Composition for prevention or treatment of epilepsy, and epilepsy therapeutic agent screening method and composition therefor

Abstract

The present invention relates to a pharmaceutical composition for the prevention or treatment of epilepsy, and an epilepsy therapeutic agent screening method and a composition therefor, and specifically provides an epilepsy therapeutic agent screening method and an epilepsy therapeutic agent screening composition using induced pluripotent stem cells (iPSCs) derived from an epilepsy patient having a genotype comprising an SCN2A gene mutation, and accordingly, a pharmaceutical composition for the prevention or treatment of epilepsy having more excellent efficacy and selectivity than conventional epilepsy therapeutic agents may be screened and provided.

Description

Composition for preventing or treating epilepsy, and method for screening for epilepsy treatment agent and composition thereof

[0001] The present invention relates to a composition for preventing or treating epilepsy, a method for screening an epilepsy treatment agent, and a composition thereof.

[0002] Epilepsy is a complex and diverse neurological disorder characterized by recurrent spontaneous seizures, estimated to affect 0.5–1% of the global population (approximately 50–100 million people). The pathophysiology of epilepsy is multifaceted and involves a complex interplay of genetic, environmental, and physiological factors, making seizure control challenging. While many existing medications are available to treat this condition, the heterogeneity of epilepsy syndromes means that more than 30–40% of patients still struggle with seizure control. Given the diverse etiology of epilepsy, there is a growing emphasis on personalized medicine tailored to each patient's individual genetic and pathophysiological phenotype.

[0003] One of the major causes of epilepsy is genetic, meaning that mutations in certain genes can cause epilepsy. SCN2A encodes the type 2 alpha subunit of the voltage-gated sodium channel subunit 1.2 (Nav 1.2). Mutations in this gene are known to cause epilepsy spectrum disorders or neurodevelopmental disorders. SCN2A is essential for the rapid depolarization phase of neuronal action potentials, and sodium channels are important for the initiation and propagation of action potentials. Gain-of-function mutations in SCN2A can cause neurons to fire more readily and excessively, leading to epilepsy. Conversely, loss-of-function mutations typically result in decreased activity, disrupting the balance of neuronal activity and causing epilepsy, but sometimes have different characteristics. Specifically, missense mutations in SCN2A are associated with early childhood epilepsy (age of seizure onset <3 months) and benign self-limited seizures. In contrast, truncation mutations often lead to late-onset epilepsy (age of seizure onset >3 months) and cause intractable developmental and epileptic encephalopathy.

[0004] Recent advances in genomics, stem cell technology, and computational drug discovery have opened new avenues for personalized treatment, enabling the advancement of treatment into the era of precision medicine. Whole genome sequencing can identify mutations within disease-causing genes. Furthermore, establishing patient-specific stem cells that reflect the identified genetic mutations present in epilepsy patients can enable the pathogenesis of the underlying mechanisms of seizure activity. Furthermore, computational methods can predict the conformation of mutant SCN2A proteins, enable virtual screening or directed design of compounds that optimally interact with the mutant protein, and suggest clinically safe drug candidates from the identified compounds.

[0005] Therefore, to ensure the success of epilepsy treatment, it is important to identify the genetic mutations involved in each patient's seizures, study the mechanisms of these mutations using patient-specific stem cell models, understand the diverse pathologies of epilepsy patients, and ultimately develop personalized antiepileptic drugs for each patient.

[0006] The purpose of the present invention is to provide a pharmaceutical composition for preventing or treating epilepsy.

[0007] Another object of the present invention is to provide a method for screening for an epilepsy treatment agent.

[0008] Another object of the present invention is to provide a composition for screening an epilepsy treatment agent.

[0009] In this study, we identified a missense mutation, c.4886G>T (R1629L), in the SCN2A gene in a 4-year-old patient with epilepsy who developed seizures 4 days after birth and was refractory to sodium channel blockers, including phenytoin. To study the pathophysiology of this patient with the SCN2A mutation in vitro, we established induced pluripotent stem cells (iPSCs) from the patient and differentiated them into inhibitory (GABAergic) and excitatory (glutamatergic) neurons. Using these differentiated cells, we investigated changes in electrophysiological activity due to the genetic mutation and successfully reproduced abnormal seizure-like behaviors in vitro. Furthermore, we performed a docking-based screening based on the mutant structure of the SCN2A protein, computationally evaluating 1.6 million compounds to identify the top five compounds with high binding affinity for the mutant SCN2A protein. Subsequently, the candidate compounds were applied to patient-derived neuronal cell models to evaluate whether electrophysiological indicators were restored. As a result, multiple compounds significantly suppressed seizure-related neural activity, confirming their potential as potential anti-seizure therapeutics targeting the SCN2A R1629L mutation.

[0010] Hereinafter, the present invention will be described in more detail.

[0011] Accordingly, the present invention provides a pharmaceutical composition for preventing or treating epilepsy, which comprises at least one compound represented by the following chemical formulas 1 to 5 as an active ingredient.

[0012] [Chemical Formula 1]

[0013]

[0014] In the above chemical formula 1, R1 is an alkyl group having 1 to 3 carbon atoms.

[0015] [Chemical Formula 2]

[0016]

[0017] [Chemical Formula 3]

[0018]

[0019] In the above chemical formula 3, R2, R3, R4 and R5 are each independently an alkyl group having 1 to 3 carbon atoms.

[0020] [Chemical Formula 4]

[0021]

[0022] In the above chemical formula 4, R6, R7 and R8 are each independently an alkyl group having 1 to 3 carbon atoms.

[0023] [Chemical Formula 5]

[0024]

[0025] In the above chemical formula 5, R9, R10 and R11 are each independently an alkyl group having 1 to 3 carbon atoms, and R12, R13 and R14 are each independently one of hydrogen (H), fluorine (F), chlorine (Cl) or bromine (Br).

[0026] Additionally, the epilepsy may be characterized as being caused by the R1629L mutation of the SCN2A (sodium voltage-gated channel alpha subunit 2) gene.

[0027] In addition, the compounds represented by the above chemical formulas 1 to 5 may be characterized by having a binding affinity of -10 kcal / mol or less for the SCN2A protein containing the R1629L mutation.

[0028] In addition, the compounds represented by the above chemical formulas 1 to 5 have IC for the SCN2A protein containing the R1629L mutation. 50 It can be characterized by a value of 18 μM or less.

[0029] In addition, the compounds represented by the above chemical formulas 1 to 5 may be characterized by exhibiting anti-seizure activity that reduces the average firing rate to 0.5 Hz or less.

[0030] In addition, the present invention provides a method for screening for an epilepsy treatment agent using induced pluripotent stem cells (iPSCs) derived from an epilepsy patient having a genetic type including a mutation in the SCN2A (sodium voltage-gated channel alpha subunit 2) gene.

[0031] (a) a step of producing induced pluripotent stem cells (iPSCs) from peripheral blood mononuclear cells of the patient;

[0032] (b) a step of differentiating the iPSC into inhibitory neurons and excitatory neurons;

[0033] (c) a step of culturing the differentiated neural cells on a micro electrode array (MEA) plate to build an in vitro neural network model;

[0034] (d) treating the candidate compound to the in vitro neural network model and measuring one or more of the following indices: activity index (average firing rate), electrode burst index, network burst index, and synchrony index; and

[0035] (e) A method is provided comprising a step of evaluating the anti-seizure activity of the candidate compound using the measured index value.

[0036] The above step (e) may be characterized by evaluating the anti-seizure activity of the candidate compound based on an index value measured in homogeneous iPSC-derived neural cells in which a mutation in the SCN2A gene has been corrected by a gene editing technology.

[0037] The above candidate compounds may be characterized by being selected through binding affinity to the SCN2A protein containing the above mutation.

[0038] The above inhibitory neurons may be characterized as being differentiated through expression of the ASCL1 and DLX2 genes, and the above excitatory neurons may be characterized as being differentiated through expression of the NGN2 gene.

