IκBβ-Mediated NF-κB Activation Confers Protection against ... Recipient...ORIGINAL RESEARCH IkBb-Mediated NF-kB Activation Confers Protection against Hyperoxic Lung Injury Katherine

ORIGINAL RESEARCH

IkBb-Mediated NF-kB Activation Confers Protection againstHyperoxic Lung InjuryKatherine A Michaelis1 Fadeke Agboke2 Thanh Liu1 Kristie Han1 Manasa Muthu2 Csaba Galambos1 Guang Yang2Phyllis A Dennery23 and Clyde J Wright1

1Section of Neonatology Department of Pediatrics University of Colorado School of Medicine Aurora Colorado 2Department ofPediatrics Division of Neonatology Childrenrsquos Hospital of Philadelphia Philadelphia Pennsylvania and 3Department of PediatricsUniversity of Pennsylvania School of Medicine Philadelphia Pennsylvania

Abstract

Supplemental oxygen is frequently used in an attempt to improveoxygen delivery however prolonged exposure results in damage tothe pulmonary endothelium and epithelium Although NF-kB hasbeen identified as a redox-responsive transcription factor whetherNF-kB activation exacerbates or attenuates hyperoxic lung injury isunclear We determined that sustained NF-kB activity mediated byIkBb attenuates lung injury and prevents mortality in adult miceexposed to greater than 95O2 Adult wild-type mice demonstratedevidence of alveolar protein leak and 100 mortality by 6 days ofhyperoxic exposure and showed NF-kB nuclear translocation thatterminated after 48hours Furthermore thesemice showed increasedexpression of NF-kBndashregulated proinflammatory and proapoptoticcytokines In contrast mice overexpressing the NF-kB inhibitoryprotein IkBb (AKBI) demonstrated significant resistance tohyperoxic lung injury with 50 surviving through 8 days ofexposure This was associated with NF-kB nuclear translocation thatpersisted through 96 hours of exposure Although induction of NF-kBndashregulated proinflammatory cytokines was not different between

wild-type and AKBI mice significant up-regulation of antiapoptoticproteins (BCL-2 BCL-XL) was found exclusively in AKBI mice Weconclude that sustained NF-kB activity mediated by IkBb protectsagainst hyperoxic lung injury through increased expression ofantiapoptotic genes

Keywords NF-kB IkB hyperoxic lung injury apoptosisinflammation

Clinical Relevance

Oxygen therapy is commonly employed in an attempt toimprove oxygen delivery and yet prolonged exposure resultsin lung injury No pharmacologic strategies to attenuatehyperoxic lung injury have been identified We show thatenhancing hyperoxia-induced NF-kB activity attenuateslung injury and thus represents a potential therapeutictarget

The first studies demonstrating the lethal effectof inspiring high concentrations of oxygenwere published over a century ago (1) Due tothis toxicity hyperoxic exposure has beenwidely used in animal studies to induce andstudy acute lung injury (2) Despite this nopharmacologic therapies to attenuatehyperoxic lung injury have been identified

Hyperoxia induces proinflammatorycytokine expression (3) and results in

pulmonary endothelial and epithelial celldeath (4) Given these facts both anti-inflammatory and antiapoptotic strategieshave been employed to attenuate hyperoxiclung injury Although anti-inflammatoryinterventions have met variable success(5ndash9) increasing the expression ofantiapoptotic factors in transgenic murinemodels results in attenuated hyperoxic lunginjury (10ndash12) These studies suggest that

preventing hyperoxia-induced apoptosiscan prevent lung injury but thetranscriptional mechanism mediatingexpression of these factors is largelyunknown

Hyperoxia-induced NF-kB activationhas been documented in both pulmonaryepithelial and endothelial cells (13)Whether this response is protective orinjurious is unclear Some in vivo studies

(Received in original form July 2 2013 accepted in final form August 7 2013 )

This work was supported by National Institutes of Health grants K08 HL098562 (CJW) and HL058752-11 (PAD)

Correspondence and requests for reprints should be addressed to Clyde J Wright MD Section of Neonatology Department of Pediatrics University ofColorado School of Medicine Mail Stop 8614 Aurora CO 80045 E-mail clydewrightucdenveredu

This article has an online supplement which is accessible from this issuersquos table of contents at wwwatsjournalsorg

Am J Respir Cell Mol Biol Vol 50 Iss 2 pp 429ndash438 Feb 2014

Copyright copy 2014 by the American Thoracic Society

Originally Published in Press as DOI 101165rcmb2013-0303OC on September 25 2013

Internet address wwwatsjournalsorg

Michaelis Agboke Liu et al IkBb and Hyperoxic Lung Injury 429

suggest that intact NF-kB signaling confersprotection against hyperoxia-induced lunginjury (14 15) whereas inhibition ofoxidant stress-induced NF-kB activityexacerbates cell death (16ndash19) In contrastNF-kB activity can potentiatea proinflammatory response to hyperoxia(20) Overall the role of NF-kB inmodulating the pulmonary response tohyperoxia is complicated For exampleNF-kB confers a protective response tooxidative stress through increasedexpression of antioxidant enzymes (egmanganese superoxide dismutase[MnSOD] glutathione peroxidase]) anti-inflammatory cytokines (IL-6 IL-11) andantiapoptotic genes (21) However NF-kBmay exacerbate lung injury throughexpression of both pro-oxidant enzymes(eg cyclo-oxygenase [COX]-2 reducednicotinamide adenine dinucleotidephosphate oxidase NOX2 [gp91 phox])and proinflammatory proteins (eg IL-1aCD40 CXCL1 and intercellular adhesionmolecule [ICAM]-1) (21) It is important toclarify whether NF-kB is cytoprotective ordetrimental in hyperoxia to devisetherapeutic strategies to obviate lung injury

