The new genetics of Dravet Syndrome

Sundance. I was asked to give a talk on the genetics of Dravet Syndrome at the Dravet Syndrome Foundation meeting in Fort Worth, Texas. I started my presentation asking the question whether there is actually anything novel to talk about given that it is well established that Dravet Syndrome is due to loss-of-function variants in SCN1A, and the challenges are in finding better treatments, not in refining SCN1A genetics. However, this is not quite true. There are several new aspects regarding the genetics of Dravet Syndrome that are worth highlighting.

Figure 1. Conceptual model of how SCN1A expression is regulated by a “poison exon”. The reduction of SCN1A expression is due to persistent expression of SCN1A exon 20N, a poison exon resulting in nonproductive splicing. Exon 20N expression is high during embryonal development and leads to suppression of SCN1A expression. The noncoding mutation maintains the embryonic pattern of exon 20N expression, resulting in reduction of SCN1A transcript. (Figure modified from Helbig and Goldberg, 2020 under a CC BY 4.0 license https://doi.org/10.1371/journal.pgen.1009214.g001)

1 – The history of Dravet Syndrome teaches important lessons
SCN1A was identified in 2001 by Claes and collaborators. However, there were quite some speculations about the etiology of Dravet Syndrome prior to this gene discovery and various other classifications for severe childhood epilepsies. For example, some studies suggested that Dravet Syndrome might be polygenic due to a high frequency of family members affected by febrile seizures. We now know that this is not the case and initial reports may have been due to recruitment bias. In addition, there was the observation that Dravet Syndrome typically started after vaccinations, raising the issue about a causal connection. Vaccine encephalopathy was demystified in 2006 by identifying SCN1A variants in a significant subset of individuals previously thought to have vaccine-related epileptic encephalopathies.

2 – Know your g./c./p.
Rules for variant interpretation in SCN1A are increasingly important, and I gave a brief overview of our work behind the scenes in the gene and variant curation group within ClinGen. During my presentation, I challenged both providers and families to “know your g./c./p.”, to be familiar with the specific variant that was identified in individuals with an SCN1A-related epilepsy. The concept “g./c./p.” refers to the exact position of the disease-causing variant. This is important as Dravet Syndrome is both underdiagnosed and overdiagnosed. I presented two situations of individuals with variants initially reported as variants of uncertain significance. In one case, the variant was causative, but the report was not retrievable, in the other case, a child had been mistakenly diagnosed with Dravet Syndrome, even though the SCN1A variant was a benign population variant and the child’s epilepsy was due to a different cause that required a different therapeutic approach.

3 – Think transcripts, not genes
Until a few years ago, I had assumed that there is nothing relevant with respect to SCN1A gene regulation that can be approached therapeutically. In my mind, SCN1A was under the regulation of a complex landscape of transcription factors that was impossible to disentangle. There was one gene, one transcript, one protein. Cognitive psychologists refer to such simplifications as “heuristics”, mental short-cuts that make our decision-making easier. However, reality is more complex and by thinking of SCN1A as just a gene, we lose track of potential therapeutic targets. In reality, there are many different SCN1A transcripts, including transcripts that include poison exons that reduce SCN1A expression. These non-productive transcripts are now major targets for antisense oligonucleotide therapies. Take home message:  genes themselves are not important, but what genes do is.

4 – Be aware of Dravet mimickers
How many other genes are there for Dravet Syndrome? Several genetic etiologies including PCDH19, CHD2, KCNA2, HCN1, and GABRG2 may result in clinical presentation that may have some overlap with Dravet Syndrome at some point during the disease course. However, in many situations, the clinical presentations and treatments are different. What is the most common gene for SCN1A-negative Dravet Syndrome? It is an SCN1A variant that was initially missed during the diagnostic work-up, as shown in our 2016 publication. The connection between the clinical presentation of Dravet Syndrome and SCN1A is so strong that it is more likely to find an SCN1A variant than any other causative gene, if the initial testing is negative. This is particularly relevant for genetic testing that had been performed several years ago when methodologies were not as established as they are today.

5 – The three-months rule – be aware of gain-of-function variants
There is an increasing awareness of SCN1A gain-of-function variants, as shown in the recent paper by Brunklaus and collaborators. Even though Dravet Syndrome with loss-of-function variants remains the most frequent presentation of SCN1A-related epilepsies, there is also an opposite disease mechanism. In this respect, SCN1A is not different than other the epilepsy sodium channel genes SCN2A, SCN3A, or SCN8A – both loss-of-function and gain-of-function mechanisms can cause disease. Traditionally SCN1A gain-of-function variants were thought to be phenotypically distinct with either hemiplegic migraine presentations or neonatal onset developmental and epileptic encephalopathy with dystonia. However, the phenotypic spectrum might be much broader and some presentations may even mimic Dravet Syndrome at some point during the disease course. However, treatment of SCN1A gain-of-function variants can be vastly different and some individuals may respond to sodium channel blocker depending on the biophysical consequences of the variant. How can we tell these conditions apart? During my talk, I introduced the “three months rule”, suggesting that every SCN1A-related epilepsy due to missense variants and age of onset before the age of three months may be revisited – some of the variants in individuals with early-onset SCN1A-related epilepsies may in fact be gain-of-function variants.

 

Ingo Helbig is a child neurologist and epilepsy genetics researcher working at the Children’s Hospital of Philadelphia (CHOP), USA.

Twitter