A pair of TALENs targeting exon 1 of rat Scn1a (Ensembl: ENSRNOG00000053122) was designed and constructed using a two-step assembly method with a Platinum Gate kit as previously reported (22). Assembled sequence was 5′-TGCAGGATGACAAGATGgagcaaacagtgcttGTACCACCAGGACCTGA-3′, where uppercase and lowercase letters indicate TALEN target sequences and spacer sequence, respectively. TALENs were microinjected into fertilized eggs of Fisher 344 (F344) rats and transferred into the oviducts of pseudopregnant female Wistar rats, as previously described (23). Genomic DNA was extracted from the tail using the GENEXTRACTOR TA-100 automatic DNA purification system (Takara Bio) and amplified with specific primer sets (forward 5′-TCCTCACTTGTTGGGTCTCA-3′, reverse 5′-TCAGGGTGACTTCAGCATTTC-3′). The polymerase chain reaction (PCR) products were directly sequenced using the BigDye terminator v3.1 cycle sequencing mix and the standard protocol for an Applied Biosystems 3130 DNA Sequencer (Life Technologies). Rats from the fifth generation or later were used in all experiments. Genotypes were assessed by running the PCR products from ear DNA on a Caliper electrophoresis system at postnatal day (P9). Only male rats were used in experiments to eliminate sex differences.

This study was approved by the Animal Research Committee of Kyoto University and the Institutional Animal Care and Use Committee of the Jikei University School of Medicine (protocol number: 2015-125C7). All procedures were conducted according to the Guidelines of the Proper Conduct of Animal Experiments of the Science Council of Japan (2006). The principles outlined in the ARRIVE guidelines, including the 3R concept, were followed when planning the experiments. P15 to 38 rats were used in this study (n =162, body weight range 21–124 g). Rats were weaned on P22 and housed in polycarbonate cages under temperature-controlled conditions (temperature: 24–25°C; relative humidity: 50–60%) with a 12/12-h light–dark cycle. They had free access to water and pelleted food (CE-2, CLEA Japan, Inc.).

Protein extraction from rat brains was performed at P13 using an ULTRARIPA kit for lipid rafts (F015; BioDynamics Laboratory Inc.); the reagent solution contained 50 mM Tris-HCl (pH 8.0), 150 mM NaCl, 1% NP-40 alternative, 0.1% SDS, and 0.5% sodium deoxycholate. Proteins were separated using SDS-PAGE and transferred to polyvinylidene difluoride membranes (Millipore). The membranes were blocked with 5% nonfat skim milk overnight at 4°C and incubated in primary antibody solution containing rabbit anti- voltage-gated sodium channel alpha subunit 1 (Na_V1.1) (ASC-001, 1:200; Alomone Labs) and mouse anti-β-actin (A1978, 1:1000; Sigma-Aldrich) for 1 h at room temperature and sequentially incubated with secondary antibodies containing anti-Rb IgG HRP (AP307P, 1:2000; Millipore) and anti-mouse IgG HRP (AP308P, 1:2000; Millipore) for 1 h at room temperature. The signal was detected using Chemiluminescence HRP Substrate (Takara Bio) on the LAS3000 Mini apparatus (FUJI FILM). Semiquantitative analysis of the signal was performed using ImageJ software (National Institute of Health).

All surgical procedures were performed under isoflurane anesthesia (4–1.5%) at P14. Epidural screw electrodes for EEG were mounted on the skull bilaterally over the somatosensory cortex (2.0 mm lateral to midline, 2.0 mm posterior to bregma) and the cerebellum (1.0 mm lateral to midline, 2.0 mm posterior to lambda) as reference electrodes. Electromyography (EMG) electrodes were inserted into the suprascapular area. Heat-induced seizures were triggered in rats as described below, and behavioral seizures, video–EEG, and EMG were recorded using a PowerLab system (AD Instruments) at P21. Data analysis was performed with LabChart software (AD Instruments).

Rats were initially anesthetized with 3% isoflurane in 30/70% O₂/N₂ in a closed chamber. Rats were laid in the prone position on a dedicated animal bed heated with warm circulating water. Anesthesia with 1.5–2.5% isoflurane in 30/70% O₂/N₂ was delivered to the spontaneously breathing animals through a snout mask, and the respiratory rate was maintained at 50–60 breaths/min. The rectal temperature was carefully monitored and maintained at 36.5 ± 0.5°C by circulating warm water under the bed.