The penultimate step of NF-kB nucleartranslocation and DNA binding involvesthe inhibitory proteins IkBa and IkBb(22) In quiescent cells IkB proteinssequester NF-kB in the cytoplasmExposure to inflammatory stress (LPSTNF-a) results in phosphorylation anddegradation of these proteins allowing NF-kB nuclear translocation Importantlyalthough both IkBa and IkBb inhibit NF-kB activation signaling mediated by theseproteins results in specific target geneexpression Because IkBa and IkBbpreferentially bind unique NF-kB dimercombinations and these dimercombinations bind to unique DNAsequences different downstream genes aretargeted (23) In addition after degradationboth newly synthesized IkBa and IkBbenter the nucleus A nuclear exportsequence found on IkBa allows it to exportDNA-bound NF-kB complexes from thenucleus (24) In contrast IkBb lacksa nuclear export sequence (25) andremains in the nucleus facilitatingsustained binding of NF-kB dimers to DNA(26) These differences have importantimplications for the expression of NF-kBtarget genes as the duration of NF-kBnuclear localization plays a role indetermining which target genes are

expressed (27) Despite the knowndifferences between IkBa and IkBb theunique roles of these proteins inmodulating hyperoxia-induced NF-kBactivation in the lung remains unexploredIn addition although studies of hyperoxia-induced NF-kB activation have investigatedthe role of IkBa no studies have evaluatedthe role of IkBb

Using AKBI mice in which IkBbcDNA has been inserted behind the IkBapromoter and thus overexpress IkBb anddo not express IkBa (28) we demonstratethat prolonged NF-kB signaling enhancesimproved survival after exposure tohyperoxia This protection occurs despiteenhanced expression of proinflammatorycytokines and is likely secondary toincreased expression of protective cytokinesand antiapoptotic genes These findingssuggest that strategies aimed at enhancingNF-kB activity may represent a therapeutictarget to prevent hyperoxic lung injury

Materials and Methods

Animal ModelICR mice were purchased from Taconic andAKBI or IkBb knockin mice were a gift ofRichard Cohen (Harvard UniversityCambridge MA) Adult mice (6ndash8 wk old)were exposed to room air or hyperoxia(O2 95) in an A-chamber (BioSpherixRedfield NY) Ambient carbon dioxide wasmaintained at under 1200 ppm Allprocedures were approved by theInstitutional Animal Care and UseCommittee at the Childrenrsquos Hospital ofPhiladelphia (Philadelphia PA)

Bronchoalveolar Lavage Analysis ofLeukocytes and Differential Proteinand CytokinesAfter exposure to hyperoxia mice wereanesthetized killed bronchoalveolar lavagefluid (BALF) obtained and fluid analyzed(29) Total protein content was determinedby the Bradford method with bovine serumalbumin as a standard BALF was assessedfor IL-6 IL-1b and IL-11 content usingELISA (IL-6 88-7064 IL-1b 88-7013[eBioscience San Diego CA] IL-11RAB0251 [Sigma-Aldrich St Louis MO])

Preparation of Cytosolic andNuclear ExtractsFreshly collected whole lungs were collectedand cytosolic andnuclear extracts prepared (15)

Immunoblot AnalysisCytosolic and nuclear extracts wereelectrophoresed on a 4ndash12 polyacrylamidegel (Invitrogen Carlsbad CA) and proteinswere transferred to an Immobilonmembrane (Millipore Billerica MA)Membranes were blotted with anti-IkBa(sc-371) anti-IkBb (sc-9130) anti-BAX(sc-493) antindashBCL-2 (sc-492) orantindashlamin B (sc-6216) antibody (SantaCruz Biotechnology Dallas TX) antindashc-Rel(4774) anti-IKKa (2682) anti-IKKb(2370) antindashvascular endothelial growthfactor receptor (VEGFR) 2 (2479) orantindashBCL-XL (2762) antibody (CellSignaling Technology Danvers MA)anti-p50 (ab7971) antibody (AbcamCambridge MA) antindashMn-SOD (06-984)or anti-tubulin antibody (05-829Millipore) Densitometric analysis wasperformed using ImageLab (Bio-RadHercules CA)

Evaluation of NF-kB Binding byElectrophoretic Mobility Shift AssayA 32P-labeled oligonucleotide with theconsensus sequence for NF-kB (59-AGTTGAGGGGACTTTCCCAGGC-39Promega Madison WI) was used toevaluate NF-kB binding ability (30) Toidentify nonspecific binding of nuclearproteins competition reactions wereperformed by addition of 50-fold excess ofthe nonradiolabeled NF-kB consensussequence to the reaction mixtures beforeelectrophoresis

Terminal DeoxynucleotidylTransferase dUTP Nick EndLabeling StainingLung sections were assessed for apoptosisusing the DeadEndFluorometric terminaldeoxynucleotidyl transferase dUTP nick endlabeling (TUNEL) system (Promega)Digital images were obtained using anOlympus IX71 fluorescence microscope(Olympus America Center Valley PA) andquantitation performed on five high-magnification fields per section withsections from three separate animalsassessed per time point

Quantitative Real-Time PCRRelative mRNA levels were evaluated byquantitative real-time PCR using theTaqMan gene expression system (AppliedBiosystems Carlsbad CA) Total lung RNAwas extracted with TRIzol Plus RNAPurification System (Ambion Carlsbad

ORIGINAL RESEARCH

430 American Journal of Respiratory Cell and Molecular Biology Volume 50 Number 2 | February 2014

CA) RNA was assessed for purity andconcentration using the NanoDrop(ThermoFisher Scientific Waltham MA)and cDNA synthesized using the VersocDNA Synthesis Kit (ThermoFisherScientific) Gene expression of CXCL1COX-2 ICAM MnSOD BAX BCL-XLBCL-2 VEGFR2 and IL-11 was assessedusing predesigned exon-spanning primers(Mm04207460_m1 Mm00478374_m1Mm00516023_m1 Mm00449726_m1Mm00432050_m1 Mm00437783_m1Mm00477631_m1 Mm01222421_m1 andMm00434162_m1 Applied Biosystems)using the StepOnePlus Real Time PCRSystem (Applied Biosystems) Relativequantitation was performed vianormalization to the endogenous controlb-2-microglobulin (cat no 433766FApplied Biosystems) using the cyclethreshold (DDCT) method

