Current state of the epilepsy drug and device pipeline

Abstract

The field of epilepsy has undergone substantial advances as we develop novel drugs and devices. Yet, considerable challenges remain in developing broadly effective, well-tolerated treatments, but also precision treatments for rare epilepsies and seizure monitoring devices. We summarize major recent and ongoing innovations in diagnostic and therapeutic products presented at the seventeenth Epilepsy Therapies and Diagnostics Development conference, which occurred May 31^st to June 2^nd, 2023, in Aventura, Florida. Therapeutics under development are targeting genetics, ion channels and other neurotransmitters, and many other potentially first-in-class interventions such as stem cells, glycogen metabolism, cholesterol, the gut microbiome, and novel modalities for delivering electrical neuromodulation.

Introduction

The need remains for innovation in epilepsy diagnostics and treatments. One-third of people with epilepsy may not have seizures controlled on antiseizure medications (ASMs).¹ Thus, developing increasingly effective and tolerated therapeutics remains critical. Additionally, to study new therapeutics, rigorous modalities are needed for seizure tracking.

This manuscript describes promising, ongoing efforts in therapeutics and device development presented at the seventeenth Epilepsy Therapies and Diagnostics Development conference in May-June 2023. We describe many therapeutics in progress, such as targeting refractory seizures or else promoting precision medicine via potentially disease-modifying approaches. We review therapeutics pertaining to a wide variety of mechanisms of action (e.g., ion channels, neurotransmitters, genetics, and other) in addition to promising devices for seizure treatment and monitoring. These efforts seek to develop more effective, selective, better-tolerated ASMs, including novel routes of administration, targeting a wide array of disease states from rare epilepsies to status epilepticus to more generally drug-resistant epilepsy.

Therapeutics

Ion channels

Voltage-dependent potassium channels

Potassium channels have emerged as a promising target. The first potassium channel opener, ezogabine (retigabine), demonstrated efficacy² leading to approval in 2011. However, warnings emerged in 2013 regarding skin and fundoscopic pigment changes³ and the drug was withdrawn in 2017 due to limited use. Three next-generation candidates targeting K_v7 are currently under development.

XEN1101 (Xenon Pharmaceuticals) is in the most advanced stage among next-generation potassium channel openers. It has demonstrated high potency at K_v7 channels in preclinical and in vitro/in vivo data, surpassing ezogabine. A recent double-blinded phase IIB study in focal epilepsy (N=325, median baseline 13–17 seizures/month, failed 6 ASMs) was encouraging.⁴ The high-dose arm of XEN1101 reduced monthly seizure counts by 53% compared to 18% in the placebo arm (p<0.001). Efficacy was seen as early as the first week at every dose, and further advantages include once-daily dosing with no titration. Adverse effects included dizziness (25%) and somnolence (16%) and rare urinary retention (N=2) and weight gain (N=7), but no tissue discoloration thus far (XEN1101 is chemically incapable of forming the phenazinium dimers implicated in ezogabine-related pigmentary abnormalities). Phase III studies enrolling adults taking 1–3 ASMs with focal (X-TOLE2 and X-TOLE3) or primary generalized (X-ACKT) epilepsy are underway. A phase II study in Major Depressive Disorder (X-NOVA) will complete before 2024.

Similarly, Biohaven is developing BHV-7000.^5,6 Preclinical data has shown that BHV-7000 hyperpolarizes neurons and increases action potential thresholds. Furthermore, BHV-7000 has chemical modifications compared with ezogabine intended to avoid causing dimers causing tissue discoloration. The compound was also intended to be as selective as possible for the potassium channel, minimizing GABAergic adverse effects. The drug is currently in a phase I randomized double-blinded placebo-controlled SAD/MAD study. Among 61 healthy adults treated at several doses for 15 days, the most common adverse effect was headache (N=7), with no dose-limiting toxicity and no somnolence or dizziness. Phase I studies are nearing completion, with Phase II-III trials planned to start by the end of 2023.

ETX-123 (Eliem Therapeutics) is a potent K_v7.2/3 channel opener with high selectivity and efficacy in preclinical studies. ETX-123 exhibits high potency at K_v7.2/3 channels, per EC50 values that are orders of magnitude more effective than ezogabine (ezogabine: 2 μM: ETX-123: 0.0007–0.002 μM).⁷ In vivo testing using the maximal electroshock seizure (MES) rat model has also demonstrated remarkable efficacy, with 2 mg/kg doses providing 50% protection, and 8 and 14 mg/kg doses offering complete protection from seizures. Tolerability assessments, such as the Rotarod test, revealed a favorable safety profile for ETX-123. Some concern has existed regarding urinary retention (K_v-7.4 subunit related), though this has not borne out when tested in a human bladder contractility assay, with EC50 values for muscle relaxation 700-fold higher than typical anticonvulsant plasma concentrations, unlike ezogabine (only 4-fold higher). With such promising preclinical data, ETX-123 awaits further investigation through additional in vivo pharmacology studies before proceeding to human trials.

