Broadening the scope of multigene panel analysis for adult epilepsy patients

Abstract

Objective

Epilepsy is a suitable target for gene panel sequencing because a considerable portion of epilepsy is now explained by genetic components, especially in syndromic cases. However, previous gene panel studies on epilepsy have mostly focused on pediatric patients.

Methods

We enrolled adult epilepsy patients meeting any of the following criteria: family history of epilepsy, seizure onset age ≤ 19 years, neuronal migration disorder, and seizure freedom not achieved by dual anti-seizure medications. We sequenced the exonic regions of 211 epilepsy genes in these patients. To confirm the pathogenicity of a novel MTOR truncating variant, we electroporated vectors with different MTOR variants into developing mouse brains.

Results

A total of 92 probands and 4 affected relatives were tested, and the proportion of intellectual disability (ID) and/or developmental disability (DD) was 21.7%. As a result, twelve probands (13.0%) had pathogenic or likely pathogenic variants in the following genes or regions: DEPDC5, 15q12-q13 duplication (n = 2), SLC6A1, SYNGAP1, EEF1A2, LGI1, MTOR, KCNQ2, MEF2C, and TSC1 (n = 1). We confirmed the functional impact of a novel truncating mutation in the MTOR gene (c.7570C > T, p.Gln2524Ter) that disrupted neuronal migration in a mouse model. The diagnostic yield was higher in patients with ID/DD or childhood-onset seizures. We also identified additional candidate variants in 20 patients that could be reassessed by further studies.

Significance

Our findings underscore the clinical utility of gene panel sequencing in adult epilepsy patients suspected of having genetic etiology, especially those with ID/DD or early-onset seizures. Gene panel sequencing could not only lead to genetic diagnosis in a substantial portion of adult epilepsy patients but also inform more precise therapeutic decisions based on their genetic background.

Plain Language Summary

This study demonstrated the effectiveness of gene panel sequencing in adults with epilepsy, revealing pathogenic or likely pathogenic variants in 13.0% of patients. Higher diagnostic yields were observed in those with neurodevelopmental disorders or childhood-onset seizures. Additionally, we have shown that expanding genetic studies into adult patients would uncover new types of pathogenic variants for epilepsy, contributing to the advancement of precision medicine for individuals with epilepsy. In conclusion, our results highlight the practical value of employing gene panel sequencing in adult epilepsy patients, particularly when genetic etiology is clinically suspected.

---

Key Points

• This study showed the clinical efficacy of gene panel sequencing in adult epilepsy patients with a suspected genetic origin.
• We identified pathogenic or likely pathogenic variants in 13.0% of our patients.
• The diagnostic yield was higher in patients with neurodevelopmental disorders or childhood-onset seizures.
• The 15q12-q13 duplication was found in two unrelated patients with neurodevelopmental disorders.
• Using a mouse model, we identified a novel type of truncating MTOR mutation that disrupts the neuronal migration process.

---

INTRODUCTION

In the last 20 years, epilepsy genetics has rapidly advanced along with the development of high-throughput screening technologies, such as next-generation sequencing (NGS) and chromosomal microarray testing.^1,2 As of 2022, approximately 170 genes are associated with epilepsy in the Online Mendelian Inheritance in Man database, and more than 1200 epilepsy-related genes are listed in the DisGeNET database.³ The increasing number of epilepsy genes implies that genetic diagnosis is becoming increasingly important in clinical practice.

NGS technology is now widely used for the genetic diagnosis of a variety of diseases.⁴ It enables us to screen variants located in many genes or loci at once. In epilepsy patients, whole exome sequencing (WES) and gene panel sequencing are increasingly utilized in the clinical setting, demonstrating their clinical efficacy.^5–7 In particular, gene panel sequencing is now considered a routine molecular diagnostic test for various clinical conditions.^4,8,9 In a previous study, we demonstrated the utility of gene panel sequencing in patients with seizure onset within the first year of life.¹⁰

Genetic studies in epilepsy have primarily focused on pediatric patients, exploring diagnostic yields using various techniques like chromosomal microarray, WES, and gene panel sequencing in numerous studies.^11–14 The clinical utility of genetic studies in adult patients with epilepsy has been rarely examined, and previous research has predominantly focused on patients with specific medical conditions such as intellectual disability (ID) and developmental disability (DD).^1,15 However, it is important to note that many individuals who experienced early-onset epilepsy with genetic alterations can survive into adulthood and require continuous management.^16–18 Despite this, the emphasis on genetic testing has been mainly on pediatric patients, where the cost-effectiveness is greatest.^19,20 Consequently, the majority of adult epilepsy patients may not have benefited from the latest advancements in genetic diagnostic technologies, leaving a significant number of them with potentially hidden genetic defects. Because some epilepsy syndromes (e.g., autosomal dominant familial neonatal seizures) have a self-limiting course, the composition of epilepsy patients will differ between the pediatric and adult populations. Therefore, data obtained from the pediatric epilepsy panel cannot be directly applied to adult patients.

