FGF13 alleviates astrocytic apoptosis via JIP2 inhibition in the hippocampus and mitigates depression-like behavior

Abstract

Major depressive disorder (MDD) is one of the leading causes of disability worldwide and significantly increases the risk of premature death and other diseases. Astrocyte loss is a key pathological hallmark of MDD, yet the underlying mechanisms remain unclear. Here, we identify fibroblast growth factor 13 (FGF13) as a critical regulator of astrocyte apoptosis in depression, which is closely associated with depression-like behaviors in mice. In depressive models, FGF13 expression is markedly reduced, particularly in astrocytes, accompanied by astrocyte apoptosis in the hippocampal region and decreased synaptic protein levels in the nervous system. Astrocyte-specific knockout of FGF13 induces astrocytic apoptosis, exacerbates inflammatory levels, and aggravates depression-like behaviors in mice. In contrast, astrocyte-specific overexpression of FGF13 significantly attenuates both astrocyte apoptosis and inflammation, and effectively ameliorates depression-like behaviors. Mechanistically, FGF13 directly binds to JIP2 protein, inhibits its activity, and subsequently blocks the downstream JIP2-JNK signaling pathway, thereby suppressing Bax/Bcl-2-mediated astrocyte apoptosis. These findings reveal a key mechanism by which FGF13 regulates astrocyte death in depression and highlight its potential as a therapeutic target for MDD, offering new insights for the development of antidepressant drugs targeting astrocytes.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12974-026-03960-7.

Introduction

Major depressive disorder (MDD) is a highly prevalent psychiatric disorder worldwide [1]. Characterized by high prevalence and disability rates, MDD represents one of the most burdensome mental illnesses globally. Its primary clinical manifestations include persistent negative mood, cognitive impairment, and anhedonia (the reduced ability to experience pleasure), which can culminate in suicide in extreme cases [2, 3]. Current prevailing hypotheses regarding MDD pathogenesis include the monoamine, hypothalamic-pituitary-adrenal (HPA) axis dysfunction, inflammation, and neuroplasticity hypotheses [4, 5]. However, none of these models fully elucidates the complex etiology of MDD. Antidepressants developed based on these hypotheses—such as monoamine oxidase inhibitors [6], selective serotonin reuptake inhibitors [7], and N-methyl-D-aspartate (NMDA) receptor antagonists—frequently suffer from a delayed onset of action [8], low remission rates, high relapse rates, and significant adverse effects. Therefore, unraveling the precise molecular mechanisms underlying MDD and exploring novel therapeutic strategies are urgently needed.

Previous studies indicate that astrocyte atrophy and loss may concurrently manifest with or significantly contribute to the progression of depression [9, 10]. For instance, individuals with MDD exhibit a marked reduction in glial cell density and number in key brain regions, including the medial prefrontal cortex (mPFC), hippocampus, and amygdala [11, 12]. As the most abundant glial cells in the brain, astrocytes are vital for regulating numerous biological processes, such as maintaining extracellular ion and neurotransmitter homeostasis, promoting neuronal and synaptic development, and modulating neuronal metabolism [13, 14]. Under pathological conditions, however, astrocytes undergo reactive transformation, secreting copious amounts of cytokines, chemokines, and growth factors that precipitate a rapid escalation in neuroinflammation [5]. MDD patients typically display elevated serum levels of pro-inflammatory cytokines, including interleukin-1 (IL-1), interleukin-6 (IL-6), and tumor necrosis factor-alpha (TNF-α), alongside endogenous metabolites like nesfatin-1 and cortisol [15]. MDD patients often exhibit elevated cortisol levels [16], whereas rodent models of depression show corresponding increases in corticosterone [17]. Astrocytes both secrete and respond to these inflammatory mediators, amplifying inflammatory cascades and skewing cerebral immune homeostasis toward a neurotoxic microenvironment [18]. Furthermore, systemic administration of these cytokines has been shown to induce depression-like behaviors in mice, such as anhedonia, fatigue, and helplessness [19, 20]. Astrocytes respond to peripheral inflammation rapidly—often within hours—and blocking their activation with broad-spectrum antagonists can ameliorate depression-like behaviors following peripheral inflammatory challenges [21, 22]. Despite these insights, the exact mechanisms triggering astrocyte loss and atrophy remain obscure, and our understanding of the cellular signaling cascades underlying astrocyte-mediated neuroinflammation is notably limited.

In recent years, the role of fibroblast growth factors (FGFs) in brain-related diseases has attracted widespread attention [23], but research on FGFs in depression is an emerging field. The FGF family is a large family of growth factors involved in early brain development as well as maintenance and repair throughout adulthood [24]. Studies have found that several FGF family members are associated with depression, including FGF2, FGF9, FGF21, and FGF22 [25, 26]. However, their functions differ; for example, FGF2 effectively reduces anxiety- and depression-like behaviors in mice [27], whereas FGF9 increases these behaviors [28]. In addition, the expression levels of FGF21 and FGF22 are negatively correlated with the severity of depression [29], and knockout of FGF22 leads to loss of hippocampal excitatory synapses and induces depression-like behaviors in mice [30, 31]. Notably, FGF13 is an intracellular non-secreted protein that is highly expressed in the human brain, where it influences brain development and cognitive function by regulating the nervous system [32]. A recent study found that FGF13 improves astrocyte function and inflammatory responses in Parkinson’s disease (PD) by regulating mitochondrial function [33]. However, whether FGF13 plays a role in major depressive disorder has not yet been reported. Given the important role of FGF13 in nervous system function, we hypothesize that FGF13 may play a significant role in the molecular mechanisms driving astrocyte loss in depression.

In the present study, we conducted a comprehensive analysis of RNA-seq data from MDD patient cohorts in the GEO database, followed by a series of in vivo and in vitro experiments. Our analysis revealed that astrocytic FGF13 expression is significantly downregulated in the brains of MDD patients and similarly reduced in the hippocampus of murine depression models of chronic restraint stress (CRS) and chronic social defeat stress (CSDS). CRS and CSDS model certain stress-related behavioral and molecular features relevant to MDD, and these models as tools for investigating stress-induced neurobiological adaptations, rather than as direct representations of the human disease. We further demonstrated that conditional knockout of FGF13 exacerbates depression-like behaviors, astrocytic apoptosis, and neuroinflammation. Mechanistically, FGF13 binds to and inhibits the phosphorylation of the JIP2 protein, effectively blocking the downstream JNK-mediated Bax/Bcl-2 apoptotic signaling cascade. These findings underscore the critical role of FGF13 in MDD, provide a novel molecular mechanism for astrocyte depletion in depressed patients, and position FGF13 as a promising therapeutic target for depression.

Materials and methods

Animal and experimental design

All animal-related experiments were strictly conducted in accordance with the Animal Care Ethics Guidelines of Wenzhou Medical University and were approved by the Ethics Committee (XMSQ2025-0259). The mice were housed in a specific pathogen-free (SPF) environment, with room temperature maintained at 22 ± 2 °C and a 12-hour light/dark cycle. The animals had free access to sterilized SPF pellet rodent feed and sterilized water. All mice used in the experiments were 8–10 weeks of age. To exclude the influence of sex on depression-like behaviors, as in previous studies, all mice were male. C57BL/6J mice (8–10 weeks) were purchased from Charles River. FGF13 ^flox/flox (FGF13 ^f/f) mice on a C57BL/6J background and GFAP-cre mice on a C57BL/6J background were generously provided by Professor Kong from Wenzhou University. The FGF13 knockdown group (FGF13-KD) mice were generated by injecting FGF13 ^f/f mice with AAV9 harboring the FGF13 knockout vector (pAAV-GfaABC1D-mCherry-P2A-Cre-WPRE) at a titer of 5 × 10¹² vg/mL, and pAAV-GfaABBC1D-mCherry-T2A-WPRE (AAV-Ctrl1) as the control vector, supplied by OBIO Technology (Shanghai). The FGF13 overexpression group (FGF13-OE) mice were generated by injecting GFAP-Cre mice with AAV9 harboring the FGF13 overexpression vector (pAAV-CMV-DIO-Fgf13-3xFLAG-T2A-EGFP-WPRE) at a titer of 5 × 10¹² vg/mL, and pAAV-CMV-DIO-3-flag-T2A-EGFP (AAV-Ctrl2) as the control vector, supplied by OBIO Technology (Shanghai). The reference sequence for FGF13 was NM_010200.3.

Stereotaxic micro-injection

As described in previous study [34], mice were anesthetized with isoflurane (3% for induction, 1.5% for maintenance) and then fixed in a digital stereotaxic instrument (RWD Life Sciences Co., Ltd., China) for virus injection. The scalp was incised, the skull was scrubbed with iodine and rinsed with saline. Unless otherwise specified, all coordinates for different sites were measured relative to bregma. The aforementioned AAVs were injected into the dorsal hippocampal region (AP -1.90 mm, ML ± 1.50 mm, DV -1.50 mm). The injection volume was 200 nl, with a flow rate controlled at 40–50 nl/min. After injection, the needle was left in place for approximately 5 min, then slowly withdrawn after sealing the needle track with saline. Following suture of the incision, the mice were transferred to a heating pad until they regained consciousness.

