The Cancer Genome Atlas (TCGA) began as a pilot study in 2006 to characterize brain tumors and ovarian cancer. It generated unprecedented amounts of data of unequaled quality, and in 2009, TCGA announced plans to produce comprehensive genomic maps of at least 20 types of cancer over the next 5 years. In this second installment, we review the genomic findings related to colorectal cancer (CRC) and the potential impact on new therapies.

Colorectal cancer is the third most common cancer and third most common cause of cancer death in both men and women in the United States.1 TCGA analyzed 224 CRC tumor and normal tissue pairs by DNA sequencing that integrated data from whole exome sequencing, DNA copy number variation, promoter methylation, and global mRNA and microRNA expression.2 CRC was the third of 20 cancer types that TCGA is genomically profiling.

TCGA analysis found that 85% of the 224 pairs had hypermutated tumors, while 15% were considered nonhypermutated. Tumors from the right colon were more likely to be hypermethylated and more commonly had hypermutation; 75% of the hypermutated samples were from the right colon. Not all of the hypermutated samples had microsatellite instability (MSI). The presence of MSI is often used to categorize CRCs.

Though patients with colon tumors are managed differently from those with rectal tumors, the nonhypermutated tumors were indistinguishable between the two sites regarding copy number, expression profile, DNA methylation, and miRNA changes. TCGA found that, at the genomic level, nonhypermutated adenocarcinomas of the colon and the rectum are the same.

A mutation in one or more members of the WNT signaling pathway was found in 93% of the colorectal cancer tumors examined by TCGA, with the mutation predominantly being in APC. WNT signaling transcriptionally represses SOX9, which TCGA found to be frequently mutated in CRC. Previously, mutations in SOX9 were unknown in any human cancer.

Since the TCGA’s analysis of CRC was published in the summer of 2012, global research has continued to make progress against this common and lethal cancer. Research teams have investigated the hypermutated subset of CRCs, explored mutations in mitochondria that alter metabolism in many cancers, and applied bioinformatics to group CRC tumors into clusters to suggest targeted therapies.


Colorectal cancers are usually subclassified based on whether or not they display microsatellite instability.3 Microsatellite DNA has simple tandem repeat sequences, and, when these sequences have genetic instability, it manifests as MSI. MSI is often associated with defective DNA mismatch repair. Several genes can be responsible for defective DNA mismatch repair, with one commonly involved gene being MLH1.

Patients with MSI-related cancers have better survival rates than those with microsatellite-stable cancers. CRC tumors with MSI are primarily located in the right colon and frequently associated with the CpG island methylator phenotype (CIMP) and hypermutation.

About 15% of CRCs have MSI, with about 75% of these cancers having hypermethylation and transcriptional silencing of the MLH1 gene. Many colorectal cancers with MSI have the oncogenic BRAF V600E mutation, and BRAF mutations are associated with increased mortality in sporadic CRCs.

The analysis by TCGA found that 85% of CRCs were nonhypermutated while 15% of CRCs were hypermutated, defined as having more than 12 nonsilent exonic mutations per megabase. The hypermutated CRCs had high rates of MLH1 gene silencing and high MSI. Lawrence A. Donehower, PhD, of Baylor College of Medicine in Houston, Texas, led an analysis of a hypermutated subset of CRC data from TCGA.3

The researchers found that identifying MLH1 expression levels in CRC is potentially useful for clinical diagnoses and therapeutic decisions. When the hypermutated CRCs (35 samples) were analyzed, the research team found that they could be usefully subcategorized into MLH1-silenced and non-MLH1-silenced subgroups.

The subgroup of MLH-1-silenced CRCs had a more definitive mutational status for APC, BRAF, and KRAS than CRCs with high MSI. In the MLH-1-silenced CRCs, 68% had BRAF mutations, making them candidates for treatment with BRAF and MEK inhibitors while suggesting that anti-EGFR molecules like cetuximab (Erbitux) would be poor choices. The authors suggested that examining MLH-1 expression status could be a useful diagnostic tool that could supplement MSI assays and influence therapeutic decisions.


