Skip to main content
NCBI home page
Neurotherapeutics logoLink to Neurotherapeutics
. 2024 Sep 7;21(6):e00447. doi: 10.1016/j.neurot.2024.e00447

Mitochondria-related HSDL2 is a potential biomarker in temporal lobe epilepsy by modulating astrocytic lipid metabolism

Xiaxin Yang a,1, Jianhang Zhang b,1, Zhihao Wang c, Zhong Yao b, Xue Yang c, Xingbang Wang d, Xiuhe Zhao c,e,⁎, Shuo Xu b,⁎
PMCID: PMC11585875  PMID: 39245623

Abstract

Temporal lobe epilepsy (TLE) is the most prevalent type of focal epilepsy in adults. While comprehensive bioinformatics analyses have facilitated the identification of novel biomarkers in animal models, similar efforts are limited for TLE patients. In the current study, a comprehensive analysis using human transcriptomics datasets GSE205661, GSE190451, and GSE186334 was conducted to reveal differentially expressed genes related to mitochondria (Mito-DEGs). Protein-protein interaction (PPI) network and Least Absolute Shrinkage and Selection Operator (LASSO) regression analyses were performed to identify hub genes. Additional GSE127871 and GSE255223 were utilized to establish the association with hippocampal sclerosis (HS) and seizure frequency, respectively. Single-cell RNA analysis, functional investigation, and clinical verification were conducted. Herein, we reported that the Mito-DEGs in human TLE were significantly enriched in metabolic processes. Through PPI and LASSO analysis, HSDL2 was identified as the hub gene, of which diagnostic potential was further confirmed using independent datasets, animal models, and clinical validation. Subsequent single-cell and functional analyses revealed that HSDL2 expression was enriched and upregulated in response to excessive lipid accumulation in astrocytes. Additionally, the diagnostic efficiency of blood HSDL2 was verified in Qilu cohort. Together, our findings highlight the translational potential of HSDL2 as a biomarker and provide a novel therapeutic perspective for human TLE.

Keywords: Epilepsy, TLE, HSDL2, Astrocytes, Biomarker

Graphical abstract

Here, a comprehensive analysis using human TLE transcriptomics datasets was conducted and HSDL2 was identified as the hub gene, which was further confirmed by animal models and clinical validation. Subsequent single-cell and functional analyses revealed that HSDL2 expression was enriched and upregulated in response to excessive lipid accumulation in astrocytes. Knockdown of HSDL2 led to increased lipid accumulation in astrocytes, suggesting the protective role of HSDL2 in TLE. Additionally, the diagnostic efficiency of blood HSDL2 was verified in Qilu TLE cohort.

Image 1

Open in a new tab

Introduction

Epilepsy is a complex neurological disorder characterized by a persistent predisposition to generate synchronous epileptic seizures, affecting over 70 million people worldwide [1]. Temporal lobe epilepsy (TLE) is the most prevalent type of focal epilepsy in adults [2], which is typically associated with hippocampal sclerosis (HS) in histopathology and leads to neurobiological, cognitive, and psychiatric comorbidities [3]. While the clinical diagnosis of TLE is well-defined on clinical semeiology, neuroimaging, and electroencephalogram (EEG), the anatomo-electro-clinical assessment largely relies on the expertise and judgment of physicians [4]. Quantitative EEG interpretation has been proven useful in reducing subjective bias and increasing diagnostic sensitivity, its application scenarios are restricted [5]. Thus, the identification of objective and reliable biomarkers is crucial for the assessment and early intervention of TLE, which has made certain achievements [[6], [7], [8]]. Recent advancements in bioinformatics have facilitated the rapid and efficient analysis of high-throughput data, providing insights into epilepsy diagnosis and predicting drug resistance. For instance, Chen et al. identified multiple transcriptome signatures to predict seizure frequency and occurrence of HS in a rat model of TLE [9]. Wang et al. highlighted the association of four inflammation-related genes with TLE in a mouse model through immunomodulation [10]. However, similar integrated analyses of transcriptomic profiling are limited for human TLE.

Approximately one-quarter of patients with mitochondrial disease also experience epilepsy [11]. Mitochondria, a membrane-bound organelle responsible for energy generation, play critical roles in fatty acid and lipid metabolism, calcium homeostasis, and the production of reactive oxygen species (ROS) [12]. Correspondingly, mitochondrial dysfunction involves various pathological consequences, including lipid accumulation, mitochondrial DNA damage, and increased ROS levels [13,14]. Given the dynamic metabolism of brain, extensive research has focused on elucidating the relationship between mitochondrial dysfunction and epilepsy. Volmering et al. demonstrated that chronic inflammation induced mitochondrial dysfunction, which further promoted epileptogenesis and neuron loss in patients with mesial TLE and HS [15]. Chan et al. described that astrocytes initiated epileptic seizures through a GABAergic pathway in an animal model of mitochondrial epilepsy [11]. However, further progress is needed to understand how mitochondria-related genes participate in the cellular activity leading to the neuronal hyperactivity and epilepsy progression.

In this study, we performed a comprehensive analysis utilizing multiple transcriptomics datasets obtained from human TLE tissues, combined with validation through animal models and clinical data. Our findings revealed that mitochondrial-related gene HSDL2 not only serves as a reliable diagnostic biomarker for TLE but also as an indicator for assessing the severity of seizures. Moreover, we uncovered a promising therapeutic approach for human TLE through subsequent functional enrichment analysis and in vitro experiments.

Materials and Methods

Data collection and TLE cohorts

Systematic GEO inquiry was performed (Gene Expression Omnibus, http://www.ncbi.nlm.nih.gov/geo). Three independent human TLE datasets, including GSE205661 (6 TLEs and 9 controls), GSE190451 (3 TLEs and 3 controls), and GSE186334 (5 TLEs and 5 controls), were enrolled for this study. Specifically, GSE205661 was used as the training set, and GSE190451 and GSE186334 were used as the external validation sets. The datasets GSE127871 (12 TLEs with seizure frequency), GSE255223 (6 TLEs with HS status) and GSE143272 with blood samples were also enrolled.

