There is no significant biological difference between colon and rectal tumors and nearly all of those tumors, have deregulation of the MYC activity, according to new results from The Cancer Genome Atlas (TCGA) colon cancer group. The research, reported in the July 10 issue of Nature (2012;487:330–337), also reveals potential new therapeutic targets, including Erbb2 (HER2) and insulin-like growth factor 2 (IGF2).
“For a tumor type for which there wasn't really much choice about therapies or clinical trials, these results open up a whole new set of possibilities for therapeutic interventions,” said the corresponding author, Raju Kucherlapati, PhD, the Paul C. Cabot Professor in the Harvard Medical School Department of Genetics and Professor in the Department of Medicine at Brigham and Women's Hospital.
The TCGA team analyzed 276 colorectal tumor samples using a wide variety of genomic techniques, including exome sequencing, DNA copy number analysis, promoter methylation assays, and messenger RNA and microRNA expression tests. Additionally the team performed whole genome sequencing on 97 samples.
“This is a huge contribution,” said Raymond Dubois, MD, PhD, Provost and Executive Vice President of The University of Texas MD Anderson Cancer Center and a colorectal cancer specialist who was not involved in the work. “It will definitely help move things forward. They rounded up the usual suspects we've known were involved in colon cancer, so that was good to validate those in a larger data set. They did find some other things that I thought were quite interesting.”
Top on the list of interesting discoveries, he said, was the fact that 93 percent of all the tumors had an alteration in the WNT--catenin pathway, including changes in the APC gene. “Thus far, it's been difficult to target the pathway, but clearly it has become a much more important target as a result of the data they collected,” Dubois told OT.
The analysis also revealed that there is little or no biological difference between colon and rectal tumors, despite the community's history of treating them differently.
“There are clearly oncologists who thought they were biologically different,” Kucherlapati said. “That is part of the reason we undertook this study—to assess whether they really are different. And when we actually do very careful analysis, we find there are no significant difference between the tumors. Biologically what this says is that these two tumors have the same biological origins.”
One interesting near commonality in all of the tumors (97%) is the deregulation of MYC activity. “MYC was known to be implicated,” he explained. “But the degree of MYC's role in this tumor type has not been appreciated before. What that means is that any drugs that might be able to act as synthetic lethal or change this strange deregulation of the pathway might be effective.”
Other important highlights from the work include the observations that:
- ERBB2 (also known as HER2/neu) is altered in 13 percent of non-hypermutated tumors and 53 percent of hypermutated ones;
- KRAS, NRAS, or BRAF were altered in a mutually exclusive pattern in 55 percent of non-hypermutated tumors;
- Approximately 22 percent of the tumors showed either amplification or overexpression of the IGF2 gene, which had not been implicated in colorectal cancer previously; and
- About one-third of tumors carried simultaneous alterations in the RAS pathway and PI3K pathway (including in IGF2).
All of those observations have implications for clinical trials, Kucherlapati noted. Patients whose tumors carry ERBB2 alterations may respond to anti-HER2 therapy, such as trastuzumab or lapatinib. Additionally, recently published data suggest that combinations of EGFR and BRAF inhibitors might be effective in patients with BRAF-mutated tumors.
Also, numerous IGF2 inhibitors are in development, and the discovery of IGF2 alterations in colorectal tumors suggests that approximately one-fifth of colorectal cancer patients might be eligible for those trials.
Finally, the observation that a substantial proportion of tumors carry simultaneous alterations in the RAS and PI3K pathways means that researchers need to think about combination therapies that inhibit both signaling cascades, as blocking only one of them is unlikely to control tumor growth.
Hypermutation from Two Sources
Overall 16 percent of the tumors tested were hypermutated. Kucherlapati says that hypermutation is most often thought to result from deficiencies in DNA mismatch repair and that microsatellite instability is used as a surrogate for the phenotype. However, nearly a quarter of the hypermutated tumors showed no evidence of microsatellite instability. Yet they carried mutations in DNA mismatch repair genes or the DNA polymerase (POLE). POLE, which is involved in both DNA replication and proofreading, has not been associated with colorectal cancer before.
The data may have implications for tumor evaluation and treatment. Hypermutated tumors don't respond as well to chemotherapy but are often associated with better clinical outcomes, he said. However, current tests use microsatellite instability as a surrogate for hypermutation. Based on the TCGA results though, the current methods would miss nearly one-fourth of the hypermutated tumors.
“I think we need new methods for detecting hypermutation,” he said. “Microsatellite instability should not be used alone. Directly sequencing a bunch of genes would be a better measure.”
Kucherlapati said that the colorectal cancer team used a wider variety of assays than the TCGA teams used in the pilot studies covering ovarian cancer and glioblastoma. Moreover, the team included more tools to integrate the data across platforms, tying together different types of genetic information to provide a more complete picture of how the tumors are genetically rewired.
“When you start these studies you don't know whether you are going to find new things or confirm things other people have found. But this study really revealed that there are a number of new things. That says that our knowledge about the cancers, even our genetic knowledge, is by no means complete, and that studies of this nature—and the whole reason for setting up studies like TCGA—are going to yield very valuable information about the tumor biology and tumor therapy.”
From Dubois' perspective, he said, the one key piece of data missing from the current analysis is information about clinical outcome: “It would have been wonderful to look at outcomes and correlate it with the fingerprints, but that is a minor weakness. Out of this, I think we could come up with a panel of things we should measure in our patients and try to correlate those with outcomes. If we could get this data correlated with clinical outcomes, that would be more helpful.”
Even so, he notes, there is so much information in the TCGA paper that it will take a while to digest it all.
Statements from Francis Collins, Harold Varmus, & Eric Green
The National Institutes of Health, National Cancer Institute, and National Human Genome Research Institute Directors all released statements about the significance of the colorectal cancer work:
- NIH Director Francis S. Collins, MD, PhD: “This finding of the true genetic nature of colon and rectal cancers is an important achievement in our quest to understand the foundations of this disease. The data and knowledge gained here have the potential to change the way we diagnose and treat certain cancers.”
- NCI Director Harold E. Varmus, MD: “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.”
- NHGRI Director Eric D. Green, MD, PhD: “It takes a critical group of researchers to conduct research at this scale and of this quality. This study is among the most comprehensive of its kind to date and vividly illustrates how TCGA data sets can shed new light on fundamental properties of human cancers.”