Ameliorating Seizures in Dravet Syndrome: A Review of Newly Approved and Investigational Drugs, RNA and Gene-Based Therapies
https://ror.org/04jr1s763grid.8404.80000 0004 1757 2304Department of Neuroscience, Pharmacology and Child Health, University of Florence, Florence, Italy
https://ror.org/04jr1s763grid.8404.80000 0004 1757 2304Neuroscience and Medical Genetics Department, Meyer Children’s Hospital IRCCS, University of Florence, Viale Pieraccini 24, 50139 Florence, Italy
https://ror.org/0370htr03grid.72163.310000 0004 0632 8656Research Department of Epilepsy, UCL Queen Square Institute of Neurology, London, UK
https://ror.org/05dbj6g52grid.410678.c0000 0000 9374 3516Department of Medicine, Epilepsy Research Centre, University of Melbourne, Austin Health, Melbourne, VIC Australia
https://ror.org/01ej9dk98grid.1008.90000 0001 2179 088XDepartment of Paediatrics, University of Melbourne, Royal Children’s Hospital, Melbourne, VIC Australia
https://ror.org/048fyec77grid.1058.c0000 0000 9442 535XMurdoch Children’s Research Institute and The Florey Institute, Melbourne, VIC Australia
Abstract
Dravet syndrome (DS) is a severe infantile-onset developmental and epileptic encephalopathy characterised by frequent drug-resistant seizures, cognitive and motor impairment, and an elevated risk of premature mortality, particularly from sudden unexpected death in epilepsy (SUDEP). Over 95% of cases are caused by loss-of-function pathogenic variants in SCN1A, the gene encoding the voltage-gated sodium channel alpha-1 subunit, NaV1.1. Despite advances in the understanding of DS pathophysiology, current treatments have primarily targeted seizure control without addressing the myriad of associated morbidities and the underlying molecular defect. Recent therapeutic developments include the approval of new antiseizure medications such as fenfluramine and pharmaceutical-grade cannabidiol, which have demonstrated efficacy and tolerability in randomised placebo-controlled trials. Investigational therapies, including selective serotonin receptor modulators and RNA-based and gene-based strategies, are expanding the treatment landscape. Antisense oligonucleotides (e.g., STK-001) aim to restore SCN1A expression, while emerging gene therapy approaches, including engineered AAV vectors (e.g., ETX101) and CRISPR-mediated transcriptional activation, seek to directly modify disease biology. Although challenges remain, such as long-term safety concerns, a broad spectrum of phenotypic severity, and efficacy and accessibility of advanced therapies, rapid progress in therapeutic research for DS offers new hope. Future directions include defining the most effective genetic therapies to improve outcomes, and ideally cure, all features of DS; optimal timing to deliver interventions as well as benefits derived from administration at later ages; ideal combinations of therapies; comparison of outcomes of targeted therapies with natural history studies and biomarker development. Together, these advances signal a paradigm shift in epilepsy management from symptomatic treatment to precision medicine for children and adults with DS.
Key Points
| New treatments such as fenfluramine have improved seizure control and quality of life for people living with Dravet syndrome. |
| Research into RNA-based and gene-based therapies is moving beyond seizure control, with early clinical data suggesting potential benefits on cognition, behaviour and adaptive functioning, in addition to seizure reduction. |
| Early genetic diagnosis and timely access to mechanism-based therapies may improve long-term seizure and non-seizure outcomes in children with Dravet syndrome, although this remains to be confirmed in prospective studies. |
Introduction
Dravet syndrome (DS) is an infantile‐onset developmental and epileptic encephalopathy (DEE) characterised by febrile and afebrile seizures, cognitive and motor impairment, and an increased mortality risk, largely due to sudden unexpected death in epilepsy (SUDEP) [1–4]. DS presents by 19 months of age [2], with an incidence of 1 per 15,500 [5] or 6.5 per 100,000 live births [6]. The clinical burden is profound, necessitating lifelong care and multidisciplinary support [7].
DS is genetically relatively homogeneous, with over 95% of patients carrying de novo pathogenic variants in SCN1A, which encodes the voltage-gated sodium channel (VGSC) α subunit Nav1.1 [2]. Although most variants arise de novo, parental germline or somatic mosaicism can occur and confers an elevated recurrence risk, with important implications for genetic counselling and family planning [8, 9]. SCN1A variants include predicted loss-of-function (LOF) microdeletions, missense and truncating variants, resulting in Nav1.1 haploinsufficiency. Studies in Scn1a knockout mouse models showed that reduced Na+-current density results in decreased excitability of parvalbumin- and somatostatin-positive inhibitory interneurons, leading to impaired inhibitory control and promoting epileptogenesis [10–12]. However, findings in certain mouse strains and developmental stages suggest that excitatory neurons may also contribute significantly to hyperexcitability [13]. Research using human-induced pluripotent stem cell (hiPSC)-derived neurons has produced conflicting data. Some studies have reported Na+-current deficits in inhibitory neurons [14, 15], while others have observed impairment across both excitatory and inhibitory populations [16, 17]. Recent studies have highlighted the prenatal impact of aberrant SCN1A, resulting in abnormal cell cycle progression, enhanced neurogenesis and abnormal lamination of excitatory pyramidal neurons, showing deficits beyond action potential generation [18].
Currently approved therapies for DS are primarily symptomatic and do not target the underlying molecular disease-causing process. Recent advances have led to the approval of several novel antiseizure medications (ASMs) and to the development of RNA-based and gene-based therapies that directly target the underlying SCN1A haploinsufficiency and associated network dysfunction.
