A universal method for elimination of hemolyzed plasma samples that improves miRNA signature performance for early detection of colorectal cancerPeter Mouritzen1, Søren J. Nielsen1, Nana Jacobsen1, Jan Stenvang2, Thorarinn Blondal1, Torben Ørntoft3, Nils Brünner2, Claus L. Andersen3, Hans J. Nielsen4 and Adam Baker1 1Exiqon A/S, Vedbaek, Denmark; 2University of Copenhagen, Copenhagen, Denmark; 3Aarhus University Hospital, Aarhus, Denmark; 4Hvidovre Hospital, Copenhagen, Denmark

Conclusions and future prospects•Aplasma/serummicroRNAPCRprofilingplatformwithunparalleledsensitivitywasdeveloped•Asimpleworkflowfromsamplepreparationtodataanalysisallowstestresultsfrom100uLof plasmawithinoneworkingday•HemolysiswasdemonstratedtobeamajordeterminantofsamplequalityformicroRNAprofiling•HemolysiscorrelatedwithhospitalIDandwiththeuseofspecificplasmacollectiontubes•Analysisofthrombocytecontaminationison-going•AplasmamicroRNAsignaturebasedonalimitednumberofmicroRNAswasdevelopedinone hospital and validated in samples from independent hospitals•Validationofthesignatureinalargersetof>1000samplesfrommultiplehospitalsison-going•Aprospectivetrialcollecting5000samplesforvalidationhasbeeninitiated

ThisworkwassupportedbytheDanishAdvancedTechnologyFoundationgrant007-2009-2

ConcerningmiRCURYLNA™UniversalRTmicroRNAPCR:NOTICETOPURCHASER:LIMITEDLICENSEPurchaseofthisproductincludesanimmunityfromsuitunderpatentsspecifiedintheproductinserttouseonlytheamountpurchasedforthepurchaser’sowninternalresearch.Nootherpatentrightsareconveyedexpressly,byimplication,orbyestoppel.FurtherinformationonpurchasinglicensesmaybeobtainedbycontactingtheDirectorofLicensing,AppliedBiosystems,850LincolnCentreDrive,FosterCity,California94404,USA.

Figure 5. Effect of hemolysis on plasma microRNA profiles. TheRBCdilutionserieswasprofiledfortheexpressionof378microRNAs.ThemicroRNAsweresortedbasedonmaximalchangeofexpressionaftertheadditionofRBClysate.microRNAsaffectedbyandimperviousto hemolysis are indicated.

Figure 9. Biomarker discovery and classifier selection. 100samplesfromhospital2wereusedtogeneratealistofcandidatemarkers.The64mostdifferentiallyexpressedmicroRNAswereusedtogenerateaPCAplot(A).AsignaturebasedonalimitedsetofmicroRNAswasdevelopedandevaluatedusingreceiver-operatorcharacteristicscurve(B).

Table 4. Performance of the microRNA signature in the different sample sets.

All samples Hospital 2(discovery)

Remaining hospitals(validation)

AUC 0.68 0.81 0.87

Sensitivity(%) 67 75 82

Specificity(%) 65 79 89

Cancer

Normal

PC2

PC1

0

0.2

0.4

0.6

0.8

1

0 0.2 0.4 0.6 0.8 1

sens

itivi

ty

1-specificity

A: PCA plot for microRNA signature performance

B: ROC curve for microRNA CRC signature

Figure 3. Superior sensitivity and linearity of the LNA™-enhanced miRNA RT-qPCR Platform. Apoolofsynthetictemplatesfor647microRNAswassubjectedtoserialdilution(thelowestinputrepresents15copiesofeachtemplateRNAinthePCRreaction)andthenassayedbyRT-qPCR.ThemedianCtvalueforallassayswasthenplottedagainsttemplateconcentration,demonstratinglinearityoftheassayplatformdownto15copiesoftemplate.

Figure 4. Effect of hemolysis on plasma miR-451 levels. Toppanel:Anon-hemolyzedplasmasamplewasspikedwithincreasingamountsofredbloodcell(RBC)lysateandanalyzedforhaemoglobincontent(A414)andmiR-451levels(qPCR).Bottompanel:Aselectionof50plasmasamplesfromthestudywasanalyzedforhaemoglobincontentandmiR-451levels.Potentialcut-offvaluesare indicated by red lines.

