Adolescent environmental enrichment prevents early-life stress-induced schizophrenia-like phenotypes via restoration of CREB-BDNF-TrkB signaling

Abstract

Early-life stress (ELS) is a major environmental risk factor for schizophrenia, yet effective preventive strategies targeting its neurodevelopmental consequences remain limited. Here, we show that adolescent environmental enrichment (EE) prevents the emergence of schizophrenia-like behavioral and cognitive deficits induced by maternal separation (MS), a validated ELS model. Mechanistically, MS induced oxidative stress, mitochondrial dysfunction, neuroinflammation, neuronal apoptosis, and synaptic deficits in the hippocampus, accompanied by suppression of the CREB-BDNF-TrkB (cAMP response element-binding protein-brain-derived neurotrophic factor-tropomyosin receptor kinase B) signaling. Adolescent EE robustly reversed these abnormalities and restored neuroplasticity-related signaling. Importantly, we provide evidence supporting a causal role of TrkB signaling in these protective effects within the experimental framework used here: pharmacological activation of TrkB rescued oxidative stress-induced synaptic deficits in vitro, whereas pharmacological inhibition of TrkB abolished the behavioral and cognitive benefits of EE in vivo. To establish translational relevance, we further show that TrkB levels are significantly reduced in neuron-derived exosomes (NDEs) from patients with schizophrenia, with moderate diagnostic performance (AUC = 0.784). Together, these findings support CREB-BDNF-TrkB signaling as a key mechanistic node linking ELS to schizophrenia-related phenotypes and suggest adolescent EE as a developmentally timed, non-pharmacological intervention with translational relevance within the experimental framework used here.

Introduction

Schizophrenia is a severe and disabling psychiatric disorder affecting approximately 1% of the global population and typically emerging in early adulthood [1, 2]. In addition to positive and negative symptoms, persistent cognitive impairments represent a core feature of the disorder and are major determinants of long-term functional outcomes [3]. Notably, these cognitive deficits remain largely resistant to current treatments, underscoring the urgent need for early intervention strategies targeting their underlying neurobiological mechanisms.

Early-life stress (ELS) is a major environmental risk factor for schizophrenia [4]. Converging evidence indicates that ELS induces long-lasting neurobiological alterations, including neuroinflammation, mitochondrial dysfunction, and neuronal apoptosis, particularly within the hippocampus, a key region for cognition and emotional regulation [5, 6]. Maternal separation (MS), a well-established rodent model of ELS, recapitulates core schizophrenia-relevant phenotypes, producing persistent structural and functional impairments in the hippocampus [6–8]. Unlike pharmacological or genetic models, MS captures the developmental stress dimension of schizophrenia risk by simulating real-world early-life adversity. It is developmentally specific and introduces no confounding direct molecular targeting, consistent with human evidence linking childhood adversity to psychosis [9, 10]. Thus, MS provides a clean, translational model for studying stress-induced vulnerability and preventive interventions. However, effective strategies to prevent or reverse these ELS-induced neurodevelopmental abnormalities remain limited.

Environmental enrichment (EE) has emerged as a promising non-pharmacological approach to enhance neuroplasticity and behavioral resilience in neurodevelopmental and psychiatric disorders [11, 12]. While EE has shown beneficial effects in several experimental models, prior studies have largely focused on behavioral outcomes, with limited mechanistic insight. In particular, the causal involvement of key neuroplasticity-related signaling pathways—such as the CREB-BDNF-TrkB (cAMP response element-binding protein-brain-derived neurotrophic factor-tropomyosin receptor kinase B) axis—remains poorly defined. Nevertheless, whether EE can prevent schizophrenia-relevant abnormalities induced by ELS, and whether its beneficial effects are confined to specific developmental windows, remain open questions.

Adolescence represents a critical neurodevelopmental window characterized by ongoing synaptic pruning, dendritic remodeling, and circuit refinement, especially within the hippocampus [13]. This heightened plasticity renders the adolescent brain particularly sensitive to environmental modulation [14], suggesting a potential window for preventive intervention. Notably, both clinical and preclinical studies indicate that schizophrenia-related phenotypes often emerge in early adulthood following earlier neurodevelopmental perturbations [15]. Consistent with this trajectory, we previously demonstrated that MS-induced behavioral abnormalities manifest in adulthood but not during adolescence [7], highlighting adolescence as a critical period for intervention.

At the molecular level, the CREB-BDNF-TrkB signaling pathway is a central regulator of synaptic plasticity and cognitive function. Reduced BDNF expression has been consistently reported in the hippocampus of patients with schizophrenia [16], while phosphorylated CREB (p-CREB) is essential for activity-dependent BDNF transcription and long-term synaptic plasticity. Importantly, upstream pathological processes induced by ELS may converge on this pathway: mitochondrial reactive oxygen species (ROS) can inhibit CREB activation [17], whereas neuroinflammation suppresses BDNF expression [18]. Consistent with these mechanisms, our previous studies demonstrated increased ROS production and elevated pro-inflammatory cytokines in the hippocampus following MS [19]. However, whether these alterations causally converge to disrupt CREB-BDNF-TrkB signaling—and whether this pathway can be developmentally targeted for intervention—remains unclear. In addition, the translational relevance of this pathway, particularly whether its disruption can be detected using peripheral biomarkers, has not been established.

In the present study, we investigated whether EE during adolescence can prevent the emergence of schizophrenia-related behavioral and neurobiological abnormalities induced by ELS. We systematically assessed cognitive function, mitochondrial integrity, neuroinflammation, neuronal apoptosis, and CREB-BDNF-TrkB signaling in the hippocampus. Combining in vivo and in vitro approaches, we further evaluated the role of TrkB signaling in the neuroprotective effects associated with EE. Finally, to enhance translational relevance, we examined TrkB levels in neuron-derived exosomes (NDEs) isolated from the peripheral blood of patients with schizophrenia. Together, this integrative approach was designed to investigate whether TrkB signaling represents an important mechanistic pathway underlying the beneficial effects of EE on ELS-related neuroplasticity deficits and to explore its potential as a peripheral biomarker candidate for schizophrenia.

Materials and methods

Animal study design

Wistar rats were housed under standard conditions (22 ± 2 °C, 50 ± 10% humidity, 12 h light/dark cycle) with ad libitum access to food and water. Timed pregnant females were individually housed, and the day of birth was defined as postnatal day (PND) 0.