Cell culture and transfection

Astrocyte C6 cells (cat. TCR1) were purchased from the Cell Bank of the Chinese Academy of Sciences and cultured in DMEM (Gibco) supplemented with 10% fetal bovine serum at 37°C in a 5% CO₂ atmosphere. After stable cell growth was achieved, C6 cells were seeded into 6-well plates (10⁵ cells/well) for infection and transfection and cultured overnight until reaching 30–40% confluence. FGF13 knockdown was performed using FGF13-specific short hairpin RNA (shRNA) and its negative control as previously described [35]. The shFGF13 sequence information was as follows: 5′-caccgcacttacactctgtttaaccttcaagagagttaacagtgtaagtgcttttg-3′ and 5′-gatccaaaaaagcacttacactctgttaacctctgaaggttaaacagtgtaagtgc-3′. The negative control sequence information was as follows: 5’-caccgttctccgaacgtgtcacgtcaagagattacgtgacacgttcggagaattttttg-3’ and 5’-gatccaaaaaattctccgaacgtgtcacgtaatctcttgacgtgacacgttcggagaac-3’. The lentiviral vectors were both constructed in the pLKO.1-EGFP-Puro backbone (Chemshine, Shanghai, China). The lentivirus titer was 6 × 10⁸ TU/mL. For FGF13 overexpression, the FGF13 coding sequence (NM_010200.3) was cloned into the pcSLenti-CMV-3xFlag-PGK-Puro-WPRE3 backbone (OBiO, Shanghai, China), and the empty vector (without FGF13) served as the control. The overexpression lentivirus titer was 8 × 10⁸ TU/mL. After lentiviral infection at a multiplicity of infection (MOI) of 10, cells were incubated with 6 µg/mL polybrene for 12 h to enhance infection efficiency. The supernatant was then replaced with fresh medium. When cell confluence reached 80%, 3 µg/mL puromycin was added to the medium to eliminate uninfected cells. After stable transduced cells emerged, the puromycin concentration in the medium was reduced to 1 µg/mL to maintain selection of infected cells. The resulting stable cell lines were designated FGF13-KD and FGF13-OE, respectively.

PC12 cells (SCSP-5259) were purchased from the Cell Bank of the Chinese Academy of Sciences and cultured in RPMI 1640 medium (Gibco) supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin (Gibco) at 37 °C in a 5% CO₂ atmosphere. BV2 cells (SCSP-5208) were purchased from the Cell Bank of the Chinese Academy of Sciences and cultured in DMEM (Gibco) containing 4.5 g/L glucose, supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin (Gibco), at 37 °C in a 5% CO₂ atmosphere.

Chronic Restraint Stress (CRS)

Mice were placed in well-ventilated 50 mL polypropylene tubes with holes, which restricted their movement for 3 h daily (from 10:00 a.m. to 1:00 p.m.) for 10 consecutive days. During the restraint period, food and water were withheld. Non-stressed control mice were also deprived of food and water during the same period.

Chronic Social Defeat Stress (CSDS)

The CSDS paradigm was performed as previously described [36]. Briefly, C57BL/6J mice were each exposed to a CD1 aggressor with a reliable attack latency for 10 min daily for 10 consecutive days. A new CD1 mouse was used for each C57 mouse each day. Following the social defeat procedure, the defeated mice were placed on the opposite side of the aggressive CD1 mouse and exposed to visual, olfactory, and auditory cues from the CD1 mouse for 24 h until the next physical defeat session. Control mice were housed in pairs and switched between ten different C57BL/6J mice for 10 consecutive days. Animal behavioral tests were conducted 24 h after the last session.

Behavioral tests

Forced Swim Test (FST): during the test, each mouse was gently placed into a glass cylinder (16 cm in diameter, 32 cm in height) containing 20 cm of water (25 ± 1 °C). Mice were video-recorded while swimming/floating for 6 min. The struggling time and immobility time during the last 5 min of the recording were recorded to assess depression-like behavior. Tail Suspension Test (TST): mice were suspended by their tails using a clip on a mold 25 cm above the floor. A partition was placed between each mouse to prevent interference. After recording for 6 min, the mice were returned to their home cages. Immobility time during the last 5 min was scored for each mouse. Sucrose Preference Test (SPT): before the formal test, mice were housed individually and habituated to two identical bottles containing 2% sucrose and water, respectively, for 24 h. To avoid side preference during the test, the positions of the bottles were switched every 12 h. Prior to the test, mice were deprived of food and water for 24 h. Following the deprivation period, each mouse was given the two bottles (one with 2% sucrose and one with water) for 24 h, with the order of the bottles randomly assigned. Bottles were weighed before and after the test to quantify the amount of sucrose solution and water consumed, and to determine the percentage of sucrose consumption. Sucrose preference was calculated using the formula: Sucrose preference (%) = [Sucrose consumption / (Water consumption + Sucrose consumption)] × 100%. Open field test (OFT): The apparatus consisted of a square arena (50 cm × 50 cm × 50 cm) illuminated with dim light (~ 20 lx). Each mouse was placed in the center of the arena and allowed to explore freely for 5 min. Total distance traveled, average velocity, and time spent in the central zone were recorded and analyzed using an automated tracking system (ANY-maze). The arena was cleaned with 75% ethanol between each trial to eliminate olfactory cues.

Immunofluorescence staining

Mice were anesthetized with 2% sodium pentobarbital and transcardially perfused with 0.9% saline, after which brain tissues were collected. One hemisphere was placed in 4% paraformaldehyde at 4 °C overnight, while the other hemisphere was dissected to isolate hippocampal and cortical tissues, which were then stored at − 80 °C. The fixed hemispheres were dehydrated sequentially using 30% and 10% sucrose gradients. Tissues were then embedded in OCT for frozen sectioning, and 30 μm brain sections were prepared using a Cryostat Microtome (CM1850, Leica). The immunostaining procedure for brain sections and cells was as follows: Brain sections were first treated with Triton solution (permeabilization agent, 0.05%), then blocked with 10% goat serum for 30 min at room temperature, followed by overnight incubation with primary antibodies (see Table S1) at 4 °C. The sections were then incubated with Alexa Fluor 488 (1:500, A-11034, Invitrogen) or Alexa Fluor 594 (1:500, R37115, Invitrogen) secondary antibodies for 1 h at room temperature, followed by three washes with PBS (10 min each). Nuclei were stained with DAPI (1:500, HY-D0814, MCE). Finally, images were acquired and analyzed using a Nikon confocal microscope (Nikon, A1R). Quantification was performed using ImageJ/Fiji software (National Institutes of Health, USA). The percentage of positive staining area was calculated as the positive signal area divided by the total field area × 100%. To ensure objective thresholding, a unified intensity threshold was applied across all images, set at 2-fold above the mean background fluorescence of each section. This threshold was kept constant for all images within the same experimental batch to avoid investigator bias. The percentage of cleaved caspase-3⁺/GFAP⁺ cells were calculated as the number of double-positive cells divided by the total number of GFAP⁺ astrocytes × 100%.

Western blotting

Brain tissues and cultured astrocytes were collected as protein samples. Samples were lysed in RIPA lysis buffer supplemented with protease inhibitor (1:100) and phosphatase inhibitor (1:100, PMSF). After centrifugation at 12,000 g for 10 min, the supernatants were collected and stored at − 80 °C. Protein concentrations were determined using a BCA kit (P0012, Beyotime Biotechnology). Twenty micrograms of protein were selected for gel electrophoresis, followed by wet transfer onto PVDF membranes. The membranes were blocked with 5% non-fat milk for 1 h at room temperature (RT) and then incubated with primary antibodies (see Table S1) overnight at 4 °C. After four washes with TBST (15 min each), the membranes were incubated with secondary antibodies for 1 h at RT, followed by four additional washes with TBST (15 min each). Antibody binding was detected using a chemiluminescent HRP substrate (WBKLS0500, Millipore). Protein bands were visualized using a GE Amersham Imager 680 imaging system and quantified with ImageJ software. All sample values were normalized to GAPDH, β-actin, or β-tubulin.

Immunoprecipitation (Co-IP)

Protein samples were incubated with primary antibodies against FGF13 (1:50) or JIP2 (1:50), along with control immunoglobulin G, overnight at 4 °C. Protein A + G agarose beads (P2012, Beyotime Biotechnology) were then added and incubated for an additional 6–8 h. After three washes with PBS (15 min each), the samples were resuspended in loading buffer and boiled for 10 min. Finally, the bound proteins were collected and analyzed by Western blotting.

ELISA

The following ELISA kits were used according to the manufacturers’ instructions: IL-1β (KE10003, Proteintech), IL-6 (KE10007, Proteintech), TNF-α (KE10002, Proteintech), and IL-18 (ml1066842, Mlbio) to measure the concentrations of IL-1β, IL-6, TNF-α, and IL-18 in serum and cell culture supernatants. In addition, glutamate levels in cell culture supernatants were measured using a glutamate assay kit (ml692212, Mlbio), and neuropeptide levels were measured using an arginine vasopressin assay kit (ml105228, Mlbio). Chemiluminescent signals were captured using an ImageQuant LAS4000 scanner (GE Healthcare, USA).

Flow cytometry

Cells were washed three times with PBS, then digested with trypsin (C25200, Gibco) for 20–30 s. After washing with PBS, cells were collected by centrifugation at 1,000 g for 5 min. Cell staining was performed according to the instructions of the Annexin V-FITC/PI Apoptosis Kit (C1062M, Beyotime Biotechnology), followed by analysis on a flow cytometer (FACSAria II, BD).

