Brd4 Is Essential for IL-1β-Induced Inflammation in Human Airway Epithelial Cells

Background Chronic inflammation and oxidative stress are key features of chronic obstructive pulmonary disease (COPD). Oxidative stress enhances COPD inflammation under the control of the pro-inflammatory redox-sensitive transcription factor nuclear factor-kappaB (NF-κB). Histone acetylation plays a critical role in chronic inflammation and bromodomain and extra terminal (BET) proteins act as “readers” of acetylated histones. Therefore, we examined the role of BET proteins in particular Brd2 and Brd4 and their inhibitors (JQ1 and PFI-1) in oxidative stress- enhanced inflammation in human bronchial epithelial cells. Methods Human primary epithelial (NHBE) cells and BEAS-2B cell lines were stimulated with IL-1β (inflammatory stimulus) in the presence or absence of H2O2 (oxidative stress) and the effect of pre-treatment with bromodomain inhibitors (JQ1 and PFI-1) was investigated. Pro-inflammatory mediators (CXCL8 and IL-6) were measured by ELISA and transcripts by RT-PCR. H3 and H4 acetylation and recruitment of p65 and Brd4 to the native IL-8 and IL-6 promoters was investigated using chromatin immunoprecipitation (ChIP). The impact of Brd2 and Brd4 siRNA knockdown on inflammatory mediators was also investigated. Result H2O2 enhanced IL1β-induced IL-6 and CXCL8 expression in NHBE and BEAS-2B cells whereas H2O2 alone did not have any affect. H3 acetylation at the IL-6 and IL-8 promoters was associated with recruitment of p65 and Brd4 proteins. Although p65 acetylation was increased this was not directly targeted by Brd4. The BET inhibitors JQ1 and PFI-1 significantly reduced IL-6 and CXCL8 expression whereas no effect was seen with the inactive enantiomer JQ1(-). Brd4, but not Brd2, knockdown markedly reduced IL-6 and CXCL8 release. JQ1 also inhibited p65 and Brd4 recruitment to the IL-6 and IL-8 promoters. Conclusion Oxidative stress enhanced IL1β-induced IL-6 and CXCL8 expression was significantly reduced by Brd4 inhibition. Brd4 plays an important role in the regulation of inflammatory genes and provides a potential novel anti-inflammatory target.


Introduction
Chronic inflammation is a core component of COPD and is associated with activation of the NF-kB signalling pathway particularly in patients with GOLD stage I-III disease [1,2]. Elevated expression of oxidants, either derived from activated immune and structural cells or from cigarette smoke, result in the high degree of oxidative stress which is found in the lungs of COPD patients [3][4][5]. Oxidative stress and inflammation are inseparably intertwined processes in these subjects. There is also a considerable evidence of oxidative stress entailed in the pathology of many other disorders, including aging, cancer, neurodegenerative and cardiovascular diseases [6,7]. Corticosteroids are frequently used in the management of inflammation in COPD patients; however, they proved to be less effective in COPD patients [8,9].
Abnormal histone acetylation (AcH) profiles have been linked to smoke exposure [10] and to relative corticosteroid unresponsiveness in COPD [11,12]. DNA is tightly packed together with histones into structural units called nucleosomes. Each nucleosome is an octamer of four core histone proteins; H2A, H2B, H3 and H4 proteins with ,146-base pair of DNA wrapped around and linked to H1 protein [13]. In transcriptionally active chromosomal regions, the chromatin unwinds allowing accessibility of transcription machinery. In contrast, the condensed heterochromatin is associated with gene suppression. This transition is achieved through reversible post-translational modifications (PMTs) such as acetylation, methylation and phosphorylation [14]. PTMs of histones play an important role in gene transcription and regulation and generally occur at histone tails [15]. Histone lysine (K) acetylation (AcK) signals the recruitment of basal transcriptional co-activators to the promoter regions of inflammatory and immunoregulatory genes [16,17]. Histone acetyltransferases (HATs) acts as writers and catalyse the addition of acetyl group to lysine residue in histone tails whereas histone deacetylases (HDACs) serve as erasers [18,19]. Acetylated histones are recognised by the bromodomain and extra-terminal (BET) proteins that are considered as readers of acetylated histones and associated with the regulation of several genes involved in cellular proliferation, cell cycle progression and apoptosis [20,21]. The BET proteins consists of Brd2, Brd3, Brd4 and testis-specific Brtd protein which all contain dual bromodomains at N-terminal regions and recognise AcK and conserved extra-terminal (ET) at C-terminal site which interacts with chromatin modifying proteins BEAS-2B cells were pre-incubated with DCFH-DA for 30 minutes in loading media followed by wash with KRH buffer. Cells were then treated with different concentrations of H 2 O 2 in KRH buffer for 2 hours and intracellular ROS was measured (A). Cells were exposed to a range of concentrations of H 2 O 2 for 2 hours and cell viability was assessed using MTT assay (B). Results are presented as mean 6 SEM. N = 4. *p,0.05; **p,0.01; ***p,0.0001; when compared to basal level (control). BEAS-2B cells were treated with H 2 O 2 for 2 hours in the absence or presence of IL-b stimulation (overnight) or left untreated as a control. IL-6 (C) and CXCL8 (D) protein levels in cell culture supernatants were quantified by ELISA. IL-6 (E) and IL-8 (F) transcript levels were quantified by comparative real-time PCR and were normalised by measuring GNB2L1 transcript levels. Results are expressed as mean 6 SEM of at least 4 independent experiments. * P, 0.05; ** P,0.01; ***P,0.001 when compared to controls. doi:10.1371/journal.pone.0095051.g001 [20,22]. Brd4 forms a complex with positive transcription elongation factor b (p-TEFb) and RNA polymerase II (RNA pol II) at the transcription start site (TSS) to transduce the AcK signal to drive gene expression [23,24].
Recent studies have implicated Brd2 and Brd4 in the regulation of inflammatory genes in murine bone marrow-derived macrophages (BMDMs) [25,26]. Zhang and colleagues have also shown that BET inhibition results in down-regulation of a subset of lineage-specific genes in human CD4+ T-cells [27]. In addition, BET inhibitors have been reported to affect NF-kB-mediated gene expression in renal tubular cells [28], HEK293 and HepG2 cells [29]. In some instances, this reflected targeting of the non-histone acetylated NF-kB p65 subunit by Brd2 rather than an effect of Brd2/4 on AcH [30]. JQ1, a small synthetic compound, has been shown to inhibit the binding of BET proteins to AcH, resulting in reduction of tumour in the mouse model of NUT midline carcinoma [31] and proliferation of c-Myc-dependent proliferation of cancer cells [32][33][34]. Similarly, PFI-1, another Brd4 inhibitor, has been shown to have anti-proliferative effects on leukemic cells lines and abrogates clonogenic growth [35]. However, the anti-inflammatory properties of these compounds yet to be demonstrated under conditions of acute oxidative stressenhanced inflammation in human airway epithelial cells.
In this study we show that both JQ1 and PFI-1, but not the inactive enantiomer of JQ1 [JQ1(-)], can suppress the NF-kBmediated induction of IL-6 and IL-8 in primary human airway epithelial (NHBE) cells and in BEAS-2B cells. This effect was mimicked by knockdown of Brd4 but not of Brd2 and associated with loss of AcH3 and p65 binding to the native gene promoters.

