Two Halves Make a Whole SCN1A

Abstract

Interneuron-Specific Dual-AAV SCN1A Gene Replacement Corrects Epileptic Phenotypes in Mouse Models of Dravet Syndrome

Mich JK, Ryu J, Wei AD, Gore BB, Guo R, Bard AM, Martinez RA, Luber EM, Liu J, Bishaw YM, Christian RJ. Sci Transl Med. 2025;17(790):p.eadn5603.

Dravet syndrome (DS) is a severe developmental epileptic encephalopathy marked by treatment-resistant seizures, developmental delay, intellectual disability, motor deficits, and a 10% to 20% rate of premature death. Most patients with DS harbor loss-of-function mutations in 1 copy of SCN1A, which encodes the voltage-gated sodium channel (NaV)1.1 alpha subunit and has been associated with inhibitory neuron dysfunction. Here, we generated a split-intein form of SCN1A and used a dual-vector delivery approach to circumvent adeno-associated virus (AAV) packaging limitations. In addition, we applied previously developed enhancer technology to produce an interneuron-specific gene replacement therapy for DS, called DLX2.0-SCN1A. The split-intein SCN1A vectors produced full-length NaV1.1 protein, and functional sodium channels were recorded in HEK293 cells in vitro. Administration of dual DLX2.0-SCN1A AAVs to wild-type mice produced full-length, reconstituted human protein by Western blot and telencephalic interneuron-specific and dose-dependent NaV1.1 expression by immunohistochemistry. These vectors also conferred strong dose-dependent protection against postnatal mortality and seizures in Scn1a_612/+;Meox2-Cre and Scn1a_+/R613X DS mouse models. Injection of single or dual DLX2.0-SCN1A AAVs into wild-type mice did not result in increased mortality, weight loss, or gliosis as measured by immunohistochemistry. In contrast, expression of SCN1A in all neurons driven by the human SYNAPSIN I promoter caused an adverse effect marked by increased mortality in the preweaning period, before disease onset. These findings demonstrate proof of concept that interneuron-specific AAV-mediated SCN1A gene replacement can rescue DS phenotypes in mouse models and suggest that it could be a therapeutic approach for patients with DS.

Commentary

Dravet syndrome (DS) is a devastating early-life epileptic encephalopathy characterized by febrile seizures, spontaneous seizures, developmental delay, intellectual disability, and an increased risk for premature mortality. Over 80% of patients with DS harbor a loss-of-function mutation in SCN1A, which encodes the voltage-gated sodium channel Na_v1.1. While most first-line treatments are ineffective or contraindicated for DS (e.g., sodium channel blockers), there are a few Food and Drug Administration-approved treatments for DS, including stiripentol, cannabidiol, and fenfluramine. Despite the progress that has been made, many patients still cannot achieve adequate seizure control and endure side effects from the prescribed medications. Thus, there remains an unmet need to identify disease-modifying treatments for DS that can improve multiple phenotypes with minimal side effects.

Loss-of-function Scn1a mouse models recapitulate many of the clinical features of DS, including increased susceptibility to hyperthermia-induced seizures, spontaneous seizures, behavioral deficits, and premature mortality.^1,2 Scn1a mouse models have been crucial for the identification of cell types that contribute to neuronal excitability in DS. Across multiple studies, it has been suggested that dysfunction of GABAergic interneurons, particularly the fast-spiking parvalbumin (PV) interneurons,^1,3,4 contribute to the phenotypes observed in DS, underscoring the possibility that PV interneurons might represent a critical therapeutic target.

As a proof-of-principle, it was previously shown that reactivation of the Scn1a gene rescues multiple phenotypes in a mouse model of DS, including reducing spontaneous seizure frequency and normalizing firing of PV interneurons and behavior. ⁵ Thus, the ideal treatment strategy for DS would be to restore SCN1A expression and function, which in turn, might reverse multiple clinical features. Scn1a mouse models have been pivotal for the recent advancements in gene therapy strategies for DS, including the use of antisense oligonucleotides (ASOs) ⁶ and adeno-associated viruses (AAVs).^7,8 Previous studies have utilized AAVs for delivery of a Scn1a-dCas9 activation system ⁷ and an engineered transcription factor that upregulates transcription of the endogenous SCN1A gene. ⁸ One of the challenges with the use of AAVs for the treatment of DS is that the human SCN1A open reading frame (ORF, >6 kb) exceeds the packaging capability of a traditional AAV (∼4.7 kb). To circumvent this limitation, a recent study used a high-capacity adenoviral vector to deliver a codon-optimized SCN1A sequence that was able to increase Scn1a mRNA levels. ⁹ Furthermore, Fadila et al., ¹⁰ demonstrated successful delivery of the SCN1A ORF using a canine adenovirus under the nonselective neuronal NSE promotor. As gene therapy strategies rapidly progress, this opens the possibility for targeting specific cell populations, such as GABAergic interneurons, that might confer greater therapeutic benefit in DS.

