Brahma-Related Gene 1-Dependent STAT3 Recruitment at IL-6-Inducible Genes¹

Chromatin remodeling is fundamental to gene regulation in eukaryotes and is mediated by two categories of enzymes that either use the energy of ATP hydrolysis to alter nucleosome structure or position, or that covalently modify histones (1). Yeast mating type switching and sucrose nonfermenting (SWI/SNF)³ is one complex that uses ATP hydrolysis to regulate the position and structure of nucleosomes, which can lead to either gene activation or repression (2). There are more than 10 components of SWI/SNF (3) which can interact with a variety of activators or repressors (2).

Brahma-related gene 1 (BRG1) is the ATPase subunit of SWI/SNF and has been linked to differentiation, proliferation, and tumorigenesis. Homozygous inactivation of BRG1 causes early embryonic lethality in mice and heterozygotes are prone to a variety of tumors (4, 5, 6). Previously, we showed that BRG1 is required for induction of a subset of IFN-γ-responsive genes (7) and subsequently others reported that it is also essential for induction of a subset of IFN-α-responsive targets (8, 9). Moreover, the BAF47 and BAF200 components of SWI/SNF are also required for IFN signaling, although the INI1 and BAF180 proteins are dispensable (10, 11, 12). These findings highlight an important role for SWI/SNF in the immune response.

In view of the findings that BRG1 binds STAT2 and STAT3 (8, 13) and that SWI/SNF typically functions downstream of primary activators in gene induction cascades (14, 15, 16, 17, 18, 19, 20, 21), it seemed logical that BRG1 might act after STAT1 recruitment at IFN targets. Instead, BRG1 is essential for STAT1 recruitment at multiple IFN-α and IFN-γ targets (22). In contrast, constitutive BRG1 recruitment is STAT1 independent (22). All of the subsequent IFN-induced events, including chromatin remodeling, histone modifications, and transcription are blocked if BRG1 is absent (10, 22). Thus, BRG1 has an apical role in IFN signaling.

IL-6 is an important immunoregulatory cytokine with multiple functions in hemopoietic proliferation, differentiation of B cells, and tumorigenesis. For example, it is essential for the growth and survival of B lymphocyte-derived murine plasmacytomas and human myelomas, and, most likely, IL-6 promotes such tumors by increasing proliferation and inhibiting apoptosis (23, 24). IL-6 triggers tyrosine phosphorylation and activation of JAKs. Activated JAKs phosphorylate STAT3 which forms either a homodimer or a STAT3/1 heterodimer, which translocates to the nucleus and bind to the STAT-binding site at target loci (25).

A direct link between IL-6 signaling and BRG1 has never been established. However, BRG1 associates with STAT3 at the p21^waf1 promoter (13), raising the possibility that, in addition to IFN targets, BRG1 may have a broader role in cytokine signaling. In this study, we provide in vivo evidence that BRG1 is required for IL-6-induced expression at a subset of target genes. As seen at IFN targets (22), BRG1 constitutively binds to IL-6-responsive promoters and is essential for all events, including IL-6-induced STAT3 recruitment, chromatin remodeling, and covalent histone modifications. Intriguingly, BRG1 has promoter-specific effects on basal chromatin remodeling and, at the IFN regulatory factor 1 (IRF1) locus, is required for STAT3 but not STAT1 recruitment. These results reveal a broad role for BRG1 in mediating the induction of STAT target genes and suggest locus and transcription factor-specific differences in its mechanism of action.

Human adenocarcinoma cell line SW13 and cervical carcinoma cell lines, HeLa-Ini-11 (HeLa) and C33A cells, were grown in α-MEM/10% FBS as described (7). Cells were treated with 50 ng/ml IL-6 (PHC0064; BioSource International).

Adenoviral vectors were based on pAdlox (26). Construction details and complete sequences are available on request. Vectors were used to generate adenovirus as described (26). Each virus was plaque purified to remove contaminating normal adenovirus. Virus was amplified in the 293-derived Cre8 cell line (26).

