RESEARCHDecreased expression of thspSheridan et al. Respiratory Research (2015) 16:54 DOI 10.1186/s12931-015-0214-6in response to smoke, including lymphocytes, epithelialMontreal, QC H4A 3J1, CanadaFull list of author information is available at the end of the articleKeywords: Inflammation, NF-κB, miRNA, COPD, Cigarette smoke, LungIntroductionChronic obstructive pulmonary disease (COPD) is a het-erogeneous disease associated with an enhanced, chronicinflammatory response due to exposure to noxious parti-cles or gases. Cigarette smoke is the single greatest riskfactor for developing COPD, with an estimated 80-90% ofCOPD cases being due to chronic smoke exposure [1,2].Cigarette smoke contributes to COPD by inciting inflam-mation, recruiting T cells, macrophages and neutrophils tothe lung in part via the induction of inflammatory media-tors, including cyclooxygenase-2 (COX-2) [3-5]. COX-2catalyzes the transformation of arachidonic acid (AA) intothromboxanes and prostaglandins (PG) such as PGE2, animmunoregulatory PG that is elevated in COPD subjects[5,6]. Inhibition of COX-2-derived PGE2 also protectsagainst the development of emphysema [7], supporting arole for COX-2 in the pathobiology of COPD. Numerouscell types within the lung are capable of producing COX-2* Correspondence: carolyn.baglole@mcgill.ca1Department of Medicine, 1001 Decarie Blvd, Montreal, QC H4A 3J1, Canada2Research Institute of the McGill University Health Centre, 1001 Decarie Blvd,RelB may be an important determinant in the aberrant, heigJared A Sheridan , Michela Zago , Parameswaran Nair , Pei Z Li , Jean Bourbeau , Wan C Tan , Qutayba Hamid ,David H Eidelman1,2, Andrea L Benedetti3 and Carolyn J Baglole1,2*AbstractBackground: Heightened inflammation, including expression of COX-2, is associated with COPD pathogenesis. RelBis an NF-κB family member that attenuates COX-2 in response to cigarette smoke by a mechanism that may involvethe miRNA miR-146a. There is no information on the expression of RelB in COPD or if RelB prevents COX-2 expressionthrough miR-146a.Methods: RelB, Cox-2 and miR-146a levels were evaluated in lung fibroblasts and blood samples derived fromnon-smokers (Normal) and smokers (At Risk) with and without COPD by qRT-PCR. RelB and COX-2 protein levelswere evaluated by western blot. Human lung fibroblasts from Normal subjects and smokers with and withoutCOPD, along with RelB knock-down (siRNA) in Normal cells, were exposed to cigarette smoke extract (CSE)in vitro and COX-2 mRNA/protein and miR-146a levels assessed.Results: Basal expression of RelB mRNA and protein were significantly lower in lung cells derived from smokerswith and without COPD, the latter of which expressed more Cox-2 mRNA and protein in response to CSE.Knock-down of RelB in Normal fibroblasts increased Cox-2 mRNA and protein induction by CSE. Basal miR-146alevels were not different between the three groups, and only Normal fibroblasts increased miR-146a expressionin response to smoke. There was a positive correlation between systemic RelB and Cox-2 mRNA levels and circulatingmiR-146a levels were higher only in GOLD stage I subjects.Conclusions: Our data indicate that RelB attenuates COX-2 expression in lung structural cells, such that loss of pulmonaryhtened inflammation associated with COPD pathogenesis.RelB in lung fibroblasts frowithout COPD potentiatesmoke-induced COX-2 ex1 1 5© 2015 Sheridan et al.; licensee BioMed CentrCommons Attribution License (http://creativecreproduction in any medium, provided the orDedication waiver (http://creativecommons.orunless otherwise stated.Open Accesse NF-κB family memberm Smokers with andcigaretteression4 4 6 1,2al. This is an Open Access article distributed under the terms of the Creativeommons.org/licenses/by/4.0), which permits unrestricted use, distribution, andiginal work is properly credited. The Creative Commons Public Domaing/publicdomain/zero/1.0/) applies to the data made available in this article,associated with the expression of Cox-2 is also not known.Sheridan et al. Respiratory Research (2015) 16:54 Page 2 of 16cells, smooth muscle cells and fibroblasts [4,8,9]. Althoughthe primary role of the fibroblast is to synthesize andmaintain extracellular matrix (ECM), thereby providingstructure and support to the lung, there is now ample evi-dence that fibroblasts contribute to chronic inflammationby producing an array of cytokines, chemokines, lipid me-diators and proteases [10-12]. Fibroblasts are among thefirst cells in the lung which encounter cigarette smoke, po-tently activating them to increase the production of IL-8,COX-2 and other mediators involved in leukocyte recruit-ment [4,13,14]. Alveolar fibroblasts also provide connec-tion between type II pneumocytes and endothelial cells,thereby providing migratory leukocytes with a directionalconduit [15]. Thus, disordered fibroblast behaviour mayregulate the switch to chronic, persistent inflammation inpart by their ability to promote the recruitment, survivaland retention of leukocytes and other immune cells to tar-get organs such as the lung due to chronic smoke expos-ure [10,16,17].Lung fibroblasts derived from patients with COPD havehigh intercellular adhesion molecule-1 (ICAM-1) expres-sion and secrete more IL-6, IL-8, MMP-9 and PGE2[2,16,18], the latter being due to increased COX-2 proteinexpression [19]. The induction of COX-2 by cigarettesmoke is regulated at least partially by the transcription fac-tor nuclear factor-κB (NF-κB) [4]. Activation of the canon-ical NF-κβ pathway- typically composed of RelA and p50subunits- increases gene transcription and the expressionof inflammatory proteins including IL-6, IL-8 and COX-2.However, activation of this NF-κB pathway also increasesRelB expression [20], the key protein in the alternativeNF-κB pathway. RelB is sequestered in the cytosol by theinhibitory NF-κB protein p100. Signal-specific processing ofp100 by the NF-κB-inducing kinase (NIK) produces p52,which together with RelB translocate to the nucleus toregulate gene expression [21]. Regulation of the alternativeNF-κB pathway during inflammatory conditions (TNF-α,LPS, cigarette smoke, etc.) can occur via stabilization ofNIK, which increases NIK expression to allow efficient pro-cessing of p100 to p52 [22]. Inflammatory stimuli can alsolead to the upregulation of RelB expression, enhanced nu-clear localization [20,23,24] or processing/cleavage of RelBprotein [25,26]. Many of these may contribute to the abilityof RelB to modulate chronic inflammation. In this regard,RelB prevents persistent non-infectious inflammation inthe liver and lung, a phenomena attributed to the suppres-sive abilities of RelB in non-lymphoid tissue, possibly fibro-blasts [27,28]. We have shown that RelB dampens cigarettesmoke-induced pulmonary inflammation both in vitro andin vivo, including the expression of COX-2 [29,30]. COX-2is over-expressed by COPD lung fibroblasts [19] and wehave shown that RelB protein is degraded by cigarettesmoke [29], rendering it possible that heightened smoke-induced COX-2 expression in COPD is associated withWe recently published that in murine lung fibroblasts,RelB suppression of cigarette smoke-induced COX-2 ex-pression is due to up-regulation of the microRNA-146a(miR-146a) [31]. miRNAs are single-stranded, non-coding,22 nucleotide-long RNA which act posttranscriptionally toinhibit protein expression [32] and are of increasing inter-est as biomarkers for COPD [33]. Cigarette smoke alterslung structural cell and circulating miRNA levels, includ-ing miR-146a [31,34,35], an anti-inflammatory miRNAthat under-expressed in cytokine-stimulated lung fibro-blasts derived from COPD patients, which ultimately re-sults in higher COX-2 expression [36]. We postulate thataltered RelB expression in COPD-derived lung fibroblastsfacilitates cigarette smoke-induced COX-2 due to dysregu-lation of miR-146a expression. Given that there is highcorrelation of expression between immune cells and circu-lating miRNA expression levels [37] and systemic expres-sion miR-146a are now indicted in several diseasesincluding rheumatoid arthritis [37], we also postulatedthat miR-146a levels would be associated with clinical fea-tures of COPD in relation to changes in circulating RelBexpression.Therefore, we first sought to investigate whether down-regulation of RelB expression by cigarette smoke renderslung cells unable to increase miR-146a levels, thereby po-tentiating the induction of COX-2. Herein we show for thefirst time that fibroblasts derived from smokers with andwithout