MS was performed once on PND9 for 24 h as previously described [6]. Briefly, dams were removed from their home cages for 24 h, while pups remained undisturbed. Control litters were left undisturbed except for routine cage maintenance. To reduce variability associated with estrous-cycle-related hormonal fluctuations and to maintain consistency with our previous studies using the same MS model, only male offspring were used in this study.

At weaning (PND21), male offspring were randomly assigned to four groups (n = 15 per group): SE-CON, EE-CON, SE-MS, and EE-MS. EE was conducted from PND28 to PND45. EE rats were housed in large cages (60 × 45 × 35 cm) equipped with tunnels, running wheels, ladders, and varied sensory objects, which were rearranged every 4 days. Standard environment (SE) rats were housed in standard cages (38 × 32 × 18 cm).

To investigate the role of TrkB signaling, ANA-12 (0.5 mg/kg, i.p.) was administered every other day during the EE period. For the ANA-12 intervention experiment, a subset of 10 animals per group was randomly selected from the original EE-MS cohort to assess the requirement of TrkB signaling. Control animals received vehicle (10% DMSO in saline). Behavioral tests were conducted in adulthood.

All animal procedures were approved by the Institutional Animal Care and Use Committee of Renmin Hospital of Wuhan University (IACUC no: 20201201) and conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. No animals were excluded from the analysis.

Primary hippocampal neuron culture

Primary hippocampal neurons were isolated from embryonic day 16-18 rats. Tissues were digested with 0.125% trypsin at 37 °C for 10 min and resuspended in DMEM containing 10% FBS to terminate digestion.

Cells were cultured in Neurobasal medium supplemented with 2% B27, 2 mM L-glutamine, and 1% penicillin-streptomycin. Half of the medium was replaced every 3 days. Neurons at day in vitro 7 were used for experiments.

To induce oxidative stress, neurons were exposed to 300 μM H₂O₂ for 3 h, followed by replacement with fresh medium with or without 100 nM 7,8-DHF for an additional 24 h.

Clinical study design

Study participants

This study represents a secondary analysis of serum samples from a previously characterized cohort of 40 schizophrenia patients (SCZ) and 40 healthy controls (HCs) [20]. SCZ diagnosis followed DSM-5 criteria. Demographic and clinical information of the participants are summarized in Supplementary Table 1. Patients were receiving routine clinical treatment, and medication status was not controlled for in this secondary analysis. This study was approved by the Ethics Committee of Renmin Hospital of Wuhan University (WDRY2023-K141), and all participants provided written informed consent.

Isolation of NDEs

Total extracellular vesicles (TEVs) were isolated from 250 μL serum using the ExoQuick kit (System Biosciences, EXOQ5A) according to the manufacturer’s instructions. Briefly, serum was thawed on ice, centrifuged at 3000 × g for 15 min to remove debris, and incubated with ExoQuick reagent overnight at 4 °C. The mixture was centrifuged at 1500 × g for 30 min, and the exosome pellet was resuspended in 200 μL PBS.

NDEs were immunocaptured using a biotinylated anti-human CD171 antibody (13-1719-82; Thermo Fisher Scientific) coupled to streptavidin-conjugated agarose resin (Thermo Fisher, 20349). TEVs were incubated with the antibody-coupled resin for 2 h at 4 °C with gentle rotation. The resin was washed five times with PBS to remove unbound exosomes, and NDEs were eluted with 0.1 M glycine-HCl (pH 2.8) followed by immediate neutralization with 1 M Tris-HCl (pH 8.0).

Characterization of NDEs

Isolated NDEs were characterized by transmission electron microscopy (TEM), nanoparticle tracking analysis (NTA), and Western blotting, with representative results shown in (Fig. 9A–C). TEVs were isolated prior to immunocapture. Following CD171-based immunoprecipitation, the unbound fraction was collected as the TEV supernatant to serve as a negative control. For TEM, 10 μL of NDE suspension was loaded onto formvar-carbon coated grids, negatively stained with 2% uranyl acetate, and imaged on a microscope. NTA was performed using a ZetaView instrument (Particle Metrix) with settings optimized for 50-200 nm particles to determine size distribution and concentration.

Fig. 9. Reduced TrkB levels in NDEs from patients with schizophrenia and its diagnostic potential.

A TEM image of NDEs isolated from a patient with schizophrenia, clearly displaying extracellular vesicle (EV)-like morphology (red arrow). B NTA showing the diameter distribution of isolated EVs (median diameter: 142.6 nm; concentration: 3.2 × 10¹⁰ particles/mL), consistent with typical EV characteristics. C Western blot analysis of TEVs, TEV supernatant, and NDEs from a healthy control. Expression of the EV markers Alix, TSG101, and CD81, the neuronal marker CD171, the astrocyte marker GLAST, and the microglial marker TMEM119 were assessed. D Comparison of the TrkB/CD81 ratio in NDEs between patients with schizophrenia (SCZ) and healthy controls (HC) at baseline. E ROC curve of the TrkB/CD81 ratio in NDEs. AUC: 0.784 (95% CI: 0.682-0.887). Data are presented as mean ± SEM. **p < 0.01 compared with the HC group. n = 40 subjects per group.

Western blotting was performed to confirm NDE purity and exclude contamination. NDE lysates were probed with the following antibodies: exosomal markers TSG101 (CSB-RA860355MA1HU, CUSABIO), CD81 (CSB-RA004960A0HU, CUSABIO), and ALIX (ab275377, Abcam); neuronal marker CD171 (20659-1-AP, Proteintech); microglial marker TMEM119 (PA5-62505, Thermo Fisher Scientific); and astrocyte marker GLAST (PA5-72895, Thermo Fisher Scientific). All antibodies were used at a 1:1000 dilution. Positive expression of exosomal and neuronal markers, along with the absence of TMEM119 and GLAST in the NDE fraction and the lack of CD171 in the TEV supernatant, confirmed the specific enrichment of NDEs.

Quantification of TrkB in NDEs

To quantify TrkB levels in NDEs, an enzyme-linked immunosorbent assay (ELISA) was performed using a commercial TrkB ELISA kit (KTE3120). CD81 levels were measured using a CD81 ELISA kit (CSB-EL004960HU) to normalize for exosome abundance. CD81 is a canonical exosomal tetraspanin marker that is stably expressed on extracellular vesicles; normalization to CD81 corrects for potential variations in exosome yield across samples. The relative TrkB level was calculated as the ratio of TrkB to CD81. All ELISA measurements were performed in duplicate, and the mean values were used for statistical analysis.