[0039] The above anti-seizure activity can be characterized by being evaluated as having effective anti-seizure activity when the average firing rate is reduced to 0.5 Hz or less.

[0040] The mutation of the above SCN2A gene may be characterized by including an R1629L mutation.

[0041] In addition, the present invention provides a composition for screening an epilepsy treatment agent comprising inhibitory neurons and excitatory neurons differentiated from induced pluripotent stem cells (iPSCs) derived from a patient having a genotype including a mutation in the SCN2A (sodium voltage-gated channel alpha subunit 2) gene.

[0042] The above-mentioned neural cells can be cultured on a microelectrode array (MEA) plate to form a neural network model exhibiting electrophysiological responsiveness.

[0043] The composition may be characterized by including a control model that can be compared, provided together with an isogenic neuronal cell in which a mutation in the SCN2A gene has been corrected.

[0044] The mutation of the above SCN2A gene may be characterized by including an R1629L mutation.

[0045] In one embodiment of the present invention, a computational structure-based screening according to the present invention was applied to 140,000 pharmacologically relevant compounds. For the screening, the structure of a mutant SCN2A protein was predicted through molecular dynamics simulation using the wild-type SCN2A protein structure (Figure 6), and structural differences compared to the wild-type SCN2A were identified (Figure 6B). The mutation R1629L is located in segment 4 of domain IV (DIV), and changing amino acid residue R to L results in the loss of a hydrogen bond between Q268 and R1629. This change affects the structural stability of the SCN2A protein, potentially contributing to the severe early childhood epilepsy phenotype. The predicted mutant SCN2A protein structure was used to estimate the binding affinities of the 140,000 compounds, and the top 100 compounds with high binding affinity for the mutant SCN2A protein were selected.

[0046] In the present invention, the SCN2A (sodium voltage-gated channel alpha subunit 2 or voltage-gated sodium channel subunit 1.2) gene may be a gene encoding a protein constituting an intracellular sodium channel. The sodium channel may be a voltage-gated sodium channel.

[0047] The above SCN2A (sodium voltage-gated channel alpha subunit 2 or voltage-gated sodium channels sub unit 1.2) gene (NCBI Locus NM_021007, NCBI accession No. NM_021007.2) may include the base sequence of sequence number 1.

[0048] In the present invention, the mutation may be a missense mutation, and more specifically, may include an R1629L mutation in which arginine (R) is substituted with leucine (L) by a base substitution of c.4886G>T.

[0049] In the present invention, the epilepsy may be pediatric epilepsy that appears in childhood.

[0050] In one embodiment of the present invention, the efficacy of compounds selected according to the present invention was evaluated and a leading drug candidate was identified as a therapeutic agent for controlling seizures in pediatric epilepsy patients with SCN2A mutation (R1629L). The selected compounds were experimentally tested on the in vitro neural network platform of the present invention derived from patient-derived iPSCs. Phenytoin was used as a control. Electrophysiological recordings of neurons were performed by applying various concentrations of compounds and IC in terms of firing rate. 50 The values ​​were calculated. Results showed that phenytoin significantly reduced the mean firing rate at concentrations starting from 10 μM, with IC 50 was 18 μM. In contrast, compounds 2786-4513 and 4333-1696 showed much higher potency, with IC 50 The values ​​were 0.25 μM and 0.2 μM, respectively. The compounds effectively reduced spontaneous neuronal firing and demonstrated potential as targeted therapies for SCN2A-associated epilepsy. In addition to firing rate, other indicators such as burst frequency, network burst frequency, and synchrony index were also evaluated. Phenytoin reduced burst activity at concentrations above 30 μM and showed efficacy in reducing network burst and synchrony at 10 μM. The new compounds, particularly 2786-4513 and 4333-1696, were effective even at lower concentrations, indicating their potential to control seizure-like activity.

[0051] Meanwhile, the corresponding features can be replaced in the above-described part, so their description is omitted.

[0052] The present invention's epilepsy treatment screening method and epilepsy treatment screening composition can be utilized to apply a personalized precision medicine approach to epilepsy treatment. This allows for the screening and provision of pharmaceutical compositions for the prevention or treatment of epilepsy with superior efficacy and selectivity compared to conventional epilepsy treatments. Furthermore, patient-derived neural network modeling and computational therapeutic screening can be used to suggest personalized treatment strategies, effectively overcoming the limitations of conventional epilepsy treatment screening for patients who do not respond to existing treatments.

[0053] Figure 1 outlines a precision medicine approach for pediatric epilepsy patients with mutations in the SCN2A gene. Figure 1 describes the research methodology, which includes identifying the genetic defect, establishing an in vitro neural network model, computationally screening drug candidates, and evaluating drug efficacy using an in vitro platform.

[0054] Figure 2a shows a timeline for reprogramming induced pluripotent stem cells (iPSCs) derived from a pediatric epilepsy patient from mononuclear cells.

[0055] Figure 2b shows pluripotency markers of iPSCs: OCT3 / 4 (right) and SSEA4 (left); NANOG (right) and TRA 1-81 (left); SOX2 (right) and TRA 1-60 (left). Scale bar = 100 μm.

[0056] Figure 2c shows the immunohistochemical results for three germ layer markers: TUJ1 (left) and PAX6 (right) for the ectoderm, alpha-SMA (left) and brachyury (BRY) (right) for the mesoderm, and AFP (left) and FOXA2 (right) for the endoderm. Scale bar = 100 μm.

[0057] Figure 2d shows the results of G-banding analysis of iPSC karyotype.

[0058] Figure 2e shows the results of STR analysis of genes in iPSCs and mononuclear cells derived from SCN2A patients.

[0059] Figure 2f shows a sequencing chromatogram of patient-derived iPSCs, confirming the same genetic mutation (R1629L).

[0060] Figure 3a shows a schematic of the overall experimental method design using CRISPR / Cas9 for correcting SCN2A mutations in iPSCs. (From top to bottom) (First row of lines) Timeline of the entire CRISPR / Cas9 study. Required cell culture factors include (second row of lines) 10 μM pifithrin-alpha, used to prevent p53-induced apoptosis; (third row of lines) 10 μM SCR7, used to enhance HDR efficiency; (fourth row, left line) FACS area; and (fourth row, right line) anti-Puro, used to ultimately select unedited cells not selected by FACS. (Left and right bars) Culture on Matrigel-coated 35π dishes; (middle bar) Culture on Matrigel-coated 96-well plates. (Short arrow) Cells were seeded onto Transwells for subculture. (Long arrow) Date of Sanger sequencing.

[0061] Figure 3b illustrates the design of the CRISPR / Cas9 vector. It contains gRNA and hCAS9, and the selection marker is EGFP / Anti-Puro. The total length is 10,603 bp.

[0062] Figure 3c shows the guide RNA location.

[0063] Figure 3d shows the ssODN (donor) position sequencing.

[0064] Figure 3e shows the results of Sanger sequencing, demonstrating successful gene editing using CRISPR / Cas9. The c.4886G>T mutation was successfully corrected in the heterozygous strain, resulting in a silent mutation.

[0065] Figure 3f shows that the corrected lineage of the SCN2A patient was confirmed by whole genome sequencing.

[0066] Figure 3g shows the off-target location confirmed.

[0067] Figure 3h identifies off-target sites similar to the sgRNA sequence of the gene. Sequences with fewer than three mismatches were identified based on the score, indicating no off-target effects.

[0068] Figure 4a shows the differentiation of patient-derived iPSCs into excitatory and inhibitory neurons, and shows the lentiviral constructs for differentiation into inhibitory (GABAergic) neurons: hASCL1 and hDLX2 genes were expressed to induce differentiation, and mCherry gene was co-expressed to confirm infection. Anti-Puro was used as a selection marker.

[0069] Figure 4b shows the lentiviral construct for differentiation into excitatory (glutamatergic) neurons: the hNGN2 transcription factor was expressed to promote differentiation, and the eGFP gene provided fluorescent confirmation of lentiviral infection.

