Clinical and functional characterization of the GABRB3 p.Met80Val variant in early-onset epilepsy with long-term follow-up

Abstract

Background

GABRB3 encodes the β3 subunit of the GABA_A receptor, which is a crucial component in inhibitory neurotransmission within the central nervous system. GABRB3 variants are associated with developmental and epileptic encephalopathy 43. Noteworthy, GABRB3 variants can result in both gain-of-function and loss-of-function effects. However, their precise functional and clinical implications remain unknown.

Methods

Whole-exome sequencing, validation of the identified GABRB3 variants using Sanger sequencing, and structural modeling were done to assess the potential impact of these variants on receptor function. Functional analyses included quantification of GABRB3 protein expression levels, subcellular localization using fluorescence microscopy, and electrophysiological recordings of α1β3γ2 and α1β3(M80V)γ2 receptor complexes to evaluate channel properties.

Results

A heterozygous de novo GABRB3 variant (NM_000814.6: c.238 A > G, p.Met80Val) was identified in a 16-year-old female who developed absence seizures at one year and exhibited persistent EEG abnormalities over the subsequent decade. She exhibited mild intellectual disability, poor academic performance, and limited language skills but maintained school attendance and social engagement. Structural modeling suggested that the p.Met80Val variant compromises the structural integrity of the protein. Functional assays revealed a 2.6-fold increase in GABRB3 protein expression and enhanced fluorescence intensity, with most of the protein localized in the cytoplasm. Electrophysiological recordings demonstrated significantly increased current amplitude, heightened GABA sensitivity, and reduced zinc sensitivity. These findings indicated that the p.Met80Val variant altered the receptor conformation or its zinc-binding site, weakening zinc-mediated inhibition.

Conclusion

This study reports the ninth case of a recurrent GABRB3 p.Met80 variant and highlights its potential as a hotspot missense variant. The findings underscore its pathogenicity after a long follow-up period of more than ten years supported by continuous EEG monitoring. These findings enhance our understanding of the functional changes among GABRB3 variants and their role in the pathogenesis of epilepsy.

Background

Epilepsy is a prevalent neurological disorder affecting 1% of the global population [1]. It is characterized by a broad spectrum of seizure types and etiologies, classified into symptomatic and genetic forms. Symptomatic neonatal epilepsy typically results from acute perinatal injuries, such as intraventricular hemorrhage, hypoxic-ischemic encephalopathy, central nervous system (CNS) infections, and transient metabolic disturbances [2]. In contrast, genetic epilepsy is primarily attributed to variants in genes encoding critical CNS components, particularly ion channels such as sodium and potassium channels and synaptic proteins. Pathogenic variants in ion channels such as SCN1A and synaptic proteins such as SYNGAP1 are well-established in epilepsy. Nonetheless, recent evidence suggests dysfunction of γ-aminobutyric acid type A receptors (GABAARs), the brain’s primary inhibitory anion channels, is a critical factor in seizure susceptibility [3–5]. GABA_ARs are heteropentameric complexes assembled from various subunits combinations, including α (1–6), β (1–3), γ (1–3), δ, ε, π, θ, and ρ (1–3). The most common isoform of GABA_ARs in the mature brain comprises two α1, two β2/3, and one γ2 subunit [6, 7]. These receptors mediate inhibitory neurotransmission by facilitating chloride ion influx upon GABA binding, thereby maintaining the excitatory-inhibitory (E-I) balance [8]. Disruption of GABA_AR assembly, trafficking, or function resulting from pathogenic variants in subunit-encoding genes disturbs this balance, leading to neuronal hyperexcitability and epilepsy [9, 10]. Notably, both the rare and common variants in GABA_AR genes contribute to epilepsy risk in complex ways. Polygenic GABAergic dysregulation is also implicated in various neuropsychiatric disorders [11]. Variants in several GABA_AR subunit genes, including GABRA1–3, GABRA5, GABRB1–3, GABRG2, and GABRD, have been associated with Mendelian disorders characterized by neurodevelopmental deficits and epilepsy (OMIM: PS308350).

The GABRB3 gene, located on chromosome 15q12, encodes the β3 subunit. The β3 subunit is highly expressed in neonatal brains and is essential for proper GABA_AR assembly [9]. GABRB3 is critical for integrating pyramidal neurons in the somatosensory cortex. Its ablation leads to a developmental reduction in GABAergic synapses, enhanced local network synchrony, and a persistent increase in connectivity among contralateral pyramidal neurons. It also leads to an elevated cortical response to tactile stimulation during the neonatal period [12]. Pathogenic GABRB3 variants cause developmental and epileptic encephalopathy-43 (DEE43, OMIM#617113). DEE43 is an autosomal dominant disorder characterized by EOE, such as febrile, generalized tonic-clonic, and absence seizures, as well as developmental delay/intellectual disability and behavioral abnormalities [13]. These variants yield varying clinical outcomes, ranging from mild febrile seizures to severe epileptic encephalopathies, suggesting residue-specific functional effects. Noteworthy, the mechanistic and phenotypic correlations remain unknown despite the knowledge of both gain-of-function (GOF) and loss-of-function (LOF) variants [14]. The p.Met80 residue in GABRB3 has emerged as a mutational hotspot, with eight distinct missense variants reported in patients [15–21]. This study integrated clinical, structural, and functional analyses to elucidate the genotype-phenotype correlations and implications for DEE43 pathogenesis of a p.Met80Val variant identified in a patient with EOE and a prolonged history of EEG abnormalities.

