Proinflammatory Genes Estrogen Receptors Occurs at a Subset of Cross Talk Between Glucocorticoid

Cross Talk between Glucocorticoid and Estrogen Receptors Occurs at a Subset of Proinflammatory Genes

This information is current as Aleksandra Cvoro, Chaoshen Yuan, Sreenivasan Paruthiyil, of September 28, 2021. Oliver H. Miller, Keith R. Yamamoto and Dale C. Leitman J Immunol 2011; 186:4354-4360; Prepublished online 28 February 2011; doi: 10.4049/jimmunol.1002205

http://www.jimmunol.org/content/186/7/4354 Downloaded from

Supplementary http://www.jimmunol.org/content/suppl/2011/02/28/jimmunol.100220 Material 5.DC1 http://www.jimmunol.org/ References This article cites 24 articles, 13 of which you can access for free at: http://www.jimmunol.org/content/186/7/4354.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision by guest on September 28, 2021 • No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2011 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Cross Talk between Glucocorticoid and Estrogen Receptors Occurs at a Subset of Proinﬂammatory Genes

Aleksandra Cvoro,*,† Chaoshen Yuan,‡ Sreenivasan Paruthiyil,*,† Oliver H. Miller,*,† Keith R. Yamamoto,x and Dale C. Leitman*,†,‡,x

Glucocorticoids exert potent anti-inflammatory effects by repressing proinflammatory genes. We previously demonstrated that estrogens repress numerous proinflammatory genes in U2OS cells. The objective of this study was to determine if cross talk occurs between the glucocorticoid receptor (GR) and estrogen receptor (ER)a. The effects of dexamethasone (Dex) and estradiol on 23 proinflammatory genes were examined in human U2OS cells stably transfected with ERa or GR. Three classes of genes were regulated by ERa and/or GR. Thirteen genes were repressed by both estradiol and Dex (ER/GR-repressed genes). Five genes were repressed by ER (ER-only repressed genes), and another five genes were repressed by GR (GR-only repressed genes). To examine if cross talk occurs between ER and GR at ER/GR-repressed genes, U2OS-GR cells were infected with an adenovirus that Downloaded from expresses ERa. The ER antagonist, ICI 182780 (ICI), blocked Dex repression of ER/GR-repressed genes. ICI did not have any effect on the GR-only repressed genes or genes activated by Dex. These results demonstrate that ICI acts on subset of proinflammatory genes in the presence of ERa but not on GR-activated genes. ICI recruited ERa to the IL-8 promoter but did not prevent Dex recruitment of GR. ICI antagonized Dex repression of the TNF response element by blocking the recruitment of nuclear coactivator 2. These findings indicate that the ICI–ERa complex blocks Dex-mediated repression by interfering with nuclear coactivator 2 recruitment to GR. Our results suggest that it might be possible to exploit ER and GR cross talk for http://www.jimmunol.org/ glucocorticoid therapies using drugs that interact with ERs. The Journal of Immunology, 2011, 186: 4354–4360.

