FFPE Quality: Impact on Downstream DNA Performance in Molecular Diagnostic Assays...
1
FFPE DNA Quality: Impact on Downstream Performance in Molecular Diagnostic Assays FFPE DNA Quality: Impact on Downstream Performance in Molecular Diagnostic Assays Jonathan M. Eibl, Julie A. Toplin, Chad D. Galderisi, Cindy S. Spittle, MolecularMD Corp., Portland, OR Introduction Personalized medicine for oncology patients is driven by mutation profiling of tumors. Improved robust DNA extraction methods that produce high quality FFPE DNA template Poster #: ID102 Figure 3a. Maxwell vs. MO BIO DNA extraction from different FFPE sample types (total DNA Yield). Overall, DNA yields were greater for l d h ll ( f l ) Figure 3b. Percent increase in DNA yield by Maxwell vs. MO BIO in different FFPE sample types (1.9‐fold or 190% increase overall). Figure 6a. Box plot graphs of Phred scores for Sanger sequences (see Figure 5; forward traces; 51‐bp window; melanoma samples). Figure 6b. Box plot graph of Phred scores for Sanger sequences (see Figure 5; reverse traces; 51‐bp window; melanoma samples). suitable for complex molecular diagnostic testing are needed. In addition to DNA fragmentation and protein‐DNA cross linking, some FFPE tissues contain inhibitors and/or quenching substances (e.g. melanin) that adversely affect downstream performance in PCR and sequencing assays. In this study, we provide a side‐by‐side evaluation of two DNA isolation methods from FFPE tissue of various tumor types: a manual, column‐based method (MO BIO) and the Maxwell CSC automated platform (Maxwell). 40 50 60 g) Maxwell MO BIO samples processed on the Maxwell CSC (30 of 34 samples, 88%). 500 600 700 800 A yield NO SEQUENCE NO SEQUENCE O SEQUENCE O SEQUENCE Materials & Methods FFPE tissue from melanoma, GIST, endometrial, and NSCLC tumor types was obtained from equivalent numbers of slides for each extraction method. Manual extractions were performed using the MO BIO BiOstic FFPE Tissue DNA Isolation Kit followed in some 10 20 30 DNA Yield (ug 100 200 300 400 500 % increase in DNA N N N N performed using the MO BIO BiOstic FFPE Tissue DNA Isolation Kit followed in some instances by an additional purification step using the Zymo Research OneStep PCR Inhibitor Removal Kit to rid DNA isolates of visible melanin. Automated extractions were carried out using the Maxwell CSC DNA FFPE Custom Kit on the Maxwell CSC instrument with no additional PCR inhibitor removal steps. DNA eluate yields and optical densities were determined via spectrophotometry. Equivalent input DNA amounts from melanoma samples were assayed using both NRAS allele‐specific qPCR (25 ng) and NRAS Sanger 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 Melanoma GIST Endometrial NSCLC 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 Melanoma GIST Endometrial NSCLC samples were assayed using both NRAS allele specific qPCR (25 ng) and NRAS Sanger sequencing (200 ng) on the BIORAD CFX platform and Applied Biosystems 3730 DNA Analyzer, respectively. NRAS qPCR data were reported as delta‐Ct values to determine qPCR differences. Sanger sequencing Phred scores were analyzed for the 51‐base pair window surrounding the NRAS codon 61 hotspot to quantitatively determine sequence‐trace quality. Quality metrics were further evaluated using the CodonCode Aligner and Sequence Scanner software packages 3 3.5 4 OD 260/280 Maxwell OD 260/280 MO BIO 2.5 3 OD 260/230 Maxwell OD 260/230 MO BIO Figure 4a. Optical densities (260/280) Maxwell vs. MO BIO DNA extraction from different FFPE sample types. Figure 4b. Optical densities (260/230) Maxwell vs. MO BIO DNA extraction from different FFPE sample types. Figures 7a – 7e. Box plots representing sequence quality metrics using CodeCodon Aligner and Sequence Scanner software. 