M. ATAROD/SCIENCE

As epilepsy drugs fail nearly one-third of patients, scientists seek root causes of seizures

Andrea VonMarkle arrived in Madison by helicopter ambulance 2 years ago, her life hanging in the balance. One month earlier she'd been a healthy 21-year-old juggling community college, waitressing, and weightlifting at a local gym. But after several weeks of feeling vaguely ill and forgetful, she was struck by a terrifying crisis.

On New Year's Eve, VonMarkle's aunt had returned to the home they shared in northern Michigan to find her niece in trouble. "The door was open, and our dog was running down the street," VonMarkle says of the scene that greeted her aunt. "I just kept saying, ‘I don't know what's going on, and I don't know why I don't know.’" Then she started seizing.

The seizures, which VonMarkle had never experienced before, didn't stop. Doctors at a local hospital were unable to quell her brain's electrical storm with powerful antiseizure medications. Because unremitting seizures can destroy brain tissue and damage other organs, the doctors put her into a medically induced coma.

"They didn't know what to do with me," she says, "so they flew me to Madison," where the University of Wisconsin hospital had more resources and staff. VonMarkle, unconscious for weeks, wouldn't find out until much later what a stroke of luck that was. She became one of the first people whose sudden-onset, life-threatening epilepsy would be treated in a whole new way: not with standard antiseizure medications, but by disabling the deeper roots of the disease. For her, that meant a drug normally used for arthritis that seemed to soothe the inflammation powering her seizures.

Although most cases of epilepsy don't carry such high stakes, struggles to find a workable treatment play out daily. Afflicting about one in 26 people, epilepsy is defined by its most visible symptom: seizures, caused by abnormal electrical activity in the brain. About two dozen medications are commonly prescribed for epilepsy, but roughly 30% of cases are classified as drug resistant because even with treatment, patients continue to have seizures. That percentage hasn't budged in decades despite many new drugs—suggesting a pivot in tactics is desperately needed.

Andrea VonMarkle was felled by a rare epilepsy syndrome until an arthritis drug subdued brain inflammation, the apparent cause of her seizures.

ADAM JOHNSON/BROCKIT

"People assume that since there's so many medications for epilepsy out there, it's taken care of," says Vicky Whittemore, a neuroscientist and program director at the National Institute of Neurological Disorders and Stroke (NINDS) in Bethesda, Maryland. "It's not."

After years of frustration, epilepsy researchers are shifting from targeting the seizures to seeking their cause, sometimes one patient at a time. Much about the condition remains a mystery, including why antiseizure medications fail so many people, among them more than 1 million in the United States. Yet leaps in gene sequencing and better animal models are changing how scientists and doctors understand, study, and—sometimes—treat the disease. The patients who stand to benefit include newborns with genetic syndromes who start to seize hours after birth and go on to suffer ruinous developmental delays, 10-year-olds who wake up wondering whether seizures will strike at school and battle difficulties with memory and learning, and adults whose head injuries put them at risk for epilepsy years later (see sidebar).

In San Francisco, California, one lab boasts rows of petri dishes filled with eyelash-size zebrafish larvae; they are genetically engineered to match some of the dozens of epilepsies that strike young children and are used to test new treatments. In Belgium, industry scientists fashioned a medication for one of the most common epilepsies, focal epilepsy, on the basis of experiments in drug and disease biology. And some researchers hope inflammation—the key to VonMarkle's condition—will help them crack other treatment-resistant cases.

"It used to be that people considered epilepsy one big bucket," Whittemore says. "What we now know is there are likely so many different causes, different ways you can get to a seizure." Converting that awareness to targeted treatments is the next frontier.

Epilepsy strikes newborns, the elderly, and everyone in between. Some cases are genetic—researchers have linked more than 100 genes to the disease. Other forms are caused by an infection such as encephalitis or by structural malformations in the brain, whether congenital or due to trauma or a stroke. Often, the cause is unknown. Despite the varied paths leading to epilepsy, all share one endpoint: disruption in electrical missives between neurons, which manifests as seizures.

