Mutations in GABAA receptor subunits associated with genetic epilepsies (2024)

Abstract

Mutations in inhibitory GABA_A receptor subunit genes (GABRA1, GABRB3, GABRG2 and GABRD) have been associated with genetic epilepsy syndromes including childhood absence epilepsy (CAE), juvenile myoclonic epilepsy (JME), pure febrile seizures (FS), generalized epilepsy with febrile seizures plus (GEFS+), and Dravet syndrome (DS)/severe myoclonic epilepsy in infancy (SMEI). These mutations are found in both translated and untranslated gene regions and have been shown to affect the GABA_A receptors by altering receptor function and/or by impairing receptor biogenesis by multiple mechanisms including reducing subunit mRNA transcription or stability, impairing subunit folding, stability, or oligomerization and by inhibiting receptor trafficking.

Robert Macdonald (centre), Jing-Qiong Kang (right) and Martin Gallagher (left) work in the Department of Neurology at Vanderbilt Medical Center and collaborate on investigation of the pathophysiological mechanisms of mutations in GABA_A receptor subunit genes that are associated with idiopathic generalized epilepsies. Their backgrounds are in molecular and cellular biophysics, mechanisms of cell genesis and degeneration and protein chemistry, respectively. Using electrophysiology, flow cytometry, confocal microscopy and protein chemistry, they have determined the molecular basis for loss of GABAergic inhibition for multiple missense and nonsense IGE mutations.

Epilepsy is diagnosed in individuals who experience two or more unprovoked seizures. It is one of the most common neurological disorders affecting 0.5–1% of the worldwide population (Hauser, 1994). Many epilepsy syndromes are believed to have a genetic basis (Hirose et al. 2005; Berkovic et al. 2006) and vary in severity from the relatively benign childhood absence epilepsy (CAE; MIM no. 600131) to the severe epilepsy syndrome Dravet syndrome (DS; MIM no. 607208). The term generalized epilepsy with febrile seizures plus (GEFS+; MIM no. 604233) has been applied to families in which several members experience febrile seizures, afebrile seizures or other generalized epilepsy syndromes. Although complex polygenic inheritance is likely to be associated with most genetic epilepsy syndromes, monogenic mutations of transmembrane ion channels have been found to cause epilepsy in several large pedigrees and in sporadic cases with de novo mutations (Mulley et al. 2003; Hirose et al. 2005; Wallace et al. 2001; Maljevic et al. 2006). Mutations of ion channels that either directly or indirectly enhance excitatory neurotransmission or reduce inhibitory neurotransmission would cause brain hyperexcitability and thereby predispose patients to seizures. In this review we will focus on mutations in inhibitory GABA_A receptor subunit genes that are associated with genetic epilepsy syndromes.

Epilepsy mutations in GABA_A receptor subunit genes

GABA_A receptors are the primary mediators of fast inhibitory synaptic transmission in the central nervous system. They are members of the cys-loop family of ligand-gated ion channels that also includes nicotinic cholinergic, serotonin 5-HT3, and glycine receptors and are formed by pentameric assemblies of different subunit subtypes (α1–α6, β1–β3, γ1–γ3, δ, ɛ, π, θ, and ρ1–ρ3) to form chloride ion channels (). The majority of GABA_A receptors contain two α subunits, two β subunits, and a γ or δ subunit (Baumann et al. 2002). GABA_A receptors mediate both phasic inhibitory synaptic transmission and tonic perisynaptic or extrasynaptic inhibition. Moreover, the mechanism of action of several antiepileptic drugs including benzodiazepines and barbiturates is to enhance GABA_A receptor currents (). Therefore, it is not surprising that several different genetic epilepsy syndromes have been associated with mutations and variants in several GABA_A receptor subunit genes including GABRA1, GABRB3, GABRD and GABRG2 (see Fig. 1, Table 1) (Macdonald et al. 2006). Investigators have designated the position of these mutant amino acids in GABRA1, GABRB3 and GABRD in the immature peptide that includes the signal sequence (i.e. GABRA1(A322D)), but mutations in GABRG2 have been reported in the mature peptide (i.e. GABRG2(R43Q) (see ). For consistency, the position of GABRG2 mutations will be designated in the immature peptide (i.e. GABRG2(R82Q)) in this article. To study the pathogenesis of these mutations, in vitro expression systems often have been used. The term ‘wild-type receptors’ has been used when only wild-type subunits were expressed in an equal 1:1:1 α_wt:β_wt:γ_wt or δ_wt subunit ratio, the term ‘hom*ozygous mutant receptors’ has been used when only mutant subunits were coexpressed with their wild-type assembly partners in an equal 1:1:1 subunit ratio and the term ‘heterozygous receptors’ has been used when mixed wild-type and mutant forms of the same subunits have been coexpressed in a 0.5:0.5 ratio with their wild-type assembly partners. For example for a mutant γ_mut subunit coexpressed with wild-type α and β subunits, heterozygous receptor expression would be at a 1:1:0.5:0.5 α:β:γ_wt:γ_mut subunit ratio. Finally, the term ‘hemizygous receptor expression’ has been used when one-half of the typical amount of wild-type subunit cDNA was expressed with full amounts of wild-type assembly partners. For example hemizygous receptor expression would be at a 1:1:0.5 α:β:γ_wt ratio; the other half γ_wt would be replaced by empty plasmid to ensure an equal total cDNA amount.

