EffectofPolyunsaturatedFattyAcid(PUFA)SupplementationonIntestinalIn - PDF document

EffectofPolyunsaturatedFattyAcid(PUFA)Su
EffectofPolyunsaturatedFattyAcid(PUFA)SupplementationonIntestinalIn¯ammationandNecrotizingEnterocolitis(NEC)inaNeonatalRatModelMICHAELS.CAPLAN,TANYARUSSELL,YUXIAO,MICHAELAMER,SUSANKAUP,ANDTAMASJILLINGDepartmentofPediatrics,EvanstonNorthwesternHealthcare,NorthwesternUniversityMedicalSchool,Evanston,Illinois60201,U.S.A.[M.S.C.,T.R.,Y.X.,M.A.,T.J.];andWyethNutritionalsInternational,Radnor,Pennsylvania10987,U.S.A.[S.K.]Inasmuchaslong-chainpolyunsaturatedfattyacids(PUFA,ReceivedSeptember28,2000;acceptedDecember20,2000.enteralnutrition,additionalanalyseshaveshownthattheinci-denceislowerinneonatesfedbreastmilkcomparedwithformula(11).Breastmilkcontainsseveralbioactivefactorsthatarenotpresentinprematureinfantformula,anditistheorizedthatthesecomponentsdown-regulatethein¯amma-torycascade,thusalteringtheincidenceorcourseofneonatalNEC(12±14).Recently,PUFAhavegeneratedsigni®cantinterestbecauseoftheirpotentialeffectsonneonatalnutrition,visualacuity,andintelligenceindevelopinginfants(15±17).Additionally,thesecompoundsarethoughttoin¯uencepros-taglandinmetabolismandcytokineactivation,andhavebeenassociatedwithareductioninneonatalNECinaprospectivesupplementationtrialofprematureinfants(18).Therefore,wehypothesizedthatPUFAsupplementationofformulainourneonatalratmodelofNECmayreducetheincidenceofdisease,andwedesignedthisstudytofurthercharacterizetheeffectsofthesecompoundsonthemetabolismofPAFandothermarkersofintestinalin¯ammation.METHODSAnimalmodel.NeonatalSprague-Dawleyratsweredeliv-eredviaabdominalincisionfromtime-datedpregnantfemales(Harlan,Inc,Indianapolis,IN,U.S.A.)onthe21stdayofgestationusingCO2anesthesia(60sinaprechargedchamber).Neonatalanimalswererecovered,dried,andwarmedinaneonatalincubator(Airshields,Hatboro,PA,U.S.A.)at37ÉCatmaximalhumidity.Animalsweredeliveredinthisfashiontopreventanymaternalmilkexposure,afactorthatimpactsonthedevelopmentofNEC.ToincreasethesusceptibilitytoNEC,neonatalanimalswereexposedtoasphyxiatwicedailyusing100%nitrogenfor50sfollowedbyexposuretocold(4ÉCfor10min).ControlanimalswerefedusingS-26/SMAprematureformula(WyethNutritionalsInternational,Radnor,PA,U.S.A.)concentratedtoprovide200kcal/kg/dwithanorogastricfeedingtube(1.9FSilasticcatheter,Bard,SaltLakeCity,UT,U.S.A.)beginningwith0.1mL/3handadvancedto0.3mL/3hby72hoflife.Usingthisprotocolofasphyxiaandformulafeeding,approximately75%ofneonatalratsdevelopclinicalandpathologicsignsofNECby72hoflife,includingabdominaldistention,bloodystools,vomiting,respiratorydis-tress,grosshemorrhagicintestinalnecrosis,andhistologicsignsvaryingbetweenmidvillousnecrosisandtransmuralnecrosis(19).ToinvestigatetheeffectofPUFAsupplementationofneo-natalformulaonthepathophysiologyofNEC,animalsweredividedintothefollowingexperimentalgroups:groupA,controlformulawithS-26/SMAat200kcal/kg/d;groupB,formulawithPUFAsupplementationcontaining34mg/100mLAAand23mg/100mLDHAtoprovideanAA/DHAratioof1.5:1.0;andgroupC,formulawithPUFAsupplementation(identicaltogroupBabove)andnucleotidesupplementation.Nucleotidesweresupplementedtoapproximatehumanmilkatthefollowingconcentrations(mg/100gformulapowder/200mLwater):CMP15.62versus3.55control,UMP9.2versus6.2control,IMP3.06versus0.73control,GMP2.08versus0.2control,andAMP3.57versus0.18control.NucleotidesweregivenwithPUFAinanadditionalexperimentalgroupbecause1)studieshaveshownnutritionalandanti-in¯ammatoryben-e®tsfromthesecompoundsthatarepresentinhumanmilk(20,21),and2)nucleotideshavebeenaddedtothecompositionofmostneonatalformulas.Allthreeformulaswerepremixedinpowderformandreconstitutedwithwater(100gpowderwith200mLwater)atthebeginningoftheexperimentalprotocol,andthecontentsofeachwereunknowntotheresearchtech-niciansandtheprincipalinvestigatorsuntilthecompletionofthestudy.Animalsweremonitoreduntil72±96hoflifeforclinicalsignsofNEC;moribundanimalsunabletoeatorwithsignsofrespiratorydistresswereeuthanized,andintestineswerefrozenforsubsequentanalysis.Inadditionalgroupsofanimals,neo-nateswereeuthanizedat24or48hoflifetoinvestigatepathophysiologicmechanismsrelatedtoNEC,includingintes-tinalapoptosis,pla

smaendotoxemia,andintestinaltissuegeneex
smaendotoxemia,andintestinaltissuegeneexpressionofkeymediatorsimplicatedinNEC,namelyPLA2,iNOS,andPAFreceptor.ThisprotocolwasreviewedandapprovedbytheInstitutionalAnimalCareandUseCommitteeofEvanstonNorthwesternHealthcare.EvaluationofNEC.Ateuthanasia,theintestinewasob-servedgrosslyforthepresenceofhemorrhagicnecrosis,andscoredasnone,mild(occasionalareasofviolaceous,in¯amedmucosa),moderate(multipleareasofnecrosis),orsevere(extensivenecrosisoccupying.50%oftheintestinallength).Subsequently,specimenswereplacedin10%formalinand®xedforsubsequenthistologicanalysis,andscoredas0(nor-malvilli),1(epithelialcellsloughing),2(midvillousnecrosis),3(completevillousnecrosis),or4(transmuralnecrosis).FluorescenceTUNELstaining.Althoughthereareseveralapproachestoquantifyapoptosis,studiesofanimaltissuesarebestaccomplishedusingaTUNELstaininsitutechnique(22).Theprincipleofthisprocedurerestsonthefactthatendonu-cleolyticdegradationofchromatinintodiscreteDNAfrag-ments(whichisthehallmarkofapoptosis)allowsincorpora-tionof¯uorescein-nucleotideconjugatesatthe3©-OHendofDNAunderthein¯uenceofTdT.Inshort,intestineisforma-lin-®xed,paraf®n-embedded,andcutontoaslide.Subse-quentlytheslideisdeparaf®nized,treatedwithproteinaseKtoimprovepermeability,andTdTenzymeaddedalongwithnucleotidemixincluding¯uorescein-dUTPconjugate.Equili-brationbufferisapplied,andthereactionisstoppedafter1h.Afterlabeling,preparationsarewashed,nucleiarecounter-stainedwithHOE33258,andthenslidesaremountedwithantifadereagent(MolecularProbes,Eugene,OR,U.S.A.).EachsampleisrunwithaTdTnegativecontroltoensurereliabilityofpositivelystainedapoptoticbodies.Slidesareexaminedwith¯uorescencedigitalimagingandassessedasmild(1,tipsofvilli),moderate(2,apoptoticfragmentsinmidvilli),orsevere(3,extensiveapoptosisthroughoutvilli).Endotoxinassay.PlasmaendotoxinwasassessedusingaspectrophotometricmethodprovidedinanLALassaycom-merciallyavailable(Biowhittaker,Walkersville,MD,U.S.A.).Brie¯y,