Paired tumor and normal tissue samples from 226 persons were used to investigate a subset of somatic mutations that allow cancer cells to shift their metabolism.4 Somatic mutations are mutations that are not passed on to children. The cancer types found in the group of samples were colon adenocarcinoma, rectal adenocarcinoma, acute myeloid leukemia, glioblastoma, and ovarian serous cystadenocarcinoma.

The researchers, led by Jonathan G. Seidman, PhD, of Harvard Medical School in Boston, Massachusetts, explained that tumor growth seems to be enhanced when tumors acquire somatic mitochondrial mutations that impact oxidative phosphorylation. Oxidative phosphorylation, the usual mechanism human metabolism uses to produce energy, is more efficient than glycolysis.

However, when tumors shift to glycolysis, they acquire a selective advantage that aids their growth, invasion, metastasis, and ability to avoid cellular senescence. These metabolic processes occur in the mitochondria of cells. The authors stated that this level of mitochondrial mutations was remarkable, since the mitochondrial genome is subjected to purifying selection.

The most frequent occurrence of high-impact mitochondrial mutations was in colon and rectal tumors, as 26% to 28% of the samples had deleterious mitochondrial DNA mutations. This data suggests that altered metabolism is a critical hallmark for many cancers, particularly in colon and rectal adenocarcinomas. The authors predicted that the mitochondrial mutations inhibit oxidative phosphorylation and provide selective growth advantages early in oncogenesis.

Additional studies in this area are suggested to characterize the molecular pathways activated by mitochondrial mutations, define the clinical use of mitochondrial mutations as tumor biomarkers, and address the therapeutic potential for targeting mutation-induced metabolic dysregulation in cancer.


An enormous volume of data has been generated by TCGA. A research team, led by Ronglai Shen, PhD, of Memorial Sloan-Kettering Cancer Center in New York, New York, applied the bioinformatics method iCLUSTER+ to integrate the full spectrum of genomic data from TCGA’s CRC dataset.5

The method revealed subgroups that are not lineage-dependent, but instead consist of different cancer types driven by a common genetic alteration. Their analysis sought to identify distinct molecular drivers that can reveal tumor subgroups of biological and clinical importance. The researchers applied their method to a subset of CRC data from TCGA with available data on MSI status, somatic copy number alterations, CIMP classification, and gene expression profiles.

When a subset of 189 CRC tumors from TCGA analysis was classified by iCLUSTER+, the tumors were grouped into three clusters. One of these clusters included the hypermutated CRCs, which was further subdivided into two clusters based on the degree of chromosomal instability.

The CRCs with the lowest chromosomal instability, considered negative for chromosomal instability, had the fewest high-stage tumors of all the subtypes of CRC. The researchers suggested that this subgroup of tumors negative for chromosomal instability may not need aggressive treatment outside of surgery.

The subgroups with high chromosomal instability were clearly separated into two additional subtypes: one having amplified levels of the genomic loci chr8q, and the other having normal levels of chr8q. Tumors with amplified chr8q had evidence suggesting expression of five genes was induced. These subgroups may be the starting point for developing targeted therapies.

Kathy Boltz is a medical writer based in Phoenix, Arizona. 


1. Siegel R, Naishadham D, Jemal A. Cancer statistics, 2012. CA Cancer J Clin. 2012;62(1):10-29.

2. Cancer Genome Atlas Network. Comprehensive molecular characterization of human colon and rectal cancer. Nature. 2012;487(7407):330-337.

3. Donehower LA, Creighton CJ, Schultz N, et al. MLH1-silenced and non-silenced subgroups of hypermutated colorectal carcinomas have distinct mutational landscapes. J Pathol. 2013;229(1):99-110.

4. Larman TC, DePalma SR, Hadjipanayis AG; Cancer Genome Atlas Research Network, et al. Spectrum of somatic mitochondrial mutations in five cancers. Proc Natl Acad Sci U S A. 2012;109(35):14087-14091.

5. Mo Q, Wang S, Seshan VE, et al. Pattern discovery and cancer gene identification in integrated cancer genomic data. Proc Natl Acad Sci U S A. 2013;110(11):4245-4250.