Increasing evidence underscores the importance of early and accurate genetic diagnosis in DS. Early identification enables avoidance of contraindicated sodium-channel-blocking ASMs, facilitates earlier access to optimal therapies, and may allow enrolment in potentially disease-modifying mechanism-based trials. Natural history studies, including BUTTERFLY and ENVISION [19, 20], demonstrate that a higher seizure burden early in life and delays in initiating effective treatment are associated with poorer long-term cognitive and adaptive outcomes. These observations, together with the appropriate increase in focus on non-seizure outcomes [21], highlight the potential for earlier intervention to improve all aspects of this multimorbidity disease.
This narrative review critically examines the clinical efficacy, safety, and pharmacological profiles of newly approved pharmacological treatments and emerging RNA-based and gene-based approaches that share the common goal of modulating disease biology rather than addressing seizures alone. We conducted a comprehensive literature search of PubMed, Embase, and the Cochrane Library for peer-reviewed articles, systematic reviews, and clinical trials published between January 1, 2017, and March 31, 2025. We used MeSH terms such as ‘Epilepsies, Myoclonic’ and ‘Nav1.1 voltage-gated sodium channel’ and subheadings such as ‘Dravet Syndrome,’ alongside free-text terms, including ‘treatment,’ ‘therapy’ and ‘clinical trials.’ Non-English studies were included to minimise language bias. We also screened the reference lists of included articles. We assessed studies for relevance, quality, and scope, and excluded duplicates or low-quality studies. A similar search for completed or ongoing interventional studies in Dravet syndrome was conducted in the ClinicalTrials.gov website. The final reference list was curated based on relevance to the themes addressed in this review.
Standard Treatment
In 2022, an international expert consensus panel developed a protocol of the order of ASM selection for patients with DS based on randomised controlled trial (RCT) data and ASM accessibility [22]. Among traditional ASMs, valproate was recommended as the first-line treatment, although its efficacy has not been confirmed through dedicated clinical trials. Second-line ASMs included clobazam, fenfluramine and stiripentol. Most importantly, the consensus statement emphasised that sodium channel blocking ASMs, such as carbamazepine, oxcarbazepine and lamotrigine are generally contraindicated for chronic use because they exacerbate seizures and may worsen developmental outcomes in DS [22]. Rare reports describe individuals in whom lamotrigine or long-term phenytoin appeared beneficial [23, 24], but these small series (n ≤ 4) must be interpreted with caution and do not justify routine use of these drugs as maintenance therapy. By contrast, intravenous phenytoin remains widely used for acute management of convulsive status epilepticus in DS when benzodiazepines and other first-line rescue medications are insufficient, in line with broader status epilepticus guidelines [23, 24].
In routine practice, most patients with DS receive valproate as first-line therapy (typically 20–40 mg/kg/day), titrated according to efficacy and tolerability. Where seizure control remains suboptimal, clobazam (0.25–1.0 mg/kg/day) is commonly added, followed by stiripentol (up to 50 mg/kg/day) in accordance with regulatory labelling [22, 25–27]. Observational and randomised data suggest that these core treatments together achieve ≥50% reduction in convulsive seizures in approximately 50–70% of patients, although complete seizure freedom remains rare [22, 28]. Polypharmacy is the rule; most individuals ultimately require three or more ASMs, and rational combinations that avoid sodium channel blockers and consider pharmacokinetic interactions are central to care [22, 28]. Additional ASMs used in practice include topiramate (5–10 mg/kg/day), ketogenic diet therapy, and clonazepam for intermittent or adjunctive use [22].
In the first randomised, double-blind, placebo-controlled phase III trial dedicated to DS [25–27], 41 children were randomised to stiripentol or placebo. This trial found a 71% (15/21 children) responder rate (defined as 50% seizure reduction) compared with 5% (1/20) children on placebo. Further studies highlighted the reduction in status epilepticus associated with stiripentol use [29, 30]. Stiripentol received its initial approval in Europe in 2007. It was later licensed in other countries, including Canada, Japan, and the United States in 2018. Of note, the European licence mandates co-administration with both valproate and clobazam, whereas in the United States, only clobazam is required. Although the precise mechanism of action of stiripentol remains incompletely understood, it is believed to enhance inhibitory γ-aminobutyric acid (GABA)-mediated neurotransmission, exert neuroprotective effects, and inhibit lactate dehydrogenase (LDH), thereby modulating neuronal metabolism [31–33]. Stiripentol also inhibits the metabolism of concurrent ASMs by targeting cytochrome P450 enzymes, specifically CYP1A2, CYP3A4 and CYP2C19 [31]. Common adverse effects observed in clinical trials include anorexia, weight loss, and sedation. In clinical practice, sedation can often be mitigated by proactive reduction of concomitant clobazam, and in some cases valproate, once stiripentol has been introduced and seizure control stabilised. Emerging evidence supports the early initiation of stiripentol before 2 years of age as both effective and well tolerated in preventing status epilepticus in patients with DS [30]. Long-term treatment has also shown sustained efficacy and safety [34].
In clinical practice, polypharmacy in DS requires careful sequencing and dose optimisation. International expert consensus recommends initiating valproate and then adding clobazam, followed by stiripentol or fenfluramine as second-line therapies, with cannabidiol generally reserved for subsequent add-on use. The consensus also suggests specific ASMs based on seizure-type specific efficacy [22]. Doses of concomitant ASMs, particularly clobazam and valproate, often require adjustment when stiripentol or cannabidiol is introduced due to pharmacokinetic interactions. Rational stepwise escalation, prioritising agents with demonstrated efficacy for convulsive seizures, is essential to minimise drug burden while achieving effective seizure control and reducing the risk of SUDEP.
Newly Approved Drugs
More recently, two new ASMs have been approved for DS after randomised double-blind placebo-controlled trials—fenfluramine and cannabidiol.
