Association of Elevated Urinary miR-126, miR-155 and miR...
Transcript of Association of Elevated Urinary miR-126, miR-155 and miR...
1
AssociationofElevatedUrinarymiR-126,miR-155andmiR-29bwithDiabeticKidneyDiseaseCristinaBeltrami,*KateSimpson,*†MarkJesky,‡AlexaWonnacott,*
ChristopherCarrington,*PeterHolmans,§LucyNewbury,*†RobertJenkins,*ThomasAshdown,*ColinDayan,¶SimonSatchell,||PeterCorish,**
PaulCockwell,‡DonaldFraser,*†‡‡andTimothyBowen*†‡‡§§* Wales Kidney Research Unit, Division of Infection and Immunity, School of Medicine,College of Biomedical and Life Sciences, Cardiff University, Heath Park Campus, CardiffCF144XN,UK
† CardiffInstituteofTissueEngineeringandRepair,CardiffUniversity,MuseumPlace,CardiffCF103BG,UK
‡ Departmentof RenalMedicine,QueenElizabethHospital Birmingham,MindelsohnWay,BirminghamB152GW,UK
§ MRC Centre for Neuropsychiatric Genetics and Genomics, Division of PsychologicalMedicine andClinicalNeurosciences, School ofMedicine, College of Biomedical and LifeSciences,CardiffUniversity,MaindyRoad,Cathays,CardiffCF244HQ,UK
¶ DiabetesResearchGroup,DivisionofInfectionandImmunity,SchoolofMedicine,CollegeofBiomedicalandLifeSciences,CardiffUniversity,HeathParkCampus,CardiffCF144XN,UK
||Bristol Renal, Bristol Medical School, University of Bristol, Dorothy Hodgkin Building,WhitsonStreet,BristolBS13NY,UK
**BBIGroup,TheCourtyard,TyGlasAvenue,CardiffCF145DX,UK‡‡TheseauthorscontributedequallytothisworkRunningtitle:UrinarymiRNAsinDiabeticNephropathy§§Towhomcorrespondenceshouldbeaddressed:TimothyBowen,WalesKidneyResearchUnit,SchoolofMedicine,CollegeofBiomedicalandLifeSciences,CardiffUniversity,HeathParkCampus,CardiffCF144XN,UK.Tel.:44-29-2074-8389;Fax:44-29-2074-8470;E-mail:[email protected]
CountsAbstract:215words;Text:15pages;Tables:3;Figures:4
Support This isapublicationof independent research fundedby theNational Institute for
HealthResearch(NIHR)InventionforInnovation(i4i)Programme(GrantReferenceno.II-LA-
0712-20003).ThePrincipalInvestigatorforthegrantisTB.Theviewsexpressedinthispaper
arethoseoftheauthorsandnotnecessarilythoseoftheNHS,theNIHR,ortheDepartment
ofHealth,UK.TheauthorsalsoacknowledgesupportfromKidneyResearchUKProjectGrant
AwardsRP44/2014(TB)and IN4/2013(SS).TheJABBSFoundationfundedcollectionofthe
RenalImpairmentinSecondaryCare(RIISC)cohort(PC).TheWalesKidneyResearchUnitis
fundedbycoresupportfromHealthandCareResearchWales(DJF).
Declarationof interestsCD:NovoNordiskAdvisor,SanofiGenzymeandAstraZenecaservices(2015-2017);SS:BoehringerIngelheimTravelSupportandUCBUKGrantSupport(2013-2015);PCor:LifeSciencesBridgingFundWalesConsultancy(2015-2017);PC:JABBSFoundationGrantSupport (2014-2015); DF and TB: Patent application PCT/GB2017/050195: Kidney diseasediagnostic(2017).
2
ABSTRACT1
Effective diabetic kidney disease (DKD) biomarkers remain elusive, and urinary2
microRNAs (miRNAs) represent a potential source of novel non-invasive disease3
sentinels.Weprofiled754miRNAsinpooledurinesamplesfromDKDpatients(n=20),4
detectingsignificantlyincreasedmiR-126,miR-155andmiR-29bcomparedtocontrols5
(n=20).Theseresultswereconfirmedinan independentcohortof89DKDpatients,6
62 diabetic patients without DKD and 41 controls: miR-126 (2.8-fold increase;7
p<0.0001),miR-155 (1.8-fold; p<0.001) andmiR-29b (4.6-fold; p = 0.024). Combined8
receiveroperatingcharacteristiccurveanalysisresultedinanareaunderthecurveof9
0.8.A relativequantification threshold equivalent to 80% sensitivity for eachmiRNA10
gaveapositivesignalfor48%ofDKDpatientscomparedto3.6%ofdiabeticpatients11
without DKD. Laser capture microdissection of renal biopsies followed by RT-qPCR12
detectedmiR-155 in glomeruli, proximal and distal tubules,whilemiR-126 andmiR-13
29b were most abundant in glomerular extracts. Subsequent experiments showed14
miR-126andmiR-29benrichmentinglomerularendothelialcells(GEnCs)comparedto15
podocytes,proximaltubularepithelialcellsandfibroblasts.SignificantlyincreasedmiR-16
126andmiR-29bweredetectedinGEnCconditionedmediuminresponsetotumour17
necrosis factor-alpha and transforming growth factor-beta 1, respectively. Our data18
revealanalteredurinarymiRNAprofileassociatedwithDKDandlinkthesevariations19
tomiRNAreleasefromGEnCs.20
21
Keywords: microRNAs, urine, biomarker, diabetic kidney disease, chronic kidney22
disease23
24
3
Abbreviations:25
ACR albumin:creatinineratio26
CKD chronickidneydisease27
DKD diabetickidneydisease28
eGFR estimatedglomerularfiltrationrate29
GEnC glomerularendothelialcell30
KDIGO kidneydisease:improvingglobaloutcomes31
NKFKDOQI NationalKidneyFoundationkidneydiseaseoutcomesqualityinitiative32
LCM lasercapturemicrodissection33
MDRD modificationofdietinrenaldisease34
miRNA microRNA35
PTC proximaltubularepithelialcell36
RIISC renalimpairmentinsecondarycarestudy37
38
39
4
Introduction40
Recentestimatessuggestthat1in12oftheglobalpopulationsuffersfromdiabetes41
mellitus,approximately40%ofthoseaffectedwillgoontodevelopdiabetickidney42
disease(DKD).1DKDistheleadingcauseofend-stagerenaldiseaseandpredisposing43
factors include genetic causes, ethnicity, hyperglycaemia, insulin resistance,44
intraglomerularhypertensionandhyperfiltration.2,345
Hyperglycaemia results in numerous deleterious consequences including46
upregulated cytokine synthesis, renin-angiotensin system activation, generation of47
advancedglycationendproductsandreactiveoxygenspecies,andincreasedprotein48
kinase C activity.4,5 Nitric oxide andNF-κB pathway-driven loss of endothelial and49
vascularmodulationhavebeen implicated in insulinresistance,andearlyDKDmay50
beassociatedwith insulinsignalingdefects specific to thepodocyte.6These insults51
result in loss of glomerular filtration rate and ultimately to renal failure from52
mesangial hyperexpansion, nodular glomerulosclerosis and tubulointerstitial53
fibrosis.754
Detection of urinary microalbuminuria currently forms the basis of DKD55
progressionmonitoring, varying from normalmean albuminuria values around 1056
