By Meredith W. Salisbury
Sometimes it feels like there’s a new one sprouting every day. You know what we mean: the Cancer Genomics Program at the MD Anderson Cancer Center, part of the University of Texas. Johns Hopkins University’s Program in Applied Genomics for cardiopulmonary disease. Stanford’s Asian Liver Center, using genomics to study liver cancer among Asians.
Everywhere you turn, there’s a new center relying on functional genomics to study one particular disease or group of diseases. What happened to Lesson Number One in genomics history: the soup-to-nuts mega-institute shall always prevail?
As genomics matures, it looks like things at NIH are heading back to business as usual: funding allocated for specific disease research, and genomic scientists — well, really, anyone looking for grant dollars — are finding ways to apply genomics to any range of diseases, rare or common. Elias Zerhouni, NIH director, told his audience at the AAAS meeting this year that he aimed to support the growing trend toward disease-specific research.
And why not? Such specific centers seem more likely to get funding from illness-based foundations: a skim of places like Alzheimer’s Association; FRAXA, the Fragile X Research Foundation; and the Cystic Fibrosis Foundation show that they’re already putting money toward genomic studies in their fields. That will certainly help make NIH’s dollar go a little further, and give more researchers the chance to work on these diseases.
It’s too early to tell if the trend is a good one. But there are obvious concerns about having these little centers working independently on particular diseases. Duplicated efforts are costly, and small institutions often don’t have the resources (large sample banks, for instance) to really come close to solving any given disease. A potential solution comes from Lesson Number Two from genomics history: the consortium model may save everything.
Time Is Right
Zerhouni may be backing the trend, but that’s no indication that other types of research centers won’t get funded, says Alan Guttmacher, NHGRI deputy director. “We’re early in this. It makes sense to support different kinds of models,” he says.
But some argue that the time is right to try out this kind of initiative. “We’re starting to come into an era where a lot of what we know about genomics can be applied to disease,” says David Bowtell, director of research at the Peter MacCallum Cancer Genomics and Genetics program in Australia. “It’s quite reasonable to take a disease focus.”
Bruce Stillman, director of the Cold Spring Harbor Laboratory, says this is just one of the byproducts of the international genomics effort. “We are in an extraordinary time at the moment because of the completion of the Human Genome Project. It’s opened a huge number of doors.”
Those doors could all open onto new research strategies, says Arthur Holden, CEO of the SNP Consortium. “There is no more profound capability to change the way we do research, the productivity of our research to get at the underlying causes of disease, than the advent of genomics.”
Guttmacher was hardly surprised by Zerhouni’s announcement this year. “NIH has always supported disease-specific centers,” he says. That’s encouraged, naturally, by the institutes’ disease-specific identities. “In some ways, this is just applying genomics to that kind of framework.”
And NIH tradition certainly has its value, points out David Altshuler of the newly formed Broad Institute, which houses the Whitehead/MIT Center for Genome Research. “Biological and medical research have done extremely well in what has become the traditional model of biomedical research,” he says, adding that NIH funding history has contributed to “the most successful [scientific] enterprise in the world.”
But Stillman, who is vice chairman of a National Research Council policy board that recently released a study on how to do large-scale biomedical research, thinks genomics may have reached a point where the regular PI-driven grant isn’t enough. Large groups of researchers or even institutions can hardly send in a grant request asking for $30 million or $100 million — numbers that might be needed to start truly integrative genomic centers working effectively on disease research. Stillman calls the lack of a good funding mechanism for sizable efforts “the biggest problem we have at the moment.” The typical NIH grant, orders of magnitude less, doesn’t cut it for these efforts: “Five hundred thousand dollars doesn’t go anywhere nowadays with large-scale science,” he says.
Whether the mechanisms are in place or not, proponents contend that funding these disease-focused research centers has to happen. “Efforts to fully understand these complex molecular events are beyond the reach of any one laboratory or group of investigators,” Zerhouni said in his April budget request to Congress. “Because of the complexity and scope” of these problems, more traditional “models must be replaced by integrated teams of specialists from numerous disciplines that were considered unrelated in the past.”
Aside from their goal of making progress toward a particular disease, research at any of these centers could prove key to myriad other health applications, Guttmacher points out. He recalls as an example a small study on familial hypercholesterolemia — designed to examine just that particular single-gene disease — which recovered so much information about the cholesterol pathway that it led to the discovery of statins, “one of the most widely used class of drugs in the country” for people at risk of various cardiovascular diseases. Even very narrowly focused biomedical research “tells us something about basic biology,” Guttmacher says.