Recent studies have suggested that adult epilepsy patients would also benefit from a precision medicine approach. When gene panel testing was conducted in 200 epilepsy patients, the vast majority of whom (91%) had comorbid ID, genetic diagnosis was achieved in 46 patients (23%).¹⁵ In particular, gene-specific treatment changes were initiated in 11 patients, and the outcome was improved. In the case of Dravet syndrome, clinical manifestations may not be evident in adulthood, and affected patients can receive inappropriate treatment.²¹ If an adult epilepsy patient is diagnosed with Dravet syndrome, sodium channel blockers should be avoided thereafter. In this manner, genetic testing in adults is also useful for therapeutic decision-making.

Currently, genetic testing is not yet routinely considered in adult epilepsy clinics, and little is known about its diagnostic yield. Criteria for genetic testing in adult epilepsy patients need to be established, and more data are needed to this end. Here, we aimed to investigate the diagnostic yield of gene panel sequencing in Korean adult epilepsy patients and to obtain additional data on the utility of genetic testing.

METHODS Study subjects

We recruited adult epilepsy patients, aged 18 years and older, who visited the neurology clinic of Seoul National University Hospital from 2016 to 2021. We enrolled patients who were suspected to have genetic etiology meeting any of the following criteria: (1) family history of epilepsy, (2) seizure onset age ≤ 19 years, (3) neuronal migration disorder, or (4) seizure freedom not achieved by dual anti-seizure medications (ASMs). Although ID or DD was a major factor in identifying genetic causes from epilepsy patients, this was not a prerequisite for enrollment in this study. Additionally, patients with epileptogenic lesions other than cell migration disorders and those with known acquired causes of epilepsy, including hippocampal sclerosis (HS), were not included in the study.

We collected clinical information from all enrolled patients, including epilepsy phenotype, seizure onset age, early development, and relevant comorbidities, such as DD and ID. Epilepsy phenotypes are classified into two types: focal and nonfocal (generalized or unknown). If there are findings of focal seizures based on electroencephalography (EEG), magnetic resonance imaging (MRI), and seizure semiology, they are classified as focal epilepsy; otherwise, they are classified as nonfocal epilepsy.²² We further investigated the patients' seizure outcomes, numbers and types of ASMs, histories of febrile seizure, and family histories of epilepsy, which defined as information about disorders present in the direct blood relatives of the patient. We documented the results of EEG and MRI.

This study protocol was approved by the Institutional Review Board of Seoul National University Hospital (1912-055-1087), and all methods were carried out in accordance with relevant guidelines and regulations. Informed consents were obtained from all study participants before collecting their blood samples.

Epilepsy gene panel sequencing

We selected 211 genes for which associations with epilepsy have been reported (Table S1). The exonic regions of the target genes were captured using a SureSelect DNA targeted sequencing panel (Agilent Technologies, Santa Clara, CA, USA) and sequenced using Illumina technology. Sequence reads were aligned to the human reference genome (GRCh37) and processed following the Genome Analysis Toolkit (GATK) best-practice pipeline.²³

Sequence variants, such as single-nucleotide variants (SNVs) and short insertions/deletions (indels), were called using HaplotypeCaller of the GATK pipeline and underwent stepwise filtration as described in Appendix S1. Candidate variants were classified according to the American College of Medical Genetics (ACMG) guidelines using the InterVar program.^24,25 Although we mainly focused on pathogenic or likely pathogenic variants, we also suggested deleterious variants of unknown significance (dVUS), which are nonsilent and extremely rare (allele frequency [AF] = 0 in all databases we used). We also utilized the CADD scores to predict the pathogenicity of SNVs, and dVUS SNVs should have CADD scores above 20.²⁶ In addition, among the pathogenic variants, we identified a novel type of nonsense mutation in MTOR which pathomechanism was unclear yet. We conducted a functional study for the mutation using a mouse model, and detailed methods were described in Appendix S1.

Copy number variants (CNVs) were called using the algorithm described in our previous study (Appendix S1).¹⁰ We utilized multiple ligation-dependent probe amplification (MLPA) for CNV validation, which was conducted according to the manufacturer's instructions (SALSA MLPA ME028 Prader-Willi/Angelman probemix, MRC-Holland, Amsterdam, Netherlands).

