OSBPL10 alleviates neuronal ferroptosis via lysosomal membrane repair in a PS-dependent manner after spinal cord injury

Abstract

Spinal cord injury (SCI) exerts severe adverse effects on patients, leading to impaired motor functions in segments remote from the injury site. Following SCI, lysosomes are damaged due to inflammatory microenvironment, and the repair of such damage requires a substantial supply of lipids. Studies have demonstrated that lipid transfer proteins (LTPs) play a pivotal role in the repair process of neuronal organelles. However, whether these proteins regulate functional recovery after SCI and their underlying mechanisms remain elusive. Through RNA sequencing and human cerebrospinal fluid enzyme-linked immunosorbent assay (ELISA) analyses, we verified that OSBPL10 is significantly downregulated in the acute phase of SCI and may serve as a protective factor against SCI. The Basso Mouse Scale (BMS), Nissl staining, and two-dimensional (2D) gait analysis were employed to evaluate functional recovery. Western blotting and immunofluorescence assays were performed to detect the expression changes of proteins associated with autophagy, ferroptosis, oxidative stress, and lysosomal membrane permeabilization (LMP) related proteins. Finally, molecular docking and rescue experiments using lipid synthesis inhibitors further confirmed that OSBPL10 ameliorate LMP by transferring PS to lysosomes, thereby promoting autophagic flux. Collectively, our findings conclude that OSBPL10 overexpression alleviates autophagic flux impairment, ferroptosis, and oxidative stress after SCI through PS-mediated lysosomal repair, thus facilitating post-injury neurological function recovery and holding promising potential for clinical application.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12974-026-03760-z.

Introduction

Spinal cord injury (SCI) is a severe central nervous system (CNS) disorder with a poor prognosis. Epidemiological investigations have shown that approximately 23 people per million population worldwide are severely affected by SCI [1]. Despite the remarkable changes in the characteristics of the SCI-affected population in the 21st century, its incidence remains persistently high. In China, the number of SCI patients has been rising year by year, with an annual incidence rate as high as 60 cases per million population [2]. In recent years, despite continuous advances in basic and clinical research, the impairment of neural regeneration post-SCI remains a key factor limiting the recovery of neurological function [3].

The pathogenesis of SCI involves two distinct phases: primary injury and secondary injury, the latter of which is mediated by a variety of pathological processes [4]. During the secondary injury phase, the activation of inflammatory responses and infiltration of immune cells lead to the impairment of neuronal lysosomal function and disruption of the local microenvironment [5, 6]. Lysosomal damage is a key cause of neuronal loss [7, 8]. Therefore, inhibiting neuronal death and maintaining the structural integrity of organelles are crucial for SCI repair. Ferroptosis is a novel form of programmed cell death distinct from apoptosis, necrosis and pyroptosis, which is driven by oxidative damage induced by intracellular iron-dependent reactive oxygen species (ROS) accumulation and lipid peroxidation [9, 10]. This death pathway is closely associated with various pathological conditions and exerts a key regulatory role particularly in traumatic CNS injuries such as SCI [11, 12], suggesting that specific targeting of neuronal ferroptosis may serve as a potential therapeutic strategy to improve neuronal survival rate after SCI.

Neuronal loss and axonal injury are the core causes of poor functional recovery after SCI, and thus neuronal survival and protection have long been the central focus of SCI research. Autophagy is a lysosome-dependent intracellular degradation pathway [13], which has been confirmed to have a clear etiological association with various diseases including neurodegenerative diseases, inflammatory diseases and cancer [14, 15], and plays a key role in the processes of normal cell proliferation, maturation and differentiation [16]. Recent studies have further confirmed that autophagy is essential for neuronal survival and functional maintenance, and exerts a prominent role especially in inhibiting programmed cell death and safeguarding the functional homeostasis of neurons [17–19]. In SCI models, studies have demonstrated that neuronal autophagic flux is impaired, and enhancing autophagic flux can significantly improve neuronal survival rate [20, 21]. The autophagic process consists of key steps including phagophore formation, autophagosome development, autophagosome-lysosome fusion and autolysosome degradation [22], and the integrity of lysosomal function is a prerequisite for the normal progression of autophagic flux. Notably, lysosomal dysfunction, particularly the increase in lysosomal membrane permeability (LMP), has been proven to be closely associated with traumatic neurological disorders such as SCI and traumatic brain injury (TBI) [7, 8]. Maintaining the normal permeability of the lysosomal membrane is critical for preserving lysosomal function and preventing cellular components from being damaged by lysosomal luminal enzymes [20]. However, how to protect lysosomal function as well as the autophagic process in SCI remains undefined.

Lipid transfer protein plays an important regulatory role in the repair of organelle damage. Among them, non-vesicular lipid transport mediated by oxysterol-binding proteins (OSBPs) and autophagy-related proteins (ATGs) has been identified as a core mechanism for organelle repair [23–26]. All members of the OSBP family contain an OSBP-related domain (ORD) with a lipid-binding pocket [27]. The classic functional model is as follows: cholesterol or phosphatidylserine (PS) is extracted from the endoplasmic reticulum and exchanged for phosphatidylinositol 4-phosphate (PI4P, generated by the phosphorylation of phosphatidylinositol catalyzed by PI4 kinases) on organelle membranes or the plasma membrane [28]. PI4P is then transported to the ER membrane and hydrolyzed by the phosphatase Sac1, and this cycle drives the accumulation of cholesterol and PS on target membranes, thereby achieving membrane structural repair. A recent study by Tan et al. has revealed the association between the OSBP family and lysosomal damage repair [26], yet the role and regulatory mechanism of the OSBP family in SCI remain unclear. Our preliminary RNA sequencing data have shown that the expression level of OSBPL10 is significantly downregulated after SCI, suggesting insufficient endogenous repair potential, and its association with lysosomal damage repair are worthy of in-depth investigation.

In summary, several critical questions remain to be addressed in current research: systematic investigations are lacking on the specific function of OSBPL10 in SCI, its direct association with the autophagy-lysosome system, the precise molecular mechanisms underlying its protective effects, as well as the spatiotemporal dynamic changes in OSBPL10 expression during the progression of SCI. The present study aims to elucidate the core mechanism by which OSBPL10 mediates lysosomal damage repair via PS after SCI. Our findings confirm that OSBPL10 primarily alleviates lysosomal membrane permeabilization through PS-dependent lysosomal membrane repair, thereby promoting the functional recovery of the autophagy-lysosome system, and ultimately exerts neuroprotective effects after SCI by inhibiting neuronal ferroptosis. The results of this study lay a solid mechanistic foundation for the development of novel OSBPL10-based therapeutic strategies for SCI.

Results

OSBPL10 is downregulated in both human and mouse samples following spinal cord injury (SCI)

First, we examined the expression profiles of lipid transfer proteins associated with the OSBP and ATG families during the acute phase following SCI. After SCI, the expression of the Osbpl10 gene at the injury site was markedly downregulated, reaching the lowest level at 3 days post-injury (Fig. 1A). It has been reported that OSBPL9/10/11 can form a transfer complex to mediate rapid lipid trafficking for lysosomal repair [26]. Accordingly, we detected the corresponding proteins via western blot analysis. The results demonstrated that the OSBPL10 protein was significantly downregulated after SCI, whereas the OSBPL9 protein expression remained unchanged and OSBPL11 protein expression was notably upregulated (Fig. 1B-E). Collectively, these findings led us to hypothesize that insufficient Osbpl10 expression may represent a critical bottleneck for endogenous repair processes after SCI.

Fig. 1.