7a. Results Both extraction methods require the same upfront steps of scraping FFPE tissue, reagent addition, and lysis/incubation, however, the Maxwell method requires fewer hands‐on, liq id handling steps (Fig 1) Additional Z mo clean p steps ere req ired for 2 man all software packages. 0.5 1 1.5 2 2.5 OD (260/280) 0.5 1 1.5 2 OD (260/230) liquid‐handling steps (Fig. 1). Additional Zymo cleanup steps were required for 2 manually‐ extracted samples due to visible melanin in the FFPE tissue/eluted DNA (Fig. 2; Samples 02 and 07). DNA yields were greater for samples processed on the Maxwell CSC (30 of 34 total samples, 88%; 1.9‐fold increase overall; Figs. 3a and 3b). Optical densities for both methods were largely congruent for the 260/280 ratio (Fig. 4a) but more variable for the 260/230 ratio (Fig. 4b). In qPCR, 11 of 12 (92%) of the Maxwell‐extracted melanoma samples crossed the threshold 1 3 Ct sooner on a erage than the corresponding man all e tracted 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 Melanoma GIST Endometrial NSCLC 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 Melanoma GIST Endometrial NSCLC Figure 5a NRAS allele‐specific PCR for DNA extracted by both methods Figure 5b Comparison of NRAS Sanger sequencing results for DNA extracted by 7b. 7c. the threshold 1.3 Ct sooner on average than the corresponding manually‐extracted templates (Fig. 5a), and one manually‐extracted, Zymo‐purified sample failed to amplify (Sample 07). Two Maxwell‐extracted samples (Samples 10 and 11) crossed the threshold markedly sooner (~2.5 Ct). Similarly, all Maxwell‐extracted melanoma DNA isolates produced robust bidirectional sequence, whereas two manually‐extracted isolates failed to yield NRAS sequence despite the inhibitor‐removal step (Fig. 5b). Aside from the sequence f il f th t lt d dditi l t lt hi h hibit d l td b li Sample C t Value ∆ C t Value (MO BIO – Maxwell) MO BIO 01 33.04 1.11 Maxwell 01 31.93 MO BIO 02 Maxwell 02 Figure 5a. NRAS allele‐specific PCR for DNA extracted by both methods from melanoma FFPE; traces shown at right for samples with ∆C t Values ˃ 2.0 (Samples 02, 10, 11) and no amplification/amplification (Sample 07). Figure 5b. Comparison of NRAS Sanger sequencing results for DNA extracted by both methods from two melanoma samples, both of which required the Zymo cleanup step for the MO BIO extraction. failures of these templates and one additional templatewhich exhibited elevated baseline noise, NRAS codon 61 mutation‐detection results for the remaining nine manually‐extracted eluates were identical to the Maxwell‐extracted samples (data not shown). Templates from both extraction methods exhibited higher median Phred scores and narrower interquartile ranges (IQR) for reverse traces than for forward traces (Figs. 6a and 6b); this characteristic is attributed to the proximal location of the forward primer relative to the NRAS codon 61 ht t S f M ll t td FFPE DNA hibit hi h Ph d MO BIO 02 35.42 2.17 Maxwell 02 33.25 MO BIO 03 29.93 0.37 Maxwell 03 29.56 MO BIO 04 32.09 0.66 Maxwell 04 31.43 MO BIO 05 32.50 0 73 MO BIO 07 Maxwell 07 Maxwell 02 Maxwell 07 F W D F R E V 7d. 7e. hotspot. Sequences from Maxwell‐extracted FFPE DNA exhibit higher mean Phred scores and less variability (insets; Figs. 6a and 6b). Overall, metrics generated using other software packages (i.e., quality values, trace scores, quality value 20+, continuous‐read length, and signal‐to‐noise ratio) illustrate equivalent or better results for the Maxwell‐extracted templates especially in regards to variability (see IQRs for boxplots; Figs. 7a –7e). Conclusions 0.73 Maxwell 05 31.77 MO BIO 06 34.64 1.94 Maxwell 06 32.70 MO BIO 07 No amplification N/A Maxwell 07 35.07 MO BIO 08 30.10 ‐0.02 Maxwell 08 