[0070] Figure 4c shows the experimental schedule for differentiating iPSCs into neural cells.

[0071] Figure 4d shows fluorescence measurements of robust gene expression 21–42 days after lentivirus infection and doxycycline treatment. Fluorescence of mCherry (first and third rows) and eGFP (second and fourth rows). Scale bar = 100 μm.

[0072] Figure 4e shows biomarker transcript levels of patient-derived mutant neurons and isogenic neurons confirming successful differentiation into the expected neuronal cell types: GAD67 [GAD1] for GABAergic neurons and vGLUT2 for glutamatergic neurons.

[0073] Figure 4f shows biomarker transcript levels of patient-derived mutant neurons and isogenic neurons confirming successful differentiation into the expected neuronal cell types: GAD67 [GAD1] for GABAergic neurons and vGLUT2 for glutamatergic neurons.

[0074] Figure 5a shows the electrophysiology of patient-derived and isogenic neurons on a microelectrode array (MEA) plate, along with fluorescent images of cultured neurons on the MEA plate taken 3–5 weeks after doxycycline treatment (see Figure 4c). Dark-colored cells represent GABAergic neurons, and light-colored cells represent glutamatergic neurons.

[0075] Figure 5b is a heatmap showing neuron firing events at individual electrodes over a specified period (15 minutes). White and red indicate high firing frequency, while blue indicates low firing frequency.

[0076] Figure 5c shows the results of spontaneous electrical signal spike measurements in a neuronal network of an MEA plate recorded for 15 minutes.

[0077] Figure 5d is a spike raster plot of each electrode, showing spikes recorded over 100 seconds. Each tick in the plot represents a spontaneous firing event, and each row corresponds to an individual electrode.

[0078] Figure 5e shows activity indices of patient-derived neurons and isogenic neurons (see Materials and Methods).

[0079] Figure 5f shows burst frequency as an electrode burst index of patient-derived neurons and isogenic neurons.

[0080] Figure 5g shows the burst percentage as an electrode burst index for patient-derived neurons and isogenic neurons.

[0081] Figure 5h shows the network burst frequency as a network burst index of patient-derived neurons and isogenic neurons.

[0082] Figure 5i shows the network burst percentage as a network burst indicator of patient-derived neurons and isogenic neurons.

[0083] Figure 5j shows the synchronization index of patient-derived neurons and isogenic neurons. (In Figures 5a - 5j, n = 6; *, p-value ≤ 0.05; **, p-value ≤ 0.01; ***, p-value ≤ 0.001; ****, p-value ≤ 0.0001; In Figures 5e - 5j, ● indicates SCN2A patient neurons, ▲ indicates isogenic neurons.)

[0084] Figure 6 shows the structures of the five screened compounds and the reference drug (phenytoin) in complex with the SCN2A(R1629L) protein. (A) Predicted structure of SCN2A(R1629L) containing four domains (DI-DIV). Each domain is color-coded, and the binding pocket is indicated by a dashed black box. (B) Structural differences between wild-type and mutant SCN2A proteins. (C-H) Compounds in complex with SCN2A(R1629L). (C) Phenytoin. (D) 4456-3635. (E) 2786-4513. (F) 4561-0590. (G) 4333-1696. (H) 6466-0224. Interactions of the compounds with the amino acids in close proximity are shown in Figures 7A-7F.

[0085] Figure 7a shows residues interacting with the screened compound (phenytoin) in SCN2A (R1629L).

[0086] Figure 7b shows the residues interacting with the screened compound (4456-3635) in SCN2A (R1629L).

[0087] Figure 7c shows the residues interacting with the screened compound (2786-4513) in SCN2A (R1629L).

[0088] Figure 7d shows residues interacting with the screened compound (4561-0590) in SCN2A (R1629L).

[0089] Figure 7e shows residues interacting with the screened compound (4333-1696) in SCN2A (R1629L).

[0090] Figure 7f shows residues interacting with the screened compound (6466-0224) in SCN2A (R1629L).

[0091] [Revised 29.08.2025 under Rule 91] Figure 8a shows the efficacy of phenytoin in SCN2A (R1629L) patient-derived neurons, showing recorded and calculated electrophysiological activities during compound treatment. (A1) Mean firing rate, (A2) IC estimated from mean firing rate. 50 As a value, this represents the concentration required for these compounds to inhibit the average ignition rate by 50%.

[0092] [Revised on 29.08.2025 under Rule 91] Figure 8b shows the efficacy of phenytoin in SCN2A (R1629L) patient-derived neurons, showing recorded and calculated electrophysiological activities during compound treatment. (A3) Burst frequency, (A4) Burst percentage.

[0093] [Revised on 29.08.2025 under Rule 91] Figure 8c shows the evaluation of the efficacy of phenytoin in SCN2A (R1629L) patient-derived neurons, showing recorded and calculated electrophysiological activities during compound treatment. (A5) Network burst frequency, (A6) Network burst duration.

[0094] [Revised on 29.08.2025 under Rule 91] Figure 8d shows the evaluation of the efficacy of phenytoin in SCN2A (R1629L) patient-derived neurons, showing the recorded and calculated electrophysiological activities during compound treatment. (A7) Motivational index.

[0095] [Revised on 29.08.2025 under Rule 91] Figure 8e shows the efficacy evaluation of 4456-3635 in SCN2A (R1629L) patient-derived neurons, showing the recorded and calculated electrophysiological activities during compound treatment. (B1) Mean firing rate, (B2) IC estimated from mean firing rate. 50 As a value, this represents the concentration required for these compounds to inhibit the average ignition rate by 50%.

[0096] [Revised 29.08.2025 under Rule 91] Figure 8f shows the efficacy evaluation of 4456-3635 in SCN2A (R1629L) patient-derived neurons, showing recorded and calculated electrophysiological activities during compound treatment. (B3) Burst frequency, (B4) Burst percentage.

[0097] [Revised on 29.08.2025 under Rule 91] Figure 8g shows the efficacy evaluation of 4456-3635 in SCN2A (R1629L) patient-derived neurons, showing recorded and calculated electrophysiological activities during compound treatment. (B5) Network burst frequency, (B6) Network burst duration.

[0098] [Revised 29.08.2025 under Rule 91] Figure 8h shows the efficacy evaluation of 4456-3635 in SCN2A (R1629L) patient-derived neurons, showing the recorded and calculated electrophysiological activity during compound treatment. (B7) Motivational index.

[0099] [Revised on 29.08.2025 by Rule 91] Figure 8i shows the efficacy evaluation of 2786-4513 in SCN2A (R1629L) patient-derived neurons, showing recorded and calculated electrophysiological activities during compound treatment. (C1) Mean firing rate, (C2) IC estimated from mean firing rate. 50 As a value, this represents the concentration required for these compounds to inhibit the average ignition rate by 50%.

[0100] [Revised 29.08.2025 under Rule 91] Figure 8j shows the efficacy evaluation of 2786-4513 in SCN2A (R1629L) patient-derived neurons, showing recorded and calculated electrophysiological activities during compound treatment. (C3) Burst frequency, (C4) Burst percentage.

[0101] [Correction under Rule 91 29.08.2025] Figure 8k shows the efficacy evaluation of 2786-4513 in SCN2A (R1629L) patient-derived neurons, showing recorded and calculated electrophysiological activities during compound treatment. (C5) Network burst frequency, (C6) Network burst duration.

[0102] [Correction under Rule 91 29.08.2025] Figure 8l shows the efficacy evaluation of 2786-4513 in SCN2A (R1629L) patient-derived neurons, showing the recorded and calculated electrophysiological activity during compound treatment. (C7) Motivational index.

[0103] [Revised 29.08.2025 under Rule 91] Figure 8m shows the efficacy evaluation of 4561-0590 in SCN2A (R1629L) patient-derived neurons, showing the recorded and calculated electrophysiological activities during compound treatment. (D1) Mean firing rate, (D2) IC estimated from mean firing rate. 50 As a value, this represents the concentration required for these compounds to inhibit the average ignition rate by 50%.