The blood and brain specimens from TLE patients were collected from June 2023 to December 2023 ​at Qilu Hospital of Shandong University. Informed consent was obtained from each participating subject. The protocol of this study have been reviewed by the Medical Ethical Committee, Qilu Hospital of Shandong University, and have therefore been performed in accordance with the ethical standards laid down in an appropriate version of the Declaration of Helsinki.

Differential expressed gene analysis

The differentially expressed genes (DEGs) between TLEs and controls were acquired using the “limma” R package, with the screening criteria of |log 2 fold change (FC)| ​> ​0.8 and adjusted P-value <0.05. Mitochondria-related genes were obtained from the MitoCarta 3.0 database (http://www.broadinstitute.org/MitoCarta) to screen out the mitochondria-related DEGs (Mito-DEGs). Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were performed using the “clusterProfiler” R package (adjusted P-value <0.05) and visualized using the “ggplot 2” R package.

Hub genes identification

For protein-protein interaction (PPI) network analysis, Mito-DEGs were uploaded to the STRING database and analyzed with Cytoscape software 3.6.1 (minimum required interaction score ​= ​0.4) [16]. The top 10 genes were selected as PPI hub genes through CytoHubba [17]. For Least Absolute Shrinkage and Selection Operator (LASSO) regression analysis, the hub genes were identified using the “glmnet” R package [18].

Single-cell RNA sequencing and single-gene set enrichment analysis (GSEA)

Single-cell transcriptomics dataset GSE190452 of human TLE was enrolled and processed by the “seurat” R package. “UMAP” was employed to visualize the different cell clusters. GSEA was performed using the “clusterProfiler” R package. The gene sets “c2.cp.kegg.v2023.1” and “c5.go.v2023.1” were downloaded from the Molecular Signatures Database (https://www.gsea-msigdb.org/gsea/msigdb/index.jsp) for GSEA (adjusted P-value <0.05).

Animal models

Pilocarpine (Pilo)- and pentylenetetrazole (PTZ)-treated mouse models of TLE were utilized as previously described [19,20]. Seizure evaluation was conducted using Racine's standard criteria (Stage 1–5) [21]. Mice with the Racine score ≥4 were included. All the procedures were conducted according to the guidelines of the National Institutes of Health on the care and use of animals and approved by the Institutional Animal Care and Use Committee of Shandong University.

Western blotting

Western blotting was performed as previously described [22,23]. Primary antibodies were as follows: HSDL2 (15631-1-AP，Proteintech, Wuhan, China), A2aR (ER1903-41, HUABIO, Hangzhou, China), PPARα (66826-1-Ig, Proteintech), and β-actin (A5441, Sigma-Aldrich, Saint Louis, USA). Secondary antibodies used are as follows: IRDye®680RD goat anti-mouse IgG (926–68070, LI-COR, Lincoln, USA), IRDye®680RD goat anti-rabbit IgG (926–68071, LI-COR), IRDye®800CW goat anti-mouse IgG (926–32210, LI-COR) and IRDye®800CW goat anti-rabbit IgG (926–32211, LI-COR). Detection was performed with the LI-COR Odyssey−DLx imaging system and quantitated with ImageJ software (National Institutes of Health, Bethesda, USA).

Quantitative real-time PCR (qRT-PCR)

Total RNA was extracted using the TRIzol reagent (15596018, Thermo Fisher Scientific, Waltham, USA) and the mRNA levels of genes were quantified by SYBR Green-based gene expression analysis (QPK201, Toyobo, Tokyo, Japan) using the.

ABI QuantStudio 3 PCR system (Thermo Fisher Scientific) as previously described [24]. Primer sequences were: 5′-ATGTTACCCAACACCGGGAG-3’ (forward) and 5′-GCTTTCAATGCAATAGCTTTGCC-3’ (reverse) for homo HSDL2; 5′-AGCGGGATGCACAGTTTTTAT-3’ (forward) and 5′-GGTCTTCGCAGCAATGACAATA-3’ (reverse) for mouse HSDL2; 5′- CGGTGACTTATCCTGTGGTCC-3’ (forward) and 5′- CCGCAGATTCTACATTCGATGTT-3’ (reverse) for homo PPARα; 5′- CGCTCCGGTACAATGGCTT-3’ (forward) and 5′- TTGTTCCAACCTAGCATGGGA-3’ (reverse) for homo A2aR; 5′-GACAGGATGCAGAAGGAGATTACT-3’ (forward) and 5′-TGATCCACATCTGCTGGAAGGT-3’ (reverse) for β-actin.

Immunofluorescence staining

Immunofluorescence staining was conducted according to a previous study [22]. Briefly, brain tissue were incubated with primary antibody at 4 ​°C overnight. Primary antibodies used were as follows: HSDL2 (PA5-83636，Thermo Fisher Scientific), GFAP (GB12090, Servicebio, Wuhan, China) and Iba1 (GB15105-100, Servicebio). Second antibodies were Alexa fluor 488-conjugated affinipure goat anti-rabbit IgG (GB25303, Servicebio) and CY3-conjugated goat anti-mouse IgG (GB21301, Servicebio). The images were captured by the fluorescent microscope (Olympus, Tokyo, Japan) and analyzed with the ImageJ software.

Cell culture and BODIPY 493/503 staining

Normal human astrocytes (NHA) cells were cultured in the Dulbecco's Modified Eagle's Medium (High glucose-DMEM, CM15019, Macgene, Beijing, China) supplemented with 10% (vol/vol) fetal bovine serum (04-001-1ACS, Biological Industries, Shanghai, China) [25]. HSDL2 siRNA (siHSDL2-1: 5′-CCAGCAGACUCAAAUACAATT-3′ and siHSDL2-2: 5′-UUGUAUUUGAGUCUGCUGGTT-3′) or corresponding negative control (siCON) were transfected into cells with Lipofectamine 3000 (L3000015, Thermo Fisher Scientific), respectively. Cells were harvested at 48 ​h after transfection to detect HSDL2 expression by qRT-PCR and western blotting. Free fatty acids (FFAs, 1:1 mixture of oleic acid and palmitic acid, 100 ​μM) were applied to cells for 48 ​h to imitate the lipid accumulation. Lipid droplets (LDs) were stained with BODIPY 493/503 staining (C2053S, Beyotime, Shanghai, China).