“The concept can work if it’s a really integrated center,” says Michael Liebman, director of computational biology at the Abramson Family Cancer Research Institute at the University of Pennsylvania. He likes the disease focus because it makes more sense from a clinical standpoint — starting from the disease and its effect in patients and stripping that down, layer by layer, to understanding “how genomics and proteomics can help differentiate between subtypes of disease” — than starting with genomics and working up from genes.
And starting with disease-driven research, rather than working from the genes up, should be a faster track to cures and other breakthroughs, argues Liebman. “If you start with patient care and work back” from symptoms down to genes, “you’ve validated that target,” he says. Starting with the genome and working up to the health level is a more winding way to do things: once a target is finally found, it “requires extensive validation” to prove its connection to a particular disease.
Because of a heightened sense of urgency toward the diseases that cause the most deaths, Guttmacher, for one, expects to see plenty of centers grouped around those. Top killers in the most recent studies, according to the National Vital Statistics Report from the National Center for Health Statistics, are heart disease, cancer, stroke, diabetes, Alzheimer’s, nephritis, liver disease, and hypertension.
The sense of excitement about these kinds of centers doesn’t hide the evident problems with the model. Like the early days of genomics, pre-HGP, lots of small centers could eat up funding and never accomplish a sweeping gesture on the order of sequencing the human genome. “We have to be very careful that we don’t duplicate things,” emphasizes Stillman. “There is not the amount of funding that is needed [already].”
Another obvious problem is scale. “If you really want to start to think about the genetics associated with common disease,” Holden says, “this is really complicated stuff.” These ideal research centers have all the latest genomics technology in addition to access to hundreds or thousands of patient samples, as well as experts in the disease of interest and a solid chemistry foundation, among numerous other factors. No small feat for a small center.
Also, smaller institutions can find it harder to keep running for any number of reasons: changes in leadership, unexpected financial changes, a shift of attention from funding agencies. Robert Lees, president of the Boston Heart Foundation, was giving presentations on the new Cardiovascular Genomics Center at that institution in 2000; today, the center no longer exists — work is still being done, but under the auspices of the BHF rather than a standalone program. (Representatives of the BHF were unable to speak with GT to explain why the center never took off.)
Wei Zhang, director of the Cancer Genomics Core Laboratory at MD Anderson’s cancer genomics program, says that it actually may be too early for larger efforts at this point — his research center, which gets some $10 million a year for operations in addition to grant money, goes it alone. “We still need to find out which platform is the best, the most consistent … who are the experts. We need to sort out all that before we centralize — otherwise we’re going to get junk.”
Beefing up with the infrastructure necessary for larger centers could be a huge problem in the long run, says Stillman, whose own Cold Spring Harbor has some 30 people working on a cancer genomics program. “We have the capability of making huge inroads into cancer now,” he says. “And it’s not being done because no one’s taking the lead” on an effort that would run in the hundreds of millions of dollars range and require infrastructure not available in even some of the larger centers, let alone the programs like Stillman’s, which operates on $6 million annually.
“The enthusiasm for these centers is justified,” says Altshuler. “But I don’t think we should go too far down this road. … There are a great many problems where we don’t know what to do and throwing people together isn’t the solution.”
Enter the consortium model. “Much of science is moving toward consortium-type approaches,” Guttmacher says. Particularly in the field of genomics, with major exemplars such as the SNP Consortium, the HGP, and now the HapMap — as well as with NHGRI, where large-scale collaboration is the “ethos of the institute,” according to Guttmacher — people are familiar with the concept.
It’s not just a matter of getting more heads focusing on a problem. A consortium “gives more statistical power, … avoids duplication, and can get you the answer more quickly,” says Muin Khoury, director of the CDC’s Office of Genomics and Disease Prevention.
And on a purely practical note, Bowtell points out that for any kind of disease research, “you need lots of samples, and to get lots of samples you need a consortium. … That’s where it’s more important to have a large consortium effort than lots of small, sort of redundant efforts that fail to achieve the critical mass.”
Liebman at Penn cites a Pennsylvania consortium, set up with the state’s tobacco funding, as a good example. Six major cancer centers have agreed to work on cancer, starting with breast, prostate, and melanoma, Liebman says. Divvying up the work in such a way means the consortium’s accomplishments could far outweigh anything the six individual centers are capable of on their own, especially with something as broad as cancer, Liebman adds.
