GABRA2 ‐related encephalopathy: Identification of two phenotypes with distinctive electroclinical features

Abstract

Pathogenic variants in γ‐aminobutyric acid type A (GABA_A) receptor subunit genes are increasingly associated with epilepsy and neurodevelopmental disorders. Pathogenic variants in GABRA2, encoding the α‐2 subunit of GABA_A receptors, have been recently reported. This study aims to better delineate the phenotypic spectrum of GABRA2 pathogenic variants. We conducted a retrospective multicenter study, analyzing six new patients with GABRA2 pathogenic variants identified through a French national collaboration. Clinical, electroencephalographic (EEG), and genetic data were reviewed alongside a literature analysis of eight previously reported cases. Two distinct electroclinical phenotypes were identified. The most severe, in four of six patients, featured early infantile developmental and epileptic encephalopathy with an EEG pattern of rapid rhythms suggestive of GABAergic hyperactivity. The milder phenotype, in two of six patients, included later onset, drug‐responsive epilepsy with moderate developmental delay. A literature review confirmed these phenotypes and supported genotype–phenotype correlations, with transmembrane domain variants more frequently associated with severe phenotypes. This study refines the phenotypic spectrum of GABRA2‐related disorders, highlighting two distinct electroclinical phenotypes. The identification of a recognizable EEG pattern of unusual rapid rhythms for age may be a biomarker for early diagnosis of a severe phenotype and suggests a potential underlying gain‐of‐function mechanism, to be confirmed by functional studies.

Key points.

• GABRA2 pathogenic variants were associated with two distinct phenotypes: early‐infantile developmental and epileptic encephalopathy (EIDEE) and milder developmental encephalopathy with late‐onset epilepsy.

• A distinctive pattern of rapid‐rhythms in the alpha/beta band, observed in the absence of GABAergic modulators, was identified in the EIDEE phenotype.

• Transmembrane domain variants were more frequently associated with the most severe phenotype (EIDEE).

1. INTRODUCTION

Pathogenic variants in genes encoding γ‐aminobutyric acid type A (GABA_A) receptor subunits are increasingly linked to a wide phenotypic spectrum, ranging from genetic epilepsy with febrile seizures plus or genetic generalized epilepsy to early infantile developmental and epileptic encephalopathy (EIDEE). ¹ , ² , ³ , ⁴ The GABRA2 gene encodes the α‐2 subunit, which is abundantly expressed in several key brain regions such as the hippocampus, frontal cortex, and amygdala. ⁵

The association of GABRA2 with epilepsy is relatively recent, with the first presumed pathogenic variant reported in 2018 in a patient with a phenotype consistent with EIDEE. ⁶ Since then, seven additional patients with GABRA2 pathogenic variants have been reported in the literature. Among these, five patients presented with EIDEE, whereas two patients had milder developmental encephalopathy with drug‐responsive epilepsy starting at 2 and 17 years of age, respectively. ⁶ , ⁷ , ⁸ , ⁹

We report here the electroclinical phenotype of six new patients carrying pathogenic GABRA2 variants and review data for previously reported patients to further delineate the phenotypic spectrum of GABRA2‐related epilepsy and explore potential genotype–phenotype correlations.

2. MATERIALS AND METHODS

2.1. Data collection

We conducted a retrospective multicenter study including patients with a pathogenic GABRA2 variant identified through the French Reference Centers of Rare Epilepsies and the EPIGENE network. We collected information on the medical history, electroencephalography (EEG), and brain imaging findings. Cognitive development was assessed by the referring physician according to developmental scales used at their institution or based on clinical evaluation. Informed consent was obtained from patients or their legal guardians, in compliance with the Declaration of Helsinki. We used a standardized survey filled in by each referring clinician. We classified seizure types and epilepsy syndromes according to the International League Against Epilepsy classification. ¹⁰ , ¹¹ , ¹²

2.2. Variant analysis

GABRA2 variants were identified by targeted next generation sequencing gene panels for five patients and genome sequencing for one patient (Patient 2). All variants were confirmed by Sanger sequencing, as was familial segregation analysis when available. We classified variants according to the guidelines of the American College of Medical Genetics and Genomics. ¹³ The genome reference sequence used was GRCh37 (hg19), and transcript NM_000807.4 was used for variant annotation.