COPD have significantly lower RelB mRNA andprotein expression compared to Normal (non-smoker) fi-broblasts. RelB expression decreases in Normal fibroblastsexposed to cigarette smoke in vitro, and siRNA knock-down of RelB potentiates Cox-2 mRNA and protein induc-tion by cigarette smoke in Normal fibroblasts. Utilizingblood samples from subjects participating in the CanadianChronic Obstructive Lung Disease (CanCOLD) platform[38], we found that there was a significant positive correl-ation between systemic RelB and Cox-2 mRNA expression.Collectively, our results highlight the differential regulationof RelB expression between lung structural cells and per-ipheral blood and suggest that dysregulation of RelB levelswithin pulmonary structural cells may be a contributingfactor in the heightened inflammatory response that ischaracteristic of individuals who smoke.Materials and methodsMaterialsAll chemicals were purchased from Sigma (St. Louis,absent/low RelB expression. Moreover, whether systemicRelB expression is also altered as a consequence of chronicsmoke exposure or COPD severity (i.e. GOLD stage) or isMO) except MG-132 which was from Tocris Bioscience(Minneapolis, MN).Cell cultureLung tissue was obtained from individuals undergoing lungresection surgery at McMaster University. Recruited indi-viduals included those with COPD, subjects without COPDbut who were current or former smokers (At Risk) or non-smokers without COPD (Normal). The clinical features ofthe subjects from which the lung fibroblasts were derivedare given in Table 1. This study was approved by the Re-search Ethics Board of St Joseph’s Healthcare Hamilton andall patients gave written, informed consent. Primary lung fi-broblasts were cultured as previously described [39] andonly tissue from cancer-free regions was used for the deriv-ation of fibroblasts. Prior to experimentation, fibroblastsSheridan et al. Respiratory Research (2015) 16:54 Page 3 of 16were characterized based on morphology and vimentinexpression as well as the absence of cytokeratin (epithelialcell marker), desmin (muscle cell marker) and α-smoothmuscle actin (α-SMA; myofibroblast marker) [39] (Figure 1).Following characterization, cells were expanded and eitherfrozen in liquid nitrogen or maintained in culture as amonolayer. For experimentation, primary fibroblasts werecultured in MEM (Life Technologies, Gaithersburg, MD)supplemented with 10% fetal bovine serum (FBS; HyCloneLaboratories, Logan, UT) and incubated in humidified 5%CO2/95% air at 37°C. All fibroblast strains were used at theearliest possible passage. For assessment of basal expressionlevels, all available fibroblast strains were cultured and ana-lyzed at the same time and were within 1 passage (passage3–4). Additional experiments were conducted with fibro-blasts from a minimum of three different individuals ofeach patient category.Canadian chronic obstructive lung disease (CanCOLD)CanCOLD is a prospective longitudinal cohort studytracking 1800 subjects and comprises 2 COPD subsets(GOLD ≥2 and GOLD 1) and 2 subsets of non-COPDpeers, i.e., normal post bronchodilator spirometry (eversmoker for those at-risk and never-smoker for the healthycontrols), matched for sex and age. The CanCOLD cohortTable 1 Patient characteristics of fibroblast strainsderived from lung resectionNormal At Risk COPDn of subjects 6 15 12Age (yr) 66.7 ± 2.6 64 ± 0.7 68.3 ± 0.9Gender (M/F) (2/4) (8/7) (8/4)Smoking# 0.0 36.5 ± 1.2 37.6 ± 1.3FEV1 (%) 88.2 ± 3.3 88.2 ± 0.9 65.8 ± 2.16FVC (%) 83 ± 3.4 92 ± 0.9 81.4 ± 2.3FEV1/FVC (%) 88 ± 2.7 76.6 ± 0.5 55.0 ± 1.7***Results presented as average ± SEM.#Denotes pack-years.***FEV1/FVC of COPD patients was significantly lower (p < 0.05) compared toeither the non-smokers or the smokers without COPD, which did not differfrom each other.is described in detail in [38]. The CanCOLD study was ap-proved by the REB at the McGill University Health Centre(MUHC) - Study # 09-025-BMC. Peripheral blood wascollected using PAXgene blood RNA tubes (PreAnalytiXGmbH, Hombrechtikon, Germany) at initial visit and fro-zen at −80°C until analysis. Clinical characteristics of thesubjects utilized in this part of the study are in Table 2.Preparation of cigarette smoke extract (CSE)Research grade cigarettes (3R4F) with a filter were obtainedfrom the Kentucky Tobacco Research Council (Lexington,KT) and CSE generated as previously described [29,40-42].Briefly, CSE was prepared by bubbling smoke from two cig-arettes into 20 ml of serum-free MEM, the pH adjusted to7.4, sterile- filtered with a 0.45-μm filter (25-mm Acrodisc;Pall Corp., Ann Arbor, MI) and was used within 30 minutesof preparation. An optical density of 0.65 (320 nm) wasconsidered to represent 100% CSE and was diluted inserum-free MEM to 2% CSE, a concentration which po-tently increases COX-2 expression in primary lung fibro-blasts deficient in RelB expression [31]. For comparison ofthe cigarette smoke response between the three patientgroups, experiments were conducted with fibroblasts cul-tured at the same time utilizing the same CSE preparationto minimize experimental variability.Analysis of gene expressionTotal RNA was isolated from media- or CSE-treated fi-broblasts using a Qiagen miRNeasy kit (Qiagen Inc.,Valencia, CA). For processing the CanCOLD samples,total RNA was isolated using PAXgene blood RNA kit(PreAnalytiX GmbH) according to the supplier’s proto-col. In each case, RNA was eluted in 30 μl RNase-freewater and RNA content and purity was measured usinga Nanodrop 1000 spectrophotometer (Thermo FisherScientific, Wilmington, DE). Reverse transcription oftotal RNA was carried out in a 20-μl reaction mixtureby iScript IITM Reverse Transcription Supermix (Bio-Rad Laboratories, Mississauga, Ontario) at 25°C for5 min, at 42°C for 30 min and at 85°C for 5 min. Realtime (qPCR) was performed with 1 μl cDNA and0.5 μM primers added in SsofastTM Eva Green® Super-mix (Bio-Rad) and PCR amplification was performedusing a CFX96 Real-Time PCR Detection System (Bio-Rad). Primer sequences for human Cox-2 are TCA-CAGGCTTCCATTGACCAG (f ) and CCGAGGCTTTTCTACCAGA (r) and for human RelB are TGTGGTGAGGATCTGCTTCCAG (f ) and GGCCCGCTTTCCTTGTTAATTC (r). Thermal cycling was initiated at95°C for 30 sec and followed by 40 cycles of denatur-ation at 95°C for 30 s and annealing for 5 s. Melt curveswere obtained to ensure that nonspecific products wereabsent. The fluorescence detection threshold was setabove the non-template control background within theFigure 1 Characterization of human lung fibroblasts. Representative immunofluorescence images of human lung fibroblasts derived from Normal(non-smoker), At Risk (Smoker without COPD) and COPD subjects. Target proteins were visualized with fluorescein (green) in combination withHoechst (blue) nuclear stain; only the merge images are shown. C2C12 (myoblasts) and A549 (epithelial) cells were used as positive controls fordesmin and keratin. Fibroblasts from each patient type were incubated with antibodies specific to vimentin (left), desmin (middle), and keratin(right). All human lung fibroblast derived by this method were positive for vimentin but negative for desmin and keratin and exhibited typicalfibroblast morphology. Magnification = 20x. Representative images are shown.Table 2 Clinical characteristics of the CanCOLD SubjectsVariable Normal At Risk GOLD 1 GOLD 2+ Overall P-valuesAge, in year [mean(sd)] 67.75 (5.84) 66.06 (8.44) 66.94 (11.20) 67.67 (10.22) 0.952Sex, male gender [n(%)] 4 (25.00) 5 (31.25) 9 (56.25) 9 (50.00) 0.215Tobacco smoking status [n(%)]Never 16 (100.00) - 7 (43.75) 4 (22.22) <0.001*Ex-smokers - 14 (87.50) 6 (37.50) 9 (50.00) <0.001*Current smokers - 2 (12.50) 3 (18.75) 5 (27.78) 0.149Cigarette smoker pack-years [mean(sd)] - 14.56 (11.26) 20.02 (26.57) 27.28 (28.78) <0.001*Post-bronchodilator spirometry [mean(sd)]FEV1, L 2.38 (0.48) 2.55 (0.63) 2.59 (0.56) 1.60 (0.70) <0.001*FEV1, % predicted 101.28 (11.33) 100.04 (10.75) 101.06 (20.55) 60.62 (16.68) <0.001*FEV1/FVC, % 76.39 (4.21) 78.23 (4.48) 67.04 (7.14) 53.18 (12.93) <0.001*Biomarker expression [mean(sd)]RelB (X10−3) 0.74 (1.08) 0.61 (0.19) 0.61 (0.28) 0.56 (0.16) 0.326COX-2 (X10−3) 7.74 (8.35) 5.62 (1.94) 6.30 (2.74) 6.76 (2.95) 0.453146a (X10-3) 0.09 (0.11) 0.46 (1.10) 0.47 (0.62) 1.22 (3.37) 0.064*indicates significance between the groups for the clinical characteristics in each row.Sheridan et al. Respiratory Research (2015) 16:54 Page 4 of 16Sheridan et al. Respiratory Research (2015) 16:54 Page 5 of 16linear phases of PCR amplifications and the cyclethreshold (Ct) of each reaction was detected. Gene ex-pression data were analyzed using the ΔΔCt methodnormalized to housekeeping (β-actin).Analysis of miR-146a expressionmiRNA expression was assessed by two-step TaqMan®RT-PCR (Applied Biosystems, Carlsbad, CA) for miR-146a and U6 snRNA, a small nuclear RNA (snRNA)used as an internal control for miRNA analysis [43,44].miRNA expression was normalized to the U6 snRNAlevels and fold-change was determined using 2−ΔΔCtmethod as we have described [31,45].Western blotFibroblasts were grown to approximately 90% conflu-ence before being treated with CSE. In separate experi-ments, Normal fibroblasts were pretreated with theproteasome inhibitor MG-132 (10 μM) for 2 hours priorto addition of 2% CSE. Total cellular protein was pre-pared using 1% IGEPAL lysis buffer [40] and 5–10 μg ofprotein were fractionated on SDS-PAGE gels andelectro-blotted onto Immun-blot PVDF membrane (Bio-Rad Laboratories, Hercules, CA). Antibodies againstRelB (1:1000; Cell Signaling), COX-2 (1:1000) (CaymanChemical, Ann Arbor, MI), p65 (1:1000, Santa Cruz) andtotal Actin (1:50,000; Milipore, Temecula, CA) wereused to assess changes in relative expression. Proteinswere visualized using HRP-conjugated secondary anti-bodies (1:10,000) followed by enhanced chemilumines-cence (ECL) and imaged using a ChemiDoc™ XRS+System (Bio-Rad).RelB knock-down in primary lung fibroblastsNormal fibroblasts (non-smoker) were seeded at 1–2 x104 cells/cm2 and transfected with 40 nM of siRNAagainst RelB (Santa Cruz, Catalogue number sc-36402)or non-targeting control siRNA (Santa Cruz, Cataloguenumber sc-37007) according to manufacturer’s instruc-tions. Six hours after the transfection, the cells wereswitched to serum-free MEM. On the next day, cellswere treated with 2% CSE for 3–24 hours and RNA orprotein collected for further analysis as described above.Verification of RelB knock-down was done by westernblot.ImmunocytochemistryLung fibroblasts from Normal, At Risk and COPD subjectswere cultured on glass chamber slides and left untreated orwere treated with 2% CSE. Following treatments, cells werewashed once with PBS/Tween, permeabilized/fixed using3% H2O2/methanol for 10 min, and blocked with UniversalBlocking Solution for 1 hour at room temperature. Theantibodies against RelB (1:300) and p65 (1:200) were dilutedin Antibody Diluent Solution (Dako) and incubated over-night at 4°C. Alexa Fluor-555 anti-goat or anti-rabbit IgGantibody was used for secondary binding (1:1000) and incu-bated for 1 hour at room temperature. Slides were thenmounted in ProLong® Gold Anti-Fade (Invitrogen), viewedon an Olympus IX71 fluorescent microscope (Olympus,Ontario, Canada) and photographed using a Retiga 2000Rcamera with ImagePro Plus software. Fluorescent images ofnuclei are visualized by Hoechst staining (1:2000). All pro-cedures were performed at the same time to minimize vari-ability in fluorescence intensity.Statistical analysisFor experimental data utilizing human lung fibroblasts,statistical analysis was performed using GraphPad Prism6 (v. 6.02; La Jolla, CA). A two-way analysis of variance(ANOVA) followed by a Newmann-Keuls test was usedto assess differences between treatment groups of morethan two factors when grouped by two variables unlessotherwise indicated. A one-way analysis of variance(ANOVA) followed by a Newman-Keuls multiple com-parisons test was used to assess differences in baselinevalues between the three subject groups. Statistical ana-lysis of blood mRNA/miRNA levels with clinical param-eters was analyzed on SAS version 9.3 (SAS Institute.Inc., Cary, N.C). Descriptive data were summarizedusing means and STD distributions or counts and per-centages for the four study groups. Statistically signifi-cant differences among the four groups were thencompared by using ANOVA analysis (or their non-parametric equivalence-Kruskal-Wallis Test) and theChi-square test as appropriate. Analysis of the correl-ation for each of the biomarker expression (Cox-2, RelB,miR-146a) with other biomarkers as well as clinical vari-ables was performed using Pearson’s correlation coeffi-cient. Results are expressed as the mean ± SEM or SD.In all cases, a p value < 0.05 is considered statisticallysignificant.ResultsLung fibroblast characterizationThe clinical features of the subjects from which the lung fi-broblasts were derived are given in Table 1. The FEV1/FVCratio after bronchodilators was significantly lower in theCOPD patients compared to either Normal (non-smokers)or smokers without COPD (At Risk). Non-smokers wereidentified as those individuals who were never-smokers(0 pack-years). There was no significant difference in pack-years between the smokers with and without COPD. Allprimary lung fibroblast used in this study had typical fibro-blast morphology (flat, elongate with oval nuclei) andexpressed vimentin (Figure 1). No staining was observedfor cytokeratin or desmin indicating that the cultured fibro-blasts did not contain cells of epithelial or muscle origin.RelB mRNA and protein expression is decreased in AtRisk- and COPD-derived lung fibroblastsOur published data show that RelB suppresses cigarettesmoke-induced COX-2 protein expression [30]. RelB is de-graded by cigarette smoke in vitro and in vivo [29,46], rais-ing the possibility that reduced RelB expression due tocigarette smoke exposure contributes to heightened COX-2expression in COPD. RelB protein at the predicted molecu-lar weight (MW ≈ 68 kDa) [47,48] was detectable in mostof the lung fibroblasts examined; also evident was a bandwith a lower MW of ≈ 55 kDa (Figure 2A, arrows) that areconsistent in size with degradation products of RelB [49].In fibroblasts derived from both At Risk and COPD sub-jects, there appeared to be lower levels of RelB. Densito-metric analysis of RelB protein at the predicted size(MW ≈ 68 kDa, analyzed in all subsequent Figures) indi-cated that there was a significant decrease in RelB proteinexpression in human lung fibroblasts from At Risk andCOPD subjects compared to Normal (Figure 2B). Therewas no significant difference in RelB protein betweensmokers with and without COPD. The relative expressionlevel of RelB mRNA was also significantly different be-tween Normal and either At Risk- and COPD-derived fi-broblasts (Figure 2C). RelB was predominantly localized tobld Mg12t RfferaasthenwSheridan et al. Respiratory Research (2015) 16:54 Page 6 of 16Figure 2 Reduced RelB mRNA and protein expression in human lung fibroexpression was detected in most lung fibroblasts examined at the predicteThere was an apparent decrease in RelB protein levels in the majority of lunfibroblasts from different individuals (Normal = 6; At Risk = 15; COPD =RelB protein expression in lung fibroblasts derived from the lungs of Acompared to fibroblasts from Normal subjects). There was no significant diexpressed as the mean ± SEM and each symbol represents fibroblasts frompredicted MW of 68 kDa. (C) RelB mRNA- Relative RelB mRNA expression wto Normal (1.0 ± 0.18; * p < 0.0044) fibroblasts. Results are expressed asfibroblasts. (D) RelB localization: Immunofluorescent imaging of quiescand COPD lung fibroblasts was localized to the cytoplasm (red colour)appeared to be qualitatively less RelB (based on intensity) in the At Risk atwo independent experiments.asts from At Risk and COPD subjects. (A) RelB western blot- RelB proteinW of 68 kDa. A faster migrating band of≈ 55 kDa was also detected.fibroblasts from At Risk and COPD subjects. Sample numbers refer to lung). (B) RelB protein densitometry- there was a significant decrease inisk (0.41 ± 0.08) as well as COPD subjects (0.38 ± 0.07) (*p = 0.0022ence in RelB levels between At Risk and COPD fibroblasts. Results aredifferent individual. Densitometry is based on the band detected at thesignificantly lower in At Risk (0.5 ± 0.06) and COPD (0.57 ± 0.07) comparede mean ± SEM (fold-change) of RelB levels normalized to the Normalt lung fibroblasts revealed that the majority of RelB in Normal, at Riskith minimal RelB evident in the nucleus (blue colour). Note that therend COPD-derived lung fibroblasts. Representative images based onthe cytoplasm in Normal, At Risk and COPD-derived lungfibroblasts cultured under basal condition, which ap-peared noticeably reduced in the At Risk and COPDcells (Figure 2D). Collectively, these data indicate thatcigarette smoke exposure is associated with reducedRelB expression.Regulation of RelB expression in human lung fibroblastsby CSEExposure to cigarette smoke, but not CD40L, results in lossof RelB protein in