Like most two-and-a-half-year-olds, Savannah Salazar could count to three. She knew a handful of colors and was becoming conversant with the subjects of dogs, cats, bikes, cars, and buses. She was potty-trained, and she loved to page through books and put together puzzles. She was on the right track for a toddler, hitting all the big milestones.
When Savannah’s parents, Ruben Salazar and Tracy Dixon-Salazar, put her to bed one night in 1995, they had no reason to believe she wouldn’t keep hitting milestones. Then they awoke in the middle of the night to the sound of their daughter choking. They rushed to her bedroom and watched in horror as Savannah’s limbs stiffened and trembled — her face white, her lips blue, bubbles of saliva at the corners of her mouth.
“I have no words to describe the terror of that moment,” says Dixon-Salazar.
When the paramedics arrived, whatever had come over Savannah had passed. The first-responders checked to make sure that Savannah’s airway was clear and told the terrified parents, “What you described sounds an awful lot like a seizure.”
Two weeks later, it happened again. But it took almost three years, hundreds more seizures, and eight normal EEGs (electroencephalograms) followed by one abnormal one for doctors to diagnose Savannah with Lennox-Gastaut syndrome — an epilepsy that includes multiple seizure types and cognitive impairment and arises before a child is 8 years old.
For Dixon-Salazar, a diagnosis wasn’t enough. She wanted to know why. The stay-at-home mom, a high-school graduate and former U.S. Navy airplane mechanic, started reading scientific papers about the causes of epilepsy. Unable to understand the papers, Dixon-Salazar decided to take some English classes at Grossmont College, near her home in San Diego. While she excelled at English, she still found herself unable to read the scientific literature. Frustrated, she registered for a biology class.
“Then I just fell in love with science,” she says. She spent the next four years working toward an associate’s degree. Next, she pursued a bachelor’s in physiology and neuroscience at the University of California, San Diego, before going on for a PhD in neurobiology.
During the more than 12 years that Dixon-Salazar was studying, Savannah’s epilepsy continued to get worse. “I couldn’t help her. There was nothing I could do for her,” she says. Savannah took multiple medications, and still she was having 300 seizures a month, some lasting up to four days at a time — a seizure known as status epilepticus. “Studying became a coping mechanism,” she recalls.
After completing her PhD, Dixon-Salazar began postdoctoral research in epilepsy genetics. It was 2011 and exome sequencing — the process of sequencing all of the protein-coding genes in the genome — had just become commercially available. Dixon-Salazar began to build a database of epilepsy exomes in order to uncover possible genetic causes of the condition. She saw a door open for her own daughter’s untreatable epilepsy when her adviser asked if she’d like to enroll Savannah in the study.
Like Savannah, some 50 million people worldwide have epilepsy. About 20 to 30 percent of them developed the condition after an event such as a stroke, head injury, or brain tumor. For the remaining 70 to 80 percent, researchers believe, epilepsy is genetic. Genome sequencing, as Dixon-Salazar was doing in the lab, has uncovered some 300 genes linked to the condition. About 50 to 100 of the genes directly cause epilepsy, while the remaining ones may combine with other factors to cause the seizure disorder.
Discovery of these genes is pushing epilepsy research forward in an unprecedented way. In some cases, gene variants can help predict the severity of the epilepsy and the course it will take. Some genes point doctors toward the most appropriate treatments and steer them away from others. The genetic cause of a particular epilepsy can inform family planning. Researchers anticipate that one day genomic research will lead to the development of more treatments that, rather than treat the seizures and cognitive problems, target the genetic defect that causes these issues in the first place.
“We are experiencing the dawn … the first glimmer of light on the horizon that shows us that a new day has begun for precision medicine in epilepsy,” says Daniel Lowenstein, the director of the UCSF Epilepsy Center and the principal investigator of the Epilepsy Phenome/Genome Project. “Now there are a growing number of examples where finding the target is leading to a modification of treatment and, in some cases, it has a beneficial effect.”