2.3. Literature review

We analyzed all cases reported to date (up to August 2024) with a pathogenic mutation in the GABRA2 gene using the PubMed website with the terms “GABRA2” and “variants” or “mutations”.

3. RESULTS

We collected data from six unpublished patients with GABRA2 variant, including four males and two females. The main clinical, electrophysiological, and therapeutic features are reported in Table 1.

TABLE 1.

Clinical, EEG, and genetic features from our series of patients with GABRA2 pathogenic variant.

Patient	1	2	3	4	5	6
Age, years/sex	1.5/F	34/M	2/M	7/F	6/M	5/M
Variant	c.799C>G	c.862A>G	c.943 T>C	c.851 T>C	c.910C>G	c.690del
(NM_000807.4)	p.(Leu267Val)	p.(Thr288Ala)	p.(315Trp>Arg)	p.(Val284Ala)	p.(Leu304Val)	p.(Phe230Leufs*3)
ACMG classification	Likely pathogenic (PS2, PM2m, PP3s)	Likely pathogenic (PM1m, PM2m, PP3s)	Likely pathogenic (PS2, PM2m, PP3s)	Pathogenic (PS2, PS3, PM2m, PP3s)	Likely pathogenic (PS2, PM2m, PP3s)	Likely pathogenic (PVS1, PM2m)
Inheritance	de novo	Not in mother	de novo	de novo	de novo	Not in mother
Seizure onset/offset	3 months/ongoing	4 months/ongoing	3 months/ongoing	3 months/ongoing	15 months/5 years	33 months/5 years
Seizure type	Clusters of focal seizures with rightward gaze deviation, left hemibody clonus, and hypersalivation	Focal seizures with apnea, cyanosis, limbs hypertonia and axial hypotonia, ocular revulsion Generalized clonic seizures	Clusters of focal seizures with autonomic signs, fixed or ocular deviation, hypertonia of one arm Spasms, head drop	Clusters of focal seizures with wide‐eyed staring, noisy breathing, erythema Focal seizures and tonic seizures, clusters of spasms	Febrile seizures Atypical absences	Impaired awareness seizures Generalized tonic–clonic seizures
EEG
Onset	Rapid rhythms predominantly on the left side	Background activity overloaded with rapid rhythms Slow waves and spikes in the left temporal region	Diffuse and unusual rapid rhythms for age, focal temporal discharges	Diffuse fast frequencies with bilateral centrotemporal predominance	Ample, synchronous theta‐band activity, bilateral spindles	Normal
Current	Diffuse slow waves, multifocal sharp waves, fast activity with a centroposterior predominance	General slowing of background activity overloaded with rapid rhythms	Occipital delta bilateral activity with diffuse rapid rhythms	Delta activity in the occipital regions and rapid rhythms Spike‐and‐wave and polyspike‐and‐wave, activated by sleep	Well‐organized background activity, diffuse slow‐wave spikes during deep sleep	Normal background, generalized spikes and waves
ASM
Previous	CBZ	PHB, CBZ, VPA, CLN, PHT, LCS	CBZ, CLN, CLB, VPA, LTG, VGB	VGB, VPA, LTG, CBD, PER, ketogenic diet	VPA, CLB, PHB, LEV	VPA, LTG
Current	CBZ	CBZ, LCS	VPA, LTG, VGB	VGB, VPA, LTG, CBD, PER	None	None
Developmental delay	Severe	Severe	Profound	Profound	Moderate	Moderate
Other	Feeding difficulties	Walking at 3 years, speech of 10 words Autoaggression, ASD	No walking, nonverbal ASD, feeding difficulties	Dystonic movements Scoliosis	Walking at 19 months, first words at 3 years ASD	Walking 21 at months ASD

Abbreviations: ACMG, American College of Medical Genetics and Genomics; ASD, autism spectrum disorder; ASM, antiseizure medication; CBD, cannabidiol; CBZ, carbamazepine; CLB, clobazam; CLN, clonazepam; EEG, electroencephalographic; F, female; LCS, lacosamide; LEV, levetiracetam; LTG, lamotrigine; M, male; PER, perampanel; PHB, phenobarbital; PHT, phenytoin; VGB, vigabatrin; VPA, valproic acid.