murine lung fibroblasts [29]. RelB is de-graded in a signal-specific manner in T cells [50], renderingit possible that chronic exposure to smoke contributes tothe loss in RelB in At Risk and COPD lung fibroblasts. Toevaluate this, we exposed lung fibroblasts to CSE via ourin vitro model of smoke exposure [29,40,51]. In lung fibro-blasts from both Normal and At Risk subjects, there was asignificant increase in RelB mRNA following 24 hours ofexposure to 2% CSE in vitro (Figure 3A, white and greybars, respectively). In COPD lung fibroblasts, RelB mRNAlevels remained unchanged throughout the 24-hour ex-posure time (Figure 3A, black bars). Consistent withthe data presented in Figure 2, basal RelB protein ex-pression remained significantly lower in At Risk andsurRNuransefibrmerobparchanSheridan et al. Respiratory Research (2015) 16:54 Page 7 of 16Figure 3 RelB protein expression is decreased by cigarette smoke expoCOPD subjects were exposed to 2% CSE for 3, 6, or 24 hours and RelB mincreased in Normal and At Risk fibroblasts exposed to 2% CSE for 24 hoaltered by CSE in COPD lung fibroblasts. Results are expressed as the mesubjects for each group. (B) RelB protein: There was a noticeable decreaBasal RelB protein levels were noticeably less in At Risk and COPD lungare shown of three experiments. (C) RelB protein- Densitometry: DensitoRelB protein expression between Normal and At Risk and COPD lung fibdecreased in Normal lung fibroblasts at 6 hours of exposure to 2% CSE com(fold-change) utilizing fibroblasts derived from three individual subjects for eawith MG-132 for 2 hours prior to cotreatment with 2% CSE for 6 hoursfollowing exposure to 2% CSE for 6 hours. MG-132 increased basal leveWestern blot image is representative of 3 separate experiments.e in Normal lung fibroblasts. Lung fibroblasts from Normal, At Risk andA and protein levels evaluated. (A) RelB mRNA-CSE: RelB mRNA wass (*p < 0.05 and **p < 0.001). RelB mRNA levels were not significantly± SEM (fold-change) utilizing fibroblasts derived from 4–6 individualin RelB protein in Normal fibroblasts exposed to 2% CSE for 6 hours.oblasts compared to Normal fibroblasts. Representative western blotstric analysis of RelB expression revealed a significant difference in basallasts in the absence of 2% CSE (media only). RelB protein expressioned to media control (* p < 0.05). Results are expressed as the mean ± SEMgroup. (D) RelB protein: MG-132: Normal lung fibroblasts were pretreatedd protein harvested for RelB expression. RelB levels were decreasedls of RelB and prevented the loss of RelB after exposure to 2% CSE.COPD fibroblasts cultured in a monolayer for the dur-ation of the in vitro experiments (Figure 3B and C).RelB protein decreased in CSE-exposed Normal fibro-blasts at 6 hours compared to media-only exposed cells(Figure 3B and C). There was no further change in RelBprotein with 2% CSE exposure in fibroblasts from AtRisk or COPD subjects. The reduction in RelB proteinlevels in Normal lung fibroblasts exposed to CSE wasdue to degradation by the proteasome. Treatment of fi-broblasts with the cell-permeable and reversible prote-asome inhibitor MG-132 [52,53] increased basal RelBprotein and also attenuated the reduction in RelB pro-tein upon exposure to 2% CSE (Figure 3D). These datasupport that exposure to cigarette smoke reduces theexpression of RelB in lung fibroblasts via the prote-asome degradation pathway.RelB attenuates cigarette smoke-induced COX-2 expres-sion in primary lung fibroblastsOur published data utilizing RelB-deficient murine fi-broblasts show that RelB potently suppresses cigarettesmoke-induced COX-2 expression [31], leading us tospeculate that lower RelB expression contribute to in-creased COX-2 in response to CSE. In response to 2%CSE, there was a transient increase in Cox-2 mRNA inpeaking at 3 hours of exposure (Figure 4A). The induc-tion in Cox-2 mRNA was significant in the COPD lungfibroblasts at all time-points examined compared tomedia-only (Figure 4A, black bars). There was minimalinduction in Cox-2 mRNA in Normal fibroblasts in re-sponse to CSE (Figure 4A, white bars) while Cox-2mRNA induction in Smoker fibroblasts was moderatelyhigher (Figure 4B, grey bars). Exposure to 2% CSE alsoincreased COX-2 protein expression only modestly in Nor-mal fibroblasts but with noticeably higher induction occur-ring in the At Risk and COPD fibroblasts (Figure 4B), bothof which also had significantly lower RelB protein levels(Figure 2).To now test whether RelB is a factor that suppressesCOX-2 induction upon direct smoke exposure, we usedsiRNA to knock-down RelB expression in Normal fi-broblasts which express relatively high levels of RelB(Figure 2A). Following confirmation of successful reductionin RelB protein (Figure 5A), we first performed qPCR forCox-2 mRNA after exposure to 2% CSE. In Normallung fibroblasts receiving Control siRNA (i.e. with RelBexpression; siCtrl) there was low induction of Cox-2mRNA (Figure 5B, open bars). Attenuation of RelB ex-pression via siRNA knockdown significantly increasedCox-2 mRNA expression when cells were exposed tomexatmRX-2Sheridan et al. Respiratory Research (2015) 16:54 Page 8 of 16At Risk and COPD lung fibroblasts, with inductionFigure 4 Cox-2 mRNA and protein expression is induced by 2% CSE in Ssignificant increase in Cox-2 mRNA in COPD lung fibroblasts (black bars)control); there was a significantly more Cox-2 mRNA in COPD fibroblastsCox-2 mRNA in At Risk fibroblasts (grey bars). Note the lack of relative Cox-2as the mean ± SEM of 4–9 fibroblasts from each subject group. (B) CONormal fibroblasts. There was noticeably more COX-2 protein in the At Risearlier COPD fibroblasts (by 3 hours). Representative western blot is sho2% CSE (Figure 5B, black bars). There was also aoker and COPD lung fibroblasts. (A) Cox-2 mRNA- CSE: There was aposed to 2% CSE for 3 hours (*** p <0.05 compared to respective6 and 24 hours of exposure. There was a trend towards increasedNA induction in Normal lung fibroblasts (white bars). Results are expressedprotein- CSE: There was a modest induction in COX-2 protein ink and COPD fibroblasts. The induction in COX-2 protein occurredwn of at least 3 different experiments.dramatic increase in COX-2 protein in RelB knock-down cells in response to 2% CSE (Figure 5C) support-ing that the attenuation of cigarette smoke-induction ofCOX-2 in vitro in Normal lung fibroblasts is associatedwith RelB expression.p65 is similar between Normal, At Risk and COPD-derivedlung fibroblastsThe promoter of Cox-2 contains two NF-κB binding sitesand thus plays an important role in the transcriptionalregulation of Cox-2 expression via the classic (p65/p50)pathway [54]. It is possible therefore that the differencein Cox-2 expression with CSE between At Risk andCOPD (but with no difference in RelB protein- Figure 2)could be due to altered p65 expression or nuclearlocalization. Total p65 protein expression in lung fibro-blasts from the three subject groups after exposure to2% CSE for up to 24 hours was not noticeably different(Figure 6A). The localization of p65 in media only cells waslargely cytoplasmic in fibroblasts from Normal, At. Riskand COPD subjects (Figure 6B). Exposure to 2% CSE for30 minutes [29] marginally increased nuclear p65 in allthree groups with little difference between them (Figure 6B).Cigarette smoke induction of miR-146a in Normal humanlung fibroblasts is independent of RelB expressionWe recently published that RelB promotes the induction ofmiR-146a in response to cigarette smoke in murine lung fi-broblasts as a mechanism through which RelB limits COX-2 protein levels [31]. COPD lung fibroblasts have reducedmiR-146a induction in response to pro-inflammatory cyto-kines, a finding that correlated with increased COX-2 ex-pression [36]. These findings led us to speculate that RelBcontrol over smoke-induced COX-2 expression occurs viamiR-146a. While there was significantly more basal miR-146a in COPD lung tissue compared to Normal (Figure 7A),there was no difference in basal levels of miR-146a betweenthe three fibroblast groups (Figure 7B). In Normal fibro-blasts, but not At Risk or COPD cells, there was a signifi-cant increase in miR-146a after exposure to 2% CSE for3 hours (Figure 8A), the peak expression time for miR-146ain response to cigarette smoke [31]. Knock-down of RelBhad no effect on miR-146a (Figure 8B) suggesting regula-tion of miR-146a expression by cigarette smoke is inde-pendent of RelB in human lung fibroblasts.Correlation between systemic RelB, Cox-2 and miR-146aexpression and