The genetics of Dravet syndrome


  • Carla Marini,

    1. Pediatric Neurology Unit and Laboratories, Children’s Hospital A. Meyer – University of Florence, Florence, Italy
    Search for more papers by this author
  • Ingrid E. Scheffer,

    1. Department of Medicine, University of Melbourne, Austin Health
    2. Department of Paediatrics, University of Melbourne, Royal Children’s Hospital, Melbourne, Victoria, Australia
    Search for more papers by this author
  • Rima Nabbout,

    1. Service de Neurologie Pédiatrique, AP-HP, Hopital Necker-Enfants Malades, centre de référence épilepsies rares, 5 Inserm U663, Paris Descartes University, Paris, France
    Search for more papers by this author
  • Arvid Suls,

    1. VIB-Department of Molecular Genetics
    2. Laboratory of Neurogenetics, Institute Born-Bunge
    3. University of Antwerp
    Search for more papers by this author
  • Peter De Jonghe,

    1. VIB-Department of Molecular Genetics
    2. Laboratory of Neurogenetics, Institute Born-Bunge
    3. University of Antwerp
    4. Division of Neurology, University Hospital Antwerp, Antwerp, Belgium
    Search for more papers by this author
  • Federico Zara,

    1. Muscular and Neurodegenerative Disease Unit, Institute G. Gaslini, University of Genova, Genova
    2. Laboratory of Genetics, E.O. Ospedale Galliera, Genova
    Search for more papers by this author
  • Renzo Guerrini

    1. IRCCS Stella Maris, Pisa, Italy
    Search for more papers by this author

Author correspondence to Dr Carla Marini, Pediatric Neurology Unit and Laboratories, Children’s Hospital A. Meyer-University of Florence, Viale Pieraccini 24, 50139 Firenze, Italy. E-mail:


Dravet syndrome (DS), otherwise known as severe myoclonic epilepsy of infancy (SMEI), is an epileptic encephalopathy presenting in the first year of life. DS has a genetic etiology: between 70% and 80% of patients carry sodium channel α1 subunit gene (SCN1A) abnormalities, and truncating mutations account for about 40% and have a significant correlation with an earlier age of seizures onset. The remaining SCN1A mutations comprise splice-site and missense mutations, most of which fall into the pore-forming region of the sodium channel. Mutations are randomly distributed across the SCN1A protein. Most mutations are de novo, but familial SCN1A mutations also occur. Somatic mosaic mutations have also been reported in some patients and might explain the phenotypical variability seen in some familial cases. SCN1A exons deletions or chromosomal rearrangements involving SCN1A and contiguous genes are also detectable in about 2–3% of patients. A small percentage of female patients with a DS-like phenotype might carry PCDH19 mutations. Rare mutations have been identified in the GABARG2 and SCN1B genes. The etiology of about 20% of DS patients remains unknown, and additional genes are likely to be implicated.

Dravet syndrome (DS) is an increasingly recognized epileptic encephalopathy in which the clinical diagnosis is supported by the finding of sodium channel gene mutations in approximately 70–80% of patients. Mutations of ion channel genes play a major role in the pathogenesis of a range of epilepsy syndromes, resulting in some epilepsies being regarded as channelopathies. Voltage-gated sodium channels (VGSCs) play an essential role in neuronal excitability; therefore, it is not surprising that many mutations associated with DS have been identified in the gene encoding a VGSC subunit.

Sodium Channel α1 Subunit Gene (SCN1A) and Dravet Syndrome

Mutations of the gene coding for the α1 subunit of the sodium channel (SCN1A) were first discovered in the epilepsy syndrome of genetic (formerly generalized) epilepsy with febrile seizure plus (GEFS+) (Escayg et al., 2000). A family history of epilepsy or febrile seizures is sometimes found in patients with DS and most affected relatives have GEFS+ phenotypes (Singh et al., 2001). In view of the predisposition to seizures with fever that typically occurs in Dravet syndrome, Claes et al. (2001) performed mutational analysis on SCN1A in seven children and all had de novo mutations.

