AttenuationofExperimentalColitisinGlutathionePeroxidase1andCatalaseDoubleKnockoutMicethroughEnhancingRegulatoryTCellFunctionHyung-RanKim1.,AnbokLee2.,Eun-JeongChoi1,Jeong-HaeKie3,WoosungLim2,HyeonKookLee2,Byung-InMoon2*,Ju-YoungSeoh1*1Department of Microbiology, Ewha Womans University Graduate School of Medicine, Seoul, Korea, 2Department of Pathology, National Health Insurance CooperationIlsanHospital, Koyang, Korea, 3DepartmentofSurgery, EwhaWomansUniversityGraduateSchool ofMedicine, Seoul, KoreaAbstractReactiveoxygenspecies(ROS)havebeenimplicatedintheprogressionofinflammatorydiseasesincludinginflammatoryboweldiseases(IBD). Meanwhile, severalstudiessuggestedtheprotectiveroleofROSinimmune-mediatedinflammatorydiseases, anditwasrecentlyreportedthatdextransodiumsulfate(DSS)-inducedcolitiswasattenuatedinmicewithanelevated level of ROS due to deficiency of peroxiredoxin II. Regulatory T cells (Tregs) are critical in the prevention of IBD andTregfunctionwasreportedtobecloselyassociatedwithROSlevel, butithasbeeninvestigatedonlyinloweredlevelsofROS so far. In the present study, in order to clarify the relationship between ROS level and Treg function, and their role in thepathogenesis of IBD, we investigated mice with an elevated level of ROS due to deficiency of both glutathione peroxidase(GPx)-1 and catalase (Cat) for the susceptibility of DSS-induced colitis in association with Treg function. The results showedthat DSS-induced colitis was attenuated and Tregs were hyperfunctional in GPx12/26Cat2/2mice. In vivo administrationof N-acetylcysteine (NAC) aggravated DSS-induced colitis and decreased Treg function to the level comparable to WT mice.Attenuated Th17 cell differentiation from na ve CD4+cells as well as impaired production of IL-6 and IL-17A by splenocytesuponstimulationsuggestedanti-inflammatorytendencyof GPx12/26Cat2/2mice. Suppressionof Stat3activationinassociation with enhancement of indoleamine 2,3-dioxygenase and FoxP3 expression might be involved in theimmunosuppressivemechanismofGPx12/26Cat2/2mice. Takentogether, itisimpliedthatROSlevel iscritical intheregulationofTregfunction, andIBDmaybeattenuatedinappropriatelyelevatedlevelsofROS.Citation:KimH-R, LeeA, ChoiE-J, KieJ-H, LimW, etal. (2014)AttenuationofExperimental ColitisinGlutathionePeroxidase1andCatalaseDoubleKnockoutMicethroughEnhancingRegulatoryTCell Function. PLoSONE9(4): e95332. doi:10.1371/journal.pone.0095332Editor:LucienneChatenoud, Universite ParisDescartes, FranceReceivedDecember5, 2013; AcceptedMarch25, 2014; PublishedApril 17, 2014Copyright: 2014 Kim etal. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricteduse, distribution, andreproductioninanymedium, providedtheoriginal authorandsourcearecredited.Funding: This study was supported by a grant of the Korean Health Technology R&D Project, Ministry of Health & Welfare, Republic of Korea (HI12C0050) (http://www.khidi.or.kr/www/run.do) andRP-Grant2011ofEwhaWomansUniversity(http://www.ewha.ac.kr/mbs/ewhaen/index.jsp). Thefundershadno role in studydesign, datacollectionandanalysis, decisiontopublish, orpreparationofthemanuscript.CompetingInterests:Theauthorshavedeclaredthatnocompetinginterestsexist.*E-mail: jyseoh@ewha.ac.kr(JYS); mbit@ewha.ac.kr(BIM).Theseauthorscontributedequallytothiswork.IntroductionReactiveoxygenspecies(ROS)arehighlyreactiveandinteractwith many bio-molecules. At high concentrations, they are likely todestroy biological structures, promoting cellular damage and tissuedestruction. Traditionally, ROShavebeenimplicatedinageingandtheprogressionof inflammatoryandautoimmune diseases,includinginflammatorybowel diseases(IBD) [1,2,3]. Meanwhile,manyrecentobservationsareopposingthetraditionalconceptonROS,suggestingtheprotectiveroleofROSinimmune-mediatedinflammatorydiseases[4].MicewithlowerlevelofROSthanWTmiceduetodefectsinROS-producingenzymesystem, suchas Ncf12/2or Nox22/2,are more susceptible to autoimmune diseases, such as arthritis andencephalomyelitis [5,6,7]. Humans withlower levels ROSthannormal persons, suchas chronic granulomatous disease (CGD)patients andcarriers, are alsomore susceptible toautoimmunediseases[8,9].Bycontrast,micewithhigherlevel ROSthanWTmice due to the defect in a ROS metabolizing enzyme, glutathioneperoxidase-1(GPx-1), areresistant toimmune-mediatedinflam-matory diseases, such as allergen-induced airway inflammationandhighfat diet-inducedatherosclerosis [10,11]. Inparticular,micewithhigherlevel of ROSduetodefect of anon-enzymaticcellular anti-oxidant, peroxiredoxin (Prx) II, are resistant todextransodiumsulfate(DSS)-inducedcolitis[12].These clinical or experimental observations implicated theimmunoregulatoryroleof ROS, andadoptive-transfer of CD4+cellsfromratswithlowerROSlevelinducedarthritisinratswithnormal ROS level, demonstrating the key role of CD4+cells in thehyperinflammatoryresponseinloweredlevels of ROS[13]. Ontheotherhand,oxidativestressinducesTcellhyporesponsivenessinseveral humanpathologies (e.g. cancer, rheumatoidarthritis,AIDS and leprosy) [14,15]. Accordingly, ROS level is supposed tobe closely associated with Tcell responsiveness. In particular,regulatory T cell (Treg) function seems to be closely linked to ROSlevel. Tregsisolatedfrommicewithlowerlevel of ROS, suchasNcf12/2mice, werehypofunctional thanWTTregs[16]. Tregswere alsohypofunctional invitroat loweredlevels of ROSbyPLOSONE | www.plosone.org 1 April2014 | Volume9 | Issue 4| e95332adding antioxidants or NADPH oxidase inhibitors. Differentiationof inducibleTreg(iTreg) seems alsolinkedtothelevel of ROS.Inductionof FoxP3+iTregwasattenuated, whereasthat of Th17cells was enhanced in lowered levels of ROS due to Nox2deficiency [6,7] or addition of apocynin [17]. By contrast,induction of FoxP3+Treg was enhanced in elevated levels ofROSdue