LAUSR

To the Editors: We were encouraged to read the recent publication by Vogel et al.1 in Epilepsia providing the first association between a de novo missense variant in GABRA4 and a neurodevelopmental disorder with earlyonset epilepsy.1 The GABRA4 gene encodes the α4 subunit of the γaminobutyric acid (GABA) type A receptor, and this subunit is relatively abundant in the cortex, hippocampus, and thalamus, all brain regions known to be involved in epilepsy.2,3 The α4 subunit primarily assembles with β and δ subunits to form, for example, α4β2δ receptors,4– 6 and these receptors are localized in extrasynaptic membranes where they respond to low ambient levels of GABA and spill over from synaptic release resulting in longlasting tonic inhibition of neuronal activity.7– 9 We recently discovered that pathogenic variants in GABRD, encoding the δ subunit, cause gainoffunction traits in α4β2δ receptors and interestingly one specific variant, GABRD p.Thr291Ile, is paralogous to the GABRA4 p.Thr300Ile variant.10 Intrigued by this observation, we extended the study of Vogel et al.1 with electrophysiological analysis of the GABRA4 p.Thr300Ile variant in combination with the δ subunit using previously described methodologies.10,11 The mean current amplitude obtained with a maximally efficacious concentration of GABA was increased by 6.2fold and the sensitivity to GABA was increased by ~10 fold for α4T300Iβ2δ vs wildtype receptors (Figure 1). Furthermore, the maximum estimated open probability was increased ~18fold, showing an increased ability of GABA to gate variant α4T300Iβ2δ receptors. Like Vogel et al.1 we observe that variant receptors display faster desensitization kinetics than wildtype receptors at high GABA concentrations (data not shown); however, α4β2δ receptors are extrasynaptic receptors that respond to low concentrations of GABA in the brain, and no obvious desensitization was observed with GABA concentrations below 1 μM. Thus despite the inherent complexity of receptor desensitization kinetics, we conclude that the increases in current amplitudes and sensitivity to GABA caused by the GABRA4 variant bestow extrasynaptic δcontaining receptors with gainoffunction properties. Of interest, the GABRA4 Thr300 amino acid position appears to be a hotspot for pathogenic variants in most if not all GABAAR subunit classes. Besides the paralog GABRD p.Thr291Ile variant mentioned above,10 we recently described the functional consequence of the paralogous variant in GABRB3 p.Thr287Ile.12,13 GABRB3 encodes the β3 subunit, and we observed that this epilepsyassociated variant also causes strong gainoffunction traits. Hence, a threonine to isoleucine substitution in this specific protein position appears to cause gainoffunction traits irrespective of the subunit type. When comparing the clinical manifestations, there are similarities as well as differences between the carriers of the GABRA4 variant1 and the paralog GABRD and GABRB3 variants.10,12 The GABRA4 variant, which was observed in mosaic state (17%) in a 5.5yearold girl, was associated with intractable nocturnal frontal lobe seizures (onset 3.5 years), dyspraxia, and attention deficit.1 In comparison, the GABRD variant was observed to cause earlyonset (1– 4 years) generalized epilepsy with intractable atypical absence seizures, various degrees of learning difficulties/intellectual disability, and attentiondeficit/ hyperactivity disorder (ADHD) in a mother and her twin sons.10 Finally, the GABRB3 variant was observed in a child with an unclassified developmental and epileptic encephalopathy (onset 3 months) with intractable tonic,