For about two-thirds of patients, drugs that modify the function of ion channels—proteins in the membranes of neurons and other brain cells that transmit electrical signals—fully restrain seizures. For the rest, drug-resistant epilepsy can be crippling. "Through the day, I would think, ‘What if I have a seizure right now, what if I have one in this moment?’" says Kristen Grip, who was almost 14 when she was diagnosed with focal epilepsy, which affects discrete regions of the brain. About 60% of epilepsy patients share this form of the disease. Medications initially quieted her seizures. But when treatment stopped working, seizures—up to 12 per month—punctuated her high school classes and basketball practice. While cycling through medications in a vain effort to find something that worked, she battled drug side effects that included personality changes and exhaustion. "I felt like I wasn't living," she says.

At 16, Grip underwent surgery in which part of her brain was excised—an option because her seizures were concentrated in a region that she could live without. Now 23, she lives in Boston and is seizure-free.

The dilemma for many patients with tough-to-treat epilepsy is that for the most part, "The new drugs do what the old drugs do," says Amy Brooks-Kayal, a pediatric neurologist at Children's Hospital Colorado and the University of Colorado in Denver. "We play whack-a-mole—we've got a bunch of different epilepsies popping out of different holes caused by different things, and we try to whack all of them with the same hammer."

The faces of epilepsyEpilepsy is defined by its most visible symptom—seizures—but varies in its causes and manifestations. Almost one-third of patients continue to sufferseizures despite taking medication. More than 100 genes have been linked to epilepsy, and some variants cause devastating symptoms in children. 1. Genetics Infections such as encephalitis can sometimes trigger epilepsy, perhaps by inducing brain inflammation. 2. Infection Traumatic brain injury and other forms of brain damage, such as stroke, can cause epilepsy. 3. Injury In about half of cases, the causeis a mystery. 4. The unknown Neurons relay messages to one another whenelectrical impulses trigger neurotransmitter release.Neurotransmitters encourage other neurons to fire orhinder their firing, keeping brain activity in check. Normal firing Scientists are working hard to understandepilepsy’s many mechanisms in hopes ofdeveloping more targeted treatments. Varying causes An imbalance in neurotransmitter signaling causesneurons to become overexcited or lose theirinhibitory control, causing abnormal bursts ofelectrical activity, resulting in a seizure. Misfiring Brain activity Time Brain activity Time Controlledbrain messaging Uncontrolledbrain activity
C. BICKEL/SCIENCE

Investigators are searching for better hammers in a gleaming glass and brick building in Salt Lake City, where hundreds of cages line the walls. This is the Epilepsy Therapy Screening Program at the University of Utah, an NINDS effort that since 1975 has used rodents to test more than 32,000 potential antiseizure drugs, including many on the market today.

Like other researchers around the world, the Utah team traditionally studied animals with healthy brains, inducing seizures one at a time with chemicals or electroshock. But as mail carriers dropped off vials of experimental drugs designed by companies and academic labs, Karen Wilcox, the pharmacologist who runs the Utah program, fretted that healthy animals might not help struggling patients. "If we never used brains that had epilepsy, we might never fix this refractory problem," she says.

So about 4 years ago, following recommendations from an NINDS working group, the program was overhauled to include animals that had chronic or spontaneous seizures. Some rodents have infection-induced epilepsy, whereas others have experienced a prolonged seizure, which leaves their brains prone to later spontaneous seizures.

These more sophisticated rodents cost five to seven times as much, because using them is more laborious and time-consuming than traditional animal models. Staff must monitor the animals around the clock and review reams of computer data chronicling their brain waves. "You don't know when they're going to have their seizures," Wilcox says. "They might have two a day, three a day, skip days—like people." Staff test only the most promising drugs on them; scientists are exploring new compounds these animals have helped identify.

The revamped rodent models reflect growing insight into epilepsy's varied roots. For every patient, "Something happened to this brain that led to this," says Annapurna Poduri, a pediatric neurologist who also runs a lab at Boston Children's Hospital. Poduri witnesses heartbreak all too often in her ninth-floor clinic: severe epilepsies in very young children that respond poorly to treatment. In about 40% of these children, genetic testing reveals a culprit, and in scattered cases, physicians are reaching for unconventional treatments. A toddler in Poduri's clinic received the Alzheimer's drug memantine because in lab studies, it partially corrected the effects of the genetic defect he carried. A little girl went from "seizing, seizing, to walking, running, jumping," Poduri says, after genetic testing revealed a rare mutation whose biochemical effects pointed to the dietary supplement uridine as a potential treatment. Still, such arresting stories are "the extreme," she cautions. "Maybe one in 100 you find one of these."