SW13 cells was lysed in radioimmunoprecipitation assay buffer (50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1 mM EDTA, 1% Nonidet P-40, 0.25% sodium deoxycholate, 1 mM Na₃VO₄, 1 mM NaF, 1 mM PMSF, 1 μg/ml aprotinin, 1 μg/ml leupeptin, and 1 μg/ml pepstatin). Fifty micrograms of cell lysate was separated by 10% SDS-PAGE gel and transferred to a nitrocellulose membrane. Proteins were incubated with Abs against STAT3 (sc-483; Santa Cruz Biotechnology), p-STAT3 (Tyr⁷⁰⁵-R; sc-7993-R; Santa Cruz Biotechnology), STAT1 (06-501; Upstate Biotechnology), BRG1 (sc-10768; Santa Cruz Biotechnology), or β-actin (A4511; Sigma-Aldrich), and detected by ECL with HRP-conjugated goat anti-mouse or goat anti-rabbit Ig G. Protein intensity was quantified by MultiImaginer (Bio-Rad) and normalized to β-actin.

RNA was isolated from cells in confluent 60-mm plates using TRIzol reagent (Invitrogen Life Technologies). A 2.5-μg aliquot of RNA was diluted in 20 μl of diethylpyrocarbonate-treated water, heated to 90°C for 5 min, then combined with 30 μl of first strand master mix (10 μg of Pd(N)₆ salt, 1× First Strand buffer (Invitrogen Life Technologies), 1 mM dNTPs, 10 mM DTT, 50 U of Superscript II reverse transcriptase (Invitrogen Life Technologies)), and incubated at 37°C for 1 h, then for 10 min at 95°C. qPCR was performed by using an Applied Biosystems PRISM 7900HT in duplicate with SYBR Green PCR master mix (Applied Biosystems) according to the manufacturer’s instructions. PCR consisted of 40 cycles of 95°C for 15 s and 55°C for 30 s. A final cycle (95°C, 15 s, 60°C) generated a dissociation curve to confirm a single product. The cycle quantity required to reach a threshold in the linear range (Qt) was determined and compared with a standard curve for each primer set generated by five 3-fold dilutions of genomic DNA samples of known concentration. Expression levels were examined using primers in the last exon of each gene or exon 6, 9, or 20 in the case of myeloid cell nuclear differentiation Ag (MNDA), IFN-γ-inducible protein 16 (IFI16), or CIITA, respectively (Table I), and normalized to those for β-actin.

HeLa or SW13 cells were cross-linked with 1% formaldehyde at room temperature for 10 min, washed twice with ice-cold PBS, collected in 1 ml of PBS and centrifuged for 5 min at 5000 rpm. Cells were resuspended in 1 ml of lysis buffer (1% SDS, 10 mM EDTA, and 50 mM Tris-HCl (pH 8)) plus proteinase inhibitors (aprotinin, leupeptin, and pepstatin), incubated on ice for 10 min and sonicated to an average size of 500 bp (Vibra Cell; Sonics and Materials). Chromatin was precleared with 25 μl of Staph A (507862; Calbiochem) at 4°C for 15 min. A 100-μl aliquot of sonicated chromatin was immunoprecipitated (IP) with 2 μg of Abs for STAT1 (06-501; Upstate Biotechnology), STAT3 (sc-7179; Santa Cruz Biotechnology), BRG1 (sc-10768; Santa Cruz Biotechnology), or acetylated histone H4 (06-866; Upstate Biotechnology), at 4°C overnight. IP samples were centrifuged at 13200 rpm and supernatant was incubated with 10 μl of Staph A at room temperature for 15 min. Precipitates were washed sequentially for 3 min in 1× dialysis buffer (2 mM EDTA, 50 mM Tris-HCl (pH 8), and 0.2% sarkosyl) twice, and IP wash buffer (1% Nonidet P-40, 100 mM Tris-HCl (pH 9), 500 mM LiCl 1% and deoxycholic acid) four times. Samples were extracted twice with 150 μl of elution buffer (1% SDS and 50 mM NaHCO₃), heated at 65°C overnight to reverse cross-links, and DNA fragments were purified with a QIAEX II Gel Extraction kit (catalog no. 20051). A 4-μl aliquot from a total of 50 μl was used in the qPCR using primers listed in Table I. Qt values were compared with a standard curve, the copy number was calculated, the amount of DNA was precipitated by an irrelevant GAL4 (06-262; Upstate Biotechnology) Ab subtracted, and the percent ChIP DNA relative to input chromatin was calculated.

