Conditional knockout of Tsc1 in RORγt-expressing cells induces brain damage and early death in mice

Tuberous sclerosis complex 1 (Tsc1) is known to regulate the development and function of various cell types, and RORγt is a critical transcription factor in the immune system. However, whether Tsc1 participates in regulating RORγt-expressing cells remains unknown. We generated a mouse model in which Tsc1 was conditionally deleted from RORγt-expressing cells (Tsc1RORγt) to study the role of RORγt-expressing cells with Tsc1 deficiency in brain homeostasis. Type 3 innate lymphoid cells (ILC3s) in Tsc1RORγt mice displayed normal development and function, and the mice showed normal Th17 cell differentiation. However, Tsc1RORγt mice exhibited spontaneous tonic-clonic seizures and died between 4 and 6 weeks after birth. At the age of 4 weeks, mice in which Tsc1 was specifically knocked out in RORγt-expressing cells had cortical neuron defects and hippocampal structural abnormalities. Notably, over-activation of neurons and astrogliosis were observed in the cortex and hippocampus of Tsc1RORγt mice. Moreover, expression of the γ-amino butyric acid (GABA) receptor in the brains of Tsc1RORγt mice was decreased, and GABA supplementation prolonged the lifespan of the mice to some extent. Further experiments revealed the presence of a group of rare RORγt-expressing cells with high metabolic activity in the mouse brain. Our study verifies the critical role of previously unnoticed RORγt-expressing cells in the brain and demonstrates that the Tsc1 signaling pathway in RORγt-expressing cells is important for maintaining brain homeostasis.


Introduction
Tuberous sclerosis complex 1 (Tsc1), which forms a heterodimeric complex with tuberous sclerosis complex 2 (Tsc2), controls and mediates major processes, including protein and lipid synthesis, autophagy, and cell survival and proliferation [1]. Disruption of the Tsc1-Tsc2 complex results in various human diseases, such as hamartomas in multiple organs, cortical dysplasia, hypomelanotic macules, and cancers [2,3]. Previous studies found that targeted homozygous Tsc1 or Tsc2 mutations lead to death during the mid-embryonic period [4]. Mice with conditional Tsc1 knockout in neurons or astrocytes exhibit abnormal brain structure and function [5]. For example, mice with neuronal Tsc1 loss develop spontaneous seizures, macroencephaly, and hydrocephalus [6,7]. Astrocyte-specific Tsc1knockout mice also display epilepsy and some alterations in brain structure [8]. Tsc1 also plays a critical role in regulating the development and functions of immune cells, such as T cells and NK cells. Deletion of Tsc1 in CD4 + cells resulted in increased Th1 and Th17 cells differentiation [9], while conditional Tsc1 deletion in Tregs impaired their suppressive activity [10]. Hematopoietic-specific deletion of Tsc1 resulted in an activated and pro-apoptotic phenotype in NK cells [11]. However, the physiological role of Tsc1, especially in innate lymphoid cells (ILCs), remains largely unknown.
According to previous studies, RORγt is a critical transcription factor for type 3 innate lymphoid cells (ILC3s) and Th17 cells [12,13]. RORγt deficiency leads to the failure of lymphoid tissue inducer cells (LTi cells) and failure to generate ILC3 subsets [14,15]. Therefore, RORγt-Cre mice are now commonly used as an animal model to study the roles of specific genes in regulating ILC3s or Th17 cells. However, whether Tsc1 participates in regulating RORγt-expressing cells remains unknown.
In this study, we found that mice with conditional Tsc1 deletion from RORγt-expressing cells (Tsc1 RORγt mice) had numbers of ILC3s in the intestinal lamina propria layer and Th17 cells in the spleen comparable to those of their control littermates. However, Tsc1 RORγt mice died in a narrow timeframe between 4 and 6 weeks after birth. As previous studies showed that Tsc1 deletion in neurons or astrocytes led to brain damage and death [6][7][8]; here, we further reveal that Tsc1 RORγt mice show obvious brain damage, including cortical neuron defects, hippocampal structural abnormalities, neuron over-activation, and astrogliosis in the cortex and hippocampus. Mechanistically, the loss of Tsc1 in RORγt-expressing cells resulted in lower γ-amino butyric acid (GABA) receptor expression in the brain, and GABA supplementation prolonged the lifespan of Tsc1 RORγt mice to some extent.

