The FASEB Journal • Research Communication

Glucocorticoid suppression of CX3CL1 (fractalkine) byreduced gene promoter recruitment of NF-�B

Pankaj K. Bhavsar,1 Maria B. Sukkar, Nadia Khorasani, Kang-Yun Lee, andKian Fan ChungExperimental Studies, Airway Disease Section, National Heart and Lung Institute, Imperial CollegeLondon, UK

ABSTRACT Glucocorticoids are an important anti-inflammatory treatment of many inflammatory diseasesincluding asthma. However, the mechanisms by whichthey mediate their suppressive effects are not fullyunderstood. Respiratory epithelial cells are a source ofCX3CL1 (fractalkine), which mediates cell adhesionand acts as a chemoattractant for monocytes, T cells,and mast cells. We show, in lung A549 epithelial cells,that the tumor necrosis factor-� (TNF-�) and IFN�synergistically induced protein release and mRNA ex-pression of CX3CL1 is inhibited by dexamethasone,without interfering with cytokine-induced nuclear trans-location of NF-�B, and by an inhibitor of I�B kinase 2,AS602868. DNA binding assays confirmed the ability ofNF-�B to bind to the proximal CX3CL1 promoter.Chromatin immunoprecipitation assays showed a 5-foldincrease in the recruitment of NF-�B to the CX3CL1gene promoter in response to IFN�/TNF-�; this toowas reversed by dexamethasone. In contrast, dexameth-asone did not displace NF-�B from the granulocyte-macrophage colony-stimulating factor gene promoter.We conclude that CX3CL1 expression is regulatedthrough the NF-�B pathway and that dexamethasoneinhibits CX3CL1 expression through a glucocorticoidreceptor-dependent (RU486 sensitive) mechanism.This study also provides support for the action ofglucocorticoids mediating their suppressive effects onexpression by interfering with the binding of transcrip-tional activators at native gene promoters.—Bhavsar,P. K., Sukkar, M. B., Khorasani, N., Lee, K.-Y., Chung,K. F. Glucocorticoid suppression of CX3CL1 (frac-talkine) by reduced gene promoter recruitment ofNF-�B. FASEB J. 22, 1807–1816 (2008)

Key Words: epithelial cells � transcription factor � displacement� chromatin

Glucocorticoids are effective anti-inflammatorytreatments for many inflammatory conditions, includ-ing asthma. However, the mechanisms by which theymediate their suppressive effects are not fully under-stood. Glucocorticoids bind to the glucocorticoid re-ceptor (GR), a ligand-dependent transcription factorbelonging to the superfamily of steroid hormone recep-tors (1). In the absence of ligand, GR is retained in the

cytoplasm by association with chaperone proteins(hsp90 and p23). Ligand binding releases the chaper-one proteins, allowing dimerization and translocationof the GR into the nucleus. In the nucleus, GR binds tothe glucocorticoid receptor element (GRE) and, de-pending on the context of the gene promoter, activatesor represses transcription. The transcriptional activityof GR depends on its ability to recruit coactivators,which regulate transcription by remodeling chromatinor by facilitating the recruitment of the basal transcrip-tional machinery (2, 3). Glucocorticoids can eitherincrease the expression of anti-inflammatory genessuch as secretory leukocyte protease inhibitor (4) andmitogen-activated kinase phosphatase-1 (5) or decreasethe expression of NF-�B and/or activating protein(AP-1) -dependent proinflammatory genes.

Respiratory epithelial cells not only act as a physicalbarrier but also contribute to airway inflammation byreleasing various mediators, such as cytokines andchemokines, and by expressing enzymes and surfacemembrane molecules involved in the adhesion of leu-kocytes (6). The capacity of airway epithelial cells tosynthesize CX3CL1 (fractalkine) may contribute to thepathophysiology of respiratory diseases such as asthmaand chronic obstructive pulmonary disease (COPD).For example, a recent study has reported an increase inexpression of CX3CL1 in airway epithelium, submu-cosa, and smooth muscle of patients with asthma com-pared with nonasthmatic control subjects (7). Further-more, the concentration of CX3CL1 in thebronchoalveolar lavage fluid has been shown to beincreased after segmental allergen challenge in asth-matic patients (8). CX3CL1 expression is also increasedin the lung tissue of patients with COPD [NHLBI/WHO Global Initiative for Chronic Obstructive LungDisease (GOLD) -2] compared with control smokers(GOLD-0), indicating a possible role for CX3CL1 in thepathophysiology of COPD (9).

Chemokines are a family of low-molecular-weightproteins that are primarily involved in leukocyte migra-

1 Correspondence: Airway Disease, National Heart andLung Institute, Guy Scadding Bldg., Imperial College Lon-don, Dovehouse St., London, SW3 6LY, UK. E-mail:p.bhavsar@imperial.ac.uk

doi: 10.1096/fj.07-094235

18070892-6638/08/0022-1807 © FASEB

tion (10). CX3CL1 is a multidomain chemokine be-longing to the CX3C family and is expressed either onthe cell surface (m-CX3CL1) or as a soluble glycopro-tein (s-CX3CL1) (11). The two forms have differingfunctions whereby m-CX3CL1 primarily mediates celladhesion and s-CX3CL1 acts as a chemoattractant.Stimulation by proinflammatory cytokines such as tu-mor necrosis factor-� (TNF-�), IFN�, and interleukin(IL) -1� can induce the expression of both forms ofCX3CL1 in a number of nonhemopoietic cells (12),vascular smooth muscle cells (13), dermal fibroblasts(14), astrocytes (15), and airway smooth muscle cells (16).

Proinflammatory cytokines, such as IFN� or TNF-�,bind to their respective receptors, resulting in the rapidactivation of the NF-�B/Rel transcription factor family(17). The activity of NF-�B is regulated by a family ofinhibitory NF-�B (I�B) proteins, which sequesterNF-�B to the cytoplasm. In response to activatingsignals, such as cytokines, degradation of I�B allowsNF-�B, typically heterodimers of p50 and p65, to trans-locate to the nucleus (18), where it binds to specificpromoter elements and induces gene transcription.Neither the functional interplay nor the order ofrecruitment of the basal transcriptional machinery andNF-�B complexes has been established at the promot-ers of genes expressed in human airway cells.