SW13 cells transduced with Ad-FG or Ad FG-BRG1 were left untreated or exposed to IL-6. Cells were trypsinized and washed twice in PBS, and once in 1× reticulocyte standard buffer (RSB; 10 mM Tris-HCl (pH 7.5), 10 mM NaCl, 3 mM MgCl₂, and 1 mm PMSF). The cell pellet was resuspended in 10 ml of 50% glycerol, 0.5% Nonidet P-40, 1 mM PMSF, and 1× RSB solution, lysed by pipetting 10 times, and incubated on ice for 5–10 min. Nuclei were spun at 4000 rpm for 10 min at 4°C, washed in 10 ml of 1× RSB plus 0.1 mM PMSF, spun at 2500 rpm for 10 min at 4°C, then resuspended in 0.3 ml of 1× RSB. DNase I stock (4 mg/ml; D5793; Sigma-Aldrich) was diluted in 1× RSB to a final concentration of 80 ng/ml. A 1/1000 volume of 1 M CaCl₂ was added and 100 μl of nuclei was digested with 2.5 Krunitz U/100 μl of DNase I for 3 min (IRF1) or 10 min (IFI16) at 37°C. A total of 100 μl of 2× STOP solution (0.6 M NaCl, 20 mM Tris-HCl (pH 8.0), 10 mM EDTA and 1% SDS) was added. Next, samples were incubated with10 μl of proteinase K (25 mg/ml) at 55°C overnight. A total of 200 μl of 1× STOP solution was added, and samples were phenol/chloroform extracted and treated with RNase (Molecular Biology Grade; Sigma-Aldrich) at 37°C overnight. Samples were phenol/chloroform extracted again, ethanol precipitated, and resuspended in 100 μl of 10 mM Tris-HCl (pH 8.0) and extracted DNA was subjected to qPCR using primers listed in Table I.

Double-stranded siRNA oligonucleotide against STAT3 (top: 5′-CAUCUGCCUAGAUCGGCUAdTdT-3′) (27) and scrambled siRNA (top: 5′-ACUCUGCGCGUUGUACAACdTdT-3′) were obtained from Dharmacon Research. SW13 cells in 12-well plates were transfected with 300 nM per siRNA for 3 days using DharmaFECT-1 (Dharmacon Research) according to the manufacturer’s instructions. One day before harvesting, cells were transduced with Ad-FG or Ad FG-BRG1 as described above.

A paired Student’s t test was used for statistical analysis.

To study the role of BRG1 in IL-6-mediated gene induction, we used SW13 and C33A cells, which do not express either BRG1 or the closely related BRM protein (3, 28, 29). These cells were transduced with an adenovirus vector (Ad FG-BRG1) expressing BRG1 fused to an N-terminal Flag-GFP tag, or a control virus (Ad FG), expressing Flag-tagged GFP. The infection was optimized in SW13 cells under conditions where >90% cells were transduced after 24 h and the level of BRG1 expression was similar to the amount of BRG1 seen in HeLa cells (22). Transduced cells were either left untreated or exposed to IL-6. qPCR was used to measure the expression levels of several potential IL-6-inducible genes using primers at the 3′ end of each transcript (Table I). These genes were selected because they are known IL-6 and/or STAT3 targets or, based on the considerable overlap between cytokine-responsive genes, were known IFN targets (30, 31, 32, 33, 34, 35, 36).

First, we studied the know IL-6-responsive gene IFI16 and the potential novel target IFN-induced transmembrane protein 3 (IFITM3) (Fig. 1,A). BRG1 increased the basal levels of both genes, especially of IFITM3 (p < 0.01; Fig. 1,A). Induction of these genes was only seen in the presence of BRG1, and was more rapid for IFI16 relative to IFITM3, which were increased at 1 or 6 h, respectively (p < 0.01; Fig. 1 A). Therefore, 6 h of IL-6 treatment was used to analyze additional targets.

