Quantification of Multiple mRNAs Quantification of Multiple mRNAs from Limited Samples by qPCRfrom Limited Samples by qPCR

Bernhard G. Zimmermann, Bernhard G. Zimmermann, JianghuaJianghua Wang, Elen Ortenberg Wang, Elen Ortenberg 22, David T. , David T. WongWong

Laboratory for Head and Neck Cancer Research, Dental Research InLaboratory for Head and Neck Cancer Research, Dental Research Institute, UCLA stitute, UCLA 22BioTrove Inc, Woburn, MABioTrove Inc, Woburn, MA

PPreamplificationreamplification

Sample sizeSample size--limited and Target concentration low:limited and Target concentration low:

Single or few cellsSingle or few cellsLCMLCM

BiopsiesBiopsiesSalivaSaliva

Demands on methodDemands on method1.1. Enable analysis of multiple targets from one aliquotEnable analysis of multiple targets from one aliquot2.2. Conserve relative abundance between targets and Conserve relative abundance between targets and

samplessamples

Saliva oral fluid

Vision

Use of Saliva for Disease Diagnostics as

well as forNormal Health Surveillance

Vision

Use of Saliva for Disease Diagnostics as

well as forNormal Health Surveillance

Saliva Diagnostics for Oral CancerSaliva Diagnostics for Oral Cancer

Patient-based proteome- and genome-wide technologies to identify molecular biomarkers for HNSCC

• Incidence : 6th highest worldwide

• 30,000 new cases annually• 5-year survival rate < 50% not

changed in the last 30y • Increased risk of developing

metastasis and second malignancies

Yang Li, et al. Journal of Dental Research, 2004Park Noh Jin et al., Clinical chemistry 2006Park Noh Jin et al. Archieves oral biology, 2007.

RNA in Saliva

UCLA

• Discovered presence of cell-free RNA in saliva• Established salivary core transcriptome by array analysis: 200 common transcripts• Integrity and stability of salivary RNA• RNA stabilization reagents• Oral cancer specific markers identified by array

Poster 108

Area Area under under ROC ROC curvecurve

SensitivitySensitivity SpecificitySpecificity

0.950.95 0.910.91 0.910.91

Using 4 Saliva RNA Biomarkers: IL-1B, OAZ1, SAT and IL-8

Using 4 Saliva RNA Biomarkers: IL-1B, OAZ1, SAT and IL-8

Proof of Principle of Salivary Transcriptome analysis for Oral Cancer Detection

Proof of Principle of Salivary Transcriptome analysis for Oral Cancer Detection

Li Y, St John MA, Zhou X, Kim Y, Sinha U, Jordan RCK, Eisele D, Abemayor E, Elashoff D, Park NH and Wong DT. Salivary Transcriptome Diagnostics for Oral Cancer Detection. Clin Cancer Res: 10: 8442-8450, 2004.

ROC Curve for Logistics Regression Model

0.00

0.20

0.40

0.60

0.80

1.00

0.00 0.20 0.40 0.60 0.80 1.001-Specificity

Saliva Transcriptome Saliva Transcriptome Diagnosis*Diagnosis*

Blood Tests**Blood Tests**

Area underArea underROC curveROC curve

0.950.95 0.880.88

Saliva and blood test of oral cancer detectionSaliva and blood test of oral cancer detection

Salivary Transcriptome Diagnosis is Better than Blood Tests for Oral Cancer DetectionSalivary Transcriptome Diagnosis is Better than Blood Tests for Oral Cancer Detection

*Yang Li, et al. Clinical Cancer Research, 10: 8442-8450, 2004.

**Yang Li, et al., J. Clin. Oncology, 24: 1754-1760, 2006.

*Yang Li, et al. Clinical Cancer Research, 10: 8442-8450, 2004.