Voltage-dependent sodium channels

Sodium channel blockade is a canonical ASM mechanism, though innovations continue, expanding targets and populations.

One problem is that currently available sodium channel blockers are nonselective, inhibiting voltage-gated sodium channel isoforms in myocardial, skeletal, and cerebral tissue.⁸ Na_V1.5 inhibition risks arrhythmogenesis given it produces the rising phase of the cardiac action potential,⁹ and Na_v1.4 inhibition may alter skeletal muscle excitability.^10–12 Within the nervous system, Na_V1.1 inhibition can aggravate certain syndromes (e.g. Dravet syndrome).¹³

Neurocrine Biosciences, Inc. is developing NBI-921352, a selective Na_v1.6 sodium channel inhibitor. Study NBI-921352-FOS2021 is a phase II randomized, double-blind, placebo-controlled study that enrolled 101 patients aged 18–65, to investigate the safety, tolerability, pharmacokinetics, and efficacy of NBI-921352 as adjunctive therapy in focal seizures. NBI-921352-DEE2012 is a phase II randomized, double-blind, placebo-controlled study that will seek to enroll 52 patients aged 2–21, to investigate NBI-921352 as adjunctive therapy for SCN8A developmental and epileptic encephalopathy (DEE).

Lacosamide is another sodium channel blocker, which enhances slow inactivation of sodium channels. While it is already approved for patients at least one month old, data are sparse for neonates. SP0968 (UCB) is a phase II/III, multicenter, open-label, randomized study, comparing lacosamide with standard of care. Eligible neonates will be at least 34 weeks of gestational age who have any confirmed electrographic seizure ≥2 minutes or ≥3 electrographic seizures and have previously received either phenobarbital, levetiracetam, or midazolam, weigh at least 2.3 kg, and have not previously received other sodium channel blockers.

Neurotransmitters

Gamma-aminobutyric acid (GABA)

OV329 (Ovid Therapeutics) is a next-generation GABA aminotransferase inhibitor in development for treating rare epilepsies. Its potency is two hundred times greater than vigabatrin, which could allow use at lower doses to avoid visual field deficits seen with vigabatrin. OV329 has demonstrated antiseizure activity in six of eight mouse and rat models thus far. Combined with safety data in rats and dogs, this suggests that OV329 may offer an improved benefit risk profile compared with vigabatrin. For example, OV329 significantly reduced seizures in a corneal kindled mouse model in a dose-dependent manner, and hippocampal paroxysmal discharges were reduced by 46% after a single oral dose in a mesial temporal lobe epilepsy mouse model (unpublished, courtesy of Ovid Therapeutics). It is currently in a Phase I Single Ascending Dose/Multiple Ascending Dose (SAD/MAD) study with no significant adverse effects thus far. Further results are expected in 2024.

Inhaled alprazolam administered via the Staccato^® breath-actuated device aims for rapid systemic exposure for seizure termination. A phase IIb randomized placebo-controlled study (N=116) recently tested its ability to abort long or repetitive seizures in an inpatient setting.¹⁴ Participants were randomized 1:1:1 to double-blinded treatment of a single seizure event with one dose of Staccato alprazolam 1 mg, 2 mg, or placebo. Both 1 and 2 mg doses demonstrated efficacy in rapidly terminating seizures (~66% responder rate in both drug arms versus 43% placebo responder rate; p=0.04). Both doses were also generally well-tolerated, the most common adverse effects being cough and somnolence (both 15%). While the initial responder rate was the same with 1 and 2 mg doses, the seizure recurrence rate was lower with 2 mg. A multicenter, randomized placebo-controlled study (N=250) remains underway to test the efficacy and safety in participants at least 12 years old with stereotypical generalized or focal prolonged seizures, beyond the inpatient setting.

The neuroactive steroid ganaxolone (Marinus Pharmaceuticals) acts as a positive allosteric modulator of GABA_A receptors.¹⁵ It is FDA-approved for CDKL5 deficiency disorder in patients 2 years and older, based on a phase III study in 100 patients.¹⁶ That study found a 31% versus 7% median reduction in major motor seizures (p<0.01) with maintenance of response over a 2-year open-label phase. The odds ratio for improvement in the Clinical Global Impression scale was 2.6 and the 50% responder rate was 24% for ganaxolone versus 10% placebo (p=0.06). Most non-seizure neurodevelopmental outcomes were unchanged. An exploratory analysis on reduction in seizure intensity and severity favored ganaxolone [OR (95%CI) = 2.56 (1.20,5.45)]. Ongoing work includes testing ganaxolone in other DEEs (Lennox-Gastaut syndrome in phase I, TSC in phase III). Additionally, an intravenous formulation is being evaluated in two placebo-controlled trials in refractory status epilepticus.¹⁷

