RESEARCHARTICLE ‘ Circulating MiRNAs of Asian Indian ’ Phenotype Identified in Subjects with Impaired Glucose Tolerance and Patients with Type 2 Diabetes ParamasivamPrabu1,SophieRome2,ChandrakumarSathishkumar1, SankaramoorthyAravind1,BalakumarMahalingam1,Coimbatore SubramanianShanthirani1,CarolineGastebois2,AudreyVillard2,ViswanathanMohan1, MuthuswamyBalasubramanyam1* 1 MadrasDiabetesResearchFoundationandDr.Mohan’sDiabetesSpecialitiesCentre,WHOCollaborating CentreforNon-CommunicableDiseasesPreventionandControl&IDFCentreofEducation,Gopalapuram, Chennai-600086,India,2 CarMeNLaboratory(INSERM1060,INRA1397,INSA),UniversityofLyon, FacultédeMédecineLyon-Sud,CheminduGrandRevoyet,69600,Oullins,France * [email protected] OPENACCESS Citation:PrabuP,RomeS,SathishkumarC,Aravind Abstract S,MahalingamB,ShanthiraniCS,etal.(2015) CirculatingMiRNAsof‘AsianIndianPhenotype’ IdentifiedinSubjectswithImpairedGlucose Severalomicstechnologiesareunderwayworldwidewithanaimtounravelthepathophysi- ToleranceandPatientswithType2Diabetes.PLoS ologyofacomplexphenotypesuchastype2diabetesmellitus(T2DM).Whilerecentstud- ONE10(5):e0128372.doi:10.1371/journal. iesimplyaclinicallyrelevantandpotentialbiomarkerroleofcirculatorymiRNAsinthe pone.0128372 etiologyofT2DM,thereislackofdataonthisaspectinIndians—anethnicpopulationchar- AcademicEditor:RatnaB.Ray,SAINTLOUIS acterizedtorepresent‘AsianIndianphenotype’knowntobemorepronetodevelopT2DM UNIVERSITY,UNITEDSTATES andcardiovasculardiseasethanEuropeans.WeperformedglobalserummiRNAprofiling Received:February14,2015 andthevalidationofcandidatemiRNAsbyqRT-PCRinacohortofsubjectscomprisedof Accepted:April26,2015 normalglucosetolerance(NGT),impairedglucosetolerance(IGT)andpatientswithT2DM. Published:May28,2015 Ourstudyrevealed4differentiallyexpressedmiRNAs(miR-128,miR-130b-3p,miR-374a- 5p,miR-423-5p)insubjectswithIGTandT2DMpatientscomparedtocontrolsubjects. Copyright:©2015Prabuetal.Thisisanopen accessarticledistributedunderthetermsofthe Theywerepositivelyornegativelycorrelatedtocholesterollevels,HbA1C,HOMA-IRand CreativeCommonsAttributionLicense,whichpermits fastinginsulin.Interestingly,circulatinglevelofmiR-128andmiR-130b-3pwerealsoaltered unrestricteduse,distribution,andreproductioninany inserumofdiet-induceddiabeticmicecomparedtocontrolanimals.Amongthealteredcir- medium,providedtheoriginalauthorandsourceare culatingmiRNAs,miR-128hadneverbeendescribedinpreviousstudies/populationsand credited. appearedtobea‘NewLead’inIndians.Itwaspositivelycorrelatedwithcholesterolbothin DataAvailabilityStatement:Allrelevantdataare prediabeticsubjectsandindiet-induceddiabeticmice,suggestingthatitsincreasedlevel withinthepaperanditsSupportingInformationfiles. mightbeassociatedwiththedevelopmentofdyslipedemiaassociatedwithT2DM.Ourfind- Funding:Authorsacknowledgethefinancialsupport ingsimplydirectionalitytowardsbiomarkerpotentialofmiRNAsintheprevention/diagnosis/ fromtheDepartmentofBiotechnology(DBT), GovernmentofIndia.Thisworkwasalsosupported treatmentoutcomesofdiabetes. bygrantsfromtheFrenchFondationpourla RechercheMédicale(FRM-20101220456)andis supportedbytheFrench—Indianmedicalresearch cooperationprogrambetweenINSERMandICMR. Theauthorsthankthestudyparticipants,andthe