Genome Scans for Cancer: What's Next?
Scanning the human genome for genetic variants involved in common cancers began to pay dividends in 2007, and the trend is likely to continue as more large studies involving new types of cancer report their results in the coming year.
For the first time, researchers have been discovering and validating genetic variants associated with common cancers such as breast, colon, and prostate. The genetics of these diseases have been exceedingly difficult to dissect, but that is starting to change, thanks to new technologies and the use of large and carefully selected patient populations.
"We're starting to unlock the genetic secrets of common, complex diseases," said Dr. Teri Manolio of the National Human Genome Research Institute, who organized a seminar on genome-wide association studies last summer. "The big lesson in cancer is that you find strong signals in parts of the genome where there aren't any known genes."
Indeed, much of the excitement surrounding genome scans comes from discoveries in places where no one had thought to look, such as a region of chromosome 8 called 8q24. The region has no known genes. But when three prostate-cancer scans published on the same day last year pointed to 8q24, it became a research priority almost overnight.
"The current excitement in the field is about what the scans will tell us about the biology of cancer," said Dr. Thomas Sellers of the Moffitt Cancer Center, who is collaborating on genome scans for ovarian cancer. "It's hard for patients to get excited about that, but this is important for progress in research and developing treatments."
Scans are underway, for instance, to find variants underlying individual differences in cancer risk, drug response, the risk of relapse, and second cancers. And as technology improves and the basic strategy is refined, scans will likely be used in new ways.
In addition to breast, colon, and prostate cancers, scans were reported last year for acute lymphoblastic leukemia. The studies largely involved patients of European descent and will need to be repeated in African and Asian populations. Scans are underway or nearing completion for lung, pancreatic, ovarian, and bladder cancers, to name a few.
A substantial number of these studies will be published for additional types of cancers in 2008, predicted Dr. David Hunter of the Harvard School of Public Health and a leader of NCI's Cancer Genetic Markers of Susceptibility (CGEMS) program. "One can expect that new, previously undiscovered associations will be found, and that should help us learn more about the inherited causes of these cancers."
For prostate, colon, and breast cancers, the story is far from over. Researchers are racing to explain associations among all three cancers and 8q24. The region may contain DNA sequences that regulate cancer-related genes, but further evidence explaining the plausibility of the findings has yet to be published.
"Something mysterious is going on in 8q24, and we are having difficulty putting the data in a rational context," said Dr. Kari Stefansson of deCODE Genetics, which first linked the region to prostate cancer in 2006. Nonetheless, he expects an answer by spring.
"Genome scans are the best tool we've ever had to pinpoint the genetic causes of disease," Dr. Stefansson added. "Never in history have we had anything that comes close. They are not perfect, but the ability to look at hundreds of thousands if not millions of variable regions of the genome at the same time is remarkable."
Scans often involve thousands of subjects. Success depends in part on having relatively homogeneous patient populations and matched comparison groups. DNA from both groups is screened for variants found more often in the affected group. Results are then tested in additional patient populations to confirm associations and avoid false-positives.
Most of the variants reported to date have modest effects on disease risk and would not be used in the clinic. With the next generation of genome scans, researchers will begin to focus on assembling panels of common risk variants.
"The goal is to quickly transition from discoveries in single studies to discoveries in several studies, and to put together strong robust findings so that researchers can test 10 to 20 common variants," said Dr. Stephen Chanock, of NCI's Division of Cancer Epidemiology and Genetics (DCEG), who directs its Laboratory of Translational Genomics.
"We are most excited about the opportunity to put them all together because we know that is where we want to be," said Dr. Chanock. Collaborative studies across the globe will be needed to achieve this goal, and efforts toward this end are underway.
By combining scans for the same cancers, it should be possible to look at subtypes of disease to see if there are specific genetic factors that one wouldn't see when looking at the total group, said Dr. Robert Hoover, also of DCEG and a leader of CGEMS.
Meanwhile, researchers are testing associations in their own study populations. Of particular interest is the increased risk of prostate cancer in African American men. "The initial scans were largely done in white populations, and we're trying to see whether the effects are the same in men of African descent," said Dr. Timothy Rebbeck of the University of Pennsylvania.
Another application is to use scans to find protective genes. One such variant has been found for breast cancer, and others may follow. An international consortium is planning to screen women with mutations in the BRCA2 gene for genetic variants that may cause some of these women to develop breast cancer at an early age and others not.
"If we find a genomic pattern that is protective in the genetic background of BRCA2 mutations then we will see if this result is generalizable," said Dr. Kenneth Offit of Memorial Sloan-Kettering Cancer Center, who is leading a study by CIMBA (Consortium of Investigators of Modifiers of BRCA1 and BRCA2).
Genome scans may indirectly lead to environmental factors involved in cancer. For instance, genes that are more strongly associated with breast cancer in women who use hormones than in nonhormone users may mediate the effects of exposure to hormones. Understanding the biological pathways involved could lead to safer uses of the therapy.
Epigenetic changes to DNA, which regulate gene activity without altering DNA sequence, also play a role in cancer. Understanding the interactions among genetic, environmental, and epigenetic factors on cancer represents an enormous challenge, and researchers stress that genome scans are merely the beginning.
In fact, genome scans often do not reveal the precise stretch of DNA responsible for an association, so more work is needed to pinpoint it. Then, functional studies are needed to truly understand the source of the association.
"Results of scans should be replicated many times by multiple groups before reaching firm conclusions or changing therapy based on the findings," cautioned Dr. William Evans of St. Jude Children's Research Hospital, who has collaborated on leukemia scans and used this approach in pharmacogenomic studies.
Dr. Mary Relling, also at St. Jude, added: "We need to be honest with the public and with our funding agencies that this process is going to be hard and will be done over decades." Nonetheless, she is optimistic and believes that genome-wide approaches are important because they may reveal factors that might otherwise be missed.
Dr. Relling is using scans in combination with other tools to study variation in drug response and toxicities among children with leukemia. And in the field of pharmacogenomics generally, many scans are underway using patient cells.
"Just as the use of scans for identifying risk factors in common diseases has exploded on the scene, you're going to see an echo of that explosion in research on variation in drug response," said Dr. Richard M. Weinshilboum of the Mayo Clinic, a former chair of the Pharmacogenetics Research Network. "Tune in next year for the results."
Researchers have seen "the first light of genetic variants associated with common diseases that we had never thought of before," said NCI's Dr. Chanock, who is a leader of CGEMS. "The excitement comes from seeing that our strategy works and that the approach is sound." The information from these studies will be used in many ways, including some that are not yet apparent.
"We are putting the CGEMS results out there and counting on people to look at the information from many different perspectives," Dr. Chanock said.
—Edward R. Winstead