OSBPL10 expression is downregulated following spinal cord injury and colocalizes with neurons. A Heatmap of bulk-RNA sequence expression profiles for lipid transport proteins in the spinal cord at different time points following injury. B-E Protein levels of OSBPL9, OSBPL10 and OSBPL11 with their quantitative analysis in spinal cord tissue at the indicated time points (n = 6; Ordinary one-way ANOVA). F The cerebrospinal fluid (CSF) of patients with clinical characteristics included in this study. G Box plot showing OSBPL10 levels in the CSF of non-spinal cord injury controls (n = 18) and spinal cord injury patients at different stages: subacute (n = 5), early chronic (n = 51), and late chronic (n = 35) (Kruskal-Wallis H test followed by Dunn’s multiple comparison test with Bonferroni correction). H The scatter plot with a linear regression line representing the levels of CSF OSBPL10 in patients across different American Spinal Injury Association Impairment scale (AIS) grades, including AIS A (n = 41), AIS B (n = 24), AIS C (n = 14), and AIS D (n = 12) (Spearman’s rank correlation test). I Immunofluorescence staining of OSBPL10 with neurons in the ventral horn gray matter of spinal cord tissue in the indicated groups (scale bar: 25 μm for original pictures and 6 μm for enlarged pictures). J The ratio of OSBPL10 positive neurons, microglial and astrocytes to their respective totals in Sham and 3 days post-SCI mice (n = 6; Two-way ANOVA). All data are presented as the means ± SD; Error bars represent SDs; ∗∗p < 0.01, ∗p < 0.05, ns indicates no significance

To investigate the dynamic changes in OSBPL10 levels in human SCI, we collected cerebrospinal fluid (CSF) samples from SCI patients at different disease stages, including the subacute phase (2 weeks to 2 months post-injury), early chronic phase (2 to 12 months post-injury), and late chronic phase (> 12 months post-injury), as well as from control subjects with spinal canal stenosis without direct SCI (Fig. 1F). Enzyme-linked immunosorbent assay (ELISA) results revealed that the CSF OSBPL10 levels in SCI patients were significantly lower than those in the control group, with the lowest levels observed during the subacute phase (Fig. 1G). To evaluate the correlation between CSF OSBPL10 levels and SCI severity, patients were stratified according to the American Spinal Injury Association (ASIA) Impairment Scale (AIS). We found that CSF OSBPL10 levels in AIS grade A patients were significantly lower than those in patients with other AIS grades (Fig. 1H). These results suggest that CSF OSBPL10 may serve as a protective factor in SCI.

To further explore the cellular distribution of OSBPL10, we analyzed its spatial localization using immunofluorescence staining. The results showed that the fluorescence intensity of OSBPL10 in neurons was significantly reduced at 3 days post-injury compared with that in the sham-operated group (Fig. 1I-J), whereas no significant changes in OSBPL10 expression were detected in other cell types (Figure S1A).

Overexpression of OSBPL10 in neurons promotes functional recovery after SCI

To further investigate the in vivo role of OSBPL10, we generated neuron-specific OSBPL10-overexpressing mice and their corresponding empty vector control mice using adeno-associated virus (AAV). Notably, gene editing did not affect the body weight of mice (Figure S2A), and no obvious abnormalities were observed in hematoxylin-eosin (HE) staining of heart, liver, spleen, lung, and kidney tissues (Figure S2B). In addition, immunofluorescence co-localization experiments of OSBPL10 with neurons verified the efficiency of OSBPL10 overexpression (Figure S2C-D).

Subsequently, we investigated the effect of neuron-specific OSBPL10 overexpression on motor function in SCI mice (Fig. 2A). Basso Mouse Scale (BMS) scores indicated that OSBPL10 overexpression promoted the recovery of hindlimb motor function in SCI model mice (Fig. 2B). To accurately record the hindlimb movement changes, we captured the gait of mice. The key body parts in the videos were labeled in two dimensions using BehaviorAtlas Posture Marker (V1.0.0; Guangdong Wanmeng Science Co., Ltd.) (see Methods section for detailed procedures), and analyzed with BehaviorAtlas Two-Dimensional Artificial Intelligence Motion Analyzer, as previously reported. Detailed kinematic analysis of the hindlimbs showed that OSBPL10 overexpression further increased the stride length and the maximum toe elevation height of SCI model mice (Fig. 2C-E). Next, electrophysiological analysis of SCI mice was performed by measuring motor evoked potentials (MEPs). The results demonstrated that OSBPL10 overexpression increased the MEP amplitude in SCI mice (Fig. 2F-G).

Fig. 2.

Overexpression of OSBPL10 promotes functional recovery after SCI. A Schematic timeline of the experimental procedure. B Measurement of BMS scores on day 0, 1, 3, 7, 14, 21, 28 post injury from the indicated groups (n = 6; Ordinary one-way ANOVA). C Gait timing photograph at 28 days post-injury, corresponding color-coded rod stick view decomposition of mouse hindlimb movements. The mobile phase is shown as red sticks, and the stationary phase as black sticks. D Quantitative analysis of stride length of mice on the 28th day post-injury from the indicated groups (n = 6; Ordinary one-way ANOVA). E Quantitative analysis of maximum toe height at 28 days post-injury in the indicated groups (n = 6; Ordinary one-way ANOVA). F Representative diagrams of motor evoked potential detection on the 28th day post-injury in the indicated groups. G Quantitative analysis for the amplitude of the first peak (mV) in (F) (n = 6; Ordinary one-way ANOVA). H Nissl staining results of the ventral horn gray matter in transverse sections of the spinal cord from the indicated groups at 28 days post-injury. Black arrows indicate the motoneurons (scale bar: 500 μm for original pictures and 100 μm for enlarged pictures). I Quantitative analysis of Nissl positive motoneurons in (H) (n = 6; Ordinary one-way ANOVA). J Immunofluorescence images showing neurons in the Z1-Z3 regions of the gray matter in coronal sections adjacent to the lesion core at 28 days post-injury. Z1, Z2, and Z3 are located 0.5 mm, 1 mm, and 1.5 mm away from the epicenter of the injury zone, respectively (scale bar: 400 μm). K Quantitative analysis of the number of NEUN⁺ cells in (J) (n = 6; Ordinary one-way ANOVA). All data are presented as the means ± SD; Error bars represent SDs; ∗∗p < 0.01, ∗p < 0.05, ns indicates no significance

The survival of spinal motor neurons is closely associated with the recovery of hindlimb motor function after SCI. Nissl staining was used to label the number of motor neurons in the anterior horn of the spinal cord in the lesion area, and NEUN immunolabeling was performed to detect the neuron density in regions at different distances from the injury center (Z1-Z4). The results showed that the number of motor neurons and NEUN⁺ neurons in the Z1-Z3 regions in the SCI + AAV-Osbpl10 group was significantly higher than that in the SCI + AAV-NC group (Fig. 2H-K), demonstrating overexpression of OSBPL10 promoted neuronal survival after SCI.

Overexpression of OSBPL10 enhances autophagy after SCI

To further explore the potential mechanism of OSBPL10 on functional recovery after SCI, we performed bulk-RNA sequencing on spinal cord samples from the SCI + AAV-NC and SCI + AAV-Osbpl10 groups. Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis showed that the top 15 upregulated pathways compared with the SCI + AAV-NC group included “phagosome” (Fig. 3A). In addition, Gene Set Enrichment Analysis (GSEA) results indicated that overexpression of OSBPL10 reduced the expression levels of genes related to oxidative stress and ferroptosis (Fig. 3B). Then, we detected changes in autophagy-related indicators using western blot, quantitative real-time polymerase chain reaction (qPCR), and immunofluorescence.​.

Fig. 3.

Overexpression of OSBPL10 expression promotes autophagy after SCI. A The top 15 KEGG enrichment results of differentially expressed up-regulated genes between the SCI + AAV-NC and the SCI + AAV-Osbpl10 group (n = 4). B-C GSEA analysis of differentially expressed pathway between the SCI + AAV-NC and the SCI + AAV-Osbpl10 group enriched by KEGG. D WB analysis of ATG5, VPS34, BECLIN1, SQSTM1/P62 and LC3 expression levels in the indicated groups. E Quantitative analysis of protein expression of (D) (n = 6; Ordinary one-way ANOVA). F The number of LC3 II puncta in neurons in the indicated groups (n = 6; Ordinary one-way ANOVA). G Quantified integrated intensity of P62 in neurons in the indicated groups (n = 6; Ordinary one-way ANOVA). H Immunofluorescence images of LC3 and NEUN in spinal cords in the indicated groups (scale bar: 20 μm). I Immunofluorescence images of P62 and NEUN in spinal cords in the indicated groups (scale bar: 20 μm). J mRNA level of the Sqstm1 gene in the spinal cord tissue in the indicated groups (n = 6; Ordinary one-way ANOVA). All data are presented as the means ± SD; Error bars represent SDs; ∗∗p < 0.01, ∗p < 0.05, ns indicates no significance

Class III phosphatidylinositol 3-kinase (especially vacuolar protein sorting 34 (VPS34)) and its binding partner (BECLIN1) jointly participate in phagophore formation, mediating the nucleation and membrane extension of phagophores during the initiation stage of autophagy [29]. ATG5 is a key component in the autophagosome elongation process, localizing to the outer membrane of phagophores and regulating the lipidation and membrane localization of LC3 family proteins [30]. Microtubule-associated protein light chain 3 (LC3) is hydrolyzed by Atg4 to generate LC3-I, which is further processed into LC3-II. LC3-II localizes to both the inner and outer membranes of autophagosomes and serves as a reliable marker for autophagosomes [31]. P62/SQSTM1 is an autophagic substrate protein that can directly bind to LC3 and be selectively degraded through autophagy [32]. Therefore, we detected the protein expression levels of VPS34, BECLIN1, ATG5, LC3 and P62 at 3 days post-SCI. Western blot results showed that OSBPL10 overexpression increased the levels of autophagy-related proteins (ATG5, VPS34, BECLIN1, LC3-II) and decreased the level of the autophagic substrate protein P62 (Fig. 3D-E).