Behavioral testing of offspring

The experimental rats were evaluated at the adult stages of development. Sample sizes were determined based on previous publications and our prior experience. For all tests, animals were first acclimated to the behavioral room for 30 min. All behavioral tests were performed between 08:00 and 22:00. The tests were administered in a fixed order: open field test (OFT), Y-maze, novel object recognition (NOR) test, Barnes maze, and finally prepulse inhibition (PPI) test, progressing from less stressful to more cognitively demanding tasks. This sequence was applied consistently across all groups. Each rat was assessed for each behavioral measure. After each trial, the apparatus was cleaned with 75% alcohol. All behavioral analyses were performed by investigators blinded to group allocation.

PPI test

PPI testing was conducted in a sound-attenuated chamber equipped with a small Plexiglas cage mounted on a load-cell platform to digitize the pressure generated by the startled rat (AniLab Scientific Instruments Co., Ltd., China. www.anilab.cn). This test was performed with a broad-band white noise of 70 dB. A high-pass (>4 kHz) white noise was used as the pulse stimulus. Following acclimatization for five minutes, a sequence of five startle stimuli (120 dB of startle stimulus for 20 ms) was applied. During the testing phase, the initial delay was 50 ms followed by a 20 ms impulse stimulus (75, 78, or 82 dB), a 100 ms delay, followed by a 40 ms startle stimulus of 120 dB randomly applied for about 40 trials by software control. The interstimulus interval ranged from 10 to 30 seconds. The percentage PPI for each prepulse intensity was calculated as follows: [1-(startle amplitude on pre-pulse trial)/(startle amplitude on pulse alone)] × 100% [21].

Open-field test (OFT)

The OFT is widely used to evaluate spontaneous locomotor activity [22]. OFT was performed in a 100 ×100 cm square arena with an indirect illumination of 60 lux. Rat movements were video-tracked using Ethovision software 11.5 (Noldus) for 10 min. Total distance traveled, time and distance in center (exclusion of a 5 cm border arena) were directly measured by the software.

Y maze test

The Y-maze test is used to evaluate attention and short-term working memory of rodents [23]. The Y-maze apparatus consists of three black arms arranged at 120-degree angles, each with a length of 50 cm, width of 10 cm, and height of 30 cm. The arms converge at a central region, and the structure is mounted 50 cm above the ground. After placing the rats in the apparatus, their movements were recorded by video and analyzed using EthoVision software. In the autonomous alternation test, each rat was placed at the starting arm and allowed to explore for 8 min. Alternation rate %= number of alternations / (total number -2) %.

Novel object recognition test (NOR)

The NOR test was performed using the open-field apparatus and had three components: habituation, training, and testing [24]. Rats were placed in the center of the open field to adapt to the apparatus for 24 h. During the training experiment, the rats were allowed to explore two identical objects for 10 min. The test experiment started 24 h later. Rats were allowed to explore one familiar and one novel object (replacing a training object) for 10 min, with locomotion tracked by video and analyzed using Panlab SMART v3.0 (Barcelona, Spain). The formula for calculating preference for exploring the novel object was: [(time spent in the novel subject) / (total time spent in two subjects) × 100%].

Barnes maze test

We examined spatial memory with the Barnes maze [25], which was elevated 140 cm above the floor and consisted of 20 holes located evenly on the surface periphery, each 10 cm in diameter. The target box was a hole that connected to a dark chamber, allowing the animal to escape from bright light exposure. The day before the formal experiment, animals were adapted to the target box for four minutes. On the first day, each animal was placed in the central black cube of the maze for five seconds and permitted to explore the maze to find the target box when the cube was removed. Once the animal entered the escape box, it was left there for 30 seconds; if it failed to find the target box within three minutes, it was taken to the target box and allowed to remain in the target box for one minute. Each animal underwent two trials during the day with an interval of four hours. The latency time to reach the target box was recorded. The test was performed over four days. The whole process was monitored by a digital camera and a computer system.

Sample collection

After completion of behavior testing, the animals were anesthetized using sodium thiopental (50 mg/kg, i.p.). Rats were sacrificed, and hippocampal tissues were collected from approximately 15 animals per group, immediately frozen on dry ice, and stored at −80 °C. For Western blotting, Nissl staining, Golgi staining, and electron microscopy, 4-5 animals per group were randomly selected from the behavioral cohort.

Nissl staining

Neuronal survival was assessed with Nissl staining. The brain sections are stained with Nissl staining solution (0.1% cresyl violet) for 10 min, then rinsed with double distilled water for 1 min. Then dehydration of samples with various concentrations of alcohol was performed, sections were transparentized with xylene for 3 min and sealed with neutral resin for observation. Nissl-positive neurons were counted in three sections per animal using ImageJ.

Golgi-Cox staining

Four rats were randomly chosen from each group at the end of the behavior tests. Brains were rapidly removed from deeply anesthetized rats. Golgi-Cox staining was performed according to the manufacturer of the FD Rapid GolgiStain™ Kit (FD Neuro Technologies Inc., USA). Brain tissue sections of 150 µm thickness were prepared along the coronal plane using a vibrating slicing device (Leica VT1200S, Germany) and placed on slides covered with gelatine. Then, the slides were prepared according to the kit protocol. Images were taken using a microscope and processed with NIH ImageJ software. Pyramidal neurons from the hippocampus were subjected to analysis of spines, including density and morphology.