Preparation of single-cell suspensions
Single-cell suspensions were prepared as previously described [18,19]. Briefly, mice were anesthetized with an isoflurane vaporizer (4-5% v/v). Spleen tissues were ground and passed through a 70-μm stainless steel mesh, and pellets were collected after centrifugation (450×g, room temperature, 10 min). Spleen mononuclear cells were separated from the pellets by lysing erythrocytes. The small intestines were opened longitudinally and washed with PBS (pH 7.4, Sigma-Aldrich, St. Louis, MO, USA) to remove the contents after Peyer's patches had been removed. Then, the intestines were cut into segments 4-5 cm in length and gently shaken in D-Hanks' solution (pH 7.4) containing 10 mM HEPES, 5 mM EDTA, 1 mM DTT, and 10% fetal bovine serum (FBS) for 20 min at 37°C. The remaining intestinal tissues were rinsed with Hanks' solution and then digested with 1 mg/ ml collagenase II (Gibco, Waltham, MA, USA) in Hanks' solution (pH 7.4) supplemented with 10% FBS for 40 min at 37°C with gentle shaking. The collected digests were filtered through a 100-μm mesh and enriched by centrifugation (450×g, room temperature, and 10 min) with a 25% Percoll solution (GE Healthcare, Pittsburgh, PA, USA) in RPMI 1640 medium (Gibco).

Antibodies and flow cytometry
All antibodies used for staining and subsequent flow cytometry analysis are listed in Supplemental Table 1. We confirmed the species reactivity for all antibodies according to the manufacturer's directions and performed preliminary experiments to determine the appropriate dilutions for all antibodies. Standard protocols were used for flow cytometry [19,20]. Briefly, single-cell suspensions were obtained from the spleen and intestinal lamina propria tissues of the mice. To analyze surface markers, 2 × 10 6 cells were incubated with staining buffer (PBS containing 2% mouse serum, 2% horse serum, and anti-CD16/CD32 blocking antibodies (BioLegend, San Diego, CA, USA)) and the indicated surface antibodies for 15 min at room temperature.
To detect IL-17A and IL-22, 2 × 10 6 cells were stimulated with IL-23 (BioLegend) or PMA/ionomycin (BD Biosciences, San Diego, CA, USA) for 4 h in the presence of BD Golgi Plug™ protein transport inhibitor (BD Biosciences) at 37°C. Then, the cells were fixed with a Fixation/Permeabilization Solution Kit (BD Biosciences) according to the manufacturer's instructions.
For RORγt staining, cells were stained with antibodies against surface markers, permeabilized with the Foxp3/ Transcription Factor Staining Buffer Set Kit (eBioscience, San Diego, CA, USA), and then stained with anti-RORγt antibody. Lineage (Lin) markers included CD3e, CD19, B220 and Gr-1. All of the isotype-matched control antibodies were purchased from BioLegend and BD and used at the same concentration as the test antibodies. All flow cytometry experiments were performed with a BD FACS Canto™ or BD FACS LSRFortessa™ instrument (BD Biosciences). Data were analyzed with FlowJo software (version 10.0, FlowJo LLC, Ashland, OR, USA). The lines indicate median values for each group.

