Editor's evaluation: Scn1a-GFP transgenic mouse revealed Nav1.1 expression in neocortical pyramidal tract projection neurons

Abstract

Full text Figures and data Side by side Abstract Editor's evaluation Introduction Results Discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract Expressions of voltage-gated sodium channels Nav1.1 and Nav1.2, encoded by SCN1A and SCN2A genes, respectively, have been reported to be mutually exclusive in most brain regions. In juvenile and adult neocortex, Nav1.1 is predominantly expressed in inhibitory neurons while Nav1.2 is in excitatory neurons. Although a distinct subpopulation of layer V (L5) neocortical excitatory neurons were also reported to express Nav1.1, their nature has been uncharacterized. In hippocampus, Nav1.1 has been proposed to be expressed only in inhibitory neurons. By using newly generated transgenic mouse lines expressing Scn1a promoter-driven green fluorescent protein (GFP), here we confirm the mutually exclusive expressions of Nav1.1 and Nav1.2 and the absence of Nav1.1 in hippocampal excitatory neurons. We also show that Nav1.1 is expressed in inhibitory and a subpopulation of excitatory neurons not only in L5 but all layers of neocortex. By using neocortical excitatory projection neuron markers including FEZF2 for L5 pyramidal tract (PT) and TBR1 for layer VI (L6) cortico-thalamic (CT) projection neurons, we further show that most L5 PT neurons and a minor subpopulation of layer II/III (L2/3) cortico-cortical (CC) neurons express Nav1.1 while the majority of L6 CT, L5/6 cortico-striatal (CS), and L2/3 CC neurons express Nav1.2. These observations now contribute to the elucidation of pathological neural circuits for diseases such as epilepsies and neurodevelopmental disorders caused by SCN1A and SCN2A mutations. Editor's evaluation Using a newly developed Scn1a promoter driven GFP mouse line, the authors convincingly show that GFP expression largely replicates the endogenous expression of Nav1.1. Additionally, they credibly identify inhibitory and excitatory neurons in the cortex that express Nav 1.1. This mouse line provides a valuable resource, especially for epilepsy and autism research, as it offers a reliable tool that can be used to identify specific cell populations that potentially cause disease-related symptoms such as seizures, ataxia, sociability deficits, learning and memory problems, and sudden unexpected death in epilepsy. https://doi.org/10.7554/eLife.87495.sa0 Decision letter Reviews on Sciety eLife's review process Introduction Voltage-gated sodium channels (VGSCs) play crucial roles in the generation and propagation of action potentials, contributing to excitability and information processing (Catterall, 2012). They consist of one main pore-forming alpha- and one or two subsidiary beta-subunits that regulate kinetics or subcellular trafficking of the alpha subunits. Human has nine alpha (Nav1.1–Nav1.9) and four beta (beta-1–beta-4) subunits. Among alphas, Nav1.1, Nav1.2, Nav1.3, and Nav1.6, encoded by SCN1A, SCN2A, SCN3A, and SCN8A, respectively, are expressed in central nervous system. SCN3A is mainly expressed embryonically (Brysch et al., 1991), and SCN1A, SCN2A, and SCN8A are major alphas after birth. Although these three genes show mutations in a wide spectrum of neurological diseases such as epilepsy, autism spectrum disorder (ASD), and intellectual disability, two of those, SCN1A and SCN2A, are major ones (reviewed in Yamakawa, 2016; Meisler et al., 2021). To understand the circuit basis of these diseases, it is indispensable to know the detailed distributions of these molecules in the brain. We previously reported that expressions of Nav1.1 and Nav1.2 seem