Results validate ambitious NIH-funded project to uncover cancer mutations

Lynda Chin, MD

An array of broken, missing, and overactive genes — some implicated for the first time — have been identified in a genetic survey of glioblastoma, the most common and deadly form of adult brain cancer, report scientists from Dana-Farber Cancer Institute and the Broad Institute of MIT and Harvard, together with their collaborating investigators at 18 institutions and organizations.

The large-scale combing of the brain cancer genome confirms the key roles of some previously known mutated genes and implicates a variety of other genetic changes that may be targets for future therapies.

The findings, posted online by Nature on Thursday, Sept. 4, help solidify and expand the "parts list" of genetic flaws linked to glioblastoma multiforme (GBM). Lynda Chin, MD, at Dana-Farber and Harvard Medical School (HMS) and Matthew Meyerson, MD, PhD, at Dana-Farber, HMS, and Broad, co-led the writing effort for the first summary of data from the $100 million pilot project of The Cancer Genome Atlas (TCGA), funded by the National Institutes of Health (NIH). The data are released to the public at TCGA's website as they are generated.

Systematic multi-dimensional genomic studies of patient samples of glioblastoma began in 2006 as the first TCGA program. The pilot is designed to determine the feasibility of a full-scale effort to systematically explore the universe of genomic changes involved in all types of human cancer and to demonstrate the values of such efforts in advancing cancer research and improving patient care.

Matthew Meyerson, MD, PhD

The current report in Nature summarizes the interim analyses of data gathered in the GBM pilot study. "The findings of significant mutations in genes that have implications for therapeutic development illustrate precisely how unbiased and systematic cancer genome analyses can lead to paradigm-shifting discoveries," said Chin, who chairs the GBM disease working group within TCGA.

An exciting example, Chin said, is an unanticipated observation of a link between DNA methylation of specific genes and DNA repair defects, leading to a hypothesis about a potential mechanism of resistance to a common chemotherapy drug used for brain cancer.

The Nature paper complements a parallel study by Johns Hopkins researchers of 22 GBM tumors, which was also published today in the journal Science.

"These data show that this approach, of looking at large numbers of tumors and a large number of genetic factors, can be done and the results are really valuable," said Meyerson. "We have made significant novel findings, and the reproducibility of the data is high."

Collaborating teams analyzed 206 specimens of glioblastoma tissue donated by patients at four medical centers. Their approach was "multidimensional" — looking for several categories of flaws simultaneously.

These included mutations — "typos" in the DNA code of a gene that alters its function; too many or too few copies of a given gene; damage to chromosomes causing loss or dislocation of pieces; gene activity that is higher or lower than normal; and changes in DNA methylation — turning genes on or off without affecting their structure.

The researchers also had access to information on how the patients who donated the samples had fared, including how they responded to certain drugs.

Automated machines at three Genome Sequencing Centers, including the Broad Institute center led by Eric S. Lander, Broad Institute director, were set to work reading the DNA messages in the cancer cells' nuclei. Of the roughly 20,000 protein-coding genes in the tumor cells, 601 genes were selected by the GBM disease working group for detailed sequencing — determining the order of chemical "letters" in the DNA — and comparison. A second installment of genes is already being sequenced, and Chin and her group are working on additional gene lists for mutational analyses.

Five major gene mutations have previously been identified in glioblastoma cells; the new sequencing effort revealed three that hadn't been discovered. One mutation affects the NF1 gene, which causes neurofibromatosis. A second mutation is in the ERBB2 gene known to be involved in breast cancer. The third affects a gene in the PIK3 signaling pathway that is abnormally activated in a number of cancers, but this particular gene, PIK3R1, had been only rarely implicated in any cancer. "Each of these mutated genes defines a new target for glioblastoma treatment," said Meyerson.

As they examined the data, the researchers found that three signaling pathways — networks of genes and proteins that act together to carry out a cellular function — were disrupted in more than three-quarters of the GBM tumors. They are known as the cyclin-dependent kinase/retinoblastoma pathway that regulates cell division; the p53 tumor suppressor pathway, which is involved in response to DNA damage and cell death; and the receptor tyrosine kinase pathway that carries signals that control cell growth.

Chin said that the most exciting finding is that this multipronged study design also enabled the scientists to make a potentially important connection between a methylation change in the glioblastoma cells and which drugs should be used for treatment. Brain tumors that contain a methylated, or silenced, form of a gene known as MGMT are known to be more susceptible to cancer drug temozolomide (Temodar). Therefore, Temodar is routinely given along with radiation to patients with MGMT methylation.

But the analysis of methylation in the glioblastoma tumors, when matched with the patients' medical history, revealed a cautionary sign. When such patients were treated with Temodar and subsequently had a recurrence of the tumor, it was very likely to become resistant to treatment because of "hypermutation" — an increased rate of gene changes that led to the tumor's ability to evade the drugs.

"This could have immediate clinical applications," said Chin.

The discoveries in the paper are only the tip of an expected iceberg, said the authors. The "most powerful impact" is expected to come from further research studies carried out by scientists who make use of the data released freely by TCGA, they said.

More than 21,000 new cases of brain cancer are expected to be diagnosed in the United States this year, and more than 13,000 people are likely to die from the disease.

"These impressive results from TCGA provide the most comprehensive view to date of the complicated genomic landscape of this deadly cancer," said NIH Director Elias A. Zerhouni, M.D. "The more we learn about the molecular basis of glioblastoma multiforme, the more swiftly we can develop better ways of helping patients with this terrible disease. Clearly, we should move ahead and apply the power of large-scale, genomic research to many other types of cancer."

Chin is co-principal investigator of a TCGA center with Raju Kucherlapati of HMS, and Meyerson is principal investigator of a TCGA Cancer Genome Characterization Center at the Broad Institute. Chin is the scientific director of the Belfer Cancer Genomics Center in the Center for Applied Cancer Science at Dana-Farber, and Meyerson directs the Center for Cancer Genome Discovery at Dana-Farber. The research was funded by grants from the NIH.

Dana-Farber Cancer Institute (www.dana-farber.org) is a principal teaching affiliate of the Harvard Medical School and is among the leading cancer research and care centers in the United States. It is a founding member of the Dana-Farber/Harvard Cancer Center (DF/HCC), designated a comprehensive cancer center by the National Cancer Institute.

The Broad Institute of MIT and Harvard was founded in 2003 to bring the power of genomics to biomedicine. It pursues this mission by empowering creative scientists to construct new and robust tools for genomic medicine, to make them accessible to the global scientific community, and to apply them to the understanding and treatment of disease. The Institute is a research collaboration that involves faculty, professional staff and students from throughout the MIT and Harvard academic and medical communities. It is jointly governed by the two universities. Organized around Scientific Programs and Scientific Platforms, the unique structure of the Broad Institute enables scientists to collaborate on transformative projects across many scientific and medical disciplines. For further information about the Broad Institute, go to http://www.broad.mit.edu.