Materials and methods

Patient and clinical evaluation

Clinical data of the proband, including age, sex, seizure onset and progression, family history, neurological examination outcomes (MRI and EEG), and antiseizure medication usage, was obtained through oral interviews and clinical assessments. This study was conducted following the statutes of the Declaration of Helsinki and was approved by the Ethics Committee of the First Affiliated Hospital of Anhui Medical University (Approval number: PJ2024-12-62). Written informed consent for participation and publication was obtained from the patient’s parents.

Whole exome sequencing

Genomic DNA was extracted from peripheral blood samples of the patient and her parents using a commercial DNA extraction kit (Biotech Corporation, Beijing, China). The xGen Exome Research Panel v1.0 (IDT, Iowa, USA) was used for exome capture, followed by sequencing on an Illumina NovaSeq 6000 system (Illumina, San Diego, CA, USA), which achieved a target sequence coverage of at least 99%. The sequence obtained was aligned to the human reference genome (GRCh37/hg19) using the Burrows-Wheeler Aligner (BWA). Variant calling, base quality score recalibration, and annotation were done using GATK, Picard, and ANNOVAR, respectively. Common variants (allele frequency > 1%) were excluded by comparing them against public databases, including dbSNP, ExAC, 1000 Genomes, ESP, and gnomAD. Variants were prioritized based on their predicted effects. They included nonsense, frameshift, splice site, in-frame, and missense variants. The variants were evaluated for pathogenicity using multiple in silico tools, including SIFT, MutationTaster, PolyPhen-2, M-CAP, PROVEAN, CADD, and REVEL. Disease associations were investigated using OMIM, HGMD, and ClinVar databases. Variant nomenclature followed HGVS guidelines (http://www.hgvs.org/), while classification was performed following the American College of Medical Genetics and Genomics (ACMG) guidelines [22].

Sanger sequencing

Suspected pathogenic variants were validated by Sanger sequencing. The p.Met80Val variant was subjected to PCR amplification of the gene using primer pair 5’-CGGCCAGGACTCCCCCGCAGGAA-3’ (forward) and 5’-CCTCCCCTCCCCCTCCGCCCTCCTC-3’ (reverse), yielding a 948-bp amplicon. The PCR products were sequenced using an ABI 3730XL DNA sequencer (Applied Biosystems, Waltham, MA, USA).

Protein modeling

Sequences of human GABRA1–6, GABRB1–3, GABRG1–3, GABRD, GABRE, GABRP, GABRQ, and GABRR1–3, as well as β3 subunit sequences from rat and mouse, were aligned using MEGA software (https://www.megasoftware.net/) to evaluate the amino acid conservation at the affected residue. The three-dimensional structure of the full-length human α1β3γ2 GABA_A receptor complex (resolved by cryo-EM in complex with diazepam, GABA, and megabody Mb38) was obtained from the Protein Data Bank (PDB: 6HUP) to assess the structural impact of the p.Met80Val variant [23]. Structural visualization and analysis were performed using UCSF Chimera (https://www.cgl.ucsf.edu/chimera/).

Cell culture and plasmid transfection

Plasmids encoding wild-type (WT) and mutant (NM_000814.6: c.238 A > G, p.Met80Val) GABRB3 were constructed using the pECMV-EGFP-N vector. The WT plasmid was generated through seamless cloning of the GABRB3 coding sequence. The PCR product was purified by agarose gel electrophoresis using a 1.5% gel, then ligated into a pECMV-3×FLAG-N vector that was digested with the restriction enzymes HindIII and XbaI. Recombinant plasmids were assembled at 50 °C for 15 min and subsequently transformed into competent cells by subjecting them to heat shock of 42 °C for 45 s. The transformed cells were plated on LB agar containing ampicillin, followed by colony verification through sequencing. The mutant plasmid was constructed using overlap PCR to amplify the mutant fragments, GABRB3-T-1 and GABRB3-T-2, under identical conditions. Transfection was performed in 293T cells at 50–60% confluence. Briefly, 2 µg of plasmid was mixed with 4 µl of P3000 reagent (Thermo Fisher Scientific, MA, USA) in 125 µl Opti-MEM for 5 min. The mixture was then combined with 7 µl of Lipofectamine 3000 (Thermo Fisher Scientific) and incubated for 15 min at room temperature (22–25 °C). The transfection complex was then added to the cells maintained at 37 °C and 5% CO₂. The medium was replaced 6 h post-transfection. Plasmids encoding the α1, β3, and γ2 subunits were co-transfected in a 1:1:1 ratio (totaling 3 µg) for the assembly of the α1β3γ2 complex following the same protocol.