While temporal lobectomy is effective for appropriate candidates, many patients remain reluctant or are not candidates for resection,¹⁸ and resection and other invasive treatments cannot guarantee cure and may require indwelling hardware. NRTX-1001 (Neurona Therapeutics) is an investigational therapy for drug-resistant unilateral mesial temporal lobe epilepsy consisting of a single MRI-guided injection of GABAergic human stem cell-derived inhibitory interneurons into the hippocampus.¹⁹ Immunosuppression is required for one year. NRTX-1001 is currently undergoing a Phase I/II clinical trial, with a primary endpoint of 1-year adverse events. Secondary endpoints include 1-year reduction in seizure frequency and responder rates, seizure severity, quality of life, neuropsychological outcomes, imaging biomarkers, and ASM dose reduction. The first participant, with data available through 9 months post-transplant, experienced no serious adverse events and a 95% reduction in seizures with stable neurocognitive scores at 6 and 9 months. The second participant, with data available up to 6 months post-transplant, has experienced no serious adverse events and a similarly excellent 94% reduction in seizures. Neurocognitive scores were stable or improved at 6 months, excluding a one-point decline in the Rey Auditory Verbal Learning Test. Recruitment remains ongoing at 11 sites.

Synaptic vesicle glycoprotein 2A (SV2A)

Currently, there are few approved monotherapies for absence epilepsy.²⁰ While levetiracetam is often used in generalized epilepsies, there may be advantages of brivaracetam due to its ten-fold higher affinity for the SV2A binding site. A 2007 subject-blinded placebo-controlled study (N=18) in patients with photosensitive epilepsy found that even a single dose of brivaracetam (10, 20, 40, or 80 mg) suppressed photoparoxysmal EEG responses.²¹ Therefore, EXPAND (UCB, N01269, started 7/2021, expected completion 2024) is a randomized, dose-finding and confirmatory, double-blinded, placebo-controlled, parallel-group, multicenter phase II/III trial evaluating the efficacy, safety, and tolerability of brivaracetam monotherapy in patients 2–25 years old with childhood or juvenile absence epilepsy. The trial consists of dose-selection and futility assessment stages, followed by an optimal-dose stage after interim analysis. This two-stage trial design will seek to investigate two potentially efficacious brivaracetam doses.²²

Glutamate receptors

Radiprodil is a negative allosteric modulator that binds to the N-methyl-D-aspartate (NMDA) receptor NR2B (GluN2B) subunit.²³ As GRIN genes encode GluN subunits of the NMDA receptor, radiprodil has the potential to become a precision therapy for GRIN-related disorders caused by gain-of-function mutations by reducing overactive NMDA receptors. In previously completed clinical studies, radiprodil has demonstrated favorable safety, tolerability, and pharmacokinetic profiles in animal models and adults. The 12 treatment-emergent adverse events (TEAE) reported by 8/10 healthy adults in a phase I study were all reported to be mild and temporary. The most common TEAEs were dizziness (40%), nausea (20%) and somnolence (20%).²⁴ Additionally, spasms responded favorably in three infants with prednisolone- and vigabatrin-resistant infantile spasms; one patient achieved sustained seizure freedom and the other two experienced a reduction in seizure clusters.²⁵ GRIN Therapeutics’ phase Ib radiprodil study is currently enrolling patients 6 months to 12 years old with gain-of-function GRIN variants in Europe to further assess safety, tolerability, pharmacokinetics, and efficacy as measured by impact on seizures, behavioral symptoms, and caregiver burden.

Basimglurant (NOE-101, Noema Pharma) is a negative allosteric modulator of mGluR5 to reduce glutamatergic function.²⁶ In addition, basimglurant may have another mechanism of action – normalizing aberrant protein synthesis via activation of the MEK/ERK pathway. Attention has therefore recently focused on its potential in reducing mTOR-dependent protein synthesis in Tuberous Sclerosis Complex (TSC). Chronic treatment with an mGluR5 modulator in a TSC mouse model (n=6) decreased seizure frequency (p=0.046) and total seizure time (p=0.044) compared with placebo (n=5).²⁶ Noema’s Galene global Phase IIb trial is enrolling pediatric and adult patients with TSC-associated seizures. Existing human trials in adults and adolescents with depression and Fragile X further support development for use in epilepsy.²⁷ Nearly two-thirds (122 out of 183) of participants with Fragile X Syndrome in the FragXis phase II trial experienced at least one (mostly gastrointestinal or psychiatric) TEAE, eight of whom discontinued the drug and the remainder resolved during the study.²⁷ This compound is also being developed for trigeminal neuralgia pain.