30.12 MO BIO 10 Maxwell 10 W D R E V MO BIO 02 FWD This comparison highlights the potential advantages of using an automated nucleic acid extraction platform for onco‐diagnostic testing: • Standardized workflow with fewer hands‐on steps minimizing the likelihood of cross‐contamination • Higher throughput for faster turn‐around‐time of results ih i ld i ii h il d f li i d f i Figure 1. Comparison of extraction method procedural steps and hands‐on time Scrape FFPE Add reagents Lyse (incubate) Steps common to both extraction methods MO BIO 09 30.11 1.72 Maxwell 09 28.39 MO BIO 10 30.93 2.60 Maxwell 10 28.33 MO BIO 11 30.80 2.55 Maxwell 11 28.25 MO BIO 12 28.44 0 05 MO BIO 11 Maxwell 11 MO BIO 07 FWD REV REV • Higher DNA yields, maximizing the potential data output from limited amounts of FFPE tumor tissue • Higher quality DNA without extra purification steps, minimizing the potential need for repeat analysis Maxwell: post‐lysis steps 1 – 12 (Hands‐on time for 16 samples ‐ ~30 min.) Cool to RT Add RNase A Incubate (5 min.) Place cartridges Remove seals Place plungers Place elution tubes Add elution water Spin FFPE digest (5 min.) Add sample to cartridge Enter sample data / run Remove elution tube Figure 2. Representative specimen (Sample 02) requiring additional Zymo cleanup steps due to high melanin content (note red boxed area); H&E staining, 4X magnification. 0.05 Maxwell 12 28.39 Acknowledgements The authors would like to thank Jeannette Nussbaum for her valuable contributions For further information please contact [email protected] or visit www.molecularmd.com . MO BIO: post‐lysis steps 1 – 22 (Hands‐on time for 16 samples ‐ ~60 min.) Heat (1 hr.) Spin (1 min.) Transfer Add FP4 Vortex Add FP5 Vortex Spin Apply to column Spin (1 min.) Transfer spin filter Add FP6 (continue) Transfer column Spin (1 min.) Add FP7 (1 of 2) Spin (1 min.) Add FP7 (2 of 2) Spin (1 min.) Spin (2 min.) Add FP8 Incubate (5 min.) Spin (1 min.) ZYMO cleanup (as needed) * The authors would like to thank Jeannette Nussbaum for her valuable contributions. *(28 total steps if Zymo needed)
Transcript of FFPE Quality: Impact on Downstream DNA Performance in Molecular Diagnostic Assays...
FFPE DNA Quality: Impact on Downstream Performance in Molecular Diagnostic AssaysFFPE DNA Quality: Impact on Downstream Performance in Molecular Diagnostic AssaysJonathan M. Eibl, Julie A. Toplin, Chad D. Galderisi, Cindy S. Spittle, MolecularMD Corp., Portland, OR
IntroductionPersonalized medicine for oncology patients is driven by mutation profiling of tumors.Improved robust DNA extraction methods that produce high quality FFPE DNA template
Poster #: ID102
Figure 3a. Maxwell vs. MO BIO DNA extraction from different FFPE sample types (total DNA Yield). Overall, DNA yields were greater for
l d h ll ( f l )
Figure 3b. Percent increase in DNA yield by Maxwell vs. MO BIO in different FFPE sample types (1.9‐fold or 190% increase overall).
Figure 6a. Box plot graphs of Phred scores for Sanger sequences (see Figure 5; forward traces; 51‐bp window; melanoma samples).
Figure 6b. Box plot graph of Phred scores for Sanger sequences (see Figure 5; reverse traces; 51‐bp window; melanoma samples). p p g q y p
suitable for complex molecular diagnostic testing are needed. In addition to DNAfragmentation and protein‐DNA cross linking, some FFPE tissues contain inhibitors and/orquenching substances (e.g. melanin) that adversely affect downstream performance in PCRand sequencing assays. In this study, we provide a side‐by‐side evaluation of two DNAisolation methods from FFPE tissue of various tumor types: a manual, column‐basedmethod (MO BIO) and the Maxwell CSC automated platform (Maxwell). 40
50
60
g)
Maxwell
MO BIO
samples processed on the Maxwell CSC (30 of 34 samples, 88%).
