January 2006: Is Genomics the Cure for Disparities?

National Human Genome Research Institute

National Institutes of Health
U.S. Department of Health and Human Services


Is Genomics the Cure for Disparities?

A presentation by NHGRI Director Francis Collins to the National Leadership Summit to Eliminate Racial and Ethnic Disparities in Health

January 10, 2006

Francis Collins, M.D., Ph.D.:

Well thank you very much Garth. Good morning everyone. If you are wandering in wondering where we are in the program. We are at the point that would have been at 9:05 but a lot of interesting things have been happening already this morning. So, we are a little behind and I will try to be concise in my comments. People who know me well will be skeptical about that but I will do my best because

I want to give the panel a chance to weigh in on this very important question. Is genomics the cure for disparities? Well let's consider the issue of inheritance and what role it plays in disease and what that has to say about health disparities, conditions that all of us are deeply concerned about and representing it by being here at this meeting.

If you look these young people, you will immediately perceive that they come from different ancestral geographic origins and they faced different futures as far as the likelihood of illness unless we do something about the disparities that are currently deeply disturbing, which brings us all here together for this historic gathering. How much of a difference do they face related to the DNA that they inherited and how much is related to environmental experiences and all manner of other issues that is attached to this very blurry concept called race. That is what I want to try to walk through recognizing that we still don't have a terribly precise answer to that question. We are learning quite a lot especially in the last couple of years.

Of course my focus because it is the thing that I know the most about is going to be on the inherited side of this and that means we need to talk about DNA, this really incredible molecule that exists within each cell of your body carrying information that you have inherited down through the generations and this particular information storage system in each in your body means that there is a complete set of instructions, a series of letters along the DNA code that you are either A, C, G, or T. and three billion of those make the human genome, the human instruction book.

It is a very remarkable thing that until very recently we didn't know a lot about that. We now have a whole instruction book in front us but let's ask the question of what is the nature of variation in that DNA. We are not all identical as we would all be one big set of identical twins and that would be pretty boring. So, what is the nature of that variation?

Well Genome project actually was not set up originally to pay as much attention to variation as it was to try to look at the part that we all share. So when the Human Genome Sequence was announced in its completed form in April of 2003, this was basically a hodge podge of DNA contributed by a number of autonomous individuals but that in itself didn't tell you much about the differences between people because any given segment of DNA came from just one person's chromosome. But clearly we need to understand that variation because varied within that portion of the genome are the risk factors for common illnesses that we really want to understand. And as it turns out that is only 0.1-percent of the DNA, 99.9-percent of the DNA is the same. But that point that 0.1-percent is obviously of great interest. To focus on that another International project was founded just a little over three years ago, the International HapMap [misspelled?] project. Again, NIH found itself in the position of being the lead in this enterprise and I found myself as the project manager. Again, trying to organize researchers from six countries to work together to try to build a catalogue of genetic variations as quickly as we could. This seemed like a very daunting project when we started and almost unaffordable because the technology hadn't matured. From the space of just three years, once again coming in sooner than expected, the HapMap delivered a really remarkable picture of DNA variation occurs across all of the chromosomes. As major sent in DNA samples, 270 of them, that while they only represented those 270 people were chosen so that a third of those come from European background, a third from Asian background and a third from West African background. That then gives us a snapshot of something about how variations operate across the world as well as how it operates across chromosomes.

This has been a gold mine of information for people who really were interested in trying to unravel the genetic contributions to diabetes and heart disease, to mental illness, to blindness, to a whole host of conditions that fill up our hospitals and our clinics with people who have illnesses that we poorly understand and frankly poorly treat. So for many of us who got into genomics in the first place with this promise to provide real understanding of those common illnesses. This is an exciting moment. Now that we have the tools from the HapMap we will over the course of the next three or fours years be discovering what are the major heredity factors in those illnesses and the big question for this conference is will those heredity factors turn out to be evenly distributed around the world or will it be some circumstances where some of them are more prevalent in one population than the other and therefore will have contributed to the health disparities. I don't know the answer at the present time.