major event in the activation of the immune response domain in the p65 subunit of NF-kB (4). However, GR does not is the synthesis and release of several cytokines that interfere with the binding of NF-kB to DNA to cause repression. A trigger inflammatory pathways. If overexpressed proin- Instead, GR inhibits the transcriptional activation function of NF- flammatory cytokines can cause excessive inflammation and tissue kB after it binds to the promoter of target genes by different damage that can lead to the development of variety of human mechanisms depending on the gene (5, 6). For instance, when GR diseases. Glucocorticoids are the most widely used immunosup- binds to NF-kB at the IL-8 promoter, it inhibits the phosphorylation by guest on September 28, 2021 pressive and anti-inflammatory agents for treating inflammatory of serine 2 (Ser2) in the C-terminal domain (CTD) of the largest disorders. The anti-inflammatory effects of glucocorticoids are subunit of RNA polymerase II (RNAPII) (7). GR does this by mediated through the glucocorticoid receptor (GR). A major mech- competing with Ser2 CTD kinase positive transcription elongation anism whereby glucocorticoids exert anti-inflammatory actions is factor b (P-TEFb) (8), which consists of a heterodimer of the C-type by causing the repression of proinflammatory genes (1). Gluco- cyclin and cyclin-dependent kinase 9. Cyclin-dependent kinase 9 corticoids initiate gene repression by binding to GR. This allows GR phosphorylates the CTD of RNAPII at Ser2 and Ser5 (9). Because k to interact directly with transcription factors, such as NF- B (2), P-TEFb is required for productive elongation of the RNAPII initi- which is a key transcription factor that binds to and activates many ation complex (9), the inhibition of P-TEFb binding by GR prevents proinflammatory genes (3). GR binds to the C-terminal activation elongation of transcripts. Although glucocorticoids are the most extensive anti-inflam- *Department of Obstetrics, Gynecology, and Reproductive Sciences, University of Cal- matory steroids studied, other compounds that interact with nu- ifornia, San Francisco, San Francisco, CA 94143; †Center for Reproductive Sciences, clear receptors also have anti-inflammatory properties. One im- ‡ University of California, San Francisco, San Francisco, CA 94143; Department of portant action of estrogens that is relatively unappreciated is its Nutritional Science and Toxicology, University of California, Berkeley, Berkeley, CA 94720; and xCellular and Molecular Pharmacology, University of California, San Fran- anti-inflammatory effects (10). Estrogens exert their effects cisco, San Francisco, CA 94143 through binding to two distinct estrogen receptors (ERs), ERa or Received for publication July 1, 2010. Accepted for publication January 27, 2011. ERb (11). Similarly to GR, it has been demonstrated that ERs exert This work was supported by Grant DK061966 from the National Institutes of Health their anti-inflammatory activity by repressing the expression of (to D.C.L.). multiple NF-kB–driven cytokine genes (12–15). Early in vitro Address correspondence and reprint requests to Dr. Dale C. Leitman, Department of studies suggested that ERs repress genes by blocking the binding of Nutritional Science and Toxicology, University of California, Berkeley, 44 Morgan Hall, Berkeley, CA 94720. E-mail address: [email protected] NF-kB to the promoter (12, 13), but other studies revealed a much The online version of this article contains supplemental material. more complex mechanism for ER-mediated repression. Our studies a Abbreviations used in this article: Ad-ERa, adenovirus expressing human estrogen with U2OS cells stably transfected with ERs (U2OS-ER and receptor a; ChIP, chromatin immunoprecipitation; CTD, C-terminal domain; Dex, U2OS-ERb) showed that TNF-a treatment leads to the recruitment dexamethasone; Dox, doxycycline; E2, estradiol; ER, estrogen receptor; GR, gluco- of unliganded ERa to the composite c-Jun/NF-kB binding site in corticoid receptor; ICI, ICI 182780; MOI, multiplicity of infection; NCOA2, nuclear coactivator A2; P-TEFb, positive transcription elongation factor b; qPCR, quantita- the TNF-a promoter (16). Once unliganded ERa associates with tive PCR; RNAPII, RNA polymerase II; Ser, serine; TNF-RE, TNF-a response NF-kB and c-Jun, it functions as a coactivator that activates the element. TNF-a gene. Estradiol (E2) treatment causes repression of the Copyright Ó 2011 by The American Association of Immunologists, Inc. 0022-1767/11/$16.00 TNF-a gene by recruiting nuclear coactivator 2 (NCOA2), which www.jimmunol.org/cgi/doi/10.4049/jimmunol.1002205 The Journal of Immunology 4355 functions as a corepressor in this context (16). A corepressor gene (21) upstream of the minimal thymidine kinase promoter in pGL4.22 a a function for NCOA2 was previously reported for GR-mediated (Promega) into U2OS-ER cells. U2OS–ER –TNF-RE cells were se- lected and maintained in 50 mg/ml hygromycin B, 500 mg/ml zeocin, and 1 repression of the collagenase-3 gene (17). The complexity of ER- mg/ml puromycin. U2OS-GR cells were a generous gift from Dr. Michael mediated transcriptional repression is also revealed on different NF- Garabedian. U2OS-GR cells were supplemented with 500 mg/ml geneticin kB response elements in MCF-7 cells (18). In the IL-6 promoter, (G418; Invitrogen). ECC-1 cells were obtained from American Type ERa displaced p65 and associated coregulators from NF-kB Culture Collection. All cell lines were cultured continuously in phenol red- binding site, whereas in the case of the MCP-1 and IL-8 promoters, free DMEM/F-12 medium (Invitrogen) containing 5% charcoal/dextran- a k stripped FBS (Gemini Bio-Products), 2 mM glutamine, 100 U/ml peni- ER displaces CREB binding protein from NF- B sites (18). The cillin and streptomycin, and 50 mg/ml fungizone. Adenoviruses expressing displacement of these transcription factors and coregulators pre- human ERa (Ad-ERa) and the control virus (Ad-LacZ) were purchased vents them from activating these proinflammatory genes. from Viraquest. The studies with GR and ER show that the two receptors share Real-time PCR some common mechanisms, but they also exploit different mechanisms to repress proinflammatory genes depending on the tar- Total RNA was prepared using the Aurum Total RNA Kit (Bio-Rad) get gene, cell type, and probably interactions with other tran- according to the manufacturer’s protocol. A total of 1 mg total RNA was reverse transcribed using the iScript Kit (Bio-Rad). Real-time quantitative scription factors. Despite the findings that both GR and ER repress PCR was performed with the Bio-Rad iCycler Thermal Cycler System proinflammatory cytokine genes, little is known about the potential using SYBR Green Supermix (Bio-Rad). The sequences of the primers are cross talk between GR and ER. However, in an experimental lung listed in Supplemental Table I. The data were collected and analyzed using inflammation rat model, the ER antagonist ICI 182780 (ICI) the comparative threshold cycle method using b-actin as the reference 6 blocked the anti-inflammatory effects of glucocorticoids (19). This gene. Experiments were performed at least three times, and the mean SE Downloaded from was calculated using the Prism curve-fitting program (GraphPad Software, finding suggests that when ER is bound to an antagonist, it can version 3.03; GraphPad). cross talk with GR to block its anti-inflammatory action, but mechanism for this effect is unknown. The purpose of this study Western blotting was to investigate if proinflammatory genes are specifically reg- U2OS-GR cells were infected with Ad-LacZ or Ad-ERa for 24 h. Total ulated by GR or ER and to investigate the molecular mechanism proteins were extracted in radioimmunoprecipitation assay buffer con- for GR-ER cross talk at proinflammatory genes. taining protease and phosphatase inhibitors. Immunoblotting of proteins http://www.jimmunol.org/ was performed following standard procedures using an Ab to ERa (1D5; DakoCytomation). Western blots for GR were done in U2OS–ERa–TNF- Materials and Methods RE cells using an Ab to GR (PA1-512; Thermo Scientific). An ECL de- Cell culture and reagents tection system (GE Healthcare) was used to visualize the proteins. U2OS-ERa cells were prepared and maintained in media supplemented Luciferase assays with 50 mg/ml hygromycin B and 500 mg/ml zeocin as previously described (20). U2OS–ERa-TNF response element (TNF-RE) cells were U2OS–ERa–TNF-RE cells were treated with doxycycline (Dox) for 18 h prepared by transfecting the TNF-a responsive element from the TNF-a to induce ERa, and then drugs were added to the medium as indicated in