toPrxII deficiency [12]. Meanwhile, the suppressivefunctionof TregshasbeeninvestigatedonlyinloweredlevelsofROS so far, and the suppressive function of GPx12/2or PrxII2/2Tregshasnotyetbeenreported.Thus, in the present study, we investigated the suppressivefunctionofTregsisolatedfrommicewithelevatedlevelsofROSdue to defects in GPx1 and catalase (Cat) [18]. The results showedthat GPx12/26Cat2/2Tregs were hyperfunctional andGPx12/26Cat2/2micewereresistant toDSS-inducedcolitis.Meanwhile, administration of n-acetylcysteine (NAC) reducedTregfunctionandmadeGPx12/26Cat2/2micesusceptibletoDSS-inducedcolitis.MaterialsandMethodsMiceC57BL/6 wild-type (WT) and GPx12/26Cat2/2mice with aC57BL/6 genetic background were housed and maintained in theanimal facilityatEwhaWomansUniversity[18]. ThisstudywasperformedaccordingtoKoreanFoodandDrugAdministrationguidelines and was specifically approved by the InstitutionalAnimal CareandUseCommitteeof EwhaWomans UniversityGraduateSchool ofMedicine(PermitNumber: 10-0133).ROSMeasurementTen million splenocytes prepared by mincing fromWTorGPx12/26Cat2/2micewereincubatedwith5 mMdichloro-fluoroscein diacetate (DC-FDA, Sigma, St. Louis, MO) for 30minat 4uCinthedark. For thesurfacestain, anti-CD4-PerCPandanti-CD25-PEoranti-CD11c-PEpurchasedfromBDBiosciences(SanDiego,CA) wereincubatedtogether.ThecellswerewashedtwicewithcoldPBSandweresuspendedinDMEM. Then, thecellswereanalyzedwithaFACSCaliburforfixedtime. After30secondfromstart of datareading, thecellswerestimulatedwith100nMphorbol 12-myristate 13-acetate (PMA, Calbiochem,Darmstadt, Germany).InductionofAcuteExperimental ColitisMales at 810 weeks of age were administered with 3% dextransodiumsulfate (DSS) purchased fromMP Biomedicals (Stras-bourg, France) in drinking water for 4 days. During the 4 days, themicewererestrictedfromwatersupplyfor12hoursandthenfedwith3%DSSwater for 12hours inaday. Eachmousedrankabout 20mL of 3% DSS water during the 4 days. Some mice wereadministeredintra-gastricallywith400 mLof40mMN-acetylcys-teine (NAC, Sigma) inwater, as mice donot like theflavor ofNAC, every day from 3 days before the treatment with DSS to theendof theexperiment, except forthe4daysduringwhichDSSwasadministered. Bodyweight wasmeasuredeverydayandthemiceweresacrificedonthe7thdayaftertreatmentwithDSS.Forhistologicalexamination,thecolonswereremoved,rolledarounda cotton swab and were fixed in 10%formaldehyde to beembedded in paraffin. The paraffin blocks were longitudinallysectioned serially with the thickness of 5 mm and were stained withFigure 1. Intracellular ROS level was higher in GPx12/26Cat2/2than WT lymphocytes. Splenocytes were stained with DC-FDA (5 mM) andsimultaneouslywithanti-CD4-PerCPandanti-CD25-PE(A)oranti-CD11c-PE(B). ThenthecellswerewashedandacquiredusingFACSCalibur. Afterstabilization for 30sec, acquisition was briefly suspended and PMA (100nM) was added to the tube, and the cells were acquired further for 482sec.ROSlevel wasexpressedasmeanfluorescenceintensity(MFI)ofDC-FDAfluorescence. Dataaremean6 SE(n =12). *, P,0.05.doi:10.1371/journal.pone.0095332.g001ROSIsCriticalinTregFunctionandExperimental ColitisPLOSONE | www.plosone.org 2 April2014 | Volume9 | Issue 4| e95332hematoxylinandeosin(H&E)toallowhistologicalexaminationofwholecolons.ImmunohistochemistrySections weredeparaffinatedinxylene, dehydratedinethanolandwashedinPBSfollowedbysuccessivepermeabilizationsteps(withTriton0.2%inPBS). Endogenous peroxidasewas blockedwithhydrogenperoxide(5%inPBS) for30minandthesectionswere subjected to heat-induced antigen retrieval step beforeincubation with a universal blocking solution (Dako, Glostrup,Denmark) for 30min. Then, the sections were incubatedwithanti-phosphotyrosine (pY)-Stat3 (Tyr705) (clone D3A7, CellSignaling Technology, Danvers, MA), anti-indoleamine 2,3,-dioxygenase (IDO, rabbit polyclonal antibody, Abcam, Cam-bridge, UK) or anti-FoxP3(rabbit polyclonal antibody, Abcam).Then, the sections were incubated with labelled streptavidin biotinreagents forrabbit, mouseandgoat antibodies (Dako, Glostrup,Denmark) anddevelopedusing 3,39-diaminobenzidine as chro-mogensubstrate. Forthecountingof FoxP3+cells, 5high-powerfields selected at random in the lesions in each slide were examinedbythreedifferentpathologists.PreparationofCellsSpleens were removed fromsacrificed mice and single cellsuspensionwaspreparedbysqueezingonacell strainer(70 mm,BDBiosciences,SanJose,CA).AftererythrocyteswerelysedwithRBClysis buffer (eBioscience, San Diego, CA), CD4+CD25+fraction was separated using a regulatory Tcell isolation kitpurchased fromMiltenyi Biotech (Auburn, CA). CD4+CD252cells were also isolated and used for effector T cells (Teffs). For thepuritycheck, thecells werestainedforsurfaceCD4andCD25,followedbyintranuclear stainingfor FoxP3. After Fc receptorsFigure 2. Body weight change during the course of DSS-induced colitis. Body weight of the mice was measured every day, after 4 days oforal administration of 3% DSS in drinking water. Some mice were administered intra-gastrically with 400 mL of 40mM NAC in water, as mice do notlike the flavor of NAC, every day from 3 days before the treatment with DSS to the end of the experiment, except for the 4 days during which DSS wasadministered. KO, GPx12/26Cat2/2. Data are mean 6SE, n =12 in each group. *, P,0.05, compared between KO mice control and KO mice treatedwith DSS. 1, P,0.05, compared between WT mice treated with DSS+NAC and WT mice treated with DSS only. ", P,0.05, compared between KO micetreatedwithDSSandWTmicetreatedwithDSS.doi:10.1371/journal.pone.0095332.g002Figure3.Colonlength was notshortenedinKO micetreated with DSSonly, but wasshortenedwith DSS andNAC. Onthe7thdayafter treatment with DSS, colon length was measured. KO, GPx12/26Cat2/2. Scale bar is 