Distinct neurodevelopmental and epileptic phenotypes associated with gain- and loss-of-function GABRB2 variants

Abstract

Background: Variants in GABRB2, encoding the β2 subunit of the γ-aminobutyric acid type A (GABA_A) receptor, can result in a diverse range of conditions, ranging from febrile seizures to severe developmental and epileptic encephalopathies. However, the mechanisms underlying the risk of developing milder vs more severe forms of disorder remain unclear. In this study, we conducted a comprehensive genotype-phenotype correlation analysis in a cohort of individuals with GABRB2 variants.

Methods: Genetic and electroclinical data of 42 individuals harbouring 26 different GABRB2 variants were collected and accompanied by electrophysiological analysis of the effects of the variants on receptor function.

Findings: Electrophysiological assessments of α1β2γ2 receptors revealed that 25/26 variants caused dysfunction to core receptor properties such as GABA sensitivity. Of these, 17 resulted in gain-of-function (GOF) while eight yielded loss-of-function traits (LOF). Genotype-phenotype correlation analysis revealed that individuals harbouring GOF variants suffered from severe developmental delay/intellectual disability (DD/ID, 74%), movement disorders such as dystonia or dyskinesia (59%), microcephaly (50%) and high risk of early mortality (26%). Conversely, LOF variants were associated with milder disease manifestations. Individuals with these variants typically exhibited fever-triggered seizures (92%), milder degrees of DD/ID (85%), and maintained ambulatory function (85%). Notably, severe movement disorders or microcephaly were not reported in individuals with loss-of-function variants.

Interpretation: The data reveals that genetic variants in GABRB2 can lead to both gain and loss-of-function, and this divergence is correlated with distinct disease manifestations. Utilising this information, we constructed a diagnostic flowchart that aids in predicting the pathogenicity of recently identified variants by considering clinical phenotypes.

Funding: This work was funded by the Australian National Health & Medical Research Council, the Novo Nordisk Foundation and The Lundbeck Foundation.

Keywords: Dystonia; Epilepsy; GABA(A) receptors; Gain-of-function; Movement disorders; Seizures.

Fig. 1

Missense GABRB2 variants associated with epilepsy or neurodevelopmental disorders. (Left) 3-D structure of a GABA_A receptor is adapted from the cryo-electron microscopy (Cryo-EM) structure of the pentameric α1β2γ2 GABA_A receptor (6x3z.pdb). (Middle) 3-D structure of β2 subunit illustrating the location of 26 presumed pathogenic variants as spheres. Variants are enriched in three important functional regions: GABA binding pocket, coupling region (Cys-loop and TM2-TM3 loop) and transmembrane helices. (Right) Membrane topology of the GABA_A receptor β2 subunit with spheres representing the relative location of 26 individual missense GABRB2 variants. Variants are colour-coded according to the functional region within the subunit: GABA binding pocket (dark blue), Cys-loop and residues in proximity (light blue), TM2-TM3 loop (yellow), and transmembrane helices TM1-TM3 (pink). Each variant is indicated by line pointing its position and amino acid substitution.

Fig. 2

Concatenated construct and representative traces (a) Top view of cryo-EM structure of the α1β2γ2 GABA_A receptor (6x3z) with bound GABA (left). Pentameric concatenated receptor design for the γ2-β2^star-α1-β2-α1 cDNA construct utilised to introduce β2 subunit missense mutations (star) in the second construct position only (middle). The five subunits are linked with four linker sequences (L) based on Alanine-Glycine-Serine repeats. The resulting pentameric receptor is portrayed with arrows indicating linkers sequences and the counterclockwise assembly orientation (right). (b) Representative electrophysiological traces depicting GABA concentration-response relationships for receptors containing the β2 wildtype (black), β2^Y181F (red) or β2^L283I (blue) subunit. Bars above the traces depict the 25-s application time with GABA concentrations indicated for each trace in μM. (c) GABA concentration-response relationships are depicted as mean ± SD for n = 9–11 biological replicates and the Hill equation was fitted to the data by non-linear regression. The GABA sensitivity is observed from the fitted EC₅₀ value and arrows indicate whether mutations cause GOF or LOF in GABA sensitivity.