Transcriptome analysis

RNA sequencing of RNA extracted from C6 cells was performed by Personalbio (Shanghai, China). Briefly, RNA integrity was first assessed using an Agilent 2100 Bioanalyzer, after which sequencing libraries were constructed. Sequencing was performed on an Illumina platform, and high-quality data were obtained in FASTQ format. Raw transcriptome reads were mapped to reference gene sequences using Hisat2 v2.0.5, and the number of reads for each gene was quantified using FeatureCounts v1.5.0-p3. Gene expression levels were calculated as FPKM based on gene length and the number of mapped reads. Differential expression analysis between the two groups was performed using the DESeq2 R package (v2.12), with an adjusted p-value < 0.05 set as the threshold for significant differential gene expression. Finally, the clusterProfiler R package (v3.8.1) was used to identify and rank GO and KEGG enrichment pathways, which were sorted by enrichment significance (p < 0.05).

Molecular docking simulation

For protein docking, molecular docking between FGF13 (Q13387) and JIP2 (Q92913) was performed using AutoDockTools 1.5.6 and AutoDock Vina. The crystal structures of the proteins were obtained from the RCSB Protein Data Bank (http://www.rcsb.org). These structures were imported into PyMOL 2.4.1 to remove water molecules and original ligands. Hydrogen atoms were added to the receptor proteins, and their charges and atom types were calculated using AutoDockTools 1.5.6. The resulting structures were saved in pdbqt format. Molecular docking and binding energy calculations were performed using AutoDock Vina. The interaction between FGF13 and the target protein SigmaR1 was visualized using PyMOL 2.4.1 and Discovery Studio 2021.

Cell viability assay

Cell viability was assessed using the Cell Counting Kit-8 (CCK-8) (HY-K0301, MCE). Briefly, different cell suspensions (100 µL/well) were seeded in 96-well plates at a density of 10⁴ cells per well. Drugs such as LPS (1 µg/ml, HY-D1056, MCE) or SP600125 (20 µM, HY-12041, MCE) were added to the cell culture medium at different concentrations, and the plates were pre-incubated in a humidified incubator at 37 °C with 5% CO₂. At the end of the drug treatment period, 10 µL of CCK-8 solution was added to each well. The plates were then incubated for 1 h in the incubator, and the absorbance at 450 nm was measured using a microplate reader.

Primary hippocampal neuron and astrocyte culture

This protocol was adapted from a previously published method [37]. Hippocampal astrocytes were isolated from postnatal day 0–2 (P0–P2) C57BL/6 mouse pups. Briefly, hippocampi were dissected, minced, and digested with 0.25% papain at 37 °C for 15 min. Cells were plated on poly-D-lysine–coated dishes (Sigma-Aldrich, USA) and maintained in DMEM/F12 supplemented with 20% fetal bovine serum (FBS) and 1% penicillin/streptomycin at 37 °C in a humidified 5% CO₂ incubator. After 5–7 days in culture, astrocytes were purified by differential trypsinization to remove microglia and oligodendrocyte precursor cells. Purity was confirmed by GFAP immunostaining (≥ 95% GFAP-positive cells). Purified astrocytes were then transferred to six-well plates for subsequent experiments. All experiments were performed using primary astrocytes at passages 2–3.

Co-culture experiment

Co-culture experiments were performed using Transwell inserts. Briefly, cultured PC12 cells or BV2 cells were first seeded into the outer compartment and allowed to adhere. Meanwhile, cultured and treated C6 cells were digested, washed, and seeded into the inner compartment. After 24 h of co-culture, the culture supernatant from the outer compartment was collected for ELISA detection.

Statistical analysis

All statistical analyses were performed using GraphPad Prism 8.0 (GraphPad Software, Inc.). Before performing any parametric tests, the normality of data distribution was verified using the Shapiro-Wilk test, and homogeneity of variance was assessed using Levene’s test. All datasets met the assumptions for parametric analysis (p > 0.05 for both tests) unless otherwise noted. For datasets that violated these assumptions, non-parametric alternatives were applied as specified below. The choice of statistical test was determined by the experimental design: (1) Single-factor two-group comparisons were analyzed using unpaired two-tailed Student’s t-test. (2) Single-factor multi-group comparisons were analyzed using One-way ANOVA followed by Bonferroni’s post-hoc test for multiple comparisons. (3) Factorial designs involving two independent variables were analyzed using Two-way ANOVA with both factors as fixed variables, including the interaction term. When a significant interaction was detected, Bonferroni’s post-hoc test was applied for pairwise comparisons between groups. All experiments were conducted with the experimenter blinded to genotype/treatment conditions during data acquisition and analysis. Animals were randomly assigned to experimental groups using a computer-generated random number table. Exact sample sizes (n) are provided in each figure legend, and all data points represent biological replicates from independent experiments or animals. Data are presented as mean ± SEM, and significance levels are indicated as *p < 0.05, **p < 0.01.

Results

Expression of FGF13 is significantly downregulated in depression and accompanied by astrocyte loss

Whether FGF13 affects the function of the central nervous system in patients with MDD remains unclear. Therefore, we first systematically analyzed publicly available sequencing data from MDD patients in the GEO database. In dataset GSE144136 (See Fig. S1 for details) [38], we examined the expression changes of FGF13 in different neural cell types in the prefrontal cortex of MDD patients (Fig. 1A). Further analysis revealed that FGF13 expression was significantly decreased in astrocytes (Fig. 1B), but showed no significant changes in microglia (Fig. 1C), excitatory neurons (Fig. 1D), inhibitory neurons (Fig. 1E), or oligodendrocytes (Fig. 1F). To determine whether the observed FGF13 downregulation in astrocytes from MDD patients reflected true transcriptional changes or was secondary to reduced astrocyte abundance, we quantified astrocyte numbers in the GSE144136 dataset. Consistent with our mouse model findings and previous reports [38], the proportion of astrocytes was significantly lower in MDD samples compared to controls (1,425 vs. 2,398 cells; Figure S1B). To control for this compositional bias, we performed pseudobulk differential expression analysis using DESeq2. Notably, FGF13 expression remained significantly downregulated in MDD astrocytes after adjusting for cell count imbalance (Figure S1C), indicating that the reduction of FGF13 is not merely a byproduct of astrocyte loss but reflects genuine transcriptional downregulation within the remaining astrocyte population. To clinically corroborate this finding, we established chronic restraint stress (CRS) and chronic social defeat stress (CSDS) murine models to recapitulate the pathological features of MDD (Fig. 1G). Successful model induction was validated via behavioral paradigms; compared to controls, both CRS and CSDS mice exhibited significantly increased immobility times in the forced swimming test (FST) (Fig. 1H) and tail suspension test (TST) (Fig. 1I), alongside diminished sucrose preference in the sucrose preference test (SPT) (Fig. 1J). By further examining FGF13 expression levels in the cerebral cortex and hippocampus of CRS and CSDS model mice, we found that FGF13 expression was significantly decreased in the hippocampus of depressed mice (Fig. 1K), while its expression in the cortex showed a decreasing trend that did not reach statistical significance (Fig. 1K). Although this finding differs from the changes in FGF13 expression in the cerebral cortex observed in the GSE144136 dataset, the hippocampus is a well-established node in the depressive circuitry [39]. Therefore, our subsequent experiments focused on the effects of hippocampal FGF13 expression on depression.

Fig. 1.

FGF13 expression is decreased in depressed patients and mouse models of depression, accompanied by astrocyte apoptosis. A. Distribution of FGF13 gene expression in brain tissues of MDD patients from a published snRNA-seq dataset (GSE144136; 17 healthy controls and 17 MDD cases). B–F. Changes in FGF13 gene expression in different cell types, including astrocytes (B), microglia (C), excitatory neurons (D), inhibitory neurons (E), and oligodendrocytes (F). G. Schematic diagram of the procedures for establishing two mouse models of depression: CRS and CSDS. H–J. Performance of wild-type (WT), CRS, and CSDS mice in the forced swimming test (H), tail suspension test (I), and sucrose preference test (J). n = 8–12 mice per group. K. Representative immunoblots of FGF13 and β-actin in hippocampal (HPC) and prefrontal cortical (PFC) tissues from WT, CRS, and CSDS mice (left), and quantitative analysis of FGF13 expression changes in each group (right). n = 4 mice per group. L. Representative immunoblots of GFAP, S100β, and GAPDH in hippocampal tissues from WT, CRS, and CSDS mice (left), and quantitative analysis of GFAP and S100β expression changes in each group (right). n = 4 mice per group. M. Representative fluorescence images of astrocytes (GFAP, green), FGF13 (red), and DAPI (blue) in the hippocampus (CA1, CA3, and DG regions) of WT, CRS, and CSDS mice. Scale bar = 100 μm. N. Representative fluorescence images of microglia (Iba1, green), FGF13 (red), and DAPI (blue) in the hippocampus (CA1, CA3, and DG regions) of WT, CRS, and CSDS mice. Scale bar = 100 μm. O–R. Quantification of the percentage of positive fluorescence signals from Figures M and N, including the percentage of FGF13-positive area in Figure M (O), the percentage of GFAP-positive astrocytes in Figure M (P), the percentage of FGF13-positive area in Figure N (Q), and the percentage of Iba1-positive microglia in Figure N (R). n = 4–6 mice per group. Data are presented as mean ± SEM. *P < 0.05, **P < 0.01