Open in a separate window

Figure 1

GABA_A receptor subunit mutations associated with genetic epilepsy syndromes

GABA_A receptor subunits are translated as a precursor protein whose signal sequence (green) is removed leaving a mature protein consisting of a large extracellular domain at the N-terminus, four transmembrane domains (M1–M4) and a large cytoplasmic domain. In this figure, we depicted GABA_A receptor α1 (blue), β3 (pink), γ2 (red) and δ (yellow) subunit mutations associated with genetic epilepsy syndromes at their appropriate protein domain within the subunit.

Mutations and variants in GABA_A receptor subunit genes associated with genetic epilepsy syndromes can be divided broadly into four classes: (1) missense mutations in coding sequences, (2) nonsense mutations in coding sequences, (3) frameshift mutations in coding sequences, and (4) mutations in non-coding sequences including intronic or 5′ upstream sequence. Here, we will review the evidence for how these different classes of mutations within different domains of affected GABA_A receptor subunits result in disinhibition, and thereby, predispose an individual to having seizures.

GABA_A receptor subunit gene missense mutations in coding sequences

Missense mutations alter codon nucleotide sequences, resulting in incorporation of a different amino acid into the subunit. If the altered amino acid is identified only in patients with the disease, it is classified as a mutation, but if the alteration is also identified in the general population, it is termed a susceptibility variant. All of the inherited GABA_A receptor subunit gene epilepsy mutations have autosomal dominant (AD) inheritance or are sporadic, and thus patients are heterozygous for the mutations and have both wild-type and mutant alleles.

The GABRG2 mutation, R82Q, is located in the distal N-terminus and is associated with an AD form of CAE and FS (Wallace et al. 2001; see Fig. 1, Table 1). It was subsequently demonstrated that the γ2(R82Q) subunit mutation alone accounted for the FS phenotype; an interaction of the γ2 subunit gene with another gene or genes is required for the CAE phenotype in this family (Marini et al. 2003). Alignment of γ2 subunit and acetylcholine binding protein sequences revealed that R82 is positioned at the γ2–β2 subunit–subunit interface, and it was demonstrated that the R82Q mutation impaired γ2 and β2 subunit oligomerization (Hales et al. 2005). Impaired oligomerization is likely to explain the effect of this mutation to reduce surface α1β2γ2 receptor levels (; ; Eugène et al. 2007; Frugier et al. 2007), cause ER retention of unassembled γ2(R82Q) subunits (; Frugier et al. 2007) and reduce GABA_A receptor currents (Bianchi et al. 2002; ). Similarly, the R82Q mutation also caused intracellular retention and reduced surface expression of GABA_A receptors in cortical pyramidal neurons (Tan et al. 2007), reduced miniature inhibitory postsynaptic currents (IPSCs) in layer II/III cortical neurons and caused electrographic and behavioural seizures in genetically modified mice. In cultured hippocampal neurons, the endogenous expression of α5 subunits was reduced when coexpressed with γ2(R82Q) subunits indicating that the γ2(R82Q) subunit conferred a dominant negative effect (Eugène et al. 2007). In addition, it is possible that a deficit in γ2 subunits, causes a compensatory increase in other subunits such as δ or β subunits. Since αβδ and αβ receptors are extrasynaptic or perisynaptic, this compensatory increase may result in a relative increase in tonic currents. Recently, it has been reported that extrasynaptic GABAergic ‘tonic’ inhibition was increased in thalamocortical neurons from both genetic and pharmacological models of absence epilepsy (Cope et al. 2009). Further study is needed to elucidate how the alteration of tonic inhibition contributes to epileptogenesis.

