Posted: May 25, 2011
CASE STUDY: Screening Cancer Genomes for Essential Drivers
“Our goal isn’t simply to understand cancer,” said Bill Hahn, M.D., Ph.D., Associate Professor of Medicine at the Dana-Farber Cancer Institute, “but to tackle it head-on in patients.” The Cancer Genome Atlas (TCGA) provides the crucial first step by doing structural analysis; in other words, it analyses the genome for each of 20 cancer types and identifies the genetic changes. “But you really need a three-legged stool for productive cancer research.”
The second leg of the stool, he explained, requires “functional studies to see what the gene changes revealed by TCGA actually do in real cells and tumors.” Only then can you work on the third leg by developing drugs that will impact the molecules that the leg one and two research identified as the important targets.
Working at the Broad Institute, Harvard Medical School and the Dana-Farber Cancer Institute, Dr. Hahn’s lab is fortifying the second leg of the stool with an important functional study of ovarian cancer. As the ovarian “page” of the atlas is finalized by the TCGA Research Network, functional studies are the crucial next step in the effort to treat and perhaps one day to cure ovarian cancer.
Ovarian is the most lethal of all reproductive system cancers and presents a real challenge, in part because it is not easy to detect in its early stages when it is very treatable. Instead, explained Dr. Hahn, “despite advances in surgery and chemotherapy, ovarian cancer in the majority of women will return and become resistant to further treatments.” This makes it a prime cancer for TCGA to target.
Dr. Hahn refers to TCGA results as a parts list of all the genetic features that are mutated or altered in each ovarian cancer. This is a prerequisite, he explains. “Without it, you’re trying to go someplace you’ve never been without a map. You don’t know how to get there because you don’t recognize the landmarks.”
The TCGA’s “Parts List”
Dr. Hahn and colleagues at Broad and the Whitehead Institute for Biomedical Research have pioneered a new way to find and describe genes that are essential for cancer to develop and progress, detailed in a landmark paper published in 20081.
It is important to realize that TCGA’s more comprehensive “parts list” does not by itself sort out which of the altered ovarian cancer genes are truly essential in the cancer’s development. In fact, TCGA researchers are finding that significant gene mutations occur in only small numbers of the tumors, usually less than one in 20. Much more promising was the finding that about 1,200 genes are amplified––too many copies of them are being created––in the ovarian cancer samples that were analyzed, compared to the normal ovarian tissue.
Since many of these amplified genes cluster in certain parts of the chromosomes, you now have a map and some landmarks, according to Dr. Hahn’s description of the TCGA first leg. “But not all of those genes can be involved in ovarian cancer,” he explains.
“In every tumor, thousands of genes will be mutated or altered, but many of these changes happen when a tumor is evolving and the genome is unstable. Many fewer genes are actually essential for the tumor to become a tumor, what we call drivers.” Driver mutations must be present for cancer to start or to spread. Consequently, he said, many hundreds of the 1,200 genes on the TCGA ovarian parts list of altered genes are surely passengers, not causing cancer but merely along for the ride.
Adding the Functional Perspective
Dr. Hahn is applying the method pioneered at Whitehead and Broad to focus on the essential driver genes for ovarian cancer. The method involves screening cancer cell lines that have been developed for research purposes, using a method that turns off certain genes (short hairpin RNA interference).
In order to find those mutations worth targeting with drug treatments, he believes, “you have to interrogate their function––see what they are actually doing in the patient’s cancer. And it’s terribly expensive and inefficient to think about doing functional experiments for each of the 1,200 genes that TCGA found to be amplified in the ovarian tumors.”
Instead, his lab took a broad, genome-wide view and screened 25 known ovarian cancer cell lines to see which genes were not working normally. They then crossed this fairly large list with the chromosomal regions identified by TCGA that contain too many copies of certain genes.
This is “one way that functional studies bring new insight to the structural TCGA results,” he explained. “We were able to narrow things down to about 65 genes––these were not only amplified, but also essential for ovarian cancer to grow.” This left them with a refined parts list of possible targets much more likely to contain driver genes because researchers found them from separate research pathways.
“Numbers really do matter,” Dr. Hahn emphasized. Compared to traditional work done in one or a few cell lines, these large-scale approaches tend to produce more compelling results. By using hundreds of carefully processed tumor and healthy (normal) samples, TCGA has established the “gold standard” for biospecimens, and Dr. Hahn’s functional interrogation of the full genome in scores of cell lines further enhances the statistical power of these conclusions.
Doing the Heavy Lifting First
Dr. Hahn sees an important lesson for cancer research in this kind of complementary work. While it is true that many of science’s greatest discoveries have come from individuals toiling away at an insoluble problem, “big science” is now a reality with great potential.
“The paradigm of single labs doing single experiments can be terribly inefficient at times,” he said, pointing to the 1,200 or so genes amplified in the TCGA ovarian atlas. “Clearly most of these 1200 genes are passengers and we know that human cancers harbor hundreds of genetic mutations that don’t really have a significant impact.”
But with parallel functional studies like his, we have “a feasible strategy to systematically identify the key genes involved in cancer initiation, maintenance, and progression. We can then refer the likely targets for therapeutic intervention to people working on the third leg,” he says.
1Luo, B., Cheung, H.W., Subramanian, A., Sharifnia, T., Okamoto, M., Yang, X., Hinkle, G., Boehm, J.S., Beroukhim, R., Weir, B.A., et al. (2008). Highly parallel identification of essential genes in cancer cells. Proc Natl Acad Sci USA. 105 (51): 20380-20385. Read the full article.