3.1. Electroclinical phenotype

All patients had epilepsy. The median age at seizure onset was 3.5 months (range = 3 months–33 months).

Four patients had EIDEE (Patients 1, 2, 3, and 4). They all had focal seizures at between 3 and 4 months of age, occurring in clusters. Autonomic manifestations (flushing, hyperventilation, pupillary dilatation, sialorrhea, and desaturation) were reported in all of them, sometimes associated with tonic or clonic features and requiring the use of rescue medication. Patients 5 and 6 developed epileptic spasms before the age of 1 year. At the time of the study, all four patients had active epilepsy, with weekly generalized seizures (tonic, atonic) in three patients (Patients 2, 3, and 4) and monthly brief focal clonic seizures under carbamazepine monotherapy in Patient 3. Of interest, awake and sleep interictal EEG pattern in these four patients was characterized by diffuse rapid rhythms in the alpha/beta‐band, in the absence of benzodiazepine treatment or other GABA modulators. In Patient 3, this observation was further supported by a quantitative analysis using power spectral density, which revealed a clear peak in the alpha frequency range (Figure 1). In Patients 3 and 4, these rapid rhythms were already present on the first EEG recordings, performed at 3 months of age as part of the initial evaluation, and before the introduction of any antiseizure medication. This excess of rapid rhythms was accentuated during sleep in three patients (Patients 1, 3, and 4). There were no physiological sleep patterns. All four patients had severe to profound intellectual disability. Three of them had severe hypotonia with poor eye contact (Patients 1, 3, and 4), whereas Patient 2 walked at age 3 years and was able to speak a few words at age 33 years. Vigabatrin (VGB) was reported as effective on spasms in Patients 3 and 4.

FIGURE 1.

Example of scalp electroencephalography (EEG) in Patient 3, recorded at 3 months of age, before initiation of any antiseizure medication. (A) Twenty‐second scalp EEG during quiet sleep, showing abnormal bilateral diffuse fast rhythms in the alpha/beta band predominant in the centroparietal regions (10/20 longitudinal montage, band‐pass filter = .5–70 Hz, sampling rate = 256 Hz); (B) F4‐C4 and F3‐C3 bipolar power spectrum density (Welch, 1‐s window length overlapping at 50%) over the same period showing a peak in the alpha frequency band around 13 Hz.

The other two patients had developmental encephalopathy with later onset and pharmacosensitive epilepsy (Patients 5 and 6). Patient 5 had stormy episodes of short generalized clonic febrile seizures from the age of 15 months and atypical absences from the age of 28 months, controlled by valproic acid and levetiracetam. EEG showed reactive high‐amplitude theta background activity with physiological anterior–posterior organization and bilateral sleep spindles, with intermittent bursts of slow delta waves during quiet sleep. Patient 6 had nonmotor seizures with behavioral arrest and ocular revulsion at 33 months, controlled by lamotrigine. He also had sporadic generalized clonic seizures between 3.5 and 5 years of age. EEG recordings exhibited normal background activity with physiological sleep patterns and rare bilateral spike‐waves. Seizures were not recorded during video‐EEG monitoring. Seizures remitted in both patients by age 5 years, allowing medication withdrawal. They had moderate developmental delay with first concerns at 2 years for Patient 5 and 9 months for Patient 6. They acquired independent walking before the age of 2 years but had expressive and receptive oral language disorder with autistic traits.

3.2. Magnetic resonance imaging

Brain magnetic resonance imaging (MRI) was normal in four of six patients. MRI of Patient 2 showed diffuse cortical atrophy, and delayed myelination was reported at 6 months in Patient 3.

