CHD4 variants are associated with childhood idiopathic epilepsy with sinus arrhythmia

Abstract Aims CHD4 gene, encoding chromodomain helicase DNA‐binding protein 4, is a vital gene for fetal development. In this study, we aimed to explore the association between CHD4 variants and idiopathic epilepsy. Methods Trios‐based whole‐exome sequencing was performed in a cohort of 482 patients with childhood idiopathic epilepsy. The Clinical Validity Framework of ClinGen and an evaluating method from five clinical‐genetic aspects were used to determine the association between CHD4 variants and epilepsy. Results Four novel heterozygous missense mutations in CHD4, including two de novo mutations (c.1597A>G/p.K533E and c.4936G>A/p.E1646K) and two inherited mutations with co‐segregation (c.856C>G/p.P286A and c.4977C>G/p.D1659E), were identified in four unrelated families with eight individuals affected. Seven affected individuals had sinus arrhythmia. From the molecular sub‐regional point of view, the missense mutations located in the central regions from SNF2‐like region to DUF1087 domain were associated with multisystem developmental disorders, while idiopathic epilepsy‐related mutations were outside this region. Strong evidence from ClinGen Clinical Validity Framework and evidences from four of the five clinical‐genetic aspects suggested an association between CHD4 variants and epilepsy. Conclusions CHD4 was potentially a candidate pathogenic gene of childhood idiopathic epilepsy with arrhythmia. The molecular sub‐regional effect of CHD4 mutations helped explaining the mechanisms underlying phenotypic variations.


| INTRODUC TI ON
CHD4 gene (OMIM* 603277) encodes chromodomain helicase DNA-binding protein 4 (CHD4), which belongs to the SNF2/RAD54 helicase family. It is an ATP-dependent chromatin remodeler that is involved in epigenetic regulation of gene transcription, DNA repair, and cell cycle progression. 1 CHD4 is ubiquitously expressed across the whole lifespan, including brain. Homozygous knockout of CHD4 in mice resulted in embryonic lethality with abnormal extraembryonic tissue physiology, absent blastocoele, and increased apoptosis.
In humans, CHD4 mutations were initially regarded to increase the risk of cancer. 2 Lately, CHD4 mutations have been reported to be associated with Sifrim-Hitz-Weiss syndrome (SIHIWES) (OMIM# 617159), which is an autosomal dominant intellectual developmental disorder with variable congenital defects involving cardiac, skeletal, and urogenital systems. 1,3 CHD4 mutations have also been detected in the patients with neurodevelopmental disorders, Rett syndrome, schizophrenia, and pituitary stalk interruption syndrome, among whom rare cases presented epilepsy. [4][5][6][7] However, the role of CHD4 in epilepsy remains largely unknown.
In this study, we performed trios-based whole-exome sequencing (WES) in a cohort of childhood epilepsy without acquired causes.
Four novel CHD4 mutations were identified in four unrelated families of childhood idiopathic epilepsy with sinus arrhythmia, suggesting that CHD4 was potentially a candidate causative gene of epilepsy with arrhythmia. We further systematically reviewed all CHD4 mutations and their corresponding phenotypes, aiming to explore the mechanisms underlying phenotypic variations.

| Subjects
A total of 482 patients with childhood unexplained epilepsy were recruited for genetic screening during January 2013-June 2020.
Seizure and epilepsy were diagnosed and classified according to the proposals by International League Against Epilepsy. [8][9][10][11] he collected clinical data included semiology, evolution of the disorders, family histories, and the data of management. Video-EEG recordings with ECG electrodes were obtained to confirm the diagnosis of epilepsy.
Echocardiography was performed for the patients with the cardiac symptoms and/or arrhythmia during EEG monitoring. Brain magnetic resonance imaging (MRI) was conducted to exclude symptomatic The studies adhered to the guidelines of the International Committee of Medical Journal Editors with regard to patient consent for research or participation and received approval from the Ethics Committee of the Second Affiliated Hospital of Guangzhou Medical University. All participants provided written informed consents.

| Identification of gene variants
The blood samples were obtained from the probands, their parents, and other available family members. Genomic DNA was extracted from peripheral blood using a QuickGene DNA whole blood kit (Fujifilm). Trios-WES was conducted on the Illumina HiSeq 2500/4000 platform by BGI-Shenzhen as previously reported. 12 Population-based filtration removed common variants presenting minor allele frequency ≥0.005 in the 1000 Genomes Projects, Exome Aggregation Consortium, and Genome Aggregation Database (gno-mAD). Then, we retained potentially pathogenic variants containing frameshift, nonsense, canonical splice site, initiation codon, and missense mutations predicted as being damaging by in silico tools (http:// varca rds.biols.ac.cn/). The potential disease-causing mutations in each case were screened under five models: epilepsy-associated gene model, dominant/de novo model, autosomal recessive inheritance model, X-linked model, and co-segregation analysis model. After excluding the known epilepsy-associated genes, the genes with de novo mutations, bi-allelic mutations, hemizygous mutations, and mutations with segregations were selected for further studies to validate as possible novel epilepsy genes. The candidate variants were validated by Sanger sequencing. Potential pathogenic variants were predicted by multiple in silico prediction tools (http:// www. varcards.biols.ac.cn/). Conservation of mutated positions was evaluated using sequence alignment of different species. All CHD4 variants were annotated based on the transcript NM_001273.3.

| Computational modeling
To evaluate the pathogenicity of candidate variants, the structure of CHD4 was modeled to predict the effect of missense mutations on protein structure by using SWISS-MODEL (https://swiss model. expasy.org/). PyMOL 2.3 software was used for three-dimensional protein structure visualization and analysis.