clinicalassistantsandadministrativepersonnelatthe PLOSONE|DOI:10.1371/journal.pone.0128372 May28,2015 1/14 CirculatoryMiRNAsofAsianIndianPhenotype MadrasDiabetesResearchFoundation,Chennai, Introduction India. SouthAsiaformsoneoftheepicentersoftheglobaldiabetesepidemic.Overthepastcoupleof CompetingInterests:Theauthorshavedeclared decades,therehasbeenaworryingincreaseintheprevalenceratesoftype2diabetesmellitus thatnocompetinginterestsexist. (T2DM)intheregionandIndiainparticular.ThehigherprevalenceofdiabetesinSouth Asiansisleastexplainedbytraditionallymeasuredriskfactors.Insulinresistanceandabnor- malitiesofinsulinsecretioninpancreaticß-cellsarethemaindefectsthatleadtoT2DM.The socalled“AsianIndianPhenotype”referstocertainuniqueclinicalandbiochemicalabnormal- itiesinIndianswhichincludeincreasedinsulinresistance,greaterabdominaladiposityi.e., higherwaistcircumferencedespitelowerbodymassindexandT2DMoccurringatayounger agethanCaucasians[1,2].SouthAsiansarealsocharacterizedbyauniquemetabolicprofile withhigherinsulinlevels[3],agreaterdegreeofinsulinresistance[4]andahigherprevalence ofdiabetes[5].InsulinresistancehasbeendemonstratedinAsianIndiansevenduringadoles- cence[6]andhyperinsulinemiaseemstobepresentinAsianIndiansevenatbirth[7].Morere- centstudiesdoimplythatSouthAsiansmayhaveanearlydeclineinß-cellfunctionsaswell [8,9,2]. GiventhepredictedexplosioninthenumberofcasesofprediabetesandT2DMworldwide andinIndia,continuedresearchisessential,particularlyinthenewerareassuchasapplication ofmicroRNA(miRNA)technologies.MiRNAsareaclassofevolutionallyconservednon-cod- ingRNAsof19–22nucleotidesandfunctionasnegativeregulatorsofgeneexpression[10].Im- portantrolesofmiRNAshaveemergedinthecontrolofmetabolicpathwaysinvolvedinlipid metabolism,adipocytedifferentiationandpancreasdevelopment,energyhomeostasis,glucose- stimulatedinsulinsecretionandinflammation[11–16].Recently,significantamountsofmiR- NAshavebeenfoundnotonlyintracellularly,butinextracellularhumanbodyfluids(e.g.; serum,plasma,saliva,urine,tears,amnioticfluidsandmilk)[17–19].Theyareremarkablysta- bledespitehighextracellularRNAseactivities.ExtracellularmiRNAsareenclosedinsmall membranousvesicles(e.g.;inexosomes,sheddingvesicles,apoptoticbodies)orassociated with,orpackagedwithinhigh-densitylipoprotein,orassociatedwithRNA-bindingproteins(e. g.;high-densitylipoprotein,Argonaute2andnucleophosmin1)[20–22].LevelsofmiRNAsin theserumofhumanshavebeenshowntobestable,reproducible,consistentamongsthealthy individualsandchangeduringpathophysiology,allowingthemtobeofpotentialvalueasbio- markersofdisease[23].Recently,novelcirculatorymiRNAsignatureshavebeenstudiedand shownassociatednotonlywiththediseasebutalsotheseverityofdiseasessuchasvariouscan- cers,liverandcardiovasculardiseases[24,25].Whereasseveralstudieshaveevaluatedtherole ofmiRNAsincancer,muchlessisknowninthefieldofdiabetes.Differentiallyexpressedcircu- latorymiRNAsignaturecharacteristicofT2DMhavebeenreportedfromrecentstudiesout- sidetheIndiabasedontheanalysisofsmallgroupofsubjects[26–33].Fromthesestudies,itis expectedthattheidentificationofpotential-specificbiomarkersofmiRNAsmayhelppredict ordetectthedevelopmentandprogressionofdiabetesanditscomplicationsatanearlystage, andthereforeallowtimelyintervention.However,thereislackofclinicallyrelevantdataonthe