Immunofluorescence results showed that the number of LC3-II puncta in neurons increased after SCI, which was further enhanced by OSBPL10 overexpression. Meanwhile, the expression of P62 in neurons of the spinal cord injury area in the SCI + AAV-Osbpl10 group was lower than that in the SCI + AAV-NC group (Fig. 3F-I). As an autophagic substrate, P62 protein is co-regulated by autophagic flux and gene expression. qPCR and WB experiments results showed that both mRNA and protein levels of P62 increased after SCI, with a more significant increase in protein levels, indicating impaired autophagic flux after SCI. In contrast, OSBPL10 overexpression led to increased P62 mRNA level but decreased protein level, suggesting that OSBPL10 overexpression restored autophagic flux in SCI mice (Fig. 3J). Collectively, these results demonstrate that SCI leads to impaired autophagic flux, while OSBPL10 overexpression can promote autophagy and restore autophagic flux.

Overexpression of OSBPL10 alleviates ferroptosis after SCI

As mentioned above, overexpression of OSBPL10 reduced the expression levels of genes related to ferroptosis and oxidative stress (Fig. 3B-C). We further detected the degree of neuronal ferroptosis and oxidative stress after SCI. Immunofluorescence results showed that in spinal cord tissues after SCI, the expression of 4-hydroxynonenal (4HNE), a ferroptosis-related marker, was significantly upregulated in neurons, while the expression of solute carrier family 7 member 11 (SLC7A11), an anti-ferroptosis marker, was significantly downregulated (Fig. 4A-D). Consistently, western blot analysis revealed that the expression levels of ferroptosis-related markers, including acyl-CoA synthetase long-chain family member 4 (ACSL4), 4HNE, and tumor protein p53 (TP53), were significantly increased in spinal cord tissues after SCI, whereas the expression levels of anti-ferroptosis markers SLC7A11 and glutathione peroxidase 4 (GPX4) were significantly decreased (Fig. 4H-I). However, overexpression of OSBPL10 restored these changes in SCI mice.

Fig. 4.

Overexpression of OSBPL10 alleviates ferroptosis and oxidative stress after SCI. A Immunofluorescence images of 4HNE and NEUN in spinal cords in the indicated groups (scale bar: 20 μm). B Immunofluorescence images of SLC7A11 and NEUN in spinal cords in the indicated groups (scale bar: 20 μm). C Quantified integrated intensity of 4HNE in neurons in the indicated groups (n = 6; Ordinary one-way ANOVA). D Quantified integrated intensity of SLC7A11 in neurons in the indicated groups (n = 6; Ordinary one-way ANOVA). E-G Quantitative analysis of Fe²⁺ concentration, GSH content, and MDA levels in spinal cord tissue (n = 6; Ordinary one-way ANOVA). H Western blot analysis of 4HNE, TP53, ACSL4, SLC7A11 and GPX4 expression levels in the indicated groups. I Quantitative analysis of protein expression of (H) (n = 6; Ordinary one-way ANOVA). J Immunofluorescence images of DHE and DAPI in spinal cords in the indicated groups (scale bar: 100 μm). K Colocalization analysis of the fluorescence intensity between DHE and DAPI in (J). All data are presented as the means ± SD; Error bars represent SDs; ∗∗p < 0.01, ∗p < 0.05, ns indicates no significance

In addition, the core characteristics of ferroptosis also include ferrous ion (Fe²⁺) accumulation, glutathione (GSH) depletion, and lipid peroxidation. Quantitative detection of these indicators showed that compared with the SCI + AAV-NC group, the SCI + AAV-Osbpl10 group significantly increased the GSH level and decreased the malondialdehyde (MDA) content and Fe²⁺ accumulation in spinal cord tissues (Fig. 4E-G). Dihydroethidium (DHE) staining was used to detect the level of ROS. The results showed that the DHE fluorescence intensity in the injury area was significantly increased after SCI, indicating massive ROS accumulation; neuron-specific OSBPL10 overexpression significantly inhibited ROS aggregation (Fig. 4J-K and Figure S3A).

Overexpression of OSBPL10 inhibits ferroptosis after SCI by activating autophagy

Previous studies have emphasized the existence of a complex interaction system between autophagy and other pathological mechanisms [20, 33]. Therefore, our subsequent study aimed to explore the correlation between changes in autophagy, oxidative stress, and ferroptosis in the injured spinal cord after OSBPL10 overexpression. We performed rescue experiments using 3-methyladenine (3MA), an inhibitor of autophagy intiation [34], comparing four groups: SCI + AAV-NC, SCI + AAV-NC/3MA, SCI + AAV-Osbpl10, and SCI + AAV-Osbpl10/3MA, to clarify the relationship among autophagy, oxidative stress, and ferroptosis. The detailed experimental procedure is shown in Figure S4A.

Immunofluorescence results showed that compared with the SCI + AAV-Osbpl10 group, the number of LC3 II puncta in neurons was significantly reduced and the accumulation intensity of P62 was increased in the SCI + AAV-Osbpl10/3MA group (Fig. 5A-D). In addition, western blot results demonstrated that compared with the SCI + AAV-Osbpl10 group, the expression levels of autophagy-related proteins (ATG5, VPS34, BECLIN1, LC3-II) were significantly decreased, while the expression of P62 was significantly increased in the SCI + AAV-Osbpl10/3MA group (Fig. 5E-F). These results indicate that 3MA effectively counteract the promoting effect of OSBPL10 overexpression on autophagy.

Fig. 5.

3MA counteracts the enhancement of autophagy and the inhibition of ferroptosis attributed to OSBPL10 overexpression. A Immunofluorescence images of LC3 and NEUN in spinal cords in the indicated groups (scale bar: 20 μm). B The number of LC3 II puncta in neurons in the indicated groups (n = 6; Ordinary one-way ANOVA). C Immunofluorescence images of P62 and NEUN in spinal cords in the indicated groups (scale bar: 20 μm). D Quantified integrated intensity of P62 in neurons in the indicated groups (n = 6; Ordinary one-way ANOVA). E Western blot analysis of ATG5, VPS34, BECLIN1, SQSTM1/P62 and LC3 expression levels in the indicated groups. F Quantitative analysis of protein expression of (E) (n = 6; Ordinary one-way ANOVA). G Immunofluorescence images of 4HNE and NEUN in spinal cords in the indicated groups (scale bar: 20 μm). H Quantified integrated intensity of 4HNE in neurons in the indicated groups (n = 6; Ordinary one-way ANOVA). I Immunofluorescence images of SLC7A11 and NEUN in spinal cords in the indicated groups (scale bar: 20 μm). J Quantified integrated intensity of SLC7A11 in neurons in the indicated groups (n = 6; Ordinary one-way ANOVA). K Western blot analysis of 4HNE, TP53, ACSL4, SLC7A11 and GPX4 expression levels in the indicated groups. L Quantitative analysis of protein expression of (K) (n = 6; Ordinary one-way ANOVA). M-O Quantitative analysis of Fe²⁺ concentration, GSH content, and MDA levels in spinal cord tissue (n = 6; Ordinary one-way ANOVA). P Quantitative analysis of DHE fluorescence intensity in spinal cord sections in (Q) (n = 6; Ordinary one-way ANOVA). Q Immunofluorescence images of DHE and DAPI in spinal cords in the indicated groups (scale bar: 100 μm). R Colocalization analysis of the fluorescence intensity between DHE and DAPI in (Q). All data are presented as the means ± SD; Error bars represent SDs; ∗∗p < 0.01, ∗p < 0.05, ns indicates no significance

Subsequently, we detected changes in the neuronal ferroptosis and oxidative stress phenotype after autophagy inhibition. Immunofluorescence results showed that compared with the SCI + AAV-Osbpl10 group, the expression of ferroptosis marker 4HNE was significantly enhanced and the anti-ferroptosis protein SLC7A11 was significantly downregulated in neurons of the SCI + AAV-Osbpl10/3MA group (Fig. 5G-J). Western blot analysis revealed that compared with the SCI + AAV-Osbpl10 group, the expression levels of ferroptosis-related markers ACSL4, 4HNE, and TP53 were significantly upregulated, while the expression levels of anti-ferroptosis markers SLC7A11 and GPX4 were significantly decreased in the SCI + AAV-Osbpl10/3MA group (Fig. 5K-L).