to be mutually exclusive in many brain regions (Yamagata et al., 2017). In adult neocortex and hippocampus, Nav1.1 is dominantly expressed in medial ganglionic eminence-derived parvalbumin-positive (PV-IN) and somatostatin-positive (SST-IN) inhibitory neurons (Ogiwara et al., 2007; Lorincz and Nusser, 2008; Ogiwara et al., 2013; Li et al., 2014; Tai et al., 2014; Tian et al., 2014; Yamagata et al., 2017). In the neocortex, some amount of Nav1.1 is also expressed in a distinct subset of layer V (L5) excitatory neurons (Ogiwara et al., 2013), but their natures were unknown. In the hippocampus, Nav1.1 seems to be expressed in inhibitory but not in excitatory neurons (Ogiwara et al., 2007; Ogiwara et al., 2013). In contrast, a major amount of Nav1.2 (~95%) is expressed in excitatory neurons including the most of neocortical and all of hippocampal ones, and a minor amount is expressed in caudal ganglionic eminence-derived inhibitory neurons such as vasoactive intestinal polypeptide (VIP)-positive ones (Lorincz and Nusser, 2010; Yamagata et al., 2017; Ogiwara et al., 2018). However, a recent study reported that a subpopulation (more than half) of VIP-positive inhibitory neurons is Nav1.1-positive (Goff and Goldberg, 2019). VGSCs are mainly localized at axons and therefore it is not always easy to identify their origins, the soma. To overcome this, here in this study we generated bacterial artificial chromosome (BAC) transgenic mouse lines that express GFP under the control of Scn1a promoters, and we carefully investigated the GFP/Nav1.1 distribution in mouse brain. Our analysis confirmed that expressions of Nav1.1 and Nav1.2 are mutually exclusive and that in neocortex Nav1.1 is expressed in both inhibitory and excitatory neurons while in hippocampus only in inhibitory but totally absent in excitatory neurons. Furthermore by using a transcription factor FEZF2 (FEZ family zinc finger protein 2 transcriptional factor), also referred to as Fezl, Fez1, Zfp312, and Fez, as a marker for L5 pyramidal tract (PT) neurons (Inoue et al., 2004; Chen et al., 2005; Chen et al., 2008; Molyneaux et al., 2005; Lodato et al., 2014; Matho et al., 2021), and a transcription factor TBR1 which suppresses FEZF2 expression and therefore does not overlap with FEZF2 (Han et al., 2011; McKenna et al., 2011; Matho et al., 2021), we found that most of L5 FEZF2-positive neurons are GFP-positive while L5/6 TBR1-positive neurons are largely GFP-negative and Nav1.2-positive. These results proposed that Nav1.1 is expressed in L5 PT while Nav1.2 in L5/6 non-PT neurons such as L5/6 cortico-striatal (CS) and L6 cortico-thalamic (CT) projection neurons. A majority of L2/3 excitatory neurons express Nav1.2 but a minor subpopulation are GFP-positive, suggesting that most of cortico-cortical (CC) projection neurons express Nav1.2 but the distinct minor population express Nav1.1. These results refine the expression loci of Nav1.1 and Nav1.2 in the brain and should contribute to the understanding of circuit mechanisms for diseases caused by SCN1A and SCN2A mutations. Results Generation and verification of Scn1a-GFP transgenic mouse lines Scn1a-GFP founder mice were generated from C57BL/6J zygotes microinjected with a modified Scn1a-GFP BAC construct harboring all, upstream and downstream, Scn1a promoters (Nakayama et al., 2010; Figure 1A) (see Materials and methods for details). Western blot analysis (Figure 1B) and immunohistochemistry (Figure 1—figure supplement 1) showed robust GFP expression and mostly normal expression levels of Nav1.1 in Scn1a-GFP mouse lines #184 and #233. Both lines showed a similar distribution of chromogenic GFP immunosignals across the entire brain (Figure 2A–H), and a