mg/daytoadiagnosisofmicroalbuminuriaat30–300mg/dayandmacroalbuminuria57
above 300 mg/day.8 Prognosis is complicated, since not all microalbuminuric58
patientsprogress toovertnephropathy.Anumberofnovelbiomarkershavebeen59
assessedforutility inDKDbutnonearebeingusedasroutineclinicalmarkers,and60
theymaylackspecificityandsensitivitytopredictindividualDKDpatientoutcomes.61
In light of the above, novel markers that can discriminate aetiology, progression62
and/orresponsetotreatmentremainhighlydesirable.63
5
MicroRNAs (miRNAs)areubiquitously-expressedshortnoncodingRNAsthat64
regulate the expression ofmost protein coding genes in the human genome, and65
detectionofmiR-192,miR-194,miR-215,miR-216,miR-146a,miR-204andmiR-88666
iselevatedinthekidney.9UrinarymiRNAsrepresentahighlypromisingnovelsource67
of non-invasive biomarkers that are stabilised via argonaute 2 protein/exosome68
associationandarerapidlyandpreciselydetectedbyRT-qPCR.1069
Reports have suggested a role for miRNAs in the pathology of DKD,11,1270
includingpreviousworkfromthislaboratoryshowingdecreasedmiR-192inbiopsies71
from late-stage DKD patients with diminished renal function.13 However,72
comparatively little is known about the abundance of urinary miRNAs in DKD73
patients.74
We hypothesised that alterations in urinary miRNA profiles would be75
associatedwithDKD.WeidentifiedcandidateDKDbiomarkersbycomparingmiRNA76
profilesinurinesamplesfromapatientdiscoverycohortwiththosefromunaffected77
controls.Selectedcandidateswerethenmeasured ina larger, independentcohort.78
Subsequently,lasercapturemicrodissectionofrenalbiopsiesandinvitrocellculture79
were used to investigate the sources of our candidate urinary miRNA DKD80
biomarkerswithrespecttonephrondomainandcelltype.81
82
6
MaterialsandMethods83
StudyParticipants84
DKDwasdefinedinaccordancewiththeNationalKidneyFoundationKidneyDisease85
Outcomes Quality Initiative (NKF KDOQI) Clinical Practice Guidelines and Clinical86
Practice Recommendations for Diabetes and Chronic Kidney Disease (CKD).1487
Accordingly, CKD should be attributable to diabetes in the presence of88
macroalbuminuria (in the absence of urinary infection), or in the presence of89
microalbuminuriawithconcomitantdiabeticretinopathy,orintype1diabetesofat90
least10yearsduration.14Theinitialprofilingstudycohortof20DKDpatientsand2091
healthy controls was obtained from the Wales Kidney Research Tissue Bank,92
UniversityHospitalofWales,Cardiff.TheDKDgroupwaspredominantlymale(85%),93
meanage72years(SD+/-8.7).DKDpatientswereCKDstage3-5(pre-dialysis),with94
meaneGFRof29ml/min/1.73m2(SD+/-8.5)andameanurinaryAlbumin:Creatinine95
ratio(ACR)of13.5mg/mmol(SD+/-14.5).Thecontrolgroup(n=20)intheprofiling96
cohortwere50%male,meanage47years (SD+/-11.0)withnomicroalbuminuria97
(ACR<3mg/mmol).ForfurtherdetailsonACRcategoriesseeTable1.98
99
The confirmation cohortwas drawn from two secondary care facilities: theWales100
KidneyResearchTissueBank(asabove)andtheRenalImpairmentInSecondaryCare101
(RIISC) study, University Hospital of Birmingham, UK.15 89 patients with DKD,102
including 3 patients with type 1 diabetes, and 41 healthy controls were recruited103
acrossthetwosites.Anadditionalcontrolgroupof62diabeticswithoutDKDwere104
recruitedfromCardiff, including17patientswithtype1diabetes. Ethicalapproval105
7
wasgrantedbytheWalesKidneyResearchTissueBankGovernanceCommitteeand106
theSouthBirminghamLocalResearchEthicsCommittee,respectively.107
108
Patient demographics and clinical parameters are shown in Table 1. All patients109
wererecruitedfromspecialistnephrologyanddiabetescareservicesatthetwosites110
during the period spanning autumn2010 to autumn2013.DKDpatients from the111
RIISCstudycohortwerepredominantlyadvancednephropathsasperRIISCprotocol112
inclusioncriteria:briefly,patientswithCKDstages4-5(pre-dialysis),orCKDstage3113
and accelerated progression and/or proteinuria as defined by the UK National114
Institute for Health and Care Excellence 2008 CKD guideline for secondary care115
review.Thediabeticpatientcontrolgroupallhadadiagnosisofdiabetesbystandard116
American Diabetes Association criteria,16 but without evidence of DKD (i.e. not117
fulfillingtheKDOQIcriteria).118
119
At initial clinic visit, renal function was recorded using estimated glomerular120
filtration rate (eGFR), calculated using the modification of diet in renal disease121
(MDRD)equation17.Urinesampleswerealiquotedforalbumin:creatinineratio(ACR)122
assessment and for RNA extraction (see below). ACR cut-offs for disease severity123
were defined as per Kidney Disease: Improving Global Outcomes (KDIGO) 2012124
guidelines.18125
126
UrineCollection,RNAIsolationandRT-qPCRanalysis127
UrinesampleswerecollectedandRNAextractionfrom350µlofurine,generationof128
cDNA from equal volumes of RNA extracts and RT-qPCRwere then carried out as129
8
described in detail elsewhere.10 TaqMan assays (Thermo Fisher Scientific, Paisley,130
UK) used in this study were: hsa-miR-29b-3p (ID 000413); hsa-miR-126-3p (ID131
002228); hsa-miR-155-5p (ID 002623); hsa-miR-191-5p (ID 002299). Relative132
quantities were calculated using the 2-ΔΔCt method, and miRNA expression was133
normalizedtohsa-miR-191-5p.10134
135
MiRNAprofilingbyTaqManArrayHumanMicroRNACards136
UrinarymiRNAswerereversetranscribedusingtheMegaplexPrimerPools(Human137
Pools A v2.1 and B v3.0, Thermo Fisher Scientific) with a predefined pool of 381138
reversetranscription(RT)primersforeachMegaplexPrimerPool.Afixedvolumeof139
3μlofRNAsolutionwasusedas input ineachRTreaction,andRTreactionswere140
performedaccordingtothemanufacturer’srecommendations.RTreactionproducts141
wereamplifiedusingMegaplexPreAmpPrimers(PrimersAv2.1andBv3.0,Thermo142
Fisher Scientific), the samples were then diluted to a final volume of 100 μl and143
controlsubjectandDKDpatientproductswerepooledasfollows.144
145
Toexcludethepossibilitythatgender,ageandeGFRstatushadextremeeffectson146