**Yang Li, et al., J. Clin. Oncology, 24: 1754-1760, 2006.

Saliva Saliva -- challengeschallenges

Limited sampleLimited samples:s:Sample access Sample access -- nnumber umber Sample vSample volumeolumeRNA concentration and integrityRNA concentration and integrity

Make the best out of one sample:Make the best out of one sample:Validation of Validation of ddisease markersisease markers, novel candidates, novel candidates and strategies and strategies for normalizationfor normalizationMarker discovery by arrayMarker discovery by arrayWhole salivaWhole saliva andand supernatant supernatant SSample processingample processing

Optimization of analytical processOptimization of analytical processAdditional Additional analytesanalytes: proteins, DNA, : proteins, DNA, miRNAmiRNA

PreamplificationPreamplification

Global amplification vs. targeted Global amplification vs. targeted amplificationamplificationEase of useEase of useRobustnessRobustnessCostCost

Hypothesis for PCR based Hypothesis for PCR based preamplificationpreamplification

1 step RT1 step RT--PCR with limited primers is unbiased if all PCR with limited primers is unbiased if all targets are amplified with the same efficiencytargets are amplified with the same efficiency

PeixotoPeixoto, et al., Quantification of multiple gene expression in individu, et al., Quantification of multiple gene expression in individual cells. Genome Res, 2004.al cells. Genome Res, 2004.

multiplex RT-PCR

qPCRDilution

““The easy wayThe easy way””11--step RTstep RT--PCR, dilution, qPCRPCR, dilution, qPCR

uRNA titration 25 ng to 100 pg (in tRNA), 15 cycles preampplification, 1/10 dilution, qPCR

Clear evidence of formation of unspecific products

Accurate quantification not possible

Problem solutionsProblem solutions1.1. Specific detection Specific detection

qPCR probes (AB: micro RNA, early access mRNA)qPCR probes (AB: micro RNA, early access mRNA)Primer extension, Primer extension, ligaligationtion etc.etc.BUT: expensive, large scatter for high multiplexes!BUT: expensive, large scatter for high multiplexes!

CleanupCleanup and and dilutiondilution

2.2. Elimination of backgroundElimination of backgroundPrimer carry over Primer carry over –– all these are all these are problem problem downstream: downstream: extended in 2nd roundextended in 2nd round qPCRqPCRPCR products of multiple targetsPCR products of multiple targetsPossibly multiple primer Possibly multiple primer dimerdimer products and unspecific products and unspecific productsproducts

Effectiveness of the cleanEffectiveness of the clean--upup

S100A8

ANXA2

Final WorkflowFinal Workflow

• 10 -50 targets

• PCR product < 130 bp

• removes leftover primers and other ssNA

• removes dNTPsadditional stabilization

• reduces overall background of products and unspecific products • qPCR of nested

product (40 – 100 bp)

•SYBR Green based

multiplex RT-PCR

Exonucleasecleanup Dilution

qPCR

Titration experimentTitration experimentIVTIVT--StandardStandard--curvescurves

17

19

21

23

25

27

29

31

2 3 4 5

Log (template)

CT

• 10-plex preamplification

• cleanup

• 10-fold dilution

• Single-plexsemi-nestedqPCR

ACTB ANXA2 CSTA GAPDH RPL37 RPS10 RPS16 RPS6 RPS7 S100A8Slope -3.30 -3.29 -3.32 Failed -3.24 -3.39 -3.28 -3.39 -3.36 -3.36RSQ 0.992 0.991 0.985 0.994 0.989 0.996 0.992 0.994 0.996

Intercept 35.11 31.82 34.00 34.55 33.66 33.74 32.75 34.94 32.56Efficiency 101.0% 101.3% 99.9% 103.4% 97.2% 101.7% 97.1% 98.5% 98.5%

Gene expression patternGene expression pattern

0

5

10

15

20

25

30

35

40

ANXA2

ACTB

CSTA

RPL37

RPS10

RPS16

RPS6

RPS7S10

0A8

Cyc

le

• Total RNA diluted between 100 ng and 6.1 pg

• 10-plex preamplification

• Cleanup

• 1000-fold dilution

• semi-nested qPCR

Replicate amplifications: 1/1000 diluted preamp product

5 replicate RT reactions. 4 Replicate qPCR of one RT.