Western blot (WB) analysis

Cells were harvested and lysed in RIPA buffer (Thermo Fisher Scientific) supplemented with protease and phosphatase inhibitors after 48 h of transfection. Lysis was conducted through intermittent mixing to ensure thorough disruption. The lysates were centrifuged at 12,000 × g for 10 min to obtain the supernatant, followed by a determination of the protein concentrations using the bicinchoninic acid assay (Thermo Fisher Scientific). Equal protein amounts (20–30 µg) were separated by SDS-PAGE and transferred onto a 0.22 μm PVDF membrane (Bio-Rad Laboratories, CA, USA). The membranes were blocked in tris-buffered saline containing tween-20 (TBST) and 5% non-fat milk for 2 h at room temperature (22–25 °C). The membranes were subsequently incubated with primary antibodies overnight at 4 °C. The antibodies were DYKDDDDK Tag (9A3) Mouse mAb (Cat#8146, 1:1000) and GAPDH (D4C6R) Mouse mAb (Cat#97166, 1:1000), both sourced from Cell Signaling Technology. The membranes were subjected to three 10-minute washes with TBST and then incubated with an HRP-conjugated secondary antibody (Anti-mouse IgG, HRP-linked, Cat#7076, 1:5000; Cell Signaling Technology) for 2 h at room temperature (22–25 °C). The membranes were again washed with TBST, and the protein bands were visualized using an ECL chemiluminescence detection kit (Thermo Fisher Scientific). The bands were quantified using Image J v1.46 software (NIH, Bethesda, USA), with GAPDH as the loading control. This assay was replicated thrice.

Immunofluorescence (IF) assay

The cells were washed thrice with phosphate-buffered saline (PBS) after 48 h of transfection and subsequently fixed with 4% paraformaldehyde for 15 min. The cells were subjected to 3 cycles of 5-minute PBS washes, after which they were permeabilized with 0.3% Triton X-100 (Yeasen Biotechnology, Shanghai, China) for 10 min and washed with PBS. The cells were then blocked in 5% bovine serum albumin (Biosharp, Hefei, China) in PBS for 1 h at room temperature (22–25 °C) and incubated overnight at 4 °C with the primary antibody (DYKDDDDK Tag (9A3) Mouse mAb, Cat#8146, 1:500; Cell Signaling Technology). The cells were then washed thrice using PBS and incubated with an Alexa Fluor 488-conjugated secondary antibody (Anti-Mouse IgG, Alexa Fluor 488, 1:1000; Invitrogen, CA, USA) for one hour at room temperature (22–25 °C) in the dark. The cells were washed with PBS, stained with Hoechst 33,342 (C1028, 1:1000, Sigma, MO, USA) for 10 min at room temperature (22–25 °C) in the dark, rewashed with PBS, and then mounted on slides. Fluorescence images were captured using a Leica TCS SP8 confocal laser scanning microscope (Leica Microsystems, Wetzlar, Germany). Signal quantification was performed using Image J v1.46, with fluorescence intensity normalized to Hoechst 33,342.

Electrophysiology

Whole-cell patch-clamp recordings were conducted in HEK293T cells cultured as a monolayer in 35-mm dishes (Corning, New York, USA) at room temperature (22–25 °C). The cells were transfected with either wild-type (α1β3γ2) or mutant (α1β3(M80V)γ2) receptor cDNAs, using 0.3 µg of each subunit plasmid per dish. The external solution contained 142 mM NaCl, 8 mM KCl, 10 mM D(+)-glucose, 10 mM HEPES, 6 mM MgCl2·6H2O, and 1 mM CaCl2 (pH 7.4, ~ 326 mOsm). In contrast, the internal solution contained 153 mM KCl, 10 mM HEPES, 5 mM EGTA, 2 mM Mg-ATP, and 1 mM MgCl2·6H2O (pH 7.3, ~ 300 mOsm). The cells were voltage-clamped at -50 mV with a chloride reversal potential near 0 mV. A 1 mM GABA stimulus was applied for six seconds using a four-channel SF-77B rapid perfusion system (Warner Instruments Corporation, CT). Whole-cell currents were amplified using an Axopatch 700B amplifier (Molecular Devices, CA, USA), followed by low-pass filtering at 2 kHz and digitization at 10 kHz using a Digidata 1550B system (Molecular Devices), and finally recorded using pCLAMP 10.2 software (Molecular Devices). Data analysis was performed offline using Clampfit 10.2 (Molecular Devices).

Statistical analysis

Statistical analyses were conducted using GraphPad Prism 8.0 (GraphPad, San Diego, CA, USA). Group comparisons were performed using the Student’s t-test. The significance threshold was set at p < 0.05.

Results

Clinical description

The proband was a 16-year-old girl delivered via cesarean section at 40 weeks of gestation due to dystocia. Her birth weight was 3,550 g, and she had no history of asphyxia or prolonged jaundice. Her father experienced convulsions during his childhood, accompanied by abnormal EEG findings. He was treated with an antiepileptic drug for over a year before discontinuing the medication and has been seizure-free ever since. In contrast, her mother was in good health with no underlying conditions. The proband’s detailed electroencephalography (EEG) findings and clinical management are presented in Fig. 1 and Table 1.

Fig. 1.