miRNAexpressionprofiles,thefollowingpoolingstrategywasfollowed.Controlpool147
(CP)1:urinesamplesfrom5femalesofaverageage44.8years;CP2:5females,57.6;148
CP3: 5 males, 35.2; CP4, 5 males, 53.2. Patient Pool (PP)1: urine samples from 5149
CKD3patientswithaneGFRbetween43.3and36mL/minper1.73m2;PP2:5stage3150
patients,35-31;PP3:5stage4/5patients,27.3-23;PP4:5stage4/5patients,22-151
12.9.152
153
9
TaqManArrayHumanMicroRNACardsAv2.1andBv3.0 (ThermoFisherScientific)154
wereusedtoquantify754humanmiRNAs.Eacharrayincluded377testmiRNAs,3155
endogenouscontrolsandanegativecontrol.Quantitative(q)PCRwascarriedouton156
an Applied Biosystems 7900HT thermo cycler (Thermo Fisher Scientific) using the157
manufacturer’srecommendedprogram.158
159
LaserCaptureMicrodissection(LCM)fromRenalBiopsySamples160
Glomeruli,proximaltubularanddistaltubularprofilesweremicrodissectedfrom6-161
μmsectionsobtainedfromfiveFFPEarchivedrenalbiopsiesfromunaffectedpeople162
using the Arcturus Pixcell IIe infrared laser enabled LCM system (Thermo Fisher163
Scientific).164
165
CellCulture166
Human conditionally immortalised glomerular endothelial cell (GEnC) and human167
podocyte(ATC)celllineswerepropagatedat33°Casdescribedpreviously.19,20After168
5(GEnC)and14(ATC)days,cellsweretransferredto37°C incubationto inactivate169
the SV40 T antigen and permit differentiation, prior to experimental use. Where170
stated, GEnCs were growth arrested for 24 h and then treated with TNF-α (10171
ng/mL)orTGF-β1(1ng/mL)ateithernormoglycaemic(5mM)orhyperglycaemic(25172
mM)D-glucoseconcentrationsfor24h.Proximaltubularepithelialcell(PTC)lineHK-173
221 and fibroblast22 cultures were maintained as described elsewhere. Cells and174
culturemediumobtainedfromeachwellwereusedforRNAextractionasdescribed175
above.176
177
10
Statisticalanalysis178
MiRNA profiling data were analysed using Thermo Fisher Scientific’s DataAssist179
Software (version 3.01), NormFinder Software (http://moma.dk/normfinder-180
software; last access 21/02/18) and GraphPad Prism 6 (version 6.0d). Pearson181
CorrelationCoefficientswasusedtodetectclustersofsimilarityinmiRNAthreshold182
cyclevaluesbetweeneachpoolgroupinpatients,andbetweeneachpoolgroupin183
controls. To identify a suitable reference gene for the normalization of miRNA184
expression in this study, the NormFinder algorithmwas applied to the expression185
dataobtained from theHumanTaqManmiRNAArrays.Analysis comparingmiRNA186
levelsbetweensubjectswithDKDandcontrolswascarriedoutusingGraphPadPrism187
version 6 version 6.0d. Values for p below 0.05 were considered statistically188
significant. MiRNA profiling data sets can be found in Gene Expression Omnibus189
(https://www.ncbi.nlm.nih.gov/geo;accessionnumberGSE114477). 190
191
11
Results192
AlteredurinarymiRNAdetectioninDKDpatients193
ToselectcandidatemiRNAsthatmayactasDKDbiomarkers,wefirstcompareddata194
from unbiased expression profiling of 754miRNAs in urine samples from 20 DKD195
patients and20unaffected controls. Analyseswereperformedon4 patient and4196
controlpools,eachcomposedofurinesamplesfrom5individualsasrecommended197
byZhangandcolleagues.23Sampleswerepooledprior toprofiling tominimize the198
contributionofsubjecttosubjectvariationandtomakesubstantivefeatureseasier199
to find, and thereby identify biomarkers common across individuals.24 Previous200
analysis suggested that40 individualsmightoptimallybepooledacross8arrays,23201
whichwasourchosenpoolingapproach.202
In Figure 1A the 12 data points in the upper right quadrant of the plot203
representthosemiRNAsforwhichstatisticallysignificantfold-changeincreaseswere204
detected inpatienturinecomparedtocontrol samples, the35points in theupper205
left quadrant the corresponding downregulatedmiRNAs. The fold-change data for206
these47miRNAsaresummarisedinFigure1B,andthe8miRNAsexhibiting>5-fold207
change were subsequently selected as potential candidate biomarkers for further208
analysis.209
Specific RT-qPCR assays were then used to analyse these miRNAs in each210
component urine sample pooled for profiling analysis. Statistically significant211
differences in miRNA detection between DKD patient and control urine samples212
werereplicatedformiR-126(4.3-fold increase;p=0.0087),miR-155(22.9-fold;p=213
0.0024)andmiR-29b(4.9-fold;p=0.0002)(Figure1,C-E).214
12
ElevatedurinarymiR-126,miR-155andmiR-29bdetectioninanindependentDKD215
patientcohort216
To test the above findings, miR-126, miR-155 and miR-29b were quantified in217
samples from an independent cohort of patients with established DKD from the218
Renal Impairment in Secondary Care Study (RIISC).15 Samples from 89 patients219
meeting the criteria established in the UK National Institute for Health and Care220
Excellence 2008 criteria were available. An additional cohort of 62 patients with221
diabetesmellitusbutwithoutproteinuriaorotherevidenceofDKDwereincluded,as222
were samples from41 peoplewithout evidence of diabetes orDKD (Table 1).We223
includeddiabetespatientswithoutDKDasa thirdgroup in thisanalysis to identify224
DKD-specificmiRNAdetectionchangesandnotpurelyhyperglycaemia-driveneffects225
fromourprofilingcomparisonofDKDpatientswithcontrolindividuals.226
Significant differenceswere again seenbetweenDKDpatients and controls227
for miR-126 (2.8-fold increase; p<0.0001; Figure 2A), miR-155 (1.8-fold; p<0.001;228
Figure2B)andmiR-29b(4.6-fold;p=0.024;Figure2C).ComparisonofDKDpatients229
withdiabeticpatientswithoutDKDwasstatisticallysignificantformiR-126(3.1-fold230
increase;p<0.0001)andmiR-155(1.6-fold;p=0.024)withatrendtoincreasedmiR-231
29b(4.1-fold;p=0.121)(Figure2A-C).232
RT-qPCR data for all 3 miRNAs were used to compare DKD patients and233
diabetic patients without DKD in the combined receiver operating characteristic234
(ROC) curve analysis shown in Figure 2D, giving an area under the curve (AUC) of235
0.80.ToanalysethecontributionsofeachmiRNAtotheaboveROCcurve,individual236