Protein extraction and western blot analysis

Total protein was extracted from hippocampus tissues with protein lysis buffer (Beyotime Biotech, China). The protein concentration was determined with a BCA kit (P0010S, Beyotime Biotech). 40 μg proteins in each sample were separated by 10% SDS-polyacrylamide gel electrophoresis and transferred onto polyvinylidene fluoride (PVDF) membranes. Western blot analysis was carried out using the following primary antibodies raised against target proteins: rabbit anti-BAX (dilution 1:1000, GB11690, Servicebio, China), rabbit anti-Bcl2 (dilution 1:1000, abs147821, Absin, China), rabbit anti-Caspase3 (dilution 1:2000, abs124299, Absin, China), rabbit anti-Cleaved-Caspase3 (dilution 1:2000, abs132005, Absin, China), rabbit anti- synaptophysin (SYN, dilution 1:1000, ab214316, Abcam), rabbit anti-postsynaptic density protein 95 (PSD-95) (dilution 1:1000, ab238135, Abcam), rabbit anti-CREB (dilution 1:1000, 9197, CST Biological Reagents, US), rabbit anti-p-CREB (dilution 1:2000, 9198, CST Biological Reagents, US), rabbit anti-BDNF (dilution 1:1000, 60071, CST Biological Reagents, US), rabbit anti-TrkB (dilution 1:1000, 4603, CST Biological Reagents, US), rabbit anti-p-TrkB (dilution 1:2000, 4619, CST Biological Reagents, US), and rabbit anti-β-actin (dilution 1:1000, ab181602, Abcam) antibodies overnight at 4 °C. The following secondary horseradish peroxidase (HRP)-conjugated antibodies were used at 1: 5000 dilutions: goat anti-rabbit HRP (12-348, Millipore, USA) and goat anti-mouse HRP (sc-2005, Santa Cruz Biotechnology, USA). The blots were visualized with enhanced chemiluminescence detection reagents using a Chemidoc TM Touch Imaging System (Bio-Rad). Protein expression levels were normalized to β-actin and quantified using ImageJ software.

RNA preparation and gene expression analysis of pro-inflammatory cytokine by quantitative real-time polymerase chain reaction (qPCR)

For total RNA isolation and PCR, RNA was isolated with TRIzol™ reagent (Invitrogen, Carlsbad, CA, USA) as recommended by the manufacturer, followed by chloroform extraction and isopropanol precipitation. The RNA was resuspended in nuclease-free water. The RNA concentration and purity of RNA were measured using a NanoDrop 2000 spectrophotometer (Thermo Scientific, USA). Then, the total RNA (2 μg) was reverse transcribed into cDNA using a Prime Script RT Kit (RR820A, Takara, Japan). The reaction mixture was added to the RNA solution, and the RNA/reagent mixture was incubated at 42 °C for 1 h, further heated at 70 °C for 5 min, and cooled to 48 °C. The qRT-PCR was performed on a BioRad Connect Real-Time PCR platform (Bio-Rad, USA). PCRs were performed in a 20 μl volume with SYBR Master Mix and 1 μl pre-primer and 1 μl post-primer PCR conditions were: 95 °C for 30 s, 40 cycles of 95 °C for 5 s and 60 °C for 30 s, 95 °C for 10 s, and 60 °C for 5 s. GAPDH was used as a housekeeping gene. The qRT-PCR primer sequences are listed in Supplementary Table 2. The Ct values were normalized to GAPDH, and relative expression levels were calculated using the 2^−ΔΔCT method.

Electron microscopy

Animals were euthanized with an overdose of inhalation anesthetic (isoflurane) and transcardially perfused with a buffered solution containing 2.5% glutaraldehyde and 2.0% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4). Coronal vibratome sections were cut at 80 μm thickness, post-fixed in 1.5% potassium ferrocyanide and 2% osmium, and stained with 1% thiocarbohydrazide followed by 2% osmium tetroxide. Sections were then stained overnight in 1% uranyl acetate, washed in distilled water at 50 °C, and stained with lead aspartate (pH 5.0) at the same temperature. After dehydration in graded alcohols, sections were embedded in Durcupan resin and hardened at 65 °C for 24 h between glass slides. The coronal section containing the hippocampus was trimmed with a razor blade and glued to an aluminum stub using conductive glue. The block was further trimmed with a glass knife, mounted inside a scanning electron microscope (Zeiss Merlin, Zeiss NTS) equipped with a block-face cutting microtome (3View, Gatan), and images were collected.

ROS detection

ROS were determined by dihydroethidium (DHE) staining. Briefly, 10 μM thick frozen brain sections or hippocampal neuronal cells were incubated with 10 mmol/L DHE (Sigma-Aldrich, St. Louis, MO, USA) at 37 °C for 30 min in a humidified chamber without light exposure. The staining was monitored under a microscope (IX53, Olympus, Tokyo, Japan). Quantitative analysis was conducted utilizing ImageJ software (National Institutes of Health, USA).

Statistical analysis

Data are presented as mean ± standard error of the mean (SEM). All statistical analyses were performed using SPSS 20.0 (IBM Corp., Armonk, NY, USA). Normality of data distribution was assessed using the Shapiro-Wilk test, and homogeneity of variances was evaluated using Levene’s test. For comparisons involving more than two groups, one-way or two-way analysis of variance (ANOVA) was used as appropriate. For the Barnes maze test, two-way repeated-measures ANOVA was performed with group as the between-subjects factor and day as the within-subjects factor. When a significant main effect or interaction was detected, post hoc comparisons were conducted using Tukey’s honestly significant difference (HSD) test. A two-tailed p-value < 0.05 was considered statistically significant. All behavioral and biochemical analyses were performed by experimenters blinded to group allocation.

Results

EE ameliorates ELS-induced schizophrenia-like behaviors and cognitive impairments

To evaluate the effects of ELS and EE on schizophrenia-like behaviors, a series of behavioral tests were conducted. In the PPI test, SE-MS rats exhibited a significant reduction in PPI compared with SE-CON rats, indicating impaired sensorimotor gating. EE treatment significantly restored PPI performance, suggesting an improvement in information processing deficits (Fig. 1B–D). In the OFT, SE-MS rats showed increased total distance traveled, reflecting hyperlocomotion. This phenotype was significantly reversed by EE treatment (Fig. 1E).

Fig. 1. EE treatment ameliorates ELS-induced schizophrenia-associated behaviors and cognitive impairments.