similar distribution was also obtained in fluorescence detection of GFP (Figure 2I–L and Figure 2—figure supplement 1). In neocortex (Figure 2B, F, J and Figure 2—figure supplement 1B, F), GFP-positive cells were distributed throughout all cortical layers. In hippocampus (Figure 2C, G, K and Figure 2—figure supplement 1C, G), cells with intense GFP signals, which are assumed to be PV-IN and SST-IN (Ogiwara et al., 2007; Tai et al., 2014) (see also Figure 8), were scattered in stratum oriens, pyramidale, radiatum, lucidum, and lacunosum-moleculare of the CA (cornu ammonis) fields, hilus and molecular layer of dentate gyrus. Of note, somata of dentate granule cells were apparently GFP-negative. CA1–3 pyramidal cells were twined around with fibrous GFP immunosignals. We previously reported that the fibrous Nav1.1 signals clinging to somata of hippocampal CA1–3 pyramidal cells were disappeared by conditional elimination of Nav1.1 in PV-INs but not in excitatory neurons, and therefore concluded that these Nav1.1-immunopositive fibers are axon terminals of PV-INs (Ogiwara et al., 2013). As such, GFP signals are fibrous but do not form cell shapes in the CA pyramidal cell layer (Figure 2C, G, K and Figure 2—figure supplement 1C, G), and therefore these CA pyramidal cells themselves are assumed to be GFP-negative. These observations further confirmed our previous proposal that hippocampal excitatory neurons are negative for Nav1.1 (Ogiwara et al., 2007; Ogiwara et al., 2013). In cerebellum (Figure 2D, H, L and Figure 2—figure supplement 1D, H), GFP signals appeared in Purkinje, basket, and deep cerebellar nuclei cells, again consistent to the previous reports (Ogiwara et al., 2007; Ogiwara et al., 2013). In the following analyses, we used the line #233 which shows stronger GFP signals than #184. Figure 1 with 1 supplement see all Download asset Open asset Generation of Scn1a-GFP mice. (A) Schematic representation of the modified bacterial artificial chromosome (BAC) construct containing the Scn1a-GFP transgene. A green fluorescent protein (GFP) reporter cassette consisting of GFP cDNA and a polyadenylation signal was inserted at the ATG initiation codon in the coding exon 1 of Scn1a. Filled and hatched boxes indicate the coding and non-coding exons of Scn9a and Scn1a. Arrows indicate the start sites and orientation of transcription of Scn1a. (B) Western blot analysis for Scn1a-GFP and endogenous Nav1.1. The whole cytosolic fractions from 5W Scn1a-GFP brains (lines #184 and #233) were probed with anti-GFP and their membrane fractions were probed with anti-Nav1.1 antibodies. β-Tubulin was used as an internal control. pA, polyadenylation signal; Tg, hemizygous Scn1a-GFP transgenic mice; Wt, wild-type littermates. Figure 1—source data 1 Raw and annotated immunoblots for Figure 1B. https://cdn.elifesciences.org/articles/87495/elife-87495-fig1-data1-v1.zip Download elife-87495-fig1-data1-v1.zip Figure 2 with 2 supplements see all Download asset Open asset Distributions of green fluorescent protein (GFP) signals in brains are similar among Scn1a-GFP mouse lines. Chromogenic immunostaining of GFP (brown) with Nissl counterstaining (violet) of lines #184 and #233 (A–H) and GFP fluorescence images of line #233 (I–L) on parasagittal sections from 5W to 6W Scn1a-GFP brains. Boxed areas in (A, E, I, B, F, J, C, G, K, D, H, L) are magnified in (B–D, F–H, J–L, B1, F1, J1, C1–3, G1–3, K1–3, D1, H1, L1). The two lines (#184 and #233) showed a similar distribution pattern of GFP-expressing cells across all brain regions (A–H), but the signals in the line #233 are more intense than the line #184. In neocortex (B, F, J), GFP-expressing cells were scattered throughout the entire