Stanford’s Asian Liver Center is another example. Founded in 1996 to study liver disease in the Asian population, the center never expected to get lots of NIH funding. Liver cancer “really doesn’t get much federal research money,” says Samuel So, director, who says so many futile studies have led most people to give up on it. So he and his research staff of five work out of the basement of one of the university’s medical buildings. They probably couldn’t make much difference alone, but they partnered with the University of Hong Kong and Peking University, which “allows us to perform these studies at a very rapid pace,” So says.
Bruce Stillman sees Australia’s Peter MacCallum cancer effort as a model of how things should work. “It’s much more than genomics, a whole integrative approach,” he says. “There was a national consensus in Australia” to attack breast cancer, Stillman says.
“It started around the development of microarrays back in about ’97,” says Bowtell. The idea was to “apply genomics right across the board in terms of the cancer sequence.” The program will study some 80 percent of women in Australia to figure out the roots of breast and ovarian cancer. Funding comes from all over — from private foundations to public institutions — and in fact, 60 percent of the institute’s money comes from Australia while 40 percent comes from overseas, Bowtell says.
One of the earliest, and considered most successful, examples of the consortium approach in genomics was the SNP Consortium. CEO Holden says one of the reasons it did so well was the emphasis on always investing money in the most cost-effective way. “Money must flow to the entities that have the greatest capability to do the research effectively,” he says. “By managing every single aspect” of the SNP Consortium in that way, “we developed a map that was 10 times greater [than expected] in 70 percent of the time with 80 percent of the cost.” (Indeed, Eric Lander noted at the genome celebration this April that the international genome sequencing consortium had also achieved better results in less time and with less money than originally planned. No wonder this is a model that has “a lot of value,” according to NHGRI’s Guttmacher.)
When Stillman’s panel was going through successful examples of large-scale biomedical research, a finding that emerged was “a strong recommendation in involving industry,” he says. Companies can be helpful and have strengths, such as training scientists as well as managers, that may not be inherent in academia. Another good consortium he looked at: NCI’s Mouse Models of Human Cancer Consortium, a group of 20 PIs affiliated with more than 40 research institutions looking into characterizing mouse models for cancer research.
A sidenote to the consortium approach is its equivalent on the funding side: expect to see more and more partnerships for providing grants and resources, says Guttmacher. NIH institutes, for instance, have been doing more and more as groups — “That’s a trend we’ll see continue,” says Guttmacher.
Despite the rise of disease-focused centers, other types of institutions are still considered valuable. Working up from the genome — while it presents unvalidated targets, Liebman’s complaint — can still provide critical information. One such example is folic acid and its pathway, says CDC’s Khoury. “It’s one of those vitamins that has been implicated in birth defects and cardiovascular health and some cancers. If you want to develop projects around cancer and folic acid or birth defects and folic acid, you could do it that way. The other approach would be to start with the genes and look outward,” he says. “There is going to be investment in both approaches.”
Guttmacher concurs. “I’m not sure we know at this point which approach will be better.” (In fact, most people concede that the genes-up approach may work best for certain types of research while the disease-down approach works for other types.)
Arthur Holden sums up the debate from a completely different perspective: “Let me comment as a taxpayer. I only want them to do one thing: I want them to do this research in the environments where they can do this most effectively and most efficiently. And I don’t care about anything else but that.”
|To Name Just a Few
A sampling of genomics-based disease research programs:
• Cancer Genomics Program, MD Anderson Cancer Center, University of Texas
• Program in Applied Genomics for cardiopulmonary disease, Johns Hopkins University
• Asian Liver Center, Stanford University
• Cancer genomics program, Cold Spring Harbor Laboratory
• Cardio Genomics, Harvard Medical School/Beth Israel Deaconess Medical Center
• TexGen, a joint program involving the University of Texas Health Science Center, Baylor College of Medicine, and the University of Texas MD Anderson Cancer Center
• Center for Human Genetics, Duke University Medical Center
• Carlsbad Genomics Institute, Carlsbad Science Foundation
• Minnesota Partnership for Biotechnology and Medical Genomics, a joint program between the University of Minnesota and the Mayo Clinic
• Ian Potter Cancer Genomics and Genetics Program, Peter MacCallum
• BayGenomics, a research consortium funded by NHLBI
• PhysGen, Medical College of Wisconsin
• Innate immune response program, University of Arizona
• Cardiovascular and lung program, Massachusetts General Hospital