Fenfluramine
Fenfluramine has been approved for the treatment of seizures associated with DS in patients aged 2 years and older in the United States, Europe, United Kingdom, Israel, Japan and Australia. Its approval followed the demonstration of efficacy and tolerability in three randomised, double-blind, placebo-controlled phase III trials involving children and adolescents aged 2–18 years; all trials counted convulsive seizures (generalised tonic–clonic, tonic, clonic and focal to bilateral tonic–clonic seizures) as the primary endpoint [35–37]. Across these trials, fenfluramine produced placebo-adjusted median reductions in monthly convulsive seizure frequency of 34–72%, with 50% responder rates (≥ 50% reduction) in the range of ~ 41–73% depending on dose, background regimen and study design. Around 21–54% of patients achieved ≥ 75% reduction in monthly seizures and up to ~ 25% experienced near seizure freedom (≤ 1 convulsive seizure over 14 weeks) [35–37]. Real-world data and early-access program data confirmed fenfluramine’s durable seizure reduction [38–40]. Up to 50% of patients were able to reduce their overall ASM load, and treatment was associated with good tolerability and reduced healthcare resource utilisation [38–41]. Post-hoc analyses demonstrated that efficacy was consistent regardless of concomitant stiripentol use or age group (< 6 years old vs ≥ 6 years old).
Beyond seizure reduction, fenfluramine has demonstrated clinically meaningful improvements in non-seizure outcomes. Post-hoc analyses and real-world studies report improvements in executive functioning, behaviour, sleep quality, and caregiver-reported quality of life, including domains such as attention, emotional regulation, and daily functioning [38–42]. These findings suggest that reductions in convulsive seizure burden, particularly tonic–clonic seizures, may translate into broader functional gains for individuals with DS. Fenfluramine was generally well tolerated. The most common treatment-emergent adverse events were decreased appetite, weight loss, fatigue, nasopharyngitis, seizures, and pyrexia [35–38, 40, 43]. Small declines in weight and BMI were observed in a minority of patients, although most children remained on expected growth trajectories [44]. Because of historical associations between high-dose fenfluramine for management of obesity and the development of valvular heart disease and pulmonary arterial hypertension, patients starting fenfluramine undergo baseline echocardiography and serial echocardiograms every 6 months initially and then annually, and 3–6 months after discontinuation, in line with risk-minimisation requirements. No valvular heart disease or pulmonary arterial hypertension has been reported in DS trials [43].
Fenfluramine is thought to act by enhancing serotonin release and signaling through 5‐HT1D, 5-HT2A and 5-HT2C receptors, and by modulating sigma‐1 receptors [45]. Given its serotonergic activity, caution is advised when co-administering serotonergic antipsychotics or antidepressants. In such cases, careful observation during treatment initiation and dose increases of fenfluramine may be necessary. Alternatively, switching to a less serotonergic antipsychotic, such as from risperidone to aripiprazole, may be appropriate [46]. A study (ORCHID, ClinicalTrials.gov identifier: NCT06118255) is currently evaluating safety, tolerability and pharmacokinetics of fenfluramine in infants aged 1 year to <2 years, with promising reports from real-world data [47].
Cannabidiol
Pharmaceutical-grade cannabidiol was approved by the FDA in June 2018 and by the EMA in 2019 (in the latter case, with required co-use of clobazam) for seizures associated with DS. Although its precise mechanism of action is unclear, it does not significantly activate the cannabinoid receptors, CB1 or CB2. It is believed to work through GABAergic modulation, intracellular calcium regulation (via TRPV, GPR-55 or VDAC), and anti-inflammatory effects (via adenosine signaling) [48]. The two pivotal randomised, double-blind, placebo-controlled trials targeted convulsive seizures as the primary endpoint, while non-convulsive seizures were monitored as secondary outcomes and demonstrated more variable responses [49, 50]. At a dosage of 20 mg/kg/day, adjunctive pharmaceutical-grade cannabidiol achieved median reductions in convulsive seizure frequency of 39–49% versus 13–27% with placebo, with 50% responder rates of 43–49%. Around 18–30% patients achieved ≥ 75% reduction in monthly convulsive seizures, and up to ~ 5% of cannabidiol-treated patients became free of convulsive seizures during the 14-week blinded period [49, 50]. In long-term open-label extension studies, sustained ≥ 50% reductions in convulsive seizures were observed for ≥ 2–3 years in roughly 40–50% of participants [51, 52]. Open-label extension data also indicate improvements in alertness, engagement, adaptive functioning and caregiver burden in some patients receiving cannabidiol [51, 52]. Although uncontrolled, these observations align with clinical experience that cannabidiol may enhance quality of life in addition to reducing seizure frequency.
Cannabidiol was generally well tolerated. The most common adverse effects included diarrhoea, vomiting, decreased appetite, fatigue and somnolence. Discontinuation due to side effects was low in the open-label extension studies [49–52]..Most adverse events associated with cannabidiol resolved within the 14-week study period [53]. Some patients (up to 19.7%) experienced a 3-fold or greater elevation in transaminases, which is more commonly seen with concurrent valproic acid. However, these elevations resolved without discontinuation of cannabidiol and were not associated with overt liver dysfunction. Small decreases in weight or BMI have been reported but are typically reversible and do not usually disrupt growth trajectories [54].
No direct comparative trials exist for stiripentol, fenfluramine and cannabidiol directly.
A network meta-analysis of randomised controlled trial data for stiripentol, fenfluramine and cannabidiol found that stiripentol and fenfluramine were more effective than cannabidiol for reducing convulsive seizures, with stiripentol having a potentially lower risk of discontinuation due to adverse events [55].