Materials and Methods
Cell culture and treatment SV-40 transformed human bronchial epithelial cells (BEAS-2B) were obtained from the American Type Culture Collection (ATCC, VA, USA) and grown to 70% confluence in keratinocyte conditioned medium (Gibco, Paisley, and UK) at 37uC and 5% CO 2 . Prior to experimentation, cells were serum-starved overnight in medium without epidermal growth factor and bovine pituitary extracts. Cells were stimulated with IL-1b (1 ng/ml) in the presence or absence of H 2 O 2 (100 mM). The expression of IL-6 and CXCL8 was also investigated in the presence or absence of the BET inhibitors (JQ1 and PFI-1) following stimulation with IL-1b alone or in combination with H 2 O 2 .

Normal Human Bronchial Epithelial (NHBE) cells Tissue culture
Normal human bronchial epithelial cells (NHBE) of nonsmoking subjects were obtained from Lonza (Lonza, Slough, UK) and grown in bronchial epithelial cells growth medium (BEGM) supplemented with growth supplements (SingleQouts kit, Lonza) as recommended by the manufacturer. Cells were passaged at passages 2-8 using the ReagentPack subculture kit (Lonza) following suppliers instructions. Cells were cultured until 80% confluent at 37uC and 5% CO 2. Prior to experiments, monolayers of cells (70-80% confluence) were incubated in basal medium (supplement free) overnight. Cells were treated with BET inhibitors (JQ1 and PFI-1) prior to stimulation with IL-1b (1 ng/ml) in the presence or absence of H 2 O 2 (100 mM). All experiments were performed at least four times.

Cell viability
Cell viability was assessed using the methylthiazolyldiphenyltetrazolium bromide (MMT) assay as described previously [36] and the Aqua LIVE/DEAD Fixable dead cell stain kit (Invitrogen, Paisley, UK) exactly according to the manufacturer's instructions. For the former assay, cells (4610 4 /well) were plated in 200 ml serum-free media overnight into 96-well culture plate before incubation with 1 mg/ml MTT solution for 30 mins. MTT solution was then removed and dimethyl sulfoxide (DMSO) was added to dissolve the formazan product to produce a purple solution. The absorbance was measured at 550 nm. The colour intensity was correlated with cell viability. In the latter assay, cells were collected from the plate by trypsinisation and neutralized with serum-free medium, washed and re-suspended in 1 ml PBS. 1 ml of Aqua LIVE/DEAD Fixable dead cell stain was added per sample and left on ice for 30 mins in dark. Cells were washed twice with cold PBS and resuspended in 1% BSA and PBS followed by flow cytometric analysis to distinguish between positive cells (heat-treated dead cells) and negative cells (alive).
Cytokine ELISA IL-6 and CXCL8 expression were quantified by sandwich ELISA (R&D Systems Europe, Abingdon, UK) according to the manufacturer's instructions.