In the current study, Mich et al., ¹¹ generated a split-intein form of SCN1A and used a dual-viral vector delivery strategy, which bypassed the packaging limitations of AAVs. Briefly, split-intein peptides were added to each half of the SCN1A ORF sequence to mediate fusion of the halves, and ultimately, the generation of full-length SCN1A. To specifically target GABAergic interneurons, the split-intein fusion gene halves of SCN1A were packaged into AAV2/PHP.eB viral vectors with the hDLXI5/6i enhancer (referred to as DLX2.0). Wild-type (WT) mice at postnatal day 2 (P2) received bilateral intracerebroventricular (ICV) injections of both viral vectors. After 20 days, the authors observed over a 50% increase in levels of Na_v1.1 compared to WT mice that received control constructs. To test whether this gene replacement strategy would be efficacious in a mouse model of DS, neonatal Scn1a^fl/+; Meox2-Cre DS mice were injected with both viral vectors (DLX2.0-SCN1A-N + C AAVs). Compared to the premature mortality observed in the control mice at P70 (50% mortality in untreated mice and 37.5% mortality in mice that received an empty or single AAV), all treated Scn1a mutant mice survived beyond P70, and a subset of the mice (N = 11) lived until at least 1 year. ¹¹ Furthermore, approximately 50% of the treated Scn1a mutant mice were protected against hyperthermia-induced myoclonic jerks, and all but one mouse was protected against generalized tonic-clonic seizures up to 42°C (N = 15 of 16 mice). Moreover, the authors observed a significant reduction in interictal spikes and no spontaneous seizures in treated Scn1a mutant mice. ¹¹ The authors also generated an independent batch of the AAVs and similarly observed protection against hyperthermia-induced seizures in Scn1a^fl/+; Meox2-Cre DS mice. Moreover, a reduction in epileptiform activity and protection against hyperthermia-induced seizures was observed in two additional treated mouse models of DS (Scn1a^+/R613X and Scn1a^fl/+; Dlx5/6-Cre mice). ¹¹ Together, the authors demonstrate that their split-intein approach is able to increase SCN1A expression in GABAergic interneurons and provide robust protection against premature mortality and improve seizure phenotypes.

As a comparison for targeting GABAergic interneurons, Mich et al., ¹¹ also used the split-intein approach but with a nonselective neuronal promotor (hSyn1). Unexpectedly, when neonatal Scn1a^fl/+; Meox2-Cre DS pups were injected with hSyn1-SCN1A-N + C AAVs, the authors noted a dose-dependent increase in preweaning mortality compared to control mice. In the surviving mice, protection against hyperthermia-induced seizures was only observed with a high dose of hSyn1-SCN1A-N + C AAVs. ¹¹ Interestingly, Fadila et al., ¹⁰ showed that delivery of the SCN1A ORF using a canine adenovirus under the nonselective neuronal NSE promotor (CAV-SCN1A) was able to protect against hyperthermia-induced seizures, epileptiform activity, behavioral deficits, and hippocampal inhibition in a mouse model of DS. ¹⁰ One notable difference between the studies is that Fadila et al., ¹⁰ performed bilateral injections of CAV-SCN1A into the hippocampus and thalamus, while the current study performed bilateral ICV injections of hSyn1-SCN1A-N + C AAVs that mainly targeted the cortex. Thus, in the current study, it is possible that the use of AAVs versus an adenovirus yields different expression levels or patterns, and SCN1A expression was increased in other cell populations throughout the brain (e.g., excitatory neurons) which might have diminished the benefit of targeting only GABAergic interneurons.

In summary, Mich et al., ¹¹ developed a novel approach for SCN1A gene replacement and found that targeting GABAergic interneurons in multiple mouse models of DS protected against several clinically relevant phenotypes, including premature mortality and hyperthermia-induced seizures. ¹¹ Compared to previous studies using CRISPR technology ⁷ or ASOs, ⁶ the current study was able to successfully increase SCN1A expression without targeting the endogenous SCN1A alleles. In future studies, it would be important to establish whether the administration of DLX2.0-SCN1A-N + C AAVs at later stages of disease progression can still be beneficial. Given that patients with DS exhibit numerous nonseizure phenotypes, it would also be valuable to investigate whether DLX2.0-SCN1A-N + C AAVs can improve behavior. Fadila et al., ¹⁰ showed that administration of CAV-SCN1A in juvenile and adolescent DS mice (Scn1a^A1783/WT) restored spatial memory. ¹⁰ While it is possible that gene therapy alone may not fully address all clinical features of DS, combining gene therapy with pharmacological treatments could potentially offer a more comprehensive therapeutic strategy. Encouragingly, a recent study reported a significant reduction in seizure frequency in an infant with SCN2A developmental and epileptic encephalopathy treated with a combination of ASO therapy (Elsunersen) and sodium channel blockers, ¹² highlighting the potential of such combinatorial approaches. However, while that treatment regimen was able to appreciably reduce seizure frequency, no amelioration of nonseizure phenotypes, including head control and hypotonia, was observed. ¹² Nevertheless, the significant progress in gene therapy strategies underscores a growing momentum in the field and provides a strong foundation for future research aimed at maximizing the therapeutic benefits of gene therapy in these severe forms of pediatric epilepsy.