Statistical analysis

All statistical comparisons were conducted using SPSS (Version 22.0, Chicago, IL, USA) and GraphPad Prism 9 (GraphPad Software, San Diego, CA). Chi-square tests or Fisher's exact tests were performed to analyze the categorical data, and p < 0.05 was considered statistically significant. The Mann–Whitney test was performed to compare the diagnostic yield according to the seizure onset age for each subgroup. The Kruskal–Wallis test with Dunn's multiple comparison method was conducted to compare multiple means when data did not follow a normal distribution.

RESULTS Patient demographics

The total number of adult patients enrolled in this study was 96, with 92 probands (Table 1), and the age range at the time of panel sequencing was from 18 to 65 years. All the probands were unrelated to each other. The cohort included 52 males, and the mean age was 33.4 years (ranging from 18 to 67 years). The seizure onset age varied from less than 1 year to 49 years, although the most common onset age was 10–19 years (49 patients, 51.0%).

TABLE 1 Patient demographics (n = 96 samples, 92 probands).

Clinical information	No. of individuals (%)
Current age
18–29 years	40 (41.7)
30–39 years	29 (30.2)
41–48 years	18 (18.8)
50–67 years	9 (9.3)
Gender
Male	52 (54.2)
Age at seizure onset
<1 year	1 (1.1)
1–9 years	32 (33.3)
10–19 years	49 (51.0)
20–29 years	10 (10.4)
30–39 years	3 (3.1)
40–49 years	1 (1.1)
Inclusion criteria
Family history	27 (28.1)
Age at onset ≤19	82 (85.4)
Neuronal migration disorder	15 (15.6)
Polytherapy	63 (65.6)
Number of ASM
≤2	35 (36.5)
3	18 (18.8)
4	27 (28.1)
≥5	16 (16.7)
Epilepsy phenotype
Focal epilepsy
TLE	40 (41.7)
FLE	14 (14.6)
PLE	5 (5.2)
OLE	8 (8.3)
Multifocal	3 (3.1)
Generalized epilepsy	22 (22.9)
Unknown	4 (4.1)
EEG abnormality
Focal spike/sharp wave	52 (54.2)
Focal rhythmic activity
Alpha	1 (1.1)
Theta	4 (4.1)
Delta	1 (1.1)
Generalized spike/sharp wave	21 (21.9)
Normal	10 (10.4)
Others	7 (7.3)
History of febrile seizure	14 (14.6)
ID/DD	20 (20.8)
Early-onset (≤10 years)	10 (10.4)
Refractory	14 (14.6)
Early-onset and refractory	7 (7.3)

Abbreviations: ASM, anti-seizure medication; DD, developmental disability; EEG others, generalized slowing, focal fast activity, no EEG; EEG, electroencephalography; FLE, frontal lobe epilepsy; ID, intellectual disability; OLE, occipital lobe epilepsy; PLE, parietal lobe epilepsy; TLE, temporal lobe epilepsy.

Among the inclusion criteria, most patients (82 patients, 85.4%) met the criteria of seizure onset age under 19 years, and 63 patients (65.6%) met the criteria of polytherapy. Focal spike/sharp wave discharges appeared most frequently (54.2%) on EEG, and temporal lobe epilepsy was the most frequent epilepsy phenotype among our patients (41.7%). Family history was confirmed in 25 patients (26.0%), history of febrile seizure in 14 patients (14.6%), and ID/DD in 20 patients (20.8%, 21.7% of probands). Among the 20 patients with ID/DD, 10 patients had seizure onset before 10 years (50.0%), 14 patients had refractory epilepsy (70.0%), and 7 patients experienced both early-onset seizures and refractory epilepsy (35.0%). Among those with and without early-onset seizures, 10 out of 38 (26.3%) and 10 out of 58 (17.2%) had ID/DD, respectively. Regarding refractory epilepsy, 14 out of 65 (21.5%) had ID/DD, compared to 6 out of 31 (19.4%) without.

Genomic variants detected by gene panel sequencing

The median average coverage depth was 623.8×, and the median % coverage of target regions (>100×) was 99.4%. Although four patients with low sequencing coverage had average coverage depths less than 150× (ranging from 62.8× to 111.1×), most of their target regions were also covered by >30× reads (89.9% ~ 98.3%). A gene panel sequencing summary is shown for each patient in Table S2.

We detected pathogenic or likely pathogenic variants in 12 probands (12/92, 13.0%), including 2 patients with 15q12-q13 duplication (Figure 1A, Table 2). Except for DEPDC5, for which mutations were identified in EP45 and EP54, every gene was affected in a single proband by a sequence variant. Three mutations, each in SLC6A1 p.Arg44Gln, DEPDC5 p.Trp1369Ter, and SYNGAP1 p.Pro562Leu, have been previously reported in ClinVar or HGMD, and the other seven variants could be novel pathogenic variants for epilepsy-causing mutations.