DCF assay for intracellular oxidative stress
Intracellular oxidative stress was detected as previously described [37]. Briefly, cells (4610 4 /well) were plated into 96well culture plate in serum-free media overnight before experimentation. Cells were washed with Krebs-Ringer-Hepes-glucoseglutamine buffer (KRH) on the day of experiment followed by incubation with dichlorofluorescine diacetate (100 mM; DCFH-DA) for 30 mins in loading medium. Cells were washed and incubated with KRH buffer with H 2 O 2 (100 mM) for 2 hrs. The fluorescence from each well was measured with excitation and emission filter set at 485 nm and 530 nm, respectively.

Nuclear and cytoplasmic extractions
Following treatments, cells were collected and nuclear and cytoplasmic proteins were extracted with nuclear extract kit (Active Motif, Carlsbad, California, USA) according to the manufacturer's instructions. The quality and purity of the subcellular fractionation was determined by immunoblotting using Figure 2. NF-kB p65 activates and translocates into the nucleus and IKK2 inhibitor AS602868 diminishes H 2 O 2 -enhanced IL-1b induction of cytokines. Cells were harvested and nuclear proteins were extracted from BEAS-2B cells after treatment with H 2 O 2 (100 mM) in the presence (+) or absence (-) of IL-1b (1ng/ml) for 2 hours. The quality and purity of the subcellular fractionation was determined by immunoblotting using anti-b-actin and TBP (A). Using Western blot analysis, NF-kB p65 nuclear protein was quantified (B). Densitometric analysis of each band is plotted above. NF-kB p65 DNA binding activity was measured using TransAM kit (C). Results are expressed as means 6 SEM as ratio of NF-kB p65/ TATA-binding protein (TBP) or relative to untreated cells. n = 4 independent experiments *p,0.05; **p,0.01; ***P,0.001 compared with unstimulated cells. Cells were pre-treated with AS602868 (5 mM, IKK2 inhibitor) followed by the treatment with H 2 O 2 with or without IL-1b (1 ng/ml) for 16 hours. IL-6 (D) and CXCL8 (E) proteins were assayed by ELISA. The release of IL-6 and CXCL8 was completely inhibited in cells pre-treated with AS602868. Levels of IL-6 (F) and IL-8 (G) mRNA were quantified by real-time RT-QPCR and were normalised with respective to GNB2LI mRNA levels. The RT-QPCR findings are consistent with the ELISA data (H). AS602868 (5 mM) did not affect cell viability using MTT assay. Results are expressed as mean 6 SEM. n = 4. ** P,0.01; ***P,0.001 when compared to H 2 O 2 +IL-1b alone. doi:10.1371/journal.pone.0095051.g002 antibodies against cytoplasmic (b-actin) and nuclear (TATA binding protein) proteins to demonstrate standardization of this method.

NF-kB activation assay
NF-kB activation after 2 hrs was measured in nuclear extracts with TransAM NF-kB activation kit (Active Motif) according to the manufacturer's instructions.

Co-Immunoprecipitation (Co-IP)
Whole cell lysates were prepared by incubating cells with IP buffer (150 mM NaCl, 50 mM Tris-HCl pH 8, 0.5% Sodium deoxycholate, 0.5% NP40, protease inhibitors cocktail [Roche Applied Science, Burgess Hill, UK) for 30 mins on ice and centrifuged at 14000 rpm for 10 mins. 500 mg of protein extract was incubated with 5 mg of antibody Brd4 (Santa Cruz Biotechnology, Santa Cruz, California USA). The immune complex was incubated overnight on rotator at 4uC. Protein A/G-plus agrose beads (20 ml 50% slurry; Santa Cruz Biotechnology) were added to the complex and left on the rotator at 4uC for 2 hours. The the presence (+) or absence (-) of IL-1b (1 ng/ml) for 2 hours, nuclear (A) and cytoplasmic (B) extracts were fractioned by Western blot and membranes were probed with anti-acetylated NF-kB p65 antibody. The blots show that acetylated-310 (Ac310) NF-kB p65 is predominantly found in the nucleus when compared with the cytoplasm. (C) Brd4 protein was immunoprecipitated from whole cell extracts following treatments and separated by SDS-PAGE and subsequently analysed by Western blotting using an anti-NF-kB p65 antibody. Each blot is representative of 3 independent experiments and densitometric analysis of each band is plotted as bar graph above it. TBP: TATA-binding protein; *p,0.05; **p,0.01 compared with control (unstimulated). doi:10.1371/journal.pone.0095051.g003 complex was then washed using IP wash buffer (50 mM Tris-HCl pH 7.4, 0.5% (v/v) NP40, 150 mM NaCl) three times. Proteins were eluted from complex using 35 ml of elution buffer (50 mM Glycine, HCl pH 2.4) followed by addition of 3.5 ml of neutralising solution (2 M Tris pH 8, 1.5 M NaCl, 1 mM EDTA). Samples were run on SDS-PAGE gel.