[For every patient,] something happened to this brain that led to this.

Annapurna Poduri, Boston’s Children Hospital

To move beyond anecdotes, neuroscientist Scott Baraban at the University of California, San Francisco, is hatching thousands of genetically engineered zebrafish. Members of each clutch carry an epilepsy-causing mutation. Baraban is especially interested in one genetic epilepsy, Dravet syndrome. It triggers prolonged seizures, developmental delays, and a host of other problems, and it is profoundly resistant to antiseizure treatments. Baraban was startled to find that an old antihistamine called clemizole blunted seizures in a fish model of Dravet, and he and his colleagues found that it worked by binding to serotonin receptors that mediate neuron excitability.

Clemizole was discontinued in the United States and elsewhere in favor of newer antihistamines, but a drug for weight loss, lorcaserin, binds to the same receptors and had the same helpful effect on Baraban's fish. The discovery surfed straight "from aquarium to bedside," Baraban says, as physicians offered the drug to five patients. Seizures were considerably diminished in four of them. A company Baraban co-founded is redeveloping a pediatric version of clemizole and recently finished a small clinical trial of the drug in Dravet.

Dravet is also the target of an eagerly awaited therapy, slated for a trial opening next year. The treatment, a form of RNA called antisense, was developed by Stoke Therapeutics in Bedford, Massachusetts, and scientists at the University of Michigan in Ann Arbor. Most children with Dravet have one normal copy of the culprit gene and one defective copy; in mice, the therapy binds to RNA. In doing so, it increases production of RNA and the healthy protein, compensating for the underlying defect.

"We're just praying these treatments have good results, they're safe, and they're available in time for Anna," says Kim Odlaug, an advertising executive in Chicago, Illinois, whose 2-year-old daughter has Dravet. In children with Dravet, developmental delays often surge between the ages of 2 and 5, and Odlaug worries about what's to come for Anna, who performs acrobatics on the family's furniture and loves assembling puzzles. Because most of Anna's more than 30 seizures—three of which left her on a ventilator—trace to excitement or illness, her parents isolate her from other children and activities that she loves, such as music class. Odlaug hopes that if the therapy appears safe, Anna can receive it in a clinical trial, because approval is uncertain and could take years.

But what of the many children with epilepsy for whom gene variants are suspected but DNA panels come back clean? Poduri and others are studying brain tissue from some of those children, usually donated after surgery. Some samples harbor DNA changes that don't appear in the blood and must have happened spontaneously during development. Such changes are harder to identify but may prove powerful in steering treatment, especially if they're shared across patients. Surprise discoveries include mutations in tumor genes and in a gene that helps sugar molecules bind to proteins.

Recently, Poduri began to collaborate with colleagues in Melbourne, Australia, who are assessing whether such DNA changes are detectable in spinal fluid. "Precision medicine," she says, "has to start with precision diagnosis."

That's not always possible today. Many cases of epilepsy can't be traced to an easily treatable cause, leaving patients like Grip to try drug after drug. The inability to hit on a precision diagnosis for many patients has led some drug company scientists to try another tack: understanding why the drugs they make don't always work and figuring out how to improve on them.

In 1999, a medication called levetiracetam (Keppra) was approved for some cases of focal epilepsy. But the drug doesn't help everyone, and scientists at the company that makes it, UCB in Braine-l'Alleud, Belgium, wanted to understand why. Experiments revealed that levetiracetam selectively binds to a protein called SV2A, which belongs to a family of proteins called SV2; the drug's binding affects release of neurotransmitters in the brain, which in turn can stanch seizures. Collaborating with scientists at the University of Liège in Belgium, the UCB team learned that another protein in the family, SV2C, is markedly elevated in some brain regions of patients with drug-resistant focal epilepsy. Months of study yielded another clue: Although levetiracetam prevents errant release of neurotransmitters, the drug was more potent when combined with compounds that target certain neurotransmitters after they're pumped out.

With that knowledge in hand, UCB scientists created padsevonil. The experimental drug acts both before and after neurotransmitters are released between nerve cells, and it targets the effects of all three members of the SV2 protein family. The company is now running two phase III trials, involving more than 1000 adults with drug-resistant focal epilepsy. "I'm not saying it's the solution to drug resistance," says Henrik Klitgaard, who consults for UCB, where he was a vice president and led epilepsy drug research for several years. But it's a "rational approach."

