Credit (and thanks) to DewDiligence for reformatting the PDF.
With New Disease Genes, A Bounty of Questions
By JENNIFER COUZIN 28 MARCH 2008 VOL 319 SCIENCE
New techniques, including genome-wide associations, are identifying new disease risk factors; researchers are uncertain what they mean—and what to advise patients
Scientists trying to understand the factors that lead to breast cancer have a wealth of new leads to follow up. During the past couple of years, they have identified an alphabet soup of variations in DNA sequences—CHEK2, FGFR2, TNRC9—that appear to increase a woman’s risk of getting breast cancer. But these new finds come with a host of questions— in particular, whether to test women for these genetic variations and what to tell those who carry them. It’s been “a matter of uncertainty about how exactly to advise people” on this, says Mark Robson, a breast oncologist and clinic director of the clinical genetics service at Memorial Sloan-Kettering Cancer Center in New York City. For the most part, Robson says, he doesn’t.
Counseling women about mutations in the widely cited BRCA1 and BRCA2 genes makes sense, says Robson, because they can increase breast cancer risk by as much as nine times— a risk “so high that it clearly exceeds most people’s threshold for action.” But what to do with a gene variant that shifts breast cancer risk from 13% to 16% over a woman’s lifetime? Or one that puts diabetes risk at 9% instead of 7%? By comparison, environmental effects can have a much bigger impact: For example, heavy smokers in their 50s have a 6% risk of dying of lung cancer over 10 years compared with about 0.2% for nonsmokers.
In the last year or so, questions about what to do with the flood of data have taken on new urgency. Genome-wide association (GWA) scans—which survey the genomes of people with a particular disease and compare them with the genomes of those with- out—are turning up dozens of DNA variations that boost risk only modestly. The results have generated enormous excitement among researchers, long frustrated by their inability to find variants driving common adult diseases. But with the data come more questions.
Researchers are finding that even with all the new details from GWA studies, much of the canvas remains obscure. For example, the function of most GWA variants hasn’t been determined; it’s also not known whether different variants that increase risk slightly for a disease might interact with or add to one another to increase risk substantially. Nor is it clear how the variants might contribute to disease mechanisms or treatment. Will a public that’s apparently hungry for genetic knowledge incorporate low-risk data into their lives? These and related questions—many of which came up at a meeting earlier this month at the U.S. National Institutes of Health in Bethesda, Maryland—have leaders in the field wondering how best to apply recent findings, where to focus the next round of studies, and how to convey often sketchy data to the public.
Risk tolerance
Some concerns about the public’s reaction have already eased. Two years ago, behavioral epidemiologist Colleen McBride and human geneticist Lawrence Brody, both at the National Human Genome Research Institute, began examining how healthy individuals respond to disease risk information. They offered volunteers in Detroit, Michigan, the chance to learn whether they carried deleterious variants for eight health conditions, including diabetes, colon cancer, and osteoporosis. Because the variants are common, virtually everyone was expected to harbor at least a couple. Those monitoring the study’s safety “were really worried, literally, that people were going to jump off bridges” when they learned that their risk of disease was increased, says McBride.
Among the 300 or so who have participated, that hasn’t happened—quite the opposite. “They’re not having big emotional responses,” says McBride. The researchers are tracking the volunteers to see whether the information affects decisions to reduce disease risk, such as seeking out a smoking-cessation program or consulting with a nutritionist.
Behavioral specialists have shifted from worrying about the devastating effects of learning about these new genetic risks to wondering whether the information will make any impression at all. In some ways, this isn’t surprising—after all, many people with high cholesterol or high blood pressure don’t make lifestyle adjustments, even though the markers have a substantial effect on disease risk. One open question is whether “people perceive the information as more accurate when DNA is being used,” says Theresa Marteau, a psychologist at King’s College London. Her analysis of published studies found that rarely is genetic information regarded fatalistically, as predicting inevitable disease. Now she’s considering whether there’s something about genetics— perhaps its uniqueness or its perceived accuracy— that can help drive healthy behavior, even if it doesn’t add much new information about risk.
Marteau is testing this hypothesis in people who have a close relative with Crohn’s disease, an inflammatory disorder of the digestive tract, aiming to recruit about 540 individuals. Because of their family histories, the participants have a Crohn’s risk of 2% to 6%, compared with 0.1% in the general population. But their risk is also high for another reason: All the people in Marteau’s study are smokers, which about doubles their chance of developing the disease.
Marteau wants to know whether adding genetic information to a risk assessment— even if it doesn’t dramatically change the actual risk of disease—makes her volunteers more likely to stop smoking. All participants will receive information based on family history; half will also be tested for a gene, Nod2, that boosts risk of Crohn’s.
