"Epilepsy: New Leaps for Ancient Disease" by Olivia Stovicek
Epilepsy, or the condition of having recurrent, unprovoked seizures, is among the world’s oldest and most difficult to treat diseases . Documentation of epilepsy’s signs and symptoms goes back nearly 4500 years, and it is the fourth most common neurological disorder in the U.S., with approximately 4% of people in the U.S. and worldwide developing epilepsy in their lifetimes [2, 3]. Nevertheless, almost a third of patients have seizures that cannot be controlled with existing antiseizure medication, and for nearly two-thirds of patients, their epilepsy has no known cause [3, 4]. Recent research has highlighted how focusing on networks—both networks of genes within cells and networks of cells within the brain—may finally enable the next leaps in the field.
Epilepsy is really many diseases rather than one; there are nearly forty known forms of epilepsy, each with distinct symptoms and treatments . Some of these, such as Dravet syndrome, are typically caused by mutations in a single gene. In Dravet syndrome, that means mutations in the gene for an important sodium channel, a protein that allows sodium ions to cross a cell’s membrane and helps determine the cell’s electrical behavior . This is unusual, however. Most epilepsy symptoms are determined by not one but many genes, often in interaction with environmental factors. Even epilepsies acquired after stroke or infection often have genetic contributions .
Since the first gene associated with epilepsy was identified in 1990, it has become increasingly clear that most forms of epilepsy without known cause are likely to have a genetic basis [4,6]. Most individuals’ specific mutations are rare, however, and many epilepsies have highly complex genetic contributions . A useful way to unravel the interactions between different gene mutations—to go beyond knowing what mutations are involved and start figuring out what they do, how all their effects combine to cause epileptic seizures—is to think about them as being part of a network, and to use computational models to look at how that network works.
A good example of how networks of genes have been explored to learn about the effects of combinations of mutations is a paper by Tara Klassen, Jeffrey Noebels, and colleagues at the Baylor College of Medicine, in which they sequenced more than half of all known ion channel genes for almost 300 people with and without epilepsy. Ion channel genes are the largest one group of genes implicated in causing epilepsy [6, 7]. Yet computational modeling the researchers did based on this sequencing indicated that whether people with mutations in ion channel genes have epilepsy, and what kind, depends not on how many of these mutations they have, but the particular pattern of them, how they each contribute to the overall network regulating the electrical properties of the cell . As we learn more about these patterns, modeling the effects of each mutation will become increasingly helpful for predicting whether a person’s particular combination of mutations will lead to epilepsy, how severe that epilepsy will be, and what drugs to treat it with .
Networks aren’t only important in epilepsy research for understanding how genes interact; they’re important for looking at neuron activity, too. Epilepsy is now understood as a “network disease,” with seizures that originate in one part of the brain (the “seizure onset zone”) occurring and spreading because of the interaction of neurons from the seizure onset zone with neurons in other parts of the brain . Improving our understanding of the network of interactions between neurons during seizures might make it possible to predict seizures better, or even to prevent them by altering the network . For example, if a particular brain region was found to control other parts of the brain in an epileptic network, perhaps it would be possible to implant a device that would stimulate that brain region to regulate seizure activity .
Some work, like a 2017 study from Christian Geier and Klaus Lehnertz, has sought to learn which parts of the brain are important in epileptic networks at what times . Geier and Lehnertz recorded brain activity for 17 patients for several days straight. Then they used that data to construct networks of epileptic brain interactions. The nodes in the network were locations on the brain, based on where the electrodes tracking brain activity had been placed, and whether nodes were linked was determined by a mathematical analysis of the extent to which activity at one node was dependent on activity at the other. The researchers analyzed which nodes or brain regions were most important, measured by how strongly a node interacted with other nodes and whether it acted as a bridge to connect many other pairs of nodes. They found that the importance of different brain regions in an epileptic network varies over time, largely in ways directed by the body’s daily rhythms. A key implication of these findings is that the seizure onset zone—often considered the most important part of the epileptic network—is not always the most important, and that other parts of the brain should be explored in looking to treat seizures.
