Cel-LDH dual-functional nanotherapy simultaneously promotes microglial efferocytosis and alleviates neuronal apoptosis for spinal cord injury recovery Authors: Nuo Chen, Lufeng Yao, Le Zou, Qiuchen Wang, Yaning Li, Junjun Luo, Mengting Zhan, Kailiang Zhou, Jian Xiao, Songlin Tong, Hongyu Zhang, Liangliang Yang, Lu Ge Published: 2026 ABSTRACT AbstractTraumatic spinal cord injury (SCI) can cause severe central nervous system damage. Efferocytosis, an intrinsic regulatory mechanism through which microglia eliminate apoptotic cells, is suppressed because of the local ischaemic and hypoxic microenvironment after spinal cord injury (SCI). In addition, hypoxia and reoxygenation (H/R) trigger mitochondrial respiratory chain electron leakage, leading to the massive generation of mitochondrial reactive oxygen species (mtROS), impairing the energy supply and causing oxidative stress damage in neurons, which induces neuronal apoptosis. Synergistic therapies targeting efferocytosis and neuronal apoptosis are important for recovery after SCI. In this study, by loading celastrol in a layered double hydroxide, a multifunctional nanoparticle, Cel-LDH, was developed to facilitate SCI recovery by concurrently normalizing efferocytosis homeostasis and inhibiting neuronal apoptosis. Cel-LDH composite nanoparticles strongly scavenged ROS and preserved mitochondrial homeostasis, thus regulating apoptosis-related proteins, including Bax and Bcl-2, and effectively inhibiting the apoptotic process of neurons. Furthermore, Cel-LDH nanoparticles promoted autophagy, inhibited the cyclic GMP-AMP (cGAMP) synthase (cGAS)-stimulator of interferon genes (Sting) pathway to modulate the anti-inflammatory phenotypic transformation of microglia, and subsequently restored the efferocytosis homeostasis of microglia, thus suppressing inflammatory cascades and creating favourable microenvironments for the repair of SCI. This approach not only remodels the dysfunctional efferocytosis capacity but also inhibits neuronal apoptosis, providing a new therapeutic strategy for SCI.Graphical Abstract Cel-LDH composite nanoparticles simultaneously promote microglial efferocytosis and alleviate neuronal apoptosis to promote recovery from spinal cord injury. Supplementary InformationThe online version contains supplementary material available at 10.1186/s12951-026-04641-6.Keywords: Spinal cord injury, Celastrol, Mitochondria, Efferocytosis, Apoptosis Introduction Spinal cord injury (SCI) is the most serious complication of spine injury and frequently causes profound impairment of both motor and sensory functions below the injury level [1, 2]. SCI not only causes serious physical and psychological harm to patients in the short term but also induces permanent sensory and motor deficits [3, 4]. SCI can be divided into two stages: primary injury and secondary injury. The primary injury phase typically occurs in acute crush contusions, lacerations, and shear injuries, followed by a secondary injury phase, which consists of a cascade of responses such as oxidative stress, neuroinflammation and so on [5–7]. Given the unpredictability and uncontrollability of primary injury, treatment of SCI focuses on blocking the generation and development of the amplifying effects of this secondary injury cascade [8]. Increasing evidence suggests that efferocytosis homeostasis is impaired after SCI, which induces inflammation and actively polarizes microglia to the M1 phenotype [9]. Efferocytosis, an intrinsically regulated pathway through which macrophages/microglia recognize, phagocytose, and clear apoptotic cells, is essential for maintaining tissue homeostasis [10, 11]. After SCI, many neurons undergo apoptosis, and these dead cells, if not cleared in a timely manner, continuously release proinflammatory factors (e.g., IL-1β and TNF-α), creating a chronic inflammatory microenvironment and exacerbating secondary injury [12]. However, in cases of severe SCI, the efferocytosis clean-up process is often dysfunctional or overwhelmed [13]. The number of apoptotic cells increases dramatically because of severe injury, far exceeding the processing capacity of phagocytes; in addition, inhibitory signalling molecules (e.g., “do not eat me” signals such as CD47) in the local microenvironment of the injury further impede the efferocytosis process [10, 14–16]. Together, the above factors lead to a large accumulation of apoptotic cells, resulting in a vicious cycle of “insufficient clearance–exacerbated inflammation–secondary injury–increased apoptosis”, which severely inhibits axon regeneration and myelin repair and ultimately hinders the recovery of neurological function. During the early phase of injury, stress signals such as hypoxia-ischemia and oxidative stress can rapidly activate the autophagic process, which manifests as an increase in autophagosome formation (e.g., an increase in LC3-II) [17, 18]. However, as the injury progresses to the chronic phase, autophagic flow becomes severely impaired, as evidenced by blocked autophagosome‒lysosome fusion and decreased lysosomal degradation (e.g., an increase in p62) [19–21]. This impairment of autophagic flow leads to impaired clearance of damaged mitochondria, which results in massive leakage of mitochondrial DNA (mtDNA). mtDNA, a typical damage-associated molecular pattern (DAMP), is recognized by cyclic GMP-AMP synthase (cGAS) in the cytoplasm, which in turn activates the Sting-dependent interferon-type I signalling pathway and promotes microglial cell polarization towards the proinflammatory M1 phenotype, further amplifying the neuroinflammatory response [22]. In addition, H/R injury drives mitochondrial electron leakage and a surge in mtROS, initiating a cascade of energy failure and oxidative stress that culminates in neuronal apoptosis [23]. This excessive mtROS not only disrupts mitochondrial energy production and exacerbates oxidative stress but also promotes the accumulation of proapoptotic proteins such as Bax on the outer mitochondrial membrane. The subsequent opening of the mitochondrial permeability transition pore (mPTP) facilitates the release of cytochrome c (Cytc) into the cytosol, where it forms an apoptosome with Apaf-1 and procaspase-9. This leads to caspase-9 activation, triggering a cascade that culminates in the cleavage of cellular substrates by caspase-3 and ultimately induces neuronal apoptosis, contributing to irreversible neurological deficits [24–26]. Therefore, timely improvement of mitochondrial function is considered one of the key strategies in the treatment of SCI. Therefore, how to effectively restore efferocytosis homeostasis in microglia and inhibit neuronal apoptosis are two important aspects of SCI recovery. Celastrol (Cel), a natural pentacyclic triterpenoid compound isolated from the traditional Chinese herb Tripterygium wilfordii, exhibits a broad range of biological activities, including anti-inflammatory, antitumour, and anti-obesity effects. It has been successfully applied in the treatment of various autoimmune and inflammatory diseases, such as rheumatoid arthritis and systemic lupus erythematosus [27–30]. Notably, Cel has shown remarkable potential for mitochondrial protection and has received widespread attention [31, 32]. Cel binds to the nuclear receptor Nur77 in mitochondria and induces mitochondrial autophagy in macrophages, thereby exerting an anti-inflammatory effect [27, 33]. Recent studies have also shown that Cel can ameliorate mitochondrial dysfunction in neurons by interacting with the EPAC-1 protein, thereby attenuating secondary brain damage after cerebral haemorrhage [34]. These results suggest that Cel, as a promising mitochondrial protective agent, can target the repair of damaged mitochondrial function, promote autophagy, and effectively block the inflammatory cascade triggered by mtDNA escape. Thus, it can inhibit M1 microglial polarization, regulate the polarization of microglia to the anti-inflammatory M2 phenotype and alleviate the neuroinflammatory storm. However, the effect of Cel on the efferocytosis process is unclear and needs to be further explored. On the basis of its regulatory effects on mitochondrial function and microglial polarization, we speculate that this drug may restore the phagocytosis of microglia, thus promoting the efferocytosis process. However, the problems of poor water solubility, low bioavailability and possible toxicity at high doses severely limited the advancement of the clinical application of Cel [35]. For this reason, the development of efficient drug delivery systems to overcome these limitations is urgently needed. Nanocarrier systems have been shown to significantly improve the solubility and targeting of Cel and reduce its toxic side effects [36]. In this study, layered double hydroxide (LDH), a layered material consisting of positively charged metal hydroxide layers with interlayer anions, was chosen as a drug delivery vehicle to form a stable structure through charge balance [37]. The unique structure of LDH gives it a number of excellent properties, including high biocompatibility and good drug loading and release ability; therefore, it has broad application prospects in the treatment of a variety of diseases [38]. The pH-responsive property of LDH enables acid-triggered controlled release, which is especially suitable for the acidic microenvironment at the site of spinal cord injury [39]. In addition, the magnesium ions in MgAl-LDH can promote nerve regeneration and neural circuit remodelling, as well as facilitate the phenotypic transformation of microglia from the pro-inflammatory M1 phenotype to the anti-inflammatory M2 phenotype [40]. Lin et al. conducted a pioneering study on the application of MgFe-LDH in the treatment of bone diseases and demonstrated that it has good osteoinductive and osseointegrative properties [41]. Wang et al. demonstrated that Mg²⁺-loaded nanomaterials exert neuroprotective effects through anti-oxidative and anti-inflammatory mechanisms, suggesting that Mg²⁺ may contribute to the regulation of microglial function and neuroinflammation [42]. In summary, MgAl-LDH not only possesses excellent biocompatibility and controllable drug loading-releasing properties but also promotes nerve regeneration, making it a highly promising vehicle for drug delivery. In this study, a layered double hydroxide (LDH)-based nanomedicine delivery system was successfully constructed for loading Cel, and its role and related mechanisms in the repair of spinal cord injury were systematically evaluated. The Cel-LDH composite nanoparticles demonstrated robust ROS-scavenging capacity and preserved mitochondrial function, consequently modulating the expression of apoptosis-related