Flip of the Switch: Targeting GABA Dysregulation to Treat Nonseizure Comorbidities in Dravet Syndrome

Commentary

Dravet syndrome (DS) is a severe neurodevelopmental disorder defined by treatment-resistant epilepsy that begins in the first year of life, prominent temperature-sensitive seizures, and a high risk of sudden unexpected death in epilepsy (SUDEP). In addition to epilepsy, DS is characterized by behavioral and cognitive impairments, including sleep disturbances, developmental delay, and prominent features of autism spectrum disorder (ASD), with a formal diagnosis of ASD in up to 67% of cases. ¹

DS is largely caused by heterozygous loss-of-function variants in SCN1A, which encodes the voltage-gated sodium channel α subunit Nav1.1. Nav1.1 is preferentially expressed in inhibitory γ-aminobutyric acid (GABAergic ergic interneurons), where it plays a critical role in mediating action potential generation and propagation. Extensive studies characterizing DS (Scn1a+/−) mouse models have demonstrated that interneurons are hypoexcitable, with impaired spike generation and synaptic transmission.^2,3 These findings support the “interneuron hypothesis” of DS, which explains how a paradoxical loss of Na + channel function can lead to network hyperexcitability.

Recent studies have built on this hypothesis, revealing multiple modes of altered GABAergic signaling in DS, a feature also implicated in other neurodevelopmental disorders.^4–6 One such locus of dysregulation is the direction of GABA action (de- or hyperpolarizing), which is determined by the chloride (Cl⁻) homeostasis of the postsynaptic cell and is maintained by the chloride transporters Na⁺–K⁺–2 Cl⁻ cotransporter-1 (NKCC1) and K⁺–Cl⁻ cotransporter-2 (KCC2). ⁴ In early development, NKCC1 creates high intracellular chloride, and activation of GABA_A receptors drives Cl-efflux, producing a depolarizing–potentially excitatory–response. As the nervous system matures, increased expression of the Cl− extruder KCC2 lowers intracellular Cl− and promotes a hyperpolarizing, inhibitory action. This developmental transition is known as the “GABA switch” and is relatively complete by the second postnatal week in rodents and 25 weeks post conception in humans. ⁴ Studies employing human tissue and rodent models show that DS samples exhibit a more depolarized GABA_A reversal potential^5,6 and an elevated NKCC1/KCC2 ratio (favoring immaturity) compared to healthy controls. ⁵ Yet, prior to the present study, how these perturbations in GABAergic signaling might contribute to the clinical manifestations of DS had not been directly explored.

Using an Scn1a +/− mouse model, Pizzamiglio et al demonstrate a stalled GABA switch in DS and suggest this delay contributes to neurodevelopmental impairment and social deficits, but not seizures/epilepsy. The authors provide evidence of delayed GABAergic maturation in Scn1a+/− mice, demonstrating decreased levels of hippocampal KCC2 protein compared to wild-type (WT) controls at postnatal day (P) 21 (ie, developing/juvenile mice). This delay manifests as an elevated ratio of spontaneous excitatory to inhibitory synaptic currents (E/I) and increased firing frequency in response to a GABA_A receptor agonist in CA3 pyramidal neurons of DS mice at P18-21, while WT neurons exhibit decreased firing. KU55933 (a kinase inhibitor that indirectly regulates KCC2 by inhibiting Ataxia Telangiectasia Mutated protein) rescues KCC2 protein levels and ameliorates the electrical abnormalities observed in CA3 pyramidal neurons. After treatment with KU55933, Scn1a+/− neurons no longer exhibit increased spike frequency in response to GABA_A receptor agonist, but do not show a net decrease in activity to approach WT controls. This suggests that dysfunction of GABA_A receptors themselves may also be involved in disease pathology. Consistent with this, reduced GABA sensitivity and altered GABA_A receptor subunit ratios have been reported in postmortem DS samples. ⁵

Pizzamiglio et al assess how the delayed GABA-switch contributes to DS pathophysiology by examining behavior and seizure phenotypes of Scn1a+/− male and female mice, with and without pharmacological intervention targeting NKCC1 and KCC2. Given that DS is diagnosed after symptom onset, the authors took a more translational approach and treated Scn1a+/− mice starting after P12 (rather than from birth). The authors conclude that neither KU nor bumetanide (an NKCC1 antagonist) is effective at treating temperature-induced or spontaneous seizures, or the premature mortality phenotype of Scn1a+/− mice. This finding is in contrast to human clinical trials reporting efficacy of bumetanide in treating neonatal seizures and temporal lobe epilepsy^7,8 as well as its success in delaying SUDEP in an Scn1b−/− mouse model of DS. ⁶ This suggests that the seizure/epilepsy phenotype of DS may be driven by another mechanism(s) which cannot be overcome solely by targeting the delayed GABA switch. However, both KU55933 and bumetanide mitigated the ASD-like behaviors observed only in male Scn1a+/− mice, which manifested as a decrease in social preference in a three-chamber social interaction test. Additionally, bumetanide improved spatial working memory (assessed by the Y-maze test) and decreased hyperactivity in DS mice. Together, these results support prior work showing that epilepsy and the behavioral/cognitive features of DS may be driven by distinct mechanisms and cell types. ⁹

Pizzamiglio et al further uncouple epilepsy from comorbidities in DS by investigating the phenotype of Scn1a+/− mice prior to seizure onset (P7-14). In a clever experiment, they demonstrate that Scn1a+/− mice are less successful at detecting their home-cage bedding than WT controls, consistent with other animal models of ASD. Additionally, they find that P11-12 CA1 pyramidal cells have reduced spine width and an elevated E/I ratio driven by decreased frequency of inhibitory postsynaptic currents. Since biochemical analysis and pharmacological intervention was restricted to one age (P15), it cannot be concluded definitively that the delayed GABA switch directly causes the behavioral and synaptic deficits observed in neonatal DS mice (P0-12). Additional mechanisms could be contributing to these phenotypes, such as disruptions to early excitatory neuron development and function reportedly caused by Nav1.1 haploinsufficiency. ¹⁰ Furthermore, since the NKCC1/KCC2 ratio was only examined at P21, it remains uncertain whether chloride homeostasis is already altered in neonatal Scn1a+/− mice. Implementing a broader temporal analysis would help establish the timing and causal contribution of the delayed GABA switch to nonseizure phenotypes. Such investigations could also provide insight into the most effective treatment window, as it is possible that early acute treatment may be sufficient to treat ASD-like behaviors. Alternatively, chronic therapy would be necessary if the GABA switch is not merely delayed but remains arrested in a permanently immature state.

Overall, this study supports the view that distinct mechanisms may underlie the diverse and complex clinical features that define DS, underscoring the existence of convergent and divergent causes of epilepsy and its co-morbidities in developmental and epileptic encephalopathies. The work by Pizzamiglio et al extends our knowledge as to the consequences of Nav1.1 loss-of-function beyond interneuron hypoexcitability per se to include a delayed switch in the sign of GABAergic signaling. Importantly, this study demonstrates that pharmacologic targeting of the switch (whether it be delayed or “stuck” in an immature position) can ameliorate cognitive and behavioral abnormalities in DS. Thus, a therapeutic “flip of the switch” controlling the direction of GABAergic signaling may serve as an effective intervention to treat nonseizure co-morbidities in DS.