[0104] [Revised 29.08.2025 under Rule 91] Figure 8n shows the efficacy evaluation of 4561-0590 in SCN2A (R1629L) patient-derived neurons, showing recorded and calculated electrophysiological activities during compound treatment. (D3) Burst frequency, (D4) Burst percentage.

[0105] [Revised 29.08.2025 under Rule 91] Figure 8o shows the efficacy evaluation of 4561-0590 in SCN2A (R1629L) patient-derived neurons, showing recorded and calculated electrophysiological activities during compound treatment. (D5) Network burst frequency, (D6) Network burst duration.

[0106] [Revised 29.08.2025 under Rule 91] Figure 8p shows the efficacy evaluation of 4561-0590 in SCN2A (R1629L) patient-derived neurons, showing the recorded and calculated electrophysiological activity during compound treatment. (D7) Motivational index.

[0107] [Revised on 29.08.2025 by Rule 91] Figure 8q shows the efficacy evaluation of 4333-1696 in SCN2A (R1629L) patient-derived neurons, showing the recorded and calculated electrophysiological activities during compound treatment. (E1) Mean firing rate, (E2) IC estimated from mean firing rate. 50 As a value, this represents the concentration required for these compounds to inhibit the average ignition rate by 50%.

[0108] [Revised on 29.08.2025 under Rule 91] Figure 8r shows the efficacy evaluation of 4333-1696 in SCN2A (R1629L) patient-derived neurons, showing the recorded and calculated electrophysiological activities during compound treatment. (E3) Burst frequency, (E4) Burst percentage.

[0109] [Revised 29.08.2025 under Rule 91] Figure 8s shows the efficacy evaluation of 4333-1696 in SCN2A (R1629L) patient-derived neurons, showing recorded and calculated electrophysiological activities during compound treatment. (E5) Network burst frequency, (E6) Network burst duration.

[0110] [Revised 29.08.2025 under Rule 91] Figure 8t shows the efficacy evaluation of 4333-1696 in SCN2A (R1629L) patient-derived neurons, showing the recorded and calculated electrophysiological activity during compound treatment. (E7) Motivational index.

[0111] [Revised 29.08.2025 under Rule 91] Figure 8u shows the efficacy evaluation of 6466-0224 in SCN2A (R1629L) patient-derived neurons, showing the recorded and calculated electrophysiological activities during compound treatment. (F1) Mean firing rate, (F2) IC estimated from mean firing rate. 50 As a value, this represents the concentration required for these compounds to inhibit the average ignition rate by 50%.

[0112] [Revised 29.08.2025 under Rule 91] Figure 8v shows the efficacy evaluation of 6466-0224 in SCN2A (R1629L) patient-derived neurons, showing recorded and calculated electrophysiological activity during compound treatment. (F3) Burst frequency, (F4) Burst percentage.

[0113] [Revised 29.08.2025 under Rule 91] Figure 8w shows the efficacy evaluation of 6466-0224 in SCN2A (R1629L) patient-derived neurons, showing recorded and calculated electrophysiological activities during compound treatment. (F5) Network burst frequency, (F6) Network burst duration.

[0114] [Revised 29.08.2025 under Rule 91] Figure 8x shows the efficacy evaluation of 6466-0224 in SCN2A (R1629L) patient-derived neurons, showing the recorded and calculated electrophysiological activities during compound treatment. (F7) Motility index. (In Figures 8a - 8x, N = 4; *, p-value ≤ 0.05; **, p-value ≤ 0.01; ***, p-value ≤ 0.001; ****, p-value ≤ 0.0001) Figure 9 shows the mutations within the transmembrane region of the repeat IV domain of the SCN2A gene in pediatric epilepsy patients with SCN2A mutations in Table 3.

[0115] Hereinafter, to aid understanding of the present invention, examples will be given in detail. However, the following examples are intended only to illustrate the scope of the present invention and are not intended to limit its scope. These examples are provided to more fully explain the present invention to those of average skill in the art.

[0116] Example 1. Experimental method

[0117] 1.1. Genomic DNA Preparation and Whole Genome Sequencing

[0118] A genetic sample was collected from a 4-year-old male patient with epilepsy who presented with developmental delay and focal seizures and who was refractory to conventional antiepileptic drugs. Genomic DNA (gDNA) was extracted using the QIAmp DNA FFPE Tissue Kit (Qiagen) according to the manufacturer's instructions. gDNA quantity and quality were assessed using the Qubit ® 2.0 photometer (Thermo Fisher Scientific) and the Qubit ® dsDNA HS Assay Kit. Barcoded samples (eight samples) were loaded onto the 316 v2 BD chip and analyzed by the Ion PGM. TM Ion PGM from (Thermo Fisher Scientific) TM Hi-Q View TM Sequencing was performed using a sequencing kit (Thermo Fisher Scientific).

[0119] 1.2. Mutation Identification

[0120] Sequence analysis was performed using Torrent Suite software v. 5.0 (Thermo Fisher Scientific). After removing low-quality reads and trimming adapter sequences, alignment to the reference genome (hg19) was performed using the Torrent Mapping Alignment Program. Variants were identified in the reference sequence using the Torrent Variant Caller plugin. To identify pathogenic variants, mutations that do not affect protein-coding regions (i.e., mutations in introns and untranslated regions, and silent mutations within exons) were filtered out. All detected variants were manually reviewed using the Integrative Genomics Viewer (IGV V.2.1, Broad Institute). The Genomic Evolutionary Rate Profiling tool was used to predict the impact of missense mutations on translated proteins and calculate conservation scores. This analysis was complemented with the ClinVar and Varsome databases. To ensure high-confidence detection of somatic mutations present in the patient gDNA, samples with coverage less than 100X and mutation frequencies less than 5% were excluded.

[0121] 1.3. Generation and maintenance of patient-derived human iPSCs, construction of clustered regularly interspaced short palindromic repeats (CRISPR) / Cas9 vectors, and editing of the SCN2A gene.

[0122] 1.3.1. iPSC reprogramming

[0123] Reprogramming was performed using mononuclear cells from SCN2A patients with Epi5 TM It was performed using the Episomal iPSC Reprogramming Kit (Invitrogen). Cells were used as StemSpan TM SFEM II and StemSpan TMCultured for 6 days in a medium containing Erythroid Expansion Supplement (100Х) and then in Opti-MEM TM I was suspended in Reduced Serum Medium (Invitrogen, Middlesex County, Massachusetts, United States). Electroporation was performed on Epi5 TM Reprogramming Vector and Epi5 TM Electroporation was performed using the NEPA21 Super Electroporator (NEPA GENE, Shiba, Japan) with p53 & EBNA vectors. The poring pulse was set to 150 V, 5 ms, and 10% duty rate, and the transfer pulse was set to 20 V, 50 ms, and 40% duty rate.

[0124] Cell culture dishes were pre-coated with Corning Matrigel Basement Membrane Matrix and DMEM / F12 (Invitrogen, Middlesex County, Massachusetts, United States). The culture medium included Neurobasal TM Medium, N-2 Supplement, B-27 TM Supplement Y-27632 dihydrochloride (MedCamExpress, Monmouth Junction, New Jersey, United States) was included at 10 nM. Electroporated PBMCs were added and cultured for 9–21 days. Basic Fibroblast Growth Factor (Merck, Darmstadt, Germany) was added to a final concentration of 100 ng / mL. iPSC colonies were visually inspected between days 9 and 21, harvested, and cultured for further characterization.

[0125] 1.3.2. iPSC culture

[0126] Human wild-type SCN2A iPSCs derived from episomal vectors were expressed in mTeSR TM 1 (STEMCELL Technologies, Vancouver, Canada). iPSCs were maintained in mTeSR in feeder-free conditions. TM 1 were cultured on Matrigel-coated dishes with a medium. Initially, iPSCs were seeded at 1X10 per 35 mm dish. 5 Cells were seeded at a density of 100 μL. Culture medium was changed daily. The CEPT cocktail (containing Chroman 1, Emricasan, Polyamine Supplement, and Trans-ISRIB) was used when thawing frozen cells and during subculture to minimize cell damage and enhance viability.