Histological analysis
The mice were anaesthetized with an isoflurane vaporizer (4-5% v/v), and all tissues were collected as described in our previous reports [18,21]. Histological structures of the heart, lung, kidney, spleen, stomach, liver, and brain were analyzed by standard hematoxylineosin (HE) staining and assessed for the presence of pathomorphological changes by trained pathologists who were blinded to the experimental groups. The Nissl bodies in the brain were detected by Nissl staining. Briefly, the tissues were fixed in 10% neutral buffered formalin for at least 24 h, embedded in paraffin, and cut into 4μm-thick sections. The paraffin sections were stained with HE or Nissl staining solutions after they had been processed according to standard protocols [22].
All antibodies used for immunofluorescence staining are listed in Supplemental Table 1. The brain sections were incubated with a polyclonal rat anti-mouse glial fibrillary acidic protein (GFAP) antibody (eBioscience) and/or polyclonal rabbit anti-mouse neuronal nuclei antigen (NeuN) antibody (eBioscience) in 10% goat serum overnight at 4°C after deparaffinization, rehydration, and antigen retrieval. The sections were washed with 1× PBST (1× PBS, 0.05% Tween-20), incubated with secondary antibodies (Alexa Fluor® 488-conjugated goat anti-rat IgG (H+L) for GFAP and Alexa Fluor® 594conjugated goat anti-rabbit IgG (H+L) for NeuN) for 1 h at room temperature and then stained with DAPI. Coverslips containing samples from all organs, except for the brain, subjected to HE staining were scanned using an Olympus VS200 slide scanner (Olympus, Japan) and visualized with OlyVIA version 3.2 software (Olympus). Coverslips containing brain samples were scanned using Pannoramic DESK (3D HISTECH, Budapest, Hungary) and visualized with CaseViewer software (3D HIST ECH). Measurements of cortical thickness and cell scattering width in the CA1, CA3, and dentate gyrus (DG) regions of the hippocampus were performed using Case-Viewer software (3D HISTECH). Quantitative analyses of the area fraction of GFAP + astrocytes, NeuN + neurons, and Nissl bodies were performed using ImageJ software version 1.52v (National Institutes of Health, Bethesda, MD, USA) [23]. The investigators who acquired and analyzed the images were blinded to the experimental groups.

Electrocardiogram measurement
The mice were anaesthetized with an isoflurane vaporizer (4-5% v/v), and the electrocardiogram (ECG) was measured and recorded by electrocardiography (Mindray, Shenzhen, China) according to the operation manual.

Detection of mRNA levels by real-time RT-PCR
Briefly, mice were anesthetized with an isoflurane vaporizer (4-5% v/v), and whole-brain parenchyma and thymus tissues then were harvested. Total RNA was extracted and reverse transcribed into cDNA using standard protocols as previously described [24]. Total RNA was extracted from the brain and thymus tissues using TRIzol reagent (Invitrogen, Waltham, MA, USA), and cDNA was synthesized using a First Stand cDNA Synthesis Kit (DBI Bioscience, Ludwigshafen, Germany). Real-time PCR was performed using Bestar® SYBR Green qPCR Master Mix (DBI Bioscience, San Diego, CA, USA). The cycle threshold (Ct) values were normalized to the internal control (gapdh). The primer sequences for qRT-PCR were obtained from the Primer Bank, and the primer pairs used in the present study are shown in Supplemental Table 2.

Seizure test
Seizures were induced with pentylenetetrazole (PTZ) as previously described [25]. PTZ (50 mg/kg, Sigma-Aldrich) was administered intraperitoneally at 10-min intervals to post-natal days (PNDs) 28 mice. The injections were administered until the mice died (1-3 injections). Seizure latency was recorded as the time from PTZ injection to seizure occurrence. Observers were blinded to the genotypes of the mice.

Single-cell data analysis
The raw gene count expression matrix and metadata from embryonic day 14.5 mouse cerebral cortex scRNAseq datasets were downloaded from GEO (https://www. ncbi.nlm.nih.gov/geo/) (accession ID: GSE123335) [26] for analysis. Dimensional reduction, clustering, and analysis of scRNA-seq data were performed with the R package Seurat version 3 [27]. Cells with a RORC gene expression level ≥ 1 were considered RORγt-positive cells. The top 25% of the median of all expressed genes in RORγt-positive cells were considered highly expressed genes and used for KEGG enrichment analysis based on ClueGO to analyze the potential functions of RORγtpositive cells [28]. Gene set variation analysis (GSVA), a nonparametric and unsupervised software algorithm, was used to assess KEGG pathway activation with the R package GSVA [29].