potentialofcirculatorymiRNAsinIndians,whoarehighlyinsulinresistantandmoreproneto developT2DMandcardiovasculardiseasethanEuropeans[5].Whilethehealthburdenof T2DMissomehowassociatedwithobesityasaco-morbidityinCaucasians,metabolicriskfac- torsareeitherhigherinAsianIndiansindependentofobesityoroperateatalowerBMIthresh- old.Therefore,thisstudyisaimedtoidentifycirculatingmiRNAsthatcouldbedifferentially expressedinsubjectswithprediabetesandpatientswithT2DMandtodissectouttheirpoten- tialbiomarkersroleinassociationwithclinicalparameters/conventionalriskfactorsofT2DM. PLOSONE|DOI:10.1371/journal.pone.0128372 May28,2015 2/14 CirculatoryMiRNAsofAsianIndianPhenotype ResearchDesignandMethods RecruitmentoftheStudySubjects SubjectswithNGT(NormalGlucoseTolerance,n=49),subjectswithIGT(ImpairedGlucose Tolerance,n=47)andpatientswithType2DiabetesMellitus(T2DM,n=49)wererecruited fromDr.Mohans’DiabetesSpecialitiesCentre,Chennaiandfromtheon-goingepidemilogical studies.ThestudywasconductedaccordingtotheprinciplesofDeclarationofHelsinkiandap- propriateapprovalbytheInstitutionalEthicsCommitteeoftheMadrasDiabetesResearch Foundation.EthicsCommitteeapprovedwritteninformedconsentwasobtainedfromall studysubjects.NGT,IGTandT2DMweredefinedusingWorldHealthOrganizationconsult- inggroupcriteria.Thosewhowereconfirmedbyoralglucosetolerancetesttohave2-hour plasmaglucosevalue11.1mmol/L(200mg/dL)ormorebasedonWorldHealthOrganization consultinggroupcriteriawerediagnosedasdiabeticpatients.Thosewith2-hourpostglucose value(cid:1)7.8mmol/L(140mg/dL)and<11.1mmol/L(200mg/dL)werediagnosedassubjects withIGT,andthosewith2-hourpostglucosevalueoflessthan7.8mmol/L(140mg/dL)as subjectswithNGT.Toruleoutthemedicationeffects,allourT2DMsubjectsareofnewly diagnosednature. AnthropometricMeasurements Anthropometricmeasurementsincludingheight,weightandwaistcircumstancewereobtained usingstandardizedtechniques.Heightwasnotedwithatapemeasuredtothenearestcentime- ter.Weightwasmeasuredwithtraditionalspringbalancethatwaskeptonafirmhorizontal surface.Thebodymassindex(BMI)wascalculatedusingtheformula,weight(kg)/height(m)2. Waistcircumferencewasmeasuredusinganonstretchablefibermeasuringtape.Bloodpres- surewasrecordedfromtherightarmofstudysubjectswhentheywererelaxedandinsitting positiontothenearest2mmHgwithamercurysphygmomanometer(DiamondDeluxeBPap- paratus,Pune,India).Tworeadingweretaken5minapartandthemeanofthetwowastaken asthebloodpressure. BiochemicalParameterInvestigations Fastingplasmaglucose(glucoseoxidase-peroxidasemethod),serumcholesterol(cholesterol oxidase-peroxidase-amidopyrinemethod),serumtriglycerides(glycerolphosphateoxidase- peroxidase-amidopyrinemethod)andHDLcholesterol(directmethod-polyethyleneglycol- pretreatedenzymes)weremeasuredusingHitachi-912Autoanalyser(Hitachi,Mannheim, Germany).Theintraandinterassayco-efficientofvariationforthebiochemicalassayswere <5%.Low-densitylipoprotein(LDL)cholesterolwascalculatedusingFriedewaldformula. Glycatedhemoglobin(HbAlc)wasestimatedbyhigh-pressureliquidchromatographyusing thevariantanalyzer(Bio-Rad,Hercules,Calif.,USA).SerumInsulinwasestimatedusingen- zyme-linkedimmunosorbentassay(Calbiotech,CA).Theintra-assayandtheinter-assaycoef- ficientsofvariationforinsulinassaywas<10%.Insulinresistancewascalculatedusingthe