Other exciting work explores how to identify the roles of important brain regions in regulating seizure networks. Seizures involve abnormal synchronization in the activity of neurons, and a 2016 study from Ankit N. Khambati, Danielle S. Bassett, and other researchers at the University of Pennsylvania sought to find out more about how this synchronization works . To explore this question, they again turned to networks; using similar methods to Geier and Lehnertz, they built models of dynamic epileptic networks right before and during a seizure. Then, they removed different nodes from their model—corresponding to different brain regions—to see how that affected the activity of the network. They were able to show that certain brain regions affected synchronization of the overall network, with some regions appearing to synchronize the activity of other neurons, and other regions desynchronizing it. The amount of network synchronization and whether a seizure happened or not was regulated by a tug of war between these different parts of the brain, including parts that were not part of the seizure-onset zone. This work further highlights how often overlooked “healthy” regions of the brain can contribute to seizures, but it also provides a way of investigating mechanisms of epilepsy regulation at the network level.
Though much of the recent work on epileptic networks has focused on figuring out how seizures work, some of this research may be closer to affecting current patients’ lives. One study released just in the last few months, for example, could make surgery easier for epilepsy patients. Surgical removal of brain regions that cause seizures can be extremely helpful for patients who aren’t helped by existing epilepsy medication, but in order to identify which regions to remove, patients often have to undergo invasive surgery so that their brain activity can be monitored in detail . They must be monitored until at least one seizure occurs, which means that patients have to wait for a seizure in a hospital, often for as long as a week, before they can have a second operation to remove seizure-causing regions . Researchers Eun-Hyoung Park and Joseph R. Madsen, however, were able to find a way to predict which parts of patients’ brains were causing seizures without having to monitor them for a seizure first . They adapted Granger causality analysis, an approach invented by an economist, to identify networks of causal relationships between activity in different parts of a brain not having seizures. Areas found to have a high concentration of causal nodes predicted quite well the areas that would be chosen for removal using the typical method. Though this technique needs refinement and further study before it can be used, it presents an exciting possibility.
These are just a few examples, but they highlight the potential of network-based investigations of epilepsy for advancing understanding and treatment of the disease. Jeffrey Noebels of Baylor has described the current moment as “the bronze age of epilepsy neurogenetics,” a time of moving past basics and building a grammar, an understanding of interactions, between genes rather than looking at them singly . Keep an eye on this area of research in the future; it just might be where we achieve the next few breakthroughs.
 Epilepsy Foundation. What Is Epilepsy? http://www.epilepsy.com/learn/epilepsy-101/what-epilepsy(accessed May 25, 2017).
 Hirtz, D., Thurman, D.J., Gwinn-Hardy, K., Mohamed, M., Chaudhuri, A.R., and Zalutsky, R. (2007) How common are the “common” neurologic disorders? Neurology 68(5), 326-337.
 Maikki, H. (2014) Epilepsy—burning questions and emerging therapies. Nat. Rev. Neurol. 10, 243.
 Thomas, R.H., and Berkovic, S.F. (2014) The hidden genetics of epilepsy—a clinically important new paradigm. Nat. Rev. Neurol. 10, 283–292.
 Berg, A.T., Berkovic, S.F., Brodie, M.J., Buchhalter, J., Cross, J.H., Van Emde Boas, W., Engel, J., French, J., Glauser, T.A., Mathern, G.W., Moshé, S.L., Nordli, D., Plouin, P., and Scheffer, I.E. (2010), Revised terminology and concepts for organization of seizures and epilepsies: Report of the ILAE Commission on Classification and Terminology, 2005–2009. Epilepsia 51, 676–685.
 Noebels, J. (2015) Pathway-driven discovery of epilepsy genes. Nat. Neurosci. 18(3), 344–350.
 Klassen, T., Davis, C., Goldman, A., Burgess, D., Chen, T., Wheeler, D., McPherson, J., Bourquin, T., Lewis, L., Villasana, D., Morgan, M., Muzny, D., Gibbs, R., and Noebels, J. (2011) Exome sequencing of ion channel genes reveals complex profiles confounding personal risk assessment in epilepsy. Cell 145, 1036–1048.
 Geier, C., and Lehnertz, K. (2017) Long-term variability of importance of brain regions in evolving epileptic brain networks. Chaos 27, 043112.
 Khambhati, A.N., Davis, K.A., Lucas, T.H., Litt, B., and Bassett, D.S. (2016) Virtual Cortical Resection Reveals Push-Pull Network Control Preceding Seizure Evolution. Neuron 91, 1170–1182.
 Park, E-H., and Madsen, J.R. (2017) Granger causality analysis of interictal iEEG predicts seizure focus and ultimate resection. Neurosurgery 0, 1–11.