3.3. GABRA2 variants

We identified six heterozygous missense variants in GABRA2, including five novel variants (four missense and one frameshift) and one previously reported missense variant (p.Val284Ala). ⁷ According to American College of Medical Genetics and Genomics guidelines, all variants were classified as likely pathogenic or pathogenic (Table 1). Four variants occurred de novo, whereas inheritance could not be confirmed for two cases (no paternal testing in Patients 2 and 6). Four variants were in the transmembrane (TM) region (Patients 1, 2, 3, and 4), a domain known to play a critical role in receptor function, ¹⁴ whereas two variants were in extracellular domains (Patients 5 and 6; Figure S1).

3.4. Literature review and phenotype–genotype correlation

The review of eight previously reported patients with GABRA2 variants highlights a phenotypic spectrum consistent with our findings (Table S1). Five patients exhibited EIDEE, with clusters of focal seizures before the age of 2 months and developmental delay evolving into severe cognitive impairment. ⁶ , ⁷ , ⁹ Two of these patients developed epileptic spasms during the course of developmental and epileptic encephalopathy. ⁷ , ⁹ Rapid rhythms on EEG were reported in two cases, ⁶ , ⁷ whereas the EEGs of the three others were described as disorganized with multifocal diffuse spikes. ⁷ , ⁹ In contrast, three patients presented a milder phenotype, characterized by normal or slowed EEG patterns, well‐controlled seizures with treatment, and moderate intellectual disability. ⁷ , ⁸ This group included two brothers who inherited the variant from their mosaic father, ⁷ whereas all other patients had de novo missense variants.

When combining our six patients with the eight previously reported cases, we observed that missense variants associated with EIDEE were predominantly located in the TM region (7/8 variants), whereas those linked to a milder phenotype were mainly found in extracellular domains (2/3 variants; Figure S1). Patient 4, carrying the recurrent p.(Val284Ala) variant, exhibited an EIDEE phenotype similar to the previously reported case. ⁷

4. DISCUSSION

This study, including our six new patients and eight previously reported cases, highlights two distinct phenotypes associated with GABRA2 variants. The most frequent, observed in nine of 14 patients, was an EIDEE beginning with a stormy phase of focal seizures before the age of 6 months. In our cohort, the detailed seizure semiology revealed prominent autonomic features, including facial flushing, hypersalivation, pupillary dilation, and respiratory abnormalities, often initially misinterpreted as nonepileptic “spells.” This clinical pattern may represent a recognizable clinical aspect of GABRA2‐related encephalopathy. In addition, the presence of diffuse, abnormal fast rhythms in the alpha/beta range was observed on interictal EEGs of six of nine patients with this severe phenotype. In two patients from our series, these rapid rhythms were reported since the first EEG recordings, before initiation of any antiseizure medication. In all patients, they were observed in the absence of benzodiazepines. Although their physiological significance remains to be fully understood, these EEG along with clinical features may serve as useful electroclinical biomarkers for early recognition of GABRA2‐related encephalopathy. Their identification could assist both clinicians in providing early prognostic guidance to families and geneticists in variant interpretation, particularly in cases involving variants of uncertain significance. Despite seizure control in some of them, these patients exhibited severe developmental delay evolving into profound and multiple disabilities. The second phenotype, seen in four of 14 patients, involved later onset, pharmacosensitive epilepsy with seizure resolution during childhood and moderate intellectual disability.