homeostasisassessmentmodel(HOMA-IR)usingtheformula:{fastinginsulin(μIU/mL)x fastingglucose(mmol/L)}/22.5. CirculatingRNAExtractionandPurification Fastingserumsample(from5mlofblood)wasobtainedbystandardvenepunctureusing VacutainerPlusPlasticSerumandSSTTubes(Becton-Dickinson,Franklin,lakes,NJ).Thesep- arationoftheserumwasperformedbycentrifugationat4,000rpmfor10min,followedby 12000rpmfor15mintocompletelyremovecelldebris.TotalRNAwasisolatedfrom0.25ml PLOSONE|DOI:10.1371/journal.pone.0128372 May28,2015 3/14 CirculatoryMiRNAsofAsianIndianPhenotype Table1. Clinicalandbiochemicalcharacteristicsofthesubjectsinvolvedinthestudy. Wholecohort Subgroupofsubjectsusedtoselectcandidate miRNAs Parameters NGT IGT T2DM NGT IGT T2DM (n=49) (n=47) (n=49) (n=12) (n=12) (n=12) Males/females (26/23) (23/24) (25/24) (6/6) (6/6) (5/7) Age(years) 44.3±6.9 44.1±7.0 44.4±8.1 38.5±3.0 41.2±4.0 40.2±5.0 BMI(kg/m2) 24.5±2.6 24.9±2.9 25.7±3.5 23.1±2.0 23.9±1.0 26.7±1.0 Fastingplasmaglucose(mg/dl) 85±11 105±18* 146±34* 83±7.0 101±7.0* 150±27* 2hrplasmaglucose(mg/dl) 96±21 166±18* 269±63* 89±24 160±20* 264±61* HbA1c(%) 5.6±0.4 6.3±0.8* 7.8±1.6* 5.2±0.5 6.0±0.5 7.9±±1.0* HOMA-IR 2.0±1.1 3.5±1.4* 6.3±1.8* 1.5±0.4 2.7±1.0 5.0±3.0* SerumCholesterol(mg/dl) 170±24 187±35 185±52 168±29 185±39 185±52 SerumTriglycerides(mg/dl) 97±38 148±76 136±62 119±65 134±69 148±67 HDLcholesterol(mg/dl) 41.2±9.0 40±5.0 37±7.0 41±9.0 41±10 40±8.0 LDLcholesterol(mg/dl) 111±22 117±32 121±46 107±26 117±33 115±44 VLDL(mg/dl) 19±7.6 30±16 25±13 24±13 27±14 29±13 FastingInsulin(μIU/ml) 7.6±2.0 11±4.0 13±6.7 9.6±4.6 13.4±4.6* 17.8±4.0* Systolicbloodpressure(mmHg) 117±9.0 121±12 124±12 123±20 124±12 126±16 Diastolicbloodpressure(mmHg) 78±10 82±5.0 81±7.0 80±10 80±8.0 84±9 Waistcircumference(cm) 87.1±1.21 93.8±1.60* 92.3±1.38* 87.3±8.4 93.5±10.9* 92.3±9.6* Allthevaluerepresentsmeanand±standarddeviation *p<0.05comparedcontrol ND,notdetermined doi:10.1371/journal.pone.0128372.t001 ofserumusingQiagenmiRNeasyMinikit(Qiagen,Valencia,CA)forRNAcollectionandpu- rificationaccordingtothemanufacturer'sprotocol.TotalRNAwascontrolledusingNanodrop 2000UVspectrophotometerandstoredat-80°Cuntiluse. GlobalSerummicroRNAProfiling GlobalserummiRNAprofilingwasdoneinadiscoverycohortcomprisedof12individuals eachofNGT,IGTandT2DM(Table1).ToassesshemolysistwomicroRNAswereused.One thatisexpressedinredbloodcells(miRNA-451),andonethatisrelativelystableinserumand plasmaandnotaffectedbyhemolysis(miRNA-23a).TheratiobetweenthesetwomiRNAscor- relatestodegreeofhemolysis.Inourexperiencesampleswithratiosabove8.0willhaveanin- creasedriskofbeingaffectedbyhemolysis.Wevalidatedthatoursampleshadlowerratiosand werenotaffectedbyhemolysis.DuetothelowlevelsofmicroRNAsandpotentiallyhighlevels ofinhibitorsinsamplesderivedfromserum/plasmawhichalsodiffersfromsampletosample, Exiqonprotocol(www.exiqon.com/serum-plasma-guidelines)recommendsusingRNA amountsbasedonstartingvolumeratherthanRNAquantity.8μLofelutedhumanplasma/ serumRNAina40μLRTreactiongenerallygiveagoodsignalwithmaximalmiRNAdetection inmiRNAPCRpanels[34].Thereforeinourstudy,8μlRNAwasreversetranscribedin40μl reactionsusingthemiRCURYLNAUniversalRTmicroRNAPCRcDNAsynthesiskit.An