FIGURE 1. Pathogenic or likely pathogenic variants identified in our patients. (A) Proportion of pathogenic (P) or likely pathogenic (LP) variants and deleterious variants of unknown significance (dVUS). (B) Increased coverage depth in the 15q12-q13 region for EP88 and EP91, calculated as log2(fold change) in comparison with the average coverage depth of study samples. Gray lines in the background indicate the coverage depth ratios of other patients in the 15q12-q13 region.

TABLE 2 Pathogenic or likely pathogenic sequence variants identified in adult epilepsy patients.

Proband	Gene symbol	Gene ID	Nucleotide change	Amino-acid change	Variant type	ACMG criteria	Previously reported	CADD score
EP05	SLC6A1	NM_003042.4	c.131G > A	p.Arg44Gln	Het	LP (PM1, PM2, PP3, PP5)	Yes (ClinVar, HGMD)	27.6
EP06	SYNGAP1	NM_006772.3	c.1685C > T	p.Pro562Leu	Het	LP (PM1, PM2, PP3, PP5)	Yes (HGMD)	31
EP09a	EEF1A2	NM_001958.5	c.497G > A	p.Arg166His	Het	LP (PM1, PM2, PP1, PP2, PP3)	No	27.7
EP23a	LGI1	NM_005097.4	c.1355 T > C	p.Ile452Thr	Het	LP (PM1, PM2, PP1, PP3)	No	26.2
EP27a	MTOR	NM_004958.4	c.7570C > T	p.Gln2524Ter	Het	P (PS2, PS3, PM2, PP3)	No	48
EP33a	KCNQ2	NM_172107.4	c.1163 T > C	p.Leu388Pro	Het	LP (PM1, PM2, PP1, PP3)	No	23.9
EP42	MEF2C	NM_001193347.1	c.1442G > A	p.Trp481Ter	Het	LP (PVS1, PM2, PP3)	No	38
EP45a	DEPDC5	NM_001242896.3	c.4107G > A	p.Trp1369Ter	Het	P (PVS1, PM2, PP1, PP3, PP5)	Yes (ClinVar, HGMD)	42
EP54	DEPDC5	NM_001242896.3	c.3805 + 1G > A	NA	Het	LP (PVS1, PM2)	No	35
EP79	TSC1	NM_000368.5	c.2271del	p.Glu757AspfsTer16	Het	LP (PVS1, PM2)	No	NA

Abbreviations: Het, heterozygous; LP, likely pathogenic; P, pathogenic; NA, not applicable; VUS, variant of unknown significance.

^aPatients with familial genetic testing conducted. The variant of EP27 was confirmed to be de novo, and variants of other patients were co-segregated in their affected family members (an older brother of EP09, two daughters of EP23, a father of EP33, and a son of EP45).

None of our study patients underwent CNV screening previously, such as microarray and karyotyping. CNV analysis based on the panel data revealed proximal 15q duplications in two patients, EP88 and EP91, which included UBE3A, GABRB3, GABRB5, and CHRNA7 among our target genes (Figure 1B). Both were validated by MLPA experiments, and EP91 was assumed to have four copies of the 15q12-q13 region, while EP88 had three copies of the CNV.

In addition to the pathogenic or likely pathogenic variants, we listed dVUS SNVs of 20 patients, located in our target regions of 211 genes, in Table S3. None of these SNVs have been reported in any public variant databases we used, and their CADD scores were high, above 20. We could not recategorize them into pathogenic or likely pathogenic variants because subsequent genotyping was not available for the family members.

Clinical characteristics of genetically confirmed patients

The clinical information of genetically confirmed patients is shown in Table 3. All patients had seizures before age 19, and 6 (50%) of them had seizures before age 10. Four patients (33.3%) had family histories of epilepsy. Neuronal migration disorder was identified in 2 patients (16.6%), displaying subependymal heterotopia. Seven patients (58.3%) maintained polytherapy because seizure-free status was not achieved with dual ASMs. While 7 patients (58.3%) had focal epilepsy, 3 of the patients (25%) had generalized epilepsy. In particular, 7 out of 12 patients (58.3%) were also affected by ID/DD.

TABLE 3 Phenotypic characteristics of patients with pathogenic or likely pathogenic variants.