500
600
700
800
A yield
NO SEQ
UEN
CE
NO SEQ
UEN
CE
O SEQ
UEN
CE
O SEQ
UEN
CE
( ) p ( )
Materials & MethodsFFPE tissue from melanoma, GIST, endometrial, and NSCLC tumor types was obtained fromequivalent numbers of slides for each extraction method. Manual extractions wereperformed using the MO BIO BiOstic FFPE Tissue DNA Isolation Kit followed in some
10
20
30
DNA Yield (ug
100
200
300
400
500
% increase in
DNA N NNN
performed using the MO BIO BiOstic FFPE Tissue DNA Isolation Kit followed in someinstances by an additional purification step using the Zymo Research OneStep PCR InhibitorRemoval Kit to rid DNA isolates of visible melanin. Automated extractions were carried outusing the Maxwell CSC DNA FFPE Custom Kit on the Maxwell CSC instrument with noadditional PCR inhibitor removal steps. DNA eluate yields and optical densities weredetermined via spectrophotometry. Equivalent input DNA amounts from melanomasamples were assayed using both NRAS allele‐specific qPCR (25 ng) and NRAS Sanger
01 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
Melanoma GIST Endometrial NSCLC
01 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
Melanoma GIST Endometrial NSCLC
samples were assayed using both NRAS allele specific qPCR (25 ng) and NRAS Sangersequencing (200 ng) on the BIORAD CFX platform and Applied Biosystems 3730 DNAAnalyzer, respectively. NRAS qPCR data were reported as delta‐Ct values to determine qPCRdifferences. Sanger sequencing Phred scores were analyzed for the 51‐base pair windowsurrounding the NRAS codon 61 hotspot to quantitatively determine sequence‐trace quality.Quality metrics were further evaluated using the CodonCode Aligner and Sequence Scannersoftware packages 3
3.5
4OD 260/280 MaxwellOD 260/280 MO BIO 2.5
3OD 260/230 MaxwellOD 260/230 MO BIO
Figure 4a. Optical densities (260/280) Maxwell vs. MO BIO DNA extraction from different FFPE sample types.
Figure 4b. Optical densities (260/230) Maxwell vs. MO BIO DNA extraction from different FFPE sample types.
Figures 7a – 7e. Box plots representing sequence quality metrics using CodeCodon Aligner and Sequence Scanner software.
7a.
ResultsBoth extraction methods require the same upfront steps of scraping FFPE tissue, reagentaddition, and lysis/incubation, however, the Maxwell method requires fewer hands‐on,liq id handling steps (Fig 1) Additional Z mo clean p steps ere req ired for 2 man all
software packages.
0.5
1
1.5
2
2.5
OD (260
/280
)
0.5
1
1.5
2
OD (260
/230
)
liquid‐handling steps (Fig. 1). Additional Zymo cleanup steps were required for 2 manually‐extracted samples due to visible melanin in the FFPE tissue/eluted DNA (Fig. 2; Samples 02and 07). DNA yields were greater for samples processed on the Maxwell CSC (30 of 34 totalsamples, 88%; 1.9‐fold increase overall; Figs. 3a and 3b). Optical densities for both methodswere largely congruent for the 260/280 ratio (Fig. 4a) but more variable for the 260/230ratio (Fig. 4b). In qPCR, 11 of 12 (92%) of the Maxwell‐extracted melanoma samples crossedthe threshold 1 3 Ct sooner on a erage than the corresponding man all e tracted
01 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
Melanoma GIST Endometrial NSCLC
01 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
Melanoma GIST Endometrial NSCLC
Figure 5a NRAS allele‐specific PCR for DNA extracted by both methods Figure 5b Comparison of NRAS Sanger sequencing results for DNA extracted by
7b. 7c.
the threshold 1.3 Ct sooner on average than the corresponding manually‐extractedtemplates (Fig. 5a), and one manually‐extracted, Zymo‐purified sample failed to amplify(Sample 07). Two Maxwell‐extracted samples (Samples 10 and 11) crossed the thresholdmarkedly sooner (~2.5 Ct). Similarly, all Maxwell‐extracted melanoma DNA isolatesproduced robust bidirectional sequence, whereas two manually‐extracted isolates failed toyield NRAS sequence despite the inhibitor‐removal step (Fig. 5b). Aside from the sequencef il f th t l t d dditi l t l t hi h hibit d l t d b li
Sample Ct Value ∆ Ct Value (MO BIO – Maxwell)
MO BIO 01 33.041.11
Maxwell 01 31.93
MO BIO 02Maxwell 02
Figure 5a. NRAS allele‐specific PCR for DNA extracted by both methods from melanoma FFPE; traces shown at right for samples with ∆Ct Values ˃ 2.0 (Samples 02, 10, 11) and no amplification/amplification (Sample 07).