region. In the hippocampus (C, G, K), GFP-positive inhibitory neurons were sparsely distributed (see also Figure 8), while excitatory neurons in stratum pyramidale and stratum granulosum are GFP-negative. In cerebellum (D, H, L), Purkinje, basket, and deep cerebellar nuclei cells were GFP-positive. IHC, immunohistochemistry; CA, cornu ammonis; DG, dentate gyrus; o, stratum oriens; p, stratum pyramidale; r, stratum radiatum; lm, stratum lacunosum-moleculare; l, stratum lucidum; m, stratum moleculare; g, stratum granulosum; h, hilus; DCN, deep cerebellar nuclei; M, molecular layer; P, Purkinje cell layer; G, granular cell layer. Scale bars: 1 mm (A, E, I), 500 µm (C, D, G, H, K, L), 100 µm (B, F, J), and 50 µm (B1, C1–3, D1, F1, G1–3, H1, J1, K1–3, L1). Quantification of Nav1.1 signals in western blot analyses of brain lysates from the Scn1a-GFP mice and their wild-type littermates (N = 5 animals per each genotype) showed no difference between genotypes, while that of GFP somehow deviated among individual Scn1a-GFP mice (Figure 2—figure supplement 2). Fluorescence imaging of the Scn1a-GFP sagittal brain sections at postnatal day 15 (P15), 4-week-old (4W) and 8W showed that GFP signals continue to be intense in caudal region such as thalamus, midbrain, and brainstem (Figure 3), which is well consistent with our previous report of Nav1.1 protein and Scn1a mRNA distributions in wild-type mouse brain (Ogiwara et al., 2007). Figure 3 with 2 supplements see all Download asset Open asset Distribution of green fluorescent protein (GFP) signals in Scn1a-GFP mouse brain are largely maintained through development. Fluorescent images of parasagittal sections from P15 (A), 4W (B), and 8W (C) Scn1a-GFP mouse brains (line #233). GFP signals were observed in multiple brain regions. APT, anterior pretectal nucleus; CPu, caudate putamen; Cx, cerebral cortex; DCN, deep cerebellar nuclei; HP, hippocampus; IC, inferior colliculus; Mo, medulla oblongata; Ob, olfactory bulb; P, pons; RT, reticular thalamic nucleus; SC, superior colliculus; STN, subthalamic nucleus; VPM, ventral posteromedial thalamic nucleus; ZI, zona incerta. Scale bars: 1 mm. Nav1.1 is expressed in both excitatory and inhibitory neurons in neocortex but only in inhibitory neurons in hippocampus In the neocortex of Scn1a-GFP mouse, a large number of cells with GFP-positive somata (GFP-positive cells) were broadly distributed across all cortical layers (Figure 3 and Figure 3—figure supplement 1). Intensities of GFP signals in primary somatosensory cortex (S1) at L2/3 are much higher than other areas such as primary motor cortex (M1) (Figure 3—figure supplement 1), however the cell population (density) of GFP-positive cells did not differ in these areas indicating that GFP signals for GFP-positive cells are stronger in S1 at L2/3. Although GFP signals are strong in PV-INs (see Figure 8), cell density of PV-INs is not specifically high at S1 area and therefore most cells with strong GFP signals in S1 at L2/3 may not be PV-INs but excitatory neurons. In order to know the ratio of GFP-positive cells among all neurons, we further performed immunohistochemical staining using NeuN-antibody on Scn1a-GFP mouse at P15 and cells were counted at M1 and S1 (Figure 3—figure supplement 2 and Supplementary file 1a). The NeuN staining showed that GFP-positive cells occupy 30% (L2/3), 32% (L5), and 22% (L6) of NeuN- and GFP-positive cells at P15 (Figure 3—figure supplement 2B and Supplementary file 1a). However, we noticed that sparsely distributed cells with intense GFP signals, which are assumed to be PV-INs (see Figure 8), were often NeuN-negative (Figure 3—figure supplement 2 – arrowheads), reminiscent of a previous