Epilepsy is a broad spectrum of seizure disorders of varying severity and seizure type. The common link is recurrent seizures that seem to arise spontaneously, as opposed to an isolated seizure that might occur after an exposure to an illicit drug or an exceedingly high fever in a child. The seizures are caused by a misfiring of neurons that transmit an electrical current in the brain. While seizures seem unprovoked, some people can identify triggers for their seizures, such as stress, bright lights, sickness, or sleep deprivation. They may be able to use these triggers to prepare for and prevent future seizures.
In the U.S., more than five million people have had a diagnosis of epilepsy or a seizure disorder. That’s just under 2 percent of adults and about 1 percent of children. Epilepsies can be classed broadly as generalized or focal. Generalized epilepsies affect multiple parts of the brain, and seizures typically come with loss of consciousness. Focal epilepsies are concentrated in one part of the brain. People may remain conscious during a focal seizure. In some cases, people around them might not even notice the seizure. Seizures can be extremely frequent and continue for a lifetime, or they can be few and far between and eventually stop.
Epilepsies may also be classified by syndrome. The more than two dozen epilepsy syndromes are based on types of seizures; other associated symptoms such as cognitive impairment, autism, and vision impairment; EEG patterns; and age of onset.
It can take some time and many tests, which may include EEGs, MRIs, CT scans, spinal taps, EKGs, or a sleep study, to diagnose the type of epilepsy one has. “A delay in diagnosis time of about three years is pretty common,” says Dixon-Salazar. Medicine controls seizures for about 70 percent of people. For those who don’t respond to conventional anti-seizure medications, drugs that target specific biochemical pathways can sometimes help.
The most severe epilepsy syndromes, including Lennox-Gastaut syndrome, are epileptic encephalopathies. These forms of epilepsy start in infancy or early childhood. They include multiple seizure types that often don’t respond to medication; they often include severe cognitive, behavioral, and neurological problems; and they can cause premature death.
Savannah’s seizures didn’t respond to medication. And after she began having them, she stopped hitting those milestones. She didn’t learn any more colors or to count past three.
When Dixon-Salazar sequenced her daughter’s exome as part of her postdoctoral research, she learned that Savannah had none of the roughly 200 gene variants that were known at the time to be associated with epilepsy. But she kept looking. “It’s what I call ‘the crazy mother analysis,’ ” she says. Savannah had 300 gene variants that, while not confirmed to be associated with epilepsy, were unusual and assumed to be damaging. The self-proclaimed crazy mother divided those 300 genes into groups according to their function or the pathways in which they signal.
In the genes that belong to the calcium-signaling pathway, Savannah had a whopping 25 mutations. None of the genes had previously been associated with epilepsy, but calcium channels had been implicated in epilepsy for several years already. Dixon-Salazar believed that some of these 25 gene mutations could be strongly related to her daughter’s epilepsy, while others might not have any effect at all.
“But the fact that she had 25 mutations in the calcium-signaling pathway suggests an overall calcium-signaling dysfunction, which is why I looked deeper into each of the 25,” she says.
Dixon-Salazar believed the dysfunction was that too much calcium was getting to Savannah’s brain cells, over-stressing the cells and causing the seizures. When she consulted with a geneticist and her daughter’s pediatric neurologist, both agreed her hypothesis was reasonable. Dixon-Salazar knew of heart medications already on the market that blocked calcium channels. She convinced Savannah’s doctor to try her on verapamil.
Until that time, Savannah had up to 10 seizures a day. “Within two weeks on verapamil, she had a 95 percent reduction in seizures and a 100 percent reduction in status epilepticus,” Dixon-Salazar says. It was a targeted therapy that seemed to get right at the genetic mechanism of the seizures.
After Savannah’s seizures stopped, she started making progress again. Now 23 years old, she talks with difficulty, but she’s learning to write and, like many girls her age, she watches music videos on her iPad.
That doesn’t mean everyone with a calcium channel defect is now taking verapamil. Savannah’s 25 calcium-related mutations are highly unusual. It would be unlikely to find the same ones in someone else and get the same response to the drug.
“It’s dramatic seizure reduction in one child,” says Lowenstein. “There are a host of other genetic factors that influence whether or not a person’s going to respond to a particular drug.”