proteins (e.g., Bax and Bcl-2) and effectively suppressing neuronal apoptosis. Furthermore, these nanoparticles enhanced autophagy, restored microglial efferocytosis, and further inhibited the cGAS–Sting pathway, which subsequently promoted an anti-inflammatory phenotypic shift in microglia, attenuated neuroinflammatory cascades, and ultimately fostered a favourable microenvironment for spinal cord injury repair. In a mouse model of SCI, the administration of Cel-LDH composite nanoparticles greatly improved motor function and encouraged axonal regeneration. This study provides a promising nanomedicine strategy for the treatment of spinal cord injury (Fig. 1). Schematic diagram of how Cel-LDH composite nanoparticles simultaneously promote microglial efferocytosis and alleviate neuronal apoptosis to promote recovery from spinal cord injury Results Synthesis and characterization of Cel-LDH Cel-loaded layered double hydroxide nanoparticles (Cel-LDH) were successfully prepared. Transmission electron microscopy (TEM) images revealed that compared with the original LDH, Cel-LDH maintained an intact nanolamellar structure but its size was increased because of the effective loading of Cel (Fig. 2A–B). Dynamic light scattering (DLS) analysis further revealed that the average particle size was 723.5 nm for LDH and 1197.1 nm for Cel-LDH, and the observed increase in particle size confirmed the successful synthesis of the composite particles (Fig. 2D–E). Energy dispersive X-ray spectroscopy (EDS) analysis confirmed the presence of uniformly distributed Mg, Al and O elements inside the LDH (Fig. 2C), and the zeta potential test results revealed that the potential shifted from + 8.93 mV to -6.24 mV, indicating the successful loading of Cel (Fig. 2F). Fourier transform infrared (FT-IR) spectroscopy provided further evidence of drug loading. The results showed that free celastrol exhibited a characteristic C = O stretching vibration peak at approximately 1700 cm⁻¹, attributed to its carboxyl group (-COOH). In the Cel-LDH conjugate, this peak completely disappeared, indicating that celastrol was not simply physically mixed or surface-adsorbed onto LDH; rather, its carboxyl group underwent deprotonation to form a carboxylate anion (-COO⁻), which interacted with the positively charged LDH layers through electrostatic interactions. Further comparison of the FT-IR spectra of LDH and Cel-LDH revealed additional changes: in the region of approximately 3400–3500 cm⁻¹, the shape and intensity of the hydroxyl absorption peak were markedly altered, suggesting the formation of a new hydrogen bonding network between the carboxylate groups of celastrol and the hydroxyl groups of the LDH layers, which changed the vibrational environment of the original hydroxyl groups. Notably, the symmetric and asymmetric stretching vibration peaks of the carboxylate anion typically appear in the regions of 1550–1610 cm⁻¹ and 1350–1420 cm⁻¹, respectively; however, these peaks may overlap or be masked by the intrinsic characteristic peaks of LDH. Collectively, these spectral changes demonstrate that celastrol was successfully loaded onto LDH, and that the loading process involves hydrogen bond formation and partial ion exchange (Fig. 2G). The absorbance of the Cel supernatant was measured at 424 nm, and a standard curve was plotted. On the basis of the standard curve, the loading rate of Cel was calculated as approximately 292 µg of Cel per 1 mg of LDH (Figure S1). An in vitro release assay revealed that the cumulative release rate of Cel-LDH reached 38.4% within 15 days (Fig. 2H), indicating that the amount of celastrol released is sufficient to achieve effective therapeutic concentrations (Figure S2). In addition, haemolysis experiments revealed that the haemolysis rate of Cel-LDH was less than 1.5%, indicating that the nanosystem has good biological safety, which provides an important basis for subsequent research (Fig. 2I). Characterization of Cel-LDH. (A) TEM image of LDH (scale bar: 500 nm). (B) TEM image of Cel-LDH (scale bar: 500 nm). (C) EDS elemental mapping of LDH (scale bar: 1 μm). (D) Particle size analysis of LDH as determined by DLS. (E) Particle size analysis of Cel-LDH as determined by DLS. (F) Zeta potential measurements for LDH and Cel-LDH obtained via DLS. (G) FT-IR spectra of LDH, Cel and Cel-LDH. (H) The release of Cel from Cel-LDH over 15 days. (I) Results of the haemolysis assays of LDH, Cel and Cel-LDH; PBS was used as a negative control, while Triton X-100 was used as a positive control Cel-LDH enhances the survival of cells under oxidative stress Following SCI, ROS levels increase sharply. Excessive ROS not only trigger intense oxidative stress but also induce an inflammatory cascade, collectively exacerbating neuronal apoptosis and disrupting the microenvironment for neural regeneration [42–46]. Thus, effectively scavenging ROS, mitigating oxidative damage, and modulating inflammatory responses to provide a conducive environment for neuronal survival represent critical strategies for promoting repair after SCI [47]. To evaluate the therapeutic potential of Cel-LDH, we first conducted a series of in vitro experiments. In an oxidative stress model, cell viability was assessed after stimulating HT22 cells with different concentrations of TBHP for 1 h. As shown in Figure S3, viability decreased to approximately 61% following 100 µM TBHP stimulation; therefore, this concentration was selected for subsequent cell damage models to simulate the acute phase of spinal cord injury. The effects of different concentrations of Cel-LDH on the proliferation of HT22 cells were then assessed using CCK-8 assays and Ki67 immunofluorescence staining. The CCK-8 results revealed a concentration-dependent increase in cell survival in the Cel-LDH-treated group compared with that in the H₂O₂-stimulated group alone (Fig. 3A). Ki67 staining further confirmed its ability to promote cell proliferation (Fig. 3E, K). Additionally, we established an inflammatory model in BV2 microglia using lipopolysaccharide (LPS) to evaluate the modulatory effect of Cel-LDH on cell activity under inflammatory conditions. Both the CCK-8 and Ki67 staining results indicated that Cel-LDH effectively promoted cell proliferation (Fig. 3B, D, J). On the basis of these findings, a concentration of 100 nM Cel-LDH was selected for subsequent experiments, as it provided significant efficacy without observable cytotoxicity. Cel-LDH enhanced the survival of cells under oxidative stress. (A) CCK-8 assay for the detection of the survival of HT22 cells treated with different concentrations of Cel-LDH under H₂O₂ conditions. (B) CCK-8 assay for the detection of the survival of BV2 cells treated with different concentrations of Cel-LDH under LPS conditions. (C) Live–dead staining assay of the different groups of HT22 cells under H2O2 conditions (scale bar: 100 μm). (D) Representative immunofluorescence images of Ki67 expression in BV2 cells (scale bar: 50 μm). (E) Representative immunofluorescence images of Ki67 expression in HT22 cells (scale bar: 50 μm). (F) Representative images of DCFH-DA staining in vitro (scale bar: 50 μm). (G) Representative images of DHE staining in vitro (scale bar: 50 μm). (H) Representative fluorescence microscopy images of DHE expression in vivo (scale bar: 25 μm). (I) HE-stained images of the heart, liver, spleen, lungs, and kidneys of mice in each group 28 days after SCI (scale bar: 200 μm). (J) Relative quantification of Ki67 fluorescence intensity in BV2 cells. K) Relative quantification of Ki67 fluorescence intensity in HT22 cells. L) Relative quantitative analysis of DCFH-DA fluorescence intensity in vitro. M) Relative quantitative analysis of DHE fluorescence intensity in vitro. N) Relative quantitative analysis of DHE fluorescence intensity in vivo. All the data are presented as the mean ± SD; n = 3; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 We next examined the ability of Cel-LDH to reduce cell death in H₂O₂-treated HT22 cells. Compared with H₂O₂ stimulation alone, LDH treatment, or Cel treatment, Cel-LDH treatment significantly increased the proportion of calcein-AM-positive cells, demonstrating notable protective effects (Fig. 3C). Together, these results clearly indicate that Cel-LDH significantly enhances cell proliferation and reduces cell death under both oxidative stress and inflammatory stress. These findings strongly suggest that loading Cel into LDH nanoparticles substantially improves its therapeutic efficacy against oxidative stress and cellular injury, highlighting a promising potential strategy for treating SCI. The levels of ROS in cells were subsequently detected using a 2,7-dichlorodihydrofluorescein diacetate (DCFH-DA) probe and a dihydroethidium (DHE) probe. After H₂O₂ treatment, the positive control group exhibited high fluorescence intensity, indicating elevated intracellular ROS levels (Fig. 3F–G, L–M). In contrast, Cel-LDH treatment markedly reduced the fluorescence intensity, demonstrating its effectiveness in attenuating ROS accumulation in HT22 cells. Consistent with the cellular findings, DHE staining of tissue sections also revealed that Cel-LDH treatment significantly decreased ROS levels (Fig. 3H, N). These findings further confirm that Cel-LDH effectively scavenges ROS not only at the cellular level but also in tissue, underscoring its potent antioxidative stress capability. In vivo, haematoxylin and eosin (H&E) staining was performed to evaluate morphological changes in the heart, liver, spleen, lungs, and kidneys of mice from the Sham, SCI, LDH, Cel, and Cel-LDH groups 28 days post-treatment (Fig. 3I). Compared with those in the Sham group, the structures in the Cel-LDH group remained intact, with no significant pathological alterations—such as inflammatory infiltration, degeneration, or fibrosis—indicating the good biocompatibility and safety of the Cel-LDH nanocomposite in vivo. These results provide an important experimental basis for its potential clinical translation. Cel-LDH protects mitochondrial homeostasis balance Following spinal cord injury, the secondary injury phase triggers excessive production of ROS, leading to severe mitochondrial dysfunction. Excessive accumulation of ROS attacks mitochondrial membrane phospholipids, proteins, and mtDNA, disrupting the integrity of the electron transport chain (ETC), reducing ATP synthesis efficiency, and inducing mitochondrial membrane potential collapse [48]. Therefore, accurate detection of reactive oxygen species levels within mitochondria is critical. The results of using MitoSOX to specifically measure superoxide levels within mitochondria are shown in Fig. 4C, G. The results indicated that