Meanwhile, the SCI + AAV-Osbpl10/3MA group exhibited increased Fe²⁺ accumulation, aggravated GSH depletion, and increased MDA production (Fig. 5M-O). DHE staining showed enhanced ROS aggregation (Fig. 5P-R). Behavioral assessments indicated that compared with the SCI + AAV-Osbpl10 group, the recovery of motor function was significantly impaired in the SCI + AAV-Osbpl10/3MA group, as evidenced by decreased BMS scores, reduced stride length, lowered maximum toe elevation height, and decreased MEP amplitude (Figure S4B-G). Histological examinations showed that the number of motor neurons and the density of NEUN⁺ neurons in the SCI + AAV-Osbpl10/3MA group were significantly reduced (Figure S4H-K).

Overexpression of OSBPL10 alleviates lysosomal membrane permeabilization (LMP) after SCI

It has been reported that LMP and the subsequent release of cathepsin B (CTSB) from lysosomes into the cytoplasm constitute a signaling pathway that induces ferroptosis [35], and lysosomal damage impairs autophagic flux in neurological diseases [7]. Given the significant role of OSBPL10 in inhibiting ferroptosis and restoring autophagic flux, we hypothesized that it may exert these effects by regulating lysosomal function after SCI. Volcano plot analysis of differentially expressed genes of bulk-RNA sequencing showed that OSBPL10 overexpression significantly downregulated the transcriptional levels of cathepsins A, B, C, D, L, S, and Z (Ctsa, Ctsb, Ctsc, Ctsd, Ctsl, Ctss and Ctsz) but had no significant effect on lysosome biogenesis-related genes (Lamp1, Lamp2, Atp6v1b2).

To further verify the relationship between OSBPL10 and lysosomal homeostasis, we isolated lysosomal and cytoplasmic protein fractions from spinal cord tissues. The results showed that the levels of cathepsins (CTSB, CTSL, CTSC) in the cytoplasmic fraction were higher in the SCI group than in the sham group, while the levels of these enzymes in the lysosomal fraction were lower (Fig. 6B-C). In addition, immunofluorescence staining revealed more neurons with diffused CTSB in SCI mice compared with the Sham group, indicating that LMP occurred after SCI and led to the leakage of lysosomal enzymes into the cytoplasm (Fig. 6D-E). In contrast, OSBPL10 overexpression significantly altered the distribution of cathepsins and reduced the number of neurons with diffused CTSB in SCI mice, suggesting that OSBPL10 overexpression alleviates SCI-induced LMP.

Fig. 6.

Overexpression of OSBPL10 attenuates lysosomal membrane permeabilization after SCI. A Volcano plot of transcriptome data from spinal cord tissue from SCI + AAV-NC and SCI + AAV-Osbpl10 group (n = 4). B Protein levels of CTSB, CTSL and CTSC in the cytoplasm and lysosomes extracted from the spinal cords of the indicated groups. The lysosomal membrane protein LAMP1 was used to identify the lysosomal fraction and as a loading control. GAPDH was used to identify the cytosolic fraction and as a loading control. C Quantitative analysis of the protein levels in (B) (n = 6; Ordinary one-way ANOVA). D Immunofluorescence staining and colocalization analysis of CTSB and LAMP1 with neurons in the ventral horn gray matter of spinal cord tissue in the indicated groups (scale bar: 25 μm for original pictures and 15 μm for enlarged pictures). E Quantitative analysis of the ratios of the diffuse CTSB cells in the anterior horn of the spinal cord tissue in the indicated groups (n = 6; Ordinary one-way ANOVA). F-G ELISA results of the activity of lysosomal enzymes CTSB and NAGLU in lysosomal or cytosolic part extracted from the spinal cord tissue (n = 6; Ordinary one-way ANOVA). H-I Immunofluorescence staining and colocalization analysis of OSBPL10 and LAMP1 with neurons in the ventral horn gray matter of spinal cord tissue in the indicated groups (scale bar: 25 μm for original pictures and 6 μm for enlarged pictures). J Quantitative analysis of the the ratio of OSBPL10 positive lysosomes to total lysosomes from each neuron in (H) (n = 6; Unpaired t-test). All data are presented as the means ± SD; Error bars represent SDs; ∗∗p < 0.01, ∗p < 0.05, ns indicates no significance

Subsequently, to further determine whether LMP affects lysosomal activity, we detected the activities of CTSB and α-N-acetylglucosaminidase (NAGLU) in lysosomal and cytoplasmic fractions. The results showed that SCI decreased the activities of CTSB and NAGLU in lysosomes while increasing their activities in the cytoplasmic fraction, and OSBPL10 overexpression significantly attenuated these SCI-induced changes (Fig. 6F-G).

It has been reported that lipid transfer protein mediates lipid trafficking and lysosomal repair by anchoring to the lysosomal surface. Immunofluorescence results showed that OSBPL10 overexpression significantly enhanced its co-localization with lysosomes, suggesting that it may directly participate in lysosomal repair (Fig. 6H-J).

Overexpression of OSBPL10 improves SCI in a phosphatidylserine (PS) not cholesterol-dependent manner

As a lipid transfer protein, OSBPL10 contains ORD and PH domains, which enable it to effectively bind various lipids and exert lipid transport functions [26]. Molecular docking analysis (see Methods section for details) was performed to explore the neuroprotective effect of OSBPL10 through its interaction with potential lipid molecules. It has been reported that a binding energy of -7.0 kcal/mol or lower indicates a stable binding mode between ligand and receptor, and such compounds have a higher probability of being potential active molecules [36]. As shown in Fig. 7A, the binding energies of OSBPL10 with cholesterol and PS were − 10.2 kcal/mol and − 7.1 kcal/mol, respectively, suggesting the existence of strong interactions.

Fig. 7.

Overexpressing neuronal OSBPL10 alleviates LMP, impairment of autophagic flux and ferroptosis in a PS-dependent manner after SCI. A Binding affinities of OSBPL10 protein to different lipid isoforms by Auto Dock Vina v.1.2.0 software. B-C The Osbpl10 protein is shown in green cartoon and white surface representations, with the interacting amino acid residues colored orange and cholesterol or phosphatidylserine displayed as pink or white sticks. D Schematic diagram of lipid synthesis for phosphatidylserine and cholesterol, and the target enzymes of lipid synthesis enzyme inhibitors. E Immunofluorescence staining and colocalization analysis of CTSB and LAMP1 with neurons in the ventral horn gray matter of spinal cord tissue in the indicated groups (scale bar: 25 μm for original pictures and 15 μm for enlarged pictures). F Western blot analysis of ATG5, VPS34, BECLIN1, SQSTM1/P62 and LC3 expression levels in the indicated groups. G Quantitative analysis of protein expression of (F) (n = 6; Ordinary one-way ANOVA). H Western blot analysis of 4HNE, TP53, ACSL4, SLC7A11 and GPX4 expression levels in the indicated groups. I Quantitative analysis of protein expression of (H) (n = 6; Ordinary one-way ANOVA). J Immunofluorescence images of DHE and DAPI in spinal cords in the indicated groups (scale bar: 100 μm). K Colocalization analysis of the fluorescence intensity between DHE and DAPI in (J) (n = 6; Ordinary one-way ANOVA). All data are presented as the means ± SD; Error bars represent SDs; ∗∗p < 0.01, ∗p < 0.05, ns indicates no significance

Furthermore, visual analysis from Fig. 7B showed that cholesterol binds to the active cavity of the OSBPL10 protein and is complementary in shape to the pocket formed by the protein. Further analysis revealed that cholesterol interacts with two amino acid residues (K479 and I601) of the target OSBPL10 protein. Visual analysis from Fig. 7C showed that phosphatidylserine binds to the active cavity of the OSBPL10 protein and is complementary in shape to the protein pocket. Additionally, phosphatidylserine interacts with six amino acid residues (K479, K723, L420, H537, V418, and N482) of the target OSBPL10 protein.

