Study subjects and samples

The study protocols were approved by the institutional review board of Yokohama City University School of Medicine. Written informed consent was obtained from all study participants. A total of 64 patients, including 18 previously reported patients without candidate variants (patients F02–F25) [24], participated in this study: FCD (n = 44), HME (n = 10), brain tumors (n = 4), hippocampal sclerosis (n = 1), and other malformations of cortical development (n = 5) (Fig. 1, Additional file 1: Figs. S1 and S2, and Table S1). Peripheral blood, saliva, frozen brain tissues, or formalin-fixed paraffin-embedded brain tissues were collected from patients. The brain lesions were surgically excised to alleviate intractable epilepsy. Genomic DNA of the peripheral blood leukocytes and FFPE brain tissues were extracted using QuickGene-610L (KURABO, Osaka, Japan) and QIAamp DNA FFPE Tissue Kit (Qiagen, Hilden, Germany), respectively. Genomic DNA samples were extracted from frozen brain tissues using the standard procedure involving proteinase K digestion, phenol–chloroform extraction, and ethanol precipitation.

Fig. 1

Brain MRI and histopathologic features of patients with somatic variants a Patient F08 at 3 years (FCD type IIB, an MTOR in-frame variant), b, c patient F61 at 2 years (FCD type IIB, germline and somatic variants of TSC2), d patient F48 at 23 years (FCD type IB, a MAP2K1 in-frame deletion), e patient F70 at 21 years (ganglioglioma, deletion of entire chromosome 9), and f patient F30 at 2 years and 7 months (hippocampal sclerosis, 19p13.3p12 deletion). a, b, c, e are T2-weighted axial brain MRIs, and d, f are fluid-attenuated inversion recovery coronal MRIs. Brain MRI showing focal irregular gyri (arrows) with blurred junctions between the cortex and white matter (a) or hyperintensity of the subcortical white matter (b–e). g Patient F67 (FCD type I, a PTPN11 missense variant). Cytoarchitectural abnormality of the cortex. h Patient F08 with FCD type IIB showing several balloon cells in the white matter. i Patient F61 with FCD type IIB exhibiting dysmorphic neurons and balloon cells (arrows) in the cortex. j, k Patient F70 with ganglioglioma. j Astrocytic cells showing a wavy, fascicular arrangement with mild hypercellularity. A dysmorphic ganglion cell (inset in k). k CD34-immunopositive cells with fine processes. g, i and inset in k: Klüver–Barrera stain. h, j hematoxylin and eosin stain. k immunostained and counterstained with hematoxylin. Scale bar = 350 μm for g, 90 μm for h and i, 180 μm for j, 50 μm for inset in k, and 140 μm for k

Deep sequencing using targeted capture

Targeted capture was conducted for DNA samples from brain tissues using the HaloPlex HS Target Enrichment System (Agilent Technologies, Santa Clara, CA, USA) and/or xGen Lockdown Probes with the xGen Prism DNA Library Prep Kit (Integrated DNA Technologies [IDT], Coralville, IA, USA). The data of HaloPlex HS were analyzed using the SureCall software (Agilent Technologies). xGen sequencing data were processed using the Picard, bwa, fgbio, and VarDict software according to the guidelines from IDT and were annotated using Annovar. The thresholds for variant allele frequency (VAF) in both analyses were set at 0.5%. The targeted captures covered 56–270 genes, depending on the capture kits used. These captures included genes associated with the mTOR and RAS/MAPK pathways and genes associated with diseases other than FCD because we used these targeted captures for diseases other than FCD. We filtered the variants based on the following conditions: exclusion of synonymous variants and other variants with a minor allele frequency (MAF) > 0.01 from 1000 Genomes Project, dbSNP138, or Human Genetic Variation Database in Haloplex; and exclusion of synonymous variants and variants with MAF > 0.01 from dbSNP138, NHLBI Exome Sequencing Project (ESP) 6500, The Exome Aggregation Consortium (ExAC), or Tohoku University Tohoku Medical Megabank Organization (ToMMo) 3.5 K in xGen. Then, the remaining variants were assessed using the prediction tools SIFT, Polyphen2, MutationTaster, and CADD, and the sequencing reads were manually checked using the Integrative Genomics Viewer (IGV). As for HaloPlex HS Target Enrichment, the mean depth of coverage was 1249 × (range, 334–6401×), and 96.8% (range, 86.3–99.3%) of the targeted regions was covered by at least 200 reads. For xGen Lockdown Probes, the mean depth of coverage was 1059× (range, 659–1567×), and 95.2% (range, 90.5–97.2%) of the coding region of the targeted Refseq genes was covered by at least 200 reads. Most cases were initially analyzed by targeted capture sequencing, and WES was applied if no candidate variants were found.

WES

We performed WES using the SureSelect XT Human All Exon v5 or v6 kit (Agilent Technologies) or Twist Human Comprehensive Exome Panel (Twist Bioscience, San Francisco, CA, USA). DNA samples from only the brain or blood and brain samples were used. Germline variants were analyzed using GATK UnifiedGenotyper, and somatic variant calling was performed using Mutect2 and Varscan2. The threshold for VAF for somatic variants was set at 5%. In somatic analysis, we focused on common genes between unrelated samples or the genes in the list we made for targeted capture. The mean depth of coverage was 187 × (range, 55–300×) for brain tissues and 176 × (range, 76–288×) for blood or saliva.