Western blotting
Nuclear and cytoplasmic extracts were fractionated by SDS-PAGE gel as previously described [38]. Proteins were transferred and immobilised on nitrocellulose membranes. Membranes were incubated with specific antibodies directed against NF-kB p65, bactin, TATA-binding protein (All from Santa Cruz Biotechnology), acetylated p65-lysine 310 (Abcam, Cambridge, UK), Brd2 and Brd4 (Bethyl Lab, Montgomery, TX, USA). After washing, membranes were incubated with horseradish peroxidase-linked anti-rabbit immunoglobulin (DAKO Ltd, Ely, UK) and detected by enhanced chemiluminescence (ECL) (GE healthcare, Amersham, UK). The membranes were exposed to X-ray and the bands density was quantified using the GelDoc Imaging System (UVP, Upland, CA). This was standardised against the loading controls and the data were represented in graphs.

Measurement of mRNA transcripts
Cells were treated as above and total RNA was isolated using an RNeasy mini kit (Qiagen, Crawley, UK) and the concentration of RNA determined using the absorbance ratio of 260 nm/280 nm by NanoDrop 2000c spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). Single stranded cDNA was synthesized using AMV-reverse transcriptase (Promega, Southampton, UK) as described previously [39]. QPCR was performed in a Rotor-Gene 3000TM PCR machine (Corbett, Research, Cambridge, UK) using a QuantiTect SYBR Green PCR Kit (Qiagen) according to the manufacturer's instructions. The RT-PCR primers used were as follows: Brd2 (NCBI: NM_005104.

Chromatin Immunoprecipitation (ChIP) assay
ChIP assay was carried out as previously described [38]. Briefly, epithelial cells (BEAS-2B) were treated with H 2 O 2 (100 mM) in the presence or absence of IL-1b (1 ng/ml) or left untreated for 2 hrs. In some experiments, cells were pre-treated with JQ1 (500 nM) for 4 hrs. After treatments, protein-DNA complex were fixed by formaldehyde (1% final concentration) and immunoprecipitated with either an anti-p65 (5 mg, Santa Cruz Biotechnology) or an anti-Brd4 (5 mg, Bethyl Lab) antibody overnight. NF-kB p65 and Brd4 binding to the IL-8 and the IL-6 promoter were quantified by real-time QPCR using a QantiTech SYBR green PCR kit (Qiagen) on Rotor gene 3000 (Corbett Research). Immunoprecipitated DNA was quantified by qPCR and compared to the level of promoter specific DNA obtained at t = 0 which was given an arbitrary value of 1. IL-8 [40,41] and IL-6 [42][43][44]  siRNA transfection BEAS-2Bs cells were seeded in 6-well plates at 0.3610 6 cells/ well and transfected with ON-Target plus SMART pool Brd2 (L-004935-00-0005) or Brd4 (L-004937-00-0010) or control (D-001810-10-05) small interfering RNAs (siRNA) at 20 nM concentration (all from Dharmacon Thermo Scientific, Chicago, IL, USA) using HiPerFect transfection reagent (Qiagen) as described by the manufacturer. After 72 hrs, nuclear extracts were examined for Brd2 and Brd4 protein expression using Western blotting. Total mRNA was extracted and converted to cDNA and Brd2 and Brd4 expression were assessed.

Statistical analysis
Data are represented as mean 6 standard error mean (SEM). Data were analysed by Student's t test for two sets of data or by one-way ANOVA/Dunn's multiple comparison test for comparing more than two sets of data. GraphPad Prism (La Jolla, CA, US) was used to evaluate the data. Differences were considered significant for P values of #0.05. Figure 4. H3 acetylation, p65 and Brd4 binding to IL-6 and IL-8 kB promoter sites. Chromatin immunoprecipitation (ChIP) assays show that IL-1b induces p65 DNA binding to both IL-6 (A) and IL-8 (B) promoters. H 2 O 2 by itself does not affect p65 DNA binding activity; however, when cotreated with IL-1b, the affinity is enhanced by 7-fold at IL-6 promoter site and 20-fold at IL-8 promoters. Brd4 is also recruited to the same kB promoter regions in the IL-6 (C) and IL-8 (D) promoters as p65. Histone 3 is acetylated at the IL-6 (E) and IL-8 (F) kB promoter sites following treatments. IgG is non-specific antibody used as a negative control. Furthermore, H3 acetylation was confirmed using confocal microscopy following IL-1b stimulation whereas H 2 O 2 had no effect on AcH3 alone or in combination with IL-1b (G). Results are representative of at least 4 independent experiments.*p,0.05, **p,0.01 compared with control (unstimulated). doi:10.1371/journal.pone.0095051.g004

