Mutations in the gene encoding the α1 subunit of the voltage gated sodium channel (SCN1A) are associated with several epilepsy syndromes, ranging from relatively mild phenotypes found in families with genetic epilepsy with febrile seizures plus (GEFS+) to the severe infant-onset epilepsy Dravet syndrome. Evidence has emerged of the consequences of SCN1α dysfunction in different neuronal networks across the brain pointing toward a channelopathy model causing the neurologic features of Dravet syndrome that is beyond purely seizure related damage. A genetic change will present according to its severity, the genetic background of the individual, and environmental factors, and will affect a variety of neuronal networks according to channel distribution. This already-vulnerable system may be susceptible to secondary aggravating events such as status epilepticus. The channelopathy model implies that pharmacologic treatment and the restoration of impaired γ-aminobutyric acid (GABA)ergic neurotransmission might not only help prevent seizures but might affect the comorbidities of the syndrome. This critical review explores recent evidence relating to the pathogenicity of SCN1A mutations in Dravet syndrome and the effect these have on the wider disease phenotype and discusses whether knowledge of specific genotypes can influence clinical practice. Genetic technology is currently advancing at unprecedented speed and will increase our knowledge of new genes and interacting genetic networks. Clinicians and geneticists will have to work in close collaboration to guarantee good delivery and counseling of genetic testing results.
A PowerPoint slide summarizing this article is available for download in the Supporting Information section here
Mutations in the gene encoding the α1 subunit of the voltage gated sodium channel (SCN1A) are associated with several epilepsy syndromes, ranging from relatively mild phenotypes found in families with genetic epilepsy with febrile seizures plus (GEFS+) to the severe infant-onset epilepsy, Dravet syndrome, previously known as severe myoclonic epilepsy of infancy (SMEI).[1-4] To date, hundreds of sequence variants of the SCN1A gene have been identified. Most mutations are novel, and when an infant presents with febrile seizures (FS) it is difficult to predict which phenotype they will develop.
Dravet syndrome typically presents around 6 months of age, in previously well children, with prolonged, febrile and afebrile, generalized clonic or hemiclonic epileptic seizures. Other seizure types including myoclonic, focal, and atypical absence seizures appear between the ages of 1 and 4 years. The epilepsy is usually not responsive to standard antiepileptic medication, and affected children develop cognitive, behavioral, and motor impairment.[7-10]
Among the whole spectrum of SCN1A-related childhood epilepsies, the strongest association of an SCN1A mutation with a clear phenotype is seen in Dravet syndrome, with 70–80% of affected individuals carrying the gene mutation.[5, 11] The discovery of this correlation between genotype and phenotype has triggered a wave of research spanning from animal models to human trials, and SCN1A has become one of the most relevant epilepsy genes today.
Dravet Syndrome—Epileptic Encephalopathy and Channelopathy
Dravet syndrome had been described as “a prototype of an epileptic encephalopathy.”[12, 13] The term “epileptic encephalopathy” was introduced by the International League Against Epilepsy (ILAE) task force on classification and terminology in 2001 representing a new concept and defined as “a condition in which the epileptiform abnormalities themselves are believed to contribute to the progressive disturbance in cerebral function.”
In 2010 the “Revised ILAE terminology and concepts for organization of seizures and epilepsies” further specified that “the epileptic activity itself may contribute to severe cognitive and behavioral impairments above and beyond what might be expected from the underlying pathology alone, and that these can worsen over time.”
This was accompanied by a comment that “epileptic encephalopathy should be viewed as a concept and a description of what is observed clinically, with the recognition that we are rapidly approaching a clear understanding of the effects of epilepsy on brain function and the potential for lasting deleterious impact in the developing brain. We must, however, recognize that the source of an apparent encephalopathy is usually unknown; it may be the product of the underlying cause, the result of an epileptic process, or a combination of both.”