Furthermore, we found that the expression of the astrocyte markers GFAP and S100β was significantly decreased in the hippocampus of CRS and CSDS model mice (Fig. 1L). We also observed a significant decrease in astrocyte density in the hippocampal regions (DG, CA1, CA3) of CRS and CSDS depressed mice (Fig. 1M and P), which closely co-localized with the spatial reduction in FGF13 expression (Fig. 1M and O). These results suggest that FGF13 deficiency may be an important factor contributing to astrocyte loss in depression; however, the precise mechanisms governing this regulation warrant further investigation. Studies have demonstrated that neuroinflammation plays a critical role in the onset and progression of depressive symptoms [5, 20]. Therefore, we further examined whether damage occurred in the central nervous system. Microglia are rapidly activated in response to various external stressors (such as inflammation, injury, and psychological stress). We detected widespread microglial activation in the hippocampal regions (DG, CA1, CA3) of CRS and CSDS depressed mice (Fig. 1N and R). In addition, the expression of the synaptic marker proteins PSD95 and Synapsin-1 in hippocampal tissue was significantly decreased in CRS and CSDS depressed mice compared with normal mice (Fig. 2A), indicating damage to synaptic structures related to brain neurological function. Further analysis showed that, compared with normal mice, the expression of the apoptosis-related protein Bax was significantly increased in the hippocampus of CRS and CSDS depressed mice (Fig. 2B), whereas the expression of the anti-apoptotic protein Bcl-2 was significantly decreased (Fig. 2B). Moreover, we measured the expression levels of inflammatory factors in serum. The results showed that, compared with normal mice, the serum levels of IL-6 (Fig. 2C), TNF-α (Fig. 2D), IL-18 (Fig. 2E), and IL-1β (Fig. 2F) were significantly increased in CRS and CSDS depressed mice. Collectively, these data suggest that astrocytic loss in depression triggers synaptic dysfunction and neuroinflammation, prompting the hypothesis that FGF13 protects astrocytes and mitigates inflammatory damage.

Fig. 2.

FGF13 regulates astrocyte apoptosis and contributes to neuronal injury. A. Representative immunoblots of PSD95, Synapsin-1, and GAPDH in hippocampal tissues from WT, CRS, and CSDS mice (left), and quantitative analysis of PSD95 and Synapsin-1 expression changes in each group (right). n = 4 mice per group. B. Representative immunoblots of Bax, Bcl-2, and β-actin in hippocampal tissues from WT, CRS, and CSDS mice (left), and quantitative analysis of Bax and Bcl-2 expression changes in each group (right). n = 4 mice per group. C–F. ELISA measurements of inflammatory factor expression levels in serum from WT, CRS, and CSDS mice, including IL-6 (C), TNF-α (D), IL-18 (E), and IL-1β (F). n = 8–10 mice per group. G–H. Representative immunostaining images of C6 astrocyte cell line after transfection with FGF13 interference virus (G), showing GFAP (red) and EGFP (green); representative images and quantitative analysis of FGF13 expression in C6 cells after lentiviral infection (H). n = 3 per group. I–J. Flow cytometry detection of astrocyte apoptosis rates in the normal control, FGF13-KD, LPS (1 µg/mL, 24 h), and LPS-treated FGF13-KD astrocyte groups (I), and statistical analysis of apoptosis rates in each group (J). n = 3 per group. K. Representative immunoblots of Bax, Bcl-2, and β-actin in cells from the control, FGF13-KD, LPS, and LPS+FGF13-KD groups (left), and quantitative analysis of Bax and Bcl-2 expression changes in each group (right). n = 3 per group. L–M. Representative immunostaining images of C6 astrocyte cell line after transfection with FGF13 overexpression adenovirus (L), showing GFAP (green) and Flag (red); representative images and quantitative analysis of FGF13 expression in C6 cells after adenoviral infection (M). n = 3 per group. N–O. Flow cytometry detection of astrocyte apoptosis rates in the normal control, FGF13-OE, LPS (1 µg/mL, 24 h), and LPS-treated FGF13-OE astrocyte groups (N), and statistical analysis of apoptosis rates in each group (O). n = 3 per group. P. Representative immunoblots of Bax, Bcl-2, and β-actin in cells from the control, FGF13-OE, LPS, and LPS+FGF13-OE groups (left), and quantitative analysis of Bax and Bcl-2 expression changes in each group (right). n = 3 per group. Data are presented as mean ± SEM. *P < 0.05, **P < 0.01

Expression of FGF13 affects the apoptosis level of astrocytes

To further investigate the correlation between FGF13 downregulation and astrocyte loss, we used LPS (lipopolysaccharide, a widely accepted neuroinflammatory agent) to induce an inflammatory response in astrocytes. We first established an FGF13-knockdown (FGF13-KD) astrocyte cell line via lentiviral short hairpin RNA (shRNA) interference (Fig. 2G and H). To explore whether FGF13 deficiency affects astrocyte apoptosis, we performed flow cytometry. The results showed that, compared with the control group, both the FGF13-KD group and the LPS group exhibited significant astrocyte apoptosis (Fig. 2I and J). Further analysis revealed that, compared with the control group, the expression of Bax was significantly increased in astrocytes in the FGF13-KD and LPS groups (Fig. 2K), while the expression of the Bcl-2 was significantly decreased (Fig. 2K). These results indicate that FGF13 deficiency significantly induces astrocyte apoptosis.

To further demonstrate that FGF13 is involved in regulating astrocyte apoptosis, we engineered an astrocyte cell line overexpressing FGF13 (FGF13-OE) (Fig. 2L and M). Flow cytometry was used to assess astrocyte apoptosis levels. While baseline apoptosis was unaffected by FGF13 overexpression, FGF13-OE robustly protected astrocytes against LPS-induced apoptosis (Fig. 2N and O). Molecularly, FGF13 overexpression suppressed the LPS-induced upregulation of Bax and restored Bcl-2 levels (Fig. 2P). These in vitro results confirm that FGF13 deficiency is sufficient to trigger astrocytic apoptosis, whereas FGF13 enrichment exerts a profound anti-apoptotic effect under inflammatory stress.

Conditional knockout of FGF13 in the hippocampus exacerbates depression-like behavior and inflammatory damage in mice

We found that decreased FGF13 expression levels in the hippocampus of depressed mice were accompanied by astrocyte loss. The hippocampus is an important brain region involved in the pathogenesis of depression. Therefore, to validate the in vivo relevance of astrocytic FGF13 in depression, we specifically inhibited FGF13 expression in astrocytes by injecting an astrocyte-specific AAV-GfaABC1D-mCherry-T2A-Cre virus into the hippocampus of FGF13-Flox mice (Figs. 3A). Four weeks post-injection, robust reporter fluorescence confirmed viral transduction (Fig. 3B), and Western blotting verified specific FGF13 depletion (Fig. 3C), designating these as FGF13-KD mice. To evaluate the effect of FGF13 on depression-like behavior in mice, we established both CRS and CSDS depression models. Behavioral profiling demonstrated that naïve FGF13-KD mice displayed basal depression-like behaviors, showing increased immobility in the FST (Fig. 3D, S2 D) and TST (Fig. 3E, S2 E), and reduced sucrose preference (Fig. 3F, S2F) relative to wild-type (WT) controls. However, the severity of these baseline deficits was milder than that observed in WT mice subjected to full CRS or CSDS paradigms. Furthermore, exposing FGF13-KD mice to CRS or CSDS did not statistically exacerbate their behavioral metrics beyond those of stressed WT mice, likely due to a behavioral ceiling effect. Furthermore, to rigorously evaluate the effects of astrocyte-specific FGF13 manipulation on CRS induced phenotypes, we also validated the specificity and efficacy of our viral approaches. For FGF13 knockdown experiments, we used pAAV-GfaABBC1D-mCherry-T2A-WPRE (AAV-Ctrl1) as the control vector, which expresses mCherry under the GFAP promoter. For FGF13 overexpression experiments, we used pAAV-CMV-DIO-3-flag-T2A-EGFP (AAV-Ctrl2) as the control vector. As shown in Figures S2 A–S2C, AAV injection alone did not significantly affect immobility time in the FST, TST and SPT in either non-stressed or CRS-exposed mice, confirming that the viral delivery procedure does not confound behavioral outcomes.

Fig. 3.