To further verify whether OSBPL10 exerts neuroprotective effects through cholesterol or PS, we performed rescue experiments using DS55980254 (abbreviated as DS5), a specific inhibitor of phosphatidylserine synthase 1 (PTDSS1), and Atorvastatin (abbreviated as Ator), an inhibitor of the key rate-limiting enzyme in cholesterol synthesis (Fig. 7D, Figure S5A). First, we directly detected the lysosomal lipid levels in different treatment groups. The results of PS and cholesterol ELISA experiments indicated that the levels of PS and cholesterol in lysosomes both decreased after SCI, while overexpression of OSBPL10 significantly promoted the enrichment of PS on lysosomes (Figure S6A-B). Moreover, additional treatment with DS55980254 or atorvastatin significantly downregulated cholesterol and PS levels on lysosomes, respectively. These results confirmed the effectiveness of the inhibitors; notably, overexpression of OSBPL10 only upregulated PS levels in lysosomes, suggesting that OSBPL10 may exert its effects solely through PS.

Immunofluorescence results showed that DS55980254 administration reversed the reduction in the number of neurons with diffused CTSB induced by OSBPL10 overexpression (Fig. 7E and Figure S6E). ELISA results demonstrated that DS55980254 administration abrogated the restorative effect of OSBPL10 overexpression on lysosomal enzyme activity (Figure S6C-D). In contrast, Atorvastatin had no effect on the regulatory role of OSBPL10 overexpression in lysosomal function. Western blot results showed that compared with the SCI + AAV-Osbpl10 group, the SCI + AAV-Osbpl10 + DS55980254 group exhibited downregulated autophagy-related and anti-ferroptosis-related proteins, as well as upregulated autophagic substrate proteins and pro-ferroptosis proteins (Fig. 7F-I). These findings indicate that DS55980254 effectively reversed OSBPL10-mediated autophagic flux restoration and ferroptosis inhibition.

Meanwhile, the SCI + AAV-Osbpl10 + DS55980254 group showed increased Fe²⁺ accumulation, aggravated GSH depletion, and increased MDA production (Figure S6F-H). DHE staining revealed enhanced ROS aggregation in the SCI + AAV-Osbpl10 + DS55980254 group (Fig. 7J-K, S6I). However, Atorvastatin did not affect the regulatory effects of OSBPL10 overexpression on oxidative stress. Behavioral assessments indicated that compared with the SCI + AAV-Osbpl10 group, the recovery of motor function was significantly impaired in the SCI + AAV-Osbpl10 + DS55980254 group, as evidenced by decreased BMS scores, reduced stride length, lowered maximum toe elevation height, and decreased MEP amplitude (Figure S5B-G). Histological examinations showed that the number of motor neurons and the density of NEUN⁺ neurons in the SCI + AAV-Osbpl10 + DS55980254 group were significantly reduced (Figure S5H-K). In contrast, Atorvastatin did not affect the improvement of motor function in SCI mice by OSBPL10 overexpression. Collectively, these results demonstrate that OSBPL10 overexpression ameliorates SCI in a PS-dependent, but not cholesterol-dependent manner.

Discussion

Traumatic spinal cord injury (SCI) leads to irreversible neurological damage, which is pathologically divided into two phases: primary injury and secondary injury [37]. To date, numerous studies have focused on the secondary injury phase, among which neuronal autophagy-lysosomal dysfunction is the primary pathological mechanism in the early stage of SCI [8]. The lysosomal repair function mediated by OSBPs has been reported in many basic medicine [26, 28], but studies on lipid transfer protein in SCI remain unreported. The present study found that injured spinal cord induces increased LMP, impaired autophagic flux, and activated ferroptosis, which further lead to neuronal damage and impaired motor function recovery. Notably, we confirmed that OSBPL10 may act as a protective factor in SCI: overexpression of OSBPL10 improves neuronal lysosomal function, restores autophagic flux, and inhibits ferroptosis in a PS-dependent manner (Fig. 8). Thus, OSBPL10 is expected to become a new target for ameliorating SCI.

Fig. 8.

Schematic diagram showing the potential protective effect of OSBPL10 against SCI

In recent years, a growing number of studies have focused on non-apoptotic cell death, especially ferroptosis [38–40]. Ferroptosis induces intracellular oxidative stress in an iron-dependent manner. Studies on the brain have shown that the susceptibility of neurons to oxidative stress is neurospecific [38]. In addition, neurodevelopmental analysis indicates that neurons are more sensitive to lipid peroxidation due to their unique lipid composition, suggesting that neurons may be more prone to ferroptosis [41]. Using mature molecular biology methods, this study confirmed that OSBPL10 overexpression significantly reduced the expression levels of pro-ferroptotic proteins and improved tissue lipid peroxidation after SCI.

Our research team has previously explored the role of autophagy in SCI [42, 43]. Basal levels of autophagy are crucial for maintaining cellular homeostasis and are essential for the function and survival of nerve cells. Current studies have shown that changes in autophagic flux after CNS injury vary with the location and severity of the injury [44], so autophagy may exert either positive or negative effects after injury. However, in either case, restoring and/or enhancing autophagic flux seems to improve cell survival and accelerate functional recovery after injury [17, 45], implying that the autophagic pathway may be a promising therapeutic target for SCI. This study detected autophagic flux and autophagy-related protein levels after SCI, and the results showed that OSBPL10 overexpression increased autophagic flux after SCI. Previous studies have confirmed that activating autophagy can inhibit cellular ferroptosis in CNS trauma [46, 47], and the results of this study also support this view—when autophagic flux is blocked, ferroptosis levels increase significantly, indicating that OSBPL10 overexpression may reduce neuronal ferroptosis in SCI by activating autophagic flux.

Increased LMP is an important type of lysosomal damage, caused by ROS, intralysosomal Fenton reaction, and other cellular stresses [48]. In recent years, researchers have begun to pay attention to the role of LMP in CNS trauma. A large body of evidence indicates that lysosomal damage leads to neuronal autophagosome accumulation and inhibits autophagic flux in neurological diseases [7, 8]. Given the previous finding that OSBPL10 overexpression restores autophagic flux after SCI and the lysosomal membrane repair function of its family proteins, this study hypothesized that OSBPL10 overexpression may interfere with the autophagic process by affecting LMP. This hypothesis was verified—OSBPL10 overexpression prevented the translocation of cathepsins from lysosomes to the cytoplasm and restored autophagy-lysosome system function.

Nevertheless, the present study has several limitations. As reported, OSBPs family members mediate lipid trafficking by forming membrane contact sites with the endoplasmic reticulum (ER). However, the technical challenges associated with precise ER localization in animal models precluded us from providing direct evidence to verify whether OSBPL10 facilitates PS translocation to lysosomes via the ER. In future studies, we will employ proximity ligation assays or immunofluorescence electron microscopy to visualize the interaction between OSBPL10 and ER/lysosome MCS-related proteins in vivo, and to directly identify the subcellular spatial colocalization of OSBPL10 with the ER and lysosomes in animal models. Second, the endosomal sorting complexes required for transport (ESCRT) machinery constitutes another critical membrane repair pathway. Thus, future experiments should detect ESCRT-related molecular markers to evaluate its reparative potential in cells after SCI. Third, this study focused exclusively on neuronal lysosomal repair pathways, while the corresponding mechanisms in other cell types (e.g., microglia and astrocytes) remain unexplored and merit further investigation. Finally, although we found the OSBPL10-PS axis as a novel signaling pathway mediating lysosomal membrane repair, more direct targeted intervention strategies are required; for instance, direct supplementation of PS to damaged lysosomes may represent a promising approach to enhance lysosomal repair efficiency after SCI.