Med

ian

Ct

28

30

29

31

32

33

34

35

36

37

-6-7 -5 -4 -3 -2 -1 0Synthetic RNA log concentration

Sensitivity and linearity of Exiqons qPCR system

Plasma or serum

R = 0.9992

All -MedianLinear (All - Median)

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

0

OD

414

nm

%RBC (v/v)

A: RBC contamination of plasma sample

Hemolysis have a strong effect on miR451 expression

0 0.2 0.4 0.6 0.8 1.0

Hemolysiscut-off

30

28

26

24

22

20

18

16

OD 414nm

miR

-451

Cp

0 0.05 0.1 0.15 0.2 0.25 0.35 0.4 0.450.3

Hemolysiscut-off

HemolyzedNot Hemolyzed

A: Effect of hemolysis on plasma miR-451 levels

15

20

25

30

35

156 microRNAs

microRNAs affected by RBC contamination microRNAs not affected by RBC contamination

Expr

essi

on le

vel (

Raw

Cq)

miR-451 miR-122

Overlaid box plot

1.0% RBC

0.032% RBC

Plasma pool

microRNA expression profile from plasma samples with different levels of Red Blood Cell (RBC) contamination

Table 3. Use of plasma collection tubes at participating hospitals. 1)GreinerVacuette,2)BDVacutainer,3)TerumoVenosafe.Hospitalswithhighincidenceofhemolysisarehighlightedinred.

Discovery and validation of a plasma-based microRNA signature for early detection of CRCAsetof325plasmasamplesfromCRCpatientsandhealthycontrolsfrom7Danishhospitals(Table2)wasprofiledacross378microRNAscommonlyfoundinhumanplasma.Initialanalysisofthedataset demonstratedpoorseparationbetweenthegroups(ROCAUC=0.68,Table4).Principalcomponentanalysis(PCA)ofallcontrolsamplesshowedmarkeddifferencesinmicroRNAprofilesbetweenthedifferenthospitals,suggestingthatthesedifferencesmayhampertheidentificationofgoodCRC-relatedbiomarkers.Wethereforeevaluatedasetofparametersrelatedtosampleacquisitionandstorage.Onesuchparameter,hemolysis(orthepresenceofredbloodcelllysateinthesample),wasshowntodiffergreatlybetweenhospitals,andtocorrelatewiththeoutlierhospitalsinourPCA(Figures7-8).

Wetheneliminatedsamplesfromtheoutlierhospitalsandusedthedatafromhospital2(N=100)todevelopasignatureofalimitednumberofmicroRNAsthatdistiguishwellbetweenCRCandcontrols(AUC=0.81,Figure9andTable4).Finally,wewereabletovalidatethesignatureinsamplesfromtheotherhospitalsnotaffectedbyhemolysis(AUC=0.87,Table4).

Table 2. Clinical material for signature discovery and validation. Plasmasampleswereassayedforthepresenceof378microRNAsbyRT-qPCR.

Cancers Controls

Mean age(Range) 70(40-94) 69(33-93)

GenderMale 86 79

Female 83 77

StageII 119 N/A

III 50 N/A

Figure 7. Variation in microRNA profile among participating hospitals. Theprofilesofthe131microRNAsexpressedinall153controlsampleswereusedasinputinaPCA. ThefirstandsecondcomponentsandthehospitalIDsareindicatedonthefigure.

321 654

PC2

PCA plot for microRNA profile among participating hospitals

PC1

Figure 8. Variation in selected microRNAs among participating hospitals. TheprofilesofmiR-122(un-affectedbyhemolysis)andmiR-451(hemolysismarker)areshown.Hospitals1,5,6allcontributedamajorproportionofhemolyzedsamples.