Of a total of 17 genes tested in SW13 cells, 4 (IFI16, IRF1, IFITM3, and ISGF3G) were IL-6 responsive, all of which exhibited BRG1 dependency except ISGF3G (Table II and Fig. 1,B). Unlike IFITM3 and IFI16, BRG1 did not increase basal expression of IRF1 (Fig. 1,B). Like IFITM3, ISGF3G is a new IL-6-responsive gene. BRG1 did not affect the levels or activation (i.e., phosphorylation) of STAT3 (Fig. 1,C), indicating that it facilitates IL-6 activity by another mechanism (see below). The BRG1 dependence of IL-6-mediated IRF1 induction is noteworthy given that IFN-γ induction of this gene is BRG1 independent (7, 22). Seven genes (p21^Waf1, cyclin D1, UCK2, JUNB, Bcl-x_L, Pim 1, and Pim 2) were constitutively expressed in SW13 cells but not further induced by IL-6, and the remaining 6 genes, such as MNDA and AIM2, were expressed at negligible levels (Table II and Fig. 1,B). None of 10 genes tested in C33A cells were induced by IL-6 (Table II). Of these, 7 (IRF1, p21^Waf1, cyclin D1, JunB, UCK2, Pim 1, and Pim 2) were constitutively expressed and 3 (CIITA, IFI27, and IFIT1) were silent (Table II).

In summary, our data reveal an important role for BRG1 in mediating induction at a subset of IL-6 targets, highlight a difference in the requirement for BRG1 in responding to IFN-γ or IL-6 at the IRF1 locus, and reveal that many genes induced by IL-6 in other cell types are unaffected in SW13 and/or C33A cells.

To confirm that STAT3 mediates IL-6 activity, and to address whether it is required for the BRG1-dependent increase in basal expression, RNA interference was used to reduce STAT3 protein levels. SW13 cells were left untransfected with STAT3 or scrambled siRNAs then transduced with BRG1 or GFP virus and left untreated or exposed to IL-6. STAT3 protein levels were significantly reduced by STAT3 but not scrambled siRNA (Fig. 2, A and B). STAT3 siRNA had no effect on the levels of STAT1 or β-actin (Fig. 2,A). As expected, STAT3 knockdown significantly reduced IL-6-induced IFITM3 and IFI16 mRNA (Fig. 2,C). In contrast, STAT3 knockdown had no effect on BRG1-enhanced basal expression of either IFITM3 or IFI16 (Fig. 2 C).

Next, we asked whether STAT3 recruitment is BRG1 dependent at IL-6-inducible genes. IFN-γ stimulates STAT1 binding to the promoters of several of the genes studied here (22), so we tested the same regions for IL-6-induced STAT3 binding. First, we used ChIP to assess the kinetics of STAT3 recruitment at two IL-6 target promoters in the presence or absence of BRG1. SW13 cells were transduced with Ad FG-BRG1 or Ad FG and left untreated or exposed to IL-6 for 0.5, 1, 3, 6, and 24 h. Sheared cross-linked chromatin was IP with STAT3 Abs and qPCR used to assess enrichment of specific target sequences.

In BRG1 but not GFP-expressing cells, STAT3 was detected at the IFITM3 and IFI16 promoters within 30 min of IL-6 treatment (Fig. 3,A). Binding diminished after 1 h, but peaked a second time at 6 h which fell again by 24 h (Fig. 3,A). This cyclical recruitment of STAT3 is reminiscent of the activated estrogen receptor which also binds in cycles coincident with transcriptional initiation (37). We also found that BRG1 was required for STAT3 recruitment at the IRF1 promoter (p < 0.05; Fig. 3,B). In contrast, STAT3 recruitment to the ISGF3G promoter was BRG1 independent (Fig. 3,B), matching the expression analysis (Fig. 1,B). The latter result proves that the inability of STAT3 to bind other targets is not due to aberrant STAT3 activity in BRG1-deficient cells. STAT3 binding at the silent MNDA promoter was negligible (Fig. 3,B). Activated STAT3 can form heterodimers with STAT1 (25), but ChIP results did not show significant STAT1 recruitment at any of the tested promoters (IFITM3, IFI16, IRF1, ISGF3G, or MNDA) (Fig. 3 B). In summary, these data indicate that BRG1-dependent induction of a subset of IL-6 targets correlates perfectly with BRG1-dependent recruitment of STAT3 to their promoters.

