Seeing through a forest of SCN2A gene variation

Image of SVN2A mutations
Marking missense. A schematic of the Nav1.2 channel encoded by SCN2A shows the locations of missense mutations associated with infantile epileptic encephalopathy (IEE) and benign infantile seizures (BIS) in red, and those associated with autism spectrum disorder/intellectual disability (ASD/ID) in blue. The cylinders represent transmembrane segments of the channel; loops joining them represent either intracellular or extracellular components of the channel. The voltage sensor is the cylinder marked 4, and the loop joining cylinders 5 and 6 within the membrane is the ion selectivity filter. Image adapted from Sanders S.J. et al.1

On August 2–3, 2019, the FamilieSCN2A Foundation held their biennial SCN2A Professional and Family meeting, in Seattle, Washington. The gathering brought together 37 families of individuals with mutations in the SCN2A gene, 60 investigators, eight clinicians and five industry groups that conduct research and/or clinical work on conditions related to this genetic change. A number of SFARI scientists and staff also attended the event.

The SCN2A family meeting was one of many events that family organizations of rare, neurodevelopmental disorders organized last summer. These meetings help families connect with others similarly affected as well as professionals working to better understand these conditions and develop new therapeutics. SFARI often attends and facilitates research opportunities carried on at these events.

SCN2A is a high-confidence autism risk gene, which encodes a subunit of a sodium channel in the brain called Nav1.2. When the channel malfunctions, conditions like epilepsy and autism follow. As part of its mission to understand the genetics and neurobiological underpinnings of autism, SFARI has awarded about $3 million for research on SCN2A, and some of this research was presented at the meeting. SFARI also supports a ‘genetics first’ initiative called Simons Searchlight (formerly known as Simons VIP), which enrolls people with a genetic diagnosis showing rare genetic changes associated with autism and related neurodevelopmental conditions, such as SCN2A.

Many stories that may reflect the different ways SCN2A can be disabled were told at the meeting. One child had his first seizure when he was days old, and now spends many of his days irritable and immobilized by dystonia. Another developed normally until his first seizure as a toddler, which seemed to wipe out all of his skills; his milestones are now hard won in the face of continuing seizures and an autism diagnosis. Another had a sudden regression at 1 year of age, and after a misdiagnosis and seizure medication, she goes to a school for children with autism. Still another suffered from relentless seizures, which robbed her of speech; she died last year at the age of 12.

So far, about 300 different variants of the SCN2A gene have been documented, and the functional consequences of many are unclear. Some researchers have developed high-throughput experiments to systematically test each of these variants, and to screen compounds that could normalize their function2. Another approach may use gene therapy to boost expression of the remaining good copy of SCN2A. Either way, finding appropriate in vitro testing grounds for these SCN2A variants is essential and may help personalize treatment approaches or identify more homogeneous patient groups for drug trials.

The meeting also underscored the power of family gatherings to push the science ahead. Investigators could see multiple examples of a rare genetic condition and engage new participants in research studies such as The Investigation of Genetic Exome Research (TIGER), a project of the University of Washington that compares phenotypes of single-gene conditions. In turn, families had the opportunity to express their concerns to scientists and infuse the research proceedings with urgency.

“My biggest takeaway from this year’s conference was the mutual inspiration between the scientists and the families,” says Leah Schust, meeting organizer and executive director of the FamilieSCN2A Foundation. Her son has a mutation in SCN2A.

“Meeting the researchers working on a cure for our kids motivates us to fight on,” Schust says. “Then the scientists all say that meeting the families inspires them to go back to their labs and work even harder.”

Family focus. The family meeting helped researchers reconsider what would be meaningful clinical endpoints for potential treatments. Schust says that most researchers and industry groups had thought seizure control was the most important issue. “After listening to us, they realized that quality of life, movement disorders and autonomic dysfunction are higher on our list of where we would like to see improvement,” she says.

SCN2A spectrum

When SCN2A mutations were first linked to autism, the gene stood out because it encodes a relatively well-understood protein, unlike many of the other identified genes. Nav1.2 is a voltage-gated channel found exclusively on excitatory neurons in the brain, where it controls the flow of sodium ions into the neuron, and thus its propensity for firing action potential. Experiments have revealed detailed pictures of Nav1.2’s structure3, and known drugs alter its function4.

SCN2A also stands out because of its high recurrence rate in autism: unlike other autism genes, SCN2A is mutated with somewhat regular frequency5 (Figure 1).

Figure 1. Frequent target. Among autism genes, SCN2A stands apart with high counts of both missense mutation (y-axis) and protein-truncating mutation (x-axis) in autism cases gathered in the largest exome-sequencing study to date. From Satterstrom F.K. et al.5

Just as understanding why a car won’t start is critical to fixing it, researchers need to understand how these SCN2A mutations alter the Nav1.2 channel. A current model1 posits that some mutations are gain-of-function, rendering the channel too active and the brain hyperexcitable, leading to infantile epilepsy; conversely, loss-of-function mutations reduce excitability and seem associated with autism and/or intellectual disability, as well as childhood-onset (as opposed to neonatal) seizures.

