Skip to main content
Premium Trial:

Request an Annual Quote

TCGA Researchers Report on Colorectal Cancer Analyses

NEW YORK (GenomeWeb News) – In a study appearing online today in Nature, members of The Cancer Genome Atlas network described the genomic patterns that they have uncovered in colorectal carcinomas, including samples originating at sites in either the colon or the rectum, revealing the similar genomic profiles that are generally present in tumors at each site..

By doing sequencing, copy number analyses, and/or methylation profiling on almost 300 colorectal carcinoma samples, the team narrowed in on key genes and pathways that tend to be altered in colorectal carcinoma, providing clues about possible avenues for treating the disease.

"[T]he data presented here provide a useful resource for understanding this deadly disease and identifying possibilities for treating it in a targeted way," corresponding author Raju Kucherlapati, a genetics researcher affiliated with Harvard Medical School and Brigham and Women's Hospital, and his co-authors explained.

"While it may take years to translate this foundational genetic data on colorectal cancers into new therapeutic strategies and surveillance methods, this genetic information unquestionably will be the springboard for determining what will be useful clinically against colorectal cancers," National Cancer Institute Director Harold Varmus, who was not directly involved in the study, said in a statement.

A subset of the colorectal tumors — most often those turning up in the right or ascending colon — had unusually high mutation levels. And more than three-quarters of these hypermutated samples showed enhanced methylation levels and microsatellite instability.

For the most part, though, the researchers found that colorectal carcinomas carried the same sorts of genetic glitches when they originated in colon as when they appeared in rectal tissue.

"[W]e found, in the non-hypermutated tumors irrespective of their anatomical origin, the same copy number, expression profile, DNA methylation, and miRNA changes," the study's authors wrote.

For the new analysis, researchers used SOLiD or Illumina sequencing platforms to sequence the exomes of 224 tumor-normal pairs to a an average depth of more than 20 times over 80 percent or more of the coding sequences targeted. With the Illumina HiSeq 2000, they also did low coverage whole-genome sequencing on 97 of the tumors and matched normal samples.

In addition, the team assessed copy number, DNA methylation, and gene expression patterns in the tumors using Affymetrix, Illumina, and Agilent arrays, respectively. The transcriptional analysis was fleshed out further through RNA sequencing and microRNA sequencing experiments.

When they brought together the various sources of genomic data together, TCGA investigators found that around 16 percent of the tumors could be classified as hypermutated, containing a median of 728 predicted somatic mutations apiece.

Along with differences in the extent of mutation and the identity of the mutated genes, the hypermutated and non-hypermutated tumors tended to contain distinct methylation and microsatellite instability patterns.

In particular, many of the most highly mutated tumors showed a corresponding jump in methylation levels. And some 77 percent of the hypermutated tumors showed microsatellite instability, a feature often found in colorectal carcinomas originating in the right colon that often corresponds to better colorectal cancer outcomes.

In the current study, researchers found that tumors from the right colon represented the majority of the hypermutated samples. But not all of the highly mutated tumors originating at that site had microsatellite stability, they noted, pointing to the possibility that mutation rate might make a more apt prognostic marker than microsatellite instability.

"As the survival rate of patients with high [microsatellite instability]-related cancers is better and these cancers are hypermutated," they wrote, "mutation rate may be a better prognostic indicator."

Within the hypermutated tumors, the team tracked down 15 genes with recurrent somatic mutations, including ACVR2A, which was mutated in 63 percent of these tumors, as well as APC and TGFBR2, each altered in 51 percent of tumors from the hypermutated group.

In the non-hypermutated tumors, meanwhile, researchers found recurrent mutations in known colorectal cancer risk genes, such as APC, TP53, PIK3CA, KRAS, and SMAD4, and several new genes, including SOX9, ARID1A, and FAM123B.

The team's array-based copy number analyses on 257 tumors revealed a range of recurrent deletions and amplifications, including amplifications affecting the ERBB2 and IGF2 genes.

From their genome sequence data, meanwhile, researchers tracked down several suspected translocation events involving bits of sequence from different chromosomes. For instance, three of the 97 tumors assessed by low-coverage genome sequencing contained a fusion linking the first exons of the chromosome 11 gene NAV2 to part of the chromosome 2 gene TCF7L1, which codes for a component in the WNT pathway, a known contributor to colorectal cancer.

Indeed, almost all of the tumors from both the hypermutated and the non-hypermutated groups included mutations expected to boost the activity of the WNT signaling pathway and to curb signaling via the TGF-beta pathway — changes that are expected to produce a jump in the activity of the MYC proto-oncogene.

Such findings fit with early reports suggesting that compounds targeting that pathway may be effective against some colon cancers, researchers noted.

"Our data suggest a number of therapeutic approaches to [colorectal cancer]," they wrote. "Included are WNT-signaling inhibitors and small-molecule [beta]-catenin inhibitors, which are showing initial promise."

Other commonly affected pathways included the RTK-RAS, MAP kinase, TP53, and PI3 kinase pathways, again pointing to potential targets for those trying to come up with new colorectal cancer treatments.