clinical features of COPD in the CanCOLDCod: 0.5woCtSheridan et al. Respiratory Research (2015) 16:54 Page 9 of 16Figure 5 siRNA-mediated knock-down of RelB potentiates CSE-inducedsiRNA: Lung fibroblasts derived from Normal subjects were transfecteby approximately 40% compared to Ctrl-transfected (siCtrl; relative-changeexperiments. (B) Cox-2 mRNA: Normal fibroblasts transfected with RelB siRNAby qRT-PCR. There was a significant increase in Cox-2 mRNA expression*p < 0.05 compared to RelB siRNA media only; $ p < 0.05 compared toTherefore alterations in p65 expression or localization can-not account for the differential regulation of Cox-2 betweenAt Risk and COPD lung fibroblasts.SEM, n = 4–5 experiments. (C) COX-2 protein- there was a corresponding and dCSE compared to the siCTRL cells. Representative western blot is shown of at lecohortWe recently published that RelB expression correlates withclinical features of COPD exacerbations [55]. To nextx-2 mRNA and protein expression in Normal lung fibroblasts. (A) RelBwith siRNA against RelB (siRelB). RelB protein levels were decreased9 ± 0.065). Results are expressed as the mean ± SEM, n = 4 independentere exposed to 2% CSE for 6 hours and Cox-2 mRNA expression evaluatednly when RelB expression is reduced (siRelB; fold-induction 15 ± 8;rl siRNA exposed to 2% CSE). Results are expressed as the mean ±ramatic increase in COX-2 protein in the siRelB fibroblasts exposed to 2%ast two independent experiments.Sheridan et al. Respiratory Research (2015) 16:54 Page 10 of 16establish if there was correlation between systemic expres-sion of RelB, Cox-2 and miR-146a in COPD, we utilized theCanCOLD cohort. Of these subjects, there was no signifi-cant difference in terms of age and gender between the 4study groups (Table 2). RelB and Cox-2 expression werenot significantly different between the subject groupswhereas the relative expression of miR-146a approachedFigure 6 p65 expression and localization in Normal, At Risk and COPD lunin p65 expression between Normal, At Risk and COPD-derived lung fibroblis shown of 3 independent experiments. (B) p65 localization: There was littfrom the three subject groups, which appeared predominantly cytoplasmicalter the localization of p65. Representative images shown are based on twFigure 7 Basal miR-146a expression in human lung fibroblasts from Normal suThere was a significant increase in miR-146a in lung tissue derived from COPD seach symbol represents a different individual (Normal, n = 10; At Risk, n = 9; COPmiR-146a expression in lung fibroblasts derived from the three patient grnormalized to the snRNA U6 and are expressed as fold-change compared toindividual (Normal, n = 5; At Risk, n = 13; COPD = 11).statistical significance (p = 0.064) (Table 2). Neither RelBnor Cox-2 mRNA expression correlated with the clinicalvariables evaluated in this population, including lung func-tion (Table 3). However, miR-146a significantly correlatedwith FEV1% predicted (p = 0.026) and trended towards sig-nificance for FEV1/FVC (p = 0.065) (Table 3). There was asignificant increase in systemic miR-146a in COPD Ig fibroblasts. (A) p65 expression: There was no perceptible differenceasts exposed to 2% CSE for up to 24 hours. Representative western blotle perceptible difference in the localization of p65 in lung fibroblasts(red colour). Exposure to 2% CSE for 30 minutes did not appreciablyo independent experiments.bjects as well as smokers with and without COPD. (A) miR-146a-lung tissue:ubjects compared to Normal. Results are expressed as mean ± SEM andD= 9). (B) miR-146a-fibroblasts: There was no significant difference in basaloups. Results are expressed as the mean ± SEM of miR-146a levelsNormal fibroblasts. Each symbol represents fibroblasts from a differentFigure 8 Regulation of miR-146a by CSE is independent of RelB. (A) miR-146a- CSE: 2% CSE significantly increased miR-146a expression only inNormal lung fibroblasts at 3 hours (fold-induction: 5.44 ± 2.4; * p < 0.05). Results are expressed as the mean ± SEM of normalized miR-146a levels;n = 4 independent experiments utilizing fibroblasts from 4 different subjects of each phenotype; post-hoc analysis was performed by Fisher’s LSD.(B) miR-146a-CSE RelB siRNA: There was no significant difference in basal or CSE-exposed (3 hrs) miR-146a levels in siRelB cells (black bars) compared tosiCtrl cells (open bars). Results are expressed as the mean ± SEM of normalized miR-146a levels from 2–4 independent experiments.Sheridan et al. Respiratory Research (2015) 16:54 Page 11 of 16(GOLD 1) compared to Normal (Figure 9). There was alsoa significant positive correlation between RelB and Cox-2expression (Table 3 and Figure 10). Collectively these datahighlight a differential role for systemic versus pulmonaryRelB expression in COPD and suggest the importance oflung structural cell RelB expression in regulating inflamma-tion caused by smoke exposure.DiscussionCOPD is an obstructive lung disease that is increasing inprevalence worldwide, affecting an estimated 200 millionpeople [56]. While the etiology of COPD is stronglylinked to smoke exposure, the underlying pathogenicmechanisms by which smoke causes chronic, aberrantpulmonary inflammation remains poorly defined. Nu-merous signal transduction pathways, including the clas-sic (p65/p50) NF-κB pathway, contribute to cigarettesmoke-induced inflammation. The pro-inflammatory ac-tivities of the classic NF-κB pathway are counterba-lanced by another REL protein called RelB, initiallyTable 3 Pearson correlation coefficients between biomarkersVariable 1 Variable 2 CorrelaRelB expression Cox-2 expression 0.86RelB expression 146a −0.08RelB expression FEV1 L −0.04RelB expression FEV1% predicted 0.05RelB expression FEV1/FVC 0.10Cox-2 expression 146a −0.12Cox-2 expression FEV1 L 0.01Cox-2 expression FEV1% predicted 0.09Cox-2 expression FEV1/FVC 0.10146a FEV1 L −0.20146a FEV1% predicted −0.28146a FEV1/FVC −0.23**indicates significant correlation between RelB and Cox-2 expression.identified as I-Rel for Inhibitory Rel because of its abilityto reduce the transcriptional activities of NF-κB [47].We have published that over-expression of RelB dimin-ishes the induction of inflammatory mediators, includingCOX-2 as well as consequent lung neutrophilia causedby cigarette smoke [29,30]. These finding highlight RelBas a potentially important anti-inflammatory REL pro-tein that protects against the deleterious effects ofcigarette smoke, raising the possibility that low/absentRelB levels in the lung may predispose some individualswho smoke to aberrant inflammation and the eventualdevelopment of COPD. To our knowledge, we are thefirst to demonstrate that lung fibroblasts from At Risk(smokers with no airflow obstruction) and COPD sub-jects have significantly less RelB mRNA and protein ex-pression compared to fibroblasts from individuals whoare non-smokers (Normal).One of the more intriguing findings from our study isthat while there was a significant decrease in RelB mRNA inlung fibroblasts from At Risk and COPD subjects (Figure 2),and clinical variablestion coefficient 95% CI P-value0.78 - 0.91 <0.001**−0.32 - 0.16 0.509−0.28 - 0.21 0.768−0.19 - 0.29 0.671−0.15 - 0.34 0.418−0.36 - 0.12 0.323−0.24 - 0.25 0.931−0.16 - 0.33 0.475−0.15 - 0.34 0.428−0.42 - 0.05 0.114−0.49 -0.04 0.026*−0.45 - 0.01 0.065D.exCf Reens.Sheridan et al. Respiratory Research (2015) 16:54 Page 12 of 16there was no significant change in systemic RelBmRNA levels based on smoking status or airflow limi-tation (Table 2). This latter observation is consistentwith our recent publication in which we demonstratedFigure 9 Systemic expression of Cox-2, RelB and miR-146a levels in COPCanCOLD project and total RNA, including miRNA, was isolated and theqRT-PCR. A total of 16 subjects from each category were analyzed from themRNA- there was no significant difference in the relative expression owas no significant difference in the relative expression of RelB mRNA betwrelative miR-146a expression in COPD Gold 1 compared to Normal subjectthat systemic RelB expression was associated withhealth outcomes during acute exacerbations in COPDbut was not associated with clinical features duringstable-state [55]. While this difference between pul-monary and systemic RelB expression may be reflectiveof the fact that the lung fibroblasts were derived fromdifferent subjects compared to those from which sys-temic RelB was measured, we postulate that this dispar-ity is reflective of fundamental differences in RelBFigure 10 Correlation between systemic RelB