IFN� and TNF-� are pleiotropic cytokines that oftenplay a critical role during an inflammatory responseand have been shown to function cooperatively orantagonistically in controlling the expression of a num-ber of cytokines and cell surface molecules (19, 20).IFN� signaling activates signal transducer and activatorof transcription 1 (STAT1) (21), and NF-�B mediatesthe TNF-�-induced transcriptional activation of cyto-kines and cell surface molecules (22). Functional syn-ergy between IFN� and TNF-� in transcriptional activa-tion may be mediated by cooperation between STAT1and NF-�B (23). Moreover, many of the genes that aresynergistically activated by IFN� and TNF-�, such asintercellular adhesion molecule-1 (24), regulated onactivation normal T cell expressed and secreted (25),and interferon regulatory factor-1 (26) have DNA bind-ing sites for both STAT1 and NF-�B. In keeping withgenes coactivated by IFN� and TNF-�, DNA sequenceanalysis of the CX3CL1 gene promoter has identifiedputative STAT1 and NF-�B cis-acting regulatory ele-ments.

The mechanisms by which glucocorticoids inhibitinflammatory gene expression, such as that of CX3CL1,are not fully understood. The ability of the GR toattenuate NF-�B-dependent gene transcription hasbeen studied, and a direct interaction between the GRand p65 had been proposed previously (27). We hy-pothesized that the attenuation of NF-�B-dependentgene transcription is associated with a reduction inNF-�B binding at specific gene promoter sites. In thisstudy, we examined the expression and regulation ofCX3CL1 in respiratory epithelial cells and a possiblemechanism through which glucocorticoids mediatetheir antiinflammatory effect.

MATERIALS AND METHODS

Reagents and antibodies

Recombinant human cytokines, goat IgG anti-humanCX3CL1 (AF-365), and biotinylated anti-human CX3CL1(BAF365) polyclonal antibodies used for ELISA assays werepurchased from R&D Systems Europe (Abingdon, UK). An-tibodies for chromatin immunoprecipitation (ChIP) assayswere purchased from Santa Cruz Biotechnology Inc. (SantaCruz, CA, USA). Dexamethasone, RU486, actinomycin D-mannitol, and all other tissue culture reagents were pur-chased from Sigma (Paisley, UK). Oligonucleotide primersfor real-time polymerase chain reaction (PCR) were synthe-sized by Sigma Genosys (Pampisford, Cambridgeshire, UK),and primers for ChIP assays were synthesized by MWG(Ebersberg, Germany). SC-514 NF-�B pathway inhibitor(#401479) was purchased from Calbiochem (San Diego, CA,USA).

Cells and culture

A549 cells (American Type Culture Collection, Manassas, VA,USA) were grown at 37°C in a 5%:95% CO2-air atmosphere inDulbecco’s modified Eagle’s medium (DMEM) supple-mented with 10% FBS, 1 mM sodium pyruvate, and 2 mMl-glutamine. The cell cultures were maintained in a 5% CO2humidified incubator at 37°C and passaged when cell densityreached 80% confluence.

A549 cells were grown to 80% confluence in 6-well tissueculture plates as described above and then serum-deprived inDMEM supplemented with 1 mM sodium pyruvate and 2 mMl-glutamine for 24 h. After serum deprivation, cells wereincubated in fresh (prewarmed, 37°C) serum-free DMEMcontaining cytokines and drugs as indicted for each experi-ment; dexamethasone was added 1 h before addition ofcytokines. ELISA assays for measurement of soluble CX3CL1in culture supernatants were performed according to themanufacturer’s instructions (R&D Systems Europe).

RNA extraction and real-time quantitative PCR

Total RNA was extracted from A549 cells using the RNeasyMini Kit (Qiagen, West Sussex, UK) according to the manu-facturer’s instructions. cDNA for real-time quantitative PCRanalysis was synthesized from 0.5 �g of total RNA usingrandom hexamer primers (Promega, Madison, WI, USA) andavian myeloblastosis virus reverse transcriptase (Promega).Real-time PCR was performed using the Rotor Gene 3000(Corbett Research, Sydney, Australia); PCR amplificationswere performed in a 20-�l volume using the SYBR GreenMaster Mix Reagent Kit (Promega). To control for variationwithin the procedure a reference mRNA, glyceraldehyde3-phosphate dehydrogenase (GAPDH), was also measured.Primers used were as follows: CX3CL1 forward primer, 5�-CCTGTAGCTTTGCTCATCCACTATC-3�; CX3CL1 reverseprimer, 5�-TCCAAGATGATT-GCGCGTT-3�; GAPDH forwardprimer, 5�-GAAGATGGTGATGGGATTTC-3�; and GAPDHreverse primer, 5�-GAAGGTGAAGGTC GGAGT-3�. PCR cy-cling conditions were 15 min at 95°C followed by 45 cycles of20 s at 94°C, 20 s at 60°C, and 20 s at 72°C. Data were analyzedusing the comparative ct method set out in User Bulletin 2(product no. 4303859; Applied Biosystems, Inc., Foster City,CA, USA).

ChIP assays

ChIP assays were performed using the ChIP Assay Kit suppliedby Upstate Biotechnology (Lake Placid, NY, USA) and the

1808 Vol. 22 June 2008 BHAVSAR ET AL.The FASEB Journal

protocol therein. Briefly, A549 cells were seeded at 2 � 106

cells/6-well plates and cultured as described above. After 48 hcells were serum-deprived for 24 h. Cells were pretreated withdexamethasone (1 �M) for 1 h and then stimulated withcytokines for 1 h, after which they were fixed in 1% formal-dehyde. Cells from two wells were pooled per sample andthen lysed, and the DNA was fragmented by sonication (5�10s). After addition of ChIP dilution buffer, 50 �M of samplewas saved as input. Two micrograms of anti-p65 was added toa 900 �M precleared chromatin solution, and the sample wasincubated overnight. Antibody/DNA complexes were cap-tured, washed, eluted, and reverse cross-linked as per proto-col. Both the DNA and input fractions were purified byphenol/chloroform/iodoacetamide wash and ethanol pre-cipitation. The precipitated DNA was resuspended in 30 �l ofnuclease-free water, and quantitative PCR was performed on8 �l of sample as described above. Sample DNA was normal-ized to input DNA. Primers used were as follows: CX3CL1NF-�B forward, 5�-GGCATGTTC CCAGCTTGTGGCAGG-3�;and CX3CL1 NF-�B reverse, 5�-GTTGCCAA GGAACCAAGC-CGGC-3�.