Statistical AnalysisFor comparison between treatment groupsthe null hypothesis that no difference existedbetween treatment means was tested by

ANOVA for multiple groups or t test fortwo groups (InStat GraphPad SoftwareInc San Diego CA) Statistical significance(P 005) between and within groups wasdetermined by means of Bonferronirsquosmethod of multiple comparisons

Results

AKBI Mice Are Resistant toHyperoxia-Induced MortalityTo evaluate the specific role of the IkB familyof proteins in NF-kB signaling IkBb knockin(AKBI) mice have been derived (Figure 1A)The AKBI mice have IkBb cDNA insertedbehind the IkBa promoter and arephenotypically indistinct from wild-type(WT) littermate control mice (28) In thisstudy AKBI mice had significantly improvedsurvival after exposure to hyperoxia whencompared with WT mice (Figure 1B) Bothgroups had 100 survival in normoxiaHowever by 5 days of hyperoxia WT micedemonstrated 100 mortality whereas AKBImice had over 90 survival Animals were

weighed daily as a measure of intake andoverall health (Figure 1C) Although WT andAKBI mice had similar weights beforehyperoxic exposure by 96 hours thesignificant weight loss observed in WT micewas absent in AKBI mice Given thesefindings animals were exposed to 96 hours ofhyperoxia for other assays

AKBI Mice Have Less Vascular Leakafter Hyperoxic ExposureIn normoxia WT and AKBI mice had similarBALF protein content Although WT micehad a 20-fold increase in BALF proteincontent after 96 hours of hyperoxic exposureAKBI mice showed no significant increase(see Figure E1A in the online supplement)This suggested decreased oxygen toxicity inthe AKBI compared with WT mice

WT and AKBI Mice Have Similar BALFCellular Content afterHyperoxic ExposureIn normoxia WT and AKBI mice hadsimilar BALF cellular content (Figure E1B)No significant difference in the total cellular

Figure 1 AKBI or IkBb knockin mice have improved survival after hyperoxic exposure (A) Schematic of IkB expression patterns in AKBI mice The IkBagene has been replaced by IkBb cDNA The IkBa promoter controls the expression of the IkBb transgenic loci Thus AKBI overexpresses IkBbwithout expressing IkBa (B) Kaplan-Meier survival analysis of wild-type (WT) and AKBI mice exposed to chronic hyperoxia ( 95) Values are expressedas the percentage of surviving animals (n = 16group) (C) Individual mouse weights through hyperoxic exposure Data expressed as means 6 SE (n =4time point) P 005 versus paired WT exposure

ORIGINAL RESEARCH

Michaelis Agboke Liu et al IkBb and Hyperoxic Lung Injury 431

BALF content or differential developedbetweenWT and AKBI mice throughout thehyperoxic exposure (Figure E1B) Furtherevaluation demonstrated that there was nosignificant increase in the difference inBALF macrophage numbers between WTand AKBI mice (Figure E1C) Both WT andAKBI mice had similar and significantlyincreased neutrophil BALF counts after96 hours of hyperoxia (Figure E1D)

WT and AKBI Mice Have SimilarPulmonary Expression of KeyNF-kB Pathway ProteinsTo evaluate the effect of the absence of IkBaandor overexpression of IkBb on proteinsinvolved in NF-kB signaling Western blotswere performed on whole-lung lysates fromWT and AKBI mice As expected AKBI micedid not express IkBa and showed IkBboverexpression when compared with WT(Figure 2A) No differences were found in theexpression of p50 p65 c-Rel IKKa and IKKbbetween WT and AKBI mice (Figure 2A)

Hyperoxia Decreases IkBbExpression Both WT and AKBI LungTo define the signaling events related toNF-kB activation levels of immunoreactive

IkBa and IkBb were evaluated in cytosolicextracts from whole-lung homogenateAfter exposure to hyperoxia IkBa in WTmice demonstrated a cyclical pattern ofdegradation and reaccumulation through96 hours of exposure (Figure 2B) Decreasesin IkBb seen in the lungs of WT mice afterexposure to 96 hours of hyperoxia were notsignificant but similar to IkBa decreasedwith prolonged exposure (Figures 2B and2D) In contrast in AKBI mice IkBbdecreased significantly throughout the 96hours of hyperoxic exposure (Figures 2Band 2D) These data suggest that hyperoxiapromotes IkB degradation and thatAKBI mice have NF-kB signaling mediatedsolely through IkBb degradation Thissuggests that IkBb-mediated NF-kBactivation dictates a protective response tohyperoxia given the difference in survivalbetween WT and AKBI mice

AKBI Mice Have Prolonged Nuclearp65 Translocation after HyperoxiaHaving noted significant decreases in lungIkBb levels in AKBI mice exposed tohyperoxia nuclear extracts were assessedfor translocation of the NF-kB subunit

p65 Western analysis revealed significantincreases in nuclear p65 in the lungs of WTmice at 48 hours which returned tobaseline by 72 hours In contrast nuclearp65 increased in the lungs of AKBI miceafter 48 hours of hyperoxia and remainedelevated through 96 hours of exposure(Figures 3A and 3B) Increased nucleartranslocation of p65 at 96 hours of exposurewas associated with increased NF-kBconsensus sequence binding asdemonstrated by electrophoretic mobilityshift assay (Figure 3C) These data suggesta prolonged and sustained level ofhyperoxia-induced NF-kB activity in AKBImice that was not present in WT mice

WT and AKBI Mice HaveSimilar Induction ofNF-kBndashRegulated CytokinesExpression of IL-6 and IL-1b is known to beregulated by NF-kB activity (21) andelevated in murine models of hyperoxiclung injury (3) Interestingly IL-6expression attenuates hyperoxic lung injury(31) whereas expression of theproinflammatory IL-1b is thought toexacerbate lung injury (3) Levels of IL-1b

