A major challenge in understanding complex idiopathic generalized epilepsies has been the characterization of their underlying molecular genetic basis. Here, we report that genetic variation within the GABRD gene, which encodes the GABAA receptor δ subunit, affects GABA current amplitude consistent with a model of polygenic susceptibility to epilepsy in humans. We have found a GABRD Glu177Ala variant which is heterozygously associated with generalized epilepsy with febrile seizures plus. We also report an Arg220His allele in GABRD which is present in the general population. Compared with wild-type receptors, α1β2Sδ GABAA receptors containing δ Glu177Ala or Arg220His have decreased GABAA receptor current amplitudes. As GABAA receptors mediate neuronal inhibition, the reduced receptor current associated with both variants is likely to be associated with increased neuronal excitability. Since δ subunit-containing receptors localize to extra- or peri-synaptic membranes and are thought to be involved in tonic inhibition, our results suggest that alteration of this process may contribute to the common generalized epilepsies.
Of the known genes for the relatively simple monogenic idiopathic epilepsies, most encode voltage- or ligand-gated ion channels (1,2). However, the genetic basis for the common generalized epilepsies is complex. Even in rare families segregating identified major genes, there are issues of incomplete penetrance, variable phenotypes and severity, and often unrelated epilepsies that complicate interpretation. Generalized epilepsies predominantly consist of a large group traditionally named idiopathic generalized epilepsies (IGEs), and a more recently recognized group termed generalized epilepsy with febrile seizures plus (GEFS+). Although these two groups form reasonably distinct clinical clusters, an overlap of phenotypes is seen in some families (3), suggesting shared genetic determinants.
Mutations in the GABAA receptor subunit genes have recently been shown to contribute to both IGE and GEFS+ in rare large families. The GABAA receptor subunit gene GABRG2 is implicated in childhood absence epilepsy (CAE), a form of IGE, febrile seizures (FS) (4,5) as well as GEFS+ (6,7). The GABAA receptor subunit gene GABRA1 is associated with juvenile myoclonic epilepsy (JME) (8), another common form of IGE. The δ subunit has been shown to assemble with α and β subunits to form a relatively non-desensitizing αβδ channel (9,10) and has an extra- or peri-synaptic localization (11,12). We hypothesized that mutation of the GABRD gene may be associated with human epilepsies.
Seventy-two unrelated IGE, 65 unrelated GEFS+ and 66 unrelated FS patients were screened for mutations in the GABRD gene. Two putative missense mutations in GABRD were identified: Glu177Ala resulting from the nucleotide substitution c. 530A→C in exon 5 was detected in a small GEFS+ family (Fig. 1A) and Arg220Cys resulting from c. 658C→T in exon 6 was detected in a second small GEFS+ family. Both changes were heterozygous and neither was detected in 192 control chromosomes derived from 96 anonymous blood donors primarily of Caucasian origin.
The two amino acid residues found to be altered, Glu177 and Arg220, are both located in the amino-terminal extracellular domain of the δ subunit protein. Glu177 is positioned immediately adjacent to one of the two invariant cysteines (C) in the extracellular domain that form a disulfide bond and Arg220 is located near the first transmembrane domain (TM1) (13). Electrophysiological studies were carried out by recording recombinant α1β2Sδ GABAA receptors expressed in HEK293T cells. However, we did not observe a significant alteration of GABA EC50 for the receptors containing the Glu177Ala, Arg220Cys or Arg220His (see below) variant when compared with wild-type receptors (data not shown), suggesting that these variants do not significantly alter agonist binding.
We tested the functional significance of the GABRD Glu177Ala and Arg220Cys amino acid changes by application of a saturating concentration of GABA (1 mM). Receptors heterozygous and homozygous for Glu177Ala were found to have significantly reduced maximal current compared with wild-type receptors (Fig. 2). As the Glu177Ala GABRD variant does not segregate monogenically with epilepsy in a large family, we hypothesize that it represents a susceptibility allele for polygenic epilepsy. An increase in neuronal excitability due to reduced GABAA receptor current is proposed to contribute to the GEFS+ phenotype of patients heterozygous for Glu177Ala. Maximal currents from receptors heterozygous and homozygous for Arg220Cys did not differ from wild-type receptors (data not shown), suggesting that more extensive studies of ion channel properties are required to explore the possibility of a functional effect, or that this change may be a neutral rare variant.