Results
Hydrogen peroxide (H 2 O 2 ) induces intracellular oxidative stress and enhances inflammatory mediator expression Exogenous H 2 O 2 (0-800 mM) enhanced intracellular ROS in BEAS-2B cells a concentration-dependent manner after 2 hours which reached significance at concentrations of 100 mM or above (Fig. 1A). However, at concentrations .200 mM there was a significant loss in cell viability (Fig. 1B) and H 2 O 2 (100 mM) was selected for all future experiments. H 2 O 2 alone had no significant effect on either IL-6 or CXCL8 release at 16 hrs; however, H 2 O 2 significantly enhanced the production of IL-6 ( Fig. 1C) and CXCL8 (Fig. 1D) release from IL-1b (1 ng/ml) co-stimulated cells. This effect was mirrored by changes in IL-6 and IL-8 mRNA expression (Fig. 1E-F).
The expression of IL-6 and IL-8 are regulated by the NF-kB signalling pathway The subcellular extraction process was demonstrated to be highly reproducible with little or no cross-contamination of cytoplasmic and nuclear extracts using Western blotting ( Fig. 2A). H 2 O 2 alone had no significant effect on p65 nuclear translocation after 2 hrs whereas it enhanced IL-1b-induced p65 nuclear import (Fig. 2B). IL-1b significantly enhanced p65 DNA binding to a consensus kB-response sequence (59-GGGACTTTCC-39) 7-fold and this was further increased (up to 9-fold greater than baseline) by pre-treatment with H 2 O 2 (Fig. 2C). In contrast, H 2 O 2 alone had no significant effect on NF-kB p65 binding activity (Fig. 2C). Furthermore, IL-1b+H 2 O 2 -induced release of both IL-6 and CXCL8 proteins (Fig. 2D, E) and the expression of IL-6 and IL-8 mRNA (Fig. 2F, G) was completely suppressed by the selective IKK2 inhibitor AS602868 [45,46]. Cell viability was not affected by AS602868 (Fig. 2H).

Effect of IL-1b and H 2 O 2 on p65 acetylation and association with Brd4
NF-kB p65 is subjected to post-translational modifications such as acetylation that modulate its activity [47,48]. Western blot analysis revealed that acetylated NF-kB p65 is found predominantly in the nuclear compartment in IL-1b stimulated cells after 2 hrs with little in the cytoplasmic compartment after any treatment (Fig. 3A, B). Although, H 2 O 2 alone did not induce p65 K310 acetylation, IL-1b-induced p65 acetylation was enhanced after IL-1b+H 2 O 2 co-stimulation (Fig. 3A). This suggests that acetylation of NF-kB at lysine-310 is associated with maximal NF-kB activation and/or translocation into the nucleus [49]. Brd4 can interact with acetylated p65 [24,28] as part of a complex with p-TEFb and RNA polymerase II [28,30]. Using coimmunoprecipitation experiments we found that there was an association between acetylated p65 and Brd4 (Fig. 3C), however, neither H 2 O 2 nor IL-1b or in combination affected this association. This suggests that acetylated p65-Brd4 association is not directly linked to ROS-induced enhancement of NF-kB function.
NF-kB p65 and Brd4 are recruited to the native IL-6 and IL-8 promoters Following 2 hrs IL-1b treatment, p65 ChIP analysis (Fig. 4A,  B) showed a marked enhancement in binding to kB elements within the native IL-6 and IL-8 promoters which was enhanced by co-stimulation with H 2 O 2 (7-and 20-fold respectively). In contrast, H 2 O 2 alone had no effect on p65 binding to either the IL-6 or IL-8 promoters. IL-1b also significantly increased binding of Brd4 to these kB sites in the IL-6 and IL-8 promoters (Fig. 4C, D). Again, this recruitment was augmented by co-stimulation with H 2 O 2 and IL-1b at the IL-6 (6-fold) and IL-8 (8-fold) promoters compared to unstimulated cells (Fig. 4C, D). Pan-histone H3 acetylation at the IL-6 and IL-8 promoter kB sites was significantly elevated following IL-1b stimulation and slightly further increased with the addition of H 2 O 2 although this did not reach significance (Fig. 4E, F). H 2 O 2 alone did not affect Brd4 nor pan-histone H3 acetylation at these sites ( Fig. 4C-F). Confocal analysis also confirmed an increased H3 acetylation in cells stimulated with IL-1b in comparison with unstimulated cells whereas H 2 O 2 alone had little effect (Fig. 4D). We did not observe H4 acetylation at either of the promoters at the 2 hr time point studied, suggesting that H4 acetylation might be time-and gene-dependent.  (Fig. 5A, B). PFI-1, a structurally distinct BET inhibitor [50], also attenuated the release of both IL-1b-induced IL-6 ( Fig. 5C) and CXCL8 (Fig. 5D) in concentration-dependent manner. The IC 50 for PFI-1 inhibition of IL-1b-induced IL-6 protein levels was 7.364.2610 -7 M and a similar effect was seen for the inhibition of IL-1binduced CXCL8 release (7.464.8610 27 M). PFI-1 inhibited IL-1b-induced IL-6 and CXCL8 release with maximal suppression of (80.761.9%) and (63.663.8%), respectively.