[0127] To characterize iPSCs, the expression levels of several markers were measured at the protein level using immunocytochemistry, including pluripotency markers (NANOG, OCT4, SOX2) and surface markers (TRA-1-60, SSEA4, TRA-1-81). The antibodies used are listed in Table 1 [Antibodies Used for iPSC Characterization].

[0128]

[0129]

[0130] Cells were passaged when they occupied more than 80% of the plate surface area or after a maximum of 5 days. The freezing medium used was 40% mTeSR1, 50% KnockOut TM Serum Replacement (KOSR) and 10% dimethyl sulfoxide (DMSO) were used. Subculture was performed according to the manufacturer's instructions using ReLeSR. TMwas performed using , and cells were transferred at a ratio of 1:10. Cells were passaged at least 5 times to stabilize them, and additional passages were performed if necessary. The remaining cells were stored at -20°C for 1 day and then deep-frozen at -70°C to -80°C before storage in liquid nitrogen.

[0131] 1.3.3. Karyotype testing

[0132] G-banding karyotyping was performed at passage 20 using a standard GTG banding protocol with 550 band resolution at DX&VX (Seoul, Korea).

[0133] 1.3.4. Immunocytochemistry

[0134] iPSCs were fixed with 4% paraformaldehyde in PBS for 15 minutes, permeabilized with 0.1% Triton X-100 in PBS for 15 minutes, and blocked with 2% bovine serum albumin in PBS for 1 hour. Cells were incubated with primary antibodies overnight at 4°C, followed by incubation with fluorescently labeled secondary antibodies for 2 hours at room temperature. Samples were mounted with VECTASHIELD® Antifade Mounting Medium containing DAPI and images were taken using an IX74 microscope equipped with a pE-300lite LED fluorometer and a DP74 digital camera (Olympus, Tokyo, Japan). Antibodies used for immunocytochemistry are listed in Tables 1 and 2.

[0135] 1.3.5. Differentiation into three germ layers

[0136] To assess in vitro differentiation into the three germ layers, AggreWell TM Embryoid bodies (EBs) were formed using 800 (STEMCELL Technologies, Vancouver, Canada). After adding Anti-Adherence Rinsing Solution and centrifuging, mTeSR and CEPT were added at 1X10 6were added together with iPSCs. After 1 day, EBs were formed and cultured for 6 days in DMEM / F12 medium containing KOSR, penicillin-streptomycin, MEM nonessential amino acids, and 2-mercaptoethanol. The medium was replaced by removing 0.5 ml of EB medium and adding fresh medium to maintain a volume of 1 ml. EBs were then transferred to Matrigel-coated dishes and cultured for 14–21 days, and differentiation into the three germ layers was confirmed through immunohistochemistry. The antibodies used to characterize the differentiated germ layers are listed in Table 2 [Antibodies Used for Characterization of Differentiated Three Germ Layers].

[0137]

[0138]

[0139] 1.3.6. Genomic DNA Extraction and Variant Identification

[0140] Genomic DNA was extracted using the Blood & Cell Culture DNA Midi Kit (QIAGEN, Hilden, Germany). For extraction, iPSCs were passaged until sufficient numbers were obtained, treated with Accutase, and cells were counted using a hemacytometer. PCR was performed using the AccuPower Taq PCR PreMix primers: 5'-TGTACTTGGCCACTGTATGC-3' and 5'-CCAACAGATGGGTTCCCACA-3'. Cycling conditions included 95°C for 5 min, followed by 35 cycles of 95°C for 30 s, 53°C for 30 s, and 72°C for 1 min, with a final extension of 72°C for 5 min.

[0141] 1.3.7. STR Analysis

[0142] Short tandem repeat (STR) analysis was performed by Cosmogentech Inc. (Seoul, Republic of Korea) to compare allelic repeats at specific loci between samples.

[0143] 1.3.8. CRISPR / CAS9 vector design

[0144] Guide RNAs (gRNAs) were designed using Merck CRISPR Design Tools and validated using Benchling's CRISPR / CAS9 Tool. CRISPR / CAS9 vectors were designed using VectorBuilder tools, and sgRNAs were generated based on the mutation sites. ssODNs were synthesized by BIONICS.

[0145] 1.3.9. Editing SCN2A using CRISPR / CAS9

[0146] We generated heterozygous indel mutations in the SCN2A gene using CRISPR / Cas9. The sgRNA was cloned into a CRISPR vector and electroporated into iPSCs. After electroporation, cells were treated with SCR7 pyrazine and the p53 inhibitor pifithrin-alpha to enhance HDR and reduce apoptosis.

[0147] 1.3.10. Isogeneic cell culture

[0148] After CRISPR / CAS9 editing, cells were screened with puromycin, transferred to 96-well plates, and cultured for 5 days. Cells were split, and gDNA was extracted for sequencing. Successfully edited cells were transferred to Matrigel-coated 6-well plates and cultured with mTeSR1 and CEPT.

[0149] 1.3.11. Off-target identification

[0150] Potential off-target sites were identified using CAS-Offinder and Benchling. Primers were designed for these sites and Sanger sequencing was performed. Whole-genome sequencing (WGS) was performed at Macrogen (Seoul, Korea) to identify on-target and off-target effects.

[0151] 1.4. Construction of lentiviruses for programmed neural cell differentiation

[0152] To construct a seizure-mimicking neuronal network, patient-derived iPSCs were differentiated into inhibitory and excitatory neurons, reflecting the forebrain cortical environment: GABAergic and glutamatergic neurons, respectively. To promote the differentiation of iPSCs into GABAergic neurons that influence GABAergic neurotransmission, iPSCs were transfected with engineered lentiviruses expressing two transcription factors, distal-less homeobox 2 (DLX2) and achaete-scute homolog 1 (ASCL1), via the Tet-On system. To differentiate iPSCs into glutamatergic neurons, iPSCs were transfected with another engineered lentivirus expressing the transcription factor Neurog2 (NGN2). To confirm transfection, both lentiviruses were engineered to express reporter genes: the former lentivirus expressing red fluorescent protein (RFP) and the latter lentivirus expressing enhanced green fluorescent protein (eGFP).

[0153] 1.5. Differentiation of patient-derived iPSCs into neural cells

[0154] The differentiation procedure was modified from a previously published method by introducing a new timeline (Fig. 4a). In brief, 1 Х 10 6 iPSCs were seeded on 6-well culture plates coated with Matrigel and cultured in neural growth medium (NGM) for more than 1 day.

[0155] NGM is a 1X dose supplement of Neurobasal Plus (Gibco, Grand Island, New York, United States) with N2Serum-Free Supplement (Gibco, Grand Island, New York, United States) and B-27 TMPlus Supplement (Gibco, Grand Island, New York, United States) at a 1X dose, L-ascorbic acid at a final concentration of 100 μM, recombinant human / murine / rat brain-derived neurotrophic factor (Peprotech, Cranbury, New Jersey, United States) at a final concentration of 10 ng / μl, recombinant human glial cell line-derived neurotrophic factor (Peprotech, Cranbury, New Jersey, United States) at a final concentration of 10 ng / μl, dibutyryl cyclic adenosine monophosphate (dbcAMP, MedCamExpress, Monmouth Junction, New Jersey, United States) at a final concentration of 250 ng / μl, and GlutaMAX TM Supplement (Gibco, Grand Island, New York, United States) was added.

[0156] Differentiation was then initiated by adding 1 titer of lentivirus containing DLX2 / ASCL1 or NGN2 and 1 titer of lentivirus containing reverse tetracycline-controlled transcription activator (rtTA) to the culture medium. Doxycycline monohydrate (Sigma-Aldrich, St. Louis, Missouri, United States) was also added to the culture medium at a final concentration of 1 μg / ml.