Serotonin

Fenfluramine’s (ZX008) antiseizure properties are proposed to be secondary to serotonin release, agonistic action at 5-HT1D, 5-HT2A and 5-HT2C receptors, and positive modulation of sigma-1 receptors.²⁸ Currently, fenfluramine is approved for treating convulsive seizures in Dravet syndrome patients Recently, Devinsky et al²⁹ demonstrated its efficacy and safety in six pediatric patients with cyclin-dependent kinase like-5 (CDKL5) deficiency disorder, particularly in those patients with generalized tonic-clonic seizures, with a median of 90% reduction in seizure frequency. The ongoing phase III GEMZ study is part of the FAiRE (Fenfluramine Assessment in Rare Epilepsy; UCB) program investigating efficacy and safety of fenfluramine in pediatric and adult patients with CDKL5 deficiency disorder-related refractory epilepsy across approximately 70 study sites globally.

LP352 (Longboard Pharmaceuticals) also acts at the serotonin receptor but is a selective 5-HT2C superagonist. The absence of action at 5H2A/B is felt to avoid side effects such as cardiac valve thickening. Like other serotonin agonists, LP352 is being developed for the treatment of DEEs. A Phase I open-label study (LP352 102; N=20) assessing the pharmacokinetics and pharmacodynamics of LP352 found that plasma and CSF concentrations increased in dose-dependent manners, confirming crossing of the blood brain barrier. Furthermore, LP352 affected quantitative EEG activity. The double-blinded Phase Ib/IIa PACIFIC study is currently evaluating LP352 for treatment of multiple DEEs, across 30 sites with three different doses, data expected early 2024.

Acetylcholine

Acetylcholine is not a typical target of existing ASMs. However, SPN-817 (formerly BIS-001ER) is a synthetic version of Huperazine A, which is an alkaloid acetylcholinesterase inhibitor from the Chinese clubmoss Huperzia serrata. It is proposed to reduce seizures by modulating the excitatory-inhibitory balance in neurons and reducing inflammation in glia via increased GABA release.³⁰ The Supernus SPN-817 Phase I study completed in 2022 demonstrated safety and good tolerability with elimination of adverse effects with slower dose titration in healthy individuals. In the subsequent phase Ib/II pilot study enrolling adults with refractory focal seizures, two of three participants were seizure-free at 24 months, with mild/transient TEAEs including nausea, insomnia, unsteadiness, and nystagmus.³¹ Now, its phase IIa study with an optional open-label extension is investigating effects on inflammatory biomarkers, cognitive performance, and EEG.

Genetic

As genetic causes of epilepsies become increasingly understood, opportunities exist for increasing therapeutic precision.

STK-001 is an ASO engineered by Stoke Therapeutics, Inc. for Dravet syndrome. STK-001 is part of a technology called Targeted Augmentation of Nuclear Gene Output to treat haploinsufficiencies like Dravet syndrome. In Dravet syndrome caused by SCN1A loss-of-function mutations, an alternative exon to pre-mRNA leads to a non-productive transcript that is degraded by nonsense-mediated mRNA decay resulting in reduced Na_v1.1 production.³² STK- 001 blocks the incorporation of this alternative exon to increase productive SCN1A mRNA expression. Initial data from 2022 suggested a 55% seizure frequency reduction in six patients at the 45 mg dose. Subsequent data in 2023 were less robust (N=16, only an 18% reduction), though the 70 mg dose has demonstrated median reductions up to 80–90%. Additionally, in vitro safety concerns surrounded off-target effects such as thrombocytopenia or complement activation,³³ which have not yet been observed clinically. The compound is currently undergoing a phase I/IIa open-label study for continued efficacy and safety information.

AMT-260 (UniQure, Inc.) is an Adeno-Associated Virus (AAV) miRNA-based gene therapy targeting GRIK2 (glutamate receptor ionotropic kainate 2) for refractory temporal lobe epilepsy. The GRIK2 gene encodes a GluK2 subunit of the kainate receptor. Aberrant GluK2-containing kainate receptors play a key role in the pathogenesis of refractory temporal lobe epilepsy.³⁴ In preclinical studies, AMT-260 has shown a dose-dependent decrease in GRIK2 gene expression in proportion to the expression of the miRNAs in the mouse hippocampus and achieved up to 75% seizure reduction. AMT-260 is injected into affected brain tissue; injection accuracy testing in primates has demonstrated post-administration transduction within the hippocampus. UniQure is planning to conduct a Phase I/II trial in adults with unilateral refractory mesial temporal lobe epilepsy in 2023. In this trial, AMT-260 will be administrated to the hippocampus under MRI-guidance following the same trajectory used for laser interstitial thermal therapy.