BET inhibitors reduce IL-1b induced inflammation
Cell viability was significantly affected at the highest concentrations of JQ1 and JQ1(-) tested (10 26 M, Fig. 5E) and does not therefore account for the reduction in IL-6 and CXCL8 release seen at 5610 27 M. High concentrations of PFI-1 (4610 26 M) also significantly reduced cell viability and were excluded from subsequent experiments (Fig. 5F). These results indicated that PFI-1 has a similar efficacy as JQ1. We used PFI-1 at 1610 26 M in subsequent experiments. The lack of effect of JQ1 and PFI-1 on cell viability as these concentrations was confirmed using FACS analysis and Live/Dead Aqua blue staining (Fig. 5G). Cells treated with DMSO (control), JQ1 (0.5 mM), JQ1(-) (0.5 mM) or with PFI-1 (1 mM) resulted in only 5% overall cell death with 95% cells being viable. **p,0.01;***P,0.001 compared with IL-1b stimulation. # p,0.05 JQ1(-) versus JQ1. £££ p,0.001when PFI-1 compared with IL-1b stimulation. $ p,0.05; $$ p,0.01when compared to control. (G) Cells were heat treated at 90uC or left untreated, mixed together and stained with LIVE/DEAD Fixable Aqua stain then analysed by flow cytometry. Cells were checked with forward scatter detector (FSC) and side scatter detector (SSC) and analysed by density graph to check cell size and granularity. Fragmented cells were excluded from the study. Histogram shows separation of live cells (left) and dead cells (right). These parameters were used to assess cell viability following treatment with JQ1 (0.5 mM) and PFI-1 (1 mM) for 16 hours. DMSO/Control (,1%) alone, PFI-1, JQ1(-) or JQ1 resulted in only 5% decrease of overall cell viability. The data is representative of 3 independent experiments. doi:10.1371/journal.pone.0095051.g005 We demonstrated above that H 2 O 2 (100 mM) alone has minimal effect on IL-6 and CXCL8 expression; however, it enhanced the expression of both cytokines (IL-6 and CXCL8) in IL-1b-induced cells. The release of IL-1b H 2 O 2 -induced IL-6 ( Fig. 6A) and CXCL8 (Fig. 6B) proteins and IL-6 ( Fig. 6C) and IL-8 (Fig. 6D) mRNA was markedly suppressed in JQ1 but not JQ1(-) treated cells. However, the levels of IL-6 and CXCL8 did not return to baseline. A similar effect was seen with PFI-1 which attenuated the release of IL-6 ( Fig. 6E) and CXCL8 (Fig. 6F) proteins and IL-6 ( Fig. 6G) and IL-8 (Fig. 6H) mRNA when compared with untreated cells. As with JQ1, mediator levels did not return to baseline in PFI-1 treated cells.
The effects of JQ1 and PFI-1 are mimicked by Brd4, but not Brd2, knockdown To determine the specificity of the JQ1 and PF-1 effects, cells were transfected with Brd4 or non-specific small interfering RNAs for 72 hours resulting in significantly decreased expression of Brd4 mRNA (71.863.2%, p,0.05) (Fig. 7A) and protein 63.2667.5%, p,0.05) expression (Fig. 7B). The release of IL-1b/H 2 O 2stimulated IL-6 ( Fig. 7C) and CXCL8 (Fig. 7D) proteins was significantly suppressed in Brd4-knockdown cells in comparison with control non-specific siRNA transfected cells. The degree of suppression of IL-6 and CXCL8 was the similar to that seen in JQ1-and PFI-1-treated cells (Fig. 6). In contrast, although there is an 80% structural homology between Brd2 and Brd4 [51,52] we were unable to show any effect of Brd2 knockdown on inflammatory gene expression. Following treatment of cells with Brd2 siRNA, there was a significant suppression of both Brd2 mRNA (Fig. 7E) and protein (Fig. 7F) expression, but no effect on IL-1b/H 2 O 2 -stimulated IL-6 ( Fig. 7G) or CXCL8 (Fig. 7H) release. These data indicate that Brd4 is central to NF-kBmediated induction of IL-6 and IL-8 expression and that JQ1 and PFI-1 preferentially target Brd4 and not Brd2 in airway epithelial cells.

JQ1 inhibits p65 and Brd4 association at IL-6 and IL-8 promoters
ChIP analysis was used to investigate the mechanism of JQ1 at the p65 to the IL-6 and IL-8 promoters. Binding of Brd4 and p65 increased 5-fold at the IL-6 promoter in H 2 O 2 +IL-1b stimulated cells when compared with unstimulated cells. This was not affected by the presence of the inactive enantiomer JQ1(-). However, JQ1 significantly reduced Brd4 (Fig. 8A) and p65 (Fig. 8B) binding at the IL-6 promoters. Similarly, Brd4 and p65 association with the IL-8 promoter increased 4-fold and 10-fold respectively in stimulated cells (H 2 O 2 +IL-1b). This association was abolished by JQ1 but was unaffected by JQ1(-) (Fig. 8C, D). The results also show that although the recruitment of p65 and Brd4 is completely abrogated at both the IL-6 and IL-8 promoters the expression of both IL-6 and IL-8 did not returned to baseline levels in JQ1 pretreated cells. This suggests that post-transcriptional factors may also play a role in driving the expression of IL-6 and CXCL8 expression in these cells.
We then examined the effect of H 2 O 2 on NHBE cells function. H 2 O 2 significantly induced intracellular ROS at concentrations . 100 mM when compared to untreated cells (Fig. 9C) although this was associated with significantly reduced cell survival at concentrations .200 mM) (Fig. 9D). Therefore, NHBE cells were treated with 100 mM of H 2 O 2 in the presence or absence of IL-1b in subsequent experiments. IL-1b-induced release of IL-6 ( Fig. 9E) and CXCL8 (Fig. 9F) was further enhanced when co-treated with H 2 O 2 which was mirrored by changes in IL-6 ( Fig. 9G) and IL-8 (Fig. 9H) mRNA transcripts. There was no significant difference between IL-6 and CXCL8 expression between JQ1(-) and JQ1 treated cells under control conditions. JQ1 markedly attenuated the release of IL-1b/H 2 O 2 -induced IL-6 ( Fig. 9E) and CXCL8 (Fig. 9F) proteins and IL-6 ( Fig. 9G) and IL-8 (Fig. 9H) mRNA expression when compared with JQ1(-) treated cells.