[0157] On the 1st day after infection, successful virus entry was confirmed by RFP or eGFP fluorescence. Puromycin was added to the medium at 1 μg / ml daily for 3 days to eliminate uninfected cells. Simultaneously, cells were treated with 1-β-D-arabinofuranosylcytosine (Ara-C, Sigma-Aldrich, St. Louis, MO, USA) at a final concentration of 1 μg / ml for 2 days to prevent cell division. In addition, cells were treated with forskolin (Sigma-Aldrich, St. Louis, MO, USA) at a final concentration of 20 μM for 2 days to stimulate adenylyl cyclase activity and increase intracellular cyclic AMP production. The culture medium was replaced with fresh medium every day after transfection.

[0158] 1.6. Measurement of neuronal maturation and electrophysiological activity

[0159] For neuronal maturation, immature neurons were co-cultured with mouse astrocytes prepared in a conventional manner (Schildge S, Bohrer C, Beck K, et al. J Vis Exp 2013(71):e50079.). For characterization, coverslips of 24-well plates were coated with Matrigel and then cultured at 5 Х 10 4 Astrocytes from dogs were loaded onto coated coverslips. One day later, immature neurons (1.5 Х 10 4 GABAergic neurons of the dog and 3.5 Х 10 4Astrocytes were co-cultured with glutamatergic neurons (GlNs) and cultured for up to 6 weeks. For electrophysiological experiments, astrocytes were cultured on microelectrode arrays (MEAs), and the next day, immature neurons were seeded onto the arrays, and the cells were cultured for up to 6 weeks. During co-culture, half of the culture medium was replaced every other day. Spontaneous electrophysiological activity was recorded for 15 min using a Maestro Edge MEA machine (Axion BioSystems, Atlanta, GA, USA) in a recording chamber maintained at 37°C, 95% O2, and 5% CO2.

[0160] To count the number of astrocytes and immature neurons, cells were treated with Accutase and counted using a hemocytometer.

[0161] 1.7. Computational Drug Candidate Screening

[0162] Drug candidates were computationally screened from the 1.6 million compounds in the ChemDiv library (San Diego, CA, USA). To exclude pharmacologically irrelevant compounds, the machine learning models LightBBB18 for blood-brain barrier (BBB) ​​permeability, CardPred for cardiotoxicity (i.e., human ether-a-go-go-related gene (hERG) inhibition), and DILI-Stk20 for hepatotoxicity were used to predict the pharmacological properties of the compounds.

[0163] For the mutant SCN2A protein structure-based prediction, the three-dimensional (3D) structure of SCN2A was retrieved from the Protein Data Bank (PDB; accession number 6J8E). After introducing the mutation (R1629L) identified in an epilepsy patient, energy minimization was performed using PyMOL. Pharmacologically useful compounds (i.e., compounds that are BBB-permeable and do not cause cardiac or hepatic side effects) were docked into the mutant SCN2A structure in the binding cavity, and binding energies were calculated using the PyRx tool. Compounds were then ranked based on their estimated binding affinities, and the top 100 compounds were selected for further analysis. The physicochemical properties, or fingerprints, of the top 100 compounds were calculated using DRAGON software, and the physiological distances, expressed as the Tanimoto distance, were calculated. Using the distance metric, the top 100 compounds were clustered using the spectral clustering algorithm. Representative compounds from each cluster were then selected to test their anti-seizure efficacy. Candidate compounds were purchased from ChemDiv (San Diego, CA, USA).

[0164] 1.8. Pharmacological Evaluation of Drug Candidates

[0165] The efficacy of the selected representative compounds and the reference compound (phenytoin) was evaluated using patient-derived neurons grown on MEA plates for 6 weeks. Before applying the selected compounds, electrical signals from the neurons were recorded for 15 minutes to serve as a baseline for the pretreatment condition. The compounds were then dissolved in dimethyl sulfoxide (DMSO) and applied to the neurons cultured on MEAs for 10 minutes. The electrical signals from the treated neurons were then recorded for 15 minutes using an MEA recorder.

[0166] Several metrics were calculated from the electrical signals. The mean firing rate is defined as the number of spikes per second. The burst frequency represents the number of spike clusters, or bursts, per second. The burst percentage represents the number of spikes in a single electrode divided by the total number of spikes. The network burst frequency represents the total number of network bursts per second. The network burst percentage is defined as the number of spikes in a network burst divided by the total number of spikes multiplied by 100. The synchrony index represents a unitless measure of synchrony ranging from 0 to 1.26, with values ​​closer to 1 indicating greater synchrony. Each metric was averaged across samples.

[0167] 1.9. Quantification and Statistical Analysis

[0168] Analyses were performed using one-way ANOVA with sequential post hoc Bonferroni correction when comparing two variables at a single time point, and a p value less than 0.05 was considered significantly different. Statistical analyses were performed using GraphPad Prism 9 (GraphPad Software Inc., Boston, MA, USA). In all graphs, data points are expressed as the mean ± standard error of the mean.

[0169] 1.10. Ethical Approval and Registration

[0170] Ethical approval for this invention was obtained from the Yonsei University Institutional Review Board (4-2023-1144).

[0171] Example 2. Experimental Results

[0172] The present invention aimed to identify effective drug candidates for pediatric epilepsy patients who do not respond well to existing sodium channel blockers due to a mutation in the SCN2A gene (R1629L). The overall experimental strategy is illustrated in Figure 1.

[0173] 2.1. Identification of SCN2A gene mutations in pediatric epilepsy patients

[0174] Detailed clinical information for a pediatric patient who did not respond effectively to sodium channel blockers is summarized in Table 3 [Pediatric Epilepsy Patients with SCN2A Mutations]. To identify the genetic defect causing the patient's epilepsy, we sequenced the patient's whole genome and detected the c.4886G>T (R1629L) mutation in the SCN2A gene. The SCN2A gene encodes a subunit of the sodium channel, and dysfunction of this gene is known to be associated with a wide range of brain disorders, including epilepsy. Of the two copies of the SCN2A gene encoded in the patient's genome, only one SCN2A gene harbored the mutation, while the other was wild-type. The mutation was located in the coding region of SCN2A, specifically in the transmembrane region of the repeat IV domain, and is described in Table 3 at the mutation location. This pathogenic mutation is known to cause benign familial childhood seizures and phenotypic manifestations, including developmental and epileptic encephalopathy.

[0175] [Correction pursuant to Rule 91, August 29, 2025]

[0176]

[0177] 2.2. Construction of patient-derived iPSCs harboring mutations in the SCN2A gene or corrected SCN2A gene.

[0178] To successfully establish patient-derived iPSCs, peripheral blood mononuclear cells from the patient were obtained using a conventional method and then reprogrammed into iPSCs (Loh YH, Agarwal S, Park IH, et al. The Journal of the American Society of Hematology 2009;113(22):5476-9. Staerk J, Dawlaty MM, Gao Q, et al. Cell Stem Cell 2010;7(1):20-4.). We confirmed that the iPSCs were successfully reprogrammed and harbored the same genetic mutation identified in the patient's genome (Fig. 2a-e). To establish an isogenic cell line with the corrected genetic mutation relative to the wild-type reference, the mutated nucleotide (c.4886G>T to c.4886G) in the SNC2A gene was corrected using homology-directed repair technology (Fig. 3a-h).

[0179] 2.3. Differentiation of patient-derived iPSCs into excitatory and inhibitory neurons

[0180] Patient-derived iPSCs were differentiated into inhibitory and excitatory neurons using a lentiviral differentiation method to reflect the neuronal network exhibiting seizure activity. Differentiation into GABAergic neurons was driven by two transcription factors, ASCL1 and DLX210, 11 (Fig. 4a), whereas differentiation into glutamatergic neurons was initiated by a single transcription factor, NGN212, 13 (Fig. 4b).