Table I. Dex and E2 repress common and unique proinﬂammatory genes by guest on September 28, 2021

U2OS-GR U2OS-ERa

Gene Fold TNF-a Percent of Fold TNF-a Percent of Symbol Accession No. Activation Repression p Value Activation Repression p Value

a TNF-a–activated genes that are repressed by E2 and Dex CD69 NM 001781 9.1 86.8 0.00393 12.4 90.1 0.00047 CSF2 NM 000758 9.3 76.1 0.00594 11.6 78.1 0.00224 C8orf NM 020130 4.9 75.4 0.00082 6.2 73.6 0.00055 GRO3 NM 002090 12.2 84.3 0.00037 13.6 88.3 0.00557 MCP-1 NM 002982 9.2 71.0 0.00253 9.2 73.2 0.00078 IL6 NM 000600 14.7 87.2 0.00685 17.0 86.3 0.00176 IL8 NM 000584 16.4 78.2 0.00224 18.1 85.5 0.00169 LTB NM 002341 3.4 73.9 0.00033 4.0 74.0 0.00331 TNFAIP6 NM 007115 3.3 50.1 0.01358 3.7 42.4 0.00021 TNFa NM 000594 18.4 69.8 0.00101 19.1 83.3 0.00054 BMP2 NM 001200 2.4 56.4 0.00200 2.9 57.3 0.00631 EFNA1 NM 004428 8.4 64.5 0.00016 7.4 61.1 0.00585 IER3 NM 003897 3.4 63.4 0.00956 3.9 61.7 0.00440 a TNF-a–activated genes that are repressed by E2 CCL20 NM 004591 11.3 0.9 0.46394 12.9 98.2 0.00239 CCL4 NM 002984 11.9 3.9 0.41457 11.6 83.6 0.00519 PTX3 NM 002852 17.5 24.2 0.43920 18.7 93.3 0.00031 TNFAIP3b NM 006290 5.3 258.9 0.02500 8.8 75.2 0.00138 EDN1 NM 001955 4.8 22.9 0.39949 4.9 81.5 0.00463 TNF-a–activated genes that are repressed by Dexa TRAF1 NM 005658 5.3 62.6 0.00809 5.7 24.0 0.27693 TNFRSF9 NM 001561 6.7 64.7 0.01316 7.7 2.1 0.40420 TNFAIP2 NM 006291 10.4 60.5 0.00726 10.7 22.3 0.20276 LIF NM 002309 2.8 78.7 0.00085 2.1 20.9 0.42512 CD83 NM 004233 3.7 50.6 0.00686 3.2 20.9 0.37140 aU2OS-GR and U2OS-ERa stable cells were treated with 2.5 ng/ml TNF-a for 2 h in the absence or presence of 10 nM Dex (U2OS-GR) or 10 nM E2 (U2OS-ERa) for 2 h, total RNA was isolated, and real-time RT-PCR was performed. Real-time RT- PCR results are shown. Each data point is the average obtained from three individual U2OS cell samples 6 SEM. Analyses were performed using the Prism curve-ﬁtting program. bTNFAIP3 is activated by Dex but repressed with E2. 4356 GR AND ER TRANSREPRESSION AND CROSS TALK

Chromatin immunoprecipitation U2OS-GR cells were infected with Ad-LacZ or Ad-ERa for 24 h. After the treatments as indicated in the figure legends, the cells were cross-linked and washed, and collected chromatin was sheared with sonication. Im- munoprecipitations were performed overnight at 4˚C with anti-ERa (1D5, or HC-20) or anti-GR Abs (N499 or PA1-510A) as previously described (16, 22). U2OS–ERa–TNF-RE cells were treated with drugs, and then chromatin immunoprecipitation (ChIP) was done using anti-GR or KAT13C/NCOA2 (ab9261-50; Abcam) Abs. Chromatin samples were purified using the QIAquick Purification Kit (Qiagen), and amplification of samples at the IL8 gene was done using quantitative PCR (qPCR). ChIP primers that spanned the TNF-RE in pGL4.22 or NF-kB element in the IL8 promoter (8) were used. Experiments were done in triplicate, the mean 6 SE was calculated, and statistical analysis was performed as previously described (23).