Discussion
In this study, we show that H 2 O 2 induces intracellular ROS in human airway epithelial cells (BEAS-2B and NHBE cells) and enhances IL-1b-induced IL-6 and CXCL8 expression via increased p65 and Brd4 promoter association at kB sites in the native gene promoters. This is in conjunction with increases in histone H3 acetylation at these kB sites. The BET inhibitors JQ1 and PFI-1 both suppressed IL-1b-and IL-1b/H 2 O 2 -induced IL-6 and CXCL8 protein and mRNA expression in a concentrationdependent manner. Knockdown studies revealed that IL-6 and CXCL8 release is significantly reduced in Brd4, but not Brd2, depleted cells. Furthermore, H 2 O 2 -enhanced IL-1b-induced recruitment of p65 and Brd4 to the IL-6 and IL-8 promoters is abolished in JQ1-, but not in JQ1(-)-, pre-treated cells.
COPD primarily affects lungs; however, it now recognised a disease with organ-specific characteristics and systemic manifestations such as chronic inflammation [53][54][55]. We have analysed the expression of two of the most important pro-inflammatory cytokines, IL-6 and CXCL8, that are elevated in plasma, BAL fluids and sputum of COPD patients and whose expression correlates with disease severity [56][57][58]. These cytokines are regulated by the NF-kB signalling pathway which also orchestrates many aspects of the cellular immune response, apoptosis, differentiation and inflammation [59,60]. High concentrations of H 2 O 2 are found in breathe condensates of COPD patients and this Figure 7. Knockdown of Brd4 reduces inflammation but not Brd2. Cells were transfected with either pooled Brd4 or Brd2 siRNAs (20 nM) or non-specific siRNAs (20 nM). 72 hours post-transfection Brd4 mRNA (A) was quantified by RT-QPCR and nuclear extracts were analysed (B) by immunoblotting using anti-Brd4 and -TATA-binding protein (TBP). IL-6 (C) and CXCL8 (D) proteins were measured by ELISA in post-transfected cells following stimulation with H 2 O 2 and IL-1b together. n = 4 independent experiments. 72 hours post-transfection, Brd2 mRNA (E) was quantified by RT-QPCR and nuclear extracts were analysed (F) by immunoblotting using anti-Brd2 and TBP. Brd2 siRNA knockdown had no effect on either IL-6 (G) or CXCL8 (H) expression. n = 4 independent experiments. Results represent mean 6 SEM *p,0.05, **p,0.01, ***p,0.001, non-specific siRNA compared to Brd2 or Brd4 siRNA. doi:10.1371/journal.pone.0095051.g007 has been linked to disease pathophysiology [61,62]. For example, H 2 O 2 can induce the release of potent chemotactic factors such as CXCL8 from epithelial cells [63] by resulting in increased hyperacetylation at the IL-8 promoter via alteration in HDAC/HAT balance [64].
NF-kB is a redox-sensitive transcription factor that converts oxidative stress signals into changes in gene expression associated with diverse cellular activities including inflammation [65]. NF-kB activation requires translocation of the NF-kB p65/p50 dimer into the nucleus where it binds to promoter regions of inflammatory genes forming a transcriptional activator complex [22,66]. Optimal induction of NF-kB activity also requires post-translation modification of p65 e.g. phosphorylation, acetylation and ubiquitination, which modulates DNA binding and transcriptional activity. For example, acetylation of p65/RelA at Lysine-310 (K310) is essential for optimal NF-kB transcriptional activity [28,30,67]. Acetylated p65 provides a docking site for Brd4 and may be an important target in the ability of BET inhibitors to suppress inflammation [30]. In the present study, we demonstrated that p65 is acetylated within the nucleus following stimulation of cells with IL-1b and H 2 O 2 although this did not affect the degree of p65:Brd4 association in the nucleus under any of the conditions tested. This data is in contrast to that observed in A549 lung adenocarcinoma-like cells where TNFa stimulation was essential for p65:Brd4 complex formation [30]. Our data suggests that in bronchial airway epithelial cells that Brd4 and p65 may be part of a pre-formed transcriptional complex.
An interesting observation made in this study is that the recruitment of Brd4 is associated with enhanced pan-acetylation of histone H3 but not of H4 at the IL-6 and IL-8 promoters confirming previous data [68][69][70]. Although, H 2 O 2 alone was unable to induce H3 acetylation directly in BEAS-2B cells, it slightly enhanced the histone acetylation seen with IL-1b alone. This may reflect either a direct effect on localised HAT activity or a reduction in HDAC activity as previously reported in BEAS-2B cells [37]. This will result in an alteration in the local chromatin environment/structure which will modulate transcription factor association with promoter regions [71]. Future studies utilising DNase1 hypersensitivity assays, for example, may reveal changes in local chromatin structure following H 2 O 2 treatment of cells which will enable greater recruitment of pTEFb by Brd4 and subsequent elongation of inflammatory gene expression by RNA polymerase II [72].