[0181] During neuronal differentiation (Fig. 4c), cells were treated with cytarabine (Ara-C) to inhibit glial proliferation and promote neuronal differentiation and maturation. Additionally, forskolin was supplemented to promote neuronal differentiation by regulating cAMP-cAMP-responsive element binding protein 1 (CREB1)-Janus kinase (JNK) signaling. To complete the expected neuronal differentiation and maturation, differentiating neurons were co-cultured with astrocytes on MEA plates 12 days after doxycycline treatment. Stable maintenance of transcription factors carried by the engineered lentivirus was confirmed for up to 6 weeks post-transfection by expression of co-transduced fluorescent reporters: mCherry with ASCL1 and DLX2 for GABAergic neuron differentiation and eGFP with NGN2 for glutamatergic neuron differentiation (Fig. 4d). Genetic markers for GABAergic neurons (GAD67) and glutamatergic neurons (vGLUT2), which indicate neuronal maturation, were strongly expressed from day 21 to up to day 42 of doxycycline treatment (Fig. 4e and 4f). GAD67, synonymous with GAD1, is a glutamate decarboxylase that catalyzes the synthesis of the inhibitory neurotransmitter GABA, and vGLUT2, synonymous with SLC17A6, is a vesicular glutamate transporter 2 located in the synaptic vesicle membrane. Electrical signals reflecting the physiological activity of the neural networks established on MEA plates were measured after full differentiation, at least 21 days after doxycycline treatment.

[0182] 2.4. Comparative Analysis of Patient Neurons and Isogeneically Corrected Neurons

[0183] The electrophysiological profiles of neurons derived from patient-specific iPSCs and mutation-corrected iPSCs were investigated using MEA plates. As shown in Figure 5a, both inhibitory and excitatory neurons were distributed across the electrode arrays of the MEA plate. By week 5, these neurons formed a widespread, interconnected network that covered the entire array. As shown in Figures 5b–5d, analysis of the electrical firing frequency of individual electrodes revealed that the frequency increased over time in both cell types, indicating robust network formation on the MEA plate. However, significant differences in firing frequency were observed between patient-derived (SCN2A mutation) and mutation-corrected (wild-type SCN2A) cells. Firing in patient-derived cells was more frequent and irregular, in contrast to the less frequent and more regular spikes of isogenic cells. Frequent and irregular electrical firing is a hallmark of epilepsy, which is generally defined in practice as two spontaneous seizures occurring more than 24 hours apart.

[0184] While the firing patterns shown in Figures 5b-5d intuitively resemble seizure patterns observed in epilepsy patients, we quantitatively investigated the characteristics of the in vitro epilepsy model by calculating three metrics from the measured spontaneous electrical firing events. The mean firing rate (Hz) represents the total number of spikes during the analysis period and is a measure used to assess the functional status and behavioral patterns of the nervous system (Figure 5e). During epileptic seizures, the firing rate of neurons increases to abnormal levels, as observed in the in vitro neural network model. This observed increase is due to the hyperexcitability of neurons due to the SCN2A mutation, and was clearly observed over time in our model. Unlike patient-derived neurons, neurons differentiated from iPSCs (isogenic synaptic cells) with the corrected mutation did not show a significant increase in firing rate, but the formation of neural networks resulted in a slight increase in firing rate over time.

[0185] Bursts or seizure clusters are defined as clusters of consecutive spikes, and bursts are clinically significant as they can be associated with increased morbidity and mortality. Therefore, we also measured burst frequency and burst percentage. Burst frequency represents the number of bursts per second, and burst percentage represents the number of spikes in a burst divided by the total number of spikes. Similar to the mean firing rate, both burst measurements increased over time due to increased firing rate (Figures 5f and 5g). Interestingly, while there was no significant difference in burst frequency between mutant and isogenic neurons, a significant difference emerged at week 5 (Figure 5f). This difference suggests that neurons must be cultured on MEA plates for at least 5 weeks to be suitable for seizure measurement in our in vitro neural network model.

[0186] In the context of epilepsy, network bursts refer to the phenomenon in which multiple neurons within a network exhibit simultaneous, high-frequency firing patterns. These bursts are characterized by long, quiet periods with little spike discharge, punctuated by brief periods of intense spike activity, during which the entire network exhibits a high firing rate. This synchronized activity results in sudden, intense bursts of electrical activity that differ from the more regular, individual firing patterns of neurons. The network burst frequency represents the total number of these network bursts per second, and the network burst percentage represents the number of spikes in network bursts divided by the total number of spikes. As shown in Figures 5h and 5i, similar to other measurements, network bursts increased over time in patient neurons, but no significant changes were observed in isogenic neurons. These results demonstrate that our in vitro model can reproduce a characteristic feature of epilepsy in patients, namely, high-frequency, synchronized firing.

[0187] The synchrony index is a quantitative measure used to assess the level of synchronization between neurons or neuronal networks during a seizure. This index ranges from 0 to 1, with values ​​closer to 1 indicating a higher level of synchronization. As shown in Figure 5j, synchronization was more frequent in patient neurons than in neurons with corrected isogenic mutations at weeks 4 and 5. Previous reports using rat brain tissue to measure the synchrony index showed indices of 0.46 at seizure onset, 0.61 during seizures, and 0.165 between seizures. While the magnitude of the synchrony index obtained in the in vitro model was lower than in previous in vivo samples, the indices at weeks 4 and 5 were significantly higher in patient-derived neurons than in neurons with corrected isogenic mutations. Taken together, these results suggest that the in vitro model can reflect in vivo seizure behavior.

[0188] 2.5. Drug Candidate Screening

[0189] We performed a computational screening of potential drug candidates targeting the mutant SCN2A protein from 1.6 million compounds available in the ChemDiv library. We initially evaluated the pharmacological properties of the compounds, including BBB permeability, cardiotoxicity, and hepatotoxicity, and eliminated compounds that were either BBB-permeable or cardiotoxic or hepatotoxic. Of the 1.6 million compounds, only 140,000 were BBB-permeable and non-toxic.

[0190] To screen potential drug candidates, we first generated a predicted structure of the mutant SCN2A protein using the 3D structure of the wild-type SCN2A protein (PDB ID: 6J8E). The R1629L mutation, identified in a pediatric epilepsy patient, was introduced into the SCN2A structure and energy minimization was performed to reflect the mutation's conformational effects. We then evaluated the binding affinity of the compounds in the binding cavity of the mutant SCN2A, particularly around residues I409, L1466, F1764, and Y1771, which form known antiepileptic drug-binding pockets. When examining the mutational effects, we found that the R1629L mutation could lead to a loss of hydrogen bonding between Q268 and the mutant residue (R1629L) compared to the wild-type. This change affects the structural stability of the SCN2A protein, slowing the rapid inactivation process of electrical signaling and resulting in a gain-of-function effect (Figures 6A and B).

[0191] Subsequently, structural screening was performed based on binding affinity, and the top 100 compounds with high affinity were selected. These top compounds exhibited similar binding affinities, suggesting similar structures and pharmacological properties. To identify distinct candidates, the top 100 compounds were clustered into five groups based on physicochemical properties. A representative compound from each group (i.e., the compound with the highest affinity in each cluster) was selected and experimentally evaluated using neurons on MEA plates. The ChemDiv IDs of these representative compounds were 4456-3635, 2786-4513, 4561-0590, 4333-1696, and 6466-0224. The structures of these compounds in complex with SCN2A(R1629L) are shown in Figures 6 and 7a–7f. Computationally, these compounds were found to reside primarily within the binding pocket of SCN2A(R1629L) via hydrophobic interactions with spatially adjacent amino acids. The predicted binding affinities for mutant SCN2A proteins are shown in Table 4 [Efficacy of phenytoin and screened compounds against the SCN2A(R1629L) mutant].