We also identified the base change c. 659G→A in GABRD exon 6, which results in the substitution of an arginine with a histidine at amino acid position 220 (Arg220His), the same residue position as the variant Arg220Cys. Arg220His was found to be carried by IGE, GEFS+ and FS patients as well as by control blood bank individuals. To explore effects of the Arg220His polymorphism we recorded currents evoked by a saturating concentration of GABA (1 mM). Receptors heterozygous for Arg220His had significantly decreased peak current in comparison with wild-type controls (Fig. 3). We propose that GABRD Arg220His combines additively as a susceptibility allele with other yet to be identified susceptibility alleles responsible for the complex epilepsies.
Next, we explored naturally occurring Arg220His homozygosity. In our screening of 203 unrelated IGE and GEFS+ patients and 96 anonymous blood bank controls, the Arg220His allele was homozygous in one individual (expect 1 in 2500), a patient with JME (Fig. 1B). This raised the possibility that homozygosity of Arg220His contributes to the pathogenesis of JME. Receptors homozygous for Arg220His showed a greater reduction in current amplitude than that recorded in heterozygotes (Fig. 3), but the difference between the heterozygotes and homozygotes did not reach significance. We screened an extra 81 JME patients for changes in GABRD, bringing the total number of JME patients screened to 110, but did not find any additional Arg220His homozygotes or any additional coding variants in the GABRD gene. Homozygosity of Arg220His or mutations in GABRD, therefore, do not appear to be a common factor in JME; however, homozygosity or heterozygosity of the Arg220His allele may increase neuronal excitability and associated susceptibility to the epilepsies with a polygenic basis.
The electrophysiological data suggest that carriers of the Glu177Ala and Arg220His variants have increased neuronal excitability. However, given the similar carrier frequencies of Arg220His in controls (4.2%) and in patients with FS (4.5%), GEFS+ (3.1%), IGE (8.3%) and JME (3.7%), Arg220His is not a common contributing factor in IGE, FS or GEFS+. Allele frequency data are homogeneous among all groups (χ24=4.3; 0.10<P<0.95) and between IGE and controls (χ21=0.1; 0.10<P<0.95) where the most extreme difference between allele frequencies (4.9% in IGE, 2.1% in controls) was observed. Much larger sample sizes would be required in order to explore the possibility of statistical associations.
It has been suggested that GABAA receptors in vivo may be composed mainly of αβγ and αβδ isoforms (14). Recombinant αβγ and αβδ receptors expressed in mammalian cells exhibit different kinetic properties: αβγ receptor currents are rapidly desensitizing but αβδ receptor currents are relatively non-desensitizing (9,15–17). Furthermore, the αβγ isoforms are mainly subsynaptic, but αβδ isoforms are selectively targeted to extra- or peri-synaptic membranes (18,12), suggesting that αβγ receptors may mediate ‘phasic’ inhibition while αβδ receptors may mediate ‘tonic’ inhibition (19–21). Previously, mutations in GABRG2 found in rare IGE and GEFS+ families suggested that compromise of phasic inhibition can cause generalized epilepsies. The present study observed that the GABRD changes in GEFS+ and IGE exhibited decreased maximal GABAA receptor currents, implicating deficiency in tonic inhibition in these epilepsies.
Our findings suggest that the variant GABRD Glu177Ala contributes to GEFS+ and that Arg220His contributes to IGE and possibly other polygenic epilepsies. It remains to be understood why two GABRD variants, which both lead to reduced GABAA receptor currents, are associated with different generalized epilepsy subtypes. Why the few pathogenic mutations seen in GABRG2 and GABRA1 are involved in generalized epilepsy phenotypes as diverse as CAE, FS, GEFS+ and JME (4–7) adds to the puzzle. At least two hypotheses are possible. Allelic heterogeneity may lead to subtly different functional effects resulting in different epilepsy syndromes; in the ion channel gene CACNA1A different allelic mutations cause highly variant phenotypes of hemiplegic migraine, episodic ataxia, progressive ataxia and seizures (22). Alternatively, since IGE subtypes such as JME and GEFS+ can co-exist in families, a similar effect of the GABRD variation on tonic GABA inhibition may contribute to different syndromes, with the precise phenotype modulated by the other as yet unidentified variants in the suite of interacting genes causing these common polygenic epilepsies.