At present, >500 mutations have been associated with DS and are randomly distributed along the gene (Mulley et al., 2006; Marini et al., 2007; Depienne et al., 2009a). The frequency of SCN1A mutations in DS is approximately 70–80%. Sequencing mutations are found in about 70% of cases and comprise truncating (40%) and missense mutations (40%) with the remaining being splice-site changes. Most mutations are de novo, but familial mutations occur in 5–10% of cases and are usually missense in nature (Claes et al., 2001; Nabbout et al., 2003; Wallace et al., 2003; Fujiwara et al., 2003). In these cases, other family members with the SCN1A mutation have mild phenotypes consistent with the GEFS+ spectrum (Nabbout et al., 2003; Fujiwara et al., 2003). An important genetic counseling issue is that several families have been reported with more than one child with DS. Somatic or germline mosaic mutations might explain the presence of an unaffected or mildly affected transmitting parent (Depienne et al., 2006; Gennaro et al., 2006; Marini et al., 2006; Morimoto et al., 2006). A recent study indicates that mosaicism is found in at least 7% of families with DS (Depienne et al., 2010). The proportion of the mutated allele in the blood of the 12 patients described varied from 0.04% to 85% and may be a significant factor underlying intrafamilial phenotypic variability (Depienne et al., 2010).

Patients with a clinical diagnosis of DS who test negative for SCN1A sequence–based mutations may still have SCN1A exonic deletions or chromosomal rearrangements involving SCN1A and contiguous genes (Madia et al., 2006; Mulley et al., 2006; Suls et al., 2006; Marini et al., 2007; Wang et al., 2008). Such abnormalities are identified with multiplex ligation-dependent probe amplification (MPLA) and further characterized by comparative genome hybridization (CGH) to determine the size of the abnormality and the identity of any additional gene(s) involved (Marini et al., 2009). Haplotype analysis with microsatellite markers and single nucleotide polymorphisms (SNPs) and multiplex amplicon quantification (MAQ) can also be used to identify small chromosomal abnormalities affecting SCN1A (Madia et al., 2006; Suls et al., 2006). Intragenic and whole gene deletions including only SCN1A and/or contiguous genes account for 2–3% of all DS cases and for about 12.5% of patients with DS who are mutation-negative on classical Sanger sequencing (Marini et al., 2009). Duplications and amplifications involving SCN1A are additional, rare, molecular mechanisms (Marini et al., 2009). Microchromosomal rearrangements extending beyond SCN1A and including variable numbers of contiguous genes are potentially associated with additional dysmorphic features, depending on the genes involved, or with a more severe epilepsy phenotype when other VGSC α subunit genes clustered on chromosome 2q such as SCN2A, SCN3A, SCN7A, and SCN9A are involved. Suls et al. (2006) reported three patients with microchromosomal rearrangements extending beyond SCN1A with two patients showing additional clinical features; however, the third patient did not have other features, suggesting that haploinsufficiency of the additional deleted genes in that case did not result in overt features. Other studies have also shown that there are no significant clinical differences between DS patients in whom the deletions involved only SCN1A and those in whom contiguous genes were also removed (Mulley et al., 2006; Marini et al., 2009). Haploinsufficiency of more than one VGSC α subunit gene might be expected to cause a more severe epilepsy phenotype; yet patients with large deletions including additional α subunit genes have a DS phenotype indistinguishable from those carrying point mutations (Marini et al., 2009). However, because the DS phenotype is usually severe, it may mask subtle clinical differences that may be due to nearby genes.

Nakayama et al. (2010) reported two patients with DS phenotype with heterozygous microdeletions removing the 5′ noncoding exons and regions with promoter activity but not affecting the coding exons. These findings further confirm the critical and predominant involvement of SCN1A in the molecular pathology of DS.

The clinical observation of a frequent family history of epilepsy and FS and familial cases of DS are difficult to reconcile with the finding that most SCN1A abnormalities in DS are de novo. In such families where the proband with DS has a de novo SCN1A mutation, the mode of inheritance is likely to be polygenic and SCN1A is one of the genetic determinants specifically associated with DS but not with the other phenotypes. Indeed, a patient with DS with variants in both SCN1A and SCN9A has been reported, suggesting that mutations of SCN9A might cause FS and be modifiers for the DS phenotype and provide support for the role of additional genes modifying a gene of major effect (SCN1A) in DS (Singh et al., 2009).

What Causes DS When There Is No Detectable SCN1A Involvement?