specificity and likelihood ratioswere calculated for relative expression (RQ) values237
13
equivalent to a sensitivity of 80%.25,26 Data displayed in Table 2 illustrate the238
magnitude of corresponding specificity values wasmiR-126 >miR-155 >miR-29b,239
and that combined miRNA data resulted in a ≥6.5% increase in specificity and240
likelihoodratiocomparedwithindividualmiRNAs.TheseRQdatawerethenusedas241
consecutive threshold values to discriminate between DKD and diabetic patients242
withoutDKD(D inTable3) fromthe independentcohort.Thediscriminatoryorder243
was miR-29b (DKD/D = 5.62) > miR-126 (3.48) > miR-155 (2.23), and RQ values244
exceedingall3thresholdswereobtainedfor48.0%ofDKDpatientscomparedwith245
3.6%ofdiabeticpatientswithoutDKD(Table3).246
247
Lasercapturemicrodissectionshows increasedglomerularabundanceofmiR-126248
andmiR-29bthatisreplicatedinGEnCculture249
Previous reports have linked changes inmiRNA expression to DKD pathology, but250
have focused on whole tissue studies. For example, we showed association of251
decreasedmiR-192 expressionwith disease progression in DKD biopsies by in situ252
hybridisation.13253
Inthepresentstudyweused lasercapturemicrodissection(LCM)to isolate254
glomeruli, proximal and distal tubules (Figure 3A) from histologically normal255
formalin-fixed, paraffin-embedded (FFPE) renal biopsy samples, and analysedmiR-256
126, miR-155 and miR-29b expression by RT-qPCR. In Figure 3B, a typical CD10-257
stained FFPE biopsy section is seen before and after LCM to isolate glomeruli,258
proximalanddistalrenaltubules.MiR-126,miR-155andmiR-29bweredetected in259
extracts from all three nephron regions (Figure 3, C-E). Increased glomerular260
14
abundanceswereobserved formiR-126 (Figure3C)andmiR-29b (Figure3E),while261
miR-155wasmostabundantinthedistaltubule(Figure3D).262
Conclusions regarding nephron region-specific miRNA expression from the263
aboveanalysesare inherently limited,however,sincetissueextractsaresubjectto264
tracecontaminationbycellsfromothernephrondomains.Therefore,cellularmiRNA265
localisationwithineachnephron regionwas subsequently investigatedbyRT-qPCR266
analysisofpodocyteandendothelialcell(GEnC)culturesfromtheglomerulus,renal267
proximal tubular epithelial cells (PTC) and fibroblasts. Detection of miR-126 was268
significantlyhigher inGEnCscomparedwithothercell types(Figure3F).MostmiR-269
155wasdetected inPTCsand least inGEnCs (Figure3G),whilemiR-29bwasmost270
abundantinGEnCs(Figure3H).271
272
GEnC release ofmiR-126 andmiR-29b in an in vitromodel of hyperglycaemia is273
drivenbyTNF-αandTGF-β1,respectively274
The above data localized themajority of miR-126 andmiR-29b expression to the275
GEnC.We next sought stimuli by whichmiRNAs are released into the glomerular276
ultrafiltrate, and hence the urine. Data from animal models of diabetes show277
increasedglomerularandPTCTNF-αexpression,andrenoprotectiveeffectsofTGF-β278
inhibitors have also been reported.27,28 GEnC expression of our candidatemiRNAs279
wasthusanalysedinvitroinresponsetoTNF-αandTGF-β1innormoglycaemiaand280
hyperglycaemia(Figure4,A-F).281
The presence of TNF-α led to significantly increased miR-126 detection in282
GEnCconditionedmediumat5mMand25mMD-glucose(Figure4B),apatternalso283
seen for miR-29b following TGF-β1 addition (Figure 4D). These cytokines did not284
15
increaseGEnCexpressionofmiR-126(Figure4A),ormiR-29b(Figure4C),apattern285
consistentwithincreasedrelease,butnotexpression,ofmiRNAs.286
NosignificantchangesinmiR-155weredetectedinresponsetoelevatedD-287
glucose with either cytokine, and data for TNF-α are shown (Figure 4, E and F).288
Similarly, changes in miR-126 following TGF-β1 addition, and for miR-29b in the289
presenceofTNF-α,werenotobserved (datanotshown).ElevatedD-glucosealone290
didnotchangemiRNAexpressioninGEnCsorconditionedmedium(Figure4,A-F).291
292
16
Discussion293
Diabetickidneydisease (DKD) is the leadingcauseofkidney failure requiring renal294
replacement therapy worldwide, but effective methods to identify and halt295
progression of disease-specific pathophysiological changes remain elusive. Current296
effective interventions such as control of blood glucose and blood pressure are297
challenging to achieve, costly and time intensive. Existing tests track DKD from298
diabetic diagnosis to kidney failure, but do not allow accurate prognosis for the299
individual patient. In addition, the absence of treatment response biomarkers300
hindersdevelopmentofemergingDKDtherapies.There is thusanunmetneed for301
additionalDKDbiomarkerstotargetinterventionandfollowresponsetotherapy.302
InthisstudywesetouttoidentifyurinarymiRNADKDbiomarkers.Increased303
detection of miR-126, miR-155 and miR-29b was observed in the urine of DKD304
patients in comparison with both unaffected individuals and diabetic patients305
without DKD. MiRNA localization and release studies further suggested specific306
releaseofmiR-126andmiR-29bfromGEnCs.Thisraisedthepossibilitythaturinary307
miRNAquantificationmightprovidedataonongoingpathologicalprocesses,andso308
aidpatientstratificationandmeasurementofresponsetotherapy.309
UrinarymiRNAbiomarkershaveseveralpotentialsignificantadvantagesover310
circulatingmiRNAsforadoptionintoexistingtreatmentpathwaysalongsidecurrent311
biomarkers,includingspeedandcostofnon-invasivesampleaccess.29However,few312
urinarymiRNADKDbiomarkerdatahavesofarbeenreported.Previousstudieshave313
focusedoncirculatingmiRNAs,andhavegeneratedconflictingdatawithrespectto314
associationofmiR-126withdiabetesmellitusand/orDKD.A recentcross-sectional315
analysis of type 2 diabetes mellitus patients found a negative association with316
17
plasma miR-126,30 and similar findings have been reported for type 1 diabetes317
mellitus and all complications.31 By contrast,miR-126 detection did not change in318
whole blood from type 2 diabetes mellitus patients and control subjects, but319
decreasedinDKDpatientsamples.32Furthermore,nochangeinplasmamiR-126was320