endotoxincatalyzestheactivationofaproenzymeinLAL,andtherateofactivationisdeterminedbytheconcen-trationofendotoxinpresent.TheendotoxininasampleiscomparedwiththatobtainedfromastandardcurveofEsche-richiacoliendotoxinsuppliedbythecompany.648CAPLANETAL.RT-PCRforPLA2,PAF-receptor,andiNOS.TotalRNAwasisolatedfromfrozentissuesusingRNASTAT-60follow-ingthemanufacturer'sinstructions,TEL-TEST,INC,Friends-wood,TX,U.S.A.GenomicDNA,apossibleimpurityintheRNAextract,wassubsequentlyremovedbyaddingRNase-freeDNase.RNAintegritywasveri®edbyelectrophoresison1%formaldehyde-agarosegel.SemiquantitativeRT-PCRwasper-formedbyplacing1±2mgofnewbornratintestinetotalRNAintotubescontaining1.0mLofreversetranscriptionreactionmixture(13rTthreversetranscriptionbuffer,1mMMnCl2,2.5UrTthDNApolymerase,200mMeachdNTP,and1mMantisenseprimer:forPLA2,orrandomhexamersforiNOSandforb-actin.Themixturewasincubatedat37ÉCfor60min,thenterminatedonice.ThePCRreactionwasperformedin50mLofPCRmixture(13chelatingbuffer,2.5mMMgCl2,and0.5mMfollowingprimers:5©-TGACAGCATGAAGGTC-CTCC-3©/5©-TGGCATCCTTGGGGGATCCT-3©forPLA2-II,5©-ATGGCTTGCCCTTGGAAGTTTCTC-3©/5©-TCCAG-GCCATCTTGGTGGCAAAGA-3©foriNOS,5©-AAGTTC-CGGAAGCACCTCAGCG-3©/5©-TTAATTTTTCAATG-GCAGGAC-3©forPAFreceptor,or5©-GATCTTGATCTT-CATGGTGCTAGG-3©/5©-TTGTAACCAACTGGGAC-GATATGG-3©forb-actin.Reactionconditionswereopti-mizedforeachmRNAspeciestoensurethattheampli®ca-tionwasinthelinearphase.AfterthePCRampli®cation,theDNAproductsareseparatedbyelectrophoresison2%agarosegelcontainingSYBRGreenI(MolecularProbes,Eugene,OR,U.S.A.).ThedensityofeachDNAbandwasevaluatedwithStormPhosphorimager(MolecularDynam-ics,Sunnyvale,CA,U.S.A.)andanalyzedbyacomputersoftwareImageQuant(MolecularDynamics,Sunnyvale,CA,U.S.A.).RatiosbetweenbandintensitiesofPLA2/b-actinandiNOS/b-actinweredetermined.Statistics.x2analysisusinga332modelwasperformedtoanalyzedifferencesingroupsforNECoutcome.ComparisonsamongthethreegroupsforcontinuousdatawereperformedusingANOVA.Forallanalyses,p,0.05wasconsideredsigni®cant.RESULTSAnimalssupplementedwithPUFAhadasigni®cantlylowerincidenceofgrossandmicroscopicNEC(8of24)comparedwithcon

trols(17of24)andwithpupssupplementedwith
trols(17of24)andwithpupssupplementedwithPUFAandnucleotides(13of23;p,0.05;Table1).Animalsthatdevelopedintestinalnecrosistypicallybecameillbetween36and72hoflife,andshowedsignsoffrequentvomiting,abdominaldistension,bloodystools,andrespiratorydistress.Thegrossintestinalexaminationrevealedsegmentalhemor-rhagicnecrosisinmostanimalswithrarepupsexhibitingalmostcompletesmallintestinalnecrosis.Microscopicabnor-malitieswerefoundinallanimalswithevidenceofgrossnecrosis;however,thestageofhistologicinjurydidnotalwayscorrelatewiththeextent(lengthofbowel)ofgross®ndings(Fig.1).Allanimalsthatdiedbeforetheendofthestudyprotocol(96h)showedhistologicsignsofintestinalnecrosisinthisstudy,whereasafewhadnoidenti®ablemacroscopiclesions.SupplementationwithPUFAwithorwithoutnucleotidesdidnotaffecttheinitiationofapoptosisinintestinalepitheliumby48to72hinthisstudy.By48h,.50%ofanimalsshowedsigni®cantapoptoticactivitycharacterizedbyabundantTUNEL-positive¯uorescenceatthemidvilluslevel;however,,25%hadhistologicsignsofNEConmicroscopicanalysis.Therewerenodifferencesamonggroupsindegreeofapoptosisorextentatanytimepointsstudied(48hmedianscoresgroupA,1.8;B,2.1;andC,1.9).Table1.EffectofPUFAandPUFA1nucleotidesonNECanddeathinaneonatalratmodelListHistologicNECGrossNECDeathNosupplement17/2414/2417/24PUFA8/24*7/24*8/24*PUFA1nucleotides13/2312/2312/23*p,0.05Fig.1.PUFAsupplementationreducedtheseverityofhistologicNEC.Ani-malsweretreatedandassignedtodifferenttreatmentgroupsasdescribedinMethods.Animalswerekilledat48hoflife,andformalin-®xed,paraf®n-embeddedintestinesweresectionedandstainedwithhematoxylinandeosin.Imagesshownarerepresentativemicrographsofhematoxylinandeosin±stainedsectionsofthesmallintestinefrommother-fedanimals(a),formula-fedandcold-asphyxia±stressedanimals(b),andfromcold-asphyxia±stressedanimalsthatwerefedwithPUFA-supplementedformula(c).649PUFASUPPLEMENTSREDUCENECINRATSPlasmaendotoxemiawassimilaramongallthreegroupsat24h,butsigni®cantlyreducedinanimalssupplementedwithPUFAby48hoflife(Fig.2).IntestinalPLA2geneexpressionwasdecreasedinPUFA-supplementedanimalsat24h,butsimilarinPUFA1nucle-otideanimalscomparedwithcontrols.Therewerenodiffer-encesamonggroupsat48hinPLA2mRNAexpression(Fig.3).ExpressionofintestinalPAFreceptorwasreducedinthePUFA-supplementedgroupat48hcomparedwithcontrolandPUFA1nucleotidegroups(1.5160.27ratiotob-actinversus1.9160.23and2.1060.20;p,0.05),butnodifferentat24h(Fig.4).AnalysisofintestinaliNOSmRNArevealedsimilarresultsamongallgroupsat24-and48-htimepoints(Fig.5).DISCUSSIONOurresultsindicatethatPUFAsupplementationofneonatalformulainthisratmodelofNECreducestheincidenceofintestinalnecrosis.AdditionaloutcomeanalysesinourstudysuggestthatPUFAreducetheintestinalexpressionofthePAF-synthesizingenzymePLA2andthePAFreceptor,anddecreaseplasmaendotoxemia,buthavenoeffectonapoptosisofintestinalepitheliumorgeneexpressionofintestinaliNOS.Althoughourdatasupporttheclinical®ndingsrecentlyre-portedonPUFAsupplementationinprematureinfants(18),theyprovidenewinsightsintothemechanismsresponsibleforthesebene®cialeffects.Althoughprematureneonatesappeartosynthesizelong-chainPUFAfromessentialfattyacidprecursors,theymayrequiredietarysupplementationtoattainnormaltissueconcen-trationsofomega-3andomega-6fattyacids(23,24).Inaddition,theselong-chainPUFAarepresentinbreastmilk,andnursinginfantsabsorbsigni®cantquantitiesduringthe®rst4to6mooflife(25).Althoughde®ciencyofessentialfattyacidsresultsinabnormalgrowthandCNSdevelopment,supplemen-tationofneonatalformulawithAAandDHAhasbeenasso-ciatedwithimprovedvisualacuityandmentaldevelopmentinseveralrecentstudies(26,27).Furthermore,inarecentran-domizedtrialexaminingtheroleofPUFAsupplementationofprematureformulaonneonatalCNSdevelopmentandvisualfunction,Carlsonetal.(18)foundthatPUFAsupplementationsigni®cantlyreducedtheincidenceofNECinthetreatedgroupofprematureinfants.These®ndingsandothershavesuggestedFig.2.TheeffectofPUFAonplasmaendotoxinlevels.Serumsampleswerecollectedfromanimalsat24or48hoflife.EndotoxinlevelsweredeterminedfromserumsamplesusingtheLALmethod.Datas