Histone lysine acetylation is linked to gene activation and previous studies indicated that STAT3 and BRG1 bind to at least one target (p21^waf2) in HepG2 cells, resulting in histone acetylation (13). We and others demonstrated that BRG1 is required for histone acetylation at IFN target promoters (10, 22). Therefore, we asked whether histone acetylation at IL-6 target promoters is also BRG1 dependent. An Ab that binds to tetra-acetylated histone H4 was used in ChIP assays with chromatin from SW13 cells infected with Ad FG-BRG1 or Ad FG, in the presence or absence of IL-6. Acetylation was induced at the IFI16 and IFITM3 promoters in the presence of BRG1 and, importantly, was blocked in its absence (Fig. 4). Thus, as seen at IFN targets, IL-6-inducible histone acetylation is downstream of and dependent on BRG1. As expected, IL-6-induced histone acetylation at the BRG1-independent ISGF3G promoter was unaffected by the absence of BRG1 and negligible levels of acetylation were seen at the silent and unresponsive MNDA promoter (Fig. 4).

BRG1 is associated with several IFN-responsive promoters even in the absence of cytokine (10, 22), so we tested whether this is also the case at IL-6 targets. These assays were performed with chromatin from HeLa cells, which express BRG1. Cells were left untreated or exposed to IL-6 for 6 h and ChIP assays performed with anti-BRG1 Ab. These assays showed that BRG1 was constitutively bound to the IRF1 and IFI16 promoters at similar levels before or after IL-6 stimulation (Fig. 5). BRG1 was also present at the IFITM1 promoter (Fig. 5), as reported previously (10). Negligible binding was observed at the irrelevant PITX2 promoter (Fig. 5). Thus, BRG1 binds constitutively to IL-6-responsive promoters.

BRG1 does not affect the basal level of DNase I and/or restriction enzyme accessibility at the IFN target genes CIITA, GBP1, or IFI27 but is essential for IFN-induced remodeling at these loci (7, 22). However, BRG1 does increase basal accessibility at the IFN-responsive IFITM1 (also called 9-27 or IFI17) locus (10). We used DNase I digestion coupled with qPCR to quantify accessibility at IFI16 and IRF1 promoters in SW13 cells in the presence or absence of BRG1 and IL-6. SW13 cells were transduced with Ad FG-BRG1 or Ad FG, left untreated or exposed to IL-6, nuclei incubated with DNase I, and levels of intact DNA at the IFI16 and IRF1 promoters were determined by qPCR. In BRG1-deficient cells, IFI16 promoter accessibility was identical plus or minus IL-6 (Fig. 6). BRG1 did not affect basal accessibility to DNase I at this promoter, but increased accessibility in the presence of IL-6 (p < 0.05; Fig. 6). At the IRF1 promoter, BRG1 increased basal accessibility (p < 0.05) and further enhanced IL-6-induced accessibility (p < 0.05; Fig. 6). In contrast, DNase I accessibility at the last exon of IRF1 was identical in untreated or IL-6-treated cells transduced with Ad FG-BRG1 or Ad FG (Fig. 6). These data suggest that, as seen at IFN target genes (10, 22), BRG1 increases basal promoter accessibility at some and IL-6-induced promoter accessibility at all tested BRG1-dependent targets.

Previously, we demonstrated that recruitment of the STAT1 homodimer to multiple IFN-γ-inducible promoters requires BRG1 (22). We also showed that recruitment of the STAT1/STAT2/IRF9 (ISGF3) trimer at various IFNα-inducible promoters is BRG1 dependent (22). In this study, we extend these observations, revealing that recruitment of a third STAT family member, STAT3, to several IL-6-responsive targets is also BRG1 dependent. Taken together, these studies reveal a broad role for BRG1 in mediating the access of STAT proteins to promoters. This finding underscores the importance of BRG1 for immune function.