RNAspike-incontrol(Sp6)wasaddedtothereversetranscriptionstep.Thiscontrolisusedto confirmthatthereversetranscriptionandamplificationoccurswithequalefficiencyinallsam- ples.cDNAwasdiluted50xandassayedin10μlPCRreactionsaccordingtotheprotocolfor miRCURYLNAUniversalRTmicroRNAPCR;eachmicroRNAwasassayedoncebyqPCRon PLOSONE|DOI:10.1371/journal.pone.0128372 May28,2015 4/14 CirculatoryMiRNAsofAsianIndianPhenotype themicroRNAReady-to-UsePCR(miRCURYLNAmicroRNAHumanpanelIExiqon).Neg- ativecontrolsexcludingtemplatefromthereversetranscriptionreactionwasperformedand profiledlikethesamples.TheamplificationwasperformedinaLightCycler480Real-Time PCRSystem(Roche)in384wellplates.Theamplificationcurveswereanalyzedusingthe RocheLCsoftware,bothfordeterminationofCp(bythe2ndderivativemethod)andformelt- ingcurveanalysis.AmplificationefficiencywascalculatedusingalgorithmssimilartotheLin- Regsoftware.AllassayswereinspectedfordistinctmeltingcurvesandtheTmwascheckedto bewithinknownspecificationsfortheassay.Furthermoreassaysmustbedetectedwith5Cp’s lessthanthenegativecontrol,andwithCp<37tobeincludedinthedataanalysis.Datathat didnotpassthesecriteriawereomittedfromanyfurtheranalysis.UsingNormFinderthebest normalizationprocedurewasfoundtobetheaverageofassaysdetectedinallsamples.Alldata werenormalizedtotheaverageofassaysdetectedinallsamples(average—assayCp).Compari- sonbetweengroupsweremadebyusingthestudentt-test(p<0.05)onnormalizeddata,tose- lectthedifferentiallyexpressedmiRNAs. ValidationofcandidatemiRNAsbyindividualqRT-PCR miRNAsshortlistedasdifferentiallyexpressedfromthediscoverysetwererevalidatedina largercohortcomprisingof145individuals(50%men,50%women)withNGT(n=49),IGT (n=47)andT2DM(n=49)byindividualPCRassays.TotalRNAfromserumwasextracted asdescribedaboveandusedforqRT-PCRbyusingtheUniversalmiRCURYLNAmicroRNA PCR,PolyadenylationandcDNAsynthesiskitIIandthemiRCURYLNAmicroRNAPCRsys- tem,ExilentSYBRgreenmastermix,fromExiqon,onABI7000AppliedBiosystemsthermo- cycler.RelativeexpressionofmicroRNAexpressionwascalculatedby2ˆ(-"DeltaCt")method.We evaluatedasuitablenumberofreferencemiRNAs,basedontheincreasedexpressionstability [28].However,wewereunabletoidentifystablemiRNAsthatdidnotdiscriminatebetween maleandfemaleorthatwerenotcorrelatedwithmetabolicparameters,orthatwasnotprevi- ouslyidentifiedasrelevantforthepathologyinpreviousstudies.Thiscouldbeexplainedpartly bythefactthatinsidethesamegroup,womenandmenmightdifferedintheirlevelofcholes- terolandtriglycerides(S1Table).Thus,inthisstudy,datawerenormalisedonthevolumeof totalRNAthatwasusedforqRT-PCR(i.e.;afixedvolumeof8μl). StatisticalMethods StatisticalanalyseswereperformedwiththeSPSSstatisticalsoftware(SPSSV12.0,Inc.,Chi- cago,IL),andtheRStatisticalSoftware(http://www.r-project.org/).ANOVAand/orpairedt- testswereperformedtostudydifferencesonquantitativevariablesbetweengroups. BioinformaticAnalysis CellularpathwaydeterminationforeachmiRNAwasdonebyusingDIANA-miRPath(http:// diana.imis.athena-innovation.gr/DianaTools/index.php?r=site/index)whichpredicted targetgenes. AnimalsMaintenance,PreclinicalCharacterizationandmiRNAProfiling MaleC57BL/6Jmice,10–12weeksold(bodyweight,18–22g)wereobtainedfromSriVenka- teswaraEnterprises(Bangalore)andmaintainedat22±1°Cundera12-hlight-darkcycle(lights