Assay of NF-�B binding to oligonucleotides

DNA oligonucleotides containing the putative NF-�B site andmutated oligonucleotides were prepared by mixing equalamounts of 100 �M sense and antisense oligonucleotides withannealing buffer (10 mM PBS, pH 7.5; 50 mM NaCl; 0.1%Tween 20; and 2.7 mM KCl), incubating in a preheated block(95°C), and leaving the solution at room temperature for 60min. Streptavidin-coated microplates (Thermo Labsystems,Waltham, MA, USA) were immobilized with 0.25 �M biotin-ylated DNA oligonucleotides derived from the CX3CL1 pro-moter in annealing buffer (10 mM Tris-HCl, pH 7.5; 50 mMNaCl; and 1 mM EDTA) overnight at 4°C. The bindingreaction was performed by incubating nuclear extracts (20�g/20 �l) from A549 cells with 30 �l of binding buffer (4 mMHEPES, pH 7.5; 120 mM KCl; 8% glycerol; 1% BSA; 2 mMdithiothreitol; and 10 �g/ml herring sperm DNA) for 1 h atroom temperature in the DNA oligonucleotide-immobilizedmicroplates. In competitive binding experiments, nonbiotiny-lated wild-type (WT) (0.2 or 2 �M) or mutated (2 �M)oligonucleotides were added to the 30 �l of binding bufferand incubated with nuclear extracts for 20 min at room tempbefore addition to the streptavidin-coated plate. After wash-ing with washing buffer, DNA oligonucleotide-bound proteinwas detected with anti-p65 (1:1000; Santa Cruz BiotechnologyInc.) and a horseradish peroxidase-conjugated goat anti-rabbit secondary antibody (1:500; DakoCytomation A/S,Copenhagen, Denmark) diluted in buffer (10 mM PBS, pH7.5; 50 mM NaCl; and 0.1% Tween 20). The colorimetricreaction was performed with 100 �l of substrate reagent(R&D Systems Europe), stopped with 50 �l of stop solution (2N H2SO4), and measured at 450 nm with a referencewavelength of 550 nm. The oligonucleotides used in thesestudies were as follows: WT �B sense, 5�-ccagcctcccGGG-GAAGGTCCCagtatgac-3�; WT �B antisense, 5�-gtcatactgggac-ctcccgggaaggctgg-3�; mutated �B sense, 5�-ccagcctcccGGG-GAAGGAGCCagtatgac-3�; and mutated �B antisense, 5�-gtcatactggctccttccccgggaggctgg-3�.

Immunofluorescent staining protocol for p65

A549 cells were seeded at a concentration of 25,000 cells/wellin an eight-well chamber slide in complete medium. Afterserum deprivation for 24 h, cells were incubated in fresh(prewarmed, 37°C) serum-free DMEM containing cytokinesand drugs as indicted for each experiment; all glucocorticoids

were added 1 h before addition of cytokines. Permeabiliza-tion was performed with 0.1% (v/v) Tween 20 in PBS plus0.1% (w/v) BSA for 15 min at room temperature. Cells wereincubated with 5% normal donkey serum in the workingbuffer (PBS containing 0.1% saponin) for 20 min at roomtemperature and incubated overnight at 4°C with dilutedprimary antibody (anti-p65 antibody), then washed (3�5min) in buffer before incubation with rhodamine (tetrameth-ylrhodamine B isothiocyanate) -conjugated donkey anti-mouse IgG (Jackson ImmunoResearch Laboratories Inc.,West Grove, PA, USA) diluted 1:100 in medium containing5% donkey serum, and left in darkness for 45 min. Slides werethen washed thoroughly in PBS (5�5 min). After thoroughwashing with buffer, nuclei were counterstained with 4�,6-diamidino-2-phenylindole (DAPI) in Vectashield (VectorLaboratories Inc., Burlingame, CA, USA) mounting mediumand stored in the dark. Slides were visualized using confocalscanning laser microscopy. The images were obtained with aLeica confocal microscope equipped with a 488/514-nmdual-band argon ion laser (Leica Microsystems, MiltonKeynes, UK) using an oil-immersion objective, and imageswere collected by using TCSNT (Leica) software.

Statistics

Data are presented as means � se. Comparison betweengroups was performed using the Student’s t test. P 0.05 wastaken as significant.

RESULTS

Cytokine-induced CX3CL1

Stimulation of A549 cells with IL-1� or IFN� or TNF-�(all at 10 ng/ml) alone for 24 h resulted in 6.6-, 2.7-,and 7-fold increases in s-CX3CL1 release, respectively,compared with unstimulated controls (Fig. 1A). Co-stimulation with IFN� and TNF-� resulted in a syner-gistic 25-fold increase in s-CX3CL1 release at 24 h.Similarly, stimulation of A549 cells with IL-1�, IFN�, orTNF-� (all at 10 ng/ml) alone for 24 h resulted in 16-,7.5-, and 11-fold increases in mRNA abundance, respec-tively, compared with unstimulated controls (Fig. 1B).Costimulation with IFN� and TNF-� resulted in asynergistic 200-fold increase in mRNA abundance at24 h. A time course of IFN�/TNF-�-induced CX3CL1gene expression showed a maximal 350-fold increase inCX3CL1 mRNA abundance at 8 h before dropping offto 200-fold at 24 h (Fig. 1C).

Effect of dexamethasone on cytokine-inducedCX3CL1 expression and its stability

A549 cells were pretreated for 1 h with dexamethasone(105–1010 M) and then costimulated with IFN� andTNF-� (both at 10 ng/ml) for 24 h. An inhibitory effectof dexamethasone was observed in a dose-dependentmanner with maximal inhibition of 50% at 106 M (Fig.2A). To determine whether the attenuation of cytokine-induced CX3CL1 expression by dexamethasone wasmediated via GRs, A549 cells were pretreated with theGR antagonist, RU486 (106–1010 M), and dexameth-

1809GLUCOCORTICOIDS AND CX3CL1 EXPRESSION

asone (106 M) in the presence of IFN� and TNF-�(both at 10 ng/ml) for 24 h. RU486 antagonized thesuppressive effects of dexamethasone (Fig. 2B) on

IFN�/TNF-�-induced CX3CL1 expression in a dose-dependent manner with complete abrogation at 1 �M.

Dexamethasone attenuated mRNA abundance to20% of that of cytokine alone (Fig. 3A). To determinewhether the observed reduction in CX3CL1 gene ex-pression may be the result of a glucocorticoid-inducedchange in mRNA stability, we used actinomycin D, aninhibitor of RNA synthesis, to examine the rate ofmRNA degradation in the presence of dexamethasone.A549 cells were stimulated with IFN�/TNF-� (10 ng/mleach) in the presence and absence of dexamethasone(1 �M) for 3 h. The rate of CX3CL1 mRNA degrada-tion after the addition of actinomycin D (5 �M) wasthen observed by real-time quantitative PCR over a 4-hperiod. The half-life of CX3CL1 mRNA in cultured lungepithelial cells stimulated with IFN�/TNF-� was 90 min(Fig. 3B). There was no statistically significant effect ofdexamethasone on the stability of CX3CL1 mRNA.

Role of NF-�B pathway in cytokine-induced CX3CL1

To investigate the role of the NF-�B pathway in theregulation of CX3CL1, we used a small molecule inhib-itor of the I�B kinase 2 (IKK2), AS602868 (28). A549cells were pretreated for 1 h with AS602868 (0.5, 1.0,and 2.5 �M) in the presence of TNF-� and IFN� (bothat 10 ng/ml) for 24 h. A dose-dependent reduction inIFN�/TNF-�-induced CX3CL1 protein expression wasobserved with the greatest inhibition at 2.5 �M (Fig.4A). At all concentrations used, an 80% inhibition ofIFN�/TNF-�-induced CX3CL1 gene expression was ob-served (Fig. 4B). AS602868 had no effect on cellviability at the concentrations used as assessed by 3-(4,5-dimethyl-2-thizolyl)-2,5-diphenyl-2H-tetrazolium bro-mide assay (data not shown). Therefore, the IFN�/TNF-�-induced increase in CX3CL1 gene expressioninvolved NF-�B activation. The role of NF-�B was alsoexamined using a commercially available inhibitor ofthe NF-�B pathway, namely SC-514. This too showed adose-dependent reduction in IFN�/TNF-�-inducedCX3CL1 protein and mRNA expression (Fig. 4C, D). Aninhibitor of the extracellular signal-regulated kinase

Figure 1. Cytokine-induced CX3CL1 expression in lung epithelial cells. A549 cells were stimulated for 24 h with cytokines (10ng/ml). *P 0.05; **P 0.01; ***P 0.001, vs. unstimulated cells. A) Effect of cytokines on CX3CL1 release. B) mRNAabundance determined using quantitative reverse transcriptase-PCR. C) Time course of expression over 24 h. Unstim, controlunstimulated cells. n � 3 independent experiments.

Figure 2. Cytokine-induced CX3CL1 release is attenuated byglucocorticoids. A549 cells were pretreated with dexametha-sone (105–1010 M) alone (A) or together (B) with RU486(106–1010 M) for 1 h and stimulated with IFN�/TNF-� (10ng/ml each) for 24 h. A) *P 0.05 vs. cells treated withIFN�/TNF-�. B) *P 0.05, compared with cells treated withIFN�/TNF-�; #P 0.05, vs. cells pretreated with dexametha-sone and IFN�/TNF-�. n � 3 independent experiments.Unstim, control unstimulated cells; Dex, dexamethasonealone.

1810 Vol. 22 June 2008 BHAVSAR ET AL.The FASEB Journal

signal pathway (PD98059) did not attenuate IFN�/TNF-�-induced CX3CL1 protein and mRNA expression(data not shown).

A putative NF-�B binding site in the proximal pro-moter of the CX3CL1 gene at position 221/209 wasidentified using the AliBaba2.1 transcription factor bind-ing site search algorithm (http://www.gene-regulation.com/pub/programs/alibaba2). ChIP assays were usedto determine the ability of transcription factors to bindspecific chromosomal sites in living cells in response toenvironmental stimuli. We chose to use and optimizeChIP assays as they offer a measure of transcriptionfactor association at DNA binding sites in the presenceof native chromatin structure at a specific gene pro-moter. We therefore performed a ChIP assay to exam-ine the ability of endogenous p65 to be recruited to theCX3CL1 gene promoter in 1) the absence of stimuliand 2) under stimulation with cytokines, IFN�, andTNF-�, for 1 h (Fig. 5). For negative controls, weperformed ChIP assays examining recruitment ofNF-�B p65 to the 3�-untranslated region (UTR), which

lacks a NF-�B binding site and immunoprecipitationsin the absence of antibody or with normal mouse IgG.We found that there was minimal occupancy of theCX3CL1 promoter by p65 in the absence of stimuli.Stimulation with IFN�/TNF-� resulted in a 5-fold in-crease in NF-�B p65 recruitment to the CX3CL1 pro-moter.

Binding of p65 to the NF-�B site in vitro

To confirm that the binding of p65 to the NF-�B site isDNA sequence-specific we performed an ELISA-basedDNA binding assay with nuclear extracts prepared fromA549 cells. With use of a specific anti-p65 antibody, thisassay was able to detect proteins binding to immobi-lized CX3CL1 NF-�B (WT �B) oligonucleotides (Fig.6B). Stimulation with IFN� and TNF-� (10 ng/ml each)dramatically increased the binding of p65 to the �Boligonucleotides (P0.01). Addition of an excess offree �B oligonucleotides (WT �B) at 0.2 and 2 �M butnot the mutated �B oligonucleotides, specifically com-peted for this binding in a concentration-dependentmanner (P0.01). We also measured the effect ofdexamethasone on the p65 binding to the CX3CL1NF-�B site and show that the addition of dexametha-sone in the presence of IFN� and TNF-� does notinhibit the in vitro binding of p65 to the �B oligonucle-otides. Taken together, these data show that costimu-lation with TNF-� and IFN� induces p65 binding to the�B site and that this binding is DNA sequence specific.

Effect of dexamethasone on recruitment of NF-�B tothe CX3CL1 gene promoter

We performed ChIP assays to examine whether p65recruitment to the CX3CL1 gene promoter is attenu-ated by dexamethasone. The expression of granulocyte-macrophage colony-stimulating factor (GM-CSF), an-other NF-�B-dependent gene (29, 30), is similarlyattenuated by glucocorticoids, and we compared p65recruitment to the CX3CL1 gene promoter with that ofthe GM-CSF gene promoter. Cells were pretreated withdexamethasone (106 M) for 1 h and then stimulatedfor a further hour with either IFN�/TNF-� or IL-1� (10ng/ml each) for CX3CL1 and GM-CSF stimulation,respectively. Pretreatment with dexamethasone attenu-ated p65 recruitment to the CX3CL1 promoter by 80%,but no loss in recruitment was observed at the GM-CSFgene promoter (Fig. 7); pretreatment with anotherglucocorticoid, namely fluticasone propionate, simi-larly only reduced recruitment of p65 at the CX3CL1gene promoter (data not shown). We also performedcoimmunoprecipitation assays, after identical experi-mental conditions, and determined that, at 1 h post-stimulation, activated p65 and GR form a proteincomplex in A549 cells (data not shown).