A Schematic illustration of the experimental design. Timeline: MS on PND9; EE/SE housing from PND28 to PND45; behavioral tests in adulthood from PND56 to PND65. B–D PPI test results. SE-MS rats exhibited a significant reduction in PPI compared to SE-CON rats; EE treatment partially restored PPI performance. E Total distance traveled in the OFT. SE-MS rats showed increased locomotor activity, which was reversed by EE. F Y maze test. All groups showed the total number of explorations during the test (left). SE-MS rats showed reduced spontaneous alternation rate in the Y-maze (right), which was improved by EE. G NOR test. All groups exhibited comparable exploration during the adaptation phase (left). SE-MS rats showed reduced discrimination after 24 h (right), which was improved by EE. H Barnes maze test. SE-MS rats displayed prolonged escape latency from Day 1 to Day 4. EE administration significantly shortened latency in EE-MS rats. ^{& &}p < 0.01, ^{& & &}p < 0.001 compared with the SE-MS group, ^#p < 0.05, ^{# #}p < 0.01, ^{# # #}p < 0.001 compared with the EE-MS group. Group sizes: n = 15 per group. Data are expressed as mean ± SEM. n.s. was not significant; *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

Cognitive function was further assessed using the Y-maze, novel object recognition (NOR), and Barnes maze tests. In the Y-maze, no significant differences were observed in total arm entries among groups, indicating comparable locomotor activity during the task. However, SE-MS rats exhibited a reduced spontaneous alternation rate, indicative of impaired working memory, which was significantly improved by EE (Fig. 1F). Similarly, in the NOR test, all groups displayed comparable exploration during the training phase. However, SE-MS rats showed a significantly reduced discrimination index at 24 h, indicating impaired recognition memory, whereas EE treatment restored this deficit (Fig. 1G). In the Barnes maze, SE-MS rats exhibited prolonged escape latency across training days (Day 1-4), reflecting impaired spatial learning and memory. EE treatment significantly shortened escape latency in EE-MS rats (Fig. 1H).

Collectively, these results demonstrate that ELS induces schizophrenia-like behavioral abnormalities, including deficits in sensorimotor gating, hyperlocomotion, and cognitive impairments, all of which are effectively ameliorated by EE. Thus, EE consistently prevented ELS-induced deficits across all behavioral measures.

EE attenuates neuronal apoptosis, restores synaptic integrity, and normalizes CREB-BDNF-TrkB signaling in MS rats

To investigate the neurobiological basis underlying the behavioral improvements induced by EE, with a particular focus on neuroplasticity-related signaling, we first assessed neuronal integrity and apoptosis in the hippocampus.

Nissl staining revealed that SE-MS rats exhibited disrupted neuronal organization and reduced neuronal density in hippocampal subregions compared with SE-CON rats. In contrast, EE-MS rats showed markedly improved neuronal morphology and increased numbers of Nissl-positive neurons, indicating that EE preserves neuronal cytoarchitecture (Fig. 2A–D).

Fig. 2. EE treatment reduces neuronal apoptosis and restores cytoarchitecture in the hippocampus.

A Nissl staining in the hippocampus. SE-MS rats showed disorganized neuronal layers, while EE-MS rats showed improved neuronal integrity. B–D Quantification of Nissl-positive cells in hippocampal subregions. E Representative Western blot bands of anti-apoptotic (Bcl-2) and pro-apoptotic (BAX, Caspase-3, Cleaved Caspase-3) proteins in the hippocampus. F–I Western blot analyses of anti-apoptotic (Bcl-2) and pro-apoptotic (BAX, Caspase-3, Cleaved Caspase-3) proteins in the hippocampus. EE reversed the apoptotic imbalance. Group sizes: n = 4 per group. Data are expressed as mean ± SEM. n.s. was not significant; *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

Consistent with these findings, Western blot analysis demonstrated that SE-MS rats exhibited a pronounced imbalance between pro- and anti-apoptotic proteins, characterized by decreased Bcl-2 expression and increased levels of BAX, Caspase-3, and Cleaved Caspase-3. EE treatment effectively reversed these alterations, indicating suppression of neuronal apoptosis (Fig. 2E–I).

Given the critical role of synaptic integrity in cognitive function, we next examined dendritic spine density and synaptic protein expression. SE-MS rats displayed a significant reduction in dendritic spine density in the hippocampus, which was restored by EE treatment (Fig. 3A, B). Similarly, the expression of synaptic proteins SYN and PSD95 was markedly decreased in SE-MS rats but significantly increased following EE exposure (Fig. 3C–E), indicating improved synaptic structure and function.

Fig. 3. EE restores synaptic protein expression and dendritic spine density in MS rats.

A, B Representative images and quantification of dendritic spine density in the hippocampus. EE reversed spine loss induced by MS. C Representative Western blot bands of SYN and PSD95 expression in the hippocampus. D, E Western blot analysis of SYN and PSD95 expression in the hippocampus. SE-MS rats showed decreased expression, which was restored in EE-MS rats. Group sizes: n = 4 per group. Data are expressed as mean ± SEM. n.s. was not significant; *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

We next assessed the CREB-BDNF-TrkB signaling pathway, a key regulator of synaptic plasticity. SE-MS rats exhibited a significantly reduced p-CREB/CREB ratio, decreased BDNF expression, and a lower p-TrkB/TrkB ratio compared with SE-CON rats, indicating impaired neuroplasticity-related signaling. Notably, EE treatment restored the activity of this pathway (Fig. 4A–F).

Fig. 4. EE treatment restores CREB-BDNF-TrkB signaling pathway in the hippocampus.

A Representative Western blot bands of CREB, p-CREB, BDNF, TrkB, and p-TrkB, in the hippocampus. B–H Western blot analysis and quantification of CREB, p-CREB, BDNF, TrkB, and p-TrkB, in the hippocampus. SE-MS rats showed significantly reduced expression of ratio of p-CREB/CREB, BDNF, and ratio of p-TrkB/TrkB compared to SE-CON rats, which were restored following EE treatment. Group sizes: n = 4 per group. Data are expressed as mean ± SEM. n.s. not significant; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Taken together, these findings indicate that ELS induces neuronal apoptosis and synaptic deficits, accompanied by suppression of CREB-BDNF-TrkB signaling, whereas EE was associated with preservation of neuronal integrity, enhanced synaptic plasticity, and restoration of this signaling pathway.

EE attenuates oxidative stress, improves mitochondrial integrity, and suppresses neuroinflammation in MS rats

To further elucidate the upstream mechanisms underlying ELS-induced neuronal and synaptic deficits, we next examined oxidative stress, mitochondrial integrity, and neuroinflammatory responses in the hippocampus.

As shown in Fig. 5A–D, SE-MS rats exhibited significantly elevated ROS levels compared with SE-CON rats, indicating increased oxidative stress. Notably, EE treatment effectively normalized ROS levels.

Fig. 5. EE treatment rescued ROS increase and improves mitochondrial function and morphology in MS rats.