Results ER and GR repress similar and unique proinﬂammatory genes

We previously identified proinflammatory genes activated by Downloaded from TNF-a that are repressed by E2 in U2OS osteosarcoma cells stably transfected with human ERa or ERb (16, 23). To determine if GR produces similar anti-inflammatory effects as ERs, we tested if glucocorticoids also repress these genes in U2OS cells stably transfected with GR. U2OS-GR or U2OS-ERa cells were treated FIGURE 1. Effect of Dex and E on repression of the inflammatory genes a 2 with TNF- for 2 h in the absence or presence of dexamethasone http://www.jimmunol.org/ a in U2OS cells. U2OS-GR and U2OS-ER stable cells were treated with 2.5 (Dex) or E2, respectively. We found that E2 and/or Dex regulated ng/ml TNF-a for 2 h in the absence or presence of 10 nM Dex (U2OS-GR) three classes of proinflammatory genes. Thirteen genes tested by a and 10 nM E2 (U2OS-ER ) for 2 h, total RNA was isolated, and real-time real-time PCR were activated by TNF-a and repressed by E and a 2 RT-PCR was performed. A, TNF- –activated genes that are repressed by Dex (ER/GR-repressed genes, Table I). Five TNF-a–activated Dex and E . Real-time RT-PCR results for representative genes (TNF-a and 2 genes were repressed only by ERa (ER-only repressed genes, CD69) repressed by both Dex and E2 are shown. Each data point is the average obtained from three individual U2OS cell samples 6 SEM. B, Table I) including TNFAIP3, which was activated by GR. Another a Differential transcriptional repression of inflammatory genes by ER and GR five genes were repressed by GR, but not ER (GR-only repressed in U2OS cells. Real-time RT-PCR results for representative genes repressed genes, Table I). Representative real-time PCR shows that Dex and

a a by guest on September 28, 2021 by Dex (LIF) or by E2 only (TNFAIP3) are shown. Each data point is the E2 blocked the induction of the TNF- activation of the TNF- average obtained from three individual U2OS cell samples 6 SEM. and CD69 genes in U2OS-GR (Fig. 1A, left panels) and U2OS- ERa (Fig. 1A, right panels) cells, respectively. The TNFAIP3 gene (Fig. 1B, upper panels) was repressed only by E2, whereas only the ﬁgure legends for 18 h. The cells were solubilized with lysis buffer, and Dex repressed the LIF gene (Fig. 1B, lower panels). These data then the plates were frozen. The samples were thawed and assayed for demonstrate that Dex and E commonly repress multiple proin- luciferase activity using the Luciferase Assay System (Promega). Tripli- 2 cate samples for each treatment were measured for luciferase activity using ﬂammatory genes, but some genes are differentially repressed by a luminometer. The experiments were repeated at least three times. ER and GR.

FIGURE 2. Effect of Dex and E2 on gene regulation in U2OS cells expressing both GR and ER. U2OS-GR stable cells were infected with 50 MOI Ad-LacZ (A) or Ad-ERa (B) for 24 h. The cells were then treated with 2.5 ng/ml TNF-a for 2 h in the absence or presence of 10 nM Dex,

10 nM E2, or both drugs for 2 h. Cells were treated with ICI 30 min prior to Dex and E2, and real-time RT-PCR for IL8 was performed. C and D, U2OS-GR stable cells were infected as in B and treated with

Dex or E2 for 4 h. ICI (1 mM) and RU486 (1 mM) were administered 30 min prior to Dex, and real-time RT-PCR for IGFBP and IRF8 genes was performed. Error bars represent the mean 6 SEM. The Journal of Immunology 4357

U2OS-GR cells infected with Ad-LacZ, which was repressed by Dex (Fig. 2A). In the absence of ER, E2 and the ERa antagonist ICI had no effect on the IL8 gene with or without Dex (Fig. 2A). When the U2OS-GR cells were infected with Ad-ERa, there was repression of the IL8 gene with both E2 and Dex (Fig. 2B). The combination of E2 and Dex did not produce a greater repression compared with the individual drugs (Fig. 2B). Similar ﬁndings were found with all 13 GR/ER-repressed genes (data not shown). The GR-antagonist RU486 did not block repression by E2 of the IL8 gene (Supplemental Fig. 1). RU486 also did not antagonize Dex repression of three TNF-a–activated genes in the presence or absence of ERa (Supplemental Fig. 2). In contrast, ICI blocked the repression of the IL8 gene by E2 (Fig. 2B). Surprisingly, ICI also blocked the Dex repression of the IL8 gene (Fig. 2B) and other ER/GR-repressed genes (data not shown). ICI did not antagonize Dex activation of the IGFBP (Fig. 2C) and IRF8 (Fig. 2D) genes in U2OS-GR cells, whereas RU486 did. These results demonstrate that RU486 blocks activation, but not repression by