Figure 3. Cytokine-induced CX3CL1 gene expression is atten-uated by glucocorticoids. A, B)A549 cells were pretreated withdexamethasone (Dex) (1 �M, 1 h) and treated with IFN�/TNF-� (10 ng/ml each) for 24 h (A) or 3 h (B). CX3CL1mRNA abundance was determined using quantitative reversetranscriptase-PCR. *P 0.05, vs. cells treated with IFN�/TNF-�. B) Assessment of CX3CL1 mRNA half-life. Actinomy-cin D (5 �g/ml) was added 3 h after IFN�/TNF-� (10 ng/mleach) stimulation. �, IFN�/TNF-�; Œ, IFN�/TNF-�/Dex.n � 4 independent experiments.

1811GLUCOCORTICOIDS AND CX3CL1 EXPRESSION

Effect of dexamethasone on p65 nucleartranslocation

To exclude the possibility that the reduction in p65recruitment to the CX3CL1 gene promoter, caused bydexamethasone, is a result of a reduction in p65translocation to the nucleus, we examined p65 translo-cation. A549 cells were stimulated with IFN�/TNF-�(10 ng/ml) in the absence and presence of dexameth-asone (106 M) for 1 h (Fig. 8). In unstimulated cells,the majority of the endogenous p65 is located withinthe cytoplasm (Fig. 8E) and was barely detectable in thenucleus (Fig. 8I). This pattern of p65 expression wasmirrored when the cells are exposed to dexamethasonealone (Fig. 8B, F, J). In cells stimulated by IFN�/TNF-�(Fig. 8C, G, K), the majority of endogenous p65 trans-

located to the nucleus (Fig. 8G, K). In cells pretreatedwith dexamethasone and IFN�/TNF-� (Fig. 8D, H, L),the majority of endogenous p65 remained translocatedto the nucleus (Fig. 8H, L).

DISCUSSION

We have shown that the proinflammatory cytokinesIL-1�, IFN�, and TNF-� induce CX3CL1 protein andgene expression (mRNA) in the respiratory A549 epi-thelial cell line and that this expression is synergisticallyenhanced when the cells are costimulated with IFN�and TNF-�. The increase in CX3CL1 protein accumu-lation is accompanied by an increase in mRNA abun-dance, indicating regulation at the level of gene tran-

Figure 4. Cytokine-induced CX3CL1 expression is mediated by the NF-�B pathway. A, B) Effect of AS602868 on IFN�/TNF-�-induced CX3CL1 release (A) and gene expression (B). CX3CL1 mRNA abundance was determined using quantitative reversetranscriptase-PCR. A549 cells were pretreated with the IKK2 kinase inhibitor, AS602868 (0.5, 1.0, and 2.5 �M) for 1 h beforestimulation with IFN�/TNF-� (10 ng/ml each) for 24 h. *P 0.05, vs. cells treated with IFN�/TNF-�. n � 3 independentexperiments. C, D) Effect of SC-514 on IFN�/TNF-�-induced CX3CL1 expression and release (C) and mRNA (D). n � 2independent experiments. Unstim, control unstimulated cells.

1812 Vol. 22 June 2008 BHAVSAR ET AL.The FASEB Journal

scription. A small molecule inhibitor of IKK2 kinase(AS602868) inhibited the cytokine-induced increase inCX3CL1 protein and mRNA. The human CX3CL1promoter contains a number of putative DNA bindingelements including ones for NF-�B. Using the chroma-tin immunoprecipitation assay, we demonstrate thatthe proinflammatory cytokines increase the recruit-ment of p65 to the endogenous CX3CL1 gene pro-moter, at the NF-�B DNA-binding site. An ELISA-basedDNA binding assay confirmed the ability of this NF-�BDNA binding site to bind p65. Cytokine-induced up-regulation of CX3CL1 expression in respiratory epithe-lial cells is therefore mediated through the NF-�Bpathway.

The expression of CX3CL1 protein and mRNA, inepithelial cells, can be attenuated by glucocorticoidswhereas in airway smooth muscle cells glucocorticoidspotentiate IFN�/TNF-�-induced CX3CL1 expression(16). Dexamethasone-induced suppression, in epithe-lial cells, is reversed by RU486, an antagonist of cytoso-lic steroid hormone receptors, confirming the effect ofdexamethasone through glucocorticoid receptor acti-vation. Our data with actinomycin D suggest that thedown-regulation of CX3CL1 gene expression in thepresence of dexamethasone is not due to an increase inthe rate of mRNA degradation. Although the meanvalues for percent mRNA levels at each time point (Fig.3B) are lower in the presence of dexamethasone (Œ),there is no statistically significant effect. Glucocorti-coids did not inhibit INF�/TNF-�-induced nucleartranslocation of p65 as measured by both nuclearstaining of p65 and by an in vitro DNA binding assay forp65; the former observation replicates the resultsshown in the study of Ito et al. (31), who also showed

that dexamethasone does not inhibit cytokine-inducedp65 translocation in A549 cells. However, attenuationof cytokine-induced CX3CL1 expression by glucocorti-coids was accompanied by a reduction in p65 recruit-ment to the CX3CL1 gene promoter, further illustrat-ing NF-�B-dependent gene regulation. In contrast, p65recruitment to the GM-CSF gene promoter, anotherNF-�B-dependent gene, is not similarly reduced in thepresence of glucocorticoids, indicating that the reduc-tion in recruitment can be specific to only certaingenes. An alternative interpretation of these data couldbe that the capacity of glucocorticoids to suppressNF-�B may differ according to the particular proinflam-matory cytokines used for stimulation. Nonetheless, weshow, at a native gene promoter, that glucocorticoidscan mediate their suppressive effects on the expressionof certain NF-�B-dependent genes, such as CX3CL1, byinterfering, by an as yet unidentified mechanism, withthe binding of the endogenous transcriptional activa-tors, namely NF-�B, at their gene promoters. Novac etal. (32) have also examined the recruitment of NF-�Bto the Fas ligand gene promoter in the presence ofdexamethasone. In this case, the NF-�B binding siteoverlaps a GRE and Novac et al. argue that dexameth-

Figure 5. ChIP assay to demonstrate that NF-�B is recruited tothe CX3CL1 gene promoter in response to IFN�/TNF-�.A549 cells were stimulated with IFN�/TNF-� (10 ng/mleach) for 1 h. Cells were then lysed, sonicated, and immuno-precipitated with anti-p65 antibody. Quantitative real-timePCR was then performed on the CX3CL1 promoter. *P 0.05, compared with unstimulated cells. n � 3 independentexperiments. Unstim, control unstimulated cells; No Ab,control for nonspecific DNA immunoprecipitation in theabsence of the p65 antibody; 3�-UTR, control for nonspecificDNA immunoprecipitation by the p65 antibody from theregion of the gene that does not contain an NF-�B bindingsite.