Case	Gene symbol	Sex	Age (year)	Seizure onset age (year)	Epilepsy phenotype	Family history	ID/DD	EEG	Brain MRI	ASM
EP05	SLC6A1	F	24	2	Generalized	No	Yes (ID)	Generalized (poly) spike and wave	Negative	LEV, LMT
EP06	SYNGAP1	M	20	3	Generalized	No	Yes (ID)	Generalized (poly) spike and wave	Negative	VPA
EP09	EEF1A2	F	24	8	Generalized	Yes (brother)	No	Generalized (poly) spike and wave	Negative	LEV, LMT, ZNS, PHT
EP23	LGI1	M	53	15	Focal (lateral TLE)	Yes (two daughters)	No	No EEG	Negative	CZP
EP27	MTOR	M	38	8	Focal (FLE)	No	Yes (ID)	Focal spike/sharp wave	Abnormal (subependymal heterotopia)	VPA, CBZ, CB, PRP
EP33	KCNQ2	M	24	12	Focal (FLE)	Yes (father)	Yes (ID, DD)	Focal spike/sharp wave	Abnormal (asymmetric lateral ventricle size)	LEV, LCS, TPM, PRP
EP42	MEF2C	M	41	8	Focal (TLE)	No	Yes (ID)	Focal spike/sharp wave	Negative	LEV, VPA, LCS, CBZ, PRP
EP45	DEPDC5	F	37	18	NA	Yes (son)	No	No EEG	No image	No
EP54	DEPDC5	F	66	8	Focal (TLE)	No	No	Focal spike/sharp wave	Negative	LCS, PGB, CBZ
EP79	TSC1	F	22	19	Focal (OLE)	No	No	Focal spike/sharp wave	Abnormal (subependymal nodule)	No
EP88	15q12-q13 duplication	M	27	16	Focal (hemispheric)	No	Yes (ID)	Rhythmic theta	Negative	VPA, LEV, LMT, CBZ
EP91	15q12-q13 duplication	M	30	18	Generalized	No	Yes (ID, autism)	Generalized (poly) spike and wave	Negative	VPA, LEV, LMT

Abbreviations: ASM, anti-seizure medication; CAE, childhood absence epilepsy; CB, clobazam; CBZ, carbamazepine; EEG, electroencephalography; FLE, frontal lobe epilepsy; ID/DD, intellectual disability/developmental disability; JAE, juvenile absence epilepsy; LCS, lacosamide; LEV, levetiracetam; LMT, lamotrigine; OLE, occipital lobe epilepsy; PGB, pregabalin; PHT, phenytoin; PRP, perampanel; TLE, temporal lobe epilepsy; TPM, topiramate; VPA, valproic acid; ZNS, zonisamide.

Of the pathogenic or likely pathogenic variants detected in our patients, LGI1 (3 patients, 1 proband), DEPDC5 (2 patients, 2 probands), and 15q12-q13 duplication (2 patients, 2 probands) were most common. Patients with LGI1 mutations were in the same family, and all three had focal epilepsy and no ID/DD. In patients with DEPDC5 mutation, one had a positive family history, and the other had no family history, but both had no ID/DD. Both patients with 15q12-q13 duplication had no family history but had ID/DD.

The patient with EEF1A2 mutation (EP09) had a family history of epilepsy in her older brother, who had a generalized epilepsy with absence seizures but had no ID or DD. Genetic testing of the patient's older brother also confirmed the presence of the EEF1A2 mutation. The patient with MTOR mutation (EP27) had a severe intellectual disability to the extent that he could not live independently and had the intellectual level of a 3–4 years old. Subependymal heterotopia was detected on his brain MRI. The patient with TSC1 mutation (EP79) had subependymal nodules on her brain MRI, as well as periungual fibromas on her right thumb and second finger. However, there were no additional clinical features suggesting tuberous sclerosis in this patient. She underwent chest and abdominal CT, and the results were normal.

Subgroup analysis of diagnostic yield

We further analyzed the diagnostic yields of 92 probands by subgroups, as illustrated in Figure 2. The diagnostic yield of patients with early-onset seizures (≤10 years) was 15.8% (6 out of 38) and that of patients with ID/DD was 35.0% (7 out of 20). Both yields were higher than that of overall patients, and the diagnostic yield of patients with ID/DD was significantly higher than that of patients without ID/DD (35.0% vs. 6.9%, p = 0.003). Moreover, among patients with early-onset seizures (≤10 years), those with ID/DD showed the highest diagnostic yield of 40.0% (4 out of 10), which was higher than that in patients without ID/DD (2 out of 28) (p = 0.031). According to multivariate analysis, ID/DD (odds ratio 8.15 [1.75–34.03]; p = 0.004) was independently associated with identification of pathogenic or likely pathogenic variants (Table S4).

FIGURE 2. Diagnostic yields differing by clinical subgroups. (A) Proportions of genetically diagnosed patients for different clinical conditions. (B) Diagnostic yields for clinical subgroups subdivided by onset age. The clinical subgroups were determined by seizure type, family history, and intellectual disability (ID)/developmental disability (DD). †This compares patients with and without ID/DD among those with seizure onset at ≤10 years.