A third example of genetic innovation includes CAP-002 (Capsida Biotherapeutics), an AAV vector-based product for syntaxin-binding protein 1 (STXBP1) mutations. STXBP1 interacts with numerous synaptic proteins in the brain playing an essential role in synaptic vesicle fusion and neurotransmitter release.³⁵ Genetic epilepsy caused by STXBP1 mutation has severe disease manifestations including early-onset refractory seizures, cognitive dysfunction, autistic behaviors, and gait disturbances. Drug development is vital, as seizures are typically refractory³⁶ and there are no currently available disease-modifying treatments. Preclinical mouse data have demonstrated that intravenous administration of a capsid carrying STXBP1 cargo produced dose-dependent improvement in spike-wave discharges and cognitive (e.g., object recognition, fear response) and motor (e.g., limb clasping) deficits in adult mice, lasting at least 12 months post therapeutic intervention. Intravenous delivery of the compound resulted in widespread transduction in neocortex, caudate, putamen and thalamus, while avoiding uptake by the liver, in non-human primates. Thus, this drug remains a promising candidate in preclinical development.

Finally, NMT.001 represents another RNA-based therapy undergoing preclinical development (NEUmiRNA Therapeutics). It inhibits miR-134, which is a brain-specific miRNA implicated in controlling neuronal microstructure. miR-134 appears upregulated in experimental models of status epilepticus.³⁷ Silencing miR-134 exerts prolonged seizure-suppression and neuroprotection in animal models.³⁸ In a proof-of-concept screening test using intra-amygdala kainate mouse models of mesial temporal lobe epilepsy, treatment with NMT.001 dramatically reduced lorazepam-resistant status epilepticus seizures

Notably, diagnostic codes do not currently exist for many rare epilepsies. With increasing availability of such treatments, further efforts advocating for unique International Classification of Diseases codes to distinguish rare epilepsies could improve access and spur research.

Other/multiple targets

Intracerebroventricular (ICV) valproate

Another avenue for innovation includes using conventional ASMs in less conventional ways. ICV drug delivery with a device implanted under the scalp already has a track record in scenarios such as treating infections, pain, and neoplasms.³⁹ It has several appealing features. In contrast to intrathecal delivery, ICV delivery avoids the need for repeated lumbar punctures. And in contrast to oral delivery, ICV delivery may reduce systemic toxicity. Valproate is an example with known systemic toxicity and teratogenicity. ICV administration of valproate could potentially avoid these pitfalls related to absorption into the bloodstream by producing a high drug concentration in the brain, but a low serum concentration. In a proof-of-concept study performed by Cerebral Therapeutics, five patients with refractory mesial temporal lobe epilepsy, who had previously not responded to oral valproate, received an implanted device delivering ICV valproate. Four subjects responded with >50% seizure reduction at the highest tested dose of 160 mg/day. All five subjects reported significant quality of life improvement.⁴⁰ This observation could eventually open up a completely new path for drug-refractory patients.

Cholesterol inhibitor

Soticlestat (Takeda) is another first-in-class investigational compound. It inhibits cholesterol-24-hydroxylase, an enzyme primarily expressed in the brain that reduces 24-hydroxcholesterol (24HC).⁴¹ 24HC is an NMDA modulator, thus implicated in hyperexcitability as well as inflammatory responses including induction of TNF-α. Phase I studies have confirmed drug-target binding and decreased 24HC via PET ligand studies and found a favorable drug-drug interaction profile. A 2022 proof of concept study (ELEKTRA) found robust reduction in convulsive seizures in Dravet syndrome (N=51; 46% median seizure reduction; p<0.001) and reduction in drop seizures in Lennox-Gastaut syndrome (N=83; 15% placebo-adjusted median seizure reduction; p=0.13),⁴² encouraging future phase III studies.

Polyglucosan

In epilepsy, the cellular machinery for glycogen metabolism may fail to synthesize homogeneous glycogen β-particles thus producing insoluble polyglucosans, which cannot be further metabolized. This process results in epileptogenesis due to detrimental neuronal effects.⁴³ This may also impair the main function of astrocytic glycogenolysis (potassium uptake), which leads to sustained depolarization.⁴³ VAL-1221 (FAB-GAA by Parasail LLC) is an antibody-enzyme fusion given every other week designed to block this pathway by clearing polyglucosan buildup. Parasail is currently fundraising to conduct expanded studies in Lafora disease, a disorder of glycogen metabolism which causes severe progressive myoclonic epilepsy and childhood dementia. Ten Lafora patients are currently being treated worldwide via Compassionate Use including three in the US under a prospective VAL-1221 Lafora Expanded Access Protocol filed as a commercial Investigational New Drug. Parasail is seeking NIH clinical grants and industry partnerships to execute additional formal studies against this rare but devastating condition without any other known disease-modifying treatment.