and Cox-2 mRNA expressionfrom the CanCOLD cohort. There was a significant positive correlation betwexpression, activation and function between structuraland immune cells. Whereas RelB is highly expressedand constitutively active in lymphoid cells [57], there isminimal RelB activity in quiescent structural cells suchPeripheral blood was obtained from subjects recruited as part of thepression of Cox-2 and RelB mRNA as well as miR-146a evaluated byanCOLD cohort; each symbol represents a different individual. (A) RelBelB mRNA between the subject categories. (B) Cox-2 mRNA: therethe subject categories. (C) miR-146a: there was a significant increase inas endothelial cells and fibroblasts [24,58]. In responseto certain inflammatory triggers however (e.g. IL-1β, lipo-polysaccharide [LPS], CD40L), RelB expression and/or ac-tivity is increased which serves to dampen the expression ofinflammation-associated proteins such as chemokines andadhesion molecules [24,29,59]. This is not the case with im-mune cells, where RelB does not control the production ofchemokines from LPS-stimulated macrophages [59].Moreover the anti-inflammatory abilities of RelB inin COPD. A total of 16 subjects from each category were analyzedeen systemic RelB and Cox-2 mRNA expression.Sheridan et al. Respiratory Research (2015) 16:54 Page 13 of 16protecting against inflammation are largely attributedto non-hematopoietic cells, particularly fibroblasts. Evi-dence in support of this includes: (1) the transfer ofnormal bone marrow into irradiated RelB-deficientmice failed to alleviated the inflammatory syndromeand (2) injection of LPS-stimulated RelB−/− fibroblastspotently induced inflammation in vivo [59]. Thus, itmay be that lung fibroblasts with low RelB expressiondue to chronic smoke exposure (rather than immunecells) are active participants in the abnormal inflamma-tion associated with cigarette smoke exposure, capableof inciting and perpetuating the pulmonary inflamma-tory response via the production of inflammatory medi-ators such as COX-2.Our observation that lung fibroblasts from individualswho smoke (At Risk) and those with airflow limitation(COPD) have reduced RelB mRNA and protein expression(Figure 2) suggests that cigarette smoke directly reducesRelB levels, a notion further supported by our in vitro ex-perimental data showing CSE exposure decreases RelB pro-tein (Figure 3). Our data also support that the mechanismby which smoke decreases RelB protein is due to degrad-ation by the 26S proteasome (Figure 3D). Consistent withthis notion, others have demonstrated that RelB is proteo-lytically degraded in T cells in a ligand-specific manner[50,60], including the appearance of the lower MW bandof ≈ 55 kDa (Figure 2A) [49] and that RelB degradation alsooccurs in pulmonary cells exposed to hypercapnia [25]. An-other unresolved question is why there was a significant de-crease in RelB mRNA in smoker-derived (At Risk) lungfibroblasts (Figure 2) whereas exposure to CSE in vitro didnot significantly decrease RelB mRNA levels in any of thesubject groups (Figure 3). It is possible that chronic long-term exposure to cigarette smoke, characteristic of individ-uals in the At Risk and COPD categories, is necessary toalter RelB mRNA expression, such that a single exposure tocigarette smoke over a 24 hour time-period, our standardin vitro exposure protocol, is insufficient to alter RelB atthe mRNA level.It is somewhat paradoxical that there was a significantcigarette smoke-induction of Cox-2 mRNA only in COPDcells (and not At Risk), despite the fact that both have simi-lar low RelB expression levels (Figures 2 and 4). At firstglance this might indicate that RelB ultimately does notcontribute to the regulation of smoke-induced COX-2.However our data utilizing siRNA to knock-down RelB inNormal cells, where there is potentiation of CSE-inducedCox-2 mRNA expression (Figure 5), expands data obtainedfrom our in vitro and in vivo models of smoke exposure[30] to show that RelB contributes to the suppression ofcigarette smoke-induced COX-2 in human lung cells. Oneexplanation for these data is that RelB is in fact critical inattenuating COX-2 in naive cells that initially encounter re-spiratory toxicants such as cigarette smoke, but that RelBalone is insufficient in counter-balancing the deleteri-ous effects associated with chronic, long-term smokeexposure in COPD subjects. It may also be that post-translational modifications of RelB protein occur inCOPD, accounting for the differential response be-tween At Risk and COPD-derived cells, both of whichhave reduced RelB protein levels but divergent tran-scriptional changes in Cox-2 mRNA. Protein ubiquiti-nation, which is implicated in COPD pathogenesis [61],augments the transactivation potential of RelB to pro-mote NF-κB-dependent transcription [62]. Thus if suchmodifications account for the difference in responsebetween At Risk and COPD cells, then alteration ofRelB levels (via siRNA) in At Risk fibroblasts would beineffectual in altering smoke-induced Cox-2 expression.There is also no difference in the expression or nuclearlocalization of p65 between the three groups in re-sponse to CSE (Figure 6), making it unlikely that RelBsuppression of Cox-2 is via alteration in the canonicalNF-κB pathway. It remains possible that co-activatorssuch as p300, necessary for transcriptional induction ofCox-2 by NF-κB [63], are repressed by RelB or alteredin COPD, thereby accounting for differences in Cox-2between At Risk and COPD fibroblasts or the initiationof Cox-2 transcription after RelB knock-down. More-over, cigarette smoke exposure can cause epigeneticchanges in the lung, leading to significant increases ininflammatory proteins associated with COPD patho-genesis [64,65]. Thus it could also be that, in additionto low RelB levels in COPD, there are further epigen-etic changes in the lung or additional protein modifica-tions to the RelB protein not identified in this studythat render it unable to exert negative control over re-peated/chronic exposures. Such epigenetic alterationscould also be why there is reduced RelB mRNA in lungfibroblasts from chronic smokers but not after a single24-hour exposure. Finally, another potential explan-ation is that a protein partner of the RelB pathway es-sential for its full inhibitory activities (e.g. p100/p52 orthe aryl hydrocarbon receptor [AhR]) [29] may also beinherently absent or defective in COPD subjects, andthus not allow for the full anti-inflammatory abilities ofRelB. These and other possibilities are currently beingexplored.Recently, a reciprocal relationship between RelB andmiR-146a in immunity and inflammation has emerged in-cluding the induction of miR-146a by IL-1β [66] and CSE[31]. McMillan and colleagues demonstrated that in adultlung fibroblasts, downregulation of RelB via siRNA de-creases the magnitude of IL-1β-induced miR-146a expres-sion [66]. Sato et al. demonstrated that COPD fibroblastsproduce less cytokine-stimulated miR-146a compared to fi-broblasts from smokers (At Risk) [36]. In our study, bothAt Risk and COPD fibroblasts also failed to significantlySheridan et al. Respiratory Research (2015) 16:54 Page 14 of 16increase miR-146a in response to CSE, and effect thatwas independent of RelB expression (Figure 8), suggest-ing that RelB does not contribute to CSE-induction ofmiR-146a in human lung fibroblasts. This differs fromour recently published data utilizing RelB−/− mouselung fibroblasts, which has significantly less miR-146acompared to RelB-expressing cells [31]. While this mayreflect species-specific differences in basal miR-146aregulation, it is equally likely that even low detectablelevels of RelB expression in At Risk and COPD fibro-blasts are sufficient to promote basal miR-146a expres-sion. Of the potential systemic markers examined inthis study, only miR-146a was associated with healthoutcomes (Table 3) and increased in COPD (Figures 7and 9). Our findings that basal miR-146a was higher inboth COPD lung tissue as well as systemically in theblood, but not in lung fibroblasts, suggests that cells ofhematopoietic origin- and not lung structural cells-contribute to this heightened expression. It is also intri-guing that miR-146a was significantly increased in COPDGold 1 but not with more severe disease (Gold 2+).Whether the increased miR-146a in COPD 1 is a compen-satory mechanism or predictive of those individuals whowill go on to develop more severe COPD is not known.Further longitudinal assessment of miR-146a utilizing theCanCOLD population may reveal the