At A Glance
Name: Ralph Snyderman
Title: Chancellor for health affairs; executive dean, Duke University School of Medicine; president and CEO, Duke University Health System.
Education: MD, Downstate Medical Center of the State University of New York
Background: Vice president and later senior vice president at Genentech; assistant professor of medicine and immunology and chief of rheumatology at Durham Veteran’s Administration Hospital
Ralph Snyderman wants a pharmacogenomics revolution. But instead of the more traditional calls to use genotyping and gene-expression tools to revamp drug-discovery and drug-development protocols, Snyderman, who is CEO of Duke University Health Systems, wants to see pharmacogenomics infiltrate the health-care continuum more quickly.
“We are now sitting on a great tsunami of biomedical research,” he told attendees of last month’s Genome Sequencing and Analysis Conference. Snyderman, who is also executive dean, Duke University School of Medicine, wants pharmacogenomics not only to treat disease, but to prevent it before it occurs — not a novel goal, to be sure, but one that may gain traction more readily thanks to a novel game plan, the help of a gene-sequencing pioneer, and some wealthy guinea pigs.
SNPtech Reporter caught up with Snyderman last week.
How do you think pharmacogenomics technologies can play a role in the health-care system in the United States, rather than just the drug-development continuum?
I think the first thing to make clear is the introduction of genotyping in regards to health care right now is more theoretical than practical. And one of the points that I made at the GSAC meeting in Savannah is that the major transformation of health care will be when we could develop and individualize risk assessment for individuals developing disease, so that we could truly try to practice preventative medicine in a personalized basis. In order to do this effectively, if you think of the time frame of disease, that most chronic disease develop over many, many years, that understanding an individual’s susceptibility — even from the time of birth — will certainly inform an individual’s personalized health plan.
So it stands to reason that at some point, genomics, by providing susceptibility information, will be part of a health-care system in which we try to intervene or prevent at the earliest possible time, rather than what we are doing now, which is treating after an event occurs.
The question is, ‘How will genomic information be introduced?’ We have a collaboration with Craig Venter to be thinking about, giving the power of sequencing anybody’s genes, what do we do with [the sequence]? And if we were to sequence an individual’s genes now, and say, ‘How does that inform the practice of medicine,’ today, that wouldn’t be very helpful because we don’t have a tremendous amount of insight as to truly what are the risk modifications in someone’s genome. So we need to figure out how we’re going to introduce genomic material in a way that makes sense.
Let me give you an example of what we’re thinking about right now. We are beginning to engage in a project to try to determine individuals at risk at a very early age for developing obesity, and, as a consequence of obesity, diabetes, cardiovascular disease, and other things that are associated with obesity. And we’re trying to develop risk-assessment tools.
I’d like to think … theoretically for a moment about how you can introduce genomics into that. To do a full sequence of everybody’s genes would be nice, but it would exceed the capability of what we would be able to do [in regards to] costs and analyzing the information. However, from experiments at looking at knock-outs in mice, for example, it’s already apparent that there are probably at least 40 genes that play a major role in determining susceptibility to obesity. These genes will probably affect different pathways and have different metabolic consequences. So what we are thinking about is based on the literature, maybe based on information with biotech companies that are doing knock-outs with many, many mice, identified genes of interest. For example, susceptibility to obesity, [or] susceptibility to diabetes. And if we are going to be dealing with a population to determine susceptibility to obesity, we would screen everybody in that group, initially for these particular genes — in other words, do genomic sequencing of these particular genes. And then follow over time, and see whether or not they really do predict the development of obesity, and, even more importantly, the association of obesity with other entities such as diabetes.
That’s one of the ways we’re thinking about introducing this now. The other thing that we will think about is, what are all the genes of interest that may be related to the development of cardiovascular disease, atherosclerosis? Then we would do a genotyping of these. What are the genes that may be associated with the development of prostate cancer? Ovarian cancer? We would do those as well.