Imaging was performed using a 9.4 T MRI 94/20 system (Bruker BioSpin GmbH) with a mouse head surface coil and a 72-mm transmit coil. We acquired T1 maps for the MRI study. The scanning parameters were as follows: T1WI for T1 map: echo time, 6.5 ms; repetition times, 525, 1050, 2100, 4200, and 8400 ms; averages, 2; scan time, 1 h 3 min 28 s 350 ms; rapid acquisition with relaxation enhancement factor, 1; slices, 24; slice thickness, 0.8 mm; image matrix, 140 × 140; field of view, 16.0 × 16.0 mm; resolution, 0.114 × 0.114 mm; and slice gap, 0.2 mm. T1 value for each ROI was calculated for each T1 map using ImageJ software (National Institutes of Health).

An isotonic solution of MnCl₂·4H₂O (203734; Sigma-Aldrich) was prepared at a concentration of 100 mM in 100 mM bicine solution, as previously described (24). Then, the MnCl₂ solution was diluted to 50 mM with saline, and the pH was adjusted to 7.4.

We intraperitoneally administered a 50 mM solution of MnCl₂ immediately after the initial MRI acquisition, twice at 1-h intervals, to get a total dose of 66 mg/kg. MEMRI acquisition was performed 24 h after the first administration of MnCl₂ based on the kinetics of intracerebral Mn²⁺ concentration (11). The imaging area was adjusted to the same position as in the initial MRI to facilitate comparison of T1 values at the same ROI.

The effective systemic dose of bumetanide on neonatal seizures in rats is 0.1–0.5 mg/kg (21, 25). We administered 0.2 mg/kg of bumetanide twice a day in repeated doses as previously described (26). Bumetanide (B3023; Sigma-Aldrich) was dissolved in 100% ethyl alcohol and then diluted with 0.9% saline to a final concentration of 0.05 mg/mL. Bumetanide solution (0.2 mg/kg) or equal volumes of 0.9% saline (4 mL/kg) was injected twice daily from P12 to P20 or P26. MEMRI was performed on P21 or P27 for each rat.

Scn1a^+/− rats were subjected to heat-induced seizures on P21 or P27 after daily administration of bumetanide or 0.9% saline from P12 to P20 or P26. Hyperthermia was induced by placing the rats in a bath filled 10 cm deep with 45°C water, as previously described (27), for a maximum of 5 min or until a seizure occurred. When a seizure occurred, the rats were removed immediately from the bath and monitored until recovery. For each rat, we recorded the latency from water contact to seizure onset, the seizure duration, and a score based on the most severe seizure observed. Seizure severity was scored based on Racine stages (28): 0, no response; 1, oral and facial movement; 2, head nodding; 3, forelimb clonus; 4, forelimb clonus with rearing; and 5, generalized tonic-clonic seizures and falling. All procedures were recorded using a video camera. We did not measure body temperature because it was difficult and would be inaccurate to measure body temperature in a water bath.

Statistical analysis was performed using SAS 9.4 software (SAS Institute Inc.) and GraphPad Prism 8 software (GraphPad Software Inc.). We used ANCOVA to compare T1_post (T1 values resulting from MEMRI T1 mapping) and adjusted for native T1 values (T1_pre), applying a split-split-plot design under different genotypes, developmental stages, and regions (29). Genotypes and developmental stages were applied to whole plots (animals), and regions were applied to split-split-plots (regions in an animal) in the MEMRI experiment at each stage of development. In the MEMRI experiment with and without bumetanide, genotype and drug (bumetanide or saline) were applied to whole plots (animals), and regions were applied to split-split-plots (regions in an animal). The replicate effect was combined into whole-plot error, including Error(A) and Error(B) (29), and Type III Sum of Squares in SAS was used as the main result. The Scheffé or Tukey multiple-comparison test was performed with split-split-plot error to compare genotypes in each region at each developmental stage or to compare bumetanide and normal saline at each region in each genotype. Differences in seizure phenotypes were assessed using the Mann–Whitney U-test. The p-values were two-sided, and p < 0.05 was considered statistically significant.

In this study, we developed a global Scn1a knockout rat model using the TALEN-mediated genome editing technique by generating a frameshift in exon 1 of the Scn1a gene (Figure 1A). TALEN mRNAs targeting Scn1a exon 1 were microinjected into fertilized eggs of F344 rats, and a 94-base-pair deletion in exon 1 was identified in a resulting founder mutant rat (F344-Scn1a^em1Kyo). Reduced Na_V1.1 protein expression was confirmed in the brains of heterozygous (Scn1a^+/−) and homozygous (Scn1a^−/−) knockout rat offspring (Figures 1B, C and Supplementary Figure 1).