The GABRG2 mutation, K328M, is located in the short extracellular loop between transmembrane domains M2 and M3 and is associated with an AD generalized epilepsy similar to GEFS+ (Baulac et al. 2001; see Fig. 1 and Table 1). Brief GABA-evoked currents recorded from hom*ozygous α1β3γ2L(K328M) receptors had unchanged current amplitudes but had accelerated deactivation (; Hales et al. 2006). In transfected hippocampal neurons, the K328M mutation also accelerated deactivation of IPSCs and thus reduced their duration (Eugène et al. 2007). Single channel currents from hom*ozygous α1β3γ2(K328M) receptors had reduced mean open times, consistent with accelerated macroscopic current deactivation (Bianchi et al. 2002). Therefore, the γ2L(K328M) subunit mutation would reduce IPSC duration by accelerating its deactivation due to impaired stability of the channel open state.

The GABRG2 mutation, R177G, is located in the N-terminus and has been associated with FS (Audenaert et al. 2006; see Fig. 1 and Table 1). The R177 residue is conserved among γ2 subunits across species, and basic residues are conserved among other γ subunits. In other cys-loop receptors, polar and charged amino acid residues occur at this position. Mutant α1β3γ2L(R177G) receptors had altered current kinetics and reduced benzodiazepine sensitivity (Audenaert et al. 2006), but the underlying molecular mechanisms are unclear. Given that impairment of receptor trafficking and glycosylation arrest are prominent defects caused by mutations of GABA_A receptor subunits, future study focusing on the mutant protein maturation and receptor trafficking is needed to characterize the molecular defect of this mutation in epilepsy.

The GABRD susceptibility variants, E177A and R220H, are located in the N terminus of the δ subunit and are associated with an AD generalized epilepsy similar to GEFS+ (Dibbens et al. 2004; see Fig. 1 and Table 1). The GABRD(E177A) variant is adjacent to one of the two cysteines that form a disulfide bond, the signature feature of cys-loop receptors, and the GABRD(R220H) variant is located between the cys-loop and the beginning of the first transmembrane domain (M1). The macroscopic current amplitudes of heterozygous and hom*ozygous α1β2δ receptors containing either δ(E177A) or δ(R220H) subunits were significantly reduced due primarily to reduced single channel mean channel open time (Feng et al. 2006). Thus, both variants reduced GABA_A receptor current by impairing channel gating.

The GABRA1 mutation, A322D, introduces a negatively charged aspartate into the middle of the M3 transmembrane helix and is associated with an AD juvenile myoclonic epilepsy (JME) (Cossette et al. 2002; see Fig. 1, Table 1, MIM no. 254770). This non-conserved mutation was shown to destabilize insertion of the M3 domain into the lipid bilayer and thus impair α1 subunit folding (Gallagher et al. 2007). When co-expressed with β2 and γ2 subunits, mutant α1(A322D) subunits reduced both total and surface α1 subunit levels and had an intermediate effect on heterozygous subunit expression. Loss of the misfolded mutant subunit was due primarily to endoplasmic reticulum (ER) associated degradation (ERAD) (Gallagher et al. 2005) and to some lysosomal degradation (Bradley et al. 2008). Consistent with the impaired folding and assembly of the mutant α1(A322D) subunits, peak GABA-evoked currents were significantly reduced with both heterozygous and hom*ozygous α1(A322D) subunit expression (Cossette et al. 2002; Fisher, 2004; Gallagher et al. 2004). It is unknown if the presence of the non-degraded, misfolded α1(A322D) subunit contributes to the JME phenotype by causing a dominant negative effect.