In recent years, variants in genes encoding GABA_A receptor subunits (including GABRG2, GABRA1, GABRA3, GABRA5, GABRB1, GABRB2, GABRB3, and GABRD) have been reported as responsible for an increasing number of monogenic epilepsies. ⁴ Although specific genotype–phenotype correlations remain elusive for most of these genes, ³ several studies have suggested that variants located in the TM region of GABA_A receptor subunits tend to be associated with more severe phenotypes, whereas those in extracellular domains are more often linked to milder forms. ³ Our findings appear to support this trend in the case of GABRA2, although this should be interpreted with caution, given the small number of patients in our study and the possibility of confounding factors such as diagnostic testing bias, whereby patients with more severe phenotypes are more likely to undergo genetic testing. Additionally, we report for the first time a GABRA2 frameshift variant associated with a moderate phenotype, likely due to a haploinsufficiency mechanism. Functional studies of other GABAA receptor genes provide insights into genotype–phenotype correlations. Interestingly, recent studies demonstrated that gain‐of‐function (GOF) variants in GABRB2 and GABRB3 were associated with a phenotype of EIDEE, whereas loss‐of‐function variants result in milder phenotypes. ² , ¹⁵ Most notably, patients with GOF variants exhibited severely disorganized EEG background activity intermixed with alpha‐band rapid activity that closely resembles the rapid‐rhythm EEG pattern identified in our patients with the EIDEE phenotype. ¹⁵ These rapid rhythms look very similar to those seen during benzodiazepine treatment, which enhances GABAergic activity and promotes the synchronization of neural firing at beta frequencies. ¹⁶ In our cohort, these rhythms were clearly present from the first EEG recordings in at least two patients, prior to any antiseizure treatment, suggesting that they reflect an intrinsic feature of the disease rather than a medication effect. Similar EEG patterns have been described in dup15q syndrome, where overexpression of clustered GABA_A receptor genes (GABRB3, GABRA5, GABRG3) is thought to cause GABAergic hyperactivity. ¹⁷ , ¹⁸ Associations between GABRA2 polymorphisms and increased beta activity on EEG have also been reported in population‐based studies. ¹⁹ , ²⁰ Finally, functional data from previously published variants (p.Val284Ala and p.Thr292Lys) showed enhanced sensitivity to GABA, ⁷ , ²¹ which may support a GOF mechanism in at least some cases. Although these observations raise the possibility that GABRA2 variants may lead to excessive GABAergic activity and that rapid EEG rhythms could serve as a functional biomarker, this hypothesis remains to be demonstrated. Larger cohorts and experimental studies will be required to validate these findings and explore their mechanistic significance.

Although the number of patients remains limited, the pharmacological response in our cohort appears variable. GABAergic enhancers such as clobazam or phenytoin did not seem effective in patients with the severe phenotype, particularly during the highly active early phase of epilepsy. However, VGB showed efficacy in the two patients who presented epileptic spasms. These observations warrant further investigation to refine therapeutic strategies tailored to the underlying functional impact of GABRA2 variants.

5. CONCLUSIONS

This study expands the phenotypic spectrum associated with GABRA2 variants, highlighting two distinct electroclinical phenotypes and suggesting genotype–phenotype correlations. In the more severe early onset form, the combination of prominent autonomic seizure manifestations and rapid interictal EEG rhythms, present before any treatment, may serve as early clinical and electrophysiological biomarkers. These features could help guide early diagnosis, inform prognostic counseling, and support variant interpretation in a diagnostic setting. Functional studies remain essential to confirm the suspected GOF mechanism and to better understand the pathophysiological basis underlying the severe phenotype. Ultimately, such advances may contribute to the development of targeted therapies tailored to variant‐specific effects in GABA_A receptor‐related encephalopathies.

AUTHOR CONTRIBUTIONS

Marie Adamo‐Croux, Mathieu Milh, and Claire Bar contributed to study design, data collection, analysis, interpretation of data, and the original draft of the manuscript. Jérôme Aupy performed EEG analysis and reviewed the final draft of the manuscript. Laurent Villard, Yorsa Halleb, Gaetan Lesca, Perrine Charles, Boris Keren, and Manon Degoutin performed genetic analysis and interpretation, and reviewed the final draft of the manuscript. Chloé Angelini, Nathalie Villeneuve, Arnaud Chefdor, Maxime Colmard, Nicole Chemaly, Mathieu Milh, and Cyril Goizet contributed to data collection and interpretation, and to revision of the manuscript for intellectual content.

CONFLICT OF INTEREST STATEMENT

None of the authors has any conflict of interest to disclose. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.

Supporting information

Figure S1.

Table S1.

Adamo‐Croux M, Angelini C, Aupy J, Villard L, Villeneuve N, Chefdor A, et al. GABRA2 ‐related encephalopathy: Identification of two phenotypes with distinctive electroclinical features. Epilepsia. 2025;66:e187–e193. 10.1111/epi.18507

DATA AVAILABILITY STATEMENT

The data supporting the findings of this study are available from the corresponding author upon reasonable request.