Figure 6. Binding of p65 to the NF-�B DNA binding site onthe fractalkine gene promoter. A) Sequences of the WT andmutated DNA oligonucleotides used in the NF-�B promoterbinding assay (only sense oligonucleotides are shown). TheNF-�B site is shown in uppercase and the mutation inunderlined. B) A549 cells were treated with IFN�/TNF-� (IT)(10 ng/ml each) in the presence and absence of dexameth-asone (106 M) for 1 h. Nuclear extracts were prepared andincubated with or without competitor DNA oligonucleotides[WT �B, WT �B, mutated (Mut) �B, and Mut �B oligonucle-otides] in 96-well plates immobilized with WT �B oligonucle-otides and were detected by an p65-specific antibody. Resultsare expressed as mean � se OD450 nm. **P 0.01, vs.unstimulated; ##P 0.01, vs. IFN�/TNF-� alone. n � 3independent experiments.

1813GLUCOCORTICOIDS AND CX3CL1 EXPRESSION

asone reduces NF-�B recruitment through sterical oc-clusion of p65 by GR. However, this mechanism doesnot apply to the CX3CL1 gene as the NF-�B does notoverlap a GRE.

The attenuation of NF-�B-dependent gene expres-sion by glucocorticoids in airway and alveolar epithelialcells has been thought to involve repressive mecha-nisms including NF-�B. One hypothesis centeredaround cytoplasmic events related to an increase in theexpression of I�B� by glucocorticoids (33, 34). How-ever, glucocorticoid treatment of endothelial cellschanged neither the expression of I�B� nor the NF-�Bactivation (35). In A549 cells, no substantial change incytokine-induced I�B� expression by dexamethasonewas observed, thereby excluding the I�B�-dependentmechanisms of GR-mediated repression. Furthermore,dexamethasone had no significant effect on the expres-sion of p65 (36). An alternative hypothesis focused onwhether glucocorticoids blocked NF-�B DNA binding.Two studies investigated this possibility by using the invitro electrophoretic gel mobility shift assays with nu-clear extracts from A549 cells. Nissen and Yamamoto(2) determined, with the IL-8 NF-�B binding site asprobe, that TNF-�-stimulated A549 cells produced areadily detectable p65/p50-DNA complex that was notdiminished by dexamethasone. Newton et al. (36) sim-

Figure 7. Glucocorticoids perturb the recruitment of NF-�Bto the native CX3CL1 gene promoter. A549 cells were stimu-lated with IFN�/TNF-� or IL-1� (10 ng/ml each) for 1 h inthe presence and absence of dexamethasone (106 M, 1-hpretreatment). Chromatin was immunoprecipitated with anti-p65 antibody, and quantitative PCR was performed for theCX3CL1 and GM-CSF promoters. *P 0.05, vs. unstimulatedcells. n � 3 independent experiments. No Ab, control fornonspecific DNA immunoprecipitation in the absence of thep65 antibody.

Figure 8. Immunofluorescence confocal microscopy of p65 nuclear translocation. A549 cells were treated with an anti-p65antibody for localization of endogenous p65 (red fluorescence). Nuclei were counterstained with DAPI (blue fluorescence,A–D) and slides were visualized using confocal microscopy. Cells were unstimulated (A, E, I) or exposed to dexamethasone(106 M; B, F, J), to IFN�/TNF-� (10 ng/ml each; C, G, K), or to IFN�/TNF-� and dexamethasone (D, H, L). Merged imagesof the red and blue fluorescence are shown (I–L). Original images �800 for all panels:. The images are representative of threeindependent preparations.

1814 Vol. 22 June 2008 BHAVSAR ET AL.The FASEB Journal

ilarly concluded that dexamethasone did not affect thebinding of NF-�B to its cognate binding site duringcorticosteroid-regulated cytokine suppression. Theyalso suggested that cytoplasmic interactions betweentranscription factors such as GRs with NF-�B mayaccount for a reduction in NF-�B translocation to thenucleus, thus leading to a reduction or repression ofNF-�B-dependent transcription (36). However, ourdata indicate that glucocorticoids do not hinder thetranslocation into the nucleus of cytokine-activated p65(Fig. 6). A further mechanism for the repression ofNF-�B-mediated gene activation may occur throughdisruption of the direct protein-protein interaction ofp65 with the basal transcriptional machinery (3), pos-sibly through the recruitment of corepressors, such ashistone deacetylases (HDACs) (37). Overall, it is stillapparent that the precise steps through which glu-cocorticoids mediate their suppressive effects on genetranscription in inflammatory diseases, such as asthma,needs further investigation.

The novelty of our findings is most readily illustratedby comparing, through the use of chromatin immuno-precipitation assays, the effects of dexamethasone onthe recruitment of native NF-�B to the endogenousIL-8 and CX3CL1 gene promoters. Nissen andYamamoto (2) have shown a 5-fold increase in therecruitment of NF-�B to the IL-8 gene promoter inresponse to TNF-� stimulation in A549 cells. Cotreat-ment with dexamethasone and TNF-� also yielded a5-fold enrichment, that is, no reduction in the recruit-ment of p65. Similarly, in the study by Garside et al. (38)performed in HeLa cells, the 6-fold increase in NF-�Brecruitment to the IL-8 gene promoter in response toTNF-� was not reversed by dexamethasone. In thesestudies, glucocorticoids did not mediate their suppres-sive effects through the displacement of NF-�B fromthe IL-8 gene promoter. In contrast, we show thepossibility that glucocorticoids can mediate its suppres-sive effects by disrupting the binding of NF-�B, asobserved at the CX3CL1 gene promoter. The reducedbinding affinity of NF-�B at the CX3CL1 gene promoterwas not due to a reduction in NF-�B availability in thenucleus because the amount of p65 translocation afterIFN�/TNF-� was not affected by dexamethasone.Rather, the displacement of NF-�B from the CX3CL1gene promoter may be a promoter-specific phenome-non, because such a loss of NF-�B DNA binding is notseen at either the native IL-8 or GM-CSF gene promot-ers. This observation is likely to be a result of theNF-�B/DNA complex formation at each site. A study byMenetski (39) has shown that the apparent DNA bind-ing affinity of a specific NF-�B binding site at eachpromoter may be a function of its nucleotide sequenceand can determine which NF-�B dimer can mostreadily bind to that site. In turn, this is dependent onthe NF-�B/Rel protein subunit composition (p50, p52,p65, c-Rel, and RelB) (2) of the NF-�B dimer itself,which in turn affects the stability of the NF-�B/DNAcomplex. It is probable therefore that dexamethasone,through GR activation, differentially affects the binding

characteristics of NF-�B, depending on both the nucle-otide sequence of the binding site and the compositionof the NF-�B dimer at a given gene promoter. To thisend, we note that the CX3CL1 and IL-8 NF-�B bindingsites have differing core nucleotide sequences. Theexact mechanism leading to dexamethasone-induceddisplacement of NF-�B has yet to be elucidated. Incontrast, the recent study by Ito et al. (37) providesevidence for a transrepression mechanism with occu-pancy of NF-�B at the native GM-CSF promoter wheredexamethasone increases HDAC2 association with thep65/NF-�B complex, which in turn leads to suppres-sion of the GM-CSF gene.

