High-throughput detection of clinically relevant mutations in archived tumor samples by multiplexed PCR and next-generation sequencing.
Author(s): Bourgon R, Lu S, Yan Y, Lackner MR, Wang W, Weigman V, Wang D, Guan Y, Ryner L, Koeppen H, Patel R, Hampton GM, Amler LC, Wang Y
Publication: Clin Cancer Res, 2014, Vol. 20, Page 2080-91
PubMed ID: 24573554 PubMed Review Paper? No
Purpose of Paper
The purpose of this paper was to determine if formalin-fixed paraffin-embedded (FFPE) specimens and frozen specimens can produce comparable next-generation sequencing (NGS) data for 88 clinically relevant cancer genes and to determine the effect of DNA copy-number and uracil-DNA glycosylase (UDG) treatment. In addition to matched specimens, the authors compared data generated using FFPE specimens to that in the Cancer Genome Atlas (TCGA).
Conclusion of Paper
FFPE specimens were shown to have 200-5800 amplifiable copies in 150 ng/DNA. The amplifiable copy count was predictive of the concordance between NGS data from frozen and FFPE specimens, with high concordance observed when the amplifiable copy number was high (5000 copies); however, there were lots of false positives for sequence variants in the FFPE specimens when the amplifiable copy number was low (<2500/copies). The major source of these false positives was deamination and could be decreased by treatment with UDG. For 116 hotspots in 11 genes, it was possible to accurately determine the genotype from FFPE specimens with low copy number even though the variant frequency differed between frozen and FFPE specimens. NGS was more sensitive than allele-specific real-time qPCR. Importantly, the mutation rates observed in 73 FFPE endometrial cancer specimens matched those reported in TCGA using 248 frozen specimens. In addition, phosphatase and tensin homolog (PTEN) levels measured by immunohistochemistry (IHC) were lower in specimens with mutations expected to lead to truncated proteins.
Studies
-
Study Purpose
The purpose of this study was to determine the effects of formalin-fixation, amplifiable copy number, UDG treatment, and using real-time PCR or IHC instead of NGS on the identification of mutations in 88 clinically relevant cancer genes. The study included matched FFPE and frozen specimens of 4 breast cancers, 4 lung cancers, 4 colon cancer and 5 ovarian cancers and FFPE specimens from an additional 73 endometrial cancers. DNA was extracted from FFPE specimens using the QIAamp FFPE kit and from frozen specimens using the QIAamp kit. For the 15 tumor suppressor genes and PIK3CA, amplicons covered all exons and splicing junctions, but for the remaining 72 cancer genes, only the hot-spots were amplified and sequenced. All amplicons were <200 bp. Amplification success of a 149 bp fragment of trafficking protein, kinesin binding 2 (TRAK2) was used to get an estimate of DNA copy number.
Summary of Findings:
The amplification assay showed FFPE specimens had 200-5800 amplifiable copies in 150 ng/DNA. The amplifiable copy count was predictive of the concordance between frozen and FFPE specimens, with high concordance observed when the amplifiable copy number was high (5000 copies); however, there were lots of false positives for sequence variants in the FFPE specimens when the copy number was low (<2500/copies). The major source of these false positives was deamination (C>T or G>A substitutions), and these increased when the amplifiable copy number decreased below 3000 copies. Treatment with UDG to eliminate uracil increased the percentage of true positives by 10-30%, but it did not affect sensitivity. NGS specificity and sensitivity for FFPE specimens increased with increasing amplifiable copy number, and the authors state that close to 100% specificity was achievable from high copy number specimens when a stringent variant reads cut-off was applied. Further, for 116 hotspots in 11 genes, it was possible to accurately determine the genotype, even from FFPE specimens with low amplifiable copy number, though the variant frequency differed between frozen and FFPE specimens. NGS was more sensitive than allele-specific real-time qPCR. Importantly, the mutation rates observed in 73 FFPE endometrial cancer specimens matched those reported in the Cancer Genome Atlas (TCGA) using 248 frozen specimens. 59% of the mutations in PTEN were nonsense and expected to lead to incomplete proteins. As many of these were novel mutations, the authors examined PTEN levels by IHC and found lower levels in specimens with these mutations. Many other expected and novel mutations in cancer genes were identified, including many in the phosphoinositide 3-kinase (PI3K) and rat sarcoma (RAS) pathways.
Biospecimens
Preservative Types
- Frozen
- Formalin
Diagnoses:
- Neoplastic - Carcinoma
Platform:
Analyte Technology Platform DNA Next generation sequencing DNA Real-time qPCR Protein Immunohistochemistry DNA MS/MS Pre-analytical Factors:
Classification Pre-analytical Factor Value(s) Biospecimen Preservation Type of fixation/preservation Formalin (buffered)
Frozen
Immunohistochemistry Specific Targeted peptide/protein PTEN
Next generation sequencing Specific Template/input amount 200-5800 amplifiable copies
Next generation sequencing Specific Template modification UDG treated
Untreated
Next generation sequencing Specific Technology platform Allele specific real-time PCR
IHC