Statistical analysis

Data were presented as Mean ​± ​SD. Student's t-test was applied for statistical comparisons between two groups and those among more than two groups were assessed with one-way ANOVA with Bonferroni's multiple comparison post hoc test. Statistical analyses were performed using Prism 9 (GraphPad, San Diego, USA), except for the bioinformatics analyses with R software. Differences in experiments were considered statistically significant at P ​< ​0.05.

Results

Mito-DEGs in TLE and functional enrichment analysis

The overall analysis workflow is depicted in Fig. 1. Following the aforementioned criteria, we identified 2349 DEGs, with 1727 upregulated-genes and 622 downregulated-genes (Fig. 2A). By intersecting these DEGs with 1136 mitochondria-related genes acquired from the MitoCarta3.0 database, 102 Mito-DEGs were then obtained, including 95 upregulated and 7 downregulated genes (Fig. 2B–C).

Fig. 1.

Fig. 1

Open in a new tab

Flowchart of the research. TLE, temporal lobe epilepsy; Mito-DEGs, mitochondria-related differentially expressed genes; GO, gene ontology; KEGG, kyoto encyclopedia of genes and genomes; PPI, protein-protein interaction; LASSO, least absolute shrinkage and selection operator; Pilo, pilocarpine; PTZ, pentylenetetrazole; GSEA, gene set enrichment analysis.

Fig. 2.

Fig. 2

Open in a new tab

Identification and functional enrichment analysis of Mito-DEGs. (A) Volcano map visualization of 2349 DEGs between TLE and control of GSE205661. (B) Venn diagram of the overlapping 102 Mito-DEGs. (C) Heat map visualization of 102 Mito-DEGs. (D–E) GO and KEGG enrichment analyses of Mito-DEGs. Mito-DEGs, mitochondria-related differentially expressed genes; DEGs, differentially expressed genes; HC, healthy control; TLE, temporal lobe epilepsy; GO, gene ontology; KEGG, kyoto encyclopedia of genes and genomes.

In terms of GO enrichment analysis, these Mito-DEGs were primarily enriched in the acetyl-CoA, thioester, acyl-CoA, and fatty acid metabolic processes (Fig. 2D). As shown in Fig. 2E, KEGG enrichment analysis indicated that the Mito-DEGs were involved in the histidine and tryptophan metabolism, as well as fatty acid degradation. These findings suggested that these Mito-DEGs might play important roles in TLE, particularly in the metabolic processes.

Identification of the hub gene HSDL2

The PPI network of Mito-DEGs was constructed using the STRING database (Fig. 3A), and 10 candidate genes (ACAA2, HSDL2, ACADSB, ACSS3, ACSS1, HADH, ALDH9A1, ACSF2, ACACB, and DLAT) were identified through the CytoHubba plug-in (Fig. 3B). Meanwhile, the Mito-DEGs were also analyzed via the LASSO regression algorithm, and 5 candidate genes (BCKDHB, HSDL2, DARS2, MMAA, and PRODH2) were determined with the optimal value of λ.1se ​= ​0.2154052 (Fig. 3C and D). By intersecting the candidate genes obtained from the different algorithms, we identified one common hub gene, HSDL2 (Fig. 3E).

Fig. 3.

Fig. 3

Open in a new tab

Identification and validation of the hub gene HSDL2. (A–B) PPI network constructed by CytoHubba (A), and 10 key genes were obtained (B). (C–D) LASSO regression analysis to acquire 5 key genes. (E) Venn diagram by intersecting the candidate genes obtained from PPI and LASSO regression algorithms. (F–H) HSDL2 expression in the training (GSE205661) and external validation datasets (GSE190451 and GSE186334). Mean ​± ​SD indicated by horizontal lines and bars were given. ∗P ​< ​0.05. PPI, protein-protein interaction; LASSO, least absolute shrinkage and selection operator; HC, healthy control; TLE, temporal lobe epilepsy.

To assess the reliability of HSDL2 as a key biomarker in human TLE, we verified its mRNA expression in the training set (GSE205661) as well as two independent validation sets (GSE190451 and GSE186334). As shown in Fig. 3F, G, and 3H, the consistently elevated expression of HSDL2 in TLE tissues was confirmed across all three datasets (P ​= ​0.0003, 0.0001, and 0.0432 for GSE205661, GSE190451, and GSE186334, respectively). Collectively, the hub gene HSDL2 was identified as a potential predictor of human TLE. The expression levels of the other 13 candidate genes identified by PPI network and LASSO regression algorithm in three datasets were also presented in Supplementary Fig. S1.

Clinical significance of HSDL2 in TLE

Given the diagnostic efficiency of HSDL2 for human TLE, we further utilized the dataset GSE127871 to explore the association between HSDL2 and seizure frequency. A positive linear correlation was observed between HSDL2 expression and seizure frequency (R ​= ​0.663, P ​= ​0.0187, Fig. 4A). Moreover, HS with dentate gyrus atrophy is a common pathological characteristic of TLE. We, therefore, examined the HSDL2 expression in the dentate gyrus from TLE patients with or without HS using the dataset GSE255223. Strikingly, TLE patients with HS exhibited higher HSDL2 expression compared to those without HS (P ​= ​0.0127, Fig. 4B).

Fig. 4.

Fig. 4

Open in a new tab

HSDL2 validation in TLE. (A) Scatter plot showing the correlation between HSDL2 and seizure frequency in patients with TLE in GSE127871. R = Spearman's rank correlation coefficient. The red line was the regression line fitted by linear modeling. (B) Bar plot showing the differential expression of HSDL2 in dental gyrus between the TLE patients with HS and those without HS (non-HS) in GSE255223. (C,E) qRT-PCR and western blotting showing elevated HSDL2 mRNA and protein expressions in Pilo-treated mouse for TLE. β-actin was used as the internal control. (D, F) qRT-PCR and western blotting showing elevated HSDL2 protein expression in resected brain specimens from TLE patients. β-actin was used as the internal control. (G) Resected specimens from TLE patients were stained with anti-HSDL2 antibody for immunohistochemistry. Representative images indicate the preferential expression of HSDL2 in human TLE specimens. Scale bar 20 ​μm. Mean ​± ​SD indicated by horizontal lines and bars were given and ∗P ​< ​0.05 for (B, C, D). HC, healthy control; TLE, temporal lobe epilepsy; HS, hippocampal sclerosis; qRT-PCR, quantitative real-time PCR; Pilo, pilocarpine.