Well what do we know about genetic variation in general? You know that from all manner of sources of information much from genetics but also from the fossil records and also from history, the migrations of the human species have been occurring in a very irregular way from our original founding population and we were once all Africans as I will come to in a moment, I think we now have very strong evidence that we were all once black Africans living in Africa. Yes? We, white people, as I will tell you in a minute, are actually mutants but we will come that. So, the estimates are that a hundred thousand years ago that there were about ten thousand homo sapiens. This was our founders; all of us, living in Africa and the variation that we see in present day human beings was largely already present in those founders. But that was a large enough group that there was a variation present. There haven't been that many new variations that have cropped up in the five thousand generations since then. Most of what is going as far as how genetic variations are organized around the world relates to the ways in which those genes were spread as people migrated out of Africa to other parts of the world.

So in this diagram for instance you will notice that the blue and the yellow circles are more prevalent outside of Africa whereas they are less so in Africa. And that is just simply a function of which people happen to migrate out and which that did not. So on that basis alone without any of these variants having any significant at all you would not expect to have actually the exact distribution, the same frequency of a particular spelling variant in one part of the world versus the other. That is what you would call genetic drift.

On the other hand, a few of the variants that are present in this founder pool might actually have provided a selected advantage in one place or another. We all know the most dramatic example of that is the sickle mutation. If you are a carrier for sickle mutation, you are protected against the most severe consequences of the malarial parasite. So, those who carry that mutation and lived in a malarial area were able to survive better. Therefore that gene occurs with high frequency and whereas, for those who did not live in an area where malaria was prominent the sickle mutation provided no advantage. In fact, it would provide a disadvantage in those who were provided a double dose; those actually would develop a disease that would become sickle cell anemia. That kind of thing has certainly been going on in other circumstances although it is less worked out in most other examples.

Okay. So that is the story about genetic variations but now let's get really murky. What is the relationship between genetic variations and race? Well, first you have to decide exactly what do you mean by race. If we asked everybody in this room to write on a piece of paper a definition of race, that would be really interesting and I suspect we would get at least a number of definitions that there are number of people in the room and maybe more because some of you might want to provide a couple alternatives. The press hasn't helped us much here in terms of explaining how race and genetics are connected. Depending on which article you read you can read something like this that says that "genes, the study of them prove that race is only a sham. No trace of race in genome project proves nothing biologically separates people". But wait a minute if that is the case how can people be making money right now selling DNA tests that will tell you where your origins are and a lot of people are in fact interested in this information. If there is nothing in DNA that tells you something about ancestry, how is it true to say that there is no biological connection whatsoever to race. Here is one of the websites for one of the places that will do that kind of DNA testing and I gather they are getting lots of customers.

Then of course, a big controversary that I will come back to a specific example of Vidal in a minute about whether race ought to be playing a role in clinical medicine in terms of identifying treatments that might work for one group and not another. What is the balance here between trying to do something about health disparities and not making the mistake of redefining the concept of race by assuming it has more biological significant than it does.

So, what does race mean anyway? I like Evelyn Higginbotham's [misspelled?] quote: "When we talk about the concept of race most people believe that they know it when they see it, but arrive at nothing short of confusion when pressed to define it." Try it out on your neighbors and they will fall into confusion, I am quite sure. Most of us if really asked to write this down have that same problem.

Because race has so much baggage it carries with it connotations of history of discrimination. It carries with it cultural and society and dietary practices. It carries a little bit of ancestral geography of course but that in fact is probably the minority of what most people are actually thinking of when the term race appears in a sentence. And is it any wonder therefore as we try to figure out the connection between genetics and race, we have trouble because we are not quite sure what it is that we are connecting to.

So, what do we know about genetics and race? Let's focus on what we know and then I will try to point out the things that we don't know. Well, we know that we are all 99.9-percent the same at the DNA level. That is proven by HapMap and a variety of other types of studies. That is a phenomenonal statement. We are much more like each other than other species on the planet. If someone arrived here from outer space and tried to write a little tree up about homo sapiens, one of the things that they would say is wow these people are really all alike. Compared to other species there is more diversity in a small group of chimpanzees living on one hillside than it is in the entire human race because we are so new on the field. It is these five thousand generations separating us from a very small pool of ancestors. That's pretty unprecedented for our species on the planet. But yet isn't that interesting that our characteristic is that we seem so focused on that difference, that 0.1-percent. That 0.1-percent of three billion is still a big number so that leaves a lot of differences between two individuals, most of those fall into places on the genome that isn't doing very much and have no consequence. And most of the variants as I said earlier that we do have were present in our common ancestors. But because of migratory events and a few other things the frequency of a particular variant where you might have a T and I might have a C, could be different depending on where your ancestors came from. Some of that is just drift. Just the sort of chance thing that happened as people migrated from one place to another and maybe brought a T over here and a C over there. And that has been absolutely no particular significance to anybody's health. Founder effects where you have maybe a very small group of founders who then result in a very large group down stream, like for instance Finland founded by a relatively small group of ancestors. Or today the Amish for instances.