When comparing tolerability across approved therapies, several practical considerations arise in using combinations of ASMs. Fenfluramine, stiripentol and cannabidiol may reduce appetite or weight, whereas clobazam/clonazepam, stiripentol and cannabidiol may contribute to sedation. In particular, as the dose of cannabidiol increases, clobazam or clonazepam may require dose reduction if the patient becomes sedated. Behavioural worsening has been reported with topiramate and benzodiazepines in some individuals with DS, which can influence drug selection and sequencing [22]. These factors, combined with known pharmacokinetic interactions, underscore the need for careful titration and rational stepwise escalation of therapy. Levetiracetam is sometimes used but has not been shown to be effective in DS and is associated with frequent behavioural adverse effects.
Several important drug–drug interactions have been identified. The combination of cannabidiol and clobazam leads to a bidirectional pharmacokinetic interaction whereby cannabidiol inhibits CYP2C19, increasing plasma concentrations of N-desmethylclobazam, while clobazam increases exposure to the active cannabidiol metabolite 7-OH-CBD [56, 57]. In clinical trials, sedation and somnolence were markedly more frequent in patients receiving both drugs [49–52]. Therefore, if the dose of clobazam is high, one should consider gently reducing it when commencing cannabidiol or, alternatively, carefully instructing the carers to do so as soon as the patient becomes sedated as the dose of cannabidiol increases, with further titration guided by clinical response. Co-administration of cannabidiol with valproate is well known to cause an increase in transaminases; however, this does not result from a pharmacokinetic interaction altering valproate or cannabidiol concentrations [57] but is likely due to a pharmacodynamic interaction at the mitochondrial level. Fenfluramine plasma concentrations are elevated in patients treated with concomitant stiripentol, resulting in a higher risk of fatigue and somnolence [58]. Therefore, when fenfluramine is combined with stiripentol, a reduction in the maximum fenfluramine dose is recommended, from 0.7 mg/kg/day (maximum 26 mg/day) without stiripentol to 0.4 mg/kg/day (maximum 17 mg/day) on stiripentol. No dose adjustment is necessary when fenfluramine is co-administered with cannabidiol.
Investigational Drugs
A number of investigational therapies targeting serotonergic signalling, glutamatergic transmission, cholesterol metabolism and calcium-channel function are being explored in DS. Most are being studied primarily in broader DEE populations, with relatively small DS subgroups and limited long-term data (Table 1). In this section, we focus on agents with specific DS data, distinguishing preclinical from clinical evidence and emphasising the current status of development.
Drug Mechanism Study type No. of DS patients Key outcome Development status Lorcaserin 5-HT2C agonist Compassionate use + retrospective series 5 + 20 ~ 47% median motor seizure reduction; appetite/weight loss common Development discontinued due to safety signals Verapamil L-type calcium channel blocker Open-label pilot 4 3/4 with 50–99% seizure reduction No ongoing development Clemizole 5-HT2A/5-HT2B modulation Zebrafish model Preclinical
Serotonergic Modulators
Among serotonergic modulators, several repurposed or novel agents are being evaluated in DS and other DEEs. These include 5-HT2C-selective agonists (bexicaserin, lorcaserin) and agents with broader serotonin receptor activity (clemizole). Their development has been driven by preclinical evidence that serotonin signalling can modulate seizure susceptibility in SCN1A-related epilepsies.
Bexicaserin (LP352)
Bexicaserin (LP352) is a selective 5-HT2C receptor superagonist designed to preferentially activate 5-HT2C and not affecting 5-HT2A and 5-HT2B receptors, thereby exhibiting antiseizure efficacy without the cardiac valve toxicity associated with non-selective serotonergic agents such as high-dose fenfluramine [59]. It is being developed for seizures associated with DEEs, including DS.
In the PACIFIC phase Ib/IIa study (NCT05364021), a randomised, double-blind, placebo-controlled trial primarily examining safety and tolerability, bexicaserin was evaluated in 53 adolescents and adults with DEEs, aged 12–65 years, including four participants with DS [60]. After a 15-day titration, patients entered a 60-day maintenance phase, followed by an open-label extension (NCT05626634). In the DS subgroup (n = 4; all allocated to bexicaserin), the median reduction in countable motor seizure frequency was approximately 75%, consistent with the robust antiseizure effect seen in the overall DEE cohort. Although this subgroup is too small to draw firm conclusions, it provides preliminary support for 5-HT2C modulation as a potential therapeutic strategy in DS.
Safety and tolerability data derive from the broader DEE population. In the PACIFIC study, treatment-emergent adverse events occurred in a similar proportion of bexicaserin- and placebo-treated participants (approximately 80–90%), were mostly mild to moderate, and led to relatively few discontinuations. The most common adverse events with bexicaserin were somnolence, decreased appetite and constipation. Serious treatment-emergent adverse events were infrequent and occurred only in bexicaserin-treated participants (ankle fracture, constipation and increased seizures). No deaths were reported and echocardiographic monitoring was not required [60].
All eligible participants in the PACIFIC study elected to continue in the open-label extension (NCT05626634), with the option of subsequent expanded access (NCT06149663). A global phase III RCT has been initiated to evaluate the efficacy, safety and tolerability of bexicaserin in paediatric, adolescent and adult patients with DEEs (DEEp-OCEAN) and with DS (DEEp-SEA).
Clemizole (EPX-100)
Clemizole is a first-generation histamine H1 receptor antagonist that was identified in a zebrafish chemical screen as a potent suppressor of behavioural and electrographic seizures in Scn1a mutant larvae [46]. Subsequent work suggested that its antiseizure effects were primarily mediated via serotonin receptors, especially 5-HT2A and/or 5-HT2B, rather than through classical antihistamine mechanisms [61]. Rapid metabolism in mice (plasma half-life <10 minutes compared with ~ 3.4 hours in humans) has limited detailed rodent pharmacodynamic studies, but the zebrafish data provide a strong preclinical rationale for clinical translation.