The chemically distinct bromodomain inhibitors (JQ1 and PFI-1) decreased recruitment of both Brd4 and p65 to the IL-6 and IL-8 promoters and reduced IL-6 and CXCL8 protein release from BEAS-2Bs and NHBE cells. This confirms data from Nicodeme and colleagues who reported that I-BET disrupts the interaction of bromodomain proteins and acetylated histones at the IL-6 promoter in LPS-stimulated macrophages [25]. Inhibition of Brd4 either by JQ1 or knockdown reduced both IL-1b-and IL- 1b/H 2 O 2 -induced IL-6 and CXCL8 expression by ,80% whereas JQ1 completely blocked p65 DNA binding and subsequent Brd4 recruitment at the kB sites studied. AS602868, an IKK2 inhibitor, almost completely abrogated IL-1b-and IL-1b/ H 2 O 2 -induced IL-6 and CXCL8 protein and mRNA expression confirming a key role for NF-kB in the transcriptional control of these genes. This suggests that other kB sites within the IL-6 and IL-8 promoters or enhancers may play a role in gene expression or that other inflammatory signalling pathways may modify the effect of JQ1 at the transcriptional and/or translational level. Future studies using ChIP-seq may address this. The failure to completely suppress CXCL8 gene expression may be important clinically as total suppression of the innate immune response may lead to opportunistic infections. The ability of JQ1 to dampen rather than ablate the immune/inflammatory response may therefore be of benefit in the treatment of patients with a chronic inflammatory disease. It will be of interest to determine whether JQ1 protects other innate immune genes in epithelial cells following stimulation with IL-1b or IL-1b/H 2 O 2 .
This is the first report to our knowledge that demonstrates a repressive effect of JQ1 and PFI-1 on IL-1b-induced CXCL8 expression. Previous studies [25,73] have showed that CXCL1 mRNA, the murine equivalent of CXCL8, was unaffected by I-BET in mouse macrophages at baseline or after LPS-stimulation. However, Huang et al. [30] showed that the Brd4 depletion in LPS-stimulated THP-1 impaired CXCL8 expression. In our study Brd2 depletion in BEAS-2B cells had no effect on inflammatory cytokine expression. In contrast, Belkina et al. have shown that, in addition to Brd4, LPS-induced inflammation in murine bone marrow-derived macrophages (BMDMs) is mediated via Brd2 which binds to the IL-6 promoter [73]. This difference may be species or cell type specific. ChIP-seq analysis indicates that Brd4 co-localizes with RNA pol II at both enhancers and promoters of all active genes in primary human CD4+ T cells. This interaction is disrupted upon JQ1 treatment leading to reduced lineagespecific gene expression and acetylated histone-bromodomain association [27]. The study implies that Brd4 could be used as a tool to identify active promoters and enhancers in a genome-wide manner in human epithelial cells. JQ1 has also been studied in NUT (Nuclear protein in testis) midline carcinoma (NMC) mouse model and various cancer cell lines, showing inhibition of cell proliferation and differentiation [31,33]. Furthermore, Brd4 is believed to be an inherited susceptibility gene for breast cancer progression and metastasis. It has been shown that deletion of Brd4 proline-rich domain at C-terminal results in a loss of contact inhibition and changed in cell morphology from epithelial to a mesenchymal-like phenotype [74].
There are some limitations to this study. Firstly, our data shows a difference in IL-6 and CXCL8 protein release in response to IL-1b and to IL-1b+H 2 O 2 in the presence and absence of JQ1 suggesting that whatever effect IL-1b has on Brd4 function this is modified at least to some extent by the presence of H 2 O 2 particularly in primary cells. However, we are unable to rule out the possibility that H 2 O 2 stimulation modifies other aspects of p65 activity. Analysis of the effects of Brd4 knockdown on IL-1b-and H 2 O 2 -stimulated cells on mediator protein, mRNA and promoter ChIP analysis would help resolve this issue. Secondly, we show a differential effect of BET mimics on IL-6 and CXCL8 protein release compared with mRNA expression data. This indicates a degree of disconnect in both BEAS-2B and primary epithelial cells. Additional experiments, using combinations of BET protein knockdowns and pharmacological inhibition, should be performed to examine non-histone-based mechanisms of translational control and secretion.
In summary, we have shown that IL-1b and IL-1b/H 2 O 2 can enhance recruitment of p65 and Brd4 to the native IL-6 and IL-8 promoters in epithelial cells. Inhibition of Brd4 by the structurally distinct bromodomain inhibitors JQ1 and PFI-1 and by genetic knockdown led to a reduction in IL-6/-8 mRNA and protein expression. This was associated with an attenuation of p65 and Brd4 binding to the native IL-6 and IL-8 promoters. The specificity and long-term side effects of bromodomain inhibitors needs to be evaluated if these agents are to be used in chronic inflammatory diseases such as COPD and potentially in the progression to lung cancer [75]. Despite these concerns, our findings suggest that inhibition of Brd4 may have therapeutic potential in inflammatory diseases where oxidative stress and NF-kB activation is present.