Global mutational profiling of formalin-fixed human colon cancers from a pathology archive

Abstract

The advent of Next-Generation sequencing technologies, which significantly increases the throughput and reduces the cost of large-scale sequencing efforts, provides an unprecedented opportunity for discovery of novel gene mutations in human cancers. However, it remains a challenge to apply Next-Generation technologies to DNA extracted from formalin-fixed paraffin-embedded cancer specimens. We describe here the successful development of a custom DNA capture method using Next-Generation for detection of 140 driver genes in five formalin-fixed paraffin-embedded human colon cancer samples using an improved extraction process to produce high-quality DNA. Isolated DNA was enriched for targeted exons and sequenced using the Illumina Next-Generation platform. An analytical pipeline using 3 software platforms to define single-nucleotide variants was used to evaluate the data output. Approximately 250 × average coverage was obtained with >96% of target bases having at least 30 sequence reads. Results were then compared with previously performed high-throughput Sanger sequencing. Using an algorithm of needing a positive call from all three callers to give a positive result, 98% of the verified Sanger sequencing somatic driver gene mutations were identified by our method with a specificity of 90%. In all, 13 insertions and deletions identified by Next-Generation were confirmed by Sanger sequencing. We also applied this technology to two components of a biphasic colon cancer, which had strikingly differing histology. Remarkably, no new driver gene mutation accumulation was identified in the more undifferentiated component. Applying this method to profiling of formalin-fixed paraffin-embedded colon cancer tissue samples yields equivalent sensitivity and specificity for mutation detection as Sanger sequencing of matched cell lines derived from these cancers. This method directly enables high-throughput comprehensive mutational profiling of colon cancer samples, and is easily extendable to enable targeted sequencing from formalin-fixed paraffin-embedded material for other tumor types.