observedinastudyofpaediatrictype1diabeticpatients.33Theseanalysesprovide321
inconsistentdataforthebiomarkerutilityofcirculatingmiR-126, incontrasttothe322
significantandreproducibleincreaseswedetectedinmiR-126,miR-155andmiR-29b323
inDKDpatienturineinthepresentstudy.324
TheDKD-specificalterationsinurinarymiRNAprofilesdetectedinthisstudy325
mayhave functional significance.Our in vitro analyses localizedmiR-126 andmiR-326
29bprincipally to theGEnC,withmiR-155expressiondistributedevenlyacross the327
nephron.GlomerularendotheliallocalizationofmiR-126mayreflecttheroleofthis328
transcript in vascular regulation. Targeted mouse miR-126 deletion resulted in329
vascular abnormalities by removing inhibition of sprouty-related EVH1 domain-330
containing protein 1 expression, thereby enhancing vascular endothelial growth331
factor (VEGF) function.34 A role in DKD pathology for VEGFA signalling between332
GEnCs and podocytes has been proposed.35 In addition, miR-126 repression of333
vascular cell adhesion molecule 1 expression in human umbilical vein endothelial334
cells regulates their response to pro-inflammatory adhesionmolecules.36MiR-126335
has also been implicated in the heterogenic inflammatory response of renal336
microvascularendothelialcells.37337
Increased expression of miR-155 has been observed in DKD patient renal338
biopsies, in close correlationwith increased serumcreatinine.38 Furthermore,miR-339
155deficiencyattenuatedrenaldamageandIL-17expressionwasdownregulatedin340
18
streptozotocin-induced DKD mice.39 Together with miR-126, miR-155 has been341
implicated in multiple forms of vascular remodelling and associated with342
cardiovasculardisease.40343
DecreasedmiR-29bhasbeenreportedinearlyandadvancedanimalmodels344
ofdiabeticrenalfibrosis.41ChenandcolleaguesfoundthatlossofrenalmiR-29bin345
db/db mice led to increased albuminuria, TGF-β-mediated fibrosis and immune346
injury,whilerestoredmiR-29bexpressioninhibitedrenalinjury.42Indeed,whilewe347
have focused on upregulatedmiRNAs in this study,we acknowledge the potential348
importanceofmiRNAdownregulationsthatwedetected.349
In the present studywe localizedmiR-29b to the glomerular endothelium.350
ReductionofcollagenandlamininsynthesishasbeenreportedfollowingforcedmiR-351
29bexpressioninhumancornealendothelialcells.43InapoEknockoutmice,miR-29b352
inducedaorticendothelialpermeability in response toahigh fatdiet,andbrought353
aboutaorticapoptosisbydirecttargetingofmelatoninreceptormt1.44 Inaddition,354
upregulated miR-29b expression has been observed in human umbilical vein355
endothelialcellsexposedtohyperglycaemia.45356
The cytokine-driven release fromGEnCs observed formiR-126 (TNF-α) and357
miR-29b (TGF-β1) reported here suggests that these cells may be the principal358
sourceofelevatedurinarymiR-126andmiR-29bdetectedinDKD.Wespeculatethat359
thisconstitutesevidencefordisease-relatedsignallingdownthenephronthatwillbe360
interesting to test in future studies. Indeed,we have demonstrated association of361
urinary miRNAs with exosomes10 and exosomal transport, which might facilitate362
passageofmiRNAsthroughthenephron,hasbeenreportedforallthreecandidate363
biomarkermiRNAs.364
19
Exosome-mediated release of miR-126 from CD34+ peripheral blood365
mononuclear cells is proangiogenic, and decreased miR-126 was detected in366
elevatedglucosecellcultureanddiabeticpatients.46MiR-155isdepletedinurinary367
exosomes frommicroalbuminuric type 1 diabetesmellitus patients.47 Endogenous368
miR-29b, spontaneously released frombeta-cellswithinexosomes, stimulatesTNF-369
αsecretionfromspleencellsisolatedfromdiabetes-proneNODmiceinvitro.48370
Insummary,wehaveusedunbiasedprofilingapproachestoidentifyaurinary371
miRNAsignatureassociatedwithDKD,andhavesubsequentlyconfirmed increased372
miR-126,miR-155andmiR-29binanindependentpatientcohort.MiR-126andmiR-373
29bwereidentifiedasenrichedinGEnCs,andreleasedfromthesecellsinresponse374
to DKD-related cytokines. Urinary miR-126, miR-155 and miR-29b are therefore375
promising DKD biomarkers, and the potential pathological significance ofmiR-126376
andmiR-29breleasefromGEnCsmeritsfurtherevaluation.377
378
20
Acknowledgments379
Wethankcontrolsubjectsandpatientsforthedonationofurinesamplesincluding380
those samples kindly provided by co-authorsDrMarkD. Jesky and Professor Paul381
Cockwell(QueenElizabethHospitalBirmingham,Birmingham,UK).382
C.B.andK.S.performed theexperimentalwork,generatedandanalyzeddata,and383
helped draft the manuscript. A.W., C.C., L.N., R.J. and T.A. performed the384
experimentalwork,generatedandanalyzeddata.M.J.,P.H.,C.D.,S.S.,P.CorandP.C.385
discussedelementsofexperimentaldesignand/orcohortcomposition.D.F.andT.B.386
designed the research. T.B. wrote themanuscript, which was edited by D.F. then387
amendedandapprovedbyeachauthor.388
389
21
References390
1. JhaV,Garcia-GarciaG,IsekiK,LiZ,NaickerS,PlattnerB,SaranR,WangAY,391
YangCW:Chronickidneydisease:globaldimensionandperspectives.Lancet392
2013,382:260–272393
2. CowieCC,PortFK,WolfeRA,SavagePJ,MollPP,HawthorneVM:Disparities394
inincidenceofdiabeticend-stagerenaldiseaseaccordingtoraceandtypeof395
diabetes.NEnglJMed1989,321:1074–1079396
3. RegeleF,JelencsicsK,ShiffmanD,PareG,McQueenMJ,MannJF,Oberbauer397
R: Genome-wide studies to identify risk factors for kidney disease with a398
focusonpatientswithdiabetes.NephrolDialTransplant2015,30:26–34399
4. Makita Z, Radoff S, Rayfield EJ, Yang Z, Skolnik E, Delaney V, Friedman EA,400
CeramiA,ViassaraH:Advancedglycosylationendproducts inpatientswith401
diabeticnephropathy.NEnglJMed1991,325:836–842402
5. Schrijvers BF, De Vriese AS, Flyvbjerg A: From hyperglycemia to diabetic403
kidneydisease:theroleofmetabolic,hemodynamic,intracellularfactorsand404
growthfactors/cytokines.EndocrRev2004,25:971–1010405
6. CowardRJ, SaleemMA:Podocytesas a targetof insulin.CurrDiabetesRev406
2011,7:22–27407