Figure 2 WT and AKBI mice have similar expression of key NF-kB pathway proteins and hyperoxia induces IkB protein degradation in WT and AKBIlung (A) Representative Western analysis of the NF-kB subunits p50 p65 and c-Rel the NF-kB inhibitory proteins IkBa and IkBb and NF-kBkinases IKKa and IKKb from WT and AKBI whole-lung homogenate (B) Representative Western blot showing IkBa and IkBb in whole-lung homogenatefrom WT and AKBI mice exposed to room air or hyperoxia (O2 95 24ndash96 h) with tubulin as loading control Densitometric evaluation of (C) IkBa and(D) IkBb Values are means 6 SEM (n = 4time point) P 005 versus unexposed control

ORIGINAL RESEARCH

432 American Journal of Respiratory Cell and Molecular Biology Volume 50 Number 2 | February 2014

and IL-6 were assessed in BALF obtainedfrom WT and AKBI mice exposed tohyperoxia for 0 24 48 72 and 96 hours Bothcytokines were significantly elevated byprolonged exposure to hyperoxia in the BALFobtained from AKBI mice when comparedwith WT mice (Figures 4A and 4B) Inaddition we assessed whole-lung mRNAexpression of the NF-kBndashregulated proteinsCXCL1 COX-2 and ICAM using quantitativereal-time RT-PCR (21) Levels of CXCL1 andCOX-2 mRNA were significantly elevated inboth WT and AKBI mice exposed tohyperoxia (Figures 4C and 4D) In additionthere was a trend toward increased ICAMexpression that did not reach significance(Figure 4E) These results are consistent withenhanced NF-kB activation and suggest thatincreased expression of proinflammatory NF-kBndashregulated targets does not exacerbatehyperoxic lung injury in AKBI mice

Antiapoptotic and Antioxidant NF-kBTarget Gene Expression Is Enhancedin AKBI MiceThe expression of NF-kBndashregulated pro-and antiapoptotic proteins was assessed on

the basis of previous reports linking themto hyperoxic exposure (10 32 33)Consistent with hyperoxia-induced NF-kB activation both WT and AKBI micedemonstrated induction of BAX mRNAand protein expression at 96 hours ofexposure (Figures 5Andash5C) Whereas noincrease in BCL-XL or BCL-2 mRNA orprotein was observed in the lungs of WTmice exposed to hyperoxia (Figures5Dndash5I) these were significantly increasedin the lungs of AKBI mice (Figures5Dndash5I)

In addition to BCL-2 and BCL-XLother NF-kBndashregulated proteins are knownto protect against hyperoxic lung injurySpecifically NF-kB regulates the inductionof IL-11 and MnSOD as well asconstitutive expression of VEGFR2 (21 34)Expression of these factors attenuateshyperoxic lung injury (11 12 35 36)Hyperoxia resulted in decreased expressionof VEGFR2 and no induction of IL-11expression in WT mice (Figures 6Andash6E) Incontrast hyperoxia-exposed AKBI micedemonstrated a significant up-regulation ofboth VEGFR2 and IL-11 mRNA and

protein (Figures 6Andash6E) In addition wefound that hyperoxia significantly inducedMnSOD mRNA and protein expressionin AKBI compared with WT mice(Figures 6Fndash6H) Together with theincreased expression of NF-kBndashregulatedantiapoptotic proteins these results suggestthat the prolonged hyperoxia-inducedNF-kB nuclear translocation seen in AKBImice tips the balance toward increasedcell survival and confers resistance tohyperoxic lung injury

Hyperoxia Increases TUNEL Stainingin WT MiceApoptosis was assessed by TUNEL stainingof WT and AKBI adult mouse lungs at96 hours of hyperoxia exposure In agreementwith the expression profile of antiapoptoticgenes hyperoxia induced apoptosis in theWT lung (Figure 7) In contrast there wasno significant increase in the number ofcells staining positive in the AKBI lungexposed to hyperoxia (Figure 7) Theseresults confirm increased apoptosis in theWT lung exposed to hyperoxia a findingthat was attenuated in the AKBI mice

Figure 3 Hyperoxia-induced NF-kB activation is prolonged in AKBI mice (A) Representative Western blot showing p65 in lung nuclear extracts from WTand AKBI mice exposed to room air or hyperoxia (O2 95 24ndash96 hr) with Lamin B as loading control (B) Densitometric evaluation of nuclear p65Values are means 6 SEM (n = 4time point) P 005 versus unexposed control daggerP 005 versus paired WT exposure (C) Representativeelectrophoretic mobility shift assay of nuclear extracts from WT and AKBI lung after exposure to room air or hyperoxia (O2 95 96 h) Bandsrepresenting NF-kB consensus sequence binding and free probe are labeled Lane 1 WT unexposed lanes 2ndash5 WT exposed to 96 hours hyperoxia lane6 WT cold WT exposed to 96-hour hyperoxia plus 50-fold excess of unlabeled oligonucleotide lane 7 AKBI unexposed lanes 8ndash11 AKBI exposed to96-hour hyperoxia lane 12 AKBI cold AKBI exposed to 96-hour hyperoxia plus 50-fold excess of unlabeled oligonucleotide

ORIGINAL RESEARCH

Michaelis Agboke Liu et al IkBb and Hyperoxic Lung Injury 433

Discussion

We found that hyperoxia-induced NF-kBactivity occurs in both WT and AKBI miceIn WT mice this induction occurs over thefirst 48 hours of exposure and results in theexpression of proinflammatory andproapoptotic proteins This is associatedwith significant mortality that occurs within5 days of exposure In contrast hyperoxia-induced NF-kB activity in AKBI micepersists through 96 hours of exposureAlthough proinflammatory andproapoptotic genes are expressed at levelsequivalent to and in some cases in excess

of WT mice up-regulation of antioxidantand antiapoptotic proteins occurs withprolonged exposure This results inattenuated lung injury and improvedsurvival Thus targeting hyperoxia-inducedNF-kB signaling represents a potentialtherapeutic target to limit lung injury