Dex. To determine if the antagonistic action of ICI was dependent Downloaded from on the levels of ERa, U2OS-GR cells were infected with increasing amounts of Ad-ERa. RT-PCR (Fig. 3A) and Western blots (Fig. 3B) show that there was a dose-dependent increase in amount of ERa in the cells with increasing multiplicity of infection (MOI) of Ad-ERa. The antagonistic action of ICI on Dex

repression was dependent on the levels of ERa in the cells (Fig. http://www.jimmunol.org/ a FIGURE 3. The antagonistic effect of ICI on Dex repression of genes is 3C). Antagonism by ICI was ﬁrst observed at 1 MOI Ad-ER , dependent on the presence of ERa. A and B, U2OS-GR cells were infected which produced very low levels of ERa as shown by the Western with increasing amount of Ad-ERa, and level of ERa expression was blot (Fig. 3B). At 50 MOI Ad-ERa, ICI abolished Dex repression measured. ERa mRNA level was measured by qPCR (A). Western blots of the IL8 gene. These results demonstrate that the ER antagonist (B) were performed using cell extracts, and ERa levels were detected with ICI but not the GR antagonist RU486 blocks GR-mediated re- ERa Ab. The + lane represents extracts from U2OS-ERa cells induced pression of ER/GR-repressed genes and that the antagonistic ef- with Dox. The 2 lane represents uninfected U2OS-GR cells. C, U2OS-GR fect of ICI is dependent on the presence of ERa in the cells. cells infected with indicated amount of Ad-ERa were treated with TNF-a m (2.5 ng/ml), Dex (10 nM), and ICI (1 M) as shown. qPCR was used to by guest on September 28, 2021 measure IL8 mRNA levels. Error bars represent the mean 6 SEM. ICI does not block GR-only repressed genes To further explore the action of ICI, we also examined its effects on an ER-only (CCL20) or a GR-only (TNFAIP2) repressed gene. ICI blocks Dex repression of ER/GR-repressed genes U2OS-GR cells were infected with Ad-ERa or Ad-LacZ and then To determine if cross talk exists between ERa and GR, we infected treated with TNF-a in the absence or presence of E2 or Dex. E2 U2OS-GR stable cells with the control adenovirus Ad-LacZ (Fig. repressed TNF-a activation of the CCL20 gene in U2OS-GR cells 2A) or Ad-ERa (Fig. 2B–D). TNF-a activated the IL8 gene in infected with Ad-ERa (Fig. 4A), but not Ad-LacZ (Fig. 4B). The

FIGURE 4. ICI blocks transcriptional repression of ER-only but not GR-only repressed genes in U2OS cells expressing both GR and ER. U2OS-GR stable cells were infected with Ad-ERa (A, C) or Ad-LacZ (B, D) as described in Fig. 2 and treated as indicated. Total RNA was isolated, and the level of CCL20 (A, B) and TNFAIP2 (C, D) mRNA was measured by real-time PCR. Each data point is the average obtained from three individual U2OS cell samples 6 SEM. 4358 GR AND ER TRANSREPRESSION AND CROSS TALK

sence (Fig. 4D)ofERa. The other GR-only genes were repressed by Dex, but not by E2 or ICI (data not shown). These data indicate that cross talk between ER and GR occurs only in the subset of genes that are repressed by both ER and GR.

ICI recruits ERa to the IL8 promoter To examine the mechanism whereby ICI inhibits GR-mediated FIGURE 5. ICI does not prevent recruitment of GR to the IL8 promoter. repression of proinﬂammatory genes, we performed ChIP to de- a U2OS-GR stable cells were infected with 50 MOI Ad-ER for 24 h. The termine the effects of ICI on ERa and GR recruitment to the IL8 cells were then treated with 2.5 ng/ml TNF-a for 2 h in the absence or a m promoter. U2OS-GR cells were infected with Ad-ER for 24 h. presence of 10 nM Dex and 1 M ICI as indicated. ChIP assays were a performed with the ERa and GR Abs. The data shown are derived from The cells were then treated with TNF- for 2 h in the absence or real-time qPCR analysis of the IL8 gene, as detailed in Materials and presence of Dex with or without ICI. ChIP was performed using Methods. The fold change was determined using the raw values from the ERa and GR Abs. Dex, ICI, or TNF-a alone did not recruit ER or untreated control. Each data point is the average obtained from three in- GR to the IL8 promoter (Fig. 5). Treatment with TNF-a and Dex dependent experiments. Error bars represent the mean 6 SEM. caused a ∼5-fold increase in GR occupancy. The addition of ICI to TNF-a and Dex induced the recruitment of ERa to the promoter, repression of the CCL20 gene by E2 was inhibited by ICI (Fig. but it did not alter the recruitment of GR (Fig. 5). These results Downloaded from 4A). E2 or ICI did not have any effect on the Dex-mediated re- indicate that ICI does not block Dex repression of genes by in- pression of the TNFAIP2 gene in the presence (Fig. 4C) or ab- terfering with GR binding to the IL8 promoter. http://www.jimmunol.org/ by guest on September 28, 2021

