Childhood absence epilepsy [CAE (ECA MIM 600131, ECA2 MIM 607681, ECA3 MIM 607682, ECA4 MIM 611136)] is a classical idiopathic generalized epilepsy (IGE) syndrome and accounts for up to 15% of childhood epilepsies (Loiseau et al., 2002). CAE usually begins in children between 2 and 8 years of age and is characterized by frequent absence seizures associated with generalized and symmetrical 3-Hz spike waves. Based on familial aggregation and twin studies, genetic factors are believed to play an important role in this epileptic syndrome (Helbig et al., 2008). Over the last decade, several genes underlying monogenic forms of epilepsy have been identified in families with absence seizures. CAE also shows significant clinical and genetic overlaps with other classical IGE syndromes. More specifically, mutations in various subunits of the γ-aminobutyric acid (GABA)A receptor have been found in families with generalized epilepsy with febrile seizure plus (GEFS+) (Baulac et al., 2001; Dibbens et al., 2004), autosomal dominant juvenile myoclonic epilepsy (Cossette et al., 2002), and childhood absences epilepsy with or without febrile seizures (Wallace et al., 2001a; Marini et al., 2003; Maljevic et al., 2006). Tanaka et al. (2008) recently reported three mutations in the GABRB3 gene, encoding for the β3 subunit of the GABAA receptor, in families with CAE. In this study, we sought to determine the frequency of mutation in this latter gene in our cohort of families with CAE and other related IGE syndromes.
Mutations in the GABRB3 have been recently associated with childhood absence epilepsy (CAE) in families from Honduras and Mexico. In this study, we aimed to determine the frequency of mutation in this gene in our cohort of families with CAE and other related idiopathic generalized epilepsy (IGE) syndromes. We screened the open reading frame of GABRB3 in 183 French-Canadian individuals with IGE, including 88 with CAE. A total of nine single nucleotide polymorphisms (SNPs) have been identified, five of which are novel. The previously described P11S missense mutation was found in three affected and one unaffected individuals from a French-Canadian family. However, the P11S variant was also found in one of our 190 control individuals of French-Canadian origin, suggesting that this variant is rather a rare polymorphism in this population. Further screening of other IGE cohorts from various ethnic origins would help to confirm the association between this rare functional variant and epilepsy.
Materials and Methods
We screened the complete open reading frame of GABRB3 in 183 French-Canadian individuals with IGE, including 88 with CAE. The methods used for the clinical characterization of this cohort have been described previously (Kinirons et al., 2008). We amplified by polymerase chain reaction (PCR) (Applied Biosystems 9700; Applied Biosystems, Foster City, CA, U.S.A.) all the coding regions of this gene. We designed all primers at least 50 bp from the exon–intron boundaries to examine the entire exonic sequences of the gene and splice sites. PCR primer pairs were designed using the software Exon primer from the UCSC Genome Browser database using NM_021912 and NM_000814 sequences. Primer sequences and PCR conditions are available upon request. The amplicons were run on the ABI3730 automatic sequencer (Applied Biosystems) and analyzed by using Mutation Surveyor software version 3.0. All mutations and variants were confirmed by resequencing. All novel single nucleotide polymorphisms (SNPs) and known mutations were tested in a French-Canadian control population of 190 individuals. In order to exclude symptoms compatible with seizures as well as any familial history of epilepsy, a systematic questionnaire was used with these control individuals. The main aim of this study was to assess whether mutations or rare variants in GABRB3 are associated with epilepsy in our cohort. We did not wish to perform a systematic association study. However, because we generated many novel genetic variants in the GABRB3 gene, we compared the relative frequencies of these genotypes between affected individuals and controls, although the relatively low sample size limits the power of these analyses. The difference in the frequency of genotypes between cases and controls was assessed by using chi-square test for SNP 1 and SNP 6. The association between these genotypes and epilepsy was estimated by a logistic regression analysis [odd ratios (ORs); 95% confidence interval (CI)]. The low frequency of minor allele for the other new variants made the statistical analysis nonapplicable.