1cm. Data are mean 6SE, n =12 in each group. *, P,0.05.doi:10.1371/journal.pone.0095332.g003ROSIsCriticalinTregFunctionandExperimental ColitisPLOSONE | www.plosone.org 3 April2014 | Volume9 | Issue 4| e95332wereblockedusinganti-mouseCD16/CD32(2.4G2, BDBiosci-ences) for 15minat 4uC, cells were incubatedwithanti-CD4-FITC(H129.19) andanti-CD25-PE(PC61) purchasedfromBDBiosciencesfor30minat4uC. Afterwashed, thecellswerefixedandpermeabilizedusingamouse regulatory Tcell stainingkit(eBiosciences, SanDiego, CA) andwerestainedwithanti-Foxp3-PerCP-Cyanine5.5 (FJK-16s). The stained cells were analyzedusingaFACSCalibur (BDBiosciences). Teffs werelabeledwithcarboxyfluorescein diacetate succinimidyl ester (CFSE, Invitrogen,Carlsbad, CA) as describedelsewhere, totracetheproliferativeresponse [19]. For the purification of dendritic cells (DCs),CD11c+cells were isolatedusinga spleendissociationmedium(StemCell technologies, Vancouver, BC), densitygradient centri-fugationover15.5%Accudenz(AccurateChemical &Scientific,Westbury, NY), and immunomagnetic selection using anti-CD11c-PE (BD Biosciences) and microbead anti-PE (Miltenyi Biotec). Thepurity of CD11c+cells was consistently over 95%. Sometimes, cellswere prepared 24 hours after the intraperitoneal injection of NAC(10mginPBS) orPBS.Figure4.Inflammatoryreaction was attenuated in thelesions ofDSS-inducedcolitisin KOmice, butwas aggravated byNAC. Onthe 7thday after oral administration of NAC, the colons were removed, rolled around a cotton swab and were fixed to be embedded in paraffin. TheparaffinblockswerelongitudinallysectionedandwerestainedwithH&Etoallowhistological examinationof thewholecolons(A). Scalebaris200 mm. Inflammatory area was measured by histological examination of the entire colon by 3 separate pathologists (B). KO, GPx12/26Cat2/2. Dataaremean 6SE, n =12ineachgroup. *, P,0.05.doi:10.1371/journal.pone.0095332.g004ROSIsCriticalinTregFunctionandExperimental ColitisPLOSONE | www.plosone.org 4 April2014 | Volume9 | Issue 4| e95332InvitroSuppressiveAssayFor the functional assessment of the suppressive activity ofTregs, DCs, Teffs and Tregs were co-cultured in the ratio of0.2:1:1. Meanwhile, a small percentage of CD4+CD25+T cells arenot Tregs and do not express FOXP3, and the isolated Tregsamples were not pure based on FOXP3 expression. In the presentstudy, 86.6,91.4% of the purified Treg fractions wereCD4+FoxP3+(Fig. S1). Toensureoptimal reproducibilityof thesuppressionassays, thedegreeof purityof theTregsampleswastaken into account when the cells were plated so that the final ratioof CD4+FOXP3+Tregs toCD4+FOXP32Teff was as closeaspossible to 1:1 [16]. For the proliferative responsiveness, 104CFSE-labeled CD4+FOXP32Teffs and 26103DCs werecultured in DMEMsupplemented with 10%FCS (Hyclone,Logan,UT)inround-bottomed96-wellplatesbystimulatingwithvarious concentrations of soluble anti-CD3e (e-Biosciences). Onthe3rddayof culture, thecells wereharvestedforstainingwithanti-CD4-PerCP.WholecellswereacquiredbyusingaFACSCa-libur and were analyzed by using Winlist software (Verity,Topsham, ME). Precursor frequency (Pf) was estimatedfor thecells exclusively gated for CD4+live cells according to thescatteringcharacteristics,usingtheproliferationwizardofModifitsoftware (Verity), as described elsewhere [19,20]. In order to assessthesuppressivefunctionof Tregs, CD4+FOXP3+104Tregs wereadded to the culture, and the Pf values in the absence and presenceof Tregs were compared to give rise to suppression (%). In order tocomparethesuppressivefunctionof WTTregsandGPx12/26Cat2/2Tregs, feeder function of WTDCs and GPx12/26Cat2/2DCs, and the proliferative activity of WTTeffs andGPx12/26Cat2/2Teffs,3 kindsof cells(DCs,TeffsandTregs)intheculturefromtheWTandGPx12/26Cat2/2micewerecross-combined to give rise to 8 kinds of combinations, as shown inFig. S2. Whennecessary, catalase(100U/mL, Sigma) or NAC(40nM) wereaddedtothecultures.InvitroTh17/TregCell DifferentiationNa veCD4+Tcells wereisolatedfromthespleenof WTorGPx12/26Cat2/2mice usinganaive CD4+Tcell isolationkit (R&D Systems, Minneapolis, MN). For the induction ofTh17cell differentiation, 16105na ve CD4+cells werestimulatedwithsoluble anti-CD3e (1 mg/mL) andsoluble anti-CD28 (1 mg/mL, e-Biosciences) in the presence of 26104CD11c+DCs for 24 hours, and were cultured further for2.5days in the presence of TGF-b1 (5ng/mL) and IL-6(20ng/mL) purchased from R&D Systems. For intracellularcytokine staining, cells were re-stimulated for 4hr with PMA(25ng/mL) andionomycin(250ng/mL, Sigma) inthepresenceof a protein trapsport inhibitor containing monensin (BDBiosciences). Then, the cells were harvested and stained forintracellular IL-17A and IFN-c using a commercial kit forfixationandpermeabilization((BDBiosciences), anti-mouse IL-17A-PE(eBioscience) andanti-mouseIFN-c-FITC(eBioscience).For the induction of iTreg differentiation, 16105na ve CD4+cells were stimulated with plate-coated anti-CD3e (50ng/well)and soluble anti-CD28 (1 mg/mL) in the presence of TGF-b1(5ng/ml) and human recombinant IL-2 (10U/mL, BDBiosciences). Afterthreedaysof culture, thecellswereharvestedforintranuclear stainingforFoxP3, usingamouseregulatoryTcell stainingkit.CytokineProductionAssayInordertoevaluatetheproducingcapabilityof inflammatorycytokines,IL-6andIL-17A,106splenocyteswerestimulatedwithsolubleanti-CD3e(1 mg/mL) andanti-CD28(1 mg/mL). After5days of culture, cytokines secreted in the supernatants wereFigure5.Stat3wasactivatedinthelesionsofDSS-inducedcolitisinWTmicebutnotinKOmice.Immunohistochemical stainingforphosphotyrosine (pY)-Stat3 (Tyr705) shows stat3 was activated in the lesions of DSS-induced colitis in WT mice. In the colons of KO mice treated withDSS,pY-Stat3isobserved butatvery weaksignalintensity inthe infiltrating cells.Meanwhile,in