A–D ROS levels in the hippocampus. SE-MS rats showed elevated ROS, which was normalized in EE-MS rats. E Representative electron microscopy images of hippocampal mitochondria. F Mitochondrial density showed no significant difference among groups in hippocampal neurons. G Quantification of mitochondrial area in the hippocampus. SE-MS rats exhibited mitochondrial swelling, which was significantly reduced by EE, whereas mitochondrial density remained unchanged across groups (F), indicating that EE primarily improves mitochondrial structural integrity without altering mitochondrial number. Group sizes: n = 4 per group. Data are expressed as mean ± SEM. n.s. was not significant; *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

To determine whether mitochondrial abnormalities contribute to the observed oxidative stress, we assessed mitochondrial morphology via electron microscopy. While mitochondrial density was comparable across all groups (Fig. 5F), SE-MS rats showed a significant increase in mitochondrial area, indicative of swelling and impaired structural integrity. EE treatment robustly reversed this swelling without affecting mitochondrial number (Fig. 5G), indicating that EE primarily preserves mitochondrial quality rather than quantity.

We next investigated whether neuroinflammation was also involved in ELS-induced pathology. Immunofluorescence staining for Iba1 revealed a marked increase in microglial activation in SE-MS rats, which was significantly attenuated by EE (Fig. 6A–D). Consistently, qPCR analysis demonstrated that pro-inflammatory cytokines, including TNF-α, IL-1β, and IL-6, were significantly upregulated in SE-MS rats, whereas EE treatment effectively suppressed their expression (Fig. 6E–G).

Fig. 6. EE treatment alleviates microglial inflammatory activation in MS rats.

A–D Immunofluorescence staining for Iba1 in the hippocampus revealed increased microglial activation in SE-MS rats, which was suppressed by EE. E–G qPCR analysis of TNF-α, IL-1β, and IL-6 expression in the hippocampus. EE reversed the upregulation of inflammatory cytokines observed in SE-MS rats. Group sizes: n = 4 per group. Data are expressed as mean ± SEM. n.s. was not significant; *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

Taken together, these findings indicate that ELS induces pronounced oxidative stress, mitochondrial dysfunction, and neuroinflammatory activation in the hippocampus, all of which are effectively alleviated by EE. These findings raise the possibility that oxidative stress may contribute to the suppression of neuroplasticity-related signaling observed in MS rats.

TrkB activation rescues oxidative stress-induced synaptic deficits in vitro

Building upon the in vivo findings that ELS induces oxidative stress and suppresses CREB-BDNF-TrkB signaling (Sections 3.2–3.3), we next sought to determine whether TrkB signaling directly mediates oxidative stress-induced synaptic impairment in vitro.

Primary hippocampal neurons were exposed to H₂O₂ in the presence or absence of the TrkB agonist 7,8-DHF. As shown in Fig. 7A, B, H₂O₂ treatment markedly increased intracellular ROS levels compared with the DMSO control, confirming the establishment of an oxidative stress model.

Fig. 7. Activation of TrkB rescues oxidative stress-induced synaptic deficits in vitro.

A, B Intracellular ROS levels. C–F Representative blots and quantification of p-TrkB/TrkB ratio. G–I Synaptic protein expression (SYN, PSD95). H2O2 induced oxidative stress, suppressed TrkB signaling, and reduced synaptic protein levels; these effects were reversed by the TrkB agonist 7,8-DHF. n = 4 per group. Data are mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001. One-way ANOVA with Tukey’s post hoc test.

We next examined TrkB signaling activity. Western blot analysis revealed that H₂O₂ exposure significantly decreased the p-TrkB/TrkB ratio, indicating suppression of TrkB activation under oxidative stress conditions (Fig. 7C–F). Importantly, 7,8-DHF treatment restored TrkB phosphorylation levels.

Consistent with these findings, H₂O₂ treatment significantly reduced the expression of synaptic proteins SYN and PSD95, whereas 7,8-DHF markedly rescued their expression (Fig. 7G–I).

Taken together, these results indicate that oxidative stress impairs synaptic integrity in association with suppression of TrkB signaling, and that pharmacological activation of TrkB is sufficient to reverse these deficits in vitro.

TrkB signaling is required for the protective effects of EE in vivo

Given the restoration of CREB-BDNF-TrkB signaling by EE observed in vivo (Section 3.2), we next investigated whether TrkB activation is required for the behavioral and cognitive benefits of EE.

To this end, the selective TrkB antagonist ANA-12 was administered to EE-MS rats. As shown in Fig. 8A–E, ANA-12 treatment significantly reduced BDNF expression and the p-TrkB/TrkB ratio compared with the EE-MS group, confirming effective inhibition of TrkB signaling in vivo.

Fig. 8. TrkB signaling is required for the beneficial effects of EE on ELS-induced behavioral and cognitive deficits.

A Representative Western blot bands of BDNF, TrkB, and phosphorylated TrkB (p-TrkB) in the hippocampus. B–E Quantitative analysis of BDNF and TrkB signaling. Compared with the EE-MS group, administration of the TrkB antagonist ANA-12 significantly reduced BDNF expression and the p-TrkB/TrkB ratio, indicating effective inhibition of TrkB signaling in vivo. F–H Prepulse inhibition (PPI) test. ANA-12 treatment significantly impaired PPI performance compared with the EE-MS group, indicating disruption of sensorimotor gating. I OFT. ANA-12-treated rats exhibited increased locomotor activity compared with the EE-MS group. J Y-maze test. No significant differences were observed in total arm entries (left). However, ANA-12 significantly reduced spontaneous alternation rate (right) compared with the EE-MS group, indicating impaired working memory. K Novel object recognition (NOR) test. No differences were observed during the training phase (left). ANA-12-treated rats exhibited a significantly reduced discrimination index at 24 h (right) compared with the EE-MS group, indicating impaired recognition memory. L Barnes maze test. ANA-12-treated rats displayed prolonged escape latency across training days (Day 1-4) compared with the EE-MS group, indicating impaired spatial learning and memory. Statistical significance: &p < 0.05, & &p < 0.01compared with the EE-MS group; ^#p < 0.05, ^{# #}p < 0.01, ^{# # #}p < 0.001 compared with the EE-MS + ANA-12 group. Group sizes: n = 10 per group. Data are expressed as mean ± SEM. n.s. was not significant; *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

We next assessed behavioral and cognitive outcomes. In the PPI test, ANA-12-treated rats exhibited a significant reduction in PPI compared with EE-MS rats, indicating impaired sensorimotor gating (Fig. 8F–H). In the OFT, ANA-12 treatment resulted in increased locomotor activity (Fig. 8I).