In recent years the Dravet mouse model, generated by knockout (KO) of the Scn1a gene (Scn1a+/−), has made a significant contribution to helping us understand the neurobiology of the disease.[16, 17] From this animal work, specific neuroanatomic correlates have been revealed, explaining the variety of symptoms seen in the mouse model. The genetic change results in altered function of Nav1.1 sodium channels in neurons clustered throughout the brain.[16, 18] Seizure susceptibility is caused by impaired γ-aminobutyric acid (GABA)ergic firing in hippocampal interneurons, which lowers the seizure threshold; impaired firing of GABAergic cerebellar Purkinje cells explains the motor disorder presenting with ataxia.[16, 17]
The incidence of sudden unexpected death in epilepsy (SUDEP) has been reported as significantly higher in Dravet syndrome compared to other epilepsy syndromes.[6, 19] However, these data are derived largely from parent surveys; prospective population-based studies have not been performed. A recent study on KO Scn1a mice suggests that SUDEP is caused by increased parasympathetic activity following tonic–clonic seizures, leading to lethal bradycardia and electrical ventricular dysfunction, thereby indicating that brain, but not cardiac KO of Scn1a, is responsible for SUDEP. However, a study of ventricular myocytes in Scn1a knock-in mice has demonstrated an independent cardiac contribution to SUDEP. This work not only showed alterations in neuronal excitability, but also cardiac electrophysiologic abnormalities contributing to the susceptibility for arrhythmogenesis and SUDEP, independent of seizures. The authors hypothesize that reductions in Nav1.1 expression may indirectly affect Nav1.5 activity and cardiac electrical function.
New evidence has emerged showing that Dravet mice have behavioral problems affecting neuronal networks in the prefrontal cortex. The mice were observed to display a whole range of abnormal behaviors that were measured using tests assessing deficits in social interaction. Many of these mirror the comorbidities of Dravet syndrome seen in humans, including hyperactivity and stereotypical behavior, cognitive deficits, and impaired social interaction.[8-10, 23] Underlying these behavioral and cognitive impairments appears to be a reduced expression of GABAergic interneurons in the forebrain; especially the prefrontal cortex tends to be affected by low Nav1.1 channel expression. Of interest, treatment with low-dose clonazepam rescued the abnormal behavior patterns in the affected mice, suggesting that these are caused by channel dysfunction rather than the epileptic activity itself as contributor to cognitive and behavioral impairment.[15, 22] This provides evidence that the medications used to control seizures may have a role to play in the overall regulation/rescuing of the deficient sodium current. Reinstating the impaired GABAergic neurotransmission may not only lead to improved seizure control but also to improved function of the prefrontal and cerebellar networks, as has been shown in mice.
Overall these findings clearly demonstrate the involvement of different neuronal networks across the brain and indicate a disease model that is beyond purely seizure-related damage.
To eliminate the effect of recurrent seizures as the sole culprit of the neurocognitive decline, further work has recently looked specifically at a model where the expression of Scn1a was selectively reduced in mice that did not have seizures. Down-regulation of Nav1.1 did not cause seizures but led to difficulties with special recognition, showing for the first time that even without seizures affected animals are impaired in their cognition by the reduced expression of SCN1α channels.
These findings point toward a channelopathy model as cause of Dravet syndrome that is independent of direct seizure damage. A genetic change will present according to its severity and modifying factors and impact on a variety of neuronal networks according to channel distribution.
Patient-induced pluripotent stem cell (iPSCs) work, an in vitro human Dravet syndrome model, has recently shown evidence of widespread increase in sodium channel-subunit expression (Nav1.1, Nav1.2, Nav1.3, Nav1.6, and Nav1.7) in maturing Dravet syndrome neurons, suggesting an overcompensation response in developing neurons.