theKO micetreated withbothDSSandNAC, pY-Stat3isobservedasstronglyasWTmicetreatedwithDSSintheinfiltratingcells. Scalebaris50 mm.doi:10.1371/journal.pone.0095332.g005ROSIsCriticalinTregFunctionandExperimental ColitisPLOSONE | www.plosone.org 5 April2014 | Volume9 | Issue 4| e95332Figure 6. KO Tregs were hyperfunctional. CFSE-labeled Teffs were stimulated in the presence of DCs and Tregs from WT or KO mice. On the 3rdday, the cells were harvested and stained for surface CD4. Live CD4+CFSE+cells were gated for the analysis of the proliferative responsiveness of Teffs.Theproliferativeresponseof Teffsinthepresenceof KOTregs was less activethaninthepresenceof WTTregs, suggestingKOTregs werehyperfunctionalinthesuppressionofTeffsthanWTTregs. KO, GPx12/26Cat2/2. Dataaremean 6SEofsixseparateexperiments.doi:10.1371/journal.pone.0095332.g006Figure 7. Simulated immune responsiveness of KO mice. Comparison of Teff proliferative responsiveness in the cultures of cognate cells fromthe same mice, i.e., between WT Teff+WT DC (+ WT Treg) and KO Teff+KO DC (+ KO Treg). In the absence of Tregs, KO Teffs were hyperproliferative,butinthepresenceofTregs, werehypoproliferativethanWTTeffs. KO, GPx12/26Cat2/2. Dataaremean 6SEofsixseparateexperiments.doi:10.1371/journal.pone.0095332.g007ROSIsCriticalinTregFunctionandExperimental ColitisPLOSONE | www.plosone.org 6 April2014 | Volume9 | Issue 4| e95332analyzedbycytometricbeadarray(BDBiosciences). Briefly, theculture supernatants were mixed with anti-mouse IL-6, anti-mouseIL-17AcapturebeadsandPEdetectionreagent for2hrat RT.The beads were washed with 1mL of wash buffer and re-suspended in 300 mLof wash buffer, and were analyzed withappropriateacquisitiontemplatesuppliedbyBDBiosciences.Statistical AnalysisData are expressed as mean6SEof more thansix separateexperiments.ComparisonofdatawasdonebyusingindependentStudents t test or ANOVA. P values less than 0.05 wereconsideredstatisticallysignificant.ResultsIntracellularROSLevel wasHigherinGPx12/26Cat2/2thanWTLymphocytesFlowcytometricanalysis of thesplenocytes stainedwithDC-FDAafter stimulationwithPMAshowedthat the intracellularROSlevelinthesplenocytesfromGPx12/26Cat2/2micewashigher than that from WT mice (Fig. 1A). Among the T cells usedinthe suppressionassay, intracellular ROSlevel was higher inCD4+CD25+Tregs than in CD4+CD252Teffs. Intracellular ROSlevel was higher in GPx12/26Cat2/2CD11c+DCs than that inWTDCs(Fig. 1B).DSS-inducedColitiswasAttenuatedinGPx12/26Cat2/2MiceAfter treatment with DSS, WT mice exhibited diarrhea, bloodystool andsignificant weight loss (Fig. 2), andcolonlengthwassignificantlyshortened(Fig. 3). Histological examinationshowedinflammatorychangesinsubstantial area(17.864.2%, n =12) ofthe entire colon, includinginflammatory cell infiltration, severeulceration, cryptic distortion and dysplastic changes (Fig. 4).Immunohistochemistry (IHC) showed enhanced expression of pY-Stat3intheinfiltratingcellsaswell asintheintestinal epithelialcells inthe lesions of DSS-inducedcolitis (Fig. 5). By contrast,GPx12/26Cat2/2micedidnot exhibit diarrhea, andweightloss was very slight only on the 5thand 6thday after treatment withDSS(Fig. 2). At the time of investigationonthe 7thday aftertreatmentwithDSS,colonlengthwasnotshortened(Fig.3),andhistological examination showed minimal inflammatory changes atrestrictedarea(2.160.8%) (Fig. 4). Expressionof pY-Stat3wasobservedintheinfiltratingcellsbut at veryweaksignal intensity(Fig. 5).AdministrationofNACAggravatedDSS-inducedColitisAdministrationof NACbeforeandafter treatment withDSSintoGPx12/26Cat2/2mice inducedsevere weight loss andcolonshorteningcomparabletoDSS-inducedcolitisinWTmice(Figs. 2&3). Meanwhile, theyrecoveredearlier thanWTmicetreatedwith DSS (Fig.2). Histological examination showed severeinflammatorychangesandexpressionof pY-Stat3comparabletoDSS-inducedcolitisinWTmice(Fig.4&5).AdministrationwithbothNACandDSSinducedearlierweightlossinWTmicethanthose treated withDSSonly (Fig. 2). Histological examinationshowedmoresevereinflammatorychangesinwiderareainWTmicetreatedwithbothNACandDSS(Fig. 4). Takentogether,invivo administration of NACaggravated DSS-induced colitisbothinWTaswell asinGPx12/26Cat2/2mice.GPx12/26Cat2/2TregswereHyperfunctionalTeffs in the cocultures with KO Tregs was less proliferative thanin cocultures with WT Tregs, suggesting KO Tregs werehyperfunctional inthesuppressionof Teff proliferationthanWTTregs (Fig. 6&Fig. S2). Meanwhile, inthe absence of Tregs,GPx12/26Cat2/2Teffs wereslightlymoreproliferativethanWT Teffs in the presence of WT CD11c+DCs, not in the presenceof GPx12/26Cat2/2DCs (Fig. S3 & Fig. S4). We alsocomparedtheproliferativeactivityofTeffsinthecocultureswithcognate cells from the same mice, i.e., between WT Teff+WT DC(+ WT Treg) and KOTeff+KODC(+ KOTreg) for thesimulationof wholeimmuneresponsiveness of mice. Theresultshowed KO Teffs were hyperproliferative in the absence of Tregs,but were hypoproliferative inthe presence of Tregs, thanWTTeffs(Fig.7).Takentogether,theattenuatedDSS-inducedcolitisin GPx12/26Cat2/2mice might reflect the hypoactive immuneresponsiveness in the presence of Tregs, suggesting the critical roleofTregsintheregulationofimmuneresponsivenessinvivo.Figure8.InvivoadministrationofNACintoKOmicereducedTreg function. Twenty four hours after the intraperitoneal injection ofNAC (10mg in PBS) or PBS (control), cells were prepared and analyzedforthesuppressivefunctionofTregsinthesamewayasinFig. 