As previously reported in Scn1a knockout mice (9, 18–20), Scn1a^−/− rats died at approximately the second postnatal week, but no spontaneous deaths were observed in Scn1a^+/− rats until at least P200 (Figure 1D). The mean body weight of Scn1a^+/− rats was not significantly different from that of wild-type rats during the experimental period (Figure 1E). Heat-induced seizures were observed in all investigated Scn1a^+/− rats (n = 26) at 45°C water, a temperature that did not induce seizures in wild-type rats (Figure 1F, Supplementary Video 1). Up to the third postnatal week, spontaneous convulsive seizures were rarely observed and EEG recordings showed no interictal epileptiform discharges or ictal activities during recording in the Scn1a^+/− rats (Figure 1G). Notably, in older Scn1a^+/− rats, some seizures were provoked by acoustic or vibrational stimuli. These findings indicate that Scn1a^+/− rats have milder symptoms than those reported in Scn1a^+/− mice, suggesting their phenotype may be closer to the genetic epilepsy with febrile seizures plus phenotype.

To identify disease-specific alterations in regional brain activities of developing Scn1a^+/− rats, we employed T1 mapping in a MEMRI experiment. This technique represents the T1 relaxation time as the T1 value for each pixel; decrease in the T1 value in MEMRI indicates increase in neural activity. Several MEMRI studies have used T1 mapping to quantitatively compare brain activity as T1 value or R1(1/T1) and successfully identified regions that show disease-specific changes in brain activity (30, 31). Therefore, we applied this technique to quantitatively compare the regional activity of wild-type and Scn1a^+/− rats. As T1 values might be influenced by myelination (32, 33), we first investigated T1_pre before performing MEMRI in wild-type (n = 44, body weight range 21–118 g) and Scn1a^+/− rats (n = 44, 25–124 g) from P15 to P38 (Supplementary Figure 2). MnCl₂ was intraperitoneally administered immediately afterwards, and MEMRI T1 mapping was performed on the same rats to investigate T1_post (Figure 2A). We compared T1_post values, adjusted for T1_pre, under different genotypes, developmental stages, and regions. Region of interests (ROIs) were defined in 17 areas in the cortex, subcortex, and cerebellum (Figure 2D). The developmental stages of rats were defined according to a previous report (34): P15–18 (neonatal period), P19–22 (infancy), P23–26 (weaning period), P27–30 (early juvenile period), P31–34 (juvenile period), and P35–38 (prepuberty). If T1_post or T1_pre could not be measured in a region, both values were excluded from the analysis. The sample size for each postnatal day and each ROI is shown in Table 1. Analysis of covariance (ANCOVA) for T1_post showed neither a significant main effect of genotype [F_{(1, 76)} = 3.19, p = 0.08] nor an interaction among the developmental stages, genotypes, and ROIs [F_{(80, 1194)} = 1.28, p = 0.05]. In Scn1a^+/− rats, however, a significant decrease in T1_post was observed compared to wild-type rats at P19–22 in all ROIs, except the cerebellar vermis, by the Tukey multiple-comparison test with respect to the interaction (p < 0.001; cerebellar vermis, p = 0.11), while the decrease was not significant at P15–18 or P23–34 (Figures 3A, B). At P35–38, a significant decrease reappeared in the hippocampal CA3, lateral habenular nucleus, and basal ganglia (CA3, p < 0.001; lateral habenular nucleus, p = 0.007; globus pallidus, p = 0.05; caudate-putamen, p < 0.001). These results demonstrate that neural activity across widespread brain regions is significantly increased in Scn1a^+/− rats at P19–22, which corresponds to the peak onset age of DS in humans (1, 4, 35).