Statistical analyses
All quantitative data were transferred to Excel, and statistical analyses were carried out with SPSS software for Windows (version 21, SPSS Inc., Chicago, Illinois, USA). Data are presented as the means ± SEMs. For comparisons between two independent experimental groups, an unpaired two-tailed Student's t test was used when data were normally distributed. When more than two independent groups were compared, one-way ANOVA followed by Tukey's test was performed. The Wilcoxon test was used to compare the difference in KEGG pathway scores estimated by GSVA between RORγt-positive and RORγt-negative cells with R (version 3.5.1) [30]. A p value less than 0.05 was considered to indicate statistical significance. Each analysis included n = 4-18 replicates per group, and the results are representative of at least two independent experiments. The sample size for each experiment is indicated in the corresponding figure legend. All graphs were produced using GraphPad Prism 5.0 for Windows (GraphPad Software Inc., La Jolla, CA, USA).

Results
Deletion of Tsc1 in RORγt-expressing cells resulted in the unexpected death of mice To investigate whether Tsc1 deletion would affect RORγtexpressing cells, we crossed Tsc1 floxed (Tsc1 fl/fl ) mice with RORγt-Cre mice to obtain Tsc1 fl/fl -RORγt Cre+ mutant mice in which Tsc1 was deleted from only RORγt-expressing cells (referred to as Tsc1 RORγt mice), while Tsc1 fl/fl -RORγt Crelittermates were used as control mice. We initially aimed to explore the effects of Tsc1 deficiency on ILC3s and Th17 cells; however, all Tsc1 RORγt mice unexpectedly died between PNDs 30 and PNDs 40, with a median survival time of 33 days (Fig. 1a, b). However, mice with heterozygous Tsc1 knockout in RORγt-expressing cells (Tsc1 fl/--RORγt Cre ) had a lifespan similar to that of control littermates (Fig. 1a). The Tsc1 RORγt mice had a normal weight and did not display gross morphological abnormalities upon comparison with control littermates before their death (Fig. 1c, d).

ILC3s and Th17 cells in Tsc1-deficient RORγt-expressing cells showed normal phenotypes
First, we postulated that the early death of the mice had been induced by the knockout of Tsc1 in ILC3s or Th17 cells, as according to the description of mice homozygous for the RORC mutant allele (RORγt GFP/GFP ) from The Jackson Laboratory, RORγt-deficient mice were reported to die at the age of 9-32 weeks [16]. Because these Tsc1 RORγt mice died between PNDs 30 and PNDs 40, all analyses of the mice were conducted at PNDs 28 unless specified otherwise. Flow cytometry analysis showed that Tsc1 RORγt mice had numbers of ILC3s and ILC3 subsets comparable to those of their control littermates (Fig. 2a, b). Moreover, Tsc1 deficiency in RORγtexpressing cells did not affect the expression of IL-17A or IL-22 (Fig. 2c, d), the main cytokines expressed in ILC3s, in this cell type [31]. As RORγt is also expressed in Th17 cells, we determined the ratio of Th17 cells in these mice and found that ratios of CD4 + T cells, CD8 + T cells, and Th17 cells in the intestinal lamina propria layer (LPL) and spleen were comparable in Tsc1 RORγt mice and their control littermates (Fig. 2e-h).
To further investigate the cause of death, we assessed pathological changes in several critical organs, including the heart, lung, kidney, spleen, stomach, and liver, of Tsc1 RORγt mice by hematoxylin-eosin (HE) staining. Tsc1 RORγt mice did not display significant pathomorphological changes that would have led to their death (Fig. 3af). Meanwhile, ECG measurements also revealed that Tsc1 RORγt mice had normal cardiac function (Fig. 3g). These results strongly suggest that the death of the mice caused by Tsc1 deficiency in RORγt-expressing cells was independent of ILC3s and Th17 cells.