Cannabidiol (CBD)

Currently, CBD is available only as a solution that suffers from poor bioavailability, poor water solubility, pharmacokinetic and dosing variability, liver toxicity, and complex drug-drug interactions.⁴⁴ London Research & Pharmaceuticals is developing LRP-01, a CBD derivative, for Dravet syndrome and LGS. LRP-01 enables CBD delivery via tablets and capsules for the first time. It has significantly better bioavailability (25%, rather than <6% for CBD) and predictable pharmacokinetics, with no first-pass metabolism. It has demonstrated up to a 25-fold increase in absorption and decreased alkaline phosphatase changes and cytochrome P450 effects (p<0.05, unpublished). These modifications may confer greater anti-seizure efficacy than typical CBD. The company projects rapid market access of this new drug through the 505 (b)(2) regulatory pathway for new drug applications.

Gut microbiome

BL-001 (BloomScience) is an orally administered live biotherapeutic product developed for Dravet syndrome and other DEEs. In a 6-Hz-induced seizure model of refractory epilepsy, the gut microbiome appeared to mediate the effect of ketogenic diet on seizures, by modulating hippocampal GABA/glutamate ratios.⁴⁵ BL-001 contains two human gut microbes that have been shown in both cell-based assays and animal studies to reduce hyperexcitability, increase GABA in the hippocampus, and significantly reduce both seizure frequency and duration to replicate the antiseizure effect of ketogenic diet. Bloom Science, Inc. has completed enrollment of 32 healthy adult participants in a phase I clinical trial, across four dose cohorts and randomized in a 3 to 1 ratio of BL-001 to placebo. Following completion of the current Phase I trial, they intend to initiate a Phase II study in patients with Dravet syndrome.

Neurostimulation

Vagus Nerve Stimulation (VNS), Responsive Neurostimulation (RNS), and Deep Brain Stimulation (DBS) all present treatment palliative options for patients with drug-resistant epilepsy who are not candidates for or who have failed or decline resective/ablative surgery. These devices have responder rates of around 60% at 4 to 9 years post-implantation, with potentially increasing efficacy over time, and smaller subsets achieving prolonged seizure freedom.^46–48

Many unanswered questions surround candidate and stimulation parameter selection. For example, in a study among 10 patients with poor response to high-frequency stimulation, transition to low-frequency stimulation was associated with a 76% median seizure reduction compared to pre-implantation baseline,⁴⁹ raising questions about how to tailor parameters. Moreover, future investigation should not just focus on optimizing stimulation parameters, but also how short-term biomarker changes might predict long-term clinical response, and how to tailor stimulation timing to an individual patient’s high-risk periods per RNS electrocorticography. To illustrate, in a retrospective analysis from 25 adults, stimulation showed differential impact on seizure frequency depending upon whether patients were in low versus high seizure risk states as determined by identifying seizure cycles.⁵⁰

There remains a pipeline of novel neurostimulation devices. For drug-resistant neocortical focal epilepsy, these include Starstim and EASEE (Epicranial Application of Stimulation Electrodes for Epilepsy).

Starstim applies transcranial Direct Current Stimulation (tDCS) via a wearable device over the scalp above the seizure onset zone to suppress cortical excitability and reduce seizure frequency. An individualized montage is developed for each patient. Children and adults in an open-label pilot study experienced a median 44% seizure reduction with the device.⁵¹ About 1% of patients experienced skin injury.⁵² Starstim is currently recruiting for a Phase III trial in children and adults. This involves a 12-week baseline, followed by 20 minutes daily of cathodal stimulation in a laboratory setting for 10 consecutive weekdays, then follow up for 10 weeks. The primary endpoint is change in seizure frequency, with a secondary endpoint of responder rate. At-home tDCS also has improved refractory depression in adults,⁵³ which also raises prospects for its use in a prevalent epilepsy comorbidity.

The EASEE device, in contrast, consists of epicranial electrodes implanted sub-scalp above the epileptogenic focus, connected via a cable through the neck under the skin to a pulse generator in the chest. The device acts via high-frequency stimulation every other second and low-frequency stimulation for twenty minutes per day. The parameters of the stimulation are programmable via a tablet by the treating physician according to patient needs. In an open-label trial of 33 patients in Europe, there was a 52% median seizure reduction from baseline to the 6^th month with 4 patients becoming seizure-free and an overall responder rate of 53%. After receiving a European CE mark, the device became commercially available in several European countries. Next steps include an IDE submission to initiate clinical work in the US.

Seizure diaries and detection devices

Electronic seizure diaries for clinical trials

Patient-reported seizure counts are imperfect.⁵⁴ Seizure diaries seek to reduce recall bias by tracking outcomes in real-time.⁵⁵ With the explosion of apps, electronic seizure diaries have become common in trials. For example, electronic seizure diaries have already played central roles in approval of fenfluramine for Lennox-Gastaut and Dravet syndromes. Despite additional benefits of electronic data entry and storage, seizure diaries still pose challenges in their user acceptance, database structure, and ease of integration into clinical trial data collection and reporting activities.⁵⁶ Data cleaning and assessments of data integrity present particular challenges, as guidelines do not yet exist for the maximum duration of data entry delay (retrospective data entry or lookback period) and what happens thereafter if data corrections are required, which oftentimes require complex auditing processes to be fit for regulatory purposes.