Selection however and new mutations are a different story. Selection as we talked about for sickle cell disease operating in a circumstance where a particular variant provides an advantage if you are a carrier for that. New mutations have occurred. For the most part, they appear and then they get lost if they have no advantage or disadvantage the chances of them hanging around for very long is pretty small, which is again why most of the common variation of the genome has been there since our founders and its still there today. But there are some interesting exceptions and one of them has gotten a lot of prominence in the press about a month ago. An article in Science Magazine heralded by a particularly interesting ? here is the cover art that was used for Science on December 9th described a finding which got a lot of people interested. Now scientifically this was a really interesting story. Investigators at Penn State Keith Chaney [misspelled?] and his colleagues were studying zebra fish. So, you see the two fish in the fishbowl there. The one on the bottom is the way zebra fish would normally appear if you went to the pet store and purchased some. But some years ago, somebody noticed in the pet store a zebra fish had has a golden color and that is the upper one. And that became a subject for scientific investigation. What exactly is the cause of this because it was clearly inherited? To make a long story short, they identified the gene that was involved in this. It was a gene that nobody thought had anything to do with pigmentation, it is an ion channel. Then the investigators, Keith Chaney went to the HapMap database to see do humans have a gene like that too. And if they do, is it the same in everybody? Or could there be actually some naturally occurring variation in that same gene that might have some role in human skin pigmentation. I am just going to show you a screen shot of what a scientist sees. This is what Keith saw when he went to the HapMap website. This is the most complicated slide that I am going to show you so don't get too alarmed; it is going to go downhill from here.

You are looking at a stretch of DNA. And there is a gene down there at the bottom seen it in yellow, called SLC24A5. That is the gene, which in the zebra fish has a glitch in it that causes the golden pigmentation. But you are not looking at the zebra fish's map. You are looking at the human map of this same gene. Yeah. We have a gene that is very much like the one on the zebra fish. Now those blue and red circles are little pie charts telling you that in HapMap samples what is the frequency of one spelling versus another in those places in this gene where there is a variation. So, each collection of four circles is a DNA variation and next to the circles that you will be able to see two letters like A or G or C or T telling what the variation is, what two possibilities that occur at that spot. I want you to pay attention to the one down there in the lower right because that is the one of interest. That is the case where the CUs samples happen to be people living in Utah who are northern European background. You can see that is our only blue, that is a hundred percent of them have the A spelling whereas the other three circles CHP which represents Chinese, living Banghani [misspelled?], yeah, that is Chinese. A GPT, which are Japanese living in Tokyo and YRI, which are Euro people living in [Inaudible] Nigeria. They have almost completely a hundred percent the G spelling. Well it turns out the A spelling actually causes this gene not to function very well and in fact, larger studies have shown us that most people who are derived from European background have that that A spelling. And this same gene which causes reduced pigmentation in the zebra fish is a major contributor to lighter skin in people who live in northern climates.

You will also notice it can't be the whole story because in this demonstration the Chinese and Japanese sample have the same spelling as the African founders so their lighter skin must be on the basis of some other genes. And in fact, it is even more not the whole story because in the European this one gene accounts for about thirty percent of skin color differences in Europeans and Africans but not a hundred percent. There are other genes involved here as well.

So, what happened here? Presumably this change from a G to an A in this gene arose somewhere back there and it provided a lighter skin color which if you were in a northern climate which is where people migrated out of Africa to go further north was a selective advantage because if you are dark skinned in a northern climate where there is not as much sun exposure, you are likely to get rickets. Somebody with rickets will have a difficult time in childbirth. Whereas if you have light skin at the Equator, you are going to end up with a very high risk of skin cancer.