To validate our findings, we constructed two mouse models of TLE treated with Pilo and PTZ, respectively. As illustrated in Fig. 4C and E, both mRNA and protein expressions of HSDL2 were significantly elevated in the temporal lobe speciemens of Pilo-treated mice. Similar results were obtained from the PTZ-treated mice (Supplementary Fig. S2). Moreover, HSDL2 upregulation was observed in the resected brain specimens from TLE patients (Fig. 4D and F). These findings were further supported by immunohistochemistry, as shown in Fig. 4G. Taken together, our results revealed that HSDL2 could serve as both a diagnostic biomarker for TLE itself and an indicator for assessing the severity of seizure.

Differential distribution of HSDL2 expression among cell subpopulations

We employed a single-cell nuclear transcriptomics dataset GSE190452 of TLE to further investigate the expression pattern of HSDL2 among different cell subpopulations. Seven major cell clusters were determined and annotated according to the representative marker gene expression, including astrocytes, excitatory neurons, inhibitory neurons, microglia, oligodendrocytes, oligodendrocyte progenitor cells (OPCs), and endothelial cells (Fig. 5A). In comparison to the healthy controls, there was a significant increase in the proportions of astrocytes and microglia in TLE (Fig. 5B–C). Notably, HSDL2 expression was abundant in astrocytes with elevated expression in TLE (Fig. 5D–E). This finding was verified in the temporal samples obtained from TLE patients by immunofluorescence staining. As illustrated in Fig. 5F, HSDL2 was predominantly co-expressed with astrocyte marker GFAP (glial fibrillary acidic protein). Preliminary GSEA analysis was also performed, suggesting that HSDL2 was closely associated with the metabolic and catabolic processes of fatty acid and lipid (Fig. 5G). Totally, our results suggested preferential expression of HSDL2 in astrocytes might play crucial roles in epileptogenesis in regulation of lipid catabolism.

Fig. 5.

Fig. 5

Open in a new tab

Single-cell RNA sequencing of HSDL2 in TLE. (A) Single-cell RNA sequencing of GSE190452 visualizing UMAP cell clusters including OPC, endothelial cell, excitatory neuron, inhibitory neuron, astrocyte, oligodendrocyte, and microglia. (B–C) The proportion of different cell types in TLE and control group from single-cell RNA sequencing of GSE190452. (D–E) UMAPs of control and TLE brain samples showing the HSDL2 expression in different cell types. (F) Immunofluorescence double staining of HSDL2 with GFAP was performed in the brain tissue of TLE patients, respectively. Scale bar 20 ​μm. (G) GSEA analysis. HC, healthy control; TLE, temporal lobe epilepsy; OPC, oligodendrocyte progenitor cell; GSEA, gene set enrichment analysis.

HSDL2 is volved in TLE by alleviating the lipid accumulation in astrocytes

Mechanically, lipids are initially generated in neurons under chronic epileptic conditions, and then transported into the astrocytes, leading to lipid droplets (LDs) accumulation and increased epileptogenesis [26]. To precisely define the HSDL2-enriched astrocytic subtypes in TLE, we performed subpopulation analysis of the astrocytes. Unsupervised subclustering further divided the astrocyte population into six subclusters (A1–A6; Fig. 6A and Supplementary Fig. S3A). Intriguingly, we observed that one unique (and the largest) astrocytic subcluster of A1 exhibited highest level of HSDL2 expression (Fig. 6B and Supplementary Fig. S3B), in which lipid metabolism-related pathways were also dominantly enriched indicated by GO pathway analysis (Fig. 6C). To mimic the excessive lipid accumulation within the astrocytes, we stimulated NHA cells with exogenous FFAs for 48 ​h and observed a significant HSDL2 upregulateion at both mRNA and protein levels, consistent with the subcluster analysis (Fig. 6D–E).

Fig. 6.

Fig. 6

Open in a new tab

HSDL2 alleviates the lipid accumulation in NHA. (A) Subclusters of astrocytes in GSE190452 visualizing by UMAP. (B) UMAP showing the HSDL2 expression in different subclusters of astrocytes. (C) GO enrichment analysis of lipid catabolic-related and metabolism-related pathways among different subclusters. (D–E) The mRNA and protein levels of HSDL2 were detected in NHA cells treated with 100 ​μM FFAs. β-actin was used as loading control. Mean ​± ​SD indicated by horizontal lines and bars were given. ∗P ​< ​0.05. (F) Heatmap showing the expression of HSDL2, PPARα and A2aR in different subclusters of astrocytes. (G–H) The mRNA and protein levels of HSDL2, PPARα and A2aR were detected in NHA cells after HSDL2 knock down. β-actin was used as loading control. Mean ​± ​SD indicated by horizontal lines and bars were given. ∗P ​< ​0.05. (I) The BODIPY 493/503 staining was applied to indicate the LDs in NHA cells after HSDL2 knock down. Scale bar 20 ​μm. (J) Schematic diagram of the protective role of HSDL2 in alleviating the lipid accumulation in astrocytes. NHA, normal human astrocyte; BD493, BODIPY 493/503 staining; LD, lipid droplet; w/o HSDL2, without HSDL2; w/HSDL2, with HSDL2.

Once lipid accumulation occurred in astrocytes, characterized by increased expression of adenosine receptor A2aR [26], activation of PPARα and subsequent lipid catabolic pathway was activated to maintain the dynamic lipid homostasis [27]. Interestingly, we observed the A1 subcluster was characterized with increased PPARα expression and decreased A2aR expression (Fig. 6F). An increase of PPARα expression was also observed in the FFAs-treated NHAs, which was significantly alleviated by HSDL2 knockdown (Fig. 6G–H). Additionally, HSDL2 silencing resulted in a significant elevation of A2aR expression, thus providing a possible explanation for the protective activity of HSDL2 in reactive astrocytes (Fig. 6G–H). As expected, knockdown of HSDL2 led to increased LD accumulation in FFAs-treated NHAs, as observed through BD493/503 staining (Fig. 6I). In summary, the upregulation of HSDL2 served as a compensatory response of astrocytes to excessive lipid stimulation, exerting an anti-seizure effect (Fig. 6J).

