Research Holds Promise for New Epilepsy Treatments


by Elizabeth Norton Lasley

May, 2008

Epilepsy is the second most common neurological disorder, after stroke, affecting about 1 percent of the population—about 50 million people worldwide, including more than 3 million Americans of all ages. Seizures are caused by sudden storms of uncontrolled brain-cell activity and usually strike without warning, disrupting the lives of patients and their loved ones.

Because about a third of epilepsy patients do not respond to medication, researchers are trying to form a clearer picture of the way seizures develop in the brain. The quest to understand the underlying biology has already resulted in a nationwide study to find culprit genes and new drug targets. On the level of treatment, brain “pacemakers” similar to deep brain stimulation are now in clinical trials.

Nationwide Genetic Dragnet

Although epilepsy has been known since ancient times to be a hereditary illness, the first genetic mutation was identified only in 1995, and mutations found since then occur in a very small group of families.

To broaden the information base, the National Institute on Neurological Disorders and Stroke launched the Epilepsy Phenome/Genome Project in 2007. This five-year study comprises 13 U.S. research centers and will include 6,750 subjects. About half will be pairs of siblings, each of whom has epilepsy; the other participants will act as controls.

Detailed descriptions of seizures from each subject, imaging studies, electroencephalograms (EEGs) and reports of experiences with medication will establish the phenome, or characteristics of the illness. The genome, or DNA, will be taken from blood samples.

“This is the largest epilepsy genetics study ever attempted,” says Dan Lowenstein of the University of California, San Francisco, who heads the project along with Ruben Kuzniecky of New York University.

The ultimate goal of the project is to identify the genes involved in epilepsy and pinpoint which ones are active in different forms of the disorder, thus leading to better-targeted, more effective treatments. The information may also yield more immediate results.

“An incredible problem in treating epilepsy is that we can’t usually tell what the best drug is for a given patient, or who will suffer side effects,” Lowenstein says. Combining DNA profiles with a detailed picture of the disease’s characteristics will help predict how well a particular patient will respond to the medications already available. 

Tracing the Problem, Finding the Correction

Clues to what goes wrong in epilepsy have been found in a handful of mutated genes in the receptor for gamma-aminobutyric acid (GABA), the brain’s chief “inhibitory” neurotransmitter. GABA’s role is to slow down the activity of neurons. The bursts of chaotic signal transmission that characterize an epileptic seizure are due in part to a lack of inhibitory control.

Several of the most common treatments work by enhancing GABA activity; however, only about a third of patients with epilepsy can control their seizures through medication. One malfunctioning receptor, designated GABAA, may be one explanation.

Researchers have identified several mutations in which some of the proteins that make up the receptor are assembled from the wrong building blocks. Theoretically, medications could be developed to correct these structural abnormalities with pinpoint accuracy.

“In an ideal world, each mutation would lead to its own designer drug, ”says Bob Macdonald, head of the neurology department at Vanderbilt University. But given the small number of affected individuals (fewer than 10, in some cases) and the economic realities of drug development, this scenario is unlikely.

A more promising route is to find mutations that are not unique to epilepsy, or even the brain, for which medications are already available or under development. For example, one abnormality in the GABAA receptor results from a different type of mutation that mistakenly halts translation of the DNA’s code into protein, thereby producing a truncated form of part of the receptor.

Mutations of this type are a common thread running through many disorders—they contribute to up to 70 percent of all inherited diseases—and are attractive targets for new therapies. In muscular dystrophy, for example, a customized compound that “reads through,” or overrides, the premature stop signal has restored muscle function in mice and, in humans, has proved safe in a phase I clinical trial. The animal studies were reported in the May 3, 2007, Nature; results of the safety trial appear in the April 2007 Journal of Clinical Pharmacology.

“By targeting malfunctions common to many disorders, we may be able to take advantage of existing medications while working on customized therapies for epilepsy,” says Macdonald. “But we won’t know where the parallels are without a detailed understanding of epilepsy genetics.”

Beyond ‘Good’ and ‘Bad’ Genes

New insights into epilepsy genes reveal that although a “bad” gene may place an individual at risk for the disease, other genes may lessen the risk by preventing the biochemical disturbance.  In fact, even a disease-producing gene can sometimes have beneficial effects.

