At A Glance
Name: David Goldstein
Title: Professor of Genetics — Department of Molecular Genetics and Microbiology, Duke University; Director of the Center for Population Genomics and Pharmacogenetics — Institute for Genome Sciences & Policy, Duke University
Background: Wolfson Professor of Genetics — Galton Laboratory in the Department of Biology — University College, London — 1999-2005
Education: PhD in Biological Sciences — Stanford University — 1994
Pharmacogenomics Reporter caught up this week with David Goldstein, a new Center leader at Duke’s Institute for Genome Sciences & Policy, to talk about gene regulation, drug interactions, and other directions he’s taking the Center for Population Genomics and Pharmacogenetics.
When he has settled in at the IGSP in June, Goldstein will have set up his own lab and hired three or four faculty members along with laboratory staff, who will eventually number about 50.
Goldstein is the author of more than 70 scholarly publications in the areas of population and medical genetics and sits on the editorial boards of Current Biology, Annals of Human Genetics, Molecular Biology and Evolution, and Human Genomics. He is the recipient of the UK Royal Society’s Wolfson Research Award for human genetics and genomics.
The center is going to focus on how genetic variation influences response to drug treatment and contributes to problems involved in drug-drug interactions. What do you plan for the Center to focus on initially?
A major focus for our work has been, and will continue to be, the response to anti-epileptic drugs. In that work, it’s not motivated primarily by thinking about drug-drug interactions, it’s motivated by just optimizing the way particular drugs are used in epilepsy, and motivated in particular by trying to understand the biological basis for refractory epilepsy, which is the big issue in epilepsy.
But, if you’re particularly thinking about drug-drug interactions, we’ve got a direction we’re trying to take to help understand when drug-drug interactions cause particular problems. So, it’s well known that environmental triggers change the expression of drug-metabolizing enzymes — of course — and that that can create a problem.
But what has been much less studied is whether individuals differ in the degree to which they respond to, for example, inhibitors of certain drug metabolizing enzymes — like CYP3A4, which metabolizes all sorts of drugs and is responsible for most drug-drug interactions.
So, one of the things that we’ll be doing is actually systematically screening for genetic differences among people that influence how much their enzymes change in response to environmental inhibitors.
So are you going to be doing much gene expression analysis?
No, actually. What we’ll do is characterize germ-line genetic differences among people, and relate that to responses to drugs, and also other environmental inducers and inhibitors.
So that’s the main focus of the center — characterization of germ-line variation, and then relating that, in the most general sense, to drug response, with a particular focus on epilepsy.
But also, as I said, systematic screens for gene variants that influence how enzymes are, in particular, inhibited, which of course is the more serious issue, but also induction as well. One of the things, for example, that we’re working on is the tolerance phenomenon with carbomazapine — so we’re trying to understand dosing requirements for the anti-epileptic drug carbomazapine. And one interesting thing there is that carbomazapine induces the enzyme that metabolizes it.
So, that’s one of the reasons it’s thought — although it’s been poorly studied — that patients need progressively higher doses of carbomazapine. So, if you start out a patient on carbomazapine with the dose that they ultimately need, it’ll knock them over — they can’t tolerate it. But after time goes by, they can tolerate it better, and in fact need a higher dose. One component of that is thought to be induction of the enzyme. But it’s likely that different individuals induce the enzyme to different degrees, and that source of genetic variation is differential response to inducers and inhibitors — it’s actually been hardly touched.
It’s a really striking thing about pharmacogenetics — everybody knows you’re not supposed to drink alcohol, people know that there are problems with grapefruit juice, and various kinds of drugs — some of which are over-the-counter. But this has applied sort-of blanket warnings, but it’s quite clearly the case that not everybody is responding to these in the same way. And so, it’s almost certain that what you’ve got is a subset of the population that is particularly sensitive where you’ve got particular concerns.
Do you expect this to depend most on germ-line genetics?
Or in the genes that mediate their induction. So this is one interesting thing about the history of pharmacogenetics. The actual enzymes themselves have been studied to death. We know them inside and out, so you will find a new mutation here and there that does something, but the genetic differences that influence how they are expressed — both cis-acting and not-cis-acting genetic differences — [those] that are actually in the gene or in some other gene that activates [it] — those have not been systematically studied.
There are ways that you can go about doing that — for example, CYP3A4 is responsible for more drug-drug interactions than anything — there are well-known environmental inducers of CYP3A4. But what’s not been done ever, is to take people and test their activities at baseline, then give them environmental inducers and inhibitors that are completely safe — like St. John’s Wort or grapefruit juice, respectively an inducer and an inhibitor — and then check to see whether there are genetic differences in the activating genes that turn CYP3A4 ‘on’ in response to St. John’s Wort, for example, that cause some people to induce it more and some people to induce it less.
And then, the idea is, finding some of these genetic differences would identify individuals that are particularly susceptible to drug-drug interactions or drug-environment interactions if it’s not another drug.
And there haven’t been screens for that kind of genetic variation, even though we’ve worked very, very hard to find structural variation in these genes that influence the activity of the enzymes. We know that these things have very, very fine, complicated regulation, and we haven’t looked for that kind of variation.
So, there will sort-of be two main lines of work in the Center. More-standard pharmacogenectics, where you get patients that respond differently to a given medicine, and then ask if you can identify the underlying genetic causes that are associated with that variation in response — that will be a major component of what we do, and epilepsy will figure prominently in that, because we’ve got a history in working with that. Another arm of work in the Center will be the use of these kind of challenge studies, not with patients, but with healthy volunteers, to try to understand genetic variation in processes that are relevant to drug response — like inhibition and induction.
Is the Center going to be looking for novel regulatory elements on its own? Or will you be testing things that are known to exist?
We will largely test things that are known to exist, although there will be some discovery aspects to what we do. You can carry out some of the association work without a hypothesis. So, for example, you can take the entire 1 Megabase interval upstream of a gene, and represent the genetic variation in that interval, and relate that to, say inhibition or induction. And if you find something, you can try to identify what precise polymorphism is responsible, and that might identify a new binding site for a transcription factor, for example.
But the core of the work will be looking through things we already know about. [Pregnane X receptor], for example, encodes the transcription factor that is responsible for induction of 3A4 in response to St. John’s Wort. So St. John’s Wort is a ligand for PXR, PXR is a nuclear receptor, which then activates CYP3A4.
So that’s a well-known mechanism, and we’d take that into account in the design of the study that searches for gene variants that might influence how much CYP3A4 is either induced — if it’s St. John’s Wort — or inhibited.
Those will be the main directions. Then the challenge side, in addition to inhibition and induction, we’re also interested in using that kind of a framework to look at responses to medicines, but not in their usual clinical setting. That’s another interesting thing — there’s not that much of that kind of work done in pharmacogenetics.
But for a lot of drugs, it’s quite safe to give them to healthy volunteers [and test their effects]. For example, many antihypertensives are safely given to healthy volunteers — in fact a lot of drugs are. That’s actually a particularly good way to ask whether multiple polymorphisms interact to influence response to a drug. So, one of the problems of pharmacogenetics right now is that it has been very difficult to get large enough sample sizes.
But if you screen very large population groups, and identify exactly those genotypes out of a large population group, then call those individuals in and see how they respond to certain medicines, then you can test for how genetic differences interact.