2.1 Ethical compliance

This study was approved by the institutional review board of the hospital (No. 1032). Informed consent was obtained from the patient's family for agreement to publish his scientific research study with unidentified individual clinical data. Clinical features, electroencephalography signals, brain magnetic resonance imaging (MRI) results, routine biochemical examination results, organ investigations, and genetic screening were also analyzed.

2.2 Exome sequencing

Peripheral blood samples were collected from the proband and his family. A total of 1.0 μg of genomic DNA from each sample was sheared into 200–300 bp fragments, which were subsequently analyzed with a SureSelect clinical research exome V2 (Agilent, USA) exon capture kit following the official guide. Sequencing was performed on the Illumina NovaSeq 6000 platform with a 150PE sequencing strategy.

After removing low-quality reads and sequencing adapters, the clean data were mapped to the human reference genome (hg19) by the Burrows–Wheeler Aligner (Li & Durbin, 2009). SNPs and short insertions or deletions were analyzed by GATK. Variants were annotated with ANNOVAR (Wang et al., 2010), and those located in exonic and splicing regions with a minor allele frequency (MAF) ≤0.0001/0.005 in the SNP database (Exome Aggregation Consortium [ExAC], 1000 Genomes, gnomAD) were obtained for further analysis. The pathogenicity of the variants was evaluated according to the ACMG guidelines (Richards et al., 2015). Candidate pathogenic variants were further validated by Sanger sequencing. All the variants detected were filtered by MAF, and the variant types (exonic and splicing) are listed in Table S1.

2.3 Wild-type and variant vector construction

In the general minigene assay of CNKSR2, the wild-type and variant fragments were cloned and inserted into an expression vector to investigate their effect on mRNA splicing. The wild-type target fragment was amplified by CNKSR2-F and CNKSR2-R, and nested PCR was subsequently performed with pECMV-CNKSR2-KpnI-F and pECMV-CNKSR2-EcoRI-R. The variant fragment of CNKSR2 was generated using an overlap PCR strategy. The 3′ fragment was amplified with the primer pair CNKSR2-mut-F and pcMINI-CNKSR2-EcoRI-R. The 5′ fragment was amplified with the primer pair pcMINI-CNKSR2-KpnI-F and CNKSR2-mut-R. The 3′ and 5′ fragments were then mixed in a 1:1 ratio and amplified with the primer pair pcMINI-CNKSR2-KpnI-F and pcMINI-CNKSR2-EcoRI-R. The primers used are as follows:

The wild-type and variant fragments and the pECMV vector were digested with the EcoRI and KpnI restriction endonucleases. The digested CNKSR2 then fragments were subsequently ligated to the pECMV vector and validated via Sanger sequencing.

2.4 Cell culture and transient transfection

The 293T cell line (GDC0187, CCTCC) was cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum. Wild-type and variant vectors were transfected into 293T cells by Lipofectamine 2000 Transfection Reagent (11668019, Invitrogen, Waltham, MA, USA) following the manufacturer's instructions.

2.5 RT-PCR

Total RNA was extracted from wild-type and variant 293T cells with an RNAiso Plus Kit (108/9109, TaKaRa). First-strand cDNA was generated with a gDNA Eraser Kit (RR047A, TaKaRa). Wild-type and variant fragments were amplified from pECMV-CNKSR2-KpnI-F and pECMV-CNKSR2-EcoRI-R, purified and investigated by Sanger sequencing. The PCR cycle condition was an initial denaturation at 98°C for 3 min; 40 cycles of 98°C 10 s, 60°C 5 min and 72°C 20 s; and a final extention at 72°C for 5 min. PCR amplification reaction system (25 μL) was 2× PrimeSTAR Max Premix 12.5 μL, primer F/R 0.5 μL (10 μM), template DNA 1 μL (100 ng/μL), and ddH₂O 10.5 μL.

3.1 Case presentation

An 8-year-old boy was admitted to the hospital for epilepsy and developmental delay. He had a tonic–clonic seizure when he was 4 years old, and the symptoms included loss of consciousness, a tilted head, an enlarged eye, cyanotic lips, and stiff limbs that had persisted for approximately 1–2 min. Four months later, another two major episodes of tonic–clonic seizures started at night and lasted for 1–2 min. His head MRI showed a right temporal polar arachnoid cyst, and a dynamic electroencephalogram (2019.06.12) indicated epileptiform discharge (mainly in the left parietooccipital area). The patient was treated with sodium valproate oral solution (40 mg/mL) gradually increasing from the initial dose (80 mg, q12h) to the control dose (90 mg, q12h). Finally, the epilepsy was well controlled.

At the ages of 5 and 7 months, the focal motor seizures were triggered by fever (body temperature of 38°C), which persisted for 10 min and manifested as right upper limb shaking. The condition was controlled again with sodium valproate oral solution (100 mg, q12h).

At the age of 7, he had onset multifocal unilateral impaired awareness seizures with no obvious cause, with symptoms of sudden falling to the right side (when standing or sitting), slow blinking of the right eye, slow shaking of the right hand, and loss of consciousness. The blood concentration of sodium valproate was 43.7 μg/mL, which was lower than the normal range (ref: 50–120 μg/mL). The treatment strategy was changed to sodium valproate tablets (200 mg, tid), and the seizures were controlled again.

At the age of 8, seizures occurred again, but the current blood concentration of valproate was 58.1 μg/mL (ref: 50–120 μg/mL); therefore, the treatment strategy was not adjusted. His parents reported frequent seizures approximately 1–4 times per month.

Reviewing his development history, we found that his development was delayed, as he walked later (1.5 years old) and only mastered simple or repeat words at 4 years of age. In his family, his brother also exhibited developmental delay; he started to walk at 3 years, spoke at 5 years, and had ADHD.

3.2 Identification of splicing variants in CNKSR2

An X-linked splicing variant of CNKSR2 (NM_014927.5: c.1657+1G>A) was identified by whole-exon sequencing (the reference genome version was hg19). In addition, no candidate CNVs associated with the patient's phenotype were detected. The variant detected in our patient was inherited from his mother and was confirmed by Sanger sequencing (Figure 1a,b). His brother, who had a similar phenotype, was also detected to have the same splicing variant. The splicing variant was not observed in the gnomAD, ExAC, or ClinVar databases. Moreover, five other splicing variants of CNKSR2 were reported (Ito & Nagata, 2022), one in the CRIC domain around the N-terminus and the other located around the PH domain (Figure 1d). c.1657+1G>A in our patient was classified as likely pathogenic according to the ACMG guidelines. The variant in our patient was uploaded to the ClinVar database (accession number: SCV004176289).

3.3 Pathogenic variant validation by minigene

To further investigate the effect of the identified CNKSR2 splicing variant, a minigene assay was designed and performed. After confirming the presence of the wild-type and variant genes in the expression vectors (Figure 2a), vector transfection, RNA isolation, reverse transcription, and target fragment amplification were performed. A larger band was observed in the variant sample, which indicated possible intron retention (Figure 2b). Sanger sequencing of the target fragments revealed a 166 bp intron retention region between exons 14 and 15 (Figure 2c,d).