Representative EEG images in the proband. (A) At 1.5 years of age, the EEG exhibited background activity at 4–5 Hz, with multiple individual spike-slow wave complexes detected in the bilateral posterior head during sleep. (B) A chronological overview of the patient’s developmental milestones and antiepileptic treatment. (C) At 16.7 years of age, the patient’s video EEG during wakefulness demonstrated background activity with mid-amplitude 10–11 Hz alpha rhythm in the bilateral occipital regions. Spike, slow spike-wave, and polyspike-slow wave discharges in the bilateral lateral occipital and mid-posterior temporal regions, with some significant involvement of the left occipital region

Table 1.

Clinical features of patients with de Novo GABRB3 Met80 variants

Patient	1(This report)	2 [15]	3 [16]	4 [17]	5 [17]	6 [18]	7 [19]	8 [20]	9 [21]	Summary
Variant	c.238 A > G p.Met80Val	c.238 A > G p.Met80Val	c.238 A > G p.Met80Val	c.238 A > T p.Met80Leu	c.238 A > T p.Met80Leu	c.239T > A p.Met80Lys	c.239T > A p.Met80Lys	c.239T > C p.Met80Thr	c.239T > G p.Met80Arg
Inheritance	De novo	NA	De novo	De novo	De novo	De novo	De novo	De novo	De novo
Sex	Female	Male	Male	Male	Male	Female	Female	Female	NA	M/F: 1/1
Age	16.7 years	4 years	NA	2.5 years	2.5 years	7 years	NA	12 years	NA
Age at onset	12 months	10.4 months	8 months	4 months	7 months	16 months	15 months	7 years	NA
Diagnosis	DEE	LGS	MAE	WS	DEE	DEE	NA	NA	NA
Seizures	Febrile, absence, GTCS	Generalized tonic	Focal, myoclonic	Epileptic spasm, focal, febrile	Focal, febrile, GTCS	Febrile, tonic/myoclonic	Focal, febrile	Bilateral tonic-clonic	NA	8/8 (100%)
EEG findings	Spikes, spike-and-slow waves, poly-spike waves (> sleep) (16y)	NA	Angelman like trace (4y)	Diffuse slowing, multifocal, hypsarrhythmia	Focal seizure discharge	Hypsarrhythmia, posterior slowing and spike-wave (> sleep)	NA	Frequent notched slow rhythmic wave discharges (sleep)	NA	6/6 (100%)
Development	Mild GDD, learning disorder	GDD/mild ID	Moderate ID	Delay	Delay	Mild motor delay/ID	Mild ID	Mild ID, common language, poor vocabulary	NA	8/8 (100%)
Brain MRI	Normal (8y)	Prominent extra-axial spaces (3y)	Normal	Normal (2 y)	Normal (1 y)	Normal (1y and 3y)	Normal	NA	NA	1/7 (14%)
Other features	No	Macrocephaly, ataxia	Ataxia, stereotypies, ogival palate, dysmorphism	No	No	No	No	No	NA	2/8 (25%)
AEDs	VPA, LEV, LTG	NA	TPM, LTG, PIR	VPA; VGT	VPA, OXC, LEV, TPM	(+/-) VPA, LEV	LEV	VPA	NA
Seizure free	9 years	NA	NA	7 months	1.4 years	NA	Yes	10 years	NA

NA, not available; EEG, electroencephalography; MRI, magnetic resonance imaging; DEE, developmental and epileptic encephalopathy; LGS, Lennox-Gastaut syndrome; WS, west syndrome; MAE, epilepsy with myoclonic-Atonic seizures; GTCS, generalized tonic-clonic seizure; GDD, global developmental delay; ID, intellectual disorder; AED, antiepileptic drug; VPA, valproic acid; LEV, levetiracetam; LTG, lamotrigine; TPM, topiramate; VGT, vigabatrin; OXC, oxcarbazepine; PIR, piracetam

The patient began experiencing convulsive seizures when 1 year old, initially presenting as absence seizures characterized by loss of consciousness, unresponsiveness, a pale complexion, cessation of movement, and a brief recovery after about 10 s. The patient received vitamin D supplementation, but the seizures persisted and gradually increased in frequency, occurring 2 to 6 times per day with episodes of loss of consciousness lasting 2 to 3 min. The patient was admitted to the First Affiliated Hospital of Anhui Medical University after 2 months for further evaluation. Physical examination revealed no significant abnormalities. The patient also had normal liver and kidney function, as well as electrolytes, immunoglobulins, and complement levels. A cranial CT scan revealed no significant abnormalities, while dynamic EEG during seizures revealed rhythmic slow-wave discharges (Fig. 1A). The patient became seizure-free after initiating sodium valproate treatment. The patient was switched to levetiracetam after eight months due to increased daytime sleepiness. The patient remained seizure-free for the next 7 years, but the EEG abnormalities persisted. Levetiracetam was discontinued in favor of lamotrigine at 8 years and 4 months. Two follow-up EEGs showed normal results six months later. However, the subsequent EEGs remained abnormal (Fig. 1B). The patient experienced two febrile convulsions with generalized tonic seizures at a body temperature of 39 °C at age 9. The seizures were characterized by loss of consciousness, upward eye rolling, clenched fists, and tonic limb contractions. These episodes lasted about 1 min before relief and subsequent sleep. A cranial MRI revealed insignificant abnormalities. However, EEG showed high-voltage spikes in the bilateral occipital regions and 3 Hz spike-and-wave discharges during sleep. Genetic testing done to determine the cause of the historic recurring seizures revealed a de novo heterozygous GABRB3 variant, confirming the diagnosis of DEE43.