onfrom6:00AMto6:00PM).Allexperimentswereperformedinaccordancewithregulations specifiedbytheCommitteeforthePurposeofControlandSupervisiononExperimentsonAn- imals(CPCSEA),GovernmentofIndiaandapprovalbytheInstitutionalAnimalEthical PLOSONE|DOI:10.1371/journal.pone.0128372 May28,2015 5/14 CirculatoryMiRNAsofAsianIndianPhenotype Committee(IAEC)oftheMadrasDiabetesResearchFoundationChennai.Food(Nutrilab, Bangalore)andwaterweregivenadlibitumtotheanimals.Highfatdiet(57%)wasprocured fromtheNationalInstituteofNutrition,Hyderabad,India.Afteracclimatization,basalfasting plasma/serumbiochemicalmarkerswereestimated.Micewererandomlydividedintotwo groupsi.e.,micefedwithnormalpelletdiet(NPD;n=5)andmicefedwithhighfatdiet(HFD; n=6),withsimilaraveragebodyweightandplasmaglucoseonday0.Allgroupswerefedwith respectivedietandwateradlibitumfor6months.Allbiochemicalandpreclinicalmeasure- mentsweredoneasperthestandardprocedures(S2Table).Attheendoftheprotocol,HFD fedanimalswerecharacterizedasglucose-intolerantandinsulin-resistantbyOGTT(OralGlu- coseToleranceTest)andITT(InsulinToleranceTest),respectively.Serumsamples(0.25ml) wereusedtoquantifycandidatemiRNAbyqRT-PCRaspreviouslydescribed. Results Inthisstudy,wehaveanalysedtheexpressionofcirculatingmicroRNAsinserumof145non obesesubjects(50%malesand50%females)sufferingfromT2DM(n=49)comparedwith controls(n=49)orpre-diabeticsubjects(n=47).AsshownonTable1,bothprediabetics (IGT)andpatientswithtype2diabetes(T2DM)hadpoorglycemiccontrolasreflectedbysig- nificanthigherfastingplasmaglucoseand2hrplasmaglucoseandhigherHbA valuescom- 1c paredwithcontrols.Subjectswithprediabetesandpatientswithtype2diabetesexhibited significantlyhigherinsulinresistanceasrevealedbyHOMA-IRestimation.Thelipidlevelsand bloodpressuredidnotsignificantlydifferamongthe3groups. ProfilingofCirculatingmiRNA SerummiRNAprofilingwasfirstperformedonasubgroupof36subjectsfromthe145subjects (50%malesand50%females).T2DMandIGTpatientsofthisidentificationsampleonlydif- feredclinicallyfromcontrolsubjectsinfastingglucoseand2hrplasmaglucoseandinfasting insulin.BothT2DMandIGTpatientshadsignificantlyhigherwaistcircumferencethanthe controlgroupofsubjects(Table1).TheglobalserummiRNAprofiling(miRCURYLNA microRNAHumanpanelIV3-Exiqon)detectedanaverageof~159miRNAspersampleof which112miRNAsweredetectedinallgroups.AsshownonS1Fig,themajorityofthe112 miRNAswereexpressedinasimilarwayandhierarchicalclusteringofthedatadidnotpermit todiscriminatebetweenthe3groups.Statisticaldataanalysisrevealedthatthemeanexpres- sionlevelof9miRNAs(i.e.;miR-128,miR-99b-5p,miR-130b-3p,miR-142-3p,miR-374a-5p, miR-423-5p,miR-484,miR-629-5p,let-7d-3p)wassignificantlydifferent(studentt-test p<0.05)acrossthestudiedgroups(Table2).TwomiRNAs(i.e.;miR-128andmiR-99b-5p) hadincreasedcirculatingconcentrationsbothinprediabeticsubjectsandpatientswithtype2 diabetescomparedtocontrolsubjects(Table2). Allthe9miRNAsdifferentiallyexpressedamongthediscoverygroupswereanalysedby quantitativeRT-PCRusingthewholecohortof145individuals(Fig1)containing50%men and50%women.AsshownonFig1,alteredexpressionsofmiR-128andmiR-423-5pinpre- diabeticpatientscomparedtocontrolswereconfirmedwithhighsignificance(p<0.05).Inad- dition,miR-374a-5pandmiR-130b-3pwerealteredinthesamedirectioninthewholegroup