hownaremean6SEMendotoxinlevelsatthetimep
hownaremean6SEMendotoxinlevelsatthetimepointsandcategoriesasindicatedonthe®gure(n58/groupforeachtimepoint).*denotesstatisticallysigni®cantdifferencefromthenosupplementgroup.Fig.3.EffectofPUFAonPLA2mRNAexpressioninthebowel.Intestineswereharvestedfromanimalsaftereuthanasiaat24or48hoflife.PLA2mRNAlevelsweredeterminedfromRNAisolatedfromintestinesusingsemiquantitativeRT-PCRasdescribedintheMethods.(a)representativeimageillustratingagarosegelelectrophoresisofPLA2andb-actinPCRproductsfromsamplestakenat48h.The764-bpbandcorrespondstob-actinandthe188-bpbandcorrespondstoPLA2.Lane1,mother-fedcontrolanimal;lane2,formula-fedandcold-asphyxia±stressedanimalwithnoformulasup-plement;andlane3,formula-fedandcold-asphyxia±stressedanimalwithPUFAsupplement.(b)mean6SEMPLA2mRNAlevelsatthetimepointsandcategoriesindicatedonthe®gure(n56/groupateachtimepoint).*representsstatisticallysigni®cantdifferencefromthenosupplementgroup.Fig.4.EffectofPUFAonPAFreceptor(PAFR)mRNAexpressioninthebowel.Intestineswereharvestedfromanimalsaftereuthanasiaat24or48hoflife.PAFRmRNAlevelsweredeterminedfromRNAisolatedfromintestinesusingsemiquantitativeRT-PCRasdescribedintheMethods.Mean6SEMPAFRmRNAlevelsatthetimepointsandcategoriesindicatedonthe®gure(n56/groupforeachtimepoint).*indicatesstatisticallysigni®cantdifferencefromthenosupplementgroup.650CAPLANETAL.thatPUFAmaymodulatethein¯ammatoryresponseinseveraltissues,includingtheintestine,andmaythereforeactasabene®cialsupplementofpretermformulaforNECprevention.ThepossiblemechanismswherebyPUFAin¯uencethein-¯ammatorycascadearenumerous.Studieshavesuggestedthat®shoilorotherlong-chainPUFAsupplementationmodulatestheproductionoreffectsofcytokinemediators(28),prosta-glandinmetabolites.includingthromboxanesandleukotrienes(29),signaltransductionproteins.includingproteinkinaseC(30),cellulartransportproteins(magnesium-ATPase)(31),andtranscriptionfactors.suchasNF-kB(32).Nonetheless,thespeci®croleofthesecompoundsinmodulatinggutin¯amma-tion(iftherearetrueeffects)remainsunclear.Arecentstudyusingashort-termhypoxia-inducedmodelofNECinmicehasshownthat®shoilsupplementation(primarilyomega-3fattyacids)reducedintestinalnecrosiswhiledecreasingintestinalPAFandleukotrieneconcentrationsintreatedanimals(33).Ourstudyshowsaclearreductionofintestinalin¯ammationandnecrosiswithAAandDHAsupplementation,andde-creasedPLA2geneexpressionat24h,therate-limitingen-zymeforPAFbiosynthesis.TheseresultsextendtheworkofAkisuetal.(33)inthatreducedintestinalPAFconcentrationsresultfromdecreasedPLA2geneexpression.Additionally,wefoundthatPUFAsupplementationattenuatedPAFreceptorexpressioninstressedintestinalhomogenates,therebyreduc-ingthetheoreticalPAFeffectongutepitheliumafteracceler-atedPAFproductionduringstress.Similartoourpreviousstudies,changesinPAF-relatedgeneexpressionpredatedtheonsetofintestinalinjury,butoccurredsimultaneouslywiththeriseinendotoxemia,therebysupportingaroleforalteredmucosalpermeabilityinthepathophysiology.Whereasinpre-viousstudiesofnewbornratNECwehaveshownthatiNOSexpression(34)andapoptosis(unpublishedobservations)areacceleratedafterformulafeedingandasphyxia,thePUFAsupplementationinthistrialhadnoeffectontheseoutcomes.Thus,thebene®cialeffectsofPUFAonintestinalhealthmayactonepithelialcellintegrityby1)reducingbacterialorendotoxintranslocation,or2)reducingmucosalPAFsynthesisandreceptoractivation,therebymodulatingthedirecteffectsofPAFonintestinalepithelium.Assuch,PAFhasbeenshowntodirectlyincreasemucosalpermeability(35,36)andisknowntostimulatetheactivationofseveralpotentsecondmessengersofin¯ammation(37±39).Inaddition,recentobservationsfromourlaboratoryinvitrohaveshownthatPAFdirectlystimulateschloridetransportinintestinalepitheliumviaactivationofamucosalPAFreceptor,leadingtointracellularacidosisandsubsequentcellulardysfunction.Insummary,PUFAsupple-mentationmodulatesPAFmetabolismandendotoxintranslo-cation,andthesemechanismsmayinpartexplainthebene®-cialeffectofthesecompoundsobservedinhumansandanimalsonintestinalnecrosis.Nonetheless,additionalstudiesevaluat-ingtheeffectofPUFAsupplemen

tationonproductionandactivityofintestina
tationonproductionandactivityofintestinalmediators(cytokines,includingIL-1,-6,-10,and-11,andprostaglandinmetabolites)wouldbeworthwhile.Dietarynucleotidesareanothergroupofbioactivemoleculesthatarepresentinhumanmilk,andhaverecentlybeenaddedtomanytermandpretermformulapreparations.Nucleotideshavebeenshowntomodulatethehumoralandcellularimmuneresponseinsupplementedinfants(20,21),andthereforemaymodulatethein¯ammatoryresponsetopathogensintheintes-tinalmilieu.Therefore,althoughtheaimofthisstudywasnottoevaluatetheroleofnucleotidesonintestinalin¯ammationornecrosis,weincludedastudygroupofPUFAandnucleotidesupplementationinasmuchasprematureinfantswilllikelybeexposedtobothpreparations,andthepresenceofnucleotidescouldin¯uencePUFAeffects.Surprisingly,wefoundthatnucleotidesupplementationabrogatedthebene®cialeffectsofPUFAsupplementationondeath,intestinalnecrosis,endotox-emia,andPLA2andPAFreceptormRNAexpressioninourneonatalmodel.Althoughthe®ndingssuggestthatnucleotidesmayattenuatetheeffectsofPUFAonintestinalin¯ammation,additionalstudiesarerequiredtoclarifyanyspeci®croletheymayplayinintestinalin¯ammationandhealth.Inconclusion,PUFAsupplementationsigni®cantlyreducedtheincidenceofNECintheneonatalratmodelwhiledown-regulatingPAFproduction,PAFreceptorsynthesis,andendo-toxintranslocationintothesystemiccirculation.Althoughthesedatasupporttheresultsofaprospectivetrialinpreterminfants,alargerrandomized,controlledtrialofPUFAsupple-Fig.5.EffectofPUFAoniNOSmRNAexpressioninthebowel.Intestineswereharvestedfromanimalsaftereuthanasiaat24or48hoflife.iNOSmRNAlevelsweredeterminedfromRNAisolatedfromintestinesusingsemiquantitativeRT-PCRasdescribedintheMethods.(a)representativeimageillustratingagarosegelelectrophoresisofiNOSandb-actinPCRproductsfromsamplestakenat48h.The764-bpbandcorrespondstob-actinandthe436-bpbandcorrespondstoiNOS(1:10dilution).Lane1,mother-fedcontrolanimal;lane2,formula-fedandcold-asphyxia±stressedanimalwithnoformulasupplement;andlane3,formula-fedandcold-asphyxia±stressedanimalwithPUFAsupplement.