Somatic Mutations in TSC1 and TSC2 Cause Focal Cortical Dysplasia

Abstract

Focal cortical dysplasia (FCD) is a major cause of the sporadic form of intractable focal epilepsies that require surgical treatment. It has recently been reported that brain somatic mutations in MTOR account for 15%-25% of FCD type II (FCDII), characterized by cortical dyslamination and dysmorphic neurons. However, the genetic etiologies of FCDII-affected individuals who lack the MTOR mutation remain unclear. Here, we performed deep hybrid capture and amplicon sequencing (read depth of 100×-20,012×) of five important mTOR pathway genes-PIK3CA, PIK3R2, AKT3, TSC1, and TSC2-by using paired brain and saliva samples from 40 FCDII individuals negative for MTOR mutations. We found that 5 of 40 individuals (12.5%) had brain somatic mutations in TSC1 (c.64C>T [p.Arg22Trp] and c.610C>T [p.Arg204Cys]) and TSC2 (c.4639G>A [p.Val1547Ile]), and these results were reproducible on two different sequencing platforms. All identified mutations induced hyperactivation of the mTOR pathway by disrupting the formation or function of the TSC1-TSC2 complex. Furthermore, in utero CRISPR-Cas9-mediated genome editing of Tsc1 or Tsc2 induced the development of spontaneous behavioral seizures, as well as cytomegalic neurons and cortical dyslamination. These results show that brain somatic mutations in TSC1 and TSC2 cause FCD and that in utero application of the CRISPR-Cas9 system is useful for generating neurodevelopmental disease models of somatic mutations in the brain.

Keywords: CRISPR-Cas9 genome editing; TSC1; TSC2; brain mosaicism; brain somatic mutation; focal cortical dysplasia; intractable epilepsy.

Figure 1

Identification of Brain Somatic Mutations in TSC1 and TSC2 in FCDII Individuals Lacking MTOR Mutations (A) Pre- and post-operative brain MRI and H&E staining of pathological samples from FCDII individuals negative for MTOR mutations. The white arrow and arrowhead indicate the cortical dysplasia and resected brain regions, respectively. The black arrow and arrowhead indicate dysmorphic neurons and balloon cells (FCDIIb), respectively. Scale bars, 100 μm. (B) Domain organization and identified mutations in TSC1 and TSC2. A recent study showed that the C terminus of TSC1 binds to the N terminus of TSC2. Abbreviations are as follows: T2BD, TSC2-binding domain; T1BD, TSC1-binding domain; coil, predicted coiled-coil domain; and GAP, GTPase-activating protein domain. (C) The identified mutation sites in TSC1 and TSC2 encode evolutionarily conserved residues.