The diagnostic yield of focal epilepsy patients (7 out of 67) was not significantly different from that of nonfocal epilepsy patients (5 out of 25) (p = 0.297) (Figure 2A). In addition, the effect of seizure onset age on the diagnostic yield was relatively small in patients with focal epilepsy or a family history of epilepsy, unlike in those with ID/DD (Figure 2B).

Pathogenic effect of a nonsense MTOR mutation on neuronal migration

As shown in Figure 3A, most of the pathogenic or likely pathogenic MTOR variants listed in ClinVar are missense variants.²⁷ However, we detected a nonsense variant from EP27 (c.7570C > T, p.Gln2524Ter), which was located in the C-terminal region of MTOR and previously not reported in public databases. Therefore, we conducted a functional assay to confirm its pathogenicity using a mouse model.

FIGURE 3. Pathogenicity confirmation of a novel nonsense mutation of MTOR. (A) In ClinVar, most of the pathogenic or likely pathogenic variants in MTOR are missense variants, except one frameshift variant located in the N-terminal region. In contrast, EP27 had a nonsense variant (p.Gln2524Ter) in the C-terminal region of MTOR. (B) Representative images of E18.5 mouse brain slices electroporated at E14.5 with vectors containing each MTOR variant. Neuronal migration was characterized by Ctip2 staining and quantified by the percentages of GFP+ cells in each layer (vector, n = 3; MTORWT, n = 3; MTORQ2524*, n = 6; MTORC1483Y, n = 3; MTORKD, n = 4; MTORQ2524* KD, n = 6; by Kruskal–Wallis test with Dunn's multiple comparison test). The error bar represents the mean ± SD. p [less than] 0.05 (*), p [less than] 0.01 (**), p [less than] 0.001 (***); scale bars, 50 μm; KD, kinase-dead; loCP, lower cortical plate; mCP, middle cortical plate; uCP, upper cortical plate.

To investigate the functional impact of the novel MTOR variant, we electroporated vectors encoding green fluorescent protein (GFP) and different MTOR variants into developing mouse brains at E14.5. The majority of GFP⁺ cells (91.88 ± 2.07%) electroporated with the control vector reached the destined upper cortical plate (uCP) region above the Ctip2⁺ layer at E18.5, whereas a significantly lower portion of GFP⁺ cells (73.83 ± 8.04%) electroporated with MTOR^Q2524* reached the uCP layer, comparable to that of MTOR^C1483Y (54.83 ± 6.77%), a previously reported gain-of-function variant known to cause neuronal migration defects.²⁸ To further test the requirement of mTOR kinase activity in mediating neuronal migration defects, we electroporated MTOR^Q2524* construct containing additional kinase-dead (KD) mutations (D2537E and V2364I), which renders mTOR null for kinase activity.²⁹ While no significant difference was observed between control and MTOR^KD overexpression, the KD mutations abolished the neuronal migration defects observed in MTOR^Q2524*, suggesting the requirement of mTOR kinase activity for migration defects. These data suggest that MTOR^Q2524* leads to a neuronal migration defect during cortical development, one of the hallmarks and underlying mechanisms of heterotopia, which may explain the subependymal heterotopia symptom observed in patient EP27.

DISCUSSION

We performed targeted gene panel sequencing in 96 Korean adult epilepsy patients (92 probands) who were likely to have genetic causes but had not previously undergone genetic testing. All the genetically confirmed patients had seizure onset before adulthood (mean age 11.3 years, between 2 and 19 years), and the mean current age was 33.9 years (ranging from 20 to 66 years). The diagnostic yield was highest in patients with ID/DD who also had childhood onset of seizure. The most frequently mutated gene in our cohort was DEPDC5, which was found in two probands, and two unrelated patients with 15q12-q13 duplication were identified. In addition, more than half of the causal variants (7 out of 12 variants) were novel variants previously not reported in ClinVar or HGMD.

We identified pathogenic or likely pathogenic variants in 13.0% of adult epilepsy probands in this study, which is comparable to findings from previous studies on large cohorts of adult patients.^30,31 The diagnostic yields of epilepsy gene panel sequencing in adults,^1,16,30 including our study, were 10%–23% and slightly lower than those of previous pediatric studies (14%–28%).^14,32,33 However, in previous studies on pediatric or adult epilepsy patients, several factors, such as the number of genes included in the gene panel, test methods for genetic diagnosis, and patient group recruitment criteria, may have affected the diagnostic yield.