ofpatientssufferingfromT2DMthanindiscoverygroup,comparedtothecontrolgroupof subjects,andmiR-130b-3pwasonlyaffectedintheT2DMgroup.AlteredexpressionsofmiR- 629a-5p,let-7d-3p,miR-142-3pandmiR-484werenotconfirmedinthewholecohort (p<0.05). ThenmiRNAexpressionswereanalysedinwomenandmenindependently(Fig2).Altered expressionsofmiR-128inpre-diabeticstatevscontrolwasfoundsignificantinthegroupof PLOSONE|DOI:10.1371/journal.pone.0128372 May28,2015 6/14 CirculatoryMiRNAsofAsianIndianPhenotype Table2. ListofdifferentiallyexpressedcirculatingmiRNAsintheidentificationgroup. Groups miRNAs Foldchanges pvalue NGTvsIGT hsa-miR-128 1.42 0.041 hsa-miR-99b-5p 1.41 0.028 IGTvsT2DM hsa-miR-130b-3p 1.41 0.035 hsa-miR-142-3p 0.57 0.0041 hsa-miR-374a-5p 0.58 0.039 hsa-miR-423-5p 1.54 0.011 hsa-miR-484 1.37 0.004 hsa-miR-629-5p 1.84 0.030 NGTvsT2DM hsa-let-7d-3p 1.26 0.029 hsa-miR-128 1.50 0.047 hsa-miR-130b-3p 1.46 0.049 hsa-miR-142-3p 0.65 0.039 Ctvalueswereusedtoevaluatedifferencesamongtheidentificationgroupandtoselectcandidate miRNAs doi:10.1371/journal.pone.0128372.t002 womenonly.Onthecontrary,alteredexpressionofmiR-423-5pwassignificantlydecreased inthegroupofmen(NGTvsIGT)only.TheincreaselevelofmiR-374a-5pinthediabetic Fig1.CirculatingmiRNAconcentrationofthe9selectedmiRNAs,inthewholecohortof145 individuals.**=pvalues<0.05.Dataareexpressedasarbitraryunits(AU). doi:10.1371/journal.pone.0128372.g001 PLOSONE|DOI:10.1371/journal.pone.0128372 May28,2015 7/14 CirculatoryMiRNAsofAsianIndianPhenotype Fig2.CirculatingmiRNAconcentrationsofthe9selectedmiRNAs,inthewholecohortof145 individuals,butconsideringmen(M)andwomen(F)independently.**=pvalues<0.05.Dataare expressedasarbitraryunits(AU).Black,controlsubjects;Grey,pre-diabeticsubjects;White, diabeticpatients. doi:10.1371/journal.pone.0128372.g002 subjectsvscontrolswasfoundsignificantinthegroupofwomenonly.Finally,miR-142-3p whichwasnotdifferentiallyexpressedconsideringthewholecohort(Fig1),wasfounddiffer- entlyexpressedwhenthegroupofwomenwasanalysedindependently(NGTvsIGT,Fig2). CorrelationofmiRNAswithMetabolicParameters CorrelationsbetweenthelevelsofthevalidatedmiRNAswiththemetabolicparameterswere calculatedinthewholecohortofsubjects(Table3).Theywerepositivelyornegativelycorrelat- edtocholesterollevels,HbA ,insulinresistanceandhyperinsulinemia.MiR-128,whichwas 1C increasedinprediabeticpatients(Fig1),waspositivelycorrelatedwithserumcholesterol.On thecontrary,miR-423-5pdecreasedintheserumofprediabeticpatients(Fig1)wasnegatively correlatedtoHDL-cholesterol. miRNATargetGenesAnalysis Targetgenesofthe4validatedmiRNAswerepredictedandsignificantcellularpathwaysaffect- edbythesegeneswereretrievedforeachmiRNA.SignificantKEGGpathwaysareshownon Fig3.Therearerelatedtocellcycle,signalingpathways,lipidmetabolism,glycanbiosynthesis, brainfunctionsandimmunesystem. CirculatorymiRNAlevelsinHFDmicevsNPDfedmice Tofurthercorroboratetheassociationsassessedinhuman,thepresentstudyalsoanalyzedby meansofqRT-PCRcirculatingmiRNAconcentrationsintheplasmaofdiet-induceddiabetic mice(HFD).Inthecontextofthewell-knowninsulin-resistanceeffects,high-fatdietledto PLOSONE|DOI:10.1371/journal.pone.0128372 May28,2015 8/14 CirculatoryMiRNAsofAsianIndianPhenotype