Other leads come from evidence that inflammation can drive some forms of the disease. Powerful anti-inflammatories such as steroids occasionally ease seizures in drug-resistant epilepsy. The ketogenic diet, an extremely high-fat, low-carbohydrate diet, can sometimes help, as well. Some researchers theorize that the diet works in part by subduing inflammation.

It used to be that people considered epilepsy one big bucket. What we now know is there are likely so many different causes, different ways you can get to a seizure.

Vicky Whittemore, National Institute of Neurological Disorders and Stroke

Two scientists pushing the epilepsy-inflammation connection forward are Eleonora Aronica and Annamaria Vezzani. In the late 1990s, Aronica, a neuropathologist at the University of Amsterdam, and Vezzani, a neuropharmacologist at the Mario Negri Institute for Pharmacological Research in Milan, Italy, grew intrigued by the presence of inflammatory molecules in brain cells in affected animals and people. Brains of fetuses that had died and had a genetic disease called tuberous sclerosis complex, which carries about an 80% chance of epilepsy, showed inflammation. So did the brains of animals with induced seizures that Vezzani was studying. When inflammatory molecules persisted in an animal's brain, they made future seizures more likely. After years of work, the pair learned that the inflammatory molecules increase neurons' excitability.

Could those links to inflammation inspire new treatments? Four years ago, a chance encounter allowed Vezzani to test the idea. At a conference, she met a doctor from the Mayo Clinic in Rochester, Minnesota. The hospital was then trying to save a toddler whose disease looked awfully like that of Vezzani's mice—the little girl had unyielding seizures that began after a fever and infection. The child's diagnosis was an extremely rare but catastrophic form of epilepsy called febrile infection-related epilepsy syndrome (FIRES), which strikes about one in 1 million people.

When asked, Vezzani recommended that doctors try an anti-inflammatory drug called anakinra, which is approved for rheumatoid arthritis. In the rodents she'd studied with brain inflammation, anakinra minimized seizures.

It worked in the little girl, too. "We had an outcome that we've never seen before," says Eric Payne, a Mayo pediatric neurologist who treated the child. She turned 7 in September, and a mild language delay and occasional seizures are the only signs of a brain that nearly destroyed itself.

Two years later came VonMarkle. "I heard about this woman in the adult neurocritical care unit who was seizing, seizing, and wouldn't respond," says David Hsu, a pediatric neurologist at the University of Wisconsin. He had trained with Payne and knew about the girl. Hsu snagged VonMarkle's medical chart, which revealed that her tale began with headaches and a fever, and she, too, was diagnosed with FIRES. At Hsu's urging, VonMarkle was started on anakinra. She had been in an induced coma for more than 1 month, but within 24 hours of starting the drug, her seizures stopped. Within another day she was weaned out of the coma.

A therapy such as anakinra embodies epilepsy's next chapter: treatments that, though not always fully understood, seem to disable the deeper mechanisms of disease in narrow groups of patients.

For anakinra, mysteries endure. The little girl Payne cared for produced a defective version of an immune protein, which anakinra supplemented, he and his colleagues later reported. That finding might help explain the drug's potency for her. VonMarkle's doctors haven't determined whether she shares this aberration or harbors another—or whether the drug's lifesaving effects for her will linger unexplained. "I was ecstatic" the drug helped her, Hsu says, but "we don't know why." At least, not yet.

Worldwide, an unknown number of children have received anakinra, including at least 25 with FIRES who are part of an unpublished cohort and have experienced varied results. Payne is trying to understand whom anakinra can help and why, and he is hunting for neuroinflammation in other tough-to-treat cases of epilepsy. He and others caution that although inflammation is worth chasing, no single drug will be a cure-all.

VonMarkle was among the lucky, though it didn't feel that way at first. After waking up, she weighed just 43 kilograms and spent another 6 weeks in the hospital. Nearly 2 years later, her short-term memory is sluggish and she hasn't returned to college, but she is waitressing, exercising, and working as a makeup artist.

She still has occasional seizures but finds them manageable. "Just reassimilating to normal life was very difficult," VonMarkle says. But, she adds, "I'm very fortunate to be able to do what I'm doing."