report that NeuN expression is absent in cerebellar inhibitory neurons such as Golgi, basket, and satellite cells in cerebellum (Weyer and Schilling, 2003). Therefore, NeuN-positive cells do not represent all neurons in neocortex as well. NeuN/GFP-double negative neurons could even exist and therefore above figure (Figure 3—figure supplement 2B) may deviate from the real ratios of GFP-positive cells among all neurons. Next, we performed triple immunostaining of Nav1.1, GFP, and ankyrinG on brains of Scn1a-GFP mouse at P15. In the neocortex (Figure 4), axon initial segments (AISs) of cells with Nav1.1-positive somata were always Nav1.1-positive but somata of cells with Nav1.1-positive AISs were occasionally Nav1.1-negative (Figure 4A–C). Cell counting revealed that 17% (L2/3), 21% (L5), and 8% (L6) of neurons (cells with ankyrinG-positive AISs) were GFP-positive (Figure 4D, left panel and Supplementary file 1b). Of note, all cells with Nav1.1-positive AISs or somata were GFP-positive, but AISs or somata for only half of GFP-positive cells were Nav1.1-positive (Figure 4D and Supplementary file 1c, d), possibly due to undetectably low levels of Nav1.1 immunosignals in a subpopulation of GFP-positive cells. The above ratios of GFP-positive cells among neurons (cells with ankyrinG-positive AISs) obtained in the triple immunostaining of Nav1.1, GFP, and ankyrinG are rather discordant to those obtained in the later experiment of triple immunostaining of Nav1.2, GFP, and ankyrinG, 23% (L2/3), 30% (L5), and 21% (L6) (see below). Therefore, we additionally performed double immunostaining of GFP and ankyrinG on brains of Scn1a-GFP mouse at P15, and the ratios of GFP-positive cells among neurons were 30% (L2/3), 26% (L5), and 9% (L6) (Figure 4—figure supplement 1 and Supplementary file 1e). Averaged ratios of GFP-positive cells among neurons of these experiments are 23% (L2/3), 26% (L5), and 13% (L6) (Figure 4—figure supplement 2 and Supplementary file 1f), which are actually significantly lower than those obtained in the NeuN staining (Figure 3—figure supplement 2 and Supplementary file 1a). Figure 4 with 2 supplements see all Download asset Open asset Nav1.1 expression at the axon initial segment (AIS) in the Scn1a-GFP mouse neocortex. (A) Triple immunofluorescent staining of parasagittal sections from P15 Scn1a-GFP mouse brain (line #233) by mouse anti-GFP (green), rabbit anti-Nav1.1 (magenta), and goat anti-ankyrinG (cyan) antibodies. Regions at primary motor cortex are shown. (B, C) Magnified images outlined in (A) are shown in (B) and (C). Arrows indicate AISs of cells with green fluorescent protein (GFP)-positive somata in which both somata and AISs are positive for Nav1.1. Arrowheads indicate AISs of cells with GFP-positive somata in which AISs but not somata are positive for Nav1.1. All images are oriented from pial surface (top) to callosal (bottom). Scale bars: 100 μm (A), 50 μm (B, C). (D) Cell counting of three Scn1a-GFP mice. Bar graphs indicating the percentage of cells with GFP- and Nav1.1-positive/negative somata and AISs per cells with ankyrinG-positive AISs (left panel), the percentage of cells with GFP-positive/negative somata per cells with ankyrinG-positive AISs and Nav1.1-positive somata and/or AISs (middle panel), and the percentage of cells with Nav1.1-positive/negative somata and/or AISs per cells with ankyrinG-positive AISs and GFP-positive somata (right panel) in L2/3, L5, and L6 (see also Supplementary file 1b–d). Only cells with ankyrinG-positive AISs were counted. Nav1.1 immunosignals were occasionally observed in somata, but in such cases Nav1.1 signals were always observed in their AISs if visible by ankyrinG staining. Note that 