Table3. CorrelationsbetweencirculatingmiRNAconcentrationsandmetabolicparametersinthestudysubjects. miRNAs Metabolicparameters Estimate StdErr tValue Probt Correlationsnotadjustedforsex miR-128 Serum-cholesterol 0.003482 0.001333 2.61 0.0099 miR-423-5p HDL-cholesterol -0.3067 0.1363 -2.25 0.0260 miR-130b-3p HbA1C 0.1902 0.04745 4.01 <.0001 miR-374a-5p HbA1C 0.1745 0.07149 2.44 0.0159 miR-374a-5p HOMA-IR 0.1093 0.04364 2.50 0.0134 miR-374a-5p Fasting-Insulin 0.04148 0.01844 2.25 0.0260 miR-128 Serum-cholesterol 0.003594 0.001323 2.72 0.0074 miR-423-5p HDL-cholesterol -0.3052 0.1266 -2.41 0.0172 Correlationsadjustedforsex miR-130b-3p HbA1C 0.1903 0.04728 4.03 <.0001 miR-374a-5p HbA1C 0.1763 0.07174 2.46 0.0152 miR-374a-5p HOMA-IR 0.1082 0.04383 2.47 0.0147 miR-374a-5p Fasting-Insulin 0.04138 0.01852 2.23 0.0270 Tocalculatecorrelations,Ctvalueswereconvertedintocopynumbers(10^((Ct-25)/-3,3)*100)totakeintoaccountthelogarithmicscaleofthedata. doi:10.1371/journal.pone.0128372.t003 Fig3.PredictedcellularpathwaysofthetargetgenesofthesignificantlyalteredcirculatingmiRNAs inserumofpre-and/ordiabeticpatientsvscontrols.OnlyvalidatedmiRNAsinthewholecohort wereconsidered. doi:10.1371/journal.pone.0128372.g003 PLOSONE|DOI:10.1371/journal.pone.0128372 May28,2015 9/14 CirculatoryMiRNAsofAsianIndianPhenotype Fig4.CirculatingmiRNAconcentrationsofthe9selectedmiRNAsidentifiedinthediscoverygroup quantifiedintheserumofmicefedwithnormalpelletdiet(NPD;n=5)orwithhighfatdiet(HFD; n=6).Dataareexpressedasarbitraryunits(AU).White,NPDmice;Black,HFD. doi:10.1371/journal.pone.0128372.g004 increasedcirculatingconcentrationsofmiR-128,miR-130b-3p,miR-99b-5p,miR-629a-5pand miR-let-7d-3pexpressioninHFDmicecomparedwithNPDfedmice(p<0.05).Conversely, miR-142-3pwassignificantlylowerinHFDanimalscomparedtocontrolmice(Fig4)Correla- tionsbetweentheexpressionsofalteredmiRs(HFDvsNPD)withthemetabolicparameters werecalculatedandareshowninTable4.Asinhumanpre-diabeticsubjects,miR-128waspos- itivelycorrelatedwithcholesterol. Discussion SouthAsiansarecharacterizedbyauniquemetabolicprofilewithhigherinsulinlevels[3],a greaterdegreeofinsulinresistance[4],greaterabdominaladiposityi.e.,higherwaistcircumfer- encedespitelowerbodymassindex[2]andahigherprevalenceofdiabetes[5].Insulinresis- tancehasbeendemonstratedinAsianIndiansevenduringadolescence[6]and hyperinsulinemiaseemstobepresentinAsianIndiansevenatbirth[7].Itappearsthatsome Table4. CorrelationsbetweencirculatingmiRNAconcentrationsandmetabolicparametersinmice. miRNAs metabolicparameters Estimate StdErr tValue Probt miR-128 cholesterol -0.0292 0.005164 -5.65 0.0005 miR-484 Glucose -0.01151 0.00478 -2.41 0.04 miR-629-5p Glucose 0,04474 0,01286 3,48 0,007 miR-128 Glucose 0,02141 0,00837 2,56 0,031 miR-130b-3p HbA1c -1.2578 0.4202 -2.99 0.019 miR-99b-5p HbA1c -2.4481 0.463 -5.29 0.001 miR-130b-3p HDL -0.03028 0.00876 -3.46 0.011 miR-99b-5p HOMA_IR 0.1122 0.04224 2.66 0.026 miR-let-7d-3p HOMA_IR 0.02725 0.01108 2.46 0.036 miR-99b-5p TGL 0.02065 0.00898 2.3 0.047 miR-let-7d-3p TGL 0.005752 0.00214 2.69 0.025 miR-99b-5p VLDL 0.1033 0.04491 2.3 0.047 miR-let-7d-3p VLDL 0.02876 0.0107 2.69 0.025 doi:10.1371/journal.pone.0128372.t004 PLOSONE|DOI:10.1371/journal.pone.0128372 May28,2015 10/14