| Statistical analysis
R statistical software (v3.4.1) was used for statistical analysis. The frequency of the CHD4 variants in general population and in the cohort of childhood idiopathic epilepsy was compared by Fisher's exact test. A p value of < 0.05 was considered to be statistically significant.  Figure 1A and 1B). The amino acid sequence alignment showed that all the four residues were highly conserved across higher vertebrates ( Figure 1C). These mutations were predicted to be damaging by multiple in silico prediction tools (Table S1). All cases had no other pathogenic or likely pathogenic mutations in the 977 genes known to be associated with seizure disorders. 15 Mutation P286A had a low allele frequency (4.09 × 10 −6 in general population and 5.567 × 10 −5 in East Asia population) in gnomAD database, but did not present in the control populations of gnomAD.
The other three mutations (K533E, E1646K, and D1659E) did not appear in gnomAD database) (Table S1). A statistical analysis on the frequency of the CHD4 variants showed a significant difference between the present cohort and the populations in gnomAD (general population, East Asian population, or controls in gnomAD (4/964 vs.

| Clinical features of the cases with CHD4 mutations
The CHD4 mutations were identified in four families with 8 individuals affected. The detailed clinical manifestations were summarized in  Figure 2E), except the brother of case 1 whose ECG was not available. Mild aortic valve regurgitation was detected in case 1 ( Figure 2F). Brain MRI was normal in the four cases. All patients had normal intelligence and development, except the proband of case 4 presented borderline intelligence (Wechsler Intelligence Scale 74). They all became seizure-free for at least one year on the routine dose of lamotrigine and/or sodium valproate.

| Structural alteration of CHD4 protein and genotype-phenotype correlation of CHD4 variants
As shown schematically in Figure   As our previous studies showed that molecular sub-regional location of the missense mutations was a critical factor to determine the pathogenicity of variants and associated with phenotypic severity, 2814

| Evaluation of epilepsy as a novel phenotype of CHD4 variants
We evaluated the CHD4 variant epilepsy correlation using ClinGen Clinical Validity Framework. The total allowable points for genetic aspect and experimental aspect were 6.3 and 6, respectively. The results of clinical validity summary matrix were 12.3 points that was categorized as "Strong," supporting the association between CHD4 variants and epilepsy (Table 2). In the five clinical-genetic dimensional evaluation, evidence from four of the five clinical-genetic aspects (repetition, genotype-phenotype correlation, inheritance pattern, and molecular sub-regional implications) was ranked "Yes," suggesting that epilepsy was a novel phenotype of CHD4 variants (Table S4). These findings indicated that CHD4 mutations are potentially a candidate pathogenic gene of epilepsy. Homozygous CHD4 knockout mice have exhibited abnormal cell cycle and even preweaning lethality with complete penetrance.

| DISCUSS ION
Clinically, the majority of CHD4 mutations were associated with varying degrees of multiple congenital anomalies (heart, skeleton, brain, etc). 21 These evidences indicate that CHD4 is essential for normal development in higher organisms. [29][30][31][32][33][34] In the central nerv- substrates are overlapping. 36 All CHD family genes were associated with neurodevelopmental disorders, while CHD1/2/3/7/8 genes were also implicated in epilepsies or seizures. 37,38 The patients with mutations in CHD1/3/7/8 genes had epilepsy as a concomitant symptom of developmental abnormalities, while CHD2 mutations caused epileptic encephalopathy as a core phenotype.
In this study, CHD4 mutations were identified in the patients with childhood idiopathic epilepsy characterized by infrequent seizures, normal background in EEG, normal neurodevelopment, and favorable outcome, indicating that CHD4 was a candidate causative gene of childhood idiopathic epilepsy. However, the mechanism of phenotypic variation caused by CHD family genes is unclear.
From the remodeling mechanisms/biological functions point of view, CHD2 acts as a monomer and is associated with relatively pure phenotype. On the contrary, CHD4 performs functions as one part of NuRD complexes, potentially explaining phenotypes of more complex clinical symptoms. Similarly, CHD7 combines with polybromo-and BRG1-associated factor-containing complex 29,39 and is associated with complex phenotypic spectrum including CHARGE syndrome. 40 In previous studies, CHD4 mutations were associated with severe multisystem developmental abnormalities mostly, while in this study, the mutations were identified in patients with idiopathic childhood epilepsy. Further analysis showed that the CHD4 missense mutations associated with multisystem developmental abnormalities were clustered in the central region from SNF2-like region to DUF 1087 domain, while the variants associated with other mild phenotypes were located outside this region, suggesting a molecular subregional effect. Moreover, seizure-related mutations were located in CHNDT-PHD1linker, CHROMO domain, and DUF1086-CHDCT2 linker, suggesting the mutations in these regions were potentially related to increased susceptibility of epilepsy.
In this study, the seizures in all patients occurred in childhood, which was consistent with increased CHD4 expression in this stage.
Clinically, it is common that CAE and BECTS consecutively or contemporarily coexist in the same patients, 41  In conclusion, this study identified four novel missense mutations in the patients with childhood idiopathic epilepsy with sinus arrhythmia. These mutations presented significantly higher frequent in case cohort than general populations. The strong supporting evidence from Clinical Validity Framework of ClinGen and evidence from four of five clinical-genetic aspects suggested the association between CHD4 variants and epilepsy, suggesting that CHD4 mutations are potentially a candidate pathogenic gene of epilepsy with sinus arrhythmia. The potentially molecular sub-regional effect of missense mutations would help understanding the underlying mechanisms of phenotypic variation.

ACK N OWLED G EM ENTS
The authors are deeply grateful to the patients and clinicians who participated in this work.

CO N FLI C T O F I NTE R E S T
All authors claim that there are no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.