For many diseases, researchers are beginning to consider whether certain combinations of gene variants might have a major impact. In January, scientists reported in the New England Journal of Medicine that for men with a family history of prostate cancer, five genetic variants together increases risk roughly nine times. This is similar to a BRCA-linked risk for breast cancer. About 2% of men harbor four or five of the prostate cancer variants in question, says Jianfeng Xu, a genetic epidemiologist at Wake Forest University School of Medicine in Winston-Salem, North Carolina, who helped lead the study. Xu and his colleagues have formed a company to commercialize the test, which he expects will be sold to doctors starting late this spring.
Although a ninefold boost in risk for prostate cancer is substantial, Xu still worries that the measure is crude. Only about 15% to 20% of men diagnosed with prostate cancer need aggressive treatment, and a much better test would be one anticipating the likelihood of aggressive disease rather than any prostate cancer. Xu is now hunting for more informative variants and hopes to add them to the test when they’ve been identified.
The public appears hungry for such information: After his paper appeared, Xu received calls and e-mails from people with a family history of prostate cancer wanting to know their genetic risk. “I don’t think we’ve been very good at anticipating that market forces were going to enter into it,” says McBride, noting that researchers and physicians have been outraged by the proliferation of companies selling gene tests directly to consumers. DeCODE Genetics in Reykjavik, Iceland, recently began marketing risk tests for type 2 diabetes, atrial fibrillation and stroke, heart attack, and prostate cancer. Its prostate cancer test, released in February, includes the five variants Xu is focusing on along with three others.
Frontier zone
While physicians worry about how to convey the results of genetic tests to patients, some scientists are thinking more about what GWA studies are not turning up. GWA has established a list of 31 genes implicated in Crohn’s disease, for example, but there’s still a mystery about how the disease is inherited. None of the 31 genes has explained why Ashkenazi Jews—a genetically cohesive group—are disproportionately at risk for Crohn’s. There must be a genetic component that “we’re just not getting yet,” says Judy Cho, head of the Inflammatory Bowel Disease Center at Yale University. Cho is beginning studies of copy number variation: duplication or loss of DNA stretches that may contribute to disease but aren’t picked up in GWA scans. In type 2 diabetes, notwithstanding some widely hailed GWA discoveries, “the proportion of heritability that we’re picking up is relatively small, and that places serious limits on how good these are as individualized predictions in diagnostics,” says Mark McCarthy, an endocrinologist at the University of Oxford in the U.K. “Despite all the giddy excitement, we’re only capturing a very small proportion of what’s out there.” He suggests looking more closely at rare mutations, which, like copy number, are territory that GWA is not designed to capture. GWA studies could also be skewed by the populations they cover. So far, most have been done on people of European origin, and it’s not clear how well they’ll translate to other groups. In heart disease, a DNA stretch called 9p21 that falls between genes has been replicated in four large GWA studies, but it’s not standing out yet in African Americans. “We’re still a little bit unclear about that,” says Ruth McPherson, an expert in cardiovascular risk prevention at the University of Ottawa Heart Institute in Canada and one of those who first discovered 9p21. She says that she and her colleagues are “trying to understand if it’s not a risk factor for disease among blacks.” Complicating matters are the logistical challenges: Many patients in GWA studies of heart disease, says McPherson, are not defined based on coronary angiography, x-ray imaging of the arteries of the heart. Finding people who have undergone angiography to serve as controls is another “difficult problem,” she says. Differences among populations and study designs suggest that failure to replicate a finding doesn’t necessarily mean it was wrong, she and others say.
That argument isn’t persuasive to David Cox, chief scientific officer of the company Perlegen Sciences in Mountain View, California. “You can come up with all sorts of reasons why [a GWA study] didn’t replicate, but if you want to use this for any kind of prediction, it has to show up over and over again,” he says. Cox worries that scientists are engaged in a love affair with GWA to the neglect of other strategies that can dissect disease. “It’s like a huge stampede,” he says.
One limitation of the technology, Cox notes, is that it likely won’t pick up genetic variation that correlates with a person’s response to medical treatments because the number of participants in a study would need to be enormous. Cox is working on this problem; his approach relies on rare gene variants to detect an effect that may be more potent than that conferred by common ones. Cox also questions how often GWA will lead researchers to new drug targets; the possibility has generated much hope and may yet pan out.
All these lingering questions are best solved by one strategy: “Find more of the genes,” says McCarthy. “Think of how little of the territory we’ve covered.” That solution—keep up the hunt—is drawing broad support.
<< |