This hypothesis is also supported by evidence from the neurogenesis and maturation of sodium channels in mouse models and human brain tissue. Early on in fetal and neonatal development, sodium channels undergo a process of functional maturation and increase their ability to generate action potentials over time.[27, 28] As Nav1.1 expression is very low in neonates, other alpha subunits such as Nav1.2 and Nav1.3 may compensate for reduced Nav1.1 function during this early stage of development. Cheah et al. demonstrated in their recent work, a physiologic decline in Nav1.3 and increase in Nav1.1 with age in normal human brain tissue. The authors hypothesize that the natural loss of Nav1.3 channel expression in brain development, coupled with the failure of increase in functional Nav1.1 channels in Dravet syndrome, might lead to widespread dysfunction of neuronal networks, intractable seizures, and comorbidities. This could explain why children with Dravet syndrome typically do not present as neonates or in very early infancy. Animal models have demonstrated a compensatory up-regulation of Nav1.3 channels in the hippocampus of Nav1.1 mutant mice, which might salvage the function of a mutant channel temporarily. However, once full Nav1.1 expression becomes essential for normal neuronal function, the deficit can no longer be compensated for by Nav1.3 up-regulation, and children start to present with seizures after several months. This might also be the reason for the marked increase in abnormal interictal electroencephalography (EEG) recordings over time.[6, 7] Whereas few abnormal recordings are seen in the first 6 months, more than three fourths of children with Dravet syndrome have an abnormal interictal EEG by the end of the third year; and early EEG abnormalities predict a worse developmental outcome. As the brain matures with age and adapts toward higher cognitive functioning, a defect in Nav1.1 channels may become more apparent and the more severe the degree of malfunction, the earlier the EEG abnormality might appear and the more unfavorable the sequelae could be.
An almost reverse pattern of seizure evolution is seen in inherited SCN2A mutations that present with benign familial neonatal-infantile seizures (BFNIS). Animal work has shown that the developmental expression of Nav1.2 in hippocampal and cortical neurons diminishes with time and is eventually replaced by the dominant channel type Nav1.6, thereby rescuing channel function. This might explain why children with BFNIS due to inherited SCN2A mutations generally present as neonates or infants and seizures remit spontaneously within the first year of life, whereas truncating or de novo SCN2A missense mutations are known to be associated with more severe phenotypes such as Ohtahara syndrome and early onset epileptic encephalopathies (EOEEs).[31, 32]
Scn8a mutations have also been shown to rescue the effect of an Scn1a mutation in the Dravet mouse model by restoring normal seizure threshold. The increase in seizure activity in Scn1a gene mutants might have been rescued by the decrease in hippocampal and cortical excitability of pyramidal cells that is caused by Nav1.6 dysfunction.[33, 34]
Seizures are only one of the many presenting features in this channelopathy alongside comorbidities such as behavioral and learning difficulties that are often equally important factors determining a patient's quality of life. Related to the genotype and independent from the occurrence of seizures, an affected individual will also be at risk of developing a motor disorder and/or a significant cognitive impairment as part of the disease course. A retrospective study examining the cognitive development of 26 children with Dravet syndrome identified the early appearance of myoclonus and absences as being associated with worse cognitive outcome. However, due to the variability of presenting symptoms and signs that are not explained by epileptic activity alone, the authors suggest that “the channelopathy itself is probably crucial in determining the phenotype.” A recent prospective neuropsychological evaluation of 67 patients with Dravet syndrome concluded that “Encephalopathy in children with Dravet syndrome is not a pure consequence of epilepsy.” The authors failed to detect a robust correlation between developmental/intellectual quotient and most epilepsy variables, apart from the presence of myoclonic or focal seizures in those children older than the age of 3 years.
