Elusive epilepsy mutations begin to yield up their secrets

mosaic epilepsy mutations concept
Fawn Gracey illustration

Anti-seizure drugs don’t work in about a third of people with epilepsy. But for people with focal epilepsy, whose seizures originate in a discrete area of the brain, surgery is sometimes an option. The diseased brain tissue that’s removed also offers a rare opportunity to discover epilepsy-related genes.

Many mutations causing epilepsy have been discovered by testing DNA taken from the blood. But it’s becoming clear that not all epilepsy mutations show up on blood tests. So-called somatic mutations can arise directly in tissues like the brain during early prenatal development. Even within the brain, these mutations may affect only a fraction of the cells — those descended from the original mutated cell. This can create a “mosaic” pattern, with affected and unaffected cells sometimes intermingling.

One of the first such mutations to be described, by Ann Poduri, MD, MPH, and colleagues at Boston Children’s Hospital in 2012, was in Dante, a young boy who was having relentless daily seizures. The entire right side of Dante’s brain was malformed and enlarged, and he underwent a drastic operation, hemispherectomy, to remove it. Only later, studying brain samples from Dante and similar children, did Poduri find the genetic cause: a mutation in the gene AKT3. It affected only about a third of Dante’s brain cells. 

Ever-more-subtle mutations

Today, Dante is in his teens and seizure-free, and Poduri directs the Epilepsy Genetics Program at Boston Children’s. Through the Epi4K consortium, which aims to analyze DNA from more than 4,000 people with epilepsy, Poduri has been part of a team investigating a range of epilepsies, including those tied to brain malformations — from large ones down to smaller focal cortical dysplasias. The list of epilepsy-associated genes has now become too large to fit on a chart.

Still, Poduri suspected that more somatic mutations could be found in patients with focal epilepsy, even if no brain malformation was visible on imaging. She partnered with Erin Heinzen, PhD, a geneticist colleague at Columbia who in turn was partnering with a multi-institutional group studying adult epilepsy. Their joint study, published recently in the Annals of Neurology, brought together geneticists, pediatric and adult epileptologists, neuropathogists, neuroradiologists and neurosurgeons.

The collaborators studied 38 children with brain malformations and 18 adults with so-called non-lesional focal epilepsy (NLFE), meaning nothing was found on brain imaging. Both children and adults had undergone epilepsy surgery. Their brain tissue was analyzed with exome sequencing and targeted deep sequencing of a multiple  epilepsy- and malformation- related genes.

In some patients with apparently normal MRI scans, microscopic examination of brain tissue revealed subtle abnormalities — such as the sample at left, which shows columns of tightly aligned neurons characteristic of a mild cortical dysplasia. (Winawer MR; et al. Ann Neurol 2018; doi:10.1002/ana.25243

These studies revealed that two of the children with malformations and three of the adults with NLFE had mutations in a gene called SLC35A2.

“We didn’t think at the outset that we would necessarily find commonalities between these two populations,” says Poduri.

But the more they dug in to the five cases, the more similar and overlapping they seemed. For example, two of the three adults classified as NLFE turned out to have subtle cortical abnormalities that didn’t show up on MRI, but could be seen when the tissue was examined microscopically. The severity of epilepsy seemed to vary in step with the percentage of brain cells bearing the mutation.

A treatable genetic cause of epilepsy?

Most previously discovered somatic mutations in epilepsy affect genes in the AKT3-PI3K-MTOR pathway, a multi-pronged cell growth pathway also implicated in tuberous sclerosis complex, a rare genetic disorder that can cause seizures. These epilepsy cases have come to be called “MTOR-opathies.” Some neurologists have begun treating them with everolimus, an MTOR inhibitor also used in tuberous sclerosis.

But SLC35A2 doesn’t seem to belong to or affect the MTOR pathway. Instead, it encodes an enzyme important for glycosylation, a fundamental chemical process of adding sugars to protein molecules.

“This is a completely new pathway we hadn’t really thought of in the context of focal brain malformations,” says Poduri. “It seems to cause a different kind of focal cortical dysplasia than MTOR causes. And because glycosylation is a biochemical pathway, it raises the possibility that a treatment could modify that pathway.”

Genetic diagnosis via spinal tap?

Poduri has begun conversations with biochemical experts to see if a treatment could indeed be developed, assuming SLC35A2 brain findings are confirmed in a clinical lab. Long term, she hopes to be able to offer a genetic diagnosis to all patients with epilepsy, including those whose mutations are somatic — and even if they don’t have brain surgery.

For example, she is exploring the possibility of making a genetic diagnosis of a somatic mutation in the brain through a lumbar puncture (spinal tap). For this, she is tapping the expertise of Maria Lehtinen, PhD, who does research on the role of cerebrospinal fluid (CSF) in brain development. Poduri and Lehtinen met during their post-doctoral studies with Christopher A. Walsh, MD, PhD, a pre-eminent scientist who investigates mosaicism in the developing brain.

While CSF doesn’t contain cells, it does contain DNA from dying cells — perhaps enough to identify a somatic mutation. It also contains proteins that could potentially be analyzed for effects of mutations. And while spinal taps are invasive, they are far less invasive than brain surgery.

“By partnering with people who are thinking all the time about spinal fluid, we might be able to move from aspiration to something clinically relevant,” Poduri says.

Poduri and her colleagues also plan to go back and analyze stored brain samples from children who previously had epilepsy surgery, but lack a genetic diagnosis. A major barrier is that much of this older tissue wasn’t frozen, so is not in the best condition for extracting DNA.

“It’s challenging to detect small changes in old tissues fixed with formaldehyde,” she explains. “But we would like our findings to benefit our patients as part of routine clinical care, and give them a diagnosis that they’ve been lacking for all these years.”

For inquiries about participating in epilepsy genetics research, email epilepsygenetics@childrens.harvard.edu.