RPS16

S100A8

ACTB

RTRT--PCR and qPCR replicatesPCR and qPCR replicates

FlexibilityFlexibility:: Cycle numberCycle number

5, 10, 15 and 20 cycles 5, 10, 15 and 20 cycles ppreamplificationreamplification

Left: S100A8 Right: ANXA2

uRNA 1/40

uRNA 1/160

• Total RNA

• 10-plex preamplification

• Cleanup

• 10-fold dilution

• qPCR

NNanoano--PCRPCR:: Promises Promises

Parallel-plexingPosters 103 and 111

NanoNano--PCRPCR

NanoNano PCRPCRPOP 1:POP 1:

Reduced copy number: saliva cDNA would be too Reduced copy number: saliva cDNA would be too dilute dilute ((14140 times less target when using same template)0 times less target when using same template)

Assay failure Assay failure 80% but80% but successful assayssuccessful assays!!

Assayamplicon length preamplification

amplicon length qPCR Tm qPCR

Tm nano PCR

CSTA 98 53 71.1 n.D.ANXA2 70 47 71.5 n.D.RPS6 82 64 73.9 73.8RPS7 75 52 72.4 n.D.RPS10 92 39 72.7 n.D.RPS16 84 56 72.8 n.D.RPL37 90 68 74.9 74.0ACTB 61 44 73.8 n.D.S100A8 77 46 72.7 n.D.GAPDH 170 140 78.6 n.D.

POP 1 POP 1 –– RPS6RPS6

Dilutions of RNA Dilutions of RNA preamplifiedpreamplified5 5 –– 20 cycles20 cycles

y = -3.81x + 16.17R2 = 0.991

y = -4.15x + 22.55R2 = 0.995

y = -4.11x + 26.50R2 = 0.992

y = -4.17x + 31.38R2 = 0.990

y = -4.18 + 33.44R2 = 0.984

0

5

10

15

20

25

30

0 1 2 3 4

log Conc

CT

5 cycles10 cycles14 cycles20 cyclesIVT

POP 1 POP 1 –– salivary mRNAsalivary mRNA

qPCR qPCR vsvs nanonano--PCRPCR

y = 1.21x - 4.99R2 = 0.85

15

17

19

21

23

25

27

17 19 21 23 25 27

CT AB

CT

BT

Averaged duplicate PCRs

ANXA2

Left: 9 replicates 0.1 ng RNA

Right: triplicates of 3 saliva RNAs

NanoNano PCR PCR –– POP 2POP 2New assay design:New assay design:

Amplicon length: Amplicon length: 70 70 –– 90 90 bpbpTm qPCR: > 75 Tm qPCR: > 75 ººCC

average Tm average Tm nanonano qPCR: 2 qPCR: 2 ººC lower!C lower!

Assay success: 100%Assay success: 100%Amp curve dilution SAT and ANXA2Amp curve dilution SAT and ANXA2

1010--plex and 22plex and 22--plex by plex by nanonano PCRPCR

0.00

5.00

10.00

15.00

20.00

25.00

30.00

ANXA2 DUSP1 H3F3A IL-1B IL-8 OAZ1 RPL37 S100A8 S100P SAT

assay

CT

0.00

5.00

10.00

15.00

20.00

25.00

30.00

ANXA2

DUSP1

H3F3A

IL-1B IL-

8

OAZ1

RPL37

S100A

8

S100P SAT

assay

CT

y = -4.0877x + 4.5924R2 = 0.9793

y = -4.1579x + 4.2154R2 = 0.9882

y = -4.2186x + 3.1352R2 = 0.9678

y = -3.9841x + 3.9637R2 = 0.995

0

5

10

15

20

25

-5 -4 -3 -2 -1 0

log conc

CT

73 - 10 assays

73 - all assays

75 - 10 assays

75 - all assays

NormalizationNormalizationSalivary internal reference genes (SIR)Salivary internal reference genes (SIR)Validation of ten candidatesValidation of ten candidates from core transcriptomefrom core transcriptome5 5 µµl RNA per samplel RNA per sample

Selection criteria:Selection criteria:No statistically significant difference between oral No statistically significant difference between oral cancer and healthycancer and healthy> 1 normalizer> 1 normalizerNo or minimal correlation between normalizer No or minimal correlation between normalizer genesgenes

