Does a SCN1A gene mutation confer earlier age of onset of febrile seizures in GEFS+?

Authors


Address correspondence to Professor Ingrid E. Scheffer, Director of Paediatrics, Austin Health, Repatriation Hospital Campus, Level 1, Neurosciences Building, Banksia Street, Heidelberg, Victoria 3081, Australia. E-mail: scheffer@unimelb.edu.au

Summary

SCN1A is the most clinically relevant epilepsy gene and is associated with generalized epilepsy and febrile seizure plus (GEFS+) and Dravet syndrome. We postulated that earlier onset of febrile seizures in the febrile seizure (FS) and febrile seizure plus (FS+) phenotypes may occur in the presence of a SCN1A mutation. This was because of the age-related onset of Dravet syndrome, which typically begins in the first year of life. We found that patients with FS and FS+ with SCN1A mutations had earlier median onset of febrile seizures compared to the population median. Patients with GABRG2 mutations had a similar early onset in contrast to patients with SCN1B mutations where onset was later. This study is the first to demonstrate that a specific genetic abnormality directly influences the FS and FS+ phenotype in terms of age of onset.

Febrile seizures are the most common form of seizures and occur in 2–5% of children. These seizures begin between the age of 3 months and 6 years, with a median age of onset of 18 months (Offringa et al., 1991). Febrile seizures have a genetic basis and usually follow complex inheritance, where multiple genes and environmental factors play a role.

Children with febrile seizures may be part of a family with the familial epilepsy syndrome of generalized epilepsy and febrile seizure plus (GEFS+) (Scheffer & Berkovic, 1997). Febrile seizure (FS) and febrile seizure plus (FS+) are the two most common phenotypes in GEFS+. Three genes have been confirmed to be associated with the GEFS+ syndrome but have been identified in only 10–20% families studied: two sodium channel subunit genes encoding the alpha 1 and beta 1 subunit (SCN1A, SCN1B) and the gamma 2 γ-aminobutyric acid (GABA)A receptor subunit gene, GABRG2 (Wallace et al., 1998, 2001a,b, 2002). The most commonly reported GEFS+ gene is SCN1A, in which missense mutations are found.

Dravet syndrome is associated with SCN1A mutations in about 70–80% cases, and represents the severe end of the GEFS+ spectrum. SCN1A mutations in Dravet syndrome include missense, truncation, or copy number variation. Marini et al. (2007) found that patients with Dravet syndrome and a SCN1A mutation had an earlier onset of seizures with fever compared to those who were mutation negative.

We, therefore, hypothesized that patients with FS and FS+ in GEFS+ families with SCN1A mutations who first presented with a febrile seizure, had an earlier age of onset compared to the age of onset of febrile seizures in the general population. We also examined the age of onset of febrile seizures in patients with SCN1B or GABRG2 mutations. Whether the nature of the GEFS+ gene alters the age of onset of febrile seizures has not been determined.

Patients and Methods

Families

Twelve multigenerational families were identified with the GEFS+ syndrome and a known mutation of major effect. There were four families with SCN1A mutations (Scheffer & Berkovic, 1997; Wallace et al., 2001b); (1 unpublished family), six families with SCN1B mutations (Wallace et al., 1998, 2002; Singh et al., 1999; Scheffer et al., 2007), and two families with GABRG2 mutations (Wallace et al., 2001a, Harkin et al., 2002;Marini et al., 2003). Patients with a pure phenotype of FS or FS+ who presented initially with a febrile seizure were included. If febrile seizures occurred after afebrile seizures, the subjects were not included in this cohort.

Epilepsy phenotypes

The phenotypes of affected family members were classified according to the international classification of epilepsy syndromes (Commission, 1989) and the description of GEFS+ phenotypes (Scheffer & Berkovic, 1997; Singh et al., 1999). In FS, seizures occurred between 3 months and 6 years of age. If seizures persisted beyond the age of 6 years and/or were accompanied by afebrile convulsions then the diagnosis was FS+. Patients with extended phenotypes were excluded, including patients with partial seizures (Scheffer et al., 2007).

Literature review

A literature search was performed to review febrile seizure onset in other families with known genetic mutations (Abou-Khalil et al., 2001; Escayg et al., 2001; Kananura, et al., 2002; Audenaert et al., 2003; Spampanato et al., 2004; Nagao et al., 2005, Pineda-Trujillo et al., 2005; Audenaert et al., 2006; Barela et al., 2006; Colosimo et al., 2007). Because some syndromes were phenotyped according to different definitions, these were modified to be consistent with the definitions used here. For example, some cases were called FS; however, seizure offset occurred after 6 years so we reclassified these cases as FS+ (Barela et al., 2006). In other cases, febrile seizures and afebrile generalized seizures were noted and we reclassified these as FS+ (Abou-Khalil et al., 2001; Barela et al., 2006). If partial seizures were documented, these patients were not included in our analysis.

Statistical analysis

Mann-Whitney U test was used to determine the difference in the median age of onset between different genes. Statistical analysis was performed using the statistical program, SPSS version 11.5 (SPSS, Chicago, IL, U.S.A.).