Three GABRB3 mutations, P11S, S15F and G32R, are associated with CAE (Tanaka et al. 2008), and the P11S mutation has also been associated with multiple autism pedigrees with some patients who also had epilepsy (Delahanty et al. 2010). P11S and S15F are in exon 1a and are located in the β3 subunit signal peptide, while G32R is in exon 2 and is located in the mature β3 subunit peptide near the N terminus. When co-expressed with α1 and γ2 subunits, all three mutant β3 subunit-containing receptors had reduced GABA-evoked whole cell peak currents (Tanaka et al. 2008; Delahanty et al. 2010). When HA-tagged β3, α1 and γ2 subunits were coexpressed, α1β3(P11S)^HAγ2 receptors had reduced β3(P11S)^HA subunit protein expression on the cell surface (Delahanty et al. 2010). Interestingly, in vitro studies suggested that these β3 subunit mutations may reduce GABA_A receptor expression and whole cell current amplitudes by altering N-linked glycosylation of the β3 subunit (Tanaka et al. 2008; Delahanty et al. 2010). Reduced β3 subunit-containing GABA_A receptor surface expression would be consistent with an epilepsy phenotype, and EEGs obtained from heterozygous and hom*ozygous GABRB3 knock-out mice demonstrate epileptiform complexes and seizures that are responsive to antiepileptic drugs (Homanics et al. 1997). However, it is quite interesting that mutations in β3 subunits, but not β1 and β2 subunits, have been associated with epilepsy. This may suggest that there is insufficient functional redundancy from other β subunits to compensate for the loss of β3 subunit in critical brain regions such as thalamus, but that there is sufficient functional overlapping for other β subunits. This is not surprising given the abundance of β3 subunit in embryonic and neonatal brain and its critical role in brain development.

GABA_A receptor subunit gene nonsense mutations in coding sequences

Nonsense mutations alter the nucleotide sequence in codons that result in introduction of a premature translation-termination codon (PTC). PTCs in the last exon of a multi-exon gene or less than 50–55 nucleotides upstream of the last exon–exon junction result in production of a truncated protein. In contrast, PTCs not in the last exon of a multi-exon gene or are more than 50–55 nucleotides upstream of the last exon–exon junction produce mRNA degradation through activation of nonsense-mediated decay (NMD), a cellular mRNA quality-control system that activates degradation of mutant mRNA to substantially reduce production of truncated proteins (Maquat, 2004).

The GABRG2 mutation, Q390X, is located in the intracellular loop between transmembrane domains M3 and M4 and was identified in a family with GEFS+ and DS (Harkin et al. 2002; see Fig. 1, Table 1). The PTC is located in the last (9th) exon, and therefore, would not be expected to activate NMD. Thus translation resulted in production of a truncated protein that lacked its C-terminal 78 amino acids and was retained in the ER (Kang et al. 2009b). Consistent with this, no GABA-evoked currents were recorded from hom*ozygous mutant receptors (Harkin et al. 2002; Kang et al. 2009b). The γ2(Q390X) subunit also caused a dominant negative effect on wild-type receptors. Currents recorded from heterozygous receptors were reduced relative to hemizygous control currents, and γ2S and γ2S(Q390X) subunits and partnering α1 and β2 subunits were all reduced more than the hemizygous condition with only one wild-type allele, suggesting that the mutation produced a loss of function of the mutant allele and a dominant negative effect of the mutant γ2S(Q390X) subunit on wild-type receptor channels.

Two GABRG2 nonsense mutations, Q40X and Q429X, have been associated with DS and GEFS+, respectively (Hirose, 2006; Sun et al. 2008). The Q40X mutation is likely to trigger NMD, although this has not yet been demonstrated. In contrast, the Q429X mutation, which generates a PTC in the last GABRG2 exon, would not be predicted to activate NMD and, therefore, would be expected to produce a truncated protein with loss of the C-terminal 39 amino acids.