In summary, we have shown that cytokine-inducedexpression of CX3CL1 in A549 respiratory epithelialcells is mediated through the NF-�B pathway, requiringrecruitment of endogenous p65 (NF-�B) to the nativeCX3CL1 gene promoter. Attenuation of cytokine-in-duced CX3CL1 expression by dexamethasone is accom-panied by a reduction in recruitment of p65 to theCX3CL1 gene promoter. This glucocorticoid mecha-nism may be specific to the CX3CL1 gene or mayhighlight a more general mode of action of glucocor-ticoids wherein they mediate their suppressive effectsthrough the displacement of p65 from the promotersof NF-�B-dependent genes.

This study was funded by an Imperial College Trust Fund.We thank Professor Ian Adcock for invaluable discussionsduring the preparation of this manuscript and Mr. Charalam-bos Michaeloudes for technical assistance.

REFERENCES

1. Rhen, T., and Cidlowski, J. A. (2005) Antiinflammatory action ofglucocorticoids—new mechanisms for old drugs. N. Engl. J. Med.353, 1711–1723

2. Nissen, R. M., and Yamamoto, K. R. (2000) The glucocorticoidreceptor inhibits NF�B by interfering with serine-2 phosphory-lation of the RNA polymerase II carboxy-terminal domain. GenesDev. 14, 2314–2329

3. De Bosscher, K., Vanden Berghe, W., Vermeulen, L., Plaisance,S., Boone, E., and Haegeman, G. (2000) Glucocorticoids repressNF-�B-driven genes by disturbing the interaction of p65 with thebasal transcription machinery, irrespective of coactivator levelsin the cell. Proc. Natl. Acad. Sci. U. S. A. 97, 3919–3924

4. Abbinante-Nissen, J. M., Simpson, L. G., and Leikauf, G. D.(1995) Corticosteroids increase secretory leukocyte proteaseinhibitor transcript levels in airway epithelial cells. Am. J. Physiol.268, L601–L606

5. Lasa, M., Abraham, S. M., Boucheron, C., Saklatvala, J., andClark, A. R. (2002) Dexamethasone causes sustained expressionof mitogen-activated protein kinase (MAPK) phosphatase 1 andphosphatase-mediated inhibition of MAPK p38. Mol. Cell. Biol.22, 7802–7811

6. Polito, A. J., and Proud, D. (1998) Epithelia cells as regulators ofairway inflammation. J. Allergy Clin. Immunol. 102, 714–718

7. El Shazly, A., Berger, P., Girodet, P. O., Ousova, O., Fayon, M.,Vernejoux, J. M., Marthan, R., and Tunon-de-Lara, J. M. (2006)Fraktalkine produced by airway smooth muscle cells contributesto mast cell recruitment in asthma. J. Immunol. 176, 1860–1868

8. Rimaniol, A. C., Till, S. J., Garcia, G., Capel, F., Godot, V.,Balabanian, K., Durand-Gasselin, I., Varga, E. M., Simonneau,G., Emilie, D., Durham, S. R., and Humbert, M. (2003) TheCX3C chemokine fractalkine in allergic asthma and rhinitis. J.Allergy Clin. Immunol. 112, 1139–1146

1815GLUCOCORTICOIDS AND CX3CL1 EXPRESSION

9. Ning, W., Li, C. J., Kaminski, N., Feghali-Bostwick, C. A., Alber,S. M., Di, Y. P., Otterbein, S. L., Song, R., Hayashi, S., Zhou, Z.,Pinsky, D. J., Watkins, S. C., Pilewski, J. M., Sciurba, F. C., Peters,D. G., Hogg, J. C., and Choi, A. M. (2004) Comprehensive geneexpression profiles reveal pathways related to the pathogenesisof chronic obstructive pulmonary disease. Proc. Natl. Acad. Sci.U. S. A. 101, 14895–14900

10. Murphy, P. M., Baggiolini, M., Charo, I. F., Hebert, C. A.,Horuk, R., Matsushima, K., Miller, L. H., Oppenheim, J. J., andPower, C. A. (2000) International union of pharmacology.XXII. Nomenclature for chemokine receptors. Pharmacol. Rev.52, 145–176

11. Garton, K. J., Gough, P. J., Blobel, C. P., Murphy, G., Greaves,D. R., Dempsey, P. J., and Raines, E. W. (2001) Tumor necrosisfactor-�-converting enzyme (ADAM17) mediates the cleavageand shedding of fractalkine (CX3CL1). J. Biol. Chem. 276,37993–38001

12. Imaizumi, T., Matsumiya, T., Fujimoto, K., Okamoto, K., Cui, X.,Ohtaki, U., Hidemi, Y., and Satoh, K. (2000) Interferon-�stimulates the expression of CX3CL1/fractalkine in culturedhuman endothelial cells. Tohoku J. Exp. Med. 192, 127–139

13. Chen, Y. M., Tu, C. J., Hung, K. Y., Wu, K. D., Tsai, T. J., andHsieh, B. S. (2003) Inhibition by pentoxifylline of TNF-�-stimulated fractalkine production in vascular smooth musclecells: evidence for mediation by NF-�B down-regulation. Br. J.Pharmacol. 138, 950–958

14. Fahy, O. L., Coates, N. J., and McColl, S. R. (2003) Inhibition ofcytokine-induced fractalkine production by bacterial invasion ofhuman-dermal fibroblasts. Lab. Invest. 83, 721–730

15. Yoshida, H., Imaizumi, T., Fujimoto, K., Matsuo, N., Kimura, K.,Cui, X., Matsumiya, T., Tanji, K., Shibata, T., Tamo, W.,Kumagai, M., and Satoh, K. (2001) Synergistic stimulation, bytumor necrosis factor-� and interferon-�, of fractalkine expres-sion in human astrocytes. Neurosci. Lett. 303, 132–136