Nevertheless good seizure control can—even after years—still be associated with better cognitive outcome, as has been demonstrated in an article on the long-term course and neuropathology findings in Dravet syndrome. The authors evaluated a cohort of 22 adult patients, including three postmortem cases, and found that even after years of drug resistance, three adult Dravet patients showed improved seizure control once they had been switched to appropriate syndrome specific mediation. This lasting reduction in seizures led to a significant additional improvement in cognitive function and quality of life in two patients. Remarkably, on histopathologic examination of the postmortem brain tissue, no striking features were seen and the neurons and interneurons in neocortex and hippocampi appeared preserved. Considering this evidence, and following the cognitive improvements that were observed with treatment change and better seizure control, the authors argued that Dravet syndrome is at least in part an epileptic encephalopathy.
Taking into account the different arguments of the encephalopathy/channelopathy debate, a number of key aspects emerge: There is overwhelming evidence that Dravet syndrome is a channelopathy causing widespread Nav1.1 dysfunction throughout the brain and this channel dysfunction contributes to the encephalopathy. It appears plausible that this already-vulnerable system may be susceptible to secondary aggravating events such as status epilepticus. Furthermore, pharmacologic treatment and the restoration of impaired GABAergic neurotransmission might not only help prevent seizures but might also recover wider neurologic functioning.
This leads us back to the original International League Against Epilepsy (ILAE) comment from 2010: “We must, however, recognize that the source of an apparent encephalopathy may be the product of the underlying cause, the result of an epileptic process, or a combination of both.”
Knowing the Genotype—Will It Benefit Patient Care?
Testing of the SCN1A gene in young children, who present with recurrent prolonged atypical febrile seizures, is becoming part of the clinical routine of child neurologists and pediatricians alike. However, when an SCN1A mutation is detected, geneticists and physicians often find themselves in the dilemma of how to interpret the test results.
Recent work has shown a number of genotype–phenotype associations among large cohorts of SCN1A mutation–positive patients.[5, 36] Although these findings demonstrate important general trends within large cohorts, these cannot be easily applied directly to the individual patient a physician encounters in clinic. Many reports have illustrated that the phenotypic presentation of SCN1A mutations can be heterogeneous within the same family, and even though an individual with a mild GEFS+ phenotype might have a 50% chance of passing the mutation on to his or her children, the resulting phenotype could be unpredictable, spanning from a normal child to a child with Dravet syndrome, illustrating what an important role the genetic background plays.
Mouse models have shown that the same Scn1a mutation expressed in a different genetic background leads to marked variation of the phenotype. Several modifying factors have been reported in animal models such as Scn8a mutations restoring normal seizure threshold and the compensatory up-regulation of Scn3a in Dravet mice.
Despite the significant genetic background effects and the abundance of confounding factors, there are nevertheless a number of significant genotype–phenotype associations: for example, the link between truncating mutations with both a young age at seizure onset and the presence of a severe phenotype. These findings, however, are trends that apply to the whole group of Dravet patients and not necessarily to the individual. When, for example, a genetic diagnosis of an SCN1A mutation is made (often between 9 and 24 months of age), it is currently impossible to accurately predict how this child will progress in the future. More than 90% of SCN1A mutations arise de novo, and the recurrence risk is low. If prenatal testing is considered, this will have to be carefully weighed against the associated risks on an individual patient basis.
Being able to make a genetic diagnosis has significant positive implications for patients.[37, 38] The benefits include a better informed antiepileptic drug (AED) choice, sparing of further unnecessary investigations, improved access to therapies/services, and the confirmation of a diagnosis that gives parents “an answer” and allows them to adjust their goals for the future. There is currently no evidence that a particular mutation type can inform AED choice. Any child with Dravet syndrome should be treated with AEDs shown to be efficacious in this syndrome, and the use of lamotrigine and/or carbamazepine is discouraged to avoid seizure exacerbation.[7, 40]
Currently, mutations are primarily judged as to whether they are pathogenic, but less so how pathogenic they are. There are a number of different prediction models and software packages in use by most genetic laboratories to help geneticists decide whether a change is pathogenic. This includes a measure of conservation across species, a measure of physicochemical properties (Grantham score [GS]), SIFT (Sorting Intolerant from Tolerant algorithm) predictions and PolyPhen (polymorphism phenotyping) predictions. The GS is a measure of physicochemical difference between amino acids including polarity, composition, and volume, and the more dissimilar the amino acid replacement the higher the GS. SIFT predicts whether an amino acid substitution affects protein function based on the degree of conservation of amino acid residues compared to closely related sequences. PolyPhen analyzes the effect of amino acid substitutions on structure and function of human proteins using physical and comparative characteristics.