Validation of blood HSDL2 expression in TLE cohort

Inspired by the diagnostic efficacy of HSDL2 in TLE samples, we wondered whether blood HSDL2 could be utilized as a biomarker to overcome the challenge of obtaining brain tissue. Firstly, HSDL2 mRNA expression was determined in dataset GSE143272 comprising of blood whole-genome mRNA expression profiling from 34 epilepsy patients and 50 healthy controls (HC). Elevated HSDL2 expression was observed in the peripheral blood of epilepsy patients (Fig. 7A, P ​< ​0.0001). Subsequently, HSDL2 expression were also examined with qRT-PCR in blood samples collected from 23 healthy donors and 20 TLE patients diagnosed on the anatomo-electro-clinical criteria in Qilu Hospital of Shandong University. As demonstrated, individuals with higher HSDL2 mRNA expression were more common in patients with TLE rather than healthy donors (P ​< ​0.0001), indicating the diagnostic potential of blood HSDL2 (Fig. 7B–C). Moreover, the ROC analysis was performed to evaluate the predictive accuracy of HSDL2, with an AUC value of 0.8478 (Fig. 7D). Interestingly, a higher proportion of blood HSDL2 expression was also observed in patients with extra-TLE (n ​= ​16) compared to healthy donors (P ​< ​0.0172), with an AUC value of 0.6712 (Fig. 7E). Collectively, the combined AUC of patients with epilepsy in our cohort, both TLE and extra-TLE, was 0.7693 (Fig. 7F), which suggested a powerful and robust predictive capacity of HSDL2 for epilepsy in general.

Fig. 7.

Fig. 7

Open in a new tab

Validation of blood HSDL2 expression as biomarker. (A) Expression of HSDL2 in dataset GSE143272 with blood samples. Mean ​± ​SD indicated by horizontal lines and bars were given. ∗P ​< ​0.05. (B) Bar plot showing the relative expression of HSDL2 in TLE (red), extra-TLE (blue) and healthy control (gray) blood samples. (C) Pie plot showing the distribution of HSDL2-high (red)/low (gray) expression in TLE (middle), extra-TLE(right) and control (left) blood samples. ∗P ​< ​0.05. (D–E) ROC curves of HSDL2 expression in TLE and extra-TLE Qilu cohort. (F) ROC curves of HSDL2 expression in Qilu cohort of epilepsy. HC, healthy control; TLE, temporal lobe epilepsy; ROC cureves, receiver operating characteristic curves.

Discussion

TLE is the most prevalent form of focal epilepsy in adults [2]. Despite extensive research efforts, the precise epileptogenic mechanisms and effective therapeutics for TLE remain elusive. In this study, we employed multiple transcriptomics datasets, including RNA sequencing and single-cell ones, to identify the pivotal role of mitochondria-related HSDL2 in human TLE. We not only verified its universality from independent validation datasets and our clinical samples, but also suggested it could be utilized as an indicator to reflect the seizure severity and HS. To the best of our knowledge, our study is the first to systematically analyze the existing RNA sequencing data acquired from TLE patients rather than animal models to elucidate the critical hub gene in human TLE. More importantly, we further identified the regulatory mechanism of HSDL2 on the lipid metabolism in astrocytes and demonstrated the diagnostic potential of blood HSDL2 in TLE.

Predominantly located in mitochondria and peroxisome [28], HSDL2 is a member of the short-chain dehydrogenase/reductase (SDR) subfamily of oxidoreductases and contains an N-terminal catalytic domain and a C-terminal sterol carrier protein type 2 (SCP-2) domain [29]. It has been defined as an oncogene in various malignancies, including esophageal adenocarcinoma [30], papillary thyroid carcinoma [31], lung adenocarcinoma [32], etc. Yang et al. proposed that HSDL2 participated in the tumorigenesis of cervical cancer by promoting cancer cell proliferation, invasion, and migration through regulation of epithelial-mesenchymal transition (EMT) and lipid metabolism [33]. On the contrary, Ma et al. showed that HSDL2 inhibited the progression of cholangiocarcinoma by modulating ferroptosis through the p53/SLC7A11 axis [34]. The role of HSDL2 within the blood-brain barrier remains to be elucidated. Chen et al. reported that the upregulation of HSDL2 existed in human gliomas, and silencing of HSDL2 suppressed glioma proliferation through an unknown mechanism [35]. A transcriptome-wide analysis focused on patients with Alzheimer's disease (AD) suggested that HSDL2 editing was associated with dementia, neuropathological measures, and longitudinal cognitive decline, which could be related to the lipid metabolism pathway [36]. Although the association between HSDL2 and epilepsy has never been reported, the aforementioned evidence, along with our comprehensive analysis, provides preliminary clues that astrocytic HSDL2 could modulate epileptic lipid dysfunction.

TLE often shows resistance to medication [37]. Repeated seizures can lead to long-term changes in epileptic dynamics and behavior, resulting in significant mental and cognitive impairments [38]. In this study, we conducted a comprehensive analysis using three independent cohorts of human TLE transcriptomics datasets to demonstrate the considerable accuracy of HSDL2 for TLE diagnosis. The upregulation of HSDL2 in TLE was then confirmed via qRT-PCR, western blotting, and IHC results. In addition, we incorporated GSE127871 and GSE255223, which contain information on seizure frequency and HS status, respectively, revealing its significant correlation with seizure severity and HS. To overcome the challenge of obtaining brain tissue, we also validated the expression of HSDL2 in blood samples from both external GEO dataset and our in-house epilepsy cohort, demonstrating the diagnostic potential of blood HSDL2 for epilepsy. Our approach combines the efficiency of high-throughput bioinformatics with the accuracy of clinical experiments to identify reliable disease biomarkers and potential therapeutics.