6. KO,GPx12/26Cat2/2. Data are mean 6 SEofsix separate experiments. Pvalue indicates significance of difference betweenKOTregandKOTreg-NAC.doi:10.1371/journal.pone.0095332.g008Figure 9. N-acetylcysteine (NAC) or catalase reduced Tregfunction invitro. Comparison of Teff proliferation in the absence andpresence of Tregs in response to soluble anti-CD3e (1 mg/mL) gave risetosuppression(%). Additionof NAC(40nM) or catalase(100U/mL)reducedthesuppressivefunctionof WTandKOTregs invitro. KO,GPx12/26Cat2/2. Dataaremean 6 SEofsixseparateexperiments.doi:10.1371/journal.pone.0095332.g009ROSIsCriticalinTregFunctionandExperimental ColitisPLOSONE | www.plosone.org 7 April2014 | Volume9 | Issue 4| e95332TregFunctionwasCloselyAssociatedwithROSLevelThe suppressive activity of Tregs isolatedfromGPx12/26Cat2/2miceonedayafterintraperitoneal (IP) injectionof NACwas reducedtothe level comparable toWTTregs (Fig. 8). IPinjection of NACinto WTmice also reduced the suppressivefunctionof Tregs (datanot shown). Inaddition, IPinjectionofvitaminC(500mg/kg) alsoreducedthesuppressivefunctionofTregs, suggestingthecritical roleof ROSlevel inTregfunction(datanot shown). Additionof NACorcatalaseintotheculturesalsoreducedthesuppressivefunctionofTregsinvitro, providinganother supportive evidence that ROS level is critical in theregulationofTregfunction(Fig. 9).Anti-inflammatoryTendencyofGPx12/26Cat2/2MiceInorder toinvestigatethe immuneresponsive pattern, na veCD4+cells isolated from the spleens of WT or GPx12/26Cat2/2micewereinducedtodifferentiateintoTh17cells or FoxP3+Tregs bystimulating in the presence of TGF-b1 and IL-6 or IL-2,respectively. TheresultsshowedthattheproportionofTh17cellsdifferentiatedfromGPx12/26Cat2/2CD4+cells was slightlybutsignificantlylowerthanthatfromWTCD4+cells, suggestingless inflammatory tendency of the GPx12/26 Cat2/2mice(Fig.10).Actually,theamountsofIL-6andIL-17AsecretedfromtheGPx12/26Cat2/2splenocytes weremuchless thanthosefromWTsplenocytes inDSS-inducedcolitis (Fig. 11). Ontheother hand, iTreg differentiationwas not significantly differentbetween WTand GPx12/26 Cat2/2CD4+cells (Fig. S5).Figure 10. Th17 cell differentiation was mitigated in KO mice. Na ve CD4+cells isolated from the spleens of WT or KO mice were induced todifferentiateintoTh17cells bystimulatinginthepresenceof TGF-b1andIL-6. KO, GPx12/26Cat2/2. Dataaremean6SEof sixseparateexperiments.doi:10.1371/journal.pone.0095332.g010Figure 11. IL-6 and IL-17A production was impaired in KO mice under inflammatory condition but was rescued by administrationwith N-acetylcysteine (NAC). Splenocytes prepared from the mice treated as indicated were stimulated with anti-CD3e and anti-CD28 for 5 days.IL-6andIL-17Asecretedinthesupernatantswereanalyzedbycytometricbeadarray. KO, GPx12/26Cat2/2. Dataaremean 6 SEofsixseparateexperiments.doi:10.1371/journal.pone.0095332.g011ROSIsCriticalinTregFunctionandExperimental ColitisPLOSONE | www.plosone.org 8 April2014 | Volume9 | Issue 4| e95332Meanwhile,administrationofNACintotheGPx12/26Cat2/2miceinducedsecretionofIL-6andIL-17A,suggestingthatNACaffectedtheGPx12/26Cat2/2micetowardpro-inflammatorytendency.IDO Expression was Induced in DSS-induced Colitis in WTMice, whileIDOwasExpressedfromtheBeginninginGPx12/26Cat2/2MiceIDOisanimmunoregulatoryenzymethat canbeinducedbyROS [21,22]. At steady state, IDO expression was rarely observedonly in lymphoid follicles in the colons of WT mice (Fig. 12). In thecolons of WT mice treated with DSS, IDOexpression wasobservedintheinfiltratingcells,andweakexpressionofIDOwasobserved in the epithelial cells. By contrast in GPx12/26Cat2/2mice, IDOexpressionwasobservedintheepithelial cells, andinparticularstronglyinendothelialcells,atsteadystate.Meanwhile,in the colons of GPx12/26Cat2/2mice treated with DSS, IDOexpressionwas not enhanced, comparedwiththosenot treated,although IDO expression was also observed in the infiltrating cells.AdditionofNACalsodidnotenhanceIDOexpression,althoughinflammatoryreactionwasaggravated.TheFrequencyofFoxP3+TregswasIncreasedintheLesions of DSS-induced Colitis in GPx12/26Cat2/2MiceFoxP3+cellswerehardlyobservedinthecolonsofWTmiceatsteadystate(Fig.13).InthelesionsofDSS-inducedcolitisinWTmice, FoxP3+cellswereobservedamongtheinfiltratingcellsbutat a very low frequency. By contrast, in the lesions of DSS-inducedcolitis inGPx12/26Cat2/2mice, FoxP3+cells wereobservedquite frequently among the infiltrating cells. Meanwhile, thefrequency of FoxP3+cells were reduced in the lesions of GPx12/26Cat2/2mice treatedwithbothNACandDSStothe levelcomparabletoWTmicetreatedwithDSS.DiscussionInthepresent study, wedemonstratedfor thefirst timethatTregs were hyperfunctional inelevatedlevel of ROSby usingGPx12/26Cat2/2Tregs. AsithasbeenalreadyreportedthatTregs werehypofunctional inloweredlevels of ROS[16,17], itcouldbearguedthatTregfunctioniscloselylinkedtoROSlevel.Actuallyinthepresentstudy,IPinjectionofNACintoGPx12/26Cat2/2micereducedthesuppressivefunctionofTregstothelevel comparabletoWTTregs (Fig. 8). Administrationof NACalsohas made GPx12/26 Cat2/2mice, whichare naturallyresistant,susceptibletoDSS-inducedcolitis,suggestingthecriticalroleofROSinthepreventionofDSS-inducedcolitis(Figs. 25).The importance of Tregs in the maintenance of intestinal immunebalance has been already shown in many other studies [23].Consequently, ROSlevel mightbecritical inthemaintenanceofintestinal immune homeostasis, providing an insight for theimmunomodulationbyROS.This argument is completelyopposingthetraditional conceptthat ROS is involved in the induction and progression of IBD [3].According to the traditional concept, anti-oxidant interventionshouldbeeffectiveinthetreatment of IBD, andlargearrayofantioxidantcompoundshavebeenshowntobeprotectiveagainstIBDinexperimental animals. Nonetheless, objectiveclinical datasupporting anti-oxidant intervention of human IBD are still scarce[24].Atthemoment,itisnecessarytoestablishanewconceptualframeworkthat cancompromisetherecent observationsandthetraditional concept. WecandrawacluefromthecomparisonofGPx-12/2miceandGPx-12/26GPx-22/2mice.Aspreviouslymentioned, immune