red fluorescence was significantly enhanced in the hydrogen peroxide-stimulated group and that mitochondrial ROS levels were significantly reduced after administration, particularly in the composite particle group. The mitochondrial membrane potential (MMP) is a key indicator of normal mitochondrial function, and JC-1 staining was used to measure the MMP. Typically, in healthy mitochondria, JC-1 aggregates show red fluorescence; in injured mitochondria, this signal is transformed to green fluorescence. Compared with the control group, the H2O2 stimulation group showed a significant decrease in red fluorescence intensity and a marked increase in green fluorescence intensity, indicating a decrease in the MMP and MPTP opening. After drug treatment, compared with that in the H2O2 group, especially in the Cel-LDH complex particle group, the red fluorescence intensity recovered, while the green fluorescence intensity weakened (Fig. 4A, E). To determine whether Cel-LDH similarly protects mitochondrial function in microglia under inflammatory conditions, we performed JC-1 staining in BV2 cells stimulated with LPS. As shown in Fig. 4B, F, Cel-LDH treatment also significantly restored the mitochondrial membrane potential in BV2 cells, as evidenced by an increased red/green fluorescence ratio of JC-1 aggregates, which is consistent with the protective effects observed in HT22 cells under oxidative stress. These results collectively demonstrate that Cel-LDH effectively preserves mitochondrial homeostasis in both neurons and microglia under different pathological conditions. Cel-LDH protected neurons and microglia against mitochondrial dysfunction and oxidative stress. (A) Representative JC-1 staining images of MMP levels in H2O2-treated HT22 cells (scale bar: 50 μm). (B) Representative JC-1 staining images of MMP levels in LPS-treated BV2 cells (scale bar: 50 μm). (C) MitoSOX staining findings showing the levels of mitochondrial ROS in H2O2-treated HT22 cells (scale bar: 50 μm). (D) Representative immunofluorescence images of ATPB in H2O2-treated HT22 cells (scale bar: 50 μm). (E) Quantification of the JC-1 aggregate/monomer fluorescence ratio in HT22 cells. (F) Quantification of the JC-1 aggregate/monomer fluorescence ratio in BV2 cells. (G) Relative quantification of MitoSOX fluorescence intensity. (H) Relative quantification of ATPB fluorescence intensity. (I) ATP levels in the different groups of HT22 cells. (J) ATP levels in the spinal cord tissues of different groups of mice. All the data are presented as the mean ± SD; n = 3; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 The core function of mitochondria is to produce ATP to meet the high energy demands of neurons [49]. ATP synthesis depends primarily on the catalytic activity of ATP synthase (ATPB). ATPB in mitochondria is the central enzyme of energy metabolism and it catalyses ATP synthesis as the final step in the mitochondrial electron transport chain. The efficient operation of ATPB is crucial for maintaining cellular energy homeostasis. In this study, ATPB expression levels in HT22 cells were detected using immunofluorescence staining, and as shown in Fig. 4D, H, the fluorescence intensity of ATPB was significantly reduced in the H2O2 stimulation group, indicating that excessive ROS damage to mitochondrial activity leads to ATPB dysfunction. To further assess the neuroprotective effects of Cel-LDH on mitochondria, ATP levels were also measured. These results were consistent with the expression levels of ATPB in neural cells across all the groups, indicating that Cel-LDH can mitigate the reduction in ATP levels induced by H2O2 (Fig. 4I). Additionally, ATP levels in mouse spinal cord tissue were measured, further confirming that Cel-LDH can restore ATP synthesis levels following spinal cord injury (Fig. 4J). In summary, Cel-LDH effectively scavenges excess ROS produced within mitochondria, alleviates oxidative stress-induced mitochondrial damage, and improves mitochondrial functional status, thus creating favourable conditions for the repair of spinal cord injury. Cel-LDH promotes neural differentiation and inhibits glial scar formation in spinal cord injury mice Next, we systematically evaluated the neuroprotective effects of Cel-LDH in SCI mice by comprehensive tissue staining. Neural stem cell differentiation plays a key regulatory role in the repair of spinal cord injury, with its differentiation fate directly influencing the regenerative potential and functional recovery of injured tissue. Under the regulation of specific microenvironmental signals, neural stem cells can be directed to differentiate into three major neural lineages, namely, neurons, astrocytes, and oligodendrocyte precursor cells. To clarify the regulatory role of Cel-LDH in neural differentiation, we monitored neuronal regeneration at key time points (7 days/28 days) after spinal cord injury using immunofluorescence staining (Fig. 5A–B, E–F). The results showed that in the early stages of injury, the density of Tuj-1-positive cells in the Cel-LDH group had already significantly increased, indicating that it accelerated the initiation of neurogenesis. At 28 days post-spinal cord injury, the density of Tuj-1-positive cells remained high. These results collectively demonstrate that the composite particles can promote the neural differentiation of endogenous neural stem cells after SCI. Cel-LDH promoted neural differentiation and inhibited glial scar formation in SCI mice. (A) Representative immunofluorescence images of Tuj-1 expression in SCI mice at 7 days (scale bar: 500 μm). (B) Representative immunofluorescence images of Tuj-1 expression in SCI mice at 28 days (scale bar: 500 μm). (C) Representative immunofluorescence images of GFAP and MAP-2 expression in SCI mice at 28 days (scale bar: 500 μm). (D) Representative immunofluorescence images of NF-200 and MBP expression in SCI mice at 28 days (scale bar: 500 μm). (E) Quantitative analysis of the Tuj-1 stained area at 7 days. (F) Quantitative analysis of the Tuj-1 stained area at 28 days. (G) Quantitative analysis of the MAP-2 stained area. (H) Quantitative analysis of the NF-200 stained area. (I) Quantitative analysis of the MBP stained area. All the data are presented as the mean ± SD; n = 3; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 Following spinal cord injury, the local microenvironment undergoes significant pathological changes, including elevated oxidative stress levels, microcirculatory disorders caused by local ischaemia and hypoxia, disruption of neural conduction pathways due to axonal rupture and demyelination, and a series of inflammatory cascades [50]. These changes collectively form an inhibitory microenvironment that disrupts neural stem cell differentiation-primarily manifested as abnormal proliferation of astrocytes, while the differentiation rate towards functional neurons is significantly reduced. This differentiation shift not only limits the potential for neural regeneration but also exacerbates glial scar formation, further impairing functional recovery [51]. Therefore, we performed immunofluorescence staining on spinal cord tissue Sect. 28 days post-injury to assess astrocyte distribution (GFAP) and neuronal growth (MAP-2) (Fig. 5C, G). The results revealed that the SCI control mice exhibited high GFAP fluorescence intensity in the injured area and surrounding regions, indicating the formation of dense glial scar tissue around the injury site. On the contrary, the GFAP fluorescence intensity around the injured area was significantly reduced in the Cel-LDH-treated mice, suggesting that these composite particles can effectively prevent astrocyte activation and glial scar formation following SCI. The MAP-2 immunofluorescence results revealed that the Cel-LDH group had a large number of MAP-2-positive neurons in the central region of the injury site, whereas the SCI group had a significantly reduced number of MAP-2-positive neurons. These results imply that by preventing glial scar formation and promoting nerve fiber regeneration at the location of spinal cord injury, Cel-LDH composite particles support functional neurological recovery. The structural repair and functional reorganization of neurons are central elements in the process of spinal cord injury repair. Following spinal cord injury, axonal rupture and demyelination are among the primary causes of neurological dysfunction [52]. Axons are the key structures responsible for transmitting electrical signals in neurons [53], whereas myelin sheaths, formed by oligodendrocytes (OLs) enveloping axons, significantly increase the conduction speed of neural impulses and maintain axonal integrity [54, 55]. Therefore, axon regeneration and myelin remodelling are critical steps in restoring neural function during the repair process of central nervous system injuries. In this study, we employed immunofluorescence staining techniques to perform dual-label detection of neurofilament protein-200 (NF-200) and myelin basic protein (MBP) to systematically assess myelinated nerve fibre regeneration in experimental animals across different groups. As shown in Fig. 5D, H–I, quantitative analysis revealed that the density of NF-200-positive axons and the degree of MBP-positive myelin coverage were significantly greater in the Cel-LDH treatment group than in the control groups. These results suggest that Cel-LDH composite nanoparticles can effectively promote axon regeneration in the injured area, improve myelin formation, and play a crucial role in facilitating the reconstruction of functional neural structures at the site of spinal cord injury. Cel-LDH promotes motor function recovery in mice with spinal cord injury The extent of recovery of hindlimb motor function in paralyzed mice following spinal cord injury directly reflects the regenerative status of functional neural structures in the injured region. Figure 6A shows a schematic diagram outlining the in vivo experimental workflow, which included a series of behavioural tests to systematically evaluate the efficacy of different treatments in promoting motor function recovery. The Basso Mouse Scale (BMS) was used to assess hindlimb motor function at 7, 14, 21, and 28 days post-injury (Fig. 6B). On day 7 post-injury, the BMS scores of all the experimental groups were less than 2, indicating severe impairment of hindlimb motor function following SCI. Over time, the motor function of the mice in the Cel-LDH group gradually recovered. From day 14 to day 28 post-injury, the BMS scores of the Cel-LDH group were significantly greater than those of the control group, with the difference becoming more pronounced by day 28. These results indicate that the Cel-LDH