Tsc1 RORγt mice exhibit spontaneous tonic-clonic seizures with neuronal defects
Unexpectedly, the Tsc1 RORγt mice often suffered megascopic spontaneous seizures characterized by generalized tonic-clonic activity before PNDs 28 (Supplemental Video 1). The seizure activity became more severe with increasing age, ultimately resulting in wild jumping and death ( Fig. 4a and Supplemental Video 2). We next assessed pathomorphological changes in the brain to further investigate the cause of death in the Tsc1 RORγt mice. The brain size and cerebral cortical thickness of the Tsc1 RORγt mice were the same as those of control littermates (Fig. 4b, c). However, neurons in the cerebral cortex of Tsc1 RORγt mice appeared swollen and necrotic, unlike those of control littermates (Fig. 4d). The hippocampus of Tsc1 RORγt mice also exhibited defects in organization. HE staining showed that structure of the dentate gyrus (DG) in the hippocampus was disorganized, and the arrangement of neurons was also disordered in the Tsc1 RORγt mice (Fig. 4e). Immunofluorescence staining for NeuN showed that the CA1 and CA3 pyramidal cell layers and DG granule cell layer were organized in an irregular fashion and that cells in these regions were discrete in Tsc1 RORγt mice. Quantitative analysis revealed increased cell scattering widths in the CA1, CA3, and DG cell layers of Tsc1 RORγt mice (Fig.  4f). These results suggest that the loss of Tsc1 in RORγtexpressing cells resulted in neuronal perturbations in the cortex and structural abnormalities in the hippocampus.