99% (L2/3), 99% (L5), and 97% (L6) of cells with Nav1.1-positive AISs have GFP-positive somata (middle panel), but only half or less of cells with GFP-positive somata have Nav1.1-positive AISs (right panel). L2/3, L5: neocortical layer II/III and V. AnkG, ankyrinG; +, positive; −, negative. Figure 4—source data 1 Numerical source data for Figure 4D. https://cdn.elifesciences.org/articles/87495/elife-87495-fig4-data1-v1.xlsx Download elife-87495-fig4-data1-v1.xlsx In contrast to the neocortex where only half of GFP-positive cells were Nav1.1-positive, in the hippocampus all GFP-positive cells were Nav1.1-positive and all Nav1.1-positive cells were GFP-positive (Figure 5). Actually, most of excitatory neurons such as CA1–3 pyramidal cells and dentate granule cells were GFP-negative. As described above (Figure 2C, G), fibrous GFP and Nav1.1 signals twining around CA1~3 pyramidal cells' somata which are assumed to be axon terminals of PV-INs were again observed (Figure 5A, B, D). Cell counting in the hippocampal CA1 region showed that 98% of cells with GFP-positive somata were Nav1.1-positive at their AISs and 100% of cells with Nav1.1-positive AISs were GFP-positive (Figure 5F and Supplementary file 1g, h). Figure 5 Download asset Open asset Nav1.1 expression at the axon initial segment (AIS) in the Scn1a-GFP mouse hippocampus. (A–D) Triple immunofluorescent staining of parasagittal sections from P15 Scn1a-GFP mouse brain (line #233) by mouse anti-GFP (green), rabbit anti-Nav1.1 (magenta), and goat anti-ankyrinG (cyan) antibodies. Regions at hippocampus were shown. Note that green fluorescent protein (GFP) and Nav1.1 immunosignals mostly overlap at somata. CA1, cornu ammonis 1; CA2, cornu ammonis 2; CA3, cornu ammonis 3; DG, dentate gyrus. Images are oriented from pial surface (top) to callosal (bottom). Scale bars: 100 μm. (E) Magnified images for co-expression of GFP and Nav1.1 in cells at CA1 region. Arrowheads indicate Nav1.1-positive AISs of GFP expression cells. Scale bar: 50 μm. (F) Bar graphs indicate the percentage of cells in hippocampal CA1 region with Nav1.1-positive/negative AISs per cells with GFP-positive somata and ankyrinG-positive AISs (left panel), and the percentage of cells with GFP-positive/negative somata per cells with Nav1.1/ankyrinG-double positive AISs (right panel) (see also Supplementary file 1g, h). Only cells with ankyrinG-positive AISs were counted. GFP/Nav1.1-double negative cells, most of which are pyramidal cells, were not counted because of the accumulated nature of their ankyrinG-positive AISs. AnkG, ankyrinG; +, positive; −, negative. Figure 5—source data 1 Numerical source data for Figure 5F. https://cdn.elifesciences.org/articles/87495/elife-87495-fig5-data1-v1.xlsx Download elife-87495-fig5-data1-v1.xlsx Double in situ hybridization of Scn1a and GFP mRNAs showed that these signals well overlap in both neocortex and hippocampus of Scn1a-GFP mice (Figure 6), further supporting that the GFP signals well represent endogenous Scn1a/Nav1.1 expression. Again, in neocortex Scn1a and GFP mRNAs seem to be expressed in a number of neurons including some of excitatory pyramidal cells, while in hippocampus they are absent in excitatory neurons such as CA1–3 pyramidal cells and dentate granule cells. All of these distributions of Scn1a and GFP mRNAs in Scn1a-GFP transgenic mouse brain are consistent to our previous report of regional distributions of Scn1a mRNA in wild-type mouse (Ogiwara et al., 2007). Figure 6 Download asset Open asset Green fluorescent protein (GFP) and Scn1a mRNAs expression mainly overlap in Scn1a-GFP mouse brain. Double in situ hybridization for Scn1a-GFP transgene mRNA and endogenous Scn1a mRNA on