material can effectively promote early recovery of motor function after SCI and can continuously improve neural function. As shown in Fig. 6C, the results of the footprint analysis were consistent with the trends in the BMS scores. In representative footprint analysis images, mice in the SCI group exhibited severe dragging of the hindlimbs, whereas those in the Cel-LDH group showed more regular footprint patterns, with significantly reduced dragging traces and markedly improved coordination between the hindlimbs and forelimbs. Gait parameters such as the rotation angle, relative position, and base of support were superior to those of the control group. Cel-LDH promoted motor function recovery in mice with SCI. (A) The in vivo experimental technique is shown in a schematic timeline. (B) BMSs of the mice at the corresponding time points. (C) Footprint analysis of SCI mice after different treatments. (D) H&E staining of spinal cord sections (scale bar: 500 μm). (E) Nissl staining of spinal cord sections (scale bar: 500 μm). (F) Macroscopic images of bladder and gastrocnemius muscle tissues collected 28 days post-surgery. (G) Gastrocnemius muscle tissue weight in each group of mice at 28 days post-surgery. (H) Bladder tissue weight in each mice group at 28 days post-surgery. (I) GO functional enrichment analysis of genes that were differentially expressed between the SCI and Cel-LDH groups. (J) KEGG analysis of genes that were differentially expressed between the SCI and Cel-LDH groups. All the data are presented as the mean ± SD; n = 3; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 Twenty-eight days after SCI, the ability of Cel-LDH to repair the spinal cord injury area was evaluated using a histopathological analysis system. H&E staining revealed that the SCI group exhibited large, irregular cavities surrounding the spinal cord tissue injury area, indicating severe tissue defects and structural damage (Fig. 6D). In contrast, the Cel-LDH treatment group showed significantly improved tissue integrity in the injured area, with a markedly reduced cavity size and more regular cavity edges. The extent of neuronal damage was further assessed using Nissl staining. Nissl bodies, important acidic protein structures within neurons, directly reflect the functional integrity of neurons. The staining results revealed that the number of Nissl-positive cells in the central region of the SCI group was significantly reduced, and the remaining neurons exhibited Nissl body dissolution, which is consistent with the mechanism of neuronal necrosis caused by excessive ROS. In contrast, the Nissl-positive staining intensity significantly increased in the Cel-LDH treatment group (Fig. 6E). These results indicate that composite nanoparticles loaded with LDH and Cel extract can promote structural regeneration of spinal cord tissue, thereby effectively facilitating the recovery of motor function in mice with spinal cord injury. Bladder dysfunction is a prevalent and serious complication of spinal cord injury (SCI) [56]. At 28 days post-injury, bladder assessments revealed that compared with the sham-operated group, the SCI group had significantly greater bladder volumes and markedly greater bladder weights (2–3 times greater). This phenomenon can be attributed to initial passive dilation due to urinary retention, followed by progressive bladder wall thickening in the chronic phase as a result of compensatory mechanisms and pathological remodelling. Notably, although the bladder mass increased, its functional capacity was substantially reduced [57]. Compared with the other treatments, treatment with Cel-LDH significantly ameliorated these pathological alterations, resulting in a pronounced reduction in both bladder volume and weight (Fig. 6F, H). Concurrently, alterations in the gastrocnemius muscle were assessed. Following SCI, the disruption of neural pathways leads to a loss of voluntary motor input and neurotrophic support, resulting in progressive and severe muscle atrophy, characterized by a significant reduction in both muscle mass and volume [58]. As shown in Fig. 6F–G, the gastrocnemius muscle in the SCI group exhibited a marked decrease in size because of atrophy. These pathological changes were ameliorated by treatment with the composite nanoparticle formulation, indicating its efficacy in improving motor function recovery in the hindlimb post-SCI. Transcriptomic analysis revealed distinctions between the SCI group and the Cel-LDH group To further elucidate the molecular mechanism by which Cel-LDH promotes spinal cord injury repair, RNA sequencing analysis was used to evaluate the therapeutic effects of Cel-LDH following SCI. All of the differentially expressed genes underwent Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology (GO) enrichment studies to further clarify the possible pathways by which Cel-LDH facilitates recovery following SCI. GO enrichment analysis (Fig. 6I) revealed that the DEGs in the Cel-LDH group relative to the SCI group showed functional enrichment in all three major categories of biological processes, cellular components, and molecular functions. Notably, genes related to phagocytosis, the response to oxidative stress, autophagy, the mitochondrial membrane, and the negative regulation of the neuronal apoptotic process were significantly enriched. These findings suggest that Cel-LDH effectively inhibits the neuronal apoptotic process by scavenging ROS and maintaining mitochondrial homeostasis, thus regulating apoptosis-related proteins. KEGG pathway enrichment analysis (Fig. 6J) further revealed the involvement of the cytosolic DNA-sensing pathway, efferocytosis and other pathways linked to inflammation in this experimental context. These results suggest that Cel-LDH may inhibit the cGAS–Sting pathway by promoting autophagy, thereby modulating the conversion of microglia to an anti-inflammatory phenotype and promoting efferocytosis, which in turn inhibits the inflammatory cascade response and creates a favourable microenvironment for the repair of spinal cord injury. Cel-LDH inhibits neuronal apoptosis To verify the role of Cel-LDH, we performed additional experiments. Mitochondrial dysfunction triggers the abnormal release of cytochrome c, which activates the apoptotic pathway and ultimately leads to massive neuronal apoptosis [59]. To assess the level of apoptosis in the spinal cord tissue, we performed TUNEL staining, as this is the standard method for detecting apoptosis levels. The results demonstrated that spinal cord tissue apoptosis was considerably decreased by treatment with LDH, Cel, and Cel-LDH. Cel-LDH outperformed LDH and Cel alone in its ability to reduce apoptosis among these therapies (Fig. 7B, F). In vitro cell experiments on HT22 cells validated the TUNEL staining results, with the Cel-LDH composite particles demonstrating potent antiapoptotic effects (Fig. 7A, E). We then further examined changes in the expression levels of the classic apoptosis-related proteins Bcl-2, BAX, Cyt c, and cleaved-caspase 3 by Western blotting. Compared with that in the spinal cord injury group, the expression of Bcl-2 in the Cel-LDH-treated group increased, whereas the expression of BAX, Cyt c, and cleaved-caspase 3 significantly decreased (Fig. 7D, H-K). To further validate these findings, we used immunofluorescence staining to detect the expression levels of the antiapoptotic marker Bcl-2 in HT22 cells. The staining results revealed that the fluorescence intensity of Bcl-2 was significantly reduced in the H2O2 stimulation group, whereas Cel-LDH treatment effectively reversed this phenomenon (Fig. 7C, G). These findings indicate that Cel-LDH can inhibit mitochondrial-dependent apoptosis, reduce neuronal apoptosis, and exert significant neuroprotective effects. Cel-LDH inhibited neuronal apoptosis. (A) TUNEL staining images of different groups of HT22 cells treated with H2O2 (scale bar: 50 μm). (B) TUNEL staining images of spinal cord tissue from different groups of mice (scale bar: 25 μm). (C) Representative immunofluorescence images of Bcl-2 in vitro (scale bar: 50 μm). (D) Representative Western blot images of Bcl-2, Bax, Cyt c and cleaved-caspase 3 expression in vivo. (E) Relative quantification of TUNEL fluorescence intensity in vitro. (F) Relative quantification of TUNEL fluorescence intensity in vivo. (G) Relative quantification of the Bcl-2 fluorescence intensity. H–K) Quantitative analysis of the Western blotting results of the Bcl-2, Bax, Cyt c and cleaved-caspase 3 protein expression levels. All the data are presented as the mean ± SD; n = 3; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 To determine the direct effect of Mg²⁺ released from LDH on neuronal survival and apoptosis, we performed CCK-8 and TUNEL assays. HT22 neurons were treated with increasing concentrations of Mg²⁺ (2.5, 5, 10, and 20 mM). The CCK-8 assay revealed that Mg²⁺ exerted certain protective effects, with 5 mM Mg²⁺ showing the highest protective effect on cell viability (Figure S4A). Therefore, 5 mM Mg²⁺ was selected for subsequent experiments. In H₂O₂-stimulated HT22 neurons, compared with the stimulation group, Mg²⁺ treatment group showed significantly increased CCK-8 cell viability and significantly reduced TUNEL-positive cells (Figure S4B-C). These results indicate that Mg²⁺ exerts direct pro-survival and anti-apoptotic effects on neurons. To further investigate whether Cel-LDH-treated microglia indirectly protect neurons through the secretion of soluble factors, we performed a conditioned medium experiment. BV2 microglial cells were divided into five groups: Control, LPS, LPS + LDH, LPS + Cel, and LPS + Cel-LDH. Conditioned media from each group were collected and applied to H₂O₂-injured HT22 neurons, and neuronal apoptosis was assessed by TUNEL staining (Figure S5). The results showed that conditioned medium from the LPS-stimulated BV2 cells exerted no protective effect on injured HT22 neurons, as evidenced by strong TUNEL-positive fluorescence signals. In contrast, conditioned media from the LPS + LDH and LPS + Cel groups exhibited moderate protective effects, with a marked reduction in the number of TUNEL-positive cells compared to the LPS group. Notably, conditioned medium from the LPS + Cel-LDH group exhibited the strongest protective effect, with a significant reduction in TUNEL-positive cells, which was markedly better than that observed with LDH or Cel alone. These findings indicate that Cel-LDH-treated microglia secrete soluble factors that indirectly protect injured neurons, and that the Cel-LDH nanocomposite exhibits superior efficacy compared to free Cel or LDH alone. Cel-LDH restores autophagic flux after spinal cord injury