In addition to STAT binding, all downstream events at IFN targets are blocked in the absence of BRG1, including chromatin remodeling, histone modifications, and recruitment of other transcription factors (7, 10, 22). Likewise, we found that both histone acetylation and chromatin remodeling require BRG1 at IL-6 target genes. Thus, BRG1 has an apical role at multiple STAT target genes and is necessary to set up the entire cascade of events that lead to transcription. This scenario is distinct from many other genes, where BRG1 acts downstream of activators that initiate promoter assembly (14, 15, 16, 17, 18, 19, 20, 21). This primary role for BRG1 at cytokine-responsive targets is possible through constitutive recruitment to IFN- and IL-6-inducible promoters. Constitutive binding of BRG1 to target genes is STAT independent (this study and Ref. 22). Numerous other mechanisms may recruit BRG1 to cytokine target genes, including association with transcription factors, various chromatin regulatory complexes, and modified histones; for a detailed discussion, see Refs. 2 and 22).

Our initial study linking BRG1 to IFN-responsive genes focused on CIITA. At this locus, we found that BRG1 did not affect basal DNase I or restriction enzyme accessibility at the promoter, but was absolutely essential for IFN-γ-induced chromatin remodeling (7). Recently, we repeated and extended this finding by showing that BRG1 does not alter basal chromatin accessibility at the CIITA, GBP1, or IFI27 promoters, but is critical for IFN-γ- induced remodeling at these loci (22). In contrast, others reported that BRG1 increases both basal and IFN-α-induced accessibility at the IFITM1 promoter (10). In this study, we showed that BRG1 increases basal DNase I accessibility at the IRF1 but not the IFI16 promoter and enhances IL-6-induced chromatin remodeling at both promoters. Thus, there appear to be two distinct classes of BRG1-dependent promoters: chromatin structure at uninduced type A promoters (e.g., CIITA, GBP1, IFI27, IFI16) is apparently unaffected by BRG1, whereas basal accessibility is increased at type B promoters (e.g., IFITM1, IRF1). BRG1 enhances cytokine-induced chromatin accessibility at both type A and B promoters.

How does BRG1 facilitate access of STAT complexes to promoters where it does not seem to affect basal chromatin structure? Because BRG1 interacts with STAT proteins (8, 13) it might act as a platform to facilitate STAT1 recruitment. However, this cannot be its only role as BRG1 ATPase activity is essential for induction of genes like CIITA (7). Another possibility is that BRG1 acts primarily on distant elements that influence promoter accessibility following cytokine treatment. However, the role of distant elements at cytokine-regulated genes is unclear, but it will be important to investigate this possibility.

Previously, we showed that IFN-γ induces IRF1 independent of BRG1 (7). Indeed, STAT1 recruitment was unimpaired at this locus in BRG1-deficient cells (22). Surprisingly, however, we found that IL-6-induced STAT3 recruitment at the IRF1 promoter was greatly enhanced by BRG1. This unexpected result suggests that STAT1 dimers and STAT3 dimers do not engage this promoter in an identical fashion. Both STAT1 and STAT3 can bind TTN₅AA motifs, but STAT3 also binds TTN₄AA motifs, and the GAS (TTCCCCGAA) in the IRF1 promoter is a TTN₅AA motif that binds well to all STAT family members (38). This observation suggests that the specific requirement for BRG1 for STAT3 access to the IRF1 promoter is not linked to the sequence of the DNA-binding motif. The crystal structures of DNA-bound STAT1 or STAT3 are virtually superimposable (39, 40), but were created using truncated molecules that lack the extreme N- and C- terminal portions. These regions bind common but also distinct proteins. For example, the N-terminal region of STAT3 but not STAT1 binds and cooperates in gene induction with c-Jun (41). Thus, it is feasible that differences in the protein complexes associated with STAT3 or STAT1 could block and/or facilitate entry in the absence of BRG1, respectively.

In summary, there is a broad role for BRG1 in regulating the access of STAT proteins to DNA. Although many promoters are BRG1 dependent, others are BRG1 independent, and among the former there are differences in the effects of BRG1 on basal chromatin accessibility and the requirement for BRG1 to promote access of difference STAT family members. Given the crucial role STAT and BRG1 proteins play in immunity and cancer, it will be important to uncover the mechanisms that underlie these differences.

The authors have no financial conflict of interest.