FGF13 deficiency induces astrocyte damage, leading to depression-like behavior and exacerbated inflammatory responses in mice. A. Schematic diagram of conditional FGF13 knockout. Astrocyte-specific AAV virus was injected into the hippocampus of FGF13-Flox mice. Fluorescence expression image after AAV injection (mCherry, red), FGF13 (cyan), Astrocyte (GFAP, Green), DAPI (blue). Scale bar = 1 mm (middle), scale bar = 300 μm (right). HPC, hippocampus. B. Schematic diagram of the procedure for generating conditional FGF13 knockout mice (FGF13-KD) and the depression model induction process. C. Detection of FGF13 expression in the hippocampus of conditional FGF13 knockout mice (FGF13-KD) and quantitative analysis of FGF13 expression. D–F. Behavioral tests in each group after CRS depression model induction, including performance in the forced swimming test (FST, D), tail suspension test (TST, E), and sucrose preference test (SPT, F). n = 8–12 mice per group. G-H. Representative immunoblots of GFAP, S100β, Bax, Bcl-2, and β-actin in hippocampal (HPC) tissues from each group after CRS modeling (G), and quantitative analysis of GFAP, S100β, Bax and Bcl-2 expression changes in each group (H). n = 3 mice per group. I. Quantification of the percentage of cleaved caspase-3⁺/GFAP⁺ cells among total GFAP⁺ astrocytes in picture J. J. Representative confocal images of hippocampal sections stained with AAV (mCherry, red), GFAP (green), cleaved caspase-3 (CCP, magenta), and DAPI (blue). White arrows indicate double-positive cells (GFAP⁺/cleaved caspase-3⁺). Scale bar = 300 μm. K-N. ELISA measurements of inflammatory factor expression levels in serum from each group after CRS modeling, including IL-6 (K), TNF-α (L), IL-18 (M), and IL-1β (N). n = 8–10 mice per group. Data are presented as mean ± SEM. *P < 0.05, **P < 0.01

To further evaluate the effect of FGF13 deficiency on the pathology of depression, we examined the expression levels of astrocyte markers in the hippocampus. The results showed that compared with normal WT mice, both FGF13-KD mice and CRS depressed mice exhibited significantly decreased expression of the astrocyte marker proteins GFAP and S100β (Fig. 3G and H). In addition, compared with normal mice, both FGF13-KD mice and CRS model mice showed significantly increased expression of the apoptosis-related protein Bax in the hippocampus (Fig. 3G and H), while the expression of the anti-apoptotic protein Bcl-2 was significantly decreased (Fig. 3G and H). Notably, compared with CRS model mice, FGF13-KD mice subjected to CRS showed a further significant decrease in Bcl-2 expression in the hippocampus (Fig. 3G and H). To directly determine whether FGF13 downregulation leads to astrocyte apoptosis in vivo, we performed double-immunofluorescence staining for cleaved caspase-3 (a reliable apoptotic marker) and GFAP (an astrocyte-specific marker) on hippocampal sections from each experimental group. As shown in Fig. 3J, CRS exposure significantly increased the proportion of cleaved caspase-3⁺/GFAP⁺ cells relative to total GFAP⁺ astrocytes in the hippocampal region compared to non-stressed controls, providing direct evidence that chronic stress triggers astrocyte apoptosis in vivo. Importantly, astrocyte-specific FGF13 knockdown further exacerbated this effect, with a significantly higher percentage of cleaved caspase-3⁺/GFAP⁺ astrocytes observed in CRS+FGF13-KD mice compared to CRS controls (Fig. 3I). These results indicate that FGF13 deficiency is an important factor contributing to astrocyte apoptosis in the hippocampus. Furthermore, we examined the effect of FGF13 deficiency on the inflammatory response. The results showed that compared with normal mice, FGF13-KD mice had significantly increased serum levels of the inflammatory cytokines IL-6 (Fig. 3K), TNF-α (Fig. 3L), IL-18 (Fig. 3M), and IL-1β (Fig. 3N); similarly, CRS model mice also showed significantly higher levels of these inflammatory cytokines compared with the normal control group (Fig. 3K and N). Notably, FGF13-KD mice subjected to CRS showed markedly increased serum inflammatory cytokine levels, with TNF-α and IL-1β levels being significantly higher than those in the CRS model group (Fig. 3L and N). Therefore, we speculate that astrocyte apoptosis induced by FGF13 deficiency may be an important factor exacerbating the inflammatory response in depressed mice.

Overexpression of FGF13 in the hippocampus effectively ameliorates depression-like behavior and inflammatory levels in mice

To further confirm the important role of FGF13 in depression, we injected an AAV-CMV-DIO-FGF13-3xflag-T2A-EGFP virus overexpressing FGF13 into the hippocampus of GFAP-Cre mice, which enabled specific expression of FGF13 in astrocytes (Figs. 4A). Four weeks after injection, abundant green fluorescence was observed in the hippocampus (Fig. 4B), and FGF13 protein expression levels were found to be significantly increased (Fig. 4C), indicating successful FGF13 overexpression; these mice were designated FGF13-OE mice. To evaluate the effect of FGF13 on depression-like behavior in mice, we established both CRS and CSDS depression models. While naïve FGF13-OE mice exhibited baseline behaviors comparable to WT controls, they demonstrated profound resilience to stress. In both the CRS and CSDS models, FGF13-OE mice demonstrated significantly reduced immobility times in the FST (Figs. 4D, S2G) and TST (Figs. 4E, S2H) and maintained normal sucrose preference (Figs. 4F, S2I) compared to their stressed WT counterparts. These results indicate that overexpression of FGF13 effectively reverses depression-like behavior in mice.

Fig. 4.

Overexpression of FGF13 protects astrocytes and ameliorates depression-like behavior and inflammatory responses in mice. A. Schematic diagram of astrocyte-specific FGF13 overexpression. AAV-DIO virus specifically recognizing astrocytes was injected into the hippocampus of GFAP-Cre mice. Fluorescence expression image after AAV injection (EGFP, green), DAPI (Blue), Astrocyte (GFAP, red), FGF13 (cyan). Scale bar = 1 mm (middle), scale bar = 200 μm (right). HPC, hippocampus. B. Schematic diagram of the procedure for generating conditional FGF13 overexpression mice (FGF13-OE) and the depression model induction process. C. Detection of FGF13 expression in the hippocampus of FGF13 overexpression mice (FGF13-OE) and quantitative analysis of FGF13 expression. D–F. Behavioral tests in each group after CRS depression model induction, including performance in the FST (D), TST (E), and SPT (F). n = 8–12 mice per group. G-H. Representative immunoblots of GFAP, S100β, Bax, Bcl-2, and β-actin in hippocampal (HPC) tissues from each group after CRS modeling (G), and quantitative analysis of GFAP, S100β, Bax and Bcl-2 expression changes in each group (H). n = 3 mice per group. I. Quantification of the percentage of cleaved caspase-3⁺/GFAP⁺ cells among total GFAP⁺ astrocytes in picture J. J. Representative confocal images of hippocampal sections stained with AAV (mCherry, red), GFAP (green), cleaved caspase-3 (CCP, magenta), and DAPI (blue). White arrows indicate double-positive cells (GFAP⁺/cleaved caspase-3⁺). Scale bar = 300 μm. K–N. ELISA measurements of inflammatory factor expression levels in serum from each group after CRS modeling, including IL-6 (K), TNF-α (L), IL-18 (M), and IL-1β (N). n = 8–10 mice per group. Data are presented as mean ± SEM. *P < 0.05, **P < 0.01

In addition, to exclude the possibility that altered locomotor activity contributed to the behavioral differences observed in the FST and TST, we performed the open field test (OFT). As shown in Figures S3A and S3B, there were no significant differences in total distance traveled or average velocity among any of the experimental groups, including CRS-induced depressive mice, FGF13-KD mice, and FGF13-OE mice, compared to their respective controls. These results confirm that the increased immobility in FST and TST reflects genuine depression-like behavior despair rather than nonspecific motor deficits. Interestingly, we observed that CRS-induced depressive mice and FGF13-KD mice spent significantly less time in the center zone of the open field compared to controls, while FGF13 overexpression rescued this phenotype (Figures S3A, S3C). Given that reduced center zone time is widely interpreted as increased anxiety-like behavior, these data suggest that CRS-induced depressive mice exhibit elevated anxiety levels, which is consistent with previous findings by Dang R et al. [40] and the well-documented clinical comorbidity between depression and anxiety [41]. Moreover, these results imply that FGF13 may exert a protective effect not only against depression-like behavior but also against anxiety-like phenotypes.

To further evaluate the effect of FGF13 on the pathology of depression in mice, we examined the expression levels of astrocyte markers in the hippocampus. The results showed that the expression of the astrocyte marker proteins GFAP and S100β was significantly decreased in CRS model mice, and FGF13 overexpression significantly increased the expression levels of GFAP and S100β (Fig. 4G and H). In addition, compared with CRS model mice, FGF13-OE mice showed significantly decreased expression of the apoptosis-related protein Bax in the hippocampus (Fig. 4G and H), while the expression of the anti-apoptotic protein Bcl-2 was significantly increased (Fig. 4G and H). Moreover, astrocyte-specific FGF13 overexpression markedly reduced the proportion of cleaved caspase-3⁺/GFAP⁺ astrocytes in CRS-treated mice, rescuing the apoptotic phenotype to levels comparable to non-stressed controls (Fig. 4J). These results indicate that overexpression of FGF13 effectively inhibits astrocyte apoptosis in the hippocampus. Furthermore, we examined the effect of FGF13 overexpression on the inflammatory response. The results showed that compared with CRS mice, FGF13-OE mice had significantly decreased serum levels of the inflammatory cytokines IL-6 (Fig. 4K), TNF-α (Fig. 4L), IL-18 (Fig. 4M), and IL-1β (Fig. 4N). These findings demonstrate that overexpression of FGF13 in hippocampal astrocytes effectively ameliorates CRS- and CSDS-induced depression, protects astrocytes, and reduces neuroinflammation.