Autophagy, as an essential intracellular quality control mechanism, effectively recognizes and selectively eliminates damaged organelles and protein aggregates, prevents the aberrant release of damage-associated molecular patterns (DAMPs), and promptly controls the subsequent inflammatory cascade, thereby creating a favorable microenvironment for neural repair following spinal cord injury [60]. Based on this, the present study further investigated whether Cel-LDH exerts neuroprotective effects by modulating the autophagic process. To clarify the direct regulatory effect of Cel-LDH on autophagy, we first validated it at the cellular level. Immunofluorescence staining was performed to detect the expression of key autophagy-related proteins in HT22 cells. The results showed that in the LPS-induced inflammatory model, the fluorescence signal of the autophagy substrate protein SQSTM1/p62 was significantly enhanced, and the fluorescence signal of the autophagosome marker protein LC3 was also increased, suggesting that post-injury inflammation induces autophagosome formation while autophagic flux is blocked at the degradation stage, leading to abnormal accumulation of both autophagosomes and substrates. Notably, following Cel-LDH treatment, the fluorescence intensity of p62 was markedly reduced, whereas LC3 fluorescence was further enhanced, indicating that Cel-LDH effectively alleviates the degradation blockade of autophagic flux, promotes the clearance of autophagic substrates, and maintains autophagosome formation, thereby restoring normal autophagic flux (Fig. 8A–B, H–I). To further validate the in vivo regulatory effect of Cel-LDH on autophagy, immunofluorescence staining was used to detect the expression of autophagy-related proteins in the injured spinal cord tissue. The results demonstrated that compared with the SCI group and other control groups, the Cel-LDH treatment group exhibited significantly reduced fluorescence signals of SQSTM1/p62 in the peri-lesion area, indicating effective clearance of accumulated autophagic substrates. Meanwhile, the fluorescence intensity of LC3 was further enhanced with more densely distributed positive signals, suggesting that autophagosome formation remains robust and autophagic flux is restored (Fig. 8C–D, J–K). Cel-LDH restores autophagic flux after SCI. (A) Representative immunofluorescence images of LC3 in vitro (scale bar: 50 μm). (B) Representative immunofluorescence images of p62 in vitro (scale bar: 50 μm). (C) Representative immunofluorescence images depicting LC3 colocalization with Iba1 in vivo (scale bar: 25 μm). (D) Representative immunofluorescence images depicting p62 colocalization with Iba1 in vivo (scale bar: 25 μm). (E) Representative Western blot images of Beclin and p62 in vivo. (F–G) Quantitative analysis of the Western blotting results of the Beclin and p62 protein expression levels. (H–I) Relative quantification of LC3 and p62 fluorescence intensity in vitro.(J–K) Relative quantification of LC3 and p62 fluorescence intensity in vivo. All the data are presented as the mean ± SD; n = 3; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 Western blot analysis was further performed to quantitatively validate the expression levels of autophagy-related proteins. The results showed that following Cel-LDH treatment, the expression level of Beclin, a key protein involved in autophagy initiation, was significantly upregulated, while the protein level of the autophagic substrate SQSTM1/p62 was markedly decreased (Fig. 8E–G). These findings are highly consistent with the observations from immunofluorescence staining, further confirming that Cel-LDH effectively activates the autophagic process and promotes the recognition and clearance of damaged cellular components. Collectively, Cel-LDH activates autophagy by upregulating Beclin expression, promoting LC3 conversion, and accelerating p62 degradation, thereby facilitating the timely clearance of damaged organelles and suppressing the inflammatory cascade. Cel-LDH promotes efferocytosis after spinal cord injury Targeted regulation of efferocytosis plays a critical role in the repair of spinal cord injury. Efferocytosis mediated by macrophages/microglia can promptly clear apoptotic cells, effectively blocking the inflammatory cascade triggered by secondary injury and promoting the formation of an anti-inflammatory microenvironment, thereby supporting neural tissue repair [61]. However, under pathological conditions, the spinal cord injury microenvironment can lead to efferocytosis dysfunction, triggering an inflammatory storm and exacerbating tissue damage [62]. As a key receptor that regulates efferocytosis, the upregulation of CD36 expression suggests that Cel-LDH may promote the clearance of apoptotic cells by enhancing the efferocytosis function of microglia [63]. Immunofluorescence staining revealed significantly increased positive signals for CD36 expression in spinal cord tissue from the Cel-LDH-treated group (Fig. 9A, F). To validate this phenomenon, we further assessed CD36 expression levels in BV2 microglia, and the results revealed that the trend in fluorescence intensity changes was highly consistent with the in vivo experimental results (Fig. 9D, J). Western blot quantitative analysis confirmed that Cel-LDH treatment significantly increased CD36 protein expression levels (Fig. 9E, H). We also assessed the uptake of Dil-labelled apoptotic neuronal cells by microglia. BV2 cells were labelled with a microglia-specific marker, and the HT22 cell membranes were labelled with Dil. Under confocal microscopy, compared with the LDH and Cel groups, the Cel-LDH group exhibited stronger intracellular red fluorescence (Fig. 9B, G). These results collectively demonstrate that Cel-LDH effectively promotes microglial efferocytosis. Cel-LDH promotes efferocytosis after SCI. (A) Representative immunofluorescence images depicting CD36 colocalization with Iba1 in vivo (scale bar: 25 μm). (B) Immunofluorescence analysis to assess the efferocytosis of BV2 cells in response to LPS treatment in different groups (scale bar: 25 μm). (C) Representative immunofluorescence images of CD36 expression in BV2 cells following different treatments (scale bar: 50 μm). BV2 cells were treated with Control, LPS, LPS + Cel-LDH and LPS + Cel-LDH + 3-MA. (D) Representative immunofluorescence images of CD36 in vitro (scale bar: 50 μm). (E) Representative Western blot images of CD36 in vivo. (F) Relative quantification of CD36 fluorescence intensity in vivo. (G) Relative quantification of microglia efferocytosis fluorescence intensity. (H) Quantitative analysis of the Western blotting results of the CD36 protein expression level. (I) Relative quantification of CD36 fluorescence intensity in Fig. 9C. (J) Relative quantification of CD36 fluorescence intensity in vitro. All the data are presented as the mean ± SD; n = 3; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 To further determine whether autophagy directly participates in the regulation of efferocytosis, we employed 3-methyladenine (3-MA), a specific autophagy inhibitor. BV2 microglial cells were treated with LPS, LPS + Cel-LDH, or LPS + Cel-LDH + 3-MA, and CD36 expression was assessed by immunofluorescence staining. As shown in Fig. 9C, I, LPS stimulation markedly reduced CD36 fluorescence intensity, indicating impaired efferocytosis capacity. Treatment with Cel-LDH significantly enhanced CD36 expression. Notably, co-treatment with 3-MA substantially attenuated the Cel-LDH-induced increase in CD36 fluorescence, suggesting that inhibition of autophagy abolishes the pro-efferocytic effect of Cel-LDH, strongly supporting a causal relationship in which autophagy acts upstream of efferocytosis. Cel-LDH inhibits the cGAS–Sting pathway Following spinal cord injury, the hypoxia-ischemia microenvironment at the injury site triggers mitochondrial dysfunction, leading to abnormal leakage of mtDNA into the cytoplasm [64]. These free mtDNA molecules can be specifically recognized and bound by cGAS, which then catalyses the synthesis of the second messenger cGAMP, activating the Sting pathway and triggering downstream immune cascades [65, 66]. Therefore, regulating mitochondrial homeostasis and rescuing damaged mitochondria to inhibit excessive activation of the cGAS–Sting signalling pathway may become important therapeutic strategies for promoting spinal cord injury repair. Immunofluorescence staining revealed that Cel-LDH effectively inhibited the upregulation of Sting and cGAS expression induced by LPS stimulation (Fig. 10A–B, E–F). Previous studies have shown that the protein expression of Sting and cGAS significantly increases after spinal cord injury and peaks on day 7 [64]. Using the microglial marker Iba-1 for double immunofluorescence staining of Sting, we further confirmed that the expression of Sting and cGAS was elevated in activated microglia in mice after SCI, whereas the fluorescence intensity of Sting and cGAS was significantly reduced in the Cel-LDH group (Fig. 10C–D, G–H). We subsequently assessed the activation level of the Sting pathway in each group via Western blotting. Protein quantification revealed that the Cel-LDH group had far lower levels of cGAS protein expression than the SCI group. Additionally, in the Cel-LDH group, the phosphorylation levels of Sting protein, TBK1 protein, and IRF3 significantly decreased (Fig. 10I–M). These results collectively indicate that Cel-LDH effectively inhibits the activation of the cGAS–Sting signalling pathway. Cel-LDH inhibited the cGAS–Sting pathway. A–B) Representative immunofluorescence images of cGAS and Sting in vitro (scale bar: 50 μm). C–D) Representative immunofluorescence images of cGAS and Sting in vivo (scale bar: 25 μm). E–F) Relative quantification of cGAS and Sting fluorescence intensity in vitro. G–H) Relative quantification of cGAS and Sting fluorescence intensity in vivo. I) Representative Western blot images of the cGAS–Sting pathway in vivo. J–M) Quantitative analysis of the Western blotting results of the cGAS–Sting pathway-related protein expression levels. All the data are presented as the mean ± SD; n = 3; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 Cel-LDH enhances microglial polarization and inhibits neuroinflammation Research has demonstrated that sustained activation of the cGAS–Sting pathway significantly upregulates IFN-β and IL-6 via the IRF3/NF-κB signalling pathway, encouraging microglia to polarize toward the M1 phenotype while decreasing the production of M2-type markers [67, 68]. This process facilitates neuroinflammation, exacerbates neuronal damage, and impedes tissue repair [69, 70]. Therefore, regulating the M1/M2 phenotype