response is attenuated inGPx-12/2mice[10,11]. The life span of GPx-12/2mice is not shortened althoughcellular DNAdamages are accumulated, suggesting that ROSlevel iselevated, at whichthemicearestill tolerable[25,26]. Incontrast, inflammatorydiseases suchas colitis developspontane-ouslyinGPx-12/26GPx-22/2mice,suggestingthatinflamma-Figure 12. IDO was expressed at steady state fromthe beginning in KOmice, but was induced after treatment with DSS in WT mice.ImmunohistochemicalstainingofthelesionsofDSS-inducedcolitisforIDO. KO, GPx12/26Cat2/2. Scalebaris50 mm.doi:10.1371/journal.pone.0095332.g012ROSIsCriticalinTregFunctionandExperimental ColitisPLOSONE | www.plosone.org 9 April2014 | Volume9 | Issue 4| e95332tionisaugmented[27]. GPx-12/26GPx-22/2miceusuallydonot grow well and lifespan is shortened, suggesting that ROS levelis elevated to a higher level that the mice could not tolerate.Spontaneous development of inflammatory diseases and shortenedlifespanobservedinGPx-12/26GPx-22/2micearecompatiblewith the traditional concept on ROS. Therefore, we think that thelevels of ROS in many evident previous observations thatcontributetoestablishthetraditional concept onROS, suchasFigure13. Thefrequencyof FoxP3+Tregswasincreasedinthelesionsof DSS-inducedcolitisinKOmice. Immunohistochemicalstaining of the lesions of DSS-induced colitis for FoxP3 (A). For the counting of FoxP3+cells, 5 high-power fields selected at random in each slide wereexaminedbythreedifferentpathologists(B). KO, GPx12/26Cat2/2. Scalebaris50 mm. Dataaremean 6 SEofsixseparateexperiments.doi:10.1371/journal.pone.0095332.g013ROSIsCriticalinTregFunctionandExperimental ColitisPLOSONE | www.plosone.org 10 April2014 | Volume9 | Issue 4| e95332vascular reperfusion injury, may be as high as that of GPx-12/26GPx-22/2mice [28,29]. Thus, we can imagine an arbitrarythresholdlevelofROS,althoughwecannotquantitativelyspecifyatthemoment,betweenthehigherlevelsthatinducedirecttissuedamage, such as GPx-12/26GPx-22/2mice, and themoderately high levels at which mice can tolerate, such asGPx12/2mice.Taken together, we hypothesize, if ROS can induce direct tissuedamages at highlevels, defensive or compensatorymechanismscounteractingthedestructiveeffectsofROSshouldbedevelopedinthe body. Actually, mice withmoderately elevated levels ofROS, such as GPx12/2, PrxII2/2and GPx12/26Cat2/2mice,were anti-inflammatory [10,11,12], whereas those withloweredROS level, such as Ncf12/2and Nox22/2mice, were pro-inflammatory [5,6,7]. From this point of view, it is quitereasonable that the suppressive function of Tregs is closelyassociatedwithROSlevel. At molecularlevel, theexpressionofanimmunoregulatoryenzyme, IDO,isalsoassociatedwithROSlevel [21,22]. IDO catabolizes the essential amino acid tryptophanintothestablemetabolite, kynurenine[30]. Consequently, IDOdepletestryptophanfromtheenvironment, thusstarvingeffectorcells. It was also found that tryptophan depletion resulted ininhibitionof Th17cell differentiationandexpansionof Foxp3+Tregs[22,31,32].IDOisprimarilyexpressedinantigen-presentingcells(APCs),suchas dendritic cells and macrophages, but IDOpathway isinducedinmanytissuesduringinflammationbecauseIDOgeneexpressionisinducedbyinterferon[22,33]. Inthepresentstudy,IDOexpression was rarely observed at steady state, but wasinducedintheepithelial cellsaswell asintheinfiltratingcellsinthelesionsofDSS-inducedcolitis(Fig.12).Thus,IDOexpressionmightbeinducedasaconsequenceoftheinflammatoryreaction,contributingtothe feedbackregulation. Onthe other hand, inGPx12/26Cat2/2mice, IDO expression was observed in manytypes of thecells, includingendothelial, epithelial andAPC-likecells, fromthe beginning. Elevated levels of ROS not onlycontribute to the induction but also enhance the enzyme activity ofIDO, as superoxide radical acts as a cofactor of IDO[30].Therefore, highexpressionandstrongactivityof IDOfromthebeginninginGPx12/26Cat2/2micemight contributetothepreparation of immunosuppressive environment preventing in-flammatory tissue damage during treatment with DSS. Actually inthepresent study, thefrequencyof FoxP3+cellswassignificantlyincreasedinparallel withsignificantly attenuatedinflammatoryreaction in the lesions of DSS-induced colitis in mice with elevatedlevel of ROS due to defects in GPx1 and Cat (Fig. 13). By contrast,IP injection of NAC significantly reduced the frequency of FoxP3+cellsandaggravatedinflammatoryreactioninthelesionsofDSS-inducedcolitis. IPinjectionof NACalsoaggravatedDSS-inducedcolitisinWTmice, inducingearlierweight lossandmoresevereinflammatory changes. This result is quite different from the reportbyAmrouche-Mekkioui &Djerdjouri (2012) thatNACimprovedDSS-inducedcolitis[24]. Meanwhile, theexperimental scheduleswere different as our model was acute colitits while theirs waschronic colitis. However, further investigation is needed tounderstandtheexactreasonsofsuchopposingresults.For the functional assay of Tregs, we used CD4+CD25+fractionisolatedfromthespleensof GPx12/26Cat2/2andWTmice.AlthoughTregwasoriginallyidentifiedinCD4+CD25+fraction,thefractionisnotspecifictoTregs[34].Atpresent,FoxP3isthemost specific to Tregs, but it is not also absolutely specific [35,36].Anyway, FoxP3+fraction would be more appropriate thanCD4+CD25+fractionfor thefunctional assayof Tregs, but it isstill technicallyimpossibletoisolateFoxP3+cellsaslivestateforfunctional assay. BecauseFoxP3is anintranuclear transcriptionfactor,itisunavoidabletodestroycytoplasmicmembranetostainFoxP3. Of course, it is possibletoisolateliveFoxP3+cells fromspecial geneticallymodifiedmice, suchas eGFP-FoxP3knock-inmice[37].However, atthemoment, theyarestill unavailableonGPx12/26Cat2/2background, insteadweusedCD4+CD25+fractionfor thefunctional assayof Tregs. Efimovaetal (2011),whodemonstratedTregswerehypofunctionalinloweredlevelsofROS, have also used CD4+CD25+fraction for the functional assayof Tregs, takingintoaccount the small FoxP32fractionintheCD4+CD25+fraction[16].Wealsofollowedthesamemethodinthepresentstudy.Concerningonthe suppressive assayof Tregs, we previouslydemonstrated that the suppressive patterns of Tregs might bedifferent accordingtothestrengthof stimulatingsignals [19,38].Thus, it is necessary to stimulate the cells in a wide range of signalstrengthfor theaccuratefunctional assessment of Tregs. Inthepresent study, weusedavarietyof concentrations of anti-CD3eantibody for different range of signal strength. Based on ourexperience of several years on Treg functional assay, we employedaprecisequantitativeparameter fortheproliferativeresponsive-ness, precursor frequency, which represents the fraction of the cellsattheinitial timepointthathavegoneintocell cycle[20,39,40].Inthepresent study, wedemonstratedanexperimental colitiswas attenuatedinelevatedlevel of ROS. Enhancement of TregfunctionandIDOexpression, investigatedinthepresent study,might be involvedinthe underlying mechanism. However, wecannot excludethepossibilityof involvement of othermoleculesandcells that havenot investigatedinthepresent study. Takentogether, the results of the present study suggest the potentialtherapeutic strategy for IBD through immunomodulation byROS.SupportingInformationFigureS1 Purityof isolatedCD4+CD25+fraction. Theisolated CD4+CD25+fraction is not pure Treg population, interms of FoxP3 expression. CD4+FoxP3+cells ranged from 86.6 ,91.4%(88.263.4%, n =12) intheCD4+CD25+fraction.(TIF)FigureS2 KOTregswerehyperfunctional.CFSE-labeledTeffswerestimulatedwithvariousconcentrationsofsolubleanti-CD3e as indicated in the presence of DCs and Tregsfrom WT orKO mice. On the 3rdday, the cells were harvested and stained forsurface CD4. Live CD4+CFSE+cells were gated for the analysis oftheproliferativeresponsivenessofTeffs.Theprolifeativeresponseof Teffs in the presence of KO Tregs (row 2, 4, 6, 8) was less activethaninthepresenceofWTTregs(row1,3,5,7),suggestingKOTregswerehyperfunctional inthesuppressionof TeffsthanWTTregs. KO, GPx12/26 Cat2/2. Numbers indicate precursorfrequency (%) of Teffs representing proliferative activity. Arepresentative series of FACSplots of six separate experimentsshowingthesamepattern.(TIF)Figure S3 KOTeffs were hyperproliferative thanWTTeffs.CFSE-labeledTeffswerestimulatedwithvariousconcen-trationsofsolubleanti-CD3easindicatedinthepresenceofDCsfromWTorKOmice. Onthe3rdday, thecellswereharvestedandstainedforsurfaceCD4. LiveCD4+CFSE+cellsweregatedfor theanalysis of theproliferativeresponsiveness of Teffs. TheprolifeativeresponseofTeffswasslightlymorevigorousthanWTTeffsinthepresenceofWTDCs,butnotinthepresenceofKODCs. KO, GPx12/26 Cat2/2. Numbers indicate precursorROSIsCriticalinTregFunctionandExperimental ColitisPLOSONE | www.plosone.org 11 April2014 | Volume9 | Issue 4| e95332frequency (%) of Teffs representing proliferative activity. Arepresentative series of FACSplots of six separate experimentsshowingthesamepattern.(TIF)Figure S4 KOTeffs were hyperproliferative thanWTTeffs. The prolifeative response of Teffs was slightly morevigorous than WT Teffs in the presence of WT DCs, but not in thepresenceofKODCs. KO, GPx12/26Cat2/2.Dataaremean6SEofsixseparateexperiments.(TIF)FigureS5 Inducible (i) Treg differentiationwas compa-rable in KO mice. Na ve CD4+cells isolated from the spleens ofWTor KOmice were inducedto differentiate intoiTregs bystimulating in the presence of TGF-b1 and IL-2. iTregdifferentiation fromKOCD4+cells seemed slightly enhancedbutnotsignificant. KO, GPx12/26Cat2/2. Dataaremean6SEofsixseparateexperiments.(TIF)AuthorContributionsConceivedanddesignedtheexperiments:HRKBIMJYS.Performedtheexperiments: HRK AL EJC JHK WSL HKL. Analyzed the data: HRK ALEJCJHKBIMJYS. Contributedreagents/materials/analysistools: JHKALWSLHKLBIMJYS.Wrotethepaper:HRKALJYS.References1. ChenAF, ChenDD, Daiber A, Faraci FM, Li H, et al. (2012) Freeradicalbiologyofthecardiovascularsystem.ClinSci (Lond) 123: 7391.2. vanHorssenJ,WitteME,SchreibeltG,deVriesHE(2011)Radicalchangesinmultiplesclerosispathogenesis. BiochimBiophysActa1812: 141150.3. Zhu H, Li YR(2012) Oxidative stress and redox signaling mechanisms ofinflammatorybowel disease: updatedexperimental andclinical evidence. ExpBiol Med(Maywood) 237: 474480.4. Hultqvist M, OlssonLM, GeldermanKA, Holmdahl R(2009) TheprotectiveroleofROSinautoimmunedisease. TrendsImmunol 30: 201208.5. HultqvistM, OlofssonP,HolmbergJ, BackstromBT, TordssonJ, etal.(2004)Enhanced autoimmunity, arthritis, and encephalomyelitis in mice with areducedoxidativeburst duetoamutationintheNcf1gene. ProcNatl AcadSci USA101: 1264612651.6. Won HY, Jang EJ, Min HJ, Hwang ES (2011) Enhancement of Allergen-inducedAirwayInflammationbyNOX2Deficiency. ImmuneNetw11: 169174.7. Lee K, Won HY, Bae MA, Hong JH, Hwang ES (2011) Spontaneous and aging-dependent development of arthritis inNADPHoxidase2deficiencythroughaltered differentiation of CD11b+ and Th/Treg cells. Proc Natl Acad Sci U S A108: 95489553.8. De RavinSS, NaumannN, CowenEW, FriendJ, Hilligoss D, et al. (2008)Chronic granulomatous disease as a risk factor for autoimmune disease. J AllergyClinImmunol 122: 10971103.9. Cale CM, Morton L, Goldblatt D(2007) Cutaneous and other lupus-likesymptomsincarriersofX-linkedchronicgranulomatousdisease:incidenceandautoimmuneserology. ClinExpImmunol 148: 7984.10. deHaanJB,WittingPK,StefanovicN,PeteJ,DaskalakisM,etal.