7. Rodriguez-IturbeB, JohnsonRJ,Herrera-Acosta J: Tubulointerstitial damage408
andprogressionofrenalfailure.KidneyIntSuppl2005,99:S82–S86409
8. ParvingH-H, Persson F, RossingP:Microalbuminuria:Aparameter thathas410
changeddiabetescare.DiabResClinPract2015,107:1-8411
9. LandgrafP,RusuM,SheridanR,SewerA,IovinoN,AravinA,PfefferS,RiceA,412
KamphorstAO,LandthalerM,LinC,SocciND,HermidaL,FulciV,ChiarettiS,413
22
Foà R, Schliwka J, Fuchs U, Novosel A, Müller RU, Schermer B, Bissels U,414
Inman J, Phan Q, Chien M, Weir DB, Choksi R, De Vita G, Frezzetti D,415
Trompeter HI, Hornung V, Teng G, Hartmann G, Palkovits M, Di Lauro R,416
WernetP,MacinoG,RoglerCE,Nagle JW, Ju J,PapavasiliouFN,BenzingT,417
LichterP,TamW,BrownsteinMJ,BosioA,BorkhardtA,Russo JJ,SanderC,418
ZavolanM,TuschlT:AmammalianmicroRNAexpressionatlasbasedonsmall419
microRNAsequencing.Cell2007,129:1404-1414420
10. BeltramiC,ClaytonA,NewburyLJ,CorishC,JenkinsRH,PhillipsAO,FraserDJ,421
Bowen T: Stabilization of urinarymicroRNAs by associationwith exosomes422
andargonaute2protein.Non-CodingRNA2015,1:151-165423
11. Trionfini P, Benigni A, Remuzzi G: MicroRNAs in kidney physiology and424
disease.NatRevNephrol201511:23-33425
12. Simpson K, Wonnacott A, Fraser DJ, Bowen T: MicroRNAs in diabetic426
nephropathy:frombiomarkerstotherapy.CurrDiabRep2016,16:35427
13. KrupaA,JenkinsR,LuoDD,LewisA,PhillipsAO,FraserDJ:LossofMicroRNA-428
192promotesfibrogenesisindiabeticnephropathy.JAmSocNephrol2010,429
21:438-447430
14. KDOQIClinicalPracticeGuideline forDiabetesandCKD:2012Update.AmJ431
KidneyDis2012,60:850-886432
15. StringerS,SharmaP,DuttonM,JeskyM,NgK,KaurO,ChappleI,DietrichT,433
FerroC,CockwellP:Thenaturalhistoryof, and risk factors for,progressive434
chronickidneydisease(CKD):theRenalImpairmentinSecondarycare(RIISC)435
study;rationaleandprotocol.BMCNephrol2013,14:95436
23
16. American Diabetes Association: Diagnosis and classification of diabetes437
mellitus.DiabetesCare2010,33(suppl1):S62-S69438
17. LeveyAS,Bosch JP, Lewis JB,GreeneT,RogersN,RothD:Amoreaccurate439
methodtoestimateglomerular filtrationratefromserumcreatinine:anew440
predictionequation.ModificationofDietinRenalDiseaseStudyGroup.Ann441
InternMed1999,130:461-470442
18. KidneyDisease:ImprovingGlobalOutcomes(KDIGO)CKDWorkGroup(2013)443
KDIGO2012ClinicalPracticeGuideline for theEvaluationandManagement444
ofChronicKidneyDisease.KidneyIntSuppl2013,3:S1-150445
19. SatchellSC,TasmanCH,SinghA,NiL,GeelenJ,vonRhulandCJ,O’HareMJ,446
SaleemMA, vandenHeuvel LP,MathiesonPW:Conditionally immortalized447
human glomerular endothelial cells expressing fenestration in response to448
VEGF.KidneyInt2006,69:1633-1640449
20. SaleemMA,O'HareMJ,ReiserJ,CowardRJ,InwardCD,FarrenT,XingCY,NiL,450
MathiesonPW,MundelP:Aconditionallyimmortalizedhumanpodocytecell451
linedemonstratingnephrinandpodocinexpression.JAmSocNephrol2002,452
13:630-638453
21. Jenkins RH, Davies LC, Taylor PR, Akiyama H, Cumbers B, Beltrami C,454
CarringtonCP,PhillipsAO,BowenT,FraserDJ:MiR-192inducesG2/Mgrowth455
arrestinaristolochicacidnephropathy.AmJPath2014,184:996-1009456
22. Midgley AC, Bowen T, Phillips AO, Steadman R: MicroRNA-7 inhibition457
rescues loss of epidermal growth factor receptor hyaluronan-dependent458
differentiationinfibroblasts.AgingCell2014,13:235-244459
24
23. Zhang SD, Gant TW: Effect of pooling samples on the efficiency of460
comparativestudiesusingmicroarrays.Bioinformatics2005,21:4378-4383461
24. Kendziorski C, Irizarry RA, Chen KS, Haag JD, Gould MN: On the utility of462
poolingbiologicalsamplesinmicroarrayexperiments.ProcNatlAcadSciUSA463
2005,102:4252-4257464
25. Prowle JR, Calzavacca P, Licari E, Ligabo EV, Echeverri JE, Bagshaw SM,465
Haase-Fielitz A, Haase M, Ostland V, Noiri E, Westerman M, Devarajan P,466
Bellomo R: Combination of biomarkers for diagnosis of acute kidney injury467
aftercardiopulmonarybypass.RenFail2015,37:408–416,2015468
26. SchleyG,KöberleC,ManuilovaE,RutzS,ForsterF,WeyandM,FormentiniI,469
Kientsch-EngelR,Kai-EckardtU,WillamC:Comparisonofplasmaandprine470
biomarkerperformanceinacutekidneyinjury.PLoSONE2015,10:e0145042471
27. Navarro J,Milena F,Mora C, Leon C, Claverie F, Flores C, Garcia J: Tumor472
necrosis factor-alphageneexpression indiabeticnephropathy:Relationship473
withurinaryalbuminexcretionandeffectofangiotensin-convertingenzyme474
inhibition.KidneyIntSuppl2005,99:S98–S102475
28. Ohga S, Shikata K, Yozai K, Okada S, Ogawa D, Usui H,Wada J, Shikata Y,476
MakinoH:Thiazolidinedioneamelioratesrenalinjuryinexperimentaldiabetic477
rats through anti-inflammatory effects mediated by inhibition of NF-κB478
activation.AmJPhysiol-RenalPhysiol2007,292:F1141–F1150479
29. Smith DA, Newbury LJ, Drago G, Bowen T, Redman JE: Electrochemical480
detection of urinary microRNAs via sulfonamide-bound antisense481
hybridisation.SensActuatorsBChem2017,253:335–341482
25
30. OlivieriF,SpazzafumoL,BonafèM,RecchioniR,PrattichizzoF,MarcheselliM,483
Micolucci L,MensàE,GiulianiA, SantiniG,GobbiM, LazzariniR,BoemiM,484
TestaR,AntonicelliR,AntonioDomenicoProcopioAD,BonfigliAR:MiR-21-5p485
and miR-126a-3p levels in plasma and circulating angiogenic cells:486
relationshipwith type2diabetes complications.Oncotarget2015,6:35372-487
35382488
31. Barutta F, Bruno G, Matullo G, Chaturvedi N, Grimaldi S, Schalkwijk C,489
StehouwerCD,FullerJH,GrudenG:MicroRNA-126andmicro-/macrovascular490
complicationsoftype1diabetesintheEURODIABprospectivecomplications491
study.ActaDiabetol2017,54:133–139492
32. Al-KafajiG,Al-MahroosG,Al-MuhtareshHA,SkrypnykC,SabryMA,Ramadan493
AR:DecreasedexpressionofcirculatingmicroRNA-126inpatientswithtype2494
diabetic nephropathy: a potential blood-based biomarker. Exp Therapeut495
Med2016,12:815-822496
33. Osipova J, Fischer D-C, Dangwal S, Volkmann I, Widera C, Schwarz K,497
Lorenzen JM, Schreiver C, Jacoby U, Heimhalt M, Thum T, Haffner D:498
Diabetes-AssociatedMicroRNAs in Pediatric PatientsWith Type 1 Diabetes499
Mellitus: A Cross-Sectional Cohort Study. J Clin Endocrinol Metab 2014,500