Mutations of protocadherin 19 (PCDH19, on chromosome Xq22) have been found in a disorder called “epilepsy limited to females with mental retardation” (EFMR) (OMIM#300088) (Scheffer et al., 2008; Dibbens et al., 2008). EFMR is a disorder with an unusual X-linked inheritance pattern, where the disorder is expressed in heterozygous females, whereas hemizygous males are unaffected carriers (Dibbens et al., 2008). The biologic role of PCDH19 is unknown; this gene is expressed in developing brains of humans and mice, and is postulated to be involved in establishing neuronal connections and signal transduction at the synaptic membrane (Yagi & Takeichi, 2000).

Depienne et al. (2009b) reported PCDH19 familial and “de novo” point mutations in 13 female patients with early onset epileptic encephalopathy mimicking DS. A male patient with a similar phenotype carried a mosaic PCDH19 mutation. The authors estimated that 16% or 25%, if only female patients were included in the calculation, of their SCN1A-negative DS patients had PCDH19 mutations and that this gene might overall account for 5% of DS (Depienne et al., 2009b). In another study mutation screening of a large cohort of female patients with variable epilepsy phenotypes with infantile onset, revealed PCDH19 mutations in 13 probands, 7 of whom exhibited Dravet syndrome and 6 had focal epilepsy (Marini et al., 2010). Clinical similarities between PCDH19 mutation–positive patients and DS, including infantile onset of febrile and afebrile seizures, occurrence of hemiclonic seizures, regression following seizure onset, suggest that PCDH19 should be tested in DS patients who are negative on screening for mutations or genomic abnormalities of SCN1A.

A single case of DS with a mutation in the γ2 subunit gene of the GABAA receptor, GABRG2, has been reported (Harkin et al., 2002). Patino et al. in 2009 reported a single patient with DS carrying a homozygous mutation in SCN1B. This is the first DS patient with an autosomal recessive pattern of inheritance (Patino et al., 2009).

Genotype–Phenotype Correlations

SCN1A mutations have been associated predominantly with DS and GEFS+ spectrum and are characterized by marked phenotypic variability including age of seizure onset, seizure types and severity, as well as variable cognitive outcome. This raises the question: why do mutations, or deletions, in the same gene result in different phenotypes? Some general genotype–phenotype correlations have been suggested: truncating, nonsense, frame shift mutations, and partial or whole gene deletions are correlated with a classical DS phenotype and appear to have a significant correlation with an earlier age of seizure onset (Marini et al., 2007). Although it seems straightforward to infer that “severe” truncation mutations result in the more severe phenotype of DS and missense mutations cause GEFS+, the data do not support this conclusion. In DS, SCN1A missense mutations account for 40% of mutations and truncation mutations account for another 40% (Depienne et al., 2009a). The severity of the phenotypes may also be correlated with the location of SCN1A missense mutations: for example, those in the pore-forming region of the sodium channel may cause DS, whereas missense changes associated with the GEFS+ spectrum may be more frequently located outside the pore-forming region (Meisler & Kearney, 2005). However, this genotype–phenotype correlation is not consistently observed. Moreover, the same SCN1A mutations and deletions cause DS in some patients and GEFS+ in others (Escayg et al., 2000; Guerrini et al., 2010; Suls et al., 2010), suggesting that modifier genes, the genetic background, and/or environmental factors may also play a role in some patients, and thus DS may sometimes follow a complex model of inheritance.

In patients with PCDH19 mutations, a phenotype–genotype correlation seems to be emerging: These patients compared to DS patients with SCN1A abnormalities, had a slightly older age of onset, less frequent status epilepticus, rare photosensitivity, a less severe seizure disorder, less severe cognitive impairment, and only in a few patients were atypical absences and myoclonic seizures observed (Depienne et al., 2009b; Marini et al., 2010). Another key observation in these individuals was the inheritance pattern. For those girls with de novo mutations, the inheritance is similar to patients with SCN1A mutations. For those with inherited mutations from either an affected mother or an unaffected father, genetic counseling is essential for their families because of the 50% risk for further affected offspring (Depienne et al., 2009b; Marini et al., 2010).

SCN1A Mutations Associated with Other Early Onset Epileptic Encephalopathies: Distinct Syndromes or Extended DS Phenotypic Spectrum?

The term “borderline severe myoclonic epilepsy of infancy” (SMEB) refers to patients who lack several of the key features of DS such as myoclonic seizures or generalized spike-wave (GSW) activity. More than 70% of patients with SMEB carry SCN1A mutations that are spread throughout the gene with a mixture of mutations including truncating, missense, and splice-site changes (Harkin et al., 2007).