These results are interesting becausethis is the first study showing that enhancedNF-kB activation mediated by IkBboverexpression confers resistance tohyperoxic lung injury Previous studieshave shown that NF-kB activity preventscell death induced by various types ofoxidant stress These include

ischemiandashreperfusion to the intestine andheart (16 17 19) oxidant stress induced byH2O2 in A549 cells (18) and glucoseoxidase in murine embryonic fibroblastcells (37) In addition enhanced hyperoxia-induced NF-kB activation explains theresistance to lung injury seen in neonatalmice in part due to inhibition of apoptosisPrevious studies have demonstrated thathyperoxic exposure results in apoptosis ofpulmonary endothelial and epithelial cells(32 38) Consistent with this findingpulmonary gene expression array analysisof mice exposed to hyperoxia showsinduction of both pro- (BAX) and

Figure 4 Hyperoxia induces cytokine expression in both WT and AKBI mice (A) IL-1b and (B) IL-6 levels in bronchoalveolar lavage fluid (BALF) obtainedfrom WT and AKBI mice exposed to hyperoxia assessed by ELISA (96 h n = 4time point) Data expressed as means 6 SE P 005 versus unexposedcontrol daggerP 005 versus paired WT exposure Expression of (C) CXCL1 (D) cyclo-oxygenase (COX)-2 and (E) intercellular adhesion moleculemRNA in WT and AKBI lung after exposure to hyperoxia (96 h n = 4time point) Data expressed as means 6 SE P 005 versus unexposed control

ORIGINAL RESEARCH

434 American Journal of Respiratory Cell and Molecular Biology Volume 50 Number 2 | February 2014

antiapoptotic (BCL-XL) proteins (33) Ofnote BAX BCL-XL and BCL-2 expressionis regulated by NF-kB (24) We found thatthe prolonged hyperoxia-induced NF-kBactivation seen in AKBI mice wasassociated with increased expression ofboth BCL-2 and BCL-XL and this wasassociated with attenuated apoptosis andimproved survival Although p5322fas22 and tnfr22 mice do notdemonstrate resistance to hyperoxic lunginjury (32 39) specific overexpression ofproteins that contain four Bcl-2 homologydomains (BCL-2 BCL-XL) and preventmitochondrial permeabilization attenuateshyperoxic lung injury whereas their

absence increases sensitivity to oxygentoxicity (10 40) Our study supports thefinding that enhancing expression ofantiapoptotic proteins via NF-kB activationto prevent mitochondrial permeabilizationprevents apoptosis and attenuateshyperoxic lung injury

In this study we evaluated theexpression of various NF-kBndashregulatedproteins pro- and anti-inflammatorycytokines and antioxidant enzymes Thesetargets were selected on the basis ofprevious reports linking these proteins tothe pulmonary response to hyperoxicinjury Previous in vivo studies have shownthat expression of MnSOD (36) IL-6 (31)

and VEGF signaling mediated throughVEGFR2 (11 35) attenuate hyperoxic lunginjury Expression of each of theseindividual factors is regulated by NF-kB(21) Here we have demonstrated that theenhanced hyperoxia-induced NF-kBactivity seen in AKBI mice is associatedwith increased expression of all of theseprotective factors Importantly NF-kBregulates the expression of additionalproteins known to attenuate hyperoxic lunginjury but not evaluated in the currentstudy These include hemeoxygenase andp21 (41 42) Of note the protectionafforded by p21 was recently demonstratedto be mediated via BCL-XL one of theNF-kBndashregulated genes assessed in thecurrent study (43) Thus targeting eventsupstream of NF-kB activation to enhanceduration of activity and thus affectingthe expression of multiple cytoprotectivegenes represents a potential therapeutictarget to be tested in future studies

Furthermore our results areconsistent with other reportsdemonstrating that disrupting signalingevents upstream of NF-kB activationexacerbates hyperoxic lung injury Theincreased susceptibility to hyperoxic lunginjury seen in TLR422 mice is associatedwith a more transient activation of NF-kB when compared with WT controlanimals as well as attenuated Aktsignaling and BCL-2 expression (14 44)Importantly Akt signaling plays a role inNF-kB activation (22) and constitutivelyactive Akt prevents hyperoxic lung injury(45) Placed in context with these previousreports our results argue for a centralrole played by NF-kB activation inattenuating hyperoxic lung injury

Hyperoxia results in increasedpulmonary expression of NF-kBndashregulatedproapoptotic proteins (BAX [46])proinflammatory cytokines (IL-1b CXCL1[3 47]) and pro-oxidant enzymes (COX-2[48]) We found hyperoxia-inducedinduction of BAX IL-1b CXCL1 andCOX-2 in both WT and AKBI mice Ourfindings show that despite expression ofthese proinflammatory proapoptoticand pro-oxidant factors AKBI micedemonstrate resistance to hyperoxic lunginjury We speculate that despiteexpression of proinflammatory signalspreventing apoptosis modulates lung injuryand confers a survival advantage tohyperoxia-exposed mice These findings areimportant as targeting the transcriptional

Figure 5 Hyperoxia induces NF-kBndashregulated antiapoptotic gene expression only in AKBI mice(A) Representative Western blot showing BAX from whole-lung homogenate from WT and AKBImice exposed to hyperoxia (O2 95 96 h) (B) Densitometric evaluation of BAX Values aremeans 6 SEM (n = 4time point) P 005 versus unexposed control (C) Pulmonary BAX mRNAexpression in WT and AKBI mice exposed to room air or hyperoxia (O2 95 96 h) Values aremeans 6 SEM (n = 4time point) P 005 versus unexposed control (D) Representative Westernblot showing BCL-XL from whole-lung homogenate from WT and AKBI mice exposed to hyperoxia(O2 95 96 h) (E) Densitometric evaluation of BCL-XL Values are means 6 SEM (n = 4timepoint) P 005 versus unexposed control and paired WT exposure (F) Pulmonary BCL-XL mRNAexpression in WT and AKBI mice exposed to room air or hyperoxia (O2 95 96 h) Values aremeans 6 SEM (n = 4time point) P 005 versus unexposed control daggerP 005 versus paired WTexposure (G) Representative Western blot showing BCL-2 from whole-lung homogenate from WTand AKBI mice exposed to hyperoxia (O2 95 96 h) (H) Densitometric evaluation of BCL-2Values are means 6 SEM (n = 4time point) P 005 versus unexposed control and paired WTexposure (I) Pulmonary BAX mRNA expression in WT and AKBI mice exposed to room air orhyperoxia (O2 95 96 h) Values are means 6 SEM (n = 4time point) P 005 versusunexposed control and paired WT exposure