FIGURE 6. ICI blocks Dex repression of the TNF-RE and recruitment of NCOA2 by GR. A, U2OS–ERa–TNF-RE cells were treated with Dox (100 ng/ ml) for 18 h, then TNF-a (5 ng/ml) was added for an additional 18 h. Luciferase activity was measured in triplicate samples. B,E2 causes a dose-dependent repression of the TNF-RE in U2OS–ERa–TNF-RE cells. U2OS–ERa–TNF-RE cells grown on 12-well plates were treated with Dox (100 ng/ml) for 18 h to induce ERa expression. The cells were then treated with TNF-a (5 ng/ml) along with increasing doses of E2 for 18 h and then assayed for luciferase activity. C, U2OS–ERa–TNF-RE cells express endogenous GR by Western blot. A rabbit anti-GR IgG was used to detect a band ∼97 kDa, which represents endogenous GR. D, Dex causes a dose-dependent repression of the TNF-RE. U2OS–ERa–TNF-RE cells were treated with Dox (100 ng/ml) for 18 h to induce ERa expression. The cells were then treated with TNF-a (5 ng/ml) along with increasing doses of Dex for 18 h and then assayed for luciferase activity. E, ICI blocks the Dex repression of the TNF-RE. U2OS–ERa–TNF-RE cells were treated with increasing amount of Dox for 18 h, and then TNF-a (5 ng/ml), Dex (10 nM), and ICI (10 nM) were added to the cells for an additional 18 h. F and G, ICI treatment prevents GR recruitment of NCOA2 to the TNF-RE. U2OS–ERa–TNF-RE were treated with Dox (100 ng/ml) for 18 h and then with TNF-a (5 ng/ml) for 2 h followed by Dex (10 nM) and ICI (10 nM) for an additional 3 h. ChIP assays were performed as described in Materials and Methods using anti-GR (F) or anti-NCOA2 (G) Ab. The data are derived from real-time qPCR. The fold changes were determined by comparison of threshold cycle values obtained from samples without Ab (control) and with speciﬁc Ab. All data are expressed as mean 6 SEM obtained from triplicate samples and are representative of several experiments. The Journal of Immunology 4359