So all this stuff about skin color that we get so excited about because of its historical significance and all of the social values that are attached to that can be perhaps be demystified in understanding this simple molecular demonstration. It is a molecular demonstration of how out of Africa hypothesis. It is a molecular demonstration of how a selection for skin color allowed people to survive better in different parts of the world. And it shows by the way that in this instance race is only skin deep. This gene doesn't seem to affect anything except skin.

So scientifically this is an incredibly interesting story and it has been very interesting to see how it plays out in terms of the public acceptance of this kind of molecular basis for something that we are all so focused on. My hope is and I think it is playing out pretty well so far is that this really does provide an opportunity to get people to think about just how silly it is to get so focused on something as the darkness of one's skin.

It also does show and I have enjoyed saying this to my friends that white people are all mutatant, that we all have this acquired glitch and that gene is not functioning very well in us. Oh, well.

So let's continue our list of what we know about genetics and race. So those differences I just showed you that one but if you look across the whole genome, there are other places where there are quite a difference in frequency of a particular spelling depending upon where your ancestors came from. Most of those just because of drift, some them maybe showing some signs of selection from where we don't know what, and you can in fact use those differences to take a DNA sample and make a guess as to where that person's ancestors came from. But you are not going to be very precise about it especially since as increasingly lots of people don't have all of their ancestors coming from one place. And importantly when people draw out descriptions of human populations, they often draw out this kind as a tree with this rooted sort of foundation in Africa and all of these branches. But that is not really correct. Gene flow don't go in one direction. Genes have been flowing back and forth as long as we humans have been wandering about. In fact, it is probably more appropriate to draw our relationships with each other as a wisteria vine with all of these interconnecting branches.

Maybe that also helps understand why it is incorrect to try to draw boundaries around any particular group and say they are biologically different. That is not justified by the science that we now have in front of us, which means except maybe in a few cases of extreme geographic isolation, the founders of our population groups are going to be blurry and imprecise.

So to get to health disparities. IF you observe then a difference in prevalence in illness between groups, this might be due to genetic differences. It could be or let's really get to the point here. It could have nothing to do with genetic differences and every thing to do with diet, culture, environment and socioeconomic factors. Or to keep our minds completely open a mix of both where some genetic susceptibilities interact with some environmental exposure resulting in an outcome.

And coming back to the skin color, we already know that is true. I am at for higher risk for getting melanoma than a lot of other people in the room because of my light skin color. That is a well-documented gene induction action if I lived in cave and never went outside then I wouldn't have that risk. But the light color of my skin predisposes to that particular environmental outcome.

Well, let's see if we can put this together in some kind of synthesis that perhaps tries to pull together what we know about genes and environment and health disparities.

Everybody in the room probably has some self-identified race or ethnic association but that is an incredible complicated mix of factors that go into that. What is the connection then to health or disease? Well it travels through several pathways. First of all, it certainly correlates with environmental exposures of different sorts of education, of access to heath care, of culture, to socioeconomic status, of stress factors, all of them things together ? none of which are related to your DNA. They are related other factors. ? play a major role in health and disease. But there is also a connection between self-identified race or ethnicity. Admittedly, a blurry one with ancestral geographical origins and in some instances that maybe tighter than others but it is probably in there as some kind of correlation.

And that in turn will correlate with variation across the genome. That in turn will correlate with the variants in specific disease genes and those too perhaps interacting with the environment may make a contribution in some instances to help make disparities. But the balance between the left and the right side of the diagram is for most disparities still not known. We have a ways to go to sort that out. But I think because of HapMap and other efforts we are close to being able to tackle it in generating some real data.

My argument here is if we really want to understand health disparities we don't want to stick to the worry proxy measurements called self-identified race or ethnicity. We could do better than that. You want to the get to the proximal causes because they are the things that you can understand and do something about. The sooner we can move from those proxy relationships the true environment influence and the true genetic influences that are true about the individual and not some poorly defined population group then the quicker we will be on the path to doing something to improving people's health.

Now one final point and then we will convene the panel. There has been much discussion about this particular drug called Vidal that was approved by the FDA last summer for treatment of heart failure in self-identified African Americans. Depending on how you look at this, this is a really good thing or it's a worrisome thing. So let's quickly walk through what happened.