ORIGINAL RESEARCH

Michaelis Agboke Liu et al IkBb and Hyperoxic Lung Injury 435

activity of NF-kB rather than one specificgene target represents a fundamentallysimpler therapeutic intervention It is clearthat hyperoxia-induced NF-kB activityoccurs in WT mice and that maintainingthis activity prevents lung injury andmortality Our results demonstrate thatalthough hyperoxia-induced NF-kBactivation affects the expression of bothprotective and injurious factors the netresult of prolonged NF-kB activity isattenuated lung injury

The possible mechanism underlyingour finding is that IkBb-mediated sustainedNF-kB activity tips the balance of geneexpression to include both antioxidant andantiapoptotic targets In quiescent cellsNF-kB remains sequestered in thecytoplasm bound to members of the IkBfamily of inhibitory proteins (22) Exposureto inflammatory stimuli results inphosphorylation of two N-terminal serineresidues on both IkBa and IkBb resultingin their degradation (22) IkBa is

transcriptionally regulated by NF-kBallowing for a well regulated negativefeedback loop that is both sensitive to andrapidly influenced by NF-kB activation(49) Newly synthesized IkBa enters thenucleus and removes DNA-bound NF-kBcomplexes (24) In contrast to IkBaIkBb is degraded more slowly it is nottranscriptionally regulated by NF-kB andafter degradation induced by inflammatorystimuli reaccumulates asa hypophosphorylated form

Figure 6 Hyperoxia induces expression of vascular endothelial growth factor receptor (VEGFR) 2 IL-11 and manganese superoxide dismutase (MnSOD)in AKBI mice (A) Representative Western blot showing VEGFR2 from whole-lung homogenate from WT and AKBI mice exposed to hyperoxia (O2 9596 h) (B) Densitometric evaluation of VEGFR2 Values are means 6 SEM (n = 4time point) P 005 versus unexposed control daggerP 005 versuspaired WT exposure (C) Pulmonary VEGFR mRNA expression in WT and AKBI mice exposed to room air or hyperoxia (O2 95 96 h) Values aremeans 6 SEM (n = 4time point) P 005 versus unexposed control and paired WT exposure (D) IL-11 fold increase in BALF obtained from WT andAKBI mice exposed to hyperoxia assessed by ELISA (72ndash96 h n = 4time point) Data expressed as means 6 SE P 005 versus unexposedcontrol and paired WT exposure (E) IL-11 mRNA in WT and AKBI lung after exposure to hyperoxia (96 h n = 4time point) Data expressed as means 6SE P 005 versus unexposed control and paired WT exposure (F) Representative Western blot showing MnSOD from lung cytosolic extracts from WTand AKBI mice exposed to room air or hyperoxia (O2 95 96 h) with calnexin as a loading control (G) Densitometric evaluation of MnSODValues are means 6 SEM (n = 4time point) P 005 versus unexposed control daggerP 005 versus paired WT exposure (H) Pulmonary MnSOD mRNAexpression in WT and AKBI mice exposed to room air or hyperoxia (O2 95 96 h) Values are means 6 SEM (n = 4time point) P 005 versusunexposed control daggerP 005 versus paired WT exposure

ORIGINAL RESEARCH

436 American Journal of Respiratory Cell and Molecular Biology Volume 50 Number 2 | February 2014

Hypophosphorylated IkBb chaperonesNF-kB dimers and facilitates DNA binding(26) In contrast to the well definedNF-kB activation cascade that occurs afterexposure to inflammatory stressa definitive pathway after exposure tooxidant stress has not been establishedWhether these same mechanisms areresponsible for prolonged NF-kB activationseen in hyperoxia-exposed AKBI miceremains to be determined

Our study is limited by the use of AKBImice From our work it is clear thatoverexpression of IkBb and lack of IkBaaffects baseline expression of NF-kBtargets including IL-1b and IL-6 (Figures6A and 6B) This is likely true for other NF-kB targets These findings are in agreementwith studies of IkBb22 mice that showincreased constitutive NF-kB activity (23)Despite this difference the AKBI mice arephenotypically indistinct from their WT

control animals and demonstrate normallung architecture In addition we have notidentified the specific cell types protectedfrom apoptosis by prolonged NF-kBactivation Whether our findings are due toenhanced NF-kB activation in resident lungcells or inflammatory cells recruited tothe lung with ongoing hyperoxic injuryremains to be determined Future studiesidentifying the specific cell type protectedfrom apoptosis and evaluating the effect ofenhancing hyperoxia-induced NF-kBactivity in WT mice are needed to confirmour results

We conclude that enhancedhyperoxia-induced NF-kB activityprevents hyperoxic lung injury andmortality in vivo Specifically NF-kBactivation results in increased expressionof cytoprotective factors includingMnSOD and antiapoptotic genes (BCL2and BCL-XL) This protection occursdespite intact proinflammatory geneexpression Our results show that thetoxic effects of hyperoxia can beattenuated by targeting the inhibitoryproteins that dictate the duration of NF-kB activity and ultimately downstreamtarget gene expression We speculate thatinterventions aimed at enhancinghyperoxia-induced NF-kB activationcould attenuate lung injury n

