molecular portraits of human breast tumours. antagonism in ER-negative breast cancer. Clin Cancer Res Nature 406: 747–752. 24: 3433–3446. 27. Horwitz KB, Costlow ME, McGuire WL (1975) MCF-7: a 34. Voss TC, Schiltz RL, Sung MH, Yen PM, human breast cancer cell line with estrogen, androgen, Stamatoyannopoulos JA, et al. (2011) Dynamic exchange progesterone, and glucocorticoid receptors. Steroids 26: at regulatory elements during chromatin remodeling 785–795. underlies assisted loading mechanism. Cell 146: 544–554. 28. Pan YF, Wansa KD, Liu MH, Zhao B, Hong SZ, et al. 35. Miranda TB, Voss TC, Sung MH, Baek S, John S, et al. (2008) Regulation of estrogen receptor-mediated long (2013) Reprogramming the chromatin landscape: interplay range transcription via evolutionarily conserved distal of the estrogen and glucocorticoid receptors at the response elements. J Biol Chem 283: 32977–32988. genomic level. Cancer Res 73: 5130–5139. 29. Sun J, Nawaz Z, Slingerland JM (2007) Long-range acti‐ 36. Nayak A, Reck A, Morsczeck C, Müller S (2017) vation of GREB1 by estrogen receptor via three distal Flightless-I governs cell fate by recruiting the SUMO iso‐ consensus estrogen-responsive elements in breast cancer peptidase SENP3 to distinct HOX genes. Epigenetics cells. Mol Endocrinol 21: 2651–2662. Chromatin 10: 15. 30. West DC, Pan D, Tonsing-Carter EY, Hernandez KM, 37. Lee YH, Campbell HD, Stallcup MR (2004) Developmen‐ Pierce CF, et al. (2016) GR and ER coactivation alters the tally essential protein flightless I is a nuclear receptor expression of differentiation genes and associates with coactivator with actin binding activity. Mol Cell Biol 24: improved ER+ breast cancer outcome. Mol Cancer Res 2103–2117.