Other ion channels — potassium and sodium channels — are also increasingly associated with epileptic syndromes. Ion channels regulate the excitability of the central nervous system. Some mutations in the gene KCNQ2 can close potassium channels and cause KCNQ2 epilepsy, a severe epileptic encephalopathy that tends to arise in the first months of life. The seizures brought on by this form of epilepsy may stop, but developmental delays remain. Children with this mutation may be nonverbal, physically impaired, and show signs of autism.
Despite its potential to cause intense side effects in some patients, the anticonvulsant retigabine can help open potassium channels and stop seizures in some individuals with a KCNQ2 mutation. While the manufacturer announced last September that it will halt the production of retigabine for commercial reasons, patient groups have been pushing hard to keep the drug available for compassionate use.
Interestingly, an experimental drug to treat tinnitus — ringing in the ears — could also treat KCNQ2 seizures. It turns out both conditions are driven by the same gene. The drug’s developers have plans to get it into clinical trials.
“A lot of us feel that our children have tinnitus, but they can’t tell us because they’re nonverbal,” says Scotty Sims, the executive secretary of KCNQ2 Cure and the mother of a 5-year-old daughter with KCNQ2 epilepsy in Denver. “Our daughter always covers her ears and likes constant noise — and people with tinnitus like that, too, to drown out the tinnitus.” She says that even if this experimental therapy proves ineffective, it’s that kind of out-of-the-box thinking about possible genetic markers that might drive drug development.
A number of epileptic encephalopathies, including Dravet syndrome and GLUT-1 deficiency, improve with a ketogenic diet. This is a high-fat, low-carb diet similar to, but far more restrictive than, the Atkins diet. It’s believed that depriving the brain of sugar, and forcing it to burn fat, can stop seizures in these types of epilepsies.
Genomic sequencing continues to uncover genes that cause epileptic encephalopathies, thanks in large part to projects like the Epilepsy Phenome Genome Project, which collected and analyzed the genomes of 4,000 people with epilepsy worldwide.
“Through gene testing, we can now find the gene that’s causing 40 [or] maybe even 50 percent of epileptic encephalopathies,” says Lowenstein. Discovery of additional genes, researchers hope, could lead to more targeted therapy.
“The goal in basic epilepsy research,” says Jeffrey Noebels, a professor of neurology, neuroscience, and molecular and human genetics at Baylor College of Medicine in Houston, “has been to identify the specific cellular circuits that are impaired and somehow repair those circuits or modulate them or stabilize them. Knowing what gene you’re dealing with can help us get there.”
Genetic information about epilepsy can also steer people away from drugs that might be harmful to them. Just as genetic variants can predict whether a particular epilepsy might respond well to a drug, variants can predict dangerous responses to drugs, too.
The gene variant HLA-B*1502, which is most prevalent in people of Chinese descent, is associated with an increased risk of a toxic reaction to the anti-seizure medication carbamazepine and related drugs. “If someone who is Chinese needs to be put on a drug, I’ll probably avoid using carbamazepine and related drugs, even without knowing their particular genotype,” says Lowenstein. “But if it’s a situation in which carbamazepine would definitely be the preferred drug, then doing the genotyping could be very worthwhile.”
Parents of children with epilepsies driven by the SCN8A gene have reported that their children got worse after taking the anti-epileptic Keppra (levetiracetam). The drug is now not recommended for this group. In PCDH19-driven epilepsies, most prevalent in females, “there are definitely certain anti-seizure drugs that you want to avoid, and there are some that have been shown to be beneficial,” says Lowenstein. In particular, people who have this form of epilepsy might benefit most from clobazam or bromide.
Some epilepsies can lead to death. “The most common cause of death in epilepsy is sudden unexpected death — not accidents, not drowning, not suicide — just being found dead in bed,” says Noebels, who is also a principal investigator in one of the National Institutes of Health’s Centers for Sudden Unexpected Death in Epilepsy (SUDEP) Research.
