ӳ����ý

Image courtesy of Bang Wong, ӳ����ý Communications

In cancer, cells divide relentlessly. Each time they split is another chance for DNA damage. Because cancer cells divide so often, their genomes become progressively more mutated. But not all of these genetic typos play an active role in causing and maintaining cancer — some are simply along for the ride. A large collaboration of researchers, led by those at the ӳ����ý of MIT and Harvard, Dana-Farber Cancer Institute, and elsewhere, has now completed a large-scale genomic effort in lung cancer to distinguish the important mutations in cancer from those that are not. The new work appears in the October 23 issue of Nature.

The aim of the Tumor Sequencing Project (TSP) is to use the precise sequence of As, Cs, Ts, and Gs in the genome of cancer cells to reveal biological pathways at work and suggest avenues for new therapies. Funded by the National Human Genome Research Institute (NHGRI), the TSP focuses on lung adenocarcinoma, the most common form of lung cancer. In the TSP’s first phase, researchers identified 50 genomic regions that are frequently gained or lost in human lung tumors. In the new work, the researchers searched for other kinds of genetic alterations, including so-called point mutations, which are single-letter substitutions of DNA or small insertions or deletions of letters in the genome of cancer cells.

To begin their search, the team gathered nearly 200 tumor samples from six medical centers across the United States. This is by far the largest cancer genome sequencing study performed to date on a single tumor type. The scientists studied the precise DNA sequence in each sample of more than 600 genes, drawing their list from those known or suspected to play a role in cancer and those in additional gene-families, such as kinases, that can be targeted by small molecules. Although most samples contained mutations, the researchers went a step further to determine which were active in the cancer, serving as so-called “driver” mutations. These DNA changes are the ones that likely give the cancer cells a survival advantage, such as being able to divide quickly or evade destruction by the immune system.

Gad Getz, a ӳ����ý computational biologist and co-first author on the new study, collaborated with fellow ӳ����ý computational biologist Mike Lawrence and scientists at Washington University to develop statistical methods of differentiating driver mutations from the background, or “passenger,” mutations. “In cancer, it seems like so many genes are mutated, that you’ll never know which ones are important,” said Getz. “But this method allows us to detect the important players.”

The method first estimates the background level of mutation in cancer cells. To do this, the team took advantage of the redundant nature of the genetic code: some mutations do not change how the DNA is made into protein, so they are unlikely to play a functional role in cancer and are considered part of the background. In turn, any gene found to be mutated at a rate higher than that of non-functional mutations is likely to be a driver gene.

More than 1,000 somatic mutations in hundreds of genes were found in the samples, but only 26 of these genes were mutated significantly above the background rate. Some of these genes were previously known to play a role in lung adenocarcinoma, but others were surprises. Some mutated genes, including NF1, APC, RB1, ATM, PTPRD, and LRP1B, are involved in other cancers, suggesting that a similar biological mechanism may be contributing to disease in many cases. The large sample size enabled the team to identify driver genes that are mutated at very low rates, occurring in as few as 5 of their samples.

The consortium also integrated different types of data — DNA sequence, gene expression, and locations of extra or missing DNA in the genome — to discern patterns. “Integrative approaches like these allow us to more clearly pinpoint important genes than a single method alone would,” said co-senior author Matthew Meyerson, a senior associate member of the ӳ����ý of MIT and Harvard and an associate professor at Dana-Farber Cancer Institute and Harvard Medical School. They found that certain genes seem to function together, forming a molecular pathway that can relay information within a cell. The pathways that are frequently impaired by mutations in lung adenocarcinoma include the MAPK, Wnt, p53, and mTOR pathways.

“One of the key findings from our study is that some of the newly discovered genes and pathways that are mutated in lung cancer are also known to be defective in other cancers,” said Meyerson. “That gives us hope that targeted therapies could be used across multiple cancer types.”

Getz explained that another realization to emerge from this study is that relatively few of the several hundred genes studied appear to play a role in lung adenocarcinoma. “Some people thought that cancer was hopelessly complicated,” said Getz. “But our results here suggest that the actual number of driver mutations in cancer may be lower than we expected, suggesting that it is a more feasible problem to tackle.”

These results, and the methods developed to generate them, will empower future studies of cancer genomics. Cancer researchers will soon rely upon “next-generation” sequencing technologies, which will greatly expand sequencing capabilities. “Our goal is to be able to sequence all 20,000 genes in all samples in a study,” said Getz, “and we think this will be feasible soon.” Until then, the team will continue with the next phase of the TSP, in which they’ll scale up to sequence several thousand genes, looking to extend even further the list of genetic changes known in lung adenocarcinoma.

The TSP is considered a pilot project for the larger NCI- and NHGRI-funded effort, The Cancer Genome Atlas (TCGA), which aims to characterize the molecular changes in several cancers using large-scale genomic technologies. Although the projects are being conducted concurrently, experimental methods and knowledge resulting from the TSP have helped to shape the ongoing TCGA work. “The TSP project, together with the TCGA, has laid the groundwork for running effective cancer sequencing studies,” said Getz.

Other ӳ����ý researchers participating in the work include Kristian Cibulskis, Carrie Sougnez, Heidi Greulich, Wendy Winckler, Jennifer Baldwin, Amit Dutt, Tim Fennell, Megan Hanna, Robert Onofrio, Barbara Weir, Xiaojun Zhao, Liuda Ziaugra, Michael Zody, Stacey Gabriel, and Eric Lander.