novelty of miR-146aas a biomarker of COPD progression.We recognize that there are several limitations with ourstudy, including the cross-sectional nature of the data ob-tained from the fibroblasts derived from the lung surgicalspecimens. This ultimately makes us unable to determine ifthe individuals who are smokers with low RelB will ultim-ately develop COPD. We also cannot exclude the possibilitythat medications (e.g. inhaled corticosteroids) taken by thesubjects in our study had an impact on the relative expres-sion levels of RelB, as corticosteroids can dampen Cox-2gene transcription via NF-κB [67]. Finally, another per-ceived limitation is the reliance on mRNA levels to correl-ate with clinical parameters, as quantification of blood RelBprotein expression remains to be determined. Despite theselimitations, our data support that RelB suppresses COX-2expression upon exposure to cigarette smoke. When con-sidered with our previous work [31], our data implies thatRelB and miR-146a may work cooperatively to suppressCOX-2 expression in response to environmental toxicants.ConclusionsTo the best of our knowledge, we are the first to reporton the expression of RelB in primary lung fibroblasts de-rived from COPD, and show that cigarette smoke con-tributes to a reduction in RelB expression in lungstructural cells. Our data further demonstrate the im-portance of RelB expression in attenuating COX-2 pro-tein in response to in vitro exposure to cigarette smokeextract. Whether low RelB levels predispose to the de-velopment of COPD in susceptible individuals is notknown but further molecular investigation into the alter-native NF-κB pathway will enhance our understandingof RelB in COPD and may contribute to the develop-ment of novel, lung-targeted anti-inflammatory treat-ments for smoke-related lung disorders.AbbreviationsCanCOLD: Canadian chronic obstructive lung disease; COPD: Chronicobstructive pulmonary disease; COX-2: Cyclooxygenase-2; miRNA: MicroRNA;NF-κB: Nuclear factor-κB.Competing interestsThe authors declare that they have no competing interests.Authors’ contributionsConception and experimental design: JAS, MZ, CJB; Data analysis andinterpretation: MZ, PN, PZL, QH, DHE, ALB, CJB; Drafting of the manuscript forimportant intellectual content: PN, JB, WCT, DHE, CJB. All authors read andapproved the final manuscript.Authors’ informationJared A Sheridan and Michela Zago co-first authorship.AcknowledgementsThis work was supported by the Canada Foundation for Innovation (CFI) andthe Canadian Institutes of Health Research (CIHR). CJB was supported by asalary award from the Fonds de recherche du Quebec-Sante (FRQ-S). MZ isthe recipient of a Meakins-Christie Post-Doctoral Fellowship Award. Dr Nair issupported by a Canada Research Chair in Airway Inflammometry.Acquisition of human lung tissue was facilitated by Katherine Radford,Department of Pathology, the Division of Thoracic Surgery and theDepartment of Pathology of St Joseph’s Healthcare Hamilton, ON.Author details1Department of Medicine, 1001 Decarie Blvd, Montreal, QC H4A 3J1, Canada.2Research Institute of the McGill University Health Centre, 1001 Decarie Blvd,Montreal, QC H4A 3J1, Canada. 3Department of Epidemiology andBiostatistics, 1001 Decarie Blvd, Montreal, QC H4A 3J1, Canada. 4RespiratoryEpidemiology and Clinical Research Unit, Montreal Chest Institute, McGillUniversity, Montreal, QC, Canada. 5Department of Medicine, McMasterUniversity, Hamilton, ON, Canada. 6The UBC James Hogg Research Centre,University of British Columbia, Vancouver, BC, Canada.Received: 15 December 2014 Accepted: 21 April 2015References1. Churg A, Cosio M, Wright JL. Mechanisms of cigarette smoke-inducedCOPD: insights from animal models. Am J Physiol Lung Cell Mol Physiol.2008;294:L612–31.2. Chung KF. Inflammatory mediators in chronic obstructive pulmonarydisease. Curr Drug Targets Inflamm Allergy. 2005;4:619–25.3. Arnson Y, Shoenfeld Y, Amital H. Effects of tobacco smoke on immunity,inflammation and autoimmunity. J Autoimmun. 2010;34:J258–65.4. Martey CA, Pollock SJ, Turner CK, O’Reilly KM, Baglole CJ, Phipps RP, et al.Cigarette smoke induces cyclooxygenase-2 and microsomal prostaglandinE2 synthase in human lung fibroblasts: implications for lung inflammationand cancer. Am J Physiol Lung Cell Mol Physiol. 2004;287:L981–91.5. Montuschi P, Kharitonov SA, Ciabattoni G, Barnes PJ. Exhaled leukotrienesand prostaglandins in COPD. Thorax. 2003;58:585–8.6. Fritscher LG, Post M, Rodrigues MT, Silverman F, Balter M, Chapman KR,et al. Profile of eicosanoids in breath condensate in asthma and COPD.J Breath Res. 2012;6:026001.7. Roh GS, Yi CO, Cho YJ, Jeon BT, Nizamudtinova IT, Kim HJ, et al.Anti-inflammatory effects of celecoxib in rat lungs with smoke-inducedemphysema. Am J Physiol Lung Cell Mol Physiol. 2010;299:L184–91.Sheridan et al. Respiratory Research (2015) 16:54 Page 15 of 168. Anto RJ, Mukhopadhyay A, Shishodia S, Gairola CG, Aggarwal BB.Cigarette smoke condensate activates nuclear transcription factor-kappaBthrough phosphorylation and degradation of IkappaB(alpha): correlationwith induction of cyclooxygenase-2. Carcinogenesis. 2002;23:1511–8.9. Petkova DK, Pang L, Range SP, Holland E, Knox AJ. Immunocytochemicallocalization of cyclo-oxygenase isoforms in cultured human airway structuralcells. Clin Exp Allergy. 1999;29:965–72.10. Buckley CD, Pilling D, Lord JM, Akbar AN, Scheel-Toellner D, Salmon M.Fibroblasts regulate the switch from acute resolving to chronic persistentinflammation. Trends Immunol. 2001;22:199–204.11. Xi X, McMillan DH, Lehmann GM, Sime PJ, Libby RT, Huxlin KR, et al. Ocularfibroblast diversity: implications for inflammation and ocular wound healing.Invest Ophthalmol Vis Sci. 2011;52:4859–65.12. Silzle T, Randolph GJ, Kreutz M, Kunz-Schughart LA. The fibroblast: sentinelcell and local immune modulator in tumor tissue. Int J Cancer.2004;108:173–80.13. D’Anna C, Cigna D, Costanzo G, Ferraro M, Siena L, Vitulo P, et al. Cigarettesmoke alters cell cycle and induces inflammation in lung fibroblasts. Life Sci.2015;126:10–8.14. Profita M, Sala A, Bonanno A, Riccobono L, Ferraro M, La Grutta S, et al.Chronic obstructive pulmonary disease and neutrophil infiltration: role ofcigarette smoke and cyclooxygenase products. Am J Physiol Lung Cell MolPhysiol. 2010;298:L261–9.15. Sirianni FE, Chu FS, Walker DC. Human alveolar wall fibroblasts directly linkepithelial type 2 cells to capillary endothelium. Am J Respir Crit Care Med.2003;168:1532–7.16. Zandvoort A, van der Geld YM, Jonker MR, Noordhoek JA, Vos JT, WesselingJ, et al. High ICAM-1 gene expression in pulmonary fibroblasts of COPDpatients: a reflection of an enhanced immunological function. Eur Respir J.2006;28:113–22.17. Parsonage G, Filer AD, Haworth O, Nash GB, Rainger GE, Salmon M, et al.A stromal address code defined by fibroblasts. Trends Immunol.2005;26:150–6.18. Laveti D, Kumar M, Hemalatha R, Sistla R, Naidu VG, Talla V, et al. Anti-Inflammatory Treatments for Chronic Diseases: A Review. Inflamm AllergyDrug Targets. 2013;12(5):346–61.19. Togo S, Holz O, Liu X, Sugiura H, Kamio K, Wang X, et al. Lung fibroblastrepair functions in patients with chronic obstructive pulmonary diseaseare altered by multiple mechanisms. Am J Respir Crit Care Med.2008;178:248–60.20. Bren GD, Solan NJ, Miyoshi H, Pennington KN, Pobst LJ, Paya CV.Transcription of the RelB gene is regulated by NF-kappaB. Oncogene.2001;20:7722–33.21. Sun SC. The noncanonical NF-kappaB pathway. Immunol Rev.2012;246:125–40.22. Kim JY, Morgan M, Kim DG, Lee JY, Bai L, Lin Y, et al. TNFalpha inducednoncanonical NF-kappaB activation is attenuated by RIP1 throughstabilization of TRAF2. J Cell Sci. 2011;124:647–56.23. Yoza BK, Hu JY, Cousart SL, Forrest LM, McCall CE. Induction of RelBparticipates in endotoxin tolerance. J Immunol. 2006;177:4080–5.24. de Souza AR, Zago M, Eidelman DH, Hamid Q, Baglole CJ. ArylHydrocarbon Receptor (AhR) Attenuation of Subchronic CigaretteSmoke-induced Pulmonary Neutrophilia Is Associated with Retention ofNuclear RelB and Suppression of Intercellular Adhesion Molecule-1(ICAM-1). Toxicol Sci. 2014;140:204–23.25. Oliver KM, Lenihan CR, Bruning U, Cheong A, Laffey JG, McLoughlin P,et al. Hypercapnia induces cleavage and nuclear localization of RelBprotein, giving insight into CO2 sensing and signaling. J Biol Chem.2012;287:14004–11.26. Brustle A, Brenner D, Knobbe CB, Lang PA, Virtanen C, Hershenfield BM,et al. The NF-kappaB regulator MALT1 determines the encephalitogenicpotential of Th17 cells. J Clin Invest. 