(b)datashownaremean6SEMiNOSmRNAlevelsatthetimepointsandcategoriesindicatedonthe®gure(n56/groupforeachtimepoint).651PUFASUPPLEMENTSREDUCENECINRATSmentationinprematureformulatopreventNECseemswarranted.REFERENCES1.KliegmanRM,FanaroffAA1984Necrotizingenterocolitis.NEnglJMed310:1093±11032.UauyRD,FanaroffAA,KoronesSB,PhillipsEA,PhillipsJB,WrightLL1991Necrotizingenterocolitisinverylowbirthweightinfants:biodemographicandclinicalcorrelates.NationalInstituteofChildHealthandHumanDevelopmentNeonatalResearchNetwork.JPediatr119:630±6383.NanthakumarNN,FusunyanRD,SandersonI,WalkerWA2000In¯ammationinthedevelopinghumanintestine:apossiblepathophysiologiccontributiontonecrotizingenterocolitis.ProcNatlAcadSciUSA97:6043±60484.KliegmanRM1990Modelsofthepathogenesisofnecrotizingenterocolitis.JPediatr117(suppl):S2±S55.CaplanMS,MacKendrickW1993Necrotizingenterocolitis:areviewofpathogeneticmechanismsandimplicationsforprevention.PediatrPathol13:357±3696.CaplanM,HsuehW,KellyA,DonovanM1990SerumPafacetylhydrolaseincreasesduringneonatalmaturation.Prostaglandins39:705±7147.HsuehW,CaplanMS,SunX,TanX,MacKendrickW,Gonzalez-CrussiF1994Platelet-activatingfactor,tumornecrosisfactor,hypoxiaandnecrotizingenterocolitis.ActaPaediatrSuppl396:11±178.Gonzalez-CrussiF,HsuehW1983Experimentalmodelofischemicbowelnecrosis.Theroleofplatelet-activatingfactorandendotoxin.AmJPathol112:127±1359.CaplanMS,HedlundE,AdlerL,LickermanM,HsuehW1997Theplatelet-activatingfactorreceptorantagonistWEB2170preventsneonatalnecrotizingentero-colitisinrats.JPediatrGastroenterolNutr24:296±30110.CaplanMS,LickermanM,AdlerL,DietschGN,YuA1997Theroleofrecombinantplatelet-activatingfactoracetylhydrolaseinaneonatalratmodelofnecrotizingenterocolitis.PediatrRes42:779±78311.LucasA,ColeTJ1990Breastmilkandneonatalnecrotisingenterocolitis.Lancet336:1519±152312.SchanlerRJ,AtkinsonSA1999Effectsofnutrientsinhumanmilkontherecipientprematureinfant.JMammaryGlandBiolNeoplasia4:297±30713.HansonLA1999Humanmilkandhostdefence:immediateandlong-termeffects.ActaPaediatrSup

pl88:42±4614.WalkerWA1997Breastmilkandth
pl88:42±4614.WalkerWA1997Breastmilkandthepreventionofneonatalandpretermgastroin-testinaldiseasestates:anewperspective.ChungHuaMinKuoHsiaoErhKoIHsuehHuiTsaChih38:321±33115.GibsonRA,MakridesM1998Theroleoflongchainpolyunsaturatedfattyacids(LCPUFA)inneonatalnutrition.ActaPaediatr87:1017±102216.CarlsonSE1999Long-chainpolyunsaturatedfattyacidsanddevelopmentofhumaninfants.ActaPaediatrSuppl88:72±7717.UauyR,HoffmanDR2000Essentialfatrequirementsofpreterminfants.AmJClinNutr71(suppl):245S±250S18.CarlsonSE,MontaltoMB,PonderDL,WerkmanSH,KoronesSB1998Lowerincidenceofnecrotizingenterocolitisininfantsfedapretermformulawitheggphospholipids.PediatrRes44:491±49819.CaplanMS,HedlundE,AdlerL,HsuehW1994Roleofasphyxiaandfeedinginaneonatalratmodelofnecrotizingenterocolitis.PediatrPathol14:1017±102820.CarverJD,PimentelB,CoxWI,BarnessLA1991Dietarynucleotideeffectsuponimmunefunctionininfants.Pediatrics88:359±36321.PickeringLK,GranoffDM,EricksonJR,MasorML,CordleCT,SchallerJP,WinshipTR,PauleCL,HiltyMD1998Modulationoftheimmunesystembyhumanmilkandinfantformulacontainingnucleotides.Pediatrics101:242±24922.GavrieliY,ShermanY,Ben-SassonSA1992Identi®cationofprogrammedcelldeathinsituviaspeci®clabelingofnuclearDNAfragmentation.JCellBiol119:493±50123.ClandininMT,ChappellJE,LeongS,HeimT,SwyerPR,ChanceGW1980Intrauterinefattyacidaccretionratesinhumanbrain:implicationsforfattyacidrequirements.EarlyHumDev4:121±12924.UauyR,MenaP1999Requirementsforlong-chainpolyunsaturatedfattyacidsinthepreterminfant.CurrOpinPediatr11:115±12025.GibsonRA,NeumannMA,MakridesM1997Effectofincreasingbreastmilkdocosahexaenoicacidonplasmaanderythrocytephospholipidfattyacidsandneuralindicesofexclusivelybreastfedinfants.EurJClinNutr51:578±58426.BirchDG,BirchEE,HoffmanDR,UauyRD1992Retinaldevelopmentinvery-low-birth-weightinfantsfeddietsdifferinginomega-3fattyacids.InvestOphthalmolVisSci33:2365±237627.CarlsonSE,WerkmanSH,TolleyEA1996Effectoflong-chainn-3fattyacidsupplementationonvisualacuityandgrowthofpreterminfantswithandwithoutbronchopulmonarydysplasia.AmJClinNutr63:687±69728.BlokWL,KatanMB,vanderMeerJW1996Modulationofin¯ammationandcytokineproductionbydietary(n-3)fattyacids.JNutr126:1515±153329.HuangMC,Craig-SchmidtMC1996Arachidonateanddocosahexaenoateaddedtoinfantformulain¯uencefattyacidcompositionandsubsequenteicosanoidproductioninneonatalpigs.JNutr126:2199±220830.RiseP,GalliC1999Arachidonicanddocosahexaenoicacidsdifferentiallyaffecttheexpressionoffattyacyl-CoAoxidase,proteinkinasecandlipidperoxidationinhepg2cells.ProstaglandinsLeukotEssentFattyAcids60:367±37031.HaagM,VermeulenF,MagadaO,KrugerMC1999PolyunsaturatedfattyacidsinhibitMg21-ATPaseinbasolateralmembranesfromratenterocytes.ProstaglandinsLeukotEssentFattyAcids61:25±2732.RusynI,BradhamCA,CohnL,SchoonhovenR,SwenbergJA,BrennerDA,ThurmanRG1999Cornoilrapidlyactivatesnuclearfactor-kbinhepaticKupffercellsbyoxidant-dependentmechanisms.Carcinogenesis20:2095±210033.AkisuM,BakaM,CokerI,KultursayN,HuseyinovA1998Effectofdietaryn-3fattyacidsonhypoxia-inducednecrotizingenterocolitisinyoungmice.N-3fattyacidsalterplatelet-activatingfactorandleukotrieneB4productionintheintestine.BiolNeonate74:31±3834.MacKendrickW,CaplanM,HsuehW1993Endogenousnitricoxideprotectsagainstplatelet-activatingfactor-inducedbowelinjuryintherat.PediatrRes34:222±22835.KubesP,ArforsKE,GrangerDN1991Platelet-activatingfactor-inducedmucosaldysfunction:roleofoxidantsandgranulocytes.AmJPhysiol260:G965±G97136.OtamiriT,LindahlM,TagessonC1988PhospholipaseA2inhibitionpreventsmucosaldamageassociatedwithsmallintestinalischaemiainrats.Gut29:489±49437.HsuehW,Gonzalez-CrussiF1988Ischemicbowelnecrosisinducedbyplatelet-activatingfactor:anexperimentalmodel.MethodsAchievExpPathol13:208±23938.HuangL,TanX,CrawfordSE,HsuehW1994Platelet-activatingfactorandendotoxininducetumournecrosisfactorgeneexpressioninratintestineandliver.Immunology83:65±6939.HsuehW,Gonzalez-CrussiF,ArroyaveJL1986Platelet-activatingfactor-inducedischemicbowelnecrosis.Aninvestigationofsecondarymediatorsinitspathogenesis.AmJPathol122:231±239652CAPLANETA