30

28

26

24

22

20

18

Hospital ID 1 2 3 4 5 6 7

Hemolysis marker: hsa-mir-451 expression profile from plasma samples

0 50 100 150 200 250 300Hem

olyz

edN

ot H

emol

yzed

36

34

32

30

28

26

24

Hospital ID 1 2 3 4 5 6 7

hsa-mir-122

microRNA expression profile from plasma samples

0 50 100 150 200 250 300Axis

text

: Exp

ress

ion

leve

l

Hospital ID 1 2 3 4 5 6 7Collection tube 1 2 2 2 3 3 2

Serum/Plasma microRNAs are promising disease biomarkers• Playimportantregulatoryroleinmanydiseases including cancers• Integratesbiologyfromentireorganismincluding diseased tissue• Minimallyinvasive• Routinelyobtainedinhospitalsandatgeneralpractitioners• Largehistoricalcollectionsexistfordiscovery• microRNAsarepresentinlowbutdetectableamounts• PlasmamicroRNAsarestableunderstandardsampling and storage conditions

Three challenges of working with serum/plasma• Serum/plasmacontainsRNasesandEnzymeinhibitors• Serum/plasmacontainslowamountsofnucleicacid• Serum/plasmaiscell-freebutmaybecontaminatedbybloodcells

Our solutions• Avoidtheuseofheparincollectiontubes• Usespike-instomonitorforco-purificationofinhibitors/RNases• UsecarrierRNAduringpurification• UseasensitiveanalyticalsystemsuchastheLNA™-enhancedmiRNA RT-qPCRplatform• Avoidtransferofcellularmaterialduringsampleacquisition• Minimizehemolysisduringacquisition• ProcessplasmaatRTwithin2hrsofphlebotomy• QCplasmaforhaemoglobincontent• MonitorPCRdatasetforsignsofhemolysis

Figure 1. Clinical Source of biomarker – 10 mL blood collection. Nucleicacidspresentindifferentbloodfractionsfroma10mlbloodsample.

Figure 2. Effect of carrier RNA.Plasma(200µL)fromtwoindividualswaspurifiedintheabsenceorpresenceofcarrierRNA(MS2phageRNA)andassayedbyRT-qPCRforthepresenceofthreemiRNAs.NotetheincreaseddetectionanddecreasedvariabilityinsamplespurifiedwithcarrierRNA.

10 mL blood sample

Blood plasma contains small amount of RNA

Plasma• ~5.5 mL• 1-50 ng RNA• <100 ng DNA ( in disease)

Red Blood Cells• ~4.5 mL• ‘low’ RNA• ‘0’ DNA

Buffy coat• <0.1 mL• 10-100 µg RNA• 200-600 µg DNA

25

27

29

31

33

35

37

39

Mea

n Cq

miR-192

Plasma I, no carrier

Plasma II, no carrier

Plasma I, with carrier

Plasma II, with carrier

let-7a miR-103

25

27

29

31

33

35

37

39

Mea

n Cq

miR-192

Plasma I, no carrier

Plasma II, no carrier

Plasma I, with carrier

Plasma II, with carrier

let-7a miR-103

25

27

29

31

33

35

37

39

Mea

n Cq

miR-192

Plasma I, no carrier

Plasma II, no carrier

Plasma I, with carrier

Plasma II, with carrier

let-7a miR-103

Colorectal cancerColorectalcancer(CRC)isamajorcauseofmortalityinthewesternworld.EarlydetectionofCRCimprovessurvivalandscreeningforCRChasbeenclinicallyproventolowerCRC-relatedmortality.However,although population screening programs have been implemented in a number of countries, screening ratesamongthe50-75yearoldsareunsatisfactory.

There is therefore a clear unmet need for a quick, sensitive, specific, and minimally invasive screening assay to select at risk individuals for definitive diagnosis by colonoscopy.

Figure 6. Colorectal cancer stages.

Stage IStage 0

Stage II

Spread to other organs

Stage III

Stage IV

Table 1. Colorectal cancer survival rates in US. EarlydetectionofCRCimprovessurvival.

CRC Stage 5 year relative survival Treatment

0-I 93% Surgery

II 80% Surgery

III 58% Surgery/adjuvantchemotherapy

IV 6.9% Chemotherapy

9206

02-

v1.

0-

8/20

12