The patient has been regularly done EEGs and outpatient visits since then, with no further clinical seizures. The most recent EEG, at 16 years and 8 months old, revealed abnormalities. The abnormalities included spikes, spike-and-slow waves, and polyspike-slow waves in the right occipital and mid-posterior temporal regions, with some involvement of the left occipital region, particularly prominent during sleep (Fig. 1C). Since then, the dose of lamotrigine has been adjusted to 75 mg twice daily.

Despite delayed intellectual development, the patient has attended primary and junior high school. The patient could lift her head at 3 months and crawl and sit steadily by 7 to 8 months. She could pronounce simple syllables like “dada” and “mama” at 11 months. However, her motor, language, and cognitive development lagged behind her peers, with reduced muscle tone after attaining 1 year. She could walk independently at 2 years but had poor balance and frequently fell. At 3 years old, she could speak a few simple words, though the pronunciation was unclear. Noteworthy, she received intermittent rehabilitation training throughout this period. She started school at age seven and has since developed a gentle personality with acceptable communication skills despite poor academic performance. She graduated from junior high school with moderate communication skills but less clear language expression.

Variant analysis

Trio-WES revealed a de novo missense variant (NM_000814.6: c.238 A > G) at chromosomal position chr15:27017551 (hg19). The variant was predicted to result in p.Met80Val in GABRB3 of the patient (Fig. 2). The variant was confirmed through Sanger sequencing and was absent in either of the healthy parents (Fig. 2A). A literature review revealed that variants at the Met80 position led to DEE. Of note, the Met80Val variant has been identified as a de novo variant in several families with related phenotypes and confirmed lineage (Table 1) (PS2_VeryStrong) [15, 16]. The Met80Val variant is located in the same codon as other missense variants, Met80Leu and Met80Lys, classified as pathogenic or likely pathogenic (PM5) [17–19]. It is very rare in the population, with a minor allele frequency of 1/1,608,726 in the gnomAD database (v4.1.0), and is listed in dbSNP as rs72708067 (PM2_Supporting). The gnomAD database highlights that the GABRB3 gene has a missense Z-score of 4.46 (≥ 3.09), underscoring its intolerance to missense variation (PP2). Multiple in silico prediction tools, including Provean, SIFT, PolyPhen2 (HDIV), PolyPhen2 (HVAR), MutationTaster, M-CAP, REVEL, and CADD, consistently classify the GABRB3 Met80Val variant as damaging, deleterious, or disease-causing (PP3). These findings support the classification of the GABRB3 Met80Val variant as pathogenic according to the ACMG guidelines (PS2_VeryStrong + PM2_Supporting + PM5 + PP2 + PP3). Trio-WES did not reveal any additional pathogenic, likely pathogenic, or copy number variations in other genes associated with the patient’s clinical phenotype.

Fig. 2.

A schematic diagram showing the GABRB3 p.Met80Val variant. (A) The Sanger sequencing indicating the presence of GABRB3 (NM_000814.6) p.Met80Val variant in the index patient. (B) Multiple sequence alignment for the GABA_AR subunits, performed using MEGA7, indicating highly conserved Met80 residue (highlighted in red boxes). (C) A schematic representation of the β3 subunit of GABA_AR, comprising an extracellular ligand binding domain, α-helix M1-M4 transmembrane bundle, and an intracellular M3-M4 loop, showing the locations of human GABRB3 variants associated with epilepsy. (D) Three-dimensional pentameric structural model of GABA_AR (PDB: 6HUP), consisting of two α1 (brown), two β3 (rose pink), and one γ2 (light purple) subunit, illustrated from the perspective of an extracellular N-terminal top and a transmembrane side. In the β3 subunit, Met80 is located in the β1-β2 coupling loop and forms seven stabilizing contacts (yellow dashed line), whereas the Val80 substitution disrupts some of these contacts

The Met80 residue in GABRB3 is highly conserved in GABRB3 and GABA_AR subunits (Fig. 2B). It is located within the extracellular N-terminal ligand-binding domain (LBD). LBD is crucial for the specific binding of GABA molecules, a key process for GABA_AR function (Fig. 2C). The receptor undergoes a conformational change that opens the chloride ion channel upon GABA binding, thereby regulating neuronal excitability. The structural impact of the missense variant (c.238 A > G, p.Met80Val) was further investigated using 3D modeling based on the CryoEM structure of the human full-length α1β3γ2 GABA_AR (PDB: 6HUP). The LBD of GABRB3 consists of ten β-sheets (β1-β10). Notably, the affected Met80 residue [auth 55] is located in the β1-β2 coupling loop (Fig. 2D). Met80 forms critical interactions with Cys161 [auth 136] and Met163 [auth 138] in the β6-β7 (Cys) loop and with Ser76 [auth 51] and Glu77 [auth 52] in the β1-β2 loop. The substitution of valine for methionine at this position disrupts essential interactions between Met80 [auth 55] and Met163 [auth 138], as well as Glu77 [auth 52]. This alteration potentially affects the protein’s overall structural integrity, thereby compromising the formation and function of receptor complexes.