There are studies ongoing now at Duke where researchers are doing careful gene-expression analysis associated with breast cancer. And what is being found — this is work being done by Joe Nevins [director of the Duke Center for Genome Technology, one of five centers that make up the school’s Institute for Genome Science and Policy] and Mike West [The Arts & Sciences Professor of statistics and decision sciences] — is that there are quite discrete patterns in gene expression in tumors that either do or do not metastasize, do or not have a very high mortality associated with them. Once we identify what these genes are, we would probably do a genotype analysis to see if the level of expression in the tumor is inherently based in the genomics of that individual. We would do this rather than just gene expression.
I guess what I’m saying is that we expect to introduce genomics analysis in a piecemeal basis based on diseases of interest, based on what we know very largely from other information as to what the genes of interest are. And then we will track people over time.
We are also thinking, but haven’t yet decided whether this makes sense in individuals who come to us for executive health physicals. These are individuals who are usually senior officers of major corporations — people who can afford to get the very best for analyzing their own health; they come to Duke, or they may go to the Mayo Clinic, or other places.
We’re thinking of offering the option of a genotypic analysis of either all their genes, or genes of special interest to them, but the reason we haven’t done it yet is that we don’t know what the individuals can usefully do with this information since, other than worry about it, there isn’t enough information there to know what is truly meaningful so that one would modify their behavior. But when Craig Venter and I initially started talking about how we might partner his … technology and understanding of genomics and Duke’s capability of delivering … healthcare, one of the things we were thinking about is doing full geno-typic analysis of all reading frames of genes for individuals who wanted to have this information. We just haven’t gone ahead and done this yet.
Tell me about the pilot study at Duke that you mentioned during your GSAC presentation.
We’re doing a few things in trying to develop tools to anticipate risk, and trying to intervene prior to adverse outcomes—what we’re calling prospective medicine. And with CMS, we were given a grant to see whether r not early intervention as a development of a risk-assessment tool so that we could provide individuals with highly personalized information saying, ‘Compared to a normal population, you have a 40-fold increased risk of having a heart attack within three years,’ whether or not that would be helpful in modifying the outcome. Initially in that study, we are not using genomic information; we are using family histories. We expect down the road that that is exactly what we are going to do.
Let’s say you wanted to know what are your risks for the top 10 major, potentially preventable diseases. What are you most susceptible to, and what do you need to do about it? Currently, we don’t know how to use genomic information to get you that risk assessment. But we feel it’s inevitable that within the next five years that more genomic analysis will be informing in determining an individual’s risk. What we’re trying to do is to develop the template in which we could insert genomic analysis. And that’s what we’re doing right now.
You said using genomic information for medicine is in the theoretical stage. What about the use of genotyping and gene-expression technologies in molecular diagnostics? Doesn’t a test based on a cytochrome P450 mutation — a test used by many reference labs today — rely on genomic tools and information?
I wouldn’t argue that there aren’t ones [products based on genomic technologies or data] already on the market. And I think that’s a very good example — the P450 series. I think that if you were to ask, ‘Where will genomic analysis have the initial impact in the practice of medicine?’ I think pharmacogenomics very likely is going to be an early use.
Back to your research with Venter: Is the aim to learn as much as you can from applying these technologies to wealthy executives, and then to move those discoveries and applications to, for lack of a better term, normal people?
We would like to. I don’t know whether the volumes will be enough. What’s ironic is that, one of the question you posed — How do you put genomic analysis into the current health-care system? — I think what you have behind that is, ‘Will the current health-care system reimburse it or support it?’ The answer is a resounding no. Absolutely no. So we have to find ways in which we have to deal with the reimbursement system and make progress.
In the executive health program, we have people who are paying their own freight. Ironically, at the other end of the spectrum — the indigent, or who have very little access to health care — there we could try to receive foundation funding, or funding from other entities to … put these analyses into large-scale pilot clinical project. Right now, the health-care system doesn’t reimburse for this.
So the answer to your question is yes; we would try to do that. In order to get sufficient volumes to study, it’s likely we’ll have to do other things — such as having focused studies on obesity, cardiovascular disease, or maybe certain forms of cancer.