Two example electronic seizure diaries in development include those by Signant Health and WCG. Both were created in collaboration with The Epilepsy Study Consortium to overcome some of these issues. Goals include standardization of data with updated terminology and common data elements,⁵⁷ complete audit trails, and improved ease of use to boost compliance. To address previously reported user challenges, both diaries allow for logging seizures with familiar wording, which are then matched to appropriate clinical classification based on ILAE terminology. Lookback periods (7 days for Signant Health and 1–7 days for WCG) allow users to add/edit entries, which can reduce or eliminate data corrections identified by the site or the sponsor. From the study team’s perspective, both electronic seizure diaries offer customizable data elements - Signant Health uses two standardized templates (“Daily Seizure Count” and “Daily Seizure Details”), and WCG’s eDiary offers pre-coded fields that can be added or removed, which allows for more efficient study startup. Data management is streamlined with visualization and periodic data reporting and alerts for both diaries.

Seizure detection devices

Seizure detection devices are available for clinical and research purposes. These devices seek to reduce morbidity and mortality associated with seizures and to address the limitations of patient reporting by improving objective documentation of seizure frequency.^58,59 Various modalities, including heart rate monitoring, photoplethysmography, electrodermal activity, electromyography sensory and limited EEGs have been validated and are often used in combination. The ILAE and IFCN recommend the use of wearable seizure detection devices for patients with tonic-clonic seizures with significant safety concerns, while emphasizing limitations for use in non-motor seizures.⁶⁰

Absence seizures represent one seizure type that is difficult to recognize clinically without video EEG. Due to their brief duration, without motor or autonomic changes, standard modalities for seizure detection have limited use for this seizure type. Epihunter, a wearable EEG device with an accompanying CE-marked smartphone-based application for behavioral testing, is designed for the detection of absence seizures, with reported 93% sensitivity and 0/hour median false alarm rates.⁶¹ This technology is being expanded to study non-convulsive seizures in rare epilepsies such as Dravet syndrome.

Diagnostic accuracy is higher with longer EEG monitoring.⁶² Though, prolonged ambulatory monitoring is limited with conventional EEG equipment. The REMI remote EEG monitoring system by Epitel uses four single patient, single use wireless sensors that transmit a single channel EEG to a mobile computing device and uses automated seizure detection using machine learning to tag events for clinician review. These hardware and machine learning algorithms are under FDA review, and a clinical trial comparing seizure count from three-day ambulatory EEG versus fourteen days of the REMI system is being planned.

EpiCare@Home from Byteflies, an end-to-end service which has received a CE mark in Europe, collects data using multi-modality (two-channel behind the ear EEG and accelerometer, heart rate, respiratory rate, and activity index) data. The two-channel EEG adhesives are hydrogel-based, biocompatible with regular replacement with intended use of up to four weeks. In a study of fifty patients, 84% had a change in clinical management using data gathered from the EpiCare@Home system.

Data from these devices could assist with changes in clinical management, seizure forecasting, seizure cycle assessment, and outcomes measurement in trials. Empatica Health Monitoring Platform, using the Embrace Plus (an FDA approved wearable seizure detection device using multiple modalities), offers continuous medical-grade data collection in real-world multi-site trials in addition to their already clinically approved available device.

National Institute of Neurological Disorders and Stroke (NINDS)

Dedicated funding mechanisms promote device development. For example, the NINDS Translational Neural Devices Program funds projects ranging from preclinical to feasibility phases. Blueprint Medtech is an NIH incubator which offers development research through its two Incubator Hubs (Center for Innovative NeuroTech Advancement & NeuroTech Harbor) and translational research through NIH funding opportunities for academic institutions (PAR-21–315) or small businesses (PAR-21–282). The Brain Research Through Advancing Innovative Neurotechnologies (BRAIN) Initiative offers several funding opportunities to develop and apply innovative neurotechnologies. Goals include better understanding of the human brain and developing new ways to treat and prevent brain disorders.

Conclusion

Despite the explosion in the number of available ASMs,⁶³ we have yet to reduce the frequency of medically-refractory patients.¹ This review highlights immense, ongoing innovation. Developing new mechanisms of action, expanding existing drugs into new populations or new routes of delivery, and refining our ability to track and detect seizures all hold promise for improving the lives of people with epilepsy.

Figure 1: Epicranial Application of Stimulation Electrodes for Epilepsy (EASEE) leads.