Figure 5b. Comparison of NRAS Sanger sequencing results for DNA extracted by both methods from two melanoma samples, both of which required the Zymocleanup step for the MO BIO extraction.
failures of these templates and one additional template which exhibited elevated baselinenoise, NRAS codon 61 mutation‐detection results for the remaining nine manually‐extractedeluates were identical to the Maxwell‐extracted samples (data not shown). Templates fromboth extraction methods exhibited higher median Phred scores and narrower interquartileranges (IQR) for reverse traces than for forward traces (Figs. 6a and 6b); this characteristic isattributed to the proximal location of the forward primer relative to the NRAS codon 61h t t S f M ll t t d FFPE DNA hibit hi h Ph d
MO BIO 02 35.422.17
Maxwell 02 33.25MO BIO 03 29.93
0.37Maxwell 03 29.56MO BIO 04 32.09
0.66Maxwell 04 31.43MO BIO 05 32.50
0 73
MO BIO 07Maxwell 07
Maxwell 02
Maxwell 07
FWD
F
REV
7d. 7e.
hotspot. Sequences from Maxwell‐extracted FFPE DNA exhibit higher mean Phred scoresand less variability (insets; Figs. 6a and 6b). Overall, metrics generated using other softwarepackages (i.e., quality values, trace scores, quality value 20+, continuous‐read length, andsignal‐to‐noise ratio) illustrate equivalent or better results for the Maxwell‐extractedtemplates especially in regards to variability (see IQRs for boxplots; Figs. 7a –7e).
Conclusions
0.73Maxwell 05 31.77MO BIO 06 34.64
1.94Maxwell 06 32.70MO BIO 07 No amplification
N/AMaxwell 07 35.07MO BIO 08 30.10
‐0.02Maxwell 08 30.12
MO BIO 10Maxwell 10
WD
REV
MO BIO 02FWD
This comparison highlights the potential advantages of using an automated nucleic acid extraction platform for onco‐diagnostictesting:
• Standardized workflow with fewer hands‐on steps minimizing the likelihood of cross‐contamination
• Higher throughput for faster turn‐around‐time of results
i h i ld i i i h i l d f li i d f i
Figure 1. Comparison of extraction method procedural steps and hands‐on time
Scrape FFPE Add reagents Lyse (incubate)
Steps common to both extraction methods
MO BIO 09 30.111.72
Maxwell 09 28.39MO BIO 10 30.93
2.60Maxwell 10 28.33MO BIO 11 30.80
2.55Maxwell 11 28.25MO BIO 12 28.44
0 05
MO BIO 11Maxwell 11
MO BIO 07FWD
REV
REV
• Higher DNA yields, maximizing the potential data output from limited amounts of FFPE tumor tissue
• Higher quality DNA without extra purification steps, minimizing the potential need for repeat analysisMaxwell: post‐lysis steps 1 – 12 (Hands‐on time for 16 samples ‐ ~30 min.)
Cool to RT Add RNase A
Incubate (5 min.)
Place cartridges
Remove seals
Place plungers
Place elution tubes
Add elution water
Spin FFPE digest (5 min.)
Add sample to cartridge
Enter sample
data / run
Remove elution tube
Figure 2. Representative specimen (Sample 02) requiring additional Zymo cleanup steps due to high melanin content (note red boxed area); H&E staining, 4X magnification.
0.05Maxwell 12 28.39
AcknowledgementsThe authors would like to thank Jeannette Nussbaum for her valuable contributions
For further information please contact [email protected] or visit www.molecularmd.com. MO BIO: post‐lysis steps 1 – 22 (Hands‐on time for 16 samples ‐ ~60 min.)
Heat (1 hr.)
Spin (1 min.) Transfer Add FP4 Vortex Add FP5 Vortex Spin Apply to
columnSpin
(1 min.)Transfer spin filter
Add FP6 (continue)
Transfer column
Spin (1 min.)
Add FP7 (1 of 2)
Spin (1 min.)
Add FP7 (2 of 2)
Spin (1 min.)
Spin (2 min.) Add FP8 Incubate
(5 min.)Spin
(1 min.)
ZYMO cleanup
(as needed)
*
The authors would like to thank Jeannette Nussbaum for her valuable contributions.
*(28 total steps if Zymo needed)