We found a c.G31A nucleotide change located in exon 1a and predicting for a P11S variation, in a proband with CAE (IV-01) (Fig. 1). This individual had typical absence seizures starting at the age of 6, and her electroencephalography (EEG) showed generalized spike and waves (GSW) discharges enhanced by hyperventilation. The GABRB3P11S variant was found in two additional affected individuals from the same family. Individual IV-02 exhibits abnormal EEG, showing posterior GSW discharges. However, this latter individual remains so far without clinical seizures at the age of 14. Another mutation carrier (III-01) has had a total of three unprovoked seizures at the age of 40 years old, for which no etiology could be identified. In particular, EEG and brain scan were normal. Interestingly, individual II-03 also had late onset (40 yo) generalized tonic–clonic seizures. Unfortunately, he died prematurely from pulmonary embolism and neither additional clinical information nor DNA was available for analysis. Finally, the GABRB3P11S variant was also found in an asymptomatic individual (II-02). EEG recording was not obtained in this latter individual. In contrast to the Mexican population (Tanaka et al., 2008), in our population the GABRB3P11S variant was present in one of the 378 control chromosomes. There is thus no significant difference in the frequency of the P11S variant between the epileptic cohort and the control group (χ2 = 0; p = 0.982; OR = 1.033). We have also identified a total of eight other SNPs, five of which are novel (one intronic, four silents). The following phenotypes are associated with these novel SNPs: CAE (n = 1) for SNP3, IGE-NOS (n = 2) and CAE (n = 1) for SNP5, IGE-NOS (n = 11) and CAE (n = 7) for SNP6, IGE-NOS (n = 1) for SNP7, and CAE (n = 1) for SNP8 (Table 1). One of these novel SNPs was found in the intron 1a close to the exon 1a and does not affect the predicted consensus splice site. Its presence in the control population was not assessed. We sought for the presence of the four novel silent variations in 190 control individuals from the French-Canadian population. The p.T201T, p.F314F, and p.G335G were all detected in a single affected individual but not in the control group. In turn, the p.S261S was found in both the affected and control groups. Unexpectedly, the frequency of this variant was significantly higher in the control group (χ2 = 33.9; p = 0.001; OR = 0.201) (Table 1). SNPs that were already listed in the NCBI database were not tested in the control population.
|SNPs||Nucleotide genomic||Allele major/minor||AA||Rs||Maf||χ2||p-Value||Odds ratio (95% CI)|
Recently, Tanaka et al. (2008) showed that loss of function mutations in GABRB3 are associated with CAE, including for the GABRB3P11S variation found in one Mexican family with four CAE individuals and one Mexican singleton. In this study, we report the same P11S variation in GABRB3 in four individuals from a French-Canadian family. In contrast to the Mexican families, in the French-Canadian family described here, only one GABRB3P11S mutation carrier is affected with CAE, whereas the other affected carrier exhibit nonspecific and late-onset generalized epilepsy. Among the two asymptomatic carriers, one shows GSW discharges. Although we cannot exclude that GSW discharges may have been present during childhood in the three additional carriers, this observation suggests that the GABRB3P11S mutation may be associated with different epilepsy phenotypes, together with low penetrance of the epileptic trait. Such a reduced penetrance of mutations associated with familial epilepsy has been reported previously (Steinlein et al., 1995; Wallace et al., 2001b). In addition, even for the CAE phenotype, the prevalence of the GABRB3 mutations is lower in our population (one mutation in 88 CAE) compared to the population of Central America (four mutations in 48 CAE). However, whereas Tanaka et al. (2008) did not find the GABRB3P11S variation in their controls, we detected it in one control individual from the French-Canadian population. Overall, our study nonetheless suggests that the P11S mutation in GABRB3 is associated with CAE and related phenotypes.
The search for common genetic variants predisposing to epilepsy allowed the identification of polymorphisms that may have a small effect on the phenotype, although a small fraction of these studies have been so far replicated. In turn, the search for rare genetic variants having a more dramatic contribution to the epilepsy phenotype such as in monogenic epilepsies was found to be much more successful (Steinlein et al., 1995; Baulac et al., 2001; Wallace et al., 2001b; Cossette et al., 2002; Tanaka et al., 2008). We propose that the rare P11S variant found in GABRB3 may fall between these two extremes. Our observation is consistent with the fact that rare missense variants (<1% of the population) are associated with mildly deleterious effects (Helbig et al., 2008). Such rare functional genetic variations may thus be one of the genetic bases for complex diseases including epilepsy.
In conclusion, we have found that GABRB3P11S is a rare missense variation segregating with CAE, asymptomatic GSW discharges, and late-onset generalized seizures in a small French-Canadian family. Screening for GABRB3 variants in additional IGE cohorts is warranted to assess the relative contribution of these genetic variants and mutations to the various IGE phenotypes.
The authors thank the families for their participation to this study. PLT and CM are supported by the Savoy Foundation. PC is a clinical scientist from the Canadian Institutes of Health Research (CIHR). This study has been funded by the Fonds de la recherché en santé du Québec (FRSQ) and the CIHR.
None of the authors has any conflict of interest to disclose. 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.