2012;122:4698–709.27. Lo D, Quill H, Burkly L, Scott B, Palmiter RD, Brinster RL. A recessive defect inlymphocyte or granulocyte function caused by an integrated transgene. AmJ Pathol. 1992;141:1237–46.28. Weih F, Carrasco D, Durham SK, Barton DS, Rizzo CA, Ryseck RP, et al.Multiorgan inflammation and hematopoietic abnormalities in mice with atargeted disruption of RelB, a member of the NF-kappa B/Rel family. Cell.1995;80:331–40.29. Baglole CJ, Maggirwar SB, Gasiewicz TA, Thatcher TH, Phipps RP, Sime PJ.The aryl hydrocarbon receptor attenuates tobacco smoke-inducedcyclooxygenase-2 and prostaglandin production in lung fibroblasts throughregulation of the NF-kappaB family member RelB. J Biol Chem.2008;283:28944–57.30. McMillan DH, Baglole CJ, Thatcher TH, Maggirwar S, Sime PJ, Phipps RP.Lung-targeted overexpression of the NF-kappaB member RelB inhibitscigarette smoke-induced inflammation. Am J Pathol. 2011;179:125–33.31. Zago M, Rico de Souza A, Hecht E, Rousseau S, Hamid Q, Eidelman DH, et al. TheNF-kappaB family member RelB regulates microRNA miR-146a to suppresscigarette smoke-induced COX-2 protein expression in lung fibroblasts. ToxicolLett. 2014;226:107–16.32. Lee RC, Ambros V. An extensive class of small RNAs in Caenorhabditis elegans.Science. 2001;294:862–4.33. Banerjee A, Luettich K. MicroRNAs as potential biomarkers of smoking-relateddiseases. Biomark Med. 2012;6:671–84.34. Schembri F, Sridhar S, Perdomo C, Gustafson AM, Zhang X, Ergun A, et al.MicroRNAs as modulators of smoking-induced gene expression changes inhuman airway epithelium. Proc Natl Acad Sci U S A. 2009;106:2319–24.35. Takahashi K, Yokota S, Tatsumi N, Fukami T, Yokoi T, Nakajima M. Cigarettesmoking substantially alters plasma microRNA profiles in healthy subjects. ToxicolAppl Pharmacol. 2013;272:154–60.36. Sato T, Liu X, Nelson A, Nakanishi M, Kanaji N, Wang X, et al. Reduced MiR-146aIncreases Prostaglandin E2 in Chronic Obstructive Pulmonary Disease Fibroblasts.Am J Respir Crit Care Med. 2010;182(8):1020–9.37. Mookherjee N, El-Gabalawy HS. High degree of correlation between whole bloodand PBMC expression levels of miR-155 and miR-146a in healthy controls andrheumatoid arthritis patients. J Immunol Methods. 2013;400–401:106–10.38. Bourbeau J, Tan WC, Benedetti A, Aaron SD, Chapman KR, Coxson HO, et al.Canadian cohort obstructive lung disease (CanCOLD): fulfilling the need forlongitudinal observational studies in COPD. COPD. 2014;11:125–32.39. Baglole CJ, Reddy SY, Pollock SJ, Feldon SE, Sime PJ, Smith TJ, et al. Isolationand phenotypic characterization of lung fibroblasts. Methods Mol Med.2005;117:115–27.40. Baglole CJ, Sime PJ, Phipps RP. Cigarette smoke-induced expression ofheme oxygenase-1 in human lung fibroblasts is regulated by intracellularglutathione. Am J Physiol Lung Cell Mol Physiol. 2008;295:L624–36.41. Rico de Souza A, Zago M, Pollock SJ, Sime PJ, Phipps RP, Baglole CJ. GeneticAblation of the Aryl Hydrocarbon Receptor Causes Cigarette Smoke-induced Mitochondrial Dysfunction and Apoptosis. J Biol Chem.2011;286:43214–28.42. Zago M, Sheridan JA, Nair P, Rico de Souza A, Gallouzi IE, Rousseau S, et al.Aryl hydrocarbon receptor-dependent retention of nuclear HuR suppressescigarette smoke-induced cyclooxygenase-2 expression independent ofDNA-binding. PLoS One. 2013;8:e74953.43. du Rieu MC, Torrisani J, Selves J, Al Saati T, Souque A, Dufresne M, et al.MicroRNA-21 is induced early in pancreatic ductal adenocarcinomaprecursor lesions. Clin Chem. 2010;56:603–12.44. Jiang J, Lee EJ, Gusev Y, Schmittgen TD. Real-time expression profiling ofmicroRNA precursors in human cancer cell lines. Nucleic Acids Res.2005;33:5394–403.45. Hecht E, Zago M, Sarill M, Souza A, Gomez A, Matthews J, et al. Arylhydrocarbon receptor-dependent regulation of miR-196a expressioncontrols lung fibroblast apoptosis but not proliferation. Toxicol Appl Pharmacol.2014;280:511–25.46. Thatcher TH, Maggirwar SB, Baglole CJ, Lakatos HF, Gasiewicz TA, Phipps RP,et al. Aryl hydrocarbon receptor-deficient mice develop heightenedinflammatory responses to cigarette smoke and endotoxin associatedwith rapid loss of the nuclear factor-kappaB component RelB. AmJ Pathol. 2007;170:855–64.47. Ruben SM, Klement JF, Coleman TA, Maher M, Chen CH, Rosen CA. I-Rel: anovel rel-related protein that inhibits NF-kappa B transcriptional activity.Genes Dev. 1992;6:745–60.48. Ryseck RP, Bull P, Takamiya M, Bours V, Siebenlist U, Dobrzanski P, et al. RelB,a new Rel family transcription activator that can interact with p50-NF-kappaB. Mol Cell Biol. 1992;12:674–84.49. Hailfinger S, Nogai H, Pelzer C, Jaworski M, Cabalzar K, Charton JE, et al.Malt1-dependent RelB cleavage promotes canonical NF-kappaB activation inlymphocytes and lymphoma cell lines. Proc Natl Acad Sci U S A.2011;108:14596–601.50. Marienfeld R, Berberich-Siebelt F, Berberich I, Denk A, Serfling E, NeumannM. Signal-specific and phosphorylation-dependent RelB degradation: apotential mechanism of NF-kappaB control. Oncogene. 2001;20:8142–7.51. Baglole CJ, Bushinsky SM, Garcia TM, Kode A, Rahman I, Sime PJ, et al.Differential induction of apoptosis by cigarette smoke extract in primaryhuman lung fibroblast strains: implications for emphysema. Am J PhysiolLung Cell Mol Physiol. 2006;291:L19–29.52. Grimm LM, Goldberg AL, Poirier GG, Schwartz LM, Osborne BA. Proteasomesplay an essential role in thymocyte apoptosis. EMBO J. 1996;15:3835–44.53. Rock KL, Gramm C, Rothstein L, Clark K, Stein R, Dick L, et al. Inhibitors ofthe proteasome block the degradation of most cell proteins and thegeneration of peptides presented on MHC class I molecules. Cell.1994;78:761–71.54. Chun KS, Surh YJ. Signal transduction pathways regulating cyclooxygenase-2 expression: potential molecular targets for chemoprevention. BiochemPharmacol. 2004;68:1089–100.55. Labonte L, Coulombe P, Zago M, Bourbeau J, Baglole CJ. Alterations in theexpression of the NF-kappaB family member RelB as a novel marker ofcardiovascular outcomes during acute exacerbations of chronic obstructiveSubmit your next manuscript to BioMed Centraland take full advantage of: • Convenient online submission• Thorough peer review• No space constraints or color figure charges• Immediate publication on acceptance• Inclusion in PubMed, CAS, Scopus and Google Scholar• Research which is freely available for redistributionSheridan et al. Respiratory Research (2015) 16:54 Page 16 of 16pulmonary disease. PLoS One. 2014;9, e112965.56. Lopez AD, Mathers CD, Ezzati M, Jamison DT, Murray CJ. Global andregional burden of disease and risk factors, 2001: systematic analysis ofpopulation health data. Lancet. 2006;367:1747–57.57. Lernbecher T, Kistler B, Wirth T. Two distinct mechanisms contribute to theconstitutive activation of RelB in lymphoid cells. EMBO J. 1994;13:4060–9.58. Olashaw NE. Inducible activation of RelB in fibroblasts. J Biol Chem.1996;271:30307–10.59. Xia Y, Pauza ME, Feng L, Lo D. RelB regulation of chemokine expressionmodulates local inflammation. Am J Pathol. 1997;151:375–87.60. Neumann M, Klar S, Wilisch-Neumann A, Hollenbach E, Kavuri S, Leverkus M,et al. Glycogen synthase kinase-3beta is a crucial mediator of signal-inducedRelB degradation. Oncogene. 2011;30:2485–92.61. Stepaniants S, Wang IM, Boie Y, Mortimer J, Kennedy B, Elliott M, et al.Genes related to emphysema are enriched for ubiquitination pathways.BMC Pulm Med. 2014;14:187.62. Leidner J, Palkowitsch L, Marienfeld U, Fischer D, Marienfeld R. Identificationof lysine residues critical for the transcriptional activity andpolyubiquitination of the NF-kappaB family member RelB. Biochem J.2008;416:117–27.63. Wu KK. Control of cyclooxygenase-2 transcriptional activation bypro-inflammatory mediators. Prostaglandins Leukot Essent Fatty Acids.2005;72:89–93.64. Talikka M, Sierro N, Ivanov NV, Chaudhary N, Peck MJ, Hoeng J, et al. Genomicimpact of cigarette smoke, with application to three smoking-related diseases.Crit Rev Toxicol. 2012;42:877–89.65. Marwick JA, Kirkham PA, Stevenson CS, Danahay H, Giddings J, Butler K, et al.Cigarette smoke alters chromatin remodeling and induces proinflammatorygenes in rat lungs. Am J Respir Cell Mol Biol. 2004;31:633–42.66. McMillan DH, Woeller CF, Thatcher TH, Spinelli SL, Maggirwar SB, Sime PJ, et al.Attenuation of inflammatory mediator production by the NF-kappaB memberRelB is mediated by microRNA-146a in lung fibroblasts. Am J Physiol Lung CellMol Physiol. 2013;304:L774–81.67. Barnes PJ. Role of HDAC2 in the pathophysiology of COPD. Annu RevPhysiol. 2009;71:451–64.Submit your manuscript at www.biomedcentral.com/submit