Neuron-astrocyte interactions have been implicated in various neurological disorders, such as stroke, AD, and epilepsy [39]. Ioannou et al. reported that toxic fatty acids produced in hyperactive neurons were transferred to and consumed by astrocytes via mitochondrial β-oxidation to trigger lipid metabolism reprogramming in acute stroke [40]. Mi et al. demonstrated that oxidative phosphorylation dysfunction of astrocytic mitochondria induced lipid droplet accumulation, followed by neuroinflammation and neurodegeneration in AD [13]. Recently, a subpopulation of lipid-accumulated reactive astrocytes (LARAs) was identified in patients with TLE, which was induced by APOE-mediated excessive lipid transfer from neurons and promoted neuronal epilepsy progression mediated by adenosine receptor A2aR [26]. Herein, our GO/KEGG analysis revealed that lipid metabolism dysregulation was enriched in human TLE. GSEA analysis and immunostaining further demonstrated that HSDL2 was abundant in astrocytes, with an intimate association of fatty acid and lipid metabolism. Mechanically, astrocytic HSDL2 upregulation could be a compensatory strategy to balance the epileptogenic effect secondary to excessive lipid transfer and accumulation. As a classic model for mimic excessive lipid accumlation in vitro, exogenous LDs induced pro-epileptogenic A2aR upregulation in astrocytes, accompanied by concurrent increase of HSDL2 expression. Intriguingly, HSDL2 knockdown further induced PPARα inhibition and A2aR upregulation. These observations indicated that neuron-derived lipid transport could elevate astrocytic HSDL2 expression, which, through the PPARα pathway, diminished the lipid accumulation and decreased the expression of A2aR, thus exerting anti-seizure effects. Although this compensatory mechanism didn't completely eliminate the excessive lipid accumulation associated with seizures, HSDL2 itself could serve as a valuable biomarker and provide potential targets for epilepsy.

Our study has several limitations. Firstly, although we utilized multiple GEO datasets to identify potential mitochondria-related hub genes for human TLE, the sample size was still limited. Future studies with larger cohorts are needed to validate our findings. Secondly, the crosstalk between the neurons and astrocytes is complex and dynamic. While our study provided insights into the lipid accumulation dysfunction in astrocytes, caution should be exercised in interpreting these findings. Thirdly, microglia, another key player that contributed to the brain network activity and epileptic pathogenesis [39,41], also exhibit considerable HSDL2 expression (Supplementary Fig. S4). Victor et al. reported that lipid-accumulated microglia weakly responded to neuronal activity and disrupted the coordinated activity of neuronal ensembles [42]. Our preliminary analysis indicated that increased microglia infiltration with high HSDL2 expression was observed in TLE (Supplementary Fig. S4A). GSEA analysis showed that HSDL2 was negatively correlated with cytokine activity and cytokine-cytokine receptor interaction (Supplementary Fig. S4B). Therefore, we supposed that HSDL2 might also participate in epileptogenesis by regulating inflammatory response in microglia. Finally, further investigation is warranted to understand the origin of blood HSDL2 and its biological significance.

Taken together, we have identified mitochondria-related HSDL2 as a potential biomarker to diagnose human TLE and reflect seizure severity via comprehensive bioinformatics analysis and clinical validation. Mechanically, elevated HSDL2 expression is induced by lipid dysregulation in astrocytes to exert a protective effect. Our results might provide promising potential for both HSDL2 blood biopsy and novel therapeutic approaches in human TLE.

Authors’ Contributions

Xiaxin Yang: Conceptualization, Methodology and Writing - Original Draft. Jianhang Zhang: Validation, Data Curation and Writing - Original Draft. Zhihao Wang: Formal analysis, Investigation. Zhong Yao and Xue Yang: Data Curation, Investigation. Xingbang Wang: Resources, Funding acquisition. Xiuhe Zhao: Resources, Funding acquisition and Supervision. Shuo Xu: Funding acquisition, Project administration, Writing - Review & Editing and Supervision.

Funding

National Natural Science Foundation of China, Grant/Award Number: 82272413; National Key Research and Development Program of China, Grant/Award Number: 2023YFC2413004 and 2022YFC2406404; The Taishan Scholarship Young Expert Program, Grant/Award Number: tsqn201909174; The Key R&D Program of Shandong Province, Grant/Award Number: 2022ZLGX03; Natural Science Foundation of Shandong Province, Grant/Award Number: ZR2021MH241. The China Postdoctoral Science Foundation, Certificate Number: 2024M751845.

Availability of Data and Materials

The GSE205661, GSE190451, GSE186334, GSE255223, GSE143272 and GSE127871 datasets were downloaded from the GEO database (https://www.ncbi.nlm.nih.gov/geo/). The data that support the findings of this study are available from the corresponding author on a reasonable request.

Ethics Approval and Consent to Participate

This study is covered under the Medical Ethical Committee, Qilu Hospital of Shandong University (KYLL-202204-014-1).

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgment

All authors gratefully acknowledge the Laboratory of Basic Medical Sciences (Qilu Hospital of Shandong University) for consultation and instrument availability.

Footnotes

Appendix A

Supplementary data to this article can be found online at https://doi.org/10.1016/j.neurot.2024.e00447.

Contributor Information

Xiuhe Zhao, Email: zhaoxiuhe@126.com.

Shuo Xu, Email: xushuo@sdu.edu.cn.

Appendix A. Supplementary data

The following are the Supplementary data to this article.

figs1.

figs1

Open in a new tab

figs2.

figs2

Open in a new tab

figs3.

figs3

Open in a new tab

figs4.