99:E1661–E1665501
34. Wang S, Aurora AB, Johnson BA, Qi X, McAnally J, Hill JA, Richardson JA,502
Bassel-DubyR,OlsonEN:Theendothelial-specificmicroRNAmiR-126governs503
vascularintegrityandangiogneneiss.DevCell2008,15:261–271504
35. BrosiusFC,CowardRJ:Podocytes,signalingpathways,andvascularfactorsin505
diabetickidneydisease.AdvChronicKidneyDis2014,21:304–310506
26
36. HarrisTA,YamakuchiM,FerlitoM,MendellJT,LowensteinCJ:MicroRNA-126507
regulates endothelial expression of vascular cell adhesionmolecule 1.Proc508
NatlAcadSciUSA2008,105:1516–1521509
37. Ásgeirsdóttir SA, van Solingen C, Kurniati NF, Zwiers PJ, Heeringa P, van510
Meurs M, Satchell SC, Saleem MA, Mathieson PW, Banas B, Kamps JA,511
RabelinkTJ,vanZonneveldAJ,MolemaG:MicroRNA-126contributestorenal512
microvascular heterogeneity of VCAM-1 protein expression in acute513
inflammation.AmJPhysiolRenalPhysiol2012,302:F1630-1639514
38. HuangY,LiuY,LiL,SuB,YangL,FanW,YinQ,ChenL,CuiT,ZhangJ,LuY,515
ChengJ,FuP,LiuF: Involvementof inflammation-relatedmiR-155andmiR-516
146aindiabeticnephropathy:implicationsforglomerularendothelialinjury.517
BMCNephrol2014,15:142518
39. Lin X, You Y, Wang J, Qin Y, Huang P, Yang F: MicroRNA-155 deficiency519
promotes nephrin acetylation and attenuates renal damage in520
hyperglycemia-inducednephropathy.Inflammation2015,38:546-544521
40. WeltenSMJ,GoossensEAC,QuaxPHA,NossentAY:Themultifactorialnature522
ofmicroRNAsinvascularremodeling.CardiovascRes2016,110:6–22523
41. WangB,KomersR,CarewR,WinbanksCE,XuB,Herman-EdelsteinM,KohP,524
Thomas M, Jandeleit-Dahm K, Gregorevic P, Cooper ME, Kantharidis P:525
Suppression of microRNA-29 expression by TGF-β1 promotes collagen526
expressionandrenalfibrosis.JAmSocNephrol23:252-265,2012527
42. Chen H-Y, Zhong X, Huang XR, Meng X-M, You Y, Chung ACK, Lan HY:528
MicroRNA-29b inhibitsdiabeticnephropathy indb/dbMice.MolTher2014,529
22:842-853530
27
43. Toyono T, Usui T, Villarreal G, Kallay L,MatthaeiM, Vianna LMM, Zhu AY,531
Kuroda M, Amano S, Jun AS: MicroRNA-29b overexpression decreases532
extracellular matrix mRNA and protein production in human corneal533
endothelialcells.Cornea2016,35:1466–1470534
44. ZhuHQ, LiQ,DongLY,ZhouQ,WangH,WangY:MicroRNA-29bpromotes535
high-fat diet-stimulated endothelial permeability and apoptosis in apoE536
knock-out mice by down-regulating MT1 expression. Int J Cardiol 2014,537
176:764-770538
45. Silambarasan M, Tan JR, Karolina DS, Armugam A, Kaur C, Jeyaseelan K:539
MicroRNAs inhyperglycemia inducedendothelial cell dysfunction. Int JMol540
Sci2016,17:518541
46. Mocharla P, Briand S, Giannotti G, Dörries C, Jakob P, Paneni F, Lüscher T,542
Landmesser U: AngiomiR-126 expression and secretion from circulating543
CD34+ and CD14+ PBMCs: role for proangiogenic effects and alterations in544
type2diabetics.Blood2013,121:226-236545
47. BaruttaF,TricaricoM,CorbelliA,AnnaratoneL,PinachS,GrimaldiS,Bruno546
G, Cimino D, Taverna D, Deregibus MC, Rastaldi MP, Perin PC, Gruden G:547
Urinary exosomal microRNAs in incipient diabetic nephropathy. PLoS ONE548
2013,8:e73798549
48. SalamaA, FichouN,AllardM,Dubreil L,DeBeaurepaire L,VielA, JégouD,550
Bösch S, Bach J-M: MicroRNA-29b modulates innate and antigen-specific551
immune responses in mouse models of autoimmunity. PLoS ONE 2014,552
9:e106153553
554
28
Figurelegends555
Figure 1UrinarymiRNAdetection inurine samples fromDKDpatientsandcontrol556
subjects.A:Volcanoplotshowingthedetectionprofileofthe377urinarymiRNAsin557
TLDACardAinDKDpatients(n=20;fourpoolsoffivepatients)andcontrols(n=20;558
fourpoolsoffivecontrols).Thedottedhorizontallinerepresentsapvalueboundary559
of 0.05. B: Fold change of miRNA detection between DKD patients and controls.560
DataAssist Software (Thermo Fisher Scientific) was used to perform relative561
quantificationforsamplecomparison,toperformt-testsamplegroupcomparisons,562
andtoproducethegraphicoutputshown. (C-E)RT-qPCRanalysisshowssignificant563
differences in detection of (C) miR-126, (D) miR-155 and (E) miR-29b between564
patients and control urine in the component urine samples pooled for profiling565
analyses(AandB).DKDpatientsversuscontrolsforC:miR-126,D:miR-155,E:miR-566
29b(n=20foreachgroup).Analysiswascarriedoutbyunpaired2-tailedt-testwith567
Welch’scorrection.ProfilingdataanalysisusingtheNormFinderalgorithmidentified568
miR-191asoptimalfornormalisationofourRT-qPCRdata.Datawerenormalizedto569
endogenouscontrolmiR-191andarepresentedasmean+/-SEM.**P<0.01,***P<570
0.005and*****P<0.0005.571
572
Figure 2 RT-qPCR detection of selected miRNAs in patients and control subjects.573
Relativeexpressionwassignificantlydifferentin89DKDpatientscomparedwith62574
diabeticpatientswithoutDKDand41controlsfor(A)miR-126and(B)miR-155,and575
significantly different in DKD patients compared with controls for (C) miR-29b. A:576
DKDpatientsversusdiabeticpatientswithoutDKDandcontrols;B:DKDpatientsand577
diabeticpatientswithoutDKD;DKDpatientsversuscontrols;C:DKDpatientsversus578
29
controls (n = 89 DKD patients, 62 diabetic patients without DKD, 41 controls).579
Analysis was carried out by unpaired 1-tailed t-test withWelch’s correction. Data580
were normalized to endogenous control miR-191 and are presented as mean +/-581
SEM.(D)CombinedreceiveroperatingcharacteristiccurveanalysisformiR-126,miR-582
155andmiR-29b,areaunderthecurve(AUC)=0.80.Dataweregeneratedusingthe583
pROCpackageinR-3.2.3.*P<0.05,****P<0.001and******P<0.0001.584
585
Figure 3 Localization ofmiRNA expression by laser capturemicrodissection (LCM)586
andcellculture.A:Keyfunctionalnephrondomainsincludetheglomerulus(G),the587
proximaltubule(PT)andthedistaltubule(DT).B:ACD10-stainedFFPErenalbiopsy588
samplebeforeandafterexcisionofglomerulibyLCM.Bars=100µm.(C-E)Relative589
expressionofmiR-126,miR-155andmiR-29b, respectively, inLCM-isolatedGs,PTs590
andDTsfrom5renalbiopsiesofhealthyindividuals.C:GversusPT;GversusDT(n=591