FGF13 targets and binds to the JIP2 protein to affect astrocyte apoptosis

To further investigate the mechanism of action of FGF13, and based on the finding that FGF13 overexpression effectively inhibits LPS-induced astrocyte apoptosis, we performed RNA-seq on LPS-treated astrocytes and FGF13-overexpressing astrocytes. A total of 728 significantly differentially expressed genes were detected, among which 627 genes were significantly upregulated and 101 genes were significantly downregulated in the FGF13-OE group (Fig. 5A). GO enrichment analysis primarily identified enrichment in apoptotic processes (Fig. 5B), such as “negative regulation of apoptotic process”, “DNA damage response”, “growth factor activity”, and “positive regulation of ERK1 and ERK2 cascade”; additionally, enrichment was found in inflammation-related processes, such as “response to transforming growth factor beta” and “cytokine activity”. Further KEGG analysis revealed enrichment in signaling pathways related to cell proliferation and inflammatory responses (Fig. 5C), including the “cell cycle”, “PI3K-Akt signaling pathway”, “cytokine-cytokine receptor interaction”, “TGF-beta signaling pathway”, and “NF-κB signaling pathway” (Fig. 5C). Moreover, we further analyzed the main pathways of apoptosis and inflammatory responses using GSEA (Fig. 5D) and found that FGF13 overexpression was strongly correlated with these pathways.

Fig. 5.

FGF13 targets and binds JIP2 to regulate astrocyte apoptosis. A. Astrocyte C6 cells and FGF13-overexpressing astrocytes were treated with LPS (1 µg/mL) for 24 h, and differentially expressed genes were analyzed by RNA-seq. MA plot showing changes in gene expression, with red dots indicating significantly upregulated genes, blue dots indicating significantly downregulated genes, and gray dots indicating genes with no significant difference between groups. B. GO enrichment analysis of differential gene functions between the control and FGF13-OE groups. C. KEGG enrichment analysis of differential gene functions between the control and FGF13-OE groups. D. Gene set enrichment analysis (GSEA) of signaling pathways between the control and FGF13-OE groups. GSEA plots show the enrichment of all identified genes in the Apoptosis, NF-κB activation pathway, Cytokine-cytokine receptor interaction, and Inflammatory response. Each vertical line on the x-axis represents a gene. E. Molecular docking simulation analysis of the interaction binding sites between FGF13 and JIP2. F. Co-IP detection of the interaction between FGF13 and JIP2. G. Expression distribution of MAPK8IP2 (encoding JIP2) in brain tissues of MDD patients from a published snRNA-seq dataset (GSE144136; 17 healthy controls and 17 MDD cases). H–K. Expression changes of MAPK8IP2 in different cell types, including excitatory neurons (H), inhibitory neurons (I), astrocytes (J), and microglia (K). Data are presented as mean ± SEM. *P < 0.05, **P < 0.01

In addition, we predicted FGF13-targeted inflammation- and cell proliferation-related targets using the STRING database. The results identified the candidate protein JIP2 (gene: MAPK8IP2), which is a scaffold protein that modulates MAPK signal cascades. Molecular docking simulation between FGF13 and JIP2 was performed using AutoDock, and the results showed a docking score of − 248.25 and a confidence score of 0.8771, indicating a good binding ability between FGF13 and JIP2 (Fig. 5E). To further confirm the targeted binding ability between FGF13 and JIP2, we performed co-immunoprecipitation (Co-IP) and found that FGF13 could bind to JIP2 (Fig. 5F). Furthermore, by analyzing a single-cell dataset from depressed patients (GSE144136), we found that the gene encoding JIP2 (MAPK8IP2) was highly expressed in neurons and was also expressed in astrocytes (Fig. 5G). Further analysis revealed that MAPK8IP2 expression was significantly increased in the brain tissue of depressed patients in excitatory neurons (Fig. 5H), inhibitory neurons (Fig. 5I), and astrocytes (Fig. 5J), but showed no significant change in microglia. Notably, JIP2 (MAPK8IP2) is a member of the JNK-interacting protein (JIP) family and is involved in regulating the JNK signaling pathway [42]. The signal transduction pathways involving JNK kinases are known to participate in regulating the pathological process of depression; for example, inflammatory factors (TNF-α, IL-1β, IL-6, etc.), oxidative stress, neuroexcitotoxicity, and glucocorticoids can all modulate JNK activity [43]. Therefore, we hypothesize that FGF13 regulates the downstream JNK signaling pathway by targeting the JIP2 protein, thereby affecting the pathological progression of depression.

FGF13 targets and regulates the JIP2-JNK signaling pathway to affect astrocyte apoptosis

To further confirm the above hypothesis, we first examined the effect of FGF13 deficiency on the JIP2-JNK signaling pathway. The results showed that, compared with normal WT mice, JIP2 expression was significantly increased in the hippocampus of FGF13-KD and CRS mice (Fig. 6A), and JIP2 expression levels were further increased in FGF13-KD mice after CRS (Fig. 6A). In addition, JIP2 protein expression levels in the hippocampus of FGF13-OE mice were similar to those in WT mice (Fig. 6B). Conversely, FGF13 overexpression completely blunted the CRS-induced upregulation of JIP2 (Fig. 6B). These results indicate that FGF13 deficiency leads to abnormally increased JIP2 protein expression, whereas FGF13 overexpression effectively inhibits JIP2 protein expression in the hippocampus of CRS model mice. Furthermore, examination of downstream JIP2 signaling revealed that, compared with WT mice, both FGF13-KD and CRS mice exhibited significantly increased phosphorylation levels of JNK protein (p-JNK/JNK) in the hippocampus, with similar phosphorylation levels between the two groups (Fig. 6A); the phosphorylation levels of c-Jun protein (p-c-Jun/c-Jun) were also increased (Fig. 6A). In contrast, FGF13-OE mice exhibited a near-total normalization of JNK and c-Jun phosphorylation despite CRS exposure (Fig. 6B). These results indicate that FGF13 deficiency leads to aberrant activation of the downstream JNK signaling pathway of JIP2, whereas FGF13 overexpression effectively inhibits the downstream JNK signaling pathway of JIP2.

Fig. 6.

FGF13 regulates astrocyte apoptosis through the JIP2-JNK signaling pathway. A. Representative immunoblots of JIP2, JNK, p-JNK, c-Jun, p-c-Jun, and GAPDH in hippocampal tissues from the control, FGF13-KD, CRS, and CRS+FGF13-KD groups (left), and quantitative analysis of JIP2, p-JNK/JNK, and p-c-Jun/c-Jun expression changes in each group (right). n = 3 mice per group. B. Representative immunoblots of JIP2, JNK, p-JNK, c-Jun, p-c-Jun, and GAPDH in hippocampal tissues from the control, FGF13-OE, CRS, and CRS+FGF13-OE groups (left), and quantitative analysis of JIP2, p-JNK/JNK, and p-c-Jun/c-Jun expression changes in each group (right). n = 3 mice per group. C. CCK-8 assay detecting the effect of LPS stimulation on astrocyte viability in each group (Groups A–F received different treatments as indicated below). LPS (1 µg/mL, 24 h), JNK inhibitor SP600125 (20 µM, 24 h). D–E. Flow cytometry detection of apoptosis levels in cells from Groups A–F (D), with FITC fluorescence intensity on the x-axis and PE fluorescence intensity on the y-axis. Statistical analysis of apoptosis levels (E). F. Schematic diagram of the FGF13-JIP2-JNK signaling pathway. FGF13 inhibits JIP2 protein expression and further suppresses downstream JNK kinase activation, ultimately affecting apoptosis-related proteins Bax/Bcl-2 and the release of inflammatory factors. G–H. Effects of FGF13 treatment on the JIP2-JNK signaling pathway in astrocytes from Groups A–F, including representative immunoblots of JIP2, JNK, p-JNK, c-Jun, p-c-Jun, Bax, Bcl-2, and GAPDH (G), and quantitative analysis of JIP2, p-JNK/JNK, p-c-Jun/c-Jun, Bax, and Bcl-2 expression changes in each group (H). n = 4 per group. I. Schematic diagram of astrocyte-neuron co-culture assay (left). ELISA detection of glutamate content changes in the culture medium from Groups A–F (middle). ELISA detection of neurotrophic factor AVP content changes in the culture medium from Groups A–F (right). J. Schematic diagram of astrocyte-microglia co-culture assay (left). ELISA detection of TNF-α content changes in the culture medium from Groups A–F (middle). ELISA detection of IL-6 content changes in the culture medium from Groups A–F (right). Data are presented as mean ± SEM. *P < 0.05, **P < 0.01

To further investigate how FGF13 participates in regulating astrocyte apoptosis through the JIP2-JNK signaling pathway, we used the CCK-8 assay and flow cytometry to assess cell viability and apoptosis levels of astrocytes under various conditions. The results showed that LPS effectively inhibited astrocyte viability (Fig. 6C) and induced apoptosis (Fig. 6D and E); FGF13 knockout further suppressed astrocyte viability (Fig. 6C) and led to more severe apoptosis (Fig. 6D and E). However, FGF13 overexpression effectively reversed the LPS-induced decrease in astrocyte viability (Fig. 6C) and effectively reduced apoptosis (Fig. 6D and E). Furthermore, treatment with the JNK inhibitor SP600125 also effectively reversed the LPS-induced decrease in astrocyte viability (Fig. 6C) and inhibited LPS-induced apoptosis (Fig. 6D and E). Aberrant activation of the JIP2-JNK signaling pathway induces downstream Bax/Bcl-2 signaling, exacerbating apoptosis, and also promotes the expression of inflammatory cytokines (IL-6 and TNF-α) through activation of c-Jun (Fig. 6F). We further examined the role of FGF13 in an LPS-stimulated astrocyte cell line model and the changes in protein expression of the JIP2-JNK signaling pathway. The results showed that, consistent with the findings in the hippocampus of depressed model mice, FGF13 deficiency led to a significant increase in JIP2 protein expression and significantly increased the phosphorylation levels of JNK and c-Jun proteins (Fig. 6G and H). In contrast, FGF13 overexpression significantly inhibited JIP2 protein expression, and both FGF13 overexpression and SP600125 significantly inhibited the phosphorylation of JNK and c-Jun proteins (Fig. 6G and H). In addition, compared with the control group, FGF13 deficiency led to a significant increase in the apoptosis-related protein Bax and a significant decrease in the anti-apoptotic protein Bcl-2 in astrocytes, whereas both FGF13 overexpression and SP600125 significantly inhibited Bax expression and significantly increased Bcl-2 expression (Fig. 6G and H).