Infants with frequent and intractable generalized tonic–clonic seizures often induced by fever and beginning before 1 year of age, have been designated as having “intractable childhood epilepsy with generalized tonic-clonic seizures (ICEGTC)” (Fujiwara et al., 2003). They follow a developmental trajectory similar to that of children with DS in that development slows in the second year of life and intellectual outcome is poor. The major feature that differentiates DS and ICEGTC is the presence of other seizure types such as myoclonic, absence, and complex partial seizures. These patients also have a high frequency of SCN1A mutations (Fujiwara et al., 2003).

Five patients with early onset multifocal seizures, developmental slowing, and abundant multifocal epileptiform activity have been designated as “severe infantile multifocal epilepsy” (SIMFE); the three with normal early development carried SCN1A mutations, whereas the two who were never normal did not (Harkin et al., 2007). The clinical keys that distinguish these patients from DS are the picture of focal epilepsy rather than generalized epilepsy with absent or rare myoclonic and absence seizures and the lack of GSWs on EEG.

Myoclonic seizures, however, are not always present in classical DS. The term “severe myoclonic epilepsy of infancy” had already been debated as potentially misleading prior to the discovery of the molecular basis of DS in 2001, and the name “severe polymorphic epilepsy of infants” had been proposed (Aicardi, 1994). Moreover, patients usually develop myoclonic seizures after 1or 2 years of age so that waiting for the evolution of myoclonic seizures could delay the identification of this syndrome. Yet, the question of whether myoclonic seizures are an essential component of DS remains controversial. The identification of SCN1A mutations in similar proportions of classical DS, SMEB, ICEGTC, and SIMFE patients (Fujiwara et al., 2003; Harkin et al., 2007) confirms that these phenotypes share the same genetic determinant in the majority of patients. Therefore, it may be better to regard these phenotypes as a continuum of the same disorder, which could be considered the DS spectrum.

Functional Effects of Epileptogenic Voltage-Gated Sodium Channel Mutations

Based on the number of epilepsy cases in which mutations have been identified, SCN1A is currently the most clinically relevant epilepsy gene. Most SCN1A mutations responsible for GEFS+ are missense, whereas mutations responsible for the most severe epileptic encephalopathies can either be missense or cause truncation that would give rise to nonfunctional channels. When studied in expression systems, missense SCN1A mutants can show loss or gain of function; these effects are consistent with either decreased or increased neuronal excitability and cannot be unified into a clear pathophysiologic mechanism (Avanzini et al., 2007). However, several lines of evidence point to loss of function as the main effect of SCN1A mutations: More than half of the identified DS mutants are predicted to be nonfunctional (Catterall et al., 2008; Meisler & Kearney, 2005; Mulley et al., 2005) and a review of the published functional effects of SCN1A missense mutations shows that most of them cause a reduction in Na+ current (Ragsdale, 2008).

Epilepsy is a disorder characterized by brain hyperexcitability and it has initially seemed puzzling that mutations in a VGSC lead to loss of function and reduced Na+ current; however, data from Scn1a knockout mice (Yu et al., 2006) and knockin mice expressing the non-sense scn1a-R1407X DS mutant (Ogiwara et al., 2007) have shown that the α1 subunit is fundamental for the excitability of at least some types of γ-aminobutyric acid (GABA)ergic interneurons in the neocortex and hippocampus. Therefore, reduced firing of inhibitory neurons and compromised network inhibition could be the major pathophysiologic mechanism causing SCN1A-related genetic epilepsies.


About 30 years after the first clinical description of the distinctive epilepsy syndrome of DS, the genetic etiology of 70–80% of patients has been solved. SCN1A harbors the vast majority of genetic abnormalities including sequence-based mutations, single or multiple exon deletions, or chromosomal rearrangements involving both the entire gene and/or other contiguous genes. Most SCN1A abnormalities are de novo, yet familial cases do exist. Phenotypic variability might be partially explained by mosaic mutations, modifier genes, and genetic background effects, as well as the variable functional effects of SCN1A mutations. A small percentage of female patients with a DS-like phenotype carry PCDH19 mutations. Rare patients have been identified with GABRG2 and SCN1B mutations. Early diagnostic recognition and more appropriate treatment choices provide higher chances of a better seizure outcome and reduced cognitive impairment. The etiology of DS in about 20% of patients with the syndrome remains unknown and additional genes are likely to be implicated.


The authors declare no relevant conflicts of interest.

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.