Demonstration that a polymorphism within an epilepsy-associated gene has an effect on ion channel properties has wider implications. Many of the ion channel subunit genes have coding single nucleotide polymorphisms (SNPs) that are potential susceptibility loci that could account for much of the underlying molecular genetic basis for common polygenic epilepsy. The detection of a rare variant with functional effect, Glu177Ala, in only one GEFS+ family and the similar frequency of the Arg220His polymorphism in the general population and in GEFS+ and IGE patients suggests that the GABRD variation is only a small component of the likely multifaceted system of rare variants (rare variant common disease model) and polymorphic susceptibility loci (common variant common disease model) likely to underlie the common IGEs. The same phenomenon is now well established for the epilepsy-associated genes initially identified from the rare large families. SCN1A, SCN1B and GABRG2 mutations account for only a small proportion of the GEFS+ cases tested (2). Similarly, CHRNA4 and CHRNB2 mutations account for only a minority of cases diagnosed with autosomal dominant nocturnal frontal lobe epilepsy (2). These observations support the likelihood of marked genetic heterogeneity among the common IGEs and reinforce the view that association studies will not have the statistical power to unravel their genetic determinants. Large-scale functional analyses of ion channel SNPs may be a more productive approach. The carrier frequency of Arg220His being ∼8% in the IGE population does, however, define the first relatively common target for pharmacogenetic intervention among hundreds of ion channels which could be involved in common polygenic diseases (23).
MATERIALS AND METHODS
Intronic HEX-labelled primers were designed for eight of nine exons of GABRD (GenBank accession no. NM_000815) (primer sequences are available on request). PCR product sizes ranged from 250 to 320 bp and the products were analysed by single-strand conformation analysis (SSCA) on a real-time gel system using a Gel-Scan 2000 DNA fragment analyser (Corbett Research). Products showing bandshifts were sequenced using the BigDye terminator cycle sequencing ready reaction kit (PE Applied Biosystems version 2.0) and the sequences were analysed on an Applied Biosystems-ABI Prism 3700 DNA analyser.
Cloning of GABAA receptor subunits and mutagenesis
We designed primers to PCR amplify the open reading frame of the GABAA receptor subunit genes GABRD, GABRB2 and GABRA1 from a fetal human brain cDNA library (Clontech) using Hot Star Taq polymerase (Qiagen). Each subunit was cloned into the vector pcDNA3.1+ (Invitrogen). The point mutations were introduced into the GABRD clone using the QuikChange site-directed mutagenesis kit (Stratagene). A column-purified oligonucleotide was used to incorporate the mutation and successful mutagenesis was confirmed by DNA sequencing.
Human embryonic kidney cells (HEK293T) were grown in Dulbecco's modified Eagle's medium (DMEM) (Invitrogen), supplemented with 10% fetal bovine serum (Invitrogen) in an incubator at 37°C with 5% CO2 and 95% air. For homozygous receptors, the cells were transfected with 2 µg each of the cDNAs encoding human α1, β2S and wild-type δ or variant δ (Arg220His, Glu177Ala and Arg220Cys) subunit using a modified calcium phosphate precipitation method (24). For heterozygous receptor transfection, 1 µg of wild-type δ subunit and 1 µg of variant δ subunit were utilized so that the total amount of cDNA was still 2 µg. For each transfection, 2 µg of pHook (Invitrogen) was used as a transfection marker. The transfected cells were separated using an immunomagnetic bead selection method (25) and were recorded 24 h later.
The number of beads on the chosen cells transfected with wild-type δ subunit was comparable to that transfected with variant δ subunit, and the whole cell currents were recorded at room temperature using the patch clamp technique with cells bathed in the external solution composed of 142 mM NaCl, 1 mM CaCl2, 6 mM MgCl2, 8 mM KCl, 10 mM glucose and 10 mM HEPES (pH 7.4, 327–330 mOsm). The resistances of the electrodes were 1.2–1.6 MΩ after being filled with an internal solution consisting of 153 mM KCl, 1 mM MgCl2, 10 mM HEPES, 5 mM EGTA and 2 mM MgATP (pH 7.3, 301–303 mOsm). Currents were recorded using an Axopatch 200A amplifier (Axon Instruments) at a holding potential of −50 mV. GABA (Sigma Chemical Co.) was applied for 4 s by gravity using an ultra-fast delivery multi-barrel tubes connected to a perfusion fast-step system (Warner Instruments). The 10–90% rise times for solution exchanges were typically 0.4–1.0 ms when the liquid junction currents were measured by switching the normal and diluted external solutions across an open electrode tip.
The Mann–Whitney test was used to compare the maximal current amplitudes among wild-type and variant δ subunit-containing GABAA receptors, and the difference was considered to be statistically significant if P<0.05.
We thank the families for their participation as well as Robert Schultz and Luyan Song for technical assistance. This study was supported by Bionomics Limited, the National Health and Medical Research Council of Australia and the National Institutes of Health (NIH).
The authors wish it to be known that, in their opinion, the first two authors should be regarded as joint First Authors.