EPX-100, a pharmaceutical formulation of clemizole, is being evaluated as adjunctive therapy in DS in the ARGUS trial (NCT04462770), a global, multicentre, randomised, double-blind, placebo-controlled, phase III study. The trial includes a 4-week observational baseline, a 16-week treatment phase and a long-term open-label extension of up to 3 years. Primary endpoints focus on changes in convulsive seizure frequency; secondary outcomes include global clinical impressions and safety. Detailed efficacy and safety data have not yet been reported. Given its serotonergic profile (including 5-HT2B engagement), long-term cardiac safety monitoring in larger cohorts will be essential.
Lorcaserin
Lorcaserin is a selective serotonin 5-HT2C receptor agonist with 8-fold to 94-fold greater affinity for 5-HT2C over 5-HT2A, and 46-fold to 168-fold greater affinity over 5-HT2B [47]. It was approved as a weight-loss medication, but was withdrawn from the market in 2020 after the CAMELLIA-TIMI 61 trial identified a small but statistically significant increase in cancer incidence compared with placebo [62]. Consequently, lorcaserin is no longer being developed as a commercial therapy, including for epilepsy.
Despite this, lorcaserin has provided useful proof-of-concept data for serotonergic modulation in DS. In a compassionate-use protocol involving five children with DS, add-on lorcaserin reduced seizures in all patients. Three had marked improvements: one became seizure-free for 3 weeks, another showed > 90% reduction in generalised tonic-clonic seizures and a third had a substantial reduction in atonic seizures (from 12 per hour to occasional seizure-free days). Over time, seizure control waned in some patients, and decreased appetite prompted discontinuation in three of five individuals after 9–14 months; two continued treatment without notable adverse effects [61].
A larger retrospective multicentre study of 36 patients with severe childhood-onset epilepsies, including 20 with DS, reported a median 47% reduction in motor seizures with lorcaserin, with 42% achieving > 50% reduction [63]. Decreased appetite and weight loss occurred in approximately 11% of patients, and five discontinued due to adverse events [63]. However, these data are uncontrolled and subject to multiple biases.
The MOMENTUM 1 trial (NCT04572243), a planned randomised, placebo-controlled study of lorcaserin in DS, was terminated early due to recruitment challenges following the global withdrawal of lorcaserin. MOMENTUM 2 (NCT04457687), an observational study of longer-term outcomes in patients who remained on lorcaserin after market withdrawal, is ongoing. Given the safety concerns and lack of active development, lorcaserin should now be regarded as a mechanistic probe rather than a future therapeutic option.
Cholesterol 24-Hydroxylase Inhibition: Soticlestat (TAK-935)
Beyond serotonergic pathways, modulation of glutamatergic signalling through inhibition of cholesterol 24-hydroxylase (CH24H) has emerged as a mechanistically distinct strategy. Soticlestat is the most advanced compound in this category, supported by strong preclinical efficacy and mixed but informative clinical trial results. Soticlestat is a potent, selective inhibitor of cholesterol 24-hydroxylase (CH24H), the brain-specific enzyme that converts cholesterol to 24S-hydroxycholesterol [64]. By lowering 24S-hydroxycholesterol levels, soticlestat is thought to stabilise lipid rafts and enhance glutamate reuptake, thereby reducing extracellular glutamate and dampening excessive excitatory neurotransmission; reduced modulation of NMDA receptors by 24S-hydroxycholesterol may further contribute to its antiseizure effects [64].
In a Scn1a+/- mouse model of DS, soticlestat reduced spontaneous and hyperthermia-induced seizures and prevented premature death consistent with SUDEP, with no deaths observed in treated animals during the observation period, whereas a substantial proportion of untreated mice died early [65]. These preclinical results strongly supported CH24H inhibition as a disease-relevant strategy in DS.
Clinically, the phase II ELEKTRA trial evaluated soticlestat as adjunctive therapy in 141 children with DS or Lennox–Gastaut syndrome (LGS), including 51 with DS [66]. In the combined DS/LGS cohort, soticlestat achieved a significant placebo-adjusted median reduction of 30–50% in motor seizure frequency over the maintenance period; in the DS subgroup, the placebo-adjusted median reduction in convulsive seizure frequency was approximately 46% [66]. Treatment-emergent adverse events were generally similar between soticlestat and placebo, with most being mild or moderate; lethargy and constipation were more common with soticlestat [66].
The phase III SKYLINE trial (NCT04940624) in 144 children and young adults with DS (aged 2–21 years) investigated soticlestat’s efficacy as adjunctive therapy. In June 2024, Takeda announced that SKYLINE narrowly missed its primary endpoint of reduction in convulsive seizure frequency, although clinically meaningful improvements were observed in several key secondary endpoints, including responder rates, caregiver and clinician global impressions of change, and seizure intensity and duration (all p ≤ 0.008). Despite these secondary findings, the company subsequently indicated that it would not pursue further development of soticlestat for epilepsy, and the molecule is no longer being advanced as a DS therapy. Nevertheless, soticlestat provides important mechanistic insight into the potential of targeting brain cholesterol metabolism and glutamatergic signalling in DS.
Calcium Channel Modulation: Verapamil
Interest has also arisen in agents targeting calcium-channel function or neurovascular mechanisms, although available clinical evidence is sparse and derives mainly from small uncontrolled series. Verapamil is a phenylalkylamine L-type voltage-gated calcium channel blocker widely used in cardiology for hypertension, angina, paroxysmal supraventricular tachycardia and atrial fibrillation. Interest in verapamil for DS arose from evidence that it may modulate blood–brain barrier function and neuroinflammation, as well as calcium-dependent neuronal excitability.