conversion of microglia is crucial for neural repair after spinal cord injury. We conducted a series of systematic immunofluorescence staining experiments to assess the regulatory effect of Cel-LDH on microglial phenotype polarization. The experiments examined the expression of the M1 phenotype markers CD86 and iNOS and the M2 phenotype markers CD206 and Arg-1 in microglia. The results of fluorescence staining revealed that the red fluorescence associated with CD86 and iNOS expression gradually decreased in the Cel-LDH composite particle group, whereas the green fluorescence associated with CD206 and Arg-1 expression correspondingly increased, directly indicating that the composite particles promoted the transition of microglia towards the M2 phenotype (Fig. 11A–D, L–O). Cel-LDH enhanced microglial polarization and inhibited neuroinflammation. A–D) Representative immunofluorescence images of CD86, CD206, iNOS and Arg-1 expression in vitro (scale bar: 50 μm). E–G) Representative immunofluorescence images of CD86, CD206, iNOS and Arg-1 expression in vivo (scale bar: 25 μm). H–I) RT‒qPCR analysis of the gene expression of anti-inflammatory cytokines (TGF-β and IL-10). J–K) RT‒qPCR analysis of the expression of proinflammatory cytokine genes (IL-1β and TNF-α). L–O) Relative quantification of CD86, CD206, iNOS and Arg-1 fluorescence intensity in vitro. P–S) Relative quantification of CD86, CD206, iNOS and Arg-1 fluorescence intensity in vivo. All the data are presented as the mean ± SD; n = 3; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 To determine whether Mg²⁺ released from LDH also participates in microglial polarization, we performed immunofluorescence staining for the M1 marker CD86 and the M2 marker CD206 in LPS-stimulated BV2 microglia treated with Mg²⁺. The results showed that Mg²⁺ treatment significantly decreased CD86 fluorescence intensity and increased CD206 fluorescence intensity compared with the LPS alone group, indicating that Mg²⁺ promotes microglial polarization from the M1 toward the M2 phenotype (Figure S4D–E, G-H). Given that M2-polarized microglia are known to exhibit enhanced efferocytosis capacity, we further assessed the effect of Mg²⁺ on CD36 expression, a key efferocytosis receptor. Immunofluorescence staining revealed that Mg²⁺ treatment significantly upregulated CD36 expression in LPS-stimulated BV2 microglia (Figure S4F, I), suggesting that Mg²⁺ promotes microglial efferocytosis by facilitating M2 polarization. Previous studies have demonstrated the role of Cel-LDH in regulating microglial polarization. Using Iba-1 as a marker for microglial localization, different treatments were given at the injury site in in vivo experiments, and the colocalization and fluorescence intensity of CD86 and CD206 with microglia were observed. The in vivo fluorescence staining results were highly consistent with the in vitro experimental results. Cel-LDH treatment significantly reduced the proportion of CD86+/Iba1+ double-positive cells and increased the proportion of CD206+/Iba1+ double-positive cells, suggesting that composite particles can reduce the excessive inflammatory responses at the injury site after SCI by increasing the proportion of M2 microglia (Fig. 11E–F, P–Q). To further validate these findings, we performed double staining for iNOS and Arg-1. iNOS-positive cells were predominant in the SCI group, whereas the proportion of Arg-1-positive cells significantly increased after Cel-LDH administration, which is consistent with the results of previous studies (Fig. 11G, R–S). M1 macrophages are linked to the continuous creation and secretion of proinflammatory factors, including TNF-α and IL-1β, whereas M2 macrophages counteract inflammation by producing IL-10 and TGF-β. Therefore, qRT‒PCR technology was used to measure the mRNA expression levels of key inflammatory factors. Cel-LDH significantly reduced the expression of the proinflammatory cytokines IL-1β and TNF-α but significantly increased the levels of the anti-inflammatory markers IL-10 and TGF-β (Fig. 11H–K). Taken together, these results indicate that LDH, Cel, and the composite particle Cel-LDH can effectively induce macrophages to transition from the proinflammatory M1 phenotype to the anti-inflammatory M2 phenotype. Notably, the Cel-LDH composite particles exhibit the most significant polarization regulation and anti-inflammatory effects. Discussion Spinal cord injury is a serious trauma to the central nervous system and often results in persistent sensory and motor dysfunction. After injury, insufficient perfusion and cellular metabolic disorders occur locally, causing hypoxia and enhanced glycolysis, resulting in the formation of an injury microenvironment. This unfavourable microenvironment not only exacerbates neuronal cell death and mitochondrial dysfunction but also promotes the release of inflammatory factors and the aberrant activation of microglia, further exacerbating the secondary injury. Here, we developed a nanodrug delivery system applicable to the local microenvironment after spinal cord injury and delivered Cel with LDH to optimize its therapeutic effect. In vitro, Cel-LDH significantly promoted autophagy and efferocytosis after oxidative stress injury and inhibited mitochondria-dependent apoptotic pathways in neurons. In in vivo experiments, Cel-LDH treatment effectively promoted motor function recovery in mice with SCI. This drug delivery strategy is important for clinical translation and provides a new therapeutic strategy for treating SCI. Improving mitochondrial function has become among the most important strategies for promoting spinal cord injury repair. Cel has been previously demonstrated to have potential in maintaining mitochondrial homeostasis and function. For example, Li et al. reported that Cel could ameliorate mitochondrial dysfunction in neurons by specifically binding to the EPAC-1 protein, thereby attenuating secondary brain damage after cerebral haemorrhage [34]. Cel also binds to Nur77, a nuclear receptor in mitochondria, and induces mitochondrial autophagy in macrophages, removing injured mitochondria and thereby inhibiting inflammation and metabolic-related diseases such as obesity [27]. Similarly, Shi et al. suggested that Cel could effectively inhibit breast cancer by modulating the mitochondrial metabolic reprogramming process [71]. However, although Cel has been shown to have significant mitochondrial protective effects in a variety of disease models, studies on its role in spinal cord injury repair are still limited. On the basis of its potential to improve mitochondrial function and regulate autophagy, the present study explored the possible use of Cel in spinal cord injury repair. The experimental results revealed that the compound was able to effectively restore mitochondrial function, increase autophagy activity, and alleviate the pathological state of the injured region, which is consistent with the effects observed in other models in the literature, further supporting its application value in the field of nerve repair. It is important to clarify a conceptual point regarding the relationship between mitochondrial function and autophagy. It is well established that mitochondrial damage triggers selective mitophagy as a quality control mechanism to eliminate dysfunctional organelles [72]. However, in the pathological context of spinal cord injury, global autophagic flux is impaired primarily at the degradation stage rather than at the initiation stage [17]. This is evidenced by our experimental data showing the concurrent accumulation of both the autophagosome marker LC3 and the autophagy substrate p62 in the injured spinal cord and LPS-stimulated microglia, indicating that although autophagosome formation is induced, the subsequent degradation of autophagic cargo is blocked. Following Cel-LDH treatment, we observed a further increase in LC3 coupled with a marked decrease in p62, suggesting that Cel-LDH effectively alleviates this degradation blockade, restores normal autophagic flux, and promotes the clearance of accumulated damaged organelles. Thus, improved mitochondrial function does not suppress mitophagy; rather, it creates a permissive cellular environment that enables the autophagic machinery to operate efficiently. This refined interpretation resolves the apparent paradox and is fully consistent with both established biological principles and our experimental data. Mitochondrial dysfunction involves a variety of pathological processes, including damage to mtDNA and its aberrant leakage to the cytoplasm [73]. An essential natural immunity signaling route, the cGas–Sting pathway detects double-stranded DNA from microorganisms and hosts in the cytoplasm and, in a Sting-dependent manner, triggers cells innate immune responses [74]. mtDNA leakage is recognized by cGas and it acts to catalyse the generation of the second messenger cGAMP; cGAMP subsequently activates Sting proteins in the endoplasmic reticulum, which in turn regulates the innate immune response and promotes inflammation [69, 75]. Studies have shown that in LPS-induced microglial inflammation models, the release of mtDNA into the cytoplasm activates the Sting pathway, and the polarization of microglia to the M1 phenotype is subsequently observed. Cai et al. reported that the expression of key receptors associated with apoptotic cell recognition (e.g., MerTK, AXL, and Tyro3) was also significantly reduced [76]. Moreover, high levels of proinflammatory factors such as TNF-α and IL-1β secreted after M1 polarization inhibited the activation of phagocytosis signals and induced the expression of “do not eat me” signals (e.g., CD47) on apoptotic cells [77]. These results suggest a close relationship between the phenotypic polarization state of microglia and their efferocytosis activity. We propose that Cel synergistically enhances mitochondrial function through multiple mechanisms, thereby indirectly promoting efferocytosis by microglia. First, it improves mitochondrial function, promotes autophagy, effectively removes damaged mitochondria, reduces the abnormal leakage of mtDNA, and subsequently inhibits overactivation of the cGAS–Sting pathway. The downregulation of Sting signalling inhibits microglial cell polarization towards the M1 type and significantly reduces proinflammatory factor expression, which alleviates the inhibition of efferocytosis by the inflammatory microenvironment. In response to the reduction in inflammation, the expression of key receptors for efferocytosis, such as MerTK and AXL, are further upregulated in microglia, enhancing their recognition and phagocytosis of apoptotic neuronal cells. Moreover, owing to improved mitochondrial