Author disclosures are available with the textof this article at wwwatsjournalsorg

References

1 Smith JL The pathological effects due to increase of oxygen tension inthe air breathed J Physiol 18992419ndash35

2 Matute-Bello G Frevert CW Martin TR Animal models of acute lunginjury Am J Physiol Lung Cell Mol Physiol 2008295L379ndashL399

3 Johnston CJ Wright TW Reed CK Finkelstein JN Comparison of adultand newborn pulmonary cytokine mRNA expression after hyperoxiaExp Lung Res 199723537ndash552

4 Crapo JD Morphologic changes in pulmonary oxygen toxicity Annu RevPhysiol 198648721ndash731

5 Tsan MF White JE Michelsen PB Wong GH Pulmonary O2 toxicity role ofendogenous tumor necrosis factor Exp Lung Res 199521589ndash597

6 Raj JU Hazinski TA Bland RD Oxygen-induced lung microvascularinjury in neutropenic rabbits and lambs J Appl Physiol 198558921ndash927

7 Guthmann F Wissel H Rustow B Early subcutaneous administration ofetanercept (enbrel) prevents from hyperoxia-induced lung injury ExpLung Res 200935770ndash780

8 Husari AW Khayat A Awdeh H Hatoum H Nasser M Mroueh SMZaatari G El-Sabban M Dbaibo GS Activated protein C attenuatesacute lung injury and apoptosis in a hyperoxic animal model Shock201033467ndash472

9 Yamada M Kubo H Kobayashi S Ishizawa K Sasaki H Interferon-gamma a key contributor to hyperoxia-induced lung injury in miceAm J Physiol Lung Cell Mol Physiol 2004287L1042ndashL1047

10 Waxman AB Kolliputi N IL-6 protects against hyperoxia-inducedmitochondrial damage via BCL-2ndashinduced BAK interactions withmitofusins Am J Respir Cell Mol Biol 200941385ndash396

11 He CH Waxman AB Lee CG Link H Rabach ME Ma B Chen QZhu Z Zhong M Nakayama K et al BCL-2ndashrelated protein A1is an endogenous and cytokine-stimulated mediator ofcytoprotection in hyperoxic acute lung injury J Clin Invest 20051151039ndash1048

12 Waxman AB Einarsson O Seres T Knickelbein RG Warshaw JBJohnston R Homer RJ Elias JA Targeted lung expression ofinterleukin-11 enhances murine tolerance of 100 oxygen anddiminishes hyperoxia-induced DNA fragmentation J Clin Invest19981011970ndash1982

13 Gore A Muralidhar M Espey MG Degenhardt K Mantell LL Hyperoxiasensing from molecular mechanisms to significance in diseaseJ Immunotoxicol 20107239ndash254

14 Qureshi ST Zhang X Aberg E Bousette N Giaid A Shan P MedzhitovRM Lee PJ Inducible activation of TLR4 confers resistance tohyperoxia-induced pulmonary apoptosis J Immunol 20061764950ndash4958

15 Yang G Abate A George AG Weng YH Dennery PA Maturationaldifferences in lung NF-kappaB activation and their role in toleranceto hyperoxia J Clin Invest 2004114669ndash678

16 Lu X Liu H Wang L Schaefer S Activation of NF-kappaB is a criticalelement in the antiapoptotic effect of anesthetic preconditioning AmJ Physiol Heart Circ Physiol 2009296H1296ndashH1304

Figure 7 Hyperoxia-induced apoptosis is attenuated in AKBI mice (A) High-magnification (403) imagesof representative terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL)- (green) and 496-diamidino-2-phenylindole (blue) -stained lung sections of WT and AKBI mice exposed to chronichyperoxia (O2 95 0 or 96 h) (B) TUNEL-positive nuclei per high-powered field in lung sections fromWT and AKBI mice exposed to hyperoxia Values are means 6 SEM (n = 4 animalstime point with 5fieldsanimal) P 005 versus unexposed control daggerP 005 versus paired WT exposure

ORIGINAL RESEARCH

Michaelis Agboke Liu et al IkBb and Hyperoxic Lung Injury 437

17 Tahepold P Vaage J Starkopf J Valen G Hyperoxia elicitsmyocardial protection through a nuclear factor kappaBndashdependentmechanism in the rat heart J Thorac Cardiovasc Surg 2003125650ndash660

18 Franek WR Horowitz S Stansberry L Kazzaz JA Koo HC Li Y Arita YDavis JM Mantell AS Scott W et al Hyperoxia inhibits oxidant-induced apoptosis in lung epithelial cells J Biol Chem 2001276569ndash575

19 Chen LW Egan L Li ZW Greten FR Kagnoff MF Karin M The twofaces of IKK and NF-kappaB inhibition prevention of systemicinflammation but increased local injury following intestinalischemiandashreperfusion Nat Med 20039575ndash581

20 Suzuki Y Nishio K Takeshita K Takeuchi O Watanabe K Sato NNaoki K Kudo H Aoki T Yamaguchi K Effect of steroid onhyperoxia-induced ICAM-1 expression in pulmonary endothelialcells Am J Physiol Lung Cell Mol Physiol 2000278L245ndashL252

21 Gilmore T NF-kB target genes [internet] Boston 9MA BostonUniversity c2013 [accessed 2013 Oct 24] Available from httpwwwbuedunf-kbgene-resourcestarget-genes

22 Perkins ND Integrating cell-signalling pathways with NF-kappaB andIKK function Nat Rev Mol Cell Biol 2007849ndash62

23 Rao P Hayden MS Long M Scott ML West AP Zhang DOeckinghaus A Lynch C Hoffmann A Baltimore D et alIkappabbeta acts to inhibit and activate gene expression during theinflammatory response Nature 20104661115ndash1119

24 Huang TT Miyamoto S Postrepression activation of NF-kappaBrequires the amino-terminal nuclear export signal specific toIkappaBalpha Mol Cell Biol 2001214737ndash4747