Genetic testing is increasingly used in epilepsy clinics, and an increasing number of genetic causes are being discovered in adult epilepsy patients. A previous study using a commercial epilepsy gene panel analyzed the genetic causes in adult epilepsy patients with ID. Mutations in epilepsy-related genes, including SCN1A (three probands), GABRB3, and UBE3A (two probands with CNVs including both genes), KANSL1, SLC2A1, KCNQ2, SLC6A1, HNRNPU, STX1B, SCN2A, PURA, and CHD2 (one proband for each gene), were pathogenic or likely pathogenic.¹ In particular, the GABRB3 and UBE3A mutations were CNVs identical to those in our results; one was a heterozygous deletion, and the other was a duplication with four copies. Another study showed a high diagnostic yield with mutations in several genes (e.g., PCDH19, SYNGAP1, TPP1, PRRT2, DEPDC5).³⁴ Similarly, we identified patients with genetic defects in SYNGAP1 and DEPDC5.

One of the most interesting findings of this study is the discovery of two unrelated adult epilepsy patients with 15q12-q13 duplication. Because of the limitation of capture-based NGS technologies, which mostly target exonic regions of interest, most studies using gene panels or WES have mainly focused on sequence variants, such as SNV and indel. However, as demonstrated in our previous study,¹⁰ we successfully discovered pathogenic CNVs using gene panel sequencing and validated them through MLPA analysis. Considering that the 15q12-q13 duplications were detected in two unrelated patients (2 out of 12 probands, 16.7%), CNVs need to be screened in addition to sequence variants when identifying genetic etiologies of adult epilepsy patients.

Another interesting finding of our study is the phenotype of patient EP09. Most of the genes identified in our study and their associated phenotypes were generally consistent with previous reports. However, patient EP09, who was confirmed to have a novel EEF1A2 mutation, exhibited a different phenotype than previously reported. Specifically, she and her older brother had a generalized epilepsy with absence seizures but no ID or DD, and did not have the characteristic features associated with EEF1A2 mutations. EEF1A2 mutations are typically associated with severe neurodevelopmental disorders and distinctive facial features.³⁵ Although a previous case series study reported mild ID/DD and characteristic facial features in a 10-year-old girl with an EEF1A2 mutation,³⁶ there have been no previous reports of EEF1A2 mutations presenting without neurodevelopmental disorders or distinctive facial features. Our findings indicate that EEF1A2 mutations may present solely with refractory epilepsy, providing important new insights into the extended phenotype of EEF1A2. Additionally, our findings suggest that, in some adult epilepsy patients, diagnosis of certain epileptic syndromes based solely on clinical symptoms can be challenging, highlighting the value of genetic testing in adult patients.

Although many patients in our study were under 30 years of age, approximately 28.1% of them (27 out of 96) were over 40 years old (Table 1). Genetic causes were identified in three of these older patients, with the oldest being 66 years old (Table 3). Among those with seizure onset over 10 years of age, genetic causes were identified in 11.1% (6 out of 54). These results suggest that the genetic factors can be identified in a significant number of adult patients, and genetic testing may also be helpful in diagnosing the genetic basis of several cases of late-onset intractable epilepsy, thereby extending the applicability of genetic testing beyond traditional pediatric settings.

We further assessed the diagnostic yields according to clinical subgroups. The diagnostic yield was higher in patients with seizure onset ≤10 years than in those with onset >10 years of age (Figure 2). In a previous study,³⁰ pathogenic or likely pathogenic variants were found in 16.7% of patients with childhood-onset epilepsy, which is similar to our results (15.8%). However, there was no statistically significant difference in diagnostic yields between childhood-onset (≤10 years) and later-onset (>10 years) epilepsy in our study (p = 0.543, Figure 2B). Although this finding may be the result of a relatively small sample size compared to previous studies, it is notable that a considerable number of patients with later-onset epilepsy were found to have genetic abnormalities. Therefore, we argue that the age of seizure onset should not be overemphasized as a determining factor for genetic testing.

Because we did not set ID or DD, one of the major factors suggesting genetic contribution, as a prerequisite for study enrollment, we believe that a comparable diagnostic yield (13.0%) is a meaningful finding in adult epilepsy patients. Approximately 20% of the patients in our study had ID/DD, while the remaining 80% did not have ID/DD. Although the majority of the genetically diagnosed patients did have ID/DD (7 out of 12), it is noteworthy that genetic diagnoses were also identified in five patients without ID/DD. Furthermore, the detected mutations were heterogeneous and different from each other. This finding suggests that various genetic abnormalities can be found when genetic analysis is performed in a broader spectrum of adult epilepsy patents, including those without ID/DD.