In a small pilot study including four children with DS and drug-resistant seizures, adjunctive verapamil was associated with partial seizure reduction (50–99%) in three of four patients over a 14-month period [67]. The drug was generally well tolerated, but the sample size was extremely small and the study uncontrolled. A phase II trial of verapamil in children and young adults with DS (NCT01607073) was completed in 2015, but the results have not been published, raising uncertainty about its efficacy and safety in a larger cohort.
Given the limited and largely anecdotal evidence, verapamil cannot currently be recommended as a routine treatment for DS. At most, it may be considered on an individual off-label basis in highly refractory cases, with careful cardiac and clinical monitoring. Its main value at present is as a proof-of-concept that modulation of calcium-channel signalling and vascular/blood–brain barrier function might influence seizure burden in DS and could inform future drug development.
RNA Therapies
Alternatively spliced exons, when included in a transcript, may introduce a premature termination codon, usually targeting the transcript for nonsense-mediated decay [68]. The role of noncoding variation in DS has been highlighted by the identification of pathogenic variants in an alternatively spliced exon introducing a premature termination codon in intron 20 (termed exon 20N) of SCN1A, which results in reduced levels of full-length SCN1A protein [69]. Subsequently, 20N was developed as a therapeutic target for antisense oligonucleotides (ASOs), leading to targeted augmentation of nuclear gene output (TANGO). This approach increased production of the translated voltage-gated sodium channel subunit NaV1.1 by reducing unproductive splicing of 20N in a DS mouse model [70, 71].
STK-001 (Zorevunersen)
The lead ASO therapy, STK-001 (also known as zorevunersen), is currently under investigation as a potential disease-modifying treatment. Two open-label phase I/IIa studies —MONARCH (NCT04442295, USA) and ADMIRAL (EudraCT 2020–006016-24, UK)—evaluated the safety, tolerability, and pharmacokinetics of STK-001 after single and multiple intrathecal doses in children and adolescents aged 2–18 years with DS. Across both studies, 81 patients received at least one dose of STK-001. These patients were highly refractory to treatment; 85% were on three or more ASMs and 54% were on four or more. About half were also taking fenfluramine. The median baseline convulsive seizure frequency among evaluable patients was 17 per 28 days (n = 77). Overall, ~30% of patients experienced treatment-emergent adverse events considered related to the drug. The most common were asymptomatic cerebrospinal fluid (CSF) protein elevation and procedural vomiting. Drug-related CSF protein elevation was reported as a treatment-emergent adverse event in 11/81 (~14%) participants in MONARCH/ADMIRAL, and in ~27% (20/74) of those in the SWALLOWTAIL/LONGWING open-label extensions [72]. When all CSF measurements were reviewed irrespective of attribution, transient elevations >50 mg/dL occurred in approximately three-quarters of patients in the extension studies, but these were not associated with clinical symptoms and rarely led to treatment discontinuation (1 patient). Serious adverse events occurred in 22% of patients but were largely unrelated to zorevunersen. Notably, however, one patient in the 70-mg cohort experienced a suspected unexpected serious adverse reaction (SUSAR) attributed to zorevunersen. Three patients experienced serious events related to CSF collection or administration procedures. One patient died from SUDEP, although this was deemed unrelated to the treatment [73]. The SWALLOWTAIL phase II open-label extension trial (NCT04740476) is ongoing, with some patients having received up to ten doses of zorevunersen without new safety signals.
A global, multicentre, randomised, double-blind, sham-controlled, phase III study (EMPEROR, NCT06872125) is now underway to evaluate zorevunersen’s efficacy, safety, and tolerability. The primary endpoint is the change from baseline in major motor seizure frequency, while secondary endpoints include behaviour, cognition, clinical status, and health-related quality of life [74]. Pharmacokinetic interactions between zorevunersen and drugs metabolised by CYP enzymes are considered unlikely, although formal drug–drug interaction studies have not yet been conducted. Emerging findings from MONARCH, ADMIRAL, SWALLOWTAIL and LONGWING also indicate potential benefits beyond seizure reduction. Caregiver- and clinician-reported measures describe improvements in behaviour, communication, sleep and overall clinical status in a subset of participants treated with STK-001. These early signals require confirmation in the EMPEROR RCT, but they support the hypothesis that RNA-based therapies may address a broader spectrum of DS-related morbidity [72–74].
ASO-84
ASO-84, a surrogate for STK-001 targeting splicing of SCN1A exon 20N, was tested in Scn1a+/- Dravet syndrome mouse models. A single intracerebroventricular injection on postnatal Day 2 restored interneuron function. In untreated DS mice, cortical parvalbumin-positive interneurons showed biphasic excitability (initial hyperexcitability followed by hypoexcitability and depolarisation block), along with reduced sodium current density and diminished GABAergic signaling to cortical pyramidal neurons. ASO-84 treatment restored action potential firing, sodium current density and GABAergic signalling in parvalbumin-positive interneurons, suggesting that interneuron excitability was selectively affected by ASO-84 [75].
Splice-Modulating Antisense Oligonucleotides Targeting Poison Exons in Introns 1 and 22 of SCN1A
Two additional alternatively spliced exons introducing a premature termination codon subject to nonsense-mediated decay in introns 1 and 22 of SCN1A have been identified through deep-coverage long-read sequencing. Splice-modulating ASOs targeting these exons successfully reduced aberrant exon inclusion in patient-derived induced pluripotent stem cells (iPSC) neurons, achieving substantial allele specificity [76].