16. Sukkar, M. B., Issa, R., Xie, S., Oltmanns, U., Newton, R., andChung, K. F. (2004) Fractalkine/CX3CL1 production by humanairway smooth muscle cells: induction by IFN-� and TNF-� andregulation by TGF-� and corticosteroids. Am. J. Physiol. 287,L1230–L1240

17. Adcock, I. M. (2001) Glucocorticoid-regulated transcriptionfactors. Pulm. Pharmacol. Ther. 14, 211–219

18. Li, Q., and Verma, I. M. (2002) NF-�B regulation in the immunesystem. Nat. Rev. Immunol. 2, 725–734

19. Barker, J. N., Sarma, V., Mitra, R. S., Dixit, V. M., and Nickoloff,B. J. (1990) Marked synergism between tumor necrosis factor-�and interferon-� in regulation of keratinocyte-derived adhesionmolecules and chemotactic factors. J. Clin. Invest. 85, 605–608

20. Johnson, D. R., and Pober, J. S. (1994) HLA class I heavy-chaingene promoter elements mediating synergy between tumornecrosis factor and interferons. Mol. Cell. Biol. 14, 1322–1332

21. Levy, D. E., Lew, D. J., Decker, T., Kessler, D. S., and Darnell,J. E., Jr. (1990) Synergistic interaction between interferon-� andinterferon-� through induced synthesis of one subunit of thetranscription factor ISGF3. EMBO J. 9, 1105–1111

22. Baeuerle, P. A., and Henkel, T. (1994) Function and activationof NF-�B in the immune system. Annu. Rev. Immunol. 12,141–179

23. Ohmori, Y., and Hamilton, T. A. (1995) The interferon-stimu-lated response element and a �B site mediate synergistic induc-tion of murine IP-10 gene transcription by IFN-� and TNF-�.J. Immunol. 154, 5235–5244

24. Jahnke, A., and Johnson, J. P. (1994) Synergistic activation ofintercellular adhesion molecule 1 (ICAM-1) by TNF-� andIFN-� is mediated by p65/p50 and p65/c-Rel and interferon-responsive factor Stat1� (p91) that can be activated by bothIFN-� and IFN-�. FEBS Lett. 354, 220–226

25. Lee, A. H., Hong, J. H., and Seo, Y. S. (2000) Tumour necrosisfactor-� and interferon-� synergistically activate the RANTESpromoter through nuclear factor �B and interferon regulatoryfactor 1 (IRF-1) transcription factors. Biochem. J. 350 (Pt 1),131–138

26. Imanishi, D., Yamamoto, K., Tsushima, H., Miyazaki, Y.,Kuriyama, K., Tomonaga, M., and Matsuyama, T. (2000) Iden-tification of a novel cytokine response element in the humanIFN regulatory factor-1 gene promoter. J. Immunol. 165, 3907–3916

27. Ray, A., and Prefontaine, K. E. (1994) Physical association andfunctional antagonism between the p65 subunit of transcriptionfactor NF-�B and the glucocorticoid receptor. Proc. Natl. Acad.Sci. U. S. A. 91, 752–756

28. Frelin, C., Imbert, V., Griessinger, E., Loubat, A., Dreano, M.,and Peyron, J. F. (2003) AS602868, a pharmacological inhibitorof IKK2, reveals the apoptotic potential of TNF-� in Jurkatleukemic cells. Oncogene 22, 8187–8194

29. Kochetkova, M., and Shannon, M. F. (1996) DNA triplexformation selectively inhibits granulocyte-macrophage colony-stimulating factor gene expression in human T cells. J. Biol.Chem. 271, 14438–14444

30. Catley, M. C., Chivers, J. E., Holden, N. S., Barnes, P. J., andNewton, R. (2005) Validation of IKK� as therapeutic target inairway inflammatory disease by adenoviral-mediated delivery ofdominant-negative IKK� to pulmonary epithelial cells. Br. J.Pharmacol. 145, 114–122

31. Ito, K., Barnes, P. J., and Adcock, I. M. (2000) Glucocorticoidreceptor recruitment of histone deacetylase 2 inhibits interleu-kin-1�-induced histone H4 acetylation on lysines 8 and 12. Mol.Cell. Biol. 20, 6891–6903

32. Novac, N., Baus, D., Dostert, A., and Heinzel, T. (2006) Com-petition between glucocorticoid receptor and NF�B for controlof the human FasL promoter. FASEB J. 20, 1074–1081

33. Scheinman, R. I., Cogswell, P. C., Lofquist, A. K., and Baldwin,A. S., Jr. (1995) Role of transcriptional activation of I�B� inmediation of immunosuppression by glucocorticoids. Science270, 283–286

34. Auphan, N., DiDonato, J. A., Rosette, C., Helmberg, A., andKarin, M. (1995) Immunosuppression by glucocorticoids: inhi-bition of NF-�B activity through induction of I�B synthesis.Science 270, 286–290

35. Brostjan, C., Anrather, J., Csizmadia, V., Stroka, D., Soares, M.,Bach, F. H., and Winkler, H. (1996) Glucocorticoid-mediatedrepression of NF�B activity in endothelial cells does not involveinduction of I�B� synthesis. J. Biol. Chem. 271, 19612–19616

36. Newton, R., Hart, L. A., Stevens, D. A., Bergmann, M., Donnelly,L. E., Adcock, I. M., and Barnes, P. J. (1998) Effect of dexa-methasone on interleukin-1�-(IL-1�)-induced nuclear factor-�B(NF-�B) and �B-dependent transcription in epithelial cells. Eur.J. Biochem. 254, 81–89

37. Ito, K., Yamamura, S., Essilfie-Quaye, S., Cosio, B., Ito, M.,Barnes, P. J., and Adcock, I. M. (2006) Histone deacetylase2-mediated deacetylation of the glucocorticoid receptor enablesNF-�B suppression. J. Exp. Med. 203, 7–13

38. Garside, H., Stevens, A., Farrow, S., Normand, C., Houle, B.,Berry, A., Maschera, B., and Ray, D. (2004) Glucocorticoidligands specify different interactions with NF-�B by allostericeffects on the glucocorticoid receptor DNA binding domain.J. Biol. Chem. 279, 50050–50059

39. Menetski, J. P. (2000) The structure of the nuclear factor-�Bprotein-DNA complex varies with DNA-binding site sequence.J. Biol. Chem. 275, 7619–7625

Received for publication July 30, 2007.Accepted for publication December 27, 2007.

1816 Vol. 22 June 2008 BHAVSAR ET AL.The FASEB Journal