25 Tam WF Sen R IkappaB family members function by differentmechanisms J Biol Chem 20012767701ndash7704

26 Suyang H Phillips R Douglas I Ghosh S Role of unphosphorylatednewly synthesized I kappa B beta in persistent activation of NF-kappa B Mol Cell Biol 1996165444ndash5449

27 Hoffmann A Levchenko A Scott ML Baltimore D The IkappaB-NF-kappaB signaling module temporal control and selective geneactivation Science 20022981241ndash1245

28 Cheng JD Ryseck RP Attar RM Dambach D Bravo R Functionalredundancy of the nuclear factor kappa B inhibitors I kappa B alphaand I kappa B beta J Exp Med 19981881055ndash1062

29 Liu Y Mei J Gonzales L Yang G Dai N Wang P Zhang P Favara MMalcolm KC Guttentag S et al IL-17A and TNF-alpha exertsynergistic effects on expression of CXCL5 by alveolar type II cellsin vivo and in vitro J Immunol 20111863197ndash3205

30 Yang G Madan A Dennery PA Maturational differences in hyperoxicAP-1 activation in rat lung Am J Physiol Lung Cell Mol Physiol 2000278L393ndashL398

31 Ward NS Waxman AB Homer RJ Mantell LL Einarsson O Du Y EliasJA Interleukin-6ndashinduced protection in hyperoxic acute lung injuryAm J Respir Cell Mol Biol 200022535ndash542

32 Barazzone C Horowitz S Donati YR Rodriguez I Piguet PF Oxygentoxicity in mouse lung pathways to cell death Am J Respir Cell MolBiol 199819573ndash581

33 Perkowski S Sun J Singhal S Santiago J Leikauf GD Albelda SMGene expression profiling of the early pulmonary response tohyperoxia in mice Am J Respir Cell Mol Biol 200328682ndash696

34 Iosef C Alastalo TP Hou Y Chen C Adams ES Lyu SC Cornfield DNAlvira CM Inhibiting NF-kappaB in the developing lung disruptsangiogenesis and alveolarization Am J Physiol Lung Cell Mol Physiol2012302L1023ndashL1036

35 Corne J Chupp G Lee CG Homer RJ Zhu Z Chen Q Ma B Du YRoux F McArdle J et al IL-13 stimulates vascular endothelial cellgrowth factor and protects against hyperoxic acute lung injury J ClinInvest 2000106783ndash791

36 Ho YS Vincent R Dey MS Slot JW Crapo JD Transgenic models for thestudy of lung antioxidant defense enhanced manganese-containingsuperoxide dismutase activity gives partial protection to B6C3 hybridmice exposed to hyperoxia Am J Respir Cell Mol Biol 199818538ndash547

37 Wright CJ Agboke F Muthu M Michaelis KA Mundy MA La P YangG Dennery PA Nuclear factor-kappaB (NF-kappaB) inhibitoryprotein IkappaBbeta determines apoptotic cell death followingexposure to oxidative stress J Biol Chem 20122876230ndash6239

38 Mantell LL Kazzaz JA Xu J Palaia TA Piedboeuf B Hall S RhodesGC Niu G Fein AF Horowitz S Unscheduled apoptosis duringacute inflammatory lung injury Cell Death Differ 19974600ndash607

39 Pryhuber GS OrsquoBrien DP Baggs R Phipps R Huyck H Sanz I NahmMH Ablation of tumor necrosis factor receptor type I (p55) altersoxygen-induced lung injury Am J Physiol Lung Cell Mol Physiol2000278L1082ndashL1090

40 Staversky RJ Vitiello PF Yee M Callahan LM Dean DA OrsquoReilly MAEpithelial ablation of BCL-XL increases sensitivity to oxygen withoutdisrupting lung development Am J Respir Cell Mol Biol 201043376ndash385

41 Otterbein LE Kolls JK Mantell LL Cook JL Alam J Choi AMExogenous administration of heme oxygenase-1 by gene transferprovides protection against hyperoxia-induced lung injury J ClinInvest 19991031047ndash1054

42 OrsquoReilly MA Staversky RJ Watkins RH Reed CK de Mesy Jensen KLFinkelstein JN Keng PC The cyclin-dependent kinase inhibitor p21protects the lung from oxidative stress Am J Respir Cell Mol Biol200124703ndash710

43 Wu YC OrsquoReilly MA BCL-X(L) is the primary mediator of p21 protectionagainst hyperoxia-induced cell death Exp Lung Res 20113782ndash91

44 Zhang X Shan P Qureshi S Homer R Medzhitov R Noble PW Lee PJCutting edge TLR4 deficiency confers susceptibility to lethal oxidantlung injury J Immunol 20051754834ndash4838

45 Lu Y Parkyn L Otterbein LE Kureishi Y Walsh K Ray A Ray PActivated Akt protects the lung from oxidant-induced injury anddelays death of mice J Exp Med 2001193545ndash549

46 Budinger GR Mutlu GM Urich D Soberanes S Buccellato LJ HawkinsK Chiarella SE Radigan KA Eisenbart J Agrawal H et al Epithelialcell death is an important contributor to oxidant-mediated acute lunginjury Am J Respir Crit Care Med 20111831043ndash1054

47 Sue RD Belperio JA Burdick MD Murray LA Xue YY Dy MC KwonJJ Keane MP Strieter RM CXCR2 is critical to hyperoxia-inducedlung injury J Immunol 20041723860ndash3868

48 Adawi A Zhang Y Baggs R Finkelstein J Phipps RP Disruption of theCD40ndashCD40 ligand system prevents an oxygen-induced respiratorydistress syndrome Am J Pathol 1998152651ndash657

49 Sun SC Ganchi PA Ballard DW Greene WC NF-kappa B controlsexpression of inhibitor I kappa B alpha evidence for an inducibleautoregulatory pathway Science 19932591912ndash1915

ORIGINAL RESEARCH

438 American Journal of Respiratory Cell and Molecular Biology Volume 50 Number 2 | February 2014