(2006)Lackof theantioxidantglutathioneperoxidase-1doesnotincreaseatherosclerosisinC57BL/J6micefedahigh-fatdiet. JLipidRes47: 11571167.11. Won HY, Sohn JH, Min HJ, Lee K, Woo HA, et al. (2010) Glutathioneperoxidase 1 deficiency attenuates allergen-induced airway inflammation bysuppressingTh2andTh17cell development. AntioxidRedoxSignal 13: 575587.12. Won HY, Jang EJ, Lee K, Oh S, KimHK, et al. (2013) Ablation ofperoxiredoxinII attenuates experimental colitis byincreasingFoxO1-inducedFoxp3+regulatoryTcells. JImmunol 191: 40294037.13. OlofssonP, HolmbergJ, Tordsson J,Lu S,AkerstromB,et al.(2003) Positionalidentification of Ncf1 as a gene that regulates arthritis severity in rats. Nat Genet33: 2532.14. Cemerski S, vanMeerwijkJP, Romagnoli P(2003) Oxidative-stress-inducedTlymphocytehyporesponsivenessiscausedbystructuralmodificationratherthanproteasomal degradationof crucial TCRsignalingmolecules. EurJ Immunol33: 21782185.15. Otsuji M, KimuraY, AoeT, OkamotoY, SaitoT(1996) Oxidativestress bytumor-derivedmacrophagessuppressestheexpressionofCD3zetachainofT-cell receptor complex andantigen-specific T-cell responses. Proc Natl AcadSci USA93: 1311913124.16. EfimovaO,SzankasiP,KelleyTW(2011)Ncf1(p47phox)isessentialfordirectregulatoryTcell mediatedsuppressionof CD4+effectorTcells. PLoSOne6:e16013.17. KraaijMD,SavageND,vanderKooijSW,KoekkoekK,WangJ,etal.(2010)Inductionof regulatoryTcellsbymacrophagesisdependentonproductionofreactiveoxygenspecies. ProcNatl AcadSci USA107: 1768617691.18. KimSJ, Lee JW, Jung YS, Kwon do Y, Park HK,et al.(2009) Ethanol-inducedliver injury and changes in sulfur amino acid metabolomics in glutathioneperoxidaseandcatalasedoubleknockoutmice. JHepatol 50: 11841191.19. JungYJ,SeohJY(2009)FeedbackloopofimmuneregulationbyCD4+CD25+Treg. Immunobiology214: 291302.20. KimHR, KieJH, LimW, MoonBI, KimSC, et al. (2012) CD4(+)CD25(+)regulatory T cells from MRL/lpr mice were functionally more active invitro butdid not prevent spontaneous as well as adriamycin-induced nephropathy invivo.Rheumatology(Oxford) 51: 13571367.21. OgasawaraN, OguroT, SakabeT, MatsushimaM, TakikawaO, etal. (2009)Hemoglobin induces the expression of indoleamine 2,3-dioxygenase in dendriticcellsthroughtheactivationofPI3K,PKC,andNF-kappaBandthegenerationofreactiveoxygenspecies. JCell Biochem108: 716725.22. Munn DH, Mellor AL(2013) Indoleamine 2,3 dioxygenase and metaboliccontrol ofimmuneresponses. TrendsImmunol 34: 137143.23. BoehmF, Martin M, Kesselring R, Schiechl G, Geissler EK, et al. (2012)Deletionof Foxp3+regulatory Tcells ingenetically targeted mice supportsdevelopmentofintestinal inflammation. BMCGastroenterol 12: 97.24. Amrouche-Mekkioui I, Djerdjouri B(2012) N-acetylcysteine improves redoxstatus, mitochondrial dysfunction, mucin-depletedcrypts andepithelial hyper-plasia in dextran sulfate sodium-induced oxidative colitis in mice.EurJPharmacol 691: 209217.25. HoYS, Magnenat JL, BronsonRT, CaoJ, GarganoM, et al. (1997) Micedeficient in cellular glutathione peroxidase develop normally and shownoincreasedsensitivitytohyperoxia. JBiol Chem272: 1664416651.26. Zhang Y, Ikeno Y, Qi W, Chaudhuri A, Li Y, et al. (2009) Mice deficient in bothMn superoxide dismutase and glutathione peroxidase-1 have increased oxidativedamage andagreater incidence of pathologybut noreductioninlongevity.JGerontol ABiol Sci MedSci 64: 12121220.27. EsworthyRS, ArandaR, MartinMG, DoroshowJH, BinderSW, etal. (2001)Mice with combined disruption of Gpx1 and Gpx2 genes have colitis.AmJPhysiol GastrointestLiverPhysiol 281: G848855.28. Kahles T, Brandes RP (2012) NADPHoxidases as therapeutic targets inischemicstroke. Cell Mol LifeSci 69: 23452363.29. Si CL, Shen T, Jiang YY, Wu L, Yu GJ, et al. (2013) Antioxidant properties andneuroprotective effects of isocampneoside II on hydrogen peroxide-inducedoxidativeinjuryinPC12cells. FoodChemToxicol 59: 145152.30. Thomas SR, Stocker R(1999) Redox reactions relatedto indoleamine 2,3-dioxygenaseandtryptophanmetabolismalongthekynureninepathway.RedoxRep4: 199220.31. Favre D, MoldJ, Hunt PW, Kanwar B, Loke P, et al. (2010) Tryptophancatabolismbyindoleamine 2,3-dioxygenase 1alters the balance of TH17toregulatoryTcellsinHIVdisease. Sci Transl Med2: 32ra36.32. TernessP, BauerTM,RoseL, DufterC, WatzlikA, etal. (2002) Inhibitionofallogeneic T cell proliferation by indoleamine 2,3-dioxygenase-expressingdendritic cells: mediation of suppression by tryptophan metabolites. J ExpMed196: 447457.33. Taylor MW, Feng GS (1991) Relationship between interferon-gamma,indoleamine2,3-dioxygenase, andtryptophancatabolism. FASEBJ 5: 25162522.34. Sakaguchi S, Sakaguchi N, Asano M, Itoh M, Toda M (1995) Immunologic self-tolerancemaintainedbyactivatedTcellsexpressingIL-2receptoralpha-chains(CD25). Breakdown of a single mechanismof self-tolerance causes variousautoimmunediseases. JImmunol 155: 11511164.35. WangJ, Ioan-FacsinayA, vander Voort EI, HuizingaTW, Toes RE(2007)Transient expressionof FOXP3inhumanactivatednonregulatoryCD4+Tcells. EurJImmunol 37: 129138.36. Martin F, Ladoire S, Mignot G, Apetoh L, Ghiringhelli F (2010) HumanFOXP3andcancer. Oncogene29: 41214129.37. KappJA, HonjoK, KappLM, GoldsmithK, BucyRP(2007) Antigen, inthepresence of TGF-beta, induces up-regulation of FoxP3gfp+ in CD4+ TCRtransgenicTcells that mediatelinkedsuppressionof CD8+Tcell responses.JImmunol 179: 21052114.38. JangMH, JungYJ, KimHR, MoonBI, SeohJY(2011) ImmuneHomeostasisMaintainedbyTregs. In: Hayashi RS, editor. RegulatoryTcells. NewYork:NovaSciencePublishers. 201223.39. Kim HR, Lee SM, Won JW, Lim W, Moon BI, et al. (2013) Functional changesin regulatory T cells during an experimental infection with sparganum(plerocercofidofSpirometramansoni).Immunology138: 5767.40. Kim HR, Moon S, Lee HK, Kang JL, Oh S, et al. (2012) Immune dysregulationinchronicstress:aquantitativeandfunctional assessmentofregulatoryTcells.Neuroimmunomodulation19: 187194.ROSIsCriticalinTregFunctionandExperimental ColitisPLOSONE | www.plosone.org 12 April2014 | Volume9 | Issue 4| e95332