These mutation classification systems try to define a generic pathogenicity by establishing how dissimilar the substituted amino acids are, whether the amino acids are conserved across species, and whether mutations affect a functionally important area of the protein. However, these models are still in their infancy and are not sufficiently refined to examine how, for example, a specific polarity change might affect a particular amino acid residue in the voltage sensor of the SCN1α protein. Functional studies have been much more useful in providing evidence that a change is pathogenic, but even this approach has its limitations. For example patch-clamp HEK cell experiments are performed in isolation of a living organism, removed from other influencing factors, and the results are not always applicable in clinical practice. Experiments in living mammals such as rodents have been much more informative but require a substantial financial investment.
Recently more detailed modeling has been undertaken to judge the significance of SCN1A mutations. By examining a segment of the voltage and pore region rather than the whole domain, there was better discrimination between phenotypes, showing that a change in polarity appeared to be particularly important in the voltage sensor.[5, 36] Now that we are acquiring a better understanding of the detailed SCN1α protein structure, we will gain more insights into how the protein functions as a whole and more importantly what significance a particular amino acid residue carries. Ultimately this information will allow the creation of more sophisticated prediction models.
It is important to bear in mind that the phenotype might not necessarily be determined only by the SCN1α protein itself, but by a number of auxiliary proteins. Sodium channels are known to constantly traffic in and out of the cell membrane, and their function will depend on a number of interacting proteins, as has been shown for the β subunits that play a crucial part in cell adhesion and channel localization. Homozygous SCN1B mutations have been found to be rarely associated with Dravet syndrome. Functional work has shown that mutations result in impaired transport of β1 subunits to the cell surface, resulting in little or no cell-surface expression. Whereas one SCN1B allele was sufficient to maintain function, the absence of both alleles led to hyperexcitability in both mice and humans. In theory, any posttranslational interference with the SCN1α protein function could result in a similar phenotype.
The Beginning of a New Era
Genetic technology is currently advancing at unprecedented speed, and instead of testing single genes we are now starting to test whole panels of genes. Next Generation Sequencing allows hundreds of different genes to be tested in parallel and is becoming increasingly available in routine clinical practice. However, such extensive epilepsy panels may simply be a stepping stone to whole exome/genome sequencing in epilepsy diagnostics. These powerful analytic techniques will increase our knowledge of new genes and will direct functional work toward identification of new pathways and interacting genetic networks. This will inevitably lead to even more complex levels of neurobiological understanding, which will further challenge the research and clinical community.
The generation of vast amounts of data will require expertise to interpret the findings, and physicians will have to become “genetically literate.” Computer technology and prediction modeling will facilitate the interpretation of results, and programs need to be designed that allow this information to be processed. Clinicians and geneticists will need to work in close collaboration to guarantee good delivery of genetic testing results and counseling. When this is done poorly it will lead to unnecessary patient misunderstanding and anxiety.[37, 38] As the ILAE genetics commission report from 2010 stated “post-test genetic counselling is crucial to help the patient understand the test result and begin to digest it in the context of life circumstances.”
Dr. Andreas Brunklaus' research fellow post was funded by the Muir Maxwell Trust. Dr. Sameer M Zuberi is an Associate Editor of The European Journal of Paediatric Neurology, has received honoraria for speaking at UCB Pharma- and Glaxo Wellcome–sponsored educational meetings and received research funding from Epilepsy Research UK and Dravet Syndrome UK. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.
Dr. Andreas Brunklaus is a research fellow in pediatric epilepsy.