In vitro functional data

Plasmids expressing the wild-type β3 subunit (NM_000814.6) and the mutant β3(M80V) were constructed to investigate the functional impact of the Met80Val variant in the β3 subunit of the GABA_AR. Western blot analysis revealed a 2.6-fold increase in protein expression of the Met80Val mutant compared to the wild-type (WT) (Fig. 3A). The Met80Val mutant also exhibited higher fluorescence intensity, primarily localized in the cytoplasm (Fig. 3B).

Fig. 3.

Functional analysis of the GABRB3 variant in vitro. (A) Western blot to detect the GFP-tagged GABRB3 expression in cells transfected with empty vector (EV), GFP-GABRB3-WT, and GFP-GABRB3-M80V. β-Actin served as the loading control. The corresponding quantification graph for the relative protein expression (normalized to β-Actin), demonstrating significantly higher expression of the p.Met80Val mutant compared to that of WT and EV (***p < 0.001). (B) Immunofluorescence images showing DAPI (blue) staining for nuclei and GFP (green) for GFP-tagged GABRB3. The top row represents the empty vector control, the middle row shows cells expressing GFP-GABRB3-WT, and the bottom row shows cells expressing GFP-GABRB3-M80V, presenting with significant cytoplasmic localization pattern. Scale bar: 75 μm. (C) Electrophysiological recordings of GABA-induced currents in cells expressing GABRB3-WT and GABRB3-M80V. The left panel displays the representative traces recorded in the presence of 1 mM GABA, while the right panel indicates the bar graphs demonstrating the quantification of peak current amplitudes and percentage of current loss upon co-application of 10 mM zinc and 1 mM GABA. The p.Met80Val mutant had larger currents and reduced current loss compared to WT (***p < 0.001)

Plasmids encoding WT β3 or β3(M80V) were co-expressed with α1 and γ2 in HEK293T cells in a 1:1:1 DNA ratio, followed by measurement of GABA-evoked currents to assess the functional consequences of the β3(M80V) variant on GABA_AR activity. Cells expressing the mutant α1β3(M80V)γ2 receptor displayed a significantly higher current amplitude compared to those expressing the WT receptor (WT: 704 ± 115 pA; M80V: 4326 ± 289 pA) under saturating GABA conditions (1 mM). The cells’ sensitivity to zinc was also evaluated because reduced zinc sensitivity indicated proper incorporation of the γ2 subunit. Noteworthy, the mutant receptor exhibited a significantly smaller reduction in current compared to the WT receptor (WT: 16.8% ± 0.8% vs. M80V: 8.0% ± 0.8%) upon co-application of 10 mM zinc with 1 mM GABA for six seconds (Fig. 3C).

Discussion

The Epi4K Consortium and Epilepsy Phenome/Genome Project identified GABRB3 as a causative gene for DEE. This conclusion was based on four unrelated patients harboring de novo heterozygous variants, p.Asn110Asp, p.Asp120Asn, p.Glu180Gly, and p.Tyr302Cys, exhibiting EOE, developmental delay, and behavioral abnormalities [13]. Functional analyses of these variants revealed reduced GABA-evoked peak currents or altered channel kinetics, leading to compromised inhibitory neurotransmission [24]. Subsequent studies have broadened the phenotypic spectrum. The studies report additional de novo variants, p.Tyr182Phe, p.Gln249Lys, p.Leu256Gln, and p.Ala305Thr. Moreover, a familial variant, p.Thr157Met, associated with genetic epilepsy with febrile seizures plus, ranging from mild febrile to severe seizures and treatment-resistant syndromes, such as Lennox-Gastaut syndrome, has also been reported. Affected individuals are characterized by various epilepsy types, including focal and generalized seizures. The patients also experience intellectual, motor, and language delays and may exhibit behavioral and psychiatric disturbances [25]. The proband in this study experienced an absence of seizures at 12 months and febrile generalized tonic-clonic seizures at 9 years with mild intellectual disability, which mirrors the core clinical features observed in GABRB3-related epilepsy. A review of nine patients with five distinct p.Met80 variants, p.Met80Val, p.Met80Leu, p.Met80Lys, p.Met80Thr, and p.Met80Arg, reveals common features (Table 1). The features included seizure onset between 4 months and 7 years with a median of 12 months, diverse seizure types including absence (1/8), febrile (5/8), and generalized tonic-clonic seizures (4/8), mild-to-moderate intellectual disability (8/8), consistent EEG abnormalities (6/6), and ataxia (2/8) and imaging anomalies (1/7) in some instances. Notably, seizure freedom was achieved in all treated cases (5/5), suggesting that the variant’s localization potentially influence both clinical severity and therapeutic response. This phenotypic dichotomy aligns with the structural-functional hierarchy of GABA_AR domains, underscoring the complex interplay between genetic variation and clinical manifestations in GABRB3-related disorders.