ICI blocks GR recruitment of NCOA2 to the TNF-RE repressed by Dex, as genes activated by Dex were not affected by To further explore the mechanism whereby ICI antagonizes GR- ICI. a mediated repression, we studied the effect of ICI on the TNF- There are several possible mechanisms whereby the ICI–ER RE in the TNF-a gene, which is one of the ER/GR-repressed complex blocks the repression of genes in response to Dex. It is genes (Table I). We previously showed that the TNF-RE is re- known that GR causes repression of proinflammatory genes by k sponsible for TNF-a activation of the TNF-a gene (21), and it binding to composite elements, such as AP-1 and NF- B elements (7, 24). Once tethered to these transcriptional factors, GR recruits mediates the repression by E2 (14). The TNF-RE consists of a composite element that has AP-1 and NF-kB sites (21), which NCOA2, which mediates the repression of inflammatory genes are known to be involved in GR-mediated repression of in- (17). Interestingly, we found that estrogens repress proinflam- flammatory genes (1, 7, 8, 24). We prepared a U2OS cell line matory genes through a similar mechanism involving NCOA2 (14, (U2OS–ERa–TNF-RE) by stably transfecting the TNF-RE up- 16). Based on these observations, the simplest model is that the a stream of the minimal thymidine kinase promoter and luciferase ICI–ER complex might bind to the same element as the Dex–GR cDNA into the Dox-inducible U2OS-ERa cells. Stable cell lines complex on the IL8 promoter and therefore compete for and were screened for the activation of the TNF-RE by TNF-a, the prevent the binding of GR. A competition model is supported by expression of endogenous GR, and the repression by E and Dex. the observation that ICI was more effective at blocking the effects 2 a TNF-a produced a large activation of the TNF-RE (Fig. 6A) that of Dex at greater levels of ER in the cells. However, the observation that ICI did not inhibit the recruitment of GR to the IL8 was repressed by E2 in a dose-dependent manner (Fig. 6B). The U2OS–ERa–TNF-RE cells expressed endogenous GR by Western gene or the TNF-RE does not support a competitive model. An- a Downloaded from blot analysis (Fig. 6C). The GR was functional in these cells as other possibility is that the ICI–ER complex directly interacts shown by a dose-dependent repression of the TNF-RE by Dex with GR to form a heterodimer on particular promoters and then (Fig. 6D). These results demonstrate that these cells can be used blocks its repression function at these genes. However, there is no the study the mechanism whereby ICI antagonizes Dex repression. evidence that ER and GR can form a heterodimer at promoters. a The ICI antagonism of the Dex repression of the TNF-RE was Finally, Dex–GR and ICI–ER complexes could bind to different a dependent on the expression level of ERa because there was sites on the promoter, allowing the ICI–ER complex to interfere a dose-dependent effect by Dox (Fig. 6E). ICI had no antagonistic with the recruitment of transcriptional factors or cofactors to http://www.jimmunol.org/ effect in the absence of Dox, but completely blocked the Dex Dex–GR that are essential for repression, such as NCOA2. Con- repression of the TNF-RE when the cells were treated with 100 sistent with this model is our finding that ICI blocked the re- ng/ml Dox to induce the production of ERa. cruitment of NCOA2 by GR at the TNF-RE, which indicates that We previously demonstrated that ER and GR recruit NCOA2 this is an important mechanism for the antagonistic effect of ICI to cytokine genes and that NCOA2 mediates repression (16, 17). on GR repression. These observations suggest that ICI might antagonize GR effects It has been demonstrated that ICI reverses the anti-inflammatory at the TNF-RE by blocking the recruitment of NCOA2. The activity produced by Dex in a rat model of carrageenan-induced U2OS–ERa–TNF-RE cells were treated with Dex and ICI after pleurisy (19). The mechanism whereby ICI inhibited the anti- by guest on September 28, 2021 ERa was induced with Dox, and then ChIP was performed using inflammatory effects of Dex in this model is unclear. Our find- Abs to GR or NCOA2. Dex induced recruitment of GR to the ings suggest that ICI might interfere with the Dex-induced re- TNF-RE by ∼6-fold (Fig. 6F). Consistent with data from the IL8 pression of proinflammatory genes by blocking the recruitment of promoter (Fig. 5), ICI did not block recruitment of GR to the NCOA2. It is clear that the anti-inflammatory effects of gluco- TNF-RE (Fig. 6F). In contrast, ICI blocked the Dex-induced re- corticoids in certain tissues, particularly in the skin and lungs, cruitment of NCOA2 to the TNF-RE (Fig. 6G). These results in- have tremendous clinical value. However, prolonged reduction in dicate that the ICI–ERa complex antagonizes GR-mediated re- cytokines by glucocorticoids could lead to adverse effects in other pression by interfering with the recruitment of NCOA2 by GR. tissues because of their beneficial role in the growth, differentia- tion, and viability of cells. By understanding the mechanism of ER and GR cross talk, it might be possible to develop unique estro- Discussion gens that could be used in conjunction with glucocorticoids to Glucocorticoids are extensively used therapeutically for numerous regulate proinflammatory genes. A combination of drugs that target inflammatory and autoimmune conditions because of their potent both GR and ER simultaneously in some tissues might reduce the anti-inflammatory properties. Although the effects of estrogens on adverse effects of glucocorticoids alone. the reproductive system are well recognized, estrogens also can produce anti-inflammatory actions (10). We previously found that estrogens repress numerous proinflammatory genes in U2OS cells Acknowledgments (16, 23). However, it is unclear if these proinflammatory genes are We thank Michael Garabedian for providing the U2OS-GR cells. also repressed by glucocorticoids and if there is cross talk between ER and GR at these genes. To explore these possibilities, we Disclosures examined the effects of Dex and E2 in U2OS cells that express The authors have no financial conflicts of interest. only ERa, only GR, or both ERa and GR. We found that Dex and/or E2 repressed three classes of proinflammatory genes. Most of the proinflammatory genes were repressed by both ERa and References GR. However, five genes were repressed by only ERa or GR. As 1. Beck, I. M., W. Vanden Berghe, L. Vermeulen, K. R. Yamamoto, G. Haegeman, and K. De Bosscher. 2009. Crosstalk in inflammation: the interplay of gluco- expected, the ER antagonist ICI blocked E2-mediated repression. corticoid receptor-based mechanisms and kinases and phosphatases. Endocr. Rev. In contrast, the GR antagonist RU486 did not block Dex-mediated 30: 830–882. repression. Instead, we unexpectedly found that ICI blocked Dex 2. Ghosh, S., M. J. May, and E. B. Kopp. 1998. NF-kappa B and Rel proteins: repression of ER/GR-repressed genes. ICI did not have any effect evolutionarily conserved mediators of immune responses. Annu. Rev. Immunol. 16: 225–260. on genes that were repressed with only GR. The effect of ICI was 3. Smale, S. T. 2010. Selective transcription in response to an inflammatory dependent on the expression of ERa and was limited to genes stimulus. Cell 140: 833–844. 4360 GR AND ER TRANSREPRESSION AND CROSS TALK