Ataluren
Ataluren is a small molecule (C₁₅H₉FN₂O₃; a 1,2,4-oxadiazole linked to fluorobenzene and benzoic acid) that promotes read‐through of premature stop codons, enabling production of full-length, functionally active proteins without altering normal stop codons or affecting mRNA stability [77]. Ataluren does not read‐through normal stop codons. Ataluren is selective for translation and does not alter levels of mRNA with premature stop codons or wild‐type mRNA; that is, it does not modify transcription or mRNA stability. Ataluren has a favorable safety profile and is approved for Duchenne muscular dystrophy caused by nonsense mutations. In DS, a single‐centre, double‐blind, placebo‐controlled crossover trial (NCT02758626) randomised seven subjects to ataluren or placebo. The study did not detect significant differences in seizure frequency, cognitive, motor, behavioural function, or quality of life. However, limitations including a small sample size and a short (12‐week) treatment duration may have prevented identification of potential disease‐modifying effects [78].
Regulatory RNAs (regRNAs) are non-coding RNAs that regulate the expression of protein-coding genes. ASO-based therapies targeting regRNA aim to upregulate endogenous SCN1A expression. In one study, an Scn8a-targeted ASO was administered to a Scn1a deletion mouse model of DS (haploinsufficient for SCN1A), in which untreated mice typically exhibit spontaneous seizures and 50% mortality by 4 weeks of age [79]. A single postnatal day 2 (P2) dose of Scn8a ASO resulted in a 50% reduction of Scn8a transcripts (without affecting Scn1a) and resulted in treated DS mice surviving beyond 5 months without behavioural or electrographic seizures [80, 81].
Gene-Based Therapies
Gene-based therapeutic strategies for DS aim to restore functional NaV1.1 expression and inhibitory interneuron activity by targeting the underlying SCN1A haploinsufficiency. Broadly, these approaches fall into two categories: gene-replacement strategies, which deliver functional SCN1A coding sequences or fragments, and gene-regulation strategies, which enhance expression of the endogenous SCN1A allele through transcriptional activation. Although differing in mechanism and delivery, these approaches converge on the common goal of normalising interneuron excitability and network inhibition.
Preclinical Gene-Replacement Strategies
Preclinical gene-replacement approaches have focused on overcoming the size constraints of adeno-associated virus (AAV) vectors, which cannot accommodate the full SCN1A coding sequence. As interneurons remain the critical pathological cell population in DS, they represent a promising target for cell class-specific therapies. Enhancers have been demonstrated to enable cell type-selective gene expression using AAVs, including specific expression in interneurons [82]. This technology could allow interneuron-specific delivery of a therapeutic transgene for DS. However, the AAV platform faces limitations due to genome size constraints (~4.7 kb), which preclude packaging the full human SCN1A open reading frame (ORF; >6.0 kb) into a single vector. To overcome this, Mich et al. [83] developed a split-intein version of SCN1A, employing a dual-vector delivery approach. By combining this strategy with interneuron-specific enhancers, the dual-AAV system was tested in two mouse models of DS, demonstrating improved postnatal survival and reductions in both febrile and spontaneous seizures [83].
Gene-Regulation Strategies
Gene-regulation strategies aim to augment expression of the intact wild type endogenous SCN1A allele rather than replacing the gene itself. In preclinical models, dead Cas9 (dCas9) fused to transcriptional activators such as VP64 and VPR enables precise upregulation of gene expression without altering DNA sequences. In DS models, dCas9 fused to transcriptional activators like VP64 and VPR to upregulate Scn1a expression in GABAergic neurons demonstrated improvement in seizure control, behaviour, and survival [84, 85]. Another study investigated the viral vector-mediated delivery of a codon-modified SCN1A open reading frame into the brain, showing improvement of DS-associated comorbidities in juvenile and adolescent Scn1a A1783V/WT mice [86]. Key findings included increased survival, reduction in epileptic spikes, protection from thermally induced seizures, normalisation of electrocorticographic background activity, amelioration of behavioural deficits, and restoration of hippocampal inhibition. Delivery involved bilateral injections into the hippocampus and/or thalamus [86]. A conditional knock-in mouse model (Scn1a Stop/+) allowing on-demand reactivation of Scn1a when symptoms had already manifested (P30) led to complete rescue of both spontaneous and thermally induced seizures, marked amelioration of behavioural abnormalities and normalisation of hippocampal fast-spiking interneuron firing. Additionally, widespread gene expression abnormalities, including those linked to astrogliosis, were reversed. Notably, reactivation at P90 (adulthood) also restored physiological Nav1.1 levels and rescued seizures, despite months of prior seizures [87].
Translational gene-regulation approaches have advanced toward clinical evaluation.
ETX101
ETX101 is a non-replicating recombinant AAV9 gene therapy that delivers an engineered transcription factor (eTF-SCN1A), regulated by a GABAergic neuron-specific enhancer element (REGABA). Unlike direct delivery of the large SCN1A gene, ETX101 enhances transcription of the endogenous wild-type SCN1A specifically within GABAergic inhibitory neurons. In a Scn1a+/- mouse model, ETX101 achieved peak expression 3–4 weeks post-injection, binding upstream of the SCN1A gene to increase NaV1.1 channel production, thereby restoring interneuron function [88]. ETX101 significantly reduced hyperthermia-induced seizure susceptibility and prolonged survival up to 470 days post-treatment. In non-human primate toxicology studies, a single intracerebroventricular dose of ETX101 was well tolerated over 28 days, with no serious adverse events or notable histopathological changes [88].