Abnormal activation of neurons and astrogliosis in the cortex and hippocampus of Tsc1 RORγt mice
Seizures are usually characterized by the abnormal and excessive synchronous firing of neurons [32]. Nissl bodies are usually regarded as an indicator of neuronal activation, in the brain [33]; therefore, we determined the number of Nissl bodies in CA1 and CA3 pyramidal cell layers and the DG granule cell layer of both control and Tsc1 RORγt mice. Compared with control littermates, Tsc1 RORγt mice exhibited an increased number of Nissl bodies with a complex microstructure, as evidenced by Fig. 3 Tsc1 RORγt mice did not display significant pathomorphological changes in various tissues and exhibited normal cardiac function. a HE staining and representative images of the heart, lung, kidney, spleen, stomach, and liver of Tsc1 RORγt mice and control littermates. b Representative ECG measurements from Tsc1 RORγt mice and control littermates. n = 5 animals per group. All HE staining images were captured from scans of whole-tissue slices. Scale bars: 500 μm (d) and 1000 μm (a, b, c, e, f) deeper staining and an increased area fraction of Nissl bodies in the cortex and hippocampus, indicating that Tsc1 RORγt mice exhibited abnormal neuronal activation (Supplemental Fig. 1a and Fig. 5a-e). However, the deletion of Tsc1 in RORγt-expressing cells did not affect the number of total neurons in the CA1 or CA3 pyramidal cell layer or the DG granule cell layer, as indicated by quantitative analysis of the area fraction of NeuN + neurons (Supplemental Fig. 1b and Fig. 5f-h).
Astrogliosis, which refers to an abnormal increase in astrocytes, usually occurs when an insult to the brain is sustained and often presents in patients with tuberous sclerosis syndrome or epilepsy [34]; elevated GFAP expression is a marker of astrogliosis [35]. We observed weakly stained astrocytes in the cerebral cortex of control mice; moreover, the area fraction of GFAP + astrocytes was obviously increased in the cortex of Tsc1 RORγt mice compared to control littermates (Fig. 5f, h). Moreover, compared with that of control littermates, increased GFAP expression was also found throughout the hippocampus of Tsc1 RORγt mice, in which GFAP + astrocytes were widely distributed, especially in the DG region (Supplemental Fig. 1b and Fig. 5g, h).
Thus, deletion of Tsc1 from RORγt-expressing cells did not affect the ratios or phenotypes of ILC3s or Th17 cells but resulted in abnormal activation of neurons and astrogliosis, which might have been responsible for the seizures observed in the Tsc1 RORγt mice.
Defects in the GABA signaling pathway are at least partially responsible for seizures and death in Tsc1 RORγt mice To further investigate the underlying mechanisms of these abnormalities in the nervous system, we measured  test (e, f, g). The locations of the magnified box are identified in the original image. All images were captured from scans of whole-brain slices. Magnification: 10x (a, f) and × 40 (b, c, d, g). Scale bars : 50 μm (b, c, d, g) and 100 μm (a, f) the gene expression levels of GABA receptor subunit genes (gabrg1, gabra2, gabrb2, gabrb1, and gabrb3), neural cadherin-like cell adhesion genes (pcdhga2, pcdhga8, and pcdhga9), voltage-gated channel subunits (kcnh7, kcna3, and scn8a), and other genes related to neuronal function (ube3a, neto2, kif5b, and erbb4) [36,37] in the whole brains of both Tsc1 RORγt and control littermate mice by quantitative real-time RT-PCR. The GABA signaling pathway is the major inhibitory neural pathway and mediates slow and prolonged inhibitory activity [36], and defects in the GABA signaling pathway increase seizure susceptibility [38]. Pcdhg genes regulate neuronal survival, cortical interneuron programmed cell death, and neural circuit assembly [39,40]. Sodium and potassium voltage-gated channel genes are also associated with neurological disorders [41][42][43]. Additionally, other genes have been reported to regulate neuronal function [44][45][46][47]. Notably, the expression levels of gabrg1, gabra2, and gabrb2 were significantly downregulated in the brains of Tsc1 RORγt mice compared with control littermates. Additionally, pcdhga2 and pcdhga8 and kcna3 levels were decreased in the brains of Tsc1 RORγt mice (Fig. 6a).