[0192]

[0193]

[0194] 2.6. Efficacy evaluation of compounds screened on an in vitro neural network platform

[0195] The five selected candidate compounds were evaluated for their efficacy in reducing spontaneous firing observed in an in vitro neural network model generated from patient-derived iPSCs. For this evaluation, neurons cultured on MEA plates for at least 5 weeks were used because they exhibit complex neural interconnectivity and intense firing-like behavior (Figures 5A–J). Because spontaneous and frequent firing is a fundamental characteristic of epileptic neurons, we primarily measured the average firing rate using the in vitro neural model during treatment with each compound. We also measured other parameters to determine the compound's effect on seizures. For efficacy comparison, we used phenytoin, a conventional sodium channel blocker commonly prescribed to control seizures but less effective in patients with epilepsy harboring the SCN2A (R1629L) mutation, as a control.

[0196] [Revised on 29.08.2025 by Article 91 of the Rules] As shown in Fig. 5e, isogenic cells harboring the wild-type SCN2A gene showed an average firing rate of about 0.5 Hz and no epileptic symptoms. Therefore, it was judged that a candidate compound that could lower the firing rate to 0.5 Hz would be effective in controlling seizures in diseased neurons. As shown in A1 of Fig. 8a, phenytoin reduced the average firing rate, and a noticeable effect was observed at about 10 μM (p value ≤ 0.001). When the activity assay results were applied, the estimated half-maximal inhibitory concentration (IC) of phenytoin 50 ) was 18 μM, indicating that phenytoin at this concentration can reduce the average firing rate by up to 50% (A2 in Fig. 8a and Table 4).

[0197] [Revised 29.08.2025 under Rule 91] Similarly, the efficacy of the compounds screened as potential blockers of mutated sodium channels was also tested (Fig. 8e, Fig. 8i, Fig. 8m, Fig. 8q, and Fig. 8u), and their estimated IC 50The values ​​are listed in Table 4. Compared to phenytoin, the IC of each compound 50 The IC values ​​ranged from approximately 18 μM to 200 nM. Specifically, compound 6466-0224 showed similar efficacy to phenytoin, but compounds 2786-4513 and 4333-1696 showed approximately 100-fold higher efficacy in suppressing seizures. The other two compounds, 4456-3635 and 4561-0590, showed approximately 10-fold higher efficacy than phenytoin. Consequently, it was found that each compound had a unique effect on spontaneous firing of neurons in vitro, and the IC 50 The two compounds with the lowest values ​​(250 nM for 2786-4513 and 200 nM for 4333-1696) emerged as potential drug candidates targeting SCN2A (R1629L) and restoring electrical firing activity and connectivity of neurons.

[0198] [Revised 29.08.2025 under Rule 91] We also evaluated other indicators to comprehensively investigate the effects of the screened compounds on seizures. Phenytoin significantly reduced bursts at concentrations above 30 μM (A3 and A4 in Figure 8b) and was effective at lower concentrations in terms of network bursts (A5 and A6 in Figure 8c). With regard to the synchrony index, an indicator of synchronization, phenytoin showed a more pronounced effect starting at a concentration of 10 μM. As shown in Figures 8e-h, compounds 4456-3635 showed a strong response even at lower concentrations, whereas phenytoin required higher concentrations to produce a similar effect. Compounds 2786-4513 also showed a pronounced response at lower concentrations than phenytoin (Figures 8i-l). Compound 4561-0590 showed a strong response even at low concentrations, but its effects on burst frequency and percentage were more variable (Figs. 8m-8p). Compound 4333-1696 showed a behavior similar to phenytoin, with a decrease in spike rate and burst frequency with increasing concentration, but a decrease in IC of mean firing rate. 50 It showed a 10-fold higher efficacy than phenytoin in terms of side effects (Figs. 8q-8t). Compound 6466-0224 showed a marked difference in its effect on the mean firing rate, but similar effects in other indicators (Figs. 8u-8x).

[0199] While specific aspects of the present invention have been described in detail above, it should be apparent to those skilled in the art that these specific descriptions merely represent preferred embodiments and are not intended to limit the scope of the present invention. In other words, the substantial scope of the present invention is defined by the appended claims and their equivalents.

[0200] By utilizing the epilepsy treatment screening method of the present invention and its composition, a pharmaceutical composition for preventing or treating epilepsy with superior efficacy and selectivity compared to conventional epilepsy treatment agents can be provided.

Claims

1. A pharmaceutical composition for preventing or treating epilepsy, comprising at least one compound represented by the following chemical formulas 1 to 5 as an active ingredient. [Chemical Formula 1] In the above chemical formula 1, R1 is an alkyl group having 1 to 3 carbon atoms. [Chemical Formula 2] [Chemical Formula 3] In the above chemical formula 3, R2, R3, R4 and R5 are each independently an alkyl group having 1 to 3 carbon atoms. [Chemical Formula 4] In the above chemical formula 4, R6, R7 and R8 are each independently an alkyl group having 1 to 3 carbon atoms. [Chemical Formula 5] In the above chemical formula 5, R9, R10 and R11 are each independently an alkyl group having 1 to 3 carbon atoms, and R12, R13 and R14 are each independently one of hydrogen (H), fluorine (F), chlorine (Cl) or bromine (Br).

2. In paragraph 1, A composition characterized in that the above epilepsy is caused by an R1629L mutation in the SCN2A (sodium voltage-gated channel alpha subunit 2) gene.

3. In paragraph 1, The compounds represented by the above chemical formulas 1 to 5 are A composition characterized in that the binding affinity for the SCN2A protein containing the R1629L mutation is -10 kcal / mol or less.

4. In paragraph 1, The compounds represented by the above chemical formulas 1 to 5 are IC for SCN2A protein containing R1629L mutation 50 A composition characterized in that the value is 18 μM or less.

5. In paragraph 1, The compounds represented by the above chemical formulas 1 to 5 are A composition characterized by exhibiting anti-seizure activity that reduces the average firing rate to 0.5 Hz or less.

6. A method for screening for an epilepsy treatment agent using induced pluripotent stem cells (iPSCs) derived from an epilepsy patient with a genetic type including a mutation in the SCN2A (sodium voltage-gated channel alpha subunit 2) gene, (a) a step of producing induced pluripotent stem cells (iPSCs) from peripheral blood mononuclear cells of the patient; (b) a step of differentiating the iPSC into inhibitory neurons and excitatory neurons; (c) a step of culturing the differentiated neural cells on a micro electrode array (MEA) plate to build an in vitro neural network model; (d) treating the candidate compound to the in vitro neural network model and measuring one or more of the following indices: activity index (average firing rate), electrode burst index, network burst index, and synchrony index; and (e) A method comprising a step of evaluating the anti-seizure activity of the candidate compound using the measured index value.

7. In paragraph 6, The above step (e) Based on the index values ​​measured in homogeneous iPSC-derived neural cells in which the mutation of the above SCN2A gene was corrected by gene editing technology, A method characterized by evaluating the anti-seizure activity of the above candidate compound.

8. In paragraph 6, A method characterized in that the candidate compound is selected based on binding affinity to the SCN2A protein containing the mutation.

9. In paragraph 6, The above inhibitory neurons are differentiated through the expression of ASCL1 and DLX2 genes, A method characterized in that the above excitatory neurons are differentiated through expression of the NGN2 gene.

10. In paragraph 6, A method characterized in that the above anti-seizure activity is evaluated as having effective anti-seizure activity when the average firing rate is reduced to 0.5 Hz or less.

11. In paragraph 6, A method characterized in that the mutation of the above SCN2A gene comprises an R1629L mutation.

12. A composition for screening an epilepsy treatment agent comprising inhibitory neurons and excitatory neurons differentiated from induced pluripotent stem cells (iPSCs) derived from a patient having a genotype including a mutation in the SCN2A (sodium voltage-gated channel alpha subunit 2) gene.

13. In paragraph 12, A composition characterized in that the above neural cells are cultured on a micro electrode array (MEA) plate to form a neural network model exhibiting electrophysiological responsiveness.

14. In paragraph 12, A composition characterized in that the composition comprises a control model for comparison, provided together with an isogenic neuronal cell in which a mutation in the SCN2A gene has been corrected.

15. In paragraph 12, A composition characterized in that the mutation of the above SCN2A gene comprises an R1629L mutation.

Images