5biopsies).(F-H)RelativeexpressionofmiR-126,miR-155andmiR-29b,respectively,592
in in vitro cultured HK-2 renal proximal tubular epithelial cells (PTCs), fibroblasts,593
podocytes and conditionally immortalized glomerular endothelial cells (GEnCs). F:594
GEnCsversusfibroblasts;GEnCsversusPTCsandpodocytes;G:PTCversusGEnCs;H:595
GEnCsversuspodocytes(n=4).Analysiswascarriedoutbyone-wayANOVAanalysis596
withTukey’smultiplecomparisontest.Datawerenormalizedtoendogenouscontrol597
miR-191 and are presented asmean +/- SEM. *P < 0.05, **P < 0.01 and ****P <598
0.001.599
600
601
30
Figure4MiRNAexpressioninGenCsandGenCconditionedmediuminresponseto602
hyperglycaemia and DKD-related cytokines. Following 24 h culture in 5mMor 25603
mM D-glucose, relative expression in GEnCs and GEnC conditioned medium,604
respectively,of(AandB)miR-126inresponseto10ng/mlTNF-α,(CandD)miR-29b605
inresponseto1ng/mlTGF-β1and(EandF)miR-155inresponseto10ng/mlTNF-α,606
andinuntreatedcells.B:5mMD-glucoseversus5mMD-glucoseplusTNF-α,and25607
mMD-glucoseversus25mMD-glucoseplusTNF- α;C:5mMD-glucoseversus5mM608
D-glucoseplusTGF-β1,and25mMD-glucoseversus25mMD-glucoseplusTGF-β1;609
D:5mMD-glucoseversus5mMD-glucoseplusTGF-β1;25mMD-glucoseversus25610
mM D-glucose plus TGF-β1 (n = 4). Analysis was carried out by one-way ANOVA611
analysis with Tukey’s multiple comparison test. Data were normalized to612
endogenouscontrolmiR-191andarepresentedasmean+/-SEM.*P<0.05and**P613
<0.01.614
615
616
617
618
619
620
621
622
623
624
625
31
Table 1Demographicandclinicalparametersofpatients recruited from2centres:626
Wales Kidney Research Tissue Bank, Cardiff (University Hospital Wales) and627
Birmingham (UniversityHospital Birmingham, Renal Impairment in Secondary care628
(RIISC)studycohort)(n=151)629
630
Feature
Patients(n=151)
Diabetic(n=62)
DKD(n=89)
Controls(n=41)
Malen(%) 37(58) 55(62) 18(44)Non-Caucasiann(%) 13(21) 33(37) MeanAge(years,SD) 52+/-16.1 62+/-13.6 55+/-15.4 eGFRml/min/1.73m2 Mean(SD) 78+/-16.3 30+/-20.9 Median(IQR) 84(72-90) 22(17-38) CKDstagen(%) NoCKD/CKDG1(eGFR≥90) 23(37) 2(2) CKDG2(eGFR=60-89) 32(52) 10(11) CKDG3(eGFR=30-59) 5(8) 17(19) CKDG4(eGFR=15-29) 2(3) 45(51) CKDG5(eGFR<15) 0 15(17) ACR*mg/mmoln(%) A1-Normal-highnormal(ACR<3) 54(87.1) 15(16.9) A2-Moderatelyincreased(ACR3-30) 8(12.9) 25(28.1) A3-Severelyincreased(ACR>30) 0(0) 49(55.0) *Albumin:Creatinine ratio (ACR) group cut-offs and nomenclature derived from631
KDIGO2012recommendations.632
633
634
635
32
Table2Specificityvalues,likelihoodratiosandRQthresholdsformiR-126,miR-155,636
miR-29bandallthreemiRNAsabovean80%ROCcurvesensitivitythreshold637
638
Sensitivity(%)
Specificity(%)
LikelihoodRatio
RQThreshold
3miRNAs 80.21 63.64 2.206 >1.148
miR-126 80.41 57.14 1.876 >0.6762
miR-155 80.61 52.00 1.679 >0.9110
miR-29b 80.61 40.00 1.344 >0.8058639
Table3DKDanddiabeticpatientswithoutDKD(D)patientnumbersandpercentages640
abovean80%ROCcurvesensitivitythresholdformiR-126,miR-155,miR-29bandall641
threemiRNAs642
643 Patientsabove80%sensitivitythreshold
miR-126miR-155miR-29b3miRNAsTotal
D 23 30 13 2 55
DKD 80 67 73 47 98
Percentageofpatientsabove80%sensitivitythresholdmiR-126miR-155miR-29b3miRNAs
D 41.8 54.6 23.6 3.6
DKD 81.6 68.4 74.5 48.0
644
Figure1
A
F o ld C h a n g e
P-v
alu
e
- 1 0 -8 -6 -4 -2 0 2 4 6 8
m iR -1 5 5
m iR -1 2 6m iR -2 9 b
0 .001
0 .01
0 .1
1 .0
0 .05
miR-155
miR-126 miR-29b
Log2(foldchange)
-Log
10(p-value
)3
2
1
0
C o n tr o ls P a tie n ts0
1
2
3
4
5
6
7
8
9
Re
lati
ve
ex
pre
ss
ion
miR
-29
b
C D
C o n tr o ls P a tie n ts0
1 0
2 0
3 0
Re
lati
ve
ex
pre
ss
ion
miR
-15
5
C o n tr o ls P a tie n ts0
1
2
3
4
5
6
7
8
9
Re
lati
ve
ex
pre
ss
ion
miR
-12
6 ** ***** ***
RelaDv
eexpression
miR-126
RelaDv
eexpression
miR-155
RelaDv
eexpression
miR-29b
ControlsPaDentsControlsPaDentsControlsPaDents
E
B
miRNA
Foldcha
nge
168
148
128
108
88
68
48
28
86
4
2
0
212 218 27a 627 202 652 328 93 30b 204 429 335 597 618 26b 194 192135b 19b 26a 331-3p let-7d 146-5p
29b 126 155 376a 331 190 30c 125b 576 885 28-5p 10a 106a 20a 374a 100135a 200b 362-3p let-7g 29c373 17 99a
miR-126
miR-155
miR-29b
D
B
Specificity
Sensitivity
0.0
0.2
0.4
0.6
0.8
1.0
1.0 0.8 0.6 0.4 0.2 0.0
AUC=0.80
A
C
Sensi0vity
Specificity
Re
lati
ve
ex
pre
ss
ion
miR
-12
6
C o n tr o ls D ia b e te s D K D0
1
2
3
4
5
Re
lati
ve
ex
pre
ss
ion
miR
-15
5
C o n tr o ls D ia b e te s D K D0
1
2
3
4
Re
lati
ve
ex
pre
ss
ion
29
b
C o n tr o ls D ia b e te s D K D0
2
4
6
8
1 0
*
************
*****
Figure2
Rela0v
eexpression
miR-126
Rela0v
eexpression
miR-155
Rela0v
eexpression
miR-29b
ControlsDiabetesDKDControlsDiabetesDKD
ControlsDiabetesDKD
Figure3
B
RenalbiopsybeforeLCM
RenalbiopsyfollowingLCM
LCM-dissectednephrondomains
Cellcultures
C
Re
lati
ve
ex
pre
ss
ion
miR
-12
6
G P T D T0
1 0
2 0
3 0
4 0
5 0
*****
Rela>v
eexpression
miR-126
GPTDT
F
P T C F ib r o b la s t P o d o c yte G E n C0
5 0
1 0 0
1 5 0
Re
lati
ve
ex
pre
ss
ion
miR
-12
6
****
******
Rela>v
eexpression
miR-126
PTCFibroblastPodocyteGEnC
D
Re
lati
ve
ex
pre
ss
ion
miR
-15
5
G P T D T0
1
2
3
4
Rela>v
eexpression
miR-155
GPTDT
P T C F ib r o b la s t P o d o c yte G E n C0
1
2
3
4
Re
lati
ve
ex
pre
ss
ion
miR
-15
5
**
Rela>v
eexpression
miR-155
PTCFibroblastPodocyteGEnC
G
E
Re
lati
ve
ex
pre
ss
ion
miR
-29
b
G P T D T0
1
2
3
4
Rela>v
eexpression
miR-29b
GPTDT
H
P T C F ib r o b la s t P o d o c yte G E n C0
1
2
3
4
Re
lati
ve
ex
pre
ss
ion
miR
-29
b
*
Rela>v
eexpression
miR-29b
PTCFibroblastPodocyteGEnC
A
LoopofHenle
Glomerulus
GlomerularcapsuleDistaltubule
Proximaltubule
Collec>ngduct
CORTEX
MEDULLA
A B
C D
**
E F
**
***
Figure4
CulturedGEnCs GEnCcondi7onedmedium
Rela7v
eexpression
miR-126
Rela7v
eexpression
miR-126
Rela7v
eexpression
miR-29b
Rela7v
eexpression
miR-29b
Rela7v
eexpression
miR-155
Rela7v
eexpression
miR-155
5mM 5mM+TNF-α
25mM 25mM+TNF-α
5mM 5mM+TNF-α
25mM 25mM+TNF-α
5mM 5mM+TNF-α
25mM 25mM+TNF-α
5mM 5mM+TNF-α
25mM 25mM+TNF-α
5mM 5mM+TGF-β1
25mM 25mM+TGF-β1
5mM 5mM+TGF-β1
25mM 25mM+TGF-β1