Copy number variant (CNV) analysis

To detect somatic CNVs, we analyzed WES data using the eXome-Hidden Markov Model (XHMM) [22, 34] and/or the SNP array with CytoScan HD array (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s instructions. CNVs were confirmed using the SNP array for patients F43 and F70 and droplet digital PCR (ddPCR) for patient F30. ddPCR was performed using the QX200 Droplet Digital PCR System and analyzed using QuantaSoft (both Bio-Rad Laboratories, Hercules, CA, USA). CACNA1A and STK11 probes for the deleted region, TSC2 probes as the non-deleted control, and TERT probes as an internal control for ddPCR were purchased from IDT. Three technical replicate measurements were repeated three times, meaning that a total of nine reactions were performed.

Validation of somatic variants

We performed deep sequencing to validate the somatic candidate variants using the 150–300 bp long PCR amplicons. PCR and sequencing library preparation were performed with PrimeSTAR GXL DNA Polymerase (TaKaRa Bio, Shiga, Japan) and ThruPLEX DNA-Seq Kit (TaKaRa Bio). Libraries were sequenced on the Illumina platform, and data were analyzed as for WES but without PCR deduplication. The read count for VAF was manually inspected using IGV. The variants c.4376C > A, c.4379T > C, and c.6644C > A in MTOR were confirmed with ddPCR, as previously described [24]. For the blood sample of patient F68, we performed allele-specific PCR for the very-low-frequency somatic 5 bp deletion, in addition to deep sequencing. Germline variants were validated using Sanger sequencing.

Plasmid construction

FLAG-HA-pcDNA3.1 vector and pcDNA3-Flag-mTOR-wt vector were purchased from Addgene (plasmid # 52535 and # 26603) [15, 35], and Halo-tagged cDNA clones of MAP2K1, PTPN11, and GAB1 were obtained from Promega (Madison, WI, USA). Site-directed mutagenesis was performed using the Prime STAR GXL polymerase (TaKaRa Bio) to introduce the variants c.4339_4353del or c.4379T > C (as a positive control) in MTOR, c.173_187del in MAP2K1, or c.178G > C in PTPN11 into each plasmid; each created mutant clone was confirmed using Sanger sequencing.

Cell culture and immunoblotting analysis

HEK293T cells were maintained in Dulbecco's modified Eagle's medium (DMEM; FUJIFILM Wako Chemicals, Osaka, Japan) supplemented with 10% fetal bovine serum and 1% penicillin–streptomycin. To evaluate the activity of the mTOR and RAS/MAPK pathways, these cells were transfected with wild-type or mutant vector of MTOR, MAP2K1, or PTPN11 with or without GAB1 using Polyethylenimine (PEI) Max (Polysciences, Warrington, PA, USA).

Twenty-four hours after transfection, the cell medium was replaced with serum-starved DMEM. After 24 h of incubation, the medium for the analysis of MTOR or MAP2K1 was replaced with Dulbecco’s phosphate-buffered saline with MgCl2 and CaCl2 (Sigma-Aldrich, St. Louis, MO, USA) for 1 h [30]. In the cells transfected with the PTPN11 vector with or without GAB1, epidermal growth factor (EGF; Sigma-Aldrich) stimulation was performed in serum-starved conditions for 5–120 min. Then, these cells were lysed using lysis buffer containing 1% Triton X-100, 20 mM Tris–HCl (pH 7.4), 150 mM NaCl, and 1 mM EDTA supplemented with cOmplete Protease Inhibitor Cocktail and PhosSTOP (both from Roche Diagnostics, Basel, Switzerland). These samples were incubated on ice for 30 min and centrifuged at 20,000×g for 20 min at 4 °C. Laemmli sample buffer (Sigma-Aldrich) was added to the supernatant, and the sample was boiled at 95 °C for 5 min.

An equal amount of protein was electrophoresed on 4–12% NuPAGE SDS–PAGE gels (Thermo Fisher Scientific) and transferred to Polyvinylidene difluoride membranes (ATTO, Tokyo, Japan). The total and phosphorylated S6 and ERK proteins were quantified using separate gels with the same amount of protein applied. The membranes were blocked using EzBlock Chemi (ATTO) for 30 min and incubated overnight with the following primary antibodies at 4 °C: anti-FLAG M2 (1:4000 dilution, A8592, Sigma-Aldrich), anti-HaloTag (1:1000 dilution, G921A, Promega), anti-S6 ribosomal protein (1:2000 dilution, #2217, Cell Signaling Technology, Danvers, MA, USA), anti-p-S6 ribosomal protein (Ser235/236; 1:4000 dilution, #4858, Cell Signaling Technology), anti-ERK1/2 (1:2000 dilution, #4695, Cell Signaling Technology), and anti-p-ERK1/2 (Thr202/Tyr204; 1:2000 dilution, #9101, Cell Signaling Technology) antibodies. Horseradish peroxidase (HRP)-conjugated anti-rabbit antibody (1:10,000 dilution, 111-035-003, Jackson ImmunoResearch Laboratories, West Grove, PA, USA) or HRP-conjugated anti-mouse antibody (1:10,000 dilution, 115-035-003, Jackson ImmunoResearch Laboratories) was used as the secondary antibody and was incubated with the membranes for 2 h at room temperature. The membranes were developed using the SuperSignal West Dura Extended Duration Substrate (Thermo Fisher Scientific), and the bands were visualized using a ChemiDoc Touch imaging system (Bio-Rad Laboratories). The data were analyzed using the Image Lab software (Bio-Rad Laboratories). The experiments were performed in triplicate or quadruplicate.

Statistical analysis

All statistical analyses were performed using GraphPad Prism 9 (GraphPad Prism Software, San Diego, CA, USA). Statistical differences were analyzed by one-way ANOVA followed by Dunnett's post-hoc test for immunoblotting analysis of MTOR and ddPCR, two-tailed paired t-test for immunoblotting of MAP2K1, and two-way repeated measurement ANOVA with Sidak’s multiple comparison test for immunoblotting of PTPN11. The level of significance was set at P < 0.05.