In cognitive assessments, ANA-12 significantly reduced spontaneous alternation in the Y-maze without affecting total arm entries, indicating impaired working memory (Fig. 8J). Similarly, in the NOR test, ANA-12-treated rats displayed a reduced discrimination index at 24 h, suggesting impaired recognition memory (Fig. 8K). In the Barnes maze, ANA-12 treatment prolonged escape latency across training days, reflecting deficits in spatial learning and memory (Fig. 8L).

Taken together, blockade of TrkB signaling abolished the behavioral and cognitive improvements induced by EE, demonstrating that TrkB signaling is required for the behavioral and cognitive benefits of EE within this model.

Reduced TrkB levels in NDEs from patients with schizophrenia

To evaluate the clinical relevance of TrkB signaling in schizophrenia, we isolated NDEs from peripheral blood of patients with schizophrenia and healthy controls using immunomagnetic beads targeting the neuron-specific marker CD171. TEM and NTA confirmed the typical morphology and size distribution of the isolated extracellular vesicles (Fig. 9A, B).

Western blot analysis further verified successful enrichment of NDEs, as indicated by positive expression of the neuronal marker CD171 and exosomal markers TSG101 and CD81, while the absence of the microglial marker TMEM119 and astrocytic marker GLAST confirmed the specificity of neuronal exosome isolation (Fig. 9C).

Quantitative analysis revealed that TrkB protein levels, normalized to CD81 were significantly reduced in NDEs derived from patients with schizophrenia compared with healthy controls (Fig. 9D). Receiver operating characteristic (ROC) curve analysis showed that exosomal TrkB levels discriminated patients from controls with an area under the curve (AUC) of 0.784, indicating a moderate discriminatory capacity (Fig. 9E).

Demographic and clinical characteristics of the study participants are summarized in Supplementary Table 1. The groups were comparable in age and sex distribution. A modest difference in BMI reached statistical significance, but the effect size was small and unlikely to confound the observed changes in exosomal TrkB. As expected, patients had fewer years of education and performed substantially worse across all BACS domains, confirming widespread cognitive impairment in this cohort.

Correlational analyses between exosomal TrkB levels and clinical measures are presented in Supplementary Table 3. In brief, the TrkB/CD81 ratio correlated inversely with total PANSS scores (r = -0.589, p < 0.001), and with the positive (r = -0.381, p = 0.020), disorganized/concrete (r = -0.403, p = 0.013), and depressive factors (r = -0.329, p = 0.047) in particular. Among cognitive domains, positive correlations emerged with token motor task (r = 0.362, p = 0.027), word fluency (r = 0.366, p = 0.026), and overall BACS composite score (r = 0.414, p = 0.011), but not with other subdomains. These associations strengthen the link between reduced peripheral TrkB signaling and core psychopathological and cognitive features in schizophrenia.

Discussion

In this study, we demonstrate that ELS, modeled by MS, induces enduring schizophrenia-like phenotypes in adulthood, including sensorimotor gating deficits, cognitive impairments, synaptic dysfunction, and hippocampal neurobiological alterations. Importantly, EE administered during adolescence effectively prevents these abnormalities. Mechanistically, our findings identify the CREB-BDNF-TrkB signaling pathway as an important mechanistic pathway associated with the protective effects of EE. These results suggest adolescence as a putative critical developmental window for intervention and support EE as a promising non-pharmacological strategy to mitigate ELS-induced vulnerability to schizophrenia-related pathology within the experimental framework used here.

At the cellular and systems level, MS-exposed animals exhibited neuronal apoptosis, synaptic loss, mitochondrial dysfunction, and heightened neuroinflammation. These alterations closely parallel key pathological features observed in schizophrenia [26, 27], and are consistent with previous reports that ELS induces oxidative stress and neuroimmune dysregulation [6, 28, 29]. Importantly, they were robustly reversed by EE, indicating multi-level neuroprotective effects. These upstream disturbances are closely associated with alterations in neuroplasticity-related signaling pathways.

A key finding of this study is evidence supporting an important role for CREB-BDNF-TrkB signaling as a converging node linking ELS-induced pathology to synaptic and behavioral deficits. We found that MS suppressed CREB phosphorylation, reduced BDNF expression, and impaired TrkB activation, consistent with clinical evidence of reduced hippocampal BDNF signaling in schizophrenia [16, 30]. Mechanistically, our data support a model in which oxidative stress and neuroinflammation are associated with suppression of this signaling axis. Direct experimental evidence from our study demonstrates that oxidative stress (induced by H₂O₂ in vitro) suppresses TrkB signaling and synaptic protein expression, effects that were reversed by the TrkB agonist 7,8-DHF. In contrast, the link between neuroinflammation and CREB-BDNF-TrkB suppression remains correlational within the present study: we observed concurrent increases in pro-inflammatory cytokines (TNF-α, IL-1β, and IL-6) and reductions in TrkB signaling in MS rats, both of which were normalized by EE. Elevated pro-inflammatory cytokines have been reported to suppress BDNF expression through NF-κB-dependent and epigenetic mechanisms [31]. Excessive mitochondrial ROS may also impair CREB activation through disruption of calcium signaling and energy metabolism [17, 32]. Together, these findings support a convergence model in which these upstream stress-related changes are associated with downstream synaptic dysfunction, consistent with the hypothesis that they may contribute to the suppression of CREB-BDNF-TrkB signaling.

Importantly, our in vitro and in vivo findings indicate that TrkB signaling is a critical mediator of the neuroprotective effects observed in this study. Pharmacological activation of TrkB using 7,8-DHF rescued oxidative stress-induced synaptic deficits in vitro, demonstrating that TrkB activation is sufficient to counteract these deficits at the cellular level. In vivo, pharmacological inhibition of TrkB with ANA-12 abolished the behavioral and cognitive benefits of EE, establishing that TrkB signaling is necessary for these protective effects. Thus, within the experimental framework used here, TrkB signaling serves as a necessary (in vivo) and, in the context of oxidative stress-induced synaptic deficits, sufficient (in vitro) mechanistic link between upstream pathological processes and downstream outcomes.