Results

Epilepsy Research Centre data

A total of 105 patients in 12 families for whom we had detailed clinical information were identified with SCN1A, SCN1B, or GABRG2 mutations. Sixty-two patients presented with a febrile seizure where the age of onset was known and had a FS or FS+ phenotype. An additional 43 patients were excluded for the following reasons: they had a more complex phenotype than FS and FS+ (31), their initial seizure was an afebrile seizure or it was unclear whether a fever was present (6), or their age of onset was imprecise (6).

The median age of onset of FS and FS+ was 12 months in subjects with SCN1A and GABRG2 mutations (Table 1, Fig. 1). The median age of onset for children with an SCN1A mutation was significantly lower than for children with an SCN1B mutation (p = 0.001, Mann–Whitney U test). There was no significant difference in onset between individuals with SCN1A and GABRG2 mutations.

Table 1.   Comparison of age of onset of FS and FS+ in SCN1A, SCN1B, and GABRG2 families
StudiesGene mutationNumber ofAge of onset (months)References
FamiliesFS subjectsFS+ subjectsTotal subjectsMedian Value Range (min–max)p-value (compared to SCN1A)
  1. FS, febrile seizure; FS+, febrile seizure plus; GEFS+, generalized epilepsy and febrile seizure plus; SCN1A, sodium channel alpha1 subunit gene; SCN1B, sodium channel beta1 subunit gene; GABRG, gamma 2 GABAA receptor subunit gene; min, minimum; max, maximum; p-value derived using Mann-Whitney U test.

  2. aStatistically significant.

Our seriesSCN1A49112012.06–108  
SCN1B615132824.04–420.001a
GABRG221131411.56–240.986
Other reported familiesSCN1A712112311.05–38 (Abou-Khalil et al., 2001; Escayg et al., 2001; Spampanato et al., 2004; Nagao et al., 2005; Pineda-Trujillo et al., 2005; Barela et al., 2006; Colosimo et al., 2007)
SCN1B131430.012–360.033a(Audenaert et al., 2003)
GABRG2230318.013–360.158(Audenaert et al., 2006; Kananura, et al. 2002)
CombinedSCN1A1121224312.05–108  
SCN1B718143224.04–42<0.001a
GABRG241431713.06–360.559
Figure 1.


Box plot showing the median age of onset of febrile seizure (FS) in individuals in our families compared with families in the literature. There is a significantly earlier age of onset in patients with SCN1A mutations compared to individuals with SCN1B mutations in both groups. Dark lines denote median, boxes denote interquartile ranges (25–75th percentiles), and error bars show 95% confidence intervals.

Analysis of families from the literature

We performed a similar analysis on the families in the literature who had SCN1A, SCN1B, or GABRG2 mutations. Of the 12 families with mutations, only 10 were included in our analysis, as there were insufficient clinical data on two families (Sugawara et al., 2001; Lossin et al., 2003). The median age of onset for mutations in each gene based on the clinical data of the reported families is shown in Table 1. Similar findings emerged to our dataset, with the median age of onset of patients with SCN1A mutations being 11 months, and a significantly earlier median onset in children with SCN1A mutations compared with SCN1B mutations.

Discussion

The critical observation was that in two independent cohorts of individuals with FS or FS+ and SCN1A mutations, the median age of onset was 11–12 months compared with the population median of 18 months (Offringa et al., 1991). This is the first demonstration that the specific gene associated with benign phenotypes such as FS or FS+ may affect age of onset of febrile seizures. This is not surprising given the frequent onset of SCN1A-related epileptic encephalopathies beginning in individuals younger than 1 year of age (Harkin, et al. 2007; Marini et al., 2007). This is likely to reflect age-related developmental expression of the SCN1A protein, rendering the individual susceptible to seizures from about 6 months. This is supported by the developmental expression pattern of SCN1A recently demonstrated in mouse brain (Ogiwara et al., 2007).

It is enticing that our data also suggest an earlier median onset in individuals with GABRG2 mutations, but not SCN1B mutations, suggesting that this is not specific to sodium channel abnormalities. However, the numbers of patients in the SCN1B and GABRG2 groups were small.

The cause of the earlier age of onset of febrile seizures in patients with SCN1A mutations remains unclear. Although the genetic heterogeneity of febrile seizures is well accepted, this is the first demonstration that a specific genetic abnormality directly influences the phenotype in terms of age of onset. This is likely to reflect the maturation dependent biologic vulnerability of the sodium channel that is critical to neuronal excitability. Functional studies examining the age of developmental expression of different ion channel subunits may help to elucidate this hypothesis.

Acknowledgments

We thank the families who have participated in our research. Dr Sijben was supported by Nationaal Epilepsie Fonds of The Netherlands. Dr Sithinamsuwan was supported by funding from Phramongkutklao Hospital, Royal Thai Army. Support was also obtained from the National Health and Medical Research Council of Australia.

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.

Disclosure: None of the authors has any conflict of interest to disclose.

Ancillary