In conclusion, this study confirms that OSBPL10 is a protective factor for SCI. Mechanistically, we clarified that the lysosomal membrane repair effect mediated by the OSBPL10-PS axis is closely associated with ferroptosis caused by impaired neuronal autophagic flux. This regulatory axis ultimately improves neuronal survival after SCI, providing a solid experimental basis for the development of novel pathway-targeted therapies for CNS injury with lysosome-targeted intervention.

Materials and methods

CSF collection from patients

CSF samples were collected from the Third Affiliated Hospital of Sun Yat-sen University, including a total of 109 patients who underwent decompression followed by spinal stabilization surgery. The patients were stratified according to the injury phase and ASIA impairment scale. All cases were diagnosed as SCI based on the combination of clinical symptoms, electrophysiological examinations, X-ray, and MRI analyses, including patients with complete or incomplete traumatic spinal cord injury in the cervicothoracic segment. CSF samples in the control group were obtained from 18 patients with spinal stenosis or lumbar disc herniation without direct spinal cord injury. CSF was collected by lumbar puncture after overnight fasting, cryopreserved at -80 °C, and then transferred to the Clinical Biobank of the Third Affiliated Hospital of Sun Yat-sen University for long-term storage. This study protocol ([2019]-08, [2019]-09, [2019]-10, [2020]-02-005-01) was approved by the Ethics Committee of the Third Affiliated Hospital of Sun Yat-sen University, and all participants signed written informed consent forms.

Animals and ethics statement

Animal use and all experiments involving animals were approved by the Ethical Committee of Guangzhou Seyotin Biotechnology Co., Ltd. (SYT2025057). Female animals are used in most SCI experimental studies, as they have shorter urethras than male mice. Thus, manual bladder expression for urine retention after SCI and overall animal care are more straightforward in females [49, 50]. Female adult C57BL/6 mice aged 8 weeks were purchased from Jiangsu GemPharmatech. All the animals were provided free access to food and water and kept in a colony room under conditions of constant temperature (25 °C), humidity (70%), and lighting (12 h light/12 h dark cycle) at the Sun Yat-sen University Animal Center.

Animal model of SCI

Prior to surgery, each animal was anesthetized via intraperitoneal injection of 1% (w: v) pentobarbital sodium at a dose of 50 mg/kg. Subsequent to anesthesia, a conventional laminectomy was performed at the T9-T10 vertebral level to expose a circular segment of the dura mater. Mice with exposed spinal cords were then subjected to a weight-drop injury to induce mild spinal cord injury, in accordance with the manufacturer’s instructions for the device (W.M. KECK+, Model III, USA). Briefly, a 5 g weight was dropped from a height of 30 mm onto the exposed spinal cord. Following injury, the muscle and skin layers were sutured sequentially using 4 − 0 silk sutures. For post-SCI care, manual bladder expression was performed three times daily to facilitate urination until the recovery of bladder reflexes. Mice in the Sham group underwent the same anesthesia and laminectomy procedures but without sustaining spinal cord trauma.

Animal treatment and groups

Part 1: C57BL/6 mice were randomly divided into the Sham group, SCI group, SCI + AAV-NC group and SCI + AAV-Osbpl10 group. Part 2: C57BL/6 mice were randomly divided into the SCI + AAV-NC group, SCI + AAV-NC+3MA group, SCI + AAV-Osbpl10 group and SCI + AAV-Osbpl10 + 3MA group. Part 3: C57BL/6 mice were randomly divided into the SCI + AAV-NC group, SCI + AAV-Osbpl10 group, SCI + AAV-Osbpl10 + DS55980254 group and SCI + AAV-Osbpl10 + Atorvastatin group. The SCI + AAV-Osbpl10 + DS55980254 group and SCI + AAV-Osbpl10 + Atorvastatin group were given 100 µL of DS55980254 (20 mg/kg/day) or Atorvastatin (10 mg/kg/day) in saline by oral gavage, whereas the mice in SCI + AAV-NC group and SCI + AAV-Osbpl10 group were given an equal volume of saline with the same protocol. The dose and time of DS55980254 and atorvastatin administration were selected according to a previous study [51–53]. The chemical reagents used are listed in Table 1.

Table 1.

Antibodies and chemical reagents resources

Antibodies	Manufacturer	Identifier
Anti-TP53	Cell Signaling Technology	2524
Anti-ACSL4	Abcam	ab155282
Anti-SLC7A11	Abcam	ab175186
Anti-GPX4	Abcam	ab125066
Anti-4HNE	Abcam	ab46545
Anti-VPS34	Proteintech	12452-1
Anti-ATG5	Affinity	DF6010
Anti-BECLIN1	Cell Signaling Technology	3738
Anti-CTSL	Santa Cruz Biotechnology	sc-390,385
Anti-CTSB	Abcam	ab214428
Anti-CTSC	Santa Cruz Biotechnology	sc-10,778
Anti-GAPDH	Affinity	AF7021
Anti-LAMP1	Abcam	ab25245
Anti- NEUN	Abcam	ab177487
Anti- NEUN	Abcam	ab134014
Anti-GFAP	Cell Signaling Technology	12,389 S
Anti-LC3	Proteintech	14600-1-ap
Anti-SQSTM1/p62	Abmart	T55546
Anti-OSBPL9	Santa Cruz Biotechnology	sc-398,961
Anti-OSBPL10	Proteintech	15491-1-AP
Anti-OSBPL11	Proteintech	11318-1-AP
Anti-mouse IgG, HRP-linked Antibody	Proteintech	SA00001-1
Anti-rabbit IgG, HRP-linked Antibody	Proteintech	SA00001-2
Goat anti-mouse IgG, Alexa Fluor 488	Invitrogen	A-11,001
Goat anti-rabbit IgG, Alexa Fluor 488	Invitrogen	A-11,008
Goat anti-mouse IgG, Alexa Fluor 594	Invitrogen	A-11,005
Goat anti-rabbit IgG, Alexa Fluor 594	Invitrogen	A-11,012
Goat anti-rat Ig(H&L), Alexa Fluor 647	Abcam	ab150167
Chemicals Reagents	Source	Identifier
RAPA	MedChemExpress	HY-10219R
3MA	MedChemExpress	HY-19,312
DS55980254	MedChemExpress	HY-158,145
Atorvastatin	MedChemExpress	HY-B0589

Adeno-associated virus (AAV) vector packaging

AAV9-hSyn-Osbpl10 (primer sequence: forward:5’ TCGAGAAGGTACCGGAATTCATGGAGAGGGCGGCCCAG-3’ and reverse: 5’ TCACCATGGTGGCGGGATCCGTGTGTCTTCCAGAGGGGGT‐3’) and AAV9-hSyn-NC were constructed and packaged by Shanghai Genechem. The detailed protocols were conducted according to previously published instructions [54]. Quantitative PCR (qPCR) demonstrated that AAV9-hSyn-Osbpl10 had 4.21E + 13 genomic copies per mL, and AAV9-hSyn-NC had 4.72E + 13 genomic copies per mL.

AAV vector injection

AAVs may need several days to take effect in vivo, and according to previous studies [55, 56], 2 weeks before SCI is considered an appropriate time point to carry out AAV injection. AAV9 harboring Osbpl10 gene (AAV9-hSyn-Osbpl10) and its control vector (AAV9-hSyn-NC) were used to infect the spinal cords of the mice via in situ injection. After injection, the needle was left in place for 3 min before removal to prevent leakage of the virus. Following injection, the muscle and skin over the exposed spinal cord were sutured. Later, all the mice were placed on a warm blanket for postoperative recovery. No mice exhibited hind limb paralysis or paresis after injection.

Quantitative real-time polymerase chain reaction

Total RNA was extracted from spinal cord tissues using TRIzol reagent in accordance with the kit manufacturer’s instructions. Quantitative analysis was performed via a two-step reaction protocol, consisting of reverse transcription (RT) and polymerase chain reaction (PCR). Each RT reaction system had a total volume of 10 µL, containing 0.5 µg of RNA, 2 µL of 5× TransScript All-in-one SuperMix for qPCR, and 0.5 µL of gDNA Remover. The reaction was carried out on a GeneAmp PCR System 9700 (Applied Biosystems, USA) under the following conditions: 42℃ for 15 min, followed by 85℃ for 5 s. The 10 µL RT reaction mixture was diluted 10-fold with nuclease-free water and stored at -20℃.