GABA_A receptor subunit gene frameshift mutation in coding sequences

Frameshift mutations occur because deletion or insertion of one or two nucleotides causes a change in downstream codons, with or without a change in the frameshifted codon. In addition to altering downstream amino acid sequence, frameshifts often introduce PTCs. The GABRA1 frameshift mutation, 975delC, S326fs328X, causes a frameshift in GABRA1 that produces a PTC in the 8th exon that is 84 base pairs upstream of intron 8 and is associated with CAE (Maljevic et al. 2006; see Fig. 1 and Table 1). Thus, this PTC activated NMD and reduced mutant α1 subunit mRNA (Kang et al. 2009a). However, NMD was incomplete, and some truncated α1 subunit protein was translated. The mutant α1(S326fs328X) subunit protein was degraded by ERAD. These results suggested that the GABRA1 mutation, S326fs328X, resulted in functional haploinsufficiency by reducing both mutant mRNA and subunit protein.

GABA_A receptor subunit gene mutations in intronic or 5′ upstream non-coding sequences

The GABRG2 mutation, IVS6 + 2T→G, is located in the splice-donor site in intron 6 and was identified in a family with CAE and FS (Kananura et al. 2002; see Fig. 1 and Table 1). The effect of this mutation on GABA_A receptor function is unknown but was predicted to lead to a non-functional protein through exon skipping, which would result in a PTC at the 5th and 7th exon junction site. Thus, it was suggested that the PTC would trigger NMD. Therefore, the underlying mechanism for this splice donor site mutation also may be haploinsufficiency.

In the GABRB3 gene, one of four haplotypes (haplotype 2) in the region from the exon 1a promoter to the beginning of intron 3, was found to have a significant association with CAE. In vitro reporter gene assays using exon 1a promoter constructs indicated that the haplotype 2 promoter caused lower transcriptional activity than the haplotype 1 promoter (Urak et al. 2006), suggesting that a thymine to cytosine substitution in the haplotype 2 promoter impaired binding of the neuron-specific transcriptional activator N-Oct-3 leading to decreased gene transcription of GABRB3 and presumably to a decrease in the β3 subunit. Since β subunits are essential for functional pentameric GABA_A receptor assembly and because β3 subunits are the most abundant β subunit in the developing brain, association of reduced transcriptional activity by the haplotype 2 promoter and CAE suggests that haplotype 2 may contribute to the pathogenesis of CAE.

Model systems for evaluating physiological changes associated with GABA_A receptor subunit gene mutations

The physiological consequences of the mutant GABA_A receptor subunits were first evaluated by overexpressing them in heterologous cells. Although this strategy was a necessary first step in the analysis of these mutations, GABA_A receptor expression and function in heterologous cells and in vivo could differ. Because synaptic GABA release and reuptake occur extremely rapidly, postsynaptic GABA_A receptor experience only very brief, non-equilibrium exposures to agonist (Mozrzymas, 2008). Although many of the studies reviewed here employed techniques to rapidly apply agonist to heterologous cells (Hinkle et al. 2003), the physiological changes identified in vitro could differ from those that would occur in vivo. In addition to not fully reproducing synaptic agonist applications, heterologous cells do not express neurone-specific GABA_A receptor-associating proteins that affect rates of GABA_A receptor biosynthesis, trafficking and cell surface stability (Jacob et al. 2008). Expressing mutant GABA_A receptor subunits in cultured neurons addresses both of these concerns. In addition to expressing neuronal specific proteins, cultured neurons also form functional GABAergic synapses which obviate the need for exogenous applications of GABA. The expression of the mutant γ2(R82Q), γ2(K328M) and γ2(Q390X) subunits in cultured neurons (Eugène et al. 2007; Frugier et al. 2007; Kang et al. 2009b) has confirmed many of the initial observations obtained in heterologous cells. For example, rapid agonist perfusion techniques applied to transfected HEK293T cells revealed that α1β2γ2(K328M) receptors deactivated significantly faster than wild-type receptors (Bianchi et al. 2002). Similarly, miniature IPSCs in hippocampal neurons transfected with γ2(K328M) subunits had significantly shorter deactivation times than untransfected neurons or those transfected with wild-type γ2 subunit (Eugène et al. 2007).