The POLARIS clinical development programme is evaluating ETX101 in children with SCN1A-positive DS across three phase I/II trials. The ENDEAVOR trial (NCT05419492) in the US and the EXPEDITION trial (NCT06283212) in the UK enrol participants aged 6 months to < 4 years, while WAYFINDER (NCT06112275) in Australia includes children aged 6 months to < 7 years. Each trial involves a single intracerebroventricular injection into the lateral ventricle via stereotactic neuronavigation, with staggered dosing, safety monitoring (blood, serum, CSF analysis, imaging, immunogenicity), and efficacy assessments based on seizure frequency, cognition, behaviour, motor function, sleep, and quality of life. Early results are promising with no treatment-related serious adverse events reported in 19 children aged 6 months to 7 years treated with a single intracerebroventricular dose of ETX101 across four dose levels. Dose-dependent and sustained reductions in monthly countable seizure frequency were observed at the first three dose levels, including a median 78% reduction at the third dose level over up to 7 months of follow-up in highly drug-resistant participants. In addition to seizure reduction, clinically meaningful improvements in neurodevelopmental outcomes were reported. Children treated before 2 years of age demonstrated accelerated cognitive skill acquisition as early as 16 weeks, with progressive gains through 52 weeks of observation, deviating from the developmental plateau typically observed in natural history cohorts [89]. These findings provide early evidence supporting the potential disease-modifying effects of ETX101, although confirmation in phase III studies is required. These will commence in 2026.
Collectively, gene-replacement and gene-regulation approaches illustrate complementary paths toward restoring NaV1.1 function in DS. While gene replacement offers direct restoration of channel expression, gene-regulation strategies leverage endogenous genomic control mechanisms and may provide greater cell-type specificity and durability. However, both approaches rely on viral vector delivery and therefore share important safety considerations, including dose-dependent immune responses, vector-related toxicity, and uncertainties regarding long-term expression and neurodevelopmental effects following early-life intervention. In addition, gene-regulation strategies introduce theoretical risks related to excessive or off-target transcriptional activation that will require careful dose selection and prolonged surveillance. Ongoing clinical studies will be critical in defining the optimal balance between efficacy, safety, durability and feasibility of these approaches.
Conclusions
Major advances over the past decade have expanded treatment options for DS with medications such as stiripentol, fenfluramine and pharmaceutical-grade cannabidiol providing meaningful reductions in convulsive seizure burden and thereby improvements in quality of life. Nonetheless, seizures remain refractory in a substantial proportion of patients, and current therapies do not fully address the complex cognitive, behavioural and adaptive impairments that characterise the disorder.
Emerging RNA-based and gene-based therapies represent a shift toward mechanism-based treatment strategies targeting SCN1A haploinsufficiency and have the potential to be disease-modifying. Early-phase clinical studies of antisense oligonucleotide approaches, such as STK-001, demonstrate acceptable safety and suggest potential benefits beyond seizure reduction, although definitive efficacy data are awaited from ongoing controlled trials. Gene-regulation strategies, including ETX101, have shown robust restoration of interneuron function and seizure protection in preclinical models, but human clinical outcomes remain under investigation.
As highlighted in recent syntheses of non-seizure outcomes in DS, including a recent review [21], cognitive, behavioural and quality-of-life impairments contribute substantially to disease burden and should be evaluated when studying emerging therapies. Continued integration of rigorous clinical trial design, natural history data and long-term follow-up will be essential to determine whether mechanism-based approaches can meaningfully modify both seizure and non-seizure outcomes across the lifespan in individuals with Dravet syndrome.
Funding
Open access funding provided by Università degli Studi di Firenze within the CRUI-CARE Agreement. No funding was received for the publication of this paper.
Declarations
Conflicts of interest
IS has served on scientific advisory boards for BioMarin, Chiesi, Eisai, Encoded Therapeutics, GSK, Knopp Biosciences, Nutricia, Takeda Pharmaceuticals, UCB, Xenon Pharmaceuticals, and Longboard Pharmaceuticals; has received speaker honoraria from GSK, UCB, BioMarin, Biocodex, Chiesi, Liva Nova, Nutricia, Zuellig Pharma, Stoke Therapeutics, Eisai, Akumentis, and Praxis; has received funding for travel from UCB, Biocodex, GlaxoSmithKline, Biomarin, Encoded Therapeutics, Stoke Therapeutics, Eisai, Longboard Pharmaceuticals; has served as an investigator for Anavex Life Sciences, Cerevel Therapeutics, Eisai, Encoded Therapeutics, EpiMinder Inc, Epygenyx, ES-Therapeutics, GW Pharma, Longboard Pharmaceuticals, Marinus, Neurocrine BioSciences, Ovid Therapeutics, SK Life Science, Takeda Pharmaceuticals, UCB, Ultragenyx, Xenon Pharmaceuticals, Zogenix, and Zynerba; has consulted for Care Beyond Diagnosis, Epilepsy Consortium, Atheneum Partners, Ovid Therapeutics, UCB, Zynerba Pharmaceuticals, BioMarin, Encoded Therapeutics, Biohaven Pharmaceuticals, Stoke Therapeutics, and Praxis; and is a Non-Executive Director of Bellberry Ltd and a Director of the Australian Academy of Health and Medical Sciences. She might accrue future revenue on pending patent WO61/010176 (filed in 2008): Therapeutic Compound; has a patent for SCN1A testing held by Bionomics and licensed to various diagnostic companies; and has a patent molecular diagnostic/theranostic target for benign familial infantile epilepsy [PRRT2] 2011904493 and 2012900190 and PCT/AU2012/001321 (TECH ID: 2012-009). SB has served on scientific advisory boards for Biocodex and Longboard Pharmaceuticals and has received speaker honoraria from Angelini, Biocodex, Eisai, Lusofarmaco, and Jazz Pharma.
Availability of data and material
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.
Ethics approval
Not applicable.
Consent to participate
Not applicable.
Consent for publication
Not applicable.
Code availability
Not applicable.