Figure 2

The Identified Mutations Induce Hyperactivation of the mTOR Pathway by Disrupting the Formation or Function of the TSC1-TSC2 Complex (A) Schematic figure showing that mTOR kinase activation is regulated by the GAP domain of the TSC complex through hydrolysis of GTP-bound Rheb. (B) Immunoblot analysis of S6K phosphorylation in TSC1 or TSC2 mutant HEK293T cells. HEK293T cells were transiently transfected with Myc-tagged wild-type TSC1 and the indicated TSC1 mutants or with FLAG-tagged wild-type TSC2 and the indicated TSC2 mutant and then treated with rapamycin (200 nM) for 1 hr. Cell lysates were subjected to immunoblot analysis with the indicated antibodies. WT, wild-type. (C) Quantification of the blotting intensity. Data represent the mean ± SEM (n = 3–5 per group). ^∗∗p < 0.01 and ^∗∗∗p < 0.001 compared with the wild-type (Student’s t test). (D) Immunoprecipitation assay of mutant TSC1 and wild-type TSC2. HEK293T cells were transiently co-transfected with Myc-tagged wild-type or mutant TSC1 and FLAG-tagged wild-type TSC2. Lysates were immunoprecipitated with anti-TSC2 antibody and subsequently immunoblotted with anti-Myc antibody. Immunoprecipitation assays of mutant TSC2 and wild-type TSC1 are presented in Figure S5. (E) Quantification of the TSC1 blotting intensity immunoprecipitated with TSC2 antibody. Data represent the mean ± SEM (n = 4 per group). ^∗p < 0.05 compared with the wild-type (Student’s t test). (F) GTP-agarose bead pull-down assay for Rheb in mutant TSC2-expressing cells. HEK293 cells were transfected or co-transfected with Myc-tagged wild-type TSC1, FLAG-tagged wild-type TSC2, or TSC2 p.Val1547Ile. Cell lysates were incubated with GTP-agarose beads, and the GTP-bound materials were analyzed by immunoblotting with anti-Rheb or anti-ARF1 antibodies. Total cell lysates were also immunoblotted with the indicated antibodies. (G) Quantification of the GTP-bound Rheb blotting intensity. Two point substitutions (TSC2 p.Asn1601Lys and p.Asn1609Ser) have been reported to abolish TSC2 GAP activity and thus were used as the enzyme-dead control. Data represent the mean ± SEM (n = 3 per group). ^∗∗p < 0.01 and ^∗∗∗p < 0.001 (Student’s t test). (H) Immunoblot analysis of phosphorylated S6 (P-S6), S6, and Tsc1 in Neuro2A cells carrying monoallelic TSC1 p.Arg22Trp. The level of phosphorylated S6 was significantly increased in the stable cell line with monoallelic TSC1 p.Arg22Trp, indicating that the mTOR pathway was hyperactivated by the heterozygous monoallelic mutation. The level of TSC1 was unaffected by the mutation. (I) Quantification of the blotting intensity of S6 phosphorylation. Data represent the mean ± SEM (n = 3 per group). ^∗∗∗p < 0.001 compared with the wild-type (Student’s t test). Cells were harvested without serum starvation for lysate extraction in all experiments.

Figure 3

The Identified Mutations Are Associated with Aberrant mTOR Activation and Cytomegalic Neurons in Individuals with FCDII (A) Co-immunostaining of pathological samples obtained from FCDII individuals carrying TSC1 or TSC2 mutations for the mTOR pathway marker P-S6, the neuronal marker NeuN, and DAPI. Non-FCD brain specimens were collected from the tumor-free margin of an individual with glioblastoma as part of a planned resection, which was pathologically confirmed as normal brain tissue without a tumor. The percentage of P-S6-positive cells and the soma size of P-S6-positive neuronal cells were measured. Scale bars, 50 μm. (B and C) The bar chart shows the percentage of P-S6-positive cells (B) and the soma size of P-S6-positive neurons (C). The bar charts correspond to an average of three to four representative cortical regions. Data represent the mean ± SEM (n = 39–134 per group). ^∗∗p < 0.01, ^∗∗∗p < 0.001, and ^∗∗∗∗p < 0.0001 compared with the non-FCD sample (Student’s t test).

Figure 4

In Utero Somatic Genome Editing of Tsc1 and Tsc2 Recapitulates the Pathology and Symptoms Observed in Individuals with FCDII (A) Schematic figure of the procedure used to generate the mouse model of brain somatic mutations induced by in utero somatic genome editing in the developing brain. In utero electroporation of CRISPR-Cas9 vectors with a dsRed reporter targeting Tsc1 or Tsc2 was performed at E14, and properly electroporated and delivered mice were screened at birth (P0). Then, they were monitored by video-electroencephalography (EEG) after 3 weeks of age. The arrow indicates the focal expression of the dsRed reporter in the embryonic mouse brain. Scale bar, 3 mm. (B) In utero electroporation of CRISPR vectors expressing selected sgRNA disrupts neuronal migration in the developing mouse neocortex. The images show coronal sections of mouse brains (>P56) electroporated with the CRISPR construct. We counted the percentage of dsRed (+) cells by dividing the number of dsRed (+) cells by the total number of cells (n = 2∼3 per group). Scale bars, 250 μm. (C) Bar charts showing the relative fluorescence intensities reflecting the distribution of electroporated cells within the cortex. Mice electroporated with the CRISPR construct without sgRNA expression served as controls. Data represent the mean ± SEM (n = 3–6 per group). ^∗p < 0.05 and ^∗∗∗p < 0.001 compared with the control (two-way ANOVA with a Bonferroni multiple-comparison test). (D) The EEG wave pattern in the ictal phase of Tsc1 and Tsc2 focal knockout (fKO) mice. EEG signals were recorded from four epidural electrodes located on the left frontal lobe (LF), right frontal lobe (RF), left temporal lobe (LT), and right temporal lobe (RT). Magnified EEG waves of the ictal phase and postictal period are presented in Figure S12. (E) The seizure frequency in genome-edited mice was measured. Control mice were transfected with a CRISPR construct without sgRNA. Data represent the mean ± SEM (n = 3–6 per group). (F) The seizure frequency in mice carrying the CRISPR construct targeting Tsc2 was dramatically reduced by rapamycin treatment. Data represent the mean ± SEM (n = 4–12 per group). ^∗p < 0.05 and ^∗∗p < 0.01 compared with the control (one-way ANOVA with a Bonferroni post-test).