To determine whether defects in the GABA signaling pathway resulted in seizures and death in the Tsc1 RORγt mice, the mice were injected with increasing doses of PTZ (a GABA receptor antagonist, 50 mg/kg, i.p.) [25], and the subsequent latency of generalized tonic-clonic seizures was determined. Compared with control littermates, Tsc1 RORγt mice showed a significant reduction in seizure latency (162.8 ± 41.5 s vs. 82.5 ± 35.0 s, respectively) after PTZ challenge (Fig. 6b). In addition, Tsc1 RORγt mice exhibited a 50% mortality rate after the first PTZ injection and 100% mortality rate after a second PTZ injection, while control littermates began to die with a 30% mortality rate after the second PTZ injection, and all died after the third PTZ injection (Fig. 6c).
To confirm the role of deficiency in the GABA signaling pathway in the death of Tsc1 RORγt mice, the mice were fed GABA through the drinking water starting at the age of 3 weeks after weaning. Although GABA supplementation did not prevent the death of the Tsc1 RORγt Fig. 6 Defects in the GABA signalling pathway are at least partially responsible for seizures and death in Tsc1 RORγt mice. a Relative mRNA expression of GABA receptor subunit genes, neural cadherin-like cell adhesion genes, voltage-gated channel genes, and other genes related to neuronal function in the whole brains of Tsc1 RORγt mice and control littermates. b, c Seizure latency (b) and survival rate (c) of Tsc1 RORγt mice and control littermates after PTZ administration. d, e Survival rate (d) and longevity (e) of Tsc1 RORγt mice, Tsc1 RORγt mice supplemented with GABA and control littermates. Each dot represents one mouse, and error bars represent SEMs; *p < 0.05, **p < 0.01, and ***p < 0.001. n = 4-7 (c) and n = 7-14 (d, e). Unpaired two-tailed Student's t test (a-c, e). One-way ANOVA (d) mice, their survival time was prolonged, with a median value of 44 days (Fig. 6d, e). These data reveal that Tsc1 deficiency in RORγt-expressing cells led to defects in the GABA signaling pathway that at least partially contributed to seizures and death in Tsc1 RORγt mice.
Presence of a group of rare RORγt-positive cells with a high metabolic level in the mouse brain The deletion of Tsc1 in RORγt-expressing cells did not affect ILC3s or Th17 cells; therefore, we speculated that a group of RORγt-expressing cells may be present in the brain and that the deletion of Tsc1 in these cells would induce brain dysfunction, resulting in seizures, and eventual death. Therefore, we first analyzed RORγt expression in the whole brain by using data from the Allen Brain Atlas (http://mouse.brain-map.org/). Indeed, in situ hybridization (ISH) analysis indicated that RORγt was expressed in the cortex of the cerebrum and cerebellum during the embryonic period (E18.5). After birth, its expression increased and peaked at PNDs 14 and then decreased at PNDs 28 (Supplemental Fig. 2). Moreover, our RT-PCR and Western blotting data showed a very low level of RORγt expression in the brain that is much lower than its expression in the thymus (Fig. 7a,  b). Therefore, we confirmed that, although rare, RORγtpositive cells are present in the brain.
By re-analyzing published single-cell RNA-seq data from the mouse brain [26], we identified 81 RORγtpositive cells among 11,069 cells (0.73% of total cells) (Fig. 7c). Functional enrichment analysis revealed that these RORγt-positive cells exhibited high expression levels of ribosomal and spliceosome-associated genes. Many of the highly expressed genes in RORγt-positive cells are also involved in both oxidative phosphorylation and neurological disorders, including Parkinson's disease, Huntington's disease, and Alzheimer's disease (Fig.  7d). Meanwhile, KEGG enrichment analysis indicated that 139 signaling pathways were significantly differently enriched between RORγt-positive cells and RORγtnegative cells (Supplemental Table 3). Among these pathways, many pathways related to metabolism, including glycosphingolipid biosynthesis and amino acid and fatty acid metabolism, were enriched in RORγt-positive cells, indicating that the metabolic level in RORγtpositive cells is enhanced compared to that of RORγtnegative cells in the brain (Fig. 7e). These data suggest the presence of a group of rare RORγt-positive cells with a high metabolic level in the mouse brain.