figs4

Open in a new tab

References

  • 1.Thijs R.D., Surges R., O'Brien T.J., Sander J.W. Epilepsy in adults. Lancet. 2019;393:689–701. doi: 10.1016/S0140-6736(18)32596-0. [DOI] [PubMed] [Google Scholar]
  • 2.Blumcke I., Spreafico R., Haaker G., Coras R., Kobow K., Bien C.G., et al. Histopathological findings in brain tissue obtained during epilepsy surgery. N Engl J Med. 2017;377:1648–1656. doi: 10.1056/NEJMoa1703784. [DOI] [PubMed] [Google Scholar]
  • 3.Hermann B., Conant L.L., Cook C.J., Hwang G., Garcia-Ramos C., Dabbs K., et al. Network, clinical and sociodemographic features of cognitive phenotypes in temporal lobe epilepsy. NeuroImage. Clinical. 2020;27 doi: 10.1016/j.nicl.2020.102341. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Johannesen K.M. From precision diagnosis to precision treatment in epilepsy. Nat Rev Neurol. 2023;19:69–70. doi: 10.1038/s41582-022-00756-0. [DOI] [PubMed] [Google Scholar]
  • 5.Abou Jaoude M., Jacobs C.S., Sarkis R.A., Jing J., Pellerin K.R., Cole A.J., et al. Noninvasive detection of hippocampal epileptiform activity on scalp electroencephalogram. JAMA Neurol. 2022;79:614–622. doi: 10.1001/jamaneurol.2022.0888. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Vezzani A., Balosso S., Ravizza T. Neuroinflammatory pathways as treatment targets and biomarkers in epilepsy. Nat Rev Neurol. 2019;15:459–472. doi: 10.1038/s41582-019-0217-x. [DOI] [PubMed] [Google Scholar]
  • 7.Walker L.E., Sills G.J., Jorgensen A., Alapirtti T., Peltola J., Brodie M.J., et al. High-mobility group box 1 as a predictive biomarker for drug-resistant epilepsy: a proof-of-concept study. Epilepsia. 2022;63:e1–e6. doi: 10.1111/epi.17116. [DOI] [PubMed] [Google Scholar]
  • 8.Lemprière S. New blood biomarker of refractory epilepsy. Nat Rev Neurol. 2023;19:326. doi: 10.1038/s41582-023-00819-w. [DOI] [PubMed] [Google Scholar]
  • 9.Chen Q.L., Xia L., Zhong S.P., Wang Q., Ding J., Wang X. Bioinformatic analysis identifies key transcriptome signatures in temporal lobe epilepsy. CNS Neurosci Ther. 2020;26:1266–1277. doi: 10.1111/cns.13470. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Wang L., Duan C., Wang R., Chen L., Wang Y. Inflammation-related genes and immune infiltration landscape identified in kainite-induced temporal lobe epilepsy based on integrated bioinformatics analysis. Front Neurosci. 2022;16 doi: 10.3389/fnins.2022.996368. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Chan F., Lax N.Z., Voss C.M., Aldana B.I., Whyte S., Jenkins A., et al. The role of astrocytes in seizure generation: insights from a novel in vitro seizure model based on mitochondrial dysfunction. Brain. 2019;142:391–411. doi: 10.1093/brain/awy320. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Giacomello M., Pyakurel A., Glytsou C., Scorrano L. The cell biology of mitochondrial membrane dynamics. Nat Rev Mol Cell Biol. 2020;21:204–224. doi: 10.1038/s41580-020-0210-7. [DOI] [PubMed] [Google Scholar]
  • 13.Mi Y., Qi G., Vitali F., Shang Y., Raikes A.C., Wang T., et al. Loss of fatty acid degradation by astrocytic mitochondria triggers neuroinflammation and neurodegeneration. Nat Metab. 2023;5:445–465. doi: 10.1038/s42255-023-00756-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Su Y., Tang M., Wang M. Mitochondrial dysfunction of astrocytes mediates lipid accumulation in temporal lobe epilepsy. Aging and disease. 2023;15:1289–1295. doi: 10.14336/AD.2023.0624. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Volmering E., Niehusmann P., Peeva V., Grote A., Zsurka G., Altmüller J., et al. Neuropathological signs of inflammation correlate with mitochondrial DNA deletions in mesial temporal lobe epilepsy. Acta Neuropathol. 2016;132:277–288. doi: 10.1007/s00401-016-1561-1. [DOI] [PubMed] [Google Scholar]
  • 16.Szklarczyk D., Gable A.L., Nastou K.C., Lyon D., Kirsch R., Pyysalo S., et al. The STRING database in 2021: customizable protein-protein networks, and functional characterization of user-uploaded gene/measurement sets. Nucleic Acids Res. 2021;49:D605–d612. doi: 10.1093/nar/gkaa1074. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Doncheva N.T., Morris J.H., Gorodkin J., Jensen L.J. Cytoscape StringApp: network analysis and visualization of proteomics data. J Proteome Res. 2019;18:623–632. doi: 10.1021/acs.jproteome.8b00702. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Engebretsen S., Bohlin J. Statistical predictions with glmnet. Clin Epigenet. 2019;11:123. doi: 10.1186/s13148-019-0730-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Wang Y., Yuan J., Yu X., Liu X., Tan C., Chen Y., et al. Vezatin regulates seizures by controlling AMPAR-mediated synaptic activity. Cell Death Dis. 2021;12:936. doi: 10.1038/s41419-021-04233-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Yang J., Jiang Q., Yu X., Xu T., Wang Y., Deng J., et al. STK24 modulates excitatory synaptic transmission in epileptic hippocampal neurons. CNS Neurosci Ther. 2020;26:851–861. doi: 10.1111/cns.13391. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Racine R.J. Modification of seizure activity by electrical stimulation. II. Motor seizure. Electroencephalogr Clin Neurophysiol. 1972;32:281–294. doi: 10.1016/0013-4694(72)90177-0. [DOI] [PubMed] [Google Scholar]
  • 22.Yang X., Yun Y., Wang P., Zhao J., Sun X. Upregulation of RCAN1.4 by HIF1α alleviates OGD-induced inflammatory response in astrocytes. Ann Clin Transl Neurol. 2022;9:1224–1240. doi: 10.1002/acn3.51624. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Xu S., Shao Q.Q., Sun J.T., Yang N., Xie Q., Wang D.H., et al. Synergy between the ectoenzymes CD39 and CD73 contributes to adenosinergic immunosuppression in human malignant gliomas. Neuro Oncol. 2013;15:1160–1172. doi: 10.1093/neuonc/not067. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Xu S., Wei J., Wang F., Kong L.Y., Ling X.Y., Nduom E., et al. Effect of miR-142-3p on the M2 macrophage and therapeutic efficacy against murine glioblastoma. J Natl Cancer Inst. 2014;106 doi: 10.1093/jnci/dju162. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Zhang Y., Fu X., Jia J., Wikerholmen T., Xi K., Kong Y., et al. Glioblastoma therapy using codelivery of cisplatin and glutathione peroxidase targeting siRNA from iron oxide nanoparticles. ACS Appl Mater Interfaces. 2020;12:43408–43421. doi: 10.1021/acsami.0c12042. [DOI] [PubMed] [Google Scholar]