function, the intracellular ATP supply is sufficient to support cytoskeletal reorganization and phagosome formation during efferocytosis. By promoting autophagy and efferocytosis, this strategy helps to eliminate apoptotic neurons in a timely manner, inhibit inflammatory factor bursts, remodel the injured microenvironment, and promote the reconstruction and repair of neural circuits. This strategy can rebuild the benign interaction network between microglia and neurons and promote functional cross-talk between them in the injured environment. Compared with the free drug, the nanodelivery system significantly enhanced the therapeutic efficacy of LDH. This is partly due to the improved solubility and targeting of the drug by the carrier and partly due to the neuroprosthetic potential of LDH itself. Various carrier systems, such as melatonin-loaded hydrogel scaffolds and growth factor-coated gelatine scaffolds, have been previously applied in therapeutic studies of spinal cord injury [78, 79]. However, most of these scaffold designs focus on drug delivery and inflammation inhibition, and the direct promotion of key regenerative processes, such as axonal regeneration and neural loop reconstruction, is still insufficient. In recent years, 3D printing technology has been used to construct biomimetic spinal cord scaffolds loaded with neural stem cells and growth factors to guide orderly axonal regeneration [80], but such technology is usually dependent on precision instruments and complicated preparation processes and still has limitations in anti-inflammatory modulation. In contrast, layered double hydroxide (LDH), as a multifunctional nanocarrier, combines the advantages of the above systems: LDH is not only biocompatible and degradable but also has an interlayer structure that allows for efficient loading and pH-responsive release of drugs to precisely match the acidic microenvironment of localized spinal cord injury (pH ≈ 6.5–6.8). In addition, Mg²⁺ ions released from LDH itself have been shown to exert multiple neuroprotective effects. Specifically, our supplemental experiments demonstrated that Mg²⁺ directly promotes neuronal survival and inhibits apoptosis, as evidenced by increased CCK-8 cell viability and decreased TUNEL-positive cells in H₂O₂-stimulated HT22 neurons. Furthermore, Mg²⁺ was found to promote microglial polarization from the pro-inflammatory M1 phenotype toward the anti-inflammatory M2 phenotype, as indicated by reduced CD86 expression and increased CD206 expression, and to enhance efferocytosis function by upregulating CD36 expression. These findings suggest that Mg²⁺ contributes to the therapeutic efficacy of the LDH carrier by directly supporting neuronal survival and modulating microglial function. In the present study, combining LDH with Cel not only improved drug delivery efficiency through nanocarriers but also promoted synergistic multiple mechanisms of carrier therapy and drug function: LDH provides physical drug delivery, microenvironmental modulation, and ionic therapy, whereas Cel protects mitochondria, inhibits apoptosis, and modulates the immune response. This strategy provides a new approach to the development of therapeutic systems for spinal cord injury that combine precise delivery, neuroregeneration and immunomodulation. However, some limitations remain. First, although this study demonstrated the overall reparative effect of Cel-LDH, the specific molecular mechanisms through which the composite nanoparticles enhance autophagy and efferocytosis have not been fully elucidated. Future studies using cell-specific knockout models could clarify the role of each pathway in neuroprotection. Second, although we observed improved recovery of motor function in mice after 28 days, to what extent Mg²⁺ contributes specifically to recovery is unclear and requires additional controlled studies. Finally, translational challenges, including the scaled-up production and storage stability of nanocomposites, need to be addressed before clinical application. Conclusion In conclusion, this study successfully loaded celastrol into LDH and constructed Cel-LDH composite nanoparticles. These multifunctional nanoparticles facilitate SCI recovery by concurrently normalizing efferocytosis homeostasis and inhibiting neuronal apoptosis, thus ultimately protecting neuronal survival and promoting axonal regeneration. This study provides novel insights into spinal cord injury repair strategies and highlights the significant translational potential of Cel-LDH for clinical applications. Experimental section Synthesis of Cel-LDH MgAl-LDH was obtained from Xianfeng Nanomaterials Technology Co., Ltd. (Jiangsu, China). To load Cel, a suspension was prepared by dispersing 10 mg of the obtained MgAl-LDH powder in 4 mL of deionized water under vigorous stirring. To this suspension, 1 mL of a Cel solution dissolved in methanol (concentration: 1 mg mL⁻¹) was added dropwise at ambient temperature. The reaction proceeded for 24 h with constant stirring to ensure complete interaction. The resulting suspension was then centrifuged (12000 rpm, 10 min), and the pellet was collected and extensively washed with deionized water (three times). The purified solid was freeze-dried overnight to obtain the final Cel-LDH nanocomposite. Characterization of Cel-LDH Several microscopic and spectroscopic methods were used to methodically analyze LDH and Cel-LDH. The morphological characteristics of the nanoparticles were examined by TEM (Tecnai G2 spirit, FEI, USA). The elemental distribution in LDH was determined by using SEM (GeminiSEM 360, ZEISS, DE). The size distribution and zeta potential were determined through DLS analysis using a Zetasizer Nano ZS90 (Malvern, UK). The chemical composition of the nanoparticles was determined using FT-IR (IRTracer-100, SHIMADZU, JPN). The Cel concentration was determined by plotting a standard curve with a UV–Vis spectrophotometer (UV-1600, CHN). Drug release in vitro Cel-LDH was added to 1 mL of PBS and placed in a constant-temperature shaker (37 °C, 120 rpm). At specified time intervals, 200 µL of release medium was taken and replaced with an equal volume of fresh medium, keeping the volume constant. The absorbance of the samples collected at different time points was measured by a UV spectrophotometer, and the cumulative release rate of Cel from Cel-LDH was calculated using the following formula: where At denotes the mass of Cel released from the start to time point t and Atotal denotes the total mass of Cel. Animal and ethical issues Eight-week-old female C57BL/6J mice weighing 20–22 g were purchased from Wenzhou Zhenghe Biotechnology Co., Ltd. The mice were then housed in an SPF-grade environment, with room temperature maintained at 20–25 °C and a 12-hour light/dark cycle. The experimental procedures adhered rigorously to the Guide for the Care and Use of Laboratory Animals as issued by the China National Institutes of Health and received approval from the Institutional Animal Care and Use Committee of Wenzhou Medical University (Approval No. XMSQ-2024-062). The experimental animals were randomly divided into the spinal cord injury group, LDH group, Cel group, and Cel-LDH group. Spinal cord injury model Preoperative weighing was performed, followed by intraperitoneal injection of 1% pentobarbital sodium at a dose of 50 mg/kg for anaesthesia. After anaesthesia was induced, a laminectomy was performed on vertebra T9 to expose the spinal cord. Then, the mouse with the exposed spinal cord was placed on the spinal cord impact device and an impact using a 2.5 mm diameter impact head was applied. The specific settings for the impact device are as follows: an impact velocity of 4 m/s, an impact depth of 0.8 mm, and an impact device dwell time of 300 m/s. Immediately following the procedure, LDH (30 mg mL− 1), Cel (10 mg mL− 1), and Cel-LDH (30 mg mL− 1) in normal saline (total volume of 10 µL) was injected in situ at the injury site (Figure S6). The mice in the untreated SCI group received 10 µL of normal saline. The wounds were subsequently disinfected with iodophor and closed layer by layer with absorbable sutures, after which the mice were placed on an electric blanket at 37 °C until awakening. They underwent manual emptying of their bladders twice daily until they were euthanized. Behavioural assessment Basso Mouse Scale (BMS) scores were recorded on days 0, 3, 7, 14, 21, and 28 after surgery. On day 28, hind limb motor function recovery was evaluated using footpad printing. All behavioral assessments were performed by two independent investigators who were blinded to the experimental group assignments. The blinding code was not broken until all data had been collected and analyzed. Inter-rater reliability was assessed using intraclass correlation coefficient (ICC > 0.9). Histopathology In accordance with previously described methods, the obtained spinal cord tissue was fixed in 4% paraformaldehyde for more than 48 h, embedded in paraffin, and sectioned using a paraffin microtome. After mounting the sections on slides, they were dewaxed using xylene, dried using a series of graded ethanol solutions, stained using different dye kits, and then subjected to histomorphological analysis. Haematoxylin and eosin (HE) staining was used to elucidate the basic cellular and tissue structures in the spinal cord sections. Nissl staining was used to evaluated the morphology and quantity of neurons, as well as to identify the nervous system impairment. Haemolysis assay Blood was taken from the orbits of each group of experimental mice using heparin-coated test tubes. The blood was centrifuged at 3000 rpm for 15 min at 4 °C, after which the plasma and platelet layers were carefully removed. The collected erythrocytes were washed three times with sterile phosphate-buffered saline (PBS, pH 7.4) and resuspended in PBS to prepare a 2% (v/v) erythrocyte suspension. Each sample was incubated with the erythrocyte suspension for 2 h at 37 °C with gentle shaking. After incubation, the samples were centrifuged at 3000 rpm for 10 min, and the absorbance of the supernatant was measured using an enzyme marker at 540 nm. PBS and Triton X-100 (1%, v/v) were used as negative and positive controls, respectively. The haemolysis rate was calculated according to the following formula: where Asample, APBS and ATriton X−100 represent the absorbances of the sample, PBS, and Triton X-100 groups, respectively. All experiments were performed in triplicate, and the results are presented as the means ± standard deviations (SDs). Cell culture HT22 cells were procured from the Cell Bank of the Type Culture Collection at the Shanghai Institute of Cell Biology, Chinese Academy of Sciences. These cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Gibco) supplemented with 10% foetal bovine serum (FBS; VivaCell) and 1% penicillin‒streptomycin solution (Solarbio) at 37 °C in a constant-temperature incubator containing 5% CO₂. BV2 cells were obtained from FuHeng Biology Co., Ltd. (Shanghai, China) and were cultured in MEM (Solarbio) supplemented with 10% FBS and 1% penicillin‒streptomycin solution at 37 °C in a constant-temperature incubator containing 5% CO₂. To simulate oxidative stress, the HT22 cells were treated with TBHP (Macklin) solution. Given that TBHP serves as a stable alternative to H₂O₂ for inducing oxidative stress, this experimental group is designated as the “H₂O₂ group” groupthroughout the manuscript for consistency with conventional terminology. The experimental groups included the control, H₂O₂, H₂O₂ + LDH, H₂O₂ + Cel, and H₂O₂ + Cel-LDH groups. BV2 cells were stimulated with 1 µg/mL LPS (Sigma) for 24 h to induce inflammation. The experimental groups included the control, LPS, LPS + LDH, LPS + Cel, and LPS + Cel-LDH groups. Calcein-AM/PI double staining assays A Calcein-AM/PI Cell Viability and Cytotoxicity Assay Kit (Engineering For Life, EFL-CLD-001) was used to simultaneously identify live and dead cells. After 24 h of treatment, the cells in the different groups were washed once with PBS and then stained using a suitable volume of calcein AM/PI detection working solution at 37 °C in the dark for 30 min. The staining effect was observed using an inverted fluorescence microscope (Leica, DM IL LED). Cell proliferation assay Cell viability was assessed using a CCK-8 cell viability and toxicity assay (Beyotime, C0037). Cells were seeded at a density of 5 × 103 cells/well in a 96-well plate and treated with different concentrations of Cel-LDH (25–1000 nM) for 24 h. The plates were then incubated in the dark with CCK-8 solution for 1 h. The absorbance was measured at 450 nm using a multimode microplate reader. Intracellular ROS scavenging To evaluate the ability of Cel-LDH to scavenge ROS in vitro, intracellular ROS levels were detected with DCFH-DA (Beyotime, S0033S) and DHE (Beyotime, S0063). HT22 cells were seeded in confocal dishes, and cell treatment was performed when the cell density reached 70%. The treated cells were washed three times with PBS and then incubated with DCFH-DA (10 µM) or DHE (1 µM) for 30 min at 37 °C in the dark. The cells were then washed three times with PBS, stained with Hoechst 33,342, and imaged using a laser confocal microscope (Leica, STELLARIS5). Mitochondrial function assay HT22 cells were seeded in confocal culture dishes and treated with the corresponding drugs. After the cells were incubated with JC-1 (5 µM) (Beyotime, C2003) and MitoSOX (5 µM) (Beyotime, S0061) for 30 min at 37 °C in the dark, the mitochondrial membrane potential and mitochondrial reactive oxygen species (ROS) production levels were assessed. After the cells were washed with fresh sterile PBS, images were immediately captured using a laser confocal microscope. An Enhanced ATP Detection Kit (Beyotime, S0027) was used to measure ATP levels in HT22 cells. HT22 cells were seeded into culture dishes (5 × 105 cells/dish), and drug treatment was performed when the cell density reached 80%. The cell samples were collected, and 300 µL of ATP lysis buffer was added to lyse the cells. The cells were harvested with a cell scraper and transferred to a 1.5 mL EP tube. The sample was subsequently centrifuged at 12,000 g at 4 °C for 5 minutes, after which the supernatant was transferred to a new EP tube. The RLU values were measured using a multimode microplate reader, and the ATP concentration was calculated on the basis of the ATP standard curve. Phagocytosis of apoptotic cells HT22 cells were induced to undergo apoptosis using TBHP, and apoptotic cells were labelled with Dil cell membrane dye (Beyotime, C1036). BV2 cells pretreated under different conditions were cocultured with Dil-labelled apoptotic cells for 12 h, fixed with 4% paraformaldehyde, washed, and blocked with 5% BSA at 37 °C for 1 h. They were then incubated with the microglia-specific marker Iba-1 at 37 °C for 2 h. After three washes with PBST, the cells were incubated with the corresponding species-specific secondary antibody at 37 °C in a shaking incubator in the dark for 1 h. After another wash, the cells were mounted with anti-fluorescence quenching mounting medium containing DAPI and imaged using a laser confocal microscope. Western blotting analysis Proteins in tissue extracts were detected using the Coomassie Brilliant Blue method. The proteins were separated by 12% SDS‒PAGE and transferred to a PVDF membrane. The membrane was blocked with 5% nonfat milk at room temperature for 2 h and incubated with the appropriate primary antibodies against CD36 (Affinity, DF13262), Beclin 1 (Proteintech, 11306-1-AP), p62/SQSTM1 (Proteintech, 18420-1-AP), IRF-3 (Cell Signaling Technology, #11904), p-IRF-3 (Cell Signaling Technology, #29047), TBK1 (Cell Signaling Technology, #3504), p-TBK1 (Cell Signaling Technology, #5483), Sting (Cell Signaling Technology, #13647), cGAS (Cell Signaling Technology, #15102), Bcl-2 (Proteintech, 26593-1-AP), Bax (Proteintech, 50599-2-Ig), Cyt c (Proteintech, 10993-1-AP), Cleaved-caspase3 (Abcam, ab2302), GAPDH (Zenbio, 380626), and β-actin (Zenbio, 380624) overnight. The corresponding secondary antibody (Zenbio, 511203) was subsequently incubated with the membranes at room temperature for 1.5 h. The protein bands were visualized using a highly sensitive ECL chemiluminescence detection kit (Zenbio, 17047) and analysed using a chemiluminescence imaging system (JS-M6). qRT‒PCR Total RNA was extracted from tissue using TRIzol, and cDNA was prepared from RNA using the RevertAid First Strand cDNA Synthesis Kit. qRT‒PCR was performed on a LightCycler96 instrument using Hieff qPCR SYBR Green Master Mix (without Rox Plus). GAPDH was used for relative expression normalization, and the 2-ΔΔCt method was employed. The primer sequences were designed using the NCBI Primer-BLAST tool, and the details are listed in Table S1. Immunofluorescence Cells were cultured on 20 × 20 mm cell monolayers in 12-well plates. After drug treatment, 4% paraformaldehyde was added to each well, and the cells were fixed at room temperature for 30 min. After fixation, the cells were washed three times with PBS. With the use of forceps and a specialized needle, the cell monolayer was transferred to a glass slide and secured with nail polish. After fixation, the monolayer was transferred to a humidified chamber and incubated at 37 °C in a shaking incubator with 5% BSA for 1 h. The cells were subsequently incubated with specific primary antibodies against Ki67 (Proteintech, 28074-1-AP), ATPB (Proteintech, 17247-1-AP), CD36 (Affinity, DF13262), LC3 (Proteintech, 14600-1-AP), p62/SQSTM1 (Proteintech, 18420-1-AP), CD86 (Abcam, ab239075), CD206 (Abcam, ab64693), iNOS (Proteintech, 22226-1-AP), Arg-1 (Proteintech, 16001-1-AP), Sting (Cell Signaling Technology, #13647), cGAS (Cell Signaling Technology, #15102), Bcl-2 (Proteintech, 26593-1-AP), Iba1 (Abcam, ab289874), Tuj-1 (Abcam, ab18207), GFAP (Proteintech, 16825-1-AP), MAP-2 (Proteintech, 17490-1-AP), NF-200 (Abcam, ab4680), and MBP (Proteintech, 10458-1-AP) overnight at 4 °C. After being washed three times with PBST, the corresponding secondary antibody was incubated with the slides at 37 °C in a shaking incubator in the dark for 1 h, discarded, and the samples were washed again with PBST in the dark, after which they were sealed under a coverslip with anti-fluorescence quenching mounting medium containing DAPI. Images were acquired using a laser confocal microscope. The tissue sections were subjected to antigen retrieval after dewaxing and rehydration. After the samples were cooled to room temperature, they were washed three times with PBS. A histochemistry pen was used to draw a circle around the tissue, after which the tissue was covered with 5% BSA and incubated at 37 °C in a constant-temperature shaking incubator for 1 h. The subsequent procedures for immunofluorescence were as described above. TUNEL assay Apoptosis was detected using the TUNEL method. The cell samples were permeabilized with 0.5% Triton X-100 for five minutes after being fixed with 4% paraformaldehyde (Solarbio, P1110) for thirty minutes. After dewaxing, the tissue slices were treated with 20 µg/ml DNase-free proteinase K for 30 min at 37 °C. Three PBS washes were then performed. After that, the sections were incubated for one hour at 37 °C in the dark using TUNEL detection solution (Beyotime, C1088). The sections were mounted using anti-fluorescence quenching mounting media and examined under a laser confocal microscope following three PBS washes. Statistical analysis The data are expressed as the means ± standard deviations. The statistical significance of differences between groups were analysed with one-way ANOVA via GraphPad Prism software, and a p value < 0.05 was considered to indicate statistical significance. Supplementary information Supplementary material 1. Author contributions Nuo Chen, Lufeng Yao and Le Zou contributed equally to this work. Nuo Chen wrote the manuscript with input from all authors. Nuo Chen, Lufeng Yao and Le Zou completed the in vitro and in vivo experiments. Qiuchen Wang and Yaning Li assisted in revising the manuscript and associated analysis. Junjun Luo and Mengting Zhan performed some experiments. Lu Ge and Liangliang Yang conceived the idea and provided advice for the experiments. Kailiang Zhou, Jian Xiao, Songlin Tong, Hongyu Zhang, Liangliang Yang and Lu Ge supervised the work. All authors read and approved the final manuscript. Funding This study was supported by the Natural Science Foundation of Zhejiang Province (No. LQN25H060010 to Lu Ge), National Natural Science Foundation of China (No. 82503074 to Lu Ge), Ningbo Natural Science Foundation Project (No. 2023J255 to Lu Ge, No. 2025J125 to Lufeng Yao, No. 2025J008 to Hongyu Zhang), Zhejiang Medical Science & Technology Program (No. 2023KY1151 to Lufeng Yao), Ningbo Yinzhou Science & Technology Program (No. 2024AS037 to Lufeng Yao). Data availability No datasets were generated or analysed during the current study. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Nuo Chen, Lufeng Yao and Le Zou contributed equally. Contributor Information Songlin Tong, Email: 13806646388@163.com. Hongyu Zhang, Email: st_hyz@126.com. Liangliang Yang, Email: liangliangyangLLL@126.com. Lu Ge, Email: gelu100466@126.com. References Associated Data Supplementary Materials Supplementary material 1. Data Availability Statement No datasets were generated or analysed during the current study.