parasagittal sections from 4W Scn1a-GFP brains (line #233). (A) Sections were hybridized with antisense (left) and sense (right) RNA probes for GFP transgene (brown) and endogenous Scn1a (blue) mRNA species and chromogenically stained. Magnified images outlined in (A) are shown in (B–D) for antisense probes, and (E–G) for sense probes. o, stratum oriens; p, stratum pyramidale; r, stratum radiatum; lm, stratum lacunosum-moleculare; m, stratum moleculare; g, stratum granulosum, CA1, cornu ammonis 1; DG, dentate gyrus. Scale bars: 500 µm (A), 50 µm (B–G). To investigate the ratio of inhibitory neurons in GFP-positive cells, we generated and examined Scn1a-GFP and vesicular GABA transporter Slc32a1 (Vgat)-Cre (Ogiwara et al., 2013) double transgenic mice in which Slc32a1-Cre is expressed in all GABAergic inhibitory neurons and visualized by floxed tdTomato transgene (Figure 7). In the neocortex at 4W, 23% (L2/3), 28% (L5), and 27% (L6) of GFP-positive cells were Tomato-positive inhibitory neurons and 73% (L2/3), 77% (L5), and 83% (L6) of Tomato-positive cells were GFP-positive (Figure 7C and Supplementary file 1i, j). These results suggest that a significant subpopulation of neocortical excitatory neurons also express Nav1.1. Our previous observation that Nav1.1 is expressed in callosal axons of neocortical excitatory neurons (Ogiwara et al., 2013) supports that a subpopulation of L2/3 CC neurons express Nav1.1. Unlike in neocortex, in the hippocampus most of GFP-positive cells were Tomato-positive, 98% (CA1) and 94% (DG), and majorities of Tomato-positive GABAergic neurons are GFP-positive, 93% (CA1) and 77% (DG). These results further confirmed that in hippocampus Nav1.1 is expressed in inhibitory neurons but not in excitatory neurons. Although somata of pyramidal cells in CA2/3 region are weakly GFP-positive in this and some other experiments (Figure 7B and Figure 2—figure supplement 1G), those were GFP-negative in other experiments (Figures 2K and 5A, D) and therefore the Nav1.1 expression in CA2/3 pyramidal cells would be minimal if any. Figure 7 Download asset Open asset One-third of green fluorescent protein (GFP)-positive cells in neocortex are inhibitory neurons, but most of GFP-positive cells in hippocampus are inhibitory neurons. (A, B) GFP (green) and Tomato (magenta) fluorescent images of parasagittal sections from 4W Scn1a-GFP/Slc32a1-cre/Ai14 mouse. Regions at primary motor cortex (A) and hippocampus (B) are shown. Scale bar: 100 μm. (C) Bar graphs indicate the percentage of cells with Tomato-positive/negative somata per cells with GFP-positive somata (left panel) (see also Supplementary file 1i) and the percentage of cells with GFP-positive/negative somata per cells with Tomato-positive somata (right panel) (see also Supplementary file 1j) in L2/3, L5, L6, CA1, and DG. Cells in primary motor cortex and hippocampus of Scn1a-GFP mouse at 4W were counted. L2/3, L5, L6, CA1, and DG: neocortical layer II/III, V, VI, cornu ammonis 1, dentate gyrus. +, positive; −, negative. Figure 7—source data 1 Numerical source data for Figure 7C. https://cdn.elifesciences.org/articles/87495/elife-87495-fig7-data1-v1.xlsx Download elife-87495-fig7-data1-v1.xlsx We further performed immunohistochemical staining of PV and SST in neocortex and hippocampus of Scn1a-GFP mice at 4W (Figure 8). PV and SST do not co-express in cells and do not overlap. PV-INs and SST-INs were both GFP-positive, and especially GFP signals in PV-INs were intense (Figure 8A). Cell counting revealed that 21% (L2/3), 37% (L5), 37% (L6), 58% (CA1), 42% (CA2/3), and 41% (DG) of GFP-positive cells were PV- or SST-positive depending on regions in neocortex and hippocampus (Figure 8B and Supplementary