Real-time PCR was performed on a LightCycler 480 II Real-Time PCR System (Roche, Switzerland). The total volume of each reaction system was 10 µL, comprising 1 µL of cDNA, 5 µL of 2× PerfectStart™ Green qPCR SuperMix, 0.2 µL of forward primer, 0.2 µL of reverse primer, and 3.6 µL of nuclease-free water. Reactions were conducted in multi-well optical plates (Roche) with the cycling parameters: initial denaturation at 94℃ for 0.5 min, followed by 45 cycles of denaturation at 94℃ for 5 s and annealing/extension at 60℃ for 30 s. All samples were analyzed in triplicate.

Upon completion of PCR cycling, melting curve analysis was performed to verify the specific amplification of the expected PCR products. Based on the mRNA sequences retrieved from the NCBI database, the following primer sequences were designed in our laboratory and synthesized by Tsingke Biotechnology Co., Ltd. (Qingdao, China):

Sqstm1: Forward 5ʹ-GATAGCCTTGGAGTCGGT-3ʹ; Reverse 5ʹ-AAATGTGTCCAGTCATCGTC-3ʹ; β-actin: Forward 5ʹ-GGCTCCTAGCACCATGAAGA-3ʹ; Reverse 5ʹ-AGCTCAGTAACAGTCCGCC-3ʹ. Finally, β-actin was used as the reference gene, and the relative mRNA expression levels of the target genes were calculated using the 2-ΔΔCt method.

Immunofluorescence staining

After cardiac perfusion with saline and 4% paraformaldehyde, the T9 spinal cord segments of anesthetized mice were completely removed. The tissues were fixed with 4% paraformaldehyde for 3 h, dehydrated with a gradient of sucrose solutions for 2 days, and then embedded in OCT compound (SAKURA, 4583). The embedded tissues were cut into 12-µm thick vertical or transverse sections using a microtome. For immunofluorescence staining, the OCT on the tissue surface was first rinsed off with PBS. The sections were permeabilized with 0.4% Triton X-100 (dissolved in PBS) for 2 h, blocked with 0.5% BSA (dissolved in PBS) for 1 h, and then incubated with primary antibodies against LC3 (1:200), SQSTM1 (1:200), CTSB (1:200), NEUN (1:1000), OSBPL10 (1:200), 4HNE (1:200), SLC7A11 (1:200), LAMP1 (1:500) and GFAP (1:1000) at 4 °C overnight. The next day, the sections were incubated with secondary antibodies at room temperature for 2 h. Finally, after being rinsed with water, the sections were mounted with ProLong Gold Antifade Reagent containing DAPI (Thermo Fisher Scientific, P36971). Immunofluorescence images of LC3 in each neuron were quantified to count specific LC3-II puncta with diameters larger than 0.2 μm but smaller than 10 μm, a size range consistent with autophagosomes [57]. The primary and secondary antibodies used are listed in Table 1.

Subcellular fractionation and preparation of lysosome-enriched fractions

For subcellular fractionation and lysosome enrichment, spinal cord tissues (approximately 100 mg per sample) were processed using the Lysosome Enrichment Kit for Tissue and Cultured Cells (BestBio, 003065) with a modified differential centrifugation protocol: For cell samples, cells were first harvested by centrifugation at 500×g for 5 min at 4 °C, washed twice with ice-cold PBS (with thorough supernatant removal after each wash), resuspended in 400 µl of ice-cold Reagent A, and incubated on ice for 10 min; for tissue samples, fresh spinal cord tissues were rinsed with ice-cold PBS to remove blood contaminants before homogenization. Both cell suspensions and tissue samples were then homogenized on ice using a Dounce tissue grinder with 30–40 gentle strokes to ensure complete cell disruption while preserving organelle integrity. The resulting homogenate was subjected to sequential differential centrifugation at 4 °C: First, the sample was centrifuged at 1,000×g for 5 min, and then, the retained supernatant was centrifuged at 3,000×g for 10 min, followed by centrifugation at 5,000×g for 10 min. The cytosolic fraction was collected from the supernatant after centrifugation at 20,000–30,000×g for 20 min, and the retained pellet was resuspended in 400 µl of ice-cold Reagent B and centrifuged again at 20,000–30,000×g for 20 min. The final pellet was resuspended in Lysosome Storage Buffer C to obtain the lysosome-enriched fraction. The isolated lysosome-enriched fractions and cytosolic fractions were stored at -80 °C for subsequent experiments.

Immunoblotting assay

The tissue was rinsed three times with ice-cold PBS and lysed in RIPA lysis buffer (Beyotime, P0013B) containing protease inhibitors (Beyotime, P1005) for 10 min on ice. The cell lysates were cleared by centrifugation at 15,000 × g for 10 min at 4 °C. The supernatants were separated by SDS-PAGE and transferred to polyvinylidene fluoride/PVDF membranes (Millipore, ISEQ00010). The membranes were blocked with 5% skim milk in Tris-buffered saline (Cell Signaling Technology, 9999 S) and incubated with primary and HRP-conjugated secondary antibodies. The band signals were observed and examined via a Bio-Rad ChemiDoc™ XRS+ Imaging System using an enhanced chemiluminescence (ECL) immunedetection instrument (Beyotime, P0018FS). The relative intensities of the bands were quantified using ImageJ (https://imagej.nih.gov). The primary and secondary antibodies used are listed in Table 1.

Enzyme-linked immunosorbent assay (ELISA)

The level of OSBPL10 in CSF was detected using a human OSBPL10 ELISA kit (abx381984, Abbexa). Add 100 µL of CSF sample to the corresponding reaction wells and operate in accordance with the manufacturer’s instructions. The activities of CTSB (Cusabio, CSB-EL006185MO) and NAGLU (MSKBIO, KT-22858) were detected by using ELISA kits according to the manufacturer’s protocols. The content of PS (bioswamp, MU30938) and cholesterol (Saipei Biotech, SP14121) were detected by using ELISA kits according to the manufacturer’s protocols. The optical density of the samples for quantification of OSBPL10, CTSB, NAGLU, PS and cholesterol was read by a microplate reader at 550 nm with a correction wavelength of 450 nm.

Determination of malondialdehyde (MDA), glutathione (GSH), and iron content

The levels of MDA and GSH were determined using commercial assay kits (MDA: Beyotime Biotechnology Co., Ltd., Shanghai, China; GSH: Sigma-Aldrich Co., Ltd.), following the manufacturers’ instructions. For the quantification of Fe²⁺ in mouse spinal cord tissues, samples from each group were homogenized and processed according to the standard protocol using a specific assay kit (Elabscience Biotechnology Co., Ltd., Wuhan, China). The absorbance was measured with a microplate reader, and the Fe²⁺ concentration was calculated based on the standard calibration curve.

Dihydroethidium (DHE) staining

Fresh frozen sections were incubated with 2 µM DHE (Thermo Fisher Scientific Inc., USA) in a humidified chamber at 37 °C for 30 min in the dark. Fluorescent images were captured using a fluorescence microscope, and the level of ROS was assessed accordingly.

Nissl and HE staining

Nissl staining was performed according to the manufacturer’s instructions (Nissl Staining Kit; Solarbio Science & Technology, G1430). After the frozen transverse sections of the spinal cord obtained from the mice 28 days after surgery were soaked in PBS, they were immersed in cresyl violet staining solution for 10 min. They were subsequently thoroughly rinsed with deionized water and differentiated with Nissl differentiation solution for 4 to 10 s until most of the staining was removed. Next, the transverse sections of the spinal cord were dehydrated in xylene and sealed with neutral resin for observation. Finally, images were acquired using an optical microscope (Olympus, Tokyo, Japan). The number of Nissl-stained ventral motor neurons in the anterior horn was counted manually in a double-blind manner. Heart, liver, spleen, lung, and kidney tissues were also collected, and HE staining was performed at week 4 post-injury.