In other cases, observations made in transfected neurons confirmed those obtained in HEK293T cells but also revealed new findings that could not have been predicted from heterologous expression alone. For example, studies in HEK293T cells revealed that the γ2(R82Q) subunit remained trapped in the endoplasmic reticulum (; ), a finding confirmed in transfected hippocampal neurons (Eugène et al. 2007; Frugier et al. 2007). However, observations made in hippocampal neurons revealed that expression of the γ2(R82Q) subunit also reduced non-synaptic ‘tonic’ GABA currents and reduced surface expression of α5 subunits (Eugène et al. 2007), two findings that would not have been observed in heterologous cells that do not express endogenous GABA_A receptor subunits or neuron-specific proteins such as radixin, which associates with the α5 subunit (Loebrich et al. 2006). Additional studies are being performed to determine the effects of other mutant GABA_A receptor subunits when expressed in cultured neurons.

Although overexpression of mutant GABA_A receptor subunits in cultured neurons is beginning to reveal new insights concerning their effects on GABAergic physiology, it is possible that these mutations will cause different effects when they are expressed at endogenous levels using the endogenous, regulated promoters. Furthermore, and perhaps more importantly, these mutations may cause different effects when expressed in different brain regions and thus shape the neuronal network differently from different locations. Although costly and time consuming, these questions must be answered by studying genetically modified animals as well as human patients who possess these mutations.

To date, only the GABRG2(R82Q) mutation has been expressed in a mouse model (Tan et al. 2007; Chiu et al. 2008). Consistent with the findings in heterologous cells and in transfected neurons, the mutation substantially reduced γ2(R82Q) subunit surface expression in mouse brain (Tan et al. 2007). However, as an example of an observation that could not be made in cultured cells, miniature IPSC amplitudes were reduced in cortical neurons, but not in thalamic relay or reticular nuclei neurons (Tan et al. 2007). As another example of a consequence of the GABRG2(R82Q) mutation that could only be detected in an intact organism, the epilepsy phenotype required γ2(R82Q) expression at critical times in development (Chiu et al. 2008).

The ultimate goal of studying these GABA_A receptor subunit gene mutations is to determine their effects on human brain physiology. Because patients with generalized epilepsy do not typically undergo resective surgical procedures and because they typically live for decades following their diagnosis, we lack surgical and postmortem tissue from epilepsy patients with GABA_A receptor subunit gene mutations. However, non-invasive neuroimaging and neurophysiology procedures have been performed on epilepsy patients with the GABRG2(R82Q) mutation (Fedi et al. 2006, 2008). These studies demonstrated that patients, like the mice, with the GABRG2(R82Q) mutation possess hyperexcitable cortices. In addition, consistent with the results from cultured cells and genetically modified mice, patients with the GABRG2(R82Q) mutation express reduced GABA_A receptors as detected by [¹¹C]flumazenil binding. However, as an example of a result found in patients, but not yet confirmed in the mice, the greatest reduction in [¹¹C]flumazenil binding occurred in the cingulate and insular cortices. The creation of genetically modified mice with other epilepsy mutations as well as more in-depth study of human patients will better enable us to determine how these GABA_A receptor subunit gene mutations alter physiology and cause the epilepsy syndromes.

Conclusions

While many unanswered questions remain, substantial advances have been made in unravelling the molecular pathogenesis of GABA_A receptor subunit gene mutations associated with genetic epilepsy syndromes. Mutations in four GABA_A receptor subunit genes in different subunit domains have been demonstrated to cause impaired channel gating and/or reduced mRNA stability, abnormal subunit folding and aberrant glycosylation that lead to impaired receptor assembly and trafficking (Table 1). These mutations result in functional haploinsufficiency of the affected gene itself and/or have a dominant negative effect on wild-type partnering subunits. However, given the phenotypical heterogeneity within pedigrees and the multiple pedigrees with the same epilepsy phenotype that harbour different mutations in GABA_A receptor subunit genes, it is likely that the genetic background or modifying genes also play a role in the epileptogenesis of genetic epilepsy syndromes. For example, the same nonsense mutation may evoke different degrees of NMD efficiency among different individuals, and the remaining misrouted mutant protein may cause different level of cellular stress. Thus, future work with knock-in mice may further explain the impact of a given GABA_A receptor subunit gene mutation in the brain with complicated neuronal networks.

Glossary

References