To validate our findings in a more physiologically relevant system, we extended our analysis to primary mouse hippocampal astrocytes (Figure S4A). Consistent with the results obtained in C6 cells, LPS stimulation significantly activated the JIP2-JNK signaling pathway in primary astrocytes, as evidenced by increased phosphorylation of JNK and its downstream targets (Figure S4B, S4C). Knockdown of FGF13 further potentiated LPS-induced JIP2-JNK activation, while FGF13 overexpression markedly reversed this effect, mirroring the inhibitory action of the JNK inhibitor SP600125 (Figure S4B, S4C). We next examined whether FGF13 modulates LPS-induced apoptosis in primary astrocytes. As shown in Figures S4B and S4C, FGF13 knockdown exacerbated LPS-induced apoptotic changes, with a significant increase in the pro-apoptotic protein Bax and a decrease in the anti-apoptotic protein Bcl-2. Conversely, FGF13 overexpression effectively reversed these apoptotic markers, reducing Bax and restoring Bcl-2 levels to near-control values (Figure S4B, S4C). These results indicate that FGF13 participates in regulating LPS-induced astrocyte apoptosis through the JIP2-JNK signaling pathway.

Furthermore, astrocytes play an important role in the central nervous system. Therefore, we further examined whether FGF13 affects nervous system function by regulating astrocytes through in vitro experiments. We established astrocyte-neuron co-culture systems (Fig. 6I) and astrocyte-microglia co-culture systems (Fig. 6J). The results showed that both FGF13 deficiency and LPS stimulation led to a significant increase in glutamate levels in the culture medium (Fig. 6I); however, both FGF13 overexpression and SP600125 effectively inhibited glutamate secretion (Fig. 6I). In addition, both FGF13 deficiency and LPS stimulation led to a significant decrease in arginine vasopressin (AVP) levels in the culture medium (Fig. 6I), whereas both FGF13 overexpression and SP600125 effectively promoted AVP secretion (Fig. 6I). Moreover, in the astrocyte-microglia co-culture system, both FGF13 deficiency and LPS stimulation led to significantly increased expression of pro-inflammatory cytokines (TNF-α and IL-6), whereas both FGF13 overexpression and SP600125 significantly inhibited the release of pro-inflammatory cytokines (TNF-α and IL-6) (Fig. 6J). These results indicate that FGF13 maintains normal astrocyte function and exerts neuroprotective effects through the JIP2-JNK signaling pathway.

Discussion

Rapid socioeconomic development has inadvertently escalated chronic psychological stress, contributing to a rising global incidence of MDD [44]. Clinical and preclinical studies consistently reveal hippocampal atrophy in MDD, accompanied by diminished GFAP expression and widespread astrocyte loss [1, 14]. As the most abundant glial cells in the central nervous system, astrocytes are indispensable for neurotransmitter cycling, metabolic support, and the modulation of synaptic plasticity, learning, and emotion [13, 22]. Despite their critical functions, astrocytes are highly vulnerable to stress. While it is widely accepted that astrocytic dysfunction precipitates depression-like behaviors, the precise sequence of pathological changes leading to astrocytic apoptosis and its definitive impact on MDD pathogenesis remain debated [18, 21]. Our study provides additional insights into the role of astrocytes in the pathogenesis of depression (Fig. 7). In this study, we found that in mouse depression models established by CRS and CSDS, astrocyte apoptosis occurred in the hippocampus and was accompanied by decreased FGF13 expression. FGF13 knockout led to astrocyte apoptosis, induced inflammatory responses, and caused depression-like behavior in mice (Fig. 7). However, astrocyte-specific overexpression of FGF13 effectively inhibited CRS- and CSDS-induced astrocyte damage and significantly alleviated depression-like behavior in mice (Fig. 7). Furthermore, we further elucidated the mechanism of astrocyte loss in depression and concluded that FGF13 plays a key role in mediating astrocyte apoptosis. These findings identify FGF13 in astrocytes as an important potential target for the treatment of depression.

Fig. 7.

Schematic diagram of the molecular mechanism of FGF13 action. Depression models were established using CSDS or CRS. FGF13 deficiency leads to astrocyte apoptosis in the hippocampus, which exacerbates the progression of depression. However, specific overexpression of FGF13 in astrocytes effectively ameliorates astrocyte apoptosis and depression-like behavior in mice. Further mechanistic analysis revealed that FGF13 targets and inhibits the JIP2 protein, which in turn inhibits the phosphorylation of the downstream signaling protein JNK, ultimately exerting anti-apoptotic and anti-inflammatory effects

The role of FGF13 in depression has not been previously reported. Previous studies have found that mutations in FGF13 often lead to neurological disorders [33, 45]; for example, FGF13 has been identified as a candidate gene for X-linked intellectual disability in human genetic studies [45]. In addition, FGF13 deficiency leads to delayed neuronal migration and impaired learning and memory function in mice [46]. FGF13 also affects motor behavioral ability; for instance, FGF13 deficiency in the cerebral cortex leads to increased spontaneous locomotion, whereas FGF13 deficiency in neural stem cells (NSCs) in the hippocampus leads to hyperactivity and anxiety-like behavior [47]. Moreover, FGF13 protects neurons from degeneration in a mouse model of Parkinson’s disease by inhibiting the transfer of stressed mitochondria from neurons to adjacent glial cells in the substantia nigra pars compacta (SN), thereby preventing uncontrollable neuroinflammation and ultimately improving PD symptoms [33]. These findings highlight the cell-specific and region-specific roles of FGF13 in brain function. In our study, the selective downregulation of FGF13 in the astrocytes of MDD patients and CRS or CSDS mice led us to hypothesize that FGF13 acts as an endogenous survival factor for glia during depressive stress, and that its deficiency directly triggers astrocytic apoptosis. Although our snRNA-seq data revealed that FGF13 is specifically downregulated in astrocytes, this cell-type–specific reduction could not be reliably resolved by conventional immunostaining of brain sections, given the extensive overlap of FGF13 signals across diverse hippocampal cell types and its modest expression levels in astrocytes. Nevertheless, using genetic manipulation to specifically regulate FGF13 expression in mouse astrocytes, we demonstrated a correlation between FGF13 function and depressive‑like behaviors. However, the current study does not definitively establish that the downregulation of FGF13 in depression is cell‑autonomously driven by astrocytes, and the precise cellular source of FGF13 dysregulation warrants further investigation in future studies.

To test the above hypothesis, we further constructed FGF13-knockout and FGF13-overexpressing astrocyte cell lines and found that FGF13 deficiency led to astrocyte apoptosis and exacerbated LPS-induced astrocyte apoptosis. However, overexpression of FGF13 effectively reversed LPS-induced astrocyte apoptosis. These findings further confirm that FGF13 is involved in regulating astrocyte apoptosis. In addition, we also found that chronic stress increases the number of cleaved caspase-3⁺/GFAP⁺ astrocytes, and that FGF13 loss-of-function exacerbates while FGF13 gain-of-function rescues this apoptotic phenotype in vivo. These results establish that FGF13 is not merely associated with, but directly regulates, astrocyte survival in the hippocampus under pathological conditions. Importantly, the consistency between our direct apoptosis measurements and the earlier indirect markers (GFAP reduction, Bax/Bcl-2 changes) validates the robustness of our findings across multiple experimental approaches. Numerous studies have shown that inhibiting astrocyte apoptosis effectively ameliorates depression-like behavior in mice. For example, Yang L et al. [48]found that activation of the Drd2/β-arrestin2 pathway eliminates inflammation-induced astrocyte loss and improves depression-like behavior in a mouse model of depression. Shiyu S et al. [9] found that IL-6 released by activated microglia exacerbates the development of depression by inducing astrocyte atrophy and apoptosis, and that inhibiting the IL-6 receptor in astrocytes ameliorates astrocyte apoptosis and depression-like behavior in mice. Zhou W et al. [49] found that overexpression of NDRG2 in astrocytes induces apoptosis and depression-like behavior in mice by inhibiting AKT and activating the FoxO3a-Puma apoptotic signaling pathway. In addition, Novakovic M et al. [50] found that loss of function of the calcium channel Orai1 effectively reduces inflammation in the mouse hippocampus, improves astrocyte function, and protects mice from a depression-like behavior phenotype following systemic inflammation. These findings indicate that reducing astrocyte apoptosis effectively ameliorates depression; therefore, we hypothesized that FGF13 protects against astrocyte damage and further ameliorates depression-like behavior in mice.