Our results are consistent with several other previous studies in adults with epilepsy.^1,16,30 McKnight et al. analyzed a genetic study of over 2000 patients and reported an overall yield of 11%.³⁰ The yield was higher in patients with ID (16%) and those with onset of epilepsy before the age of 10 years (17%). In two previous studies of adult epilepsy patients with ID/DD, the diagnostic yields were 21.8% (14 out of 64)¹ and 23% (46 out of 200).¹⁵ Most genetically diagnosed patients in these two studies had childhood-onset epilepsy. In a study of pediatric patients with intractable epilepsy, global DD, and cognitive dysfunction,³³ the diagnostic yield was 28% (31 out of 110). Among the 31 patients, 8 patients (7%) were identified with inherited metabolic disorders causing epileptic encephalopathy in childhood, and the average seizure onset age was 4.1 ± 8.1 SD months (ranging from 1 h to 2 years). The remaining 23 patients (21%) were identified with genetic causes of epileptic encephalopathy in childhood, and the average seizure onset age was 27.23 ± 44.05 SD months (ranging from 1 week to 15 years). These results support that genetic testing should be conducted more actively in adult epilepsy patents with ID/DD, especially if seizures occur before age 10.

In our study, dVUS was additionally reported in Table S3 and was identified in 21.7% of patients (20 out of 92). As we applied strict criteria to select dVUS, including AF = 0 in multiple variant databases and high CADD scores above 20, some may be true causal variants for epilepsy. Most selected dVUS could be recategorized as likely pathogenic variants according to the ACMG guidelines when they are confirmed to be de novo or shared by affected family members. Unfortunately, conducting parental genetic testing is difficult in adult cohorts, unlike in pediatric patients, who usually visit clinics with their parents. Furthermore, most of our patients were analyzed by singleton panel sequencing. The diagnostic yield may have been higher if the genetic tests of parents or siblings of adult patients had been performed.

Functional validation could be a good alternative method to determine variant pathogenicity, especially for those identified in adult patients. In this study, we confirmed the pathogenic effect of a novel MTOR mutation using a mouse model. While most previously reported MTOR mutations were missense variants (Figure 3A), a nonsense mutation (c.7570C > T, p.Gln2524Ter) was identified from EP27. Considering its location in the C-terminal region, we can assume that it might act as a gain-of-function mutation rather than a loss-of-function mutation. Our experimental data support that MTOR^Q2524* disrupts the neuronal migration process during cortical development, similar to MTOR^C1483Y, which is a previously reported gain-of-function variant causing neuronal migration defects.²⁸ Increased mTOR kinase activity and hyperactivation of the PI3K-AKT–mTOR pathway have been well-characterized in neurodevelopmental disorders such as intractable epilepsy and focal cortical dysplasia, affecting neuronal excitation and migration.^28,37–39 In line with the previous observations, our data also showed that additional KD mutations abolished migration deficits caused by MTOR^Q2524*, supporting the requirement of mTOR kinase activity in neuronal migration defects caused by MTOR^Q2524*.

The data presented here highlight the need to increase awareness of and access to genetic testing for adult epilepsy patients. As our understanding of the genetic causes of epilepsy expands, the use of a multigene panel, exome sequencing, and/or whole genome sequencing, as well as the increasing availability of precision medicine (including gene-targeted therapies) will continue to expand, resulting in an increased chance of finding clinically actionable genes.

In conclusion, our findings highlight the importance of expanding genetic testing to include adult epilepsy patients, especially when a genetic etiology is clinically suspected, and underscore the need for increased awareness among neurologists and epileptologists managing such patients. In particular, identifying undiagnosed genetic disorders in adult patients who have transitioned from pediatric care may be crucial. We believe that a more systematic approach and refinement of the inclusion criteria for genetic testing for adult epilepsy patients have the potential to improve diagnosis and treatment outcomes, ultimately leading to better patient care.

AUTHOR CONTRIBUTIONS

BCL and JM: Conceptualization and design of the study. KC, SKL, KJK, BCL, and JM: Data acquisition. SL, M-KK, and SJ: Data analysis. KHS, RJ, and STB: Functional study. SL and M-KK: Original draft and figures. SL, M-KK, KHS, RJ, YWS, SJ, JGY, SK, MK, KC, SKL, KJK, STB, BCL, and JM: Review and editing of the manuscript.

ACKNOWLEDGMENTS

This research was supported by a fund (2020-ER6904-01) by Research of Korea Centers for Disease Control and Prevention and also supported by the Seoul National University Hospital Research Fund (0320200350). Kon Chu was supported by research grant from Samjin Pharmaceutical (0620201110). Jangsup Moon was supported by the Seoul National University Hospital Research Fund (0320210170).

CONFLICT OF INTEREST STATEMENT

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this study. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.

ETHICS STATEMENT

This study protocol was approved by the Institutional Review Board of Seoul National University Hospital (1912–055-1087).

PATIENT CONSENT STATEMENT

All participants provided written informed consent.