Noebels’s lab and others have identified genes that may predict who is at greatest risk of sudden death. “These genes are expressed not only in your brain, where they can cause epilepsy, but in your heart, where they can also cause abnormal and irregular heartbeats,” he says. One such gene is KCNQ1. Identifying variants in these genes could allow doctors to consider treating the potential heart problem, which is often unrecognized, and to reassure the vast majority of other patients with epilepsy but without such variants that they are not at increased risk of SUDEP.
Genetic insight into the cause of epilepsy doesn’t always point people to a better treatment. But genetics can provide a wealth of information to people with epilepsy and their families. Sometimes parents, like Dixon-Salazar, just want to know what caused the epilepsy, even if there’s nothing more to be done. Parents might also want to know the risk of having another child with epilepsy.
In epileptic encephalopathies, the majority of driver genes are de novo, Lowenstein says. That means the epilepsy-causing mutation appears for the first time in the person with epilepsy and wasn’t inherited from a parent. Recent research from Lowenstein and his colleagues has found that these de novo mutations tend to appear in a set of approximately 4,000 genes that are highly intolerant to variation. Mutations that occur in genes that are intolerant to variation are more likely to have detrimental effects. Knowing that a child’s epilepsy is the result of a de novo mutation can give parents confidence that they have little chance of having another child with epilepsy.
But that’s not true of all epilepsies. Family planning is one of the main reasons young adults seek counseling at the Adult Epilepsy Genetic Clinic at Baylor College of Medicine in Houston. “They are concerned about potentially passing on their epilepsy predisposition to their offspring,” says Alica Goldman, who runs the clinic.
Inheritance depends on many factors. Some forms of epilepsy are autosomal dominant, which means that if one parent has the variant, each of his or her children has a 50 percent chance of having the condition, too. Autosomal recessive epilepsies are passed on to biological children of two carriers 25 percent of the time.
“What’s more difficult to predict is when there’s a rare case of epilepsy in your family. The likelihood of you or your child developing epilepsy is increased, but it’s nowhere near 25 or 50 percent,” says Lowenstein.
Stephanie Rensland’s epilepsy is caused by hereditary cavernous angioma. The blood vessel abnormality, most common among people of Mexican descent, causes seizures. Rensland, who is Mexican-American, inherited the condition from her mother. She has aunts, uncles, and cousins with the condition as well. She has partial seizures several times a week.
“They’re like panic attacks, except much worse,” she says. “My body temperature changes; I get so scared I might jump in your lap. I do things I don’t know I’m doing, like grab at people or pull on my clothes.” During seizures, her husband has to lock the door of their Katy, Texas, home so she doesn’t run out.
Rensland dearly wanted a baby. “Why could my mom have one, and basically everybody else in my family but me?” Besides the 50 percent chance of passing the condition on to her child, Rensland’s doctors had told her that pregnancy could be risky. “I knew there were risks, but I just wanted a doctor who would be a little bit positive.”
Pregnancy could exacerbate her condition, doctors told her. And a generalized seizure could cause a miscarriage. But Goldman, Rensland’s epilepsy doctor, told her that they could manage her condition during pregnancy.
Rensland’s daughter is now 7 months old, and she’s scheduled to be tested for cavernous angioma soon even though fetal ultrasounds assured Rensland that her child would not be born with any of the seizure-causing angiomas — abnormal clusters of blood vessels typically found on the brain or spinal cord. Rensland doesn’t plan to have any more children. “I came out of this healthy, and if she doesn’t have it, I’m just going to be so happy. I wouldn’t want to risk passing it on to another child.”
If Rensland’s daughter has the seizure disorder, catching it early may help. Dixon-Salazar always wondered how life would have been different if Savannah had gotten the right drug to begin with. “I think if we had tried the verapamil by itself earlier, before all of the seizures and all of the medicine, it probably would have been sufficient, but at this point the brain is so damaged,” she says.
Genome sequencing at birth or sooner, followed by targeted treatments, could one day stop seizures before they start.
“For people who are living a life like we were living,” Dixon-Salazar says, “where their kid has been seizing for years, precision medicine can still work for you. There is hope.”