GABA_ARs are pentameric, Cys-loop ligand-gated chloride channels assembled from various subunits. The subunits include α (GABRA1–6), β (GABRB1–3), γ (GABRG1–3), and other subunits. Of note, the β3 subunit (473 amino acids) comprises an extracellular N-terminal LBD, four transmembrane domains (TMDs), and a cytoplasmic loop between TMD3 and TMD4 [7]. Pathogenic variants in critical regions, such as the GABA binding site or TMDs, disrupt chloride ion flow and neuronal inhibition [26]. Large-scale analyses suggest that extracellular domain variants in GABRB3 typically correlate with generalized epilepsy with a median onset of 12 months and mild-to-moderate intellectual disability. In contrast, transmembrane domain variants are associated with earlier-onset focal epilepsy with a median onset of 4 months and more severe intellectual disability [14]. Herein, the de novo p.Met80Val variant in the LBD disrupted the hydrophobic interactions within the β1–β2 loop and Cys-loop, which form part of the coupling region that mediates the conversion of the GABA binding event into the gating of the channel (Fig. 2). Patients with the p.His83Asn variant in the same loop experience mild epilepsy with febrile seizures and achieve seizure freedom with monotherapy [27]. Analogous variants in paralogous subunits such as GABRB2 p.Met79Thr underscore the critical role of this region in maintaining receptor function and modulating seizure susceptibility [28, 29]. These findings suggest that the extracellular domains mediating ligand binding are crucial determinants of channel activity and contribute significantly to the observed clinical variability.

Recent studies postulate that genetic variants in GABA_AR subunits lead to GOF or LOF effects, resulting in distinct clinical manifestations, even among homologous substitutions [3]. Pathogenic GABRB3 variants are broadly classified as GOF or LOF mutations. GOF mutations enhance GABA sensitivity, exemplified by p.Thr281Ala, and are associated with severe EOE, focal epilepsy, profound intellectual disability, and poor treatment response. LOF mutations, such as mutations in p.Thr281Ile, typically lead to milder epilepsy, later-onset seizures, moderate intellectual disability, and better responsiveness to antiepileptic drugs [28, 30]. Animal models highlight the critical role of GABRB3 in seizure disorders. For instance, global and forebrain-specific knockout mice exhibit behavioral deficits. In contrast, knock-in models harboring variants such as p.Asp120Asn, p.Asn110Asp, and p.Asn328Asp recapitulate human DEE phenotypes, including spontaneous seizures, diverse ictal EEG patterns, and cognitive and behavioral impairments [31–33]. Specifically, Gabrb3+/Asp120Asn mice exhibit reduced miniature inhibitory postsynaptic current amplitudes and prolonged thalamocortical oscillations. In contrast, Gabrb3+/Asn110Asp mice exhibit decreased cortical mIPSCs and extended oscillatory firing, while Gabrb3+/Asn328Asp mice exhibit reduced β3 subunit expression in crucial brain regions, collectively mirroring Lennox-Gastaut syndrome [31–33]. Herein, the p.Met80Val variant exhibited increased protein expression relative to the wild-type. Immunofluorescence analysis revealed enhanced fluorescence predominantly in the cytoplasm, suggesting impaired receptor trafficking and reduced membrane localization, potentially contributing to epilepsy phenotypes. Previous studies postulate that specific GABA_AR variants reduce surface expression, influencing epilepsy pathogenesis [34]. Functionally, the mutant α1β3(M80V)γ2 receptor displayed increased sensitivity to GABA, possibly because of altered channel gating, resulting in enhanced currents. Notably, the p.Met80Val variant exhibited significantly lower current loss upon zinc application, which is a known GABA_AR inhibitor. This finding suggested altered zinc binding or conformational changes, potentially influenced by γ2 subunit incorporation. Residues within extracellular coupling loops play a crucial role in channel gating. Noteworthy, both GOF and LOF variants are identified in these regions. Variants in the β1-β2 loop include GOF (β3-Glu77Lys and β3-Val78Phe), exhibiting a 2.5- to 4.4-fold increase in GABA sensitivity, and LOF (β3-Ser76Cys, β3-Met80Thr, and β3-Met80Lys), exhibiting a 2.5- to 3.6-fold decrease. Similarly, the β6-β7 loop contains GOF (β3-Leu170Arg), exhibiting a 1.8-fold increase, and LOF (β3-Leu165Gln and β3-Arg166Ser), exhibiting a 2.0- to 10-fold reduction in GABA sensitivity [20]. Traditionally, reduced GABA_AR-mediated inhibition (LOF) was linked to epilepsy. However, emerging evidence suggests that GOF and LOF variants, often near the GABA binding site, can produce overlapping clinical phenotypes. These findings underscore the complexity of genotype-phenotype correlations and indicate that the observed phenotypic variability in GABRB3-related disorders is potentially influenced by additional factors, such as modifier genes or environmental influences [3, 20, 35]. Notably, a model attributing phenotype to impaired surface expression or gating deficits is conceptually intuitive because absent or non-functional receptors would render other in vitro findings irrelevant. Nonetheless, it remains challenging to determine whether the reduced receptor currents observed in heterologous systems accurately reflect synaptic dysfunction in neurons.