Basso mouse scale scoring

On days 0, 1, 3, 7, 14, 21, and 28 following SCI, BMS scoring was performed. The BMS ranges from 0 to 9, where 0 indicates complete paralysis and 9 indicates fully normal motor function [58]. Each assessment was conducted by two trained observers who were blinded to the genotype and treatment. The BMS score of mice in an open field was determined on the basis of hindlimb joint movements, body weight support, trunk position and stability, gait coordination, paw position, and tail control.

BehaviorAtlas 2D gait motion data acquisition and analysis

After adjusting the lighting conditions and preparing the environment before the experiment, the mice 28 days after the surgery were placed in the experimental field, and videos were collected using a mobile phone device at a frame rate of 60 fps and a resolution of 848 × 480. The collection duration for each mouse was approximately 3 min. The collected videos were used for subsequent two-dimensional pose estimation and gait analysis. The BehaviourAtlas Pose Labeler (V1.0.0, Guangdong BayONE Scientific Co., Ltd.) was used to label the key body parts in the videos for two-dimensional poses, and the BehaviorAtlas 2D-AI Motor Analyzer (V1.0.0, Guangdong BayONE Scientific Co., Ltd.) was used to preprocess the pose data and extract gait parameters. Lateral views of the hindlimbs and the angles of rotational movements were obtained in a double-blind manner through MATLAB software. The function of the motor nervous system of the lower limbs was tested by motor evoked potential assessment on the 28th day after SCI injury. A BL-420 A/F Data Acquisition Analysis System (TECHMAN SOFT) was used. The motor cortex of each mouse was exposed and touched by stimulating electrodes. The recording electrodes were placed on the contralateral sciatic nerve. The latency and amplitude of the first evoked peak were selected as parameters for assessing the function of the motor neuron.

Bulk-RNA sequencing and differential expressed gene analysis

Bulk-RNA sequencing data for the SCI mouse post injury on day 0, 1, 3, 5 and 7 were downloaded from GEO (website: https://www.ncbi.nlm.nih.gov/sra/?term=PRJNA753793; SRA: PRJNA753793). The “DESeq2” package in R (4.4.1) software was used to analyse the differentially expressed genes (DEGs) in the high-throughput sequencing data.

Total RNA was extracted using TRIzol reagent (provided by Life Technologies Corp.) and treated with DNase to eliminate genomic DNA contamination. mRNA was purified with the NEBNext PolyA mRNA Magnetic Isolation Module (supplied by New England Biolabs, Ipswich, MA, USA), followed by the construction of RNA sequencing libraries using the NEBNext Ultra Directional RNA Library Prep Kit (also from New England Biolabs). These libraries were subjected to paired-end 2 × 150 sequencing using Illumina sequencing technology. Raw sequencing data were processed with Cutadapt and Trimmomatic to remove sequencing adapters, discard reads shorter than 35 bp, and filter out low-quality sequences, resulting in high-quality clean reads. Quality control checks of the refined reads were performed using FastQC. The optimized reads were aligned to the mouse reference genome sequence (GRCm38.p6) via HISAT2. StringTie was used to calculate the Fragments Per Kilobase of transcript per Million mapped reads (FPKM) value for each gene to estimate gene expression levels. Differential gene expression analysis was conducted using the edgeR v3.24.2 package in R language, and p-values were adjusted using the false discovery rate control method for multiple testing scenarios to assess the significance of differential expression. In this study, genes with an adjusted p-value < 0.05 and an absolute log2 fold change (|log2FC|) ≥ 1 were selected for further analysis. Gene annotations were obtained from the Ensembl Genome Browser Release 96 database (URL: http://www.ensembl.org/index.html). Functional annotation and enrichment analysis of KEGG pathways for differentially expressed genes were performed using ClusterProfiler. The entire process of RNA sequencing was completed by LuMing Biological Technology Co. Ltd. (Shanghai, China).

Preparation and preprocessing for molecular docking

Use AlphaFold3 (https://alphafoldserver.com/) for homologous modeling of the target protein structure, and download the molecular structure files of ligands cholesterol and phosphatidylserine from the PubChem database. Use PyMOL 2.3.0 software to perform operations such as removing water molecules and original ligands from the downloaded target protein. Use Chem3D (2020 version) software to perform molecular mechanics optimization on the optimal conformation of small molecules, and finally obtain the energy-minimized optimal conformation. Docking process: Use AutoDock Tools 1.5.6 to preprocess the preprocessed target protein to obtain pdbqt files. Use AutoDock Vina v.1.2.0 software to perform molecular simulation docking between the target protein and ligand molecules. The docking algorithm is the Lamarckian genetic algorithm, the docking mode is semi-flexible docking, exhaustiveness is set to 16, and the maximum number of output conformations is set to 9.

Statistical analysis and reproducibility

All experiments were carried out with at least three biological replicates, and successful reproducibility was shown. All the data are reported as the mean ± standard deviation (SD) of at least three independent experiments. All sample sizes are presented in the figure legends. Statistical analysis between two groups was performed using an unpaired t test. Statistical analysis between multiple groups was performed by one-way ANOVA or Tukey’s multiple comparison test. Statistical analysis involving multiple variables and analysis of their interactive effects was performed by two-way ANOVA. Post-hoc comparisons were performed using Tukey’s honestly significant difference (HSD) test to evaluate pairwise differences between groups after significant two-way ANOVA interactions. All the data were analysed using GraphPad Software. A two-sided p value < 0.05 was considered to indicate statistical significance. The level of significance is defined as p < 0.05 (*), p < 0.01 (**), and ns indicates no significance, p > 0.05.

Abbreviations

AAV

Adeno-associated virus

ATGs

Autophagy-related proteins

Ator

Atorvastatin

BMS

Basso mouse scale

CNS

Central nervous system

CTSB

Cathepsin B

CSF

Cerebrospinal fluid

DHE

Dihydroethidium

DS5

DS55980254

ELISA

Enzyme-linked immunosorbent assay

ER

Endoplasmic reticulum

ESCRT

Endosomal sorting complexes required for transport

GSEA

Gene Set Enrichment Analysis

HE

Hematoxylin-eosin

3MA

3-methyladenine

4HNE

4-hydroxynonenal

KEGG

Kyoto Encyclopedia of Genes and Genomes

LC3

Microtubule-associated protein light chain 3

LMP

Lysosomal membrane permeabilization

MDA

Malondialdehyde

MEPs

Motor evoked potentials

NAGLU

α-N-acetylglucosaminidase

OSBPs

Oxysterol-binding proteins

ORD

OSBP-related domain

PI4P

Phosphatidylinositol 4-phosphate

PS

Phosphatidylserine

qPCR

Quantitative real-time polymerase chain reaction

ROS

Reactive oxygen species

SCI

Spinal cord injury

SLC7A11

Solute carrier family 7 member 11

TBI

Traumatic brain injury

TP53

Tumor protein p53

VPS34

Vacuolar protein sorting 34

Authors’ contributions

Conceptualization: Limin Rong, Haojie Zhang, Senyu Yao; Methodology: Yiqian Luo, Tianlun Zhao, Yu Kang, Daoqiang Huang, Xuantao Hu; Investigation: Haojie Zhang, Tianlun Zhao, Daoqiang Huang, Mao Pang; Visualisation: Yiqian Luo, Jiawei Di, Yu Kang; Supervision: Limin Rong, Bin Liu, Mao Pang; Writing-original draft: Haojie Zhang, Senyu Yao; Writing-review and editing: Haojie Zhang, Senyu Yao, Yu Kang.

Funding

This work was supported by the National Natural Science Foundation of China (82572748, U22A20297, 82172433, 82372400, 823B2060); Key Technologies Research and Development Program of Guangzhou Municipality (202206060003); Guangdong Basic and Applied Basic Research Foundation (2024A1515012766, 2023A1515010313); Guangzhou Municipal Science and Technology Program key projects (2023A03J0203); Guangdong Provincial Drug Administration Science and Technology Innovation Project (2024ZDZ13); Postdoctoral Fellowship Program of China Postdoctoral Science Foundation (GZB20240904) and the China Postdoctoral Science Foundation (2024M763769, 2025T180653, 2025M782367).

Data availability

Authors declare that data supporting the findings of this study are available within the article and its Supplementary information files or from the corresponding author upon request.

Declarations

Competing interests

The authors declare no competing interests.