file 1k–m). All PV-INs were GFP-positive (Figure 8B, middle), and most of SST-INs were GFP-positive (Figure 8B, right). Comparison of these results with those of Slc32a1-Cre mouse (Figure 7) suggests that GFP-positive GABAergic neurons in neocortex are mostly PV- or SST-positive, while in hippocampus a half of those are PV/SST-negative GABAergic neurons. Higher ratios of PV- or SST-positive cells (Figure 8B) compared with those of Slc32a1-Cre-positive cells (Figure 7C) among GFP-positive cells would be explained by us counting a cell as PV-positive if their PV immunosignals are moderate and a significant subpopulation of such cells are known to be excitatory neurons (Jinno and Kosaka, 2004; Tanahira et al., 2009; Matho et al., 2021). Quantitative analysis of GFP signal intensity and area size of cells revealed that GFP signal intensities in PV-positive cells were significantly higher than those in PV-negative cells and GFP signal intensities in SST-positive cells were lower than those in PV-positive cells but similar to PV/SST-double negative cells (Figure 9 and Supplementary file 1n–p). These results indicate that Nav1.1 expression level in PV-INs is significantly higher than those in excitatory neurons and PV-negative GABAergic neurons including SST-INs. Figure 8 Download asset Open asset Parvalbumin- or somatostatin-positive inhibitory neurons are green fluorescent protein (GFP)-positive in Scn1a-GFP mouse neocortex and hippocampus. (A) Triple immunofluorescent staining of parasagittal sections from 4W Scn1a-GFP mouse (line #233) by mouse anti-GFP (green), rabbit anti-parvalbumin (PV) (magenta), and goat anti-somatostatin (SST) (cyan) antibodies. Regions at neocortex and hippocampus are shown. Merged images were shown in the right columns. Arrows indicate SST/GFP-double positive cells. Arrowheads indicate PV/GFP-double positive. o, stratum oriens; p, stratum pyramidale; r, stratum radiatum; h, hilus; g, stratum granulosum; m, stratum moleculare. All images are oriented from pial surface (top) to callosal (bottom). Scale bars: 50 μm. (B) Bar graphs indicate the percentage of cells with PV- and SST-positive/negative somata per cells with GFP-positive somata (left panel) (see also Supplementary file 1k), the percentage of cells with GFP-positive/negative somata per cells with PV-positive somata (middle panel) (see also Supplementary file 1l), and the percentage of cells with GFP-positive/negative somata per cells with SST-positive somata (right panel) (see also Supplementary file 1m) in L2/3, L5, L6, CA1, CA2/3, and DG. Cells in neocortex and hippocampus of Scn1a-GFP mouse at 4W were counted. L2/3, L5, L6, CA1, CA2/3, and DG: neocortical layer II/III, V, VI, cornu ammonis 1, 2 plus 3, dentate gyrus. +, positive; −, negative. Figure 8—source data 1 Numerical source data for Figure 8B. https://cdn.elifesciences.org/articles/87495/elife-87495-fig8-data1-v1.xlsx Download elife-87495-fig8-data1-v1.xlsx Figure 9 Download asset Open asset Green fluorescent protein (GFP) signals in parvalbumin-positive inhibitory neurons are higher than PV-negative/GFP-positive cells in Scn1a-GFP mouse neocortex. (A) Scatter plots of intensities and area sizes of GFP immunosignals in GFP-positive cells with PV- or SST-positive or -negative somata. Cells at primary motor cortex (upper panels) and hippocampus (lower panels) in parasagittal sections from 4W Scn1a-GFP mouse (line #233) were analyzed. PV-positive (magenta circles) or SST-positive (black circles) and -negative (green circles) cells in neocortical L2/3, L5, and L6 or hippocampal CA1, CA2/3, and DG are plotted (see also Supplementary file 1n). (B) Box plots represent values for the intensity and area size in