Discussion
Our study shows that Tsc1 expression in RORγtexpressing cells is dispensable for ILC3 development and function but critical for brain homeostasis. Loss of Tsc1 in RORγt-expressing cells in the mouse brain induced neuronal defects, astrogliosis in the cortex and hippocampus, spontaneous tonic-clonic seizures, and death. Notably, GABA supplementation delayed death and prolonged lifespan to some extent.
According to previous studies, Th17 cells can infiltrate the brains of patients of Parkinson's disease and induce neuronal cell death [48], and Tsc1 deficiency affects T cell development and function [9,10]. However, our data showed that knockout of Tsc1 in RORγt-expressing cells did not affect the development of Th17 cells; however, Tsc1 knockout caused severe brain damage followed by death at the age of 4 to 6 weeks. Previous studies reported that neuron-specific inactivation of Tsc1 resulted in the increased generation of neural progeny and death with a median survival time of 18 days [7], while the specific inactivation of Tsc1 in GFAP-positive astrocytes led to astrogliosis and death between 11 and 22 weeks after birth [8]. These mice with astrocyte-and neuronspecific inactivation of Tsc1 had an enlarged brain and edema. However, the brains of our model Tsc1 RORγt mice showed a normal appearance. The loss of Tsc1 in RORγt-expressing cells not only caused neuronal defects but also induced astrogliosis, indicating that epilepsy and death caused by the inactivation of Tsc1 in RORγtexpressing cells might be due to both the abnormal discharge of neurons and activated astrocytes. Because RORγt is expressed in the brain during the embryonic period (E18.5), the Tsc1 gene was knocked out in RORγt-expressing cells in utero. Based on the findings, abnormal brain structures might appear during embryonic development, and a vicious circle between a disordered brain structure and seizures further aggravates brain damage with increasing age.
GABA, the major neurotransmitter in the central nervous system, is an important regulator of neuronal inhibition as its binds GABA A receptors, which are ligandgated anion channels. Dysfunction or mutation of GABA A receptors has been identified in patients with various types of epilepsy [49]. Tsc1 RORγt mice exhibited a decreased latency to PTZ-induced seizures and were more susceptible to PTZ than control littermates, potentially due to decreased expression of GABA A receptor subunits in the brain. Moreover, GABA supplementation prolonged the lifespan of Tsc1 RORγt mice, which further suggests that the impairment of GABA-GABA A receptor interactions exacerbated seizures. Meanwhile, the loss of Tsc1 in RORγt-expressing cells also decreased the expression levels of neuronal function-related genes (pcdhga2 and pcdhga8) and ion channel gene expression (kcna3). Pcdhg genes were reported to regulate neuronal survival, synaptic maintenance, and neural circuit assembly [40,50], which might explain why GABA supplementation alone did not improve the survival rate of Tsc1 RORγt mice. The Kcna3 gene encodes the voltage-gated potassium channel Kv1.3, which protects against neuro-inflammation [43,[51][52][53]. Thus, we proposed that the decreased expression of kcna3 was a compensatory mechanism when neuronal defects and astrogliosis occurred in the brains of Tsc1 RORγt mice.
According to the Allen Brain Atlas, RORγt is expressed in the brain during the embryonic period, and its expression increases after birth. In our study, the expression of RORγt in the brain parenchyma was confirmed by RT-PCR and Western blot analysis and then verified by analysis of published single-cell RNA-seq data. Interestingly, these rare RORγt-positive cells were found to be critical for brain homeostasis. Compared with RORγt-negative cells, RORγt-positive cells exhibited an increased metabolic level, as evidenced by higher levels of ribosomal gene expression and oxidative phosphorylation. Oxidative phosphorylation is an essential process for cell function, and ribosomes are targeted for autophagy-mediated degradation to provide supplemental amino acids and nucleotides under conditions of nutrient starvation [54].
A recent study identified a cluster of RORγt-positive γδ T cells in the meninges that regulates anxiety-like behavior via IL-17a signaling in neurons [55]. These meningeal γδ T cells are virtually absent from the prenatal dural meninges but undergo progressive seeding after birth. However, we found that RORγt is expressed in the brain parenchyma during the embryonic period. In other words, the RORγt-positive cells in the brain parenchyma identified in the present study might not originate from the meninges. However, we were unable to exclude the possibility that Tsc1 deletion in these meningeal γδ17 T Fig. 7 A group of rare RORγt-positive cells with a high metabolic level exists in the mouse brain. a, b Relative expression of the RORγt mRNA in the whole brain and thymus (a) and RORγt protein expression in the whole brains (b) of mice at PNDs 14. c Proportion of RORγt-positive cells among total brain cells. d Functional enrichment analysis of the top 25% most highly expressed genes in RORγt-positive cells. e Functional gene sets identified by KEGG pathway enrichment analysis indicated increased metabolic pathway enrichment in RORγt-positive cells compared with RORγt-negative cells. Each dot represents one mouse, and error bars represent SEMs; ***p < 0.001. Unpaired two-tailed Student's t test (a). Wilcoxon test (e) cells would also contribute to brain damage and death, as meningeal γδ17 T cells also express RORγt at high levels. A limitation of the study is that we did not determine the location, type or source of these RORγtpositive cells.

Conclusions
Taken together, our results verify the critical role of previously unnoticed RORγt-expressing cells in the brain and indicate that the Tsc1 signaling pathway in RORγtexpressing cells is important for maintaining neuron and astrocyte homeostasis.