Epilepsy & The Genome
Neurobiology Chairman Jim McNamaraThis article originally appeared in the May 2005 issue of GenomeLife.
It's like an all-star history team: Socrates, Alexander the Great, Joan of Arc, Napoleon, Lord Byron, Pope Pius IX, Dostoyevsky, Alfred Nobel, and van Gogh, just to name one imaginary starting line-up. What did they have in common? Epilepsy. According to the Epilepsy Foundation, 2.5 million Americans are afflicted with epilepsy, and ten times that number will experience a seizure at some point during their lives.
But despite the relatively high incidence of epilepsy, our understanding of what causes the electrical signals in the brain to become disrupted and provoke seizures remains incomplete. We are also unable to prevent it - current therapies treat the symptoms - not the causes - of epilepsy, much as insulin treats the symptoms of diabetes. What's more, 30 percent of patients don't respond to anti-seizure medication. Recent and ongoing work at Duke suggests that those treatment-resistant patients may have cause for hope.
Epilepsy: Of Genes and Drugs
Dozens of studies have demonstrated that genes are involved in epilepsy. However, as with other complex diseases like cancer and heart disease, single-gene versions of epilepsy account for only a tiny fraction of the disease burden. From a genome scientist's perspective, the rest is kind of a mess.
"That's exactly how I'd characterize it," says David Goldstein, Director of the IGSP's Center for Population Genomics & Pharmacogenetics. "But I would make one additional point: the nice thing about studying epilepsy is that you have access to the relevant tissue."
Goldstein is referring to patients whose epilepsy does not respond to medication and must be surgically treated. In such cases (~ 30 per year treated at Duke University Hospital), the brain tissue where the seizures originate is removed. "What that allows us to do is look at those cells and see what's different about them. That might help us to understand what's going on."
But building up large numbers of tissue samples for study takes time. For now, Goldstein is using a pharmacogenetic approach to get at the molecular basis of epilepsy and, he hopes, to influence the way anti-epileptic drugs are used. In a recent paper published in The Proceedings of the National Academy of Sciences, Goldstein and his former colleagues at University College in London identified variants in two genes that are strongly associated with the maximum dose needed to control seizures with two common medications, carbamazepine and phenytoin.
How did he decide which genes to check for such an association? "We looked at the very first genes you would look at if you were starting a pharmacogenetics project: the genes encoding the primary drug-metabolizing enzymes, the one encoding the transporters that keep the drug out of the brain, and the ones encoding the targets that the drug physically interacts with."
For Goldstein, the take-home message is to not overlook the "obvious genes," many of which will harbor variants in them that may be clinically meaningful. And because the majority of prescription drugs' modes of action are well understood or at least strongly suspected by pharmacologists, he expects that the approach he took with carbamazepine and phenytoin will be broadly applicable to other conditions and other drugs. "There's a minority of drugs where we just don't know how they work," he says. "But in most cases we have a very good idea where to start looking."
A Real Knockout
Goldstein's work on epilepsy finds him in good company. Jim McNamara, the Carl R. Deane Professor and Chair of Neurobiology, has been studying epilepsy for over 30 years. In the 1990s, his lab discovered genes that, when knocked out in mice, either caused epilepsy or made the animals more seizure-prone. Last year, he was somewhat taken aback to find a gene called trkB (pronounced "track B") that, when knocked out, actually prevents mice from becoming epileptic.
"Epileptogenesis is the process by which a normal brain becomes epileptic," he says. "And this is the only gene that's been demonstrated to be required for epileptogenesis in a mouse model. I figured that there would always be adaptive mechanisms in the mammalian brain that would circumvent epileptogenesis - you might slow the process but you'd never be able to stop it altogether. [The existence of] trkB really surprised me."
McNamara is hopeful that interfering with human trkB might offer a way of preventing epileptogenesis, be it in patients predisposed to sporadic epilepsies or in those with familial varieties. But first, he will need to determine whether disruption of trkB can prevent it in mice known to be epileptic, i.e., mice carrying mutations in a gene that has made them susceptible to spontaneous seizures. "If so, I think that would give us a good sense of the importance of the trkB pathway."
Synergy and Overlap
McNamara's interests are not limited to mouse models. He has been collecting material on human epilepsy families for some time and is looking forward to pooling his resources with Goldstein's. "We have [identified a genetic] locus and now we have to track down the genes within that locus. We're going to do this by analyzing additional families. It's a nice convergence because David was interested in this independently [before coming to Duke]."
Goldstein sees the collaboration on family studies with McNamara dovetailing seamlessly with his own work in pharmacogenetics. "It makes perfect sense to do those two things together because there's a great deal of overlap in the genes you test. If you think about it, one of the best sets of candidate genes for predisposition in epilepsy are precisely those genes that drugs act on and that [reside in] the pathways they affect. You can easily imagine that dysfunction in those pathways is one of the things that contributes to epilepsy."