Yet the functional consequences of most SCN2A mutations remain unknown, and some may not fall neatly into a loss-of-function or gain-of-function category. A way of making sense of these mutations may come from looking at the working parts of Nav1.2, said Arthur Campbell of the Broad Institute of MIT and Harvard. For example, missense SCN2A variants linked to epilepsy seem to hit the channel randomly. But when marking their location on a crystal structure model of the channel, the missense variants cluster in several places: on the voltage sensor, on the linker helix responsible for conveying voltage sensor movement to the channel pore, on an area thought to interact with the beta-subunits involved in chaperoning the channel to the right place, and on the inactivation gate, which closes the pore off from sodium ion flow. He suggested that this knowledge, combined with the structural similarities between all sodium channels, may help drug development for SCN2A-related conditions.

High-throughput systems that can assay hundreds of cells at a time are helping researchers systematically explore SCN2A mutation, explained SFARI Investigator Al George of Northwestern University. While conventional electrophysiology would require weeks of work to characterize a single SCN2A variant, George’s group uses an automated patch-clamp system that can characterize multiple variants transfected into non-neuronal cell lines in a day. Using this system, two variants associated with neonatal seizures both exhibited an exceptional willingness to activate and a slowness to inactivate, which are properties consistent with a gain-of-function interpretation.

The high-throughput set up also promises to expedite the hunt for drugs to normalize SCN2A function: George described a 384-well plate design that allows measurement of the effects of two different drugs, at four different concentrations, on the SCN2A variant and control channels simultaneously. A known drug (carbamazepine) and an experimental drug (PRX-330) shifted channel inactivation to more hyperpolarized voltages, which could help quiet channels with gain-of-function mutations.

To narrow in on potentially therapeutic compounds, Jeff Cottrell and colleagues at the Broad Institute of MIT and Harvard have come up with a two-stage screen to find small molecule activators or inhibitors of Nav1.2 channels. First, compounds are initially tested on non-neural cells transfected with Nav1.2 sodium channels and potassium channels, which enables them to spike. The cells in 384-well plates are stimulated in parallel, and voltage-sensitive dyes give a readout of spiking activity; remarkably, Cottrell’s system allows data collection from up to 96 wells simultaneously. Any compounds that modulate spiking would then be subjected to the second stage, in a high-throughput electrophysiology assay similar to that described by George. Compounds with helpful mechanisms would then be tested for selectivity for Nav1.2 versus other sodium channels. A selective compound would then be tested in neurons, first in vitro then in vivo. This step-wise process has identified an activating compound that makes Nav1.2 more likely to open at rest and has potent effects on action potentials in brain slices and on electroencephalogram (EEG) traces from mice engineered to carry a disabled copy of SCN2A; however, Cottrell said this particular compound is not a therapeutic candidate in part because it broadens the action potential in a way that could promote seizures. A full screen is underway, and so far has identified 378 modulators from a library of 77,000 compounds.

Beyond academia, J.P. Johnson Jr. of Xenon in Burnaby, British Columbia, discussed the company’s work to create sodium channel inhibitors for treating epilepsy. To obtain selective compounds, the group targets the voltage-sensing domain because its structure is the most diverse region of sodium channels. Xenon uses a trial-and-error method to optimize sodium channel inhibitor potency and selectivity. The methodical process has yielded some interesting compounds, including both selective Nav1.6 inhibitors and dual Nav1.6 and Nav1.2 inhibitors. Both quashed spiking in mouse excitatory pyramidal neurons, which contain only Nav1.2 and Nav1.6, but they did not alter spiking in Nav1.1-containing inhibitory neurons. A Nav1.6 selective inhibitor, XEN901, is currently undergoing safety trials in humans.

Gene therapies

Kathrin Meyer of Nationwide Children’s Hospital in Columbus, Ohio, addressed the possibility of using gene therapy to normalize malfunctioning Nav1.2 channels. Meyer has been involved in several gene-therapy trials for neuromuscular disorders, including a successful one for infant-onset spinal muscular atrophy type6. Gene therapy for brain diseases was spurred by the discovery of adeno-associated virus 9 (AAV9), which can cross the blood–brain barrier to deliver genetic material to the central nervous system. AAV9 is small, cannot replicate, does not integrate into host DNA and seems not to cause disease in humans. In considering gene therapy for SCN2A-related conditions, Meyer emphasized an approach that adds back a working copy of the gene, thus sidestepping the need for gene editing to make mutation-specific corrections. Such a treatment would only apply to those with loss-of-function mutations.