The decrease in T1_post in Scn1a^+/− rats occurred transiently at P19–22, and later reached the same level as wild-type rats. Thus, we speculated that the impairment of a brain developmental process that occurs at this age might contribute to this phenomenon. We focused on the maturation process of inhibitory neural circuits, which is influenced by developmental changes in intracellular chloride concentration ([Cl⁻]_i). A decrease in [Cl⁻]_i promotes depolarizing to hyperpolarizing (D-H) switching and further enhances the inhibitory effect of type A gamma-aminobutyric acid (GABA_A) receptor signaling (36, 37). On the other hand, elevated [Cl⁻]_i causes depolarizing GABA_A receptor signaling, which is observed in immature neurons primarily as a result of higher expression of the chloride importer NKCC1 (32, 36). Conversely, mature neurons express higher levels of the chloride exporter KCC2, which results in decreased [Cl⁻]_i, leading to hyperpolarizing GABA_A receptor signaling (32, 36). In rodents, [Cl⁻]_i generally decreases by P20 (38). As a result, the D-H switch is completed by P14 (36, 37, 39), and especially by P17 in males (40, 41). Therefore, we hypothesized that the transient increase in brain activity is associated with delayed [Cl⁻]_i decrease. We performed a rescue experiment by administering bumetanide, an NKCC1 inhibitor (32), to Scn1a^+/− rats which would decrease [Cl⁻]_i and promote the maturation of inhibitory neural network. Bumetanide or normal saline (control) was administered intraperitoneally every day from P12–20 in wild-type rats (bumetanide-treated: n = 6, body weight range 31–39 g; saline-treated: n = 6, 31–38 g) and Scn1a^+/− rats (bumetanide-treated: n = 6, 32–40 g; saline-treated: n = 6, 33–37 g), and MEMRI was performed on P21 (Figure 2B). ANCOVA for T1_post showed no significant interaction among genotypes, ROIs, and drug administration [F_{(16, 319)} = 1.58, p = 0.07]. However, Scheffé multiple-comparison test with respect to the interaction showed that T1_post significantly increased in all ROIs, except the lateral habenular nucleus, in bumetanide-treated Scn1a^+/− rats compared to those in saline-treated Scn1a^+/− rats (p < 0.05; lateral habenular nucleus, p = 0.06), while no significant difference was observed between bumetanide- and saline-treated wild-type rats (Figure 4A). The caudate-putamen, lateral globus pallidus, retrosplenial granular b cortex, dentate gyrus, primary somatosensory cortex, reticular thalamic nucleus, and cerebellar hemisphere of Scn1a^+/− rats had remarkable increases in T1_post (p < 0.001). Based on these results, we speculated that the immature inhibitory neural network due to delayed decrease in [Cl⁻]_i is involved in the increased brain activity in Scn1a^+/− rats at P19–22.

To investigate whether bumetanide also relieves seizures in Scn1a^+/− rats, we examined changes in seizure characteristics in bumetanide-treated Scn1a^+/− rats. After the administration of bumetanide (n = 6, body weight range 36–39 g) or normal saline (n = 6, 32–41 g) to Scn1a^+/− rats as previously described, seizures were induced on P21 by a 45°C-hot water bath (Figure 2C). Seizure latency, seizure duration, and Racine's seizure score were compared between bumetanide- and saline-treated Scn1a^+/− rats. Although there was no significant difference in seizure duration (166.5 s [135.3–204.8] vs. 144.0 s [135.8–163.0], Mann–Whitney U test, U = 12.0, Z = −0.962, p = 0.34) or seizure score (4.0 [3.8–4.0] vs. 3.5 [2.8–4.0], Mann–Whitney U test, U = 11.5, Z = −1.251, p = 0.21), the seizure latency was significantly prolonged in bumetanide-treated rats (183.0 s [173.3–196.0] vs. 158.0 s [154.0–170.0], Mann–Whitney U test, U = 2.0, Z = −2.567, p = 0.01) (median [interquartile range]) (Figure 4B). These results indicate that bumetanide increased the seizure threshold but did not mitigate seizure severity in Scn1a^+/− rats.

We conducted a similar experiment in Scn1a^+/− rats at P27 to confirm whether the transiently observed decrease in T1_post meant that [Cl⁻]_i decrease was completed after P21. After daily bumetanide or normal saline administration from P12 to P26, we compared T1_post between wild-type (bumetanide-treated: n = 6, 44–58 g; saline-treated: n = 6, 44–63 g) and Scn1a^+/− (bumetanide-treated: n = 6, 62–64 g; saline-treated: n = 6, 60–64 g) rats using MEMRI at P27 (Figure 2B). ANCOVA for T1_post showed no significant interaction among genotypes, ROIs, and drug administration [F_{(16, 319)} = 0.43, p = 0.97]. Tukey multiple-comparison test with respect to the interaction showed no significant difference in T1_post between bumetanide- and saline-treated Scn1a^+/− rats in any ROI (Figure 5A). We also compared heat-induced seizure characteristics of bumetanide-treated (n = 7, 56–66 g) and saline-treated (n = 7, 58–69 g) Scn1a^+/− rats on P27 (Figure 2C). There were no significant differences in seizure latency (113.0 s [106.0–130.0] vs. 120.0 s [113.0–148.0], Mann–Whitney U-test, U = 16.5, Z = −1.027, p = 0.31), seizure duration (157.0 s [122.0–182.0] vs. 168.0 s [124.0–209.0], Mann–Whitney U test, U = 21.0, Z = −0.447, p = 0.66), or Racine's seizure score (5.0 [5.0–5.0] vs. 5.0 [5.0–5.0], Mann–Whitney U-test, U = 24.5, Z = 0.000, p = 1.00) (Figure 5B). These results support the hypothesis that in Scn1a^+/− rats, the [Cl⁻]_i decreased in widespread brain regions during the fourth postnatal week contributing to an overall improvement in seizure threshold.