Fig. 3

GABA sensitivity changes and maximal GABA-evoked current amplitudes for α1β2γ2 receptors with β2 subunit mutations. (Left) Changes in GABA sensitivity between wildtype (WT) and receptors with β2 subunit mutations are presented as mean ΔLogEC₅₀ ± SD for n = 222 (WT) or n = 9–22 (β2 mutations) experiments with individual datapoints shown (light grey). Blue indicates mutations that significantly increase GABA sensitivity (GOF), red indicate mutations that significantly reduce GABA sensitivity (LOF) and grey indicate mutations with no significant change. Significance was determined using one-way ANOVA with corrected Dunnetts’ multiple comparisons post hoc test (∗∗∗∗, P < 0.0001; full detail in the Supplementary table). A change from darker to lighter vertical blue shading at ΔLogEC₅₀ = 0.7 guides the separation between High-shift (ΔLogEC₅₀ > 0.7) and Low-shift (ΔLogEC₅₀ < 0.7) GOF variants. (Right) Normalised maximal GABA-evoked current amplitudes are presented as median with IQR for n = 390 (WT) or n = 17–42 (β2 mutations) experiments. Red indicates mutations causing LOF whereas grey indicates mutations with no significant change. Significance was determined using Mann–Whitney U test (∗∗∗∗, P < 0.0001; full detail in the Supplementary table). Stars indicate a significantly reduced current amplitude that additionally is below the set threshold 0.5 level (red shading, see methods).

Fig. 4

Clinical phenotypes of individuals with LOF and GOF GABRB2 variants. Selected clinical parameters were assessed for their association with the molecular phenotype of the respective variants. (a) Comparison of LOF (n = 13) vs GOF (n = 27) sub-cohorts. (b) Comparison of High-shift GOF (n = 9) vs Low-shift GOF (n = 18) groups. Odds ratio (OR) analyses of phenotype–genotype associations are presented with the centre circle denoting the OR and 95% confidence interval. Blue indicates significant enrichment in individuals with GOF (a) or High-shift GOF (b), red significant enrichment in individuals with LOF and grey no significant difference between compared individuals. Open circles without confidence intervals indicate data where one category contains 0 or 100% of individuals and the OR and CI cannot be determined. Statistical analyses were performed using two-sided Fisher’s exact test resulting in the indicated P values. Statistics for survival analyses, seizure onset and seizure risk were performed using Mantel–Cox, Mann–Whitney and Mantel–Cox tests, respectively, with the obtained P values indicated. Number of individuals at risk in survival and seizure risk analyses are indicated at five timepoints, “m” refers to months in age of seizure onset and “movement disorder∗” refers to dystonia, dyskinesia, hyperkinesia, or chorea.

Fig. 5

GABA sensitivity changes, maximal current amplitudes, and desensitisation properties for β2^L255V-containing receptors. (a) Two pentameric concatenated constructs were designed with β2^L255V mutations (star) to reflect a heterozygous patient condition where receptors can have a single (β2^L255V, β2^wt) or two (β2^L255V, β2^L255V) variant subunits. (b) Changes in GABA sensitivity for the indicated receptor types are presented as mean ΔLogEC₅₀ ± SD for n = 15–17 experiments and statistical analysis was performed using one-way ANOVA with corrected Dunnetts’ post-hoc test values depicted in the panel (∗∗∗∗, P < 0.0001). (c) Normalised maximal GABA-evoked current amplitudes are depicted as median with IQR for n = 26–33 experiments for the indicated receptor types. Statistical analysis was determined using Mann–Whitney U test and no significant differences were observed (Supplementary table). (d) Representative traces of responses to 150-s applications of 3 mM GABA at β2^wt, β2^wt (black), β2^L255, β2^wt (blue) and β2^L255V, β2^L255V (dark blue) receptors for illustration of current decay rates and steady-state current amplitudes. Traces were fitted to an exponential decay function to assess the initial current decay rate constant (k) and the estimated open probability at equilibrium (Est. P_O (_{SS, max})). (e, f) Current decay rates (e) and steady-state current amplitudes at equilibrium (f) are presented for the indicated receptor types as mean ± SD for n = 10–13 experiments. Statistical analysis was performed using a non-parametric one-way ANOVA Kruskal–Wallis test with Dunn’s post hoc test values depicted in the panels (∗∗∗∗, P < 0.0001).

Fig. 6

Clinical indicators for GOF and LOF phenotypes. In cases where functional analysis is not available, clinical biomarkers may be used to predict whether a newly identified variant causes GOF or LOF. Age of seizure onset is a strong indicator of GOF disease in cases of early infantile onset and is associated with variants causing High-shift GOF. Severe movement disorders and microcephaly are strongly associated with GOF disease while epilepsy syndromes in the GEFS+ spectrum with fever sensitivity are strongly associated with LOF disease. As the child ages, intellectual disability, independent mobility, and language development support the early indicators. For cases that are not clearly defined by these indicators, functional analysis is needed to determine the effects of the variant.

Supplementary Fig. S1

Supplementary Fig. S1. 2-D topology of a GABA_A receptor β2 subunit with mapping of GOF and LOF variants.