Our results showed that specific knockout of FGF13 in the mouse hippocampus led to depression-like behavior in mice. Although FGF13-KD mice did not exhibit a statistically significant exacerbation of depression-like behaviors following CRS or CSDS compared to stressed wild-type mice, we attribute this to the ceiling effects inherent in behavioral assays like the FST, TST, and SPT. Notably, specific overexpression of FGF13 in astrocytes significantly reversed depression-like behavior in mice, and FGF13-OE mice did not exhibit depression-like behavior even after CRS or CSDS treatment. These results indicate that FGF13 effectively exerts antidepressant effects. Furthermore, FGF13 deficiency led to extensive astrocyte damage and higher levels of inflammation, whereas FGF13 overexpression effectively inhibited astrocyte apoptosis and the inflammatory response. It is well known that neuroinflammation generally refers to inflammatory responses within the central nervous system caused by various pathological injuries [51]. This process is characterized by the expression of pro-inflammatory cytokines. These cytokines can lead to synaptic dysfunction, neuronal death, and impaired neurogenesis [52]. However, neuroinflammation is often a chronic process that cannot resolve spontaneously, and persistent neuroinflammation can induce depression-like behavior or promote the progression of major depressive disorder [53]. For example, it was found a strong correlation between depression and upregulation of peripheral inflammatory markers such as interleukin-6 (IL-6) and C-reactive protein (CRP) in cerebrospinal fluid samples [54]. Moreover, clinical use of various anti-inflammatory drugs, including nonsteroidal anti-inflammatory drugs (NSAIDs), cytokine inhibitors, statins, and minocycline, significantly improves depressive symptoms in some patients with major depressive disorder [55, 56]. Astrocytes are closely related to the occurrence of neuroinflammation. In addition to secreting pro-inflammatory factors themselves, astrocyte-derived C-C motif chemokine ligand 2 (CCL2) and C-X-C motif chemokine ligand 10 (CXCL10) can induce the extravasation and infiltration of peripheral macrophages, monocytes, and T lymphocytes and promote their subsequent transdifferentiation into a pro-inflammatory phenotype [57, 58]. Therefore, we speculate that in addition to inhibiting astrocyte apoptosis, FGF13 in astrocytes is also involved in the inflammatory response process they induce (Fig. 7). To further explore the mechanism of action of FGF13, we performed RNA-seq analysis, which further confirmed that FGF13 is involved in astrocyte apoptosis and inflammatory responses.

However, an important conceptual question raised by our findings is the directionality of the relationship between FGF13 downregulation and neuroinflammation. While our data demonstrate a strong correlation between reduced FGF13 expression and increased inflammatory responses in mouse models, the causal hierarchy between these two events remains incompletely resolved. On one hand, our astrocyte-specific loss-of-function experiments provide direct evidence that FGF13 downregulation in astrocytes is sufficient to drive inflammatory cytokine production and astrocyte apoptosis, suggesting that FGF13 loss acts as an upstream trigger of inflammation. This is consistent with the established role of FGF13 in maintaining cellular homeostasis and suppressing stress-activated signaling pathways, such as the JIP2-JNK axis. On the other hand, we acknowledge the plausible alternative scenario in which inflammation itself may suppress FGF13 expression. Inflammatory stimuli such as LPS, IL-1β, and TNF-α are known to modulate the expression of various intracellular signaling molecules, and our in vitro data showing LPS-induced FGF13 reduction support this possibility. If inflammation feeds back to further downregulate FGF13, this would establish a vicious cycle FGF13 loss promotes inflammation, and inflammation exacerbates FGF13 reduction—that could amplify astrocyte dysfunction and contribute to the progression of depressive pathology.

Regarding the specific mechanism of action of FGF13, FGF13 is a non-secreted growth factor that is synthesized and functions intracellularly, thus having multiple potential targets. For example, Nan S et al. [33] found that FGF13 binds to the mitochondrial outer membrane protein MCHT2 to maintain mitochondrial function and ensure a stable energy supply for astrocytes. Qiao Y et al. [47] found that FGF13 participates in regulating neuronal differentiation and NSC self-renewal by binding to ARID1B (a unit of the Brahma-associated factor chromatin remodeling complex) in neural stem cells, thereby maintaining hippocampal neurogenesis. We found that FGF13 targets and binds to JIP2 (MAPK8IP2, a scaffold protein that modulates MAPK signal cascades) and further regulates the JIP2-JNK signaling pathway. JNK kinase (c-Jun N-terminal kinase) is a member of the stress-induced mitogen-activated protein (MAP) kinase family [42]. Studies have found that many regulatory functions of JIP2-JNK are highly relevant to the pathological mechanisms of depression [59]. Many pathological factors of depression, such as neuroinflammation, neuroexcitotoxicity, oxidative stress, cerebral ischemia, and psychological stress, can promote abnormally elevated JNK activity. For example, H Mohammad et al. [60] found that JNK knockout effectively improves hippocampal neuronal function and reduces depression-like behavior in mice. Hai Z et al. [61] found that aluminum trichloride caused hippocampal neural cell death and subsequent depression-like behavior in rats via activation of the IL-1β/JNK signaling pathway. We detected significant activation of the JIP2-JNK signaling pathway in the hippocampus of CRS-induced depressed mice, and we also detected aberrant activation of the JIP2-JNK signaling pathway in the hippocampus of FGF13-knockout mice. However, overexpression of FGF13 significantly inhibited the JIP2-JNK signaling pathway (Fig. 7). In addition, we detected the same results in an astrocyte cell line and primary astrocytes in vitro: FGF13 deficiency led to aberrant activation of the JIP2-JNK signaling pathway in astrocytes and induced astrocyte apoptosis, whereas both FGF13 overexpression and the JNK inhibitor SP600125 effectively inhibited the JIP2-JNK signaling pathway and reduced LPS-induced apoptosis. Notably, the regulation of the JIP2-JNK pathway and apoptotic markers by FGF13 in primary astrocytes was highly consistent with our C6 cell data. These results indicate that FGF13 reduces astrocyte apoptosis by inhibiting the JIP2-JNK signaling pathway and ultimately ameliorates depression-like behavior in mice.

In addition, it is important to note that the JIP2-JNK pathway may not represent the exclusive mechanism through which FGF13 exerts its anti-apoptotic effects on astrocytes. As an intracellular FGF, FGF13 lacks a signal peptide and functions in a non-canonical, autocrine/intracrine manner [62]. Emerging evidence has linked FGF13 to multiple signaling networks beyond the JNK cascade. For example, in A549 lung cancer cells, FGF13 activates the AKT-GSK3 pathway to inhibit p21 and p27 activity, thereby promoting cell proliferation [63]. In neural stem cells, FGF13 physically associates with ARID1B, a component of the BAF chromatin remodeling complex, to regulate hippocampal neurogenesis [47]. Additionally, FGF13 has been shown to modulate Parkin-mediated mitophagy and apoptosis in glomerular endothelial cells [64]. Given that these pathways are all intimately involved in cell survival and programmed cell death, it is plausible that FGF13 may concurrently engage one or more of these signaling modules to regulate astrocyte homeostasis under stress conditions. Future studies employing unbiased phosphoproteomic or transcriptomic approaches, as well as pathway-specific pharmacological inhibitors, will be required to systematically dissect the relative contributions of these candidate pathways to FGF13-mediated astrocyte protection and depression-like behavior.

However, our study has several limitations that need to be considered. First, our RNA-seq analysis and signaling pathway detection were performed in the C6 astrocyte cell line, which may differ from the transcriptional profiles of primary astrocytes and in vivo astrocytes. Second, our experiments included only male mice, and it is currently unclear to what extent the regulation of inflammation and depression by astrocytic FGF13 is similar between male and female mice. There are physiological differences between males and females that may lead to sex-based differences in experimental results. Furthermore, our behavioral analyses focused only on depression-like behavior in mice. Because it is difficult to extrapolate behaviors observed in mice to humans, it is not necessarily reasonable to extrapolate our conclusions to broader depressive syndromes such as major depressive disorder, and that our findings require cautious translational interpretation. This limitation underscores the need for broader clinical studies to validate and extend these insights in future research.

Conclusion

In summary, we have elucidated the important molecular mechanism by which FGF13 regulates astrocytes in depression. In depression, the lack of FGF13 in astrocytes leads to cell apoptosis. FGF13 improves astrocyte apoptosis levels by inhibiting the aberrantly activated JIP2-JNK signaling pathway and ultimately ameliorates depression-like behavior in mice. Our findings reveal a critical role for FGF13 in astrocyte-mediated depression pathophysiology, establish FGF13 as a potential therapeutic target, and highlight astrocyte-based strategies for improved depression treatment.

Authors’ contributions

All authors contributed to the study design. RZ, YJ Z, YW Land ZT analyzed the data, the data collection was performed by ZT, HX L, XH C, ZY H, LC S and YW L. The first draft of the manuscript was written by ZG W, YY Z and RZ. YYZ and ZG W helped to revise the manuscript. All authors read and approved the final manuscript.

Funding

This work was supported by Wenzhou Municipal Science and Technology Bureau (Y20240311).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Clinical trial number: not applicable. All animal studies have been approved by Animal Care Ethics Guidelines of Wenzhou Medical University and were approved by the Ethics Committee (XMSQ2025-0259) and were in accordance with the ARRIVE guidelines on the Care and Use of Experimental Animals.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.