Precision medicine holds significant promise for managing neonatal epilepsy [36]. Spagnoli et al. emphasized the importance of early diagnosis and intervention in neonatal seizures, suggesting that prompt treatment can significantly improve long-term outcomes [2]. These reports underscore the necessity of developing targeted therapeutic strategies based on specific genetic factors. Of note, effective therapeutic strategies for GABRB3-related DEE should account for the functional impact of pathogenic variants because approximately half result in GOF while the other half result in LOF, disrupting GABA_AR activity [17]. These functional differences influence clinical presentation and treatment response. Patients with LOF variants often achieve seizure control with antiseizure medications (ASMs), such as valproate (VPA) and levetiracetam. However, persistent EEG abnormalities may remain, as observed in this study. Despite the effectiveness of VPA, it requires careful monitoring due to risks such as dose-dependent thrombocytopenia [37]. Targeted therapies, including GABA_AR potentiators such as vinpocetine, effectively restore inhibitory currents in LOF variants in preclinical models [38]. Conversely, GOF variants may benefit from GABA_AR antagonists, such as bicuculline, because GABA-enhancing drugs, such as vigabatrin, exacerbate symptoms [39, 40]. The patient herein experienced recurrent absence seizures at 12 months, initially controlled with VPA, followed by levetiracetam and later lamotrigine (100 mg BID), which achieved seizure freedom despite persistent EEG abnormalities. Moreover, the patient exhibited mild intellectual disability and communicated adequately, though with unclear speech. Vitamin D reduces seizure frequency through various mechanisms, including modulating neurotransmitter synthesis, promoting neurotrophic factor production, and inhibiting oxidative stress [41]. Herein, the patient’s seizures remained uncontrolled despite vitamin D supplementation. This phenomenon highlights the differences in clinical management between GABRB3-related epilepsy and symptomatic neonatal epilepsy. The core issue in GABRB3-related epilepsy is the inherent receptor defect in GABA-mediated inhibition rather than transient metabolic disturbances. In contrast, symptomatic epilepsy is often associated with electrolyte imbalances that alter neuronal excitability, which can be effectively alleviated by acute corrective measures, such as calcium and glucose infusions [2]. GABRB3-related epilepsy thus requires more precise medical strategies rather than simply correcting metabolic disturbances.

One limitation of this study is that it was based single-patient cohort and reliance on in vitro models. Future studies should thus focus on (1) expanding the cohort sizes to strengthen genotype-phenotype correlations; (2) developing Gabrb3 knock-in animal models to assess the long-term effects of p.Met80Val; (3) determining the modifier genes and environmental factors contributing to phenotypic variability; and (4) exploring novel therapeutic candidates such as magnolol in preclinical models to identify more effective treatment options [42]. Moreover, large-scale studies are needed to stratify neonatal seizures into symptomatic and genetic categories and to map corresponding electrolyte and receptor profiles. This step may prompt more systematic screening protocols in neonatal intensive care units to facilitate early identification and intervention for neonatal seizures. These efforts would advance precision medicine for GABRB3-related DEE and provide a more scientific basis for treating neonatal seizures.

Conclusions

This study provides crucial mechanistic insights into GABA_AR dysfunction caused by GABRB3 variants. It analyzes the ninth reported case of a recurrent mutation at the p.Met80 residue and underscores the necessity of integrating genetic, functional, and clinical data to advance precision medicine. GABRB3 variants consistently lead to mild developmental delay/intellectual disability and EOE, with typical favorable responses to ASMs. This study comprehensively reviews Met80-related variants and deciphers the functional consequences of p.Met80Val through structural analysis and in vitro assays assessing receptor stability and function to highlight their pathogenic relevance. The findings of this study underscore GABRB3 as a crucial gene in epilepsy pathogenesis and identify p.Met80 as a mutational hotspot. It provides critical insights for diagnosis, management, and therapeutic development in GABRB3-related disorders.

Abbreviations

DEE

Developmental and epileptic encephalopathy

EOE

Early-onset epilepsy

EEG

Electroencephalography

AED

Antiepileptic drug

WES

Whole-exome sequencing

ACMG

American academy of medical genetics and genomics

WB

Western blot

IF

Immunofluorescence

Author contributions

LY conceptualized the study. ZZ curated the data and developed the methodology. JJ conducted the formal analysis and validated the results. BW acquired the funding and contributed to the visualization. RH investigated the topic and wrote the original draft. RH also provided resources. JW developed the software. DW supervised the project and reviewed and edited the manuscript. All authors read and approved the final manuscript.

Funding

The work was funded by the National Natural Science Foundation of China (Grant No. 81472167).

Data availability

The datasets used and analyzed in the current study are available in the article.

Declarations

Ethics approval and consent to participate

Informed consent documents for the study were obtained from the patient’s parents. The study was approved by the Medical Ethics Committee of the First Affiliated Hospital of Anhui Medical University (No. PJ2024-12-62) and was conducted by HIPAA regulations and with the tenets of the Declaration of Helsinki.

Consent for publication

Consent for publication was obtained from the institution and the patient’s parents.

Competing interests

All authors have declared no conflicts of interest.