4. Ray, A., and K. E. Prefontaine. 1994. Physical association and functional an- 15. An, J., C. Tzagarakis-Foster, T. C. Scharschmidt, N. Lomri, and D. C. Leitman. tagonism between the p65 subunit of transcription factor NF-kappa B and the 2001. Estrogen receptor beta-selective transcriptional activity and recruitment of glucocorticoid receptor. Proc. Natl. Acad. Sci. USA 91: 752–756. coregulators by phytoestrogens. J. Biol. Chem. 276: 17808–17814. 5. Webster, J. C., and J. A. Cidlowski. 1999. Mechanisms of Glucocorticoid-receptor- 16. Cvoro, A., C. Tzagarakis-Foster, D. Tatomer, S. Paruthiyil, M. S. Fox, and mediated Repression of Gene Expression. Trends Endocrinol. Metab. 10: 396–402. D. C. Leitman. 2006. Distinct roles of unliganded and liganded estrogen 6. De Bosscher, K., and G. Haegeman. 2009. Minireview: latest perspectives on receptors in transcriptional repression. Mol. Cell 21: 555–564. antiinflammatory actions of glucocorticoids. Mol. Endocrinol. 23: 281–291. 17. Rogatsky, I., K. A. Zarember, and K. R. Yamamoto. 2001. Factor recruitment 7. Nissen, R. M., and K. R. Yamamoto. 2000. The glucocorticoid receptor inhibits and TIF2/GRIP1 corepressor activity at a collagenase-3 response element that mediates regulation by phorbol esters and hormones. EMBO J. 20: 6071–6083. NFkappaB by interfering with serine-2 phosphorylation of the RNA polymerase 18. Nettles, K. W., G. Gil, J. Nowak, R. Me´tivier, V. B. Sharma, and G. L. Greene. II carboxy-terminal domain. Genes Dev. 14: 2314–2329. 2008. CBP Is a dosage-dependent regulator of nuclear factor-kappaB suppres- 8. Luecke, H. F., and K. R. Yamamoto. 2005. The glucocorticoid receptor blocks sion by the estrogen receptor. Mol. Endocrinol. 22: 263–272. P-TEFb recruitment by NFkappaB to effect promoter-specific transcriptional 19. Cuzzocrea, S., S. Bruscoli, C. Crisafulli, E. Mazzon, M. Agostini, C. Muia`, repression. Genes Dev. 19: 1116–1127. E. Esposito, R. Di Virgilio, R. Meli, E. Vegeto, et al. 2007. Estrogen receptor 9. Price, D. H. 2000. P-TEFb, a cyclin-dependent kinase controlling elongation by antagonist fulvestrant (ICI 182,780) inhibits the anti-inflammatory effect of RNA polymerase II. Mol. Cell. Biol. 20: 2629–2634. glucocorticoids. Mol. Pharmacol. 71: 132–144. 10. Straub, R. H. 2007. The complex role of estrogens in inflammation. Endocr. Rev. 20. Kian Tee, M., I. Rogatsky, C. Tzagarakis-Foster, A. Cvoro, J. An, R. J. Christy, 28: 521–574. K. R. Yamamoto, and D. C. Leitman. 2004. Estradiol and selective estrogen 11. Heldring, N., A. Pike, S. Andersson, J. Matthews, G. Cheng, J. Hartman, receptor modulators differentially regulate target genes with estrogen receptors M. Tujague, A. Stro¨m, E. Treuter, M. Warner, and J. A. Gustafsson. 2007. Estrogen alpha and beta. Mol. Biol. Cell 15: 1262–1272. receptors: how do they signal and what are their targets. Physiol. Rev. 87: 905–931. 21. Leitman, D. C., R. C. Ribeiro, E. R. Mackow, J. D. Baxter, and B. L. West. 1991. 12. Stein, B., and M. X. Yang. 1995. Repression of the interleukin-6 promoter by Identification of a tumor necrosis factor-responsive element in the tumor necrosis estrogen receptor is mediated by NF-kappa B and C/EBP beta. Mol. Cell. Biol. factor alpha gene. J. Biol. Chem. 266: 9343–9346. 15: 4971–4979. 22. Levy, N., D. Tatomer, C. B. Herber, X. Zhao, H. Tang, T. Sargeant, L. J. Ball, J. Summers, T. P. Speed, and D. C. Leitman. 2008. Differential regulation of

13. Ray, P., S. K. Ghosh, D. H. Zhang, and A. Ray. 1997. Repression of interleukin-6 Downloaded from native estrogen receptor-regulatory elements by estradiol, tamoxifen, and gene expression by 17 beta-estradiol: inhibition of the DNA-binding activity of raloxifene. Mol. Endocrinol. 22: 287–303. the transcription factors NF-IL6 and NF-kappa B by the estrogen receptor. FEBS 23. Cvoro, A., D. Tatomer, M. K. Tee, T. Zogovic, H. A. Harris, and D. C. Leitman. Lett. 409: 79–85. 2008. Selective estrogen receptor-beta agonists repress transcription of proin- 14. An, J., R. C. Ribeiro, P. Webb, J. A. Gustafsson, P. J. Kushner, J. D. Baxter, and ﬂammatory genes. J. Immunol. 180: 630–636. D. C. Leitman. 1999. Estradiol repression of tumor necrosis factor-alpha tran- 24. Diamond, M. I., J. N. Miner, S. K. Yoshinaga, and K. R. Yamamoto. 1990. scription requires estrogen receptor activation function-2 and is enhanced by Transcription factor interactions: selectors of positive or negative regulation coactivators. Proc. Natl. Acad. Sci. USA 96: 15161–15166. from a single DNA element. Science 249: 1266–1272. http://www.jimmunol.org/ by guest on September 28, 2021