To evaluate the clinical relevance of TrkB signaling in schizophrenia, we isolated NDEs from peripheral blood of patients with schizophrenia. NDEs provide a potentially informative window into central nervous system pathophysiology because they are enriched for neuronal cargo and can be detected in peripheral blood, enabling minimally invasive assessment of neuronal molecular signatures [33–35]. Although CD171-based immunocapture is widely used for neuronal exosome enrichment, it may not fully represent all neuronal subtypes; therefore, the isolated vesicles should be interpreted as neuron-enriched exosomes rather than a complete representation of the neuronal population. TrkB levels, normalized to the exosomal marker CD81, were significantly reduced in NDEs from patients compared with healthy controls, with moderate diagnostic performance (AUC = 0.784). While phosphorylated TrkB would provide a more direct assessment of receptor activation, reliable quantification of p-TrkB in serum-derived NDEs remains technically challenging because of its low abundance and potential instability. Therefore, total TrkB was used as an initial translational indicator of pathway alterations. This observation aligns with our preclinical findings of suppressed TrkB signaling in the hippocampus following ELS, supporting the notion that peripheral exosomal TrkB may reflect central neuroplasticity deficits. Although the diagnostic accuracy is moderate (AUC = 0.784) and does not support the use of exosomal TrkB as a standalone diagnostic tool, this finding suggests potential clinical utility as one component of a multi-marker panel for risk stratification, or as a state-related indicator of neuroplasticity dysfunction and treatment response. Future studies incorporating additional exosomal or circulating biomarkers are needed to improve diagnostic performance. Beyond its diagnostic potential, TrkB signaling also represents a developmentally targetable pathway for intervention, particularly given our demonstration that adolescent EE restores TrkB activation and prevents schizophrenia-related phenotypes. These findings support TrkB signaling as a candidate biomarker pathway and a potential therapeutic target.

Our results also suggest adolescence as a potential window for preventive intervention. Schizophrenia typically emerges in late adolescence or early adulthood, following earlier neurodevelopmental disruptions [36–38], indicating a period of latent vulnerability that may be modifiable. Consistent with this trajectory, we previously reported that MS-induced behavioral abnormalities manifest in adulthood but not during adolescence [7]. The hippocampus undergoes pronounced structural changes during adolescence, including synaptic pruning and circuit refinement that extend into early adulthood [13, 14, 39, 40]. This prolonged maturation renders the adolescent period particularly sensitive to environmental modulation [15, 41]. Within this framework, EE represents a multi-dimensional intervention capable of enhancing neural plasticity through sensory, cognitive, and social stimulation. Unlike interventions targeting a single molecular pathway, EE may simultaneously modulate oxidative stress, neuroinflammation, mitochondrial function, and neuroplasticity-related signaling, thereby promoting resilience against the long-term consequences of ELS. Because EE consists of multiple interacting components (physical, cognitive, sensory, and social stimulation), the present study was not designed to determine the relative contribution of individual elements. Future work may help identify which aspects of enrichment are most critical for mediating the observed neuroprotective effects. Notably, the beneficial effects of EE were preferentially observed in stressed animals, consistent with previous reports that EE exerts stronger effects under pathological conditions [42].

Several limitations should be acknowledged. First, only male rats were included to reduce hormonal variability and maintain consistency with our previous MS studies. Therefore, whether similar effects occur in females remains unclear. Future studies incorporating female animals are needed to determine whether the protective effects of EE and the role of TrkB are sex dependent. Second, medication status was not controlled in the clinical cohort, and potential effects of antipsychotic treatment on exosomal TrkB levels cannot be excluded. Third, the present study employed a single MS paradigm and did not compare EE interventions across different developmental stages. Future studies using additional ELS models and alternative intervention windows will help determine the generalizability of the TrkB-dependent mechanism identified here. Finally, although BMI differed modestly between groups, the magnitude of this difference was small and unlikely to confound the observed alterations in exosomal TrkB levels. Nevertheless, given the limited sample size, a potential contribution of BMI-related factors cannot be entirely excluded and should be examined in larger future studies.

Conclusions

In summary, ELS induces persistent schizophrenia-related behavioral and neurobiological abnormalities, which are prevented by adolescent EE. Mechanistically, EE was associated with reduced oxidative stress and neuroinflammation, improved mitochondrial integrity, and restoration of CREB-BDNF-TrkB signaling (Fig. 10). Our findings support an important role for TrkB signaling in these protective effects, as TrkB activity was required for the behavioral benefits of EE in vivo and was sufficient to rescue oxidative stress-induced synaptic deficits in vitro. The observation of reduced TrkB levels in neuron-derived exosomes from patients with schizophrenia further supports the translational relevance of this pathway. Together, these findings support a model in which alterations in neuroplasticity-related TrkB signaling contribute to the link between ELS and later schizophrenia-related vulnerability, and highlight adolescence as a potentially modifiable developmental period for preventive intervention.

Fig. 10. Mechanistic framework linking ELS, adolescent EE, and TrkB-dependent neuroplasticity in schizophrenia.

ELS, modeled by MS, induces oxidative stress and neuroinflammation in the hippocampus, which suppress CREB-BDNF-TrkB signaling and lead to synaptic dysfunction and schizophrenia-like behavioral abnormalities. EE during adolescence, a critical developmental window, counteracts these effects by restoring CREB-BDNF-TrkB pathway activity and improving synaptic integrity and cognitive function. Mechanistically, TrkB signaling serves as a central mediator of EE-induced neuroprotection. Inhibition of TrkB with ANA-12 abolishes the beneficial effects of EE on behavior and cognition (necessity), whereas activation with 7,8-DHF rescues oxidative stress-induced synaptic deficits (sufficiency), supporting TrkB as a key converging node within the experimental framework.

Author contributions

K-K H, H-L W, and F-S C designed the study and wrote the protocol. F-S C and K-K H performed the experiments and analyzed the data. K-K H, X-Y L, X-W W, Q-Y Y, H-H, and G-P C performed literature searches and drew the figures. K-K H and F-S C wrote the manuscript. H-L W revised the manuscript.

Funding

The study was supported by the National Natural Science Foundation of China (Grant No. 82471523) and the National Natural Science Foundation of China (Grant No. 82501834).

Data availability

Data generated during the current study are available from the corresponding author upon reasonable request.

Competing interests

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.

Ethics approval and consent to participate

This work was carried out based on the Regulations of Experimental Animal Administration issued by the State Committee of Science and Technology of the People’s Republic of China, with the approval and consent to participate of the Ethics Committee in Renmin Hospital of Wuhan University.

Consent for publication

Not applicable.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41398-026-04388-6.