Vidal is a combination of two generic drugs. Isorbide nitrate and hydralazine. It acts we think as a nitrate oxide enhancer. Nitrate oxide dilates blood vessels so this is a good thing if you have heart failure. It reduces the resistance and you would expect that would improve the functioning of some body who has congestive heart disease. So, Vidal was originally tested way back in the 1980s and 1990s in clinical trials involving Veterans Hospital patients in a mix of individuals in different backgrounds. It showed possible benefit, was brought to the FDA but the FDA ruled in 1997, the results were unconvincing and did not approve the drug. At that time they went back and looked retrospectively at the data and became convinced that the benefit that was seen with this drug was greatest in African American patients and less in those that self-identified as Caucasians. Interestingly and this is I think an important part of the story, the patent on Vidal was at that point within only a few years from expiring. A new patent was filed based on this retrospective analysis on the use of this drug specifically for African Americans. The patent office approved that. The first time as far as I know that a patent for a drug for a specific racial group was approved without in fact a whole lot of discussion occurring at the time.

Of course what that meant was that the patent was now refreshed for another seventeen years. So there is an opportunity here for an expanded market opportunity. So what happened? Nitromed, the company that picked up the drug, then conducted a clinical trial solely on African Americans, didn't involve any other groups, eleven hundred of them. A very well conducted clinical trial. The trial was supported by the Association of Black Cardiologists and the National Medical Association. It showed in The New England Journal article reporting this about a year ago absolutely convincing benefit. So the solid line there the Isorbide nitrate plus hydralazine that is Vidal. You can see the overall survival compared to placebo is substantially improved and this is also in patients who are getting other medicines for congestion heart failure and Vidal is being added on top of that. So that is the type of thing that you would except to happen out in practice. So really exciting news.

There is clearly benefits to a disease that affects African Americans at a higher frequency than many other groups and for which treatment has been unsatisfactory and here is a clinical trial showing substantial benefit of this new drug combination. So as you might expect FDA looking at that, approved it by the summer of last year for the treatment of congestive heart failure. The package insert says the following Vidal was indicated for the treatment of heart failure as an adjunct to standard treatment in self-identified black patients to improve survival, to reduce hospitalization for heart failure and to improve patient reported functional status. No comment made about whether this drug should or should not be used for other groups. So was this is a good thing. It is a good thing that we have a drug that treats individuals with congestive heart failure and clearly improves their survival. Nobody could argue that that is not a really good thing. But are we sure that this came about actually in a way that made the most sense. Are we sure that this drug would not have benefited other groups? If you go back and look at the original studies, it doesn't look as if it was benefit in the patients in the VA study who did not identify as black but it was a relative modest size study and there could have very well been some benefits in others.

Are we sure that this has anything to do with being African American or could it be that since African Americans tend to have heart failure on the basis of hypertension that this says this drug is for hypertensive heart failure and not as well heart failure from coronary artery disease, which is perhaps more common in other groups. We know that from epidemiology. Or fail for every black person looking at a lump together theory individuals. But by lumping them together, there is a real risk here that this is interpreted ah, well that means that black people really are biologically different. After all, there is this drug it only works for them. And that is unjustified by the science of what has been done here.

So this is a real interesting example and a two edged sword. More examples like this are potentially coming along. I don't think this is where we want to go. I think we want to go in the direction of figuring out, okay, if this drug works for some people and not others, why is that. What specific DNA variants are responsible for that variation response? And okay if they have them differentially distributed around the world, that is fine, who cares, lets check the individual and find out where they are likely to respond to the drug or not. And not use this very murky and potentially misleading and damaging proxy called race and pretend that we are practicing really upscale medicine. We can do better than that.

So, in summary here race is a term that carries a whole lot of baggage. I think we all agree. Much is being learned about genetic variation around the world. Neither of these statements is correct. Race has absolutely no biological basis. Not really true. There is a connection in ancestral geographic origin. It's murky but it's not absent. But it is certainly true if you can't draw sharp boundaries around groups based on genetics. So again, if you want to understand health disparities we should go after the proximal causes and not depend upon these muddy and misleading proxies. And to answer the question posed in this session is genomics the cure for health disparities probably not but in some instances, it may play a role in terms of heritable factors but almost always in concert with environmental factors. If we really want to know, we need more data. The good news is in another two or three years we are going to have a lot more data on this subject. I think therefore we will be much better poised to do something about it.

Thank you very much.

Microsoft PowerPoint Microsoft PowerPoint presentation. Download the latest version of PowerPoint Viewer.

Top of page

Last Reviewed: March 17, 2012