The large size of the SCN2A gene precludes its delivery by AAV9, however. As a workaround, Meyer suggested that SCN2A’s mRNA transcript could be targeted in an attempt to replace only the affected area of the mRNA. So far, such strategies have not been very efficient, but there are new ideas that might address some of the difficulties. Because access to tissue samples of patients with neurological disorders is limited, the development and testing of new therapies is complicated. Meyer suggested developing gene therapies in vitro using neurons reprogrammed from skin cells of patients. This might help identify which patients would react best to a certain treatment. “There is likely not a one-fit-for-all situation,” she said.

SFARI deputy scientific director John Spiro underscored the need for in vitro systems, citing the organization’s initiative to bank blood cells to systematically generate induced pluripotent stem cells from individuals with autism. Simons Searchlight is also a resource of many different biospecimens for researchers. So far, 186 families with SCN2A-related changes have registered, and 83 of these have completed consent, lab reports and medical histories — with a large number of blood samples as well. (On the sidelines of the meeting, 18 parents, 11 of their children with SCN2A mutations, and three unaffected siblings donated blood toward this initiative.) Spiro also stressed a need to come up with more quantitative methods of phenotyping, such as wearable electronics that can monitor sleep and circadian rhythms. Data that can be collected longitudinally and at home might provide sensitive outcome measures for clinical trials.

A new role for Nav1.2 has been revealed in recent work described by SFARI Investigator Kevin Bender of the University of California, San Francisco: the channels mediate back-propagating action potentials, which travel into the dendritic trees of neurons. Mice engineered to lack one copy of SCN2A — a situation that mimics people with truncating SCN2A mutations that render the resulting Nav1.2 channels useless — had cortical neurons with slower action potentials, reduced dendritic excitability and immature synapses based on their shape and function7. This role for Nav1.2 was particularly important later in development: when conditional knockout mice lost an SCN2A copy later in life, their cortical neurons exhibited immature synapses, though their density remained normal. Preliminary experiments suggest that adding back a working copy of SCN2A later in life — through transgenic methods or by upregulating transcription of the remaining good copy of SCN2A via CRISPR techniques — can restore action potential velocity and synaptic maturity.

Images of neurons
Figure 2. Dual roles. Mouse cortical neurons lacking one copy of SCN2A have trouble initiating action potentials early in development (left). Later in development, the neurons exhibit impaired dendritic excitability and synapses (right); those lacking a copy of SCN2A are sparser and thinner than controls (inset). Image adapted from Spratt P.W.E et al.7

Bender stressed how interacting with the SCN2A family group helped focus his research on important aspects of their children’s conditions. For example, parents have noted sensory hypersensitivity in their children, leading Bender to collaborate with colleague Evan Feinberg to use an eye-tracking assay in mice to measure their visual responses. He noted that SCN2A haploinsufficient mice were more sensitive to certain visual stimuli than control mice; if the assay is robust, it could help bridge the gap between SCN2A-related phenotypes in humans and behaviors measured in mice.

Future directions

As meeting attendees sorted through the new findings, therapeutic questions lingered. An important issue for any therapy, whether drug or gene, will be how early in development one will have to intervene to help someone with an SCN2A mutation. Bender noted that synaptic properties could be rescued in his mice when they were 30 days old — equivalent to a 10-year-old human — but these and other experiments will have to probe the time periods during which therapies will be maximally effective. To find good measures of efficacy also means understanding the full complement of conditions that beset people with SCN2A mutations. For example, though seizures afflict many, Keith Coffman of Children’s Mercy Hospital in Kansas City, Missouri, suggested that, in some cases, these represent a movement disorder rather than epilepsy. Basic descriptive knowledge like this is imperative for guiding future treatment approaches.

Another smaller SCN2A meeting is planned for this year from July 30 to August 2, in Columbus, Ohio. This will be more family focused, says Schust, and there will be opportunities to participate in research.

“There is clearly a lot more work to do before all the terrific basic research that was discussed at this meeting produces meaningful results for families, but it is extremely gratifying to see how much progress has been made on so many fronts and how many new good ideas are emerging,” Spiro says. “And it’s terrific to witness firsthand the positive cycle of how families drive researchers and vice versa.”

Attendees at the SCN2A family meeting, Seattle, Washington, August 2–3, 2019.


  1. Sanders S.J. et al. Trends Neurosci. 41, 442-456 (2018) PubMed
  2. Pan J.Q. et al., Methods Mol. Biol.1787, 235-252 (2018) PubMed
  3. Pan X. et al., Science 363, 1309-1313 (2019) PubMed
  4. De Lera Ruiz M. and Kraus R.L. J. Med. Chem. 58, 7093-7118 (2015) PubMed
  5. Satterstrom F.K. et al. Cell 180, 568-584 (2019) PubMed
  6. Mendell J.R. et al. N. Eng. J. Med. 377, 1713-1722 (2017) PubMed
  7. Spratt P.W.E et al. Neuron 103, 673-685 (2019) PubMed
Recent Workshop and Meeting Reports