Figure 5

Cytomegalic Neurons Contain the Correctly Edited Genome Sequence of Tsc1 and Tsc2 Targeted by the CRISPR-Cas9 Construct (A) Co-immunostaining for NeuN and DAPI. NeuN staining of dsRed-positive cells is presented. The soma size of dsRed-positive neurons was measured. Scale bars, 20 μm. (B) Bar chart of the neuronal size in each group shows an increased soma size in affected cortical regions. Data represent the mean ± SEM (n = 27–299 per group). ^∗∗∗p < 0.001 compared with dsRed-negative neurons (Student’s t test). (C) Schematic figure of the microdissection procedure used to isolate dsRed-positive cytomegalic neurons and subsequent targeted deep amplicon sequencing. The dsRed-positive cytomegalic neurons are denoted by a white circle in the LCM image (“targeting”). Labeled neurons were microdissected, and 10∼20 cells were collected in an adhesive cap tube. Genomic DNA was then extracted from the collected cells and submitted for deep amplicon sequencing. Scale bars, 50 μm. (D) Representative view of Tsc1 and Tsc2 indels induced by Tsc1 and Tsc2 target sgRNA from the Integrative Genomic Viewer. Black and purple bars indicate deletions and insertions, respectively. The arrowheads indicate the theoretical cutting site of Cas9. PAM, protospacer adjacent motif. (E) The frequency (relative to total reads) and position of indels are shown. The 0 position indicates the theoretical cutting site of Cas9 guided by sgRNA. The detailed amplicon sequencing procedure and deep sequencing data for the control group are presented in Figure S14.

Figure 6

In Utero Application of Cas9n with Modified ssODN Models Cytomegalic Neurons Carrying the TSC2 p.Val1526Ile Variant (A) Co-immunostaining for NeuN and DAPI in mouse brains at P28, which were electroporated in utero with Cas9n-selected sgRNA and modified single-stranded oligonucleotides (ssODNs) targeting mouse TSC2 c.4576G>A (p.Val1526Ile), corresponding to human TSC2 c.4639G>A (p.Val1547Ile). Among the dsRed-positive neurons, a small subset of cytomegalic neurons (“affected cells”) were observed. Scale bars, 25 μm. (B) LCM was used to isolate dsRed-positive cytomegalic neurons. We discerned the affected cells from the unaffected cells by carefully examining the cell size. Prior to capturing the cells, we analyzed the image and selected target cells with a size two times larger than the average neuron size. The white arrowhead and arrow indicate the unaffected and affected cells, respectively. Scale bars, 100 μm. (C) Bar chart of the neuronal size in unaffected and affected cells. Data represent the mean ± SEM (n = 10–21 per group). ^∗∗p < 0.01 compared with dsRed-positive unaffected neurons (Student’s t test). (D) Representative view of Tsc2 targeted by modified ssODNs from the collapsed view of the Integrative Genomic Viewer. The green and yellow bars on the reference genomic sequence indicate the PAM and the target site of sgRNA, respectively. To discerned HDR from random mutations, we inserted several point mutations in ssODNs (asterisk) to induce silence mutations.