“In the brain, the dividing line between good mutations and bad is unclear, and two wrongs sometimes make a right,” says Jeffrey Noebels, a professor of neurology and molecular genetics at Baylor College of Medicine.

Noebels’s research is providing a comprehensive map of the ion channels involved in human epilepsy. Ion channels are gates in the cell membrane through which neurons exchange electrical impulses. (The GABA receptor is also an ion channel, but one that opens in response to a chemical messenger as opposed to a direct electrical signal.)

In a paper published in the November 2007 Nature Neuroscience, Noebels and colleagues developed a line of mice that carried mutations in two different ion channels, each of which plays a role in human epilepsy. One, a mutant potassium channel, leads to “tonic-clonic” seizures—the stiffening and jerking of the limbs also known as grand mal seizures. The other, a defective calcium channel, contributes to childhood “absence epilepsy,” or brief periods of blank staring.

“Normally, one would predict that adding two genes together might produce a complex seizure disorder with both seizure types,” says Noebels. “The genes in the study were chosen because they have mutually opposing effects in the excitability, or firing patterns, of nerve cells—canceling each other out at the electrical level.”

In mice carrying both mutations at once, the two “bad” genes to a large extent canceled each other out. Tonic clonic seizures decreased by about 60 percent, while absence seizures (detected in mice by EEG recordings and “behavioral arrest,” or sudden lack of activity) were virtually eliminated.

“Our work shows that the mere presence of a suspect gene does not predict the severity or even the occurrence of the disease. The whole profile of genes must be analyzed to accurately predict risk,” says Noebels. He adds that a clear picture of the underlying genetics is important for counseling families and prospective parents—“especially in the case of epilepsy, which still carries a heavy social stigma.”

Early-warning Devices

Particularly for patients whose seizures cannot be medically controlled, pinpointing gene activity is but one approach to finding better-targeted and more-effective treatments. Another method is leading to a new generation of pacemaker-type devices now in clinical trials, which may soon be able to cut seizures short or even prevent them entirely.

One, the Intercept Epilepsy Control System, is being developed by Medtronic Inc. of Minneapolis (the company that developed a deep brain stimulator for Parkinson’s disease).  A form of deep brain stimulation, the Intercept system uses electrodes surgically implanted in the thalamus, which are connected by a wire down the patient’s neck to a battery pack in the chest. The electrodes act as a real-time EEG, detecting abnormal electrical activity in the brain and sending a corrective signal to the thalamus (an area that regulates movement and is thought to play a role in controlling seizures).

A second device, the Responsive Neurostimulator System, is under development at Neuropace Inc. of Mountain View, Calif. It is essentially a microchip beneath the skull, connected to electrodesthat can be implanted in a variety of locations and deliver bursts of electricity to disrupt seizure activity.

Both devices have undergone safety trials and have been well tolerated, with few side effects aside from soreness due to the surgery. Most patients reported a reduction in the frequency and severity of seizures, though the studies were set up to assess safety only. The clinical trials are focusing on how well the devices work.

Other types of implantable devices under scrutiny can forestall seizures by cooling target areas or by squirting micro-amounts of medication into just the right spot.

The next generation of technology may be able to predict a seizure and stop it before it even begins, says Brian Litt of the University of Pennsylvania’s Neuroengineering Research Lab. “Seizures don’t just occur abruptly. Growing evidence suggests that they’re generated over time and involve whole networks of neurons,” says Litt.

Research from Litt’s lab contributed to both of the devices now in clinical trials. He is using microelectrodes to find the smallest unit of the brain that’s capable of generating a seizure. Working in collaboration with Gregory Worrell’s team at the Mayo Clinic and many other labs, Litt’s group is using high-frequency EEG recordings—“way beyond what hospital machines use, which are based on old pen-and-paper EEG machines”—to track the development of a seizure from a single neuron all the way to entire brain networks.

“Our hope is that with an understanding of the basic neurophysiology, we can spot ‘microseizures’ in tiny islands of brain tissue that would otherwise coalesce in to full-blown clinical events. This will provide the keys to better diagnostic and therapeutic devices,” Litt says.