SIR normalizationSIR normalization

TT--testtest BoxplotBoxplot::control group (n=14) control group (n=14)

vs. vs. oral cancer (n=18)oral cancer (n=18)

fold t pval wil pvalANXA2 1.37 0.663 0.489RPS6 2.3 0.303 0.125RPS7 2 0.395 0.293RPS10 2.24 0.299 0.097RPS16 2.07 0.356 0.211RPL37 1.86 0.318 0.293ACTB 1.47 0.575 0.594S100A8 1.27 0.74 0.767CSTA 0.9 0.902 0.953

SIR normalizationSIR normalizationClustering reveals 3 clusters Clustering reveals 3 clusters –– 1 gene from each selected1 gene from each selected

ANXA2 RPS6 RPS7 RPS10 RPS16 RPL37 ACTB S100A8 CSTAANXA 1 0.82 0.85 0.91 0.95 0.71 0.8 0.4 0.19RPS6 0.82 1 0.9 0.92 0.88 0.65 0.63 0.06 -0.1RPS7 0.85 0.9 1 0.96 0.95 0.51 0.55 0.04 -0.15RPS10 0.91 0.92 0.96 1 0.97 0.66 0.68 0.16 -0.02RPS16 0.95 0.88 0.95 0.97 1 0.67 0.72 0.23 0.02RPL37 0.71 0.65 0.51 0.66 0.67 1 0.92 0.67 0.59ACTB 0.8 0.63 0.55 0.68 0.72 0.92 1 0.71 0.54S100A8 0.4 0.06 0.04 0.16 0.23 0.67 0.71 1 0.9CSTA 0.19 -0.1 -0.15 -0.02 0.02 0.59 0.54 0.9 1

SIR normalizationSIR normalization

-4

-2

0

2

4

6

8

AN

XA2

DU

SP1

H3F

3A

IL1B IL

8

OA

Z1

RPL

37

S100

A8

S100

P

SAT

delta

CT

Dilution of total RNADilution of total RNANormalization of each Normalization of each marker by geometric marker by geometric mean of 3 mean of 3 SIRsSIRs::∆∆CCTT = C= CTT –– CCT3SIRsT3SIRs

Pattern Pattern independentindependentof RNA inputof RNA inputStdDevStdDev of replicate of replicate RTRT--PCR reducedPCR reducedSD between different SD between different RNA input RNA input ~ replicates ~ replicates at same inputat same input

QC: New concept:QC: New concept:9 e9 exogenous mRNA Spikesxogenous mRNA Spikes

• Added as RT-PCR controls to reaction mixAlso:

• Sample processing control (extraction)

Commercial source: IVT mRNAs from E. coli

(ArrayControl RNA Spikes and RNA Control 250, Ambion, Austin, TX)

101044545420022002Spike 9Spike 9

101044545420262026Spike 8Spike 8

101011555514741474Spike 7Spike 7

101022555512761276Spike 6Spike 6

101033535310591059Spike 5Spike 5

101044505010271027Spike 4Spike 4

101044565610261026Spike 3Spike 3

1010445454776776Spike 2Spike 2

1010445050776776Spike 1Spike 1

copies per preamplificationcopies per preamplificationGC content ( % )GC content ( % )length (bases)length (bases)

101044545420022002Spike 9Spike 9

101044545420262026Spike 8Spike 8

101011555514741474Spike 7Spike 7

101022555512761276Spike 6Spike 6

101033535310591059Spike 5Spike 5

101044505010271027Spike 4Spike 4

101044565610261026Spike 3Spike 3

1010445454776776Spike 2Spike 2

1010445050776776Spike 1Spike 1

copies per preamplificationcopies per preamplificationGC content ( % )GC content ( % )length (bases)length (bases)

Control and Calibration by SpikesControl and Calibration by Spikes

-12

-10

-8

-6

-4

-2

0

2

4

6

8

10

AN

XA2

DU

SP1

H3F

3A

IL1B IL

8

OA

Z1

RPL

37

S100

A8

S100

P

SAT

3 SI

Rs

delta

CT

5

7

9

11

13

15

17

19

21

23

25

27

29

AN

XA2

DU

SP1

H3F

3A

IL1B IL

8

OA

Z1

RPL

37

S100

A8

S100

P

SAT

3 SI

Rs

CT

Before After

ANXA2 DUSP1 H3F3A IL1B IL8 OAZ1 RPL37S100A

8 S100P SAT 3SIRsslope -3.68 -3.63 -3.43 -3.76 -3.62 -3.50 -3.90 -3.63 -3.63 -3.64 -3.77RSQ 0.986 0.982 0.979 0.965 0.979 0.982 0.987 0.985 0.981 0.988 0.988median SD 0.30 0.52 0.52 0.77 0.30 0.21 0.45 0.47 0.42 0.25 0.35