  • 26.Chen Z.P., Wang S., Zhao X., Fang W., Wang Z., Ye H., et al. Lipid-accumulated reactive astrocytes promote disease progression in epilepsy. Nat Neurosci. 2023;26:542–554. doi: 10.1038/s41593-023-01288-6. [DOI] [PubMed] [Google Scholar]
  • 27.Su Y., Tang M., Wang M. Mitochondrial dysfunction of astrocytes mediates lipid accumulation in temporal lobe epilepsy. Aging Dis. 2024;15:1289–1295. doi: 10.14336/AD.2023.0624. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Cheng Z., Li Y., Sui C., Sun X., Xie Y. Synthesis, purification and crystallographic studies of the C-terminal sterol carrier protein type 2 (SCP-2) domain of human hydroxysteroid dehydrogenase-like protein 2. Acta crystallographica. Section F, Structural biology communications. 2015;71:901–905. doi: 10.1107/S2053230X15008559. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Gallegos A.M., Atshaves B.P., Storey S.M., Starodub O., Petrescu A.D., Huang H., et al. Gene structure, intracellular localization, and functional roles of sterol carrier protein-2. Prog Lipid Res. 2001;40:498–563. doi: 10.1016/s0163-7827(01)00015-7. [DOI] [PubMed] [Google Scholar]
  • 30.Wang F., Zhang L., Xu Y., Xie Y., Li S. Comprehensive analysis and identification of key driver genes for distinguishing between esophageal adenocarcinoma and squamous cell carcinoma. Front Cell Dev Biol. 2021;9 doi: 10.3389/fcell.2021.676156. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Zeng J., Ma X., Wang J., Liu R., Shao Y., Hou Y., et al. Down-regulated HSDL2 expression suppresses cell proliferation and promotes apoptosis in papillary thyroid carcinoma. Biosci Rep. 2019;39 doi: 10.1042/BSR20190425. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Shi Y., Mao Z., Huang Y., Sun Y., Cao Q., Yin X., et al. Knockdown of HSDL2 inhibits lung adenocarcinoma progression via down-regulating AKT2 expression. Biosci Rep. 2020;40 doi: 10.1042/BSR20200348. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Yang Y., Han A., Wang X., Yin X., Cui M., Lin Z. Lipid metabolism regulator human hydroxysteroid dehydrogenase-like 2 (HSDL2) modulates cervical cancer cell proliferation and metastasis. J Cell Mol Med. 2021;25:4846–4859. doi: 10.1111/jcmm.16461. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Ma S., Ma Y., Qi F., Lei J., Chen F., Sun W., et al. HSDL2 knockdown promotes the progression of cholangiocarcinoma by inhibiting ferroptosis through the P53/SLC7A11 axis. World J Surg Oncol. 2023;21:293. doi: 10.1186/s12957-023-03176-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Ruokun C., Yake X., Fengdong Y., Xinting W., Laijun S., Xianzhi L. Lentivirus-mediated silencing of HSDL2 suppresses cell proliferation in human gliomas. Tumour biology : the journal of the International Society for Oncodevelopmental Biology and Medicine. 2016;37:15065–15077. doi: 10.1007/s13277-016-5402-6. [DOI] [PubMed] [Google Scholar]
  • 36.Ma Y., Dammer E.B., Felsky D., Duong D.M., Klein H.U., White C.C., et al. Atlas of RNA editing events affecting protein expression in aged and Alzheimer's disease human brain tissue. Nat Commun. 2021;12:7035. doi: 10.1038/s41467-021-27204-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Lamberink H.J., Otte W.M., Blümcke I., Braun K.P.J. Seizure outcome and use of antiepileptic drugs after epilepsy surgery according to histopathological diagnosis: a retrospective multicentre cohort study. Lancet Neurol. 2020;19:748–757. doi: 10.1016/S1474-4422(20)30220-9. [DOI] [PubMed] [Google Scholar]
  • 38.Larivière S., Royer J., Rodríguez-Cruces R., Paquola C., Caligiuri M.E., Gambardella A., et al. Structural network alterations in focal and generalized epilepsy assessed in a worldwide ENIGMA study follow axes of epilepsy risk gene expression. Nat Commun. 2022;13:4320. doi: 10.1038/s41467-022-31730-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Patel D.C., Tewari B.P., Chaunsali L., Sontheimer H. Neuron-glia interactions in the pathophysiology of epilepsy. Nat Rev Neurosci. 2019;20:282–297. doi: 10.1038/s41583-019-0126-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Ioannou M.S., Jackson J., Sheu S.H., Chang C.L., Weigel A.V., Liu H., et al. Neuron-astrocyte metabolic coupling protects against activity-induced fatty acid toxicity. Cell. 2019;177:1522–1535.e14. doi: 10.1016/j.cell.2019.04.001. [DOI] [PubMed] [Google Scholar]
  • 41.Kumar P., Lim A., Hazirah S.N., Chua C.J.H., Ngoh A., Poh S.L., et al. Single-cell transcriptomics and surface epitope detection in human brain epileptic lesions identifies pro-inflammatory signaling. Nat Neurosci. 2022;25:956–966. doi: 10.1038/s41593-022-01095-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Victor M.B., Leary N., Luna X., Meharena H.S., Scannail A.N., Bozzelli P.L., et al. Lipid accumulation induced by APOE4 impairs microglial surveillance of neuronal-network activity. Cell Stem Cell. 2022;29:1197–1212.e8. doi: 10.1016/j.stem.2022.07.005. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

The GSE205661, GSE190451, GSE186334, GSE255223, GSE143272 and GSE127871 datasets were downloaded from the GEO database (https://www.ncbi.nlm.nih.gov/geo/). The data that support the findings of this study are available from the corresponding author on a reasonable request.

Articles from Neurotherapeutics are provided here courtesy of Elsevier

RESOURCES