These leads are implanted underneath the scalp above the epileptogenic focus and are connected via a cable under the skin to a pulse generator in the chest. This applies high-frequency direct stimulation every other minute and direct current stimulation for twenty minutes per day. Potential advantages of this device include being able to avoid intracranial surgery during implantation, and the device is easily removable and repositioned if needed.

Table 1:

Drugs

Target	Drug	Company	Phase	Population
Ion channels
Volt-dept. K channels	XEN1101	Xenon Pharmaceuticals	III	Focal and generalized epilepsy
Volt-dept. K channels	BHV-7000	Biohaven Pharmaceuticals	I	Drug-refractory focal and generalized epilepsy
Volt-dept. K channels	ETX-123	Eliem Therapeutics	Preclinical	Drug-refractory focal and generalized epilepsy
Volt-dept. Na channels	NBI-921352	Neurocrine Biosciences, Inc.	II	Focal onset seizures
Volt-dept. Na channels	Lacosamide	UCB	II/III	Neonates
Neurotransmitters
GABA	OV329	Ovid Therapeutics	I	Rare adult and pediatric epilepsies
GABA	Alprazolam	UCB	III	Long or clustering seizures
GABA	Ganaxolone	Marinus Pharmaceuticals, Inc.	IV	CDKL5 deficiency disorder (approved), TSC, LGS, refractory SE
Stem cell-derived GABAergic interneurons	NRTX-1001	Neurona Therapeutics	I/II	Unilateral drug-refractory mesial temporal lobe epilepsy
SV2A	Brivaracetam	UCB	II/III	Absence epilepsy
NMDA subtype of glutamate receptors	Radiprodil	GRIN Therapeutics	IIb	Pediatric patients with gain-of-function GRIN variants: seizure cohort, behavioral cohort
Glutamate	Basimglurant	Noema Pharma		Pediatric and adult patients with epilepsy associated with TSC
Serotonin	LP352	Longboard Pharmaceuticals	Ib/Iia	DEE’s
Serotonin	Fenfluramine	UCB	III	Pediatric and adult patients with CDKL5 deficiency disorder and refractory epilepsy
Acetylcholine	SPN-817	Supernus Pharmaceuticals	Iia	Adult patients with refractory focal impaired awareness epilepsy
Genetic
TANGO ASO	SKT-001	Stoke	I/Iia	SCN1A Dravet syndrome
AAV-miRNA	AMT-260	uniQure	Preclinical	Temporal lobe epilepsy
Anti-miRNA	NMT.001	NEUmiRNA Therapeutics	Preclinical	Drug refractory epilepsy
DNA replacement	CAP-002	Capsida Biotherapeutics	Preclinical	STXBP1 DEE
Other/multiple targets
Valproate	Intraventricular delivery	Cerebral Therapeutics	II	Drug-refractory adults
Cholesterol-24-hydroxylase	Soticlestat	Takeda Pharmaceuticals	I	Dravet syndrome, LGS
Polyglucosan	VAL-1221	Parasail LLC	II	Lafora disease
Multiple	Cannabidiol	London Research and Pharmaceuticals		Anticipated 505(b)(2) regulatory pathway
Multiple	BL-001/gut bacteria	Bloom Science	I	Dravet syndrome

Abbreviations: AAV-miRNA (Adeno-associated Virus microRNA), ASO (antisense oligonucleotide), DEE (developmental and epileptic encephalopathy), LGS (Lennox-Gastaut syndrome), SE (status epilepticus), TANGO (Targeted Augmentation of Nuclear Gene Output), TSC (Tuberous Sclerosis Complex), Volt-dept. K channels (voltage-dependent potassium channels), Volt-dept. Na channels (voltage-dependent sodium channels).

Table 2:

Devices

Purpose	Device	Company	Phase	Population
Stimulation	EASEE	Precisis GmbH	II	Drug-refractory epilepsy
	tDCS	Neuroelectrics	III	Drug-refractory epilepsy
Electronic seizure diaries		Signant Health		All
		WCG Clinical Endpoint Solutions		All
Seizure detection devices	EmbarcePlus	Empatica, Inc.		All
	EpiCare@Home	EpiCare@Home		All
	REMI	Epitel, Inc.		All
	AWARE	Epihunter		All

Abbreviations: AWARE (Automated Awareness Testing in Absence Epilepsy), EASEE (Epicranial Application of Stimulation Electrodes for Epilepsy), REMI (Remote EEG Ambulatory Monitoring), tDCS (transcranial Direct Current Stimulation).

Key points box.

• This review discusses the latest drugs and devices in the epilepsy diagnostics and therapeutics pipeline.

• Many antiseizure medications plus other genetic and non-genetic therapeutics are currently under development, including numerous that are first-in-class.

• Seizure tracking, detection, and stimulation devices all continue to make advances as devices enter the pipeline.