slope -3.08 -3.03 -2.99 -3.02 -2.98 -2.90 -3.30 -3.03 -3.03 -3.04 -3.17RSQ 0.993 0.994 0.985 0.982 0.979 0.987 0.996 0.996 0.992 0.993 0.997median SD 0.23 0.25 0.32 0.38 0.29 0.23 0.16 0.19 0.30 0.24 0.13

CT values

∆CT values (spike corrected)

total RNA Effecttotal RNA Effect

RNA below 10ng: RNA below 10ng: CCTT increases by increases by 0.8 cycles per 0.8 cycles per 10x RNA 10x RNA concentrationconcentration

Carrier Carrier tRNAtRNA15

17

19

21

23

25

27

29

SPIK

E 1

SPIK

E 2

SPIK

E 3

SPIK

E 4

SPIK

E 5

SPIK

E 6

SPIK

E 7

SPIK

E 8

SPIK

E 9

CT

4 fold diluted RNA between 100 ng and 6 pg

ConclusionConclusionssMultiplex Multiplex preamplificationpreamplification::

Conserves expression profiles for > 50 targetsConserves expression profiles for > 50 targetsExcellent quantitative performance Excellent quantitative performance and novel controlsand novel controls

Endogenous mRNAs for normalizationEndogenous mRNAs for normalizationNovel exogenous spikeNovel exogenous spike--in controls (& calibration)in controls (& calibration)

One reaction One reaction –– all targets: variabilityall targets: variabilityAll targets and controls in one reactionAll targets and controls in one reactionSensitivity Sensitivity –– high input, no dilution after RThigh input, no dilution after RTSpecificity of nested PCRSpecificity of nested PCR

SStraightforward and flexibletraightforward and flexibleAssay designAssay designCycle number for preampCycle number for preampTarget input > 6 orders of magnitudeTarget input > 6 orders of magnitudeCostCost: : preamppreamplificationlification in 10 in 10 µµl RTl RT--PCR mix and enzymatic cleanup of 5 PCR mix and enzymatic cleanup of 5 µµl:l:< 2 $ < 2 $ per reaction per reaction -- plus primersplus primers< 2 hours for RT< 2 hours for RT--PCR PCR preamplificationpreamplificationStability of PCR product vs. cDNA Stability of PCR product vs. cDNA Amenable to automationAmenable to automation

Take home messageTake home messageEnzymatic cleanup removes backgroundEnzymatic cleanup removes backgroundNew dimensions of (RTNew dimensions of (RT--)PCR)PCR

Limited samples Limited samples –– multiple targetsmultiple targetsNanoNano--PCR PCR –– high throughputhigh throughputQC: Normalization and controls QC: Normalization and controls –– research and research and diagnosticsdiagnostics

Applicable to a wide range of Applicable to a wide range of Applications: research and clinicalApplications: research and clinicalSampleSamples s –– any sample and concentrationany sample and concentrationTargetsTargets –– any nucleic acidsany nucleic acidsDDownstream applicationsownstream applications

PCR PCR –– qPCR et al.qPCR et al.LigationLigationHybridizationHybridizationPrimer extension: sequencing, MS etc.Primer extension: sequencing, MS etc.

AcknowledgementsAcknowledgements

David WongDavid WongJianghuaJianghua WangWangNoh Jin ParkNoh Jin ParkZhanzhiZhanzhi Mike Mike HuHu, Brad Henson, Brad HensonTianweiTianwei YuYuDavid ElashoffDavid ElashoffBioTrove BioTrove –– ElenElen OrtenbergOrtenbergAmpTech AmpTech –– Guido KruppGuido Krupp

Funding support: NIH grantFunding support: NIH grantss RO1 DE 15970 and T32 RO1 DE 15970 and T32 DE 07296DE 07296